Hi Dave,

By the way, in http://lists.linaro.org/pipermail/linaro-dev/2011-September/thread.html#7560, we discussed about oprofile timer mode granularity for platforms, which had the PMU bug that loses PMU interrupts (4430, 4460 and maybe 4470). Was this timer granularity proposed/introduced ? Is it in fact part of perf infrastructure (don't know if perf relies directly on oprofile drivers/oprofile infrastructuret) ?

Regards Fred

OMAP Platform Business Unit - OMAP System Engineering - Platform Enablement - System Multimedia

Texas Instruments France SA, 821 Avenue Jack Kilby, 06270 Villeneuve Loubet. 036 420 040 R.C.S Antibes. Capital de EUR 753.920

-----Original Message-----

From: linaro-dev-bounces@lists.linaro.org [mailto:linaro-dev-bounces@lists.linaro.org] On Behalf Of Dave Martin Sent: Thursday, March 01, 2012 9:57 AM To: zachary6626 Cc: linaro-dev Subject: Re: Do ARM Cortex-A9 PMU supoort oprofile

On Wed, Feb 29, 2012 at 11:21:58PM +0800, zachary6626 wrote:

...

Sorry disturbing you!

I have ask google for do ARM Cortex-A9 PMU supoort oprofile, but it do not have a answer definitely

.

...

So i ask you do ARM Cortex-A9 PMU supoort oprofile.

Can u help me?

In general, yes it does work. But board support is needed in addition to the CPU PMU support.

If you're trying to work on omap4 (panda) you may need some extra patches:

For more information on this, take a look at the following thread where this issue was discussed recently:

http://lists.linaro.org/pipermail/linaro-dev/2012-February/010326.html

Cheers ---Dave

---

linaro-dev mailing list linaro-dev@lists.linaro.org http://lists.linaro.org/mailman/listinfo/linaro-dev

Use hrtimer ;-) Well, to make PMU work, look at linaro-dev mail archives, Lei Ming and Antipov Dmitry had a discussion recently on the right patches

Regards Fred

OMAP Platform Business Unit - OMAP System Engineering - Platform Enablement - System Multimedia

Texas Instruments France SA, 821 Avenue Jack Kilby, 06270 Villeneuve Loubet. 036 420 040 R.C.S Antibes. Capital de EUR 753.920

-----Original Message-----

From: zachary6626 [mailto:zachary6626@gmail.com] Sent: Thursday, March 01, 2012 4:45 PM To: Turgis, Frederic; Dave Martin Cc: linaro-dev Subject: Re: RE: Do ARM Cortex-A9 PMU supoort oprofile

Dear Dave and Turgis,

  Thanks for your replying.
  I have run oprofile on OMAP4460, but there are many problems.

  My test env is:
  oprofile version: 0.9.7
  Android ver: ICS
  kernel ver: 3.0
  CPU:    OMAP4460

  Oprofile ran abnormally after I patched the kernel with a patch from

http://omappedia.org/wiki/Android_Debugging#OProfile_on_OMAP4; . Step1: # opcontrol --setup

  Step2:
  # opcontrol --vmlinux=/vmlinux --kernel-range 0xc0046000,0xc009a000 --event CPU_CYCLES:25000000
  I analysed the log after send the command "dmesg", found that the PMU interrupt enable control

bit has been modified, the oprofile can not work in PMU mode.

  But when I patch the kernel with a patch from http://ssvb.github.com/2011/08/23/yet-another-

oprofile-tutorial.html. The oprofile can work in hrtimer mode.

  Could you share some suggestions?

---

zachary6626

...

Hi Dave,

By the way, in http://lists.linaro.org/pipermail/linaro-dev/2011-September/thread.html#7560, we

discussed about oprofile timer mode granularity for platforms, which had the PMU bug that loses PMU interrupts (4430, 4460 and maybe 4470).

...

Was this timer granularity proposed/introduced ? Is it in fact part of perf infrastructure (don't

know if perf relies directly on oprofile drivers/oprofile infrastructuret) ?

...

Regards Fred

OMAP Platform Business Unit - OMAP System Engineering - Platform Enablement - System Multimedia

...

Texas Instruments France SA, 821 Avenue Jack Kilby, 06270 Villeneuve Loubet. 036 420 040 R.C.S

Antibes. Capital de EUR 753.920

...

-----Original Message-----

...

From: linaro-dev-bounces@lists.linaro.org [mailto:linaro-dev-bounces@lists.linaro.org] On Behalf Of Dave Martin Sent: Thursday, March 01, 2012 9:57 AM To: zachary6626 Cc: linaro-dev Subject: Re: Do ARM Cortex-A9 PMU supoort oprofile

On Wed, Feb 29, 2012 at 11:21:58PM +0800, zachary6626 wrote:

...

Sorry disturbing you!

I have ask google for do ARM Cortex-A9 PMU supoort oprofile, but it do not have a answer

definitely

...

...

.

...

So i ask you do ARM Cortex-A9 PMU supoort oprofile.

Can u help me?

In general, yes it does work. But board support is needed in addition to the CPU PMU support.

If you're trying to work on omap4 (panda) you may need some extra patches:

For more information on this, take a look at the following thread where this issue was discussed recently:

http://lists.linaro.org/pipermail/linaro-dev/2012-February/010326.html

Cheers ---Dave

---

linaro-dev mailing list linaro-dev@lists.linaro.org http://lists.linaro.org/mailman/listinfo/linaro-dev

On Thu, 2012-03-01 at 16:59 +0000, Turgis, Frederic wrote:

Wrong Dave, I meant Dave Long ;-)

OMAP Platform Business Unit - OMAP System Engineering - Platform Enablement - System Multimedia

...

Hi Dave,

By the way, in http://lists.linaro.org/pipermail/linaro-dev/2011-September/thread.html#7560, we discussed about oprofile timer mode granularity for platforms, which had the PMU bug that loses PMU interrupts (4430, 4460 and maybe 4470). Was this timer granularity proposed/introduced ? Is it in fact part of perf infrastructure (don't know if perf relies directly on oprofile drivers/oprofile infrastructuret) ?

Sorry for the delay in responding.

As far as I know this has not happened. Such a change would also have to include making the use of timer interrupts (as opposed to PMU counter interrupts) selectable at runtime. I believe the use of timer interrupts for oprofile data collection is still a configuration/boot option. Alternatively we could just decide to deconfigure PMU hardware counters by default to make timers accessible.

-dl