We often use the symbol__annotate2() to annotate a specified symbol.
While annotating may take some time, so in order to avoid annotating
the same symbol repeatedly, the patch creates a new flag to indicate
the symbol has been annotated.
Signed-off-by: Jin Yao
---
tools/perf/util/annotate.c | 1
iri's patch to support stdio mode
2. Add a new patch "perf annotate: Create a annotate2 flag
in struct symbol" which records if the symbol has been
annotated yet.
3. Minor update such as adding { } for multiline code in 'if'
condition.
Jin Yao (4):
p
On 11/29/2018 9:27 PM, Ingo Molnar wrote:
* Jin Yao wrote:
Add explanations for new columns "IPC" and "IPC coverage" in perf
documentation.
Signed-off-by: Jin Yao
---
tools/perf/Documentation/perf-report.txt | 8
1 file changed, 8 insertions(+)
dif
On 11/29/2018 7:33 PM, Jiri Olsa wrote:
On Thu, Nov 29, 2018 at 11:39:34PM +0800, Jin Yao wrote:
Add supporting of displaying the average IPC and IPC coverage
percentage per function.
For example,
$ perf record -b ...
$ perf report -s symbol or
perf report -s symbol --stdio
Overhead
On 11/29/2018 6:13 PM, Jiri Olsa wrote:
On Thu, Nov 29, 2018 at 02:24:27PM +0800, Jin, Yao wrote:
On 11/28/2018 6:18 PM, Jiri Olsa wrote:
On Wed, Nov 28, 2018 at 11:17:57AM +0100, Jiri Olsa wrote:
On Wed, Nov 28, 2018 at 11:14:55PM +0800, Jin Yao wrote:
Add supporting of displaying the
Add explanations for new columns "IPC" and "IPC coverage" in perf
documentation.
v5:
---
Update the description according to Ingo's comments.
Signed-off-by: Jin Yao
Reviewed-by: Jiri Olsa
Reviewed-by: Ingo Molnar
---
tools/perf/Documentation/perf-report.txt | 8 ++
We often use the symbol__annotate2() to annotate a specified symbol.
While annotating may take some time, so in order to avoid annotating
the same symbol repeatedly, the patch creates a new flag to indicate
the symbol has been annotated.
Signed-off-by: Jin Yao
Reviewed-by: Jiri Olsa
---
tools
nabled when "symbol"
is specified.
v2:
---
Merge in Jiri's patch to support stdio mode
Signed-off-by: Jin Yao
Reviewed-by: Jiri Olsa
---
tools/perf/builtin-report.c | 26 ---
tools/perf/util/hist.h | 1 +
tools/perf/util/sort.c | 61
__lll_unlock_wake_private
add$0x80,%rsp
21.56 2.908a: movslq 0xc(%rsp),%rax
2.90 add$0x18,%rsp
9.03 2.90 1 ← retq
It shows for this symbol the average IPC is 2.30 and the IPC coverage
is 54.8%.
Signed-off-by: Jin Yao
Reviewed-by: Jiri
2 flag
in struct symbol" which records if the symbol has been
annotated yet.
3. Minor update such as adding { } for multiline code in 'if'
condition.
Jin Yao (4):
perf annotate: Compute average IPC and IPC coverage per symbol
perf annotate: Create a annotate2 flag
On 6/15/2018 7:36 PM, Mark Rutland wrote:
On Fri, Jun 15, 2018 at 06:03:22PM +0800, Jin Yao wrote:
When doing sampling, for example:
perf record -e cycles:u ...
On workloads that do a lot of kernel entry/exits we see kernel
samples, even though :u is specified. This is due to skid existing
On 6/18/2018 6:45 PM, Peter Zijlstra wrote:
On Mon, Jun 18, 2018 at 02:55:32PM +0800, Jin, Yao wrote:
Thanks for providing the patch. I understand this approach.
In my opinion, the skid window is from counter overflow to interrupt
delivered. While if the skid window is too *big* (e.g. user
e)
transaction (from sample->transaction, raw value)
Signed-off-by: Jin Yao
---
.../util/scripting-engines/trace-event-python.c| 212 -
1 file changed, 211 insertions(+), 1 deletion(-)
diff --git a/tools/perf/util/scripting-engines/trace-event-python.c
b/tools/perf/util
On 5/31/2018 1:43 AM, Andi Kleen wrote:
On Wed, May 30, 2018 at 10:20:45PM +0800, Jin Yao wrote:
When doing pmu sampling and then running a script with
perf script -s script.py, the process_event function gets
dictionary with some fields from the perf ring buffer
(like ip, sym, callchain etc
Add the brief introduction of fields to perf-script-python.txt.
It may help python script developer easily finding what fields
are supported.
Signed-off-by: Jin Yao
---
tools/perf/Documentation/perf-script-python.txt | 26 +
1 file changed, 26 insertions(+)
diff --git
,iregs,uregs,and etc.
This patch adds these fields for perf script python.
Jin Yao (3):
perf script python: Move dsoname code to a new function
perf script python: Add more PMU fields
perf script python: Add fields introduction to Documentation
tools/perf/Documentation/perf-script-python.txt
This patch creates a new function get_dsoname() and move the
code which gets the dsoname string to this function.
That's because in next patch, when we process LBR data, we will
also need get_dsoname() to return dsoname for branch from/to.
Signed-off-by: Jin Yao
---
.../util/scripting-en
(converted string)
key : datasrc
key : datasrc_decode (decoded string)
key : iregs
key : uregs
key : weight
key : transaction
v2:
---
Add new fields for dso.
Use PyBool_FromLong() for mispred/predicted/in_tx/abort
Signed-off-by: Jin Yao
---
.../util/scripting-engines/trace-event-python.c
Introduce a new sysctl /sys/devices/cpu/perf_allow_sample_leakage, which
turns on/off dropping leaked kernel samples.
Signed-off-by: Jin Yao
---
tools/perf/Documentation/perf-record.txt | 14 ++
1 file changed, 14 insertions(+)
diff --git a/tools/perf/Documentation/perf-record.txt
ked kernel samples.
1 - don't drop the leaked kernel samples.
For rr it can write 1 to /sys/devices/cpu/perf_allow_sample_leakage to
keep original system behavior.
Jin Yao (2):
perf/core: Use sysctl to turn on/off dropping leaked kernel samples
perf Documentation: Intro
linux] [k] apic_timer_interrupt
0.00% div libc-2.23.so [.] intel_check_word
0.00% div ld-2.23.so[.] brk
0.00% div [kernel.vmlinux] [k] page_fault
0.00% div ld-2.23.so[.] _start
We can see the kernel symbols apic_timer_interrupt and
if (event->attr.exclude_kernel && !user_mode(regs))
return false;
}
But rr needs to add PERF_SAMPLE_ALLOW_LEAKAGE to sample_type since by
default the bit is not set.
3. Sysctl is a more flexible way. It provides us with an option when we
want to see if skid is existing, we can use sysctl to turn on that.
Thanks
Jin Yao
On 6/15/2018 1:59 PM, Stephane Eranian wrote:
On Thu, Jun 14, 2018 at 7:10 PM Jin Yao wrote:
When doing sampling, for example:
perf record -e cycles:u ...
On workloads that do a lot of kernel entry/exits we see kernel
samples, even though :u is specified. This is due to skid existing
On 6/15/2018 3:45 PM, Peter Zijlstra wrote:
On Fri, Jun 15, 2018 at 06:03:21PM +0800, Jin Yao wrote:
On workloads that do a lot of kernel entry/exits we see kernel
samples, even though :u is specified. This is due to skid existing.
This might be a security issue because it can leak kernel
On 6/15/2018 4:12 PM, Peter Zijlstra wrote:
On Fri, Jun 15, 2018 at 04:01:45PM +0800, Jin, Yao wrote:
Bring more overhead to kernel if we zero the bits considering the number of
leaked samples may be not too small?
Keeping the samples at least allows you to know how many samples
happened
has potential conflicts with some
applications.
Is this proposal reasonable?
Thanks
Jin Yao
On 6/16/2018 1:34 AM, Robert O'Callahan wrote:
On Fri, Jun 15, 2018 at 10:16 AM, Kyle Huey wrote:
If you want a sysctl for your own reasons that's fine. But we don't
want a sysctl. We wa
On 6/16/2018 8:56 AM, Kyle Huey wrote:
On Fri, Jun 15, 2018 at 5:50 PM, Jin, Yao wrote:
Hi All,
This patch raised many questions, I was prepared. :)
I'd like to try another proposal that it adds a special flag in the returned
perf_sample_data to indicate the perf binary that this s
Hi,
Any comments for this fix?
Thanks
Jin Yao
On 5/10/2018 9:59 PM, Jin Yao wrote:
See example,
perf record -e cycles,branches ./div
perf annotate main --stdio or
perf annotate main --stdio2 or
perf annotate main
The "perf annotate" should show both cycles and branches on the
left
│↓ jne29
│↓ jmp43
│1.1020: cmpxchg %esi,__abort_msg@@GLIBC_PRIVATE+0x8a0
8.93 │1.10 1 ↓ je 43
Signed-off-by: Jin Yao
---
tools/perf/ui/browsers/annotate.c | 8
tools/perf/util/annotate.c| 37 +++
p43
│1.1020: cmpxchg %esi,__abort_msg@@GLIBC_PRIVATE+0x8a0
8.93 │1.10 1 ↓ je 43
Jin Yao (2):
perf annotate: Record the max/min cycles
perf annotate: Create hotkey 'c' to show max/min cycles
tools/perf/ui/browsers/annotate.c | 8 ++
tools/pe
cycles but also the max and min cycles.
This patch records the max and min cycles.
Signed-off-by: Jin Yao
---
tools/perf/util/annotate.c | 14 +-
tools/perf/util/annotate.h | 4
2 files changed, 17 insertions(+), 1 deletion(-)
diff --git a/tools/perf/util/annotate.c b/tools/perf
On 5/18/2018 4:06 AM, Arnaldo Carvalho de Melo wrote:
Em Thu, May 17, 2018 at 10:58:38PM +0800, Jin Yao escreveu:
In perf annotate view, a new hotkey 'c' is created for
showing the max/min cycles.
I just changed everything from "max/min" to "min/max", as i
On 5/18/2018 4:27 AM, Jiri Olsa wrote:
On Thu, May 17, 2018 at 01:27:50PM -0300, Arnaldo Carvalho de Melo wrote:
Em Thu, May 10, 2018 at 09:59:22PM +0800, Jin Yao escreveu:
See example,
perf record -e cycles,branches ./div
perf annotate main --stdio or
perf annotate main --stdio2 or
perf
--
Post the patch 'perf annotate: Support multiple events without group'
Jin Yao (3):
perf evlist: Create a new function perf_evlist_forced_leader
perf report: Use perf_evlist_forced_leader to support '--group'
perf annotate: Support '--group' option
For non-explicit group, perf report supports a option '--group'
which can enable group output. We also need to support perf annotate
with the same '--group'.
Create a new function perf_evlist_forced_leader which contains
common code to force setting the group leader.
Si
Since we have created a new function perf_evlist_forced_leader,
so now remove the old code and use perf_evlist_forced_leader
instead.
Signed-off-by: Jin Yao
---
tools/perf/builtin-report.c | 8 ++--
1 file changed, 2 insertions(+), 6 deletions(-)
diff --git a/tools/perf/builtin-report.c b
or%edi,%edi
:srand(s_randseed);
0.00 : 4004b3: mov$0x77359400,%ebx
:
:return i;
:}
Signed-off-by: Jin Yao
---
tools/perf/builtin-annotate.c | 7 +++
1 file changed, 7 insertions(+)
di
On 5/18/2018 3:04 PM, Jiri Olsa wrote:
On Fri, May 18, 2018 at 08:57:39PM +0800, Jin Yao wrote:
Since we have created a new function perf_evlist_forced_leader,
so now remove the old code and use perf_evlist_forced_leader
instead.
