[SRU][J][PATCH 0/1] CVE-2024-53220

[Cengiz Can]

https://ubuntu.com/security/CVE-2024-53220

[ Impact ]

In the Linux kernel, the following vulnerability has been resolved:

f2fs: fix to account dirty data in __get_secs_required()

It will trigger system panic w/ testcase in [1]:

------------[ cut here ]------------ kernel BUG at fs/f2fs/segment.c:2752! RIP:
0010:new_curseg+0xc81/0x2110 Call Trace: f2fs_allocate_data_block+0x1c91/0x4540
do_write_page+0x163/0xdf0 f2fs_outplace_write_data+0x1aa/0x340
f2fs_do_write_data_page+0x797/0x2280 f2fs_write_single_data_page+0x16cd/0x2190
f2fs_write_cache_pages+0x994/0x1c80 f2fs_write_data_pages+0x9cc/0xea0
do_writepages+0x194/0x7a0 filemap_fdatawrite_wbc+0x12b/0x1a0
__filemap_fdatawrite_range+0xbb/0xf0 file_write_and_wait_range+0xa1/0x110
f2fs_do_sync_file+0x26f/0x1c50 f2fs_sync_file+0x12b/0x1d0
vfs_fsync_range+0xfa/0x230 do_fsync+0x3d/0x80 __x64_sys_fsync+0x37/0x50
x64_sys_call+0x1e88/0x20d0 do_syscall_64+0x4b/0x110
entry_SYSCALL_64_after_hwframe+0x76/0x7e

The root cause is if checkpoint_disabling and lfs_mode are both on, it will
trigger OPU for all overwritten data, it may cost more free segment than
expected, so f2fs must account those data correctly to calculate cosumed free
segments later, and return ENOSPC earlier to avoid run out of free segment
during block allocation.

[1] https://lore.kernel.org/fstests/20241015025106.3203676-1-chao@kernel.org/

[ Fix ]

jammy/linux: backported from 1acd73edbbfe

The fix folds the missing dirty-data accounting into the free-segment
calculation so that overwritten data blocks are counted when checkpoint
disabling and LFS mode both force out-of-place updates. This lets f2fs return
ENOSPC before the segment allocator runs out of free segments and hits the
BUG_ON in new_curseg().

This tree has no __get_secs_required() helper; the same lower/upper/curseg
logic lives inline in has_not_enough_free_secs(), so the dirty-data accounting
was folded there. has_curseg_enough_space() keeps this tree's per-segment
accounting (f2fs_usable_blks_in_seg - get_seg_entry()->ckpt_valid_blocks)
instead of upstream's CAP_BLKS_PER_SEC/get_ckpt_valid_blocks while extending
the loop over CURSEG_HOT_DATA..NR_PERSISTENT_LOG and picking data_blocks vs
node_blocks.

[ Test Plan ]

Build and boot tested.

[ Where Problems Could Occur ]

A regression would only be visible on systems using the f2fs filesystem, and
in practice only when it is mounted with LFS mode together with checkpoint
disabling (typically flash-based storage on mobile and embedded devices). A
bad fix could miscalculate free-segment reservation and either return ENOSPC
too early, wasting usable space, or too late, reintroducing the allocator
panic under heavy fsync/overwrite workloads. Systems not using f2fs, or f2fs
mounts without both LFS mode and checkpoint disabling, are not affected.

[ Other Info ]

Kybele flow-v11-21-g0f091652. Reference: 10e3b253/v1