aboutsummaryrefslogtreecommitdiff
path: root/arch/powerpc/oprofile
diff options
context:
space:
mode:
Diffstat (limited to 'arch/powerpc/oprofile')
-rw-r--r--arch/powerpc/oprofile/Makefile2
-rw-r--r--arch/powerpc/oprofile/cell/spu_task_sync.c15
-rw-r--r--arch/powerpc/oprofile/common.c28
-rw-r--r--arch/powerpc/oprofile/op_model_7450.c1
-rw-r--r--arch/powerpc/oprofile/op_model_cell.c4
-rw-r--r--arch/powerpc/oprofile/op_model_fsl_emb.c31
-rw-r--r--arch/powerpc/oprofile/op_model_pa6t.c1
-rw-r--r--arch/powerpc/oprofile/op_model_power4.c117
-rw-r--r--arch/powerpc/oprofile/op_model_rs64.c1
9 files changed, 157 insertions, 43 deletions
diff --git a/arch/powerpc/oprofile/Makefile b/arch/powerpc/oprofile/Makefile
index 73456c4cec2..751ec7bd501 100644
--- a/arch/powerpc/oprofile/Makefile
+++ b/arch/powerpc/oprofile/Makefile
@@ -1,6 +1,6 @@
subdir-ccflags-$(CONFIG_PPC_WERROR) := -Werror
-ccflags-$(CONFIG_PPC64) := -mno-minimal-toc
+ccflags-$(CONFIG_PPC64) := $(NO_MINIMAL_TOC)
obj-$(CONFIG_OPROFILE) += oprofile.o
diff --git a/arch/powerpc/oprofile/cell/spu_task_sync.c b/arch/powerpc/oprofile/cell/spu_task_sync.c
index 642fca137cc..28f1af2db1f 100644
--- a/arch/powerpc/oprofile/cell/spu_task_sync.c
+++ b/arch/powerpc/oprofile/cell/spu_task_sync.c
@@ -304,7 +304,7 @@ static inline unsigned long fast_get_dcookie(struct path *path)
return cookie;
}
-/* Look up the dcookie for the task's first VM_EXECUTABLE mapping,
+/* Look up the dcookie for the task's mm->exe_file,
* which corresponds loosely to "application name". Also, determine
* the offset for the SPU ELF object. If computed offset is
* non-zero, it implies an embedded SPU object; otherwise, it's a
@@ -321,7 +321,6 @@ get_exec_dcookie_and_offset(struct spu *spu, unsigned int *offsetp,
{
unsigned long app_cookie = 0;
unsigned int my_offset = 0;
- struct file *app = NULL;
struct vm_area_struct *vma;
struct mm_struct *mm = spu->mm;
@@ -330,16 +329,10 @@ get_exec_dcookie_and_offset(struct spu *spu, unsigned int *offsetp,
down_read(&mm->mmap_sem);
- for (vma = mm->mmap; vma; vma = vma->vm_next) {
- if (!vma->vm_file)
- continue;
- if (!(vma->vm_flags & VM_EXECUTABLE))
- continue;
- app_cookie = fast_get_dcookie(&vma->vm_file->f_path);
+ if (mm->exe_file) {
+ app_cookie = fast_get_dcookie(&mm->exe_file->f_path);
pr_debug("got dcookie for %s\n",
- vma->vm_file->f_dentry->d_name.name);
- app = vma->vm_file;
- break;
+ mm->exe_file->f_dentry->d_name.name);
}
for (vma = mm->mmap; vma; vma = vma->vm_next) {
diff --git a/arch/powerpc/oprofile/common.c b/arch/powerpc/oprofile/common.c
index 4f51025f5b0..c77348c5d46 100644
--- a/arch/powerpc/oprofile/common.c
+++ b/arch/powerpc/oprofile/common.c
@@ -119,7 +119,7 @@ static void op_powerpc_stop(void)
model->global_stop();
}
-static int op_powerpc_create_files(struct super_block *sb, struct dentry *root)
+static int op_powerpc_create_files(struct dentry *root)
{
int i;
@@ -128,9 +128,9 @@ static int op_powerpc_create_files(struct super_block *sb, struct dentry *root)
* There is one mmcr0, mmcr1 and mmcra for setting the events for
* all of the counters.
*/
- oprofilefs_create_ulong(sb, root, "mmcr0", &sys.mmcr0);
- oprofilefs_create_ulong(sb, root, "mmcr1", &sys.mmcr1);
- oprofilefs_create_ulong(sb, root, "mmcra", &sys.mmcra);
+ oprofilefs_create_ulong(root, "mmcr0", &sys.mmcr0);
+ oprofilefs_create_ulong(root, "mmcr1", &sys.mmcr1);
+ oprofilefs_create_ulong(root, "mmcra", &sys.mmcra);
#ifdef CONFIG_OPROFILE_CELL
/* create a file the user tool can check to see what level of profiling
* support exits with this kernel. Initialize bit mask to indicate
@@ -142,7 +142,7 @@ static int op_powerpc_create_files(struct super_block *sb, struct dentry *root)
* If the file does not exist, then the kernel only supports SPU
* cycle profiling, PPU event and cycle profiling.
*/
- oprofilefs_create_ulong(sb, root, "cell_support", &sys.cell_support);
+ oprofilefs_create_ulong(root, "cell_support", &sys.cell_support);
sys.cell_support = 0x1; /* Note, the user OProfile tool must check
* that this bit is set before attempting to
* user SPU event profiling. Older kernels
@@ -160,11 +160,11 @@ static int op_powerpc_create_files(struct super_block *sb, struct dentry *root)
char buf[4];
snprintf(buf, sizeof buf, "%d", i);
- dir = oprofilefs_mkdir(sb, root, buf);
+ dir = oprofilefs_mkdir(root, buf);
- oprofilefs_create_ulong(sb, dir, "enabled", &ctr[i].enabled);
- oprofilefs_create_ulong(sb, dir, "event", &ctr[i].event);
- oprofilefs_create_ulong(sb, dir, "count", &ctr[i].count);
+ oprofilefs_create_ulong(dir, "enabled", &ctr[i].enabled);
+ oprofilefs_create_ulong(dir, "event", &ctr[i].event);
+ oprofilefs_create_ulong(dir, "count", &ctr[i].count);
/*
* Classic PowerPC doesn't support per-counter
@@ -173,14 +173,14 @@ static int op_powerpc_create_files(struct super_block *sb, struct dentry *root)
* Book-E style performance monitors, we do
* support them.
*/
- oprofilefs_create_ulong(sb, dir, "kernel", &ctr[i].kernel);
- oprofilefs_create_ulong(sb, dir, "user", &ctr[i].user);
+ oprofilefs_create_ulong(dir, "kernel", &ctr[i].kernel);
+ oprofilefs_create_ulong(dir, "user", &ctr[i].user);
- oprofilefs_create_ulong(sb, dir, "unit_mask", &ctr[i].unit_mask);
+ oprofilefs_create_ulong(dir, "unit_mask", &ctr[i].unit_mask);
}
- oprofilefs_create_ulong(sb, root, "enable_kernel", &sys.enable_kernel);
- oprofilefs_create_ulong(sb, root, "enable_user", &sys.enable_user);
+ oprofilefs_create_ulong(root, "enable_kernel", &sys.enable_kernel);
+ oprofilefs_create_ulong(root, "enable_user", &sys.enable_user);
/* Default to tracing both kernel and user */
sys.enable_kernel = 1;
diff --git a/arch/powerpc/oprofile/op_model_7450.c b/arch/powerpc/oprofile/op_model_7450.c
index ff617246d12..d29b6e4e5e7 100644
--- a/arch/powerpc/oprofile/op_model_7450.c
+++ b/arch/powerpc/oprofile/op_model_7450.c
@@ -16,7 +16,6 @@
*/
#include <linux/oprofile.h>
-#include <linux/init.h>
#include <linux/smp.h>
#include <asm/ptrace.h>
#include <asm/processor.h>
diff --git a/arch/powerpc/oprofile/op_model_cell.c b/arch/powerpc/oprofile/op_model_cell.c
index b9589c19ccd..863d89386f6 100644
--- a/arch/powerpc/oprofile/op_model_cell.c
