Merge tag 'riscv-for-linus-6.7-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/riscv/linux

Pull RISC-V updates from Palmer Dabbelt:

 - Support for cbo.zero in userspace

 - Support for CBOs on ACPI-based systems

 - A handful of improvements for the T-Head cache flushing ops

 - Support for software shadow call stacks

 - Various cleanups and fixes

* tag 'riscv-for-linus-6.7-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/riscv/linux: (31 commits)
  RISC-V: hwprobe: Fix vDSO SIGSEGV
  riscv: configs: defconfig: Enable configs required for RZ/Five SoC
  riscv: errata: prefix T-Head mnemonics with th.
  riscv: put interrupt entries into .irqentry.text
  riscv: mm: Update the comment of CONFIG_PAGE_OFFSET
  riscv: Using TOOLCHAIN_HAS_ZIHINTPAUSE marco replace zihintpause
  riscv/mm: Fix the comment for swap pte format
  RISC-V: clarify the QEMU workaround in ISA parser
  riscv: correct pt_level name via pgtable_l5/4_enabled
  RISC-V: Provide pgtable_l5_enabled on rv32
  clocksource: timer-riscv: Increase rating of clock_event_device for Sstc
  clocksource: timer-riscv: Don't enable/disable timer interrupt
  lkdtm: Fix CFI_BACKWARD on RISC-V
  riscv: Use separate IRQ shadow call stacks
  riscv: Implement Shadow Call Stack
  riscv: Move global pointer loading to a macro
  riscv: Deduplicate IRQ stack switching
  riscv: VMAP_STACK overflow detection thread-safe
  RISC-V: cacheflush: Initialize CBO variables on ACPI systems
  RISC-V: ACPI: RHCT: Add function to get CBO block sizes
  ...

3 files changed, 116 insertions, 7 deletions

diff --git a/drivers/acpi/riscv/rhct.c b/drivers/acpi/riscv/rhct.c
index b280b3e9c7d9..caa2c16e1697 100644
--- a/drivers/acpi/riscv/rhct.c
+++ b/drivers/acpi/riscv/rhct.c
@@ -8,8 +8,9 @@
 #define pr_fmt(fmt)     "ACPI: RHCT: " fmt
 
 #include <linux/acpi.h>
+#include <linux/bits.h>
 
-static struct acpi_table_header *acpi_get_rhct(void)
+static struct acpi_table_rhct *acpi_get_rhct(void)
 {
 	static struct acpi_table_header *rhct;
 	acpi_status status;
@@ -26,7 +27,7 @@ static struct acpi_table_header *acpi_get_rhct(void)
 		}
 	}
 
-	return rhct;
+	return (struct acpi_table_rhct *)rhct;
 }
 
 /*
@@ -48,7 +49,7 @@ int acpi_get_riscv_isa(struct acpi_table_header *table, unsigned int cpu, const
 	BUG_ON(acpi_disabled);
 
 	if (!table) {
-		rhct = (struct acpi_table_rhct *)acpi_get_rhct();
+		rhct = acpi_get_rhct();
 		if (!rhct)
 			return -ENOENT;
 	} else {
@@ -81,3 +82,89 @@ int acpi_get_riscv_isa(struct acpi_table_header *table, unsigned int cpu, const
 
 	return -1;
 }
+
+static void acpi_parse_hart_info_cmo_node(struct acpi_table_rhct *rhct,
+					  struct acpi_rhct_hart_info *hart_info,
+					  u32 *cbom_size, u32 *cboz_size, u32 *cbop_size)
+{
+	u32 size_hartinfo = sizeof(struct acpi_rhct_hart_info);
+	u32 size_hdr = sizeof(struct acpi_rhct_node_header);
+	struct acpi_rhct_node_header *ref_node;
+	struct acpi_rhct_cmo_node *cmo_node;
+	u32 *hart_info_node_offset;
+
+	hart_info_node_offset = ACPI_ADD_PTR(u32, hart_info, size_hartinfo);
+	for (int i = 0; i < hart_info->num_offsets; i++) {
+		ref_node = ACPI_ADD_PTR(struct acpi_rhct_node_header,
+					rhct, hart_info_node_offset[i]);
+		if (ref_node->type == ACPI_RHCT_NODE_TYPE_CMO) {
+			cmo_node = ACPI_ADD_PTR(struct acpi_rhct_cmo_node,
+						ref_node, size_hdr);
+			if (cbom_size && cmo_node->cbom_size <= 30) {
+				if (!*cbom_size)
+					*cbom_size = BIT(cmo_node->cbom_size);
+				else if (*cbom_size != BIT(cmo_node->cbom_size))
+					pr_warn("CBOM size is not the same across harts\n");
+			}
+
+			if (cboz_size && cmo_node->cboz_size <= 30) {
+				if (!*cboz_size)
+					*cboz_size = BIT(cmo_node->cboz_size);
+				else if (*cboz_size != BIT(cmo_node->cboz_size))
+					pr_warn("CBOZ size is not the same across harts\n");
+			}
+
+			if (cbop_size && cmo_node->cbop_size <= 30) {
+				if (!*cbop_size)
+					*cbop_size = BIT(cmo_node->cbop_size);
+				else if (*cbop_size != BIT(cmo_node->cbop_size))
+					pr_warn("CBOP size is not the same across harts\n");
+			}
+		}
+	}
+}
+
+/*
+ * During early boot, the caller should call acpi_get_table() and pass its pointer to
+ * these functions (and free up later). At run time, since this table can be used
+ * multiple times, pass NULL so that the table remains in memory.
+ */
+void acpi_get_cbo_block_size(struct acpi_table_header *table, u32 *cbom_size,
+			     u32 *cboz_size, u32 *cbop_size)
+{
+	u32 size_hdr = sizeof(struct acpi_rhct_node_header);
+	struct acpi_rhct_node_header *node, *end;
+	struct acpi_rhct_hart_info *hart_info;
+	struct acpi_table_rhct *rhct;
+
+	if (acpi_disabled)
+		return;
+
+	if (table) {
+		rhct = (struct acpi_table_rhct *)table;
+	} else {
+		rhct = acpi_get_rhct();
+		if (!rhct)
+			return;
+	}
+
+	if (cbom_size)
+		*cbom_size = 0;
+
+	if (cboz_size)
+		*cboz_size = 0;
+
+	if (cbop_size)
+		*cbop_size = 0;
+
+	end = ACPI_ADD_PTR(struct acpi_rhct_node_header, rhct, rhct->header.length);
+	for (node = ACPI_ADD_PTR(struct acpi_rhct_node_header, rhct, rhct->node_offset);
+	     node < end;
+	     node = ACPI_ADD_PTR(struct acpi_rhct_node_header, node, node->length)) {
+		if (node->type == ACPI_RHCT_NODE_TYPE_HART_INFO) {
+			hart_info = ACPI_ADD_PTR(struct acpi_rhct_hart_info, node, size_hdr);
+			acpi_parse_hart_info_cmo_node(rhct, hart_info, cbom_size,
+						      cboz_size, cbop_size);
+		}
+	}
+}
diff --git a/drivers/clocksource/timer-riscv.c b/drivers/clocksource/timer-riscv.c
index 50198657230e..06f5bad3c3e0 100644
--- a/drivers/clocksource/timer-riscv.c
+++ b/drivers/clocksource/timer-riscv.c
@@ -22,6 +22,7 @@
 #include <linux/io-64-nonatomic-lo-hi.h>
 #include <linux/interrupt.h>
 #include <linux/of_irq.h>
+#include <linux/limits.h>
 #include <clocksource/timer-riscv.h>
 #include <asm/smp.h>
 #include <asm/hwcap.h>
@@ -31,12 +32,22 @@
 static DEFINE_STATIC_KEY_FALSE(riscv_sstc_available);
 static bool riscv_timer_cannot_wake_cpu;
 
