Re: [PATCH 4/6] KVM MMU: optimize for writing cr4

From: Avi Kivity <avi@redhat.com>
To: Xiao Guangrong <xiaoguangrong@cn.fujitsu.com>
Cc: Marcelo Tosatti <mtosatti@redhat.com>,
	KVM list <kvm@vger.kernel.org>,
	LKML <linux-kernel@vger.kernel.org>
Subject: Re: [PATCH 4/6] KVM MMU: optimize for writing cr4
Date: Mon, 12 Apr 2010 14:22:27 +0300	[thread overview]
Message-ID: <4BC30273.3050406@redhat.com> (raw)
In-Reply-To: <4BC2F928.2060407@cn.fujitsu.com>

On 04/12/2010 01:42 PM, Xiao Guangrong wrote:
> Hi Avi,
>
> Thanks for your comments.
>
> Avi Kivity wrote:
>
>    
>> Later we have:
>>
>>      
>>>          kvm_x86_ops->set_cr4(vcpu, cr4);
>>>          vcpu->arch.cr4 = cr4;
>>>          vcpu->arch.mmu.base_role.cr4_pge = (cr4&  X86_CR4_PGE)&&
>>> !tdp_enabled;
>>>        
>> All of which depend on cr4.
>>      
> Oh, destroy_kvm_mmu() is not really destroyed cr3 and we can reload it later
> form shadow page cache, so, maybe this patch is unnecessary.
>
> But, i have a another question here, why we need encode 'cr4&  X86_CR4_PGE' into
> base_role.cr4_gpe? Why we need allocation different shadow page for global page
> and no-global page?
>    

See 6364a3918cb.  It was reverted later due to a problem with the 
implementation.  I'm not sure whether I want to fix the bug and restore 
that patch, or to drop it altogether and give the guest ownership of 
cr4.pge.  See cr4_guest_owned_bits (currently only used on ept).

> As i know, global page is not static in TLB, and x86 cpu also may flush them form TLB,
> maybe we no need treat global page specially... Am i miss something? :-(
>    

You can't read reverted patches? :)

-- 
I have a truly marvellous patch that fixes the bug which this
signature is too narrow to contain.