summaryrefslogtreecommitdiff
diff options
context:
space:
mode:
Diffstat
-rw-r--r--Documentation/filesystems/xfs.txt29
-rw-r--r--fs/xfs/libxfs/xfs_alloc.c104
-rw-r--r--fs/xfs/libxfs/xfs_bmap.c125
-rw-r--r--fs/xfs/libxfs/xfs_btree.c24
-rw-r--r--fs/xfs/libxfs/xfs_dir2_data.c39
-rw-r--r--fs/xfs/libxfs/xfs_ialloc.c42
-rw-r--r--fs/xfs/libxfs/xfs_sb.c8
-rw-r--r--fs/xfs/xfs_bmap_util.c31
-rw-r--r--fs/xfs/xfs_buf_item.c4
-rw-r--r--fs/xfs/xfs_discard.c2
-rw-r--r--fs/xfs/xfs_error.h8
-rw-r--r--fs/xfs/xfs_file.c84
-rw-r--r--fs/xfs/xfs_icache.c4
-rw-r--r--fs/xfs/xfs_inode.c554
-rw-r--r--fs/xfs/xfs_inode.h58
-rw-r--r--fs/xfs/xfs_ioctl.c5
-rw-r--r--fs/xfs/xfs_iops.c125
-rw-r--r--fs/xfs/xfs_iops.h2
-rw-r--r--fs/xfs/xfs_itable.c2
-rw-r--r--fs/xfs/xfs_pnfs.c4
-rw-r--r--fs/xfs/xfs_qm.c18
-rw-r--r--fs/xfs/xfs_super.c23
-rw-r--r--fs/xfs/xfs_symlink.c58
-rw-r--r--fs/xfs/xfs_trace.h3
24 files changed, 776 insertions, 580 deletions
diff --git a/Documentation/filesystems/xfs.txt b/Documentation/filesystems/xfs.txt
index 0bfafe108357..5a5a05582b58 100644
--- a/Documentation/filesystems/xfs.txt
+++ b/Documentation/filesystems/xfs.txt
@@ -228,30 +228,19 @@ default behaviour.
Deprecated Mount Options
========================
- delaylog/nodelaylog
- Delayed logging is the only logging method that XFS supports
- now, so these mount options are now ignored.
-
- Due for removal in 3.12.
-
- ihashsize=value
- In memory inode hashes have been removed, so this option has
- no function as of August 2007. Option is deprecated.
-
- Due for removal in 3.12.
+None at present.
- irixsgid
- This behaviour is now controlled by a sysctl, so the mount
- option is ignored.
- Due for removal in 3.12.
+Removed Mount Options
+=====================
- osyncisdsync
- osyncisosync
- O_SYNC and O_DSYNC are fully supported, so there is no need
- for these options any more.
+ Name Removed
+ ---- -------
+ delaylog/nodelaylog v3.20
+ ihashsize v3.20
+ irixsgid v3.20
+ osyncisdsync/osyncisosync v3.20
- Due for removal in 3.12.
sysctls
=======
diff --git a/fs/xfs/libxfs/xfs_alloc.c b/fs/xfs/libxfs/xfs_alloc.c
index a6fbf4472017..516162be1398 100644
--- a/fs/xfs/libxfs/xfs_alloc.c
+++ b/fs/xfs/libxfs/xfs_alloc.c
@@ -260,6 +260,7 @@ xfs_alloc_fix_len(
rlen = rlen - (k - args->mod);
else
rlen = rlen - args->prod + (args->mod - k);
+ /* casts to (int) catch length underflows */
if ((int)rlen < (int)args->minlen)
return;
ASSERT(rlen >= args->minlen && rlen <= args->maxlen);
@@ -286,7 +287,8 @@ xfs_alloc_fix_minleft(
if (diff >= 0)
return 1;
args->len += diff; /* shrink the allocated space */
- if (args->len >= args->minlen)
+ /* casts to (int) catch length underflows */
+ if ((int)args->len >= (int)args->minlen)
return 1;
args->agbno = NULLAGBLOCK;
return 0;
@@ -315,6 +317,9 @@ xfs_alloc_fixup_trees(
xfs_agblock_t nfbno2; /* second new free startblock */
xfs_extlen_t nflen1=0; /* first new free length */
xfs_extlen_t nflen2=0; /* second new free length */
+ struct xfs_mount *mp;
+
+ mp = cnt_cur->bc_mp;
/*
* Look up the record in the by-size tree if necessary.
@@ -323,13 +328,13 @@ xfs_alloc_fixup_trees(
#ifdef DEBUG
if ((error = xfs_alloc_get_rec(cnt_cur, &nfbno1, &nflen1, &i)))
return error;
- XFS_WANT_CORRUPTED_RETURN(
+ XFS_WANT_CORRUPTED_RETURN(mp,
i == 1 && nfbno1 == fbno && nflen1 == flen);
#endif
} else {
if ((error = xfs_alloc_lookup_eq(cnt_cur, fbno, flen, &i)))
return error;
- XFS_WANT_CORRUPTED_RETURN(i == 1);
+ XFS_WANT_CORRUPTED_RETURN(mp, i == 1);
}
/*
* Look up the record in the by-block tree if necessary.
@@ -338,13 +343,13 @@ xfs_alloc_fixup_trees(
#ifdef DEBUG
if ((error = xfs_alloc_get_rec(bno_cur, &nfbno1, &nflen1, &i)))
return error;
- XFS_WANT_CORRUPTED_RETURN(
+ XFS_WANT_CORRUPTED_RETURN(mp,
i == 1 && nfbno1 == fbno && nflen1 == flen);
#endif
} else {
if ((error = xfs_alloc_lookup_eq(bno_cur, fbno, flen, &i)))
return error;
- XFS_WANT_CORRUPTED_RETURN(i == 1);
+ XFS_WANT_CORRUPTED_RETURN(mp, i == 1);
}
#ifdef DEBUG
@@ -355,7 +360,7 @@ xfs_alloc_fixup_trees(
bnoblock = XFS_BUF_TO_BLOCK(bno_cur->bc_bufs[0]);
cntblock = XFS_BUF_TO_BLOCK(cnt_cur->bc_bufs[0]);
- XFS_WANT_CORRUPTED_RETURN(
+ XFS_WANT_CORRUPTED_RETURN(mp,
bnoblock->bb_numrecs == cntblock->bb_numrecs);
}
#endif
@@ -386,25 +391,25 @@ xfs_alloc_fixup_trees(
*/
if ((error = xfs_btree_delete(cnt_cur, &i)))
return error;
- XFS_WANT_CORRUPTED_RETURN(i == 1);
+ XFS_WANT_CORRUPTED_RETURN(mp, i == 1);
/*
* Add new by-size btree entry(s).
*/
if (nfbno1 != NULLAGBLOCK) {
if ((error = xfs_alloc_lookup_eq(cnt_cur, nfbno1, nflen1, &i)))
return error;
- XFS_WANT_CORRUPTED_RETURN(i == 0);
+ XFS_WANT_CORRUPTED_RETURN(mp, i == 0);
if ((error = xfs_btree_insert(cnt_cur, &i)))
return error;
- XFS_WANT_CORRUPTED_RETURN(i == 1);
+ XFS_WANT_CORRUPTED_RETURN(mp, i == 1);
}
if (nfbno2 != NULLAGBLOCK) {
if ((error = xfs_alloc_lookup_eq(cnt_cur, nfbno2, nflen2, &i)))
return error;
- XFS_WANT_CORRUPTED_RETURN(i == 0);
+ XFS_WANT_CORRUPTED_RETURN(mp, i == 0);
if ((error = xfs_btree_insert(cnt_cur, &i)))
return error;
- XFS_WANT_CORRUPTED_RETURN(i == 1);
+ XFS_WANT_CORRUPTED_RETURN(mp, i == 1);
}
/*
* Fix up the by-block btree entry(s).
@@ -415,7 +420,7 @@ xfs_alloc_fixup_trees(
*/
if ((error = xfs_btree_delete(bno_cur, &i)))
return error;
- XFS_WANT_CORRUPTED_RETURN(i == 1);
+ XFS_WANT_CORRUPTED_RETURN(mp, i == 1);
} else {
/*
* Update the by-block entry to start later|be shorter.
@@ -429,10 +434,10 @@ xfs_alloc_fixup_trees(
*/
if ((error = xfs_alloc_lookup_eq(bno_cur, nfbno2, nflen2, &i)))
return error;
- XFS_WANT_CORRUPTED_RETURN(i == 0);
+ XFS_WANT_CORRUPTED_RETURN(mp, i == 0);
if ((error = xfs_btree_insert(bno_cur, &i)))
return error;
- XFS_WANT_CORRUPTED_RETURN(i == 1);
+ XFS_WANT_CORRUPTED_RETURN(mp, i == 1);
}
return 0;
}
@@ -682,7 +687,7 @@ xfs_alloc_ag_vextent_exact(
error = xfs_alloc_get_rec(bno_cur, &fbno, &flen, &i);
if (error)
goto error0;
- XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
+ XFS_WANT_CORRUPTED_GOTO(args->mp, i == 1, error0);
ASSERT(fbno <= args->agbno);
/*
@@ -783,7 +788,7 @@ xfs_alloc_find_best_extent(
error = xfs_alloc_get_rec(*scur, sbno, slen, &i);
if (error)
goto error0;
- XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
+ XFS_WANT_CORRUPTED_GOTO(args->mp, i == 1, error0);
xfs_alloc_compute_aligned(args, *sbno, *slen, sbnoa, slena);
/*
@@ -946,7 +951,7 @@ restart:
if ((error = xfs_alloc_get_rec(cnt_cur, &ltbno,
&ltlen, &i)))
goto error0;
- XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
+ XFS_WANT_CORRUPTED_GOTO(args->mp, i == 1, error0);
if (ltlen >= args->minlen)
break;
if ((error = xfs_btree_increment(cnt_cur, 0, &i)))
@@ -966,7 +971,7 @@ restart:
*/
if ((error = xfs_alloc_get_rec(cnt_cur, &ltbno, &ltlen, &i)))
goto error0;
- XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
+ XFS_WANT_CORRUPTED_GOTO(args->mp, i == 1, error0);
xfs_alloc_compute_aligned(args, ltbno, ltlen,
&ltbnoa, &ltlena);
if (ltlena < args->minlen)
@@ -999,7 +1004,7 @@ restart:
cnt_cur->bc_ptrs[0] = besti;
if ((error = xfs_alloc_get_rec(cnt_cur, &ltbno, &ltlen, &i)))
goto error0;
- XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
+ XFS_WANT_CORRUPTED_GOTO(args->mp, i == 1, error0);
ASSERT(ltbno + ltlen <= be32_to_cpu(XFS_BUF_TO_AGF(args->agbp)->agf_length));
args->len = blen;
if (!xfs_alloc_fix_minleft(args)) {
@@ -1088,7 +1093,7 @@ restart:
if (bno_cur_lt) {
if ((error = xfs_alloc_get_rec(bno_cur_lt, &ltbno, &ltlen, &i)))
goto error0;
- XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
+ XFS_WANT_CORRUPTED_GOTO(args->mp, i == 1, error0);
xfs_alloc_compute_aligned(args, ltbno, ltlen,
&ltbnoa, &ltlena);
if (ltlena >= args->minlen)
@@ -1104,7 +1109,7 @@ restart:
if (bno_cur_gt) {
if ((error = xfs_alloc_get_rec(bno_cur_gt, &gtbno, &gtlen, &i)))
goto error0;
- XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
+ XFS_WANT_CORRUPTED_GOTO(args->mp, i == 1, error0);
xfs_alloc_compute_aligned(args, gtbno, gtlen,
&gtbnoa, &gtlena);
if (gtlena >= args->minlen)
@@ -1303,7 +1308,7 @@ restart:
error = xfs_alloc_get_rec(cnt_cur, &fbno, &flen, &i);
if (error)
goto error0;
- XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
+ XFS_WANT_CORRUPTED_GOTO(args->mp, i == 1, error0);
xfs_alloc_compute_aligned(args, fbno, flen,
&rbno, &rlen);
@@ -1342,7 +1347,7 @@ restart:
* This can't happen in the second case above.
*/
rlen = XFS_EXTLEN_MIN(args->maxlen, rlen);
- XFS_WANT_CORRUPTED_GOTO(rlen == 0 ||
+ XFS_WANT_CORRUPTED_GOTO(args->mp, rlen == 0 ||
(rlen <= flen && rbno + rlen <= fbno + flen), error0);
if (rlen < args->maxlen) {
xfs_agblock_t bestfbno;
@@ -1362,13 +1367,13 @@ restart:
if ((error = xfs_alloc_get_rec(cnt_cur, &fbno, &flen,
&i)))
goto error0;
- XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
+ XFS_WANT_CORRUPTED_GOTO(args->mp, i == 1, error0);
if (flen < bestrlen)
break;
xfs_alloc_compute_aligned(args, fbno, flen,
&rbno, &rlen);
rlen = XFS_EXTLEN_MIN(args->maxlen, rlen);
- XFS_WANT_CORRUPTED_GOTO(rlen == 0 ||
+ XFS_WANT_CORRUPTED_GOTO(args->mp, rlen == 0 ||
(rlen <= flen && rbno + rlen <= fbno + flen),
error0);
if (rlen > bestrlen) {
@@ -1383,7 +1388,7 @@ restart:
if ((error = xfs_alloc_lookup_eq(cnt_cur, bestfbno, bestflen,
&i)))
goto error0;
- XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
+ XFS_WANT_CORRUPTED_GOTO(args->mp, i == 1, error0);
rlen = bestrlen;
rbno = bestrbno;
flen = bestflen;
@@ -1408,7 +1413,7 @@ restart:
if (!xfs_alloc_fix_minleft(args))
goto out_nominleft;
rlen = args->len;
- XFS_WANT_CORRUPTED_GOTO(rlen <= flen, error0);
+ XFS_WANT_CORRUPTED_GOTO(args->mp, rlen <= flen, error0);
/*
* Allocate and initialize a cursor for the by-block tree.
