* [6.1.y] udf: VAT inode size checks
@ 2026-11-21 18:16 Aymen T
0 siblings, 0 replies; only message in thread
From: Aymen T @ 2026-11-21 18:16 UTC (permalink / raw)
To: stable; +Cc: jack
Hi,
Please consider these two for 6.1.y:
5163e6ee1ea7 ("udf: validate VAT inode size for old VAT format")
49f771d14578 ("udf: validate vat20->lengthHeader and VAT inode size in udf_load_vat()")
6.1.192 still computes the old-format VAT entry count without checking
the inode size:
map->s_type_specific.s_virtual.s_num_entries =
(sbi->s_vat_inode->i_size - 36) >> 2;
so a VAT inode smaller than 36 bytes gives a huge s_num_entries, and
udf_get_pblock_virt15() then reads a VAT entry far past the VAT data.
On 6.1.180 in QEMU, with a crafted image on a USB stick, that value
became the LBA of the next read.
5163e6ee1ea7 applies cleanly to 6.1.192. 49f771d14578 needs fuzz on
its first hunk, because 6.1.y doesn't have the udf_block_map() error
handling that surrounds it in mainline; the resulting change is the
same two checks. I can send it as a backport if that's easier.
Thanks,
Aymen
^ permalink raw reply [flat|nested] only message in thread
only message in thread, other threads:[~2026-11-21 18:16 UTC | newest]
Thread overview: (only message) (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2026-11-21 18:16 [6.1.y] udf: VAT inode size checks Aymen T
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox