2017-12-13 04:34:42

by Wanpeng Li

Subject: [PATCH v2] KVM: X86: Fix load bad host fpu state

From: Wanpeng Li <[email protected]>

------------[ cut here ]------------
Bad FPU state detected at kvm_put_guest_fpu+0xd8/0x2d0 [kvm], reinitializing FPU registers.
WARNING: CPU: 1 PID: 4594 at arch/x86/mm/extable.c:103 ex_handler_fprestore+0x88/0x90
CPU: 1 PID: 4594 Comm: qemu-system-x86 Tainted: G B OE 4.15.0-rc2+ #10
RIP: 0010:ex_handler_fprestore+0x88/0x90
Call Trace:
fixup_exception+0x4e/0x60
do_general_protection+0xff/0x270
general_protection+0x22/0x30
RIP: 0010:kvm_put_guest_fpu+0xd8/0x2d0 [kvm]
RSP: 0018:ffff8803d5627810 EFLAGS: 00010246
kvm_vcpu_reset+0x3b4/0x3c0 [kvm]
kvm_apic_accept_events+0x1c0/0x240 [kvm]
kvm_arch_vcpu_ioctl_run+0x1658/0x2fb0 [kvm]
kvm_vcpu_ioctl+0x479/0x880 [kvm]
do_vfs_ioctl+0x142/0x9a0
SyS_ioctl+0x74/0x80
do_syscall_64+0x15f/0x600

This can be reproduced by running any testcase in kvm-unit-tests since
the qemu userspace FPU context is not initialized, which results in the
init path from kvm_apic_accept_events() will load/put qemu userspace
FPU context w/o initialized. In addition, w/o this splatting we still
should initialize vcpu->arch.user_fpu instead of current->thread.fpu.
This patch fixes it by initializing qemu user FPU context during vCPU
creation.

Cc: Paolo Bonzini <[email protected]>
Cc: Radim Krčmář <[email protected]>
Cc: Rik van Riel <[email protected]>
Fixes: f775b13eedee (x86,kvm: move qemu/guest FPU switching out to vcpu_run)
Signed-off-by: Wanpeng Li <[email protected]>
---
v1 -> v2:
* initialize user_fpu in fx_init

arch/x86/kvm/x86.c | 4 +---
1 file changed, 1 insertion(+), 3 deletions(-)

diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c
index 5134a85..6a837fa 100644
--- a/arch/x86/kvm/x86.c
+++ b/arch/x86/kvm/x86.c
@@ -7278,11 +7278,8 @@ static int complete_emulated_mmio(struct kvm_vcpu *vcpu)

int kvm_arch_vcpu_ioctl_run(struct kvm_vcpu *vcpu, struct kvm_run *kvm_run)
{
- struct fpu *fpu = &current->thread.fpu;
int r;

- fpu__initialize(fpu);
-
kvm_sigset_activate(vcpu);

if (unlikely(vcpu->arch.mp_state == KVM_MP_STATE_UNINITIALIZED)) {
@@ -7709,6 +7706,7 @@ int kvm_arch_vcpu_ioctl_set_fpu(struct kvm_vcpu *vcpu, struct kvm_fpu *fpu)
static void fx_init(struct kvm_vcpu *vcpu)
{
fpstate_init(&vcpu->arch.guest_fpu.state);
+ fpstate_init(&vcpu->arch.user_fpu.state);
if (boot_cpu_has(X86_FEATURE_XSAVES))
vcpu->arch.guest_fpu.state.xsave.header.xcomp_bv =
host_xcr0 | XSTATE_COMPACTION_ENABLED;
--
2.7.4

2017-12-13 10:00:11

by David Hildenbrand

Subject: Re: [PATCH v2] KVM: X86: Fix load bad host fpu state

> if (unlikely(vcpu->arch.mp_state == KVM_MP_STATE_UNINITIALIZED)) {
> @@ -7709,6 +7706,7 @@ int kvm_arch_vcpu_ioctl_set_fpu(struct kvm_vcpu *vcpu, struct kvm_fpu *fpu)
> static void fx_init(struct kvm_vcpu *vcpu)
> {
> fpstate_init(&vcpu->arch.guest_fpu.state);
> + fpstate_init(&vcpu->arch.user_fpu.state);

Won't at least one of them trigger the WARN in fpstate_init?

> if (boot_cpu_has(X86_FEATURE_XSAVES))
> vcpu->arch.guest_fpu.state.xsave.header.xcomp_bv =
> host_xcr0 | XSTATE_COMPACTION_ENABLED;
>


--

Thanks,

David / dhildenb

2017-12-13 12:01:44

by Wanpeng Li

Subject: Re: [PATCH v2] KVM: X86: Fix load bad host fpu state

2017-12-13 17:59 GMT+08:00 David Hildenbrand <[email protected]>:
>
>> if (unlikely(vcpu->arch.mp_state == KVM_MP_STATE_UNINITIALIZED)) {
>> @@ -7709,6 +7706,7 @@ int kvm_arch_vcpu_ioctl_set_fpu(struct kvm_vcpu *vcpu, struct kvm_fpu *fpu)
>> static void fx_init(struct kvm_vcpu *vcpu)
>> {
>> fpstate_init(&vcpu->arch.guest_fpu.state);
>> + fpstate_init(&vcpu->arch.user_fpu.state);
>
> Won't at least one of them trigger the WARN in fpstate_init?

Nothing trigger the WARN during my testing on a Skylake box. And I
will not continue to push this one since there is an alternative
solution in kvm/queue.

Regards,
Wanpeng Li