All of lore.kernel.org
 help / color / mirror / Atom feed
From: "Darrick J. Wong" <djwong@kernel.org>
To: djwong@kernel.org
Cc: Omar Sandoval <osandov@fb.com>, Christoph Hellwig <hch@lst.de>,
	linux-xfs@vger.kernel.org, osandov@osandov.com, osandov@fb.com,
	hch@lst.de
Subject: [PATCH 8/9] xfs: don't try redundant allocations in xfs_rtallocate_extent_near()
Date: Thu, 19 Oct 2023 09:30:40 -0700	[thread overview]
Message-ID: <169773211847.225862.10218692865813765251.stgit@frogsfrogsfrogs> (raw)
In-Reply-To: <169773211712.225862.9408784830071081083.stgit@frogsfrogsfrogs>

From: Omar Sandoval <osandov@fb.com>

xfs_rtallocate_extent_near() tries to find a free extent as close to a
target bitmap block given by bbno as possible, which may be before or
after bbno. Searching backwards has a complication: the realtime summary
accounts for free space _starting_ in a bitmap block, but not straddling
or ending in a bitmap block. So, when the negative search finds a free
extent in the realtime summary, in order to end up closer to the target,
it looks for the end of the free extent. For example, if bbno - 2 has a
free extent, then it will check bbno - 1, then bbno - 2. But then if
bbno - 3 has a free extent, it will check bbno - 1 again, then bbno - 2
again, and then bbno - 3. This results in a quadratic loop, which is
completely pointless since the repeated checks won't find anything new.

Fix it by remembering where we last checked up to and continue from
there. This also obviates the need for a check of the realtime summary.

Signed-off-by: Omar Sandoval <osandov@fb.com>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
---
 fs/xfs/xfs_rtalloc.c |   54 ++++++--------------------------------------------
 1 file changed, 7 insertions(+), 47 deletions(-)


diff --git a/fs/xfs/xfs_rtalloc.c b/fs/xfs/xfs_rtalloc.c
index 3aa9634a9e76..b743da885ed6 100644
--- a/fs/xfs/xfs_rtalloc.c
+++ b/fs/xfs/xfs_rtalloc.c
@@ -477,6 +477,7 @@ xfs_rtallocate_extent_near(
 	}
 	bbno = xfs_rtx_to_rbmblock(mp, start);
 	i = 0;
+	j = -1;
 	ASSERT(minlen != 0);
 	log2len = xfs_highbit32(minlen);
 	/*
@@ -527,33 +528,13 @@ xfs_rtallocate_extent_near(
 			 */
 			else {		/* i < 0 */
 				/*
-				 * Loop backwards through the bitmap blocks from
-				 * the starting point-1 up to where we are now.
-				 * There should be an extent which ends in this
-				 * bitmap block and is long enough.
+				 * Loop backwards through the bitmap blocks
+				 * from where we last checked down to where we
+				 * are now.  There should be an extent which
+				 * ends in this bitmap block and is long
+				 * enough.
 				 */
-				for (j = -1; j > i; j--) {
-					/*
-					 * Grab the summary information for
-					 * this bitmap block.
-					 */
-					error = xfs_rtany_summary(args,
-							log2len,
-							mp->m_rsumlevels - 1,
-							bbno + j, &maxlog);
-					if (error) {
-						return error;
-					}
-					/*
-					 * If there's no extent given in the
-					 * summary that means the extent we
-					 * found must carry over from an
-					 * earlier block.  If there is an
-					 * extent given, we've already tried
-					 * that allocation, don't do it again.
-					 */
-					if (maxlog >= 0)
-						continue;
+				for (; j >= i; j--) {
 					error = xfs_rtallocate_extent_block(args,
 							bbno + j, minlen,
 							maxavail, len, &n, prod,
@@ -569,27 +550,6 @@ xfs_rtallocate_extent_near(
 						return 0;
 					}
 				}
-				/*
-				 * There weren't intervening bitmap blocks
-				 * with a long enough extent, or the
-				 * allocation didn't work for some reason
-				 * (i.e. it's a little * too short).
-				 * Try to allocate from the summary block
-				 * that we found.
-				 */
-				error = xfs_rtallocate_extent_block(args,
-						bbno + i, minlen, maxavail, len,
-						&n, prod, &r);
-				if (error) {
-					return error;
-				}
-				/*
-				 * If it works, return the extent.
-				 */
-				if (r != NULLRTEXTNO) {
-					*rtx = r;
-					return 0;
-				}
 			}
 		}
 		/*


  parent reply	other threads:[~2023-10-19 16:30 UTC|newest]

Thread overview: 47+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2023-10-19 16:07 [ANNOUNCE] xfs-linux: realtime work branch rtalloc-speedups-6.7 updated to e0f7422f54b0 Darrick J. Wong
2023-10-19 16:19 ` [PATCHSET v1.1 0/4] xfs: minor bugfixes for rt stuff Darrick J. Wong
2023-10-19 16:21   ` [PATCH 1/4] xfs: bump max fsgeom struct version Darrick J. Wong
2023-10-19 16:21   ` [PATCH 2/4] xfs: hoist freeing of rt data fork extent mappings Darrick J. Wong
2023-10-19 16:22   ` [PATCH 3/4] xfs: prevent rt growfs when quota is enabled Darrick J. Wong
2023-10-19 16:22   ` [PATCH 4/4] xfs: rt stubs should return negative errnos when rt disabled Darrick J. Wong
2023-10-19 16:20 ` [PATCHSET v1.1 0/8] xfs: clean up realtime type usage Darrick J. Wong
2023-10-19 16:22   ` [PATCH 1/8] xfs: fix units conversion error in xfs_bmap_del_extent_delay Darrick J. Wong
2023-10-19 16:22   ` [PATCH 2/8] xfs: make sure maxlen is still congruent with prod when rounding down Darrick J. Wong
2023-10-19 16:23   ` [PATCH 3/8] xfs: move the xfs_rtbitmap.c declarations to xfs_rtbitmap.h Darrick J. Wong
2023-10-19 16:23   ` [PATCH 4/8] xfs: convert xfs_extlen_t to xfs_rtxlen_t in the rt allocator Darrick J. Wong
2023-10-19 16:23   ` [PATCH 5/8] xfs: convert rt bitmap/summary block numbers to xfs_fileoff_t Darrick J. Wong
2023-10-19 16:23   ` [PATCH 6/8] xfs: convert rt bitmap extent lengths to xfs_rtbxlen_t Darrick J. Wong
2023-10-19 16:24   ` [PATCH 7/8] xfs: rename xfs_verify_rtext to xfs_verify_rtbext Darrick J. Wong
2023-10-19 16:24   ` [PATCH 8/8] xfs: convert rt extent numbers to xfs_rtxnum_t Darrick J. Wong
2023-10-19 16:20 ` [PATCHSET v1.1 0/7] xfs: refactor rt extent unit conversions Darrick J. Wong
2023-10-19 16:24   ` [PATCH 1/7] xfs: create a helper to convert rtextents to rtblocks Darrick J. Wong
2023-10-19 16:24   ` [PATCH 2/7] xfs: create a helper to compute leftovers of realtime extents Darrick J. Wong
2023-10-19 16:25   ` [PATCH 3/7] xfs: create a helper to convert extlen to rtextlen Darrick J. Wong
2023-10-19 16:25   ` [PATCH 4/7] xfs: create helpers to convert rt block numbers to rt extent numbers Darrick J. Wong
2023-10-19 16:25   ` [PATCH 5/7] xfs: convert do_div calls to xfs_rtb_to_rtx helper calls Darrick J. Wong
2023-10-19 16:25   ` [PATCH 6/7] xfs: create rt extent rounding helpers for realtime extent blocks Darrick J. Wong
2023-10-19 16:26   ` [PATCH 7/7] xfs: use shifting and masking when converting rt extents, if possible Darrick J. Wong
2023-10-19 16:20 ` [PATCHSET v1.1 0/5] xfs: refactor rtbitmap/summary macros Darrick J. Wong
2023-10-19 16:26   ` [PATCH 1/5] xfs: convert the rtbitmap block and bit macros to static inline functions Darrick J. Wong