Signed-off-by: Jin Yao
---
tools/perf/builtin-report.c | 8
patch 'perf annotate: Support multiple events without group'
Jin Yao (3):
perf evlist: Create a new function perf_evlist_forced_leader
perf report: Use perf_evlist_forced_leader to support '--group'
perf annotate: Support '--group' option
tools/perf/builtin
or%edi,%edi
:srand(s_randseed);
0.00 : 4004b3: mov$0x77359400,%ebx
:
:return i;
:}
Signed-off-by: Jin Yao
---
tools/perf/builtin-annotate.c | 7 +++
1 file changed, 7 insertions(+)
di
Since we have created a new function perf_evlist_forced_leader,
so now remove the old code and use perf_evlist_forced_leader
instead.
Signed-off-by: Jin Yao
---
tools/perf/builtin-report.c | 13 ++---
1 file changed, 2 insertions(+), 11 deletions(-)
diff --git a/tools/perf/builtin
For non-explicit group, perf report supports a option '--group'
which can enable group output. We also need to support perf annotate
with the same '--group'.
Create a new function perf_evlist_forced_leader which contains
common code to force setting the group leader.
Si
On 5/18/2018 10:18 PM, Arnaldo Carvalho de Melo wrote:
Em Sat, May 19, 2018 at 12:00:32AM +0800, Jin Yao escreveu:
For non-explicit group, perf report supports a option '--group'
which can enable group output. We also need to support perf annotate
with the same '--group&
On 1/4/2019 8:54 PM, Jiri Olsa wrote:
On Fri, Jan 04, 2019 at 10:28:17AM +0800, Jin Yao wrote:
Following test shows the stat keeps running even if no longer
task to monitor (mgen exits at ~5s).
perf stat -e cycles -p `pgrep mgen` -I1000 -- sleep 10
time counts unit
On 1/6/2019 9:25 PM, Jiri Olsa wrote:
On Sat, Jan 05, 2019 at 11:16:40AM +0800, Jin, Yao wrote:
On 1/4/2019 8:54 PM, Jiri Olsa wrote:
On Fri, Jan 04, 2019 at 10:28:17AM +0800, Jin Yao wrote:
Following test shows the stat keeps running even if no longer
task to monitor (mgen exits at ~5s
ntial race condition window, that is the
child_pid set with -1 and kill(-1, SIGTERM) may happen.
Jiri provides a simple method which stores the child_pid to
a local variable and check the local variable before kill().
At least, we can avoid the kill(-1, SIGTERM) case.
Signed-off-by: Jin Yao
---
because
it's waiting for the child process "sleep 1000" to be end. So perf
stat doesn't return even 5s passed.
This patch lets the perf stat return when the specified child process
is end (in this case, specified child process is "sleep 5").
Signed-off-by: Jin Yao
---
cycles
3.000609910 1,297,047,663 cycles
4.000807545 1,297,215,816 cycles
5.001001578 1,297,208,032 cycles
6.001390345582,343,659 cycles
sleep: Terminated
Now the stat exits immediately when the monitored tasks ends.
Signed-off-by: Jin Yao
les:1 iter:1 avg_cycles:23)
avg_cycles:23 is the average cycles of this iteration.
Fixes: c4ee06251d42 ("perf report: Calculate the average cycles of iterations")
Signed-off-by: Jin Yao
---
tools/perf/util/callchain.c | 32
tools/perf/util/callchain.
s:400)
0.58% pick_next_task_fair (cycles:47)
+ 0.56% i915_request_retire (cycles:2)
+ 0.52% do_sys_poll (cycles:4)
Signed-off-by: Jin Yao
---
tools/perf/builtin-top.c | 3 +++
1 file changed, 3 insertions(+)
diff --git a/tools/perf/builtin-top.c b/tools/perf/builtin-top.c
Hi,
Any comments for this patch?
Thanks
Jin Yao
On 10/31/2018 7:06 PM, Jin Yao wrote:
Perf report has supported the displaying of LBR stats
(such as cycles, predicted%) in callchain entry.
For example,
perf report --branch-history --stdio
--1.01%--intel_idle mwait.h:29
intel_idle
Hi Arnaldo, Hi Jiri,
Kan's patch series for AlderLake perf core support has been upstreamed, so the interface will not be
changed any more.
