From: Xiao Guangrong <xiaoguangrong@cn.fujitsu.com>
To: Avi Kivity <avi@redhat.com>
Cc: Marcelo Tosatti <mtosatti@redhat.com>,
LKML <linux-kernel@vger.kernel.org>, KVM <kvm@vger.kernel.org>
Subject: Re: [PATCH 04/11] KVM: MMU: do not mark access bit on pte write path
Date: Wed, 27 Jul 2011 18:04:05 +0800 [thread overview]
Message-ID: <4E2FE295.5070905@cn.fujitsu.com> (raw)
In-Reply-To: <4E2FD578.2090803@redhat.com>
On 07/27/2011 05:08 PM, Avi Kivity wrote:
> On 07/26/2011 02:28 PM, Xiao Guangrong wrote:
>> In current code, the accessed bit is always set when page fault occurred,
>> do not need to set it on pte write path
>>
>
> Is this true? a write with pte.w=pte.a=0 sets pte.a?
>
Generally, we will call set_spte() with speculative = false, then Accessed bit
is set on page fault path, but there has two case:
- if pte.d = 1, everything is ok
- if pte.d = 0, so this write can change pte.d = 1, the access permission of
shadow page can be changed from read-only to writable. So, we will find a
new sp to establish the mapping. In this case, the original spte is not
set Accessed bit.
However, we used Accessed bit to detect write flooding, so the first write can
set the dirty bit, and the later write can not cause sp changed. I think it is
not too bad.
next prev parent reply other threads:[~2011-07-27 10:02 UTC|newest]
Thread overview: 34+ messages / expand[flat|nested] mbox.gz Atom feed top
2011-07-26 11:24 [PATCH 0/11] KVM: x86: optimize for guest page written Xiao Guangrong
2011-07-26 11:25 ` [PATCH 01/11] KVM: MMU: avoid pte_list_desc run out in kvm_mmu_pte_write Xiao Guangrong
2011-07-27 9:00 ` Avi Kivity
2011-07-27 9:37 ` Xiao Guangrong
2011-07-26 11:25 ` [PATCH 02/11] KVM: x86: cleanup pio/pout emulated Xiao Guangrong
2011-07-26 11:26 ` [PATCH 03/11] KVM: x86: fast emulate repeat string write instructions Xiao Guangrong
2011-07-26 12:27 ` Gleb Natapov
2011-07-26 13:53 ` Avi Kivity
2011-07-27 1:47 ` Xiao Guangrong
2011-07-27 4:26 ` Gleb Natapov
2011-07-27 6:32 ` Xiao Guangrong
2011-07-27 7:51 ` Gleb Natapov
2011-07-27 9:36 ` Xiao Guangrong
2011-07-27 9:04 ` Avi Kivity
2011-07-27 9:37 ` Xiao Guangrong
2011-07-26 11:28 ` [PATCH 04/11] KVM: MMU: do not mark access bit on pte write path Xiao Guangrong
2011-07-27 9:08 ` Avi Kivity
2011-07-27 10:04 ` Xiao Guangrong [this message]
2011-07-26 11:28 ` [PATCH 05/11] KVM: MMU: cleanup FNAME(invlpg) Xiao Guangrong
2011-07-26 11:29 ` [PATCH 06/11] KVM: MMU: fast prefetch spte on invlpg path Xiao Guangrong
2011-07-26 11:29 ` [PATCH 07/11] KVM: MMU: remove unnecessary kvm_mmu_free_some_pages Xiao Guangrong
2011-07-26 11:30 ` [PATCH 08/11] KVM: MMU: split kvm_mmu_pte_write function Xiao Guangrong
2011-07-26 11:31 ` [PATCH 09/11] KVM: MMU: remove the mismatch shadow page Xiao Guangrong
2011-07-27 9:11 ` Avi Kivity
2011-07-27 9:13 ` Avi Kivity
2011-07-27 10:05 ` Xiao Guangrong
2011-07-26 11:31 ` [PATCH 10/11] KVM: MMU: fix detecting misaligned accessed Xiao Guangrong
2011-07-27 9:15 ` Avi Kivity
2011-07-27 10:10 ` Xiao Guangrong
2011-07-26 11:32 ` [PATCH 11/11] KVM: MMU: improve write flooding detected Xiao Guangrong
2011-07-27 9:23 ` Avi Kivity
2011-07-27 10:20 ` Xiao Guangrong
2011-07-27 11:08 ` Avi Kivity
2011-07-28 2:43 ` Xiao Guangrong
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=4E2FE295.5070905@cn.fujitsu.com \
--to=xiaoguangrong@cn.fujitsu.com \
--cc=avi@redhat.com \
--cc=kvm@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=mtosatti@redhat.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox