All of lore.kernel.org
 help / color / mirror / Atom feed
From: Christoph Hellwig <hch@lst.de>
Cc: linux-xfs@vger.kernel.org, "Darrick J. Wong" <darrick.wong@oracle.com>
Subject: [PATCH 09/26] xfs: allow unwritten extents in the CoW fork
Date: Mon, 27 Mar 2017 10:38:20 +0200	[thread overview]
Message-ID: <20170327083837.11027-10-hch@lst.de> (raw)
In-Reply-To: <20170327083837.11027-1-hch@lst.de>

From: "Darrick J. Wong" <darrick.wong@oracle.com>

commit 05a630d76bd3f39baf0eecfa305bed2820796dee upstream.

In the data fork, we only allow extents to perform the following state
transitions:

delay -> real <-> unwritten

There's no way to move directly from a delalloc reservation to an
/unwritten/ allocated extent.  However, for the CoW fork we want to be
able to do the following to each extent:

delalloc -> unwritten -> written -> remapped to data fork

This will help us to avoid a race in the speculative CoW preallocation
code between a first thread that is allocating a CoW extent and a second
thread that is remapping part of a file after a write.  In order to do
this, however, we need two things: first, we have to be able to
transition from da to unwritten, and second the function that converts
between real and unwritten has to be made aware of the cow fork.  Do
both of those things.

Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
---
 fs/xfs/libxfs/xfs_bmap.c | 80 ++++++++++++++++++++++++++++++------------------
 1 file changed, 50 insertions(+), 30 deletions(-)

diff --git a/fs/xfs/libxfs/xfs_bmap.c b/fs/xfs/libxfs/xfs_bmap.c
index d79c66d1215d..5b626a4f1d13 100644
--- a/fs/xfs/libxfs/xfs_bmap.c
+++ b/fs/xfs/libxfs/xfs_bmap.c
@@ -1861,6 +1861,7 @@ xfs_bmap_add_extent_delay_real(
 		 */
 		trace_xfs_bmap_pre_update(bma->ip, bma->idx, state, _THIS_IP_);
 		xfs_bmbt_set_startblock(ep, new->br_startblock);
+		xfs_bmbt_set_state(ep, new->br_state);
 		trace_xfs_bmap_post_update(bma->ip, bma->idx, state, _THIS_IP_);
 
 		(*nextents)++;
@@ -2199,6 +2200,7 @@ STATIC int				/* error */
 xfs_bmap_add_extent_unwritten_real(
 	struct xfs_trans	*tp,
 	xfs_inode_t		*ip,	/* incore inode pointer */
+	int			whichfork,
 	xfs_extnum_t		*idx,	/* extent number to update/insert */
 	xfs_btree_cur_t		**curp,	/* if *curp is null, not a btree */
 	xfs_bmbt_irec_t		*new,	/* new data to add to file extents */
@@ -2218,12 +2220,14 @@ xfs_bmap_add_extent_unwritten_real(
 					/* left is 0, right is 1, prev is 2 */
 	int			rval=0;	/* return value (logging flags) */
 	int			state = 0;/* state bits, accessed thru macros */
-	struct xfs_mount	*mp = tp->t_mountp;
+	struct xfs_mount	*mp = ip->i_mount;
 
 	*logflagsp = 0;
 
 	cur = *curp;
-	ifp = XFS_IFORK_PTR(ip, XFS_DATA_FORK);
+	ifp = XFS_IFORK_PTR(ip, whichfork);
+	if (whichfork == XFS_COW_FORK)
+		state |= BMAP_COWFORK;
 
 	ASSERT(*idx >= 0);
 	ASSERT(*idx <= xfs_iext_count(ifp));
@@ -2282,7 +2286,7 @@ xfs_bmap_add_extent_unwritten_real(
 	 * Don't set contiguous if the combined extent would be too large.
 	 * Also check for all-three-contiguous being too large.
 	 */
-	if (*idx < xfs_iext_count(&ip->i_df) - 1) {
+	if (*idx < xfs_iext_count(ifp) - 1) {
 		state |= BMAP_RIGHT_VALID;
 		xfs_bmbt_get_all(xfs_iext_get_ext(ifp, *idx + 1), &RIGHT);
 		if (isnullstartblock(RIGHT.br_startblock))
@@ -2322,7 +2326,8 @@ xfs_bmap_add_extent_unwritten_real(
 		trace_xfs_bmap_post_update(ip, *idx, state, _THIS_IP_);
 
 		xfs_iext_remove(ip, *idx + 1, 2, state);
-		ip->i_d.di_nextents -= 2;
+		XFS_IFORK_NEXT_SET(ip, whichfork,
+				XFS_IFORK_NEXTENTS(ip, whichfork) - 2);
 		if (cur == NULL)
 			rval = XFS_ILOG_CORE | XFS_ILOG_DEXT;
 		else {
@@ -2365,7 +2370,8 @@ xfs_bmap_add_extent_unwritten_real(
 		trace_xfs_bmap_post_update(ip, *idx, state, _THIS_IP_);
 
 		xfs_iext_remove(ip, *idx + 1, 1, state);
-		ip->i_d.di_nextents--;
+		XFS_IFORK_NEXT_SET(ip, whichfork,
+				XFS_IFORK_NEXTENTS(ip, whichfork) - 1);
 		if (cur == NULL)
 			rval = XFS_ILOG_CORE | XFS_ILOG_DEXT;
 		else {
@@ -2400,7 +2406,8 @@ xfs_bmap_add_extent_unwritten_real(
 		xfs_bmbt_set_state(ep, newext);
 		trace_xfs_bmap_post_update(ip, *idx, state, _THIS_IP_);
 		xfs_iext_remove(ip, *idx + 1, 1, state);
-		ip->i_d.di_nextents--;
+		XFS_IFORK_NEXT_SET(ip, whichfork,
+				XFS_IFORK_NEXTENTS(ip, whichfork) - 1);
 		if (cur == NULL)
 			rval = XFS_ILOG_CORE | XFS_ILOG_DEXT;
 		else {
@@ -2512,7 +2519,8 @@ xfs_bmap_add_extent_unwritten_real(
 		trace_xfs_bmap_post_update(ip, *idx, state, _THIS_IP_);
 
 		xfs_iext_insert(ip, *idx, 1, new, state);
-		ip->i_d.di_nextents++;
+		XFS_IFORK_NEXT_SET(ip, whichfork,
+				XFS_IFORK_NEXTENTS(ip, whichfork) + 1);
 		if (cur == NULL)
 			rval = XFS_ILOG_CORE | XFS_ILOG_DEXT;
 		else {
@@ -2590,7 +2598,8 @@ xfs_bmap_add_extent_unwritten_real(
 		++*idx;
 		xfs_iext_insert(ip, *idx, 1, new, state);
 
-		ip->i_d.di_nextents++;
+		XFS_IFORK_NEXT_SET(ip, whichfork,
+				XFS_IFORK_NEXTENTS(ip, whichfork) + 1);
 		if (cur == NULL)
 			rval = XFS_ILOG_CORE | XFS_ILOG_DEXT;
 		else {
@@ -2638,7 +2647,8 @@ xfs_bmap_add_extent_unwritten_real(
 		++*idx;
 		xfs_iext_insert(ip, *idx, 2, &r[0], state);
 
-		ip->i_d.di_nextents += 2;
+		XFS_IFORK_NEXT_SET(ip, whichfork,
+				XFS_IFORK_NEXTENTS(ip, whichfork) + 2);
 		if (cur == NULL)
 			rval = XFS_ILOG_CORE | XFS_ILOG_DEXT;
 		else {
@@ -2692,17 +2702,17 @@ xfs_bmap_add_extent_unwritten_real(
 	}
 
 	/* update reverse mappings */
-	error = xfs_rmap_convert_extent(mp, dfops, ip, XFS_DATA_FORK, new);
+	error = xfs_rmap_convert_extent(mp, dfops, ip, whichfork, new);
 	if (error)
 		goto done;
 
 	/* convert to a btree if necessary */
-	if (xfs_bmap_needs_btree(ip, XFS_DATA_FORK)) {
+	if (xfs_bmap_needs_btree(ip, whichfork)) {
 		int	tmp_logflags;	/* partial log flag return val */
 
 		ASSERT(cur == NULL);
 		error = xfs_bmap_extents_to_btree(tp, ip, first, dfops, &cur,
-				0, &tmp_logflags, XFS_DATA_FORK);
+				0, &tmp_logflags, whichfork);
 		*logflagsp |= tmp_logflags;
 		if (error)
 			goto done;
@@ -2714,7 +2724,7 @@ xfs_bmap_add_extent_unwritten_real(
 		*curp = cur;
 	}
 
-	xfs_bmap_check_leaf_extents(*curp, ip, XFS_DATA_FORK);
+	xfs_bmap_check_leaf_extents(*curp, ip, whichfork);
 done:
 	*logflagsp |= rval;
 	return error;
@@ -4365,10 +4375,16 @@ xfs_bmapi_allocate(
 	bma->got.br_state = XFS_EXT_NORM;
 
