Re: [PATCH 5/9] KVM: cpuid: set CPUID(EAX=0xd,ECX=1).EBX correctly

[Wanpeng Li <wanpeng.li@linux.intel.com>]

Hi Paolo,
On Thu, Dec 04, 2014 at 04:57:10PM +0100, Paolo Bonzini wrote:
>This is the size of the XSAVES area.  This starts providing guest support
>for XSAVES (with no support yet for supervisor states, i.e. XSS == 0
>always in guests for now).
>
>Wanpeng Li suggested testing XSAVEC as well as XSAVES, since in practice
>no real processor exists that only has one of them, and there is no
>other way for userspace programs to compute the area of the XSAVEC
>save area.  CPUID(EAX=0xd,ECX=1).EBX provides an upper bound.
>
>Suggested-by: Radim Krčmář <rkrcmar@redhat.com>
>Reviewed-by: Radim Krčmář <rkrcmar@redhat.com>
>Signed-off-by: Wanpeng Li <wanpeng.li@linux.intel.com>
>Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
>---
> arch/x86/kvm/cpuid.c | 22 ++++++++++++++++------
> 1 file changed, 16 insertions(+), 6 deletions(-)
>
>diff --git a/arch/x86/kvm/cpuid.c b/arch/x86/kvm/cpuid.c
>index e24df01ab118..2f7bc2de9915 100644
>--- a/arch/x86/kvm/cpuid.c
>+++ b/arch/x86/kvm/cpuid.c
>@@ -23,7 +23,7 @@
> #include "mmu.h"
> #include "trace.h"
> 
>-static u32 xstate_required_size(u64 xstate_bv)
>+static u32 xstate_required_size(u64 xstate_bv, bool compacted)
> {
> 	int feature_bit = 0;
> 	u32 ret = XSAVE_HDR_SIZE + XSAVE_HDR_OFFSET;
>@@ -31,9 +31,10 @@ static u32 xstate_required_size(u64 xstate_bv)
> 	xstate_bv &= XSTATE_EXTEND_MASK;
> 	while (xstate_bv) {
> 		if (xstate_bv & 0x1) {
>-		        u32 eax, ebx, ecx, edx;
>+		        u32 eax, ebx, ecx, edx, offset;
> 		        cpuid_count(0xD, feature_bit, &eax, &ebx, &ecx, &edx);
>-			ret = max(ret, eax + ebx);
>+			offset = compacted ? ret : ebx;
>+			ret = max(ret, offset + eax);
> 		}
> 
> 		xstate_bv >>= 1;
>@@ -87,9 +88,13 @@ int kvm_update_cpuid(struct kvm_vcpu *vcpu)
> 			(best->eax | ((u64)best->edx << 32)) &
> 			kvm_supported_xcr0();
> 		vcpu->arch.guest_xstate_size = best->ebx =
>-			xstate_required_size(vcpu->arch.xcr0);
>+			xstate_required_size(vcpu->arch.xcr0, false);
> 	}
> 
>+	best = kvm_find_cpuid_entry(vcpu, 0xD, 1);
>+	if (best && (best->eax & (F(XSAVES)|F(XSAVEC))))
>+		best->ebx = xstate_required_size(vcpu->arch.xcr0, true);
>+
> 	/*
> 	 * The existing code assumes virtual address is 48-bit in the canonical
> 	 * address checks; exit if it is ever changed.
>@@ -470,9 +475,14 @@ static inline int __do_cpuid_ent(struct kvm_cpuid_entry2 *entry, u32 function,
> 				goto out;
> 
> 			do_cpuid_1_ent(&entry[i], function, idx);
>-			if (idx == 1)

When if (idx == 1) is added in this patchset? I suspect that you miss 
to add your patch "kvm: x86: mask out XSAVES" in this patchset.

Regards,
Wanpeng Li 

>+			if (idx == 1) {
> 				entry[i].eax &= kvm_supported_word10_x86_features;
>-			else if (entry[i].eax == 0 || !(supported & mask))
>+				entry[i].ebx = 0;
>+				if (entry[i].eax & (F(XSAVES)|F(XSAVEC)))
>+					entry[i].ebx =
>+						xstate_required_size(supported,
>+								     true);
>+			} else if (entry[i].eax == 0 || !(supported & mask))
> 				continue;
> 			entry[i].flags |=
> 			       KVM_CPUID_FLAG_SIGNIFCANT_INDEX;
>-- 
>1.8.3.1
>