| Age | Commit message (Collapse) | Author | Files | Lines |
|---|
|
Ingo reported this small 'perf bench sched pipe' output problem:
| $ ./perf bench sched pipe
| (executing 1000000 pipe operations between two tasks)
|
| Total time:4.898 sec
| $ 4.898586 usecs/op
| 204140 ops/sec
|
| the shell prompt came back before the usecs/op and ops/sec line
| was printed. Process teardown race, lack of wait() or so?
This caused by lack of calling waitpid() by parent process,
so I added it.
Signed-off-by: Hitoshi Mitake <[email protected]>
Cc: Rusty Russell <[email protected]>
Cc: Peter Zijlstra <[email protected]>
Cc: Paul Mackerras <[email protected]>
Cc: Mike Galbraith <[email protected]>
Cc: Arnaldo Carvalho de Melo <[email protected]>
Cc: Frederic Weisbecker <[email protected]>
Cc: Jiri Kosina <[email protected]>
LKML-Reference: <[email protected]>
Signed-off-by: Ingo Molnar <[email protected]>
|
|
modify perf.c get_debugfs_mntpnt() to use the util/debugfs.c
debugfs_find_mountpoint()
modify util/parse-events.c to use debugfs_valid_mountpoint().
Signed-off-by: Clark Williams <[email protected]>
Cc: Arnaldo Carvalho de Melo <[email protected]>
Cc: Peter Zijlstra <[email protected]>
LKML-Reference: <20091101155720.624cc87e@torg>
Signed-off-by: Ingo Molnar <[email protected]>
|
|
Add routines to locate the debugfs mount point and to manage the
mounting and unmounting of the debugfs.
Signed-off-by: Clark Williams <[email protected]>
Cc: Arnaldo Carvalho de Melo <[email protected]>
Cc: Peter Zijlstra <[email protected]>
LKML-Reference: <20091101155621.2b3503ee@torg>
Signed-off-by: Ingo Molnar <[email protected]>
|
|
The perf_event_open() system call returns EACCES if the user is
not root which results in a very confusing error message:
$ perf record -A -a -f
Error: perfcounter syscall returned with -1 (Permission denied)
Fatal: No CONFIG_PERF_EVENTS=y kernel support configured?
It turns out that's because perf tools are checking only for
EPERM. Fix that up to get a much better error message:
$ perf record -A -a -f
Fatal: Permission error - are you root?
Signed-off-by: Pekka Enberg <[email protected]>
Cc: Peter Zijlstra <[email protected]>
Cc: Paul Mackerras <[email protected]>
Cc: Frederic Weisbecker <[email protected]>
LKML-Reference: <[email protected]>
Signed-off-by: Ingo Molnar <[email protected]>
|
|
With this change 'perf record' will intercept PERF_RECORD_MMAP
calls, creating a linked list of DSOs, then when the session
finishes, it will traverse this list and read the buildids,
stashing them at the end of the file and will set up a new
feature bit in the header bitmask.
'perf report' will then notice this feature and populate the
'dsos' list and set the build ids.
When reading the symtabs it will refuse to load from a file that
doesn't have the same build id. This improves the
reliability of the profiler output, as symbols and profiling
data is more guaranteed to match.
Example:
[root@doppio ~]# perf report | head
/home/acme/bin/perf with build id b1ea544ac3746e7538972548a09aadecc5753868 not found, continuing without symbols
# Samples: 2621434559
#
# Overhead Command Shared Object Symbol
# ........ ............... ............................. ......
#
7.91% init [kernel] [k] read_hpet
7.64% init [kernel] [k] mwait_idle_with_hints
7.60% swapper [kernel] [k] read_hpet
7.60% swapper [kernel] [k] mwait_idle_with_hints
3.65% init [kernel] [k] 0xffffffffa02339d9
[root@doppio ~]#
In this case the 'perf' binary was an older one, vanished,
so its symbols probably wouldn't match or would cause subtly
different (and misleading) output.
Next patches will support the kernel as well, reading the build
id notes for it and the modules from /sys.
Another patch should also introduce a new plumbing command:
'perf list-buildids'
that will then be used in porcelain that is distro specific to
fetch -debuginfo packages where such buildids are present. This
will in turn allow for one to run 'perf record' in one machine
and 'perf report' in another.
Future work on having the buildid sent directly from the kernel
in the PERF_RECORD_MMAP event is needed to close races, as the
DSO can be changed during a 'perf record' session, but this
patch at least helps with non-corner cases and current/older
kernels.
