Описание
In the Linux kernel, the following vulnerability has been resolved:
libceph: fix OOB read in decode_watchers() via missing bounds check
ceph_start_decoding() validates that struct_len bytes remain in the buffer after the encoding header, but accepts struct_len=0 as valid: ceph_decode_need(p, end, 0, bad) always passes. When a malicious or compromised OSD sends an obj_list_watch_response_t reply with struct_len=0, ceph_start_decoding() returns success with p == end, leaving zero bytes guaranteed for subsequent reads.
The immediately following ceph_decode_32(p) in decode_watchers() has no preceding bounds check. With p == end this is a 4-byte read past the validated buffer boundary. The garbage value is then passed directly to kzalloc_objs() as the watcher count.
The sibling function decode_watcher() already uses the safe variants (ceph_decode_copy_safe, ceph_decode_64_safe, ceph_decode_skip_32) after its own ceph_start_decoding() call. decode_watchers() is the only site that uses...
In the Linux kernel, the following vulnerability has been resolved:
libceph: fix OOB read in decode_watchers() via missing bounds check
ceph_start_decoding() validates that struct_len bytes remain in the buffer after the encoding header, but accepts struct_len=0 as valid: ceph_decode_need(p, end, 0, bad) always passes. When a malicious or compromised OSD sends an obj_list_watch_response_t reply with struct_len=0, ceph_start_decoding() returns success with p == end, leaving zero bytes guaranteed for subsequent reads.
The immediately following ceph_decode_32(p) in decode_watchers() has no preceding bounds check. With p == end this is a 4-byte read past the validated buffer boundary. The garbage value is then passed directly to kzalloc_objs() as the watcher count.
The sibling function decode_watcher() already uses the safe variants (ceph_decode_copy_safe, ceph_decode_64_safe, ceph_decode_skip_32) after its own ceph_start_decoding() call. decode_watchers() is the only site that uses the bare variant, confirming an oversight.
Fix by replacing ceph_decode_32(p) with ceph_decode_32_safe(p, end, *num_watchers, bad), consistent with the established pattern.
Attacker model: a malicious or compromised OSD in a multi-tenant Ceph deployment (e.g. cloud) can trigger this against any kernel client that calls CEPH_OSD_OP_LIST_WATCHERS, without any further privileges beyond OSD session establishment.
[ idryomov: trim changelog ]
Ссылки
- https://nvd.nist.gov/vuln/detail/CVE-2026-80557
- https://git.kernel.org/stable/c/00ead17c7de137a692edee59f2772e6af687e8eb
- https://git.kernel.org/stable/c/1c824e7c75bb4adf19553dd4ea944a5d83096be8
- https://git.kernel.org/stable/c/7130d94846dadbb97b6b7f4d78a3a7bba6e3daa1
- https://git.kernel.org/stable/c/85479b7d65b4ebcb07fbbe57230976793974ab4a
- https://git.kernel.org/stable/c/c59219a6b62d74936963983e5815524c3de8dd79
- https://git.kernel.org/stable/c/cb8246e5846dbbe34930903a90c7a90dd8e5910b
- https://git.kernel.org/stable/c/eab3eeb68bfc639d74f27256f05546af5c4f787d
- https://git.kernel.org/stable/c/f161be39201eb5f9b1f58fb8f90b8a9cd3931eb6
Связанные уязвимости
In the Linux kernel, the following vulnerability has been resolved: libceph: fix OOB read in decode_watchers() via missing bounds check ceph_start_decoding() validates that struct_len bytes remain in the buffer after the encoding header, but accepts struct_len=0 as valid: ceph_decode_need(p, end, 0, bad) always passes. When a malicious or compromised OSD sends an obj_list_watch_response_t reply with struct_len=0, ceph_start_decoding() returns success with p == end, leaving zero bytes guaranteed for subsequent reads. The immediately following ceph_decode_32(p) in decode_watchers() has no preceding bounds check. With p == end this is a 4-byte read past the validated buffer boundary. The garbage value is then passed directly to kzalloc_objs() as the watcher count. The sibling function decode_watcher() already uses the safe variants (ceph_decode_copy_safe, ceph_decode_64_safe, ceph_decode_skip_32) after its own ceph_start_decoding() call. decode_watchers() is the only site that uses the...
In the Linux kernel, the following vulnerability has been resolved: libceph: fix OOB read in decode_watchers() via missing bounds check ceph_start_decoding() validates that struct_len bytes remain in the buffer after the encoding header, but accepts struct_len=0 as valid: ceph_decode_need(p, end, 0, bad) always passes. When a malicious or compromised OSD sends an obj_list_watch_response_t reply with struct_len=0, ceph_start_decoding() returns success with p == end, leaving zero bytes guaranteed for subsequent reads. The immediately following ceph_decode_32(p) in decode_watchers() has no preceding bounds check. With p == end this is a 4-byte read past the validated buffer boundary. The garbage value is then passed directly to kzalloc_objs() as the watcher count. The sibling function decode_watcher() already uses the safe variants (ceph_decode_copy_safe, ceph_decode_64_safe, ceph_decode_skip_32) after its own ceph_start_decoding() call. decode_watchers() is the only site that uses the...
In the Linux kernel, the following vulnerability has been resolved: libceph: fix OOB read in decode_watchers() via missing bounds check ceph_start_decoding() validates that struct_len bytes remain in the buffer after the encoding header, but accepts struct_len=0 as valid: ceph_decode_need(p, end, 0, bad) always passes. When a malicious or compromised OSD sends an obj_list_watch_response_t reply with struct_len=0, ceph_start_decoding() returns success with p == end, leaving zero bytes guaranteed for subsequent reads. The immediately following ceph_decode_32(p) in decode_watchers() has no preceding bounds check. With p == end this is a 4-byte read past the validated buffer boundary. The garbage value is then passed directly to kzalloc_objs() as the watcher count. The sibling function decode_watcher() already uses the safe variants (ceph_decode_copy_safe, ceph_decode_64_safe, ceph_decode_skip_32) after its own ceph_start_decoding() call. decode_watchers() is the only site that uses th
In the Linux kernel, the following vulnerability has been resolved: l ...