*/
@@ -1422,7 +1427,7 @@ restart:
cnt_cur = bno_cur = NULL;
args->len = rlen;
args->agbno = rbno;
- XFS_WANT_CORRUPTED_GOTO(
+ XFS_WANT_CORRUPTED_GOTO(args->mp,
args->agbno + args->len <=
be32_to_cpu(XFS_BUF_TO_AGF(args->agbp)->agf_length),
error0);
@@ -1467,7 +1472,7 @@ xfs_alloc_ag_vextent_small(
if (i) {
if ((error = xfs_alloc_get_rec(ccur, &fbno, &flen, &i)))
goto error0;
- XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
+ XFS_WANT_CORRUPTED_GOTO(args->mp, i == 1, error0);
}
/*
* Nothing in the btree, try the freelist. Make sure
@@ -1493,7 +1498,7 @@ xfs_alloc_ag_vextent_small(
}
args->len = 1;
args->agbno = fbno;
- XFS_WANT_CORRUPTED_GOTO(
+ XFS_WANT_CORRUPTED_GOTO(args->mp,
args->agbno + args->len <=
be32_to_cpu(XFS_BUF_TO_AGF(args->agbp)->agf_length),
error0);
@@ -1579,7 +1584,7 @@ xfs_free_ag_extent(
*/
if ((error = xfs_alloc_get_rec(bno_cur, &ltbno, &ltlen, &i)))
goto error0;
- XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
+ XFS_WANT_CORRUPTED_GOTO(mp, i == 1, error0);
/*
* It's not contiguous, though.
*/
@@ -1591,7 +1596,8 @@ xfs_free_ag_extent(
* space was invalid, it's (partly) already free.
* Very bad.
*/
- XFS_WANT_CORRUPTED_GOTO(ltbno + ltlen <= bno, error0);
+ XFS_WANT_CORRUPTED_GOTO(mp,
+ ltbno + ltlen <= bno, error0);
}
}
/*
@@ -1606,7 +1612,7 @@ xfs_free_ag_extent(
*/
if ((error = xfs_alloc_get_rec(bno_cur, &gtbno, &gtlen, &i)))
goto error0;
- XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
+ XFS_WANT_CORRUPTED_GOTO(mp, i == 1, error0);
/*
* It's not contiguous, though.
*/
@@ -1618,7 +1624,7 @@ xfs_free_ag_extent(
* space was invalid, it's (partly) already free.
* Very bad.
*/
- XFS_WANT_CORRUPTED_GOTO(gtbno >= bno + len, error0);
+ XFS_WANT_CORRUPTED_GOTO(mp, gtbno >= bno + len, error0);
}
}
/*
@@ -1635,31 +1641,31 @@ xfs_free_ag_extent(
*/
if ((error = xfs_alloc_lookup_eq(cnt_cur, ltbno, ltlen, &i)))
goto error0;
- XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
+ XFS_WANT_CORRUPTED_GOTO(mp, i == 1, error0);
if ((error = xfs_btree_delete(cnt_cur, &i)))
goto error0;
- XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
+ XFS_WANT_CORRUPTED_GOTO(mp, i == 1, error0);
/*
* Delete the old by-size entry on the right.
*/
if ((error = xfs_alloc_lookup_eq(cnt_cur, gtbno, gtlen, &i)))
goto error0;
- XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
+ XFS_WANT_CORRUPTED_GOTO(mp, i == 1, error0);
if ((error = xfs_btree_delete(cnt_cur, &i)))
goto error0;
- XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
+ XFS_WANT_CORRUPTED_GOTO(mp, i == 1, error0);
/*
* Delete the old by-block entry for the right block.
*/
if ((error = xfs_btree_delete(bno_cur, &i)))
goto error0;
- XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
+ XFS_WANT_CORRUPTED_GOTO(mp, i == 1, error0);
/*
* Move the by-block cursor back to the left neighbor.
*/
if ((error = xfs_btree_decrement(bno_cur, 0, &i)))
goto error0;
- XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
+ XFS_WANT_CORRUPTED_GOTO(mp, i == 1, error0);
#ifdef DEBUG
/*
* Check that this is the right record: delete didn't
@@ -1672,7 +1678,7 @@ xfs_free_ag_extent(
if ((error = xfs_alloc_get_rec(bno_cur, &xxbno, &xxlen,
&i)))
goto error0;
- XFS_WANT_CORRUPTED_GOTO(
+ XFS_WANT_CORRUPTED_GOTO(mp,
i == 1 && xxbno == ltbno && xxlen == ltlen,
error0);
}
@@ -1695,17 +1701,17 @@ xfs_free_ag_extent(
*/
if ((error = xfs_alloc_lookup_eq(cnt_cur, ltbno, ltlen, &i)))
goto error0;
- XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
+ XFS_WANT_CORRUPTED_GOTO(mp, i == 1, error0);
if ((error = xfs_btree_delete(cnt_cur, &i)))
goto error0;
- XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
+ XFS_WANT_CORRUPTED_GOTO(mp, i == 1, error0);
/*
* Back up the by-block cursor to the left neighbor, and
* update its length.
*/
if ((error = xfs_btree_decrement(bno_cur, 0, &i)))
goto error0;
- XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
+ XFS_WANT_CORRUPTED_GOTO(mp, i == 1, error0);
nbno = ltbno;
nlen = len + ltlen;
if ((error = xfs_alloc_update(bno_cur, nbno, nlen)))
@@ -1721,10 +1727,10 @@ xfs_free_ag_extent(
*/
if ((error = xfs_alloc_lookup_eq(cnt_cur, gtbno, gtlen, &i)))
goto error0;
- XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
+ XFS_WANT_CORRUPTED_GOTO(mp, i == 1, error0);
if ((error = xfs_btree_delete(cnt_cur, &i)))
goto error0;
- XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
+ XFS_WANT_CORRUPTED_GOTO(mp, i == 1, error0);
/*
* Update the starting block and length of the right
* neighbor in the by-block tree.
@@ -1743,7 +1749,7 @@ xfs_free_ag_extent(
nlen = len;
if ((error = xfs_btree_insert(bno_cur, &i)))
goto error0;
- XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
+ XFS_WANT_CORRUPTED_GOTO(mp, i == 1, error0);
}
xfs_btree_del_cursor(bno_cur, XFS_BTREE_NOERROR);
bno_cur = NULL;
@@ -1752,10 +1758,10 @@ xfs_free_ag_extent(
*/
if ((error = xfs_alloc_lookup_eq(cnt_cur, nbno, nlen, &i)))
goto error0;
- XFS_WANT_CORRUPTED_GOTO(i == 0, error0);
+ XFS_WANT_CORRUPTED_GOTO(mp, i == 0, error0);
if ((error = xfs_btree_insert(cnt_cur, &i)))
goto error0;
- XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
+ XFS_WANT_CORRUPTED_GOTO(mp, i == 1, error0);
xfs_btree_del_cursor(cnt_cur, XFS_BTREE_NOERROR);
cnt_cur = NULL;
diff --git a/fs/xfs/libxfs/xfs_bmap.c b/fs/xfs/libxfs/xfs_bmap.c
index 990595548958..b872e9cedb69 100644
--- a/fs/xfs/libxfs/xfs_bmap.c
+++ b/fs/xfs/libxfs/xfs_bmap.c
@@ -397,7 +397,8 @@ xfs_bmap_check_leaf_extents(
xfs_check_block(block, mp, 0, 0);
pp = XFS_BMBT_PTR_ADDR(mp, block, 1, mp->m_bmap_dmxr[1]);
bno = be64_to_cpu(*pp);
- XFS_WANT_CORRUPTED_GOTO(XFS_FSB_SANITY_CHECK(mp, bno), error0);
+ XFS_WANT_CORRUPTED_GOTO(mp,
+ XFS_FSB_SANITY_CHECK(mp, bno), error0);
if (bp_release) {
bp_release = 0;
xfs_trans_brelse(NULL, bp);
@@ -1002,7 +1003,7 @@ xfs_bmap_add_attrfork_btree(
if ((error = xfs_bmbt_lookup_ge(cur, 0, 0, 0, &stat)))
goto error0;
/* must be at least one entry */
- XFS_WANT_CORRUPTED_GOTO(stat == 1, error0);
+ XFS_WANT_CORRUPTED_GOTO(mp, stat == 1, error0);
if ((error = xfs_btree_new_iroot(cur, flags, &stat)))
goto error0;
if (stat == 0) {
@@ -1288,7 +1289,8 @@ xfs_bmap_read_extents(
break;
pp = XFS_BMBT_PTR_ADDR(mp, block, 1, mp->m_bmap_dmxr[1]);
bno = be64_to_cpu(*pp);
- XFS_WANT_CORRUPTED_GOTO(XFS_FSB_SANITY_CHECK(mp, bno), error0);
+ XFS_WANT_CORRUPTED_GOTO(mp,
+ XFS_FSB_SANITY_CHECK(mp, bno), error0);
xfs_trans_brelse(tp, bp);
}
/*
@@ -1722,7 +1724,9 @@ xfs_bmap_add_extent_delay_real(
xfs_filblks_t temp=0; /* value for da_new calculations */
xfs_filblks_t temp2=0;/* value for da_new calculations */
int tmp_rval; /* partial logging flags */
+ struct xfs_mount *mp;
+ mp = bma->tp ? bma->tp->t_mountp : NULL;
ifp = XFS_IFORK_PTR(bma->ip, XFS_DATA_FORK);
ASSERT(bma->idx >= 0);
@@ -1833,15 +1837,15 @@ xfs_bmap_add_extent_delay_real(
RIGHT.br_blockcount, &i);
if (error)
goto done;
- XFS_WANT_CORRUPTED_GOTO(i == 1, done);
+ XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done);
error = xfs_btree_delete(bma->cur, &i);
if (error)
goto done;
- XFS_WANT_CORRUPTED_GOTO(i == 1, done);
+ XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done);
error = xfs_btree_decrement(bma->cur, 0, &i);
if (error)
goto done;
- XFS_WANT_CORRUPTED_GOTO(i == 1, done);
+ XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done);
error = xfs_bmbt_update(bma->cur, LEFT.br_startoff,
LEFT.br_startblock,
LEFT.br_blockcount +
@@ -1874,7 +1878,7 @@ xfs_bmap_add_extent_delay_real(
&i);
if (error)
goto done;
- XFS_WANT_CORRUPTED_GOTO(i == 1, done);
+ XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done);
error = xfs_bmbt_update(bma->cur, LEFT.br_startoff,
LEFT.br_startblock,
LEFT.br_blockcount +
@@ -1905,7 +1909,7 @@ xfs_bmap_add_extent_delay_real(
RIGHT.br_blockcount, &i);
if (error)
goto done;
- XFS_WANT_CORRUPTED_GOTO(i == 1, done);
+ XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done);
error = xfs_bmbt_update(bma->cur, PREV.br_startoff,
new->br_startblock,
PREV.br_blockcount +
@@ -1935,12 +1939,12 @@ xfs_bmap_add_extent_delay_real(
&i);
if (error)
goto done;
- XFS_WANT_CORRUPTED_GOTO(i == 0, done);
+ XFS_WANT_CORRUPTED_GOTO(mp, i == 0, done);
bma->cur->bc_rec.b.br_state = XFS_EXT_NORM;
error = xfs_btree_insert(bma->cur, &i);
if (error)
goto done;
- XFS_WANT_CORRUPTED_GOTO(i == 1, done);
+ XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done);
}
break;
@@ -1968,7 +1972,7 @@ xfs_bmap_add_extent_delay_real(
&i);
if (error)
goto done;
- XFS_WANT_CORRUPTED_GOTO(i == 1, done);
+ XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done);
error = xfs_bmbt_update(bma->cur, LEFT.br_startoff,
LEFT.br_startblock,
LEFT.br_blockcount +
@@ -2005,12 +2009,12 @@ xfs_bmap_add_extent_delay_real(
&i);
if (error)
goto done;
- XFS_WANT_CORRUPTED_GOTO(i == 0, done);
+ XFS_WANT_CORRUPTED_GOTO(mp, i == 0, done);
bma->cur->bc_rec.b.br_state = XFS_EXT_NORM;
error = xfs_btree_insert(bma->cur, &i);
if (error)
goto done;
- XFS_WANT_CORRUPTED_GOTO(i == 1, done);
+ XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done);
}
if (xfs_bmap_needs_btree(bma->ip, XFS_DATA_FORK)) {
@@ -2051,7 +2055,7 @@ xfs_bmap_add_extent_delay_real(
RIGHT.br_blockcount, &i);
if (error)
goto done;
- XFS_WANT_CORRUPTED_GOTO(i == 1, done);
+ XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done);
error = xfs_bmbt_update(bma->cur, new->br_startoff,
new->br_startblock,
new->br_blockcount +
@@ -2089,12 +2093,12 @@ xfs_bmap_add_extent_delay_real(
&i);
if (error)
goto done;
- XFS_WANT_CORRUPTED_GOTO(i == 0, done);
+ XFS_WANT_CORRUPTED_GOTO(mp, i == 0, done);
bma->cur->bc_rec.b.br_state = XFS_EXT_NORM;
error = xfs_btree_insert(bma->cur, &i);
if (error)
goto done;
- XFS_WANT_CORRUPTED_GOTO(i == 1, done);
+ XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done);
}
if (xfs_bmap_needs_btree(bma->ip, XFS_DATA_FORK)) {
@@ -2158,12 +2162,12 @@ xfs_bmap_add_extent_delay_real(
&i);
if (error)
goto done;
- XFS_WANT_CORRUPTED_GOTO(i == 0, done);
+ XFS_WANT_CORRUPTED_GOTO(mp, i == 0, done);
bma->cur->bc_rec.b.br_state = XFS_EXT_NORM;
error = xfs_btree_insert(bma->cur, &i);
if (error)
goto done;
- XFS_WANT_CORRUPTED_GOTO(i == 1, done);
+ XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done);
}
if (xfs_bmap_needs_btree(bma->ip, XFS_DATA_FORK)) {
@@ -2274,6 +2278,7 @@ xfs_bmap_add_extent_unwritten_real(
/* left is 0, right is 1, prev is 2 */
int rval=0; /* return value (logging flags) */
int state = 0;/* state bits, accessed thru macros */
+ struct xfs_mount *mp = tp->t_mountp;
*logflagsp = 0;
@@ -2386,19 +2391,19 @@ xfs_bmap_add_extent_unwritten_real(
RIGHT.br_startblock,
RIGHT.br_blockcount, &i)))
goto done;
- XFS_WANT_CORRUPTED_GOTO(i == 1, done);
+ XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done);
if ((error = xfs_btree_delete(cur, &i)))
goto done;
- XFS_WANT_CORRUPTED_GOTO(i == 1, done);
+ XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done);
if ((error = xfs_btree_decrement(cur, 0, &i)))
goto done;
- XFS_WANT_CORRUPTED_GOTO(i == 1, done);
+ XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done);
if ((error = xfs_btree_delete(cur, &i)))
goto done;
- XFS_WANT_CORRUPTED_GOTO(i == 1, done);
+ XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done);
if ((error = xfs_btree_decrement(cur, 0, &i)))
goto done;
- XFS_WANT_CORRUPTED_GOTO(i == 1, done);
+ XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done);
if ((error = xfs_bmbt_update(cur, LEFT.br_startoff,
LEFT.br_startblock,
LEFT.br_blockcount + PREV.br_blockcount +
@@ -2429,13 +2434,13 @@ xfs_bmap_add_extent_unwritten_real(
PREV.br_startblock, PREV.br_blockcount,
&i)))
goto done;
- XFS_WANT_CORRUPTED_GOTO(i == 1, done);
+ XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done);
if ((error = xfs_btree_delete(cur, &i)))
goto done;
- XFS_WANT_CORRUPTED_GOTO(i == 1, done);
+ XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done);
if ((error = xfs_btree_decrement(cur, 0, &i)))
goto done;
- XFS_WANT_CORRUPTED_GOTO(i == 1, done);
+ XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done);
if ((error = xfs_bmbt_update(cur, LEFT.br_startoff,
LEFT.br_startblock,
LEFT.br_blockcount + PREV.br_blockcount,
@@ -2464,13 +2469,13 @@ xfs_bmap_add_extent_unwritten_real(
RIGHT.br_startblock,
RIGHT.br_blockcount, &i)))
goto done;
- XFS_WANT_CORRUPTED_GOTO(i == 1, done);
+ XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done);
if ((error = xfs_btree_delete(cur, &i)))
goto done;
- XFS_WANT_CORRUPTED_GOTO(i == 1, done);
+ XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done);
if ((error = xfs_btree_decrement(cur, 0, &i)))
goto done;
- XFS_WANT_CORRUPTED_GOTO(i == 1, done);
+ XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done);
if ((error = xfs_bmbt_update(cur, new->br_startoff,
new->br_startblock,
new->br_blockcount + RIGHT.br_blockcount,
@@ -2497,7 +2502,7 @@ xfs_bmap_add_extent_unwritten_real(
new->br_startblock, new->br_blockcount,
&i)))
goto done;
- XFS_WANT_CORRUPTED_GOTO(i == 1, done);
+ XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done);
if ((error = xfs_bmbt_update(cur, new->br_startoff,
new->br_startblock, new->br_blockcount,
newext)))
@@ -2534,7 +2539,7 @@ xfs_bmap_add_extent_unwritten_real(
PREV.br_startblock, PREV.br_blockcount,
&i)))
goto done;
- XFS_WANT_CORRUPTED_GOTO(i == 1, done);
+ XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done);
if ((error = xfs_bmbt_update(cur,
PREV.br_startoff + new->br_blockcount,
PREV.br_startblock + new->br_blockcount,
@@ -2576,7 +2581,7 @@ xfs_bmap_add_extent_unwritten_real(
PREV.br_startblock, PREV.br_blockcount,
&i)))
goto done;
- XFS_WANT_CORRUPTED_GOTO(i == 1, done);
+ XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done);
if ((error = xfs_bmbt_update(cur,
PREV.br_startoff + new->br_blockcount,
PREV.br_startblock + new->br_blockcount,
@@ -2586,7 +2591,7 @@ xfs_bmap_add_extent_unwritten_real(
cur->bc_rec.b = *new;
if ((error = xfs_btree_insert(cur, &i)))
goto done;
- XFS_WANT_CORRUPTED_GOTO(i == 1, done);
+ XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done);
}
break;
@@ -2616,7 +2621,7 @@ xfs_bmap_add_extent_unwritten_real(
PREV.br_startblock,
PREV.br_blockcount, &i)))
goto done;
- XFS_WANT_CORRUPTED_GOTO(i == 1, done);
+ XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done);
if ((error = xfs_bmbt_update(cur, PREV.br_startoff,
PREV.br_startblock,
PREV.br_blockcount - new->br_blockcount,
@@ -2654,7 +2659,7 @@ xfs_bmap_add_extent_unwritten_real(
PREV.br_startblock, PREV.br_blockcount,
&i)))
goto done;
- XFS_WANT_CORRUPTED_GOTO(i == 1, done);
+ XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done);
if ((error = xfs_bmbt_update(cur, PREV.br_startoff,
PREV.br_startblock,
PREV.br_blockcount - new->br_blockcount,
@@ -2664,11 +2669,11 @@ xfs_bmap_add_extent_unwritten_real(
new->br_startblock, new->br_blockcount,
&i)))
goto done;
- XFS_WANT_CORRUPTED_GOTO(i == 0, done);
+ XFS_WANT_CORRUPTED_GOTO(mp, i == 0, done);
cur->bc_rec.b.br_state = XFS_EXT_NORM;
if ((error = xfs_btree_insert(cur, &i)))
goto done;
- XFS_WANT_CORRUPTED_GOTO(i == 1, done);
+ XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done);
}
break;
@@ -2702,7 +2707,7 @@ xfs_bmap_add_extent_unwritten_real(
PREV.br_startblock, PREV.br_blockcount,
&i)))
goto done;
- XFS_WANT_CORRUPTED_GOTO(i == 1, done);
+ XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done);
/* new right extent - oldext */
if ((error = xfs_bmbt_update(cur, r[1].br_startoff,
r[1].br_startblock, r[1].br_blockcount,
@@ -2714,7 +2719,7 @@ xfs_bmap_add_extent_unwritten_real(
new->br_startoff - PREV.br_startoff;
if ((error = xfs_btree_insert(cur, &i)))
goto done;
- XFS_WANT_CORRUPTED_GOTO(i == 1, done);
+ XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done);
/*
* Reset the cursor to the position of the new extent
* we are about to insert as we can't trust it after
@@ -2724,12 +2729,12 @@ xfs_bmap_add_extent_unwritten_real(
new->br_startblock, new->br_blockcount,
&i)))
goto done;
- XFS_WANT_CORRUPTED_GOTO(i == 0, done);
+ XFS_WANT_CORRUPTED_GOTO(mp, i == 0, done);
/* new middle extent - newext */
cur->bc_rec.b.br_state = new->br_state;
if ((error = xfs_btree_insert(cur, &i)))
goto done;
- XFS_WANT_CORRUPTED_GOTO(i == 1, done);
+ XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done);
}
break;
@@ -2933,7 +2938,9 @@ xfs_bmap_add_extent_hole_real(
xfs_bmbt_irec_t right; /* right neighbor extent entry */
int rval=0; /* return value (logging flags) */
int state; /* state bits, accessed thru macros */
+ struct xfs_mount *mp;
+ mp = bma->tp ? bma->tp->t_mountp : NULL;
ifp = XFS_IFORK_PTR(bma->ip, whichfork);
ASSERT(bma->idx >= 0);
@@ -3021,15 +3028,15 @@ xfs_bmap_add_extent_hole_real(
&i);
if (error)
goto done;
- XFS_WANT_CORRUPTED_GOTO(i == 1, done);
+ XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done);
error = xfs_btree_delete(bma->cur, &i);
if (error)
goto done;
- XFS_WANT_CORRUPTED_GOTO(i == 1, done);
+ XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done);
error = xfs_btree_decrement(bma->cur, 0, &i);
if (error)
goto done;
- XFS_WANT_CORRUPTED_GOTO(i == 1, done);
+ XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done);
error = xfs_bmbt_update(bma->cur, left.br_startoff,
left.br_startblock,
left.br_blockcount +
@@ -3062,7 +3069,7 @@ xfs_bmap_add_extent_hole_real(
&i);
if (error)
goto done;
- XFS_WANT_CORRUPTED_GOTO(i == 1, done);
+ XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done);
error = xfs_bmbt_update(bma->cur, left.br_startoff,
left.br_startblock,
left.br_blockcount +
@@ -3096,7 +3103,7 @@ xfs_bmap_add_extent_hole_real(
right.br_blockcount, &i);
if (error)
goto done;
- XFS_WANT_CORRUPTED_GOTO(i == 1, done);
+ XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done);
error = xfs_bmbt_update(bma->cur, new->br_startoff,
new->br_startblock,
new->br_blockcount +
@@ -3126,12 +3133,12 @@ xfs_bmap_add_extent_hole_real(
new->br_blockcount, &i);
if (error)
goto done;
- XFS_WANT_CORRUPTED_GOTO(i == 0, done);
+ XFS_WANT_CORRUPTED_GOTO(mp, i == 0, done);
bma->cur->bc_rec.b.br_state = new->br_state;
error = xfs_btree_insert(bma->cur, &i);
if (error)
goto done;
- XFS_WANT_CORRUPTED_GOTO(i == 1, done);
+ XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done);
}
break;
}
@@ -4763,7 +4770,7 @@ xfs_bmap_del_extent(
got.br_startblock, got.br_blockcount,
&i)))
goto done;
- XFS_WANT_CORRUPTED_GOTO(i == 1, done);
+ XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done);
}
da_old = da_new = 0;
} else {
@@ -4797,7 +4804,7 @@ xfs_bmap_del_extent(
}
if ((error = xfs_btree_delete(cur, &i)))
goto done;
- XFS_WANT_CORRUPTED_GOTO(i == 1, done);
+ XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done);
break;
case 2:
@@ -4897,7 +4904,8 @@ xfs_bmap_del_extent(
got.br_startblock,
temp, &i)))
goto done;
- XFS_WANT_CORRUPTED_GOTO(i == 1, done);
+ XFS_WANT_CORRUPTED_GOTO(mp,
+ i == 1, done);
/*
* Update the btree record back
* to the original value.
@@ -4918,7 +4926,7 @@ xfs_bmap_del_extent(
error = -ENOSPC;
goto done;
}
- XFS_WANT_CORRUPTED_GOTO(i == 1, done);
+ XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done);
} else
flags |= xfs_ilog_fext(whichfork);
XFS_IFORK_NEXT_SET(ip, whichfork,
@@ -5412,6 +5420,7 @@ xfs_bmse_merge(
struct xfs_bmbt_irec left;
xfs_filblks_t blockcount;
int error, i;
+ struct xfs_mount *mp = ip->i_mount;
xfs_bmbt_get_all(gotp, &got);
xfs_bmbt_get_all(leftp, &left);
@@ -5446,19 +5455,19 @@ xfs_bmse_merge(
got.br_blockcount, &i);
if (error)
return error;
- XFS_WANT_CORRUPTED_RETURN(i == 1);
+ XFS_WANT_CORRUPTED_RETURN(mp, i == 1);
error = xfs_btree_delete(cur, &i);
if (error)
return error;
- XFS_WANT_CORRUPTED_RETURN(i == 1);
+ XFS_WANT_CORRUPTED_RETURN(mp, i == 1);
/* lookup and update size of the previous extent */
error = xfs_bmbt_lookup_eq(cur, left.br_startoff, left.br_startblock,
left.br_blockcount, &i);
if (error)
return error;
- XFS_WANT_CORRUPTED_RETURN(i == 1);
+ XFS_WANT_CORRUPTED_RETURN(mp, i == 1);
left.br_blockcount = blockcount;
@@ -5480,6 +5489,7 @@ xfs_bmse_shift_one(
int *logflags)
{
struct xfs_ifork *ifp;
+ struct xfs_mount *mp;
xfs_fileoff_t startoff;
struct xfs_bmbt_rec_host *leftp;
struct xfs_bmbt_irec got;
@@ -5487,13 +5497,14 @@ xfs_bmse_shift_one(
int error;
int i;
+ mp = ip->i_mount;
ifp = XFS_IFORK_PTR(ip, whichfork);
xfs_bmbt_get_all(gotp, &got);
startoff = got.br_startoff - offset_shift_fsb;
/* delalloc extents should be prevented by caller */
- XFS_WANT_CORRUPTED_RETURN(!isnullstartblock(got.br_startblock));
+ XFS_WANT_CORRUPTED_RETURN(mp, !isnullstartblock(got.br_startblock));
/*
* Check for merge if we've got an extent to the left, otherwise make
@@ -5532,7 +5543,7 @@ xfs_bmse_shift_one(
got.br_blockcount, &i);
if (error)
return error;
- XFS_WANT_CORRUPTED_RETURN(i == 1);
+ XFS_WANT_CORRUPTED_RETURN(mp, i == 1);
got.br_startoff = startoff;
return xfs_bmbt_update(cur, got.br_startoff, got.br_startblock,
diff --git a/fs/xfs/libxfs/xfs_btree.c b/fs/xfs/libxfs/xfs_btree.c
index 81cad433df85..c72283dd8d44 100644
--- a/fs/xfs/libxfs/xfs_btree.c
+++ b/fs/xfs/libxfs/xfs_btree.c
@@ -168,7 +168,7 @@ xfs_btree_check_lptr(
xfs_fsblock_t bno, /* btree block disk address */
int level) /* btree block level */
{
- XFS_WANT_CORRUPTED_RETURN(
+ XFS_WANT_CORRUPTED_RETURN(cur->bc_mp,
level > 0 &&
bno != NULLFSBLOCK &&
XFS_FSB_SANITY_CHECK(cur->bc_mp, bno));
@@ -187,7 +187,7 @@ xfs_btree_check_sptr(
{
xfs_agblock_t agblocks = cur->bc_mp->m_sb.sb_agblocks;
- XFS_WANT_CORRUPTED_RETURN(
+ XFS_WANT_CORRUPTED_RETURN(cur->bc_mp,
level > 0 &&
bno != NULLAGBLOCK &&
bno != 0 &&
@@ -1825,7 +1825,7 @@ xfs_btree_lookup(
error = xfs_btree_increment(cur, 0, &i);
if (error)
goto error0;
- XFS_WANT_CORRUPTED_RETURN(i == 1);
+ XFS_WANT_CORRUPTED_RETURN(cur->bc_mp, i == 1);
XFS_BTREE_TRACE_CURSOR(cur, XBT_EXIT);
*stat = 1;
return 0;
@@ -2285,7 +2285,7 @@ xfs_btree_rshift(
if (error)
goto error0;
i = xfs_btree_lastrec(tcur, level);
- XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
+ XFS_WANT_CORRUPTED_GOTO(cur->bc_mp, i == 1, error0);
error = xfs_btree_increment(tcur, level, &i);
if (error)
@@ -3138,7 +3138,7 @@ xfs_btree_insert(
goto error0;
}
- XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
+ XFS_WANT_CORRUPTED_GOTO(cur->bc_mp, i == 1, error0);
level++;
/*
@@ -3582,15 +3582,15 @@ xfs_btree_delrec(
* Actually any entry but the first would suffice.
*/
i = xfs_btree_lastrec(tcur, level);
- XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
+ XFS_WANT_CORRUPTED_GOTO(cur->bc_mp, i == 1, error0);
error = xfs_btree_increment(tcur, level, &i);
if (error)
goto error0;
- XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
+ XFS_WANT_CORRUPTED_GOTO(cur->bc_mp, i == 1, error0);
i = xfs_btree_lastrec(tcur, level);
- XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
+ XFS_WANT_CORRUPTED_GOTO(cur->bc_mp, i == 1, error0);
/* Grab a pointer to the block. */
right = xfs_btree_get_block(tcur, level, &rbp);
@@ -3634,12 +3634,12 @@ xfs_btree_delrec(
rrecs = xfs_btree_get_numrecs(right);
if (!xfs_btree_ptr_is_null(cur, &lptr)) {
i = xfs_btree_firstrec(tcur, level);
- XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
+ XFS_WANT_CORRUPTED_GOTO(cur->bc_mp, i == 1, error0);
error = xfs_btree_decrement(tcur, level, &i);
if (error)
goto error0;
- XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
+ XFS_WANT_CORRUPTED_GOTO(cur->bc_mp, i == 1, error0);
}
}
@@ -3653,13 +3653,13 @@ xfs_btree_delrec(
* previous block.
*/
i = xfs_btree_firstrec(tcur, level);
- XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
+ XFS_WANT_CORRUPTED_GOTO(cur->bc_mp, i == 1, error0);
error = xfs_btree_decrement(tcur, level, &i);
if (error)
goto error0;
i = xfs_btree_firstrec(tcur, level);
- XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
+ XFS_WANT_CORRUPTED_GOTO(cur->bc_mp, i == 1, error0);
/* Grab a pointer to the block. */
left = xfs_btree_get_block(tcur, level, &lbp);
diff --git a/fs/xfs/libxfs/xfs_dir2_data.c b/fs/xfs/libxfs/xfs_dir2_data.c
index 5ff31be9b1cd..de1ea16f5748 100644
--- a/fs/xfs/libxfs/xfs_dir2_data.c
+++ b/fs/xfs/libxfs/xfs_dir2_data.c
@@ -89,7 +89,7 @@ __xfs_dir3_data_check(
* so just ensure that the count falls somewhere inside the
* block right now.
*/
- XFS_WANT_CORRUPTED_RETURN(be32_to_cpu(btp->count) <
+ XFS_WANT_CORRUPTED_RETURN(mp, be32_to_cpu(btp->count) <
((char *)btp - p) / sizeof(struct xfs_dir2_leaf_entry));
break;
case cpu_to_be32(XFS_DIR3_DATA_MAGIC):
@@ -107,21 +107,21 @@ __xfs_dir3_data_check(
bf = ops->data_bestfree_p(hdr);
count = lastfree = freeseen = 0;
if (!bf[0].length) {
- XFS_WANT_CORRUPTED_RETURN(!bf[0].offset);
+ XFS_WANT_CORRUPTED_RETURN(mp, !bf[0].offset);
freeseen |= 1 << 0;
}
if (!bf[1].length) {
- XFS_WANT_CORRUPTED_RETURN(!bf[1].offset);
+ XFS_WANT_CORRUPTED_RETURN(mp, !bf[1].offset);
freeseen |= 1 << 1;
}
if (!bf[2].length) {
- XFS_WANT_CORRUPTED_RETURN(!bf[2].offset);
+ XFS_WANT_CORRUPTED_RETURN(mp, !bf[2].offset);
freeseen |= 1 << 2;
}
- XFS_WANT_CORRUPTED_RETURN(be16_to_cpu(bf[0].length) >=
+ XFS_WANT_CORRUPTED_RETURN(mp, be16_to_cpu(bf[0].length) >=
be16_to_cpu(bf[1].length));
- XFS_WANT_CORRUPTED_RETURN(be16_to_cpu(bf[1].length) >=
+ XFS_WANT_CORRUPTED_RETURN(mp, be16_to_cpu(bf[1].length) >=
be16_to_cpu(bf[2].length));
/*
* Loop over the data/unused entries.
@@ -134,18 +134,18 @@ __xfs_dir3_data_check(
* doesn't need to be there.
*/
if (be16_to_cpu(dup->freetag) == XFS_DIR2_DATA_FREE_TAG) {
- XFS_WANT_CORRUPTED_RETURN(lastfree == 0);
- XFS_WANT_CORRUPTED_RETURN(
+ XFS_WANT_CORRUPTED_RETURN(mp, lastfree == 0);
+ XFS_WANT_CORRUPTED_RETURN(mp,
be16_to_cpu(*xfs_dir2_data_unused_tag_p(dup)) ==
(char *)dup - (char *)hdr);
dfp = xfs_dir2_data_freefind(hdr, bf, dup);
if (dfp) {
i = (int)(dfp - bf);
- XFS_WANT_CORRUPTED_RETURN(
+ XFS_WANT_CORRUPTED_RETURN(mp,
(freeseen & (1 << i)) == 0);
freeseen |= 1 << i;
} else {
- XFS_WANT_CORRUPTED_RETURN(
+ XFS_WANT_CORRUPTED_RETURN(mp,
be16_to_cpu(dup->length) <=
be16_to_cpu(bf[2].length));
}
@@ -160,13 +160,13 @@ __xfs_dir3_data_check(
* The linear search is crude but this is DEBUG code.
*/
dep = (xfs_dir2_data_entry_t *)p;
- XFS_WANT_CORRUPTED_RETURN(dep->namelen != 0);
- XFS_WANT_CORRUPTED_RETURN(
+ XFS_WANT_CORRUPTED_RETURN(mp, dep->namelen != 0);
+ XFS_WANT_CORRUPTED_RETURN(mp,
!xfs_dir_ino_validate(mp, be64_to_cpu(dep->inumber)));
- XFS_WANT_CORRUPTED_RETURN(
+ XFS_WANT_CORRUPTED_RETURN(mp,
be16_to_cpu(*ops->data_entry_tag_p(dep)) ==
(char *)dep - (char *)hdr);
- XFS_WANT_CORRUPTED_RETURN(
+ XFS_WANT_CORRUPTED_RETURN(mp,
ops->data_get_ftype(dep) < XFS_DIR3_FT_MAX);
count++;
lastfree = 0;
@@ -183,14 +183,15 @@ __xfs_dir3_data_check(
be32_to_cpu(lep[i].hashval) == hash)
break;
}
- XFS_WANT_CORRUPTED_RETURN(i < be32_to_cpu(btp->count));
+ XFS_WANT_CORRUPTED_RETURN(mp,
+ i < be32_to_cpu(btp->count));
}
p += ops->data_entsize(dep->namelen);
}
/*
* Need to have seen all the entries and all the bestfree slots.
*/
- XFS_WANT_CORRUPTED_RETURN(freeseen == 7);
+ XFS_WANT_CORRUPTED_RETURN(mp, freeseen == 7);
if (hdr->magic == cpu_to_be32(XFS_DIR2_BLOCK_MAGIC) ||
hdr->magic == cpu_to_be32(XFS_DIR3_BLOCK_MAGIC)) {
for (i = stale = 0; i < be32_to_cpu(btp->count); i++) {
@@ -198,13 +199,13 @@ __xfs_dir3_data_check(
cpu_to_be32(XFS_DIR2_NULL_DATAPTR))
stale++;
if (i > 0)
- XFS_WANT_CORRUPTED_RETURN(
+ XFS_WANT_CORRUPTED_RETURN(mp,
be32_to_cpu(lep[i].hashval) >=
be32_to_cpu(lep[i - 1].hashval));
}
- XFS_WANT_CORRUPTED_RETURN(count ==
+ XFS_WANT_CORRUPTED_RETURN(mp, count ==
be32_to_cpu(btp->count) - be32_to_cpu(btp->stale));
- XFS_WANT_CORRUPTED_RETURN(stale == be32_to_cpu(btp->stale));
+ XFS_WANT_CORRUPTED_RETURN(mp, stale == be32_to_cpu(btp->stale));
}
return 0;
}
diff --git a/fs/xfs/libxfs/xfs_ialloc.c b/fs/xfs/libxfs/xfs_ialloc.c
index 5b4ba9f6b37d..07349a183a11 100644
--- a/fs/xfs/libxfs/xfs_ialloc.c
+++ b/fs/xfs/libxfs/xfs_ialloc.c
@@ -701,7 +701,7 @@ xfs_ialloc_next_rec(
error = xfs_inobt_get_rec(cur, rec, &i);
if (error)
return error;
- XFS_WANT_CORRUPTED_RETURN(i == 1);
+ XFS_WANT_CORRUPTED_RETURN(cur->bc_mp, i == 1);
}
return 0;
@@ -725,7 +725,7 @@ xfs_ialloc_get_rec(
error = xfs_inobt_get_rec(cur, rec, &i);
if (error)
return error;
- XFS_WANT_CORRUPTED_RETURN(i == 1);
+ XFS_WANT_CORRUPTED_RETURN(cur->bc_mp, i == 1);
}
return 0;
@@ -784,12 +784,12 @@ xfs_dialloc_ag_inobt(
error = xfs_inobt_lookup(cur, pagino, XFS_LOOKUP_LE, &i);
if (error)
goto error0;
- XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
+ XFS_WANT_CORRUPTED_GOTO(mp, i == 1, error0);
error = xfs_inobt_get_rec(cur, &rec, &j);
if (error)
goto error0;
- XFS_WANT_CORRUPTED_GOTO(j == 1, error0);
+ XFS_WANT_CORRUPTED_GOTO(mp, j == 1, error0);
if (rec.ir_freecount > 0) {
/*
@@ -945,19 +945,19 @@ newino:
error = xfs_inobt_lookup(cur, 0, XFS_LOOKUP_GE, &i);
if (error)
goto error0;
- XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
+ XFS_WANT_CORRUPTED_GOTO(mp, i == 1, error0);
for (;;) {
error = xfs_inobt_get_rec(cur, &rec, &i);
if (error)
goto error0;
- XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
+ XFS_WANT_CORRUPTED_GOTO(mp, i == 1, error0);
if (rec.ir_freecount > 0)
break;
error = xfs_btree_increment(cur, 0, &i);
if (error)
goto error0;
- XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
+ XFS_WANT_CORRUPTED_GOTO(mp, i == 1, error0);
}
alloc_inode:
@@ -1017,7 +1017,7 @@ xfs_dialloc_ag_finobt_near(
error = xfs_inobt_get_rec(lcur, rec, &i);
if (error)
return error;
- XFS_WANT_CORRUPTED_RETURN(i == 1);
+ XFS_WANT_CORRUPTED_RETURN(lcur->bc_mp, i == 1);
/*
* See if we've landed in the parent inode record. The finobt
@@ -1040,10 +1040,10 @@ xfs_dialloc_ag_finobt_near(
error = xfs_inobt_get_rec(rcur, &rrec, &j);
if (error)
goto error_rcur;
- XFS_WANT_CORRUPTED_GOTO(j == 1, error_rcur);
+ XFS_WANT_CORRUPTED_GOTO(lcur->bc_mp, j == 1, error_rcur);
}
- XFS_WANT_CORRUPTED_GOTO(i == 1 || j == 1, error_rcur);
+ XFS_WANT_CORRUPTED_GOTO(lcur->bc_mp, i == 1 || j == 1, error_rcur);
if (i == 1 && j == 1) {
/*
* Both the left and right records are valid. Choose the closer
@@ -1096,7 +1096,7 @@ xfs_dialloc_ag_finobt_newino(
error = xfs_inobt_get_rec(cur, rec, &i);
if (error)
return error;
- XFS_WANT_CORRUPTED_RETURN(i == 1);
+ XFS_WANT_CORRUPTED_RETURN(cur->bc_mp, i == 1);
return 0;
}
}
@@ -1107,12 +1107,12 @@ xfs_dialloc_ag_finobt_newino(
error = xfs_inobt_lookup(cur, 0, XFS_LOOKUP_GE, &i);
if (error)
return error;
- XFS_WANT_CORRUPTED_RETURN(i == 1);
+ XFS_WANT_CORRUPTED_RETURN(cur->bc_mp, i == 1);
error = xfs_inobt_get_rec(cur, rec, &i);
if (error)
return error;
- XFS_WANT_CORRUPTED_RETURN(i == 1);
+ XFS_WANT_CORRUPTED_RETURN(cur->bc_mp, i == 1);
return 0;
}
@@ -1134,19 +1134,19 @@ xfs_dialloc_ag_update_inobt(
error = xfs_inobt_lookup(cur, frec->ir_startino, XFS_LOOKUP_EQ, &i);
if (error)
return error;
- XFS_WANT_CORRUPTED_RETURN(i == 1);
+ XFS_WANT_CORRUPTED_RETURN(cur->bc_mp, i == 1);
error = xfs_inobt_get_rec(cur, &rec, &i);
if (error)
return error;
- XFS_WANT_CORRUPTED_RETURN(i == 1);
+ XFS_WANT_CORRUPTED_RETURN(cur->bc_mp, i == 1);
ASSERT((XFS_AGINO_TO_OFFSET(cur->bc_mp, rec.ir_startino) %
XFS_INODES_PER_CHUNK) == 0);
rec.ir_free &= ~XFS_INOBT_MASK(offset);
rec.ir_freecount--;
- XFS_WANT_CORRUPTED_RETURN((rec.ir_free == frec->ir_free) &&
+ XFS_WANT_CORRUPTED_RETURN(cur->bc_mp, (rec.ir_free == frec->ir_free) &&
(rec.ir_freecount == frec->ir_freecount));
return xfs_inobt_update(cur, &rec);
@@ -1477,14 +1477,14 @@ xfs_difree_inobt(
__func__, error);
goto error0;
}
- XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
+ XFS_WANT_CORRUPTED_GOTO(mp, i == 1, error0);
error = xfs_inobt_get_rec(cur, &rec, &i);
if (error) {
xfs_warn(mp, "%s: xfs_inobt_get_rec() returned error %d.",
__func__, error);
goto error0;
}
- XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
+ XFS_WANT_CORRUPTED_GOTO(mp, i == 1, error0);
/*
* Get the offset in the inode chunk.
*/
@@ -1594,7 +1594,7 @@ xfs_difree_finobt(
* freed an inode in a previously fully allocated chunk. If not,
* something is out of sync.
*/
- XFS_WANT_CORRUPTED_GOTO(ibtrec->ir_freecount == 1, error);
+ XFS_WANT_CORRUPTED_GOTO(mp, ibtrec->ir_freecount == 1, error);
error = xfs_inobt_insert_rec(cur, ibtrec->ir_freecount,
ibtrec->ir_free, &i);
@@ -1615,12 +1615,12 @@ xfs_difree_finobt(
error = xfs_inobt_get_rec(cur, &rec, &i);
if (error)
goto error;
- XFS_WANT_CORRUPTED_GOTO(i == 1, error);
+ XFS_WANT_CORRUPTED_GOTO(mp, i == 1, error);
rec.ir_free |= XFS_INOBT_MASK(offset);
rec.ir_freecount++;
- XFS_WANT_CORRUPTED_GOTO((rec.ir_free == ibtrec->ir_free) &&
+ XFS_WANT_CORRUPTED_GOTO(mp, (rec.ir_free == ibtrec->ir_free) &&
(rec.ir_freecount == ibtrec->ir_freecount),
error);
diff --git a/fs/xfs/libxfs/xfs_sb.c b/fs/xfs/libxfs/xfs_sb.c
index a270095ec3c0..dc4bfc5d88fc 100644
--- a/fs/xfs/libxfs/xfs_sb.c
+++ b/fs/xfs/libxfs/xfs_sb.c
@@ -111,14 +111,6 @@ xfs_mount_validate_sb(
bool check_inprogress,
bool check_version)
{
-
- /*
- * If the log device and data device have the
- * same device number, the log is internal.
