public inbox for linux-xfs@vger.kernel.org
 help / color / mirror / Atom feed
From: Dave Chinner <david@fromorbit.com>
To: xfs@oss.sgi.com
Subject: [PATCH 08/27] xfs: Introduce an ordered buffer item
Date: Wed, 12 Jun 2013 20:22:28 +1000	[thread overview]
Message-ID: <1371032567-21772-9-git-send-email-david@fromorbit.com> (raw)
In-Reply-To: <1371032567-21772-1-git-send-email-david@fromorbit.com>

If we have a buffer that we have modified but we do not wish to
physically log in a transaction (e.g. we've logged a logical
change), we still need to ensure that transactional integrity is
maintained. Hence we must not move the tail of the log past the
transaction that the buffer is associated with before the buffer is
written to disk.

This means these special buffers still need to be included in the
transaction and added to the AIL just like a normal buffer, but we
do not want the modifications to the buffer written into the
transaction. IOWs, what we want is an "ordered buffer" that
maintains the same transactional life cycle as a physically logged
buffer, just without the transcribing of the modifications to the
log.

Hence we need to flag the buffer as an "ordered buffer" to avoid
including it in vector size calculations or formatting during the
transaction. Once the transaction is committed, the buffer appears
for all intents to be the same as a physically logged buffer as it
transitions through the log and AIL.

Relogging will also work just fine for such an ordered buffer - the
logical transaction will be replayed before the subsequent
modifications that relog the buffer, so everything will be
reconstructed correctly by recovery.

Signed-off-by: Dave Chinner <david@fromorbit.com>
---
 fs/xfs/xfs_buf_item.c  |   75 +++++++++++++++++++++++++++++++-----------------
 fs/xfs/xfs_buf_item.h  |    4 ++-
 fs/xfs/xfs_trace.h     |    4 +++
 fs/xfs/xfs_trans.h     |    1 +
 fs/xfs/xfs_trans_buf.c |   34 ++++++++++++++++++++--
 5 files changed, 87 insertions(+), 31 deletions(-)

diff --git a/fs/xfs/xfs_buf_item.c b/fs/xfs/xfs_buf_item.c
index 4ec4317..61f6876 100644
--- a/fs/xfs/xfs_buf_item.c
+++ b/fs/xfs/xfs_buf_item.c
@@ -140,6 +140,16 @@ xfs_buf_item_size(
 
 	ASSERT(bip->bli_flags & XFS_BLI_LOGGED);
 
+	if (bip->bli_flags & XFS_BLI_ORDERED) {
+		/*
+		 * The buffer has been logged just to order it.
+		 * It is not being included in the transaction
+		 * commit, so no vectors are used at all.
+		 */
+		trace_xfs_buf_item_size_ordered(bip);
+		return XFS_LOG_VEC_ORDERED;
+	}
+
 	/*
 	 * the vector count is based on the number of buffer vectors we have
 	 * dirty bits in. This will only be greater than one when we have a
@@ -212,6 +222,7 @@ xfs_buf_item_format_segment(
 		goto out;
 	}
 
+
 	/*
 	 * Fill in an iovec for each set of contiguous chunks.
 	 */
@@ -311,6 +322,16 @@ xfs_buf_item_format(
 		bip->bli_flags &= ~XFS_BLI_INODE_BUF;
 	}
 
+	if ((bip->bli_flags & (XFS_BLI_ORDERED|XFS_BLI_STALE)) ==
+							XFS_BLI_ORDERED) {
+		/*
+		 * The buffer has been logged just to order it.  It is not being
+		 * included in the transaction commit, so don't format it.
+		 */
+		trace_xfs_buf_item_format_ordered(bip);
+		return;
+	}
+
 	for (i = 0; i < bip->bli_format_count; i++) {
 		vecp = xfs_buf_item_format_segment(bip, vecp, offset,
 						&bip->bli_formats[i]);
@@ -340,6 +361,7 @@ xfs_buf_item_pin(
 
