Add basic plumbing for TopDown in perf stat
TopDown is intended to replace the frontend cycles idle/ backend cycles
idle metrics in standard perf stat output. These metrics are not
reliable in many workloads, due to out of order effects.
This implements a new --topdown mode in perf stat (similar to
--transaction) that measures the pipe line bottlenecks using
standardized formulas. The measurement can be all done with 5 counters
(one fixed counter)
The result are four metrics:
FrontendBound, BackendBound, BadSpeculation, Retiring
that describe the CPU pipeline behavior on a high level.
The full top down methology has many hierarchical metrics. This
implementation only supports level 1 which can be collected without
multiplexing. A full implementation of top down on top of perf is
available in pmu-tools toplev. (http://github.com/andikleen/pmu-tools)
The current version works on Intel Core CPUs starting with Sandy Bridge,
and Atom CPUs starting with Silvermont. In principle the generic
metrics should be also implementable on other out of order CPUs.
TopDown level 1 uses a set of abstracted metrics which are generic to
out of order CPU cores (although some CPUs may not implement all of
them):
topdown-total-slots Available slots in the pipeline
topdown-slots-issued Slots issued into the pipeline
topdown-slots-retired Slots successfully retired
topdown-fetch-bubbles Pipeline gaps in the frontend
topdown-recovery-bubbles Pipeline gaps during recovery
from misspeculation
These metrics then allow to compute four useful metrics:
FrontendBound, BackendBound, Retiring, BadSpeculation.
Add a new --topdown options to enable events. When --topdown is
specified set up events for all topdown events supported by the kernel.
Add topdown-* as a special case to the event parser, as is needed for
all events containing -.
The actual code to compute the metrics is in follow-on patches.
v2: Use standard sysctl read function.
v3: Move x86 specific code to arch/
v4: Enable --metric-only implicitly for topdown.
v5: Add --single-thread option to not force per core mode
v6: Fix output order of topdown metrics
v7: Allow combining with -d
v8: Remove --single-thread again
v9: Rename functions, adding arch_ and topdown_.
v10: Expand man page and describe TopDown better
Paste intro into commit description.
Print error when malloc fails.
Signed-off-by: Andi Kleen <ak@linux.intel.com>
Acked-by: Jiri Olsa <jolsa@kernel.org>
Link: http://lkml.kernel.org/r/1464119559-17203-1-git-send-email-andi@firstfloor.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Without this patch, the last output doesn't have timestamp appended if
--timestamp-filename is not explicitly provided. For example:
# perf record -a --switch-output &
[1] 11224
# kill -s SIGUSR2 11224
[ perf record: dump data: Woken up 1 times ]
# [ perf record: Dump perf.data.2015122622372823 ]
# fg
perf record -a --switch-output
^C[ perf record: Woken up 1 times to write data ]
[ perf record: Captured and wrote 0.027 MB perf.data (540 samples) ]
# ls -l
total 836
-rw------- 1 root root 33256 Dec 26 22:37 perf.data <---- *Odd*
-rw------- 1 root root 817156 Dec 26 22:37 perf.data.2015122622372823
Signed-off-by: Wang Nan <wangnan0@huawei.com>
Tested-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Masami Hiramatsu <mhiramat@kernel.org>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Zefan Li <lizefan@huawei.com>
Cc: pi3orama@163.com
Link: http://lkml.kernel.org/r/1461178794-40467-5-git-send-email-wangnan0@huawei.com
Signed-off-by: He Kuang <hekuang@huawei.com>
[ Updated man page, that also got an entry for --timestamp-filename ]
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Now, one can print the call chain for every encountered sys_exit event,
e.g.:
$ perf trace -e nanosleep --call-graph dwarf path/to/ex_sleep
1005.757 (1000.090 ms): ex_sleep/13167 nanosleep(...) = 0
syscall_slow_exit_work ([kernel.kallsyms])
syscall_return_slowpath ([kernel.kallsyms])
int_ret_from_sys_call ([kernel.kallsyms])
__nanosleep (/usr/lib/libc-2.23.so)
[unknown] (/usr/lib/libQt5Core.so.5.6.0)
QThread::sleep (/usr/lib/libQt5Core.so.5.6.0)
main (path/to/ex_sleep)
__libc_start_main (/usr/lib/libc-2.23.so)
_start (path/to/ex_sleep)
Note that it is advised to increase the number of mmap pages to prevent
event losses when using this new feature. Often, adding `-m 10M` to the
`perf trace` invocation is enough.
