[PATCH 00/12] various updates for 2.6.36

[PATCH 00/12] various updates for 2.6.36

[Christoph Hellwig]

This is my current set of pending patches, with the writeback code updates,
various tracing updates, the write_inode speedup and a couple of cleanups.
Some of the patches have already been sent out individually, but this replaces
everything since may last big series on May 31st.  It requires that
patchset ("[PATCH 00/17] pending patches") to be fully applied and Dave's
"[PATCH 0/4, V2] xfs: validate inode numbers in file handles correctly"
from today to apply. 

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

[PATCH 01/12] [PATCH] xfs: avoid synchronous transaction in xfs_fs_write_inode

[Christoph Hellwig]

[-- Attachment #1: xfs-optimize-write_inode --]
[-- Type: text/plain, Size: 5085 bytes --]

We already rely on the fact that the sync code will cause a synchronous
log force later on (currently via xfs_fs_sync_fs -> xfs_quiesce_data ->
xfs_sync_data), so no need to do this here.  This allows us to avoid
a lot of synchronous log forces during sync, which pays of especially
with delayed logging enabled.   Some compilebench numbers that show
this:

xfs (delayed logging, 256k logbufs)
===================================

intial create		  25.94 MB/s	  25.75 MB/s	  25.64 MB/s
create			   8.54 MB/s	   9.12 MB/s	   9.15 MB/s
patch			   2.47 MB/s	   2.47 MB/s	   3.17 MB/s
compile			  29.65 MB/s	  30.51 MB/s	  27.33 MB/s
clean			  90.92 MB/s	  98.83 MB/s	 128.87 MB/s
read tree		  11.90 MB/s	  11.84 MB/s	   8.56 MB/s
read compiled		  28.75 MB/s	  29.96 MB/s	  24.25 MB/s
delete tree		8.39 seconds	8.12 seconds	8.46 seconds
delete compiled		8.35 seconds	8.44 seconds	5.11 seconds
stat tree		6.03 seconds	5.59 seconds	5.19 seconds
stat compiled tree	9.00 seconds	9.52 seconds	8.49 seconds


xfs + write_inode log_force removal
===================================
intial create		  25.87 MB/s	  25.76 MB/s	  25.87 MB/s
create			  15.18 MB/s	  14.80 MB/s	  14.94 MB/s
patch			   3.13 MB/s	   3.14 MB/s	   3.11 MB/s
compile			  36.74 MB/s	  37.17 MB/s	  36.84 MB/s
clean			 226.02 MB/s	 222.58 MB/s	 217.94 MB/s
read tree		  15.14 MB/s	  15.02 MB/s	  15.14 MB/s
read compiled tree	  29.30 MB/s	  29.31 MB/s	  29.32 MB/s
delete tree		6.22 seconds	6.14 seconds	6.15 seconds
delete compiled tree	5.75 seconds	5.92 seconds	5.81 seconds
stat tree		4.60 seconds	4.51 seconds	4.56 seconds
stat compiled tree	4.07 seconds	3.87 seconds	3.96 seconds

In addition to that also remove the delwri inode flush that is unessecary
now that bulkstat is always coherent.

Signed-off-by: Christoph Hellwig <hch@lst.de>

Index: xfs/fs/xfs/linux-2.6/xfs_super.c
===================================================================
--- xfs.orig/fs/xfs/linux-2.6/xfs_super.c	2010-06-21 09:31:43.638254092 +0200
+++ xfs/fs/xfs/linux-2.6/xfs_super.c	2010-06-21 09:37:01.982255907 +0200
@@ -1025,7 +1025,6 @@ xfs_log_inode(
 	 */
 	xfs_trans_ijoin(tp, ip);
 	xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE);
-	xfs_trans_set_sync(tp);
 	error = xfs_trans_commit(tp, 0);
 	xfs_ilock_demote(ip, XFS_ILOCK_EXCL);
 
@@ -1048,20 +1047,11 @@ xfs_fs_write_inode(
 
 	if (wbc->sync_mode == WB_SYNC_ALL) {
 		/*
-		 * Make sure the inode has hit stable storage.  By using the
-		 * log and the fsync transactions we reduce the IOs we have
-		 * to do here from two (log and inode) to just the log.
-		 *
-		 * Note: We still need to do a delwri write of the inode after
-		 * this to flush it to the backing buffer so that bulkstat
-		 * works properly if this is the first time the inode has been
-		 * written.  Because we hold the ilock atomically over the
-		 * transaction commit and the inode flush we are guaranteed
-		 * that the inode is not pinned when it returns. If the flush
-		 * lock is already held, then the inode has already been
-		 * flushed once and we don't need to flush it again.  Hence
-		 * the code will only flush the inode if it isn't already
-		 * being flushed.
+		 * Make sure the inode has made it it into the log.  Instead
+		 * of forcing it all the way to stable storage using a
+		 * synchronous transaction we let the log force inside the
+		 * ->sync_fs call do that for thus, which reduces the number
+		 * of synchronous log foces dramatically.
 		 */
 		xfs_ioend_wait(ip);
 		xfs_ilock(ip, XFS_ILOCK_SHARED);
@@ -1075,27 +1065,29 @@ xfs_fs_write_inode(
 		 * We make this non-blocking if the inode is contended, return
 		 * EAGAIN to indicate to the caller that they did not succeed.
 		 * This prevents the flush path from blocking on inodes inside
-		 * another operation right now, they get caught later by xfs_sync.
+		 * another operation right now, they get caught later by
+		 * xfs_sync.
 		 */
 		if (!xfs_ilock_nowait(ip, XFS_ILOCK_SHARED))
 			goto out;
-	}
 
-	if (xfs_ipincount(ip) || !xfs_iflock_nowait(ip))
-		goto out_unlock;
+		if (xfs_ipincount(ip) || !xfs_iflock_nowait(ip))
+			goto out_unlock;
 
-	/*
-	 * Now we have the flush lock and the inode is not pinned, we can check
-	 * if the inode is really clean as we know that there are no pending
-	 * transaction completions, it is not waiting on the delayed write
-	 * queue and there is no IO in progress.
-	 */
-	if (xfs_inode_clean(ip)) {
-		xfs_ifunlock(ip);
-		error = 0;
-		goto out_unlock;
+		/*
+		 * Now we have the flush lock and the inode is not pinned, we
+		 * can check if the inode is really clean as we know that
+		 * there are no pending transaction completions, it is not
+		 * waiting on the delayed write queue and there is no IO in
+		 * progress.
+		 */
+		if (xfs_inode_clean(ip)) {
+			xfs_ifunlock(ip);
+			error = 0;
+			goto out_unlock;
+		}
+		error = xfs_iflush(ip, 0);
 	}
-	error = xfs_iflush(ip, 0);
 
  out_unlock:
 	xfs_iunlock(ip, XFS_ILOCK_SHARED);

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

[PATCH 02/12] xfs: fix corruption case for block size < page size

[Christoph Hellwig]

[-- Attachment #1: xfs-fix-194 --]
[-- Type: text/plain, Size: 1621 bytes --]

xfstests 194 first truncats a file back and then extends it again by
truncating it to a larger size.  This causes discard_buffer to drop
the mapped, but not the uptodate bit and thus creates something that
xfs_page_state_convert takes for unmapped space created by mmap because
it doesn't check for the dirty bit, which also gets cleared by
discard_buffer and checked by other ->writepage implementations like
block_write_full_page.  Handle this kind of buffers early, and unlike
Eric's first version of the patch simply ASSERT that the buffers is
dirty, given that the mmap write case can't happen anymore since the
introduction of ->page_mkwrite.  The now dead code dealing with that
will be deleted in a follow on patch.

Signed-off-by: Eric Sandeen <sandeen@sandeen.net>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Dave Chinner <dchinner@redhat.com>

Index: xfs/fs/xfs/linux-2.6/xfs_aops.c
===================================================================
--- xfs.orig/fs/xfs/linux-2.6/xfs_aops.c	2010-06-10 08:59:49.598004335 +0200
+++ xfs/fs/xfs/linux-2.6/xfs_aops.c	2010-06-11 09:20:31.292254712 +0200
@@ -1125,6 +1125,16 @@ xfs_page_state_convert(
 			continue;
 		}
 
+		/*
+		 * A hole may still be marked uptodate because discard_buffer
+		 * leaves the flag set.
+		 */
+		if (!buffer_mapped(bh) && buffer_uptodate(bh)) {
+			ASSERT(!buffer_dirty(bh));
+			imap_valid = 0;
+			continue;
+		}
+
 		if (imap_valid)
 			imap_valid = xfs_imap_valid(inode, &imap, offset);
 

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

[PATCH 03/12] xfs: simplify xfs_vm_releasepage

[Christoph Hellwig]

[-- Attachment #1: xfs-simplify-releasepage --]
[-- Type: text/plain, Size: 17104 bytes --]

Currently the xfs releasepage implementation has code to deal with converting
delayed allocated and unwritten space.  But we never get called for those as
we always convert delayed and unwritten space when cleaning a page, or drop
the state from the buffers in block_invalidatepage.  We still keep a WARN_ON
on those cases for now, but remove all the case dealing with it, which allows
to fold xfs_page_state_convert into xfs_vm_writepage and remove the !startio
case from the whole writeback path.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Dave Chinner <dchinner@redhat.com>

Index: xfs/fs/xfs/linux-2.6/xfs_aops.c
===================================================================
--- xfs.orig/fs/xfs/linux-2.6/xfs_aops.c	2010-06-11 09:22:13.864253873 +0200
+++ xfs/fs/xfs/linux-2.6/xfs_aops.c	2010-06-11 09:27:27.602003840 +0200
@@ -754,7 +754,6 @@ xfs_convert_page(
 	struct xfs_bmbt_irec	*imap,
 	xfs_ioend_t		**ioendp,
 	struct writeback_control *wbc,
-	int			startio,
 	int			all_bh)
 {
 	struct buffer_head	*bh, *head;
@@ -825,19 +824,14 @@ xfs_convert_page(
 			ASSERT(imap->br_startblock != DELAYSTARTBLOCK);
 
 			xfs_map_at_offset(inode, bh, imap, offset);
-			if (startio) {
-				xfs_add_to_ioend(inode, bh, offset,
-						type, ioendp, done);
-			} else {
-				set_buffer_dirty(bh);
-				unlock_buffer(bh);
-				mark_buffer_dirty(bh);
-			}
+			xfs_add_to_ioend(inode, bh, offset, type,
+					 ioendp, done);
+
 			page_dirty--;
 			count++;
 		} else {
 			type = IO_NEW;
-			if (buffer_mapped(bh) && all_bh && startio) {
+			if (buffer_mapped(bh) && all_bh) {
 				lock_buffer(bh);
 				xfs_add_to_ioend(inode, bh, offset,
 						type, ioendp, done);
@@ -852,14 +846,12 @@ xfs_convert_page(
 	if (uptodate && bh == head)
 		SetPageUptodate(page);
 
-	if (startio) {
-		if (count) {
-			wbc->nr_to_write--;
-			if (wbc->nr_to_write <= 0)
-				done = 1;
-		}
-		xfs_start_page_writeback(page, !page_dirty, count);
+	if (count) {
+		wbc->nr_to_write--;
+		if (wbc->nr_to_write <= 0)
+			done = 1;
 	}
+	xfs_start_page_writeback(page, !page_dirty, count);
 
 	return done;
  fail_unlock_page:
@@ -879,7 +871,6 @@ xfs_cluster_write(
 	struct xfs_bmbt_irec	*imap,
 	xfs_ioend_t		**ioendp,
 	struct writeback_control *wbc,
-	int			startio,
 	int			all_bh,
 	pgoff_t			tlast)
 {
@@ -895,7 +886,7 @@ xfs_cluster_write(
 
 		for (i = 0; i < pagevec_count(&pvec); i++) {
 			done = xfs_convert_page(inode, pvec.pages[i], tindex++,
-					imap, ioendp, wbc, startio, all_bh);
+					imap, ioendp, wbc, all_bh);
 			if (done)
 				break;
 		}
@@ -1025,51 +1016,101 @@ out_invalidate:
 }
 
 /*
- * Calling this without startio set means we are being asked to make a dirty
- * page ready for freeing it's buffers.  When called with startio set then
- * we are coming from writepage.
+ * Write out a dirty page.
  *
- * When called with startio set it is important that we write the WHOLE
- * page if possible.
- * The bh->b_state's cannot know if any of the blocks or which block for
- * that matter are dirty due to mmap writes, and therefore bh uptodate is
- * only valid if the page itself isn't completely uptodate.  Some layers
- * may clear the page dirty flag prior to calling write page, under the
- * assumption the entire page will be written out; by not writing out the
- * whole page the page can be reused before all valid dirty data is
- * written out.  Note: in the case of a page that has been dirty'd by
- * mapwrite and but partially setup by block_prepare_write the
- * bh->b_states's will not agree and only ones setup by BPW/BCW will have
- * valid state, thus the whole page must be written out thing.
+ * For delalloc space on the page we need to allocate space and flush it.
+ * For unwritten space on the page we need to start the conversion to
+ * regular allocated space.
+ * For unmapped buffer heads on the page we should allocate space if the
+ * page is uptodate.
+ * For any other dirty buffer heads on the page we should flush them.
+ *
+ * If we detect that a transaction would be required to flush the page, we
+ * have to check the process flags first, if we are already in a transaction
+ * or disk I/O during allocations is off, we need to fail the writepage and
+ * redirty the page.
+ *
+ * The bh->b_state's cannot know if any of the blocks or which block for that
+ * matter are dirty due to mmap writes, and therefore bh uptodate is only
+ * valid if the page itself isn't completely uptodate.
  */
-
 STATIC int
-xfs_page_state_convert(
-	struct inode	*inode,
-	struct page	*page,
-	struct writeback_control *wbc,
-	int		startio,
-	int		unmapped) /* also implies page uptodate */
+xfs_vm_writepage(
+	struct page		*page,
+	struct writeback_control *wbc)
 {
+	struct inode		*inode = page->mapping->host;
+	int			need_trans;
+	int			delalloc, unmapped, unwritten;
 	struct buffer_head	*bh, *head;
 	struct xfs_bmbt_irec	imap;
 	xfs_ioend_t		*ioend = NULL, *iohead = NULL;
 	loff_t			offset;
-	unsigned long           p_offset = 0;
 	unsigned int		type;
 	__uint64_t              end_offset;
 	pgoff_t                 end_index, last_index;
 	ssize_t			size, len;
 	int			flags, err, imap_valid = 0, uptodate = 1;
-	int			page_dirty, count = 0;
-	int			trylock = 0;
-	int			all_bh = unmapped;
-
-	if (startio) {
-		if (wbc->sync_mode == WB_SYNC_NONE && wbc->nonblocking)
-			trylock |= BMAPI_TRYLOCK;
+	int			count = 0;
+	int			all_bh;
+
+	trace_xfs_writepage(inode, page, 0);
+
+	/*
+	 * Refuse to write the page out if we are called from reclaim context.
+	 *
+	 * This is primarily to avoid stack overflows when called from deep
+	 * used stacks in random callers for direct reclaim, but disabling
+	 * reclaim for kswap is a nice side-effect as kswapd causes rather
+	 * suboptimal I/O patters, too.
+	 *
+	 * This should really be done by the core VM, but until that happens
+	 * filesystems like XFS, btrfs and ext4 have to take care of this
+	 * by themselves.
+	 */
+	if (current->flags & PF_MEMALLOC)
+		goto out_fail;
+
+	/*
+	 * We need a transaction if:
+	 *  1. There are delalloc buffers on the page
+	 *  2. The page is uptodate and we have unmapped buffers
+	 *  3. The page is uptodate and we have no buffers
+	 *  4. There are unwritten buffers on the page
+	 */
+	if (!page_has_buffers(page)) {
+		unmapped = 1;
+		need_trans = 1;
+	} else {
+		xfs_count_page_state(page, &delalloc, &unmapped, &unwritten);
+		if (!PageUptodate(page))
+			unmapped = 0;
+		need_trans = delalloc + unmapped + unwritten;
 	}
 
+	/*
+	 * If we need a transaction and the process flags say
+	 * we are already in a transaction, or no IO is allowed
+	 * then mark the page dirty again and leave the page
+	 * as is.
+	 */
+	if (current_test_flags(PF_FSTRANS) && need_trans)
+		goto out_fail;
+
+	/*
+	 * Delay hooking up buffer heads until we have
+	 * made our go/no-go decision.
+	 */
+	if (!page_has_buffers(page))
+		create_empty_buffers(page, 1 << inode->i_blkbits, 0);
+
+	/*
+	 * VM calculation for nr_to_write seems off.  Bump it way
+	 * up, this gets simple streaming writes zippy again.
+	 * To be reviewed again after Jens' writeback changes.
+	 */
+	wbc->nr_to_write *= 4;
+
 	/* Is this page beyond the end of the file? */
 	offset = i_size_read(inode);
 	end_index = offset >> PAGE_CACHE_SHIFT;
@@ -1077,53 +1118,27 @@ xfs_page_state_convert(
 	if (page->index >= end_index) {
 		if ((page->index >= end_index + 1) ||
 		    !(i_size_read(inode) & (PAGE_CACHE_SIZE - 1))) {
-			if (startio)
-				unlock_page(page);
+			unlock_page(page);
 			return 0;
 		}
 	}
 
-	/*
-	 * page_dirty is initially a count of buffers on the page before
-	 * EOF and is decremented as we move each into a cleanable state.
-	 *
-	 * Derivation:
-	 *
-	 * End offset is the highest offset that this page should represent.
-	 * If we are on the last page, (end_offset & (PAGE_CACHE_SIZE - 1))
-	 * will evaluate non-zero and be less than PAGE_CACHE_SIZE and
-	 * hence give us the correct page_dirty count. On any other page,
-	 * it will be zero and in that case we need page_dirty to be the
-	 * count of buffers on the page.
- 	 */
 	end_offset = min_t(unsigned long long,
 			(xfs_off_t)(page->index + 1) << PAGE_CACHE_SHIFT, offset);
 	len = 1 << inode->i_blkbits;
-	p_offset = min_t(unsigned long, end_offset & (PAGE_CACHE_SIZE - 1),
-					PAGE_CACHE_SIZE);
-	p_offset = p_offset ? roundup(p_offset, len) : PAGE_CACHE_SIZE;
-	page_dirty = p_offset / len;
 
 	bh = head = page_buffers(page);
 	offset = page_offset(page);
 	flags = BMAPI_READ;
 	type = IO_NEW;
 
-	/* TODO: cleanup count and page_dirty */
+	all_bh = unmapped;
 
 	do {
 		if (offset >= end_offset)
 			break;
 		if (!buffer_uptodate(bh))
 			uptodate = 0;
-		if (!(PageUptodate(page) || buffer_uptodate(bh)) && !startio) {
-			/*
-			 * the iomap is actually still valid, but the ioend
-			 * isn't.  shouldn't happen too often.
-			 */
-			imap_valid = 0;
-			continue;
-		}
 
 		/*
 		 * A hole may still be marked uptodate because discard_buffer
@@ -1150,7 +1165,7 @@ xfs_page_state_convert(
 		 */
 		if (buffer_unwritten(bh) || buffer_delay(bh) ||
 		    ((buffer_uptodate(bh) || PageUptodate(page)) &&
-		     !buffer_mapped(bh) && (unmapped || startio))) {
+		     !buffer_mapped(bh))) {
 			int new_ioend = 0;
 
 			/*
@@ -1164,7 +1179,11 @@ xfs_page_state_convert(
 				flags = BMAPI_WRITE | BMAPI_IGNSTATE;
 			} else if (buffer_delay(bh)) {
 				type = IO_DELAY;
-				flags = BMAPI_ALLOCATE | trylock;
+				flags = BMAPI_ALLOCATE;
+
+				if (wbc->sync_mode == WB_SYNC_NONE &&
+				    wbc->nonblocking)
+					flags |= BMAPI_TRYLOCK;
 			} else {
 				type = IO_NEW;
 				flags = BMAPI_WRITE | BMAPI_MMAP;
@@ -1196,19 +1215,11 @@ xfs_page_state_convert(
 			}
 			if (imap_valid) {
 				xfs_map_at_offset(inode, bh, &imap, offset);
-				if (startio) {
-					xfs_add_to_ioend(inode, bh, offset,
-							type, &ioend,
-							new_ioend);
-				} else {
-					set_buffer_dirty(bh);
-					unlock_buffer(bh);
-					mark_buffer_dirty(bh);
-				}
-				page_dirty--;
+				xfs_add_to_ioend(inode, bh, offset, type,
+						 &ioend, new_ioend);
 				count++;
 			}
-		} else if (buffer_uptodate(bh) && startio) {
+		} else if (buffer_uptodate(bh)) {
 			/*
 			 * we got here because the buffer is already mapped.
 			 * That means it must already have extents allocated
@@ -1241,13 +1252,11 @@ xfs_page_state_convert(
 					all_bh = 1;
 				xfs_add_to_ioend(inode, bh, offset, type,
 						&ioend, !imap_valid);
-				page_dirty--;
 				count++;
 			} else {
 				imap_valid = 0;
 			}
-		} else if ((buffer_uptodate(bh) || PageUptodate(page)) &&
-			   (unmapped || startio)) {
+		} else if (PageUptodate(page)) {
 			imap_valid = 0;
 		}
 
@@ -1259,8 +1268,7 @@ xfs_page_state_convert(
 	if (uptodate && bh == head)
 		SetPageUptodate(page);
 
-	if (startio)
-		xfs_start_page_writeback(page, 1, count);
+	xfs_start_page_writeback(page, 1, count);
 
 	if (ioend && imap_valid) {
 		xfs_off_t		end_index;
@@ -1278,139 +1286,28 @@ xfs_page_state_convert(
 			end_index = last_index;
 
 		xfs_cluster_write(inode, page->index + 1, &imap, &ioend,
-					wbc, startio, all_bh, end_index);
+					wbc, all_bh, end_index);
 	}
 
 	if (iohead)
 		xfs_submit_ioend(wbc, iohead);
 
-	return page_dirty;
+	return 0;
 
 error:
 	if (iohead)
 		xfs_cancel_ioend(iohead);
 
-	/*
-	 * If it's delalloc and we have nowhere to put it,
-	 * throw it away, unless the lower layers told
-	 * us to try again.
-	 */
-	if (err != -EAGAIN) {
-		if (!unmapped)
-			xfs_aops_discard_page(page);
-		ClearPageUptodate(page);
-	}
+	if (!unmapped)
+		xfs_aops_discard_page(page);
+	ClearPageUptodate(page);
+	unlock_page(page);
 	return err;
-}
-
-/*
- * writepage: Called from one of two places:
- *
- * 1. we are flushing a delalloc buffer head.
- *
- * 2. we are writing out a dirty page. Typically the page dirty
- *    state is cleared before we get here. In this case is it
- *    conceivable we have no buffer heads.
- *
- * For delalloc space on the page we need to allocate space and
- * flush it. For unmapped buffer heads on the page we should
- * allocate space if the page is uptodate. For any other dirty
- * buffer heads on the page we should flush them.
- *
- * If we detect that a transaction would be required to flush
- * the page, we have to check the process flags first, if we
- * are already in a transaction or disk I/O during allocations
- * is off, we need to fail the writepage and redirty the page.
- */
-
-STATIC int
-xfs_vm_writepage(
-	struct page		*page,
-	struct writeback_control *wbc)
-{
-	int			error;
-	int			need_trans;
-	int			delalloc, unmapped, unwritten;
-	struct inode		*inode = page->mapping->host;
-
-	trace_xfs_writepage(inode, page, 0);
-
-	/*
-	 * Refuse to write the page out if we are called from reclaim context.
-	 *
-	 * This is primarily to avoid stack overflows when called from deep
-	 * used stacks in random callers for direct reclaim, but disabling
-	 * reclaim for kswap is a nice side-effect as kswapd causes rather
-	 * suboptimal I/O patters, too.
-	 *
-	 * This should really be done by the core VM, but until that happens
-	 * filesystems like XFS, btrfs and ext4 have to take care of this
-	 * by themselves.
-	 */
-	if (current->flags & PF_MEMALLOC)
-		goto out_fail;
-
-	/*
-	 * We need a transaction if:
-	 *  1. There are delalloc buffers on the page
-	 *  2. The page is uptodate and we have unmapped buffers
-	 *  3. The page is uptodate and we have no buffers
-	 *  4. There are unwritten buffers on the page
-	 */
-
-	if (!page_has_buffers(page)) {
-		unmapped = 1;
-		need_trans = 1;
-	} else {
-		xfs_count_page_state(page, &delalloc, &unmapped, &unwritten);
-		if (!PageUptodate(page))
-			unmapped = 0;
-		need_trans = delalloc + unmapped + unwritten;
-	}
-
-	/*
-	 * If we need a transaction and the process flags say
-	 * we are already in a transaction, or no IO is allowed
-	 * then mark the page dirty again and leave the page
-	 * as is.
-	 */
-	if (current_test_flags(PF_FSTRANS) && need_trans)
-		goto out_fail;
-
-	/*
-	 * Delay hooking up buffer heads until we have
-	 * made our go/no-go decision.
-	 */
-	if (!page_has_buffers(page))
-		create_empty_buffers(page, 1 << inode->i_blkbits, 0);
-
-
-	/*
-	 *  VM calculation for nr_to_write seems off.  Bump it way
-	 *  up, this gets simple streaming writes zippy again.
-	 *  To be reviewed again after Jens' writeback changes.
-	 */
-	wbc->nr_to_write *= 4;
-
-	/*
-	 * Convert delayed allocate, unwritten or unmapped space
-	 * to real space and flush out to disk.
-	 */
-	error = xfs_page_state_convert(inode, page, wbc, 1, unmapped);
-	if (error == -EAGAIN)
-		goto out_fail;
-	if (unlikely(error < 0))
-		goto out_unlock;
-
-	return 0;
 
 out_fail:
 	redirty_page_for_writepage(wbc, page);
 	unlock_page(page);
 	return 0;
-out_unlock:
-	unlock_page(page);
-	return error;
 }
 
 STATIC int
@@ -1424,65 +1321,27 @@ xfs_vm_writepages(
 
 /*
  * Called to move a page into cleanable state - and from there
- * to be released. Possibly the page is already clean. We always
+ * to be released. The page should already be clean. We always
  * have buffer heads in this call.
  *
- * Returns 0 if the page is ok to release, 1 otherwise.
- *
- * Possible scenarios are:
- *
- * 1. We are being called to release a page which has been written
- *    to via regular I/O. buffer heads will be dirty and possibly
- *    delalloc. If no delalloc buffer heads in this case then we
- *    can just return zero.
- *
- * 2. We are called to release a page which has been written via
- *    mmap, all we need to do is ensure there is no delalloc
- *    state in the buffer heads, if not we can let the caller
- *    free them and we should come back later via writepage.
+ * Returns 1 if the page is ok to release, 0 otherwise.
  */
 STATIC int
 xfs_vm_releasepage(
 	struct page		*page,
 	gfp_t			gfp_mask)
 {
-	struct inode		*inode = page->mapping->host;
-	int			dirty, delalloc, unmapped, unwritten;
-	struct writeback_control wbc = {
-		.sync_mode = WB_SYNC_ALL,
-		.nr_to_write = 1,
-	};
-
-	trace_xfs_releasepage(inode, page, 0);
+	int			delalloc, unmapped, unwritten;
 
-	if (!page_has_buffers(page))
-		return 0;
+	trace_xfs_releasepage(page->mapping->host, page, 0);
 
 	xfs_count_page_state(page, &delalloc, &unmapped, &unwritten);
-	if (!delalloc && !unwritten)
-		goto free_buffers;
 
-	if (!(gfp_mask & __GFP_FS))
+	if (WARN_ON(delalloc))
 		return 0;
-
-	/* If we are already inside a transaction or the thread cannot
-	 * do I/O, we cannot release this page.
-	 */
-	if (current_test_flags(PF_FSTRANS))
+	if (WARN_ON(unwritten))
 		return 0;
 
-	/*
-	 * Convert delalloc space to real space, do not flush the
-	 * data out to disk, that will be done by the caller.
-	 * Never need to allocate space here - we will always
-	 * come back to writepage in that case.
-	 */
-	dirty = xfs_page_state_convert(inode, page, &wbc, 0, 0);
-	if (dirty == 0 && !unwritten)
-		goto free_buffers;
-	return 0;
-
-free_buffers:
 	return try_to_free_buffers(page);
 }
 

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

[PATCH 04/12] xfs: simplify xfs_vm_writepage

[Christoph Hellwig]

[-- Attachment #1: xfs-simplify-writepage --]
[-- Type: text/plain, Size: 11788 bytes --]

The writepage implementation in XFS still tries to deal with dirty but
unmapped buffers which used to caused by writes through shared mmaps.  Since
the introduction of ->page_mkwrite these can't happen anymore, so remove the
code dealing with them.

Note that the all_bh variable which causes us to start I/O on all buffers on
the pages was controlled by the count of unmapped buffers, which also
included those not actually dirty.  It's now unconditionally initialized to
0 but set to 1 for the case of small file size extensions.  It probably can
be removed entirely, but that's left for another patch.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Dave Chinner <dchinner@redhat.com>

Index: xfs/fs/xfs/linux-2.6/xfs_aops.c
===================================================================
--- xfs.orig/fs/xfs/linux-2.6/xfs_aops.c	2010-06-11 09:27:27.602003840 +0200
+++ xfs/fs/xfs/linux-2.6/xfs_aops.c	2010-06-11 09:28:45.408003978 +0200
@@ -85,18 +85,15 @@ void
 xfs_count_page_state(
 	struct page		*page,
 	int			*delalloc,
-	int			*unmapped,
 	int			*unwritten)
 {
 	struct buffer_head	*bh, *head;
 
-	*delalloc = *unmapped = *unwritten = 0;
+	*delalloc = *unwritten = 0;
 
 	bh = head = page_buffers(page);
 	do {
-		if (buffer_uptodate(bh) && !buffer_mapped(bh))
-			(*unmapped) = 1;
-		else if (buffer_unwritten(bh))
+		if (buffer_unwritten(bh))
 			(*unwritten) = 1;
 		else if (buffer_delay(bh))
 			(*delalloc) = 1;
@@ -607,31 +604,30 @@ xfs_map_at_offset(
 STATIC unsigned int
 xfs_probe_page(
 	struct page		*page,
-	unsigned int		pg_offset,
-	int			mapped)
+	unsigned int		pg_offset)
 {
+	struct buffer_head	*bh, *head;
 	int			ret = 0;
 
 	if (PageWriteback(page))
 		return 0;
+	if (!PageDirty(page))
+		return 0;
+	if (!page->mapping)
+		return 0;
+	if (!page_has_buffers(page))
+		return 0;
 
-	if (page->mapping && PageDirty(page)) {
-		if (page_has_buffers(page)) {
-			struct buffer_head	*bh, *head;
-
-			bh = head = page_buffers(page);
-			do {
-				if (!buffer_uptodate(bh))
-					break;
-				if (mapped != buffer_mapped(bh))
-					break;
-				ret += bh->b_size;
-				if (ret >= pg_offset)
-					break;
-			} while ((bh = bh->b_this_page) != head);
-		} else
-			ret = mapped ? 0 : PAGE_CACHE_SIZE;
-	}
+	bh = head = page_buffers(page);
+	do {
+		if (!buffer_uptodate(bh))
+			break;
+		if (!buffer_mapped(bh))
+			break;
+		ret += bh->b_size;
+		if (ret >= pg_offset)
+			break;
+	} while ((bh = bh->b_this_page) != head);
 
 	return ret;
 }
@@ -641,8 +637,7 @@ xfs_probe_cluster(
 	struct inode		*inode,
 	struct page		*startpage,
 	struct buffer_head	*bh,
-	struct buffer_head	*head,
-	int			mapped)
+	struct buffer_head	*head)
 {
 	struct pagevec		pvec;
 	pgoff_t			tindex, tlast, tloff;
@@ -651,7 +646,7 @@ xfs_probe_cluster(
 
 	/* First sum forwards in this page */
 	do {
-		if (!buffer_uptodate(bh) || (mapped != buffer_mapped(bh)))
+		if (!buffer_uptodate(bh) || !buffer_mapped(bh))
 			return total;
 		total += bh->b_size;
 	} while ((bh = bh->b_this_page) != head);
@@ -685,7 +680,7 @@ xfs_probe_cluster(
 				pg_offset = PAGE_CACHE_SIZE;
 
 			if (page->index == tindex && trylock_page(page)) {
-				pg_len = xfs_probe_page(page, pg_offset, mapped);
+				pg_len = xfs_probe_page(page, pg_offset);
 				unlock_page(page);
 			}
 
@@ -1021,18 +1016,12 @@ out_invalidate:
  * For delalloc space on the page we need to allocate space and flush it.
  * For unwritten space on the page we need to start the conversion to
  * regular allocated space.
- * For unmapped buffer heads on the page we should allocate space if the
- * page is uptodate.
  * For any other dirty buffer heads on the page we should flush them.
  *
  * If we detect that a transaction would be required to flush the page, we
  * have to check the process flags first, if we are already in a transaction
  * or disk I/O during allocations is off, we need to fail the writepage and
  * redirty the page.
- *
- * The bh->b_state's cannot know if any of the blocks or which block for that
- * matter are dirty due to mmap writes, and therefore bh uptodate is only
- * valid if the page itself isn't completely uptodate.
  */
 STATIC int
 xfs_vm_writepage(
@@ -1040,8 +1029,7 @@ xfs_vm_writepage(
 	struct writeback_control *wbc)
 {
 	struct inode		*inode = page->mapping->host;
-	int			need_trans;
-	int			delalloc, unmapped, unwritten;
+	int			delalloc, unwritten;
 	struct buffer_head	*bh, *head;
 	struct xfs_bmbt_irec	imap;
 	xfs_ioend_t		*ioend = NULL, *iohead = NULL;
@@ -1052,10 +1040,12 @@ xfs_vm_writepage(
 	ssize_t			size, len;
 	int			flags, err, imap_valid = 0, uptodate = 1;
 	int			count = 0;
-	int			all_bh;
+	int			all_bh = 0;
 
 	trace_xfs_writepage(inode, page, 0);
 
+	ASSERT(page_has_buffers(page));
+
 	/*
 	 * Refuse to write the page out if we are called from reclaim context.
 	 *
@@ -1072,39 +1062,18 @@ xfs_vm_writepage(
 		goto out_fail;
 
 	/*
-	 * We need a transaction if:
-	 *  1. There are delalloc buffers on the page
-	 *  2. The page is uptodate and we have unmapped buffers
-	 *  3. The page is uptodate and we have no buffers
-	 *  4. There are unwritten buffers on the page
-	 */
-	if (!page_has_buffers(page)) {
-		unmapped = 1;
-		need_trans = 1;
-	} else {
-		xfs_count_page_state(page, &delalloc, &unmapped, &unwritten);
-		if (!PageUptodate(page))
-			unmapped = 0;
-		need_trans = delalloc + unmapped + unwritten;
-	}
-
-	/*
-	 * If we need a transaction and the process flags say
-	 * we are already in a transaction, or no IO is allowed
-	 * then mark the page dirty again and leave the page
-	 * as is.
+	 * We need a transaction if there are delalloc or unwritten buffers
+	 * on the page.
+	 *
+	 * If we need a transaction and the process flags say we are already
+	 * in a transaction, or no IO is allowed then mark the page dirty
+	 * again and leave the page as is.
 	 */
-	if (current_test_flags(PF_FSTRANS) && need_trans)
+	xfs_count_page_state(page, &delalloc, &unwritten);
+	if ((current->flags & PF_FSTRANS) && (delalloc || unwritten))
 		goto out_fail;
 
 	/*
-	 * Delay hooking up buffer heads until we have
-	 * made our go/no-go decision.
-	 */
-	if (!page_has_buffers(page))
-		create_empty_buffers(page, 1 << inode->i_blkbits, 0);
-
-	/*
 	 * VM calculation for nr_to_write seems off.  Bump it way
 	 * up, this gets simple streaming writes zippy again.
 	 * To be reviewed again after Jens' writeback changes.
@@ -1124,7 +1093,8 @@ xfs_vm_writepage(
 	}
 
 	end_offset = min_t(unsigned long long,
-			(xfs_off_t)(page->index + 1) << PAGE_CACHE_SHIFT, offset);
+			(xfs_off_t)(page->index + 1) << PAGE_CACHE_SHIFT,
+			offset);
 	len = 1 << inode->i_blkbits;
 
 	bh = head = page_buffers(page);
@@ -1132,8 +1102,6 @@ xfs_vm_writepage(
 	flags = BMAPI_READ;
 	type = IO_NEW;
 
-	all_bh = unmapped;
-
 	do {
 		if (offset >= end_offset)
 			break;
@@ -1153,19 +1121,7 @@ xfs_vm_writepage(
 		if (imap_valid)
 			imap_valid = xfs_imap_valid(inode, &imap, offset);
 
-		/*
-		 * First case, map an unwritten extent and prepare for
-		 * extent state conversion transaction on completion.
-		 *
-		 * Second case, allocate space for a delalloc buffer.
-		 * We can return EAGAIN here in the release page case.
-		 *
-		 * Third case, an unmapped buffer was found, and we are
-		 * in a path where we need to write the whole page out.
-		 */
-		if (buffer_unwritten(bh) || buffer_delay(bh) ||
-		    ((buffer_uptodate(bh) || PageUptodate(page)) &&
-		     !buffer_mapped(bh))) {
+		if (buffer_unwritten(bh) || buffer_delay(bh)) {
 			int new_ioend = 0;
 
