* [PATCH] partitions: aix: bound the pp_count scan to the ppe array
@ 2026-06-07 6:41 Bryam Vargas
2026-06-08 11:46 ` Philippe De Muyter
2026-06-08 14:34 ` Jens Axboe
0 siblings, 2 replies; 3+ messages in thread
From: Bryam Vargas @ 2026-06-07 6:41 UTC (permalink / raw)
To: Jens Axboe
Cc: Philippe De Muyter, Kees Cook, Michael Bommarito, linux-block,
linux-kernel
aix_partition() reads the physical volume descriptor into a fixed-size
struct pvd and then scans its physical-partition-extent array:
int numpps = be16_to_cpu(pvd->pp_count);
...
for (i = 0; i < numpps; i += 1) {
struct ppe *p = pvd->ppe + i;
...
lp_ix = be16_to_cpu(p->lp_ix);
pvd points at a single kmalloc()'d struct pvd whose ppe[] member holds a
fixed ARRAY_SIZE(pvd->ppe) (1016) entries, but the loop runs up to the
on-disk pp_count. pp_count is an unvalidated __be16 read straight from
the descriptor, so a crafted AIX image with pp_count larger than 1016
drives the loop to read pvd->ppe[i] past the end of the allocation (up to
65535 entries, ~2 MB out of bounds).
The partition scan runs without mounting anything, when a block device
with a crafted AIX/IBM partition table appears (an attacker-supplied
image attached with losetup -P, or a device auto-scanned by udev), via
msdos_partition() -> aix_partition().
Clamp the scan to the number of entries the ppe[] array can hold.
Fixes: 6ceea22bbbc8 ("partitions: add aix lvm partition support files")
Cc: stable@vger.kernel.org
Signed-off-by: Bryam Vargas <hexlabsecurity@proton.me>
---
Reproduced on v7.1-rc6 with KASAN (CONFIG_PARTITION_ADVANCED +
CONFIG_AIX_PARTITION). A crafted disk image whose AIX/IBM partition table
sets pp_count to 0xffff, attached with `losetup -fP image.img` (in-kernel
partition scan, no mount), is reported by KASAN:
BUG: KASAN: slab-out-of-bounds in aix_partition+0xb6e/0xee0
Read of size 2 at addr ... by task losetup
aix_partition
msdos_partition
bdev_disk_changed
loop_reread_partitions
loop_configure
lo_ioctl
__x64_sys_ioctl
i.e. a read past the end of the kmalloc(sizeof(struct pvd)) object. A control
image with pp_count == 1016 (== ARRAY_SIZE(pvd->ppe)) is clean. With this
patch the crafted image is parsed with no out-of-bounds access.
This is the read-loop sibling of the lvd scan bounded by Michael Bommarito's
"partitions: aix: bound the lvd scan to one sector"; that change does not
touch the pp_count/ppe[] loop, so the two are complementary (separate hunks).
block/partitions/aix.c | 9 +++++++++
1 file changed, 9 insertions(+)
diff --git a/block/partitions/aix.c b/block/partitions/aix.c
index 29b8f4cebb63..f3c4174e003e 100644
--- a/block/partitions/aix.c
+++ b/block/partitions/aix.c
@@ -226,6 +226,15 @@ int aix_partition(struct parsed_partitions *state)
int next_lp_ix = 1;
int lp_ix;
+ /*
+ * pvd was read into a fixed-size struct pvd whose ppe[] array
+ * holds ARRAY_SIZE(pvd->ppe) entries. pp_count is an
+ * unvalidated on-disk __be16, so clamp the scan to the array
+ * size to avoid walking past the allocation.
+ */
+ if (numpps > ARRAY_SIZE(pvd->ppe))
+ numpps = ARRAY_SIZE(pvd->ppe);
+
for (i = 0; i < numpps; i += 1) {
struct ppe *p = pvd->ppe + i;
unsigned int lv_ix;
--
2.43.0
^ permalink raw reply [flat|nested] 3+ messages in thread* Re: [PATCH] partitions: aix: bound the pp_count scan to the ppe array
2026-06-07 6:41 [PATCH] partitions: aix: bound the pp_count scan to the ppe array Bryam Vargas
@ 2026-06-08 11:46 ` Philippe De Muyter
2026-06-08 14:34 ` Jens Axboe
1 sibling, 0 replies; 3+ messages in thread
From: Philippe De Muyter @ 2026-06-08 11:46 UTC (permalink / raw)
To: Bryam Vargas
Cc: Jens Axboe, Kees Cook, Michael Bommarito, linux-block, linux-kernel
Hello Bryam,
On Sun, Jun 07, 2026 at 06:41:43AM +0000, Bryam Vargas wrote:
> aix_partition() reads the physical volume descriptor into a fixed-size
> struct pvd and then scans its physical-partition-extent array:
>
> int numpps = be16_to_cpu(pvd->pp_count);
> ...
> for (i = 0; i < numpps; i += 1) {
> struct ppe *p = pvd->ppe + i;
> ...
> lp_ix = be16_to_cpu(p->lp_ix);
>
> pvd points at a single kmalloc()'d struct pvd whose ppe[] member holds a
> fixed ARRAY_SIZE(pvd->ppe) (1016) entries, but the loop runs up to the
> on-disk pp_count. pp_count is an unvalidated __be16 read straight from
> the descriptor, so a crafted AIX image with pp_count larger than 1016
> drives the loop to read pvd->ppe[i] past the end of the allocation (up to
> 65535 entries, ~2 MB out of bounds).
>
> The partition scan runs without mounting anything, when a block device
> with a crafted AIX/IBM partition table appears (an attacker-supplied
> image attached with losetup -P, or a device auto-scanned by udev), via
> msdos_partition() -> aix_partition().
>
> Clamp the scan to the number of entries the ppe[] array can hold.
>
> Fixes: 6ceea22bbbc8 ("partitions: add aix lvm partition support files")
> Cc: stable@vger.kernel.org
> Signed-off-by: Bryam Vargas <hexlabsecurity@proton.me>
> ---
> Reproduced on v7.1-rc6 with KASAN (CONFIG_PARTITION_ADVANCED +
> CONFIG_AIX_PARTITION). A crafted disk image whose AIX/IBM partition table
> sets pp_count to 0xffff, attached with `losetup -fP image.img` (in-kernel
> partition scan, no mount), is reported by KASAN:
>
> BUG: KASAN: slab-out-of-bounds in aix_partition+0xb6e/0xee0
> Read of size 2 at addr ... by task losetup
> aix_partition
> msdos_partition
> bdev_disk_changed
> loop_reread_partitions
> loop_configure
> lo_ioctl
> __x64_sys_ioctl
>
> i.e. a read past the end of the kmalloc(sizeof(struct pvd)) object. A control
> image with pp_count == 1016 (== ARRAY_SIZE(pvd->ppe)) is clean. With this
> patch the crafted image is parsed with no out-of-bounds access.
>
> This is the read-loop sibling of the lvd scan bounded by Michael Bommarito's
> "partitions: aix: bound the lvd scan to one sector"; that change does not
> touch the pp_count/ppe[] loop, so the two are complementary (separate hunks).
>
> block/partitions/aix.c | 9 +++++++++
> 1 file changed, 9 insertions(+)
>
> diff --git a/block/partitions/aix.c b/block/partitions/aix.c
> index 29b8f4cebb63..f3c4174e003e 100644
> --- a/block/partitions/aix.c
> +++ b/block/partitions/aix.c
> @@ -226,6 +226,15 @@ int aix_partition(struct parsed_partitions *state)
> int next_lp_ix = 1;
> int lp_ix;
>
> + /*
> + * pvd was read into a fixed-size struct pvd whose ppe[] array
> + * holds ARRAY_SIZE(pvd->ppe) entries. pp_count is an
> + * unvalidated on-disk __be16, so clamp the scan to the array
> + * size to avoid walking past the allocation.
> + */
> + if (numpps > ARRAY_SIZE(pvd->ppe))
> + numpps = ARRAY_SIZE(pvd->ppe);
> +
> for (i = 0; i < numpps; i += 1) {
> struct ppe *p = pvd->ppe + i;
> unsigned int lv_ix;
> --
> 2.43.0
Thank you for your patch.
Acked-by: Philippe De Muyter <phdm@macqel.be>
Best regards
Philippe
^ permalink raw reply [flat|nested] 3+ messages in thread* Re: [PATCH] partitions: aix: bound the pp_count scan to the ppe array
2026-06-07 6:41 [PATCH] partitions: aix: bound the pp_count scan to the ppe array Bryam Vargas
2026-06-08 11:46 ` Philippe De Muyter
@ 2026-06-08 14:34 ` Jens Axboe
1 sibling, 0 replies; 3+ messages in thread
From: Jens Axboe @ 2026-06-08 14:34 UTC (permalink / raw)
To: Bryam Vargas
Cc: Philippe De Muyter, Kees Cook, Michael Bommarito, linux-block,
linux-kernel
On Sun, 07 Jun 2026 06:41:43 +0000, Bryam Vargas wrote:
> aix_partition() reads the physical volume descriptor into a fixed-size
> struct pvd and then scans its physical-partition-extent array:
>
> int numpps = be16_to_cpu(pvd->pp_count);
> ...
> for (i = 0; i < numpps; i += 1) {
> struct ppe *p = pvd->ppe + i;
> ...
> lp_ix = be16_to_cpu(p->lp_ix);
>
> [...]
Applied, thanks!
[1/1] partitions: aix: bound the pp_count scan to the ppe array
commit: 2dc0bfd2fe355fb930de63c2f2eb8ced8570c579
Best regards,
--
Jens Axboe
^ permalink raw reply [flat|nested] 3+ messages in thread
end of thread, other threads:[~2026-06-08 14:34 UTC | newest]
Thread overview: 3+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2026-06-07 6:41 [PATCH] partitions: aix: bound the pp_count scan to the ppe array Bryam Vargas
2026-06-08 11:46 ` Philippe De Muyter
2026-06-08 14:34 ` Jens Axboe
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox
all inboxes | Powered by JetHome®