linux-kernel - Re: [PATCH 3/7] KVM: X86: Zap the invalid list after remote tlb flushing

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

Message-ID: <YTFIFlU9wtVHf4xd@google.com>
Date:   Thu, 2 Sep 2021 21:54:30 +0000
From:   Sean Christopherson <seanjc@...gle.com>
To:     Lai Jiangshan <jiangshanlai@...il.com>
Cc:     linux-kernel@...r.kernel.org, Paolo Bonzini <pbonzini@...hat.com>,
        Lai Jiangshan <laijs@...ux.alibaba.com>,
        Vitaly Kuznetsov <vkuznets@...hat.com>,
        Wanpeng Li <wanpengli@...cent.com>,
        Jim Mattson <jmattson@...gle.com>,
        Joerg Roedel <joro@...tes.org>,
        Thomas Gleixner <tglx@...utronix.de>,
        Ingo Molnar <mingo@...hat.com>, Borislav Petkov <bp@...en8.de>,
        x86@...nel.org, "H. Peter Anvin" <hpa@...or.com>,
        kvm@...r.kernel.org
Subject: Re: [PATCH 3/7] KVM: X86: Zap the invalid list after remote tlb
 flushing

On Tue, Aug 24, 2021, Lai Jiangshan wrote:
> From: Lai Jiangshan <laijs@...ux.alibaba.com>
> 
> In mmu_sync_children(), it can zap the invalid list after remote tlb flushing.
> Emptifying the invalid list ASAP might help reduce a remote tlb flushing
> in some cases.
> 
> Signed-off-by: Lai Jiangshan <laijs@...ux.alibaba.com>
> ---
>  arch/x86/kvm/mmu/mmu.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/arch/x86/kvm/mmu/mmu.c b/arch/x86/kvm/mmu/mmu.c
> index 987953a901d2..a165eb8713bc 100644
> --- a/arch/x86/kvm/mmu/mmu.c
> +++ b/arch/x86/kvm/mmu/mmu.c
> @@ -2050,7 +2050,7 @@ static bool mmu_sync_children(struct kvm_vcpu *vcpu,
>  			protected |= rmap_write_protect(vcpu, sp->gfn);
>  
>  		if (protected) {
> -			kvm_flush_remote_tlbs(vcpu->kvm);
> +			kvm_mmu_flush_or_zap(vcpu, &invalid_list, true, flush);

This can just be 

			kvm_mmu_remote_flush_or_zap(vcpu, &invalid_list, true);

since a remote flush always does a local flush too.

Related to the tlbs_dirty revert, to avoid overzealous flushing, kvm_sync_page()
can pass back a "remote_flush" flag instead of doing the flush itself.  Something
like the below, or maybe multiplex an 'int' return.

diff --git a/arch/x86/kvm/mmu/mmu.c b/arch/x86/kvm/mmu/mmu.c
index ac260e01e9d8..f61de53de55a 100644
--- a/arch/x86/kvm/mmu/mmu.c
+++ b/arch/x86/kvm/mmu/mmu.c
@@ -2041,7 +2041,7 @@ static bool mmu_sync_children(struct kvm_vcpu *vcpu,
        struct mmu_page_path parents;
        struct kvm_mmu_pages pages;
        LIST_HEAD(invalid_list);
-       bool flush = false;
+       bool flush = false, remote_flush = false;

        while (mmu_unsync_walk(parent, &pages)) {
                bool protected = false;
@@ -2050,17 +2050,17 @@ static bool mmu_sync_children(struct kvm_vcpu *vcpu,
                        protected |= rmap_write_protect(vcpu, sp->gfn);

                if (protected) {
-                       kvm_mmu_flush_or_zap(vcpu, &invalid_list, true, flush);
-                       flush = false;
+                       kvm_mmu_remote_flush_or_zap(vcpu, &invalid_list, true);
+                       remote_flush = flush = false;
                }

                for_each_sp(pages, sp, parents, i) {
                        kvm_unlink_unsync_page(vcpu->kvm, sp);
-                       flush |= kvm_sync_page(vcpu, sp, &invalid_list);
+                       flush |= kvm_sync_page(vcpu, sp, &invalid_list, &remote_flush);
                        mmu_pages_clear_parents(&parents);
                }
                if (need_resched() || rwlock_needbreak(&vcpu->kvm->mmu_lock)) {
-                       kvm_mmu_flush_or_zap(vcpu, &invalid_list, false, flush);
+                       kvm_mmu_flush_or_zap(vcpu, &invalid_list, remote_flush, flush);
                        cond_resched_rwlock_write(&vcpu->kvm->mmu_lock);
                        /*
                         * If @parent is not root, the caller doesn't have
@@ -2074,7 +2074,7 @@ static bool mmu_sync_children(struct kvm_vcpu *vcpu,
                }
        }

-       kvm_mmu_flush_or_zap(vcpu, &invalid_list, false, flush);
+       kvm_mmu_flush_or_zap(vcpu, &invalid_list, remote_flush, flush);
        return true;
 }

>  			flush = false;
>  		}
>  
> -- 
> 2.19.1.6.gb485710b
>