xfs: rmap btree requires more reserved free space

Originally-From: Dave Chinner <dchinner@redhat.com>

The rmap btree is allocated from the AGFL, which means we have to
ensure ENOSPC is reported to userspace before we run out of free
space in each AG. The last allocation in an AG can cause a full
height rmap btree split, and that means we have to reserve at least
this many blocks *in each AG* to be placed on the AGFL at ENOSPC.
Update the various space calculation functions to handle this.

Also, because the macros are now executing conditional code and are
called quite frequently, convert them to functions that initialise
variables in the struct xfs_mount, use the new variables everywhere
and document the calculations better.

[darrick.wong@oracle.com: don't reserve blocks if !rmap]
[dchinner@redhat.com: update m_ag_max_usable after growfs]

Signed-off-by: Dave Chinner <dchinner@redhat.com>
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: Dave Chinner <dchinner@redhat.com>
Signed-off-by: Dave Chinner <david@fromorbit.com>

1 files changed, 69 insertions, 0 deletions

diff --git a/fs/xfs/libxfs/xfs_alloc.c b/fs/xfs/libxfs/xfs_alloc.c
index dba7ce4b4be8..73ab1ea1270d 100644
--- a/fs/xfs/libxfs/xfs_alloc.c
+++ b/fs/xfs/libxfs/xfs_alloc.c
@@ -63,6 +63,70 @@ xfs_prealloc_blocks(
 }
 
 /*
+ * In order to avoid ENOSPC-related deadlock caused by out-of-order locking of
+ * AGF buffer (PV 947395), we place constraints on the relationship among
+ * actual allocations for data blocks, freelist blocks, and potential file data
+ * bmap btree blocks. However, these restrictions may result in no actual space
+ * allocated for a delayed extent, for example, a data block in a certain AG is
+ * allocated but there is no additional block for the additional bmap btree
+ * block due to a split of the bmap btree of the file. The result of this may
+ * lead to an infinite loop when the file gets flushed to disk and all delayed
+ * extents need to be actually allocated. To get around this, we explicitly set
+ * aside a few blocks which will not be reserved in delayed allocation.
+ *
+ * When rmap is disabled, we need to reserve 4 fsbs _per AG_ for the freelist
+ * and 4 more to handle a potential split of the file's bmap btree.
+ *
+ * When rmap is enabled, we must also be able to handle two rmap btree inserts
+ * to record both the file data extent and a new bmbt block.  The bmbt block
+ * might not be in the same AG as the file data extent.  In the worst case
+ * the bmap btree splits multiple levels and all the new blocks come from
+ * different AGs, so set aside enough to handle rmap btree splits in all AGs.
+ */
+unsigned int
+xfs_alloc_set_aside(
+	struct xfs_mount	*mp)
+{
+	unsigned int		blocks;
+
+	blocks = 4 + (mp->m_sb.sb_agcount * XFS_ALLOC_AGFL_RESERVE);
+	if (xfs_sb_version_hasrmapbt(&mp->m_sb))
+		blocks += mp->m_sb.sb_agcount * mp->m_rmap_maxlevels;
+	return blocks;
+}
+
+/*
+ * When deciding how much space to allocate out of an AG, we limit the
+ * allocation maximum size to the size the AG. However, we cannot use all the
+ * blocks in the AG - some are permanently used by metadata. These
+ * blocks are generally:
+ *	- the AG superblock, AGF, AGI and AGFL
+ *	- the AGF (bno and cnt) and AGI btree root blocks, and optionally
+ *	  the AGI free inode and rmap btree root blocks.
+ *	- blocks on the AGFL according to xfs_alloc_set_aside() limits
+ *	- the rmapbt root block
+ *
+ * The AG headers are sector sized, so the amount of space they take up is
+ * dependent on filesystem geometry. The others are all single blocks.
+ */
+unsigned int
+xfs_alloc_ag_max_usable(
+	struct xfs_mount	*mp)
+{
+	unsigned int		blocks;
+
+	blocks = XFS_BB_TO_FSB(mp, XFS_FSS_TO_BB(mp, 4)); /* ag headers */
+	blocks += XFS_ALLOC_AGFL_RESERVE;
+	blocks += 3;			/* AGF, AGI btree root blocks */
+	if (xfs_sb_version_hasfinobt(&mp->m_sb))
+		blocks++;		/* finobt root block */
+	if (xfs_sb_version_hasrmapbt(&mp->m_sb))
+		blocks++; 		/* rmap root block */
+
+	return mp->m_sb.sb_agblocks - blocks;
+}
+
+/*
  * Lookup the record equal to [bno, len] in the btree given by cur.
  */
 STATIC int				/* error */
@@ -1904,6 +1968,11 @@ xfs_alloc_min_freelist(
 	/* space needed by-size freespace btree */
 	min_free += min_t(unsigned int, pag->pagf_levels[XFS_BTNUM_CNTi] + 1,
 				       mp->m_ag_maxlevels);
+	/* space needed reverse mapping used space btree */
+	if (xfs_sb_version_hasrmapbt(&mp->m_sb))
+		min_free += min_t(unsigned int,
+				  pag->pagf_levels[XFS_BTNUM_RMAPi] + 1,
+				  mp->m_rmap_maxlevels);
 
 	return min_free;
 }