Instituto Nacional de ciberseguridad. Sección Incibe
Instituto Nacional de Ciberseguridad. Sección INCIBE-CERT

CVE-2025-40026

Gravedad:
Pendiente de análisis
Tipo:
No Disponible / Otro tipo
Fecha de publicación:
28/10/2025
Última modificación:
30/10/2025

Descripción

*** Pendiente de traducción *** In the Linux kernel, the following vulnerability has been resolved:<br /> <br /> KVM: x86: Don&amp;#39;t (re)check L1 intercepts when completing userspace I/O<br /> <br /> When completing emulation of instruction that generated a userspace exit<br /> for I/O, don&amp;#39;t recheck L1 intercepts as KVM has already finished that<br /> phase of instruction execution, i.e. has already committed to allowing L2<br /> to perform I/O. If L1 (or host userspace) modifies the I/O permission<br /> bitmaps during the exit to userspace, KVM will treat the access as being<br /> intercepted despite already having emulated the I/O access.<br /> <br /> Pivot on EMULTYPE_NO_DECODE to detect that KVM is completing emulation.<br /> Of the three users of EMULTYPE_NO_DECODE, only complete_emulated_io() (the<br /> intended "recipient") can reach the code in question. gp_interception()&amp;#39;s<br /> use is mutually exclusive with is_guest_mode(), and<br /> complete_emulated_insn_gp() unconditionally pairs EMULTYPE_NO_DECODE with<br /> EMULTYPE_SKIP.<br /> <br /> The bad behavior was detected by a syzkaller program that toggles port I/O<br /> interception during the userspace I/O exit, ultimately resulting in a WARN<br /> on vcpu-&gt;arch.pio.count being non-zero due to KVM no completing emulation<br /> of the I/O instruction.<br /> <br /> WARNING: CPU: 23 PID: 1083 at arch/x86/kvm/x86.c:8039 emulator_pio_in_out+0x154/0x170 [kvm]<br /> Modules linked in: kvm_intel kvm irqbypass<br /> CPU: 23 UID: 1000 PID: 1083 Comm: repro Not tainted 6.16.0-rc5-c1610d2d66b1-next-vm #74 NONE<br /> Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS 0.0.0 02/06/2015<br /> RIP: 0010:emulator_pio_in_out+0x154/0x170 [kvm]<br /> PKRU: 55555554<br /> Call Trace:<br /> <br /> kvm_fast_pio+0xd6/0x1d0 [kvm]<br /> vmx_handle_exit+0x149/0x610 [kvm_intel]<br /> kvm_arch_vcpu_ioctl_run+0xda8/0x1ac0 [kvm]<br /> kvm_vcpu_ioctl+0x244/0x8c0 [kvm]<br /> __x64_sys_ioctl+0x8a/0xd0<br /> do_syscall_64+0x5d/0xc60<br /> entry_SYSCALL_64_after_hwframe+0x4b/0x53<br />

Impacto