AZL-96462

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

ext4: validate donor file superblock early in EXT4IOCMOVE_EXT

Reject the EXT4IOCMOVEEXT ioctl early if the donor file does not belong to the same superblock as the original file. Currently, this validation is performed inside ext4moveextents() by mextcheckvalidity(), but only after locktwo_nondirectories() has already acquired the inode locks. When the donor fd refers to a file on a different filesystem (e.g., overlayfs), this late validation creates a circular lock dependency:

CPU0 (overlayfs write) CPU1 (ext4 ioctl) ---- ---- inodelock(ovlinode) mntwantwritefile(filp) sbstartwrite(ext4sb) [sbwriters] backingfilewriteiter() vfsiterwrite(realfile) filestartwrite(realfile) sbstartwrite(ext4sb) [blocked by freeze] locktwonondirectories() inodelock(ovl_inode) [blocked]

With a concurrent freeze operation holding sbwriters write side, this forms a deadlock cycle: CPU0 waits for freeze to complete, freeze waits for CPU1's sbwriters reader to exit, CPU1 waits for CPU0's inode lock.

Since EXT4IOCMOVEEXT exchanges physical extents between two files, it fundamentally requires both files to reside on the same ext4 filesystem. Moving the superblock check before any lock acquisition is both semantically correct and eliminates the circular dependency by ensuring that cross-filesystem donor fds are rejected before sbwriters or inode locks are taken.