Messages in this thread Patch in this message | | | Subject | Re: [PATCH 1/4] add task handling notifier: base definitions | From | Matt Helsley <> | Date | Wed, 23 Jan 2008 11:11:25 -0800 |
| |
On Wed, 2008-01-09 at 09:46 +0000, Jan Beulich wrote: > >> +BLOCKING_NOTIFIER_HEAD(task_notifier_list); > >> +EXPORT_SYMBOL_GPL(task_notifier_list); > >> +ATOMIC_NOTIFIER_HEAD(atomic_task_notifier_list); > >> +EXPORT_SYMBOL_GPL(atomic_task_notifier_list); > >> + > > > >When these global notifier lists were proposed years ago folks at SGI > >loudly objected with concerns over anticipated cache line bouncing on > >512+ cpu machines. Is that no longer a concern? > > I can't see an alternative, since the serialization is unavoidable.
There are definitely alternatives. Naturally they are all much more complex than using a single notifier chain. You could do per-cpu chains for example (ugly, I know..).
> >> @@ -121,6 +127,9 @@ void __put_task_struct(struct task_struc > >> WARN_ON(atomic_read(&tsk->usage)); > >> WARN_ON(tsk == current); > >> > >> + atomic_notifier_call_chain(&atomic_task_notifier_list, > >> + TASK_DELETE, tsk); > >> + > >> security_task_free(tsk); > >> free_uid(tsk->user); > >> put_group_info(tsk->group_info); > > > >Would the atomic notifier call chain be necessary if you hooked into an > >earlier section of do_exit() instead? > > I'm afraid it is, as I was told that sleeping in the do_exit() path is not > generally possible. > > Jan
Odd. Last I checked I thought I saw a bunch of calls in do_exit() that could sleep. Only in certain sections and at the end did it appear to prevent sleeping.
Sorry for the late reply.
Cheers, -Matt Helsley
| |