From: Sean Christopherson <sean.j.christopherson@intel.com>
To: Yang Weijiang <weijiang.yang@intel.com>
Cc: kvm@vger.kernel.org, linux-kernel@vger.kernel.org,
pbonzini@redhat.com, mst@redhat.com, rkrcmar@redhat.com,
jmattson@google.com
Subject: Re: [PATCH v6 7/8] KVM: x86: Load Guest fpu state when accessing MSRs managed by XSAVES
Date: Mon, 12 Aug 2019 16:04:50 -0700 [thread overview]
Message-ID: <20190812230450.GD4996@linux.intel.com> (raw)
In-Reply-To: <20190812230203.GC4996@linux.intel.com>
On Mon, Aug 12, 2019 at 04:02:03PM -0700, Sean Christopherson wrote:
> On Thu, Jul 25, 2019 at 11:12:45AM +0800, Yang Weijiang wrote:
> > From: Sean Christopherson <sean.j.christopherson@intel.com>
> >
> > A handful of CET MSRs are not context switched through "traditional"
> > methods, e.g. VMCS or manual switching, but rather are passed through
> > to the guest and are saved and restored by XSAVES/XRSTORS, i.e. the
> > guest's FPU state.
> >
> > Load the guest's FPU state if userspace is accessing MSRs whose values
> > are managed by XSAVES so that the MSR helper, e.g. vmx_{get,set}_msr(),
> > can simply do {RD,WR}MSR to access the guest's value.
> >
> > Note that guest_cpuid_has() is not queried as host userspace is allowed
> > to access MSRs that have not been exposed to the guest, e.g. it might do
> > KVM_SET_MSRS prior to KVM_SET_CPUID2.
> >
> > Signed-off-by: Sean Christopherson <sean.j.christopherson@intel.com>
> > Co-developed-by: Yang Weijiang <weijiang.yang@intel.com>
> > Signed-off-by: Yang Weijiang <weijiang.yang@intel.com>
> > ---
> > arch/x86/kvm/x86.c | 29 ++++++++++++++++++++++++++++-
> > 1 file changed, 28 insertions(+), 1 deletion(-)
> >
> > diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c
> > index fafd81d2c9ea..c657e6a56527 100644
> > --- a/arch/x86/kvm/x86.c
> > +++ b/arch/x86/kvm/x86.c
> > @@ -102,6 +102,8 @@ static void enter_smm(struct kvm_vcpu *vcpu);
> > static void __kvm_set_rflags(struct kvm_vcpu *vcpu, unsigned long rflags);
> > static void store_regs(struct kvm_vcpu *vcpu);
> > static int sync_regs(struct kvm_vcpu *vcpu);
> > +static void kvm_load_guest_fpu(struct kvm_vcpu *vcpu);
> > +static void kvm_put_guest_fpu(struct kvm_vcpu *vcpu);
> >
> > struct kvm_x86_ops *kvm_x86_ops __read_mostly;
> > EXPORT_SYMBOL_GPL(kvm_x86_ops);
> > @@ -2959,6 +2961,12 @@ int kvm_get_msr_common(struct kvm_vcpu *vcpu, struct msr_data *msr_info)
> > }
> > EXPORT_SYMBOL_GPL(kvm_get_msr_common);
> >
> > +static bool is_xsaves_msr(u32 index)
> > +{
> > + return index == MSR_IA32_U_CET ||
> > + (index >= MSR_IA32_PL0_SSP && index <= MSR_IA32_PL3_SSP);
> > +}
> > +
> > /*
> > * Read or write a bunch of msrs. All parameters are kernel addresses.
> > *
> > @@ -2969,11 +2977,30 @@ static int __msr_io(struct kvm_vcpu *vcpu, struct kvm_msrs *msrs,
> > int (*do_msr)(struct kvm_vcpu *vcpu,
> > unsigned index, u64 *data))
> > {
> > + bool fpu_loaded = false;
> > int i;
> > + u64 cet_bits = XFEATURE_MASK_CET_USER | XFEATURE_MASK_CET_KERNEL;
>
> Dunno if the compiler will actually generate different code, but this can be a
> const.
>
> > + u64 host_xss = 0;
> > +
> > + for (i = 0; i < msrs->nmsrs; ++i) {
> > + if (!fpu_loaded && is_xsaves_msr(entries[i].index)) {
> > + if (!kvm_x86_ops->xsaves_supported() ||
> > + !kvm_x86_ops->supported_xss())
>
> The "!kvm_x86_ops->supported_xss()" is redundant with the host_xss check
> below.
>
> > + continue;
>
> Hmm, vmx_set_msr() should be checking host_xss, arguably we should call
> do_msr() and let it handle the bad MSR access. I don't have a strong
> opinion either way, practically speaking the end result will be the same.
>
> If we do want to handle a misbehaving userspace here, this should be
> 'break' instead of 'continue'.
