linux-kernel - Re: I just got got another Oops

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: <49B9E725.3060103@davidnewall.com>
Date:	Fri, 13 Mar 2009 15:25:01 +1030
From:	David Newall <davidn@...idnewall.com>
To:	Gene Heskett <gene.heskett@...il.com>
CC:	"linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>
Subject: Re: I just got got another Oops

Gene,

I think it's a null page pointer passed to prep_new_page.  It's
difficult to be certain, because it's compiled with -O2, but that's what
it looks like.

Let's test that hypothesis.  Let's also compile page_alloc.c with
debugging.  Do that by executing the commands in the attached file, and
then doing a make install (or whatever you do.)

But first, let's also sidestep the optimiser by making prep_new_page
public. Do that by changing:

   static int prep_new_page(struct page *page, int order, gfp_t gfp_flags)**

in mm/page_alloc.c, near line 637, to:

   int my_prep_new_page(struct page *page, int order, gfp_t gfp_flags)**

and by changing:

	if (prep_new_page(page, order, gfp_flags))

near line 1137, to:

	if (my_prep_new_page(page, order, gfp_flags))

This will prevent the optimiser from inlining that function, and we
would expect to see the error change from
"get_page_from_freelist+0x24b/0x4c0" to "my_prep_new_page+0x0yy/0x0zz".

David

View attachment "compile_page_alloc" of type "text/plain" (993 bytes)