anges
* add missing bswap_32 on reading nr_groups (Jiri)
* remove perf_evlist__recalc_nr_groups() in favor of list_is_last (Jiri)
v1 -> v2:
* save group relation to header (Jiri)
[1] https://lkml.org/lkml/2012/7/24/81
Namhyung Kim (18):
perf evsel: Set leader evsel's ->leader to i
anges
* add missing bswap_32 on reading nr_groups (Jiri)
* remove perf_evlist__recalc_nr_groups() in favor of list_is_last (Jiri)
v1 -> v2:
* save group relation to header (Jiri)
[1] https://lkml.org/lkml/2012/7/24/81
Namhyung Kim (18):
perf evsel: Set leader evsel's ->leader to i
From: Namhyung Kim
Show group members' overhead also when showing the leader's if event
group is enabled. Use macro for defining hpp functions which looks
almost identical.
Cc: Jiri Olsa
Cc: Stephane Eranian
Signed-off-by: Namhyung Kim
---
tools/perf/ui/hist.c
From: Namhyung Kim
Show group members' overhead also when showing the leader's if event
group is enabled. Use macro for defining hpp functions which looks
almost identical.
Cc: Jiri Olsa
Cc: Stephane Eranian
Signed-off-by: Namhyung Kim
---
tools/perf/ui/browsers/his
From: Namhyung Kim
Since we have all necessary information in the leader events and
other members don't, bypass members. Member events will be shown
along with the leaders if event group is enabled.
Cc: Jiri Olsa
Cc: Stephane Eranian
Cc: Pekka Enberg
Acked-by: Jiri Olsa
Signed-o
From: Namhyung Kim
Show group members' overhead also when showing the leader's if event
group is enabled. Use macro for defining hpp functions which looks
almost identical.
Cc: Jiri Olsa
Cc: Stephane Eranian
Cc: Pekka Enberg
Signed-off-by: Namhyung Kim
---
tools/perf/ui/gtk/brow
From: Namhyung Kim
When using event group viewer, it's better to show the group
description rather than the leader information alone.
If a leader did not contain any member, it's a non-group event.
Cc: Jiri Olsa
Cc: Stephane Eranian
Cc: Pekka Enberg
Acked-by: Jiri Olsa
Sig
From: Namhyung Kim
Currently only non-leader members are set ->leader to the leader evsel
of the group and the leader has set NULL. Thus it requires special
casing for leader evsels. Set ->leader to itself will remove this.
Suggested-by: Arnaldo Carvalho de Melo
Cc: Jiri Olsa
Cc: St
From: Namhyung Kim
Add report.group config option for setting default value of event
group view. It affects the report output only if perf.data contains
event group info.
A user can write .perfconfig file like below to enable group view by
default:
$ cat ~/.perfconfig
[report]
group
From: Namhyung Kim
Add --group option to enable event grouping. When enabled, all the
group members information will be shown together with the leader.
Cc: Jiri Olsa
Cc: Stephane Eranian
Acked-by: Jiri Olsa
Signed-off-by: Namhyung Kim
---
tools/perf/Documentation/perf-report.txt | 3
From: Namhyung Kim
Convert perf_evsel__is_group_member to perf_evsel__is_group_leader.
This is because the most usecases are using negative form to check
whether the given evsel is a leader or not and it's IMHO somewhat
ambiguous - leader also *is* a member of the group.
Signed-off-by: Nam
From: Namhyung Kim
Add a few of group-related field in struct perf_{evlist,evsel} so that
the group information in a evlist can be known easily. It only counts
groups which have more than 1 members since leader-only groups are
treated as non-group events.
Cc: Jiri Olsa
Cc: Stephane Eranian
From: Namhyung Kim
When event group is enabled, sorting hist entries on periods for
output should consider groups members' period also. To do that, build
period table using link/pair information and compare the table.
Cc: Jiri Olsa
Cc: Stephane Eranian
Signed-off-by: Namhyung Kim
---
From: Namhyung Kim
Now the event grouping viewing requires linking all member hists in a
group to the leader's. Thus hists__output_resort should be called
after linking all events in evlist.
