Merge tag 'xfs-5.6-merge-8' of git://git.kernel.org/pub/scm/fs/xfs/xfs-linux

Pull moar xfs updates from Darrick Wong:
 "This contains the buffer error code refactoring I mentioned last week,
  now that it has had extra time to complete the full xfs fuzz testing
  suite to make sure there aren't any obvious new bugs"

* tag 'xfs-5.6-merge-8' of git://git.kernel.org/pub/scm/fs/xfs/xfs-linux:
  xfs: fix xfs_buf_ioerror_alert location reporting
  xfs: remove unnecessary null pointer checks from _read_agf callers
  xfs: make xfs_*read_agf return EAGAIN to ALLOC_FLAG_TRYLOCK callers
  xfs: remove the xfs_btree_get_buf[ls] functions
  xfs: make xfs_trans_get_buf return an error code
  xfs: make xfs_trans_get_buf_map return an error code
  xfs: make xfs_buf_read return an error code
  xfs: make xfs_buf_get_uncached return an error code
  xfs: make xfs_buf_get return an error code
  xfs: make xfs_buf_read_map return an error code
  xfs: make xfs_buf_get_map return an error code
  xfs: make xfs_buf_alloc return an error code

fs/xfs/libxfs/xfs_ag.c

@@ -23,25 +23,28 @@
 #include "xfs_ag_resv.h"
 #include "xfs_health.h"
 
-static struct xfs_buf *
+static int
 xfs_get_aghdr_buf(
 	struct xfs_mount	*mp,
 	xfs_daddr_t		blkno,
 	size_t			numblks,
+	struct xfs_buf		**bpp,
 	const struct xfs_buf_ops *ops)
 {
 	struct xfs_buf		*bp;
+	int			error;
 
-	bp = xfs_buf_get_uncached(mp->m_ddev_targp, numblks, 0);
-	if (!bp)
-		return NULL;
+	error = xfs_buf_get_uncached(mp->m_ddev_targp, numblks, 0, &bp);
+	if (error)
+		return error;
 
 	xfs_buf_zero(bp, 0, BBTOB(bp->b_length));
 	bp->b_bn = blkno;
 	bp->b_maps[0].bm_bn = blkno;
 	bp->b_ops = ops;
 
-	return bp;
+	*bpp = bp;
+	return 0;
 }
 
 static inline bool is_log_ag(struct xfs_mount *mp, struct aghdr_init_data *id)
@@ -340,13 +343,13 @@ xfs_ag_init_hdr(
 	struct aghdr_init_data	*id,
 	aghdr_init_work_f	work,
 	const struct xfs_buf_ops *ops)
-
 {
 	struct xfs_buf		*bp;
+	int			error;
 
-	bp = xfs_get_aghdr_buf(mp, id->daddr, id->numblks, ops);
-	if (!bp)
-		return -ENOMEM;
+	error = xfs_get_aghdr_buf(mp, id->daddr, id->numblks, &bp, ops);
+	if (error)
+		return error;
 
 	(*work)(mp, bp, id);
 

fs/xfs/libxfs/xfs_alloc.c

@@ -1070,11 +1070,11 @@ xfs_alloc_ag_vextent_small(
 	if (args->datatype & XFS_ALLOC_USERDATA) {
 		struct xfs_buf	*bp;
 
-		bp = xfs_btree_get_bufs(args->mp, args->tp, args->agno, fbno);
-		if (XFS_IS_CORRUPT(args->mp, !bp)) {
-			error = -EFSCORRUPTED;
+		error = xfs_trans_get_buf(args->tp, args->mp->m_ddev_targp,
+				XFS_AGB_TO_DADDR(args->mp, args->agno, fbno),
+				args->mp->m_bsize, 0, &bp);
+		if (error)
 			goto error;
-		}
 		xfs_trans_binval(args->tp, bp);
 	}
 	*fbnop = args->agbno = fbno;
@@ -2347,9 +2347,11 @@ xfs_free_agfl_block(
 	if (error)
 		return error;
 
-	bp = xfs_btree_get_bufs(tp->t_mountp, tp, agno, agbno);
-	if (XFS_IS_CORRUPT(tp->t_mountp, !bp))
-		return -EFSCORRUPTED;
+	error = xfs_trans_get_buf(tp, tp->t_mountp->m_ddev_targp,
+			XFS_AGB_TO_DADDR(tp->t_mountp, agno, agbno),
+			tp->t_mountp->m_bsize, 0, &bp);
+	if (error)
+		return error;
 	xfs_trans_binval(tp, bp);
 
 	return 0;
@@ -2500,12 +2502,11 @@ xfs_alloc_fix_freelist(
 
 	if (!pag->pagf_init) {
 		error = xfs_alloc_read_agf(mp, tp, args->agno, flags, &agbp);
-		if (error)
+		if (error) {
+			/* Couldn't lock the AGF so skip this AG. */
+			if (error == -EAGAIN)
+				error = 0;
 			goto out_no_agbp;
-		if (!pag->pagf_init) {
-			ASSERT(flags & XFS_ALLOC_FLAG_TRYLOCK);
-			ASSERT(!(flags & XFS_ALLOC_FLAG_FREEING));
-			goto out_agbp_relse;
 		}
 	}
 
@@ -2531,11 +2532,10 @@ xfs_alloc_fix_freelist(
 	 */
 	if (!agbp) {
 		error = xfs_alloc_read_agf(mp, tp, args->agno, flags, &agbp);
-		if (error)
-			goto out_no_agbp;
-		if (!agbp) {
-			ASSERT(flags & XFS_ALLOC_FLAG_TRYLOCK);
-			ASSERT(!(flags & XFS_ALLOC_FLAG_FREEING));
+		if (error) {
+			/* Couldn't lock the AGF so skip this AG. */
+			if (error == -EAGAIN)
+				error = 0;
 			goto out_no_agbp;
 		}
 	}
@@ -2766,11 +2766,10 @@ xfs_alloc_pagf_init(
 	xfs_buf_t		*bp;
 	int			error;
 
-	if ((error = xfs_alloc_read_agf(mp, tp, agno, flags, &bp)))
-		return error;
-	if (bp)
+	error = xfs_alloc_read_agf(mp, tp, agno, flags, &bp);
+	if (!error)
 		xfs_trans_brelse(tp, bp);
-	return 0;
+	return error;
 }
 
 /*
@@ -2956,14 +2955,11 @@ xfs_read_agf(
 	trace_xfs_read_agf(mp, agno);
 
 	ASSERT(agno != NULLAGNUMBER);
-	error = xfs_trans_read_buf(
-			mp, tp, mp->m_ddev_targp,
+	error = xfs_trans_read_buf(mp, tp, mp->m_ddev_targp,
 			XFS_AG_DADDR(mp, agno, XFS_AGF_DADDR(mp)),
 			XFS_FSS_TO_BB(mp, 1), flags, bpp, &xfs_agf_buf_ops);
 	if (error)
 		return error;
-	if (!*bpp)
-		return 0;
 
