| lists.openwall.net | lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC | |
|
Open Source and information security mailing list archives
| ||
|
Message-ID: <CABayD+f9o1CZTdak-ktKXpJnxcOAP4KPnYCDBzry91QcK6WVcw@mail.gmail.com>
Date: Tue, 6 Apr 2021 13:14:32 -0700
From: Steve Rutherford <srutherford@...gle.com>
To: Ashish Kalra <ashish.kalra@....com>
Cc: Sean Christopherson <seanjc@...gle.com>,
Paolo Bonzini <pbonzini@...hat.com>,
Thomas Gleixner <tglx@...utronix.de>,
Ingo Molnar <mingo@...hat.com>,
"H. Peter Anvin" <hpa@...or.com>, Joerg Roedel <joro@...tes.org>,
Borislav Petkov <bp@...e.de>,
Tom Lendacky <thomas.lendacky@....com>,
X86 ML <x86@...nel.org>, KVM list <kvm@...r.kernel.org>,
LKML <linux-kernel@...r.kernel.org>,
Venu Busireddy <venu.busireddy@...cle.com>,
Brijesh Singh <brijesh.singh@....com>,
Will Deacon <will@...nel.org>, maz@...nel.org,
Quentin Perret <qperret@...gle.com>
Subject: Re: [PATCH v11 08/13] KVM: X86: Introduce KVM_HC_PAGE_ENC_STATUS hypercall
On Tue, Apr 6, 2021 at 9:08 AM Ashish Kalra <ashish.kalra@....com> wrote:
>
> On Tue, Apr 06, 2021 at 03:48:20PM +0000, Sean Christopherson wrote:
> > On Mon, Apr 05, 2021, Ashish Kalra wrote:
> > > From: Ashish Kalra <ashish.kalra@....com>
> >
> > ...
> >
> > > diff --git a/arch/x86/include/asm/kvm_host.h b/arch/x86/include/asm/kvm_host.h
> > > index 3768819693e5..78284ebbbee7 100644
> > > --- a/arch/x86/include/asm/kvm_host.h
> > > +++ b/arch/x86/include/asm/kvm_host.h
> > > @@ -1352,6 +1352,8 @@ struct kvm_x86_ops {
> > > int (*complete_emulated_msr)(struct kvm_vcpu *vcpu, int err);
> > >
> > > void (*vcpu_deliver_sipi_vector)(struct kvm_vcpu *vcpu, u8 vector);
> > > + int (*page_enc_status_hc)(struct kvm_vcpu *vcpu, unsigned long gpa,
> > > + unsigned long sz, unsigned long mode);
> > > };
> > >
> > > struct kvm_x86_nested_ops {
> > > diff --git a/arch/x86/kvm/svm/sev.c b/arch/x86/kvm/svm/sev.c
> > > index c9795a22e502..fb3a315e5827 100644
> > > --- a/arch/x86/kvm/svm/sev.c
> > > +++ b/arch/x86/kvm/svm/sev.c
> > > @@ -1544,6 +1544,67 @@ static int sev_receive_finish(struct kvm *kvm, struct kvm_sev_cmd *argp)
> > > return ret;
> > > }
> > >
> > > +static int sev_complete_userspace_page_enc_status_hc(struct kvm_vcpu *vcpu)
> > > +{
> > > + vcpu->run->exit_reason = 0;
> > > + kvm_rax_write(vcpu, vcpu->run->dma_sharing.ret);
> > > + ++vcpu->stat.hypercalls;
> > > + return kvm_skip_emulated_instruction(vcpu);
> > > +}
> > > +
> > > +int svm_page_enc_status_hc(struct kvm_vcpu *vcpu, unsigned long gpa,
> > > + unsigned long npages, unsigned long enc)
> > > +{
> > > + kvm_pfn_t pfn_start, pfn_end;
> > > + struct kvm *kvm = vcpu->kvm;
> > > + gfn_t gfn_start, gfn_end;
> > > +
> > > + if (!sev_guest(kvm))
> > > + return -EINVAL;
> > > +
> > > + if (!npages)
> > > + return 0;
> >
> > Parth of me thinks passing a zero size should be an error not a nop. Either way
> > works, just feels a bit weird to allow this to be a nop.
> >
> > > +
> > > + gfn_start = gpa_to_gfn(gpa);
> >
> > This should check that @gpa is aligned.
> >
> > > + gfn_end = gfn_start + npages;
> > > +
> > > + /* out of bound access error check */
> > > + if (gfn_end <= gfn_start)
> > > + return -EINVAL;
> > > +
> > > + /* lets make sure that gpa exist in our memslot */
> > > + pfn_start = gfn_to_pfn(kvm, gfn_start);
> > > + pfn_end = gfn_to_pfn(kvm, gfn_end);
> > > +
> > > + if (is_error_noslot_pfn(pfn_start) && !is_noslot_pfn(pfn_start)) {
> > > + /*
> > > + * Allow guest MMIO range(s) to be added
> > > + * to the shared pages list.
> > > + */
> > > + return -EINVAL;
> > > + }
> > > +
> > > + if (is_error_noslot_pfn(pfn_end) && !is_noslot_pfn(pfn_end)) {
> > > + /*
> > > + * Allow guest MMIO range(s) to be added
> > > + * to the shared pages list.
> > > + */
> > > + return -EINVAL;
> > > + }
> >
> > I don't think KVM should do any checks beyond gfn_end <= gfn_start. Just punt
> > to userspace and give userspace full say over what is/isn't legal.
