From 60197e8df364df326dcbb987519f367ad0ee1a11 Mon Sep 17 00:00:00 2001
From: Christoph Hellwig <hch@infradead.org>
Date: Thu, 30 Oct 2008 17:11:19 +1100
Subject: [XFS] Cleanup maxrecs calculation.

Clean up the way the maximum and minimum records for the btree blocks are
calculated. For the alloc and inobt btrees all the values are
pre-calculated in xfs_mount_common, and we switch the current loop around
the ugly generic macros that use cpp token pasting to generate type names
to two small helpers in normal C code. For the bmbt and bmdr trees these
helpers also exist, but can be called during runtime, too. Here we also
kill various macros dealing with them and inline the logic into the
get_minrecs / get_maxrecs / get_dmaxrecs methods in xfs_bmap_btree.c.

Note that all these new helpers take an xfs_mount * argument which will be
needed to determine the size of a btree block once we add support for
extended btree blocks with CRCs and other RAS information.

SGI-PV: 988146

SGI-Modid: xfs-linux-melb:xfs-kern:32292a

Signed-off-by: Christoph Hellwig <hch@infradead.org>
Signed-off-by: Donald Douwsma <donaldd@sgi.com>
Signed-off-by: Lachlan McIlroy <lachlan@sgi.com>
---
 fs/xfs/xfs_log_recover.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

(limited to 'fs/xfs/xfs_log_recover.c')

diff --git a/fs/xfs/xfs_log_recover.c b/fs/xfs/xfs_log_recover.c
index 82d46ce69d5f..23c3a782a9e7 100644
--- a/fs/xfs/xfs_log_recover.c
+++ b/fs/xfs/xfs_log_recover.c
@@ -2452,7 +2452,7 @@ xlog_recover_do_inode_trans(
 		break;
 
 	case XFS_ILOG_DBROOT:
-		xfs_bmbt_to_bmdr((xfs_bmbt_block_t *)src, len,
+		xfs_bmbt_to_bmdr(mp, (xfs_bmbt_block_t *)src, len,
 				 &(dip->di_u.di_bmbt),
 				 XFS_DFORK_DSIZE(dip, mp));
 		break;
@@ -2490,7 +2490,7 @@ xlog_recover_do_inode_trans(
 
 		case XFS_ILOG_ABROOT:
 			dest = XFS_DFORK_APTR(dip);
-			xfs_bmbt_to_bmdr((xfs_bmbt_block_t *)src, len,
+			xfs_bmbt_to_bmdr(mp, (xfs_bmbt_block_t *)src, len,
 					 (xfs_bmdr_block_t*)dest,
 					 XFS_DFORK_ASIZE(dip, mp));
 			break;
-- 
cgit v1.2.2


From 7cc95a821df8f09a5d37a923cf8c3a7c3ee00c29 Mon Sep 17 00:00:00 2001
From: Christoph Hellwig <hch@infradead.org>
Date: Thu, 30 Oct 2008 17:14:34 +1100
Subject: [XFS] Always use struct xfs_btree_block instead of short / longform
 structures.

Always use the generic xfs_btree_block type instead of the short / long
structures. Add XFS_BTREE_SBLOCK_LEN / XFS_BTREE_LBLOCK_LEN defines for
the length of a short / long form block. The rationale for this is that we
will grow more btree block header variants to support CRCs and other RAS
information, and always accessing them through the same datatype with
unions for the short / long form pointers makes implementing this much
easier.

SGI-PV: 988146

SGI-Modid: xfs-linux-melb:xfs-kern:32300a

Signed-off-by: Christoph Hellwig <hch@infradead.org>
Signed-off-by: Donald Douwsma <donaldd@sgi.com>
Signed-off-by: David Chinner <david@fromorbit.com>
Signed-off-by: Lachlan McIlroy <lachlan@sgi.com>
---
 fs/xfs/xfs_log_recover.c | 8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)

(limited to 'fs/xfs/xfs_log_recover.c')

diff --git a/fs/xfs/xfs_log_recover.c b/fs/xfs/xfs_log_recover.c
index 23c3a782a9e7..199c8ea36474 100644
--- a/fs/xfs/xfs_log_recover.c
+++ b/fs/xfs/xfs_log_recover.c
@@ -2452,8 +2452,8 @@ xlog_recover_do_inode_trans(
 		break;
 
 	case XFS_ILOG_DBROOT:
-		xfs_bmbt_to_bmdr(mp, (xfs_bmbt_block_t *)src, len,
-				 &(dip->di_u.di_bmbt),
+		xfs_bmbt_to_bmdr(mp, (struct xfs_btree_block *)src, len,
+				 &dip->di_u.di_bmbt,
 				 XFS_DFORK_DSIZE(dip, mp));
 		break;
 
@@ -2490,8 +2490,8 @@ xlog_recover_do_inode_trans(
 
 		case XFS_ILOG_ABROOT:
 			dest = XFS_DFORK_APTR(dip);
-			xfs_bmbt_to_bmdr(mp, (xfs_bmbt_block_t *)src, len,
-					 (xfs_bmdr_block_t*)dest,
+			xfs_bmbt_to_bmdr(mp, (struct xfs_btree_block *)src,
+					 len, (xfs_bmdr_block_t*)dest,
 					 XFS_DFORK_ASIZE(dip, mp));
 			break;
 
-- 
cgit v1.2.2


From 27d8d5fe0ef9daeaafbdd32b14b32a2211930062 Mon Sep 17 00:00:00 2001
From: David Chinner <david@fromorbit.com>
Date: Thu, 30 Oct 2008 17:38:39 +1100
Subject: [XFS] Use a cursor for AIL traversal.

To replace the current generation number ensuring sanity of the AIL
traversal, replace it with an external cursor that is linked to the AIL.

Basically, we store the next item in the cursor whenever we want to drop
the AIL lock to do something to the current item. When we regain the lock.
the current item may already be free, so we can't reference it, but the
next item in the traversal is already held in the cursor.

When we move or delete an object, we search all the active cursors and if
there is an item match we clear the cursor(s) that point to the object.
This forces the traversal to restart transparently.

We don't invalidate the cursor on insert because the cursor still points
to a valid item. If the intem is inserted between the current item and the
cursor it does not matter; the traversal is considered to be past the
insertion point so it will be picked up in the next traversal.

Hence traversal restarts pretty much disappear altogether with this method
of traversal, which should substantially reduce the overhead of pushing on
a busy AIL.

Version 2 o add restart logic o comment cursor interface o minor cleanups

SGI-PV: 988143

SGI-Modid: xfs-linux-melb:xfs-kern:32347a

Signed-off-by: David Chinner <david@fromorbit.com>
Signed-off-by: Lachlan McIlroy <lachlan@sgi.com>
Signed-off-by: Christoph Hellwig <hch@infradead.org>
---
 fs/xfs/xfs_log_recover.c | 61 +++++++++++++++++-------------------------------
 1 file changed, 21 insertions(+), 40 deletions(-)

(limited to 'fs/xfs/xfs_log_recover.c')

diff --git a/fs/xfs/xfs_log_recover.c b/fs/xfs/xfs_log_recover.c
index 199c8ea36474..37ba4899f3e6 100644
--- a/fs/xfs/xfs_log_recover.c
+++ b/fs/xfs/xfs_log_recover.c
@@ -54,10 +54,8 @@ STATIC void	xlog_recover_insert_item_backq(xlog_recover_item_t **q,
 					       xlog_recover_item_t *item);
 #if defined(DEBUG)
 STATIC void	xlog_recover_check_summary(xlog_t *);
-STATIC void	xlog_recover_check_ail(xfs_mount_t *, xfs_log_item_t *, int);
 #else
 #define	xlog_recover_check_summary(log)
-#define	xlog_recover_check_ail(mp, lip, gen)
 #endif
 
 
@@ -2710,8 +2708,8 @@ xlog_recover_do_efd_trans(
 	xfs_efd_log_format_t	*efd_formatp;
 	xfs_efi_log_item_t	*efip = NULL;
 	xfs_log_item_t		*lip;
-	int			gen;
 	__uint64_t		efi_id;
+	struct xfs_ail_cursor	cur;
 
 	if (pass == XLOG_RECOVER_PASS1) {
 		return;
@@ -2730,7 +2728,8 @@ xlog_recover_do_efd_trans(
 	 */
 	mp = log->l_mp;
 	spin_lock(&mp->m_ail_lock);
-	lip = xfs_trans_first_ail(mp, &gen);
+	xfs_trans_ail_cursor_init(mp->m_ail, &cur);
+	lip = xfs_trans_first_ail(mp, &cur);
 	while (lip != NULL) {
 		if (lip->li_type == XFS_LI_EFI) {
 			efip = (xfs_efi_log_item_t *)lip;
@@ -2741,11 +2740,13 @@ xlog_recover_do_efd_trans(
 				 */
 				xfs_trans_delete_ail(mp, lip);
 				xfs_efi_item_free(efip);
-				return;
+				spin_lock(&mp->m_ail_lock);
+				break;
 			}
 		}
-		lip = xfs_trans_next_ail(mp, lip, &gen, NULL);
+		lip = xfs_trans_next_ail(mp, &cur);
 	}
+	xfs_trans_ail_cursor_done(mp->m_ail, &cur);
 	spin_unlock(&mp->m_ail_lock);
 }
 
