Skip to content

Commit

Permalink
Merge branch 'perf-urgent-for-linus' of git://git.kernel.org/pub/scm/…
Browse files Browse the repository at this point in the history
…linux/kernel/git/tip/tip

Pull perf updates from Ingo Molnar:
 "This tree includes some late late perf items that missed the first
  round:

  tools:

   - Bash auto completion improvements, now we can auto complete the
     tools long options, tracepoint event names, etc, from Namhyung Kim.

   - Look up thread using tid instead of pid in 'perf sched'.

   - Move global variables into a perf_kvm struct, from David Ahern.

   - Hists refactorings, preparatory for improved 'diff' command, from
     Jiri Olsa.

   - Hists refactorings, preparatory for event group viewieng work, from
     Namhyung Kim.

   - Remove double negation on optional feature macro definitions, from
     Namhyung Kim.

   - Remove several cases of needless global variables, on most
     builtins.

   - misc fixes

  kernel:

   - sysfs support for IBS on AMD CPUs, from Robert Richter.

   - Support for an upcoming Intel CPU, the Xeon-Phi / Knights Corner
     HPC blade PMU, from Vince Weaver.

   - misc fixes"

* 'perf-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip: (46 commits)
  perf: Fix perf_cgroup_switch for sw-events
  perf: Clarify perf_cpu_context::active_pmu usage by renaming it to ::unique_pmu
  perf/AMD/IBS: Add sysfs support
  perf hists: Add more helpers for hist entry stat
  perf hists: Move he->stat.nr_events initialization to a template
  perf hists: Introduce struct he_stat
  perf diff: Removing the total_period argument from output code
  perf tool: Add hpp interface to enable/disable hpp column
  perf tools: Removing hists pair argument from output path
  perf hists: Separate overhead and baseline columns
  perf diff: Refactor diff displacement possition info
  perf hists: Add struct hists pointer to struct hist_entry
  perf tools: Complete tracepoint event names
  perf/x86: Add support for Intel Xeon-Phi Knights Corner PMU
  perf evlist: Remove some unused methods
  perf evlist: Introduce add_newtp method
  perf kvm: Move global variables into a perf_kvm struct
  perf tools: Convert to BACKTRACE_SUPPORT
  perf tools: Long option completion support for each subcommands
  perf tools: Complete long option names of perf command
  ...
  • Loading branch information
torvalds committed Oct 13, 2012
2 parents 871a059 + 95cf59e commit ade0899
Show file tree
Hide file tree
Showing 55 changed files with 1,517 additions and 1,198 deletions.
5 changes: 5 additions & 0 deletions arch/x86/include/asm/msr-index.h
Original file line number Diff line number Diff line change
Expand Up @@ -121,6 +121,11 @@
#define MSR_P6_EVNTSEL0 0x00000186
#define MSR_P6_EVNTSEL1 0x00000187

#define MSR_KNC_PERFCTR0 0x00000020
#define MSR_KNC_PERFCTR1 0x00000021
#define MSR_KNC_EVNTSEL0 0x00000028
#define MSR_KNC_EVNTSEL1 0x00000029

/* AMD64 MSRs. Not complete. See the architecture manual for a more
complete list. */

Expand Down
2 changes: 1 addition & 1 deletion arch/x86/kernel/cpu/Makefile
Original file line number Diff line number Diff line change
Expand Up @@ -32,7 +32,7 @@ obj-$(CONFIG_PERF_EVENTS) += perf_event.o

ifdef CONFIG_PERF_EVENTS
obj-$(CONFIG_CPU_SUP_AMD) += perf_event_amd.o
obj-$(CONFIG_CPU_SUP_INTEL) += perf_event_p6.o perf_event_p4.o
obj-$(CONFIG_CPU_SUP_INTEL) += perf_event_p6.o perf_event_knc.o perf_event_p4.o
obj-$(CONFIG_CPU_SUP_INTEL) += perf_event_intel_lbr.o perf_event_intel_ds.o perf_event_intel.o
obj-$(CONFIG_CPU_SUP_INTEL) += perf_event_intel_uncore.o
endif
Expand Down
2 changes: 2 additions & 0 deletions arch/x86/kernel/cpu/perf_event.h
Original file line number Diff line number Diff line change
Expand Up @@ -626,6 +626,8 @@ int p4_pmu_init(void);

int p6_pmu_init(void);

int knc_pmu_init(void);