Cc: Jiri Olsa
Cc: Stephane Eranian
Signed-off-by: Namhyung Kim
---
tools/perf/builtin-report.c
From: Namhyung Kim
The event_group field is for enabling event group view on perf report
and other commands. It requires collapsing hist entries since every
member in a group needs to be linked before final output resorting.
Cc: Jiri Olsa
Cc: Stephane Eranian
Signed-off-by: Namhyung Kim
From: Namhyung Kim
The current hists__match/link functions are worked on the output tree
(hists->entries). However for event group view it needs to be
matched/linked before the final resort. Thus add new helpers for
doing it on the collapsed tree.
Cc: Jiri Olsa
Cc: Stephane Eranian
Sig
anges
* add missing bswap_32 on reading nr_groups (Jiri)
* remove perf_evlist__recalc_nr_groups() in favor of list_is_last (Jiri)
v1 -> v2:
* save group relation to header (Jiri)
[1] https://lkml.org/lkml/2012/7/24/81
Namhyung Kim (18):
perf evsel: Set leader evsel's ->leader to i
From: Namhyung Kim
When comparing entries for collapsing put the given entry first, and
then the iterated entry. This is for the sake of consistency and will
be required by the event group report.
Cc: Jiri Olsa
Cc: Stephane Eranian
Signed-off-by: Namhyung Kim
---
tools/perf/util/hist.c | 2
From: Namhyung Kim
Current hists__match/link() link a leader to its pair, so if multiple
pairs were linked, the leader will lose pointer to previous pairs
since it was overwritten. Fix it by making leader the list head.
Cc: Jiri Olsa
Cc: Stephane Eranian
Signed-off-by: Namhyung Kim
From: Namhyung Kim
Save group relationship information so that it can be restored when
perf report is running.
Cc: Jiri Olsa
Cc: Stephane Eranian
Acked-by: Jiri Olsa
Signed-off-by: Namhyung Kim
---
tools/perf/builtin-record.c | 3 +
tools/perf/util/header.c| 153
From: Namhyung Kim
Fix a misplaced underscore. In this case, 'hist_entry' is the name of
data structure and we usually put double underscores between data
structure and actual function name.
Signed-off-by: Namhyung Kim
---
tools/perf/util/hist.c | 4 ++--
tools/perf/util/sort.h
Hi Jiri,
On Wed, 28 Nov 2012 14:52:36 +0100, Jiri Olsa wrote:
> Adding perf_hpp__list list to register and contain all period
> related columns the command is interested in.
>
> This way we get rid of static array holding all possible
> columns and enable commands to register their own columns.
B
a
> Cc: Ingo Molnar
> Cc: Paul Mackerras
> Cc: Corey Ashford
> Cc: Frederic Weisbecker
> Cc: Namhyung Kim
> ---
> tools/perf/ui/hist.c | 8 ++--
> 1 file changed, 2 insertions(+), 6 deletions(-)
>
> diff --git a/tools/perf/ui/hist.c b/tools/perf/ui/hist.c
>
e Melo
> Cc: Peter Zijlstra
> Cc: Ingo Molnar
> Cc: Paul Mackerras
> Cc: Corey Ashford
> Cc: Frederic Weisbecker
Acked-by: Namhyung Kim
Thanks,
Namhyung
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to m
Hi,
On Thu, 29 Nov 2012 16:44:59 +0900, Namhyung Kim wrote:
> Hi Jiri,
>
> On Wed, 28 Nov 2012 14:52:36 +0100, Jiri Olsa wrote:
>> Adding perf_hpp__list list to register and contain all period
>> related columns the command is interested in.