 	/*
-	 * A wasdelay extent has been initialized, so shouldn't be flagged
-	 * as unwritten.
+	 * In the data fork, a wasdelay extent has been initialized, so
+	 * shouldn't be flagged as unwritten.
+	 *
+	 * For the cow fork, however, we convert delalloc reservations
+	 * (extents allocated for speculative preallocation) to
+	 * allocated unwritten extents, and only convert the unwritten
+	 * extents to real extents when we're about to write the data.
 	 */
-	if (!bma->wasdel && (bma->flags & XFS_BMAPI_PREALLOC) &&
+	if ((!bma->wasdel || (bma->flags & XFS_BMAPI_COWFORK)) &&
+	    (bma->flags & XFS_BMAPI_PREALLOC) &&
 	    xfs_sb_version_hasextflgbit(&mp->m_sb))
 		bma->got.br_state = XFS_EXT_UNWRITTEN;
 
@@ -4419,8 +4435,6 @@ xfs_bmapi_convert_unwritten(
 			(XFS_BMAPI_PREALLOC | XFS_BMAPI_CONVERT))
 		return 0;
 
-	ASSERT(whichfork != XFS_COW_FORK);
-
 	/*
 	 * Modify (by adding) the state flag, if writing.
 	 */
@@ -4445,8 +4459,8 @@ xfs_bmapi_convert_unwritten(
 			return error;
 	}
 
-	error = xfs_bmap_add_extent_unwritten_real(bma->tp, bma->ip, &bma->idx,
-			&bma->cur, mval, bma->firstblock, bma->dfops,
+	error = xfs_bmap_add_extent_unwritten_real(bma->tp, bma->ip, whichfork,
+			&bma->idx, &bma->cur, mval, bma->firstblock, bma->dfops,
 			&tmp_logflags);
 	/*
 	 * Log the inode core unconditionally in the unwritten extent conversion
@@ -4455,8 +4469,12 @@ xfs_bmapi_convert_unwritten(
 	 * in the transaction for the sake of fsync(), even if nothing has
 	 * changed, because fsync() will not force the log for this transaction
 	 * unless it sees the inode pinned.
+	 *
+	 * Note: If we're only converting cow fork extents, there aren't
+	 * any on-disk updates to make, so we don't need to log anything.
 	 */
-	bma->logflags |= tmp_logflags | XFS_ILOG_CORE;
+	if (whichfork != XFS_COW_FORK)
+		bma->logflags |= tmp_logflags | XFS_ILOG_CORE;
 	if (error)
 		return error;
 
@@ -4530,15 +4548,15 @@ xfs_bmapi_write(
 	ASSERT(*nmap >= 1);
 	ASSERT(*nmap <= XFS_BMAP_MAX_NMAP);
 	ASSERT(!(flags & XFS_BMAPI_IGSTATE));
-	ASSERT(tp != NULL);
+	ASSERT(tp != NULL ||
+	       (flags & (XFS_BMAPI_CONVERT | XFS_BMAPI_COWFORK)) ==
+			(XFS_BMAPI_CONVERT | XFS_BMAPI_COWFORK));
 	ASSERT(len > 0);
 	ASSERT(XFS_IFORK_FORMAT(ip, whichfork) != XFS_DINODE_FMT_LOCAL);
 	ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL));
 	ASSERT(!(flags & XFS_BMAPI_REMAP) || whichfork == XFS_DATA_FORK);
 	ASSERT(!(flags & XFS_BMAPI_PREALLOC) || !(flags & XFS_BMAPI_REMAP));
 	ASSERT(!(flags & XFS_BMAPI_CONVERT) || !(flags & XFS_BMAPI_REMAP));
-	ASSERT(!(flags & XFS_BMAPI_PREALLOC) || whichfork != XFS_COW_FORK);
-	ASSERT(!(flags & XFS_BMAPI_CONVERT) || whichfork != XFS_COW_FORK);
 
 	/* zeroing is for currently only for data extents, not metadata */
 	ASSERT((flags & (XFS_BMAPI_METADATA | XFS_BMAPI_ZERO)) !=
@@ -5553,8 +5571,8 @@ __xfs_bunmapi(
 			}
 			del.br_state = XFS_EXT_UNWRITTEN;
 			error = xfs_bmap_add_extent_unwritten_real(tp, ip,
-					&lastx, &cur, &del, firstblock, dfops,
-					&logflags);
+					whichfork, &lastx, &cur, &del,
+					firstblock, dfops, &logflags);
 			if (error)
 				goto error0;
 			goto nodelete;
@@ -5607,8 +5625,9 @@ __xfs_bunmapi(
 				prev.br_state = XFS_EXT_UNWRITTEN;
 				lastx--;
 				error = xfs_bmap_add_extent_unwritten_real(tp,
-						ip, &lastx, &cur, &prev,
-						firstblock, dfops, &logflags);
+						ip, whichfork, &lastx, &cur,
+						&prev, firstblock, dfops,
+						&logflags);
 				if (error)
 					goto error0;
 				goto nodelete;
@@ -5616,8 +5635,9 @@ __xfs_bunmapi(
 				ASSERT(del.br_state == XFS_EXT_NORM);
 				del.br_state = XFS_EXT_UNWRITTEN;
 				error = xfs_bmap_add_extent_unwritten_real(tp,
-						ip, &lastx, &cur, &del,
-						firstblock, dfops, &logflags);
+						ip, whichfork, &lastx, &cur,
+						&del, firstblock, dfops,
+						&logflags);
 				if (error)
 					goto error0;
 				goto nodelete;
-- 
2.11.0