>
> > +
> > + host_xss = kvm_x86_ops->supported_xss();
> >
> > - for (i = 0; i < msrs->nmsrs; ++i)
> > + if ((host_xss & cet_bits) != cet_bits)
>
> I'm pretty sure this should check for either CET bit being set, not both,
> e.g. I assume it's possible to enable and expose XFEATURE_MASK_CET_USER
> but not XFEATURE_MASK_CET_KERNEL.
>
> So something like
>
> const u64 cet_bits = XFEATURE_MASK_CET_USER | XFEATURE_MASK_CET_KERNEL;
> const bool cet_supported = kvm_x86_ops->xsaves_supported() &&
> (kvm_x86_ops->supported_xss() & cet_bits);
Oh, and this should use kvm_supported_xss(), which masks ->supported_xss()
with KVM_SUPPORTED_XSS .
>
> for (i = 0; i < msrs->nmsrs; ++i) {
> if (!fpu_loaded && cet_supported &&
> is_xsaves_msr(entries[i].index)) {
> kvm_load_guest_fpu(vcpu);
> fpu_loaded = true;
> }
> if (do_msr(vcpu, entries[i].index, &entries[i].data))
> break;
> }
>
> or
>
> const u64 cet_bits = XFEATURE_MASK_CET_USER | XFEATURE_MASK_CET_KERNEL;
>
> for (i = 0; i < msrs->nmsrs; ++i) {
> if (!fpu_loaded && is_xsaves_msr(entries[i].index)) {
> if (!kvm_x86_ops->supported_xss() ||
> !(kvm_x86_ops->supported_xss() & cet_bits))
> break;
> kvm_load_guest_fpu(vcpu);
> fpu_loaded = true;
> }
> if (do_msr(vcpu, entries[i].index, &entries[i].data))
> break;
> }
>
>
> > + continue;
> > +
> > + kvm_load_guest_fpu(vcpu);
> > + fpu_loaded = true;
> > + }
> > if (do_msr(vcpu, entries[i].index, &entries[i].data))
> > break;
> > + }
> > + if (fpu_loaded)
> > + kvm_put_guest_fpu(vcpu);
> >
> > return i;
> > }
> > --
> > 2.17.2
> >
next prev parent reply other threads:[~2019-08-12 23:04 UTC|newest]
Thread overview: 24+ messages / expand[flat|nested] mbox.gz Atom feed top
2019-07-25 3:12 [PATCH v6 0/8] Introduce support for Guest CET feature Yang Weijiang
2019-07-25 3:12 ` [PATCH v6 1/8] KVM: VMX: Define CET VMCS fields and control bits Yang Weijiang
2019-07-25 3:12 ` [PATCH v6 2/8] KVM: x86: Add a helper function for CPUID(0xD,n>=1) enumeration Yang Weijiang
2019-08-12 22:18 ` Sean Christopherson
2019-08-13 6:11 ` Yang Weijiang
2019-07-25 3:12 ` [PATCH v6 3/8] KVM: x86: Implement CET CPUID enumeration for Guest Yang Weijiang
2019-08-13 0:06 ` Sean Christopherson
2019-08-13 5:27 ` Yang Weijiang
2019-07-25 3:12 ` [PATCH v6 4/8] KVM: VMX: Pass through CET related MSRs to Guest Yang Weijiang
2019-08-12 23:53 ` Sean Christopherson
2019-08-13 5:49 ` Yang Weijiang
2019-07-25 3:12 ` [PATCH v6 5/8] KVM: VMX: Load Guest CET via VMCS when CET is enabled in Guest Yang Weijiang
2019-08-12 23:56 ` Sean Christopherson
2019-08-13 5:38 ` Yang Weijiang
2019-07-25 3:12 ` [PATCH v6 6/8] KVM: x86: Add CET bits setting in CR4 and XSS Yang Weijiang
2019-07-25 3:12 ` [PATCH v6 7/8] KVM: x86: Load Guest fpu state when accessing MSRs managed by XSAVES Yang Weijiang
2019-08-12 23:02 ` Sean Christopherson
2019-08-12 23:04 ` Sean Christopherson [this message]
2019-08-12 23:29 ` Sean Christopherson
2019-08-13 6:06 ` Yang Weijiang
2019-08-13 6:05 ` Yang Weijiang
2019-07-25 3:12 ` [PATCH v6 8/8] KVM: x86: Add user-space access interface for CET MSRs Yang Weijiang
2019-08-12 23:43 ` Sean Christopherson
2019-08-13 5:41 ` Yang Weijiang
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20190812230450.GD4996@linux.intel.com \
--to=sean.j.christopherson@intel.com \
--cc=jmattson@google.com \
--cc=kvm@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=mst@redhat.com \
--cc=pbonzini@redhat.com \
--cc=rkrcmar@redhat.com \
--cc=weijiang.yang@intel.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).