Signed-off-by: Arnaldo Carvalho de Melo <[email protected]>
Cc: Ananth N Mavinakayanahalli <[email protected]>
Cc: Christoph Hellwig <[email protected]>
Cc: Frank Ch. Eigler <[email protected]>
Cc: Frederic Weisbecker <[email protected]>
Cc: Jason Baron <[email protected]>
Cc: Jim Keniston <[email protected]>
Cc: K. Prasad <[email protected]>
Cc: Masami Hiramatsu <[email protected]>
Cc: Peter Zijlstra <[email protected]>
Cc: Roland McGrath <[email protected]>
Cc: Srikar Dronamraju <[email protected]>
Cc: Steven Rostedt <[email protected]>
LKML-Reference: <[email protected]>
Signed-off-by: Ingo Molnar <[email protected]>
|
|
This patch modifies Makefile for new files related to 'bench'
subcommand. The new code is active from this point on.
Signed-off-by: Hitoshi Mitake <[email protected]>
Cc: Rusty Russell <[email protected]>
Cc: Peter Zijlstra <[email protected]>
Cc: Mike Galbraith <[email protected]>
Cc: Arnaldo Carvalho de Melo <[email protected]>
Cc: [email protected]
Cc: Jiri Kosina <[email protected]>
LKML-Reference: <[email protected]>
Signed-off-by: Ingo Molnar <[email protected]>
|
|
This patch modifies perf.c for invoking 'bench' subcommand.
Signed-off-by: Hitoshi Mitake <[email protected]>
Cc: Rusty Russell <[email protected]>
Cc: Peter Zijlstra <[email protected]>
Cc: Mike Galbraith <[email protected]>
Cc: Arnaldo Carvalho de Melo <[email protected]>
Cc: [email protected]
Cc: Jiri Kosina <[email protected]>
LKML-Reference: <[email protected]>
Signed-off-by: Ingo Molnar <[email protected]>
|
|
This patch modifies builtin.h to add prototype of cmd_bench().
Signed-off-by: Hitoshi Mitake <[email protected]>
Cc: Rusty Russell <[email protected]>
Cc: Peter Zijlstra <[email protected]>
Cc: Mike Galbraith <[email protected]>
Cc: Arnaldo Carvalho de Melo <[email protected]>
Cc: [email protected]
Cc: Jiri Kosina <[email protected]>
LKML-Reference: <[email protected]>
Signed-off-by: Ingo Molnar <[email protected]>
|
|
This patch adds builtin-bench.c
builtin-bench.c is a general framework for benchmark suites.
Signed-off-by: Hitoshi Mitake <[email protected]>
Cc: Rusty Russell <[email protected]>
Cc: Peter Zijlstra <[email protected]>
Cc: Mike Galbraith <[email protected]>
Cc: Arnaldo Carvalho de Melo <[email protected]>
Cc: [email protected]
Cc: Jiri Kosina <[email protected]>
LKML-Reference: <[email protected]>
Signed-off-by: Ingo Molnar <[email protected]>
|
|
This patch adds bench/sched-pipe.c.
bench/sched-pipe.c is a benchmark program
to measure performance of pipe() system call.
This benchmark is based on pipe-test-1m.c by Ingo Molnar:
http://people.redhat.com/mingo/cfs-scheduler/tools/pipe-test-1m.c
Example of use:
% perf bench sched pipe
(executing 1000000 pipe operations between two tasks)
Total time:4.499 sec
4.499179 usecs/op
222262 ops/sec
% perf bench sched pipe -s -l 1000
0.015
Signed-off-by: Hitoshi Mitake <[email protected]>
Cc: Rusty Russell <[email protected]>
Cc: Peter Zijlstra <[email protected]>
Cc: Mike Galbraith <[email protected]>
Cc: Arnaldo Carvalho de Melo <[email protected]>
Cc: [email protected]
Cc: Jiri Kosina <[email protected]>
LKML-Reference: <[email protected]>
Signed-off-by: Ingo Molnar <[email protected]>
|
|
mechanisms based on hackbench
This patch adds bench/sched-messaging.c.
This benchmark measures performance of scheduler and IPC
mechanisms, and is based on hackbench by Rusty Russell.
Example of usage:
% perf bench sched messaging -g 20 -l 1000 -s
5.432 # in sec
% perf bench sched messaging # run with default
options (20 sender and receiver processes per group)
(10 groups == 400 processes run)
Total time:0.308 sec
% perf bench sched messaging -t -g 20 # # be multi-thread,
with 20 groups (20 sender and receiver threads per group)
(20 groups == 800 threads run)
Total time:0.582 sec
( Rusty is the original author of hackbench.c and he said the code is
and was under the GPLv2 so fine to be merged. )
Signed-off-by: Hitoshi Mitake <[email protected]>
Acked-by: Rusty Russell <[email protected]>
Cc: Peter Zijlstra <[email protected]>
Cc: Mike Galbraith <[email protected]>
Cc: Arnaldo Carvalho de Melo <[email protected]>
Cc: [email protected]
Cc: Jiri Kosina <[email protected]>
LKML-Reference: <[email protected]>
Signed-off-by: Ingo Molnar <[email protected]>
|
|
This patch adds bench/ directory and bench/bench.h.
bench/ directory will contain modules for bench subcommand.
bench/bench.h is for listing prototypes of module functions.
Signed-off-by: Hitoshi Mitake <[email protected]>
Cc: Rusty Russell <[email protected]>
Cc: Peter Zijlstra <[email protected]>
Cc: Mike Galbraith <[email protected]>
Cc: Arnaldo Carvalho de Melo <[email protected]>
Cc: [email protected]
Cc: Jiri Kosina <[email protected]>
LKML-Reference: <[email protected]>
Signed-off-by: Ingo Molnar <[email protected]>
|
|
Rename the group name of perf probe events to 'probe'.
Signed-off-by: Masami Hiramatsu <[email protected]>
Acked-by: Frederic Weisbecker <[email protected]>
Cc: Steven Rostedt <[email protected]>
Cc: Jim Keniston <[email protected]>
Cc: Ananth N Mavinakayanahalli <[email protected]>
Cc: Christoph Hellwig <[email protected]>
Cc: Frank Ch. Eigler <[email protected]>
Cc: Jason Baron <[email protected]>
Cc: K.Prasad <[email protected]>
Cc: Peter Zijlstra <[email protected]>
Cc: Srikar Dronamraju <[email protected]>
LKML-Reference: <20091104001238.3454.70508.stgit@harusame>
Signed-off-by: Ingo Molnar <[email protected]>
|
|
Fall back to non-dwarf probe point if the probe definition may
not need dwarf analysis, when perf can't find vmlinux/debuginfo.
This might skip some inlined code of target function.
Signed-off-by: Masami Hiramatsu <[email protected]>
Acked-by: Frederic Weisbecker <[email protected]>
Cc: Steven Rostedt <[email protected]>
Cc: Jim Keniston <[email protected]>
Cc: Ananth N Mavinakayanahalli <[email protected]>
Cc: Christoph Hellwig <[email protected]>
Cc: Frank Ch. Eigler <[email protected]>
Cc: Jason Baron <[email protected]>
Cc: K.Prasad <[email protected]>
Cc: Peter Zijlstra <[email protected]>
Cc: Srikar Dronamraju <[email protected]>
LKML-Reference: <20091104001229.3454.63987.stgit@harusame>
Signed-off-by: Ingo Molnar <[email protected]>
|
|
Improve error messages in perf-probe so that users can figure
out problems easily.
Reported-by: Ingo Molnar <[email protected]>
Signed-off-by: Masami Hiramatsu <[email protected]>
Acked-by: Frederic Weisbecker <[email protected]>
Cc: Steven Rostedt <[email protected]>
Cc: Jim Keniston <[email protected]>
Cc: Ananth N Mavinakayanahalli <[email protected]>
Cc: Christoph Hellwig <[email protected]>
Cc: Frank Ch. Eigler <[email protected]>
Cc: Jason Baron <[email protected]>
Cc: K.Prasad <[email protected]>
Cc: Peter Zijlstra <[email protected]>
Cc: Srikar Dronamraju <[email protected]>
LKML-Reference: <20091104001221.3454.52030.stgit@harusame>
Signed-off-by: Ingo Molnar <[email protected]>
|
|
Update Documentation/perf-probe.txt accoding to recent
syntax changes.
Signed-off-by: Masami Hiramatsu <[email protected]>
Acked-by: Frederic Weisbecker <[email protected]>
Cc: Steven Rostedt <[email protected]>
Cc: Jim Keniston <[email protected]>
Cc: Ananth N Mavinakayanahalli <[email protected]>
Cc: Christoph Hellwig <[email protected]>
Cc: Frank Ch. Eigler <[email protected]>
Cc: Jason Baron <[email protected]>
Cc: K.Prasad <[email protected]>
Cc: Peter Zijlstra <[email protected]>
Cc: Srikar Dronamraju <[email protected]>
LKML-Reference: <20091104001212.3454.19415.stgit@harusame>
Signed-off-by: Ingo Molnar <[email protected]>
|
|
So that we can run it without having a DSO instance.
Signed-off-by: Arnaldo Carvalho de Melo <[email protected]>
Cc: Frederic Weisbecker <[email protected]>
Cc: Peter Zijlstra <[email protected]>
Cc: Paul Mackerras <[email protected]>
Cc: Mike Galbraith <[email protected]>
LKML-Reference: <[email protected]>
Signed-off-by: Ingo Molnar <[email protected]>
|
|
Conflicts:
tools/perf/Makefile
Merge reason: Resolve the conflict, merge to upstream and merge in
perf fixes so we can add a dependent patch.