 	ASSERT(atomic_read(&bip->bli_refcount) > 0);
 	ASSERT((bip->bli_flags & XFS_BLI_LOGGED) ||
+	       (bip->bli_flags & XFS_BLI_ORDERED) ||
 	       (bip->bli_flags & XFS_BLI_STALE));
 
 	trace_xfs_buf_item_pin(bip);
@@ -512,8 +534,9 @@ xfs_buf_item_unlock(
 {
 	struct xfs_buf_log_item	*bip = BUF_ITEM(lip);
 	struct xfs_buf		*bp = bip->bli_buf;
-	int			aborted, clean, i;
-	uint			hold;
+	bool			clean;
+	bool			aborted;
+	int			flags;
 
 	/* Clear the buffer's association with this transaction. */
 	bp->b_transp = NULL;
@@ -524,23 +547,21 @@ xfs_buf_item_unlock(
 	 * (cancelled) buffers at unpin time, but we'll never go through the
 	 * pin/unpin cycle if we abort inside commit.
 	 */
-	aborted = (lip->li_flags & XFS_LI_ABORTED) != 0;
-
+	aborted = (lip->li_flags & XFS_LI_ABORTED) ? true : false;
 	/*
-	 * Before possibly freeing the buf item, determine if we should
-	 * release the buffer at the end of this routine.
+	 * Before possibly freeing the buf item, copy the per-transaction state
+	 * so we can reference it safely later after clearing it from the
+	 * buffer log item.
 	 */
-	hold = bip->bli_flags & XFS_BLI_HOLD;
-
-	/* Clear the per transaction state. */
-	bip->bli_flags &= ~(XFS_BLI_LOGGED | XFS_BLI_HOLD);
+	flags = bip->bli_flags;
+	bip->bli_flags &= ~(XFS_BLI_LOGGED | XFS_BLI_HOLD | XFS_BLI_ORDERED);
 
 	/*
 	 * If the buf item is marked stale, then don't do anything.  We'll
 	 * unlock the buffer and free the buf item when the buffer is unpinned
 	 * for the last time.
 	 */
-	if (bip->bli_flags & XFS_BLI_STALE) {
+	if (flags & XFS_BLI_STALE) {
 		trace_xfs_buf_item_unlock_stale(bip);
 		ASSERT(bip->__bli_format.blf_flags & XFS_BLF_CANCEL);
 		if (!aborted) {
@@ -557,13 +578,19 @@ xfs_buf_item_unlock(
 	 * be the only reference to the buf item, so we free it anyway
 	 * regardless of whether it is dirty or not. A dirty abort implies a
 	 * shutdown, anyway.
+	 *
+	 * Ordered buffers are dirty but may have no recorded changes, so ensure
+	 * we only release clean items here.
 	 */
-	clean = 1;
-	for (i = 0; i < bip->bli_format_count; i++) {
-		if (!xfs_bitmap_empty(bip->bli_formats[i].blf_data_map,
-			     bip->bli_formats[i].blf_map_size)) {
-			clean = 0;
-			break;
+	clean = (flags & XFS_BLI_DIRTY) ? false : true;
+	if (clean) {
+		int i;
+		for (i = 0; i < bip->bli_format_count; i++) {
+			if (!xfs_bitmap_empty(bip->bli_formats[i].blf_data_map,
+				     bip->bli_formats[i].blf_map_size)) {
+				clean = false;
+				break;
+			}
 		}
 	}
 	if (clean)
@@ -576,7 +603,7 @@ xfs_buf_item_unlock(
 	} else
 		atomic_dec(&bip->bli_refcount);
 
-	if (!hold)
+	if (!(flags & XFS_BLI_HOLD))
 		xfs_buf_relse(bp);
 }
 
@@ -842,12 +869,6 @@ xfs_buf_item_log(
 	struct xfs_buf		*bp = bip->bli_buf;
 
 	/*
-	 * Mark the item as having some dirty data for
-	 * quick reference in xfs_buf_item_dirty.
-	 */
-	bip->bli_flags |= XFS_BLI_DIRTY;
-
-	/*
 	 * walk each buffer segment and mark them dirty appropriately.
 	 */
 	start = 0;
@@ -873,7 +894,7 @@ xfs_buf_item_log(
 