#else /* CONFIG_CPU_SUP_INTEL */

static inline void reserve_ds_buffers(void)
Expand Down
61 changes: 49 additions & 12 deletions arch/x86/kernel/cpu/perf_event_amd_ibs.c
Original file line number Diff line number Diff line change
Expand Up @@ -41,17 +41,22 @@ struct cpu_perf_ibs {
};

struct perf_ibs {
struct pmu pmu;
unsigned int msr;
u64 config_mask;
u64 cnt_mask;
u64 enable_mask;
u64 valid_mask;
u64 max_period;
unsigned long offset_mask[1];
int offset_max;
struct cpu_perf_ibs __percpu *pcpu;
u64 (*get_count)(u64 config);
struct pmu pmu;
unsigned int msr;
u64 config_mask;
u64 cnt_mask;
u64 enable_mask;
u64 valid_mask;
u64 max_period;
unsigned long offset_mask[1];
int offset_max;
struct cpu_perf_ibs __percpu *pcpu;

struct attribute **format_attrs;
struct attribute_group format_group;
const struct attribute_group *attr_groups[2];

u64 (*get_count)(u64 config);
};

struct perf_ibs_data {
Expand Down Expand Up @@ -446,6 +451,19 @@ static void perf_ibs_del(struct perf_event *event, int flags)

static void perf_ibs_read(struct perf_event *event) { }

PMU_FORMAT_ATTR(rand_en, "config:57");
PMU_FORMAT_ATTR(cnt_ctl, "config:19");

static struct attribute *ibs_fetch_format_attrs[] = {
&format_attr_rand_en.attr,
NULL,
};

static struct attribute *ibs_op_format_attrs[] = {
NULL, /* &format_attr_cnt_ctl.attr if IBS_CAPS_OPCNT */
NULL,
};

static struct perf_ibs perf_ibs_fetch = {
.pmu = {
.task_ctx_nr = perf_invalid_context,
Expand All @@ -465,6 +483,7 @@ static struct perf_ibs perf_ibs_fetch = {
.max_period = IBS_FETCH_MAX_CNT << 4,
.offset_mask = { MSR_AMD64_IBSFETCH_REG_MASK },
.offset_max = MSR_AMD64_IBSFETCH_REG_COUNT,
.format_attrs = ibs_fetch_format_attrs,

.get_count = get_ibs_fetch_count,
};
Expand All @@ -488,6 +507,7 @@ static struct perf_ibs perf_ibs_op = {
.max_period = IBS_OP_MAX_CNT << 4,
.offset_mask = { MSR_AMD64_IBSOP_REG_MASK },
.offset_max = MSR_AMD64_IBSOP_REG_COUNT,
.format_attrs = ibs_op_format_attrs,

.get_count = get_ibs_op_count,
};
Expand Down Expand Up @@ -597,6 +617,17 @@ static __init int perf_ibs_pmu_init(struct perf_ibs *perf_ibs, char *name)

perf_ibs->pcpu = pcpu;

/* register attributes */
if (perf_ibs->format_attrs[0]) {
memset(&perf_ibs->format_group, 0, sizeof(perf_ibs->format_group));
perf_ibs->format_group.name = "format";
perf_ibs->format_group.attrs = perf_ibs->format_attrs;

memset(&perf_ibs->attr_groups, 0, sizeof(perf_ibs->attr_groups));
perf_ibs->attr_groups[0] = &perf_ibs->format_group;
perf_ibs->pmu.attr_groups = perf_ibs->attr_groups;
}

ret = perf_pmu_register(&perf_ibs->pmu, name, -1);
if (ret) {
perf_ibs->pcpu = NULL;
Expand All @@ -608,13 +639,19 @@ static __init int perf_ibs_pmu_init(struct perf_ibs *perf_ibs, char *name)

static __init int perf_event_ibs_init(void)
{
struct attribute **attr = ibs_op_format_attrs;

if (!ibs_caps)
return -ENODEV; /* ibs not supported by the cpu */

perf_ibs_pmu_init(&perf_ibs_fetch, "ibs_fetch");
if (ibs_caps & IBS_CAPS_OPCNT)

if (ibs_caps & IBS_CAPS_OPCNT) {
perf_ibs_op.config_mask |= IBS_OP_CNT_CTL;
*attr++ = &format_attr_cnt_ctl.attr;
}
perf_ibs_pmu_init(&perf_ibs_op, "ibs_op");

register_nmi_handler(NMI_LOCAL, perf_ibs_nmi_handler, 0, "perf_ibs");
printk(KERN_INFO "perf: AMD IBS detected (0x%08x)\n", ibs_caps);

Expand Down
2 changes: 2 additions & 0 deletions arch/x86/kernel/cpu/perf_event_intel.c
Original file line number Diff line number Diff line change
Expand Up @@ -1906,6 +1906,8 @@ __init int intel_pmu_init(void)
switch (boot_cpu_data.x86) {
case 0x6:
return p6_pmu_init();
case 0xb:
return knc_pmu_init();
case 0xf:
return p4_pmu_init();
}
Expand Down
248 changes: 248 additions & 0 deletions arch/x86/kernel/cpu/perf_event_knc.c
Original file line number Diff line number Diff line change
@@ -0,0 +1,248 @@
/* Driver for Intel Xeon Phi "Knights Corner" PMU */

#include <linux/perf_event.h>
#include <linux/types.h>

#include "perf_event.h"

static const u64 knc_perfmon_event_map[] =
{
[PERF_COUNT_HW_CPU_CYCLES] = 0x002a,
[PERF_COUNT_HW_INSTRUCTIONS] = 0x0016,
[PERF_COUNT_HW_CACHE_REFERENCES] = 0x0028,
[PERF_COUNT_HW_CACHE_MISSES] = 0x0029,
[PERF_COUNT_HW_BRANCH_INSTRUCTIONS] = 0x0012,
[PERF_COUNT_HW_BRANCH_MISSES] = 0x002b,
};

static __initconst u64 knc_hw_cache_event_ids
[PERF_COUNT_HW_CACHE_MAX]
[PERF_COUNT_HW_CACHE_OP_MAX]
[PERF_COUNT_HW_CACHE_RESULT_MAX] =
{
[ C(L1D) ] = {
[ C(OP_READ) ] = {
/* On Xeon Phi event "0" is a valid DATA_READ */
/* (L1 Data Cache Reads) Instruction. */
/* We code this as ARCH_PERFMON_EVENTSEL_INT as this */
/* bit will always be set in x86_pmu_hw_config(). */
[ C(RESULT_ACCESS) ] = ARCH_PERFMON_EVENTSEL_INT,
/* DATA_READ */
[ C(RESULT_MISS) ] = 0x0003, /* DATA_READ_MISS */
},
[ C(OP_WRITE) ] = {
[ C(RESULT_ACCESS) ] = 0x0001, /* DATA_WRITE */
[ C(RESULT_MISS) ] = 0x0004, /* DATA_WRITE_MISS */
},
[ C(OP_PREFETCH) ] = {
[ C(RESULT_ACCESS) ] = 0x0011, /* L1_DATA_PF1 */
[ C(RESULT_MISS) ] = 0x001c, /* L1_DATA_PF1_MISS */
},
},
[ C(L1I ) ] = {
[ C(OP_READ) ] = {
[ C(RESULT_ACCESS) ] = 0x000c, /* CODE_READ */
[ C(RESULT_MISS) ] = 0x000e, /* CODE_CACHE_MISS */
},
[ C(OP_WRITE) ] = {
[ C(RESULT_ACCESS) ] = -1,
[ C(RESULT_MISS) ] = -1,
},
[ C(OP_PREFETCH) ] = {
[ C(RESULT_ACCESS) ] = 0x0,
[ C(RESULT_MISS) ] = 0x0,
},
},
[ C(LL ) ] = {
[ C(OP_READ) ] = {
[ C(RESULT_ACCESS) ] = 0,
[ C(RESULT_MISS) ] = 0x10cb, /* L2_READ_MISS */
},
[ C(OP_WRITE) ] = {
[ C(RESULT_ACCESS) ] = 0x10cc, /* L2_WRITE_HIT */
[ C(RESULT_MISS) ] = 0,
},
[ C(OP_PREFETCH) ] = {
[ C(RESULT_ACCESS) ] = 0x10fc, /* L2_DATA_PF2 */
[ C(RESULT_MISS) ] = 0x10fe, /* L2_DATA_PF2_MISS */
},
},
[ C(DTLB) ] = {
[ C(OP_READ) ] = {
[ C(RESULT_ACCESS) ] = ARCH_PERFMON_EVENTSEL_INT,
/* DATA_READ */
/* see note on L1 OP_READ */
[ C(RESULT_MISS) ] = 0x0002, /* DATA_PAGE_WALK */
},
[ C(OP_WRITE) ] = {
[ C(RESULT_ACCESS) ] = 0x0001, /* DATA_WRITE */
[ C(RESULT_MISS) ] = 0x0002, /* DATA_PAGE_WALK */
},
[ C(OP_PREFETCH) ] = {
[ C(RESULT_ACCESS) ] = 0x0,
[ C(RESULT_MISS) ] = 0x0,
},
},
[ C(ITLB) ] = {
[ C(OP_READ) ] = {
[ C(RESULT_ACCESS) ] = 0x000c, /* CODE_READ */
[ C(RESULT_MISS) ] = 0x000d, /* CODE_PAGE_WALK */
},
[ C(OP_WRITE) ] = {
[ C(RESULT_ACCESS) ] = -1,
[ C(RESULT_MISS) ] = -1,
},
[ C(OP_PREFETCH) ] = {
[ C(RESULT_ACCESS) ] = -1,
[ C(RESULT_MISS) ] = -1,
},
},
[ C(BPU ) ] = {
[ C(OP_READ) ] = {
[ C(RESULT_ACCESS) ] = 0x0012, /* BRANCHES */
[ C(RESULT_MISS) ] = 0x002b, /* BRANCHES_MISPREDICTED */
},
[ C(OP_WRITE) ] = {
[ C(RESULT_ACCESS) ] = -1,
[ C(RESULT_MISS) ] = -1,
},
[ C(OP_PREFETCH) ] = {
[ C(RESULT_ACCESS) ] = -1,
[ C(RESULT_MISS) ] = -1,
},
},
};


