perf tools: Enable AUX area tracing snapshots using a FIFO
Adrian Hunter
reviewercover perf tools: Enable AUX area tracing snapshots using a FIFO
Hi
Here are some patches to enable AUX area tracing snapshots using a FIFO. Presently SIGUSR2 can be used but the advantage of the FIFO is that access is governed by access to the FIFO. Refer to the example in patch 5.
The first 4 patches are preparation, including patch 4 which enhances the --control option to accept file names.
Patch 5 adds snapshot control command.
The final patch updates Intel PT documentation.
Adrian Hunter (6):
perf tools: Consolidate --control option parsing into one function perf tools: Handle read errors from ctl_fd
perf tools: Use AsciiDoc formatting for --control option documentation perf tools: Add FIFO file names as alternative options to --control perf record: Add 'snapshot' control command
perf intel-pt: Document snapshot control command
tools/perf/Documentation/perf-intel-pt.txt | 23 ++++++- tools/perf/Documentation/perf-record.txt | 56 ++++++++--------- tools/perf/Documentation/perf-stat.txt | 48 ++++++++------- tools/perf/builtin-record.c | 74 +++++++++++++---------- tools/perf/builtin-stat.c | 35 +++++------ tools/perf/util/evlist.c | 96 +++++++++++++++++++++++++++--- tools/perf/util/evlist.h | 6 +- tools/perf/util/record.h | 1 + tools/perf/util/stat.h | 1 + 9 files changed, 227 insertions(+), 113 deletions(-)
Regards
Adrian
Adrian Hunter
reviewerPATCH 1/6 perf tools: Consolidate --control option parsing into one function
Consolidate --control option parsing into one function, in preparation for adding FIFO file name options.
--- tools/perf/builtin-record.c | 22 ++-------------------- tools/perf/builtin-stat.c | 22 ++-------------------- tools/perf/util/evlist.c | 24 ++++++++++++++++++++++++ tools/perf/util/evlist.h | 1 + 4 files changed, 29 insertions(+), 40 deletions(-)
tools/perf/builtin-record.c +2 -20
@@ -2234,27 +2234,9 @@ static int parse_control_option(const struct option *opt, const char *str, int unset __maybe_unused) { - char *comma = NULL, *endptr = NULL; -struct record_opts *config = (struct record_opts *)opt->value; - -if (strncmp (str,"fd:" ,3 )) -return -EINVAL ; - - config->ctl_fd =strtoul (&str[3 ], &endptr,0 ); -if (endptr == &str[3 ]) -return -EINVAL ; - - comma =strchr (str,',' ); -if (comma) { -if (endptr != comma) -return -EINVAL ; - - config->ctl_fd_ack =strtoul (comma +1 , &endptr,0 ); -if (endptr == comma +1 || *endptr !='\0' ) -return -EINVAL ; - } +struct record_opts *opts = opt->value; -return 0 ; +return evlist__parse_control (str, &opts->ctl_fd, &opts->ctl_fd_ack); }static void switch_output_size_warn (struct record *rec)
tools/perf/builtin-stat.c +2 -20
@@ -1045,27 +1045,9 @@ static int parse_control_option(const struct option *opt, const char *str, int unset __maybe_unused) { - char *comma = NULL, *endptr = NULL; -struct perf_stat_config *config = (struct perf_stat_config *)opt->value; +struct perf_stat_config *config = opt->value; -if (strncmp (str,"fd:" ,3 )) -return -EINVAL ; - - config->ctl_fd =strtoul (&str[3 ], &endptr,0 ); -if (endptr == &str[3 ]) -return -EINVAL ; - - comma =strchr (str,',' ); -if (comma) { -if (endptr != comma) -return -EINVAL ; - - config->ctl_fd_ack =strtoul (comma +1 , &endptr,0 ); -if (endptr == comma +1 || *endptr !='\0' ) -return -EINVAL ; - } - -return 0 ; +return evlist__parse_control (str, &config->ctl_fd, &config->ctl_fd_ack); }static struct option stat_options[] = {
tools/perf/util/evlist.c +24 -0
@@ -1727,6 +1727,30 @@ struct evsel *perf_evlist__reset_weak_group(struct evlist *evsel_list,return leader; } +int evlist__parse_control (const char *str,int *ctl_fd,int *ctl_fd_ack) +{ +char *comma =NULL , *endptr =NULL ; + +if (strncmp (str,"fd:" ,3 )) +return -EINVAL ; + + *ctl_fd =strtoul (&str[3 ], &endptr,0 ); +if (endptr == &str[3 ]) +return -EINVAL ; + + comma =strchr (str,',' ); +if (comma) { +if (endptr != comma) +return -EINVAL ; + + *ctl_fd_ack =strtoul (comma +1 , &endptr,0 ); +if (endptr == comma +1 || *endptr !='\0' ) +return -EINVAL ; + } + +return 0 ; +} +int evlist__initialize_ctlfd (struct evlist *evlist,int fd,int ack) {if (fd ==-1 ) {
tools/perf/util/evlist.h +1 -0
@@ -373,6 +373,7 @@ enum evlist_ctl_cmd { EVLIST_CTL_CMD_ACK }; +int evlist__parse_control (const char *str,int *ctl_fd,int *ctl_fd_ack);int evlist__initialize_ctlfd (struct evlist *evlist,int ctl_fd,int ctl_fd_ack);int evlist__finalize_ctlfd (struct evlist *evlist);bool evlist__ctlfd_initialized (struct evlist *evlist);
Signature · 1 line
2.17.1
Adrian Hunter
reviewerPATCH 2/6 perf tools: Handle read errors from ctl_fd
In reply to Adrian HunterHandle read errors from ctl_fd such as EINTR, EAGAIN and EWOULDBLOCK.
