{"id":"CVE-2026-90275","published":"2026-09-17T17:17:24.130","lastModified":"2026-09-17T17:17:24.130","description":"In the Linux kernel, the following vulnerability has been resolved:\n\nmd/raid1: don't set array_frozen in raid1_takeover()\n\nraid1_takeover() sets conf->array_frozen = 1 on the newly-allocated\nr1conf and nothing ever clears it, so every I/O to the array stalls\npermanently once _wait_barrier() sees it stuck at 1.\n\nThis used to be harmless: level_store() called mddev_resume() right\nafter pers->run(), which called raid1_quiesce(mddev, 0) and cleared\narray_frozen back to 0 regardless of what raid1_takeover() set. Commit\nb39f35ebe86d (\"md: don't quiesce in mddev_suspend()\") removed that\nquiesce(mddev, 0) call, so the pre-set now sticks.\n\nsetup_conf() already zero-initializes the new r1conf via kzalloc, so\njust don't set array_frozen here.\n\nSame class of bug as commit 892da88d1cd9 (\"md/raid10: fix a\n'conf->barrier' leakage in raid10_takeover()\"), also triggered by\nb39f35ebe86d.","cvssScore":null,"cvssSeverity":null,"cvssVector":null,"cwes":[],"vendors":[],"products":[],"references":[{"url":"https://git.kernel.org/stable/c/668d99ba47210ab6ceb18f1cf022c0accca7e3bb","tags":[]},{"url":"https://git.kernel.org/stable/c/b6ec4bf1ec4446c0213c08ea493e48e6b83cd1af","tags":[]},{"url":"https://git.kernel.org/stable/c/c57ec2749ee03cace564eec7bb37c89dc3d8d47b","tags":[]},{"url":"https://git.kernel.org/stable/c/dc386aa0ac0a3ec06c9a3ea9b064b073fb72a916","tags":[]},{"url":"https://git.kernel.org/stable/c/f9cdb5bb8efbf401e84d850828271af87e6be2dc","tags":[]}],"exploitRefs":[],"hasPoc":false,"ai":null}