@@ -3029,33 +3030,6 @@ abort_error:
 	return error;
 }
 
-/*
- * Verify that once we've encountered something other than an EFI
- * in the AIL that there are no more EFIs in the AIL.
- */
-#if defined(DEBUG)
-STATIC void
-xlog_recover_check_ail(
-	xfs_mount_t		*mp,
-	xfs_log_item_t		*lip,
-	int			gen)
-{
-	int			orig_gen = gen;
-
-	do {
-		ASSERT(lip->li_type != XFS_LI_EFI);
-		lip = xfs_trans_next_ail(mp, lip, &gen, NULL);
-		/*
-		 * The check will be bogus if we restart from the
-		 * beginning of the AIL, so ASSERT that we don't.
-		 * We never should since we're holding the AIL lock
-		 * the entire time.
-		 */
-		ASSERT(gen == orig_gen);
-	} while (lip != NULL);
-}
-#endif	/* DEBUG */
-
 /*
  * When this is called, all of the EFIs which did not have
  * corresponding EFDs should be in the AIL.  What we do now
@@ -3080,20 +3054,25 @@ xlog_recover_process_efis(
 {
 	xfs_log_item_t		*lip;
 	xfs_efi_log_item_t	*efip;
-	int			gen;
 	xfs_mount_t		*mp;
 	int			error = 0;
+	struct xfs_ail_cursor	cur;
 
 	mp = log->l_mp;
 	spin_lock(&mp->m_ail_lock);
 
-	lip = xfs_trans_first_ail(mp, &gen);
+	xfs_trans_ail_cursor_init(mp->m_ail, &cur);
+	lip = xfs_trans_first_ail(mp, &cur);
 	while (lip != NULL) {
 		/*
 		 * We're done when we see something other than an EFI.
+		 * There should be no EFIs left in the AIL now.
 		 */
 		if (lip->li_type != XFS_LI_EFI) {
-			xlog_recover_check_ail(mp, lip, gen);
+#ifdef DEBUG
+			for (; lip; lip = xfs_trans_next_ail(mp, &cur))
+				ASSERT(lip->li_type != XFS_LI_EFI);
+#endif
 			break;
 		}
 
@@ -3102,17 +3081,19 @@ xlog_recover_process_efis(
 		 */
 		efip = (xfs_efi_log_item_t *)lip;
 		if (efip->efi_flags & XFS_EFI_RECOVERED) {
-			lip = xfs_trans_next_ail(mp, lip, &gen, NULL);
+			lip = xfs_trans_next_ail(mp, &cur);
 			continue;
 		}
 
 		spin_unlock(&mp->m_ail_lock);
 		error = xlog_recover_process_efi(mp, efip);
-		if (error)
-			return error;
 		spin_lock(&mp->m_ail_lock);
-		lip = xfs_trans_next_ail(mp, lip, &gen, NULL);
+		if (error)
+			goto out;
+		lip = xfs_trans_next_ail(mp, &cur);
 	}
+out:
+	xfs_trans_ail_cursor_done(mp->m_ail, &cur);
 	spin_unlock(&mp->m_ail_lock);
 	return error;
 }
-- 
cgit v1.2.2


From 5b00f14fbd60d42441f78c0e414a539cbfba5cb9 Mon Sep 17 00:00:00 2001
From: David Chinner <david@fromorbit.com>
Date: Thu, 30 Oct 2008 17:39:00 +1100
Subject: [XFS] move the AIl traversal over to a consistent interface

With the new cursor interface, it makes sense to make all the traversing
code use the cursor interface and make the old one go away. This means
more of the AIL interfacing is done by passing struct xfs_ail pointers
around the place instead of struct xfs_mount pointers.

We can replace the use of xfs_trans_first_ail() in xfs_log_need_covered()
as it is only checking if the AIL is empty. We can do that with a call to
xfs_trans_ail_tail() instead, where a zero LSN returned indicates and
empty AIL...

SGI-PV: 988143

SGI-Modid: xfs-linux-melb:xfs-kern:32348a

Signed-off-by: David Chinner <david@fromorbit.com>
Signed-off-by: Lachlan McIlroy <lachlan@sgi.com>
Signed-off-by: Christoph Hellwig <hch@infradead.org>
---
 fs/xfs/xfs_log_recover.c | 15 +++++++--------
 1 file changed, 7 insertions(+), 8 deletions(-)

(limited to 'fs/xfs/xfs_log_recover.c')

diff --git a/fs/xfs/xfs_log_recover.c b/fs/xfs/xfs_log_recover.c
index 37ba4899f3e6..45ea0d950138 100644
--- a/fs/xfs/xfs_log_recover.c
+++ b/fs/xfs/xfs_log_recover.c
@@ -2728,8 +2728,7 @@ xlog_recover_do_efd_trans(
 	 */
 	mp = log->l_mp;
 	spin_lock(&mp->m_ail_lock);
-	xfs_trans_ail_cursor_init(mp->m_ail, &cur);
-	lip = xfs_trans_first_ail(mp, &cur);
+	lip = xfs_trans_ail_cursor_first(mp->m_ail, &cur, 0);
 	while (lip != NULL) {
 		if (lip->li_type == XFS_LI_EFI) {
 			efip = (xfs_efi_log_item_t *)lip;
@@ -2744,7 +2743,7 @@ xlog_recover_do_efd_trans(
 				break;
 			}
 		}
-		lip = xfs_trans_next_ail(mp, &cur);
+		lip = xfs_trans_ail_cursor_next(mp->m_ail, &cur);
 	}
 	xfs_trans_ail_cursor_done(mp->m_ail, &cur);
 	spin_unlock(&mp->m_ail_lock);
@@ -3061,8 +3060,7 @@ xlog_recover_process_efis(
 	mp = log->l_mp;
 	spin_lock(&mp->m_ail_lock);
 
-	xfs_trans_ail_cursor_init(mp->m_ail, &cur);
-	lip = xfs_trans_first_ail(mp, &cur);
+	lip = xfs_trans_ail_cursor_first(mp->m_ail, &cur, 0);
 	while (lip != NULL) {
 		/*
 		 * We're done when we see something other than an EFI.
@@ -3070,7 +3068,8 @@ xlog_recover_process_efis(
 		 */
 		if (lip->li_type != XFS_LI_EFI) {
 #ifdef DEBUG
-			for (; lip; lip = xfs_trans_next_ail(mp, &cur))
+			for (; lip;
+			       lip = xfs_trans_ail_cursor_next(mp->m_ail, &cur))
 				ASSERT(lip->li_type != XFS_LI_EFI);
 #endif
 			break;
@@ -3081,7 +3080,7 @@ xlog_recover_process_efis(
 		 */
 		efip = (xfs_efi_log_item_t *)lip;
 		if (efip->efi_flags & XFS_EFI_RECOVERED) {
-			lip = xfs_trans_next_ail(mp, &cur);
+			lip = xfs_trans_ail_cursor_next(mp->m_ail, &cur);
 			continue;
 		}
 
@@ -3090,7 +3089,7 @@ xlog_recover_process_efis(
 		spin_lock(&mp->m_ail_lock);
 		if (error)
 			goto out;
-		lip = xfs_trans_next_ail(mp, &cur);
+		lip = xfs_trans_ail_cursor_next(mp->m_ail, &cur);
 	}
 out:
 	xfs_trans_ail_cursor_done(mp->m_ail, &cur);
-- 
cgit v1.2.2


From c7e8f268278a292d3823b4352182fa7755a71410 Mon Sep 17 00:00:00 2001
From: David Chinner <david@fromorbit.com>
Date: Thu, 30 Oct 2008 17:39:23 +1100
Subject: [XFS] Move the AIL lock into the struct xfs_ail

Bring the ail lock inside the struct xfs_ail. This means the AIL can be
entirely manipulated via the struct xfs_ail rather than needing both the
struct xfs_mount and the struct xfs_ail.

SGI-PV: 988143

SGI-Modid: xfs-linux-melb:xfs-kern:32350a

Signed-off-by: David Chinner <david@fromorbit.com>
Signed-off-by: Lachlan McIlroy <lachlan@sgi.com>
Signed-off-by: Christoph Hellwig <hch@infradead.org>
---
 fs/xfs/xfs_log_recover.c | 16 ++++++++--------
 1 file changed, 8 insertions(+), 8 deletions(-)

(limited to 'fs/xfs/xfs_log_recover.c')

diff --git a/fs/xfs/xfs_log_recover.c b/fs/xfs/xfs_log_recover.c
index 45ea0d950138..a484febb9ec6 100644
--- a/fs/xfs/xfs_log_recover.c
+++ b/fs/xfs/xfs_log_recover.c
@@ -2681,7 +2681,7 @@ xlog_recover_do_efi_trans(
 	efip->efi_next_extent = efi_formatp->efi_nextents;
 	efip->efi_flags |= XFS_EFI_COMMITTED;
 
-	spin_lock(&mp->m_ail_lock);
+	spin_lock(&mp->m_ail->xa_lock);
 	/*
 	 * xfs_trans_update_ail() drops the AIL lock.
 	 */
@@ -2727,7 +2727,7 @@ xlog_recover_do_efd_trans(
 	 * in the AIL.
 	 */
 	mp = log->l_mp;
-	spin_lock(&mp->m_ail_lock);
+	spin_lock(&mp->m_ail->xa_lock);
 	lip = xfs_trans_ail_cursor_first(mp->m_ail, &cur, 0);
 	while (lip != NULL) {
 		if (lip->li_type == XFS_LI_EFI) {
@@ -2739,14 +2739,14 @@ xlog_recover_do_efd_trans(
 				 */
 				xfs_trans_delete_ail(mp, lip);
 				xfs_efi_item_free(efip);
-				spin_lock(&mp->m_ail_lock);
+				spin_lock(&mp->m_ail->xa_lock);
 				break;
 			}
 		}
 		lip = xfs_trans_ail_cursor_next(mp->m_ail, &cur);
 	}
 	xfs_trans_ail_cursor_done(mp->m_ail, &cur);
-	spin_unlock(&mp->m_ail_lock);
+	spin_unlock(&mp->m_ail->xa_lock);
 }
 
 /*
@@ -3058,7 +3058,7 @@ xlog_recover_process_efis(
 	struct xfs_ail_cursor	cur;
 
 	mp = log->l_mp;
-	spin_lock(&mp->m_ail_lock);
+	spin_lock(&mp->m_ail->xa_lock);
 
 	lip = xfs_trans_ail_cursor_first(mp->m_ail, &cur, 0);
 	while (lip != NULL) {
@@ -3084,16 +3084,16 @@ xlog_recover_process_efis(
 			continue;
 		}
 
-		spin_unlock(&mp->m_ail_lock);
+		spin_unlock(&mp->m_ail->xa_lock);
 		error = xlog_recover_process_efi(mp, efip);
-		spin_lock(&mp->m_ail_lock);
+		spin_lock(&mp->m_ail->xa_lock);
 		if (error)
 			goto out;
 		lip = xfs_trans_ail_cursor_next(mp->m_ail, &cur);
 	}
 out:
 	xfs_trans_ail_cursor_done(mp->m_ail, &cur);
-	spin_unlock(&mp->m_ail_lock);
+	spin_unlock(&mp->m_ail->xa_lock);
 	return error;
 }
 
-- 
cgit v1.2.2


From a9c21c1b9deaced836034e77fe25fe0b55c21f02 Mon Sep 17 00:00:00 2001
From: David Chinner <david@fromorbit.com>
Date: Thu, 30 Oct 2008 17:39:35 +1100
Subject: [XFS] Given the log a pointer to the AIL

When we need to go from the log to the AIL, we have to go via the
xfs_mount. Add a xfs_ail pointer to the log so we can go directly to the
AIL associated with the log.

SGI-PV: 988143

SGI-Modid: xfs-linux-melb:xfs-kern:32351a

Signed-off-by: David Chinner <david@fromorbit.com>
Signed-off-by: Lachlan McIlroy <lachlan@sgi.com>
Signed-off-by: Christoph Hellwig <hch@infradead.org>
---
 fs/xfs/xfs_log_recover.c | 42 +++++++++++++++++++++---------------------
 1 file changed, 21 insertions(+), 21 deletions(-)

(limited to 'fs/xfs/xfs_log_recover.c')

diff --git a/fs/xfs/xfs_log_recover.c b/fs/xfs/xfs_log_recover.c
index a484febb9ec6..0bbde7b84fc9 100644
--- a/fs/xfs/xfs_log_recover.c
+++ b/fs/xfs/xfs_log_recover.c
@@ -2681,7 +2681,7 @@ xlog_recover_do_efi_trans(
 	efip->efi_next_extent = efi_formatp->efi_nextents;
 	efip->efi_flags |= XFS_EFI_COMMITTED;
 
-	spin_lock(&mp->m_ail->xa_lock);
+	spin_lock(&log->l_ailp->xa_lock);
 	/*
 	 * xfs_trans_update_ail() drops the AIL lock.
 	 */
@@ -2710,6 +2710,7 @@ xlog_recover_do_efd_trans(
 	xfs_log_item_t		*lip;
 	__uint64_t		efi_id;
 	struct xfs_ail_cursor	cur;
+	struct xfs_ail		*ailp;
 
 	if (pass == XLOG_RECOVER_PASS1) {
 		return;
@@ -2727,8 +2728,9 @@ xlog_recover_do_efd_trans(
 	 * in the AIL.
 	 */
 	mp = log->l_mp;
-	spin_lock(&mp->m_ail->xa_lock);
-	lip = xfs_trans_ail_cursor_first(mp->m_ail, &cur, 0);
+	ailp = log->l_ailp;
+	spin_lock(&ailp->xa_lock);
+	lip = xfs_trans_ail_cursor_first(ailp, &cur, 0);
 	while (lip != NULL) {
 		if (lip->li_type == XFS_LI_EFI) {
 			efip = (xfs_efi_log_item_t *)lip;
@@ -2739,14 +2741,14 @@ xlog_recover_do_efd_trans(
 				 */
 				xfs_trans_delete_ail(mp, lip);
 				xfs_efi_item_free(efip);
-				spin_lock(&mp->m_ail->xa_lock);
+				spin_lock(&ailp->xa_lock);
 				break;
 			}
 		}
-		lip = xfs_trans_ail_cursor_next(mp->m_ail, &cur);
+		lip = xfs_trans_ail_cursor_next(ailp, &cur);
 	}
-	xfs_trans_ail_cursor_done(mp->m_ail, &cur);
-	spin_unlock(&mp->m_ail->xa_lock);
+	xfs_trans_ail_cursor_done(ailp, &cur);
+	spin_unlock(&ailp->xa_lock);
 }
 
 /*
@@ -3053,14 +3055,13 @@ xlog_recover_process_efis(
 {
 	xfs_log_item_t		*lip;
 	xfs_efi_log_item_t	*efip;
-	xfs_mount_t		*mp;
 	int			error = 0;
 	struct xfs_ail_cursor	cur;
+	struct xfs_ail		*ailp;
 
-	mp = log->l_mp;
-	spin_lock(&mp->m_ail->xa_lock);
-
-	lip = xfs_trans_ail_cursor_first(mp->m_ail, &cur, 0);
+	ailp = log->l_ailp;
+	spin_lock(&ailp->xa_lock);
+	lip = xfs_trans_ail_cursor_first(ailp, &cur, 0);
 	while (lip != NULL) {
 		/*
 		 * We're done when we see something other than an EFI.
@@ -3068,8 +3069,7 @@ xlog_recover_process_efis(
 		 */
 		if (lip->li_type != XFS_LI_EFI) {
 #ifdef DEBUG
-			for (; lip;
-			       lip = xfs_trans_ail_cursor_next(mp->m_ail, &cur))
+			for (; lip; lip = xfs_trans_ail_cursor_next(ailp, &cur))
 				ASSERT(lip->li_type != XFS_LI_EFI);
 #endif
 			break;
@@ -3080,20 +3080,20 @@ xlog_recover_process_efis(
 		 */
 		efip = (xfs_efi_log_item_t *)lip;
 		if (efip->efi_flags & XFS_EFI_RECOVERED) {
-			lip = xfs_trans_ail_cursor_next(mp->m_ail, &cur);
+			lip = xfs_trans_ail_cursor_next(ailp, &cur);
 			continue;
 		}
 
-		spin_unlock(&mp->m_ail->xa_lock);
-		error = xlog_recover_process_efi(mp, efip);
-		spin_lock(&mp->m_ail->xa_lock);
+		spin_unlock(&ailp->xa_lock);
+		error = xlog_recover_process_efi(log->l_mp, efip);
+		spin_lock(&ailp->xa_lock);
 		if (error)
 			goto out;
-		lip = xfs_trans_ail_cursor_next(mp->m_ail, &cur);
+		lip = xfs_trans_ail_cursor_next(ailp, &cur);
 	}
 out:
-	xfs_trans_ail_cursor_done(mp->m_ail, &cur);
-	spin_unlock(&mp->m_ail->xa_lock);
+	xfs_trans_ail_cursor_done(ailp, &cur);
+	spin_unlock(&ailp->xa_lock);
 	return error;
 }
 
-- 
cgit v1.2.2


From 783a2f656f9674c31d4019708a94af93fa1d1c22 Mon Sep 17 00:00:00 2001
From: David Chinner <david@fromorbit.com>
Date: Thu, 30 Oct 2008 17:39:58 +1100
Subject: [XFS] Finish removing the mount pointer from the AIL API

Change all the remaining AIL API functions that are passed struct
xfs_mount pointers to pass pointers directly to the struct xfs_ail being
used. With this conversion, all external access to the AIL is via the
struct xfs_ail. Hence the operation and referencing of the AIL is almost
entirely independent of the xfs_mount that is using it - it is now much
more tightly tied to the log and the items it is tracking in the log than
it is tied to the xfs_mount.

SGI-PV: 988143

SGI-Modid: xfs-linux-melb:xfs-kern:32353a

Signed-off-by: David Chinner <david@fromorbit.com>
Signed-off-by: Lachlan McIlroy <lachlan@sgi.com>
Signed-off-by: Christoph Hellwig <hch@infradead.org>
---
 fs/xfs/xfs_log_recover.c | 13 +++++--------
 1 file changed, 5 insertions(+), 8 deletions(-)

(limited to 'fs/xfs/xfs_log_recover.c')

diff --git a/fs/xfs/xfs_log_recover.c b/fs/xfs/xfs_log_recover.c
index 0bbde7b84fc9..cff901efc24b 100644
--- a/fs/xfs/xfs_log_recover.c
+++ b/fs/xfs/xfs_log_recover.c
@@ -2683,9 +2683,9 @@ xlog_recover_do_efi_trans(
 
 	spin_lock(&log->l_ailp->xa_lock);
 	/*
-	 * xfs_trans_update_ail() drops the AIL lock.
+	 * xfs_trans_ail_update() drops the AIL lock.
 	 */
-	xfs_trans_update_ail(mp, (xfs_log_item_t *)efip, lsn);
+	xfs_trans_ail_update(log->l_ailp, (xfs_log_item_t *)efip, lsn);
 	return 0;
 }
 
@@ -2704,13 +2704,12 @@ xlog_recover_do_efd_trans(
 	xlog_recover_item_t	*item,
 	int			pass)
 {
-	xfs_mount_t		*mp;
 	xfs_efd_log_format_t	*efd_formatp;
 	xfs_efi_log_item_t	*efip = NULL;
 	xfs_log_item_t		*lip;
 	__uint64_t		efi_id;
 	struct xfs_ail_cursor	cur;
-	struct xfs_ail		*ailp;
+	struct xfs_ail		*ailp = log->l_ailp;
 
 	if (pass == XLOG_RECOVER_PASS1) {
 		return;
@@ -2727,8 +2726,6 @@ xlog_recover_do_efd_trans(
 	 * Search for the efi with the id in the efd format structure
 	 * in the AIL.
 	 */
-	mp = log->l_mp;
-	ailp = log->l_ailp;
 	spin_lock(&ailp->xa_lock);
 	lip = xfs_trans_ail_cursor_first(ailp, &cur, 0);
 	while (lip != NULL) {
@@ -2736,10 +2733,10 @@ xlog_recover_do_efd_trans(
 			efip = (xfs_efi_log_item_t *)lip;
 			if (efip->efi_format.efi_id == efi_id) {
 				/*
-				 * xfs_trans_delete_ail() drops the
+				 * xfs_trans_ail_delete() drops the
 				 * AIL lock.
 				 */
-				xfs_trans_delete_ail(mp, lip);
+				xfs_trans_ail_delete(ailp, lip);
 				xfs_efi_item_free(efip);
 				spin_lock(&ailp->xa_lock);
 				break;
-- 
cgit v1.2.2


From 5a792c4579af8466246408e38fd4eff45d8493b8 Mon Sep 17 00:00:00 2001
From: David Chinner <david@fromorbit.com>
Date: Thu, 30 Oct 2008 17:40:09 +1100
Subject: [XFS] XFS: Check for valid transaction headers in recovery

When we are about to add a new item to a transaction in recovery, we need
to check that it is valid first. Currently we just assert that header
magic number matches, but in production systems that is not present and we
add a corrupted transaction to the list to be processed. This results in a
kernel oops later when processing the corrupted transaction.

Instead, if we detect a corrupted transaction, abort recovery and leave
the user to clean up the mess that has occurred.

SGI-PV: 988145

SGI-Modid: xfs-linux-melb:xfs-kern:32356a

Signed-off-by: David Chinner <david@fromorbit.com>
Signed-off-by: Tim Shimmin <tes@sgi.com>
Signed-off-by: Eric Sandeen <sandeen@sandeen.net>
Signed-off-by: Lachlan McIlroy <lachlan@sgi.com>
---
 fs/xfs/xfs_log_recover.c | 8 +++++++-
 1 file changed, 7 insertions(+), 1 deletion(-)

(limited to 'fs/xfs/xfs_log_recover.c')

diff --git a/fs/xfs/xfs_log_recover.c b/fs/xfs/xfs_log_recover.c
index cff901efc24b..b411d4947318 100644
--- a/fs/xfs/xfs_log_recover.c
+++ b/fs/xfs/xfs_log_recover.c
@@ -1417,7 +1417,13 @@ xlog_recover_add_to_trans(
 		return 0;
 	item = trans->r_itemq;
 	if (item == NULL) {
-		ASSERT(*(uint *)dp == XFS_TRANS_HEADER_MAGIC);
+		/* we need to catch log corruptions here */
+		if (*(uint *)dp != XFS_TRANS_HEADER_MAGIC) {
+			xlog_warn("XFS: xlog_recover_add_to_trans: "
+				  "bad header magic number");
+			ASSERT(0);
+			return XFS_ERROR(EIO);
+		}
 		if (len == sizeof(xfs_trans_header_t))
 			xlog_recover_add_item(&trans->r_itemq);
 		memcpy(&trans->r_theader, dp, len); /* d, s, l */
-- 
cgit v1.2.2


From 5e1be0fb1a3950597aeda448698e85b0595a2e92 Mon Sep 17 00:00:00 2001
From: Christoph Hellwig <hch@lst.de>
Date: Fri, 28 Nov 2008 14:23:37 +1100
Subject: [XFS] factor out xfs_read_agi helper

Add a helper to read the AGI header and perform basic verification.
Based on hunks from a larger patch from Dave Chinner.

(First sent on Juli 23rd)

Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Dave Chinner <david@fromorbit.com>
Signed-off-by: Niv Sardi <xaiki@sgi.com>
---
 fs/xfs/xfs_log_recover.c | 72 +++++++++++++++++-------------------------------
 1 file changed, 26 insertions(+), 46 deletions(-)

(limited to 'fs/xfs/xfs_log_recover.c')

diff --git a/fs/xfs/xfs_log_recover.c b/fs/xfs/xfs_log_recover.c
index b411d4947318..b552676ca5c4 100644
--- a/fs/xfs/xfs_log_recover.c
+++ b/fs/xfs/xfs_log_recover.c
@@ -3117,19 +3117,16 @@ xlog_recover_clear_agi_bucket(
 	int		error;
 
 	tp = xfs_trans_alloc(mp, XFS_TRANS_CLEAR_AGI_BUCKET);
-	error = xfs_trans_reserve(tp, 0, XFS_CLEAR_AGI_BUCKET_LOG_RES(mp), 0, 0, 0);
-	if (!error)
-		error = xfs_trans_read_buf(mp, tp, mp->m_ddev_targp,
-				   XFS_AG_DADDR(mp, agno, XFS_AGI_DADDR(mp)),
-				   XFS_FSS_TO_BB(mp, 1), 0, &agibp);
+	error = xfs_trans_reserve(tp, 0, XFS_CLEAR_AGI_BUCKET_LOG_RES(mp),
+				  0, 0, 0);
 	if (error)
 		goto out_abort;
 
-	error = EINVAL;
-	agi = XFS_BUF_TO_AGI(agibp);
-	if (be32_to_cpu(agi->agi_magicnum) != XFS_AGI_MAGIC)
+	error = xfs_read_agi(mp, tp, agno, &agibp);
+	if (error)
 		goto out_abort;
 
+	agi = XFS_BUF_TO_AGI(agibp);
 	agi->agi_unlinked[bucket] = cpu_to_be32(NULLAGINO);
 	offset = offsetof(xfs_agi_t, agi_unlinked) +
 		 (sizeof(xfs_agino_t) * bucket);
@@ -3190,16 +3187,17 @@ xlog_recover_process_iunlinks(
 		/*
 		 * Find the agi for this ag.
 		 */
-		agibp = xfs_buf_read(mp->m_ddev_targp,
-				XFS_AG_DADDR(mp, agno, XFS_AGI_DADDR(mp)),
-				XFS_FSS_TO_BB(mp, 1), 0);
-		if (XFS_BUF_ISERROR(agibp)) {
-			xfs_ioerror_alert("xlog_recover_process_iunlinks(#1)",
-				log->l_mp, agibp,
-				XFS_AG_DADDR(mp, agno, XFS_AGI_DADDR(mp)));
+		error = xfs_read_agi(mp, NULL, agno, &agibp);
+		if (error) {
+			/*
+			 * AGI is b0rked. Don't process it.
+			 *
+			 * We should probably mark the filesystem as corrupt
+			 * after we've recovered all the ag's we can....
+			 */
+			continue;
 		}
 		agi = XFS_BUF_TO_AGI(agibp);
-		ASSERT(XFS_AGI_MAGIC == be32_to_cpu(agi->agi_magicnum));
 
 		for (bucket = 0; bucket < XFS_AGI_UNLINKED_BUCKETS; bucket++) {
 
@@ -3278,22 +3276,12 @@ xlog_recover_process_iunlinks(
 
 				/*
 				 * Reacquire the agibuffer and continue around
-				 * the loop.
+				 * the loop. This should never fail as we know
+				 * the buffer was good earlier on.
 				 */
-				agibp = xfs_buf_read(mp->m_ddev_targp,
-						XFS_AG_DADDR(mp, agno,
-							XFS_AGI_DADDR(mp)),
-						XFS_FSS_TO_BB(mp, 1), 0);
-				if (XFS_BUF_ISERROR(agibp)) {
-					xfs_ioerror_alert(
-				"xlog_recover_process_iunlinks(#2)",
-						log->l_mp, agibp,
-						XFS_AG_DADDR(mp, agno,
-							XFS_AGI_DADDR(mp)));
-				}
+				error = xfs_read_agi(mp, NULL, agno, &agibp);
+				ASSERT(error == 0);
 				agi = XFS_BUF_TO_AGI(agibp);
-				ASSERT(XFS_AGI_MAGIC == be32_to_cpu(
-					agi->agi_magicnum));
 			}
 		}
 
@@ -3980,11 +3968,9 @@ xlog_recover_check_summary(
 {
 	xfs_mount_t	*mp;
 	xfs_agf_t	*agfp;
-	xfs_agi_t	*agip;
 	xfs_buf_t	*agfbp;
 	xfs_buf_t	*agibp;
 	xfs_daddr_t	agfdaddr;
-	xfs_daddr_t	agidaddr;
 	xfs_buf_t	*sbbp;
 #ifdef XFS_LOUD_RECOVERY
 	xfs_sb_t	*sbp;
@@ -3993,6 +3979,7 @@ xlog_recover_check_summary(
 	__uint64_t	freeblks;
 	__uint64_t	itotal;
 	__uint64_t	ifree;
+	int		error;
 
 	mp = log->l_mp;
 
@@ -4016,21 +4003,14 @@ xlog_recover_check_summary(
 			    be32_to_cpu(agfp->agf_flcount);
 		xfs_buf_relse(agfbp);
 
-		agidaddr = XFS_AG_DADDR(mp, agno, XFS_AGI_DADDR(mp));
-		agibp = xfs_buf_read(mp->m_ddev_targp, agidaddr,
-				XFS_FSS_TO_BB(mp, 1), 0);
-		if (XFS_BUF_ISERROR(agibp)) {
-			xfs_ioerror_alert("xlog_recover_check_summary(agi)",
-					  mp, agibp, agidaddr);
-		}
-		agip = XFS_BUF_TO_AGI(agibp);
-		ASSERT(XFS_AGI_MAGIC == be32_to_cpu(agip->agi_magicnum));
-		ASSERT(XFS_AGI_GOOD_VERSION(be32_to_cpu(agip->agi_versionnum)));
-		ASSERT(be32_to_cpu(agip->agi_seqno) == agno);
+		error = xfs_read_agi(mp, NULL, agno, &agibp);
+		if (!error) {
+			struct xfs_agi	*agi = XFS_BUF_TO_AGI(agibp);
 
-		itotal += be32_to_cpu(agip->agi_count);
-		ifree += be32_to_cpu(agip->agi_freecount);
-		xfs_buf_relse(agibp);
+			itotal += be32_to_cpu(agi->agi_count);
+			ifree += be32_to_cpu(agi->agi_freecount);
+			xfs_buf_relse(agibp);
+		}
 	}
 
 	sbbp = xfs_getsb(mp, 0);
-- 
cgit v1.2.2


From 4805621a37d9b2b16641b5c68597651419e9e252 Mon Sep 17 00:00:00 2001
From: "From: Christoph Hellwig" <hch@lst.de>
Date: Fri, 28 Nov 2008 14:23:38 +1100
Subject: [XFS] factor out xfs_read_agf helper

Add a helper to read the AGF header and perform basic verification.
Based on hunks from a larger patch from Dave Chinner.

(First sent on Juli 23rd)

Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Dave Chinner <david@fromorbit.com>
Signed-off-by: Niv Sardi <xaiki@sgi.com>
---
 fs/xfs/xfs_log_recover.c | 26 +++++++++++---------------
 1 file changed, 11 insertions(+), 15 deletions(-)

(limited to 'fs/xfs/xfs_log_recover.c')

diff --git a/fs/xfs/xfs_log_recover.c b/fs/xfs/xfs_log_recover.c
index b552676ca5c4..d94987915d5e 100644
--- a/fs/xfs/xfs_log_recover.c
+++ b/fs/xfs/xfs_log_recover.c
@@ -3970,7 +3970,6 @@ xlog_recover_check_summary(
 	xfs_agf_t	*agfp;
 	xfs_buf_t	*agfbp;
 	xfs_buf_t	*agibp;
-	xfs_daddr_t	agfdaddr;
 	xfs_buf_t	*sbbp;
 #ifdef XFS_LOUD_RECOVERY
 	xfs_sb_t	*sbp;
@@ -3987,21 +3986,18 @@ xlog_recover_check_summary(
 	itotal = 0LL;
 	ifree = 0LL;
 	for (agno = 0; agno < mp->m_sb.sb_agcount; agno++) {
-		agfdaddr = XFS_AG_DADDR(mp, agno, XFS_AGF_DADDR(mp));
-		agfbp = xfs_buf_read(mp->m_ddev_targp, agfdaddr,
-				XFS_FSS_TO_BB(mp, 1), 0);
-		if (XFS_BUF_ISERROR(agfbp)) {
-			xfs_ioerror_alert("xlog_recover_check_summary(agf)",
-						mp, agfbp, agfdaddr);
+		error = xfs_read_agf(mp, NULL, agno, 0, &agfbp);
+		if (error) {
+			xfs_fs_cmn_err(CE_ALERT, mp,
+					"xlog_recover_check_summary(agf)"
+					"agf read failed agno %d error %d",
+							agno, error);
+		} else {
+			agfp = XFS_BUF_TO_AGF(agfbp);
+			freeblks += be32_to_cpu(agfp->agf_freeblks) +
+				    be32_to_cpu(agfp->agf_flcount);
+			xfs_buf_relse(agfbp);
 		}
-		agfp = XFS_BUF_TO_AGF(agfbp);
-		ASSERT(XFS_AGF_MAGIC == be32_to_cpu(agfp->agf_magicnum));
-		ASSERT(XFS_AGF_GOOD_VERSION(be32_to_cpu(agfp->agf_versionnum)));
-		ASSERT(be32_to_cpu(agfp->agf_seqno) == agno);
-
-		freeblks += be32_to_cpu(agfp->agf_freeblks) +
-			    be32_to_cpu(agfp->agf_flcount);
-		xfs_buf_relse(agfbp);
 
 		error = xfs_read_agi(mp, NULL, agno, &agibp);
 		if (!error) {
-- 
cgit v1.2.2


From b28708d6a0a3ed65a68f0dcd8e6d1c09f14e5cf3 Mon Sep 17 00:00:00 2001
From: Christoph Hellwig <hch@lst.de>
Date: Fri, 28 Nov 2008 14:23:38 +1100
Subject: [XFS] sanitize xlog_in_core_t definition

Move all fields from xlog_iclog_fields_t into xlog_in_core_t instead of having
them in a substructure and the using #defines to make it look like they were
directly in xlog_in_core_t.  Also document that xlog_in_core_2_t is grossly
misnamed, and make all references to it typesafe.

(First sent on Semptember 15th)

Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Dave Chinner <david@fromorbit.com>
Signed-off-by: Niv Sardi <xaiki@sgi.com>
---
 fs/xfs/xfs_log_recover.c | 7 +++----
 1 file changed, 3 insertions(+), 4 deletions(-)

(limited to 'fs/xfs/xfs_log_recover.c')

diff --git a/fs/xfs/xfs_log_recover.c b/fs/xfs/xfs_log_recover.c
index d94987915d5e..9abb96a7674d 100644
--- a/fs/xfs/xfs_log_recover.c
+++ b/fs/xfs/xfs_log_recover.c
@@ -3332,7 +3332,6 @@ xlog_pack_data(
 	int			size = iclog->ic_offset + roundoff;
 	__be32			cycle_lsn;
 	xfs_caddr_t		dp;
-	xlog_in_core_2_t	*xhdr;
 
 	xlog_pack_data_checksum(log, iclog, size);
 
@@ -3347,7 +3346,8 @@ xlog_pack_data(
 	}
 
 	if (xfs_sb_version_haslogv2(&log->l_mp->m_sb)) {
-		xhdr = (xlog_in_core_2_t *)&iclog->ic_header;
+		xlog_in_core_2_t *xhdr = iclog->ic_data;
+
 		for ( ; i < BTOBB(size); i++) {
 			j = i / (XLOG_HEADER_CYCLE_SIZE / BBSIZE);
 			k = i % (XLOG_HEADER_CYCLE_SIZE / BBSIZE);
@@ -3405,7 +3405,6 @@ xlog_unpack_data(
 	xlog_t			*log)
 {
 	int			i, j, k;
-	xlog_in_core_2_t	*xhdr;
 
 	for (i = 0; i < BTOBB(be32_to_cpu(rhead->h_len)) &&
 		  i < (XLOG_HEADER_CYCLE_SIZE / BBSIZE); i++) {
@@ -3414,7 +3413,7 @@ xlog_unpack_data(
 	}
 
 	if (xfs_sb_version_haslogv2(&log->l_mp->m_sb)) {
-		xhdr = (xlog_in_core_2_t *)rhead;
+		xlog_in_core_2_t *xhdr = (xlog_in_core_2_t *)rhead;
 		for ( ; i < BTOBB(be32_to_cpu(rhead->h_len)); i++) {
 			j = i / (XLOG_HEADER_CYCLE_SIZE / BBSIZE);
 			k = i % (XLOG_HEADER_CYCLE_SIZE / BBSIZE);
-- 
cgit v1.2.2


From 81591fe2db19d0fc1ec2aaaa6a790a5ab97ac3ab Mon Sep 17 00:00:00 2001
From: Christoph Hellwig <hch@lst.de>
Date: Fri, 28 Nov 2008 14:23:39 +1100
Subject: [XFS] kill xfs_dinode_core_t

Now that we have a separate xfs_icdinode_t for the in-core inode which
gets logged there is no need anymore for the xfs_dinode vs xfs_dinode_core
split - the fact that part of the structure gets logged through the inode
log item and a small part not can better be described in a comment.

