for bpftool which is
currently broken on debian (libbfd feature shows OFF, but we still
unconditionally link against -lbfd and it works).
Tested on debian with only -lbfd installed (without -liberty); I'd
appreciate if somebody on the other systems can test this new detection
method.
Signed-off
On Fri, Nov 16, 2018 at 5:34 AM Jiri Olsa wrote:
>
> On Thu, Nov 15, 2018 at 04:32:01PM -0800, Stanislav Fomichev wrote:
> > Current libbfd feature test unconditionally links against -liberty and -lz.
> > While it's required on some systems (e.g. opensuse), it's compl
send another patchset for that.
Not sure whether Oleg or Andrii plan to take a look, but LGTM:
Acked-by: Stanislav Fomichev
On 09/06, Florian Kauer wrote:
> The current xdp_devmap_attach test attaches a program
> that redirects to another program via devmap.
>
> It is, however, never executed, so do that to catch
> any bugs that might occur during execution.
>
> Also, execute the same for a veth pair so that we
> also
On 09/06, Stanislav Fomichev wrote:
> On 09/06, Florian Kauer wrote:
> > The current xdp_devmap_attach test attaches a program
> > that redirects to another program via devmap.
> >
> > It is, however, never executed, so do that to catch
> > any bugs
No, Thomas didn't yet push the fix.
27.12.2014, 02:34, "Stephen Boyd" :
> On 06/22/2014 07:46 AM, Thomas Gleixner wrote:
>>> + for (i = 0; i < HRTIMER_MAX_CLOCK_BASES; i++) {
>>> + ktime_t expires;
>> So this adds the third incarnation of finding the next expiring timer
>> to the code. Not rea
On 04/16, Jiri Olsa wrote:
> Moving file specific code in dso_cache__read function
> into separate file_read function. I'll add bpf specific
> code in following patches.
>
> Link: http://lkml.kernel.org/n/tip-7f7d717uzrqt5ka2xp29i...@git.kernel.org
> Signed-off-by: Jiri Olsa
> ---
> tools/perf/u
Thanks, the merge resolution looks good to me!
On Sun, Jan 24, 2021 at 4:12 PM Stephen Rothwell wrote:
>
> Hi all,
>
> Today's linux-next merge of the bpf-next tree got a conflict in:
>
> net/ipv4/tcp.c
>
> between commit:
>
> 7eeba1706eba ("tcp: Add receive timestamp support for receive zero
; Workload failed: No such file or directory
>
> This happens because of an off-by-one error while preparing the argv for
> cmd_record(): it attempts to execute the command 'status' and complains
> that it doesn't exist. Fix this error.
>
> Cc: Stanislav Fomichev
find
next expiry time for all clocks.
Signed-off-by: Stanislav Fomichev
---
kernel/hrtimer.c | 38 --
1 file changed, 24 insertions(+), 14 deletions(-)
diff --git a/kernel/hrtimer.c b/kernel/hrtimer.c
index 09094361dce5..63805f9f9531 100644
--- a/kernel/hrtim
find
next expiry time for all clocks.
Signed-off-by: Stanislav Fomichev
---
include/linux/hrtimer.h | 2 ++
kernel/hrtimer.c| 41 +++--
2 files changed, 33 insertions(+), 10 deletions(-)
diff --git a/include/linux/hrtimer.h b/include/linux/hrtim
> Passes my testing, but would appreciate a tested-by.
Tested-by: Stanislav Fomichev
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Pl
> This makes the new tool stop processing old files, can you try the patch
> attached instead?
I see two downsides to your approach:
1) with your patch I'm now required to run 'perf timechart record' and
'perf timechart' on the same machine (otherwise, on the 'perf timechart'
machine we may have wr
> Se above, if before this patch the format_field info was obtained from
> the perf.data file, why should it now get it from the local machine?
Yes, sorry for confusion, I didn't know that we can obtain trace events
format from the perf.data. I hence have no concerns regarding your
implementation :
continue working with old perf.data. Instead, we now
calculate event data offset dynamically using offset of first non-common
event field in the perf.data.
Signed-off-by: Stanislav Fomichev
---
tools/perf/builtin-timechart.c | 56 +++---
tools/perf/util/evsel.c
> Nacked-by: Namhyung Kim
Thanks for a review. I'll try to convert all binary structures
to perf_evsel__intval() and perf_evsel__strval() and will rebase on
Arnaldo's perf/core branch.
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majord...@
coded copy of fields layout from the kernel.
Signed-off-by: Stanislav Fomichev
---
tools/perf/builtin-timechart.c | 119 +++--
1 file changed, 42 insertions(+), 77 deletions(-)
diff --git a/tools/perf/builtin-timechart.c b/tools/perf/builtin-timechart.
> 1. introduce while loop: it's a behavioral change so that it can be a
> separate patch. But it seems not checking the process filter - in that
> case the loop is almost useless IMHO.
Could you please elaborate on 'not checking the process filter'?
This loop is for the case when process filter i
> I don't see how it does same thing. You mean it by skipping
> draw_process_bars() and draw_wakeups() on patch 1? It'd be better
> changelog explains the details.
Yes, it depends on the previous patch, I'll add more info to the
changelog.
> And I'd like to keep it as BOOLEAN option - it's simpl
> Hmm.. double negation is always confusing to me. :-/ How about making
> it a boolean option that sets "task_only" variable?
Ok.
> Also need to update the doc. And what if user gives -P and -T options
> at the same time?
Is something like this is enough (or you'd like some printout with error
d
> For consistency, wouldn't it be better renaming svg_sample() to
> svg_running()?
Makes sense. I'll resend the whole series with updates later.
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majord...@vger.kernel.org
More majordomo info at h
> It should be ARRAY_SIZE(old_power_args).
Thanks :-)
> Do we really need to separate the option and usage for record? AFAICS
> it does exactly same thing..
Yes, we do. Some options which are defined for timechart are undefined
for record (like -i, -o, etc), so we want to get an error in case of
If we don't want either power or task events we may use -T or -P
with the `perf timechart record` command to filter out events while
recording to keep perf.data small.
Signed-off-by: Stanislav Fomichev
---
tools/perf/Documentation/perf-timechart.txt | 15 -
tools/perf/builtin-timech
Don't use special flag to indicate power-only mode,
just set proc_num to 0.
Signed-off-by: Stanislav Fomichev
---
tools/perf/builtin-timechart.c | 12
1 file changed, 8 insertions(+), 4 deletions(-)
diff --git a/tools/perf/builtin-timechart.c b/tools/perf/builtin-timechart.c
ted out by Namhyung Kim
Stanislav Fomichev (7):
perf timechart: always try to print at least 15 tasks
perf timechart: add option to limit number of tasks
perf timechart: use proc_num to implement --power-only
perf timechart: add support for displaying only tasks related data
perf timech
Add -g flag to `perf timechart record` which saves callchain info in
the perf.data.
When generating SVG, add backtrace information to the figure details,
so now it's possible to see which code path woke up the task and why some
task went to sleep.
Signed-off-by: Stanislav Fomichev
---
Always try to print at least 15 tasks no matter how long they run.
Signed-off-by: Stanislav Fomichev
---
tools/perf/builtin-timechart.c | 12 +++-
1 file changed, 7 insertions(+), 5 deletions(-)
diff --git a/tools/perf/builtin-timechart.c b/tools/perf/builtin-timechart.c
index
In order to make SVG smaller and faster to browse add possibility to
switch off power related information with -T switch.
Signed-off-by: Stanislav Fomichev
---
tools/perf/Documentation/perf-timechart.txt | 3 +++
tools/perf/builtin-timechart.c | 11 ++-
2 files changed, 13
Add -n option to specify min. number of tasks to print.
Signed-off-by: Stanislav Fomichev
---
tools/perf/Documentation/perf-timechart.txt | 4
tools/perf/builtin-timechart.c | 5 -
2 files changed, 8 insertions(+), 1 deletion(-)
diff --git a/tools/perf/Documentation/perf
tails.
