Hi,
With libpfm4, I have developed a set of examples which are using
perf_events to demonstrate typical counting and sampling scenarios.
In two of the sampling program, it appears time_enabled, time_running
included in the samples are incorrect.
The program notify_self.c uses IOC_REFRESH to get
Gary,
On Tue, Oct 6, 2009 at 1:36 AM, wrote:
>
> I have been asked to provide a perfmon2 kernel patch for a 2.6.30 fedora
> kernel. Wondered if you would recommend trying to port the 2.6.29 patch or
> if it would be better to work from your git tree to pick up the changes you
> have already mad
* David Miller wrote:
> From: stephane eranian
> Date: Wed, 7 Oct 2009 14:31:58 +0200
>
> > What PPC does is probably the only way to do this given the interface
> > between
> > generic and machine-specific code. The one advantage I see is that it works
> > inside an event group but also acro
On Thu, Oct 8, 2009 at 10:08 PM, Ingo Molnar wrote:
>
> * David Miller wrote:
>
>> From: stephane eranian
>> Date: Wed, 7 Oct 2009 14:31:58 +0200
>>
>> > What PPC does is probably the only way to do this given the interface
>> > between
>> > generic and machine-specific code. The one advantage
Stephane
Thanks, the path you provided was missing a small part but as they say "it
was close
enough for government work" and I am pulling from the git now.
For anyone else who may be watching, the correct command is:
git clone
git://git.kernel.org/pub/scm/linux/kernel/git/eranian/linux-2.6.git
stephane eranian writes:
> I am not an expert on PPC PMU register constraints but I took a quick look
> at the code and in particular hw_perf_enable() where the action seems to be.
>
> Given that in kernel/perf_events.c, the PMU specific layer is invoked on a per
> event basis in event_sched_in()