 			/*
@@ -1184,14 +1140,11 @@ xfs_vm_writepage(
 				if (wbc->sync_mode == WB_SYNC_NONE &&
 				    wbc->nonblocking)
 					flags |= BMAPI_TRYLOCK;
-			} else {
-				type = IO_NEW;
-				flags = BMAPI_WRITE | BMAPI_MMAP;
 			}
 
 			if (!imap_valid) {
 				/*
-				 * if we didn't have a valid mapping then we
+				 * If we didn't have a valid mapping then we
 				 * need to ensure that we put the new mapping
 				 * in a new ioend structure. This needs to be
 				 * done to ensure that the ioends correctly
@@ -1199,14 +1152,7 @@ xfs_vm_writepage(
 				 * for unwritten extent conversion.
 				 */
 				new_ioend = 1;
-				if (type == IO_NEW) {
-					size = xfs_probe_cluster(inode,
-							page, bh, head, 0);
-				} else {
-					size = len;
-				}
-
-				err = xfs_map_blocks(inode, offset, size,
+				err = xfs_map_blocks(inode, offset, len,
 						&imap, flags);
 				if (err)
 					goto error;
@@ -1227,8 +1173,7 @@ xfs_vm_writepage(
 			 */
 			if (!imap_valid || flags != BMAPI_READ) {
 				flags = BMAPI_READ;
-				size = xfs_probe_cluster(inode, page, bh,
-								head, 1);
+				size = xfs_probe_cluster(inode, page, bh, head);
 				err = xfs_map_blocks(inode, offset, size,
 						&imap, flags);
 				if (err)
@@ -1247,7 +1192,6 @@ xfs_vm_writepage(
 			 */
 			type = IO_NEW;
 			if (trylock_buffer(bh)) {
-				ASSERT(buffer_mapped(bh));
 				if (imap_valid)
 					all_bh = 1;
 				xfs_add_to_ioend(inode, bh, offset, type,
@@ -1257,6 +1201,7 @@ xfs_vm_writepage(
 				imap_valid = 0;
 			}
 		} else if (PageUptodate(page)) {
+			ASSERT(buffer_mapped(bh));
 			imap_valid = 0;
 		}
 
@@ -1298,8 +1243,7 @@ error:
 	if (iohead)
 		xfs_cancel_ioend(iohead);
 
-	if (!unmapped)
-		xfs_aops_discard_page(page);
+	xfs_aops_discard_page(page);
 	ClearPageUptodate(page);
 	unlock_page(page);
 	return err;
@@ -1331,11 +1275,11 @@ xfs_vm_releasepage(
 	struct page		*page,
 	gfp_t			gfp_mask)
 {
-	int			delalloc, unmapped, unwritten;
+	int			delalloc, unwritten;
 
 	trace_xfs_releasepage(page->mapping->host, page, 0);
 
-	xfs_count_page_state(page, &delalloc, &unmapped, &unwritten);
+	xfs_count_page_state(page, &delalloc, &unwritten);
 
 	if (WARN_ON(delalloc))
 		return 0;
Index: xfs/fs/xfs/linux-2.6/xfs_aops.h
===================================================================
--- xfs.orig/fs/xfs/linux-2.6/xfs_aops.h	2010-02-13 22:45:45.000000000 +0100
+++ xfs/fs/xfs/linux-2.6/xfs_aops.h	2010-06-11 09:27:44.075004398 +0200
@@ -45,6 +45,6 @@ extern int xfs_get_blocks(struct inode *
 extern void xfs_ioend_init(void);
 extern void xfs_ioend_wait(struct xfs_inode *);
 
-extern void xfs_count_page_state(struct page *, int *, int *, int *);
+extern void xfs_count_page_state(struct page *, int *, int *);
 
 #endif /* __XFS_AOPS_H__ */
Index: xfs/fs/xfs/linux-2.6/xfs_trace.h
===================================================================
--- xfs.orig/fs/xfs/linux-2.6/xfs_trace.h	2010-06-04 15:12:52.000000000 +0200
+++ xfs/fs/xfs/linux-2.6/xfs_trace.h	2010-06-11 09:27:44.083003979 +0200
@@ -829,33 +829,29 @@ DECLARE_EVENT_CLASS(xfs_page_class,
 		__field(loff_t, size)
 		__field(unsigned long, offset)
 		__field(int, delalloc)
-		__field(int, unmapped)
 		__field(int, unwritten)
 	),
 	TP_fast_assign(
-		int delalloc = -1, unmapped = -1, unwritten = -1;
+		int delalloc = -1, unwritten = -1;
 
 		if (page_has_buffers(page))
-			xfs_count_page_state(page, &delalloc,
-					     &unmapped, &unwritten);
+			xfs_count_page_state(page, &delalloc, &unwritten);
 		__entry->dev = inode->i_sb->s_dev;
 		__entry->ino = XFS_I(inode)->i_ino;
 		__entry->pgoff = page_offset(page);
 		__entry->size = i_size_read(inode);
 		__entry->offset = off;
 		__entry->delalloc = delalloc;
-		__entry->unmapped = unmapped;
 		__entry->unwritten = unwritten;
 	),
 	TP_printk("dev %d:%d ino 0x%llx pgoff 0x%lx size 0x%llx offset %lx "
-		  "delalloc %d unmapped %d unwritten %d",
+		  "delalloc %d unwritten %d",
 		  MAJOR(__entry->dev), MINOR(__entry->dev),
 		  __entry->ino,
 		  __entry->pgoff,
 		  __entry->size,
 		  __entry->offset,
 		  __entry->delalloc,
-		  __entry->unmapped,
 		  __entry->unwritten)
 )
 

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

[PATCH 05/12] xfs: reduce stack usage in xfs_iomap

[Christoph Hellwig]

[-- Attachment #1: xfs-iomap-more-stack-reduction --]
[-- Type: text/plain, Size: 4713 bytes --]

xfs_iomap passes a xfs_bmbt_irec pointer to xfs_iomap_write_direct and
xfs_iomap_write_allocate to give them the results of our read-only
xfs_bmapi query.  Instead of allocating a new xfs_bmbt_irec on stack
for the next call to xfs_bmapi re use the one we got passed as it's not
used after this point.

Signed-off-by: Christoph Hellwig <hch@lst.de>

Index: xfs/fs/xfs/xfs_iomap.c
===================================================================
--- xfs.orig/fs/xfs/xfs_iomap.c	2010-06-18 09:58:41.210004416 +0200
+++ xfs/fs/xfs/xfs_iomap.c	2010-06-19 14:38:03.535254031 +0200
@@ -242,7 +242,7 @@ xfs_iomap_write_direct(
 	xfs_off_t	offset,
 	size_t		count,
 	int		flags,
-	xfs_bmbt_irec_t *ret_imap,
+	xfs_bmbt_irec_t *imap,
 	int		*nmaps)
 {
 	xfs_mount_t	*mp = ip->i_mount;
@@ -256,7 +256,6 @@ xfs_iomap_write_direct(
 	int		quota_flag;
 	int		rt;
 	xfs_trans_t	*tp;
-	xfs_bmbt_irec_t imap;
 	xfs_bmap_free_t free_list;
 	uint		qblocks, resblks, resrtextents;
 	int		committed;
@@ -280,10 +279,10 @@ xfs_iomap_write_direct(
 		if (error)
 			goto error_out;
 	} else {
-		if (*nmaps && (ret_imap->br_startblock == HOLESTARTBLOCK))
+		if (*nmaps && (imap->br_startblock == HOLESTARTBLOCK))
 			last_fsb = MIN(last_fsb, (xfs_fileoff_t)
-					ret_imap->br_blockcount +
-					ret_imap->br_startoff);
+					imap->br_blockcount +
+					imap->br_startoff);
 	}
 	count_fsb = last_fsb - offset_fsb;
 	ASSERT(count_fsb > 0);
@@ -336,12 +335,15 @@ xfs_iomap_write_direct(
 		bmapi_flag |= XFS_BMAPI_PREALLOC;
 
 	/*
-	 * Issue the xfs_bmapi() call to allocate the blocks
+	 * Issue the xfs_bmapi() call to allocate the blocks.
+	 *
+	 * From this point onwards we overwrite the imap pointer that the
+	 * caller gave to us.
 	 */
 	xfs_bmap_init(&free_list, &firstfsb);
 	nimaps = 1;
 	error = xfs_bmapi(tp, ip, offset_fsb, count_fsb, bmapi_flag,
-		&firstfsb, 0, &imap, &nimaps, &free_list);
+		&firstfsb, 0, imap, &nimaps, &free_list);
 	if (error)
 		goto error0;
 
@@ -363,12 +365,11 @@ xfs_iomap_write_direct(
 		goto error_out;
 	}
 
-	if (!(imap.br_startblock || XFS_IS_REALTIME_INODE(ip))) {
-		error = xfs_cmn_err_fsblock_zero(ip, &imap);
+	if (!(imap->br_startblock || XFS_IS_REALTIME_INODE(ip))) {
+		error = xfs_cmn_err_fsblock_zero(ip, imap);
 		goto error_out;
 	}
 
-	*ret_imap = imap;
 	*nmaps = 1;
 	return 0;
 
@@ -542,7 +543,7 @@ xfs_iomap_write_allocate(
 	xfs_inode_t	*ip,
 	xfs_off_t	offset,
 	size_t		count,
-	xfs_bmbt_irec_t *map,
+	xfs_bmbt_irec_t *imap,
 	int		*retmap)
 {
 	xfs_mount_t	*mp = ip->i_mount;
@@ -551,7 +552,6 @@ xfs_iomap_write_allocate(
 	xfs_fsblock_t	first_block;
 	xfs_bmap_free_t	free_list;
 	xfs_filblks_t	count_fsb;
-	xfs_bmbt_irec_t	imap;
 	xfs_trans_t	*tp;
 	int		nimaps, committed;
 	int		error = 0;
@@ -567,8 +567,8 @@ xfs_iomap_write_allocate(
 		return XFS_ERROR(error);
 
 	offset_fsb = XFS_B_TO_FSBT(mp, offset);
-	count_fsb = map->br_blockcount;
-	map_start_fsb = map->br_startoff;
+	count_fsb = imap->br_blockcount;
+	map_start_fsb = imap->br_startoff;
 
 	XFS_STATS_ADD(xs_xstrat_bytes, XFS_FSB_TO_B(mp, count_fsb));
 
@@ -647,10 +647,15 @@ xfs_iomap_write_allocate(
 				}
 			}
 
-			/* Go get the actual blocks */
+			/*
+			 * Go get the actual blocks.
+	 	 	 *
+			 * From this point onwards we overwrite the imap
+			 * pointer that the caller gave to us.
+			 */
 			error = xfs_bmapi(tp, ip, map_start_fsb, count_fsb,
 					XFS_BMAPI_WRITE, &first_block, 1,
-					&imap, &nimaps, &free_list);
+					imap, &nimaps, &free_list);
 			if (error)
 				goto trans_cancel;
 
@@ -669,13 +674,12 @@ xfs_iomap_write_allocate(
 		 * See if we were able to allocate an extent that
 		 * covers at least part of the callers request
 		 */
-		if (!(imap.br_startblock || XFS_IS_REALTIME_INODE(ip)))
-			return xfs_cmn_err_fsblock_zero(ip, &imap);
+		if (!(imap->br_startblock || XFS_IS_REALTIME_INODE(ip)))
+			return xfs_cmn_err_fsblock_zero(ip, imap);
 
-		if ((offset_fsb >= imap.br_startoff) &&
-		    (offset_fsb < (imap.br_startoff +
-				   imap.br_blockcount))) {
-			*map = imap;
+		if ((offset_fsb >= imap->br_startoff) &&
+		    (offset_fsb < (imap->br_startoff +
+				   imap->br_blockcount))) {
 			*retmap = 1;
 			XFS_STATS_INC(xs_xstrat_quick);
 			return 0;
@@ -685,8 +689,8 @@ xfs_iomap_write_allocate(
 		 * So far we have not mapped the requested part of the
 		 * file, just surrounding data, try again.
 		 */
-		count_fsb -= imap.br_blockcount;
-		map_start_fsb = imap.br_startoff + imap.br_blockcount;
+		count_fsb -= imap->br_blockcount;
+		map_start_fsb = imap->br_startoff + imap->br_blockcount;
 	}
 
 trans_cancel:

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

[PATCH 06/12] xfs: small cleanups for xfs_iomap / __xfs_get_blocks

[Christoph Hellwig]

[-- Attachment #1: xfs-iomap-cleanup --]
[-- Type: text/plain, Size: 3001 bytes --]

Remove the flags argument to  __xfs_get_blocks as we can easily derive
it from the direct argument, and remove the unused BMAPI_MMAP flag.