 
 /*
- * Return 1 if the buffer has some data that has been logged (at any
+ * Return 1 if the buffer has been logged or ordered in a transaction (at any
  * point, not just the current transaction) and 0 if not.
  */
 uint
@@ -907,11 +928,11 @@ void
 xfs_buf_item_relse(
 	xfs_buf_t	*bp)
 {
-	xfs_buf_log_item_t	*bip;
+	xfs_buf_log_item_t	*bip = bp->b_fspriv;
 
 	trace_xfs_buf_item_relse(bp, _RET_IP_);
+	ASSERT(!(bip->bli_item.li_flags & XFS_LI_IN_AIL));
 
-	bip = bp->b_fspriv;
 	bp->b_fspriv = bip->bli_item.li_bio_list;
 	if (bp->b_fspriv == NULL)
 		bp->b_iodone = NULL;
diff --git a/fs/xfs/xfs_buf_item.h b/fs/xfs/xfs_buf_item.h
index 2573d2a..0f1c247 100644
--- a/fs/xfs/xfs_buf_item.h
+++ b/fs/xfs/xfs_buf_item.h
@@ -120,6 +120,7 @@ xfs_blft_from_flags(struct xfs_buf_log_format *blf)
 #define	XFS_BLI_INODE_ALLOC_BUF	0x10
 #define XFS_BLI_STALE_INODE	0x20
 #define	XFS_BLI_INODE_BUF	0x40
+#define	XFS_BLI_ORDERED		0x80
 
 #define XFS_BLI_FLAGS \
 	{ XFS_BLI_HOLD,		"HOLD" }, \
@@ -128,7 +129,8 @@ xfs_blft_from_flags(struct xfs_buf_log_format *blf)
 	{ XFS_BLI_LOGGED,	"LOGGED" }, \
 	{ XFS_BLI_INODE_ALLOC_BUF, "INODE_ALLOC" }, \
 	{ XFS_BLI_STALE_INODE,	"STALE_INODE" }, \
-	{ XFS_BLI_INODE_BUF,	"INODE_BUF" }
+	{ XFS_BLI_INODE_BUF,	"INODE_BUF" }, \
+	{ XFS_BLI_ORDERED,	"ORDERED" }
 
 
 #ifdef __KERNEL__
diff --git a/fs/xfs/xfs_trace.h b/fs/xfs/xfs_trace.h
index aa4db33..34422b6 100644
--- a/fs/xfs/xfs_trace.h
+++ b/fs/xfs/xfs_trace.h
@@ -486,9 +486,12 @@ DEFINE_EVENT(xfs_buf_item_class, name, \
 	TP_PROTO(struct xfs_buf_log_item *bip), \
 	TP_ARGS(bip))
 DEFINE_BUF_ITEM_EVENT(xfs_buf_item_size);
+DEFINE_BUF_ITEM_EVENT(xfs_buf_item_size_ordered);
 DEFINE_BUF_ITEM_EVENT(xfs_buf_item_size_stale);
 DEFINE_BUF_ITEM_EVENT(xfs_buf_item_format);
+DEFINE_BUF_ITEM_EVENT(xfs_buf_item_format_ordered);
 DEFINE_BUF_ITEM_EVENT(xfs_buf_item_format_stale);
+DEFINE_BUF_ITEM_EVENT(xfs_buf_item_ordered);
 DEFINE_BUF_ITEM_EVENT(xfs_buf_item_pin);
 DEFINE_BUF_ITEM_EVENT(xfs_buf_item_unpin);
 DEFINE_BUF_ITEM_EVENT(xfs_buf_item_unpin_stale);
@@ -508,6 +511,7 @@ DEFINE_BUF_ITEM_EVENT(xfs_trans_bjoin);
 DEFINE_BUF_ITEM_EVENT(xfs_trans_bhold);
 DEFINE_BUF_ITEM_EVENT(xfs_trans_bhold_release);
 DEFINE_BUF_ITEM_EVENT(xfs_trans_binval);
+DEFINE_BUF_ITEM_EVENT(xfs_trans_buf_ordered);
 
