Date:   Fri, 17 Mar 2023 15:59:55 -0700
In-Reply-To: <20230211014626.3659152-3-vipinsh@google.com>
Mime-Version: 1.0
References: <20230211014626.3659152-1-vipinsh@google.com> <20230211014626.3659152-3-vipinsh@google.com>
Message-ID: <ZBTw6ymNg0gYxhtW@google.com>
Subject: Re: [Patch v3 2/7] KVM: x86/mmu: Atomically clear SPTE dirty state in
 the clear-dirty-log flow
From:   Sean Christopherson <seanjc@google.com>
To:     Vipin Sharma <vipinsh@google.com>
Cc:     pbonzini@redhat.com, bgardon@google.com, dmatlack@google.com,
        kvm@vger.kernel.org, linux-kernel@vger.kernel.org
Content-Type: text/plain; charset="us-ascii"
Precedence: bulk

On Fri, Feb 10, 2023, Vipin Sharma wrote:
> @@ -1677,8 +1670,13 @@ bool kvm_tdp_mmu_clear_dirty_slot(struct kvm *kvm,
>  static void clear_dirty_pt_masked(struct kvm *kvm, struct kvm_mmu_page *root,
>  				  gfn_t gfn, unsigned long mask, bool wrprot)
>  {
> +	/*
> +	 * Either all SPTEs in TDP MMU will need write protection or none. This
> +	 * contract will not be modified for TDP MMU pages.
> +	 */
> +	u64 clear_bit = (wrprot || !kvm_ad_enabled()) ? PT_WRITABLE_MASK :
> +							shadow_dirty_mask;

Switching from spte_ad_need_write_protect() to kvm_ad_enabled() belongs in a
separate.  In the unlikely event that the above assertion/contracts is invalid,
then any issues should bisect to the switch, not to a much more complex patch.

I'll make that happen when applying.