Re: [PATCH 15/18] perf tools: Make the mmap length autotuning more robust

[Jiri Olsa <jolsa@redhat.com>]

On Tue, Dec 16, 2014 at 01:57:17PM -0300, Arnaldo Carvalho de Melo wrote:
> From: Arnaldo Carvalho de Melo <acme@redhat.com>
> 
> If /proc/sys/kernel/perf_event_mlock_kb is not (power of 2 + PAGE_SIZE_in_kb)
> and we let the perf tools do mmap length autosizing based on that, then, for
> non-CAP_IPC_LOCK users when /proc/sys/kernel/perf_event_paranoid is > -1, then
> we get an -EINVAL that ends up in:
> 
>   [acme@ssdandy linux]$ trace usleep 1
>   Invalid argument
>   [acme@ssdandy linux]$ perf record usleep 1
>   failed to mmap with 22 (Invalid argument)
> 
> After this fix:
> 
>   [acme@ssdandy linux]$ trace usleep 1
>   <SNIP>
>    0.806 ( 0.006 ms): munmap(addr: 0x7f7e4740a000, len: 66467) = 0
>    0.869 ( 0.002 ms): brk(                                   ) = 0x7bb000
>    0.873 ( 0.003 ms): brk(brk: 0x7dc000                      ) = 0x7dc000
>    0.877 ( 0.001 ms): brk(                                   ) = 0x7dc000
>    0.953 ( 0.058 ms): nanosleep(rqtp: 0x7fff26ab9420         ) = 0
>    0.959 ( 0.000 ms): exit_group(
>   [acme@ssdandy linux]$ perf record usleep 1
>   [ perf record: Woken up 1 times to write data ]
>   [ perf record: Captured and wrote 0.017 MB perf.data (~759 samples) ]
>   [acme@ssdandy linux]$
> 
> Cc: Adrian Hunter <adrian.hunter@intel.com>
> Cc: Borislav Petkov <bp@suse.de>
> Cc: David Ahern <dsahern@gmail.com>
> Cc: Don Zickus <dzickus@redhat.com>
> Cc: Frederic Weisbecker <fweisbec@gmail.com>
> Cc: Jiri Olsa <jolsa@redhat.com>
> Cc: Mike Galbraith <efault@gmx.de>
> Cc: Namhyung Kim <namhyung@kernel.org>
> Cc: Paul Mackerras <paulus@samba.org>
> Cc: Peter Zijlstra <peterz@infradead.org>
> Cc: Stephane Eranian <eranian@google.com>
> Link: http://lkml.kernel.org/n/tip-6p6l5ou6jev6o7ymc4nn1n2a@git.kernel.org
> Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
> ---
>  tools/perf/util/evlist.c | 2 ++
>  1 file changed, 2 insertions(+)
> 
> diff --git a/tools/perf/util/evlist.c b/tools/perf/util/evlist.c
> index 8eb92cf5d552..7d88327d94a9 100644
> --- a/tools/perf/util/evlist.c
> +++ b/tools/perf/util/evlist.c
> @@ -909,6 +909,8 @@ static size_t perf_evlist__mmap_size(unsigned long pages)
>  		}
>  
>  		pages = (max * 1024) / page_size;
> +		if (!is_power_of_2(pages))
> +			pages = rounddown_pow_of_two(pages);

ok then ;-)

jirka