Jiaming Zhang

**Name of the Vulnerable Software and Affected Versions** Linux kernel (affected versions not specified) **Description** The Linux kernel contains a flaw within the XFS file system related to the revalidation of two btrees during file system repair. Specifically, the issue arises when rebuilding both the free space and inode btrees concurrently. The first btree revalidation can nullify a cursor required by the second revalidation, leading to a null pointer dereference and a system crash. This occurs because the first call to `xchk allocbt()` nullifies the cursor needed by the second call. The issue affects the `xrep revalidate allocbt()` function, which is called through a chain including `xfs file ioctl()`, `xfs ioc scrubv metadata()`, and `xfs scrub metadata()`. The function `sc->ops->repair eval(sc)` is also involved. The vulnerability is triggered when the first btree revalidation fails on a cross-reference attempt, resulting in the deletion of the cursor used by the second btree. The same issue also affects the `xrep revalidate iallocbt` function. **Recommendations** At the moment, there is no information about a newer version that contains a fix for this vulnerability.

**Name of the Vulnerable Software and Affected Versions** Linux kernel versions 6.6 through 6.14 **Description** The Linux kernel contains an issue where the `xchk xfile * descr` macros utilize `kasprintf`, which may fail to allocate memory if the resulting formatted string exceeds a certain length. This can occur because the formatted strings can exceed the memory allocation limits. The issue was identified through fuzzing with syzbot. The `xchk xfile * descr` macros are used for debugging purposes and are not essential for functionality. **Recommendations** Versions 6.6 through 6.14 should be updated to a later version that addresses this issue.

**Name of the Vulnerable Software and Affected Versions** Linux kernel (affected versions not specified) **Description** A flaw exists in the Linux kernel's f2fs implementation related to out-of-boundary access in dnode pages. The issue stems from a corrupted image where a dnode shares a node ID with its inode. During the `f2fs get dnode of data()` function, an attempt is made to access a block address within the dnode at an offset, but the dnode is parsed as an inode node, leading to an invalid memory access calculation (360 + 934 * 4 = 4096 with 4 bytes). The root cause is a lack of sanity checks for node IDs of direct nodes during the `f2fs get dnode of data()` function. **Recommendations** At the moment, there is no information about a newer version that contains a fix for this vulnerability.