From 5015b49448cbe5352b9cc232333ab26f3e608a07 Mon Sep 17 00:00:00 2001 From: Stephen Rothwell Date: Mon, 31 Oct 2005 18:39:20 +1100 Subject: powerpc: fix __strnlen_user in merge tree Change USER/KERNEL_DS so that the merged version of __strnlen_user can be used which allows us to complete the removal of arch/ppc64/lib/. Signed-off-by: Stephen Rothwell --- arch/ppc64/Makefile | 1 - arch/ppc64/lib/Makefile | 5 -- arch/ppc64/lib/string.S | 179 ------------------------------------------------ 3 files changed, 185 deletions(-) delete mode 100644 arch/ppc64/lib/Makefile delete mode 100644 arch/ppc64/lib/string.S (limited to 'arch/ppc64') diff --git a/arch/ppc64/Makefile b/arch/ppc64/Makefile index fdbd6f44adc0..a55a82d145d4 100644 --- a/arch/ppc64/Makefile +++ b/arch/ppc64/Makefile @@ -86,7 +86,6 @@ head-y := arch/ppc64/kernel/head.o head-y += arch/powerpc/kernel/fpu.o head-y += arch/powerpc/kernel/entry_64.o -libs-y += arch/ppc64/lib/ core-y += arch/ppc64/kernel/ arch/powerpc/kernel/ core-y += arch/powerpc/mm/ core-y += arch/powerpc/sysdev/ diff --git a/arch/ppc64/lib/Makefile b/arch/ppc64/lib/Makefile deleted file mode 100644 index 42d5295bf345..000000000000 --- a/arch/ppc64/lib/Makefile +++ /dev/null @@ -1,5 +0,0 @@ -# -# Makefile for ppc64-specific library files.. -# - -lib-y := string.o diff --git a/arch/ppc64/lib/string.S b/arch/ppc64/lib/string.S deleted file mode 100644 index e21a0038a4d6..000000000000 --- a/arch/ppc64/lib/string.S +++ /dev/null @@ -1,179 +0,0 @@ -/* - * String handling functions for PowerPC. - * - * Copyright (C) 1996 Paul Mackerras. - * - * This program is free software; you can redistribute it and/or - * modify it under the terms of the GNU General Public License - * as published by the Free Software Foundation; either version - * 2 of the License, or (at your option) any later version. - */ -#include -#include -#include - -_GLOBAL(strcpy) - addi r5,r3,-1 - addi r4,r4,-1 -1: lbzu r0,1(r4) - cmpwi 0,r0,0 - stbu r0,1(r5) - bne 1b - blr - -_GLOBAL(strncpy) - cmpwi 0,r5,0 - beqlr - mtctr r5 - addi r6,r3,-1 - addi r4,r4,-1 -1: lbzu r0,1(r4) - cmpwi 0,r0,0 - stbu r0,1(r6) - bdnzf 2,1b /* dec ctr, branch if ctr != 0 && !cr0.eq */ - blr - -_GLOBAL(strcat) - addi r5,r3,-1 - addi r4,r4,-1 -1: lbzu r0,1(r5) - cmpwi 0,r0,0 - bne 1b - addi r5,r5,-1 -1: lbzu r0,1(r4) - cmpwi 0,r0,0 - stbu r0,1(r5) - bne 1b - blr - -_GLOBAL(strcmp) - addi r5,r3,-1 - addi r4,r4,-1 -1: lbzu r3,1(r5) - cmpwi 1,r3,0 - lbzu r0,1(r4) - subf. r3,r0,r3 - beqlr 1 - beq 1b - blr - -_GLOBAL(strlen) - addi r4,r3,-1 -1: lbzu r0,1(r4) - cmpwi 0,r0,0 - bne 1b - subf r3,r3,r4 - blr - -_GLOBAL(memcmp) - cmpwi 0,r5,0 - ble- 2f - mtctr r5 - addi r6,r3,-1 - addi r4,r4,-1 -1: lbzu r3,1(r6) - lbzu r0,1(r4) - subf. r3,r0,r3 - bdnzt 2,1b - blr -2: li r3,0 - blr - -_GLOBAL(memchr) - cmpwi 0,r5,0 - ble- 2f - mtctr r5 - addi r3,r3,-1 -1: lbzu r0,1(r3) - cmpw 0,r0,r4 - bdnzf 2,1b - beqlr -2: li r3,0 - blr - -_GLOBAL(__clear_user) - addi r6,r3,-4 - li r3,0 - li r5,0 - cmplwi 0,r4,4 - blt 7f - /* clear a single word */ -11: stwu r5,4(r6) - beqlr - /* clear word sized chunks */ - andi. r0,r6,3 - add r4,r0,r4 - subf r6,r0,r6 - srwi r0,r4,2 - andi. r4,r4,3 - mtctr r0 - bdz 7f -1: stwu r5,4(r6) - bdnz 1b - /* clear byte sized chunks */ -7: cmpwi 0,r4,0 - beqlr - mtctr r4 - addi r6,r6,3 -8: stbu r5,1(r6) - bdnz 8b - blr -90: mr r3,r4 - blr -91: mfctr r3 - slwi r3,r3,2 - add r3,r3,r4 - blr -92: mfctr r3 - blr - - .section __ex_table,"a" - .align 3 - .llong 11b,90b - .llong 1b,91b - .llong 8b,92b - .text - -/* r3 = dst, r4 = src, r5 = count */ -_GLOBAL(__strncpy_from_user) - addi r6,r3,-1 - addi r4,r4,-1 - cmpwi 0,r5,0 - beq 2f - mtctr r5 -1: lbzu r0,1(r4) - cmpwi 0,r0,0 - stbu r0,1(r6) - bdnzf 2,1b /* dec ctr, branch if ctr != 0 && !cr0.eq */ - beq 3f -2: addi r6,r6,1 -3: subf r3,r3,r6 - blr -99: li r3,-EFAULT - blr - - .section __ex_table,"a" - .align 3 - .llong 1b,99b - .text - -/* r3 = str, r4 = len (> 0) */ -_GLOBAL(__strnlen_user) - addi r7,r3,-1 - mtctr r4 /* ctr = len */ -1: lbzu r0,1(r7) /* get next byte */ - cmpwi 0,r0,0 - bdnzf 2,1b /* loop if --ctr != 0 && byte != 0 */ - addi r7,r7,1 - subf r3,r3,r7 /* number of bytes we have looked at */ - beqlr /* return if we found a 0 byte */ - cmpw 0,r3,r4 /* did we look at all len bytes? */ - blt 99f /* if not, must have hit top */ - addi r3,r4,1 /* return len + 1 to indicate no null found */ - blr -99: li r3,0 /* bad address, return 0 */ - blr - - .section __ex_table,"a" - .align 3 - .llong 1b,99b -- cgit v1.2.3 From f3f66f599db131ea57dc567ffd931d269dbc690e Mon Sep 17 00:00:00 2001 From: Arnd Bergmann Date: Mon, 31 Oct 2005 20:08:37 -0500 Subject: [PATCH] powerpc: Rename BPA to Cell The official name for BPA is now CBEA (Cell Broadband Engine Architecture). This patch renames all occurences of the term BPA to 'Cell' for easier recognition. Signed-off-by: Arnd Bergmann Signed-off-by: Paul Mackerras --- arch/powerpc/Kconfig | 8 +++--- arch/powerpc/kernel/prom_init.c | 2 +- arch/powerpc/kernel/setup_64.c | 6 ++--- arch/powerpc/kernel/traps.c | 4 +-- arch/ppc64/kernel/Makefile | 2 +- arch/ppc64/kernel/bpa_iic.c | 8 +++--- arch/ppc64/kernel/bpa_iic.h | 8 +++--- arch/ppc64/kernel/bpa_iommu.c | 44 +++++++++++++++--------------- arch/ppc64/kernel/bpa_iommu.h | 10 +++---- arch/ppc64/kernel/bpa_nvram.c | 60 ++++++++++++++++++++--------------------- arch/ppc64/kernel/bpa_setup.c | 46 +++++++++++++++---------------- arch/ppc64/kernel/irq.c | 2 +- arch/ppc64/kernel/proc_ppc64.c | 2 +- arch/ppc64/kernel/prom_init.c | 4 +-- arch/ppc64/kernel/spider-pic.c | 2 +- include/asm-powerpc/irq.h | 2 +- include/asm-powerpc/processor.h | 2 +- include/asm-ppc64/nvram.h | 2 +- 18 files changed, 107 insertions(+), 107 deletions(-) (limited to 'arch/ppc64') diff --git a/arch/powerpc/Kconfig b/arch/powerpc/Kconfig index 967ecf92d6a7..0bd315de156b 100644 --- a/arch/powerpc/Kconfig +++ b/arch/powerpc/Kconfig @@ -319,8 +319,8 @@ config PPC_MAPLE This option enables support for the Maple 970FX Evaluation Board. For more informations, refer to -config PPC_BPA - bool " Broadband Processor Architecture" +config PPC_CELL + bool " Cell Broadband Processor Architecture" depends on PPC_MULTIPLATFORM && PPC64 select PPC_RTAS @@ -358,8 +358,8 @@ config MPIC_BROKEN_U3 depends on PPC_MAPLE default y -config BPA_IIC - depends on PPC_BPA +config CELL_IIC + depends on PPC_CELL bool default y diff --git a/arch/powerpc/kernel/prom_init.c b/arch/powerpc/kernel/prom_init.c index 9750b3cd8ecd..c758b6624d7b 100644 --- a/arch/powerpc/kernel/prom_init.c +++ b/arch/powerpc/kernel/prom_init.c @@ -2000,7 +2000,7 @@ unsigned long __init prom_init(unsigned long r3, unsigned long r4, #endif /* - * On pSeries and BPA, copy the CPU hold code + * Copy the CPU hold code */ if (RELOC(of_platform) != PLATFORM_POWERMAC) copy_and_flush(0, KERNELBASE + offset, 0x100, 0); diff --git a/arch/powerpc/kernel/setup_64.c b/arch/powerpc/kernel/setup_64.c index 19530ce9cd27..70ead7d0d12e 100644 --- a/arch/powerpc/kernel/setup_64.c +++ b/arch/powerpc/kernel/setup_64.c @@ -296,7 +296,7 @@ static void __init setup_cpu_maps(void) extern struct machdep_calls pSeries_md; extern struct machdep_calls pmac_md; extern struct machdep_calls maple_md; -extern struct machdep_calls bpa_md; +extern struct machdep_calls cell_md; extern struct machdep_calls iseries_md; /* Ultimately, stuff them in an elf section like initcalls... */ @@ -310,8 +310,8 @@ static struct machdep_calls __initdata *machines[] = { #ifdef CONFIG_PPC_MAPLE &maple_md, #endif /* CONFIG_PPC_MAPLE */ -#ifdef CONFIG_PPC_BPA - &bpa_md, +#ifdef CONFIG_PPC_CELL + &cell_md, #endif #ifdef CONFIG_PPC_ISERIES &iseries_md, diff --git a/arch/powerpc/kernel/traps.c b/arch/powerpc/kernel/traps.c index 5d638ecddbd0..bb0d00284a7e 100644 --- a/arch/powerpc/kernel/traps.c +++ b/arch/powerpc/kernel/traps.c @@ -147,8 +147,8 @@ int die(const char *str, struct pt_regs *regs, long err) printk("POWERMAC "); nl = 1; break; - case PLATFORM_BPA: - printk("BPA "); + case PLATFORM_CELL: + printk("CELL "); nl = 1; break; } diff --git a/arch/ppc64/kernel/Makefile b/arch/ppc64/kernel/Makefile index 327c08ce4291..a01a9de9eb56 100644 --- a/arch/ppc64/kernel/Makefile +++ b/arch/ppc64/kernel/Makefile @@ -31,7 +31,7 @@ endif obj-$(CONFIG_PPC_PSERIES) += rtasd.o udbg_16550.o -obj-$(CONFIG_PPC_BPA) += bpa_setup.o bpa_iommu.o bpa_nvram.o \ +obj-$(CONFIG_PPC_CELL) += bpa_setup.o bpa_iommu.o bpa_nvram.o \ bpa_iic.o spider-pic.o obj-$(CONFIG_KEXEC) += machine_kexec.o diff --git a/arch/ppc64/kernel/bpa_iic.c b/arch/ppc64/kernel/bpa_iic.c index 0aaa878e19d3..7fbe78a9327d 100644 --- a/arch/ppc64/kernel/bpa_iic.c +++ b/arch/ppc64/kernel/bpa_iic.c @@ -1,5 +1,5 @@ /* - * BPA Internal Interrupt Controller + * Cell Internal Interrupt Controller * * (C) Copyright IBM Deutschland Entwicklung GmbH 2005 * @@ -31,7 +31,7 @@ #include #include -#include "bpa_iic.h" +#include "interrupt.h" struct iic_pending_bits { u32 data; @@ -89,7 +89,7 @@ static void iic_end(unsigned int irq) } static struct hw_interrupt_type iic_pic = { - .typename = " BPA-IIC ", + .typename = " CELL-IIC ", .startup = iic_startup, .enable = iic_enable, .disable = iic_disable, @@ -106,7 +106,7 @@ static int iic_external_get_irq(struct iic_pending_bits pending) irq = -1; /* - * This mapping is specific to the Broadband + * This mapping is specific to the Cell Broadband * Engine. We might need to get the numbers * from the device tree to support future CPUs. */ diff --git a/arch/ppc64/kernel/bpa_iic.h b/arch/ppc64/kernel/bpa_iic.h index 6833c3022166..37d58e6fd0c6 100644 --- a/arch/ppc64/kernel/bpa_iic.h +++ b/arch/ppc64/kernel/bpa_iic.h @@ -1,5 +1,5 @@ -#ifndef ASM_BPA_IIC_H -#define ASM_BPA_IIC_H +#ifndef ASM_CELL_PIC_H +#define ASM_CELL_PIC_H #ifdef __KERNEL__ /* * Mapping of IIC pending bits into per-node @@ -21,7 +21,7 @@ * + node number * * don't care * - * A node consists of a Broadband Engine and an optional + * A node consists of a Cell Broadband Engine and an optional * south bridge device providing a maximum of 64 IRQs. * The south bridge may be connected to either IOIF0 * or IOIF1. @@ -59,4 +59,4 @@ extern void spider_init_IRQ(void); extern int spider_get_irq(unsigned long int_pending); #endif -#endif /* ASM_BPA_IIC_H */ +#endif /* ASM_CELL_PIC_H */ diff --git a/arch/ppc64/kernel/bpa_iommu.c b/arch/ppc64/kernel/bpa_iommu.c index da1b4b7a3269..74f999b4ac9e 100644 --- a/arch/ppc64/kernel/bpa_iommu.c +++ b/arch/ppc64/kernel/bpa_iommu.c @@ -1,5 +1,5 @@ /* - * IOMMU implementation for Broadband Processor Architecture + * IOMMU implementation for Cell Broadband Processor Architecture * We just establish a linear mapping at boot by setting all the * IOPT cache entries in the CPU. * The mapping functions should be identical to pci_direct_iommu, @@ -41,7 +41,7 @@ #include #include -#include "bpa_iommu.h" +#include "iommu.h" static inline unsigned long get_iopt_entry(unsigned long real_address, unsigned long ioid, @@ -276,7 +276,7 @@ static void iommu_dev_setup_null(struct pci_dev *d) { } * for each DMA window used by any device. For now, we * happen to know that there is only one DMA window in use, * starting at iopt_phys_offset. */ -static void bpa_map_iommu(void) +static void cell_map_iommu(void) { unsigned long address; void __iomem *base; @@ -309,7 +309,7 @@ static void bpa_map_iommu(void) } -static void *bpa_alloc_coherent(struct device *hwdev, size_t size, +static void *cell_alloc_coherent(struct device *hwdev, size_t size, dma_addr_t *dma_handle, gfp_t flag) { void *ret; @@ -317,65 +317,65 @@ static void *bpa_alloc_coherent(struct device *hwdev, size_t size, ret = (void *)__get_free_pages(flag, get_order(size)); if (ret != NULL) { memset(ret, 0, size); - *dma_handle = virt_to_abs(ret) | BPA_DMA_VALID; + *dma_handle = virt_to_abs(ret) | CELL_DMA_VALID; } return ret; } -static void bpa_free_coherent(struct device *hwdev, size_t size, +static void cell_free_coherent(struct device *hwdev, size_t size, void *vaddr, dma_addr_t dma_handle) { free_pages((unsigned long)vaddr, get_order(size)); } -static dma_addr_t bpa_map_single(struct device *hwdev, void *ptr, +static dma_addr_t cell_map_single(struct device *hwdev, void *ptr, size_t size, enum dma_data_direction direction) { - return virt_to_abs(ptr) | BPA_DMA_VALID; + return virt_to_abs(ptr) | CELL_DMA_VALID; } -static void bpa_unmap_single(struct device *hwdev, dma_addr_t dma_addr, +static void cell_unmap_single(struct device *hwdev, dma_addr_t dma_addr, size_t size, enum dma_data_direction direction) { } -static int bpa_map_sg(struct device *hwdev, struct scatterlist *sg, +static int cell_map_sg(struct device *hwdev, struct scatterlist *sg, int nents, enum dma_data_direction direction) { int i; for (i = 0; i < nents; i++, sg++) { sg->dma_address = (page_to_phys(sg->page) + sg->offset) - | BPA_DMA_VALID; + | CELL_DMA_VALID; sg->dma_length = sg->length; } return nents; } -static void bpa_unmap_sg(struct device *hwdev, struct scatterlist *sg, +static void cell_unmap_sg(struct device *hwdev, struct scatterlist *sg, int nents, enum dma_data_direction direction) { } -static int bpa_dma_supported(struct device *dev, u64 mask) +static int cell_dma_supported(struct device *dev, u64 mask) { return mask < 0x100000000ull; } -void bpa_init_iommu(void) +void cell_init_iommu(void) { - bpa_map_iommu(); + cell_map_iommu(); /* Direct I/O, IOMMU off */ ppc_md.iommu_dev_setup = iommu_dev_setup_null; ppc_md.iommu_bus_setup = iommu_bus_setup_null; - pci_dma_ops.alloc_coherent = bpa_alloc_coherent; - pci_dma_ops.free_coherent = bpa_free_coherent; - pci_dma_ops.map_single = bpa_map_single; - pci_dma_ops.unmap_single = bpa_unmap_single; - pci_dma_ops.map_sg = bpa_map_sg; - pci_dma_ops.unmap_sg = bpa_unmap_sg; - pci_dma_ops.dma_supported = bpa_dma_supported; + pci_dma_ops.alloc_coherent = cell_alloc_coherent; + pci_dma_ops.free_coherent = cell_free_coherent; + pci_dma_ops.map_single = cell_map_single; + pci_dma_ops.unmap_single = cell_unmap_single; + pci_dma_ops.map_sg = cell_map_sg; + pci_dma_ops.unmap_sg = cell_unmap_sg; + pci_dma_ops.dma_supported = cell_dma_supported; } diff --git a/arch/ppc64/kernel/bpa_iommu.h b/arch/ppc64/kernel/bpa_iommu.h index e547d77dfa04..490d77abfe85 100644 --- a/arch/ppc64/kernel/bpa_iommu.h +++ b/arch/ppc64/kernel/bpa_iommu.h @@ -1,5 +1,5 @@ -#ifndef BPA_IOMMU_H -#define BPA_IOMMU_H +#ifndef CELL_IOMMU_H +#define CELL_IOMMU_H /* some constants */ enum { @@ -55,11 +55,11 @@ enum { /* The high bit needs to be set on every DMA address, only 2GB are addressable */ - BPA_DMA_VALID = 0x80000000, - BPA_DMA_MASK = 0x7fffffff, + CELL_DMA_VALID = 0x80000000, + CELL_DMA_MASK = 0x7fffffff, }; -void bpa_init_iommu(void); +void cell_init_iommu(void); #endif diff --git a/arch/ppc64/kernel/bpa_nvram.c b/arch/ppc64/kernel/bpa_nvram.c index 06a119cfceb5..74e0d31a3559 100644 --- a/arch/ppc64/kernel/bpa_nvram.c +++ b/arch/ppc64/kernel/bpa_nvram.c @@ -1,5 +1,5 @@ /* - * NVRAM for CPBW + * memory mapped NVRAM * * (C) Copyright IBM Corp. 2005 * @@ -30,54 +30,54 @@ #include #include -static void __iomem *bpa_nvram_start; -static long bpa_nvram_len; -static spinlock_t bpa_nvram_lock = SPIN_LOCK_UNLOCKED; +static void __iomem *mmio_nvram_start; +static long mmio_nvram_len; +static spinlock_t mmio_nvram_lock = SPIN_LOCK_UNLOCKED; -static ssize_t bpa_nvram_read(char *buf, size_t count, loff_t *index) +static ssize_t mmio_nvram_read(char *buf, size_t count, loff_t *index) { unsigned long flags; - if (*index >= bpa_nvram_len) + if (*index >= mmio_nvram_len) return 0; - if (*index + count > bpa_nvram_len) - count = bpa_nvram_len - *index; + if (*index + count > mmio_nvram_len) + count = mmio_nvram_len - *index; - spin_lock_irqsave(&bpa_nvram_lock, flags); + spin_lock_irqsave(&mmio_nvram_lock, flags); - memcpy_fromio(buf, bpa_nvram_start + *index, count); + memcpy_fromio(buf, mmio_nvram_start + *index, count); - spin_unlock_irqrestore(&bpa_nvram_lock, flags); + spin_unlock_irqrestore(&mmio_nvram_lock, flags); *index += count; return count; } -static ssize_t bpa_nvram_write(char *buf, size_t count, loff_t *index) +static ssize_t mmio_nvram_write(char *buf, size_t count, loff_t *index) { unsigned long flags; - if (*index >= bpa_nvram_len) + if (*index >= mmio_nvram_len) return 0; - if (*index + count > bpa_nvram_len) - count = bpa_nvram_len - *index; + if (*index + count > mmio_nvram_len) + count = mmio_nvram_len - *index; - spin_lock_irqsave(&bpa_nvram_lock, flags); + spin_lock_irqsave(&mmio_nvram_lock, flags); - memcpy_toio(bpa_nvram_start + *index, buf, count); + memcpy_toio(mmio_nvram_start + *index, buf, count); - spin_unlock_irqrestore(&bpa_nvram_lock, flags); + spin_unlock_irqrestore(&mmio_nvram_lock, flags); *index += count; return count; } -static ssize_t bpa_nvram_get_size(void) +static ssize_t mmio_nvram_get_size(void) { - return bpa_nvram_len; + return mmio_nvram_len; } -int __init bpa_nvram_init(void) +int __init mmio_nvram_init(void) { struct device_node *nvram_node; unsigned long *buffer; @@ -97,20 +97,20 @@ int __init bpa_nvram_init(void) ret = -ENODEV; nvram_addr = buffer[0]; - bpa_nvram_len = buffer[1]; - if ( (!bpa_nvram_len) || (!nvram_addr) ) + mmio_nvram_len = buffer[1]; + if ( (!mmio_nvram_len) || (!nvram_addr) ) goto out; - bpa_nvram_start = ioremap(nvram_addr, bpa_nvram_len); - if (!bpa_nvram_start) + mmio_nvram_start = ioremap(nvram_addr, mmio_nvram_len); + if (!mmio_nvram_start) goto out; - printk(KERN_INFO "BPA NVRAM, %luk mapped to %p\n", - bpa_nvram_len >> 10, bpa_nvram_start); + printk(KERN_INFO "mmio NVRAM, %luk mapped to %p\n", + mmio_nvram_len >> 10, mmio_nvram_start); - ppc_md.nvram_read = bpa_nvram_read; - ppc_md.nvram_write = bpa_nvram_write; - ppc_md.nvram_size = bpa_nvram_get_size; + ppc_md.nvram_read = mmio_nvram_read; + ppc_md.nvram_write = mmio_nvram_write; + ppc_md.nvram_size = mmio_nvram_get_size; out: of_node_put(nvram_node); diff --git a/arch/ppc64/kernel/bpa_setup.c b/arch/ppc64/kernel/bpa_setup.c index c2dc8f282eb8..9a495634d0c2 100644 --- a/arch/ppc64/kernel/bpa_setup.c +++ b/arch/ppc64/kernel/bpa_setup.c @@ -1,11 +1,11 @@ /* - * linux/arch/ppc/kernel/bpa_setup.c + * linux/arch/powerpc/platforms/cell/cell_setup.c * * Copyright (C) 1995 Linus Torvalds * Adapted from 'alpha' version by Gary Thomas * Modified by Cort Dougan (cort@cs.nmt.edu) * Modified by PPC64 Team, IBM Corp - * Modified by BPA Team, IBM Deutschland Entwicklung GmbH + * Modified by Cell Team, IBM Deutschland Entwicklung GmbH * * This program is free software; you can redistribute it and/or * modify it under the terms of the GNU General Public License @@ -46,8 +46,8 @@ #include #include -#include "bpa_iic.h" -#include "bpa_iommu.h" +#include "interrupt.h" +#include "iommu.h" #ifdef DEBUG #define DBG(fmt...) udbg_printf(fmt) @@ -55,7 +55,7 @@ #define DBG(fmt...) #endif -void bpa_show_cpuinfo(struct seq_file *m) +void cell_show_cpuinfo(struct seq_file *m) { struct device_node *root; const char *model = ""; @@ -63,22 +63,22 @@ void bpa_show_cpuinfo(struct seq_file *m) root = of_find_node_by_path("/"); if (root) model = get_property(root, "model", NULL); - seq_printf(m, "machine\t\t: BPA %s\n", model); + seq_printf(m, "machine\t\t: CHRP %s\n", model); of_node_put(root); } -static void bpa_progress(char *s, unsigned short hex) +static void cell_progress(char *s, unsigned short hex) { printk("*** %04x : %s\n", hex, s ? s : ""); } -static void __init bpa_setup_arch(void) +static void __init cell_setup_arch(void) { ppc_md.init_IRQ = iic_init_IRQ; ppc_md.get_irq = iic_get_irq; #ifdef CONFIG_SMP - smp_init_pSeries(); + smp_init_cell(); #endif /* init to some ~sane value until calibrate_delay() runs */ @@ -97,39 +97,39 @@ static void __init bpa_setup_arch(void) conswitchp = &dummy_con; #endif - bpa_nvram_init(); + mmio_nvram_init(); } /* * Early initialization. Relocation is on but do not reference unbolted pages */ -static void __init bpa_init_early(void) +static void __init cell_init_early(void) { - DBG(" -> bpa_init_early()\n"); + DBG(" -> cell_init_early()\n"); hpte_init_native(); - bpa_init_iommu(); + cell_init_iommu(); - ppc64_interrupt_controller = IC_BPA_IIC; + ppc64_interrupt_controller = IC_CELL_PIC; - DBG(" <- bpa_init_early()\n"); + DBG(" <- cell_init_early()\n"); } -static int __init bpa_probe(int platform) +static int __init cell_probe(int platform) { - if (platform != PLATFORM_BPA) + if (platform != PLATFORM_CELL) return 0; return 1; } -struct machdep_calls __initdata bpa_md = { - .probe = bpa_probe, - .setup_arch = bpa_setup_arch, - .init_early = bpa_init_early, - .show_cpuinfo = bpa_show_cpuinfo, +struct machdep_calls __initdata cell_md = { + .probe = cell_probe, + .setup_arch = cell_setup_arch, + .init_early = cell_init_early, + .show_cpuinfo = cell_show_cpuinfo, .restart = rtas_restart, .power_off = rtas_power_off, .halt = rtas_halt, @@ -137,5 +137,5 @@ struct machdep_calls __initdata bpa_md = { .get_rtc_time = rtas_get_rtc_time, .set_rtc_time = rtas_set_rtc_time, .calibrate_decr = generic_calibrate_decr, - .progress = bpa_progress, + .progress = cell_progress, }; diff --git a/arch/ppc64/kernel/irq.c b/arch/ppc64/kernel/irq.c index f41afe545045..b072ed6f77ad 100644 --- a/arch/ppc64/kernel/irq.c +++ b/arch/ppc64/kernel/irq.c @@ -392,7 +392,7 @@ int virt_irq_create_mapping(unsigned int real_irq) if (ppc64_interrupt_controller == IC_OPEN_PIC) return real_irq; /* no mapping for openpic (for now) */ - if (ppc64_interrupt_controller == IC_BPA_IIC) + if (ppc64_interrupt_controller == IC_CELL_PIC) return real_irq; /* no mapping for iic either */ /* don't map interrupts < MIN_VIRT_IRQ */ diff --git a/arch/ppc64/kernel/proc_ppc64.c b/arch/ppc64/kernel/proc_ppc64.c index a87c66a9652a..24e955ee9487 100644 --- a/arch/ppc64/kernel/proc_ppc64.c +++ b/arch/ppc64/kernel/proc_ppc64.c @@ -53,7 +53,7 @@ static int __init proc_ppc64_create(void) if (!root) return 1; - if (!(systemcfg->platform & (PLATFORM_PSERIES | PLATFORM_BPA))) + if (!(systemcfg->platform & (PLATFORM_PSERIES | PLATFORM_CELL))) return 0; if (!proc_mkdir("rtas", root)) diff --git a/arch/ppc64/kernel/prom_init.c b/arch/ppc64/kernel/prom_init.c index 69924ba4d7d9..a4bbca6dbb8b 100644 --- a/arch/ppc64/kernel/prom_init.c +++ b/arch/ppc64/kernel/prom_init.c @@ -1939,9 +1939,9 @@ unsigned long __init prom_init(unsigned long r3, unsigned long r4, unsigned long prom_send_capabilities(); /* - * On pSeries and BPA, copy the CPU hold code + * On pSeries and Cell, copy the CPU hold code */ - if (RELOC(of_platform) & (PLATFORM_PSERIES | PLATFORM_BPA)) + if (RELOC(of_platform) & (PLATFORM_PSERIES | PLATFORM_CELL)) copy_and_flush(0, KERNELBASE - offset, 0x100, 0); /* diff --git a/arch/ppc64/kernel/spider-pic.c b/arch/ppc64/kernel/spider-pic.c index d5c9a02fb119..e74132188bdf 100644 --- a/arch/ppc64/kernel/spider-pic.c +++ b/arch/ppc64/kernel/spider-pic.c @@ -27,7 +27,7 @@ #include #include -#include "bpa_iic.h" +#include "interrupt.h" /* register layout taken from Spider spec, table 7.4-4 */ enum { diff --git a/include/asm-powerpc/irq.h b/include/asm-powerpc/irq.h index c7c3f912a3c2..b3935ea28fff 100644 --- a/include/asm-powerpc/irq.h +++ b/include/asm-powerpc/irq.h @@ -73,7 +73,7 @@ extern unsigned int real_irq_to_virt_slowpath(unsigned int real_irq); #define IC_INVALID 0 #define IC_OPEN_PIC 1 #define IC_PPC_XIC 2 -#define IC_BPA_IIC 3 +#define IC_CELL_PIC 3 #define IC_ISERIES 4 extern u64 ppc64_interrupt_controller; diff --git a/include/asm-powerpc/processor.h b/include/asm-powerpc/processor.h index eee954a001fd..1dc4bf7b52b3 100644 --- a/include/asm-powerpc/processor.h +++ b/include/asm-powerpc/processor.h @@ -70,7 +70,7 @@ extern unsigned char ucBoardRevMaj, ucBoardRevMin; #define PLATFORM_LPAR 0x0001 #define PLATFORM_POWERMAC 0x0400 #define PLATFORM_MAPLE 0x0500 -#define PLATFORM_BPA 0x1000 +#define PLATFORM_CELL 0x1000 /* Compatibility with drivers coming from PPC32 world */ #define _machine (systemcfg->platform) diff --git a/include/asm-ppc64/nvram.h b/include/asm-ppc64/nvram.h index dfaa21566c9a..def47d720d3d 100644 --- a/include/asm-ppc64/nvram.h +++ b/include/asm-ppc64/nvram.h @@ -70,7 +70,7 @@ extern struct nvram_partition *nvram_find_partition(int sig, const char *name); extern int pSeries_nvram_init(void); extern int pmac_nvram_init(void); -extern int bpa_nvram_init(void); +extern int mmio_nvram_init(void); /* PowerMac specific nvram stuffs */ -- cgit v1.2.3 From edf03c1ee84e5095e8f4a4b8ad6e0784118cb176 Mon Sep 17 00:00:00 2001 From: Arnd Bergmann Date: Mon, 31 Oct 2005 20:08:40 -0500 Subject: [PATCH] powerpc: move mmio_nvram.c over to arch/powerpc The nvram code formally known as bpa_nvram.c is rather generic really, so it is quite likely to be useful to future boards not based on cell. This patch puts it into arch/powerpc/sysdev. Signed-off-by: Arnd Bergmann Signed-off-by: Paul Mackerras --- arch/powerpc/Kconfig | 5 ++ arch/powerpc/sysdev/Makefile | 1 + arch/powerpc/sysdev/mmio_nvram.c | 118 +++++++++++++++++++++++++++++++++++++++ arch/ppc64/kernel/Makefile | 2 +- arch/ppc64/kernel/bpa_nvram.c | 118 --------------------------------------- 5 files changed, 125 insertions(+), 119 deletions(-) create mode 100644 arch/powerpc/sysdev/mmio_nvram.c delete mode 100644 arch/ppc64/kernel/bpa_nvram.c (limited to 'arch/ppc64') diff --git a/arch/powerpc/Kconfig b/arch/powerpc/Kconfig index aa13a9d95ea2..1c44a1dac421 100644 --- a/arch/powerpc/Kconfig +++ b/arch/powerpc/Kconfig @@ -325,6 +325,7 @@ config PPC_CELL depends on PPC_MULTIPLATFORM && PPC64 select PPC_RTAS select RTAS_FW + select MMIO_NVRAM config PPC_OF bool @@ -360,6 +361,10 @@ config RTAS_FW depends on PPC_RTAS default n +config MMIO_NVRAM + bool + default n + config MPIC_BROKEN_U3 bool depends on PPC_MAPLE diff --git a/arch/powerpc/sysdev/Makefile b/arch/powerpc/sysdev/Makefile index 8acd21dee05d..6b7efcfc352a 100644 --- a/arch/powerpc/sysdev/Makefile +++ b/arch/powerpc/sysdev/Makefile @@ -5,3 +5,4 @@ obj-$(CONFIG_PPC_MPC106) += grackle.o obj-$(CONFIG_BOOKE) += dcr.o obj-$(CONFIG_40x) += dcr.o obj-$(CONFIG_U3_DART) += u3_iommu.o +obj-$(CONFIG_MMIO_NVRAM) += mmio_nvram.o diff --git a/arch/powerpc/sysdev/mmio_nvram.c b/arch/powerpc/sysdev/mmio_nvram.c new file mode 100644 index 000000000000..74e0d31a3559 --- /dev/null +++ b/arch/powerpc/sysdev/mmio_nvram.c @@ -0,0 +1,118 @@ +/* + * memory mapped NVRAM + * + * (C) Copyright IBM Corp. 2005 + * + * Authors : Utz Bacher + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2, or (at your option) + * any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program; if not, write to the Free Software + * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA. + */ + +#include +#include +#include +#include +#include + +#include +#include +#include + +static void __iomem *mmio_nvram_start; +static long mmio_nvram_len; +static spinlock_t mmio_nvram_lock = SPIN_LOCK_UNLOCKED; + +static ssize_t mmio_nvram_read(char *buf, size_t count, loff_t *index) +{ + unsigned long flags; + + if (*index >= mmio_nvram_len) + return 0; + if (*index + count > mmio_nvram_len) + count = mmio_nvram_len - *index; + + spin_lock_irqsave(&mmio_nvram_lock, flags); + + memcpy_fromio(buf, mmio_nvram_start + *index, count); + + spin_unlock_irqrestore(&mmio_nvram_lock, flags); + + *index += count; + return count; +} + +static ssize_t mmio_nvram_write(char *buf, size_t count, loff_t *index) +{ + unsigned long flags; + + if (*index >= mmio_nvram_len) + return 0; + if (*index + count > mmio_nvram_len) + count = mmio_nvram_len - *index; + + spin_lock_irqsave(&mmio_nvram_lock, flags); + + memcpy_toio(mmio_nvram_start + *index, buf, count); + + spin_unlock_irqrestore(&mmio_nvram_lock, flags); + + *index += count; + return count; +} + +static ssize_t mmio_nvram_get_size(void) +{ + return mmio_nvram_len; +} + +int __init mmio_nvram_init(void) +{ + struct device_node *nvram_node; + unsigned long *buffer; + int proplen; + unsigned long nvram_addr; + int ret; + + ret = -ENODEV; + nvram_node = of_find_node_by_type(NULL, "nvram"); + if (!nvram_node) + goto out; + + ret = -EIO; + buffer = (unsigned long *)get_property(nvram_node, "reg", &proplen); + if (proplen != 2*sizeof(unsigned long)) + goto out; + + ret = -ENODEV; + nvram_addr = buffer[0]; + mmio_nvram_len = buffer[1]; + if ( (!mmio_nvram_len) || (!nvram_addr) ) + goto out; + + mmio_nvram_start = ioremap(nvram_addr, mmio_nvram_len); + if (!mmio_nvram_start) + goto out; + + printk(KERN_INFO "mmio NVRAM, %luk mapped to %p\n", + mmio_nvram_len >> 10, mmio_nvram_start); + + ppc_md.nvram_read = mmio_nvram_read; + ppc_md.nvram_write = mmio_nvram_write; + ppc_md.nvram_size = mmio_nvram_get_size; + +out: + of_node_put(nvram_node); + return ret; +} diff --git a/arch/ppc64/kernel/Makefile b/arch/ppc64/kernel/Makefile index a01a9de9eb56..afd445114446 100644 --- a/arch/ppc64/kernel/Makefile +++ b/arch/ppc64/kernel/Makefile @@ -31,7 +31,7 @@ endif obj-$(CONFIG_PPC_PSERIES) += rtasd.o udbg_16550.o -obj-$(CONFIG_PPC_CELL) += bpa_setup.o bpa_iommu.o bpa_nvram.o \ +obj-$(CONFIG_PPC_CELL) += bpa_setup.o bpa_iommu.o \ bpa_iic.o spider-pic.o obj-$(CONFIG_KEXEC) += machine_kexec.o diff --git a/arch/ppc64/kernel/bpa_nvram.c b/arch/ppc64/kernel/bpa_nvram.c deleted file mode 100644 index 74e0d31a3559..000000000000 --- a/arch/ppc64/kernel/bpa_nvram.c +++ /dev/null @@ -1,118 +0,0 @@ -/* - * memory mapped NVRAM - * - * (C) Copyright IBM Corp. 2005 - * - * Authors : Utz Bacher - * - * This program is free software; you can redistribute it and/or modify - * it under the terms of the GNU General Public License as published by - * the Free Software Foundation; either version 2, or (at your option) - * any later version. - * - * This program is distributed in the hope that it will be useful, - * but WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the - * GNU General Public License for more details. - * - * You should have received a copy of the GNU General Public License - * along with this program; if not, write to the Free Software - * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA. - */ - -#include -#include -#include -#include -#include - -#include -#include -#include - -static void __iomem *mmio_nvram_start; -static long mmio_nvram_len; -static spinlock_t mmio_nvram_lock = SPIN_LOCK_UNLOCKED; - -static ssize_t mmio_nvram_read(char *buf, size_t count, loff_t *index) -{ - unsigned long flags; - - if (*index >= mmio_nvram_len) - return 0; - if (*index + count > mmio_nvram_len) - count = mmio_nvram_len - *index; - - spin_lock_irqsave(&mmio_nvram_lock, flags); - - memcpy_fromio(buf, mmio_nvram_start + *index, count); - - spin_unlock_irqrestore(&mmio_nvram_lock, flags); - - *index += count; - return count; -} - -static ssize_t mmio_nvram_write(char *buf, size_t count, loff_t *index) -{ - unsigned long flags; - - if (*index >= mmio_nvram_len) - return 0; - if (*index + count > mmio_nvram_len) - count = mmio_nvram_len - *index; - - spin_lock_irqsave(&mmio_nvram_lock, flags); - - memcpy_toio(mmio_nvram_start + *index, buf, count); - - spin_unlock_irqrestore(&mmio_nvram_lock, flags); - - *index += count; - return count; -} - -static ssize_t mmio_nvram_get_size(void) -{ - return mmio_nvram_len; -} - -int __init mmio_nvram_init(void) -{ - struct device_node *nvram_node; - unsigned long *buffer; - int proplen; - unsigned long nvram_addr; - int ret; - - ret = -ENODEV; - nvram_node = of_find_node_by_type(NULL, "nvram"); - if (!nvram_node) - goto out; - - ret = -EIO; - buffer = (unsigned long *)get_property(nvram_node, "reg", &proplen); - if (proplen != 2*sizeof(unsigned long)) - goto out; - - ret = -ENODEV; - nvram_addr = buffer[0]; - mmio_nvram_len = buffer[1]; - if ( (!mmio_nvram_len) || (!nvram_addr) ) - goto out; - - mmio_nvram_start = ioremap(nvram_addr, mmio_nvram_len); - if (!mmio_nvram_start) - goto out; - - printk(KERN_INFO "mmio NVRAM, %luk mapped to %p\n", - mmio_nvram_len >> 10, mmio_nvram_start); - - ppc_md.nvram_read = mmio_nvram_read; - ppc_md.nvram_write = mmio_nvram_write; - ppc_md.nvram_size = mmio_nvram_get_size; - -out: - of_node_put(nvram_node); - return ret; -} -- cgit v1.2.3 From 811d4176f4366ca13a00c1fbb57587d210d38f19 Mon Sep 17 00:00:00 2001 From: Arnd Bergmann Date: Mon, 31 Oct 2005 20:08:41 -0500 Subject: [PATCH] powerpc: move arch/ppc64/kernel/bpa* to arch/powerpc/platforms/cell This patch simply moves files over to arch/powerpc without making any changes to them. Signed-off-by: Arnd Bergmann Signed-off-by: Paul Mackerras --- arch/powerpc/platforms/cell/Makefile | 1 + arch/powerpc/platforms/cell/interrupt.c | 284 +++++++++++++++++++++++ arch/powerpc/platforms/cell/interrupt.h | 62 +++++ arch/powerpc/platforms/cell/iommu.c | 381 +++++++++++++++++++++++++++++++ arch/powerpc/platforms/cell/iommu.h | 65 ++++++ arch/powerpc/platforms/cell/setup.c | 141 ++++++++++++ arch/powerpc/platforms/cell/spider-pic.c | 191 ++++++++++++++++ arch/ppc64/kernel/Makefile | 3 - arch/ppc64/kernel/bpa_iic.c | 284 ----------------------- arch/ppc64/kernel/bpa_iic.h | 62 ----- arch/ppc64/kernel/bpa_iommu.c | 381 ------------------------------- arch/ppc64/kernel/bpa_iommu.h | 65 ------ arch/ppc64/kernel/bpa_setup.c | 141 ------------ arch/ppc64/kernel/spider-pic.c | 191 ---------------- 14 files changed, 1125 insertions(+), 1127 deletions(-) create mode 100644 arch/powerpc/platforms/cell/interrupt.c create mode 100644 arch/powerpc/platforms/cell/interrupt.h create mode 100644 arch/powerpc/platforms/cell/iommu.c create mode 100644 arch/powerpc/platforms/cell/iommu.h create mode 100644 arch/powerpc/platforms/cell/setup.c create mode 100644 arch/powerpc/platforms/cell/spider-pic.c delete mode 100644 arch/ppc64/kernel/bpa_iic.c delete mode 100644 arch/ppc64/kernel/bpa_iic.h delete mode 100644 arch/ppc64/kernel/bpa_iommu.c delete mode 100644 arch/ppc64/kernel/bpa_iommu.h delete mode 100644 arch/ppc64/kernel/bpa_setup.c delete mode 100644 arch/ppc64/kernel/spider-pic.c (limited to 'arch/ppc64') diff --git a/arch/powerpc/platforms/cell/Makefile b/arch/powerpc/platforms/cell/Makefile index 7710d2e08037..55e094b96bc0 100644 --- a/arch/powerpc/platforms/cell/Makefile +++ b/arch/powerpc/platforms/cell/Makefile @@ -1 +1,2 @@ +obj-y += interrupt.o iommu.o setup.o spider-pic.o obj-$(CONFIG_SMP) += smp.o diff --git a/arch/powerpc/platforms/cell/interrupt.c b/arch/powerpc/platforms/cell/interrupt.c new file mode 100644 index 000000000000..7fbe78a9327d --- /dev/null +++ b/arch/powerpc/platforms/cell/interrupt.c @@ -0,0 +1,284 @@ +/* + * Cell Internal Interrupt Controller + * + * (C) Copyright IBM Deutschland Entwicklung GmbH 2005 + * + * Author: Arnd Bergmann + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2, or (at your option) + * any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program; if not, write to the Free Software + * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA. + */ + +#include +#include +#include +#include +#include + +#include +#include +#include +#include + +#include "interrupt.h" + +struct iic_pending_bits { + u32 data; + u8 flags; + u8 class; + u8 source; + u8 prio; +}; + +enum iic_pending_flags { + IIC_VALID = 0x80, + IIC_IPI = 0x40, +}; + +struct iic_regs { + struct iic_pending_bits pending; + struct iic_pending_bits pending_destr; + u64 generate; + u64 prio; +}; + +struct iic { + struct iic_regs __iomem *regs; +}; + +static DEFINE_PER_CPU(struct iic, iic); + +void iic_local_enable(void) +{ + out_be64(&__get_cpu_var(iic).regs->prio, 0xff); +} + +void iic_local_disable(void) +{ + out_be64(&__get_cpu_var(iic).regs->prio, 0x0); +} + +static unsigned int iic_startup(unsigned int irq) +{ + return 0; +} + +static void iic_enable(unsigned int irq) +{ + iic_local_enable(); +} + +static void iic_disable(unsigned int irq) +{ +} + +static void iic_end(unsigned int irq) +{ + iic_local_enable(); +} + +static struct hw_interrupt_type iic_pic = { + .typename = " CELL-IIC ", + .startup = iic_startup, + .enable = iic_enable, + .disable = iic_disable, + .end = iic_end, +}; + +static int iic_external_get_irq(struct iic_pending_bits pending) +{ + int irq; + unsigned char node, unit; + + node = pending.source >> 4; + unit = pending.source & 0xf; + irq = -1; + + /* + * This mapping is specific to the Cell Broadband + * Engine. We might need to get the numbers + * from the device tree to support future CPUs. + */ + switch (unit) { + case 0x00: + case 0x0b: + /* + * One of these units can be connected + * to an external interrupt controller. + */ + if (pending.prio > 0x3f || + pending.class != 2) + break; + irq = IIC_EXT_OFFSET + + spider_get_irq(pending.prio + node * IIC_NODE_STRIDE) + + node * IIC_NODE_STRIDE; + break; + case 0x01 ... 0x04: + case 0x07 ... 0x0a: + /* + * These units are connected to the SPEs + */ + if (pending.class > 2) + break; + irq = IIC_SPE_OFFSET + + pending.class * IIC_CLASS_STRIDE + + node * IIC_NODE_STRIDE + + unit; + break; + } + if (irq == -1) + printk(KERN_WARNING "Unexpected interrupt class %02x, " + "source %02x, prio %02x, cpu %02x\n", pending.class, + pending.source, pending.prio, smp_processor_id()); + return irq; +} + +/* Get an IRQ number from the pending state register of the IIC */ +int iic_get_irq(struct pt_regs *regs) +{ + struct iic *iic; + int irq; + struct iic_pending_bits pending; + + iic = &__get_cpu_var(iic); + *(unsigned long *) &pending = + in_be64((unsigned long __iomem *) &iic->regs->pending_destr); + + irq = -1; + if (pending.flags & IIC_VALID) { + if (pending.flags & IIC_IPI) { + irq = IIC_IPI_OFFSET + (pending.prio >> 4); +/* + if (irq > 0x80) + printk(KERN_WARNING "Unexpected IPI prio %02x" + "on CPU %02x\n", pending.prio, + smp_processor_id()); +*/ + } else { + irq = iic_external_get_irq(pending); + } + } + return irq; +} + +static struct iic_regs __iomem *find_iic(int cpu) +{ + struct device_node *np; + int nodeid = cpu / 2; + unsigned long regs; + struct iic_regs __iomem *iic_regs; + + for (np = of_find_node_by_type(NULL, "cpu"); + np; + np = of_find_node_by_type(np, "cpu")) { + if (nodeid == *(int *)get_property(np, "node-id", NULL)) + break; + } + + if (!np) { + printk(KERN_WARNING "IIC: CPU %d not found\n", cpu); + iic_regs = NULL; + } else { + regs = *(long *)get_property(np, "iic", NULL); + + /* hack until we have decided on the devtree info */ + regs += 0x400; + if (cpu & 1) + regs += 0x20; + + printk(KERN_DEBUG "IIC for CPU %d at %lx\n", cpu, regs); + iic_regs = __ioremap(regs, sizeof(struct iic_regs), + _PAGE_NO_CACHE); + } + return iic_regs; +} + +#ifdef CONFIG_SMP + +/* Use the highest interrupt priorities for IPI */ +static inline int iic_ipi_to_irq(int ipi) +{ + return IIC_IPI_OFFSET + IIC_NUM_IPIS - 1 - ipi; +} + +static inline int iic_irq_to_ipi(int irq) +{ + return IIC_NUM_IPIS - 1 - (irq - IIC_IPI_OFFSET); +} + +void iic_setup_cpu(void) +{ + out_be64(&__get_cpu_var(iic).regs->prio, 0xff); +} + +void iic_cause_IPI(int cpu, int mesg) +{ + out_be64(&per_cpu(iic, cpu).regs->generate, (IIC_NUM_IPIS - 1 - mesg) << 4); +} + +static irqreturn_t iic_ipi_action(int irq, void *dev_id, struct pt_regs *regs) +{ + smp_message_recv(iic_irq_to_ipi(irq), regs); + return IRQ_HANDLED; +} + +static void iic_request_ipi(int ipi, const char *name) +{ + int irq; + + irq = iic_ipi_to_irq(ipi); + /* IPIs are marked SA_INTERRUPT as they must run with irqs + * disabled */ + get_irq_desc(irq)->handler = &iic_pic; + get_irq_desc(irq)->status |= IRQ_PER_CPU; + request_irq(irq, iic_ipi_action, SA_INTERRUPT, name, NULL); +} + +void iic_request_IPIs(void) +{ + iic_request_ipi(PPC_MSG_CALL_FUNCTION, "IPI-call"); + iic_request_ipi(PPC_MSG_RESCHEDULE, "IPI-resched"); +#ifdef CONFIG_DEBUGGER + iic_request_ipi(PPC_MSG_DEBUGGER_BREAK, "IPI-debug"); +#endif /* CONFIG_DEBUGGER */ +} +#endif /* CONFIG_SMP */ + +static void iic_setup_spe_handlers(void) +{ + int be, isrc; + + /* Assume two threads per BE are present */ + for (be=0; be < num_present_cpus() / 2; be++) { + for (isrc = 0; isrc < IIC_CLASS_STRIDE * 3; isrc++) { + int irq = IIC_NODE_STRIDE * be + IIC_SPE_OFFSET + isrc; + get_irq_desc(irq)->handler = &iic_pic; + } + } +} + +void iic_init_IRQ(void) +{ + int cpu, irq_offset; + struct iic *iic; + + irq_offset = 0; + for_each_cpu(cpu) { + iic = &per_cpu(iic, cpu); + iic->regs = find_iic(cpu); + if (iic->regs) + out_be64(&iic->regs->prio, 0xff); + } + iic_setup_spe_handlers(); +} diff --git a/arch/powerpc/platforms/cell/interrupt.h b/arch/powerpc/platforms/cell/interrupt.h new file mode 100644 index 000000000000..37d58e6fd0c6 --- /dev/null +++ b/arch/powerpc/platforms/cell/interrupt.h @@ -0,0 +1,62 @@ +#ifndef ASM_CELL_PIC_H +#define ASM_CELL_PIC_H +#ifdef __KERNEL__ +/* + * Mapping of IIC pending bits into per-node + * interrupt numbers. + * + * IRQ FF CC SS PP FF CC SS PP Description + * + * 00-3f 80 02 +0 00 - 80 02 +0 3f South Bridge + * 00-3f 80 02 +b 00 - 80 02 +b 3f South Bridge + * 41-4a 80 00 +1 ** - 80 00 +a ** SPU Class 0 + * 51-5a 80 01 +1 ** - 80 01 +a ** SPU Class 1 + * 61-6a 80 02 +1 ** - 80 02 +a ** SPU Class 2 + * 70-7f C0 ** ** 00 - C0 ** ** 0f IPI + * + * F flags + * C class + * S source + * P Priority + * + node number + * * don't care + * + * A node consists of a Cell Broadband Engine and an optional + * south bridge device providing a maximum of 64 IRQs. + * The south bridge may be connected to either IOIF0 + * or IOIF1. + * Each SPE is represented as three IRQ lines, one per + * interrupt class. + * 16 IRQ numbers are reserved for inter processor + * interruptions, although these are only used in the + * range of the first node. + * + * This scheme needs 128 IRQ numbers per BIF node ID, + * which means that with the total of 512 lines + * available, we can have a maximum of four nodes. + */ + +enum { + IIC_EXT_OFFSET = 0x00, /* Start of south bridge IRQs */ + IIC_NUM_EXT = 0x40, /* Number of south bridge IRQs */ + IIC_SPE_OFFSET = 0x40, /* Start of SPE interrupts */ + IIC_CLASS_STRIDE = 0x10, /* SPE IRQs per class */ + IIC_IPI_OFFSET = 0x70, /* Start of IPI IRQs */ + IIC_NUM_IPIS = 0x10, /* IRQs reserved for IPI */ + IIC_NODE_STRIDE = 0x80, /* Total IRQs per node */ +}; + +extern void iic_init_IRQ(void); +extern int iic_get_irq(struct pt_regs *regs); +extern void iic_cause_IPI(int cpu, int mesg); +extern void iic_request_IPIs(void); +extern void iic_setup_cpu(void); +extern void iic_local_enable(void); +extern void iic_local_disable(void); + + +extern void spider_init_IRQ(void); +extern int spider_get_irq(unsigned long int_pending); + +#endif +#endif /* ASM_CELL_PIC_H */ diff --git a/arch/powerpc/platforms/cell/iommu.c b/arch/powerpc/platforms/cell/iommu.c new file mode 100644 index 000000000000..74f999b4ac9e --- /dev/null +++ b/arch/powerpc/platforms/cell/iommu.c @@ -0,0 +1,381 @@ +/* + * IOMMU implementation for Cell Broadband Processor Architecture + * We just establish a linear mapping at boot by setting all the + * IOPT cache entries in the CPU. + * The mapping functions should be identical to pci_direct_iommu, + * except for the handling of the high order bit that is required + * by the Spider bridge. These should be split into a separate + * file at the point where we get a different bridge chip. + * + * Copyright (C) 2005 IBM Deutschland Entwicklung GmbH, + * Arnd Bergmann + * + * Based on linear mapping + * Copyright (C) 2003 Benjamin Herrenschmidt (benh@kernel.crashing.org) + * + * This program is free software; you can redistribute it and/or + * modify it under the terms of the GNU General Public License + * as published by the Free Software Foundation; either version + * 2 of the License, or (at your option) any later version. + */ + +#undef DEBUG + +#include +#include +#include +#include +#include +#include +#include +#include + +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include + +#include "iommu.h" + +static inline unsigned long +get_iopt_entry(unsigned long real_address, unsigned long ioid, + unsigned long prot) +{ + return (prot & IOPT_PROT_MASK) + | (IOPT_COHERENT) + | (IOPT_ORDER_VC) + | (real_address & IOPT_RPN_MASK) + | (ioid & IOPT_IOID_MASK); +} + +typedef struct { + unsigned long val; +} ioste; + +static inline ioste +mk_ioste(unsigned long val) +{ + ioste ioste = { .val = val, }; + return ioste; +} + +static inline ioste +get_iost_entry(unsigned long iopt_base, unsigned long io_address, unsigned page_size) +{ + unsigned long ps; + unsigned long iostep; + unsigned long nnpt; + unsigned long shift; + + switch (page_size) { + case 0x1000000: + ps = IOST_PS_16M; + nnpt = 0; /* one page per segment */ + shift = 5; /* segment has 16 iopt entries */ + break; + + case 0x100000: + ps = IOST_PS_1M; + nnpt = 0; /* one page per segment */ + shift = 1; /* segment has 256 iopt entries */ + break; + + case 0x10000: + ps = IOST_PS_64K; + nnpt = 0x07; /* 8 pages per io page table */ + shift = 0; /* all entries are used */ + break; + + case 0x1000: + ps = IOST_PS_4K; + nnpt = 0x7f; /* 128 pages per io page table */ + shift = 0; /* all entries are used */ + break; + + default: /* not a known compile time constant */ + { + /* BUILD_BUG_ON() is not usable here */ + extern void __get_iost_entry_bad_page_size(void); + __get_iost_entry_bad_page_size(); + } + break; + } + + iostep = iopt_base + + /* need 8 bytes per iopte */ + (((io_address / page_size * 8) + /* align io page tables on 4k page boundaries */ + << shift) + /* nnpt+1 pages go into each iopt */ + & ~(nnpt << 12)); + + nnpt++; /* this seems to work, but the documentation is not clear + about wether we put nnpt or nnpt-1 into the ioste bits. + In theory, this can't work for 4k pages. */ + return mk_ioste(IOST_VALID_MASK + | (iostep & IOST_PT_BASE_MASK) + | ((nnpt << 5) & IOST_NNPT_MASK) + | (ps & IOST_PS_MASK)); +} + +/* compute the address of an io pte */ +static inline unsigned long +get_ioptep(ioste iost_entry, unsigned long io_address) +{ + unsigned long iopt_base; + unsigned long page_size; + unsigned long page_number; + unsigned long iopt_offset; + + iopt_base = iost_entry.val & IOST_PT_BASE_MASK; + page_size = iost_entry.val & IOST_PS_MASK; + + /* decode page size to compute page number */ + page_number = (io_address & 0x0fffffff) >> (10 + 2 * page_size); + /* page number is an offset into the io page table */ + iopt_offset = (page_number << 3) & 0x7fff8ul; + return iopt_base + iopt_offset; +} + +/* compute the tag field of the iopt cache entry */ +static inline unsigned long +get_ioc_tag(ioste iost_entry, unsigned long io_address) +{ + unsigned long iopte = get_ioptep(iost_entry, io_address); + + return IOPT_VALID_MASK + | ((iopte & 0x00000000000000ff8ul) >> 3) + | ((iopte & 0x0000003fffffc0000ul) >> 9); +} + +/* compute the hashed 6 bit index for the 4-way associative pte cache */ +static inline unsigned long +get_ioc_hash(ioste iost_entry, unsigned long io_address) +{ + unsigned long iopte = get_ioptep(iost_entry, io_address); + + return ((iopte & 0x000000000000001f8ul) >> 3) + ^ ((iopte & 0x00000000000020000ul) >> 17) + ^ ((iopte & 0x00000000000010000ul) >> 15) + ^ ((iopte & 0x00000000000008000ul) >> 13) + ^ ((iopte & 0x00000000000004000ul) >> 11) + ^ ((iopte & 0x00000000000002000ul) >> 9) + ^ ((iopte & 0x00000000000001000ul) >> 7); +} + +/* same as above, but pretend that we have a simpler 1-way associative + pte cache with an 8 bit index */ +static inline unsigned long +get_ioc_hash_1way(ioste iost_entry, unsigned long io_address) +{ + unsigned long iopte = get_ioptep(iost_entry, io_address); + + return ((iopte & 0x000000000000001f8ul) >> 3) + ^ ((iopte & 0x00000000000020000ul) >> 17) + ^ ((iopte & 0x00000000000010000ul) >> 15) + ^ ((iopte & 0x00000000000008000ul) >> 13) + ^ ((iopte & 0x00000000000004000ul) >> 11) + ^ ((iopte & 0x00000000000002000ul) >> 9) + ^ ((iopte & 0x00000000000001000ul) >> 7) + ^ ((iopte & 0x0000000000000c000ul) >> 8); +} + +static inline ioste +get_iost_cache(void __iomem *base, unsigned long index) +{ + unsigned long __iomem *p = (base + IOC_ST_CACHE_DIR); + return mk_ioste(in_be64(&p[index])); +} + +static inline void +set_iost_cache(void __iomem *base, unsigned long index, ioste ste) +{ + unsigned long __iomem *p = (base + IOC_ST_CACHE_DIR); + pr_debug("ioste %02lx was %016lx, store %016lx", index, + get_iost_cache(base, index).val, ste.val); + out_be64(&p[index], ste.val); + pr_debug(" now %016lx\n", get_iost_cache(base, index).val); +} + +static inline unsigned long +get_iopt_cache(void __iomem *base, unsigned long index, unsigned long *tag) +{ + unsigned long __iomem *tags = (void *)(base + IOC_PT_CACHE_DIR); + unsigned long __iomem *p = (void *)(base + IOC_PT_CACHE_REG); + + *tag = tags[index]; + rmb(); + return *p; +} + +static inline void +set_iopt_cache(void __iomem *base, unsigned long index, + unsigned long tag, unsigned long val) +{ + unsigned long __iomem *tags = base + IOC_PT_CACHE_DIR; + unsigned long __iomem *p = base + IOC_PT_CACHE_REG; + pr_debug("iopt %02lx was v%016lx/t%016lx, store v%016lx/t%016lx\n", + index, get_iopt_cache(base, index, &oldtag), oldtag, val, tag); + + out_be64(p, val); + out_be64(&tags[index], tag); +} + +static inline void +set_iost_origin(void __iomem *base) +{ + unsigned long __iomem *p = base + IOC_ST_ORIGIN; + unsigned long origin = IOSTO_ENABLE | IOSTO_SW; + + pr_debug("iost_origin %016lx, now %016lx\n", in_be64(p), origin); + out_be64(p, origin); +} + +static inline void +set_iocmd_config(void __iomem *base) +{ + unsigned long __iomem *p = base + 0xc00; + unsigned long conf; + + conf = in_be64(p); + pr_debug("iost_conf %016lx, now %016lx\n", conf, conf | IOCMD_CONF_TE); + out_be64(p, conf | IOCMD_CONF_TE); +} + +/* FIXME: get these from the device tree */ +#define ioc_base 0x20000511000ull +#define ioc_mmio_base 0x20000510000ull +#define ioid 0x48a +#define iopt_phys_offset (- 0x20000000) /* We have a 512MB offset from the SB */ +#define io_page_size 0x1000000 + +static unsigned long map_iopt_entry(unsigned long address) +{ + switch (address >> 20) { + case 0x600: + address = 0x24020000000ull; /* spider i/o */ + break; + default: + address += iopt_phys_offset; + break; + } + + return get_iopt_entry(address, ioid, IOPT_PROT_RW); +} + +static void iommu_bus_setup_null(struct pci_bus *b) { } +static void iommu_dev_setup_null(struct pci_dev *d) { } + +/* initialize the iommu to support a simple linear mapping + * for each DMA window used by any device. For now, we + * happen to know that there is only one DMA window in use, + * starting at iopt_phys_offset. */ +static void cell_map_iommu(void) +{ + unsigned long address; + void __iomem *base; + ioste ioste; + unsigned long index; + + base = __ioremap(ioc_base, 0x1000, _PAGE_NO_CACHE); + pr_debug("%lx mapped to %p\n", ioc_base, base); + set_iocmd_config(base); + iounmap(base); + + base = __ioremap(ioc_mmio_base, 0x1000, _PAGE_NO_CACHE); + pr_debug("%lx mapped to %p\n", ioc_mmio_base, base); + + set_iost_origin(base); + + for (address = 0; address < 0x100000000ul; address += io_page_size) { + ioste = get_iost_entry(0x10000000000ul, address, io_page_size); + if ((address & 0xfffffff) == 0) /* segment start */ + set_iost_cache(base, address >> 28, ioste); + index = get_ioc_hash_1way(ioste, address); + pr_debug("addr %08lx, index %02lx, ioste %016lx\n", + address, index, ioste.val); + set_iopt_cache(base, + get_ioc_hash_1way(ioste, address), + get_ioc_tag(ioste, address), + map_iopt_entry(address)); + } + iounmap(base); +} + + +static void *cell_alloc_coherent(struct device *hwdev, size_t size, + dma_addr_t *dma_handle, gfp_t flag) +{ + void *ret; + + ret = (void *)__get_free_pages(flag, get_order(size)); + if (ret != NULL) { + memset(ret, 0, size); + *dma_handle = virt_to_abs(ret) | CELL_DMA_VALID; + } + return ret; +} + +static void cell_free_coherent(struct device *hwdev, size_t size, + void *vaddr, dma_addr_t dma_handle) +{ + free_pages((unsigned long)vaddr, get_order(size)); +} + +static dma_addr_t cell_map_single(struct device *hwdev, void *ptr, + size_t size, enum dma_data_direction direction) +{ + return virt_to_abs(ptr) | CELL_DMA_VALID; +} + +static void cell_unmap_single(struct device *hwdev, dma_addr_t dma_addr, + size_t size, enum dma_data_direction direction) +{ +} + +static int cell_map_sg(struct device *hwdev, struct scatterlist *sg, + int nents, enum dma_data_direction direction) +{ + int i; + + for (i = 0; i < nents; i++, sg++) { + sg->dma_address = (page_to_phys(sg->page) + sg->offset) + | CELL_DMA_VALID; + sg->dma_length = sg->length; + } + + return nents; +} + +static void cell_unmap_sg(struct device *hwdev, struct scatterlist *sg, + int nents, enum dma_data_direction direction) +{ +} + +static int cell_dma_supported(struct device *dev, u64 mask) +{ + return mask < 0x100000000ull; +} + +void cell_init_iommu(void) +{ + cell_map_iommu(); + + /* Direct I/O, IOMMU off */ + ppc_md.iommu_dev_setup = iommu_dev_setup_null; + ppc_md.iommu_bus_setup = iommu_bus_setup_null; + + pci_dma_ops.alloc_coherent = cell_alloc_coherent; + pci_dma_ops.free_coherent = cell_free_coherent; + pci_dma_ops.map_single = cell_map_single; + pci_dma_ops.unmap_single = cell_unmap_single; + pci_dma_ops.map_sg = cell_map_sg; + pci_dma_ops.unmap_sg = cell_unmap_sg; + pci_dma_ops.dma_supported = cell_dma_supported; +} diff --git a/arch/powerpc/platforms/cell/iommu.h b/arch/powerpc/platforms/cell/iommu.h new file mode 100644 index 000000000000..490d77abfe85 --- /dev/null +++ b/arch/powerpc/platforms/cell/iommu.h @@ -0,0 +1,65 @@ +#ifndef CELL_IOMMU_H +#define CELL_IOMMU_H + +/* some constants */ +enum { + /* segment table entries */ + IOST_VALID_MASK = 0x8000000000000000ul, + IOST_TAG_MASK = 0x3000000000000000ul, + IOST_PT_BASE_MASK = 0x000003fffffff000ul, + IOST_NNPT_MASK = 0x0000000000000fe0ul, + IOST_PS_MASK = 0x000000000000000ful, + + IOST_PS_4K = 0x1, + IOST_PS_64K = 0x3, + IOST_PS_1M = 0x5, + IOST_PS_16M = 0x7, + + /* iopt tag register */ + IOPT_VALID_MASK = 0x0000000200000000ul, + IOPT_TAG_MASK = 0x00000001fffffffful, + + /* iopt cache register */ + IOPT_PROT_MASK = 0xc000000000000000ul, + IOPT_PROT_NONE = 0x0000000000000000ul, + IOPT_PROT_READ = 0x4000000000000000ul, + IOPT_PROT_WRITE = 0x8000000000000000ul, + IOPT_PROT_RW = 0xc000000000000000ul, + IOPT_COHERENT = 0x2000000000000000ul, + + IOPT_ORDER_MASK = 0x1800000000000000ul, + /* order access to same IOID/VC on same address */ + IOPT_ORDER_ADDR = 0x0800000000000000ul, + /* similar, but only after a write access */ + IOPT_ORDER_WRITES = 0x1000000000000000ul, + /* Order all accesses to same IOID/VC */ + IOPT_ORDER_VC = 0x1800000000000000ul, + + IOPT_RPN_MASK = 0x000003fffffff000ul, + IOPT_HINT_MASK = 0x0000000000000800ul, + IOPT_IOID_MASK = 0x00000000000007fful, + + IOSTO_ENABLE = 0x8000000000000000ul, + IOSTO_ORIGIN = 0x000003fffffff000ul, + IOSTO_HW = 0x0000000000000800ul, + IOSTO_SW = 0x0000000000000400ul, + + IOCMD_CONF_TE = 0x0000800000000000ul, + + /* memory mapped registers */ + IOC_PT_CACHE_DIR = 0x000, + IOC_ST_CACHE_DIR = 0x800, + IOC_PT_CACHE_REG = 0x910, + IOC_ST_ORIGIN = 0x918, + IOC_CONF = 0x930, + + /* The high bit needs to be set on every DMA address, + only 2GB are addressable */ + CELL_DMA_VALID = 0x80000000, + CELL_DMA_MASK = 0x7fffffff, +}; + + +void cell_init_iommu(void); + +#endif diff --git a/arch/powerpc/platforms/cell/setup.c b/arch/powerpc/platforms/cell/setup.c new file mode 100644 index 000000000000..9a495634d0c2 --- /dev/null +++ b/arch/powerpc/platforms/cell/setup.c @@ -0,0 +1,141 @@ +/* + * linux/arch/powerpc/platforms/cell/cell_setup.c + * + * Copyright (C) 1995 Linus Torvalds + * Adapted from 'alpha' version by Gary Thomas + * Modified by Cort Dougan (cort@cs.nmt.edu) + * Modified by PPC64 Team, IBM Corp + * Modified by Cell Team, IBM Deutschland Entwicklung GmbH + * + * This program is free software; you can redistribute it and/or + * modify it under the terms of the GNU General Public License + * as published by the Free Software Foundation; either version + * 2 of the License, or (at your option) any later version. + */ +#undef DEBUG + +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include + +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include + +#include "interrupt.h" +#include "iommu.h" + +#ifdef DEBUG +#define DBG(fmt...) udbg_printf(fmt) +#else +#define DBG(fmt...) +#endif + +void cell_show_cpuinfo(struct seq_file *m) +{ + struct device_node *root; + const char *model = ""; + + root = of_find_node_by_path("/"); + if (root) + model = get_property(root, "model", NULL); + seq_printf(m, "machine\t\t: CHRP %s\n", model); + of_node_put(root); +} + +static void cell_progress(char *s, unsigned short hex) +{ + printk("*** %04x : %s\n", hex, s ? s : ""); +} + +static void __init cell_setup_arch(void) +{ + ppc_md.init_IRQ = iic_init_IRQ; + ppc_md.get_irq = iic_get_irq; + +#ifdef CONFIG_SMP + smp_init_cell(); +#endif + + /* init to some ~sane value until calibrate_delay() runs */ + loops_per_jiffy = 50000000; + + if (ROOT_DEV == 0) { + printk("No ramdisk, default root is /dev/hda2\n"); + ROOT_DEV = Root_HDA2; + } + + /* Find and initialize PCI host bridges */ + init_pci_config_tokens(); + find_and_init_phbs(); + spider_init_IRQ(); +#ifdef CONFIG_DUMMY_CONSOLE + conswitchp = &dummy_con; +#endif + + mmio_nvram_init(); +} + +/* + * Early initialization. Relocation is on but do not reference unbolted pages + */ +static void __init cell_init_early(void) +{ + DBG(" -> cell_init_early()\n"); + + hpte_init_native(); + + cell_init_iommu(); + + ppc64_interrupt_controller = IC_CELL_PIC; + + DBG(" <- cell_init_early()\n"); +} + + +static int __init cell_probe(int platform) +{ + if (platform != PLATFORM_CELL) + return 0; + + return 1; +} + +struct machdep_calls __initdata cell_md = { + .probe = cell_probe, + .setup_arch = cell_setup_arch, + .init_early = cell_init_early, + .show_cpuinfo = cell_show_cpuinfo, + .restart = rtas_restart, + .power_off = rtas_power_off, + .halt = rtas_halt, + .get_boot_time = rtas_get_boot_time, + .get_rtc_time = rtas_get_rtc_time, + .set_rtc_time = rtas_set_rtc_time, + .calibrate_decr = generic_calibrate_decr, + .progress = cell_progress, +}; diff --git a/arch/powerpc/platforms/cell/spider-pic.c b/arch/powerpc/platforms/cell/spider-pic.c new file mode 100644 index 000000000000..e74132188bdf --- /dev/null +++ b/arch/powerpc/platforms/cell/spider-pic.c @@ -0,0 +1,191 @@ +/* + * External Interrupt Controller on Spider South Bridge + * + * (C) Copyright IBM Deutschland Entwicklung GmbH 2005 + * + * Author: Arnd Bergmann + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2, or (at your option) + * any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program; if not, write to the Free Software + * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA. + */ + +#include +#include + +#include +#include +#include + +#include "interrupt.h" + +/* register layout taken from Spider spec, table 7.4-4 */ +enum { + TIR_DEN = 0x004, /* Detection Enable Register */ + TIR_MSK = 0x084, /* Mask Level Register */ + TIR_EDC = 0x0c0, /* Edge Detection Clear Register */ + TIR_PNDA = 0x100, /* Pending Register A */ + TIR_PNDB = 0x104, /* Pending Register B */ + TIR_CS = 0x144, /* Current Status Register */ + TIR_LCSA = 0x150, /* Level Current Status Register A */ + TIR_LCSB = 0x154, /* Level Current Status Register B */ + TIR_LCSC = 0x158, /* Level Current Status Register C */ + TIR_LCSD = 0x15c, /* Level Current Status Register D */ + TIR_CFGA = 0x200, /* Setting Register A0 */ + TIR_CFGB = 0x204, /* Setting Register B0 */ + /* 0x208 ... 0x3ff Setting Register An/Bn */ + TIR_PPNDA = 0x400, /* Packet Pending Register A */ + TIR_PPNDB = 0x404, /* Packet Pending Register B */ + TIR_PIERA = 0x408, /* Packet Output Error Register A */ + TIR_PIERB = 0x40c, /* Packet Output Error Register B */ + TIR_PIEN = 0x444, /* Packet Output Enable Register */ + TIR_PIPND = 0x454, /* Packet Output Pending Register */ + TIRDID = 0x484, /* Spider Device ID Register */ + REISTIM = 0x500, /* Reissue Command Timeout Time Setting */ + REISTIMEN = 0x504, /* Reissue Command Timeout Setting */ + REISWAITEN = 0x508, /* Reissue Wait Control*/ +}; + +static void __iomem *spider_pics[4]; + +static void __iomem *spider_get_pic(int irq) +{ + int node = irq / IIC_NODE_STRIDE; + irq %= IIC_NODE_STRIDE; + + if (irq >= IIC_EXT_OFFSET && + irq < IIC_EXT_OFFSET + IIC_NUM_EXT && + spider_pics) + return spider_pics[node]; + return NULL; +} + +static int spider_get_nr(unsigned int irq) +{ + return (irq % IIC_NODE_STRIDE) - IIC_EXT_OFFSET; +} + +static void __iomem *spider_get_irq_config(int irq) +{ + void __iomem *pic; + pic = spider_get_pic(irq); + return pic + TIR_CFGA + 8 * spider_get_nr(irq); +} + +static void spider_enable_irq(unsigned int irq) +{ + void __iomem *cfg = spider_get_irq_config(irq); + irq = spider_get_nr(irq); + + out_be32(cfg, in_be32(cfg) | 0x3107000eu); + out_be32(cfg + 4, in_be32(cfg + 4) | 0x00020000u | irq); +} + +static void spider_disable_irq(unsigned int irq) +{ + void __iomem *cfg = spider_get_irq_config(irq); + irq = spider_get_nr(irq); + + out_be32(cfg, in_be32(cfg) & ~0x30000000u); +} + +static unsigned int spider_startup_irq(unsigned int irq) +{ + spider_enable_irq(irq); + return 0; +} + +static void spider_shutdown_irq(unsigned int irq) +{ + spider_disable_irq(irq); +} + +static void spider_end_irq(unsigned int irq) +{ + spider_enable_irq(irq); +} + +static void spider_ack_irq(unsigned int irq) +{ + spider_disable_irq(irq); + iic_local_enable(); +} + +static struct hw_interrupt_type spider_pic = { + .typename = " SPIDER ", + .startup = spider_startup_irq, + .shutdown = spider_shutdown_irq, + .enable = spider_enable_irq, + .disable = spider_disable_irq, + .ack = spider_ack_irq, + .end = spider_end_irq, +}; + + +int spider_get_irq(unsigned long int_pending) +{ + void __iomem *regs = spider_get_pic(int_pending); + unsigned long cs; + int irq; + + cs = in_be32(regs + TIR_CS); + + irq = cs >> 24; + if (irq != 63) + return irq; + + return -1; +} + +void spider_init_IRQ(void) +{ + int node; + struct device_node *dn; + unsigned int *property; + long spiderpic; + int n; + +/* FIXME: detect multiple PICs as soon as the device tree has them */ + for (node = 0; node < 1; node++) { + dn = of_find_node_by_path("/"); + n = prom_n_addr_cells(dn); + property = (unsigned int *) get_property(dn, + "platform-spider-pic", NULL); + + if (!property) + continue; + for (spiderpic = 0; n > 0; --n) + spiderpic = (spiderpic << 32) + *property++; + printk(KERN_DEBUG "SPIDER addr: %lx\n", spiderpic); + spider_pics[node] = __ioremap(spiderpic, 0x800, _PAGE_NO_CACHE); + for (n = 0; n < IIC_NUM_EXT; n++) { + int irq = n + IIC_EXT_OFFSET + node * IIC_NODE_STRIDE; + get_irq_desc(irq)->handler = &spider_pic; + + /* do not mask any interrupts because of level */ + out_be32(spider_pics[node] + TIR_MSK, 0x0); + + /* disable edge detection clear */ + /* out_be32(spider_pics[node] + TIR_EDC, 0x0); */ + + /* enable interrupt packets to be output */ + out_be32(spider_pics[node] + TIR_PIEN, + in_be32(spider_pics[node] + TIR_PIEN) | 0x1); + + /* Enable the interrupt detection enable bit. Do this last! */ + out_be32(spider_pics[node] + TIR_DEN, + in_be32(spider_pics[node] +TIR_DEN) | 0x1); + + } + } +} diff --git a/arch/ppc64/kernel/Makefile b/arch/ppc64/kernel/Makefile index afd445114446..247d2fc6a8ed 100644 --- a/arch/ppc64/kernel/Makefile +++ b/arch/ppc64/kernel/Makefile @@ -31,9 +31,6 @@ endif obj-$(CONFIG_PPC_PSERIES) += rtasd.o udbg_16550.o -obj-$(CONFIG_PPC_CELL) += bpa_setup.o bpa_iommu.o \ - bpa_iic.o spider-pic.o - obj-$(CONFIG_KEXEC) += machine_kexec.o obj-$(CONFIG_EEH) += eeh.o obj-$(CONFIG_PROC_FS) += proc_ppc64.o diff --git a/arch/ppc64/kernel/bpa_iic.c b/arch/ppc64/kernel/bpa_iic.c deleted file mode 100644 index 7fbe78a9327d..000000000000 --- a/arch/ppc64/kernel/bpa_iic.c +++ /dev/null @@ -1,284 +0,0 @@ -/* - * Cell Internal Interrupt Controller - * - * (C) Copyright IBM Deutschland Entwicklung GmbH 2005 - * - * Author: Arnd Bergmann - * - * This program is free software; you can redistribute it and/or modify - * it under the terms of the GNU General Public License as published by - * the Free Software Foundation; either version 2, or (at your option) - * any later version. - * - * This program is distributed in the hope that it will be useful, - * but WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the - * GNU General Public License for more details. - * - * You should have received a copy of the GNU General Public License - * along with this program; if not, write to the Free Software - * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA. - */ - -#include -#include -#include -#include -#include - -#include -#include -#include -#include - -#include "interrupt.h" - -struct iic_pending_bits { - u32 data; - u8 flags; - u8 class; - u8 source; - u8 prio; -}; - -enum iic_pending_flags { - IIC_VALID = 0x80, - IIC_IPI = 0x40, -}; - -struct iic_regs { - struct iic_pending_bits pending; - struct iic_pending_bits pending_destr; - u64 generate; - u64 prio; -}; - -struct iic { - struct iic_regs __iomem *regs; -}; - -static DEFINE_PER_CPU(struct iic, iic); - -void iic_local_enable(void) -{ - out_be64(&__get_cpu_var(iic).regs->prio, 0xff); -} - -void iic_local_disable(void) -{ - out_be64(&__get_cpu_var(iic).regs->prio, 0x0); -} - -static unsigned int iic_startup(unsigned int irq) -{ - return 0; -} - -static void iic_enable(unsigned int irq) -{ - iic_local_enable(); -} - -static void iic_disable(unsigned int irq) -{ -} - -static void iic_end(unsigned int irq) -{ - iic_local_enable(); -} - -static struct hw_interrupt_type iic_pic = { - .typename = " CELL-IIC ", - .startup = iic_startup, - .enable = iic_enable, - .disable = iic_disable, - .end = iic_end, -}; - -static int iic_external_get_irq(struct iic_pending_bits pending) -{ - int irq; - unsigned char node, unit; - - node = pending.source >> 4; - unit = pending.source & 0xf; - irq = -1; - - /* - * This mapping is specific to the Cell Broadband - * Engine. We might need to get the numbers - * from the device tree to support future CPUs. - */ - switch (unit) { - case 0x00: - case 0x0b: - /* - * One of these units can be connected - * to an external interrupt controller. - */ - if (pending.prio > 0x3f || - pending.class != 2) - break; - irq = IIC_EXT_OFFSET - + spider_get_irq(pending.prio + node * IIC_NODE_STRIDE) - + node * IIC_NODE_STRIDE; - break; - case 0x01 ... 0x04: - case 0x07 ... 0x0a: - /* - * These units are connected to the SPEs - */ - if (pending.class > 2) - break; - irq = IIC_SPE_OFFSET - + pending.class * IIC_CLASS_STRIDE - + node * IIC_NODE_STRIDE - + unit; - break; - } - if (irq == -1) - printk(KERN_WARNING "Unexpected interrupt class %02x, " - "source %02x, prio %02x, cpu %02x\n", pending.class, - pending.source, pending.prio, smp_processor_id()); - return irq; -} - -/* Get an IRQ number from the pending state register of the IIC */ -int iic_get_irq(struct pt_regs *regs) -{ - struct iic *iic; - int irq; - struct iic_pending_bits pending; - - iic = &__get_cpu_var(iic); - *(unsigned long *) &pending = - in_be64((unsigned long __iomem *) &iic->regs->pending_destr); - - irq = -1; - if (pending.flags & IIC_VALID) { - if (pending.flags & IIC_IPI) { - irq = IIC_IPI_OFFSET + (pending.prio >> 4); -/* - if (irq > 0x80) - printk(KERN_WARNING "Unexpected IPI prio %02x" - "on CPU %02x\n", pending.prio, - smp_processor_id()); -*/ - } else { - irq = iic_external_get_irq(pending); - } - } - return irq; -} - -static struct iic_regs __iomem *find_iic(int cpu) -{ - struct device_node *np; - int nodeid = cpu / 2; - unsigned long regs; - struct iic_regs __iomem *iic_regs; - - for (np = of_find_node_by_type(NULL, "cpu"); - np; - np = of_find_node_by_type(np, "cpu")) { - if (nodeid == *(int *)get_property(np, "node-id", NULL)) - break; - } - - if (!np) { - printk(KERN_WARNING "IIC: CPU %d not found\n", cpu); - iic_regs = NULL; - } else { - regs = *(long *)get_property(np, "iic", NULL); - - /* hack until we have decided on the devtree info */ - regs += 0x400; - if (cpu & 1) - regs += 0x20; - - printk(KERN_DEBUG "IIC for CPU %d at %lx\n", cpu, regs); - iic_regs = __ioremap(regs, sizeof(struct iic_regs), - _PAGE_NO_CACHE); - } - return iic_regs; -} - -#ifdef CONFIG_SMP - -/* Use the highest interrupt priorities for IPI */ -static inline int iic_ipi_to_irq(int ipi) -{ - return IIC_IPI_OFFSET + IIC_NUM_IPIS - 1 - ipi; -} - -static inline int iic_irq_to_ipi(int irq) -{ - return IIC_NUM_IPIS - 1 - (irq - IIC_IPI_OFFSET); -} - -void iic_setup_cpu(void) -{ - out_be64(&__get_cpu_var(iic).regs->prio, 0xff); -} - -void iic_cause_IPI(int cpu, int mesg) -{ - out_be64(&per_cpu(iic, cpu).regs->generate, (IIC_NUM_IPIS - 1 - mesg) << 4); -} - -static irqreturn_t iic_ipi_action(int irq, void *dev_id, struct pt_regs *regs) -{ - smp_message_recv(iic_irq_to_ipi(irq), regs); - return IRQ_HANDLED; -} - -static void iic_request_ipi(int ipi, const char *name) -{ - int irq; - - irq = iic_ipi_to_irq(ipi); - /* IPIs are marked SA_INTERRUPT as they must run with irqs - * disabled */ - get_irq_desc(irq)->handler = &iic_pic; - get_irq_desc(irq)->status |= IRQ_PER_CPU; - request_irq(irq, iic_ipi_action, SA_INTERRUPT, name, NULL); -} - -void iic_request_IPIs(void) -{ - iic_request_ipi(PPC_MSG_CALL_FUNCTION, "IPI-call"); - iic_request_ipi(PPC_MSG_RESCHEDULE, "IPI-resched"); -#ifdef CONFIG_DEBUGGER - iic_request_ipi(PPC_MSG_DEBUGGER_BREAK, "IPI-debug"); -#endif /* CONFIG_DEBUGGER */ -} -#endif /* CONFIG_SMP */ - -static void iic_setup_spe_handlers(void) -{ - int be, isrc; - - /* Assume two threads per BE are present */ - for (be=0; be < num_present_cpus() / 2; be++) { - for (isrc = 0; isrc < IIC_CLASS_STRIDE * 3; isrc++) { - int irq = IIC_NODE_STRIDE * be + IIC_SPE_OFFSET + isrc; - get_irq_desc(irq)->handler = &iic_pic; - } - } -} - -void iic_init_IRQ(void) -{ - int cpu, irq_offset; - struct iic *iic; - - irq_offset = 0; - for_each_cpu(cpu) { - iic = &per_cpu(iic, cpu); - iic->regs = find_iic(cpu); - if (iic->regs) - out_be64(&iic->regs->prio, 0xff); - } - iic_setup_spe_handlers(); -} diff --git a/arch/ppc64/kernel/bpa_iic.h b/arch/ppc64/kernel/bpa_iic.h deleted file mode 100644 index 37d58e6fd0c6..000000000000 --- a/arch/ppc64/kernel/bpa_iic.h +++ /dev/null @@ -1,62 +0,0 @@ -#ifndef ASM_CELL_PIC_H -#define ASM_CELL_PIC_H -#ifdef __KERNEL__ -/* - * Mapping of IIC pending bits into per-node - * interrupt numbers. - * - * IRQ FF CC SS PP FF CC SS PP Description - * - * 00-3f 80 02 +0 00 - 80 02 +0 3f South Bridge - * 00-3f 80 02 +b 00 - 80 02 +b 3f South Bridge - * 41-4a 80 00 +1 ** - 80 00 +a ** SPU Class 0 - * 51-5a 80 01 +1 ** - 80 01 +a ** SPU Class 1 - * 61-6a 80 02 +1 ** - 80 02 +a ** SPU Class 2 - * 70-7f C0 ** ** 00 - C0 ** ** 0f IPI - * - * F flags - * C class - * S source - * P Priority - * + node number - * * don't care - * - * A node consists of a Cell Broadband Engine and an optional - * south bridge device providing a maximum of 64 IRQs. - * The south bridge may be connected to either IOIF0 - * or IOIF1. - * Each SPE is represented as three IRQ lines, one per - * interrupt class. - * 16 IRQ numbers are reserved for inter processor - * interruptions, although these are only used in the - * range of the first node. - * - * This scheme needs 128 IRQ numbers per BIF node ID, - * which means that with the total of 512 lines - * available, we can have a maximum of four nodes. - */ - -enum { - IIC_EXT_OFFSET = 0x00, /* Start of south bridge IRQs */ - IIC_NUM_EXT = 0x40, /* Number of south bridge IRQs */ - IIC_SPE_OFFSET = 0x40, /* Start of SPE interrupts */ - IIC_CLASS_STRIDE = 0x10, /* SPE IRQs per class */ - IIC_IPI_OFFSET = 0x70, /* Start of IPI IRQs */ - IIC_NUM_IPIS = 0x10, /* IRQs reserved for IPI */ - IIC_NODE_STRIDE = 0x80, /* Total IRQs per node */ -}; - -extern void iic_init_IRQ(void); -extern int iic_get_irq(struct pt_regs *regs); -extern void iic_cause_IPI(int cpu, int mesg); -extern void iic_request_IPIs(void); -extern void iic_setup_cpu(void); -extern void iic_local_enable(void); -extern void iic_local_disable(void); - - -extern void spider_init_IRQ(void); -extern int spider_get_irq(unsigned long int_pending); - -#endif -#endif /* ASM_CELL_PIC_H */ diff --git a/arch/ppc64/kernel/bpa_iommu.c b/arch/ppc64/kernel/bpa_iommu.c deleted file mode 100644 index 74f999b4ac9e..000000000000 --- a/arch/ppc64/kernel/bpa_iommu.c +++ /dev/null @@ -1,381 +0,0 @@ -/* - * IOMMU implementation for Cell Broadband Processor Architecture - * We just establish a linear mapping at boot by setting all the - * IOPT cache entries in the CPU. - * The mapping functions should be identical to pci_direct_iommu, - * except for the handling of the high order bit that is required - * by the Spider bridge. These should be split into a separate - * file at the point where we get a different bridge chip. - * - * Copyright (C) 2005 IBM Deutschland Entwicklung GmbH, - * Arnd Bergmann - * - * Based on linear mapping - * Copyright (C) 2003 Benjamin Herrenschmidt (benh@kernel.crashing.org) - * - * This program is free software; you can redistribute it and/or - * modify it under the terms of the GNU General Public License - * as published by the Free Software Foundation; either version - * 2 of the License, or (at your option) any later version. - */ - -#undef DEBUG - -#include -#include -#include -#include -#include -#include -#include -#include - -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include - -#include "iommu.h" - -static inline unsigned long -get_iopt_entry(unsigned long real_address, unsigned long ioid, - unsigned long prot) -{ - return (prot & IOPT_PROT_MASK) - | (IOPT_COHERENT) - | (IOPT_ORDER_VC) - | (real_address & IOPT_RPN_MASK) - | (ioid & IOPT_IOID_MASK); -} - -typedef struct { - unsigned long val; -} ioste; - -static inline ioste -mk_ioste(unsigned long val) -{ - ioste ioste = { .val = val, }; - return ioste; -} - -static inline ioste -get_iost_entry(unsigned long iopt_base, unsigned long io_address, unsigned page_size) -{ - unsigned long ps; - unsigned long iostep; - unsigned long nnpt; - unsigned long shift; - - switch (page_size) { - case 0x1000000: - ps = IOST_PS_16M; - nnpt = 0; /* one page per segment */ - shift = 5; /* segment has 16 iopt entries */ - break; - - case 0x100000: - ps = IOST_PS_1M; - nnpt = 0; /* one page per segment */ - shift = 1; /* segment has 256 iopt entries */ - break; - - case 0x10000: - ps = IOST_PS_64K; - nnpt = 0x07; /* 8 pages per io page table */ - shift = 0; /* all entries are used */ - break; - - case 0x1000: - ps = IOST_PS_4K; - nnpt = 0x7f; /* 128 pages per io page table */ - shift = 0; /* all entries are used */ - break; - - default: /* not a known compile time constant */ - { - /* BUILD_BUG_ON() is not usable here */ - extern void __get_iost_entry_bad_page_size(void); - __get_iost_entry_bad_page_size(); - } - break; - } - - iostep = iopt_base + - /* need 8 bytes per iopte */ - (((io_address / page_size * 8) - /* align io page tables on 4k page boundaries */ - << shift) - /* nnpt+1 pages go into each iopt */ - & ~(nnpt << 12)); - - nnpt++; /* this seems to work, but the documentation is not clear - about wether we put nnpt or nnpt-1 into the ioste bits. - In theory, this can't work for 4k pages. */ - return mk_ioste(IOST_VALID_MASK - | (iostep & IOST_PT_BASE_MASK) - | ((nnpt << 5) & IOST_NNPT_MASK) - | (ps & IOST_PS_MASK)); -} - -/* compute the address of an io pte */ -static inline unsigned long -get_ioptep(ioste iost_entry, unsigned long io_address) -{ - unsigned long iopt_base; - unsigned long page_size; - unsigned long page_number; - unsigned long iopt_offset; - - iopt_base = iost_entry.val & IOST_PT_BASE_MASK; - page_size = iost_entry.val & IOST_PS_MASK; - - /* decode page size to compute page number */ - page_number = (io_address & 0x0fffffff) >> (10 + 2 * page_size); - /* page number is an offset into the io page table */ - iopt_offset = (page_number << 3) & 0x7fff8ul; - return iopt_base + iopt_offset; -} - -/* compute the tag field of the iopt cache entry */ -static inline unsigned long -get_ioc_tag(ioste iost_entry, unsigned long io_address) -{ - unsigned long iopte = get_ioptep(iost_entry, io_address); - - return IOPT_VALID_MASK - | ((iopte & 0x00000000000000ff8ul) >> 3) - | ((iopte & 0x0000003fffffc0000ul) >> 9); -} - -/* compute the hashed 6 bit index for the 4-way associative pte cache */ -static inline unsigned long -get_ioc_hash(ioste iost_entry, unsigned long io_address) -{ - unsigned long iopte = get_ioptep(iost_entry, io_address); - - return ((iopte & 0x000000000000001f8ul) >> 3) - ^ ((iopte & 0x00000000000020000ul) >> 17) - ^ ((iopte & 0x00000000000010000ul) >> 15) - ^ ((iopte & 0x00000000000008000ul) >> 13) - ^ ((iopte & 0x00000000000004000ul) >> 11) - ^ ((iopte & 0x00000000000002000ul) >> 9) - ^ ((iopte & 0x00000000000001000ul) >> 7); -} - -/* same as above, but pretend that we have a simpler 1-way associative - pte cache with an 8 bit index */ -static inline unsigned long -get_ioc_hash_1way(ioste iost_entry, unsigned long io_address) -{ - unsigned long iopte = get_ioptep(iost_entry, io_address); - - return ((iopte & 0x000000000000001f8ul) >> 3) - ^ ((iopte & 0x00000000000020000ul) >> 17) - ^ ((iopte & 0x00000000000010000ul) >> 15) - ^ ((iopte & 0x00000000000008000ul) >> 13) - ^ ((iopte & 0x00000000000004000ul) >> 11) - ^ ((iopte & 0x00000000000002000ul) >> 9) - ^ ((iopte & 0x00000000000001000ul) >> 7) - ^ ((iopte & 0x0000000000000c000ul) >> 8); -} - -static inline ioste -get_iost_cache(void __iomem *base, unsigned long index) -{ - unsigned long __iomem *p = (base + IOC_ST_CACHE_DIR); - return mk_ioste(in_be64(&p[index])); -} - -static inline void -set_iost_cache(void __iomem *base, unsigned long index, ioste ste) -{ - unsigned long __iomem *p = (base + IOC_ST_CACHE_DIR); - pr_debug("ioste %02lx was %016lx, store %016lx", index, - get_iost_cache(base, index).val, ste.val); - out_be64(&p[index], ste.val); - pr_debug(" now %016lx\n", get_iost_cache(base, index).val); -} - -static inline unsigned long -get_iopt_cache(void __iomem *base, unsigned long index, unsigned long *tag) -{ - unsigned long __iomem *tags = (void *)(base + IOC_PT_CACHE_DIR); - unsigned long __iomem *p = (void *)(base + IOC_PT_CACHE_REG); - - *tag = tags[index]; - rmb(); - return *p; -} - -static inline void -set_iopt_cache(void __iomem *base, unsigned long index, - unsigned long tag, unsigned long val) -{ - unsigned long __iomem *tags = base + IOC_PT_CACHE_DIR; - unsigned long __iomem *p = base + IOC_PT_CACHE_REG; - pr_debug("iopt %02lx was v%016lx/t%016lx, store v%016lx/t%016lx\n", - index, get_iopt_cache(base, index, &oldtag), oldtag, val, tag); - - out_be64(p, val); - out_be64(&tags[index], tag); -} - -static inline void -set_iost_origin(void __iomem *base) -{ - unsigned long __iomem *p = base + IOC_ST_ORIGIN; - unsigned long origin = IOSTO_ENABLE | IOSTO_SW; - - pr_debug("iost_origin %016lx, now %016lx\n", in_be64(p), origin); - out_be64(p, origin); -} - -static inline void -set_iocmd_config(void __iomem *base) -{ - unsigned long __iomem *p = base + 0xc00; - unsigned long conf; - - conf = in_be64(p); - pr_debug("iost_conf %016lx, now %016lx\n", conf, conf | IOCMD_CONF_TE); - out_be64(p, conf | IOCMD_CONF_TE); -} - -/* FIXME: get these from the device tree */ -#define ioc_base 0x20000511000ull -#define ioc_mmio_base 0x20000510000ull -#define ioid 0x48a -#define iopt_phys_offset (- 0x20000000) /* We have a 512MB offset from the SB */ -#define io_page_size 0x1000000 - -static unsigned long map_iopt_entry(unsigned long address) -{ - switch (address >> 20) { - case 0x600: - address = 0x24020000000ull; /* spider i/o */ - break; - default: - address += iopt_phys_offset; - break; - } - - return get_iopt_entry(address, ioid, IOPT_PROT_RW); -} - -static void iommu_bus_setup_null(struct pci_bus *b) { } -static void iommu_dev_setup_null(struct pci_dev *d) { } - -/* initialize the iommu to support a simple linear mapping - * for each DMA window used by any device. For now, we - * happen to know that there is only one DMA window in use, - * starting at iopt_phys_offset. */ -static void cell_map_iommu(void) -{ - unsigned long address; - void __iomem *base; - ioste ioste; - unsigned long index; - - base = __ioremap(ioc_base, 0x1000, _PAGE_NO_CACHE); - pr_debug("%lx mapped to %p\n", ioc_base, base); - set_iocmd_config(base); - iounmap(base); - - base = __ioremap(ioc_mmio_base, 0x1000, _PAGE_NO_CACHE); - pr_debug("%lx mapped to %p\n", ioc_mmio_base, base); - - set_iost_origin(base); - - for (address = 0; address < 0x100000000ul; address += io_page_size) { - ioste = get_iost_entry(0x10000000000ul, address, io_page_size); - if ((address & 0xfffffff) == 0) /* segment start */ - set_iost_cache(base, address >> 28, ioste); - index = get_ioc_hash_1way(ioste, address); - pr_debug("addr %08lx, index %02lx, ioste %016lx\n", - address, index, ioste.val); - set_iopt_cache(base, - get_ioc_hash_1way(ioste, address), - get_ioc_tag(ioste, address), - map_iopt_entry(address)); - } - iounmap(base); -} - - -static void *cell_alloc_coherent(struct device *hwdev, size_t size, - dma_addr_t *dma_handle, gfp_t flag) -{ - void *ret; - - ret = (void *)__get_free_pages(flag, get_order(size)); - if (ret != NULL) { - memset(ret, 0, size); - *dma_handle = virt_to_abs(ret) | CELL_DMA_VALID; - } - return ret; -} - -static void cell_free_coherent(struct device *hwdev, size_t size, - void *vaddr, dma_addr_t dma_handle) -{ - free_pages((unsigned long)vaddr, get_order(size)); -} - -static dma_addr_t cell_map_single(struct device *hwdev, void *ptr, - size_t size, enum dma_data_direction direction) -{ - return virt_to_abs(ptr) | CELL_DMA_VALID; -} - -static void cell_unmap_single(struct device *hwdev, dma_addr_t dma_addr, - size_t size, enum dma_data_direction direction) -{ -} - -static int cell_map_sg(struct device *hwdev, struct scatterlist *sg, - int nents, enum dma_data_direction direction) -{ - int i; - - for (i = 0; i < nents; i++, sg++) { - sg->dma_address = (page_to_phys(sg->page) + sg->offset) - | CELL_DMA_VALID; - sg->dma_length = sg->length; - } - - return nents; -} - -static void cell_unmap_sg(struct device *hwdev, struct scatterlist *sg, - int nents, enum dma_data_direction direction) -{ -} - -static int cell_dma_supported(struct device *dev, u64 mask) -{ - return mask < 0x100000000ull; -} - -void cell_init_iommu(void) -{ - cell_map_iommu(); - - /* Direct I/O, IOMMU off */ - ppc_md.iommu_dev_setup = iommu_dev_setup_null; - ppc_md.iommu_bus_setup = iommu_bus_setup_null; - - pci_dma_ops.alloc_coherent = cell_alloc_coherent; - pci_dma_ops.free_coherent = cell_free_coherent; - pci_dma_ops.map_single = cell_map_single; - pci_dma_ops.unmap_single = cell_unmap_single; - pci_dma_ops.map_sg = cell_map_sg; - pci_dma_ops.unmap_sg = cell_unmap_sg; - pci_dma_ops.dma_supported = cell_dma_supported; -} diff --git a/arch/ppc64/kernel/bpa_iommu.h b/arch/ppc64/kernel/bpa_iommu.h deleted file mode 100644 index 490d77abfe85..000000000000 --- a/arch/ppc64/kernel/bpa_iommu.h +++ /dev/null @@ -1,65 +0,0 @@ -#ifndef CELL_IOMMU_H -#define CELL_IOMMU_H - -/* some constants */ -enum { - /* segment table entries */ - IOST_VALID_MASK = 0x8000000000000000ul, - IOST_TAG_MASK = 0x3000000000000000ul, - IOST_PT_BASE_MASK = 0x000003fffffff000ul, - IOST_NNPT_MASK = 0x0000000000000fe0ul, - IOST_PS_MASK = 0x000000000000000ful, - - IOST_PS_4K = 0x1, - IOST_PS_64K = 0x3, - IOST_PS_1M = 0x5, - IOST_PS_16M = 0x7, - - /* iopt tag register */ - IOPT_VALID_MASK = 0x0000000200000000ul, - IOPT_TAG_MASK = 0x00000001fffffffful, - - /* iopt cache register */ - IOPT_PROT_MASK = 0xc000000000000000ul, - IOPT_PROT_NONE = 0x0000000000000000ul, - IOPT_PROT_READ = 0x4000000000000000ul, - IOPT_PROT_WRITE = 0x8000000000000000ul, - IOPT_PROT_RW = 0xc000000000000000ul, - IOPT_COHERENT = 0x2000000000000000ul, - - IOPT_ORDER_MASK = 0x1800000000000000ul, - /* order access to same IOID/VC on same address */ - IOPT_ORDER_ADDR = 0x0800000000000000ul, - /* similar, but only after a write access */ - IOPT_ORDER_WRITES = 0x1000000000000000ul, - /* Order all accesses to same IOID/VC */ - IOPT_ORDER_VC = 0x1800000000000000ul, - - IOPT_RPN_MASK = 0x000003fffffff000ul, - IOPT_HINT_MASK = 0x0000000000000800ul, - IOPT_IOID_MASK = 0x00000000000007fful, - - IOSTO_ENABLE = 0x8000000000000000ul, - IOSTO_ORIGIN = 0x000003fffffff000ul, - IOSTO_HW = 0x0000000000000800ul, - IOSTO_SW = 0x0000000000000400ul, - - IOCMD_CONF_TE = 0x0000800000000000ul, - - /* memory mapped registers */ - IOC_PT_CACHE_DIR = 0x000, - IOC_ST_CACHE_DIR = 0x800, - IOC_PT_CACHE_REG = 0x910, - IOC_ST_ORIGIN = 0x918, - IOC_CONF = 0x930, - - /* The high bit needs to be set on every DMA address, - only 2GB are addressable */ - CELL_DMA_VALID = 0x80000000, - CELL_DMA_MASK = 0x7fffffff, -}; - - -void cell_init_iommu(void); - -#endif diff --git a/arch/ppc64/kernel/bpa_setup.c b/arch/ppc64/kernel/bpa_setup.c deleted file mode 100644 index 9a495634d0c2..000000000000 --- a/arch/ppc64/kernel/bpa_setup.c +++ /dev/null @@ -1,141 +0,0 @@ -/* - * linux/arch/powerpc/platforms/cell/cell_setup.c - * - * Copyright (C) 1995 Linus Torvalds - * Adapted from 'alpha' version by Gary Thomas - * Modified by Cort Dougan (cort@cs.nmt.edu) - * Modified by PPC64 Team, IBM Corp - * Modified by Cell Team, IBM Deutschland Entwicklung GmbH - * - * This program is free software; you can redistribute it and/or - * modify it under the terms of the GNU General Public License - * as published by the Free Software Foundation; either version - * 2 of the License, or (at your option) any later version. - */ -#undef DEBUG - -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include - -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include - -#include "interrupt.h" -#include "iommu.h" - -#ifdef DEBUG -#define DBG(fmt...) udbg_printf(fmt) -#else -#define DBG(fmt...) -#endif - -void cell_show_cpuinfo(struct seq_file *m) -{ - struct device_node *root; - const char *model = ""; - - root = of_find_node_by_path("/"); - if (root) - model = get_property(root, "model", NULL); - seq_printf(m, "machine\t\t: CHRP %s\n", model); - of_node_put(root); -} - -static void cell_progress(char *s, unsigned short hex) -{ - printk("*** %04x : %s\n", hex, s ? s : ""); -} - -static void __init cell_setup_arch(void) -{ - ppc_md.init_IRQ = iic_init_IRQ; - ppc_md.get_irq = iic_get_irq; - -#ifdef CONFIG_SMP - smp_init_cell(); -#endif - - /* init to some ~sane value until calibrate_delay() runs */ - loops_per_jiffy = 50000000; - - if (ROOT_DEV == 0) { - printk("No ramdisk, default root is /dev/hda2\n"); - ROOT_DEV = Root_HDA2; - } - - /* Find and initialize PCI host bridges */ - init_pci_config_tokens(); - find_and_init_phbs(); - spider_init_IRQ(); -#ifdef CONFIG_DUMMY_CONSOLE - conswitchp = &dummy_con; -#endif - - mmio_nvram_init(); -} - -/* - * Early initialization. Relocation is on but do not reference unbolted pages - */ -static void __init cell_init_early(void) -{ - DBG(" -> cell_init_early()\n"); - - hpte_init_native(); - - cell_init_iommu(); - - ppc64_interrupt_controller = IC_CELL_PIC; - - DBG(" <- cell_init_early()\n"); -} - - -static int __init cell_probe(int platform) -{ - if (platform != PLATFORM_CELL) - return 0; - - return 1; -} - -struct machdep_calls __initdata cell_md = { - .probe = cell_probe, - .setup_arch = cell_setup_arch, - .init_early = cell_init_early, - .show_cpuinfo = cell_show_cpuinfo, - .restart = rtas_restart, - .power_off = rtas_power_off, - .halt = rtas_halt, - .get_boot_time = rtas_get_boot_time, - .get_rtc_time = rtas_get_rtc_time, - .set_rtc_time = rtas_set_rtc_time, - .calibrate_decr = generic_calibrate_decr, - .progress = cell_progress, -}; diff --git a/arch/ppc64/kernel/spider-pic.c b/arch/ppc64/kernel/spider-pic.c deleted file mode 100644 index e74132188bdf..000000000000 --- a/arch/ppc64/kernel/spider-pic.c +++ /dev/null @@ -1,191 +0,0 @@ -/* - * External Interrupt Controller on Spider South Bridge - * - * (C) Copyright IBM Deutschland Entwicklung GmbH 2005 - * - * Author: Arnd Bergmann - * - * This program is free software; you can redistribute it and/or modify - * it under the terms of the GNU General Public License as published by - * the Free Software Foundation; either version 2, or (at your option) - * any later version. - * - * This program is distributed in the hope that it will be useful, - * but WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the - * GNU General Public License for more details. - * - * You should have received a copy of the GNU General Public License - * along with this program; if not, write to the Free Software - * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA. - */ - -#include -#include - -#include -#include -#include - -#include "interrupt.h" - -/* register layout taken from Spider spec, table 7.4-4 */ -enum { - TIR_DEN = 0x004, /* Detection Enable Register */ - TIR_MSK = 0x084, /* Mask Level Register */ - TIR_EDC = 0x0c0, /* Edge Detection Clear Register */ - TIR_PNDA = 0x100, /* Pending Register A */ - TIR_PNDB = 0x104, /* Pending Register B */ - TIR_CS = 0x144, /* Current Status Register */ - TIR_LCSA = 0x150, /* Level Current Status Register A */ - TIR_LCSB = 0x154, /* Level Current Status Register B */ - TIR_LCSC = 0x158, /* Level Current Status Register C */ - TIR_LCSD = 0x15c, /* Level Current Status Register D */ - TIR_CFGA = 0x200, /* Setting Register A0 */ - TIR_CFGB = 0x204, /* Setting Register B0 */ - /* 0x208 ... 0x3ff Setting Register An/Bn */ - TIR_PPNDA = 0x400, /* Packet Pending Register A */ - TIR_PPNDB = 0x404, /* Packet Pending Register B */ - TIR_PIERA = 0x408, /* Packet Output Error Register A */ - TIR_PIERB = 0x40c, /* Packet Output Error Register B */ - TIR_PIEN = 0x444, /* Packet Output Enable Register */ - TIR_PIPND = 0x454, /* Packet Output Pending Register */ - TIRDID = 0x484, /* Spider Device ID Register */ - REISTIM = 0x500, /* Reissue Command Timeout Time Setting */ - REISTIMEN = 0x504, /* Reissue Command Timeout Setting */ - REISWAITEN = 0x508, /* Reissue Wait Control*/ -}; - -static void __iomem *spider_pics[4]; - -static void __iomem *spider_get_pic(int irq) -{ - int node = irq / IIC_NODE_STRIDE; - irq %= IIC_NODE_STRIDE; - - if (irq >= IIC_EXT_OFFSET && - irq < IIC_EXT_OFFSET + IIC_NUM_EXT && - spider_pics) - return spider_pics[node]; - return NULL; -} - -static int spider_get_nr(unsigned int irq) -{ - return (irq % IIC_NODE_STRIDE) - IIC_EXT_OFFSET; -} - -static void __iomem *spider_get_irq_config(int irq) -{ - void __iomem *pic; - pic = spider_get_pic(irq); - return pic + TIR_CFGA + 8 * spider_get_nr(irq); -} - -static void spider_enable_irq(unsigned int irq) -{ - void __iomem *cfg = spider_get_irq_config(irq); - irq = spider_get_nr(irq); - - out_be32(cfg, in_be32(cfg) | 0x3107000eu); - out_be32(cfg + 4, in_be32(cfg + 4) | 0x00020000u | irq); -} - -static void spider_disable_irq(unsigned int irq) -{ - void __iomem *cfg = spider_get_irq_config(irq); - irq = spider_get_nr(irq); - - out_be32(cfg, in_be32(cfg) & ~0x30000000u); -} - -static unsigned int spider_startup_irq(unsigned int irq) -{ - spider_enable_irq(irq); - return 0; -} - -static void spider_shutdown_irq(unsigned int irq) -{ - spider_disable_irq(irq); -} - -static void spider_end_irq(unsigned int irq) -{ - spider_enable_irq(irq); -} - -static void spider_ack_irq(unsigned int irq) -{ - spider_disable_irq(irq); - iic_local_enable(); -} - -static struct hw_interrupt_type spider_pic = { - .typename = " SPIDER ", - .startup = spider_startup_irq, - .shutdown = spider_shutdown_irq, - .enable = spider_enable_irq, - .disable = spider_disable_irq, - .ack = spider_ack_irq, - .end = spider_end_irq, -}; - - -int spider_get_irq(unsigned long int_pending) -{ - void __iomem *regs = spider_get_pic(int_pending); - unsigned long cs; - int irq; - - cs = in_be32(regs + TIR_CS); - - irq = cs >> 24; - if (irq != 63) - return irq; - - return -1; -} - -void spider_init_IRQ(void) -{ - int node; - struct device_node *dn; - unsigned int *property; - long spiderpic; - int n; - -/* FIXME: detect multiple PICs as soon as the device tree has them */ - for (node = 0; node < 1; node++) { - dn = of_find_node_by_path("/"); - n = prom_n_addr_cells(dn); - property = (unsigned int *) get_property(dn, - "platform-spider-pic", NULL); - - if (!property) - continue; - for (spiderpic = 0; n > 0; --n) - spiderpic = (spiderpic << 32) + *property++; - printk(KERN_DEBUG "SPIDER addr: %lx\n", spiderpic); - spider_pics[node] = __ioremap(spiderpic, 0x800, _PAGE_NO_CACHE); - for (n = 0; n < IIC_NUM_EXT; n++) { - int irq = n + IIC_EXT_OFFSET + node * IIC_NODE_STRIDE; - get_irq_desc(irq)->handler = &spider_pic; - - /* do not mask any interrupts because of level */ - out_be32(spider_pics[node] + TIR_MSK, 0x0); - - /* disable edge detection clear */ - /* out_be32(spider_pics[node] + TIR_EDC, 0x0); */ - - /* enable interrupt packets to be output */ - out_be32(spider_pics[node] + TIR_PIEN, - in_be32(spider_pics[node] + TIR_PIEN) | 0x1); - - /* Enable the interrupt detection enable bit. Do this last! */ - out_be32(spider_pics[node] + TIR_DEN, - in_be32(spider_pics[node] +TIR_DEN) | 0x1); - - } - } -} -- cgit v1.2.3 From a0e60b2033b30a6bb8479629001cf98e58e4079a Mon Sep 17 00:00:00 2001 From: David Gibson Date: Tue, 1 Nov 2005 17:28:10 +1100 Subject: [PATCH] powerpc: Merge bitops.h Here's a revised version. This re-introduces the set_bits() function from ppc64, which I removed because I thought it was unused (it exists on no other arch). In fact it is used in the powermac interrupt code (but not on pSeries). - We use LARXL/STCXL macros to generate the right (32 or 64 bit) instructions, similar to LDL/STL from ppc_asm.h, used in fpu.S - ppc32 previously used a full "sync" barrier at the end of test_and_*_bit(), whereas ppc64 used an "isync". The merged version uses "isync", since I believe that's sufficient. - The ppc64 versions of then minix_*() bitmap functions have changed semantics. Previously on ppc64, these functions were big-endian (that is bit 0 was the LSB in the first 64-bit, big-endian word). On ppc32 (and x86, for that matter, they were little-endian. As far as I can tell, the big-endian usage was simply wrong - I guess no-one ever tried to use minixfs on ppc64. - On ppc32 find_next_bit() and find_next_zero_bit() are no longer inline (they were already out-of-line on ppc64). - For ppc64, sched_find_first_bit() has moved from mmu_context.h to the merged bitops. What it was doing in mmu_context.h in the first place, I have no idea. - The fls() function is now implemented using the cntlzw instruction on ppc64, instead of generic_fls(), as it already was on ppc32. - For ARCH=ppc, this patch requires adding arch/powerpc/lib to the arch/ppc/Makefile. This in turn requires some changes to arch/powerpc/lib/Makefile which didn't correctly handle ARCH=ppc. Built and running on G5. Signed-off-by: David Gibson Signed-off-by: Paul Mackerras --- arch/powerpc/kernel/ppc_ksyms.c | 9 - arch/powerpc/lib/Makefile | 9 +- arch/powerpc/lib/bitops.c | 150 +++++++++++++ arch/ppc/Makefile | 3 +- arch/ppc/kernel/bitops.c | 126 ----------- arch/ppc64/kernel/Makefile | 2 +- arch/ppc64/kernel/bitops.c | 147 ------------- include/asm-powerpc/bitops.h | 437 ++++++++++++++++++++++++++++++++++++++ include/asm-ppc/bitops.h | 460 ---------------------------------------- include/asm-ppc64/bitops.h | 360 ------------------------------- include/asm-ppc64/mmu_context.h | 15 -- 11 files changed, 595 insertions(+), 1123 deletions(-) create mode 100644 arch/powerpc/lib/bitops.c delete mode 100644 arch/ppc/kernel/bitops.c delete mode 100644 arch/ppc64/kernel/bitops.c create mode 100644 include/asm-powerpc/bitops.h delete mode 100644 include/asm-ppc/bitops.h delete mode 100644 include/asm-ppc64/bitops.h (limited to 'arch/ppc64') diff --git a/arch/powerpc/kernel/ppc_ksyms.c b/arch/powerpc/kernel/ppc_ksyms.c index 8bc540337ba0..47d6f7e2ea9f 100644 --- a/arch/powerpc/kernel/ppc_ksyms.c +++ b/arch/powerpc/kernel/ppc_ksyms.c @@ -81,15 +81,6 @@ EXPORT_SYMBOL(_prep_type); EXPORT_SYMBOL(ucSystemType); #endif -#if !defined(__INLINE_BITOPS) -EXPORT_SYMBOL(set_bit); -EXPORT_SYMBOL(clear_bit); -EXPORT_SYMBOL(change_bit); -EXPORT_SYMBOL(test_and_set_bit); -EXPORT_SYMBOL(test_and_clear_bit); -EXPORT_SYMBOL(test_and_change_bit); -#endif /* __INLINE_BITOPS */ - EXPORT_SYMBOL(strcpy); EXPORT_SYMBOL(strncpy); EXPORT_SYMBOL(strcat); diff --git a/arch/powerpc/lib/Makefile b/arch/powerpc/lib/Makefile index dfb33915ad61..34f5c2e074c9 100644 --- a/arch/powerpc/lib/Makefile +++ b/arch/powerpc/lib/Makefile @@ -3,13 +3,14 @@ # ifeq ($(CONFIG_PPC_MERGE),y) -obj-y := string.o +obj-y := string.o strcase.o +obj-$(CONFIG_PPC32) += div64.o copy_32.o checksum_32.o endif -obj-y += strcase.o -obj-$(CONFIG_PPC32) += div64.o copy_32.o checksum_32.o +obj-y += bitops.o obj-$(CONFIG_PPC64) += checksum_64.o copypage_64.o copyuser_64.o \ - memcpy_64.o usercopy_64.o mem_64.o string.o + memcpy_64.o usercopy_64.o mem_64.o string.o \ + strcase.o obj-$(CONFIG_PPC_ISERIES) += e2a.o obj-$(CONFIG_XMON) += sstep.o diff --git a/arch/powerpc/lib/bitops.c b/arch/powerpc/lib/bitops.c new file mode 100644 index 000000000000..b67ce3004ebf --- /dev/null +++ b/arch/powerpc/lib/bitops.c @@ -0,0 +1,150 @@ +#include +#include +#include +#include + +/** + * find_next_bit - find the next set bit in a memory region + * @addr: The address to base the search on + * @offset: The bitnumber to start searching at + * @size: The maximum size to search + */ +unsigned long find_next_bit(const unsigned long *addr, unsigned long size, + unsigned long offset) +{ + const unsigned long *p = addr + BITOP_WORD(offset); + unsigned long result = offset & ~(BITS_PER_LONG-1); + unsigned long tmp; + + if (offset >= size) + return size; + size -= result; + offset %= BITS_PER_LONG; + if (offset) { + tmp = *(p++); + tmp &= (~0UL << offset); + if (size < BITS_PER_LONG) + goto found_first; + if (tmp) + goto found_middle; + size -= BITS_PER_LONG; + result += BITS_PER_LONG; + } + while (size & ~(BITS_PER_LONG-1)) { + if ((tmp = *(p++))) + goto found_middle; + result += BITS_PER_LONG; + size -= BITS_PER_LONG; + } + if (!size) + return result; + tmp = *p; + +found_first: + tmp &= (~0UL >> (64 - size)); + if (tmp == 0UL) /* Are any bits set? */ + return result + size; /* Nope. */ +found_middle: + return result + __ffs(tmp); +} +EXPORT_SYMBOL(find_next_bit); + +/* + * This implementation of find_{first,next}_zero_bit was stolen from + * Linus' asm-alpha/bitops.h. + */ +unsigned long find_next_zero_bit(const unsigned long *addr, unsigned long size, + unsigned long offset) +{ + const unsigned long *p = addr + BITOP_WORD(offset); + unsigned long result = offset & ~(BITS_PER_LONG-1); + unsigned long tmp; + + if (offset >= size) + return size; + size -= result; + offset %= BITS_PER_LONG; + if (offset) { + tmp = *(p++); + tmp |= ~0UL >> (BITS_PER_LONG - offset); + if (size < BITS_PER_LONG) + goto found_first; + if (~tmp) + goto found_middle; + size -= BITS_PER_LONG; + result += BITS_PER_LONG; + } + while (size & ~(BITS_PER_LONG-1)) { + if (~(tmp = *(p++))) + goto found_middle; + result += BITS_PER_LONG; + size -= BITS_PER_LONG; + } + if (!size) + return result; + tmp = *p; + +found_first: + tmp |= ~0UL << size; + if (tmp == ~0UL) /* Are any bits zero? */ + return result + size; /* Nope. */ +found_middle: + return result + ffz(tmp); +} +EXPORT_SYMBOL(find_next_zero_bit); + +static inline unsigned int ext2_ilog2(unsigned int x) +{ + int lz; + + asm("cntlzw %0,%1": "=r"(lz):"r"(x)); + return 31 - lz; +} + +static inline unsigned int ext2_ffz(unsigned int x) +{ + u32 rc; + if ((x = ~x) == 0) + return 32; + rc = ext2_ilog2(x & -x); + return rc; +} + +unsigned long find_next_zero_le_bit(const unsigned long *addr, + unsigned long size, unsigned long offset) +{ + const unsigned int *p = ((const unsigned int *)addr) + (offset >> 5); + unsigned int result = offset & ~31; + unsigned int tmp; + + if (offset >= size) + return size; + size -= result; + offset &= 31; + if (offset) { + tmp = cpu_to_le32p(p++); + tmp |= ~0U >> (32 - offset); /* bug or feature ? */ + if (size < 32) + goto found_first; + if (tmp != ~0) + goto found_middle; + size -= 32; + result += 32; + } + while (size >= 32) { + if ((tmp = cpu_to_le32p(p++)) != ~0) + goto found_middle; + result += 32; + size -= 32; + } + if (!size) + return result; + tmp = cpu_to_le32p(p); +found_first: + tmp |= ~0 << size; + if (tmp == ~0) /* Are any bits zero? */ + return result + size; /* Nope. */ +found_middle: + return result + ext2_ffz(tmp); +} +EXPORT_SYMBOL(find_next_zero_le_bit); diff --git a/arch/ppc/Makefile b/arch/ppc/Makefile index 94d5716fa7c3..e719a4933af1 100644 --- a/arch/ppc/Makefile +++ b/arch/ppc/Makefile @@ -66,7 +66,8 @@ head-$(CONFIG_PPC_FPU) += arch/powerpc/kernel/fpu.o core-y += arch/ppc/kernel/ arch/powerpc/kernel/ \ arch/ppc/platforms/ \ arch/ppc/mm/ arch/ppc/lib/ \ - arch/ppc/syslib/ arch/powerpc/sysdev/ + arch/ppc/syslib/ arch/powerpc/sysdev/ \ + arch/powerpc/lib/ core-$(CONFIG_4xx) += arch/ppc/platforms/4xx/ core-$(CONFIG_83xx) += arch/ppc/platforms/83xx/ core-$(CONFIG_85xx) += arch/ppc/platforms/85xx/ diff --git a/arch/ppc/kernel/bitops.c b/arch/ppc/kernel/bitops.c deleted file mode 100644 index 7f53d193968b..000000000000 --- a/arch/ppc/kernel/bitops.c +++ /dev/null @@ -1,126 +0,0 @@ -/* - * Copyright (C) 1996 Paul Mackerras. - */ - -#include -#include - -/* - * If the bitops are not inlined in bitops.h, they are defined here. - * -- paulus - */ -#if !__INLINE_BITOPS -void set_bit(int nr, volatile void * addr) -{ - unsigned long old; - unsigned long mask = 1 << (nr & 0x1f); - unsigned long *p = ((unsigned long *)addr) + (nr >> 5); - - __asm__ __volatile__(SMP_WMB "\n\ -1: lwarx %0,0,%3 \n\ - or %0,%0,%2 \n" - PPC405_ERR77(0,%3) -" stwcx. %0,0,%3 \n\ - bne 1b" - SMP_MB - : "=&r" (old), "=m" (*p) - : "r" (mask), "r" (p), "m" (*p) - : "cc" ); -} - -void clear_bit(int nr, volatile void *addr) -{ - unsigned long old; - unsigned long mask = 1 << (nr & 0x1f); - unsigned long *p = ((unsigned long *)addr) + (nr >> 5); - - __asm__ __volatile__(SMP_WMB "\n\ -1: lwarx %0,0,%3 \n\ - andc %0,%0,%2 \n" - PPC405_ERR77(0,%3) -" stwcx. %0,0,%3 \n\ - bne 1b" - SMP_MB - : "=&r" (old), "=m" (*p) - : "r" (mask), "r" (p), "m" (*p) - : "cc"); -} - -void change_bit(int nr, volatile void *addr) -{ - unsigned long old; - unsigned long mask = 1 << (nr & 0x1f); - unsigned long *p = ((unsigned long *)addr) + (nr >> 5); - - __asm__ __volatile__(SMP_WMB "\n\ -1: lwarx %0,0,%3 \n\ - xor %0,%0,%2 \n" - PPC405_ERR77(0,%3) -" stwcx. %0,0,%3 \n\ - bne 1b" - SMP_MB - : "=&r" (old), "=m" (*p) - : "r" (mask), "r" (p), "m" (*p) - : "cc"); -} - -int test_and_set_bit(int nr, volatile void *addr) -{ - unsigned int old, t; - unsigned int mask = 1 << (nr & 0x1f); - volatile unsigned int *p = ((volatile unsigned int *)addr) + (nr >> 5); - - __asm__ __volatile__(SMP_WMB "\n\ -1: lwarx %0,0,%4 \n\ - or %1,%0,%3 \n" - PPC405_ERR77(0,%4) -" stwcx. %1,0,%4 \n\ - bne 1b" - SMP_MB - : "=&r" (old), "=&r" (t), "=m" (*p) - : "r" (mask), "r" (p), "m" (*p) - : "cc"); - - return (old & mask) != 0; -} - -int test_and_clear_bit(int nr, volatile void *addr) -{ - unsigned int old, t; - unsigned int mask = 1 << (nr & 0x1f); - volatile unsigned int *p = ((volatile unsigned int *)addr) + (nr >> 5); - - __asm__ __volatile__(SMP_WMB "\n\ -1: lwarx %0,0,%4 \n\ - andc %1,%0,%3 \n" - PPC405_ERR77(0,%4) -" stwcx. %1,0,%4 \n\ - bne 1b" - SMP_MB - : "=&r" (old), "=&r" (t), "=m" (*p) - : "r" (mask), "r" (p), "m" (*p) - : "cc"); - - return (old & mask) != 0; -} - -int test_and_change_bit(int nr, volatile void *addr) -{ - unsigned int old, t; - unsigned int mask = 1 << (nr & 0x1f); - volatile unsigned int *p = ((volatile unsigned int *)addr) + (nr >> 5); - - __asm__ __volatile__(SMP_WMB "\n\ -1: lwarx %0,0,%4 \n\ - xor %1,%0,%3 \n" - PPC405_ERR77(0,%4) -" stwcx. %1,0,%4 \n\ - bne 1b" - SMP_MB - : "=&r" (old), "=&r" (t), "=m" (*p) - : "r" (mask), "r" (p), "m" (*p) - : "cc"); - - return (old & mask) != 0; -} -#endif /* !