Re: [PATCH] fix-flush_workqueue-vs-cpu_dead-race-update

[Ingo Molnar]

* Andrew Morton <akpm@xxxxxxxx> wrote:

> > FYI, the lock_cpu_hotplug() rewrite proposed by Gautham at 
> > http://lkml.org/lkml/2006/10/26/65 may still need refinement to 
> > avoid all the kind of deadlocks we have unearthed with workqueue 
> > example. I can review that design with Gautham if there is some 
> > interest to revive lock_cpu_hotplug() ..
> 
> Has anyone thought seriously about using the process freezer in the 
> cpu-down/cpu-up paths?  That way we don't need to lock anything 
> anywhere?

yes, yes, yes - lets please do that! The process freezer is already used 
for suspend, for hibernate and recently for kprobes - so its performance 
and robustness is being relied on and verified from multiple angles. I 
can see no reason why it couldnt be made really fast even on large 
boxes, if the need arises. (but even the current one is fast enough for 
any human-driven CPU hotplug stuff)

	Ingo
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/