>>
>> This way we ge
o Carvalho de Melo
> Cc: Peter Zijlstra
> Cc: Ingo Molnar
> Cc: Paul Mackerras
> Cc: Corey Ashford
> Cc: Frederic Weisbecker
Acked-by: Namhyung Kim
Thanks,
Namhyung
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to ma
o Carvalho de Melo
> Cc: Peter Zijlstra
> Cc: Ingo Molnar
> Cc: Paul Mackerras
> Cc: Corey Ashford
> Cc: Frederic Weisbecker
Acked-by: Namhyung Kim
Thanks,
Namhyung
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to ma
On Wed, 28 Nov 2012 14:52:43 +0100, Jiri Olsa wrote:
> Adding diff command the flexibility to specify multiple data
> files on input. If not input file is given the standard behaviour
> stands and diff inspects 'perf.data' and 'perf.data.old' files.
>
> Also changing the processing and output displ
On Wed, 28 Nov 2012 14:52:44 +0100, Jiri Olsa wrote:
> +COMPARISON
> +--
> +The comparison is governed by the baseline file. The baseline perf.data
> +file is iterated for samples. All other perf.data files specified on
> +the command line are searched for the baseline sample pair. If the p
On Wed, 28 Nov 2012 14:52:45 +0100, Jiri Olsa wrote:
> Now when diff command is separated from other standard outputs,
> we can use perf_hpp__init to initialize all standard columns.
>
> Moving PERF_HPP__OVERHEAD column init back to perf_hpp__init,
> and removing extra enable calls.
Why was this n
On Wed, 28 Nov 2012 14:52:49 +0100, Jiri Olsa wrote:
> Adding option 'o' to allow sorting based on the
> input file number.
[snip]
> hist_entry__cmp_compute(struct hist_entry *left, struct hist_entry *right,
> int c)
> {
> - int i;
> + struct hist_entry **pairs_left
2012-11-29 (목), 15:33 +0100, Jiri Olsa:
> On Thu, Nov 29, 2012 at 03:38:31PM +0900, Namhyung Kim wrote:
> > From: Namhyung Kim
>
> SNIP
>
> > --- a/tools/perf/util/evlist.h
> > +++ b/tools/perf/util/evlist.h
> > @@ -21,6 +21,7 @@ struct perf_evlis
2012-11-29 (목), 16:02 +0100, Jiri Olsa:
> On Thu, Nov 29, 2012 at 11:58:54PM +0900, Namhyung Kim wrote:
> > 2012-11-29 (목), 15:33 +0100, Jiri Olsa:
> > > On Thu, Nov 29, 2012 at 03:38:31PM +0900, Namhyung Kim wrote:
> > > > From: Namhyung Kim
> > >
> >
Hi Jiri,
On Thu, 29 Nov 2012 13:13:19 +0100, Jiri Olsa wrote:
> On Thu, Nov 29, 2012 at 08:55:47PM +0900, Namhyung Kim wrote:
>> On Wed, 28 Nov 2012 14:52:45 +0100, Jiri Olsa wrote:
>> > Now when diff command is separated from other standard outputs,
>> > we can use p
From: Namhyung Kim
When cpu/task clock events are initialized, their sampling frequencies
are converted to have a fixed value. However it missed to update the
hwc->last_period which was set to 1 for initial sampling frequency
calibration.
Because this hwc->last_period value is used as a
From: Namhyung Kim
This test case checks frequency conversion of hrtimer-based software
clock events (cpu-clock, task-clock) have valid (non-1) periods.
Cc: Jiri Olsa
Signed-off-by: Namhyung Kim
---
tools/perf/Makefile | 1 +
tools/perf/tests/builtin-test.c | 4 ++
tools/perf
From: Namhyung Kim
It's a per-cpu data structure but missed the __percpu annotation.
Cc: Tejun Heo
Signed-off-by: Namhyung Kim
---
include/linux/perf_event.h | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/include/linux/perf_event.h b/include/linux/perf_event.h
Hi Steve,
On Fri, 15 Mar 2013 23:24:21 -0400, Steven Rostedt wrote:
> From: "Steven Rostedt (Red Hat)"
>
> Update the README file in debugfs/tracing to something more useful.
