To get the latest fixes in the perf tools including perf stat output,
dlfilter and LLVM feature detection.
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
The changes in ("perf evsel: Rename evsel__increase_rlimit to
rlimit__increase_nofile") ended up breaking the python binding that now
references the rlimit__increase_nofile function, add the util/rlimit.o
to the tools/perf/util/python-ext-sources to cure that.
This was detected by the 'perf test python' regression test:
$ perf test python
14: 'import perf' in python : FAILED!
$ perf test -v python
Couldn't bump rlimit(MEMLOCK), failures may take place when creating BPF maps, etc
14: 'import perf' in python :
--- start ---
test child forked, pid 2912462
python usage test: "echo "import sys ; sys.path.insert(0, '/tmp/build/perf-tools-next/python'); import perf" | '/usr/bin/python3' "
Traceback (most recent call last):
File "<stdin>", line 1, in <module>
ImportError: /tmp/build/perf-tools-next/python/perf.cpython-311-x86_64-linux-gnu.so: undefined symbol: rlimit__increase_nofile
test child finished with -1
---- end ----
'import perf' in python: FAILED!
$
Fixes: e093a222d7 ("perf evsel: Rename evsel__increase_rlimit to rlimit__increase_nofile")
Acked-by: Namhyung Kim <namhyung@kernel.org>
Acked-by: Yang Jihong <yangjihong1@huawei.com>
Link: https://lore.kernel.org/lkml/ZTrCS5Z3PZAmfPdV@kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
Currently lock contention timestamp is maintained in a hash map keyed by
pid. That means it needs to get and release a map element (which is
proctected by spinlock!) on each contention begin and end pair. This
can impact on performance if there are a lot of contention (usually from
spinlocks).
It used to go with task local storage but it had an issue on memory
allocation in some critical paths. Although it's addressed in recent
kernels IIUC, the tool should support old kernels too. So it cannot
simply switch to the task local storage at least for now.
As spinlocks create lots of contention and they disabled preemption
during the spinning, it can use per-cpu array to keep the timestamp to
avoid overhead in hashmap update and delete.
In contention_begin, it's easy to check the lock types since it can see
the flags. But contention_end cannot see it. So let's try to per-cpu
array first (unconditionally) if it has an active element (lock != 0).
Then it should be used and per-task tstamp map should not be used until
the per-cpu array element is cleared which means nested spinlock
contention (if any) was finished and it nows see (the outer) lock.
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
Acked-by: Ian Rogers <irogers@google.com>
Cc: Hao Luo <haoluo@google.com>
Cc: Song Liu <song@kernel.org>
Cc: bpf@vger.kernel.org
Link: https://lore.kernel.org/r/20231020204741.1869520-3-namhyung@kernel.org
It checks the current lock to calculated the delta of contention time.
The address is saved in the tstamp map which is allocated at begining of
contention and released at end of contention.
But it's possible for bpf_map_delete_elem() to fail. In that case, the
element in the tstamp map kept for the current lock and it makes the
next contention for the same lock tracked incorrectly. Specificially
the next contention begin will see the existing element for the task and
it'd just return. Then the next contention end will see the element and
calculate the time using the timestamp for the previous begin.
This can result in a large value for two small contentions happened from
time to time. Let's clear the lock address so that it can be updated
next time even if the bpf_map_delete_elem() failed.
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
Acked-by: Ian Rogers <irogers@google.com>
Cc: Hao Luo <haoluo@google.com>
Cc: Song Liu <song@kernel.org>
Cc: bpf@vger.kernel.org
Link: https://lore.kernel.org/r/20231020204741.1869520-1-namhyung@kernel.org
evsel__increase_rlimit() helper does nothing with evsel, and description
of the functionality is inaccurate, rename it and move to util/rlimit.c.
