Christoph Egger
2010-Aug-05 15:03 UTC
[Xen-devel] [PATCH 08/14] Nested Virtualization: efer
Signed-off-by: Christoph Egger <Christoph.Egger@amd.com> -- ---to satisfy European Law for business letters: Advanced Micro Devices GmbH Einsteinring 24, 85609 Dornach b. Muenchen Geschaeftsfuehrer: Alberto Bozzo, Andrew Bowd Sitz: Dornach, Gemeinde Aschheim, Landkreis Muenchen Registergericht Muenchen, HRB Nr. 43632 _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
Dong, Eddie
2010-Aug-19 03:33 UTC
RE: [Xen-devel] [PATCH 08/14] Nested Virtualization: efer
> # HG changeset patch > # User cegger > # Date 1281003856 -7200 > Allow guest to enable SVM in EFER > > diff -r 7789f1d7d7e7 -r 8c7c597735c6 xen/arch/x86/hvm/hvm.c > --- a/xen/arch/x86/hvm/hvm.c > +++ b/xen/arch/x86/hvm/hvm.c > @@ -616,6 +616,7 @@ static bool_t hvm_efer_valid(uint64_t va > { > return !((value & ~efer_validbits) || > ((sizeof(long) != 8) && (value & EFER_LME)) || > + (!cpu_has_svm && (value & EFER_SVME)) || > (!cpu_has_nx && (value & EFER_NX)) || > (!cpu_has_syscall && (value & EFER_SCE)) || > (!cpu_has_lmsl && (value & EFER_LMSLE)) || > @@ -630,6 +631,7 @@ static int hvm_load_cpu_ctxt(struct doma > struct hvm_hw_cpu ctxt; > struct segment_register seg; > struct vcpu_guest_context *vc; > + uint64_t efer_validbits; > > /* Which vcpu is this? */ > vcpuid = hvm_load_instance(h); > @@ -669,9 +671,12 @@ static int hvm_load_cpu_ctxt(struct doma > return -EINVAL; > } > > - if ( !hvm_efer_valid( > - ctxt.msr_efer, > - EFER_FFXSE | EFER_LMSLE | EFER_LME | EFER_LMA | EFER_NX | > EFER_SCE) ) + efer_validbits = EFER_FFXSE | EFER_LMSLE | EFER_LME > | EFER_LMA + | EFER_NX | EFER_SCE; > + if ( nestedhvm_enabled(v->domain) ) > + efer_validbits |= EFER_SVME;Check it is running on SVM first.> + > + if ( !hvm_efer_valid(ctxt.msr_efer, efer_validbits) ) > { > gdprintk(XENLOG_ERR, "HVM restore: bad EFER 0x%"PRIx64"\n", > ctxt.msr_efer); > @@ -1081,14 +1086,18 @@ bool_t hvm_hap_nested_page_fault(unsigne > int hvm_set_efer(uint64_t value) > { > struct vcpu *v = current; > + uint64_t efer_validbits; > > value &= ~EFER_LMA; > > - if ( !hvm_efer_valid(value, > - EFER_FFXSE | EFER_LMSLE | EFER_LME | EFER_NX | EFER_SCE) > ) + efer_validbits = EFER_FFXSE | EFER_LMSLE | EFER_LME | EFER_NX > | EFER_SCE; + if ( nestedhvm_enabled(v->domain) ) > + efer_validbits |= EFER_SVME;ditto> + > + if ( !hvm_efer_valid(value, efer_validbits) ) > { > gdprintk(XENLOG_WARNING, "Trying to set reserved bit in " > - "EFER: %"PRIx64"\n", value); > + "EFER: 0x%"PRIx64"\n", value); > hvm_inject_exception(TRAP_gp_fault, 0, 0); > return X86EMUL_EXCEPTION; > } > diff -r 7789f1d7d7e7 -r 8c7c597735c6 xen/include/asm-x86/cpufeature.h > --- a/xen/include/asm-x86/cpufeature.h > +++ b/xen/include/asm-x86/cpufeature.h > @@ -199,6 +199,8 @@ > > #define cpu_has_rdtscp boot_cpu_has(X86_FEATURE_RDTSCP) > > +#define cpu_has_svm boot_cpu_has(X86_FEATURE_SVME) > + > #endif /* __ASM_I386_CPUFEATURE_H */ > > /*_______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
Tim Deegan
2010-Aug-19 08:41 UTC
Re: [Xen-devel] [PATCH 08/14] Nested Virtualization: efer
