{"id":"CVE-2026-89986","published":"2026-09-16T11:17:09.653","lastModified":"2026-09-16T15:18:22.557","description":"In the Linux kernel, the following vulnerability has been resolved:\n\nmm/mempolicy: fix sleeping allocation in alloc_pages_bulk_weighted_interleave()\n\nsyzbot reported a sleeping function called from invalid context splat in\nbucket_table_alloc().\n\nWhen rhashtable_insert_slow() rehashes the table under rcu_read_lock(), it\ncalls bucket_table_alloc(..., GFP_ATOMIC | __GFP_NOWARN).  If the bucket\ntable allocation uses vmalloc, __vmalloc_node_range_noprof() invokes\nvm_area_alloc_pages() -> alloc_pages_bulk_mempolicy_noprof() with the\npassed GFP_ATOMIC flags.\n\nIf the current task has an MPOL_WEIGHTED_INTERLEAVE mempolicy,\nalloc_pages_bulk_weighted_interleave() is called and currently hardcodes\nGFP_KERNEL when allocating the temporary weights array, triggering a\nmight_alloc() splat in atomic/RCU contexts.\n\nPass the gfp flags (masked with GFP_RECLAIM_MASK to strip page-allocator\nzone modifiers like __GFP_HIGHMEM) received by\nalloc_pages_bulk_weighted_interleave() to kmalloc() instead of hardcoding\nGFP_KERNEL.  Since the weights buffer is immediately initialized in full,\nkmalloc() is sufficient.","cvssScore":7.8,"cvssSeverity":"HIGH","cvssVector":"CVSS:3.1/AV:L/AC:L/PR:L/UI:N/S:U/C:H/I:H/A:H","cwes":[],"vendors":[],"products":[],"references":[{"url":"https://git.kernel.org/stable/c/0ceda28f371df9e0bbdaa29214f71fe8298f23d8","tags":[]},{"url":"https://git.kernel.org/stable/c/2943f1f4b7f2816177060eb9f551f2e6d8b629ba","tags":[]},{"url":"https://git.kernel.org/stable/c/540e583b66d6402bf556fde5e53c817a54c1afe5","tags":[]},{"url":"https://git.kernel.org/stable/c/bcb3d0c867ee40dc48e9c085bf328fbc679b6656","tags":[]}],"exploitRefs":[],"hasPoc":false,"ai":null}