* [PATCH V2 0/2] Improvement and fixes for BHRB
@ 2013-05-28 4:47 Anshuman Khandual
2013-05-28 4:47 ` [PATCH V2 1/2] powerpc, perf: Ignore separate BHRB privilege state filter request Anshuman Khandual
2013-05-28 4:47 ` [PATCH V2 2/2] powerpc, perf: BHRB filter configuration should follow the task Anshuman Khandual
0 siblings, 2 replies; 6+ messages in thread
From: Anshuman Khandual @ 2013-05-28 4:47 UTC (permalink / raw)
To: linuxppc-dev; +Cc: mikey
(1) The first patch fixes a situation like this
Before patch:-
------------
./perf record -j any -e branch-misses:k ls
Error:
The sys_perf_event_open() syscall returned with 95 (Operation not supported) for event (branch-misses:k).
/bin/dmesg may provide additional information.
No CONFIG_PERF_EVENTS=y kernel support configured?
Here 'perf record' actually copies over ':k' filter request into BHRB
privilege state filter config and our previous check in kernel would
fail that.
After patch:-
-------------
/perf record -j any -e branch-misses:k ls
perf perf.data perf.data.old test-mmap-ring
[ perf record: Woken up 1 times to write data ]
[ perf record: Captured and wrote 0.002 MB perf.data (~102 samples) ]
(2) The second patch fixes context migration for BHRB filter configuration
Changes in V2
-------------
(1) Updated the comment section of the code
(2) Removed the informational print
(3) Updated the commit section of the first patch
Anshuman Khandual (2):
powerpc, perf: Ignore separate BHRB privilege state filter request
powerpc, perf: BHRB filter configuration should follow the task
arch/powerpc/perf/core-book3s.c | 5 ++++-
arch/powerpc/perf/power8-pmu.c | 12 ++++--------
2 files changed, 8 insertions(+), 9 deletions(-)
--
1.7.11.7
^ permalink raw reply [flat|nested] 6+ messages in thread
* [PATCH V2 1/2] powerpc, perf: Ignore separate BHRB privilege state filter request
2013-05-28 4:47 [PATCH V2 0/2] Improvement and fixes for BHRB Anshuman Khandual
@ 2013-05-28 4:47 ` Anshuman Khandual
2013-06-06 4:56 ` Michael Neuling
2013-05-28 4:47 ` [PATCH V2 2/2] powerpc, perf: BHRB filter configuration should follow the task Anshuman Khandual
1 sibling, 1 reply; 6+ messages in thread
From: Anshuman Khandual @ 2013-05-28 4:47 UTC (permalink / raw)
To: linuxppc-dev; +Cc: mikey
Completely ignore BHRB privilege state filter request as we are
already configuring that with privilege state filtering attribute
for the accompanying PMU event. This would help achieve cleaner
user space interaction for BHRB.
This patch fixes a situation like this
Before patch:-
------------
./perf record -j any -e branch-misses:k ls
Error:
The sys_perf_event_open() syscall returned with 95 (Operation not supported) for event (branch-misses:k).
/bin/dmesg may provide additional information.
No CONFIG_PERF_EVENTS=y kernel support configured?
Here 'perf record' actually copies over ':k' filter request into BHRB
privilege state filter config and our previous check in kernel would
fail that.
After patch:-
-------------
./perf record -j any -e branch-misses:k ls
perf perf.data perf.data.old test-mmap-ring
[ perf record: Woken up 1 times to write data ]
[ perf record: Captured and wrote 0.002 MB perf.data (~102 samples) ]
Signed-off-by: Anshuman Khandual <khandual@linux.vnet.ibm.com>
---
arch/powerpc/perf/power8-pmu.c | 12 ++++--------
1 file changed, 4 insertions(+), 8 deletions(-)
diff --git a/arch/powerpc/perf/power8-pmu.c b/arch/powerpc/perf/power8-pmu.c
index f7d1c4f..3a58416 100644
--- a/arch/powerpc/perf/power8-pmu.c
+++ b/arch/powerpc/perf/power8-pmu.c
@@ -525,16 +525,12 @@ static u64 power8_bhrb_filter_map(u64 branch_sample_type)
u64 pmu_bhrb_filter = 0;
u64 br_privilege = branch_sample_type & ONLY_PLM;
- /* BHRB and regular PMU events share the same prvillege state
+ /* BHRB and regular PMU events share the same privilege state
* filter configuration. BHRB is always recorded along with a
- * regular PMU event. So privilege state filter criteria for BHRB
- * and the companion PMU events has to be the same. As a default
- * "perf record" tool sets all privillege bits ON when no filter
- * criteria is provided in the command line. So as along as all
- * privillege bits are ON or they are OFF, we are good to go.
+ * regular PMU event. As the privilege state filter is handled
+ * in the basic PMC configuration of the accompanying regular
+ * PMU event, we ignore any separate BHRB specific request.
*/
- if ((br_privilege != 7) && (br_privilege != 0))
- return -1;
/* No branch filter requested */
if (branch_sample_type & PERF_SAMPLE_BRANCH_ANY)
--
1.7.11.7
^ permalink raw reply related [flat|nested] 6+ messages in thread
* Re: [PATCH V2 1/2] powerpc, perf: Ignore separate BHRB privilege state filter request
2013-05-28 4:47 ` [PATCH V2 1/2] powerpc, perf: Ignore separate BHRB privilege state filter request Anshuman Khandual
@ 2013-06-06 4:56 ` Michael Neuling
2013-06-06 6:04 ` Anshuman Khandual
0 siblings, 1 reply; 6+ messages in thread
From: Michael Neuling @ 2013-06-06 4:56 UTC (permalink / raw)
To: Anshuman Khandual; +Cc: linuxppc-dev
Anshuman Khandual <khandual@linux.vnet.ibm.com> wrote:
> Completely ignore BHRB privilege state filter request as we are
> already configuring that with privilege state filtering attribute
> for the accompanying PMU event. This would help achieve cleaner
> user space interaction for BHRB.
>=20
> This patch fixes a situation like this
>=20
> Before patch:-
> ------------
> ./perf record -j any -e branch-misses:k ls
> Error:
> The sys_perf_event_open() syscall returned with 95 (Operation not support=
ed) for event (branch-misses:k).
> /bin/dmesg may provide additional information.
> No CONFIG_PERF_EVENTS=3Dy kernel support configured?
>=20
> Here 'perf record' actually copies over ':k' filter request into BHRB
> privilege state filter config and our previous check in kernel would
> fail that.
>=20
> After patch:-
> -------------
> ./perf record -j any -e branch-misses:k ls
> perf perf.data perf.data.old test-mmap-ring
> [ perf record: Woken up 1 times to write data ]
> [ perf record: Captured and wrote 0.002 MB perf.data (~102 samples) ]
>=20
> Signed-off-by: Anshuman Khandual <khandual@linux.vnet.ibm.com>
> ---
> arch/powerpc/perf/power8-pmu.c | 12 ++++--------
> 1 file changed, 4 insertions(+), 8 deletions(-)
>=20
> diff --git a/arch/powerpc/perf/power8-pmu.c b/arch/powerpc/perf/power8-pm=
u.c
> index f7d1c4f..3a58416 100644
> --- a/arch/powerpc/perf/power8-pmu.c
> +++ b/arch/powerpc/perf/power8-pmu.c
> @@ -525,16 +525,12 @@ static u64 power8_bhrb_filter_map(u64 branch_sample=
_type)
> u64 pmu_bhrb_filter =3D 0;
> u64 br_privilege =3D branch_sample_type & ONLY_PLM;
>=20=20
> - /* BHRB and regular PMU events share the same prvillege state
> + /* BHRB and regular PMU events share the same privilege state
> * filter configuration. BHRB is always recorded along with a
> - * regular PMU event. So privilege state filter criteria for BHRB
> - * and the companion PMU events has to be the same. As a default
> - * "perf record" tool sets all privillege bits ON when no filter
> - * criteria is provided in the command line. So as along as all
> - * privillege bits are ON or they are OFF, we are good to go.