> What's currently in the file is very old and what it shows doesn't
> have much use. Heck, it tells you how to mount debu
Hi Borislav,
On Fri, 15 Mar 2013 14:06:26 +0100, Borislav Petkov wrote:
> From: Borislav Petkov
>
> Yeah,
>
> here's a refresh of the persistent events deal, accessing those is much
> cleaner now. Here's how:
>
> So kernel code initializes and enables the event at its convenience
> (during boot,
On Fri, 15 Mar 2013 14:06:28 +0100, Borislav Petkov wrote:
> Add a barebones implementation for registering persistent events with
> perf. For that, we don't destroy the buffers when they're unmapped;
> also, we map them read-only so that multiple agents can access them.
>
> Also, we allocate the e
From: Namhyung Kim
It's a per-cpu data structure but missed the __percpu annotation.
Cc: Tejun Heo
Cc: Li Zefan
Signed-off-by: Namhyung Kim
---
kernel/events/core.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/kernel/events/core.c b/kernel/events/core.c
Hi Arnaldo,
On Mon, 18 Mar 2013 11:42:43 -0300, Arnaldo Carvalho de Melo wrote:
> Em Mon, Mar 18, 2013 at 11:41:47AM +0900, Namhyung Kim escreveu:
>> +static int __test__sw_clock_freq(enum perf_sw_ids clock_id)
>> +{
>> +int i, err = -1;
>> +volatile int tmp = 0
From: Namhyung Kim
It's for calculating whole trace data size during reading. However
relation functions are called only in this file, no need to
conditionalize it with tricky +1 offset and rename the variable to
more meaningful name like trace_data_size.
Cc: Steven Rostedt
Cc: Fre
From: Namhyung Kim
Check return value of write and fail if error.
Cc: Steven Rostedt
Cc: Frederic Weisbecker
Signed-off-by: Namhyung Kim
---
tools/perf/util/trace-event-info.c | 169 +
tools/perf/util/trace-event.h | 2 +-
2 files changed, 116
From: Namhyung Kim
Convert them to pr_debug() and propagate error code.
Cc: Steven Rostedt
Cc: Frederic Weisbecker
Signed-off-by: Namhyung Kim
---
tools/perf/util/trace-event-read.c | 44 +-
1 file changed, 29 insertions(+), 15 deletions(-)
diff --git a
From: Namhyung Kim
Check return value of malloc and fail if NULL.
Cc: Steven Rostedt
Cc: Frederic Weisbecker
Signed-off-by: Namhyung Kim
---
tools/perf/util/trace-event-info.c | 48 --
1 file changed, 30 insertions(+), 18 deletions(-)
diff --git a/tools
From: Namhyung Kim
Rename it to do_read and original do_read to __do_read, and check
their return value.
Cc: Steven Rostedt
Cc: Frederic Weisbecker
Signed-off-by: Namhyung Kim
---
tools/perf/util/trace-event-read.c | 80 +++---
1 file changed, 57 insertions
From: Namhyung Kim
Check return value of malloc() and fail if error. Now read_string()
can return NULL also check its return value and bail out.
Cc: Steven Rostedt
Cc: Frederic Weisbecker
Signed-off-by: Namhyung Kim
---
tools/perf/util/trace-event-read.c | 90
From: Namhyung Kim
So that it can be used by other places.
Cc: Steven Rostedt
Cc: Frederic Weisbecker
Signed-off-by: Namhyung Kim
---
tools/perf/util/trace-event-info.c | 25 ++---
1 file changed, 22 insertions(+), 3 deletions(-)
diff --git a/tools/perf/util/trace-event
From: Namhyung Kim
Now remove all remaining die() calls and convert them to check return
value and propagate it.
Cc: Steven Rostedt
Cc: Frederic Weisbecker
Signed-off-by: Namhyung Kim
---
tools/perf/util/trace-event-info.c | 114 +++--
1 file changed, 72
From: Namhyung Kim
If pevent allocation in read_trace_init() fails, trace_report() will
return -1 and *ppevent is set to NULL. Its callers should check this
case and handle it properly.
This is also a preparation for the removal of *die() calls.
Cc: Steven Rostedt
Cc: Frederic Weisbecker
sting more!
Thanks,
Namhyung
Namhyung Kim (9):
perf util: Let get_tracing_file() can return NULL
perf util: Get rid of malloc_or_die() in trace-event-info.c
perf util: Get rid of write_or_die() from trace-event-info.c
perf util: Get rid of die() calls from trace-event-info.c
perf util: Hand
From: Namhyung Kim
We define it in the Makefile so no need to duplicate it.
Signed-off-by: Namhyung Kim
---
tools/perf/builtin-record.c| 2 --
tools/perf/util/header.c | 2 --
tools/perf/util/session.c | 2 --
tools/perf/util/trace-event-read.c | 2 --
tools/perf
Hi Steve,
On Tue, 19 Mar 2013 09:54:21 -0400, Steven Rostedt wrote:
> On Tue, 2013-03-19 at 17:53 +0900, Namhyung Kim wrote:
>> From: Namhyung Kim
>>
>> So that it can be used by other places.
>>
>> Cc: Steven Rostedt
>> Cc: Frederic Wei
On Tue, 19 Mar 2013 16:04:03 +0100, Peter Zijlstra wrote:
> On Tue, 2013-03-19 at 10:59 -0400, Steven Rostedt wrote:
>> On Tue, 2013-03-19 at 15:49 +0100, Peter Zijlstra wrote:
>> > On Tue, 2013-03-19 at 10:35 -0400, Steven Rostedt wrote:
>> > > What about:
>> > > int err = 0;
>> > >
>> >
On Tue, 19 Mar 2013 10:47:53 -0400, Steven Rostedt wrote:
> On Tue, 2013-03-19 at 17:53 +0900, Namhyung Kim wrote:
>
>> @@ -3100,7 +3105,7 @@ int perf_event__process_tracing_data(union perf_event
>> *event,
>> }
>> }
>>
>> -if
On Tue, 19 Mar 2013 10:50:02 -0400, Steven Rostedt wrote:
> On Tue, 2013-03-19 at 17:53 +0900, Namhyung Kim wrote:
>> free(version);
>> @@ -331,11 +354,12 @@ ssize_t trace_report(int fd, struct pevent **ppevent,
>> bool __repipe)
>>
>&
On Tue, 19 Mar 2013 10:54:27 -0400, Steven Rostedt wrote:
> On Tue, 2013-03-19 at 17:53 +0900, Namhyung Kim wrote:
>> From: Namhyung Kim
>>
>> Rename it to do_read and original do_read to __do_read, and check
>> their return value.
>>
>> Cc: Steven Rostedt
On Tue, 19 Mar 2013 10:55:28 -0400, Steven Rostedt wrote:
> On Tue, 2013-03-19 at 17:53 +0900, Namhyung Kim wrote:
>> From: Namhyung Kim
>>
>> Convert them to pr_debug() and propagate error code.
>
> Shouldn't they be pr_err(). I mean, if the old code would kill
On Tue, 19 Mar 2013 21:55:02 -0400, Steven Rostedt wrote:
> On Wed, 2013-03-20 at 10:14 +0900, Namhyung Kim wrote:
>> On Tue, 19 Mar 2013 10:50:02 -0400, Steven Rostedt wrote:
>> > On Tue, 2013-03-19 at 17:53 +0900, Namhyung Kim wrote:
>> >> free(version);
>&
From: Namhyung Kim
So that it can be used by other places.
Cc: Steven Rostedt
Cc: Frederic Weisbecker
Signed-off-by: Namhyung Kim
---
tools/perf/util/trace-event-info.c | 25 ++---
1 file changed, 22 insertions(+), 3 deletions(-)
diff --git a/tools/perf/util/trace-event
From: Namhyung Kim
Check return value of malloc() and fail if error. Now read_string()
can return NULL also check its return value and bail out.
Cc: Steven Rostedt
Cc: Frederic Weisbecker
Signed-off-by: Namhyung Kim
---
tools/perf/util/trace-event-read.c | 100
From: Namhyung Kim
Now remove all remaining die() calls and convert them to check return
value and propagate it.