 	ASSERT(!(*bpp)->b_error);
 	xfs_buf_set_ref(*bpp, XFS_AGF_REF);
@@ -2987,14 +2983,15 @@ xfs_alloc_read_agf(
 
 	trace_xfs_alloc_read_agf(mp, agno);
 
+	/* We don't support trylock when freeing. */
+	ASSERT((flags & (XFS_ALLOC_FLAG_FREEING | XFS_ALLOC_FLAG_TRYLOCK)) !=
+			(XFS_ALLOC_FLAG_FREEING | XFS_ALLOC_FLAG_TRYLOCK));
 	ASSERT(agno != NULLAGNUMBER);
 	error = xfs_read_agf(mp, tp, agno,
 			(flags & XFS_ALLOC_FLAG_TRYLOCK) ? XBF_TRYLOCK : 0,
 			bpp);
 	if (error)
 		return error;
-	if (!*bpp)
-		return 0;
 	ASSERT(!(*bpp)->b_error);
 
 	agf = XFS_BUF_TO_AGF(*bpp);

fs/xfs/libxfs/xfs_attr_remote.c

@@ -418,20 +418,10 @@ xfs_attr_rmtval_get(
 			       (map[i].br_startblock != HOLESTARTBLOCK));
 			dblkno = XFS_FSB_TO_DADDR(mp, map[i].br_startblock);
 			dblkcnt = XFS_FSB_TO_BB(mp, map[i].br_blockcount);
-			bp = xfs_buf_read(mp->m_ddev_targp, dblkno, dblkcnt, 0,
-					&xfs_attr3_rmt_buf_ops);
-			if (!bp)
-				return -ENOMEM;
-			error = bp->b_error;
-			if (error) {
-				xfs_buf_ioerror_alert(bp, __func__);
-				xfs_buf_relse(bp);
-
-				/* bad CRC means corrupted metadata */
-				if (error == -EFSBADCRC)
-					error = -EFSCORRUPTED;
+			error = xfs_buf_read(mp->m_ddev_targp, dblkno, dblkcnt,
+					0, &bp, &xfs_attr3_rmt_buf_ops);
+			if (error)
 				return error;
-			}
 
 			error = xfs_attr_rmtval_copyout(mp, bp, args->dp->i_ino,
 							&offset, &valuelen,
@@ -555,9 +545,9 @@ xfs_attr_rmtval_set(
 		dblkno = XFS_FSB_TO_DADDR(mp, map.br_startblock),
 		dblkcnt = XFS_FSB_TO_BB(mp, map.br_blockcount);
 
-		bp = xfs_buf_get(mp->m_ddev_targp, dblkno, dblkcnt);
-		if (!bp)
-			return -ENOMEM;
+		error = xfs_buf_get(mp->m_ddev_targp, dblkno, dblkcnt, &bp);
+		if (error)
+			return error;
 		bp->b_ops = &xfs_attr3_rmt_buf_ops;
 
 		xfs_attr_rmtval_copyin(mp, bp, args->dp->i_ino, &offset,

fs/xfs/libxfs/xfs_bmap.c

@@ -730,11 +730,11 @@ xfs_bmap_extents_to_btree(
 	cur->bc_private.b.allocated++;
 	ip->i_d.di_nblocks++;
 	xfs_trans_mod_dquot_byino(tp, ip, XFS_TRANS_DQ_BCOUNT, 1L);
-	abp = xfs_btree_get_bufl(mp, tp, args.fsbno);
-	if (XFS_IS_CORRUPT(mp, !abp)) {
-		error = -EFSCORRUPTED;
+	error = xfs_trans_get_buf(tp, mp->m_ddev_targp,
+			XFS_FSB_TO_DADDR(mp, args.fsbno),
+			mp->m_bsize, 0, &abp);
+	if (error)
 		goto out_unreserve_dquot;
-	}
 
 	/*
 	 * Fill in the child block.
@@ -878,7 +878,11 @@ xfs_bmap_local_to_extents(
 	ASSERT(args.fsbno != NULLFSBLOCK);
 	ASSERT(args.len == 1);
 	tp->t_firstblock = args.fsbno;
-	bp = xfs_btree_get_bufl(args.mp, tp, args.fsbno);
+	error = xfs_trans_get_buf(tp, args.mp->m_ddev_targp,
+			XFS_FSB_TO_DADDR(args.mp, args.fsbno),
+			args.mp->m_bsize, 0, &bp);
+	if (error)
+		goto done;
 
 	/*
 	 * Initialize the block, copy the data and log the remote buffer.
@@ -3307,11 +3311,12 @@ xfs_bmap_longest_free_extent(
 	pag = xfs_perag_get(mp, ag);
 	if (!pag->pagf_init) {
 		error = xfs_alloc_pagf_init(mp, tp, ag, XFS_ALLOC_FLAG_TRYLOCK);
-		if (error)
-			goto out;
-
-		if (!pag->pagf_init) {
+		if (error) {
+			/* Couldn't lock the AGF, so skip this AG. */
+			if (error == -EAGAIN) {
 				*notinit = 1;
+				error = 0;
+			}
 			goto out;
 		}
 	}

fs/xfs/libxfs/xfs_btree.c

@@ -678,42 +678,6 @@ xfs_btree_get_block(
 	return XFS_BUF_TO_BLOCK(*bpp);
 }
 
-/*
- * Get a buffer for the block, return it with no data read.
- * Long-form addressing.
- */
-xfs_buf_t *				/* buffer for fsbno */
-xfs_btree_get_bufl(
-	xfs_mount_t	*mp,		/* file system mount point */
-	xfs_trans_t	*tp,		/* transaction pointer */
-	xfs_fsblock_t	fsbno)		/* file system block number */
-{
-	xfs_daddr_t		d;		/* real disk block address */
-
-	ASSERT(fsbno != NULLFSBLOCK);
-	d = XFS_FSB_TO_DADDR(mp, fsbno);
-	return xfs_trans_get_buf(tp, mp->m_ddev_targp, d, mp->m_bsize, 0);
-}
-
-/*
- * Get a buffer for the block, return it with no data read.
- * Short-form addressing.
- */
-xfs_buf_t *				/* buffer for agno/agbno */
-xfs_btree_get_bufs(
-	xfs_mount_t	*mp,		/* file system mount point */
-	xfs_trans_t	*tp,		/* transaction pointer */
-	xfs_agnumber_t	agno,		/* allocation group number */
-	xfs_agblock_t	agbno)		/* allocation group block number */
-{
-	xfs_daddr_t		d;		/* real disk block address */
-
-	ASSERT(agno != NULLAGNUMBER);
-	ASSERT(agbno != NULLAGBLOCK);
-	d = XFS_AGB_TO_DADDR(mp, agno, agbno);
-	return xfs_trans_get_buf(tp, mp->m_ddev_targp, d, mp->m_bsize, 0);
-}
-
 /*
  * Change the cursor to point to the first record at the given level.
  * Other levels are unaffected.
@@ -1270,11 +1234,10 @@ xfs_btree_get_buf_block(
 	error = xfs_btree_ptr_to_daddr(cur, ptr, &d);
 	if (error)
 		return error;
-	*bpp = xfs_trans_get_buf(cur->bc_tp, mp->m_ddev_targp, d,
-				 mp->m_bsize, 0);
-
-	if (!*bpp)
-		return -ENOMEM;
+	error = xfs_trans_get_buf(cur->bc_tp, mp->m_ddev_targp, d, mp->m_bsize,
+			0, bpp);
+	if (error)
+		return error;
 
 	(*bpp)->b_ops = cur->bc_ops->buf_ops;
 	*block = XFS_BUF_TO_BLOCK(*bpp);

fs/xfs/libxfs/xfs_btree.h

@@ -296,27 +296,6 @@ xfs_btree_dup_cursor(
 	xfs_btree_cur_t		*cur,	/* input cursor */
 	xfs_btree_cur_t		**ncur);/* output cursor */
 
-/*
- * Get a buffer for the block, return it with no data read.
- * Long-form addressing.
- */
-struct xfs_buf *				/* buffer for fsbno */
-xfs_btree_get_bufl(
-	struct xfs_mount	*mp,	/* file system mount point */
-	struct xfs_trans	*tp,	/* transaction pointer */
-	xfs_fsblock_t		fsbno);	/* file system block number */
-
-/*
- * Get a buffer for the block, return it with no data read.
- * Short-form addressing.
- */
-struct xfs_buf *				/* buffer for agno/agbno */
-xfs_btree_get_bufs(
-	struct xfs_mount	*mp,	/* file system mount point */
-	struct xfs_trans	*tp,	/* transaction pointer */
-	xfs_agnumber_t		agno,	/* allocation group number */
-	xfs_agblock_t		agbno);	/* allocation group block number */
-
 /*
  * Compute first and last byte offsets for the fields given.
  * Interprets the offsets table, which contains struct field offsets.

fs/xfs/libxfs/xfs_da_btree.c

@@ -2591,13 +2591,9 @@ xfs_da_get_buf(
 	if (error || nmap == 0)
 		goto out_free;
 
-	bp = xfs_trans_get_buf_map(tp, mp->m_ddev_targp, mapp, nmap, 0);
-	error = bp ? bp->b_error : -EIO;
-	if (error) {
-		if (bp)
-			xfs_trans_brelse(tp, bp);
+	error = xfs_trans_get_buf_map(tp, mp->m_ddev_targp, mapp, nmap, 0, &bp);
+	if (error)
 		goto out_free;
-	}
 
 	*bpp = bp;
 

fs/xfs/libxfs/xfs_ialloc.c

@@ -276,6 +276,7 @@ xfs_ialloc_inode_init(
 	int			i, j;
 	xfs_daddr_t		d;
 	xfs_ino_t		ino = 0;
+	int			error;
 
 	/*
 	 * Loop over the new block(s), filling in the inodes.  For small block
@@ -327,12 +328,11 @@ xfs_ialloc_inode_init(
 		 */
 		d = XFS_AGB_TO_DADDR(mp, agno, agbno +
 				(j * M_IGEO(mp)->blocks_per_cluster));
-		fbuf = xfs_trans_get_buf(tp, mp->m_ddev_targp, d,
-					 mp->m_bsize *
-					 M_IGEO(mp)->blocks_per_cluster,
-					 XBF_UNMAPPED);
-		if (!fbuf)
-			return -ENOMEM;
+		error = xfs_trans_get_buf(tp, mp->m_ddev_targp, d,
+				mp->m_bsize * M_IGEO(mp)->blocks_per_cluster,
+				XBF_UNMAPPED, &fbuf);
+		if (error)
+			return error;
 
 		/* Initialize the inode buffers and log them appropriately. */
 		fbuf->b_ops = &xfs_inode_buf_ops;

fs/xfs/libxfs/xfs_refcount.c

@@ -1177,8 +1177,6 @@ xfs_refcount_finish_one(
 				XFS_ALLOC_FLAG_FREEING, &agbp);
 		if (error)
 			return error;
-		if (XFS_IS_CORRUPT(tp->t_mountp, !agbp))
-			return -EFSCORRUPTED;
 
 		rcur = xfs_refcountbt_init_cursor(mp, tp, agbp, agno);
 		if (!rcur) {
@@ -1718,10 +1716,6 @@ xfs_refcount_recover_cow_leftovers(
 	error = xfs_alloc_read_agf(mp, tp, agno, 0, &agbp);
 	if (error)
 		goto out_trans;
-	if (!agbp) {
-		error = -ENOMEM;
-		goto out_trans;
-	}
 	cur = xfs_refcountbt_init_cursor(mp, tp, agbp, agno);
 
 	/* Find all the leftover CoW staging extents. */

fs/xfs/libxfs/xfs_sb.c

@@ -985,9 +985,9 @@ xfs_update_secondary_sbs(
 	for (agno = 1; agno < mp->m_sb.sb_agcount; agno++) {
 		struct xfs_buf		*bp;
 
-		bp = xfs_buf_get(mp->m_ddev_targp,
+		error = xfs_buf_get(mp->m_ddev_targp,
 				 XFS_AG_DADDR(mp, agno, XFS_SB_DADDR),
-				 XFS_FSS_TO_BB(mp, 1));
+				 XFS_FSS_TO_BB(mp, 1), &bp);
 		/*
 		 * If we get an error reading or writing alternate superblocks,
 		 * continue.  xfs_repair chooses the "best" superblock based
@@ -995,12 +995,12 @@ xfs_update_secondary_sbs(
 		 * superblocks un-updated than updated, and xfs_repair may
 		 * pick them over the properly-updated primary.
 		 */
-		if (!bp) {
+		if (error) {
 			xfs_warn(mp,
 		"error allocating secondary superblock for ag %d",
 				agno);
 			if (!saved_error)
-				saved_error = -ENOMEM;
+				saved_error = error;
 			continue;
 		}
 
@@ -1185,13 +1185,14 @@ xfs_sb_get_secondary(
 	struct xfs_buf		**bpp)
 {
 	struct xfs_buf		*bp;
+	int			error;
 
 	ASSERT(agno != 0 && agno != NULLAGNUMBER);
-	bp = xfs_trans_get_buf(tp, mp->m_ddev_targp,
+	error = xfs_trans_get_buf(tp, mp->m_ddev_targp,
 			XFS_AG_DADDR(mp, agno, XFS_SB_BLOCK(mp)),
-			XFS_FSS_TO_BB(mp, 1), 0);
-	if (!bp)
-		return -ENOMEM;
+			XFS_FSS_TO_BB(mp, 1), 0, &bp);
+	if (error)
+		return error;
 	bp->b_ops = &xfs_sb_buf_ops;
 	xfs_buf_oneshot(bp);
 	*bpp = bp;

fs/xfs/scrub/agheader_repair.c

@@ -659,8 +659,6 @@ xrep_agfl(
 	error = xfs_alloc_read_agf(mp, sc->tp, sc->sa.agno, 0, &agf_bp);
 	if (error)
 		return error;
-	if (!agf_bp)
-		return -ENOMEM;
 
 	/*
 	 * Make sure we have the AGFL buffer, as scrub might have decided it
@@ -735,8 +733,6 @@ xrep_agi_find_btrees(
 	error = xfs_alloc_read_agf(mp, sc->tp, sc->sa.agno, 0, &agf_bp);
 	if (error)
 		return error;
-	if (!agf_bp)
-		return -ENOMEM;
 
 	/* Find the btree roots. */
 	error = xrep_find_ag_btree_roots(sc, agf_bp, fab, NULL);

fs/xfs/scrub/fscounters.c

@@ -83,9 +83,6 @@ xchk_fscount_warmup(
 		error = xfs_alloc_read_agf(mp, sc->tp, agno, 0, &agf_bp);
 		if (error)
 			break;
-		error = -ENOMEM;
-		if (!agf_bp || !agi_bp)
-			break;
 
 		/*
 		 * These are supposed to be initialized by the header read

fs/xfs/scrub/repair.c

@@ -341,13 +341,17 @@ xrep_init_btblock(
 	struct xfs_trans		*tp = sc->tp;
 	struct xfs_mount		*mp = sc->mp;
 	struct xfs_buf			*bp;
+	int				error;
 
 	trace_xrep_init_btblock(mp, XFS_FSB_TO_AGNO(mp, fsb),
 			XFS_FSB_TO_AGBNO(mp, fsb), btnum);
 
 	ASSERT(XFS_FSB_TO_AGNO(mp, fsb) == sc->sa.agno);
-	bp = xfs_trans_get_buf(tp, mp->m_ddev_targp, XFS_FSB_TO_DADDR(mp, fsb),
-			XFS_FSB_TO_BB(mp, 1), 0);
+	error = xfs_trans_get_buf(tp, mp->m_ddev_targp,
+			XFS_FSB_TO_DADDR(mp, fsb), XFS_FSB_TO_BB(mp, 1), 0,
+			&bp);
+	if (error)
+		return error;
 	xfs_buf_zero(bp, 0, BBTOB(bp->b_length));
 	xfs_btree_init_block(mp, bp, btnum, 0, 0, sc->sa.agno);
 	xfs_trans_buf_set_type(tp, bp, XFS_BLFT_BTREE_BUF);
@@ -542,8 +546,6 @@ xrep_reap_block(
 		error = xfs_alloc_read_agf(sc->mp, sc->tp, agno, 0, &agf_bp);
 		if (error)
 			return error;
-		if (!agf_bp)
-			return -ENOMEM;
 	} else {
 		agf_bp = sc->sa.agf_bp;
 	}

fs/xfs/xfs_attr_inactive.c

@@ -205,11 +205,12 @@ xfs_attr3_node_inactive(
 		/*
 		 * Remove the subsidiary block from the cache and from the log.
 		 */
-		child_bp = xfs_trans_get_buf(*trans, mp->m_ddev_targp,
+		error = xfs_trans_get_buf(*trans, mp->m_ddev_targp,
 				child_blkno,
-				XFS_FSB_TO_BB(mp, mp->m_attr_geo->fsbcount), 0);
-		if (!child_bp)
-			return -EIO;
+				XFS_FSB_TO_BB(mp, mp->m_attr_geo->fsbcount), 0,
+				&child_bp);
+		if (error)
+			return error;
 		error = bp->b_error;
 		if (error) {
 			xfs_trans_brelse(*trans, child_bp);
@@ -298,10 +299,10 @@ xfs_attr3_root_inactive(
 	/*
 	 * Invalidate the incore copy of the root block.
 	 */
-	bp = xfs_trans_get_buf(*trans, mp->m_ddev_targp, blkno,
-			XFS_FSB_TO_BB(mp, mp->m_attr_geo->fsbcount), 0);
-	if (!bp)
-		return -EIO;
+	error = xfs_trans_get_buf(*trans, mp->m_ddev_targp, blkno,
+			XFS_FSB_TO_BB(mp, mp->m_attr_geo->fsbcount), 0, &bp);
+	if (error)
+		return error;
 	error = bp->b_error;
 	if (error) {
 		xfs_trans_brelse(*trans, bp);

fs/xfs/xfs_buf.c

@@ -198,20 +198,22 @@ xfs_buf_free_maps(
 	}
 }
 
-static struct xfs_buf *
+static int
 _xfs_buf_alloc(
 	struct xfs_buftarg	*target,
 	struct xfs_buf_map	*map,
 	int			nmaps,
-	xfs_buf_flags_t		flags)
+	xfs_buf_flags_t		flags,
+	struct xfs_buf		**bpp)
 {
 	struct xfs_buf		*bp;
 	int			error;
 	int			i;
 
+	*bpp = NULL;
 	bp = kmem_zone_zalloc(xfs_buf_zone, KM_NOFS);
 	if (unlikely(!bp))
-		return NULL;
+		return -ENOMEM;
 
 	/*
 	 * We don't want certain flags to appear in b_flags unless they are
@@ -239,7 +241,7 @@ _xfs_buf_alloc(
 	error = xfs_buf_get_maps(bp, nmaps);
 	if (error)  {
 		kmem_cache_free(xfs_buf_zone, bp);
-		return NULL;
+		return error;
 	}
 
 	bp->b_bn = map[0].bm_bn;
@@ -256,7 +258,8 @@ _xfs_buf_alloc(
 	XFS_STATS_INC(bp->b_mount, xb_create);
 	trace_xfs_buf_init(bp, _RET_IP_);
 
-	return bp;
+	*bpp = bp;
+	return 0;
 }
 
 /*
@@ -682,53 +685,39 @@ xfs_buf_incore(
  * cache hits, as metadata intensive workloads will see 3 orders of magnitude
  * more hits than misses.
  */
-struct xfs_buf *
+int
 xfs_buf_get_map(
 	struct xfs_buftarg	*target,
 	struct xfs_buf_map	*map,
 	int			nmaps,
-	xfs_buf_flags_t		flags)
+	xfs_buf_flags_t		flags,
+	struct xfs_buf		**bpp)
 {
 	struct xfs_buf		*bp;
 	struct xfs_buf		*new_bp;
 	int			error = 0;
 
+	*bpp = NULL;
 	error = xfs_buf_find(target, map, nmaps, flags, NULL, &bp);
-
-	switch (error) {
-	case 0:
-		/* cache hit */
+	if (!error)
 		goto found;
-	case -EAGAIN:
-		/* cache hit, trylock failure, caller handles failure */
-		ASSERT(flags & XBF_TRYLOCK);
-		return NULL;
-	case -ENOENT:
-		/* cache miss, go for insert */
-		break;
-	case -EFSCORRUPTED:
-	default:
-		/*
-		 * None of the higher layers understand failure types
-		 * yet, so return NULL to signal a fatal lookup error.
-		 */
-		return NULL;
-	}
+	if (error != -ENOENT)
+		return error;
 
