* [PATCH] Powerpc/perf: Wire up PMI throttling
@ 2018-11-14 14:14 Ravi Bangoria
2018-11-15 12:43 ` Michael Ellerman
0 siblings, 1 reply; 3+ messages in thread
From: Ravi Bangoria @ 2018-11-14 14:14 UTC (permalink / raw)
To: mpe, linuxppc-dev
Cc: benh, paulus, ebiederm, muriloo, christophe.leroy, npiggin,
leitao, aneesh.kumar, linux-kernel, naveen.n.rao, Ravi Bangoria
Commit 14c63f17b1fde ("perf: Drop sample rate when sampling is too
slow") introduced a way to throttle PMU interrupts if we're spending
too much time just processing those. Wire up powerpc PMI handler to
use this infrastructure.
Signed-off-by: Ravi Bangoria <ravi.bangoria@linux.ibm.com>
---
arch/powerpc/kernel/traps.c | 4 ++++
1 file changed, 4 insertions(+)
diff --git a/arch/powerpc/kernel/traps.c b/arch/powerpc/kernel/traps.c
index 9a86572db1ef..44f85fa22356 100644
--- a/arch/powerpc/kernel/traps.c
+++ b/arch/powerpc/kernel/traps.c
@@ -18,6 +18,7 @@
#include <linux/errno.h>
#include <linux/sched.h>
#include <linux/sched/debug.h>
+#include <linux/sched/clock.h>
#include <linux/kernel.h>
#include <linux/mm.h>
#include <linux/pkeys.h>
@@ -1803,9 +1804,12 @@ void vsx_unavailable_tm(struct pt_regs *regs)
void performance_monitor_exception(struct pt_regs *regs)
{
+ u64 start_clock;
__this_cpu_inc(irq_stat.pmu_irqs);
+ start_clock = sched_clock();
perf_irq(regs);
+ perf_sample_event_took(sched_clock() - start_clock);
}
#ifdef CONFIG_PPC_ADV_DEBUG_REGS
--
2.19.1
^ permalink raw reply related [flat|nested] 3+ messages in thread
* Re: [PATCH] Powerpc/perf: Wire up PMI throttling
2018-11-14 14:14 [PATCH] Powerpc/perf: Wire up PMI throttling Ravi Bangoria
@ 2018-11-15 12:43 ` Michael Ellerman
2018-11-16 9:56 ` Ravi Bangoria
0 siblings, 1 reply; 3+ messages in thread
From: Michael Ellerman @ 2018-11-15 12:43 UTC (permalink / raw)
To: Ravi Bangoria, linuxppc-dev
Cc: benh, paulus, ebiederm, muriloo, christophe.leroy, npiggin,
leitao, aneesh.kumar, linux-kernel, naveen.n.rao, Ravi Bangoria
Ravi Bangoria <ravi.bangoria@linux.ibm.com> writes:
> Commit 14c63f17b1fde ("perf: Drop sample rate when sampling is too
> slow") introduced a way to throttle PMU interrupts if we're spending
> too much time just processing those. Wire up powerpc PMI handler to
> use this infrastructure.
To be clear we have throttling of the *rate* of interrupts, but this
adds throttling based on the *time taken* to process the interrupts. Or
at least that's my understanding?
> diff --git a/arch/powerpc/kernel/traps.c b/arch/powerpc/kernel/traps.c
> index 9a86572db1ef..44f85fa22356 100644
> --- a/arch/powerpc/kernel/traps.c
> +++ b/arch/powerpc/kernel/traps.c
> @@ -18,6 +18,7 @@
> #include <linux/errno.h>
> #include <linux/sched.h>
> #include <linux/sched/debug.h>
> +#include <linux/sched/clock.h>
> #include <linux/kernel.h>
> #include <linux/mm.h>
> #include <linux/pkeys.h>
> @@ -1803,9 +1804,12 @@ void vsx_unavailable_tm(struct pt_regs *regs)
>
> void performance_monitor_exception(struct pt_regs *regs)
> {
> + u64 start_clock;
> __this_cpu_inc(irq_stat.pmu_irqs);
>
> + start_clock = sched_clock();
> perf_irq(regs);
> + perf_sample_event_took(sched_clock() - start_clock);
> }
Despite the name, perf_irq() may not actually be the perf IRQ handler :)
It's a function pointer which might call perf or might call oprofile, or
a dummy handler.
