public inbox for [email protected]
 help / color / mirror / Atom feed
From: Jens Axboe <[email protected]>
To: Keith Busch <[email protected]>,
	[email protected], [email protected],
	[email protected], [email protected]
Cc: [email protected], [email protected],
	[email protected], Keith Busch <[email protected]>
Subject: Re: [PATCHv7 1/6] io_uring/rw: move fixed buffer import to issue path
Date: Wed, 26 Feb 2025 12:04:04 -0700	[thread overview]
Message-ID: <[email protected]> (raw)
In-Reply-To: <[email protected]>

On 2/26/25 11:20 AM, Keith Busch wrote:
> From: Keith Busch <[email protected]>
> 
> Registered buffers may depend on a linked command, which makes the prep
> path too early to import. Move to the issue path when the node is
> actually needed like all the other users of fixed buffers.

Conceptually I think this patch is fine, but it does bother me with
random bool arguments. We could fold in something like the (totally
tested) below diff to get rid of that. What do you think?

diff --git a/io_uring/rw.c b/io_uring/rw.c
index 728d695d2552..a8a46a32f20d 100644
--- a/io_uring/rw.c
+++ b/io_uring/rw.c
@@ -248,8 +248,8 @@ static int io_prep_rw_pi(struct io_kiocb *req, struct io_rw *rw, int ddir,
 	return ret;
 }
 
-static int io_prep_rw(struct io_kiocb *req, const struct io_uring_sqe *sqe,
-		      int ddir, bool do_import)
+static int __io_prep_rw(struct io_kiocb *req, const struct io_uring_sqe *sqe,
+			int ddir)
 {
 	struct io_rw *rw = io_kiocb_to_cmd(req, struct io_rw);
 	unsigned ioprio;
@@ -285,14 +285,6 @@ static int io_prep_rw(struct io_kiocb *req, const struct io_uring_sqe *sqe,
 	rw->len = READ_ONCE(sqe->len);
 	rw->flags = READ_ONCE(sqe->rw_flags);
 
-	if (do_import && !io_do_buffer_select(req)) {
-		struct io_async_rw *io = req->async_data;
-
-		ret = io_import_rw_buffer(ddir, req, io, 0);
-		if (unlikely(ret))
-			return ret;
-	}
-
 	attr_type_mask = READ_ONCE(sqe->attr_type_mask);
 	if (attr_type_mask) {
 		u64 attr_ptr;
@@ -307,27 +299,52 @@ static int io_prep_rw(struct io_kiocb *req, const struct io_uring_sqe *sqe,
 	return ret;
 }
 
+static int io_rw_do_import(struct io_kiocb *req, int ddir)
+{
+	if (!io_do_buffer_select(req)) {
+		struct io_async_rw *io = req->async_data;
+		int ret;
+
+		ret = io_import_rw_buffer(ddir, req, io, 0);
+		if (unlikely(ret))
+			return ret;
+	}
+
+	return 0;
+}
+
+static int io_prep_rw(struct io_kiocb *req, const struct io_uring_sqe *sqe,
+		      int ddir)
+{
+	int ret;
+
+	ret = __io_prep_rw(req, sqe, ddir);
+	if (unlikely(ret))
+		return ret;
+
+	return io_rw_do_import(req, ITER_DEST);
+}
+
 int io_prep_read(struct io_kiocb *req, const struct io_uring_sqe *sqe)
 {
-	return io_prep_rw(req, sqe, ITER_DEST, true);
+	return io_prep_rw(req, sqe, ITER_DEST);
 }
 
 int io_prep_write(struct io_kiocb *req, const struct io_uring_sqe *sqe)
 {
-	return io_prep_rw(req, sqe, ITER_SOURCE, true);
+	return io_prep_rw(req, sqe, ITER_SOURCE);
 }
 
 static int io_prep_rwv(struct io_kiocb *req, const struct io_uring_sqe *sqe,
 		       int ddir)
 {
-	const bool do_import = !(req->flags & REQ_F_BUFFER_SELECT);
 	int ret;
 
-	ret = io_prep_rw(req, sqe, ddir, do_import);
+	ret = io_prep_rw(req, sqe, ddir);
 	if (unlikely(ret))
 		return ret;
-	if (do_import)
-		return 0;
+	if (!(req->flags & REQ_F_BUFFER_SELECT))
+		return io_rw_do_import(req, ddir);
 
 	/*
 	 * Have to do this validation here, as this is in io_read() rw->len
@@ -364,12 +381,12 @@ static int io_init_rw_fixed(struct io_kiocb *req, unsigned int issue_flags,
 
 int io_prep_read_fixed(struct io_kiocb *req, const struct io_uring_sqe *sqe)
 {
-	return io_prep_rw(req, sqe, ITER_DEST, false);
+	return io_prep_rw(req, sqe, ITER_DEST);
 }
 
 int io_prep_write_fixed(struct io_kiocb *req, const struct io_uring_sqe *sqe)
 {
-	return io_prep_rw(req, sqe, ITER_SOURCE, false);
+	return io_prep_rw(req, sqe, ITER_SOURCE);
 }
 
 /*
@@ -385,7 +402,7 @@ int io_read_mshot_prep(struct io_kiocb *req, const struct io_uring_sqe *sqe)
 	if (!(req->flags & REQ_F_BUFFER_SELECT))
 		return -EINVAL;
 
-	ret = io_prep_rw(req, sqe, ITER_DEST, false);
+	ret = io_prep_rw(req, sqe, ITER_DEST);
 	if (unlikely(ret))
 		return ret;
 

-- 
Jens Axboe

  reply	other threads:[~2025-02-26 19:04 UTC|newest]

Thread overview: 21+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2025-02-26 18:20 [PATCHv7 0/6] ublk zero copy support Keith Busch
2025-02-26 18:20 ` [PATCHv7 1/6] io_uring/rw: move fixed buffer import to issue path Keith Busch
2025-02-26 19:04   ` Jens Axboe [this message]
2025-02-26 20:20     ` Jens Axboe
2025-02-27 21:46       ` Keith Busch
2025-02-26 18:20 ` [PATCHv7 2/6] io_uring: add support for kernel registered bvecs Keith Busch
2025-02-26 20:36   ` Jens Axboe
2025-02-27 15:54   ` Pavel Begunkov
2025-02-27 16:04     ` Jens Axboe
2025-02-27 22:45     ` Keith Busch
2025-02-28 12:19       ` Pavel Begunkov
2025-02-26 18:20 ` [PATCHv7 3/6] nvme: map uring_cmd data even if address is 0 Keith Busch
2025-02-26 18:20 ` [PATCHv7 4/6] ublk: zc register/unregister bvec Keith Busch
2025-02-26 18:40   ` Jens Axboe
2025-02-26 18:21 ` [PATCHv7 5/6] io_uring: add abstraction for buf_table rsrc data Keith Busch
2025-02-26 18:21 ` [PATCHv7 6/6] io_uring: cache nodes and mapped buffers Keith Busch
2025-02-26 18:48   ` Jens Axboe
2025-02-26 19:36   ` Jens Axboe
2025-02-26 19:43     ` Jens Axboe
2025-02-26 19:58       ` Jens Axboe
2025-02-26 20:00         ` Keith Busch

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 \
    [email protected] \
    [email protected] \
    [email protected] \
    [email protected] \
    [email protected] \
    [email protected] \
    [email protected] \
    [email protected] \
    [email protected] \
    [email protected] \
    [email protected] \
    /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