Cc: Steven Rostedt
Cc: Frederic Weisbecker
Signed-off-by: Namhyung Kim
---
tools/perf/util/trace-event-info.c | 114 +++--
1 file changed, 72
From: Namhyung Kim
Check return value of write and fail if error.
Cc: Steven Rostedt
Cc: Frederic Weisbecker
Signed-off-by: Namhyung Kim
---
tools/perf/util/trace-event-info.c | 180 ++---
tools/perf/util/trace-event.h | 2 +-
2 files changed, 127
From: Namhyung Kim
If pevent allocation in read_trace_init() fails, trace_report() will
return -1 and *ppevent is set to NULL. Its callers should check this
case and handle it properly.
This is also a preparation for the removal of *die() calls.
Cc: Steven Rostedt
Cc: Frederic Weisbecker
From: Namhyung Kim
It's for calculating whole trace data size during reading. However
relation functions are called only in this file, no need to
conditionalize it with tricky +1 offset and rename the variable to
more meaningful name like trace_data_size.
Cc: Steven Rostedt
Cc: Fre
From: Namhyung Kim
Rename it to do_read and original do_read to __do_read, and check
their return value.
Cc: Steven Rostedt
Cc: Frederic Weisbecker
Signed-off-by: Namhyung Kim
---
tools/perf/util/trace-event-read.c | 80 +++---
1 file changed, 57 insertions
From: Namhyung Kim
Convert them to pr_debug() and propagate error code.
Cc: Steven Rostedt
Cc: Frederic Weisbecker
Signed-off-by: Namhyung Kim
---
tools/perf/util/trace-event-read.c | 44 +-
1 file changed, 29 insertions(+), 15 deletions(-)
diff --git a
Hi,
I updated this series with previous feedbacks from Steven, Peter and
Arnaldo. Thanks for all your comments.
* v2 changes
- fix a grammar in an error message
- change some pr_err() -> pr_debug()
- use kernel error checking style consistently
Namhyung Kim (9):
perf util:
From: Namhyung Kim
Check return value of malloc and fail if NULL.
Cc: Steven Rostedt
Cc: Frederic Weisbecker
Signed-off-by: Namhyung Kim
---
tools/perf/util/trace-event-info.c | 48 --
1 file changed, 30 insertions(+), 18 deletions(-)
diff --git a/tools
Hi Jiri,
On Mon, 4 Feb 2013 13:32:55 +0100, Jiri Olsa wrote:
> Currently we don't display group members' values for raw
> columns like 'Samples' and 'Period' when in group report
> mode.
Oops, it was lost during the rework. Thanks for fixing this.
Thanks,
Namhyung
--
To unsubscribe from this l
fix:
> http://lkml.org/lkml/2013/2/4/122
>
> and is available also at:
> git://git.kernel.org/pub/scm/linux/kernel/git/jolsa/linux.git
> perf/group6
For the whole series:
Acked-by: Namhyung Kim
Thanks,
Namhyung
--
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/
From: Namhyung Kim
Current _sort__sym_cmp() function is used for comparing symbols
between two hist entries on symbol, symbol_from and symbol_to sort
keys. Those functions pass addresses of symbols but it's meaningless
since it gets over-written inside of the _sort__sym_cmp function to a
From: Namhyung Kim
When setup_sorting() is called, 'str' is passed to strtok_r() but it's
not checked to have a valid pointer. As strtok_r() accepts NULL
pointer on a first argument and use the third argument in that case,
it can cause a trouble since our third argument, tmp, is
From: Namhyung Kim
Currently the setup_sorting() is called for parsing sort keys and
exits if it failed to add the sort key. As it's included in libperf
it'd be better returning an error code rather than exiting application
inside of the library.
Suggested-by: Arnaldo Carvalho d
Hi Tejun and Hillf,
On Wed, 6 Feb 2013 10:47:49 -0800, Tejun Heo wrote:
> On Wed, Feb 06, 2013 at 08:38:43PM +0800, Hillf Danton wrote:
>> As handled by the kernel thread, work is dequeued first for further actions.
>
> Ditto as the previous patch.
