Attributes (struct perf_event_attr) are recorded separately in the
perf.data file. perf script uses them to set up output options.
However attributes can also be in the event stream, for example when the
input is a pipe (i.e. live mode). This patch makes perf script process
in-stream attributes in the same way as on-file attributes.
Here is an example:
Before this patch:
$ perf record uname | perf script
Linux
[ perf record: Woken up 1 times to write data ]
[ perf record: Captured and wrote 0.015 MB (null) (~655 samples) ]
:4220 4220 [-01] 2933367.838906: cycles:
:4220 4220 [-01] 2933367.838910: cycles:
:4220 4220 [-01] 2933367.838912: cycles:
:4220 4220 [-01] 2933367.838914: cycles:
:4220 4220 [-01] 2933367.838916: cycles:
:4220 4220 [-01] 2933367.838918: cycles:
uname 4220 [-01] 2933367.838938: cycles:
uname 4220 [-01] 2933367.839207: cycles:
After this patch:
$ perf record uname | perf script
Linux
[ perf record: Woken up 1 times to write data ]
[ perf record: Captured and wrote 0.015 MB (null) (~655 samples) ]
:4582 4582 2933425.707724: cycles: ffffffff81043ffa native_write_msr_safe ([kernel.kallsyms])
:4582 4582 2933425.707728: cycles: ffffffff81043ffa native_write_msr_safe ([kernel.kallsyms])
:4582 4582 2933425.707730: cycles: ffffffff81043ffa native_write_msr_safe ([kernel.kallsyms])
:4582 4582 2933425.707732: cycles: ffffffff81043ffa native_write_msr_safe ([kernel.kallsyms])
:4582 4582 2933425.707734: cycles: ffffffff81043ffa native_write_msr_safe ([kernel.kallsyms])
:4582 4582 2933425.707736: cycles: ffffffff81309a24 memcpy ([kernel.kallsyms])
uname 4582 2933425.707760: cycles: ffffffff8109c1c7 enqueue_task_fair ([kernel.kallsyms])
uname 4582 2933425.707978: cycles: ffffffff81308457 clear_page_c ([kernel.kallsyms])
Signed-off-by: Adrian Hunter <adrian.hunter@intel.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Namhyung Kim <namhyung@gmail.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/r/1383313899-15987-3-git-send-email-adrian.hunter@intel.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
This new COMM infrastructure provides two features:
1) It keeps track of all comms lifecycle for a given thread. This way we
can associate a timeframe to any thread COMM, as long as
PERF_SAMPLE_TIME samples are joined to COMM and fork events.
As a result we should have more precise COMM sorted hists with seperated
entries for pre and post exec time after a fork.
2) It also makes sure that a given COMM string is not duplicated but
rather shared among the threads that refer to it. This way the threads
COMM can be compared against pointer values from the sort
infrastructure.
Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com>
Tested-by: Jiri Olsa <jolsa@redhat.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-hwjf70b2wve9m2kosxiq8bb3@git.kernel.org
[ Rename some accessor functions ]
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
[ Use __ as separator for class__method for private comm_str methods ]
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
This way we can later delimit a lifecycle for the COMM and map a hist to
a precise COMM:timeslice couple.
PERF_RECORD_COMM and PERF_RECORD_FORK events that don't have
PERF_SAMPLE_TIME samples can only send 0 value as a timestamp and thus
should overwrite any previous COMM on a given thread because there is no
sensible way to keep track of all the comms lifecycles in a thread
without time informations.
Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com>
Tested-by: Jiri Olsa <jolsa@redhat.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-6tyow99vgmmtt9qwr2u2lqd7@git.kernel.org
[ Made it cope with PERF_RECORD_MMAP2 ]
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
The "struct perf_event_attr setup" entry in 'perf test' is in fact a
series of tests that will exec the tools, passing different sets of
command line arguments to then intercept the sys_perf_event_open
syscall, in user space, to check that the perf_event_attr->sample_type
and other feature request bits are setup as expected.
We recently restored the callchain requesting command line argument, -g,
to not require a parameter ("dwarf" or "fp"), instead using a default
("fp" for now) and making the long option variant, --call-chain, be the
one to be used when a different callchain collection method is
preferred.
The "struct perf_event_attr setup" test failed because we forgot to
update the tests involving callchains, not switching from, '-g dwarf' to
'--call-chain dwarf', making 'perf test' detect it:
[root@sandy ~]# perf test -v 13
13: struct perf_event_attr setup :
--- start ---
running '/home/acme/libexec/perf-core/tests/attr/test-record-basic'
running '/home/acme/libexec/perf-core/tests/attr/test-record-branch-any'
<SNIP>
running '/home/acme/libexec/perf-core/tests/attr/test-record-graph-default'
running '/home/acme/libexec/perf-core/tests/attr/test-record-graph-dwarf'
expected sample_type=12583, got 295
expected exclude_callchain_user=1, got 0
expected sample_stack_user=8192, got 0
FAILED '/home/acme/libexec/perf-core/tests/attr/test-record-graph-dwarf' - match failure
---- end ----
struct perf_event_attr setup: FAILED!
[root@sandy ~]#
Fix all of them now to use --call-chain when explicitely specifying a
method.
There is still work to do, as '-g fp', for instance, passed without
problems.
In that case 'perf test' saw no problems as the intercepted syscall got
the bits as expected, i.e. the default is 'fp', but the fact that 'fp'
may be an existing program and the specified workload would then be
passed as a parameter to it is an usability problem that needs fixing.
Next merge window tho.
Acked-by: Jiri Olsa <jolsa@redhat.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-jr3oq1k5iywnp7vvqlslzydm@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
There are two warnings in bench/numa, when building this on 32-bit
machine.
The warning output is attached:
bench/numa.c:1113:20: error: comparison between signed and unsigned integer expressions [-Werror=sign-compare]
bench/numa.c:1161:6: error: format ‘%lx’ expects argument of t'long unsigned int’, but argument 5 has type ‘u64’ [-Werror=format]
This patch fixes these two warnings.
Signed-off-by: Wei Yang <weiyang@linux.vnet.ibm.com>
Acked-by: Ingo Molnar <mingo@kernel.org>
Cc: Ingo Molnar <mingo@redhat.com>
Link: http://lkml.kernel.org/r/1379839764-9245-1-git-send-email-weiyang@linux.vnet.ibm.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
We are using the Python scripting interface in perf to extract kernel
events relevant for performance analysis of HPC codes. We noticed that
the "perf script" call allocates a significant amount of memory (in the
order of several 100 MiB) during it's run, e.g. 125 MiB for a 25 MiB
input file:
$> perf record -o perf.data -a -R -g fp \
-e power:cpu_frequency -e sched:sched_switch \
-e sched:sched_migrate_task -e sched:sched_process_exit \
-e sched:sched_process_fork -e sched:sched_process_exec \
-e cycles -m 4096 --freq 4000
$> /usr/bin/time perf script -i perf.data -s dummy_script.py
0.84user 0.13system 0:01.92elapsed 51%CPU (0avgtext+0avgdata
125532maxresident)k
73072inputs+0outputs (57major+33086minor)pagefaults 0swaps
Upon further investigation using the valgrind massif tool, we noticed
that Python objects that are created in trace-event-python.c via
PyString_FromString*() (and their Integer and Long counterparts) are
never free'd.
The reason for this seem to be missing Py_DECREF calls on the objects
that are returned by these functions and stored in the Python
dictionaries. The Python dictionaries do not steal references (as
opposed to Python tuples and lists) but instead add their own reference.
Hence, the reference that is returned by these object creation functions
is never released and the memory is leaked. (see [1,2])
The attached patch fixes this by wrapping all relevant calls to
PyDict_SetItemString() and decrementing the reference counter
immediately after the Python function call.
This reduces the allocated memory to a reasonable amount:
$> /usr/bin/time perf script -i perf.data -s dummy_script.py
0.73user 0.05system 0:00.79elapsed 99%CPU (0avgtext+0avgdata
49132maxresident)k
0inputs+0outputs (0major+14045minor)pagefaults 0swaps
For comparison, with a 120 MiB input file the memory consumption
reported by time drops from almost 600 MiB to 146 MiB.
The patch has been tested using Linux 3.8.2 with Python 2.7.4 and Linux
3.11.6 with Python 2.7.5.
Please let me know if you need any further information.
[1] http://docs.python.org/2/c-api/tuple.html#PyTuple_SetItem
[2] http://docs.python.org/2/c-api/dict.html#PyDict_SetItemString
Signed-off-by: Joseph Schuchart <joseph.schuchart@tu-dresden.de>
Reviewed-by: Tom Zanussi <tom.zanussi@linux.intel.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Tom Zanussi <tom.zanussi@linux.intel.com>
Link: http://lkml.kernel.org/r/1381468543-25334-4-git-send-email-namhyung@kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Before this patch, looking at 'perf bench sched pipe' behavior over
'top' only told us that something related to perf is running:
PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
19934 mingo 20 0 54836 1296 952 R 18.6 0.0 0:00.56 perf
19935 mingo 20 0 54836 384 36 S 18.6 0.0 0:00.56 perf
After the patch it's clearly visible what's going on:
PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
19744 mingo 20 0 125m 3536 2644 R 68.2 0.0 0:01.12 sched-pipe
19745 mingo 20 0 125m 1172 276 R 68.2 0.0 0:01.12 sched-pipe
The benchmark-subsystem name is concatenated with the individual
testcase name.
Unfortunately 'perf top' does not show the reconfigured name, possibly
because it caches ->comm[] values and does not recognize changes to
them?
Also clean up a few bits in builtin-bench.c while at it and reorganize
the code and the output strings to be consistent.
Use iterators to access the various arrays. Rename 'suites' concept to
'benchmark collection' and the 'bench_suite' to 'benchmark/bench'. The
many repetitions of 'suite' made the code harder to read and understand.
The new output is:
comet:~/tip/tools/perf> ./perf bench
Usage:
perf bench [<common options>] <collection> <benchmark> [<options>]
# List of all available benchmark collections:
sched: Scheduler and IPC benchmarks
mem: Memory access benchmarks
numa: NUMA scheduling and MM benchmarks
all: All benchmarks
comet:~/tip/tools/perf> ./perf bench sched
# List of available benchmarks for collection 'sched':
messaging: Benchmark for scheduling and IPC
pipe: Benchmark for pipe() between two processes
all: Test all scheduler benchmarks
comet:~/tip/tools/perf> ./perf bench mem
# List of available benchmarks for collection 'mem':
memcpy: Benchmark for memcpy()
memset: Benchmark for memset() tests
all: Test all memory benchmarks
comet:~/tip/tools/perf> ./perf bench numa
# List of available benchmarks for collection 'numa':
mem: Benchmark for NUMA workloads
all: Test all NUMA benchmarks
Individual benchmark modules were not touched.
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Cc: David Ahern <dsahern@gmail.com>
Cc: Hitoshi Mitake <h.mitake@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Namhyung Kim <namhyung.kim@lge.com>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Link: http://lkml.kernel.org/r/20131023123756.GA17871@gmail.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
At this point, --fentry (mcount function entry) option for gcc fuzzes
the debuginfo variable locations by skipping the mcount instruction
offset (on x86, this is a 5 byte call instruction).
This makes variable searching fail at the entry of functions which
are mcount'ed.
e.g.)
Available variables at vfs_read
@<vfs_read+0>
(No matched variables)
This patch adds additional location search at the function entry point
to solve this issue, which tries to find the earliest address for the
variable location.
Note that this only works with function parameters (formal parameters)
because any local variables should not exist on the function entry
address (those are not initialized yet).
With this patch, perf probe shows correct parameters if possible;
# perf probe --vars vfs_read
Available variables at vfs_read
@<vfs_read+0>
char* buf
loff_t* pos
size_t count
struct file* file
Signed-off-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Link: http://lkml.kernel.org/r/20131011071025.15557.13275.stgit@udc4-manage.rcp.hitachi.co.jp
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