> + * regular PMU event. As the privilege state filter is handled
> + * in the basic PMC configuration of the accompanying regular
> + * PMU event, we ignore any separate BHRB specific request.
> */
> - if ((br_privilege !=3D 7) && (br_privilege !=3D 0))
> - return -1;
br_privilege is now unused which causes this compile error.
arch/powerpc/perf/power8-pmu.c:526:6: error: unused variable =E2=80=98br_pr=
ivilege=E2=80=99 [-Werror=3Dunused-variable]
I assume since you didn't compile test this code, you also didn't
runtime test it either?!!?!?!?
Mikey
>=20=20
> /* No branch filter requested */
> if (branch_sample_type & PERF_SAMPLE_BRANCH_ANY)
> --=20
> 1.7.11.7
>=20
> _______________________________________________
> Linuxppc-dev mailing list
> Linuxppc-dev@lists.ozlabs.org
> https://lists.ozlabs.org/listinfo/linuxppc-dev
>=20
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [PATCH V2 1/2] powerpc, perf: Ignore separate BHRB privilege state filter request
2013-06-06 4:56 ` Michael Neuling
@ 2013-06-06 6:04 ` Anshuman Khandual
2013-06-06 23:46 ` Michael Neuling
0 siblings, 1 reply; 6+ messages in thread
From: Anshuman Khandual @ 2013-06-06 6:04 UTC (permalink / raw)
To: Michael Neuling; +Cc: linuxppc-dev
On 06/06/2013 10:26 AM, Michael Neuling wrote:
> Anshuman Khandual <khandual@linux.vnet.ibm.com> wrote:
>
>> Completely ignore BHRB privilege state filter request as we are
>> already configuring that with privilege state filtering attribute
>> for the accompanying PMU event. This would help achieve cleaner
>> user space interaction for BHRB.
>>
>> This patch fixes a situation like this
>>
>> Before patch:-
>> ------------
>> ./perf record -j any -e branch-misses:k ls
>> Error:
>> The sys_perf_event_open() syscall returned with 95 (Operation not supported) for event (branch-misses:k).
>> /bin/dmesg may provide additional information.
>> No CONFIG_PERF_EVENTS=y kernel support configured?
>>
>> Here 'perf record' actually copies over ':k' filter request into BHRB
>> privilege state filter config and our previous check in kernel would
>> fail that.
>>
>> After patch:-
>> -------------
>> ./perf record -j any -e branch-misses:k ls
>> perf perf.data perf.data.old test-mmap-ring
>> [ perf record: Woken up 1 times to write data ]
>> [ perf record: Captured and wrote 0.002 MB perf.data (~102 samples) ]
>>
>> Signed-off-by: Anshuman Khandual <khandual@linux.vnet.ibm.com>
>> ---
>> arch/powerpc/perf/power8-pmu.c | 12 ++++--------
>> 1 file changed, 4 insertions(+), 8 deletions(-)
>>
>> diff --git a/arch/powerpc/perf/power8-pmu.c b/arch/powerpc/perf/power8-pmu.c
>> index f7d1c4f..3a58416 100644
>> --- a/arch/powerpc/perf/power8-pmu.c
>> +++ b/arch/powerpc/perf/power8-pmu.c
>> @@ -525,16 +525,12 @@ static u64 power8_bhrb_filter_map(u64 branch_sample_type)
>> u64 pmu_bhrb_filter = 0;
>> u64 br_privilege = branch_sample_type & ONLY_PLM;
>>
>> - /* BHRB and regular PMU events share the same prvillege state
>> + /* BHRB and regular PMU events share the same privilege state
>> * filter configuration. BHRB is always recorded along with a
>> - * regular PMU event. So privilege state filter criteria for BHRB
>> - * and the companion PMU events has to be the same. As a default
>> - * "perf record" tool sets all privillege bits ON when no filter
>> - * criteria is provided in the command line. So as along as all
>> - * privillege bits are ON or they are OFF, we are good to go.