- * Consequently, the sb_logstart should be non-zero. If
- * we have a zero sb_logstart in this case, we may be trying to mount
- * a volume filesystem in a non-volume manner.
- */
if (sbp->sb_magicnum != XFS_SB_MAGIC) {
xfs_warn(mp, "bad magic number");
return -EWRONGFS;
diff --git a/fs/xfs/xfs_bmap_util.c b/fs/xfs/xfs_bmap_util.c
index 22a5dcb70b32..7efa23e72a90 100644
--- a/fs/xfs/xfs_bmap_util.c
+++ b/fs/xfs/xfs_bmap_util.c
@@ -1599,13 +1599,6 @@ xfs_swap_extent_flush(
/* Verify O_DIRECT for ftmp */
if (VFS_I(ip)->i_mapping->nrpages)
return -EINVAL;
-
- /*
- * Don't try to swap extents on mmap()d files because we can't lock
- * out races against page faults safely.
- */
- if (mapping_mapped(VFS_I(ip)->i_mapping))
- return -EBUSY;
return 0;
}
@@ -1633,13 +1626,14 @@ xfs_swap_extents(
}
/*
- * Lock up the inodes against other IO and truncate to begin with.
- * Then we can ensure the inodes are flushed and have no page cache
- * safely. Once we have done this we can take the ilocks and do the rest
- * of the checks.
+ * Lock the inodes against other IO, page faults and truncate to
+ * begin with. Then we can ensure the inodes are flushed and have no
+ * page cache safely. Once we have done this we can take the ilocks and
+ * do the rest of the checks.
*/
- lock_flags = XFS_IOLOCK_EXCL;
+ lock_flags = XFS_IOLOCK_EXCL | XFS_MMAPLOCK_EXCL;
xfs_lock_two_inodes(ip, tip, XFS_IOLOCK_EXCL);
+ xfs_lock_two_inodes(ip, tip, XFS_MMAPLOCK_EXCL);
/* Verify that both files have the same format */
if ((ip->i_d.di_mode & S_IFMT) != (tip->i_d.di_mode & S_IFMT)) {
@@ -1666,8 +1660,16 @@ xfs_swap_extents(
xfs_trans_cancel(tp, 0);
goto out_unlock;
}
+
+ /*
+ * Lock and join the inodes to the tansaction so that transaction commit
+ * or cancel will unlock the inodes from this point onwards.
+ */
xfs_lock_two_inodes(ip, tip, XFS_ILOCK_EXCL);
lock_flags |= XFS_ILOCK_EXCL;
+ xfs_trans_ijoin(tp, ip, lock_flags);
+ xfs_trans_ijoin(tp, tip, lock_flags);
+
/* Verify all data are being swapped */
if (sxp->sx_offset != 0 ||
@@ -1720,9 +1722,6 @@ xfs_swap_extents(
goto out_trans_cancel;
}
- xfs_trans_ijoin(tp, ip, lock_flags);
- xfs_trans_ijoin(tp, tip, lock_flags);
-
/*
* Before we've swapped the forks, lets set the owners of the forks
* appropriately. We have to do this as we are demand paging the btree
@@ -1856,5 +1855,5 @@ out_unlock:
out_trans_cancel:
xfs_trans_cancel(tp, 0);
- goto out_unlock;
+ goto out;
}
diff --git a/fs/xfs/xfs_buf_item.c b/fs/xfs/xfs_buf_item.c
index 507d96a57ac7..092d652bc03d 100644
--- a/fs/xfs/xfs_buf_item.c
+++ b/fs/xfs/xfs_buf_item.c
@@ -537,9 +537,9 @@ xfs_buf_item_push(
/* has a previous flush failed due to IO errors? */
if ((bp->b_flags & XBF_WRITE_FAIL) &&
- ___ratelimit(&xfs_buf_write_fail_rl_state, "XFS:")) {
+ ___ratelimit(&xfs_buf_write_fail_rl_state, "XFS: Failing async write")) {
xfs_warn(bp->b_target->bt_mount,
-"Detected failing async write on buffer block 0x%llx. Retrying async write.",
+"Failing async write on buffer block 0x%llx. Retrying async write.",
(long long)bp->b_bn);
}
diff --git a/fs/xfs/xfs_discard.c b/fs/xfs/xfs_discard.c
index 799e5a2d334d..e85a9519a5ae 100644
--- a/fs/xfs/xfs_discard.c
+++ b/fs/xfs/xfs_discard.c
@@ -84,7 +84,7 @@ xfs_trim_extents(
error = xfs_alloc_get_rec(cur, &fbno, &flen, &i);
if (error)
goto out_del_cursor;
- XFS_WANT_CORRUPTED_GOTO(i == 1, out_del_cursor);
+ XFS_WANT_CORRUPTED_GOTO(mp, i == 1, out_del_cursor);
ASSERT(flen <= be32_to_cpu(XFS_BUF_TO_AGF(agbp)->agf_longest));
/*
diff --git a/fs/xfs/xfs_error.h b/fs/xfs/xfs_error.h
index 279a76e52791..c0394ed126fc 100644
--- a/fs/xfs/xfs_error.h
+++ b/fs/xfs/xfs_error.h
@@ -40,25 +40,25 @@ extern void xfs_verifier_error(struct xfs_buf *bp);
/*
* Macros to set EFSCORRUPTED & return/branch.
*/
-#define XFS_WANT_CORRUPTED_GOTO(x,l) \
+#define XFS_WANT_CORRUPTED_GOTO(mp, x, l) \
{ \
int fs_is_ok = (x); \
ASSERT(fs_is_ok); \
if (unlikely(!fs_is_ok)) { \
XFS_ERROR_REPORT("XFS_WANT_CORRUPTED_GOTO", \
- XFS_ERRLEVEL_LOW, NULL); \
+ XFS_ERRLEVEL_LOW, mp); \
error = -EFSCORRUPTED; \
goto l; \
} \
}
-#define XFS_WANT_CORRUPTED_RETURN(x) \
+#define XFS_WANT_CORRUPTED_RETURN(mp, x) \
{ \
int fs_is_ok = (x); \
ASSERT(fs_is_ok); \
if (unlikely(!fs_is_ok)) { \
XFS_ERROR_REPORT("XFS_WANT_CORRUPTED_RETURN", \
- XFS_ERRLEVEL_LOW, NULL); \
+ XFS_ERRLEVEL_LOW, mp); \
return -EFSCORRUPTED; \
} \
}
diff --git a/fs/xfs/xfs_file.c b/fs/xfs/xfs_file.c
index ce615d12fb44..b101e80f2862 100644
--- a/fs/xfs/xfs_file.c
+++ b/fs/xfs/xfs_file.c
@@ -397,7 +397,8 @@ STATIC int /* error (positive) */
xfs_zero_last_block(
struct xfs_inode *ip,
xfs_fsize_t offset,
- xfs_fsize_t isize)
+ xfs_fsize_t isize,
+ bool *did_zeroing)
{
struct xfs_mount *mp = ip->i_mount;
xfs_fileoff_t last_fsb = XFS_B_TO_FSBT(mp, isize);
@@ -425,6 +426,7 @@ xfs_zero_last_block(
zero_len = mp->m_sb.sb_blocksize - zero_offset;
if (isize + zero_len > offset)
zero_len = offset - isize;
+ *did_zeroing = true;
return xfs_iozero(ip, isize, zero_len);
}
@@ -443,7 +445,8 @@ int /* error (positive) */
xfs_zero_eof(
struct xfs_inode *ip,
xfs_off_t offset, /* starting I/O offset */
- xfs_fsize_t isize) /* current inode size */
+ xfs_fsize_t isize, /* current inode size */
+ bool *did_zeroing)
{
struct xfs_mount *mp = ip->i_mount;
xfs_fileoff_t start_zero_fsb;
@@ -465,7 +468,7 @@ xfs_zero_eof(
* We only zero a part of that block so it is handled specially.
*/
if (XFS_B_FSB_OFFSET(mp, isize) != 0) {
- error = xfs_zero_last_block(ip, offset, isize);
+ error = xfs_zero_last_block(ip, offset, isize, did_zeroing);
if (error)
return error;
}
@@ -525,6 +528,7 @@ xfs_zero_eof(
if (error)
return error;
+ *did_zeroing = true;
start_zero_fsb = imap.br_startoff + imap.br_blockcount;
ASSERT(start_zero_fsb <= (end_zero_fsb + 1));
}
@@ -567,13 +571,15 @@ restart:
* having to redo all checks before.
*/
if (*pos > i_size_read(inode)) {
+ bool zero = false;
+
if (*iolock == XFS_IOLOCK_SHARED) {
xfs_rw_iunlock(ip, *iolock);
*iolock = XFS_IOLOCK_EXCL;
xfs_rw_ilock(ip, *iolock);
goto restart;
}
- error = xfs_zero_eof(ip, *pos, i_size_read(inode));
+ error = xfs_zero_eof(ip, *pos, i_size_read(inode), &zero);
if (error)
return error;
}
@@ -841,6 +847,9 @@ xfs_file_fallocate(
if (error)
goto out_unlock;
+ xfs_ilock(ip, XFS_MMAPLOCK_EXCL);
+ iolock |= XFS_MMAPLOCK_EXCL;
+
if (mode & FALLOC_FL_PUNCH_HOLE) {
error = xfs_free_file_space(ip, offset, len);
if (error)
@@ -991,20 +1000,6 @@ xfs_file_mmap(
}
/*
- * mmap()d file has taken write protection fault and is being made
- * writable. We can set the page state up correctly for a writable
- * page, which means we can do correct delalloc accounting (ENOSPC
- * checking!) and unwritten extent mapping.
- */
-STATIC int
-xfs_vm_page_mkwrite(
- struct vm_area_struct *vma,
- struct vm_fault *vmf)
-{
- return block_page_mkwrite(vma, vmf, xfs_get_blocks);
-}
-
-/*
* This type is designed to indicate the type of offset we would like
* to search from page cache for xfs_seek_hole_data().
*/
@@ -1379,6 +1374,55 @@ xfs_file_llseek(
}
}
+/*
+ * Locking for serialisation of IO during page faults. This results in a lock
+ * ordering of:
+ *
+ * mmap_sem (MM)
+ * i_mmap_lock (XFS - truncate serialisation)
+ * page_lock (MM)
+ * i_lock (XFS - extent map serialisation)
+ */
+STATIC int
+xfs_filemap_fault(
+ struct vm_area_struct *vma,
+ struct vm_fault *vmf)
+{
+ struct xfs_inode *ip = XFS_I(vma->vm_file->f_mapping->host);
+ int error;
+
+ trace_xfs_filemap_fault(ip);
+
+ xfs_ilock(ip, XFS_MMAPLOCK_SHARED);
+ error = filemap_fault(vma, vmf);
+ xfs_iunlock(ip, XFS_MMAPLOCK_SHARED);
+
+ return error;
+}
+
+/*
+ * mmap()d file has taken write protection fault and is being made writable. We
+ * can set the page state up correctly for a writable page, which means we can
+ * do correct delalloc accounting (ENOSPC checking!) and unwritten extent
+ * mapping.
+ */
+STATIC int
+xfs_filemap_page_mkwrite(
+ struct vm_area_struct *vma,
+ struct vm_fault *vmf)
+{
+ struct xfs_inode *ip = XFS_I(vma->vm_file->f_mapping->host);
+ int error;
+
+ trace_xfs_filemap_page_mkwrite(ip);
+
+ xfs_ilock(ip, XFS_MMAPLOCK_SHARED);
+ error = block_page_mkwrite(vma, vmf, xfs_get_blocks);
+ xfs_iunlock(ip, XFS_MMAPLOCK_SHARED);
+
+ return error;
+}
+
const struct file_operations xfs_file_operations = {
.llseek = xfs_file_llseek,
.read = new_sync_read,
@@ -1411,7 +1455,7 @@ const struct file_operations xfs_dir_file_operations = {
};
static const struct vm_operations_struct xfs_file_vm_ops = {
- .fault = filemap_fault,
+ .fault = xfs_filemap_fault,
.map_pages = filemap_map_pages,
- .page_mkwrite = xfs_vm_page_mkwrite,
+ .page_mkwrite = xfs_filemap_page_mkwrite,
};
diff --git a/fs/xfs/xfs_icache.c b/fs/xfs/xfs_icache.c
index 9771b7ef62ed..76a9f2783282 100644
--- a/fs/xfs/xfs_icache.c
+++ b/fs/xfs/xfs_icache.c
@@ -439,11 +439,11 @@ again:
*ipp = ip;
/*
- * If we have a real type for an on-disk inode, we can set ops(&unlock)
+ * If we have a real type for an on-disk inode, we can setup the inode
* now. If it's a new inode being created, xfs_ialloc will handle it.
*/
if (xfs_iflags_test(ip, XFS_INEW) && ip->i_d.di_mode != 0)
- xfs_setup_inode(ip);
+ xfs_setup_existing_inode(ip);
return 0;
out_error_or_again:
diff --git a/fs/xfs/xfs_inode.c b/fs/xfs/xfs_inode.c
index b7064f20de7f..d6ebc85192b7 100644
--- a/fs/xfs/xfs_inode.c
+++ b/fs/xfs/xfs_inode.c
@@ -117,24 +117,34 @@ xfs_ilock_attr_map_shared(
}
/*
- * The xfs inode contains 2 locks: a multi-reader lock called the
- * i_iolock and a multi-reader lock called the i_lock. This routine
- * allows either or both of the locks to be obtained.
+ * The xfs inode contains 3 multi-reader locks: the i_iolock the i_mmap_lock and
+ * the i_lock. This routine allows various combinations of the locks to be
+ * obtained.
*
- * The 2 locks should always be ordered so that the IO lock is
- * obtained first in order to prevent deadlock.
+ * The 3 locks should always be ordered so that the IO lock is obtained first,
+ * the mmap lock second and the ilock last in order to prevent deadlock.
*
- * ip -- the inode being locked
- * lock_flags -- this parameter indicates the inode's locks
- * to be locked. It can be:
- * XFS_IOLOCK_SHARED,
- * XFS_IOLOCK_EXCL,
- * XFS_ILOCK_SHARED,
- * XFS_ILOCK_EXCL,
- * XFS_IOLOCK_SHARED | XFS_ILOCK_SHARED,
- * XFS_IOLOCK_SHARED | XFS_ILOCK_EXCL,
- * XFS_IOLOCK_EXCL | XFS_ILOCK_SHARED,
- * XFS_IOLOCK_EXCL | XFS_ILOCK_EXCL
+ * Basic locking order:
+ *
+ * i_iolock -> i_mmap_lock -> page_lock -> i_ilock
+ *
+ * mmap_sem locking order:
+ *
+ * i_iolock -> page lock -> mmap_sem
+ * mmap_sem -> i_mmap_lock -> page_lock
+ *
+ * The difference in mmap_sem locking order mean that we cannot hold the
+ * i_mmap_lock over syscall based read(2)/write(2) based IO. These IO paths can
+ * fault in pages during copy in/out (for buffered IO) or require the mmap_sem
+ * in get_user_pages() to map the user pages into the kernel address space for
+ * direct IO. Similarly the i_iolock cannot be taken inside a page fault because
+ * page faults already hold the mmap_sem.
+ *
+ * Hence to serialise fully against both syscall and mmap based IO, we need to
+ * take both the i_iolock and the i_mmap_lock. These locks should *only* be both
+ * taken in places where we need to invalidate the page cache in a race
+ * free manner (e.g. truncate, hole punch and other extent manipulation
+ * functions).
*/
void
xfs_ilock(
@@ -150,6 +160,8 @@ xfs_ilock(
*/
ASSERT((lock_flags & (XFS_IOLOCK_SHARED | XFS_IOLOCK_EXCL)) !=
(XFS_IOLOCK_SHARED | XFS_IOLOCK_EXCL));
+ ASSERT((lock_flags & (XFS_MMAPLOCK_SHARED | XFS_MMAPLOCK_EXCL)) !=
+ (XFS_MMAPLOCK_SHARED | XFS_MMAPLOCK_EXCL));
ASSERT((lock_flags & (XFS_ILOCK_SHARED | XFS_ILOCK_EXCL)) !=
(XFS_ILOCK_SHARED | XFS_ILOCK_EXCL));
ASSERT((lock_flags & ~(XFS_LOCK_MASK | XFS_LOCK_DEP_MASK)) == 0);
@@ -159,6 +171,11 @@ xfs_ilock(
else if (lock_flags & XFS_IOLOCK_SHARED)
mraccess_nested(&ip->i_iolock, XFS_IOLOCK_DEP(lock_flags));
+ if (lock_flags & XFS_MMAPLOCK_EXCL)
+ mrupdate_nested(&ip->i_mmaplock, XFS_MMAPLOCK_DEP(lock_flags));
+ else if (lock_flags & XFS_MMAPLOCK_SHARED)
+ mraccess_nested(&ip->i_mmaplock, XFS_MMAPLOCK_DEP(lock_flags));
+
if (lock_flags & XFS_ILOCK_EXCL)
mrupdate_nested(&ip->i_lock, XFS_ILOCK_DEP(lock_flags));
else if (lock_flags & XFS_ILOCK_SHARED)
@@ -191,6 +208,8 @@ xfs_ilock_nowait(
*/
ASSERT((lock_flags & (XFS_IOLOCK_SHARED | XFS_IOLOCK_EXCL)) !=
(XFS_IOLOCK_SHARED | XFS_IOLOCK_EXCL));
+ ASSERT((lock_flags & (XFS_MMAPLOCK_SHARED | XFS_MMAPLOCK_EXCL)) !=
+ (XFS_MMAPLOCK_SHARED | XFS_MMAPLOCK_EXCL));
ASSERT((lock_flags & (XFS_ILOCK_SHARED | XFS_ILOCK_EXCL)) !=
(XFS_ILOCK_SHARED | XFS_ILOCK_EXCL));
ASSERT((lock_flags & ~(XFS_LOCK_MASK | XFS_LOCK_DEP_MASK)) == 0);
@@ -202,21 +221,35 @@ xfs_ilock_nowait(
if (!mrtryaccess(&ip->i_iolock))
goto out;
}
+
+ if (lock_flags & XFS_MMAPLOCK_EXCL) {
+ if (!mrtryupdate(&ip->i_mmaplock))
+ goto out_undo_iolock;
+ } else if (lock_flags & XFS_MMAPLOCK_SHARED) {
+ if (!mrtryaccess(&ip->i_mmaplock))
+ goto out_undo_iolock;
+ }
+
if (lock_flags & XFS_ILOCK_EXCL) {
if (!mrtryupdate(&ip->i_lock))
- goto out_undo_iolock;
+ goto out_undo_mmaplock;
} else if (lock_flags & XFS_ILOCK_SHARED) {
if (!mrtryaccess(&ip->i_lock))
- goto out_undo_iolock;
+ goto out_undo_mmaplock;
}
return 1;
- out_undo_iolock:
+out_undo_mmaplock:
+ if (lock_flags & XFS_MMAPLOCK_EXCL)
+ mrunlock_excl(&ip->i_mmaplock);
+ else if (lock_flags & XFS_MMAPLOCK_SHARED)
+ mrunlock_shared(&ip->i_mmaplock);
+out_undo_iolock:
if (lock_flags & XFS_IOLOCK_EXCL)
mrunlock_excl(&ip->i_iolock);
else if (lock_flags & XFS_IOLOCK_SHARED)
mrunlock_shared(&ip->i_iolock);
- out:
+out:
return 0;
}
@@ -244,6 +277,8 @@ xfs_iunlock(
*/
ASSERT((lock_flags & (XFS_IOLOCK_SHARED | XFS_IOLOCK_EXCL)) !=
(XFS_IOLOCK_SHARED | XFS_IOLOCK_EXCL));
+ ASSERT((lock_flags & (XFS_MMAPLOCK_SHARED | XFS_MMAPLOCK_EXCL)) !=
+ (XFS_MMAPLOCK_SHARED | XFS_MMAPLOCK_EXCL));
ASSERT((lock_flags & (XFS_ILOCK_SHARED | XFS_ILOCK_EXCL)) !=
(XFS_ILOCK_SHARED | XFS_ILOCK_EXCL));
ASSERT((lock_flags & ~(XFS_LOCK_MASK | XFS_LOCK_DEP_MASK)) == 0);
@@ -254,6 +289,11 @@ xfs_iunlock(
else if (lock_flags & XFS_IOLOCK_SHARED)
mrunlock_shared(&ip->i_iolock);
+ if (lock_flags & XFS_MMAPLOCK_EXCL)
+ mrunlock_excl(&ip->i_mmaplock);
+ else if (lock_flags & XFS_MMAPLOCK_SHARED)
+ mrunlock_shared(&ip->i_mmaplock);
+
if (lock_flags & XFS_ILOCK_EXCL)
mrunlock_excl(&ip->i_lock);
else if (lock_flags & XFS_ILOCK_SHARED)
@@ -271,11 +311,14 @@ xfs_ilock_demote(
xfs_inode_t *ip,
uint lock_flags)
{
- ASSERT(lock_flags & (XFS_IOLOCK_EXCL|XFS_ILOCK_EXCL));
- ASSERT((lock_flags & ~(XFS_IOLOCK_EXCL|XFS_ILOCK_EXCL)) == 0);
+ ASSERT(lock_flags & (XFS_IOLOCK_EXCL|XFS_MMAPLOCK_EXCL|XFS_ILOCK_EXCL));
+ ASSERT((lock_flags &
+ ~(XFS_IOLOCK_EXCL|XFS_MMAPLOCK_EXCL|XFS_ILOCK_EXCL)) == 0);
if (lock_flags & XFS_ILOCK_EXCL)
mrdemote(&ip->i_lock);
+ if (lock_flags & XFS_MMAPLOCK_EXCL)
+ mrdemote(&ip->i_mmaplock);
if (lock_flags & XFS_IOLOCK_EXCL)
mrdemote(&ip->i_iolock);
@@ -294,6 +337,12 @@ xfs_isilocked(
return rwsem_is_locked(&ip->i_lock.mr_lock);
}
+ if (lock_flags & (XFS_MMAPLOCK_EXCL|XFS_MMAPLOCK_SHARED)) {
+ if (!(lock_flags & XFS_MMAPLOCK_SHARED))
+ return !!ip->i_mmaplock.mr_writer;
+ return rwsem_is_locked(&ip->i_mmaplock.mr_lock);
+ }
+
if (lock_flags & (XFS_IOLOCK_EXCL|XFS_IOLOCK_SHARED)) {
if (!(lock_flags & XFS_IOLOCK_SHARED))
return !!ip->i_iolock.mr_writer;
@@ -314,14 +363,27 @@ int xfs_lock_delays;
#endif
/*
- * Bump the subclass so xfs_lock_inodes() acquires each lock with
- * a different value
+ * Bump the subclass so xfs_lock_inodes() acquires each lock with a different
+ * value. This shouldn't be called for page fault locking, but we also need to
+ * ensure we don't overrun the number of lockdep subclasses for the iolock or
+ * mmaplock as that is limited to 12 by the mmap lock lockdep annotations.
*/
static inline int
xfs_lock_inumorder(int lock_mode, int subclass)
{
- if (lock_mode & (XFS_IOLOCK_SHARED|XFS_IOLOCK_EXCL))
+ if (lock_mode & (XFS_IOLOCK_SHARED|XFS_IOLOCK_EXCL)) {
+ ASSERT(subclass + XFS_LOCK_INUMORDER <
+ (1 << (XFS_MMAPLOCK_SHIFT - XFS_IOLOCK_SHIFT)));
lock_mode |= (subclass + XFS_LOCK_INUMORDER) << XFS_IOLOCK_SHIFT;
+ }
+
+ if (lock_mode & (XFS_MMAPLOCK_SHARED|XFS_MMAPLOCK_EXCL)) {
+ ASSERT(subclass + XFS_LOCK_INUMORDER <
+ (1 << (XFS_ILOCK_SHIFT - XFS_MMAPLOCK_SHIFT)));
+ lock_mode |= (subclass + XFS_LOCK_INUMORDER) <<
+ XFS_MMAPLOCK_SHIFT;
+ }
+
if (lock_mode & (XFS_ILOCK_SHARED|XFS_ILOCK_EXCL))
lock_mode |= (subclass + XFS_LOCK_INUMORDER) << XFS_ILOCK_SHIFT;
@@ -329,15 +391,14 @@ xfs_lock_inumorder(int lock_mode, int subclass)
}
/*
- * The following routine will lock n inodes in exclusive mode.
- * We assume the caller calls us with the inodes in i_ino order.
+ * The following routine will lock n inodes in exclusive mode. We assume the
+ * caller calls us with the inodes in i_ino order.
*
- * We need to detect deadlock where an inode that we lock
- * is in the AIL and we start waiting for another inode that is locked
- * by a thread in a long running transaction (such as truncate). This can
- * result in deadlock since the long running trans might need to wait
- * for the inode we just locked in order to push the tail and free space
- * in the log.
+ * We need to detect deadlock where an inode that we lock is in the AIL and we
+ * start waiting for another inode that is locked by a thread in a long running
+ * transaction (such as truncate). This can result in deadlock since the long
+ * running trans might need to wait for the inode we just locked in order to
+ * push the tail and free space in the log.
*/
void
xfs_lock_inodes(
@@ -348,30 +409,27 @@ xfs_lock_inodes(
int attempts = 0, i, j, try_lock;
xfs_log_item_t *lp;
- ASSERT(ips && (inodes >= 2)); /* we need at least two */
+ /* currently supports between 2 and 5 inodes */
+ ASSERT(ips && inodes >= 2 && inodes <= 5);
try_lock = 0;
i = 0;
-
again:
for (; i < inodes; i++) {
ASSERT(ips[i]);
- if (i && (ips[i] == ips[i-1])) /* Already locked */
+ if (i && (ips[i] == ips[i - 1])) /* Already locked */
continue;
/*
- * If try_lock is not set yet, make sure all locked inodes
- * are not in the AIL.
- * If any are, set try_lock to be used later.
+ * If try_lock is not set yet, make sure all locked inodes are
+ * not in the AIL. If any are, set try_lock to be used later.
*/
-
if (!try_lock) {
for (j = (i - 1); j >= 0 && !try_lock; j--) {
lp = (xfs_log_item_t *)ips[j]->i_itemp;
- if (lp && (lp->li_flags & XFS_LI_IN_AIL)) {
+ if (lp && (lp->li_flags & XFS_LI_IN_AIL))
try_lock++;
- }
}
}
@@ -381,51 +439,42 @@ again:
* we can't get any, we must release all we have
* and try again.
*/
+ if (!try_lock) {
+ xfs_ilock(ips[i], xfs_lock_inumorder(lock_mode, i));
+ continue;
+ }
+
+ /* try_lock means we have an inode locked that is in the AIL. */
+ ASSERT(i != 0);
+ if (xfs_ilock_nowait(ips[i], xfs_lock_inumorder(lock_mode, i)))
+ continue;
- if (try_lock) {
- /* try_lock must be 0 if i is 0. */
+ /*
+ * Unlock all previous guys and try again. xfs_iunlock will try
+ * to push the tail if the inode is in the AIL.
+ */
+ attempts++;
+ for (j = i - 1; j >= 0; j--) {
/*
- * try_lock means we have an inode locked
- * that is in the AIL.
+ * Check to see if we've already unlocked this one. Not
+ * the first one going back, and the inode ptr is the
+ * same.
*/
- ASSERT(i != 0);
- if (!xfs_ilock_nowait(ips[i], xfs_lock_inumorder(lock_mode, i))) {
- attempts++;
-
- /*
- * Unlock all previous guys and try again.
- * xfs_iunlock will try to push the tail
- * if the inode is in the AIL.
- */
-
- for(j = i - 1; j >= 0; j--) {
-
- /*
- * Check to see if we've already
- * unlocked this one.
- * Not the first one going back,
- * and the inode ptr is the same.
- */
- if ((j != (i - 1)) && ips[j] ==
- ips[j+1])
- continue;
-
- xfs_iunlock(ips[j], lock_mode);
- }
+ if (j != (i - 1) && ips[j] == ips[j + 1])
+ continue;
+
+ xfs_iunlock(ips[j], lock_mode);
+ }
- if ((attempts % 5) == 0) {
- delay(1); /* Don't just spin the CPU */
+ if ((attempts % 5) == 0) {
+ delay(1); /* Don't just spin the CPU */
#ifdef DEBUG
- xfs_lock_delays++;
+ xfs_lock_delays++;
#endif
- }
- i = 0;
- try_lock = 0;
- goto again;
- }
- } else {
- xfs_ilock(ips[i], xfs_lock_inumorder(lock_mode, i));
}
+ i = 0;
+ try_lock = 0;
+ goto again;
}
#ifdef DEBUG
@@ -440,10 +489,10 @@ again:
}
/*
- * xfs_lock_two_inodes() can only be used to lock one type of lock
- * at a time - the iolock or the ilock, but not both at once. If
- * we lock both at once, lockdep will report false positives saying
- * we have violated locking orders.
