If git is installed we'll have a 'perf --version' output of this form:
$ make -j8 -C tools/perf/ O=/home/acme/git/build/perf install
$ perf --version
perf version 3.7.rc3.g3afad6
Now on a machine without git installed:
$ mv /home/acme/bin/git /home/acme/bin/git.OFF
$ make -j8 -C tools/perf/ O=/home/acme/git/build/perf install
$ perf --version
perf version 3.7.0-rc2
That is, no error message due to git not being installed will appear on the
screen and instead the version string in the top level Makefile will be
used.
Requested-by: Ingo Molnar <mingo@kernel.org>
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: Namhyung Kim <namhyung@gmail.com>
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-am6yp6phvxyjmyndxogpunjv@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
There's another source of overhead in the perf version string generator:
git update-index -q --refresh
... which will iterate the whole checked out tree. This can be pretty
slow on NFS volumes, but takes some time even with local SSD disks and a
fully cached kernel tree:
$ perf stat --null --repeat 3 --pre "rm -f PERF-VERSION-FILE" util/PERF-VERSION-GEN
PERF_VERSION = 3.7.rc3.g5399b3b.dirty
PERF_VERSION = 3.7.rc3.g5399b3b.dirty
PERF_VERSION = 3.7.rc3.g5399b3b.dirty
Performance counter stats for 'util/PERF-VERSION-GEN' (3 runs):
0.306999221 seconds time elapsed ( +- 0.56% )
So remove the .dirty differentiator as well - it adds little information
because locally patched git trees are common, but seldom are the perf
tools modified.
So a lot of version strings are reported as 'dirty' while in fact they
are pristine perf builds. For example 99% of my perf builds are not
patched but the kernel tree is slightly patched, which adds the .dirty
tag.
Eliminating that tag speeds up version generation by another order of
magnitude:
$ perf stat --null --repeat 3 --sync --pre "rm -f PERF-VERSION-FILE" util/PERF-VERSION-GEN
PERF_VERSION = 3.7.rc3.g4b0bd3
PERF_VERSION = 3.7.rc3.g4b0bd3
PERF_VERSION = 3.7.rc3.g4b0bd3
Performance counter stats for 'util/PERF-VERSION-GEN' (3 runs):
0.021270923 seconds time elapsed ( +- 1.94% )
(Also clean up some of the comments around this code.)
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Cc: Andrew Vagin <avagin@openvz.org>
Cc: Borislav Petkov <bp@amd64.org>
Cc: David Howells <dhowells@redhat.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Stephane Eranian <eranian@gmail.com>
Cc: Steven Rostedt <rostedt@goodmis.org>
Link: http://lkml.kernel.org/r/20121030085441.GC8245@gmail.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Building perf is pretty slow on trees that have a lot of commits
relative to the nearest Git tag. This slowness manifests itself during
version string generation:
$ perf stat --null --repeat 3 --sync --pre "rm -f PERF-VERSION-FILE" util/PERF-VERSION-GEN
PERF_VERSION = 3.7.rc3.1458.g5399b3b
PERF_VERSION = 3.7.rc3.1458.g5399b3b
PERF_VERSION = 3.7.rc3.1458.g5399b3b
Performance counter stats for 'util/PERF-VERSION-GEN' (3 runs):
2.857503976 seconds time elapsed ( +- 0.22% )
The build can be even slower than that, when one over NFS volumes.
The reason for the slowness is that util/PERF-VERSION-GEN uses "git
describe" to generate the string, which has to count the "number of
commits distance" from the nearest tag - the ".1458." count in the
output above. For that Git had to extract and decompress 1458 Git
objects, which takes time and bandwidth.
But this "number of commits" value is mostly irrelevant in practice. We
either want to know an approximate tag name, or we want to know the
precise sha1.
So this patch simplifies the version string to:
PERF_VERSION = 3.7.rc3.g5399b3b.dirty
which speeds up the version string generation script by an order of
magnitude:
$ perf stat --null --repeat 3 --sync --pre "rm -f PERF-VERSION-FILE" util/PERF-VERSION-GEN
PERF_VERSION = 3.7.rc3.g5399b3b.dirty
PERF_VERSION = 3.7.rc3.g5399b3b.dirty
PERF_VERSION = 3.7.rc3.g5399b3b.dirty
Performance counter stats for 'util/PERF-VERSION-GEN' (3 runs):
0.307633559 seconds time elapsed ( +- 0.84% )
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Cc: Andrew Vagin <avagin@openvz.org>
Cc: Borislav Petkov <bp@amd64.org>
Cc: David Howells <dhowells@redhat.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Stephane Eranian <eranian@gmail.com>
Cc: Steven Rostedt <rostedt@goodmis.org>
Link: http://lkml.kernel.org/r/20121030084600.GB8245@gmail.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
This was found during chasing down the header output regression. The
strbuf_addf() was checking buffer length with a result of vscnprintf()
which cannot be greater than that of strbuf_avail().
Since numa topology and pmu mapping info in header were converted to use
strbuf, it sometimes caused uninteresting behaviors with the broken
strbuf.
Fix it by using vsnprintf() which returns desired output string length
regardless of the available buffer size and grow the buffer if needed.
Reported-by: Andrew Jones <drjones@redhat.com>
Tested-by: Andrew Jones <drjones@redhat.com>
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
Cc: Andrew Jones <drjones@redhat.com>
Link: http://lkml.kernel.org/r/1350999890-6920-2-git-send-email-namhyung@kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Create a script browser, so that user can check all the available
scripts for current perf data file and run them inside the main perf
report or annotation browsers, for all perf samples or for samples
belong to one thread/symbol.
Please be noted: current script browser is only for report use, and
doesn't cover the record phase, IOW it must run against one existing
perf data file.
The work flow is, users can use function key to list all the available
scripts for current perf data file in system and chose one, which will
be executed with popen("perf script -s xxx.xx",) and all the output
lines are put into one ui browser, pressing 'q' or left arrow key will
make it return to previous browser.
Signed-off-by: Feng Tang <feng.tang@intel.com>
Cc: Andi Kleen <andi@firstfloor.org>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/r/1351569369-26732-4-git-send-email-feng.tang@intel.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
There's a portion in the "perf list" output refering to the exact
specification of raw hardware events.
Since this description is in the perf-list manpage, try to build and
install the man pages, warning the user when that is not possible
due to missing packages (xmlto and asciidoc).
Signed-off-by: Borislav Petkov <borislav.petkov@amd.com>
Tested-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Link: http://lkml.kernel.org/n/tip-ij71ysszkdvz3fy3wr331bke@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Using the same strategies as in the tmp.perf/trace2, i.e. the 'trace'
tool implemented by tglx, just updated to the current codebase.
Example:
[root@sandy linux]# perf trace usleep 1 | tail
2.003: mmap(addr: 0, len: 4096, prot: 3, flags: 34, fd: 4294967295, off: 0 ) = -2128396288
2.017: mmap(addr: 0, len: 4096, prot: 3, flags: 34, fd: 4294967295, off: 0 ) = -2128400384
2.029: arch_prctl(option: 4098, arg2: 140146949441280, arg3: 140146949435392, arg4: 34, arg5: 4294967295) = 0
2.084: mprotect(start: 208741634048, len: 16384, prot: 1 ) = 0
2.098: mprotect(start: 208735956992, len: 4096, prot: 1 ) = 0
2.122: munmap(addr: 140146949447680, len: 91882 ) = 0
2.359: brk(brk: 0 ) = 28987392
2.371: brk(brk: 29122560 ) = 29122560
2.490: nanosleep(rqtp: 140735694241504, rmtp: 0 ) = 0
2.507: exit_group(error_code: 0
[root@sandy linux]#
For now the timestamp and duration are always on, will be selectable.
Also if multiple threads are being monitored, its tid will appear.
The ret output continues to be interpreted a la strace.
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: Namhyung Kim <namhyung@gmail.com>
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-ly9ulroru4my5isn0xe9gr0m@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Event parsing tests are broken by following commit:
perf tool: Precise mode requires exclude_guest
commit 1342798cc1
Author: David Ahern <dsahern@gmail.com>
Date: Thu Sep 13 14:59:13 2012 -0600
which enables 'exclude_guest' modifier any time the 'precise'
modifier is detected.
Fixing related tests and adding special comment.
Reported-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Signed-off-by: Jiri Olsa <jolsa@redhat.com>
Tested-by: David Ahern <dsahern@gmail.com>
Acked-by: David Ahern <dsahern@gmail.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Namhyung Kim <namhyung@gmail.com>
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Namhyung Kim reported that the build fails with:
GEN python/perf.so
gcc: error: python_ext_build/tmp//../../libtraceevent.a: No such file or directory
error: command 'gcc' failed with exit status 1
cp: cannot stat `python_ext_build/lib/perf.so': No such file or directory
make: *** [python/perf.so] Error 1
We need to propagate the TE_PATH variable to the setup.py file.
Reported-by: Namhyung Kim <namhyung@kernel.org>
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Link: http://lkml.kernel.org/n/tip-8umiPbm4sxpknKivbjgykhut@git.kernel.org
[ Fixed superfluous variable build error. ]
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Pull perf/urgent fixes from Arnaldo Carvalho de Melo:
* The python binding needs to link with libtraceevent and to initialize
the 'page_size' variable so that mmaping works again.
* The callchain folding character that appears on the TUI just before
the overhead had disappeared due to recent changes, add it back.
* Intel PEBS in VT-x context uses the DS address as a guest linear address,
even though its programmed by the host as a host linear address. This either
results in guest memory corruption and or the hardware faulting and 'crashing'
the virtual machine. Therefore we have to disable PEBS on VT-x enter and
re-enable on VT-x exit, enforcing a strict exclude_guest.
Kernel side enforcement fix by Peter Zijlstra, tooling side fix by David Ahern.
* Fix build on sparc due to UAPI, fix from David Miller.
* Fixes for the srclike sort key for unresolved symbols and when processing
samples in JITted code, where we don't have an ELF file, just an special
symbol table, fixes from Namhyung Kim.
* Fix some leaks in libtraceevent, from Steven Rostedt.
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Summary of events per Peter:
"Intel PEBS in VT-x context uses the DS address as a guest linear address,
even though its programmed by the host as a host linear address. This
either results in guest memory corruption and or the hardware faulting and
'crashing' the virtual machine. Therefore we have to disable PEBS on VT-x
enter and re-enable on VT-x exit, enforcing a strict exclude_guest.
AMB IBS does work but doesn't currently support exclude_* at all,
setting an exclude_* bit will make it fail."
This patch handles userspace perf command, setting the exclude_guest
attribute if precise mode is requested, but only if a user has not
specified a request for guest or host only profiling (G or H attribute).
Kernel side AMD currently ignores all exclude_* bits, so there is no impact
to existing IBS code paths. Robert Richter has a patch where IBS code will
return EINVAL if an exclude_* bit is set. When this goes in it means use
of :p on AMD with IBS will first fail with EINVAL (because exclude_guest
will be set). Then the existing fallback code within perf will unset
exclude_guest and try again. The second attempt will succeed if the CPU
supports IBS profiling.
Signed-off-by: David Ahern <dsahern@gmail.com>
Acked-by: Peter Zijlstra <peterz@infradead.org>
Acked-by: Robert Richter <robert.richter@amd.com>
Tested-by: Robert Richter <robert.richter@amd.com>
Reviewed-by: Robert Richter <robert.richter@amd.com>
Cc: Avi Kivity <avi@redhat.com>
Cc: Gleb Natapov <gleb@redhat.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Robert Richter <robert.richter@amd.com>
Link: http://lkml.kernel.org/r/1347569955-54626-2-git-send-email-dsahern@gmail.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
The UAPI commits forgot to test tooling builds such as tools/perf/,
and this fixes the fallout.
Manual conversion.
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Pull perf updates from Ingo Molnar:
"This tree includes some late late perf items that missed the first
round:
tools:
- Bash auto completion improvements, now we can auto complete the
tools long options, tracepoint event names, etc, from Namhyung Kim.
- Look up thread using tid instead of pid in 'perf sched'.
- Move global variables into a perf_kvm struct, from David Ahern.
- Hists refactorings, preparatory for improved 'diff' command, from
Jiri Olsa.
- Hists refactorings, preparatory for event group viewieng work, from
Namhyung Kim.
- Remove double negation on optional feature macro definitions, from
Namhyung Kim.
- Remove several cases of needless global variables, on most
builtins.
- misc fixes
kernel:
- sysfs support for IBS on AMD CPUs, from Robert Richter.
- Support for an upcoming Intel CPU, the Xeon-Phi / Knights Corner
HPC blade PMU, from Vince Weaver.
- misc fixes"
* 'perf-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip: (46 commits)
perf: Fix perf_cgroup_switch for sw-events
perf: Clarify perf_cpu_context::active_pmu usage by renaming it to ::unique_pmu
perf/AMD/IBS: Add sysfs support
perf hists: Add more helpers for hist entry stat
perf hists: Move he->stat.nr_events initialization to a template
perf hists: Introduce struct he_stat
perf diff: Removing the total_period argument from output code
perf tool: Add hpp interface to enable/disable hpp column
perf tools: Removing hists pair argument from output path
perf hists: Separate overhead and baseline columns
perf diff: Refactor diff displacement possition info
perf hists: Add struct hists pointer to struct hist_entry
perf tools: Complete tracepoint event names
perf/x86: Add support for Intel Xeon-Phi Knights Corner PMU
perf evlist: Remove some unused methods
perf evlist: Introduce add_newtp method
perf kvm: Move global variables into a perf_kvm struct
perf tools: Convert to BACKTRACE_SUPPORT
perf tools: Long option completion support for each subcommands
perf tools: Complete long option names of perf command
...
In __rb_erase_color(), we were always setting a node to black after
exiting the main loop. And in one case, after fixing up the tree to
satisfy all rbtree invariants, we were setting the current node to root
just to guarantee a loop exit, at which point the root would be set to
black. However this is not necessary, as the root of an rbtree is already
known to be black. The only case where the color flip is required is when
we exit the loop due to the current node being red, and it's easiest to
just do the flip at that point instead of doing it after the loop.
[adrian.hunter@intel.com: perf tools: fix build for another rbtree.c change]
Signed-off-by: Michel Lespinasse <walken@google.com>
Cc: Andrea Arcangeli <aarcange@redhat.com>
Acked-by: David Woodhouse <David.Woodhouse@intel.com>
Cc: Rik van Riel <riel@redhat.com>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Daniel Santos <daniel.santos@pobox.com>
Cc: Jens Axboe <axboe@kernel.dk>
Cc: "Eric W. Biederman" <ebiederm@xmission.com>
Signed-off-by: Adrian Hunter <adrian.hunter@intel.com>
Cc: Alexander Shishkin <alexander.shishkin@intel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
When we were processing a PERF_RECORD_EXIT event we first used
machine__findnew_thread for both the thread exiting and for its parent,
only to use just the thread struct associated with the one exiting, and
to just delete it.
If it existed, i.e. not created at this very moment in
machine__findnew_thread, it will be moved to the machine->dead_threads
linked list, because we may have hist_entries pointing to it, but if it
was created just do be deleted, it will just sit there with no
references at all.
Use the new machine__find_thread() method so that if it is not there, we
don't create it.
As a bonus the parent thread will also not be created at this point.
Create process_fork() and process_exit() helpers to use this and make
the builtins use it instead of the generic process_task(), ditched by
this patch.
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: Namhyung Kim <namhyung@gmail.com>
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-z7n2y98ebjyrvmytaope4vdl@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
