From: JeffleXu <jefflexu@linux.alibaba.com>
To: Miklos Szeredi <miklos@szeredi.hu>
Cc: Vivek Goyal <vgoyal@redhat.com>,
Stefan Hajnoczi <stefanha@redhat.com>,
linux-fsdevel@vger.kernel.org,
virtualization@lists.linux-foundation.org,
virtio-fs-list <virtio-fs@redhat.com>,
Joseph Qi <joseph.qi@linux.alibaba.com>,
Liu Bo <bo.liu@linux.alibaba.com>
Subject: Re: [PATCH v4 0/8] fuse,virtiofs: support per-file DAX
Date: Wed, 18 Aug 2021 11:39:54 +0800 [thread overview]
Message-ID: <1100b711-012d-d68b-7078-20eea6fa4bab@linux.alibaba.com> (raw)
In-Reply-To: <CAJfpegv01k5hEyJ3LPDWJoqB+vL8hwTan9dLu1pkkD0xoRuFzw@mail.gmail.com>
On 8/17/21 10:08 PM, Miklos Szeredi wrote:
> On Tue, 17 Aug 2021 at 15:22, JeffleXu <jefflexu@linux.alibaba.com> wrote:
>>
>>
>>
>> On 8/17/21 8:39 PM, Vivek Goyal wrote:
>>> On Tue, Aug 17, 2021 at 10:06:53AM +0200, Miklos Szeredi wrote:
>>>> On Tue, 17 Aug 2021 at 04:22, Jeffle Xu <jefflexu@linux.alibaba.com> wrote:
>>>>>
>>>>> This patchset adds support of per-file DAX for virtiofs, which is
>>>>> inspired by Ira Weiny's work on ext4[1] and xfs[2].
>>>>
>>>> Can you please explain the background of this change in detail?
>>>>
>>>> Why would an admin want to enable DAX for a particular virtiofs file
>>>> and not for others?
>>>
>>> Initially I thought that they needed it because they are downloading
>>> files on the fly from server. So they don't want to enable dax on the file
>>> till file is completely downloaded.
>>
>> Right, it's our initial requirement.
>>
>>
>>> But later I realized that they should
>>> be able to block in FUSE_SETUPMAPPING call and make sure associated
>>> file section has been downloaded before returning and solve the problem.
>>> So that can't be the primary reason.
>>
>> Saying we want to access 4KB of one file inside guest, if it goes
>> through FUSE request routine, then the fuse daemon only need to download
>> this 4KB from remote server. But if it goes through DAX, then the fuse
>> daemon need to download the whole DAX window (e.g., 2MB) from remote
>> server, so called amplification. Maybe we could decrease the DAX window
>> size, but it's a trade off.
>
> That could be achieved with a plain fuse filesystem on the host (which
> will get 4k READ requests for accesses to mapped area inside guest).
> Since this can be done selectively for files which are not yet
> downloaded, the extra layer wouldn't be a performance problem.
I'm not sure if I fully understand your idea. Then in this case, host
daemon only prepares 4KB while guest thinks that the whole DAX window
(e.g., 2MB) has been fully mapped. Then when guest really accesses the
remained part (2MB - 4KB), page fault is triggered, and now host daemon
is responsible for downloading the remained part?
>
> Is there a reason why that wouldn't work?
>
> Thanks,
> Miklos
>
--
Thanks,
Jeffle
WARNING: multiple messages have this Message-ID (diff)
From: JeffleXu <jefflexu@linux.alibaba.com>
To: Miklos Szeredi <miklos@szeredi.hu>
Cc: virtualization@lists.linux-foundation.org,
virtio-fs-list <virtio-fs@redhat.com>,
Joseph Qi <joseph.qi@linux.alibaba.com>,
Liu Bo <bo.liu@linux.alibaba.com>,
Stefan Hajnoczi <stefanha@redhat.com>,
linux-fsdevel@vger.kernel.org, Vivek Goyal <vgoyal@redhat.com>
Subject: Re: [PATCH v4 0/8] fuse,virtiofs: support per-file DAX
Date: Wed, 18 Aug 2021 11:39:54 +0800 [thread overview]
Message-ID: <1100b711-012d-d68b-7078-20eea6fa4bab@linux.alibaba.com> (raw)
In-Reply-To: <CAJfpegv01k5hEyJ3LPDWJoqB+vL8hwTan9dLu1pkkD0xoRuFzw@mail.gmail.com>
On 8/17/21 10:08 PM, Miklos Szeredi wrote:
> On Tue, 17 Aug 2021 at 15:22, JeffleXu <jefflexu@linux.alibaba.com> wrote:
>>
>>
>>
>> On 8/17/21 8:39 PM, Vivek Goyal wrote:
>>> On Tue, Aug 17, 2021 at 10:06:53AM +0200, Miklos Szeredi wrote:
>>>> On Tue, 17 Aug 2021 at 04:22, Jeffle Xu <jefflexu@linux.alibaba.com> wrote:
>>>>>
>>>>> This patchset adds support of per-file DAX for virtiofs, which is
>>>>> inspired by Ira Weiny's work on ext4[1] and xfs[2].
>>>>
>>>> Can you please explain the background of this change in detail?
>>>>
>>>> Why would an admin want to enable DAX for a particular virtiofs file
>>>> and not for others?
>>>
>>> Initially I thought that they needed it because they are downloading
>>> files on the fly from server. So they don't want to enable dax on the file
>>> till file is completely downloaded.
>>
>> Right, it's our initial requirement.
>>
>>
>>> But later I realized that they should
>>> be able to block in FUSE_SETUPMAPPING call and make sure associated
>>> file section has been downloaded before returning and solve the problem.
>>> So that can't be the primary reason.
>>
>> Saying we want to access 4KB of one file inside guest, if it goes
>> through FUSE request routine, then the fuse daemon only need to download
>> this 4KB from remote server. But if it goes through DAX, then the fuse
>> daemon need to download the whole DAX window (e.g., 2MB) from remote
>> server, so called amplification. Maybe we could decrease the DAX window
>> size, but it's a trade off.
>
> That could be achieved with a plain fuse filesystem on the host (which
> will get 4k READ requests for accesses to mapped area inside guest).
> Since this can be done selectively for files which are not yet
> downloaded, the extra layer wouldn't be a performance problem.
I'm not sure if I fully understand your idea. Then in this case, host
daemon only prepares 4KB while guest thinks that the whole DAX window
(e.g., 2MB) has been fully mapped. Then when guest really accesses the
remained part (2MB - 4KB), page fault is triggered, and now host daemon
is responsible for downloading the remained part?
>
> Is there a reason why that wouldn't work?
>
> Thanks,
> Miklos
>
--
Thanks,
Jeffle
_______________________________________________
Virtualization mailing list
Virtualization@lists.linux-foundation.org
https://lists.linuxfoundation.org/mailman/listinfo/virtualization
WARNING: multiple messages have this Message-ID (diff)
From: JeffleXu <jefflexu@linux.alibaba.com>
To: Miklos Szeredi <miklos@szeredi.hu>
Cc: virtualization@lists.linux-foundation.org,
virtio-fs-list <virtio-fs@redhat.com>,
Joseph Qi <joseph.qi@linux.alibaba.com>,
linux-fsdevel@vger.kernel.org, Vivek Goyal <vgoyal@redhat.com>
Subject: Re: [Virtio-fs] [PATCH v4 0/8] fuse,virtiofs: support per-file DAX
Date: Wed, 18 Aug 2021 11:39:54 +0800 [thread overview]
Message-ID: <1100b711-012d-d68b-7078-20eea6fa4bab@linux.alibaba.com> (raw)
In-Reply-To: <CAJfpegv01k5hEyJ3LPDWJoqB+vL8hwTan9dLu1pkkD0xoRuFzw@mail.gmail.com>
On 8/17/21 10:08 PM, Miklos Szeredi wrote:
> On Tue, 17 Aug 2021 at 15:22, JeffleXu <jefflexu@linux.alibaba.com> wrote:
>>
>>
>>
>> On 8/17/21 8:39 PM, Vivek Goyal wrote:
>>> On Tue, Aug 17, 2021 at 10:06:53AM +0200, Miklos Szeredi wrote:
>>>> On Tue, 17 Aug 2021 at 04:22, Jeffle Xu <jefflexu@linux.alibaba.com> wrote:
>>>>>
>>>>> This patchset adds support of per-file DAX for virtiofs, which is
>>>>> inspired by Ira Weiny's work on ext4[1] and xfs[2].
>>>>
>>>> Can you please explain the background of this change in detail?
>>>>
>>>> Why would an admin want to enable DAX for a particular virtiofs file
>>>> and not for others?
>>>
>>> Initially I thought that they needed it because they are downloading
>>> files on the fly from server. So they don't want to enable dax on the file
>>> till file is completely downloaded.
>>
>> Right, it's our initial requirement.
>>
>>
>>> But later I realized that they should
>>> be able to block in FUSE_SETUPMAPPING call and make sure associated
>>> file section has been downloaded before returning and solve the problem.
>>> So that can't be the primary reason.
>>
>> Saying we want to access 4KB of one file inside guest, if it goes
>> through FUSE request routine, then the fuse daemon only need to download
>> this 4KB from remote server. But if it goes through DAX, then the fuse
>> daemon need to download the whole DAX window (e.g., 2MB) from remote
>> server, so called amplification. Maybe we could decrease the DAX window
>> size, but it's a trade off.
>
> That could be achieved with a plain fuse filesystem on the host (which
> will get 4k READ requests for accesses to mapped area inside guest).
> Since this can be done selectively for files which are not yet
> downloaded, the extra layer wouldn't be a performance problem.
I'm not sure if I fully understand your idea. Then in this case, host
daemon only prepares 4KB while guest thinks that the whole DAX window
(e.g., 2MB) has been fully mapped. Then when guest really accesses the
remained part (2MB - 4KB), page fault is triggered, and now host daemon
is responsible for downloading the remained part?
>
> Is there a reason why that wouldn't work?
>
> Thanks,
> Miklos
>
--
Thanks,
Jeffle
next prev parent reply other threads:[~2021-08-18 3:39 UTC|newest]
Thread overview: 151+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-08-17 2:22 [PATCH v4 0/8] fuse,virtiofs: support per-file DAX Jeffle Xu
2021-08-17 2:22 ` [Virtio-fs] " Jeffle Xu
2021-08-17 2:22 ` Jeffle Xu
2021-08-17 2:22 ` [PATCH v4 1/8] fuse: add fuse_should_enable_dax() helper Jeffle Xu
2021-08-17 2:22 ` [Virtio-fs] " Jeffle Xu
2021-08-17 2:22 ` Jeffle Xu
2021-08-17 2:22 ` [PATCH v4 2/8] fuse: Make DAX mount option a tri-state Jeffle Xu
2021-08-17 2:22 ` [Virtio-fs] " Jeffle Xu
2021-08-17 2:22 ` Jeffle Xu
2021-08-17 2:22 ` [PATCH v4 3/8] fuse: support per-file DAX Jeffle Xu
2021-08-17 2:22 ` [Virtio-fs] " Jeffle Xu
2021-08-17 2:22 ` Jeffle Xu
2021-08-17 2:22 ` [PATCH v4 4/8] fuse: negotiate if server/client supports " Jeffle Xu
2021-08-17 2:22 ` [Virtio-fs] " Jeffle Xu
2021-08-17 2:22 ` Jeffle Xu
2021-08-17 2:22 ` [PATCH v4 5/8] fuse: enable " Jeffle Xu
2021-08-17 2:22 ` [Virtio-fs] " Jeffle Xu
2021-08-17 2:22 ` Jeffle Xu
2021-08-17 2:22 ` [PATCH v4 6/8] fuse: mark inode DONT_CACHE when per-file DAX indication changes Jeffle Xu
2021-08-17 2:22 ` [Virtio-fs] " Jeffle Xu
2021-08-17 2:22 ` Jeffle Xu
2021-08-17 10:26 ` [Virtio-fs] " Dr. David Alan Gilbert
2021-08-17 10:26 ` Dr. David Alan Gilbert
2021-08-17 10:26 ` Dr. David Alan Gilbert
2021-08-17 13:23 ` JeffleXu
2021-08-17 13:23 ` JeffleXu
2021-08-17 13:23 ` JeffleXu
2021-08-17 2:22 ` [PATCH v4 7/8] fuse: support changing per-file DAX flag inside guest Jeffle Xu
2021-08-17 2:22 ` [Virtio-fs] " Jeffle Xu
2021-08-17 2:22 ` Jeffle Xu
2021-08-17 2:22 ` [PATCH v4 8/8] fuse: show '-o dax=inode' option only when FUSE server supports Jeffle Xu
2021-08-17 2:22 ` [Virtio-fs] " Jeffle Xu
2021-08-17 2:22 ` Jeffle Xu
2021-08-17 2:23 ` [virtiofsd PATCH v4 0/4] virtiofsd: support per-file DAX Jeffle Xu
2021-08-17 2:23 ` [Virtio-fs] " Jeffle Xu
2021-08-17 2:23 ` Jeffle Xu
2021-08-17 2:23 ` [virtiofsd PATCH v4 1/4] virtiofsd: add .ioctl() support Jeffle Xu
2021-08-17 2:23 ` [Virtio-fs] " Jeffle Xu
2021-08-17 2:23 ` Jeffle Xu
2021-08-18 17:33 ` Vivek Goyal
2021-08-18 17:33 ` [Virtio-fs] " Vivek Goyal
2021-08-18 17:33 ` Vivek Goyal
2021-08-17 2:23 ` [virtiofsd PATCH v4 2/4] virtiofsd: expand fuse protocol to support per-file DAX Jeffle Xu
2021-08-17 2:23 ` [Virtio-fs] " Jeffle Xu
2021-08-17 2:23 ` Jeffle Xu
2021-08-17 2:23 ` [virtiofsd PATCH v4 3/4] virtiofsd: support per-file DAX negotiation in FUSE_INIT Jeffle Xu
2021-08-17 2:23 ` [Virtio-fs] " Jeffle Xu
2021-08-17 2:23 ` Jeffle Xu
2021-08-17 17:15 ` [Virtio-fs] " Dr. David Alan Gilbert
2021-08-17 17:15 ` Dr. David Alan Gilbert
2021-08-17 17:15 ` Dr. David Alan Gilbert
2021-08-18 5:28 ` JeffleXu
2021-08-18 5:28 ` JeffleXu
2021-08-18 5:28 ` JeffleXu
2021-08-19 13:57 ` Dr. David Alan Gilbert
2021-08-19 13:57 ` Dr. David Alan Gilbert
2021-08-19 13:57 ` Dr. David Alan Gilbert
2021-08-18 17:30 ` Vivek Goyal
2021-08-18 17:30 ` Vivek Goyal
2021-08-18 17:30 ` Vivek Goyal
2021-08-17 2:23 ` [virtiofsd PATCH v4 4/4] virtiofsd: support per-file DAX in FUSE_LOOKUP Jeffle Xu
2021-08-17 2:23 ` [Virtio-fs] " Jeffle Xu
2021-08-17 2:23 ` Jeffle Xu
2021-08-17 19:00 ` [Virtio-fs] " Dr. David Alan Gilbert
2021-08-17 19:00 ` Dr. David Alan Gilbert
2021-08-17 19:00 ` Dr. David Alan Gilbert
2021-08-18 5:46 ` JeffleXu
2021-08-18 5:46 ` JeffleXu
2021-08-18 5:46 ` JeffleXu
2021-08-19 13:08 ` Dr. David Alan Gilbert
2021-08-19 13:08 ` Dr. David Alan Gilbert
2021-08-19 13:08 ` Dr. David Alan Gilbert
2021-08-20 5:03 ` JeffleXu
2021-08-20 5:03 ` JeffleXu
2021-08-20 5:03 ` JeffleXu
2021-08-24 10:15 ` Greg Kurz
2021-08-24 10:15 ` Greg Kurz
2021-08-24 10:15 ` Greg Kurz
2021-09-08 10:34 ` JeffleXu
2021-09-08 10:34 ` JeffleXu
2021-09-08 10:34 ` JeffleXu
2021-08-17 8:06 ` [PATCH v4 0/8] fuse,virtiofs: support per-file DAX Miklos Szeredi
2021-08-17 8:06 ` [Virtio-fs] " Miklos Szeredi
2021-08-17 9:32 ` Dr. David Alan Gilbert
2021-08-17 9:32 ` Dr. David Alan Gilbert
2021-08-17 9:32 ` Dr. David Alan Gilbert
2021-08-17 10:09 ` Miklos Szeredi
2021-08-17 10:09 ` Miklos Szeredi
2021-08-17 10:37 ` Dr. David Alan Gilbert
2021-08-17 10:37 ` Dr. David Alan Gilbert
2021-08-17 10:37 ` Dr. David Alan Gilbert
2021-08-17 13:08 ` JeffleXu
2021-08-17 13:08 ` JeffleXu
2021-08-17 13:08 ` JeffleXu
2021-08-17 14:11 ` Miklos Szeredi
2021-08-17 14:11 ` Miklos Szeredi
2021-08-17 15:19 ` Vivek Goyal
2021-08-17 15:19 ` Vivek Goyal
2021-08-17 15:19 ` Vivek Goyal
2021-08-17 14:54 ` Vivek Goyal
2021-08-17 14:54 ` Vivek Goyal
2021-08-17 14:54 ` Vivek Goyal
2021-08-18 5:10 ` JeffleXu
2021-08-18 5:10 ` JeffleXu
2021-08-18 5:10 ` JeffleXu
2021-08-19 6:14 ` JeffleXu
2021-08-19 6:14 ` JeffleXu
2021-08-19 6:14 ` JeffleXu
2021-08-17 12:40 ` Vivek Goyal
2021-08-17 12:40 ` Vivek Goyal
2021-08-17 12:40 ` Vivek Goyal
2021-09-16 8:21 ` JeffleXu
2021-09-16 8:21 ` JeffleXu
2021-09-16 8:21 ` JeffleXu
2021-09-18 3:06 ` JeffleXu
2021-09-18 3:06 ` JeffleXu
2021-09-18 3:06 ` JeffleXu
2021-09-19 19:45 ` Vivek Goyal
2021-09-19 19:45 ` Vivek Goyal
2021-09-19 19:45 ` Vivek Goyal
2021-09-22 8:16 ` JeffleXu
2021-09-22 8:16 ` JeffleXu
2021-09-22 8:16 ` JeffleXu
2021-08-17 12:39 ` Vivek Goyal
2021-08-17 12:39 ` [Virtio-fs] " Vivek Goyal
2021-08-17 12:39 ` Vivek Goyal
2021-08-17 13:22 ` JeffleXu
2021-08-17 13:22 ` [Virtio-fs] " JeffleXu
2021-08-17 13:22 ` JeffleXu
2021-08-17 14:08 ` Miklos Szeredi
2021-08-17 14:08 ` [Virtio-fs] " Miklos Szeredi
2021-08-18 3:39 ` JeffleXu [this message]
2021-08-18 3:39 ` JeffleXu
2021-08-18 3:39 ` JeffleXu
2021-08-18 5:08 ` Miklos Szeredi
2021-08-18 5:08 ` [Virtio-fs] " Miklos Szeredi
2021-08-18 16:58 ` Vivek Goyal
2021-08-18 16:58 ` [Virtio-fs] " Vivek Goyal
2021-08-18 16:58 ` Vivek Goyal
2021-09-03 5:30 ` JeffleXu
2021-09-03 5:30 ` [Virtio-fs] " JeffleXu
2021-09-03 5:30 ` JeffleXu
2021-09-07 14:51 ` Miklos Szeredi
2021-09-07 14:51 ` [Virtio-fs] " Miklos Szeredi
2021-08-17 14:57 ` Vivek Goyal
2021-08-17 14:57 ` [Virtio-fs] " Vivek Goyal
2021-08-17 14:57 ` Vivek Goyal
2021-08-18 5:20 ` JeffleXu
2021-08-18 5:20 ` [Virtio-fs] " JeffleXu
2021-08-18 5:20 ` JeffleXu
2021-08-30 23:31 ` [Virtio-fs] " Liu Bo
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=1100b711-012d-d68b-7078-20eea6fa4bab@linux.alibaba.com \
--to=jefflexu@linux.alibaba.com \
--cc=bo.liu@linux.alibaba.com \
--cc=joseph.qi@linux.alibaba.com \
--cc=linux-fsdevel@vger.kernel.org \
--cc=miklos@szeredi.hu \
--cc=stefanha@redhat.com \
--cc=vgoyal@redhat.com \
--cc=virtio-fs@redhat.com \
--cc=virtualization@lists.linux-foundation.org \
/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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.