[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <CAJpZYjVrTsuSwT-31hBYihzmK94j6BWGD4XxudTXMB0W+Hcb+g@mail.gmail.com>
Date: Tue, 5 Nov 2024 16:32:22 -0800
From: Chun-Tse Shao <ctshao@...gle.com>
To: Namhyung Kim <namhyung@...nel.org>
Cc: linux-kernel@...r.kernel.org, Peter Zijlstra <peterz@...radead.org>,
Ingo Molnar <mingo@...hat.com>, Arnaldo Carvalho de Melo <acme@...nel.org>, Mark Rutland <mark.rutland@....com>,
Alexander Shishkin <alexander.shishkin@...ux.intel.com>, Jiri Olsa <jolsa@...nel.org>,
Ian Rogers <irogers@...gle.com>, Adrian Hunter <adrian.hunter@...el.com>,
Liang Kan <kan.liang@...ux.intel.com>, Ze Gao <zegao2021@...il.com>,
Weilin Wang <weilin.wang@...el.com>, linux-perf-users@...r.kernel.org
Subject: Re: [PATCH v2 2/3] perf: Reveal PMU type in fdinfo
Thanks for your review, Namhyung! Here is the link to patch v3:
https://lore.kernel.org/all/20241106003007.2112584-2-ctshao@google.com/
On Mon, Nov 4, 2024 at 4:12 PM Namhyung Kim <namhyung@...nel.org> wrote:
>
> On Fri, Nov 01, 2024 at 09:17:56PM +0000, Chun-Tse Shao wrote:
> > It gives useful info on knowing which PMUs are reserved by this process.
> > Also add extra attributes which would be useful.
> >
> > ```
> > Testing cycles
> > $ ./perf stat -e cycles &
> > $ cat /proc/`pidof perf`/fdinfo/3
> > pos: 0
> > flags: 02000002
> > mnt_id: 16
> > ino: 3081
> > perf_event-attr.type: 0
>
> Maybe 'perf_event_attr' would be appropriate as it's the name of the
> struct.
>
Agree.
> > perf_event-attr.config: 0
> > perf_event-attr.config1: 0
> > perf_event-attr.config2: 0
> > perf_event-attr.config3: 0
>
> It's hard to pick which fields to show here but I'd say that those
> config[123] are not used frequently at least for regular events.
> Maybe just showing type and config is fine.
>
Agree, it should be fine to add more if needed.
> >
> > Testing L1-dcache-load-misses//
> > $ ./perf stat -e L1-dcache-load-misses &
> > $ cat /proc/`pidof perf`/fdinfo/3
> > pos: 0
> > flags: 02000002
> > mnt_id: 16
> > ino: 1072
> > perf_event-attr.type: 3
> > perf_event-attr.config: 65536
> > perf_event-attr.config1: 0
> > perf_event-attr.config2: 0
> > perf_event-attr.config3: 0
> > ```
> >
> > Signed-off-by: Chun-Tse Shao <ctshao@...gle.com>
> > ---
> > kernel/events/core.c | 13 +++++++++++++
> > 1 file changed, 13 insertions(+)
>
> FYI usually the kernel changes are applied to a different tree than the
> tools.
>
Thanks for the heads up!
> >
> > diff --git a/kernel/events/core.c b/kernel/events/core.c
> > index cdd09769e6c56..c950b6fc92cda 100644
> > --- a/kernel/events/core.c
> > +++ b/kernel/events/core.c
> > @@ -55,6 +55,7 @@
> > #include <linux/pgtable.h>
> > #include <linux/buildid.h>
> > #include <linux/task_work.h>
> > +#include <linux/seq_file.h>
> >
> > #include "internal.h"
> >
> > @@ -6820,6 +6821,17 @@ static int perf_fasync(int fd, struct file *filp, int on)
> > return 0;
> > }
> >
> > +static void perf_show_fdinfo(struct seq_file *m, struct file *f)
> > +{
> > + struct perf_event *event = f->private_data;
> > +
> > + seq_printf(m, "perf_event-attr.type:\t%u\n", event->orig_type);
> > + seq_printf(m, "perf_event-attr.config:\t%llu\n", event->attr.config);
>
> I'm not sure if all archs are happy with treating it as %llu.
>
I think with type casting to `unsigned long long`, it will make it
print in `%llu` for all archs.
> Thanks,
> Namhyung
>
>
> > + seq_printf(m, "perf_event-attr.config1:\t%llu\n", event->attr.config1);
> > + seq_printf(m, "perf_event-attr.config2:\t%llu\n", event->attr.config2);
> > + seq_printf(m, "perf_event-attr.config3:\t%llu\n", event->attr.config3);
> > +}
> > +
> > static const struct file_operations perf_fops = {
> > .release = perf_release,
> > .read = perf_read,
> > @@ -6828,6 +6840,7 @@ static const struct file_operations perf_fops = {
> > .compat_ioctl = perf_compat_ioctl,
> > .mmap = perf_mmap,
> > .fasync = perf_fasync,
> > + .show_fdinfo = perf_show_fdinfo,
> > };
> >
> > /*
> > --
> > 2.47.0.163.g1226f6d8fa-goog
> >
Powered by blists - more mailing lists