--- tools/perf/util/evlist.c | 16 +++++++++++----- 1 file changed, 11 insertions(+), 5 deletions(-)
tools/perf/util/evlist.c +11 -5
@@ -1802,6 +1802,7 @@ static int evlist__ctlfd_recv(struct evlist *evlist, enum evlist_ctl_cmd *cmd,char c;size_t bytes_read =0 ; + *cmd =EVLIST_CTL_CMD_UNSUPPORTED ;memset (cmd_data,0 , data_size); data_size--;
@@ -1813,17 +1814,22 @@ static int evlist__ctlfd_recv(struct evlist *evlist, enum evlist_ctl_cmd *cmd, cmd_data[bytes_read++] = c;if (bytes_read == data_size)break ; - }else { -if (err ==-1 ) +continue ; + }else if (err ==-1 ) { +if (errno ==EINTR ) +continue ; +if (errno ==EAGAIN || errno ==EWOULDBLOCK ) + err =0 ; +else pr_err ("Failed to read from ctlfd %d: %m\n" , evlist->ctl_fd.fd); -break ; } +break ; }while (1 );pr_debug ("Message from ctl_fd: \"%s%s\"\n" , cmd_data, bytes_read == data_size ?"" : c =='\n' ?"\\n" :"\\0" ); -if (err >0 ) { +if (bytes_read >0 ) {if (!strncmp (cmd_data,EVLIST_CTL_CMD_ENABLE_TAG , (sizeof (EVLIST_CTL_CMD_ENABLE_TAG )-1 ))) { *cmd =EVLIST_CTL_CMD_ENABLE ;
@@ -1833,7 +1839,7 @@ static int evlist__ctlfd_recv(struct evlist *evlist, enum evlist_ctl_cmd *cmd, } } -return err; +return bytes_read ? (int )bytes_read : err; }static int evlist__ctlfd_ack (struct evlist *evlist)
Signature · 1 line
2.17.1
Adrian Hunter
reviewerPATCH 3/6 perf tools: Use AsciiDoc formatting for --control option documentation
In reply to Adrian HunterThe --control option does not display well in man pages unless AsciiDoc formatting is used.