By the way, fix a checkppatch warning about misplaced license tag:
WARNING: Misplaced SPDX-License-Identifier tag - use line 1 instead
#160: FILE: tools/perf/util/rlimit.h:3:
/* SPDX-License-Identifier: LGPL-2.1 */
No functional change.
Signed-off-by: Yang Jihong <yangjihong1@huawei.com>
Link: https://lore.kernel.org/r/20231023033144.1011896-1-yangjihong1@huawei.com
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
If using parallel threads to collect data, perf record needs at least 6 fds
per CPU. (one for sys_perf_event_open, four for pipe msg and ack of the
pipe, see record__thread_data_open_pipes(), and one for open perf.data.XXX)
For an environment with more than 100 cores, if perf record uses both
`-a` and `--threads` options, it is easy to exceed the upper limit of the
file descriptor number, when we run out of them try to increase the limits.
Before:
$ ulimit -n
1024
$ lscpu | grep 'On-line CPU(s)'
On-line CPU(s) list: 0-159
$ perf record --threads -a sleep 1
Failed to create data directory: Too many open files
After:
$ ulimit -n
1024
$ lscpu | grep 'On-line CPU(s)'
On-line CPU(s) list: 0-159
$ perf record --threads -a sleep 1
[ perf record: Woken up 1 times to write data ]
[ perf record: Captured and wrote 0.394 MB perf.data (1576 samples) ]
Signed-off-by: Yang Jihong <yangjihong1@huawei.com>
Acked-by: Namhyung Kim <namhyung@kernel.org>
Link: https://lore.kernel.org/r/20231013075945.698874-1-yangjihong1@huawei.com
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
The default config is computed during creation of the PMU and may do
things like scanning sysfs, when the PMU may just be used as part of
scanning. Change default_config to perf_event_attr_init_default, a
callback that is used when a default config needs initializing. This
avoids holding onto the memory for a perf_event_attr and copying.
On a tigerlake laptop running the pmu-scan benchmark:
Before:
Running 'internals/pmu-scan' benchmark:
Computing performance of sysfs PMU event scan for 100 times
Average core PMU scanning took: 28.780 usec (+- 0.503 usec)
Average PMU scanning took: 283.480 usec (+- 18.471 usec)
Number of openat syscalls: 30,227
After:
Running 'internals/pmu-scan' benchmark:
Computing performance of sysfs PMU event scan for 100 times
Average core PMU scanning took: 27.880 usec (+- 0.169 usec)
Average PMU scanning took: 245.260 usec (+- 15.758 usec)
Number of openat syscalls: 28,914
Over 3 runs it is a nearly 12% reduction in execution time and a 4.3%
of openat calls.
Signed-off-by: Ian Rogers <irogers@google.com>
Reviewed-by: Adrian Hunter <adrian.hunter@intel.com>
Cc: Ravi Bangoria <ravi.bangoria@amd.com>
Cc: James Clark <james.clark@arm.com>
Cc: Suzuki K Poulose <suzuki.poulose@arm.com>
Cc: Yang Jihong <yangjihong1@huawei.com>
Cc: Will Deacon <will@kernel.org>
Cc: Leo Yan <leo.yan@linaro.org>
Cc: Mike Leach <mike.leach@linaro.org>
Cc: Jing Zhang <renyu.zj@linux.alibaba.com>
Cc: Kajol Jain <kjain@linux.ibm.com>
Cc: Thomas Richter <tmricht@linux.ibm.com>
Cc: Kan Liang <kan.liang@linux.intel.com>
Cc: John Garry <john.g.garry@oracle.com>
Cc: linux-arm-kernel@lists.infradead.org
Cc: coresight@lists.linaro.org
Link: https://lore.kernel.org/r/20231012175645.1849503-8-irogers@google.com
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
Raw events can be strings like 'r0xead' but the 0x is optional so they
can also be 'read'. On IcelakeX uncore_imc_free_running has an event
called 'read' which may be programmed as:
```
$ perf stat -e 'uncore_imc_free_running/event=read/' -a sleep 1
```
However, the PE_RAW type isn't allowed on the right of a term, even
though in this case we just want to interpret it as a string. This
leads to the following error on IcelakeX:
```
$ perf stat -e 'uncore_imc_free_running/event=read/' -a sleep 1
event syntax error: '..nning/event=read/'
\___ parser error
Run 'perf list' for a list of valid events
Usage: perf stat [<options>] [<command>]
-e, --event <event> event selector. use 'perf list' to list available events
```
Fix this by allowing raw types on the right of terms and treat them as
strings, just as is already done for PE_LEGACY_CACHE. Make this
consistent by just entirely removing name_or_legacy and always using
name_or_raw that covers all three cases.
