Re: more signal locking bugs?

[Linus Torvalds]

On Sun, 16 Feb 2003, Manfred Spraul wrote:
> But these lines are not in 2.4 or 2.5.61.
> The current rule to nesting tasklist_lock and task_lock is
> - read_lock(&tasklist_lock) and task_lock can be mixed in any order.
> - write_lock_irq(&tasklist_lock) and task_lock are incompatible.

Oh, you're right, and you're right exactly _because_ "task->signal" isn't
protected by the task lock right now. Aurgh. I had already mentally done
that protection, which is why I thought we already had the bug.

So never mind. 2.4.x is obviously also ok.

> What about this minimal patch? The performance critical operation is 
> signal delivery - we should fix the synchronization between signal 
> delivery and exec first.

The patch looks ok, although I'd also remove the locking and testing from
collect_sigign_sigcatch() once it is done at a higher level.

And yeah, what about signal delivery? Put back the same lock there?

		Linus