I don't think we should be calling perf_sample_event_took() if we're not
actually using perf, that is wasteful at best.
So the timing logic should go in the perf specific handler I think.
ie. perf_event_interrupt().
cheers
^ permalink raw reply [flat|nested] 3+ messages in thread
* Re: [PATCH] Powerpc/perf: Wire up PMI throttling
2018-11-15 12:43 ` Michael Ellerman
@ 2018-11-16 9:56 ` Ravi Bangoria
0 siblings, 0 replies; 3+ messages in thread
From: Ravi Bangoria @ 2018-11-16 9:56 UTC (permalink / raw)
To: Michael Ellerman
Cc: linuxppc-dev, benh, paulus, ebiederm, muriloo, christophe.leroy,
npiggin, leitao, aneesh.kumar, linux-kernel, naveen.n.rao,
Ravi Bangoria
On 11/15/18 6:13 PM, Michael Ellerman wrote:
> Ravi Bangoria <ravi.bangoria@linux.ibm.com> writes:
>
>> Commit 14c63f17b1fde ("perf: Drop sample rate when sampling is too
>> slow") introduced a way to throttle PMU interrupts if we're spending
>> too much time just processing those. Wire up powerpc PMI handler to
>> use this infrastructure.
>
> To be clear we have throttling of the *rate* of interrupts, but this
> adds throttling based on the *time taken* to process the interrupts. Or
> at least that's my understanding?
Right. This throttles based on time taken to process the interrupts.
>
>> diff --git a/arch/powerpc/kernel/traps.c b/arch/powerpc/kernel/traps.c
>> index 9a86572db1ef..44f85fa22356 100644
>> --- a/arch/powerpc/kernel/traps.c
>> +++ b/arch/powerpc/kernel/traps.c
>> @@ -18,6 +18,7 @@
>> #include <linux/errno.h>
>> #include <linux/sched.h>
>> #include <linux/sched/debug.h>
>> +#include <linux/sched/clock.h>
>> #include <linux/kernel.h>
>> #include <linux/mm.h>
>> #include <linux/pkeys.h>
>> @@ -1803,9 +1804,12 @@ void vsx_unavailable_tm(struct pt_regs *regs)
>>
>> void performance_monitor_exception(struct pt_regs *regs)
>> {
>> + u64 start_clock;
>> __this_cpu_inc(irq_stat.pmu_irqs);
>>
>> + start_clock = sched_clock();
>> perf_irq(regs);
>> + perf_sample_event_took(sched_clock() - start_clock);
>> }
>
> Despite the name, perf_irq() may not actually be the perf IRQ handler :)
>
> It's a function pointer which might call perf or might call oprofile, or
> a dummy handler.
>
> I don't think we should be calling perf_sample_event_took() if we're not
> actually using perf, that is wasteful at best.
>
> So the timing logic should go in the perf specific handler I think.
> ie. perf_event_interrupt().
Makes sense. I'll re-send with that change.
Thanks,
Ravi
^ permalink raw reply [flat|nested] 3+ messages in thread
end of thread, other threads:[~2018-11-16 9:56 UTC | newest]
Thread overview: 3+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2018-11-14 14:14 [PATCH] Powerpc/perf: Wire up PMI throttling Ravi Bangoria
2018-11-15 12:43 ` Michael Ellerman
2018-11-16 9:56 ` Ravi Bangoria
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).