linux-kernel - Re: sigaltstack fun (was Re: new execve/kernel

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

Message-ID: <20121121015302.GQ16916@ZenIV.linux.org.uk>
Date:	Wed, 21 Nov 2012 01:53:02 +0000
From:	Al Viro <viro@...IV.linux.org.uk>
To:	Linus Torvalds <torvalds@...ux-foundation.org>
Cc:	Michal Simek <monstr@...str.eu>,
	Linux Kernel Mailing List <linux-kernel@...r.kernel.org>,
	"linux-arch@...r.kernel.org" <linux-arch@...r.kernel.org>
Subject: Re: sigaltstack fun (was Re: new execve/kernel_thread design)

On Sun, Nov 18, 2012 at 08:45:43AM -1000, Linus Torvalds wrote:
> On Sat, Nov 17, 2012 at 7:45 PM, Al Viro <viro@...iv.linux.org.uk> wrote:
> >
> > Linus, do you have any objections to the above?  FWIW, I've a tentative
> > patchset in that direction (most of it from the last cycle); right now
> > it + stuff currently in signal.git#for-next is at -3.4KLoC and I hadn't
> > dealt with the biarch side of things yet...
> 
> I have absolutely no objections. sigaltstack has always been kind of
> messy, and made worse by the fact that it gets effectively no testing
> (because it's generally not used by normal code and even code that
> uses it tends to use it only for very uncommon events). So forcing all
> the sigaltstack code into generic code and at least avoiding the
> "different architectures can get things subtly - or not so subtly -
> wrong in different ways" sounds like a good thing.

OK...  Intermediate state (do_sigaltstack() guts still need to be cleaned up
and distributed between the 4 callers) is in signal.git#master right now;
_not_ for merge in that form.  Crap found in process:
	* microblaze, sparc64 and tile (compat side) have the same bug (see
above).  Fixed in that queue.
	* score, sh64 and openrisc all tried call do do_sigaltstack() passing
it an address of on-stack copy; no, set_fs() was not used.  Fixed in that queue.
	* alpha and c6x do not bother to restore saved altstack settings on
sigreturn.
	* c6x, cris and hexagon don't bother *saving* altstack settings on
signal arrival.  And yes, it does mean that cris and hexagon call
do_sigaltstack() on an unitialized chunk of userland stack.  Needs to be
fixed (and in a way that could go into -stable, unfortunately ;-/)

I hadn't done anything serious with altstack users yet; I'm afraid that
this code (i.e. sigframe allocation logics) will also bring a lot of
fun weirdness... ;-/
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/