>> + * regular PMU event. As the privilege state filter is handled
>> + * in the basic PMC configuration of the accompanying regular
>> + * PMU event, we ignore any separate BHRB specific request.
>> */
>> - if ((br_privilege != 7) && (br_privilege != 0))
>> - return -1;
>
> br_privilege is now unused which causes this compile error.
>
> arch/powerpc/perf/power8-pmu.c:526:6: error: unused variable ‘br_privilege’ [-Werror=unused-variable]
>
> I assume since you didn't compile test this code, you also didn't
> runtime test it either?!!?!?!?
Always compile with the cross compiler "ARCH=powerpc CROSS_COMPILE=powerpc64-linux-gnu- make"
I copied over arch/powerpc/config/pseries_defconfig and compiled it. Some how it missed
to be a 64 bit kernel config which resulted in shutting down compilation of arch/powerpc/perf/power8-pmu.c
file. So never encountered this compiler warning message before. Will fix this.
Regards
Anshuman
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [PATCH V2 1/2] powerpc, perf: Ignore separate BHRB privilege state filter request
2013-06-06 6:04 ` Anshuman Khandual
@ 2013-06-06 23:46 ` Michael Neuling
0 siblings, 0 replies; 6+ messages in thread
From: Michael Neuling @ 2013-06-06 23:46 UTC (permalink / raw)
To: Anshuman Khandual; +Cc: linuxppc-dev
Anshuman Khandual <khandual@linux.vnet.ibm.com> wrote:
> On 06/06/2013 10:26 AM, Michael Neuling wrote:
> > Anshuman Khandual <khandual@linux.vnet.ibm.com> wrote:
> >=20
> >> Completely ignore BHRB privilege state filter request as we are
> >> already configuring that with privilege state filtering attribute
> >> for the accompanying PMU event. This would help achieve cleaner
> >> user space interaction for BHRB.
> >>
> >> This patch fixes a situation like this
> >>
> >> Before patch:-
> >> ------------
> >> ./perf record -j any -e branch-misses:k ls
> >> Error:
> >> The sys_perf_event_open() syscall returned with 95 (Operation not supp=
orted) for event (branch-misses:k).
> >> /bin/dmesg may provide additional information.
> >> No CONFIG_PERF_EVENTS=3Dy kernel support configured?
> >>
> >> Here 'perf record' actually copies over ':k' filter request into BHRB
> >> privilege state filter config and our previous check in kernel would
> >> fail that.
> >>
> >> After patch:-
> >> -------------
> >> ./perf record -j any -e branch-misses:k ls
> >> perf perf.data perf.data.old test-mmap-ring
> >> [ perf record: Woken up 1 times to write data ]
> >> [ perf record: Captured and wrote 0.002 MB perf.data (~102 samples) ]
> >>
> >> Signed-off-by: Anshuman Khandual <khandual@linux.vnet.ibm.com>
> >> ---
> >> arch/powerpc/perf/power8-pmu.c | 12 ++++--------
> >> 1 file changed, 4 insertions(+), 8 deletions(-)
> >>
> >> diff --git a/arch/powerpc/perf/power8-pmu.c b/arch/powerpc/perf/power8=
-pmu.c
> >> index f7d1c4f..3a58416 100644
> >> --- a/arch/powerpc/perf/power8-pmu.c
> >> +++ b/arch/powerpc/perf/power8-pmu.c
> >> @@ -525,16 +525,12 @@ static u64 power8_bhrb_filter_map(u64 branch_sam=
ple_type)
> >> u64 pmu_bhrb_filter =3D 0;
> >> u64 br_privilege =3D branch_sample_type & ONLY_PLM;
> >>=20=20
> >> - /* BHRB and regular PMU events share the same prvillege state
> >> + /* BHRB and regular PMU events share the same privilege state
> >> * filter configuration. BHRB is always recorded along with a
> >> - * regular PMU event. So privilege state filter criteria for BHRB
> >> - * and the companion PMU events has to be the same. As a default
> >> - * "perf record" tool sets all privillege bits ON when no filter
> >> - * criteria is provided in the command line. So as along as all
> >> - * privillege bits are ON or they are OFF, we are good to go.
