Currently branch stacks can be only shown as edge histograms for
individual branches. I never found this display particularly useful.
This implements an alternative mode that creates histograms over
complete branch traces, instead of individual branches, similar to how
normal callgraphs are handled. This is done by putting it in front of
the normal callgraph and then using the normal callgraph histogram
infrastructure to unify them.
This way in complex functions we can understand the control flow that
lead to a particular sample, and may even see some control flow in the
caller for short functions.
Example (simplified, of course for such simple code this is usually not
needed), please run this after the whole patchkit is in, as at this
point in the patch order there is no --branch-history, that will be
added in a patch after this one:
tcall.c:
volatile a = 10000, b = 100000, c;
__attribute__((noinline)) f2()
{
c = a / b;
}
__attribute__((noinline)) f1()
{
f2();
f2();
}
main()
{
int i;
for (i = 0; i < 1000000; i++)
f1();
}
% perf record -b -g ./tsrc/tcall
[ perf record: Woken up 1 times to write data ]
[ perf record: Captured and wrote 0.044 MB perf.data (~1923 samples) ]
% perf report --no-children --branch-history
...
54.91% tcall.c:6 [.] f2 tcall
|
|--65.53%-- f2 tcall.c:5
| |
| |--70.83%-- f1 tcall.c:11
| | f1 tcall.c:10
| | main tcall.c:18
| | main tcall.c:18
| | main tcall.c:17
| | main tcall.c:17
| | f1 tcall.c:13
| | f1 tcall.c:13
| | f2 tcall.c:7
| | f2 tcall.c:5
| | f1 tcall.c:12
| | f1 tcall.c:12
| | f2 tcall.c:7
| | f2 tcall.c:5
| | f1 tcall.c:11
| |
| --29.17%-- f1 tcall.c:12
| f1 tcall.c:12
| f2 tcall.c:7
| f2 tcall.c:5
| f1 tcall.c:11
| f1 tcall.c:10
| main tcall.c:18
| main tcall.c:18
| main tcall.c:17
| main tcall.c:17
| f1 tcall.c:13
| f1 tcall.c:13
| f2 tcall.c:7
| f2 tcall.c:5
| f1 tcall.c:12
The default output is unchanged.
This is only implemented in perf report, no change to record or anywhere
else.
This adds the basic code to report:
- add a new "branch" option to the -g option parser to enable this mode
- when the flag is set include the LBR into the callstack in machine.c.
The rest of the history code is unchanged and doesn't know the
difference between LBR entry and normal call entry.
- detect overlaps with the callchain
- remove small loop duplicates in the LBR
Current limitations:
- The LBR flags (mispredict etc.) are not shown in the history
and LBR entries have no special marker.
- It would be nice if annotate marked the LBR entries somehow
(e.g. with arrows)
v2: Various fixes.
v3: Merge further patches into this one. Fix white space.
v4: Improve manpage. Address review feedback.
v5: Rename functions. Better error message without -g. Fix crash without
-b.
v6: Rebase
v7: Rebase. Use NO_ENTRY in memset.
v8: Port to latest tip. Move add_callchain_ip to separate
patch. Skip initial entries in callchain. Minor cleanups.
Signed-off-by: Andi Kleen <ak@linux.intel.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Link: http://lkml.kernel.org/r/1415844328-4884-3-git-send-email-andi@firstfloor.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
We need to define bfd_demangle() to either a wrapper for
cplus_demangle() or to a stub when NO_DEMANGLE is defined.
That is at odds with using bfd.h for some other reason, as it defines
bfd_demangle() and then if code that wants to use symbol.h, where the
above stubbing/wrapping is done, and bfd.h for other reasons, we end up
with a build error where bfd_demangle() is found to be redefined.
Avoid that by moving the stubbing/wrapping to symbol-elf.c, that is the
only user of such function. If we ever get to a point where there are
more valid users, we can then introduce a header for that.
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Andi Kleen <ak@linux.intel.com>
Cc: Borislav Petkov <bp@suse.de>
Cc: David Ahern <dsahern@gmail.com>
Cc: Don Zickus <dzickus@redhat.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-6wzjpe2fy9xtgchshulixlzw@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
When perf record finishes a session, it pre-processes samples in order
to write build-id info from DSOs that had samples.
During this process it'll call map__load() for the kernel map, and it
ends up calling dso__load_vmlinux_path() which replaces dso->long_name.
But this function checks kernel's build-id before searching vmlinux path
so it'll end up with a cryptic name, the pathname for the entry in the
~/.debug cache, which can be confusing to users.
