[SRU][J][PATCH 0/1] CVE-2024-27407
Massimiliano Pellizzer
massimiliano.pellizzer at canonical.com
Mon Aug 4 17:19:15 UTC 2025
https://ubuntu.com/security/CVE-2024-27407
[ Impact ]
CVE‑2024‑27407 stems from an incorrect arithmetic overflow check in mi_enum_attr().
The original code used:
if (t16 + le32_to_cpu(attr->res.data_size) > asize)
This check was flawed: t16 is the current offset and attr->res.data_size is the size of the attribute data.
Summing them before comparison allows overflow wrap‑around, potentially bypassing the guard and enabling a buffer overflow.
The correct logic should compute:
if (le32_to_cpu(attr->res.data_size) > asize - t16)
This ensures that data_size itself doesn’t exceed the remaining space (asize ‑ t16),
preventing mis‑reporting sizes that would overflow the buffer.
[ Fix ]
Plucky: Not affected
Noble: Not affected
Jammy: Backported fix commit
[ Test Plan ]
Compile and boot tested.
Tested ntfs3 stability with stress-ng:
$ sudo mkfs.ntfs -f /dev/sdb1
Cluster size has been automatically set to 4096 bytes.
Creating NTFS volume structures.
mkntfs completed successfully.
$ sudo mkdir -p /mnt/ntfs3
$ sudo mount -t ntfs3 /dev/sdb1 /mnt/ntfs3
$ lsmod | grep ntfs
ntfs3 274432 1
$ cd /mnt/ntf3
$ sudo stress-ng --hdd 4 --aggressive --dentry 4 --timeout 5m --metrics-brief
stress-ng: info: [1393] setting to a 300 second (5 mins, 0.00 secs) run per stressor
stress-ng: info: [1393] dispatching hogs: 4 hdd, 4 dentry
stress-ng: info: [1398] stress-ng-dentry: 16329719 dentries allocated
stress-ng: info: [1393] successful run completed in 310.72s (5 mins, 10.72 secs)
stress-ng: info: [1393] stressor bogo ops real time usr time sys time bogo ops/s bogo ops/s
stress-ng: info: [1393] (secs) (secs) (secs) (real time) (usr+sys time)
stress-ng: info: [1393] hdd 526150 309.39 10.30 39.26 1700.61 10616.42
stress-ng: info: [1393] dentry 16355282 306.77 54.78 791.49 53315.06 19326.32
[ Where Problems Could Occur ]
A regression here is unlikely due to the very limited scope
of the patch.
More information about the kernel-team
mailing list