Signed-off-by: Christoph Hellwig <hch@lst.de>

Index: xfs/fs/xfs/linux-2.6/xfs_aops.c
===================================================================
--- xfs.orig/fs/xfs/linux-2.6/xfs_aops.c	2010-06-19 14:37:17.299004136 +0200
+++ xfs/fs/xfs/linux-2.6/xfs_aops.c	2010-06-19 14:39:34.313254590 +0200
@@ -1295,9 +1295,9 @@ __xfs_get_blocks(
 	sector_t		iblock,
 	struct buffer_head	*bh_result,
 	int			create,
-	int			direct,
-	bmapi_flags_t		flags)
+	int			direct)
 {
+	int			flags = create ? BMAPI_WRITE : BMAPI_READ;
 	struct xfs_bmbt_irec	imap;
 	xfs_off_t		offset;
 	ssize_t			size;
@@ -1312,8 +1312,11 @@ __xfs_get_blocks(
 	if (!create && direct && offset >= i_size_read(inode))
 		return 0;
 
-	error = xfs_iomap(XFS_I(inode), offset, size,
-			     create ? flags : BMAPI_READ, &imap, &nimap, &new);
+	if (direct && create)
+		flags |= BMAPI_DIRECT;
+
+	error = xfs_iomap(XFS_I(inode), offset, size, flags, &imap, &nimap,
+			  &new);
 	if (error)
 		return -error;
 	if (nimap == 0)
@@ -1393,8 +1396,7 @@ xfs_get_blocks(
 	struct buffer_head	*bh_result,
 	int			create)
 {
-	return __xfs_get_blocks(inode, iblock,
-				bh_result, create, 0, BMAPI_WRITE);
+	return __xfs_get_blocks(inode, iblock, bh_result, create, 0);
 }
 
 STATIC int
@@ -1404,8 +1406,7 @@ xfs_get_blocks_direct(
 	struct buffer_head	*bh_result,
 	int			create)
 {
-	return __xfs_get_blocks(inode, iblock,
-				bh_result, create, 1, BMAPI_WRITE|BMAPI_DIRECT);
+	return __xfs_get_blocks(inode, iblock, bh_result, create, 1);
 }
 
 STATIC void
Index: xfs/fs/xfs/xfs_iomap.c
===================================================================
--- xfs.orig/fs/xfs/xfs_iomap.c	2010-06-19 14:38:03.535254031 +0200
+++ xfs/fs/xfs/xfs_iomap.c	2010-06-19 14:39:34.313254590 +0200
@@ -133,7 +133,7 @@ xfs_iomap(
 			break;
 		}
 
-		if (flags & (BMAPI_DIRECT|BMAPI_MMAP)) {
+		if (flags & BMAPI_DIRECT) {
 			error = xfs_iomap_write_direct(ip, offset, count, flags,
 						       imap, nimaps);
 		} else {
Index: xfs/fs/xfs/xfs_iomap.h
===================================================================
--- xfs.orig/fs/xfs/xfs_iomap.h	2010-06-19 14:37:17.318004066 +0200
+++ xfs/fs/xfs/xfs_iomap.h	2010-06-19 14:39:34.318254380 +0200
@@ -26,7 +26,6 @@ typedef enum {
 	/* modifiers */
 	BMAPI_IGNSTATE = (1 << 4),	/* ignore unwritten state on read */
 	BMAPI_DIRECT = (1 << 5),	/* direct instead of buffered write */
-	BMAPI_MMAP = (1 << 6),		/* allocate for mmap write */
 	BMAPI_TRYLOCK = (1 << 7),	/* non-blocking request */
 } bmapi_flags_t;
 
@@ -36,7 +35,6 @@ typedef enum {
 	{ BMAPI_ALLOCATE,	"ALLOCATE" }, \
 	{ BMAPI_IGNSTATE,	"IGNSTATE" }, \
 	{ BMAPI_DIRECT,		"DIRECT" }, \
-	{ BMAPI_MMAP,		"MMAP" }, \
 	{ BMAPI_TRYLOCK,	"TRYLOCK" }
 
 struct xfs_inode;

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

[PATCH 07/12] xfs: remove explicit xfs_sync_data/xfs_sync_attr calls on umount

[Christoph Hellwig]

[-- Attachment #1: xfs-remove-sync-on-umount --]
[-- Type: text/plain, Size: 2355 bytes --]

On the final put of a superblock the VFS already calls sync_filesystem
for us to write out all data and wait for it.  No need to start another
asynchronous writenack inside ->put_super.


Signed-off-by: Christoph Hellwig <hch@lst.de>

Index: xfs/fs/xfs/linux-2.6/xfs_super.c
===================================================================
--- xfs.orig/fs/xfs/linux-2.6/xfs_super.c	2010-06-19 11:54:54.072006021 +0200
+++ xfs/fs/xfs/linux-2.6/xfs_super.c	2010-06-19 12:26:48.388003856 +0200
@@ -1160,18 +1160,6 @@ xfs_fs_put_super(
 
 	xfs_syncd_stop(mp);
 
-	if (!(sb->s_flags & MS_RDONLY)) {
-		/*
-		 * XXX(hch): this should be SYNC_WAIT.
-		 *
-		 * Or more likely not needed at all because the VFS is already
-		 * calling ->sync_fs after shutting down all filestem
-		 * operations and just before calling ->put_super.
-		 */
-		xfs_sync_data(mp, 0);
-		xfs_sync_attr(mp, 0);
-	}
-
 	/*
 	 * Blow away any referenced inode in the filestreams cache.
 	 * This can and will cause log traffic as inodes go inactive
Index: xfs/fs/xfs/linux-2.6/xfs_sync.c
===================================================================
--- xfs.orig/fs/xfs/linux-2.6/xfs_sync.c	2010-06-19 10:41:57.626004207 +0200
+++ xfs/fs/xfs/linux-2.6/xfs_sync.c	2010-06-19 12:26:48.389003298 +0200
@@ -274,7 +274,7 @@ xfs_sync_inode_attr(
 /*
  * Write out pagecache data for the whole filesystem.
  */
-int
+STATIC int
 xfs_sync_data(
 	struct xfs_mount	*mp,
 	int			flags)
@@ -295,7 +295,7 @@ xfs_sync_data(
 /*
  * Write out inode metadata (attributes) for the whole filesystem.
  */
-int
+STATIC int
 xfs_sync_attr(
 	struct xfs_mount	*mp,
 	int			flags)
Index: xfs/fs/xfs/linux-2.6/xfs_sync.h
===================================================================
--- xfs.orig/fs/xfs/linux-2.6/xfs_sync.h	2010-06-19 10:41:57.647253683 +0200
+++ xfs/fs/xfs/linux-2.6/xfs_sync.h	2010-06-19 12:26:48.391003717 +0200
@@ -35,9 +35,6 @@ typedef struct xfs_sync_work {
 int xfs_syncd_init(struct xfs_mount *mp);
 void xfs_syncd_stop(struct xfs_mount *mp);
 
-int xfs_sync_attr(struct xfs_mount *mp, int flags);
-int xfs_sync_data(struct xfs_mount *mp, int flags);
-
 int xfs_quiesce_data(struct xfs_mount *mp);
 void xfs_quiesce_attr(struct xfs_mount *mp);
 

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

[PATCH 08/12] xfs: do not use emums for flags used in tracing

[Christoph Hellwig]

[-- Attachment #1: xfs-tracing-doesnt-like-enums --]
[-- Type: text/plain, Size: 8344 bytes --]

The tracing code can't print flags defined as enums.  Most flags that
we want to print are defines as macros already, but move the few remaining
ones over to make the trace output more useful.

Signed-off-by: Christoph Hellwig <hch@lst.de>

Index: xfs/fs/xfs/xfs_alloc.h
===================================================================
--- xfs.orig/fs/xfs/xfs_alloc.h	2010-06-16 14:32:30.000000000 +0200
+++ xfs/fs/xfs/xfs_alloc.h	2010-06-16 14:41:06.701005358 +0200
@@ -27,16 +27,16 @@ struct xfs_busy_extent;
 /*
  * Freespace allocation types.  Argument to xfs_alloc_[v]extent.
  */
-typedef enum xfs_alloctype
-{
-	XFS_ALLOCTYPE_ANY_AG,		/* allocate anywhere, use rotor */
-	XFS_ALLOCTYPE_FIRST_AG,		/* ... start at ag 0 */
-	XFS_ALLOCTYPE_START_AG,		/* anywhere, start in this a.g. */
-	XFS_ALLOCTYPE_THIS_AG,		/* anywhere in this a.g. */
-	XFS_ALLOCTYPE_START_BNO,	/* near this block else anywhere */
-	XFS_ALLOCTYPE_NEAR_BNO,		/* in this a.g. and near this block */
-	XFS_ALLOCTYPE_THIS_BNO		/* at exactly this block */
-} xfs_alloctype_t;
+#define XFS_ALLOCTYPE_ANY_AG	0x01	/* allocate anywhere, use rotor */
+#define XFS_ALLOCTYPE_FIRST_AG	0x02	/* ... start at ag 0 */
+#define XFS_ALLOCTYPE_START_AG	0x04	/* anywhere, start in this a.g. */
+#define XFS_ALLOCTYPE_THIS_AG	0x08	/* anywhere in this a.g. */
+#define XFS_ALLOCTYPE_START_BNO	0x10	/* near this block else anywhere */
+#define XFS_ALLOCTYPE_NEAR_BNO	0x20	/* in this a.g. and near this block */
+#define XFS_ALLOCTYPE_THIS_BNO	0x40	/* at exactly this block */
+
+/* this should become an enum again when the tracing code is fixed */
+typedef unsigned int xfs_alloctype_t;
 
 #define XFS_ALLOC_TYPES \
 	{ XFS_ALLOCTYPE_ANY_AG,		"ANY_AG" }, \
Index: xfs/fs/xfs/xfs_iomap.h
===================================================================
--- xfs.orig/fs/xfs/xfs_iomap.h	2010-06-16 14:33:09.000000000 +0200
+++ xfs/fs/xfs/xfs_iomap.h	2010-06-16 14:39:50.906009896 +0200
@@ -18,16 +18,16 @@
 #ifndef __XFS_IOMAP_H__
 #define __XFS_IOMAP_H__
 
-typedef enum {
-	/* base extent manipulation calls */
-	BMAPI_READ = (1 << 0),		/* read extents */
-	BMAPI_WRITE = (1 << 1),		/* create extents */
-	BMAPI_ALLOCATE = (1 << 2),	/* delayed allocate to real extents */
-	/* modifiers */
-	BMAPI_IGNSTATE = (1 << 4),	/* ignore unwritten state on read */
-	BMAPI_DIRECT = (1 << 5),	/* direct instead of buffered write */
-	BMAPI_TRYLOCK = (1 << 7),	/* non-blocking request */
-} bmapi_flags_t;
+/* base extent manipulation calls */
+#define BMAPI_READ	(1 << 0)	/* read extents */
+#define BMAPI_WRITE	(1 << 1)	/* create extents */
+#define BMAPI_ALLOCATE	(1 << 2)	/* delayed allocate to real extents */
+
+/* modifiers */
+#define BMAPI_IGNSTATE	(1 << 4)	/* ignore unwritten state on read */
+#define BMAPI_DIRECT	(1 << 5)	/* direct instead of buffered write */
+#define BMAPI_MMA	(1 << 6)	/* allocate for mmap write */
+#define BMAPI_TRYLOCK	(1 << 7)	/* non-blocking request */
 
 #define BMAPI_FLAGS \
 	{ BMAPI_READ,		"READ" }, \
Index: xfs/fs/xfs/linux-2.6/xfs_buf.h
===================================================================
--- xfs.orig/fs/xfs/linux-2.6/xfs_buf.h	2010-06-16 14:41:35.278003751 +0200
+++ xfs/fs/xfs/linux-2.6/xfs_buf.h	2010-06-18 09:57:22.366004277 +0200
@@ -44,57 +44,57 @@ typedef enum {
 	XBRW_ZERO = 3,			/* Zero target memory */
 } xfs_buf_rw_t;
 