This feature is also available in strace when built with libunwind via
`strace -k`. Performance wise, this solution is much better:
$ time find path/to/linux &> /dev/null
real 0m0.051s
user 0m0.013s
sys 0m0.037s
$ time perf trace -m 800M --call-graph dwarf find path/to/linux &> /dev/null
real 0m2.624s
user 0m1.203s
sys 0m1.333s
$ time strace -k find path/to/linux &> /dev/null
real 0m35.398s
user 0m10.403s
sys 0m23.173s
Note that it is currently not possible to configure the print output.
Adding such a feature, similar to what is available in `perf script` via
its `--fields` knob can be added later on.
Signed-off-by: Milian Wolff <milian.wolff@kdab.com>
Tested-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Wang Nan <wangnan0@huawei.com>
LPU-Reference: 1460115255-17648-1-git-send-email-milian.wolff@kdab.com
[ Split from a larger patch, do not print the IP, left align,
remove dup call symbol__init(), added man page entry ]
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Document some features for specifying events in the perf list manpage:
- Event groups
- Leader sampling
- How to specify raw PMU events in the new syntax
- Global versus per process PMUs.
- Access restrictions
- Fix Intel SDM URL
v2: Lots of new content. address review feedback.
Signed-off-by: Andi Kleen <ak@linux.intel.com>
Acked-by: Jiri Olsa <jolsa@kernel.org>
Link: http://lkml.kernel.org/r/1459810686-15913-1-git-send-email-andi@firstfloor.org
[ Add quotes to some keywords, such as "any" ]
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
When using 'perf script' to look at PT traces it is often useful to
ignore the initialization code at the beginning.
On larger traces which may have many millions of instructions in
initialization code doing that in a pipeline can be very slow, with perf
script spending a lot of CPU time calling printf and writing data.
This patch adds an extension to the --itrace argument that skips 'n'
events (instructions, branches or transactions) at the beginning. This
is much more efficient.
v2:
Add support for BTS (Adrian Hunter)
Document in itrace.txt
Fix branch check
Check transactions and instructions too
Committer note:
To test intel_pt one needs to make sure VT-x isn't active, i.e.
stopping KVM guests on the test machine, as described by Andi Kleen
at http://lkml.kernel.org/r/20160301234953.GD23621@tassilo.jf.intel.com
Signed-off-by: Andi Kleen <ak@linux.intel.com>
Tested-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/r/1459187142-20035-1-git-send-email-andi@firstfloor.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Add a new mode to only print metrics. Sometimes we don't care about the
raw values, just want the computed metrics. This allows more compact
printing, so with -I each sample is only a single line. This also
allows easier plotting and processing with other tools.
The main target is with using --topdown, but it also works with -T and
standard perf stat. A few metrics are not supported.
To avoiding having to hardcode all the metrics in the code it uses a two
pass approach: first compute dummy metrics and only print the headers in
the print_metric callback. Then use the callback to print the actual
values.
There are some additional changes in the stat printout code to handle
all metrics being on a single line.
One issue is that the column code doesn't know in advance what events
are not supported by the CPU, and it would be hard to find out as this
could change based on dynamic conditions. That causes empty columns in
some cases.
The output can be fairly wide, often you may need more than 80 columns.
Example:
% perf stat -a -I 1000 --metric-only
1.001452803 frontend cycles idle insn per cycle stalled cycles per insn branch-misses of all branches
1.001452803 158.91% 0.66 2.39 2.92%
2.002192321 180.63% 0.76 2.08 2.96%
3.003088282 150.59% 0.62 2.57 2.84%
4.004369835 196.20% 0.98 1.62 3.79%
5.005227314 231.98% 0.84 1.90 4.71%
v2: Lots of updates.
v3: Use slightly narrower columns
v4: Add comment
Signed-off-by: Andi Kleen <ak@linux.intel.com>
Acked-by: Jiri Olsa <jolsa@kernel.org>
Link: http://lkml.kernel.org/r/1457049458-28956-6-git-send-email-andi@firstfloor.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
With all the recently added fields in the perf stat CSV output we should
finally document them in the man page. Do this here.
v2: Fix fields in documentation (Jiri)
v3: fix order of fields again (Jiri)
v4: Change order again.
v5: Document more fields (Jiri)
v6: Move time stamp first
v7: More fixes (Jiri)
Signed-off-by: Andi Kleen <ak@linux.intel.com>
Acked-by: Jiri Olsa <jolsa@kernel.org>
Link: http://lkml.kernel.org/r/1457049458-28956-5-git-send-email-andi@firstfloor.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
The --hierarchy option is to show output in hierarchy mode. It extends
folding/unfolding in the TUI and GTK browsers to support sort items as
well as callchains. Users can toggle the items to see the performance
result at wanted level.
$ perf report --hierarchy --tui
Overhead Command / Shared Object / Symbol
--------------------------------------------------
+ 32.96% gnome-shell
- 15.11% swapper
- 14.97% [kernel.vmlinux]
6.82% [k] intel_idle
0.66% [k] menu_select
0.43% [k] __hrtimer_start_range_ns
...
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
Acked-by: Pekka Enberg <penberg@kernel.org>
Cc: Andi Kleen <andi@firstfloor.org>
Cc: David Ahern <dsahern@gmail.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Cc: Wang Nan <wangnan0@huawei.com>
Link: http://lkml.kernel.org/r/1456326830-30456-17-git-send-email-namhyung@kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Allow user to easily switch all events to user or kernel space with simple
--all-user or --all-kernel options.
This will be handy within perf mem/c2c wrappers to switch easily monitoring
modes.
Committer note:
Testing it:
# perf record --all-kernel --all-user -a sleep 2
Error: option `all-user' cannot be used with all-kernel
Usage: perf record [<options>] [<command>]
or: perf record [<options>] -- <command> [<options>]
--all-user Configure all used events to run in user space.
--all-kernel Configure all used events to run in kernel space.
# perf record --all-user --all-kernel -a sleep 2
Error: option `all-kernel' cannot be used with all-user
Usage: perf record [<options>] [<command>]
or: perf record [<options>] -- <command> [<options>]
--all-kernel Configure all used events to run in kernel space.
--all-user Configure all used events to run in user space.
# perf record --all-user -a sleep 1
[ perf record: Woken up 1 times to write data ]
[ perf record: Captured and wrote 1.416 MB perf.data (162 samples) ]
# perf report | grep '\[k\]'
# perf record --all-kernel -a sleep 1
[ perf record: Woken up 1 times to write data ]
[ perf record: Captured and wrote 1.423 MB perf.data (296 samples) ]
# perf report | grep '\[\.\]'
#
Signed-off-by: Jiri Olsa <jolsa@kernel.org>
Tested-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Andi Kleen <ak@linux.intel.com>
Cc: David Ahern <dsahern@gmail.com>
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/r/1455525293-8671-2-git-send-email-jolsa@kernel.org
[ Made those options to be mutually exclusive ]
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
This patch adds a --jit/-j option to perf inject.
This options injects MMAP records into the perf.data file to cover the
jitted code mmaps. It also emits ELF images for each function in the
jidump file. Those images are created where the jitdump file is. The
MMAP records point to that location as well.
Typical flow:
$ perf record -k mono -- java -agentpath:libpjvmti.so java_class
$ perf inject --jit -i perf.data -o perf.data.jitted
$ perf report -i perf.data.jitted
Note that jitdump.h support is not limited to Java, it works with any
jitted environment modified to emit the jitdump file format, include
those where code can be jitted multiple times and moved around.
The jitdump.h format is adapted from the Oprofile project.
The genelf.c (ELF binary generation) depends on MD5 hash encoding for
the buildid. To enable this, libssl-dev must be installed. If not, then
genelf.c defaults to using urandom to generate the buildid, which is not
ideal. The Makefile auto-detects the presence on libssl-dev.
This version mmaps the jitdump file to create a marker MMAP record in
the perf.data file. The marker is used to detect jitdump and cause perf
inject to inject the jitted mmaps and generate ELF images for jitted
functions.
In V8, the following fixes and changes were made among other things:
- the jidump header format include a new flags field to be used
to carry information about the configuration of the runtime agent.
Contributed by: Adrian Hunter <adrian.hunter@intel.com>
- Fix mmap pgoff: MMAP event pgoff must be the offset within the ELF file
at which the code resides.
Contributed by: Adrian Hunter <adrian.hunter@intel.com>
- Fix ELF virtual addresses: perf tools expect the ELF virtual addresses of dynamic
objects to match the file offset.
Contributed by: Adrian Hunter <adrian.hunter@intel.com>
- JIT MMAP injection does not obey finished_round semantics. JIT MMAP injection injects all
MMAP events in one go, so it does not obey finished_round semantics, so drop the
finished_round events from the output perf.data file.
Contributed by: Adrian Hunter <adrian.hunter@intel.com>
Signed-off-by: Stephane Eranian <eranian@google.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Andi Kleen <ak@linux.intel.com>
Cc: Carl Love <cel@us.ibm.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: John McCutchan <johnmccutchan@google.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Pawel Moll <pawel.moll@arm.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Sonny Rao <sonnyrao@chromium.org>
Cc: Sukadev Bhattiprolu <sukadev@linux.vnet.ibm.com>
Link: http://lkml.kernel.org/r/1448874143-7269-3-git-send-email-eranian@google.com
[ Moved inject.build_ids ordering bits to a separate patch, fixed the NO_LIBELF=1 build ]
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
