2014-06-15 05:58:41

by Vince Weaver

[permalink] [raw]
Subject: [PATCH 3/6] metag,perf: Use common PMU interrupt disabled code


Transition to using the new generic PERF_PMU_CAP_NO_INTERRUPT method for
failing a sampling event when no PMU interrupt is available.

Signed-off-by: Vince Weaver <[email protected]>

diff --git a/arch/metag/kernel/perf/perf_event.c b/arch/metag/kernel/perf/perf_event.c
index 5cc4d4d..02c0873 100644
--- a/arch/metag/kernel/perf/perf_event.c
+++ b/arch/metag/kernel/perf/perf_event.c
@@ -568,16 +568,6 @@ static int _hw_perf_event_init(struct perf_event *event)
return -EINVAL;

/*
- * Early cores have "limited" counters - they have no overflow
- * interrupts - and so are unable to do sampling without extra work
- * and timer assistance.
- */
- if (metag_pmu->max_period == 0) {
- if (hwc->sample_period)
- return -EINVAL;
- }
-
- /*
* Don't assign an index until the event is placed into the hardware.
* -1 signifies that we're still deciding where to put it. On SMP
* systems each core has its own set of counters, so we can't do any
@@ -866,6 +856,15 @@ static int __init init_hw_perf_events(void)
pr_info("enabled with %s PMU driver, %d counters available\n",
metag_pmu->name, metag_pmu->max_events);

+ /*
+ * Early cores have "limited" counters - they have no overflow
+ * interrupts - and so are unable to do sampling without extra work
+ * and timer assistance.
+ */
+ if (metag_pmu->max_period == 0) {
+ metag_pmu->pmu.capabilities |= PERF_PMU_CAP_NO_INTERRUPT;
+ }
+
/* Initialise the active events and reservation mutex */
atomic_set(&metag_pmu->active_events, 0);
mutex_init(&metag_pmu->reserve_mutex);


2014-06-16 14:57:05

by James Hogan

[permalink] [raw]
Subject: Re: [PATCH 3/6] metag,perf: Use common PMU interrupt disabled code

Hi Vince,

On 15/06/14 07:03, Vince Weaver wrote:
>
> Transition to using the new generic PERF_PMU_CAP_NO_INTERRUPT method for
> failing a sampling event when no PMU interrupt is available.
>
> Signed-off-by: Vince Weaver <[email protected]>


> @@ -866,6 +856,15 @@ static int __init init_hw_perf_events(void)
> pr_info("enabled with %s PMU driver, %d counters available\n",
> metag_pmu->name, metag_pmu->max_events);
>
> + /*
> + * Early cores have "limited" counters - they have no overflow
> + * interrupts - and so are unable to do sampling without extra work
> + * and timer assistance.
> + */
> + if (metag_pmu->max_period == 0) {
> + metag_pmu->pmu.capabilities |= PERF_PMU_CAP_NO_INTERRUPT;
> + }

The braces could be dropped.

Otherwise looks good to me,

Acked-by: James Hogan <[email protected]>

Cheers
James

Subject: [tip:perf/core] metag, perf: Use common PMU interrupt disabled code

Commit-ID: 1b92722fafe2003e49ec904f33c0020ade36bd16
Gitweb: http://git.kernel.org/tip/1b92722fafe2003e49ec904f33c0020ade36bd16
Author: Vince Weaver <[email protected]>
AuthorDate: Sun, 15 Jun 2014 02:03:13 -0400
Committer: Ingo Molnar <[email protected]>
CommitDate: Wed, 18 Jun 2014 18:43:45 +0200

metag, perf: Use common PMU interrupt disabled code

Transition to using the new generic PERF_PMU_CAP_NO_INTERRUPT method for
failing a sampling event when no PMU interrupt is available.

Signed-off-by: Vince Weaver <[email protected]>
Acked-by: James Hogan <[email protected]>
Signed-off-by: Peter Zijlstra <[email protected]>
Cc: Arnaldo Carvalho de Melo <[email protected]>
Cc: Paul Mackerras <[email protected]>
Cc: [email protected]
Link: http://lkml.kernel.org/r/alpine.DEB.2.10.1406150202050.16738@vincent-weaver-1.umelst.maine.edu
Signed-off-by: Ingo Molnar <[email protected]>
---
arch/metag/kernel/perf/perf_event.c | 19 +++++++++----------
1 file changed, 9 insertions(+), 10 deletions(-)

diff --git a/arch/metag/kernel/perf/perf_event.c b/arch/metag/kernel/perf/perf_event.c
index 5cc4d4d..02c0873 100644
--- a/arch/metag/kernel/perf/perf_event.c
+++ b/arch/metag/kernel/perf/perf_event.c
@@ -568,16 +568,6 @@ static int _hw_perf_event_init(struct perf_event *event)
return -EINVAL;

/*
- * Early cores have "limited" counters - they have no overflow
- * interrupts - and so are unable to do sampling without extra work
- * and timer assistance.
- */
- if (metag_pmu->max_period == 0) {
- if (hwc->sample_period)
- return -EINVAL;
- }
-
- /*
* Don't assign an index until the event is placed into the hardware.
* -1 signifies that we're still deciding where to put it. On SMP
* systems each core has its own set of counters, so we can't do any
@@ -866,6 +856,15 @@ static int __init init_hw_perf_events(void)
pr_info("enabled with %s PMU driver, %d counters available\n",
metag_pmu->name, metag_pmu->max_events);

+ /*
+ * Early cores have "limited" counters - they have no overflow
+ * interrupts - and so are unable to do sampling without extra work
+ * and timer assistance.
+ */
+ if (metag_pmu->max_period == 0) {
+ metag_pmu->pmu.capabilities |= PERF_PMU_CAP_NO_INTERRUPT;
+ }
+
/* Initialise the active events and reservation mutex */
atomic_set(&metag_pmu->active_events, 0);
mutex_init(&metag_pmu->reserve_mutex);