{"id":"AZL-101556","summary":"CVE-2026-89956 affecting package kernel 6.6.150.1-1","details":"In the Linux kernel, the following vulnerability has been resolved:\n\ns390/vfio-ap: Fix missing lock required to access list of ap_matrix_mdev objects\n\nIn order to traverse or add/remove ap_matrix_mdev objects in the\nmatrix_dev-\u003emdev_list, the matrix_dev-\u003eguests_lock mutex must be held.\nThere are two functions that access the list without holding the mutex:\n\nvfio_ap_mdev_probe function\n~~~~~~~~~~~~~~~~~~~~~~~~~~~\nThe vfio_ap_mdev_probe function uses the matrix_dev-\u003emdevs_lock\nmutex to guard the add of a newly created ap_matrix_mdev object to the\nmatrix_dev-\u003emdev_list. This mutex does not protect list access; its purpose\nis to guard against concurrent access to fields contained in an\nap_matrix_mdev object. This could lead to kernel memory corruption or\nuse-after-free if another mdev is created or removed concurrently.\n\nThe adding of an ap_matrix_mdev object to matrix_dev-\u003emdev_list\nis now guarded by the matrix_dev-\u003eguests_lock which is the correct\nway to protect against concurrent mdev_list access.\n\nAlso removed the following two lines of code because the matrix_mdev is\nallocated via vfio_alloc_device macro which uses kzalloc, so req_trigger\nand cfg_chg_trigger are already zero-initialised when the struct is\nallocated before the call to vfio_register_emulated_iommu_dev. This\nprevents a window whereby these triggers are set to NULL after\nthe device is exposed to userspace.\n\nmatrix_mdev-\u003ereq_trigger = NULL;\nmatrix_mdev-\u003ecfg_chg_trigger = NULL;\n\nvfio_ap_mdev_for_queue function\n~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~\nThe status_show function that supports display of the status attribute of\nthe devices in /sys/bus/ap/devices calls the vfio_ap_mdev_for_queue\nfunction which iterates the matrix_dev-\u003emdev_list to find the object\nrepresenting the queue device whose status is to be displayed. In order to\ntraverse this list, the matrix_dev-\u003eguests_lock mutex must be held.\n\nTo fix this, the guests_lock mutex is taken prior to taking the\nmatrix_dev-\u003emdevs_lock mutex in the status_show function. It is taken\nthere rather than the vfio_ap_mdev_for_queue function - where it is\nneeded - because it must be taken prior to the mdevs_lock mutex in order to\nadhere to the proper locking order and prevent a lockdep splat; also\nbecause the mdevs_lock is needed there to access fields within\nthe matrix_mdev object in that function.\n\nSee the vfio-ap-locking.rst in the linux kernel tree.","modified":"2026-09-17T14:15:39.296634474Z","published":"2026-09-16T11:17:05Z","upstream":["CVE-2026-89956"],"references":[{"type":"WEB","url":"https://nvd.nist.gov/vuln/detail/CVE-2026-89956"}],"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-101556.json"}}],"schema_version":"1.9.0"}