__INLINE_BITOPS */ diff --git a/arch/ppc64/kernel/Makefile b/arch/ppc64/kernel/Makefile index 247d2fc6a8ed..990df0905c87 100644 --- a/arch/ppc64/kernel/Makefile +++ b/arch/ppc64/kernel/Makefile @@ -13,7 +13,7 @@ endif obj-y += irq.o idle.o dma.o \ signal.o \ - align.o bitops.o pacaData.o \ + align.o pacaData.o \ udbg.o ioctl32.o \ rtc.o \ cpu_setup_power4.o \ diff --git a/arch/ppc64/kernel/bitops.c b/arch/ppc64/kernel/bitops.c deleted file mode 100644 index ae329e8b4acb..000000000000 --- a/arch/ppc64/kernel/bitops.c +++ /dev/null @@ -1,147 +0,0 @@ -/* - * These are too big to be inlined. - */ - -#include -#include -#include -#include - -unsigned long find_next_zero_bit(const unsigned long *addr, unsigned long size, - unsigned long offset) -{ - const unsigned long *p = addr + (offset >> 6); - unsigned long result = offset & ~63UL; - unsigned long tmp; - - if (offset >= size) - return size; - size -= result; - offset &= 63UL; - if (offset) { - tmp = *(p++); - tmp |= ~0UL >> (64 - offset); - if (size < 64) - goto found_first; - if (~tmp) - goto found_middle; - size -= 64; - result += 64; - } - while (size & ~63UL) { - if (~(tmp = *(p++))) - goto found_middle; - result += 64; - size -= 64; - } - if (!size) - return result; - tmp = *p; - -found_first: - tmp |= ~0UL << size; - if (tmp == ~0UL) /* Are any bits zero? */ - return result + size; /* Nope. */ -found_middle: - return result + ffz(tmp); -} - -EXPORT_SYMBOL(find_next_zero_bit); - -unsigned long find_next_bit(const unsigned long *addr, unsigned long size, - unsigned long offset) -{ - const unsigned long *p = addr + (offset >> 6); - unsigned long result = offset & ~63UL; - unsigned long tmp; - - if (offset >= size) - return size; - size -= result; - offset &= 63UL; - if (offset) { - tmp = *(p++); - tmp &= (~0UL << offset); - if (size < 64) - goto found_first; - if (tmp) - goto found_middle; - size -= 64; - result += 64; - } - while (size & ~63UL) { - if ((tmp = *(p++))) - goto found_middle; - result += 64; - size -= 64; - } - if (!size) - return result; - tmp = *p; - -found_first: - tmp &= (~0UL >> (64 - size)); - if (tmp == 0UL) /* Are any bits set? */ - return result + size; /* Nope. */ -found_middle: - return result + __ffs(tmp); -} - -EXPORT_SYMBOL(find_next_bit); - -static inline unsigned int ext2_ilog2(unsigned int x) -{ - int lz; - - asm("cntlzw %0,%1": "=r"(lz):"r"(x)); - return 31 - lz; -} - -static inline unsigned int ext2_ffz(unsigned int x) -{ - u32 rc; - if ((x = ~x) == 0) - return 32; - rc = ext2_ilog2(x & -x); - return rc; -} - -unsigned long find_next_zero_le_bit(const unsigned long *addr, unsigned long size, - unsigned long offset) -{ - const unsigned int *p = ((const unsigned int *)addr) + (offset >> 5); - unsigned int result = offset & ~31; - unsigned int tmp; - - if (offset >= size) - return size; - size -= result; - offset &= 31; - if (offset) { - tmp = cpu_to_le32p(p++); - tmp |= ~0U >> (32 - offset); /* bug or feature ? */ - if (size < 32) - goto found_first; - if (tmp != ~0) - goto found_middle; - size -= 32; - result += 32; - } - while (size >= 32) { - if ((tmp = cpu_to_le32p(p++)) != ~0) - goto found_middle; - result += 32; - size -= 32; - } - if (!size) - return result; - tmp = cpu_to_le32p(p); -found_first: - tmp |= ~0 << size; - if (tmp == ~0) /* Are any bits zero? */ - return result + size; /* Nope. */ -found_middle: - return result + ext2_ffz(tmp); -} - -EXPORT_SYMBOL(find_next_zero_le_bit); diff --git a/include/asm-powerpc/bitops.h b/include/asm-powerpc/bitops.h new file mode 100644 index 000000000000..dc25c53704d5 --- /dev/null +++ b/include/asm-powerpc/bitops.h @@ -0,0 +1,437 @@ +/* + * PowerPC atomic bit operations. + * + * Merged version by David Gibson . + * Based on ppc64 versions by: Dave Engebretsen, Todd Inglett, Don + * Reed, Pat McCarthy, Peter Bergner, Anton Blanchard. They + * originally took it from the ppc32 code. + * + * Within a word, bits are numbered LSB first. Lot's of places make + * this assumption by directly testing bits with (val & (1< 1 word) bitmaps on a + * big-endian system because, unlike little endian, the number of each + * bit depends on the word size. + * + * The bitop functions are defined to work on unsigned longs, so for a + * ppc64 system the bits end up numbered: + * |63..............0|127............64|191...........128|255...........196| + * and on ppc32: + * |31.....0|63....31|95....64|127...96|159..128|191..160|223..192|255..224| + * + * There are a few little-endian macros used mostly for filesystem + * bitmaps, these work on similar bit arrays layouts, but + * byte-oriented: + * |7...0|15...8|23...16|31...24|39...32|47...40|55...48|63...56| + * + * The main difference is that bit 3-5 (64b) or 3-4 (32b) in the bit + * number field needs to be reversed compared to the big-endian bit + * fields. This can be achieved by XOR with 0x38 (64b) or 0x18 (32b). + * + * This program is free software; you can redistribute it and/or + * modify it under the terms of the GNU General Public License + * as published by the Free Software Foundation; either version + * 2 of the License, or (at your option) any later version. + */ + +#ifndef _ASM_POWERPC_BITOPS_H +#define _ASM_POWERPC_BITOPS_H + +#ifdef __KERNEL__ + +#include +#include +#include + +/* + * clear_bit doesn't imply a memory barrier + */ +#define smp_mb__before_clear_bit() smp_mb() +#define smp_mb__after_clear_bit() smp_mb() + +#define BITOP_MASK(nr) (1UL << ((nr) % BITS_PER_LONG)) +#define BITOP_WORD(nr) ((nr) / BITS_PER_LONG) +#define BITOP_LE_SWIZZLE ((BITS_PER_LONG-1) & ~0x7) + +#ifdef CONFIG_PPC64 +#define LARXL "ldarx" +#define STCXL "stdcx." +#define CNTLZL "cntlzd" +#else +#define LARXL "lwarx" +#define STCXL "stwcx." +#define CNTLZL "cntlzw" +#endif + +static __inline__ void set_bit(int nr, volatile unsigned long *addr) +{ + unsigned long old; + unsigned long mask = BITOP_MASK(nr); + unsigned long *p = ((unsigned long *)addr) + BITOP_WORD(nr); + + __asm__ __volatile__( +"1:" LARXL " %0,0,%3 # set_bit\n" + "or %0,%0,%2\n" + PPC405_ERR77(0,%3) + STCXL " %0,0,%3\n" + "bne- 1b" + : "=&r"(old), "=m"(*p) + : "r"(mask), "r"(p), "m"(*p) + : "cc" ); +} + +static __inline__ void clear_bit(int nr, volatile unsigned long *addr) +{ + unsigned long old; + unsigned long mask = BITOP_MASK(nr); + unsigned long *p = ((unsigned long *)addr) + BITOP_WORD(nr); + + __asm__ __volatile__( +"1:" LARXL " %0,0,%3 # set_bit\n" + "andc %0,%0,%2\n" + PPC405_ERR77(0,%3) + STCXL " %0,0,%3\n" + "bne- 1b" + : "=&r"(old), "=m"(*p) + : "r"(mask), "r"(p), "m"(*p) + : "cc" ); +} + +static __inline__ void change_bit(int nr, volatile unsigned long *addr) +{ + unsigned long old; + unsigned long mask = BITOP_MASK(nr); + unsigned long *p = ((unsigned long *)addr) + BITOP_WORD(nr); + + __asm__ __volatile__( +"1:" LARXL " %0,0,%3 # set_bit\n" + "xor %0,%0,%2\n" + PPC405_ERR77(0,%3) + STCXL " %0,0,%3\n" + "bne- 1b" + : "=&r"(old), "=m"(*p) + : "r"(mask), "r"(p), "m"(*p) + : "cc" ); +} + +static __inline__ int test_and_set_bit(unsigned long nr, + volatile unsigned long *addr) +{ + unsigned long old, t; + unsigned long mask = BITOP_MASK(nr); + unsigned long *p = ((unsigned long *)addr) + BITOP_WORD(nr); + + __asm__ __volatile__( + EIEIO_ON_SMP +"1:" LARXL " %0,0,%3 # test_and_set_bit\n" + "or %1,%0,%2 \n" + PPC405_ERR77(0,%3) + STCXL " %1,0,%3 \n" + "bne- 1b" + ISYNC_ON_SMP + : "=&r" (old), "=&r" (t) + : "r" (mask), "r" (p) + : "cc", "memory"); + + return (old & mask) != 0; +} + +static __inline__ int test_and_clear_bit(unsigned long nr, + volatile unsigned long *addr) +{ + unsigned long old, t; + unsigned long mask = BITOP_MASK(nr); + unsigned long *p = ((unsigned long *)addr) + BITOP_WORD(nr); + + __asm__ __volatile__( + EIEIO_ON_SMP +"1:" LARXL " %0,0,%3 # test_and_clear_bit\n" + "andc %1,%0,%2 \n" + PPC405_ERR77(0,%3) + STCXL " %1,0,%3 \n" + "bne- 1b" + ISYNC_ON_SMP + : "=&r" (old), "=&r" (t) + : "r" (mask), "r" (p) + : "cc", "memory"); + + return (old & mask) != 0; +} + +static __inline__ int test_and_change_bit(unsigned long nr, + volatile unsigned long *addr) +{ + unsigned long old, t; + unsigned long mask = BITOP_MASK(nr); + unsigned long *p = ((unsigned long *)addr) + BITOP_WORD(nr); + + __asm__ __volatile__( + EIEIO_ON_SMP +"1:" LARXL " %0,0,%3 # test_and_change_bit\n" + "xor %1,%0,%2 \n" + PPC405_ERR77(0,%3) + STCXL " %1,0,%3 \n" + "bne- 1b" + ISYNC_ON_SMP + : "=&r" (old), "=&r" (t) + : "r" (mask), "r" (p) + : "cc", "memory"); + + return (old & mask) != 0; +} + +static __inline__ void set_bits(unsigned long mask, unsigned long *addr) +{ + unsigned long old; + + __asm__ __volatile__( +"1:" LARXL " %0,0,%3 # set_bit\n" + "or %0,%0,%2\n" + STCXL " %0,0,%3\n" + "bne- 1b" + : "=&r" (old), "=m" (*addr) + : "r" (mask), "r" (addr), "m" (*addr) + : "cc"); +} + +/* Non-atomic versions */ +static __inline__ int test_bit(unsigned long nr, + __const__ volatile unsigned long *addr) +{ + return 1UL & (addr[BITOP_WORD(nr)] >> (nr & (BITS_PER_LONG-1))); +} + +static __inline__ void __set_bit(unsigned long nr, + volatile unsigned long *addr) +{ + unsigned long mask = BITOP_MASK(nr); + unsigned long *p = ((unsigned long *)addr) + BITOP_WORD(nr); + + *p |= mask; +} + +static __inline__ void __clear_bit(unsigned long nr, + volatile unsigned long *addr) +{ + unsigned long mask = BITOP_MASK(nr); + unsigned long *p = ((unsigned long *)addr) + BITOP_WORD(nr); + + *p &= ~mask; +} + +static __inline__ void __change_bit(unsigned long nr, + volatile unsigned long *addr) +{ + unsigned long mask = BITOP_MASK(nr); + unsigned long *p = ((unsigned long *)addr) + BITOP_WORD(nr); + + *p ^= mask; +} + +static __inline__ int __test_and_set_bit(unsigned long nr, + volatile unsigned long *addr) +{ + unsigned long mask = BITOP_MASK(nr); + unsigned long *p = ((unsigned long *)addr) + BITOP_WORD(nr); + unsigned long old = *p; + + *p = old | mask; + return (old & mask) != 0; +} + +static __inline__ int __test_and_clear_bit(unsigned long nr, + volatile unsigned long *addr) +{ + unsigned long mask = BITOP_MASK(nr); + unsigned long *p = ((unsigned long *)addr) + BITOP_WORD(nr); + unsigned long old = *p; + + *p = old & ~mask; + return (old & mask) != 0; +} + +static __inline__ int __test_and_change_bit(unsigned long nr, + volatile unsigned long *addr) +{ + unsigned long mask = BITOP_MASK(nr); + unsigned long *p = ((unsigned long *)addr) + BITOP_WORD(nr); + unsigned long old = *p; + + *p = old ^ mask; + return (old & mask) != 0; +} + +/* + * Return the zero-based bit position (LE, not IBM bit numbering) of + * the most significant 1-bit in a double word. + */ +static __inline__ int __ilog2(unsigned long x) +{ + int lz; + + asm (CNTLZL " %0,%1" : "=r" (lz) : "r" (x)); + return BITS_PER_LONG - 1 - lz; +} + +/* + * Determines the bit position of the least significant 0 bit in the + * specified double word. The returned bit position will be + * zero-based, starting from the right side (63/31 - 0). + */ +static __inline__ unsigned long ffz(unsigned long x) +{ + /* no zero exists anywhere in the 8 byte area. */ + if ((x = ~x) == 0) + return BITS_PER_LONG; + + /* + * Calculate the bit position of the least signficant '1' bit in x + * (since x has been changed this will actually be the least signficant + * '0' bit in * the original x). Note: (x & -x) gives us a mask that + * is the least significant * (RIGHT-most) 1-bit of the value in x. + */ + return __ilog2(x & -x); +} + +static __inline__ int __ffs(unsigned long x) +{ + return __ilog2(x & -x); +} + +/* + * ffs: find first bit set. This is defined the same way as + * the libc and compiler builtin ffs routines, therefore + * differs in spirit from the above ffz (man ffs). + */ +static __inline__ int ffs(int x) +{ + unsigned long i = (unsigned long)x; + return __ilog2(i & -i) + 1; +} + +/* + * fls: find last (most-significant) bit set. + * Note fls(0) = 0, fls(1) = 1, fls(0x80000000) = 32. + */ +static __inline__ int fls(unsigned int x) +{ + int lz; + + asm ("cntlzw %0,%1" : "=r" (lz) : "r" (x)); + return 32 - lz; +} + +/* + * hweightN: returns the hamming weight (i.e. the number + * of bits set) of a N-bit word + */ +#define hweight64(x) generic_hweight64(x) +#define hweight32(x) generic_hweight32(x) +#define hweight16(x) generic_hweight16(x) +#define hweight8(x) generic_hweight8(x) + +#define find_first_zero_bit(addr, size) find_next_zero_bit((addr), (size), 0) +unsigned long find_next_zero_bit(const unsigned long *addr, + unsigned long size, unsigned long offset); +/** + * find_first_bit - find the first set bit in a memory region + * @addr: The address to start the search at + * @size: The maximum size to search + * + * Returns the bit-number of the first set bit, not the number of the byte + * containing a bit. + */ +#define find_first_bit(addr, size) find_next_bit((addr), (size), 0) +unsigned long find_next_bit(const unsigned long *addr, + unsigned long size, unsigned long offset); + +/* Little-endian versions */ + +static __inline__ int test_le_bit(unsigned long nr, + __const__ unsigned long *addr) +{ + __const__ unsigned char *tmp = (__const__ unsigned char *) addr; + return (tmp[nr >> 3] >> (nr & 7)) & 1; +} + +#define __set_le_bit(nr, addr) \ + __set_bit((nr) ^ BITOP_LE_SWIZZLE, (addr)) +#define __clear_le_bit(nr, addr) \ + __clear_bit((nr) ^ BITOP_LE_SWIZZLE, (addr)) + +#define test_and_set_le_bit(nr, addr) \ + test_and_set_bit((nr) ^ BITOP_LE_SWIZZLE, (addr)) +#define test_and_clear_le_bit(nr, addr) \ + test_and_clear_bit((nr) ^ BITOP_LE_SWIZZLE, (addr)) + +#define __test_and_set_le_bit(nr, addr) \ + __test_and_set_bit((nr) ^ BITOP_LE_SWIZZLE, (addr)) +#define __test_and_clear_le_bit(nr, addr) \ + __test_and_clear_bit((nr) ^ BITOP_LE_SWIZZLE, (addr)) + +#define find_first_zero_le_bit(addr, size) find_next_zero_le_bit((addr), (size), 0) +unsigned long find_next_zero_le_bit(const unsigned long *addr, + unsigned long size, unsigned long offset); + +/* Bitmap functions for the ext2 filesystem */ + +#define ext2_set_bit(nr,addr) \ + __test_and_set_le_bit((nr), (unsigned long*)addr) +#define ext2_clear_bit(nr, addr) \ + __test_and_clear_le_bit((nr), (unsigned long*)addr) + +#define ext2_set_bit_atomic(lock, nr, addr) \ + test_and_set_le_bit((nr), (unsigned long*)addr) +#define ext2_clear_bit_atomic(lock, nr, addr) \ + test_and_clear_le_bit((nr), (unsigned long*)addr) + +#define ext2_test_bit(nr, addr) test_le_bit((nr),(unsigned long*)addr) + +#define ext2_find_first_zero_bit(addr, size) \ + find_first_zero_le_bit((unsigned long*)addr, size) +#define ext2_find_next_zero_bit(addr, size, off) \ + find_next_zero_le_bit((unsigned long*)addr, size, off) + +/* Bitmap functions for the minix filesystem. */ + +#define minix_test_and_set_bit(nr,addr) \ + __test_and_set_le_bit(nr, (unsigned long *)addr) +#define minix_set_bit(nr,addr) \ + __set_le_bit(nr, (unsigned long *)addr) +#define minix_test_and_clear_bit(nr,addr) \ + __test_and_clear_le_bit(nr, (unsigned long *)addr) +#define minix_test_bit(nr,addr) \ + test_le_bit(nr, (unsigned long *)addr) + +#define minix_find_first_zero_bit(addr,size) \ + find_first_zero_le_bit((unsigned long *)addr, size) + +/* + * Every architecture must define this function. It's the fastest + * way of searching a 140-bit bitmap where the first 100 bits are + * unlikely to be set. It's guaranteed that at least one of the 140 + * bits is cleared. + */ +static inline int sched_find_first_bit(const unsigned long *b) +{ +#ifdef CONFIG_PPC64 + if (unlikely(b[0])) + return __ffs(b[0]); + if (unlikely(b[1])) + return __ffs(b[1]) + 64; + return __ffs(b[2]) + 128; +#else + if (unlikely(b[0])) + return __ffs(b[0]); + if (unlikely(b[1])) + return __ffs(b[1]) + 32; + if (unlikely(b[2])) + return __ffs(b[2]) + 64; + if (b[3]) + return __ffs(b[3]) + 96; + return __ffs(b[4]) + 128; +#endif +} + +#endif /* __KERNEL__ */ + +#endif /* _ASM_POWERPC_BITOPS_H */ diff --git a/include/asm-ppc/bitops.h b/include/asm-ppc/bitops.h deleted file mode 100644 index e30f536fd830..000000000000 --- a/include/asm-ppc/bitops.h +++ /dev/null @@ -1,460 +0,0 @@ -/* - * bitops.h: Bit string operations on the ppc - */ - -#ifdef __KERNEL__ -#ifndef _PPC_BITOPS_H -#define _PPC_BITOPS_H - -#include -#include -#include -#include - -/* - * The test_and_*_bit operations are taken to imply a memory barrier - * on SMP systems. - */ -#ifdef CONFIG_SMP -#define SMP_WMB "eieio\n" -#define SMP_MB "\nsync" -#else -#define SMP_WMB -#define SMP_MB -#endif /* CONFIG_SMP */ - -static __inline__ void set_bit(int nr, volatile unsigned long * addr) -{ - unsigned long old; - unsigned long mask = 1 << (nr & 0x1f); - unsigned long *p = ((unsigned long *)addr) + (nr >> 5); - - __asm__ __volatile__("\n\ -1: lwarx %0,0,%3 \n\ - or %0,%0,%2 \n" - PPC405_ERR77(0,%3) -" stwcx. %0,0,%3 \n\ - bne- 1b" - : "=&r" (old), "=m" (*p) - : "r" (mask), "r" (p), "m" (*p) - : "cc" ); -} - -/* - * non-atomic version - */ -static __inline__ void __set_bit(int nr, volatile unsigned long *addr) -{ - unsigned long mask = 1 << (nr & 0x1f); - unsigned long *p = ((unsigned long *)addr) + (nr >> 5); - - *p |= mask; -} - -/* - * clear_bit doesn't imply a memory barrier - */ -#define smp_mb__before_clear_bit() smp_mb() -#define smp_mb__after_clear_bit() smp_mb() - -static __inline__ void clear_bit(int nr, volatile unsigned long *addr) -{ - unsigned long old; - unsigned long mask = 1 << (nr & 0x1f); - unsigned long *p = ((unsigned long *)addr) + (nr >> 5); - - __asm__ __volatile__("\n\ -1: lwarx %0,0,%3 \n\ - andc %0,%0,%2 \n" - PPC405_ERR77(0,%3) -" stwcx. %0,0,%3 \n\ - bne- 1b" - : "=&r" (old), "=m" (*p) - : "r" (mask), "r" (p), "m" (*p) - : "cc"); -} - -/* - * non-atomic version - */ -static __inline__ void __clear_bit(int nr, volatile unsigned long *addr) -{ - unsigned long mask = 1 << (nr & 0x1f); - unsigned long *p = ((unsigned long *)addr) + (nr >> 5); - - *p &= ~mask; -} - -static __inline__ void change_bit(int nr, volatile unsigned long *addr) -{ - unsigned long old; - unsigned long mask = 1 << (nr & 0x1f); - unsigned long *p = ((unsigned long *)addr) + (nr >> 5); - - __asm__ __volatile__("\n\ -1: lwarx %0,0,%3 \n\ - xor %0,%0,%2 \n" - PPC405_ERR77(0,%3) -" stwcx. %0,0,%3 \n\ - bne- 1b" - : "=&r" (old), "=m" (*p) - : "r" (mask), "r" (p), "m" (*p) - : "cc"); -} - -/* - * non-atomic version - */ -static __inline__ void __change_bit(int nr, volatile unsigned long *addr) -{ - unsigned long mask = 1 << (nr & 0x1f); - unsigned long *p = ((unsigned long *)addr) + (nr >> 5); - - *p ^= mask; -} - -/* - * test_and_*_bit do imply a memory barrier (?) - */ -static __inline__ int test_and_set_bit(int nr, volatile unsigned long *addr) -{ - unsigned int old, t; - unsigned int mask = 1 << (nr & 0x1f); - volatile unsigned int *p = ((volatile unsigned int *)addr) + (nr >> 5); - - __asm__ __volatile__(SMP_WMB "\n\ -1: lwarx %0,0,%4 \n\ - or %1,%0,%3 \n" - PPC405_ERR77(0,%4) -" stwcx. %1,0,%4 \n\ - bne 1b" - SMP_MB - : "=&r" (old), "=&r" (t), "=m" (*p) - : "r" (mask), "r" (p), "m" (*p) - : "cc", "memory"); - - return (old & mask) != 0; -} - -/* - * non-atomic version - */ -static __inline__ int __test_and_set_bit(int nr, volatile unsigned long *addr) -{ - unsigned long mask = 1 << (nr & 0x1f); - unsigned long *p = ((unsigned long *)addr) + (nr >> 5); - unsigned long old = *p; - - *p = old | mask; - return (old & mask) != 0; -} - -static __inline__ int test_and_clear_bit(int nr, volatile unsigned long *addr) -{ - unsigned int old, t; - unsigned int mask = 1 << (nr & 0x1f); - volatile unsigned int *p = ((volatile unsigned int *)addr) + (nr >> 5); - - __asm__ __volatile__(SMP_WMB "\n\ -1: lwarx %0,0,%4 \n\ - andc %1,%0,%3 \n" - PPC405_ERR77(0,%4) -" stwcx. %1,0,%4 \n\ - bne 1b" - SMP_MB - : "=&r" (old), "=&r" (t), "=m" (*p) - : "r" (mask), "r" (p), "m" (*p) - : "cc", "memory"); - - return (old & mask) != 0; -} - -/* - * non-atomic version - */ -static __inline__ int __test_and_clear_bit(int nr, volatile unsigned long *addr) -{ - unsigned long mask = 1 << (nr & 0x1f); - unsigned long *p = ((unsigned long *)addr) + (nr >> 5); - unsigned long old = *p; - - *p = old & ~mask; - return (old & mask) != 0; -} - -static __inline__ int test_and_change_bit(int nr, volatile unsigned long *addr) -{ - unsigned int old, t; - unsigned int mask = 1 << (nr & 0x1f); - volatile unsigned int *p = ((volatile unsigned int *)addr) + (nr >> 5); - - __asm__ __volatile__(SMP_WMB "\n\ -1: lwarx %0,0,%4 \n\ - xor %1,%0,%3 \n" - PPC405_ERR77(0,%4) -" stwcx. %1,0,%4 \n\ - bne 1b" - SMP_MB - : "=&r" (old), "=&r" (t), "=m" (*p) - : "r" (mask), "r" (p), "m" (*p) - : "cc", "memory"); - - return (old & mask) != 0; -} - -/* - * non-atomic version - */ -static __inline__ int __test_and_change_bit(int nr, volatile unsigned long *addr) -{ - unsigned long mask = 1 << (nr & 0x1f); - unsigned long *p = ((unsigned long *)addr) + (nr >> 5); - unsigned long old = *p; - - *p = old ^ mask; - return (old & mask) != 0; -} - -static __inline__ int test_bit(int nr, __const__ volatile unsigned long *addr) -{ - return ((addr[nr >> 5] >> (nr & 0x1f)) & 1) != 0; -} - -/* Return the bit position of the most significant 1 bit in a word */ -static __inline__ int __ilog2(unsigned long x) -{ - int lz; - - asm ("cntlzw %0,%1" : "=r" (lz) : "r" (x)); - return 31 - lz; -} - -static __inline__ int ffz(unsigned long x) -{ - if ((x = ~x) == 0) - return 32; - return __ilog2(x & -x); -} - -static inline int __ffs(unsigned long x) -{ - return __ilog2(x & -x); -} - -/* - * ffs: find first bit set. This is defined the same way as - * the libc and compiler builtin ffs routines, therefore - * differs in spirit from the above ffz (man ffs). - */ -static __inline__ int ffs(int x) -{ - return __ilog2(x & -x) + 1; -} - -/* - * fls: find last (most-significant) bit set. - * Note fls(0) = 0, fls(1) = 1, fls(0x80000000) = 32. - */ -static __inline__ int fls(unsigned int x) -{ - int lz; - - asm ("cntlzw %0,%1" : "=r" (lz) : "r" (x)); - return 32 - lz; -} - -/* - * hweightN: returns the hamming weight (i.e. the number - * of bits set) of a N-bit word - */ - -#define hweight32(x) generic_hweight32(x) -#define hweight16(x) generic_hweight16(x) -#define hweight8(x) generic_hweight8(x) - -/* - * Find the first bit set in a 140-bit bitmap. - * The first 100 bits are unlikely to be set. - */ -static inline int sched_find_first_bit(const unsigned long *b) -{ - if (unlikely(b[0])) - return __ffs(b[0]); - if (unlikely(b[1])) - return __ffs(b[1]) + 32; - if (unlikely(b[2])) - return __ffs(b[2]) + 64; - if (b[3]) - return __ffs(b[3]) + 96; - return __ffs(b[4]) + 128; -} - -/** - * find_next_bit - find the next set bit in a memory region - * @addr: The address to base the search on - * @offset: The bitnumber to start searching at - * @size: The maximum size to search - */ -static __inline__ unsigned long find_next_bit(const unsigned long *addr, - unsigned long size, unsigned long offset) -{ - unsigned int *p = ((unsigned int *) addr) + (offset >> 5); - unsigned int result = offset & ~31UL; - unsigned int tmp; - - if (offset >= size) - return size; - size -= result; - offset &= 31UL; - if (offset) { - tmp = *p++; - tmp &= ~0UL << offset; - if (size < 32) - goto found_first; - if (tmp) - goto found_middle; - size -= 32; - result += 32; - } - while (size >= 32) { - if ((tmp = *p++) != 0) - goto found_middle; - result += 32; - size -= 32; - } - if (!size) - return result; - tmp = *p; - -found_first: - tmp &= ~0UL >> (32 - size); - if (tmp == 0UL) /* Are any bits set? */ - return result + size; /* Nope. */ -found_middle: - return result + __ffs(tmp); -} - -/** - * find_first_bit - find the first set bit in a memory region - * @addr: The address to start the search at - * @size: The maximum size to search - * - * Returns the bit-number of the first set bit, not the number of the byte - * containing a bit. - */ -#define find_first_bit(addr, size) \ - find_next_bit((addr), (size), 0) - -/* - * This implementation of find_{first,next}_zero_bit was stolen from - * Linus' asm-alpha/bitops.h. - */ -#define find_first_zero_bit(addr, size) \ - find_next_zero_bit((addr), (size), 0) - -static __inline__ unsigned long find_next_zero_bit(const unsigned long *addr, - unsigned long size, unsigned long offset) -{ - unsigned int * p = ((unsigned int *) addr) + (offset >> 5); - unsigned int result = offset & ~31UL; - unsigned int tmp; - - if (offset >= size) - return size; - size -= result; - offset &= 31UL; - if (offset) { - tmp = *p++; - tmp |= ~0UL >> (32-offset); - if (size < 32) - goto found_first; - if (tmp != ~0U) - goto found_middle; - size -= 32; - result += 32; - } - while (size >= 32) { - if ((tmp = *p++) != ~0U) - goto found_middle; - result += 32; - size -= 32; - } - if (!size) - return result; - tmp = *p; -found_first: - tmp |= ~0UL << size; - if (tmp == ~0UL) /* Are any bits zero? */ - return result + size; /* Nope. */ -found_middle: - return result + ffz(tmp); -} - - -#define ext2_set_bit(nr, addr) __test_and_set_bit((nr) ^ 0x18, (unsigned long *)(addr)) -#define ext2_set_bit_atomic(lock, nr, addr) test_and_set_bit((nr) ^ 0x18, (unsigned long *)(addr)) -#define ext2_clear_bit(nr, addr) __test_and_clear_bit((nr) ^ 0x18, (unsigned long *)(addr)) -#define ext2_clear_bit_atomic(lock, nr, addr) test_and_clear_bit((nr) ^ 0x18, (unsigned long *)(addr)) - -static __inline__ int ext2_test_bit(int nr, __const__ void * addr) -{ - __const__ unsigned char *ADDR = (__const__ unsigned char *) addr; - - return (ADDR[nr >> 3] >> (nr & 7)) & 1; -} - -/* - * This implementation of ext2_find_{first,next}_zero_bit was stolen from - * Linus' asm-alpha/bitops.h and modified for a big-endian machine. - */ - -#define ext2_find_first_zero_bit(addr, size) \ - ext2_find_next_zero_bit((addr), (size), 0) - -static __inline__ unsigned long ext2_find_next_zero_bit(const void *addr, - unsigned long size, unsigned long offset) -{ - unsigned int *p = ((unsigned int *) addr) + (offset >> 5); - unsigned int result = offset & ~31UL; - unsigned int tmp; - - if (offset >= size) - return size; - size -= result; - offset &= 31UL; - if (offset) { - tmp = cpu_to_le32p(p++); - tmp |= ~0UL >> (32-offset); - if (size < 32) - goto found_first; - if (tmp != ~0U) - goto found_middle; - size -= 32; - result += 32; - } - while (size >= 32) { - if ((tmp = cpu_to_le32p(p++)) != ~0U) - goto found_middle; - result += 32; - size -= 32; - } - if (!size) - return result; - tmp = cpu_to_le32p(p); -found_first: - tmp |= ~0U << size; - if (tmp == ~0UL) /* Are any bits zero? */ - return result + size; /* Nope. */ -found_middle: - return result + ffz(tmp); -} - -/* Bitmap functions for the minix filesystem. */ -#define minix_test_and_set_bit(nr,addr) ext2_set_bit(nr,addr) -#define minix_set_bit(nr,addr) ((void)ext2_set_bit(nr,addr)) -#define minix_test_and_clear_bit(nr,addr) ext2_clear_bit(nr,addr) -#define minix_test_bit(nr,addr) ext2_test_bit(nr,addr) -#define minix_find_first_zero_bit(addr,size) ext2_find_first_zero_bit(addr,size) - -#endif /* _PPC_BITOPS_H */ -#endif /* __KERNEL__ */ diff --git a/include/asm-ppc64/bitops.h b/include/asm-ppc64/bitops.h deleted file mode 100644 index dbfa42ef4a99..000000000000 --- a/include/asm-ppc64/bitops.h +++ /dev/null @@ -1,360 +0,0 @@ -/* - * PowerPC64 atomic bit operations. - * Dave Engebretsen, Todd Inglett, Don Reed, Pat McCarthy, Peter Bergner, - * Anton Blanchard - * - * Originally taken from the 32b PPC code. Modified to use 64b values for - * the various counters & memory references. - * - * Bitops are odd when viewed on big-endian systems. They were designed - * on little endian so the size of the bitset doesn't matter (low order bytes - * come first) as long as the bit in question is valid. - * - * Bits are "tested" often using the C expression (val & (1< - -/* - * clear_bit doesn't imply a memory barrier - */ -#define smp_mb__before_clear_bit() smp_mb() -#define smp_mb__after_clear_bit() smp_mb() - -static __inline__ int test_bit(unsigned long nr, __const__ volatile unsigned long *addr) -{ - return (1UL & (addr[nr >> 6] >> (nr & 63))); -} - -static __inline__ void set_bit(unsigned long nr, volatile unsigned long *addr) -{ - unsigned long old; - unsigned long mask = 1UL << (nr & 0x3f); - unsigned long *p = ((unsigned long *)addr) + (nr >> 6); - - __asm__ __volatile__( -"1: ldarx %0,0,%3 # set_bit\n\ - or %0,%0,%2\n\ - stdcx. %0,0,%3\n\ - bne- 1b" - : "=&r" (old), "=m" (*p) - : "r" (mask), "r" (p), "m" (*p) - : "cc"); -} - -static __inline__ void clear_bit(unsigned long nr, volatile unsigned long *addr) -{ - unsigned long old; - unsigned long mask = 1UL << (nr & 0x3f); - unsigned long *p = ((unsigned long *)addr) + (nr >> 6); - - __asm__ __volatile__( -"1: ldarx %0,0,%3 # clear_bit\n\ - andc %0,%0,%2\n\ - stdcx. %0,0,%3\n\ - bne- 1b" - : "=&r" (old), "=m" (*p) - : "r" (mask), "r" (p), "m" (*p) - : "cc"); -} - -static __inline__ void change_bit(unsigned long nr, volatile unsigned long *addr) -{ - unsigned long old; - unsigned long mask = 1UL << (nr & 0x3f); - unsigned long *p = ((unsigned long *)addr) + (nr >> 6); - - __asm__ __volatile__( -"1: ldarx %0,0,%3 # change_bit\n\ - xor %0,%0,%2\n\ - stdcx. %0,0,%3\n\ - bne- 1b" - : "=&r" (old), "=m" (*p) - : "r" (mask), "r" (p), "m" (*p) - : "cc"); -} - -static __inline__ int test_and_set_bit(unsigned long nr, volatile unsigned long *addr) -{ - unsigned long old, t; - unsigned long mask = 1UL << (nr & 0x3f); - unsigned long *p = ((unsigned long *)addr) + (nr >> 6); - - __asm__ __volatile__( - EIEIO_ON_SMP -"1: ldarx %0,0,%3 # test_and_set_bit\n\ - or %1,%0,%2 \n\ - stdcx. %1,0,%3 \n\ - bne- 1b" - ISYNC_ON_SMP - : "=&r" (old), "=&r" (t) - : "r" (mask), "r" (p) - : "cc", "memory"); - - return (old & mask) != 0; -} - -static __inline__ int test_and_clear_bit(unsigned long nr, volatile unsigned long *addr) -{ - unsigned long old, t; - unsigned long mask = 1UL << (nr & 0x3f); - unsigned long *p = ((unsigned long *)addr) + (nr >> 6); - - __asm__ __volatile__( - EIEIO_ON_SMP -"1: ldarx %0,0,%3 # test_and_clear_bit\n\ - andc %1,%0,%2\n\ - stdcx. %1,0,%3\n\ - bne- 1b" - ISYNC_ON_SMP - : "=&r" (old), "=&r" (t) - : "r" (mask), "r" (p) - : "cc", "memory"); - - return (old & mask) != 0; -} - -static __inline__ int test_and_change_bit(unsigned long nr, volatile unsigned long *addr) -{ - unsigned long old, t; - unsigned long mask = 1UL << (nr & 0x3f); - unsigned long *p = ((unsigned long *)addr) + (nr >> 6); - - __asm__ __volatile__( - EIEIO_ON_SMP -"1: ldarx %0,0,%3 # test_and_change_bit\n\ - xor %1,%0,%2\n\ - stdcx. %1,0,%3\n\ - bne- 1b" - ISYNC_ON_SMP - : "=&r" (old), "=&r" (t) - : "r" (mask), "r" (p) - : "cc", "memory"); - - return (old & mask) != 0; -} - -static __inline__ void set_bits(unsigned long mask, unsigned long *addr) -{ - unsigned long old; - - __asm__ __volatile__( -"1: ldarx %0,0,%3 # set_bit\n\ - or %0,%0,%2\n\ - stdcx. %0,0,%3\n\ - bne- 1b" - : "=&r" (old), "=m" (*addr) - : "r" (mask), "r" (addr), "m" (*addr) - : "cc"); -} - -/* - * non-atomic versions - */ -static __inline__ void __set_bit(unsigned long nr, volatile unsigned long *addr) -{ - unsigned long mask = 1UL << (nr & 0x3f); - unsigned long *p = ((unsigned long *)addr) + (nr >> 6); - - *p |= mask; -} - -static __inline__ void __clear_bit(unsigned long nr, volatile unsigned long *addr) -{ - unsigned long mask = 1UL << (nr & 0x3f); - unsigned long *p = ((unsigned long *)addr) + (nr >> 6); - - *p &= ~mask; -} - -static __inline__ void __change_bit(unsigned long nr, volatile unsigned long *addr) -{ - unsigned long mask = 1UL << (nr & 0x3f); - unsigned long *p = ((unsigned long *)addr) + (nr >> 6); - - *p ^= mask; -} - -static __inline__ int __test_and_set_bit(unsigned long nr, volatile unsigned long *addr) -{ - unsigned long mask = 1UL << (nr & 0x3f); - unsigned long *p = ((unsigned long *)addr) + (nr >> 6); - unsigned long old = *p; - - *p = old | mask; - return (old & mask) != 0; -} - -static __inline__ int __test_and_clear_bit(unsigned long nr, volatile unsigned long *addr) -{ - unsigned long mask = 1UL << (nr & 0x3f); - unsigned long *p = ((unsigned long *)addr) + (nr >> 6); - unsigned long old = *p; - - *p = old & ~mask; - return (old & mask) != 0; -} - -static __inline__ int __test_and_change_bit(unsigned long nr, volatile unsigned long *addr) -{ - unsigned long mask = 1UL << (nr & 0x3f); - unsigned long *p = ((unsigned long *)addr) + (nr >> 6); - unsigned long old = *p; - - *p = old ^ mask; - return (old & mask) != 0; -} - -/* - * Return the zero-based bit position (from RIGHT TO LEFT, 63 -> 0) of the - * most significant (left-most) 1-bit in a double word. - */ -static __inline__ int __ilog2(unsigned long x) -{ - int lz; - - asm ("cntlzd %0,%1" : "=r" (lz) : "r" (x)); - return 63 - lz; -} - -/* - * Determines the bit position of the least significant (rightmost) 0 bit - * in the specified double word. The returned bit position will be zero-based, - * starting from the right side (63 - 0). - */ -static __inline__ unsigned long ffz(unsigned long x) -{ - /* no zero exists anywhere in the 8 byte area. */ - if ((x = ~x) == 0) - return 64; - - /* - * Calculate the bit position of the least signficant '1' bit in x - * (since x has been changed this will actually be the least signficant - * '0' bit in * the original x). Note: (x & -x) gives us a mask that - * is the least significant * (RIGHT-most) 1-bit of the value in x. - */ - return __ilog2(x & -x); -} - -static __inline__ int __ffs(unsigned long x) -{ - return __ilog2(x & -x); -} - -/* - * ffs: find first bit set. This is defined the same way as - * the libc and compiler builtin ffs routines, therefore - * differs in spirit from the above ffz (man ffs). - */ -static __inline__ int ffs(int x) -{ - unsigned long i = (unsigned long)x; - return __ilog2(i & -i) + 1; -} - -/* - * fls: find last (most-significant) bit set. - * Note fls(0) = 0, fls(1) = 1, fls(0x80000000) = 32. - */ -#define fls(x) generic_fls(x) - -/* - * hweightN: returns the hamming weight (i.e. the number - * of bits set) of a N-bit word - */ -#define hweight64(x) generic_hweight64(x) -#define hweight32(x) generic_hweight32(x) -#define hweight16(x) generic_hweight16(x) -#define hweight8(x) generic_hweight8(x) - -extern unsigned long find_next_zero_bit(const unsigned long *addr, unsigned long size, unsigned long offset); -#define find_first_zero_bit(addr, size) \ - find_next_zero_bit((addr), (size), 0) - -extern unsigned long find_next_bit(const unsigned long *addr, unsigned long size, unsigned long offset); -#define find_first_bit(addr, size) \ - find_next_bit((addr), (size), 0) - -extern unsigned long find_next_zero_le_bit(const unsigned long *addr, unsigned long size, unsigned long offset); -#define find_first_zero_le_bit(addr, size) \ - find_next_zero_le_bit((addr), (size), 0) - -static __inline__ int test_le_bit(unsigned long nr, __const__ unsigned long * addr) -{ - __const__ unsigned char *ADDR = (__const__ unsigned char *) addr; - return (ADDR[nr >> 3] >> (nr & 7)) & 1; -} - -#define test_and_clear_le_bit(nr, addr) \ - test_and_clear_bit((nr) ^ 0x38, (addr)) -#define test_and_set_le_bit(nr, addr) \ - test_and_set_bit((nr) ^ 0x38, (addr)) - -/* - * non-atomic versions - */ - -#define __set_le_bit(nr, addr) \ - __set_bit((nr) ^ 0x38, (addr)) -#define __clear_le_bit(nr, addr) \ - __clear_bit((nr) ^ 0x38, (addr)) -#define __test_and_clear_le_bit(nr, addr) \ - __test_and_clear_bit((nr) ^ 0x38, (addr)) -#define __test_and_set_le_bit(nr, addr) \ - __test_and_set_bit((nr) ^ 0x38, (addr)) - -#define ext2_set_bit(nr,addr) \ - __test_and_set_le_bit((nr), (unsigned long*)addr) -#define ext2_clear_bit(nr, addr) \ - __test_and_clear_le_bit((nr), (unsigned long*)addr) - -#define ext2_set_bit_atomic(lock, nr, addr) \ - test_and_set_le_bit((nr), (unsigned long*)addr) -#define ext2_clear_bit_atomic(lock, nr, addr) \ - test_and_clear_le_bit((nr), (unsigned long*)addr) - - -#define ext2_test_bit(nr, addr) test_le_bit((nr),(unsigned long*)addr) -#define ext2_find_first_zero_bit(addr, size) \ - find_first_zero_le_bit((unsigned long*)addr, size) -#define ext2_find_next_zero_bit(addr, size, off) \ - find_next_zero_le_bit((unsigned long*)addr, size, off) - -#define minix_test_and_set_bit(nr,addr) test_and_set_bit(nr,addr) -#define minix_set_bit(nr,addr) set_bit(nr,addr) -#define minix_test_and_clear_bit(nr,addr) test_and_clear_bit(nr,addr) -#define minix_test_bit(nr,addr) test_bit(nr,addr) -#define minix_find_first_zero_bit(addr,size) find_first_zero_bit(addr,size) - -#endif /* __KERNEL__ */ -#endif /* _PPC64_BITOPS_H */ diff --git a/include/asm-ppc64/mmu_context.h b/include/asm-ppc64/mmu_context.h index 77a743402db4..820dd729b895 100644 --- a/include/asm-ppc64/mmu_context.h +++ b/include/asm-ppc64/mmu_context.h @@ -16,21 +16,6 @@ * 2 of the License, or (at your option) any later version. */ -/* - * Every architecture must define this function. It's the fastest - * way of searching a 140-bit bitmap where the first 100 bits are - * unlikely to be set. It's guaranteed that at least one of the 140 - * bits is cleared. - */ -static inline int sched_find_first_bit(unsigned long *b) -{ - if (unlikely(b[0])) - return __ffs(b[0]); - if (unlikely(b[1])) - return __ffs(b[1]) + 64; - return __ffs(b[2]) + 128; -} - static inline void enter_lazy_tlb(struct mm_struct *mm, struct task_struct *tsk) { } -- cgit v1.2.3 From 15b17189489f6d759fa2d61e7b6c87c55eeffb2c Mon Sep 17 00:00:00 2001 From: Kelly Daly Date: Wed, 2 Nov 2005 11:55:28 +1100 Subject: merge filename and modify reference to iseries/hv_lp_config.h Signed-off-by: Kelly Daly --- arch/powerpc/platforms/iseries/hvlpconfig.c | 2 +- arch/powerpc/platforms/iseries/mf.c | 2 +- arch/powerpc/platforms/iseries/setup.c | 2 +- arch/powerpc/platforms/iseries/vio.c | 2 +- arch/powerpc/platforms/iseries/viopath.c | 2 +- arch/ppc64/kernel/lparcfg.c | 2 +- drivers/block/viodasd.c | 2 +- drivers/char/viocons.c | 2 +- drivers/char/viotape.c | 2 +- drivers/net/iseries_veth.c | 2 +- drivers/scsi/ibmvscsi/iseries_vscsi.c | 2 +- include/asm-powerpc/iseries/hv_lp_config.h | 138 ++++++++++++++++++++++++++++ include/asm-ppc64/iSeries/HvLpConfig.h | 138 ---------------------------- 13 files changed, 149 insertions(+), 149 deletions(-) create mode 100644 include/asm-powerpc/iseries/hv_lp_config.h delete mode 100644 include/asm-ppc64/iSeries/HvLpConfig.h (limited to 'arch/ppc64') diff --git a/arch/powerpc/platforms/iseries/hvlpconfig.c b/arch/powerpc/platforms/iseries/hvlpconfig.c index dc28621aea0d..663a1affb4bb 100644 --- a/arch/powerpc/platforms/iseries/hvlpconfig.c +++ b/arch/powerpc/platforms/iseries/hvlpconfig.c @@ -17,7 +17,7 @@ */ #include -#include +#include HvLpIndex HvLpConfig_getLpIndex_outline(void) { diff --git a/arch/powerpc/platforms/iseries/mf.c b/arch/powerpc/platforms/iseries/mf.c index e5de31aa0015..027df204446d 100644 --- a/arch/powerpc/platforms/iseries/mf.c +++ b/arch/powerpc/platforms/iseries/mf.c @@ -40,7 +40,7 @@ #include #include #include -#include +#include #include #include "setup.h" diff --git a/arch/powerpc/platforms/iseries/setup.c b/arch/powerpc/platforms/iseries/setup.c index 669e63174ecc..638747c772f8 100644 --- a/arch/powerpc/platforms/iseries/setup.c +++ b/arch/powerpc/platforms/iseries/setup.c @@ -45,7 +45,7 @@ #include #include #include -#include +#include #include #include #include diff --git a/arch/powerpc/platforms/iseries/vio.c b/arch/powerpc/platforms/iseries/vio.c index c64741e50cb2..4e2dafc83bb2 100644 --- a/arch/powerpc/platforms/iseries/vio.c +++ b/arch/powerpc/platforms/iseries/vio.c @@ -19,7 +19,7 @@ #include #include #include -#include +#include #include struct device *iSeries_vio_dev = &vio_bus_device.dev; diff --git a/arch/powerpc/platforms/iseries/viopath.c b/arch/powerpc/platforms/iseries/viopath.c index c0c767bd37f1..8e33a90bedb9 100644 --- a/arch/powerpc/platforms/iseries/viopath.c +++ b/arch/powerpc/platforms/iseries/viopath.c @@ -44,7 +44,7 @@ #include #include #include -#include +#include #include #include diff --git a/arch/ppc64/kernel/lparcfg.c b/arch/ppc64/kernel/lparcfg.c index cae19bbd5acd..e4236f2707e1 100644 --- a/arch/ppc64/kernel/lparcfg.c +++ b/arch/ppc64/kernel/lparcfg.c @@ -26,7 +26,7 @@ #include #include #include -#include +#include #include #include #include diff --git a/drivers/block/viodasd.c b/drivers/block/viodasd.c index 709f809f79f1..5c907b747648 100644 --- a/drivers/block/viodasd.c +++ b/drivers/block/viodasd.c @@ -47,7 +47,7 @@ #include #include #include -#include +#include #include MODULE_DESCRIPTION("iSeries Virtual DASD"); diff --git a/drivers/char/viocons.c b/drivers/char/viocons.c index ecf5c0953ae4..33d7820eaab9 100644 --- a/drivers/char/viocons.c +++ b/drivers/char/viocons.c @@ -48,7 +48,7 @@ #include #include -#include +#include #include #ifdef CONFIG_VT diff --git a/drivers/char/viotape.c b/drivers/char/viotape.c index 232da8c5e66a..296797e5a2b5 100644 --- a/drivers/char/viotape.c +++ b/drivers/char/viotape.c @@ -57,7 +57,7 @@ #include #include #include -#include +#include #define VIOTAPE_VERSION "1.2" #define VIOTAPE_MAXREQ 1 diff --git a/drivers/net/iseries_veth.c b/drivers/net/iseries_veth.c index f5ea39ff1017..aaabaf30d4b8 100644 --- a/drivers/net/iseries_veth.c +++ b/drivers/net/iseries_veth.c @@ -75,7 +75,7 @@ #include #include -#include +#include #include #include #include diff --git a/drivers/scsi/ibmvscsi/iseries_vscsi.c b/drivers/scsi/ibmvscsi/iseries_vscsi.c index e9202f2a8276..c652779c1a32 100644 --- a/drivers/scsi/ibmvscsi/iseries_vscsi.c +++ b/drivers/scsi/ibmvscsi/iseries_vscsi.c @@ -31,7 +31,7 @@ #include #include #include -#include +#include #include #include #include "ibmvscsi.h" diff --git a/include/asm-powerpc/iseries/hv_lp_config.h b/include/asm-powerpc/iseries/hv_lp_config.h new file mode 100644 index 000000000000..9077fc8e3f6c --- /dev/null +++ b/include/asm-powerpc/iseries/hv_lp_config.h @@ -0,0 +1,138 @@ +/* + * HvLpConfig.h + * Copyright (C) 2001 Mike Corrigan IBM Corporation + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program; if not, write to the Free Software + * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA + */ +#ifndef _HVLPCONFIG_H +#define _HVLPCONFIG_H + +/* + * This file contains the interface to the LPAR configuration data + * to determine which resources should be allocated to each partition. + */ + +#include +#include +#include + +enum { + HvCallCfg_Cur = 0, + HvCallCfg_Init = 1, + HvCallCfg_Max = 2, + HvCallCfg_Min = 3 +}; + +#define HvCallCfgGetSystemPhysicalProcessors HvCallCfg + 6 +#define HvCallCfgGetPhysicalProcessors HvCallCfg + 7 +#define HvCallCfgGetMsChunks HvCallCfg + 9 +#define HvCallCfgGetSharedPoolIndex HvCallCfg + 20 +#define HvCallCfgGetSharedProcUnits HvCallCfg + 21 +#define HvCallCfgGetNumProcsInSharedPool HvCallCfg + 22 +#define HvCallCfgGetVirtualLanIndexMap HvCallCfg + 30 +#define HvCallCfgGetHostingLpIndex HvCallCfg + 32 + +extern HvLpIndex HvLpConfig_getLpIndex_outline(void); + +static inline HvLpIndex HvLpConfig_getLpIndex(void) +{ + return itLpNaca.xLpIndex; +} + +static inline HvLpIndex HvLpConfig_getPrimaryLpIndex(void) +{ + return itLpNaca.xPrimaryLpIndex; +} + +static inline u64 HvLpConfig_getMsChunks(void) +{ + return HvCall2(HvCallCfgGetMsChunks, HvLpConfig_getLpIndex(), + HvCallCfg_Cur); +} + +static inline u64 HvLpConfig_getSystemPhysicalProcessors(void) +{ + return HvCall0(HvCallCfgGetSystemPhysicalProcessors); +} + +static inline u64 HvLpConfig_getNumProcsInSharedPool(HvLpSharedPoolIndex sPI) +{ + return (u16)HvCall1(HvCallCfgGetNumProcsInSharedPool, sPI); +} + +static inline u64 HvLpConfig_getPhysicalProcessors(void) +{ + return HvCall2(HvCallCfgGetPhysicalProcessors, HvLpConfig_getLpIndex(), + HvCallCfg_Cur); +} + +static inline HvLpSharedPoolIndex HvLpConfig_getSharedPoolIndex(void) +{ + return HvCall1(HvCallCfgGetSharedPoolIndex, HvLpConfig_getLpIndex()); +} + +static inline u64 HvLpConfig_getSharedProcUnits(void) +{ + return HvCall2(HvCallCfgGetSharedProcUnits, HvLpConfig_getLpIndex(), + HvCallCfg_Cur); +} + +static inline u64 HvLpConfig_getMaxSharedProcUnits(void) +{ + return HvCall2(HvCallCfgGetSharedProcUnits, HvLpConfig_getLpIndex(), + HvCallCfg_Max); +} + +static inline u64 HvLpConfig_getMaxPhysicalProcessors(void) +{ + return HvCall2(HvCallCfgGetPhysicalProcessors, HvLpConfig_getLpIndex(), + HvCallCfg_Max); +} + +static inline HvLpVirtualLanIndexMap HvLpConfig_getVirtualLanIndexMapForLp( + HvLpIndex lp) +{ + /* + * This is a new function in V5R1 so calls to this on older + * hypervisors will return -1 + */ + u64 retVal = HvCall1(HvCallCfgGetVirtualLanIndexMap, lp); + if (retVal == -1) + retVal = 0; + return retVal; +} + +static inline HvLpVirtualLanIndexMap HvLpConfig_getVirtualLanIndexMap(void) +{ + return HvLpConfig_getVirtualLanIndexMapForLp( + HvLpConfig_getLpIndex_outline()); +} + +static inline int HvLpConfig_doLpsCommunicateOnVirtualLan(HvLpIndex lp1, + HvLpIndex lp2) +{ + HvLpVirtualLanIndexMap virtualLanIndexMap1 = + HvLpConfig_getVirtualLanIndexMapForLp(lp1); + HvLpVirtualLanIndexMap virtualLanIndexMap2 = + HvLpConfig_getVirtualLanIndexMapForLp(lp2); + return ((virtualLanIndexMap1 & virtualLanIndexMap2) != 0); +} + +static inline HvLpIndex HvLpConfig_getHostingLpIndex(HvLpIndex lp) +{ + return HvCall1(HvCallCfgGetHostingLpIndex, lp); +} + +#endif /* _HVLPCONFIG_H */ diff --git a/include/asm-ppc64/iSeries/HvLpConfig.h b/include/asm-ppc64/iSeries/HvLpConfig.h deleted file mode 100644 index 9077fc8e3f6c..000000000000 --- a/include/asm-ppc64/iSeries/HvLpConfig.h +++ /dev/null @@ -1,138 +0,0 @@ -/* - * HvLpConfig.h - * Copyright (C) 2001 Mike Corrigan IBM Corporation - * - * This program is free software; you can redistribute it and/or modify - * it under the terms of the GNU General Public License as published by - * the Free Software Foundation; either version 2 of the License, or - * (at your option) any later version. - * - * This program is distributed in the hope that it will be useful, - * but WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the - * GNU General Public License for more details. - * - * You should have received a copy of the GNU General Public License - * along with this program; if not, write to the Free Software - * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA - */ -#ifndef _HVLPCONFIG_H -#define _HVLPCONFIG_H - -/* - * This file contains the interface to the LPAR configuration data - * to determine which resources should be allocated to each partition. - */ - -#include -#include -#include - -enum { - HvCallCfg_Cur = 0, - HvCallCfg_Init = 1, - HvCallCfg_Max = 2, - HvCallCfg_Min = 3 -}; - -#define HvCallCfgGetSystemPhysicalProcessors HvCallCfg + 6 -#define HvCallCfgGetPhysicalProcessors HvCallCfg + 7 -#define HvCallCfgGetMsChunks HvCallCfg + 9 -#define HvCallCfgGetSharedPoolIndex HvCallCfg + 20 -#define HvCallCfgGetSharedProcUnits HvCallCfg + 21 -#define HvCallCfgGetNumProcsInSharedPool HvCallCfg + 22 -#define HvCallCfgGetVirtualLanIndexMap HvCallCfg + 30 -#define HvCallCfgGetHostingLpIndex HvCallCfg + 32 - -extern HvLpIndex HvLpConfig_getLpIndex_outline(void); - -static inline HvLpIndex HvLpConfig_getLpIndex(void) -{ - return itLpNaca.xLpIndex; -} - -static inline HvLpIndex HvLpConfig_getPrimaryLpIndex(void) -{ - return itLpNaca.xPrimaryLpIndex; -} - -static inline u64 HvLpConfig_getMsChunks(void) -{ - return HvCall2(HvCallCfgGetMsChunks, HvLpConfig_getLpIndex(), - HvCallCfg_Cur); -} - -static inline u64 HvLpConfig_getSystemPhysicalProcessors(void) -{ - return HvCall0(HvCallCfgGetSystemPhysicalProcessors); -} - -static inline u64 HvLpConfig_getNumProcsInSharedPool(HvLpSharedPoolIndex sPI) -{ - return (u16)HvCall1(HvCallCfgGetNumProcsInSharedPool, sPI); -} - -static inline u64 HvLpConfig_getPhysicalProcessors(void) -{ - return HvCall2(HvCallCfgGetPhysicalProcessors, HvLpConfig_getLpIndex(), - HvCallCfg_Cur); -} - -static inline HvLpSharedPoolIndex HvLpConfig_getSharedPoolIndex(void) -{ - return HvCall1(HvCallCfgGetSharedPoolIndex, HvLpConfig_getLpIndex()); -} - -static inline u64 HvLpConfig_getSharedProcUnits(void) -{ - return HvCall2(HvCallCfgGetSharedProcUnits, HvLpConfig_getLpIndex(), - HvCallCfg_Cur); -} - -static inline u64 HvLpConfig_getMaxSharedProcUnits(void) -{ - return HvCall2(HvCallCfgGetSharedProcUnits, HvLpConfig_getLpIndex(), - HvCallCfg_Max); -} - -static inline u64 HvLpConfig_getMaxPhysicalProcessors(void) -{ - return HvCall2(HvCallCfgGetPhysicalProcessors, HvLpConfig_getLpIndex(), - HvCallCfg_Max); -} - -static inline HvLpVirtualLanIndexMap HvLpConfig_getVirtualLanIndexMapForLp( - HvLpIndex lp) -{ - /* - * This is a new function in V5R1 so calls to this on older - * hypervisors will return -1 - */ - u64 retVal = HvCall1(HvCallCfgGetVirtualLanIndexMap, lp); - if (retVal == -1) - retVal = 0; - return retVal; -} - -static inline HvLpVirtualLanIndexMap HvLpConfig_getVirtualLanIndexMap(void) -{ - return HvLpConfig_getVirtualLanIndexMapForLp( - HvLpConfig_getLpIndex_outline()); -} - -static inline int HvLpConfig_doLpsCommunicateOnVirtualLan(HvLpIndex lp1, - HvLpIndex lp2) -{ - HvLpVirtualLanIndexMap virtualLanIndexMap1 = - HvLpConfig_getVirtualLanIndexMapForLp(lp1); - HvLpVirtualLanIndexMap virtualLanIndexMap2 = - HvLpConfig_getVirtualLanIndexMapForLp(lp2); - return ((virtualLanIndexMap1 & virtualLanIndexMap2) != 0); -} - -static inline HvLpIndex HvLpConfig_getHostingLpIndex(HvLpIndex lp) -{ - return HvCall1(HvCallCfgGetHostingLpIndex, lp); -} - -#endif /* _HVLPCONFIG_H */ -- cgit v1.2.3 From e45423eac2e191a6cfdacdf61cb931976d73cc0b Mon Sep 17 00:00:00 2001 From: Kelly Daly Date: Wed, 2 Nov 2005 12:08:31 +1100 Subject: merge filename and modify references to iseries/hv_lp_event.h Signed-off-by: Kelly Daly --- arch/powerpc/kernel/asm-offsets.c | 2 +- arch/powerpc/platforms/iseries/irq.c | 2 +- arch/powerpc/platforms/iseries/lpevents.c | 2 +- arch/powerpc/platforms/iseries/setup.c | 2 +- arch/powerpc/platforms/iseries/viopath.c | 2 +- arch/ppc64/kernel/asm-offsets.c | 2 +- drivers/block/viodasd.c | 2 +- drivers/cdrom/viocd.c | 2 +- drivers/char/viocons.c | 2 +- drivers/char/viotape.c | 2 +- drivers/net/iseries_veth.c | 2 +- drivers/scsi/ibmvscsi/iseries_vscsi.c | 2 +- include/asm-powerpc/iseries/hv_lp_event.h | 142 ++++++++++++++++++++++++++++++ include/asm-ppc64/iSeries/HvLpEvent.h | 142 ------------------------------ include/asm-ppc64/iSeries/vio.h | 2 +- 15 files changed, 155 insertions(+), 155 deletions(-) create mode 100644 include/asm-powerpc/iseries/hv_lp_event.h delete mode 100644 include/asm-ppc64/iSeries/HvLpEvent.h (limited to 'arch/ppc64') diff --git a/arch/powerpc/kernel/asm-offsets.c b/arch/powerpc/kernel/asm-offsets.c index 330cd783206f..bc5a3689cc05 100644 --- a/arch/powerpc/kernel/asm-offsets.c +++ b/arch/powerpc/kernel/asm-offsets.c @@ -40,7 +40,7 @@ #ifdef CONFIG_PPC64 #include #include -#include +#include #include #include #include diff --git a/arch/powerpc/platforms/iseries/irq.c b/arch/powerpc/platforms/iseries/irq.c index b1aacf64fb8e..55ecc8d3e73f 100644 --- a/arch/powerpc/platforms/iseries/irq.c +++ b/arch/powerpc/platforms/iseries/irq.c @@ -37,7 +37,7 @@ #include #include -#include +#include #include #include "irq.h" diff --git a/arch/powerpc/platforms/iseries/lpevents.c b/arch/powerpc/platforms/iseries/lpevents.c index 96aa0dd72fb0..39ee881c36c3 100644 --- a/arch/powerpc/platforms/iseries/lpevents.c +++ b/arch/powerpc/platforms/iseries/lpevents.c @@ -18,7 +18,7 @@ #include #include #include -#include +#include #include #include diff --git a/arch/powerpc/platforms/iseries/setup.c b/arch/powerpc/platforms/iseries/setup.c index 638747c772f8..881d81fcf3b8 100644 --- a/arch/powerpc/platforms/iseries/setup.c +++ b/arch/powerpc/platforms/iseries/setup.c @@ -50,7 +50,7 @@ #include #include #include -#include +#include #include #include "setup.h" diff --git a/arch/powerpc/platforms/iseries/viopath.c b/arch/powerpc/platforms/iseries/viopath.c index 8e33a90bedb9..7b1a9daae206 100644 --- a/arch/powerpc/platforms/iseries/viopath.c +++ b/arch/powerpc/platforms/iseries/viopath.c @@ -43,7 +43,7 @@ #include #include #include -#include +#include #include #include #include diff --git a/arch/ppc64/kernel/asm-offsets.c b/arch/ppc64/kernel/asm-offsets.c index 5e6046cb414e..504dee836d29 100644 --- a/arch/ppc64/kernel/asm-offsets.c +++ b/arch/ppc64/kernel/asm-offsets.c @@ -31,7 +31,7 @@ #include #include -#include +#include #include #include #include diff --git a/drivers/block/viodasd.c b/drivers/block/viodasd.c index 5c907b747648..3af1885a8200 100644 --- a/drivers/block/viodasd.c +++ b/drivers/block/viodasd.c @@ -46,7 +46,7 @@ #include #include #include -#include +#include #include #include diff --git a/drivers/cdrom/viocd.c b/drivers/cdrom/viocd.c index 36f31d202223..da78729413f4 100644 --- a/drivers/cdrom/viocd.c +++ b/drivers/cdrom/viocd.c @@ -47,7 +47,7 @@ #include #include #include -#include +#include #include #define VIOCD_DEVICE "iseries/vcd" diff --git a/drivers/char/viocons.c b/drivers/char/viocons.c index 33d7820eaab9..694bab3b7bf7 100644 --- a/drivers/char/viocons.c +++ b/drivers/char/viocons.c @@ -46,7 +46,7 @@ #include -#include +#include #include #include #include diff --git a/drivers/char/viotape.c b/drivers/char/viotape.c index 296797e5a2b5..3242b241d789 100644 --- a/drivers/char/viotape.c +++ b/drivers/char/viotape.c @@ -55,7 +55,7 @@ #include #include -#include +#include #include #include diff --git a/drivers/net/iseries_veth.c b/drivers/net/iseries_veth.c index aaabaf30d4b8..57ba1a11e405 100644 --- a/drivers/net/iseries_veth.c +++ b/drivers/net/iseries_veth.c @@ -77,7 +77,7 @@ #include #include -#include +#include #include #include diff --git a/drivers/scsi/ibmvscsi/iseries_vscsi.c b/drivers/scsi/ibmvscsi/iseries_vscsi.c index c652779c1a32..78397d855d34 100644 --- a/drivers/scsi/ibmvscsi/iseries_vscsi.c +++ b/drivers/scsi/ibmvscsi/iseries_vscsi.c @@ -29,7 +29,7 @@ */ #include -#include +#include #include #include #include diff --git a/include/asm-powerpc/iseries/hv_lp_event.h b/include/asm-powerpc/iseries/hv_lp_event.h new file mode 100644 index 000000000000..75ff1b3ed2e5 --- /dev/null +++ b/include/asm-powerpc/iseries/hv_lp_event.h @@ -0,0 +1,142 @@ +/* + * HvLpEvent.h + * Copyright (C) 2001 Mike Corrigan IBM Corporation + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program; if not, write to the Free Software + * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA + */ + +/* This file contains the class for HV events in the system. */ + +#ifndef _HVLPEVENT_H +#define _HVLPEVENT_H + +#include +#include +#include +#include + +/* + * HvLpEvent is the structure for Lp Event messages passed between + * partitions through PLIC. + */ + +struct HvEventFlags { + u8 xValid:1; /* Indicates a valid request x00-x00 */ + u8 xRsvd1:4; /* Reserved ... */ + u8 xAckType:1; /* Immediate or deferred ... */ + u8 xAckInd:1; /* Indicates if ACK required ... */ + u8 xFunction:1; /* Interrupt or Acknowledge ... */ +}; + + +struct HvLpEvent { + struct HvEventFlags xFlags; /* Event flags x00-x00 */ + u8 xType; /* Type of message x01-x01 */ + u16 xSubtype; /* Subtype for event x02-x03 */ + u8 xSourceLp; /* Source LP x04-x04 */ + u8 xTargetLp; /* Target LP x05-x05 */ + u8 xSizeMinus1; /* Size of Derived class - 1 x06-x06 */ + u8 xRc; /* RC for Ack flows x07-x07 */ + u16 xSourceInstanceId; /* Source sides instance id x08-x09 */ + u16 xTargetInstanceId; /* Target sides instance id x0A-x0B */ + union { + u32 xSubtypeData; /* Data usable by the subtype x0C-x0F */ + u16 xSubtypeDataShort[2]; /* Data as 2 shorts */ + u8 xSubtypeDataChar[4]; /* Data as 4 chars */ + } x; + + u64 xCorrelationToken; /* Unique value for source/type x10-x17 */ +}; + +typedef void (*LpEventHandler)(struct HvLpEvent *, struct pt_regs *); + +/* Register a handler for an event type - returns 0 on success */ +extern int HvLpEvent_registerHandler(HvLpEvent_Type eventType, + LpEventHandler hdlr); + +/* + * Unregister a handler for an event type + * + * This call will sleep until the handler being removed is guaranteed to + * be no longer executing on any CPU. Do not call with locks held. + * + * returns 0 on success + * Unregister will fail if there are any paths open for the type + */ +extern int HvLpEvent_unregisterHandler(HvLpEvent_Type eventType); + +/* + * Open an Lp Event Path for an event type + * returns 0 on success + * openPath will fail if there is no handler registered for the event type. + * The lpIndex specified is the partition index for the target partition + * (for VirtualIo, VirtualLan and SessionMgr) other types specify zero) + */ +extern int HvLpEvent_openPath(HvLpEvent_Type eventType, HvLpIndex lpIndex); + +/* + * Close an Lp Event Path for a type and partition + * returns 0 on sucess + */ +extern int HvLpEvent_closePath(HvLpEvent_Type eventType, HvLpIndex lpIndex); + +#define HvLpEvent_Type_Hypervisor 0 +#define HvLpEvent_Type_MachineFac 1 +#define HvLpEvent_Type_SessionMgr 2 +#define HvLpEvent_Type_SpdIo 3 +#define HvLpEvent_Type_VirtualBus 4 +#define HvLpEvent_Type_PciIo 5 +#define HvLpEvent_Type_RioIo 6 +#define HvLpEvent_Type_VirtualLan 7 +#define HvLpEvent_Type_VirtualIo 8 +#define HvLpEvent_Type_NumTypes 9 + +#define HvLpEvent_Rc_Good 0 +#define HvLpEvent_Rc_BufferNotAvailable 1 +#define HvLpEvent_Rc_Cancelled 2 +#define HvLpEvent_Rc_GenericError 3 +#define HvLpEvent_Rc_InvalidAddress 4 +#define HvLpEvent_Rc_InvalidPartition 5 +#define HvLpEvent_Rc_InvalidSize 6 +#define HvLpEvent_Rc_InvalidSubtype 7 +#define HvLpEvent_Rc_InvalidSubtypeData 8 +#define HvLpEvent_Rc_InvalidType 9 +#define HvLpEvent_Rc_PartitionDead 10 +#define HvLpEvent_Rc_PathClosed 11 +#define HvLpEvent_Rc_SubtypeError 12 + +#define HvLpEvent_Function_Ack 0 +#define HvLpEvent_Function_Int 1 + +#define HvLpEvent_AckInd_NoAck 0 +#define HvLpEvent_AckInd_DoAck 1 + +#define HvLpEvent_AckType_ImmediateAck 0 +#define HvLpEvent_AckType_DeferredAck 1 + +#define HvLpDma_Direction_LocalToRemote 0 +#define HvLpDma_Direction_RemoteToLocal 1 + +#define HvLpDma_AddressType_TceIndex 0 +#define HvLpDma_AddressType_RealAddress 1 + +#define HvLpDma_Rc_Good 0 +#define HvLpDma_Rc_Error 1 +#define HvLpDma_Rc_PartitionDead 2 +#define HvLpDma_Rc_PathClosed 3 +#define HvLpDma_Rc_InvalidAddress 4 +#define HvLpDma_Rc_InvalidLength 5 + +#endif /* _HVLPEVENT_H */ diff --git a/include/asm-ppc64/iSeries/HvLpEvent.h b/include/asm-ppc64/iSeries/HvLpEvent.h deleted file mode 100644 index 75ff1b3ed2e5..000000000000 --- a/include/asm-ppc64/iSeries/HvLpEvent.h +++ /dev/null @@ -1,142 +0,0 @@ -/* - * HvLpEvent.h - * Copyright (C) 2001 Mike Corrigan IBM Corporation - * - * This program is free software; you can redistribute it and/or modify - * it under the terms of the GNU General Public License as published by - * the Free Software Foundation; either version 2 of the License, or - * (at your option) any later version. - * - * This program is distributed in the hope that it will be useful, - * but WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the - * GNU General Public License for more details. - * - * You should have received a copy of the GNU General Public License - * along with this program; if not, write to the Free Software - * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA - */ - -/* This file contains the class for HV events in the system. */ - -#ifndef _HVLPEVENT_H -#define _HVLPEVENT_H - -#include -#include -#include -#include - -/* - * HvLpEvent is the structure for Lp Event messages passed between - * partitions through PLIC. - */ - -struct HvEventFlags { - u8 xValid:1; /* Indicates a valid request x00-x00 */ - u8 xRsvd1:4; /* Reserved ... */ - u8 xAckType:1; /* Immediate or deferred ... */ - u8 xAckInd:1; /* Indicates if ACK required ... */ - u8 xFunction:1; /* Interrupt or Acknowledge ... */ -}; - - -struct HvLpEvent { - struct HvEventFlags xFlags; /* Event flags x00-x00 */ - u8 xType; /* Type of message x01-x01 */ - u16 xSubtype; /* Subtype for event x02-x03 */ - u8 xSourceLp; /* Source LP x04-x04 */ - u8 xTargetLp; /* Target LP x05-x05 */ - u8 xSizeMinus1; /* Size of Derived class - 1 x06-x06 */ - u8 xRc; /* RC for Ack flows x07-x07 */ - u16 xSourceInstanceId; /* Source sides instance id x08-x09 */ - u16 xTargetInstanceId; /* Target sides instance id x0A-x0B */ - union { - u32 xSubtypeData; /* Data usable by the subtype x0C-x0F */ - u16 xSubtypeDataShort[2]; /* Data as 2 shorts */ - u8 xSubtypeDataChar[4]; /* Data as 4 chars */ - } x; - - u64 xCorrelationToken; /* Unique value for source/type x10-x17 */ -}; - -typedef void (*LpEventHandler)(struct HvLpEvent *, struct pt_regs *); - -/* Register a handler for an event type - returns 0 on success */ -extern int HvLpEvent_registerHandler(HvLpEvent_Type eventType, - LpEventHandler hdlr); - -/* - * Unregister a handler for an event type - * - * This call will sleep until the handler being removed is guaranteed to - * be no longer executing on any CPU. Do not call with locks held. - * - * returns 0 on success - * Unregister will fail if there are any paths open for the type - */ -extern int HvLpEvent_unregisterHandler(HvLpEvent_Type eventType); - -/* - * Open an Lp Event Path for an event type - * returns 0 on success - * openPath will fail if there is no handler registered for the event type. - * The lpIndex specified is the partition index for the target partition - * (for VirtualIo, VirtualLan and SessionMgr) other types specify zero) - */ -extern int HvLpEvent_openPath(HvLpEvent_Type eventType, HvLpIndex lpIndex); - -/* - * Close an Lp Event Path for a type and partition - * returns 0 on sucess - */ -extern int HvLpEvent_closePath(HvLpEvent_Type eventType, HvLpIndex lpIndex); - -#define HvLpEvent_Type_Hypervisor 0 -#define HvLpEvent_Type_MachineFac 1 -#define HvLpEvent_Type_SessionMgr 2 -#define HvLpEvent_Type_SpdIo 3 -#define HvLpEvent_Type_VirtualBus 4 -#define HvLpEvent_Type_PciIo 5 -#define HvLpEvent_Type_RioIo 6 -#define HvLpEvent_Type_VirtualLan 7 -#define HvLpEvent_Type_VirtualIo 8 -#define HvLpEvent_Type_NumTypes 9 - -#define HvLpEvent_Rc_Good 0 -#define HvLpEvent_Rc_BufferNotAvailable 1 -#define HvLpEvent_Rc_Cancelled 2 -#define HvLpEvent_Rc_GenericError 3 -#define HvLpEvent_Rc_InvalidAddress 4 -#define HvLpEvent_Rc_InvalidPartition 5 -#define HvLpEvent_Rc_InvalidSize 6 -#define HvLpEvent_Rc_InvalidSubtype 7 -#define HvLpEvent_Rc_InvalidSubtypeData 8 -#define HvLpEvent_Rc_InvalidType 9 -#define HvLpEvent_Rc_PartitionDead 10 -#define HvLpEvent_Rc_PathClosed 11 -#define HvLpEvent_Rc_SubtypeError 12 - -#define HvLpEvent_Function_Ack 0 -#define HvLpEvent_Function_Int 1 - -#define HvLpEvent_AckInd_NoAck 0 -#define HvLpEvent_AckInd_DoAck 1 - -#define HvLpEvent_AckType_ImmediateAck 0 -#define HvLpEvent_AckType_DeferredAck 1 - -#define HvLpDma_Direction_LocalToRemote 0 -#define HvLpDma_Direction_RemoteToLocal 1 - -#define HvLpDma_AddressType_TceIndex 0 -#define HvLpDma_AddressType_RealAddress 1 - -#define HvLpDma_Rc_Good 0 -#define HvLpDma_Rc_Error 1 -#define HvLpDma_Rc_PartitionDead 2 -#define HvLpDma_Rc_PathClosed 3 -#define HvLpDma_Rc_InvalidAddress 4 -#define HvLpDma_Rc_InvalidLength 5 - -#endif /* _HVLPEVENT_H */ diff --git a/include/asm-ppc64/iSeries/vio.h b/include/asm-ppc64/iSeries/vio.h index 6c05e6257f53..4d39ce6c687a 100644 --- a/include/asm-ppc64/iSeries/vio.h +++ b/include/asm-ppc64/iSeries/vio.h @@ -42,7 +42,7 @@ #define _ISERIES_VIO_H #include -#include +#include /* * iSeries virtual I/O events use the subtype field in -- cgit v1.2.3 From 7b487bb801b4bd790aa4bf7c157889f9d05e239c Mon Sep 17 00:00:00 2001 From: Kelly Daly Date: Wed, 2 Nov 2005 13:48:25 +1100 Subject: merge filename and modify references to iseries/it_exp_vpd_panel.h Signed-off-by: Kelly Daly --- arch/powerpc/platforms/iseries/lpardata.c | 2 +- arch/powerpc/platforms/iseries/viopath.c | 2 +- arch/ppc64/kernel/lparcfg.c | 2 +- include/asm-powerpc/iseries/it_exp_vpd_panel.h | 52 ++++++++++++++++++++++++++ include/asm-ppc64/iSeries/ItExtVpdPanel.h | 52 -------------------------- 5 files changed, 55 insertions(+), 55 deletions(-) create mode 100644 include/asm-powerpc/iseries/it_exp_vpd_panel.h delete mode 100644 include/asm-ppc64/iSeries/ItExtVpdPanel.h (limited to 'arch/ppc64') diff --git a/arch/powerpc/platforms/iseries/lpardata.c b/arch/powerpc/platforms/iseries/lpardata.c index ed2ffee6f731..9c8c21252142 100644 --- a/arch/powerpc/platforms/iseries/lpardata.c +++ b/arch/powerpc/platforms/iseries/lpardata.c @@ -20,7 +20,7 @@ #include #include #include -#include +#include #include #include "vpd_areas.h" diff --git a/arch/powerpc/platforms/iseries/viopath.c b/arch/powerpc/platforms/iseries/viopath.c index 17bf4e70d5cc..6fa39ae5f12c 100644 --- a/arch/powerpc/platforms/iseries/viopath.c +++ b/arch/powerpc/platforms/iseries/viopath.c @@ -42,7 +42,7 @@ #include #include #include -#include +#include #include #include #include diff --git a/arch/ppc64/kernel/lparcfg.c b/arch/ppc64/kernel/lparcfg.c index e4236f2707e1..e86155770bbc 100644 --- a/arch/ppc64/kernel/lparcfg.c +++ b/arch/ppc64/kernel/lparcfg.c @@ -33,7 +33,7 @@ #include #include #include -#include +#include #include #define MODULE_VERS "1.6" diff --git a/include/asm-powerpc/iseries/it_exp_vpd_panel.h b/include/asm-powerpc/iseries/it_exp_vpd_panel.h new file mode 100644 index 000000000000..4c546a8802b4 --- /dev/null +++ b/include/asm-powerpc/iseries/it_exp_vpd_panel.h @@ -0,0 +1,52 @@ +/* + * ItExtVpdPanel.h + * Copyright (C) 2002 Dave Boutcher IBM Corporation + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program; if not, write to the Free Software + * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA + */ +#ifndef _ITEXTVPDPANEL_H +#define _ITEXTVPDPANEL_H + +/* + * This struct maps the panel information + * + * Warning: + * This data must match the architecture for the panel information + */ + +#include + +struct ItExtVpdPanel { + /* Definition of the Extended Vpd On Panel Data Area */ + char systemSerial[8]; + char mfgID[4]; + char reserved1[24]; + char machineType[4]; + char systemID[6]; + char somUniqueCnt[4]; + char serialNumberCount; + char reserved2[7]; + u16 bbu3; + u16 bbu2; + u16 bbu1; + char xLocationLabel[8]; + u8 xRsvd1[6]; + u16 xFrameId; + u8 xRsvd2[48]; +}; + +extern struct ItExtVpdPanel xItExtVpdPanel; + +#endif /* _ITEXTVPDPANEL_H */ diff --git a/include/asm-ppc64/iSeries/ItExtVpdPanel.h b/include/asm-ppc64/iSeries/ItExtVpdPanel.h deleted file mode 100644 index 4c546a8802b4..000000000000 --- a/include/asm-ppc64/iSeries/ItExtVpdPanel.h +++ /dev/null @@ -1,52 +0,0 @@ -/* - * ItExtVpdPanel.h - * Copyright (C) 2002 Dave Boutcher IBM Corporation - * - * This program is free software; you can redistribute it and/or modify - * it under the terms of the GNU General Public License as published by - * the Free Software Foundation; either version 2 of the License, or - * (at your option) any later version. - * - * This program is distributed in the hope that it will be useful, - * but WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the - * GNU General Public License for more details. - * - * You should have received a copy of the GNU General Public License - * along with this program; if not, write to the Free Software - * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA - */ -#ifndef _ITEXTVPDPANEL_H -#define _ITEXTVPDPANEL_H - -/* - * This struct maps the panel information - * - * Warning: - * This data must match the architecture for the panel information - */ - -#include - -struct ItExtVpdPanel { - /* Definition of the Extended Vpd On Panel Data Area */ - char systemSerial[8]; - char mfgID[4]; - char reserved1[24]; - char machineType[4]; - char systemID[6]; - char somUniqueCnt[4]; - char serialNumberCount; - char reserved2[7]; - u16 bbu3; - u16 bbu2; - u16 bbu1; - char xLocationLabel[8]; - u8 xRsvd1[6]; - u16 xFrameId; - u8 xRsvd2[48]; -}; - -extern struct ItExtVpdPanel xItExtVpdPanel; - -#endif /* _ITEXTVPDPANEL_H */ -- cgit v1.2.3 From 8875ccfb7a6bd69d95a4e889ab36adda06c30d9e Mon Sep 17 00:00:00 2001 From: Kelly Daly Date: Wed, 2 Nov 2005 14:13:34 +1100 Subject: merge filename and modify references to iseries/it_lp_queue.h Signed-off-by: Kelly Daly --- arch/powerpc/kernel/time.c | 2 +- arch/powerpc/platforms/iseries/lpardata.c | 2 +- arch/powerpc/platforms/iseries/lpevents.c | 2 +- arch/powerpc/platforms/iseries/mf.c | 2 +- arch/powerpc/platforms/iseries/proc.c | 2 +- arch/powerpc/platforms/iseries/setup.c | 2 +- arch/ppc64/kernel/irq.c | 2 +- arch/ppc64/kernel/pacaData.c | 2 +- include/asm-powerpc/iseries/it_lp_queue.h | 81 +++++++++++++++++++++++++++++++ include/asm-ppc64/iSeries/ItLpQueue.h | 81 ------------------------------- 10 files changed, 89 insertions(+), 89 deletions(-) create mode 100644 include/asm-powerpc/iseries/it_lp_queue.h delete mode 100644 include/asm-ppc64/iSeries/ItLpQueue.h (limited to 'arch/ppc64') diff --git a/arch/powerpc/kernel/time.c b/arch/powerpc/kernel/time.c index 348746c55ebd..6996a593dcb3 100644 --- a/arch/powerpc/kernel/time.c +++ b/arch/powerpc/kernel/time.c @@ -66,7 +66,7 @@ #include #endif #ifdef CONFIG_PPC_ISERIES -#include +#include #include #endif diff --git a/arch/powerpc/platforms/iseries/lpardata.c b/arch/powerpc/platforms/iseries/lpardata.c index d0937613e776..dfe38eb4036d 100644 --- a/arch/powerpc/platforms/iseries/lpardata.c +++ b/arch/powerpc/platforms/iseries/lpardata.c @@ -21,7 +21,7 @@ #include #include #include -#include +#include #include "vpd_areas.h" #include "spcomm_area.h" diff --git a/arch/powerpc/platforms/iseries/lpevents.c b/arch/powerpc/platforms/iseries/lpevents.c index f0040d469a49..e9fb98bf895f 100644 --- a/arch/powerpc/platforms/iseries/lpevents.c +++ b/arch/powerpc/platforms/iseries/lpevents.c @@ -17,7 +17,7 @@ #include #include -#include +#include #include #include #include diff --git a/arch/powerpc/platforms/iseries/mf.c b/arch/powerpc/platforms/iseries/mf.c index 027df204446d..689f65a49c00 100644 --- a/arch/powerpc/platforms/iseries/mf.c +++ b/arch/powerpc/platforms/iseries/mf.c @@ -41,7 +41,7 @@ #include #include #include -#include +#include #include "setup.h" diff --git a/arch/powerpc/platforms/iseries/proc.c b/arch/powerpc/platforms/iseries/proc.c index beba0f62e653..e68b6b5fa89f 100644 --- a/arch/powerpc/platforms/iseries/proc.c +++ b/arch/powerpc/platforms/iseries/proc.c @@ -24,7 +24,7 @@ #include #include #include -#include +#include #include #include "processor_vpd.h" diff --git a/arch/powerpc/platforms/iseries/setup.c b/arch/powerpc/platforms/iseries/setup.c index 881d81fcf3b8..9a9e54b736d3 100644 --- a/arch/powerpc/platforms/iseries/setup.c +++ b/arch/powerpc/platforms/iseries/setup.c @@ -48,7 +48,7 @@ #include #include #include -#include +#include #include #include #include diff --git a/arch/ppc64/kernel/irq.c b/arch/ppc64/kernel/irq.c index f41afe545045..310931dbd4ae 100644 --- a/arch/ppc64/kernel/irq.c +++ b/arch/ppc64/kernel/irq.c @@ -52,7 +52,7 @@ #include #include #include -#include +#include #include #include diff --git a/arch/ppc64/kernel/pacaData.c b/arch/ppc64/kernel/pacaData.c index 33a2d8db3f21..5e27e5a6a35d 100644 --- a/arch/ppc64/kernel/pacaData.c +++ b/arch/ppc64/kernel/pacaData.c @@ -17,7 +17,7 @@ #include #include -#include +#include #include static union { diff --git a/include/asm-powerpc/iseries/it_lp_queue.h b/include/asm-powerpc/iseries/it_lp_queue.h new file mode 100644 index 000000000000..69b26ad74135 --- /dev/null +++ b/include/asm-powerpc/iseries/it_lp_queue.h @@ -0,0 +1,81 @@ +/* + * ItLpQueue.h + * Copyright (C) 2001 Mike Corrigan IBM Corporation + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program; if not, write to the Free Software + * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA + */ +#ifndef _ITLPQUEUE_H +#define _ITLPQUEUE_H + +/* + * This control block defines the simple LP queue structure that is + * shared between the hypervisor (PLIC) and the OS in order to send + * events to an LP. + */ + +#include +#include + +struct HvLpEvent; + +#define ITMaxLpQueues 8 + +#define NotUsed 0 // Queue will not be used by PLIC +#define DedicatedIo 1 // Queue dedicated to IO processor specified +#define DedicatedLp 2 // Queue dedicated to LP specified +#define Shared 3 // Queue shared for both IO and LP + +#define LpEventStackSize 4096 +#define LpEventMaxSize 256 +#define LpEventAlign 64 + +struct hvlpevent_queue { +/* + * The xSlicCurEventPtr is the pointer to the next event stack entry + * that will become valid. The OS must peek at this entry to determine + * if it is valid. PLIC will set the valid indicator as the very last + * store into that entry. + * + * When the OS has completed processing of the event then it will mark + * the event as invalid so that PLIC knows it can store into that event + * location again. + * + * If the event stack fills and there are overflow events, then PLIC + * will set the xPlicOverflowIntPending flag in which case the OS will + * have to fetch the additional LP events once they have drained the + * event stack. + * + * The first 16-bytes are known by both the OS and PLIC. The remainder + * of the cache line is for use by the OS. + */ + u8 xPlicOverflowIntPending;// 0x00 Overflow events are pending + u8 xPlicStatus; // 0x01 DedicatedIo or DedicatedLp or NotUsed + u16 xSlicLogicalProcIndex; // 0x02 Logical Proc Index for correlation + u8 xPlicRsvd[12]; // 0x04 + char *xSlicCurEventPtr; // 0x10 + char *xSlicLastValidEventPtr; // 0x18 + char *xSlicEventStackPtr; // 0x20 + u8 xIndex; // 0x28 unique sequential index. + u8 xSlicRsvd[3]; // 0x29-2b + spinlock_t lock; +}; + +extern struct hvlpevent_queue hvlpevent_queue; + +extern int hvlpevent_is_pending(void); +extern void process_hvlpevents(struct pt_regs *); +extern void setup_hvlpevent_queue(void); + +#endif /* _ITLPQUEUE_H */ diff --git a/include/asm-ppc64/iSeries/ItLpQueue.h b/include/asm-ppc64/iSeries/ItLpQueue.h deleted file mode 100644 index 69b26ad74135..000000000000 --- a/include/asm-ppc64/iSeries/ItLpQueue.h +++ /dev/null @@ -1,81 +0,0 @@ -/* - * ItLpQueue.h - * Copyright (C) 2001 Mike Corrigan IBM Corporation - * - * This program is free software; you can redistribute it and/or modify - * it under the terms of the GNU General Public License as published by - * the Free Software Foundation; either version 2 of the License, or - * (at your option) any later version. - * - * This program is distributed in the hope that it will be useful, - * but WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the - * GNU General Public License for more details. - * - * You should have received a copy of the GNU General Public License - * along with this program; if not, write to the Free Software - * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA - */ -#ifndef _ITLPQUEUE_H -#define _ITLPQUEUE_H - -/* - * This control block defines the simple LP queue structure that is - * shared between the hypervisor (PLIC) and the OS in order to send - * events to an LP. - */ - -#include -#include - -struct HvLpEvent; - -#define ITMaxLpQueues 8 - -#define NotUsed 0 // Queue will not be used by PLIC -#define DedicatedIo 1 // Queue dedicated to IO processor specified -#define DedicatedLp 2 // Queue dedicated to LP specified -#define Shared 3 // Queue shared for both IO and LP - -#define LpEventStackSize 4096 -#define LpEventMaxSize 256 -#define LpEventAlign 64 - -struct hvlpevent_queue { -/* - * The xSlicCurEventPtr is the pointer to the next event stack entry - * that will become valid. The OS must peek at this entry to determine - * if it is valid. PLIC will set the valid indicator as the very last - * store into that entry. - * - * When the OS has completed processing of the event then it will mark - * the event as invalid so that PLIC knows it can store into that event - * location again. - * - * If the event stack fills and there are overflow events, then PLIC - * will set the xPlicOverflowIntPending flag in which case the OS will - * have to fetch the additional LP events once they have drained the - * event stack. - * - * The first 16-bytes are known by both the OS and PLIC. The remainder - * of the cache line is for use by the OS. - */ - u8 xPlicOverflowIntPending;// 0x00 Overflow events are pending - u8 xPlicStatus; // 0x01 DedicatedIo or DedicatedLp or NotUsed - u16 xSlicLogicalProcIndex; // 0x02 Logical Proc Index for correlation - u8 xPlicRsvd[12]; // 0x04 - char *xSlicCurEventPtr; // 0x10 - char *xSlicLastValidEventPtr; // 0x18 - char *xSlicEventStackPtr; // 0x20 - u8 xIndex; // 0x28 unique sequential index. - u8 xSlicRsvd[3]; // 0x29-2b - spinlock_t lock; -}; - -extern struct hvlpevent_queue hvlpevent_queue; - -extern int hvlpevent_is_pending(void); -extern void process_hvlpevents(struct pt_regs *); -extern void setup_hvlpevent_queue(void); - -#endif /* _ITLPQUEUE_H */ -- cgit v1.2.3 From c43a55ff4e55d1bcfdf0cadfe2f00281e156c436 Mon Sep 17 00:00:00 2001 From: Kelly Daly Date: Wed, 2 Nov 2005 15:02:47 +1100 Subject: merge filename and modify references to iseries/lpar_map.h Signed-off-by: Kelly Daly --- arch/powerpc/kernel/head_64.S | 2 +- arch/powerpc/kernel/lparmap.c | 2 +- arch/powerpc/platforms/iseries/lpardata.c | 2 +- arch/powerpc/platforms/iseries/setup.c | 2 +- arch/ppc64/kernel/head.S | 2 +- include/asm-powerpc/iseries/lpar_map.h | 83 +++++++++++++++++++++++++++++++ include/asm-ppc64/iSeries/LparMap.h | 83 ------------------------------- 7 files changed, 88 insertions(+), 88 deletions(-) create mode 100644 include/asm-powerpc/iseries/lpar_map.h delete mode 100644 include/asm-ppc64/iSeries/LparMap.h (limited to 'arch/ppc64') diff --git a/arch/powerpc/kernel/head_64.S b/arch/powerpc/kernel/head_64.S index 147215a0d6c0..ba30e3177b05 100644 --- a/arch/powerpc/kernel/head_64.S +++ b/arch/powerpc/kernel/head_64.S @@ -35,7 +35,7 @@ #include #include #include -#include +#include #include #ifdef CONFIG_PPC_ISERIES diff --git a/arch/powerpc/kernel/lparmap.c b/arch/powerpc/kernel/lparmap.c index b81de286df5e..eded971d1bf9 100644 --- a/arch/powerpc/kernel/lparmap.c +++ b/arch/powerpc/kernel/lparmap.c @@ -8,7 +8,7 @@ */ #include #include -#include +#include const struct LparMap __attribute__((__section__(".text"))) xLparMap = { .xNumberEsids = HvEsidsToMap, diff --git a/arch/powerpc/platforms/iseries/lpardata.c b/arch/powerpc/platforms/iseries/lpardata.c index 6cd25b9b5e5a..608431a22607 100644 --- a/arch/powerpc/platforms/iseries/lpardata.c +++ b/arch/powerpc/platforms/iseries/lpardata.c @@ -19,7 +19,7 @@ #include #include #include -#include +#include #include #include diff --git a/arch/powerpc/platforms/iseries/setup.c b/arch/powerpc/platforms/iseries/setup.c index 9a9e54b736d3..758f5b50628a 100644 --- a/arch/powerpc/platforms/iseries/setup.c +++ b/arch/powerpc/platforms/iseries/setup.c @@ -51,7 +51,7 @@ #include #include #include -#include +#include #include "setup.h" #include "irq.h" diff --git a/arch/ppc64/kernel/head.S b/arch/ppc64/kernel/head.S index 929f9f42cf7a..0b65efe1682e 100644 --- a/arch/ppc64/kernel/head.S +++ b/arch/ppc64/kernel/head.S @@ -35,7 +35,7 @@ #include #include #include -#include +#include #include #ifdef CONFIG_PPC_ISERIES diff --git a/include/asm-powerpc/iseries/lpar_map.h b/include/asm-powerpc/iseries/lpar_map.h new file mode 100644 index 000000000000..a6840b186d03 --- /dev/null +++ b/include/asm-powerpc/iseries/lpar_map.h @@ -0,0 +1,83 @@ +/* + * LparMap.h + * Copyright (C) 2001 Mike Corrigan IBM Corporation + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program; if not, write to the Free Software + * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA + */ +#ifndef _LPARMAP_H +#define _LPARMAP_H + +#ifndef __ASSEMBLY__ + +#include + +/* + * The iSeries hypervisor will set up mapping for one or more + * ESID/VSID pairs (in SLB/segment registers) and will set up + * mappings of one or more ranges of pages to VAs. + * We will have the hypervisor set up the ESID->VSID mapping + * for the four kernel segments (C-F). With shared processors, + * the hypervisor will clear all segment registers and reload + * these four whenever the processor is switched from one + * partition to another. + */ + +/* The Vsid and Esid identified below will be used by the hypervisor + * to set up a memory mapping for part of the load area before giving + * control to the Linux kernel. The load area is 64 MB, but this must + * not attempt to map the whole load area. The Hashed Page Table may + * need to be located within the load area (if the total partition size + * is 64 MB), but cannot be mapped. Typically, this should specify + * to map half (32 MB) of the load area. + * + * The hypervisor will set up page table entries for the number of + * pages specified. + * + * In 32-bit mode, the hypervisor will load all four of the + * segment registers (identified by the low-order four bits of the + * Esid field. In 64-bit mode, the hypervisor will load one SLB + * entry to map the Esid to the Vsid. +*/ + +#define HvEsidsToMap 2 +#define HvRangesToMap 1 + +/* Hypervisor initially maps 32MB of the load area */ +#define HvPagesToMap 8192 + +struct LparMap { + u64 xNumberEsids; // Number of ESID/VSID pairs + u64 xNumberRanges; // Number of VA ranges to map + u64 xSegmentTableOffs; // Page number within load area of seg table + u64 xRsvd[5]; + struct { + u64 xKernelEsid; // Esid used to map kernel load + u64 xKernelVsid; // Vsid used to map kernel load + } xEsids[HvEsidsToMap]; + struct { + u64 xPages; // Number of pages to be mapped + u64 xOffset; // Offset from start of load area + u64 xVPN; // Virtual Page Number + } xRanges[HvRangesToMap]; +}; + +extern const struct LparMap xLparMap; + +#endif /* __ASSEMBLY__ */ + +/* the fixed address where the LparMap exists */ +#define LPARMAP_PHYS 0x7000 + +#endif /* _LPARMAP_H */ diff --git a/include/asm-ppc64/iSeries/LparMap.h b/include/asm-ppc64/iSeries/LparMap.h deleted file mode 100644 index a6840b186d03..000000000000 --- a/include/asm-ppc64/iSeries/LparMap.h +++ /dev/null @@ -1,83 +0,0 @@ -/* - * LparMap.h - * Copyright (C) 2001 Mike Corrigan IBM Corporation - * - * This program is free software; you can redistribute it and/or modify - * it under the terms of the GNU General Public License as published by - * the Free Software Foundation; either version 2 of the License, or - * (at your option) any later version. - * - * This program is distributed in the hope that it will be useful, - * but WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the - * GNU General Public License for more details. - * - * You should have received a copy of the GNU General Public License - * along with this program; if not, write to the Free Software - * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA - */ -#ifndef _LPARMAP_H -#define _LPARMAP_H - -#ifndef __ASSEMBLY__ - -#include - -/* - * The iSeries hypervisor will set up mapping for one or more - * ESID/VSID pairs (in SLB/segment registers) and will set up - * mappings of one or more ranges of pages to VAs. - * We will have the hypervisor set up the ESID->VSID mapping - * for the four kernel segments (C-F). With shared processors, - * the hypervisor will clear all segment registers and reload - * these four whenever the processor is switched from one - * partition to another. - */ - -/* The Vsid and Esid identified below will be used by the hypervisor - * to set up a memory mapping for part of the load area before giving - * control to the Linux kernel. The load area is 64 MB, but this must - * not attempt to map the whole load area. The Hashed Page Table may - * need to be located within the load area (if the total partition size - * is 64 MB), but cannot be mapped. Typically, this should specify - * to map half (32 MB) of the load area. - * - * The hypervisor will set up page table entries for the number of - * pages specified. - * - * In 32-bit mode, the hypervisor will load all four of the - * segment registers (identified by the low-order four bits of the - * Esid field. In 64-bit mode, the hypervisor will load one SLB - * entry to map the Esid to the Vsid. -*/ - -#define HvEsidsToMap 2 -#define HvRangesToMap 1 - -/* Hypervisor initially maps 32MB of the load area */ -#define HvPagesToMap 8192 - -struct LparMap { - u64 xNumberEsids; // Number of ESID/VSID pairs - u64 xNumberRanges; // Number of VA ranges to map - u64 xSegmentTableOffs; // Page number within load area of seg table - u64 xRsvd[5]; - struct { - u64 xKernelEsid; // Esid used to map kernel load - u64 xKernelVsid; // Vsid used to map kernel load - } xEsids[HvEsidsToMap]; - struct { - u64 xPages; // Number of pages to be mapped - u64 xOffset; // Offset from start of load area - u64 xVPN; // Virtual Page Number - } xRanges[HvRangesToMap]; -}; - -extern const struct LparMap xLparMap; - -#endif /* __ASSEMBLY__ */ - -/* the fixed address where the LparMap exists */ -#define LPARMAP_PHYS 0x7000 - -#endif /* _LPARMAP_H */ -- cgit v1.2.3 From f11b7bd88f8b41f0986498ffa11ffff1c3e513ca Mon Sep 17 00:00:00 2001 From: David Gibson Date: Tue, 1 Nov 2005 15:30:26 +1100 Subject: [PATCH] powerpc: Move naca.h to platforms/iseries These days, the NACA only exists on iSeries. Therefore, this patch moves naca.h from include/asm-ppc64 to arch/powerpc/platforms/iseries. There was one file including naca.h outside of platforms/iseries - arch/ppc64/kernel/udbg_scc.c. However, that's obviously a hangover from older days. The include is not necessary, so this patch simply removes it. Built and booted on iSeries, built for G5 (which uses udbg_scc.o). Signed-off-by: David Gibson Signed-off-by: Paul Mackerras --- arch/powerpc/platforms/iseries/lpardata.c | 2 +- arch/powerpc/platforms/iseries/naca.h | 24 ++++++++++++++++++++++++ arch/powerpc/platforms/iseries/release_data.h | 2 +- arch/powerpc/platforms/iseries/setup.c | 2 +- arch/ppc64/kernel/udbg_scc.c | 1 - include/asm-ppc64/naca.h | 24 ------------------------ 6 files changed, 27 insertions(+), 28 deletions(-) create mode 100644 arch/powerpc/platforms/iseries/naca.h delete mode 100644 include/asm-ppc64/naca.h (limited to 'arch/ppc64') diff --git a/arch/powerpc/platforms/iseries/lpardata.c b/arch/powerpc/platforms/iseries/lpardata.c index ed2ffee6f731..e856947bd16d 100644 --- a/arch/powerpc/platforms/iseries/lpardata.c +++ b/arch/powerpc/platforms/iseries/lpardata.c @@ -13,7 +13,6 @@ #include #include #include -#include #include #include #include @@ -23,6 +22,7 @@ #include #include +#include "naca.h" #include "vpd_areas.h" #include "spcomm_area.h" #include "ipl_parms.h" diff --git a/arch/powerpc/platforms/iseries/naca.h b/arch/powerpc/platforms/iseries/naca.h new file mode 100644 index 000000000000..ab2372eb8d2e --- /dev/null +++ b/arch/powerpc/platforms/iseries/naca.h @@ -0,0 +1,24 @@ +#ifndef _PLATFORMS_ISERIES_NACA_H +#define _PLATFORMS_ISERIES_NACA_H + +/* + * c 2001 PPC 64 Team, IBM Corp + * + * This program is free software; you can redistribute it and/or + * modify it under the terms of the GNU General Public License + * as published by the Free Software Foundation; either version + * 2 of the License, or (at your option) any later version. + */ + +#include + +struct naca_struct { + /* Kernel only data - undefined for user space */ + void *xItVpdAreas; /* VPD Data 0x00 */ + void *xRamDisk; /* iSeries ramdisk 0x08 */ + u64 xRamDiskSize; /* In pages 0x10 */ +}; + +extern struct naca_struct naca; + +#endif /* _PLATFORMS_ISERIES_NACA_H */ diff --git a/arch/powerpc/platforms/iseries/release_data.h b/arch/powerpc/platforms/iseries/release_data.h index c68b9c3e5caf..66189fd2e32d 100644 --- a/arch/powerpc/platforms/iseries/release_data.h +++ b/arch/powerpc/platforms/iseries/release_data.h @@ -24,7 +24,7 @@ * address of the OS's NACA). */ #include -#include +#include "naca.h" /* * When we IPL a secondary partition, we will check if if the diff --git a/arch/powerpc/platforms/iseries/setup.c b/arch/powerpc/platforms/iseries/setup.c index 06431e642533..7bc19a594cc6 100644 --- a/arch/powerpc/platforms/iseries/setup.c +++ b/arch/powerpc/platforms/iseries/setup.c @@ -41,7 +41,6 @@ #include #include -#include #include #include #include @@ -54,6 +53,7 @@ #include #include +#include "naca.h" #include "setup.h" #include "irq.h" #include "vpd_areas.h" diff --git a/arch/ppc64/kernel/udbg_scc.c b/arch/ppc64/kernel/udbg_scc.c index c47fd6c63531..820c53551507 100644 --- a/arch/ppc64/kernel/udbg_scc.c +++ b/arch/ppc64/kernel/udbg_scc.c @@ -12,7 +12,6 @@ #include #include #include -#include #include #include #include diff --git a/include/asm-ppc64/naca.h b/include/asm-ppc64/naca.h deleted file mode 100644 index d2afe6447597..000000000000 --- a/include/asm-ppc64/naca.h +++ /dev/null @@ -1,24 +0,0 @@ -#ifndef _NACA_H -#define _NACA_H - -/* - * c 2001 PPC 64 Team, IBM Corp - * - * This program is free software; you can redistribute it and/or - * modify it under the terms of the GNU General Public License - * as published by the Free Software Foundation; either version - * 2 of the License, or (at your option) any later version. - */ - -#include - -struct naca_struct { - /* Kernel only data - undefined for user space */ - void *xItVpdAreas; /* VPD Data 0x00 */ - void *xRamDisk; /* iSeries ramdisk 0x08 */ - u64 xRamDiskSize; /* In pages 0x10 */ -}; - -extern struct naca_struct naca; - -#endif /* _NACA_H */ -- cgit v1.2.3 From f4fcbbe9a7fdf11305fea657202b954bdc2228ec Mon Sep 17 00:00:00 2001 From: Paul Mackerras Date: Thu, 3 Nov 2005 14:41:19 +1100 Subject: powerpc: Merge remaining RTAS code This moves rtas-proc.c and rtas_flash.c into arch/powerpc/kernel, since cell wants them as well as pseries (and chrp can use rtas-proc.c too, at least in principle). rtas_fw.c is gone, with its bits moved into rtas_flash.c and rtas.c. Signed-off-by: Paul Mackerras --- arch/powerpc/Kconfig | 12 +- arch/powerpc/kernel/Makefile | 3 +- arch/powerpc/kernel/rtas-proc.c | 808 +++++++++++++++++++++++++++++++ arch/powerpc/kernel/rtas.c | 19 +- arch/powerpc/kernel/rtas_flash.c | 834 ++++++++++++++++++++++++++++++++ arch/powerpc/kernel/rtas_fw.c | 136 ------ arch/powerpc/platforms/pseries/Kconfig | 9 - arch/powerpc/platforms/pseries/Makefile | 2 +- arch/powerpc/platforms/pseries/rtasd.c | 527 ++++++++++++++++++++ arch/powerpc/platforms/pseries/setup.c | 6 +- arch/ppc64/kernel/Makefile | 4 +- arch/ppc64/kernel/rtas-proc.c | 808 ------------------------------- arch/ppc64/kernel/rtas_flash.c | 725 --------------------------- arch/ppc64/kernel/rtasd.c | 527 -------------------- include/asm-powerpc/rtas.h | 28 +- 15 files changed, 2205 insertions(+), 2243 deletions(-) create mode 100644 arch/powerpc/kernel/rtas-proc.c create mode 100644 arch/powerpc/kernel/rtas_flash.c delete mode 100644 arch/powerpc/kernel/rtas_fw.c create mode 100644 arch/powerpc/platforms/pseries/rtasd.c delete mode 100644 arch/ppc64/kernel/rtas-proc.c delete mode 100644 arch/ppc64/kernel/rtas_flash.c delete mode 100644 arch/ppc64/kernel/rtasd.c (limited to 'arch/ppc64') diff --git a/arch/powerpc/Kconfig b/arch/powerpc/Kconfig index 1c44a1dac421..3cf03ab46113 100644 --- a/arch/powerpc/Kconfig +++ b/arch/powerpc/Kconfig @@ -278,7 +278,6 @@ config PPC_PSERIES select PPC_I8259 select PPC_RTAS select RTAS_ERROR_LOGGING - select RTAS_FW default y config PPC_CHRP @@ -324,7 +323,6 @@ config PPC_CELL bool " Cell Broadband Processor Architecture" depends on PPC_MULTIPLATFORM && PPC64 select PPC_RTAS - select RTAS_FW select MMIO_NVRAM config PPC_OF @@ -356,10 +354,14 @@ config RTAS_ERROR_LOGGING depends on PPC_RTAS default n -config RTAS_FW - bool +config RTAS_PROC + bool "Proc interface to RTAS" depends on PPC_RTAS - default n + default y + +config RTAS_FLASH + tristate "Firmware flash interface" + depends on PPC64 && RTAS_PROC config MMIO_NVRAM bool diff --git a/arch/powerpc/kernel/Makefile b/arch/powerpc/kernel/Makefile index abad3059a21a..601ddbf1c20b 100644 --- a/arch/powerpc/kernel/Makefile +++ b/arch/powerpc/kernel/Makefile @@ -18,7 +18,8 @@ obj-$(CONFIG_ALTIVEC) += vecemu.o vector.o obj-$(CONFIG_POWER4) += idle_power4.o obj-$(CONFIG_PPC_OF) += of_device.o obj-$(CONFIG_PPC_RTAS) += rtas.o -obj-$(CONFIG_RTAS_FW) += rtas_fw.o +obj-$(CONFIG_RTAS_FLASH) += rtas_flash.o +obj-$(CONFIG_RTAS_PROC) += rtas-proc.o obj-$(CONFIG_IBMVIO) += vio.o ifeq ($(CONFIG_PPC_MERGE),y) diff --git a/arch/powerpc/kernel/rtas-proc.c b/arch/powerpc/kernel/rtas-proc.c new file mode 100644 index 000000000000..5bdd5b079d96 --- /dev/null +++ b/arch/powerpc/kernel/rtas-proc.c @@ -0,0 +1,808 @@ +/* + * arch/ppc64/kernel/rtas-proc.c + * Copyright (C) 2000 Tilmann Bitterberg + * (tilmann@bitterberg.de) + * + * RTAS (Runtime Abstraction Services) stuff + * Intention is to provide a clean user interface + * to use the RTAS. + * + * TODO: + * Split off a header file and maybe move it to a different + * location. Write Documentation on what the /proc/rtas/ entries + * actually do. + */ + +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include + +#include +#include +#include +#include +#include +#include /* for ppc_md */ +#include +#include + +/* Token for Sensors */ +#define KEY_SWITCH 0x0001 +#define ENCLOSURE_SWITCH 0x0002 +#define THERMAL_SENSOR 0x0003 +#define LID_STATUS 0x0004 +#define POWER_SOURCE 0x0005 +#define BATTERY_VOLTAGE 0x0006 +#define BATTERY_REMAINING 0x0007 +#define BATTERY_PERCENTAGE 0x0008 +#define EPOW_SENSOR 0x0009 +#define BATTERY_CYCLESTATE 0x000a +#define BATTERY_CHARGING 0x000b + +/* IBM specific sensors */ +#define IBM_SURVEILLANCE 0x2328 /* 9000 */ +#define IBM_FANRPM 0x2329 /* 9001 */ +#define IBM_VOLTAGE 0x232a /* 9002 */ +#define IBM_DRCONNECTOR 0x232b /* 9003 */ +#define IBM_POWERSUPPLY 0x232c /* 9004 */ + +/* Status return values */ +#define SENSOR_CRITICAL_HIGH 13 +#define SENSOR_WARNING_HIGH 12 +#define SENSOR_NORMAL 11 +#define SENSOR_WARNING_LOW 10 +#define SENSOR_CRITICAL_LOW 9 +#define SENSOR_SUCCESS 0 +#define SENSOR_HW_ERROR -1 +#define SENSOR_BUSY -2 +#define SENSOR_NOT_EXIST -3 +#define SENSOR_DR_ENTITY -9000 + +/* Location Codes */ +#define LOC_SCSI_DEV_ADDR 'A' +#define LOC_SCSI_DEV_LOC 'B' +#define LOC_CPU 'C' +#define LOC_DISKETTE 'D' +#define LOC_ETHERNET 'E' +#define LOC_FAN 'F' +#define LOC_GRAPHICS 'G' +/* reserved / not used 'H' */ +#define LOC_IO_ADAPTER 'I' +/* reserved / not used 'J' */ +#define LOC_KEYBOARD 'K' +#define LOC_LCD 'L' +#define LOC_MEMORY 'M' +#define LOC_NV_MEMORY 'N' +#define LOC_MOUSE 'O' +#define LOC_PLANAR 'P' +#define LOC_OTHER_IO 'Q' +#define LOC_PARALLEL 'R' +#define LOC_SERIAL 'S' +#define LOC_DEAD_RING 'T' +#define LOC_RACKMOUNTED 'U' /* for _u_nit is rack mounted */ +#define LOC_VOLTAGE 'V' +#define LOC_SWITCH_ADAPTER 'W' +#define LOC_OTHER 'X' +#define LOC_FIRMWARE 'Y' +#define LOC_SCSI 'Z' + +/* Tokens for indicators */ +#define TONE_FREQUENCY 0x0001 /* 0 - 1000 (HZ)*/ +#define TONE_VOLUME 0x0002 /* 0 - 100 (%) */ +#define SYSTEM_POWER_STATE 0x0003 +#define WARNING_LIGHT 0x0004 +#define DISK_ACTIVITY_LIGHT 0x0005 +#define HEX_DISPLAY_UNIT 0x0006 +#define BATTERY_WARNING_TIME 0x0007 +#define CONDITION_CYCLE_REQUEST 0x0008 +#define SURVEILLANCE_INDICATOR 0x2328 /* 9000 */ +#define DR_ACTION 0x2329 /* 9001 */ +#define DR_INDICATOR 0x232a /* 9002 */ +/* 9003 - 9004: Vendor specific */ +/* 9006 - 9999: Vendor specific */ + +/* other */ +#define MAX_SENSORS 17 /* I only know of 17 sensors */ +#define MAX_LINELENGTH 256 +#define SENSOR_PREFIX "ibm,sensor-" +#define cel_to_fahr(x) ((x*9/5)+32) + + +/* Globals */ +static struct rtas_sensors sensors; +static struct device_node *rtas_node = NULL; +static unsigned long power_on_time = 0; /* Save the time the user set */ +static char progress_led[MAX_LINELENGTH]; + +static unsigned long rtas_tone_frequency = 1000; +static unsigned long rtas_tone_volume = 0; + +/* ****************STRUCTS******************************************* */ +struct individual_sensor { + unsigned int token; + unsigned int quant; +}; + +struct rtas_sensors { + struct individual_sensor sensor[MAX_SENSORS]; + unsigned int quant; +}; + +/* ****************************************************************** */ +/* Declarations */ +static int ppc_rtas_sensors_show(struct seq_file *m, void *v); +static int ppc_rtas_clock_show(struct seq_file *m, void *v); +static ssize_t ppc_rtas_clock_write(struct file *file, + const char __user *buf, size_t count, loff_t *ppos); +static int ppc_rtas_progress_show(struct seq_file *m, void *v); +static ssize_t ppc_rtas_progress_write(struct file *file, + const char __user *buf, size_t count, loff_t *ppos); +static int ppc_rtas_poweron_show(struct seq_file *m, void *v); +static ssize_t ppc_rtas_poweron_write(struct file *file, + const char __user *buf, size_t count, loff_t *ppos); + +static ssize_t ppc_rtas_tone_freq_write(struct file *file, + const char __user *buf, size_t count, loff_t *ppos); +static int ppc_rtas_tone_freq_show(struct seq_file *m, void *v); +static ssize_t ppc_rtas_tone_volume_write(struct file *file, + const char __user *buf, size_t count, loff_t *ppos); +static int ppc_rtas_tone_volume_show(struct seq_file *m, void *v); +static int ppc_rtas_rmo_buf_show(struct seq_file *m, void *v); + +static int sensors_open(struct inode *inode, struct file *file) +{ + return single_open(file, ppc_rtas_sensors_show, NULL); +} + +struct file_operations ppc_rtas_sensors_operations = { + .open = sensors_open, + .read = seq_read, + .llseek = seq_lseek, + .release = single_release, +}; + +static int poweron_open(struct inode *inode, struct file *file) +{ + return single_open(file, ppc_rtas_poweron_show, NULL); +} + +struct file_operations ppc_rtas_poweron_operations = { + .open = poweron_open, + .read = seq_read, + .llseek = seq_lseek, + .write = ppc_rtas_poweron_write, + .release = single_release, +}; + +static int progress_open(struct inode *inode, struct file *file) +{ + return single_open(file, ppc_rtas_progress_show, NULL); +} + +struct file_operations ppc_rtas_progress_operations = { + .open = progress_open, + .read = seq_read, + .llseek = seq_lseek, + .write = ppc_rtas_progress_write, + .release = single_release, +}; + +static int clock_open(struct inode *inode, struct file *file) +{ + return single_open(file, ppc_rtas_clock_show, NULL); +} + +struct file_operations ppc_rtas_clock_operations = { + .open = clock_open, + .read = seq_read, + .llseek = seq_lseek, + .write = ppc_rtas_clock_write, + .release = single_release, +}; + +static int tone_freq_open(struct inode *inode, struct file *file) +{ + return single_open(file, ppc_rtas_tone_freq_show, NULL); +} + +struct file_operations ppc_rtas_tone_freq_operations = { + .open = tone_freq_open, + .read = seq_read, + .llseek = seq_lseek, + .write = ppc_rtas_tone_freq_write, + .release = single_release, +}; + +static int tone_volume_open(struct inode *inode, struct file *file) +{ + return single_open(file, ppc_rtas_tone_volume_show, NULL); +} + +struct file_operations ppc_rtas_tone_volume_operations = { + .open = tone_volume_open, + .read = seq_read, + .llseek = seq_lseek, + .write = ppc_rtas_tone_volume_write, + .release = single_release, +}; + +static int rmo_buf_open(struct inode *inode, struct file *file) +{ + return single_open(file, ppc_rtas_rmo_buf_show, NULL); +} + +struct file_operations ppc_rtas_rmo_buf_ops = { + .open = rmo_buf_open, + .read = seq_read, + .llseek = seq_lseek, + .release = single_release, +}; + +static int ppc_rtas_find_all_sensors(void); +static void ppc_rtas_process_sensor(struct seq_file *m, + struct individual_sensor *s, int state, int error, char *loc); +static char *ppc_rtas_process_error(int error); +static void get_location_code(struct seq_file *m, + struct individual_sensor *s, char *loc); +static void check_location_string(struct seq_file *m, char *c); +static void check_location(struct seq_file *m, char *c); + +static int __init proc_rtas_init(void) +{ + struct proc_dir_entry *entry; + + if (!(systemcfg->platform & PLATFORM_PSERIES)) + return 1; + + rtas_node = of_find_node_by_name(NULL, "rtas"); + if (rtas_node == NULL) + return 1; + + entry = create_proc_entry("ppc64/rtas/progress", S_IRUGO|S_IWUSR, NULL); + if (entry) + entry->proc_fops = &ppc_rtas_progress_operations; + + entry = create_proc_entry("ppc64/rtas/clock", S_IRUGO|S_IWUSR, NULL); + if (entry) + entry->proc_fops = &ppc_rtas_clock_operations; + + entry = create_proc_entry("ppc64/rtas/poweron", S_IWUSR|S_IRUGO, NULL); + if (entry) + entry->proc_fops = &ppc_rtas_poweron_operations; + + entry = create_proc_entry("ppc64/rtas/sensors", S_IRUGO, NULL); + if (entry) + entry->proc_fops = &ppc_rtas_sensors_operations; + + entry = create_proc_entry("ppc64/rtas/frequency", S_IWUSR|S_IRUGO, + NULL); + if (entry) + entry->proc_fops = &ppc_rtas_tone_freq_operations; + + entry = create_proc_entry("ppc64/rtas/volume", S_IWUSR|S_IRUGO, NULL); + if (entry) + entry->proc_fops = &ppc_rtas_tone_volume_operations; + + entry = create_proc_entry("ppc64/rtas/rmo_buffer", S_IRUSR, NULL); + if (entry) + entry->proc_fops = &ppc_rtas_rmo_buf_ops; + + return 0; +} + +__initcall(proc_rtas_init); + +static int parse_number(const char __user *p, size_t count, unsigned long *val) +{ + char buf[40]; + char *end; + + if (count > 39) + return -EINVAL; + + if (copy_from_user(buf, p, count)) + return -EFAULT; + + buf[count] = 0; + + *val = simple_strtoul(buf, &end, 10); + if (*end && *end != '\n') + return -EINVAL; + + return 0; +} + +/* ****************************************************************** */ +/* POWER-ON-TIME */ +/* ****************************************************************** */ +static ssize_t ppc_rtas_poweron_write(struct file *file, + const char __user *buf, size_t count, loff_t *ppos) +{ + struct rtc_time tm; + unsigned long nowtime; + int error = parse_number(buf, count, &nowtime); + if (error) + return error; + + power_on_time = nowtime; /* save the time */ + + to_tm(nowtime, &tm); + + error = rtas_call(rtas_token("set-time-for-power-on"), 7, 1, NULL, + tm.tm_year, tm.tm_mon, tm.tm_mday, + tm.tm_hour, tm.tm_min, tm.tm_sec, 0 /* nano */); + if (error) + printk(KERN_WARNING "error: setting poweron time returned: %s\n", + ppc_rtas_process_error(error)); + return count; +} +/* ****************************************************************** */ +static int ppc_rtas_poweron_show(struct seq_file *m, void *v) +{ + if (power_on_time == 0) + seq_printf(m, "Power on time not set\n"); + else + seq_printf(m, "%lu\n",power_on_time); + return 0; +} + +/* ****************************************************************** */ +/* PROGRESS */ +/* ****************************************************************** */ +static ssize_t ppc_rtas_progress_write(struct file *file, + const char __user *buf, size_t count, loff_t *ppos) +{ + unsigned long hex; + + if (count >= MAX_LINELENGTH) + count = MAX_LINELENGTH -1; + if (copy_from_user(progress_led, buf, count)) { /* save the string */ + return -EFAULT; + } + progress_led[count] = 0; + + /* Lets see if the user passed hexdigits */ + hex = simple_strtoul(progress_led, NULL, 10); + + rtas_progress ((char *)progress_led, hex); + return count; + + /* clear the line */ + /* rtas_progress(" ", 0xffff);*/ +} +/* ****************************************************************** */ +static int ppc_rtas_progress_show(struct seq_file *m, void *v) +{ + if (progress_led) + seq_printf(m, "%s\n", progress_led); + return 0; +} + +/* ****************************************************************** */ +/* CLOCK */ +/* ****************************************************************** */ +static ssize_t ppc_rtas_clock_write(struct file *file, + const char __user *buf, size_t count, loff_t *ppos) +{ + struct rtc_time tm; + unsigned long nowtime; + int error = parse_number(buf, count, &nowtime); + if (error) + return error; + + to_tm(nowtime, &tm); + error = rtas_call(rtas_token("set-time-of-day"), 7, 1, NULL, + tm.tm_year, tm.tm_mon, tm.tm_mday, + tm.tm_hour, tm.tm_min, tm.tm_sec, 0); + if (error) + printk(KERN_WARNING "error: setting the clock returned: %s\n", + ppc_rtas_process_error(error)); + return count; +} +/* ****************************************************************** */ +static int ppc_rtas_clock_show(struct seq_file *m, void *v) +{ + int ret[8]; + int error = rtas_call(rtas_token("get-time-of-day"), 0, 8, ret); + + if (error) { + printk(KERN_WARNING "error: reading the clock returned: %s\n", + ppc_rtas_process_error(error)); + seq_printf(m, "0"); + } else { + unsigned int year, mon, day, hour, min, sec; + year = ret[0]; mon = ret[1]; day = ret[2]; + hour = ret[3]; min = ret[4]; sec = ret[5]; + seq_printf(m, "%lu\n", + mktime(year, mon, day, hour, min, sec)); + } + return 0; +} + +/* ****************************************************************** */ +/* SENSOR STUFF */ +/* ****************************************************************** */ +static int ppc_rtas_sensors_show(struct seq_file *m, void *v) +{ + int i,j; + int state, error; + int get_sensor_state = rtas_token("get-sensor-state"); + + seq_printf(m, "RTAS (RunTime Abstraction Services) Sensor Information\n"); + seq_printf(m, "Sensor\t\tValue\t\tCondition\tLocation\n"); + seq_printf(m, "********************************************************\n"); + + if (ppc_rtas_find_all_sensors() != 0) { + seq_printf(m, "\nNo sensors are available\n"); + return 0; + } + + for (i=0; itoken); + loc = (char *) get_property(rtas_node, rstr, &llen); + + /* A sensor may have multiple instances */ + for (j = 0, offs = 0; j <= p->quant; j++) { + error = rtas_call(get_sensor_state, 2, 2, &state, + p->token, j); + + ppc_rtas_process_sensor(m, p, state, error, loc); + seq_putc(m, '\n'); + if (loc) { + offs += strlen(loc) + 1; + loc += strlen(loc) + 1; + if (offs >= llen) + loc = NULL; + } + } + } + return 0; +} + +/* ****************************************************************** */ + +static int ppc_rtas_find_all_sensors(void) +{ + unsigned int *utmp; + int len, i; + + utmp = (unsigned int *) get_property(rtas_node, "rtas-sensors", &len); + if (utmp == NULL) { + printk (KERN_ERR "error: could not get rtas-sensors\n"); + return 1; + } + + sensors.quant = len / 8; /* int + int */ + + for (i=0; itoken) { + case KEY_SWITCH: + seq_printf(m, "Key switch:\t"); + num_states = sizeof(key_switch) / sizeof(char *); + if (state < num_states) { + seq_printf(m, "%s\t", key_switch[state]); + have_strings = 1; + } + break; + case ENCLOSURE_SWITCH: + seq_printf(m, "Enclosure switch:\t"); + num_states = sizeof(enclosure_switch) / sizeof(char *); + if (state < num_states) { + seq_printf(m, "%s\t", + enclosure_switch[state]); + have_strings = 1; + } + break; + case THERMAL_SENSOR: + seq_printf(m, "Temp. (C/F):\t"); + temperature = 1; + break; + case LID_STATUS: + seq_printf(m, "Lid status:\t"); + num_states = sizeof(lid_status) / sizeof(char *); + if (state < num_states) { + seq_printf(m, "%s\t", lid_status[state]); + have_strings = 1; + } + break; + case POWER_SOURCE: + seq_printf(m, "Power source:\t"); + num_states = sizeof(power_source) / sizeof(char *); + if (state < num_states) { + seq_printf(m, "%s\t", + power_source[state]); + have_strings = 1; + } + break; + case BATTERY_VOLTAGE: + seq_printf(m, "Battery voltage:\t"); + break; + case BATTERY_REMAINING: + seq_printf(m, "Battery remaining:\t"); + num_states = sizeof(battery_remaining) / sizeof(char *); + if (state < num_states) + { + seq_printf(m, "%s\t", + battery_remaining[state]); + have_strings = 1; + } + break; + case BATTERY_PERCENTAGE: + seq_printf(m, "Battery percentage:\t"); + break; + case EPOW_SENSOR: + seq_printf(m, "EPOW Sensor:\t"); + num_states = sizeof(epow_sensor) / sizeof(char *); + if (state < num_states) { + seq_printf(m, "%s\t", epow_sensor[state]); + have_strings = 1; + } + break; + case BATTERY_CYCLESTATE: + seq_printf(m, "Battery cyclestate:\t"); + num_states = sizeof(battery_cyclestate) / + sizeof(char *); + if (state < num_states) { + seq_printf(m, "%s\t", + battery_cyclestate[state]); + have_strings = 1; + } + break; + case BATTERY_CHARGING: + seq_printf(m, "Battery Charging:\t"); + num_states = sizeof(battery_charging) / sizeof(char *); + if (state < num_states) { + seq_printf(m, "%s\t", + battery_charging[state]); + have_strings = 1; + } + break; + case IBM_SURVEILLANCE: + seq_printf(m, "Surveillance:\t"); + break; + case IBM_FANRPM: + seq_printf(m, "Fan (rpm):\t"); + break; + case IBM_VOLTAGE: + seq_printf(m, "Voltage (mv):\t"); + break; + case IBM_DRCONNECTOR: + seq_printf(m, "DR connector:\t"); + num_states = sizeof(ibm_drconnector) / sizeof(char *); + if (state < num_states) { + seq_printf(m, "%s\t", + ibm_drconnector[state]); + have_strings = 1; + } + break; + case IBM_POWERSUPPLY: + seq_printf(m, "Powersupply:\t"); + break; + default: + seq_printf(m, "Unknown sensor (type %d), ignoring it\n", + s->token); + unknown = 1; + have_strings = 1; + break; + } + if (have_strings == 0) { + if (temperature) { + seq_printf(m, "%4d /%4d\t", state, cel_to_fahr(state)); + } else + seq_printf(m, "%10d\t", state); + } + if (unknown == 0) { + seq_printf(m, "%s\t", ppc_rtas_process_error(error)); + get_location_code(m, s, loc); + } +} + +/* ****************************************************************** */ + +static void check_location(struct seq_file *m, char *c) +{ + switch (c[0]) { + case LOC_PLANAR: + seq_printf(m, "Planar #%c", c[1]); + break; + case LOC_CPU: + seq_printf(m, "CPU #%c", c[1]); + break; + case LOC_FAN: + seq_printf(m, "Fan #%c", c[1]); + break; + case LOC_RACKMOUNTED: + seq_printf(m, "Rack #%c", c[1]); + break; + case LOC_VOLTAGE: + seq_printf(m, "Voltage #%c", c[1]); + break; + case LOC_LCD: + seq_printf(m, "LCD #%c", c[1]); + break; + case '.': + seq_printf(m, "- %c", c[1]); + break; + default: + seq_printf(m, "Unknown location"); + break; + } +} + + +/* ****************************************************************** */ +/* + * Format: + * ${LETTER}${NUMBER}[[-/]${LETTER}${NUMBER} [ ... ] ] + * the '.' may be an abbrevation + */ +static void check_location_string(struct seq_file *m, char *c) +{ + while (*c) { + if (isalpha(*c) || *c == '.') + check_location(m, c); + else if (*c == '/' || *c == '-') + seq_printf(m, " at "); + c++; + } +} + + +/* ****************************************************************** */ + +static void get_location_code(struct seq_file *m, struct individual_sensor *s, char *loc) +{ + if (!loc || !*loc) { + seq_printf(m, "---");/* does not have a location */ + } else { + check_location_string(m, loc); + } + seq_putc(m, ' '); +} +/* ****************************************************************** */ +/* INDICATORS - Tone Frequency */ +/* ****************************************************************** */ +static ssize_t ppc_rtas_tone_freq_write(struct file *file, + const char __user *buf, size_t count, loff_t *ppos) +{ + unsigned long freq; + int error = parse_number(buf, count, &freq); + if (error) + return error; + + rtas_tone_frequency = freq; /* save it for later */ + error = rtas_call(rtas_token("set-indicator"), 3, 1, NULL, + TONE_FREQUENCY, 0, freq); + if (error) + printk(KERN_WARNING "error: setting tone frequency returned: %s\n", + ppc_rtas_process_error(error)); + return count; +} +/* ****************************************************************** */ +static int ppc_rtas_tone_freq_show(struct seq_file *m, void *v) +{ + seq_printf(m, "%lu\n", rtas_tone_frequency); + return 0; +} +/* ****************************************************************** */ +/* INDICATORS - Tone Volume */ +/* ****************************************************************** */ +static ssize_t ppc_rtas_tone_volume_write(struct file *file, + const char __user *buf, size_t count, loff_t *ppos) +{ + unsigned long volume; + int error = parse_number(buf, count, &volume); + if (error) + return error; + + if (volume > 100) + volume = 100; + + rtas_tone_volume = volume; /* save it for later */ + error = rtas_call(rtas_token("set-indicator"), 3, 1, NULL, + TONE_VOLUME, 0, volume); + if (error) + printk(KERN_WARNING "error: setting tone volume returned: %s\n", + ppc_rtas_process_error(error)); + return count; +} +/* ****************************************************************** */ +static int ppc_rtas_tone_volume_show(struct seq_file *m, void *v) +{ + seq_printf(m, "%lu\n", rtas_tone_volume); + return 0; +} + +#define RMO_READ_BUF_MAX 30 + +/* RTAS Userspace access */ +static int ppc_rtas_rmo_buf_show(struct seq_file *m, void *v) +{ + seq_printf(m, "%016lx %x\n", rtas_rmo_buf, RTAS_RMOBUF_MAX); + return 0; +} diff --git a/arch/powerpc/kernel/rtas.c b/arch/powerpc/kernel/rtas.c index 4d22eeeeb91d..b7fc2d884950 100644 --- a/arch/powerpc/kernel/rtas.c +++ b/arch/powerpc/kernel/rtas.c @@ -42,6 +42,13 @@ DEFINE_SPINLOCK(rtas_data_buf_lock); char rtas_data_buf[RTAS_DATA_BUF_SIZE] __cacheline_aligned; unsigned long rtas_rmo_buf; +/* + * If non-NULL, this gets called when the kernel terminates. + * This is done like this so rtas_flash can be a module. + */ +void (*rtas_flash_term_hook)(int); +EXPORT_SYMBOL(rtas_flash_term_hook); + /* * call_rtas_display_status and call_rtas_display_status_delay * are designed only for very early low-level debugging, which @@ -206,6 +213,7 @@ void rtas_progress(char *s, unsigned short hex) spin_unlock(&progress_lock); } +EXPORT_SYMBOL(rtas_progress); /* needed by rtas_flash module */ int rtas_token(const char *service) { @@ -492,6 +500,8 @@ int rtas_set_indicator(int indicator, int index, int new_value) void rtas_restart(char *cmd) { + if (rtas_flash_term_hook) + rtas_flash_term_hook(SYS_RESTART); printk("RTAS system-reboot returned %d\n", rtas_call(rtas_token("system-reboot"), 0, 1, NULL)); for (;;); @@ -499,6 +509,8 @@ void rtas_restart(char *cmd) void rtas_power_off(void) { + if (rtas_flash_term_hook) + rtas_flash_term_hook(SYS_POWER_OFF); /* allow power on only with power button press */ printk("RTAS power-off returned %d\n", rtas_call(rtas_token("power-off"), 2, 1, NULL, -1, -1)); @@ -507,7 +519,12 @@ void rtas_power_off(void) void rtas_halt(void) { - rtas_power_off(); + if (rtas_flash_term_hook) + rtas_flash_term_hook(SYS_HALT); + /* allow power on only with power button press */ + printk("RTAS power-off returned %d\n", + rtas_call(rtas_token("power-off"), 2, 1, NULL, -1, -1)); + for (;;); } /* Must be in the RMO region, so we place it here */ diff --git a/arch/powerpc/kernel/rtas_flash.c b/arch/powerpc/kernel/rtas_flash.c new file mode 100644 index 000000000000..50500093c97f --- /dev/null +++ b/arch/powerpc/kernel/rtas_flash.c @@ -0,0 +1,834 @@ +/* + * c 2001 PPC 64 Team, IBM Corp + * + * This program is free software; you can redistribute it and/or + * modify it under the terms of the GNU General Public License + * as published by the Free Software Foundation; either version + * 2 of the License, or (at your option) any later version. + * + * /proc/ppc64/rtas/firmware_flash interface + * + * This file implements a firmware_flash interface to pump a firmware + * image into the kernel. At reboot time rtas_restart() will see the + * firmware image and flash it as it reboots (see rtas.c). + */ + +#include +#include +#include +#include +#include +#include +#include + +#define MODULE_VERS "1.0" +#define MODULE_NAME "rtas_flash" + +#define FIRMWARE_FLASH_NAME "firmware_flash" +#define FIRMWARE_UPDATE_NAME "firmware_update" +#define MANAGE_FLASH_NAME "manage_flash" +#define VALIDATE_FLASH_NAME "validate_flash" + +/* General RTAS Status Codes */ +#define RTAS_RC_SUCCESS 0 +#define RTAS_RC_HW_ERR -1 +#define RTAS_RC_BUSY -2 + +/* Flash image status values */ +#define FLASH_AUTH -9002 /* RTAS Not Service Authority Partition */ +#define FLASH_NO_OP -1099 /* No operation initiated by user */ +#define FLASH_IMG_SHORT -1005 /* Flash image shorter than expected */ +#define FLASH_IMG_BAD_LEN -1004 /* Bad length value in flash list block */ +#define FLASH_IMG_NULL_DATA -1003 /* Bad data value in flash list block */ +#define FLASH_IMG_READY 0 /* Firmware img ready for flash on reboot */ + +/* Manage image status values */ +#define MANAGE_AUTH -9002 /* RTAS Not Service Authority Partition */ +#define MANAGE_ACTIVE_ERR -9001 /* RTAS Cannot Overwrite Active Img */ +#define MANAGE_NO_OP -1099 /* No operation initiated by user */ +#define MANAGE_PARAM_ERR -3 /* RTAS Parameter Error */ +#define MANAGE_HW_ERR -1 /* RTAS Hardware Error */ + +/* Validate image status values */ +#define VALIDATE_AUTH -9002 /* RTAS Not Service Authority Partition */ +#define VALIDATE_NO_OP -1099 /* No operation initiated by the user */ +#define VALIDATE_INCOMPLETE -1002 /* User copied < VALIDATE_BUF_SIZE */ +#define VALIDATE_READY -1001 /* Firmware image ready for validation */ +#define VALIDATE_PARAM_ERR -3 /* RTAS Parameter Error */ +#define VALIDATE_HW_ERR -1 /* RTAS Hardware Error */ +#define VALIDATE_TMP_UPDATE 0 /* Validate Return Status */ +#define VALIDATE_FLASH_AUTH 1 /* Validate Return Status */ +#define VALIDATE_INVALID_IMG 2 /* Validate Return Status */ +#define VALIDATE_CUR_UNKNOWN 3 /* Validate Return Status */ +#define VALIDATE_TMP_COMMIT_DL 4 /* Validate Return Status */ +#define VALIDATE_TMP_COMMIT 5 /* Validate Return Status */ +#define VALIDATE_TMP_UPDATE_DL 6 /* Validate Return Status */ + +/* ibm,manage-flash-image operation tokens */ +#define RTAS_REJECT_TMP_IMG 0 +#define RTAS_COMMIT_TMP_IMG 1 + +/* Array sizes */ +#define VALIDATE_BUF_SIZE 4096 +#define RTAS_MSG_MAXLEN 64 + +struct flash_block { + char *data; + unsigned long length; +}; + +/* This struct is very similar but not identical to + * that needed by the rtas flash update. + * All we need to do for rtas is rewrite num_blocks + * into a version/length and translate the pointers + * to absolute. + */ +#define FLASH_BLOCKS_PER_NODE ((PAGE_SIZE - 16) / sizeof(struct flash_block)) +struct flash_block_list { + unsigned long num_blocks; + struct flash_block_list *next; + struct flash_block blocks[FLASH_BLOCKS_PER_NODE]; +}; +struct flash_block_list_header { /* just the header of flash_block_list */ + unsigned long num_blocks; + struct flash_block_list *next; +}; + +static struct flash_block_list_header rtas_firmware_flash_list = {0, NULL}; + +#define FLASH_BLOCK_LIST_VERSION (1UL) + +/* Local copy of the flash block list. + * We only allow one open of the flash proc file and create this + * list as we go. This list will be put in the + * rtas_firmware_flash_list var once it is fully read. + * + * For convenience as we build the list we use virtual addrs, + * we do not fill in the version number, and the length field + * is treated as the number of entries currently in the block + * (i.e. not a byte count). This is all fixed on release. + */ + +/* Status int must be first member of struct */ +struct rtas_update_flash_t +{ + int status; /* Flash update status */ + struct flash_block_list *flist; /* Local copy of flash block list */ +}; + +/* Status int must be first member of struct */ +struct rtas_manage_flash_t +{ + int status; /* Returned status */ + unsigned int op; /* Reject or commit image */ +}; + +/* Status int must be first member of struct */ +struct rtas_validate_flash_t +{ + int status; /* Returned status */ + char buf[VALIDATE_BUF_SIZE]; /* Candidate image buffer */ + unsigned int buf_size; /* Size of image buf */ + unsigned int update_results; /* Update results token */ +}; + +static DEFINE_SPINLOCK(flash_file_open_lock); +static struct proc_dir_entry *firmware_flash_pde; +static struct proc_dir_entry *firmware_update_pde; +static struct proc_dir_entry *validate_pde; +static struct proc_dir_entry *manage_pde; + +/* Do simple sanity checks on the flash image. */ +static int flash_list_valid(struct flash_block_list *flist) +{ + struct flash_block_list *f; + int i; + unsigned long block_size, image_size; + + /* Paranoid self test here. We also collect the image size. */ + image_size = 0; + for (f = flist; f; f = f->next) { + for (i = 0; i < f->num_blocks; i++) { + if (f->blocks[i].data == NULL) { + return FLASH_IMG_NULL_DATA; + } + block_size = f->blocks[i].length; + if (block_size <= 0 || block_size > PAGE_SIZE) { + return FLASH_IMG_BAD_LEN; + } + image_size += block_size; + } + } + + if (image_size < (256 << 10)) { + if (image_size < 2) + return FLASH_NO_OP; + } + + printk(KERN_INFO "FLASH: flash image with %ld bytes stored for hardware flash on reboot\n", image_size); + + return FLASH_IMG_READY; +} + +static void free_flash_list(struct flash_block_list *f) +{ + struct flash_block_list *next; + int i; + + while (f) { + for (i = 0; i < f->num_blocks; i++) + free_page((unsigned long)(f->blocks[i].data)); + next = f->next; + free_page((unsigned long)f); + f = next; + } +} + +static int rtas_flash_release(struct inode *inode, struct file *file) +{ + struct proc_dir_entry *dp = PDE(file->f_dentry->d_inode); + struct rtas_update_flash_t *uf; + + uf = (struct rtas_update_flash_t *) dp->data; + if (uf->flist) { + /* File was opened in write mode for a new flash attempt */ + /* Clear saved list */ + if (rtas_firmware_flash_list.next) { + free_flash_list(rtas_firmware_flash_list.next); + rtas_firmware_flash_list.next = NULL; + } + + if (uf->status != FLASH_AUTH) + uf->status = flash_list_valid(uf->flist); + + if (uf->status == FLASH_IMG_READY) + rtas_firmware_flash_list.next = uf->flist; + else + free_flash_list(uf->flist); + + uf->flist = NULL; + } + + atomic_dec(&dp->count); + return 0; +} + +static void get_flash_status_msg(int status, char *buf) +{ + char *msg; + + switch (status) { + case FLASH_AUTH: + msg = "error: this partition does not have service authority\n"; + break; + case FLASH_NO_OP: + msg = "info: no firmware image for flash\n"; + break; + case FLASH_IMG_SHORT: + msg = "error: flash image short\n"; + break; + case FLASH_IMG_BAD_LEN: + msg = "error: internal error bad length\n"; + break; + case FLASH_IMG_NULL_DATA: + msg = "error: internal error null data\n"; + break; + case FLASH_IMG_READY: + msg = "ready: firmware image ready for flash on reboot\n"; + break; + default: + sprintf(buf, "error: unexpected status value %d\n", status); + return; + } + + strcpy(buf, msg); +} + +/* Reading the proc file will show status (not the firmware contents) */ +static ssize_t rtas_flash_read(struct file *file, char __user *buf, + size_t count, loff_t *ppos) +{ + struct proc_dir_entry *dp = PDE(file->f_dentry->d_inode); + struct rtas_update_flash_t *uf; + char msg[RTAS_MSG_MAXLEN]; + int msglen; + + uf = (struct rtas_update_flash_t *) dp->data; + + if (!strcmp(dp->name, FIRMWARE_FLASH_NAME)) { + get_flash_status_msg(uf->status, msg); + } else { /* FIRMWARE_UPDATE_NAME */ + sprintf(msg, "%d\n", uf->status); + } + msglen = strlen(msg); + if (msglen > count) + msglen = count; + + if (ppos && *ppos != 0) + return 0; /* be cheap */ + + if (!access_ok(VERIFY_WRITE, buf, msglen)) + return -EINVAL; + + if (copy_to_user(buf, msg, msglen)) + return -EFAULT; + + if (ppos) + *ppos = msglen; + return msglen; +} + +/* We could be much more efficient here. But to keep this function + * simple we allocate a page to the block list no matter how small the + * count is. If the system is low on memory it will be just as well + * that we fail.... + */ +static ssize_t rtas_flash_write(struct file *file, const char __user *buffer, + size_t count, loff_t *off) +{ + struct proc_dir_entry *dp = PDE(file->f_dentry->d_inode); + struct rtas_update_flash_t *uf; + char *p; + int next_free; + struct flash_block_list *fl; + + uf = (struct rtas_update_flash_t *) dp->data; + + if (uf->status == FLASH_AUTH || count == 0) + return count; /* discard data */ + + /* In the case that the image is not ready for flashing, the memory + * allocated for the block list will be freed upon the release of the + * proc file + */ + if (uf->flist == NULL) { + uf->flist = (struct flash_block_list *) get_zeroed_page(GFP_KERNEL); + if (!uf->flist) + return -ENOMEM; + } + + fl = uf->flist; + while (fl->next) + fl = fl->next; /* seek to last block_list for append */ + next_free = fl->num_blocks; + if (next_free == FLASH_BLOCKS_PER_NODE) { + /* Need to allocate another block_list */ + fl->next = (struct flash_block_list *)get_zeroed_page(GFP_KERNEL); + if (!fl->next) + return -ENOMEM; + fl = fl->next; + next_free = 0; + } + + if (count > PAGE_SIZE) + count = PAGE_SIZE; + p = (char *)get_zeroed_page(GFP_KERNEL); + if (!p) + return -ENOMEM; + + if(copy_from_user(p, buffer, count)) { + free_page((unsigned long)p); + return -EFAULT; + } + fl->blocks[next_free].data = p; + fl->blocks[next_free].length = count; + fl->num_blocks++; + + return count; +} + +static int rtas_excl_open(struct inode *inode, struct file *file) +{ + struct proc_dir_entry *dp = PDE(inode); + + /* Enforce exclusive open with use count of PDE */ + spin_lock(&flash_file_open_lock); + if (atomic_read(&dp->count) > 1) { + spin_unlock(&flash_file_open_lock); + return -EBUSY; + } + + atomic_inc(&dp->count); + spin_unlock(&flash_file_open_lock); + + return 0; +} + +static int rtas_excl_release(struct inode *inode, struct file *file) +{ + struct proc_dir_entry *dp = PDE(inode); + + atomic_dec(&dp->count); + + return 0; +} + +static void manage_flash(struct rtas_manage_flash_t *args_buf) +{ + unsigned int wait_time; + s32 rc; + + while (1) { + rc = rtas_call(rtas_token("ibm,manage-flash-image"), 1, + 1, NULL, args_buf->op); + if (rc == RTAS_RC_BUSY) + udelay(1); + else if (rtas_is_extended_busy(rc)) { + wait_time = rtas_extended_busy_delay_time(rc); + udelay(wait_time * 1000); + } else + break; + } + + args_buf->status = rc; +} + +static ssize_t manage_flash_read(struct file *file, char __user *buf, + size_t count, loff_t *ppos) +{ + struct proc_dir_entry *dp = PDE(file->f_dentry->d_inode); + struct rtas_manage_flash_t *args_buf; + char msg[RTAS_MSG_MAXLEN]; + int msglen; + + args_buf = (struct rtas_manage_flash_t *) dp->data; + if (args_buf == NULL) + return 0; + + msglen = sprintf(msg, "%d\n", args_buf->status); + if (msglen > count) + msglen = count; + + if (ppos && *ppos != 0) + return 0; /* be cheap */ + + if (!access_ok(VERIFY_WRITE, buf, msglen)) + return -EINVAL; + + if (copy_to_user(buf, msg, msglen)) + return -EFAULT; + + if (ppos) + *ppos = msglen; + return msglen; +} + +static ssize_t manage_flash_write(struct file *file, const char __user *buf, + size_t count, loff_t *off) +{ + struct proc_dir_entry *dp = PDE(file->f_dentry->d_inode); + struct rtas_manage_flash_t *args_buf; + const char reject_str[] = "0"; + const char commit_str[] = "1"; + char stkbuf[10]; + int op; + + args_buf = (struct rtas_manage_flash_t *) dp->data; + if ((args_buf->status == MANAGE_AUTH) || (count == 0)) + return count; + + op = -1; + if (buf) { + if (count > 9) count = 9; + if (copy_from_user (stkbuf, buf, count)) { + return -EFAULT; + } + if (strncmp(stkbuf, reject_str, strlen(reject_str)) == 0) + op = RTAS_REJECT_TMP_IMG; + else if (strncmp(stkbuf, commit_str, strlen(commit_str)) == 0) + op = RTAS_COMMIT_TMP_IMG; + } + + if (op == -1) /* buf is empty, or contains invalid string */ + return -EINVAL; + + args_buf->op = op; + manage_flash(args_buf); + + return count; +} + +static void validate_flash(struct rtas_validate_flash_t *args_buf) +{ + int token = rtas_token("ibm,validate-flash-image"); + unsigned int wait_time; + int update_results; + s32 rc; + + rc = 0; + while(1) { + spin_lock(&rtas_data_buf_lock); + memcpy(rtas_data_buf, args_buf->buf, VALIDATE_BUF_SIZE); + rc = rtas_call(token, 2, 2, &update_results, + (u32) __pa(rtas_data_buf), args_buf->buf_size); + memcpy(args_buf->buf, rtas_data_buf, VALIDATE_BUF_SIZE); + spin_unlock(&rtas_data_buf_lock); + + if (rc == RTAS_RC_BUSY) + udelay(1); + else if (rtas_is_extended_busy(rc)) { + wait_time = rtas_extended_busy_delay_time(rc); + udelay(wait_time * 1000); + } else + break; + } + + args_buf->status = rc; + args_buf->update_results = update_results; +} + +static int get_validate_flash_msg(struct rtas_validate_flash_t *args_buf, + char *msg) +{ + int n; + + if (args_buf->status >= VALIDATE_TMP_UPDATE) { + n = sprintf(msg, "%d\n", args_buf->update_results); + if ((args_buf->update_results >= VALIDATE_CUR_UNKNOWN) || + (args_buf->update_results == VALIDATE_TMP_UPDATE)) + n += sprintf(msg + n, "%s\n", args_buf->buf); + } else { + n = sprintf(msg, "%d\n", args_buf->status); + } + return n; +} + +static ssize_t validate_flash_read(struct file *file, char __user *buf, + size_t count, loff_t *ppos) +{ + struct proc_dir_entry *dp = PDE(file->f_dentry->d_inode); + struct rtas_validate_flash_t *args_buf; + char msg[RTAS_MSG_MAXLEN]; + int msglen; + + args_buf = (struct rtas_validate_flash_t *) dp->data; + + if (ppos && *ppos != 0) + return 0; /* be cheap */ + + msglen = get_validate_flash_msg(args_buf, msg); + if (msglen > count) + msglen = count; + + if (!access_ok(VERIFY_WRITE, buf, msglen)) + return -EINVAL; + + if (copy_to_user(buf, msg, msglen)) + return -EFAULT; + + if (ppos) + *ppos = msglen; + return msglen; +} + +static ssize_t validate_flash_write(struct file *file, const char __user *buf, + size_t count, loff_t *off) +{ + struct proc_dir_entry *dp = PDE(file->f_dentry->d_inode); + struct rtas_validate_flash_t *args_buf; + int rc; + + args_buf = (struct rtas_validate_flash_t *) dp->data; + + if (dp->data == NULL) { + dp->data = kmalloc(sizeof(struct rtas_validate_flash_t), + GFP_KERNEL); + if (dp->data == NULL) + return -ENOMEM; + } + + /* We are only interested in the first 4K of the + * candidate image */ + if ((*off >= VALIDATE_BUF_SIZE) || + (args_buf->status == VALIDATE_AUTH)) { + *off += count; + return count; + } + + if (*off + count >= VALIDATE_BUF_SIZE) { + count = VALIDATE_BUF_SIZE - *off; + args_buf->status = VALIDATE_READY; + } else { + args_buf->status = VALIDATE_INCOMPLETE; + } + + if (!access_ok(VERIFY_READ, buf, count)) { + rc = -EFAULT; + goto done; + } + if (copy_from_user(args_buf->buf + *off, buf, count)) { + rc = -EFAULT; + goto done; + } + + *off += count; + rc = count; +done: + if (rc < 0) { + kfree(dp->data); + dp->data = NULL; + } + return rc; +} + +static int validate_flash_release(struct inode *inode, struct file *file) +{ + struct proc_dir_entry *dp = PDE(file->f_dentry->d_inode); + struct rtas_validate_flash_t *args_buf; + + args_buf = (struct rtas_validate_flash_t *) dp->data; + + if (args_buf->status == VALIDATE_READY) { + args_buf->buf_size = VALIDATE_BUF_SIZE; + validate_flash(args_buf); + } + + /* The matching atomic_inc was in rtas_excl_open() */ + atomic_dec(&dp->count); + + return 0; +} + +static void rtas_flash_firmware(int reboot_type) +{ + unsigned long image_size; + struct flash_block_list *f, *next, *flist; + unsigned long rtas_block_list; + int i, status, update_token; + + if (rtas_firmware_flash_list.next == NULL) + return; /* nothing to do */ + + if (reboot_type != SYS_RESTART) { + printk(KERN_ALERT "FLASH: firmware flash requires a reboot\n"); + printk(KERN_ALERT "FLASH: the firmware image will NOT be flashed\n"); + return; + } + + update_token = rtas_token("ibm,update-flash-64-and-reboot"); + if (update_token == RTAS_UNKNOWN_SERVICE) { + printk(KERN_ALERT "FLASH: ibm,update-flash-64-and-reboot " + "is not available -- not a service partition?\n"); + printk(KERN_ALERT "FLASH: firmware will not be flashed\n"); + return; + } + + /* NOTE: the "first" block list is a global var with no data + * blocks in the kernel data segment. We do this because + * we want to ensure this block_list addr is under 4GB. + */ + rtas_firmware_flash_list.num_blocks = 0; + flist = (struct flash_block_list *)&rtas_firmware_flash_list; + rtas_block_list = virt_to_abs(flist); + if (rtas_block_list >= 4UL*1024*1024*1024) { + printk(KERN_ALERT "FLASH: kernel bug...flash list header addr above 4GB\n"); + return; + } + + printk(KERN_ALERT "FLASH: preparing saved firmware image for flash\n"); + /* Update the block_list in place. */ + image_size = 0; + for (f = flist; f; f = next) { + /* Translate data addrs to absolute */ + for (i = 0; i < f->num_blocks; i++) { + f->blocks[i].data = (char *)virt_to_abs(f->blocks[i].data); + image_size += f->blocks[i].length; + } + next = f->next; + /* Don't translate NULL pointer for last entry */ + if (f->next) + f->next = (struct flash_block_list *)virt_to_abs(f->next); + else + f->next = NULL; + /* make num_blocks into the version/length field */ + f->num_blocks = (FLASH_BLOCK_LIST_VERSION << 56) | ((f->num_blocks+1)*16); + } + + printk(KERN_ALERT "FLASH: flash image is %ld bytes\n", image_size); + printk(KERN_ALERT "FLASH: performing flash and reboot\n"); + rtas_progress("Flashing \n", 0x0); + rtas_progress("Please Wait... ", 0x0); + printk(KERN_ALERT "FLASH: this will take several minutes. Do not power off!\n"); + status = rtas_call(update_token, 1, 1, NULL, rtas_block_list); + switch (status) { /* should only get "bad" status */ + case 0: + printk(KERN_ALERT "FLASH: success\n"); + break; + case -1: + printk(KERN_ALERT "FLASH: hardware error. Firmware may not be not flashed\n"); + break; + case -3: + printk(KERN_ALERT "FLASH: image is corrupt or not correct for this platform. Firmware not flashed\n"); + break; + case -4: + printk(KERN_ALERT "FLASH: flash failed when partially complete. System may not reboot\n"); + break; + default: + printk(KERN_ALERT "FLASH: unknown flash return code %d\n", status); + break; + } +} + +static void remove_flash_pde(struct proc_dir_entry *dp) +{ + if (dp) { + if (dp->data != NULL) + kfree(dp->data); + dp->owner = NULL; + remove_proc_entry(dp->name, dp->parent); + } +} + +static int initialize_flash_pde_data(const char *rtas_call_name, + size_t buf_size, + struct proc_dir_entry *dp) +{ + int *status; + int token; + + dp->data = kmalloc(buf_size, GFP_KERNEL); + if (dp->data == NULL) { + remove_flash_pde(dp); + return -ENOMEM; + } + + memset(dp->data, 0, buf_size); + + /* + * This code assumes that the status int is the first member of the + * struct + */ + status = (int *) dp->data; + token = rtas_token(rtas_call_name); + if (token == RTAS_UNKNOWN_SERVICE) + *status = FLASH_AUTH; + else + *status = FLASH_NO_OP; + + return 0; +} + +static struct proc_dir_entry *create_flash_pde(const char *filename, + struct file_operations *fops) +{ + struct proc_dir_entry *ent = NULL; + + ent = create_proc_entry(filename, S_IRUSR | S_IWUSR, NULL); + if (ent != NULL) { + ent->nlink = 1; + ent->proc_fops = fops; + ent->owner = THIS_MODULE; + } + + return ent; +} + +static struct file_operations rtas_flash_operations = { + .read = rtas_flash_read, + .write = rtas_flash_write, + .open = rtas_excl_open, + .release = rtas_flash_release, +}; + +static struct file_operations manage_flash_operations = { + .read = manage_flash_read, + .write = manage_flash_write, + .open = rtas_excl_open, + .release = rtas_excl_release, +}; + +static struct file_operations validate_flash_operations = { + .read = validate_flash_read, + .write = validate_flash_write, + .open = rtas_excl_open, + .release = validate_flash_release, +}; + +int __init rtas_flash_init(void) +{ + int rc; + + if (rtas_token("ibm,update-flash-64-and-reboot") == + RTAS_UNKNOWN_SERVICE) { + printk(KERN_ERR "rtas_flash: no firmware flash support\n"); + return 1; + } + + firmware_flash_pde = create_flash_pde("ppc64/rtas/" + FIRMWARE_FLASH_NAME, + &rtas_flash_operations); + if (firmware_flash_pde == NULL) { + rc = -ENOMEM; + goto cleanup; + } + + rc = initialize_flash_pde_data("ibm,update-flash-64-and-reboot", + sizeof(struct rtas_update_flash_t), + firmware_flash_pde); + if (rc != 0) + goto cleanup; + + firmware_update_pde = create_flash_pde("ppc64/rtas/" + FIRMWARE_UPDATE_NAME, + &rtas_flash_operations); + if (firmware_update_pde == NULL) { + rc = -ENOMEM; + goto cleanup; + } + + rc = initialize_flash_pde_data("ibm,update-flash-64-and-reboot", + sizeof(struct rtas_update_flash_t), + firmware_update_pde); + if (rc != 0) + goto cleanup; + + validate_pde = create_flash_pde("ppc64/rtas/" VALIDATE_FLASH_NAME, + &validate_flash_operations); + if (validate_pde == NULL) { + rc = -ENOMEM; + goto cleanup; + } + + rc = initialize_flash_pde_data("ibm,validate-flash-image", + sizeof(struct rtas_validate_flash_t), + validate_pde); + if (rc != 0) + goto cleanup; + + manage_pde = create_flash_pde("ppc64/rtas/" MANAGE_FLASH_NAME, + &manage_flash_operations); + if (manage_pde == NULL) { + rc = -ENOMEM; + goto cleanup; + } + + rc = initialize_flash_pde_data("ibm,manage-flash-image", + sizeof(struct rtas_manage_flash_t), + manage_pde); + if (rc != 0) + goto cleanup; + + rtas_flash_term_hook = rtas_flash_firmware; + return 0; + +cleanup: + remove_flash_pde(firmware_flash_pde); + remove_flash_pde(firmware_update_pde); + remove_flash_pde(validate_pde); + remove_flash_pde(manage_pde); + + return rc; +} + +void __exit rtas_flash_cleanup(void) +{ + rtas_flash_term_hook = NULL; + remove_flash_pde(firmware_flash_pde); + remove_flash_pde(firmware_update_pde); + remove_flash_pde(validate_pde); + remove_flash_pde(manage_pde); +} + +module_init(rtas_flash_init); +module_exit(rtas_flash_cleanup); +MODULE_LICENSE("GPL"); diff --git a/arch/powerpc/kernel/rtas_fw.c b/arch/powerpc/kernel/rtas_fw.c deleted file mode 100644 index 448922e8af1b..000000000000 --- a/arch/powerpc/kernel/rtas_fw.c +++ /dev/null @@ -1,136 +0,0 @@ -/* - * - * Procedures for firmware flash updates. - * - * Peter Bergner, IBM March 2001. - * Copyright (C) 2001 IBM. - * - * This program is free software; you can redistribute it and/or - * modify it under the terms of the GNU General Public License - * as published by the Free Software Foundation; either version - * 2 of the License, or (at your option) any later version. - */ - -#include -#include -#include -#include -#include -#include - -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include - -struct flash_block_list_header rtas_firmware_flash_list = {0, NULL}; - -#define FLASH_BLOCK_LIST_VERSION (1UL) - -static void rtas_flash_firmware(void) -{ - unsigned long image_size; - struct flash_block_list *f, *next, *flist; - unsigned long rtas_block_list; - int i, status, update_token; - - update_token = rtas_token("ibm,update-flash-64-and-reboot"); - if (update_token == RTAS_UNKNOWN_SERVICE) { - printk(KERN_ALERT "FLASH: ibm,update-flash-64-and-reboot is not available -- not a service partition?\n"); - printk(KERN_ALERT "FLASH: firmware will not be flashed\n"); - return; - } - - /* NOTE: the "first" block list is a global var with no data - * blocks in the kernel data segment. We do this because - * we want to ensure this block_list addr is under 4GB. - */ - rtas_firmware_flash_list.num_blocks = 0; - flist = (struct flash_block_list *)&rtas_firmware_flash_list; - rtas_block_list = virt_to_abs(flist); - if (rtas_block_list >= 4UL*1024*1024*1024) { - printk(KERN_ALERT "FLASH: kernel bug...flash list header addr above 4GB\n"); - return; - } - - printk(KERN_ALERT "FLASH: preparing saved firmware image for flash\n"); - /* Update the block_list in place. */ - image_size = 0; - for (f = flist; f; f = next) { - /* Translate data addrs to absolute */ - for (i = 0; i < f->num_blocks; i++) { - f->blocks[i].data = (char *)virt_to_abs(f->blocks[i].data); - image_size += f->blocks[i].length; - } - next = f->next; - /* Don't translate NULL pointer for last entry */ - if (f->next) - f->next = (struct flash_block_list *)virt_to_abs(f->next); - else - f->next = NULL; - /* make num_blocks into the version/length field */ - f->num_blocks = (FLASH_BLOCK_LIST_VERSION << 56) | ((f->num_blocks+1)*16); - } - - printk(KERN_ALERT "FLASH: flash image is %ld bytes\n", image_size); - printk(KERN_ALERT "FLASH: performing flash and reboot\n"); - rtas_progress("Flashing \n", 0x0); - rtas_progress("Please Wait... ", 0x0); - printk(KERN_ALERT "FLASH: this will take several minutes. Do not power off!\n"); - status = rtas_call(update_token, 1, 1, NULL, rtas_block_list); - switch (status) { /* should only get "bad" status */ - case 0: - printk(KERN_ALERT "FLASH: success\n"); - break; - case -1: - printk(KERN_ALERT "FLASH: hardware error. Firmware may not be not flashed\n"); - break; - case -3: - printk(KERN_ALERT "FLASH: image is corrupt or not correct for this platform. Firmware not flashed\n"); - break; - case -4: - printk(KERN_ALERT "FLASH: flash failed when partially complete. System may not reboot\n"); - break; - default: - printk(KERN_ALERT "FLASH: unknown flash return code %d\n", status); - break; - } -} - -void rtas_flash_bypass_warning(void) -{ - printk(KERN_ALERT "FLASH: firmware flash requires a reboot\n"); - printk(KERN_ALERT "FLASH: the firmware image will NOT be flashed\n"); -} - - -void rtas_fw_restart(char *cmd) -{ - if (rtas_firmware_flash_list.next) - rtas_flash_firmware(); - rtas_restart(cmd); -} - -void rtas_fw_power_off(void) -{ - if (rtas_firmware_flash_list.next) - rtas_flash_bypass_warning(); - rtas_power_off(); -} - -void rtas_fw_halt(void) -{ - if (rtas_firmware_flash_list.next) - rtas_flash_bypass_warning(); - rtas_halt(); -} - -EXPORT_SYMBOL(rtas_firmware_flash_list); diff --git a/arch/powerpc/platforms/pseries/Kconfig b/arch/powerpc/platforms/pseries/Kconfig index 2d57f588151d..e3fc3407bb1f 100644 --- a/arch/powerpc/platforms/pseries/Kconfig +++ b/arch/powerpc/platforms/pseries/Kconfig @@ -21,15 +21,6 @@ config EEH depends on PPC_PSERIES default y if !EMBEDDED -config RTAS_PROC - bool "Proc interface to RTAS" - depends on PPC_RTAS - default y - -config RTAS_FLASH - tristate "Firmware flash interface" - depends on PPC64 && RTAS_PROC - config SCANLOG tristate "Scanlog dump interface" depends on RTAS_PROC && PPC_PSERIES diff --git a/arch/powerpc/platforms/pseries/Makefile b/arch/powerpc/platforms/pseries/Makefile index 91909a844736..b9938fece781 100644 --- a/arch/powerpc/platforms/pseries/Makefile +++ b/arch/powerpc/platforms/pseries/Makefile @@ -1,5 +1,5 @@ obj-y := pci.o lpar.o hvCall.o nvram.o reconfig.o \ - setup.o iommu.o ras.o + setup.o iommu.o ras.o rtasd.o obj-$(CONFIG_SMP) += smp.o obj-$(CONFIG_IBMVIO) += vio.o obj-$(CONFIG_XICS) += xics.o diff --git a/arch/powerpc/platforms/pseries/rtasd.c b/arch/powerpc/platforms/pseries/rtasd.c new file mode 100644 index 000000000000..e26b0420b6dd --- /dev/null +++ b/arch/powerpc/platforms/pseries/rtasd.c @@ -0,0 +1,527 @@ +/* + * Copyright (C) 2001 Anton Blanchard , IBM + * + * This program is free software; you can redistribute it and/or + * modify it under the terms of the GNU General Public License + * as published by the Free Software Foundation; either version + * 2 of the License, or (at your option) any later version. + * + * Communication to userspace based on kernel/printk.c + */ + +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include + +#include +#include +#include +#include +#include +#include +#include + +#if 0 +#define DEBUG(A...) printk(KERN_ERR A) +#else +#define DEBUG(A...) +#endif + +static DEFINE_SPINLOCK(rtasd_log_lock); + +DECLARE_WAIT_QUEUE_HEAD(rtas_log_wait); + +static char *rtas_log_buf; +static unsigned long rtas_log_start; +static unsigned long rtas_log_size; + +static int surveillance_timeout = -1; +static unsigned int rtas_event_scan_rate; +static unsigned int rtas_error_log_max; +static unsigned int rtas_error_log_buffer_max; + +static int full_rtas_msgs = 0; + +extern int no_logging; + +volatile int error_log_cnt = 0; + +/* + * Since we use 32 bit RTAS, the physical address of this must be below + * 4G or else bad things happen. Allocate this in the kernel data and + * make it big enough. + */ +static unsigned char logdata[RTAS_ERROR_LOG_MAX]; + +static int get_eventscan_parms(void); + +static char *rtas_type[] = { + "Unknown", "Retry", "TCE Error", "Internal Device Failure", + "Timeout", "Data Parity", "Address Parity", "Cache Parity", + "Address Invalid", "ECC Uncorrected", "ECC Corrupted", +}; + +static char *rtas_event_type(int type) +{ + if ((type > 0) && (type < 11)) + return rtas_type[type]; + + switch (type) { + case RTAS_TYPE_EPOW: + return "EPOW"; + case RTAS_TYPE_PLATFORM: + return "Platform Error"; + case RTAS_TYPE_IO: + return "I/O Event"; + case RTAS_TYPE_INFO: + return "Platform Information Event"; + case RTAS_TYPE_DEALLOC: + return "Resource Deallocation Event"; + case RTAS_TYPE_DUMP: + return "Dump Notification Event"; + } + + return rtas_type[0]; +} + +/* To see this info, grep RTAS /var/log/messages and each entry + * will be collected together with obvious begin/end. + * There will be a unique identifier on the begin and end lines. + * This will persist across reboots. + * + * format of error logs returned from RTAS: + * bytes (size) : contents + * -------------------------------------------------------- + * 0-7 (8) : rtas_error_log + * 8-47 (40) : extended info + * 48-51 (4) : vendor id + * 52-1023 (vendor specific) : location code and debug data + */ +static void printk_log_rtas(char *buf, int len) +{ + + int i,j,n = 0; + int perline = 16; + char buffer[64]; + char * str = "RTAS event"; + + if (full_rtas_msgs) { + printk(RTAS_DEBUG "%d -------- %s begin --------\n", + error_log_cnt, str); + + /* + * Print perline bytes on each line, each line will start + * with RTAS and a changing number, so syslogd will + * print lines that are otherwise the same. Separate every + * 4 bytes with a space. + */ + for (i = 0; i < len; i++) { + j = i % perline; + if (j == 0) { + memset(buffer, 0, sizeof(buffer)); + n = sprintf(buffer, "RTAS %d:", i/perline); + } + + if ((i % 4) == 0) + n += sprintf(buffer+n, " "); + + n += sprintf(buffer+n, "%02x", (unsigned char)buf[i]); + + if (j == (perline-1)) + printk(KERN_DEBUG "%s\n", buffer); + } + if ((i % perline) != 0) + printk(KERN_DEBUG "%s\n", buffer); + + printk(RTAS_DEBUG "%d -------- %s end ----------\n", + error_log_cnt, str); + } else { + struct rtas_error_log *errlog = (struct rtas_error_log *)buf; + + printk(RTAS_DEBUG "event: %d, Type: %s, Severity: %d\n", + error_log_cnt, rtas_event_type(errlog->type), + errlog->severity); + } +} + +static int log_rtas_len(char * buf) +{ + int len; + struct rtas_error_log *err; + + /* rtas fixed header */ + len = 8; + err = (struct rtas_error_log *)buf; + if (err->extended_log_length) { + + /* extended header */ + len += err->extended_log_length; + } + + if (rtas_error_log_max == 0) { + get_eventscan_parms(); + } + if (len > rtas_error_log_max) + len = rtas_error_log_max; + + return len; +} + +/* + * First write to nvram, if fatal error, that is the only + * place we log the info. The error will be picked up + * on the next reboot by rtasd. If not fatal, run the + * method for the type of error. Currently, only RTAS + * errors have methods implemented, but in the future + * there might be a need to store data in nvram before a + * call to panic(). + * + * XXX We write to nvram periodically, to indicate error has + * been written and sync'd, but there is a possibility + * that if we don't shutdown correctly, a duplicate error + * record will be created on next reboot. + */ +void pSeries_log_error(char *buf, unsigned int err_type, int fatal) +{ + unsigned long offset; + unsigned long s; + int len = 0; + + DEBUG("logging event\n"); + if (buf == NULL) + return; + + spin_lock_irqsave(&rtasd_log_lock, s); + + /* get length and increase count */ + switch (err_type & ERR_TYPE_MASK) { + case ERR_TYPE_RTAS_LOG: + len = log_rtas_len(buf); + if (!(err_type & ERR_FLAG_BOOT)) + error_log_cnt++; + break; + case ERR_TYPE_KERNEL_PANIC: + default: + spin_unlock_irqrestore(&rtasd_log_lock, s); + return; + } + + /* Write error to NVRAM */ + if (!no_logging && !(err_type & ERR_FLAG_BOOT)) + nvram_write_error_log(buf, len, err_type); + + /* + * rtas errors can occur during boot, and we do want to capture + * those somewhere, even if nvram isn't ready (why not?), and even + * if rtasd isn't ready. Put them into the boot log, at least. + */ + if ((err_type & ERR_TYPE_MASK) == ERR_TYPE_RTAS_LOG) + printk_log_rtas(buf, len); + + /* Check to see if we need to or have stopped logging */ + if (fatal || no_logging) { + no_logging = 1; + spin_unlock_irqrestore(&rtasd_log_lock, s); + return; + } + + /* call type specific method for error */ + switch (err_type & ERR_TYPE_MASK) { + case ERR_TYPE_RTAS_LOG: + offset = rtas_error_log_buffer_max * + ((rtas_log_start+rtas_log_size) & LOG_NUMBER_MASK); + + /* First copy over sequence number */ + memcpy(&rtas_log_buf[offset], (void *) &error_log_cnt, sizeof(int)); + + /* Second copy over error log data */ + offset += sizeof(int); + memcpy(&rtas_log_buf[offset], buf, len); + + if (rtas_log_size < LOG_NUMBER) + rtas_log_size += 1; + else + rtas_log_start += 1; + + spin_unlock_irqrestore(&rtasd_log_lock, s); + wake_up_interruptible(&rtas_log_wait); + break; + case ERR_TYPE_KERNEL_PANIC: + default: + spin_unlock_irqrestore(&rtasd_log_lock, s); + return; + } + +} + + +static int rtas_log_open(struct inode * inode, struct file * file) +{ + return 0; +} + +static int rtas_log_release(struct inode * inode, struct file * file) +{ + return 0; +} + +/* This will check if all events are logged, if they are then, we + * know that we can safely clear the events in NVRAM. + * Next we'll sit and wait for something else to log. + */ +static ssize_t rtas_log_read(struct file * file, char __user * buf, + size_t count, loff_t *ppos) +{ + int error; + char *tmp; + unsigned long s; + unsigned long offset; + + if (!buf || count < rtas_error_log_buffer_max) + return -EINVAL; + + count = rtas_error_log_buffer_max; + + if (!access_ok(VERIFY_WRITE, buf, count)) + return -EFAULT; + + tmp = kmalloc(count, GFP_KERNEL); + if (!tmp) + return -ENOMEM; + + + spin_lock_irqsave(&rtasd_log_lock, s); + /* if it's 0, then we know we got the last one (the one in NVRAM) */ + if (rtas_log_size == 0 && !no_logging) + nvram_clear_error_log(); + spin_unlock_irqrestore(&rtasd_log_lock, s); + + + error = wait_event_interruptible(rtas_log_wait, rtas_log_size); + if (error) + goto out; + + spin_lock_irqsave(&rtasd_log_lock, s); + offset = rtas_error_log_buffer_max * (rtas_log_start & LOG_NUMBER_MASK); + memcpy(tmp, &rtas_log_buf[offset], count); + + rtas_log_start += 1; + rtas_log_size -= 1; + spin_unlock_irqrestore(&rtasd_log_lock, s); + + error = copy_to_user(buf, tmp, count) ? -EFAULT : count; +out: + kfree(tmp); + return error; +} + +static unsigned int rtas_log_poll(struct file *file, poll_table * wait) +{ + poll_wait(file, &rtas_log_wait, wait); + if (rtas_log_size) + return POLLIN | POLLRDNORM; + return 0; +} + +struct file_operations proc_rtas_log_operations = { + .read = rtas_log_read, + .poll = rtas_log_poll, + .open = rtas_log_open, + .release = rtas_log_release, +}; + +static int enable_surveillance(int timeout) +{ + int error; + + error = rtas_set_indicator(SURVEILLANCE_TOKEN, 0, timeout); + + if (error == 0) + return 0; + + if (error == -EINVAL) { + printk(KERN_INFO "rtasd: surveillance not supported\n"); + return 0; + } + + printk(KERN_ERR "rtasd: could not update surveillance\n"); + return -1; +} + +static int get_eventscan_parms(void) +{ + struct device_node *node; + int *ip; + + node = of_find_node_by_path("/rtas"); + + ip = (int *)get_property(node, "rtas-event-scan-rate", NULL); + if (ip == NULL) { + printk(KERN_ERR "rtasd: no rtas-event-scan-rate\n"); + of_node_put(node); + return -1; + } + rtas_event_scan_rate = *ip; + DEBUG("rtas-event-scan-rate %d\n", rtas_event_scan_rate); + + /* Make room for the sequence number */ + rtas_error_log_max = rtas_get_error_log_max(); + rtas_error_log_buffer_max = rtas_error_log_max + sizeof(int); + + of_node_put(node); + + return 0; +} + +static void do_event_scan(int event_scan) +{ + int error; + do { + memset(logdata, 0, rtas_error_log_max); + error = rtas_call(event_scan, 4, 1, NULL, + RTAS_EVENT_SCAN_ALL_EVENTS, 0, + __pa(logdata), rtas_error_log_max); + if (error == -1) { + printk(KERN_ERR "event-scan failed\n"); + break; + } + + if (error == 0) + pSeries_log_error(logdata, ERR_TYPE_RTAS_LOG, 0); + + } while(error == 0); +} + +static void do_event_scan_all_cpus(long delay) +{ + int cpu; + + lock_cpu_hotplug(); + cpu = first_cpu(cpu_online_map); + for (;;) { + set_cpus_allowed(current, cpumask_of_cpu(cpu)); + do_event_scan(rtas_token("event-scan")); + set_cpus_allowed(current, CPU_MASK_ALL); + + /* Drop hotplug lock, and sleep for the specified delay */ + unlock_cpu_hotplug(); + msleep_interruptible(delay); + lock_cpu_hotplug(); + + cpu = next_cpu(cpu, cpu_online_map); + if (cpu == NR_CPUS) + break; + } + unlock_cpu_hotplug(); +} + +static int rtasd(void *unused) +{ + unsigned int err_type; + int event_scan = rtas_token("event-scan"); + int rc; + + daemonize("rtasd"); + + if (event_scan == RTAS_UNKNOWN_SERVICE || get_eventscan_parms() == -1) + goto error; + + rtas_log_buf = vmalloc(rtas_error_log_buffer_max*LOG_NUMBER); + if (!rtas_log_buf) { + printk(KERN_ERR "rtasd: no memory\n"); + goto error; + } + + printk(KERN_INFO "RTAS daemon started\n"); + + DEBUG("will sleep for %d milliseconds\n", (30000/rtas_event_scan_rate)); + + /* See if we have any error stored in NVRAM */ + memset(logdata, 0, rtas_error_log_max); + + rc = nvram_read_error_log(logdata, rtas_error_log_max, &err_type); + + /* We can use rtas_log_buf now */ + no_logging = 0; + + if (!rc) { + if (err_type != ERR_FLAG_ALREADY_LOGGED) { + pSeries_log_error(logdata, err_type | ERR_FLAG_BOOT, 0); + } + } + + /* First pass. */ + do_event_scan_all_cpus(1000); + + if (surveillance_timeout != -1) { + DEBUG("enabling surveillance\n"); + enable_surveillance(surveillance_timeout); + DEBUG("surveillance enabled\n"); + } + + /* Delay should be at least one second since some + * machines have problems if we call event-scan too + * quickly. */ + for (;;) + do_event_scan_all_cpus(30000/rtas_event_scan_rate); + +error: + /* Should delete proc entries */ + return -EINVAL; +} + +static int __init rtas_init(void) +{ + struct proc_dir_entry *entry; + + /* No RTAS, only warn if we are on a pSeries box */ + if (rtas_token("event-scan") == RTAS_UNKNOWN_SERVICE) { + if (systemcfg->platform & PLATFORM_PSERIES) + printk(KERN_INFO "rtasd: no event-scan on system\n"); + return 1; + } + + entry = create_proc_entry("ppc64/rtas/error_log", S_IRUSR, NULL); + if (entry) + entry->proc_fops = &proc_rtas_log_operations; + else + printk(KERN_ERR "Failed to create error_log proc entry\n"); + + if (kernel_thread(rtasd, NULL, CLONE_FS) < 0) + printk(KERN_ERR "Failed to start RTAS daemon\n"); + + return 0; +} + +static int __init surveillance_setup(char *str) +{ + int i; + + if (get_option(&str,&i)) { + if (i >= 0 && i <= 255) + surveillance_timeout = i; + } + + return 1; +} + +static int __init rtasmsgs_setup(char *str) +{ + if (strcmp(str, "on") == 0) + full_rtas_msgs = 1; + else if (strcmp(str, "off") == 0) + full_rtas_msgs = 0; + + return 1; +} +__initcall(rtas_init); +__setup("surveillance=", surveillance_setup); +__setup("rtasmsgs=", rtasmsgs_setup); diff --git a/arch/powerpc/platforms/pseries/setup.c b/arch/powerpc/platforms/pseries/setup.c index c0a3d918148a..f73d69143d35 100644 --- a/arch/powerpc/platforms/pseries/setup.c +++ b/arch/powerpc/platforms/pseries/setup.c @@ -589,9 +589,9 @@ struct machdep_calls __initdata pSeries_md = { .pcibios_fixup = pSeries_final_fixup, .pci_probe_mode = pSeries_pci_probe_mode, .irq_bus_setup = pSeries_irq_bus_setup, - .restart = rtas_fw_restart, - .power_off = rtas_fw_power_off, - .halt = rtas_fw_halt, + .restart = rtas_restart, + .power_off = rtas_power_off, + .halt = rtas_halt, .panic = rtas_os_term, .cpu_die = pSeries_mach_cpu_die, .get_boot_time = rtas_get_boot_time, diff --git a/arch/ppc64/kernel/Makefile b/arch/ppc64/kernel/Makefile index 990df0905c87..74892ad032ba 100644 --- a/arch/ppc64/kernel/Makefile +++ b/arch/ppc64/kernel/Makefile @@ -29,19 +29,17 @@ ifneq ($(CONFIG_PPC_MERGE),y) obj-$(CONFIG_PPC_MULTIPLATFORM) += prom_init.o endif -obj-$(CONFIG_PPC_PSERIES) += rtasd.o udbg_16550.o +obj-$(CONFIG_PPC_PSERIES) += udbg_16550.o obj-$(CONFIG_KEXEC) += machine_kexec.o obj-$(CONFIG_EEH) += eeh.o obj-$(CONFIG_PROC_FS) += proc_ppc64.o -obj-$(CONFIG_RTAS_FLASH) += rtas_flash.o obj-$(CONFIG_SMP) += smp.o obj-$(CONFIG_MODULES) += module.o ifneq ($(CONFIG_PPC_MERGE),y) obj-$(CONFIG_MODULES) += ppc_ksyms.o endif obj-$(CONFIG_PPC_RTAS) += rtas_pci.o -obj-$(CONFIG_RTAS_PROC) += rtas-proc.o obj-$(CONFIG_SCANLOG) += scanlog.o obj-$(CONFIG_LPARCFG) += lparcfg.o obj-$(CONFIG_HVC_CONSOLE) += hvconsole.o diff --git a/arch/ppc64/kernel/rtas-proc.c b/arch/ppc64/kernel/rtas-proc.c deleted file mode 100644 index 5bdd5b079d96..000000000000 --- a/arch/ppc64/kernel/rtas-proc.c +++ /dev/null @@ -1,808 +0,0 @@ -/* - * arch/ppc64/kernel/rtas-proc.c - * Copyright (C) 2000 Tilmann Bitterberg - * (tilmann@bitterberg.de) - * - * RTAS (Runtime Abstraction Services) stuff - * Intention is to provide a clean user interface - * to use the RTAS. - * - * TODO: - * Split off a header file and maybe move it to a different - * location. Write Documentation on what the /proc/rtas/ entries - * actually do. - */ - -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include - -#include -#include -#include -#include -#include -#include /* for ppc_md */ -#include -#include - -/* Token for Sensors */ -#define KEY_SWITCH 0x0001 -#define ENCLOSURE_SWITCH 0x0002 -#define THERMAL_SENSOR 0x0003 -#define LID_STATUS 0x0004 -#define POWER_SOURCE 0x0005 -#define BATTERY_VOLTAGE 0x0006 -#define BATTERY_REMAINING 0x0007 -#define BATTERY_PERCENTAGE 0x0008 -#define EPOW_SENSOR 0x0009 -#define BATTERY_CYCLESTATE 0x000a -#define BATTERY_CHARGING 0x000b - -/* IBM specific sensors */ -#define IBM_SURVEILLANCE 0x2328 /* 9000 */ -#define IBM_FANRPM 0x2329 /* 9001 */ -#define IBM_VOLTAGE 0x232a /* 9002 */ -#define IBM_DRCONNECTOR 0x232b /* 9003 */ -#define IBM_POWERSUPPLY 0x232c /* 9004 */ - -/* Status return values */ -#define SENSOR_CRITICAL_HIGH 13 -#define SENSOR_WARNING_HIGH 12 -#define SENSOR_NORMAL 11 -#define SENSOR_WARNING_LOW 10 -#define SENSOR_CRITICAL_LOW 9 -#define SENSOR_SUCCESS 0 -#define SENSOR_HW_ERROR -1 -#define SENSOR_BUSY -2 -#define SENSOR_NOT_EXIST -3 -#define SENSOR_DR_ENTITY -9000 - -/* Location Codes */ -#define LOC_SCSI_DEV_ADDR 'A' -#define LOC_SCSI_DEV_LOC 'B' -#define LOC_CPU 'C' -#define LOC_DISKETTE 'D' -#define LOC_ETHERNET 'E' -#define LOC_FAN 'F' -#define LOC_GRAPHICS 'G' -/* reserved / not used 'H' */ -#define LOC_IO_ADAPTER 'I' -/* reserved / not used 'J' */ -#define LOC_KEYBOARD 'K' -#define LOC_LCD 'L' -#define LOC_MEMORY 'M' -#define LOC_NV_MEMORY 'N' -#define LOC_MOUSE 'O' -#define LOC_PLANAR 'P' -#define LOC_OTHER_IO 'Q' -#define LOC_PARALLEL 'R' -#define LOC_SERIAL 'S' -#define LOC_DEAD_RING 'T' -#define LOC_RACKMOUNTED 'U' /* for _u_nit is rack mounted */ -#define LOC_VOLTAGE 'V' -#define LOC_SWITCH_ADAPTER 'W' -#define LOC_OTHER 'X' -#define LOC_FIRMWARE 'Y' -#define LOC_SCSI 'Z' - -/* Tokens for indicators */ -#define TONE_FREQUENCY 0x0001 /* 0 - 1000 (HZ)*/ -#define TONE_VOLUME 0x0002 /* 0 - 100 (%) */ -#define SYSTEM_POWER_STATE 0x0003 -#define WARNING_LIGHT 0x0004 -#define DISK_ACTIVITY_LIGHT 0x0005 -#define HEX_DISPLAY_UNIT 0x0006 -#define BATTERY_WARNING_TIME 0x0007 -#define CONDITION_CYCLE_REQUEST 0x0008 -#define SURVEILLANCE_INDICATOR 0x2328 /* 9000 */ -#define DR_ACTION 0x2329 /* 9001 */ -#define DR_INDICATOR 0x232a /* 9002 */ -/* 9003 - 9004: Vendor specific */ -/* 9006 - 9999: Vendor specific */ - -/* other */ -#define MAX_SENSORS 17 /* I only know of 17 sensors */ -#define MAX_LINELENGTH 256 -#define SENSOR_PREFIX "ibm,sensor-" -#define cel_to_fahr(x) ((x*9/5)+32) - - -/* Globals */ -static struct rtas_sensors sensors; -static struct device_node *rtas_node = NULL; -static unsigned long power_on_time = 0; /* Save the time the user set */ -static char progress_led[MAX_LINELENGTH]; - -static unsigned long rtas_tone_frequency = 1000; -static unsigned long rtas_tone_volume = 0; - -/* ****************STRUCTS******************************************* */ -struct individual_sensor { - unsigned int token; - unsigned int quant; -}; - -struct rtas_sensors { - struct individual_sensor sensor[MAX_SENSORS]; - unsigned int quant; -}; - -/* ****************************************************************** */ -/* Declarations */ -static int ppc_rtas_sensors_show(struct seq_file *m, void *v); -static int ppc_rtas_clock_show(struct seq_file *m, void *v); -static ssize_t ppc_rtas_clock_write(struct file *file, - const char __user *buf, size_t count, loff_t *ppos); -static int ppc_rtas_progress_show(struct seq_file *m, void *v); -static ssize_t ppc_rtas_progress_write(struct file *file, - const char __user *buf, size_t count, loff_t *ppos); -static int ppc_rtas_poweron_show(struct seq_file *m, void *v); -static ssize_t ppc_rtas_poweron_write(struct file *file, - const char __user *buf, size_t count, loff_t *ppos); - -static ssize_t ppc_rtas_tone_freq_write(struct file *file, - const char __user *buf, size_t count, loff_t *ppos); -static int ppc_rtas_tone_freq_show(struct seq_file *m, void *v); -static ssize_t ppc_rtas_tone_volume_write(struct file *file, - const char __user *buf, size_t count, loff_t *ppos); -static int ppc_rtas_tone_volume_show(struct seq_file *m, void *v); -static int ppc_rtas_rmo_buf_show(struct seq_file *m, void *v); - -static int sensors_open(struct inode *inode, struct file *file) -{ - return single_open(file, ppc_rtas_sensors_show, NULL); -} - -struct file_operations ppc_rtas_sensors_operations = { - .open = sensors_open, - .read = seq_read, - .llseek = seq_lseek, - .release = single_release, -}; - -static int poweron_open(struct inode *inode, struct file *file) -{ - return single_open(file, ppc_rtas_poweron_show, NULL); -} - -struct file_operations ppc_rtas_poweron_operations = { - .open = poweron_open, - .read = seq_read, - .llseek = seq_lseek, - .write = ppc_rtas_poweron_write, - .release = single_release, -}; - -static int progress_open(struct inode *inode, struct file *file) -{ - return single_open(file, ppc_rtas_progress_show, NULL); -} - -struct file_operations ppc_rtas_progress_operations = { - .open = progress_open, - .read = seq_read, - .llseek = seq_lseek, - .write = ppc_rtas_progress_write, - .release = single_release, -}; - -static int clock_open(struct inode *inode, struct file *file) -{ - return single_open(file, ppc_rtas_clock_show, NULL); -} - -struct file_operations ppc_rtas_clock_operations = { - .open = clock_open, - .read = seq_read, - .llseek = seq_lseek, - .write = ppc_rtas_clock_write, - .release = single_release, -}; - -static int tone_freq_open(struct inode *inode, struct file *file) -{ - return single_open(file, ppc_rtas_tone_freq_show, NULL); -} - -struct file_operations ppc_rtas_tone_freq_operations = { - .open = tone_freq_open, - .read = seq_read, - .llseek = seq_lseek, - .write = ppc_rtas_tone_freq_write, - .release = single_release, -}; - -static int tone_volume_open(struct inode *inode, struct file *file) -{ - return single_open(file, ppc_rtas_tone_volume_show, NULL); -} - -struct file_operations ppc_rtas_tone_volume_operations = { - .open = tone_volume_open, - .read = seq_read, - .llseek = seq_lseek, - .write = ppc_rtas_tone_volume_write, - .release = single_release, -}; - -static int rmo_buf_open(struct inode *inode, struct file *file) -{ - return single_open(file, ppc_rtas_rmo_buf_show, NULL); -} - -struct file_operations ppc_rtas_rmo_buf_ops = { - .open = rmo_buf_open, - .read = seq_read, - .llseek = seq_lseek, - .release = single_release, -}; - -static int ppc_rtas_find_all_sensors(void); -static void ppc_rtas_process_sensor(struct seq_file *m, - struct individual_sensor *s, int state, int error, char *loc); -static char *ppc_rtas_process_error(int error); -static void get_location_code(struct seq_file *m, - struct individual_sensor *s, char *loc); -static void check_location_string(struct seq_file *m, char *c); -static void check_location(struct seq_file *m, char *c); - -static int __init proc_rtas_init(void) -{ - struct proc_dir_entry *entry; - - if (!(systemcfg->platform & PLATFORM_PSERIES)) - return 1; - - rtas_node = of_find_node_by_name(NULL, "rtas"); - if (rtas_node == NULL) - return 1; - - entry = create_proc_entry("ppc64/rtas/progress", S_IRUGO|S_IWUSR, NULL); - if (entry) - entry->proc_fops = &ppc_rtas_progress_operations; - - entry = create_proc_entry("ppc64/rtas/clock", S_IRUGO|S_IWUSR, NULL); - if (entry) - entry->proc_fops = &ppc_rtas_clock_operations; - - entry = create_proc_entry("ppc64/rtas/poweron", S_IWUSR|S_IRUGO, NULL); - if (entry) - entry->proc_fops = &ppc_rtas_poweron_operations; - - entry = create_proc_entry("ppc64/rtas/sensors", S_IRUGO, NULL); - if (entry) - entry->proc_fops = &ppc_rtas_sensors_operations; - - entry = create_proc_entry("ppc64/rtas/frequency", S_IWUSR|S_IRUGO, - NULL); - if (entry) - entry->proc_fops = &ppc_rtas_tone_freq_operations; - - entry = create_proc_entry("ppc64/rtas/volume", S_IWUSR|S_IRUGO, NULL); - if (entry) - entry->proc_fops = &ppc_rtas_tone_volume_operations; - - entry = create_proc_entry("ppc64/rtas/rmo_buffer", S_IRUSR, NULL); - if (entry) - entry->proc_fops = &ppc_rtas_rmo_buf_ops; - - return 0; -} - -__initcall(proc_rtas_init); - -static int parse_number(const char __user *p, size_t count, unsigned long *val) -{ - char buf[40]; - char *end; - - if (count > 39) - return -EINVAL; - - if (copy_from_user(buf, p, count)) - return -EFAULT; - - buf[count] = 0; - - *val = simple_strtoul(buf, &end, 10); - if (*end && *end != '\n') - return -EINVAL; - - return 0; -} - -/* ****************************************************************** */ -/* POWER-ON-TIME */ -/* ****************************************************************** */ -static ssize_t ppc_rtas_poweron_write(struct file *file, - const char __user *buf, size_t count, loff_t *ppos) -{ - struct rtc_time tm; - unsigned long nowtime; - int error = parse_number(buf, count, &nowtime); - if (error) - return error; - - power_on_time = nowtime; /* save the time */ - - to_tm(nowtime, &tm); - - error = rtas_call(rtas_token("set-time-for-power-on"), 7, 1, NULL, - tm.tm_year, tm.tm_mon, tm.tm_mday, - tm.tm_hour, tm.tm_min, tm.tm_sec, 0 /* nano */); - if (error) - printk(KERN_WARNING "error: setting poweron time returned: %s\n", - ppc_rtas_process_error(error)); - return count; -} -/* ****************************************************************** */ -static int ppc_rtas_poweron_show(struct seq_file *m, void *v) -{ - if (power_on_time == 0) - seq_printf(m, "Power on time not set\n"); - else - seq_printf(m, "%lu\n",power_on_time); - return 0; -} - -/* ****************************************************************** */ -/* PROGRESS */ -/* ****************************************************************** */ -static ssize_t ppc_rtas_progress_write(struct file *file, - const char __user *buf, size_t count, loff_t *ppos) -{ - unsigned long hex; - - if (count >= MAX_LINELENGTH) - count = MAX_LINELENGTH -1; - if (copy_from_user(progress_led, buf, count)) { /* save the string */ - return -EFAULT; - } - progress_led[count] = 0; - - /* Lets see if the user passed hexdigits */ - hex = simple_strtoul(progress_led, NULL, 10); - - rtas_progress ((char *)progress_led, hex); - return count; - - /* clear the line */ - /* rtas_progress(" ", 0xffff);*/ -} -/* ****************************************************************** */ -static int ppc_rtas_progress_show(struct seq_file *m, void *v) -{ - if (progress_led) - seq_printf(m, "%s\n", progress_led); - return 0; -} - -/* ****************************************************************** */ -/* CLOCK */ -/* ****************************************************************** */ -static ssize_t ppc_rtas_clock_write(struct file *file, - const char __user *buf, size_t count, loff_t *ppos) -{ - struct rtc_time tm; - unsigned long nowtime; - int error = parse_number(buf, count, &nowtime); - if (error) - return error; - - to_tm(nowtime, &tm); - error = rtas_call(rtas_token("set-time-of-day"), 7, 1, NULL, - tm.tm_year, tm.tm_mon, tm.tm_mday, - tm.tm_hour, tm.tm_min, tm.tm_sec, 0); - if (error) - printk(KERN_WARNING "error: setting the clock returned: %s\n", - ppc_rtas_process_error(error)); - return count; -} -/* ****************************************************************** */ -static int ppc_rtas_clock_show(struct seq_file *m, void *v) -{ - int ret[8]; - int error = rtas_call(rtas_token("get-time-of-day"), 0, 8, ret); - - if (error) { - printk(KERN_WARNING "error: reading the clock returned: %s\n", - ppc_rtas_process_error(error)); - seq_printf(m, "0"); - } else { - unsigned int year, mon, day, hour, min, sec; - year = ret[0]; mon = ret[1]; day = ret[2]; - hour = ret[3]; min = ret[4]; sec = ret[5]; - seq_printf(m, "%lu\n", - mktime(year, mon, day, hour, min, sec)); - } - return 0; -} - -/* ****************************************************************** */ -/* SENSOR STUFF */ -/* ****************************************************************** */ -static int ppc_rtas_sensors_show(struct seq_file *m, void *v) -{ - int i,j; - int state, error; - int get_sensor_state = rtas_token("get-sensor-state"); - - seq_printf(m, "RTAS (RunTime Abstraction Services) Sensor Information\n"); - seq_printf(m, "Sensor\t\tValue\t\tCondition\tLocation\n"); - seq_printf(m, "********************************************************\n"); - - if (ppc_rtas_find_all_sensors() != 0) { - seq_printf(m, "\nNo sensors are available\n"); - return 0; - } - - for (i=0; itoken); - loc = (char *) get_property(rtas_node, rstr, &llen); - - /* A sensor may have multiple instances */ - for (j = 0, offs = 0; j <= p->quant; j++) { - error = rtas_call(get_sensor_state, 2, 2, &state, - p->token, j); - - ppc_rtas_process_sensor(m, p, state, error, loc); - seq_putc(m, '\n'); - if (loc) { - offs += strlen(loc) + 1; - loc += strlen(loc) + 1; - if (offs >= llen) - loc = NULL; - } - } - } - return 0; -} - -/* ****************************************************************** */ - -static int ppc_rtas_find_all_sensors(void) -{ - unsigned int *utmp; - int len, i; - - utmp = (unsigned int *) get_property(rtas_node, "rtas-sensors", &len); - if (utmp == NULL) { - printk (KERN_ERR "error: could not get rtas-sensors\n"); - return 1; - } - - sensors.quant = len / 8; /* int + int */ - - for (i=0; itoken) { - case KEY_SWITCH: - seq_printf(m, "Key switch:\t"); - num_states = sizeof(key_switch) / sizeof(char *); - if (state < num_states) { - seq_printf(m, "%s\t", key_switch[state]); - have_strings = 1; - } - break; - case ENCLOSURE_SWITCH: - seq_printf(m, "Enclosure switch:\t"); - num_states = sizeof(enclosure_switch) / sizeof(char *); - if (state < num_states) { - seq_printf(m, "%s\t", - enclosure_switch[state]); - have_strings = 1; - } - break; - case THERMAL_SENSOR: - seq_printf(m, "Temp. (C/F):\t"); - temperature = 1; - break; - case LID_STATUS: - seq_printf(m, "Lid status:\t"); - num_states = sizeof(lid_status) / sizeof(char *); - if (state < num_states) { - seq_printf(m, "%s\t", lid_status[state]); - have_strings = 1; - } - break; - case POWER_SOURCE: - seq_printf(m, "Power source:\t"); - num_states = sizeof(power_source) / sizeof(char *); - if (state < num_states) { - seq_printf(m, "%s\t", - power_source[state]); - have_strings = 1; - } - break; - case BATTERY_VOLTAGE: - seq_printf(m, "Battery voltage:\t"); - break; - case BATTERY_REMAINING: - seq_printf(m, "Battery remaining:\t"); - num_states = sizeof(battery_remaining) / sizeof(char *); - if (state < num_states) - { - seq_printf(m, "%s\t", - battery_remaining[state]); - have_strings = 1; - } - break; - case BATTERY_PERCENTAGE: - seq_printf(m, "Battery percentage:\t"); - break; - case EPOW_SENSOR: - seq_printf(m, "EPOW Sensor:\t"); - num_states = sizeof(epow_sensor) / sizeof(char *); - if (state < num_states) { - seq_printf(m, "%s\t", epow_sensor[state]); - have_strings = 1; - } - break; - case BATTERY_CYCLESTATE: - seq_printf(m, "Battery cyclestate:\t"); - num_states = sizeof(battery_cyclestate) / - sizeof(char *); - if (state < num_states) { - seq_printf(m, "%s\t", - battery_cyclestate[state]); - have_strings = 1; - } - break; - case BATTERY_CHARGING: - seq_printf(m, "Battery Charging:\t"); - num_states = sizeof(battery_charging) / sizeof(char *); - if (state < num_states) { - seq_printf(m, "%s\t", - battery_charging[state]); - have_strings = 1; - } - break; - case IBM_SURVEILLANCE: - seq_printf(m, "Surveillance:\t"); - break; - case IBM_FANRPM: - seq_printf(m, "Fan (rpm):\t"); - break; - case IBM_VOLTAGE: - seq_printf(m, "Voltage (mv):\t"); - break; - case IBM_DRCONNECTOR: - seq_printf(m, "DR connector:\t"); - num_states = sizeof(ibm_drconnector) / sizeof(char *); - if (state < num_states) { - seq_printf(m, "%s\t", - ibm_drconnector[state]); - have_strings = 1; - } - break; - case IBM_POWERSUPPLY: - seq_printf(m, "Powersupply:\t"); - break; - default: - seq_printf(m, "Unknown sensor (type %d), ignoring it\n", - s->token); - unknown = 1; - have_strings = 1; - break; - } - if (have_strings == 0) { - if (temperature) { - seq_printf(m, "%4d /%4d\t", state, cel_to_fahr(state)); - } else - seq_printf(m, "%10d\t", state); - } - if (unknown == 0) { - seq_printf(m, "%s\t", ppc_rtas_process_error(error)); - get_location_code(m, s, loc); - } -} - -/* ****************************************************************** */ - -static void check_location(struct seq_file *m, char *c) -{ - switch (c[0]) { - case LOC_PLANAR: - seq_printf(m, "Planar #%c", c[1]); - break; - case LOC_CPU: - seq_printf(m, "CPU #%c", c[1]); - break; - case LOC_FAN: - seq_printf(m, "Fan #%c", c[1]); - break; - case LOC_RACKMOUNTED: - seq_printf(m, "Rack #%c", c[1]); - break; - case LOC_VOLTAGE: - seq_printf(m, "Voltage #%c", c[1]); - break; - case LOC_LCD: - seq_printf(m, "LCD #%c", c[1]); - break; - case '.': - seq_printf(m, "- %c", c[1]); - break; - default: - seq_printf(m, "Unknown location"); - break; - } -} - - -/* ****************************************************************** */ -/* - * Format: - * ${LETTER}${NUMBER}[[-/]${LETTER}${NUMBER} [ ... ] ] - * the '.' may be an abbrevation - */ -static void check_location_string(struct seq_file *m, char *c) -{ - while (*c) { - if (isalpha(*c) || *c == '.') - check_location(m, c); - else if (*c == '/' || *c == '-') - seq_printf(m, " at "); - c++; - } -} - - -/* ****************************************************************** */ - -static void get_location_code(struct seq_file *m, struct individual_sensor *s, char *loc) -{ - if (!loc || !*loc) { - seq_printf(m, "---");/* does not have a location */ - } else { - check_location_string(m, loc); - } - seq_putc(m, ' '); -} -/* ****************************************************************** */ -/* INDICATORS - Tone Frequency */ -/* ****************************************************************** */ -static ssize_t ppc_rtas_tone_freq_write(struct file *file, - const char __user *buf, size_t count, loff_t *ppos) -{ - unsigned long freq; - int error = parse_number(buf, count, &freq); - if (error) - return error; - - rtas_tone_frequency = freq; /* save it for later */ - error = rtas_call(rtas_token("set-indicator"), 3, 1, NULL, - TONE_FREQUENCY, 0, freq); - if (error) - printk(KERN_WARNING "error: setting tone frequency returned: %s\n", - ppc_rtas_process_error(error)); - return count; -} -/* ****************************************************************** */ -static int ppc_rtas_tone_freq_show(struct seq_file *m, void *v) -{ - seq_printf(m, "%lu\n", rtas_tone_frequency); - return 0; -} -/* ****************************************************************** */ -/* INDICATORS - Tone Volume */ -/* ****************************************************************** */ -static ssize_t ppc_rtas_tone_volume_write(struct file *file, - const char __user *buf, size_t count, loff_t *ppos) -{ - unsigned long volume; - int error = parse_number(buf, count, &volume); - if (error) - return error; - - if (volume > 100) - volume = 100; - - rtas_tone_volume = volume; /* save it for later */ - error = rtas_call(rtas_token("set-indicator"), 3, 1, NULL, - TONE_VOLUME, 0, volume); - if (error) - printk(KERN_WARNING "error: setting tone volume returned: %s\n", - ppc_rtas_process_error(error)); - return count; -} -/* ****************************************************************** */ -static int ppc_rtas_tone_volume_show(struct seq_file *m, void *v) -{ - seq_printf(m, "%lu\n", rtas_tone_volume); - return 0; -} - -#define RMO_READ_BUF_MAX 30 - -/* RTAS Userspace access */ -static int ppc_rtas_rmo_buf_show(struct seq_file *m, void *v) -{ - seq_printf(m, "%016lx %x\n", rtas_rmo_buf, RTAS_RMOBUF_MAX); - return 0; -} diff --git a/arch/ppc64/kernel/rtas_flash.c b/arch/ppc64/kernel/rtas_flash.c deleted file mode 100644 index 923e2e201a70..000000000000 --- a/arch/ppc64/kernel/rtas_flash.c +++ /dev/null @@ -1,725 +0,0 @@ -/* - * c 2001 PPC 64 Team, IBM Corp - * - * This program is free software; you can redistribute it and/or - * modify it under the terms of the GNU General Public License - * as published by the Free Software Foundation; either version - * 2 of the License, or (at your option) any later version. - * - * /proc/ppc64/rtas/firmware_flash interface - * - * This file implements a firmware_flash interface to pump a firmware - * image into the kernel. At reboot time rtas_restart() will see the - * firmware image and flash it as it reboots (see rtas.c). - */ - -#include -#include -#include -#include -#include -#include - -#define MODULE_VERS "1.0" -#define MODULE_NAME "rtas_flash" - -#define FIRMWARE_FLASH_NAME "firmware_flash" -#define FIRMWARE_UPDATE_NAME "firmware_update" -#define MANAGE_FLASH_NAME "manage_flash" -#define VALIDATE_FLASH_NAME "validate_flash" - -/* General RTAS Status Codes */ -#define RTAS_RC_SUCCESS 0 -#define RTAS_RC_HW_ERR -1 -#define RTAS_RC_BUSY -2 - -/* Flash image status values */ -#define FLASH_AUTH -9002 /* RTAS Not Service Authority Partition */ -#define FLASH_NO_OP -1099 /* No operation initiated by user */ -#define FLASH_IMG_SHORT -1005 /* Flash image shorter than expected */ -#define FLASH_IMG_BAD_LEN -1004 /* Bad length value in flash list block */ -#define FLASH_IMG_NULL_DATA -1003 /* Bad data value in flash list block */ -#define FLASH_IMG_READY 0 /* Firmware img ready for flash on reboot */ - -/* Manage image status values */ -#define MANAGE_AUTH -9002 /* RTAS Not Service Authority Partition */ -#define MANAGE_ACTIVE_ERR -9001 /* RTAS Cannot Overwrite Active Img */ -#define MANAGE_NO_OP -1099 /* No operation initiated by user */ -#define MANAGE_PARAM_ERR -3 /* RTAS Parameter Error */ -#define MANAGE_HW_ERR -1 /* RTAS Hardware Error */ - -/* Validate image status values */ -#define VALIDATE_AUTH -9002 /* RTAS Not Service Authority Partition */ -#define VALIDATE_NO_OP -1099 /* No operation initiated by the user */ -#define VALIDATE_INCOMPLETE -1002 /* User copied < VALIDATE_BUF_SIZE */ -#define VALIDATE_READY -1001 /* Firmware image ready for validation */ -#define VALIDATE_PARAM_ERR -3 /* RTAS Parameter Error */ -#define VALIDATE_HW_ERR -1 /* RTAS Hardware Error */ -#define VALIDATE_TMP_UPDATE 0 /* Validate Return Status */ -#define VALIDATE_FLASH_AUTH 1 /* Validate Return Status */ -#define VALIDATE_INVALID_IMG 2 /* Validate Return Status */ -#define VALIDATE_CUR_UNKNOWN 3 /* Validate Return Status */ -#define VALIDATE_TMP_COMMIT_DL 4 /* Validate Return Status */ -#define VALIDATE_TMP_COMMIT 5 /* Validate Return Status */ -#define VALIDATE_TMP_UPDATE_DL 6 /* Validate Return Status */ - -/* ibm,manage-flash-image operation tokens */ -#define RTAS_REJECT_TMP_IMG 0 -#define RTAS_COMMIT_TMP_IMG 1 - -/* Array sizes */ -#define VALIDATE_BUF_SIZE 4096 -#define RTAS_MSG_MAXLEN 64 - -/* Local copy of the flash block list. - * We only allow one open of the flash proc file and create this - * list as we go. This list will be put in the kernel's - * rtas_firmware_flash_list global var once it is fully read. - * - * For convenience as we build the list we use virtual addrs, - * we do not fill in the version number, and the length field - * is treated as the number of entries currently in the block - * (i.e. not a byte count). This is all fixed on release. - */ - -/* Status int must be first member of struct */ -struct rtas_update_flash_t -{ - int status; /* Flash update status */ - struct flash_block_list *flist; /* Local copy of flash block list */ -}; - -/* Status int must be first member of struct */ -struct rtas_manage_flash_t -{ - int status; /* Returned status */ - unsigned int op; /* Reject or commit image */ -}; - -/* Status int must be first member of struct */ -struct rtas_validate_flash_t -{ - int status; /* Returned status */ - char buf[VALIDATE_BUF_SIZE]; /* Candidate image buffer */ - unsigned int buf_size; /* Size of image buf */ - unsigned int update_results; /* Update results token */ -}; - -static DEFINE_SPINLOCK(flash_file_open_lock); -static struct proc_dir_entry *firmware_flash_pde; -static struct proc_dir_entry *firmware_update_pde; -static struct proc_dir_entry *validate_pde; -static struct proc_dir_entry *manage_pde; - -/* Do simple sanity checks on the flash image. */ -static int flash_list_valid(struct flash_block_list *flist) -{ - struct flash_block_list *f; - int i; - unsigned long block_size, image_size; - - /* Paranoid self test here. We also collect the image size. */ - image_size = 0; - for (f = flist; f; f = f->next) { - for (i = 0; i < f->num_blocks; i++) { - if (f->blocks[i].data == NULL) { - return FLASH_IMG_NULL_DATA; - } - block_size = f->blocks[i].length; - if (block_size <= 0 || block_size > PAGE_SIZE) { - return FLASH_IMG_BAD_LEN; - } - image_size += block_size; - } - } - - if (image_size < (256 << 10)) { - if (image_size < 2) - return FLASH_NO_OP; - } - - printk(KERN_INFO "FLASH: flash image with %ld bytes stored for hardware flash on reboot\n", image_size); - - return FLASH_IMG_READY; -} - -static void free_flash_list(struct flash_block_list *f) -{ - struct flash_block_list *next; - int i; - - while (f) { - for (i = 0; i < f->num_blocks; i++) - free_page((unsigned long)(f->blocks[i].data)); - next = f->next; - free_page((unsigned long)f); - f = next; - } -} - -static int rtas_flash_release(struct inode *inode, struct file *file) -{ - struct proc_dir_entry *dp = PDE(file->f_dentry->d_inode); - struct rtas_update_flash_t *uf; - - uf = (struct rtas_update_flash_t *) dp->data; - if (uf->flist) { - /* File was opened in write mode for a new flash attempt */ - /* Clear saved list */ - if (rtas_firmware_flash_list.next) { - free_flash_list(rtas_firmware_flash_list.next); - rtas_firmware_flash_list.next = NULL; - } - - if (uf->status != FLASH_AUTH) - uf->status = flash_list_valid(uf->flist); - - if (uf->status == FLASH_IMG_READY) - rtas_firmware_flash_list.next = uf->flist; - else - free_flash_list(uf->flist); - - uf->flist = NULL; - } - - atomic_dec(&dp->count); - return 0; -} - -static void get_flash_status_msg(int status, char *buf) -{ - char *msg; - - switch (status) { - case FLASH_AUTH: - msg = "error: this partition does not have service authority\n"; - break; - case FLASH_NO_OP: - msg = "info: no firmware image for flash\n"; - break; - case FLASH_IMG_SHORT: - msg = "error: flash image short\n"; - break; - case FLASH_IMG_BAD_LEN: - msg = "error: internal error bad length\n"; - break; - case FLASH_IMG_NULL_DATA: - msg = "error: internal error null data\n"; - break; - case FLASH_IMG_READY: - msg = "ready: firmware image ready for flash on reboot\n"; - break; - default: - sprintf(buf, "error: unexpected status value %d\n", status); - return; - } - - strcpy(buf, msg); -} - -/* Reading the proc file will show status (not the firmware contents) */ -static ssize_t rtas_flash_read(struct file *file, char __user *buf, - size_t count, loff_t *ppos) -{ - struct proc_dir_entry *dp = PDE(file->f_dentry->d_inode); - struct rtas_update_flash_t *uf; - char msg[RTAS_MSG_MAXLEN]; - int msglen; - - uf = (struct rtas_update_flash_t *) dp->data; - - if (!strcmp(dp->name, FIRMWARE_FLASH_NAME)) { - get_flash_status_msg(uf->status, msg); - } else { /* FIRMWARE_UPDATE_NAME */ - sprintf(msg, "%d\n", uf->status); - } - msglen = strlen(msg); - if (msglen > count) - msglen = count; - - if (ppos && *ppos != 0) - return 0; /* be cheap */ - - if (!access_ok(VERIFY_WRITE, buf, msglen)) - return -EINVAL; - - if (copy_to_user(buf, msg, msglen)) - return -EFAULT; - - if (ppos) - *ppos = msglen; - return msglen; -} - -/* We could be much more efficient here. But to keep this function - * simple we allocate a page to the block list no matter how small the - * count is. If the system is low on memory it will be just as well - * that we fail.... - */ -static ssize_t rtas_flash_write(struct file *file, const char __user *buffer, - size_t count, loff_t *off) -{ - struct proc_dir_entry *dp = PDE(file->f_dentry->d_inode); - struct rtas_update_flash_t *uf; - char *p; - int next_free; - struct flash_block_list *fl; - - uf = (struct rtas_update_flash_t *) dp->data; - - if (uf->status == FLASH_AUTH || count == 0) - return count; /* discard data */ - - /* In the case that the image is not ready for flashing, the memory - * allocated for the block list will be freed upon the release of the - * proc file - */ - if (uf->flist == NULL) { - uf->flist = (struct flash_block_list *) get_zeroed_page(GFP_KERNEL); - if (!uf->flist) - return -ENOMEM; - } - - fl = uf->flist; - while (fl->next) - fl = fl->next; /* seek to last block_list for append */ - next_free = fl->num_blocks; - if (next_free == FLASH_BLOCKS_PER_NODE) { - /* Need to allocate another block_list */ - fl->next = (struct flash_block_list *)get_zeroed_page(GFP_KERNEL); - if (!fl->next) - return -ENOMEM; - fl = fl->next; - next_free = 0; - } - - if (count > PAGE_SIZE) - count = PAGE_SIZE; - p = (char *)get_zeroed_page(GFP_KERNEL); - if (!p) - return -ENOMEM; - - if(copy_from_user(p, buffer, count)) { - free_page((unsigned long)p); - return -EFAULT; - } - fl->blocks[next_free].data = p; - fl->blocks[next_free].length = count; - fl->num_blocks++; - - return count; -} - -static int rtas_excl_open(struct inode *inode, struct file *file) -{ - struct proc_dir_entry *dp = PDE(inode); - - /* Enforce exclusive open with use count of PDE */ - spin_lock(&flash_file_open_lock); - if (atomic_read(&dp->count) > 1) { - spin_unlock(&flash_file_open_lock); - return -EBUSY; - } - - atomic_inc(&dp->count); - spin_unlock(&flash_file_open_lock); - - return 0; -} - -static int rtas_excl_release(struct inode *inode, struct file *file) -{ - struct proc_dir_entry *dp = PDE(inode); - - atomic_dec(&dp->count); - - return 0; -} - -static void manage_flash(struct rtas_manage_flash_t *args_buf) -{ - unsigned int wait_time; - s32 rc; - - while (1) { - rc = rtas_call(rtas_token("ibm,manage-flash-image"), 1, - 1, NULL, args_buf->op); - if (rc == RTAS_RC_BUSY) - udelay(1); - else if (rtas_is_extended_busy(rc)) { - wait_time = rtas_extended_busy_delay_time(rc); - udelay(wait_time * 1000); - } else - break; - } - - args_buf->status = rc; -} - -static ssize_t manage_flash_read(struct file *file, char __user *buf, - size_t count, loff_t *ppos) -{ - struct proc_dir_entry *dp = PDE(file->f_dentry->d_inode); - struct rtas_manage_flash_t *args_buf; - char msg[RTAS_MSG_MAXLEN]; - int msglen; - - args_buf = (struct rtas_manage_flash_t *) dp->data; - if (args_buf == NULL) - return 0; - - msglen = sprintf(msg, "%d\n", args_buf->status); - if (msglen > count) - msglen = count; - - if (ppos && *ppos != 0) - return 0; /* be cheap */ - - if (!access_ok(VERIFY_WRITE, buf, msglen)) - return -EINVAL; - - if (copy_to_user(buf, msg, msglen)) - return -EFAULT; - - if (ppos) - *ppos = msglen; - return msglen; -} - -static ssize_t manage_flash_write(struct file *file, const char __user *buf, - size_t count, loff_t *off) -{ - struct proc_dir_entry *dp = PDE(file->f_dentry->d_inode); - struct rtas_manage_flash_t *args_buf; - const char reject_str[] = "0"; - const char commit_str[] = "1"; - char stkbuf[10]; - int op; - - args_buf = (struct rtas_manage_flash_t *) dp->data; - if ((args_buf->status == MANAGE_AUTH) || (count == 0)) - return count; - - op = -1; - if (buf) { - if (count > 9) count = 9; - if (copy_from_user (stkbuf, buf, count)) { - return -EFAULT; - } - if (strncmp(stkbuf, reject_str, strlen(reject_str)) == 0) - op = RTAS_REJECT_TMP_IMG; - else if (strncmp(stkbuf, commit_str, strlen(commit_str)) == 0) - op = RTAS_COMMIT_TMP_IMG; - } - - if (op == -1) /* buf is empty, or contains invalid string */ - return -EINVAL; - - args_buf->op = op; - manage_flash(args_buf); - - return count; -} - -static void validate_flash(struct rtas_validate_flash_t *args_buf) -{ - int token = rtas_token("ibm,validate-flash-image"); - unsigned int wait_time; - int update_results; - s32 rc; - - rc = 0; - while(1) { - spin_lock(&rtas_data_buf_lock); - memcpy(rtas_data_buf, args_buf->buf, VALIDATE_BUF_SIZE); - rc = rtas_call(token, 2, 2, &update_results, - (u32) __pa(rtas_data_buf), args_buf->buf_size); - memcpy(args_buf->buf, rtas_data_buf, VALIDATE_BUF_SIZE); - spin_unlock(&rtas_data_buf_lock); - - if (rc == RTAS_RC_BUSY) - udelay(1); - else if (rtas_is_extended_busy(rc)) { - wait_time = rtas_extended_busy_delay_time(rc); - udelay(wait_time * 1000); - } else - break; - } - - args_buf->status = rc; - args_buf->update_results = update_results; -} - -static int get_validate_flash_msg(struct rtas_validate_flash_t *args_buf, - char *msg) -{ - int n; - - if (args_buf->status >= VALIDATE_TMP_UPDATE) { - n = sprintf(msg, "%d\n", args_buf->update_results); - if ((args_buf->update_results >= VALIDATE_CUR_UNKNOWN) || - (args_buf->update_results == VALIDATE_TMP_UPDATE)) - n += sprintf(msg + n, "%s\n", args_buf->buf); - } else { - n = sprintf(msg, "%d\n", args_buf->status); - } - return n; -} - -static ssize_t validate_flash_read(struct file *file, char __user *buf, - size_t count, loff_t *ppos) -{ - struct proc_dir_entry *dp = PDE(file->f_dentry->d_inode); - struct rtas_validate_flash_t *args_buf; - char msg[RTAS_MSG_MAXLEN]; - int msglen; - - args_buf = (struct rtas_validate_flash_t *) dp->data; - - if (ppos && *ppos != 0) - return 0; /* be cheap */ - - msglen = get_validate_flash_msg(args_buf, msg); - if (msglen > count) - msglen = count; - - if (!access_ok(VERIFY_WRITE, buf, msglen)) - return -EINVAL; - - if (copy_to_user(buf, msg, msglen)) - return -EFAULT; - - if (ppos) - *ppos = msglen; - return msglen; -} - -static ssize_t validate_flash_write(struct file *file, const char __user *buf, - size_t count, loff_t *off) -{ - struct proc_dir_entry *dp = PDE(file->f_dentry->d_inode); - struct rtas_validate_flash_t *args_buf; - int rc; - - args_buf = (struct rtas_validate_flash_t *) dp->data; - - if (dp->data == NULL) { - dp->data = kmalloc(sizeof(struct rtas_validate_flash_t), - GFP_KERNEL); - if (dp->data == NULL) - return -ENOMEM; - } - - /* We are only interested in the first 4K of the - * candidate image */ - if ((*off >= VALIDATE_BUF_SIZE) || - (args_buf->status == VALIDATE_AUTH)) { - *off += count; - return count; - } - - if (*off + count >= VALIDATE_BUF_SIZE) { - count = VALIDATE_BUF_SIZE - *off; - args_buf->status = VALIDATE_READY; - } else { - args_buf->status = VALIDATE_INCOMPLETE; - } - - if (!access_ok(VERIFY_READ, buf, count)) { - rc = -EFAULT; - goto done; - } - if (copy_from_user(args_buf->buf + *off, buf, count)) { - rc = -EFAULT; - goto done; - } - - *off += count; - rc = count; -done: - if (rc < 0) { - kfree(dp->data); - dp->data = NULL; - } - return rc; -} - -static int validate_flash_release(struct inode *inode, struct file *file) -{ - struct proc_dir_entry *dp = PDE(file->f_dentry->d_inode); - struct rtas_validate_flash_t *args_buf; - - args_buf = (struct rtas_validate_flash_t *) dp->data; - - if (args_buf->status == VALIDATE_READY) { - args_buf->buf_size = VALIDATE_BUF_SIZE; - validate_flash(args_buf); - } - - /* The matching atomic_inc was in rtas_excl_open() */ - atomic_dec(&dp->count); - - return 0; -} - -static void remove_flash_pde(struct proc_dir_entry *dp) -{ - if (dp) { - if (dp->data != NULL) - kfree(dp->data); - dp->owner = NULL; - remove_proc_entry(dp->name, dp->parent); - } -} - -static int initialize_flash_pde_data(const char *rtas_call_name, - size_t buf_size, - struct proc_dir_entry *dp) -{ - int *status; - int token; - - dp->data = kmalloc(buf_size, GFP_KERNEL); - if (dp->data == NULL) { - remove_flash_pde(dp); - return -ENOMEM; - } - - memset(dp->data, 0, buf_size); - - /* - * This code assumes that the status int is the first member of the - * struct - */ - status = (int *) dp->data; - token = rtas_token(rtas_call_name); - if (token == RTAS_UNKNOWN_SERVICE) - *status = FLASH_AUTH; - else - *status = FLASH_NO_OP; - - return 0; -} - -static struct proc_dir_entry *create_flash_pde(const char *filename, - struct file_operations *fops) -{ - struct proc_dir_entry *ent = NULL; - - ent = create_proc_entry(filename, S_IRUSR | S_IWUSR, NULL); - if (ent != NULL) { - ent->nlink = 1; - ent->proc_fops = fops; - ent->owner = THIS_MODULE; - } - - return ent; -} - -static struct file_operations rtas_flash_operations = { - .read = rtas_flash_read, - .write = rtas_flash_write, - .open = rtas_excl_open, - .release = rtas_flash_release, -}; - -static struct file_operations manage_flash_operations = { - .read = manage_flash_read, - .write = manage_flash_write, - .open = rtas_excl_open, - .release = rtas_excl_release, -}; - -static struct file_operations validate_flash_operations = { - .read = validate_flash_read, - .write = validate_flash_write, - .open = rtas_excl_open, - .release = validate_flash_release, -}; - -int __init rtas_flash_init(void) -{ - int rc; - - if (rtas_token("ibm,update-flash-64-and-reboot") == - RTAS_UNKNOWN_SERVICE) { - printk(KERN_ERR "rtas_flash: no firmware flash support\n"); - return 1; - } - - firmware_flash_pde = create_flash_pde("ppc64/rtas/" - FIRMWARE_FLASH_NAME, - &rtas_flash_operations); - if (firmware_flash_pde == NULL) { - rc = -ENOMEM; - goto cleanup; - } - - rc = initialize_flash_pde_data("ibm,update-flash-64-and-reboot", - sizeof(struct rtas_update_flash_t), - firmware_flash_pde); - if (rc != 0) - goto cleanup; - - firmware_update_pde = create_flash_pde("ppc64/rtas/" - FIRMWARE_UPDATE_NAME, - &rtas_flash_operations); - if (firmware_update_pde == NULL) { - rc = -ENOMEM; - goto cleanup; - } - - rc = initialize_flash_pde_data("ibm,update-flash-64-and-reboot", - sizeof(struct rtas_update_flash_t), - firmware_update_pde); - if (rc != 0) - goto cleanup; - - validate_pde = create_flash_pde("ppc64/rtas/" VALIDATE_FLASH_NAME, - &validate_flash_operations); - if (validate_pde == NULL) { - rc = -ENOMEM; - goto cleanup; - } - - rc = initialize_flash_pde_data("ibm,validate-flash-image", - sizeof(struct rtas_validate_flash_t), - validate_pde); - if (rc != 0) - goto cleanup; - - manage_pde = create_flash_pde("ppc64/rtas/" MANAGE_FLASH_NAME, - &manage_flash_operations); - if (manage_pde == NULL) { - rc = -ENOMEM; - goto cleanup; - } - - rc = initialize_flash_pde_data("ibm,manage-flash-image", - sizeof(struct rtas_manage_flash_t), - manage_pde); - if (rc != 0) - goto cleanup; - - return 0; - -cleanup: - remove_flash_pde(firmware_flash_pde); - remove_flash_pde(firmware_update_pde); - remove_flash_pde(validate_pde); - remove_flash_pde(manage_pde); - - return rc; -} - -void __exit rtas_flash_cleanup(void) -{ - remove_flash_pde(firmware_flash_pde); - remove_flash_pde(firmware_update_pde); - remove_flash_pde(validate_pde); - remove_flash_pde(manage_pde); -} - -module_init(rtas_flash_init); -module_exit(rtas_flash_cleanup); -MODULE_LICENSE("GPL"); diff --git a/arch/ppc64/kernel/rtasd.c b/arch/ppc64/kernel/rtasd.c deleted file mode 100644 index e26b0420b6dd..000000000000 --- a/arch/ppc64/kernel/rtasd.c +++ /dev/null @@ -1,527 +0,0 @@ -/* - * Copyright (C) 2001 Anton Blanchard , IBM - * - * This program is free software; you can redistribute it and/or - * modify it under the terms of the GNU General Public License - * as published by the Free Software Foundation; either version - * 2 of the License, or (at your option) any later version. - * - * Communication to userspace based on kernel/printk.c - */ - -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include - -#include -#include -#include -#include -#include -#include -#include - -#if 0 -#define DEBUG(A...) printk(KERN_ERR A) -#else -#define DEBUG(A...) -#endif - -static DEFINE_SPINLOCK(rtasd_log_lock); - -DECLARE_WAIT_QUEUE_HEAD(rtas_log_wait); - -static char *rtas_log_buf; -static unsigned long rtas_log_start; -static unsigned long rtas_log_size; - -static int surveillance_timeout = -1; -static unsigned int rtas_event_scan_rate; -static unsigned int rtas_error_log_max; -static unsigned int rtas_error_log_buffer_max; - -static int full_rtas_msgs = 0; - -extern int no_logging; - -volatile int error_log_cnt = 0; - -/* - * Since we use 32 bit RTAS, the physical address of this must be below - * 4G or else bad things happen. Allocate this in the kernel data and - * make it big enough. - */ -static unsigned char logdata[RTAS_ERROR_LOG_MAX]; - -static int get_eventscan_parms(void); - -static char *rtas_type[] = { - "Unknown", "Retry", "TCE Error", "Internal Device Failure", - "Timeout", "Data Parity", "Address Parity", "Cache Parity", - "Address Invalid", "ECC Uncorrected", "ECC Corrupted", -}; - -static char *rtas_event_type(int type) -{ - if ((type > 0) && (type < 11)) - return rtas_type[type]; - - switch (type) { - case RTAS_TYPE_EPOW: - return "EPOW"; - case RTAS_TYPE_PLATFORM: - return "Platform Error"; - case RTAS_TYPE_IO: - return "I/O Event"; - case RTAS_TYPE_INFO: - return "Platform Information Event"; - case RTAS_TYPE_DEALLOC: - return "Resource Deallocation Event"; - case RTAS_TYPE_DUMP: - return "Dump Notification Event"; - } - - return rtas_type[0]; -} - -/* To see this info, grep RTAS /var/log/messages and each entry - * will be collected together with obvious begin/end. - * There will be a unique identifier on the begin and end lines. - * This will persist across reboots. - * - * format of error logs returned from RTAS: - * bytes (size) : contents - * -------------------------------------------------------- - * 0-7 (8) : rtas_error_log - * 8-47 (40) : extended info - * 48-51 (4) : vendor id - * 52-1023 (vendor specific) : location code and debug data - */ -static void printk_log_rtas(char *buf, int len) -{ - - int i,j,n = 0; - int perline = 16; - char buffer[64]; - char * str = "RTAS event"; - - if (full_rtas_msgs) { - printk(RTAS_DEBUG "%d -------- %s begin --------\n", - error_log_cnt, str); - - /* - * Print perline bytes on each line, each line will start - * with RTAS and a changing number, so syslogd will - * print lines that are otherwise the same. Separate every - * 4 bytes with a space. - */ - for (i = 0; i < len; i++) { - j = i % perline; - if (j == 0) { - memset(buffer, 0, sizeof(buffer)); - n = sprintf(buffer, "RTAS %d:", i/perline); - } - - if ((i % 4) == 0) - n += sprintf(buffer+n, " "); - - n += sprintf(buffer+n, "%02x", (unsigned char)buf[i]); - - if (j == (perline-1)) - printk(KERN_DEBUG "%s\n", buffer); - } - if ((i % perline) != 0) - printk(KERN_DEBUG "%s\n", buffer); - - printk(RTAS_DEBUG "%d -------- %s end ----------\n", - error_log_cnt, str); - } else { - struct rtas_error_log *errlog = (struct rtas_error_log *)buf; - - printk(RTAS_DEBUG "event: %d, Type: %s, Severity: %d\n", - error_log_cnt, rtas_event_type(errlog->type), - errlog->severity); - } -} - -static int log_rtas_len(char * buf) -{ - int len; - struct rtas_error_log *err; - - /* rtas fixed header */ - len = 8; - err = (struct rtas_error_log *)buf; - if (err->extended_log_length) { - - /* extended header */ - len += err->extended_log_length; - } - - if (rtas_error_log_max == 0) { - get_eventscan_parms(); - } - if (len > rtas_error_log_max) - len = rtas_error_log_max; - - return len; -} - -/* - * First write to nvram, if fatal error, that is the only - * place we log the info. The error will be picked up - * on the next reboot by rtasd. If not fatal, run the - * method for the type of error. Currently, only RTAS - * errors have methods implemented, but in the future - * there might be a need to store data in nvram before a - * call to panic(). - * - * XXX We write to nvram periodically, to indicate error has - * been written and sync'd, but there is a possibility - * that if we don't shutdown correctly, a duplicate error - * record will be created on next reboot. - */ -void pSeries_log_error(char *buf, unsigned int err_type, int fatal) -{ - unsigned long offset; - unsigned long s; - int len = 0; - - DEBUG("logging event\n"); - if (buf == NULL) - return; - - spin_lock_irqsave(&rtasd_log_lock, s); - - /* get length and increase count */ - switch (err_type & ERR_TYPE_MASK) { - case ERR_TYPE_RTAS_LOG: - len = log_rtas_len(buf); - if (!(err_type & ERR_FLAG_BOOT)) - error_log_cnt++; - break; - case ERR_TYPE_KERNEL_PANIC: - default: - spin_unlock_irqrestore(&rtasd_log_lock, s); - return; - } - - /* Write error to NVRAM */ - if (!no_logging && !