linux-kernel - Re: Signal handling in a page fault handler

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [day] [month] [year] [list]

Date:   Tue, 3 Apr 2018 06:48:54 -0700
From:   Matthew Wilcox <willy@...radead.org>
To:     Chris Wilson <chris@...is-wilson.co.uk>
Cc:     dri-devel@...ts.freedesktop.org, linux-mm@...ck.org,
        Souptick Joarder <jrdr.linux@...il.com>,
        linux-kernel@...r.kernel.org
Subject: Re: Signal handling in a page fault handler

On Tue, Apr 03, 2018 at 02:20:43PM +0100, Chris Wilson wrote:
> Quoting Matthew Wilcox (2018-04-03 14:10:25)
> > On Tue, Apr 03, 2018 at 01:33:15PM +0100, Chris Wilson wrote:
> > > Quoting Matthew Wilcox (2018-04-02 15:10:58)
> > > > I don't think the graphics drivers really want to be interrupted by
> > > > any signal.
> > > 
> > > Assume the worst case and we may block for 10s. Even a 10ms delay may be
> > > unacceptable to some signal handlers (one presumes). For the number one
> > > ^C usecase, yes that may be reduced to only bother if it's killable, but
> > > I wonder if there are not timing loops (e.g. sigitimer in Xorg < 1.19)
> > > that want to be able to interrupt random blockages.
> > 
> > Ah, setitimer / SIGALRM.  So what do we want to have happen if that
> > signal handler touches the mmaped device memory?
> 
> Burn in a great ball of fire :) Isn't that what usually happens if you
> do anything in a signal handler?

I don't know.  My mummy and daddy don't let me play with sharp things
like signals.

> Hmm, if SIGBUS has a handler does that count as a killable signal? The
> ddx does have code to service SIGBUS emitted when accessing the mmapped
> pointer that may result from the page insertion failing with no memory
> (or other random error). There we stop accessing via the pointer and
> use another indirect method.

Any signal with a handler is non-fatal, and so a call to
mutex_lock_killable() would not return if SIGBUS was delivered to a thread
blocking in a page fault.  mutex_lock_interruptible() would return -EINTR.