mirror of
https://mirrors.bfsu.edu.cn/git/linux.git
synced 2024-12-30 06:25:03 +08:00
2a57d40832
On the Intel Sapphire Rapids server, an auxiliary event has to be
enabled simultaneously with the load latency event to retrieve complete
Memory Info.
Add X86 specific perf_mem_events__name() to handle the auxiliary event.
- Users are only interested in the samples of the mem-loads event.
Sample read the auxiliary event.
- The auxiliary event must be in front of the load latency event in a
group. Assume the second event to sample if the auxiliary event is the
leader.
- Add a weak is_mem_loads_aux_event() to check the auxiliary event for
X86. For other ARCHs, it always return false.
Parse the unique event name, mem-loads-aux, for the auxiliary event.
Committer notes:
According to 61b985e3e7
("perf/x86/intel: Add perf core PMU
support for Sapphire Rapids"), ENODATA is only returned by
sys_perf_event_open() when used with these auxiliary events, with this
in evsel__open_strerror():
case ENODATA:
return scnprintf(msg, size, "Cannot collect data source with the load latency event alone. "
"Please add an auxiliary event in front of the load latency event.");
This is Ok at this point in time, but fragile long term, I pointed this
out in the e-mail thread, requesting a follow up patch to check if
ENODATA is really for this specific case.
Fixed up sizeof(MEM_LOADS_AUX_NAME) bug pointed out by Namhyung.
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/20210205152648.GC920417@kernel.org
Link: http://lore.kernel.org/lkml/1612296553-21962-3-git-send-email-kan.liang@linux.intel.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
290 lines
7.3 KiB
C
290 lines
7.3 KiB
C
// SPDX-License-Identifier: GPL-2.0
|
|
#include "debug.h"
|
|
#include "evlist.h"
|
|
#include "evsel.h"
|
|
#include "evsel_config.h"
|
|
#include "parse-events.h"
|
|
#include <errno.h>
|
|
#include <limits.h>
|
|
#include <stdlib.h>
|
|
#include <api/fs/fs.h>
|
|
#include <subcmd/parse-options.h>
|
|
#include <perf/cpumap.h>
|
|
#include "cloexec.h"
|
|
#include "util/perf_api_probe.h"
|
|
#include "record.h"
|
|
#include "../perf-sys.h"
|
|
#include "topdown.h"
|
|
#include "map_symbol.h"
|
|
#include "mem-events.h"
|
|
|
|
/*
|
|
* evsel__config_leader_sampling() uses special rules for leader sampling.
|
|
* However, if the leader is an AUX area event, then assume the event to sample
|
|
* is the next event.
|
|
*/
|
|
static struct evsel *evsel__read_sampler(struct evsel *evsel, struct evlist *evlist)
|
|
{
|
|
struct evsel *leader = evsel->leader;
|
|
|
|
if (evsel__is_aux_event(leader) || arch_topdown_sample_read(leader) ||
|
|
is_mem_loads_aux_event(leader)) {
|
|
evlist__for_each_entry(evlist, evsel) {
|
|
if (evsel->leader == leader && evsel != evsel->leader)
|
|
return evsel;
|
|
}
|
|
}
|
|
|
|
return leader;
|
|
}
|
|
|
|
static u64 evsel__config_term_mask(struct evsel *evsel)
|
|
{
|
|
struct evsel_config_term *term;
|
|
struct list_head *config_terms = &evsel->config_terms;
|
|
u64 term_types = 0;
|
|
|
|
list_for_each_entry(term, config_terms, list) {
|
|
term_types |= 1 << term->type;
|
|
}
|
|
return term_types;
|
|
}
|
|
|
|
static void evsel__config_leader_sampling(struct evsel *evsel, struct evlist *evlist)
|
|
{
|
|
struct perf_event_attr *attr = &evsel->core.attr;
|
|
struct evsel *leader = evsel->leader;
|
|
struct evsel *read_sampler;
|
|
u64 term_types, freq_mask;
|
|
|
|
if (!leader->sample_read)
|
|
return;
|
|
|
|
read_sampler = evsel__read_sampler(evsel, evlist);
|
|
|
|
if (evsel == read_sampler)
|
|
return;
|
|
|
|
term_types = evsel__config_term_mask(evsel);
|
|
/*
|
|
* Disable sampling for all group members except those with explicit
|
|
* config terms or the leader. In the case of an AUX area event, the 2nd
|
|
* event in the group is the one that 'leads' the sampling.
|
|
*/
|
|
freq_mask = (1 << EVSEL__CONFIG_TERM_FREQ) | (1 << EVSEL__CONFIG_TERM_PERIOD);
|
|
if ((term_types & freq_mask) == 0) {
|
|
attr->freq = 0;
|
|
attr->sample_freq = 0;
|
|
attr->sample_period = 0;
|
|
}
|
|
if ((term_types & (1 << EVSEL__CONFIG_TERM_OVERWRITE)) == 0)
|
|
attr->write_backward = 0;
|
|
|
|
/*
|
|
* We don't get a sample for slave events, we make them when delivering
|
|
* the group leader sample. Set the slave event to follow the master
|
|
* sample_type to ease up reporting.
|
|
* An AUX area event also has sample_type requirements, so also include
|
|
* the sample type bits from the leader's sample_type to cover that
|
|
* case.
|
|
*/
|
|
attr->sample_type = read_sampler->core.attr.sample_type |
|
|
leader->core.attr.sample_type;
|
|
}
|
|
|
|
void evlist__config(struct evlist *evlist, struct record_opts *opts, struct callchain_param *callchain)
|
|
{
|
|
struct evsel *evsel;
|
|
bool use_sample_identifier = false;
|
|
bool use_comm_exec;
|
|
bool sample_id = opts->sample_id;
|
|
|
|
/*
|
|
* Set the evsel leader links before we configure attributes,
|
|
* since some might depend on this info.
|
|
*/
|
|
if (opts->group)
|
|
evlist__set_leader(evlist);
|
|
|
|
if (evlist->core.cpus->map[0] < 0)
|
|
opts->no_inherit = true;
|
|
|
|
use_comm_exec = perf_can_comm_exec();
|
|
|
|
evlist__for_each_entry(evlist, evsel) {
|
|
evsel__config(evsel, opts, callchain);
|
|
if (evsel->tracking && use_comm_exec)
|
|
evsel->core.attr.comm_exec = 1;
|
|
}
|
|
|
|
/* Configure leader sampling here now that the sample type is known */
|
|
evlist__for_each_entry(evlist, evsel)
|
|
evsel__config_leader_sampling(evsel, evlist);
|
|
|
|
if (opts->full_auxtrace) {
|
|
/*
|
|
* Need to be able to synthesize and parse selected events with
|
|
* arbitrary sample types, which requires always being able to
|
|
* match the id.
|
|
*/
|
|
use_sample_identifier = perf_can_sample_identifier();
|
|
sample_id = true;
|
|
} else if (evlist->core.nr_entries > 1) {
|
|
struct evsel *first = evlist__first(evlist);
|
|
|
|
evlist__for_each_entry(evlist, evsel) {
|
|
if (evsel->core.attr.sample_type == first->core.attr.sample_type)
|
|
continue;
|
|
use_sample_identifier = perf_can_sample_identifier();
|
|
break;
|
|
}
|
|
sample_id = true;
|
|
}
|
|
|
|
if (sample_id) {
|
|
evlist__for_each_entry(evlist, evsel)
|
|
evsel__set_sample_id(evsel, use_sample_identifier);
|
|
}
|
|
|
|
evlist__set_id_pos(evlist);
|
|
}
|
|
|
|
static int get_max_rate(unsigned int *rate)
|
|
{
|
|
return sysctl__read_int("kernel/perf_event_max_sample_rate", (int *)rate);
|
|
}
|
|
|
|
static int record_opts__config_freq(struct record_opts *opts)
|
|
{
|
|
bool user_freq = opts->user_freq != UINT_MAX;
|
|
unsigned int max_rate;
|
|
|
|
if (opts->user_interval != ULLONG_MAX)
|
|
opts->default_interval = opts->user_interval;
|
|
if (user_freq)
|
|
opts->freq = opts->user_freq;
|
|
|
|
/*
|
|
* User specified count overrides default frequency.