>
>> Signed-off-by: Hillf Danton
>> ---
>>
>> -
Hi Stephane,
On Wed, 6 Feb 2013 15:46:00 +0100, Stephane Eranian wrote:
> This patch adds per-processor socket count aggregation
> for system-wide mode measurements. This is a useful
> mode to detect imbalance between sockets for uniform
> workloads.
>
> To enable this mode, use --aggr-socket in
Hi Oleg,
On Wed, 6 Feb 2013 20:42:18 +0100, Oleg Nesterov wrote:
> OK, builtin-record.c:record sets .target->uses_mmap == T, this is correct,
> we are going to use perf_mmap().
>
> But why do we need it? It is for perf_evlist__create_maps() to ensure we
> do not use cpu_map__dummy_new().
>
> But.
Hi Sasha,
On Wed, 6 Feb 2013 17:11:34 -0500, Sasha Levin wrote:
> This is a simple wrapper to make using liblockdep on existing applications
> much easier.
>
> After running 'make && make install', it becomes quite simple to test things
> with liblockdep. For example, to try it on perf:
>
>
On Wed, 6 Feb 2013 17:11:34 -0500, Sasha Levin wrote:
> diff --git a/tools/lib/lockdep/lockdep b/tools/lib/lockdep/lockdep
> new file mode 100755
> index 000..616bf9a
> --- /dev/null
> +++ b/tools/lib/lockdep/lockdep
> @@ -0,0 +1,3 @@
> +#! /bin/bash
> +
> +LD_PRELOAD=liblockdep.so "$@"
Just
* squash the last (add --gtk option) patch into the patch 1 (Arnaldo)
* make --gtk option does not fallback to stdio for a piped stdout (Borislav)
Namhyung Kim (7):
perf ui/gtk: Implement basic GTK2 annotation browser
perf gtk/annotate: Support multiple event annotation
perf gtk/annotate: Show
From: Namhyung Kim
In order to differentiate source lines from asm line, print them with
gray color. To do this, it needs to be escaped since sometimes it
contains "<" and/or ">" characters so that it should not be considered
as a markup tags. Use glib's g_mar
From: Namhyung Kim
Don't need to setup a browser window if annotate cannot work.
Cc: Pekka Enberg
Signed-off-by: Namhyung Kim
---
tools/perf/ui/gtk/annotate.c | 16
1 file changed, 8 insertions(+), 8 deletions(-)
diff --git a/tools/perf/ui/gtk/annotate.c b/tools/perf/u
From: Namhyung Kim
Add --skip-missing option for skipping symbols that cannot be used for
annotation. It's the case of kernel symbols that user doesn't have a
vmlinux image file.
Signed-off-by: Namhyung Kim
---
tools/perf/builtin-annotate.c | 17 +++--
1 file c
From: Namhyung Kim
Show multiple annotation result for each evsel. Each result
represents the most frquently sampled symbol/function for the evsel
and it will be shown in a tab window.
For this add a reference to main container (notebook) to the pgctx.
At the first call to annotate browser
From: Namhyung Kim
When perf annotate runs with no vmlinux file it cannot annotate kernel
symbols because the kallsyms only provides symbol addresses. So it
recommends to run perf buildid-cache to install proper vmlinux image.
But running perf buildid-cache -av vmlinux as the message gives me
Basic implementation of perf annotate on GTK2. Currently only
shows first symbol. Add a new --gtk option to use it.
Cc: Borislav Petkov
Cc: Pekka Enberg
Signed-off-by: Namhyung Kim
---
tools/perf/Documentation/perf-annotate.txt | 4 +-
tools/perf/Makefile| 1
From: Namhyung Kim
When adding vmlinux file to build-id cache, it'd be fail since
kallsyms dso with a same build-id was already added by perf record.
So one needs to remove the kallsyms first to add vmlinux into the
cache. Add --update option for doing it at once.