-	new_bp = _xfs_buf_alloc(target, map, nmaps, flags);
-	if (unlikely(!new_bp))
-		return NULL;
+	error = _xfs_buf_alloc(target, map, nmaps, flags, &new_bp);
+	if (error)
+		return error;
 
 	error = xfs_buf_allocate_memory(new_bp, flags);
 	if (error) {
 		xfs_buf_free(new_bp);
-		return NULL;
+		return error;
 	}
 
 	error = xfs_buf_find(target, map, nmaps, flags, new_bp, &bp);
 	if (error) {
 		xfs_buf_free(new_bp);
-		return NULL;
+		return error;
 	}
 
 	if (bp != new_bp)
@@ -741,7 +730,7 @@ xfs_buf_get_map(
 			xfs_warn(target->bt_mount,
 				"%s: failed to map pagesn", __func__);
 			xfs_buf_relse(bp);
-			return NULL;
+			return error;
 		}
 	}
 
@@ -754,7 +743,8 @@ xfs_buf_get_map(
 
 	XFS_STATS_INC(target->bt_mount, xb_get);
 	trace_xfs_buf_get(bp, flags, _RET_IP_);
-	return bp;
+	*bpp = bp;
+	return 0;
 }
 
 STATIC int
@@ -806,46 +796,77 @@ xfs_buf_reverify(
 	return bp->b_error;
 }
 
-xfs_buf_t *
+int
 xfs_buf_read_map(
 	struct xfs_buftarg	*target,
 	struct xfs_buf_map	*map,
 	int			nmaps,
 	xfs_buf_flags_t		flags,
-	const struct xfs_buf_ops *ops)
+	struct xfs_buf		**bpp,
+	const struct xfs_buf_ops *ops,
+	xfs_failaddr_t		fa)
 {
 	struct xfs_buf		*bp;
+	int			error;
 
 	flags |= XBF_READ;
+	*bpp = NULL;
 
-	bp = xfs_buf_get_map(target, map, nmaps, flags);
-	if (!bp)
-		return NULL;
+	error = xfs_buf_get_map(target, map, nmaps, flags, &bp);
+	if (error)
+		return error;
 
 	trace_xfs_buf_read(bp, flags, _RET_IP_);
 
 	if (!(bp->b_flags & XBF_DONE)) {
+		/* Initiate the buffer read and wait. */
 		XFS_STATS_INC(target->bt_mount, xb_get_read);
 		bp->b_ops = ops;
-		_xfs_buf_read(bp, flags);
-		return bp;
-	}
+		error = _xfs_buf_read(bp, flags);
 
-	xfs_buf_reverify(bp, ops);
+		/* Readahead iodone already dropped the buffer, so exit. */
+		if (flags & XBF_ASYNC)
+			return 0;
+	} else {
+		/* Buffer already read; all we need to do is check it. */
+		error = xfs_buf_reverify(bp, ops);
 
+		/* Readahead already finished; drop the buffer and exit. */
 		if (flags & XBF_ASYNC) {
-		/*
-		 * Read ahead call which is already satisfied,
-		 * drop the buffer
-		 */
 			xfs_buf_relse(bp);
-		return NULL;
+			return 0;
 		}
 
 		/* We do not want read in the flags */
 		bp->b_flags &= ~XBF_READ;
 		ASSERT(bp->b_ops != NULL || ops == NULL);
-	return bp;
+	}
+
+	/*
+	 * If we've had a read error, then the contents of the buffer are
+	 * invalid and should not be used. To ensure that a followup read tries
+	 * to pull the buffer from disk again, we clear the XBF_DONE flag and
+	 * mark the buffer stale. This ensures that anyone who has a current
+	 * reference to the buffer will interpret it's contents correctly and
+	 * future cache lookups will also treat it as an empty, uninitialised
+	 * buffer.
+	 */
+	if (error) {
+		if (!XFS_FORCED_SHUTDOWN(target->bt_mount))
+			xfs_buf_ioerror_alert(bp, fa);
+
+		bp->b_flags &= ~XBF_DONE;
+		xfs_buf_stale(bp);
+		xfs_buf_relse(bp);
+
+		/* bad CRC means corrupted metadata */
+		if (error == -EFSBADCRC)
+			error = -EFSCORRUPTED;
+		return error;
+	}
+
+	*bpp = bp;
+	return 0;
 }
 
 /*
@@ -859,11 +880,14 @@ xfs_buf_readahead_map(
 	int			nmaps,
 	const struct xfs_buf_ops *ops)
 {
+	struct xfs_buf		*bp;
+
 	if (bdi_read_congested(target->bt_bdev->bd_bdi))
 		return;
 
 	xfs_buf_read_map(target, map, nmaps,
-		     XBF_TRYLOCK|XBF_ASYNC|XBF_READ_AHEAD, ops);
+		     XBF_TRYLOCK | XBF_ASYNC | XBF_READ_AHEAD, &bp, ops,
+		     __this_address);
 }
 
 /*
@@ -880,12 +904,13 @@ xfs_buf_read_uncached(
 	const struct xfs_buf_ops *ops)
 {
 	struct xfs_buf		*bp;
+	int			error;
 
 	*bpp = NULL;
 
-	bp = xfs_buf_get_uncached(target, numblks, flags);
-	if (!bp)
-		return -ENOMEM;
+	error = xfs_buf_get_uncached(target, numblks, flags, &bp);
+	if (error)
+		return error;
 
 	/* set up the buffer for a read IO */
 	ASSERT(bp->b_map_count == 1);
@@ -896,7 +921,7 @@ xfs_buf_read_uncached(
 
 	xfs_buf_submit(bp);
 	if (bp->b_error) {
-		int	error = bp->b_error;
+		error = bp->b_error;
 		xfs_buf_relse(bp);
 		return error;
 	}
@@ -905,20 +930,23 @@ xfs_buf_read_uncached(
 	return 0;
 }
 
-xfs_buf_t *
+int
 xfs_buf_get_uncached(
 	struct xfs_buftarg	*target,
 	size_t			numblks,
-	int			flags)
+	int			flags,
+	struct xfs_buf		**bpp)
 {
 	unsigned long		page_count;
 	int			error, i;
 	struct xfs_buf		*bp;
 	DEFINE_SINGLE_BUF_MAP(map, XFS_BUF_DADDR_NULL, numblks);
 