-typedef enum {
-	XBF_READ = (1 << 0),	/* buffer intended for reading from device */
-	XBF_WRITE = (1 << 1),	/* buffer intended for writing to device   */
-	XBF_MAPPED = (1 << 2),  /* buffer mapped (b_addr valid)            */
-	XBF_ASYNC = (1 << 4),   /* initiator will not wait for completion  */
-	XBF_DONE = (1 << 5),    /* all pages in the buffer uptodate	   */
-	XBF_DELWRI = (1 << 6),  /* buffer has dirty pages                  */
-	XBF_STALE = (1 << 7),	/* buffer has been staled, do not find it  */
-	XBF_FS_MANAGED = (1 << 8),  /* filesystem controls freeing memory  */
- 	XBF_ORDERED = (1 << 11),    /* use ordered writes		   */
-	XBF_READ_AHEAD = (1 << 12), /* asynchronous read-ahead		   */
-	XBF_LOG_BUFFER = (1 << 13), /* this is a buffer used for the log   */
-
-	/* flags used only as arguments to access routines */
-	XBF_LOCK = (1 << 14),       /* lock requested			   */
-	XBF_TRYLOCK = (1 << 15),    /* lock requested, but do not wait	   */
-	XBF_DONT_BLOCK = (1 << 16), /* do not block in current thread	   */
-
-	/* flags used only internally */
-	_XBF_PAGE_CACHE = (1 << 17),/* backed by pagecache		   */
-	_XBF_PAGES = (1 << 18),	    /* backed by refcounted pages	   */
-	_XBF_RUN_QUEUES = (1 << 19),/* run block device task queue	   */
-	_XBF_DELWRI_Q = (1 << 21),   /* buffer on delwri queue		   */
-
-	/*
-	 * Special flag for supporting metadata blocks smaller than a FSB.
-	 *
-	 * In this case we can have multiple xfs_buf_t on a single page and
-	 * need to lock out concurrent xfs_buf_t readers as they only
-	 * serialise access to the buffer.
-	 *
-	 * If the FSB size >= PAGE_CACHE_SIZE case, we have no serialisation
-	 * between reads of the page. Hence we can have one thread read the
-	 * page and modify it, but then race with another thread that thinks
-	 * the page is not up-to-date and hence reads it again.
-	 *
-	 * The result is that the first modifcation to the page is lost.
-	 * This sort of AGF/AGI reading race can happen when unlinking inodes
-	 * that require truncation and results in the AGI unlinked list
-	 * modifications being lost.
-	 */
-	_XBF_PAGE_LOCKED = (1 << 22),
-
-	/*
-	 * If we try a barrier write, but it fails we have to communicate
-	 * this to the upper layers.  Unfortunately b_error gets overwritten
-	 * when the buffer is re-issued so we have to add another flag to
-	 * keep this information.
-	 */
-	_XFS_BARRIER_FAILED = (1 << 23),
-} xfs_buf_flags_t;
+#define XBF_READ	(1 << 0) /* buffer intended for reading from device */
+#define XBF_WRITE	(1 << 1) /* buffer intended for writing to device */
+#define XBF_MAPPED	(1 << 2) /* buffer mapped (b_addr valid) */
+#define XBF_ASYNC	(1 << 4) /* initiator will not wait for completion */
+#define XBF_DONE	(1 << 5) /* all pages in the buffer uptodate */
+#define XBF_DELWRI	(1 << 6) /* buffer has dirty pages */
+#define XBF_STALE	(1 << 7) /* buffer has been staled, do not find it */
+#define XBF_FS_MANAGED	(1 << 8) /* filesystem controls freeing memory */
+#define XBF_ORDERED	(1 << 11)/* use ordered writes */
+#define XBF_READ_AHEAD	(1 << 12)/* asynchronous read-ahead */
+#define XBF_LOG_BUFFER	(1 << 13)/* this is a buffer used for the log */
+
+/* flags used only as arguments to access routines */
+#define XBF_LOCK	(1 << 14)/* lock requested */
+#define XBF_TRYLOCK	(1 << 15)/* lock requested, but do not wait */
+#define XBF_DONT_BLOCK	(1 << 16)/* do not block in current thread */
+
+/* flags used only internally */
+#define _XBF_PAGE_CACHE	(1 << 17)/* backed by pagecache */
+#define _XBF_PAGES	(1 << 18)/* backed by refcounted pages */
+#define	_XBF_RUN_QUEUES	(1 << 19)/* run block device task queue	*/
+#define _XBF_DELWRI_Q	(1 << 21)/* buffer on delwri queue */
+
+/*
+ * Special flag for supporting metadata blocks smaller than a FSB.
+ *
+ * In this case we can have multiple xfs_buf_t on a single page and
+ * need to lock out concurrent xfs_buf_t readers as they only
+ * serialise access to the buffer.
+ *
+ * If the FSB size >= PAGE_CACHE_SIZE case, we have no serialisation
+ * between reads of the page. Hence we can have one thread read the
+ * page and modify it, but then race with another thread that thinks
+ * the page is not up-to-date and hence reads it again.
+ *
+ * The result is that the first modifcation to the page is lost.
+ * This sort of AGF/AGI reading race can happen when unlinking inodes
+ * that require truncation and results in the AGI unlinked list
+ * modifications being lost.
+ */
+#define _XBF_PAGE_LOCKED	(1 << 22)
+
+/*
+ * If we try a barrier write, but it fails we have to communicate
+ * this to the upper layers.  Unfortunately b_error gets overwritten
+ * when the buffer is re-issued so we have to add another flag to
+ * keep this information.
+ */
+#define _XFS_BARRIER_FAILED	(1 << 23)
+
+typedef unsigned int xfs_buf_flags_t;
 
 #define XFS_BUF_FLAGS \
 	{ XBF_READ,		"READ" }, \

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

[PATCH 09/12] xfs: some iget tracing cleanups / fixes

[Christoph Hellwig]

[-- Attachment #1: xfs-cleanup-iget-tracing --]
[-- Type: text/plain, Size: 2661 bytes --]

The xfs_iget_alloc/found tracepoints are a bit misnamed and misplaced.
Rename them to xfs_iget_hit/xfs_iget_miss and move them to the beggining
of the xfs_iget_cache_hit/miss functions.  Add a new xfs_iget_reclaim_fail
tracepoint for the case where we fail to re-initialize a VFS inode,
and add a second instance of the xfs_iget_skip tracepoint for the case
of a failed igrab() call.

Signed-off-by: Christoph Hellwig <hch@lst.de>

Index: xfs/fs/xfs/linux-2.6/xfs_trace.h
===================================================================
--- xfs.orig/fs/xfs/linux-2.6/xfs_trace.h	2010-06-19 10:42:37.374266953 +0200
+++ xfs/fs/xfs/linux-2.6/xfs_trace.h	2010-06-19 14:37:04.680004137 +0200
@@ -558,8 +558,9 @@ DEFINE_EVENT(xfs_iget_class, name, \
 	TP_ARGS(ip))
 DEFINE_IGET_EVENT(xfs_iget_skip);
 DEFINE_IGET_EVENT(xfs_iget_reclaim);
-DEFINE_IGET_EVENT(xfs_iget_found);
-DEFINE_IGET_EVENT(xfs_iget_alloc);
+DEFINE_IGET_EVENT(xfs_iget_reclaim_fail);
+DEFINE_IGET_EVENT(xfs_iget_hit);
+DEFINE_IGET_EVENT(xfs_iget_miss);
 
 DECLARE_EVENT_CLASS(xfs_inode_class,
 	TP_PROTO(struct xfs_inode *ip, unsigned long caller_ip),
Index: xfs/fs/xfs/xfs_iget.c
===================================================================
--- xfs.orig/fs/xfs/xfs_iget.c	2010-06-19 10:44:00.955254311 +0200
+++ xfs/fs/xfs/xfs_iget.c	2010-06-19 10:54:39.000000000 +0200
@@ -208,7 +208,7 @@ xfs_iget_cache_hit(
 			ip->i_flags &= ~XFS_INEW;
 			ip->i_flags |= XFS_IRECLAIMABLE;
 			__xfs_inode_set_reclaim_tag(pag, ip);
-			trace_xfs_iget_reclaim(ip);
+			trace_xfs_iget_reclaim_fail(ip);
 			goto out_error;
 		}
 
@@ -223,6 +223,7 @@ xfs_iget_cache_hit(
 	} else {
 		/* If the VFS inode is being torn down, pause and try again. */
 		if (!igrab(inode)) {
+			trace_xfs_iget_skip(ip);
 			error = EAGAIN;
 			goto out_error;
 		}
@@ -230,6 +231,7 @@ xfs_iget_cache_hit(
 		/* We've got a live one. */
 		spin_unlock(&ip->i_flags_lock);
 		read_unlock(&pag->pag_ici_lock);
+		trace_xfs_iget_hit(ip);
 	}
 
 	if (lock_flags != 0)
@@ -238,7 +240,6 @@ xfs_iget_cache_hit(
 	xfs_iflags_clear(ip, XFS_ISTALE);
 	XFS_STATS_INC(xs_ig_found);
 
-	trace_xfs_iget_found(ip);
 	return 0;
 
 out_error:
@@ -271,7 +272,7 @@ xfs_iget_cache_miss(
 	if (error)
 		goto out_destroy;
 
-	xfs_itrace_entry(ip);
+	trace_xfs_iget_miss(ip);
 
 	if ((ip->i_d.di_mode == 0) && !(flags & XFS_IGET_CREATE)) {
 		error = ENOENT;
@@ -317,7 +318,6 @@ xfs_iget_cache_miss(
 	write_unlock(&pag->pag_ici_lock);
 	radix_tree_preload_end();
 
-	trace_xfs_iget_alloc(ip);
 	*ipp = ip;
 	return 0;
 

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

[PATCH 10/12] [PATCH] xfs: remove xfs_iput_new

[Christoph Hellwig]

[-- Attachment #1: xfs-kill-xfs_iput_new --]
[-- Type: text/plain, Size: 3193 bytes --]

We never get an i_mode of 0 or a locked VFS inode until we pass in the
XFS_IGET_CREATE flag to xfs_iget, which makes xfs_iput_new equivalent to
xfs_iput for the only caller.  In addition to that xfs_nfs_get_inode
does not even need to lock the inode given that the generation never changes
for a life inode, so just pass a 0 lock_flags to xfs_iget and release
the inode using IRELE in the error path.

Signed-off-by: Christoph Hellwig <hch@lst.de>

Index: xfs/fs/xfs/linux-2.6/xfs_export.c
===================================================================
--- xfs.orig/fs/xfs/linux-2.6/xfs_export.c	2010-06-19 23:06:53.086254161 +0200
+++ xfs/fs/xfs/linux-2.6/xfs_export.c	2010-06-19 23:07:10.000000000 +0200
@@ -29,6 +29,7 @@
 #include "xfs_bmap_btree.h"
 #include "xfs_inode.h"
 #include "xfs_inode_item.h"
+#include "xfs_trace.h"
 
 /*
  * Note that we only accept fileids which are long enough rather than allow
@@ -131,8 +132,7 @@ xfs_nfs_get_inode(
 	 * fine and not an indication of a corrupted filesystem as clients can
 	 * send invalid file handles and we have to handle it gracefully..
 	 */
-	error = xfs_iget(mp, NULL, ino, XFS_IGET_UNTRUSTED,
-			 XFS_ILOCK_SHARED, &ip);
+	error = xfs_iget(mp, NULL, ino, XFS_IGET_UNTRUSTED, 0, &ip);
 	if (error) {
 		/*
 		 * EINVAL means the inode cluster doesn't exist anymore.
@@ -147,11 +147,10 @@ xfs_nfs_get_inode(
 	}
 
 	if (ip->i_d.di_gen != generation) {
-		xfs_iput_new(ip, XFS_ILOCK_SHARED);
+		IRELE(ip);
 		return ERR_PTR(-ENOENT);
 	}
 
-	xfs_iunlock(ip, XFS_ILOCK_SHARED);
 	return VFS_I(ip);
 }
 
Index: xfs/fs/xfs/xfs_iget.c
===================================================================
--- xfs.orig/fs/xfs/xfs_iget.c	2010-06-19 23:06:57.008263659 +0200
+++ xfs/fs/xfs/xfs_iget.c	2010-06-21 09:13:49.643254301 +0200
@@ -435,29 +435,6 @@ xfs_iput(xfs_inode_t	*ip,
 }
 
 /*
- * Special iput for brand-new inodes that are still locked
- */
-void
-xfs_iput_new(
-	xfs_inode_t	*ip,
-	uint		lock_flags)
-{
-	struct inode	*inode = VFS_I(ip);
-
-	xfs_itrace_entry(ip);
-
-	if ((ip->i_d.di_mode == 0)) {
-		ASSERT(!xfs_iflags_test(ip, XFS_IRECLAIMABLE));
-		make_bad_inode(inode);
-	}
-	if (inode->i_state & I_NEW)
-		unlock_new_inode(inode);
-	if (lock_flags)
-		xfs_iunlock(ip, lock_flags);
-	IRELE(ip);
-}
-
-/*
  * This is called free all the memory associated with an inode.
  * It must free the inode itself and any buffers allocated for
  * if_extents/if_data and if_broot.  It must also free the lock
Index: xfs/fs/xfs/xfs_inode.h
===================================================================
--- xfs.orig/fs/xfs/xfs_inode.h	2010-06-19 23:06:53.112024311 +0200
+++ xfs/fs/xfs/xfs_inode.h	2010-06-21 09:13:49.669254301 +0200
@@ -444,7 +444,6 @@ static inline void xfs_ifunlock(xfs_inod
 int		xfs_iget(struct xfs_mount *, struct xfs_trans *, xfs_ino_t,
 			 uint, uint, xfs_inode_t **);
 void		xfs_iput(xfs_inode_t *, uint);
-void		xfs_iput_new(xfs_inode_t *, uint);
 void		xfs_ilock(xfs_inode_t *, uint);
 int		xfs_ilock_nowait(xfs_inode_t *, uint);
 void		xfs_iunlock(xfs_inode_t *, uint);

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

[PATCH 11/12] xfs: remove xfs_iput

[Christoph Hellwig]

[-- Attachment #1: xfs-remove-xfs_iput --]
[-- Type: text/plain, Size: 4741 bytes --]

xfs_iput is just a small wrapper for xfs_iunlock + IRELE.  Having this
out of line wrapper means the trace events in those two can't track
their caller properly.  So just remove the wrapper and opencode the
unlock + rele in the few callers.