For this perf tool series (v4), do you have any comments?
Thanks
Jin Yao
On 4/16/2021 10:04 PM, Jin Yao wrote:
AlderLake uses a hybrid archite
For HSW, a patch "[PATCH] ASoC: Intel: Remove invalid kfree of devm
allocated data" to fix this issue has been applied.
But yes, we also need a similar patch for Baytrail.
Thanks
Jin Yao
On 2015/4/16 22:08, Jarkko Nikula wrote:
> On 04/16/2015 04:46 PM, weiyj...@163.com wrote:
llow.
Thanks
JIn Yao
On 2015/4/13 3:06, kbuild test robot wrote:
> sound/soc/intel/common/sst-ipc.c:287:2-7: WARNING: NULL check before freeing
> functions like kfree, debugfs_remove, debugfs_remove_recursive or
> usb_free_urb is not needed. Maybe consider reorganizing relevant code to
zlib: [ on ]
lzma: [ on ]
get_cpuid: [ on ]
bpf: [ on ]
[ on ]: library is compiled-in
[ OFF ]: library is disabled in make configuration
OR library is not installed in build environment
Jin Yao (3):
perf config: A
uot;OFF".
v2:
---
1. Use IS_BUILTIN macro to replace #ifdef/#endif block.
2. Print color for on/OFF.
Signed-off-by: Jin Yao
---
tools/perf/builtin-version.c | 68
1 file changed, 68 insertions(+)
diff --git a/tools/perf/builtin-version.c b/tool
R library is not installed in build environment
v2:
---
Use a global variable version_verbose to count the number of 'v'.
Signed-off-by: Jin Yao
---
tools/perf/perf.c | 22 ++
tools/perf/perf.h | 1 +
2 files changed, 19 insertions(+), 4 deletions(-)
diff --git
From: Jiri Olsa
Adding IS_BUILTIN macro and its dependencies into
tools world.
It's taken from kernel's include/linux/kconfig.h,
which can't be taken completely due to its kconfig
dependencies.
Signed-off-by: Jiri Olsa
---
tools/include/tools/config.h | 34 ++
On 3/28/2018 9:10 PM, Jiri Olsa wrote:
On Wed, Mar 28, 2018 at 10:57:07PM +0800, Jin Yao wrote:
SNIP
+
+static void library_status(void)
+{
+ STATUS(HAVE_DWARF_SUPPORT, dwarf);
+ STATUS(HAVE_DWARF_GETLOCATIONS, dwarf_getlocations);
+ STATUS(HAVE_GLIBC_SUPPORT, glibc
On 3/28/2018 9:22 PM, Jiri Olsa wrote:
On Wed, Mar 28, 2018 at 10:57:08PM +0800, Jin Yao wrote:
We keep having bug reports that when users build perf on their own,
but they don't install some needed libraries such as libelf,
libbfd/libibery.
The perf can build, but it is missing impo
On 3/28/2018 10:19 PM, Jiri Olsa wrote:
On Wed, Mar 28, 2018 at 10:00:32PM +0800, Jin, Yao wrote:
On 3/28/2018 9:22 PM, Jiri Olsa wrote:
On Wed, Mar 28, 2018 at 10:57:08PM +0800, Jin Yao wrote:
We keep having bug reports that when users build perf on their own,
but they don't install
On 3/28/2018 11:46 PM, Ingo Molnar wrote:
* Jin Yao wrote:
$ ./perf -vv or ./perf -version --build-options
perf version 4.13.rc5.gcb1183
My suggestion was to add the 'version' subcommand like Git has, not a "-version"
option:
$ git version
git version 2.14.1
T
--
Use a global variable version_verbose to record the number of 'v'.
Signed-off-by: Jin Yao
---
tools/perf/perf.c | 6 ++
tools/perf/perf.h | 1 +
2 files changed, 7 insertions(+)
diff --git a/tools/perf/perf.c b/tools/perf/perf.c
index 1b3fc8e..1659029 100644
--- a/tools/perf/pe
From: Jiri Olsa
Adding IS_BUILTIN macro and its dependencies into
tools world.
It's taken from kernel's include/linux/kconfig.h,
which can't be taken completely due to its kconfig
dependencies.
Signed-off-by: Jiri Olsa
---
tools/include/tools/config.h | 34 ++
arse_options'.
v2:
---
1. Use IS_BUILTIN macro to replace #ifdef/#endif block.
2. Print color for on/OFF.
Signed-off-by: Jin Yao
---
tools/perf/builtin-version.c | 81 +++-
1 file changed, 80 insertions(+), 1 deletion(-)
diff --git a/tools/perf/builtin-
Since a new option '--build-options' is created for 'perf version',
so we need to document it.
Signed-off-by: Jin Yao
---
tools/perf/Documentation/perf-version.txt | 24
1 file changed, 24 insertions(+)
create mode 100644 tools/perf/Documentation/pe
[ on ]
bpf: [ on ]
Jin Yao (4):
perf config: Add some new -DHAVE_XXX to CFLAGS
perf version: Print the compiled-in status of libraries
perf: Support perf -vv
perf Documentation: Create perf-version.txt
Jiri Olsa (1):
tools include: Add config.h header file
tools/in
On 3/29/2018 3:20 PM, Ingo Molnar wrote:
* Jin Yao wrote:
For example:
$ ./perf version --build-options
or
./perf --version --build-options
or
./perf -v --build-options
or
./perf -vv
Thank you, the UI looks perfect to me!
perf version 4.13.rc5.g4c1df1
On 6/7/2018 2:36 AM, Arnaldo Carvalho de Melo wrote:
Em Fri, Jun 01, 2018 at 05:01:02PM +0800, Jin Yao escreveu:
+static int get_symoff(struct symbol *sym, struct addr_location *al,
+ bool print_off, char *bf, int size)
+{
+ unsigned long offset;
+
+ if
efore parsing the time string.
Signed-off-by: Jin Yao
---
tools/perf/builtin-report.c | 38 +++--
tools/perf/builtin-script.c | 39 +++--
tools/perf/util/time-utils.c | 51 +++-
tools/perf/util/time-utils.h
On 3/1/2019 1:57 AM, Arnaldo Carvalho de Melo wrote:
Em Thu, Feb 28, 2019 at 10:30:09PM +0800, Jin Yao escreveu:
Jiri points out that we don't need any time checking and time string
parsing if the --time option is not set. That makes sense.
This patch refactors the time range parsing
eport --time 0%-10%,30%-40%
perf script --time 0%-10%,30%-40%
Select the time slices from timestamp 3971 to 3973
perf report --time 3971,3973
perf script --time 3971,3973
v2:
---
Add some examples in patch description.
No functional changes.
Signed-off-by: Jin Yao
---
tools/perf/bu
On 4/16/2019 3:26 AM, Arnaldo Carvalho de Melo wrote:
Em Mon, Apr 15, 2019 at 08:09:09PM +0530, Ravi Bangoria escreveu:
On 4/12/19 7:29 PM, Jin Yao wrote:
diff --git a/tools/perf/Documentation/perf-stat.txt
b/tools/perf/Documentation/perf-stat.txt
index 39c05f8..1e312c2 100644
--- a
For example,
perf record -b ...
perf annotate
The before/after output should be no change.
Signed-off-by: Jin Yao
---
tools/perf/builtin-annotate.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/tools/perf/builtin-annotate.c b/tools/perf/builtin-annotate.c
index 67
Any comments for this patch or any issue is found? In testing, I just
feel it accelerates the perf annotate data processing.
Thanks
Jin Yao
On 3/13/2019 4:22 AM, Jin Yao wrote:
The hist__account_cycles is executed when the hist_iter__branch_callback
is called. But it looks it's not nece
On 3/14/2019 8:04 PM, Jiri Olsa wrote:
On Wed, Mar 13, 2019 at 04:22:54AM +0800, Jin Yao wrote:
The hist__account_cycles is executed when the hist_iter__branch_callback
is called. But it looks it's not necessary. In hist__account_cycles, it
already walks on all branch entries.