2023-10-19 16:26   ` [PATCH 2/5] xfs: remove XFS_BLOCKWSIZE and XFS_BLOCKWMASK macros Darrick J. Wong
2023-10-19 16:27   ` [PATCH 3/5] xfs: convert open-coded xfs_rtword_t pointer accesses to helper Darrick J. Wong
2023-10-19 16:27   ` [PATCH 4/5] xfs: convert rt summary macros to helpers Darrick J. Wong
2023-10-19 16:27   ` [PATCH 5/5] xfs: create helpers for rtbitmap block/wordcount computations Darrick J. Wong
2023-10-19 16:21 ` [PATCHSET v1.2 0/4] xfs: refactor rtbitmap/summary accessors Darrick J. Wong
2023-10-19 16:27   ` [PATCH 1/4] xfs: create a helper to handle logging parts of rt bitmap/summary blocks Darrick J. Wong
2023-10-20  6:20     ` Christoph Hellwig
2023-10-20 15:20       ` Darrick J. Wong
2023-10-19 16:28   ` [PATCH 2/4] xfs: use accessor functions for bitmap words Darrick J. Wong
2023-10-19 16:28   ` [PATCH 3/4] xfs: create helpers for rtsummary block/wordcount computations Darrick J. Wong
2023-10-19 16:28   ` [PATCH 4/4] xfs: use accessor functions for summary info words Darrick J. Wong
2023-10-19 16:21 ` [PATCHSET v2.3 0/9] xfs: CPU usage optimizations for realtime allocator Darrick J. Wong
2023-10-19 16:28   ` [PATCH 1/9] xfs: consolidate realtime allocation arguments Darrick J. Wong
2023-10-19 16:29   ` [PATCH 2/9] xfs: cache last bitmap block in realtime allocator Darrick J. Wong
2023-10-19 16:29   ` [PATCH 3/9] xfs: simplify xfs_rtbuf_get calling conventions Darrick J. Wong
2023-10-19 16:29   ` [PATCH 4/9] xfs: simplify rt bitmap/summary block accessor functions Darrick J. Wong
2023-10-19 16:29   ` [PATCH 5/9] xfs: invert the realtime summary cache Darrick J. Wong
2023-10-19 16:30   ` [PATCH 6/9] xfs: return maximum free size from xfs_rtany_summary() Darrick J. Wong
2023-10-19 16:30   ` [PATCH 7/9] xfs: limit maxlen based on available space in xfs_rtallocate_extent_near() Darrick J. Wong
2023-10-19 16:30   ` Darrick J. Wong [this message]
2023-10-19 16:30   ` [PATCH 9/9] xfs: don't look for end of extent further than necessary " Darrick J. Wong
  -- strict thread matches above, loose matches on Subject: below --
2023-10-19  0:00 [PATCHSET v2.2 0/9] xfs: CPU usage optimizations for realtime allocator Darrick J. Wong
2023-10-19  0:01 ` [PATCH 8/9] xfs: don't try redundant allocations in xfs_rtallocate_extent_near() Darrick J. Wong

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=169773211847.225862.10218692865813765251.stgit@frogsfrogsfrogs \
    --to=djwong@kernel.org \
    --cc=hch@lst.de \
    --cc=linux-xfs@vger.kernel.org \
    --cc=osandov@fb.com \
    --cc=osandov@osandov.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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.