qemu-devel.nongnu.org archive mirror
 help / color / mirror / Atom feed
From: Anthony Liguori <aliguori@linux.vnet.ibm.com>
To: "Zheng, Jiajia" <jiajia.zheng@intel.com>
Cc: "lmr@redhat.com" <lmr@redhat.com>,
	"Li, Haicheng" <haicheng.li@intel.com>,
	Max Asbock <masbock@linux.vnet.ibm.com>,
	"qemu-devel@nongnu.org" <qemu-devel@nongnu.org>,
	"You, Yongkang" <yongkang.you@intel.com>,
	"avi@redhat.com" <avi@redhat.com>,
	"Kleen, Andi" <andi.kleen@intel.com>
Subject: Re: [Qemu-devel] [Patch] Support translating Guest physical	address to Host virtual address.
Date: Wed, 03 Feb 2010 07:41:24 -0600	[thread overview]
Message-ID: <4B697D04.7070507@linux.vnet.ibm.com> (raw)
In-Reply-To: <DA586906BA1FFC4384FCFD6429ECE860352F27B5@shzsmsx502.ccr.corp.intel.com>

On 02/02/2010 10:04 PM, Zheng, Jiajia wrote:
> Hi,
> Any futher comments for this patch so that we can modify?
>    

Unfortunately, I see no way to modify this feature so that it would be 
acceptable to upstream.  I think you're going to have to carry this 
patch for your own debugging.

The problem is, the approach you're taking to testing is extremely 
racy.  Even just with the p2v command, the mapping within qemu that 
translates guest physical to host virtual is not stable over any given 
period of time.  By the time the monitor prints out the virtual mapping 
of a host page, it's already potentially wrong.

Likewise, the virtual->physical mapping within the host is also not 
stable so using that virtual address to lookup a physical address in the 
kernel is inherently racy.

If you want to integrate MCE testing into qemu using the host to inject 
MCEs, you'll need to figure out a way to do it that allows the whole 
thing to be done without potentially killing the wrong process or 
injecting it to within the wrong address within qemu.  I suspect that's 
going to be very difficult to achieve.

Regards,

Anthony Liguori

  reply	other threads:[~2010-02-03 13:41 UTC|newest]

Thread overview: 17+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2010-01-27  3:25 [Qemu-devel] [Patch] Support translating Guest physical address to Host virtual address Zheng, Jiajia
2010-01-27 21:39 ` Anthony Liguori
2010-01-27 22:31   ` Max Asbock
2010-02-03  4:04     ` Zheng, Jiajia
2010-02-03 13:41       ` Anthony Liguori [this message]
2010-02-03 14:11         ` Kleen, Andi
2010-02-03 14:23           ` Anthony Liguori
2010-02-03 15:29             ` Lucas Meneghel Rodrigues
2010-02-03 15:49             ` Kleen, Andi
2010-02-03 16:14               ` Anthony Liguori
2010-02-05  2:07                 ` Zheng, Jiajia
2010-02-07 14:03                 ` Avi Kivity
2010-02-07 16:23                   ` Anthony Liguori
2010-02-07 16:31                     ` Avi Kivity
2010-02-07 22:09                       ` Anthony Liguori
2010-02-08  3:38                         ` Zheng, Jiajia
2010-02-08  8:43                         ` Avi Kivity

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=4B697D04.7070507@linux.vnet.ibm.com \
    --to=aliguori@linux.vnet.ibm.com \
    --cc=andi.kleen@intel.com \
    --cc=avi@redhat.com \
    --cc=haicheng.li@intel.com \
    --cc=jiajia.zheng@intel.com \
    --cc=lmr@redhat.com \
    --cc=masbock@linux.vnet.ibm.com \
    --cc=qemu-devel@nongnu.org \
    --cc=yongkang.you@intel.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;
as well as URLs for NNTP newsgroup(s).