All sizeof operations on the dinode_core either really wanted the
icdinode and are switched to that one, or had already added the size
of the agi unlinked list pointer.  Later both will be replaced with
helpers once we get the larger CRC-enabled dinode.

Removing the data and attribute fork unions also has the advantage that
xfs_dinode.h doesn't need to pull in every header under the sun.

While we're at it also add some more comments describing the dinode
structure.

(First sent on October 7th)

Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Dave Chinner <david@fromorbit.com>
Signed-off-by: Niv Sardi <xaiki@sgi.com>
---
 fs/xfs/xfs_log_recover.c | 29 +++++++++++++++--------------
 1 file changed, 15 insertions(+), 14 deletions(-)

(limited to 'fs/xfs/xfs_log_recover.c')

diff --git a/fs/xfs/xfs_log_recover.c b/fs/xfs/xfs_log_recover.c
index 9abb96a7674d..4099618f5fa0 100644
--- a/fs/xfs/xfs_log_recover.c
+++ b/fs/xfs/xfs_log_recover.c
@@ -2320,7 +2320,7 @@ xlog_recover_do_inode_trans(
 	 * Make sure the place we're flushing out to really looks
 	 * like an inode!
 	 */
-	if (unlikely(be16_to_cpu(dip->di_core.di_magic) != XFS_DINODE_MAGIC)) {
+	if (unlikely(be16_to_cpu(dip->di_magic) != XFS_DINODE_MAGIC)) {
 		xfs_buf_relse(bp);
 		xfs_fs_cmn_err(CE_ALERT, mp,
 			"xfs_inode_recover: Bad inode magic number, dino ptr = 0x%p, dino bp = 0x%p, ino = %Ld",
@@ -2343,12 +2343,12 @@ xlog_recover_do_inode_trans(
 	}
 
 	/* Skip replay when the on disk inode is newer than the log one */
-	if (dicp->di_flushiter < be16_to_cpu(dip->di_core.di_flushiter)) {
+	if (dicp->di_flushiter < be16_to_cpu(dip->di_flushiter)) {
 		/*
 		 * Deal with the wrap case, DI_MAX_FLUSH is less
 		 * than smaller numbers
 		 */
-		if (be16_to_cpu(dip->di_core.di_flushiter) == DI_MAX_FLUSH &&
+		if (be16_to_cpu(dip->di_flushiter) == DI_MAX_FLUSH &&
 		    dicp->di_flushiter < (DI_MAX_FLUSH >> 1)) {
 			/* do nothing */
 		} else {
@@ -2408,7 +2408,7 @@ xlog_recover_do_inode_trans(
 		error = EFSCORRUPTED;
 		goto error;
 	}
-	if (unlikely(item->ri_buf[1].i_len > sizeof(xfs_dinode_core_t))) {
+	if (unlikely(item->ri_buf[1].i_len > sizeof(struct xfs_icdinode))) {
 		XFS_CORRUPTION_ERROR("xlog_recover_do_inode_trans(7)",
 				     XFS_ERRLEVEL_LOW, mp, dicp);
 		xfs_buf_relse(bp);
@@ -2420,23 +2420,24 @@ xlog_recover_do_inode_trans(
 	}
 
 	/* The core is in in-core format */
-	xfs_dinode_to_disk(&dip->di_core,
-		(xfs_icdinode_t *)item->ri_buf[1].i_addr);
+	xfs_dinode_to_disk(dip, (xfs_icdinode_t *)item->ri_buf[1].i_addr);
 
 	/* the rest is in on-disk format */
-	if (item->ri_buf[1].i_len > sizeof(xfs_dinode_core_t)) {
-		memcpy((xfs_caddr_t) dip + sizeof(xfs_dinode_core_t),
-			item->ri_buf[1].i_addr + sizeof(xfs_dinode_core_t),
-			item->ri_buf[1].i_len  - sizeof(xfs_dinode_core_t));
+	if (item->ri_buf[1].i_len > sizeof(struct xfs_icdinode)) {
+		memcpy((xfs_caddr_t) dip + sizeof(struct xfs_icdinode),
+			item->ri_buf[1].i_addr + sizeof(struct xfs_icdinode),
+			item->ri_buf[1].i_len  - sizeof(struct xfs_icdinode));
 	}
 
 	fields = in_f->ilf_fields;
 	switch (fields & (XFS_ILOG_DEV | XFS_ILOG_UUID)) {
 	case XFS_ILOG_DEV:
-		dip->di_u.di_dev = cpu_to_be32(in_f->ilf_u.ilfu_rdev);
+		xfs_dinode_put_rdev(dip, in_f->ilf_u.ilfu_rdev);
 		break;
 	case XFS_ILOG_UUID:
-		dip->di_u.di_muuid = in_f->ilf_u.ilfu_uuid;
+		memcpy(XFS_DFORK_DPTR(dip),
+		       &in_f->ilf_u.ilfu_uuid,
+		       sizeof(uuid_t));
 		break;
 	}
 
@@ -2452,12 +2453,12 @@ xlog_recover_do_inode_trans(
 	switch (fields & XFS_ILOG_DFORK) {
 	case XFS_ILOG_DDATA:
 	case XFS_ILOG_DEXT:
-		memcpy(&dip->di_u, src, len);
+		memcpy(XFS_DFORK_DPTR(dip), src, len);
 		break;
 
 	case XFS_ILOG_DBROOT:
 		xfs_bmbt_to_bmdr(mp, (struct xfs_btree_block *)src, len,
-				 &dip->di_u.di_bmbt,
+				 (xfs_bmdr_block_t *)XFS_DFORK_DPTR(dip),
 				 XFS_DFORK_DSIZE(dip, mp));
 		break;
 
-- 
cgit v1.2.2


From 23fac50f959a87febf7ce4ae9d47525121f10c7a Mon Sep 17 00:00:00 2001
From: Christoph Hellwig <hch@lst.de>
Date: Fri, 28 Nov 2008 14:23:40 +1100
Subject: [XFS] split up xlog_recover_process_iunlinks

Split out the body of the main loop into a separate helper to make the
code readable.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Dave Chinner <david@fromorbit.com>
Signed-off-by: Niv Sardi <xaiki@sgi.com>
---
 fs/xfs/xfs_log_recover.c | 134 +++++++++++++++++++++++------------------------
 1 file changed, 66 insertions(+), 68 deletions(-)

(limited to 'fs/xfs/xfs_log_recover.c')

diff --git a/fs/xfs/xfs_log_recover.c b/fs/xfs/xfs_log_recover.c
index 4099618f5fa0..841398d24211 100644
--- a/fs/xfs/xfs_log_recover.c
+++ b/fs/xfs/xfs_log_recover.c
@@ -3147,6 +3147,70 @@ out_error:
 	return;
 }
 