+++ b/arch/powerpc/oprofile/op_model_cell.c
@@ -16,7 +16,6 @@
#include <linux/cpufreq.h>
#include <linux/delay.h>
-#include <linux/init.h>
#include <linux/jiffies.h>
#include <linux/kthread.h>
#include <linux/oprofile.h>
@@ -1122,8 +1121,7 @@ oprof_cpufreq_notify(struct notifier_block *nb, unsigned long val, void *data)
int ret = 0;
struct cpufreq_freqs *frq = data;
if ((val == CPUFREQ_PRECHANGE && frq->old < frq->new) ||
- (val == CPUFREQ_POSTCHANGE && frq->old > frq->new) ||
- (val == CPUFREQ_RESUMECHANGE || val == CPUFREQ_SUSPENDCHANGE))
+ (val == CPUFREQ_POSTCHANGE && frq->old > frq->new))
set_spu_profiling_frequency(frq->new, spu_cycle_reset);
return ret;
}
diff --git a/arch/powerpc/oprofile/op_model_fsl_emb.c b/arch/powerpc/oprofile/op_model_fsl_emb.c
index ccc1daa33ae..14cf86fddda 100644
--- a/arch/powerpc/oprofile/op_model_fsl_emb.c
+++ b/arch/powerpc/oprofile/op_model_fsl_emb.c
@@ -14,7 +14,6 @@
*/
#include <linux/oprofile.h>
-#include <linux/init.h>
#include <linux/smp.h>
#include <asm/ptrace.h>
#include <asm/processor.h>
@@ -46,6 +45,12 @@ static inline u32 get_pmlca(int ctr)
case 3:
pmlca = mfpmr(PMRN_PMLCA3);
break;
+ case 4:
+ pmlca = mfpmr(PMRN_PMLCA4);
+ break;
+ case 5:
+ pmlca = mfpmr(PMRN_PMLCA5);
+ break;
default:
panic("Bad ctr number\n");
}
@@ -68,6 +73,12 @@ static inline void set_pmlca(int ctr, u32 pmlca)
case 3:
mtpmr(PMRN_PMLCA3, pmlca);
break;
+ case 4:
+ mtpmr(PMRN_PMLCA4, pmlca);
+ break;
+ case 5:
+ mtpmr(PMRN_PMLCA5, pmlca);
+ break;
default:
panic("Bad ctr number\n");
}
@@ -84,6 +95,10 @@ static inline unsigned int ctr_read(unsigned int i)
return mfpmr(PMRN_PMC2);
case 3:
return mfpmr(PMRN_PMC3);
+ case 4:
+ return mfpmr(PMRN_PMC4);
+ case 5:
+ return mfpmr(PMRN_PMC5);
default:
return 0;
}
@@ -104,6 +119,12 @@ static inline void ctr_write(unsigned int i, unsigned int val)
case 3:
mtpmr(PMRN_PMC3, val);
break;
+ case 4:
+ mtpmr(PMRN_PMC4, val);
+ break;
+ case 5:
+ mtpmr(PMRN_PMC5, val);
+ break;
default:
break;
}
@@ -133,6 +154,14 @@ static void init_pmc_stop(int ctr)
mtpmr(PMRN_PMLCA3, pmlca);
mtpmr(PMRN_PMLCB3, pmlcb);
break;
+ case 4:
+ mtpmr(PMRN_PMLCA4, pmlca);
+ mtpmr(PMRN_PMLCB4, pmlcb);
+ break;
+ case 5:
+ mtpmr(PMRN_PMLCA5, pmlca);
+ mtpmr(PMRN_PMLCB5, pmlcb);
+ break;
default:
panic("Bad ctr number!\n");
}
diff --git a/arch/powerpc/oprofile/op_model_pa6t.c b/arch/powerpc/oprofile/op_model_pa6t.c
index 42f778dff91..a114a7c22d4 100644
--- a/arch/powerpc/oprofile/op_model_pa6t.c
+++ b/arch/powerpc/oprofile/op_model_pa6t.c
@@ -22,7 +22,6 @@
*/
#include <linux/oprofile.h>
-#include <linux/init.h>
#include <linux/smp.h>
#include <linux/percpu.h>
#include <asm/processor.h>
diff --git a/arch/powerpc/oprofile/op_model_power4.c b/arch/powerpc/oprofile/op_model_power4.c
index 95ae77dec3f..962fe7b3e3f 100644
--- a/arch/powerpc/oprofile/op_model_power4.c
+++ b/arch/powerpc/oprofile/op_model_power4.c
@@ -10,7 +10,6 @@
*/
#include <linux/oprofile.h>
-#include <linux/init.h>
#include <linux/smp.h>
#include <asm/firmware.h>
#include <asm/ptrace.h>
@@ -21,6 +20,13 @@
#include <asm/reg.h>
#define dbg(args...)
