Messages in this thread |  | | | Date | Sun, 27 Jul 2008 23:37:46 -0500 | | From | Paul Jackson <> | | Subject | Re: [PATCH 1/3] cpuset: clean up cpuset hierarchy traversal code | |
Li Zefan wrote:
> + list_add(&top_cpuset.stack_list, &q);
Can you figure out if we are holding any system wide lock, such as
cgroup_mutex, at this point? There is only one top_cpuset.stack_list
in the system, so I would think that we need to single thread its usage
somehow.
The hotplug hooks, such as common_cpu_mem_hotplug_unplug() that were
added sometime in the last few months (while the so called maintainer,
who's initials seem to be 'pj', was asleep ;) don't describe what
locking applies to them very well, and they call rebuild_sched_domains(),
where the above line of code lives.
--
I won't rest till it's the best ...
Programmer, Linux Scalability
Paul Jackson <pj@sgi.com> 1.940.382.4214
|  |