Signed-off-by: Stanislav Fomichev
---
tools/perf/builtin-timechart.c | 6 ++---
tools/perf/util/svghelper.c| 56 --
tools/perf/util/svghelper.h| 5 ++--
3 files changed, 60 insertions(+), 7 deletions(-)
diff --git a/tools/perf/bu
ils
(hence miss required libbfd version) and I still want an option to build perf
which works on any machine regardless of binutils version.
Signed-off-by: Stanislav Fomichev
---
tools/perf/config/Makefile | 8 +++-
1 file changed, 3 insertions(+), 5 deletions(-)
diff --git a/tools/perf/config/M
Add PID to the figures of CPU usage timechart.
Signed-off-by: Stanislav Fomichev
---
tools/perf/builtin-timechart.c | 1 +
tools/perf/util/svghelper.c| 4 ++--
tools/perf/util/svghelper.h| 2 +-
3 files changed, 4 insertions(+), 3 deletions(-)
diff --git a/tools/perf/builtin
on to sort CPUs (on the CPU overview chart) topologically (draw
adjacent SMT threads next to each other)
- adds '-e' option to highlight long events on the CPU overview chart
Stanislav Fomichev (5):
perf timechart: add backtrace support to CPU info
perf timechart: print pid a
Add -t switch to sort CPUs topologically.
Signed-off-by: Stanislav Fomichev
---
tools/perf/Documentation/perf-timechart.txt | 3 +
tools/perf/builtin-timechart.c | 17 +++-
tools/perf/util/svghelper.c | 132 +++-
tools/perf/util
Add backtrace info to the CPU usage timechart.
Signed-off-by: Stanislav Fomichev
---
tools/perf/builtin-timechart.c | 10 --
tools/perf/util/svghelper.c| 4 +++-
tools/perf/util/svghelper.h| 2 +-
3 files changed, 12 insertions(+), 4 deletions(-)
diff --git a/tools/perf
Print all CPUs, even if there were no events (use perf header to get number
of CPUs).
This is required to support topology in the next patch.
Signed-off-by: Stanislav Fomichev
---
tools/perf/builtin-timechart.c | 32 ++--
1 file changed, 26 insertions(+), 6 deletions
Free previously allocated map.
Signed-off-by: Stanislav Fomichev
---
tools/perf/util/session.c | 2 ++
1 file changed, 2 insertions(+)
diff --git a/tools/perf/util/session.c b/tools/perf/util/session.c
index 4ce146bae552..ee717ee9ca3a 100644
--- a/tools/perf/util/session.c
+++ b/tools/perf
This option highlights tasks (using different color) that run more than given
duration or tasks with given name.
Signed-off-by: Stanislav Fomichev
---
tools/perf/Documentation/perf-timechart.txt | 4
tools/perf/builtin-timechart.c | 21 -
tools/perf/util
umentation.
Attached new patch below.
---
This option highlights tasks (using different color) that run more than given
duration or tasks with given name.
Signed-off-by: Stanislav Fomichev
---
tools/perf/Documentation/perf-timechart.txt | 14 ++
tools/perf/builtin-timechart.c |
t color) that run more than given
duration or tasks with given name.
Signed-off-by: Stanislav Fomichev
---
tools/perf/Documentation/perf-timechart.txt | 13 +
tools/perf/builtin-timechart.c | 21 -
tools/perf/util/svghelper.c
graph tracer we register
handler which reinitializes tracer depending on tracing_thresh.
Signed-off-by: Stanislav Fomichev
---
kernel/trace/trace.c | 65
kernel/trace/trace.h | 2 ++
kernel/trace/trace_functions_graph.c | 7 ++
Signed-off-by: Stanislav Fomichev
---
tools/perf/util/session.c | 2 ++
1 file changed, 2 insertions(+)
diff --git a/tools/perf/util/session.c b/tools/perf/util/session.c
index 7acc03e8f3b2..03815af30b16 100644
--- a/tools/perf/util/session.c
+++ b/tools/perf/util/session.c
@@ -1600,6 +1600,7
If we call perf timechart with -p 0 arguments, it means we don't
want any tasks related data. It works, but space for tasks data is
reserved in the generated SVG. Remove this unused empty space via
passing 0 as count to the open_svg.