Signed-off-by: Christoph Hellwig <hch@lst.de>

Index: xfs/fs/xfs/xfs_iget.c
===================================================================
--- xfs.orig/fs/xfs/xfs_iget.c	2010-06-19 23:06:59.269003778 +0200
+++ xfs/fs/xfs/xfs_iget.c	2010-06-19 23:08:14.611253952 +0200
@@ -418,23 +418,6 @@ out_error_or_again:
 }
 
 /*
- * Decrement reference count of an inode structure and unlock it.
- *
- * ip -- the inode being released
- * lock_flags -- this parameter indicates the inode's locks to be
- *       to be released.  See the comment on xfs_iunlock() for a list
- *	 of valid values.
- */
-void
-xfs_iput(xfs_inode_t	*ip,
-	 uint		lock_flags)
-{
-	xfs_itrace_entry(ip);
-	xfs_iunlock(ip, lock_flags);
-	IRELE(ip);
-}
-
-/*
  * This is called free all the memory associated with an inode.
  * It must free the inode itself and any buffers allocated for
  * if_extents/if_data and if_broot.  It must also free the lock
Index: xfs/fs/xfs/quota/xfs_qm.c
===================================================================
--- xfs.orig/fs/xfs/quota/xfs_qm.c	2010-06-19 23:06:53.000000000 +0200
+++ xfs/fs/xfs/quota/xfs_qm.c	2010-06-19 23:08:14.612254022 +0200
@@ -1662,7 +1662,8 @@ xfs_qm_dqusage_adjust(
 	 * making us disable quotas for the file system.
 	 */
 	if ((error = xfs_qm_dqget_noattach(ip, &udqp, &gdqp))) {
-		xfs_iput(ip, XFS_ILOCK_EXCL);
+		xfs_iunlock(ip, XFS_ILOCK_EXCL);
+		IRELE(ip);
 		*res = BULKSTAT_RV_GIVEUP;
 		return error;
 	}
@@ -1675,7 +1676,8 @@ xfs_qm_dqusage_adjust(
 		 * Walk thru the extent list and count the realtime blocks.
 		 */
 		if ((error = xfs_qm_get_rtblks(ip, &rtblks))) {
-			xfs_iput(ip, XFS_ILOCK_EXCL);
+			xfs_iunlock(ip, XFS_ILOCK_EXCL);
+			IRELE(ip);
 			if (udqp)
 				xfs_qm_dqput(udqp);
 			if (gdqp)
Index: xfs/fs/xfs/quota/xfs_qm_syscalls.c
===================================================================
--- xfs.orig/fs/xfs/quota/xfs_qm_syscalls.c	2010-06-19 23:06:53.000000000 +0200
+++ xfs/fs/xfs/quota/xfs_qm_syscalls.c	2010-06-19 23:08:14.613254161 +0200
@@ -861,8 +861,9 @@ xfs_dqrele_inode(
 		xfs_qm_dqrele(ip->i_gdquot);
 		ip->i_gdquot = NULL;
 	}
-	xfs_iput(ip, XFS_ILOCK_EXCL);
+	xfs_iunlock(ip, XFS_ILOCK_EXCL);
 
+	IRELE(ip);
 	return 0;
 }
 
@@ -1129,7 +1130,8 @@ xfs_qm_internalqcheck_adjust(
 	 * of those now.
 	 */
 	if (! ipreleased) {
-		xfs_iput(ip, lock_flags);
+		xfs_iunlock(ip, lock_flags);
+		IRELE(ip);
 		ipreleased = B_TRUE;
 		goto again;
 	}
@@ -1146,7 +1148,8 @@ xfs_qm_internalqcheck_adjust(
 		ASSERT(gd);
 		xfs_qm_internalqcheck_dqadjust(ip, gd);
 	}
-	xfs_iput(ip, lock_flags);
+	xfs_iunlock(ip, lock_flags);
+	IRELE(ip);
 	*res = BULKSTAT_RV_DIDONE;
 	return (0);
 }
Index: xfs/fs/xfs/xfs_inode.h
===================================================================
--- xfs.orig/fs/xfs/xfs_inode.h	2010-06-19 23:06:59.000000000 +0200
+++ xfs/fs/xfs/xfs_inode.h	2010-06-19 23:08:14.618253882 +0200
@@ -443,7 +443,6 @@ static inline void xfs_ifunlock(xfs_inod
  */
 int		xfs_iget(struct xfs_mount *, struct xfs_trans *, xfs_ino_t,
 			 uint, uint, xfs_inode_t **);
-void		xfs_iput(xfs_inode_t *, uint);
 void		xfs_ilock(xfs_inode_t *, uint);
 int		xfs_ilock_nowait(xfs_inode_t *, uint);
 void		xfs_iunlock(xfs_inode_t *, uint);
Index: xfs/fs/xfs/xfs_inode_item.c
===================================================================
--- xfs.orig/fs/xfs/xfs_inode_item.c	2010-06-19 23:06:49.000000000 +0200
+++ xfs/fs/xfs/xfs_inode_item.c	2010-06-19 23:08:14.623253952 +0200
@@ -660,8 +660,10 @@ xfs_inode_item_unlock(
 
 	lock_flags = iip->ili_lock_flags;
 	iip->ili_lock_flags = 0;
-	if (lock_flags)
-		xfs_iput(iip->ili_inode, lock_flags);
+	if (lock_flags) {
+		xfs_iunlock(iip->ili_inode, lock_flags);
+		IRELE(iip->ili_inode);
+	}
 }
 
 /*
Index: xfs/fs/xfs/xfs_itable.c
===================================================================
--- xfs.orig/fs/xfs/xfs_itable.c	2010-06-19 23:06:53.000000000 +0200
+++ xfs/fs/xfs/xfs_itable.c	2010-06-19 23:08:14.627253952 +0200
@@ -34,6 +34,7 @@
 #include "xfs_itable.h"
 #include "xfs_error.h"
 #include "xfs_btree.h"
+#include "xfs_trace.h"
 
 STATIC int
 xfs_internal_inum(
@@ -139,7 +140,8 @@ xfs_bulkstat_one_int(
 		buf->bs_blocks = dic->di_nblocks + ip->i_delayed_blks;
 		break;
 	}
-	xfs_iput(ip, XFS_ILOCK_SHARED);
+	xfs_iunlock(ip, XFS_ILOCK_SHARED);
+	IRELE(ip);
 
 	error = formatter(buffer, ubsize, ubused, buf);
 

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

[PATCH 12/12] xfs: split xfs_itrace_entry

[Christoph Hellwig]

[-- Attachment #1: xfs-split-itrace --]
[-- Type: text/plain, Size: 13260 bytes --]

Replace the xfs_itrace_entry catchall with specific trace points.  For
most simple callers we now use the simple inode class, which used to
be the iget class, but add more details tracing for namespace events,
which now includes the name of the directory entries manipulated.

Remove the xfs_inactive trace point, which is a duplicate of the clear_inode
one, and the xfs_change_file_space trace point, which is immediately
followed by the more specific alloc/free space trace points.

Signed-off-by: Christoph Hellwig <hch@lst.de>

Index: xfs/fs/xfs/linux-2.6/xfs_acl.c
===================================================================
--- xfs.orig/fs/xfs/linux-2.6/xfs_acl.c	2010-06-21 09:31:42.750254231 +0200
+++ xfs/fs/xfs/linux-2.6/xfs_acl.c	2010-06-21 09:38:38.800255768 +0200
@@ -225,7 +225,7 @@ xfs_check_acl(struct inode *inode, int m
 	struct posix_acl *acl;
 	int error = -EAGAIN;
 
-	xfs_itrace_entry(ip);
+	trace_xfs_check_acl(ip);
 
 	/*
 	 * If there is no attribute fork no ACL exists on this inode and
Index: xfs/fs/xfs/linux-2.6/xfs_aops.c
===================================================================
--- xfs.orig/fs/xfs/linux-2.6/xfs_aops.c	2010-06-21 09:38:38.667254022 +0200
+++ xfs/fs/xfs/linux-2.6/xfs_aops.c	2010-06-21 09:38:38.802285241 +0200
@@ -1513,7 +1513,7 @@ xfs_vm_bmap(
 	struct inode		*inode = (struct inode *)mapping->host;
 	struct xfs_inode	*ip = XFS_I(inode);
 
-	xfs_itrace_entry(XFS_I(inode));
+	trace_xfs_vm_bmap(XFS_I(inode));
 	xfs_ilock(ip, XFS_IOLOCK_SHARED);
 	xfs_flush_pages(ip, (xfs_off_t)0, -1, 0, FI_REMAPF);
 	xfs_iunlock(ip, XFS_IOLOCK_SHARED);
Index: xfs/fs/xfs/linux-2.6/xfs_file.c
===================================================================
--- xfs.orig/fs/xfs/linux-2.6/xfs_file.c	2010-06-21 09:31:42.765253742 +0200
+++ xfs/fs/xfs/linux-2.6/xfs_file.c	2010-06-21 09:38:38.807317438 +0200
@@ -100,7 +100,7 @@ xfs_file_fsync(
 	int			error = 0;
 	int			log_flushed = 0;
 
-	xfs_itrace_entry(ip);
+	trace_xfs_file_fsync(ip);
 
 	if (XFS_FORCED_SHUTDOWN(ip->i_mount))
 		return -XFS_ERROR(EIO);
Index: xfs/fs/xfs/linux-2.6/xfs_ioctl.c
===================================================================
--- xfs.orig/fs/xfs/linux-2.6/xfs_ioctl.c	2010-06-21 09:31:42.775254720 +0200
+++ xfs/fs/xfs/linux-2.6/xfs_ioctl.c	2010-06-21 09:38:38.810292505 +0200
@@ -899,7 +899,7 @@ xfs_ioctl_setattr(
 	struct xfs_dquot	*olddquot = NULL;
 	int			code;
 
-	xfs_itrace_entry(ip);
+	trace_xfs_ioctl_setattr(ip);
 
 	if (mp->m_flags & XFS_MOUNT_RDONLY)
 		return XFS_ERROR(EROFS);
@@ -1282,7 +1282,7 @@ xfs_file_ioctl(
 	if (filp->f_mode & FMODE_NOCMTIME)
 		ioflags |= IO_INVIS;
 
-	xfs_itrace_entry(ip);
+	trace_xfs_file_ioctl(ip);
 
 	switch (cmd) {
 	case XFS_IOC_ALLOCSP:
Index: xfs/fs/xfs/linux-2.6/xfs_ioctl32.c
===================================================================
--- xfs.orig/fs/xfs/linux-2.6/xfs_ioctl32.c	2010-06-21 09:31:47.810003149 +0200
+++ xfs/fs/xfs/linux-2.6/xfs_ioctl32.c	2010-06-21 09:38:38.815255279 +0200
@@ -540,7 +540,7 @@ xfs_file_compat_ioctl(
 	if (filp->f_mode & FMODE_NOCMTIME)
 		ioflags |= IO_INVIS;
 
-	xfs_itrace_entry(ip);
+	trace_xfs_file_compat_ioctl(ip);
 
 	switch (cmd) {
 	/* No size or alignment issues on any arch */
Index: xfs/fs/xfs/linux-2.6/xfs_iops.c
===================================================================
--- xfs.orig/fs/xfs/linux-2.6/xfs_iops.c	2010-06-21 09:31:42.790254790 +0200
+++ xfs/fs/xfs/linux-2.6/xfs_iops.c	2010-06-21 09:38:38.820255558 +0200
@@ -488,7 +488,7 @@ xfs_vn_getattr(
 	struct xfs_inode	*ip = XFS_I(inode);
 	struct xfs_mount	*mp = ip->i_mount;
 
-	xfs_itrace_entry(ip);
+	trace_xfs_getattr(ip);
 
 	if (XFS_FORCED_SHUTDOWN(mp))
 		return XFS_ERROR(EIO);
Index: xfs/fs/xfs/linux-2.6/xfs_super.c
===================================================================
--- xfs.orig/fs/xfs/linux-2.6/xfs_super.c	2010-06-21 09:38:38.683275324 +0200
+++ xfs/fs/xfs/linux-2.6/xfs_super.c	2010-06-21 09:38:38.827255768 +0200
@@ -907,7 +907,7 @@ xfs_fs_destroy_inode(
 {
 	struct xfs_inode	*ip = XFS_I(inode);
 
-	xfs_itrace_entry(ip);
+	trace_xfs_destroy_inode(ip);
 
 	XFS_STATS_INC(vn_reclaim);
 
@@ -1040,7 +1040,7 @@ xfs_fs_write_inode(
 	struct xfs_mount	*mp = ip->i_mount;
 	int			error = EAGAIN;
 
-	xfs_itrace_entry(ip);
+	trace_xfs_write_inode(ip);
 
 	if (XFS_FORCED_SHUTDOWN(mp))
 		return XFS_ERROR(EIO);
@@ -1107,7 +1107,8 @@ xfs_fs_clear_inode(
 {
 	xfs_inode_t		*ip = XFS_I(inode);
 
-	xfs_itrace_entry(ip);
+	trace_xfs_clear_inode(ip);
+
 	XFS_STATS_INC(vn_rele);
 	XFS_STATS_INC(vn_remove);
 	XFS_STATS_DEC(vn_active);
Index: xfs/fs/xfs/linux-2.6/xfs_trace.h
===================================================================
--- xfs.orig/fs/xfs/linux-2.6/xfs_trace.h	2010-06-21 09:38:38.711253323 +0200
+++ xfs/fs/xfs/linux-2.6/xfs_trace.h	2010-06-21 09:38:38.834282797 +0200
@@ -536,7 +536,7 @@ DEFINE_LOCK_EVENT(xfs_ilock_nowait);
 DEFINE_LOCK_EVENT(xfs_ilock_demote);
 DEFINE_LOCK_EVENT(xfs_iunlock);
 
