linux-kernel - Re: [PATCH 02/10] KVM: SVM: Use AVIC_HPA_MASK when initializing vCPU's Physical ID entry

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

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Message-ID: <74477cff348512fd88da8bb8fa901207fa29e316.camel@redhat.com>
Date:   Thu, 05 Oct 2023 15:49:57 +0300
From:   Maxim Levitsky <mlevitsk@...hat.com>
To:     Sean Christopherson <seanjc@...gle.com>,
        Paolo Bonzini <pbonzini@...hat.com>,
        Joerg Roedel <joro@...tes.org>
Cc:     kvm@...r.kernel.org, iommu@...ts.linux.dev,
        linux-kernel@...r.kernel.org
Subject: Re: [PATCH 02/10] KVM: SVM: Use AVIC_HPA_MASK when initializing
 vCPU's Physical ID entry

У вт, 2023-08-15 у 14:35 -0700, Sean Christopherson пише:
> Use AVIC_HPA_MASK instead of AVIC_PHYSICAL_ID_ENTRY_BACKING_PAGE_MASK when
> initializing a vCPU's Physical ID table entry, the two masks are identical.

Masks are identical, but they refer to different things. 
AVIC_HPA_MASK is about avic's fields in vmcb, while AVIC_PHYSICAL_ID_ENTRY_BACKING_PAGE_MASK
is about the portion of the physical id entry which has the address.

They are the same currently but for documentation purposes I don't understand
why removal of one of them makes sense.

IMHO, it's best to define them to the same value, e.g:

#define AVIC_HPA_MASK	GENMASK_ULL(51, 12)

#define AVIC_PHYSICAL_ID_ENTRY_BACKING_PAGE_MASK GENMASK_ULL(51, 12)

Best regards,
	Maxim Levitsky


> Keep both #defines for now, along with a few new static asserts.  A future
> change will clean up the entire mess (spoiler alert, the masks are
> pointless).
> 
> Opportunisitically move the bitwise-OR of AVIC_PHYSICAL_ID_ENTRY_VALID_MASK
> outside of the call to __sme_set(), again to pave the way for code
> deduplication.  __sme_set() is purely additive, i.e. ORing in the valid
> bit before or after the C-bit does not change the end result.
> 
> No functional change intended.
> 
> Signed-off-by: Sean Christopherson <seanjc@...gle.com>
> ---
>  arch/x86/include/asm/svm.h | 2 ++
>  arch/x86/kvm/svm/avic.c    | 5 ++---
>  2 files changed, 4 insertions(+), 3 deletions(-)
> 
> diff --git a/arch/x86/include/asm/svm.h b/arch/x86/include/asm/svm.h
> index 1e70600e84f7..609c9b596399 100644
> --- a/arch/x86/include/asm/svm.h
> +++ b/arch/x86/include/asm/svm.h
> @@ -285,6 +285,8 @@ static_assert((AVIC_MAX_PHYSICAL_ID & AVIC_PHYSICAL_MAX_INDEX_MASK) == AVIC_MAX_
>  static_assert((X2AVIC_MAX_PHYSICAL_ID & AVIC_PHYSICAL_MAX_INDEX_MASK) == X2AVIC_MAX_PHYSICAL_ID);
>  
>  #define AVIC_HPA_MASK	~((0xFFFULL << 52) | 0xFFF)
> +static_assert(AVIC_PHYSICAL_ID_ENTRY_BACKING_PAGE_MASK == AVIC_HPA_MASK);
> +static_assert(AVIC_PHYSICAL_ID_ENTRY_BACKING_PAGE_MASK == GENMASK_ULL(51, 12));
>  
>  #define SVM_SEV_FEAT_DEBUG_SWAP                        BIT(5)
>  
> diff --git a/arch/x86/kvm/svm/avic.c b/arch/x86/kvm/svm/avic.c
> index 7062164e4041..442c58ef8158 100644
> --- a/arch/x86/kvm/svm/avic.c
> +++ b/arch/x86/kvm/svm/avic.c
> @@ -308,9 +308,8 @@ static int avic_init_backing_page(struct kvm_vcpu *vcpu)
>  	if (!entry)
>  		return -EINVAL;
>  
> -	new_entry = __sme_set((page_to_phys(svm->avic_backing_page) &
> -			      AVIC_PHYSICAL_ID_ENTRY_BACKING_PAGE_MASK) |
> -			      AVIC_PHYSICAL_ID_ENTRY_VALID_MASK);
> +	new_entry = __sme_set(page_to_phys(svm->avic_backing_page) & AVIC_HPA_MASK) |
> +		    AVIC_PHYSICAL_ID_ENTRY_VALID_MASK;
>  	WRITE_ONCE(*entry, new_entry);
>  
>  	svm->avic_physical_id_cache = entry;