This patch adds a flag to skip the build-id check during record, so
that it'll have the original vmlinux path for the kernel dso->long_name,
not the entry in the ~/.debug cache.
Before:
# perf record -va sleep 3
mmap size 528384B
[ perf record: Woken up 1 times to write data ]
[ perf record: Captured and wrote 0.196 MB perf.data (~8545 samples) ]
Looking at the vmlinux_path (7 entries long)
Using /home/namhyung/.debug/.build-id/f0/6e17aa50adf4d00b88925e03775de107611551 for symbols
After:
# perf record -va sleep 3
mmap size 528384B
[ perf record: Woken up 1 times to write data ]
[ perf record: Captured and wrote 0.193 MB perf.data (~8432 samples) ]
Looking at the vmlinux_path (7 entries long)
Using /lib/modules/3.16.4-1-ARCH/build/vmlinux for symbols
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Namhyung Kim <namhyung.kim@lge.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/1415063674-17206-7-git-send-email-namhyung@kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Some C++ symbols have very long name and they make column length longer.
Most of them are about parameters including templates and we can ignore
such info most of time IMHO.
This patch passes DMGL_NO_OPTS by default when calling bfd_demangle().
One can still see full symbols with -v/--verbose option.
before:
JS_CallFunctionValue(JSContext*, JSObject*, JS::Value, unsigned int, JS::Value*, JS::Value*)
after:
JS_CallFunctionValue
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Namhyung Kim <namhyung.kim@lge.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Link: http://lkml.kernel.org/r/1406785662-5534-9-git-send-email-namhyung@kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Currently vmlinux_path__init() only tries to find vmlinux file from
current directory, /boot and some canonical directories with version
number of the running kernel. This can be a problem when reporting old
data recorded on a kernel version not running currently.
We can use --symfs option for this but it's annoying for user to do it
always. As we already have the info in the perf.data file, it can be
changed to use it for the search automatically.
Before:
$ perf report
...
# Samples: 4K of event 'cpu-clock'
# Event count (approx.): 1067250000
#
# Overhead Command Shared Object Symbol
# ........ .......... ................. ..............................
71.87% swapper [kernel.kallsyms] [k] recover_probed_instruction
After:
# Overhead Command Shared Object Symbol
# ........ .......... ................. ....................
71.87% swapper [kernel.kallsyms] [k] native_safe_halt
This requires to change signature of symbol__init() to receive struct
perf_session_env *.
Reported-by: Minchan Kim <minchan@kernel.org>
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com>
Cc: Minchan Kim <minchan@kernel.org>
Cc: Namhyung Kim <namhyung.kim@lge.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/1407825645-24586-14-git-send-email-namhyung@kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
The --percentage option is for controlling overhead percentage
displayed. It can only receive either of "relative" or "absolute".
"relative" means it's relative to filtered entries only so that the
sum of shown entries will be always 100%. "absolute" means it retains
the original value before and after the filter is applied.
$ perf report -s comm
# Overhead Command
# ........ ............
#
74.19% cc1
7.61% gcc
6.11% as
4.35% sh
4.14% make
1.13% fixdep
...
$ perf report -s comm -c cc1,gcc --percentage absolute
# Overhead Command
# ........ ............
#
74.19% cc1
7.61% gcc
$ perf report -s comm -c cc1,gcc --percentage relative
# Overhead Command
# ........ ............
#
90.69% cc1
9.31% gcc
Note that it has zero effect if no filter was applied.
Suggested-by: Arnaldo Carvalho de Melo <acme@ghostprotocols.net>
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
Link: http://lkml.kernel.org/r/1397145720-8063-3-git-send-email-namhyung@kernel.org
Signed-off-by: Jiri Olsa <jolsa@redhat.com>
Allow to add events on the local functions without debuginfo.
(With the debuginfo, we can add events even on inlined functions)
Currently, probing on local functions requires debuginfo to
locate actual address. It is also possible without debuginfo since
we have symbol maps.
Without this change;
----
# ./perf probe -a t_show
Added new event:
probe:t_show (on t_show)
You can now use it in all perf tools, such as:
perf record -e probe:t_show -aR sleep 1
# ./perf probe -x perf -a identity__map_ip
no symbols found in /kbuild/ksrc/linux-3/tools/perf/perf, maybe install a debug package?
Failed to load map.
Error: Failed to add events. (-22)
----
As the above results, perf probe just put one event
on the first found symbol for kprobe event. Moreover,
for uprobe event, perf probe failed to find local
functions.
With this change;
----
# ./perf probe -a t_show
Added new events:
probe:t_show (on t_show)
probe:t_show_1 (on t_show)
probe:t_show_2 (on t_show)
probe:t_show_3 (on t_show)
You can now use it in all perf tools, such as:
perf record -e probe:t_show_3 -aR sleep 1
# ./perf probe -x perf -a identity__map_ip
Added new events:
probe_perf:identity__map_ip (on identity__map_ip in /kbuild/ksrc/linux-3/tools/perf/perf)
probe_perf:identity__map_ip_1 (on identity__map_ip in /kbuild/ksrc/linux-3/tools/perf/perf)
probe_perf:identity__map_ip_2 (on identity__map_ip in /kbuild/ksrc/linux-3/tools/perf/perf)
probe_perf:identity__map_ip_3 (on identity__map_ip in /kbuild/ksrc/linux-3/tools/perf/perf)
You can now use it in all perf tools, such as:
perf record -e probe_perf:identity__map_ip_3 -aR sleep 1
----
Now we succeed to put events on every given local functions
for both kprobes and uprobes. :)
Note that this also introduces some symbol rbtree
iteration macros; symbols__for_each, dso__for_each_symbol,
and map__for_each_symbol. These are for walking through
the symbol list in a map.
Changes from v2:
- Fix add_exec_to_probe_trace_events() not to convert address
to tp->symbol any more.
- Fix to set kernel probes based on ref_reloc_sym.
Signed-off-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: "David A. Long" <dave.long@linaro.org>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Oleg Nesterov <oleg@redhat.com>
Cc: Srikar Dronamraju <srikar@linux.vnet.ibm.com>
Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: yrl.pp-manager.tt@hitachi.com
Link: http://lkml.kernel.org/r/20140206053225.29635.15026.stgit@kbuild-fedora.yrl.intra.hitachi.co.jp
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
kcore can be used to view the running kernel object code. However,
kcore changes as modules are loaded and unloaded, and when the kernel
decides to modify its own code. Consequently it is useful to create a
copy of kcore at a particular time. Unlike vmlinux, kcore is not unique
for a given build-id. And in addition, the kallsyms and modules files
are also needed. The tool therefore creates a directory:
~/.debug/[kernel.kcore]/<build-id>/<YYYYmmddHHMMSShh>
which contains: kcore, kallsyms and modules.
Note that the copied kcore contains only code sections. See the
kcore_copy() function for how that is determined.
The tool will not make additional copies of kcore if there is already
one with the same modules at the same addresses.
Currently, perf tools will not look for kcore in the cache. That is
addressed in another patch.
Signed-off-by: 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: 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/525BF849.5030405@intel.com
[ renamed 'index' to 'idx' to avoid shadowing string.h symbol in f12,
use at least one member initializer when initializing a struct to
zeros, also to fix the build on f12 ]
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Running "perf top" on a machine with possibly invalid or non-matching
vmlinux at the various places results in no symbol resolving despite
/proc/kallsyms being present and valid.
Add a new option --ignore-vmlinux to explicitly indicate that we do not
want to use these kernels and just use what we have (kallsyms).
Signed-off-by: Willy Tarreau <w@1wt.eu>
Cc: Ingo Molnar <mingo@redhat.com>
Link: http://lkml.kernel.org/r/20130914083259.GA3418@1wt.eu
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
This patch adds the sorting and histogram support
functions to enable profiling of memory accesses.
The following sorting orders are added:
- symbol_daddr: data address symbol (or raw address)
- dso_daddr: data address shared object
- locked: access uses locked transaction
- tlb : TLB access
- mem : memory level of the access (L1, L2, L3, RAM, ...)
- snoop: access snoop mode
Signed-off-by: Stephane Eranian <eranian@google.com>
Cc: Andi Kleen <ak@linux.intel.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Namhyung Kim <namhyung.kim@lge.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/r/1359040242-8269-12-git-send-email-eranian@google.com
[ committer note: changed to cope with fc5871ed, the move of methods to
machine.[ch], and the rename of dsrc to data_src, to match the change
made in the PERF_SAMPLE_DSRC in a previous patch. ]
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
The elf.h header file is used for NO_LIBELF build too so it should be
included anyway. Also remove duplicated include of the header file in
symbol-*.c. This patch fixes following build error on NO_LIBELF build:
CC tests/hists_link.o
tests/hists_link.c: In function ‘setup_fake_machine’:
tests/hists_link.c:132:8: error: ‘STB_GLOBAL’ undeclared (first use in this function)
tests/hists_link.c:132:8: note: each undeclared identifier is reported only once for each function it appears in
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
Cc: David Ahern <dsahern@gmail.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Link: http://lkml.kernel.org/r/1356679009-32122-1-git-send-email-namhyung@kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
perf defines both __used and __unused variables to use for marking
unused variables. The variable __used is defined to
__attribute__((__unused__)), which contradicts the kernel definition to
__attribute__((__used__)) for new gcc versions. On Android, __used is
also defined in system headers and this leads to warnings like: warning:
'__used__' attribute ignored
__unused is not defined in the kernel and is not a standard definition.
If __unused is included everywhere instead of __used, this leads to
conflicts with glibc headers, since glibc has a variables with this name
in its headers.
The best approach is to use __maybe_unused, the definition used in the
kernel for __attribute__((unused)). In this way there is only one
definition in perf sources (instead of 2 definitions that point to the
same thing: __used and __unused) and it works on both Linux and Android.
This patch simply replaces all instances of __used and __unused with
__maybe_unused.
Signed-off-by: Irina Tirdea <irina.tirdea@intel.com>
Acked-by: Pekka Enberg <penberg@kernel.org>
Cc: David Ahern <dsahern@gmail.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Namhyung Kim <namhyung.kim@lge.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Steven Rostedt <rostedt@goodmis.org>
Link: http://lkml.kernel.org/r/1347315303-29906-7-git-send-email-irina.tirdea@intel.com
[ committer note: fixed up conflict with a116e05 in builtin-sched.c ]
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
perf uses the glibc version of basename(), by defining _GNU_SOURCE,
including string.h and not including libgen.h. The glibc version of
basename is better than the POSIX version since it does not modify its
argument.
Android has only one version of basename which is defined in libgen.h.
This version is the same as the glibc version.
Error on Android:
util/annotate.c: In function 'symbol__annotate_printf':
util/annotate.c:503:3: error: implicit declaration of function 'basename'
[-Werror=implicit-function-declaration]
util/annotate.c:503:3: error: nested extern declaration of 'basename'
[-Werror=nested-externs]
util/annotate.c:503:14: error: assignment makes pointer from integer without
a cast [-Werror]
On Android libgen.h should be included to define basename.
Signed-off-by: Irina Tirdea <irina.tirdea@intel.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Namhyung Kim <namhyung.kim@lge.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Steven Rostedt <rostedt@goodmis.org>
Link: http://lkml.kernel.org/r/1347065004-15306-6-git-send-email-irina.tirdea@intel.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
We keep both a 'runtime' elf image as well as a 'debug' elf image around
and generate symbols by looking at both of these.
This eliminates the need for the want_symtab/goto restart mechanism
combined with iterating over and reopening the elf images a second time.
Also give dso__synthsize_plt_symbols() the runtime image (which has
dynsyms) instead of the symbol image (which may only have a symtab and
no dynsyms).
Previously if a debug image was found all runtime images were ignored.
This fixes 2 issues:
- Symbol resolution to failure on PowerPC systems with debug symbols
installed, as the debug images lack a '.opd' section which contains
function descriptors.
- On all archs, plt synthesis failed when a debug image was loaded and
that debug image lacks a dynsym section while a runtime image has a
dynsym section.
Assumptions:
- If a .opd section exists, it is contained in the highest priority
image with a dynsym section.
- This generally implies that the debug image lacks a dynsym section
(ie: it is marked as NO_BITS).
Signed-off-by: Cody P Schafer <cody@linux.vnet.ibm.com>
Cc: David Hansen <dave@linux.vnet.ibm.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Matt Hellsley <matthltc@us.ibm.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Sukadev Bhattiprolu <sukadev@linux.vnet.ibm.com>
Link: http://lkml.kernel.org/r/1344637382-22789-17-git-send-email-cody@linux.vnet.ibm.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
To properly handle platforms with an opd section, both a runtime image
(which contains the opd section but possibly lacks symbols) and a symbol
image (which probably lacks an opd section but has symbols).
The next patch ("perf symbol: use both runtime and debug images")
adjusts the callsite in dso__load() to take advantage of being able to
pass both runtime & debug images.
Assumptions made here:
- The opd section, if it exists in the runtime image, has headers in
both the runtime image and the debug/syms image.
- The index of the opd section (again, only if it exists in the runtime
image) is the same in both the runtime and debug/symbols image.
Both of these are true on RHEL, but it is unclear how accurate they are
in general (on platforms with function descriptors in opd sections).
Signed-off-by: Cody P Schafer <cody@linux.vnet.ibm.com>
Cc: David Hansen <dave@linux.vnet.ibm.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Matt Hellsley <matthltc@us.ibm.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Sukadev Bhattiprolu <sukadev@linux.vnet.ibm.com>
Link: http://lkml.kernel.org/r/1344637382-22789-16-git-send-email-cody@linux.vnet.ibm.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
