{"id":"CVE-2026-89917","published":"2026-09-16T11:17:00.763","lastModified":"2026-09-16T11:17:00.763","description":"In the Linux kernel, the following vulnerability has been resolved:\n\nKVM: arm64: Handle VNCR TLB invalidation race with vcpu_put() VNCR unmapping\n\nWhile VNCR TLB invalidation always occurs under the MMU lock,\nvcpu_put() doesn't, while it unmaps the VNCR page.\n\nThe problem is that the invalidation evaluates vncr_tlb::cpu to\ndecide whether an unmapping needs to take place (cpu != -1) before\nperforming it. On the other hand, this_cpu_reset_vncr_fixmap()\nunconditionally unmaps if L1_VNCR_MAPPED is set.\n\nThese two obviously can race, with a TOCTOU pattern on the TLBI\npath, and a BUG_ON() on the vcpu_put() path. And the two can end-up\ncalling vncr_fixmap(-1), with extra lethal effects.\n\nMove the reset of vncr_tlb::cpu to -1 to a common function, and make\nthis update atomic so that only a single thread can reset the field\nand perform the corresponding unmap. The vcpu_put() still need to\nunconditionally unmap the current VNCR to close another ugly race.\n\nFinally, the assignment of vncr_tlb::cpu is moved to be kept in sync\nwith the actual mapping, similar to L1_VNCR_MAPPED being set.","cvssScore":null,"cvssSeverity":null,"cvssVector":null,"cwes":[],"vendors":[],"products":[],"references":[{"url":"https://git.kernel.org/stable/c/38640bc32be3fcf9526d477155bc19d3f146231f","tags":[]},{"url":"https://git.kernel.org/stable/c/5cbd8761d001c11ada73b1753d772c4745a70e72","tags":[]},{"url":"https://git.kernel.org/stable/c/9065c1261f8a05c6a7a2d90048c7c0665ab2896d","tags":[]}],"exploitRefs":[],"hasPoc":false,"ai":null}