Re: [PATCH] perf cgroups: Don't rotate events for cgroups unnecessarily

2019-10-14 Thread Ganapatrao Kulkarni
Hi Peter, On Wed, Sep 18, 2019 at 12:51 PM Ganapatrao Kulkarni wrote: > > On Fri, Aug 23, 2019 at 6:33 PM Peter Zijlstra wrote: > > > > On Fri, Aug 23, 2019 at 06:26:34PM +0530, Ganapatrao Kulkarni wrote: > > > On Fri, Aug 23, 2019 at 5:29 PM Peter Zijlstra > > > wrote: > > > > On Fri, Aug

Re: [PATCH] perf cgroups: Don't rotate events for cgroups unnecessarily

2019-09-18 Thread Ganapatrao Kulkarni
On Fri, Aug 23, 2019 at 6:33 PM Peter Zijlstra wrote: > > On Fri, Aug 23, 2019 at 06:26:34PM +0530, Ganapatrao Kulkarni wrote: > > On Fri, Aug 23, 2019 at 5:29 PM Peter Zijlstra wrote: > > > On Fri, Aug 23, 2019 at 04:13:46PM +0530, Ganapatrao Kulkarni wrote: > > > > > > > We are seeing

Re: [PATCH] perf cgroups: Don't rotate events for cgroups unnecessarily

2019-08-23 Thread Peter Zijlstra
On Fri, Aug 23, 2019 at 06:26:34PM +0530, Ganapatrao Kulkarni wrote: > On Fri, Aug 23, 2019 at 5:29 PM Peter Zijlstra wrote: > > On Fri, Aug 23, 2019 at 04:13:46PM +0530, Ganapatrao Kulkarni wrote: > > > > > We are seeing regression with our uncore perf driver(Marvell's > > > ThunderX2, ARM64

Re: [PATCH] perf cgroups: Don't rotate events for cgroups unnecessarily

2019-08-23 Thread Ganapatrao Kulkarni
Hi Peter, On Fri, Aug 23, 2019 at 5:29 PM Peter Zijlstra wrote: > > > A: Because it messes up the order in which people normally read text. > Q: Why is top-posting such a bad thing? > A: Top-posting. > Q: What is the most annoying thing in e-mail? Sorry for the top-posting. > > On Fri, Aug 23,

Re: [PATCH] perf cgroups: Don't rotate events for cgroups unnecessarily

2019-08-23 Thread Peter Zijlstra
A: Because it messes up the order in which people normally read text. Q: Why is top-posting such a bad thing? A: Top-posting. Q: What is the most annoying thing in e-mail? On Fri, Aug 23, 2019 at 04:13:46PM +0530, Ganapatrao Kulkarni wrote: > We are seeing regression with our uncore perf

Re: [PATCH] perf cgroups: Don't rotate events for cgroups unnecessarily

2019-08-23 Thread Ganapatrao Kulkarni
Hi, We are seeing regression with our uncore perf driver(Marvell's ThunderX2, ARM64 server platform) on 5.3-Rc1. After bisecting, it turned out to be this patch causing the issue. Test case: Load module and run perf for more than 4 events( we have 4 counters, event multiplexing takes place for

Re: [PATCH] perf cgroups: Don't rotate events for cgroups unnecessarily

2019-06-27 Thread Ian Rogers
group_index On Mon, Jun 24, 2019 at 12:55 AM Peter Zijlstra wrote: > > On Fri, Jun 21, 2019 at 11:01:29AM -0700, Ian Rogers wrote: > > On Fri, Jun 21, 2019 at 1:24 AM Peter Zijlstra wrote: > > > > > > On Sat, Jun 01, 2019 at 01:27:22AM -0700, Ian Rogers wrote: > > > > @@ -3325,6 +3331,15 @@

Re: [PATCH] perf cgroups: Don't rotate events for cgroups unnecessarily

2019-06-24 Thread Peter Zijlstra
On Fri, Jun 21, 2019 at 11:01:29AM -0700, Ian Rogers wrote: > On Fri, Jun 21, 2019 at 1:24 AM Peter Zijlstra wrote: > > > > On Sat, Jun 01, 2019 at 01:27:22AM -0700, Ian Rogers wrote: > > > @@ -3325,6 +3331,15 @@ static int flexible_sched_in(struct perf_event > > > *event, void *data) > > >

Re: [PATCH] perf cgroups: Don't rotate events for cgroups unnecessarily

2019-06-21 Thread Ian Rogers
On Fri, Jun 21, 2019 at 1:24 AM Peter Zijlstra wrote: > > On Sat, Jun 01, 2019 at 01:27:22AM -0700, Ian Rogers wrote: > > @@ -3325,6 +3331,15 @@ static int flexible_sched_in(struct perf_event > > *event, void *data) > > sid->can_add_hw = 0; > > } > > > > + /* > >

Re: [PATCH] perf cgroups: Don't rotate events for cgroups unnecessarily

2019-06-21 Thread Peter Zijlstra
On Sat, Jun 01, 2019 at 01:27:22AM -0700, Ian Rogers wrote: > @@ -3325,6 +3331,15 @@ static int flexible_sched_in(struct perf_event *event, > void *data) > sid->can_add_hw = 0; > } > > + /* > + * If the group wasn't scheduled then set that multiplexing is

Re: [PATCH] perf cgroups: Don't rotate events for cgroups unnecessarily

2019-06-20 Thread Ian Rogers
On Fri, Jun 14, 2019 at 2:43 PM Liang, Kan wrote: > > > > On 6/14/2019 3:10 PM, Stephane Eranian wrote: > > On Thu, Jun 13, 2019 at 9:13 AM Liang, Kan > > wrote: > >> > >> > >> > >> On 6/1/2019 4:27 AM, Ian Rogers wrote: > >>> Currently perf_rotate_context assumes that if the context's

Re: [PATCH] perf cgroups: Don't rotate events for cgroups unnecessarily

2019-06-14 Thread Liang, Kan
On 6/14/2019 3:10 PM, Stephane Eranian wrote: On Thu, Jun 13, 2019 at 9:13 AM Liang, Kan wrote: On 6/1/2019 4:27 AM, Ian Rogers wrote: Currently perf_rotate_context assumes that if the context's nr_events != nr_active a rotation is necessary for perf event multiplexing. With cgroups,

Re: [PATCH] perf cgroups: Don't rotate events for cgroups unnecessarily

2019-06-14 Thread Stephane Eranian
On Thu, Jun 13, 2019 at 9:13 AM Liang, Kan wrote: > > > > On 6/1/2019 4:27 AM, Ian Rogers wrote: > > Currently perf_rotate_context assumes that if the context's nr_events != > > nr_active a rotation is necessary for perf event multiplexing. With > > cgroups, nr_events is the total count of events

Re: [PATCH] perf cgroups: Don't rotate events for cgroups unnecessarily

2019-06-13 Thread Liang, Kan
On 6/1/2019 4:27 AM, Ian Rogers wrote: Currently perf_rotate_context assumes that if the context's nr_events != nr_active a rotation is necessary for perf event multiplexing. With cgroups, nr_events is the total count of events for all cgroups and nr_active will not include events in a cgroup

[PATCH] perf cgroups: Don't rotate events for cgroups unnecessarily

2019-06-01 Thread Ian Rogers
Currently perf_rotate_context assumes that if the context's nr_events != nr_active a rotation is necessary for perf event multiplexing. With cgroups, nr_events is the total count of events for all cgroups and nr_active will not include events in a cgroup other than the current task's. This makes