+static void riscv_clock_event_stop(void)
+{
+	if (static_branch_likely(&riscv_sstc_available)) {
+		csr_write(CSR_STIMECMP, ULONG_MAX);
+		if (IS_ENABLED(CONFIG_32BIT))
+			csr_write(CSR_STIMECMPH, ULONG_MAX);
+	} else {
+		sbi_set_timer(U64_MAX);
+	}
+}
+
 static int riscv_clock_next_event(unsigned long delta,
 		struct clock_event_device *ce)
 {
 	u64 next_tval = get_cycles64() + delta;
 
-	csr_set(CSR_IE, IE_TIE);
 	if (static_branch_likely(&riscv_sstc_available)) {
 #if defined(CONFIG_32BIT)
 		csr_write(CSR_STIMECMP, next_tval & 0xFFFFFFFF);
@@ -94,6 +105,8 @@ static int riscv_timer_starting_cpu(unsigned int cpu)
 	ce->irq = riscv_clock_event_irq;
 	if (riscv_timer_cannot_wake_cpu)
 		ce->features |= CLOCK_EVT_FEAT_C3STOP;
+	if (static_branch_likely(&riscv_sstc_available))
+		ce->rating = 450;
 	clockevents_config_and_register(ce, riscv_timebase, 100, 0x7fffffff);
 
 	enable_percpu_irq(riscv_clock_event_irq,
@@ -119,7 +132,7 @@ static irqreturn_t riscv_timer_interrupt(int irq, void *dev_id)
 {
 	struct clock_event_device *evdev = this_cpu_ptr(&riscv_clock_event);
 
-	csr_clear(CSR_IE, IE_TIE);
+	riscv_clock_event_stop();
 	evdev->event_handler(evdev);
 
 	return IRQ_HANDLED;
diff --git a/drivers/misc/lkdtm/cfi.c b/drivers/misc/lkdtm/cfi.c
index fc28714ae3a6..6a33889d0902 100644
--- a/drivers/misc/lkdtm/cfi.c
+++ b/drivers/misc/lkdtm/cfi.c
@@ -68,12 +68,20 @@ static void lkdtm_CFI_FORWARD_PROTO(void)
 #define no_pac_addr(addr)      \
 	((__force __typeof__(addr))((uintptr_t)(addr) | PAGE_OFFSET))
 
+#ifdef CONFIG_RISCV
+/* https://github.com/riscv-non-isa/riscv-elf-psabi-doc/blob/master/riscv-cc.adoc#frame-pointer-convention */
+#define FRAME_RA_OFFSET		(-1)
+#else
+#define FRAME_RA_OFFSET		1
+#endif
+
 /* The ultimate ROP gadget. */
 static noinline __no_ret_protection
 void set_return_addr_unchecked(unsigned long *expected, unsigned long *addr)
 {
 	/* Use of volatile is to make sure final write isn't seen as a dead store. */
-	unsigned long * volatile *ret_addr = (unsigned long **)__builtin_frame_address(0) + 1;
+	unsigned long * volatile *ret_addr =
+		(unsigned long **)__builtin_frame_address(0) + FRAME_RA_OFFSET;
 
 	/* Make sure we've found the right place on the stack before writing it. */
 	if (no_pac_addr(*ret_addr) == expected)
@@ -88,7 +96,8 @@ static noinline
 void set_return_addr(unsigned long *expected, unsigned long *addr)
 {
 	/* Use of volatile is to make sure final write isn't seen as a dead store. */
-	unsigned long * volatile *ret_addr = (unsigned long **)__builtin_frame_address(0) + 1;
+	unsigned long * volatile *ret_addr =
+		(unsigned long **)__builtin_frame_address(0) + FRAME_RA_OFFSET;
 
 	/* Make sure we've found the right place on the stack before writing it. */
 	if (no_pac_addr(*ret_addr) == expected)