{"id":"CVE-2026-93233","published":"2026-09-24T16:17:18.887","lastModified":"2026-09-24T16:17:18.887","description":"In the Linux kernel, the following vulnerability has been resolved:\n\ndrm/nouveau/dmem: fix callocated underflow on large folio split\n\nnouveau_dmem_folio_free() drops chunk->callocated once per freed folio,\nwhile a large (compound) device-private folio is only counted once when\nit is allocated.  When such a folio is split, the mm core invokes\n->folio_split() (nouveau_dmem_folio_split()) once for each new\nsub-folio, but the hook only fixes up the sub-folio metadata and leaves\nchunk->callocated unchanged.\n\nEach resulting sub-folio is later freed separately, so after a split\nthe single allocation (+1) is met by N frees (-N), leaving\nchunk->callocated short by N-1.  On the first split/free cycle it\nunderflows: WARN_ON(!chunk->callocated) fires, the unsigned counter\nwraps and never returns to zero, so the chunk can no longer be\nreclaimed (nouveau_dmem_fini() also warns on the leaked count).\n\nAccount for the new sub-folio in the split hook, under the same lock as\nnouveau_dmem_folio_free(), so the count stays balanced.","cvssScore":null,"cvssSeverity":null,"cvssVector":null,"cwes":[],"vendors":[],"products":[],"references":[{"url":"https://git.kernel.org/stable/c/0163b92946d0a7c816f619c57a4d9867954dd106","tags":[]},{"url":"https://git.kernel.org/stable/c/c2256c044a1df39c8aad4dd2d6f709b2533e2d7a","tags":[]}],"exploitRefs":[],"hasPoc":false,"ai":null}