public inbox for linux-nfs@vger.kernel.org
 help / color / mirror / Atom feed
From: trondmy@kernel.org
To: linux-nfs@vger.kernel.org
Subject: [PATCH 08/16] SUNRPC: Avoid unnecessary copies in xdr_buf_pages_copy_left/right()
Date: Wed,  9 Dec 2020 09:47:53 -0500	[thread overview]
Message-ID: <20201209144801.700778-9-trondmy@kernel.org> (raw)
In-Reply-To: <20201209144801.700778-8-trondmy@kernel.org>

From: Trond Myklebust <trond.myklebust@hammerspace.com>

Moving data in the XDR page arrays can be expensive, since it can
involve touching up to a megabyte of data. Try to avoid doing so for the
case where the server returned less data than we preallocated.

Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>
---
 net/sunrpc/xdr.c | 16 ++++++++++++----
 1 file changed, 12 insertions(+), 4 deletions(-)

diff --git a/net/sunrpc/xdr.c b/net/sunrpc/xdr.c
index 7ca6208e7623..86a48c4cdcfc 100644
--- a/net/sunrpc/xdr.c
+++ b/net/sunrpc/xdr.c
@@ -600,14 +600,20 @@ static void xdr_buf_tail_shift_right(const struct xdr_buf *buf,
 static void xdr_buf_pages_shift_right(const struct xdr_buf *buf,
 				      unsigned int base, unsigned int shift)
 {
+	struct xdr_buf subbuf;
+
 	if (!shift)
 		return;
 	if (base >= buf->page_len) {
 		xdr_buf_tail_shift_right(buf, base - buf->page_len, shift);
 		return;
 	}
-	xdr_buf_tail_shift_right(buf, 0, shift);
-	xdr_buf_pages_copy_right(buf, base, shift);
+	base += buf->head->iov_len;
+	if (base >= buf->len)
+		return;
+	xdr_buf_subsegment(buf, &subbuf, base, buf->len - base);
+	xdr_buf_tail_shift_right(&subbuf, 0, shift);
+	xdr_buf_pages_copy_right(&subbuf, 0, shift);
 }
 
 static void xdr_buf_head_shift_right(const struct xdr_buf *buf,
@@ -710,14 +716,16 @@ static void xdr_buf_tail_shift_left(const struct xdr_buf *buf,
 static void xdr_buf_pages_shift_left(const struct xdr_buf *buf,
 				     unsigned int base, unsigned int shift)
 {
+	struct xdr_buf subbuf;
 	if (!shift)
 		return;
 	if (base >= buf->page_len) {
 		xdr_buf_tail_shift_left(buf, base - buf->page_len, shift);
 		return;
 	}
-	xdr_buf_pages_copy_left(buf, base, shift);
-	xdr_buf_tail_shift_left(buf, 0, shift);
+	xdr_buf_subsegment(buf, &subbuf, 0, buf->len);
+	xdr_buf_pages_copy_left(&subbuf, base, shift);
+	xdr_buf_tail_shift_left(&subbuf, 0, shift);
 }
 
 /**
-- 
2.29.2


  reply	other threads:[~2020-12-09 14:49 UTC|newest]

Thread overview: 21+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-12-09 14:47 [PATCH 00/16] Fixes for the NFSv4.2 READ_PLUS operation trondmy
2020-12-09 14:47 ` [PATCH 01/16] SUNRPC: _shift_data_left/right_pages should check the shift length trondmy
2020-12-09 14:47   ` [PATCH 02/16] SUNRPC: Fixes for xdr_align_data() trondmy
2020-12-09 14:47     ` [PATCH 03/16] SUNRPC: Fix xdr_expand_hole() trondmy
2020-12-09 14:47       ` [PATCH 04/16] SUNRPC: Cleanup xdr_shrink_bufhead() trondmy
2020-12-09 14:47         ` [PATCH 05/16] SUNRPC: _copy_to/from_pages() now check for zero length trondmy
2020-12-09 14:47           ` [PATCH 06/16] SUNRPC: Clean up open coded setting of the xdr_stream 'nwords' field trondmy
2020-12-09 14:47             ` [PATCH 07/16] SUNRPC: Cleanup - constify a number of xdr_buf helpers trondmy
2020-12-09 14:47               ` trondmy [this message]
2020-12-09 14:47                 ` [PATCH 09/16] NFSv4.2: Ensure we always reset the result->count in decode_read_plus() trondmy
2020-12-09 14:47                   ` [PATCH 10/16] NFSv4.2: decode_read_plus_data() must skip padding after data segment trondmy
2020-12-09 14:47                     ` [PATCH 11/16] NFSv4.2: decode_read_plus_hole() needs to check the extent offset trondmy
2020-12-09 14:47                       ` [PATCH 12/16] NFSv4.2: Handle hole lengths that exceed the READ_PLUS read buffer trondmy
2020-12-09 14:47                         ` [PATCH 13/16] NFSv4.2: Don't error when exiting early on a READ_PLUS buffer overflow trondmy
2020-12-09 14:47                           ` [PATCH 14/16] NFSv4.2: Deal with potential READ_PLUS data extent " trondmy
2020-12-09 14:48                             ` [PATCH 15/16] nfsd: Fixes for nfsd4_encode_read_plus_data() trondmy
2020-12-09 14:48                               ` [PATCH 16/16] nfsd: Don't set eof on a truncated READ_PLUS trondmy
2020-12-09 16:16                               ` [PATCH 15/16] nfsd: Fixes for nfsd4_encode_read_plus_data() Chuck Lever
2020-12-09 16:39                                 ` Trond Myklebust
2020-12-09 16:57                                   ` Chuck Lever
2020-12-09 17:01                                     ` Trond Myklebust

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=20201209144801.700778-9-trondmy@kernel.org \
    --to=trondmy@kernel.org \
    --cc=linux-nfs@vger.kernel.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 a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox