CVE-2023-53848: 
Linux Kernel vulnerability analysis and mitigation

CVE-2023-53848 is a deadlock vulnerability in the Linux kernel's md/raid5-cache subsystem, specifically within the r5l_exit_log() function. The flaw was introduced by commit b13015af94cf ("md/raid5-cache: Clear conf->log after finishing work") and causes a deadlock condition when r5l_exit_log() is called while holding reconfig_mutex. It affects Linux kernel versions from the introduction of commit b13015af94cf up to the patched stable releases (before 6.1.53, 6.4.16, 6.5.3, and 6.6). The vulnerability was published on December 9, 2025, and carries an estimated CVSS severity of Medium with an EPSS score of 0.00018 (Feedly, EUVD).

The root cause is a concurrency/locking defect (CWE-667: Improper Locking) in the RAID5 write-back cache log exit path. When r5l_exit_log() is called by a caller holding reconfig_mutex, it invokes flush_work(&log->disable_writeback_work), which internally calls r5c_disable_writeback_async(). That function uses wait_event() waiting for conf->log to become NULL, but because conf->log is only set to NULL after the flush_work() call completes, and mddev_trylock() fails due to the already-held reconfig_mutex, the wait_event() condition can never be satisfied — resulting in a permanent deadlock. The fix reorders operations so that conf->log is set to NULL before wake_up() is called (restoring the original ordering), and moves md_unregister_thread() forward to also preserve the null-pointer dereference fix from the prior commit (Feedly).

Successful triggering of this vulnerability causes the affected Linux kernel to deadlock in the RAID5 cache log exit path, resulting in a denial of service — the system or the RAID subsystem becomes unresponsive and cannot recover without a reboot. There is no evidence of confidentiality or integrity impact; the primary consequence is an availability impact on systems using RAID5 with write-back cache (journal) configurations. Affected systems include servers and storage appliances running vulnerable Linux kernel versions with md/raid5-cache enabled (Feedly).

Apply the upstream Linux kernel patches that fix the ordering of conf->log = NULL relative to wake_up() in r5l_exit_log(). Fixed versions are available in stable kernel releases: 6.1.53, 6.4.16, 6.5.3, and 6.6 and later. Users of affected distributions should apply vendor-provided kernel updates; SUSE has issued advisories (SUSE-2026-0281-1 and SUSE-2026-20876-1) addressing this issue (Feedly, Linux Security). As a workaround where patching is not immediately possible, avoiding the use of RAID5 write-back cache (journal) mode can prevent exposure to this code path.