Since the branch ip has been added to call stack for easier browsing,
this patch adds more branch information. For example, add a flag to
indicate if this ip is a branch, and also add with the branch flag.
Then we can know if the cursor node represents a branch and know what
the branch flag it has.
The branch history code has a loop detection pass that removes loops. It
would be nice for knowing how many loops were removed then in next
steps, we can compute out the average number of iterations.
For example:
Before remove_loops(),
entry0: from = 0x100, to = 0x200
entry1: from = 0x300, to = 0x250
entry2: from = 0x300, to = 0x250
entry3: from = 0x300, to = 0x250
entry4: from = 0x700, to = 0x800
After remove_loops()
entry0: from = 0x100, to = 0x200
entry1: from = 0x300, to = 0x250
entry2: from = 0x700, to = 0x800
The original entry2 and entry3 are removed. So the number of iterations
(from = 0x300, to = 0x250) is equal to removed number + 1 (2 + 1).
iterations = removed number + 1;
average iteractions = Sum(iteractions) / number of samples
This formula ignores other cases, for example, iterations cross multiple
buffers and one buffer contains 2+ loops. Because in practice, it's good
enough.
Signed-off-by: Yao Jin <yao.jin@linux.intel.com>
Acked-by: Andi Kleen <ak@linux.intel.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Kan Liang <kan.liang@intel.com>
Cc: Linux-kernel@vger.kernel.org
Cc: Yao Jin <yao.jin@linux.intel.com>
Link: http://lkml.kernel.org/n/1477876794-30749-2-git-send-email-yao.jin@linux.intel.com
[ Renamed 'iter' to 'nr_loop_iter' for clarity ]
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Since 841e3558b2 ("perf callchain: Recording 'dwarf' callchains do not
need DWARF unwinding support"), --call-graph dwarf is allowed in 'perf
record' even without unwind support. A couple of other places don't
reflect this yet though: the help text should list dwarf as a valid
record mode and the dump_size config should be respected too.
Signed-off-by: Rabin Vincent <rabinv@axis.com>
Cc: He Kuang <hekuang@huawei.com>
Fixes: 841e3558b2 ("perf callchain: Recording 'dwarf' callchains do not need DWARF unwinding support")
Link: http://lkml.kernel.org/r/1470837148-7642-1-git-send-email-rabin.vincent@axis.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
The recent perf_evsel__fprintf_callchain() move to evsel.c added several
new symbol requirements to the python binding, for instance:
# perf test -v python
16: Try 'import perf' in python, checking link problems :
--- start ---
test child forked, pid 18030
Traceback (most recent call last):
File "<stdin>", line 1, in <module>
ImportError: /tmp/build/perf/python/perf.so: undefined symbol:
callchain_cursor
test child finished with -1
---- end ----
Try 'import perf' in python, checking link problems: FAILED!
#
This would require linking against callchain.c to access to the global
callchain_cursor variables.
Since lots of functions already receive as a parameter a
callchain_cursor struct pointer, make that be the case for some more
function so that we can start phasing out usage of yet another global
variable.
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Wang Nan <wangnan0@huawei.com>
Link: http://lkml.kernel.org/n/tip-djko3097eyg2rn66v2qcqfvn@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
The flat callchain mode is to print all chains in a single, simple
hierarchy so make it easy to see.
Currently perf report --tui doesn't show flat callchains properly. With
flat callchains, only leaf nodes are added to the final rbtree so it
should show entries in parent nodes. To do that, add parent_val list to
struct callchain_node and show them along with the (normal) val list.
For example, consider following callchains with '-g graph'.
$ perf report -g graph
- 39.93% swapper [kernel.vmlinux] [k] intel_idle
intel_idle
cpuidle_enter_state
cpuidle_enter
call_cpuidle
- cpu_startup_entry
28.63% start_secondary
- 11.30% rest_init
start_kernel
x86_64_start_reservations
x86_64_start_kernel
Before:
$ perf report -g flat
- 39.93% swapper [kernel.vmlinux] [k] intel_idle
28.63% start_secondary
- 11.30% rest_init
start_kernel
x86_64_start_reservations
x86_64_start_kernel
After:
$ perf report -g flat
- 39.93% swapper [kernel.vmlinux] [k] intel_idle
- 28.63% intel_idle
cpuidle_enter_state
cpuidle_enter
call_cpuidle
cpu_startup_entry
start_secondary
- 11.30% intel_idle
cpuidle_enter_state
cpuidle_enter
call_cpuidle
cpu_startup_entry
start_kernel
x86_64_start_reservations
x86_64_start_kernel
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
Tested-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Tested-by: Brendan Gregg <brendan.d.gregg@gmail.com>
Cc: Andi Kleen <andi@firstfloor.org>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Kan Liang <kan.liang@intel.com>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Link: http://lkml.kernel.org/r/1447047946-1691-8-git-send-email-namhyung@kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Now -g/--call-graph option supports how to display callchain values.
Possible values are 'percent', 'period' and 'count'. The percent is
same as before and it's the default behavior. The period displays the
raw period value rather than the percentage. The count displays the
number of occurrences.
$ perf report --no-children --stdio -g percent
...
39.93% swapper [kernel.vmlinux] [k] intel_idel
|
---intel_idle
cpuidle_enter_state
cpuidle_enter
call_cpuidle
cpu_startup_entry
|
|--28.63%-- start_secondary
|
--11.30%-- rest_init
$ perf report --no-children --show-total-period --stdio -g period
...
39.93% 13018705 swapper [kernel.vmlinux] [k] intel_idel
|
---intel_idle
cpuidle_enter_state
cpuidle_enter
call_cpuidle
cpu_startup_entry
|
|--9334403-- start_secondary
|
--3684302-- rest_init
$ perf report --no-children --show-nr-samples --stdio -g count
...
39.93% 80 swapper [kernel.vmlinux] [k] intel_idel
|
---intel_idle
cpuidle_enter_state
cpuidle_enter
call_cpuidle
cpu_startup_entry
|
|--57-- start_secondary
|
--23-- rest_init
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
Acked-by: Brendan Gregg <brendan.d.gregg@gmail.com>
Cc: Andi Kleen <andi@firstfloor.org>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Kan Liang <kan.liang@intel.com>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Link: http://lkml.kernel.org/r/1447047946-1691-6-git-send-email-namhyung@kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Add new call chain option (-g) 'folded' to print callchains in a line.
The callchains are separated by semicolons, and preceded by (absolute)
percent values and a space.
For example, the following 20 lines can be printed in 3 lines with the
folded output mode:
$ perf report -g flat --no-children | grep -v ^# | head -20
60.48% swapper [kernel.vmlinux] [k] intel_idle
54.60%
intel_idle
cpuidle_enter_state
cpuidle_enter
call_cpuidle
cpu_startup_entry
start_secondary
5.88%
intel_idle
cpuidle_enter_state
cpuidle_enter
call_cpuidle
cpu_startup_entry
rest_init
start_kernel
x86_64_start_reservations
x86_64_start_kernel
$ perf report -g folded --no-children | grep -v ^# | head -3
60.48% swapper [kernel.vmlinux] [k] intel_idle
54.60% intel_idle;cpuidle_enter_state;cpuidle_enter;call_cpuidle;cpu_startup_entry;start_secondary
5.88% intel_idle;cpuidle_enter_state;cpuidle_enter;call_cpuidle;cpu_startup_entry;rest_init;start_kernel;x86_64_start_reservations;x86_64_start_kernel
This mode is supported only for --stdio now and intended to be used by
some scripts like in FlameGraphs[1]. Support for other UI might be
added later.
[1] http://www.brendangregg.com/FlameGraphs/cpuflamegraphs.html
Requested-and-Tested-by: Brendan Gregg <brendan.d.gregg@gmail.com>
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
Tested-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Acked-by: Jiri Olsa <jolsa@kernel.org>
Cc: Andi Kleen <andi@firstfloor.org>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Kan Liang <kan.liang@intel.com>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Link: http://lkml.kernel.org/r/1447047946-1691-2-git-send-email-namhyung@kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Currently 'perf top --call-graph' option is same as 'perf record'. But
'perf top' also need to receive display options in 'perf report'. To do
that, change parse_callchain_report_opt() to allow record options too.
Now perf top can receive display options like below:
$ perf top --call-graph
Error: option `call-graph' requires a value
Usage: perf top [<options>]
--call-graph
<mode[,dump_size],output_type,min_percent[,print_limit],call_order[,branch]>
setup and enables call-graph (stack chain/backtrace)
recording: fp dwarf lbr, output_type (graph, flat,
fractal, or none), min percent threshold, optional
print limit, callchain order, key (function or
address), add branches
$ perf top --call-graph callee,graph,fp
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Borislav Petkov <bp@suse.de>
Cc: Brendan Gregg <brendan.d.gregg@gmail.com>
Cc: Chandler Carruth <chandlerc@gmail.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Stephane Eranian <eranian@google.com>
Cc: Wang Nan <wangnan0@huawei.com>
Link: http://lkml.kernel.org/r/1445495330-25416-2-git-send-email-namhyung@kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
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>
For lbr-as-callgraph we need to see the line number in the history,
because many LBR entries can be in a single function, and just
showing the same function name many times is not useful.
When the history code is configured to sort by address, also try to
resolve the address to a file:srcline and display this in the browser.
If that doesn't work still display the address.
This can be also useful without LBRs for understanding which call in a large
function (or in which inlined function) called something else.
Contains fixes from Namhyung Kim
v2: Refactor code into common function
v3: Fix GTK build
v4: Rebase
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/1415844328-4884-7-git-send-email-andi@firstfloor.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
When a new callchain child branch matches an existing one in the rbtree,
the comparison of its first entry is performed twice:
1) From append_chain_children() on branch lookup
2) If 1) reports a match, append_chain() then compares all entries of
the new branch against the matching node in the rbtree, and this
comparison includes the first entry of the new branch again.
Lets shortcut this by performing the whole comparison only from
append_chain() which then returns the result of the comparison between
the first entry of the new branch and the iterating node in the rbtree.
If the first entry matches, the lookup on the current level of siblings
stops and propagates to the children of the matching nodes.
This results in less comparisons performed by the CPU.
Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com>
Acked-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@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/r/1389713836-13375-3-git-send-email-fweisbec@gmail.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Current collapse stage has a scalability problem which can be reproduced
easily with a parallel kernel build.
This is because it needs to traverse every children of callchains
linearly during the collapse/merge stage.
Converting it to a rbtree reduced the overhead significantly.
On my 400MB perf.data file which recorded with make -j32 kernel build:
$ time perf --no-pager report --stdio > /dev/null
before:
real 6m22.073s
user 6m18.683s
sys 0m0.706s
after:
real 0m20.780s
user 0m19.962s
sys 0m0.689s
During the perf report the overhead on append_chain_children went down
from 96.69% to 18.16%:
- 18.16% perf perf [.] append_chain_children
- append_chain_children
- 77.48% append_chain_children
+ 69.79% merge_chain_branch
- 22.96% append_chain_children
+ 67.44% merge_chain_branch
+ 30.15% append_chain_children
+ 2.41% callchain_append
+ 7.25% callchain_append
+ 12.26% callchain_append
+ 10.22% merge_chain_branch
+ 11.58% perf perf [.] dso__find_symbol
+ 8.02% perf perf [.] sort__comm_cmp
+ 5.48% perf libc-2.17.so [.] malloc_consolidate
Reported-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Link: http://lkml.kernel.org/r/1381468543-25334-2-git-send-email-namhyung@kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
With programs with very large functions it can be useful to distinguish
the callgraph nodes on more than just function names. So for example if
you have multiple calls to the same function, it ends up being separate
nodes in the chain.
This patch adds a new key field to the callgraph options, that allows
comparing nodes on functions (as today, default) and addresses.
Longer term it would be nice to also handle src lines, but that would
need more changes and address is a reasonable proxy for it today.
I right now reference the global params, as there was no simple way to
register a params pointer.
Signed-off-by: Andi Kleen <ak@linux.intel.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Link: http://lkml.kernel.org/n/tip-0uskktybf0e7wrnoi5e9b9it@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
A sweep of the kernel for regex "kcalloc(sizeof" turned up 2 reversed
args, fixed in commit d3d09e1820 ("EDAC:
Fix kcalloc argument order") and also fixed in the networking commit
a1b1add07f ("gro: Fix kcalloc argument
order").
I know that was the regex used, because on seeing the 1st of these
changes, I wondered "how many other instances of this are there" and I
happened to just use "calloc(sizeof" as a regex and it in turn found
these additional reversed args instances in the perf code.
In the kcalloc cases, the changes are cosmetic, since the numbers are
simply multiplied. I had no desire to go data mining in userspace to
see if the same thing held true there, however.
Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Joe Perches <joe@perches.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Link: http://lkml.kernel.org/r/1359594349-25912-1-git-send-email-paul.gortmaker@windriver.com
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>
