Re: [PATCH] Fix signal race during process exit

[Rusty Russell]

On Thu, 2004-06-10 at 11:48, Roland McGrath wrote:
> The concern I had is basically that this might not be true of all cases.
> The only problem case that has come up is the current task's own interrupt
> handlers calling signal code while interrupting release_task.  I know for
> the case of posix-timers it's not an issue because their cleanup is handled
> with special synchronization in __exit_signal.  What I'm not sure about is
> all other sources of asynchronous signals that use task_struct pointers
> rather than PID lookups and so might do one while release_task is in progress.
> e.g. async IO signals triggered via driver interrupts, etc.

Yes.  In 2.4 we explicitly checked in the signal code.

Why don't we do the sane thing and just do release_task() from
__put_task_struct(), rather than the current two-stage thing?

Rusty.
-- 
Anyone who quotes me in their signature is an idiot -- Rusty Russell

-
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/