Signed-off-by: Stanislav Fomichev
---
tools/perf/bu
ils
(hence miss required libbfd version) and I still want an option to build perf
which works on any machine regardless of binutils version.
Signed-off-by: Stanislav Fomichev
---
tools/perf/config/Makefile | 8 +++-
1 file changed, 3 insertions(+), 5 deletions(-)
diff --git a/tools/perf/config/M
without libbfd (because on 3.10 lts NO_DEMANGLE=1
does not link with libbfd and I'd like to preserve this behavior).
Stanislav Fomichev (3):
perf timechart: fix wrong SVG height
perf util: free cpu_map in perf_session__cpu_bitmap
perf tools: bring back old behavior when NO_DEMAG
Hi,
Any comment on the v2?
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
It will be reused by perf trace in the following commit.
Signed-off-by: Stanislav Fomichev
---
tools/perf/builtin-kvm.c | 88 +++
tools/perf/util/session.c | 85 +
tools/perf/util/session.h | 5 +++
3
Signed-off-by: Stanislav Fomichev
---
tools/perf/Documentation/perf-trace.txt | 7 ++
tools/perf/builtin-trace.c | 39 ++---
2 files changed, 33 insertions(+), 13 deletions(-)
diff --git a/tools/perf/Documentation/perf-trace.txt
b/tools/perf
cept, and I'd like to get some comments about
where and what I got wrong and what additional useful information I can
expose in the trace.
Stanislav Fomichev (5):
perf trace: add support for pagefault tracing
perf trace: add pagefault statistics
perf trace: add possibility to switch off s
Use session to cache and order events from all CPUs so we get
correct ordering of comms/mmaps and faults.
Signed-off-by: Stanislav Fomichev
---
tools/perf/builtin-trace.c | 117 ++---
1 file changed, 46 insertions(+), 71 deletions(-)
diff --git a/tools
Signed-off-by: Stanislav Fomichev
---
tools/perf/builtin-trace.c | 11 +++
1 file changed, 11 insertions(+)
diff --git a/tools/perf/builtin-trace.c b/tools/perf/builtin-trace.c
index 6c7ae048db59..535f0b504950 100644
--- a/tools/perf/builtin-trace.c
+++ b/tools/perf/builtin-trace.c
Signed-off-by: Stanislav Fomichev
---
tools/perf/Documentation/perf-trace.txt | 12 ++
tools/perf/builtin-trace.c | 198 ++--
2 files changed, 176 insertions(+), 34 deletions(-)
diff --git a/tools/perf/Documentation/perf-trace.txt
b/tools/perf
> I try, when possible, to not use short options that are used in
> 'strace', so what if we use 'F' here?
Agreed, will change.
> And:
>
>trace --pgfaults --pgfaults
>
> for major and min page faults looks ugly, what if we instead use --pf
> for both, and allow passing min or maj as args?
>
> Hmm.. it looks like an unrelated issue. I can see the bars in inkscape,
> but still see empty grey bars in firefox.
I'll try to have a look; maybe I'm writing some non-standard svg
instructions which fail on (probably) standard-compliant firefox and
work on more robust inkscape.
--
To unsubscrib
> So make it default to the most common case :-)
Ok, agreed, will make --pf=all|maj|min and will default it to maj.
> That is my expectation, yes, if I ask for the map where address N is, it
> should return just that, where have you found this bug?
>
> The thread__find_addr_map will end up callin
> > Hmm.. it looks like an unrelated issue. I can see the bars in inkscape,
> > but still see empty grey bars in firefox.
> I'll try to have a look; maybe I'm writing some non-standard svg
> instructions which fail on (probably) standard-compliant firefox and
> work on more robust inkscape.
Found
--io-skip-eagain - don't show EAGAIN errors
--io-min-time- make small io bursts visible
--io-merge-dist - merge adjacent events
Signed-off-by: Stanislav Fomichev
---
tools/perf/Documentation/perf-timechart.txt | 13 +
tools/perf/builtin-timechart.c
start_time fix into distinct commit
- separated --io-xyz options into distinct commit
- fixed 'error' shadowing
- initialized variables in declaration when possible
- removed assert usage
- fprintf -> pr_err
- removed wrongly placed __maybe_unused
Stanislav Fomichev (3):
p
We don't need to overwrite current task start_time on fork, so update it
only if it's zero.