At 04:33 +0100 on 19 Aug (1282192384), Dong, Eddie wrote:> > # HG changeset patch > > # User cegger > > # Date 1281003856 -7200 > > Allow guest to enable SVM in EFER > > > > diff -r 7789f1d7d7e7 -r 8c7c597735c6 xen/arch/x86/hvm/hvm.c > > --- a/xen/arch/x86/hvm/hvm.c > > +++ b/xen/arch/x86/hvm/hvm.c > > @@ -616,6 +616,7 @@ static bool_t hvm_efer_valid(uint64_t va > > { > > return !((value & ~efer_validbits) || > > ((sizeof(long) != 8) && (value & EFER_LME)) || > > + (!cpu_has_svm && (value & EFER_SVME)) || > > (!cpu_has_nx && (value & EFER_NX)) || > > (!cpu_has_syscall && (value & EFER_SCE)) || > > (!cpu_has_lmsl && (value & EFER_LMSLE)) || > > @@ -630,6 +631,7 @@ static int hvm_load_cpu_ctxt(struct doma > > struct hvm_hw_cpu ctxt; > > struct segment_register seg; > > struct vcpu_guest_context *vc; > > + uint64_t efer_validbits; > > > > /* Which vcpu is this? */ > > vcpuid = hvm_load_instance(h); > > @@ -669,9 +671,12 @@ static int hvm_load_cpu_ctxt(struct doma > > return -EINVAL; > > } > > > > - if ( !hvm_efer_valid( > > - ctxt.msr_efer, > > - EFER_FFXSE | EFER_LMSLE | EFER_LME | EFER_LMA | EFER_NX | > > EFER_SCE) ) + efer_validbits = EFER_FFXSE | EFER_LMSLE | EFER_LME > > | EFER_LMA + | EFER_NX | EFER_SCE; > > + if ( nestedhvm_enabled(v->domain) ) > > + efer_validbits |= EFER_SVME; > > Check it is running on SVM first.This check is made in hvm_efer_valid(). But it might be less confusing to have hvm_efer_valid() check nestedhvm_enabled() as well rather than do half the check in the caller and half in the function. Tim. -- Tim Deegan <Tim.Deegan@citrix.com> Principal Software Engineer, XenServer Engineering Citrix Systems UK Ltd. (Company #02937203, SL9 0BG) _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
Christoph Egger
2010-Aug-19 10:18 UTC
Re: [Xen-devel] [PATCH 08/14] Nested Virtualization: efer
On Thursday 19 August 2010 10:41:03 Tim Deegan wrote:> At 04:33 +0100 on 19 Aug (1282192384), Dong, Eddie wrote: > > > # HG changeset patch > > > # User cegger > > > # Date 1281003856 -7200 > > > Allow guest to enable SVM in EFER > > > > > > diff -r 7789f1d7d7e7 -r 8c7c597735c6 xen/arch/x86/hvm/hvm.c > > > --- a/xen/arch/x86/hvm/hvm.c > > > +++ b/xen/arch/x86/hvm/hvm.c > > > @@ -616,6 +616,7 @@ static bool_t hvm_efer_valid(uint64_t va > > > { > > > return !((value & ~efer_validbits) || > > > ((sizeof(long) != 8) && (value & EFER_LME)) || > > > + (!cpu_has_svm && (value & EFER_SVME)) || > > > (!cpu_has_nx && (value & EFER_NX)) || > > > (!cpu_has_syscall && (value & EFER_SCE)) || > > > (!cpu_has_lmsl && (value & EFER_LMSLE)) || > > > @@ -630,6 +631,7 @@ static int hvm_load_cpu_ctxt(struct doma > > > struct hvm_hw_cpu ctxt; > > > struct segment_register seg; > > > struct vcpu_guest_context *vc; > > > + uint64_t efer_validbits; > > > > > > /* Which vcpu is this? */ > > > vcpuid = hvm_load_instance(h); > > > @@ -669,9 +671,12 @@ static int hvm_load_cpu_ctxt(struct doma > > > return -EINVAL; > > > } > > > > > > - if ( !hvm_efer_valid( > > > - ctxt.msr_efer, > > > - EFER_FFXSE | EFER_LMSLE | EFER_LME | EFER_LMA | EFER_NX | > > > EFER_SCE) ) + efer_validbits = EFER_FFXSE | EFER_LMSLE | EFER_LME > > > > > > | EFER_LMA + | EFER_NX | EFER_SCE; > > > > > > + if ( nestedhvm_enabled(v->domain) ) > > > + efer_validbits |= EFER_SVME; > > > > Check it is running on SVM first. > > This check is made in hvm_efer_valid(). But it might be less confusing > to have hvm_efer_valid() check nestedhvm_enabled() as well rather than > do half the check in the caller and half in the function.Good point. Fixed in local tree. Christoph -- ---to satisfy European Law for business letters: Advanced Micro Devices GmbH Einsteinring 24, 85609 Dornach b. Muenchen Geschaeftsfuehrer: Alberto Bozzo, Andrew Bowd Sitz: Dornach, Gemeinde Aschheim, Landkreis Muenchen Registergericht Muenchen, HRB Nr. 43632 _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel