Messages in this thread | | | Date | Wed, 12 Oct 2011 14:05:21 -0400 | From | Ben Blum <> | Subject | Re: [PATCH 3/4] threadgroup: extend threadgroup_lock() to cover exit and exec |
| |
On Wed, Oct 12, 2011 at 07:51:04PM +0200, Oleg Nesterov wrote: > Hi, > > On 10/10, Tejun Heo wrote: > > > > Hope you can still remember some > > of this one. :) > > I am not sure ;) > > > On Sun, Sep 18, 2011 at 07:37:23PM +0200, Oleg Nesterov wrote: > > > > With this change, threadgroup_lock() guarantees that the target > > > > threadgroup will remain stable - no new task will be added, no new > > > > PF_EXITING will be set and exec won't happen. > > > > > > To me, this is the only "contradictory" change, > > > > What do you mean "contradictory"? Can you please elaborate? > > Because, iirc, with this patch do_exit() does (almost) everything > under rw_sem. OK, down_read() should be cheap, but still. > > See also below. > > > > > + /* > > > > + * Release threadgroup and make sure we are holding no locks. > > > > + */ > > > > + threadgroup_change_done(tsk); > > > > > > I am wondering, can't we narrow the scope of threadgroup_change_begin/done > > > in do_exit() path? > > > > > > The code after 4/4 still has to check PF_EXITING, this is correct. And yes, > > > with this patch PF_EXITING becomes stable under ->group_rwsem. But, it seems, > > > we do not really need this? > > > > > > I mean, can't we change cgroup_exit() to do threadgroup_change_begin/done > > > instead? We do not really care about PF_EXITING, we only need to ensure that > > > we can't race with cgroup_exit(), right? > > > > If we confine our usage to cgroup, excluding just against > > cgroup_exit() might work although this is still a bit nasty. ie. some > > callbacks might not expect half torn-down tasks in methods other than > > the exit callback. > > Oh, sorry, I don't understand... I already forgot the details. > > > Also, it makes the mechanism unnecessarily cgroup-specific without > > gaining much if anything. > > Yes! And _personally_ I think it should be cgroup-specific, that is > why I dislike the very fact do_exit() uses it directly. To me it would > be cleaner to shift it into cgroup hooks. Yes, sure, this is subjective.
In the fork path, threadgroup_fork_read_...() is also called directly, not through cgroups. Would that change too?
> > In fact I still hope we can kill this sem altogether, but so far I have > no idea how we can do this. We do need the new per-process lock to > protect (in particular) ->thread_group. It is quite possible that it > should be rw_semaphore. But in this case we down_write(), not _read > in exit/fork paths, and its scope should be small.
I'm confused - taking a big rwsem for writing in the fork/exit paths?
The point here is that even though fork/exit modify thread_group, they are logical "readers" while cgroups is a "writer", since cgroups needs a stable view that excludes all fork/exit, but fork/exit can go together.
For clarity: In the fork path it is not meant to protect thread_group; it is meant to protect the window between cgroup_fork() and cgroup_post_fork().
> > I do not think the current lock should have more users. Of course I > can be wrong. And what exactly it protects? I mean copy_process(). > Almost everything, but this simply connects to cgroup fork hooks. > > Just my opinion, I am not going to insist. > > Oleg. > >
| |