perf tools: Support PERF_SAMPLE_WEIGHT_STRUCT
The new sample type, PERF_SAMPLE_WEIGHT_STRUCT, is an alternative of the PERF_SAMPLE_WEIGHT sample type. Users can apply either the PERF_SAMPLE_WEIGHT sample type or the PERF_SAMPLE_WEIGHT_STRUCT sample type to retrieve the sample weight, but they cannot apply both sample types simultaneously. The new sample type shares the same space as the PERF_SAMPLE_WEIGHT sample type. The lower 32 bits are exactly the same for both sample type. The higher 32 bits may be different for different architecture. Add arch specific arch_evsel__set_sample_weight() to set the new sample type for X86. Only store the lower 32 bits for the sample->weight if the new sample type is applied. In practice, no memory access could last than 4G cycles. No data will be lost. If the kernel doesn't support the new sample type. Fall back to the PERF_SAMPLE_WEIGHT sample type. There is no impact for other architectures. Committer notes: Fixup related to PERF_SAMPLE_CODE_PAGE_SIZE, present in acme/perf/core but not upstream yet. Signed-off-by: Kan Liang <kan.liang@linux.intel.com> Cc: Andi Kleen <ak@linux.intel.com> Cc: Jin Yao <yao.jin@linux.intel.com> Cc: Jiri Olsa <jolsa@redhat.com> Cc: Madhavan Srinivasan <maddy@linux.vnet.ibm.com> Cc: Namhyung Kim <namhyung@kernel.org> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Stephane Eranian <eranian@google.com> Link: http://lore.kernel.org/lkml/1612296553-21962-6-git-send-email-kan.liang@linux.intel.com Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
This commit is contained in:
Родитель
d9d5d767b2
Коммит
ea8d0ed6ea
|
@ -8,6 +8,7 @@ perf-y += machine.o
|
|||
perf-y += event.o
|
||||
perf-y += evlist.o
|
||||
perf-y += mem-events.o
|
||||
perf-y += evsel.o
|
||||
|
||||
perf-$(CONFIG_DWARF) += dwarf-regs.o
|
||||
perf-$(CONFIG_BPF_PROLOGUE) += dwarf-regs.o
|
||||
|
|
|
@ -0,0 +1,8 @@
|
|||
// SPDX-License-Identifier: GPL-2.0
|
||||
#include <stdio.h>
|
||||
#include "util/evsel.h"
|
||||
|
||||
void arch_evsel__set_sample_weight(struct evsel *evsel)
|
||||
{
|
||||
evsel__set_sample_bit(evsel, WEIGHT_STRUCT);
|
||||
}
|
|
@ -1014,6 +1014,11 @@ struct evsel_config_term *__evsel__get_config_term(struct evsel *evsel, enum evs
|
|||
return found_term;
|
||||
}
|
||||
|
||||
void __weak arch_evsel__set_sample_weight(struct evsel *evsel)
|
||||
{
|
||||
evsel__set_sample_bit(evsel, WEIGHT);
|
||||
}
|
||||
|
||||
/*
|
||||
* The enable_on_exec/disabled value strategy:
|
||||
*
|
||||
|
@ -1168,7 +1173,7 @@ void evsel__config(struct evsel *evsel, struct record_opts *opts,
|
|||
}
|
||||
|
||||
if (opts->sample_weight)
|
||||
evsel__set_sample_bit(evsel, WEIGHT);
|
||||
arch_evsel__set_sample_weight(evsel);
|
||||
|
||||
attr->task = track;
|
||||
attr->mmap = track;
|
||||
|
@ -1743,6 +1748,10 @@ static int evsel__open_cpu(struct evsel *evsel, struct perf_cpu_map *cpus,
|
|||
}
|
||||
|
||||
fallback_missing_features:
|
||||
if (perf_missing_features.weight_struct) {
|
||||
evsel__set_sample_bit(evsel, WEIGHT);
|
||||
evsel__reset_sample_bit(evsel, WEIGHT_STRUCT);
|
||||
}
|
||||
if (perf_missing_features.clockid_wrong)
|
||||
evsel->core.attr.clockid = CLOCK_MONOTONIC; /* should always work */
|
||||
if (perf_missing_features.clockid) {
|
||||
|
@ -1883,7 +1892,12 @@ try_fallback:
|
|||
* Must probe features in the order they were added to the
|
||||
* perf_event_attr interface.