> >
> > > +
> > > + if (enc)
> > > + vcpu->run->exit_reason = KVM_EXIT_DMA_UNSHARE;
> > > + else
> > > + vcpu->run->exit_reason = KVM_EXIT_DMA_SHARE;
> >
> > Use a single exit and pass "enc" via kvm_run. I also strongly dislike "DMA",
> > there's no guarantee the guest is sharing memory for DMA.
> >
> > I think we can usurp KVM_EXIT_HYPERCALL for this? E.g.
> >
>
> I see the following in Documentation/virt/kvm/api.rst :
> ..
> ..
> /* KVM_EXIT_HYPERCALL */
> struct {
> __u64 nr;
> __u64 args[6];
> __u64 ret;
> __u32 longmode;
> __u32 pad;
> } hypercall;
>
> Unused. This was once used for 'hypercall to userspace'. To implement
> such functionality, use KVM_EXIT_IO (x86) or KVM_EXIT_MMIO (all except s390).
>
> This mentions this exitcode to be unused and implementing this
> functionality using KVM_EXIT_IO for x86?
I suspect this description is historical. It was originally from 2009.
KVM_EXIT_IO is used for IO port reads/writes.
Personally, I would prefer to stay the course and use a name similar
to KVM_EXIT_DMA_SHARE, say KVM_EXIT_MEM_SHARE and
KVM_EXIT_MEM_UNSHARE. These just seem very clear, which I appreciate.
Reusing hypercall would work, but shoehorning this into
KVM_EXIT_HYPERCALL when we don't have generic hypercall exits feels a
bit off in my mind. Note: that preference isn't particularly strong.
Steve
>
> Thanks,
> Ashish
>
> > vcpu->run->exit_reason = KVM_EXIT_HYPERCALL;
> > vcpu->run->hypercall.nr = KVM_HC_PAGE_ENC_STATUS;
> > vcpu->run->hypercall.args[0] = gfn_start << PAGE_SHIFT;
> > vcpu->run->hypercall.args[1] = npages * PAGE_SIZE;
> > vcpu->run->hypercall.args[2] = enc;
> > vcpu->run->hypercall.longmode = is_64_bit_mode(vcpu);
> >
> > > +
> > > + vcpu->run->dma_sharing.addr = gfn_start;
> >
> > Addresses and pfns are not the same thing. If you're passing the size in bytes,
> > then it's probably best to pass the gpa, not the gfn. Same for the params from
> > the guest, they should be in the same "domain".
> >
> > > + vcpu->run->dma_sharing.len = npages * PAGE_SIZE;
> > > + vcpu->arch.complete_userspace_io =
> > > + sev_complete_userspace_page_enc_status_hc;
> >
> > I vote to drop the "userspace" part, it's already quite verbose.
> >
> > vcpu->arch.complete_userspace_io = sev_complete_page_enc_status_hc;
> >
> > > +
> > > + return 0;
> > > +}
> > > +
> >
> > ..
> >
> > > diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c
> > > index f7d12fca397b..ef5c77d59651 100644
> > > --- a/arch/x86/kvm/x86.c
> > > +++ b/arch/x86/kvm/x86.c
> > > @@ -8273,6 +8273,18 @@ int kvm_emulate_hypercall(struct kvm_vcpu *vcpu)
> > > kvm_sched_yield(vcpu->kvm, a0);
> > > ret = 0;
> > > break;
> > > + case KVM_HC_PAGE_ENC_STATUS: {
> > > + int r;
> > > +
> > > + ret = -KVM_ENOSYS;
> > > + if (kvm_x86_ops.page_enc_status_hc) {
> > > + r = kvm_x86_ops.page_enc_status_hc(vcpu, a0, a1, a2);
> >
> > Use static_call().
> >
> > > + if (r >= 0)
> > > + return r;
> > > + ret = r;
> > > + }
> >
> > Hmm, an alternative to adding a kvm_x86_ops hook would be to tag the VM as
> > supporting/allowing the hypercall. That would clean up this code, ensure VMX
> > and SVM don't end up creating a different userspace ABI, and make it easier to
> > reuse the hypercall in the future (I'm still hopeful :-) ). E.g.
> >
> > case KVM_HC_PAGE_ENC_STATUS: {
> > u64 gpa = a0, nr_bytes = a1;
> >
> > if (!vcpu->kvm->arch.page_enc_hc_enable)
> > break;
> >
> > if (!PAGE_ALIGNED(gpa) || !PAGE_ALIGNED(nr_bytes) ||
> > !nr_bytes || gpa + nr_bytes <= gpa)) {
> > ret = -EINVAL;
> > break;
> > }
> >
> > vcpu->run->exit_reason = KVM_EXIT_HYPERCALL;
> > vcpu->run->hypercall.nr = KVM_HC_PAGE_ENC_STATUS;
> > vcpu->run->hypercall.args[0] = gpa;
> > vcpu->run->hypercall.args[1] = nr_bytes;
> > vcpu->run->hypercall.args[2] = enc;
> > vcpu->run->hypercall.longmode = op_64_bit;
> > vcpu->arch.complete_userspace_io = complete_page_enc_hc;
> > return 0;
> > }
> >
> >
> > > + break;
> > > + }
> > > default:
> > > ret = -KVM_ENOSYS;
> > > break;
Powered by blists - more mailing lists