From mboxrd@z Thu Jan  1 00:00:00 1970
From: Gleb Natapov <gleb@redhat.com>
Subject: Re: [PATCH 08/23] KVM: x86 emulator: cleanup some direct calls
 into kvm to use existing callbacks
Date: Wed, 28 Apr 2010 12:33:10 +0300
Message-ID: <20100428093310.GN10044@redhat.com>
References: <1272370524-1295-1-git-send-email-gleb@redhat.com>
 <1272370524-1295-9-git-send-email-gleb@redhat.com>
 <4BD7F90A.8060107@redhat.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Cc: mtosatti@redhat.com, kvm@vger.kernel.org
To: Avi Kivity <avi@redhat.com>
Return-path: <kvm-owner@vger.kernel.org>
Received: from mx1.redhat.com ([209.132.183.28]:7026 "EHLO mx1.redhat.com"
	rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP
	id S1751006Ab0D1KSk (ORCPT <rfc822;kvm@vger.kernel.org>);
	Wed, 28 Apr 2010 06:18:40 -0400
Received: from int-mx04.intmail.prod.int.phx2.redhat.com (int-mx04.intmail.prod.int.phx2.redhat.com [10.5.11.17])
	by mx1.redhat.com (8.13.8/8.13.8) with ESMTP id o3S9Y1Tw008257
	(version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=OK)
	for <kvm@vger.kernel.org>; Wed, 28 Apr 2010 05:34:01 -0400
Content-Disposition: inline
In-Reply-To: <4BD7F90A.8060107@redhat.com>
Sender: kvm-owner@vger.kernel.org
List-ID: <kvm.vger.kernel.org>

On Wed, Apr 28, 2010 at 11:59:54AM +0300, Avi Kivity wrote:
> On 04/27/2010 03:15 PM, Gleb Natapov wrote:
> >Use callbacks from x86_emulate_ops to access segments instead of calling
> >into kvm directly.
> >
> >
> >
> >-static unsigned long seg_base(struct x86_emulate_ctxt *ctxt, int seg)
> >+static unsigned long seg_base(struct x86_emulate_ctxt *ctxt,
> >+			      struct x86_emulate_ops *ops, int seg)
> >  {
> >-	if (ctxt->mode == X86EMUL_MODE_PROT64&&  seg<  VCPU_SREG_FS)
> >-		return 0;
> >+	unsigned long base;
> >
> >-	return kvm_x86_ops->get_segment_base(ctxt->vcpu, seg);
> 
> get_segment_base() is only one vmread on intel, but you replace it
> with reading the entire segment.
> 
Didn't what to have separate x86_emulate_ops callback for reading
segment base. If this is serious performance concern it can be added

> >+	if (ctxt->mode == X86EMUL_MODE_PROT64) {
> >+		u64 val;
> >+		switch (seg) {
> >+		case VCPU_SREG_FS:
> >+			ops->get_msr(ctxt->vcpu, MSR_FS_BASE,&val);
> >+			break;
> >+		case VCPU_SREG_GS:
> >+			ops->get_msr(ctxt->vcpu, MSR_GS_BASE,&val);
> >+			break;
> >+		default:
> >+			val = 0;
> >+			break;
> >+		}
> 
> Why this ugliness?  get_cached_descriptor() should do this.
> 
get_cached_descriptor() returns struct desc_struct which is 32 bit only.
There is not 64bit segment descriptors.

> >
> >  static unsigned long seg_override_base(struct x86_emulate_ctxt *ctxt,
> >+				       struct x86_emulate_ops *ops,
> >  				       struct decode_cache *c)
> >  {
> >  	if (!c->has_seg_override)
> >  		return 0;
> >
> >-	return seg_base(ctxt, c->seg_override);
> >+	return seg_base(ctxt, ops, c->seg_override);
> >  }
> 
> Sticking ops into ctxt would reduce the size of these patches.
> 
But it will introduce intermediate huge patch. But I agree that this
should be done eventually. Planned to do it somewhere at the end. Near
the patch that change x86_emulate_ops callbacks to get ctxt instead of
vcpu as a parameter.

--
			Gleb.