|
||||
*/
|
||||
if (!perf_missing_features.code_page_size &&
|
||||
if (!perf_missing_features.weight_struct &&
|
||||
(evsel->core.attr.sample_type & PERF_SAMPLE_WEIGHT_STRUCT)) {
|
||||
perf_missing_features.weight_struct = true;
|
||||
pr_debug2("switching off weight struct support\n");
|
||||
goto fallback_missing_features;
|
||||
} else if (!perf_missing_features.code_page_size &&
|
||||
(evsel->core.attr.sample_type & PERF_SAMPLE_CODE_PAGE_SIZE)) {
|
||||
perf_missing_features.code_page_size = true;
|
||||
pr_debug2_peo("Kernel has no PERF_SAMPLE_CODE_PAGE_SIZE support, bailing out\n");
|
||||
|
@ -2331,9 +2345,15 @@ int evsel__parse_sample(struct evsel *evsel, union perf_event *event,
|
|||
}
|
||||
}
|
||||
|
||||
if (type & PERF_SAMPLE_WEIGHT) {
|
||||
if (type & PERF_SAMPLE_WEIGHT_TYPE) {
|
||||
union perf_sample_weight weight;
|
||||
|
||||
OVERFLOW_CHECK_u64(array);
|
||||
data->weight = *array;
|
||||
weight.full = *array;
|
||||
if (type & PERF_SAMPLE_WEIGHT)
|
||||
data->weight = weight.full;
|
||||
else
|
||||
data->weight = weight.var1_dw;
|
||||
array++;
|
||||
}
|
||||
|
||||
|
|
|
@ -150,6 +150,7 @@ struct perf_missing_features {
|
|||
bool cgroup;
|
||||
bool data_page_size;
|
||||
bool code_page_size;
|
||||
bool weight_struct;
|
||||
};
|
||||
|
||||
extern struct perf_missing_features perf_missing_features;
|
||||
|
@ -244,6 +245,8 @@ void __evsel__reset_sample_bit(struct evsel *evsel, enum perf_event_sample_forma
|
|||
|
||||
void evsel__set_sample_id(struct evsel *evsel, bool use_sample_identifier);
|
||||
|
||||
void arch_evsel__set_sample_weight(struct evsel *evsel);
|
||||
|
||||
int evsel__set_filter(struct evsel *evsel, const char *filter);
|
||||
int evsel__append_tp_filter(struct evsel *evsel, const char *filter);
|
||||
int evsel__append_addr_filter(struct evsel *evsel, const char *filter);
|
||||
|
|
|
@ -1853,13 +1853,29 @@ static int intel_pt_synth_pebs_sample(struct intel_pt_queue *ptq)
|
|||
if (sample_type & PERF_SAMPLE_ADDR && items->has_mem_access_address)
|
||||
sample.addr = items->mem_access_address;
|
||||
|
||||
if (sample_type & PERF_SAMPLE_WEIGHT) {
|
||||
if (sample_type & PERF_SAMPLE_WEIGHT_TYPE) {
|
||||
/*
|
||||
* Refer kernel's setup_pebs_adaptive_sample_data() and
|
||||
* intel_hsw_weight().
|
||||
*/
|
||||
if (items->has_mem_access_latency)
|
||||
sample.weight = items->mem_access_latency;
|
||||
if (items->has_mem_access_latency) {
|
||||
u64 weight = items->mem_access_latency >> 32;
|
||||
|
||||
/*
|
||||
* Starts from SPR, the mem access latency field
|
||||
* contains both cache latency [47:32] and instruction
|
||||
* latency [15:0]. The cache latency is the same as the
|
||||
* mem access latency on previous platforms.
|
||||
*
|
||||
* In practice, no memory access could last than 4G
|
||||
* cycles. Use latency >> 32 to distinguish the
|
||||
* different format of the mem access latency field.
|
||||
*/
|
||||
if (weight > 0)
|
||||
sample.weight = weight & 0xffff;
|
||||
else
|
||||
sample.weight = items->mem_access_latency;
|
||||
}
|
||||
if (!sample.weight && items->has_tsx_aux_info) {
|
||||
/* Cycles last block */
|
||||
sample.weight = (u32)items->tsx_aux_info;
|
||||
|
|
|
@ -36,6 +36,7 @@ static void __p_sample_type(char *buf, size_t size, u64 value)
|
|||
bit_name(IDENTIFIER), bit_name(REGS_INTR), bit_name(DATA_SRC),
|
||||
bit_name(WEIGHT), bit_name(PHYS_ADDR), bit_name(AUX),
|
||||
bit_name(CGROUP), bit_name(DATA_PAGE_SIZE), bit_name(CODE_PAGE_SIZE),
|
||||
bit_name(WEIGHT_STRUCT),
|
||||
{ .name = NULL, }
|
||||
};
|
||||
#undef bit_name
|
||||
|
|
|
@ -1300,7 +1300,7 @@ static void dump_sample(struct evsel *evsel, union perf_event *event,
|
|||
if (sample_type & PERF_SAMPLE_STACK_USER)
|
||||
stack_user__printf(&sample->user_stack);
|
||||
|
||||
if (sample_type & PERF_SAMPLE_WEIGHT)
|
||||
if (sample_type & PERF_SAMPLE_WEIGHT_TYPE)
|
||||
printf("... weight: %" PRIu64 "\n", sample->weight);
|
||||
|
||||
if (sample_type & PERF_SAMPLE_DATA_SRC)
|
||||
|
|
|
@ -1468,7 +1468,7 @@ size_t perf_event__sample_event_size(const struct perf_sample *sample, u64 type,
|
|||
}
|
||||
}
|
||||
|
||||
if (type & PERF_SAMPLE_WEIGHT)
|
||||
if (type & PERF_SAMPLE_WEIGHT_TYPE)
|
||||
result += sizeof(u64);
|
||||
|
||||
if (type & PERF_SAMPLE_DATA_SRC)
|
||||
|
@ -1642,8 +1642,10 @@ int perf_event__synthesize_sample(union perf_event *event, u64 type, u64 read_fo
|
|||
}
|
||||
}
|
||||
|
||||
if (type & PERF_SAMPLE_WEIGHT) {
|
||||
if (type & PERF_SAMPLE_WEIGHT_TYPE) {
|
||||
*array = sample->weight;
|
||||
if (type & PERF_SAMPLE_WEIGHT_STRUCT)
|
||||
*array &= 0xffffffff;
|
||||
array++;
|
||||
}
|
||||
|
||||
|
|
Загрузка…
Ссылка в новой задаче