Signed-off-by: Ingo Molnar <[email protected]>
|
|
Brown paper bag bug introduced in:
66bd8424cc05e800db384053bf7ab967e4658468 ("perf tools: Delay
loading symtabs till we hit a map with it")
Without this we were not loading any symtabs that happened to be
on a DSO for which the allocated memory for ->loaded was !0.
Signed-off-by: Arnaldo Carvalho de Melo <[email protected]>
Cc: Frederic Weisbecker <[email protected]>
Cc: Peter Zijlstra <[email protected]>
Cc: Paul Mackerras <[email protected]>
Cc: Mike Galbraith <[email protected]>
LKML-Reference: <[email protected]>
Signed-off-by: Ingo Molnar <[email protected]>
|
|
Pointed out by checkpatch.
Signed-off-by: Thiago Farina <[email protected]>
Cc: [email protected]
Cc: [email protected]
Cc: [email protected]
LKML-Reference: <[email protected]>
Signed-off-by: Ingo Molnar <[email protected]>
|
|
Annotate away this false positive warning on older GCCs:
cc1: warnings being treated as errors
builtin-probe.c: In function ‘parse_probe_event’:
builtin-probe.c:72: warning: ‘nc’ is used uninitialized in this function
Signed-off-by: Arnaldo Carvalho de Melo <[email protected]>
Acked-by: Masami Hiramatsu <[email protected]>
LKML-Reference: <[email protected]>
Signed-off-by: Ingo Molnar <[email protected]>
|
|
Fix:
util/map.c: In function ‘map__find_symbol’:
util/map.c:97: error: field precision should have type ‘int’, but argument 3 has type ‘size_t’
Also clean up some line wrap damage - we dont line-wrap printk
messages.
Cc: Arnaldo Carvalho de Melo <[email protected]>
Cc: Frederic Weisbecker <[email protected]>
Cc: Peter Zijlstra <[email protected]>
Cc: Paul Mackerras <[email protected]>
Cc: Mike Galbraith <[email protected]>
LKML-Reference: <[email protected]>
Signed-off-by: Ingo Molnar <[email protected]>
|
|
Instead of:
no symbols found in /usr/lib/gstreamer-0.10/libgsttypefindfunctions.so (deleted), maybe install a debug package?
no symbols found in /usr/lib/gstreamer-0.10/libgstaudioconvert.so (deleted), maybe install a debug package?
We now emit:
/usr/lib/gstreamer-0.10/libgsttypefindfunctions.so was updated, restart the long running apps that use it!
/usr/lib/gstreamer-0.10/libgstaudioconvert.so was updated, restart the long running apps that use it!
Which is far less misleading about what the cause of the
symbol mismatch is.
Signed-off-by: Arnaldo Carvalho de Melo <[email protected]>
Cc: Frederic Weisbecker <[email protected]>
Cc: Peter Zijlstra <[email protected]>
Cc: Paul Mackerras <[email protected]>
Cc: Mike Galbraith <[email protected]>
LKML-Reference: <[email protected]>
Signed-off-by: Ingo Molnar <[email protected]>
|
|
Before we were storing this in the DSO, but in fact this is a
property of the 'symbol' class, not something that will vary
among DSOs, so move it to a global variable and initialize it
using the existing symbol__init routine.
Signed-off-by: Arnaldo Carvalho de Melo <[email protected]>
Cc: Frederic Weisbecker <[email protected]>
Cc: Peter Zijlstra <[email protected]>
Cc: Paul Mackerras <[email protected]>
Cc: Mike Galbraith <[email protected]>
LKML-Reference: <[email protected]>
Signed-off-by: Ingo Molnar <[email protected]>
|
|
Signed-off-by: Arnaldo Carvalho de Melo <[email protected]>
Cc: Frederic Weisbecker <[email protected]>
Cc: Peter Zijlstra <[email protected]>
Cc: Paul Mackerras <[email protected]>
Cc: Mike Galbraith <[email protected]>
LKML-Reference: <[email protected]>
Signed-off-by: Ingo Molnar <[email protected]>
|
|
Add function-entry relative line number specifying support to
perf-probe. This allows users to define probes by line number
from entry of the function.
e.g.
perf probe schedule:16
Signed-off-by: Masami Hiramatsu <[email protected]>
Cc: Steven Rostedt <[email protected]>
Cc: Jim Keniston <[email protected]>
Cc: Ananth N Mavinakayanahalli <[email protected]>
Cc: Christoph Hellwig <[email protected]>
Cc: Frank Ch. Eigler <[email protected]>
Cc: Frederic Weisbecker <[email protected]>
Cc: Jason Baron <[email protected]>
Cc: K.Prasad <[email protected]>
Cc: Peter Zijlstra <[email protected]>
Cc: Srikar Dronamraju <[email protected]>
LKML-Reference: <20091027204319.30545.30678.stgit@harusame>
Signed-off-by: Ingo Molnar <[email protected]>
|
|
This changes probe point syntax of perf-probe as below
<SRC>[:ABS_LN] [ARGS]
or
<FUNC>[+OFFS|%return][@SRC] [ARGS]
And event name and event group name are automatically
generated based on probe-symbol and offset as below.
perfprobes/SYMBOL_OFFSET[_NUM]
Where SYMBOL is the probing symbol and OFFSET is
the byte offset from the symbol.