+	*bpp = NULL;
+
 	/* flags might contain irrelevant bits, pass only what we care about */
-	bp = _xfs_buf_alloc(target, &map, 1, flags & XBF_NO_IOACCT);
-	if (unlikely(bp == NULL))
+	error = _xfs_buf_alloc(target, &map, 1, flags & XBF_NO_IOACCT, &bp);
+	if (error)
 		goto fail;
 
 	page_count = PAGE_ALIGN(numblks << BBSHIFT) >> PAGE_SHIFT;
@@ -928,9 +956,11 @@ xfs_buf_get_uncached(
 
 	for (i = 0; i < page_count; i++) {
 		bp->b_pages[i] = alloc_page(xb_to_gfp(flags));
-		if (!bp->b_pages[i])
+		if (!bp->b_pages[i]) {
+			error = -ENOMEM;
 			goto fail_free_mem;
 		}
+	}
 	bp->b_flags |= _XBF_PAGES;
 
 	error = _xfs_buf_map_pages(bp, 0);
@@ -941,7 +971,8 @@ xfs_buf_get_uncached(
 	}
 
 	trace_xfs_buf_get_uncached(bp, _RET_IP_);
-	return bp;
+	*bpp = bp;
+	return 0;
 
  fail_free_mem:
 	while (--i >= 0)
@@ -951,7 +982,7 @@ xfs_buf_get_uncached(
 	xfs_buf_free_maps(bp);
 	kmem_cache_free(xfs_buf_zone, bp);
  fail:
-	return NULL;
+	return error;
 }
 
 /*
@@ -1205,10 +1236,10 @@ __xfs_buf_ioerror(
 void
 xfs_buf_ioerror_alert(
 	struct xfs_buf		*bp,
-	const char		*func)
+	xfs_failaddr_t		func)
 {
 	xfs_alert(bp->b_mount,
-"metadata I/O error in \"%s\" at daddr 0x%llx len %d error %d",
+"metadata I/O error in \"%pS\" at daddr 0x%llx len %d error %d",
 			func, (uint64_t)XFS_BUF_ADDR(bp), bp->b_length,
 			-bp->b_error);
 }

fs/xfs/xfs_buf.h

@@ -192,37 +192,40 @@ struct xfs_buf *xfs_buf_incore(struct xfs_buftarg *target,
 			   xfs_daddr_t blkno, size_t numblks,
 			   xfs_buf_flags_t flags);
 
-struct xfs_buf *xfs_buf_get_map(struct xfs_buftarg *target,
-			       struct xfs_buf_map *map, int nmaps,
-			       xfs_buf_flags_t flags);
-struct xfs_buf *xfs_buf_read_map(struct xfs_buftarg *target,
-			       struct xfs_buf_map *map, int nmaps,
-			       xfs_buf_flags_t flags,
-			       const struct xfs_buf_ops *ops);
+int xfs_buf_get_map(struct xfs_buftarg *target, struct xfs_buf_map *map,
+		int nmaps, xfs_buf_flags_t flags, struct xfs_buf **bpp);
+int xfs_buf_read_map(struct xfs_buftarg *target, struct xfs_buf_map *map,
+		int nmaps, xfs_buf_flags_t flags, struct xfs_buf **bpp,
+		const struct xfs_buf_ops *ops, xfs_failaddr_t fa);
 void xfs_buf_readahead_map(struct xfs_buftarg *target,
 			       struct xfs_buf_map *map, int nmaps,
 			       const struct xfs_buf_ops *ops);
 
-static inline struct xfs_buf *
+static inline int
 xfs_buf_get(
 	struct xfs_buftarg	*target,
 	xfs_daddr_t		blkno,
-	size_t			numblks)
+	size_t			numblks,
+	struct xfs_buf		**bpp)
 {
 	DEFINE_SINGLE_BUF_MAP(map, blkno, numblks);
-	return xfs_buf_get_map(target, &map, 1, 0);
+
+	return xfs_buf_get_map(target, &map, 1, 0, bpp);
 }
 
-static inline struct xfs_buf *
+static inline int
 xfs_buf_read(
 	struct xfs_buftarg	*target,
 	xfs_daddr_t		blkno,
 	size_t			numblks,
 	xfs_buf_flags_t		flags,
+	struct xfs_buf		**bpp,
 	const struct xfs_buf_ops *ops)
 {
 	DEFINE_SINGLE_BUF_MAP(map, blkno, numblks);
-	return xfs_buf_read_map(target, &map, 1, flags, ops);
+
+	return xfs_buf_read_map(target, &map, 1, flags, bpp, ops,
+			__builtin_return_address(0));
 }
 
 static inline void
@@ -236,8 +239,8 @@ xfs_buf_readahead(
 	return xfs_buf_readahead_map(target, &map, 1, ops);
 }
 
-struct xfs_buf *xfs_buf_get_uncached(struct xfs_buftarg *target, size_t numblks,
-				int flags);
+int xfs_buf_get_uncached(struct xfs_buftarg *target, size_t numblks, int flags,
+		struct xfs_buf **bpp);
 int xfs_buf_read_uncached(struct xfs_buftarg *target, xfs_daddr_t daddr,
 			  size_t numblks, int flags, struct xfs_buf **bpp,
 			  const struct xfs_buf_ops *ops);
@@ -259,7 +262,7 @@ extern void xfs_buf_ioend(struct xfs_buf *bp);
 extern void __xfs_buf_ioerror(struct xfs_buf *bp, int error,
 		xfs_failaddr_t failaddr);
 #define xfs_buf_ioerror(bp, err) __xfs_buf_ioerror((bp), (err), __this_address)
-extern void xfs_buf_ioerror_alert(struct xfs_buf *, const char *func);
+extern void xfs_buf_ioerror_alert(struct xfs_buf *bp, xfs_failaddr_t fa);
 
 extern int __xfs_buf_submit(struct xfs_buf *bp, bool);
 static inline int xfs_buf_submit(struct xfs_buf *bp)

fs/xfs/xfs_buf_item.c

@@ -1113,7 +1113,7 @@ xfs_buf_iodone_callback_error(
 	if (bp->b_target != lasttarg ||
 	    time_after(jiffies, (lasttime + 5*HZ))) {
 		lasttime = jiffies;
-		xfs_buf_ioerror_alert(bp, __func__);
+		xfs_buf_ioerror_alert(bp, __this_address);
 	}
 	lasttarg = bp->b_target;
 

fs/xfs/xfs_discard.c

@@ -45,7 +45,7 @@ xfs_trim_extents(
 	xfs_log_force(mp, XFS_LOG_SYNC);
 
 	error = xfs_alloc_read_agf(mp, NULL, agno, 0, &agbp);
-	if (error || !agbp)
+	if (error)
 		goto out_put_perag;
 
 	cur = xfs_allocbt_init_cursor(mp, NULL, agbp, agno, XFS_BTNUM_CNT);

fs/xfs/xfs_dquot.c

@@ -320,10 +320,10 @@ xfs_dquot_disk_alloc(
 	dqp->q_blkno = XFS_FSB_TO_DADDR(mp, map.br_startblock);
 
 	/* now we can just get the buffer (there's nothing to read yet) */
-	bp = xfs_trans_get_buf(tp, mp->m_ddev_targp, dqp->q_blkno,
-			mp->m_quotainfo->qi_dqchunklen, 0);
-	if (!bp)
-		return -ENOMEM;
+	error = xfs_trans_get_buf(tp, mp->m_ddev_targp, dqp->q_blkno,
+			mp->m_quotainfo->qi_dqchunklen, 0, &bp);
+	if (error)
+		return error;
 	bp->b_ops = &xfs_dquot_buf_ops;
 
