{"id":"CVE-2026-89985","aliases":[],"url":"https://o3.security/vulnerability/CVE-2026-89985","summary":"In the Linux kernel, the following vulnerability has been resolved:\n\nmemcg: keep folio's objcg same as its node\n\nmemcg_reparent_objcgs() has an inherent assumption that a folio's objcg…","details":"In the Linux kernel, the following vulnerability has been resolved:\n\nmemcg: keep folio's objcg same as its node\n\nmemcg_reparent_objcgs() has an inherent assumption that a folio's objcg is\nthe objcg of the folio's node.  Folio migration across nodes breaks that\nassumption: the new folio simply inherits the old folio's objcg while\nliving on a different node.\n\nOnce the assumption is broken, the reparenting of the folio's objcg and\nthe reparenting of the folio's LRU list are no longer atomic. \nmemcg_reparent_objcgs() handles one node per iteration and drops all the\nlocks in between, so the objcg gets reparented in the iteration for the\nobjcg's node while the LRU list gets spliced in the iteration for the\nfolio's node.  Any LRU operation on that folio in between resolves its\nlruvec through the objcg, and thus takes the lru_lock of the wrong memcg,\nnot the lru_lock of the list the folio is actually on.\n\nFix this by selecting the objcg by folio_nid() at charge time, and by\nre-deriving it for the destination node in mem_cgroup_migrate() and\nmem_cgroup_replace_folio().","published":"2026-09-16T11:17:09.547","modified":"2026-09-16T11:17:09.547","cvss":null,"epss":null,"cisaKev":null,"exploitsKnown":null,"affectedPackages":[],"fix":null,"references":[{"type":"WEB","url":"https://git.kernel.org/stable/c/6165478eaa3094eaee1d5e33b12520064faf043d"},{"type":"WEB","url":"https://git.kernel.org/stable/c/bf4ade7dbd76d4ec8697840e4ebb15ed77c5ec26"}],"provenance":{"sources":["OSV.dev","NVD","FIRST.org (EPSS)"],"lastVerified":"2026-09-16T11:17:09.547"}}