Signed-off-by: Namhyun
Hi Arnaldo,
On Mon, 4 Feb 2013 22:12:47 -0300, Arnaldo Carvalho de Melo wrote:
> Em Thu, Feb 07, 2013 at 06:02:14PM +0900, Namhyung Kim escreveu:
>> From: Namhyung Kim
>>
>> Add --skip-missing option for skipping symbols that cannot be used for
>> annotation. It
; valid when updating work result.
The only user of stop_one_cpu_nowait() is the scheduler active balancer
and active_load_balance_cpu_stop() always returns 0. But I think this
change makes its sense and the code evolves, so:
Reviewed-by: Namhyung Kim
Thanks,
Namhyung
>
> Signed-of
On Fri, 8 Feb 2013 11:42:43 +0800, Hillf Danton wrote:
> As checked with BUG_ON in the case of CPU_UP_PREPARE, we have to dequeue
> work first for further actions, then stopper reaches sane and clear state.
Reviewed-by: Namhyung Kim
Thanks,
Namhyung
--
To unsubscribe from this list: se
Hi,
On Mon, 11 Feb 2013 10:54:58 +0100, Ingo Molnar wrote:
> * Clark Williams wrote:
>
>> I figured that was coming. :)
>
> ;-)
>
>> I'll look at it again and see about pulling the
>> autogroup/cgroup stuff into it's own header. After that it's
>> probably going to require some serious changes.
Hi Alex,
On Sat, 30 Mar 2013 22:34:57 +0800, Alex Shi wrote:
> Since the rt task priority is higher than fair tasks, cfs_rq utilization
> is just the left of rt utilization.
>
> When there are some cfs tasks in queue, the potential utilization may
> be yielded, so mulitiplying cfs task number to g
Hi Alex,
On Sat, 30 Mar 2013 22:35:00 +0800, Alex Shi wrote:
> Sleeping task has no utiliation, when they were bursty waked up, the
> zero utilization make scheduler out of balance, like aim7 benchmark.
>
> rq->avg_idle is 'to used to accommodate bursty loads in a dirt simple
> dirt cheap manner'
Hi Jiri,
On Tue, 2 Apr 2013 10:40:14 +0200, Jiri Olsa wrote:
> On Tue, Apr 02, 2013 at 11:15:23AM +0900, Namhyung Kim wrote:
>> On Mon, 1 Apr 2013 22:40:23 +0200, Jiri Olsa wrote:
>> > On Mon, Apr 01, 2013 at 08:35:22PM +0900, Namhyung Kim wrote:
>> >> From: Namhy
okay
to do it though. I also cleaned up sort__has_sym and sort key eliding
code in this version.
You can get it from 'perf/sort-v4' branch on my tree at:
git://git.kernel.org/pub/scm/linux/kernel/namhyung/linux-perf.git
Any comments are welcome, thanks.
Namhyung
Namhyung Kim (10):
From: Namhyung Kim
The same code was duplicate to places, factor them out to common
sort__setup_elide().
Signed-off-by: Namhyung Kim
---
tools/perf/builtin-diff.c | 4 +---
tools/perf/builtin-report.c | 20 +---
tools/perf/builtin-top.c| 4 +---
tools/perf/util/sort.c
From: Namhyung Kim
New addr sort key provides a way to sort the entries by the symbol
addresses. It can be helpful to figure out symbol resolution problem
when a dso cannot do it properly as well as finding hotpath in a dso
and/or a function.
Suggested-by: Arnaldo Carvalho de Melo
Buglink
From: Namhyung Kim
The TUI hist browser had a similar variable has_symbols for the same
purpose. Let's get rid of the duplication.
Signed-off-by: Namhyung Kim
---
tools/perf/ui/browsers/hists.c | 9 ++---
1 file changed, 2 insertions(+), 7 deletions(-)
diff --git a/tools/pe
From: Namhyung Kim
perf top had a similar variable sort_has_symbols for the same purpose.
Signed-off-by: Namhyung Kim
---
tools/perf/builtin-top.c | 12 +++-
tools/perf/util/top.h| 1 -
2 files changed, 3 insertions(+), 10 deletions(-)
diff --git a/tools/perf/builtin-top.c b
1 - 100 of 6630 matches
Mail list logo