+ * xfs_lock_two_inodes() can only be used to lock one type of lock at a time -
+ * the iolock, the mmaplock or the ilock, but not more than one at a time. If we
+ * lock more than one at a time, lockdep will report false positives saying we
+ * have violated locking orders.
*/
void
xfs_lock_two_inodes(
@@ -455,8 +504,12 @@ xfs_lock_two_inodes(
int attempts = 0;
xfs_log_item_t *lp;
- if (lock_mode & (XFS_IOLOCK_SHARED|XFS_IOLOCK_EXCL))
- ASSERT((lock_mode & (XFS_ILOCK_SHARED|XFS_ILOCK_EXCL)) == 0);
+ if (lock_mode & (XFS_IOLOCK_SHARED|XFS_IOLOCK_EXCL)) {
+ ASSERT(!(lock_mode & (XFS_MMAPLOCK_SHARED|XFS_MMAPLOCK_EXCL)));
+ ASSERT(!(lock_mode & (XFS_ILOCK_SHARED|XFS_ILOCK_EXCL)));
+ } else if (lock_mode & (XFS_MMAPLOCK_SHARED|XFS_MMAPLOCK_EXCL))
+ ASSERT(!(lock_mode & (XFS_ILOCK_SHARED|XFS_ILOCK_EXCL)));
+
ASSERT(ip0->i_ino != ip1->i_ino);
if (ip0->i_ino > ip1->i_ino) {
@@ -818,7 +871,7 @@ xfs_ialloc(
xfs_trans_ijoin(tp, ip, XFS_ILOCK_EXCL);
xfs_trans_log_inode(tp, ip, flags);
- /* now that we have an i_mode we can setup inode ops and unlock */
+ /* now that we have an i_mode we can setup the inode structure */
xfs_setup_inode(ip);
*ipp = ip;
@@ -1235,12 +1288,14 @@ xfs_create(
xfs_trans_cancel(tp, cancel_flags);
out_release_inode:
/*
- * Wait until after the current transaction is aborted to
- * release the inode. This prevents recursive transactions
- * and deadlocks from xfs_inactive.
+ * Wait until after the current transaction is aborted to finish the
+ * setup of the inode and release the inode. This prevents recursive
+ * transactions and deadlocks from xfs_inactive.
*/
- if (ip)
+ if (ip) {
+ xfs_finish_inode_setup(ip);
IRELE(ip);
+ }
xfs_qm_dqrele(udqp);
xfs_qm_dqrele(gdqp);
@@ -1345,12 +1400,14 @@ xfs_create_tmpfile(
xfs_trans_cancel(tp, cancel_flags);
out_release_inode:
/*
- * Wait until after the current transaction is aborted to
- * release the inode. This prevents recursive transactions
- * and deadlocks from xfs_inactive.
+ * Wait until after the current transaction is aborted to finish the
+ * setup of the inode and release the inode. This prevents recursive
+ * transactions and deadlocks from xfs_inactive.
*/
- if (ip)
+ if (ip) {
+ xfs_finish_inode_setup(ip);
IRELE(ip);
+ }
xfs_qm_dqrele(udqp);
xfs_qm_dqrele(gdqp);
@@ -2611,19 +2668,22 @@ xfs_remove(
/*
* Enter all inodes for a rename transaction into a sorted array.
*/
+#define __XFS_SORT_INODES 5
STATIC void
xfs_sort_for_rename(
- xfs_inode_t *dp1, /* in: old (source) directory inode */
- xfs_inode_t *dp2, /* in: new (target) directory inode */
- xfs_inode_t *ip1, /* in: inode of old entry */
- xfs_inode_t *ip2, /* in: inode of new entry, if it
- already exists, NULL otherwise. */
- xfs_inode_t **i_tab,/* out: array of inode returned, sorted */
- int *num_inodes) /* out: number of inodes in array */
+ struct xfs_inode *dp1, /* in: old (source) directory inode */
+ struct xfs_inode *dp2, /* in: new (target) directory inode */
+ struct xfs_inode *ip1, /* in: inode of old entry */
+ struct xfs_inode *ip2, /* in: inode of new entry */
+ struct xfs_inode *wip, /* in: whiteout inode */
+ struct xfs_inode **i_tab,/* out: sorted array of inodes */
+ int *num_inodes) /* in/out: inodes in array */
{
- xfs_inode_t *temp;
int i, j;
+ ASSERT(*num_inodes == __XFS_SORT_INODES);
+ memset(i_tab, 0, *num_inodes * sizeof(struct xfs_inode *));
+
/*
* i_tab contains a list of pointers to inodes. We initialize
* the table here & we'll sort it. We will then use it to
@@ -2631,25 +2691,24 @@ xfs_sort_for_rename(
*
* Note that the table may contain duplicates. e.g., dp1 == dp2.
*/
- i_tab[0] = dp1;
- i_tab[1] = dp2;
- i_tab[2] = ip1;
- if (ip2) {
- *num_inodes = 4;
- i_tab[3] = ip2;
- } else {
- *num_inodes = 3;
- i_tab[3] = NULL;
- }
+ i = 0;
+ i_tab[i++] = dp1;
+ i_tab[i++] = dp2;
+ i_tab[i++] = ip1;
+ if (ip2)
+ i_tab[i++] = ip2;
+ if (wip)
+ i_tab[i++] = wip;
+ *num_inodes = i;
/*
* Sort the elements via bubble sort. (Remember, there are at
- * most 4 elements to sort, so this is adequate.)
+ * most 5 elements to sort, so this is adequate.)
*/
for (i = 0; i < *num_inodes; i++) {
for (j = 1; j < *num_inodes; j++) {
if (i_tab[j]->i_ino < i_tab[j-1]->i_ino) {
- temp = i_tab[j];
+ struct xfs_inode *temp = i_tab[j];
i_tab[j] = i_tab[j-1];
i_tab[j-1] = temp;
}
@@ -2657,6 +2716,31 @@ xfs_sort_for_rename(
}
}
+static int
+xfs_finish_rename(
+ struct xfs_trans *tp,
+ struct xfs_bmap_free *free_list)
+{
+ int committed = 0;
+ int error;
+
+ /*
+ * If this is a synchronous mount, make sure that the rename transaction
+ * goes to disk before returning to the user.
+ */
+ if (tp->t_mountp->m_flags & (XFS_MOUNT_WSYNC|XFS_MOUNT_DIRSYNC))
+ xfs_trans_set_sync(tp);
+
+ error = xfs_bmap_finish(&tp, free_list, &committed);
+ if (error) {
+ xfs_bmap_cancel(free_list);
+ xfs_trans_cancel(tp, XFS_TRANS_RELEASE_LOG_RES|XFS_TRANS_ABORT);
+ return error;
+ }
+
+ return xfs_trans_commit(tp, XFS_TRANS_RELEASE_LOG_RES);
+}
+
/*
* xfs_cross_rename()
*
@@ -2685,14 +2769,14 @@ xfs_cross_rename(
ip2->i_ino,
first_block, free_list, spaceres);
if (error)
- goto out;
+ goto out_trans_abort;
/* Swap inode number for dirent in second parent */
error = xfs_dir_replace(tp, dp2, name2,
ip1->i_ino,
first_block, free_list, spaceres);
if (error)
- goto out;
+ goto out_trans_abort;
/*
* If we're renaming one or more directories across different parents,
@@ -2707,16 +2791,16 @@ xfs_cross_rename(
dp1->i_ino, first_block,
free_list, spaceres);
if (error)
- goto out;
+ goto out_trans_abort;
/* transfer ip2 ".." reference to dp1 */
if (!S_ISDIR(ip1->i_d.di_mode)) {
error = xfs_droplink(tp, dp2);
if (error)
- goto out;
+ goto out_trans_abort;
error = xfs_bumplink(tp, dp1);
if (error)
- goto out;
+ goto out_trans_abort;
}
/*
@@ -2734,16 +2818,16 @@ xfs_cross_rename(
dp2->i_ino, first_block,
free_list, spaceres);
if (error)
- goto out;
+ goto out_trans_abort;
/* transfer ip1 ".." reference to dp2 */
if (!S_ISDIR(ip2->i_d.di_mode)) {
error = xfs_droplink(tp, dp1);
if (error)
- goto out;
+ goto out_trans_abort;
error = xfs_bumplink(tp, dp2);
if (error)
- goto out;
+ goto out_trans_abort;
}
/*
@@ -2771,66 +2855,108 @@ xfs_cross_rename(
}
xfs_trans_ichgtime(tp, dp1, XFS_ICHGTIME_MOD | XFS_ICHGTIME_CHG);
xfs_trans_log_inode(tp, dp1, XFS_ILOG_CORE);
-out:
+ return xfs_finish_rename(tp, free_list);
+
+out_trans_abort:
+ xfs_bmap_cancel(free_list);
+ xfs_trans_cancel(tp, XFS_TRANS_RELEASE_LOG_RES|XFS_TRANS_ABORT);
return error;
}
/*
+ * xfs_rename_alloc_whiteout()
+ *
+ * Return a referenced, unlinked, unlocked inode that that can be used as a
+ * whiteout in a rename transaction. We use a tmpfile inode here so that if we
+ * crash between allocating the inode and linking it into the rename transaction
+ * recovery will free the inode and we won't leak it.
+ */
+static int
+xfs_rename_alloc_whiteout(
+ struct xfs_inode *dp,
+ struct xfs_inode **wip)
+{
+ struct xfs_inode *tmpfile;
+ int error;
+
+ error = xfs_create_tmpfile(dp, NULL, S_IFCHR | WHITEOUT_MODE, &tmpfile);
+ if (error)
+ return error;
+
+ /* Satisfy xfs_bumplink that this is a real tmpfile */
+ xfs_finish_inode_setup(tmpfile);
+ VFS_I(tmpfile)->i_state |= I_LINKABLE;
+
+ *wip = tmpfile;
+ return 0;
+}
+
+/*
* xfs_rename
*/
int
xfs_rename(
- xfs_inode_t *src_dp,
- struct xfs_name *src_name,
- xfs_inode_t *src_ip,
- xfs_inode_t *target_dp,
- struct xfs_name *target_name,
- xfs_inode_t *target_ip,
- unsigned int flags)
+ struct xfs_inode *src_dp,
+ struct xfs_name *src_name,
+ struct xfs_inode *src_ip,
+ struct xfs_inode *target_dp,
+ struct xfs_name *target_name,
+ struct xfs_inode *target_ip,
+ unsigned int flags)
{
- xfs_trans_t *tp = NULL;
- xfs_mount_t *mp = src_dp->i_mount;
- bool new_parent; /* moving to a new dir */
- bool src_is_directory; /* src_name is a directory */
- int error;
- xfs_bmap_free_t free_list;
- xfs_fsblock_t first_block;
- int cancel_flags;
- int committed;
- xfs_inode_t *inodes[4];
- int spaceres;
- int num_inodes;
+ struct xfs_mount *mp = src_dp->i_mount;
+ struct xfs_trans *tp;
+ struct xfs_bmap_free free_list;
+ xfs_fsblock_t first_block;
+ struct xfs_inode *wip = NULL; /* whiteout inode */
+ struct xfs_inode *inodes[__XFS_SORT_INODES];
+ int num_inodes = __XFS_SORT_INODES;
+ bool new_parent = (src_dp != target_dp);
+ bool src_is_directory = S_ISDIR(src_ip->i_d.di_mode);
+ int cancel_flags = 0;
+ int spaceres;
+ int error;
trace_xfs_rename(src_dp, target_dp, src_name, target_name);
- new_parent = (src_dp != target_dp);
- src_is_directory = S_ISDIR(src_ip->i_d.di_mode);
+ if ((flags & RENAME_EXCHANGE) && !target_ip)
+ return -EINVAL;
+
+ /*
+ * If we are doing a whiteout operation, allocate the whiteout inode
+ * we will be placing at the target and ensure the type is set
+ * appropriately.
+ */
+ if (flags & RENAME_WHITEOUT) {
+ ASSERT(!(flags & (RENAME_NOREPLACE | RENAME_EXCHANGE)));
+ error = xfs_rename_alloc_whiteout(target_dp, &wip);
+ if (error)
+ return error;
+
+ /* setup target dirent info as whiteout */
+ src_name->type = XFS_DIR3_FT_CHRDEV;
+ }
- xfs_sort_for_rename(src_dp, target_dp, src_ip, target_ip,
+ xfs_sort_for_rename(src_dp, target_dp, src_ip, target_ip, wip,
inodes, &num_inodes);
- xfs_bmap_init(&free_list, &first_block);
tp = xfs_trans_alloc(mp, XFS_TRANS_RENAME);
- cancel_flags = XFS_TRANS_RELEASE_LOG_RES;
spaceres = XFS_RENAME_SPACE_RES(mp, target_name->len);
error = xfs_trans_reserve(tp, &M_RES(mp)->tr_rename, spaceres, 0);
if (error == -ENOSPC) {
spaceres = 0;
error = xfs_trans_reserve(tp, &M_RES(mp)->tr_rename, 0, 0);
}
- if (error) {
- xfs_trans_cancel(tp, 0);
- goto std_return;
- }
+ if (error)
+ goto out_trans_cancel;
+ cancel_flags = XFS_TRANS_RELEASE_LOG_RES;
/*
* Attach the dquots to the inodes
*/
error = xfs_qm_vop_rename_dqattach(inodes);
- if (error) {
- xfs_trans_cancel(tp, cancel_flags);
- goto std_return;
- }
+ if (error)
+ goto out_trans_cancel;
/*
* Lock all the participating inodes. Depending upon whether
@@ -2851,6 +2977,8 @@ xfs_rename(
xfs_trans_ijoin(tp, src_ip, XFS_ILOCK_EXCL);
if (target_ip)
xfs_trans_ijoin(tp, target_ip, XFS_ILOCK_EXCL);
+ if (wip)
+ xfs_trans_ijoin(tp, wip, XFS_ILOCK_EXCL);
/*
* If we are using project inheritance, we only allow renames
@@ -2860,20 +2988,16 @@ xfs_rename(
if (unlikely((target_dp->i_d.di_flags & XFS_DIFLAG_PROJINHERIT) &&
(xfs_get_projid(target_dp) != xfs_get_projid(src_ip)))) {
error = -EXDEV;
- goto error_return;
+ goto out_trans_cancel;
}
- /*
- * Handle RENAME_EXCHANGE flags
- */
- if (flags & RENAME_EXCHANGE) {
- error = xfs_cross_rename(tp, src_dp, src_name, src_ip,
- target_dp, target_name, target_ip,
- &free_list, &first_block, spaceres);
- if (error)
- goto abort_return;
- goto finish_rename;
- }
+ xfs_bmap_init(&free_list, &first_block);
+
+ /* RENAME_EXCHANGE is unique from here on. */
+ if (flags & RENAME_EXCHANGE)
+ return xfs_cross_rename(tp, src_dp, src_name, src_ip,
+ target_dp, target_name, target_ip,
+ &free_list, &first_block, spaceres);
/*
* Set up the target.