static u64 knc_pmu_event_map(int hw_event)
{
return knc_perfmon_event_map[hw_event];
}

static struct event_constraint knc_event_constraints[] =
{
INTEL_EVENT_CONSTRAINT(0xc3, 0x1), /* HWP_L2HIT */
INTEL_EVENT_CONSTRAINT(0xc4, 0x1), /* HWP_L2MISS */
INTEL_EVENT_CONSTRAINT(0xc8, 0x1), /* L2_READ_HIT_E */
INTEL_EVENT_CONSTRAINT(0xc9, 0x1), /* L2_READ_HIT_M */
INTEL_EVENT_CONSTRAINT(0xca, 0x1), /* L2_READ_HIT_S */
INTEL_EVENT_CONSTRAINT(0xcb, 0x1), /* L2_READ_MISS */
INTEL_EVENT_CONSTRAINT(0xcc, 0x1), /* L2_WRITE_HIT */
INTEL_EVENT_CONSTRAINT(0xce, 0x1), /* L2_STRONGLY_ORDERED_STREAMING_VSTORES_MISS */
INTEL_EVENT_CONSTRAINT(0xcf, 0x1), /* L2_WEAKLY_ORDERED_STREAMING_VSTORE_MISS */
INTEL_EVENT_CONSTRAINT(0xd7, 0x1), /* L2_VICTIM_REQ_WITH_DATA */
INTEL_EVENT_CONSTRAINT(0xe3, 0x1), /* SNP_HITM_BUNIT */
INTEL_EVENT_CONSTRAINT(0xe6, 0x1), /* SNP_HIT_L2 */
INTEL_EVENT_CONSTRAINT(0xe7, 0x1), /* SNP_HITM_L2 */
INTEL_EVENT_CONSTRAINT(0xf1, 0x1), /* L2_DATA_READ_MISS_CACHE_FILL */
INTEL_EVENT_CONSTRAINT(0xf2, 0x1), /* L2_DATA_WRITE_MISS_CACHE_FILL */
INTEL_EVENT_CONSTRAINT(0xf6, 0x1), /* L2_DATA_READ_MISS_MEM_FILL */
INTEL_EVENT_CONSTRAINT(0xf7, 0x1), /* L2_DATA_WRITE_MISS_MEM_FILL */
INTEL_EVENT_CONSTRAINT(0xfc, 0x1), /* L2_DATA_PF2 */
INTEL_EVENT_CONSTRAINT(0xfd, 0x1), /* L2_DATA_PF2_DROP */
INTEL_EVENT_CONSTRAINT(0xfe, 0x1), /* L2_DATA_PF2_MISS */
INTEL_EVENT_CONSTRAINT(0xff, 0x1), /* L2_DATA_HIT_INFLIGHT_PF2 */
EVENT_CONSTRAINT_END
};