  parent reply	other threads:[~2017-03-27  8:39 UTC|newest]

Thread overview: 28+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2017-03-27  8:38 Christoph Hellwig
2017-03-27  8:38 ` [PATCH 01/26] xfs: pull up iolock from xfs_free_eofblocks() Christoph Hellwig
2017-03-27  8:38 ` [PATCH 02/26] xfs: sync eofblocks scans under iolock are livelock prone Christoph Hellwig
2017-03-27  8:38 ` [PATCH 03/26] xfs: fix eofblocks race with file extending async dio writes Christoph Hellwig
2017-03-27  8:38 ` [PATCH 04/26] xfs: fix toctou race when locking an inode to access the data map Christoph Hellwig
2017-03-27  8:38 ` [PATCH 05/26] xfs: fail _dir_open when readahead fails Christoph Hellwig
2017-03-27  8:38 ` [PATCH 06/26] xfs: filter out obviously bad btree pointers Christoph Hellwig
2017-03-27  8:38 ` [PATCH 07/26] xfs: check for obviously bad level values in the bmbt root Christoph Hellwig
2017-03-27  8:38 ` [PATCH 08/26] xfs: verify free block header fields Christoph Hellwig
2017-03-27  8:38 ` Christoph Hellwig [this message]
2017-03-27  8:38 ` [PATCH 10/26] xfs: mark speculative prealloc CoW fork extents unwritten Christoph Hellwig
2017-03-27  8:38 ` [PATCH 11/26] xfs: reset b_first_retry_time when clear the retry status of xfs_buf_t Christoph Hellwig
2017-03-27  8:38 ` [PATCH 12/26] xfs: reject all unaligned direct writes to reflinked files Christoph Hellwig
2017-03-27  8:38 ` [PATCH 13/26] xfs: update ctime and mtime on clone destinatation inodes Christoph Hellwig
2017-03-27  8:38 ` [PATCH 14/26] xfs: correct null checks and error processing in xfs_initialize_perag Christoph Hellwig
2017-03-27  8:38 ` [PATCH 15/26] xfs: don't fail xfs_extent_busy allocation Christoph Hellwig
2017-03-27  8:38 ` [PATCH 16/26] xfs: handle indlen shortage on delalloc extent merge Christoph Hellwig
2017-03-27  8:38 ` [PATCH 17/26] xfs: split indlen reservations fairly when under reserved Christoph Hellwig
2017-03-27  8:38 ` [PATCH 18/26] xfs: fix uninitialized variable in _reflink_convert_cow Christoph Hellwig
2017-03-27  8:38 ` [PATCH 19/26] xfs: don't reserve blocks for right shift transactions Christoph Hellwig
2017-03-27  8:38 ` [PATCH 20/26] xfs: Use xfs_icluster_size_fsb() to calculate inode chunk alignment Christoph Hellwig
2017-03-27  8:38 ` [PATCH 21/26] xfs: tune down agno asserts in the bmap code Christoph Hellwig
2017-03-27  8:38 ` [PATCH 22/26] xfs: only reclaim unwritten COW extents periodically Christoph Hellwig
2017-03-27  8:38 ` [PATCH 23/26] xfs: fix and streamline error handling in xfs_end_io Christoph Hellwig
2017-03-27  8:38 ` [PATCH 24/26] xfs: Use xfs_icluster_size_fsb() to calculate inode alignment mask Christoph Hellwig
2017-03-27  8:38 ` [PATCH 25/26] xfs: use iomap new flag for newly allocated delalloc blocks Christoph Hellwig
2017-03-27  8:38 ` [PATCH 26/26] xfs: try any AG when allocating the first btree block when reflinking Christoph Hellwig
2017-04-01  6:34 4.10-stable updates for XFS Christoph Hellwig
2017-04-01  6:34 ` [PATCH 09/26] xfs: allow unwritten extents in the CoW fork Christoph Hellwig

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=20170327083837.11027-10-hch@lst.de \
    --to=hch@lst.de \
    --cc=darrick.wong@oracle.com \
    --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 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.