Fixes: 6fd1e51915 ("perf parse-events: Support PMUs for legacy cache events")
Signed-off-by: Ian Rogers <irogers@google.com>
Cc: James Clark <james.clark@arm.com>
Cc: Kan Liang <kan.liang@linux.intel.com>
Link: https://lore.kernel.org/r/20230928004431.1926969-1-irogers@google.com
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
This is a longstanding to do list entry: we need a way to see that a
sample took place while in idle state, as the current way to do it is
to infer that by the name of the functions that in such state have
more samples, IOW: a hack.
Maybe we can do flip a bit in samples that take place inside the
enter/exit idle section in do_idle()?
But till then, add one more :-\
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Acked-by: Namhyung Kim <namhyung@kernel.org>
Cc: Frederic Weisbecker <frederic@kernel.org>
Link: https://lore.kernel.org/r/ZR66Qgbcltt+zG7F@kernel.org
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
We specify that a "num_hex" comprises 1 or more digits, however, that
allows strtoull to fail with ERANGE. Limit the number of hex digits to
being between 1 and 16.
Before:
```
$ perf stat -e 'cpu/rE7574c47490475745/' true
perf: util/parse-events.c:215: fix_raw: Assertion `errno == 0' failed.
Aborted (core dumped)
```
After:
```
$ perf stat -e 'cpu/rE7574c47490475745/' true
event syntax error: 'cpu/rE7574c47490475745/'
\___ Bad event or PMU
Unable to find PMU or event on a PMU of 'cpu'
Initial error:
event syntax error: 'cpu/rE7574c47490475745/'
\___ unknown term 'rE7574c47490475745' for pmu 'cpu'
valid terms: event,pc,edge,offcore_rsp,ldlat,inv,umask,frontend,cmask,config,config1,config2,config3,name,period,percore,metric-id
Run 'perf list' for a list of valid events
Usage: perf stat [<options>] [<command>]
-e, --event <event> event selector. use 'perf list' to list available events
```
Issue found through fuzz testing.
Signed-off-by: Ian Rogers <irogers@google.com>
Acked-by: Namhyung Kim <namhyung@kernel.org>
Link: https://lore.kernel.org/r/20230907210533.3712979-1-irogers@google.com
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
To pick up the 'perf bench sched-seccomp-notify' changes to allow us to
continue build testing perf-tools-next with the set of distro
containers, where some older ones don't have a recent enough seccomp.h
UAPI header that contains defines needed by this new 'perf bench'
workload.
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Update "struct dso" to include new member "is_kmod".
This new field will determine if the file is a kernel
module or not.
To resolve the address from a sample, perf looks at the
DSO maps. In case of address from a kernel module, there
were some address found to be not resolved. This was
observed while running perf test for "Object code reading".
Though the ip falls beteen the start address of the loaded
module (perf map->start ) and end address ( perf map->end),
it was unresolved.
This was happening because in some cases for kernel
modules, address from sample points to stub instructions.
To identify if the DSO is a kernel module, the new field
"is_kmod" is added to "struct dso".
Reported-by: Disha Goel <disgoel@linux.ibm.com>
Signed-off-by: Athira Rajeev <atrajeev@linux.vnet.ibm.com>
Acked-by: Namhyung Kim <namhyung@kernel.org>
Cc: kjain@linux.ibm.com
Cc: maddy@linux.ibm.com
Cc: disgoel@linux.vnet.ibm.com
Cc: linuxppc-dev@lists.ozlabs.org
Link: https://lore.kernel.org/r/20230928075213.84392-2-atrajeev@linux.vnet.ibm.com
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
Update "struct dso" to include new member "text_end".
This new field will represent the offset for end of text
section for a dso. For elf, this value is derived as:
sh_size (Size of section in byes) + sh_offset (Section file
offst) of the elf header for text.
For bfd, this value is derived as:
1. For PE file,
section->size + ( section->vma - dso->text_offset)
2. Other cases:
section->filepos (file position) + section->size (size of
section)
To resolve the address from a sample, perf looks at the
DSO maps. In case of address from a kernel module, there
were some address found to be not resolved. This was
observed while running perf test for "Object code reading".
Though the ip falls beteen the start address of the loaded
module (perf map->start ) and end address ( perf map->end),
it was unresolved.
Example:
Reading object code for memory address: 0xc008000007f0142c
File is: /lib/modules/6.5.0-rc3+/kernel/fs/xfs/xfs.ko
On file address is: 0x1114cc
Objdump command is: objdump -z -d --start-address=0x11142c --stop-address=0x1114ac /lib/modules/6.5.0-rc3+/kernel/fs/xfs/xfs.ko
objdump read too few bytes: 128
test child finished with -1
Here, module is loaded at:
# cat /proc/modules | grep xfs
xfs 2228224 3 - Live 0xc008000007d00000
From objdump for xfs module, text section is:
text 0010f7bc 0000000000000000 0000000000000000 000000a0 2**4
Here the offset for 0xc008000007f0142c ie 0x112074 falls out
.text section which is up to 0x10f7bc.
In this case for module, the address 0xc008000007e11fd4 is pointing
to stub instructions. This address range represents the module stubs
which is allocated on module load and hence is not part of DSO offset.
To identify such address, which falls out of text
section and within module end, added the new field "text_end" to
"struct dso".
Reported-by: Disha Goel <disgoel@linux.ibm.com>
Signed-off-by: Athira Rajeev <atrajeev@linux.vnet.ibm.com>
Reviewed-by: Adrian Hunter <adrian.hunter@intel.com>
Reviewed-by: Kajol Jain <kjain@linux.ibm.com>
Acked-by: Namhyung Kim <namhyung@kernel.org>
Cc: maddy@linux.ibm.com
Cc: disgoel@linux.vnet.ibm.com
Cc: linuxppc-dev@lists.ozlabs.org
Link: https://lore.kernel.org/r/20230928075213.84392-1-atrajeev@linux.vnet.ibm.com
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
PMU alias names were computed when the first perf_pmu is created,
scanning all PMUs in event sources for a file called alias that
generally doesn't exist. Switch to trying to load the file when all
PMU related files are loaded in lookup. This would cause a PMU name
lookup of an alias name to fail if no PMUs were loaded, so in that
case all PMUs are loaded and the find repeated. The overhead is
similar but in the (very) general case not all PMUs are scanned for
the alias file.
As the overhead occurs once per invocation it doesn't show in perf
bench internals pmu-scan. On a tigerlake machine, the number of openat
system calls for an event of cpu/cycles/ with perf stat reduces from
94 to 69 (ie 25 fewer openat calls).
Signed-off-by: Ian Rogers <irogers@google.com>
Acked-by: Namhyung Kim <namhyung@kernel.org>
Cc: Ravi Bangoria <ravi.bangoria@amd.com>
Cc: James Clark <james.clark@arm.com>
Cc: Leo Yan <leo.yan@linaro.org>
Cc: Kan Liang <kan.liang@linux.intel.com>
Link: https://lore.kernel.org/r/20230925062323.840799-1-irogers@google.com
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