 DECLARE_EVENT_CLASS(xfs_lock_class,
 	TP_PROTO(struct xfs_inode *ip, unsigned lock_flags,
diff --git a/fs/xfs/xfs_trans.h b/fs/xfs/xfs_trans.h
index a44dba5..850fe78 100644
--- a/fs/xfs/xfs_trans.h
+++ b/fs/xfs/xfs_trans.h
@@ -503,6 +503,7 @@ void		xfs_trans_bhold_release(xfs_trans_t *, struct xfs_buf *);
 void		xfs_trans_binval(xfs_trans_t *, struct xfs_buf *);
 void		xfs_trans_inode_buf(xfs_trans_t *, struct xfs_buf *);
 void		xfs_trans_stale_inode_buf(xfs_trans_t *, struct xfs_buf *);
+void		xfs_trans_ordered_buf(xfs_trans_t *, struct xfs_buf *);
 void		xfs_trans_dquot_buf(xfs_trans_t *, struct xfs_buf *, uint);
 void		xfs_trans_inode_alloc_buf(xfs_trans_t *, struct xfs_buf *);
 void		xfs_trans_ichgtime(struct xfs_trans *, struct xfs_inode *, int);
diff --git a/fs/xfs/xfs_trans_buf.c b/fs/xfs/xfs_trans_buf.c
index 73a5fa4..aa5a04b 100644
--- a/fs/xfs/xfs_trans_buf.c
+++ b/fs/xfs/xfs_trans_buf.c
@@ -397,7 +397,6 @@ shutdown_abort:
 	return XFS_ERROR(EIO);
 }
 
-
 /*
  * Release the buffer bp which was previously acquired with one of the
  * xfs_trans_... buffer allocation routines if the buffer has not
@@ -603,8 +602,14 @@ xfs_trans_log_buf(xfs_trans_t	*tp,
 
 	tp->t_flags |= XFS_TRANS_DIRTY;
 	bip->bli_item.li_desc->lid_flags |= XFS_LID_DIRTY;
-	bip->bli_flags |= XFS_BLI_LOGGED;
-	xfs_buf_item_log(bip, first, last);
+
+	/*
+	 * If we have an ordered buffer we are not logging any dirty range but
+	 * it still needs to be marked dirty and that it has been logged.
+	 */
+	bip->bli_flags |= XFS_BLI_DIRTY | XFS_BLI_LOGGED;
+	if (!(bip->bli_flags & XFS_BLI_ORDERED))
+		xfs_buf_item_log(bip, first, last);
 }
 
 
@@ -757,6 +762,29 @@ xfs_trans_inode_alloc_buf(
 }
 
 /*
+ * Mark the buffer as ordered for this transaction. This means
+ * that the contents of the buffer are not recorded in the transaction
+ * but it is tracked in the AIL as though it was. This allows us
+ * to record logical changes in transactions rather than the physical
+ * changes we make to the buffer without changing writeback ordering
+ * constraints of metadata buffers.
+ */
+void
+xfs_trans_ordered_buf(
+	struct xfs_trans	*tp,
+	struct xfs_buf		*bp)
+{
+	struct xfs_buf_log_item	*bip = bp->b_fspriv;
+
+	ASSERT(bp->b_transp == tp);
+	ASSERT(bip != NULL);
+	ASSERT(atomic_read(&bip->bli_refcount) > 0);
+
+	bip->bli_flags |= XFS_BLI_ORDERED;
+	trace_xfs_buf_item_ordered(bip);
+}
+
+/*
  * Set the type of the buffer for log recovery so that it can correctly identify
  * and hence attach the correct buffer ops to the buffer after replay.
  */
-- 
1.7.10.4

_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs

  parent reply	other threads:[~2013-06-12 10:23 UTC|newest]