> >> + * regular PMU event. As the privilege state filter is handled
> >> + * in the basic PMC configuration of the accompanying regular
> >> + * PMU event, we ignore any separate BHRB specific request.
> >> */
> >> - if ((br_privilege !=3D 7) && (br_privilege !=3D 0))
> >> - return -1;
> >=20
> > br_privilege is now unused which causes this compile error.
> >=20
> > arch/powerpc/perf/power8-pmu.c:526:6: error: unused variable =E2=80=98b=
r_privilege=E2=80=99 [-Werror=3Dunused-variable]
> >=20
> > I assume since you didn't compile test this code, you also didn't
> > runtime test it either?!!?!?!?
>=20
>=20
> Always compile with the cross compiler "ARCH=3Dpowerpc
> CROSS_COMPILE=3Dpowerpc64-linux-gnu- make" I copied over
> arch/powerpc/config/pseries_defconfig and compiled it. Some how it
> missed to be a 64 bit kernel config which resulted in shutting down
> compilation of arch/powerpc/perf/power8-pmu.c file. So never
> encountered this compiler warning message before. Will fix this.
I don't understand this at all... pseries_defconfig is a 64bit kernel.
BTW. you don't need to copy it. Just do "make pseries_defconfig && make
vmlinux".
Mikey
^ permalink raw reply [flat|nested] 6+ messages in thread
* [PATCH V2 2/2] powerpc, perf: BHRB filter configuration should follow the task
2013-05-28 4:47 [PATCH V2 0/2] Improvement and fixes for BHRB Anshuman Khandual
2013-05-28 4:47 ` [PATCH V2 1/2] powerpc, perf: Ignore separate BHRB privilege state filter request Anshuman Khandual
@ 2013-05-28 4:47 ` Anshuman Khandual
1 sibling, 0 replies; 6+ messages in thread
From: Anshuman Khandual @ 2013-05-28 4:47 UTC (permalink / raw)
To: linuxppc-dev; +Cc: mikey
When the task moves around the system, the corresponding cpuhw
per cpu strcuture should be popullated with the BHRB filter
request value so that PMU could be configured appropriately with
that during the next call into power_pmu_enable().
Signed-off-by: Anshuman Khandual <khandual@linux.vnet.ibm.com>
---
arch/powerpc/perf/core-book3s.c | 5 ++++-
1 file changed, 4 insertions(+), 1 deletion(-)
diff --git a/arch/powerpc/perf/core-book3s.c b/arch/powerpc/perf/core-book3s.c
index 426180b..48c68a8 100644
--- a/arch/powerpc/perf/core-book3s.c
+++ b/arch/powerpc/perf/core-book3s.c
@@ -1122,8 +1122,11 @@ nocheck:
ret = 0;
out:
- if (has_branch_stack(event))
+ if (has_branch_stack(event)) {
power_pmu_bhrb_enable(event);
+ cpuhw->bhrb_filter = ppmu->bhrb_filter_map(
+ event->attr.branch_sample_type);
+ }
perf_pmu_enable(event->pmu);
local_irq_restore(flags);
--
1.7.11.7
^ permalink raw reply related [flat|nested] 6+ messages in thread
end of thread, other threads:[~2013-06-06 23:46 UTC | newest]
Thread overview: 6+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2013-05-28 4:47 [PATCH V2 0/2] Improvement and fixes for BHRB Anshuman Khandual
2013-05-28 4:47 ` [PATCH V2 1/2] powerpc, perf: Ignore separate BHRB privilege state filter request Anshuman Khandual
2013-06-06 4:56 ` Michael Neuling
2013-06-06 6:04 ` Anshuman Khandual
2013-06-06 23:46 ` Michael Neuling
2013-05-28 4:47 ` [PATCH V2 2/2] powerpc, perf: BHRB filter configuration should follow the task Anshuman Khandual
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).