mirror of https://lore.kernel.org/lkml/
 help / color / mirror / Atom feed
From: Horst Birthelmer <horst@birthelmer.de>
To: Miklos Szeredi <miklos@szeredi.hu>
Cc: Horst Birthelmer <horst@birthelmer.com>,
	 Bernd Schubert <bschubert@ddn.com>,
	Joanne Koong <joannelkoong@gmail.com>,
	 Luis Henriques <luis@igalia.com>,
	linux-kernel@vger.kernel.org, linux-fsdevel@vger.kernel.org,
	 Horst Birthelmer <hbirthelmer@ddn.com>
Subject: Re: Re: [PATCH v6 1/3] fuse: add compound command to combine multiple requests
Date: Fri, 6 Mar 2026 15:27:06 +0100	[thread overview]
Message-ID: <aarhiEz5nuveOQdP@fedora.fritz.box> (raw)
In-Reply-To: <CAJfpegsNpWb-miyx+P-W_=11dB3Shz6ikNOQ6Qp_hyOp1DqE9A@mail.gmail.com>

On Fri, Feb 27, 2026 at 10:45:36AM +0100, Miklos Szeredi wrote:
> On Thu, 26 Feb 2026 at 17:43, Horst Birthelmer <horst@birthelmer.com> wrote:
> 
> > +
> > +fallback_separate:
> > +       /* Kernel tries to fallback to separate requests */
> > +       if (!(compound->compound_header.flags & FUSE_COMPOUND_ATOMIC))
> > +               ret = fuse_compound_fallback_separate(compound);
> > +
> > +out:
> > +       kfree(resp_payload_buffer);
> > +out_free_buffer:
> > +       kfree(buffer);
> > +       return ret;
> > +}
> 
> If we go with the list of fuse_args, then all the above logic could go
> into the lower layer (dev.c) which already handles fuse_args ->
> request -> fuse_args conversion.  What's needed is mostly just a loop
> that repeats this for all the sub requests.
> 
> 

I have actually implemented this idea and avoided any memory allocation.
So the short version is, it can be done.

But to me this looks kinda ugly and a bit wrong. I have to check in the 
lower layer for an opcode from the upper layer and 'stream' the args.
(in fuse_dev_do_read() or somewhere in that region there has to be 
a check for FUSE_COMPOUND and then call into different code)

When handled on that level it has to be handled for io-uring
slightly differently as well. 

I will test this a bit more and provide a new version unless someone 
tells me that this is not the right direction.

> 
> Thanks,
> Miklos

Thanks,
Horst

  parent reply	other threads:[~2026-03-06 14:27 UTC|newest]

Thread overview: 31+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-02-26 16:43 [PATCH v6 0/3] fuse: compound commands Horst Birthelmer
2026-02-26 16:43 ` [PATCH v6 1/3] fuse: add compound command to combine multiple requests Horst Birthelmer
2026-02-26 23:05   ` Joanne Koong
2026-02-27  9:45   ` Miklos Szeredi
2026-02-27 10:48     ` Horst Birthelmer
2026-02-27 11:29       ` Miklos Szeredi
2026-02-27 11:37         ` Horst Birthelmer
2026-02-27 11:58           ` Miklos Szeredi
2026-03-02  9:56     ` Horst Birthelmer
2026-03-02 11:03       ` Miklos Szeredi
2026-03-02 13:19         ` Horst Birthelmer
2026-03-02 13:30           ` Miklos Szeredi
2026-03-06 14:27     ` Horst Birthelmer [this message]
2026-02-26 16:43 ` [PATCH v6 2/3] fuse: create helper functions for filling in fuse args for open and getattr Horst Birthelmer
2026-02-26 16:43 ` [PATCH v6 3/3] fuse: add an implementation of open+getattr Horst Birthelmer
2026-02-26 19:12   ` Joanne Koong
2026-02-27  7:48     ` Horst Birthelmer
2026-02-27 17:51       ` Joanne Koong
2026-02-27 18:07         ` Joanne Koong
2026-02-28  8:14           ` Horst Birthelmer
2026-03-02 18:56             ` Joanne Koong
2026-03-02 20:03               ` Bernd Schubert
2026-03-03  5:06                 ` Darrick J. Wong
2026-03-03  7:26                   ` Horst Birthelmer
2026-03-03 10:03                   ` Miklos Szeredi
2026-03-03 10:38                     ` Horst Birthelmer
2026-03-03 21:19                       ` Joanne Koong
2026-03-04  9:11                         ` Horst Birthelmer
2026-03-04 21:42                           ` Joanne Koong
2026-03-03 23:13                     ` Joanne Koong
2026-03-04  9:37                       ` Miklos Szeredi

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=aarhiEz5nuveOQdP@fedora.fritz.box \
    --to=horst@birthelmer.de \
    --cc=bschubert@ddn.com \
    --cc=hbirthelmer@ddn.com \
    --cc=horst@birthelmer.com \
    --cc=joannelkoong@gmail.com \
    --cc=linux-fsdevel@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=luis@igalia.com \
    --cc=miklos@szeredi.hu \
    /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

all inboxes | Powered by JetHome®