Thread overview: 56+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2013-06-12 10:22 [PATCH 00/27] xfs: current patch queue for 3.11 Dave Chinner
2013-06-12 10:22 ` [PATCH 01/27] xfs: update mount options documentation Dave Chinner
2013-06-13 13:34   ` Eric Sandeen
2013-06-14  0:40     ` Dave Chinner
2013-06-14  0:53       ` Eric Sandeen
2013-06-12 10:22 ` [PATCH 02/27] xfs: add pluging for bulkstat readahead Dave Chinner
2013-06-12 10:22 ` [PATCH 03/27] xfs: plug directory buffer readahead Dave Chinner
2013-06-12 10:22 ` [PATCH 04/27] xfs: don't use speculative prealloc for small files Dave Chinner
2013-06-12 16:10   ` Brian Foster
2013-06-13  0:50     ` Dave Chinner
2013-06-12 10:22 ` [PATCH 05/27] xfs: don't do IO when creating an new inode Dave Chinner
2013-06-12 10:22 ` [PATCH 06/27] xfs: xfs_ifree doesn't need to modify the inode buffer Dave Chinner
2013-06-12 10:22 ` [PATCH 07/27] xfs: Introduce ordered log vector support Dave Chinner
2013-06-12 10:22 ` Dave Chinner [this message]
2013-06-12 10:22 ` [PATCH 09/27] xfs: Inode create log items Dave Chinner
2013-06-12 10:22 ` [PATCH 10/27] xfs: Inode create transaction reservations Dave Chinner
2013-06-12 10:22 ` [PATCH 11/27] xfs: Inode create item recovery Dave Chinner
2013-06-12 10:22 ` [PATCH 12/27] xfs: Use inode create transaction Dave Chinner
2013-06-12 10:22 ` [PATCH 13/27] xfs: remove local fork format handling from xfs_bmapi_write() Dave Chinner
2013-06-12 10:22 ` [PATCH 14/27] xfs: move getdents code into it's own file Dave Chinner
2013-06-12 10:22 ` [PATCH 15/27] xfs: reshuffle dir2 definitions around for userspace Dave Chinner
2013-06-17 16:05   ` Christoph Hellwig
2013-06-18 21:12     ` Dave Chinner
2013-06-18 21:35       ` Dave Chinner
2013-06-12 10:22 ` [PATCH 16/27] xfs: split out attribute listing code into separate file Dave Chinner
2013-06-12 10:22 ` [PATCH 17/27] xfs: split out attribute fork truncation " Dave Chinner
2013-06-12 10:22 ` [PATCH 18/27] xfs: split out xfs inode operations " Dave Chinner
2013-06-12 14:05   ` Christoph Hellwig
2013-06-13  1:14     ` Dave Chinner
2013-06-13  8:00       ` Dave Chinner
2013-06-17 15:56         ` Christoph Hellwig
2013-06-17 18:14           ` Ben Myers
2013-06-18 20:40             ` Dave Chinner
2013-06-18 21:37               ` Ben Myers
2013-06-18 22:02                 ` Dave Chinner
2013-06-12 10:22 ` [PATCH 19/27] xfs: consolidate xfs_vnodeops.c into xfs_inode_ops.c Dave Chinner
2013-06-12 13:59   ` Christoph Hellwig
2013-06-13  1:39     ` Dave Chinner
2013-06-17 16:02       ` Christoph Hellwig
2013-06-18 20:55         ` Dave Chinner
2013-06-12 10:22 ` [PATCH 20/27] xfs: move xfs_getbmap to xfs_extent_ops.c Dave Chinner
2013-06-12 10:22 ` [PATCH 21/27] xfs: introduce xfs_sb.c for sharing with libxfs Dave Chinner
2013-06-12 10:22 ` [PATCH 22/27] xfs: move xfs_trans_reservations to xfs_trans.h Dave Chinner
2013-06-12 10:22 ` [PATCH 23/27] xfs: sync minor header differences needed by userspace Dave Chinner
2013-06-12 10:22 ` [PATCH 24/27] xfs: move xfs_bmap_punch_delalloc() to xfs_aops.c Dave Chinner
2013-06-12 14:06   ` Christoph Hellwig
2013-06-13  1:39     ` Dave Chinner
2013-06-12 10:22 ` [PATCH 25/27] xfs: split out transaction reservation code Dave Chinner
2013-06-12 10:22 ` [PATCH 26/27] xfs: minor cleanups Dave Chinner
2013-06-12 10:22 ` [PATCH 27/27] xfs: fix issues that cause userspace warnings Dave Chinner
2013-06-17 19:32   ` Brian Foster
2013-06-18 21:42     ` Dave Chinner
2013-06-12 13:06 ` [PATCH 00/27] xfs: current patch queue for 3.11 Brian Foster
2013-06-13  1:40   ` Dave Chinner
2013-06-12 14:17 ` Ben Myers
2013-06-13  1:58   ` Dave Chinner

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=1371032567-21772-9-git-send-email-david@fromorbit.com \
    --to=david@fromorbit.com \
    --cc=xfs@oss.sgi.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