DMARC-Filter: OpenDMARC Filter v1.3.2 mx1.redhat.com BED4D883AD
Date: Wed, 2 Aug 2017 22:13:38 +0200
From: Radim =?utf-8?B?S3LEjW3DocWZ?= <rkrcmar@redhat.com>
To: Paolo Bonzini <pbonzini@redhat.com>
Cc: linux-kernel@vger.kernel.org, kvm@vger.kernel.org, rkrcmar@redhat.com,
        me@kylehuey.com
Subject: Re: [PATCH] KVM: svm: support single stepping over emulated
 instructions
Message-ID: <20170802201338.GA32318@flask>
References: <1501689810-20638-1-git-send-email-pbonzini@redhat.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <1501689810-20638-1-git-send-email-pbonzini@redhat.com>
Sender: linux-kernel-owner@vger.kernel.org
Content-Length: 873
Lines: 28

2017-08-02 18:03+0200, Paolo Bonzini:
> Use kvm_skip_emulated_instruction so that debug exceptions are
> properly injected.  This fixes debug.flat on AMD machines.
> 
> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
> ---

We already have b742c1e6e79d ("KVM: SVM: handle singlestep exception
when skipping emulated instructions"), so the only applicable part of
this patch is

> diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c
> index 0e846f0cb83b..931ba449456e 100644
> --- a/arch/x86/kvm/x86.c
> +++ b/arch/x86/kvm/x86.c
> @@ -5548,6 +5548,7 @@ static void kvm_vcpu_do_singlestep(struct kvm_vcpu *vcpu, int *r)
>  		 */
>  		vcpu->arch.dr6 &= ~15;
>  		vcpu->arch.dr6 |= DR6_BS | DR6_RTM;
> +		kvm_update_dr6(vcpu);
>  		kvm_queue_exception(vcpu, DB_VECTOR);
>  	}
>  }

which makes sense.  And shouldn't we do the same for
kvm_vcpu_check_breakpoint()?

Thanks.