{"id":"CVE-2026-90260","published":"2026-09-17T17:17:22.433","lastModified":"2026-09-18T18:17:51.297","description":"In the Linux kernel, the following vulnerability has been resolved:\n\nbtrfs: zoned: don't clobber the extent buffer when zeroing it out\n\nOn a zoned filesystem a freed-but-still-dirty tree block is written out\nas zeros (EXTENT_BUFFER_ZONED_ZEROOUT) only to keep the zone write\npointer advancing. btree_csum_one_bio() implemented this by memzeroing\nthe extent buffer's own folios before submission.\n\nThat destroys the in-memory buffer while it may still be referenced. In\nparticular btrfs_free_tree_block() can run on it afterwards and reads\nthe header to add a delayed reference; once the header has been zeroed\nit frees bytenr 0 and corrupts the extent tree (the\nbtrfs_header_bytenr(buf) != 0 ASSERT in btrfs_free_tree_block(), or an\n\"unable to find ref\" abort). It is flaky and reproduces under fsstress,\ne.g. generic/461 and generic/013.\n\nWrite the zeros to disk from the shared zero page instead and leave the\nextent buffer content untouched, so any later reference - including the\ndelayed reference from btrfs_free_tree_block() - still sees a valid\nheader. end_bbio_meta_write() now clears writeback on the buffer's own\nfolios, as the bio no longer carries them.","cvssScore":7.1,"cvssSeverity":"HIGH","cvssVector":"CVSS:3.1/AV:L/AC:L/PR:L/UI:N/S:U/C:N/I:H/A:H","cwes":[],"vendors":[],"products":[],"references":[{"url":"https://git.kernel.org/stable/c/92484ad014f2e2b0e0e7bbc1610951e978cf557c","tags":[]},{"url":"https://git.kernel.org/stable/c/98e3747587f9f87f010d8d7e55786216249a94af","tags":[]},{"url":"https://git.kernel.org/stable/c/db4b9eefc8ee0bcaeee4d5e6a7313905f6a2fe7c","tags":[]}],"exploitRefs":[],"hasPoc":false,"ai":null}