qemu-devel.nongnu.org archive mirror
 help / color / mirror / Atom feed
From: Jason Wang <jasowang@redhat.com>
To: Paolo Bonzini <pbonzini@redhat.com>,
	"Tian, Kevin" <kevin.tian@intel.com>,
	"Zhao, Yan Y" <yan.y.zhao@intel.com>
Cc: 'Alex Williamson' <alex.williamson@redhat.com>,
	"qemu-devel@nongnu.org" <qemu-devel@nongnu.org>
Subject: Re: [Qemu-devel] vhost, iova, and dirty page tracking
Date: Thu, 19 Sep 2019 20:39:28 +0800	[thread overview]
Message-ID: <00084b36-3281-7c8d-5057-427eaabfb174@redhat.com> (raw)
In-Reply-To: <1ec55b2e-6a59-f1df-0604-5b524da0f001@redhat.com>


On 2019/9/19 下午7:14, Paolo Bonzini wrote:
> On 19/09/19 09:16, Tian, Kevin wrote:
>>>> why GPA1 and GPA2 should be both dirty?
>>>> even they have the same HVA due to overlaping virtual address space in
>>>> two processes, they still correspond to two physical pages.
>>>> don't get what's your meaning :)
>>> The point is not leave any corner case that is hard to debug or fix in
>>> the future.
>>>
>>> Let's just start by a single process, the API allows userspace to maps
>>> HVA to both GPA1 and GPA2. Since it knows GPA1 and GPA2 are equivalent,
>>> it's ok to sync just through GPA1. That means if you only log GPA2, it
>>> won't work.
>> I noted KVM itself doesn't consider such situation (one HVA is mapped
>> to multiple GPAs), when doing its dirty page tracking. If you look at
>> kvm_vcpu_mark_page_dirty, it simply finds the unique memslot which
>> contains the dirty gfn and then set the dirty bit within that slot. It
>> doesn't attempt to walk all memslots to find out any other GPA which
>> may be mapped to the same HVA.
>>
>> So there must be some disconnect here. let's hear from Paolo first and
>> understand the rationale behind such situation.
> In general, userspace cannot assume that it's okay to sync just through
> GPA1.  It must sync the host page if *either* GPA1 or GPA2 are marked dirty.


Maybe we need document this somewhere.


>
> The situation really only arises in special cases.  For example,
> 0xfffe0000..0xffffffff and 0xe0000..0xfffff might be the same memory.
>  From "info mtree" before the guest boots:
>
>      0000000000000000-ffffffffffffffff (prio -1, i/o): pci
>        00000000000e0000-00000000000fffff (prio 1, i/o): alias isa-bios
> @pc.bios 0000000000020000-000000000003ffff
>        00000000fffc0000-00000000ffffffff (prio 0, rom): pc.bios
>
> However, non-x86 machines may have other cases of aliased memory so it's
> a case that you should cover.
>
> Paolo


Any other issue that still need to be covered consider userspace need to 
sync both GPAs?

Thanks



  reply	other threads:[~2019-09-19 13:29 UTC|newest]

Thread overview: 40+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2019-09-16  1:51 [Qemu-devel] vhost, iova, and dirty page tracking Tian, Kevin
2019-09-16  8:33 ` Jason Wang
2019-09-17  8:48   ` Tian, Kevin
2019-09-17 10:36     ` Jason Wang
2019-09-18  1:44       ` Tian, Kevin
2019-09-18  6:10         ` Jason Wang
2019-09-18  7:41           ` Tian, Kevin
2019-09-18  8:37           ` Tian, Kevin
2019-09-19  1:05             ` Jason Wang
2019-09-19  5:28               ` Yan Zhao
2019-09-19  6:09                 ` Jason Wang
2019-09-19  6:17                   ` Yan Zhao
2019-09-19  6:32                     ` Jason Wang
2019-09-19  6:29                       ` Yan Zhao
2019-09-19  6:32                         ` Yan Zhao
2019-09-19  9:35                           ` Jason Wang
2019-09-19  9:36                             ` Yan Zhao
2019-09-19 10:08                               ` Jason Wang
2019-09-19 10:06                         ` Jason Wang
2019-09-19 10:16                           ` Yan Zhao
2019-09-19 12:14                             ` Jason Wang
2019-09-19  7:16                       ` Tian, Kevin
2019-09-19  9:37                         ` Jason Wang
2019-09-19 14:06                           ` Michael S. Tsirkin
2019-09-20  1:15                             ` Jason Wang
2019-09-20 10:02                               ` Michael S. Tsirkin
2019-09-19 11:14                         ` Paolo Bonzini
2019-09-19 12:39                           ` Jason Wang [this message]
2019-09-19 12:45                             ` Paolo Bonzini
2019-09-19 22:54                           ` Tian, Kevin
2019-09-20  1:18                             ` Jason Wang
2019-09-24  2:02                               ` Tian, Kevin
2019-09-25  3:46                                 ` Jason Wang
2019-09-17 14:54     ` Alex Williamson
2019-09-18  1:31       ` Tian, Kevin
2019-09-18  6:03         ` Jason Wang
2019-09-18  7:21           ` Tian, Kevin
2019-09-19 17:20             ` Alex Williamson
2019-09-19 22:40               ` Tian, Kevin
     [not found]       ` <AADFC41AFE54684AB9EE6CBC0274A5D19D57AFB7@SHSMSX104.ccr.corp.intel.com>
2019-09-18  2:15         ` Tian, Kevin

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=00084b36-3281-7c8d-5057-427eaabfb174@redhat.com \
    --to=jasowang@redhat.com \
    --cc=alex.williamson@redhat.com \
    --cc=kevin.tian@intel.com \
    --cc=pbonzini@redhat.com \
    --cc=qemu-devel@nongnu.org \
    --cc=yan.y.zhao@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).