(err_type & ERR_FLAG_BOOT)) - nvram_write_error_log(buf, len, err_type); - - /* - * rtas errors can occur during boot, and we do want to capture - * those somewhere, even if nvram isn't ready (why not?), and even - * if rtasd isn't ready. Put them into the boot log, at least. - */ - if ((err_type & ERR_TYPE_MASK) == ERR_TYPE_RTAS_LOG) - printk_log_rtas(buf, len); - - /* Check to see if we need to or have stopped logging */ - if (fatal || no_logging) { - no_logging = 1; - spin_unlock_irqrestore(&rtasd_log_lock, s); - return; - } - - /* call type specific method for error */ - switch (err_type & ERR_TYPE_MASK) { - case ERR_TYPE_RTAS_LOG: - offset = rtas_error_log_buffer_max * - ((rtas_log_start+rtas_log_size) & LOG_NUMBER_MASK); - - /* First copy over sequence number */ - memcpy(&rtas_log_buf[offset], (void *) &error_log_cnt, sizeof(int)); - - /* Second copy over error log data */ - offset += sizeof(int); - memcpy(&rtas_log_buf[offset], buf, len); - - if (rtas_log_size < LOG_NUMBER) - rtas_log_size += 1; - else - rtas_log_start += 1; - - spin_unlock_irqrestore(&rtasd_log_lock, s); - wake_up_interruptible(&rtas_log_wait); - break; - case ERR_TYPE_KERNEL_PANIC: - default: - spin_unlock_irqrestore(&rtasd_log_lock, s); - return; - } - -} - - -static int rtas_log_open(struct inode * inode, struct file * file) -{ - return 0; -} - -static int rtas_log_release(struct inode * inode, struct file * file) -{ - return 0; -} - -/* This will check if all events are logged, if they are then, we - * know that we can safely clear the events in NVRAM. - * Next we'll sit and wait for something else to log. - */ -static ssize_t rtas_log_read(struct file * file, char __user * buf, - size_t count, loff_t *ppos) -{ - int error; - char *tmp; - unsigned long s; - unsigned long offset; - - if (!buf || count < rtas_error_log_buffer_max) - return -EINVAL; - - count = rtas_error_log_buffer_max; - - if (!access_ok(VERIFY_WRITE, buf, count)) - return -EFAULT; - - tmp = kmalloc(count, GFP_KERNEL); - if (!tmp) - return -ENOMEM; - - - spin_lock_irqsave(&rtasd_log_lock, s); - /* if it's 0, then we know we got the last one (the one in NVRAM) */ - if (rtas_log_size == 0 && !no_logging) - nvram_clear_error_log(); - spin_unlock_irqrestore(&rtasd_log_lock, s); - - - error = wait_event_interruptible(rtas_log_wait, rtas_log_size); - if (error) - goto out; - - spin_lock_irqsave(&rtasd_log_lock, s); - offset = rtas_error_log_buffer_max * (rtas_log_start & LOG_NUMBER_MASK); - memcpy(tmp, &rtas_log_buf[offset], count); - - rtas_log_start += 1; - rtas_log_size -= 1; - spin_unlock_irqrestore(&rtasd_log_lock, s); - - error = copy_to_user(buf, tmp, count) ? -EFAULT : count; -out: - kfree(tmp); - return error; -} - -static unsigned int rtas_log_poll(struct file *file, poll_table * wait) -{ - poll_wait(file, &rtas_log_wait, wait); - if (rtas_log_size) - return POLLIN | POLLRDNORM; - return 0; -} - -struct file_operations proc_rtas_log_operations = { - .read = rtas_log_read, - .poll = rtas_log_poll, - .open = rtas_log_open, - .release = rtas_log_release, -}; - -static int enable_surveillance(int timeout) -{ - int error; - - error = rtas_set_indicator(SURVEILLANCE_TOKEN, 0, timeout); - - if (error == 0) - return 0; - - if (error == -EINVAL) { - printk(KERN_INFO "rtasd: surveillance not supported\n"); - return 0; - } - - printk(KERN_ERR "rtasd: could not update surveillance\n"); - return -1; -} - -static int get_eventscan_parms(void) -{ - struct device_node *node; - int *ip; - - node = of_find_node_by_path("/rtas"); - - ip = (int *)get_property(node, "rtas-event-scan-rate", NULL); - if (ip == NULL) { - printk(KERN_ERR "rtasd: no rtas-event-scan-rate\n"); - of_node_put(node); - return -1; - } - rtas_event_scan_rate = *ip; - DEBUG("rtas-event-scan-rate %d\n", rtas_event_scan_rate); - - /* Make room for the sequence number */ - rtas_error_log_max = rtas_get_error_log_max(); - rtas_error_log_buffer_max = rtas_error_log_max + sizeof(int); - - of_node_put(node); - - return 0; -} - -static void do_event_scan(int event_scan) -{ - int error; - do { - memset(logdata, 0, rtas_error_log_max); - error = rtas_call(event_scan, 4, 1, NULL, - RTAS_EVENT_SCAN_ALL_EVENTS, 0, - __pa(logdata), rtas_error_log_max); - if (error == -1) { - printk(KERN_ERR "event-scan failed\n"); - break; - } - - if (error == 0) - pSeries_log_error(logdata, ERR_TYPE_RTAS_LOG, 0); - - } while(error == 0); -} - -static void do_event_scan_all_cpus(long delay) -{ - int cpu; - - lock_cpu_hotplug(); - cpu = first_cpu(cpu_online_map); - for (;;) { - set_cpus_allowed(current, cpumask_of_cpu(cpu)); - do_event_scan(rtas_token("event-scan")); - set_cpus_allowed(current, CPU_MASK_ALL); - - /* Drop hotplug lock, and sleep for the specified delay */ - unlock_cpu_hotplug(); - msleep_interruptible(delay); - lock_cpu_hotplug(); - - cpu = next_cpu(cpu, cpu_online_map); - if (cpu == NR_CPUS) - break; - } - unlock_cpu_hotplug(); -} - -static int rtasd(void *unused) -{ - unsigned int err_type; - int event_scan = rtas_token("event-scan"); - int rc; - - daemonize("rtasd"); - - if (event_scan == RTAS_UNKNOWN_SERVICE || get_eventscan_parms() == -1) - goto error; - - rtas_log_buf = vmalloc(rtas_error_log_buffer_max*LOG_NUMBER); - if (!rtas_log_buf) { - printk(KERN_ERR "rtasd: no memory\n"); - goto error; - } - - printk(KERN_INFO "RTAS daemon started\n"); - - DEBUG("will sleep for %d milliseconds\n", (30000/rtas_event_scan_rate)); - - /* See if we have any error stored in NVRAM */ - memset(logdata, 0, rtas_error_log_max); - - rc = nvram_read_error_log(logdata, rtas_error_log_max, &err_type); - - /* We can use rtas_log_buf now */ - no_logging = 0; - - if (!rc) { - if (err_type != ERR_FLAG_ALREADY_LOGGED) { - pSeries_log_error(logdata, err_type | ERR_FLAG_BOOT, 0); - } - } - - /* First pass. */ - do_event_scan_all_cpus(1000); - - if (surveillance_timeout != -1) { - DEBUG("enabling surveillance\n"); - enable_surveillance(surveillance_timeout); - DEBUG("surveillance enabled\n"); - } - - /* Delay should be at least one second since some - * machines have problems if we call event-scan too - * quickly. */ - for (;;) - do_event_scan_all_cpus(30000/rtas_event_scan_rate); - -error: - /* Should delete proc entries */ - return -EINVAL; -} - -static int __init rtas_init(void) -{ - struct proc_dir_entry *entry; - - /* No RTAS, only warn if we are on a pSeries box */ - if (rtas_token("event-scan") == RTAS_UNKNOWN_SERVICE) { - if (systemcfg->platform & PLATFORM_PSERIES) - printk(KERN_INFO "rtasd: no event-scan on system\n"); - return 1; - } - - entry = create_proc_entry("ppc64/rtas/error_log", S_IRUSR, NULL); - if (entry) - entry->proc_fops = &proc_rtas_log_operations; - else - printk(KERN_ERR "Failed to create error_log proc entry\n"); - - if (kernel_thread(rtasd, NULL, CLONE_FS) < 0) - printk(KERN_ERR "Failed to start RTAS daemon\n"); - - return 0; -} - -static int __init surveillance_setup(char *str) -{ - int i; - - if (get_option(&str,&i)) { - if (i >= 0 && i <= 255) - surveillance_timeout = i; - } - - return 1; -} - -static int __init rtasmsgs_setup(char *str) -{ - if (strcmp(str, "on") == 0) - full_rtas_msgs = 1; - else if (strcmp(str, "off") == 0) - full_rtas_msgs = 0; - - return 1; -} -__initcall(rtas_init); -__setup("surveillance=", surveillance_setup); -__setup("rtasmsgs=", rtasmsgs_setup); diff --git a/include/asm-powerpc/rtas.h b/include/asm-powerpc/rtas.h index d9fd7866927f..d1bb611ea626 100644 --- a/include/asm-powerpc/rtas.h +++ b/include/asm-powerpc/rtas.h @@ -149,31 +149,11 @@ struct rtas_error_log { unsigned char buffer[1]; }; -struct flash_block { - char *data; - unsigned long length; -}; - -/* This struct is very similar but not identical to - * that needed by the rtas flash update. - * All we need to do for rtas is rewrite num_blocks - * into a version/length and translate the pointers - * to absolute. +/* + * This can be set by the rtas_flash module so that it can get called + * as the absolutely last thing before the kernel terminates. */ -#define FLASH_BLOCKS_PER_NODE ((PAGE_SIZE - 16) / sizeof(struct flash_block)) -struct flash_block_list { - unsigned long num_blocks; - struct flash_block_list *next; - struct flash_block blocks[FLASH_BLOCKS_PER_NODE]; -}; -struct flash_block_list_header { /* just the header of flash_block_list */ - unsigned long num_blocks; - struct flash_block_list *next; -}; -extern struct flash_block_list_header rtas_firmware_flash_list; -void rtas_fw_restart(char *cmd); -void rtas_fw_power_off(void); -void rtas_fw_halt(void); +extern void (*rtas_flash_term_hook)(int); extern struct rtas_t rtas; -- cgit v1.2.3 From 62f4f47da64a10630d5f6dabc616477b0e09b612 Mon Sep 17 00:00:00 2001 From: Stephen Rothwell Date: Thu, 3 Nov 2005 15:26:23 +1100 Subject: powerpc: move ppc64/kernel/signal.c to arch/powerpc Signed-off-by: Stephen Rothwell --- arch/powerpc/kernel/Makefile | 2 +- arch/powerpc/kernel/signal_64.c | 581 ++++++++++++++++++++++++++++++++++++++++ arch/ppc64/kernel/Makefile | 1 - arch/ppc64/kernel/signal.c | 581 ---------------------------------------- 4 files changed, 582 insertions(+), 583 deletions(-) create mode 100644 arch/powerpc/kernel/signal_64.c delete mode 100644 arch/ppc64/kernel/signal.c (limited to 'arch/ppc64') diff --git a/arch/powerpc/kernel/Makefile b/arch/powerpc/kernel/Makefile index 601ddbf1c20b..7a3e1155ac9a 100644 --- a/arch/powerpc/kernel/Makefile +++ b/arch/powerpc/kernel/Makefile @@ -13,7 +13,7 @@ endif obj-y := semaphore.o cputable.o ptrace.o syscalls.o \ signal_32.o pmc.o obj-$(CONFIG_PPC64) += setup_64.o binfmt_elf32.o sys_ppc32.o \ - ptrace32.o systbl.o + signal_64.o ptrace32.o systbl.o obj-$(CONFIG_ALTIVEC) += vecemu.o vector.o obj-$(CONFIG_POWER4) += idle_power4.o obj-$(CONFIG_PPC_OF) += of_device.o diff --git a/arch/powerpc/kernel/signal_64.c b/arch/powerpc/kernel/signal_64.c new file mode 100644 index 000000000000..ec9d0984b6a0 --- /dev/null +++ b/arch/powerpc/kernel/signal_64.c @@ -0,0 +1,581 @@ +/* + * linux/arch/ppc64/kernel/signal.c + * + * PowerPC version + * Copyright (C) 1995-1996 Gary Thomas (gdt@linuxppc.org) + * + * Derived from "arch/i386/kernel/signal.c" + * Copyright (C) 1991, 1992 Linus Torvalds + * 1997-11-28 Modified for POSIX.1b signals by Richard Henderson + * + * This program is free software; you can redistribute it and/or + * modify it under the terms of the GNU General Public License + * as published by the Free Software Foundation; either version + * 2 of the License, or (at your option) any later version. + */ + +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include + +#include +#include +#include +#include +#include +#include +#include +#include + +#define DEBUG_SIG 0 + +#define _BLOCKABLE (~(sigmask(SIGKILL) | sigmask(SIGSTOP))) + +#define GP_REGS_SIZE min(sizeof(elf_gregset_t), sizeof(struct pt_regs)) +#define FP_REGS_SIZE sizeof(elf_fpregset_t) + +#define TRAMP_TRACEBACK 3 +#define TRAMP_SIZE 6 + +/* + * When we have signals to deliver, we set up on the user stack, + * going down from the original stack pointer: + * 1) a rt_sigframe struct which contains the ucontext + * 2) a gap of __SIGNAL_FRAMESIZE bytes which acts as a dummy caller + * frame for the signal handler. + */ + +struct rt_sigframe { + /* sys_rt_sigreturn requires the ucontext be the first field */ + struct ucontext uc; + unsigned long _unused[2]; + unsigned int tramp[TRAMP_SIZE]; + struct siginfo *pinfo; + void *puc; + struct siginfo info; + /* 64 bit ABI allows for 288 bytes below sp before decrementing it. */ + char abigap[288]; +} __attribute__ ((aligned (16))); + + +/* + * Atomically swap in the new signal mask, and wait for a signal. + */ +long sys_rt_sigsuspend(sigset_t __user *unewset, size_t sigsetsize, int p3, int p4, + int p6, int p7, struct pt_regs *regs) +{ + sigset_t saveset, newset; + + /* XXX: Don't preclude handling different sized sigset_t's. */ + if (sigsetsize != sizeof(sigset_t)) + return -EINVAL; + + if (copy_from_user(&newset, unewset, sizeof(newset))) + return -EFAULT; + sigdelsetmask(&newset, ~_BLOCKABLE); + + spin_lock_irq(¤t->sighand->siglock); + saveset = current->blocked; + current->blocked = newset; + recalc_sigpending(); + spin_unlock_irq(¤t->sighand->siglock); + + regs->result = -EINTR; + regs->gpr[3] = EINTR; + regs->ccr |= 0x10000000; + while (1) { + current->state = TASK_INTERRUPTIBLE; + schedule(); + if (do_signal(&saveset, regs)) + return 0; + } +} + +long sys_sigaltstack(const stack_t __user *uss, stack_t __user *uoss, unsigned long r5, + unsigned long r6, unsigned long r7, unsigned long r8, + struct pt_regs *regs) +{ + return do_sigaltstack(uss, uoss, regs->gpr[1]); +} + + +/* + * Set up the sigcontext for the signal frame. + */ + +static long setup_sigcontext(struct sigcontext __user *sc, struct pt_regs *regs, + int signr, sigset_t *set, unsigned long handler) +{ + /* When CONFIG_ALTIVEC is set, we _always_ setup v_regs even if the + * process never used altivec yet (MSR_VEC is zero in pt_regs of + * the context). This is very important because we must ensure we + * don't lose the VRSAVE content that may have been set prior to + * the process doing its first vector operation + * Userland shall check AT_HWCAP to know wether it can rely on the + * v_regs pointer or not + */ +#ifdef CONFIG_ALTIVEC + elf_vrreg_t __user *v_regs = (elf_vrreg_t __user *)(((unsigned long)sc->vmx_reserve + 15) & ~0xful); +#endif + long err = 0; + + flush_fp_to_thread(current); + + /* Make sure signal doesn't get spurrious FP exceptions */ + current->thread.fpscr.val = 0; + +#ifdef CONFIG_ALTIVEC + err |= __put_user(v_regs, &sc->v_regs); + + /* save altivec registers */ + if (current->thread.used_vr) { + flush_altivec_to_thread(current); + /* Copy 33 vec registers (vr0..31 and vscr) to the stack */ + err |= __copy_to_user(v_regs, current->thread.vr, 33 * sizeof(vector128)); + /* set MSR_VEC in the MSR value in the frame to indicate that sc->v_reg) + * contains valid data. + */ + regs->msr |= MSR_VEC; + } + /* We always copy to/from vrsave, it's 0 if we don't have or don't + * use altivec. + */ + err |= __put_user(current->thread.vrsave, (u32 __user *)&v_regs[33]); +#else /* CONFIG_ALTIVEC */ + err |= __put_user(0, &sc->v_regs); +#endif /* CONFIG_ALTIVEC */ + err |= __put_user(&sc->gp_regs, &sc->regs); + err |= __copy_to_user(&sc->gp_regs, regs, GP_REGS_SIZE); + err |= __copy_to_user(&sc->fp_regs, ¤t->thread.fpr, FP_REGS_SIZE); + err |= __put_user(signr, &sc->signal); + err |= __put_user(handler, &sc->handler); + if (set != NULL) + err |= __put_user(set->sig[0], &sc->oldmask); + + return err; +} + +/* + * Restore the sigcontext from the signal frame. + */ + +static long restore_sigcontext(struct pt_regs *regs, sigset_t *set, int sig, + struct sigcontext __user *sc) +{ +#ifdef CONFIG_ALTIVEC + elf_vrreg_t __user *v_regs; +#endif + unsigned long err = 0; + unsigned long save_r13 = 0; + elf_greg_t *gregs = (elf_greg_t *)regs; +#ifdef CONFIG_ALTIVEC + unsigned long msr; +#endif + int i; + + /* If this is not a signal return, we preserve the TLS in r13 */ + if (!sig) + save_r13 = regs->gpr[13]; + + /* copy everything before MSR */ + err |= __copy_from_user(regs, &sc->gp_regs, + PT_MSR*sizeof(unsigned long)); + + /* skip MSR and SOFTE */ + for (i = PT_MSR+1; i <= PT_RESULT; i++) { + if (i == PT_SOFTE) + continue; + err |= __get_user(gregs[i], &sc->gp_regs[i]); + } + + if (!sig) + regs->gpr[13] = save_r13; + err |= __copy_from_user(¤t->thread.fpr, &sc->fp_regs, FP_REGS_SIZE); + if (set != NULL) + err |= __get_user(set->sig[0], &sc->oldmask); + +#ifdef CONFIG_ALTIVEC + err |= __get_user(v_regs, &sc->v_regs); + err |= __get_user(msr, &sc->gp_regs[PT_MSR]); + if (err) + return err; + /* Copy 33 vec registers (vr0..31 and vscr) from the stack */ + if (v_regs != 0 && (msr & MSR_VEC) != 0) + err |= __copy_from_user(current->thread.vr, v_regs, + 33 * sizeof(vector128)); + else if (current->thread.used_vr) + memset(current->thread.vr, 0, 33 * sizeof(vector128)); + /* Always get VRSAVE back */ + if (v_regs != 0) + err |= __get_user(current->thread.vrsave, (u32 __user *)&v_regs[33]); + else + current->thread.vrsave = 0; +#endif /* CONFIG_ALTIVEC */ + +#ifndef CONFIG_SMP + preempt_disable(); + if (last_task_used_math == current) + last_task_used_math = NULL; + if (last_task_used_altivec == current) + last_task_used_altivec = NULL; + preempt_enable(); +#endif + /* Force reload of FP/VEC */ + regs->msr &= ~(MSR_FP | MSR_FE0 | MSR_FE1 | MSR_VEC); + + return err; +} + +/* + * Allocate space for the signal frame + */ +static inline void __user * get_sigframe(struct k_sigaction *ka, struct pt_regs *regs, + size_t frame_size) +{ + unsigned long newsp; + + /* Default to using normal stack */ + newsp = regs->gpr[1]; + + if (ka->sa.sa_flags & SA_ONSTACK) { + if (! on_sig_stack(regs->gpr[1])) + newsp = (current->sas_ss_sp + current->sas_ss_size); + } + + return (void __user *)((newsp - frame_size) & -16ul); +} + +/* + * Setup the trampoline code on the stack + */ +static long setup_trampoline(unsigned int syscall, unsigned int __user *tramp) +{ + int i; + long err = 0; + + /* addi r1, r1, __SIGNAL_FRAMESIZE # Pop the dummy stackframe */ + err |= __put_user(0x38210000UL | (__SIGNAL_FRAMESIZE & 0xffff), &tramp[0]); + /* li r0, __NR_[rt_]sigreturn| */ + err |= __put_user(0x38000000UL | (syscall & 0xffff), &tramp[1]); + /* sc */ + err |= __put_user(0x44000002UL, &tramp[2]); + + /* Minimal traceback info */ + for (i=TRAMP_TRACEBACK; i < TRAMP_SIZE ;i++) + err |= __put_user(0, &tramp[i]); + + if (!err) + flush_icache_range((unsigned long) &tramp[0], + (unsigned long) &tramp[TRAMP_SIZE]); + + return err; +} + +/* + * Restore the user process's signal mask (also used by signal32.c) + */ +void restore_sigmask(sigset_t *set) +{ + sigdelsetmask(set, ~_BLOCKABLE); + spin_lock_irq(¤t->sighand->siglock); + current->blocked = *set; + recalc_sigpending(); + spin_unlock_irq(¤t->sighand->siglock); +} + + +/* + * Handle {get,set,swap}_context operations + */ +int sys_swapcontext(struct ucontext __user *old_ctx, + struct ucontext __user *new_ctx, + long ctx_size, long r6, long r7, long r8, struct pt_regs *regs) +{ + unsigned char tmp; + sigset_t set; + + /* Context size is for future use. Right now, we only make sure + * we are passed something we understand + */ + if (ctx_size < sizeof(struct ucontext)) + return -EINVAL; + + if (old_ctx != NULL) { + if (!access_ok(VERIFY_WRITE, old_ctx, sizeof(*old_ctx)) + || setup_sigcontext(&old_ctx->uc_mcontext, regs, 0, NULL, 0) + || __copy_to_user(&old_ctx->uc_sigmask, + ¤t->blocked, sizeof(sigset_t))) + return -EFAULT; + } + if (new_ctx == NULL) + return 0; + if (!access_ok(VERIFY_READ, new_ctx, sizeof(*new_ctx)) + || __get_user(tmp, (u8 __user *) new_ctx) + || __get_user(tmp, (u8 __user *) (new_ctx + 1) - 1)) + return -EFAULT; + + /* + * If we get a fault copying the context into the kernel's + * image of the user's registers, we can't just return -EFAULT + * because the user's registers will be corrupted. For instance + * the NIP value may have been updated but not some of the + * other registers. Given that we have done the access_ok + * and successfully read the first and last bytes of the region + * above, this should only happen in an out-of-memory situation + * or if another thread unmaps the region containing the context. + * We kill the task with a SIGSEGV in this situation. + */ + + if (__copy_from_user(&set, &new_ctx->uc_sigmask, sizeof(set))) + do_exit(SIGSEGV); + restore_sigmask(&set); + if (restore_sigcontext(regs, NULL, 0, &new_ctx->uc_mcontext)) + do_exit(SIGSEGV); + + /* This returns like rt_sigreturn */ + return 0; +} + + +/* + * Do a signal return; undo the signal stack. + */ + +int sys_rt_sigreturn(unsigned long r3, unsigned long r4, unsigned long r5, + unsigned long r6, unsigned long r7, unsigned long r8, + struct pt_regs *regs) +{ + struct ucontext __user *uc = (struct ucontext __user *)regs->gpr[1]; + sigset_t set; + + /* Always make any pending restarted system calls return -EINTR */ + current_thread_info()->restart_block.fn = do_no_restart_syscall; + + if (!access_ok(VERIFY_READ, uc, sizeof(*uc))) + goto badframe; + + if (__copy_from_user(&set, &uc->uc_sigmask, sizeof(set))) + goto badframe; + restore_sigmask(&set); + if (restore_sigcontext(regs, NULL, 1, &uc->uc_mcontext)) + goto badframe; + + /* do_sigaltstack expects a __user pointer and won't modify + * what's in there anyway + */ + do_sigaltstack(&uc->uc_stack, NULL, regs->gpr[1]); + + return regs->result; + +badframe: +#if DEBUG_SIG + printk("badframe in sys_rt_sigreturn, regs=%p uc=%p &uc->uc_mcontext=%p\n", + regs, uc, &uc->uc_mcontext); +#endif + force_sig(SIGSEGV, current); + return 0; +} + +static int setup_rt_frame(int signr, struct k_sigaction *ka, siginfo_t *info, + sigset_t *set, struct pt_regs *regs) +{ + /* Handler is *really* a pointer to the function descriptor for + * the signal routine. The first entry in the function + * descriptor is the entry address of signal and the second + * entry is the TOC value we need to use. + */ + func_descr_t __user *funct_desc_ptr; + struct rt_sigframe __user *frame; + unsigned long newsp = 0; + long err = 0; + + frame = get_sigframe(ka, regs, sizeof(*frame)); + + if (!access_ok(VERIFY_WRITE, frame, sizeof(*frame))) + goto badframe; + + err |= __put_user(&frame->info, &frame->pinfo); + err |= __put_user(&frame->uc, &frame->puc); + err |= copy_siginfo_to_user(&frame->info, info); + if (err) + goto badframe; + + /* Create the ucontext. */ + err |= __put_user(0, &frame->uc.uc_flags); + err |= __put_user(0, &frame->uc.uc_link); + err |= __put_user(current->sas_ss_sp, &frame->uc.uc_stack.ss_sp); + err |= __put_user(sas_ss_flags(regs->gpr[1]), + &frame->uc.uc_stack.ss_flags); + err |= __put_user(current->sas_ss_size, &frame->uc.uc_stack.ss_size); + err |= setup_sigcontext(&frame->uc.uc_mcontext, regs, signr, NULL, + (unsigned long)ka->sa.sa_handler); + err |= __copy_to_user(&frame->uc.uc_sigmask, set, sizeof(*set)); + if (err) + goto badframe; + + /* Set up to return from userspace. */ + if (vdso64_rt_sigtramp && current->thread.vdso_base) { + regs->link = current->thread.vdso_base + vdso64_rt_sigtramp; + } else { + err |= setup_trampoline(__NR_rt_sigreturn, &frame->tramp[0]); + if (err) + goto badframe; + regs->link = (unsigned long) &frame->tramp[0]; + } + funct_desc_ptr = (func_descr_t __user *) ka->sa.sa_handler; + + /* Allocate a dummy caller frame for the signal handler. */ + newsp = (unsigned long)frame - __SIGNAL_FRAMESIZE; + err |= put_user(regs->gpr[1], (unsigned long __user *)newsp); + + /* Set up "regs" so we "return" to the signal handler. */ + err |= get_user(regs->nip, &funct_desc_ptr->entry); + regs->gpr[1] = newsp; + err |= get_user(regs->gpr[2], &funct_desc_ptr->toc); + regs->gpr[3] = signr; + regs->result = 0; + if (ka->sa.sa_flags & SA_SIGINFO) { + err |= get_user(regs->gpr[4], (unsigned long __user *)&frame->pinfo); + err |= get_user(regs->gpr[5], (unsigned long __user *)&frame->puc); + regs->gpr[6] = (unsigned long) frame; + } else { + regs->gpr[4] = (unsigned long)&frame->uc.uc_mcontext; + } + if (err) + goto badframe; + + if (test_thread_flag(TIF_SINGLESTEP)) + ptrace_notify(SIGTRAP); + + return 1; + +badframe: +#if DEBUG_SIG + printk("badframe in setup_rt_frame, regs=%p frame=%p newsp=%lx\n", + regs, frame, newsp); +#endif + force_sigsegv(signr, current); + return 0; +} + + +/* + * OK, we're invoking a handler + */ +static int handle_signal(unsigned long sig, struct k_sigaction *ka, + siginfo_t *info, sigset_t *oldset, struct pt_regs *regs) +{ + int ret; + + /* Set up Signal Frame */ + ret = setup_rt_frame(sig, ka, info, oldset, regs); + + if (ret) { + spin_lock_irq(¤t->sighand->siglock); + sigorsets(¤t->blocked, ¤t->blocked, &ka->sa.sa_mask); + if (!(ka->sa.sa_flags & SA_NODEFER)) + sigaddset(¤t->blocked,sig); + recalc_sigpending(); + spin_unlock_irq(¤t->sighand->siglock); + } + + return ret; +} + +static inline void syscall_restart(struct pt_regs *regs, struct k_sigaction *ka) +{ + switch ((int)regs->result) { + case -ERESTART_RESTARTBLOCK: + case -ERESTARTNOHAND: + /* ERESTARTNOHAND means that the syscall should only be + * restarted if there was no handler for the signal, and since + * we only get here if there is a handler, we dont restart. + */ + regs->result = -EINTR; + break; + case -ERESTARTSYS: + /* ERESTARTSYS means to restart the syscall if there is no + * handler or the handler was registered with SA_RESTART + */ + if (!(ka->sa.sa_flags & SA_RESTART)) { + regs->result = -EINTR; + break; + } + /* fallthrough */ + case -ERESTARTNOINTR: + /* ERESTARTNOINTR means that the syscall should be + * called again after the signal handler returns. + */ + regs->gpr[3] = regs->orig_gpr3; + regs->nip -= 4; + regs->result = 0; + break; + } +} + +/* + * Note that 'init' is a special process: it doesn't get signals it doesn't + * want to handle. Thus you cannot kill init even with a SIGKILL even by + * mistake. + */ +int do_signal(sigset_t *oldset, struct pt_regs *regs) +{ + siginfo_t info; + int signr; + struct k_sigaction ka; + + /* + * If the current thread is 32 bit - invoke the + * 32 bit signal handling code + */ + if (test_thread_flag(TIF_32BIT)) + return do_signal32(oldset, regs); + + if (!oldset) + oldset = ¤t->blocked; + + signr = get_signal_to_deliver(&info, &ka, regs, NULL); + if (signr > 0) { + /* Whee! Actually deliver the signal. */ + if (TRAP(regs) == 0x0C00) + syscall_restart(regs, &ka); + + /* + * Reenable the DABR before delivering the signal to + * user space. The DABR will have been cleared if it + * triggered inside the kernel. + */ + if (current->thread.dabr) + set_dabr(current->thread.dabr); + + return handle_signal(signr, &ka, &info, oldset, regs); + } + + if (TRAP(regs) == 0x0C00) { /* System Call! */ + if ((int)regs->result == -ERESTARTNOHAND || + (int)regs->result == -ERESTARTSYS || + (int)regs->result == -ERESTARTNOINTR) { + regs->gpr[3] = regs->orig_gpr3; + regs->nip -= 4; /* Back up & retry system call */ + regs->result = 0; + } else if ((int)regs->result == -ERESTART_RESTARTBLOCK) { + regs->gpr[0] = __NR_restart_syscall; + regs->nip -= 4; + regs->result = 0; + } + } + + return 0; +} +EXPORT_SYMBOL(do_signal); diff --git a/arch/ppc64/kernel/Makefile b/arch/ppc64/kernel/Makefile index 74892ad032ba..430cb3900a07 100644 --- a/arch/ppc64/kernel/Makefile +++ b/arch/ppc64/kernel/Makefile @@ -12,7 +12,6 @@ obj-y := misc.o prom.o endif obj-y += irq.o idle.o dma.o \ - signal.o \ align.o pacaData.o \ udbg.o ioctl32.o \ rtc.o \ diff --git a/arch/ppc64/kernel/signal.c b/arch/ppc64/kernel/signal.c deleted file mode 100644 index ec9d0984b6a0..000000000000 --- a/arch/ppc64/kernel/signal.c +++ /dev/null @@ -1,581 +0,0 @@ -/* - * linux/arch/ppc64/kernel/signal.c - * - * PowerPC version - * Copyright (C) 1995-1996 Gary Thomas (gdt@linuxppc.org) - * - * Derived from "arch/i386/kernel/signal.c" - * Copyright (C) 1991, 1992 Linus Torvalds - * 1997-11-28 Modified for POSIX.1b signals by Richard Henderson - * - * This program is free software; you can redistribute it and/or - * modify it under the terms of the GNU General Public License - * as published by the Free Software Foundation; either version - * 2 of the License, or (at your option) any later version. - */ - -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include - -#include -#include -#include -#include -#include -#include -#include -#include - -#define DEBUG_SIG 0 - -#define _BLOCKABLE (~(sigmask(SIGKILL) | sigmask(SIGSTOP))) - -#define GP_REGS_SIZE min(sizeof(elf_gregset_t), sizeof(struct pt_regs)) -#define FP_REGS_SIZE sizeof(elf_fpregset_t) - -#define TRAMP_TRACEBACK 3 -#define TRAMP_SIZE 6 - -/* - * When we have signals to deliver, we set up on the user stack, - * going down from the original stack pointer: - * 1) a rt_sigframe struct which contains the ucontext - * 2) a gap of __SIGNAL_FRAMESIZE bytes which acts as a dummy caller - * frame for the signal handler. - */ - -struct rt_sigframe { - /* sys_rt_sigreturn requires the ucontext be the first field */ - struct ucontext uc; - unsigned long _unused[2]; - unsigned int tramp[TRAMP_SIZE]; - struct siginfo *pinfo; - void *puc; - struct siginfo info; - /* 64 bit ABI allows for 288 bytes below sp before decrementing it. */ - char abigap[288]; -} __attribute__ ((aligned (16))); - - -/* - * Atomically swap in the new signal mask, and wait for a signal. - */ -long sys_rt_sigsuspend(sigset_t __user *unewset, size_t sigsetsize, int p3, int p4, - int p6, int p7, struct pt_regs *regs) -{ - sigset_t saveset, newset; - - /* XXX: Don't preclude handling different sized sigset_t's. */ - if (sigsetsize != sizeof(sigset_t)) - return -EINVAL; - - if (copy_from_user(&newset, unewset, sizeof(newset))) - return -EFAULT; - sigdelsetmask(&newset, ~_BLOCKABLE); - - spin_lock_irq(¤t->sighand->siglock); - saveset = current->blocked; - current->blocked = newset; - recalc_sigpending(); - spin_unlock_irq(¤t->sighand->siglock); - - regs->result = -EINTR; - regs->gpr[3] = EINTR; - regs->ccr |= 0x10000000; - while (1) { - current->state = TASK_INTERRUPTIBLE; - schedule(); - if (do_signal(&saveset, regs)) - return 0; - } -} - -long sys_sigaltstack(const stack_t __user *uss, stack_t __user *uoss, unsigned long r5, - unsigned long r6, unsigned long r7, unsigned long r8, - struct pt_regs *regs) -{ - return do_sigaltstack(uss, uoss, regs->gpr[1]); -} - - -/* - * Set up the sigcontext for the signal frame. - */ - -static long setup_sigcontext(struct sigcontext __user *sc, struct pt_regs *regs, - int signr, sigset_t *set, unsigned long handler) -{ - /* When CONFIG_ALTIVEC is set, we _always_ setup v_regs even if the - * process never used altivec yet (MSR_VEC is zero in pt_regs of - * the context). This is very important because we must ensure we - * don't lose the VRSAVE content that may have been set prior to - * the process doing its first vector operation - * Userland shall check AT_HWCAP to know wether it can rely on the - * v_regs pointer or not - */ -#ifdef CONFIG_ALTIVEC - elf_vrreg_t __user *v_regs = (elf_vrreg_t __user *)(((unsigned long)sc->vmx_reserve + 15) & ~0xful); -#endif - long err = 0; - - flush_fp_to_thread(current); - - /* Make sure signal doesn't get spurrious FP exceptions */ - current->thread.fpscr.val = 0; - -#ifdef CONFIG_ALTIVEC - err |= __put_user(v_regs, &sc->v_regs); - - /* save altivec registers */ - if (current->thread.used_vr) { - flush_altivec_to_thread(current); - /* Copy 33 vec registers (vr0..31 and vscr) to the stack */ - err |= __copy_to_user(v_regs, current->thread.vr, 33 * sizeof(vector128)); - /* set MSR_VEC in the MSR value in the frame to indicate that sc->v_reg) - * contains valid data. - */ - regs->msr |= MSR_VEC; - } - /* We always copy to/from vrsave, it's 0 if we don't have or don't - * use altivec. - */ - err |= __put_user(current->thread.vrsave, (u32 __user *)&v_regs[33]); -#else /* CONFIG_ALTIVEC */ - err |= __put_user(0, &sc->v_regs); -#endif /* CONFIG_ALTIVEC */ - err |= __put_user(&sc->gp_regs, &sc->regs); - err |= __copy_to_user(&sc->gp_regs, regs, GP_REGS_SIZE); - err |= __copy_to_user(&sc->fp_regs, ¤t->thread.fpr, FP_REGS_SIZE); - err |= __put_user(signr, &sc->signal); - err |= __put_user(handler, &sc->handler); - if (set != NULL) - err |= __put_user(set->sig[0], &sc->oldmask); - - return err; -} - -/* - * Restore the sigcontext from the signal frame. - */ - -static long restore_sigcontext(struct pt_regs *regs, sigset_t *set, int sig, - struct sigcontext __user *sc) -{ -#ifdef CONFIG_ALTIVEC - elf_vrreg_t __user *v_regs; -#endif - unsigned long err = 0; - unsigned long save_r13 = 0; - elf_greg_t *gregs = (elf_greg_t *)regs; -#ifdef CONFIG_ALTIVEC - unsigned long msr; -#endif - int i; - - /* If this is not a signal return, we preserve the TLS in r13 */ - if (!sig) - save_r13 = regs->gpr[13]; - - /* copy everything before MSR */ - err |= __copy_from_user(regs, &sc->gp_regs, - PT_MSR*sizeof(unsigned long)); - - /* skip MSR and SOFTE */ - for (i = PT_MSR+1; i <= PT_RESULT; i++) { - if (i == PT_SOFTE) - continue; - err |= __get_user(gregs[i], &sc->gp_regs[i]); - } - - if (!sig) - regs->gpr[13] = save_r13; - err |= __copy_from_user(¤t->thread.fpr, &sc->fp_regs, FP_REGS_SIZE); - if (set != NULL) - err |= __get_user(set->sig[0], &sc->oldmask); - -#ifdef CONFIG_ALTIVEC - err |= __get_user(v_regs, &sc->v_regs); - err |= __get_user(msr, &sc->gp_regs[PT_MSR]); - if (err) - return err; - /* Copy 33 vec registers (vr0..31 and vscr) from the stack */ - if (v_regs != 0 && (msr & MSR_VEC) != 0) - err |= __copy_from_user(current->thread.vr, v_regs, - 33 * sizeof(vector128)); - else if (current->thread.used_vr) - memset(current->thread.vr, 0, 33 * sizeof(vector128)); - /* Always get VRSAVE back */ - if (v_regs != 0) - err |= __get_user(current->thread.vrsave, (u32 __user *)&v_regs[33]); - else - current->thread.vrsave = 0; -#endif /* CONFIG_ALTIVEC */ - -#ifndef CONFIG_SMP - preempt_disable(); - if (last_task_used_math == current) - last_task_used_math = NULL; - if (last_task_used_altivec == current) - last_task_used_altivec = NULL; - preempt_enable(); -#endif - /* Force reload of FP/VEC */ - regs->msr &= ~(MSR_FP | MSR_FE0 | MSR_FE1 | MSR_VEC); - - return err; -} - -/* - * Allocate space for the signal frame - */ -static inline void __user * get_sigframe(struct k_sigaction *ka, struct pt_regs *regs, - size_t frame_size) -{ - unsigned long newsp; - - /* Default to using normal stack */ - newsp = regs->gpr[1]; - - if (ka->sa.sa_flags & SA_ONSTACK) { - if (! on_sig_stack(regs->gpr[1])) - newsp = (current->sas_ss_sp + current->sas_ss_size); - } - - return (void __user *)((newsp - frame_size) & -16ul); -} - -/* - * Setup the trampoline code on the stack - */ -static long setup_trampoline(unsigned int syscall, unsigned int __user *tramp) -{ - int i; - long err = 0; - - /* addi r1, r1, __SIGNAL_FRAMESIZE # Pop the dummy stackframe */ - err |= __put_user(0x38210000UL | (__SIGNAL_FRAMESIZE & 0xffff), &tramp[0]); - /* li r0, __NR_[rt_]sigreturn| */ - err |= __put_user(0x38000000UL | (syscall & 0xffff), &tramp[1]); - /* sc */ - err |= __put_user(0x44000002UL, &tramp[2]); - - /* Minimal traceback info */ - for (i=TRAMP_TRACEBACK; i < TRAMP_SIZE ;i++) - err |= __put_user(0, &tramp[i]); - - if (!err) - flush_icache_range((unsigned long) &tramp[0], - (unsigned long) &tramp[TRAMP_SIZE]); - - return err; -} - -/* - * Restore the user process's signal mask (also used by signal32.c) - */ -void restore_sigmask(sigset_t *set) -{ - sigdelsetmask(set, ~_BLOCKABLE); - spin_lock_irq(¤t->sighand->siglock); - current->blocked = *set; - recalc_sigpending(); - spin_unlock_irq(¤t->sighand->siglock); -} - - -/* - * Handle {get,set,swap}_context operations - */ -int sys_swapcontext(struct ucontext __user *old_ctx, - struct ucontext __user *new_ctx, - long ctx_size, long r6, long r7, long r8, struct pt_regs *regs) -{ - unsigned char tmp; - sigset_t set; - - /* Context size is for future use. Right now, we only make sure - * we are passed something we understand - */ - if (ctx_size < sizeof(struct ucontext)) - return -EINVAL; - - if (old_ctx != NULL) { - if (!access_ok(VERIFY_WRITE, old_ctx, sizeof(*old_ctx)) - || setup_sigcontext(&old_ctx->uc_mcontext, regs, 0, NULL, 0) - || __copy_to_user(&old_ctx->uc_sigmask, - ¤t->blocked, sizeof(sigset_t))) - return -EFAULT; - } - if (new_ctx == NULL) - return 0; - if (!access_ok(VERIFY_READ, new_ctx, sizeof(*new_ctx)) - || __get_user(tmp, (u8 __user *) new_ctx) - || __get_user(tmp, (u8 __user *) (new_ctx + 1) - 1)) - return -EFAULT; - - /* - * If we get a fault copying the context into the kernel's - * image of the user's registers, we can't just return -EFAULT - * because the user's registers will be corrupted. For instance - * the NIP value may have been updated but not some of the - * other registers. Given that we have done the access_ok - * and successfully read the first and last bytes of the region - * above, this should only happen in an out-of-memory situation - * or if another thread unmaps the region containing the context. - * We kill the task with a SIGSEGV in this situation. - */ - - if (__copy_from_user(&set, &new_ctx->uc_sigmask, sizeof(set))) - do_exit(SIGSEGV); - restore_sigmask(&set); - if (restore_sigcontext(regs, NULL, 0, &new_ctx->uc_mcontext)) - do_exit(SIGSEGV); - - /* This returns like rt_sigreturn */ - return 0; -} - - -/* - * Do a signal return; undo the signal stack. - */ - -int sys_rt_sigreturn(unsigned long r3, unsigned long r4, unsigned long r5, - unsigned long r6, unsigned long r7, unsigned long r8, - struct pt_regs *regs) -{ - struct ucontext __user *uc = (struct ucontext __user *)regs->gpr[1]; - sigset_t set; - - /* Always make any pending restarted system calls return -EINTR */ - current_thread_info()->restart_block.fn = do_no_restart_syscall; - - if (!access_ok(VERIFY_READ, uc, sizeof(*uc))) - goto badframe; - - if (__copy_from_user(&set, &uc->uc_sigmask, sizeof(set))) - goto badframe; - restore_sigmask(&set); - if (restore_sigcontext(regs, NULL, 1, &uc->uc_mcontext)) - goto badframe; - - /* do_sigaltstack expects a __user pointer and won't modify - * what's in there anyway - */ - do_sigaltstack(&uc->uc_stack, NULL, regs->gpr[1]); - - return regs->result; - -badframe: -#if DEBUG_SIG - printk("badframe in sys_rt_sigreturn, regs=%p uc=%p &uc->uc_mcontext=%p\n", - regs, uc, &uc->uc_mcontext); -#endif - force_sig(SIGSEGV, current); - return 0; -} - -static int setup_rt_frame(int signr, struct k_sigaction *ka, siginfo_t *info, - sigset_t *set, struct pt_regs *regs) -{ - /* Handler is *really* a pointer to the function descriptor for - * the signal routine. The first entry in the function - * descriptor is the entry address of signal and the second - * entry is the TOC value we need to use. - */ - func_descr_t __user *funct_desc_ptr; - struct rt_sigframe __user *frame; - unsigned long newsp = 0; - long err = 0; - - frame = get_sigframe(ka, regs, sizeof(*frame)); - - if (!access_ok(VERIFY_WRITE, frame, sizeof(*frame))) - goto badframe; - - err |= __put_user(&frame->info, &frame->pinfo); - err |= __put_user(&frame->uc, &frame->puc); - err |= copy_siginfo_to_user(&frame->info, info); - if (err) - goto badframe; - - /* Create the ucontext. */ - err |= __put_user(0, &frame->uc.uc_flags); - err |= __put_user(0, &frame->uc.uc_link); - err |= __put_user(current->sas_ss_sp, &frame->uc.uc_stack.ss_sp); - err |= __put_user(sas_ss_flags(regs->gpr[1]), - &frame->uc.uc_stack.ss_flags); - err |= __put_user(current->sas_ss_size, &frame->uc.uc_stack.ss_size); - err |= setup_sigcontext(&frame->uc.uc_mcontext, regs, signr, NULL, - (unsigned long)ka->sa.sa_handler); - err |= __copy_to_user(&frame->uc.uc_sigmask, set, sizeof(*set)); - if (err) - goto badframe; - - /* Set up to return from userspace. */ - if (vdso64_rt_sigtramp && current->thread.vdso_base) { - regs->link = current->thread.vdso_base + vdso64_rt_sigtramp; - } else { - err |= setup_trampoline(__NR_rt_sigreturn, &frame->tramp[0]); - if (err) - goto badframe; - regs->link = (unsigned long) &frame->tramp[0]; - } - funct_desc_ptr = (func_descr_t __user *) ka->sa.sa_handler; - - /* Allocate a dummy caller frame for the signal handler. */ - newsp = (unsigned long)frame - __SIGNAL_FRAMESIZE; - err |= put_user(regs->gpr[1], (unsigned long __user *)newsp); - - /* Set up "regs" so we "return" to the signal handler. */ - err |= get_user(regs->nip, &funct_desc_ptr->entry); - regs->gpr[1] = newsp; - err |= get_user(regs->gpr[2], &funct_desc_ptr->toc); - regs->gpr[3] = signr; - regs->result = 0; - if (ka->sa.sa_flags & SA_SIGINFO) { - err |= get_user(regs->gpr[4], (unsigned long __user *)&frame->pinfo); - err |= get_user(regs->gpr[5], (unsigned long __user *)&frame->puc); - regs->gpr[6] = (unsigned long) frame; - } else { - regs->gpr[4] = (unsigned long)&frame->uc.uc_mcontext; - } - if (err) - goto badframe; - - if (test_thread_flag(TIF_SINGLESTEP)) - ptrace_notify(SIGTRAP); - - return 1; - -badframe: -#if DEBUG_SIG - printk("badframe in setup_rt_frame, regs=%p frame=%p newsp=%lx\n", - regs, frame, newsp); -#endif - force_sigsegv(signr, current); - return 0; -} - - -/* - * OK, we're invoking a handler - */ -static int handle_signal(unsigned long sig, struct k_sigaction *ka, - siginfo_t *info, sigset_t *oldset, struct pt_regs *regs) -{ - int ret; - - /* Set up Signal Frame */ - ret = setup_rt_frame(sig, ka, info, oldset, regs); - - if (ret) { - spin_lock_irq(¤t->sighand->siglock); - sigorsets(¤t->blocked, ¤t->blocked, &ka->sa.sa_mask); - if (!(ka->sa.sa_flags & SA_NODEFER)) - sigaddset(¤t->blocked,sig); - recalc_sigpending(); - spin_unlock_irq(¤t->sighand->siglock); - } - - return ret; -} - -static inline void syscall_restart(struct pt_regs *regs, struct k_sigaction *ka) -{ - switch ((int)regs->result) { - case -ERESTART_RESTARTBLOCK: - case -ERESTARTNOHAND: - /* ERESTARTNOHAND means that the syscall should only be - * restarted if there was no handler for the signal, and since - * we only get here if there is a handler, we dont restart. - */ - regs->result = -EINTR; - break; - case -ERESTARTSYS: - /* ERESTARTSYS means to restart the syscall if there is no - * handler or the handler was registered with SA_RESTART - */ - if (!(ka->sa.sa_flags & SA_RESTART)) { - regs->result = -EINTR; - break; - } - /* fallthrough */ - case -ERESTARTNOINTR: - /* ERESTARTNOINTR means that the syscall should be - * called again after the signal handler returns. - */ - regs->gpr[3] = regs->orig_gpr3; - regs->nip -= 4; - regs->result = 0; - break; - } -} - -/* - * Note that 'init' is a special process: it doesn't get signals it doesn't - * want to handle. Thus you cannot kill init even with a SIGKILL even by - * mistake. - */ -int do_signal(sigset_t *oldset, struct pt_regs *regs) -{ - siginfo_t info; - int signr; - struct k_sigaction ka; - - /* - * If the current thread is 32 bit - invoke the - * 32 bit signal handling code - */ - if (test_thread_flag(TIF_32BIT)) - return do_signal32(oldset, regs); - - if (!oldset) - oldset = ¤t->blocked; - - signr = get_signal_to_deliver(&info, &ka, regs, NULL); - if (signr > 0) { - /* Whee! Actually deliver the signal. */ - if (TRAP(regs) == 0x0C00) - syscall_restart(regs, &ka); - - /* - * Reenable the DABR before delivering the signal to - * user space. The DABR will have been cleared if it - * triggered inside the kernel. - */ - if (current->thread.dabr) - set_dabr(current->thread.dabr); - - return handle_signal(signr, &ka, &info, oldset, regs); - } - - if (TRAP(regs) == 0x0C00) { /* System Call! */ - if ((int)regs->result == -ERESTARTNOHAND || - (int)regs->result == -ERESTARTSYS || - (int)regs->result == -ERESTARTNOINTR) { - regs->gpr[3] = regs->orig_gpr3; - regs->nip -= 4; /* Back up & retry system call */ - regs->result = 0; - } else if ((int)regs->result == -ERESTART_RESTARTBLOCK) { - regs->gpr[0] = __NR_restart_syscall; - regs->nip -= 4; - regs->result = 0; - } - } - - return 0; -} -EXPORT_SYMBOL(do_signal); -- cgit v1.2.3 From 8902e87f1fc72740a40686a31ca07e6c2cf1f0f2 Mon Sep 17 00:00:00 2001 From: John Rose Date: Wed, 2 Nov 2005 10:29:55 -0600 Subject: [PATCH] ppc64: Fix add notifier crashes The extraction of PCI stuff from struct device_node left some false assumptions in notifier code. As a result, dynamic add crashes when non-PCI nodes are added. This patch fixes these assumptions. Signed-off-by: John Rose Signed-off-by: Paul Mackerras --- arch/powerpc/platforms/pseries/iommu.c | 2 +- arch/ppc64/kernel/pci_dn.c | 5 +++-- 2 files changed, 4 insertions(+), 3 deletions(-) (limited to 'arch/ppc64') diff --git a/arch/powerpc/platforms/pseries/iommu.c b/arch/powerpc/platforms/pseries/iommu.c index 9e90d41131d8..a73faafaac8c 100644 --- a/arch/powerpc/platforms/pseries/iommu.c +++ b/arch/powerpc/platforms/pseries/iommu.c @@ -498,7 +498,7 @@ static int iommu_reconfig_notifier(struct notifier_block *nb, unsigned long acti switch (action) { case PSERIES_RECONFIG_REMOVE: - if (pci->iommu_table && + if (pci && pci->iommu_table && get_property(np, "ibm,dma-window", NULL)) iommu_free_table(np); break; diff --git a/arch/ppc64/kernel/pci_dn.c b/arch/ppc64/kernel/pci_dn.c index 493bbe43f5b4..1a443a7ada4c 100644 --- a/arch/ppc64/kernel/pci_dn.c +++ b/arch/ppc64/kernel/pci_dn.c @@ -181,13 +181,14 @@ EXPORT_SYMBOL(fetch_dev_dn); static int pci_dn_reconfig_notifier(struct notifier_block *nb, unsigned long action, void *node) { struct device_node *np = node; - struct pci_dn *pci; + struct pci_dn *pci = NULL; int err = NOTIFY_OK; switch (action) { case PSERIES_RECONFIG_ADD: pci = np->parent->data; - update_dn_pci_info(np, pci->phb); + if (pci) + update_dn_pci_info(np, pci->phb); break; default: err = NOTIFY_DONE; -- cgit v1.2.3 From b8f510219edc719d4c305918e16edc578bcfc16f Mon Sep 17 00:00:00 2001 From: Michael Ellerman Date: Fri, 4 Nov 2005 12:09:42 +1100 Subject: powerpc: Implement smp_release_cpus() in C not asm There's no reason for smp_release_cpus() to be asm, and most people can make more sense of C code. Add an extern declaration to smp.h and remove the custom one in machine_kexec.c Signed-off-by: Michael Ellerman Signed-off-by: Stephen Rothwell --- arch/powerpc/kernel/head_64.S | 18 ------------------ arch/powerpc/kernel/setup_64.c | 23 +++++++++++++++++++++-- arch/ppc64/kernel/head.S | 18 ------------------ arch/ppc64/kernel/machine_kexec.c | 1 - include/asm-ppc64/smp.h | 1 + 5 files changed, 22 insertions(+), 39 deletions(-) (limited to 'arch/ppc64') diff --git a/arch/powerpc/kernel/head_64.S b/arch/powerpc/kernel/head_64.S index ba30e3177b05..45d81976987f 100644 --- a/arch/powerpc/kernel/head_64.S +++ b/arch/powerpc/kernel/head_64.S @@ -1914,24 +1914,6 @@ _GLOBAL(hmt_start_secondary) blr #endif -#if defined(CONFIG_KEXEC) || defined(CONFIG_SMP) -_GLOBAL(smp_release_cpus) - /* All secondary cpus are spinning on a common - * spinloop, release them all now so they can start - * to spin on their individual paca spinloops. - * For non SMP kernels, the secondary cpus never - * get out of the common spinloop. - * XXX This does nothing useful on iSeries, secondaries are - * already waiting on their paca. - */ - li r3,1 - LOADADDR(r5,__secondary_hold_spinloop) - std r3,0(r5) - sync - blr -#endif /* CONFIG_SMP */ - - /* * We put a few things here that have to be page-aligned. * This stuff goes at the beginning of the bss, which is page-aligned. diff --git a/arch/powerpc/kernel/setup_64.c b/arch/powerpc/kernel/setup_64.c index d4a3c5dd1a21..275d86ddd612 100644 --- a/arch/powerpc/kernel/setup_64.c +++ b/arch/powerpc/kernel/setup_64.c @@ -103,8 +103,6 @@ extern void htab_initialize(void); extern void early_init_devtree(void *flat_dt); extern void unflatten_device_tree(void); -extern void smp_release_cpus(void); - int have_of = 1; int boot_cpuid = 0; int boot_cpuid_phys = 0; @@ -400,6 +398,27 @@ void __init early_setup(unsigned long dt_ptr) } +#if defined(CONFIG_SMP) || defined(CONFIG_KEXEC) +void smp_release_cpus(void) +{ + extern unsigned long __secondary_hold_spinloop; + + DBG(" -> smp_release_cpus()\n"); + + /* All secondary cpus are spinning on a common spinloop, release them + * all now so they can start to spin on their individual paca + * spinloops. For non SMP kernels, the secondary cpus never get out + * of the common spinloop. + * This is useless but harmless on iSeries, secondaries are already + * waiting on their paca spinloops. */ + + __secondary_hold_spinloop = 1; + mb(); + + DBG(" <- smp_release_cpus()\n"); +} +#endif /* CONFIG_SMP || CONFIG_KEXEC */ + /* * Initialize some remaining members of the ppc64_caches and systemcfg structures * (at least until we get rid of them completely). This is mostly some diff --git a/arch/ppc64/kernel/head.S b/arch/ppc64/kernel/head.S index 0b65efe1682e..db1cf397be2d 100644 --- a/arch/ppc64/kernel/head.S +++ b/arch/ppc64/kernel/head.S @@ -1914,24 +1914,6 @@ _GLOBAL(hmt_start_secondary) blr #endif -#if defined(CONFIG_KEXEC) || defined(CONFIG_SMP) -_GLOBAL(smp_release_cpus) - /* All secondary cpus are spinning on a common - * spinloop, release them all now so they can start - * to spin on their individual paca spinloops. - * For non SMP kernels, the secondary cpus never - * get out of the common spinloop. - * XXX This does nothing useful on iSeries, secondaries are - * already waiting on their paca. - */ - li r3,1 - LOADADDR(r5,__secondary_hold_spinloop) - std r3,0(r5) - sync - blr -#endif /* CONFIG_SMP */ - - /* * We put a few things here that have to be page-aligned. * This stuff goes at the beginning of the bss, which is page-aligned. diff --git a/arch/ppc64/kernel/machine_kexec.c b/arch/ppc64/kernel/machine_kexec.c index bf7cc4f8210f..ff8679f260f3 100644 --- a/arch/ppc64/kernel/machine_kexec.c +++ b/arch/ppc64/kernel/machine_kexec.c @@ -244,7 +244,6 @@ static void kexec_prepare_cpus(void) static void kexec_prepare_cpus(void) { - extern void smp_release_cpus(void); /* * move the secondarys to us so that we can copy * the new kernel 0-0x100 safely diff --git a/include/asm-ppc64/smp.h b/include/asm-ppc64/smp.h index 46c04046952d..ba0f5c8bbb22 100644 --- a/include/asm-ppc64/smp.h +++ b/include/asm-ppc64/smp.h @@ -79,6 +79,7 @@ extern int smt_enabled_at_boot; extern int smp_mpic_probe(void); extern void smp_mpic_setup_cpu(int cpu); extern void smp_generic_kick_cpu(int nr); +extern void smp_release_cpus(void); extern void smp_generic_give_timebase(void); extern void smp_generic_take_timebase(void); -- cgit v1.2.3 From 8ad200d7b7c8fac77cf705831e90e889360d7030 Mon Sep 17 00:00:00 2001 From: Paul Mackerras Date: Fri, 4 Nov 2005 13:28:58 +1100 Subject: powerpc: Merge smp-tbsync.c (the generic timebase sync routine) Signed-off-by: Paul Mackerras --- arch/powerpc/Kconfig | 7 ++ arch/powerpc/kernel/Makefile | 1 + arch/powerpc/kernel/smp-tbsync.c | 171 +++++++++++++++++++++++++++++++++++++ arch/ppc64/Kconfig | 5 ++ arch/ppc64/kernel/Makefile | 5 -- arch/ppc64/kernel/smp-tbsync.c | 179 --------------------------------------- 6 files changed, 184 insertions(+), 184 deletions(-) create mode 100644 arch/powerpc/kernel/smp-tbsync.c delete mode 100644 arch/ppc64/kernel/smp-tbsync.c (limited to 'arch/ppc64') diff --git a/arch/powerpc/Kconfig b/arch/powerpc/Kconfig index 3cf03ab46113..f4e25c648fbb 100644 --- a/arch/powerpc/Kconfig +++ b/arch/powerpc/Kconfig @@ -300,6 +300,7 @@ config PPC_PMAC64 bool depends on PPC_PMAC && POWER4 select U3_DART + select GENERIC_TBSYNC default y config PPC_PREP @@ -314,6 +315,7 @@ config PPC_MAPLE bool " Maple 970FX Evaluation Board" select U3_DART select MPIC_BROKEN_U3 + select GENERIC_TBSYNC default n help This option enables support for the Maple 970FX Evaluation Board. @@ -386,6 +388,11 @@ config PPC_MPC106 bool default n +config GENERIC_TBSYNC + bool + default y if CONFIG_PPC32 && CONFIG_SMP + default n + source "drivers/cpufreq/Kconfig" config CPU_FREQ_PMAC diff --git a/arch/powerpc/kernel/Makefile b/arch/powerpc/kernel/Makefile index 7a3e1155ac9a..631149ea93db 100644 --- a/arch/powerpc/kernel/Makefile +++ b/arch/powerpc/kernel/Makefile @@ -21,6 +21,7 @@ obj-$(CONFIG_PPC_RTAS) += rtas.o obj-$(CONFIG_RTAS_FLASH) += rtas_flash.o obj-$(CONFIG_RTAS_PROC) += rtas-proc.o obj-$(CONFIG_IBMVIO) += vio.o +obj-$(CONFIG_GENERIC_TBSYNC) += smp-tbsync.o ifeq ($(CONFIG_PPC_MERGE),y) diff --git a/arch/powerpc/kernel/smp-tbsync.c b/arch/powerpc/kernel/smp-tbsync.c new file mode 100644 index 000000000000..9adef3bddad4 --- /dev/null +++ b/arch/powerpc/kernel/smp-tbsync.c @@ -0,0 +1,171 @@ +/* + * Smp timebase synchronization for ppc. + * + * Copyright (C) 2003 Samuel Rydh (samuel@ibrium.se) + * + */ + +#include +#include +#include +#include +#include +#include +#include +#include +#include + +#define NUM_ITER 300 + +enum { + kExit=0, kSetAndTest, kTest +}; + +static struct { + volatile u64 tb; + volatile u64 mark; + volatile int cmd; + volatile int handshake; + int filler[2]; + + volatile int ack; + int filler2[7]; + + volatile int race_result; +} *tbsync; + +static volatile int running; + +static void __devinit enter_contest(u64 mark, long add) +{ + while (get_tb() < mark) + tbsync->race_result = add; +} + +void __devinit smp_generic_take_timebase(void) +{ + int cmd; + u64 tb; + + local_irq_disable(); + while (!running) + barrier(); + rmb(); + + for (;;) { + tbsync->ack = 1; + while (!tbsync->handshake) + barrier(); + rmb(); + + cmd = tbsync->cmd; + tb = tbsync->tb; + mb(); + tbsync->ack = 0; + if (cmd == kExit) + break; + + while (tbsync->handshake) + barrier(); + if (cmd == kSetAndTest) + set_tb(tb >> 32, tb & 0xfffffffful); + enter_contest(tbsync->mark, -1); + } + local_irq_enable(); +} + +static int __devinit start_contest(int cmd, long offset, int num) +{ + int i, score=0; + u64 tb; + long mark; + + tbsync->cmd = cmd; + + local_irq_disable(); + for (i = -3; i < num; ) { + tb = get_tb() + 400; + tbsync->tb = tb + offset; + tbsync->mark = mark = tb + 400; + + wmb(); + + tbsync->handshake = 1; + while (tbsync->ack) + barrier(); + + while (get_tb() <= tb) + barrier(); + tbsync->handshake = 0; + enter_contest(mark, 1); + + while (!tbsync->ack) + barrier(); + + if (i++ > 0) + score += tbsync->race_result; + } + local_irq_enable(); + return score; +} + +void __devinit smp_generic_give_timebase(void) +{ + int i, score, score2, old, min=0, max=5000, offset=1000; + + printk("Synchronizing timebase\n"); + + /* if this fails then this kernel won't work anyway... */ + tbsync = kmalloc( sizeof(*tbsync), GFP_KERNEL ); + memset( tbsync, 0, sizeof(*tbsync) ); + mb(); + running = 1; + + while (!tbsync->ack) + barrier(); + + printk("Got ack\n"); + + /* binary search */ + for (old = -1; old != offset ; offset = (min+max) / 2) { + score = start_contest(kSetAndTest, offset, NUM_ITER); + + printk("score %d, offset %d\n", score, offset ); + + if( score > 0 ) + max = offset; + else + min = offset; + old = offset; + } + score = start_contest(kSetAndTest, min, NUM_ITER); + score2 = start_contest(kSetAndTest, max, NUM_ITER); + + printk("Min %d (score %d), Max %d (score %d)\n", + min, score, max, score2); + score = abs(score); + score2 = abs(score2); + offset = (score < score2) ? min : max; + + /* guard against inaccurate mttb */ + for (i = 0; i < 10; i++) { + start_contest(kSetAndTest, offset, NUM_ITER/10); + + if ((score2 = start_contest(kTest, offset, NUM_ITER)) < 0) + score2 = -score2; + if (score2 <= score || score2 < 20) + break; + } + printk("Final offset: %d (%d/%d)\n", offset, score2, NUM_ITER ); + + /* exiting */ + tbsync->cmd = kExit; + wmb(); + tbsync->handshake = 1; + while (tbsync->ack) + barrier(); + tbsync->handshake = 0; + kfree(tbsync); + tbsync = NULL; + running = 0; +} diff --git a/arch/ppc64/Kconfig b/arch/ppc64/Kconfig index 42677cc96508..b987164fca4c 100644 --- a/arch/ppc64/Kconfig +++ b/arch/ppc64/Kconfig @@ -89,12 +89,14 @@ config PPC_PMAC bool " Apple G5 based machines" default y select U3_DART + select GENERIC_TBSYNC config PPC_MAPLE depends on PPC_MULTIPLATFORM bool " Maple 970FX Evaluation Board" select U3_DART select MPIC_BROKEN_U3 + select GENERIC_TBSYNC default n help This option enables support for the Maple 970FX Evaluation Board. @@ -182,6 +184,9 @@ config MPIC_BROKEN_U3 depends on PPC_MAPLE default y +config GENERIC_TBSYNC + def_bool n + config PPC_PMAC64 bool depends on PPC_PMAC diff --git a/arch/ppc64/kernel/Makefile b/arch/ppc64/kernel/Makefile index 430cb3900a07..f597c2954b71 100644 --- a/arch/ppc64/kernel/Makefile +++ b/arch/ppc64/kernel/Makefile @@ -51,11 +51,6 @@ obj-$(CONFIG_PPC_PMAC) += udbg_scc.o obj-$(CONFIG_PPC_MAPLE) += udbg_16550.o -ifdef CONFIG_SMP -obj-$(CONFIG_PPC_PMAC) += smp-tbsync.o -obj-$(CONFIG_PPC_MAPLE) += smp-tbsync.o -endif - obj-$(CONFIG_KPROBES) += kprobes.o CFLAGS_ioctl32.o += -Ifs/ diff --git a/arch/ppc64/kernel/smp-tbsync.c b/arch/ppc64/kernel/smp-tbsync.c deleted file mode 100644 index 7d8ec9996b3e..000000000000 --- a/arch/ppc64/kernel/smp-tbsync.c +++ /dev/null @@ -1,179 +0,0 @@ -/* - * Smp timebase synchronization for ppc. - * - * Copyright (C) 2003 Samuel Rydh (samuel@ibrium.se) - * - */ - -#include -#include -#include -#include -#include -#include -#include -#include -#include - -#define NUM_ITER 300 - -enum { - kExit=0, kSetAndTest, kTest -}; - -static struct { - volatile long tb; - volatile long mark; - volatile int cmd; - volatile int handshake; - int filler[3]; - - volatile int ack; - int filler2[7]; - - volatile int race_result; -} *tbsync; - -static volatile int running; - -static void __devinit -enter_contest( long mark, long add ) -{ - while( (long)(mftb() - mark) < 0 ) - tbsync->race_result = add; -} - -void __devinit -smp_generic_take_timebase( void ) -{ - int cmd; - long tb; - - local_irq_disable(); - while( !running ) - ; - rmb(); - - for( ;; ) { - tbsync->ack = 1; - while( !tbsync->handshake ) - ; - rmb(); - - cmd = tbsync->cmd; - tb = tbsync->tb; - tbsync->ack = 0; - if( cmd == kExit ) - return; - - if( cmd == kSetAndTest ) { - while( tbsync->handshake ) - ; - asm volatile ("mttbl %0" :: "r" (tb & 0xfffffffful) ); - asm volatile ("mttbu %0" :: "r" (tb >> 32) ); - } else { - while( tbsync->handshake ) - ; - } - enter_contest( tbsync->mark, -1 ); - } - local_irq_enable(); -} - -static int __devinit -start_contest( int cmd, long offset, long num ) -{ - int i, score=0; - long tb, mark; - - tbsync->cmd = cmd; - - local_irq_disable(); - for( i=-3; itb = tb + offset; - tbsync->mark = mark = tb + 400; - - wmb(); - - tbsync->handshake = 1; - while( tbsync->ack ) - ; - - while( (long)(mftb() - tb) <= 0 ) - ; - tbsync->handshake = 0; - enter_contest( mark, 1 ); - - while( !tbsync->ack ) - ; - - if ((tbsync->tb ^ (long)mftb()) & 0x8000000000000000ul) - continue; - if( i++ > 0 ) - score += tbsync->race_result; - } - local_irq_enable(); - return score; -} - -void __devinit -smp_generic_give_timebase( void ) -{ - int i, score, score2, old, min=0, max=5000, offset=1000; - - printk("Synchronizing timebase\n"); - - /* if this fails then this kernel won't work anyway... */ - tbsync = kmalloc( sizeof(*tbsync), GFP_KERNEL ); - memset( tbsync, 0, sizeof(*tbsync) ); - mb(); - running = 1; - - while( !tbsync->ack ) - ; - - printk("Got ack\n"); - - /* binary search */ - for( old=-1 ; old != offset ; offset=(min+max)/2 ) { - score = start_contest( kSetAndTest, offset, NUM_ITER ); - - printk("score %d, offset %d\n", score, offset ); - - if( score > 0 ) - max = offset; - else - min = offset; - old = offset; - } - score = start_contest( kSetAndTest, min, NUM_ITER ); - score2 = start_contest( kSetAndTest, max, NUM_ITER ); - - printk( "Min %d (score %d), Max %d (score %d)\n", min, score, max, score2 ); - score = abs( score ); - score2 = abs( score2 ); - offset = (score < score2) ? min : max; - - /* guard against inaccurate mttb */ - for( i=0; i<10; i++ ) { - start_contest( kSetAndTest, offset, NUM_ITER/10 ); - - if( (score2=start_contest(kTest, offset, NUM_ITER)) < 0 ) - score2 = -score2; - if( score2 <= score || score2 < 20 ) - break; - } - printk("Final offset: %d (%d/%d)\n", offset, score2, NUM_ITER ); - - /* exiting */ - tbsync->cmd = kExit; - wmb(); - tbsync->handshake = 1; - while( tbsync->ack ) - ; - tbsync->handshake = 0; - kfree( tbsync ); - tbsync = NULL; - running = 0; -} -- cgit v1.2.3 From 5ad570786158e327a1c5d32dd3d66f26d8de6340 Mon Sep 17 00:00:00 2001 From: Paul Mackerras Date: Sat, 5 Nov 2005 10:33:55 +1100 Subject: powerpc: Merge smp.c and smp.h This also moves setup_cpu_maps to setup-common.c (calling it smp_setup_cpu_maps) and uses it on both 32-bit and 64-bit. Signed-off-by: Paul Mackerras --- arch/powerpc/kernel/Makefile | 4 +- arch/powerpc/kernel/setup-common.c | 119 +++++++- arch/powerpc/kernel/setup_32.c | 2 + arch/powerpc/kernel/setup_64.c | 120 +------- arch/powerpc/kernel/smp.c | 565 ++++++++++++++++++++++++++++++++++++ arch/powerpc/platforms/chrp/smp.c | 39 +-- arch/ppc/kernel/Makefile | 1 - arch/ppc/kernel/irq.c | 2 +- arch/ppc64/kernel/Makefile | 1 - arch/ppc64/kernel/smp.c | 568 ------------------------------------- include/asm-powerpc/smp.h | 119 ++++++++ include/asm-ppc64/smp.h | 92 ------ 12 files changed, 812 insertions(+), 820 deletions(-) create mode 100644 arch/powerpc/kernel/smp.c delete mode 100644 arch/ppc64/kernel/smp.c create mode 100644 include/asm-powerpc/smp.h delete mode 100644 include/asm-ppc64/smp.h (limited to 'arch/ppc64') diff --git a/arch/powerpc/kernel/Makefile b/arch/powerpc/kernel/Makefile index 631149ea93db..b3ae2993efb8 100644 --- a/arch/powerpc/kernel/Makefile +++ b/arch/powerpc/kernel/Makefile @@ -41,6 +41,7 @@ obj-$(CONFIG_PPC_OF) += prom_init.o obj-$(CONFIG_MODULES) += ppc_ksyms.o obj-$(CONFIG_BOOTX_TEXT) += btext.o obj-$(CONFIG_6xx) += idle_6xx.o +obj-$(CONFIG_SMP) += smp.o ifeq ($(CONFIG_PPC_ISERIES),y) $(obj)/head_64.o: $(obj)/lparmap.s @@ -49,8 +50,9 @@ endif else # stuff used from here for ARCH=ppc or ARCH=ppc64 +smpobj-$(CONFIG_SMP) += smp.o obj-$(CONFIG_PPC64) += traps.o process.o init_task.o time.o \ - setup-common.o + setup-common.o $(smpobj-y) endif diff --git a/arch/powerpc/kernel/setup-common.c b/arch/powerpc/kernel/setup-common.c index 14ebe3bc48c3..d43fa8c0e5ac 100644 --- a/arch/powerpc/kernel/setup-common.c +++ b/arch/powerpc/kernel/setup-common.c @@ -170,11 +170,7 @@ static int show_cpuinfo(struct seq_file *m, void *v) } #ifdef CONFIG_SMP -#ifdef CONFIG_PPC64 /* XXX for now */ pvr = per_cpu(pvr, cpu_id); -#else - pvr = cpu_data[cpu_id].pvr; -#endif #else pvr = mfspr(SPRN_PVR); #endif @@ -408,3 +404,118 @@ static int __init set_preferred_console(void) } console_initcall(set_preferred_console); #endif /* CONFIG_PPC_MULTIPLATFORM */ + +#ifdef CONFIG_SMP + +/** + * setup_cpu_maps - initialize the following cpu maps: + * cpu_possible_map + * cpu_present_map + * cpu_sibling_map + * + * Having the possible map set up early allows us to restrict allocations + * of things like irqstacks to num_possible_cpus() rather than NR_CPUS. + * + * We do not initialize the online map here; cpus set their own bits in + * cpu_online_map as they come up. + * + * This function is valid only for Open Firmware systems. finish_device_tree + * must be called before using this. + * + * While we're here, we may as well set the "physical" cpu ids in the paca. + */ +void __init smp_setup_cpu_maps(void) +{ + struct device_node *dn = NULL; + int cpu = 0; + int swap_cpuid = 0; + + while ((dn = of_find_node_by_type(dn, "cpu")) && cpu < NR_CPUS) { + int *intserv; + int j, len = sizeof(u32), nthreads = 1; + + intserv = (int *)get_property(dn, "ibm,ppc-interrupt-server#s", + &len); + if (intserv) + nthreads = len / sizeof(int); + else { + intserv = (int *) get_property(dn, "reg", NULL); + if (!intserv) + intserv = &cpu; /* assume logical == phys */ + } + + for (j = 0; j < nthreads && cpu < NR_CPUS; j++) { + cpu_set(cpu, cpu_present_map); + set_hard_smp_processor_id(cpu, intserv[j]); + + if (intserv[j] == boot_cpuid_phys) + swap_cpuid = cpu; + cpu_set(cpu, cpu_possible_map); + cpu++; + } + } + + /* Swap CPU id 0 with boot_cpuid_phys, so we can always assume that + * boot cpu is logical 0. + */ + if (boot_cpuid_phys != get_hard_smp_processor_id(0)) { + u32 tmp; + tmp = get_hard_smp_processor_id(0); + set_hard_smp_processor_id(0, boot_cpuid_phys); + set_hard_smp_processor_id(swap_cpuid, tmp); + } + +#ifdef CONFIG_PPC64 + /* + * On pSeries LPAR, we need to know how many cpus + * could possibly be added to this partition. + */ + if (systemcfg->platform == PLATFORM_PSERIES_LPAR && + (dn = of_find_node_by_path("/rtas"))) { + int num_addr_cell, num_size_cell, maxcpus; + unsigned int *ireg; + + num_addr_cell = prom_n_addr_cells(dn); + num_size_cell = prom_n_size_cells(dn); + + ireg = (unsigned int *) + get_property(dn, "ibm,lrdr-capacity", NULL); + + if (!ireg) + goto out; + + maxcpus = ireg[num_addr_cell + num_size_cell]; + + /* Double maxcpus for processors which have SMT capability */ + if (cpu_has_feature(CPU_FTR_SMT)) + maxcpus *= 2; + + if (maxcpus > NR_CPUS) { + printk(KERN_WARNING + "Partition configured for %d cpus, " + "operating system maximum is %d.\n", + maxcpus, NR_CPUS); + maxcpus = NR_CPUS; + } else + printk(KERN_INFO "Partition configured for %d cpus.\n", + maxcpus); + + for (cpu = 0; cpu < maxcpus; cpu++) + cpu_set(cpu, cpu_possible_map); + out: + of_node_put(dn); + } + + /* + * Do the sibling map; assume only two threads per processor. + */ + for_each_cpu(cpu) { + cpu_set(cpu, cpu_sibling_map[cpu]); + if (cpu_has_feature(CPU_FTR_SMT)) + cpu_set(cpu ^ 0x1, cpu_sibling_map[cpu]); + } + + systemcfg->processorCount = num_present_cpus(); +#endif /* CONFIG_PPC64 */ +} +#endif /* CONFIG_SMP */ diff --git a/arch/powerpc/kernel/setup_32.c b/arch/powerpc/kernel/setup_32.c index 9680ae99b084..b45eedbb4b3a 100644 --- a/arch/powerpc/kernel/setup_32.c +++ b/arch/powerpc/kernel/setup_32.c @@ -288,6 +288,8 @@ void __init setup_arch(char **cmdline_p) unflatten_device_tree(); finish_device_tree(); + smp_setup_cpu_maps(); + #ifdef CONFIG_BOOTX_TEXT init_boot_display(); #endif diff --git a/arch/powerpc/kernel/setup_64.c b/arch/powerpc/kernel/setup_64.c index 275d86ddd612..6b52cce872be 100644 --- a/arch/powerpc/kernel/setup_64.c +++ b/arch/powerpc/kernel/setup_64.c @@ -181,114 +181,8 @@ static int __init early_smt_enabled(char *p) } early_param("smt-enabled", early_smt_enabled); -/** - * setup_cpu_maps - initialize the following cpu maps: - * cpu_possible_map - * cpu_present_map - * cpu_sibling_map - * - * Having the possible map set up early allows us to restrict allocations - * of things like irqstacks to num_possible_cpus() rather than NR_CPUS. - * - * We do not initialize the online map here; cpus set their own bits in - * cpu_online_map as they come up. - * - * This function is valid only for Open Firmware systems. finish_device_tree - * must be called before using this. - * - * While we're here, we may as well set the "physical" cpu ids in the paca. - */ -static void __init setup_cpu_maps(void) -{ - struct device_node *dn = NULL; - int cpu = 0; - int swap_cpuid = 0; - - check_smt_enabled(); - - while ((dn = of_find_node_by_type(dn, "cpu")) && cpu < NR_CPUS) { - u32 *intserv; - int j, len = sizeof(u32), nthreads; - - intserv = (u32 *)get_property(dn, "ibm,ppc-interrupt-server#s", - &len); - if (!intserv) - intserv = (u32 *)get_property(dn, "reg", NULL); - - nthreads = len / sizeof(u32); - - for (j = 0; j < nthreads && cpu < NR_CPUS; j++) { - cpu_set(cpu, cpu_present_map); - set_hard_smp_processor_id(cpu, intserv[j]); - - if (intserv[j] == boot_cpuid_phys) - swap_cpuid = cpu; - cpu_set(cpu, cpu_possible_map); - cpu++; - } - } - - /* Swap CPU id 0 with boot_cpuid_phys, so we can always assume that - * boot cpu is logical 0. - */ - if (boot_cpuid_phys != get_hard_smp_processor_id(0)) { - u32 tmp; - tmp = get_hard_smp_processor_id(0); - set_hard_smp_processor_id(0, boot_cpuid_phys); - set_hard_smp_processor_id(swap_cpuid, tmp); - } - - /* - * On pSeries LPAR, we need to know how many cpus - * could possibly be added to this partition. - */ - if (systemcfg->platform == PLATFORM_PSERIES_LPAR && - (dn = of_find_node_by_path("/rtas"))) { - int num_addr_cell, num_size_cell, maxcpus; - unsigned int *ireg; - - num_addr_cell = prom_n_addr_cells(dn); - num_size_cell = prom_n_size_cells(dn); - - ireg = (unsigned int *) - get_property(dn, "ibm,lrdr-capacity", NULL); - - if (!ireg) - goto out; - - maxcpus = ireg[num_addr_cell + num_size_cell]; - - /* Double maxcpus for processors which have SMT capability */ - if (cpu_has_feature(CPU_FTR_SMT)) - maxcpus *= 2; - - if (maxcpus > NR_CPUS) { - printk(KERN_WARNING - "Partition configured for %d cpus, " - "operating system maximum is %d.\n", - maxcpus, NR_CPUS); - maxcpus = NR_CPUS; - } else - printk(KERN_INFO "Partition configured for %d cpus.\n", - maxcpus); - - for (cpu = 0; cpu < maxcpus; cpu++) - cpu_set(cpu, cpu_possible_map); - out: - of_node_put(dn); - } - - /* - * Do the sibling map; assume only two threads per processor. - */ - for_each_cpu(cpu) { - cpu_set(cpu, cpu_sibling_map[cpu]); - if (cpu_has_feature(CPU_FTR_SMT)) - cpu_set(cpu ^ 0x1, cpu_sibling_map[cpu]); - } - - systemcfg->processorCount = num_present_cpus(); -} +#else +#define check_smt_enabled() #endif /* CONFIG_SMP */ extern struct machdep_calls pSeries_md; @@ -417,6 +311,8 @@ void smp_release_cpus(void) DBG(" <- smp_release_cpus()\n"); } +#else +#define smp_release_cpus() #endif /* CONFIG_SMP || CONFIG_KEXEC */ /* @@ -608,17 +504,13 @@ void __init setup_system(void) parse_early_param(); -#ifdef CONFIG_SMP - /* - * iSeries has already initialized the cpu maps at this point. - */ - setup_cpu_maps(); + check_smt_enabled(); + smp_setup_cpu_maps(); /* Release secondary cpus out of their spinloops at 0x60 now that * we can map physical -> logical CPU ids */ smp_release_cpus(); -#endif printk("Starting Linux PPC64 %s\n", system_utsname.version); diff --git a/arch/powerpc/kernel/smp.c b/arch/powerpc/kernel/smp.c new file mode 100644 index 000000000000..1794a694a928 --- /dev/null +++ b/arch/powerpc/kernel/smp.c @@ -0,0 +1,565 @@ +/* + * SMP support for ppc. + * + * Written by Cort Dougan (cort@cs.nmt.edu) borrowing a great + * deal of code from the sparc and intel versions. + * + * Copyright (C) 1999 Cort Dougan + * + * PowerPC-64 Support added by Dave Engebretsen, Peter Bergner, and + * Mike Corrigan {engebret|bergner|mikec}@us.ibm.com + * + * This program is free software; you can redistribute it and/or + * modify it under the terms of the GNU General Public License + * as published by the Free Software Foundation; either version + * 2 of the License, or (at your option) any later version. + */ + +#undef DEBUG + +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include + +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#ifdef CONFIG_PPC64 +#include +#endif + +int smp_hw_index[NR_CPUS]; +struct thread_info *secondary_ti; + +#ifdef DEBUG +#define DBG(fmt...) udbg_printf(fmt) +#else +#define DBG(fmt...) +#endif + +cpumask_t cpu_possible_map = CPU_MASK_NONE; +cpumask_t cpu_online_map = CPU_MASK_NONE; +cpumask_t cpu_sibling_map[NR_CPUS] = { [0 ... NR_CPUS-1] = CPU_MASK_NONE }; + +EXPORT_SYMBOL(cpu_online_map); +EXPORT_SYMBOL(cpu_possible_map); + +/* SMP operations for this machine */ +struct smp_ops_t *smp_ops; + +static volatile unsigned int cpu_callin_map[NR_CPUS]; + +void smp_call_function_interrupt(void); + +int smt_enabled_at_boot = 1; + +#ifdef CONFIG_MPIC +int __init smp_mpic_probe(void) +{ + int nr_cpus; + + DBG("smp_mpic_probe()...\n"); + + nr_cpus = cpus_weight(cpu_possible_map); + + DBG("nr_cpus: %d\n", nr_cpus); + + if (nr_cpus > 1) + mpic_request_ipis(); + + return nr_cpus; +} + +void __devinit smp_mpic_setup_cpu(int cpu) +{ + mpic_setup_this_cpu(); +} +#endif /* CONFIG_MPIC */ + +#ifdef CONFIG_PPC64 +void __devinit smp_generic_kick_cpu(int nr) +{ + BUG_ON(nr < 0 || nr >= NR_CPUS); + + /* + * The processor is currently spinning, waiting for the + * cpu_start field to become non-zero After we set cpu_start, + * the processor will continue on to secondary_start + */ + paca[nr].cpu_start = 1; + smp_mb(); +} +#endif + +void smp_message_recv(int msg, struct pt_regs *regs) +{ + switch(msg) { + case PPC_MSG_CALL_FUNCTION: + smp_call_function_interrupt(); + break; + case PPC_MSG_RESCHEDULE: + /* XXX Do we have to do this? */ + set_need_resched(); + break; +#ifdef CONFIG_DEBUGGER + case PPC_MSG_DEBUGGER_BREAK: + debugger_ipi(regs); + break; +#endif + default: + printk("SMP %d: smp_message_recv(): unknown msg %d\n", + smp_processor_id(), msg); + break; + } +} + +void smp_send_reschedule(int cpu) +{ + smp_ops->message_pass(cpu, PPC_MSG_RESCHEDULE); +} + +#ifdef CONFIG_DEBUGGER +void smp_send_debugger_break(int cpu) +{ + smp_ops->message_pass(cpu, PPC_MSG_DEBUGGER_BREAK); +} +#endif + +static void stop_this_cpu(void *dummy) +{ + local_irq_disable(); + while (1) + ; +} + +void smp_send_stop(void) +{ + smp_call_function(stop_this_cpu, NULL, 1, 0); +} + +/* + * Structure and data for smp_call_function(). This is designed to minimise + * static memory requirements. It also looks cleaner. + * Stolen from the i386 version. + */ +static __cacheline_aligned_in_smp DEFINE_SPINLOCK(call_lock); + +static struct call_data_struct { + void (*func) (void *info); + void *info; + atomic_t started; + atomic_t finished; + int wait; +} *call_data; + +/* delay of at least 8 seconds */ +#define SMP_CALL_TIMEOUT 8 + +/* + * This function sends a 'generic call function' IPI to all other CPUs + * in the system. + * + * [SUMMARY] Run a function on all other CPUs. + * The function to run. This must be fast and non-blocking. + * An arbitrary pointer to pass to the function. + * currently unused. + * If true, wait (atomically) until function has completed on other CPUs. + * [RETURNS] 0 on success, else a negative status code. Does not return until + * remote CPUs are nearly ready to execute <> or are or have executed. + * + * You must not call this function with disabled interrupts or from a + * hardware interrupt handler or from a bottom half handler. + */ +int smp_call_function (void (*func) (void *info), void *info, int nonatomic, + int wait) +{ + struct call_data_struct data; + int ret = -1, cpus; + u64 timeout; + + /* Can deadlock when called with interrupts disabled */ + WARN_ON(irqs_disabled()); + + data.func = func; + data.info = info; + atomic_set(&data.started, 0); + data.wait = wait; + if (wait) + atomic_set(&data.finished, 0); + + spin_lock(&call_lock); + /* Must grab online cpu count with preempt disabled, otherwise + * it can change. */ + cpus = num_online_cpus() - 1; + if (!cpus) { + ret = 0; + goto out; + } + + call_data = &data; + smp_wmb(); + /* Send a message to all other CPUs and wait for them to respond */ + smp_ops->message_pass(MSG_ALL_BUT_SELF, PPC_MSG_CALL_FUNCTION); + + timeout = get_tb() + (u64) SMP_CALL_TIMEOUT * tb_ticks_per_sec; + + /* Wait for response */ + while (atomic_read(&data.started) != cpus) { + HMT_low(); + if (get_tb() >= timeout) { + printk("smp_call_function on cpu %d: other cpus not " + "responding (%d)\n", smp_processor_id(), + atomic_read(&data.started)); + debugger(NULL); + goto out; + } + } + + if (wait) { + while (atomic_read(&data.finished) != cpus) { + HMT_low(); + if (get_tb() >= timeout) { + printk("smp_call_function on cpu %d: other " + "cpus not finishing (%d/%d)\n", + smp_processor_id(), + atomic_read(&data.finished), + atomic_read(&data.started)); + debugger(NULL); + goto out; + } + } + } + + ret = 0; + + out: + call_data = NULL; + HMT_medium(); + spin_unlock(&call_lock); + return ret; +} + +EXPORT_SYMBOL(smp_call_function); + +void smp_call_function_interrupt(void) +{ + void (*func) (void *info); + void *info; + int wait; + + /* call_data will be NULL if the sender timed out while + * waiting on us to receive the call. + */ + if (!call_data) + return; + + func = call_data->func; + info = call_data->info; + wait = call_data->wait; + + if (!wait) + smp_mb__before_atomic_inc(); + + /* + * Notify initiating CPU that I've grabbed the data and am + * about to execute the function + */ + atomic_inc(&call_data->started); + /* + * At this point the info structure may be out of scope unless wait==1 + */ + (*func)(info); + if (wait) { + smp_mb__before_atomic_inc(); + atomic_inc(&call_data->finished); + } +} + +extern struct gettimeofday_struct do_gtod; + +struct thread_info *current_set[NR_CPUS]; + +DECLARE_PER_CPU(unsigned int, pvr); + +static void __devinit smp_store_cpu_info(int id) +{ + per_cpu(pvr, id) = mfspr(SPRN_PVR); +} + +static void __init smp_create_idle(unsigned int cpu) +{ + struct task_struct *p; + + /* create a process for the processor */ + p = fork_idle(cpu); + if (IS_ERR(p)) + panic("failed fork for CPU %u: %li", cpu, PTR_ERR(p)); +#ifdef CONFIG_PPC64 + paca[cpu].__current = p; +#endif + current_set[cpu] = p->thread_info; + p->thread_info->cpu = cpu; +} + +void __init smp_prepare_cpus(unsigned int max_cpus) +{ + unsigned int cpu; + + DBG("smp_prepare_cpus\n"); + + /* + * setup_cpu may need to be called on the boot cpu. We havent + * spun any cpus up but lets be paranoid. + */ + BUG_ON(boot_cpuid != smp_processor_id()); + + /* Fixup boot cpu */ + smp_store_cpu_info(boot_cpuid); + cpu_callin_map[boot_cpuid] = 1; + + max_cpus = smp_ops->probe(); + + smp_space_timers(max_cpus); + + for_each_cpu(cpu) + if (cpu != boot_cpuid) + smp_create_idle(cpu); +} + +void __devinit smp_prepare_boot_cpu(void) +{ + BUG_ON(smp_processor_id() != boot_cpuid); + + cpu_set(boot_cpuid, cpu_online_map); +#ifdef CONFIG_PPC64 + paca[boot_cpuid].__current = current; +#endif + current_set[boot_cpuid] = current->thread_info; +} + +#ifdef CONFIG_HOTPLUG_CPU +/* State of each CPU during hotplug phases */ +DEFINE_PER_CPU(int, cpu_state) = { 0 }; + +int generic_cpu_disable(void) +{ + unsigned int cpu = smp_processor_id(); + + if (cpu == boot_cpuid) + return -EBUSY; + + systemcfg->processorCount--; + cpu_clear(cpu, cpu_online_map); + fixup_irqs(cpu_online_map); + return 0; +} + +int generic_cpu_enable(unsigned int cpu) +{ + /* Do the normal bootup if we haven't + * already bootstrapped. */ + if (system_state != SYSTEM_RUNNING) + return -ENOSYS; + + /* get the target out of it's holding state */ + per_cpu(cpu_state, cpu) = CPU_UP_PREPARE; + smp_wmb(); + + while (!cpu_online(cpu)) + cpu_relax(); + + fixup_irqs(cpu_online_map); + /* counter the irq disable in fixup_irqs */ + local_irq_enable(); + return 0; +} + +void generic_cpu_die(unsigned int cpu) +{ + int i; + + for (i = 0; i < 100; i++) { + smp_rmb(); + if (per_cpu(cpu_state, cpu) == CPU_DEAD) + return; + msleep(100); + } + printk(KERN_ERR "CPU%d didn't die...\n", cpu); +} + +void generic_mach_cpu_die(void) +{ + unsigned int cpu; + + local_irq_disable(); + cpu = smp_processor_id(); + printk(KERN_DEBUG "CPU%d offline\n", cpu); + __get_cpu_var(cpu_state) = CPU_DEAD; + smp_wmb(); + while (__get_cpu_var(cpu_state) != CPU_UP_PREPARE) + cpu_relax(); + + flush_tlb_pending(); + cpu_set(cpu, cpu_online_map); + local_irq_enable(); +} +#endif + +static int __devinit cpu_enable(unsigned int cpu) +{ + if (smp_ops->cpu_enable) + return smp_ops->cpu_enable(cpu); + + return -ENOSYS; +} + +int __devinit __cpu_up(unsigned int cpu) +{ + int c; + + secondary_ti = current_set[cpu]; + if (!cpu_enable(cpu)) + return 0; + + if (smp_ops->cpu_bootable && !smp_ops->cpu_bootable(cpu)) + return -EINVAL; + +#ifdef CONFIG_PPC64 + paca[cpu].default_decr = tb_ticks_per_jiffy; +#endif + + /* Make sure callin-map entry is 0 (can be leftover a CPU + * hotplug + */ + cpu_callin_map[cpu] = 0; + + /* The information for processor bringup must + * be written out to main store before we release + * the processor. + */ + smp_mb(); + + /* wake up cpus */ + DBG("smp: kicking cpu %d\n", cpu); + smp_ops->kick_cpu(cpu); + + /* + * wait to see if the cpu made a callin (is actually up). + * use this value that I found through experimentation. + * -- Cort + */ + if (system_state < SYSTEM_RUNNING) + for (c = 5000; c && !cpu_callin_map[cpu]; c--) + udelay(100); +#ifdef CONFIG_HOTPLUG_CPU + else + /* + * CPUs can take much longer to come up in the + * hotplug case. Wait five seconds. + */ + for (c = 25; c && !cpu_callin_map[cpu]; c--) { + msleep(200); + } +#endif + + if (!cpu_callin_map[cpu]) { + printk("Processor %u is stuck.\n", cpu); + return -ENOENT; + } + + printk("Processor %u found.\n", cpu); + + if (smp_ops->give_timebase) + smp_ops->give_timebase(); + + /* Wait until cpu puts itself in the online map */ + while (!cpu_online(cpu)) + cpu_relax(); + + return 0; +} + + +/* Activate a secondary processor. */ +int __devinit start_secondary(void *unused) +{ + unsigned int cpu = smp_processor_id(); + + atomic_inc(&init_mm.mm_count); + current->active_mm = &init_mm; + + smp_store_cpu_info(cpu); + set_dec(tb_ticks_per_jiffy); + cpu_callin_map[cpu] = 1; + + smp_ops->setup_cpu(cpu); + if (smp_ops->take_timebase) + smp_ops->take_timebase(); + + spin_lock(&call_lock); + cpu_set(cpu, cpu_online_map); + spin_unlock(&call_lock); + + local_irq_enable(); + + cpu_idle(); + return 0; +} + +int setup_profiling_timer(unsigned int multiplier) +{ + return 0; +} + +void __init smp_cpus_done(unsigned int max_cpus) +{ + cpumask_t old_mask; + + /* We want the setup_cpu() here to be called from CPU 0, but our + * init thread may have been "borrowed" by another CPU in the meantime + * se we pin us down to CPU 0 for a short while + */ + old_mask = current->cpus_allowed; + set_cpus_allowed(current, cpumask_of_cpu(boot_cpuid)); + + smp_ops->setup_cpu(boot_cpuid); + + set_cpus_allowed(current, old_mask); +} + +#ifdef CONFIG_HOTPLUG_CPU +int __cpu_disable(void) +{ + if (smp_ops->cpu_disable) + return smp_ops->cpu_disable(); + + return -ENOSYS; +} + +void __cpu_die(unsigned int cpu) +{ + if (smp_ops->cpu_die) + smp_ops->cpu_die(cpu); +} +#endif diff --git a/arch/powerpc/platforms/chrp/smp.c b/arch/powerpc/platforms/chrp/smp.c index 31ee49c25014..bb2315997d45 100644 --- a/arch/powerpc/platforms/chrp/smp.c +++ b/arch/powerpc/platforms/chrp/smp.c @@ -35,43 +35,6 @@ #include #include -extern unsigned long smp_chrp_cpu_nr; - -static int __init smp_chrp_probe(void) -{ - struct device_node *cpus = NULL; - unsigned int *reg; - int reglen; - int ncpus = 0; - int cpuid; - unsigned int phys; - - /* Count CPUs in the device-tree */ - cpuid = 1; /* the boot cpu is logical cpu 0 */ - while ((cpus = of_find_node_by_type(cpus, "cpu")) != NULL) { - phys = ncpus; - reg = (unsigned int *) get_property(cpus, "reg", ®len); - if (reg && reglen >= sizeof(unsigned int)) - /* hmmm, not having a reg property would be bad */ - phys = *reg; - if (phys != boot_cpuid_phys) { - set_hard_smp_processor_id(cpuid, phys); - ++cpuid; - } - ++ncpus; - } - - printk(KERN_INFO "CHRP SMP probe found %d cpus\n", ncpus); - - /* Nothing more to do if less than 2 of them */ - if (ncpus <= 1) - return 1; - - mpic_request_ipis(); - - return ncpus; -} - static void __devinit smp_chrp_kick_cpu(int nr) { *(unsigned long *)KERNELBASE = nr; @@ -114,7 +77,7 @@ void __devinit smp_chrp_take_timebase(void) /* CHRP with openpic */ struct smp_ops_t chrp_smp_ops = { .message_pass = smp_mpic_message_pass, - .probe = smp_chrp_probe, + .probe = smp_mpic_probe, .kick_cpu = smp_chrp_kick_cpu, .setup_cpu = smp_chrp_setup_cpu, .give_timebase = smp_chrp_give_timebase, diff --git a/arch/ppc/kernel/Makefile b/arch/ppc/kernel/Makefile index b35346df1e37..c610ca933a25 100644 --- a/arch/ppc/kernel/Makefile +++ b/arch/ppc/kernel/Makefile @@ -45,7 +45,6 @@ obj-$(CONFIG_MODULES) += module.o obj-$(CONFIG_NOT_COHERENT_CACHE) += dma-mapping.o obj-$(CONFIG_PCI) += pci.o obj-$(CONFIG_KGDB) += ppc-stub.o -obj-$(CONFIG_SMP) += smp.o smp-tbsync.o obj-$(CONFIG_TAU) += temp.o ifndef CONFIG_E200 obj-$(CONFIG_FSL_BOOKE) += perfmon_fsl_booke.o diff --git a/arch/ppc/kernel/irq.c b/arch/ppc/kernel/irq.c index 772e428aaa59..fbb2b9f8922c 100644 --- a/arch/ppc/kernel/irq.c +++ b/arch/ppc/kernel/irq.c @@ -126,7 +126,7 @@ skip: seq_puts(p, " PowerPC Thermal Assist (cpu temp)\n"); } #endif -#ifdef CONFIG_SMP +#if defined(CONFIG_SMP) && !defined(CONFIG_PPC_MERGE) /* should this be per processor send/receive? */ seq_printf(p, "IPI (recv/sent): %10u/%u\n", atomic_read(&ipi_recv), atomic_read(&ipi_sent)); diff --git a/arch/ppc64/kernel/Makefile b/arch/ppc64/kernel/Makefile index f597c2954b71..c441aebe7648 100644 --- a/arch/ppc64/kernel/Makefile +++ b/arch/ppc64/kernel/Makefile @@ -33,7 +33,6 @@ obj-$(CONFIG_PPC_PSERIES) += udbg_16550.o obj-$(CONFIG_KEXEC) += machine_kexec.o obj-$(CONFIG_EEH) += eeh.o obj-$(CONFIG_PROC_FS) += proc_ppc64.o -obj-$(CONFIG_SMP) += smp.o obj-$(CONFIG_MODULES) += module.o ifneq ($(CONFIG_PPC_MERGE),y) obj-$(CONFIG_MODULES) += ppc_ksyms.o diff --git a/arch/ppc64/kernel/smp.c b/arch/ppc64/kernel/smp.c deleted file mode 100644 index 017c12919832..000000000000 --- a/arch/ppc64/kernel/smp.c +++ /dev/null @@ -1,568 +0,0 @@ -/* - * SMP support for ppc. - * - * Written by Cort Dougan (cort@cs.nmt.edu) borrowing a great - * deal of code from the sparc and intel versions. - * - * Copyright (C) 1999 Cort Dougan - * - * PowerPC-64 Support added by Dave Engebretsen, Peter Bergner, and - * Mike Corrigan {engebret|bergner|mikec}@us.ibm.com - * - * This program is free software; you can redistribute it and/or - * modify it under the terms of the GNU General Public License - * as published by the Free Software Foundation; either version - * 2 of the License, or (at your option) any later version. - */ - -#undef DEBUG - -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include - -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include - -#ifdef DEBUG -#define DBG(fmt...) udbg_printf(fmt) -#else -#define DBG(fmt...) -#endif - -cpumask_t cpu_possible_map = CPU_MASK_NONE; -cpumask_t cpu_online_map = CPU_MASK_NONE; -cpumask_t cpu_sibling_map[NR_CPUS] = { [0 ... NR_CPUS-1] = CPU_MASK_NONE }; - -EXPORT_SYMBOL(cpu_online_map); -EXPORT_SYMBOL(cpu_possible_map); - -struct smp_ops_t *smp_ops; - -static volatile unsigned int cpu_callin_map[NR_CPUS]; - -void smp_call_function_interrupt(void); - -int smt_enabled_at_boot = 1; - -#ifdef CONFIG_MPIC -int __init smp_mpic_probe(void) -{ - int nr_cpus; - - DBG("smp_mpic_probe()...\n"); - - nr_cpus = cpus_weight(cpu_possible_map); - - DBG("nr_cpus: %d\n", nr_cpus); - - if (nr_cpus > 1) - mpic_request_ipis(); - - return nr_cpus; -} - -void __devinit smp_mpic_setup_cpu(int cpu) -{ - mpic_setup_this_cpu(); -} - -void __devinit smp_generic_kick_cpu(int nr) -{ - BUG_ON(nr < 0 || nr >= NR_CPUS); - - /* - * The processor is currently spinning, waiting for the - * cpu_start field to become non-zero After we set cpu_start, - * the processor will continue on to secondary_start - */ - paca[nr].cpu_start = 1; - smp_mb(); -} - -#endif /* CONFIG_MPIC */ - -void smp_message_recv(int msg, struct pt_regs *regs) -{ - switch(msg) { - case PPC_MSG_CALL_FUNCTION: - smp_call_function_interrupt(); - break; - case PPC_MSG_RESCHEDULE: - /* XXX Do we have to do this? */ - set_need_resched(); - break; -#if 0 - case PPC_MSG_MIGRATE_TASK: - /* spare */ - break; -#endif -#ifdef CONFIG_DEBUGGER - case PPC_MSG_DEBUGGER_BREAK: - debugger_ipi(regs); - break; -#endif - default: - printk("SMP %d: smp_message_recv(): unknown msg %d\n", - smp_processor_id(), msg); - break; - } -} - -void smp_send_reschedule(int cpu) -{ - smp_ops->message_pass(cpu, PPC_MSG_RESCHEDULE); -} - -#ifdef CONFIG_DEBUGGER -void smp_send_debugger_break(int cpu) -{ - smp_ops->message_pass(cpu, PPC_MSG_DEBUGGER_BREAK); -} -#endif - -static void stop_this_cpu(void *dummy) -{ - local_irq_disable(); - while (1) - ; -} - -void smp_send_stop(void) -{ - smp_call_function(stop_this_cpu, NULL, 1, 0); -} - -/* - * Structure and data for smp_call_function(). This is designed to minimise - * static memory requirements. It also looks cleaner. - * Stolen from the i386 version. - */ -static __cacheline_aligned_in_smp DEFINE_SPINLOCK(call_lock); - -static struct call_data_struct { - void (*func) (void *info); - void *info; - atomic_t started; - atomic_t finished; - int wait; -} *call_data; - -/* delay of at least 8 seconds on 1GHz cpu */ -#define SMP_CALL_TIMEOUT (1UL << (30 + 3)) - -/* - * This function sends a 'generic call function' IPI to all other CPUs - * in the system. - * - * [SUMMARY] Run a function on all other CPUs. - * The function to run. This must be fast and non-blocking. - * An arbitrary pointer to pass to the function. - * currently unused. - * If true, wait (atomically) until function has completed on other CPUs. - * [RETURNS] 0 on success, else a negative status code. Does not return until - * remote CPUs are nearly ready to execute <> or are or have executed. - * - * You must not call this function with disabled interrupts or from a - * hardware interrupt handler or from a bottom half handler. - */ -int smp_call_function (void (*func) (void *info), void *info, int nonatomic, - int wait) -{ - struct call_data_struct data; - int ret = -1, cpus; - unsigned long timeout; - - /* Can deadlock when called with interrupts disabled */ - WARN_ON(irqs_disabled()); - - data.func = func; - data.info = info; - atomic_set(&data.started, 0); - data.wait = wait; - if (wait) - atomic_set(&data.finished, 0); - - spin_lock(&call_lock); - /* Must grab online cpu count with preempt disabled, otherwise - * it can change. */ - cpus = num_online_cpus() - 1; - if (!cpus) { - ret = 0; - goto out; - } - - call_data = &data; - smp_wmb(); - /* Send a message to all other CPUs and wait for them to respond */ - smp_ops->message_pass(MSG_ALL_BUT_SELF, PPC_MSG_CALL_FUNCTION); - - /* Wait for response */ - timeout = SMP_CALL_TIMEOUT; - while (atomic_read(&data.started) != cpus) { - HMT_low(); - if (--timeout == 0) { - printk("smp_call_function on cpu %d: other cpus not " - "responding (%d)\n", smp_processor_id(), - atomic_read(&data.started)); - debugger(NULL); - goto out; - } - } - - if (wait) { - timeout = SMP_CALL_TIMEOUT; - while (atomic_read(&data.finished) != cpus) { - HMT_low(); - if (--timeout == 0) { - printk("smp_call_function on cpu %d: other " - "cpus not finishing (%d/%d)\n", - smp_processor_id(), - atomic_read(&data.finished), - atomic_read(&data.started)); - debugger(NULL); - goto out; - } - } - } - - ret = 0; - -out: - call_data = NULL; - HMT_medium(); - spin_unlock(&call_lock); - return ret; -} - -EXPORT_SYMBOL(smp_call_function); - -void smp_call_function_interrupt(void) -{ - void (*func) (void *info); - void *info; - int wait; - - /* call_data will be NULL if the sender timed out while - * waiting on us to receive the call. - */ - if (!call_data) - return; - - func = call_data->func; - info = call_data->info; - wait = call_data->wait; - - if (!wait) - smp_mb__before_atomic_inc(); - - /* - * Notify initiating CPU that I've grabbed the data and am - * about to execute the function - */ - atomic_inc(&call_data->started); - /* - * At this point the info structure may be out of scope unless wait==1 - */ - (*func)(info); - if (wait) { - smp_mb__before_atomic_inc(); - atomic_inc(&call_data->finished); - } -} - -extern struct gettimeofday_struct do_gtod; - -struct thread_info *current_set[NR_CPUS]; - -DECLARE_PER_CPU(unsigned int, pvr); - -static void __devinit smp_store_cpu_info(int id) -{ - per_cpu(pvr, id) = mfspr(SPRN_PVR); -} - -static void __init smp_create_idle(unsigned int cpu) -{ - struct task_struct *p; - - /* create a process for the processor */ - p = fork_idle(cpu); - if (IS_ERR(p)) - panic("failed fork for CPU %u: %li", cpu, PTR_ERR(p)); - paca[cpu].__current = p; - current_set[cpu] = p->thread_info; -} - -void __init smp_prepare_cpus(unsigned int max_cpus) -{ - unsigned int cpu; - - DBG("smp_prepare_cpus\n"); - - /* - * setup_cpu may need to be called on the boot cpu. We havent - * spun any cpus up but lets be paranoid. - */ - BUG_ON(boot_cpuid != smp_processor_id()); - - /* Fixup boot cpu */ - smp_store_cpu_info(boot_cpuid); - cpu_callin_map[boot_cpuid] = 1; - -#ifndef CONFIG_PPC_ISERIES - paca[boot_cpuid].next_jiffy_update_tb = tb_last_stamp = get_tb(); - - /* - * Should update do_gtod.stamp_xsec. - * For now we leave it which means the time can be some - * number of msecs off until someone does a settimeofday() - */ - do_gtod.varp->tb_orig_stamp = tb_last_stamp; - systemcfg->tb_orig_stamp = tb_last_stamp; -#endif - - max_cpus = smp_ops->probe(); - - smp_space_timers(max_cpus); - - for_each_cpu(cpu) - if (cpu != boot_cpuid) - smp_create_idle(cpu); -} - -void __devinit smp_prepare_boot_cpu(void) -{ - BUG_ON(smp_processor_id() != boot_cpuid); - - cpu_set(boot_cpuid, cpu_online_map); - - paca[boot_cpuid].__current = current; - current_set[boot_cpuid] = current->thread_info; -} - -#ifdef CONFIG_HOTPLUG_CPU -/* State of each CPU during hotplug phases */ -DEFINE_PER_CPU(int, cpu_state) = { 0 }; - -int generic_cpu_disable(void) -{ - unsigned int cpu = smp_processor_id(); - - if (cpu == boot_cpuid) - return -EBUSY; - - systemcfg->processorCount--; - cpu_clear(cpu, cpu_online_map); - fixup_irqs(cpu_online_map); - return 0; -} - -int generic_cpu_enable(unsigned int cpu) -{ - /* Do the normal bootup if we haven't - * already bootstrapped. */ - if (system_state != SYSTEM_RUNNING) - return -ENOSYS; - - /* get the target out of it's holding state */ - per_cpu(cpu_state, cpu) = CPU_UP_PREPARE; - smp_wmb(); - - while (!cpu_online(cpu)) - cpu_relax(); - - fixup_irqs(cpu_online_map); - /* counter the irq disable in fixup_irqs */ - local_irq_enable(); - return 0; -} - -void generic_cpu_die(unsigned int cpu) -{ - int i; - - for (i = 0; i < 100; i++) { - smp_rmb(); - if (per_cpu(cpu_state, cpu) == CPU_DEAD) - return; - msleep(100); - } - printk(KERN_ERR "CPU%d didn't die...\n", cpu); -} - -void generic_mach_cpu_die(void) -{ - unsigned int cpu; - - local_irq_disable(); - cpu = smp_processor_id(); - printk(KERN_DEBUG "CPU%d offline\n", cpu); - __get_cpu_var(cpu_state) = CPU_DEAD; - smp_wmb(); - while (__get_cpu_var(cpu_state) != CPU_UP_PREPARE) - cpu_relax(); - - flush_tlb_pending(); - cpu_set(cpu, cpu_online_map); - local_irq_enable(); -} -#endif - -static int __devinit cpu_enable(unsigned int cpu) -{ - if (smp_ops->cpu_enable) - return smp_ops->cpu_enable(cpu); - - return -ENOSYS; -} - -int __devinit __cpu_up(unsigned int cpu) -{ - int c; - - if (!cpu_enable(cpu)) - return 0; - - if (smp_ops->cpu_bootable && !smp_ops->cpu_bootable(cpu)) - return -EINVAL; - - paca[cpu].default_decr = tb_ticks_per_jiffy; - - /* Make sure callin-map entry is 0 (can be leftover a CPU - * hotplug - */ - cpu_callin_map[cpu] = 0; - - /* The information for processor bringup must - * be written out to main store before we release - * the processor. - */ - smp_mb(); - - /* wake up cpus */ - DBG("smp: kicking cpu %d\n", cpu); - smp_ops->kick_cpu(cpu); - - /* - * wait to see if the cpu made a callin (is actually up). - * use this value that I found through experimentation. - * -- Cort - */ - if (system_state < SYSTEM_RUNNING) - for (c = 5000; c && !cpu_callin_map[cpu]; c--) - udelay(100); -#ifdef CONFIG_HOTPLUG_CPU - else - /* - * CPUs can take much longer to come up in the - * hotplug case. Wait five seconds. - */ - for (c = 25; c && !cpu_callin_map[cpu]; c--) { - msleep(200); - } -#endif - - if (!cpu_callin_map[cpu]) { - printk("Processor %u is stuck.\n", cpu); - return -ENOENT; - } - - printk("Processor %u found.\n", cpu); - - if (smp_ops->give_timebase) - smp_ops->give_timebase(); - - /* Wait until cpu puts itself in the online map */ - while (!cpu_online(cpu)) - cpu_relax(); - - return 0; -} - - -/* Activate a secondary processor. */ -int __devinit start_secondary(void *unused) -{ - unsigned int cpu = smp_processor_id(); - - atomic_inc(&init_mm.mm_count); - current->active_mm = &init_mm; - - smp_store_cpu_info(cpu); - set_dec(paca[cpu].default_decr); - cpu_callin_map[cpu] = 1; - - smp_ops->setup_cpu(cpu); - if (smp_ops->take_timebase) - smp_ops->take_timebase(); - - spin_lock(&call_lock); - cpu_set(cpu, cpu_online_map); - spin_unlock(&call_lock); - - local_irq_enable(); - - cpu_idle(); - return 0; -} - -int setup_profiling_timer(unsigned int multiplier) -{ - return 0; -} - -void __init smp_cpus_done(unsigned int max_cpus) -{ - cpumask_t old_mask; - - /* We want the setup_cpu() here to be called from CPU 0, but our - * init thread may have been "borrowed" by another CPU in the meantime - * se we pin us down to CPU 0 for a short while - */ - old_mask = current->cpus_allowed; - set_cpus_allowed(current, cpumask_of_cpu(boot_cpuid)); - - smp_ops->setup_cpu(boot_cpuid); - - set_cpus_allowed(current, old_mask); -} - -#ifdef CONFIG_HOTPLUG_CPU -int __cpu_disable(void) -{ - if (smp_ops->cpu_disable) - return smp_ops->cpu_disable(); - - return -ENOSYS; -} - -void __cpu_die(unsigned int cpu) -{ - if (smp_ops->cpu_die) - smp_ops->cpu_die(cpu); -} -#endif diff --git a/include/asm-powerpc/smp.h b/include/asm-powerpc/smp.h new file mode 100644 index 000000000000..8bcdd0faefea --- /dev/null +++ b/include/asm-powerpc/smp.h @@ -0,0 +1,119 @@ +/* + * smp.h: PowerPC-specific SMP code. + * + * Original was a copy of sparc smp.h. Now heavily modified + * for PPC. + * + * Copyright (C) 1996 David S. Miller (davem@caip.rutgers.edu) + * Copyright (C) 1996-2001 Cort Dougan + * + * This program is free software; you can redistribute it and/or + * modify it under the terms of the GNU General Public License + * as published by the Free Software Foundation; either version + * 2 of the License, or (at your option) any later version. + */ + +#ifndef _ASM_POWERPC_SMP_H +#define _ASM_POWERPC_SMP_H +#ifdef __KERNEL__ + +#include +#include +#include +#include + +#ifndef __ASSEMBLY__ + +#ifdef CONFIG_PPC64 +#include +#endif + +extern int boot_cpuid; +extern int boot_cpuid_phys; + +extern void cpu_die(void); + +#ifdef CONFIG_SMP + +extern void smp_send_debugger_break(int cpu); +struct pt_regs; +extern void smp_message_recv(int, struct pt_regs *); + +#ifdef CONFIG_HOTPLUG_CPU +extern void fixup_irqs(cpumask_t map); +int generic_cpu_disable(void); +int generic_cpu_enable(unsigned int cpu); +void generic_cpu_die(unsigned int cpu); +void generic_mach_cpu_die(void); +#endif + +#ifdef CONFIG_PPC64 +#define raw_smp_processor_id() (get_paca()->paca_index) +#define hard_smp_processor_id() (get_paca()->hw_cpu_id) +#else +/* 32-bit */ +extern int smp_hw_index[]; + +#define raw_smp_processor_id() (current_thread_info()->cpu) +#define hard_smp_processor_id() (smp_hw_index[smp_processor_id()]) +#define get_hard_smp_processor_id(cpu) (smp_hw_index[(cpu)]) +#define set_hard_smp_processor_id(cpu, phys)\ + (smp_hw_index[(cpu)] = (phys)) +#endif + +extern cpumask_t cpu_sibling_map[NR_CPUS]; + +/* Since OpenPIC has only 4 IPIs, we use slightly different message numbers. + * + * Make sure this matches openpic_request_IPIs in open_pic.c, or what shows up + * in /proc/interrupts will be wrong!!! --Troy */ +#define PPC_MSG_CALL_FUNCTION 0 +#define PPC_MSG_RESCHEDULE 1 +/* This is unused now */ +#if 0 +#define PPC_MSG_MIGRATE_TASK 2 +#endif +#define PPC_MSG_DEBUGGER_BREAK 3 + +void smp_init_iSeries(void); +void smp_init_pSeries(void); +void smp_init_cell(void); +void smp_setup_cpu_maps(void); + +extern int __cpu_disable(void); +extern void __cpu_die(unsigned int cpu); + +#else +/* for UP */ +#define smp_setup_cpu_maps() +#define smp_release_cpus() + +#endif /* CONFIG_SMP */ + +#ifdef CONFIG_PPC64 +#define get_hard_smp_processor_id(CPU) (paca[(CPU)].hw_cpu_id) +#define set_hard_smp_processor_id(CPU, VAL) \ + do { (paca[(CPU)].hw_cpu_id = (VAL)); } while (0) +#else +/* 32-bit */ +#ifndef CONFIG_SMP +#define get_hard_smp_processor_id(cpu) boot_cpuid_phys +#define set_hard_smp_processor_id(cpu, phys) +#endif +#endif + +extern int smt_enabled_at_boot; + +extern int smp_mpic_probe(void); +extern void smp_mpic_setup_cpu(int cpu); +extern void smp_generic_kick_cpu(int nr); + +extern void smp_generic_give_timebase(void); +extern void smp_generic_take_timebase(void); + +extern struct smp_ops_t *smp_ops; + +#endif /* __ASSEMBLY__ */ + +#endif /* __KERNEL__ */ +#endif /* _ASM_POWERPC_SMP_H) */ diff --git a/include/asm-ppc64/smp.h b/include/asm-ppc64/smp.h deleted file mode 100644 index ba0f5c8bbb22..000000000000 --- a/include/asm-ppc64/smp.h +++ /dev/null @@ -1,92 +0,0 @@ -/* - * smp.h: PPC64 specific SMP code. - * - * Original was a copy of sparc smp.h. Now heavily modified - * for PPC. - * - * Copyright (C) 1996 David S. Miller (davem@caip.rutgers.edu) - * Copyright (C) 1996-2001 Cort Dougan - * - * This program is free software; you can redistribute it and/or - * modify it under the terms of the GNU General Public License - * as published by the Free Software Foundation; either version - * 2 of the License, or (at your option) any later version. - */ - -#ifdef __KERNEL__ -#ifndef _PPC64_SMP_H -#define _PPC64_SMP_H - -#include -#include -#include -#include - -#ifndef __ASSEMBLY__ - -#include - -extern int boot_cpuid; -extern int boot_cpuid_phys; - -extern void cpu_die(void); - -#ifdef CONFIG_SMP - -extern void smp_send_debugger_break(int cpu); -struct pt_regs; -extern void smp_message_recv(int, struct pt_regs *); - -#ifdef CONFIG_HOTPLUG_CPU -extern void fixup_irqs(cpumask_t map); -int generic_cpu_disable(void); -int generic_cpu_enable(unsigned int cpu); -void generic_cpu_die(unsigned int cpu); -void generic_mach_cpu_die(void); -#endif - -#define raw_smp_processor_id() (get_paca()->paca_index) -#define hard_smp_processor_id() (get_paca()->hw_cpu_id) - -extern cpumask_t cpu_sibling_map[NR_CPUS]; - -/* Since OpenPIC has only 4 IPIs, we use slightly different message numbers. - * - * Make sure this matches openpic_request_IPIs in open_pic.c, or what shows up - * in /proc/interrupts will be wrong!!! --Troy */ -#define PPC_MSG_CALL_FUNCTION 0 -#define PPC_MSG_RESCHEDULE 1 -/* This is unused now */ -#if 0 -#define PPC_MSG_MIGRATE_TASK 2 -#endif -#define PPC_MSG_DEBUGGER_BREAK 3 - -void smp_init_iSeries(void); -void smp_init_pSeries(void); -void smp_init_cell(void); - -extern int __cpu_disable(void); -extern void __cpu_die(unsigned int cpu); -#endif /* CONFIG_SMP */ - -#define get_hard_smp_processor_id(CPU) (paca[(CPU)].hw_cpu_id) -#define set_hard_smp_processor_id(CPU, VAL) \ - do { (paca[(CPU)].hw_cpu_id = (VAL)); } while (0) - -extern int smt_enabled_at_boot; - -extern int smp_mpic_probe(void); -extern void smp_mpic_setup_cpu(int cpu); -extern void smp_generic_kick_cpu(int nr); -extern void smp_release_cpus(void); - -extern void smp_generic_give_timebase(void); -extern void smp_generic_take_timebase(void); - -extern struct smp_ops_t *smp_ops; - -#endif /* __ASSEMBLY__ */ - -#endif /* !(_PPC64_SMP_H) */ -#endif /* __KERNEL__ */ -- cgit v1.2.3 From 3c726f8dee6f55e96475574e9f645327e461884c Mon Sep 17 00:00:00 2001 From: Benjamin Herrenschmidt Date: Mon, 7 Nov 2005 11:06:55 +1100 Subject: [PATCH] ppc64: support 64k pages Adds a new CONFIG_PPC_64K_PAGES which, when enabled, changes the kernel base page size to 64K. The resulting kernel still boots on any hardware. On current machines with 4K pages support only, the kernel will maintain 16 "subpages" for each 64K page transparently. Note that while real 64K capable HW has been tested, the current patch will not enable it yet as such hardware is not released yet, and I'm still verifying with the firmware architects the proper to get the information from the newer hypervisors. Signed-off-by: Benjamin Herrenschmidt Signed-off-by: Linus Torvalds --- arch/powerpc/Kconfig | 9 + arch/powerpc/kernel/asm-offsets.c | 3 + arch/powerpc/kernel/cputable.c | 4 +- arch/powerpc/kernel/head_64.S | 300 +++++++++++---- arch/powerpc/kernel/lparmap.c | 2 +- arch/powerpc/kernel/process.c | 6 +- arch/powerpc/kernel/prom.c | 76 ++-- arch/powerpc/kernel/setup_64.c | 31 +- arch/powerpc/lib/copypage_64.S | 2 +- arch/powerpc/lib/copyuser_64.S | 4 +- arch/powerpc/mm/hash_low_64.S | 613 +++++++++++++++++++++++++++++-- arch/powerpc/mm/hash_native_64.c | 377 +++++++++++-------- arch/powerpc/mm/hash_utils_64.c | 532 +++++++++++++++++++++------ arch/powerpc/mm/hugetlbpage.c | 134 ++++--- arch/powerpc/mm/init_64.c | 18 +- arch/powerpc/mm/mem.c | 56 ++- arch/powerpc/mm/pgtable_64.c | 22 +- arch/powerpc/mm/ppc_mmu_32.c | 15 + arch/powerpc/mm/slb.c | 102 ++++- arch/powerpc/mm/slb_low.S | 220 +++++++---- arch/powerpc/mm/stab.c | 30 +- arch/powerpc/mm/tlb_64.c | 32 +- arch/powerpc/platforms/iseries/htab.c | 65 ++-- arch/powerpc/platforms/iseries/hvlog.c | 4 +- arch/powerpc/platforms/iseries/iommu.c | 74 ++-- arch/powerpc/platforms/iseries/setup.c | 13 +- arch/powerpc/platforms/iseries/vio.c | 39 +- arch/powerpc/platforms/iseries/viopath.c | 16 +- arch/powerpc/platforms/pseries/lpar.c | 115 +++--- arch/ppc64/Kconfig | 13 + arch/ppc64/kernel/asm-offsets.c | 3 + arch/ppc64/kernel/head.S | 300 +++++++++++---- arch/ppc64/kernel/pacaData.c | 2 +- arch/ppc64/kernel/prom.c | 94 ++--- include/asm-powerpc/cputable.h | 3 + include/asm-powerpc/iommu.h | 5 + include/asm-powerpc/machdep.h | 10 +- include/asm-powerpc/prom.h | 8 + include/asm-powerpc/system.h | 2 +- include/asm-powerpc/thread_info.h | 20 +- include/asm-powerpc/tlbflush.h | 9 +- include/asm-ppc64/mmu.h | 208 ++++++----- include/asm-ppc64/mmu_context.h | 15 +- include/asm-ppc64/paca.h | 13 +- include/asm-ppc64/page.h | 147 ++++++-- include/asm-ppc64/pgalloc.h | 47 ++- include/asm-ppc64/pgtable-4k.h | 88 +++++ include/asm-ppc64/pgtable-64k.h | 87 +++++ include/asm-ppc64/pgtable.h | 160 +++----- include/asm-ppc64/prom.h | 8 + include/asm-ppc64/system.h | 2 +- mm/hugetlb.c | 3 + 52 files changed, 2949 insertions(+), 1212 deletions(-) create mode 100644 include/asm-ppc64/pgtable-4k.h create mode 100644 include/asm-ppc64/pgtable-64k.h (limited to 'arch/ppc64') diff --git a/arch/powerpc/Kconfig b/arch/powerpc/Kconfig index f4e25c648fbb..ca7acb0c79f0 100644 --- a/arch/powerpc/Kconfig +++ b/arch/powerpc/Kconfig @@ -603,6 +603,15 @@ config NODES_SPAN_OTHER_NODES def_bool y depends on NEED_MULTIPLE_NODES +config PPC_64K_PAGES + bool "64k page size" + help + This option changes the kernel logical page size to 64k. On machines + without processor support for 64k pages, the kernel will simulate + them by loading each individual 4k page on demand transparently, + while on hardware with such support, it will be used to map + normal application pages. + config SCHED_SMT bool "SMT (Hyperthreading) scheduler support" depends on PPC64 && SMP diff --git a/arch/powerpc/kernel/asm-offsets.c b/arch/powerpc/kernel/asm-offsets.c index bc5a3689cc05..b75757251994 100644 --- a/arch/powerpc/kernel/asm-offsets.c +++ b/arch/powerpc/kernel/asm-offsets.c @@ -125,6 +125,9 @@ int main(void) DEFINE(PACASLBCACHE, offsetof(struct paca_struct, slb_cache)); DEFINE(PACASLBCACHEPTR, offsetof(struct paca_struct, slb_cache_ptr)); DEFINE(PACACONTEXTID, offsetof(struct paca_struct, context.id)); +#ifdef CONFIG_PPC_64K_PAGES + DEFINE(PACAPGDIR, offsetof(struct paca_struct, pgdir)); +#endif #ifdef CONFIG_HUGETLB_PAGE DEFINE(PACALOWHTLBAREAS, offsetof(struct paca_struct, context.low_htlb_areas)); DEFINE(PACAHIGHHTLBAREAS, offsetof(struct paca_struct, context.high_htlb_areas)); diff --git a/arch/powerpc/kernel/cputable.c b/arch/powerpc/kernel/cputable.c index b91345fa0805..33c63bcf69f8 100644 --- a/arch/powerpc/kernel/cputable.c +++ b/arch/powerpc/kernel/cputable.c @@ -240,7 +240,7 @@ struct cpu_spec cpu_specs[] = { .oprofile_model = &op_model_power4, #endif }, - { /* Power5 */ + { /* Power5 GR */ .pvr_mask = 0xffff0000, .pvr_value = 0x003a0000, .cpu_name = "POWER5 (gr)", @@ -255,7 +255,7 @@ struct cpu_spec cpu_specs[] = { .oprofile_model = &op_model_power4, #endif }, - { /* Power5 */ + { /* Power5 GS */ .pvr_mask = 0xffff0000, .pvr_value = 0x003b0000, .cpu_name = "POWER5 (gs)", diff --git a/arch/powerpc/kernel/head_64.S b/arch/powerpc/kernel/head_64.S index 45d81976987f..16ab40daa738 100644 --- a/arch/powerpc/kernel/head_64.S +++ b/arch/powerpc/kernel/head_64.S @@ -195,11 +195,11 @@ exception_marker: #define EX_R12 24 #define EX_R13 32 #define EX_SRR0 40 -#define EX_R3 40 /* SLB miss saves R3, but not SRR0 */ #define EX_DAR 48 -#define EX_LR 48 /* SLB miss saves LR, but not DAR */ #define EX_DSISR 56 #define EX_CCR 60 +#define EX_R3 64 +#define EX_LR 72 #define EXCEPTION_PROLOG_PSERIES(area, label) \ mfspr r13,SPRN_SPRG3; /* get paca address into r13 */ \ @@ -419,17 +419,22 @@ data_access_slb_pSeries: mtspr SPRN_SPRG1,r13 RUNLATCH_ON(r13) mfspr r13,SPRN_SPRG3 /* get paca address into r13 */ + std r3,PACA_EXSLB+EX_R3(r13) + mfspr r3,SPRN_DAR std r9,PACA_EXSLB+EX_R9(r13) /* save r9 - r12 */ + mfcr r9 +#ifdef __DISABLED__ + /* Keep that around for when we re-implement dynamic VSIDs */ + cmpdi r3,0 + bge slb_miss_user_pseries +#endif /* __DISABLED__ */ std r10,PACA_EXSLB+EX_R10(r13) std r11,PACA_EXSLB+EX_R11(r13) std r12,PACA_EXSLB+EX_R12(r13) - std r3,PACA_EXSLB+EX_R3(r13) - mfspr r9,SPRN_SPRG1 - std r9,PACA_EXSLB+EX_R13(r13) - mfcr r9 + mfspr r10,SPRN_SPRG1 + std r10,PACA_EXSLB+EX_R13(r13) mfspr r12,SPRN_SRR1 /* and SRR1 */ - mfspr r3,SPRN_DAR - b .do_slb_miss /* Rel. branch works in real mode */ + b .slb_miss_realmode /* Rel. branch works in real mode */ STD_EXCEPTION_PSERIES(0x400, instruction_access) @@ -440,17 +445,22 @@ instruction_access_slb_pSeries: mtspr SPRN_SPRG1,r13 RUNLATCH_ON(r13) mfspr r13,SPRN_SPRG3 /* get paca address into r13 */ + std r3,PACA_EXSLB+EX_R3(r13) + mfspr r3,SPRN_SRR0 /* SRR0 is faulting address */ std r9,PACA_EXSLB+EX_R9(r13) /* save r9 - r12 */ + mfcr r9 +#ifdef __DISABLED__ + /* Keep that around for when we re-implement dynamic VSIDs */ + cmpdi r3,0 + bge slb_miss_user_pseries +#endif /* __DISABLED__ */ std r10,PACA_EXSLB+EX_R10(r13) std r11,PACA_EXSLB+EX_R11(r13) std r12,PACA_EXSLB+EX_R12(r13) - std r3,PACA_EXSLB+EX_R3(r13) - mfspr r9,SPRN_SPRG1 - std r9,PACA_EXSLB+EX_R13(r13) - mfcr r9 + mfspr r10,SPRN_SPRG1 + std r10,PACA_EXSLB+EX_R13(r13) mfspr r12,SPRN_SRR1 /* and SRR1 */ - mfspr r3,SPRN_SRR0 /* SRR0 is faulting address */ - b .do_slb_miss /* Rel. branch works in real mode */ + b .slb_miss_realmode /* Rel. branch works in real mode */ STD_EXCEPTION_PSERIES(0x500, hardware_interrupt) STD_EXCEPTION_PSERIES(0x600, alignment) @@ -508,6 +518,38 @@ _GLOBAL(do_stab_bolted_pSeries) mfspr r12,SPRN_SPRG2 EXCEPTION_PROLOG_PSERIES(PACA_EXSLB, .do_stab_bolted) +/* + * We have some room here we use that to put + * the peries slb miss user trampoline code so it's reasonably + * away from slb_miss_user_common to avoid problems with rfid + * + * This is used for when the SLB miss handler has to go virtual, + * which doesn't happen for now anymore but will once we re-implement + * dynamic VSIDs for shared page tables + */ +#ifdef __DISABLED__ +slb_miss_user_pseries: + std r10,PACA_EXGEN+EX_R10(r13) + std r11,PACA_EXGEN+EX_R11(r13) + std r12,PACA_EXGEN+EX_R12(r13) + mfspr r10,SPRG1 + ld r11,PACA_EXSLB+EX_R9(r13) + ld r12,PACA_EXSLB+EX_R3(r13) + std r10,PACA_EXGEN+EX_R13(r13) + std r11,PACA_EXGEN+EX_R9(r13) + std r12,PACA_EXGEN+EX_R3(r13) + clrrdi r12,r13,32 + mfmsr r10 + mfspr r11,SRR0 /* save SRR0 */ + ori r12,r12,slb_miss_user_common@l /* virt addr of handler */ + ori r10,r10,MSR_IR|MSR_DR|MSR_RI + mtspr SRR0,r12 + mfspr r12,SRR1 /* and SRR1 */ + mtspr SRR1,r10 + rfid + b . /* prevent spec. execution */ +#endif /* __DISABLED__ */ + /* * Vectors for the FWNMI option. Share common code. */ @@ -559,22 +601,59 @@ END_FTR_SECTION_IFCLR(CPU_FTR_SLB) .globl data_access_slb_iSeries data_access_slb_iSeries: mtspr SPRN_SPRG1,r13 /* save r13 */ - EXCEPTION_PROLOG_ISERIES_1(PACA_EXSLB) + mfspr r13,SPRN_SPRG3 /* get paca address into r13 */ std r3,PACA_EXSLB+EX_R3(r13) - ld r12,PACALPPACA+LPPACASRR1(r13) mfspr r3,SPRN_DAR - b .do_slb_miss + std r9,PACA_EXSLB+EX_R9(r13) + mfcr r9 +#ifdef __DISABLED__ + cmpdi r3,0 + bge slb_miss_user_iseries +#endif + std r10,PACA_EXSLB+EX_R10(r13) + std r11,PACA_EXSLB+EX_R11(r13) + std r12,PACA_EXSLB+EX_R12(r13) + mfspr r10,SPRN_SPRG1 + std r10,PACA_EXSLB+EX_R13(r13) + ld r12,PACALPPACA+LPPACASRR1(r13); + b .slb_miss_realmode STD_EXCEPTION_ISERIES(0x400, instruction_access, PACA_EXGEN) .globl instruction_access_slb_iSeries instruction_access_slb_iSeries: mtspr SPRN_SPRG1,r13 /* save r13 */ - EXCEPTION_PROLOG_ISERIES_1(PACA_EXSLB) + mfspr r13,SPRN_SPRG3 /* get paca address into r13 */ std r3,PACA_EXSLB+EX_R3(r13) - ld r12,PACALPPACA+LPPACASRR1(r13) - ld r3,PACALPPACA+LPPACASRR0(r13) - b .do_slb_miss + ld r3,PACALPPACA+LPPACASRR0(r13) /* get SRR0 value */ + std r9,PACA_EXSLB+EX_R9(r13) + mfcr r9 +#ifdef __DISABLED__ + cmpdi r3,0 + bge .slb_miss_user_iseries +#endif + std r10,PACA_EXSLB+EX_R10(r13) + std r11,PACA_EXSLB+EX_R11(r13) + std r12,PACA_EXSLB+EX_R12(r13) + mfspr r10,SPRN_SPRG1 + std r10,PACA_EXSLB+EX_R13(r13) + ld r12,PACALPPACA+LPPACASRR1(r13); + b .slb_miss_realmode + +#ifdef __DISABLED__ +slb_miss_user_iseries: + std r10,PACA_EXGEN+EX_R10(r13) + std r11,PACA_EXGEN+EX_R11(r13) + std r12,PACA_EXGEN+EX_R12(r13) + mfspr r10,SPRG1 + ld r11,PACA_EXSLB+EX_R9(r13) + ld r12,PACA_EXSLB+EX_R3(r13) + std r10,PACA_EXGEN+EX_R13(r13) + std r11,PACA_EXGEN+EX_R9(r13) + std r12,PACA_EXGEN+EX_R3(r13) + EXCEPTION_PROLOG_ISERIES_2 + b slb_miss_user_common +#endif MASKABLE_EXCEPTION_ISERIES(0x500, hardware_interrupt) STD_EXCEPTION_ISERIES(0x600, alignment, PACA_EXGEN) @@ -809,6 +888,126 @@ instruction_access_common: li r5,0x400 b .do_hash_page /* Try to handle as hpte fault */ +/* + * Here is the common SLB miss user that is used when going to virtual + * mode for SLB misses, that is currently not used + */ +#ifdef __DISABLED__ + .align 7 + .globl slb_miss_user_common +slb_miss_user_common: + mflr r10 + std r3,PACA_EXGEN+EX_DAR(r13) + stw r9,PACA_EXGEN+EX_CCR(r13) + std r10,PACA_EXGEN+EX_LR(r13) + std r11,PACA_EXGEN+EX_SRR0(r13) + bl .slb_allocate_user + + ld r10,PACA_EXGEN+EX_LR(r13) + ld r3,PACA_EXGEN+EX_R3(r13) + lwz r9,PACA_EXGEN+EX_CCR(r13) + ld r11,PACA_EXGEN+EX_SRR0(r13) + mtlr r10 + beq- slb_miss_fault + + andi. r10,r12,MSR_RI /* check for unrecoverable exception */ + beq- unrecov_user_slb + mfmsr r10 + +.machine push +.machine "power4" + mtcrf 0x80,r9 +.machine pop + + clrrdi r10,r10,2 /* clear RI before setting SRR0/1 */ + mtmsrd r10,1 + + mtspr SRR0,r11 + mtspr SRR1,r12 + + ld r9,PACA_EXGEN+EX_R9(r13) + ld r10,PACA_EXGEN+EX_R10(r13) + ld r11,PACA_EXGEN+EX_R11(r13) + ld r12,PACA_EXGEN+EX_R12(r13) + ld r13,PACA_EXGEN+EX_R13(r13) + rfid + b . + +slb_miss_fault: + EXCEPTION_PROLOG_COMMON(0x380, PACA_EXGEN) + ld r4,PACA_EXGEN+EX_DAR(r13) + li r5,0 + std r4,_DAR(r1) + std r5,_DSISR(r1) + b .handle_page_fault + +unrecov_user_slb: + EXCEPTION_PROLOG_COMMON(0x4200, PACA_EXGEN) + DISABLE_INTS + bl .save_nvgprs +1: addi r3,r1,STACK_FRAME_OVERHEAD + bl .unrecoverable_exception + b 1b + +#endif /* __DISABLED__ */ + + +/* + * r13 points to the PACA, r9 contains the saved CR, + * r12 contain the saved SRR1, SRR0 is still ready for return + * r3 has the faulting address + * r9 - r13 are saved in paca->exslb. + * r3 is saved in paca->slb_r3 + * We assume we aren't going to take any exceptions during this procedure. + */ +_GLOBAL(slb_miss_realmode) + mflr r10 + + stw r9,PACA_EXSLB+EX_CCR(r13) /* save CR in exc. frame */ + std r10,PACA_EXSLB+EX_LR(r13) /* save LR */ + + bl .slb_allocate_realmode + + /* All done -- return from exception. */ + + ld r10,PACA_EXSLB+EX_LR(r13) + ld r3,PACA_EXSLB+EX_R3(r13) + lwz r9,PACA_EXSLB+EX_CCR(r13) /* get saved CR */ +#ifdef CONFIG_PPC_ISERIES + ld r11,PACALPPACA+LPPACASRR0(r13) /* get SRR0 value */ +#endif /* CONFIG_PPC_ISERIES */ + + mtlr r10 + + andi. r10,r12,MSR_RI /* check for unrecoverable exception */ + beq- unrecov_slb + +.machine push +.machine "power4" + mtcrf 0x80,r9 + mtcrf 0x01,r9 /* slb_allocate uses cr0 and cr7 */ +.machine pop + +#ifdef CONFIG_PPC_ISERIES + mtspr SPRN_SRR0,r11 + mtspr SPRN_SRR1,r12 +#endif /* CONFIG_PPC_ISERIES */ + ld r9,PACA_EXSLB+EX_R9(r13) + ld r10,PACA_EXSLB+EX_R10(r13) + ld r11,PACA_EXSLB+EX_R11(r13) + ld r12,PACA_EXSLB+EX_R12(r13) + ld r13,PACA_EXSLB+EX_R13(r13) + rfid + b . /* prevent speculative execution */ + +unrecov_slb: + EXCEPTION_PROLOG_COMMON(0x4100, PACA_EXSLB) + DISABLE_INTS + bl .save_nvgprs +1: addi r3,r1,STACK_FRAME_OVERHEAD + bl .unrecoverable_exception + b 1b + .align 7 .globl hardware_interrupt_common .globl hardware_interrupt_entry @@ -1138,62 +1337,6 @@ _GLOBAL(do_stab_bolted) rfid b . /* prevent speculative execution */ -/* - * r13 points to the PACA, r9 contains the saved CR, - * r11 and r12 contain the saved SRR0 and SRR1. - * r3 has the faulting address - * r9 - r13 are saved in paca->exslb. - * r3 is saved in paca->slb_r3 - * We assume we aren't going to take any exceptions during this procedure. - */ -_GLOBAL(do_slb_miss) - mflr r10 - - stw r9,PACA_EXSLB+EX_CCR(r13) /* save CR in exc. frame */ - std r10,PACA_EXSLB+EX_LR(r13) /* save LR */ - - bl .slb_allocate /* handle it */ - - /* All done -- return from exception. */ - - ld r10,PACA_EXSLB+EX_LR(r13) - ld r3,PACA_EXSLB+EX_R3(r13) - lwz r9,PACA_EXSLB+EX_CCR(r13) /* get saved CR */ -#ifdef CONFIG_PPC_ISERIES - ld r11,PACALPPACA+LPPACASRR0(r13) /* get SRR0 value */ -#endif /* CONFIG_PPC_ISERIES */ - - mtlr r10 - - andi. r10,r12,MSR_RI /* check for unrecoverable exception */ - beq- unrecov_slb - -.machine push -.machine "power4" - mtcrf 0x80,r9 - mtcrf 0x01,r9 /* slb_allocate uses cr0 and cr7 */ -.machine pop - -#ifdef CONFIG_PPC_ISERIES - mtspr SPRN_SRR0,r11 - mtspr SPRN_SRR1,r12 -#endif /* CONFIG_PPC_ISERIES */ - ld r9,PACA_EXSLB+EX_R9(r13) - ld r10,PACA_EXSLB+EX_R10(r13) - ld r11,PACA_EXSLB+EX_R11(r13) - ld r12,PACA_EXSLB+EX_R12(r13) - ld r13,PACA_EXSLB+EX_R13(r13) - rfid - b . /* prevent speculative execution */ - -unrecov_slb: - EXCEPTION_PROLOG_COMMON(0x4100, PACA_EXSLB) - DISABLE_INTS - bl .save_nvgprs -1: addi r3,r1,STACK_FRAME_OVERHEAD - bl .unrecoverable_exception - b 1b - /* * Space for CPU0's segment table. * @@ -1569,7 +1712,10 @@ _GLOBAL(__secondary_start) #endif /* Initialize the first segment table (or SLB) entry */ ld r3,PACASTABVIRT(r13) /* get addr of segment table */ +BEGIN_FTR_SECTION bl .stab_initialize +END_FTR_SECTION_IFCLR(CPU_FTR_SLB) + bl .slb_initialize /* Initialize the kernel stack. Just a repeat for iSeries. */ LOADADDR(r3,current_set) diff --git a/arch/powerpc/kernel/lparmap.c b/arch/powerpc/kernel/lparmap.c index eded971d1bf9..5a05a797485f 100644 --- a/arch/powerpc/kernel/lparmap.c +++ b/arch/powerpc/kernel/lparmap.c @@ -25,7 +25,7 @@ const struct LparMap __attribute__((__section__(".text"))) xLparMap = { .xRanges = { { .xPages = HvPagesToMap, .xOffset = 0, - .xVPN = KERNEL_VSID(KERNELBASE) << (SID_SHIFT - PAGE_SHIFT), + .xVPN = KERNEL_VSID(KERNELBASE) << (SID_SHIFT - HW_PAGE_SHIFT), }, }, }; diff --git a/arch/powerpc/kernel/process.c b/arch/powerpc/kernel/process.c index 96843211cc5c..7f64f0464d44 100644 --- a/arch/powerpc/kernel/process.c +++ b/arch/powerpc/kernel/process.c @@ -554,12 +554,10 @@ int copy_thread(int nr, unsigned long clone_flags, unsigned long usp, #ifdef CONFIG_PPC64 if (cpu_has_feature(CPU_FTR_SLB)) { unsigned long sp_vsid = get_kernel_vsid(sp); + unsigned long llp = mmu_psize_defs[mmu_linear_psize].sllp; sp_vsid <<= SLB_VSID_SHIFT; - sp_vsid |= SLB_VSID_KERNEL; - if (cpu_has_feature(CPU_FTR_16M_PAGE)) - sp_vsid |= SLB_VSID_L; - + sp_vsid |= SLB_VSID_KERNEL | llp; p->thread.ksp_vsid = sp_vsid; } diff --git a/arch/powerpc/kernel/prom.c b/arch/powerpc/kernel/prom.c index eec2da695508..3675ef4bac90 100644 --- a/arch/powerpc/kernel/prom.c +++ b/arch/powerpc/kernel/prom.c @@ -724,10 +724,10 @@ static inline char *find_flat_dt_string(u32 offset) * used to extract the memory informations at boot before we can * unflatten the tree */ -static int __init scan_flat_dt(int (*it)(unsigned long node, - const char *uname, int depth, - void *data), - void *data) +int __init of_scan_flat_dt(int (*it)(unsigned long node, + const char *uname, int depth, + void *data), + void *data) { unsigned long p = ((unsigned long)initial_boot_params) + initial_boot_params->off_dt_struct; @@ -784,8 +784,8 @@ static int __init scan_flat_dt(int (*it)(unsigned long node, * This function can be used within scan_flattened_dt callback to get * access to properties */ -static void* __init get_flat_dt_prop(unsigned long node, const char *name, - unsigned long *size) +void* __init of_get_flat_dt_prop(unsigned long node, const char *name, + unsigned long *size) { unsigned long p = node; @@ -1087,7 +1087,7 @@ void __init unflatten_device_tree(void) static int __init early_init_dt_scan_cpus(unsigned long node, const char *uname, int depth, void *data) { - char *type = get_flat_dt_prop(node, "device_type", NULL); + char *type = of_get_flat_dt_prop(node, "device_type", NULL); u32 *prop; unsigned long size = 0; @@ -1095,19 +1095,6 @@ static int __init early_init_dt_scan_cpus(unsigned long node, if (type == NULL || strcmp(type, "cpu") != 0) return 0; -#ifdef CONFIG_PPC_PSERIES - /* On LPAR, look for the first ibm,pft-size property for the hash table size - */ - if (systemcfg->platform == PLATFORM_PSERIES_LPAR && ppc64_pft_size == 0) { - u32 *pft_size; - pft_size = get_flat_dt_prop(node, "ibm,pft-size", NULL); - if (pft_size != NULL) { - /* pft_size[0] is the NUMA CEC cookie */ - ppc64_pft_size = pft_size[1]; - } - } -#endif - boot_cpuid = 0; boot_cpuid_phys = 0; if (initial_boot_params && initial_boot_params->version >= 2) { @@ -1117,8 +1104,9 @@ static int __init early_init_dt_scan_cpus(unsigned long node, boot_cpuid_phys = initial_boot_params->boot_cpuid_phys; } else { /* Check if it's the boot-cpu, set it's hw index now */ - if (get_flat_dt_prop(node, "linux,boot-cpu", NULL) != NULL) { - prop = get_flat_dt_prop(node, "reg", NULL); + if (of_get_flat_dt_prop(node, + "linux,boot-cpu", NULL) != NULL) { + prop = of_get_flat_dt_prop(node, "reg", NULL); if (prop != NULL) boot_cpuid_phys = *prop; } @@ -1127,14 +1115,14 @@ static int __init early_init_dt_scan_cpus(unsigned long node, #ifdef CONFIG_ALTIVEC /* Check if we have a VMX and eventually update CPU features */ - prop = (u32 *)get_flat_dt_prop(node, "ibm,vmx", &size); + prop = (u32 *)of_get_flat_dt_prop(node, "ibm,vmx", &size); if (prop && (*prop) > 0) { cur_cpu_spec->cpu_features |= CPU_FTR_ALTIVEC; cur_cpu_spec->cpu_user_features |= PPC_FEATURE_HAS_ALTIVEC; } /* Same goes for Apple's "altivec" property */ - prop = (u32 *)get_flat_dt_prop(node, "altivec", NULL); + prop = (u32 *)of_get_flat_dt_prop(node, "altivec", NULL); if (prop) { cur_cpu_spec->cpu_features |= CPU_FTR_ALTIVEC; cur_cpu_spec->cpu_user_features |= PPC_FEATURE_HAS_ALTIVEC; @@ -1147,7 +1135,7 @@ static int __init early_init_dt_scan_cpus(unsigned long node, * this by looking at the size of the ibm,ppc-interrupt-server#s * property */ - prop = (u32 *)get_flat_dt_prop(node, "ibm,ppc-interrupt-server#s", + prop = (u32 *)of_get_flat_dt_prop(node, "ibm,ppc-interrupt-server#s", &size); cur_cpu_spec->cpu_features &= ~CPU_FTR_SMT; if (prop && ((size / sizeof(u32)) > 1)) @@ -1170,7 +1158,7 @@ static int __init early_init_dt_scan_chosen(unsigned long node, return 0; /* get platform type */ - prop = (u32 *)get_flat_dt_prop(node, "linux,platform", NULL); + prop = (u32 *)of_get_flat_dt_prop(node, "linux,platform", NULL); if (prop == NULL) return 0; #ifdef CONFIG_PPC64 @@ -1183,21 +1171,21 @@ static int __init early_init_dt_scan_chosen(unsigned long node, #ifdef CONFIG_PPC64 /* check if iommu is forced on or off */ - if (get_flat_dt_prop(node, "linux,iommu-off", NULL) != NULL) + if (of_get_flat_dt_prop(node, "linux,iommu-off", NULL) != NULL) iommu_is_off = 1; - if (get_flat_dt_prop(node, "linux,iommu-force-on", NULL) != NULL) + if (of_get_flat_dt_prop(node, "linux,iommu-force-on", NULL) != NULL) iommu_force_on = 1; #endif - lprop = get_flat_dt_prop(node, "linux,memory-limit", NULL); + lprop = of_get_flat_dt_prop(node, "linux,memory-limit", NULL); if (lprop) memory_limit = *lprop; #ifdef CONFIG_PPC64 - lprop = get_flat_dt_prop(node, "linux,tce-alloc-start", NULL); + lprop = of_get_flat_dt_prop(node, "linux,tce-alloc-start", NULL); if (lprop) tce_alloc_start = *lprop; - lprop = get_flat_dt_prop(node, "linux,tce-alloc-end", NULL); + lprop = of_get_flat_dt_prop(node, "linux,tce-alloc-end", NULL); if (lprop) tce_alloc_end = *lprop; #endif @@ -1209,9 +1197,9 @@ static int __init early_init_dt_scan_chosen(unsigned long node, { u64 *basep, *entryp; - basep = get_flat_dt_prop(node, "linux,rtas-base", NULL); - entryp = get_flat_dt_prop(node, "linux,rtas-entry", NULL); - prop = get_flat_dt_prop(node, "linux,rtas-size", NULL); + basep = of_get_flat_dt_prop(node, "linux,rtas-base", NULL); + entryp = of_get_flat_dt_prop(node, "linux,rtas-entry", NULL); + prop = of_get_flat_dt_prop(node, "linux,rtas-size", NULL); if (basep && entryp && prop) { rtas.base = *basep; rtas.entry = *entryp; @@ -1232,11 +1220,11 @@ static int __init early_init_dt_scan_root(unsigned long node, if (depth != 0) return 0; - prop = get_flat_dt_prop(node, "#size-cells", NULL); + prop = of_get_flat_dt_prop(node, "#size-cells", NULL); dt_root_size_cells = (prop == NULL) ? 1 : *prop; DBG("dt_root_size_cells = %x\n", dt_root_size_cells); - prop = get_flat_dt_prop(node, "#address-cells", NULL); + prop = of_get_flat_dt_prop(node, "#address-cells", NULL); dt_root_addr_cells = (prop == NULL) ? 2 : *prop; DBG("dt_root_addr_cells = %x\n", dt_root_addr_cells); @@ -1271,7 +1259,7 @@ static unsigned long __init dt_mem_next_cell(int s, cell_t **cellp) static int __init early_init_dt_scan_memory(unsigned long node, const char *uname, int depth, void *data) { - char *type = get_flat_dt_prop(node, "device_type", NULL); + char *type = of_get_flat_dt_prop(node, "device_type", NULL); cell_t *reg, *endp; unsigned long l; @@ -1279,7 +1267,7 @@ static int __init early_init_dt_scan_memory(unsigned long node, if (type == NULL || strcmp(type, "memory") != 0) return 0; - reg = (cell_t *)get_flat_dt_prop(node, "reg", &l); + reg = (cell_t *)of_get_flat_dt_prop(node, "reg", &l); if (reg == NULL) return 0; @@ -1343,12 +1331,12 @@ void __init early_init_devtree(void *params) * device-tree, including the platform type, initrd location and * size, TCE reserve, and more ... */ - scan_flat_dt(early_init_dt_scan_chosen, NULL); + of_scan_flat_dt(early_init_dt_scan_chosen, NULL); /* Scan memory nodes and rebuild LMBs */ lmb_init(); - scan_flat_dt(early_init_dt_scan_root, NULL); - scan_flat_dt(early_init_dt_scan_memory, NULL); + of_scan_flat_dt(early_init_dt_scan_root, NULL); + of_scan_flat_dt(early_init_dt_scan_memory, NULL); lmb_enforce_memory_limit(memory_limit); lmb_analyze(); #ifdef CONFIG_PPC64 @@ -1363,10 +1351,10 @@ void __init early_init_devtree(void *params) DBG("Scanning CPUs ...\n"); - /* Retreive hash table size from flattened tree plus other - * CPU related informations (altivec support, boot CPU ID, ...) + /* Retreive CPU related informations from the flat tree + * (altivec support, boot CPU ID, ...) */ - scan_flat_dt(early_init_dt_scan_cpus, NULL); + of_scan_flat_dt(early_init_dt_scan_cpus, NULL); DBG(" <- early_init_devtree()\n"); } diff --git a/arch/powerpc/kernel/setup_64.c b/arch/powerpc/kernel/setup_64.c index 6b52cce872be..b0994050024f 100644 --- a/arch/powerpc/kernel/setup_64.c +++ b/arch/powerpc/kernel/setup_64.c @@ -277,16 +277,21 @@ void __init early_setup(unsigned long dt_ptr) DBG("Found, Initializing memory management...\n"); /* - * Initialize stab / SLB management + * Initialize the MMU Hash table and create the linear mapping + * of memory. Has to be done before stab/slb initialization as + * this is currently where the page size encoding is obtained */ - if (!firmware_has_feature(FW_FEATURE_ISERIES)) - stab_initialize(lpaca->stab_real); + htab_initialize(); /* - * Initialize the MMU Hash table and create the linear mapping - * of memory + * Initialize stab / SLB management except on iSeries */ - htab_initialize(); + if (!firmware_has_feature(FW_FEATURE_ISERIES)) { + if (cpu_has_feature(CPU_FTR_SLB)) + slb_initialize(); + else + stab_initialize(lpaca->stab_real); + } DBG(" <- early_setup()\n"); } @@ -552,10 +557,12 @@ static void __init irqstack_early_init(void) * SLB misses on them. */ for_each_cpu(i) { - softirq_ctx[i] = (struct thread_info *)__va(lmb_alloc_base(THREAD_SIZE, - THREAD_SIZE, 0x10000000)); - hardirq_ctx[i] = (struct thread_info *)__va(lmb_alloc_base(THREAD_SIZE, - THREAD_SIZE, 0x10000000)); + softirq_ctx[i] = (struct thread_info *) + __va(lmb_alloc_base(THREAD_SIZE, + THREAD_SIZE, 0x10000000)); + hardirq_ctx[i] = (struct thread_info *) + __va(lmb_alloc_base(THREAD_SIZE, + THREAD_SIZE, 0x10000000)); } } #else @@ -583,8 +590,8 @@ static void __init emergency_stack_init(void) limit = min(0x10000000UL, lmb.rmo_size); for_each_cpu(i) - paca[i].emergency_sp = __va(lmb_alloc_base(PAGE_SIZE, 128, - limit)) + PAGE_SIZE; + paca[i].emergency_sp = + __va(lmb_alloc_base(HW_PAGE_SIZE, 128, limit)) + HW_PAGE_SIZE; } /* diff --git a/arch/powerpc/lib/copypage_64.S b/arch/powerpc/lib/copypage_64.S index 733d61618bbf..40523b140109 100644 --- a/arch/powerpc/lib/copypage_64.S +++ b/arch/powerpc/lib/copypage_64.S @@ -11,7 +11,7 @@ #include #include -_GLOBAL(copy_page) +_GLOBAL(copy_4K_page) std r31,-8(1) std r30,-16(1) std r29,-24(1) diff --git a/arch/powerpc/lib/copyuser_64.S b/arch/powerpc/lib/copyuser_64.S index a0b3fbbd6fb1..6d69ef39b7df 100644 --- a/arch/powerpc/lib/copyuser_64.S +++ b/arch/powerpc/lib/copyuser_64.S @@ -24,7 +24,7 @@ _GLOBAL(__copy_tofrom_user) std r4,-16(r1) std r5,-8(r1) dcbt 0,r4 - beq .Lcopy_page + beq .Lcopy_page_4K andi. r6,r6,7 mtcrf 0x01,r5 blt cr1,.Lshort_copy @@ -366,7 +366,7 @@ _GLOBAL(__copy_tofrom_user) * above (following the .Ldst_aligned label) but it runs slightly * slower on POWER3. */ -.Lcopy_page: +.Lcopy_page_4K: std r31,-32(1) std r30,-40(1) std r29,-48(1) diff --git a/arch/powerpc/mm/hash_low_64.S b/arch/powerpc/mm/hash_low_64.S index d6ed9102eeea..e0d02c4a2615 100644 --- a/arch/powerpc/mm/hash_low_64.S +++ b/arch/powerpc/mm/hash_low_64.S @@ -1,7 +1,7 @@ /* * ppc64 MMU hashtable management routines * - * (c) Copyright IBM Corp. 2003 + * (c) Copyright IBM Corp. 2003, 2005 * * Maintained by: Benjamin Herrenschmidt * @@ -10,6 +10,7 @@ * described in the kernel's COPYING file. */ +#include #include #include #include @@ -42,14 +43,24 @@ /* Save non-volatile offsets */ #define STK_REG(i) (112 + ((i)-14)*8) + +#ifndef CONFIG_PPC_64K_PAGES + +/***************************************************************************** + * * + * 4K SW & 4K HW pages implementation * + * * + *****************************************************************************/ + + /* - * _hash_page(unsigned long ea, unsigned long access, unsigned long vsid, - * pte_t *ptep, unsigned long trap, int local) + * _hash_page_4K(unsigned long ea, unsigned long access, unsigned long vsid, + * pte_t *ptep, unsigned long trap, int local) * - * Adds a page to the hash table. This is the non-LPAR version for now + * Adds a 4K page to the hash table in a segment of 4K pages only */ -_GLOBAL(__hash_page) +_GLOBAL(__hash_page_4K) mflr r0 std r0,16(r1) stdu r1,-STACKFRAMESIZE(r1) @@ -88,7 +99,8 @@ _GLOBAL(__hash_page) /* If so, just bail out and refault if needed. Someone else * is changing this PTE anyway and might hash it. */ - bne- bail_ok + bne- htab_bail_ok + /* Prepare new PTE value (turn access RW into DIRTY, then * add BUSY,HASHPTE and ACCESSED) */ @@ -118,10 +130,10 @@ _GLOBAL(__hash_page) /* Convert linux PTE bits into HW equivalents */ andi. r3,r30,0x1fe /* Get basic set of flags */ - xori r3,r3,HW_NO_EXEC /* _PAGE_EXEC -> NOEXEC */ + xori r3,r3,HPTE_R_N /* _PAGE_EXEC -> NOEXEC */ rlwinm r0,r30,32-9+1,30,30 /* _PAGE_RW -> _PAGE_USER (r0) */ rlwinm r4,r30,32-7+1,30,30 /* _PAGE_DIRTY -> _PAGE_USER (r4) */ - and r0,r0,r4 /* _PAGE_RW & _PAGE_DIRTY -> r0 bit 30 */ + and r0,r0,r4 /* _PAGE_RW & _PAGE_DIRTY ->r0 bit 30*/ andc r0,r30,r0 /* r0 = pte & ~r0 */ rlwimi r3,r0,32-1,31,31 /* Insert result into PP lsb */ @@ -158,19 +170,21 @@ htab_insert_pte: andc r30,r30,r0 ori r30,r30,_PAGE_HASHPTE - /* page number in r5 */ - rldicl r5,r31,64-PTE_SHIFT,PTE_SHIFT + /* physical address r5 */ + rldicl r5,r31,64-PTE_RPN_SHIFT,PTE_RPN_SHIFT + sldi r5,r5,PAGE_SHIFT /* Calculate primary group hash */ and r0,r28,r27 - rldicr r3,r0,3,63-3 /* r0 = (hash & mask) << 3 */ + rldicr r3,r0,3,63-3 /* r3 = (hash & mask) << 3 */ /* Call ppc_md.hpte_insert */ - ld r7,STK_PARM(r4)(r1) /* Retreive new pp bits */ + ld r6,STK_PARM(r4)(r1) /* Retreive new pp bits */ mr r4,r29 /* Retreive va */ - li r6,0 /* no vflags */ + li r7,0 /* !bolted, !secondary */ + li r8,MMU_PAGE_4K /* page size */ _GLOBAL(htab_call_hpte_insert1) - bl . /* Will be patched by htab_finish_init() */ + bl . /* Patched by htab_finish_init() */ cmpdi 0,r3,0 bge htab_pte_insert_ok /* Insertion successful */ cmpdi 0,r3,-2 /* Critical failure */ @@ -178,19 +192,21 @@ _GLOBAL(htab_call_hpte_insert1) /* Now try secondary slot */ - /* page number in r5 */ - rldicl r5,r31,64-PTE_SHIFT,PTE_SHIFT + /* physical address r5 */ + rldicl r5,r31,64-PTE_RPN_SHIFT,PTE_RPN_SHIFT + sldi r5,r5,PAGE_SHIFT /* Calculate secondary group hash */ andc r0,r27,r28 rldicr r3,r0,3,63-3 /* r0 = (~hash & mask) << 3 */ /* Call ppc_md.hpte_insert */ - ld r7,STK_PARM(r4)(r1) /* Retreive new pp bits */ + ld r6,STK_PARM(r4)(r1) /* Retreive new pp bits */ mr r4,r29 /* Retreive va */ - li r6,HPTE_V_SECONDARY@l /* secondary slot */ + li r7,HPTE_V_SECONDARY /* !bolted, secondary */ + li r8,MMU_PAGE_4K /* page size */ _GLOBAL(htab_call_hpte_insert2) - bl . /* Will be patched by htab_finish_init() */ + bl . /* Patched by htab_finish_init() */ cmpdi 0,r3,0 bge+ htab_pte_insert_ok /* Insertion successful */ cmpdi 0,r3,-2 /* Critical failure */ @@ -207,14 +223,14 @@ _GLOBAL(htab_call_hpte_insert2) rldicr r3,r0,3,63-3 /* r0 = (hash & mask) << 3 */ /* Call ppc_md.hpte_remove */ _GLOBAL(htab_call_hpte_remove) - bl . /* Will be patched by htab_finish_init() */ + bl . /* Patched by htab_finish_init() */ /* Try all again */ b htab_insert_pte -bail_ok: +htab_bail_ok: li r3,0 - b bail + b htab_bail htab_pte_insert_ok: /* Insert slot number & secondary bit in PTE */ @@ -227,7 +243,7 @@ htab_write_out_pte: ld r6,STK_PARM(r6)(r1) std r30,0(r6) li r3, 0 -bail: +htab_bail: ld r27,STK_REG(r27)(r1) ld r28,STK_REG(r28)(r1) ld r29,STK_REG(r29)(r1) @@ -256,10 +272,10 @@ htab_modify_pte: /* Call ppc_md.hpte_updatepp */ mr r5,r29 /* va */ - li r6,0 /* large is 0 */ + li r6,MMU_PAGE_4K /* page size */ ld r7,STK_PARM(r8)(r1) /* get "local" param */ _GLOBAL(htab_call_hpte_updatepp) - bl . /* Will be patched by htab_finish_init() */ + bl . /* Patched by htab_finish_init() */ /* if we failed because typically the HPTE wasn't really here * we try an insertion. @@ -276,13 +292,556 @@ htab_wrong_access: /* Bail out clearing reservation */ stdcx. r31,0,r6 li r3,1 - b bail + b htab_bail + +htab_pte_insert_failure: + /* Bail out restoring old PTE */ + ld r6,STK_PARM(r6)(r1) + std r31,0(r6) + li r3,-1 + b htab_bail + + +#else /* CONFIG_PPC_64K_PAGES */ + + +/***************************************************************************** + * * + * 64K SW & 4K or 64K HW in a 4K segment pages implementation * + * * + *****************************************************************************/ + +/* _hash_page_4K(unsigned long ea, unsigned long access, unsigned long vsid, + * pte_t *ptep, unsigned long trap, int local) + */ + +/* + * For now, we do NOT implement Admixed pages + */ +_GLOBAL(__hash_page_4K) + mflr r0 + std r0,16(r1) + stdu r1,-STACKFRAMESIZE(r1) + /* Save all params that we need after a function call */ + std r6,STK_PARM(r6)(r1) + std r8,STK_PARM(r8)(r1) + + /* Add _PAGE_PRESENT to access */ + ori r4,r4,_PAGE_PRESENT + + /* Save non-volatile registers. + * r31 will hold "old PTE" + * r30 is "new PTE" + * r29 is "va" + * r28 is a hash value + * r27 is hashtab mask (maybe dynamic patched instead ?) + * r26 is the hidx mask + * r25 is the index in combo page + */ + std r25,STK_REG(r25)(r1) + std r26,STK_REG(r26)(r1) + std r27,STK_REG(r27)(r1) + std r28,STK_REG(r28)(r1) + std r29,STK_REG(r29)(r1) + std r30,STK_REG(r30)(r1) + std r31,STK_REG(r31)(r1) + + /* Step 1: + * + * Check permissions, atomically mark the linux PTE busy + * and hashed. + */ +1: + ldarx r31,0,r6 + /* Check access rights (access & ~(pte_val(*ptep))) */ + andc. r0,r4,r31 + bne- htab_wrong_access + /* Check if PTE is busy */ + andi. r0,r31,_PAGE_BUSY + /* If so, just bail out and refault if needed. Someone else + * is changing this PTE anyway and might hash it. + */ + bne- htab_bail_ok + /* Prepare new PTE value (turn access RW into DIRTY, then + * add BUSY and ACCESSED) + */ + rlwinm r30,r4,32-9+7,31-7,31-7 /* _PAGE_RW -> _PAGE_DIRTY */ + or r30,r30,r31 + ori r30,r30,_PAGE_BUSY | _PAGE_ACCESSED | _PAGE_HASHPTE + /* Write the linux PTE atomically (setting busy) */ + stdcx. r30,0,r6 + bne- 1b + isync + + /* Step 2: + * + * Insert/Update the HPTE in the hash table. At this point, + * r4 (access) is re-useable, we use it for the new HPTE flags + */ + + /* Load the hidx index */ + rldicl r25,r3,64-12,60 + + /* Calc va and put it in r29 */ + rldicr r29,r5,28,63-28 /* r29 = (vsid << 28) */ + rldicl r3,r3,0,36 /* r3 = (ea & 0x0fffffff) */ + or r29,r3,r29 /* r29 = va + + /* Calculate hash value for primary slot and store it in r28 */ + rldicl r5,r5,0,25 /* vsid & 0x0000007fffffffff */ + rldicl r0,r3,64-12,48 /* (ea >> 12) & 0xffff */ + xor r28,r5,r0 + + /* Convert linux PTE bits into HW equivalents */ + andi. r3,r30,0x1fe /* Get basic set of flags */ + xori r3,r3,HPTE_R_N /* _PAGE_EXEC -> NOEXEC */ + rlwinm r0,r30,32-9+1,30,30 /* _PAGE_RW -> _PAGE_USER (r0) */ + rlwinm r4,r30,32-7+1,30,30 /* _PAGE_DIRTY -> _PAGE_USER (r4) */ + and r0,r0,r4 /* _PAGE_RW & _PAGE_DIRTY ->r0 bit 30*/ + andc r0,r30,r0 /* r0 = pte & ~r0 */ + rlwimi r3,r0,32-1,31,31 /* Insert result into PP lsb */ + + /* We eventually do the icache sync here (maybe inline that + * code rather than call a C function...) + */ +BEGIN_FTR_SECTION + mr r4,r30 + mr r5,r7 + bl .hash_page_do_lazy_icache +END_FTR_SECTION(CPU_FTR_NOEXECUTE|CPU_FTR_COHERENT_ICACHE, CPU_FTR_NOEXECUTE) + + /* At this point, r3 contains new PP bits, save them in + * place of "access" in the param area (sic) + */ + std r3,STK_PARM(r4)(r1) + + /* Get htab_hash_mask */ + ld r4,htab_hash_mask@got(2) + ld r27,0(r4) /* htab_hash_mask -> r27 */ + + /* Check if we may already be in the hashtable, in this case, we + * go to out-of-line code to try to modify the HPTE. We look for + * the bit at (1 >> (index + 32)) + */ + andi. r0,r31,_PAGE_HASHPTE + li r26,0 /* Default hidx */ + beq htab_insert_pte + ld r6,STK_PARM(r6)(r1) + ori r26,r6,0x8000 /* Load the hidx mask */ + ld r26,0(r26) + addi r5,r25,36 /* Check actual HPTE_SUB bit, this */ + rldcr. r0,r31,r5,0 /* must match pgtable.h definition */ + bne htab_modify_pte + +htab_insert_pte: + /* real page number in r5, PTE RPN value + index */ + rldicl r5,r31,64-PTE_RPN_SHIFT,PTE_RPN_SHIFT + sldi r5,r5,PAGE_SHIFT-HW_PAGE_SHIFT + add r5,r5,r25 + sldi r5,r5,HW_PAGE_SHIFT + + /* Calculate primary group hash */ + and r0,r28,r27 + rldicr r3,r0,3,63-3 /* r0 = (hash & mask) << 3 */ + + /* Call ppc_md.hpte_insert */ + ld r6,STK_PARM(r4)(r1) /* Retreive new pp bits */ + mr r4,r29 /* Retreive va */ + li r7,0 /* !bolted, !secondary */ + li r8,MMU_PAGE_4K /* page size */ +_GLOBAL(htab_call_hpte_insert1) + bl . /* patched by htab_finish_init() */ + cmpdi 0,r3,0 + bge htab_pte_insert_ok /* Insertion successful */ + cmpdi 0,r3,-2 /* Critical failure */ + beq- htab_pte_insert_failure + + /* Now try secondary slot */ + + /* real page number in r5, PTE RPN value + index */ + rldicl r5,r31,64-PTE_RPN_SHIFT,PTE_RPN_SHIFT + sldi r5,r5,PAGE_SHIFT-HW_PAGE_SHIFT + add r5,r5,r25 + sldi r5,r5,HW_PAGE_SHIFT + + /* Calculate secondary group hash */ + andc r0,r27,r28 + rldicr r3,r0,3,63-3 /* r0 = (~hash & mask) << 3 */ + + /* Call ppc_md.hpte_insert */ + ld r6,STK_PARM(r4)(r1) /* Retreive new pp bits */ + mr r4,r29 /* Retreive va */ + li r7,HPTE_V_SECONDARY /* !bolted, secondary */ + li r8,MMU_PAGE_4K /* page size */ +_GLOBAL(htab_call_hpte_insert2) + bl . /* patched by htab_finish_init() */ + cmpdi 0,r3,0 + bge+ htab_pte_insert_ok /* Insertion successful */ + cmpdi 0,r3,-2 /* Critical failure */ + beq- htab_pte_insert_failure + + /* Both are full, we need to evict something */ + mftb r0 + /* Pick a random group based on TB */ + andi. r0,r0,1 + mr r5,r28 + bne 2f + not r5,r5 +2: and r0,r5,r27 + rldicr r3,r0,3,63-3 /* r0 = (hash & mask) << 3 */ + /* Call ppc_md.hpte_remove */ +_GLOBAL(htab_call_hpte_remove) + bl . /* patched by htab_finish_init() */ + + /* Try all again */ + b htab_insert_pte + +htab_bail_ok: + li r3,0 + b htab_bail + +htab_pte_insert_ok: + /* Insert slot number & secondary bit in PTE second half, + * clear _PAGE_BUSY and set approriate HPTE slot bit + */ + ld r6,STK_PARM(r6)(r1) + li r0,_PAGE_BUSY + andc r30,r30,r0 + /* HPTE SUB bit */ + li r0,1 + subfic r5,r25,27 /* Must match bit position in */ + sld r0,r0,r5 /* pgtable.h */ + or r30,r30,r0 + /* hindx */ + sldi r5,r25,2 + sld r3,r3,r5 + li r4,0xf + sld r4,r4,r5 + andc r26,r26,r4 + or r26,r26,r3 + ori r5,r6,0x8000 + std r26,0(r5) + lwsync + std r30,0(r6) + li r3, 0 +htab_bail: + ld r25,STK_REG(r25)(r1) + ld r26,STK_REG(r26)(r1) + ld r27,STK_REG(r27)(r1) + ld r28,STK_REG(r28)(r1) + ld r29,STK_REG(r29)(r1) + ld r30,STK_REG(r30)(r1) + ld r31,STK_REG(r31)(r1) + addi r1,r1,STACKFRAMESIZE + ld r0,16(r1) + mtlr r0 + blr + +htab_modify_pte: + /* Keep PP bits in r4 and slot idx from the PTE around in r3 */ + mr r4,r3 + sldi r5,r25,2 + srd r3,r26,r5 + + /* Secondary group ? if yes, get a inverted hash value */ + mr r5,r28 + andi. r0,r3,0x8 /* page secondary ? */ + beq 1f + not r5,r5 +1: andi. r3,r3,0x7 /* extract idx alone */ + + /* Calculate proper slot value for ppc_md.hpte_updatepp */ + and r0,r5,r27 + rldicr r0,r0,3,63-3 /* r0 = (hash & mask) << 3 */ + add r3,r0,r3 /* add slot idx */ + + /* Call ppc_md.hpte_updatepp */ + mr r5,r29 /* va */ + li r6,MMU_PAGE_4K /* page size */ + ld r7,STK_PARM(r8)(r1) /* get "local" param */ +_GLOBAL(htab_call_hpte_updatepp) + bl . /* patched by htab_finish_init() */ + + /* if we failed because typically the HPTE wasn't really here + * we try an insertion. + */ + cmpdi 0,r3,-1 + beq- htab_insert_pte + + /* Clear the BUSY bit and Write out the PTE */ + li r0,_PAGE_BUSY + andc r30,r30,r0 + ld r6,STK_PARM(r6)(r1) + std r30,0(r6) + li r3,0 + b htab_bail + +htab_wrong_access: + /* Bail out clearing reservation */ + stdcx. r31,0,r6 + li r3,1 + b htab_bail htab_pte_insert_failure: /* Bail out restoring old PTE */ ld r6,STK_PARM(r6)(r1) std r31,0(r6) li r3,-1 - b bail + b htab_bail + + +/***************************************************************************** + * * + * 64K SW & 64K HW in a 64K segment pages implementation * + * * + *****************************************************************************/ + +_GLOBAL(__hash_page_64K) + mflr r0 + std r0,16(r1) + stdu r1,-STACKFRAMESIZE(r1) + /* Save all params that we need after a function call */ + std r6,STK_PARM(r6)(r1) + std r8,STK_PARM(r8)(r1) + + /* Add _PAGE_PRESENT to access */ + ori r4,r4,_PAGE_PRESENT + + /* Save non-volatile registers. + * r31 will hold "old PTE" + * r30 is "new PTE" + * r29 is "va" + * r28 is a hash value + * r27 is hashtab mask (maybe dynamic patched instead ?) + */ + std r27,STK_REG(r27)(r1) + std r28,STK_REG(r28)(r1) + std r29,STK_REG(r29)(r1) + std r30,STK_REG(r30)(r1) + std r31,STK_REG(r31)(r1) + + /* Step 1: + * + * Check permissions, atomically mark the linux PTE busy + * and hashed. + */ +1: + ldarx r31,0,r6 + /* Check access rights (access & ~(pte_val(*ptep))) */ + andc. r0,r4,r31 + bne- ht64_wrong_access + /* Check if PTE is busy */ + andi. r0,r31,_PAGE_BUSY + /* If so, just bail out and refault if needed. Someone else + * is changing this PTE anyway and might hash it. + */ + bne- ht64_bail_ok + /* Prepare new PTE value (turn access RW into DIRTY, then + * add BUSY,HASHPTE and ACCESSED) + */ + rlwinm r30,r4,32-9+7,31-7,31-7 /* _PAGE_RW -> _PAGE_DIRTY */ + or r30,r30,r31 + ori r30,r30,_PAGE_BUSY | _PAGE_ACCESSED | _PAGE_HASHPTE + /* Write the linux PTE atomically (setting busy) */ + stdcx. r30,0,r6 + bne- 1b + isync + + /* Step 2: + * + * Insert/Update the HPTE in the hash table. At this point, + * r4 (access) is re-useable, we use it for the new HPTE flags + */ + + /* Calc va and put it in r29 */ + rldicr r29,r5,28,63-28 + rldicl r3,r3,0,36 + or r29,r3,r29 + + /* Calculate hash value for primary slot and store it in r28 */ + rldicl r5,r5,0,25 /* vsid & 0x0000007fffffffff */ + rldicl r0,r3,64-16,52 /* (ea >> 16) & 0xfff */ + xor r28,r5,r0 + + /* Convert linux PTE bits into HW equivalents */ + andi. r3,r30,0x1fe /* Get basic set of flags */ + xori r3,r3,HPTE_R_N /* _PAGE_EXEC -> NOEXEC */ + rlwinm r0,r30,32-9+1,30,30 /* _PAGE_RW -> _PAGE_USER (r0) */ + rlwinm r4,r30,32-7+1,30,30 /* _PAGE_DIRTY -> _PAGE_USER (r4) */ + and r0,r0,r4 /* _PAGE_RW & _PAGE_DIRTY ->r0 bit 30*/ + andc r0,r30,r0 /* r0 = pte & ~r0 */ + rlwimi r3,r0,32-1,31,31 /* Insert result into PP lsb */ + + /* We eventually do the icache sync here (maybe inline that + * code rather than call a C function...) + */ +BEGIN_FTR_SECTION + mr r4,r30 + mr r5,r7 + bl .hash_page_do_lazy_icache +END_FTR_SECTION(CPU_FTR_NOEXECUTE|CPU_FTR_COHERENT_ICACHE, CPU_FTR_NOEXECUTE) + + /* At this point, r3 contains new PP bits, save them in + * place of "access" in the param area (sic) + */ + std r3,STK_PARM(r4)(r1) + + /* Get htab_hash_mask */ + ld r4,htab_hash_mask@got(2) + ld r27,0(r4) /* htab_hash_mask -> r27 */ + + /* Check if we may already be in the hashtable, in this case, we + * go to out-of-line code to try to modify the HPTE + */ + andi. r0,r31,_PAGE_HASHPTE + bne ht64_modify_pte + +ht64_insert_pte: + /* Clear hpte bits in new pte (we also clear BUSY btw) and + * add _PAGE_HASHPTE + */ + lis r0,_PAGE_HPTEFLAGS@h + ori r0,r0,_PAGE_HPTEFLAGS@l + andc r30,r30,r0 + ori r30,r30,_PAGE_HASHPTE + + /* Phyical address in r5 */ + rldicl r5,r31,64-PTE_RPN_SHIFT,PTE_RPN_SHIFT + sldi r5,r5,PAGE_SHIFT + + /* Calculate primary group hash */ + and r0,r28,r27 + rldicr r3,r0,3,63-3 /* r0 = (hash & mask) << 3 */ + + /* Call ppc_md.hpte_insert */ + ld r6,STK_PARM(r4)(r1) /* Retreive new pp bits */ + mr r4,r29 /* Retreive va */ + li r7,0 /* !bolted, !secondary */ + li r8,MMU_PAGE_64K +_GLOBAL(ht64_call_hpte_insert1) + bl . /* patched by htab_finish_init() */ + cmpdi 0,r3,0 + bge ht64_pte_insert_ok /* Insertion successful */ + cmpdi 0,r3,-2 /* Critical failure */ + beq- ht64_pte_insert_failure + + /* Now try secondary slot */ + + /* Phyical address in r5 */ + rldicl r5,r31,64-PTE_RPN_SHIFT,PTE_RPN_SHIFT + sldi r5,r5,PAGE_SHIFT + + /* Calculate secondary group hash */ + andc r0,r27,r28 + rldicr r3,r0,3,63-3 /* r0 = (~hash & mask) << 3 */ + + /* Call ppc_md.hpte_insert */ + ld r6,STK_PARM(r4)(r1) /* Retreive new pp bits */ + mr r4,r29 /* Retreive va */ + li r7,HPTE_V_SECONDARY /* !bolted, secondary */ + li r8,MMU_PAGE_64K +_GLOBAL(ht64_call_hpte_insert2) + bl . /* patched by htab_finish_init() */ + cmpdi 0,r3,0 + bge+ ht64_pte_insert_ok /* Insertion successful */ + cmpdi 0,r3,-2 /* Critical failure */ + beq- ht64_pte_insert_failure + + /* Both are full, we need to evict something */ + mftb r0 + /* Pick a random group based on TB */ + andi. r0,r0,1 + mr r5,r28 + bne 2f + not r5,r5 +2: and r0,r5,r27 + rldicr r3,r0,3,63-3 /* r0 = (hash & mask) << 3 */ + /* Call ppc_md.hpte_remove */ +_GLOBAL(ht64_call_hpte_remove) + bl . /* patched by htab_finish_init() */ + + /* Try all again */ + b ht64_insert_pte + +ht64_bail_ok: + li r3,0 + b ht64_bail + +ht64_pte_insert_ok: + /* Insert slot number & secondary bit in PTE */ + rldimi r30,r3,12,63-15 + + /* Write out the PTE with a normal write + * (maybe add eieio may be good still ?) + */ +ht64_write_out_pte: + ld r6,STK_PARM(r6)(r1) + std r30,0(r6) + li r3, 0 +ht64_bail: + ld r27,STK_REG(r27)(r1) + ld r28,STK_REG(r28)(r1) + ld r29,STK_REG(r29)(r1) + ld r30,STK_REG(r30)(r1) + ld r31,STK_REG(r31)(r1) + addi r1,r1,STACKFRAMESIZE + ld r0,16(r1) + mtlr r0 + blr + +ht64_modify_pte: + /* Keep PP bits in r4 and slot idx from the PTE around in r3 */ + mr r4,r3 + rlwinm r3,r31,32-12,29,31 + + /* Secondary group ? if yes, get a inverted hash value */ + mr r5,r28 + andi. r0,r31,_PAGE_F_SECOND + beq 1f + not r5,r5 +1: + /* Calculate proper slot value for ppc_md.hpte_updatepp */ + and r0,r5,r27 + rldicr r0,r0,3,63-3 /* r0 = (hash & mask) << 3 */ + add r3,r0,r3 /* add slot idx */ + + /* Call ppc_md.hpte_updatepp */ + mr r5,r29 /* va */ + li r6,MMU_PAGE_64K + ld r7,STK_PARM(r8)(r1) /* get "local" param */ +_GLOBAL(ht64_call_hpte_updatepp) + bl . /* patched by htab_finish_init() */ + + /* if we failed because typically the HPTE wasn't really here + * we try an insertion. + */ + cmpdi 0,r3,-1 + beq- ht64_insert_pte + + /* Clear the BUSY bit and Write out the PTE */ + li r0,_PAGE_BUSY + andc r30,r30,r0 + b ht64_write_out_pte + +ht64_wrong_access: + /* Bail out clearing reservation */ + stdcx. r31,0,r6 + li r3,1 + b ht64_bail + +ht64_pte_insert_failure: + /* Bail out restoring old PTE */ + ld r6,STK_PARM(r6)(r1) + std r31,0(r6) + li r3,-1 + b ht64_bail + + +#endif /* CONFIG_PPC_64K_PAGES */ +/***************************************************************************** + * * + * Huge pages implementation is in hugetlbpage.c * + * * + *****************************************************************************/ diff --git a/arch/powerpc/mm/hash_native_64.c b/arch/powerpc/mm/hash_native_64.c index 174d14576c28..d96bcfe4c6f6 100644 --- a/arch/powerpc/mm/hash_native_64.c +++ b/arch/powerpc/mm/hash_native_64.c @@ -9,6 +9,9 @@ * as published by the Free Software Foundation; either version * 2 of the License, or (at your option) any later version. */ + +#undef DEBUG_LOW + #include #include #include @@ -22,11 +25,84 @@ #include #include #include +#include + +#ifdef DEBUG_LOW +#define DBG_LOW(fmt...) udbg_printf(fmt) +#else +#define DBG_LOW(fmt...) +#endif #define HPTE_LOCK_BIT 3 static DEFINE_SPINLOCK(native_tlbie_lock); +static inline void __tlbie(unsigned long va, unsigned int psize) +{ + unsigned int penc; + + /* clear top 16 bits, non SLS segment */ + va &= ~(0xffffULL << 48); + + switch (psize) { + case MMU_PAGE_4K: + va &= ~0xffful; + asm volatile("tlbie %0,0" : : "r" (va) : "memory"); + break; + default: + penc = mmu_psize_defs[psize].penc; + va &= ~((1ul << mmu_psize_defs[psize].shift) - 1); + va |= (0x7f >> (8 - penc)) << 12; + asm volatile("tlbie %0,1" : : "r" (va) : "memory"); + break; + } +} + +static inline void __tlbiel(unsigned long va, unsigned int psize) +{ + unsigned int penc; + + /* clear top 16 bits, non SLS segment */ + va &= ~(0xffffULL << 48); + + switch (psize) { + case MMU_PAGE_4K: + va &= ~0xffful; + asm volatile(".long 0x7c000224 | (%0 << 11) | (0 << 21)" + : : "r"(va) : "memory"); + break; + default: + penc = mmu_psize_defs[psize].penc; + va &= ~((1ul << mmu_psize_defs[psize].shift) - 1); + va |= (0x7f >> (8 - penc)) << 12; + asm volatile(".long 0x7c000224 | (%0 << 11) | (1 << 21)" + : : "r"(va) : "memory"); + break; + } + +} + +static inline void tlbie(unsigned long va, int psize, int local) +{ + unsigned int use_local = local && cpu_has_feature(CPU_FTR_TLBIEL); + int lock_tlbie = !cpu_has_feature(CPU_FTR_LOCKLESS_TLBIE); + + if (use_local) + use_local = mmu_psize_defs[psize].tlbiel; + if (lock_tlbie && !use_local) + spin_lock(&native_tlbie_lock); + asm volatile("ptesync": : :"memory"); + if (use_local) { + __tlbiel(va, psize); + asm volatile("ptesync": : :"memory"); + } else { + __tlbie(va, psize); + asm volatile("eieio; tlbsync; ptesync": : :"memory"); + } + if (lock_tlbie && !use_local) + spin_unlock(&native_tlbie_lock); +} + static inline void native_lock_hpte(hpte_t *hptep) { unsigned long *word = &hptep->v; @@ -48,13 +124,19 @@ static inline void native_unlock_hpte(hpte_t *hptep) } long native_hpte_insert(unsigned long hpte_group, unsigned long va, - unsigned long prpn, unsigned long vflags, - unsigned long rflags) + unsigned long pa, unsigned long rflags, + unsigned long vflags, int psize) { hpte_t *hptep = htab_address + hpte_group; unsigned long hpte_v, hpte_r; int i; + if (!(vflags & HPTE_V_BOLTED)) { + DBG_LOW(" insert(group=%lx, va=%016lx, pa=%016lx," + " rflags=%lx, vflags=%lx, psize=%d)\n", + hpte_group, va, pa, rflags, vflags, psize); + } + for (i = 0; i < HPTES_PER_GROUP; i++) { if (! (hptep->v & HPTE_V_VALID)) { /* retry with lock held */ @@ -70,10 +152,13 @@ long native_hpte_insert(unsigned long hpte_group, unsigned long va, if (i == HPTES_PER_GROUP) return -1; - hpte_v = (va >> 23) << HPTE_V_AVPN_SHIFT | vflags | HPTE_V_VALID; - if (vflags & HPTE_V_LARGE) - va &= ~(1UL << HPTE_V_AVPN_SHIFT); - hpte_r = (prpn << HPTE_R_RPN_SHIFT) | rflags; + hpte_v = hpte_encode_v(va, psize) | vflags | HPTE_V_VALID; + hpte_r = hpte_encode_r(pa, psize) | rflags; + + if (!(vflags & HPTE_V_BOLTED)) { + DBG_LOW(" i=%x hpte_v=%016lx, hpte_r=%016lx\n", + i, hpte_v, hpte_r); + } hptep->r = hpte_r; /* Guarantee the second dword is visible before the valid bit */ @@ -96,6 +181,8 @@ static long native_hpte_remove(unsigned long hpte_group) int slot_offset; unsigned long hpte_v; + DBG_LOW(" remove(group=%lx)\n", hpte_group); + /* pick a random entry to start at */ slot_offset = mftb() & 0x7; @@ -126,34 +213,51 @@ static long native_hpte_remove(unsigned long hpte_group) return i; } -static inline void set_pp_bit(unsigned long pp, hpte_t *addr) +static long native_hpte_updatepp(unsigned long slot, unsigned long newpp, + unsigned long va, int psize, int local) { - unsigned long old; - unsigned long *p = &addr->r; - - __asm__ __volatile__( - "1: ldarx %0,0,%3\n\ - rldimi %0,%2,0,61\n\ - stdcx. %0,0,%3\n\ - bne 1b" - : "=&r" (old), "=m" (*p) - : "r" (pp), "r" (p), "m" (*p) - : "cc"); + hpte_t *hptep = htab_address + slot; + unsigned long hpte_v, want_v; + int ret = 0; + + want_v = hpte_encode_v(va, psize); + + DBG_LOW(" update(va=%016lx, avpnv=%016lx, hash=%016lx, newpp=%x)", + va, want_v & HPTE_V_AVPN, slot, newpp); + + native_lock_hpte(hptep); + + hpte_v = hptep->v; + + /* Even if we miss, we need to invalidate the TLB */ + if (!HPTE_V_COMPARE(hpte_v, want_v) || !(hpte_v & HPTE_V_VALID)) { + DBG_LOW(" -> miss\n"); + native_unlock_hpte(hptep); + ret = -1; + } else { + DBG_LOW(" -> hit\n"); + /* Update the HPTE */ + hptep->r = (hptep->r & ~(HPTE_R_PP | HPTE_R_N)) | + (newpp & (HPTE_R_PP | HPTE_R_N)); + native_unlock_hpte(hptep); + } + + /* Ensure it is out of the tlb too. */ + tlbie(va, psize, local); + + return ret; } -/* - * Only works on small pages. Yes its ugly to have to check each slot in - * the group but we only use this during bootup. - */ -static long native_hpte_find(unsigned long vpn) +static long native_hpte_find(unsigned long va, int psize) { hpte_t *hptep; unsigned long hash; unsigned long i, j; long slot; - unsigned long hpte_v; + unsigned long want_v, hpte_v; - hash = hpt_hash(vpn, 0); + hash = hpt_hash(va, mmu_psize_defs[psize].shift); + want_v = hpte_encode_v(va, psize); for (j = 0; j < 2; j++) { slot = (hash & htab_hash_mask) * HPTES_PER_GROUP; @@ -161,7 +265,7 @@ static long native_hpte_find(unsigned long vpn) hptep = htab_address + slot; hpte_v = hptep->v; - if ((HPTE_V_AVPN_VAL(hpte_v) == (vpn >> 11)) + if (HPTE_V_COMPARE(hpte_v, want_v) && (hpte_v & HPTE_V_VALID) && ( !!(hpte_v & HPTE_V_SECONDARY) == j)) { /* HPTE matches */ @@ -177,120 +281,92 @@ static long native_hpte_find(unsigned long vpn) return -1; } -static long native_hpte_updatepp(unsigned long slot, unsigned long newpp, - unsigned long va, int large, int local) -{ - hpte_t *hptep = htab_address + slot; - unsigned long hpte_v; - unsigned long avpn = va >> 23; - int ret = 0; - - if (large) - avpn &= ~1; - - native_lock_hpte(hptep); - - hpte_v = hptep->v; - - /* Even if we miss, we need to invalidate the TLB */ - if ((HPTE_V_AVPN_VAL(hpte_v) != avpn) - || !(hpte_v & HPTE_V_VALID)) { - native_unlock_hpte(hptep); - ret = -1; - } else { - set_pp_bit(newpp, hptep); - native_unlock_hpte(hptep); - } - - /* Ensure it is out of the tlb too */ - if (cpu_has_feature(CPU_FTR_TLBIEL) && !large && local) { - tlbiel(va); - } else { - int lock_tlbie = !cpu_has_feature(CPU_FTR_LOCKLESS_TLBIE); - - if (lock_tlbie) - spin_lock(&native_tlbie_lock); - tlbie(va, large); - if (lock_tlbie) - spin_unlock(&native_tlbie_lock); - } - - return ret; -} - /* * Update the page protection bits. Intended to be used to create * guard pages for kernel data structures on pages which are bolted * in the HPT. Assumes pages being operated on will not be stolen. - * Does not work on large pages. * * No need to lock here because we should be the only user. */ -static void native_hpte_updateboltedpp(unsigned long newpp, unsigned long ea) +static void native_hpte_updateboltedpp(unsigned long newpp, unsigned long ea, + int psize) { - unsigned long vsid, va, vpn, flags = 0; + unsigned long vsid, va; long slot; hpte_t *hptep; - int lock_tlbie = !cpu_has_feature(CPU_FTR_LOCKLESS_TLBIE); vsid = get_kernel_vsid(ea); va = (vsid << 28) | (ea & 0x0fffffff); - vpn = va >> PAGE_SHIFT; - slot = native_hpte_find(vpn); + slot = native_hpte_find(va, psize); if (slot == -1) panic("could not find page to bolt\n"); hptep = htab_address + slot; - set_pp_bit(newpp, hptep); + /* Update the HPTE */ + hptep->r = (hptep->r & ~(HPTE_R_PP | HPTE_R_N)) | + (newpp & (HPTE_R_PP | HPTE_R_N)); - /* Ensure it is out of the tlb too */ - if (lock_tlbie) - spin_lock_irqsave(&native_tlbie_lock, flags); - tlbie(va, 0); - if (lock_tlbie) - spin_unlock_irqrestore(&native_tlbie_lock, flags); + /* Ensure it is out of the tlb too. */ + tlbie(va, psize, 0); } static void native_hpte_invalidate(unsigned long slot, unsigned long va, - int large, int local) + int psize, int local) { hpte_t *hptep = htab_address + slot; unsigned long hpte_v; - unsigned long avpn = va >> 23; + unsigned long want_v; unsigned long flags; - int lock_tlbie = !cpu_has_feature(CPU_FTR_LOCKLESS_TLBIE); - - if (large) - avpn &= ~1; local_irq_save(flags); - native_lock_hpte(hptep); + DBG_LOW(" invalidate(va=%016lx, hash: %x)\n", va, slot); + + want_v = hpte_encode_v(va, psize); + native_lock_hpte(hptep); hpte_v = hptep->v; /* Even if we miss, we need to invalidate the TLB */ - if ((HPTE_V_AVPN_VAL(hpte_v) != avpn) - || !(hpte_v & HPTE_V_VALID)) { + if (!HPTE_V_COMPARE(hpte_v, want_v) || !(hpte_v & HPTE_V_VALID)) native_unlock_hpte(hptep); - } else { + else /* Invalidate the hpte. NOTE: this also unlocks it */ hptep->v = 0; - } - /* Invalidate the tlb */ - if (cpu_has_feature(CPU_FTR_TLBIEL) && !large && local) { - tlbiel(va); - } else { - if (lock_tlbie) - spin_lock(&native_tlbie_lock); - tlbie(va, large); - if (lock_tlbie) - spin_unlock(&native_tlbie_lock); - } + /* Invalidate the TLB */ + tlbie(va, psize, local); + local_irq_restore(flags); } +/* + * XXX This need fixing based on page size. It's only used by + * native_hpte_clear() for now which needs fixing too so they + * make a good pair... + */ +static unsigned long slot2va(unsigned long hpte_v, unsigned long slot) +{ + unsigned long avpn = HPTE_V_AVPN_VAL(hpte_v); + unsigned long va; + + va = avpn << 23; + + if (! (hpte_v & HPTE_V_LARGE)) { + unsigned long vpi, pteg; + + pteg = slot / HPTES_PER_GROUP; + if (hpte_v & HPTE_V_SECONDARY) + pteg = ~pteg; + + vpi = ((va >> 28) ^ pteg) & htab_hash_mask; + + va |= vpi << PAGE_SHIFT; + } + + return va; +} + /* * clear all mappings on kexec. All cpus are in real mode (or they will * be when they isi), and we are the only one left. We rely on our kernel @@ -298,6 +374,8 @@ static void native_hpte_invalidate(unsigned long slot, unsigned long va, * * TODO: add batching support when enabled. remember, no dynamic memory here, * athough there is the control page available... + * + * XXX FIXME: 4k only for now ! */ static void native_hpte_clear(void) { @@ -327,7 +405,7 @@ static void native_hpte_clear(void) if (hpte_v & HPTE_V_VALID) { hptep->v = 0; - tlbie(slot2va(hpte_v, slot), hpte_v & HPTE_V_LARGE); + tlbie(slot2va(hpte_v, slot), MMU_PAGE_4K, 0); } } @@ -335,59 +413,59 @@ static void native_hpte_clear(void) local_irq_restore(flags); } +/* + * Batched hash table flush, we batch the tlbie's to avoid taking/releasing + * the lock all the time + */ static void native_flush_hash_range(unsigned long number, int local) { - unsigned long va, vpn, hash, secondary, slot, flags, avpn; - int i, j; + unsigned long va, hash, index, hidx, shift, slot; hpte_t *hptep; unsigned long hpte_v; + unsigned long want_v; + unsigned long flags; + real_pte_t pte; struct ppc64_tlb_batch *batch = &__get_cpu_var(ppc64_tlb_batch); - unsigned long large = batch->large; + unsigned long psize = batch->psize; + int i; local_irq_save(flags); - j = 0; for (i = 0; i < number; i++) { - va = batch->vaddr[j]; - if (large) - vpn = va >> HPAGE_SHIFT; - else - vpn = va >> PAGE_SHIFT; - hash = hpt_hash(vpn, large); - secondary = (pte_val(batch->pte[i]) & _PAGE_SECONDARY) >> 15; - if (secondary) - hash = ~hash; - slot = (hash & htab_hash_mask) * HPTES_PER_GROUP; - slot += (pte_val(batch->pte[i]) & _PAGE_GROUP_IX) >> 12; - - hptep = htab_address + slot; - - avpn = va >> 23; - if (large) - avpn &= ~0x1UL; - - native_lock_hpte(hptep); - - hpte_v = hptep->v; - - /* Even if we miss, we need to invalidate the TLB */ - if ((HPTE_V_AVPN_VAL(hpte_v) != avpn) - || !(hpte_v & HPTE_V_VALID)) { - native_unlock_hpte(hptep); - } else { - /* Invalidate the hpte. NOTE: this also unlocks it */ - hptep->v = 0; - } - - j++; + va = batch->vaddr[i]; + pte = batch->pte[i]; + + pte_iterate_hashed_subpages(pte, psize, va, index, shift) { + hash = hpt_hash(va, shift); + hidx = __rpte_to_hidx(pte, index); + if (hidx & _PTEIDX_SECONDARY) + hash = ~hash; + slot = (hash & htab_hash_mask) * HPTES_PER_GROUP; + slot += hidx & _PTEIDX_GROUP_IX; + hptep = htab_address + slot; + want_v = hpte_encode_v(va, psize); + native_lock_hpte(hptep); + hpte_v = hptep->v; + if (!HPTE_V_COMPARE(hpte_v, want_v) || + !(hpte_v & HPTE_V_VALID)) + native_unlock_hpte(hptep); + else + hptep->v = 0; + } pte_iterate_hashed_end(); } - if (cpu_has_feature(CPU_FTR_TLBIEL) && !large && local) { + if (cpu_has_feature(CPU_FTR_TLBIEL) && + mmu_psize_defs[psize].tlbiel && local) { asm volatile("ptesync":::"memory"); - - for (i = 0; i < j; i++) - __tlbiel(batch->vaddr[i]); - + for (i = 0; i < number; i++) { + va = batch->vaddr[i]; + pte = batch->pte[i]; + + pte_iterate_hashed_subpages(pte, psize, va, index, + shift) { + __tlbiel(va, psize); + } pte_iterate_hashed_end(); + } asm volatile("ptesync":::"memory"); } else { int lock_tlbie = !cpu_has_feature(CPU_FTR_LOCKLESS_TLBIE); @@ -396,10 +474,15 @@ static void native_flush_hash_range(unsigned long number, int local) spin_lock(&native_tlbie_lock); asm volatile("ptesync":::"memory"); - - for (i = 0; i < j; i++) - __tlbie(batch->vaddr[i], large); - + for (i = 0; i < number; i++) { + va = batch->vaddr[i]; + pte = batch->pte[i]; + + pte_iterate_hashed_subpages(pte, psize, va, index, + shift) { + __tlbie(va, psize); + } pte_iterate_hashed_end(); + } asm volatile("eieio; tlbsync; ptesync":::"memory"); if (lock_tlbie) diff --git a/arch/powerpc/mm/hash_utils_64.c b/arch/powerpc/mm/hash_utils_64.c index 6e9e05cce02c..b2f3dbca6952 100644 --- a/arch/powerpc/mm/hash_utils_64.c +++ b/arch/powerpc/mm/hash_utils_64.c @@ -19,6 +19,7 @@ */ #undef DEBUG +#undef DEBUG_LOW #include #include @@ -59,6 +60,15 @@ #define DBG(fmt...) #endif +#ifdef DEBUG_LOW +#define DBG_LOW(fmt...) udbg_printf(fmt) +#else +#define DBG_LOW(fmt...) +#endif + +#define KB (1024) +#define MB (1024*KB) + /* * Note: pte --> Linux PTE * HPTE --> PowerPC Hashed Page Table Entry @@ -77,91 +87,290 @@ extern unsigned long dart_tablebase; hpte_t *htab_address; unsigned long htab_hash_mask; - unsigned long _SDR1; +struct mmu_psize_def mmu_psize_defs[MMU_PAGE_COUNT]; +int mmu_linear_psize = MMU_PAGE_4K; +int mmu_virtual_psize = MMU_PAGE_4K; +#ifdef CONFIG_HUGETLB_PAGE +int mmu_huge_psize = MMU_PAGE_16M; +unsigned int HPAGE_SHIFT; +#endif -#define KB (1024) -#define MB (1024*KB) - -static inline void loop_forever(void) -{ - volatile unsigned long x = 1; - for(;x;x|=1) - ; -} +/* There are definitions of page sizes arrays to be used when none + * is provided by the firmware. + */ -static inline void create_pte_mapping(unsigned long start, unsigned long end, - unsigned long mode, int large) +/* Pre-POWER4 CPUs (4k pages only) + */ +struct mmu_psize_def mmu_psize_defaults_old[] = { + [MMU_PAGE_4K] = { + .shift = 12, + .sllp = 0, + .penc = 0, + .avpnm = 0, + .tlbiel = 0, + }, +}; + +/* POWER4, GPUL, POWER5 + * + * Support for 16Mb large pages + */ +struct mmu_psize_def mmu_psize_defaults_gp[] = { + [MMU_PAGE_4K] = { + .shift = 12, + .sllp = 0, + .penc = 0, + .avpnm = 0, + .tlbiel = 1, + }, + [MMU_PAGE_16M] = { + .shift = 24, + .sllp = SLB_VSID_L, + .penc = 0, + .avpnm = 0x1UL, + .tlbiel = 0, + }, +}; + + +int htab_bolt_mapping(unsigned long vstart, unsigned long vend, + unsigned long pstart, unsigned long mode, int psize) { - unsigned long addr; - unsigned int step; + unsigned long vaddr, paddr; + unsigned int step, shift; unsigned long tmp_mode; - unsigned long vflags; + int ret = 0; - if (large) { - step = 16*MB; - vflags = HPTE_V_BOLTED | HPTE_V_LARGE; - } else { - step = 4*KB; - vflags = HPTE_V_BOLTED; - } + shift = mmu_psize_defs[psize].shift; + step = 1 << shift; - for (addr = start; addr < end; addr += step) { + for (vaddr = vstart, paddr = pstart; vaddr < vend; + vaddr += step, paddr += step) { unsigned long vpn, hash, hpteg; - unsigned long vsid = get_kernel_vsid(addr); - unsigned long va = (vsid << 28) | (addr & 0xfffffff); - int ret = -1; - - if (large) - vpn = va >> HPAGE_SHIFT; - else - vpn = va >> PAGE_SHIFT; - + unsigned long vsid = get_kernel_vsid(vaddr); + unsigned long va = (vsid << 28) | (vaddr & 0x0fffffff); + vpn = va >> shift; tmp_mode = mode; /* Make non-kernel text non-executable */ - if (!in_kernel_text(addr)) - tmp_mode = mode | HW_NO_EXEC; - - hash = hpt_hash(vpn, large); + if (!in_kernel_text(vaddr)) + tmp_mode = mode | HPTE_R_N; + hash = hpt_hash(va, shift); hpteg = ((hash & htab_hash_mask) * HPTES_PER_GROUP); + /* The crap below can be cleaned once ppd_md.probe() can + * set up the hash callbacks, thus we can just used the + * normal insert callback here. + */ #ifdef CONFIG_PPC_ISERIES - if (systemcfg->platform & PLATFORM_ISERIES_LPAR) - ret = iSeries_hpte_bolt_or_insert(hpteg, va, - virt_to_abs(addr) >> PAGE_SHIFT, - vflags, tmp_mode); + if (systemcfg->platform == PLATFORM_ISERIES_LPAR) + ret = iSeries_hpte_insert(hpteg, va, + virt_to_abs(paddr), + tmp_mode, + HPTE_V_BOLTED, + psize); else #endif #ifdef CONFIG_PPC_PSERIES if (systemcfg->platform & PLATFORM_LPAR) ret = pSeries_lpar_hpte_insert(hpteg, va, - virt_to_abs(addr) >> PAGE_SHIFT, - vflags, tmp_mode); + virt_to_abs(paddr), + tmp_mode, + HPTE_V_BOLTED, + psize); else #endif #ifdef CONFIG_PPC_MULTIPLATFORM ret = native_hpte_insert(hpteg, va, - virt_to_abs(addr) >> PAGE_SHIFT, - vflags, tmp_mode); + virt_to_abs(paddr), + tmp_mode, HPTE_V_BOLTED, + psize); #endif + if (ret < 0) + break; + } + return ret < 0 ? ret : 0; +} - if (ret == -1) { - ppc64_terminate_msg(0x20, "create_pte_mapping"); - loop_forever(); +static int __init htab_dt_scan_page_sizes(unsigned long node, + const char *uname, int depth, + void *data) +{ + char *type = of_get_flat_dt_prop(node, "device_type", NULL); + u32 *prop; + unsigned long size = 0; + + /* We are scanning "cpu" nodes only */ + if (type == NULL || strcmp(type, "cpu") != 0) + return 0; + + prop = (u32 *)of_get_flat_dt_prop(node, + "ibm,segment-page-sizes", &size); + if (prop != NULL) { + DBG("Page sizes from device-tree:\n"); + size /= 4; + cur_cpu_spec->cpu_features &= ~(CPU_FTR_16M_PAGE); + while(size > 0) { + unsigned int shift = prop[0]; + unsigned int slbenc = prop[1]; + unsigned int lpnum = prop[2]; + unsigned int lpenc = 0; + struct mmu_psize_def *def; + int idx = -1; + + size -= 3; prop += 3; + while(size > 0 && lpnum) { + if (prop[0] == shift) + lpenc = prop[1]; + prop += 2; size -= 2; + lpnum--; + } + switch(shift) { + case 0xc: + idx = MMU_PAGE_4K; + break; + case 0x10: + idx = MMU_PAGE_64K; + break; + case 0x14: + idx = MMU_PAGE_1M; + break; + case 0x18: + idx = MMU_PAGE_16M; + cur_cpu_spec->cpu_features |= CPU_FTR_16M_PAGE; + break; + case 0x22: + idx = MMU_PAGE_16G; + break; + } + if (idx < 0) + continue; + def = &mmu_psize_defs[idx]; + def->shift = shift; + if (shift <= 23) + def->avpnm = 0; + else + def->avpnm = (1 << (shift - 23)) - 1; + def->sllp = slbenc; + def->penc = lpenc; + /* We don't know for sure what's up with tlbiel, so + * for now we only set it for 4K and 64K pages + */ + if (idx == MMU_PAGE_4K || idx == MMU_PAGE_64K) + def->tlbiel = 1; + else + def->tlbiel = 0; + + DBG(" %d: shift=%02x, sllp=%04x, avpnm=%08x, " + "tlbiel=%d, penc=%d\n", + idx, shift, def->sllp, def->avpnm, def->tlbiel, + def->penc); } + return 1; + } + return 0; +} + + +static void __init htab_init_page_sizes(void) +{ + int rc; + + /* Default to 4K pages only */ + memcpy(mmu_psize_defs, mmu_psize_defaults_old, + sizeof(mmu_psize_defaults_old)); + + /* + * Try to find the available page sizes in the device-tree + */ + rc = of_scan_flat_dt(htab_dt_scan_page_sizes, NULL); + if (rc != 0) /* Found */ + goto found; + + /* + * Not in the device-tree, let's fallback on known size + * list for 16M capable GP & GR + */ + if ((systemcfg->platform != PLATFORM_ISERIES_LPAR) && + cpu_has_feature(CPU_FTR_16M_PAGE)) + memcpy(mmu_psize_defs, mmu_psize_defaults_gp, + sizeof(mmu_psize_defaults_gp)); + found: + /* + * Pick a size for the linear mapping. Currently, we only support + * 16M, 1M and 4K which is the default + */ + if (mmu_psize_defs[MMU_PAGE_16M].shift) + mmu_linear_psize = MMU_PAGE_16M; + else if (mmu_psize_defs[MMU_PAGE_1M].shift) + mmu_linear_psize = MMU_PAGE_1M; + + /* + * Pick a size for the ordinary pages. Default is 4K, we support + * 64K if cache inhibited large pages are supported by the + * processor + */ +#ifdef CONFIG_PPC_64K_PAGES + if (mmu_psize_defs[MMU_PAGE_64K].shift && + cpu_has_feature(CPU_FTR_CI_LARGE_PAGE)) + mmu_virtual_psize = MMU_PAGE_64K; +#endif + + printk(KERN_INFO "Page orders: linear mapping = %d, others = %d\n", + mmu_psize_defs[mmu_linear_psize].shift, + mmu_psize_defs[mmu_virtual_psize].shift); + +#ifdef CONFIG_HUGETLB_PAGE + /* Init large page size. Currently, we pick 16M or 1M depending + * on what is available + */ + if (mmu_psize_defs[MMU_PAGE_16M].shift) + mmu_huge_psize = MMU_PAGE_16M; + else if (mmu_psize_defs[MMU_PAGE_1M].shift) + mmu_huge_psize = MMU_PAGE_1M; + + /* Calculate HPAGE_SHIFT and sanity check it */ + if (mmu_psize_defs[mmu_huge_psize].shift > 16 && + mmu_psize_defs[mmu_huge_psize].shift < 28) + HPAGE_SHIFT = mmu_psize_defs[mmu_huge_psize].shift; + else + HPAGE_SHIFT = 0; /* No huge pages dude ! */ +#endif /* CONFIG_HUGETLB_PAGE */ +} + +static int __init htab_dt_scan_pftsize(unsigned long node, + const char *uname, int depth, + void *data) +{ + char *type = of_get_flat_dt_prop(node, "device_type", NULL); + u32 *prop; + + /* We are scanning "cpu" nodes only */ + if (type == NULL || strcmp(type, "cpu") != 0) + return 0; + + prop = (u32 *)of_get_flat_dt_prop(node, "ibm,pft-size", NULL); + if (prop != NULL) { + /* pft_size[0] is the NUMA CEC cookie */ + ppc64_pft_size = prop[1]; + return 1; } + return 0; } -static unsigned long get_hashtable_size(void) +static unsigned long __init htab_get_table_size(void) { unsigned long rnd_mem_size, pteg_count; - /* If hash size wasn't obtained in prom.c, we calculate it now based on - * the total RAM size + /* If hash size isn't already provided by the platform, we try to + * retreive it from the device-tree. If it's not there neither, we + * calculate it now based on the total RAM size */ + if (ppc64_pft_size == 0) + of_scan_flat_dt(htab_dt_scan_pftsize, NULL); if (ppc64_pft_size) return 1UL << ppc64_pft_size; @@ -181,17 +390,21 @@ void __init htab_initialize(void) unsigned long table, htab_size_bytes; unsigned long pteg_count; unsigned long mode_rw; - int i, use_largepages = 0; unsigned long base = 0, size = 0; + int i; + extern unsigned long tce_alloc_start, tce_alloc_end; DBG(" -> htab_initialize()\n"); + /* Initialize page sizes */ + htab_init_page_sizes(); + /* * Calculate the required size of the htab. We want the number of * PTEGs to equal one half the number of real pages. */ - htab_size_bytes = get_hashtable_size(); + htab_size_bytes = htab_get_table_size(); pteg_count = htab_size_bytes >> 7; /* For debug, make the HTAB 1/8 as big as it normally would be. */ @@ -211,14 +424,11 @@ void __init htab_initialize(void) * the absolute address space. */ table = lmb_alloc(htab_size_bytes, htab_size_bytes); + BUG_ON(table == 0); DBG("Hash table allocated at %lx, size: %lx\n", table, htab_size_bytes); - if ( !table ) { - ppc64_terminate_msg(0x20, "hpt space"); - loop_forever(); - } htab_address = abs_to_virt(table); /* htab absolute addr + encoded htabsize */ @@ -234,8 +444,6 @@ void __init htab_initialize(void) * _NOT_ map it to avoid cache paradoxes as it's remapped non * cacheable later on */ - if (cpu_has_feature(CPU_FTR_16M_PAGE)) - use_largepages = 1; /* create bolted the linear mapping in the hash table */ for (i=0; i < lmb.memory.cnt; i++) { @@ -246,27 +454,32 @@ void __init htab_initialize(void) #ifdef CONFIG_U3_DART /* Do not map the DART space. Fortunately, it will be aligned - * in such a way that it will not cross two lmb regions and will - * fit within a single 16Mb page. - * The DART space is assumed to be a full 16Mb region even if we - * only use 2Mb of that space. We will use more of it later for - * AGP GART. We have to use a full 16Mb large page. + * in such a way that it will not cross two lmb regions and + * will fit within a single 16Mb page. + * The DART space is assumed to be a full 16Mb region even if + * we only use 2Mb of that space. We will use more of it later + * for AGP GART. We have to use a full 16Mb large page. */ DBG("DART base: %lx\n", dart_tablebase); if (dart_tablebase != 0 && dart_tablebase >= base && dart_tablebase < (base + size)) { if (base != dart_tablebase) - create_pte_mapping(base, dart_tablebase, mode_rw, - use_largepages); + BUG_ON(htab_bolt_mapping(base, dart_tablebase, + base, mode_rw, + mmu_linear_psize)); if ((base + size) > (dart_tablebase + 16*MB)) - create_pte_mapping(dart_tablebase + 16*MB, base + size, - mode_rw, use_largepages); + BUG_ON(htab_bolt_mapping(dart_tablebase+16*MB, + base + size, + dart_tablebase+16*MB, + mode_rw, + mmu_linear_psize)); continue; } #endif /* CONFIG_U3_DART */ - create_pte_mapping(base, base + size, mode_rw, use_largepages); - } + BUG_ON(htab_bolt_mapping(base, base + size, base, + mode_rw, mmu_linear_psize)); + } /* * If we have a memory_limit and we've allocated TCEs then we need to @@ -282,8 +495,9 @@ void __init htab_initialize(void) if (base + size >= tce_alloc_start) tce_alloc_start = base + size + 1; - create_pte_mapping(tce_alloc_start, tce_alloc_end, - mode_rw, use_largepages); + BUG_ON(htab_bolt_mapping(tce_alloc_start, tce_alloc_end, + tce_alloc_start, mode_rw, + mmu_linear_psize)); } DBG(" <- htab_initialize()\n"); @@ -298,9 +512,6 @@ unsigned int hash_page_do_lazy_icache(unsigned int pp, pte_t pte, int trap) { struct page *page; - if (!pfn_valid(pte_pfn(pte))) - return pp; - page = pte_page(pte); /* page is dirty */ @@ -309,7 +520,7 @@ unsigned int hash_page_do_lazy_icache(unsigned int pp, pte_t pte, int trap) __flush_dcache_icache(page_address(page)); set_bit(PG_arch_1, &page->flags); } else - pp |= HW_NO_EXEC; + pp |= HPTE_R_N; } return pp; } @@ -325,94 +536,169 @@ int hash_page(unsigned long ea, unsigned long access, unsigned long trap) unsigned long vsid; struct mm_struct *mm; pte_t *ptep; - int ret; - int user_region = 0; - int local = 0; cpumask_t tmp; + int rc, user_region = 0, local = 0; - if ((ea & ~REGION_MASK) >= PGTABLE_RANGE) - return 1; + DBG_LOW("hash_page(ea=%016lx, access=%lx, trap=%lx\n", + ea, access, trap); + if ((ea & ~REGION_MASK) >= PGTABLE_RANGE) { + DBG_LOW(" out of pgtable range !\n"); + return 1; + } + + /* Get region & vsid */ switch (REGION_ID(ea)) { case USER_REGION_ID: user_region = 1; mm = current->mm; - if (! mm) + if (! mm) { + DBG_LOW(" user region with no mm !\n"); return 1; - + } vsid = get_vsid(mm->context.id, ea); break; case VMALLOC_REGION_ID: mm = &init_mm; vsid = get_kernel_vsid(ea); break; -#if 0 - case KERNEL_REGION_ID: - /* - * Should never get here - entire 0xC0... region is bolted. - * Send the problem up to do_page_fault - */ -#endif default: /* Not a valid range * Send the problem up to do_page_fault */ return 1; - break; } + DBG_LOW(" mm=%p, mm->pgdir=%p, vsid=%016lx\n", mm, mm->pgd, vsid); + /* Get pgdir */ pgdir = mm->pgd; - if (pgdir == NULL) return 1; + /* Check CPU locality */ tmp = cpumask_of_cpu(smp_processor_id()); if (user_region && cpus_equal(mm->cpu_vm_mask, tmp)) local = 1; - /* Is this a huge page ? */ - if (unlikely(in_hugepage_area(mm->context, ea))) - ret = hash_huge_page(mm, access, ea, vsid, local); - else { - ptep = find_linux_pte(pgdir, ea); - if (ptep == NULL) - return 1; - ret = __hash_page(ea, access, vsid, ptep, trap, local); + /* Handle hugepage regions */ + if (unlikely(in_hugepage_area(mm->context, ea))) { + DBG_LOW(" -> huge page !\n"); + return hash_huge_page(mm, access, ea, vsid, local); + } + + /* Get PTE and page size from page tables */ + ptep = find_linux_pte(pgdir, ea); + if (ptep == NULL || !pte_present(*ptep)) { + DBG_LOW(" no PTE !\n"); + return 1; + } + +#ifndef CONFIG_PPC_64K_PAGES + DBG_LOW(" i-pte: %016lx\n", pte_val(*ptep)); +#else + DBG_LOW(" i-pte: %016lx %016lx\n", pte_val(*ptep), + pte_val(*(ptep + PTRS_PER_PTE))); +#endif + /* Pre-check access permissions (will be re-checked atomically + * in __hash_page_XX but this pre-check is a fast path + */ + if (access & ~pte_val(*ptep)) { + DBG_LOW(" no access !\n"); + return 1; } - return ret; + /* Do actual hashing */ +#ifndef CONFIG_PPC_64K_PAGES + rc = __hash_page_4K(ea, access, vsid, ptep, trap, local); +#else + if (mmu_virtual_psize == MMU_PAGE_64K) + rc = __hash_page_64K(ea, access, vsid, ptep, trap, local); + else + rc = __hash_page_4K(ea, access, vsid, ptep, trap, local); +#endif /* CONFIG_PPC_64K_PAGES */ + +#ifndef CONFIG_PPC_64K_PAGES + DBG_LOW(" o-pte: %016lx\n", pte_val(*ptep)); +#else + DBG_LOW(" o-pte: %016lx %016lx\n", pte_val(*ptep), + pte_val(*(ptep + PTRS_PER_PTE))); +#endif + DBG_LOW(" -> rc=%d\n", rc); + return rc; } -void flush_hash_page(unsigned long va, pte_t pte, int local) +void hash_preload(struct mm_struct *mm, unsigned long ea, + unsigned long access, unsigned long trap) { - unsigned long vpn, hash, secondary, slot; - unsigned long huge = pte_huge(pte); + unsigned long vsid; + void *pgdir; + pte_t *ptep; + cpumask_t mask; + unsigned long flags; + int local = 0; + + /* We don't want huge pages prefaulted for now + */ + if (unlikely(in_hugepage_area(mm->context, ea))) + return; + + DBG_LOW("hash_preload(mm=%p, mm->pgdir=%p, ea=%016lx, access=%lx," + " trap=%lx\n", mm, mm->pgd, ea, access, trap); - if (huge) - vpn = va >> HPAGE_SHIFT; + /* Get PTE, VSID, access mask */ + pgdir = mm->pgd; + if (pgdir == NULL) + return; + ptep = find_linux_pte(pgdir, ea); + if (!ptep) + return; + vsid = get_vsid(mm->context.id, ea); + + /* Hash it in */ + local_irq_save(flags); + mask = cpumask_of_cpu(smp_processor_id()); + if (cpus_equal(mm->cpu_vm_mask, mask)) + local = 1; +#ifndef CONFIG_PPC_64K_PAGES + __hash_page_4K(ea, access, vsid, ptep, trap, local); +#else + if (mmu_virtual_psize == MMU_PAGE_64K) + __hash_page_64K(ea, access, vsid, ptep, trap, local); else - vpn = va >> PAGE_SHIFT; - hash = hpt_hash(vpn, huge); - secondary = (pte_val(pte) & _PAGE_SECONDARY) >> 15; - if (secondary) - hash = ~hash; - slot = (hash & htab_hash_mask) * HPTES_PER_GROUP; - slot += (pte_val(pte) & _PAGE_GROUP_IX) >> 12; - - ppc_md.hpte_invalidate(slot, va, huge, local); + __hash_page_4K(ea, access, vsid, ptep, trap, local); +#endif /* CONFIG_PPC_64K_PAGES */ + local_irq_restore(flags); +} + +void flush_hash_page(unsigned long va, real_pte_t pte, int psize, int local) +{ + unsigned long hash, index, shift, hidx, slot; + + DBG_LOW("flush_hash_page(va=%016x)\n", va); + pte_iterate_hashed_subpages(pte, psize, va, index, shift) { + hash = hpt_hash(va, shift); + hidx = __rpte_to_hidx(pte, index); + if (hidx & _PTEIDX_SECONDARY) + hash = ~hash; + slot = (hash & htab_hash_mask) * HPTES_PER_GROUP; + slot += hidx & _PTEIDX_GROUP_IX; + DBG_LOW(" sub %d: hash=%x, hidx=%x\n", index, slot, hidx); + ppc_md.hpte_invalidate(slot, va, psize, local); + } pte_iterate_hashed_end(); } void flush_hash_range(unsigned long number, int local) { - if (ppc_md.flush_hash_range) { + if (ppc_md.flush_hash_range) ppc_md.flush_hash_range(number, local); - } else { + else { int i; struct ppc64_tlb_batch *batch = &__get_cpu_var(ppc64_tlb_batch); for (i = 0; i < number; i++) - flush_hash_page(batch->vaddr[i], batch->pte[i], local); + flush_hash_page(batch->vaddr[i], batch->pte[i], + batch->psize, local); } } @@ -452,6 +738,18 @@ void __init htab_finish_init(void) extern unsigned int *htab_call_hpte_remove; extern unsigned int *htab_call_hpte_updatepp; +#ifdef CONFIG_PPC_64K_PAGES + extern unsigned int *ht64_call_hpte_insert1; + extern unsigned int *ht64_call_hpte_insert2; + extern unsigned int *ht64_call_hpte_remove; + extern unsigned int *ht64_call_hpte_updatepp; + + make_bl(ht64_call_hpte_insert1, ppc_md.hpte_insert); + make_bl(ht64_call_hpte_insert2, ppc_md.hpte_insert); + make_bl(ht64_call_hpte_remove, ppc_md.hpte_remove); + make_bl(ht64_call_hpte_updatepp, ppc_md.hpte_updatepp); +#endif /* CONFIG_PPC_64K_PAGES */ + make_bl(htab_call_hpte_insert1, ppc_md.hpte_insert); make_bl(htab_call_hpte_insert2, ppc_md.hpte_insert); make_bl(htab_call_hpte_remove, ppc_md.hpte_remove); diff --git a/arch/powerpc/mm/hugetlbpage.c b/arch/powerpc/mm/hugetlbpage.c index 0ea0994ed974..0073a04047e4 100644 --- a/arch/powerpc/mm/hugetlbpage.c +++ b/arch/powerpc/mm/hugetlbpage.c @@ -47,10 +47,25 @@ pte_t *huge_pte_offset(struct mm_struct *mm, unsigned long addr) pu = pud_offset(pg, addr); if (!pud_none(*pu)) { pm = pmd_offset(pu, addr); +#ifdef CONFIG_PPC_64K_PAGES + /* Currently, we use the normal PTE offset within full + * size PTE pages, thus our huge PTEs are scattered in + * the PTE page and we do waste some. We may change + * that in the future, but the current mecanism keeps + * things much simpler + */ + if (!pmd_none(*pm)) { + /* Note: pte_offset_* are all equivalent on + * ppc64 as we don't have HIGHMEM + */ + pt = pte_offset_kernel(pm, addr); + return pt; + } +#else /* CONFIG_PPC_64K_PAGES */ + /* On 4k pages, we put huge PTEs in the PMD page */ pt = (pte_t *)pm; - BUG_ON(!pmd_none(*pm) - && !(pte_present(*pt) && pte_huge(*pt))); return pt; +#endif /* CONFIG_PPC_64K_PAGES */ } } @@ -74,9 +89,16 @@ pte_t *huge_pte_alloc(struct mm_struct *mm, unsigned long addr) if (pu) { pm = pmd_alloc(mm, pu, addr); if (pm) { +#ifdef CONFIG_PPC_64K_PAGES + /* See comment in huge_pte_offset. Note that if we ever + * want to put the page size in the PMD, we would have + * to open code our own pte_alloc* function in order + * to populate and set the size atomically + */ + pt = pte_alloc_map(mm, pm, addr); +#else /* CONFIG_PPC_64K_PAGES */ pt = (pte_t *)pm; - BUG_ON(!pmd_none(*pm) - && !(pte_present(*pt) && pte_huge(*pt))); +#endif /* CONFIG_PPC_64K_PAGES */ return pt; } } @@ -84,35 +106,29 @@ pte_t *huge_pte_alloc(struct mm_struct *mm, unsigned long addr) return NULL; } -#define HUGEPTE_BATCH_SIZE (HPAGE_SIZE / PMD_SIZE) - void set_huge_pte_at(struct mm_struct *mm, unsigned long addr, pte_t *ptep, pte_t pte) { - int i; - if (pte_present(*ptep)) { - pte_clear(mm, addr, ptep); + /* We open-code pte_clear because we need to pass the right + * argument to hpte_update (huge / !huge) + */ + unsigned long old = pte_update(ptep, ~0UL); + if (old & _PAGE_HASHPTE) + hpte_update(mm, addr & HPAGE_MASK, ptep, old, 1); flush_tlb_pending(); } - - for (i = 0; i < HUGEPTE_BATCH_SIZE; i++) { - *ptep = __pte(pte_val(pte) & ~_PAGE_HPTEFLAGS); - ptep++; - } + *ptep = __pte(pte_val(pte) & ~_PAGE_HPTEFLAGS); } pte_t huge_ptep_get_and_clear(struct mm_struct *mm, unsigned long addr, pte_t *ptep) { unsigned long old = pte_update(ptep, ~0UL); - int i; if (old & _PAGE_HASHPTE) - hpte_update(mm, addr, old, 0); - - for (i = 1; i < HUGEPTE_BATCH_SIZE; i++) - ptep[i] = __pte(0); + hpte_update(mm, addr & HPAGE_MASK, ptep, old, 1); + *ptep = __pte(0); return __pte(old); } @@ -563,6 +579,8 @@ unsigned long hugetlb_get_unmapped_area(struct file *file, unsigned long addr, int lastshift; u16 areamask, curareas; + if (HPAGE_SHIFT == 0) + return -EINVAL; if (len & ~HPAGE_MASK) return -EINVAL; @@ -619,19 +637,15 @@ int hash_huge_page(struct mm_struct *mm, unsigned long access, unsigned long ea, unsigned long vsid, int local) { pte_t *ptep; - unsigned long va, vpn; - pte_t old_pte, new_pte; - unsigned long rflags, prpn; + unsigned long old_pte, new_pte; + unsigned long va, rflags, pa; long slot; int err = 1; - spin_lock(&mm->page_table_lock); - ptep = huge_pte_offset(mm, ea); /* Search the Linux page table for a match with va */ va = (vsid << 28) | (ea & 0x0fffffff); - vpn = va >> HPAGE_SHIFT; /* * If no pte found or not present, send the problem up to @@ -640,8 +654,6 @@ int hash_huge_page(struct mm_struct *mm, unsigned long access, if (unlikely(!ptep || pte_none(*ptep))) goto out; -/* BUG_ON(pte_bad(*ptep)); */ - /* * Check the user's access rights to the page. If access should be * prevented then send the problem up to do_page_fault. @@ -661,58 +673,64 @@ int hash_huge_page(struct mm_struct *mm, unsigned long access, */ - old_pte = *ptep; - new_pte = old_pte; - - rflags = 0x2 | (! (pte_val(new_pte) & _PAGE_RW)); + do { + old_pte = pte_val(*ptep); + if (old_pte & _PAGE_BUSY) + goto out; + new_pte = old_pte | _PAGE_BUSY | + _PAGE_ACCESSED | _PAGE_HASHPTE; + } while(old_pte != __cmpxchg_u64((unsigned long *)ptep, + old_pte, new_pte)); + + rflags = 0x2 | (!(new_pte & _PAGE_RW)); /* _PAGE_EXEC -> HW_NO_EXEC since it's inverted */ - rflags |= ((pte_val(new_pte) & _PAGE_EXEC) ? 0 : HW_NO_EXEC); + rflags |= ((new_pte & _PAGE_EXEC) ? 0 : HPTE_R_N); /* Check if pte already has an hpte (case 2) */ - if (unlikely(pte_val(old_pte) & _PAGE_HASHPTE)) { + if (unlikely(old_pte & _PAGE_HASHPTE)) { /* There MIGHT be an HPTE for this pte */ unsigned long hash, slot; - hash = hpt_hash(vpn, 1); - if (pte_val(old_pte) & _PAGE_SECONDARY) + hash = hpt_hash(va, HPAGE_SHIFT); + if (old_pte & _PAGE_F_SECOND) hash = ~hash; slot = (hash & htab_hash_mask) * HPTES_PER_GROUP; - slot += (pte_val(old_pte) & _PAGE_GROUP_IX) >> 12; + slot += (old_pte & _PAGE_F_GIX) >> 12; if (ppc_md.hpte_updatepp(slot, rflags, va, 1, local) == -1) - pte_val(old_pte) &= ~_PAGE_HPTEFLAGS; + old_pte &= ~_PAGE_HPTEFLAGS; } - if (likely(!(pte_val(old_pte) & _PAGE_HASHPTE))) { - unsigned long hash = hpt_hash(vpn, 1); + if (likely(!(old_pte & _PAGE_HASHPTE))) { + unsigned long hash = hpt_hash(va, HPAGE_SHIFT); unsigned long hpte_group; - prpn = pte_pfn(old_pte); + pa = pte_pfn(__pte(old_pte)) << PAGE_SHIFT; repeat: hpte_group = ((hash & htab_hash_mask) * HPTES_PER_GROUP) & ~0x7UL; - /* Update the linux pte with the HPTE slot */ - pte_val(new_pte) &= ~_PAGE_HPTEFLAGS; - pte_val(new_pte) |= _PAGE_HASHPTE; + /* clear HPTE slot informations in new PTE */ + new_pte = (new_pte & ~_PAGE_HPTEFLAGS) | _PAGE_HASHPTE; /* Add in WIMG bits */ /* XXX We should store these in the pte */ + /* --BenH: I think they are ... */ rflags |= _PAGE_COHERENT; - slot = ppc_md.hpte_insert(hpte_group, va, prpn, - HPTE_V_LARGE, rflags); + /* Insert into the hash table, primary slot */ + slot = ppc_md.hpte_insert(hpte_group, va, pa, rflags, 0, + mmu_huge_psize); /* Primary is full, try the secondary */ if (unlikely(slot == -1)) { - pte_val(new_pte) |= _PAGE_SECONDARY; + new_pte |= _PAGE_F_SECOND; hpte_group = ((~hash & htab_hash_mask) * HPTES_PER_GROUP) & ~0x7UL; - slot = ppc_md.hpte_insert(hpte_group, va, prpn, - HPTE_V_LARGE | + slot = ppc_md.hpte_insert(hpte_group, va, pa, rflags, HPTE_V_SECONDARY, - rflags); + mmu_huge_psize); if (slot == -1) { if (mftb() & 0x1) hpte_group = ((hash & htab_hash_mask) * @@ -726,20 +744,18 @@ repeat: if (unlikely(slot == -2)) panic("hash_huge_page: pte_insert failed\n"); - pte_val(new_pte) |= (slot<<12) & _PAGE_GROUP_IX; - - /* - * No need to use ldarx/stdcx here because all who - * might be updating the pte will hold the - * page_table_lock - */ - *ptep = new_pte; + new_pte |= (slot << 12) & _PAGE_F_GIX; } + /* + * No need to use ldarx/stdcx here because all who + * might be updating the pte will hold the + * page_table_lock + */ + *ptep = __pte(new_pte & ~_PAGE_BUSY); + err = 0; out: - spin_unlock(&mm->page_table_lock); - return err; } diff --git a/arch/powerpc/mm/init_64.c b/arch/powerpc/mm/init_64.c index b0fc822ec29f..dfe7fa37b41a 100644 --- a/arch/powerpc/mm/init_64.c +++ b/arch/powerpc/mm/init_64.c @@ -188,12 +188,21 @@ static void zero_ctor(void *addr, kmem_cache_t *cache, unsigned long flags) memset(addr, 0, kmem_cache_size(cache)); } +#ifdef CONFIG_PPC_64K_PAGES +static const int pgtable_cache_size[2] = { + PTE_TABLE_SIZE, PGD_TABLE_SIZE +}; +static const char *pgtable_cache_name[ARRAY_SIZE(pgtable_cache_size)] = { + "pte_pmd_cache", "pgd_cache", +}; +#else static const int pgtable_cache_size[2] = { PTE_TABLE_SIZE, PMD_TABLE_SIZE }; static const char *pgtable_cache_name[ARRAY_SIZE(pgtable_cache_size)] = { "pgd_pte_cache", "pud_pmd_cache", }; +#endif /* CONFIG_PPC_64K_PAGES */ kmem_cache_t *pgtable_cache[ARRAY_SIZE(pgtable_cache_size)]; @@ -201,19 +210,14 @@ void pgtable_cache_init(void) { int i; - BUILD_BUG_ON(PTE_TABLE_SIZE != pgtable_cache_size[PTE_CACHE_NUM]); - BUILD_BUG_ON(PMD_TABLE_SIZE != pgtable_cache_size[PMD_CACHE_NUM]); - BUILD_BUG_ON(PUD_TABLE_SIZE != pgtable_cache_size[PUD_CACHE_NUM]); - BUILD_BUG_ON(PGD_TABLE_SIZE != pgtable_cache_size[PGD_CACHE_NUM]); - for (i = 0; i < ARRAY_SIZE(pgtable_cache_size); i++) { int size = pgtable_cache_size[i]; const char *name = pgtable_cache_name[i]; pgtable_cache[i] = kmem_cache_create(name, size, size, - SLAB_HWCACHE_ALIGN - | SLAB_MUST_HWCACHE_ALIGN, + SLAB_HWCACHE_ALIGN | + SLAB_MUST_HWCACHE_ALIGN, zero_ctor, NULL); if (! pgtable_cache[i]) diff --git a/arch/powerpc/mm/mem.c b/arch/powerpc/mm/mem.c index 117b00012e14..7faa46b71f21 100644 --- a/arch/powerpc/mm/mem.c +++ b/arch/powerpc/mm/mem.c @@ -61,6 +61,9 @@ int init_bootmem_done; int mem_init_done; unsigned long memory_limit; +extern void hash_preload(struct mm_struct *mm, unsigned long ea, + unsigned long access, unsigned long trap); + /* * This is called by /dev/mem to know if a given address has to * be mapped non-cacheable or not @@ -493,18 +496,10 @@ EXPORT_SYMBOL(flush_icache_user_range); void update_mmu_cache(struct vm_area_struct *vma, unsigned long address, pte_t pte) { - /* handle i-cache coherency */ - unsigned long pfn = pte_pfn(pte); -#ifdef CONFIG_PPC32 - pmd_t *pmd; -#else - unsigned long vsid; - void *pgdir; - pte_t *ptep; - int local = 0; - cpumask_t tmp; - unsigned long flags; +#ifdef CONFIG_PPC_STD_MMU + unsigned long access = 0, trap; #endif + unsigned long pfn = pte_pfn(pte); /* handle i-cache coherency */ if (!cpu_has_feature(CPU_FTR_COHERENT_ICACHE) && @@ -535,30 +530,21 @@ void update_mmu_cache(struct vm_area_struct *vma, unsigned long address, /* We only want HPTEs for linux PTEs that have _PAGE_ACCESSED set */ if (!pte_young(pte) || address >= TASK_SIZE) return; -#ifdef CONFIG_PPC32 - if (Hash == 0) - return; - pmd = pmd_offset(pgd_offset(vma->vm_mm, address), address); - if (!pmd_none(*pmd)) - add_hash_page(vma->vm_mm->context, address, pmd_val(*pmd)); -#else - pgdir = vma->vm_mm->pgd; - if (pgdir == NULL) - return; - ptep = find_linux_pte(pgdir, address); - if (!ptep) + /* We try to figure out if we are coming from an instruction + * access fault and pass that down to __hash_page so we avoid + * double-faulting on execution of fresh text. We have to test + * for regs NULL since init will get here first thing at boot + * + * We also avoid filling the hash if not coming from a fault + */ + if (current->thread.regs == NULL) return; - - vsid = get_vsid(vma->vm_mm->context.id, address); - - local_irq_save(flags); - tmp = cpumask_of_cpu(smp_processor_id()); - if (cpus_equal(vma->vm_mm->cpu_vm_mask, tmp)) - local = 1; - - __hash_page(address, 0, vsid, ptep, 0x300, local); - local_irq_restore(flags); -#endif -#endif + trap = TRAP(current->thread.regs); + if (trap == 0x400) + access |= _PAGE_EXEC; + else if (trap != 0x300) + return; + hash_preload(vma->vm_mm, address, access, trap); +#endif /* CONFIG_PPC_STD_MMU */ } diff --git a/arch/powerpc/mm/pgtable_64.c b/arch/powerpc/mm/pgtable_64.c index b79a78206135..51b786940971 100644 --- a/arch/powerpc/mm/pgtable_64.c +++ b/arch/powerpc/mm/pgtable_64.c @@ -101,7 +101,6 @@ static int map_io_page(unsigned long ea, unsigned long pa, int flags) pud_t *pudp; pmd_t *pmdp; pte_t *ptep; - unsigned long vsid; if (mem_init_done) { pgdp = pgd_offset_k(ea); @@ -117,28 +116,15 @@ static int map_io_page(unsigned long ea, unsigned long pa, int flags) set_pte_at(&init_mm, ea, ptep, pfn_pte(pa >> PAGE_SHIFT, __pgprot(flags))); } else { - unsigned long va, vpn, hash, hpteg; - /* * If the mm subsystem is not fully up, we cannot create a * linux page table entry for this mapping. Simply bolt an * entry in the hardware page table. + * */ - vsid = get_kernel_vsid(ea); - va = (vsid << 28) | (ea & 0xFFFFFFF); - vpn = va >> PAGE_SHIFT; - - hash = hpt_hash(vpn, 0); - - hpteg = ((hash & htab_hash_mask) * HPTES_PER_GROUP); - - /* Panic if a pte grpup is full */ - if (ppc_md.hpte_insert(hpteg, va, pa >> PAGE_SHIFT, - HPTE_V_BOLTED, - _PAGE_NO_CACHE|_PAGE_GUARDED|PP_RWXX) - == -1) { - panic("map_io_page: could not insert mapping"); - } + if (htab_bolt_mapping(ea, ea + PAGE_SIZE, pa, flags, + mmu_virtual_psize)) + panic("Can't map bolted IO mapping"); } return 0; } diff --git a/arch/powerpc/mm/ppc_mmu_32.c b/arch/powerpc/mm/ppc_mmu_32.c index cef9e83cc7e9..d137abd241ff 100644 --- a/arch/powerpc/mm/ppc_mmu_32.c +++ b/arch/powerpc/mm/ppc_mmu_32.c @@ -178,6 +178,21 @@ void __init setbat(int index, unsigned long virt, unsigned long phys, bat_addrs[index].phys = phys; } +/* + * Preload a translation in the hash table + */ +void hash_preload(struct mm_struct *mm, unsigned long ea, + unsigned long access, unsigned long trap) +{ + pmd_t *pmd; + + if (Hash == 0) + return; + pmd = pmd_offset(pgd_offset(vma->vm_mm, address), address); + if (!pmd_none(*pmd)) + add_hash_page(vma->vm_mm->context, address, pmd_val(*pmd)); +} + /* * Initialize the hash table and patch the instructions in hashtable.S. */ diff --git a/arch/powerpc/mm/slb.c b/arch/powerpc/mm/slb.c index 0473953f6a37..60e852f2f8e5 100644 --- a/arch/powerpc/mm/slb.c +++ b/arch/powerpc/mm/slb.c @@ -14,14 +14,32 @@ * 2 of the License, or (at your option) any later version. */ +#undef DEBUG + #include #include #include #include #include #include +#include + +#ifdef DEBUG +#define DBG(fmt...) udbg_printf(fmt) +#else +#define DBG(fmt...) +#endif -extern void slb_allocate(unsigned long ea); +extern void slb_allocate_realmode(unsigned long ea); +extern void slb_allocate_user(unsigned long ea); + +static void slb_allocate(unsigned long ea) +{ + /* Currently, we do real mode for all SLBs including user, but + * that will change if we bring back dynamic VSIDs + */ + slb_allocate_realmode(ea); +} static inline unsigned long mk_esid_data(unsigned long ea, unsigned long slot) { @@ -46,13 +64,15 @@ static void slb_flush_and_rebolt(void) { /* If you change this make sure you change SLB_NUM_BOLTED * appropriately too. */ - unsigned long ksp_flags = SLB_VSID_KERNEL; + unsigned long linear_llp, virtual_llp, lflags, vflags; unsigned long ksp_esid_data; WARN_ON(!irqs_disabled()); - if (cpu_has_feature(CPU_FTR_16M_PAGE)) - ksp_flags |= SLB_VSID_L; + linear_llp = mmu_psize_defs[mmu_linear_psize].sllp; + virtual_llp = mmu_psize_defs[mmu_virtual_psize].sllp; + lflags = SLB_VSID_KERNEL | linear_llp; + vflags = SLB_VSID_KERNEL | virtual_llp; ksp_esid_data = mk_esid_data(get_paca()->kstack, 2); if ((ksp_esid_data & ESID_MASK) == KERNELBASE) @@ -67,9 +87,9 @@ static void slb_flush_and_rebolt(void) /* Slot 2 - kernel stack */ "slbmte %2,%3\n" "isync" - :: "r"(mk_vsid_data(VMALLOCBASE, SLB_VSID_KERNEL)), + :: "r"(mk_vsid_data(VMALLOCBASE, vflags)), "r"(mk_esid_data(VMALLOCBASE, 1)), - "r"(mk_vsid_data(ksp_esid_data, ksp_flags)), + "r"(mk_vsid_data(ksp_esid_data, lflags)), "r"(ksp_esid_data) : "memory"); } @@ -102,6 +122,9 @@ void switch_slb(struct task_struct *tsk, struct mm_struct *mm) get_paca()->slb_cache_ptr = 0; get_paca()->context = mm->context; +#ifdef CONFIG_PPC_64K_PAGES + get_paca()->pgdir = mm->pgd; +#endif /* CONFIG_PPC_64K_PAGES */ /* * preload some userspace segments into the SLB. @@ -131,28 +154,77 @@ void switch_slb(struct task_struct *tsk, struct mm_struct *mm) slb_allocate(unmapped_base); } +static inline void patch_slb_encoding(unsigned int *insn_addr, + unsigned int immed) +{ + /* Assume the instruction had a "0" immediate value, just + * "or" in the new value + */ + *insn_addr |= immed; + flush_icache_range((unsigned long)insn_addr, 4+ + (unsigned long)insn_addr); +} + void slb_initialize(void) { + unsigned long linear_llp, virtual_llp; + static int slb_encoding_inited; + extern unsigned int *slb_miss_kernel_load_linear; + extern unsigned int *slb_miss_kernel_load_virtual; + extern unsigned int *slb_miss_user_load_normal; +#ifdef CONFIG_HUGETLB_PAGE + extern unsigned int *slb_miss_user_load_huge; + unsigned long huge_llp; + + huge_llp = mmu_psize_defs[mmu_huge_psize].sllp; +#endif + + /* Prepare our SLB miss handler based on our page size */ + linear_llp = mmu_psize_defs[mmu_linear_psize].sllp; + virtual_llp = mmu_psize_defs[mmu_virtual_psize].sllp; + if (!slb_encoding_inited) { + slb_encoding_inited = 1; + patch_slb_encoding(slb_miss_kernel_load_linear, + SLB_VSID_KERNEL | linear_llp); + patch_slb_encoding(slb_miss_kernel_load_virtual, + SLB_VSID_KERNEL | virtual_llp); + patch_slb_encoding(slb_miss_user_load_normal, + SLB_VSID_USER | virtual_llp); + + DBG("SLB: linear LLP = %04x\n", linear_llp); + DBG("SLB: virtual LLP = %04x\n", virtual_llp); +#ifdef CONFIG_HUGETLB_PAGE + patch_slb_encoding(slb_miss_user_load_huge, + SLB_VSID_USER | huge_llp); + DBG("SLB: huge LLP = %04x\n", huge_llp); +#endif + } + /* On iSeries the bolted entries have already been set up by * the hypervisor from the lparMap data in head.S */ #ifndef CONFIG_PPC_ISERIES - unsigned long flags = SLB_VSID_KERNEL; + { + unsigned long lflags, vflags; - /* Invalidate the entire SLB (even slot 0) & all the ERATS */ - if (cpu_has_feature(CPU_FTR_16M_PAGE)) - flags |= SLB_VSID_L; + lflags = SLB_VSID_KERNEL | linear_llp; + vflags = SLB_VSID_KERNEL | virtual_llp; - asm volatile("isync":::"memory"); - asm volatile("slbmte %0,%0"::"r" (0) : "memory"); + /* Invalidate the entire SLB (even slot 0) & all the ERATS */ + asm volatile("isync":::"memory"); + asm volatile("slbmte %0,%0"::"r" (0) : "memory"); asm volatile("isync; slbia; isync":::"memory"); - create_slbe(KERNELBASE, flags, 0); - create_slbe(VMALLOCBASE, SLB_VSID_KERNEL, 1); + create_slbe(KERNELBASE, lflags, 0); + + /* VMALLOC space has 4K pages always for now */ + create_slbe(VMALLOCBASE, vflags, 1); + /* We don't bolt the stack for the time being - we're in boot, * so the stack is in the bolted segment. By the time it goes * elsewhere, we'll call _switch() which will bolt in the new * one. */ asm volatile("isync":::"memory"); -#endif + } +#endif /* CONFIG_PPC_ISERIES */ get_paca()->stab_rr = SLB_NUM_BOLTED; } diff --git a/arch/powerpc/mm/slb_low.S b/arch/powerpc/mm/slb_low.S index a3a03da503bc..3e18241b6f35 100644 --- a/arch/powerpc/mm/slb_low.S +++ b/arch/powerpc/mm/slb_low.S @@ -18,61 +18,28 @@ #include #include -#include -#include #include #include #include +#include +#include +#include -/* void slb_allocate(unsigned long ea); +/* void slb_allocate_realmode(unsigned long ea); * * Create an SLB entry for the given EA (user or kernel). * r3 = faulting address, r13 = PACA * r9, r10, r11 are clobbered by this function * No other registers are examined or changed. */ -_GLOBAL(slb_allocate) - /* - * First find a slot, round robin. Previously we tried to find - * a free slot first but that took too long. Unfortunately we - * dont have any LRU information to help us choose a slot. - */ -#ifdef CONFIG_PPC_ISERIES - /* - * On iSeries, the "bolted" stack segment can be cast out on - * shared processor switch so we need to check for a miss on - * it and restore it to the right slot. - */ - ld r9,PACAKSAVE(r13) - clrrdi r9,r9,28 - clrrdi r11,r3,28 - li r10,SLB_NUM_BOLTED-1 /* Stack goes in last bolted slot */ - cmpld r9,r11 - beq 3f -#endif /* CONFIG_PPC_ISERIES */ - - ld r10,PACASTABRR(r13) - addi r10,r10,1 - /* use a cpu feature mask if we ever change our slb size */ - cmpldi r10,SLB_NUM_ENTRIES - - blt+ 4f - li r10,SLB_NUM_BOLTED - -4: - std r10,PACASTABRR(r13) -3: - /* r3 = faulting address, r10 = entry */ +_GLOBAL(slb_allocate_realmode) + /* r3 = faulting address */ srdi r9,r3,60 /* get region */ - srdi r3,r3,28 /* get esid */ + srdi r10,r3,28 /* get esid */ cmpldi cr7,r9,0xc /* cmp KERNELBASE for later use */ - rldimi r10,r3,28,0 /* r10= ESID<<28 | entry */ - oris r10,r10,SLB_ESID_V@h /* r10 |= SLB_ESID_V */ - - /* r3 = esid, r10 = esid_data, cr7 = <>KERNELBASE */ - + /* r3 = address, r10 = esid, cr7 = <>KERNELBASE */ blt cr7,0f /* user or kernel? */ /* kernel address: proto-VSID = ESID */ @@ -81,43 +48,161 @@ _GLOBAL(slb_allocate) * top segment. That's ok, the scramble below will translate * it to VSID 0, which is reserved as a bad VSID - one which * will never have any pages in it. */ - li r11,SLB_VSID_KERNEL -BEGIN_FTR_SECTION - bne cr7,9f - li r11,(SLB_VSID_KERNEL|SLB_VSID_L) -END_FTR_SECTION_IFSET(CPU_FTR_16M_PAGE) - b 9f -0: /* user address: proto-VSID = context<<15 | ESID */ - srdi. r9,r3,USER_ESID_BITS + /* Check if hitting the linear mapping of the vmalloc/ioremap + * kernel space + */ + bne cr7,1f + + /* Linear mapping encoding bits, the "li" instruction below will + * be patched by the kernel at boot + */ +_GLOBAL(slb_miss_kernel_load_linear) + li r11,0 + b slb_finish_load + +1: /* vmalloc/ioremap mapping encoding bits, the "li" instruction below + * will be patched by the kernel at boot + */ +_GLOBAL(slb_miss_kernel_load_virtual) + li r11,0 + b slb_finish_load + + +0: /* user address: proto-VSID = context << 15 | ESID. First check + * if the address is within the boundaries of the user region + */ + srdi. r9,r10,USER_ESID_BITS bne- 8f /* invalid ea bits set */ + /* Figure out if the segment contains huge pages */ #ifdef CONFIG_HUGETLB_PAGE BEGIN_FTR_SECTION + b 1f +END_FTR_SECTION_IFCLR(CPU_FTR_16M_PAGE) lhz r9,PACAHIGHHTLBAREAS(r13) - srdi r11,r3,(HTLB_AREA_SHIFT-SID_SHIFT) + srdi r11,r10,(HTLB_AREA_SHIFT-SID_SHIFT) srd r9,r9,r11 lhz r11,PACALOWHTLBAREAS(r13) - srd r11,r11,r3 - or r9,r9,r11 -END_FTR_SECTION_IFSET(CPU_FTR_16M_PAGE) + srd r11,r11,r10 + or. r9,r9,r11 + beq 1f +_GLOBAL(slb_miss_user_load_huge) + li r11,0 + b 2f +1: #endif /* CONFIG_HUGETLB_PAGE */ - li r11,SLB_VSID_USER +_GLOBAL(slb_miss_user_load_normal) + li r11,0 -#ifdef CONFIG_HUGETLB_PAGE -BEGIN_FTR_SECTION - rldimi r11,r9,8,55 /* shift masked bit into SLB_VSID_L */ -END_FTR_SECTION_IFSET(CPU_FTR_16M_PAGE) -#endif /* CONFIG_HUGETLB_PAGE */ +2: + ld r9,PACACONTEXTID(r13) + rldimi r10,r9,USER_ESID_BITS,0 + b slb_finish_load + +8: /* invalid EA */ + li r10,0 /* BAD_VSID */ + li r11,SLB_VSID_USER /* flags don't much matter */ + b slb_finish_load + +#ifdef __DISABLED__ + +/* void slb_allocate_user(unsigned long ea); + * + * Create an SLB entry for the given EA (user or kernel). + * r3 = faulting address, r13 = PACA + * r9, r10, r11 are clobbered by this function + * No other registers are examined or changed. + * + * It is called with translation enabled in order to be able to walk the + * page tables. This is not currently used. + */ +_GLOBAL(slb_allocate_user) + /* r3 = faulting address */ + srdi r10,r3,28 /* get esid */ + + crset 4*cr7+lt /* set "user" flag for later */ + + /* check if we fit in the range covered by the pagetables*/ + srdi. r9,r3,PGTABLE_EADDR_SIZE + crnot 4*cr0+eq,4*cr0+eq + beqlr + /* now we need to get to the page tables in order to get the page + * size encoding from the PMD. In the future, we'll be able to deal + * with 1T segments too by getting the encoding from the PGD instead + */ + ld r9,PACAPGDIR(r13) + cmpldi cr0,r9,0 + beqlr + rlwinm r11,r10,8,25,28 + ldx r9,r9,r11 /* get pgd_t */ + cmpldi cr0,r9,0 + beqlr + rlwinm r11,r10,3,17,28 + ldx r9,r9,r11 /* get pmd_t */ + cmpldi cr0,r9,0 + beqlr + + /* build vsid flags */ + andi. r11,r9,SLB_VSID_LLP + ori r11,r11,SLB_VSID_USER + + /* get context to calculate proto-VSID */ ld r9,PACACONTEXTID(r13) - rldimi r3,r9,USER_ESID_BITS,0 + rldimi r10,r9,USER_ESID_BITS,0 + + /* fall through slb_finish_load */ + +#endif /* __DISABLED__ */ -9: /* r3 = protovsid, r11 = flags, r10 = esid_data, cr7 = <>KERNELBASE */ - ASM_VSID_SCRAMBLE(r3,r9) - rldimi r11,r3,SLB_VSID_SHIFT,16 /* combine VSID and flags */ +/* + * Finish loading of an SLB entry and return + * + * r3 = EA, r10 = proto-VSID, r11 = flags, clobbers r9, cr7 = <>KERNELBASE + */ +slb_finish_load: + ASM_VSID_SCRAMBLE(r10,r9) + rldimi r11,r10,SLB_VSID_SHIFT,16 /* combine VSID and flags */ + + /* r3 = EA, r11 = VSID data */ + /* + * Find a slot, round robin. Previously we tried to find a + * free slot first but that took too long. Unfortunately we + * dont have any LRU information to help us choose a slot. + */ +#ifdef CONFIG_PPC_ISERIES + /* + * On iSeries, the "bolted" stack segment can be cast out on + * shared processor switch so we need to check for a miss on + * it and restore it to the right slot. + */ + ld r9,PACAKSAVE(r13) + clrrdi r9,r9,28 + clrrdi r3,r3,28 + li r10,SLB_NUM_BOLTED-1 /* Stack goes in last bolted slot */ + cmpld r9,r3 + beq 3f +#endif /* CONFIG_PPC_ISERIES */ + + ld r10,PACASTABRR(r13) + addi r10,r10,1 + /* use a cpu feature mask if we ever change our slb size */ + cmpldi r10,SLB_NUM_ENTRIES + + blt+ 4f + li r10,SLB_NUM_BOLTED + +4: + std r10,PACASTABRR(r13) + +3: + rldimi r3,r10,0,36 /* r3= EA[0:35] | entry */ + oris r10,r3,SLB_ESID_V@h /* r3 |= SLB_ESID_V */ + + /* r3 = ESID data, r11 = VSID data */ /* * No need for an isync before or after this slbmte. The exception @@ -125,7 +210,9 @@ END_FTR_SECTION_IFSET(CPU_FTR_16M_PAGE) */ slbmte r11,r10 - bgelr cr7 /* we're done for kernel addresses */ + /* we're done for kernel addresses */ + crclr 4*cr0+eq /* set result to "success" */ + bgelr cr7 /* Update the slb cache */ lhz r3,PACASLBCACHEPTR(r13) /* offset = paca->slb_cache_ptr */ @@ -143,9 +230,6 @@ END_FTR_SECTION_IFSET(CPU_FTR_16M_PAGE) li r3,SLB_CACHE_ENTRIES+1 2: sth r3,PACASLBCACHEPTR(r13) /* paca->slb_cache_ptr = offset */ + crclr 4*cr0+eq /* set result to "success" */ blr -8: /* invalid EA */ - li r3,0 /* BAD_VSID */ - li r11,SLB_VSID_USER /* flags don't much matter */ - b 9b diff --git a/arch/powerpc/mm/stab.c b/arch/powerpc/mm/stab.c index 1b83f002bf27..fa325dbf98fc 100644 --- a/arch/powerpc/mm/stab.c +++ b/arch/powerpc/mm/stab.c @@ -26,7 +26,6 @@ struct stab_entry { unsigned long vsid_data; }; -/* Both the segment table and SLB code uses the following cache */ #define NR_STAB_CACHE_ENTRIES 8 DEFINE_PER_CPU(long, stab_cache_ptr); DEFINE_PER_CPU(long, stab_cache[NR_STAB_CACHE_ENTRIES]); @@ -186,7 +185,7 @@ void switch_stab(struct task_struct *tsk, struct mm_struct *mm) /* Never flush the first entry. */ ste += 1; for (entry = 1; - entry < (PAGE_SIZE / sizeof(struct stab_entry)); + entry < (HW_PAGE_SIZE / sizeof(struct stab_entry)); entry++, ste++) { unsigned long ea; ea = ste->esid_data & ESID_MASK; @@ -200,6 +199,10 @@ void switch_stab(struct task_struct *tsk, struct mm_struct *mm) __get_cpu_var(stab_cache_ptr) = 0; +#ifdef CONFIG_PPC_64K_PAGES + get_paca()->pgdir = mm->pgd; +#endif /* CONFIG_PPC_64K_PAGES */ + /* Now preload some entries for the new task */ if (test_tsk_thread_flag(tsk, TIF_32BIT)) unmapped_base = TASK_UNMAPPED_BASE_USER32; @@ -223,8 +226,6 @@ void switch_stab(struct task_struct *tsk, struct mm_struct *mm) asm volatile("sync" : : : "memory"); } -extern void slb_initialize(void); - /* * Allocate segment tables for secondary CPUs. These must all go in * the first (bolted) segment, so that do_stab_bolted won't get a @@ -243,18 +244,21 @@ void stabs_alloc(void) if (cpu == 0) continue; /* stab for CPU 0 is statically allocated */ - newstab = lmb_alloc_base(PAGE_SIZE, PAGE_SIZE, 1< #include #include @@ -30,7 +31,7 @@ #include #include #include -#include +#include DEFINE_PER_CPU(struct ppc64_tlb_batch, ppc64_tlb_batch); @@ -126,28 +127,46 @@ void pgtable_free_tlb(struct mmu_gather *tlb, pgtable_free_t pgf) * (if we remove it we should clear the _PTE_HPTEFLAGS bits). */ void hpte_update(struct mm_struct *mm, unsigned long addr, - unsigned long pte, int wrprot) + pte_t *ptep, unsigned long pte, int huge) { struct ppc64_tlb_batch *batch = &__get_cpu_var(ppc64_tlb_batch); unsigned long vsid; + unsigned int psize = mmu_virtual_psize; int i; i = batch->index; + /* We mask the address for the base page size. Huge pages will + * have applied their own masking already + */ + addr &= PAGE_MASK; + + /* Get page size (maybe move back to caller) */ + if (huge) { +#ifdef CONFIG_HUGETLB_PAGE + psize = mmu_huge_psize; +#else + BUG(); +#endif + } + /* * This can happen when we are in the middle of a TLB batch and * we encounter memory pressure (eg copy_page_range when it tries * to allocate a new pte). If we have to reclaim memory and end * up scanning and resetting referenced bits then our batch context * will change mid stream. + * + * We also need to ensure only one page size is present in a given + * batch */ - if (i != 0 && (mm != batch->mm || batch->large != pte_huge(pte))) { + if (i != 0 && (mm != batch->mm || batch->psize != psize)) { flush_tlb_pending(); i = 0; } if (i == 0) { batch->mm = mm; - batch->large = pte_huge(pte); + batch->psize = psize; } if (addr < KERNELBASE) { vsid = get_vsid(mm->context.id, addr); @@ -155,7 +174,7 @@ void hpte_update(struct mm_struct *mm, unsigned long addr, } else vsid = get_kernel_vsid(addr); batch->vaddr[i] = (vsid << 28 ) | (addr & 0x0fffffff); - batch->pte[i] = __pte(pte); + batch->pte[i] = __real_pte(__pte(pte), ptep); batch->index = ++i; if (i >= PPC64_TLB_BATCH_NR) flush_tlb_pending(); @@ -177,7 +196,8 @@ void __flush_tlb_pending(struct ppc64_tlb_batch *batch) local = 1; if (i == 1) - flush_hash_page(batch->vaddr[0], batch->pte[0], local); + flush_hash_page(batch->vaddr[0], batch->pte[0], + batch->psize, local); else flush_hash_range(i, local); batch->index = 0; diff --git a/arch/powerpc/platforms/iseries/htab.c b/arch/powerpc/platforms/iseries/htab.c index b3c6c3374ca6..30bdcf3925d9 100644 --- a/arch/powerpc/platforms/iseries/htab.c +++ b/arch/powerpc/platforms/iseries/htab.c @@ -39,15 +39,16 @@ static inline void iSeries_hunlock(unsigned long slot) spin_unlock(&iSeries_hlocks[(slot >> 4) & 0x3f]); } -static long iSeries_hpte_insert(unsigned long hpte_group, unsigned long va, - unsigned long prpn, unsigned long vflags, - unsigned long rflags) +long iSeries_hpte_insert(unsigned long hpte_group, unsigned long va, + unsigned long pa, unsigned long rflags, + unsigned long vflags, int psize) { - unsigned long arpn; long slot; hpte_t lhpte; int secondary = 0; + BUG_ON(psize != MMU_PAGE_4K); + /* * The hypervisor tries both primary and secondary. * If we are being called to insert in the secondary, @@ -59,8 +60,19 @@ static long iSeries_hpte_insert(unsigned long hpte_group, unsigned long va, iSeries_hlock(hpte_group); - slot = HvCallHpt_findValid(&lhpte, va >> PAGE_SHIFT); - BUG_ON(lhpte.v & HPTE_V_VALID); + slot = HvCallHpt_findValid(&lhpte, va >> HW_PAGE_SHIFT); + if (unlikely(lhpte.v & HPTE_V_VALID)) { + if (vflags & HPTE_V_BOLTED) { + HvCallHpt_setSwBits(slot, 0x10, 0); + HvCallHpt_setPp(slot, PP_RWXX); + iSeries_hunlock(hpte_group); + if (slot < 0) + return 0x8 | (slot & 7); + else + return slot & 7; + } + BUG(); + } if (slot == -1) { /* No available entry found in either group */ iSeries_hunlock(hpte_group); @@ -73,10 +85,9 @@ static long iSeries_hpte_insert(unsigned long hpte_group, unsigned long va, slot &= 0x7fffffffffffffff; } - arpn = phys_to_abs(prpn << PAGE_SHIFT) >> PAGE_SHIFT; - lhpte.v = (va >> 23) << HPTE_V_AVPN_SHIFT | vflags | HPTE_V_VALID; - lhpte.r = (arpn << HPTE_R_RPN_SHIFT) | rflags; + lhpte.v = hpte_encode_v(va, MMU_PAGE_4K) | vflags | HPTE_V_VALID; + lhpte.r = hpte_encode_r(phys_to_abs(pa), MMU_PAGE_4K) | rflags; /* Now fill in the actual HPTE */ HvCallHpt_addValidate(slot, secondary, &lhpte); @@ -86,25 +97,6 @@ static long iSeries_hpte_insert(unsigned long hpte_group, unsigned long va, return (secondary << 3) | (slot & 7); } -long iSeries_hpte_bolt_or_insert(unsigned long hpte_group, - unsigned long va, unsigned long prpn, unsigned long vflags, - unsigned long rflags) -{ - long slot; - hpte_t lhpte; - - slot = HvCallHpt_findValid(&lhpte, va >> PAGE_SHIFT); - - if (lhpte.v & HPTE_V_VALID) { - /* Bolt the existing HPTE */ - HvCallHpt_setSwBits(slot, 0x10, 0); - HvCallHpt_setPp(slot, PP_RWXX); - return 0; - } - - return iSeries_hpte_insert(hpte_group, va, prpn, vflags, rflags); -} - static unsigned long iSeries_hpte_getword0(unsigned long slot) { hpte_t hpte; @@ -150,15 +142,17 @@ static long iSeries_hpte_remove(unsigned long hpte_group) * bits 61..63 : PP2,PP1,PP0 */ static long iSeries_hpte_updatepp(unsigned long slot, unsigned long newpp, - unsigned long va, int large, int local) + unsigned long va, int psize, int local) { hpte_t hpte; - unsigned long avpn = va >> 23; + unsigned long want_v; iSeries_hlock(slot); HvCallHpt_get(&hpte, slot); - if ((HPTE_V_AVPN_VAL(hpte.v) == avpn) && (hpte.v & HPTE_V_VALID)) { + want_v = hpte_encode_v(va, MMU_PAGE_4K); + + if (HPTE_V_COMPARE(hpte.v, want_v) && (hpte.v & HPTE_V_VALID)) { /* * Hypervisor expects bits as NPPP, which is * different from how they are mapped in our PP. @@ -210,14 +204,17 @@ static long iSeries_hpte_find(unsigned long vpn) * * No need to lock here because we should be the only user. */ -static void iSeries_hpte_updateboltedpp(unsigned long newpp, unsigned long ea) +static void iSeries_hpte_updateboltedpp(unsigned long newpp, unsigned long ea, + int psize) { unsigned long vsid,va,vpn; long slot; + BUG_ON(psize != MMU_PAGE_4K); + vsid = get_kernel_vsid(ea); va = (vsid << 28) | (ea & 0x0fffffff); - vpn = va >> PAGE_SHIFT; + vpn = va >> HW_PAGE_SHIFT; slot = iSeries_hpte_find(vpn); if (slot == -1) panic("updateboltedpp: Could not find page to bolt\n"); @@ -225,7 +222,7 @@ static void iSeries_hpte_updateboltedpp(unsigned long newpp, unsigned long ea) } static void iSeries_hpte_invalidate(unsigned long slot, unsigned long va, - int large, int local) + int psize, int local) { unsigned long hpte_v; unsigned long avpn = va >> 23; diff --git a/arch/powerpc/platforms/iseries/hvlog.c b/arch/powerpc/platforms/iseries/hvlog.c index 62ec73479687..f476d71194fa 100644 --- a/arch/powerpc/platforms/iseries/hvlog.c +++ b/arch/powerpc/platforms/iseries/hvlog.c @@ -22,7 +22,7 @@ void HvCall_writeLogBuffer(const void *buffer, u64 len) while (len) { hv_buf.addr = cur; - left_this_page = ((cur & PAGE_MASK) + PAGE_SIZE) - cur; + left_this_page = ((cur & HW_PAGE_MASK) + HW_PAGE_SIZE) - cur; if (left_this_page > len) left_this_page = len; hv_buf.len = left_this_page; @@ -30,6 +30,6 @@ void HvCall_writeLogBuffer(const void *buffer, u64 len) HvCall2(HvCallBaseWriteLogBuffer, virt_to_abs(&hv_buf), left_this_page); - cur = (cur & PAGE_MASK) + PAGE_SIZE; + cur = (cur & HW_PAGE_MASK) + HW_PAGE_SIZE; } } diff --git a/arch/powerpc/platforms/iseries/iommu.c b/arch/powerpc/platforms/iseries/iommu.c index 1a6845b5c5a4..bf081b345820 100644 --- a/arch/powerpc/platforms/iseries/iommu.c +++ b/arch/powerpc/platforms/iseries/iommu.c @@ -43,9 +43,12 @@ static void tce_build_iSeries(struct iommu_table *tbl, long index, long npages, u64 rc; union tce_entry tce; + index <<= TCE_PAGE_FACTOR; + npages <<= TCE_PAGE_FACTOR; + while (npages--) { tce.te_word = 0; - tce.te_bits.tb_rpn = virt_to_abs(uaddr) >> PAGE_SHIFT; + tce.te_bits.tb_rpn = virt_to_abs(uaddr) >> TCE_SHIFT; if (tbl->it_type == TCE_VB) { /* Virtual Bus */ @@ -66,7 +69,7 @@ static void tce_build_iSeries(struct iommu_table *tbl, long index, long npages, panic("PCI_DMA: HvCallXm_setTce failed, Rc: 0x%lx\n", rc); index++; - uaddr += PAGE_SIZE; + uaddr += TCE_PAGE_SIZE; } } @@ -74,6 +77,9 @@ static void tce_free_iSeries(struct iommu_table *tbl, long index, long npages) { u64 rc; + npages <<= TCE_PAGE_FACTOR; + index <<= TCE_PAGE_FACTOR; + while (npages--) { rc = HvCallXm_setTce((u64)tbl->it_index, (u64)index, 0); if (rc) @@ -83,27 +89,6 @@ static void tce_free_iSeries(struct iommu_table *tbl, long index, long npages) } } -#ifdef CONFIG_PCI -/* - * This function compares the known tables to find an iommu_table - * that has already been built for hardware TCEs. - */ -static struct iommu_table *iommu_table_find(struct iommu_table * tbl) -{ - struct pci_dn *pdn; - - list_for_each_entry(pdn, &iSeries_Global_Device_List, Device_List) { - struct iommu_table *it = pdn->iommu_table; - if ((it != NULL) && - (it->it_type == TCE_PCI) && - (it->it_offset == tbl->it_offset) && - (it->it_index == tbl->it_index) && - (it->it_size == tbl->it_size)) - return it; - } - return NULL; -} - /* * Call Hv with the architected data structure to get TCE table info. * info. Put the returned data into the Linux representation of the @@ -113,8 +98,10 @@ static struct iommu_table *iommu_table_find(struct iommu_table * tbl) * 2. TCE table per Bus. * 3. TCE Table per IOA. */ -static void iommu_table_getparms(struct pci_dn *pdn, - struct iommu_table* tbl) +void iommu_table_getparms_iSeries(unsigned long busno, + unsigned char slotno, + unsigned char virtbus, + struct iommu_table* tbl) { struct iommu_table_cb *parms; @@ -124,9 +111,9 @@ static void iommu_table_getparms(struct pci_dn *pdn, memset(parms, 0, sizeof(*parms)); - parms->itc_busno = pdn->busno; - parms->itc_slotno = pdn->LogicalSlot; - parms->itc_virtbus = 0; + parms->itc_busno = busno; + parms->itc_slotno = slotno; + parms->itc_virtbus = virtbus; HvCallXm_getTceTableParms(iseries_hv_addr(parms)); @@ -134,17 +121,40 @@ static void iommu_table_getparms(struct pci_dn *pdn, panic("PCI_DMA: parms->size is zero, parms is 0x%p", parms); /* itc_size is in pages worth of table, it_size is in # of entries */ - tbl->it_size = (parms->itc_size * PAGE_SIZE) / sizeof(union tce_entry); + tbl->it_size = ((parms->itc_size * TCE_PAGE_SIZE) / + sizeof(union tce_entry)) >> TCE_PAGE_FACTOR; tbl->it_busno = parms->itc_busno; - tbl->it_offset = parms->itc_offset; + tbl->it_offset = parms->itc_offset >> TCE_PAGE_FACTOR; tbl->it_index = parms->itc_index; tbl->it_blocksize = 1; - tbl->it_type = TCE_PCI; + tbl->it_type = virtbus ? TCE_VB : TCE_PCI; kfree(parms); } +#ifdef CONFIG_PCI +/* + * This function compares the known tables to find an iommu_table + * that has already been built for hardware TCEs. + */ +static struct iommu_table *iommu_table_find(struct iommu_table * tbl) +{ + struct pci_dn *pdn; + + list_for_each_entry(pdn, &iSeries_Global_Device_List, Device_List) { + struct iommu_table *it = pdn->iommu_table; + if ((it != NULL) && + (it->it_type == TCE_PCI) && + (it->it_offset == tbl->it_offset) && + (it->it_index == tbl->it_index) && + (it->it_size == tbl->it_size)) + return it; + } + return NULL; +} + + void iommu_devnode_init_iSeries(struct device_node *dn) { struct iommu_table *tbl; @@ -152,7 +162,7 @@ void iommu_devnode_init_iSeries(struct device_node *dn) tbl = kmalloc(sizeof(struct iommu_table), GFP_KERNEL); - iommu_table_getparms(pdn, tbl); + iommu_table_getparms_iSeries(pdn->busno, pdn->LogicalSlot, 0, tbl); /* Look for existing tce table */ pdn->iommu_table = iommu_table_find(tbl); diff --git a/arch/powerpc/platforms/iseries/setup.c b/arch/powerpc/platforms/iseries/setup.c index fda712b42168..c5207064977d 100644 --- a/arch/powerpc/platforms/iseries/setup.c +++ b/arch/powerpc/platforms/iseries/setup.c @@ -320,11 +320,11 @@ static void __init iSeries_init_early(void) */ if (naca.xRamDisk) { initrd_start = (unsigned long)__va(naca.xRamDisk); - initrd_end = initrd_start + naca.xRamDiskSize * PAGE_SIZE; + initrd_end = initrd_start + naca.xRamDiskSize * HW_PAGE_SIZE; initrd_below_start_ok = 1; // ramdisk in kernel space ROOT_DEV = Root_RAM0; - if (((rd_size * 1024) / PAGE_SIZE) < naca.xRamDiskSize) - rd_size = (naca.xRamDiskSize * PAGE_SIZE) / 1024; + if (((rd_size * 1024) / HW_PAGE_SIZE) < naca.xRamDiskSize) + rd_size = (naca.xRamDiskSize * HW_PAGE_SIZE) / 1024; } else #endif /* CONFIG_BLK_DEV_INITRD */ { @@ -470,13 +470,14 @@ static void __init build_iSeries_Memory_Map(void) */ hptFirstChunk = (u32)addr_to_chunk(HvCallHpt_getHptAddress()); hptSizePages = (u32)HvCallHpt_getHptPages(); - hptSizeChunks = hptSizePages >> (MSCHUNKS_CHUNK_SHIFT - PAGE_SHIFT); + hptSizeChunks = hptSizePages >> + (MSCHUNKS_CHUNK_SHIFT - HW_PAGE_SHIFT); hptLastChunk = hptFirstChunk + hptSizeChunks - 1; printk("HPT absolute addr = %016lx, size = %dK\n", chunk_to_addr(hptFirstChunk), hptSizeChunks * 256); - ppc64_pft_size = __ilog2(hptSizePages * PAGE_SIZE); + ppc64_pft_size = __ilog2(hptSizePages * HW_PAGE_SIZE); /* * The actual hashed page table is in the hypervisor, @@ -629,7 +630,7 @@ static void __init iSeries_fixup_klimit(void) */ if (naca.xRamDisk) klimit = KERNELBASE + (u64)naca.xRamDisk + - (naca.xRamDiskSize * PAGE_SIZE); + (naca.xRamDiskSize * HW_PAGE_SIZE); else { /* * No ram disk was included - check and see if there diff --git a/arch/powerpc/platforms/iseries/vio.c b/arch/powerpc/platforms/iseries/vio.c index c27a66876c2c..384360ee06ec 100644 --- a/arch/powerpc/platforms/iseries/vio.c +++ b/arch/powerpc/platforms/iseries/vio.c @@ -30,41 +30,14 @@ static struct iommu_table vio_iommu_table; static void __init iommu_vio_init(void) { - struct iommu_table *t; - struct iommu_table_cb cb; - unsigned long cbp; - unsigned long itc_entries; + iommu_table_getparms_iSeries(255, 0, 0xff, &veth_iommu_table); + veth_iommu_table.it_size /= 2; + vio_iommu_table = veth_iommu_table; + vio_iommu_table.it_offset += veth_iommu_table.it_size; - cb.itc_busno = 255; /* Bus 255 is the virtual bus */ - cb.itc_virtbus = 0xff; /* Ask for virtual bus */ - - cbp = virt_to_abs(&cb); - HvCallXm_getTceTableParms(cbp); - - itc_entries = cb.itc_size * PAGE_SIZE / sizeof(union tce_entry); - veth_iommu_table.it_size = itc_entries / 2; - veth_iommu_table.it_busno = cb.itc_busno; - veth_iommu_table.it_offset = cb.itc_offset; - veth_iommu_table.it_index = cb.itc_index; - veth_iommu_table.it_type = TCE_VB; - veth_iommu_table.it_blocksize = 1; - - t = iommu_init_table(&veth_iommu_table); - - if (!t) + if (!iommu_init_table(&veth_iommu_table)) printk("Virtual Bus VETH TCE table failed.\n"); - - vio_iommu_table.it_size = itc_entries - veth_iommu_table.it_size; - vio_iommu_table.it_busno = cb.itc_busno; - vio_iommu_table.it_offset = cb.itc_offset + - veth_iommu_table.it_size; - vio_iommu_table.it_index = cb.itc_index; - vio_iommu_table.it_type = TCE_VB; - vio_iommu_table.it_blocksize = 1; - - t = iommu_init_table(&vio_iommu_table); - - if (!t) + if (!iommu_init_table(&vio_iommu_table)) printk("Virtual Bus VIO TCE table failed.\n"); } diff --git a/arch/powerpc/platforms/iseries/viopath.c b/arch/powerpc/platforms/iseries/viopath.c index fe97bfbf7463..842672695598 100644 --- a/arch/powerpc/platforms/iseries/viopath.c +++ b/arch/powerpc/platforms/iseries/viopath.c @@ -68,7 +68,8 @@ static DEFINE_SPINLOCK(statuslock); * For each kind of event we allocate a buffer that is * guaranteed not to cross a page boundary */ -static unsigned char event_buffer[VIO_MAX_SUBTYPES * 256] __page_aligned; +static unsigned char event_buffer[VIO_MAX_SUBTYPES * 256] + __attribute__((__aligned__(4096))); static atomic_t event_buffer_available[VIO_MAX_SUBTYPES]; static int event_buffer_initialised; @@ -116,12 +117,12 @@ static int proc_viopath_show(struct seq_file *m, void *v) HvLpEvent_Rc hvrc; DECLARE_MUTEX_LOCKED(Semaphore); - buf = kmalloc(PAGE_SIZE, GFP_KERNEL); + buf = kmalloc(HW_PAGE_SIZE, GFP_KERNEL); if (!buf) return 0; - memset(buf, 0, PAGE_SIZE); + memset(buf, 0, HW_PAGE_SIZE); - handle = dma_map_single(iSeries_vio_dev, buf, PAGE_SIZE, + handle = dma_map_single(iSeries_vio_dev, buf, HW_PAGE_SIZE, DMA_FROM_DEVICE); hvrc = HvCallEvent_signalLpEventFast(viopath_hostLp, @@ -131,7 +132,7 @@ static int proc_viopath_show(struct seq_file *m, void *v) viopath_sourceinst(viopath_hostLp), viopath_targetinst(viopath_hostLp), (u64)(unsigned long)&Semaphore, VIOVERSION << 16, - ((u64)handle) << 32, PAGE_SIZE, 0, 0); + ((u64)handle) << 32, HW_PAGE_SIZE, 0, 0); if (hvrc != HvLpEvent_Rc_Good) printk(VIOPATH_KERN_WARN "hv error on op %d\n", (int)hvrc); @@ -140,7 +141,7 @@ static int proc_viopath_show(struct seq_file *m, void *v) vlanMap = HvLpConfig_getVirtualLanIndexMap(); - buf[PAGE_SIZE-1] = '\0'; + buf[HW_PAGE_SIZE-1] = '\0'; seq_printf(m, "%s", buf); seq_printf(m, "AVAILABLE_VETH=%x\n", vlanMap); seq_printf(m, "SRLNBR=%c%c%c%c%c%c%c\n", @@ -152,7 +153,8 @@ static int proc_viopath_show(struct seq_file *m, void *v) e2a(xItExtVpdPanel.systemSerial[4]), e2a(xItExtVpdPanel.systemSerial[5])); - dma_unmap_single(iSeries_vio_dev, handle, PAGE_SIZE, DMA_FROM_DEVICE); + dma_unmap_single(iSeries_vio_dev, handle, HW_PAGE_SIZE, + DMA_FROM_DEVICE); kfree(buf); return 0; diff --git a/arch/powerpc/platforms/pseries/lpar.c b/arch/powerpc/platforms/pseries/lpar.c index e384a5a91796..ab0c6dd6ec94 100644 --- a/arch/powerpc/platforms/pseries/lpar.c +++ b/arch/powerpc/platforms/pseries/lpar.c @@ -19,7 +19,7 @@ * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA */ -#define DEBUG +#undef DEBUG_LOW #include #include @@ -41,10 +41,10 @@ #include "plpar_wrappers.h" -#ifdef DEBUG -#define DBG(fmt...) udbg_printf(fmt) +#ifdef DEBUG_LOW +#define DBG_LOW(fmt...) do { udbg_printf(fmt); } while(0) #else -#define DBG(fmt...) +#define DBG_LOW(fmt...) do { } while(0) #endif /* in pSeries_hvCall.S */ @@ -276,8 +276,9 @@ void vpa_init(int cpu) } long pSeries_lpar_hpte_insert(unsigned long hpte_group, - unsigned long va, unsigned long prpn, - unsigned long vflags, unsigned long rflags) + unsigned long va, unsigned long pa, + unsigned long rflags, unsigned long vflags, + int psize) { unsigned long lpar_rc; unsigned long flags; @@ -285,11 +286,28 @@ long pSeries_lpar_hpte_insert(unsigned long hpte_group, unsigned long hpte_v, hpte_r; unsigned long dummy0, dummy1; - hpte_v = ((va >> 23) << HPTE_V_AVPN_SHIFT) | vflags | HPTE_V_VALID; - if (vflags & HPTE_V_LARGE) - hpte_v &= ~(1UL << HPTE_V_AVPN_SHIFT); - - hpte_r = (prpn << HPTE_R_RPN_SHIFT) | rflags; + if (!(vflags & HPTE_V_BOLTED)) + DBG_LOW("hpte_insert(group=%lx, va=%016lx, pa=%016lx, " + "rflags=%lx, vflags=%lx, psize=%d)\n", + hpte_group, va, pa, rflags, vflags, psize); + + hpte_v = hpte_encode_v(va, psize) | vflags | HPTE_V_VALID; + hpte_r = hpte_encode_r(pa, psize) | rflags; + + if (!(vflags & HPTE_V_BOLTED)) + DBG_LOW(" hpte_v=%016lx, hpte_r=%016lx\n", hpte_v, hpte_r); + +#if 1 + { + int i; + for (i=0;i<8;i++) { + unsigned long w0, w1; + plpar_pte_read(0, hpte_group, &w0, &w1); + BUG_ON (HPTE_V_COMPARE(hpte_v, w0) + && (w0 & HPTE_V_VALID)); + } + } +#endif /* Now fill in the actual HPTE */ /* Set CEC cookie to 0 */ @@ -299,23 +317,30 @@ long pSeries_lpar_hpte_insert(unsigned long hpte_group, /* Exact = 0 */ flags = 0; - /* XXX why is this here? - Anton */ + /* Make pHyp happy */ if (rflags & (_PAGE_GUARDED|_PAGE_NO_CACHE)) hpte_r &= ~_PAGE_COHERENT; lpar_rc = plpar_hcall(H_ENTER, flags, hpte_group, hpte_v, hpte_r, &slot, &dummy0, &dummy1); - - if (unlikely(lpar_rc == H_PTEG_Full)) + if (unlikely(lpar_rc == H_PTEG_Full)) { + if (!(vflags & HPTE_V_BOLTED)) + DBG_LOW(" full\n"); return -1; + } /* * Since we try and ioremap PHBs we don't own, the pte insert * will fail. However we must catch the failure in hash_page * or we will loop forever, so return -2 in this case. */ - if (unlikely(lpar_rc != H_Success)) + if (unlikely(lpar_rc != H_Success)) { + if (!(vflags & HPTE_V_BOLTED)) + DBG_LOW(" lpar err %d\n", lpar_rc); return -2; + } + if (!(vflags & HPTE_V_BOLTED)) + DBG_LOW(" -> slot: %d\n", slot & 7); /* Because of iSeries, we have to pass down the secondary * bucket bit here as well @@ -340,10 +365,8 @@ static long pSeries_lpar_hpte_remove(unsigned long hpte_group) /* don't remove a bolted entry */ lpar_rc = plpar_pte_remove(H_ANDCOND, hpte_group + slot_offset, (0x1UL << 4), &dummy1, &dummy2); - if (lpar_rc == H_Success) return i; - BUG_ON(lpar_rc != H_Not_Found); slot_offset++; @@ -371,20 +394,28 @@ static void pSeries_lpar_hptab_clear(void) * We can probably optimize here and assume the high bits of newpp are * already zero. For now I am paranoid. */ -static long pSeries_lpar_hpte_updatepp(unsigned long slot, unsigned long newpp, - unsigned long va, int large, int local) +static long pSeries_lpar_hpte_updatepp(unsigned long slot, + unsigned long newpp, + unsigned long va, + int psize, int local) { unsigned long lpar_rc; unsigned long flags = (newpp & 7) | H_AVPN; - unsigned long avpn = va >> 23; + unsigned long want_v; - if (large) - avpn &= ~0x1UL; + want_v = hpte_encode_v(va, psize); - lpar_rc = plpar_pte_protect(flags, slot, (avpn << 7)); + DBG_LOW(" update: avpnv=%016lx, hash=%016lx, f=%x, psize: %d ... ", + want_v & HPTE_V_AVPN, slot, flags, psize); - if (lpar_rc == H_Not_Found) + lpar_rc = plpar_pte_protect(flags, slot, want_v & HPTE_V_AVPN); + + if (lpar_rc == H_Not_Found) { + DBG_LOW("not found !\n"); return -1; + } + + DBG_LOW("ok\n"); BUG_ON(lpar_rc != H_Success); @@ -410,21 +441,22 @@ static unsigned long pSeries_lpar_hpte_getword0(unsigned long slot) return dword0; } -static long pSeries_lpar_hpte_find(unsigned long vpn) +static long pSeries_lpar_hpte_find(unsigned long va, int psize) { unsigned long hash; unsigned long i, j; long slot; - unsigned long hpte_v; + unsigned long want_v, hpte_v; - hash = hpt_hash(vpn, 0); + hash = hpt_hash(va, mmu_psize_defs[psize].shift); + want_v = hpte_encode_v(va, psize); for (j = 0; j < 2; j++) { slot = (hash & htab_hash_mask) * HPTES_PER_GROUP; for (i = 0; i < HPTES_PER_GROUP; i++) { hpte_v = pSeries_lpar_hpte_getword0(slot); - if ((HPTE_V_AVPN_VAL(hpte_v) == (vpn >> 11)) + if (HPTE_V_COMPARE(hpte_v, want_v) && (hpte_v & HPTE_V_VALID) && (!!(hpte_v & HPTE_V_SECONDARY) == j)) { /* HPTE matches */ @@ -441,17 +473,15 @@ static long pSeries_lpar_hpte_find(unsigned long vpn) } static void pSeries_lpar_hpte_updateboltedpp(unsigned long newpp, - unsigned long ea) + unsigned long ea, + int psize) { - unsigned long lpar_rc; - unsigned long vsid, va, vpn, flags; - long slot; + unsigned long lpar_rc, slot, vsid, va, flags; vsid = get_kernel_vsid(ea); va = (vsid << 28) | (ea & 0x0fffffff); - vpn = va >> PAGE_SHIFT; - slot = pSeries_lpar_hpte_find(vpn); + slot = pSeries_lpar_hpte_find(va, psize); BUG_ON(slot == -1); flags = newpp & 7; @@ -461,18 +491,18 @@ static void pSeries_lpar_hpte_updateboltedpp(unsigned long newpp, } static void pSeries_lpar_hpte_invalidate(unsigned long slot, unsigned long va, - int large, int local) + int psize, int local) { - unsigned long avpn = va >> 23; + unsigned long want_v; unsigned long lpar_rc; unsigned long dummy1, dummy2; - if (large) - avpn &= ~0x1UL; - - lpar_rc = plpar_pte_remove(H_AVPN, slot, (avpn << 7), &dummy1, - &dummy2); + DBG_LOW(" inval : slot=%lx, va=%016lx, psize: %d, local: %d", + slot, va, psize, local); + want_v = hpte_encode_v(va, psize); + lpar_rc = plpar_pte_remove(H_AVPN, slot, want_v & HPTE_V_AVPN, + &dummy1, &dummy2); if (lpar_rc == H_Not_Found) return; @@ -494,7 +524,8 @@ void pSeries_lpar_flush_hash_range(unsigned long number, int local) spin_lock_irqsave(&pSeries_lpar_tlbie_lock, flags); for (i = 0; i < number; i++) - flush_hash_page(batch->vaddr[i], batch->pte[i], local); + flush_hash_page(batch->vaddr[i], batch->pte[i], + batch->psize, local); if (lock_tlbie) spin_unlock_irqrestore(&pSeries_lpar_tlbie_lock, flags); diff --git a/arch/ppc64/Kconfig b/arch/ppc64/Kconfig index b987164fca4c..2130cc315957 100644 --- a/arch/ppc64/Kconfig +++ b/arch/ppc64/Kconfig @@ -47,6 +47,10 @@ config ARCH_MAY_HAVE_PC_FDC bool default y +config PPC_STD_MMU + bool + default y + # We optimistically allocate largepages from the VM, so make the limit # large enough (16MB). This badly named config option is actually # max order + 1 @@ -294,6 +298,15 @@ config NODES_SPAN_OTHER_NODES def_bool y depends on NEED_MULTIPLE_NODES +config PPC_64K_PAGES + bool "64k page size" + help + This option changes the kernel logical page size to 64k. On machines + without processor support for 64k pages, the kernel will simulate + them by loading each individual 4k page on demand transparently, + while on hardware with such support, it will be used to map + normal application pages. + config SCHED_SMT bool "SMT (Hyperthreading) scheduler support" depends on SMP diff --git a/arch/ppc64/kernel/asm-offsets.c b/arch/ppc64/kernel/asm-offsets.c index 504dee836d29..bce9065da6cb 100644 --- a/arch/ppc64/kernel/asm-offsets.c +++ b/arch/ppc64/kernel/asm-offsets.c @@ -93,6 +93,9 @@ int main(void) DEFINE(PACASLBCACHE, offsetof(struct paca_struct, slb_cache)); DEFINE(PACASLBCACHEPTR, offsetof(struct paca_struct, slb_cache_ptr)); DEFINE(PACACONTEXTID, offsetof(struct paca_struct, context.id)); +#ifdef CONFIG_PPC_64K_PAGES + DEFINE(PACAPGDIR, offsetof(struct paca_struct, pgdir)); +#endif #ifdef CONFIG_HUGETLB_PAGE DEFINE(PACALOWHTLBAREAS, offsetof(struct paca_struct, context.low_htlb_areas)); DEFINE(PACAHIGHHTLBAREAS, offsetof(struct paca_struct, context.high_htlb_areas)); diff --git a/arch/ppc64/kernel/head.S b/arch/ppc64/kernel/head.S index db1cf397be2d..9e8050ea1225 100644 --- a/arch/ppc64/kernel/head.S +++ b/arch/ppc64/kernel/head.S @@ -195,11 +195,11 @@ exception_marker: #define EX_R12 24 #define EX_R13 32 #define EX_SRR0 40 -#define EX_R3 40 /* SLB miss saves R3, but not SRR0 */ #define EX_DAR 48 -#define EX_LR 48 /* SLB miss saves LR, but not DAR */ #define EX_DSISR 56 #define EX_CCR 60 +#define EX_R3 64 +#define EX_LR 72 #define EXCEPTION_PROLOG_PSERIES(area, label) \ mfspr r13,SPRN_SPRG3; /* get paca address into r13 */ \ @@ -419,17 +419,22 @@ data_access_slb_pSeries: mtspr SPRN_SPRG1,r13 RUNLATCH_ON(r13) mfspr r13,SPRN_SPRG3 /* get paca address into r13 */ + std r3,PACA_EXSLB+EX_R3(r13) + mfspr r3,SPRN_DAR std r9,PACA_EXSLB+EX_R9(r13) /* save r9 - r12 */ + mfcr r9 +#ifdef __DISABLED__ + /* Keep that around for when we re-implement dynamic VSIDs */ + cmpdi r3,0 + bge slb_miss_user_pseries +#endif /* __DISABLED__ */ std r10,PACA_EXSLB+EX_R10(r13) std r11,PACA_EXSLB+EX_R11(r13) std r12,PACA_EXSLB+EX_R12(r13) - std r3,PACA_EXSLB+EX_R3(r13) - mfspr r9,SPRN_SPRG1 - std r9,PACA_EXSLB+EX_R13(r13) - mfcr r9 + mfspr r10,SPRN_SPRG1 + std r10,PACA_EXSLB+EX_R13(r13) mfspr r12,SPRN_SRR1 /* and SRR1 */ - mfspr r3,SPRN_DAR - b .do_slb_miss /* Rel. branch works in real mode */ + b .slb_miss_realmode /* Rel. branch works in real mode */ STD_EXCEPTION_PSERIES(0x400, instruction_access) @@ -440,17 +445,22 @@ instruction_access_slb_pSeries: mtspr SPRN_SPRG1,r13 RUNLATCH_ON(r13) mfspr r13,SPRN_SPRG3 /* get paca address into r13 */ + std r3,PACA_EXSLB+EX_R3(r13) + mfspr r3,SPRN_SRR0 /* SRR0 is faulting address */ std r9,PACA_EXSLB+EX_R9(r13) /* save r9 - r12 */ + mfcr r9 +#ifdef __DISABLED__ + /* Keep that around for when we re-implement dynamic VSIDs */ + cmpdi r3,0 + bge slb_miss_user_pseries +#endif /* __DISABLED__ */ std r10,PACA_EXSLB+EX_R10(r13) std r11,PACA_EXSLB+EX_R11(r13) std r12,PACA_EXSLB+EX_R12(r13) - std r3,PACA_EXSLB+EX_R3(r13) - mfspr r9,SPRN_SPRG1 - std r9,PACA_EXSLB+EX_R13(r13) - mfcr r9 + mfspr r10,SPRN_SPRG1 + std r10,PACA_EXSLB+EX_R13(r13) mfspr r12,SPRN_SRR1 /* and SRR1 */ - mfspr r3,SPRN_SRR0 /* SRR0 is faulting address */ - b .do_slb_miss /* Rel. branch works in real mode */ + b .slb_miss_realmode /* Rel. branch works in real mode */ STD_EXCEPTION_PSERIES(0x500, hardware_interrupt) STD_EXCEPTION_PSERIES(0x600, alignment) @@ -508,6 +518,38 @@ _GLOBAL(do_stab_bolted_pSeries) mfspr r12,SPRN_SPRG2 EXCEPTION_PROLOG_PSERIES(PACA_EXSLB, .do_stab_bolted) +/* + * We have some room here we use that to put + * the peries slb miss user trampoline code so it's reasonably + * away from slb_miss_user_common to avoid problems with rfid + * + * This is used for when the SLB miss handler has to go virtual, + * which doesn't happen for now anymore but will once we re-implement + * dynamic VSIDs for shared page tables + */ +#ifdef __DISABLED__ +slb_miss_user_pseries: + std r10,PACA_EXGEN+EX_R10(r13) + std r11,PACA_EXGEN+EX_R11(r13) + std r12,PACA_EXGEN+EX_R12(r13) + mfspr r10,SPRG1 + ld r11,PACA_EXSLB+EX_R9(r13) + ld r12,PACA_EXSLB+EX_R3(r13) + std r10,PACA_EXGEN+EX_R13(r13) + std r11,PACA_EXGEN+EX_R9(r13) + std r12,PACA_EXGEN+EX_R3(r13) + clrrdi r12,r13,32 + mfmsr r10 + mfspr r11,SRR0 /* save SRR0 */ + ori r12,r12,slb_miss_user_common@l /* virt addr of handler */ + ori r10,r10,MSR_IR|MSR_DR|MSR_RI + mtspr SRR0,r12 + mfspr r12,SRR1 /* and SRR1 */ + mtspr SRR1,r10 + rfid + b . /* prevent spec. execution */ +#endif /* __DISABLED__ */ + /* * Vectors for the FWNMI option. Share common code. */ @@ -559,22 +601,59 @@ END_FTR_SECTION_IFCLR(CPU_FTR_SLB) .globl data_access_slb_iSeries data_access_slb_iSeries: mtspr SPRN_SPRG1,r13 /* save r13 */ - EXCEPTION_PROLOG_ISERIES_1(PACA_EXSLB) + mfspr r13,SPRN_SPRG3 /* get paca address into r13 */ std r3,PACA_EXSLB+EX_R3(r13) - ld r12,PACALPPACA+LPPACASRR1(r13) mfspr r3,SPRN_DAR - b .do_slb_miss + std r9,PACA_EXSLB+EX_R9(r13) + mfcr r9 +#ifdef __DISABLED__ + cmpdi r3,0 + bge slb_miss_user_iseries +#endif + std r10,PACA_EXSLB+EX_R10(r13) + std r11,PACA_EXSLB+EX_R11(r13) + std r12,PACA_EXSLB+EX_R12(r13) + mfspr r10,SPRN_SPRG1 + std r10,PACA_EXSLB+EX_R13(r13) + ld r12,PACALPPACA+LPPACASRR1(r13); + b .slb_miss_realmode STD_EXCEPTION_ISERIES(0x400, instruction_access, PACA_EXGEN) .globl instruction_access_slb_iSeries instruction_access_slb_iSeries: mtspr SPRN_SPRG1,r13 /* save r13 */ - EXCEPTION_PROLOG_ISERIES_1(PACA_EXSLB) + mfspr r13,SPRN_SPRG3 /* get paca address into r13 */ std r3,PACA_EXSLB+EX_R3(r13) - ld r12,PACALPPACA+LPPACASRR1(r13) - ld r3,PACALPPACA+LPPACASRR0(r13) - b .do_slb_miss + ld r3,PACALPPACA+LPPACASRR0(r13) /* get SRR0 value */ + std r9,PACA_EXSLB+EX_R9(r13) + mfcr r9 +#ifdef __DISABLED__ + cmpdi r3,0 + bge .slb_miss_user_iseries +#endif + std r10,PACA_EXSLB+EX_R10(r13) + std r11,PACA_EXSLB+EX_R11(r13) + std r12,PACA_EXSLB+EX_R12(r13) + mfspr r10,SPRN_SPRG1 + std r10,PACA_EXSLB+EX_R13(r13) + ld r12,PACALPPACA+LPPACASRR1(r13); + b .slb_miss_realmode + +#ifdef __DISABLED__ +slb_miss_user_iseries: + std r10,PACA_EXGEN+EX_R10(r13) + std r11,PACA_EXGEN+EX_R11(r13) + std r12,PACA_EXGEN+EX_R12(r13) + mfspr r10,SPRG1 + ld r11,PACA_EXSLB+EX_R9(r13) + ld r12,PACA_EXSLB+EX_R3(r13) + std r10,PACA_EXGEN+EX_R13(r13) + std r11,PACA_EXGEN+EX_R9(r13) + std r12,PACA_EXGEN+EX_R3(r13) + EXCEPTION_PROLOG_ISERIES_2 + b slb_miss_user_common +#endif MASKABLE_EXCEPTION_ISERIES(0x500, hardware_interrupt) STD_EXCEPTION_ISERIES(0x600, alignment, PACA_EXGEN) @@ -809,6 +888,126 @@ instruction_access_common: li r5,0x400 b .do_hash_page /* Try to handle as hpte fault */ +/* + * Here is the common SLB miss user that is used when going to virtual + * mode for SLB misses, that is currently not used + */ +#ifdef __DISABLED__ + .align 7 + .globl slb_miss_user_common +slb_miss_user_common: + mflr r10 + std r3,PACA_EXGEN+EX_DAR(r13) + stw r9,PACA_EXGEN+EX_CCR(r13) + std r10,PACA_EXGEN+EX_LR(r13) + std r11,PACA_EXGEN+EX_SRR0(r13) + bl .slb_allocate_user + + ld r10,PACA_EXGEN+EX_LR(r13) + ld r3,PACA_EXGEN+EX_R3(r13) + lwz r9,PACA_EXGEN+EX_CCR(r13) + ld r11,PACA_EXGEN+EX_SRR0(r13) + mtlr r10 + beq- slb_miss_fault + + andi. r10,r12,MSR_RI /* check for unrecoverable exception */ + beq- unrecov_user_slb + mfmsr r10 + +.machine push +.machine "power4" + mtcrf 0x80,r9 +.machine pop + + clrrdi r10,r10,2 /* clear RI before setting SRR0/1 */ + mtmsrd r10,1 + + mtspr SRR0,r11 + mtspr SRR1,r12 + + ld r9,PACA_EXGEN+EX_R9(r13) + ld r10,PACA_EXGEN+EX_R10(r13) + ld r11,PACA_EXGEN+EX_R11(r13) + ld r12,PACA_EXGEN+EX_R12(r13) + ld r13,PACA_EXGEN+EX_R13(r13) + rfid + b . + +slb_miss_fault: + EXCEPTION_PROLOG_COMMON(0x380, PACA_EXGEN) + ld r4,PACA_EXGEN+EX_DAR(r13) + li r5,0 + std r4,_DAR(r1) + std r5,_DSISR(r1) + b .handle_page_fault + +unrecov_user_slb: + EXCEPTION_PROLOG_COMMON(0x4200, PACA_EXGEN) + DISABLE_INTS + bl .save_nvgprs +1: addi r3,r1,STACK_FRAME_OVERHEAD + bl .unrecoverable_exception + b 1b + +#endif /* __DISABLED__ */ + + +/* + * r13 points to the PACA, r9 contains the saved CR, + * r12 contain the saved SRR1, SRR0 is still ready for return + * r3 has the faulting address + * r9 - r13 are saved in paca->exslb. + * r3 is saved in paca->slb_r3 + * We assume we aren't going to take any exceptions during this procedure. + */ +_GLOBAL(slb_miss_realmode) + mflr r10 + + stw r9,PACA_EXSLB+EX_CCR(r13) /* save CR in exc. frame */ + std r10,PACA_EXSLB+EX_LR(r13) /* save LR */ + + bl .slb_allocate_realmode + + /* All done -- return from exception. */ + + ld r10,PACA_EXSLB+EX_LR(r13) + ld r3,PACA_EXSLB+EX_R3(r13) + lwz r9,PACA_EXSLB+EX_CCR(r13) /* get saved CR */ +#ifdef CONFIG_PPC_ISERIES + ld r11,PACALPPACA+LPPACASRR0(r13) /* get SRR0 value */ +#endif /* CONFIG_PPC_ISERIES */ + + mtlr r10 + + andi. r10,r12,MSR_RI /* check for unrecoverable exception */ + beq- unrecov_slb + +.machine push +.machine "power4" + mtcrf 0x80,r9 + mtcrf 0x01,r9 /* slb_allocate uses cr0 and cr7 */ +.machine pop + +#ifdef CONFIG_PPC_ISERIES + mtspr SPRN_SRR0,r11 + mtspr SPRN_SRR1,r12 +#endif /* CONFIG_PPC_ISERIES */ + ld r9,PACA_EXSLB+EX_R9(r13) + ld r10,PACA_EXSLB+EX_R10(r13) + ld r11,PACA_EXSLB+EX_R11(r13) + ld r12,PACA_EXSLB+EX_R12(r13) + ld r13,PACA_EXSLB+EX_R13(r13) + rfid + b . /* prevent speculative execution */ + +unrecov_slb: + EXCEPTION_PROLOG_COMMON(0x4100, PACA_EXSLB) + DISABLE_INTS + bl .save_nvgprs +1: addi r3,r1,STACK_FRAME_OVERHEAD + bl .unrecoverable_exception + b 1b + .align 7 .globl hardware_interrupt_common .globl hardware_interrupt_entry @@ -1138,62 +1337,6 @@ _GLOBAL(do_stab_bolted) rfid b . /* prevent speculative execution */ -/* - * r13 points to the PACA, r9 contains the saved CR, - * r11 and r12 contain the saved SRR0 and SRR1. - * r3 has the faulting address - * r9 - r13 are saved in paca->exslb. - * r3 is saved in paca->slb_r3 - * We assume we aren't going to take any exceptions during this procedure. - */ -_GLOBAL(do_slb_miss) - mflr r10 - - stw r9,PACA_EXSLB+EX_CCR(r13) /* save CR in exc. frame */ - std r10,PACA_EXSLB+EX_LR(r13) /* save LR */ - - bl .slb_allocate /* handle it */ - - /* All done -- return from exception. */ - - ld r10,PACA_EXSLB+EX_LR(r13) - ld r3,PACA_EXSLB+EX_R3(r13) - lwz r9,PACA_EXSLB+EX_CCR(r13) /* get saved CR */ -#ifdef CONFIG_PPC_ISERIES - ld r11,PACALPPACA+LPPACASRR0(r13) /* get SRR0 value */ -#endif /* CONFIG_PPC_ISERIES */ - - mtlr r10 - - andi. r10,r12,MSR_RI /* check for unrecoverable exception */ - beq- unrecov_slb - -.machine push -.machine "power4" - mtcrf 0x80,r9 - mtcrf 0x01,r9 /* slb_allocate uses cr0 and cr7 */ -.machine pop - -#ifdef CONFIG_PPC_ISERIES - mtspr SPRN_SRR0,r11 - mtspr SPRN_SRR1,r12 -#endif /* CONFIG_PPC_ISERIES */ - ld r9,PACA_EXSLB+EX_R9(r13) - ld r10,PACA_EXSLB+EX_R10(r13) - ld r11,PACA_EXSLB+EX_R11(r13) - ld r12,PACA_EXSLB+EX_R12(r13) - ld r13,PACA_EXSLB+EX_R13(r13) - rfid - b . /* prevent speculative execution */ - -unrecov_slb: - EXCEPTION_PROLOG_COMMON(0x4100, PACA_EXSLB) - DISABLE_INTS - bl .save_nvgprs -1: addi r3,r1,STACK_FRAME_OVERHEAD - bl .unrecoverable_exception - b 1b - /* * Space for CPU0's segment table. * @@ -1569,7 +1712,10 @@ _GLOBAL(__secondary_start) #endif /* Initialize the first segment table (or SLB) entry */ ld r3,PACASTABVIRT(r13) /* get addr of segment table */ +BEGIN_FTR_SECTION bl .stab_initialize +END_FTR_SECTION_IFCLR(CPU_FTR_SLB) + bl .slb_initialize /* Initialize the kernel stack. Just a repeat for iSeries. */ LOADADDR(r3,current_set) diff --git a/arch/ppc64/kernel/pacaData.c b/arch/ppc64/kernel/pacaData.c index 5e27e5a6a35d..3133c72b28ec 100644 --- a/arch/ppc64/kernel/pacaData.c +++ b/arch/ppc64/kernel/pacaData.c @@ -23,7 +23,7 @@ static union { struct systemcfg data; u8 page[PAGE_SIZE]; -} systemcfg_store __page_aligned; +} systemcfg_store __attribute__((__section__(".data.page.aligned"))); struct systemcfg *systemcfg = &systemcfg_store.data; EXPORT_SYMBOL(systemcfg); diff --git a/arch/ppc64/kernel/prom.c b/arch/ppc64/kernel/prom.c index 97bfceb5353b..dece31e58bc4 100644 --- a/arch/ppc64/kernel/prom.c +++ b/arch/ppc64/kernel/prom.c @@ -635,10 +635,10 @@ static inline char *find_flat_dt_string(u32 offset) * used to extract the memory informations at boot before we can * unflatten the tree */ -static int __init scan_flat_dt(int (*it)(unsigned long node, - const char *uname, int depth, - void *data), - void *data) +int __init of_scan_flat_dt(int (*it)(unsigned long node, + const char *uname, int depth, + void *data), + void *data) { unsigned long p = ((unsigned long)initial_boot_params) + initial_boot_params->off_dt_struct; @@ -695,8 +695,8 @@ static int __init scan_flat_dt(int (*it)(unsigned long node, * This function can be used within scan_flattened_dt callback to get * access to properties */ -static void* __init get_flat_dt_prop(unsigned long node, const char *name, - unsigned long *size) +void* __init of_get_flat_dt_prop(unsigned long node, const char *name, + unsigned long *size) { unsigned long p = node; @@ -996,7 +996,7 @@ void __init unflatten_device_tree(void) static int __init early_init_dt_scan_cpus(unsigned long node, const char *uname, int depth, void *data) { - char *type = get_flat_dt_prop(node, "device_type", NULL); + char *type = of_get_flat_dt_prop(node, "device_type", NULL); u32 *prop; unsigned long size; @@ -1004,17 +1004,6 @@ static int __init early_init_dt_scan_cpus(unsigned long node, if (type == NULL || strcmp(type, "cpu") != 0) return 0; - /* On LPAR, look for the first ibm,pft-size property for the hash table size - */ - if (systemcfg->platform == PLATFORM_PSERIES_LPAR && ppc64_pft_size == 0) { - u32 *pft_size; - pft_size = (u32 *)get_flat_dt_prop(node, "ibm,pft-size", NULL); - if (pft_size != NULL) { - /* pft_size[0] is the NUMA CEC cookie */ - ppc64_pft_size = pft_size[1]; - } - } - if (initial_boot_params && initial_boot_params->version >= 2) { /* version 2 of the kexec param format adds the phys cpuid * of booted proc. @@ -1023,8 +1012,9 @@ static int __init early_init_dt_scan_cpus(unsigned long node, boot_cpuid = 0; } else { /* Check if it's the boot-cpu, set it's hw index in paca now */ - if (get_flat_dt_prop(node, "linux,boot-cpu", NULL) != NULL) { - u32 *prop = get_flat_dt_prop(node, "reg", NULL); + if (of_get_flat_dt_prop(node, "linux,boot-cpu", NULL) + != NULL) { + u32 *prop = of_get_flat_dt_prop(node, "reg", NULL); set_hard_smp_processor_id(0, prop == NULL ? 0 : *prop); boot_cpuid_phys = get_hard_smp_processor_id(0); } @@ -1032,14 +1022,14 @@ static int __init early_init_dt_scan_cpus(unsigned long node, #ifdef CONFIG_ALTIVEC /* Check if we have a VMX and eventually update CPU features */ - prop = (u32 *)get_flat_dt_prop(node, "ibm,vmx", NULL); + prop = (u32 *)of_get_flat_dt_prop(node, "ibm,vmx", NULL); if (prop && (*prop) > 0) { cur_cpu_spec->cpu_features |= CPU_FTR_ALTIVEC; cur_cpu_spec->cpu_user_features |= PPC_FEATURE_HAS_ALTIVEC; } /* Same goes for Apple's "altivec" property */ - prop = (u32 *)get_flat_dt_prop(node, "altivec", NULL); + prop = (u32 *)of_get_flat_dt_prop(node, "altivec", NULL); if (prop) { cur_cpu_spec->cpu_features |= CPU_FTR_ALTIVEC; cur_cpu_spec->cpu_user_features |= PPC_FEATURE_HAS_ALTIVEC; @@ -1051,7 +1041,7 @@ static int __init early_init_dt_scan_cpus(unsigned long node, * this by looking at the size of the ibm,ppc-interrupt-server#s * property */ - prop = (u32 *)get_flat_dt_prop(node, "ibm,ppc-interrupt-server#s", + prop = (u32 *)of_get_flat_dt_prop(node, "ibm,ppc-interrupt-server#s", &size); cur_cpu_spec->cpu_features &= ~CPU_FTR_SMT; if (prop && ((size / sizeof(u32)) > 1)) @@ -1072,26 +1062,26 @@ static int __init early_init_dt_scan_chosen(unsigned long node, return 0; /* get platform type */ - prop = (u32 *)get_flat_dt_prop(node, "linux,platform", NULL); + prop = (u32 *)of_get_flat_dt_prop(node, "linux,platform", NULL); if (prop == NULL) return 0; systemcfg->platform = *prop; /* check if iommu is forced on or off */ - if (get_flat_dt_prop(node, "linux,iommu-off", NULL) != NULL) + if (of_get_flat_dt_prop(node, "linux,iommu-off", NULL) != NULL) iommu_is_off = 1; - if (get_flat_dt_prop(node, "linux,iommu-force-on", NULL) != NULL) + if (of_get_flat_dt_prop(node, "linux,iommu-force-on", NULL) != NULL) iommu_force_on = 1; - prop64 = (u64*)get_flat_dt_prop(node, "linux,memory-limit", NULL); + prop64 = (u64*)of_get_flat_dt_prop(node, "linux,memory-limit", NULL); if (prop64) memory_limit = *prop64; - prop64 = (u64*)get_flat_dt_prop(node, "linux,tce-alloc-start", NULL); + prop64 = (u64*)of_get_flat_dt_prop(node, "linux,tce-alloc-start",NULL); if (prop64) tce_alloc_start = *prop64; - prop64 = (u64*)get_flat_dt_prop(node, "linux,tce-alloc-end", NULL); + prop64 = (u64*)of_get_flat_dt_prop(node, "linux,tce-alloc-end", NULL); if (prop64) tce_alloc_end = *prop64; @@ -1102,9 +1092,12 @@ static int __init early_init_dt_scan_chosen(unsigned long node, { u64 *basep, *entryp; - basep = (u64*)get_flat_dt_prop(node, "linux,rtas-base", NULL); - entryp = (u64*)get_flat_dt_prop(node, "linux,rtas-entry", NULL); - prop = (u32*)get_flat_dt_prop(node, "linux,rtas-size", NULL); + basep = (u64*)of_get_flat_dt_prop(node, + "linux,rtas-base", NULL); + entryp = (u64*)of_get_flat_dt_prop(node, + "linux,rtas-entry", NULL); + prop = (u32*)of_get_flat_dt_prop(node, + "linux,rtas-size", NULL); if (basep && entryp && prop) { rtas.base = *basep; rtas.entry = *entryp; @@ -1125,11 +1118,11 @@ static int __init early_init_dt_scan_root(unsigned long node, if (depth != 0) return 0; - prop = (u32 *)get_flat_dt_prop(node, "#size-cells", NULL); + prop = (u32 *)of_get_flat_dt_prop(node, "#size-cells", NULL); dt_root_size_cells = (prop == NULL) ? 1 : *prop; DBG("dt_root_size_cells = %x\n", dt_root_size_cells); - prop = (u32 *)get_flat_dt_prop(node, "#address-cells", NULL); + prop = (u32 *)of_get_flat_dt_prop(node, "#address-cells", NULL); dt_root_addr_cells = (prop == NULL) ? 2 : *prop; DBG("dt_root_addr_cells = %x\n", dt_root_addr_cells); @@ -1161,7 +1154,7 @@ static unsigned long __init dt_mem_next_cell(int s, cell_t **cellp) static int __init early_init_dt_scan_memory(unsigned long node, const char *uname, int depth, void *data) { - char *type = get_flat_dt_prop(node, "device_type", NULL); + char *type = of_get_flat_dt_prop(node, "device_type", NULL); cell_t *reg, *endp; unsigned long l; @@ -1169,7 +1162,7 @@ static int __init early_init_dt_scan_memory(unsigned long node, if (type == NULL || strcmp(type, "memory") != 0) return 0; - reg = (cell_t *)get_flat_dt_prop(node, "reg", &l); + reg = (cell_t *)of_get_flat_dt_prop(node, "reg", &l); if (reg == NULL) return 0; @@ -1225,19 +1218,16 @@ void __init early_init_devtree(void *params) /* Setup flat device-tree pointer */ initial_boot_params = params; - /* By default, hash size is not set */ - ppc64_pft_size = 0; - /* Retreive various informations from the /chosen node of the * device-tree, including the platform type, initrd location and * size, TCE reserve, and more ... */ - scan_flat_dt(early_init_dt_scan_chosen, NULL); + of_scan_flat_dt(early_init_dt_scan_chosen, NULL); /* Scan memory nodes and rebuild LMBs */ lmb_init(); - scan_flat_dt(early_init_dt_scan_root, NULL); - scan_flat_dt(early_init_dt_scan_memory, NULL); + of_scan_flat_dt(early_init_dt_scan_root, NULL); + of_scan_flat_dt(early_init_dt_scan_memory, NULL); lmb_enforce_memory_limit(memory_limit); lmb_analyze(); systemcfg->physicalMemorySize = lmb_phys_mem_size(); @@ -1253,26 +1243,8 @@ void __init early_init_devtree(void *params) /* Retreive hash table size from flattened tree plus other * CPU related informations (altivec support, boot CPU ID, ...) */ - scan_flat_dt(early_init_dt_scan_cpus, NULL); - - /* If hash size wasn't obtained above, we calculate it now based on - * the total RAM size - */ - if (ppc64_pft_size == 0) { - unsigned long rnd_mem_size, pteg_count; - - /* round mem_size up to next power of 2 */ - rnd_mem_size = 1UL << __ilog2(systemcfg->physicalMemorySize); - if (rnd_mem_size < systemcfg->physicalMemorySize) - rnd_mem_size <<= 1; - - /* # pages / 2 */ - pteg_count = max(rnd_mem_size >> (12 + 1), 1UL << 11); - - ppc64_pft_size = __ilog2(pteg_count << 7); - } + of_scan_flat_dt(early_init_dt_scan_cpus, NULL); - DBG("Hash pftSize: %x\n", (int)ppc64_pft_size); DBG(" <- early_init_devtree()\n"); } diff --git a/include/asm-powerpc/cputable.h b/include/asm-powerpc/cputable.h index c019501daceb..79a0556a0ab8 100644 --- a/include/asm-powerpc/cputable.h +++ b/include/asm-powerpc/cputable.h @@ -101,6 +101,7 @@ extern void do_cpu_ftr_fixups(unsigned long offset); #define CPU_FTR_COHERENT_ICACHE ASM_CONST(0x0000020000000000) #define CPU_FTR_LOCKLESS_TLBIE ASM_CONST(0x0000040000000000) #define CPU_FTR_MMCRA_SIHV ASM_CONST(0x0000080000000000) +#define CPU_FTR_CI_LARGE_PAGE ASM_CONST(0x0000100000000000) #else /* ensure on 32b processors the flags are available for compiling but * don't do anything */ @@ -116,6 +117,7 @@ extern void do_cpu_ftr_fixups(unsigned long offset); #define CPU_FTR_COHERENT_ICACHE ASM_CONST(0x0) #define CPU_FTR_LOCKLESS_TLBIE ASM_CONST(0x0) #define CPU_FTR_MMCRA_SIHV ASM_CONST(0x0) +#define CPU_FTR_CI_LARGE_PAGE ASM_CONST(0x0) #endif #ifndef __ASSEMBLY__ @@ -339,6 +341,7 @@ enum { #ifdef __powerpc64__ CPU_FTRS_POWER3 | CPU_FTRS_RS64 | CPU_FTRS_POWER4 | CPU_FTRS_PPC970 | CPU_FTRS_POWER5 | CPU_FTRS_CELL | + CPU_FTR_CI_LARGE_PAGE | #endif 0, diff --git a/include/asm-powerpc/iommu.h b/include/asm-powerpc/iommu.h index 9d91bdd667ae..6a35e6570ccd 100644 --- a/include/asm-powerpc/iommu.h +++ b/include/asm-powerpc/iommu.h @@ -74,6 +74,11 @@ extern void iommu_devnode_init_pSeries(struct device_node *dn); /* Creates table for an individual device node */ extern void iommu_devnode_init_iSeries(struct device_node *dn); +/* Get table parameters from HV */ +extern void iommu_table_getparms_iSeries(unsigned long busno, + unsigned char slotno, + unsigned char virtbus, + struct iommu_table* tbl); #endif /* CONFIG_PPC_ISERIES */ diff --git a/include/asm-powerpc/machdep.h b/include/asm-powerpc/machdep.h index 629ca964b974..fa03864d06eb 100644 --- a/include/asm-powerpc/machdep.h +++ b/include/asm-powerpc/machdep.h @@ -47,20 +47,22 @@ struct machdep_calls { #ifdef CONFIG_PPC64 void (*hpte_invalidate)(unsigned long slot, unsigned long va, - int large, + int psize, int local); long (*hpte_updatepp)(unsigned long slot, unsigned long newpp, unsigned long va, - int large, + int pize, int local); void (*hpte_updateboltedpp)(unsigned long newpp, - unsigned long ea); + unsigned long ea, + int psize); long (*hpte_insert)(unsigned long hpte_group, unsigned long va, unsigned long prpn, + unsigned long rflags, unsigned long vflags, - unsigned long rflags); + int psize); long (*hpte_remove)(unsigned long hpte_group); void (*flush_hash_range)(unsigned long number, int local); diff --git a/include/asm-powerpc/prom.h b/include/asm-powerpc/prom.h index 3a0104fa0462..7587bf5f38c6 100644 --- a/include/asm-powerpc/prom.h +++ b/include/asm-powerpc/prom.h @@ -178,6 +178,14 @@ extern struct device_node *of_get_next_child(const struct device_node *node, extern struct device_node *of_node_get(struct device_node *node); extern void of_node_put(struct device_node *node); +/* For scanning the flat device-tree at boot time */ +int __init of_scan_flat_dt(int (*it)(unsigned long node, + const char *uname, int depth, + void *data), + void *data); +void* __init of_get_flat_dt_prop(unsigned long node, const char *name, + unsigned long *size); + /* For updating the device tree at runtime */ extern void of_attach_node(struct device_node *); extern void of_detach_node(const struct device_node *); diff --git a/include/asm-powerpc/system.h b/include/asm-powerpc/system.h index b5da0b851e02..3536a5cd7a2d 100644 --- a/include/asm-powerpc/system.h +++ b/include/asm-powerpc/system.h @@ -289,7 +289,7 @@ __cmpxchg_u32(volatile unsigned int *p, unsigned long old, unsigned long new) #ifdef CONFIG_PPC64 static __inline__ unsigned long -__cmpxchg_u64(volatile long *p, unsigned long old, unsigned long new) +__cmpxchg_u64(volatile unsigned long *p, unsigned long old, unsigned long new) { unsigned long prev; diff --git a/include/asm-powerpc/thread_info.h b/include/asm-powerpc/thread_info.h index ab17db79f69d..e525f49bd179 100644 --- a/include/asm-powerpc/thread_info.h +++ b/include/asm-powerpc/thread_info.h @@ -65,23 +65,27 @@ struct thread_info { /* thread information allocation */ -#ifdef CONFIG_DEBUG_STACK_USAGE -#define THREAD_INFO_GFP GFP_KERNEL | __GFP_ZERO -#else -#define THREAD_INFO_GFP GFP_KERNEL -#endif - #if THREAD_SHIFT >= PAGE_SHIFT #define THREAD_ORDER (THREAD_SHIFT - PAGE_SHIFT) +#ifdef CONFIG_DEBUG_STACK_USAGE #define alloc_thread_info(tsk) \ - ((struct thread_info *)__get_free_pages(THREAD_INFO_GFP, THREAD_ORDER)) + ((struct thread_info *)__get_free_pages(GFP_KERNEL | \ + __GFP_ZERO, THREAD_ORDER)) +#else +#define alloc_thread_info(tsk) \ + ((struct thread_info *)__get_free_pages(GFP_KERNEL, THREAD_ORDER)) +#endif #define free_thread_info(ti) free_pages((unsigned long)ti, THREAD_ORDER) #else /* THREAD_SHIFT < PAGE_SHIFT */ -#define alloc_thread_info(tsk) kmalloc(THREAD_SIZE, THREAD_INFO_GFP) +#ifdef CONFIG_DEBUG_STACK_USAGE +#define alloc_thread_info(tsk) kzalloc(THREAD_SIZE, GFP_KERNEL) +#else +#define alloc_thread_info(tsk) kmalloc(THREAD_SIZE, GFP_KERNEL) +#endif #define free_thread_info(ti) kfree(ti) #endif /* THREAD_SHIFT < PAGE_SHIFT */ diff --git a/include/asm-powerpc/tlbflush.h b/include/asm-powerpc/tlbflush.h index ca3655672bbc..a2998eee37bb 100644 --- a/include/asm-powerpc/tlbflush.h +++ b/include/asm-powerpc/tlbflush.h @@ -31,9 +31,9 @@ struct mm_struct; struct ppc64_tlb_batch { unsigned long index; struct mm_struct *mm; - pte_t pte[PPC64_TLB_BATCH_NR]; + real_pte_t pte[PPC64_TLB_BATCH_NR]; unsigned long vaddr[PPC64_TLB_BATCH_NR]; - unsigned int large; + unsigned int psize; }; DECLARE_PER_CPU(struct ppc64_tlb_batch, ppc64_tlb_batch); @@ -48,8 +48,9 @@ static inline void flush_tlb_pending(void) put_cpu_var(ppc64_tlb_batch); } -extern void flush_hash_page(unsigned long va, pte_t pte, int local); -void flush_hash_range(unsigned long number, int local); +extern void flush_hash_page(unsigned long va, real_pte_t pte, int psize, + int local); +extern void flush_hash_range(unsigned long number, int local); #else /* CONFIG_PPC64 */ diff --git a/include/asm-ppc64/mmu.h b/include/asm-ppc64/mmu.h index e0505acb77d9..4c18a5cb69f5 100644 --- a/include/asm-ppc64/mmu.h +++ b/include/asm-ppc64/mmu.h @@ -48,13 +48,21 @@ extern char initial_stab[]; /* Bits in the SLB VSID word */ #define SLB_VSID_SHIFT 12 +#define SLB_VSID_B ASM_CONST(0xc000000000000000) +#define SLB_VSID_B_256M ASM_CONST(0x0000000000000000) +#define SLB_VSID_B_1T ASM_CONST(0x4000000000000000) #define SLB_VSID_KS ASM_CONST(0x0000000000000800) #define SLB_VSID_KP ASM_CONST(0x0000000000000400) #define SLB_VSID_N ASM_CONST(0x0000000000000200) /* no-execute */ -#define SLB_VSID_L ASM_CONST(0x0000000000000100) /* largepage */ +#define SLB_VSID_L ASM_CONST(0x0000000000000100) #define SLB_VSID_C ASM_CONST(0x0000000000000080) /* class */ -#define SLB_VSID_LS ASM_CONST(0x0000000000000070) /* size of largepage */ - +#define SLB_VSID_LP ASM_CONST(0x0000000000000030) +#define SLB_VSID_LP_00 ASM_CONST(0x0000000000000000) +#define SLB_VSID_LP_01 ASM_CONST(0x0000000000000010) +#define SLB_VSID_LP_10 ASM_CONST(0x0000000000000020) +#define SLB_VSID_LP_11 ASM_CONST(0x0000000000000030) +#define SLB_VSID_LLP (SLB_VSID_L|SLB_VSID_LP) + #define SLB_VSID_KERNEL (SLB_VSID_KP) #define SLB_VSID_USER (SLB_VSID_KP|SLB_VSID_KS|SLB_VSID_C) @@ -69,6 +77,7 @@ extern char initial_stab[]; #define HPTE_V_AVPN_SHIFT 7 #define HPTE_V_AVPN ASM_CONST(0xffffffffffffff80) #define HPTE_V_AVPN_VAL(x) (((x) & HPTE_V_AVPN) >> HPTE_V_AVPN_SHIFT) +#define HPTE_V_COMPARE(x,y) (!(((x) ^ (y)) & HPTE_V_AVPN)) #define HPTE_V_BOLTED ASM_CONST(0x0000000000000010) #define HPTE_V_LOCK ASM_CONST(0x0000000000000008) #define HPTE_V_LARGE ASM_CONST(0x0000000000000004) @@ -81,6 +90,7 @@ extern char initial_stab[]; #define HPTE_R_RPN ASM_CONST(0x3ffffffffffff000) #define HPTE_R_FLAGS ASM_CONST(0x00000000000003ff) #define HPTE_R_PP ASM_CONST(0x0000000000000003) +#define HPTE_R_N ASM_CONST(0x0000000000000004) /* Values for PP (assumes Ks=0, Kp=1) */ /* pp0 will always be 0 for linux */ @@ -99,100 +109,120 @@ typedef struct { extern hpte_t *htab_address; extern unsigned long htab_hash_mask; -static inline unsigned long hpt_hash(unsigned long vpn, int large) +/* + * Page size definition + * + * shift : is the "PAGE_SHIFT" value for that page size + * sllp : is a bit mask with the value of SLB L || LP to be or'ed + * directly to a slbmte "vsid" value + * penc : is the HPTE encoding mask for the "LP" field: + * + */ +struct mmu_psize_def { - unsigned long vsid; - unsigned long page; - - if (large) { - vsid = vpn >> 4; - page = vpn & 0xf; - } else { - vsid = vpn >> 16; - page = vpn & 0xffff; - } + unsigned int shift; /* number of bits */ + unsigned int penc; /* HPTE encoding */ + unsigned int tlbiel; /* tlbiel supported for that page size */ + unsigned long avpnm; /* bits to mask out in AVPN in the HPTE */ + unsigned long sllp; /* SLB L||LP (exact mask to use in slbmte) */ +}; - return (vsid & 0x7fffffffffUL) ^ page; -} - -static inline void __tlbie(unsigned long va, int large) -{ - /* clear top 16 bits, non SLS segment */ - va &= ~(0xffffULL << 48); - - if (large) { - va &= HPAGE_MASK; - asm volatile("tlbie %0,1" : : "r"(va) : "memory"); - } else { - va &= PAGE_MASK; - asm volatile("tlbie %0,0" : : "r"(va) : "memory"); - } -} +#endif /* __ASSEMBLY__ */ -static inline void tlbie(unsigned long va, int large) -{ - asm volatile("ptesync": : :"memory"); - __tlbie(va, large); - asm volatile("eieio; tlbsync; ptesync": : :"memory"); -} +/* + * The kernel use the constants below to index in the page sizes array. + * The use of fixed constants for this purpose is better for performances + * of the low level hash refill handlers. + * + * A non supported page size has a "shift" field set to 0 + * + * Any new page size being implemented can get a new entry in here. Whether + * the kernel will use it or not is a different matter though. The actual page + * size used by hugetlbfs is not defined here and may be made variable + */ -static inline void __tlbiel(unsigned long va) -{ - /* clear top 16 bits, non SLS segment */ - va &= ~(0xffffULL << 48); - va &= PAGE_MASK; - - /* - * Thanks to Alan Modra we are now able to use machine specific - * assembly instructions (like tlbiel) by using the gas -many flag. - * However we have to support older toolchains so for the moment - * we hardwire it. - */ -#if 0 - asm volatile("tlbiel %0" : : "r"(va) : "memory"); -#else - asm volatile(".long 0x7c000224 | (%0 << 11)" : : "r"(va) : "memory"); -#endif -} +#define MMU_PAGE_4K 0 /* 4K */ +#define MMU_PAGE_64K 1 /* 64K */ +#define MMU_PAGE_64K_AP 2 /* 64K Admixed (in a 4K segment) */ +#define MMU_PAGE_1M 3 /* 1M */ +#define MMU_PAGE_16M 4 /* 16M */ +#define MMU_PAGE_16G 5 /* 16G */ +#define MMU_PAGE_COUNT 6 -static inline void tlbiel(unsigned long va) -{ - asm volatile("ptesync": : :"memory"); - __tlbiel(va); - asm volatile("ptesync": : :"memory"); -} +#ifndef __ASSEMBLY__ -static inline unsigned long slot2va(unsigned long hpte_v, unsigned long slot) -{ - unsigned long avpn = HPTE_V_AVPN_VAL(hpte_v); - unsigned long va; +/* + * The current system page sizes + */ +extern struct mmu_psize_def mmu_psize_defs[MMU_PAGE_COUNT]; +extern int mmu_linear_psize; +extern int mmu_virtual_psize; - va = avpn << 23; +#ifdef CONFIG_HUGETLB_PAGE +/* + * The page size index of the huge pages for use by hugetlbfs + */ +extern int mmu_huge_psize; - if (! (hpte_v & HPTE_V_LARGE)) { - unsigned long vpi, pteg; +#endif /* CONFIG_HUGETLB_PAGE */ - pteg = slot / HPTES_PER_GROUP; - if (hpte_v & HPTE_V_SECONDARY) - pteg = ~pteg; +/* + * This function sets the AVPN and L fields of the HPTE appropriately + * for the page size + */ +static inline unsigned long hpte_encode_v(unsigned long va, int psize) +{ + unsigned long v = + v = (va >> 23) & ~(mmu_psize_defs[psize].avpnm); + v <<= HPTE_V_AVPN_SHIFT; + if (psize != MMU_PAGE_4K) + v |= HPTE_V_LARGE; + return v; +} - vpi = ((va >> 28) ^ pteg) & htab_hash_mask; +/* + * This function sets the ARPN, and LP fields of the HPTE appropriately + * for the page size. We assume the pa is already "clean" that is properly + * aligned for the requested page size + */ +static inline unsigned long hpte_encode_r(unsigned long pa, int psize) +{ + unsigned long r; - va |= vpi << PAGE_SHIFT; + /* A 4K page needs no special encoding */ + if (psize == MMU_PAGE_4K) + return pa & HPTE_R_RPN; + else { + unsigned int penc = mmu_psize_defs[psize].penc; + unsigned int shift = mmu_psize_defs[psize].shift; + return (pa & ~((1ul << shift) - 1)) | (penc << 12); } - - return va; + return r; } /* - * Handle a fault by adding an HPTE. If the address can't be determined - * to be valid via Linux page tables, return 1. If handled return 0 + * This hashes a virtual address for a 256Mb segment only for now */ -extern int __hash_page(unsigned long ea, unsigned long access, - unsigned long vsid, pte_t *ptep, unsigned long trap, - int local); + +static inline unsigned long hpt_hash(unsigned long va, unsigned int shift) +{ + return ((va >> 28) & 0x7fffffffffUL) ^ ((va & 0x0fffffffUL) >> shift); +} + +extern int __hash_page_4K(unsigned long ea, unsigned long access, + unsigned long vsid, pte_t *ptep, unsigned long trap, + unsigned int local); +extern int __hash_page_64K(unsigned long ea, unsigned long access, + unsigned long vsid, pte_t *ptep, unsigned long trap, + unsigned int local); +struct mm_struct; +extern int hash_huge_page(struct mm_struct *mm, unsigned long access, + unsigned long ea, unsigned long vsid, int local); extern void htab_finish_init(void); +extern int htab_bolt_mapping(unsigned long vstart, unsigned long vend, + unsigned long pstart, unsigned long mode, + int psize); extern void hpte_init_native(void); extern void hpte_init_lpar(void); @@ -200,17 +230,21 @@ extern void hpte_init_iSeries(void); extern long pSeries_lpar_hpte_insert(unsigned long hpte_group, unsigned long va, unsigned long prpn, - unsigned long vflags, - unsigned long rflags); -extern long native_hpte_insert(unsigned long hpte_group, unsigned long va, - unsigned long prpn, - unsigned long vflags, unsigned long rflags); + unsigned long rflags, + unsigned long vflags, int psize); + +extern long native_hpte_insert(unsigned long hpte_group, + unsigned long va, unsigned long prpn, + unsigned long rflags, + unsigned long vflags, int psize); -extern long iSeries_hpte_bolt_or_insert(unsigned long hpte_group, - unsigned long va, unsigned long prpn, - unsigned long vflags, unsigned long rflags); +extern long iSeries_hpte_insert(unsigned long hpte_group, + unsigned long va, unsigned long prpn, + unsigned long rflags, + unsigned long vflags, int psize); extern void stabs_alloc(void); +extern void slb_initialize(void); #endif /* __ASSEMBLY__ */ diff --git a/include/asm-ppc64/mmu_context.h b/include/asm-ppc64/mmu_context.h index 820dd729b895..4f512e9fa6b8 100644 --- a/include/asm-ppc64/mmu_context.h +++ b/include/asm-ppc64/mmu_context.h @@ -16,8 +16,16 @@ * 2 of the License, or (at your option) any later version. */ -static inline void enter_lazy_tlb(struct mm_struct *mm, struct task_struct *tsk) +/* + * Getting into a kernel thread, there is no valid user segment, mark + * paca->pgdir NULL so that SLB miss on user addresses will fault + */ +static inline void enter_lazy_tlb(struct mm_struct *mm, + struct task_struct *tsk) { +#ifdef CONFIG_PPC_64K_PAGES + get_paca()->pgdir = NULL; +#endif /* CONFIG_PPC_64K_PAGES */ } #define NO_CONTEXT 0 @@ -40,8 +48,13 @@ static inline void switch_mm(struct mm_struct *prev, struct mm_struct *next, cpu_set(smp_processor_id(), next->cpu_vm_mask); /* No need to flush userspace segments if the mm doesnt change */ +#ifdef CONFIG_PPC_64K_PAGES + if (prev == next && get_paca()->pgdir == next->pgd) + return; +#else if (prev == next) return; +#endif /* CONFIG_PPC_64K_PAGES */ #ifdef CONFIG_ALTIVEC if (cpu_has_feature(CPU_FTR_ALTIVEC)) diff --git a/include/asm-ppc64/paca.h b/include/asm-ppc64/paca.h index f68fe91debaf..bccacd6aa93a 100644 --- a/include/asm-ppc64/paca.h +++ b/include/asm-ppc64/paca.h @@ -72,10 +72,15 @@ struct paca_struct { /* * Now, starting in cacheline 2, the exception save areas */ - u64 exgen[8] __attribute__((aligned(0x80))); /* used for most interrupts/exceptions */ - u64 exmc[8]; /* used for machine checks */ - u64 exslb[8]; /* used for SLB/segment table misses - * on the linear mapping */ + /* used for most interrupts/exceptions */ + u64 exgen[10] __attribute__((aligned(0x80))); + u64 exmc[10]; /* used for machine checks */ + u64 exslb[10]; /* used for SLB/segment table misses + * on the linear mapping */ +#ifdef CONFIG_PPC_64K_PAGES + pgd_t *pgdir; +#endif /* CONFIG_PPC_64K_PAGES */ + mm_context_t context; u16 slb_cache[SLB_CACHE_ENTRIES]; u16 slb_cache_ptr; diff --git a/include/asm-ppc64/page.h b/include/asm-ppc64/page.h index d404431f0a9a..82ce187e5be8 100644 --- a/include/asm-ppc64/page.h +++ b/include/asm-ppc64/page.h @@ -13,32 +13,59 @@ #include #include /* for ASM_CONST */ -/* PAGE_SHIFT determines the page size */ -#define PAGE_SHIFT 12 -#define PAGE_SIZE (ASM_CONST(1) << PAGE_SHIFT) -#define PAGE_MASK (~(PAGE_SIZE-1)) +/* + * We support either 4k or 64k software page size. When using 64k pages + * however, wether we are really supporting 64k pages in HW or not is + * irrelevant to those definitions. We always define HW_PAGE_SHIFT to 12 + * as use of 64k pages remains a linux kernel specific, every notion of + * page number shared with the firmware, TCEs, iommu, etc... still assumes + * a page size of 4096. + */ +#ifdef CONFIG_PPC_64K_PAGES +#define PAGE_SHIFT 16 +#else +#define PAGE_SHIFT 12 +#endif -#define SID_SHIFT 28 -#define SID_MASK 0xfffffffffUL -#define ESID_MASK 0xfffffffff0000000UL -#define GET_ESID(x) (((x) >> SID_SHIFT) & SID_MASK) +#define PAGE_SIZE (ASM_CONST(1) << PAGE_SHIFT) +#define PAGE_MASK (~(PAGE_SIZE-1)) -#define HPAGE_SHIFT 24 -#define HPAGE_SIZE ((1UL) << HPAGE_SHIFT) -#define HPAGE_MASK (~(HPAGE_SIZE - 1)) +/* HW_PAGE_SHIFT is always 4k pages */ +#define HW_PAGE_SHIFT 12 +#define HW_PAGE_SIZE (ASM_CONST(1) << HW_PAGE_SHIFT) +#define HW_PAGE_MASK (~(HW_PAGE_SIZE-1)) -#ifdef CONFIG_HUGETLB_PAGE +/* PAGE_FACTOR is the number of bits factor between PAGE_SHIFT and + * HW_PAGE_SHIFT, that is 4k pages + */ +#define PAGE_FACTOR (PAGE_SHIFT - HW_PAGE_SHIFT) + +/* Segment size */ +#define SID_SHIFT 28 +#define SID_MASK 0xfffffffffUL +#define ESID_MASK 0xfffffffff0000000UL +#define GET_ESID(x) (((x) >> SID_SHIFT) & SID_MASK) +/* Large pages size */ + +#ifndef __ASSEMBLY__ +extern unsigned int HPAGE_SHIFT; +#define HPAGE_SIZE ((1UL) << HPAGE_SHIFT) +#define HPAGE_MASK (~(HPAGE_SIZE - 1)) #define HUGETLB_PAGE_ORDER (HPAGE_SHIFT - PAGE_SHIFT) +#endif /* __ASSEMBLY__ */ + +#ifdef CONFIG_HUGETLB_PAGE + #define HTLB_AREA_SHIFT 40 #define HTLB_AREA_SIZE (1UL << HTLB_AREA_SHIFT) #define GET_HTLB_AREA(x) ((x) >> HTLB_AREA_SHIFT) -#define LOW_ESID_MASK(addr, len) (((1U << (GET_ESID(addr+len-1)+1)) \ - - (1U << GET_ESID(addr))) & 0xffff) -#define HTLB_AREA_MASK(addr, len) (((1U << (GET_HTLB_AREA(addr+len-1)+1)) \ - - (1U << GET_HTLB_AREA(addr))) & 0xffff) +#define LOW_ESID_MASK(addr, len) (((1U << (GET_ESID(addr+len-1)+1)) \ + - (1U << GET_ESID(addr))) & 0xffff) +#define HTLB_AREA_MASK(addr, len) (((1U << (GET_HTLB_AREA(addr+len-1)+1)) \ + - (1U << GET_HTLB_AREA(addr))) & 0xffff) #define ARCH_HAS_HUGEPAGE_ONLY_RANGE #define ARCH_HAS_PREPARE_HUGEPAGE_RANGE @@ -114,7 +141,25 @@ static __inline__ void clear_page(void *addr) : "ctr", "memory"); } -extern void copy_page(void *to, void *from); +extern void copy_4K_page(void *to, void *from); + +#ifdef CONFIG_PPC_64K_PAGES +static inline void copy_page(void *to, void *from) +{ + unsigned int i; + for (i=0; i < (1 << (PAGE_SHIFT - 12)); i++) { + copy_4K_page(to, from); + to += 4096; + from += 4096; + } +} +#else /* CONFIG_PPC_64K_PAGES */ +static inline void copy_page(void *to, void *from) +{ + copy_4K_page(to, from); +} +#endif /* CONFIG_PPC_64K_PAGES */ + struct page; extern void clear_user_page(void *page, unsigned long vaddr, struct page *pg); extern void copy_user_page(void *to, void *from, unsigned long vaddr, struct page *p); @@ -124,43 +169,75 @@ extern void copy_user_page(void *to, void *from, unsigned long vaddr, struct pag * These are used to make use of C type-checking. * Entries in the pte table are 64b, while entries in the pgd & pmd are 32b. */ -typedef struct { unsigned long pte; } pte_t; -typedef struct { unsigned long pmd; } pmd_t; -typedef struct { unsigned long pud; } pud_t; -typedef struct { unsigned long pgd; } pgd_t; -typedef struct { unsigned long pgprot; } pgprot_t; +/* PTE level */ +typedef struct { unsigned long pte; } pte_t; #define pte_val(x) ((x).pte) -#define pmd_val(x) ((x).pmd) -#define pud_val(x) ((x).pud) -#define pgd_val(x) ((x).pgd) -#define pgprot_val(x) ((x).pgprot) - #define __pte(x) ((pte_t) { (x) }) + +/* 64k pages additionally define a bigger "real PTE" type that gathers + * the "second half" part of the PTE for pseudo 64k pages + */ +#ifdef CONFIG_PPC_64K_PAGES +typedef struct { pte_t pte; unsigned long hidx; } real_pte_t; +#else +typedef struct { pte_t pte; } real_pte_t; +#endif + +/* PMD level */ +typedef struct { unsigned long pmd; } pmd_t; +#define pmd_val(x) ((x).pmd) #define __pmd(x) ((pmd_t) { (x) }) + +/* PUD level exusts only on 4k pages */ +#ifndef CONFIG_PPC_64K_PAGES +typedef struct { unsigned long pud; } pud_t; +#define pud_val(x) ((x).pud) #define __pud(x) ((pud_t) { (x) }) +#endif + +/* PGD level */ +typedef struct { unsigned long pgd; } pgd_t; +#define pgd_val(x) ((x).pgd) #define __pgd(x) ((pgd_t) { (x) }) + +/* Page protection bits */ +typedef struct { unsigned long pgprot; } pgprot_t; +#define pgprot_val(x) ((x).pgprot) #define __pgprot(x) ((pgprot_t) { (x) }) #else + /* * .. while these make it easier on the compiler */ -typedef unsigned long pte_t; -typedef unsigned long pmd_t; -typedef unsigned long pud_t; -typedef unsigned long pgd_t; -typedef unsigned long pgprot_t; +typedef unsigned long pte_t; #define pte_val(x) (x) +#define __pte(x) (x) + +#ifdef CONFIG_PPC_64K_PAGES +typedef struct { pte_t pte; unsigned long hidx; } real_pte_t; +#else +typedef unsigned long real_pte_t; +#endif + + +typedef unsigned long pmd_t; #define pmd_val(x) (x) +#define __pmd(x) (x) + +#ifndef CONFIG_PPC_64K_PAGES +typedef unsigned long pud_t; #define pud_val(x) (x) +#define __pud(x) (x) +#endif + +typedef unsigned long pgd_t; #define pgd_val(x) (x) #define pgprot_val(x) (x) -#define __pte(x) (x) -#define __pmd(x) (x) -#define __pud(x) (x) +typedef unsigned long pgprot_t; #define __pgd(x) (x) #define __pgprot(x) (x) diff --git a/include/asm-ppc64/pgalloc.h b/include/asm-ppc64/pgalloc.h index 26bc49c1108d..98da0e4262bd 100644 --- a/include/asm-ppc64/pgalloc.h +++ b/include/asm-ppc64/pgalloc.h @@ -8,10 +8,16 @@ extern kmem_cache_t *pgtable_cache[]; +#ifdef CONFIG_PPC_64K_PAGES +#define PTE_CACHE_NUM 0 +#define PMD_CACHE_NUM 0 +#define PGD_CACHE_NUM 1 +#else #define PTE_CACHE_NUM 0 #define PMD_CACHE_NUM 1 #define PUD_CACHE_NUM 1 #define PGD_CACHE_NUM 0 +#endif /* * This program is free software; you can redistribute it and/or @@ -30,6 +36,8 @@ static inline void pgd_free(pgd_t *pgd) kmem_cache_free(pgtable_cache[PGD_CACHE_NUM], pgd); } +#ifndef CONFIG_PPC_64K_PAGES + #define pgd_populate(MM, PGD, PUD) pgd_set(PGD, PUD) static inline pud_t *pud_alloc_one(struct mm_struct *mm, unsigned long addr) @@ -43,7 +51,30 @@ static inline void pud_free(pud_t *pud) kmem_cache_free(pgtable_cache[PUD_CACHE_NUM], pud); } -#define pud_populate(MM, PUD, PMD) pud_set(PUD, PMD) +static inline void pud_populate(struct mm_struct *mm, pud_t *pud, pmd_t *pmd) +{ + pud_set(pud, (unsigned long)pmd); +} + +#define pmd_populate(mm, pmd, pte_page) \ + pmd_populate_kernel(mm, pmd, page_address(pte_page)) +#define pmd_populate_kernel(mm, pmd, pte) pmd_set(pmd, (unsigned long)(pte)) + + +#else /* CONFIG_PPC_64K_PAGES */ + +#define pud_populate(mm, pud, pmd) pud_set(pud, (unsigned long)pmd) + +static inline void pmd_populate_kernel(struct mm_struct *mm, pmd_t *pmd, + pte_t *pte) +{ + pmd_set(pmd, (unsigned long)pte); +} + +#define pmd_populate(mm, pmd, pte_page) \ + pmd_populate_kernel(mm, pmd, page_address(pte_page)) + +#endif /* CONFIG_PPC_64K_PAGES */ static inline pmd_t *pmd_alloc_one(struct mm_struct *mm, unsigned long addr) { @@ -56,17 +87,15 @@ static inline void pmd_free(pmd_t *pmd) kmem_cache_free(pgtable_cache[PMD_CACHE_NUM], pmd); } -#define pmd_populate_kernel(mm, pmd, pte) pmd_set(pmd, pte) -#define pmd_populate(mm, pmd, pte_page) \ - pmd_populate_kernel(mm, pmd, page_address(pte_page)) - -static inline pte_t *pte_alloc_one_kernel(struct mm_struct *mm, unsigned long address) +static inline pte_t *pte_alloc_one_kernel(struct mm_struct *mm, + unsigned long address) { return kmem_cache_alloc(pgtable_cache[PTE_CACHE_NUM], GFP_KERNEL|__GFP_REPEAT); } -static inline struct page *pte_alloc_one(struct mm_struct *mm, unsigned long address) +static inline struct page *pte_alloc_one(struct mm_struct *mm, + unsigned long address) { return virt_to_page(pte_alloc_one_kernel(mm, address)); } @@ -103,7 +132,7 @@ static inline void pgtable_free(pgtable_free_t pgf) kmem_cache_free(pgtable_cache[cachenum], p); } -void pgtable_free_tlb(struct mmu_gather *tlb, pgtable_free_t pgf); +extern void pgtable_free_tlb(struct mmu_gather *tlb, pgtable_free_t pgf); #define __pte_free_tlb(tlb, ptepage) \ pgtable_free_tlb(tlb, pgtable_free_cache(page_address(ptepage), \ @@ -111,9 +140,11 @@ void pgtable_free_tlb(struct mmu_gather *tlb, pgtable_free_t pgf); #define __pmd_free_tlb(tlb, pmd) \ pgtable_free_tlb(tlb, pgtable_free_cache(pmd, \ PMD_CACHE_NUM, PMD_TABLE_SIZE-1)) +#ifndef CONFIG_PPC_64K_PAGES #define __pud_free_tlb(tlb, pmd) \ pgtable_free_tlb(tlb, pgtable_free_cache(pud, \ PUD_CACHE_NUM, PUD_TABLE_SIZE-1)) +#endif /* CONFIG_PPC_64K_PAGES */ #define check_pgt_cache() do { } while (0) diff --git a/include/asm-ppc64/pgtable-4k.h b/include/asm-ppc64/pgtable-4k.h new file mode 100644 index 000000000000..c883a2748558 --- /dev/null +++ b/include/asm-ppc64/pgtable-4k.h @@ -0,0 +1,88 @@ +/* + * Entries per page directory level. The PTE level must use a 64b record + * for each page table entry. The PMD and PGD level use a 32b record for + * each entry by assuming that each entry is page aligned. + */ +#define PTE_INDEX_SIZE 9 +#define PMD_INDEX_SIZE 7 +#define PUD_INDEX_SIZE 7 +#define PGD_INDEX_SIZE 9 + +#define PTE_TABLE_SIZE (sizeof(pte_t) << PTE_INDEX_SIZE) +#define PMD_TABLE_SIZE (sizeof(pmd_t) << PMD_INDEX_SIZE) +#define PUD_TABLE_SIZE (sizeof(pud_t) << PUD_INDEX_SIZE) +#define PGD_TABLE_SIZE (sizeof(pgd_t) << PGD_INDEX_SIZE) + +#define PTRS_PER_PTE (1 << PTE_INDEX_SIZE) +#define PTRS_PER_PMD (1 << PMD_INDEX_SIZE) +#define PTRS_PER_PUD (1 << PMD_INDEX_SIZE) +#define PTRS_PER_PGD (1 << PGD_INDEX_SIZE) + +/* PMD_SHIFT determines what a second-level page table entry can map */ +#define PMD_SHIFT (PAGE_SHIFT + PTE_INDEX_SIZE) +#define PMD_SIZE (1UL << PMD_SHIFT) +#define PMD_MASK (~(PMD_SIZE-1)) + +/* PUD_SHIFT determines what a third-level page table entry can map */ +#define PUD_SHIFT (PMD_SHIFT + PMD_INDEX_SIZE) +#define PUD_SIZE (1UL << PUD_SHIFT) +#define PUD_MASK (~(PUD_SIZE-1)) + +/* PGDIR_SHIFT determines what a fourth-level page table entry can map */ +#define PGDIR_SHIFT (PUD_SHIFT + PUD_INDEX_SIZE) +#define PGDIR_SIZE (1UL << PGDIR_SHIFT) +#define PGDIR_MASK (~(PGDIR_SIZE-1)) + +/* PTE bits */ +#define _PAGE_SECONDARY 0x8000 /* software: HPTE is in secondary group */ +#define _PAGE_GROUP_IX 0x7000 /* software: HPTE index within group */ +#define _PAGE_F_SECOND _PAGE_SECONDARY +#define _PAGE_F_GIX _PAGE_GROUP_IX + +/* PTE flags to conserve for HPTE identification */ +#define _PAGE_HPTEFLAGS (_PAGE_BUSY | _PAGE_HASHPTE | \ + _PAGE_SECONDARY | _PAGE_GROUP_IX) + +/* PAGE_MASK gives the right answer below, but only by accident */ +/* It should be preserving the high 48 bits and then specifically */ +/* preserving _PAGE_SECONDARY | _PAGE_GROUP_IX */ +#define _PAGE_CHG_MASK (PAGE_MASK | _PAGE_ACCESSED | _PAGE_DIRTY | \ + _PAGE_HPTEFLAGS) + +/* Bits to mask out from a PMD to get to the PTE page */ +#define PMD_MASKED_BITS 0 +/* Bits to mask out from a PUD to get to the PMD page */ +#define PUD_MASKED_BITS 0 +/* Bits to mask out from a PGD to get to the PUD page */ +#define PGD_MASKED_BITS 0 + +/* shift to put page number into pte */ +#define PTE_RPN_SHIFT (17) + +#define __real_pte(e,p) ((real_pte_t)(e)) +#define __rpte_to_pte(r) (r) +#define __rpte_to_hidx(r,index) (pte_val((r)) >> 12) + +#define pte_iterate_hashed_subpages(rpte, psize, va, index, shift) \ + do { \ + index = 0; \ + shift = mmu_psize_defs[psize].shift; \ + +#define pte_iterate_hashed_end() } while(0) + +/* + * 4-level page tables related bits + */ + +#define pgd_none(pgd) (!pgd_val(pgd)) +#define pgd_bad(pgd) (pgd_val(pgd) == 0) +#define pgd_present(pgd) (pgd_val(pgd) != 0) +#define pgd_clear(pgdp) (pgd_val(*(pgdp)) = 0) +#define pgd_page(pgd) (pgd_val(pgd) & ~PGD_MASKED_BITS) + +#define pud_offset(pgdp, addr) \ + (((pud_t *) pgd_page(*(pgdp))) + \ + (((addr) >> PUD_SHIFT) & (PTRS_PER_PUD - 1))) + +#define pud_ERROR(e) \ + printk("%s:%d: bad pmd %08lx.\n", __FILE__, __LINE__, pud_val(e)) diff --git a/include/asm-ppc64/pgtable-64k.h b/include/asm-ppc64/pgtable-64k.h new file mode 100644 index 000000000000..c5f437c86b3c --- /dev/null +++ b/include/asm-ppc64/pgtable-64k.h @@ -0,0 +1,87 @@ +#include + + +#define PTE_INDEX_SIZE 12 +#define PMD_INDEX_SIZE 12 +#define PUD_INDEX_SIZE 0 +#define PGD_INDEX_SIZE 4 + +#define PTE_TABLE_SIZE (sizeof(real_pte_t) << PTE_INDEX_SIZE) +#define PMD_TABLE_SIZE (sizeof(pmd_t) << PMD_INDEX_SIZE) +#define PGD_TABLE_SIZE (sizeof(pgd_t) << PGD_INDEX_SIZE) + +#define PTRS_PER_PTE (1 << PTE_INDEX_SIZE) +#define PTRS_PER_PMD (1 << PMD_INDEX_SIZE) +#define PTRS_PER_PGD (1 << PGD_INDEX_SIZE) + +/* PMD_SHIFT determines what a second-level page table entry can map */ +#define PMD_SHIFT (PAGE_SHIFT + PTE_INDEX_SIZE) +#define PMD_SIZE (1UL << PMD_SHIFT) +#define PMD_MASK (~(PMD_SIZE-1)) + +/* PGDIR_SHIFT determines what a third-level page table entry can map */ +#define PGDIR_SHIFT (PMD_SHIFT + PMD_INDEX_SIZE) +#define PGDIR_SIZE (1UL << PGDIR_SHIFT) +#define PGDIR_MASK (~(PGDIR_SIZE-1)) + +/* Additional PTE bits (don't change without checking asm in hash_low.S) */ +#define _PAGE_HPTE_SUB 0x0ffff000 /* combo only: sub pages HPTE bits */ +#define _PAGE_HPTE_SUB0 0x08000000 /* combo only: first sub page */ +#define _PAGE_COMBO 0x10000000 /* this is a combo 4k page */ +#define _PAGE_F_SECOND 0x00008000 /* full page: hidx bits */ +#define _PAGE_F_GIX 0x00007000 /* full page: hidx bits */ + +/* PTE flags to conserve for HPTE identification */ +#define _PAGE_HPTEFLAGS (_PAGE_BUSY | _PAGE_HASHPTE | _PAGE_HPTE_SUB |\ + _PAGE_COMBO) + +/* Shift to put page number into pte. + * + * That gives us a max RPN of 32 bits, which means a max of 48 bits + * of addressable physical space. + * We could get 3 more bits here by setting PTE_RPN_SHIFT to 29 but + * 32 makes PTEs more readable for debugging for now :) + */ +#define PTE_RPN_SHIFT (32) +#define PTE_RPN_MAX (1UL << (64 - PTE_RPN_SHIFT)) +#define PTE_RPN_MASK (~((1UL<> ((index)<<2)) & 0xf) : ((pte_val((r).pte) >> 12) & 0xf)) +#define __rpte_to_pte(r) ((r).pte) +#define __rpte_sub_valid(rpte, index) \ + (pte_val(rpte.pte) & (_PAGE_HPTE_SUB0 >> (index))) + + +/* Trick: we set __end to va + 64k, which happens works for + * a 16M page as well as we want only one iteration + */ +#define pte_iterate_hashed_subpages(rpte, psize, va, index, shift) \ + do { \ + unsigned long __end = va + PAGE_SIZE; \ + unsigned __split = (psize == MMU_PAGE_4K || \ + psize == MMU_PAGE_64K_AP); \ + shift = mmu_psize_defs[psize].shift; \ + for (index = 0; va < __end; index++, va += (1 << shift)) { \ + if (!__split || __rpte_sub_valid(rpte, index)) do { \ + +#define pte_iterate_hashed_end() } while(0); } } while(0) + + +#endif /* __ASSEMBLY__ */ diff --git a/include/asm-ppc64/pgtable.h b/include/asm-ppc64/pgtable.h index 8c3f574046b6..fde93ec36abc 100644 --- a/include/asm-ppc64/pgtable.h +++ b/include/asm-ppc64/pgtable.h @@ -15,40 +15,11 @@ #include #endif /* __ASSEMBLY__ */ -/* - * Entries per page directory level. The PTE level must use a 64b record - * for each page table entry. The PMD and PGD level use a 32b record for - * each entry by assuming that each entry is page aligned. - */ -#define PTE_INDEX_SIZE 9 -#define PMD_INDEX_SIZE 7 -#define PUD_INDEX_SIZE 7 -#define PGD_INDEX_SIZE 9 - -#define PTE_TABLE_SIZE (sizeof(pte_t) << PTE_INDEX_SIZE) -#define PMD_TABLE_SIZE (sizeof(pmd_t) << PMD_INDEX_SIZE) -#define PUD_TABLE_SIZE (sizeof(pud_t) << PUD_INDEX_SIZE) -#define PGD_TABLE_SIZE (sizeof(pgd_t) << PGD_INDEX_SIZE) - -#define PTRS_PER_PTE (1 << PTE_INDEX_SIZE) -#define PTRS_PER_PMD (1 << PMD_INDEX_SIZE) -#define PTRS_PER_PUD (1 << PMD_INDEX_SIZE) -#define PTRS_PER_PGD (1 << PGD_INDEX_SIZE) - -/* PMD_SHIFT determines what a second-level page table entry can map */ -#define PMD_SHIFT (PAGE_SHIFT + PTE_INDEX_SIZE) -#define PMD_SIZE (1UL << PMD_SHIFT) -#define PMD_MASK (~(PMD_SIZE-1)) - -/* PUD_SHIFT determines what a third-level page table entry can map */ -#define PUD_SHIFT (PMD_SHIFT + PMD_INDEX_SIZE) -#define PUD_SIZE (1UL << PUD_SHIFT) -#define PUD_MASK (~(PUD_SIZE-1)) - -/* PGDIR_SHIFT determines what a fourth-level page table entry can map */ -#define PGDIR_SHIFT (PUD_SHIFT + PUD_INDEX_SIZE) -#define PGDIR_SIZE (1UL << PGDIR_SHIFT) -#define PGDIR_MASK (~(PGDIR_SIZE-1)) +#ifdef CONFIG_PPC_64K_PAGES +#include +#else +#include +#endif #define FIRST_USER_ADDRESS 0 @@ -75,8 +46,9 @@ #define VMALLOC_END (VMALLOC_START + VMALLOC_SIZE) /* - * Bits in a linux-style PTE. These match the bits in the - * (hardware-defined) PowerPC PTE as closely as possible. + * Common bits in a linux-style PTE. These match the bits in the + * (hardware-defined) PowerPC PTE as closely as possible. Additional + * bits may be defined in pgtable-*.h */ #define _PAGE_PRESENT 0x0001 /* software: pte contains a translation */ #define _PAGE_USER 0x0002 /* matches one of the PP bits */ @@ -91,15 +63,6 @@ #define _PAGE_RW 0x0200 /* software: user write access allowed */ #define _PAGE_HASHPTE 0x0400 /* software: pte has an associated HPTE */ #define _PAGE_BUSY 0x0800 /* software: PTE & hash are busy */ -#define _PAGE_SECONDARY 0x8000 /* software: HPTE is in secondary group */ -#define _PAGE_GROUP_IX 0x7000 /* software: HPTE index within group */ -#define _PAGE_HUGE 0x10000 /* 16MB page */ -/* Bits 0x7000 identify the index within an HPT Group */ -#define _PAGE_HPTEFLAGS (_PAGE_BUSY | _PAGE_HASHPTE | _PAGE_SECONDARY | _PAGE_GROUP_IX) -/* PAGE_MASK gives the right answer below, but only by accident */ -/* It should be preserving the high 48 bits and then specifically */ -/* preserving _PAGE_SECONDARY | _PAGE_GROUP_IX */ -#define _PAGE_CHG_MASK (PAGE_MASK | _PAGE_ACCESSED | _PAGE_DIRTY | _PAGE_HPTEFLAGS) #define _PAGE_BASE (_PAGE_PRESENT | _PAGE_ACCESSED | _PAGE_COHERENT) @@ -122,10 +85,10 @@ #define PAGE_AGP __pgprot(_PAGE_BASE | _PAGE_WRENABLE | _PAGE_NO_CACHE) #define HAVE_PAGE_AGP -/* - * This bit in a hardware PTE indicates that the page is *not* executable. - */ -#define HW_NO_EXEC _PAGE_EXEC +/* PTEIDX nibble */ +#define _PTEIDX_SECONDARY 0x8 +#define _PTEIDX_GROUP_IX 0x7 + /* * POWER4 and newer have per page execute protection, older chips can only @@ -164,21 +127,10 @@ extern unsigned long empty_zero_page[PAGE_SIZE/sizeof(unsigned long)]; #define ZERO_PAGE(vaddr) (virt_to_page(empty_zero_page)) #endif /* __ASSEMBLY__ */ -/* shift to put page number into pte */ -#define PTE_SHIFT (17) - #ifdef CONFIG_HUGETLB_PAGE -#ifndef __ASSEMBLY__ -int hash_huge_page(struct mm_struct *mm, unsigned long access, - unsigned long ea, unsigned long vsid, int local); -#endif /* __ASSEMBLY__ */ - #define HAVE_ARCH_UNMAPPED_AREA #define HAVE_ARCH_UNMAPPED_AREA_TOPDOWN -#else - -#define hash_huge_page(mm,a,ea,vsid,local) -1 #endif @@ -197,7 +149,7 @@ static inline pte_t pfn_pte(unsigned long pfn, pgprot_t pgprot) pte_t pte; - pte_val(pte) = (pfn << PTE_SHIFT) | pgprot_val(pgprot); + pte_val(pte) = (pfn << PTE_RPN_SHIFT) | pgprot_val(pgprot); return pte; } @@ -209,30 +161,25 @@ static inline pte_t pfn_pte(unsigned long pfn, pgprot_t pgprot) /* pte_clear moved to later in this file */ -#define pte_pfn(x) ((unsigned long)((pte_val(x) >> PTE_SHIFT))) +#define pte_pfn(x) ((unsigned long)((pte_val(x)>>PTE_RPN_SHIFT))) #define pte_page(x) pfn_to_page(pte_pfn(x)) -#define pmd_set(pmdp, ptep) ({BUG_ON((u64)ptep < KERNELBASE); pmd_val(*(pmdp)) = (unsigned long)(ptep);}) +#define pmd_set(pmdp, pmdval) (pmd_val(*(pmdp)) = (pmdval)) #define pmd_none(pmd) (!pmd_val(pmd)) #define pmd_bad(pmd) (pmd_val(pmd) == 0) #define pmd_present(pmd) (pmd_val(pmd) != 0) #define pmd_clear(pmdp) (pmd_val(*(pmdp)) = 0) -#define pmd_page_kernel(pmd) (pmd_val(pmd)) +#define pmd_page_kernel(pmd) (pmd_val(pmd) & ~PMD_MASKED_BITS) #define pmd_page(pmd) virt_to_page(pmd_page_kernel(pmd)) -#define pud_set(pudp, pmdp) (pud_val(*(pudp)) = (unsigned long)(pmdp)) +#define pud_set(pudp, pudval) (pud_val(*(pudp)) = (pudval)) #define pud_none(pud) (!pud_val(pud)) #define pud_bad(pud) ((pud_val(pud)) == 0) #define pud_present(pud) (pud_val(pud) != 0) #define pud_clear(pudp) (pud_val(*(pudp)) = 0) -#define pud_page(pud) (pud_val(pud)) +#define pud_page(pud) (pud_val(pud) & ~PUD_MASKED_BITS) #define pgd_set(pgdp, pudp) ({pgd_val(*(pgdp)) = (unsigned long)(pudp);}) -#define pgd_none(pgd) (!pgd_val(pgd)) -#define pgd_bad(pgd) (pgd_val(pgd) == 0) -#define pgd_present(pgd) (pgd_val(pgd) != 0) -#define pgd_clear(pgdp) (pgd_val(*(pgdp)) = 0) -#define pgd_page(pgd) (pgd_val(pgd)) /* * Find an entry in a page-table-directory. We combine the address region @@ -243,9 +190,6 @@ static inline pte_t pfn_pte(unsigned long pfn, pgprot_t pgprot) #define pgd_offset(mm, address) ((mm)->pgd + pgd_index(address)) -#define pud_offset(pgdp, addr) \ - (((pud_t *) pgd_page(*(pgdp))) + (((addr) >> PUD_SHIFT) & (PTRS_PER_PUD - 1))) - #define pmd_offset(pudp,addr) \ (((pmd_t *) pud_page(*(pudp))) + (((addr) >> PMD_SHIFT) & (PTRS_PER_PMD - 1))) @@ -271,7 +215,6 @@ static inline int pte_exec(pte_t pte) { return pte_val(pte) & _PAGE_EXEC;} static inline int pte_dirty(pte_t pte) { return pte_val(pte) & _PAGE_DIRTY;} static inline int pte_young(pte_t pte) { return pte_val(pte) & _PAGE_ACCESSED;} static inline int pte_file(pte_t pte) { return pte_val(pte) & _PAGE_FILE;} -static inline int pte_huge(pte_t pte) { return pte_val(pte) & _PAGE_HUGE;} static inline void pte_uncache(pte_t pte) { pte_val(pte) |= _PAGE_NO_CACHE; } static inline void pte_cache(pte_t pte) { pte_val(pte) &= ~_PAGE_NO_CACHE; } @@ -286,7 +229,6 @@ static inline pte_t pte_mkclean(pte_t pte) { pte_val(pte) &= ~(_PAGE_DIRTY); return pte; } static inline pte_t pte_mkold(pte_t pte) { pte_val(pte) &= ~_PAGE_ACCESSED; return pte; } - static inline pte_t pte_mkread(pte_t pte) { pte_val(pte) |= _PAGE_USER; return pte; } static inline pte_t pte_mkexec(pte_t pte) { @@ -298,7 +240,7 @@ static inline pte_t pte_mkdirty(pte_t pte) { static inline pte_t pte_mkyoung(pte_t pte) { pte_val(pte) |= _PAGE_ACCESSED; return pte; } static inline pte_t pte_mkhuge(pte_t pte) { - pte_val(pte) |= _PAGE_HUGE; return pte; } + return pte; } /* Atomic PTE updates */ static inline unsigned long pte_update(pte_t *p, unsigned long clr) @@ -321,11 +263,13 @@ static inline unsigned long pte_update(pte_t *p, unsigned long clr) /* PTE updating functions, this function puts the PTE in the * batch, doesn't actually triggers the hash flush immediately, * you need to call flush_tlb_pending() to do that. + * Pass -1 for "normal" size (4K or 64K) */ -extern void hpte_update(struct mm_struct *mm, unsigned long addr, unsigned long pte, - int wrprot); +extern void hpte_update(struct mm_struct *mm, unsigned long addr, + pte_t *ptep, unsigned long pte, int huge); -static inline int __ptep_test_and_clear_young(struct mm_struct *mm, unsigned long addr, pte_t *ptep) +static inline int __ptep_test_and_clear_young(struct mm_struct *mm, + unsigned long addr, pte_t *ptep) { unsigned long old; @@ -333,7 +277,7 @@ static inline int __ptep_test_and_clear_young(struct mm_struct *mm, unsigned lon return 0; old = pte_update(ptep, _PAGE_ACCESSED); if (old & _PAGE_HASHPTE) { - hpte_update(mm, addr, old, 0); + hpte_update(mm, addr, ptep, old, 0); flush_tlb_pending(); } return (old & _PAGE_ACCESSED) != 0; @@ -351,7 +295,8 @@ static inline int __ptep_test_and_clear_young(struct mm_struct *mm, unsigned lon * moment we always flush but we need to fix hpte_update and test if the * optimisation is worth it. */ -static inline int __ptep_test_and_clear_dirty(struct mm_struct *mm, unsigned long addr, pte_t *ptep) +static inline int __ptep_test_and_clear_dirty(struct mm_struct *mm, + unsigned long addr, pte_t *ptep) { unsigned long old; @@ -359,7 +304,7 @@ static inline int __ptep_test_and_clear_dirty(struct mm_struct *mm, unsigned lon return 0; old = pte_update(ptep, _PAGE_DIRTY); if (old & _PAGE_HASHPTE) - hpte_update(mm, addr, old, 0); + hpte_update(mm, addr, ptep, old, 0); return (old & _PAGE_DIRTY) != 0; } #define __HAVE_ARCH_PTEP_TEST_AND_CLEAR_DIRTY @@ -371,7 +316,8 @@ static inline int __ptep_test_and_clear_dirty(struct mm_struct *mm, unsigned lon }) #define __HAVE_ARCH_PTEP_SET_WRPROTECT -static inline void ptep_set_wrprotect(struct mm_struct *mm, unsigned long addr, pte_t *ptep) +static inline void ptep_set_wrprotect(struct mm_struct *mm, unsigned long addr, + pte_t *ptep) { unsigned long old; @@ -379,7 +325,7 @@ static inline void ptep_set_wrprotect(struct mm_struct *mm, unsigned long addr, return; old = pte_update(ptep, _PAGE_RW); if (old & _PAGE_HASHPTE) - hpte_update(mm, addr, old, 0); + hpte_update(mm, addr, ptep, old, 0); } /* @@ -408,21 +354,23 @@ static inline void ptep_set_wrprotect(struct mm_struct *mm, unsigned long addr, }) #define __HAVE_ARCH_PTEP_GET_AND_CLEAR -static inline pte_t ptep_get_and_clear(struct mm_struct *mm, unsigned long addr, pte_t *ptep) +static inline pte_t ptep_get_and_clear(struct mm_struct *mm, + unsigned long addr, pte_t *ptep) { unsigned long old = pte_update(ptep, ~0UL); if (old & _PAGE_HASHPTE) - hpte_update(mm, addr, old, 0); + hpte_update(mm, addr, ptep, old, 0); return __pte(old); } -static inline void pte_clear(struct mm_struct *mm, unsigned long addr, pte_t * ptep) +static inline void pte_clear(struct mm_struct *mm, unsigned long addr, + pte_t * ptep) { unsigned long old = pte_update(ptep, ~0UL); if (old & _PAGE_HASHPTE) - hpte_update(mm, addr, old, 0); + hpte_update(mm, addr, ptep, old, 0); } /* @@ -435,7 +383,14 @@ static inline void set_pte_at(struct mm_struct *mm, unsigned long addr, pte_clear(mm, addr, ptep); flush_tlb_pending(); } - *ptep = __pte(pte_val(pte) & ~_PAGE_HPTEFLAGS); + pte = __pte(pte_val(pte) & ~_PAGE_HPTEFLAGS); + +#ifdef CONFIG_PPC_64K_PAGES + if (mmu_virtual_psize != MMU_PAGE_64K) + pte = __pte(pte_val(pte) | _PAGE_COMBO); +#endif /* CONFIG_PPC_64K_PAGES */ + + *ptep = pte; } /* Set the dirty and/or accessed bits atomically in a linux PTE, this @@ -482,8 +437,6 @@ extern pgprot_t phys_mem_access_prot(struct file *file, unsigned long pfn, printk("%s:%d: bad pte %08lx.\n", __FILE__, __LINE__, pte_val(e)) #define pmd_ERROR(e) \ printk("%s:%d: bad pmd %08lx.\n", __FILE__, __LINE__, pmd_val(e)) -#define pud_ERROR(e) \ - printk("%s:%d: bad pud %08lx.\n", __FILE__, __LINE__, pud_val(e)) #define pgd_ERROR(e) \ printk("%s:%d: bad pgd %08lx.\n", __FILE__, __LINE__, pgd_val(e)) @@ -509,12 +462,12 @@ extern void update_mmu_cache(struct vm_area_struct *, unsigned long, pte_t); /* Encode and de-code a swap entry */ #define __swp_type(entry) (((entry).val >> 1) & 0x3f) #define __swp_offset(entry) ((entry).val >> 8) -#define __swp_entry(type, offset) ((swp_entry_t) { ((type) << 1) | ((offset) << 8) }) -#define __pte_to_swp_entry(pte) ((swp_entry_t) { pte_val(pte) >> PTE_SHIFT }) -#define __swp_entry_to_pte(x) ((pte_t) { (x).val << PTE_SHIFT }) -#define pte_to_pgoff(pte) (pte_val(pte) >> PTE_SHIFT) -#define pgoff_to_pte(off) ((pte_t) {((off) << PTE_SHIFT)|_PAGE_FILE}) -#define PTE_FILE_MAX_BITS (BITS_PER_LONG - PTE_SHIFT) +#define __swp_entry(type, offset) ((swp_entry_t){((type)<< 1)|((offset)<<8)}) +#define __pte_to_swp_entry(pte) ((swp_entry_t){pte_val(pte) >> PTE_RPN_SHIFT}) +#define __swp_entry_to_pte(x) ((pte_t) { (x).val << PTE_RPN_SHIFT }) +#define pte_to_pgoff(pte) (pte_val(pte) >> PTE_RPN_SHIFT) +#define pgoff_to_pte(off) ((pte_t) {((off) << PTE_RPN_SHIFT)|_PAGE_FILE}) +#define PTE_FILE_MAX_BITS (BITS_PER_LONG - PTE_RPN_SHIFT) /* * kern_addr_valid is intended to indicate whether an address is a valid @@ -532,29 +485,22 @@ void pgtable_cache_init(void); /* * find_linux_pte returns the address of a linux pte for a given * effective address and directory. If not found, it returns zero. - */ -static inline pte_t *find_linux_pte(pgd_t *pgdir, unsigned long ea) + */static inline pte_t *find_linux_pte(pgd_t *pgdir, unsigned long ea) { pgd_t *pg; pud_t *pu; pmd_t *pm; pte_t *pt = NULL; - pte_t pte; pg = pgdir + pgd_index(ea); if (!pgd_none(*pg)) { pu = pud_offset(pg, ea); if (!pud_none(*pu)) { pm = pmd_offset(pu, ea); - if (pmd_present(*pm)) { + if (pmd_present(*pm)) pt = pte_offset_kernel(pm, ea); - pte = *pt; - if (!pte_present(pte)) - pt = NULL; - } } } - return pt; } diff --git a/include/asm-ppc64/prom.h b/include/asm-ppc64/prom.h index e8d0d2ab4c0f..bdb47174ff0e 100644 --- a/include/asm-ppc64/prom.h +++ b/include/asm-ppc64/prom.h @@ -188,6 +188,14 @@ extern struct device_node *of_get_next_child(const struct device_node *node, extern struct device_node *of_node_get(struct device_node *node); extern void of_node_put(struct device_node *node); +/* For scanning the flat device-tree at boot time */ +int __init of_scan_flat_dt(int (*it)(unsigned long node, + const char *uname, int depth, + void *data), + void *data); +void* __init of_get_flat_dt_prop(unsigned long node, const char *name, + unsigned long *size); + /* For updating the device tree at runtime */ extern void of_attach_node(struct device_node *); extern void of_detach_node(const struct device_node *); diff --git a/include/asm-ppc64/system.h b/include/asm-ppc64/system.h index 99b8ca52f101..0cdd66c9f4b7 100644 --- a/include/asm-ppc64/system.h +++ b/include/asm-ppc64/system.h @@ -248,7 +248,7 @@ __cmpxchg_u32(volatile unsigned int *p, unsigned long old, unsigned long new) } static __inline__ unsigned long -__cmpxchg_u64(volatile long *p, unsigned long old, unsigned long new) +__cmpxchg_u64(volatile unsigned long *p, unsigned long old, unsigned long new) { unsigned long prev; diff --git a/mm/hugetlb.c b/mm/hugetlb.c index c9b43360fd33..9a565808da3f 100644 --- a/mm/hugetlb.c +++ b/mm/hugetlb.c @@ -103,6 +103,9 @@ static int __init hugetlb_init(void) unsigned long i; struct page *page; + if (HPAGE_SHIFT == 0) + return 0; + for (i = 0; i < MAX_NUMNODES; ++i) INIT_LIST_HEAD(&hugepage_freelists[i]); -- cgit v1.2.3 From ead837174d265b24334f0721a84cbdda8ccd89a7 Mon Sep 17 00:00:00 2001 From: John Rose Date: Fri, 4 Nov 2005 15:30:56 -0600 Subject: [PATCH] dlpar enable for OF pci probe This patch contains the arch/ppc64 bits for enabling DLPAR and PCI Hotplug for the new OF-based PCI probe mechanism. This code path is currently broken. Signed-off-by: John Rose Signed-off-by: Paul Mackerras --- arch/ppc64/kernel/pci.c | 17 +++++++++-------- arch/ppc64/kernel/rtas_pci.c | 6 +----- include/asm-powerpc/ppc-pci.h | 1 + include/asm-ppc64/pci.h | 8 ++++++++ 4 files changed, 19 insertions(+), 13 deletions(-) (limited to 'arch/ppc64') diff --git a/arch/ppc64/kernel/pci.c b/arch/ppc64/kernel/pci.c index 3d2106b022a1..30247ff74972 100644 --- a/arch/ppc64/kernel/pci.c +++ b/arch/ppc64/kernel/pci.c @@ -295,8 +295,8 @@ static void pci_parse_of_addrs(struct device_node *node, struct pci_dev *dev) } } -static struct pci_dev *of_create_pci_dev(struct device_node *node, - struct pci_bus *bus, int devfn) +struct pci_dev *of_create_pci_dev(struct device_node *node, + struct pci_bus *bus, int devfn) { struct pci_dev *dev; const char *type; @@ -354,10 +354,9 @@ static struct pci_dev *of_create_pci_dev(struct device_node *node, return dev; } +EXPORT_SYMBOL(of_create_pci_dev); -static void of_scan_pci_bridge(struct device_node *node, struct pci_dev *dev); - -static void __devinit of_scan_bus(struct device_node *node, +void __devinit of_scan_bus(struct device_node *node, struct pci_bus *bus) { struct device_node *child = NULL; @@ -381,9 +380,10 @@ static void __devinit of_scan_bus(struct device_node *node, do_bus_setup(bus); } +EXPORT_SYMBOL(of_scan_bus); -static void __devinit of_scan_pci_bridge(struct device_node *node, - struct pci_dev *dev) +void __devinit of_scan_pci_bridge(struct device_node *node, + struct pci_dev *dev) { struct pci_bus *bus; u32 *busrange, *ranges; @@ -464,9 +464,10 @@ static void __devinit of_scan_pci_bridge(struct device_node *node, else if (mode == PCI_PROBE_NORMAL) pci_scan_child_bus(bus); } +EXPORT_SYMBOL(of_scan_pci_bridge); #endif /* CONFIG_PPC_MULTIPLATFORM */ -static void __devinit scan_phb(struct pci_controller *hose) +void __devinit scan_phb(struct pci_controller *hose) { struct pci_bus *bus; struct device_node *node = hose->arch_data; diff --git a/arch/ppc64/kernel/rtas_pci.c b/arch/ppc64/kernel/rtas_pci.c index 3ad15c90fbbd..3c3f19192fcc 100644 --- a/arch/ppc64/kernel/rtas_pci.c +++ b/arch/ppc64/kernel/rtas_pci.c @@ -440,7 +440,6 @@ struct pci_controller * __devinit init_phb_dynamic(struct device_node *dn) struct device_node *root = of_find_node_by_path("/"); unsigned int root_size_cells = 0; struct pci_controller *phb; - struct pci_bus *bus; int primary; root_size_cells = prom_n_size_cells(root); @@ -456,10 +455,7 @@ struct pci_controller * __devinit init_phb_dynamic(struct device_node *dn) of_node_put(root); pci_devs_phb_init_dynamic(phb); - phb->last_busno = 0xff; - bus = pci_scan_bus(phb->first_busno, phb->ops, phb->arch_data); - phb->bus = bus; - phb->last_busno = bus->subordinate; + scan_phb(phb); return phb; } diff --git a/include/asm-powerpc/ppc-pci.h b/include/asm-powerpc/ppc-pci.h index a88728fba8f6..13aacff755f3 100644 --- a/include/asm-powerpc/ppc-pci.h +++ b/include/asm-powerpc/ppc-pci.h @@ -34,6 +34,7 @@ void *traverse_pci_devices(struct device_node *start, traverse_func pre, void pci_devs_phb_init(void); void pci_devs_phb_init_dynamic(struct pci_controller *phb); +void __devinit scan_phb(struct pci_controller *hose); /* PCI address cache management routines */ void pci_addr_cache_insert_device(struct pci_dev *dev); diff --git a/include/asm-ppc64/pci.h b/include/asm-ppc64/pci.h index 342e2d755550..fafdf885a3cc 100644 --- a/include/asm-ppc64/pci.h +++ b/include/asm-ppc64/pci.h @@ -162,6 +162,14 @@ pcibios_fixup_device_resources(struct pci_dev *dev, struct pci_bus *bus); extern struct pci_controller *init_phb_dynamic(struct device_node *dn); +extern struct pci_dev *of_create_pci_dev(struct device_node *node, + struct pci_bus *bus, int devfn); + +extern void of_scan_pci_bridge(struct device_node *node, + struct pci_dev *dev); + +extern void of_scan_bus(struct device_node *node, struct pci_bus *bus); + extern int pci_read_irq_line(struct pci_dev *dev); extern void pcibios_add_platform_entries(struct pci_dev *dev); -- cgit v1.2.3 From dcad47fc423ac9f4934579af814fa2dad5c8081b Mon Sep 17 00:00:00 2001 From: David Gibson Date: Mon, 7 Nov 2005 09:49:43 +1100 Subject: [PATCH] powerpc: Kill ppcdebug The ancient ppcdebug/PPCDBG mechanism is now only used in two places. First, in the hash setup code, one of the bits allows the size of the hash table to be reduced by a factor of 8 - which would be better accomplished with a command line option for that purpose. The other was a bunch of bus walking related messages in the iSeries code, which would seem to be insufficient reason to keep the mechanism. This patch removes the last traces of this mechanism. Built and booted on iSeries and pSeries POWER5 LPAR (ARCH=powerpc). Signed-off-by: David Gibson Signed-off-by: Paul Mackerras --- arch/powerpc/kernel/setup_64.c | 9 +-- arch/powerpc/kernel/signal_32.c | 1 - arch/powerpc/kernel/signal_64.c | 1 - arch/powerpc/mm/hash_utils_64.c | 7 --- arch/powerpc/mm/init_64.c | 1 - arch/powerpc/mm/pgtable_64.c | 1 - arch/powerpc/platforms/iseries/irq.c | 5 -- arch/powerpc/platforms/iseries/pci.c | 37 ----------- arch/powerpc/platforms/iseries/setup.c | 4 -- arch/powerpc/platforms/iseries/smp.c | 1 - arch/powerpc/platforms/pseries/iommu.c | 1 - arch/powerpc/platforms/pseries/lpar.c | 2 +- arch/powerpc/platforms/pseries/ras.c | 2 +- arch/powerpc/sysdev/u3_iommu.c | 1 - arch/ppc64/Kconfig.debug | 4 -- arch/ppc64/kernel/prom.c | 1 - arch/ppc64/kernel/prom_init.c | 1 - arch/ppc64/kernel/udbg.c | 55 ----------------- include/asm-ppc64/ppcdebug.h | 108 --------------------------------- include/asm-ppc64/udbg.h | 3 - 20 files changed, 3 insertions(+), 242 deletions(-) delete mode 100644 include/asm-ppc64/ppcdebug.h (limited to 'arch/ppc64') diff --git a/arch/powerpc/kernel/setup_64.c b/arch/powerpc/kernel/setup_64.c index 6b52cce872be..5f8154f95f96 100644 --- a/arch/powerpc/kernel/setup_64.c +++ b/arch/powerpc/kernel/setup_64.c @@ -41,7 +41,6 @@ #include #include #include -#include #include #include #include @@ -60,6 +59,7 @@ #include #include #include +#include #ifdef DEBUG #define DBG(fmt...) udbg_printf(fmt) @@ -243,12 +243,6 @@ void __init early_setup(unsigned long dt_ptr) DBG(" -> early_setup()\n"); - /* - * Fill the default DBG level (do we want to keep - * that old mecanism around forever ?) - */ - ppcdbg_initialize(); - /* * Do early initializations using the flattened device * tree, like retreiving the physical memory map or @@ -516,7 +510,6 @@ void __init setup_system(void) printk("-----------------------------------------------------\n"); printk("ppc64_pft_size = 0x%lx\n", ppc64_pft_size); - printk("ppc64_debug_switch = 0x%lx\n", ppc64_debug_switch); printk("ppc64_interrupt_controller = 0x%ld\n", ppc64_interrupt_controller); printk("systemcfg = 0x%p\n", systemcfg); printk("systemcfg->platform = 0x%x\n", systemcfg->platform); diff --git a/arch/powerpc/kernel/signal_32.c b/arch/powerpc/kernel/signal_32.c index 876c57c11365..081d931eae48 100644 --- a/arch/powerpc/kernel/signal_32.c +++ b/arch/powerpc/kernel/signal_32.c @@ -44,7 +44,6 @@ #include #ifdef CONFIG_PPC64 #include "ppc32.h" -#include #include #include #else diff --git a/arch/powerpc/kernel/signal_64.c b/arch/powerpc/kernel/signal_64.c index ec9d0984b6a0..58194e150711 100644 --- a/arch/powerpc/kernel/signal_64.c +++ b/arch/powerpc/kernel/signal_64.c @@ -33,7 +33,6 @@ #include #include #include -#include #include #include #include diff --git a/arch/powerpc/mm/hash_utils_64.c b/arch/powerpc/mm/hash_utils_64.c index 6e9e05cce02c..b6970c96d96f 100644 --- a/arch/powerpc/mm/hash_utils_64.c +++ b/arch/powerpc/mm/hash_utils_64.c @@ -32,7 +32,6 @@ #include #include -#include #include #include #include @@ -194,12 +193,6 @@ void __init htab_initialize(void) htab_size_bytes = get_hashtable_size(); pteg_count = htab_size_bytes >> 7; - /* For debug, make the HTAB 1/8 as big as it normally would be. */ - ifppcdebug(PPCDBG_HTABSIZE) { - pteg_count >>= 3; - htab_size_bytes = pteg_count << 7; - } - htab_hash_mask = pteg_count - 1; if (systemcfg->platform & PLATFORM_LPAR) { diff --git a/arch/powerpc/mm/init_64.c b/arch/powerpc/mm/init_64.c index b0fc822ec29f..aa98b79e734c 100644 --- a/arch/powerpc/mm/init_64.c +++ b/arch/powerpc/mm/init_64.c @@ -57,7 +57,6 @@ #include #include #include -#include #include #include #include diff --git a/arch/powerpc/mm/pgtable_64.c b/arch/powerpc/mm/pgtable_64.c index b79a78206135..d1fbd3fe684f 100644 --- a/arch/powerpc/mm/pgtable_64.c +++ b/arch/powerpc/mm/pgtable_64.c @@ -59,7 +59,6 @@ #include #include #include -#include #include #include #include diff --git a/arch/powerpc/platforms/iseries/irq.c b/arch/powerpc/platforms/iseries/irq.c index c1135912cc05..a06603d84a45 100644 --- a/arch/powerpc/platforms/iseries/irq.c +++ b/arch/powerpc/platforms/iseries/irq.c @@ -35,7 +35,6 @@ #include #include -#include #include #include #include @@ -227,8 +226,6 @@ static void iSeries_enable_IRQ(unsigned int irq) /* Unmask secondary INTA */ mask = 0x80000000; HvCallPci_unmaskInterrupts(bus, subBus, deviceId, mask); - PPCDBG(PPCDBG_BUSWALK, "iSeries_enable_IRQ 0x%02X.%02X.%02X 0x%04X\n", - bus, subBus, deviceId, irq); } /* This is called by iSeries_activate_IRQs */ @@ -310,8 +307,6 @@ static void iSeries_disable_IRQ(unsigned int irq) /* Mask secondary INTA */ mask = 0x80000000; HvCallPci_maskInterrupts(bus, subBus, deviceId, mask); - PPCDBG(PPCDBG_BUSWALK, "iSeries_disable_IRQ 0x%02X.%02X.%02X 0x%04X\n", - bus, subBus, deviceId, irq); } /* diff --git a/arch/powerpc/platforms/iseries/pci.c b/arch/powerpc/platforms/iseries/pci.c index 7d7d5884343f..4b75131773a6 100644 --- a/arch/powerpc/platforms/iseries/pci.c +++ b/arch/powerpc/platforms/iseries/pci.c @@ -32,7 +32,6 @@ #include #include #include -#include #include #include @@ -207,10 +206,6 @@ static struct device_node *build_device_node(HvBusNumber Bus, struct device_node *node; struct pci_dn *pdn; - PPCDBG(PPCDBG_BUSWALK, - "-build_device_node 0x%02X.%02X.%02X Function: %02X\n", - Bus, SubBus, AgentId, Function); - node = kmalloc(sizeof(struct device_node), GFP_KERNEL); if (node == NULL) return NULL; @@ -243,8 +238,6 @@ unsigned long __init find_and_init_phbs(void) struct pci_controller *phb; HvBusNumber bus; - PPCDBG(PPCDBG_BUSWALK, "find_and_init_phbs Entry\n"); - /* Check all possible buses. */ for (bus = 0; bus < 256; bus++) { int ret = HvCallXm_testBus(bus); @@ -261,9 +254,6 @@ unsigned long __init find_and_init_phbs(void) phb->last_busno = bus; phb->ops = &iSeries_pci_ops; - PPCDBG(PPCDBG_BUSWALK, "PCI:Create iSeries pci_controller(%p), Bus: %04X\n", - phb, bus); - /* Find and connect the devices. */ scan_PHB_slots(phb); } @@ -285,11 +275,9 @@ unsigned long __init find_and_init_phbs(void) */ void iSeries_pcibios_init(void) { - PPCDBG(PPCDBG_BUSWALK, "iSeries_pcibios_init Entry.\n"); iomm_table_initialize(); find_and_init_phbs(); io_page_mask = -1; - PPCDBG(PPCDBG_BUSWALK, "iSeries_pcibios_init Exit.\n"); } /* @@ -301,8 +289,6 @@ void __init iSeries_pci_final_fixup(void) struct device_node *node; int DeviceCount = 0; - PPCDBG(PPCDBG_BUSWALK, "iSeries_pcibios_fixup Entry.\n"); - /* Fix up at the device node and pci_dev relationship */ mf_display_src(0xC9000100); @@ -316,9 +302,6 @@ void __init iSeries_pci_final_fixup(void) ++DeviceCount; pdev->sysdata = (void *)node; PCI_DN(node)->pcidev = pdev; - PPCDBG(PPCDBG_BUSWALK, - "pdev 0x%p <==> DevNode 0x%p\n", - pdev, node); allocate_device_bars(pdev); iSeries_Device_Information(pdev, DeviceCount); iommu_devnode_init_iSeries(node); @@ -333,13 +316,10 @@ void __init iSeries_pci_final_fixup(void) void pcibios_fixup_bus(struct pci_bus *PciBus) { - PPCDBG(PPCDBG_BUSWALK, "iSeries_pcibios_fixup_bus(0x%04X) Entry.\n", - PciBus->number); } void pcibios_fixup_resources(struct pci_dev *pdev) { - PPCDBG(PPCDBG_BUSWALK, "fixup_resources pdev %p\n", pdev); } /* @@ -401,9 +381,6 @@ static void scan_EADS_bridge(HvBusNumber bus, HvSubBusNumber SubBus, printk("found device at bus %d idsel %d func %d (AgentId %x)\n", bus, IdSel, Function, AgentId); /* Connect EADs: 0x18.00.12 = 0x00 */ - PPCDBG(PPCDBG_BUSWALK, - "PCI:Connect EADs: 0x%02X.%02X.%02X\n", - bus, SubBus, AgentId); HvRc = HvCallPci_getBusUnitInfo(bus, SubBus, AgentId, iseries_hv_addr(BridgeInfo), sizeof(struct HvCallPci_BridgeInfo)); @@ -414,14 +391,6 @@ static void scan_EADS_bridge(HvBusNumber bus, HvSubBusNumber SubBus, BridgeInfo->maxAgents, BridgeInfo->maxSubBusNumber, BridgeInfo->logicalSlotNumber); - PPCDBG(PPCDBG_BUSWALK, - "PCI: BridgeInfo, Type:0x%02X, SubBus:0x%02X, MaxAgents:0x%02X, MaxSubBus: 0x%02X, LSlot: 0x%02X\n", - BridgeInfo->busUnitInfo.deviceType, - BridgeInfo->subBusNumber, - BridgeInfo->maxAgents, - BridgeInfo->maxSubBusNumber, - BridgeInfo->logicalSlotNumber); - if (BridgeInfo->busUnitInfo.deviceType == HvCallPci_BridgeDevice) { /* Scan_Bridge_Slot...: 0x18.00.12 */ @@ -454,9 +423,6 @@ static int scan_bridge_slot(HvBusNumber Bus, /* iSeries_allocate_IRQ.: 0x18.00.12(0xA3) */ Irq = iSeries_allocate_IRQ(Bus, 0, EADsIdSel); - PPCDBG(PPCDBG_BUSWALK, - "PCI:- allocate and assign IRQ 0x%02X.%02X.%02X = 0x%02X\n", - Bus, 0, EADsIdSel, Irq); /* * Connect all functions of any device found. @@ -482,9 +448,6 @@ static int scan_bridge_slot(HvBusNumber Bus, printk("read vendor ID: %x\n", VendorId); /* FoundDevice: 0x18.28.10 = 0x12AE */ - PPCDBG(PPCDBG_BUSWALK, - "PCI:- FoundDevice: 0x%02X.%02X.%02X = 0x%04X, irq %d\n", - Bus, SubBus, AgentId, VendorId, Irq); HvRc = HvCallPci_configStore8(Bus, SubBus, AgentId, PCI_INTERRUPT_LINE, Irq); if (HvRc != 0) diff --git a/arch/powerpc/platforms/iseries/setup.c b/arch/powerpc/platforms/iseries/setup.c index fda712b42168..36f89e9ec7d0 100644 --- a/arch/powerpc/platforms/iseries/setup.c +++ b/arch/powerpc/platforms/iseries/setup.c @@ -71,8 +71,6 @@ extern void hvlog(char *fmt, ...); #endif /* Function Prototypes */ -extern void ppcdbg_initialize(void); - static void build_iSeries_Memory_Map(void); static void iseries_shared_idle(void); static void iseries_dedicated_idle(void); @@ -309,8 +307,6 @@ static void __init iSeries_init_early(void) ppc64_firmware_features = FW_FEATURE_ISERIES; - ppcdbg_initialize(); - ppc64_interrupt_controller = IC_ISERIES; #if defined(CONFIG_BLK_DEV_INITRD) diff --git a/arch/powerpc/platforms/iseries/smp.c b/arch/powerpc/platforms/iseries/smp.c index 3336bad67724..fcb094ec6aec 100644 --- a/arch/powerpc/platforms/iseries/smp.c +++ b/arch/powerpc/platforms/iseries/smp.c @@ -40,7 +40,6 @@ #include #include #include -#include #include #include #include diff --git a/arch/powerpc/platforms/pseries/iommu.c b/arch/powerpc/platforms/pseries/iommu.c index 513e27231493..8a99902c7ef7 100644 --- a/arch/powerpc/platforms/pseries/iommu.c +++ b/arch/powerpc/platforms/pseries/iommu.c @@ -37,7 +37,6 @@ #include #include #include -#include #include #include #include diff --git a/arch/powerpc/platforms/pseries/lpar.c b/arch/powerpc/platforms/pseries/lpar.c index e384a5a91796..d84d76f0026b 100644 --- a/arch/powerpc/platforms/pseries/lpar.c +++ b/arch/powerpc/platforms/pseries/lpar.c @@ -31,13 +31,13 @@ #include #include #include -#include #include #include #include #include #include #include +#include #include "plpar_wrappers.h" diff --git a/arch/powerpc/platforms/pseries/ras.c b/arch/powerpc/platforms/pseries/ras.c index 6562ff4b0a82..fbd214d68b07 100644 --- a/arch/powerpc/platforms/pseries/ras.c +++ b/arch/powerpc/platforms/pseries/ras.c @@ -48,7 +48,7 @@ #include #include #include -#include +#include static unsigned char ras_log_buf[RTAS_ERROR_LOG_MAX]; static DEFINE_SPINLOCK(ras_log_buf_lock); diff --git a/arch/powerpc/sysdev/u3_iommu.c b/arch/powerpc/sysdev/u3_iommu.c index 607722178c1a..543d65909812 100644 --- a/arch/powerpc/sysdev/u3_iommu.c +++ b/arch/powerpc/sysdev/u3_iommu.c @@ -37,7 +37,6 @@ #include #include #include -#include #include #include #include diff --git a/arch/ppc64/Kconfig.debug b/arch/ppc64/Kconfig.debug index f16a5030527b..b258c9314a1b 100644 --- a/arch/ppc64/Kconfig.debug +++ b/arch/ppc64/Kconfig.debug @@ -55,10 +55,6 @@ config XMON_DEFAULT xmon is normally disabled unless booted with 'xmon=on'. Use 'xmon=off' to disable xmon init during runtime. -config PPCDBG - bool "Include PPCDBG realtime debugging" - depends on DEBUG_KERNEL - config IRQSTACKS bool "Use separate kernel stacks when processing interrupts" help diff --git a/arch/ppc64/kernel/prom.c b/arch/ppc64/kernel/prom.c index 97bfceb5353b..a4cb97a25819 100644 --- a/arch/ppc64/kernel/prom.c +++ b/arch/ppc64/kernel/prom.c @@ -46,7 +46,6 @@ #include #include #include -#include #include #include #include diff --git a/arch/ppc64/kernel/prom_init.c b/arch/ppc64/kernel/prom_init.c index a4bbca6dbb8b..e72b3f9061f0 100644 --- a/arch/ppc64/kernel/prom_init.c +++ b/arch/ppc64/kernel/prom_init.c @@ -44,7 +44,6 @@ #include #include #include -#include #include #include #include diff --git a/arch/ppc64/kernel/udbg.c b/arch/ppc64/kernel/udbg.c index d49c3613c8ec..0d878e72fc44 100644 --- a/arch/ppc64/kernel/udbg.c +++ b/arch/ppc64/kernel/udbg.c @@ -10,12 +10,10 @@ */ #include -#define WANT_PPCDBG_TAB /* Only defined here */ #include #include #include #include -#include #include void (*udbg_putc)(unsigned char c); @@ -89,59 +87,6 @@ void udbg_printf(const char *fmt, ...) va_end(args); } -/* PPCDBG stuff */ - -u64 ppc64_debug_switch; - -/* Special print used by PPCDBG() macro */ -void udbg_ppcdbg(unsigned long debug_flags, const char *fmt, ...) -{ - unsigned long active_debugs = debug_flags & ppc64_debug_switch; - - if (active_debugs) { - va_list ap; - unsigned char buf[UDBG_BUFSIZE]; - unsigned long i, len = 0; - - for (i=0; i < PPCDBG_NUM_FLAGS; i++) { - if (((1U << i) & active_debugs) && - trace_names[i]) { - len += strlen(trace_names[i]); - udbg_puts(trace_names[i]); - break; - } - } - - snprintf(buf, UDBG_BUFSIZE, " [%s]: ", current->comm); - len += strlen(buf); - udbg_puts(buf); - - while (len < 18) { - udbg_puts(" "); - len++; - } - - va_start(ap, fmt); - vsnprintf(buf, UDBG_BUFSIZE, fmt, ap); - udbg_puts(buf); - va_end(ap); - } -} - -unsigned long udbg_ifdebug(unsigned long flags) -{ - return (flags & ppc64_debug_switch); -} - -/* - * Initialize the PPCDBG state. Called before relocation has been enabled. - */ -void __init ppcdbg_initialize(void) -{ - ppc64_debug_switch = PPC_DEBUG_DEFAULT; /* | PPCDBG_BUSWALK | */ - /* PPCDBG_PHBINIT | PPCDBG_MM | PPCDBG_MMINIT | PPCDBG_TCEINIT | PPCDBG_TCE */; -} - /* * Early boot console based on udbg */ diff --git a/include/asm-ppc64/ppcdebug.h b/include/asm-ppc64/ppcdebug.h deleted file mode 100644 index fd7f696065c4..000000000000 --- a/include/asm-ppc64/ppcdebug.h +++ /dev/null @@ -1,108 +0,0 @@ -#ifndef __PPCDEBUG_H -#define __PPCDEBUG_H -/******************************************************************** - * Author: Adam Litke, IBM Corp - * (c) 2001 - * - * This file contains definitions and macros for a runtime debugging - * system for ppc64 (This should also work on 32 bit with a few - * adjustments. - * - * This program is free software; you can redistribute it and/or - * modify it under the terms of the GNU General Public License - * as published by the Free Software Foundation; either version - * 2 of the License, or (at your option) any later version. - * - ********************************************************************/ - -#include -#include -#include -#include - -#define PPCDBG_BITVAL(X) ((1UL)<<((unsigned long)(X))) - -/* Defined below are the bit positions of various debug flags in the - * ppc64_debug_switch variable. - * -- When adding new values, please enter them into trace names below -- - * - * Values 62 & 63 can be used to stress the hardware page table management - * code. They must be set statically, any attempt to change them dynamically - * would be a very bad idea. - */ -#define PPCDBG_MMINIT PPCDBG_BITVAL(0) -#define PPCDBG_MM PPCDBG_BITVAL(1) -#define PPCDBG_SYS32 PPCDBG_BITVAL(2) -#define PPCDBG_SYS32NI PPCDBG_BITVAL(3) -#define PPCDBG_SYS32X PPCDBG_BITVAL(4) -#define PPCDBG_SYS32M PPCDBG_BITVAL(5) -#define PPCDBG_SYS64 PPCDBG_BITVAL(6) -#define PPCDBG_SYS64NI PPCDBG_BITVAL(7) -#define PPCDBG_SYS64X PPCDBG_BITVAL(8) -#define PPCDBG_SIGNAL PPCDBG_BITVAL(9) -#define PPCDBG_SIGNALXMON PPCDBG_BITVAL(10) -#define PPCDBG_BINFMT32 PPCDBG_BITVAL(11) -#define PPCDBG_BINFMT64 PPCDBG_BITVAL(12) -#define PPCDBG_BINFMTXMON PPCDBG_BITVAL(13) -#define PPCDBG_BINFMT_32ADDR PPCDBG_BITVAL(14) -#define PPCDBG_ALIGNFIXUP PPCDBG_BITVAL(15) -#define PPCDBG_TCEINIT PPCDBG_BITVAL(16) -#define PPCDBG_TCE PPCDBG_BITVAL(17) -#define PPCDBG_PHBINIT PPCDBG_BITVAL(18) -#define PPCDBG_SMP PPCDBG_BITVAL(19) -#define PPCDBG_BOOT PPCDBG_BITVAL(20) -#define PPCDBG_BUSWALK PPCDBG_BITVAL(21) -#define PPCDBG_PROM PPCDBG_BITVAL(22) -#define PPCDBG_RTAS PPCDBG_BITVAL(23) -#define PPCDBG_HTABSTRESS PPCDBG_BITVAL(62) -#define PPCDBG_HTABSIZE PPCDBG_BITVAL(63) -#define PPCDBG_NONE (0UL) -#define PPCDBG_ALL (0xffffffffUL) - -/* The default initial value for the debug switch */ -#define PPC_DEBUG_DEFAULT 0 -/* #define PPC_DEBUG_DEFAULT PPCDBG_ALL */ - -#define PPCDBG_NUM_FLAGS 64 - -extern u64 ppc64_debug_switch; - -#ifdef WANT_PPCDBG_TAB -/* A table of debug switch names to allow name lookup in xmon - * (and whoever else wants it. - */ -char *trace_names[PPCDBG_NUM_FLAGS] = { - /* Known debug names */ - "mminit", "mm", - "syscall32", "syscall32_ni", "syscall32x", "syscall32m", - "syscall64", "syscall64_ni", "syscall64x", - "signal", "signal_xmon", - "binfmt32", "binfmt64", "binfmt_xmon", "binfmt_32addr", - "alignfixup", "tceinit", "tce", "phb_init", - "smp", "boot", "buswalk", "prom", - "rtas" -}; -#else -extern char *trace_names[64]; -#endif /* WANT_PPCDBG_TAB */ - -#ifdef CONFIG_PPCDBG -/* Macro to conditionally print debug based on debug_switch */ -#define PPCDBG(...) udbg_ppcdbg(__VA_ARGS__) - -/* Macro to conditionally call a debug routine based on debug_switch */ -#define PPCDBGCALL(FLAGS,FUNCTION) ifppcdebug(FLAGS) FUNCTION - -/* Macros to test for debug states */ -#define ifppcdebug(FLAGS) if (udbg_ifdebug(FLAGS)) -#define ppcdebugset(FLAGS) (udbg_ifdebug(FLAGS)) -#define PPCDBG_BINFMT (test_thread_flag(TIF_32BIT) ? PPCDBG_BINFMT32 : PPCDBG_BINFMT64) - -#else -#define PPCDBG(...) do {;} while (0) -#define PPCDBGCALL(FLAGS,FUNCTION) do {;} while (0) -#define ifppcdebug(...) if (0) -#define ppcdebugset(FLAGS) (0) -#endif /* CONFIG_PPCDBG */ - -#endif /*__PPCDEBUG_H */ diff --git a/include/asm-ppc64/udbg.h b/include/asm-ppc64/udbg.h index 8192fb8541cc..e3b927991851 100644 --- a/include/asm-ppc64/udbg.h +++ b/include/asm-ppc64/udbg.h @@ -23,9 +23,6 @@ extern int udbg_read(char *buf, int buflen); extern void register_early_udbg_console(void); extern void udbg_printf(const char *fmt, ...); -extern void udbg_ppcdbg(unsigned long flags, const char *fmt, ...); -extern unsigned long udbg_ifdebug(unsigned long flags); -extern void __init ppcdbg_initialize(void); extern void udbg_init_uart(void __iomem *comport, unsigned int speed); -- cgit v1.2.3 From 2249ca9d60d3a8a1f6f223f0f0a0283fcb7ce33e Mon Sep 17 00:00:00 2001 From: Paul Mackerras Date: Mon, 7 Nov 2005 13:18:13 +1100 Subject: powerpc: Various UP build fixes Mostly this involves adding #include , since that defines things like boot_cpuid[_phys] and [gs]et_hard_smp_processor_id, which are SMP-related but still needed on UP. This incorporates fixes posted by Olof Johansson and Heikki Lindholm. Signed-off-by: Paul Mackerras --- arch/powerpc/kernel/rtas.c | 2 -- arch/powerpc/kernel/time.c | 5 +---- arch/powerpc/lib/locks.c | 1 + arch/powerpc/mm/numa.c | 1 + arch/powerpc/platforms/pseries/iommu.c | 1 + arch/powerpc/platforms/pseries/lpar.c | 1 + arch/powerpc/platforms/pseries/setup.c | 1 + arch/ppc64/kernel/idle.c | 1 + arch/ppc64/kernel/machine_kexec.c | 1 + arch/ppc64/kernel/sysfs.c | 1 + include/asm-powerpc/smp.h | 4 +++- 11 files changed, 12 insertions(+), 7 deletions(-) (limited to 'arch/ppc64') diff --git a/arch/powerpc/kernel/rtas.c b/arch/powerpc/kernel/rtas.c index b7fc2d884950..751f5ddea913 100644 --- a/arch/powerpc/kernel/rtas.c +++ b/arch/powerpc/kernel/rtas.c @@ -608,7 +608,6 @@ asmlinkage int ppc_rtas(struct rtas_args __user *uargs) return 0; } -#ifdef CONFIG_SMP /* This version can't take the spinlock, because it never returns */ struct rtas_args rtas_stop_self_args = { @@ -633,7 +632,6 @@ void rtas_stop_self(void) panic("Alas, I survived.\n"); } -#endif /* * Call early during boot, before mem init or bootmem, to retreive the RTAS diff --git a/arch/powerpc/kernel/time.c b/arch/powerpc/kernel/time.c index 6996a593dcb3..1c7ea5e06311 100644 --- a/arch/powerpc/kernel/time.c +++ b/arch/powerpc/kernel/time.c @@ -61,6 +61,7 @@ #include #include #include +#include #ifdef CONFIG_PPC64 #include #include @@ -118,10 +119,6 @@ static unsigned adjusting_time = 0; unsigned long ppc_proc_freq; unsigned long ppc_tb_freq; -#ifdef CONFIG_PPC32 /* XXX for now */ -#define boot_cpuid 0 -#endif - u64 tb_last_jiffy __cacheline_aligned_in_smp; unsigned long tb_last_stamp; diff --git a/arch/powerpc/lib/locks.c b/arch/powerpc/lib/locks.c index 2a912f411eb4..35bd03c41dd1 100644 --- a/arch/powerpc/lib/locks.c +++ b/arch/powerpc/lib/locks.c @@ -23,6 +23,7 @@ #if defined(CONFIG_PPC_SPLPAR) || defined(CONFIG_PPC_ISERIES) #include #include +#include void __spin_yield(raw_spinlock_t *lock) { diff --git a/arch/powerpc/mm/numa.c b/arch/powerpc/mm/numa.c index 4035cad8d7f1..da09ba03c424 100644 --- a/arch/powerpc/mm/numa.c +++ b/arch/powerpc/mm/numa.c @@ -21,6 +21,7 @@ #include #include #include +#include static int numa_enabled = 1; diff --git a/arch/powerpc/platforms/pseries/iommu.c b/arch/powerpc/platforms/pseries/iommu.c index 8a99902c7ef7..fcc50bfd43fd 100644 --- a/arch/powerpc/platforms/pseries/iommu.c +++ b/arch/powerpc/platforms/pseries/iommu.c @@ -46,6 +46,7 @@ #include #include #include +#include #include "plpar_wrappers.h" diff --git a/arch/powerpc/platforms/pseries/lpar.c b/arch/powerpc/platforms/pseries/lpar.c index d84d76f0026b..8a42006370c5 100644 --- a/arch/powerpc/platforms/pseries/lpar.c +++ b/arch/powerpc/platforms/pseries/lpar.c @@ -38,6 +38,7 @@ #include #include #include +#include #include "plpar_wrappers.h" diff --git a/arch/powerpc/platforms/pseries/setup.c b/arch/powerpc/platforms/pseries/setup.c index 65bee939eecc..934d7006353f 100644 --- a/arch/powerpc/platforms/pseries/setup.c +++ b/arch/powerpc/platforms/pseries/setup.c @@ -65,6 +65,7 @@ #include #include #include +#include #include "plpar_wrappers.h" diff --git a/arch/ppc64/kernel/idle.c b/arch/ppc64/kernel/idle.c index 8abd2ad92832..8fec27469802 100644 --- a/arch/ppc64/kernel/idle.c +++ b/arch/ppc64/kernel/idle.c @@ -28,6 +28,7 @@ #include #include #include +#include extern void power4_idle(void); diff --git a/arch/ppc64/kernel/machine_kexec.c b/arch/ppc64/kernel/machine_kexec.c index ff8679f260f3..07ea03598c00 100644 --- a/arch/ppc64/kernel/machine_kexec.c +++ b/arch/ppc64/kernel/machine_kexec.c @@ -24,6 +24,7 @@ #include #include /* _end */ #include +#include #define HASH_GROUP_SIZE 0x80 /* size of each hash group, asm/mmu.h */ diff --git a/arch/ppc64/kernel/sysfs.c b/arch/ppc64/kernel/sysfs.c index 6654b350979c..e99ec62c2c52 100644 --- a/arch/ppc64/kernel/sysfs.c +++ b/arch/ppc64/kernel/sysfs.c @@ -20,6 +20,7 @@ #include #include #include +#include static DEFINE_PER_CPU(struct cpu, cpu_devices); diff --git a/include/asm-powerpc/smp.h b/include/asm-powerpc/smp.h index 8bcdd0faefea..98581e5a8279 100644 --- a/include/asm-powerpc/smp.h +++ b/include/asm-powerpc/smp.h @@ -86,7 +86,6 @@ extern void __cpu_die(unsigned int cpu); #else /* for UP */ #define smp_setup_cpu_maps() -#define smp_release_cpus() #endif /* CONFIG_SMP */ @@ -94,6 +93,9 @@ extern void __cpu_die(unsigned int cpu); #define get_hard_smp_processor_id(CPU) (paca[(CPU)].hw_cpu_id) #define set_hard_smp_processor_id(CPU, VAL) \ do { (paca[(CPU)].hw_cpu_id = (VAL)); } while (0) + +extern void smp_release_cpus(void); + #else /* 32-bit */ #ifndef CONFIG_SMP -- cgit v1.2.3 From 732ee21f2894819781766a0cd88e32bdd630d11e Mon Sep 17 00:00:00 2001 From: Olof Johansson Date: Mon, 7 Nov 2005 00:57:55 -0800 Subject: [PATCH] POWERPC/PPC64: Fix CONFIG_SMP=n build for ppc64 Two CONFIG_SMP=n build fixes due to missing includes. Signed-off-by: Olof Johansson Cc: Benjamin Herrenschmidt Cc: Paul Mackerras Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- arch/powerpc/kernel/time.c | 1 + arch/ppc64/kernel/sysfs.c | 1 + 2 files changed, 2 insertions(+) (limited to 'arch/ppc64') diff --git a/arch/powerpc/kernel/time.c b/arch/powerpc/kernel/time.c index 6996a593dcb3..b1c89bc4bf90 100644 --- a/arch/powerpc/kernel/time.c +++ b/arch/powerpc/kernel/time.c @@ -69,6 +69,7 @@ #include #include #endif +#include /* keep track of when we need to update the rtc */ time_t last_rtc_update; diff --git a/arch/ppc64/kernel/sysfs.c b/arch/ppc64/kernel/sysfs.c index 6654b350979c..e99ec62c2c52 100644 --- a/arch/ppc64/kernel/sysfs.c +++ b/arch/ppc64/kernel/sysfs.c @@ -20,6 +20,7 @@ #include #include #include +#include static DEFINE_PER_CPU(struct cpu, cpu_devices); -- cgit v1.2.3 From c8e3c8b21bd7a317d071ab8cf478880e7a4f92d6 Mon Sep 17 00:00:00 2001 From: Benjamin Herrenschmidt Date: Mon, 7 Nov 2005 00:57:58 -0800 Subject: [PATCH] ppc64: Fix zImage boot The zImage wrapper has a bug where it doesn't claim() the memory for the kernel properly, it forgets to take into account the offset between the ELF header and the kernel itself. This results on some machines, like G5s, into a kernel that crashes at boot when clearing the BSS. Signed-off-by: Benjamin Herrenschmidt Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- arch/ppc64/boot/main.c | 11 +++++++++-- 1 file changed, 9 insertions(+), 2 deletions(-) (limited to 'arch/ppc64') diff --git a/arch/ppc64/boot/main.c b/arch/ppc64/boot/main.c index c1dc876bccab..e0dde24a72ce 100644 --- a/arch/ppc64/boot/main.c +++ b/arch/ppc64/boot/main.c @@ -203,8 +203,15 @@ void start(unsigned long a1, unsigned long a2, void *promptr, void *sp) if (elf64ph->p_type == PT_LOAD && elf64ph->p_offset != 0) break; } - vmlinux.size = (unsigned long)elf64ph->p_filesz; - vmlinux.memsize = (unsigned long)elf64ph->p_memsz; + vmlinux.size = (unsigned long)elf64ph->p_filesz + + (unsigned long)elf64ph->p_offset; + /* We need to claim the memsize plus the file offset since gzip + * will expand the header (file offset), then the kernel, then + * possible rubbish we don't care about. But the kernel bss must + * be claimed (it will be zero'd by the kernel itself) + */ + vmlinux.memsize = (unsigned long)elf64ph->p_memsz + + (unsigned long)elf64ph->p_offset; printf("Allocating 0x%lx bytes for kernel ...\n\r", vmlinux.memsize); vmlinux.addr = try_claim(vmlinux.memsize); if (vmlinux.addr == 0) { -- cgit v1.2.3 From 66ff2d0691e00e1e7bfdf398a970310c9a0fe671 Mon Sep 17 00:00:00 2001 From: Ananth N Mavinakayanahalli Date: Mon, 7 Nov 2005 01:00:07 -0800 Subject: [PATCH] Kprobes: rearrange preempt_disable/enable() calls The following set of patches are aimed at improving kprobes scalability. We currently serialize kprobe registration, unregistration and handler execution using a single spinlock - kprobe_lock. With these changes, kprobe handlers can run without any locks held. It also allows for simultaneous kprobe handler executions on different processors as we now track kprobe execution on a per processor basis. It is now necessary that the handlers be re-entrant since handlers can run concurrently on multiple processors. All changes have been tested on i386, ia64, ppc64 and x86_64, while sparc64 has been compile tested only. The patches can be viewed as 3 logical chunks: patch 1: Reorder preempt_(dis/en)able calls patches 2-7: Introduce per_cpu data areas to track kprobe execution patches 8-9: Use RCU to synchronize kprobe (un)registration and handler execution. Thanks to Maneesh Soni, James Keniston and Anil Keshavamurthy for their review and suggestions. Thanks again to Anil, Hien Nguyen and Kevin Stafford for testing the patches. This patch: Reorder preempt_disable/enable() calls in arch kprobes files in preparation to introduce locking changes. No functional changes introduced by this patch. Signed-off-by: Ananth N Mavinakayahanalli Signed-off-by: Anil S Keshavamurthy Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- arch/i386/kernel/kprobes.c | 35 ++++++++++++++++++----------------- arch/ia64/kernel/kprobes.c | 22 ++++++++++++++-------- arch/ppc64/kernel/kprobes.c | 11 ++++++----- arch/sparc64/kernel/kprobes.c | 25 +++++++++++++------------ arch/x86_64/kernel/kprobes.c | 28 ++++++++++++++-------------- 5 files changed, 65 insertions(+), 56 deletions(-) (limited to 'arch/ppc64') diff --git a/arch/i386/kernel/kprobes.c b/arch/i386/kernel/kprobes.c index 6345b430b105..fd35039859e6 100644 --- a/arch/i386/kernel/kprobes.c +++ b/arch/i386/kernel/kprobes.c @@ -158,8 +158,6 @@ static int __kprobes kprobe_handler(struct pt_regs *regs) kprobe_opcode_t *addr = NULL; unsigned long *lp; - /* We're in an interrupt, but this is clear and BUG()-safe. */ - preempt_disable(); /* Check if the application is using LDT entry for its code segment and * calculate the address by reading the base address from the LDT entry. */ @@ -232,6 +230,11 @@ static int __kprobes kprobe_handler(struct pt_regs *regs) goto no_kprobe; } + /* + * This preempt_disable() matches the preempt_enable_no_resched() + * in post_kprobe_handler() + */ + preempt_disable(); kprobe_status = KPROBE_HIT_ACTIVE; set_current_kprobe(p, regs); @@ -245,7 +248,6 @@ ss_probe: return 1; no_kprobe: - preempt_enable_no_resched(); return ret; } @@ -313,11 +315,11 @@ int __kprobes trampoline_probe_handler(struct kprobe *p, struct pt_regs *regs) unlock_kprobes(); preempt_enable_no_resched(); - /* - * By returning a non-zero value, we are telling - * kprobe_handler() that we have handled unlocking - * and re-enabling preemption. - */ + /* + * By returning a non-zero value, we are telling + * kprobe_handler() that we have handled unlocking + * and re-enabling preemption + */ return 1; } @@ -453,29 +455,29 @@ int __kprobes kprobe_exceptions_notify(struct notifier_block *self, unsigned long val, void *data) { struct die_args *args = (struct die_args *)data; + int ret = NOTIFY_DONE; + + preempt_disable(); switch (val) { case DIE_INT3: if (kprobe_handler(args->regs)) - return NOTIFY_STOP; + ret = NOTIFY_STOP; break; case DIE_DEBUG: if (post_kprobe_handler(args->regs)) - return NOTIFY_STOP; + ret = NOTIFY_STOP; break; case DIE_GPF: - if (kprobe_running() && - kprobe_fault_handler(args->regs, args->trapnr)) - return NOTIFY_STOP; - break; case DIE_PAGE_FAULT: if (kprobe_running() && kprobe_fault_handler(args->regs, args->trapnr)) - return NOTIFY_STOP; + ret = NOTIFY_STOP; break; default: break; } - return NOTIFY_DONE; + preempt_enable(); + return ret; } int __kprobes setjmp_pre_handler(struct kprobe *p, struct pt_regs *regs) @@ -502,7 +504,6 @@ int __kprobes setjmp_pre_handler(struct kprobe *p, struct pt_regs *regs) void __kprobes jprobe_return(void) { - preempt_enable_no_resched(); asm volatile (" xchgl %%ebx,%%esp \n" " int3 \n" " .globl jprobe_return_end \n" diff --git a/arch/ia64/kernel/kprobes.c b/arch/ia64/kernel/kprobes.c index 471086b808a4..1e80ec80dd21 100644 --- a/arch/ia64/kernel/kprobes.c +++ b/arch/ia64/kernel/kprobes.c @@ -395,7 +395,7 @@ int __kprobes trampoline_probe_handler(struct kprobe *p, struct pt_regs *regs) /* * By returning a non-zero value, we are telling * kprobe_handler() that we have handled unlocking - * and re-enabling preemption. + * and re-enabling preemption */ return 1; } @@ -607,8 +607,6 @@ static int __kprobes pre_kprobes_handler(struct die_args *args) struct pt_regs *regs = args->regs; kprobe_opcode_t *addr = (kprobe_opcode_t *)instruction_pointer(regs); - preempt_disable(); - /* Handle recursion cases */ if (kprobe_running()) { p = get_kprobe(addr); @@ -665,6 +663,11 @@ static int __kprobes pre_kprobes_handler(struct die_args *args) goto no_kprobe; } + /* + * This preempt_disable() matches the preempt_enable_no_resched() + * in post_kprobes_handler() + */ + preempt_disable(); kprobe_status = KPROBE_HIT_ACTIVE; set_current_kprobe(p); @@ -682,7 +685,6 @@ ss_probe: return 1; no_kprobe: - preempt_enable_no_resched(); return ret; } @@ -733,22 +735,26 @@ int __kprobes kprobe_exceptions_notify(struct notifier_block *self, unsigned long val, void *data) { struct die_args *args = (struct die_args *)data; + int ret = NOTIFY_DONE; + + preempt_disable(); switch(val) { case DIE_BREAK: if (pre_kprobes_handler(args)) - return NOTIFY_STOP; + ret = NOTIFY_STOP; break; case DIE_SS: if (post_kprobes_handler(args->regs)) - return NOTIFY_STOP; + ret = NOTIFY_STOP; break; case DIE_PAGE_FAULT: if (kprobes_fault_handler(args->regs, args->trapnr)) - return NOTIFY_STOP; + ret = NOTIFY_STOP; default: break; } - return NOTIFY_DONE; + preempt_enable(); + return ret; } int __kprobes setjmp_pre_handler(struct kprobe *p, struct pt_regs *regs) diff --git a/arch/ppc64/kernel/kprobes.c b/arch/ppc64/kernel/kprobes.c index ed876a5178ae..6071ee99f5cb 100644 --- a/arch/ppc64/kernel/kprobes.c +++ b/arch/ppc64/kernel/kprobes.c @@ -209,6 +209,11 @@ static inline int kprobe_handler(struct pt_regs *regs) goto no_kprobe; } + /* + * This preempt_disable() matches the preempt_enable_no_resched() + * in post_kprobe_handler(). + */ + preempt_disable(); kprobe_status = KPROBE_HIT_ACTIVE; current_kprobe = p; kprobe_saved_msr = regs->msr; @@ -219,11 +224,6 @@ static inline int kprobe_handler(struct pt_regs *regs) ss_probe: prepare_singlestep(p, regs); kprobe_status = KPROBE_HIT_SS; - /* - * This preempt_disable() matches the preempt_enable_no_resched() - * in post_kprobe_handler(). - */ - preempt_disable(); return 1; no_kprobe: @@ -293,6 +293,7 @@ int __kprobes trampoline_probe_handler(struct kprobe *p, struct pt_regs *regs) regs->nip = orig_ret_address; unlock_kprobes(); + preempt_enable_no_resched(); /* * By returning a non-zero value, we are telling diff --git a/arch/sparc64/kernel/kprobes.c b/arch/sparc64/kernel/kprobes.c index 0d66d07c8c6e..755a0d7d887f 100644 --- a/arch/sparc64/kernel/kprobes.c +++ b/arch/sparc64/kernel/kprobes.c @@ -118,8 +118,6 @@ static int __kprobes kprobe_handler(struct pt_regs *regs) void *addr = (void *) regs->tpc; int ret = 0; - preempt_disable(); - if (kprobe_running()) { /* We *are* holding lock here, so this is safe. * Disarm the probe we just hit, and ignore it. @@ -171,6 +169,11 @@ static int __kprobes kprobe_handler(struct pt_regs *regs) goto no_kprobe; } + /* + * This preempt_disable() matches the preempt_enable_no_resched() + * in post_kprobes_handler() + */ + preempt_disable(); set_current_kprobe(p, regs); kprobe_status = KPROBE_HIT_ACTIVE; if (p->pre_handler && p->pre_handler(p, regs)) @@ -182,7 +185,6 @@ ss_probe: return 1; no_kprobe: - preempt_enable_no_resched(); return ret; } @@ -322,29 +324,29 @@ int __kprobes kprobe_exceptions_notify(struct notifier_block *self, unsigned long val, void *data) { struct die_args *args = (struct die_args *)data; + int ret = NOTIFY_DONE; + + preempt_disable(); switch (val) { case DIE_DEBUG: if (kprobe_handler(args->regs)) - return NOTIFY_STOP; + ret = NOTIFY_STOP; break; case DIE_DEBUG_2: if (post_kprobe_handler(args->regs)) - return NOTIFY_STOP; + ret = NOTIFY_STOP; break; case DIE_GPF: - if (kprobe_running() && - kprobe_fault_handler(args->regs, args->trapnr)) - return NOTIFY_STOP; - break; case DIE_PAGE_FAULT: if (kprobe_running() && kprobe_fault_handler(args->regs, args->trapnr)) - return NOTIFY_STOP; + ret = NOTIFY_STOP; break; default: break; } - return NOTIFY_DONE; + preempt_enable(); + return ret; } asmlinkage void __kprobes kprobe_trap(unsigned long trap_level, @@ -396,7 +398,6 @@ int __kprobes setjmp_pre_handler(struct kprobe *p, struct pt_regs *regs) void __kprobes jprobe_return(void) { - preempt_enable_no_resched(); __asm__ __volatile__( ".globl jprobe_return_trap_instruction\n" "jprobe_return_trap_instruction:\n\t" diff --git a/arch/x86_64/kernel/kprobes.c b/arch/x86_64/kernel/kprobes.c index 76a28b007be9..ebfa2c9241ca 100644 --- a/arch/x86_64/kernel/kprobes.c +++ b/arch/x86_64/kernel/kprobes.c @@ -302,9 +302,6 @@ int __kprobes kprobe_handler(struct pt_regs *regs) int ret = 0; kprobe_opcode_t *addr = (kprobe_opcode_t *)(regs->rip - sizeof(kprobe_opcode_t)); - /* We're in an interrupt, but this is clear and BUG()-safe. */ - preempt_disable(); - /* Check we're not actually recursing */ if (kprobe_running()) { /* We *are* holding lock here, so this is safe. @@ -372,6 +369,11 @@ int __kprobes kprobe_handler(struct pt_regs *regs) goto no_kprobe; } + /* + * This preempt_disable() matches the preempt_enable_no_resched() + * in post_kprobe_handler() + */ + preempt_disable(); kprobe_status = KPROBE_HIT_ACTIVE; set_current_kprobe(p, regs); @@ -385,7 +387,6 @@ ss_probe: return 1; no_kprobe: - preempt_enable_no_resched(); return ret; } @@ -456,7 +457,7 @@ int __kprobes trampoline_probe_handler(struct kprobe *p, struct pt_regs *regs) /* * By returning a non-zero value, we are telling * kprobe_handler() that we have handled unlocking - * and re-enabling preemption. + * and re-enabling preemption */ return 1; } @@ -599,29 +600,29 @@ int __kprobes kprobe_exceptions_notify(struct notifier_block *self, unsigned long val, void *data) { struct die_args *args = (struct die_args *)data; + int ret = NOTIFY_DONE; + + preempt_disable(); switch (val) { case DIE_INT3: if (kprobe_handler(args->regs)) - return NOTIFY_STOP; + ret = NOTIFY_STOP; break; case DIE_DEBUG: if (post_kprobe_handler(args->regs)) - return NOTIFY_STOP; + ret = NOTIFY_STOP; break; case DIE_GPF: - if (kprobe_running() && - kprobe_fault_handler(args->regs, args->trapnr)) - return NOTIFY_STOP; - break; case DIE_PAGE_FAULT: if (kprobe_running() && kprobe_fault_handler(args->regs, args->trapnr)) - return NOTIFY_STOP; + ret = NOTIFY_STOP; break; default: break; } - return NOTIFY_DONE; + preempt_enable(); + return ret; } int __kprobes setjmp_pre_handler(struct kprobe *p, struct pt_regs *regs) @@ -647,7 +648,6 @@ int __kprobes setjmp_pre_handler(struct kprobe *p, struct pt_regs *regs) void __kprobes jprobe_return(void) { - preempt_enable_no_resched(); asm volatile (" xchg %%rbx,%%rsp \n" " int3 \n" " .globl jprobe_return_end \n" -- cgit v1.2.3 From 0dc036c91ac11b2b76bb91b59d8c7af919aa4a8d Mon Sep 17 00:00:00 2001 From: Ananth N Mavinakayanahalli Date: Mon, 7 Nov 2005 01:00:10 -0800 Subject: [PATCH] Kprobes: Track kprobe on a per_cpu basis - ppc64 changes PPC64 changes to track kprobe execution on a per-cpu basis. We now track the kprobe state machine independently on each cpu using an arch specific kprobe control block. Signed-off-by: Ananth N Mavinakayanahalli Signed-off-by: Anil S Keshavamurthy Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- arch/ppc64/kernel/kprobes.c | 94 +++++++++++++++++++++++++------------------ include/asm-powerpc/kprobes.h | 15 +++++++ 2 files changed, 69 insertions(+), 40 deletions(-) (limited to 'arch/ppc64') diff --git a/arch/ppc64/kernel/kprobes.c b/arch/ppc64/kernel/kprobes.c index 6071ee99f5cb..3f89f3e5584a 100644 --- a/arch/ppc64/kernel/kprobes.c +++ b/arch/ppc64/kernel/kprobes.c @@ -37,12 +37,8 @@ #include static DECLARE_MUTEX(kprobe_mutex); - -static struct kprobe *current_kprobe; -static unsigned long kprobe_status, kprobe_saved_msr; -static struct kprobe *kprobe_prev; -static unsigned long kprobe_status_prev, kprobe_saved_msr_prev; -static struct pt_regs jprobe_saved_regs; +DEFINE_PER_CPU(struct kprobe *, current_kprobe) = NULL; +DEFINE_PER_CPU(struct kprobe_ctlblk, kprobe_ctlblk); int __kprobes arch_prepare_kprobe(struct kprobe *p) { @@ -108,18 +104,25 @@ static inline void prepare_singlestep(struct kprobe *p, struct pt_regs *regs) regs->nip = (unsigned long)p->ainsn.insn; } -static inline void save_previous_kprobe(void) +static inline void save_previous_kprobe(struct kprobe_ctlblk *kcb) +{ + kcb->prev_kprobe.kp = kprobe_running(); + kcb->prev_kprobe.status = kcb->kprobe_status; + kcb->prev_kprobe.saved_msr = kcb->kprobe_saved_msr; +} + +static inline void restore_previous_kprobe(struct kprobe_ctlblk *kcb) { - kprobe_prev = current_kprobe; - kprobe_status_prev = kprobe_status; - kprobe_saved_msr_prev = kprobe_saved_msr; + __get_cpu_var(current_kprobe) = kcb->prev_kprobe.kp; + kcb->kprobe_status = kcb->prev_kprobe.status; + kcb->kprobe_saved_msr = kcb->prev_kprobe.saved_msr; } -static inline void restore_previous_kprobe(void) +static inline void set_current_kprobe(struct kprobe *p, struct pt_regs *regs, + struct kprobe_ctlblk *kcb) { - current_kprobe = kprobe_prev; - kprobe_status = kprobe_status_prev; - kprobe_saved_msr = kprobe_saved_msr_prev; + __get_cpu_var(current_kprobe) = p; + kcb->kprobe_saved_msr = regs->msr; } void __kprobes arch_prepare_kretprobe(struct kretprobe *rp, @@ -145,6 +148,7 @@ static inline int kprobe_handler(struct pt_regs *regs) struct kprobe *p; int ret = 0; unsigned int *addr = (unsigned int *)regs->nip; + struct kprobe_ctlblk *kcb = get_kprobe_ctlblk(); /* Check we're not actually recursing */ if (kprobe_running()) { @@ -153,10 +157,10 @@ static inline int kprobe_handler(struct pt_regs *regs) p = get_kprobe(addr); if (p) { kprobe_opcode_t insn = *p->ainsn.insn; - if (kprobe_status == KPROBE_HIT_SS && + if (kcb->kprobe_status == KPROBE_HIT_SS && is_trap(insn)) { regs->msr &= ~MSR_SE; - regs->msr |= kprobe_saved_msr; + regs->msr |= kcb->kprobe_saved_msr; unlock_kprobes(); goto no_kprobe; } @@ -166,15 +170,15 @@ static inline int kprobe_handler(struct pt_regs *regs) * just single step on the instruction of the new probe * without calling any user handlers. */ - save_previous_kprobe(); - current_kprobe = p; - kprobe_saved_msr = regs->msr; + save_previous_kprobe(kcb); + set_current_kprobe(p, regs, kcb); + kcb->kprobe_saved_msr = regs->msr; p->nmissed++; prepare_singlestep(p, regs); - kprobe_status = KPROBE_REENTER; + kcb->kprobe_status = KPROBE_REENTER; return 1; } else { - p = current_kprobe; + p = __get_cpu_var(current_kprobe); if (p->break_handler && p->break_handler(p, regs)) { goto ss_probe; } @@ -214,16 +218,15 @@ static inline int kprobe_handler(struct pt_regs *regs) * in post_kprobe_handler(). */ preempt_disable(); - kprobe_status = KPROBE_HIT_ACTIVE; - current_kprobe = p; - kprobe_saved_msr = regs->msr; + kcb->kprobe_status = KPROBE_HIT_ACTIVE; + set_current_kprobe(p, regs, kcb); if (p->pre_handler && p->pre_handler(p, regs)) /* handler has already set things up, so skip ss setup */ return 1; ss_probe: prepare_singlestep(p, regs); - kprobe_status = KPROBE_HIT_SS; + kcb->kprobe_status = KPROBE_HIT_SS; return 1; no_kprobe: @@ -292,6 +295,7 @@ int __kprobes trampoline_probe_handler(struct kprobe *p, struct pt_regs *regs) BUG_ON(!orig_ret_address || (orig_ret_address == trampoline_address)); regs->nip = orig_ret_address; + reset_current_kprobe(); unlock_kprobes(); preempt_enable_no_resched(); @@ -324,22 +328,26 @@ static void __kprobes resume_execution(struct kprobe *p, struct pt_regs *regs) static inline int post_kprobe_handler(struct pt_regs *regs) { - if (!kprobe_running()) + struct kprobe *cur = kprobe_running(); + struct kprobe_ctlblk *kcb = get_kprobe_ctlblk(); + + if (!cur) return 0; - if ((kprobe_status != KPROBE_REENTER) && current_kprobe->post_handler) { - kprobe_status = KPROBE_HIT_SSDONE; - current_kprobe->post_handler(current_kprobe, regs, 0); + if ((kcb->kprobe_status != KPROBE_REENTER) && cur->post_handler) { + kcb->kprobe_status = KPROBE_HIT_SSDONE; + cur->post_handler(cur, regs, 0); } - resume_execution(current_kprobe, regs); - regs->msr |= kprobe_saved_msr; + resume_execution(cur, regs); + regs->msr |= kcb->kprobe_saved_msr; /*Restore back the original saved kprobes variables and continue. */ - if (kprobe_status == KPROBE_REENTER) { - restore_previous_kprobe(); + if (kcb->kprobe_status == KPROBE_REENTER) { + restore_previous_kprobe(kcb); goto out; } + reset_current_kprobe(); unlock_kprobes(); out: preempt_enable_no_resched(); @@ -358,15 +366,18 @@ out: /* Interrupts disabled, kprobe_lock held. */ static inline int kprobe_fault_handler(struct pt_regs *regs, int trapnr) { - if (current_kprobe->fault_handler - && current_kprobe->fault_handler(current_kprobe, regs, trapnr)) + struct kprobe *cur = kprobe_running(); + struct kprobe_ctlblk *kcb = get_kprobe_ctlblk(); + + if (cur->fault_handler && cur->fault_handler(cur, regs, trapnr)) return 1; - if (kprobe_status & KPROBE_HIT_SS) { - resume_execution(current_kprobe, regs); + if (kcb->kprobe_status & KPROBE_HIT_SS) { + resume_execution(cur, regs); regs->msr &= ~MSR_SE; - regs->msr |= kprobe_saved_msr; + regs->msr |= kcb->kprobe_saved_msr; + reset_current_kprobe(); unlock_kprobes(); preempt_enable_no_resched(); } @@ -411,8 +422,9 @@ int __kprobes kprobe_exceptions_notify(struct notifier_block *self, int __kprobes setjmp_pre_handler(struct kprobe *p, struct pt_regs *regs) { struct jprobe *jp = container_of(p, struct jprobe, kp); + struct kprobe_ctlblk *kcb = get_kprobe_ctlblk(); - memcpy(&jprobe_saved_regs, regs, sizeof(struct pt_regs)); + memcpy(&kcb->jprobe_saved_regs, regs, sizeof(struct pt_regs)); /* setup return addr to the jprobe handler routine */ regs->nip = (unsigned long)(((func_descr_t *)jp->entry)->entry); @@ -432,12 +444,14 @@ void __kprobes jprobe_return_end(void) int __kprobes longjmp_break_handler(struct kprobe *p, struct pt_regs *regs) { + struct kprobe_ctlblk *kcb = get_kprobe_ctlblk(); + /* * FIXME - we should ideally be validating that we got here 'cos * of the "trap" in jprobe_return() above, before restoring the * saved regs... */ - memcpy(regs, &jprobe_saved_regs, sizeof(struct pt_regs)); + memcpy(regs, &kcb->jprobe_saved_regs, sizeof(struct pt_regs)); return 1; } diff --git a/include/asm-powerpc/kprobes.h b/include/asm-powerpc/kprobes.h index b2f09f17fbe0..6cd0a3bfa280 100644 --- a/include/asm-powerpc/kprobes.h +++ b/include/asm-powerpc/kprobes.h @@ -27,6 +27,7 @@ */ #include #include +#include struct pt_regs; @@ -53,6 +54,20 @@ struct arch_specific_insn { kprobe_opcode_t *insn; }; +struct prev_kprobe { + struct kprobe *kp; + unsigned long status; + unsigned long saved_msr; +}; + +/* per-cpu kprobe control block */ +struct kprobe_ctlblk { + unsigned long kprobe_status; + unsigned long kprobe_saved_msr; + struct pt_regs jprobe_saved_regs; + struct prev_kprobe prev_kprobe; +}; + #ifdef CONFIG_KPROBES extern int kprobe_exceptions_notify(struct notifier_block *self, unsigned long val, void *data); -- cgit v1.2.3 From 991a51d83a3d9bebfafdd1e692cf310899d60791 Mon Sep 17 00:00:00 2001 From: Ananth N Mavinakayanahalli Date: Mon, 7 Nov 2005 01:00:14 -0800 Subject: [PATCH] Kprobes: Use RCU for (un)register synchronization - arch changes Changes to the arch kprobes infrastructure to take advantage of the locking changes introduced by usage of RCU for synchronization. All handlers are now run without any locks held, so they have to be re-entrant or provide their own synchronization. Signed-off-by: Ananth N Mavinakayanahalli Signed-off-by: Anil S Keshavamurthy Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- arch/i386/kernel/kprobes.c | 22 +++++++--------------- arch/ia64/kernel/kprobes.c | 16 ++++++---------- arch/ppc64/kernel/kprobes.c | 24 ++++++------------------ arch/sparc64/kernel/kprobes.c | 14 ++------------ arch/x86_64/kernel/kprobes.c | 25 ++++++------------------- 5 files changed, 27 insertions(+), 74 deletions(-) (limited to 'arch/ppc64') diff --git a/arch/i386/kernel/kprobes.c b/arch/i386/kernel/kprobes.c index 99565a66915d..ad469299267a 100644 --- a/arch/i386/kernel/kprobes.c +++ b/arch/i386/kernel/kprobes.c @@ -31,7 +31,6 @@ #include #include #include -#include #include #include #include @@ -123,6 +122,7 @@ static inline void prepare_singlestep(struct kprobe *p, struct pt_regs *regs) regs->eip = (unsigned long)&p->ainsn.insn; } +/* Called with kretprobe_lock held */ void __kprobes arch_prepare_kretprobe(struct kretprobe *rp, struct pt_regs *regs) { @@ -168,15 +168,12 @@ static int __kprobes kprobe_handler(struct pt_regs *regs) } /* Check we're not actually recursing */ if (kprobe_running()) { - /* We *are* holding lock here, so this is safe. - Disarm the probe we just hit, and ignore it. */ p = get_kprobe(addr); if (p) { if (kcb->kprobe_status == KPROBE_HIT_SS && *p->ainsn.insn == BREAKPOINT_INSTRUCTION) { regs->eflags &= ~TF_MASK; regs->eflags |= kcb->kprobe_saved_eflags; - unlock_kprobes(); goto no_kprobe; } /* We have reentered the kprobe_handler(), since @@ -197,14 +194,11 @@ static int __kprobes kprobe_handler(struct pt_regs *regs) goto ss_probe; } } - /* If it's not ours, can't be delete race, (we hold lock). */ goto no_kprobe; } - lock_kprobes(); p = get_kprobe(addr); if (!p) { - unlock_kprobes(); if (regs->eflags & VM_MASK) { /* We are in virtual-8086 mode. Return 0 */ goto no_kprobe; @@ -268,9 +262,10 @@ int __kprobes trampoline_probe_handler(struct kprobe *p, struct pt_regs *regs) struct kretprobe_instance *ri = NULL; struct hlist_head *head; struct hlist_node *node, *tmp; - unsigned long orig_ret_address = 0; + unsigned long flags, orig_ret_address = 0; unsigned long trampoline_address =(unsigned long)&kretprobe_trampoline; + spin_lock_irqsave(&kretprobe_lock, flags); head = kretprobe_inst_table_head(current); /* @@ -310,7 +305,7 @@ int __kprobes trampoline_probe_handler(struct kprobe *p, struct pt_regs *regs) regs->eip = orig_ret_address; reset_current_kprobe(); - unlock_kprobes(); + spin_unlock_irqrestore(&kretprobe_lock, flags); preempt_enable_no_resched(); /* @@ -395,7 +390,7 @@ static void __kprobes resume_execution(struct kprobe *p, /* * Interrupts are disabled on entry as trap1 is an interrupt gate and they - * remain disabled thoroughout this function. And we hold kprobe lock. + * remain disabled thoroughout this function. */ static inline int post_kprobe_handler(struct pt_regs *regs) { @@ -419,7 +414,6 @@ static inline int post_kprobe_handler(struct pt_regs *regs) goto out; } reset_current_kprobe(); - unlock_kprobes(); out: preempt_enable_no_resched(); @@ -434,7 +428,6 @@ out: return 1; } -/* Interrupts disabled, kprobe_lock held. */ static inline int kprobe_fault_handler(struct pt_regs *regs, int trapnr) { struct kprobe *cur = kprobe_running(); @@ -448,7 +441,6 @@ static inline int kprobe_fault_handler(struct pt_regs *regs, int trapnr) regs->eflags |= kcb->kprobe_old_eflags; reset_current_kprobe(); - unlock_kprobes(); preempt_enable_no_resched(); } return 0; @@ -463,7 +455,7 @@ int __kprobes kprobe_exceptions_notify(struct notifier_block *self, struct die_args *args = (struct die_args *)data; int ret = NOTIFY_DONE; - preempt_disable(); + rcu_read_lock(); switch (val) { case DIE_INT3: if (kprobe_handler(args->regs)) @@ -482,7 +474,7 @@ int __kprobes kprobe_exceptions_notify(struct notifier_block *self, default: break; } - preempt_enable(); + rcu_read_unlock(); return ret; } diff --git a/arch/ia64/kernel/kprobes.c b/arch/ia64/kernel/kprobes.c index 17e70b1b8d79..fddbac32d44a 100644 --- a/arch/ia64/kernel/kprobes.c +++ b/arch/ia64/kernel/kprobes.c @@ -26,7 +26,6 @@ #include #include #include -#include #include #include #include @@ -343,10 +342,11 @@ int __kprobes trampoline_probe_handler(struct kprobe *p, struct pt_regs *regs) struct kretprobe_instance *ri = NULL; struct hlist_head *head; struct hlist_node *node, *tmp; - unsigned long orig_ret_address = 0; + unsigned long flags, orig_ret_address = 0; unsigned long trampoline_address = ((struct fnptr *)kretprobe_trampoline)->ip; + spin_lock_irqsave(&kretprobe_lock, flags); head = kretprobe_inst_table_head(current); /* @@ -386,7 +386,7 @@ int __kprobes trampoline_probe_handler(struct kprobe *p, struct pt_regs *regs) regs->cr_iip = orig_ret_address; reset_current_kprobe(); - unlock_kprobes(); + spin_unlock_irqrestore(&kretprobe_lock, flags); preempt_enable_no_resched(); /* @@ -397,6 +397,7 @@ int __kprobes trampoline_probe_handler(struct kprobe *p, struct pt_regs *regs) return 1; } +/* Called with kretprobe_lock held */ void __kprobes arch_prepare_kretprobe(struct kretprobe *rp, struct pt_regs *regs) { @@ -612,7 +613,6 @@ static int __kprobes pre_kprobes_handler(struct die_args *args) if ((kcb->kprobe_status == KPROBE_HIT_SS) && (p->ainsn.inst_flag == INST_FLAG_BREAK_INST)) { ia64_psr(regs)->ss = 0; - unlock_kprobes(); goto no_kprobe; } /* We have reentered the pre_kprobe_handler(), since @@ -641,10 +641,8 @@ static int __kprobes pre_kprobes_handler(struct die_args *args) } } - lock_kprobes(); p = get_kprobe(addr); if (!p) { - unlock_kprobes(); if (!is_ia64_break_inst(regs)) { /* * The breakpoint instruction was removed right @@ -707,7 +705,6 @@ static int __kprobes post_kprobes_handler(struct pt_regs *regs) goto out; } reset_current_kprobe(); - unlock_kprobes(); out: preempt_enable_no_resched(); @@ -728,7 +725,6 @@ static int __kprobes kprobes_fault_handler(struct pt_regs *regs, int trapnr) if (kcb->kprobe_status & KPROBE_HIT_SS) { resume_execution(cur, regs); reset_current_kprobe(); - unlock_kprobes(); preempt_enable_no_resched(); } @@ -741,7 +737,7 @@ int __kprobes kprobe_exceptions_notify(struct notifier_block *self, struct die_args *args = (struct die_args *)data; int ret = NOTIFY_DONE; - preempt_disable(); + rcu_read_lock(); switch(val) { case DIE_BREAK: if (pre_kprobes_handler(args)) @@ -757,7 +753,7 @@ int __kprobes kprobe_exceptions_notify(struct notifier_block *self, default: break; } - preempt_enable(); + rcu_read_unlock(); return ret; } diff --git a/arch/ppc64/kernel/kprobes.c b/arch/ppc64/kernel/kprobes.c index 3f89f3e5584a..e0a25b35437f 100644 --- a/arch/ppc64/kernel/kprobes.c +++ b/arch/ppc64/kernel/kprobes.c @@ -30,7 +30,6 @@ #include #include #include -#include #include #include #include @@ -125,6 +124,7 @@ static inline void set_current_kprobe(struct kprobe *p, struct pt_regs *regs, kcb->kprobe_saved_msr = regs->msr; } +/* Called with kretprobe_lock held */ void __kprobes arch_prepare_kretprobe(struct kretprobe *rp, struct pt_regs *regs) { @@ -152,8 +152,6 @@ static inline int kprobe_handler(struct pt_regs *regs) /* Check we're not actually recursing */ if (kprobe_running()) { - /* We *are* holding lock here, so this is safe. - Disarm the probe we just hit, and ignore it. */ p = get_kprobe(addr); if (p) { kprobe_opcode_t insn = *p->ainsn.insn; @@ -161,7 +159,6 @@ static inline int kprobe_handler(struct pt_regs *regs) is_trap(insn)) { regs->msr &= ~MSR_SE; regs->msr |= kcb->kprobe_saved_msr; - unlock_kprobes(); goto no_kprobe; } /* We have reentered the kprobe_handler(), since @@ -183,14 +180,11 @@ static inline int kprobe_handler(struct pt_regs *regs) goto ss_probe; } } - /* If it's not ours, can't be delete race, (we hold lock). */ goto no_kprobe; } - lock_kprobes(); p = get_kprobe(addr); if (!p) { - unlock_kprobes(); if (*addr != BREAKPOINT_INSTRUCTION) { /* * PowerPC has multiple variants of the "trap" @@ -254,9 +248,10 @@ int __kprobes trampoline_probe_handler(struct kprobe *p, struct pt_regs *regs) struct kretprobe_instance *ri = NULL; struct hlist_head *head; struct hlist_node *node, *tmp; - unsigned long orig_ret_address = 0; + unsigned long flags, orig_ret_address = 0; unsigned long trampoline_address =(unsigned long)&kretprobe_trampoline; + spin_lock_irqsave(&kretprobe_lock, flags); head = kretprobe_inst_table_head(current); /* @@ -296,7 +291,7 @@ int __kprobes trampoline_probe_handler(struct kprobe *p, struct pt_regs *regs) regs->nip = orig_ret_address; reset_current_kprobe(); - unlock_kprobes(); + spin_unlock_irqrestore(&kretprobe_lock, flags); preempt_enable_no_resched(); /* @@ -348,7 +343,6 @@ static inline int post_kprobe_handler(struct pt_regs *regs) goto out; } reset_current_kprobe(); - unlock_kprobes(); out: preempt_enable_no_resched(); @@ -363,7 +357,6 @@ out: return 1; } -/* Interrupts disabled, kprobe_lock held. */ static inline int kprobe_fault_handler(struct pt_regs *regs, int trapnr) { struct kprobe *cur = kprobe_running(); @@ -378,7 +371,6 @@ static inline int kprobe_fault_handler(struct pt_regs *regs, int trapnr) regs->msr |= kcb->kprobe_saved_msr; reset_current_kprobe(); - unlock_kprobes(); preempt_enable_no_resched(); } return 0; @@ -393,11 +385,7 @@ int __kprobes kprobe_exceptions_notify(struct notifier_block *self, struct die_args *args = (struct die_args *)data; int ret = NOTIFY_DONE; - /* - * Interrupts are not disabled here. We need to disable - * preemption, because kprobe_running() uses smp_processor_id(). - */ - preempt_disable(); + rcu_read_lock(); switch (val) { case DIE_BPT: if (kprobe_handler(args->regs)) @@ -415,7 +403,7 @@ int __kprobes kprobe_exceptions_notify(struct notifier_block *self, default: break; } - preempt_enable_no_resched(); + rcu_read_unlock(); return ret; } diff --git a/arch/sparc64/kernel/kprobes.c b/arch/sparc64/kernel/kprobes.c index b95984154dba..58a815e90373 100644 --- a/arch/sparc64/kernel/kprobes.c +++ b/arch/sparc64/kernel/kprobes.c @@ -116,15 +116,11 @@ static int __kprobes kprobe_handler(struct pt_regs *regs) struct kprobe_ctlblk *kcb = get_kprobe_ctlblk(); if (kprobe_running()) { - /* We *are* holding lock here, so this is safe. - * Disarm the probe we just hit, and ignore it. - */ p = get_kprobe(addr); if (p) { if (kcb->kprobe_status == KPROBE_HIT_SS) { regs->tstate = ((regs->tstate & ~TSTATE_PIL) | kcb->kprobe_orig_tstate_pil); - unlock_kprobes(); goto no_kprobe; } /* We have reentered the kprobe_handler(), since @@ -144,14 +140,11 @@ static int __kprobes kprobe_handler(struct pt_regs *regs) if (p->break_handler && p->break_handler(p, regs)) goto ss_probe; } - /* If it's not ours, can't be delete race, (we hold lock). */ goto no_kprobe; } - lock_kprobes(); p = get_kprobe(addr); if (!p) { - unlock_kprobes(); if (*(u32 *)addr != BREAKPOINT_INSTRUCTION) { /* * The breakpoint instruction was removed right @@ -296,14 +289,12 @@ static inline int post_kprobe_handler(struct pt_regs *regs) goto out; } reset_current_kprobe(); - unlock_kprobes(); out: preempt_enable_no_resched(); return 1; } -/* Interrupts disabled, kprobe_lock held. */ static inline int kprobe_fault_handler(struct pt_regs *regs, int trapnr) { struct kprobe *cur = kprobe_running(); @@ -316,7 +307,6 @@ static inline int kprobe_fault_handler(struct pt_regs *regs, int trapnr) resume_execution(cur, regs, kcb); reset_current_kprobe(); - unlock_kprobes(); preempt_enable_no_resched(); } return 0; @@ -331,7 +321,7 @@ int __kprobes kprobe_exceptions_notify(struct notifier_block *self, struct die_args *args = (struct die_args *)data; int ret = NOTIFY_DONE; - preempt_disable(); + rcu_read_lock(); switch (val) { case DIE_DEBUG: if (kprobe_handler(args->regs)) @@ -350,7 +340,7 @@ int __kprobes kprobe_exceptions_notify(struct notifier_block *self, default: break; } - preempt_enable(); + rcu_read_unlock(); return ret; } diff --git a/arch/x86_64/kernel/kprobes.c b/arch/x86_64/kernel/kprobes.c index 6cb40d133b7c..9bef2c8dc12c 100644 --- a/arch/x86_64/kernel/kprobes.c +++ b/arch/x86_64/kernel/kprobes.c @@ -34,7 +34,6 @@ #include #include #include -#include #include #include #include @@ -266,6 +265,7 @@ static void __kprobes prepare_singlestep(struct kprobe *p, struct pt_regs *regs) regs->rip = (unsigned long)p->ainsn.insn; } +/* Called with kretprobe_lock held */ void __kprobes arch_prepare_kretprobe(struct kretprobe *rp, struct pt_regs *regs) { @@ -299,15 +299,12 @@ int __kprobes kprobe_handler(struct pt_regs *regs) /* Check we're not actually recursing */ if (kprobe_running()) { - /* We *are* holding lock here, so this is safe. - Disarm the probe we just hit, and ignore it. */ p = get_kprobe(addr); if (p) { if (kcb->kprobe_status == KPROBE_HIT_SS && *p->ainsn.insn == BREAKPOINT_INSTRUCTION) { regs->eflags &= ~TF_MASK; regs->eflags |= kcb->kprobe_saved_rflags; - unlock_kprobes(); goto no_kprobe; } else if (kcb->kprobe_status == KPROBE_HIT_SSDONE) { /* TODO: Provide re-entrancy from @@ -340,14 +337,11 @@ int __kprobes kprobe_handler(struct pt_regs *regs) goto ss_probe; } } - /* If it's not ours, can't be delete race, (we hold lock). */ goto no_kprobe; } - lock_kprobes(); p = get_kprobe(addr); if (!p) { - unlock_kprobes(); if (*addr != BREAKPOINT_INSTRUCTION) { /* * The breakpoint instruction was removed right @@ -406,9 +400,10 @@ int __kprobes trampoline_probe_handler(struct kprobe *p, struct pt_regs *regs) struct kretprobe_instance *ri = NULL; struct hlist_head *head; struct hlist_node *node, *tmp; - unsigned long orig_ret_address = 0; + unsigned long flags, orig_ret_address = 0; unsigned long trampoline_address =(unsigned long)&kretprobe_trampoline; + spin_lock_irqsave(&kretprobe_lock, flags); head = kretprobe_inst_table_head(current); /* @@ -448,7 +443,7 @@ int __kprobes trampoline_probe_handler(struct kprobe *p, struct pt_regs *regs) regs->rip = orig_ret_address; reset_current_kprobe(); - unlock_kprobes(); + spin_unlock_irqrestore(&kretprobe_lock, flags); preempt_enable_no_resched(); /* @@ -536,10 +531,6 @@ static void __kprobes resume_execution(struct kprobe *p, } } -/* - * Interrupts are disabled on entry as trap1 is an interrupt gate and they - * remain disabled thoroughout this function. And we hold kprobe lock. - */ int __kprobes post_kprobe_handler(struct pt_regs *regs) { struct kprobe *cur = kprobe_running(); @@ -560,8 +551,6 @@ int __kprobes post_kprobe_handler(struct pt_regs *regs) if (kcb->kprobe_status == KPROBE_REENTER) { restore_previous_kprobe(kcb); goto out; - } else { - unlock_kprobes(); } reset_current_kprobe(); out: @@ -578,7 +567,6 @@ out: return 1; } -/* Interrupts disabled, kprobe_lock held. */ int __kprobes kprobe_fault_handler(struct pt_regs *regs, int trapnr) { struct kprobe *cur = kprobe_running(); @@ -592,7 +580,6 @@ int __kprobes kprobe_fault_handler(struct pt_regs *regs, int trapnr) regs->eflags |= kcb->kprobe_old_rflags; reset_current_kprobe(); - unlock_kprobes(); preempt_enable_no_resched(); } return 0; @@ -607,7 +594,7 @@ int __kprobes kprobe_exceptions_notify(struct notifier_block *self, struct die_args *args = (struct die_args *)data; int ret = NOTIFY_DONE; - preempt_disable(); + rcu_read_lock(); switch (val) { case DIE_INT3: if (kprobe_handler(args->regs)) @@ -626,7 +613,7 @@ int __kprobes kprobe_exceptions_notify(struct notifier_block *self, default: break; } - preempt_enable(); + rcu_read_unlock(); return ret; } -- cgit v1.2.3 From d217d5450f11d8c907c0458d175b0dc999b4d06d Mon Sep 17 00:00:00 2001 From: Ananth N Mavinakayanahalli Date: Mon, 7 Nov 2005 01:00:14 -0800 Subject: [PATCH] Kprobes: preempt_disable/enable() simplification Reorganize the preempt_disable/enable calls to eliminate the extra preempt depth. Changes based on Paul McKenney's review suggestions for the kprobes RCU changeset. Signed-off-by: Ananth N Mavinakayanahalli Signed-off-by: Anil S Keshavamurthy Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- arch/i386/kernel/kprobes.c | 25 +++++++++++++++---------- arch/ia64/kernel/kprobes.c | 37 ++++++++++++++++++++----------------- arch/ppc64/kernel/kprobes.c | 25 +++++++++++++++---------- arch/sparc64/kernel/kprobes.c | 21 +++++++++++++-------- arch/x86_64/kernel/kprobes.c | 29 +++++++++++++++-------------- include/linux/kprobes.h | 2 +- kernel/kprobes.c | 2 +- 7 files changed, 80 insertions(+), 61 deletions(-) (limited to 'arch/ppc64') diff --git a/arch/i386/kernel/kprobes.c b/arch/i386/kernel/kprobes.c index ad469299267a..32b0c24ab9a6 100644 --- a/arch/i386/kernel/kprobes.c +++ b/arch/i386/kernel/kprobes.c @@ -153,7 +153,14 @@ static int __kprobes kprobe_handler(struct pt_regs *regs) int ret = 0; kprobe_opcode_t *addr = NULL; unsigned long *lp; - struct kprobe_ctlblk *kcb = get_kprobe_ctlblk(); + struct kprobe_ctlblk *kcb; + + /* + * We don't want to be preempted for the entire + * duration of kprobe processing + */ + preempt_disable(); + kcb = get_kprobe_ctlblk(); /* Check if the application is using LDT entry for its code segment and * calculate the address by reading the base address from the LDT entry. @@ -221,11 +228,6 @@ static int __kprobes kprobe_handler(struct pt_regs *regs) goto no_kprobe; } - /* - * This preempt_disable() matches the preempt_enable_no_resched() - * in post_kprobe_handler() - */ - preempt_disable(); set_current_kprobe(p, regs, kcb); kcb->kprobe_status = KPROBE_HIT_ACTIVE; @@ -239,6 +241,7 @@ ss_probe: return 1; no_kprobe: + preempt_enable_no_resched(); return ret; } @@ -310,8 +313,8 @@ int __kprobes trampoline_probe_handler(struct kprobe *p, struct pt_regs *regs) /* * By returning a non-zero value, we are telling - * kprobe_handler() that we have handled unlocking - * and re-enabling preemption + * kprobe_handler() that we don't want the post_handler + * to run (and have re-enabled preemption) */ return 1; } @@ -455,7 +458,6 @@ int __kprobes kprobe_exceptions_notify(struct notifier_block *self, struct die_args *args = (struct die_args *)data; int ret = NOTIFY_DONE; - rcu_read_lock(); switch (val) { case DIE_INT3: if (kprobe_handler(args->regs)) @@ -467,14 +469,16 @@ int __kprobes kprobe_exceptions_notify(struct notifier_block *self, break; case DIE_GPF: case DIE_PAGE_FAULT: + /* kprobe_running() needs smp_processor_id() */ + preempt_disable(); if (kprobe_running() && kprobe_fault_handler(args->regs, args->trapnr)) ret = NOTIFY_STOP; + preempt_enable(); break; default: break; } - rcu_read_unlock(); return ret; } @@ -537,6 +541,7 @@ int __kprobes longjmp_break_handler(struct kprobe *p, struct pt_regs *regs) *regs = kcb->jprobe_saved_regs; memcpy((kprobe_opcode_t *) stack_addr, kcb->jprobes_stack, MIN_STACK_SIZE(stack_addr)); + preempt_enable_no_resched(); return 1; } return 0; diff --git a/arch/ia64/kernel/kprobes.c b/arch/ia64/kernel/kprobes.c index fddbac32d44a..96736a119c91 100644 --- a/arch/ia64/kernel/kprobes.c +++ b/arch/ia64/kernel/kprobes.c @@ -389,11 +389,11 @@ int __kprobes trampoline_probe_handler(struct kprobe *p, struct pt_regs *regs) spin_unlock_irqrestore(&kretprobe_lock, flags); preempt_enable_no_resched(); - /* - * By returning a non-zero value, we are telling - * kprobe_handler() that we have handled unlocking - * and re-enabling preemption - */ + /* + * By returning a non-zero value, we are telling + * kprobe_handler() that we don't want the post_handler + * to run (and have re-enabled preemption) + */ return 1; } @@ -604,7 +604,14 @@ static int __kprobes pre_kprobes_handler(struct die_args *args) int ret = 0; struct pt_regs *regs = args->regs; kprobe_opcode_t *addr = (kprobe_opcode_t *)instruction_pointer(regs); - struct kprobe_ctlblk *kcb = get_kprobe_ctlblk(); + struct kprobe_ctlblk *kcb; + + /* + * We don't want to be preempted for the entire + * duration of kprobe processing + */ + preempt_disable(); + kcb = get_kprobe_ctlblk(); /* Handle recursion cases */ if (kprobe_running()) { @@ -659,11 +666,6 @@ static int __kprobes pre_kprobes_handler(struct die_args *args) goto no_kprobe; } - /* - * This preempt_disable() matches the preempt_enable_no_resched() - * in post_kprobes_handler() - */ - preempt_disable(); set_current_kprobe(p, kcb); kcb->kprobe_status = KPROBE_HIT_ACTIVE; @@ -681,6 +683,7 @@ ss_probe: return 1; no_kprobe: + preempt_enable_no_resched(); return ret; } @@ -716,9 +719,6 @@ static int __kprobes kprobes_fault_handler(struct pt_regs *regs, int trapnr) struct kprobe *cur = kprobe_running(); struct kprobe_ctlblk *kcb = get_kprobe_ctlblk(); - if (!cur) - return 0; - if (cur->fault_handler && cur->fault_handler(cur, regs, trapnr)) return 1; @@ -737,7 +737,6 @@ int __kprobes kprobe_exceptions_notify(struct notifier_block *self, struct die_args *args = (struct die_args *)data; int ret = NOTIFY_DONE; - rcu_read_lock(); switch(val) { case DIE_BREAK: if (pre_kprobes_handler(args)) @@ -748,12 +747,15 @@ int __kprobes kprobe_exceptions_notify(struct notifier_block *self, ret = NOTIFY_STOP; break; case DIE_PAGE_FAULT: - if (kprobes_fault_handler(args->regs, args->trapnr)) + /* kprobe_running() needs smp_processor_id() */ + preempt_disable(); + if (kprobe_running() && + kprobes_fault_handler(args->regs, args->trapnr)) ret = NOTIFY_STOP; + preempt_enable(); default: break; } - rcu_read_unlock(); return ret; } @@ -785,6 +787,7 @@ int __kprobes longjmp_break_handler(struct kprobe *p, struct pt_regs *regs) struct kprobe_ctlblk *kcb = get_kprobe_ctlblk(); *regs = kcb->jprobe_saved_regs; + preempt_enable_no_resched(); return 1; } diff --git a/arch/ppc64/kernel/kprobes.c b/arch/ppc64/kernel/kprobes.c index e0a25b35437f..511af54e6230 100644 --- a/arch/ppc64/kernel/kprobes.c +++ b/arch/ppc64/kernel/kprobes.c @@ -148,7 +148,14 @@ static inline int kprobe_handler(struct pt_regs *regs) struct kprobe *p; int ret = 0; unsigned int *addr = (unsigned int *)regs->nip; - struct kprobe_ctlblk *kcb = get_kprobe_ctlblk(); + struct kprobe_ctlblk *kcb; + + /* + * We don't want to be preempted for the entire + * duration of kprobe processing + */ + preempt_disable(); + kcb = get_kprobe_ctlblk(); /* Check we're not actually recursing */ if (kprobe_running()) { @@ -207,11 +214,6 @@ static inline int kprobe_handler(struct pt_regs *regs) goto no_kprobe; } - /* - * This preempt_disable() matches the preempt_enable_no_resched() - * in post_kprobe_handler(). - */ - preempt_disable(); kcb->kprobe_status = KPROBE_HIT_ACTIVE; set_current_kprobe(p, regs, kcb); if (p->pre_handler && p->pre_handler(p, regs)) @@ -224,6 +226,7 @@ ss_probe: return 1; no_kprobe: + preempt_enable_no_resched(); return ret; } @@ -296,8 +299,8 @@ int __kprobes trampoline_probe_handler(struct kprobe *p, struct pt_regs *regs) /* * By returning a non-zero value, we are telling - * kprobe_handler() that we have handled unlocking - * and re-enabling preemption. + * kprobe_handler() that we don't want the post_handler + * to run (and have re-enabled preemption) */ return 1; } @@ -385,7 +388,6 @@ int __kprobes kprobe_exceptions_notify(struct notifier_block *self, struct die_args *args = (struct die_args *)data; int ret = NOTIFY_DONE; - rcu_read_lock(); switch (val) { case DIE_BPT: if (kprobe_handler(args->regs)) @@ -396,14 +398,16 @@ int __kprobes kprobe_exceptions_notify(struct notifier_block *self, ret = NOTIFY_STOP; break; case DIE_PAGE_FAULT: + /* kprobe_running() needs smp_processor_id() */ + preempt_disable(); if (kprobe_running() && kprobe_fault_handler(args->regs, args->trapnr)) ret = NOTIFY_STOP; + preempt_enable(); break; default: break; } - rcu_read_unlock(); return ret; } @@ -440,6 +444,7 @@ int __kprobes longjmp_break_handler(struct kprobe *p, struct pt_regs *regs) * saved regs... */ memcpy(regs, &kcb->jprobe_saved_regs, sizeof(struct pt_regs)); + preempt_enable_no_resched(); return 1; } diff --git a/arch/sparc64/kernel/kprobes.c b/arch/sparc64/kernel/kprobes.c index 58a815e90373..96bd09b098f4 100644 --- a/arch/sparc64/kernel/kprobes.c +++ b/arch/sparc64/kernel/kprobes.c @@ -113,7 +113,14 @@ static int __kprobes kprobe_handler(struct pt_regs *regs) struct kprobe *p; void *addr = (void *) regs->tpc; int ret = 0; - struct kprobe_ctlblk *kcb = get_kprobe_ctlblk(); + struct kprobe_ctlblk *kcb; + + /* + * We don't want to be preempted for the entire + * duration of kprobe processing + */ + preempt_disable(); + kcb = get_kprobe_ctlblk(); if (kprobe_running()) { p = get_kprobe(addr); @@ -159,11 +166,6 @@ static int __kprobes kprobe_handler(struct pt_regs *regs) goto no_kprobe; } - /* - * This preempt_disable() matches the preempt_enable_no_resched() - * in post_kprobes_handler() - */ - preempt_disable(); set_current_kprobe(p, regs, kcb); kcb->kprobe_status = KPROBE_HIT_ACTIVE; if (p->pre_handler && p->pre_handler(p, regs)) @@ -175,6 +177,7 @@ ss_probe: return 1; no_kprobe: + preempt_enable_no_resched(); return ret; } @@ -321,7 +324,6 @@ int __kprobes kprobe_exceptions_notify(struct notifier_block *self, struct die_args *args = (struct die_args *)data; int ret = NOTIFY_DONE; - rcu_read_lock(); switch (val) { case DIE_DEBUG: if (kprobe_handler(args->regs)) @@ -333,14 +335,16 @@ int __kprobes kprobe_exceptions_notify(struct notifier_block *self, break; case DIE_GPF: case DIE_PAGE_FAULT: + /* kprobe_running() needs smp_processor_id() */ + preempt_disable(); if (kprobe_running() && kprobe_fault_handler(args->regs, args->trapnr)) ret = NOTIFY_STOP; + preempt_enable(); break; default: break; } - rcu_read_unlock(); return ret; } @@ -426,6 +430,7 @@ int __kprobes longjmp_break_handler(struct kprobe *p, struct pt_regs *regs) &(kcb->jprobe_saved_stack), sizeof(kcb->jprobe_saved_stack)); + preempt_enable_no_resched(); return 1; } return 0; diff --git a/arch/x86_64/kernel/kprobes.c b/arch/x86_64/kernel/kprobes.c index 9bef2c8dc12c..dddeb678b440 100644 --- a/arch/x86_64/kernel/kprobes.c +++ b/arch/x86_64/kernel/kprobes.c @@ -286,16 +286,19 @@ void __kprobes arch_prepare_kretprobe(struct kretprobe *rp, } } -/* - * Interrupts are disabled on entry as trap3 is an interrupt gate and they - * remain disabled thorough out this function. - */ int __kprobes kprobe_handler(struct pt_regs *regs) { struct kprobe *p; int ret = 0; kprobe_opcode_t *addr = (kprobe_opcode_t *)(regs->rip - sizeof(kprobe_opcode_t)); - struct kprobe_ctlblk *kcb = get_kprobe_ctlblk(); + struct kprobe_ctlblk *kcb; + + /* + * We don't want to be preempted for the entire + * duration of kprobe processing + */ + preempt_disable(); + kcb = get_kprobe_ctlblk(); /* Check we're not actually recursing */ if (kprobe_running()) { @@ -359,11 +362,6 @@ int __kprobes kprobe_handler(struct pt_regs *regs) goto no_kprobe; } - /* - * This preempt_disable() matches the preempt_enable_no_resched() - * in post_kprobe_handler() - */ - preempt_disable(); set_current_kprobe(p, regs, kcb); kcb->kprobe_status = KPROBE_HIT_ACTIVE; @@ -377,6 +375,7 @@ ss_probe: return 1; no_kprobe: + preempt_enable_no_resched(); return ret; } @@ -448,8 +447,8 @@ int __kprobes trampoline_probe_handler(struct kprobe *p, struct pt_regs *regs) /* * By returning a non-zero value, we are telling - * kprobe_handler() that we have handled unlocking - * and re-enabling preemption + * kprobe_handler() that we don't want the post_handler + * to run (and have re-enabled preemption) */ return 1; } @@ -594,7 +593,6 @@ int __kprobes kprobe_exceptions_notify(struct notifier_block *self, struct die_args *args = (struct die_args *)data; int ret = NOTIFY_DONE; - rcu_read_lock(); switch (val) { case DIE_INT3: if (kprobe_handler(args->regs)) @@ -606,14 +604,16 @@ int __kprobes kprobe_exceptions_notify(struct notifier_block *self, break; case DIE_GPF: case DIE_PAGE_FAULT: + /* kprobe_running() needs smp_processor_id() */ + preempt_disable(); if (kprobe_running() && kprobe_fault_handler(args->regs, args->trapnr)) ret = NOTIFY_STOP; + preempt_enable(); break; default: break; } - rcu_read_unlock(); return ret; } @@ -675,6 +675,7 @@ int __kprobes longjmp_break_handler(struct kprobe *p, struct pt_regs *regs) *regs = kcb->jprobe_saved_regs; memcpy((kprobe_opcode_t *) stack_addr, kcb->jprobes_stack, MIN_STACK_SIZE(stack_addr)); + preempt_enable_no_resched(); return 1; } return 0; diff --git a/include/linux/kprobes.h b/include/linux/kprobes.h index cff281cf70cf..e373c4a9de53 100644 --- a/include/linux/kprobes.h +++ b/include/linux/kprobes.h @@ -159,7 +159,7 @@ extern void show_registers(struct pt_regs *regs); extern kprobe_opcode_t *get_insn_slot(void); extern void free_insn_slot(kprobe_opcode_t *slot); -/* Get the kprobe at this addr (if any) - called under a rcu_read_lock() */ +/* Get the kprobe at this addr (if any) - called with preemption disabled */ struct kprobe *get_kprobe(void *addr); struct hlist_head * kretprobe_inst_table_head(struct task_struct *tsk); diff --git a/kernel/kprobes.c b/kernel/kprobes.c index cfef426e4cdc..5beda378cc75 100644 --- a/kernel/kprobes.c +++ b/kernel/kprobes.c @@ -167,7 +167,7 @@ static inline void reset_kprobe_instance(void) * This routine is called either: * - under the kprobe_lock spinlock - during kprobe_[un]register() * OR - * - under an rcu_read_lock() - from arch/xxx/kernel/kprobes.c + * - with preemption disabled - from arch/xxx/kernel/kprobes.c */ struct kprobe __kprobes *get_kprobe(void *addr) { -- cgit v1.2.3 From b2325fe1b7e5654fac9e9419423aa2c58a3dbd83 Mon Sep 17 00:00:00 2001 From: Jesper Juhl Date: Mon, 7 Nov 2005 01:01:35 -0800 Subject: [PATCH] kfree cleanup: arch This is the arch/ part of the big kfree cleanup patch. Remove pointless checks for NULL prior to calling kfree() in arch/. Signed-off-by: Jesper Juhl Acked-by: Grant Grundler Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- arch/arm/mach-integrator/impd1.c | 3 +-- arch/cris/arch-v32/drivers/cryptocop.c | 14 +++++++------- arch/ia64/kernel/perfmon.c | 2 +- arch/powerpc/platforms/pseries/reconfig.c | 6 ++---- arch/ppc/8xx_io/cs4218_tdm.c | 3 +-- arch/ppc/syslib/prom.c | 6 ++---- arch/ppc64/kernel/lparcfg.c | 4 +--- arch/ppc64/kernel/scanlog.c | 3 +-- arch/s390/mm/extmem.c | 8 ++++---- arch/sparc64/kernel/us2e_cpufreq.c | 7 ++----- arch/sparc64/kernel/us3_cpufreq.c | 7 ++----- arch/um/kernel/sigio_user.c | 2 +- 12 files changed, 25 insertions(+), 40 deletions(-) (limited to 'arch/ppc64') diff --git a/arch/arm/mach-integrator/impd1.c b/arch/arm/mach-integrator/impd1.c index a1b153d1626c..a4bafee77a06 100644 --- a/arch/arm/mach-integrator/impd1.c +++ b/arch/arm/mach-integrator/impd1.c @@ -420,8 +420,7 @@ static int impd1_probe(struct lm_device *dev) free_impd1: if (impd1 && impd1->base) iounmap(impd1->base); - if (impd1) - kfree(impd1); + kfree(impd1); release_lm: release_mem_region(dev->resource.start, SZ_4K); return ret; diff --git a/arch/cris/arch-v32/drivers/cryptocop.c b/arch/cris/arch-v32/drivers/cryptocop.c index ca72076c630a..501fa52d8d3a 100644 --- a/arch/cris/arch-v32/drivers/cryptocop.c +++ b/arch/cris/arch-v32/drivers/cryptocop.c @@ -277,7 +277,7 @@ struct file_operations cryptocop_fops = { static void free_cdesc(struct cryptocop_dma_desc *cdesc) { DEBUG(printk("free_cdesc: cdesc 0x%p, from_pool=%d\n", cdesc, cdesc->from_pool)); - if (cdesc->free_buf) kfree(cdesc->free_buf); + kfree(cdesc->free_buf); if (cdesc->from_pool) { unsigned long int flags; @@ -2950,15 +2950,15 @@ static int cryptocop_ioctl_process(struct inode *inode, struct file *filp, unsig put_page(outpages[i]); } - if (digest_result) kfree(digest_result); - if (inpages) kfree(inpages); - if (outpages) kfree(outpages); + kfree(digest_result); + kfree(inpages); + kfree(outpages); if (cop){ - if (cop->tfrm_op.indata) kfree(cop->tfrm_op.indata); - if (cop->tfrm_op.outdata) kfree(cop->tfrm_op.outdata); + kfree(cop->tfrm_op.indata); + kfree(cop->tfrm_op.outdata); kfree(cop); } - if (jc) kfree(jc); + kfree(jc); DEBUG(print_lock_status()); diff --git a/arch/ia64/kernel/perfmon.c b/arch/ia64/kernel/perfmon.c index f7dfc107cb7b..410d4804fa6e 100644 --- a/arch/ia64/kernel/perfmon.c +++ b/arch/ia64/kernel/perfmon.c @@ -4940,7 +4940,7 @@ abort_locked: if (call_made && PFM_CMD_RW_ARG(cmd) && copy_to_user(arg, args_k, base_sz*count)) ret = -EFAULT; error_args: - if (args_k) kfree(args_k); + kfree(args_k); DPRINT(("cmd=%s ret=%ld\n", PFM_CMD_NAME(cmd), ret)); diff --git a/arch/powerpc/platforms/pseries/reconfig.c b/arch/powerpc/platforms/pseries/reconfig.c index 58c61219d08e..d7d400339458 100644 --- a/arch/powerpc/platforms/pseries/reconfig.c +++ b/arch/powerpc/platforms/pseries/reconfig.c @@ -286,10 +286,8 @@ static struct property *new_property(const char *name, const int length, return new; cleanup: - if (new->name) - kfree(new->name); - if (new->value) - kfree(new->value); + kfree(new->name); + kfree(new->value); kfree(new); return NULL; } diff --git a/arch/ppc/8xx_io/cs4218_tdm.c b/arch/ppc/8xx_io/cs4218_tdm.c index 532caa388dc2..49eb2a7e65c0 100644 --- a/arch/ppc/8xx_io/cs4218_tdm.c +++ b/arch/ppc/8xx_io/cs4218_tdm.c @@ -1013,8 +1013,7 @@ static void CS_IrqCleanup(void) */ cpm_free_handler(CPMVEC_SMC2); - if (beep_buf) - kfree(beep_buf); + kfree(beep_buf); kd_mksound = orig_mksound; } #endif /* MODULE */ diff --git a/arch/ppc/syslib/prom.c b/arch/ppc/syslib/prom.c index 278da6ee62ea..1b9aa0d6a924 100644 --- a/arch/ppc/syslib/prom.c +++ b/arch/ppc/syslib/prom.c @@ -1335,10 +1335,8 @@ release_OF_resource(struct device_node* node, int index) if (!res) return -ENODEV; - if (res->name) { - kfree(res->name); - res->name = NULL; - } + kfree(res->name); + res->name = NULL; release_resource(res); kfree(res); diff --git a/arch/ppc64/kernel/lparcfg.c b/arch/ppc64/kernel/lparcfg.c index e86155770bbc..3e7b2f28ec83 100644 --- a/arch/ppc64/kernel/lparcfg.c +++ b/arch/ppc64/kernel/lparcfg.c @@ -599,9 +599,7 @@ int __init lparcfg_init(void) void __exit lparcfg_cleanup(void) { if (proc_ppc64_lparcfg) { - if (proc_ppc64_lparcfg->data) { - kfree(proc_ppc64_lparcfg->data); - } + kfree(proc_ppc64_lparcfg->data); remove_proc_entry("lparcfg", proc_ppc64_lparcfg->parent); } } diff --git a/arch/ppc64/kernel/scanlog.c b/arch/ppc64/kernel/scanlog.c index 215bf8900304..2edc947f7c44 100644 --- a/arch/ppc64/kernel/scanlog.c +++ b/arch/ppc64/kernel/scanlog.c @@ -225,8 +225,7 @@ int __init scanlog_init(void) void __exit scanlog_cleanup(void) { if (proc_ppc64_scan_log_dump) { - if (proc_ppc64_scan_log_dump->data) - kfree(proc_ppc64_scan_log_dump->data); + kfree(proc_ppc64_scan_log_dump->data); remove_proc_entry("scan-log-dump", proc_ppc64_scan_log_dump->parent); } } diff --git a/arch/s390/mm/extmem.c b/arch/s390/mm/extmem.c index c5348108ca3c..506a33b51e4f 100644 --- a/arch/s390/mm/extmem.c +++ b/arch/s390/mm/extmem.c @@ -234,8 +234,8 @@ query_segment_type (struct dcss_segment *seg) rc = 0; out_free: - if (qin) kfree(qin); - if (qout) kfree(qout); + kfree(qin); + kfree(qout); return rc; } @@ -394,7 +394,7 @@ __segment_load (char *name, int do_nonshared, unsigned long *addr, unsigned long segtype_string[seg->vm_segtype]); goto out; out_free: - kfree (seg); + kfree(seg); out: return rc; } @@ -505,7 +505,7 @@ segment_modify_shared (char *name, int do_nonshared) list_del(&seg->list); dcss_diag(DCSS_PURGESEG, seg->dcss_name, &dummy, &dummy); - kfree (seg); + kfree(seg); out_unlock: spin_unlock(&dcss_lock); return rc; diff --git a/arch/sparc64/kernel/us2e_cpufreq.c b/arch/sparc64/kernel/us2e_cpufreq.c index 686e526bec04..b35dc8dc995a 100644 --- a/arch/sparc64/kernel/us2e_cpufreq.c +++ b/arch/sparc64/kernel/us2e_cpufreq.c @@ -388,10 +388,8 @@ err_out: kfree(driver); cpufreq_us2e_driver = NULL; } - if (us2e_freq_table) { - kfree(us2e_freq_table); - us2e_freq_table = NULL; - } + kfree(us2e_freq_table); + us2e_freq_table = NULL; return ret; } @@ -402,7 +400,6 @@ static void __exit us2e_freq_exit(void) { if (cpufreq_us2e_driver) { cpufreq_unregister_driver(cpufreq_us2e_driver); - kfree(cpufreq_us2e_driver); cpufreq_us2e_driver = NULL; kfree(us2e_freq_table); diff --git a/arch/sparc64/kernel/us3_cpufreq.c b/arch/sparc64/kernel/us3_cpufreq.c index 0340041f6143..6d1f9a3c464f 100644 --- a/arch/sparc64/kernel/us3_cpufreq.c +++ b/arch/sparc64/kernel/us3_cpufreq.c @@ -249,10 +249,8 @@ err_out: kfree(driver); cpufreq_us3_driver = NULL; } - if (us3_freq_table) { - kfree(us3_freq_table); - us3_freq_table = NULL; - } + kfree(us3_freq_table); + us3_freq_table = NULL; return ret; } @@ -263,7 +261,6 @@ static void __exit us3_freq_exit(void) { if (cpufreq_us3_driver) { cpufreq_unregister_driver(cpufreq_us3_driver); - kfree(cpufreq_us3_driver); cpufreq_us3_driver = NULL; kfree(us3_freq_table); diff --git a/arch/um/kernel/sigio_user.c b/arch/um/kernel/sigio_user.c index 3fbfd956bfe7..48b1f644b9a6 100644 --- a/arch/um/kernel/sigio_user.c +++ b/arch/um/kernel/sigio_user.c @@ -224,7 +224,7 @@ static int need_poll(int n) next_poll.used = n; return(0); } - if(next_poll.poll != NULL) kfree(next_poll.poll); + kfree(next_poll.poll); next_poll.poll = um_kmalloc_atomic(n * sizeof(struct pollfd)); if(next_poll.poll == NULL){ printk("need_poll : failed to allocate new pollfds\n"); -- cgit v1.2.3 From 4350147a816b9c5b40fa59e4fa23f17490630b79 Mon Sep 17 00:00:00 2001 From: Benjamin Herrenschmidt Date: Mon, 7 Nov 2005 14:27:33 +1100 Subject: [PATCH] ppc64: SMU based macs cpufreq support CPU freq support using 970FX powertune facility for iMac G5 and SMU based single CPU desktop. Signed-off-by: Benjamin Herrenschmidt Signed-off-by: Paul Mackerras --- arch/powerpc/Kconfig | 8 + arch/powerpc/kernel/misc_64.S | 70 +++ arch/powerpc/platforms/powermac/Makefile | 3 +- arch/powerpc/platforms/powermac/cpufreq.c | 726 -------------------------- arch/powerpc/platforms/powermac/cpufreq_32.c | 727 +++++++++++++++++++++++++++ arch/powerpc/platforms/powermac/cpufreq_64.c | 323 ++++++++++++ arch/powerpc/platforms/powermac/setup.c | 13 - arch/ppc64/Kconfig | 10 + arch/ppc64/kernel/misc.S | 72 ++- drivers/macintosh/smu.c | 12 + include/asm-powerpc/reg.h | 9 +- include/asm-powerpc/smu.h | 60 ++- 12 files changed, 1290 insertions(+), 743 deletions(-) delete mode 100644 arch/powerpc/platforms/powermac/cpufreq.c create mode 100644 arch/powerpc/platforms/powermac/cpufreq_32.c create mode 100644 arch/powerpc/platforms/powermac/cpufreq_64.c (limited to 'arch/ppc64') diff --git a/arch/powerpc/Kconfig b/arch/powerpc/Kconfig index 6ffae2d2b3fa..3ac9195dab35 100644 --- a/arch/powerpc/Kconfig +++ b/arch/powerpc/Kconfig @@ -404,6 +404,14 @@ config CPU_FREQ_PMAC this currently includes some models of iBook & Titanium PowerBook. +config CPU_FREQ_PMAC64 + bool "Support for some Apple G5s" + depends on CPU_FREQ && PMAC_SMU && PPC64 + select CPU_FREQ_TABLE + help + This adds support for frequency switching on Apple iMac G5, + and some of the more recent desktop G5 machines as well. + config PPC601_SYNC_FIX bool "Workarounds for PPC601 bugs" depends on 6xx && (PPC_PREP || PPC_PMAC) diff --git a/arch/powerpc/kernel/misc_64.S b/arch/powerpc/kernel/misc_64.S index b3e95ff0dba0..ae1433da09b2 100644 --- a/arch/powerpc/kernel/misc_64.S +++ b/arch/powerpc/kernel/misc_64.S @@ -603,6 +603,76 @@ _GLOBAL(real_writeb) blr #endif /* defined(CONFIG_PPC_PMAC) || defined(CONFIG_PPC_MAPLE) */ +/* + * SCOM access functions for 970 (FX only for now) + * + * unsigned long scom970_read(unsigned int address); + * void scom970_write(unsigned int address, unsigned long value); + * + * The address passed in is the 24 bits register address. This code + * is 970 specific and will not check the status bits, so you should + * know what you are doing. + */ +_GLOBAL(scom970_read) + /* interrupts off */ + mfmsr r4 + ori r0,r4,MSR_EE + xori r0,r0,MSR_EE + mtmsrd r0,1 + + /* rotate 24 bits SCOM address 8 bits left and mask out it's low 8 bits + * (including parity). On current CPUs they must be 0'd, + * and finally or in RW bit + */ + rlwinm r3,r3,8,0,15 + ori r3,r3,0x8000 + + /* do the actual scom read */ + sync + mtspr SPRN_SCOMC,r3 + isync + mfspr r3,SPRN_SCOMD + isync + mfspr r0,SPRN_SCOMC + isync + + /* XXX: fixup result on some buggy 970's (ouch ! we lost a bit, bah + * that's the best we can do). Not implemented yet as we don't use + * the scom on any of the bogus CPUs yet, but may have to be done + * ultimately + */ + + /* restore interrupts */ + mtmsrd r4,1 + blr + + +_GLOBAL(scom970_write) + /* interrupts off */ + mfmsr r5 + ori r0,r5,MSR_EE + xori r0,r0,MSR_EE + mtmsrd r0,1 + + /* rotate 24 bits SCOM address 8 bits left and mask out it's low 8 bits + * (including parity). On current CPUs they must be 0'd. + */ + + rlwinm r3,r3,8,0,15 + + sync + mtspr SPRN_SCOMD,r4 /* write data */ + isync + mtspr SPRN_SCOMC,r3 /* write command */ + isync + mfspr 3,SPRN_SCOMC + isync + + /* restore interrupts */ + mtmsrd r5,1 + blr + + /* * Create a kernel thread * kernel_thread(fn, arg, flags) diff --git a/arch/powerpc/platforms/powermac/Makefile b/arch/powerpc/platforms/powermac/Makefile index 4369676f1d54..c9df44fcf571 100644 --- a/arch/powerpc/platforms/powermac/Makefile +++ b/arch/powerpc/platforms/powermac/Makefile @@ -1,7 +1,8 @@ obj-y += pic.o setup.o time.o feature.o pci.o \ sleep.o low_i2c.o cache.o obj-$(CONFIG_PMAC_BACKLIGHT) += backlight.o -obj-$(CONFIG_CPU_FREQ_PMAC) += cpufreq.o +obj-$(CONFIG_CPU_FREQ_PMAC) += cpufreq_32.o +obj-$(CONFIG_CPU_FREQ_PMAC64) += cpufreq_64.o obj-$(CONFIG_NVRAM) += nvram.o # ppc64 pmac doesn't define CONFIG_NVRAM but needs nvram stuff obj-$(CONFIG_PPC64) += nvram.o diff --git a/arch/powerpc/platforms/powermac/cpufreq.c b/arch/powerpc/platforms/powermac/cpufreq.c deleted file mode 100644 index c47f8b69725c..000000000000 --- a/arch/powerpc/platforms/powermac/cpufreq.c +++ /dev/null @@ -1,726 +0,0 @@ -/* - * arch/ppc/platforms/pmac_cpufreq.c - * - * Copyright (C) 2002 - 2005 Benjamin Herrenschmidt - * Copyright (C) 2004 John Steele Scott - * - * This program is free software; you can redistribute it and/or modify - * it under the terms of the GNU General Public License version 2 as - * published by the Free Software Foundation. - * - * TODO: Need a big cleanup here. Basically, we need to have different - * cpufreq_driver structures for the different type of HW instead of the - * current mess. We also need to better deal with the detection of the - * type of machine. - * - */ - -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include - -/* WARNING !!! This will cause calibrate_delay() to be called, - * but this is an __init function ! So you MUST go edit - * init/main.c to make it non-init before enabling DEBUG_FREQ - */ -#undef DEBUG_FREQ - -/* - * There is a problem with the core cpufreq code on SMP kernels, - * it won't recalculate the Bogomips properly - */ -#ifdef CONFIG_SMP -#warning "WARNING, CPUFREQ not recommended on SMP kernels" -#endif - -extern void low_choose_7447a_dfs(int dfs); -extern void low_choose_750fx_pll(int pll); -extern void low_sleep_handler(void); - -/* - * Currently, PowerMac cpufreq supports only high & low frequencies - * that are set by the firmware - */ -static unsigned int low_freq; -static unsigned int hi_freq; -static unsigned int cur_freq; -static unsigned int sleep_freq; - -/* - * Different models uses different mecanisms to switch the frequency - */ -static int (*set_speed_proc)(int low_speed); -static unsigned int (*get_speed_proc)(void); - -/* - * Some definitions used by the various speedprocs - */ -static u32 voltage_gpio; -static u32 frequency_gpio; -static u32 slew_done_gpio; -static int no_schedule; -static int has_cpu_l2lve; -static int is_pmu_based; - -/* There are only two frequency states for each processor. Values - * are in kHz for the time being. - */ -#define CPUFREQ_HIGH 0 -#define CPUFREQ_LOW 1 - -static struct cpufreq_frequency_table pmac_cpu_freqs[] = { - {CPUFREQ_HIGH, 0}, - {CPUFREQ_LOW, 0}, - {0, CPUFREQ_TABLE_END}, -}; - -static struct freq_attr* pmac_cpu_freqs_attr[] = { - &cpufreq_freq_attr_scaling_available_freqs, - NULL, -}; - -static inline void local_delay(unsigned long ms) -{ - if (no_schedule) - mdelay(ms); - else - msleep(ms); -} - -#ifdef DEBUG_FREQ -static inline void debug_calc_bogomips(void) -{ - /* This will cause a recalc of bogomips and display the - * result. We backup/restore the value to avoid affecting the - * core cpufreq framework's own calculation. - */ - extern void calibrate_delay(void); - - unsigned long save_lpj = loops_per_jiffy; - calibrate_delay(); - loops_per_jiffy = save_lpj; -} -#endif /* DEBUG_FREQ */ - -/* Switch CPU speed under 750FX CPU control - */ -static int cpu_750fx_cpu_speed(int low_speed) -{ - u32 hid2; - - if (low_speed == 0) { - /* ramping up, set voltage first */ - pmac_call_feature(PMAC_FTR_WRITE_GPIO, NULL, voltage_gpio, 0x05); - /* Make sure we sleep for at least 1ms */ - local_delay(10); - - /* tweak L2 for high voltage */ - if (has_cpu_l2lve) { - hid2 = mfspr(SPRN_HID2); - hid2 &= ~0x2000; - mtspr(SPRN_HID2, hid2); - } - } -#ifdef CONFIG_6xx - low_choose_750fx_pll(low_speed); -#endif - if (low_speed == 1) { - /* tweak L2 for low voltage */ - if (has_cpu_l2lve) { - hid2 = mfspr(SPRN_HID2); - hid2 |= 0x2000; - mtspr(SPRN_HID2, hid2); - } - - /* ramping down, set voltage last */ - pmac_call_feature(PMAC_FTR_WRITE_GPIO, NULL, voltage_gpio, 0x04); - local_delay(10); - } - - return 0; -} - -static unsigned int cpu_750fx_get_cpu_speed(void) -{ - if (mfspr(SPRN_HID1) & HID1_PS) - return low_freq; - else - return hi_freq; -} - -/* Switch CPU speed using DFS */ -static int dfs_set_cpu_speed(int low_speed) -{ - if (low_speed == 0) { - /* ramping up, set voltage first */ - pmac_call_feature(PMAC_FTR_WRITE_GPIO, NULL, voltage_gpio, 0x05); - /* Make sure we sleep for at least 1ms */ - local_delay(1); - } - - /* set frequency */ -#ifdef CONFIG_6xx - low_choose_7447a_dfs(low_speed); -#endif - udelay(100); - - if (low_speed == 1) { - /* ramping down, set voltage last */ - pmac_call_feature(PMAC_FTR_WRITE_GPIO, NULL, voltage_gpio, 0x04); - local_delay(1); - } - - return 0; -} - -static unsigned int dfs_get_cpu_speed(void) -{ - if (mfspr(SPRN_HID1) & HID1_DFS) - return low_freq; - else - return hi_freq; -} - - -/* Switch CPU speed using slewing GPIOs - */ -static int gpios_set_cpu_speed(int low_speed) -{ - int gpio, timeout = 0; - - /* If ramping up, set voltage first */ - if (low_speed == 0) { - pmac_call_feature(PMAC_FTR_WRITE_GPIO, NULL, voltage_gpio, 0x05); - /* Delay is way too big but it's ok, we schedule */ - local_delay(10); - } - - /* Set frequency */ - gpio = pmac_call_feature(PMAC_FTR_READ_GPIO, NULL, frequency_gpio, 0); - if (low_speed == ((gpio & 0x01) == 0)) - goto skip; - - pmac_call_feature(PMAC_FTR_WRITE_GPIO, NULL, frequency_gpio, - low_speed ? 0x04 : 0x05); - udelay(200); - do { - if (++timeout > 100) - break; - local_delay(1); - gpio = pmac_call_feature(PMAC_FTR_READ_GPIO, NULL, slew_done_gpio, 0); - } while((gpio & 0x02) == 0); - skip: - /* If ramping down, set voltage last */ - if (low_speed == 1) { - pmac_call_feature(PMAC_FTR_WRITE_GPIO, NULL, voltage_gpio, 0x04); - /* Delay is way too big but it's ok, we schedule */ - local_delay(10); - } - -#ifdef DEBUG_FREQ - debug_calc_bogomips(); -#endif - - return 0; -} - -/* Switch CPU speed under PMU control - */ -static int pmu_set_cpu_speed(int low_speed) -{ - struct adb_request req; - unsigned long save_l2cr; - unsigned long save_l3cr; - unsigned int pic_prio; - unsigned long flags; - - preempt_disable(); - -#ifdef DEBUG_FREQ - printk(KERN_DEBUG "HID1, before: %x\n", mfspr(SPRN_HID1)); -#endif - pmu_suspend(); - - /* Disable all interrupt sources on openpic */ - pic_prio = mpic_cpu_get_priority(); - mpic_cpu_set_priority(0xf); - - /* Make sure the decrementer won't interrupt us */ - asm volatile("mtdec %0" : : "r" (0x7fffffff)); - /* Make sure any pending DEC interrupt occuring while we did - * the above didn't re-enable the DEC */ - mb(); - asm volatile("mtdec %0" : : "r" (0x7fffffff)); - - /* We can now disable MSR_EE */ - local_irq_save(flags); - - /* Giveup the FPU & vec */ - enable_kernel_fp(); - -#ifdef CONFIG_ALTIVEC - if (cpu_has_feature(CPU_FTR_ALTIVEC)) - enable_kernel_altivec(); -#endif /* CONFIG_ALTIVEC */ - - /* Save & disable L2 and L3 caches */ - save_l3cr = _get_L3CR(); /* (returns -1 if not available) */ - save_l2cr = _get_L2CR(); /* (returns -1 if not available) */ - - /* Send the new speed command. My assumption is that this command - * will cause PLL_CFG[0..3] to be changed next time CPU goes to sleep - */ - pmu_request(&req, NULL, 6, PMU_CPU_SPEED, 'W', 'O', 'O', 'F', low_speed); - while (!req.complete) - pmu_poll(); - - /* Prepare the northbridge for the speed transition */ - pmac_call_feature(PMAC_FTR_SLEEP_STATE,NULL,1,1); - - /* Call low level code to backup CPU state and recover from - * hardware reset - */ - low_sleep_handler(); - - /* Restore the northbridge */ - pmac_call_feature(PMAC_FTR_SLEEP_STATE,NULL,1,0); - - /* Restore L2 cache */ - if (save_l2cr != 0xffffffff && (save_l2cr & L2CR_L2E) != 0) - _set_L2CR(save_l2cr); - /* Restore L3 cache */ - if (save_l3cr != 0xffffffff && (save_l3cr & L3CR_L3E) != 0) - _set_L3CR(save_l3cr); - - /* Restore userland MMU context */ - set_context(current->active_mm->context, current->active_mm->pgd); - -#ifdef DEBUG_FREQ - printk(KERN_DEBUG "HID1, after: %x\n", mfspr(SPRN_HID1)); -#endif - - /* Restore low level PMU operations */ - pmu_unlock(); - - /* Restore decrementer */ - wakeup_decrementer(); - - /* Restore interrupts */ - mpic_cpu_set_priority(pic_prio); - - /* Let interrupts flow again ... */ - local_irq_restore(flags); - -#ifdef DEBUG_FREQ - debug_calc_bogomips(); -#endif - - pmu_resume(); - - preempt_enable(); - - return 0; -} - -static int do_set_cpu_speed(int speed_mode, int notify) -{ - struct cpufreq_freqs freqs; - unsigned long l3cr; - static unsigned long prev_l3cr; - - freqs.old = cur_freq; - freqs.new = (speed_mode == CPUFREQ_HIGH) ? hi_freq : low_freq; - freqs.cpu = smp_processor_id(); - - if (freqs.old == freqs.new) - return 0; - - if (notify) - cpufreq_notify_transition(&freqs, CPUFREQ_PRECHANGE); - if (speed_mode == CPUFREQ_LOW && - cpu_has_feature(CPU_FTR_L3CR)) { - l3cr = _get_L3CR(); - if (l3cr & L3CR_L3E) { - prev_l3cr = l3cr; - _set_L3CR(0); - } - } - set_speed_proc(speed_mode == CPUFREQ_LOW); - if (speed_mode == CPUFREQ_HIGH && - cpu_has_feature(CPU_FTR_L3CR)) { - l3cr = _get_L3CR(); - if ((prev_l3cr & L3CR_L3E) && l3cr != prev_l3cr) - _set_L3CR(prev_l3cr); - } - if (notify) - cpufreq_notify_transition(&freqs, CPUFREQ_POSTCHANGE); - cur_freq = (speed_mode == CPUFREQ_HIGH) ? hi_freq : low_freq; - - return 0; -} - -static unsigned int pmac_cpufreq_get_speed(unsigned int cpu) -{ - return cur_freq; -} - -static int pmac_cpufreq_verify(struct cpufreq_policy *policy) -{ - return cpufreq_frequency_table_verify(policy, pmac_cpu_freqs); -} - -static int pmac_cpufreq_target( struct cpufreq_policy *policy, - unsigned int target_freq, - unsigned int relation) -{ - unsigned int newstate = 0; - - if (cpufreq_frequency_table_target(policy, pmac_cpu_freqs, - target_freq, relation, &newstate)) - return -EINVAL; - - return do_set_cpu_speed(newstate, 1); -} - -unsigned int pmac_get_one_cpufreq(int i) -{ - /* Supports only one CPU for now */ - return (i == 0) ? cur_freq : 0; -} - -static int pmac_cpufreq_cpu_init(struct cpufreq_policy *policy) -{ - if (policy->cpu != 0) - return -ENODEV; - - policy->governor = CPUFREQ_DEFAULT_GOVERNOR; - policy->cpuinfo.transition_latency = CPUFREQ_ETERNAL; - policy->cur = cur_freq; - - cpufreq_frequency_table_get_attr(pmac_cpu_freqs, policy->cpu); - return cpufreq_frequency_table_cpuinfo(policy, pmac_cpu_freqs); -} - -static u32 read_gpio(struct device_node *np) -{ - u32 *reg = (u32 *)get_property(np, "reg", NULL); - u32 offset; - - if (reg == NULL) - return 0; - /* That works for all keylargos but shall be fixed properly - * some day... The problem is that it seems we can't rely - * on the "reg" property of the GPIO nodes, they are either - * relative to the base of KeyLargo or to the base of the - * GPIO space, and the device-tree doesn't help. - */ - offset = *reg; - if (offset < KEYLARGO_GPIO_LEVELS0) - offset += KEYLARGO_GPIO_LEVELS0; - return offset; -} - -static int pmac_cpufreq_suspend(struct cpufreq_policy *policy, pm_message_t pmsg) -{ - /* Ok, this could be made a bit smarter, but let's be robust for now. We - * always force a speed change to high speed before sleep, to make sure - * we have appropriate voltage and/or bus speed for the wakeup process, - * and to make sure our loops_per_jiffies are "good enough", that is will - * not cause too short delays if we sleep in low speed and wake in high - * speed.. - */ - no_schedule = 1; - sleep_freq = cur_freq; - if (cur_freq == low_freq && !is_pmu_based) - do_set_cpu_speed(CPUFREQ_HIGH, 0); - return 0; -} - -static int pmac_cpufreq_resume(struct cpufreq_policy *policy) -{ - /* If we resume, first check if we have a get() function */ - if (get_speed_proc) - cur_freq = get_speed_proc(); - else - cur_freq = 0; - - /* We don't, hrm... we don't really know our speed here, best - * is that we force a switch to whatever it was, which is - * probably high speed due to our suspend() routine - */ - do_set_cpu_speed(sleep_freq == low_freq ? - CPUFREQ_LOW : CPUFREQ_HIGH, 0); - - no_schedule = 0; - return 0; -} - -static struct cpufreq_driver pmac_cpufreq_driver = { - .verify = pmac_cpufreq_verify, - .target = pmac_cpufreq_target, - .get = pmac_cpufreq_get_speed, - .init = pmac_cpufreq_cpu_init, - .suspend = pmac_cpufreq_suspend, - .resume = pmac_cpufreq_resume, - .flags = CPUFREQ_PM_NO_WARN, - .attr = pmac_cpu_freqs_attr, - .name = "powermac", - .owner = THIS_MODULE, -}; - - -static int pmac_cpufreq_init_MacRISC3(struct device_node *cpunode) -{ - struct device_node *volt_gpio_np = of_find_node_by_name(NULL, - "voltage-gpio"); - struct device_node *freq_gpio_np = of_find_node_by_name(NULL, - "frequency-gpio"); - struct device_node *slew_done_gpio_np = of_find_node_by_name(NULL, - "slewing-done"); - u32 *value; - - /* - * Check to see if it's GPIO driven or PMU only - * - * The way we extract the GPIO address is slightly hackish, but it - * works well enough for now. We need to abstract the whole GPIO - * stuff sooner or later anyway - */ - - if (volt_gpio_np) - voltage_gpio = read_gpio(volt_gpio_np); - if (freq_gpio_np) - frequency_gpio = read_gpio(freq_gpio_np); - if (slew_done_gpio_np) - slew_done_gpio = read_gpio(slew_done_gpio_np); - - /* If we use the frequency GPIOs, calculate the min/max speeds based - * on the bus frequencies - */ - if (frequency_gpio && slew_done_gpio) { - int lenp, rc; - u32 *freqs, *ratio; - - freqs = (u32 *)get_property(cpunode, "bus-frequencies", &lenp); - lenp /= sizeof(u32); - if (freqs == NULL || lenp != 2) { - printk(KERN_ERR "cpufreq: bus-frequencies incorrect or missing\n"); - return 1; - } - ratio = (u32 *)get_property(cpunode, "processor-to-bus-ratio*2", NULL); - if (ratio == NULL) { - printk(KERN_ERR "cpufreq: processor-to-bus-ratio*2 missing\n"); - return 1; - } - - /* Get the min/max bus frequencies */ - low_freq = min(freqs[0], freqs[1]); - hi_freq = max(freqs[0], freqs[1]); - - /* Grrrr.. It _seems_ that the device-tree is lying on the low bus - * frequency, it claims it to be around 84Mhz on some models while - * it appears to be approx. 101Mhz on all. Let's hack around here... - * fortunately, we don't need to be too precise - */ - if (low_freq < 98000000) - low_freq = 101000000; - - /* Convert those to CPU core clocks */ - low_freq = (low_freq * (*ratio)) / 2000; - hi_freq = (hi_freq * (*ratio)) / 2000; - - /* Now we get the frequencies, we read the GPIO to see what is out current - * speed - */ - rc = pmac_call_feature(PMAC_FTR_READ_GPIO, NULL, frequency_gpio, 0); - cur_freq = (rc & 0x01) ? hi_freq : low_freq; - - set_speed_proc = gpios_set_cpu_speed; - return 1; - } - - /* If we use the PMU, look for the min & max frequencies in the - * device-tree - */ - value = (u32 *)get_property(cpunode, "min-clock-frequency", NULL); - if (!value) - return 1; - low_freq = (*value) / 1000; - /* The PowerBook G4 12" (PowerBook6,1) has an error in the device-tree - * here */ - if (low_freq < 100000) - low_freq *= 10; - - value = (u32 *)get_property(cpunode, "max-clock-frequency", NULL); - if (!value) - return 1; - hi_freq = (*value) / 1000; - set_speed_proc = pmu_set_cpu_speed; - is_pmu_based = 1; - - return 0; -} - -static int pmac_cpufreq_init_7447A(struct device_node *cpunode) -{ - struct device_node *volt_gpio_np; - - if (get_property(cpunode, "dynamic-power-step", NULL) == NULL) - return 1; - - volt_gpio_np = of_find_node_by_name(NULL, "cpu-vcore-select"); - if (volt_gpio_np) - voltage_gpio = read_gpio(volt_gpio_np); - if (!voltage_gpio){ - printk(KERN_ERR "cpufreq: missing cpu-vcore-select gpio\n"); - return 1; - } - - /* OF only reports the high frequency */ - hi_freq = cur_freq; - low_freq = cur_freq/2; - - /* Read actual frequency from CPU */ - cur_freq = dfs_get_cpu_speed(); - set_speed_proc = dfs_set_cpu_speed; - get_speed_proc = dfs_get_cpu_speed; - - return 0; -} - -static int pmac_cpufreq_init_750FX(struct device_node *cpunode) -{ - struct device_node *volt_gpio_np; - u32 pvr, *value; - - if (get_property(cpunode, "dynamic-power-step", NULL) == NULL) - return 1; - - hi_freq = cur_freq; - value = (u32 *)get_property(cpunode, "reduced-clock-frequency", NULL); - if (!value) - return 1; - low_freq = (*value) / 1000; - - volt_gpio_np = of_find_node_by_name(NULL, "cpu-vcore-select"); - if (volt_gpio_np) - voltage_gpio = read_gpio(volt_gpio_np); - - pvr = mfspr(SPRN_PVR); - has_cpu_l2lve = !((pvr & 0xf00) == 0x100); - - set_speed_proc = cpu_750fx_cpu_speed; - get_speed_proc = cpu_750fx_get_cpu_speed; - cur_freq = cpu_750fx_get_cpu_speed(); - - return 0; -} - -/* Currently, we support the following machines: - * - * - Titanium PowerBook 1Ghz (PMU based, 667Mhz & 1Ghz) - * - Titanium PowerBook 800 (PMU based, 667Mhz & 800Mhz) - * - Titanium PowerBook 400 (PMU based, 300Mhz & 400Mhz) - * - Titanium PowerBook 500 (PMU based, 300Mhz & 500Mhz) - * - iBook2 500/600 (PMU based, 400Mhz & 500/600Mhz) - * - iBook2 700 (CPU based, 400Mhz & 700Mhz, support low voltage) - * - Recent MacRISC3 laptops - * - All new machines with 7447A CPUs - */ -static int __init pmac_cpufreq_setup(void) -{ - struct device_node *cpunode; - u32 *value; - - if (strstr(cmd_line, "nocpufreq")) - return 0; - - /* Assume only one CPU */ - cpunode = find_type_devices("cpu"); - if (!cpunode) - goto out; - - /* Get current cpu clock freq */ - value = (u32 *)get_property(cpunode, "clock-frequency", NULL); - if (!value) - goto out; - cur_freq = (*value) / 1000; - - /* Check for 7447A based MacRISC3 */ - if (machine_is_compatible("MacRISC3") && - get_property(cpunode, "dynamic-power-step", NULL) && - PVR_VER(mfspr(SPRN_PVR)) == 0x8003) { - pmac_cpufreq_init_7447A(cpunode); - /* Check for other MacRISC3 machines */ - } else if (machine_is_compatible("PowerBook3,4") || - machine_is_compatible("PowerBook3,5") || - machine_is_compatible("MacRISC3")) { - pmac_cpufreq_init_MacRISC3(cpunode); - /* Else check for iBook2 500/600 */ - } else if (machine_is_compatible("PowerBook4,1")) { - hi_freq = cur_freq; - low_freq = 400000; - set_speed_proc = pmu_set_cpu_speed; - is_pmu_based = 1; - } - /* Else check for TiPb 550 */ - else if (machine_is_compatible("PowerBook3,3") && cur_freq == 550000) { - hi_freq = cur_freq; - low_freq = 500000; - set_speed_proc = pmu_set_cpu_speed; - is_pmu_based = 1; - } - /* Else check for TiPb 400 & 500 */ - else if (machine_is_compatible("PowerBook3,2")) { - /* We only know about the 400 MHz and the 500Mhz model - * they both have 300 MHz as low frequency - */ - if (cur_freq < 350000 || cur_freq > 550000) - goto out; - hi_freq = cur_freq; - low_freq = 300000; - set_speed_proc = pmu_set_cpu_speed; - is_pmu_based = 1; - } - /* Else check for 750FX */ - else if (PVR_VER(mfspr(SPRN_PVR)) == 0x7000) - pmac_cpufreq_init_750FX(cpunode); -out: - if (set_speed_proc == NULL) - return -ENODEV; - - pmac_cpu_freqs[CPUFREQ_LOW].frequency = low_freq; - pmac_cpu_freqs[CPUFREQ_HIGH].frequency = hi_freq; - - printk(KERN_INFO "Registering PowerMac CPU frequency driver\n"); - printk(KERN_INFO "Low: %d Mhz, High: %d Mhz, Boot: %d Mhz\n", - low_freq/1000, hi_freq/1000, cur_freq/1000); - - return cpufreq_register_driver(&pmac_cpufreq_driver); -} - -module_init(pmac_cpufreq_setup); - diff --git a/arch/powerpc/platforms/powermac/cpufreq_32.c b/arch/powerpc/platforms/powermac/cpufreq_32.c new file mode 100644 index 000000000000..7960a7bfca0a --- /dev/null +++ b/arch/powerpc/platforms/powermac/cpufreq_32.c @@ -0,0 +1,727 @@ +/* + * arch/ppc/platforms/pmac_cpufreq.c + * + * Copyright (C) 2002 - 2005 Benjamin Herrenschmidt + * Copyright (C) 2004 John Steele Scott + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License version 2 as + * published by the Free Software Foundation. + * + * TODO: Need a big cleanup here. Basically, we need to have different + * cpufreq_driver structures for the different type of HW instead of the + * current mess. We also need to better deal with the detection of the + * type of machine. + * + */ + +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include + +/* WARNING !!! This will cause calibrate_delay() to be called, + * but this is an __init function ! So you MUST go edit + * init/main.c to make it non-init before enabling DEBUG_FREQ + */ +#undef DEBUG_FREQ + +/* + * There is a problem with the core cpufreq code on SMP kernels, + * it won't recalculate the Bogomips properly + */ +#ifdef CONFIG_SMP +#warning "WARNING, CPUFREQ not recommended on SMP kernels" +#endif + +extern void low_choose_7447a_dfs(int dfs); +extern void low_choose_750fx_pll(int pll); +extern void low_sleep_handler(void); + +/* + * Currently, PowerMac cpufreq supports only high & low frequencies + * that are set by the firmware + */ +static unsigned int low_freq; +static unsigned int hi_freq; +static unsigned int cur_freq; +static unsigned int sleep_freq; + +/* + * Different models uses different mecanisms to switch the frequency + */ +static int (*set_speed_proc)(int low_speed); +static unsigned int (*get_speed_proc)(void); + +/* + * Some definitions used by the various speedprocs + */ +static u32 voltage_gpio; +static u32 frequency_gpio; +static u32 slew_done_gpio; +static int no_schedule; +static int has_cpu_l2lve; +static int is_pmu_based; + +/* There are only two frequency states for each processor. Values + * are in kHz for the time being. + */ +#define CPUFREQ_HIGH 0 +#define CPUFREQ_LOW 1 + +static struct cpufreq_frequency_table pmac_cpu_freqs[] = { + {CPUFREQ_HIGH, 0}, + {CPUFREQ_LOW, 0}, + {0, CPUFREQ_TABLE_END}, +}; + +static struct freq_attr* pmac_cpu_freqs_attr[] = { + &cpufreq_freq_attr_scaling_available_freqs, + NULL, +}; + +static inline void local_delay(unsigned long ms) +{ + if (no_schedule) + mdelay(ms); + else + msleep(ms); +} + +#ifdef DEBUG_FREQ +static inline void debug_calc_bogomips(void) +{ + /* This will cause a recalc of bogomips and display the + * result. We backup/restore the value to avoid affecting the + * core cpufreq framework's own calculation. + */ + extern void calibrate_delay(void); + + unsigned long save_lpj = loops_per_jiffy; + calibrate_delay(); + loops_per_jiffy = save_lpj; +} +#endif /* DEBUG_FREQ */ + +/* Switch CPU speed under 750FX CPU control + */ +static int cpu_750fx_cpu_speed(int low_speed) +{ + u32 hid2; + + if (low_speed == 0) { + /* ramping up, set voltage first */ + pmac_call_feature(PMAC_FTR_WRITE_GPIO, NULL, voltage_gpio, 0x05); + /* Make sure we sleep for at least 1ms */ + local_delay(10); + + /* tweak L2 for high voltage */ + if (has_cpu_l2lve) { + hid2 = mfspr(SPRN_HID2); + hid2 &= ~0x2000; + mtspr(SPRN_HID2, hid2); + } + } +#ifdef CONFIG_6xx + low_choose_750fx_pll(low_speed); +#endif + if (low_speed == 1) { + /* tweak L2 for low voltage */ + if (has_cpu_l2lve) { + hid2 = mfspr(SPRN_HID2); + hid2 |= 0x2000; + mtspr(SPRN_HID2, hid2); + } + + /* ramping down, set voltage last */ + pmac_call_feature(PMAC_FTR_WRITE_GPIO, NULL, voltage_gpio, 0x04); + local_delay(10); + } + + return 0; +} + +static unsigned int cpu_750fx_get_cpu_speed(void) +{ + if (mfspr(SPRN_HID1) & HID1_PS) + return low_freq; + else + return hi_freq; +} + +/* Switch CPU speed using DFS */ +static int dfs_set_cpu_speed(int low_speed) +{ + if (low_speed == 0) { + /* ramping up, set voltage first */ + pmac_call_feature(PMAC_FTR_WRITE_GPIO, NULL, voltage_gpio, 0x05); + /* Make sure we sleep for at least 1ms */ + local_delay(1); + } + + /* set frequency */ +#ifdef CONFIG_6xx + low_choose_7447a_dfs(low_speed); +#endif + udelay(100); + + if (low_speed == 1) { + /* ramping down, set voltage last */ + pmac_call_feature(PMAC_FTR_WRITE_GPIO, NULL, voltage_gpio, 0x04); + local_delay(1); + } + + return 0; +} + +static unsigned int dfs_get_cpu_speed(void) +{ + if (mfspr(SPRN_HID1) & HID1_DFS) + return low_freq; + else + return hi_freq; +} + + +/* Switch CPU speed using slewing GPIOs + */ +static int gpios_set_cpu_speed(int low_speed) +{ + int gpio, timeout = 0; + + /* If ramping up, set voltage first */ + if (low_speed == 0) { + pmac_call_feature(PMAC_FTR_WRITE_GPIO, NULL, voltage_gpio, 0x05); + /* Delay is way too big but it's ok, we schedule */ + local_delay(10); + } + + /* Set frequency */ + gpio = pmac_call_feature(PMAC_FTR_READ_GPIO, NULL, frequency_gpio, 0); + if (low_speed == ((gpio & 0x01) == 0)) + goto skip; + + pmac_call_feature(PMAC_FTR_WRITE_GPIO, NULL, frequency_gpio, + low_speed ? 