--- tools/perf/Documentation/perf-record.txt | 46 ++++++++++++------------ tools/perf/Documentation/perf-stat.txt | 46 ++++++++++++------------ 2 files changed, 46 insertions(+), 46 deletions(-)
tools/perf/Documentation/perf-record.txt +23 -23
@@ -627,43 +627,43 @@ option. The -e option and this one can be mixed and matched. Events can be grouped using the {} notation. endif::HAVE_LIBPFM[] ---control fd:ctl-fd[,ack-fd] +--control=fd:ctl-fd[,ack-fd]:: Listen on ctl-fd descriptor for command to control measurement ('enable': enable events, 'disable': disable events). Measurements can be started with events disabled using --delay=-1 option. Optionally send control command completion ('ack\n') to ack-fd descriptor to synchronize with the controlling process. Example of bash shell script to enable and disable events during measurements: -#!/bin/bash + #!/bin/bash -ctl_dir=/tmp/ + ctl_dir=/tmp/ -ctl_fifo=${ctl_dir}perf_ctl.fifo -test -p ${ctl_fifo} && unlink ${ctl_fifo} -mkfifo ${ctl_fifo} -exec {ctl_fd}<>${ctl_fifo} + ctl_fifo=${ctl_dir}perf_ctl.fifo + test -p ${ctl_fifo} && unlink ${ctl_fifo} + mkfifo ${ctl_fifo} + exec {ctl_fd}<>${ctl_fifo} -ctl_ack_fifo=${ctl_dir}perf_ctl_ack.fifo -test -p ${ctl_ack_fifo} && unlink ${ctl_ack_fifo} -mkfifo ${ctl_ack_fifo} -exec {ctl_fd_ack}<>${ctl_ack_fifo} + ctl_ack_fifo=${ctl_dir}perf_ctl_ack.fifo + test -p ${ctl_ack_fifo} && unlink ${ctl_ack_fifo} + mkfifo ${ctl_ack_fifo} + exec {ctl_fd_ack}<>${ctl_ack_fifo} -perf record -D -1 -e cpu-cycles -a \ - --control fd:${ctl_fd},${ctl_fd_ack} \ - -- sleep 30 & -perf_pid=$! + perf record -D -1 -e cpu-cycles -a \ + --control fd:${ctl_fd},${ctl_fd_ack} \ + -- sleep 30 & + perf_pid=$! -sleep 5 && echo 'enable' >&${ctl_fd} && read -u ${ctl_fd_ack} e1 && echo "enabled(${e1})" -sleep 10 && echo 'disable' >&${ctl_fd} && read -u ${ctl_fd_ack} d1 && echo "disabled(${d1})" + sleep 5 && echo 'enable' >&${ctl_fd} && read -u ${ctl_fd_ack} e1 && echo "enabled(${e1})" + sleep 10 && echo 'disable' >&${ctl_fd} && read -u ${ctl_fd_ack} d1 && echo "disabled(${d1})" -exec {ctl_fd_ack}>&- -unlink ${ctl_ack_fifo} + exec {ctl_fd_ack}>&- + unlink ${ctl_ack_fifo} -exec {ctl_fd}>&- -unlink ${ctl_fifo} + exec {ctl_fd}>&- + unlink ${ctl_fifo} -wait -n ${perf_pid} -exit $? + wait -n ${perf_pid} + exit $? SEE ALSO
tools/perf/Documentation/perf-stat.txt +23 -23
66 lines not on this page · raw message
Signature · 1 line
2.17.1
Adrian Hunter
reviewerPATCH 4/6 perf tools: Add FIFO file names as alternative options to --control
In reply to Adrian HunterEnable the --control option to accept file names as an alternative to file descriptors.
Example:
$ mkfifo perf.control $ mkfifo perf.ack $ cat perf.ack & [1] 6808 $ perf record --control perf.control,perf.ack -- sleep 300 & [2] 6810 $ echo disable > perf.control $ Events disabled ack
$ echo enable > perf.control $ Events enabled ack
$ echo disable > perf.control $ Events disabled ack
$ kill %2 [ perf record: Woken up 4 times to write data ] $ [ perf record: Captured and wrote 0.018 MB perf.data (7 samples) ]
[1]- Done cat perf.ack [2]+ Terminated perf record --control perf.control,perf.ack -- sleep 300
$
--- tools/perf/Documentation/perf-record.txt | 2 + tools/perf/Documentation/perf-stat.txt | 2 + tools/perf/builtin-record.c | 34 ++++++++++++---- tools/perf/builtin-stat.c | 18 +++++++-- tools/perf/util/evlist.c | 49 +++++++++++++++++++++++- tools/perf/util/evlist.h | 2 +- tools/perf/util/record.h | 1 + tools/perf/util/stat.h | 1 + 8 files changed, 95 insertions(+), 14 deletions(-)
tools/perf/Documentation/perf-record.txt +2 -0
@@ -627,7 +627,9 @@ option. The -e option and this one can be mixed and matched. Events can be grouped using the {} notation. endif::HAVE_LIBPFM[] +--control=ctl-fifo[,ack-fifo]:: --control=fd:ctl-fd[,ack-fd]:: +ctl-fifo / ack-fifo are opened and used as ctl-fd / ack-fd as follows. Listen on ctl-fd descriptor for command to control measurement ('enable': enable events, 'disable': disable events). Measurements can be started with events disabled using --delay=-1 option. Optionally send control command completion ('ack\n') to ack-fd descriptor
tools/perf/Documentation/perf-stat.txt +2 -0
@@ -176,7 +176,9 @@ with it. --append may be used here. Examples: 3>results perf stat --log-fd 3 -- $cmd 3>>results perf stat --log-fd 3 --append -- $cmd +--control=ctl-fifo[,ack-fifo]:: --control=fd:ctl-fd[,ack-fd]:: +ctl-fifo / ack-fifo are opened and used as ctl-fd / ack-fd as follows. Listen on ctl-fd descriptor for command to control measurement ('enable': enable events, 'disable': disable events). Measurements can be started with events disabled using --delay=-1 option. Optionally send control command completion ('ack\n') to ack-fd descriptor
tools/perf/builtin-record.c +26 -8
@@ -2236,7 +2236,17 @@ static int parse_control_option(const struct option *opt, {struct record_opts *opts = opt->value; -return evlist__parse_control (str, &opts->ctl_fd, &opts->ctl_fd_ack); +return evlist__parse_control (str, &opts->ctl_fd, &opts->ctl_fd_ack, &opts->ctl_fd_close); +} + +static void close_control_option (struct record_opts *opts) +{ +if (opts->ctl_fd_close) { + opts->ctl_fd_close = false; +close (opts->ctl_fd); +if (opts->ctl_fd_ack >=0 ) +close (opts->ctl_fd_ack); + } }static void switch_output_size_warn (struct record *rec)
@@ -2578,9 +2588,10 @@ static struct option __record_options[] = { "libpfm4 event selector. use 'perf list' to list available events", parse_libpfm_events_option),#endif -OPT_CALLBACK (0 ,"control" , &record.opts,"fd:ctl-fd[,ack-fd]" , +OPT_CALLBACK (0 ,"control" , &record.opts,"fd:ctl-fd[,ack-fd] or ctl-fifo[,ack-fifo]" ,"Listen on ctl-fd descriptor for command to control measurement ('enable': enable events, 'disable': disable events).\n" -"\t\t\t Optionally send control command completion ('ack\\n') to ack-fd descriptor." , +"\t\t\t Optionally send control command completion ('ack\\n') to ack-fd descriptor.\n" +"\t\t\t Alternatively, ctl-fifo / ack-fifo will be opened and used as ctl-fd / ack-fd." , parse_control_option), OPT_END() };
@@ -2653,12 +2664,14 @@ int cmd_record(int argc, const char **argv) !perf_can_record_switch_events()) {ui__error ("kernel does not support recording context switch events\n" );parse_options_usage (record_usage, record_options,"switch-events" ,0 ); -return -EINVAL ; + err = -EINVAL ; + goto out_opts; }if (switch_output_setup (rec)) {parse_options_usage (record_usage, record_options,"switch-output" ,0 ); -return -EINVAL ; + err = -EINVAL ; + goto out_opts; }if (rec->switch_output.time) {
@@ -2669,8 +2682,10 @@ int cmd_record(int argc, const char **argv)if (rec->switch_output.num_files) { rec->switch_output.filenames =calloc (sizeof (char *), rec->switch_output.num_files); -if (!rec->switch_output.filenames) -return -EINVAL ; +if (!rec->switch_output.filenames) { + err = -EINVAL ; + goto out_opts; + } }/*
@@ -2686,7 +2701,8 @@ int cmd_record(int argc, const char **argv) rec->affinity_mask.bits =bitmap_alloc (rec->affinity_mask.nbits);if (!rec->affinity_mask.bits) {pr_err ("Failed to allocate thread mask for %zd cpus\n" , rec->affinity_mask.nbits); -return -ENOMEM ; + err = -ENOMEM ; + goto out_opts; }pr_debug2 ("thread mask[%zd]: empty\n" , rec->affinity_mask.nbits); }
@@ -2817,6 +2833,8 @@ int cmd_record(int argc, const char **argv)evlist__delete (rec->evlist);symbol__exit ();auxtrace_record__free (rec->itr); +out_opts: +close_control_option (&rec->opts);return err; }
tools/perf/builtin-stat.c +15 -3
