[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <6530eb94-b937-415c-8457-f5c598d94e7b@redhat.com>
Date: Wed, 14 Aug 2024 19:50:01 +0200
From: Paolo Bonzini <pbonzini@...hat.com>
To: Sean Christopherson <seanjc@...gle.com>
Cc: kvm@...r.kernel.org, linux-kernel@...r.kernel.org,
Peter Gonda <pgonda@...gle.com>, Michael Roth <michael.roth@....com>,
Vishal Annapurve <vannapurve@...gle.com>,
Ackerly Tng <ackerleytng@...gle.com>
Subject: Re: [PATCH 16/22] KVM: x86: Remove manual pfn lookup when retrying
#PF after failed emulation
On 8/9/24 21:03, Sean Christopherson wrote:
> Drop the manual pfn look when retrying an instruction that KVM failed to
> emulation in response to a #PF due to a write-protected gfn. Now that KVM
> sets EMULTYPE_PF if and only if the page fault it a write-protected gfn,
Pointing out where this happened will likely help a few years from now:
With the introduction of RET_PF_WRITE_PROTECTED, KVM sets EMULTYPE_PF if
and only if the page fault it a write-protected gfn, i.e. if and only if
there's a writable memslot. KVM will never try to redo an instruction
that failed on emulated MMIO (no slot, or a write to a read-only slot),
so therefore there's no redo the lookup in reexecute_instruction().
Paolo
> i.e. if and only if there's a writable memslot, there's no need to redo
> the lookup to avoid retrying an instruction that failed on emulated MMIO
> (no slot, or a write to a read-only slot).
>
> Signed-off-by: Sean Christopherson <seanjc@...gle.com>
> ---
> arch/x86/kvm/x86.c | 18 ------------------
> 1 file changed, 18 deletions(-)
>
> diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c
> index 771e67381fce..67f9871990fb 100644
> --- a/arch/x86/kvm/x86.c
> +++ b/arch/x86/kvm/x86.c
> @@ -8867,7 +8867,6 @@ static bool reexecute_instruction(struct kvm_vcpu *vcpu, gpa_t cr2_or_gpa,
> int emulation_type)
> {
> gpa_t gpa = cr2_or_gpa;
> - kvm_pfn_t pfn;
>
> if (!(emulation_type & EMULTYPE_ALLOW_RETRY_PF))
> return false;
> @@ -8887,23 +8886,6 @@ static bool reexecute_instruction(struct kvm_vcpu *vcpu, gpa_t cr2_or_gpa,
> return true;
> }
>
> - /*
> - * Do not retry the unhandleable instruction if it faults on the
> - * readonly host memory, otherwise it will goto a infinite loop:
> - * retry instruction -> write #PF -> emulation fail -> retry
> - * instruction -> ...
> - */
> - pfn = gfn_to_pfn(vcpu->kvm, gpa_to_gfn(gpa));
> -
> - /*
> - * If the instruction failed on the error pfn, it can not be fixed,
> - * report the error to userspace.
> - */
> - if (is_error_noslot_pfn(pfn))
> - return false;
> -
> - kvm_release_pfn_clean(pfn);
> -
> /*
> * If emulation may have been triggered by a write to a shadowed page
> * table, unprotect the gfn (zap any relevant SPTEs) and re-enter the
Powered by blists - more mailing lists