+#define OPROFILE_PM_PMCSEL_MSK 0xffULL
+#define OPROFILE_PM_UNIT_SHIFT 60
+#define OPROFILE_PM_UNIT_MSK 0xfULL
+#define OPROFILE_MAX_PMC_NUM 3
+#define OPROFILE_PMSEL_FIELD_WIDTH 8
+#define OPROFILE_UNIT_FIELD_WIDTH 4
+#define MMCRA_SIAR_VALID_MASK 0x10000000ULL
static unsigned long reset_value[OP_MAX_COUNTER];
@@ -31,6 +37,61 @@ static int use_slot_nums;
static u32 mmcr0_val;
static u64 mmcr1_val;
static u64 mmcra_val;
+static u32 cntr_marked_events;
+
+static int power7_marked_instr_event(u64 mmcr1)
+{
+ u64 psel, unit;
+ int pmc, cntr_marked_events = 0;
+
+ /* Given the MMCR1 value, look at the field for each counter to
+ * determine if it is a marked event. Code based on the function
+ * power7_marked_instr_event() in file arch/powerpc/perf/power7-pmu.c.
+ */
+ for (pmc = 0; pmc < 4; pmc++) {
+ psel = mmcr1 & (OPROFILE_PM_PMCSEL_MSK
+ << (OPROFILE_MAX_PMC_NUM - pmc)
+ * OPROFILE_PMSEL_FIELD_WIDTH);
+ psel = (psel >> ((OPROFILE_MAX_PMC_NUM - pmc)
+ * OPROFILE_PMSEL_FIELD_WIDTH)) & ~1ULL;
+ unit = mmcr1 & (OPROFILE_PM_UNIT_MSK
+ << (OPROFILE_PM_UNIT_SHIFT
+ - (pmc * OPROFILE_PMSEL_FIELD_WIDTH )));
+ unit = unit >> (OPROFILE_PM_UNIT_SHIFT
+ - (pmc * OPROFILE_PMSEL_FIELD_WIDTH));
+
+ switch (psel >> 4) {
+ case 2:
+ cntr_marked_events |= (pmc == 1 || pmc == 3) << pmc;
+ break;
+ case 3:
+ if (psel == 0x3c) {
+ cntr_marked_events |= (pmc == 0) << pmc;
+ break;
+ }
+
+ if (psel == 0x3e) {
+ cntr_marked_events |= (pmc != 1) << pmc;
+ break;
+ }
+
+ cntr_marked_events |= 1 << pmc;
+ break;
+ case 4:
+ case 5:
+ cntr_marked_events |= (unit == 0xd) << pmc;
+ break;
+ case 6:
+ if (psel == 0x64)
+ cntr_marked_events |= (pmc >= 2) << pmc;
+ break;
+ case 8:
+ cntr_marked_events |= (unit == 0xd) << pmc;
+ break;
+ }
+ }
+ return cntr_marked_events;
+}
static int power4_reg_setup(struct op_counter_config *ctr,
struct op_system_config *sys,
@@ -47,6 +108,23 @@ static int power4_reg_setup(struct op_counter_config *ctr,
mmcr1_val = sys->mmcr1;
mmcra_val = sys->mmcra;
+ /* Power 7+ and newer architectures:
+ * Determine which counter events in the group (the group of events is
+ * specified by the bit settings in the MMCR1 register) are marked
+ * events for use in the interrupt handler. Do the calculation once
+ * before OProfile starts. Information is used in the interrupt
+ * handler. Starting with Power 7+ we only record the sample for
+ * marked events if the SIAR valid bit is set. For non marked events
+ * the sample is always recorded.
+ */
+ if (pvr_version_is(PVR_POWER7p))
+ cntr_marked_events = power7_marked_instr_event(mmcr1_val);
+ else
+ cntr_marked_events = 0; /* For older processors, set the bit map
+ * to zero so the sample will always be
+ * be recorded.
+ */
+
for (i = 0; i < cur_cpu_spec->num_pmcs; ++i)
reset_value[i] = 0x80000000UL - ctr[i].count;
@@ -61,10 +139,10 @@ static int power4_reg_setup(struct op_counter_config *ctr,
else
mmcr0_val |= MMCR0_PROBLEM_DISABLE;
- if (__is_processor(PV_POWER4) || __is_processor(PV_POWER4p) ||
- __is_processor(PV_970) || __is_processor(PV_970FX) ||
- __is_processor(PV_970MP) || __is_processor(PV_970GX) ||
- __is_processor(PV_POWER5) || __is_processor(PV_POWER5p))
+ if (pvr_version_is(PVR_POWER4) || pvr_version_is(PVR_POWER4p) ||
+ pvr_version_is(PVR_970) || pvr_version_is(PVR_970FX) ||
+ pvr_version_is(PVR_970MP) || pvr_version_is(PVR_970GX) ||
+ pvr_version_is(PVR_POWER5) || pvr_version_is(PVR_POWER5p))
use_slot_nums = 1;
return 0;
@@ -84,9 +162,9 @@ extern void ppc_enable_pmcs(void);
*/
static inline int mmcra_must_set_sample(void)
{
- if (__is_processor(PV_POWER4) || __is_processor(PV_POWER4p) ||
- __is_processor(PV_970) || __is_processor(PV_970FX) ||
- __is_processor(PV_970MP) || __is_processor(PV_970GX))
+ if (pvr_version_is(PVR_POWER4) || pvr_version_is(PVR_POWER4p) ||
+ pvr_version_is(PVR_970) || pvr_version_is(PVR_970FX) ||
+ pvr_version_is(PVR_970MP) || pvr_version_is(PVR_970GX))
return 1;
return 0;
@@ -276,7 +354,7 @@ static bool pmc_overflow(unsigned long val)
* PMCs because a user might set a period of less than 256 and we
* don't want to mistakenly reset them.
*/
- if (__is_processor(PV_POWER7) && ((0x80000000 - val) <= 256))
+ if (pvr_version_is(PVR_POWER7) && ((0x80000000 - val) <= 256))
return true;
return false;
@@ -291,6 +369,7 @@ static void power4_handle_interrupt(struct pt_regs *regs,
int i;
unsigned int mmcr0;
unsigned long mmcra;
+ bool siar_valid = false;
mmcra = mfspr(SPRN_MMCRA);
@@ -300,11 +379,29 @@ static void power4_handle_interrupt(struct pt_regs *regs,
/* set the PMM bit (see comment below) */
mtmsrd(mfmsr() | MSR_PMM);
+ /* Check that the SIAR valid bit in MMCRA is set to 1. */
+ if ((mmcra & MMCRA_SIAR_VALID_MASK) == MMCRA_SIAR_VALID_MASK)
+ siar_valid = true;
+
for (i = 0; i < cur_cpu_spec->num_pmcs; ++i) {
val = classic_ctr_read(i);
if (pmc_overflow(val)) {
if (oprofile_running && ctr[i].enabled) {
- oprofile_add_ext_sample(pc, regs, i, is_kernel);
+ /* Power 7+ and newer architectures:
+ * If the event is a marked event, then only
+ * save the sample if the SIAR valid bit is
+ * set. If the event is not marked, then
+ * always save the sample.
+ * Note, the Sample enable bit in the MMCRA
+ * register must be set to 1 if the group
+ * contains a marked event.
+ */
+ if ((siar_valid &&
+ (cntr_marked_events & (1 << i)))
+ || !(cntr_marked_events & (1 << i)))
+ oprofile_add_ext_sample(pc, regs, i,
+ is_kernel);
+
classic_ctr_write(i, reset_value[i]);
} else {
classic_ctr_write(i, 0);
diff --git a/arch/powerpc/oprofile/op_model_rs64.c b/arch/powerpc/oprofile/op_model_rs64.c
index 9b801b8c8c5..7e5b8ed3a1b 100644
--- a/arch/powerpc/oprofile/op_model_rs64.c
+++ b/arch/powerpc/oprofile/op_model_rs64.c
@@ -8,7 +8,6 @@
*/
#include <linux/oprofile.h>
-#include <linux/init.h>
#include <linux/smp.h>
#include <asm/ptrace.h>
#include <asm/processor.h>
generated by cgit v1.2.3-70-g09d2 (git 2.46.0) at 2025-12-15 19:41:42 +0000