From: Brian Foster <bfoster@redhat.com>
To: Christoph Hellwig <hch@lst.de>
Cc: linux-xfs@vger.kernel.org, linux-fsdevel@vger.kernel.org,
linux-mm@kvack.org
Subject: Re: [PATCH 03/18] xfs: simplify xfs_bmap_punch_delalloc_range
Date: Wed, 30 May 2018 09:35:39 -0400 [thread overview]
Message-ID: <20180530133538.GC112411@bfoster.bfoster> (raw)
In-Reply-To: <20180530100013.31358-4-hch@lst.de>
On Wed, May 30, 2018 at 11:59:58AM +0200, Christoph Hellwig wrote:
> Instead of using xfs_bmapi_read to find delalloc extents and then punch
> them out using xfs_bunmapi, opencode the loop to iterate over the extents
> and call xfs_bmap_del_extent_delay directly. This both simplifies the
> code and reduces the number of extent tree lookups required.
>
> Signed-off-by: Christoph Hellwig <hch@lst.de>
> ---
> fs/xfs/xfs_bmap_util.c | 84 ++++++++++++++----------------------------
> 1 file changed, 28 insertions(+), 56 deletions(-)
>
> diff --git a/fs/xfs/xfs_bmap_util.c b/fs/xfs/xfs_bmap_util.c
> index 06badcbadeb4..f2b87873612d 100644
> --- a/fs/xfs/xfs_bmap_util.c
> +++ b/fs/xfs/xfs_bmap_util.c
> @@ -695,12 +695,10 @@ xfs_getbmap(
> }
>
> /*
> - * dead simple method of punching delalyed allocation blocks from a range in
> - * the inode. Walks a block at a time so will be slow, but is only executed in
> - * rare error cases so the overhead is not critical. This will always punch out
> - * both the start and end blocks, even if the ranges only partially overlap
> - * them, so it is up to the caller to ensure that partial blocks are not
> - * passed in.
> + * Dead simple method of punching delalyed allocation blocks from a range in
> + * the inode. This will always punch out both the start and end blocks, even
> + * if the ranges only partially overlap them, so it is up to the caller to
> + * ensure that partial blocks are not passed in.
> */
> int
> xfs_bmap_punch_delalloc_range(
> @@ -708,63 +706,37 @@ xfs_bmap_punch_delalloc_range(
> xfs_fileoff_t start_fsb,
> xfs_fileoff_t length)
> {
...
> + if (!xfs_iext_lookup_extent_before(ip, ifp, &end_fsb, &icur, &got))
> + return 0;
>
> - /*
> - * Note: while we initialise the firstblock/dfops pair, they
> - * should never be used because blocks should never be
> - * allocated or freed for a delalloc extent and hence we need
> - * don't cancel or finish them after the xfs_bunmapi() call.
> - */
> - xfs_defer_init(&dfops, &firstblock);
> - error = xfs_bunmapi(NULL, ip, start_fsb, 1, 0, 1, &firstblock,
> - &dfops, &done);
> - if (error)
> - break;
> + while (got.br_startoff + got.br_blockcount > start_fsb) {
> + del = got;
> + xfs_trim_extent(&del, start_fsb, length);
>
> - ASSERT(!xfs_defer_has_unfinished_work(&dfops));
> -next_block:
> - start_fsb++;
> - remaining--;
> - } while(remaining > 0);
> + if (del.br_blockcount && isnullstartblock(del.br_startblock)) {
I think there's subtle behavior here that warrants a comment (and
describes the somewhat funky logic). E.g., something like:
/*
* got might point to the extent after del in some cases. The next
* iteration will detect this and step back to the previous extent.
*/
Alternatively, I find separating the if/else a bit more readable (see
the appended hunk). But otherwise looks fine:
Reviewed-by: Brian Foster <bfoster@redhat.com>
> + error = xfs_bmap_del_extent_delay(ip, XFS_DATA_FORK,
> + &icur, &got, &del);
> + if (error || !xfs_iext_get_extent(ifp, &icur, &got))
> + break;
> + } else {
> + if (!xfs_iext_prev_extent(ifp, &icur, &got))
> + break;
> + }
> + }
>
> return error;
> }
> --
> 2.17.0
--- 8< ---
diff --git a/fs/xfs/xfs_bmap_util.c b/fs/xfs/xfs_bmap_util.c
index f2b87873612d..0070b877ed94 100644
--- a/fs/xfs/xfs_bmap_util.c
+++ b/fs/xfs/xfs_bmap_util.c
@@ -727,15 +727,22 @@ xfs_bmap_punch_delalloc_range(
del = got;
xfs_trim_extent(&del, start_fsb, length);
- if (del.br_blockcount && isnullstartblock(del.br_startblock)) {
- error = xfs_bmap_del_extent_delay(ip, XFS_DATA_FORK,
- &icur, &got, &del);
- if (error || !xfs_iext_get_extent(ifp, &icur, &got))
- break;
- } else {
+ /*
+ * A delete can push the cursor forward. Step back to the
+ * previous extent on non-delalloc or extents outside the
+ * target range.
+ */
+ if (!del.br_blockcount ||
+ !isnullstartblock(del.br_startblock)) {
if (!xfs_iext_prev_extent(ifp, &icur, &got))
break;
+ continue;
}
+
+ error = xfs_bmap_del_extent_delay(ip, XFS_DATA_FORK, &icur,
+ &got, &del);
+ if (error || !xfs_iext_get_extent(ifp, &icur, &got))
+ break;
}
return error;
next prev parent reply other threads:[~2018-05-30 13:35 UTC|newest]
Thread overview: 61+ messages / expand[flat|nested] mbox.gz Atom feed top
2018-05-30 9:59 buffered writes without buffer heads in xfs and iomap v4 Christoph Hellwig
2018-05-30 9:59 ` [PATCH 01/18] fs: factor out a __generic_write_end helper Christoph Hellwig
2018-05-30 13:34 ` Brian Foster
2018-05-30 16:26 ` Darrick J. Wong
2018-05-30 9:59 ` [PATCH 02/18] iomap: add initial support for writes without buffer heads Christoph Hellwig
2018-05-30 13:34 ` Brian Foster
2018-05-30 16:33 ` Darrick J. Wong
2018-05-30 9:59 ` [PATCH 03/18] xfs: simplify xfs_bmap_punch_delalloc_range Christoph Hellwig
2018-05-30 13:35 ` Brian Foster [this message]
2018-05-30 16:55 ` Darrick J. Wong
2018-05-30 9:59 ` [PATCH 04/18] xfs: simplify xfs_aops_discard_page Christoph Hellwig
2018-05-30 13:35 ` Brian Foster
2018-05-30 16:40 ` Darrick J. Wong
2018-05-30 10:00 ` [PATCH 05/18] xfs: move locking into xfs_bmap_punch_delalloc_range Christoph Hellwig
2018-05-30 13:35 ` Brian Foster
2018-05-30 16:58 ` Darrick J. Wong
2018-05-30 17:40 ` Brian Foster
2018-05-30 10:00 ` [PATCH 06/18] xfs: make xfs_writepage_map extent map centric Christoph Hellwig
2018-05-30 17:19 ` Darrick J. Wong
2018-05-30 17:39 ` Brian Foster
2018-05-31 16:11 ` Christoph Hellwig
2018-05-31 16:12 ` Christoph Hellwig
2018-05-30 10:00 ` [PATCH 07/18] xfs: remove the now unused XFS_BMAPI_IGSTATE flag Christoph Hellwig
2018-05-30 17:19 ` Darrick J. Wong
2018-05-31 13:46 ` Brian Foster
2018-05-31 16:11 ` Christoph Hellwig
2018-05-30 10:00 ` [PATCH 08/18] xfs: remove xfs_reflink_find_cow_mapping Christoph Hellwig
2018-05-30 17:23 ` Darrick J. Wong
2018-05-31 13:46 ` Brian Foster
2018-05-30 10:00 ` [PATCH 09/18] xfs: remove xfs_reflink_trim_irec_to_next_cow Christoph Hellwig
2018-05-30 17:27 ` Darrick J. Wong
2018-05-31 13:46 ` Brian Foster
2018-05-30 10:00 ` [PATCH 10/18] xfs: simplify xfs_map_blocks by using xfs_iext_lookup_extent directly Christoph Hellwig
2018-05-30 17:32 ` Darrick J. Wong
2018-05-31 13:46 ` Brian Foster
2018-05-30 10:00 ` [PATCH 11/18] xfs: don't clear imap_valid for a non-uptodate buffers Christoph Hellwig
2018-05-30 17:34 ` Darrick J. Wong
2018-05-31 13:46 ` Brian Foster
2018-05-30 10:00 ` [PATCH 12/18] xfs: remove the imap_valid flag Christoph Hellwig
2018-05-30 17:44 ` Darrick J. Wong
2018-05-31 13:47 ` Brian Foster
2018-05-30 10:00 ` [PATCH 13/18] xfs: don't look at buffer heads in xfs_add_to_ioend Christoph Hellwig
2018-05-30 17:55 ` Darrick J. Wong
2018-05-31 7:03 ` Christoph Hellwig
2018-05-31 13:49 ` Brian Foster
2018-05-30 10:00 ` [PATCH 14/18] xfs: move all writeback buffer_head manipulation into xfs_map_at_offset Christoph Hellwig
2018-05-30 17:56 ` Darrick J. Wong
2018-05-31 13:49 ` Brian Foster
2018-05-30 10:00 ` [PATCH 15/18] xfs: remove xfs_start_page_writeback Christoph Hellwig
2018-05-30 17:57 ` Darrick J. Wong
2018-05-31 13:49 ` Brian Foster
2018-05-30 10:00 ` [PATCH 16/18] xfs: refactor the tail of xfs_writepage_map Christoph Hellwig
2018-05-30 18:00 ` Darrick J. Wong
2018-05-31 13:49 ` Brian Foster
2018-05-30 10:00 ` [PATCH 17/18] xfs: do not set the page uptodate in xfs_writepage_map Christoph Hellwig
2018-05-30 18:08 ` Darrick J. Wong
2018-05-31 7:04 ` Christoph Hellwig
2018-05-31 13:49 ` Brian Foster
2018-05-30 10:00 ` [PATCH 18/18] xfs: allow writeback on pages without buffer heads Christoph Hellwig
2018-05-30 18:08 ` Darrick J. Wong
2018-05-31 13:49 ` Brian Foster
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=20180530133538.GC112411@bfoster.bfoster \
--to=bfoster@redhat.com \
--cc=hch@lst.de \
--cc=linux-fsdevel@vger.kernel.org \
--cc=linux-mm@kvack.org \
--cc=linux-xfs@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;
as well as URLs for NNTP newsgroup(s).