{"id":"CVE-2026-64283","summary":"KVM: guest_memfd: Treat memslot binding offset+size as unsigned values","details":"In the Linux kernel, the following vulnerability has been resolved:\n\nKVM: guest_memfd: Treat memslot binding offset+size as unsigned values\n\nWhen binding a memslot to a guest_memfd file, treat the offset and size as\nunsigned values to fix a bug where the sum of the two can result in a false\nnegative when checking for overflow against the size of the file.  Passing\nunsigned values also avoids relying on somewhat obscure checks in other\nflows for safety, and tracks the offset and size as they are intended to be\ntracked, as unsigned values.\n\nOn 64-bit kernels, the number of pages a memslot contains and thus the size\n(and offset) of its guest_memfd binding are unsigned 64-bit values.  Taking\nthe offset+size as an loff_t instead of a uoff_t inadvertently converts\nthe unsigned value to a signed value if the offset and/or size is massive.\n\nLocally storing the offset and size as signed values is benign in and of\nitself (though even that is *extremely* difficult to discern), but\noperating on their sum is not.\n\nFor the offset, KVM explicitly checks against a negative value, which might\nseem like a bug as KVM could incorrectly reject a legitimate binding, but\nthat's not actually the case as KVM_CREATE_GUEST_MEMFD takes a signed value\nfor its size, i.e. a would-be-negative offset is also greater than the\nmaximum possible size of any guest_memfd file.\n\nRegarding the size, while KVM lacks an explicit check for a negative value,\ni.e. seemingly has a flawed overflow check, KVM restricts the number of\npages in a single memslot to the largest positive signed 32-bit value:\n\n        if (id \u003c KVM_USER_MEM_SLOTS &&\n            (mem-\u003ememory_size \u003e\u003e PAGE_SHIFT) \u003e KVM_MEM_MAX_NR_PAGES)\n                return -EINVAL;\n\nand so that maximum \"size\" will ever be is 0x7fffffff000.\n\nThe sum of the two is, however, problematic.  While the size is restricted\nby KVM's memslot logic, the offset is not, i.e. the offset is completely\nunchecked until the \"offset + size \u003e i_size_read(inode)\" check.  If the\noffset is the (nearly) largest possible _positive_ value, then adding size\nto the offset can result in a signed, negative 64-bit value.  When compared\nagainst the size of the file (guaranteed to be positive), the negative sum\nis always smaller, and KVM incorrectly allows the absurd offset.\n\nOpportunistically add missing includes in kvm_mm.h (instead of relying on\nits parents).","modified":"2026-07-27T04:03:20.973000396Z","published":"2026-07-25T08:49:25.642Z","database_specific":{"osv_generated_from":"https://github.com/CVEProject/cvelistV5/tree/main/cves/2026/64xxx/CVE-2026-64283.json","cna_assigner":"Linux"},"references":[{"type":"WEB","url":"https://git.kernel.org/stable/c/eba85fee7fc6cf28fec38a5bf3c378bef9a79ca6"},{"type":"WEB","url":"https://git.kernel.org/stable/c/f3a98d5881b9bd4807f49156143565f6aabcef1e"},{"type":"ADVISORY","url":"https://github.com/CVEProject/cvelistV5/tree/main/cves/2026/64xxx/CVE-2026-64283.json"},{"type":"ADVISORY","url":"https://nvd.nist.gov/vuln/detail/CVE-2026-64283"},{"type":"PACKAGE","url":"https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git"}],"affected":[{"ranges":[{"type":"GIT","repo":"https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git","events":[{"introduced":"a7800aa80ea4d5356b8474c2302812e9d4926fa6"},{"fixed":"f3a98d5881b9bd4807f49156143565f6aabcef1e"},{"fixed":"eba85fee7fc6cf28fec38a5bf3c378bef9a79ca6"}]}],"database_specific":{"source":"https://storage.googleapis.com/cve-osv-conversion/osv-output/CVE-2026-64283.json"}},{"package":{"name":"Kernel","ecosystem":"Linux"},"ranges":[{"type":"ECOSYSTEM","events":[{"introduced":"6.8.0"},{"fixed":"7.1.4"}]}],"database_specific":{"source":"https://storage.googleapis.com/cve-osv-conversion/osv-output/CVE-2026-64283.json"}}],"schema_version":"1.7.5"}