{ "dataType": "CVE_RECORD", "dataVersion": "5.2", "cveMetadata": { "cveId": "CVE-2025-38367", "assignerOrgId": "416baaa9-dc9f-4396-8d5f-8c081fb06d67", "state": "PUBLISHED", "assignerShortName": "Linux", "dateReserved": "2025-04-16T04:51:24.008Z", "datePublished": "2025-07-25T12:47:41.965Z", "dateUpdated": "2026-08-05T12:01:24.525Z" }, "containers": { "cna": { "providerMetadata": { "orgId": "416baaa9-dc9f-4396-8d5f-8c081fb06d67", "shortName": "Linux", "dateUpdated": "2026-08-05T12:01:24.525Z" }, "descriptions": [ { "lang": "en", "value": "In the Linux kernel, the following vulnerability has been resolved:\n\nLoongArch: KVM: Avoid overflow with array index\n\nThe variable index is modified and reused as array index when modify\nregister EIOINTC_ENABLE. There will be array index overflow problem." } ], "metrics": [ { "cvssV3_1": { "version": "3.1", "vectorString": "CVSS:3.1/AV:L/AC:L/PR:L/UI:N/S:C/C:H/I:H/A:H", "baseScore": 8.8, "baseSeverity": "HIGH" }, "scenarios": [ { "lang": "en", "value": "AV:L - The vulnerable code is reached only by a guest vCPU executing a privileged `iocsrwr*` instruction against the in-kernel EIOINTC at IOCSR 0x1400, trapped locally on the host via `kvm_emu_iocsr()` → `kvm_io_bus_write()`. The attacker must be executing code on the physical host inside a VM, not sending network traffic.\nAC:L - A single deterministic 2/4/8-byte IOCSR write to offset 0x200–0x21f reproduces the out-of-bounds indexing every time, with no race, no timing window, and no dependence on memory layout the attacker cannot influence. The attacker also pre-seeds the out-of-bounds source bytes by writing the adjacent COREMAP/COREISR registers first.\nPR:L - IOCSR instructions require PLV0, so the attacker needs kernel privilege inside the guest — which is exactly what an ordinary cloud/virtualization tenant already possesses in its own VM, and no privilege at all on the host. From the host security authority being attacked, this is a low-privileged, untrusted principal.\nUI:N - The guest triggers the flaw entirely on its own by executing one instruction; no host operator, VMM action, or victim interaction is needed.\nS:C - The vulnerable component is the host kernel's KVM EIOINTC emulation, while the attacker's authority is confined to the guest; a guest-initiated write causes out-of-bounds accesses in host kernel memory, crossing the guest→hypervisor security boundary.\nC:H - Guest input causes the host kernel to read up to ~192 bytes beyond the 32-byte `enable`/`isr` arrays, and those out-of-bounds host-memory values directly determine which interrupts are asserted/deasserted, giving the guest an observable oracle over host kernel memory it is not entitled to read.\nI:H - Values harvested out of bounds are fed into `eiointc_enable_irq()`/`eiointc_update_irq()`, which then modify host-maintained state — `coreisr` words, `sw_coreisr` bitmaps, and queued vCPU interrupt-pending bits — so attacker-influenced out-of-bounds data corrupts host kernel interrupt-controller state and forces arbitrary interrupt injection/withdrawal.\nA:H - The corrupted indexing drives `eiointc_update_irq()` for IRQs the guest could not otherwise reach, where `kvm_get_vcpu(s->kvm, s->sw_coremap[irq])` can return NULL and `kvm_vcpu_ioctl_interrupt()` dereferences it while `eiointc->lock` is held with interrupts disabled, producing a host kernel oops/hang; even short of that, spurious or permanently lost interrupt delivery wedges the VM." } ] } ], "affected": [ { "product": "Linux", "vendor": "Linux", "defaultStatus": "unaffected", "repo": "https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git", "programFiles": [ "arch/loongarch/kvm/intc/eiointc.c" ], "versions": [ { "version": "3956a52bc05bd811082a3c9d2b423ee957e6fefc", "lessThan": "2cc84c4b0d70d42e291862ecc848890d18e1004a", "status": "affected", "versionType": "git" }, { "version": "3956a52bc05bd811082a3c9d2b423ee957e6fefc", "lessThan": "080e8d2ecdfde588897aa8a87a8884061f4dbbbb", "status": "affected", "versionType": "git" } ] }, { "product": "Linux", "vendor": "Linux", "defaultStatus": "affected", "repo": "https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git", "programFiles": [ "arch/loongarch/kvm/intc/eiointc.c" ], "versions": [ { "version": "6.13", "status": "affected" }, { "version": "0", "lessThan": "6.13", "status": "unaffected", "versionType": "semver" }, { "version": "6.15.5", "lessThanOrEqual": "6.15.*", "status": "unaffected", "versionType": "semver" }, { "version": "6.16", "lessThanOrEqual": "*", "status": "unaffected", "versionType": "original_commit_for_fix" } ] } ], "cpeApplicability": [ { "nodes": [ { "operator": "OR", "negate": false, "cpeMatch": [ { "vulnerable": true, "criteria": "cpe:2.3:o:linux:linux_kernel:*:*:*:*:*:*:*:*", "versionStartIncluding": "6.13", "versionEndExcluding": "6.15.5" }, { "vulnerable": true, "criteria": "cpe:2.3:o:linux:linux_kernel:*:*:*:*:*:*:*:*", "versionStartIncluding": "6.13", "versionEndExcluding": "6.16" } ] } ] } ], "references": [ { "url": "https://git.kernel.org/stable/c/2cc84c4b0d70d42e291862ecc848890d18e1004a" }, { "url": "https://git.kernel.org/stable/c/080e8d2ecdfde588897aa8a87a8884061f4dbbbb" } ], "title": "LoongArch: KVM: Avoid overflow with array index", "x_generator": { "engine": "bippy-1.2.0" } } } }