From: Eric Sesterhenn <snakebyte@gmx.de>
To: "Jörn Engel" <joern@logfs.org>
Cc: Tom Rini <trini@kernel.crashing.org>,
chris@zankel.net, phillip@lougher.demon.co.uk,
linuxppc-dev@ozlabs.org, rpurdie@rpsys.net,
linux-fsdevel@vger.kernel.org
Subject: Re: [Patch] NULL pointer deref with corrupted squashfs image
Date: Tue, 20 Jan 2009 17:47:14 +0100 [thread overview]
Message-ID: <20090120163957.GA21339@alice> (raw)
In-Reply-To: <20090117134916.GA29641@logfs.org>
* J=C3=B6rn Engel (joern@logfs.org) wrote:
> On Fri, 16 January 2009 16:07:00 -0700, Tom Rini wrote:
> >=20
> > Sounds like a plan to me, except maybe zlib_inflate_unsafe() and a
> > comment above the wrapper saying what/why is going on?
>=20
> Eric, will you do the honors? Since you did all the hard work before,
> you derserve the fame as well. :)
Since I am not sure either about xtensa I added chris to the cc list.
Some callees of zlib_inflate() might accidentally pass a NULL
pointer in strm->next_out which zlib_inflate() should catch.
Others like the powerpc gunzip_partial expect to be able
to extract a zImage to memory location 0. This introduces
zlib_inflate_usafe() for those and adds a check for the rest
Signed-off-by: Eric Sesterhenn <snakebyte@gmx.de>
--- linux/arch/powerpc/boot/gunzip_util.c.orig 2009-01-20 10:23:11.00000000=
0 +0100
+++ linux/arch/powerpc/boot/gunzip_util.c 2009-01-20 10:23:31.000000000 +01=
00
@@ -109,7 +109,7 @@ int gunzip_partial(struct gunzip_state *
=20
state->s.next_out =3D dst;
state->s.avail_out =3D dstlen;
- r =3D zlib_inflate(&state->s, Z_FULL_FLUSH);
+ r =3D zlib_inflate_unsafe(&state->s, Z_FULL_FLUSH);
if (r !=3D Z_OK && r !=3D Z_STREAM_END)
fatal("inflate returned %d msg: %s\n\r", r, state->s.msg);
len =3D state->s.next_out - (unsigned char *)dst;
--- linux/include/linux/zlib.h.orig 2009-01-20 10:11:33.000000000 +0100
+++ linux/include/linux/zlib.h 2009-01-20 10:19:56.000000000 +0100
@@ -329,7 +329,23 @@ extern int zlib_inflateInit (z_streamp s
*/
=20
=20
-extern int zlib_inflate (z_streamp strm, int flush);
+extern int __zlib_inflate (z_streamp strm, int flush, int check_out);
+/*
+ These two wrappers decide wheter strm->next_out gets checked for NULL.
+ The zlib_inflate_unsafe() version got added because the PPC zImage
+ gets extracted to memory address 0 and therefore
+ we avoid this check for zlib_inflate_unsafe()
+*/
+static inline int zlib_inflate(z_streamp strm, int flush)
+{
+ return __zlib_inflate(strm, flush, 1);
+}
+
+static inline int zlib_inflate_unsafe(z_streamp strm, int flush)
+{
+ return __zlib_inflate(strm, flush, 0);
+}
+
/*
inflate decompresses as much data as possible, and stops when the input
buffer becomes empty or the output buffer becomes full. It may introduce
--- linux/lib/zlib_inflate/inflate_syms.c.orig 2009-01-20 10:22:21.00000000=
0 +0100
+++ linux/lib/zlib_inflate/inflate_syms.c 2009-01-20 10:22:32.000000000 +01=
00
@@ -11,7 +11,7 @@
#include <linux/zlib.h>
=20
EXPORT_SYMBOL(zlib_inflate_workspacesize);
-EXPORT_SYMBOL(zlib_inflate);
+EXPORT_SYMBOL(__zlib_inflate);
EXPORT_SYMBOL(zlib_inflateInit2);
EXPORT_SYMBOL(zlib_inflateEnd);
EXPORT_SYMBOL(zlib_inflateReset);
--- linux/lib/zlib_inflate/inflate.c.orig 2009-01-20 10:20:14.000000000 +01=
00
+++ linux/lib/zlib_inflate/inflate.c 2009-01-20 10:22:03.000000000 +0100
@@ -329,7 +329,7 @@ static int zlib_inflateSyncPacket(z_stre
will return Z_BUF_ERROR if it has not reached the end of the stream.
*/
=20
-int zlib_inflate(z_streamp strm, int flush)
+int __zlib_inflate(z_streamp strm, int flush, int check_out)
{
struct inflate_state *state;
const unsigned char *next; /* next input */
@@ -347,8 +347,10 @@ int zlib_inflate(z_streamp strm, int flu
static const unsigned short order[19] =3D /* permutation of code lengt=
hs */
{16, 17, 18, 0, 8, 7, 9, 6, 10, 5, 11, 4, 12, 3, 13, 2, 14, 1, 15};
=20
- /* Do not check for strm->next_out =3D=3D NULL here as ppc zImage
- inflates to strm->next_out =3D 0 */
+ /* We only check strm->next_out if the callee requests it,
+ since ppc extracts the ppc zImage to 0 */
+ if (check_out && !strm->next_out)
+ return Z_STREAM_ERROR;
=20
if (strm =3D=3D NULL || strm->state =3D=3D NULL ||
(strm->next_in =3D=3D NULL && strm->avail_in !=3D 0))
WARNING: multiple messages have this Message-ID (diff)
From: Eric Sesterhenn <snakebyte@gmx.de>
To: "Jörn Engel" <joern@logfs.org>
Cc: Tom Rini <trini@kernel.crashing.org>,
phillip@lougher.demon.co.uk, linux-fsdevel@vger.kernel.org,
jacmet@sunsite.dk, rpurdie@rpsys.net, linuxppc-dev@ozlabs.org,
chris@zankel.net
Subject: Re: [Patch] NULL pointer deref with corrupted squashfs image
Date: Tue, 20 Jan 2009 17:47:14 +0100 [thread overview]
Message-ID: <20090120163957.GA21339@alice> (raw)
In-Reply-To: <20090117134916.GA29641@logfs.org>
* Jörn Engel (joern@logfs.org) wrote:
> On Fri, 16 January 2009 16:07:00 -0700, Tom Rini wrote:
> >
> > Sounds like a plan to me, except maybe zlib_inflate_unsafe() and a
> > comment above the wrapper saying what/why is going on?
>
> Eric, will you do the honors? Since you did all the hard work before,
> you derserve the fame as well. :)
Since I am not sure either about xtensa I added chris to the cc list.
Some callees of zlib_inflate() might accidentally pass a NULL
pointer in strm->next_out which zlib_inflate() should catch.
Others like the powerpc gunzip_partial expect to be able
to extract a zImage to memory location 0. This introduces
zlib_inflate_usafe() for those and adds a check for the rest
Signed-off-by: Eric Sesterhenn <snakebyte@gmx.de>
--- linux/arch/powerpc/boot/gunzip_util.c.orig 2009-01-20 10:23:11.000000000 +0100
+++ linux/arch/powerpc/boot/gunzip_util.c 2009-01-20 10:23:31.000000000 +0100
@@ -109,7 +109,7 @@ int gunzip_partial(struct gunzip_state *
state->s.next_out = dst;
state->s.avail_out = dstlen;
- r = zlib_inflate(&state->s, Z_FULL_FLUSH);
+ r = zlib_inflate_unsafe(&state->s, Z_FULL_FLUSH);
if (r != Z_OK && r != Z_STREAM_END)
fatal("inflate returned %d msg: %s\n\r", r, state->s.msg);
len = state->s.next_out - (unsigned char *)dst;
--- linux/include/linux/zlib.h.orig 2009-01-20 10:11:33.000000000 +0100
+++ linux/include/linux/zlib.h 2009-01-20 10:19:56.000000000 +0100
@@ -329,7 +329,23 @@ extern int zlib_inflateInit (z_streamp s
*/
-extern int zlib_inflate (z_streamp strm, int flush);
+extern int __zlib_inflate (z_streamp strm, int flush, int check_out);
+/*
+ These two wrappers decide wheter strm->next_out gets checked for NULL.
+ The zlib_inflate_unsafe() version got added because the PPC zImage
+ gets extracted to memory address 0 and therefore
+ we avoid this check for zlib_inflate_unsafe()
+*/
+static inline int zlib_inflate(z_streamp strm, int flush)
+{
+ return __zlib_inflate(strm, flush, 1);
+}
+
+static inline int zlib_inflate_unsafe(z_streamp strm, int flush)
+{
+ return __zlib_inflate(strm, flush, 0);
+}
+
/*
inflate decompresses as much data as possible, and stops when the input
buffer becomes empty or the output buffer becomes full. It may introduce
--- linux/lib/zlib_inflate/inflate_syms.c.orig 2009-01-20 10:22:21.000000000 +0100
+++ linux/lib/zlib_inflate/inflate_syms.c 2009-01-20 10:22:32.000000000 +0100
@@ -11,7 +11,7 @@
#include <linux/zlib.h>
EXPORT_SYMBOL(zlib_inflate_workspacesize);
-EXPORT_SYMBOL(zlib_inflate);
+EXPORT_SYMBOL(__zlib_inflate);
EXPORT_SYMBOL(zlib_inflateInit2);
EXPORT_SYMBOL(zlib_inflateEnd);
EXPORT_SYMBOL(zlib_inflateReset);
--- linux/lib/zlib_inflate/inflate.c.orig 2009-01-20 10:20:14.000000000 +0100
+++ linux/lib/zlib_inflate/inflate.c 2009-01-20 10:22:03.000000000 +0100
@@ -329,7 +329,7 @@ static int zlib_inflateSyncPacket(z_stre
will return Z_BUF_ERROR if it has not reached the end of the stream.
*/
-int zlib_inflate(z_streamp strm, int flush)
+int __zlib_inflate(z_streamp strm, int flush, int check_out)
{
struct inflate_state *state;
const unsigned char *next; /* next input */
@@ -347,8 +347,10 @@ int zlib_inflate(z_streamp strm, int flu
static const unsigned short order[19] = /* permutation of code lengths */
{16, 17, 18, 0, 8, 7, 9, 6, 10, 5, 11, 4, 12, 3, 13, 2, 14, 1, 15};
- /* Do not check for strm->next_out == NULL here as ppc zImage
- inflates to strm->next_out = 0 */
+ /* We only check strm->next_out if the callee requests it,
+ since ppc extracts the ppc zImage to 0 */
+ if (check_out && !strm->next_out)
+ return Z_STREAM_ERROR;
if (strm == NULL || strm->state == NULL ||
(strm->next_in == NULL && strm->avail_in != 0))
--
To unsubscribe from this list: send the line "unsubscribe linux-fsdevel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
next prev parent reply other threads:[~2009-01-20 16:47 UTC|newest]
Thread overview: 18+ messages / expand[flat|nested] mbox.gz Atom feed top
2009-01-13 12:40 Bug with corrupted squashfs image Eric Sesterhenn
2009-01-16 17:45 ` [Patch] NULL pointer deref " Eric Sesterhenn
2009-01-16 19:07 ` Jörn Engel
2009-01-16 23:07 ` Tom Rini
2009-01-17 13:49 ` Jörn Engel
2009-01-17 19:38 ` Eric Sesterhenn
2009-01-20 16:47 ` Eric Sesterhenn [this message]
2009-01-20 16:47 ` Eric Sesterhenn
2009-01-20 17:57 ` Jörn Engel
2009-01-20 17:57 ` Jörn Engel
2009-01-20 18:47 ` Tom Rini
2009-01-20 18:47 ` Tom Rini
2009-01-21 8:34 ` Eric Sesterhenn
2009-01-21 8:34 ` Eric Sesterhenn
2009-01-21 12:31 ` Phillip Lougher
2009-01-21 12:31 ` Phillip Lougher
2009-01-22 2:48 ` Phillip Lougher
2009-01-22 9:46 ` Jörn Engel
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20090120163957.GA21339@alice \
--to=snakebyte@gmx.de \
--cc=chris@zankel.net \
--cc=joern@logfs.org \
--cc=linux-fsdevel@vger.kernel.org \
--cc=linuxppc-dev@ozlabs.org \
--cc=phillip@lougher.demon.co.uk \
--cc=rpurdie@rpsys.net \
--cc=trini@kernel.crashing.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.