Open MPI logo

Hardware Locality Development Mailing List Archives

  |   Home   |   Support   |   FAQ   |   all Hardware Locality Development mailing list

Subject: Re: [hwloc-devel] Cpuset problem
From: Brice Goglin (Brice.Goglin_at_[hidden])
Date: 2010-07-23 03:21:07


Let me know if this patch helps. It looks for the the noprefix option
and marks the mount point as cpuset instead of cgroup if found.
It also contains a couple cleanups that make the parsing safer.

Brice

Le 22/07/2010 19:30, Michael Raymond a écrit :
> /proc/mounts:
> cpuset /dev/cpuset cgroup
> rw,relatime,cpuset,noprefix,release_agent=/sbin/cpuset_release_agent 0 0
>
> The third element there is "cgroup" which the code specifically looks
> for in hwloc_find_linux_cpuset_mntpnt().
>
> /etc/fstab
> ..
> cpuset /dev/cpuset cpuset defaults
> 0 0
> ..
>
>
>
> Brice Goglin wrote:
>
>> Le 22/07/2010 14:53, Michael Raymond a écrit :
>>
>>> I was doing some testing on what's in SVN and I found that the
>>> topology discovery code has a problem when running within a cpuset. On
>>> my 2.6.32 SLES11SP1 box the code calls hwloc_read_linux_cpuset_mask()
>>> with a cgroup mount point and tries to read
>>> /dev/cpuset/<blah>/cpuset.[cpus|mems]. Those files don't exist but
>>> <blah>/[cpus|mems] do. If I change the code to use those files instead
>>> then everything else works.
>>>
>>>
>> Hello Michael,
>> This probably means that we found a cgroup mount point in /dev/cpuset
>> while its actually a cpuset mount point. What does your /proc/mounts
>> contain? Can you try to understand why hwloc_find_linux_cpuset_mntpnt
>> would find a wrong mount type?
>> thanks,
>> Brice
>>
>