public inbox for linux-kernel@vger.kernel.org
 help / color / mirror / Atom feed
From: Oleg Nesterov <oleg@redhat.com>
To: K Prateek Nayak <kprateek.nayak@amd.com>
Cc: Christian Brauner <brauner@kernel.org>,
	Jeff Layton <jlayton@kernel.org>,
	David Howells <dhowells@redhat.com>,
	"Gautham R. Shenoy" <gautham.shenoy@amd.com>,
	Mateusz Guzik <mjguzik@gmail.com>,
	Neeraj Upadhyay <Neeraj.Upadhyay@amd.com>,
	Oliver Sang <oliver.sang@intel.com>,
	Swapnil Sapkal <swapnil.sapkal@amd.com>,
	WangYuli <wangyuli@uniontech.com>,
	linux-fsdevel@vger.kernel.org, linux-kernel@vger.kernel.org,
	Linus Torvalds <torvalds@linux-foundation.org>
Subject: Re: [PATCH 1/2] pipe: change pipe_write() to never add a zero-sized buffer
Date: Mon, 10 Feb 2025 18:22:00 +0100	[thread overview]
Message-ID: <20250210172200.GA16955@redhat.com> (raw)
In-Reply-To: <b050f92e-4117-4e93-8ec6-ec595fd8570a@amd.com>

Hi Prateek,

On 02/10, K Prateek Nayak wrote:
>
>  1-groups     1.00 [ -0.00]( 7.19)                0.95 [  4.90](12.39)
>  2-groups     1.00 [ -0.00]( 3.54)                1.02 [ -1.92]( 6.55)
>  4-groups     1.00 [ -0.00]( 2.78)                1.01 [ -0.85]( 2.18)
>  8-groups     1.00 [ -0.00]( 1.04)                0.99 [  0.63]( 0.77)
> 16-groups     1.00 [ -0.00]( 1.02)                1.00 [ -0.26]( 0.98)
>
> I don't see any regression / improvements from a performance standpoint

Yes, this patch shouldn't make any difference performance-wise, at least
in this case. Although I was thinking the same thing when I sent "pipe_read:
don't wake up the writer if the pipe is still full" ;)

> Tested-by: K Prateek Nayak <kprateek.nayak@amd.com>

Thanks! Please see v2, I've included you tag.

Any chance you can also test the patch below?

To me it looks like a cleanup which makes the "merge small writes" logic
more understandable. And note that "page-align the rest of the writes"
doesn't work anyway if "total_len & (PAGE_SIZE-1)" can't fit in the last
buffer.

However, in this particular case with DATASIZE=100 this patch can increase
the number of copy_page_from_iter()'s in pipe_write(). And with this change
receiver() can certainly get the short reads, so this can increase the
number of sys_read() calls.

So I am just curious if this change can cause any noticeable regression on
your machine.

Thank you!

Oleg.

--- a/fs/pipe.c
+++ b/fs/pipe.c
@@ -459,16 +459,16 @@ anon_pipe_write(struct kiocb *iocb, struct iov_iter *from)
 	was_empty = pipe_empty(head, pipe->tail);
 	chars = total_len & (PAGE_SIZE-1);
 	if (chars && !was_empty) {
-		unsigned int mask = pipe->ring_size - 1;
-		struct pipe_buffer *buf = &pipe->bufs[(head - 1) & mask];
+		struct pipe_buffer *buf = pipe_buf(pipe, head - 1);
 		int offset = buf->offset + buf->len;
+		int avail = PAGE_SIZE - offset;
 
-		if ((buf->flags & PIPE_BUF_FLAG_CAN_MERGE) &&
-		    offset + chars <= PAGE_SIZE) {
+		if (avail && (buf->flags & PIPE_BUF_FLAG_CAN_MERGE)) {
 			ret = pipe_buf_confirm(pipe, buf);
 			if (ret)
 				goto out;
 
+			chars = min_t(ssize_t, chars, avail);
 			ret = copy_page_from_iter(buf->page, offset, chars, from);
 			if (unlikely(ret < chars)) {
 				ret = -EFAULT;


  reply	other threads:[~2025-02-10 17:22 UTC|newest]

Thread overview: 16+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2025-02-09 15:07 [PATCH 0/2] pipe: change pipe_write() to never add a zero-sized buffer Oleg Nesterov
2025-02-09 15:07 ` [PATCH 1/2] " Oleg Nesterov
2025-02-09 17:26   ` Linus Torvalds
2025-02-09 18:02     ` Oleg Nesterov
2025-02-09 18:17       ` Linus Torvalds
2025-02-09 18:44         ` Oleg Nesterov
2025-02-09 18:52           ` Linus Torvalds
2025-02-09 19:15             ` Oleg Nesterov
2025-02-09 23:39               ` K Prateek Nayak
2025-02-10 17:22                 ` Oleg Nesterov [this message]
2025-02-10 17:37                   ` Linus Torvalds
2025-02-10 18:36                     ` Oleg Nesterov
2025-02-11  3:59                   ` K Prateek Nayak
2025-02-09 15:08 ` [PATCH 2/2] splice: add some pipe_buf_assert_len() checks Oleg Nesterov
2025-02-10 11:40 ` [PATCH v2 1/1] pipe: change pipe_write() to never add a zero-sized buffer Oleg Nesterov
2025-02-10 12:52 ` [PATCH 0/2] " Christian Brauner

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=20250210172200.GA16955@redhat.com \
    --to=oleg@redhat.com \
    --cc=Neeraj.Upadhyay@amd.com \
    --cc=brauner@kernel.org \
    --cc=dhowells@redhat.com \
    --cc=gautham.shenoy@amd.com \
    --cc=jlayton@kernel.org \
    --cc=kprateek.nayak@amd.com \
    --cc=linux-fsdevel@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=mjguzik@gmail.com \
    --cc=oliver.sang@intel.com \
    --cc=swapnil.sapkal@amd.com \
    --cc=torvalds@linux-foundation.org \
    --cc=wangyuli@uniontech.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