linux-kernel - Re: [BUG] SLOB's krealloc() seems bust

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] [day] [month] [year] [list]

Date:	Wed, 08 Oct 2008 14:51:57 -0500
From:	Matt Mackall <mpm@...enic.com>
To:	Linus Torvalds <torvalds@...ux-foundation.org>
Cc:	Peter Zijlstra <a.p.zijlstra@...llo.nl>,
	Christoph Lameter <cl@...ux-foundation.org>,
	linux-mm <linux-mm@...ck.org>,
	Nick Piggin <nickpiggin@...oo.com.au>,
	Ingo Molnar <mingo@...e.hu>,
	linux-kernel <linux-kernel@...r.kernel.org>,
	akpm <akpm@...uxfoundation.org>,
	Pekka J Enberg <penberg@...helsinki.fi>
Subject: Re: [BUG] SLOB's krealloc() seems bust

On Tue, 2008-10-07 at 11:18 -0700, Linus Torvalds wrote:
> 
> On Tue, 7 Oct 2008, Peter Zijlstra wrote:
> 
> > On Tue, 2008-10-07 at 10:57 -0700, Linus Torvalds wrote:
> > 
> > > Peter - can you check with that
> > > 
> > > >  	if (slob_page(sp))
> > > > -		return ((slob_t *)block - 1)->units + SLOB_UNIT;
> > > > +		return (((slob_t *)block - 1)->units - 1) * SLOB_UNIT;
> > > 
> > > thing using
> > > 
> > > -		return ((slob_t *)block - 1)->units + SLOB_UNIT;
> > > +		return ((slob_t *)block - 1)->units * SLOB_UNIT;
> > > 
> > > instead? 
> > 
> > went splat on the second run...
> 
> Well, that makes it simple. I'll take Matt's patch as being "tested", and 
> somebody can hopefully explain where the extra unit comes from later.

Ok, I think we've gotten to the bottom of this. Here's an incremental
patch that doesn't work by dumb luck. Please apply.


SLOB: fix bogus ksize calculation fix

This fixes the previous fix, which was completely wrong on closer
inspection. This version has been manually tested with a user-space
test harness and generates sane values. A nearly identical patch has
been boot-tested.

The problem arose from changing how kmalloc/kfree handled alignment
padding without updating ksize to match. This brings it in sync.

Signed-off-by: Matt Mackall <mpm@...enic.com>

diff -r 3dd2424d4c32 -r 73d55a1b6c10 mm/slob.c
--- a/mm/slob.c	Tue Oct 07 23:00:11 2008 +0000
+++ b/mm/slob.c	Wed Oct 08 14:48:45 2008 -0500
@@ -514,9 +514,11 @@
 		return 0;
 
 	sp = (struct slob_page *)virt_to_page(block);
-	if (slob_page(sp))
-		return (((slob_t *)block - 1)->units - 1) * SLOB_UNIT;
-	else
+	if (slob_page(sp)) {
+		int align = max(ARCH_KMALLOC_MINALIGN, ARCH_SLAB_MINALIGN);
+		unsigned int *m = (unsigned int *)(block - align);
+		return SLOB_UNITS(*m) * SLOB_UNIT;
+	} else
 		return sp->page.private;
 }

-- 
Mathematics is the supreme nostalgia of our time.

--
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/