Signed-off-by: Masami Hiramatsu <[email protected]>
Cc: Steven Rostedt <[email protected]>
Cc: Jim Keniston <[email protected]>
Cc: Ananth N Mavinakayanahalli <[email protected]>
Cc: Christoph Hellwig <[email protected]>
Cc: Frank Ch. Eigler <[email protected]>
Cc: Frederic Weisbecker <[email protected]>
Cc: Jason Baron <[email protected]>
Cc: K.Prasad <[email protected]>
Cc: Peter Zijlstra <[email protected]>
Cc: Srikar Dronamraju <[email protected]>
LKML-Reference: <20091027204310.30545.84984.stgit@harusame>
Signed-off-by: Ingo Molnar <[email protected]>
|
|
Change command-line option from -P to --add, and accepting
probes without --add too.
perf probe --add "probe-define"
or, just:
perf probe "probe-define"
Signed-off-by: Masami Hiramatsu <[email protected]>
Cc: Steven Rostedt <[email protected]>
Cc: Jim Keniston <[email protected]>
Cc: Ananth N Mavinakayanahalli <[email protected]>
Cc: Christoph Hellwig <[email protected]>
Cc: Frank Ch. Eigler <[email protected]>
Cc: Frederic Weisbecker <[email protected]>
Cc: Jason Baron <[email protected]>
Cc: K.Prasad <[email protected]>
Cc: Peter Zijlstra <[email protected]>
Cc: Srikar Dronamraju <[email protected]>
LKML-Reference: <20091027204301.30545.48600.stgit@harusame>
Signed-off-by: Ingo Molnar <[email protected]>
|
|
Exit searching after finding real (not-inlined) function,
because there should be no same symbol in that CU.
Signed-off-by: Masami Hiramatsu <[email protected]>
Cc: Steven Rostedt <[email protected]>
Cc: Jim Keniston <[email protected]>
Cc: Ananth N Mavinakayanahalli <[email protected]>
Cc: Christoph Hellwig <[email protected]>
Cc: Frank Ch. Eigler <[email protected]>
Cc: Frederic Weisbecker <[email protected]>
Cc: Jason Baron <[email protected]>
Cc: K.Prasad <[email protected]>
Cc: Peter Zijlstra <[email protected]>
Cc: Srikar Dronamraju <[email protected]>
LKML-Reference: <20091027204252.30545.19251.stgit@harusame>
Signed-off-by: Ingo Molnar <[email protected]>
|
|
So that we can have a quicker start on perf top and even
speedups in the other tools, as we can have maps with no hits,
so no need to load its symtabs.
Signed-off-by: Arnaldo Carvalho de Melo <[email protected]>
Cc: Frederic Weisbecker <[email protected]>
Cc: Peter Zijlstra <[email protected]>
Cc: Paul Mackerras <[email protected]>
Cc: Mike Galbraith <[email protected]>
LKML-Reference: <[email protected]>
Signed-off-by: Ingo Molnar <[email protected]>
|
|
Add two more software events that are common to many cpus.
Alignment faults: When a load or store is not aligned properly.
Emulation faults: When an instruction is emulated in software.
Both cause a very significant slowdown (100x or worse), so identifying and
fixing them is very important.
Signed-off-by: Anton Blanchard <[email protected]>
Signed-off-by: Paul Mackerras <[email protected]>
|
|
Writing to stdout is probably the expected behavior because the
user explicitly asked for a list.
Signed-off-by: Marti Raudsepp <[email protected]>
Cc: Peter Zijlstra <[email protected]>
Cc: Paul Mackerras <[email protected]>
LKML-Reference: <4ebb59420ef057972167.1256603585@localhost>
Signed-off-by: Ingo Molnar <[email protected]>
|
|
Previously no indication was given about what went wrong.
