From d53e4307c2f3856167407a1d9b8f8fa001286066 Mon Sep 17 00:00:00 2001 From: Giuseppe CAVALLARO Date: Wed, 17 Nov 2010 06:50:17 +0000 Subject: sh: Use GCC __builtin_prefetch() to implement prefetch(). GCC's __builtin_prefetch() was introduced a long time ago, all supported GCC versions have it. So this patch is to use it for implementing the prefetch on SH2A and SH4. The current prefetch implementation is almost equivalent with __builtin_prefetch. The third parameter in the __builtin_prefetch is the locality that it's not supported on SH architectures. It has been set to three and it should be verified if it's suitable for SH2A as well. I didn't test on this architecture. The builtin usage should be more efficient that an __asm__ because less barriers, and because the compiler doesn't see the inst as a "black box" allowing better code generation. This has been already done on other architectures (see the commit: 0453fb3c528c5eb3483441a466b24a4cb409eec5). Many thanks to Christian Bruel for his support on evaluate the impact of the gcc built-in on SH4 arch. No regressions found while testing with LMbench on STLinux targets. Signed-off-by: Giuseppe Cavallaro Signed-off-by: Stuart Menefy Signed-off-by: Paul Mundt --- arch/sh/include/asm/processor_32.h | 7 +++++-- 1 file changed, 5 insertions(+), 2 deletions(-) (limited to 'arch/sh/include') diff --git a/arch/sh/include/asm/processor_32.h b/arch/sh/include/asm/processor_32.h index 46d5179c9f49..e3c73cdd8c90 100644 --- a/arch/sh/include/asm/processor_32.h +++ b/arch/sh/include/asm/processor_32.h @@ -199,10 +199,13 @@ extern unsigned long get_wchan(struct task_struct *p); #define ARCH_HAS_PREFETCHW static inline void prefetch(void *x) { - __asm__ __volatile__ ("pref @%0\n\t" : : "r" (x) : "memory"); + __builtin_prefetch(x, 0, 3); } -#define prefetchw(x) prefetch(x) +static inline void prefetchw(void *x) +{ + __builtin_prefetch(x, 1, 3); +} #endif #endif /* __KERNEL__ */ -- cgit v1.2.3 From 4bd5d259e451d35380de2fcd1a3510ae22d000a3 Mon Sep 17 00:00:00 2001 From: Kuninori Morimoto Date: Fri, 26 Nov 2010 09:40:22 +0000 Subject: sh: Add div6_reparent_clks to clock framework for FSI Current clk_ops doesn't support .init which is used to select external clock on ecovec without CONFIG_SH_CLK_CPG_LEGACY. To solve this problem, this patch add div6_reparent_clks to clock-sh7724. This patch solve compile error too. Signed-off-by: Kuninori Morimoto Signed-off-by: Paul Mundt --- arch/sh/boards/mach-ecovec24/setup.c | 34 ++++---------------------- arch/sh/include/cpu-sh4/cpu/sh7724.h | 3 +++ arch/sh/kernel/cpu/sh4a/clock-sh7724.c | 44 ++++++++++++++++++++++++++++++---- 3 files changed, 46 insertions(+), 35 deletions(-) (limited to 'arch/sh/include') diff --git a/arch/sh/boards/mach-ecovec24/setup.c b/arch/sh/boards/mach-ecovec24/setup.c index 2eaeb9e59585..f48c492a68d3 100644 --- a/arch/sh/boards/mach-ecovec24/setup.c +++ b/arch/sh/boards/mach-ecovec24/setup.c @@ -720,32 +720,6 @@ static struct platform_device camera_devices[] = { }; /* FSI */ -/* - * FSI-B use external clock which came from da7210. - * So, we should change parent of fsi - */ -#define FCLKBCR 0xa415000c -static void fsimck_init(struct clk *clk) -{ - u32 status = __raw_readl(clk->enable_reg); - - /* use external clock */ - status &= ~0x000000ff; - status |= 0x00000080; - - __raw_writel(status, clk->enable_reg); -} - -static struct clk_ops fsimck_clk_ops = { - .init = fsimck_init, -}; - -static struct clk fsimckb_clk = { - .ops = &fsimck_clk_ops, - .enable_reg = (void __iomem *)FCLKBCR, - .rate = 0, /* unknown */ -}; - static struct sh_fsi_platform_info fsi_info = { .portb_flags = SH_FSI_BRS_INV | SH_FSI_OUT_SLAVE_MODE | @@ -1264,10 +1238,10 @@ static int __init arch_setup(void) /* change parent of FSI B */ clk = clk_get(NULL, "fsib_clk"); if (!IS_ERR(clk)) { - clk_register(&fsimckb_clk); - clk_set_parent(clk, &fsimckb_clk); - clk_set_rate(clk, 11000); - clk_set_rate(&fsimckb_clk, 11000); + /* 48kHz dummy clock was used to make sure 1/1 divide */ + clk_set_rate(&sh7724_fsimckb_clk, 48000); + clk_set_parent(clk, &sh7724_fsimckb_clk); + clk_set_rate(clk, 48000); clk_put(clk); } diff --git a/arch/sh/include/cpu-sh4/cpu/sh7724.h b/arch/sh/include/cpu-sh4/cpu/sh7724.h index 4c27b68789b3..7eb435999426 100644 --- a/arch/sh/include/cpu-sh4/cpu/sh7724.h +++ b/arch/sh/include/cpu-sh4/cpu/sh7724.h @@ -303,4 +303,7 @@ enum { SHDMA_SLAVE_SDHI1_RX, }; +extern struct clk sh7724_fsimcka_clk; +extern struct clk sh7724_fsimckb_clk; + #endif /* __ASM_SH7724_H__ */ diff --git a/arch/sh/kernel/cpu/sh4a/clock-sh7724.c b/arch/sh/kernel/cpu/sh4a/clock-sh7724.c index 0fe2e9329cb2..271c0b325a9a 100644 --- a/arch/sh/kernel/cpu/sh4a/clock-sh7724.c +++ b/arch/sh/kernel/cpu/sh4a/clock-sh7724.c @@ -111,12 +111,21 @@ static struct clk div3_clk = { .parent = &pll_clk, }; +/* External input clock (pin name: FSIMCKA/FSIMCKB ) */ +struct clk sh7724_fsimcka_clk = { +}; + +struct clk sh7724_fsimckb_clk = { +}; + static struct clk *main_clks[] = { &r_clk, &extal_clk, &fll_clk, &pll_clk, &div3_clk, + &sh7724_fsimcka_clk, + &sh7724_fsimckb_clk, }; static void div4_kick(struct clk *clk) @@ -154,16 +163,38 @@ struct clk div4_clks[DIV4_NR] = { [DIV4_M1] = DIV4(FRQCRB, 4, 0x2f7c, CLK_ENABLE_ON_INIT), }; -enum { DIV6_V, DIV6_FA, DIV6_FB, DIV6_I, DIV6_S, DIV6_NR }; +enum { DIV6_V, DIV6_I, DIV6_S, DIV6_NR }; static struct clk div6_clks[DIV6_NR] = { [DIV6_V] = SH_CLK_DIV6(&div3_clk, VCLKCR, 0), - [DIV6_FA] = SH_CLK_DIV6(&div3_clk, FCLKACR, 0), - [DIV6_FB] = SH_CLK_DIV6(&div3_clk, FCLKBCR, 0), [DIV6_I] = SH_CLK_DIV6(&div3_clk, IRDACLKCR, 0), [DIV6_S] = SH_CLK_DIV6(&div3_clk, SPUCLKCR, CLK_ENABLE_ON_INIT), }; +enum { DIV6_FA, DIV6_FB, DIV6_REPARENT_NR }; + +/* Indices are important - they are the actual src selecting values */ +static struct clk *fclkacr_parent[] = { + [0] = &div3_clk, + [1] = NULL, + [2] = &sh7724_fsimcka_clk, + [3] = NULL, +}; + +static struct clk *fclkbcr_parent[] = { + [0] = &div3_clk, + [1] = NULL, + [2] = &sh7724_fsimckb_clk, + [3] = NULL, +}; + +static struct clk div6_reparent_clks[DIV6_REPARENT_NR] = { + [DIV6_FA] = SH_CLK_DIV6_EXT(&div3_clk, FCLKACR, 0, + fclkacr_parent, ARRAY_SIZE(fclkacr_parent), 6, 2), + [DIV6_FB] = SH_CLK_DIV6_EXT(&div3_clk, FCLKBCR, 0, + fclkbcr_parent, ARRAY_SIZE(fclkbcr_parent), 6, 2), +}; + static struct clk mstp_clks[HWBLK_NR] = { SH_HWBLK_CLK(HWBLK_TLB, &div4_clks[DIV4_I], CLK_ENABLE_ON_INIT), SH_HWBLK_CLK(HWBLK_IC, &div4_clks[DIV4_I], CLK_ENABLE_ON_INIT), @@ -240,8 +271,8 @@ static struct clk_lookup lookups[] = { /* DIV6 clocks */ CLKDEV_CON_ID("video_clk", &div6_clks[DIV6_V]), - CLKDEV_CON_ID("fsia_clk", &div6_clks[DIV6_FA]), - CLKDEV_CON_ID("fsib_clk", &div6_clks[DIV6_FB]), + CLKDEV_CON_ID("fsia_clk", &div6_reparent_clks[DIV6_FA]), + CLKDEV_CON_ID("fsib_clk", &div6_reparent_clks[DIV6_FB]), CLKDEV_CON_ID("irda_clk", &div6_clks[DIV6_I]), CLKDEV_CON_ID("spu_clk", &div6_clks[DIV6_S]), @@ -375,6 +406,9 @@ int __init arch_clk_init(void) if (!ret) ret = sh_clk_div6_register(div6_clks, DIV6_NR); + if (!ret) + ret = sh_clk_div6_reparent_register(div6_reparent_clks, DIV6_REPARENT_NR); + if (!ret) ret = sh_hwblk_clk_register(mstp_clks, HWBLK_NR); -- cgit v1.2.3 From 55661fc1f105ed75852e937bf8ea408270eb0cca Mon Sep 17 00:00:00 2001 From: Paul Mundt Date: Wed, 1 Dec 2010 15:39:51 +0900 Subject: sh: Assume new page cache pages have dirty dcache lines. This follows the ARM change c01778001a4f5ad9c62d882776235f3f31922fdd ("ARM: 6379/1: Assume new page cache pages have dirty D-cache") for the same rationale: There are places in Linux where writes to newly allocated page cache pages happen without a subsequent call to flush_dcache_page() (several PIO drivers including USB HCD). This patch changes the meaning of PG_arch_1 to be PG_dcache_clean and always flush the D-cache for a newly mapped page in update_mmu_cache(). This addresses issues seen with executing binaries from MMC, in addition to some of the other HCDs that don't explicitly do cache management for their pipe-in buffers. Requested-by: Yoshihiro Shimoda Signed-off-by: Paul Mundt --- arch/sh/include/asm/cacheflush.h | 2 +- arch/sh/mm/cache-sh4.c | 4 ++-- arch/sh/mm/cache-sh7705.c | 2 +- arch/sh/mm/cache.c | 14 +++++++------- arch/sh/mm/kmap.c | 2 +- 5 files changed, 12 insertions(+), 12 deletions(-) (limited to 'arch/sh/include') diff --git a/arch/sh/include/asm/cacheflush.h b/arch/sh/include/asm/cacheflush.h index 1f4e562c5e8c..82e1eabeac98 100644 --- a/arch/sh/include/asm/cacheflush.h +++ b/arch/sh/include/asm/cacheflush.h @@ -96,7 +96,7 @@ void kmap_coherent_init(void); void *kmap_coherent(struct page *page, unsigned long addr); void kunmap_coherent(void *kvaddr); -#define PG_dcache_dirty PG_arch_1 +#define PG_dcache_clean PG_arch_1 void cpu_cache_init(void); diff --git a/arch/sh/mm/cache-sh4.c b/arch/sh/mm/cache-sh4.c index 2cfae81914aa..92eb98633ab0 100644 --- a/arch/sh/mm/cache-sh4.c +++ b/arch/sh/mm/cache-sh4.c @@ -114,7 +114,7 @@ static void sh4_flush_dcache_page(void *arg) struct address_space *mapping = page_mapping(page); if (mapping && !mapping_mapped(mapping)) - set_bit(PG_dcache_dirty, &page->flags); + clear_bit(PG_dcache_clean, &page->flags); else #endif flush_cache_one(CACHE_OC_ADDRESS_ARRAY | @@ -239,7 +239,7 @@ static void sh4_flush_cache_page(void *args) * another ASID than the current one. */ map_coherent = (current_cpu_data.dcache.n_aliases && - !test_bit(PG_dcache_dirty, &page->flags) && + test_bit(PG_dcache_clean, &page->flags) && page_mapped(page)); if (map_coherent) vaddr = kmap_coherent(page, address); diff --git a/arch/sh/mm/cache-sh7705.c b/arch/sh/mm/cache-sh7705.c index f498da1cce7a..7729cca727eb 100644 --- a/arch/sh/mm/cache-sh7705.c +++ b/arch/sh/mm/cache-sh7705.c @@ -139,7 +139,7 @@ static void sh7705_flush_dcache_page(void *arg) struct address_space *mapping = page_mapping(page); if (mapping && !mapping_mapped(mapping)) - set_bit(PG_dcache_dirty, &page->flags); + clear_bit(PG_dcache_clean, &page->flags); else __flush_dcache_page(__pa(page_address(page))); } diff --git a/arch/sh/mm/cache.c b/arch/sh/mm/cache.c index ba401d137bb9..88d3dc3d30d5 100644 --- a/arch/sh/mm/cache.c +++ b/arch/sh/mm/cache.c @@ -60,14 +60,14 @@ void copy_to_user_page(struct vm_area_struct *vma, struct page *page, unsigned long len) { if (boot_cpu_data.dcache.n_aliases && page_mapped(page) && - !test_bit(PG_dcache_dirty, &page->flags)) { + test_bit(PG_dcache_clean, &page->flags)) { void *vto = kmap_coherent(page, vaddr) + (vaddr & ~PAGE_MASK); memcpy(vto, src, len); kunmap_coherent(vto); } else { memcpy(dst, src, len); if (boot_cpu_data.dcache.n_aliases) - set_bit(PG_dcache_dirty, &page->flags); + clear_bit(PG_dcache_clean, &page->flags); } if (vma->vm_flags & VM_EXEC) @@ -79,14 +79,14 @@ void copy_from_user_page(struct vm_area_struct *vma, struct page *page, unsigned long len) { if (boot_cpu_data.dcache.n_aliases && page_mapped(page) && - !test_bit(PG_dcache_dirty, &page->flags)) { + test_bit(PG_dcache_clean, &page->flags)) { void *vfrom = kmap_coherent(page, vaddr) + (vaddr & ~PAGE_MASK); memcpy(dst, vfrom, len); kunmap_coherent(vfrom); } else { memcpy(dst, src, len); if (boot_cpu_data.dcache.n_aliases) - set_bit(PG_dcache_dirty, &page->flags); + clear_bit(PG_dcache_clean, &page->flags); } } @@ -98,7 +98,7 @@ void copy_user_highpage(struct page *to, struct page *from, vto = kmap_atomic(to, KM_USER1); if (boot_cpu_data.dcache.n_aliases && page_mapped(from) && - !test_bit(PG_dcache_dirty, &from->flags)) { + test_bit(PG_dcache_clean, &from->flags)) { vfrom = kmap_coherent(from, vaddr); copy_page(vto, vfrom); kunmap_coherent(vfrom); @@ -141,7 +141,7 @@ void __update_cache(struct vm_area_struct *vma, page = pfn_to_page(pfn); if (pfn_valid(pfn)) { - int dirty = test_and_clear_bit(PG_dcache_dirty, &page->flags); + int dirty = !test_and_set_bit(PG_dcache_clean, &page->flags); if (dirty) __flush_purge_region(page_address(page), PAGE_SIZE); } @@ -153,7 +153,7 @@ void __flush_anon_page(struct page *page, unsigned long vmaddr) if (pages_do_alias(addr, vmaddr)) { if (boot_cpu_data.dcache.n_aliases && page_mapped(page) && - !test_bit(PG_dcache_dirty, &page->flags)) { + test_bit(PG_dcache_clean, &page->flags)) { void *kaddr; kaddr = kmap_coherent(page, vmaddr); diff --git a/arch/sh/mm/kmap.c b/arch/sh/mm/kmap.c index 15d74ea42094..ec29e14ec5a8 100644 --- a/arch/sh/mm/kmap.c +++ b/arch/sh/mm/kmap.c @@ -34,7 +34,7 @@ void *kmap_coherent(struct page *page, unsigned long addr) enum fixed_addresses idx; unsigned long vaddr; - BUG_ON(test_bit(PG_dcache_dirty, &page->flags)); + BUG_ON(!test_bit(PG_dcache_clean, &page->flags)); pagefault_disable(); -- cgit v1.2.3 From 21b6e4c7106b2d68a6710506d8706608272fd78b Mon Sep 17 00:00:00 2001 From: Carmelo AMOROSO Date: Mon, 13 Dec 2010 10:20:26 +0000 Subject: sh: wire up accept4 syscall (non-multiplexed path) Signed-off-by: Carmelo Amoroso Signed-off-by: Paul Mundt --- arch/sh/include/asm/unistd_32.h | 3 ++- arch/sh/kernel/syscalls_32.S | 1 + 2 files changed, 3 insertions(+), 1 deletion(-) (limited to 'arch/sh/include') diff --git a/arch/sh/include/asm/unistd_32.h b/arch/sh/include/asm/unistd_32.h index 903cd618eb74..d6741fca89a4 100644 --- a/arch/sh/include/asm/unistd_32.h +++ b/arch/sh/include/asm/unistd_32.h @@ -368,8 +368,9 @@ #define __NR_sendmsg 355 #define __NR_recvmsg 356 #define __NR_recvmmsg 357 +#define __NR_accept4 358 -#define NR_syscalls 358 +#define NR_syscalls 359 #ifdef __KERNEL__ diff --git a/arch/sh/kernel/syscalls_32.S b/arch/sh/kernel/syscalls_32.S index e872e81add8a..6fc347ebe59d 100644 --- a/arch/sh/kernel/syscalls_32.S +++ b/arch/sh/kernel/syscalls_32.S @@ -375,3 +375,4 @@ ENTRY(sys_call_table) .long sys_sendmsg /* 355 */ .long sys_recvmsg .long sys_recvmmsg + .long sys_accept4 -- cgit v1.2.3