{"id":"CVE-2026-98445","published":"2026-10-18T16:16:54.837","lastModified":"2026-10-19T10:18:53.613","description":"In the Linux kernel, the following vulnerability has been resolved:\n\nxfs: don't let memory failures leak blocks and kill repairs\n\nLOLLM complains that a memory allocation failure in\nxrep_newbt_add_blocks results in online repair leaking blocks that were\npreviously allocated to write a new btree, but the problem is worse than\nthat -- a limitation of the codebase is that the callers cannot undo the\ntransaction /and/ return the error -- either you undo all changes and\ncommit the transaction, or you error out and the filesystem goes down.\n\nHowever, the new btree space reservation object isn't that big (~48\nbytes).  Let's just do a NOFAIL allocation and the problem goes away.","cvssScore":null,"cvssSeverity":null,"cvssVector":null,"cwes":[],"vendors":[],"products":[],"references":[{"url":"https://git.kernel.org/stable/c/0e0d3372a84029da1768c7a4c4601484af5b58a2","tags":[]},{"url":"https://git.kernel.org/stable/c/3380151cf0ce8ad0c1d356fed51e4b69e1d28e50","tags":[]},{"url":"https://git.kernel.org/stable/c/ab1c416d2377cdc16123ef521aac4da1c468c3d4","tags":[]},{"url":"https://git.kernel.org/stable/c/b15c79fca28060fd9c7e77f2c3ebc4ba023974e8","tags":[]}],"exploitRefs":[],"hasPoc":false,"ai":null}