Messages in this thread | | | Date | Wed, 17 Dec 2014 11:00:42 -0300 | From | Arnaldo Carvalho de Melo <> | Subject | Re: [PATCH 05/18] perf evlist: Do not use hard coded value for a mmap_pages default |
| |
Em Wed, Dec 17, 2014 at 02:23:30PM +0100, Jiri Olsa escreveu: > On Tue, Dec 16, 2014 at 01:57:07PM -0300, Arnaldo Carvalho de Melo wrote: > > + if (pages == UINT_MAX) { > > + int max; > > + > > + if (sysctl__read_int("kernel/perf_event_mlock_kb", &max) < 0) { > > + /* > > + * Pick a once upon a time good value, i.e. things look > > + * strange since we can't read a sysctl value, but lets not > > + * die yet... > > + */ > > + max = 512; > > + } else { > > + max -= (page_size / 1024); > > + }
> so this way you depend on value in perf_event_mlock_kb being power 2, > otherwise:
> # echo 1000 > /proc/sys/kernel/perf_event_mlock_kb > $ perf record ls > failed to mmap with 22 (Invalid argument)
Right, I thought that perf_event_mlock_kb input was validated, realized that it is not, see the following patchkit that makes it validate that after reading, using rounddown_pow_of_two if not.
- Arnaldo
> maybe you could use the logic/code from parse_pages_arg function > > jirka
| |