Signed-off-by: Stanislav Fomichev
---
tools/perf/builtin-timechart.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/tools/perf/builtin-timechart.c b/tools/perf/builtin-t
In IO mode timechart shows any disk/network activity.
Signed-off-by: Stanislav Fomichev
---
tools/perf/Documentation/perf-timechart.txt | 25 +-
tools/perf/builtin-timechart.c | 613 +++-
tools/perf/util/svghelper.c | 98 -
tools/perf
> It (and others below) should be 'rec_argc -= 4' - i.e. I still see the
> segfault. :)
Damn, stupid me.
> Looks like this change should be applied to the exsting (for cpu and
> power events) boxes too. Care to send it as a separate fix?
Yes, will add it as a first patch to the series, so it can
--io-skip-eagain - don't show EAGAIN errors
--io-min-time- make small io bursts visible
--io-merge-dist - merge adjacent events
Signed-off-by: Stanislav Fomichev
---
tools/perf/Documentation/perf-timechart.txt | 13 +
tools/perf/builtin-timechart.c
We don't need to overwrite current task start_time on fork, so update it
only if it's zero.
Signed-off-by: Stanislav Fomichev
---
tools/perf/builtin-timechart.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/tools/perf/builtin-timechart.c b/tools/perf/builtin-t
Firefox doesn't correctly handle cases where we specify number in
quotes and have some padding around the number, like the following:
In this case, it doesn't draw the figure. This patch removes 'field width'
component from fprintf strings to fix it.
Signed-off-by: Stanislav
n possible
- removed assert usage
- fprintf -> pr_err
- removed wrongly placed __maybe_unused
Stanislav Fomichev (4):
perf timechart: fix rendering in Firefox
perf timechart: implement IO mode
perf timechart: conditionally update start_time on fork
perf timechart: add more options to IO
In IO mode timechart shows any disk/network activity.
Signed-off-by: Stanislav Fomichev
---
tools/perf/Documentation/perf-timechart.txt | 25 +-
tools/perf/builtin-timechart.c | 613 +++-
tools/perf/util/svghelper.c | 98 -
tools/perf
It will be reused by perf trace in the following commit.
Signed-off-by: Stanislav Fomichev
---
tools/perf/builtin-kvm.c | 88 +++
tools/perf/util/session.c | 85 +
tools/perf/util/session.h | 5 +++
3
'perf trace' can show summary of events using -S option. This commit
also reports number of major/minor pagefault events in this summary.
Signed-off-by: Stanislav Fomichev
---
tools/perf/builtin-trace.c | 11 +++
1 file changed, 11 insertions(+)
diff --git a/tools/perf/built
( 0.000 ms): curl/5937 majfault [0x7fa7261978b6] =>
/usr/lib/x86_64-linux-gnu/libkrb5.so.26.0.0+0x85288 (d.)
1862866.036 ( 0.000 ms): wget/8460 majfault [__clear_user+0x3f] => 0x659cb4 (?k)
Signed-off-by: Stanislav Fomichev
---
tools/perf/Documentation/perf-trace.txt | 12
tools/perf
It will be used by next pagefault tracing patches in the series.
Signed-off-by: Stanislav Fomichev
---
tools/perf/builtin-trace.c | 11 ---
1 file changed, 8 insertions(+), 3 deletions(-)
diff --git a/tools/perf/builtin-trace.c b/tools/perf/builtin-trace.c
index f954c26de231
Previous commit added live pagefault trace support, this one adds
record and replay support.
Signed-off-by: Stanislav Fomichev
---
tools/perf/builtin-trace.c | 63 +-
1 file changed, 45 insertions(+), 18 deletions(-)
diff --git a/tools/perf/builtin
Currently, we may either trace syscalls or syscalls+pagefaults. We'd
like to be able to trace *only* pagefaults and this commit implements
this feature.
