Re: Fix signalfd interaction with thread-private signals

[Davide Libenzi <davidel@xmailserver.org>]

On Wed, 20 Jun 2007, Benjamin Herrenschmidt wrote:

> On Tue, 2007-06-19 at 19:15 -0700, Davide Libenzi wrote:
> 
> > Ok, why instead don't we go for something like the attached patch?
> > We exclude sync signals from signalfd, but we don't limit signalfd to 
> > shared signals. Ie, we should be able to fetch a signal sent with 
> > sys_tkill() to threads different from "current", that otherwise we would 
> > not be able to fetch.
> > Ben, sorry but my memory sucks ... the "notifier" thing was fine in that 
> > case, no?
> 
> I'm generally nervous about the idea of letting signalfd dequeue
> somebody else private signals... even if we filter out SEGV's and
> friends but I'll let Linus decide there.
> 
> Regarding the notifier, it's dodgy in most cases I'd say but I suppose
> it should be allright to only worry about "current" and not the target
> task there.

I believe that once we exclude synchronous signals from being dequeued, we 
should be fine. Limiting all signals sent with sys_tkill() from being 
dequeued with a signalfd is a too restricting behaviour IMO.


- Davide