Re: [34-longterm 007/196] perf: Better fit max unprivileged mlockpages for tools needs

[Stephane Eranian]

On Tue, Mar 13, 2012 at 1:12 AM, Paul Gortmaker
<paul.gortmaker@xxxxxxxxxxxxx> wrote:
> From: Frederic Weisbecker <fweisbec@xxxxxxxxx>
>
> Â Â Â Â Â Â Â Â Â -------------------
> Â ÂThis is a commit scheduled for the next v2.6.34 longterm release.
> Â ÂIf you see a problem with using this for longterm, please comment.
> Â Â Â Â Â Â Â Â Â -------------------
>
> commit 880f57318450dbead6a03f9e31a1468924d6dd88 upstream.
>
> The maximum kilobytes of locked memory that an unprivileged user
> can reserve is of 512 kB = 128 pages by default, scaled to the
> number of onlined CPUs, which fits well with the tools that use
> 128 data pages by default.
>
> However tools actually use 129 pages, because they need one more
> for the user control page. Thus the default mlock threshold is
> not sufficient for the default tools needs and we always end up
> to evaluate the constant mlock rlimit policy, which doesn't have
> this scaling with the number of online CPUs.
>
> Hence, on systems that have more than 16 CPUs, we overlap the
> rlimit threshold and fail to mmap:
>
> Â Â Â Â$ perf record ls
> Â Â Â ÂError: failed to mmap with 1 (Operation not permitted)
>
But isn't that simply coming from the fact that perf chose to request
128 pages by default? I don't know how this size was chosen. If perf
were to allocate 127 (data) pages, you would not have that problem
with the current default limit, would you?


> Just increase the max unprivileged mlock threshold by one page
> so that it supports well perf tools even after 16 CPUs.
>
> Reported-by: Han Pingtian <phan@xxxxxxxxxx>
> Reported-by: Peter Zijlstra <a.p.zijlstra@xxxxxxxxx>
> Reported-by: Arnaldo Carvalho de Melo <acme@xxxxxxxxxx>
> Signed-off-by: Frederic Weisbecker <fweisbec@xxxxxxxxx>
> Acked-by: Arnaldo Carvalho de Melo <acme@xxxxxxxxxx>
> Cc: Stephane Eranian <eranian@xxxxxxxxxx>
> LKML-Reference: <1300904979-5508-1-git-send-email-fweisbec@xxxxxxxxx>
> Signed-off-by: Ingo Molnar <mingo@xxxxxxx>
> Signed-off-by: Paul Gortmaker <paul.gortmaker@xxxxxxxxxxxxx>
> ---
> Âkernel/perf_event.c | Â Â3 ++-
> Â1 file changed, 2 insertions(+), 1 deletion(-)
>
> diff --git a/kernel/perf_event.c b/kernel/perf_event.c
> index 2357b19..b203546 100644
> --- a/kernel/perf_event.c
> +++ b/kernel/perf_event.c
> @@ -57,7 +57,8 @@ static atomic_t nr_task_events __read_mostly;
> Â*/
> Âint sysctl_perf_event_paranoid __read_mostly = 1;
>
> -int sysctl_perf_event_mlock __read_mostly = 512; /* 'free' kb per user */
> +/* Minimum for 128 pages + 1 for the user control page */
> +int sysctl_perf_event_mlock __read_mostly = 516; /* 'free' kb per user */
>
> Â/*
> Â* max perf event sample rate
> --
> 1.7.9.3
>
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/