0x04 : 0x05); + udelay(200); + do { + if (++timeout > 100) + break; + local_delay(1); + gpio = pmac_call_feature(PMAC_FTR_READ_GPIO, NULL, slew_done_gpio, 0); + } while((gpio & 0x02) == 0); + skip: + /* If ramping down, set voltage last */ + if (low_speed == 1) { + pmac_call_feature(PMAC_FTR_WRITE_GPIO, NULL, voltage_gpio, 0x04); + /* Delay is way too big but it's ok, we schedule */ + local_delay(10); + } + +#ifdef DEBUG_FREQ + debug_calc_bogomips(); +#endif + + return 0; +} + +/* Switch CPU speed under PMU control + */ +static int pmu_set_cpu_speed(int low_speed) +{ + struct adb_request req; + unsigned long save_l2cr; + unsigned long save_l3cr; + unsigned int pic_prio; + unsigned long flags; + + preempt_disable(); + +#ifdef DEBUG_FREQ + printk(KERN_DEBUG "HID1, before: %x\n", mfspr(SPRN_HID1)); +#endif + pmu_suspend(); + + /* Disable all interrupt sources on openpic */ + pic_prio = mpic_cpu_get_priority(); + mpic_cpu_set_priority(0xf); + + /* Make sure the decrementer won't interrupt us */ + asm volatile("mtdec %0" : : "r" (0x7fffffff)); + /* Make sure any pending DEC interrupt occuring while we did + * the above didn't re-enable the DEC */ + mb(); + asm volatile("mtdec %0" : : "r" (0x7fffffff)); + + /* We can now disable MSR_EE */ + local_irq_save(flags); + + /* Giveup the FPU & vec */ + enable_kernel_fp(); + +#ifdef CONFIG_ALTIVEC + if (cpu_has_feature(CPU_FTR_ALTIVEC)) + enable_kernel_altivec(); +#endif /* CONFIG_ALTIVEC */ + + /* Save & disable L2 and L3 caches */ + save_l3cr = _get_L3CR(); /* (returns -1 if not available) */ + save_l2cr = _get_L2CR(); /* (returns -1 if not available) */ + + /* Send the new speed command. My assumption is that this command + * will cause PLL_CFG[0..3] to be changed next time CPU goes to sleep + */ + pmu_request(&req, NULL, 6, PMU_CPU_SPEED, 'W', 'O', 'O', 'F', low_speed); + while (!req.complete) + pmu_poll(); + + /* Prepare the northbridge for the speed transition */ + pmac_call_feature(PMAC_FTR_SLEEP_STATE,NULL,1,1); + + /* Call low level code to backup CPU state and recover from + * hardware reset + */ + low_sleep_handler(); + + /* Restore the northbridge */ + pmac_call_feature(PMAC_FTR_SLEEP_STATE,NULL,1,0); + + /* Restore L2 cache */ + if (save_l2cr != 0xffffffff && (save_l2cr & L2CR_L2E) != 0) + _set_L2CR(save_l2cr); + /* Restore L3 cache */ + if (save_l3cr != 0xffffffff && (save_l3cr & L3CR_L3E) != 0) + _set_L3CR(save_l3cr); + + /* Restore userland MMU context */ + set_context(current->active_mm->context, current->active_mm->pgd); + +#ifdef DEBUG_FREQ + printk(KERN_DEBUG "HID1, after: %x\n", mfspr(SPRN_HID1)); +#endif + + /* Restore low level PMU operations */ + pmu_unlock(); + + /* Restore decrementer */ + wakeup_decrementer(); + + /* Restore interrupts */ + mpic_cpu_set_priority(pic_prio); + + /* Let interrupts flow again ... */ + local_irq_restore(flags); + +#ifdef DEBUG_FREQ + debug_calc_bogomips(); +#endif + + pmu_resume(); + + preempt_enable(); + + return 0; +} + +static int do_set_cpu_speed(int speed_mode, int notify) +{ + struct cpufreq_freqs freqs; + unsigned long l3cr; + static unsigned long prev_l3cr; + + freqs.old = cur_freq; + freqs.new = (speed_mode == CPUFREQ_HIGH) ? hi_freq : low_freq; + freqs.cpu = smp_processor_id(); + + if (freqs.old == freqs.new) + return 0; + + if (notify) + cpufreq_notify_transition(&freqs, CPUFREQ_PRECHANGE); + if (speed_mode == CPUFREQ_LOW && + cpu_has_feature(CPU_FTR_L3CR)) { + l3cr = _get_L3CR(); + if (l3cr & L3CR_L3E) { + prev_l3cr = l3cr; + _set_L3CR(0); + } + } + set_speed_proc(speed_mode == CPUFREQ_LOW); + if (speed_mode == CPUFREQ_HIGH && + cpu_has_feature(CPU_FTR_L3CR)) { + l3cr = _get_L3CR(); + if ((prev_l3cr & L3CR_L3E) && l3cr != prev_l3cr) + _set_L3CR(prev_l3cr); + } + if (notify) + cpufreq_notify_transition(&freqs, CPUFREQ_POSTCHANGE); + cur_freq = (speed_mode == CPUFREQ_HIGH) ? hi_freq : low_freq; + + return 0; +} + +static unsigned int pmac_cpufreq_get_speed(unsigned int cpu) +{ + return cur_freq; +} + +static int pmac_cpufreq_verify(struct cpufreq_policy *policy) +{ + return cpufreq_frequency_table_verify(policy, pmac_cpu_freqs); +} + +static int pmac_cpufreq_target( struct cpufreq_policy *policy, + unsigned int target_freq, + unsigned int relation) +{ + unsigned int newstate = 0; + int rc; + + if (cpufreq_frequency_table_target(policy, pmac_cpu_freqs, + target_freq, relation, &newstate)) + return -EINVAL; + + rc = do_set_cpu_speed(newstate, 1); + + ppc_proc_freq = cur_freq * 1000ul; + return rc; +} + +static int pmac_cpufreq_cpu_init(struct cpufreq_policy *policy) +{ + if (policy->cpu != 0) + return -ENODEV; + + policy->governor = CPUFREQ_DEFAULT_GOVERNOR; + policy->cpuinfo.transition_latency = CPUFREQ_ETERNAL; + policy->cur = cur_freq; + + cpufreq_frequency_table_get_attr(pmac_cpu_freqs, policy->cpu); + return cpufreq_frequency_table_cpuinfo(policy, pmac_cpu_freqs); +} + +static u32 read_gpio(struct device_node *np) +{ + u32 *reg = (u32 *)get_property(np, "reg", NULL); + u32 offset; + + if (reg == NULL) + return 0; + /* That works for all keylargos but shall be fixed properly + * some day... The problem is that it seems we can't rely + * on the "reg" property of the GPIO nodes, they are either + * relative to the base of KeyLargo or to the base of the + * GPIO space, and the device-tree doesn't help. + */ + offset = *reg; + if (offset < KEYLARGO_GPIO_LEVELS0) + offset += KEYLARGO_GPIO_LEVELS0; + return offset; +} + +static int pmac_cpufreq_suspend(struct cpufreq_policy *policy, pm_message_t pmsg) +{ + /* Ok, this could be made a bit smarter, but let's be robust for now. We + * always force a speed change to high speed before sleep, to make sure + * we have appropriate voltage and/or bus speed for the wakeup process, + * and to make sure our loops_per_jiffies are "good enough", that is will + * not cause too short delays if we sleep in low speed and wake in high + * speed.. + */ + no_schedule = 1; + sleep_freq = cur_freq; + if (cur_freq == low_freq && !is_pmu_based) + do_set_cpu_speed(CPUFREQ_HIGH, 0); + return 0; +} + +static int pmac_cpufreq_resume(struct cpufreq_policy *policy) +{ + /* If we resume, first check if we have a get() function */ + if (get_speed_proc) + cur_freq = get_speed_proc(); + else) + cur_freq = 0; + + /* We don't, hrm... we don't really know our speed here, best + * is that we force a switch to whatever it was, which is + * probably high speed due to our suspend() routine + */ + do_set_cpu_speed(sleep_freq == low_freq ? + CPUFREQ_LOW : CPUFREQ_HIGH, 0); + + ppc_proc_freq = cur_freq * 1000ul; + + no_schedule = 0; + return 0; +} + +static struct cpufreq_driver pmac_cpufreq_driver = { + .verify = pmac_cpufreq_verify, + .target = pmac_cpufreq_target, + .get = pmac_cpufreq_get_speed, + .init = pmac_cpufreq_cpu_init, + .suspend = pmac_cpufreq_suspend, + .resume = pmac_cpufreq_resume, + .flags = CPUFREQ_PM_NO_WARN, + .attr = pmac_cpu_freqs_attr, + .name = "powermac", + .owner = THIS_MODULE, +}; + + +static int pmac_cpufreq_init_MacRISC3(struct device_node *cpunode) +{ + struct device_node *volt_gpio_np = of_find_node_by_name(NULL, + "voltage-gpio"); + struct device_node *freq_gpio_np = of_find_node_by_name(NULL, + "frequency-gpio"); + struct device_node *slew_done_gpio_np = of_find_node_by_name(NULL, + "slewing-done"); + u32 *value; + + /* + * Check to see if it's GPIO driven or PMU only + * + * The way we extract the GPIO address is slightly hackish, but it + * works well enough for now. We need to abstract the whole GPIO + * stuff sooner or later anyway + */ + + if (volt_gpio_np) + voltage_gpio = read_gpio(volt_gpio_np); + if (freq_gpio_np) + frequency_gpio = read_gpio(freq_gpio_np); + if (slew_done_gpio_np) + slew_done_gpio = read_gpio(slew_done_gpio_np); + + /* If we use the frequency GPIOs, calculate the min/max speeds based + * on the bus frequencies + */ + if (frequency_gpio && slew_done_gpio) { + int lenp, rc; + u32 *freqs, *ratio; + + freqs = (u32 *)get_property(cpunode, "bus-frequencies", &lenp); + lenp /= sizeof(u32); + if (freqs == NULL || lenp != 2) { + printk(KERN_ERR "cpufreq: bus-frequencies incorrect or missing\n"); + return 1; + } + ratio = (u32 *)get_property(cpunode, "processor-to-bus-ratio*2", NULL); + if (ratio == NULL) { + printk(KERN_ERR "cpufreq: processor-to-bus-ratio*2 missing\n"); + return 1; + } + + /* Get the min/max bus frequencies */ + low_freq = min(freqs[0], freqs[1]); + hi_freq = max(freqs[0], freqs[1]); + + /* Grrrr.. It _seems_ that the device-tree is lying on the low bus + * frequency, it claims it to be around 84Mhz on some models while + * it appears to be approx. 101Mhz on all. Let's hack around here... + * fortunately, we don't need to be too precise + */ + if (low_freq < 98000000) + low_freq = 101000000; + + /* Convert those to CPU core clocks */ + low_freq = (low_freq * (*ratio)) / 2000; + hi_freq = (hi_freq * (*ratio)) / 2000; + + /* Now we get the frequencies, we read the GPIO to see what is out current + * speed + */ + rc = pmac_call_feature(PMAC_FTR_READ_GPIO, NULL, frequency_gpio, 0); + cur_freq = (rc & 0x01) ? hi_freq : low_freq; + + set_speed_proc = gpios_set_cpu_speed; + return 1; + } + + /* If we use the PMU, look for the min & max frequencies in the + * device-tree + */ + value = (u32 *)get_property(cpunode, "min-clock-frequency", NULL); + if (!value) + return 1; + low_freq = (*value) / 1000; + /* The PowerBook G4 12" (PowerBook6,1) has an error in the device-tree + * here */ + if (low_freq < 100000) + low_freq *= 10; + + value = (u32 *)get_property(cpunode, "max-clock-frequency", NULL); + if (!value) + return 1; + hi_freq = (*value) / 1000; + set_speed_proc = pmu_set_cpu_speed; + is_pmu_based = 1; + + return 0; +} + +static int pmac_cpufreq_init_7447A(struct device_node *cpunode) +{ + struct device_node *volt_gpio_np; + + if (get_property(cpunode, "dynamic-power-step", NULL) == NULL) + return 1; + + volt_gpio_np = of_find_node_by_name(NULL, "cpu-vcore-select"); + if (volt_gpio_np) + voltage_gpio = read_gpio(volt_gpio_np); + if (!voltage_gpio){ + printk(KERN_ERR "cpufreq: missing cpu-vcore-select gpio\n"); + return 1; + } + + /* OF only reports the high frequency */ + hi_freq = cur_freq; + low_freq = cur_freq/2; + + /* Read actual frequency from CPU */ + cur_freq = dfs_get_cpu_speed(); + set_speed_proc = dfs_set_cpu_speed; + get_speed_proc = dfs_get_cpu_speed; + + return 0; +} + +static int pmac_cpufreq_init_750FX(struct device_node *cpunode) +{ + struct device_node *volt_gpio_np; + u32 pvr, *value; + + if (get_property(cpunode, "dynamic-power-step", NULL) == NULL) + return 1; + + hi_freq = cur_freq; + value = (u32 *)get_property(cpunode, "reduced-clock-frequency", NULL); + if (!value) + return 1; + low_freq = (*value) / 1000; + + volt_gpio_np = of_find_node_by_name(NULL, "cpu-vcore-select"); + if (volt_gpio_np) + voltage_gpio = read_gpio(volt_gpio_np); + + pvr = mfspr(SPRN_PVR); + has_cpu_l2lve = !((pvr & 0xf00) == 0x100); + + set_speed_proc = cpu_750fx_cpu_speed; + get_speed_proc = cpu_750fx_get_cpu_speed; + cur_freq = cpu_750fx_get_cpu_speed(); + + return 0; +} + +/* Currently, we support the following machines: + * + * - Titanium PowerBook 1Ghz (PMU based, 667Mhz & 1Ghz) + * - Titanium PowerBook 800 (PMU based, 667Mhz & 800Mhz) + * - Titanium PowerBook 400 (PMU based, 300Mhz & 400Mhz) + * - Titanium PowerBook 500 (PMU based, 300Mhz & 500Mhz) + * - iBook2 500/600 (PMU based, 400Mhz & 500/600Mhz) + * - iBook2 700 (CPU based, 400Mhz & 700Mhz, support low voltage) + * - Recent MacRISC3 laptops + * - All new machines with 7447A CPUs + */ +static int __init pmac_cpufreq_setup(void) +{ + struct device_node *cpunode; + u32 *value; + + if (strstr(cmd_line, "nocpufreq")) + return 0; + + /* Assume only one CPU */ + cpunode = find_type_devices("cpu"); + if (!cpunode) + goto out; + + /* Get current cpu clock freq */ + value = (u32 *)get_property(cpunode, "clock-frequency", NULL); + if (!value) + goto out; + cur_freq = (*value) / 1000; + + /* Check for 7447A based MacRISC3 */ + if (machine_is_compatible("MacRISC3") && + get_property(cpunode, "dynamic-power-step", NULL) && + PVR_VER(mfspr(SPRN_PVR)) == 0x8003) { + pmac_cpufreq_init_7447A(cpunode); + /* Check for other MacRISC3 machines */ + } else if (machine_is_compatible("PowerBook3,4") || + machine_is_compatible("PowerBook3,5") || + machine_is_compatible("MacRISC3")) { + pmac_cpufreq_init_MacRISC3(cpunode); + /* Else check for iBook2 500/600 */ + } else if (machine_is_compatible("PowerBook4,1")) { + hi_freq = cur_freq; + low_freq = 400000; + set_speed_proc = pmu_set_cpu_speed; + is_pmu_based = 1; + } + /* Else check for TiPb 550 */ + else if (machine_is_compatible("PowerBook3,3") && cur_freq == 550000) { + hi_freq = cur_freq; + low_freq = 500000; + set_speed_proc = pmu_set_cpu_speed; + is_pmu_based = 1; + } + /* Else check for TiPb 400 & 500 */ + else if (machine_is_compatible("PowerBook3,2")) { + /* We only know about the 400 MHz and the 500Mhz model + * they both have 300 MHz as low frequency + */ + if (cur_freq < 350000 || cur_freq > 550000) + goto out; + hi_freq = cur_freq; + low_freq = 300000; + set_speed_proc = pmu_set_cpu_speed; + is_pmu_based = 1; + } + /* Else check for 750FX */ + else if (PVR_VER(mfspr(SPRN_PVR)) == 0x7000) + pmac_cpufreq_init_750FX(cpunode); +out: + if (set_speed_proc == NULL) + return -ENODEV; + + pmac_cpu_freqs[CPUFREQ_LOW].frequency = low_freq; + pmac_cpu_freqs[CPUFREQ_HIGH].frequency = hi_freq; + ppc_proc_freq = cur_freq * 1000ul; + + printk(KERN_INFO "Registering PowerMac CPU frequency driver\n"); + printk(KERN_INFO "Low: %d Mhz, High: %d Mhz, Boot: %d Mhz\n", + low_freq/1000, hi_freq/1000, cur_freq/1000); + + return cpufreq_register_driver(&pmac_cpufreq_driver); +} + +module_init(pmac_cpufreq_setup); + diff --git a/arch/powerpc/platforms/powermac/cpufreq_64.c b/arch/powerpc/platforms/powermac/cpufreq_64.c new file mode 100644 index 000000000000..39150342c6f1 --- /dev/null +++ b/arch/powerpc/platforms/powermac/cpufreq_64.c @@ -0,0 +1,323 @@ +/* + * Copyright (C) 2002 - 2005 Benjamin Herrenschmidt + * and Markus Demleitner + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License version 2 as + * published by the Free Software Foundation. + * + * This driver adds basic cpufreq support for SMU & 970FX based G5 Macs, + * that is iMac G5 and latest single CPU desktop. + */ + +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include + +#undef DEBUG + +#ifdef DEBUG +#define DBG(fmt...) printk(fmt) +#else +#define DBG(fmt...) +#endif + +/* see 970FX user manual */ + +#define SCOM_PCR 0x0aa001 /* PCR scom addr */ + +#define PCR_HILO_SELECT 0x80000000U /* 1 = PCR, 0 = PCRH */ +#define PCR_SPEED_FULL 0x00000000U /* 1:1 speed value */ +#define PCR_SPEED_HALF 0x00020000U /* 1:2 speed value */ +#define PCR_SPEED_QUARTER 0x00040000U /* 1:4 speed value */ +#define PCR_SPEED_MASK 0x000e0000U /* speed mask */ +#define PCR_SPEED_SHIFT 17 +#define PCR_FREQ_REQ_VALID 0x00010000U /* freq request valid */ +#define PCR_VOLT_REQ_VALID 0x00008000U /* volt request valid */ +#define PCR_TARGET_TIME_MASK 0x00006000U /* target time */ +#define PCR_STATLAT_MASK 0x00001f00U /* STATLAT value */ +#define PCR_SNOOPLAT_MASK 0x000000f0U /* SNOOPLAT value */ +#define PCR_SNOOPACC_MASK 0x0000000fU /* SNOOPACC value */ + +#define SCOM_PSR 0x408001 /* PSR scom addr */ +/* warning: PSR is a 64 bits register */ +#define PSR_CMD_RECEIVED 0x2000000000000000U /* command received */ +#define PSR_CMD_COMPLETED 0x1000000000000000U /* command completed */ +#define PSR_CUR_SPEED_MASK 0x0300000000000000U /* current speed */ +#define PSR_CUR_SPEED_SHIFT (56) + +/* + * The G5 only supports two frequencies (Quarter speed is not supported) + */ +#define CPUFREQ_HIGH 0 +#define CPUFREQ_LOW 1 + +static struct cpufreq_frequency_table g5_cpu_freqs[] = { + {CPUFREQ_HIGH, 0}, + {CPUFREQ_LOW, 0}, + {0, CPUFREQ_TABLE_END}, +}; + +static struct freq_attr* g5_cpu_freqs_attr[] = { + &cpufreq_freq_attr_scaling_available_freqs, + NULL, +}; + +/* Power mode data is an array of the 32 bits PCR values to use for + * the various frequencies, retreived from the device-tree + */ +static u32 *g5_pmode_data; +static int g5_pmode_max; +static int g5_pmode_cur; + +static DECLARE_MUTEX(g5_switch_mutex); + + +static struct smu_sdbp_fvt *g5_fvt_table; /* table of op. points */ +static int g5_fvt_count; /* number of op. points */ +static int g5_fvt_cur; /* current op. point */ + +/* ----------------- real hardware interface */ + +static void g5_switch_volt(int speed_mode) +{ + struct smu_simple_cmd cmd; + + DECLARE_COMPLETION(comp); + smu_queue_simple(&cmd, SMU_CMD_POWER_COMMAND, 8, smu_done_complete, + &comp, 'V', 'S', 'L', 'E', 'W', + 0xff, g5_fvt_cur+1, speed_mode); + wait_for_completion(&comp); +} + +static int g5_switch_freq(int speed_mode) +{ + struct cpufreq_freqs freqs; + int to; + + if (g5_pmode_cur == speed_mode) + return 0; + + down(&g5_switch_mutex); + + freqs.old = g5_cpu_freqs[g5_pmode_cur].frequency; + freqs.new = g5_cpu_freqs[speed_mode].frequency; + freqs.cpu = 0; + + cpufreq_notify_transition(&freqs, CPUFREQ_PRECHANGE); + + /* If frequency is going up, first ramp up the voltage */ + if (speed_mode < g5_pmode_cur) + g5_switch_volt(speed_mode); + + /* Clear PCR high */ + scom970_write(SCOM_PCR, 0); + /* Clear PCR low */ + scom970_write(SCOM_PCR, PCR_HILO_SELECT | 0); + /* Set PCR low */ + scom970_write(SCOM_PCR, PCR_HILO_SELECT | + g5_pmode_data[speed_mode]); + + /* Wait for completion */ + for (to = 0; to < 10; to++) { + unsigned long psr = scom970_read(SCOM_PSR); + + if ((psr & PSR_CMD_RECEIVED) == 0 && + (((psr >> PSR_CUR_SPEED_SHIFT) ^ + (g5_pmode_data[speed_mode] >> PCR_SPEED_SHIFT)) & 0x3) + == 0) + break; + if (psr & PSR_CMD_COMPLETED) + break; + udelay(100); + } + + /* If frequency is going down, last ramp the voltage */ + if (speed_mode > g5_pmode_cur) + g5_switch_volt(speed_mode); + + g5_pmode_cur = speed_mode; + ppc_proc_freq = g5_cpu_freqs[speed_mode].frequency * 1000ul; + + cpufreq_notify_transition(&freqs, CPUFREQ_POSTCHANGE); + + up(&g5_switch_mutex); + + return 0; +} + +static int g5_query_freq(void) +{ + unsigned long psr = scom970_read(SCOM_PSR); + int i; + + for (i = 0; i <= g5_pmode_max; i++) + if ((((psr >> PSR_CUR_SPEED_SHIFT) ^ + (g5_pmode_data[i] >> PCR_SPEED_SHIFT)) & 0x3) == 0) + break; + return i; +} + +/* ----------------- cpufreq bookkeeping */ + +static int g5_cpufreq_verify(struct cpufreq_policy *policy) +{ + return cpufreq_frequency_table_verify(policy, g5_cpu_freqs); +} + +static int g5_cpufreq_target(struct cpufreq_policy *policy, + unsigned int target_freq, unsigned int relation) +{ + unsigned int newstate = 0; + + if (cpufreq_frequency_table_target(policy, g5_cpu_freqs, + target_freq, relation, &newstate)) + return -EINVAL; + + return g5_switch_freq(newstate); +} + +static unsigned int g5_cpufreq_get_speed(unsigned int cpu) +{ + return g5_cpu_freqs[g5_pmode_cur].frequency; +} + +static int g5_cpufreq_cpu_init(struct cpufreq_policy *policy) +{ + if (policy->cpu != 0) + return -ENODEV; + + policy->governor = CPUFREQ_DEFAULT_GOVERNOR; + policy->cpuinfo.transition_latency = CPUFREQ_ETERNAL; + policy->cur = g5_cpu_freqs[g5_query_freq()].frequency; + cpufreq_frequency_table_get_attr(g5_cpu_freqs, policy->cpu); + + return cpufreq_frequency_table_cpuinfo(policy, + g5_cpu_freqs); +} + + +static struct cpufreq_driver g5_cpufreq_driver = { + .name = "powermac", + .owner = THIS_MODULE, + .flags = CPUFREQ_CONST_LOOPS, + .init = g5_cpufreq_cpu_init, + .verify = g5_cpufreq_verify, + .target = g5_cpufreq_target, + .get = g5_cpufreq_get_speed, + .attr = g5_cpu_freqs_attr, +}; + + +static int __init g5_cpufreq_init(void) +{ + struct device_node *cpunode; + unsigned int psize, ssize; + struct smu_sdbp_header *shdr; + unsigned long max_freq; + u32 *valp; + int rc = -ENODEV; + + /* Look for CPU and SMU nodes */ + cpunode = of_find_node_by_type(NULL, "cpu"); + if (!cpunode) { + DBG("No CPU node !\n"); + return -ENODEV; + } + + /* Check 970FX for now */ + valp = (u32 *)get_property(cpunode, "cpu-version", NULL); + if (!valp) { + DBG("No cpu-version property !\n"); + goto bail_noprops; + } + if (((*valp) >> 16) != 0x3c) { + DBG("Wrong CPU version: %08x\n", *valp); + goto bail_noprops; + } + + /* Look for the powertune data in the device-tree */ + g5_pmode_data = (u32 *)get_property(cpunode, "power-mode-data",&psize); + if (!g5_pmode_data) { + DBG("No power-mode-data !\n"); + goto bail_noprops; + } + g5_pmode_max = psize / sizeof(u32) - 1; + + /* Look for the FVT table */ + shdr = smu_get_sdb_partition(SMU_SDB_FVT_ID, NULL); + if (!shdr) + goto bail_noprops; + g5_fvt_table = (struct smu_sdbp_fvt *)&shdr[1]; + ssize = (shdr->len * sizeof(u32)) - sizeof(struct smu_sdbp_header); + g5_fvt_count = ssize / sizeof(struct smu_sdbp_fvt); + g5_fvt_cur = 0; + + /* Sanity checking */ + if (g5_fvt_count < 1 || g5_pmode_max < 1) + goto bail_noprops; + + /* + * From what I see, clock-frequency is always the maximal frequency. + * The current driver can not slew sysclk yet, so we really only deal + * with powertune steps for now. We also only implement full freq and + * half freq in this version. So far, I haven't yet seen a machine + * supporting anything else. + */ + valp = (u32 *)get_property(cpunode, "clock-frequency", NULL); + if (!valp) + return -ENODEV; + max_freq = (*valp)/1000; + g5_cpu_freqs[0].frequency = max_freq; + g5_cpu_freqs[1].frequency = max_freq/2; + + /* Check current frequency */ + g5_pmode_cur = g5_query_freq(); + if (g5_pmode_cur > 1) + /* We don't support anything but 1:1 and 1:2, fixup ... */ + g5_pmode_cur = 1; + + /* Force apply current frequency to make sure everything is in + * sync (voltage is right for example). Firmware may leave us with + * a strange setting ... + */ + g5_switch_freq(g5_pmode_cur); + + printk(KERN_INFO "Registering G5 CPU frequency driver\n"); + printk(KERN_INFO "Low: %d Mhz, High: %d Mhz, Cur: %d MHz\n", + g5_cpu_freqs[1].frequency/1000, + g5_cpu_freqs[0].frequency/1000, + g5_cpu_freqs[g5_pmode_cur].frequency/1000); + + rc = cpufreq_register_driver(&g5_cpufreq_driver); + + /* We keep the CPU node on hold... hopefully, Apple G5 don't have + * hotplug CPU with a dynamic device-tree ... + */ + return rc; + + bail_noprops: + of_node_put(cpunode); + + return rc; +} + +module_init(g5_cpufreq_init); + + +MODULE_LICENSE("GPL"); diff --git a/arch/powerpc/platforms/powermac/setup.c b/arch/powerpc/platforms/powermac/setup.c index 80b58c1ec412..7acb0546671f 100644 --- a/arch/powerpc/platforms/powermac/setup.c +++ b/arch/powerpc/platforms/powermac/setup.c @@ -193,18 +193,6 @@ static void pmac_show_cpuinfo(struct seq_file *m) pmac_newworld ? "NewWorld" : "OldWorld"); } -static void pmac_show_percpuinfo(struct seq_file *m, int i) -{ -#ifdef CONFIG_CPU_FREQ_PMAC - extern unsigned int pmac_get_one_cpufreq(int i); - unsigned int freq = pmac_get_one_cpufreq(i); - if (freq != 0) { - seq_printf(m, "clock\t\t: %dMHz\n", freq/1000); - return; - } -#endif /* CONFIG_CPU_FREQ_PMAC */ -} - #ifndef CONFIG_ADB_CUDA int find_via_cuda(void) { @@ -767,7 +755,6 @@ struct machdep_calls __initdata pmac_md = { .setup_arch = pmac_setup_arch, .init_early = pmac_init_early, .show_cpuinfo = pmac_show_cpuinfo, - .show_percpuinfo = pmac_show_percpuinfo, .init_IRQ = pmac_pic_init, .get_irq = mpic_get_irq, /* changed later */ .pcibios_fixup = pmac_pcibios_fixup, diff --git a/arch/ppc64/Kconfig b/arch/ppc64/Kconfig index 2130cc315957..94a81279deac 100644 --- a/arch/ppc64/Kconfig +++ b/arch/ppc64/Kconfig @@ -173,6 +173,16 @@ config KEXEC support. As of this writing the exact hardware interface is strongly in flux, so no good recommendation can be made. +source "drivers/cpufreq/Kconfig" + +config CPU_FREQ_PMAC64 + bool "Support for some Apple G5s" + depends on CPU_FREQ && PMAC_SMU && PPC64 + select CPU_FREQ_TABLE + help + This adds support for frequency switching on Apple iMac G5, + and some of the more recent desktop G5 machines as well. + config IBMVIO depends on PPC_PSERIES || PPC_ISERIES bool diff --git a/arch/ppc64/kernel/misc.S b/arch/ppc64/kernel/misc.S index 077507ffbab8..914632ec587d 100644 --- a/arch/ppc64/kernel/misc.S +++ b/arch/ppc64/kernel/misc.S @@ -560,7 +560,7 @@ _GLOBAL(real_readb) isync blr - /* +/* * Do an IO access in real mode */ _GLOBAL(real_writeb) @@ -592,6 +592,76 @@ _GLOBAL(real_writeb) blr #endif /* defined(CONFIG_PPC_PMAC) || defined(CONFIG_PPC_MAPLE) */ +/* + * SCOM access functions for 970 (FX only for now) + * + * unsigned long scom970_read(unsigned int address); + * void scom970_write(unsigned int address, unsigned long value); + * + * The address passed in is the 24 bits register address. This code + * is 970 specific and will not check the status bits, so you should + * know what you are doing. + */ +_GLOBAL(scom970_read) + /* interrupts off */ + mfmsr r4 + ori r0,r4,MSR_EE + xori r0,r0,MSR_EE + mtmsrd r0,1 + + /* rotate 24 bits SCOM address 8 bits left and mask out it's low 8 bits + * (including parity). On current CPUs they must be 0'd, + * and finally or in RW bit + */ + rlwinm r3,r3,8,0,15 + ori r3,r3,0x8000 + + /* do the actual scom read */ + sync + mtspr SPRN_SCOMC,r3 + isync + mfspr r3,SPRN_SCOMD + isync + mfspr r0,SPRN_SCOMC + isync + + /* XXX: fixup result on some buggy 970's (ouch ! we lost a bit, bah + * that's the best we can do). Not implemented yet as we don't use + * the scom on any of the bogus CPUs yet, but may have to be done + * ultimately + */ + + /* restore interrupts */ + mtmsrd r4,1 + blr + + +_GLOBAL(scom970_write) + /* interrupts off */ + mfmsr r5 + ori r0,r5,MSR_EE + xori r0,r0,MSR_EE + mtmsrd r0,1 + + /* rotate 24 bits SCOM address 8 bits left and mask out it's low 8 bits + * (including parity). On current CPUs they must be 0'd. + */ + + rlwinm r3,r3,8,0,15 + + sync + mtspr SPRN_SCOMD,r4 /* write data */ + isync + mtspr SPRN_SCOMC,r3 /* write command */ + isync + mfspr 3,SPRN_SCOMC + isync + + /* restore interrupts */ + mtmsrd r5,1 + blr + + /* * Create a kernel thread * kernel_thread(fn, arg, flags) diff --git a/drivers/macintosh/smu.c b/drivers/macintosh/smu.c index 34f3c7e2d832..a931e508feb6 100644 --- a/drivers/macintosh/smu.c +++ b/drivers/macintosh/smu.c @@ -845,6 +845,18 @@ int smu_queue_i2c(struct smu_i2c_cmd *cmd) return 0; } +struct smu_sdbp_header *smu_get_sdb_partition(int id, unsigned int *size) +{ + char pname[32]; + + if (!smu) + return NULL; + + sprintf(pname, "sdb-partition-%02x", id); + return (struct smu_sdbp_header *)get_property(smu->of_node, + pname, size); +} +EXPORT_SYMBOL(smu_get_sdb_partition); /* diff --git a/include/asm-powerpc/reg.h b/include/asm-powerpc/reg.h index da848412f11b..489cf4c99c21 100644 --- a/include/asm-powerpc/reg.h +++ b/include/asm-powerpc/reg.h @@ -396,6 +396,9 @@ #define SPRN_VRSAVE 0x100 /* Vector Register Save Register */ #define SPRN_XER 0x001 /* Fixed Point Exception Register */ +#define SPRN_SCOMC 0x114 /* SCOM Access Control */ +#define SPRN_SCOMD 0x115 /* SCOM Access DATA */ + /* Performance monitor SPRs */ #ifdef CONFIG_PPC64 #define SPRN_MMCR0 795 @@ -594,7 +597,11 @@ static inline void ppc64_runlatch_off(void) mtspr(SPRN_CTRLT, ctrl); } } -#endif + +extern unsigned long scom970_read(unsigned int address); +extern void scom970_write(unsigned int address, unsigned long value); + +#endif /* CONFIG_PPC64 */ #define __get_SP() ({unsigned long sp; \ asm volatile("mr %0,1": "=r" (sp)); sp;}) diff --git a/include/asm-powerpc/smu.h b/include/asm-powerpc/smu.h index dee8eefe47bc..959bad660233 100644 --- a/include/asm-powerpc/smu.h +++ b/include/asm-powerpc/smu.h @@ -144,7 +144,11 @@ * - lenght 8 ("VSLEWxyz") has 3 additional bytes appended, and is * used to set the voltage slewing point. The SMU replies with "DONE" * I yet have to figure out their exact meaning of those 3 bytes in - * both cases. + * both cases. They seem to be: + * x = processor mask + * y = op. point index + * z = processor freq. step index + * I haven't yet decyphered result codes * */ #define SMU_CMD_POWER_COMMAND 0xaa @@ -333,6 +337,60 @@ extern int smu_queue_i2c(struct smu_i2c_cmd *cmd); #endif /* __KERNEL__ */ + +/* + * - SMU "sdb" partitions informations - + */ + + +/* + * Partition header format + */ +struct smu_sdbp_header { + __u8 id; + __u8 len; + __u8 version; + __u8 flags; +}; + +/* + * 32 bits integers are usually encoded with 2x16 bits swapped, + * this demangles them + */ +#define SMU_U32_MIX(x) ((((x) << 16) & 0xffff0000u) | (((x) >> 16) & 0xffffu)) + +/* This is the definition of the SMU sdb-partition-0x12 table (called + * CPU F/V/T operating points in Darwin). The definition for all those + * SMU tables should be moved to some separate file + */ +#define SMU_SDB_FVT_ID 0x12 + +struct smu_sdbp_fvt { + __u32 sysclk; /* Base SysClk frequency in Hz for + * this operating point + */ + __u8 pad; + __u8 maxtemp; /* Max temp. supported by this + * operating point + */ + + __u16 volts[3]; /* CPU core voltage for the 3 + * PowerTune modes, a mode with + * 0V = not supported. + */ +}; + +#ifdef __KERNEL__ +/* + * This returns the pointer to an SMU "sdb" partition data or NULL + * if not found. The data format is described below + */ +extern struct smu_sdbp_header *smu_get_sdb_partition(int id, + unsigned int *size); + +#endif /* __KERNEL__ */ + + /* * - Userland interface - */ -- cgit v1.2.3 From 183d020258dfd08178a05c6793dae10409db8abb Mon Sep 17 00:00:00 2001 From: Benjamin Herrenschmidt Date: Mon, 7 Nov 2005 14:29:02 +1100 Subject: [PATCH] ppc64: SMU partition recovery This patch adds the ability to the SMU driver to recover missing calibration partitions from the SMU chip itself. It also adds some dynamic mecanism to /proc/device-tree so that new properties are visible to userland. Signed-off-by: Benjamin Herrenschmidt Signed-off-by: Paul Mackerras --- arch/powerpc/kernel/prom.c | 21 +++++- arch/ppc/syslib/prom.c | 4 +- arch/ppc64/kernel/prom.c | 24 +++++-- drivers/macintosh/smu.c | 164 ++++++++++++++++++++++++++++++++++++++++++--- fs/proc/proc_devtree.c | 57 ++++++++++------ include/asm-powerpc/prom.h | 2 +- include/asm-powerpc/smu.h | 153 +++++++++++++++++++++++++++++++++++++++--- include/asm-ppc/prom.h | 2 +- include/asm-ppc64/prom.h | 2 +- include/linux/proc_fs.h | 9 +-- 10 files changed, 381 insertions(+), 57 deletions(-) (limited to 'arch/ppc64') diff --git a/arch/powerpc/kernel/prom.c b/arch/powerpc/kernel/prom.c index 3675ef4bac90..f645adb57534 100644 --- a/arch/powerpc/kernel/prom.c +++ b/arch/powerpc/kernel/prom.c @@ -1974,14 +1974,29 @@ EXPORT_SYMBOL(get_property); /* * Add a property to a node */ -void prom_add_property(struct device_node* np, struct property* prop) +int prom_add_property(struct device_node* np, struct property* prop) { - struct property **next = &np->properties; + struct property **next; prop->next = NULL; - while (*next) + write_lock(&devtree_lock); + next = &np->properties; + while (*next) { + if (strcmp(prop->name, (*next)->name) == 0) { + /* duplicate ! don't insert it */ + write_unlock(&devtree_lock); + return -1; + } next = &(*next)->next; + } *next = prop; + write_unlock(&devtree_lock); + + /* try to add to proc as well if it was initialized */ + if (np->pde) + proc_device_tree_add_prop(np->pde, prop); + + return 0; } /* I quickly hacked that one, check against spec ! */ diff --git a/arch/ppc/syslib/prom.c b/arch/ppc/syslib/prom.c index 1b9aa0d6a924..03b1fc9b9501 100644 --- a/arch/ppc/syslib/prom.c +++ b/arch/ppc/syslib/prom.c @@ -1165,7 +1165,7 @@ get_property(struct device_node *np, const char *name, int *lenp) /* * Add a property to a node */ -void +int prom_add_property(struct device_node* np, struct property* prop) { struct property **next = &np->properties; @@ -1174,6 +1174,8 @@ prom_add_property(struct device_node* np, struct property* prop) while (*next) next = &(*next)->next; *next = prop; + + return 0; } /* I quickly hacked that one, check against spec ! */ diff --git a/arch/ppc64/kernel/prom.c b/arch/ppc64/kernel/prom.c index 0e8961dea3bc..3402fbee62c7 100644 --- a/arch/ppc64/kernel/prom.c +++ b/arch/ppc64/kernel/prom.c @@ -31,6 +31,7 @@ #include #include #include +#include #include #include @@ -1865,17 +1866,32 @@ get_property(struct device_node *np, const char *name, int *lenp) EXPORT_SYMBOL(get_property); /* - * Add a property to a node + * Add a property to a node. */ -void +int prom_add_property(struct device_node* np, struct property* prop) { - struct property **next = &np->properties; + struct property **next; prop->next = NULL; - while (*next) + write_lock(&devtree_lock); + next = &np->properties; + while (*next) { + if (strcmp(prop->name, (*next)->name) == 0) { + /* duplicate ! don't insert it */ + write_unlock(&devtree_lock); + return -1; + } next = &(*next)->next; + } *next = prop; + write_unlock(&devtree_lock); + + /* try to add to proc as well if it was initialized */ + if (np->pde) + proc_device_tree_add_prop(np->pde, prop); + + return 0; } #if 0 diff --git a/drivers/macintosh/smu.c b/drivers/macintosh/smu.c index a931e508feb6..a83c4acf5710 100644 --- a/drivers/macintosh/smu.c +++ b/drivers/macintosh/smu.c @@ -47,13 +47,13 @@ #include #include -#define VERSION "0.6" +#define VERSION "0.7" #define AUTHOR "(c) 2005 Benjamin Herrenschmidt, IBM Corp." #undef DEBUG_SMU #ifdef DEBUG_SMU -#define DPRINTK(fmt, args...) do { printk(KERN_DEBUG fmt , ##args); } while (0) +#define DPRINTK(fmt, args...) do { udbg_printf(KERN_DEBUG fmt , ##args); } while (0) #else #define DPRINTK(fmt, args...) do { } while (0) #endif @@ -92,7 +92,7 @@ struct smu_device { * for now, just hard code that */ static struct smu_device *smu; - +static DECLARE_MUTEX(smu_part_access); /* * SMU driver low level stuff @@ -113,9 +113,11 @@ static void smu_start_cmd(void) DPRINTK("SMU: starting cmd %x, %d bytes data\n", cmd->cmd, cmd->data_len); - DPRINTK("SMU: data buffer: %02x %02x %02x %02x ...\n", + DPRINTK("SMU: data buffer: %02x %02x %02x %02x %02x %02x %02x %02x\n", ((u8 *)cmd->data_buf)[0], ((u8 *)cmd->data_buf)[1], - ((u8 *)cmd->data_buf)[2], ((u8 *)cmd->data_buf)[3]); + ((u8 *)cmd->data_buf)[2], ((u8 *)cmd->data_buf)[3], + ((u8 *)cmd->data_buf)[4], ((u8 *)cmd->data_buf)[5], + ((u8 *)cmd->data_buf)[6], ((u8 *)cmd->data_buf)[7]); /* Fill the SMU command buffer */ smu->cmd_buf->cmd = cmd->cmd; @@ -440,7 +442,7 @@ int smu_present(void) EXPORT_SYMBOL(smu_present); -int smu_init (void) +int __init smu_init (void) { struct device_node *np; u32 *data; @@ -845,16 +847,154 @@ int smu_queue_i2c(struct smu_i2c_cmd *cmd) return 0; } -struct smu_sdbp_header *smu_get_sdb_partition(int id, unsigned int *size) +/* + * Handling of "partitions" + */ + +static int smu_read_datablock(u8 *dest, unsigned int addr, unsigned int len) +{ + DECLARE_COMPLETION(comp); + unsigned int chunk; + struct smu_cmd cmd; + int rc; + u8 params[8]; + + /* We currently use a chunk size of 0xe. We could check the + * SMU firmware version and use bigger sizes though + */ + chunk = 0xe; + + while (len) { + unsigned int clen = min(len, chunk); + + cmd.cmd = SMU_CMD_MISC_ee_COMMAND; + cmd.data_len = 7; + cmd.data_buf = params; + cmd.reply_len = chunk; + cmd.reply_buf = dest; + cmd.done = smu_done_complete; + cmd.misc = ∁ + params[0] = SMU_CMD_MISC_ee_GET_DATABLOCK_REC; + params[1] = 0x4; + *((u32 *)¶ms[2]) = addr; + params[6] = clen; + + rc = smu_queue_cmd(&cmd); + if (rc) + return rc; + wait_for_completion(&comp); + if (cmd.status != 0) + return rc; + if (cmd.reply_len != clen) { + printk(KERN_DEBUG "SMU: short read in " + "smu_read_datablock, got: %d, want: %d\n", + cmd.reply_len, clen); + return -EIO; + } + len -= clen; + addr += clen; + dest += clen; + } + return 0; +} + +static struct smu_sdbp_header *smu_create_sdb_partition(int id) +{ + DECLARE_COMPLETION(comp); + struct smu_simple_cmd cmd; + unsigned int addr, len, tlen; + struct smu_sdbp_header *hdr; + struct property *prop; + + /* First query the partition info */ + smu_queue_simple(&cmd, SMU_CMD_PARTITION_COMMAND, 2, + smu_done_complete, &comp, + SMU_CMD_PARTITION_LATEST, id); + wait_for_completion(&comp); + + /* Partition doesn't exist (or other error) */ + if (cmd.cmd.status != 0 || cmd.cmd.reply_len != 6) + return NULL; + + /* Fetch address and length from reply */ + addr = *((u16 *)cmd.buffer); + len = cmd.buffer[3] << 2; + /* Calucluate total length to allocate, including the 17 bytes + * for "sdb-partition-XX" that we append at the end of the buffer + */ + tlen = sizeof(struct property) + len + 18; + + prop = kcalloc(tlen, 1, GFP_KERNEL); + if (prop == NULL) + return NULL; + hdr = (struct smu_sdbp_header *)(prop + 1); + prop->name = ((char *)prop) + tlen - 18; + sprintf(prop->name, "sdb-partition-%02x", id); + prop->length = len; + prop->value = (unsigned char *)hdr; + prop->next = NULL; + + /* Read the datablock */ + if (smu_read_datablock((u8 *)hdr, addr, len)) { + printk(KERN_DEBUG "SMU: datablock read failed while reading " + "partition %02x !\n", id); + goto failure; + } + + /* Got it, check a few things and create the property */ + if (hdr->id != id) { + printk(KERN_DEBUG "SMU: Reading partition %02x and got " + "%02x !\n", id, hdr->id); + goto failure; + } + if (prom_add_property(smu->of_node, prop)) { + printk(KERN_DEBUG "SMU: Failed creating sdb-partition-%02x " + "property !\n", id); + goto failure; + } + + return hdr; + failure: + kfree(prop); + return NULL; +} + +/* Note: Only allowed to return error code in pointers (using ERR_PTR) + * when interruptible is 1 + */ +struct smu_sdbp_header *__smu_get_sdb_partition(int id, unsigned int *size, + int interruptible) { char pname[32]; + struct smu_sdbp_header *part; if (!smu) return NULL; sprintf(pname, "sdb-partition-%02x", id); - return (struct smu_sdbp_header *)get_property(smu->of_node, + + if (interruptible) { + int rc; + rc = down_interruptible(&smu_part_access); + if (rc) + return ERR_PTR(rc); + } else + down(&smu_part_access); + + part = (struct smu_sdbp_header *)get_property(smu->of_node, pname, size); + if (part == NULL) { + part = smu_create_sdb_partition(id); + if (part != NULL && size) + *size = part->len << 2; + } + up(&smu_part_access); + return part; +} + +struct smu_sdbp_header *smu_get_sdb_partition(int id, unsigned int *size) +{ + return __smu_get_sdb_partition(id, size, 0); } EXPORT_SYMBOL(smu_get_sdb_partition); @@ -930,6 +1070,14 @@ static ssize_t smu_write(struct file *file, const char __user *buf, else if (hdr.cmdtype == SMU_CMDTYPE_WANTS_EVENTS) { pp->mode = smu_file_events; return 0; + } else if (hdr.cmdtype == SMU_CMDTYPE_GET_PARTITION) { + struct smu_sdbp_header *part; + part = __smu_get_sdb_partition(hdr.cmd, NULL, 1); + if (part == NULL) + return -EINVAL; + else if (IS_ERR(part)) + return PTR_ERR(part); + return 0; } else if (hdr.cmdtype != SMU_CMDTYPE_SMU) return -EINVAL; else if (pp->mode != smu_file_commands) diff --git a/fs/proc/proc_devtree.c b/fs/proc/proc_devtree.c index 6fd57f154197..fb117b74809e 100644 --- a/fs/proc/proc_devtree.c +++ b/fs/proc/proc_devtree.c @@ -48,6 +48,39 @@ static int property_read_proc(char *page, char **start, off_t off, * and "@10" to it. */ +/* + * Add a property to a node + */ +static struct proc_dir_entry * +__proc_device_tree_add_prop(struct proc_dir_entry *de, struct property *pp) +{ + struct proc_dir_entry *ent; + + /* + * Unfortunately proc_register puts each new entry + * at the beginning of the list. So we rearrange them. + */ + ent = create_proc_read_entry(pp->name, + strncmp(pp->name, "security-", 9) + ? S_IRUGO : S_IRUSR, de, + property_read_proc, pp); + if (ent == NULL) + return NULL; + + if (!strncmp(pp->name, "security-", 9)) + ent->size = 0; /* don't leak number of password chars */ + else + ent->size = pp->length; + + return ent; +} + + +void proc_device_tree_add_prop(struct proc_dir_entry *pde, struct property *prop) +{ + __proc_device_tree_add_prop(pde, prop); +} + /* * Process a node, adding entries for its children and its properties. */ @@ -57,11 +90,9 @@ void proc_device_tree_add_node(struct device_node *np, struct property *pp; struct proc_dir_entry *ent; struct device_node *child; - struct proc_dir_entry *list = NULL, **lastp; const char *p; set_node_proc_entry(np, de); - lastp = &list; for (child = NULL; (child = of_get_next_child(np, child));) { p = strrchr(child->full_name, '/'); if (!p) @@ -71,9 +102,6 @@ void proc_device_tree_add_node(struct device_node *np, ent = proc_mkdir(p, de); if (ent == 0) break; - *lastp = ent; - ent->next = NULL; - lastp = &ent->next; proc_device_tree_add_node(child, ent); } of_node_put(child); @@ -84,7 +112,7 @@ void proc_device_tree_add_node(struct device_node *np, * properties are quite unimportant for us though, thus we * simply "skip" them here, but we do have to check. */ - for (ent = list; ent != NULL; ent = ent->next) + for (ent = de->subdir; ent != NULL; ent = ent->next) if (!strcmp(ent->name, pp->name)) break; if (ent != NULL) { @@ -94,25 +122,10 @@ void proc_device_tree_add_node(struct device_node *np, continue; } - /* - * Unfortunately proc_register puts each new entry - * at the beginning of the list. So we rearrange them. - */ - ent = create_proc_read_entry(pp->name, - strncmp(pp->name, "security-", 9) - ? S_IRUGO : S_IRUSR, de, - property_read_proc, pp); + ent = __proc_device_tree_add_prop(de, pp); if (ent == 0) break; - if (!strncmp(pp->name, "security-", 9)) - ent->size = 0; /* don't leak number of password chars */ - else - ent->size = pp->length; - ent->next = NULL; - *lastp = ent; - lastp = &ent->next; } - de->subdir = list; } /* diff --git a/include/asm-powerpc/prom.h b/include/asm-powerpc/prom.h index 7587bf5f38c6..f999df1c5c90 100644 --- a/include/asm-powerpc/prom.h +++ b/include/asm-powerpc/prom.h @@ -203,7 +203,7 @@ extern int prom_n_addr_cells(struct device_node* np); extern int prom_n_size_cells(struct device_node* np); extern int prom_n_intr_cells(struct device_node* np); extern void prom_get_irq_senses(unsigned char *senses, int off, int max); -extern void prom_add_property(struct device_node* np, struct property* prop); +extern int prom_add_property(struct device_node* np, struct property* prop); #ifdef CONFIG_PPC32 /* diff --git a/include/asm-powerpc/smu.h b/include/asm-powerpc/smu.h index 959bad660233..76c29a9784dd 100644 --- a/include/asm-powerpc/smu.h +++ b/include/asm-powerpc/smu.h @@ -20,16 +20,52 @@ /* * Partition info commands * - * I do not know what those are for at this point + * These commands are used to retreive the sdb-partition-XX datas from + * the SMU. The lenght is always 2. First byte is the subcommand code + * and second byte is the partition ID. + * + * The reply is 6 bytes: + * + * - 0..1 : partition address + * - 2 : a byte containing the partition ID + * - 3 : length (maybe other bits are rest of header ?) + * + * The data must then be obtained with calls to another command: + * SMU_CMD_MISC_ee_GET_DATABLOCK_REC (described below). */ #define SMU_CMD_PARTITION_COMMAND 0x3e +#define SMU_CMD_PARTITION_LATEST 0x01 +#define SMU_CMD_PARTITION_BASE 0x02 +#define SMU_CMD_PARTITION_UPDATE 0x03 /* * Fan control * - * This is a "mux" for fan control commands, first byte is the - * "sub" command. + * This is a "mux" for fan control commands. The command seem to + * act differently based on the number of arguments. With 1 byte + * of argument, this seem to be queries for fans status, setpoint, + * etc..., while with 0xe arguments, we will set the fans speeds. + * + * Queries (1 byte arg): + * --------------------- + * + * arg=0x01: read RPM fans status + * arg=0x02: read RPM fans setpoint + * arg=0x11: read PWM fans status + * arg=0x12: read PWM fans setpoint + * + * the "status" queries return the current speed while the "setpoint" ones + * return the programmed/target speed. It _seems_ that the result is a bit + * mask in the first byte of active/available fans, followed by 6 words (16 + * bits) containing the requested speed. + * + * Setpoint (14 bytes arg): + * ------------------------ + * + * first arg byte is 0 for RPM fans and 0x10 for PWM. Second arg byte is the + * mask of fans affected by the command. Followed by 6 words containing the + * setpoint value for selected fans in the mask (or 0 if mask value is 0) */ #define SMU_CMD_FAN_COMMAND 0x4a @@ -156,6 +192,14 @@ #define SMU_CMD_POWER_SHUTDOWN "SHUTDOWN" #define SMU_CMD_POWER_VOLTAGE_SLEW "VSLEW" +/* + * Read ADC sensors + * + * This command takes one byte of parameter: the sensor ID (or "reg" + * value in the device-tree) and returns a 16 bits value + */ +#define SMU_CMD_READ_ADC 0xd8 + /* Misc commands * * This command seem to be a grab bag of various things @@ -176,6 +220,25 @@ * Misc commands * * This command seem to be a grab bag of various things + * + * SMU_CMD_MISC_ee_GET_DATABLOCK_REC is used, among others, to + * transfer blocks of data from the SMU. So far, I've decrypted it's + * usage to retreive partition data. In order to do that, you have to + * break your transfer in "chunks" since that command cannot transfer + * more than a chunk at a time. The chunk size used by OF is 0xe bytes, + * but it seems that the darwin driver will let you do 0x1e bytes if + * your "PMU" version is >= 0x30. You can get the "PMU" version apparently + * either in the last 16 bits of property "smu-version-pmu" or as the 16 + * bytes at offset 1 of "smu-version-info" + * + * For each chunk, the command takes 7 bytes of arguments: + * byte 0: subcommand code (0x02) + * byte 1: 0x04 (always, I don't know what it means, maybe the address + * space to use or some other nicety. It's hard coded in OF) + * byte 2..5: SMU address of the chunk (big endian 32 bits) + * byte 6: size to transfer (up to max chunk size) + * + * The data is returned directly */ #define SMU_CMD_MISC_ee_COMMAND 0xee #define SMU_CMD_MISC_ee_GET_DATABLOCK_REC 0x02 @@ -353,21 +416,26 @@ struct smu_sdbp_header { __u8 flags; }; -/* - * 32 bits integers are usually encoded with 2x16 bits swapped, - * this demangles them + + /* + * demangle 16 and 32 bits integer in some SMU partitions + * (currently, afaik, this concerns only the FVT partition + * (0x12) */ -#define SMU_U32_MIX(x) ((((x) << 16) & 0xffff0000u) | (((x) >> 16) & 0xffffu)) +#define SMU_U16_MIX(x) le16_to_cpu(x); +#define SMU_U32_MIX(x) ((((x) & 0xff00ff00u) >> 8)|(((x) & 0x00ff00ffu) << 8)) + /* This is the definition of the SMU sdb-partition-0x12 table (called * CPU F/V/T operating points in Darwin). The definition for all those * SMU tables should be moved to some separate file */ -#define SMU_SDB_FVT_ID 0x12 +#define SMU_SDB_FVT_ID 0x12 struct smu_sdbp_fvt { __u32 sysclk; /* Base SysClk frequency in Hz for - * this operating point + * this operating point. Value need to + * be unmixed with SMU_U32_MIX() */ __u8 pad; __u8 maxtemp; /* Max temp. supported by this @@ -376,10 +444,73 @@ struct smu_sdbp_fvt { __u16 volts[3]; /* CPU core voltage for the 3 * PowerTune modes, a mode with - * 0V = not supported. + * 0V = not supported. Value need + * to be unmixed with SMU_U16_MIX() */ }; +/* This partition contains voltage & current sensor calibration + * informations + */ +#define SMU_SDB_CPUVCP_ID 0x21 + +struct smu_sdbp_cpuvcp { + __u16 volt_scale; /* u4.12 fixed point */ + __s16 volt_offset; /* s4.12 fixed point */ + __u16 curr_scale; /* u4.12 fixed point */ + __s16 curr_offset; /* s4.12 fixed point */ + __s32 power_quads[3]; /* s4.28 fixed point */ +}; + +/* This partition contains CPU thermal diode calibration + */ +#define SMU_SDB_CPUDIODE_ID 0x18 + +struct smu_sdbp_cpudiode { + __u16 m_value; /* u1.15 fixed point */ + __s16 b_value; /* s10.6 fixed point */ + +}; + +/* This partition contains Slots power calibration + */ +#define SMU_SDB_SLOTSPOW_ID 0x78 + +struct smu_sdbp_slotspow { + __u16 pow_scale; /* u4.12 fixed point */ + __s16 pow_offset; /* s4.12 fixed point */ +}; + +/* This partition contains machine specific version information about + * the sensor/control layout + */ +#define SMU_SDB_SENSORTREE_ID 0x25 + +struct smu_sdbp_sensortree { + u8 model_id; + u8 unknown[3]; +}; + +/* This partition contains CPU thermal control PID informations. So far + * only single CPU machines have been seen with an SMU, so we assume this + * carries only informations for those + */ +#define SMU_SDB_CPUPIDDATA_ID 0x17 + +struct smu_sdbp_cpupiddata { + u8 unknown1; + u8 target_temp_delta; + u8 unknown2; + u8 history_len; + s16 power_adj; + u16 max_power; + s32 gp,gr,gd; +}; + + +/* Other partitions without known structures */ +#define SMU_SDB_DEBUG_SWITCHES_ID 0x05 + #ifdef __KERNEL__ /* * This returns the pointer to an SMU "sdb" partition data or NULL @@ -423,8 +554,10 @@ struct smu_user_cmd_hdr __u32 cmdtype; #define SMU_CMDTYPE_SMU 0 /* SMU command */ #define SMU_CMDTYPE_WANTS_EVENTS 1 /* switch fd to events mode */ +#define SMU_CMDTYPE_GET_PARTITION 2 /* retreive an sdb partition */ __u8 cmd; /* SMU command byte */ + __u8 pad[3]; /* padding */ __u32 data_len; /* Lenght of data following */ }; diff --git a/include/asm-ppc/prom.h b/include/asm-ppc/prom.h index 75c0637acdc8..3e39827ed566 100644 --- a/include/asm-ppc/prom.h +++ b/include/asm-ppc/prom.h @@ -93,7 +93,7 @@ extern int device_is_compatible(struct device_node *device, const char *); extern int machine_is_compatible(const char *compat); extern unsigned char *get_property(struct device_node *node, const char *name, int *lenp); -extern void prom_add_property(struct device_node* np, struct property* prop); +extern int prom_add_property(struct device_node* np, struct property* prop); extern void prom_get_irq_senses(unsigned char *, int, int); extern int prom_n_addr_cells(struct device_node* np); extern int prom_n_size_cells(struct device_node* np); diff --git a/include/asm-ppc64/prom.h b/include/asm-ppc64/prom.h index bdb47174ff0e..76bb0266d67c 100644 --- a/include/asm-ppc64/prom.h +++ b/include/asm-ppc64/prom.h @@ -213,6 +213,6 @@ extern int prom_n_addr_cells(struct device_node* np); extern int prom_n_size_cells(struct device_node* np); extern int prom_n_intr_cells(struct device_node* np); extern void prom_get_irq_senses(unsigned char *senses, int off, int max); -extern void prom_add_property(struct device_node* np, struct property* prop); +extern int prom_add_property(struct device_node* np, struct property* prop); #endif /* _PPC64_PROM_H */ diff --git a/include/linux/proc_fs.h b/include/linux/proc_fs.h index 0563581e3a02..65ceeaa30652 100644 --- a/include/linux/proc_fs.h +++ b/include/linux/proc_fs.h @@ -139,15 +139,12 @@ extern void proc_tty_unregister_driver(struct tty_driver *driver); /* * proc_devtree.c */ +#ifdef CONFIG_PROC_DEVICETREE struct device_node; +struct property; extern void proc_device_tree_init(void); -#ifdef CONFIG_PROC_DEVICETREE extern void proc_device_tree_add_node(struct device_node *, struct proc_dir_entry *); -#else /* !CONFIG_PROC_DEVICETREE */ -static inline void proc_device_tree_add_node(struct device_node *np, struct proc_dir_entry *pde) -{ - return; -} +extern void proc_device_tree_add_prop(struct proc_dir_entry *pde, struct property *prop); #endif /* CONFIG_PROC_DEVICETREE */ extern struct proc_dir_entry *proc_symlink(const char *, -- cgit v1.2.3 From 7d49697ef92bd2cf84ab53bd4cea82fefb197fb9 Mon Sep 17 00:00:00 2001 From: Benjamin Herrenschmidt Date: Mon, 7 Nov 2005 14:36:21 +1100 Subject: [PATCH] ppc64: More U3 device-tree fixes Some more U3 revisions have the missing "interrupts" property in U3, this adds them to the fixup code in prom_init.c Signed-off-by: Benjamin Herrenschmidt Signed-off-by: Paul Mackerras --- arch/powerpc/kernel/prom_init.c | 2 +- arch/ppc64/kernel/prom_init.c | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-) (limited to 'arch/ppc64') diff --git a/arch/powerpc/kernel/prom_init.c b/arch/powerpc/kernel/prom_init.c index c758b6624d7b..0d91961f9433 100644 --- a/arch/powerpc/kernel/prom_init.c +++ b/arch/powerpc/kernel/prom_init.c @@ -1872,7 +1872,7 @@ static void __init fixup_device_tree(void) if (prom_getprop(u3, "device-rev", &u3_rev, sizeof(u3_rev)) == PROM_ERROR) return; - if (u3_rev != 0x35 && u3_rev != 0x37) + if (u3_rev < 0x35 || u3_rev > 0x39) return; /* does it need fixup ? */ if (prom_getproplen(i2c, "interrupts") > 0) diff --git a/arch/ppc64/kernel/prom_init.c b/arch/ppc64/kernel/prom_init.c index e72b3f9061f0..e4c880dab997 100644 --- a/arch/ppc64/kernel/prom_init.c +++ b/arch/ppc64/kernel/prom_init.c @@ -1824,7 +1824,7 @@ static void __init fixup_device_tree(void) if (prom_getprop(u3, "device-rev", &u3_rev, sizeof(u3_rev)) == PROM_ERROR) return; - if (u3_rev != 0x35 && u3_rev != 0x37) + if (u3_rev < 0x35 || u3_rev > 0x39) return; /* does it need fixup ? */ if (prom_getproplen(i2c, "interrupts") > 0) -- cgit v1.2.3 From 0286486783a75ef991df3ee250917efb55df75b3 Mon Sep 17 00:00:00 2001 From: Mike Kravetz Date: Mon, 7 Nov 2005 13:48:59 -0800 Subject: [PATCH] Memory Add Fixes for ppc64 On Tue, Nov 08, 2005 at 08:12:56AM +1100, Benjamin Herrenschmidt wrote: > Yes, the MAX_ORDER should be different indeed. But can Kconfig do that ? > That is have the default value be different based on a Kconfig option ? > I don't see that ... We may have to do things differently here... This seems to be done in other parts of the Kconfig file. Using those as an example, this should keep the MAX_ORDER block size at 16MB. Signed-off-by: Mike Kravetz Signed-off-by: Paul Mackerras --- arch/powerpc/Kconfig | 1 + arch/ppc64/Kconfig | 1 + 2 files changed, 2 insertions(+) (limited to 'arch/ppc64') diff --git a/arch/powerpc/Kconfig b/arch/powerpc/Kconfig index 3ac9195dab35..1493c7896fe3 100644 --- a/arch/powerpc/Kconfig +++ b/arch/powerpc/Kconfig @@ -492,6 +492,7 @@ source "fs/Kconfig.binfmt" config FORCE_MAX_ZONEORDER int depends on PPC64 + default "9" if PPC_64K_PAGES default "13" config MATH_EMULATION diff --git a/arch/ppc64/Kconfig b/arch/ppc64/Kconfig index 94a81279deac..29552348e581 100644 --- a/arch/ppc64/Kconfig +++ b/arch/ppc64/Kconfig @@ -56,6 +56,7 @@ config PPC_STD_MMU # max order + 1 config FORCE_MAX_ZONEORDER int + default "9" if PPC_64K_PAGES default "13" source "init/Kconfig" -- cgit v1.2.3 From 5bfb5d690f36d316a5f3b4f7775fda996faa6b12 Mon Sep 17 00:00:00 2001 From: Nick Piggin Date: Tue, 8 Nov 2005 21:39:01 -0800 Subject: [PATCH] sched: disable preempt in idle tasks Run idle threads with preempt disabled. Also corrected a bugs in arm26's cpu_idle (make it actually call schedule()). How did it ever work before? Might fix the CPU hotplugging hang which Nigel Cunningham noted. We think the bug hits if the idle thread is preempted after checking need_resched() and before going to sleep, then the CPU offlined. After calling stop_machine_run, the CPU eventually returns from preemption and into the idle thread and goes to sleep. The CPU will continue executing previous idle and have no chance to call play_dead. By disabling preemption until we are ready to explicitly schedule, this bug is fixed and the idle threads generally become more robust. From: alexs PPC build fix From: Yoichi Yuasa MIPS build fix Signed-off-by: Nick Piggin Signed-off-by: Yoichi Yuasa Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- arch/arm/kernel/process.c | 4 ++-- arch/arm/kernel/smp.c | 5 ++++- arch/arm26/kernel/process.c | 12 +++++------- arch/cris/arch-v32/kernel/smp.c | 1 + arch/cris/kernel/process.c | 2 ++ arch/frv/kernel/process.c | 6 +++++- arch/h8300/kernel/process.c | 28 +++++++++++++++------------- arch/i386/kernel/process.c | 4 +++- arch/i386/kernel/smpboot.c | 1 + arch/ia64/kernel/process.c | 2 ++ arch/ia64/kernel/smpboot.c | 1 + arch/m32r/kernel/process.c | 2 ++ arch/m32r/kernel/smpboot.c | 1 + arch/m68k/kernel/process.c | 2 ++ arch/mips/kernel/process.c | 2 ++ arch/mips/kernel/smp.c | 4 +++- arch/parisc/kernel/process.c | 2 ++ arch/parisc/kernel/smp.c | 1 + arch/powerpc/platforms/iseries/setup.c | 4 ++++ arch/powerpc/platforms/pseries/setup.c | 4 ++++ arch/ppc/kernel/idle.c | 17 ++++++++++++----- arch/ppc/kernel/smp.c | 1 + arch/ppc64/kernel/idle.c | 4 ++++ arch/s390/kernel/process.c | 11 ++++++++--- arch/s390/kernel/smp.c | 1 + arch/sh/kernel/process.c | 2 ++ arch/sh/kernel/smp.c | 5 ++++- arch/sh64/kernel/process.c | 2 ++ arch/sparc/kernel/process.c | 4 ++++ arch/sparc64/kernel/process.c | 4 ++++ arch/sparc64/kernel/smp.c | 3 +++ arch/v850/kernel/process.c | 16 ++++++++++------ arch/x86_64/kernel/process.c | 2 ++ arch/x86_64/kernel/smpboot.c | 1 + arch/xtensa/kernel/process.c | 3 ++- init/main.c | 4 +++- 36 files changed, 125 insertions(+), 43 deletions(-) (limited to 'arch/ppc64') diff --git a/arch/arm/kernel/process.c b/arch/arm/kernel/process.c index ba298277becd..93dd92cc12f8 100644 --- a/arch/arm/kernel/process.c +++ b/arch/arm/kernel/process.c @@ -116,13 +116,13 @@ void cpu_idle(void) if (!idle) idle = default_idle; - preempt_disable(); leds_event(led_idle_start); while (!need_resched()) idle(); leds_event(led_idle_end); - preempt_enable(); + preempt_enable_no_resched(); schedule(); + preempt_disable(); } } diff --git a/arch/arm/kernel/smp.c b/arch/arm/kernel/smp.c index 77e2e9ca89fa..e55ea952f7aa 100644 --- a/arch/arm/kernel/smp.c +++ b/arch/arm/kernel/smp.c @@ -256,7 +256,9 @@ void __cpuexit cpu_die(void) asmlinkage void __cpuinit secondary_start_kernel(void) { struct mm_struct *mm = &init_mm; - unsigned int cpu = smp_processor_id(); + unsigned int cpu; + + cpu = smp_processor_id(); printk("CPU%u: Booted secondary processor\n", cpu); @@ -273,6 +275,7 @@ asmlinkage void __cpuinit secondary_start_kernel(void) local_flush_tlb_all(); cpu_init(); + preempt_disable(); /* * Give the platform a chance to do its own initialisation. diff --git a/arch/arm26/kernel/process.c b/arch/arm26/kernel/process.c index 9eb9964d32a7..15833a0057dd 100644 --- a/arch/arm26/kernel/process.c +++ b/arch/arm26/kernel/process.c @@ -74,15 +74,13 @@ __setup("hlt", hlt_setup); void cpu_idle(void) { /* endless idle loop with no priority at all */ - preempt_disable(); while (1) { - while (!need_resched()) { - local_irq_disable(); - if (!need_resched() && !hlt_counter) - local_irq_enable(); - } + while (!need_resched()) + cpu_relax(); + preempt_enable_no_resched(); + schedule(); + preempt_disable(); } - schedule(); } static char reboot_mode = 'h'; diff --git a/arch/cris/arch-v32/kernel/smp.c b/arch/cris/arch-v32/kernel/smp.c index 957f551ba5ce..13867f4fad16 100644 --- a/arch/cris/arch-v32/kernel/smp.c +++ b/arch/cris/arch-v32/kernel/smp.c @@ -161,6 +161,7 @@ void __init smp_callin(void) REG_WR(intr_vect, irq_regs[cpu], rw_mask, vect_mask); unmask_irq(IPI_INTR_VECT); unmask_irq(TIMER_INTR_VECT); + preempt_disable(); local_irq_enable(); cpu_set(cpu, cpu_online_map); diff --git a/arch/cris/kernel/process.c b/arch/cris/kernel/process.c index 949a0e40e03c..7c80afb10460 100644 --- a/arch/cris/kernel/process.c +++ b/arch/cris/kernel/process.c @@ -218,7 +218,9 @@ void cpu_idle (void) idle = default_idle; idle(); } + preempt_enable_no_resched(); schedule(); + preempt_disable(); } } diff --git a/arch/frv/kernel/process.c b/arch/frv/kernel/process.c index 3001b82b1514..54a452136f00 100644 --- a/arch/frv/kernel/process.c +++ b/arch/frv/kernel/process.c @@ -77,16 +77,20 @@ void (*idle)(void) = core_sleep_idle; */ void cpu_idle(void) { + int cpu = smp_processor_id(); + /* endless idle loop with no priority at all */ while (1) { while (!need_resched()) { - irq_stat[smp_processor_id()].idle_timestamp = jiffies; + irq_stat[cpu].idle_timestamp = jiffies; if (!frv_dma_inprogress && idle) idle(); } + preempt_enable_no_resched(); schedule(); + preempt_disable(); } } diff --git a/arch/h8300/kernel/process.c b/arch/h8300/kernel/process.c index 27f1fce64ce4..fe21adf3e75e 100644 --- a/arch/h8300/kernel/process.c +++ b/arch/h8300/kernel/process.c @@ -53,22 +53,18 @@ asmlinkage void ret_from_fork(void); #if !defined(CONFIG_H8300H_SIM) && !defined(CONFIG_H8S_SIM) void default_idle(void) { - while(1) { - if (!need_resched()) { - local_irq_enable(); - __asm__("sleep"); - local_irq_disable(); - } - schedule(); - } + local_irq_disable(); + if (!need_resched()) { + local_irq_enable(); + /* XXX: race here! What if need_resched() gets set now? */ + __asm__("sleep"); + } else + local_irq_enable(); } #else void default_idle(void) { - while(1) { - if (need_resched()) - schedule(); - } + cpu_relax(); } #endif void (*idle)(void) = default_idle; @@ -81,7 +77,13 @@ void (*idle)(void) = default_idle; */ void cpu_idle(void) { - idle(); + while (1) { + while (!need_resched()) + idle(); + preempt_enable_no_resched(); + schedule(); + preempt_disable(); + } } void machine_restart(char * __unused) diff --git a/arch/i386/kernel/process.c b/arch/i386/kernel/process.c index 7a14fdfd3af9..5296e284ea36 100644 --- a/arch/i386/kernel/process.c +++ b/arch/i386/kernel/process.c @@ -179,7 +179,7 @@ static inline void play_dead(void) */ void cpu_idle(void) { - int cpu = raw_smp_processor_id(); + int cpu = smp_processor_id(); /* endless idle loop with no priority at all */ while (1) { @@ -201,7 +201,9 @@ void cpu_idle(void) __get_cpu_var(irq_stat).idle_timestamp = jiffies; idle(); } + preempt_enable_no_resched(); schedule(); + preempt_disable(); } } diff --git a/arch/i386/kernel/smpboot.c b/arch/i386/kernel/smpboot.c index 47ec76794d02..bc5a9d97466b 100644 --- a/arch/i386/kernel/smpboot.c +++ b/arch/i386/kernel/smpboot.c @@ -485,6 +485,7 @@ static void __devinit start_secondary(void *unused) * things done here to the most necessary things. */ cpu_init(); + preempt_disable(); smp_callin(); while (!cpu_isset(smp_processor_id(), smp_commenced_mask)) rep_nop(); diff --git a/arch/ia64/kernel/process.c b/arch/ia64/kernel/process.c index 051e050359e4..4c621fc3c3b9 100644 --- a/arch/ia64/kernel/process.c +++ b/arch/ia64/kernel/process.c @@ -292,7 +292,9 @@ cpu_idle (void) #ifdef CONFIG_SMP normal_xtp(); #endif + preempt_enable_no_resched(); schedule(); + preempt_disable(); check_pgt_cache(); if (cpu_is_offline(smp_processor_id())) play_dead(); diff --git a/arch/ia64/kernel/smpboot.c b/arch/ia64/kernel/smpboot.c index 400a48987124..8f44e7d2df66 100644 --- a/arch/ia64/kernel/smpboot.c +++ b/arch/ia64/kernel/smpboot.c @@ -399,6 +399,7 @@ start_secondary (void *unused) Dprintk("start_secondary: starting CPU 0x%x\n", hard_smp_processor_id()); efi_map_pal_code(); cpu_init(); + preempt_disable(); smp_callin(); cpu_idle(); diff --git a/arch/m32r/kernel/process.c b/arch/m32r/kernel/process.c index ea13a8f4d8b0..cc4b571e5db7 100644 --- a/arch/m32r/kernel/process.c +++ b/arch/m32r/kernel/process.c @@ -104,7 +104,9 @@ void cpu_idle (void) idle(); } + preempt_enable_no_resched(); schedule(); + preempt_disable(); } } diff --git a/arch/m32r/kernel/smpboot.c b/arch/m32r/kernel/smpboot.c index 640d592ea072..b90c54169fa5 100644 --- a/arch/m32r/kernel/smpboot.c +++ b/arch/m32r/kernel/smpboot.c @@ -426,6 +426,7 @@ void __init smp_cpus_done(unsigned int max_cpus) int __init start_secondary(void *unused) { cpu_init(); + preempt_disable(); smp_callin(); while (!cpu_isset(smp_processor_id(), smp_commenced_mask)) cpu_relax(); diff --git a/arch/m68k/kernel/process.c b/arch/m68k/kernel/process.c index 11b1b90ba6ba..13d109328a42 100644 --- a/arch/m68k/kernel/process.c +++ b/arch/m68k/kernel/process.c @@ -102,7 +102,9 @@ void cpu_idle(void) while (1) { while (!need_resched()) idle(); + preempt_enable_no_resched(); schedule(); + preempt_disable(); } } diff --git a/arch/mips/kernel/process.c b/arch/mips/kernel/process.c index 4fe3d5715c41..dd725779d91f 100644 --- a/arch/mips/kernel/process.c +++ b/arch/mips/kernel/process.c @@ -52,7 +52,9 @@ ATTRIB_NORET void cpu_idle(void) while (!need_resched()) if (cpu_wait) (*cpu_wait)(); + preempt_enable_no_resched(); schedule(); + preempt_disable(); } } diff --git a/arch/mips/kernel/smp.c b/arch/mips/kernel/smp.c index fcacf1aae98a..25472fcaf715 100644 --- a/arch/mips/kernel/smp.c +++ b/arch/mips/kernel/smp.c @@ -82,7 +82,7 @@ extern ATTRIB_NORET void cpu_idle(void); */ asmlinkage void start_secondary(void) { - unsigned int cpu = smp_processor_id(); + unsigned int cpu; cpu_probe(); cpu_report(); @@ -95,6 +95,8 @@ asmlinkage void start_secondary(void) */ calibrate_delay(); + preempt_disable(); + cpu = smp_processor_id(); cpu_data[cpu].udelay_val = loops_per_jiffy; prom_smp_finish(); diff --git a/arch/parisc/kernel/process.c b/arch/parisc/kernel/process.c index 7fdca87ef647..f482f78de435 100644 --- a/arch/parisc/kernel/process.c +++ b/arch/parisc/kernel/process.c @@ -92,7 +92,9 @@ void cpu_idle(void) while (1) { while (!need_resched()) barrier(); + preempt_enable_no_resched(); schedule(); + preempt_disable(); check_pgt_cache(); } } diff --git a/arch/parisc/kernel/smp.c b/arch/parisc/kernel/smp.c index 5db3be4e2704..a9ecf6465784 100644 --- a/arch/parisc/kernel/smp.c +++ b/arch/parisc/kernel/smp.c @@ -463,6 +463,7 @@ void __init smp_callin(void) #endif smp_cpu_init(slave_id); + preempt_disable(); #if 0 /* NOT WORKING YET - see entry.S */ istack = (void *)__get_free_pages(GFP_KERNEL,ISTACK_ORDER); diff --git a/arch/powerpc/platforms/iseries/setup.c b/arch/powerpc/platforms/iseries/setup.c index d3e4bf756c83..0130f2619dac 100644 --- a/arch/powerpc/platforms/iseries/setup.c +++ b/arch/powerpc/platforms/iseries/setup.c @@ -694,7 +694,9 @@ static void iseries_shared_idle(void) if (hvlpevent_is_pending()) process_iSeries_events(); + preempt_enable_no_resched(); schedule(); + preempt_disable(); } } @@ -726,7 +728,9 @@ static void iseries_dedicated_idle(void) } ppc64_runlatch_on(); + preempt_enable_no_resched(); schedule(); + preempt_disable(); } } diff --git a/arch/powerpc/platforms/pseries/setup.c b/arch/powerpc/platforms/pseries/setup.c index e78c39368841..4854f5eb5c3d 100644 --- a/arch/powerpc/platforms/pseries/setup.c +++ b/arch/powerpc/platforms/pseries/setup.c @@ -539,7 +539,9 @@ static void pseries_dedicated_idle(void) lpaca->lppaca.idle = 0; ppc64_runlatch_on(); + preempt_enable_no_resched(); schedule(); + preempt_disable(); if (cpu_is_offline(cpu) && system_state == SYSTEM_RUNNING) cpu_die(); @@ -583,7 +585,9 @@ static void pseries_shared_idle(void) lpaca->lppaca.idle = 0; ppc64_runlatch_on(); + preempt_enable_no_resched(); schedule(); + preempt_disable(); if (cpu_is_offline(cpu) && system_state == SYSTEM_RUNNING) cpu_die(); diff --git a/arch/ppc/kernel/idle.c b/arch/ppc/kernel/idle.c index 11e5b44713f7..a6141f05c919 100644 --- a/arch/ppc/kernel/idle.c +++ b/arch/ppc/kernel/idle.c @@ -53,10 +53,6 @@ void default_idle(void) } #endif } - if (need_resched()) - schedule(); - if (cpu_is_offline(cpu) && system_state == SYSTEM_RUNNING) - cpu_die(); } /* @@ -64,11 +60,22 @@ void default_idle(void) */ void cpu_idle(void) { - for (;;) + int cpu = smp_processor_id(); + + for (;;) { if (ppc_md.idle != NULL) ppc_md.idle(); else default_idle(); + if (cpu_is_offline(cpu) && system_state == SYSTEM_RUNNING) + cpu_die(); + if (need_resched()) { + preempt_enable_no_resched(); + schedule(); + preempt_disable(); + } + + } } #if defined(CONFIG_SYSCTL) && defined(CONFIG_6xx) diff --git a/arch/ppc/kernel/smp.c b/arch/ppc/kernel/smp.c index bc5bf1124836..43b8fc2ca591 100644 --- a/arch/ppc/kernel/smp.c +++ b/arch/ppc/kernel/smp.c @@ -341,6 +341,7 @@ int __devinit start_secondary(void *unused) cpu = smp_processor_id(); smp_store_cpu_info(cpu); set_dec(tb_ticks_per_jiffy); + preempt_disable(); cpu_callin_map[cpu] = 1; printk("CPU %d done callin...\n", cpu); diff --git a/arch/ppc64/kernel/idle.c b/arch/ppc64/kernel/idle.c index 8fec27469802..909ea669af91 100644 --- a/arch/ppc64/kernel/idle.c +++ b/arch/ppc64/kernel/idle.c @@ -61,7 +61,9 @@ void default_idle(void) } ppc64_runlatch_on(); + preempt_enable_no_resched(); schedule(); + preempt_disable(); if (cpu_is_offline(cpu) && system_state == SYSTEM_RUNNING) cpu_die(); } @@ -77,7 +79,9 @@ void native_idle(void) if (need_resched()) { ppc64_runlatch_on(); + preempt_enable_no_resched(); schedule(); + preempt_disable(); } if (cpu_is_offline(smp_processor_id()) && diff --git a/arch/s390/kernel/process.c b/arch/s390/kernel/process.c index 9f3dff6c0b72..66ca5757e368 100644 --- a/arch/s390/kernel/process.c +++ b/arch/s390/kernel/process.c @@ -102,7 +102,6 @@ void default_idle(void) local_irq_disable(); if (need_resched()) { local_irq_enable(); - schedule(); return; } @@ -139,8 +138,14 @@ void default_idle(void) void cpu_idle(void) { - for (;;) - default_idle(); + for (;;) { + while (!need_resched()) + default_idle(); + + preempt_enable_no_resched(); + schedule(); + preempt_disable(); + } } void show_regs(struct pt_regs *regs) diff --git a/arch/s390/kernel/smp.c b/arch/s390/kernel/smp.c index e13c87b446b2..5856b3fda6bf 100644 --- a/arch/s390/kernel/smp.c +++ b/arch/s390/kernel/smp.c @@ -533,6 +533,7 @@ int __devinit start_secondary(void *cpuvoid) { /* Setup the cpu */ cpu_init(); + preempt_disable(); /* init per CPU timer */ init_cpu_timer(); #ifdef CONFIG_VIRT_TIMER diff --git a/arch/sh/kernel/process.c b/arch/sh/kernel/process.c index 6dce9d0b81f8..1cbc26b796ad 100644 --- a/arch/sh/kernel/process.c +++ b/arch/sh/kernel/process.c @@ -64,7 +64,9 @@ void default_idle(void) cpu_sleep(); } + preempt_enable_no_resched(); schedule(); + preempt_disable(); } } diff --git a/arch/sh/kernel/smp.c b/arch/sh/kernel/smp.c index 5ecefc02896a..59e49b18252c 100644 --- a/arch/sh/kernel/smp.c +++ b/arch/sh/kernel/smp.c @@ -112,7 +112,9 @@ int __cpu_up(unsigned int cpu) int start_secondary(void *unused) { - unsigned int cpu = smp_processor_id(); + unsigned int cpu; + + cpu = smp_processor_id(); atomic_inc(&init_mm.mm_count); current->active_mm = &init_mm; @@ -120,6 +122,7 @@ int start_secondary(void *unused) smp_store_cpu_info(cpu); __smp_slave_init(cpu); + preempt_disable(); per_cpu_trap_init(); atomic_inc(&cpus_booted); diff --git a/arch/sh64/kernel/process.c b/arch/sh64/kernel/process.c index efde41c0cd66..0c09537449b3 100644 --- a/arch/sh64/kernel/process.c +++ b/arch/sh64/kernel/process.c @@ -334,7 +334,9 @@ void default_idle(void) } local_irq_enable(); } + preempt_enable_no_resched(); schedule(); + preempt_disable(); } } diff --git a/arch/sparc/kernel/process.c b/arch/sparc/kernel/process.c index 29e72b57d4fd..c39f4d01096d 100644 --- a/arch/sparc/kernel/process.c +++ b/arch/sparc/kernel/process.c @@ -120,7 +120,9 @@ void cpu_idle(void) (*pm_idle)(); } + preempt_enable_no_resched(); schedule(); + preempt_disable(); check_pgt_cache(); } } @@ -133,7 +135,9 @@ void cpu_idle(void) /* endless idle loop with no priority at all */ while(1) { if(need_resched()) { + preempt_enable_no_resched(); schedule(); + preempt_disable(); check_pgt_cache(); } barrier(); /* or else gcc optimizes... */ diff --git a/arch/sparc64/kernel/process.c b/arch/sparc64/kernel/process.c index 7d10b0397091..2f89206e008f 100644 --- a/arch/sparc64/kernel/process.c +++ b/arch/sparc64/kernel/process.c @@ -74,7 +74,9 @@ void cpu_idle(void) while (!need_resched()) barrier(); + preempt_enable_no_resched(); schedule(); + preempt_disable(); check_pgt_cache(); } } @@ -93,7 +95,9 @@ void cpu_idle(void) if (need_resched()) { unidle_me(); clear_thread_flag(TIF_POLLING_NRFLAG); + preempt_enable_no_resched(); schedule(); + preempt_disable(); set_thread_flag(TIF_POLLING_NRFLAG); check_pgt_cache(); } diff --git a/arch/sparc64/kernel/smp.c b/arch/sparc64/kernel/smp.c index 5d90ee9aebf1..8aca4b1dc04e 100644 --- a/arch/sparc64/kernel/smp.c +++ b/arch/sparc64/kernel/smp.c @@ -168,6 +168,9 @@ void __init smp_callin(void) rmb(); cpu_set(cpuid, cpu_online_map); + + /* idle thread is expected to have preempt disabled */ + preempt_disable(); } void cpu_panic(void) diff --git a/arch/v850/kernel/process.c b/arch/v850/kernel/process.c index 9c708c32c1f0..39cf247cdae4 100644 --- a/arch/v850/kernel/process.c +++ b/arch/v850/kernel/process.c @@ -36,11 +36,8 @@ extern void ret_from_fork (void); /* The idle loop. */ void default_idle (void) { - while (1) { - while (! need_resched ()) - asm ("halt; nop; nop; nop; nop; nop" ::: "cc"); - schedule (); - } + while (! need_resched ()) + asm ("halt; nop; nop; nop; nop; nop" ::: "cc"); } void (*idle)(void) = default_idle; @@ -54,7 +51,14 @@ void (*idle)(void) = default_idle; void cpu_idle (void) { /* endless idle loop with no priority at all */ - (*idle) (); + while (1) { + while (!need_resched()) + (*idle) (); + + preempt_enable_no_resched(); + schedule(); + preempt_disable(); + } } /* diff --git a/arch/x86_64/kernel/process.c b/arch/x86_64/kernel/process.c index b5a89c0bdf59..571f9fe490ce 100644 --- a/arch/x86_64/kernel/process.c +++ b/arch/x86_64/kernel/process.c @@ -204,7 +204,9 @@ void cpu_idle (void) idle(); } + preempt_enable_no_resched(); schedule(); + preempt_disable(); } } diff --git a/arch/x86_64/kernel/smpboot.c b/arch/x86_64/kernel/smpboot.c index 4b5b088ec102..c4e59bbdc187 100644 --- a/arch/x86_64/kernel/smpboot.c +++ b/arch/x86_64/kernel/smpboot.c @@ -472,6 +472,7 @@ void __cpuinit start_secondary(void) * things done here to the most necessary things. */ cpu_init(); + preempt_disable(); smp_callin(); /* otherwise gcc will move up the smp_processor_id before the cpu_init */ diff --git a/arch/xtensa/kernel/process.c b/arch/xtensa/kernel/process.c index 08ef6d82ee51..6a44b54ae817 100644 --- a/arch/xtensa/kernel/process.c +++ b/arch/xtensa/kernel/process.c @@ -96,8 +96,9 @@ void cpu_idle(void) while (1) { while (!need_resched()) platform_idle(); - preempt_enable(); + preempt_enable_no_resched(); schedule(); + preempt_disable(); } } diff --git a/init/main.c b/init/main.c index f142d4035341..27f97f9b4636 100644 --- a/init/main.c +++ b/init/main.c @@ -394,14 +394,16 @@ static void noinline rest_init(void) kernel_thread(init, NULL, CLONE_FS | CLONE_SIGHAND); numa_default_policy(); unlock_kernel(); - preempt_enable_no_resched(); /* * The boot idle thread must execute schedule() * at least one to get things moving: */ + preempt_enable_no_resched(); schedule(); + preempt_disable(); + /* Call into cpu_idle with preempt disabled */ cpu_idle(); } -- cgit v1.2.3 From 64c7c8f88559624abdbe12b5da6502e8879f8d28 Mon Sep 17 00:00:00 2001 From: Nick Piggin Date: Tue, 8 Nov 2005 21:39:04 -0800 Subject: [PATCH] sched: resched and cpu_idle rework Make some changes to the NEED_RESCHED and POLLING_NRFLAG to reduce confusion, and make their semantics rigid. Improves efficiency of resched_task and some cpu_idle routines. * In resched_task: - TIF_NEED_RESCHED is only cleared with the task's runqueue lock held, and as we hold it during resched_task, then there is no need for an atomic test and set there. The only other time this should be set is when the task's quantum expires, in the timer interrupt - this is protected against because the rq lock is irq-safe. - If TIF_NEED_RESCHED is set, then we don't need to do anything. It won't get unset until the task get's schedule()d off. - If we are running on the same CPU as the task we resched, then set TIF_NEED_RESCHED and no further action is required. - If we are running on another CPU, and TIF_POLLING_NRFLAG is *not* set after TIF_NEED_RESCHED has been set, then we need to send an IPI. Using these rules, we are able to remove the test and set operation in resched_task, and make clear the previously vague semantics of POLLING_NRFLAG. * In idle routines: - Enter cpu_idle with preempt disabled. When the need_resched() condition becomes true, explicitly call schedule(). This makes things a bit clearer (IMO), but haven't updated all architectures yet. - Many do a test and clear of TIF_NEED_RESCHED for some reason. According to the resched_task rules, this isn't needed (and actually breaks the assumption that TIF_NEED_RESCHED is only cleared with the runqueue lock held). So remove that. Generally one less locked memory op when switching to the idle thread. - Many idle routines clear TIF_POLLING_NRFLAG, and only set it in the inner most polling idle loops. The above resched_task semantics allow it to be set until before the last time need_resched() is checked before going into a halt requiring interrupt wakeup. Many idle routines simply never enter such a halt, and so POLLING_NRFLAG can be always left set, completely eliminating resched IPIs when rescheduling the idle task. POLLING_NRFLAG width can be increased, to reduce the chance of resched IPIs. Signed-off-by: Nick Piggin Cc: Ingo Molnar Cc: Con Kolivas Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- Documentation/sched-arch.txt | 89 ++++++++++++++++++++++++++++++++++ arch/alpha/kernel/process.c | 10 ++-- arch/arm/kernel/process.c | 14 ++++-- arch/i386/kernel/apm.c | 20 +++++++- arch/i386/kernel/process.c | 64 +++++++++++------------- arch/ia64/kernel/process.c | 32 ++++++------ arch/parisc/kernel/process.c | 2 + arch/powerpc/platforms/iseries/setup.c | 10 +--- arch/powerpc/platforms/pseries/setup.c | 12 ++--- arch/ppc/kernel/idle.c | 20 ++++---- arch/ppc64/kernel/idle.c | 11 +---- arch/s390/kernel/process.c | 13 ++--- arch/sh/kernel/process.c | 12 ++--- arch/sh64/kernel/process.c | 14 ++---- arch/sparc/kernel/process.c | 35 ++++++------- arch/sparc64/kernel/process.c | 20 +++++--- arch/sparc64/kernel/smp.c | 13 +---- arch/x86_64/kernel/process.c | 67 ++++++++++++------------- drivers/acpi/processor_idle.c | 37 ++++++++------ kernel/sched.c | 21 +++++--- 20 files changed, 296 insertions(+), 220 deletions(-) create mode 100644 Documentation/sched-arch.txt (limited to 'arch/ppc64') diff --git a/Documentation/sched-arch.txt b/Documentation/sched-arch.txt new file mode 100644 index 000000000000..941615a9769b --- /dev/null +++ b/Documentation/sched-arch.txt @@ -0,0 +1,89 @@ + CPU Scheduler implementation hints for architecture specific code + + Nick Piggin, 2005 + +Context switch +============== +1. Runqueue locking +By default, the switch_to arch function is called with the runqueue +locked. This is usually not a problem unless switch_to may need to +take the runqueue lock. This is usually due to a wake up operation in +the context switch. See include/asm-ia64/system.h for an example. + +To request the scheduler call switch_to with the runqueue unlocked, +you must `#define __ARCH_WANT_UNLOCKED_CTXSW` in a header file +(typically the one where switch_to is defined). + +Unlocked context switches introduce only a very minor performance +penalty to the core scheduler implementation in the CONFIG_SMP case. + +2. Interrupt status +By default, the switch_to arch function is called with interrupts +disabled. Interrupts may be enabled over the call if it is likely to +introduce a significant interrupt latency by adding the line +`#define __ARCH_WANT_INTERRUPTS_ON_CTXSW` in the same place as for +unlocked context switches. This define also implies +`__ARCH_WANT_UNLOCKED_CTXSW`. See include/asm-arm/system.h for an +example. + + +CPU idle +======== +Your cpu_idle routines need to obey the following rules: + +1. Preempt should now disabled over idle routines. Should only + be enabled to call schedule() then disabled again. + +2. need_resched/TIF_NEED_RESCHED is only ever set, and will never + be cleared until the running task has called schedule(). Idle + threads need only ever query need_resched, and may never set or + clear it. + +3. When cpu_idle finds (need_resched() == 'true'), it should call + schedule(). It should not call schedule() otherwise. + +4. The only time interrupts need to be disabled when checking + need_resched is if we are about to sleep the processor until + the next interrupt (this doesn't provide any protection of + need_resched, it prevents losing an interrupt). + + 4a. Common problem with this type of sleep appears to be: + local_irq_disable(); + if (!need_resched()) { + local_irq_enable(); + *** resched interrupt arrives here *** + __asm__("sleep until next interrupt"); + } + +5. TIF_POLLING_NRFLAG can be set by idle routines that do not + need an interrupt to wake them up when need_resched goes high. + In other words, they must be periodically polling need_resched, + although it may be reasonable to do some background work or enter + a low CPU priority. + + 5a. If TIF_POLLING_NRFLAG is set, and we do decide to enter + an interrupt sleep, it needs to be cleared then a memory + barrier issued (followed by a test of need_resched with + interrupts disabled, as explained in 3). + +arch/i386/kernel/process.c has examples of both polling and +sleeping idle functions. + + +Possible arch/ problems +======================= + +Possible arch problems I found (and either tried to fix or didn't): + +h8300 - Is such sleeping racy vs interrupts? (See #4a). + The H8/300 manual I found indicates yes, however disabling IRQs + over the sleep mean only NMIs can wake it up, so can't fix easily + without doing spin waiting. + +ia64 - is safe_halt call racy vs interrupts? (does it sleep?) (See #4a) + +sh64 - Is sleeping racy vs interrupts? (See #4a) + +sparc - IRQs on at this point(?), change local_irq_save to _disable. + - TODO: needs secondary CPUs to disable preempt (See #1) + diff --git a/arch/alpha/kernel/process.c b/arch/alpha/kernel/process.c index eb20c3afff58..a8682612abc0 100644 --- a/arch/alpha/kernel/process.c +++ b/arch/alpha/kernel/process.c @@ -43,21 +43,17 @@ #include "proto.h" #include "pci_impl.h" -void default_idle(void) -{ - barrier(); -} - void cpu_idle(void) { + set_thread_flag(TIF_POLLING_NRFLAG); + while (1) { - void (*idle)(void) = default_idle; /* FIXME -- EV6 and LCA45 know how to power down the CPU. */ while (!need_resched()) - idle(); + cpu_relax(); schedule(); } } diff --git a/arch/arm/kernel/process.c b/arch/arm/kernel/process.c index 93dd92cc12f8..c0f6a119de3b 100644 --- a/arch/arm/kernel/process.c +++ b/arch/arm/kernel/process.c @@ -86,12 +86,16 @@ EXPORT_SYMBOL(pm_power_off); */ void default_idle(void) { - local_irq_disable(); - if (!need_resched() && !hlt_counter) { - timer_dyn_reprogram(); - arch_idle(); + if (hlt_counter) + cpu_relax(); + else { + local_irq_disable(); + if (!need_resched()) { + timer_dyn_reprogram(); + arch_idle(); + } + local_irq_enable(); } - local_irq_enable(); } /* diff --git a/arch/i386/kernel/apm.c b/arch/i386/kernel/apm.c index 86e80c551478..003548b8735f 100644 --- a/arch/i386/kernel/apm.c +++ b/arch/i386/kernel/apm.c @@ -769,8 +769,26 @@ static int set_system_power_state(u_short state) static int apm_do_idle(void) { u32 eax; + u8 ret = 0; + int idled = 0; + int polling; + + polling = test_thread_flag(TIF_POLLING_NRFLAG); + if (polling) { + clear_thread_flag(TIF_POLLING_NRFLAG); + smp_mb__after_clear_bit(); + } + if (!need_resched()) { + idled = 1; + ret = apm_bios_call_simple(APM_FUNC_IDLE, 0, 0, &eax); + } + if (polling) + set_thread_flag(TIF_POLLING_NRFLAG); + + if (!idled) + return 0; - if (apm_bios_call_simple(APM_FUNC_IDLE, 0, 0, &eax)) { + if (ret) { static unsigned long t; /* This always fails on some SMP boards running UP kernels. diff --git a/arch/i386/kernel/process.c b/arch/i386/kernel/process.c index 5296e284ea36..1cb261f225d5 100644 --- a/arch/i386/kernel/process.c +++ b/arch/i386/kernel/process.c @@ -99,14 +99,22 @@ EXPORT_SYMBOL(enable_hlt); */ void default_idle(void) { + local_irq_enable(); + if (!hlt_counter && boot_cpu_data.hlt_works_ok) { - local_irq_disable(); - if (!need_resched()) - safe_halt(); - else - local_irq_enable(); + clear_thread_flag(TIF_POLLING_NRFLAG); + smp_mb__after_clear_bit(); + while (!need_resched()) { + local_irq_disable(); + if (!need_resched()) + safe_halt(); + else + local_irq_enable(); + } + set_thread_flag(TIF_POLLING_NRFLAG); } else { - cpu_relax(); + while (!need_resched()) + cpu_relax(); } } #ifdef CONFIG_APM_MODULE @@ -120,29 +128,14 @@ EXPORT_SYMBOL(default_idle); */ static void poll_idle (void) { - int oldval; - local_irq_enable(); - /* - * Deal with another CPU just having chosen a thread to - * run here: - */ - oldval = test_and_clear_thread_flag(TIF_NEED_RESCHED); - - if (!oldval) { - set_thread_flag(TIF_POLLING_NRFLAG); - asm volatile( - "2:" - "testl %0, %1;" - "rep; nop;" - "je 2b;" - : : "i"(_TIF_NEED_RESCHED), "m" (current_thread_info()->flags)); - - clear_thread_flag(TIF_POLLING_NRFLAG); - } else { - set_need_resched(); - } + asm volatile( + "2:" + "testl %0, %1;" + "rep; nop;" + "je 2b;" + : : "i"(_TIF_NEED_RESCHED), "m" (current_thread_info()->flags)); } #ifdef CONFIG_HOTPLUG_CPU @@ -181,6 +174,8 @@ void cpu_idle(void) { int cpu = smp_processor_id(); + set_thread_flag(TIF_POLLING_NRFLAG); + /* endless idle loop with no priority at all */ while (1) { while (!need_resched()) { @@ -246,15 +241,12 @@ static void mwait_idle(void) { local_irq_enable(); - if (!need_resched()) { - set_thread_flag(TIF_POLLING_NRFLAG); - do { - __monitor((void *)¤t_thread_info()->flags, 0, 0); - if (need_resched()) - break; - __mwait(0, 0); - } while (!need_resched()); - clear_thread_flag(TIF_POLLING_NRFLAG); + while (!need_resched()) { + __monitor((void *)¤t_thread_info()->flags, 0, 0); + smp_mb(); + if (need_resched()) + break; + __mwait(0, 0); } } diff --git a/arch/ia64/kernel/process.c b/arch/ia64/kernel/process.c index 4c621fc3c3b9..640d6908f8ec 100644 --- a/arch/ia64/kernel/process.c +++ b/arch/ia64/kernel/process.c @@ -197,11 +197,15 @@ void default_idle (void) { local_irq_enable(); - while (!need_resched()) - if (can_do_pal_halt) - safe_halt(); - else + while (!need_resched()) { + if (can_do_pal_halt) { + local_irq_disable(); + if (!need_resched()) + safe_halt(); + local_irq_enable(); + } else cpu_relax(); + } } #ifdef CONFIG_HOTPLUG_CPU @@ -263,16 +267,16 @@ void __attribute__((noreturn)) cpu_idle (void) { void (*mark_idle)(int) = ia64_mark_idle; + int cpu = smp_processor_id(); + set_thread_flag(TIF_POLLING_NRFLAG); /* endless idle loop with no priority at all */ while (1) { + if (!need_resched()) { + void (*idle)(void); #ifdef CONFIG_SMP - if (!need_resched()) min_xtp(); #endif - while (!need_resched()) { - void (*idle)(void); - if (__get_cpu_var(cpu_idle_state)) __get_cpu_var(cpu_idle_state) = 0; @@ -284,19 +288,17 @@ cpu_idle (void) if (!idle) idle = default_idle; (*idle)(); - } - - if (mark_idle) - (*mark_idle)(0); - + if (mark_idle) + (*mark_idle)(0); #ifdef CONFIG_SMP - normal_xtp(); + normal_xtp(); #endif + } preempt_enable_no_resched(); schedule(); preempt_disable(); check_pgt_cache(); - if (cpu_is_offline(smp_processor_id())) + if (cpu_is_offline(cpu)) play_dead(); } } diff --git a/arch/parisc/kernel/process.c b/arch/parisc/kernel/process.c index f482f78de435..fee4f1f09adc 100644 --- a/arch/parisc/kernel/process.c +++ b/arch/parisc/kernel/process.c @@ -88,6 +88,8 @@ void default_idle(void) */ void cpu_idle(void) { + set_thread_flag(TIF_POLLING_NRFLAG); + /* endless idle loop with no priority at all */ while (1) { while (!need_resched()) diff --git a/arch/powerpc/platforms/iseries/setup.c b/arch/powerpc/platforms/iseries/setup.c index 0130f2619dac..7f8f0cda6a74 100644 --- a/arch/powerpc/platforms/iseries/setup.c +++ b/arch/powerpc/platforms/iseries/setup.c @@ -703,13 +703,10 @@ static void iseries_shared_idle(void) static void iseries_dedicated_idle(void) { long oldval; + set_thread_flag(TIF_POLLING_NRFLAG); while (1) { - oldval = test_and_clear_thread_flag(TIF_NEED_RESCHED); - - if (!oldval) { - set_thread_flag(TIF_POLLING_NRFLAG); - + if (!need_resched()) { while (!need_resched()) { ppc64_runlatch_off(); HMT_low(); @@ -722,9 +719,6 @@ static void iseries_dedicated_idle(void) } HMT_medium(); - clear_thread_flag(TIF_POLLING_NRFLAG); - } else { - set_need_resched(); } ppc64_runlatch_on(); diff --git a/arch/powerpc/platforms/pseries/setup.c b/arch/powerpc/platforms/pseries/setup.c index 4854f5eb5c3d..a093a0d4dd69 100644 --- a/arch/powerpc/platforms/pseries/setup.c +++ b/arch/powerpc/platforms/pseries/setup.c @@ -469,6 +469,7 @@ static inline void dedicated_idle_sleep(unsigned int cpu) * more. */ clear_thread_flag(TIF_POLLING_NRFLAG); + smp_mb__after_clear_bit(); /* * SMT dynamic mode. Cede will result in this thread going @@ -481,6 +482,7 @@ static inline void dedicated_idle_sleep(unsigned int cpu) cede_processor(); else local_irq_enable(); + set_thread_flag(TIF_POLLING_NRFLAG); } else { /* * Give the HV an opportunity at the processor, since we are @@ -492,11 +494,11 @@ static inline void dedicated_idle_sleep(unsigned int cpu) static void pseries_dedicated_idle(void) { - long oldval; struct paca_struct *lpaca = get_paca(); unsigned int cpu = smp_processor_id(); unsigned long start_snooze; unsigned long *smt_snooze_delay = &__get_cpu_var(smt_snooze_delay); + set_thread_flag(TIF_POLLING_NRFLAG); while (1) { /* @@ -505,10 +507,7 @@ static void pseries_dedicated_idle(void) */ lpaca->lppaca.idle = 1; - oldval = test_and_clear_thread_flag(TIF_NEED_RESCHED); - if (!oldval) { - set_thread_flag(TIF_POLLING_NRFLAG); - + if (!need_resched()) { start_snooze = __get_tb() + *smt_snooze_delay * tb_ticks_per_usec; @@ -531,9 +530,6 @@ static void pseries_dedicated_idle(void) } HMT_medium(); - clear_thread_flag(TIF_POLLING_NRFLAG); - } else { - set_need_resched(); } lpaca->lppaca.idle = 0; diff --git a/arch/ppc/kernel/idle.c b/arch/ppc/kernel/idle.c index a6141f05c919..3c4e4cb61074 100644 --- a/arch/ppc/kernel/idle.c +++ b/arch/ppc/kernel/idle.c @@ -63,18 +63,18 @@ void cpu_idle(void) int cpu = smp_processor_id(); for (;;) { - if (ppc_md.idle != NULL) - ppc_md.idle(); - else - default_idle(); - if (cpu_is_offline(cpu) && system_state == SYSTEM_RUNNING) - cpu_die(); - if (need_resched()) { - preempt_enable_no_resched(); - schedule(); - preempt_disable(); + while (need_resched()) { + if (ppc_md.idle != NULL) + ppc_md.idle(); + else + default_idle(); } + if (cpu_is_offline(cpu) && system_state == SYSTEM_RUNNING) + cpu_die(); + preempt_enable_no_resched(); + schedule(); + preempt_disable(); } } diff --git a/arch/ppc64/kernel/idle.c b/arch/ppc64/kernel/idle.c index 909ea669af91..715bc0e71e0f 100644 --- a/arch/ppc64/kernel/idle.c +++ b/arch/ppc64/kernel/idle.c @@ -34,15 +34,11 @@ extern void power4_idle(void); void default_idle(void) { - long oldval; unsigned int cpu = smp_processor_id(); + set_thread_flag(TIF_POLLING_NRFLAG); while (1) { - oldval = test_and_clear_thread_flag(TIF_NEED_RESCHED); - - if (!oldval) { - set_thread_flag(TIF_POLLING_NRFLAG); - + if (!need_resched()) { while (!need_resched() && !cpu_is_offline(cpu)) { ppc64_runlatch_off(); @@ -55,9 +51,6 @@ void default_idle(void) } HMT_medium(); - clear_thread_flag(TIF_POLLING_NRFLAG); - } else { - set_need_resched(); } ppc64_runlatch_on(); diff --git a/arch/s390/kernel/process.c b/arch/s390/kernel/process.c index 66ca5757e368..78b64fe5e7c2 100644 --- a/arch/s390/kernel/process.c +++ b/arch/s390/kernel/process.c @@ -99,14 +99,15 @@ void default_idle(void) { int cpu, rc; + /* CPU is going idle. */ + cpu = smp_processor_id(); + local_irq_disable(); - if (need_resched()) { + if (need_resched()) { local_irq_enable(); - return; - } + return; + } - /* CPU is going idle. */ - cpu = smp_processor_id(); rc = notifier_call_chain(&idle_chain, CPU_IDLE, (void *)(long) cpu); if (rc != NOTIFY_OK && rc != NOTIFY_DONE) BUG(); @@ -119,7 +120,7 @@ void default_idle(void) __ctl_set_bit(8, 15); #ifdef CONFIG_HOTPLUG_CPU - if (cpu_is_offline(smp_processor_id())) + if (cpu_is_offline(cpu)) cpu_die(); #endif diff --git a/arch/sh/kernel/process.c b/arch/sh/kernel/process.c index 1cbc26b796ad..fd4f240b833d 100644 --- a/arch/sh/kernel/process.c +++ b/arch/sh/kernel/process.c @@ -51,14 +51,13 @@ void enable_hlt(void) EXPORT_SYMBOL(enable_hlt); -void default_idle(void) +void cpu_idle(void) { /* endless idle loop with no priority at all */ while (1) { if (hlt_counter) { - while (1) - if (need_resched()) - break; + while (!need_resched()) + cpu_relax(); } else { while (!need_resched()) cpu_sleep(); @@ -70,11 +69,6 @@ void default_idle(void) } } -void cpu_idle(void) -{ - default_idle(); -} - void machine_restart(char * __unused) { /* SR.BL=1 and invoke address error to let CPU reset (manual reset) */ diff --git a/arch/sh64/kernel/process.c b/arch/sh64/kernel/process.c index 0c09537449b3..b95d04141855 100644 --- a/arch/sh64/kernel/process.c +++ b/arch/sh64/kernel/process.c @@ -307,23 +307,19 @@ __setup("hlt", hlt_setup); static inline void hlt(void) { - if (hlt_counter) - return; - __asm__ __volatile__ ("sleep" : : : "memory"); } /* * The idle loop on a uniprocessor SH.. */ -void default_idle(void) +void cpu_idle(void) { /* endless idle loop with no priority at all */ while (1) { if (hlt_counter) { - while (1) - if (need_resched()) - break; + while (!need_resched()) + cpu_relax(); } else { local_irq_disable(); while (!need_resched()) { @@ -338,11 +334,7 @@ void default_idle(void) schedule(); preempt_disable(); } -} -void cpu_idle(void) -{ - default_idle(); } void machine_restart(char * __unused) diff --git a/arch/sparc/kernel/process.c b/arch/sparc/kernel/process.c index c39f4d01096d..ea8647411462 100644 --- a/arch/sparc/kernel/process.c +++ b/arch/sparc/kernel/process.c @@ -67,13 +67,6 @@ extern void fpsave(unsigned long *, unsigned long *, void *, unsigned long *); struct task_struct *last_task_used_math = NULL; struct thread_info *current_set[NR_CPUS]; -/* - * default_idle is new in 2.5. XXX Review, currently stolen from sparc64. - */ -void default_idle(void) -{ -} - #ifndef CONFIG_SMP #define SUN4C_FAULT_HIGH 100 @@ -92,12 +85,11 @@ void cpu_idle(void) static unsigned long fps; unsigned long now; unsigned long faults; - unsigned long flags; extern unsigned long sun4c_kernel_faults; extern void sun4c_grow_kernel_ring(void); - local_irq_save(flags); + local_irq_disable(); now = jiffies; count -= (now - last_jiffies); last_jiffies = now; @@ -113,13 +105,16 @@ void cpu_idle(void) sun4c_grow_kernel_ring(); } } - local_irq_restore(flags); + local_irq_enable(); } - while((!need_resched()) && pm_idle) { - (*pm_idle)(); + if (pm_idle) { + while (!need_resched()) + (*pm_idle)(); + } else { + while (!need_resched()) + cpu_relax(); } - preempt_enable_no_resched(); schedule(); preempt_disable(); @@ -132,15 +127,15 @@ void cpu_idle(void) /* This is being executed in task 0 'user space'. */ void cpu_idle(void) { + set_thread_flag(TIF_POLLING_NRFLAG); /* endless idle loop with no priority at all */ while(1) { - if(need_resched()) { - preempt_enable_no_resched(); - schedule(); - preempt_disable(); - check_pgt_cache(); - } - barrier(); /* or else gcc optimizes... */ + while (!need_resched()) + cpu_relax(); + preempt_enable_no_resched(); + schedule(); + preempt_disable(); + check_pgt_cache(); } } diff --git a/arch/sparc64/kernel/process.c b/arch/sparc64/kernel/process.c index 2f89206e008f..02f9dec1d459 100644 --- a/arch/sparc64/kernel/process.c +++ b/arch/sparc64/kernel/process.c @@ -85,23 +85,31 @@ void cpu_idle(void) /* * the idle loop on a UltraMultiPenguin... + * + * TIF_POLLING_NRFLAG is set because we do not sleep the cpu + * inside of the idler task, so an interrupt is not needed + * to get a clean fast response. + * + * XXX Reverify this assumption... -DaveM + * + * Addendum: We do want it to do something for the signal + * delivery case, we detect that by just seeing + * if we are trying to send this to an idler or not. */ -#define idle_me_harder() (cpu_data(smp_processor_id()).idle_volume += 1) -#define unidle_me() (cpu_data(smp_processor_id()).idle_volume = 0) void cpu_idle(void) { + cpuinfo_sparc *cpuinfo = &local_cpu_data(); set_thread_flag(TIF_POLLING_NRFLAG); + while(1) { if (need_resched()) { - unidle_me(); - clear_thread_flag(TIF_POLLING_NRFLAG); + cpuinfo->idle_volume = 0; preempt_enable_no_resched(); schedule(); preempt_disable(); - set_thread_flag(TIF_POLLING_NRFLAG); check_pgt_cache(); } - idle_me_harder(); + cpuinfo->idle_volume++; /* The store ordering is so that IRQ handlers on * other cpus see our increasing idleness for the buddy diff --git a/arch/sparc64/kernel/smp.c b/arch/sparc64/kernel/smp.c index 8aca4b1dc04e..797a65493fb8 100644 --- a/arch/sparc64/kernel/smp.c +++ b/arch/sparc64/kernel/smp.c @@ -1152,20 +1152,9 @@ void __init smp_cpus_done(unsigned int max_cpus) (bogosum/(5000/HZ))%100); } -/* This needn't do anything as we do not sleep the cpu - * inside of the idler task, so an interrupt is not needed - * to get a clean fast response. - * - * XXX Reverify this assumption... -DaveM - * - * Addendum: We do want it to do something for the signal - * delivery case, we detect that by just seeing - * if we are trying to send this to an idler or not. - */ void smp_send_reschedule(int cpu) { - if (cpu_data(cpu).idle_volume == 0) - smp_receive_signal(cpu); + smp_receive_signal(cpu); } /* This is a nop because we capture all other cpus diff --git a/arch/x86_64/kernel/process.c b/arch/x86_64/kernel/process.c index 571f9fe490ce..59be85d9a4bc 100644 --- a/arch/x86_64/kernel/process.c +++ b/arch/x86_64/kernel/process.c @@ -86,12 +86,22 @@ EXPORT_SYMBOL(enable_hlt); */ void default_idle(void) { + local_irq_enable(); + if (!atomic_read(&hlt_counter)) { - local_irq_disable(); - if (!need_resched()) - safe_halt(); - else - local_irq_enable(); + clear_thread_flag(TIF_POLLING_NRFLAG); + smp_mb__after_clear_bit(); + while (!need_resched()) { + local_irq_disable(); + if (!need_resched()) + safe_halt(); + else + local_irq_enable(); + } + set_thread_flag(TIF_POLLING_NRFLAG); + } else { + while (!need_resched()) + cpu_relax(); } } @@ -102,30 +112,16 @@ void default_idle(void) */ static void poll_idle (void) { - int oldval; - local_irq_enable(); - /* - * Deal with another CPU just having chosen a thread to - * run here: - */ - oldval = test_and_clear_thread_flag(TIF_NEED_RESCHED); - - if (!oldval) { - set_thread_flag(TIF_POLLING_NRFLAG); - asm volatile( - "2:" - "testl %0,%1;" - "rep; nop;" - "je 2b;" - : : - "i" (_TIF_NEED_RESCHED), - "m" (current_thread_info()->flags)); - clear_thread_flag(TIF_POLLING_NRFLAG); - } else { - set_need_resched(); - } + asm volatile( + "2:" + "testl %0,%1;" + "rep; nop;" + "je 2b;" + : : + "i" (_TIF_NEED_RESCHED), + "m" (current_thread_info()->flags)); } void cpu_idle_wait(void) @@ -187,6 +183,8 @@ static inline void play_dead(void) */ void cpu_idle (void) { + set_thread_flag(TIF_POLLING_NRFLAG); + /* endless idle loop with no priority at all */ while (1) { while (!need_resched()) { @@ -221,15 +219,12 @@ static void mwait_idle(void) { local_irq_enable(); - if (!need_resched()) { - set_thread_flag(TIF_POLLING_NRFLAG); - do { - __monitor((void *)¤t_thread_info()->flags, 0, 0); - if (need_resched()) - break; - __mwait(0, 0); - } while (!need_resched()); - clear_thread_flag(TIF_POLLING_NRFLAG); + while (!need_resched()) { + __monitor((void *)¤t_thread_info()->flags, 0, 0); + smp_mb(); + if (need_resched()) + break; + __mwait(0, 0); } } diff --git a/drivers/acpi/processor_idle.c b/drivers/acpi/processor_idle.c index 161db4acfb91..573b6a97bb1f 100644 --- a/drivers/acpi/processor_idle.c +++ b/drivers/acpi/processor_idle.c @@ -167,6 +167,19 @@ acpi_processor_power_activate(struct acpi_processor *pr, return; } +static void acpi_safe_halt(void) +{ + int polling = test_thread_flag(TIF_POLLING_NRFLAG); + if (polling) { + clear_thread_flag(TIF_POLLING_NRFLAG); + smp_mb__after_clear_bit(); + } + if (!need_resched()) + safe_halt(); + if (polling) + set_thread_flag(TIF_POLLING_NRFLAG); +} + static atomic_t c3_cpu_count; static void acpi_processor_idle(void) @@ -177,7 +190,7 @@ static void acpi_processor_idle(void) int sleep_ticks = 0; u32 t1, t2 = 0; - pr = processors[raw_smp_processor_id()]; + pr = processors[smp_processor_id()]; if (!pr) return; @@ -197,8 +210,13 @@ static void acpi_processor_idle(void) } cx = pr->power.state; - if (!cx) - goto easy_out; + if (!cx) { + if (pm_idle_save) + pm_idle_save(); + else + acpi_safe_halt(); + return; + } /* * Check BM Activity @@ -278,7 +296,8 @@ static void acpi_processor_idle(void) if (pm_idle_save) pm_idle_save(); else - safe_halt(); + acpi_safe_halt(); + /* * TBD: Can't get time duration while in C1, as resumes * go to an ISR rather than here. Need to instrument @@ -414,16 +433,6 @@ static void acpi_processor_idle(void) */ if (next_state != pr->power.state) acpi_processor_power_activate(pr, next_state); - - return; - - easy_out: - /* do C1 instead of busy loop */ - if (pm_idle_save) - pm_idle_save(); - else - safe_halt(); - return; } static int acpi_processor_set_power_policy(struct acpi_processor *pr) diff --git a/kernel/sched.c b/kernel/sched.c index 0f2def822296..ac3f5cc3bb51 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -864,21 +864,28 @@ static void deactivate_task(struct task_struct *p, runqueue_t *rq) #ifdef CONFIG_SMP static void resched_task(task_t *p) { - int need_resched, nrpolling; + int cpu; assert_spin_locked(&task_rq(p)->lock); - /* minimise the chance of sending an interrupt to poll_idle() */ - nrpolling = test_tsk_thread_flag(p,TIF_POLLING_NRFLAG); - need_resched = test_and_set_tsk_thread_flag(p,TIF_NEED_RESCHED); - nrpolling |= test_tsk_thread_flag(p,TIF_POLLING_NRFLAG); + if (unlikely(test_tsk_thread_flag(p, TIF_NEED_RESCHED))) + return; + + set_tsk_thread_flag(p, TIF_NEED_RESCHED); + + cpu = task_cpu(p); + if (cpu == smp_processor_id()) + return; - if (!need_resched && !nrpolling && (task_cpu(p) != smp_processor_id())) - smp_send_reschedule(task_cpu(p)); + /* NEED_RESCHED must be visible before we test POLLING_NRFLAG */ + smp_mb(); + if (!test_tsk_thread_flag(p, TIF_POLLING_NRFLAG)) + smp_send_reschedule(cpu); } #else static inline void resched_task(task_t *p) { + assert_spin_locked(&task_rq(p)->lock); set_tsk_need_resched(p); } #endif -- cgit v1.2.3