{"id":"CVE-2026-63799","title":"In the Linux kernel, the following vulnerability has been resolved:\n\nsched/mmcid: Fix OOB clear_bit when CID is MM_CID_UNSET in fixup path\n\nIn mm_cid_fixup_cpus_to_tasks(), when rq->curr has the target mm and\nmm_cid.active is set, the CI…","summary":"In the Linux kernel, the following vulnerability has been resolved:\n\nsched/mmcid: Fix OOB clear_bit when CID is MM_CID_UNSET in fixup path\n\nIn mm_cid_fixup_cpus_to_tasks(), when rq->curr has the target mm and\nmm_cid.active is set, the CI…","severity":"none","published":"2026-07-19","updated":"2026-07-19","source":"NVD","sourceUrl":"https://nvd.nist.gov/vuln/detail/CVE-2026-63799","references":[{"url":"https://git.kernel.org/stable/c/8d32856fb72ba976d9c87ba405fd17e80419934c","label":"416baaa9-dc9f-4396-8d5f-8c081fb06d67"},{"url":"https://git.kernel.org/stable/c/de3ab9bd3133899efb92e4cd05ba4203e58fc0a3","label":"416baaa9-dc9f-4396-8d5f-8c081fb06d67"}],"tags":["nvd"],"ingestedAt":"2026-07-20T00:35:12.296Z","epss":0.00128,"epssPercentile":0.02016,"slug":"CVE-2026-63799","body":"## Overview\n\nIn the Linux kernel, the following vulnerability has been resolved:\n\nsched/mmcid: Fix OOB clear_bit when CID is MM_CID_UNSET in fixup path\n\nIn mm_cid_fixup_cpus_to_tasks(), when rq->curr has the target mm and\nmm_cid.active is set, the CID is checked with cid_in_transit() before\nsetting the transition bit.  In per-CPU mode a newly forked or exec'd\ntask can be running with mm_cid.cid == MM_CID_UNSET because CIDs are\nassigned lazily on schedule-in.  With cid_in_transit() the guard passes\nfor MM_CID_UNSET (no transit bit), converts it to MM_CID_UNSET |\nMM_CID_TRANSIT and stores it back; later mm_cid_schedout() feeds this\nto clear_bit() with MM_CID_UNSET as the bit number, triggering an\nout-of-bounds write.\n\nSymptoms: this is genuine memory corruption, but a bounded out-of-bounds\nwrite, not an arbitrary one.  MM_CID_UNSET is the fixed sentinel BIT(31),\nso once the bad value reaches mm_cid_schedout() the cid_from_transit_cid()\nstrip leaves MM_CID_UNSET, which fails the \"cid < max_cids\" convergence\ntest and falls into mm_drop_cid() -> clear_bit(MM_CID_UNSET,\nmm_cidmask(mm)).  The cid bitmap is embedded in the mm_struct slab object\n(after cpu_bitmap and mm_cpus_allowed) and is only num_possible_cpus()\nbits wide, so clearing bit 31 is a deterministic OOB bit-clear at a\nfixed offset of 2^31 / 8 == 256 MiB past the bitmap base.  The address is\nnot attacker-influenced (fixed sentinel -> fixed offset) and the op only\nclears a single bit; what sits 256 MiB further along the direct map is\nwhatever kernel object happens to live there, so this corrupts one bit of\nunpredictable kernel memory -- it is not an arbitrary-address or\narbitrary-value write.\n\nIt triggers only in per-CPU CID mode, when a CPU is running an active\ntask of the target mm whose cid is still MM_CID_UNSET -- the\nfork()/execve() window before that task's next schedule-in assigns it a\nreal CID -- and a per-CPU -> per-task fixup walks over it (the mode\nfallback driven by a thread exit, sched_mm_cid_exit(), or by the deferred\nmax_cids recompute in mm_cid_work_fn()).\n\nIn practice syzkaller surfaced it as a KASAN use-after-free reported in\n__schedule -> mm_cid_switch_to, where the offending clear_bit() is inlined\nvia mm_cid_schedout() -> mm_drop_cid().\n\nGuard the transition-bit assignment against MM_CID_UNSET, in addition to\nthe existing cid_in_transit() check, so the bit is only set on a genuine\ntask-owned CID.  A CPU-owned (MM_CID_ONCPU) CID of a running active task\nis handled by the cid_on_cpu(pcp->cid) branch above and never reaches\nthis path, so excluding MM_CID_UNSET (and the already-transitioning case)\nis sufficient.\n\n## Remediation\n\nRefer to the linked advisories for vendor-supplied fixes and affected version ranges.","depth":"sunlit","depthScore":3,"depthScoreParts":{"impact":2.8,"likelihood":0,"exploitation":0,"ransomware":0},"changes":[]}