|
|
*/
|
|
if (opts->default_interval)
|
|
opts->freq = 0;
|
|
else if (opts->freq) {
|
|
opts->default_interval = opts->freq;
|
|
} else {
|
|
pr_err("frequency and count are zero, aborting\n");
|
|
return -1;
|
|
}
|
|
|
|
if (get_max_rate(&max_rate))
|
|
return 0;
|
|
|
|
/*
|
|
* User specified frequency is over current maximum.
|
|
*/
|
|
if (user_freq && (max_rate < opts->freq)) {
|
|
if (opts->strict_freq) {
|
|
pr_err("error: Maximum frequency rate (%'u Hz) exceeded.\n"
|
|
" Please use -F freq option with a lower value or consider\n"
|
|
" tweaking /proc/sys/kernel/perf_event_max_sample_rate.\n",
|
|
max_rate);
|
|
return -1;
|
|
} else {
|
|
pr_warning("warning: Maximum frequency rate (%'u Hz) exceeded, throttling from %'u Hz to %'u Hz.\n"
|
|
" The limit can be raised via /proc/sys/kernel/perf_event_max_sample_rate.\n"
|
|
" The kernel will lower it when perf's interrupts take too long.\n"
|
|
" Use --strict-freq to disable this throttling, refusing to record.\n",
|
|
max_rate, opts->freq, max_rate);
|
|
|
|
opts->freq = max_rate;
|
|
}
|
|
}
|
|
|
|
/*
|
|
* Default frequency is over current maximum.
|
|
*/
|
|
if (max_rate < opts->freq) {
|
|
pr_warning("Lowering default frequency rate from %u to %u.\n"
|
|
"Please consider tweaking "
|
|
"/proc/sys/kernel/perf_event_max_sample_rate.\n",
|
|
opts->freq, max_rate);
|
|
opts->freq = max_rate;
|
|
}
|
|
|
|
return 0;
|
|
}
|
|
|
|
int record_opts__config(struct record_opts *opts)
|
|
{
|
|
return record_opts__config_freq(opts);
|
|
}
|
|
|
|
bool evlist__can_select_event(struct evlist *evlist, const char *str)
|
|
{
|
|
struct evlist *temp_evlist;
|
|
struct evsel *evsel;
|
|
int err, fd, cpu;
|
|
bool ret = false;
|
|
pid_t pid = -1;
|
|
|
|
temp_evlist = evlist__new();
|
|
if (!temp_evlist)
|
|
return false;
|
|
|
|
err = parse_events(temp_evlist, str, NULL);
|
|
if (err)
|
|
goto out_delete;
|
|
|
|
evsel = evlist__last(temp_evlist);
|
|
|
|
if (!evlist || perf_cpu_map__empty(evlist->core.cpus)) {
|
|
struct perf_cpu_map *cpus = perf_cpu_map__new(NULL);
|
|
|
|
cpu = cpus ? cpus->map[0] : 0;
|
|
perf_cpu_map__put(cpus);
|
|
} else {
|
|
cpu = evlist->core.cpus->map[0];
|
|
}
|
|
|
|
while (1) {
|
|
fd = sys_perf_event_open(&evsel->core.attr, pid, cpu, -1,
|
|
perf_event_open_cloexec_flag());
|
|
if (fd < 0) {
|
|
if (pid == -1 && errno == EACCES) {
|
|
pid = 0;
|
|
continue;
|
|
}
|
|
goto out_delete;
|
|
}
|
|
break;
|
|
}
|
|
close(fd);
|
|
ret = true;
|
|
|
|
out_delete:
|
|
evlist__delete(temp_evlist);
|
|
return ret;
|
|
}
|
|
|
|
int record__parse_freq(const struct option *opt, const char *str, int unset __maybe_unused)
|
|
{
|
|
unsigned int freq;
|
|
struct record_opts *opts = opt->value;
|
|
|
|
if (!str)
|
|
return -EINVAL;
|
|
|
|
if (strcasecmp(str, "max") == 0) {
|
|
if (get_max_rate(&freq)) {
|
|
pr_err("couldn't read /proc/sys/kernel/perf_event_max_sample_rate\n");
|
|
return -1;
|
|
}
|
|
pr_info("info: Using a maximum frequency rate of %'d Hz\n", freq);
|
|
} else {
|
|
freq = atoi(str);
|
|
}
|
|
|
|
opts->user_freq = freq;
|
|
return 0;
|
|
}
|