Signed-off-by: Stanislav Fomichev
---
tools/perf/Documentation/perf-trace.txt | 7
tools/perf/builtin-trace.c
into additional patch
- separated record/replay into additional patch
- other fixes pointed out by Arnaldo Carvalho de Melo
Stanislav Fomichev (7):
perf trace: add perf_event parameter to tracepoint_handler
perf trace: add support for pagefault tracing
perf trace: add pagefaults record
Use session to cache and order events from all CPUs so we get
correct ordering of comms/mmaps and faults.
Signed-off-by: Stanislav Fomichev
---
tools/perf/builtin-trace.c | 117 ++---
1 file changed, 46 insertions(+), 71 deletions(-)
diff --git a/tools
> I.e. what we need from perf_session is just the ordered_samples bits,
> perhaps in its current form, perhaps rewritten, see (renewed) discussion
> involving David Ahern and Jiri Olsa.
Do you mean you want to pull (at some point) Jiri's patches and want to use
just event ordering part in perf trac
> Just to clarify here, those slides came from slides I made and in turn
> the whole idea about pagefaults tracing I got from the trace prototype
> that Thomas Gleixner implemented in his 'trace' utility, described
> here:
>
> Announcing a new utility: 'trace'
> http://lwn.net/Articles/415728
> > 1756272.905 ( 0.000 ms): curl/5937 majfault [0x7fa7261978b6] =>
> > /usr/lib/x86_64-linux-gnu/libkrb5.so.26.0.0+0x85288 (d.)
>
> Things like the above may be better expressed using dso@symbol+offset,
> even more when the symbol is not resolved within a DSO, due to
> insufficient expressivity
> Hey, haven't you seen the vfs_getname probe? Idea is to hook on where
> the relevant copy_from_user is done and insert that into the ring
> buffer, as we already do for mapping fd -> pathname.
I saw it but didn't actually try because it needs all the debugging
stuff enabled and in place.
> I.e.
Hi Steven,
Did you have some spare time to check why I'm getting triple fault with the
jump labels inside function tracers? Or will you just pull my another
patch without them?
On Thu, Jun 05, 2014 at 11:38:39AM +0400, Stanislav Fomichev wrote:
> > You mean inside the function trace
--io-skip-eagain - don't show EAGAIN errors
--io-min-time- make small io bursts visible
--io-merge-dist - merge adjacent events
Signed-off-by: Stanislav Fomichev
---
tools/perf/Documentation/perf-timechart.txt | 13 +
tools/perf/builtin-timechart.c
polls).
Signed-off-by: Stanislav Fomichev
---
tools/perf/Documentation/perf-timechart.txt | 25 +-
tools/perf/builtin-timechart.c | 613 +++-
tools/perf/util/svghelper.c | 98 -
tools/perf/util/svghelper.h | 6 +-
4 files ch
e fix into distinct commit
- separated --io-xyz options into distinct commit
- fixed 'error' shadowing
- initialized variables in declaration when possible
- removed assert usage
- fprintf -> pr_err
- removed wrongly placed __maybe_unused
Stanislav Fomichev (4):
perf time
Firefox doesn't correctly handle cases where we specify number in
quotes and have some padding around the number, like the following:
In this case, it doesn't draw the figure. This patch removes 'field width'
component from fprintf strings to fix it.
Signed-off-by: Stanislav
We don't need to overwrite current task start_time on fork, so update it
only if it's zero.
Signed-off-by: Stanislav Fomichev
---
tools/perf/builtin-timechart.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/tools/perf/builtin-timechart.c b/tools/perf/builtin-t
Currently, we may either trace syscalls or syscalls+pagefaults. We'd
like to be able to trace *only* pagefaults and this commit implements
this feature.
Signed-off-by: Stanislav Fomichev
---
tools/perf/Documentation/perf-trace.txt | 7
tools/perf/builtin-trace.c
Previous commit added live pagefault trace support, this one adds
record and replay support.
Signed-off-by: Stanislav Fomichev
---
tools/perf/builtin-trace.c | 63 +-
1 file changed, 45 insertions(+), 18 deletions(-)
diff --git a/tools/perf/builtin
'perf trace' can show summary of events using -S option. This commit
also reports number of major/minor pagefault events in this summary.
