{"id":"AZL-115726","summary":"CVE-2026-99797 affecting package kernel 6.6.161.1-1","details":"In the Linux kernel, the following vulnerability has been resolved:\n\nio_uring/cancel: drop sqd-\u003elock while SQPOLL waits\n\nAn exiting SQPOLL thread calls io_uring_cancel_generic() with sqd-\u003elock\nheld, and keeps it across cond_resched() and the sleep until nothing is\nin flight anymore. A zero-copy send notification can't be canceled and\nstays pending as long as the peer doesn't read, so the lock is held for\nas long as a user cares to. io_ring_exit_work() needs it to park the\nthread, and so do IORING_REGISTER_IOWQ_AFF and IOWQ_MAX_WORKERS. With\n64 such rings sharing the thread, every iou_exit slot sleeps on the lock\nand no ring gets freed for anyone.\n\nDrop the lock around cond_resched() and the sleep, like the main loop of\nthe thread does. Each walk of ctx_list is still under one hold, and\nnothing is kept across the unlock. The thread drops its io-wq, frees its\ntctx and clears sqd-\u003ethread under one hold after the loop, so whoever\nholds the lock and finds sqd-\u003ethread set can still use them.\n\n[axboe: adapted to the schedule_timeout() based loop]","modified":"2026-11-21T14:17:26.858473475Z","published":"2026-11-20T08:18:37Z","upstream":["CVE-2026-99797"],"references":[{"type":"WEB","url":"https://nvd.nist.gov/vuln/detail/CVE-2026-99797"}],"affected":[{"package":{"name":"kernel","ecosystem":"Azure Linux:3","purl":"pkg:rpm/azure-linux/kernel"},"ranges":[{"type":"ECOSYSTEM","events":[{"introduced":"0"},{"last_affected":"6.6.161.1-1"}]}],"database_specific":{"source":"https://github.com/microsoft/AzureLinuxVulnerabilityData/blob/main/osv/AZL-115726.json"}}],"schema_version":"1.9.2"}