{"id":"AZL-100265","summary":"CVE-2026-89491 affecting package kernel 6.6.150.1-1","details":"In the Linux kernel, the following vulnerability has been resolved:\n\nocfs2: cluster: don't sleep while holding o2hb_live_lock in o2hb_region_pin()\n\nPatch series \"ocfs2: cluster: o2hb_region_pin() fixes\", v2.\n\nThis series fixes three related issues in o2hb_region_pin(), all are from\nthe original implementation in commit: 58a3158a5d17 (\"ocfs2/cluster:\nPin/unpin o2hb regions\"):\n\n1) It is called with o2hb_live_lock (a spinlock) held, but the\n   underlying configfs_depend_item() sleeps (takes inode rwsem and\n   pins the filesystem).  This triggers BUG under\n   CONFIG_DEBUG_ATOMIC_SLEEP.\n\n2) When called from the configfs drop_item callback, it creates a\n   lock order inversion: parent inode_lock -\u003e configfs root\n   inode_lock, which can deadlock against subsystem unregistration\n   paths taking root -\u003e parent.\n\n3) If pinning fails partway through o2hb_region_inc_user(), the\n   o2hb_dependent_users counter is leaked and partially-pinned\n   regions are never released, leaving heartbeat regions\n   unprotected on subsequent mounts.\n\nPatch 1 reworks o2hb_region_pin() to drop o2hb_live_lock across each\nsleeping configfs_depend_item() call, using a config_item reference to\nkeep the region alive while unlocked.\n\nPatch 2 adds a from_callback parameter to select\nconfigfs_depend_item_unlocked() when called from configfs context,\navoiding the inode_lock nesting.\n\nPatch 3 fixes the error path in o2hb_region_inc_user() to unpin and\ndecrement the counter on failure.\n\n\nThis patch (of 3):\n\no2hb_region_pin() is always called with the o2hb_live_lock spinlock held\n(from o2hb_region_inc_user() and o2hb_heartbeat_group_drop_item()), but it\ncalls o2nm_depend_item() -\u003e configfs_depend_item(), which sleeps: it pins\nthe configfs filesystem and takes the configfs root inode rwsem.  Under\nCONFIG_DEBUG_ATOMIC_SLEEP this triggers:\n\n  BUG: sleeping function called from invalid context at kernel/locking/rwsem.c\n  in_atomic(): 1, ... name: mount.ocfs2\n    down_write\n    configfs_depend_item\n    o2hb_region_pin\n    o2hb_region_inc_user\n    o2hb_register_callback\n    dlm_register_domain_handlers\n    ...\n    ocfs2_dlm_init\n    ocfs2_mount_volume\n    ocfs2_fill_super\n\nRework o2hb_region_pin() to pin one region at a time with the lock dropped\nacross the sleeping call: under o2hb_live_lock find the next eligible\nregion and take a config_item reference to keep it alive, drop the lock,\ncall o2nm_depend_item(), then retake the lock and record the pin.  The\nconfig_item_put() is done with the lock released as well, since\no2hb_region_release() also acquires o2hb_live_lock and can sleep.  The\nregion list may change while unlocked, so the scan restarts from the top\nafter each pin.  Local heartbeat still pins only the matching region;\nglobal heartbeat pins all eligible regions.\n\nThe unpin path is unaffected: configfs_undepend_item() only takes a\nspinlock and does not sleep.","modified":"2026-09-12T14:15:18.671007055Z","published":"2026-09-11T20:19:30Z","upstream":["CVE-2026-89491"],"references":[{"type":"WEB","url":"https://nvd.nist.gov/vuln/detail/CVE-2026-89491"}],"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-100265.json"}}],"schema_version":"1.9.0"}