 	/*

fs/xfs/xfs_filestream.c

@@ -159,16 +159,15 @@ xfs_filestream_pick_ag(
 
 		if (!pag->pagf_init) {
 			err = xfs_alloc_pagf_init(mp, NULL, ag, trylock);
-			if (err && !trylock) {
+			if (err) {
 				xfs_perag_put(pag);
+				if (err != -EAGAIN)
 					return err;
+				/* Couldn't lock the AGF, skip this AG. */
+				continue;
 			}
 		}
 
-		/* Might fail sometimes during the 1st pass with trylock set. */
-		if (!pag->pagf_init)
-			goto next_ag;
-
 		/* Keep track of the AG with the most free blocks. */
 		if (pag->pagf_freeblks > maxfree) {
 			maxfree = pag->pagf_freeblks;

fs/xfs/xfs_inode.c

@@ -2546,6 +2546,7 @@ xfs_ifree_cluster(
 	struct xfs_perag	*pag;
 	struct xfs_ino_geometry	*igeo = M_IGEO(mp);
 	xfs_ino_t		inum;
+	int			error;
 
 	inum = xic->first_ino;
 	pag = xfs_perag_get(mp, XFS_INO_TO_AGNO(mp, inum));
@@ -2574,12 +2575,11 @@ xfs_ifree_cluster(
 		 * complete before we get a lock on it, and hence we may fail
 		 * to mark all the active inodes on the buffer stale.
 		 */
-		bp = xfs_trans_get_buf(tp, mp->m_ddev_targp, blkno,
+		error = xfs_trans_get_buf(tp, mp->m_ddev_targp, blkno,
 				mp->m_bsize * igeo->blocks_per_cluster,
-					XBF_UNMAPPED);
-
-		if (!bp)
-			return -ENOMEM;
+				XBF_UNMAPPED, &bp);
+		if (error)
+			return error;
 
 		/*
 		 * This buffer may not have been correctly initialised as we

fs/xfs/xfs_log_recover.c

@@ -294,7 +294,7 @@ xlog_recover_iodone(
 		 * this during recovery. One strike!
 		 */
 		if (!XFS_FORCED_SHUTDOWN(bp->b_mount)) {
-			xfs_buf_ioerror_alert(bp, __func__);
+			xfs_buf_ioerror_alert(bp, __this_address);
 			xfs_force_shutdown(bp->b_mount, SHUTDOWN_META_IO_ERROR);
 		}
 	}
@@ -2745,15 +2745,10 @@ xlog_recover_buffer_pass2(
 	if (buf_f->blf_flags & XFS_BLF_INODE_BUF)
 		buf_flags |= XBF_UNMAPPED;
 
-	bp = xfs_buf_read(mp->m_ddev_targp, buf_f->blf_blkno, buf_f->blf_len,
-			  buf_flags, NULL);
-	if (!bp)
-		return -ENOMEM;
-	error = bp->b_error;
-	if (error) {
-		xfs_buf_ioerror_alert(bp, "xlog_recover_do..(read#1)");
-		goto out_release;
-	}
+	error = xfs_buf_read(mp->m_ddev_targp, buf_f->blf_blkno, buf_f->blf_len,
+			  buf_flags, &bp, NULL);
+	if (error)
+		return error;
 
 	/*
 	 * Recover the buffer only if we get an LSN from it and it's less than
@@ -2950,17 +2945,10 @@ xlog_recover_inode_pass2(
 	}
 	trace_xfs_log_recover_inode_recover(log, in_f);
 
-	bp = xfs_buf_read(mp->m_ddev_targp, in_f->ilf_blkno, in_f->ilf_len, 0,
-			  &xfs_inode_buf_ops);
-	if (!bp) {
-		error = -ENOMEM;
+	error = xfs_buf_read(mp->m_ddev_targp, in_f->ilf_blkno, in_f->ilf_len,
+			0, &bp, &xfs_inode_buf_ops);
+	if (error)
 		goto error;
-	}
-	error = bp->b_error;
-	if (error) {
-		xfs_buf_ioerror_alert(bp, "xlog_recover_do..(read#2)");
-		goto out_release;
-	}
 	ASSERT(in_f->ilf_fields & XFS_ILOG_CORE);
 	dip = xfs_buf_offset(bp, in_f->ilf_boffset);
 
@@ -5639,7 +5627,7 @@ xlog_do_recover(
 	error = xfs_buf_submit(bp);
 	if (error) {
 		if (!XFS_FORCED_SHUTDOWN(mp)) {
-			xfs_buf_ioerror_alert(bp, __func__);
+			xfs_buf_ioerror_alert(bp, __this_address);
 			ASSERT(0);
 		}
 		xfs_buf_relse(bp);

fs/xfs/xfs_reflink.c

@@ -143,8 +143,6 @@ xfs_reflink_find_shared(
 	error = xfs_alloc_read_agf(mp, tp, agno, 0, &agbp);
 	if (error)
 		return error;
-	if (!agbp)
-		return -ENOMEM;
 
 	cur = xfs_refcountbt_init_cursor(mp, tp, agbp, agno);
 

fs/xfs/xfs_rtalloc.c

@@ -826,12 +826,10 @@ xfs_growfs_rt_alloc(
 			 * Get a buffer for the block.
 			 */
 			d = XFS_FSB_TO_DADDR(mp, fsbno);
-			bp = xfs_trans_get_buf(tp, mp->m_ddev_targp, d,
-				mp->m_bsize, 0);
-			if (bp == NULL) {
-				error = -EIO;
+			error = xfs_trans_get_buf(tp, mp->m_ddev_targp, d,
+					mp->m_bsize, 0, &bp);
+			if (error)
 				goto out_trans_cancel;
-			}
 			memset(bp->b_addr, 0, mp->m_sb.sb_blocksize);
 			xfs_trans_log_buf(tp, bp, 0, mp->m_sb.sb_blocksize - 1);
 			/*

fs/xfs/xfs_symlink.c

@@ -53,20 +53,10 @@ xfs_readlink_bmap_ilocked(
 		d = XFS_FSB_TO_DADDR(mp, mval[n].br_startblock);
 		byte_cnt = XFS_FSB_TO_B(mp, mval[n].br_blockcount);
 
