{"id":"AZL-101367","summary":"CVE-2026-89857 affecting package kernel 6.6.150.1-1","details":"In the Linux kernel, the following vulnerability has been resolved:\n\nscsi: qla2xxx: Hold qpair lock when sending NVMe LS reject\n\nqla_nvme_ls_reject_iocb() allocates from and advances the request ring\nthrough __qla2x00_alloc_iocbs() (which assumes the hardware_lock is\nheld) and qla2x00_start_iocbs() (which advances the ring and rings the\nrequest-in doorbell), but takes no lock itself. Two of its callers\ninvoke it without the producer lock held:\n\n - qla_nvme_xmt_ls_rsp(), the NVMe-FC .xmt_ls_rsp transport callback, on\n   its error path, and\n\n - qla2xxx_process_purls_pkt(), run from the purex work/DPC context.\n\nBoth use ha-\u003ebase_qpair, whose qp_lock_ptr is hardware_lock, so they can\nrun concurrently with normal I/O submission on the base ring and corrupt\nthe ring producer state, leading to duplicated or dropped commands. The\nthird caller, qla2xxx_process_purls_iocb(), runs inside\nqla24xx_process_response_queue() with the qpair lock already held and is\nsafe; that is also why the lock cannot be taken inside the helper itself\n(it would recursively re-acquire hardware_lock on the response path).\n\nTake qp_lock_ptr around the two unlocked callers and document the helper\nas caller-locked. Both run in process context, so spin_lock_irqsave() is\nused and nothing in the locked region sleeps.","modified":"2026-09-18T05:37:08Z","published":"2026-09-16T11:16:53Z","upstream":["CVE-2026-89857"],"references":[{"type":"WEB","url":"https://nvd.nist.gov/vuln/detail/CVE-2026-89857"}],"affected":[{"package":{"name":"kernel","ecosystem":"Azure Linux:3","purl":"pkg:rpm/azure-linux/kernel"},"ranges":[{"type":"ECOSYSTEM","events":[{"introduced":"0"},{"last_affected":"6.6.150.1-1"}]}],"database_specific":{"source":"https://github.com/microsoft/AzureLinuxVulnerabilityData/blob/main/osv/AZL-101367.json"}}],"schema_version":"1.9.0"}