Signed-off-by: Stanislav Fomichev
---
tools/perf/builtin-trace.c | 11 +++
1 file changed, 11 insertions(+)
diff --git a/tools/perf/built
It will be used by next pagefault tracing patches in the series.
Signed-off-by: Stanislav Fomichev
---
tools/perf/builtin-trace.c | 11 ---
1 file changed, 8 insertions(+), 3 deletions(-)
diff --git a/tools/perf/builtin-trace.c b/tools/perf/builtin-trace.c
index f954c26de231
( 0.000 ms): curl/5937 majfault [0x7fa7261978b6] =>
/usr/lib/x86_64-linux-gnu/libkrb5.so.26.0.0@0x85288 (d.)
1862866.036 ( 0.000 ms): wget/8460 majfault [__clear_user+0x3f] => 0x659cb4 (?k)
Signed-off-by: Stanislav Fomichev
---
tools/perf/Documentation/perf-trace.txt | 37 ++
to
-pf=[all|min|maj])
- separated tracepoint_handler changes into additional patch
- separated record/replay into additional patch
- other fixes pointed out by Arnaldo Carvalho de Melo
Stanislav Fomichev (5):
perf trace: add perf_event parameter to tracepoint_handler
perf trace: add suppo
> Where is it?
> usleep (19155), 151 events, 84.4%, 0:68 faults, 0.000 msec
^^^
maj:min
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majord...@vger.kernel.o
On Fri, Jun 27, 2014 at 10:53:30AM -0300, Arnaldo Carvalho de Melo wrote:
> Em Fri, Jun 27, 2014 at 12:43:07PM +0400, Stanislav Fomichev escreveu:
> > > Where is it?
> > > usleep (19155), 151 events, 84.4%, 0:
We don't need to overwrite current task start_time on fork, so update it
only if it's zero.
Signed-off-by: Stanislav Fomichev
---
tools/perf/builtin-timechart.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/tools/perf/builtin-timechart.c b/tools/perf/builtin-t
t commit
- fixed 'error' shadowing
- initialized variables in declaration when possible
- removed assert usage
- fprintf -> pr_err
- removed wrongly placed __maybe_unused
Stanislav Fomichev (4):
perf timechart: fix rendering in Firefox
perf timechart: implement IO mode
Firefox doesn't correctly handle cases where we specify number in
quotes and have some padding around the number, like the following:
In this case, it doesn't draw the figure. This patch removes 'field width'
component from fprintf strings to fix it.
Signed-off-by: Stanislav
--io-skip-eagain - don't show EAGAIN errors
--io-min-time- make small io bursts visible
--io-merge-dist - merge adjacent events
Signed-off-by: Stanislav Fomichev
---
tools/perf/Documentation/perf-timechart.txt | 13 +
tools/perf/builtin-timechart.c
polls).
Signed-off-by: Stanislav Fomichev
---
tools/perf/Documentation/perf-timechart.txt | 25 +-
tools/perf/builtin-timechart.c | 620 +++-
tools/perf/util/svghelper.c | 98 -
tools/perf/util/svghelper.h | 6 +-
4 files ch
0.000 0.033 0.246 91.00%
arch_prctl 1 0.001 0.001 0.001 0.00%
Signed-off-by: Stanislav Fomichev
---
tools/perf/builtin-trace.c | 17 +
1 file changed, 17 insertions(+)
diff --git a/tools/perf/builtin-trace.c b/tools/perf/builtin-tra
graph tracer we register
handler which reinitializes tracer depending on tracing_thresh.
Signed-off-by: Stanislav Fomichev
---
kernel/trace/trace.c | 65
kernel/trace/trace.h | 2 ++
kernel/trace/trace_functions_graph.c
perf timechart: fix documentation
moves misplaced options into appropriate place
perf timechart: implement IO mode
adds io mode, instead of process time we record io syscalls and then
draw timechart of writes/reads/tx/rx/polls
Stanislav Fomichev (2):
perf timechart: fix documentation
1 - 100 of 325 matches
Mail list logo