@@ -1047,7 +1047,17 @@ static int parse_control_option(const struct option *opt, {struct perf_stat_config *config = opt->value; -return evlist__parse_control (str, &config->ctl_fd, &config->ctl_fd_ack); +return evlist__parse_control (str, &config->ctl_fd, &config->ctl_fd_ack, &config->ctl_fd_close); +} + +static void close_control_option (struct perf_stat_config *config) +{ +if (config->ctl_fd_close) { + config->ctl_fd_close = false; +close (config->ctl_fd); +if (config->ctl_fd_ack >=0 ) +close (config->ctl_fd_ack); + } }static struct option stat_options[] = {
@@ -1151,9 +1161,10 @@ static struct option stat_options[] = { "libpfm4 event selector. use 'perf list' to list available events", parse_libpfm_events_option),#endif -OPT_CALLBACK (0 ,"control" , &stat_config,"fd:ctl-fd[,ack-fd]" , +OPT_CALLBACK (0 ,"control" , &stat_config,"fd:ctl-fd[,ack-fd] or ctl-fifo[,ack-fifo]" ,"Listen on ctl-fd descriptor for command to control measurement ('enable': enable events, 'disable': disable events).\n" -"\t\t\t Optionally send control command completion ('ack\\n') to ack-fd descriptor." , +"\t\t\t Optionally send control command completion ('ack\\n') to ack-fd descriptor.\n" +"\t\t\t Alternatively, ctl-fifo / ack-fifo will be opened and used as ctl-fd / ack-fd." , parse_control_option), OPT_END() };
@@ -2396,6 +2407,7 @@ int cmd_stat(int argc, const char **argv)metricgroup__rblist_exit (&stat_config.metric_events);runtime_stat_delete (&stat_config); +close_control_option (&stat_config);return status; }
tools/perf/util/evlist.c +47 -2
59 lines not on this page · raw message
tools/perf/util/evlist.h +1 -1
@@ -373,7 +373,7 @@ enum evlist_ctl_cmd { EVLIST_CTL_CMD_ACK }; -int evlist__parse_control (const char *str,int *ctl_fd,int *ctl_fd_ack); +int evlist__parse_control (const char *str,int *ctl_fd,int *ctl_fd_ack,bool *ctl_fd_close);int evlist__initialize_ctlfd (struct evlist *evlist,int ctl_fd,int ctl_fd_ack);int evlist__finalize_ctlfd (struct evlist *evlist);bool evlist__ctlfd_initialized (struct evlist *evlist);
tools/perf/util/record.h +1 -0
@@ -73,6 +73,7 @@ struct record_opts {unsigned int nr_threads_synthesize;int ctl_fd;int ctl_fd_ack; +bool ctl_fd_close; };extern const char *const *record_usage;
tools/perf/util/stat.h +1 -0
@@ -135,6 +135,7 @@ struct perf_stat_config {struct rblist metric_events;int ctl_fd;int ctl_fd_ack; +bool ctl_fd_close; };void perf_stat__set_big_num (int set);
Signature · 1 line
2.17.1
Re: [PATCH 4/6] perf tools: Add FIFO file names as alternative options to --control
On Sat, Aug 29, 2020 at 01:50:13PM +0300, Adrian Hunter wrote:
SNIP
Quoted 28 lines
tools/perf/util/evlist.c +28
-0+ *p ='\0' ; + +/* +* O_RDWR avoids POLLHUPs which is necessary to allow the other +* end of a FIFO to be repeatedly opened and closed. +*/ + fd =open (s,O_RDWR |O_NONBLOCK |O_CLOEXEC ); +if (fd <0 ) { +pr_err ("Failed to open '%s'\n" , s); + ret = -errno; + goto out_free; + } + *ctl_fd = fd; + *ctl_fd_close = true; + +if (p && *++p) { +/* O_RDWR | O_NONBLOCK means the other end need not be open */ + fd =open (p,O_RDWR |O_NONBLOCK |O_CLOEXEC ); +if (fd <0 ) { +pr_err ("Failed to open '%s'\n" , p); + ret = -errno; + goto out_free; + } + *ctl_fd_ack = fd; + } + +out_free: +free (s);
tools/perf/util/evlist.c struct evsel *perf_evlist__reset_weak_group(struct evlist *evsel_list, · line 1780
Quoted 10 lines
tools/perf/util/evlist.c +5
-1+return ret; +} + +int evlist__parse_control (const char *str,int *ctl_fd,int *ctl_fd_ack,bool *ctl_fd_close) {char *comma =NULL , *endptr =NULL ;if (strncmp (str,"fd:" ,3 )) -return -EINVAL ; +return evlist__parse_control_names (str, ctl_fd, ctl_fd_ack, ctl_fd_close);
do we want to mention somewhere that the fifo name is everything except for 'fd:' ?