This
On 3/14/2019 10:17 PM, Jin, Yao wrote:
On 3/14/2019 8:04 PM, Jiri Olsa wrote:
On Wed, Mar 13, 2019 at 04:22:54AM +0800, Jin Yao wrote:
The hist__account_cycles is executed when the hist_iter__branch_callback
is called. But it looks it's not necessary. In hist__account_cycles, it
al
nge.
v2:
---
1. Cover the similar perf report
2. Remove the checking code "ch->reset >= ch->num / 2"
Signed-off-by: Jin Yao
---
tools/perf/builtin-annotate.c | 4 ++--
tools/perf/builtin-report.c | 9 +++--
tools/perf/util/annotate.c| 2 +-
3 files changed, 6 i
ed to support this per-core counting on a event level.
This can be implemented in only the user tool, no kernel support needed.
Signed-off-by: Jin Yao
---
tools/perf/util/evsel.c| 2 ++
tools/perf/util/evsel.h| 3 +++
tools/perf/util/parse-events.c | 27 +++
an already support per-core counting with --per-core, but it's
often useful to do this together with other metrics that are collected
per CPU (per hardware thread). So this patch series supports this
per-core counting on a event level.
Jin Yao (3):
perf: Add a coresum event qualifier
per
It's a simple test which just checks if parser works.
Signed-off-by: Jin Yao
---
tools/perf/tests/parse-events.c | 10 +-
1 file changed, 9 insertions(+), 1 deletion(-)
diff --git a/tools/perf/tests/parse-events.c b/tools/perf/tests/parse-events.c
index 4a69c07..8d741d4 100644
r-core and per-CPU in
one perf stat command-line. From the output, we can see:
S0-C0 = CPU0 + CPU4
S0-C1 = CPU1 + CPU5
S0-C2 = CPU2 + CPU6
S0-C3 = CPU3 + CPU7
So the result is expected (tiny difference is ignored).
Signed-off-by: Jin Yao
---
tools/perf/Documentation/perf-stat.txt | 4
On 3/15/2019 8:54 PM, Jiri Olsa wrote:
On Fri, Mar 15, 2019 at 09:46:01PM +0800, Jin Yao wrote:
The hist__account_cycles is executed when the hist_iter__branch_callback
is called. But it looks it's not necessary. In hist__account_cycles, it
already walks on all branch entries.
This
o change.
v3:
---
Fix the crash in stdio mode.
Like previous code, it needs the checking of ui__has_annotation()
before hist__account_cycles()
v2:
---
1. Cover the similar perf report
2. Remove the checking code "ch->reset >= ch->num / 2"
Signed-off-by: Jin Yao
---
too
Hi Arnaldo,
Can this patch be accepted?
Thanks
Jin Yao
On 3/19/2019 6:20 PM, Jiri Olsa wrote:
On Tue, Mar 19, 2019 at 04:56:52PM +0800, Jin Yao wrote:
The 'percore' event qualifier which sums up the event counts for both
hardware threads in a core. For example,
perf stat -e c
On 4/10/2019 8:54 PM, Arnaldo Carvalho de Melo wrote:
Em Wed, Apr 10, 2019 at 09:36:41AM -0300, Arnaldo Carvalho de Melo escreveu:
Em Tue, Mar 19, 2019 at 04:56:53PM +0800, Jin Yao escreveu:
Add a 'percore' event qualifier, like cpu/event=0,umask=0x3,percore=1/,
that sums up
On 2/25/2019 9:38 PM, Jiri Olsa wrote:
On Mon, Feb 25, 2019 at 09:42:42PM +0800, Jin Yao wrote:
SNIP
COMPARISON
--
The comparison is governed by the baseline file. The baseline perf.data
diff --git a/tools/perf/builtin-diff.c b/tools/perf/builtin-diff.c
index 751e197..ddc41e7
On 2/25/2019 10:19 PM, Jiri Olsa wrote:
On Mon, Feb 25, 2019 at 10:12:08PM +0800, Jin, Yao wrote:
On 2/25/2019 9:38 PM, Jiri Olsa wrote:
On Mon, Feb 25, 2019 at 09:42:42PM +0800, Jin Yao wrote:
SNIP
COMPARISON
--
The comparison is governed by the baseline file. The
For better filtering support for perf diff, it would be useful to
add --pid and --tid filter options.
For example,
perf diff --tid 13965
It only diff the samples for thread 13965.
v2:
---
No functional change.
Signed-off-by: Jin Yao
---
tools/perf/Documentation/perf-diff.txt | 6
e, put the time string
processing code to separate functions and align the memebers in
perf_diff.
Signed-off-by: Jin Yao
---
tools/perf/Documentation/perf-diff.txt | 41
tools/perf/builtin-diff.c | 167 ++---
2 files changed, 195 insertion
samples for CPU0 and CPU1.
v2:
---
No functional change.
Signed-off-by: Jin Yao
---
tools/perf/Documentation/perf-diff.txt | 5 +
tools/perf/builtin-diff.c | 16
2 files changed, 21 insertions(+)
diff --git a/tools/perf/Documentation/perf-diff.txt
b/tools
diff: Support --time filter option" according to
Jiri's comments. Others are no functional changes.
Jin Yao (3):
perf diff: Support --time filter option
perf diff: Support --cpu filter option
perf diff: Support --pid/--tid filter options
tools/perf/Documentation/
On 2/27/2019 5:28 PM, Jiri Olsa wrote:
On Tue, Feb 26, 2019 at 08:11:07PM +0800, Jin Yao wrote:
SNIP
+ .ordered_events = true,
+ .ordering_requires_timestamps = true,
+ },
};
static struct perf_evsel *evsel_match(struct perf_evsel *evsel
On 2/27/2019 5:27 PM, Jiri Olsa wrote:
On Tue, Feb 26, 2019 at 08:11:07PM +0800, Jin Yao wrote:
SNIP
+ abstime_tmp = abstime_ostr;
data__for_each_file(i, d) {
- d->session = perf_session__new(&d->data, false, &tool);
+ d->session =
On 2/27/2019 9:10 PM, Jiri Olsa wrote:
On Wed, Feb 27, 2019 at 08:51:44PM +0800, Jin, Yao wrote:
On 2/27/2019 5:28 PM, Jiri Olsa wrote:
On Tue, Feb 26, 2019 at 08:11:07PM +0800, Jin Yao wrote:
SNIP
+ .ordered_events = true,
+ .ordering_requires_timestamps
.
So the easy way is to rename "software-gt-awake-time" to "gt-awake-time", right? Otherwise we have
to hardcode something in perf tool.
What do you think of this issue and the solution?
BTW, as the robot reported, it's related to the
On 2/12/2021 12:22 AM, Liang, Kan wrote:
On 2/11/2021 6:40 AM, Jiri Olsa wrote:
On Mon, Feb 08, 2021 at 07:24:57AM -0800, kan.li...@linux.intel.com wrote:
SNIP
Jin Yao (24):
perf jevents: Support unit value "cpu_core" and "cpu_atom"
perf util: Sav
Hi Arnaldo,
On 2/1/2021 6:27 AM, Jiri Olsa wrote:
On Thu, Jan 28, 2021 at 09:34:17AM +0800, Jin Yao wrote:
Uncore becomes die-scope on Xeon Cascade Lake-AP and perf has supported
--per-die aggregation yet.
One issue is found in check_per_pkg() for uncore events running on
AP system. On
ordered_events for raw dump")
Signed-off-by: Jin Yao
---
tools/perf/Documentation/perf-report.txt | 3 +++
tools/perf/builtin-report.c | 5 -
2 files changed, 7 insertions(+), 1 deletion(-)
diff --git a/tools/perf/Documentation/perf-report.txt
b/tools/perf/Documentation/per
Hi John,
On 2/2/2021 4:15 PM, John Garry wrote:
On 02/02/2021 02:24, Jin Yao wrote:
Hit a segmentation fault for DRAM_BW_Use on SKL/KBL.
# perf stat -M DRAM_BW_Use -a -- sleep 1
Segmentation fault (core dumped)
(gdb) backtrace
#0 __strcmp_sse2_unaligned () at
../sysdeps/x86_64
101 - 200 of 1316 matches
Mail list logo
