Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
MIME-Version: 1.0
References: <20210430123822.13825-1-brijesh.singh@amd.com> <20210430123822.13825-33-brijesh.singh@amd.com>
 <CAMkAt6oYhRmqsKzDev3V5yMMePAR7ZzpEDRLadKhhCrb9Fq2=g@mail.gmail.com> <84f52e0e-b034-ebcf-e787-7ef9e3baae2f@amd.com>
In-Reply-To: <84f52e0e-b034-ebcf-e787-7ef9e3baae2f@amd.com>
From:   Peter Gonda <pgonda@google.com>
Date:   Mon, 10 May 2021 13:59:08 -0600
Message-ID: <CAMkAt6rfnjqPZHUqJM1r6GjnBMy-Gbw8VtGv4gd-G0Rjq+ALnQ@mail.gmail.com>
Subject: Re: [PATCH Part2 RFC v2 32/37] KVM: SVM: Add support to handle MSR
 based Page State Change VMGEXIT
To:     Brijesh Singh <brijesh.singh@amd.com>
Cc:     x86@kernel.org, linux-kernel@vger.kernel.org,
        kvm list <kvm@vger.kernel.org>,
        Thomas Gleixner <tglx@linutronix.de>,
        Borislav Petkov <bp@alien8.de>, jroedel@suse.de,
        "Lendacky, Thomas" <thomas.lendacky@amd.com>,
        Paolo Bonzini <pbonzini@redhat.com>,
        Ingo Molnar <mingo@redhat.com>,
        Dave Hansen <dave.hansen@intel.com>,
        David Rientjes <rientjes@google.com>,
        Sean Christopherson <seanjc@google.com>, peterz@infradead.org,
        "H. Peter Anvin" <hpa@zytor.com>, tony.luck@intel.com
Content-Type: text/plain; charset="UTF-8"
Precedence: bulk

On Mon, May 10, 2021 at 11:51 AM Brijesh Singh <brijesh.singh@amd.com> wrote:
>
> Hi Peter,
>
> On 5/10/21 12:30 PM, Peter Gonda wrote:
> >> +static int snp_make_page_shared(struct kvm_vcpu *vcpu, gpa_t gpa, kvm_pfn_t pfn, int level)
> >> +{
> >> +       struct rmpupdate val;
> >> +       int rc, rmp_level;
> >> +       struct rmpentry *e;
> >> +
> >> +       e = snp_lookup_page_in_rmptable(pfn_to_page(pfn), &rmp_level);
> >> +       if (!e)
> >> +               return -EINVAL;
> >> +
> >> +       if (!rmpentry_assigned(e))
> >> +               return 0;
> >> +
> >> +       /* Log if the entry is validated */
> >> +       if (rmpentry_validated(e))
> >> +               pr_debug_ratelimited("Remove RMP entry for a validated gpa 0x%llx\n", gpa);
> >> +
> >> +       /*
> >> +        * Is the page part of an existing 2M RMP entry ? Split the 2MB into multiple
> >> +        * of 4K-page before making the memory shared.
> >> +        */
> >> +       if ((level == PG_LEVEL_4K) && (rmp_level == PG_LEVEL_2M)) {
> >> +               rc = snp_rmptable_psmash(vcpu, pfn);
> >> +               if (rc)
> >> +                       return rc;
> >> +       }
> >> +
> >> +       memset(&val, 0, sizeof(val));
> >> +       val.pagesize = X86_TO_RMP_PG_LEVEL(level);
> > This is slightly different from Rev 2.00 of the GHCB spec. This
> > defaults to 2MB page sizes, when the spec says the only valid settings
> > for level are 0 -> 4k pages or 1 -> 2MB pages. Should this enforce the
> > same strictness as the spec?
>
>
> The caller of the snp_make_page_shared() must pass the x86 page level.
> We should reach here after all the guest provide value have passed
> through checks.
>
> The call sequence in this case should be:
>
> snp_handle_vmgexit_msr_protocol()
>
>  __snp_handle_page_state_change(vcpu, gfn_to_gpa(gfn), PG_LEVEL_4K)
>
>   snp_make_page_shared(..., level)
>
> Am I missing something  ?

Thanks Brijesh. I think my comment was misplaced. Looking at 33/37

+static unsigned long snp_handle_page_state_change(struct vcpu_svm
*svm, struct ghcb *ghcb)
+{
...
+ while (info->header.cur_entry <= info->header.end_entry) {
+ entry = &info->entry[info->header.cur_entry];
+ gpa = gfn_to_gpa(entry->gfn);
+ level = RMP_TO_X86_PG_LEVEL(entry->pagesize);
+ op = entry->operation;

This call to RMP_TO_X86_PG_LEVEL is not as strict as the spec. Is that OK?

>
> >> +       return rmpupdate(pfn_to_page(pfn), &val);
> >> +}
> >> +
> >> +static int snp_make_page_private(struct kvm_vcpu *vcpu, gpa_t gpa, kvm_pfn_t pfn, int level)
> >> +{
> >> +       struct kvm_sev_info *sev = &to_kvm_svm(vcpu->kvm)->sev_info;
> >> +       struct rmpupdate val;
> >> +       struct rmpentry *e;
> >> +       int rmp_level;
> >> +
> >> +       e = snp_lookup_page_in_rmptable(pfn_to_page(pfn), &rmp_level);
> >> +       if (!e)
> >> +               return -EINVAL;
> >> +
> >> +       /* Log if the entry is validated */
> >> +       if (rmpentry_validated(e))
> >> +               pr_err_ratelimited("Asked to make a pre-validated gpa %llx private\n", gpa);
> >> +
> >> +       memset(&val, 0, sizeof(val));
> >> +       val.gpa = gpa;
> >> +       val.asid = sev->asid;
> >> +       val.pagesize = X86_TO_RMP_PG_LEVEL(level);
> > Same comment as above.
>
> See my above response.
>
>
> >
> >> +       val.assigned = true;
> >> +
> >> +       return rmpupdate(pfn_to_page(pfn), &val);
> >> +}