CVE-2026-97926 PUBLISHED

ufs: validate cylinder group metadata before caching it

Assigner: Linux
Reserved: 25.09.2026 Published: 25.09.2026 Updated: 25.09.2026

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

ufs: validate cylinder group metadata before caching it

ufs_read_cylinder() copies the cylinder group index and the rotor positions straight from the on-disk group and caches them without any check:

<pre>ucpi->c_cgx = fs32_to_cpu(sb, ucg->cg_cgx); ucpi->c_rotor = fs32_to_cpu(sb, ucg->cg_rotor); ucpi->c_frotor = fs32_to_cpu(sb, ucg->cg_frotor); ucpi->c_irotor = fs32_to_cpu(sb, ucg->cg_irotor); </pre>

They are then used as indices during allocation and free:

  • c_cgx indexes the cylinder summary array as UFS_SB(sb)->fs_cs(ucpi->c_cgx), so a value past s_ncg writes a 32 bit count outside the s_csp allocation.

  • c_frotor becomes a bitmap scan start, start = c_frotor >> 3, and then length = ((s_fpg + 7) >> 3) - start. A start beyond the block bitmap wraps the unsigned length to a huge value, so ubh_scanc() walks far past the cylinder group buffers. c_irotor drives the inode bitmap the same way.

A crafted image can set any of these freely, turning an ordinary allocation into an out of bounds access.

Reject a cylinder group whose recorded index does not match the group being read, or whose rotors fall outside the group, before the metadata is cached. Valid filesystems keep cg_cgx equal to the group number and the rotors within the group, so only malformed images are rejected.

Metrics

CVSS Vector: CVSS:3.1/AV:L/AC:H/PR:N/UI:R/S:U/C:H/I:H/A:H
CVSS Score: 7

AV:L - The bad cg_cgx/cg_rotor/cg_frotor/cg_irotor values come from a crafted UFS image's cylinder group block, read by ufs_read_cylinder(). They are then used by the local VFS allocate/free paths (ufs_new_fragments, ufs_free_fragments, ufs_new_inode, ufs_free_inode). No network protocol carries these bytes. AC:H - ufs_load_cylinder() is called only from the allocate/free paths in balloc.c and ialloc.c. These need a read-write mount, and ufs_fill_super() returns -EROFS for one unless the non-default CONFIG_UFS_FS_WRITE ("DANGEROUS") is set. Major distros leave it off and only a few MIPS defconfigs set it, so reaching the bug depends on a rare kernel config. PR:N - The attacker only has to craft the image. The ufs file_system_type sets FS_REQUIRES_DEV without FS_USERNS_MOUNT, so the mount is done by a privileged victim, not the attacker. After that, creating or extending a file triggers the allocation. UI:R - A victim has to mount the attacker-supplied UFS image read-write before ufs_read_cylinder() caches the malformed cylinder group fields. S:U - The out-of-bounds accesses stay in kernel heap memory in the same kernel security authority. No guest/host or hardware boundary is crossed. C:H - If c_frotor/c_irotor lie outside the group, the unsigned length in ufs_bitmap_search() wraps. ubh_scanc() then reads far past the cylinder group buffers, an unbounded out-of-bounds read of kernel memory. I:H - ufs_add_fragments()/ufs_free_* call fs32_add() on UFS_SB(sb)->fs_cs(ucpi->c_cgx). An attacker-chosen c_cgx beyond s_ncg adds to a 32-bit value outside the s_csp heap allocation, an out-of-bounds heap write at a chosen index. A:H - The wild ubh_scanc() walk and the out-of-bounds fs_cs() writes can oops the kernel or corrupt the heap, crashing the system.

Product Status

Vendor Linux
Product Linux
Versions Default: unaffected
  • affected from 1da177e4c3f41524e886b7f1b8a0c1fc7321cac2 to 5902a95066883cf96fa15b2680694fc5dd0c7d11 (excl.)
  • affected from 1da177e4c3f41524e886b7f1b8a0c1fc7321cac2 to 87b12dc360a002eb2d498aa4f01e84347019612d (excl.)
  • affected from 1da177e4c3f41524e886b7f1b8a0c1fc7321cac2 to abde9eb33106850dfa367ad3965d3588bb8558d5 (excl.)
  • affected from 1da177e4c3f41524e886b7f1b8a0c1fc7321cac2 to c9d263be26806d388129fab8c6904bed197fc6af (excl.)
Vendor Linux
Product Linux
Versions Default: affected
  • Version 2.6.12 is affected
  • unaffected from 0 to 2.6.12 (excl.)
  • unaffected from 6.12.111 to 6.12.* (incl.)
  • unaffected from 6.18.53 to 6.18.* (incl.)
  • unaffected from 7.2.7 to 7.2.* (incl.)
  • unaffected from 7.3-rc3 to * (incl.)

References