2014-12-03 15:26+0100, Paolo Bonzini:
>
>
> On 03/12/2014 15:23, Nadav Amit wrote:
> > I think it is better just to replace the last line with:
> >
> > *(u64 *)(dest + XSAVE_HDR_OFFSET) = xsave->xsave_hdr.xstate_bv
Yeah, or we can use this value for xstate_bv to save some copying too,
diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c
index 19e5e8f..ba2b7bd 100644
--- a/arch/x86/kvm/x86.c
+++ b/arch/x86/kvm/x86.c
@@ -3137,7 +3137,7 @@ static int kvm_vcpu_ioctl_x86_set_debugregs(struct
kvm_vcpu *vcpu,
static void fill_xsave(u8 *dest, struct kvm_vcpu *vcpu)
{
struct xsave_struct *xsave = &vcpu->arch.guest_fpu.state->xsave;
- u64 xstate_bv = vcpu->arch.guest_supported_xcr0 | XSTATE_FPSSE;
+ u64 xstate_bv = xsave->xsave_hdr.xstate_bv;
u64 valid;
/*
> Right, this matches
>
> u64 xstate_bv = *(u64 *)(src + XSAVE_HDR_OFFSET);
> ...
> xsave->xsave_hdr.xstate_bv = xstate_bv;
>
> in load_xsave.
Btw, we don't care about crashers from userspace?
---8<---
KVM: x86: prevent #GP with malicious xsave
XRSTORS throws #GP when XSTATE_BV isn't a subset of XCOMP_BV.
Make it so.
SDM: XRSTORS Exceptions
#GP If a bit in the XCOMP_BV field in the XSAVE header is 0 and the
corresponding bit in the XSTATE_BV field is 1.
(Also in SDM: 13.11 OPERATION OF XRSTORS)
Signed-off-by: Radim Krčmář <rkrc...@redhat.com>
---
arch/x86/kvm/x86.c | 4 +++-
1 file changed, 3 insertions(+), 1 deletion(-)
diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c
index ca26681..19e5e8f 100644
--- a/arch/x86/kvm/x86.c
+++ b/arch/x86/kvm/x86.c
@@ -3184,8 +3184,10 @@ static void load_xsave(struct kvm_vcpu *vcpu, u8 *src)
/* Set XSTATE_BV and possibly XCOMP_BV. */
xsave->xsave_hdr.xstate_bv = xstate_bv;
- if (cpu_has_xsaves)
+ if (cpu_has_xsaves) {
xsave->xsave_hdr.xcomp_bv = host_xcr0 |
XSTATE_COMPACTION_ENABLED;
+ xsave->xsave_hdr.xstate_bv &= xsave->xsave_hdr.xcomp_bv;
+ }
/*
* Copy each region from the non-compacted offset to the
--
2.2.0
--
To unsubscribe from this list: send the line "unsubscribe kvm" in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
