lkml.org 
[lkml]   [2014]   [Oct]   [23]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
From
SubjectRe: [PATCH] KVM: x86: Enable Intel AVX-512 for guest
On Wed, Oct 22, 2014 at 12:17:33PM +0200, Paolo Bonzini wrote:
> On 10/22/2014 11:35 AM, Chao Peng wrote:
> > Expose Intel AVX-512 feature bits to guest. Also add checks for
> > xcr0 AVX512 related bits according to spec:
> > http://download-software.intel.com/sites/default/files/managed/71/2e/319433-017.pdf
> >
> > Signed-off-by: Chao Peng <chao.p.peng@linux.intel.com>
>
> The patch looks good, but you also have to patch QEMU in order to
> save/restore the values of the registers. IIRC the manual already
> details where the registers are in the XSAVE area, so it should be easy
> to get them in and out. You can look at the MPX patches for an example.
>
> In the meanwhile, kernel bits are
>
> Reviewed-by: Paolo Bonzini <pbonzini@redhat.com>
>
> Paolo

Thanks Paolo.

QEMU side patch is already sent out to QEMU list in the other thread.
Also accessible from url:
http://lists.nongnu.org/archive/html/qemu-devel/2014-10/msg02681.html

Chao
>
> > ---
> > arch/x86/include/asm/xsave.h | 1 +
> > arch/x86/kvm/cpuid.c | 3 ++-
> > arch/x86/kvm/x86.c | 6 ++++++
> > arch/x86/kvm/x86.h | 3 ++-
> > 4 files changed, 11 insertions(+), 2 deletions(-)
> >
> > diff --git a/arch/x86/include/asm/xsave.h b/arch/x86/include/asm/xsave.h
> > index 7e7a79a..5fa9770 100644
> > --- a/arch/x86/include/asm/xsave.h
> > +++ b/arch/x86/include/asm/xsave.h
> > @@ -16,6 +16,7 @@
> > #define XSTATE_Hi16_ZMM 0x80
> >
> > #define XSTATE_FPSSE (XSTATE_FP | XSTATE_SSE)
> > +#define XSTATE_AVX512 (XSTATE_OPMASK | XSTATE_ZMM_Hi256 | XSTATE_Hi16_ZMM)
> > /* Bit 63 of XCR0 is reserved for future expansion */
> > #define XSTATE_EXTEND_MASK (~(XSTATE_FPSSE | (1ULL << 63)))
> >
> > diff --git a/arch/x86/kvm/cpuid.c b/arch/x86/kvm/cpuid.c
> > index 976e3a5..20d8321 100644
> > --- a/arch/x86/kvm/cpuid.c
> > +++ b/arch/x86/kvm/cpuid.c
> > @@ -317,7 +317,8 @@ static inline int __do_cpuid_ent(struct kvm_cpuid_entry2 *entry, u32 function,
> > const u32 kvm_supported_word9_x86_features =
> > F(FSGSBASE) | F(BMI1) | F(HLE) | F(AVX2) | F(SMEP) |
> > F(BMI2) | F(ERMS) | f_invpcid | F(RTM) | f_mpx | F(RDSEED) |
> > - F(ADX) | F(SMAP);
> > + F(ADX) | F(SMAP) | F(AVX512F) | F(AVX512PF) | F(AVX512ER) |
> > + F(AVX512CD);
> >
> > /* all calls to cpuid_count() should be made on the same cpu */
> > get_cpu();
> > diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c
> > index 5430e4b..3d77b88 100644
> > --- a/arch/x86/kvm/x86.c
> > +++ b/arch/x86/kvm/x86.c
> > @@ -651,6 +651,12 @@ int __kvm_set_xcr(struct kvm_vcpu *vcpu, u32 index, u64 xcr)
> > if ((!(xcr0 & XSTATE_BNDREGS)) != (!(xcr0 & XSTATE_BNDCSR)))
> > return 1;
> >
> > + if (xcr0 & XSTATE_AVX512) {
> > + if (!(xcr0 & XSTATE_YMM))
> > + return 1;
> > + if ((xcr0 & XSTATE_AVX512) != XSTATE_AVX512)
> > + return 1;
> > + }
> > kvm_put_guest_xcr0(vcpu);
> > vcpu->arch.xcr0 = xcr0;
> >
> > diff --git a/arch/x86/kvm/x86.h b/arch/x86/kvm/x86.h
> > index 7cb9c45..cc1d61a 100644
> > --- a/arch/x86/kvm/x86.h
> > +++ b/arch/x86/kvm/x86.h
> > @@ -162,7 +162,8 @@ int kvm_write_guest_virt_system(struct x86_emulate_ctxt *ctxt,
> > bool kvm_mtrr_valid(struct kvm_vcpu *vcpu, u32 msr, u64 data);
> >
> > #define KVM_SUPPORTED_XCR0 (XSTATE_FP | XSTATE_SSE | XSTATE_YMM \
> > - | XSTATE_BNDREGS | XSTATE_BNDCSR)
> > + | XSTATE_BNDREGS | XSTATE_BNDCSR \
> > + | XSTATE_AVX512)
> > extern u64 host_xcr0;
> >
> > extern u64 kvm_supported_xcr0(void);
> >


\
 
 \ /
  Last update: 2014-10-23 08:21    [W:0.045 / U:0.008 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site