linux-fsdevel.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Hao Xu <hao.xu@linux.dev>
To: Bernd Schubert <bernd.schubert@fastmail.fm>,
	"fuse-devel@lists.sourceforge.net"
	<fuse-devel@lists.sourceforge.net>,
	miklos@szeredi.hu
Cc: Antonio SJ Musumeci <trapexit@spawn.link>, linux-fsdevel@vger.kernel.org
Subject: Re: [RFC] FUSE: add another flag to support shared mmap in FOPEN_DIRECT_IO mode
Date: Wed, 3 May 2023 13:09:15 +0800	[thread overview]
Message-ID: <75f7fa44-8bf8-f986-63f5-993b7fc863f7@linux.dev> (raw)
In-Reply-To: <45ad47ae-5471-3d44-d3d6-2760dee0945d@fastmail.fm>

Hi Bernd,


On 5/2/23 15:37, Bernd Schubert wrote:
> Hi Hao,
>
> On 5/2/23 09:28, Hao Xu wrote:
>> Hi all,
>>
>>  From discussion with Bernd, I get that FOPEN_DIRECT_IO is designed 
>> for those user cases where users want strong coherency like network 
>> filesystems, where one server serves multiple remote clients. And 
>> thus shared mmap is disabled since local page cache existence breaks 
>> this kind of coherency.
>>
>> But here our use case is one virtiofs daemon serve one guest vm, We 
>> use FOPEN_DIRECT_IO to reduce memory footprint not for coherency. So 
>> we expect shared mmap works in this case. Here I suggest/am 
>> implementing adding another flag to indicate this kind of 
>> cases----use FOPEN_DIRECT_IO not for coherency----so that shared mmap 
>> works.
>
> Yeah it should work, but I think what you want is "DAX" - can you try 
> to enable it?
>
> fuse_i.h:    FUSE_DAX_ALWAYS,    /* "-o dax=always" */
> fuse_i.h:    FUSE_DAX_NEVER,        /* "-o dax=never" */
> fuse_i.h:    FUSE_DAX_INODE_USER,    /* "-o dax=inode" */
>
>
>
> Hope it helps,
> Bernd


Thanks for your suggestion, the thing is most IO in our case are small 
random read to small files. I believe Dax should help but doubt it may 
be not too much.

Another reason we are not leveraging DAX is the hypervisor software we 
use doesn't support DAX well. So I think the feature in this RFC should 
be good for

some non-DAX use cases (like ours)


Thanks,

Hao


      reply	other threads:[~2023-05-03  5:20 UTC|newest]

Thread overview: 3+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2023-05-02  7:28 [RFC] FUSE: add another flag to support shared mmap in FOPEN_DIRECT_IO mode Hao Xu
2023-05-02  7:37 ` Bernd Schubert
2023-05-03  5:09   ` Hao Xu [this message]

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=75f7fa44-8bf8-f986-63f5-993b7fc863f7@linux.dev \
    --to=hao.xu@linux.dev \
    --cc=bernd.schubert@fastmail.fm \
    --cc=fuse-devel@lists.sourceforge.net \
    --cc=linux-fsdevel@vger.kernel.org \
    --cc=miklos@szeredi.hu \
    --cc=trapexit@spawn.link \
    /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).