Signed-off-by: Marti Raudsepp <[email protected]>
Cc: Peter Zijlstra <[email protected]>
Cc: Paul Mackerras <[email protected]>
LKML-Reference: <03ec9ee96f17cef05424.1256603584@localhost>
Signed-off-by: Ingo Molnar <[email protected]>
|
|
Example:
Compiling the kernel with 'make -k 22 allyesconfig'
[root@emilia linux-2.6-tip]# perf top -r 90
------------------------------------------------------------------------------
PerfTop: 3669 irqs/sec kernel:59.9% [1000Hz cycles], (all, 8 CPUs)
------------------------------------------------------------------------------
samples pcnt function DSO
_______ _____ ________________________________ ________________
3062.00 6.5% clear_page_c [kernel]
2233.00 4.8% _int_malloc /lib64/libc-2.5.so
2100.00 4.5% yylex /home/acme/git/build/allyesconfig/scripts/genksyms/genksyms
2029.00 4.3% memset /lib64/libc-2.5.so
1224.00 2.6% page_fault [kernel]
1075.00 2.3% __GI_strlen /lib64/libc-2.5.so
863.00 1.8% sub_preempt_count [kernel]
822.00 1.8% __GI_memcpy /lib64/libc-2.5.so
810.00 1.7% __GI_vfprintf /lib64/libc-2.5.so
786.00 1.7% _int_free /lib64/libc-2.5.so
775.00 1.7% __GI_strcmp /lib64/libc-2.5.so
748.00 1.6% _spin_lock [kernel]
699.00 1.5% main /home/acme/git/build/allyesconfig/scripts/basic/fixdep
659.00 1.4% add_preempt_count [kernel]
649.00 1.4% yyparse /home/acme/git/build/allyesconfig/scripts/genksyms/genksyms
645.00 1.4% preempt_trace [kernel]
635.00 1.4% __GI___libc_free /lib64/libc-2.5.so
597.00 1.3% trace_preempt_on [kernel]
551.00 1.2% __GI___libc_malloc /lib64/libc-2.5.so
516.00 1.1% _spin_lock_irqsave [kernel]
481.00 1.0% copy_user_generic_string [kernel]
479.00 1.0% unmap_vmas [kernel]
429.00 0.9% _IO_file_xsputn_internal /lib64/libc-2.5.so
425.00 0.9% __GI_strncpy /lib64/libc-2.5.so
416.00 0.9% get_page_from_freelist [kernel]
414.00 0.9% malloc_consolidate /lib64/libc-2.5.so
406.00 0.9% get_parent_ip [kernel]
362.00 0.8% __rmqueue [kernel]
347.00 0.7% in_lock_functions [kernel]
316.00 0.7% __d_lookup [kernel]
[root@emilia linux-2.6-tip]#
More polishing is needed to print just DSO basename when not
--verbose, etc.
Supporting a 'comm' column requires some more reworking of 'perf
top' internals as we will need to use something like the hist
entries 'perf report' uses and will be done in another patch.
Signed-off-by: Arnaldo Carvalho de Melo <[email protected]>
Cc: Frederic Weisbecker <[email protected]>
Cc: Peter Zijlstra <[email protected]>
Cc: Paul Mackerras <[email protected]>
Cc: Mike Galbraith <[email protected]>
LKML-Reference: <[email protected]>
Signed-off-by: Ingo Molnar <[email protected]>
|
|
Because we will need it in 'perf top' to support userspace
symbols for existing threads.
Now we pass a callback that will receive the synthesized event
and then write it to the output file in 'perf record' and in the
upcoming patch for 'perf top' we will just immediatelly create
the in memory representation of threads and maps.
Signed-off-by: Arnaldo Carvalho de Melo <[email protected]>
Cc: Frederic Weisbecker <[email protected]>
Cc: Peter Zijlstra <[email protected]>
Cc: Paul Mackerras <[email protected]>
Cc: Mike Galbraith <[email protected]>
LKML-Reference: <[email protected]>
Signed-off-by: Ingo Molnar <[email protected]>
|
|
/proc/pid/tasks
Signed-off-by: Arnaldo Carvalho de Melo <[email protected]>
Cc: Frederic Weisbecker <[email protected]>
Cc: Peter Zijlstra <[email protected]>
Cc: Paul Mackerras <[email protected]>
Cc: Mike Galbraith <[email protected]>
LKML-Reference: <[email protected]>
Signed-off-by: Ingo Molnar <[email protected]>
|
|
For the perf tool the patch implements an Alpha specific section
in the perf.h header file.
Signed-off-by: Michael Cree <[email protected]>
Cc: Richard Henderson <[email protected]>
Cc: Ivan Kokshaysky <[email protected]>
Cc: Peter Zijlstra <[email protected]>
Cc: Paul Mackerras <[email protected]>
LKML-Reference: <[email protected]>
Signed-off-by: Ingo Molnar <[email protected]>
|
|
The present use of -Wcast-align causes the build to blow up on
SH due to generating a "cast increases required alignment of
target type" error on each invocation of list_for_each_entry().
It seems that this was previously reported and killed off in the
ia64 support patch, but nothing seems to have happened with
that. Presumably the same problem still remains there, too.
Signed-off-by: Paul Mundt <[email protected]>
LKML-Reference: <[email protected]>
Signed-off-by: Ingo Molnar <[email protected]>
|
|
The Makefile now automatically defines LIBELF_NO_MMAP when
libelf 0.8.x is detected. libelf 0.8 is still maintained and
some distributions such as Arch Linux use it instead of
elfutils.