#define MSR_KNC_IA32_PERF_GLOBAL_STATUS 0x0000002d
#define MSR_KNC_IA32_PERF_GLOBAL_OVF_CONTROL 0x0000002e
#define MSR_KNC_IA32_PERF_GLOBAL_CTRL 0x0000002f

#define KNC_ENABLE_COUNTER0 0x00000001
#define KNC_ENABLE_COUNTER1 0x00000002

static void knc_pmu_disable_all(void)
{
u64 val;

rdmsrl(MSR_KNC_IA32_PERF_GLOBAL_CTRL, val);
val &= ~(KNC_ENABLE_COUNTER0|KNC_ENABLE_COUNTER1);
wrmsrl(MSR_KNC_IA32_PERF_GLOBAL_CTRL, val);
}

static void knc_pmu_enable_all(int added)
{
u64 val;

rdmsrl(MSR_KNC_IA32_PERF_GLOBAL_CTRL, val);
val |= (KNC_ENABLE_COUNTER0|KNC_ENABLE_COUNTER1);
wrmsrl(MSR_KNC_IA32_PERF_GLOBAL_CTRL, val);
}

static inline void
knc_pmu_disable_event(struct perf_event *event)
{
struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events);
struct hw_perf_event *hwc = &event->hw;
u64 val;

val = hwc->config;
if (cpuc->enabled)
val &= ~ARCH_PERFMON_EVENTSEL_ENABLE;

(void)wrmsrl_safe(hwc->config_base + hwc->idx, val);
}

static void knc_pmu_enable_event(struct perf_event *event)
{
struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events);
struct hw_perf_event *hwc = &event->hw;
u64 val;

val = hwc->config;
if (cpuc->enabled)
val |= ARCH_PERFMON_EVENTSEL_ENABLE;

(void)wrmsrl_safe(hwc->config_base + hwc->idx, val);
}

PMU_FORMAT_ATTR(event, "config:0-7" );
PMU_FORMAT_ATTR(umask, "config:8-15" );
PMU_FORMAT_ATTR(edge, "config:18" );
PMU_FORMAT_ATTR(inv, "config:23" );
PMU_FORMAT_ATTR(cmask, "config:24-31" );

static struct attribute *intel_knc_formats_attr[] = {
&format_attr_event.attr,
&format_attr_umask.attr,
&format_attr_edge.attr,
&format_attr_inv.attr,
&format_attr_cmask.attr,
NULL,
};

static __initconst struct x86_pmu knc_pmu = {
.name = "knc",
.handle_irq = x86_pmu_handle_irq,
.disable_all = knc_pmu_disable_all,
.enable_all = knc_pmu_enable_all,
.enable = knc_pmu_enable_event,
.disable = knc_pmu_disable_event,
.hw_config = x86_pmu_hw_config,
.schedule_events = x86_schedule_events,
.eventsel = MSR_KNC_EVNTSEL0,
.perfctr = MSR_KNC_PERFCTR0,
.event_map = knc_pmu_event_map,
.max_events = ARRAY_SIZE(knc_perfmon_event_map),
.apic = 1,
.max_period = (1ULL << 31) - 1,
.version = 0,
.num_counters = 2,
/* in theory 40 bits, early silicon is buggy though */
.cntval_bits = 32,
.cntval_mask = (1ULL << 32) - 1,
.get_event_constraints = x86_get_event_constraints,
.event_constraints = knc_event_constraints,
.format_attrs = intel_knc_formats_attr,
};

__init int knc_pmu_init(void)
{
x86_pmu = knc_pmu;

memcpy(hw_cache_event_ids, knc_hw_cache_event_ids,
sizeof(hw_cache_event_ids));

return 0;
}
Loading

0 comments on commit ade0899

Please sign in to comment.