-		bp = xfs_buf_read(mp->m_ddev_targp, d, BTOBB(byte_cnt), 0,
-				  &xfs_symlink_buf_ops);
-		if (!bp)
-			return -ENOMEM;
-		error = bp->b_error;
-		if (error) {
-			xfs_buf_ioerror_alert(bp, __func__);
-			xfs_buf_relse(bp);
-
-			/* bad CRC means corrupted metadata */
-			if (error == -EFSBADCRC)
-				error = -EFSCORRUPTED;
-			goto out;
-		}
+		error = xfs_buf_read(mp->m_ddev_targp, d, BTOBB(byte_cnt), 0,
+				&bp, &xfs_symlink_buf_ops);
+		if (error)
+			return error;
 		byte_cnt = XFS_SYMLINK_BUF_SPACE(mp, byte_cnt);
 		if (pathlen < byte_cnt)
 			byte_cnt = pathlen;
@@ -290,12 +280,10 @@ xfs_symlink(
 
 			d = XFS_FSB_TO_DADDR(mp, mval[n].br_startblock);
 			byte_cnt = XFS_FSB_TO_B(mp, mval[n].br_blockcount);
-			bp = xfs_trans_get_buf(tp, mp->m_ddev_targp, d,
-					       BTOBB(byte_cnt), 0);
-			if (!bp) {
-				error = -ENOMEM;
+			error = xfs_trans_get_buf(tp, mp->m_ddev_targp, d,
+					       BTOBB(byte_cnt), 0, &bp);
+			if (error)
 				goto out_trans_cancel;
-			}
 			bp->b_ops = &xfs_symlink_buf_ops;
 
 			byte_cnt = XFS_SYMLINK_BUF_SPACE(mp, byte_cnt);
@@ -433,13 +421,12 @@ xfs_inactive_symlink_rmt(
 	 * Invalidate the block(s). No validation is done.
 	 */
 	for (i = 0; i < nmaps; i++) {
-		bp = xfs_trans_get_buf(tp, mp->m_ddev_targp,
+		error = xfs_trans_get_buf(tp, mp->m_ddev_targp,
 				XFS_FSB_TO_DADDR(mp, mval[i].br_startblock),
-			XFS_FSB_TO_BB(mp, mval[i].br_blockcount), 0);
-		if (!bp) {
-			error = -ENOMEM;
+				XFS_FSB_TO_BB(mp, mval[i].br_blockcount), 0,
+				&bp);
+		if (error)
 			goto error_trans_cancel;
-		}
 		xfs_trans_binval(tp, bp);
 	}
 	/*

fs/xfs/xfs_trans.h

@@ -169,21 +169,21 @@ int		xfs_trans_alloc_empty(struct xfs_mount *mp,
 			struct xfs_trans **tpp);
 void		xfs_trans_mod_sb(xfs_trans_t *, uint, int64_t);
 
-struct xfs_buf	*xfs_trans_get_buf_map(struct xfs_trans *tp,
-				       struct xfs_buftarg *target,
-				       struct xfs_buf_map *map, int nmaps,
-				       uint flags);
+int xfs_trans_get_buf_map(struct xfs_trans *tp, struct xfs_buftarg *target,
+		struct xfs_buf_map *map, int nmaps, xfs_buf_flags_t flags,
+		struct xfs_buf **bpp);
 
-static inline struct xfs_buf *
+static inline int
 xfs_trans_get_buf(
 	struct xfs_trans	*tp,
 	struct xfs_buftarg	*target,
 	xfs_daddr_t		blkno,
 	int			numblks,
-	uint			flags)
+	uint			flags,
+	struct xfs_buf		**bpp)
 {
 	DEFINE_SINGLE_BUF_MAP(map, blkno, numblks);
-	return xfs_trans_get_buf_map(tp, target, &map, 1, flags);
+	return xfs_trans_get_buf_map(tp, target, &map, 1, flags, bpp);
 }
 
 int		xfs_trans_read_buf_map(struct xfs_mount *mp,

fs/xfs/xfs_trans_buf.c

@@ -112,19 +112,22 @@ xfs_trans_bjoin(
  * If the transaction pointer is NULL, make this just a normal
  * get_buf() call.
  */
-struct xfs_buf *
+int
 xfs_trans_get_buf_map(
 	struct xfs_trans	*tp,
 	struct xfs_buftarg	*target,
 	struct xfs_buf_map	*map,
 	int			nmaps,
-	xfs_buf_flags_t		flags)
+	xfs_buf_flags_t		flags,
+	struct xfs_buf		**bpp)
 {
 	xfs_buf_t		*bp;
 	struct xfs_buf_log_item	*bip;
+	int			error;
 
+	*bpp = NULL;
 	if (!tp)
-		return xfs_buf_get_map(target, map, nmaps, flags);
+		return xfs_buf_get_map(target, map, nmaps, flags, bpp);
 
 	/*
 	 * If we find the buffer in the cache with this transaction
@@ -146,19 +149,20 @@ xfs_trans_get_buf_map(
 		ASSERT(atomic_read(&bip->bli_refcount) > 0);
 		bip->bli_recur++;
 		trace_xfs_trans_get_buf_recur(bip);
-		return bp;
+		*bpp = bp;
+		return 0;
 	}
 
-	bp = xfs_buf_get_map(target, map, nmaps, flags);
-	if (bp == NULL) {
-		return NULL;
-	}
+	error = xfs_buf_get_map(target, map, nmaps, flags, &bp);
+	if (error)
+		return error;
 
 	ASSERT(!bp->b_error);
 
 	_xfs_trans_bjoin(tp, bp, 1);
 	trace_xfs_trans_get_buf(bp->b_log_item);
-	return bp;
+	*bpp = bp;
+	return 0;
 }
 
 /*
@@ -276,7 +280,7 @@ xfs_trans_read_buf_map(
 		ASSERT(bp->b_ops != NULL);
 		error = xfs_buf_reverify(bp, ops);
 		if (error) {
-			xfs_buf_ioerror_alert(bp, __func__);
+			xfs_buf_ioerror_alert(bp, __return_address);
 
 			if (tp->t_flags & XFS_TRANS_DIRTY)
 				xfs_force_shutdown(tp->t_mountp,
@@ -298,36 +302,17 @@ xfs_trans_read_buf_map(
 		return 0;
 	}
 
-	bp = xfs_buf_read_map(target, map, nmaps, flags, ops);
-	if (!bp) {
-		if (!(flags & XBF_TRYLOCK))
-			return -ENOMEM;
-		return tp ? 0 : -EAGAIN;
-	}
-
-	/*
-	 * If we've had a read error, then the contents of the buffer are
-	 * invalid and should not be used. To ensure that a followup read tries
-	 * to pull the buffer from disk again, we clear the XBF_DONE flag and
-	 * mark the buffer stale. This ensures that anyone who has a current
-	 * reference to the buffer will interpret it's contents correctly and
-	 * future cache lookups will also treat it as an empty, uninitialised
-	 * buffer.
-	 */
-	if (bp->b_error) {
-		error = bp->b_error;
-		if (!XFS_FORCED_SHUTDOWN(mp))
-			xfs_buf_ioerror_alert(bp, __func__);
-		bp->b_flags &= ~XBF_DONE;
-		xfs_buf_stale(bp);
-
+	error = xfs_buf_read_map(target, map, nmaps, flags, &bp, ops,
+			__return_address);
+	switch (error) {
+	case 0:
+		break;
+	default:
 		if (tp && (tp->t_flags & XFS_TRANS_DIRTY))
 			xfs_force_shutdown(tp->t_mountp, SHUTDOWN_META_IO_ERROR);
-		xfs_buf_relse(bp);
-
-		/* bad CRC means corrupted metadata */
-		if (error == -EFSBADCRC)
-			error = -EFSCORRUPTED;
+		/* fall through */
+	case -ENOMEM:
+	case -EAGAIN:
 		return error;
 	}