also how likely is that we will add another channel type that will need another keyword (likd 'fd:')? I originaly thought we'd use 'fifo:filename' for this ... would be great to somehow avoid future confusions
thanks,
jirka
Adrian Hunter
reviewerRe: [PATCH 4/6] perf tools: Add FIFO file names as alternative options to --control
Quoted 39 lines · Jiri Olsa
On Sat, Aug 29, 2020 at 01:50:13PM +0300, Adrian Hunter wrote:
SNIP
Quoted 35 lines
(quoted diff) +32
-0+ *p = '\0'; + + /* + * O_RDWR avoids POLLHUPs which is necessary to allow the other + * end of a FIFO to be repeatedly opened and closed. + */ + fd = open(s, O_RDWR | O_NONBLOCK | O_CLOEXEC); + if (fd < 0) { + pr_err("Failed to open '%s'\n", s); + ret = -errno; + goto out_free; + } + *ctl_fd = fd; + *ctl_fd_close = true; + + if (p && *++p) { + /* O_RDWR | O_NONBLOCK means the other end need not be open */ + fd = open(p, O_RDWR | O_NONBLOCK | O_CLOEXEC); + if (fd < 0) { + pr_err("Failed to open '%s'\n", p); + ret = -errno; + goto out_free; + } + *ctl_fd_ack = fd; + } + +out_free: + free(s); + return ret; +} + +int evlist__parse_control(const char *str, int *ctl_fd, int *ctl_fd_ack, bool *ctl_fd_close) { char *comma = NULL, *endptr = NULL;
Quoted 6 lines
Review
Quoted 3 lines
(quoted diff) +1
-1if (strncmp(str, "fd:", 3)) - return -EINVAL; + return evlist__parse_control_names(str, ctl_fd, ctl_fd_ack, ctl_fd_close);do we want to mention somewhere that the fifo name is everything except for 'fd:' ?
It is only mentioned in the documentation i.e.
tools/perf/Documentation/perf-record.txt +2 -0
@@ -627,7 +627,9 @@ option. The -e option and this one can be mixed and matched. Events can be grouped using the {} notation. endif::HAVE_LIBPFM[] +--control=ctl-fifo[,ack-fifo]:: --control=fd:ctl-fd[,ack-fd]:: +ctl-fifo / ack-fifo are opened and used as ctl-fd / ack-fd as follows. Listen on ctl-fd descriptor for command to control measurement ('enable': enable events, 'disable': disable events). Measurements can be started with events disabled using --delay=-1 option. Optionally send control command completion ('ack\n') to ack-fd descriptor
Quoted 5 lines
also how likely is that we will add another channel type that will need another keyword (likd 'fd:')? I originaly thought we'd use 'fifo:filename' for this ... would be great to somehow avoid future confusions
Sure, I will add fifo: in V2
Adrian Hunter
reviewerPATCH 5/6 perf record: Add 'snapshot' control command
In reply to Adrian HunterAdd 'snapshot' control command to create an AUX area tracing snapshot the same as if sending SIGUSR2. The advantage of the FIFO is that access is governed by access to the FIFO.
Example:
$ mkfifo perf.control $ mkfifo perf.ack $ cat perf.ack & [1] 15235 $ sudo ~/bin/perf record --control perf.control,perf.ack -S -e intel_pt//u -- sleep 60 & [2] 15243 $ ps -e | grep perf 15244 pts/1 00:00:00 perf $ kill -USR2 15244 bash: kill: (15244) - Operation not permitted $ echo snapshot > perf.control ack
--- tools/perf/Documentation/perf-record.txt | 8 ++++---- tools/perf/builtin-record.c | 24 +++++++++++++++++------- tools/perf/builtin-stat.c | 1 + tools/perf/util/evlist.c | 11 +++++++++-- tools/perf/util/evlist.h | 5 ++++- 5 files changed, 35 insertions(+), 14 deletions(-)
tools/perf/Documentation/perf-record.txt +4 -4
@@ -631,10 +631,10 @@ endif::HAVE_LIBPFM[] --control=fd:ctl-fd[,ack-fd]:: ctl-fifo / ack-fifo are opened and used as ctl-fd / ack-fd as follows. Listen on ctl-fd descriptor for command to control measurement ('enable': enable events, -'disable': disable events). Measurements can be started with events disabled using ---delay=-1 option. Optionally send control command completion ('ack\n') to ack-fd descriptor -to synchronize with the controlling process. Example of bash shell script to enable and -disable events during measurements: +'disable': disable events, 'snapshot': AUX area tracing snapshot). Measurements can be +started with events disabled using --delay=-1 option. Optionally send control command +completion ('ack\n') to ack-fd descriptor to synchronize with the controlling process. +Example of bash shell script to enable and disable events during measurements: #!/bin/bash
