On 08/02/2018 08:32, Wanpeng Li wrote: > From: Wanpeng Li <wanpengli@xxxxxxxxxxx> > > Reported by syzkaller: > > WARNING: CPU: 6 PID: 2434 at arch/x86/kvm/vmx.c:6660 handle_ept_misconfig+0x54/0x1e0 [kvm_intel] > CPU: 6 PID: 2434 Comm: repro_test Not tainted 4.15.0+ #4 > RIP: 0010:handle_ept_misconfig+0x54/0x1e0 [kvm_intel] > Call Trace: > vmx_handle_exit+0xbd/0xe20 [kvm_intel] > kvm_arch_vcpu_ioctl_run+0xdaf/0x1d50 [kvm] > kvm_vcpu_ioctl+0x3e9/0x720 [kvm] > do_vfs_ioctl+0xa4/0x6a0 > SyS_ioctl+0x79/0x90 > entry_SYSCALL_64_fastpath+0x25/0x9c > > The syzkaller creates a former thread to issue KVM_SMI ioctl, and then creates > a latter thread to mmap and operate on the same vCPU, rsm emulation will not be > executed since there is no something like seabios which implements smi handler > when running syzkaller directly. This triggers a race condition when running > the testcase with multiple threads. Sometimes one thread exit w/ SHUTDOWN > reason, another thread mmaps and operates on the same vCPU, it continues to > use CS=0x30000, IP=0x8000 to access the address of SMI handler which results > in the above ept misconfig. This patch fixes it by returning RET_PF_EMULATE > in kvm_handle_bad_page() which would cause an emulation failure and then an > exit with KVM_EXIT_INTERNAL_ERROR, the VM will stop instead of resetting and > be taken advantage of. > > Reported-by: syzbot+c1d9517cab094dae65e446c0c5b4de6c40f4dc58@xxxxxxxxxxxxxxxxxxxxxxxxx > Cc: Paolo Bonzini <pbonzini@xxxxxxxxxx> > Cc: Radim Krčmář <rkrcmar@xxxxxxxxxx> > Signed-off-by: Wanpeng Li <wanpengli@xxxxxxxxxxx> > --- > v1 -> v2: > * return RET_PF_EMULATE > > arch/x86/kvm/mmu.c | 2 +- > 1 file changed, 1 insertion(+), 1 deletion(-) > > diff --git a/arch/x86/kvm/mmu.c b/arch/x86/kvm/mmu.c > index 8eca1d0..6c5a82c 100644 > --- a/arch/x86/kvm/mmu.c > +++ b/arch/x86/kvm/mmu.c > @@ -3029,7 +3029,7 @@ static int kvm_handle_bad_page(struct kvm_vcpu *vcpu, gfn_t gfn, kvm_pfn_t pfn) > return RET_PF_RETRY; > } > > - return -EFAULT; > + return RET_PF_EMULATE; > } > > static void transparent_hugepage_adjust(struct kvm_vcpu *vcpu, > Reviewed-by: Paolo Bonzini <pbonzini@xxxxxxxxxx>
