Re: [PATCH] perf/core: Enable the bp only if the .disable field is 0.

From: Milind Chabbi <chabbi.milind@gmail.com>
To: Milind Chabbi <chabbi.milind@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>,
	Peter Zijlstra <a.p.zijlstra@chello.nl>,
	Paul Mackerras <paulus@samba.org>, Ingo Molnar <mingo@redhat.com>,
	Arnaldo Carvalho de Melo <acme@ghostprotocols.net>,
	linux-kernel@vger.kernel.org
Subject: Re: [PATCH] perf/core: Enable the bp only if the .disable field is 0.
Date: Sun, 26 Nov 2017 22:50:27 -0800	[thread overview]
Message-ID: <CAMmz+YnSSn3t=qSjfMX5H5Mp6QuTePkuLMjV7FO7at8z4TXu1Q@mail.gmail.com> (raw)
In-Reply-To: <1511765030-20132-1-git-send-email-chabbi.milind@gmail.com>

On Sun, Nov 26, 2017 at 10:43 PM, Milind Chabbi <chabbi.milind@gmail.com> wrote:
> Signed-off-by: Milind Chabbi <chabbi.milind@gmail.com>
> ---
>  kernel/events/core.c | 3 ++-
>  1 file changed, 2 insertions(+), 1 deletion(-)
>
> diff --git a/kernel/events/core.c b/kernel/events/core.c
> index 35747a58ffb4..1b8eae85e9de 100644
> --- a/kernel/events/core.c
> +++ b/kernel/events/core.c
> @@ -2659,7 +2659,8 @@ static int perf_event_modify_breakpoint(struct perf_event *bp,
>                 return err;
>         }
>
> -       _perf_event_enable(bp);
> +       if (!attr->disabled)
> +               _perf_event_enable(bp);
>         return 0;
>  }
>
> --
> 2.14.1
>

Hi Jirka,
Thanks for your changes for proper accounting of the bp.
This additional change is needed so that we do not enable the bp if
the user has not asked to enable it.
I did the testing for ioctl and it continues to show the significant
speedups that I had originally seen.

-Milind