+STATIC xfs_agino_t
+xlog_recover_process_one_iunlink(
+	struct xfs_mount		*mp,
+	xfs_agnumber_t			agno,
+	xfs_agino_t			agino,
+	int				bucket)
+{
+	struct xfs_buf			*ibp;
+	struct xfs_dinode		*dip;
+	struct xfs_inode		*ip;
+	xfs_ino_t			ino;
+	int				error;
+
+	ino = XFS_AGINO_TO_INO(mp, agno, agino);
+	error = xfs_iget(mp, NULL, ino, 0, 0, &ip, 0);
+	if (error)
+		goto fail;
+
+	/*
+	 * Get the on disk inode to find the next inode in the bucket.
+	 */
+	ASSERT(ip != NULL);
+	error = xfs_itobp(mp, NULL, ip, &dip, &ibp, 0, 0, XFS_BUF_LOCK);
+	if (error)
+		goto fail;
+
+	ASSERT(dip != NULL);
+	ASSERT(ip->i_d.di_nlink == 0);
+
+	/* setup for the next pass */
+	agino = be32_to_cpu(dip->di_next_unlinked);
+	xfs_buf_relse(ibp);
+
+	/*
+	 * Prevent any DMAPI event from being sent when the reference on
+	 * the inode is dropped.
+	 */
+	ip->i_d.di_dmevmask = 0;
+
+	/*
+	 * If this is a new inode, handle it specially.  Otherwise, just
+	 * drop our reference to the inode.  If there are no other
+	 * references, this will send the inode to xfs_inactive() which
+	 * will truncate the file and free the inode.
+	 */
+	if (ip->i_d.di_mode == 0)
+		xfs_iput_new(ip, 0);
+	else
+		IRELE(ip);
+	return agino;
+
+ fail:
+	/*
+	 * We can't read in the inode this bucket points to, or this inode
+	 * is messed up.  Just ditch this bucket of inodes.  We will lose
+	 * some inodes and space, but at least we won't hang.
+	 *
+	 * Call xlog_recover_clear_agi_bucket() to perform a transaction to
+	 * clear the inode pointer in the bucket.
+	 */
+	xlog_recover_clear_agi_bucket(mp, agno, bucket);
+	return NULLAGINO;
+}
+
 /*
  * xlog_iunlink_recover
  *
@@ -3167,11 +3231,7 @@ xlog_recover_process_iunlinks(
 	xfs_agnumber_t	agno;
 	xfs_agi_t	*agi;
 	xfs_buf_t	*agibp;
-	xfs_buf_t	*ibp;
-	xfs_dinode_t	*dip;
-	xfs_inode_t	*ip;
 	xfs_agino_t	agino;
-	xfs_ino_t	ino;
 	int		bucket;
 	int		error;
 	uint		mp_dmevmask;
@@ -3201,10 +3261,8 @@ xlog_recover_process_iunlinks(
 		agi = XFS_BUF_TO_AGI(agibp);
 
 		for (bucket = 0; bucket < XFS_AGI_UNLINKED_BUCKETS; bucket++) {
-
 			agino = be32_to_cpu(agi->agi_unlinked[bucket]);
 			while (agino != NULLAGINO) {
-
 				/*
 				 * Release the agi buffer so that it can
 				 * be acquired in the normal course of the
@@ -3212,68 +3270,8 @@ xlog_recover_process_iunlinks(
 				 */
 				xfs_buf_relse(agibp);
 
-				ino = XFS_AGINO_TO_INO(mp, agno, agino);
-				error = xfs_iget(mp, NULL, ino, 0, 0, &ip, 0);
-				ASSERT(error || (ip != NULL));
-
-				if (!error) {
-					/*
-					 * Get the on disk inode to find the
-					 * next inode in the bucket.
-					 */
-					error = xfs_itobp(mp, NULL, ip, &dip,
-							&ibp, 0, 0,
-							XFS_BUF_LOCK);
-					ASSERT(error || (dip != NULL));
-				}
-
-				if (!error) {
-					ASSERT(ip->i_d.di_nlink == 0);
-
-					/* setup for the next pass */
-					agino = be32_to_cpu(
-							dip->di_next_unlinked);
-					xfs_buf_relse(ibp);
-					/*
-					 * Prevent any DMAPI event from
-					 * being sent when the
-					 * reference on the inode is
-					 * dropped.
-					 */
-					ip->i_d.di_dmevmask = 0;
-
-					/*
-					 * If this is a new inode, handle
-					 * it specially.  Otherwise,
-					 * just drop our reference to the
-					 * inode.  If there are no
-					 * other references, this will
-					 * send the inode to
-					 * xfs_inactive() which will
-					 * truncate the file and free
-					 * the inode.
-					 */
-					if (ip->i_d.di_mode == 0)
-						xfs_iput_new(ip, 0);
-					else
-						IRELE(ip);
-				} else {
-					/*
-					 * We can't read in the inode
-					 * this bucket points to, or
-					 * this inode is messed up.  Just
-					 * ditch this bucket of inodes.  We
-					 * will lose some inodes and space,
-					 * but at least we won't hang.  Call
-					 * xlog_recover_clear_agi_bucket()
-					 * to perform a transaction to clear
-					 * the inode pointer in the bucket.
-					 */
-					xlog_recover_clear_agi_bucket(mp, agno,
-							bucket);
-
-					agino = NULLAGINO;
-				}
+				agino = xlog_recover_process_one_iunlink(mp,
+							agno, agino, bucket);
 
 				/*
 				 * Reacquire the agibuffer and continue around
-- 
cgit v1.2.2


From 76d8b277f7b715f78ee3cb09ee112563639693a5 Mon Sep 17 00:00:00 2001
From: Christoph Hellwig <hch@lst.de>
Date: Fri, 28 Nov 2008 14:23:40 +1100
Subject: [XFS] stop using xfs_itobp in xfs_iread

The only caller of xfs_itobp that doesn't have i_blkno setup is now
the initial inode read.  It needs access to the whole xfs_imap so using
xfs_inotobp is not an option.  Instead opencode the buffer lookup in
xfs_iread and kill all the functionality for the initial map from
xfs_itobp.

(First sent on October 21st)

Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Dave Chinner <david@fromorbit.com>
Signed-off-by: Niv Sardi <xaiki@sgi.com>
---
 fs/xfs/xfs_log_recover.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

(limited to 'fs/xfs/xfs_log_recover.c')

diff --git a/fs/xfs/xfs_log_recover.c b/fs/xfs/xfs_log_recover.c
index 841398d24211..48bdfa4dc290 100644
--- a/fs/xfs/xfs_log_recover.c
+++ b/fs/xfs/xfs_log_recover.c
@@ -3169,7 +3169,7 @@ xlog_recover_process_one_iunlink(
 	 * Get the on disk inode to find the next inode in the bucket.
 	 */
 	ASSERT(ip != NULL);
-	error = xfs_itobp(mp, NULL, ip, &dip, &ibp, 0, 0, XFS_BUF_LOCK);
+	error = xfs_itobp(mp, NULL, ip, &dip, &ibp, XFS_BUF_LOCK);
 	if (error)
 		goto fail;
 
-- 
cgit v1.2.2


From a1941895034cda2bffa23ba845607c82138ccf52 Mon Sep 17 00:00:00 2001
From: Christoph Hellwig <hch@lst.de>
Date: Fri, 28 Nov 2008 14:23:40 +1100
Subject: [XFS] remove dead code for old inode item recovery

We have removed the support for old-style inode items a while ago and
xlog_recover_do_inode_trans is now only called for XFS_LI_INODE items.
That means we can remove the call to xfs_imap there and with it the
XFS_IMAP_LOOKUP that is set by all other callers.  We can also mark
xfs_imap static now.

(First sent on October 21st)

Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Dave Chinner <david@fromorbit.com>
Signed-off-by: Niv Sardi <xaiki@sgi.com>
---
 fs/xfs/xfs_log_recover.c | 32 ++++++--------------------------
 1 file changed, 6 insertions(+), 26 deletions(-)

(limited to 'fs/xfs/xfs_log_recover.c')

diff --git a/fs/xfs/xfs_log_recover.c b/fs/xfs/xfs_log_recover.c
index 48bdfa4dc290..bf8573b5a7d8 100644
--- a/fs/xfs/xfs_log_recover.c
+++ b/fs/xfs/xfs_log_recover.c
@@ -2245,7 +2245,6 @@ xlog_recover_do_inode_trans(
 	xfs_inode_log_format_t	*in_f;
 	xfs_mount_t		*mp;
 	xfs_buf_t		*bp;
-	xfs_imap_t		imap;
 	xfs_dinode_t		*dip;
 	xfs_ino_t		ino;
 	int			len;
@@ -2273,48 +2272,29 @@ xlog_recover_do_inode_trans(
 	}
 	ino = in_f->ilf_ino;
 	mp = log->l_mp;
-	if (ITEM_TYPE(item) == XFS_LI_INODE) {
-		imap.im_blkno = (xfs_daddr_t)in_f->ilf_blkno;
-		imap.im_len = in_f->ilf_len;
-		imap.im_boffset = in_f->ilf_boffset;
-	} else {
-		/*
-		 * It's an old inode format record.  We don't know where
-		 * its cluster is located on disk, and we can't allow
-		 * xfs_imap() to figure it out because the inode btrees
-		 * are not ready to be used.  Therefore do not pass the
-		 * XFS_IMAP_LOOKUP flag to xfs_imap().  This will give
-		 * us only the single block in which the inode lives
-		 * rather than its cluster, so we must make sure to
-		 * invalidate the buffer when we write it out below.
-		 */
-		imap.im_blkno = 0;
-		error = xfs_imap(log->l_mp, NULL, ino, &imap, 0);
-		if (error)
-			goto error;
-	}
 
 	/*
 	 * Inode buffers can be freed, look out for it,
 	 * and do not replay the inode.
 	 */
-	if (xlog_check_buffer_cancelled(log, imap.im_blkno, imap.im_len, 0)) {
+	if (xlog_check_buffer_cancelled(log, in_f->ilf_blkno,
+					in_f->ilf_len, 0)) {
 		error = 0;
 		goto error;
 	}
 
-	bp = xfs_buf_read_flags(mp->m_ddev_targp, imap.im_blkno, imap.im_len,
-								XFS_BUF_LOCK);
+	bp = xfs_buf_read_flags(mp->m_ddev_targp, in_f->ilf_blkno,
+				in_f->ilf_len, XFS_BUF_LOCK);
 	if (XFS_BUF_ISERROR(bp)) {
 		xfs_ioerror_alert("xlog_recover_do..(read#2)", mp,
-				  bp, imap.im_blkno);
+				  bp, in_f->ilf_blkno);
 		error = XFS_BUF_GETERROR(bp);
 		xfs_buf_relse(bp);
 		goto error;
 	}
 	error = 0;
 	ASSERT(in_f->ilf_fields & XFS_ILOG_CORE);
-	dip = (xfs_dinode_t *)xfs_buf_offset(bp, imap.im_boffset);
+	dip = (xfs_dinode_t *)xfs_buf_offset(bp, in_f->ilf_boffset);
 
 	/*
 	 * Make sure the place we're flushing out to really looks
-- 
cgit v1.2.2


From 92bfc6e7c4eabbbd15e7d6d49123b296d05dcfd1 Mon Sep 17 00:00:00 2001
From: Christoph Hellwig <hch@lst.de>
Date: Fri, 28 Nov 2008 14:23:41 +1100
Subject: [XFS] embededd struct xfs_imap into xfs_inode

Most uses of struct xfs_imap are to map and inode to a buffer.  To avoid
copying around the inode location information we should just embedd a
strcut xfs_imap into the xfs_inode.  To make sure it doesn't bloat an
inode the im_len is changed to a ushort, which is fine as that's what
the users exepect anyway.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Dave Chinner <david@fromorbit.com>
Signed-off-by: Niv Sardi <xaiki@sgi.com>
---
 fs/xfs/xfs_log_recover.c | 1 -
 1 file changed, 1 deletion(-)

(limited to 'fs/xfs/xfs_log_recover.c')

diff --git a/fs/xfs/xfs_log_recover.c b/fs/xfs/xfs_log_recover.c
index bf8573b5a7d8..ce6e907bec62 100644
--- a/fs/xfs/xfs_log_recover.c
+++ b/fs/xfs/xfs_log_recover.c
@@ -36,7 +36,6 @@
 #include "xfs_dinode.h"
 #include "xfs_inode.h"
 #include "xfs_inode_item.h"
-#include "xfs_imap.h"
 #include "xfs_alloc.h"
 #include "xfs_ialloc.h"
 #include "xfs_log_priv.h"
-- 
cgit v1.2.2


From 0e446673a15a4e9c336b67c1a638eb12c21d0993 Mon Sep 17 00:00:00 2001
From: Christoph Hellwig <hch@lst.de>
Date: Fri, 28 Nov 2008 14:23:42 +1100
Subject: [XFS] fix error handling in xlog_recover_process_one_iunlink

If we fail after xfs_iget we have to drop the reference count, spotted
by Dave Chinner.  Also remove some useless asserts and stop trying to
deal with di_mode == 0 inodes because never gets those without passing
the IGET_CREATE flag to xfs_iget.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Eric Sandeen <sandeen@sandeen.net>
Signed-off-by: Niv Sardi <xaiki@sgi.com>
---
 fs/xfs/xfs_log_recover.c | 18 +++++-------------
 1 file changed, 5 insertions(+), 13 deletions(-)

(limited to 'fs/xfs/xfs_log_recover.c')

diff --git a/fs/xfs/xfs_log_recover.c b/fs/xfs/xfs_log_recover.c
index ce6e907bec62..51412cced010 100644
--- a/fs/xfs/xfs_log_recover.c
+++ b/fs/xfs/xfs_log_recover.c
@@ -3147,13 +3147,12 @@ xlog_recover_process_one_iunlink(
 	/*
 	 * Get the on disk inode to find the next inode in the bucket.
 	 */
-	ASSERT(ip != NULL);
 	error = xfs_itobp(mp, NULL, ip, &dip, &ibp, XFS_BUF_LOCK);
 	if (error)
-		goto fail;
+		goto fail_iput;
 
-	ASSERT(dip != NULL);
 	ASSERT(ip->i_d.di_nlink == 0);
+	ASSERT(ip->i_d.di_mode != 0);
 
 	/* setup for the next pass */
 	agino = be32_to_cpu(dip->di_next_unlinked);
@@ -3165,18 +3164,11 @@ xlog_recover_process_one_iunlink(
 	 */
 	ip->i_d.di_dmevmask = 0;
 
-	/*
-	 * If this is a new inode, handle it specially.  Otherwise, just
-	 * drop our reference to the inode.  If there are no other
-	 * references, this will send the inode to xfs_inactive() which
-	 * will truncate the file and free the inode.
-	 */
-	if (ip->i_d.di_mode == 0)
-		xfs_iput_new(ip, 0);
-	else
-		IRELE(ip);
+	IRELE(ip);
 	return agino;
 
+ fail_iput:
+	IRELE(ip);
  fail:
 	/*
 	 * We can't read in the inode this bucket points to, or this inode
-- 
cgit v1.2.2


From 15ac08a8b2c129abccf1be47b6ab09491e013db2 Mon Sep 17 00:00:00 2001
From: Christoph Hellwig <hch@infradead.org>
Date: Tue, 9 Dec 2008 04:47:30 -0500
Subject: [XFS] replace b_fspriv with b_mount

Replace the b_fspriv pointer and it's ugly accessors with a properly types
xfs_mount pointer.  Also switch log reocvery over to it instead of using
b_fspriv for the mount pointer.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Dave Chinner <david@fromorbit.com>
Signed-off-by: Lachlan McIlroy <lachlan@sgi.com>
---
 fs/xfs/xfs_log_recover.c | 26 +++++++++-----------------
 1 file changed, 9 insertions(+), 17 deletions(-)

(limited to 'fs/xfs/xfs_log_recover.c')

diff --git a/fs/xfs/xfs_log_recover.c b/fs/xfs/xfs_log_recover.c
index 51412cced010..35cca98bd94c 100644
--- a/fs/xfs/xfs_log_recover.c
+++ b/fs/xfs/xfs_log_recover.c
@@ -267,21 +267,16 @@ STATIC void
 xlog_recover_iodone(
 	struct xfs_buf	*bp)
 {
-	xfs_mount_t	*mp;
-
-	ASSERT(XFS_BUF_FSPRIVATE(bp, void *));
-
 	if (XFS_BUF_GETERROR(bp)) {
 		/*
 		 * We're not going to bother about retrying
 		 * this during recovery. One strike!
 		 */
-		mp = XFS_BUF_FSPRIVATE(bp, xfs_mount_t *);
 		xfs_ioerror_alert("xlog_recover_iodone",
-				  mp, bp, XFS_BUF_ADDR(bp));
-		xfs_force_shutdown(mp, SHUTDOWN_META_IO_ERROR);
+				  bp->b_mount, bp, XFS_BUF_ADDR(bp));
+		xfs_force_shutdown(bp->b_mount, SHUTDOWN_META_IO_ERROR);
 	}
-	XFS_BUF_SET_FSPRIVATE(bp, NULL);
+	bp->b_mount = NULL;
 	XFS_BUF_CLR_IODONE_FUNC(bp);
 	xfs_biodone(bp);
 }
@@ -2225,9 +2220,8 @@ xlog_recover_do_buffer_trans(
 		XFS_BUF_STALE(bp);
 		error = xfs_bwrite(mp, bp);
 	} else {
-		ASSERT(XFS_BUF_FSPRIVATE(bp, void *) == NULL ||
-		       XFS_BUF_FSPRIVATE(bp, xfs_mount_t *) == mp);
-		XFS_BUF_SET_FSPRIVATE(bp, mp);
+		ASSERT(bp->b_mount == NULL || bp->b_mount == mp);
+		bp->b_mount = mp;
 		XFS_BUF_SET_IODONE_FUNC(bp, xlog_recover_iodone);
 		xfs_bdwrite(mp, bp);
 	}
@@ -2490,9 +2484,8 @@ xlog_recover_do_inode_trans(
 
 write_inode_buffer:
 	if (ITEM_TYPE(item) == XFS_LI_INODE) {
-		ASSERT(XFS_BUF_FSPRIVATE(bp, void *) == NULL ||
-		       XFS_BUF_FSPRIVATE(bp, xfs_mount_t *) == mp);
-		XFS_BUF_SET_FSPRIVATE(bp, mp);
+		ASSERT(bp->b_mount == NULL || bp->b_mount == mp);
+		bp->b_mount = mp;
 		XFS_BUF_SET_IODONE_FUNC(bp, xlog_recover_iodone);
 		xfs_bdwrite(mp, bp);
 	} else {
@@ -2623,9 +2616,8 @@ xlog_recover_do_dquot_trans(
 	memcpy(ddq, recddq, item->ri_buf[1].i_len);
 
 	ASSERT(dq_f->qlf_size == 2);
-	ASSERT(XFS_BUF_FSPRIVATE(bp, void *) == NULL ||
-	       XFS_BUF_FSPRIVATE(bp, xfs_mount_t *) == mp);
-	XFS_BUF_SET_FSPRIVATE(bp, mp);
+	ASSERT(bp->b_mount == NULL || bp->b_mount == mp);
+	bp->b_mount = mp;
 	XFS_BUF_SET_IODONE_FUNC(bp, xlog_recover_iodone);
 	xfs_bdwrite(mp, bp);
 
-- 
cgit v1.2.2