-DECLARE_EVENT_CLASS(xfs_iget_class,
+DECLARE_EVENT_CLASS(xfs_inode_class,
 	TP_PROTO(struct xfs_inode *ip),
 	TP_ARGS(ip),
 	TP_STRUCT__entry(
@@ -552,17 +552,36 @@ DECLARE_EVENT_CLASS(xfs_iget_class,
 		  __entry->ino)
 )
 
-#define DEFINE_IGET_EVENT(name) \
-DEFINE_EVENT(xfs_iget_class, name, \
+#define DEFINE_INODE_EVENT(name) \
+DEFINE_EVENT(xfs_inode_class, name, \
 	TP_PROTO(struct xfs_inode *ip), \
 	TP_ARGS(ip))
-DEFINE_IGET_EVENT(xfs_iget_skip);
-DEFINE_IGET_EVENT(xfs_iget_reclaim);
-DEFINE_IGET_EVENT(xfs_iget_reclaim_fail);
-DEFINE_IGET_EVENT(xfs_iget_hit);
-DEFINE_IGET_EVENT(xfs_iget_miss);
+DEFINE_INODE_EVENT(xfs_iget_skip);
+DEFINE_INODE_EVENT(xfs_iget_reclaim);
+DEFINE_INODE_EVENT(xfs_iget_reclaim_fail);
+DEFINE_INODE_EVENT(xfs_iget_hit);
+DEFINE_INODE_EVENT(xfs_iget_miss);
+
+DEFINE_INODE_EVENT(xfs_getattr);
+DEFINE_INODE_EVENT(xfs_setattr);
+DEFINE_INODE_EVENT(xfs_readlink);
+DEFINE_INODE_EVENT(xfs_alloc_file_space);
+DEFINE_INODE_EVENT(xfs_free_file_space);
+DEFINE_INODE_EVENT(xfs_readdir);
+DEFINE_INODE_EVENT(xfs_check_acl);
+DEFINE_INODE_EVENT(xfs_vm_bmap);
+DEFINE_INODE_EVENT(xfs_file_ioctl);
+DEFINE_INODE_EVENT(xfs_file_compat_ioctl);
+DEFINE_INODE_EVENT(xfs_ioctl_setattr);
+DEFINE_INODE_EVENT(xfs_file_fsync);
+DEFINE_INODE_EVENT(xfs_destroy_inode);
+DEFINE_INODE_EVENT(xfs_write_inode);
+DEFINE_INODE_EVENT(xfs_clear_inode);
 
-DECLARE_EVENT_CLASS(xfs_inode_class,
+DEFINE_INODE_EVENT(xfs_dquot_dqalloc);
+DEFINE_INODE_EVENT(xfs_dquot_dqdetach);
+
+DECLARE_EVENT_CLASS(xfs_iref_class,
 	TP_PROTO(struct xfs_inode *ip, unsigned long caller_ip),
 	TP_ARGS(ip, caller_ip),
 	TP_STRUCT__entry(
@@ -587,20 +606,71 @@ DECLARE_EVENT_CLASS(xfs_inode_class,
 		  (char *)__entry->caller_ip)
 )
 
-#define DEFINE_INODE_EVENT(name) \
-DEFINE_EVENT(xfs_inode_class, name, \
+#define DEFINE_IREF_EVENT(name) \
+DEFINE_EVENT(xfs_iref_class, name, \
 	TP_PROTO(struct xfs_inode *ip, unsigned long caller_ip), \
 	TP_ARGS(ip, caller_ip))
-DEFINE_INODE_EVENT(xfs_ihold);
-DEFINE_INODE_EVENT(xfs_irele);
-DEFINE_INODE_EVENT(xfs_inode_pin);
-DEFINE_INODE_EVENT(xfs_inode_unpin);
-DEFINE_INODE_EVENT(xfs_inode_unpin_nowait);
-
-/* the old xfs_itrace_entry tracer - to be replaced by s.th. in the VFS */
-DEFINE_INODE_EVENT(xfs_inode);
-#define xfs_itrace_entry(ip)    \
-	trace_xfs_inode(ip, _THIS_IP_)
+DEFINE_IREF_EVENT(xfs_ihold);
+DEFINE_IREF_EVENT(xfs_irele);
+DEFINE_IREF_EVENT(xfs_inode_pin);
+DEFINE_IREF_EVENT(xfs_inode_unpin);
+DEFINE_IREF_EVENT(xfs_inode_unpin_nowait);
+
+DECLARE_EVENT_CLASS(xfs_namespace_class,
+	TP_PROTO(struct xfs_inode *dp, struct xfs_name *name),
+	TP_ARGS(dp, name),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(xfs_ino_t, dp_ino)
+		__dynamic_array(char, name, name->len)
+	),
+	TP_fast_assign(
+		__entry->dev = VFS_I(dp)->i_sb->s_dev;
+		__entry->dp_ino = dp->i_ino;
+		memcpy(__get_str(name), name->name, name->len);
+	),
+	TP_printk("dev %d:%d dp ino 0x%llx name %s",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->dp_ino,
+		  __get_str(name))
+)
+
+#define DEFINE_NAMESPACE_EVENT(name) \
+DEFINE_EVENT(xfs_namespace_class, name, \
+	TP_PROTO(struct xfs_inode *dp, struct xfs_name *name), \
+	TP_ARGS(dp, name))
+DEFINE_NAMESPACE_EVENT(xfs_remove);
+DEFINE_NAMESPACE_EVENT(xfs_link);
+DEFINE_NAMESPACE_EVENT(xfs_lookup);
+DEFINE_NAMESPACE_EVENT(xfs_create);
+DEFINE_NAMESPACE_EVENT(xfs_symlink);
+
+TRACE_EVENT(xfs_rename,
+	TP_PROTO(struct xfs_inode *src_dp, struct xfs_inode *target_dp,
+		 struct xfs_name *src_name, struct xfs_name *target_name),
+	TP_ARGS(src_dp, target_dp, src_name, target_name),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(xfs_ino_t, src_dp_ino)
+		__field(xfs_ino_t, target_dp_ino)
+		__dynamic_array(char, src_name, src_name->len)
+		__dynamic_array(char, target_name, target_name->len)
+	),
+	TP_fast_assign(
+		__entry->dev = VFS_I(src_dp)->i_sb->s_dev;
+		__entry->src_dp_ino = src_dp->i_ino;
+		__entry->target_dp_ino = target_dp->i_ino;
+		memcpy(__get_str(src_name), src_name->name, src_name->len);
+		memcpy(__get_str(target_name), target_name->name, target_name->len);
+	),
+	TP_printk("dev %d:%d src dp ino 0x%llx target dp ino 0x%llx"
+		  " src name %s target name %s",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->src_dp_ino,
+		  __entry->target_dp_ino,
+		  __get_str(src_name),
+		  __get_str(target_name))
+)
 
 DECLARE_EVENT_CLASS(xfs_dquot_class,
 	TP_PROTO(struct xfs_dquot *dqp),
@@ -680,9 +750,6 @@ DEFINE_DQUOT_EVENT(xfs_dqrele);
 DEFINE_DQUOT_EVENT(xfs_dqflush);
 DEFINE_DQUOT_EVENT(xfs_dqflush_force);
 DEFINE_DQUOT_EVENT(xfs_dqflush_done);
-/* not really iget events, but we re-use the format */
-DEFINE_IGET_EVENT(xfs_dquot_dqalloc);
-DEFINE_IGET_EVENT(xfs_dquot_dqdetach);
 
 DECLARE_EVENT_CLASS(xfs_loggrant_class,
 	TP_PROTO(struct log *log, struct xlog_ticket *tic),
Index: xfs/fs/xfs/xfs_dir2.c
===================================================================
--- xfs.orig/fs/xfs/xfs_dir2.c	2010-06-21 09:31:42.811254930 +0200
+++ xfs/fs/xfs/xfs_dir2.c	2010-06-21 09:38:38.840032692 +0200
@@ -380,7 +380,7 @@ xfs_readdir(
 	int		rval;		/* return value */
 	int		v;		/* type-checking value */
 
-	xfs_itrace_entry(dp);
+	trace_xfs_readdir(dp);
 
 	if (XFS_FORCED_SHUTDOWN(dp->i_mount))
 		return XFS_ERROR(EIO);
Index: xfs/fs/xfs/xfs_vnodeops.c
===================================================================
--- xfs.orig/fs/xfs/xfs_vnodeops.c	2010-06-21 09:31:47.890255838 +0200
+++ xfs/fs/xfs/xfs_vnodeops.c	2010-06-21 09:38:38.848255488 +0200
@@ -68,7 +68,7 @@ xfs_setattr(
 	struct xfs_dquot	*udqp, *gdqp, *olddquot1, *olddquot2;
 	int			need_iolock = 1;
 
-	xfs_itrace_entry(ip);
+	trace_xfs_setattr(ip);
 
 	if (mp->m_flags & XFS_MOUNT_RDONLY)
 		return XFS_ERROR(EROFS);
@@ -533,7 +533,7 @@ xfs_readlink(
 	int		pathlen;
 	int		error = 0;
 
-	xfs_itrace_entry(ip);
+	trace_xfs_readlink(ip);
 
 	if (XFS_FORCED_SHUTDOWN(mp))
 		return XFS_ERROR(EIO);
@@ -1005,8 +1005,6 @@ xfs_inactive(
 	int		error;
 	int		truncate;
 
-	xfs_itrace_entry(ip);
-
 	/*
 	 * If the inode is already free, then there can be nothing
 	 * to clean up here.
@@ -1221,7 +1219,7 @@ xfs_lookup(
 	int			error;
 	uint			lock_mode;
 
-	xfs_itrace_entry(dp);
+	trace_xfs_lookup(dp, name);
 
 	if (XFS_FORCED_SHUTDOWN(dp->i_mount))
 		return XFS_ERROR(EIO);
@@ -1273,7 +1271,7 @@ xfs_create(
 	uint			log_res;
 	uint			log_count;
 
-	xfs_itrace_entry(dp);
+	trace_xfs_create(dp, name);
 
 	if (XFS_FORCED_SHUTDOWN(mp))
 		return XFS_ERROR(EIO);
@@ -1670,8 +1668,7 @@ xfs_remove(
 	uint			resblks;
 	uint			log_count;
 
-	xfs_itrace_entry(dp);
-	xfs_itrace_entry(ip);
+	trace_xfs_remove(dp, name);
 
 	if (XFS_FORCED_SHUTDOWN(mp))
 		return XFS_ERROR(EIO);
@@ -1832,8 +1829,7 @@ xfs_link(
 	int			committed;
 	int			resblks;
 
-	xfs_itrace_entry(tdp);
-	xfs_itrace_entry(sip);
+	trace_xfs_link(tdp, target_name);
 
 	ASSERT(!S_ISDIR(sip->i_d.di_mode));
 
@@ -1966,7 +1962,7 @@ xfs_symlink(
 	ip = NULL;
 	tp = NULL;
 
-	xfs_itrace_entry(dp);
+	trace_xfs_symlink(dp, link_name);
 
 	if (XFS_FORCED_SHUTDOWN(mp))
 		return XFS_ERROR(EIO);
@@ -2256,7 +2252,7 @@ xfs_alloc_file_space(
 	int			committed;
 	int			error;
 
-	xfs_itrace_entry(ip);
+	trace_xfs_alloc_file_space(ip);
 
 	if (XFS_FORCED_SHUTDOWN(mp))
 		return XFS_ERROR(EIO);
@@ -2517,7 +2513,7 @@ xfs_free_file_space(
 
 	mp = ip->i_mount;
 
-	xfs_itrace_entry(ip);
+	trace_xfs_free_file_space(ip);
 
 	error = xfs_qm_dqattach(ip, 0);
 	if (error)
@@ -2707,8 +2703,6 @@ xfs_change_file_space(
 	xfs_trans_t	*tp;
 	struct iattr	iattr;
 
-	xfs_itrace_entry(ip);
-
 	if (!S_ISREG(ip->i_d.di_mode))
 		return XFS_ERROR(EINVAL);
 
Index: xfs/fs/xfs/xfs_rename.c
===================================================================
--- xfs.orig/fs/xfs/xfs_rename.c	2010-06-21 09:31:42.833254441 +0200
+++ xfs/fs/xfs/xfs_rename.c	2010-06-21 09:38:38.853285101 +0200
@@ -113,8 +113,7 @@ xfs_rename(
 	int		spaceres;
 	int		num_inodes;
 
-	xfs_itrace_entry(src_dp);
-	xfs_itrace_entry(target_dp);
+	trace_xfs_rename(src_dp, target_dp, src_name, target_name);
 
 	new_parent = (src_dp != target_dp);
 	src_is_directory = ((src_ip->i_d.di_mode & S_IFMT) == S_IFDIR);

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