Signed-off-by: Marti Raudsepp <[email protected]>
Cc: Lucas De Marchi <[email protected]>
Cc: Peter Zijlstra <[email protected]>
Cc: Paul Mackerras <[email protected]>
Cc: Frederic Weisbecker <[email protected]>
Cc: Arnaldo Carvalho de Melo <[email protected]>
Cc: Arjan van de Ven <[email protected]>
Cc: Mike Galbraith <[email protected]>
LKML-Reference: <1256400636.3007.16.camel@newn>
Signed-off-by: Ingo Molnar <[email protected]>
|
|
Use the new pr_{err,warning,debug,etc} printout methods, just
like in the kernel.
Signed-off-by: Arnaldo Carvalho de Melo <[email protected]>
Cc: Masami Hiramatsu <[email protected]>
Cc: Frederic Weisbecker <[email protected]>
Cc: Peter Zijlstra <[email protected]>
Cc: Paul Mackerras <[email protected]>
Cc: Mike Galbraith <[email protected]>
LKML-Reference: <[email protected]>
[ Split this patch out, to keep perf/probes separate. ]
Signed-off-by: Ingo Molnar <[email protected]>
|
|
Conflicts:
tools/perf/Makefile
Merge reason:
- fix the conflict
- pick up the pr_*() infrastructure to queue up dependent patch
Signed-off-by: Ingo Molnar <[email protected]>
|
|
We were using eprintf in some places, that looks at a global
'verbose' level, and at other places passing a 'v' parameter to
specify the verbosity level, unify it by introducing
pr_{err,warning,debug,etc}, just like in the kernel.
Signed-off-by: Arnaldo Carvalho de Melo <[email protected]>
Cc: Frederic Weisbecker <[email protected]>
Cc: Peter Zijlstra <[email protected]>
Cc: Paul Mackerras <[email protected]>
Cc: Mike Galbraith <[email protected]>
LKML-Reference: <[email protected]>
Signed-off-by: Ingo Molnar <[email protected]>
|
|
Wrapping the kernel headers is dangerous when it comes to arch
headers. Once we wrap asm/types.h, it will also replace the
glibc asm/types.h, not only the kernel one.
This results in build errors on some machines.
Drop this wrapper and do its work from linux/types.h wrapper,
also the glibc asm/types.h can already handle most of the type
definition it was doing (typedef __u64, __u32, etc...).
Todo: Check the others asm/*.h wrappers to prevent from other
conflicts.
Reported-by: Ingo Molnar <[email protected]>
Signed-off-by: Frederic Weisbecker <[email protected]>
Cc: Peter Zijlstra <[email protected]>
Cc: Arnaldo Carvalho de Melo <[email protected]>
Cc: Mike Galbraith <[email protected]>
Cc: Paul Mackerras <[email protected]>
Cc: Anton Blanchard <[email protected]>
LKML-Reference: <[email protected]>
Signed-off-by: Ingo Molnar <[email protected]>
|
|
Currently, the callchains are displayed using a constant left
margin. So depending on the current sort dimension
configuration, callchains may appear to be well attached to the
first sort dimension column field which is mostly the case,
except when the first dimension of sorting is done by comm,
because these are right aligned.
This patch binds the callchain to the first letter in the first
column, whatever type of column it is (dso, comm, symbol).
Before:
0.80% perf [k] __lock_acquire
__lock_acquire
lock_acquire
|
|--58.33%-- _spin_lock
| |
| |--28.57%-- inotify_should_send_event
| | fsnotify
| | __fsnotify_parent
After:
0.80% perf [k] __lock_acquire
__lock_acquire
lock_acquire
|
|--58.33%-- _spin_lock
| |
| |--28.57%-- inotify_should_send_event
| | fsnotify
| | __fsnotify_parent
Also, for clarity, we don't put anymore the callchain as is but:
- If we have a top level ancestor in the callchain, start it
with a first ascii hook.
Before:
0.80% perf [kernel] [k] __lock_acquire
__lock_acquire
lock_acquire
|
|--58.33%-- _spin_lock
| |
| |--28.57%-- inotify_should_send_event
| | fsnotify
[..] [..]
After:
0.80% perf [kernel] [k] __lock_acquire
|
--- __lock_acquire
lock_acquire
|
|--58.33%-- _spin_lock
| |
| |--28.57%-- inotify_should_send_event
| | fsnotify
[..] [..]
- Otherwise, if we have several top level ancestors, then
display these like we did before:
1.69% Xorg
|
|--21.21%-- vread_hpet
| 0x7fffd85b46fc
| 0x7fffd85b494d
| 0x7f4fafb4e54d
|
|--15.15%-- exaOffscreenAlloc
|
|--9.09%-- I830WaitLpRing
Signed-off-by: Frederic Weisbecker <[email protected]>
Cc: Peter Zijlstra <[email protected]>
Cc: Arnaldo Carvalho de Melo <[email protected]>
Cc: Mike Galbraith <[email protected]>
Cc: Paul Mackerras <[email protected]>
Cc: Anton Blanchard <[email protected]>
LKML-Reference: <[email protected]>
Signed-off-by: Ingo Molnar <[email protected]>
|
|
While recursively printing the branches of each callchains, we
forget to display the root. It is never printed.
Say we have:
symbol
f1
f2
|
-------- f3
| f4
|
---------f5
f6
Actually we never see that, instead it displays:
symbol
|
--------- f3
| f4
|
--------- f5
f6
However f1 is always the same than "symbol" and if we are
sorting by symbols first then "symbol", f1 and f2 will be well
aligned like in the above example, so displaying f1 looks
redundant here.
But if we are sorting by something else first (dso, comm,
etc...), displaying f1 doesn't look redundant but rather
necessary because the symbol is not well aligned anymore with
its callchain:
comm dso symbol
f1
f2
|
--------- [...]
And we want the callchain to be obvious.
So we fix the bug by printing the root branch, but we also
filter its first entry if we are sorting by symbols first.
Reported-by: Anton Blanchard <[email protected]>
Signed-off-by: Frederic Weisbecker <[email protected]>
Cc: Peter Zijlstra <[email protected]>
Cc: Arnaldo Carvalho de Melo <[email protected]>
Cc: Mike Galbraith <[email protected]>
Cc: Paul Mackerras <[email protected]>
LKML-Reference: <[email protected]>
Signed-off-by: Ingo Molnar <[email protected]>
|
|
The second argument in the strtok_r() function is not to be used
generically and can have different implementations. Currently
the function parsing of the perf trace code uses the second
argument to copy data from. This can crash the tool or just have
unpredictable results.
The correct solution is to use strsep() which has a defined
result.
I also added a check to see if the result was correct, and will
break out of the loop in case it fails to parse as expected.
Reported-by: Arnaldo Carvalho de Melo <[email protected]>
Signed-off-by: Steven Rostedt <[email protected]>
Cc: Peter Zijlstra <[email protected]>
Cc: Frederic Weisbecker <[email protected]>
LKML-Reference: <[email protected]>
Signed-off-by: Ingo Molnar <[email protected]>
|
|
When using gdb to debug perf, it is practically impossible to
use when perf is compiled with -O6. For developers, this patch
adds the DEBUG feature to the make command line so that a
developer can easily remove the optimization flag.
LKML-Reference: <1255590330.8392.446.camel@twins>
Signed-off-by: Steven Rostedt <[email protected]>
Cc: Peter Zijlstra <[email protected]>
Cc: Frederic Weisbecker <[email protected]>
Cc: Arnaldo Carvalho de Melo <[email protected]>
LKML-Reference: <[email protected]>
Signed-off-by: Ingo Molnar <[email protected]>
|
|
We need to use map->unmap_ip() here too to match section
relative symbol address to the absolute address needed to match
objdump -dS addresses.
Reported-by: Mike Galbraith <[email protected]>
Signed-off-by: Arnaldo Carvalho de Melo <[email protected]>
Cc: Frederic Weisbecker <[email protected]>
Cc: Peter Zijlstra <[email protected]>
Cc: Paul Mackerras <[email protected]>
LKML-Reference: <[email protected]>
Signed-off-by: Ingo Molnar <[email protected]>
|
|
If the user doesn't pass a symbol name to annotate, it will
annotate all the symbols that have hits, in order, just like
'perf report -s comm,dso,symbol'.
This is a natural followup patch to the one that uses
output_hists to find the symbols with hits.
The common case is to annotate the first few entries at the top
of a perf report, so lets type less characters.
Signed-off-by: Arnaldo Carvalho de Melo <[email protected]>
Cc: Frederic Weisbecker <[email protected]>
Cc: Peter Zijlstra <[email protected]>
Cc: Paul Mackerras <[email protected]>
Cc: Mike Galbraith <[email protected]>
LKML-Reference: <[email protected]>
Signed-off-by: Ingo Molnar <[email protected]>
|
|
We have this sym_priv_size mechanism for attaching private areas
to struct symbol entries but annotate wasn't using it, adding
private areas to struct symbol in addition to a ->priv pointer.
Scrap all that and use the sym_priv_size mechanism.
Signed-off-by: Arnaldo Carvalho de Melo <[email protected]>
Cc: Frederic Weisbecker <[email protected]>
Cc: Peter Zijlstra <[email protected]>
Cc: Paul Mackerras <[email protected]>
Cc: Mike Galbraith <[email protected]>
LKML-Reference: <[email protected]>
Signed-off-by: Ingo Molnar <[email protected]>
|