{"id":"CVE-2026-90262","published":"2026-09-17T17:17:22.640","lastModified":"2026-09-17T17:17:22.640","description":"In the Linux kernel, the following vulnerability has been resolved:\n\nbtrfs: retry verity reads for not-uptodate Merkle folios\n\nbtrfs_read_merkle_tree_page() can find a folio in the mapping that is not\nuptodate.  After taking the folio lock, the current code treats that state\nas a read error and returns -EIO.\n\nThat can make a previous transient read failure sticky.  If the failed read\nleft a not-uptodate folio in the mapping, later callers find that folio and\nfail instead of retrying the read.\n\nKeep the existing page-cache insertion and locking order, but retry the\nMerkle item read when a not-uptodate folio is found in the mapping.  Also\nunlock the folio when read_key_bytes() fails so that a later caller can\nlock it and retry the read.","cvssScore":null,"cvssSeverity":null,"cvssVector":null,"cwes":[],"vendors":[],"products":[],"references":[{"url":"https://git.kernel.org/stable/c/12b6d1a1715cbced2e445ca353f9c9987b8636e2","tags":[]},{"url":"https://git.kernel.org/stable/c/81241f734f0f662378f5ffc53882b012923e6fe5","tags":[]},{"url":"https://git.kernel.org/stable/c/8cc569696dac51fc62bb39b3b8f530582b916d29","tags":[]},{"url":"https://git.kernel.org/stable/c/90e9eae1b5907fa36620ffb7f4f1a4afa9333427","tags":[]},{"url":"https://git.kernel.org/stable/c/c1fa005cdf3b7ff14cdfd7d512830088a3fc256b","tags":[]}],"exploitRefs":[],"hasPoc":false,"ai":null}