{"cve":"CVE-2026-68173","nvd":{"published":"2026-08-10T13:20:04.067","modified":"2026-08-17T05:18:18.770","status":"Received","description":"In the Linux kernel, the following vulnerability has been resolved:\n\nublk: wait on ublk_dev_ready() instead of ub->completion\n\nub->completion is only re-armed by a successful START_USER_RECOVERY. If\nthe ublk server sends END_USER_RECOVERY without one - e.g. its START\nfailed with -EBUSY and the error was ignored - the wait is satisfied by\nthe stale completion of the previous recovery cycle, and the device is\nmarked LIVE and the requeue list kicked while the FETCH stream is still\nrunning and ubq->canceling is still set. The kick redispatches a\npreviously requeued request, __ublk_queue_rq_common() sees ->canceling\nand parks it again via __ublk_abort_rq(), and after the last FETCH\nclears ->canceling nothing ever kicks the requeue list again: the\nrequest is stranded there while holding its tag. If it is the flush\nmachinery's flush_rq, every subsequent fsync piles up in uninterruptible\nsleep and teardown hangs on tag draining. This matches a report of a\nlost PREFLUSH with ext4 on top of ublk after daemon crash recovery.\n\nub->completion is an edge-triggered latch used as a proxy for the level\ncondition \"every queue has fetched all I/O commands\", which can regress\n(F_BATCH's UNPREP, daemon death) and whose re-arm can be skipped. Drop\nit and wait on the real condition instead: the new helper\nublk_wait_dev_ready_and_lock() waits on ublk_dev_ready() via\nwait_var_event_interruptible(), woken from ublk_mark_io_ready(), then\nre-checks it under ub->mutex, waiting again on regression, and returns\nwith the mutex held and readiness guaranteed.\n\nReadiness becomes true in the same ub->mutex critical section that\nclears the last queue's ->canceling, so END_USER_RECOVERY marks the\ndevice LIVE and kicks the requeue list strictly after ->canceling\nclears. The wait stays interruptible, so a server whose daemon died can\nstill be signalled out. For ublk_ctrl_start_dev() this replaces the\nfail-fast -EINVAL on an F_BATCH ready->UNPREP regression with waiting\nuntil the device is ready again.","cvss_v4":null,"cvss_v3":{"version":"3.1","vector":"CVSS:3.1/AV:L/AC:L/PR:L/UI:N/S:U/C:N/I:H/A:H","score":7.1},"cwes":[],"references":["https://git.kernel.org/stable/c/432a9b2780c0a01caf547bd1fc2fcf28aeb8d173","https://git.kernel.org/stable/c/7dd26adf7e7d482af524e3a0cca4a81ef7c159d0","https://git.kernel.org/stable/c/8f188dd11a1c2ad94caeaee36ef68bb8221d4a12"],"products":[],"source":"nvd","synced_at":"2026-10-19T22:47:54.990491175+00:00"},"epss":{"cve":"CVE-2026-68173","score":0.00162,"percentile":0.05374,"date":"2026-10-31"},"kev":null,"osv":[],"attribution":["This product uses the NVD API but is not endorsed or certified by the NVD.","EPSS scores: FIRST Exploit Prediction Scoring System, https://www.first.org/epss","Known exploited vulnerabilities: CISA Known Exploited Vulnerabilities Catalog, https://www.cisa.gov/known-exploited-vulnerabilities-catalog","Package advisories: OSV (https://osv.dev), from the GitHub Advisory Database, PyPA, Go, RustSec and other ecosystem databases, under CC-BY 4.0 or CC0","CVE records: the CVE Program's CVE List (https://www.cve.org), under its terms of use (https://www.cve.org/Legal/TermsOfUse). CVE is a registered trademark of The MITRE Corporation; this product is not affiliated with the CVE Program."]}