tools/perf/builtin-record.c +17 -7
@@ -1593,6 +1593,16 @@ static int record__init_clock(struct record *rec)return 0 ; } +static void hit_auxtrace_snapshot_trigger (struct record *rec) +{ +if (trigger_is_ready (&auxtrace_snapshot_trigger)) { +trigger_hit (&auxtrace_snapshot_trigger); + auxtrace_record__snapshot_started =1 ; +if (auxtrace_record__snapshot_start (rec->itr)) +trigger_error (&auxtrace_snapshot_trigger); + } +} +static int __cmd_record (struct record *rec,int argc,const char **argv) {int err;
@@ -1937,6 +1947,10 @@ static int __cmd_record(struct record *rec, int argc, const char **argv)case EVLIST_CTL_CMD_DISABLE :pr_info (EVLIST_DISABLED_MSG );break ; +case EVLIST_CTL_CMD_SNAPSHOT : +hit_auxtrace_snapshot_trigger (rec); +evlist__ctlfd_ack (rec->evlist); +break ;case EVLIST_CTL_CMD_ACK :case EVLIST_CTL_CMD_UNSUPPORTED :default :
@@ -2589,7 +2603,8 @@ static struct option __record_options[] = { parse_libpfm_events_option), #endifOPT_CALLBACK (0 ,"control" , &record.opts,"fd:ctl-fd[,ack-fd] or ctl-fifo[,ack-fifo]" , -"Listen on ctl-fd descriptor for command to control measurement ('enable': enable events, 'disable': disable events).\n" +"Listen on ctl-fd descriptor for command to control measurement ('enable': enable events, 'disable': disable events,\n" +"\t\t\t 'snapshot': AUX area tracing snapshot).\n" "\t\t\t Optionally send control command completion ('ack\\n') to ack-fd descriptor.\n" "\t\t\t Alternatively, ctl-fifo / ack-fifo will be opened and used as ctl-fd / ack-fd." , parse_control_option),
@@ -2842,12 +2857,7 @@ static void snapshot_sig_handler(int sig __maybe_unused) {struct record *rec = &record; -if (trigger_is_ready (&auxtrace_snapshot_trigger)) { -trigger_hit (&auxtrace_snapshot_trigger); - auxtrace_record__snapshot_started =1 ; -if (auxtrace_record__snapshot_start (record.itr)) -trigger_error (&auxtrace_snapshot_trigger); - } +hit_auxtrace_snapshot_trigger (rec);if (switch_output_signal (rec))trigger_hit (&switch_output_trigger);
tools/perf/builtin-stat.c +1 -0
@@ -578,6 +578,7 @@ static void process_evlist(struct evlist *evlist, unsigned int interval)process_interval ();pr_info (EVLIST_DISABLED_MSG );break ; +case EVLIST_CTL_CMD_SNAPSHOT :case EVLIST_CTL_CMD_ACK :case EVLIST_CTL_CMD_UNSUPPORTED :default :
tools/perf/util/evlist.c +9 -2
@@ -1881,13 +1881,17 @@ static int evlist__ctlfd_recv(struct evlist *evlist, enum evlist_ctl_cmd *cmd, }else if (!strncmp (cmd_data,EVLIST_CTL_CMD_DISABLE_TAG , (sizeof (EVLIST_CTL_CMD_DISABLE_TAG )-1 ))) { *cmd =EVLIST_CTL_CMD_DISABLE ; + }else if (!strncmp (cmd_data,EVLIST_CTL_CMD_SNAPSHOT_TAG , + (sizeof (EVLIST_CTL_CMD_SNAPSHOT_TAG )-1 ))) { + *cmd =EVLIST_CTL_CMD_SNAPSHOT ; +pr_debug ("is snapshot\n" ); } }return bytes_read ? (int )bytes_read : err; } -static int evlist__ctlfd_ack (struct evlist *evlist) +int evlist__ctlfd_ack (struct evlist *evlist) {int err;
@@ -1923,13 +1927,16 @@ int evlist__ctlfd_process(struct evlist *evlist, enum evlist_ctl_cmd *cmd)case EVLIST_CTL_CMD_DISABLE :evlist__disable (evlist);break ; +case EVLIST_CTL_CMD_SNAPSHOT : +break ;case EVLIST_CTL_CMD_ACK :case EVLIST_CTL_CMD_UNSUPPORTED :default :pr_debug ("ctlfd: unsupported %d\n" , *cmd);break ; } -if (!(*cmd ==EVLIST_CTL_CMD_ACK || *cmd ==EVLIST_CTL_CMD_UNSUPPORTED )) +if (!(*cmd ==EVLIST_CTL_CMD_ACK || *cmd ==EVLIST_CTL_CMD_UNSUPPORTED || + *cmd ==EVLIST_CTL_CMD_SNAPSHOT ))evlist__ctlfd_ack (evlist); } }
tools/perf/util/evlist.h +4 -1
