{"id":"AZL-96141","summary":"CVE-2026-74572 affecting package kernel 6.6.150.1-1","details":"In the Linux kernel, the following vulnerability has been resolved:\n\nbtrfs: zoned: fix deadlock between metadata writeback and transaction commit\n\nWhen writing out metadata extent buffers in a zoned filesystem,\nbtree_writepages() holds fs_info-\u003ezoned_meta_io_lock across the whole\nwriteback loop, including the call to btrfs_check_meta_write_pointer() -\u003e\ncheck_bg_is_active().\n\nFor the tree-log block group, check_bg_is_active() may fail to activate\nthe zone and fall back to btrfs_zone_finish_one_bg() to free an active\nzone. That path waits for the running transaction to commit while still\nholding zoned_meta_io_lock, but the committer needs that same lock to\nwrite out the tree extents, so the two tasks deadlock:\n\n  Task A (kworker, metadata writeback)      Task B (fsstress, transaction commit)\n  ------------------------------------      -------------------------------------\n  wb_workfn()                               btrfs_commit_transaction(T)\n   btree_writepages()                        btrfs_write_and_wait_transaction()\n    btrfs_zoned_meta_io_lock()                btrfs_write_marked_extents()\n    btrfs_check_meta_write_pointer()           btree_writepages()\n     check_bg_is_active() [treelog_bg]          btrfs_zoned_meta_io_lock()\n      btrfs_zone_finish_one_bg()               \u003cblocks on zoned_meta_io_lock,\n       btrfs_zone_finish()                      held by Task A\u003e\n        do_zone_finish()\n         btrfs_inc_block_group_ro()\n          btrfs_wait_for_commit()\n           \u003cblocks waiting for commit\n            of transaction T, done by\n            Task B\u003e\n\nThe sibling branch in check_bg_is_active() already drops zoned_meta_io_lock\naround do_zone_finish() for this exact reason. Do the same in the tree-log\nbranch: release the lock around btrfs_zone_finish_one_bg() and re-acquire\nit afterwards. The lock only protects fs_info-\u003eactive_{meta,system}_bg,\nwhich this branch does not touch, and ctx-\u003ezoned_bg keeps a reference to\nthe block group across the unlock, so nothing is lost while the lock\nis dropped.\n\nThis hang occasionally reproduces with fstests generic/475 on a zoned\nbtrfs filesystem.","modified":"2026-08-30T05:26:50Z","published":"2026-08-15T13:18:02Z","upstream":["CVE-2026-74572"],"references":[{"type":"WEB","url":"https://nvd.nist.gov/vuln/detail/CVE-2026-74572"}],"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-96141.json"}}],"schema_version":"1.9.0"}