@@ -363,6 +363,7 @@ struct evsel *perf_evlist__reset_weak_group(struct evlist *evlist,#define EVLIST_CTL_CMD_ENABLE_TAG "enable"#define EVLIST_CTL_CMD_DISABLE_TAG "disable"#define EVLIST_CTL_CMD_ACK_TAG "ack\n" +#define EVLIST_CTL_CMD_SNAPSHOT_TAG "snapshot"#define EVLIST_CTL_CMD_MAX_LEN 64
@@ -370,7 +371,8 @@ enum evlist_ctl_cmd {EVLIST_CTL_CMD_UNSUPPORTED =0 ,EVLIST_CTL_CMD_ENABLE ,EVLIST_CTL_CMD_DISABLE , - EVLIST_CTL_CMD_ACK +EVLIST_CTL_CMD_ACK , + EVLIST_CTL_CMD_SNAPSHOT, };int evlist__parse_control (const char *str,int *ctl_fd,int *ctl_fd_ack,bool *ctl_fd_close);
@@ -378,6 +380,7 @@ int evlist__initialize_ctlfd(struct evlist *evlist, int ctl_fd, int ctl_fd_ack);int evlist__finalize_ctlfd (struct evlist *evlist);bool evlist__ctlfd_initialized (struct evlist *evlist);int evlist__ctlfd_process (struct evlist *evlist,enum evlist_ctl_cmd *cmd); +int evlist__ctlfd_ack (struct evlist *evlist);#define EVLIST_ENABLED_MSG "Events enabled\n"#define EVLIST_DISABLED_MSG "Events disabled\n"
Signature · 1 line
2.17.1
Re: [PATCH 5/6] perf record: Add 'snapshot' control command
Quoted 3 lines · Adrian Hunter
Add 'snapshot' control command to create an AUX area tracing snapshot the same as if sending SIGUSR2. The advantage of the FIFO is that access is governed by access to the FIFO.
How about the --switch-output SIGUSR2 switch? Could that be handled too?
Other than that it looks great. Thanks.
-Andi
Adrian Hunter
reviewerRe: [PATCH 5/6] perf record: Add 'snapshot' control command
Quoted 6 lines · Andi Kleen
Quoted 3 lines · Adrian Hunter
Add 'snapshot' control command to create an AUX area tracing snapshot the same as if sending SIGUSR2. The advantage of the FIFO is that access is governed by access to the FIFO.
How about the --switch-output SIGUSR2 switch? Could that be handled too?
Presumably, but I have never used --switch-out because Intel PT does not yet support it (it is on the list of things to do). Perhaps someone more familiar with it can add that :-)
Quoted 5 lines
Other than that it looks great. Thanks.
-Andi
Adrian Hunter
reviewerPATCH 6/6 perf intel-pt: Document snapshot control command
In reply to Adrian HunterThe documentation descibes snapshot mode. Update it to include the new snapshot control command.
--- tools/perf/Documentation/perf-intel-pt.txt | 23 +++++++++++++++++++--- 1 file changed, 20 insertions(+), 3 deletions(-)
tools/perf/Documentation/perf-intel-pt.txt +20 -3
@@ -558,7 +558,7 @@ The mmap size and auxtrace mmap size are displayed if the -vv option is used e.g Intel PT modes of operation ~~~~~~~~~~~~~~~~~~~~~~~~~~~ -Intel PT can be used in 2 modes: +Intel PT can be used in 3 modes: full-trace mode sample mode snapshot mode
@@ -571,7 +571,8 @@ Sample mode attaches a Intel PT sample to other events e.g. perf record --aux-sample -e intel_pt//u -e branch-misses:u -Snapshot mode captures the available data when a signal is sent e.g. +Snapshot mode captures the available data when a signal is sent or "snapshot" +control command is issued. e.g. using a signal perf record -v -e intel_pt//u -S ./loopy 1000000000 & [1] 11435
@@ -582,7 +583,23 @@ Note that the signal sent is SIGUSR2. Note that "Recording AUX area tracing snapshot" is displayed because the -v option is used. -The 2 modes cannot be used together. +The advantage of using "snapshot" control command is that the access is +controlled by access to a FIFO e.g. + + $ mkfifo perf.control + $ mkfifo perf.ack + $ cat perf.ack & + [1] 15235 + $ sudo ~/bin/perf record --control perf.control,perf.ack -S -e intel_pt//u -- sleep 60 & + [2] 15243 + $ ps -e | grep perf + 15244 pts/1 00:00:00 perf + $ kill -USR2 15244 + bash: kill: (15244) - Operation not permitted + $ echo snapshot > perf.control + ack + +The 3 Intel PT modes of operation cannot be used together. Buffer handling
Signature · 1 line
2.17.1