@@ -2886,7 +3010,7 @@ xfs_rename(
if (!spaceres) {
error = xfs_dir_canenter(tp, target_dp, target_name);
if (error)
- goto error_return;
+ goto out_trans_cancel;
}
/*
* If target does not exist and the rename crosses
@@ -2897,9 +3021,9 @@ xfs_rename(
src_ip->i_ino, &first_block,
&free_list, spaceres);
if (error == -ENOSPC)
- goto error_return;
+ goto out_bmap_cancel;
if (error)
- goto abort_return;
+ goto out_trans_abort;
xfs_trans_ichgtime(tp, target_dp,
XFS_ICHGTIME_MOD | XFS_ICHGTIME_CHG);
@@ -2907,7 +3031,7 @@ xfs_rename(
if (new_parent && src_is_directory) {
error = xfs_bumplink(tp, target_dp);
if (error)
- goto abort_return;
+ goto out_trans_abort;
}
} else { /* target_ip != NULL */
/*
@@ -2922,7 +3046,7 @@ xfs_rename(
if (!(xfs_dir_isempty(target_ip)) ||
(target_ip->i_d.di_nlink > 2)) {
error = -EEXIST;
- goto error_return;
+ goto out_trans_cancel;
}
}
@@ -2939,7 +3063,7 @@ xfs_rename(
src_ip->i_ino,
&first_block, &free_list, spaceres);
if (error)
- goto abort_return;
+ goto out_trans_abort;
xfs_trans_ichgtime(tp, target_dp,
XFS_ICHGTIME_MOD | XFS_ICHGTIME_CHG);
@@ -2950,7 +3074,7 @@ xfs_rename(
*/
error = xfs_droplink(tp, target_ip);
if (error)
- goto abort_return;
+ goto out_trans_abort;
if (src_is_directory) {
/*
@@ -2958,7 +3082,7 @@ xfs_rename(
*/
error = xfs_droplink(tp, target_ip);
if (error)
- goto abort_return;
+ goto out_trans_abort;
}
} /* target_ip != NULL */
@@ -2975,7 +3099,7 @@ xfs_rename(
&first_block, &free_list, spaceres);
ASSERT(error != -EEXIST);
if (error)
- goto abort_return;
+ goto out_trans_abort;
}
/*
@@ -3001,49 +3125,67 @@ xfs_rename(
*/
error = xfs_droplink(tp, src_dp);
if (error)
- goto abort_return;
+ goto out_trans_abort;
}
- error = xfs_dir_removename(tp, src_dp, src_name, src_ip->i_ino,
+ /*
+ * For whiteouts, we only need to update the source dirent with the
+ * inode number of the whiteout inode rather than removing it
+ * altogether.
+ */
+ if (wip) {
+ error = xfs_dir_replace(tp, src_dp, src_name, wip->i_ino,
&first_block, &free_list, spaceres);
+ } else
+ error = xfs_dir_removename(tp, src_dp, src_name, src_ip->i_ino,
+ &first_block, &free_list, spaceres);
if (error)
- goto abort_return;
-
- xfs_trans_ichgtime(tp, src_dp, XFS_ICHGTIME_MOD | XFS_ICHGTIME_CHG);
- xfs_trans_log_inode(tp, src_dp, XFS_ILOG_CORE);
- if (new_parent)
- xfs_trans_log_inode(tp, target_dp, XFS_ILOG_CORE);
+ goto out_trans_abort;
-finish_rename:
/*
- * If this is a synchronous mount, make sure that the
- * rename transaction goes to disk before returning to
- * the user.
+ * For whiteouts, we need to bump the link count on the whiteout inode.
+ * This means that failures all the way up to this point leave the inode
+ * on the unlinked list and so cleanup is a simple matter of dropping
+ * the remaining reference to it. If we fail here after bumping the link
+ * count, we're shutting down the filesystem so we'll never see the
+ * intermediate state on disk.
*/
- if (mp->m_flags & (XFS_MOUNT_WSYNC|XFS_MOUNT_DIRSYNC)) {
- xfs_trans_set_sync(tp);
- }
+ if (wip) {
+ ASSERT(wip->i_d.di_nlink == 0);
+ error = xfs_bumplink(tp, wip);
+ if (error)
+ goto out_trans_abort;
+ error = xfs_iunlink_remove(tp, wip);
+ if (error)
+ goto out_trans_abort;
+ xfs_trans_log_inode(tp, wip, XFS_ILOG_CORE);
- error = xfs_bmap_finish(&tp, &free_list, &committed);
- if (error) {
- xfs_bmap_cancel(&free_list);
- xfs_trans_cancel(tp, (XFS_TRANS_RELEASE_LOG_RES |
- XFS_TRANS_ABORT));
- goto std_return;
+ /*
+ * Now we have a real link, clear the "I'm a tmpfile" state
+ * flag from the inode so it doesn't accidentally get misused in
+ * future.
+ */
+ VFS_I(wip)->i_state &= ~I_LINKABLE;
}
- /*
- * trans_commit will unlock src_ip, target_ip & decrement
- * the vnode references.
- */
- return xfs_trans_commit(tp, XFS_TRANS_RELEASE_LOG_RES);
+ xfs_trans_ichgtime(tp, src_dp, XFS_ICHGTIME_MOD | XFS_ICHGTIME_CHG);
+ xfs_trans_log_inode(tp, src_dp, XFS_ILOG_CORE);
+ if (new_parent)
+ xfs_trans_log_inode(tp, target_dp, XFS_ILOG_CORE);
- abort_return:
+ error = xfs_finish_rename(tp, &free_list);
+ if (wip)
+ IRELE(wip);
+ return error;
+
+out_trans_abort:
cancel_flags |= XFS_TRANS_ABORT;
- error_return:
+out_bmap_cancel:
xfs_bmap_cancel(&free_list);
+out_trans_cancel:
xfs_trans_cancel(tp, cancel_flags);
- std_return:
+ if (wip)
+ IRELE(wip);
return error;
}
diff --git a/fs/xfs/xfs_inode.h b/fs/xfs/xfs_inode.h
index 86cd6b39bed7..8f22d20368d8 100644
--- a/fs/xfs/xfs_inode.h
+++ b/fs/xfs/xfs_inode.h
@@ -56,6 +56,7 @@ typedef struct xfs_inode {
struct xfs_inode_log_item *i_itemp; /* logging information */
mrlock_t i_lock; /* inode lock */
mrlock_t i_iolock; /* inode IO lock */
+ mrlock_t i_mmaplock; /* inode mmap IO lock */
atomic_t i_pincount; /* inode pin count */
spinlock_t i_flags_lock; /* inode i_flags lock */
/* Miscellaneous state. */
@@ -263,15 +264,20 @@ static inline int xfs_isiflocked(struct xfs_inode *ip)
#define XFS_IOLOCK_SHARED (1<<1)
#define XFS_ILOCK_EXCL (1<<2)
#define XFS_ILOCK_SHARED (1<<3)
+#define XFS_MMAPLOCK_EXCL (1<<4)
+#define XFS_MMAPLOCK_SHARED (1<<5)
#define XFS_LOCK_MASK (XFS_IOLOCK_EXCL | XFS_IOLOCK_SHARED \
- | XFS_ILOCK_EXCL | XFS_ILOCK_SHARED)
+ | XFS_ILOCK_EXCL | XFS_ILOCK_SHARED \
+ | XFS_MMAPLOCK_EXCL | XFS_MMAPLOCK_SHARED)
#define XFS_LOCK_FLAGS \
{ XFS_IOLOCK_EXCL, "IOLOCK_EXCL" }, \
{ XFS_IOLOCK_SHARED, "IOLOCK_SHARED" }, \
{ XFS_ILOCK_EXCL, "ILOCK_EXCL" }, \
- { XFS_ILOCK_SHARED, "ILOCK_SHARED" }
+ { XFS_ILOCK_SHARED, "ILOCK_SHARED" }, \
+ { XFS_MMAPLOCK_EXCL, "MMAPLOCK_EXCL" }, \
+ { XFS_MMAPLOCK_SHARED, "MMAPLOCK_SHARED" }
/*
@@ -302,17 +308,26 @@ static inline int xfs_isiflocked(struct xfs_inode *ip)
#define XFS_IOLOCK_SHIFT 16
#define XFS_IOLOCK_PARENT (XFS_LOCK_PARENT << XFS_IOLOCK_SHIFT)
+#define XFS_MMAPLOCK_SHIFT 20
+
#define XFS_ILOCK_SHIFT 24
#define XFS_ILOCK_PARENT (XFS_LOCK_PARENT << XFS_ILOCK_SHIFT)
#define XFS_ILOCK_RTBITMAP (XFS_LOCK_RTBITMAP << XFS_ILOCK_SHIFT)
#define XFS_ILOCK_RTSUM (XFS_LOCK_RTSUM << XFS_ILOCK_SHIFT)
-#define XFS_IOLOCK_DEP_MASK 0x00ff0000
+#define XFS_IOLOCK_DEP_MASK 0x000f0000
+#define XFS_MMAPLOCK_DEP_MASK 0x00f00000
#define XFS_ILOCK_DEP_MASK 0xff000000
-#define XFS_LOCK_DEP_MASK (XFS_IOLOCK_DEP_MASK | XFS_ILOCK_DEP_MASK)
+#define XFS_LOCK_DEP_MASK (XFS_IOLOCK_DEP_MASK | \
+ XFS_MMAPLOCK_DEP_MASK | \
+ XFS_ILOCK_DEP_MASK)
-#define XFS_IOLOCK_DEP(flags) (((flags) & XFS_IOLOCK_DEP_MASK) >> XFS_IOLOCK_SHIFT)
-#define XFS_ILOCK_DEP(flags) (((flags) & XFS_ILOCK_DEP_MASK) >> XFS_ILOCK_SHIFT)
+#define XFS_IOLOCK_DEP(flags) (((flags) & XFS_IOLOCK_DEP_MASK) \
+ >> XFS_IOLOCK_SHIFT)
+#define XFS_MMAPLOCK_DEP(flags) (((flags) & XFS_MMAPLOCK_DEP_MASK) \
+ >> XFS_MMAPLOCK_SHIFT)
+#define XFS_ILOCK_DEP(flags) (((flags) & XFS_ILOCK_DEP_MASK) \
+ >> XFS_ILOCK_SHIFT)
/*
* For multiple groups support: if S_ISGID bit is set in the parent
@@ -384,11 +399,34 @@ enum xfs_prealloc_flags {
XFS_PREALLOC_INVISIBLE = (1 << 4),
};
-int xfs_update_prealloc_flags(struct xfs_inode *,
- enum xfs_prealloc_flags);
-int xfs_zero_eof(struct xfs_inode *, xfs_off_t, xfs_fsize_t);
-int xfs_iozero(struct xfs_inode *, loff_t, size_t);
+int xfs_update_prealloc_flags(struct xfs_inode *ip,
+ enum xfs_prealloc_flags flags);
+int xfs_zero_eof(struct xfs_inode *ip, xfs_off_t offset,
+ xfs_fsize_t isize, bool *did_zeroing);
+int xfs_iozero(struct xfs_inode *ip, loff_t pos, size_t count);
+
+/* from xfs_iops.c */
+/*
+ * When setting up a newly allocated inode, we need to call
+ * xfs_finish_inode_setup() once the inode is fully instantiated at
+ * the VFS level to prevent the rest of the world seeing the inode
+ * before we've completed instantiation. Otherwise we can do it
+ * the moment the inode lookup is complete.
+ */
+extern void xfs_setup_inode(struct xfs_inode *ip);
+static inline void xfs_finish_inode_setup(struct xfs_inode *ip)
+{
+ xfs_iflags_clear(ip, XFS_INEW);
+ barrier();
+ unlock_new_inode(VFS_I(ip));
+}
+
+static inline void xfs_setup_existing_inode(struct xfs_inode *ip)
+{
+ xfs_setup_inode(ip);
+ xfs_finish_inode_setup(ip);
+}
#define IHOLD(ip) \
do { \
diff --git a/fs/xfs/xfs_ioctl.c b/fs/xfs/xfs_ioctl.c
index ac4feae45eb3..4ee44ddfdfb7 100644
--- a/fs/xfs/xfs_ioctl.c
+++ b/fs/xfs/xfs_ioctl.c
@@ -631,7 +631,7 @@ xfs_ioc_space(
if (filp->f_flags & O_DSYNC)
flags |= XFS_PREALLOC_SYNC;
- if (ioflags & XFS_IO_INVIS)
+ if (ioflags & XFS_IO_INVIS)
flags |= XFS_PREALLOC_INVISIBLE;
error = mnt_want_write_file(filp);
@@ -643,6 +643,9 @@ xfs_ioc_space(
if (error)
goto out_unlock;
+ xfs_ilock(ip, XFS_MMAPLOCK_EXCL);
+ iolock |= XFS_MMAPLOCK_EXCL;
+
switch (bf->l_whence) {
case 0: /*SEEK_SET*/
break;
diff --git a/fs/xfs/xfs_iops.c b/fs/xfs/xfs_iops.c
index d919ad7b16bf..015d6a366b16 100644
--- a/fs/xfs/xfs_iops.c
+++ b/fs/xfs/xfs_iops.c
@@ -187,6 +187,8 @@ xfs_generic_create(
else
d_instantiate(dentry, inode);
+ xfs_finish_inode_setup(ip);
+
out_free_acl:
if (default_acl)
posix_acl_release(default_acl);
@@ -195,6 +197,7 @@ xfs_generic_create(
return error;
out_cleanup_inode:
+ xfs_finish_inode_setup(ip);
if (!tmpfile)
xfs_cleanup_inode(dir, inode, dentry);
iput(inode);
@@ -367,9 +370,11 @@ xfs_vn_symlink(
goto out_cleanup_inode;
d_instantiate(dentry, inode);
+ xfs_finish_inode_setup(cip);
return 0;
out_cleanup_inode:
+ xfs_finish_inode_setup(cip);
xfs_cleanup_inode(dir, inode, dentry);
iput(inode);
out:
@@ -389,7 +394,7 @@ xfs_vn_rename(
struct xfs_name oname;
struct xfs_name nname;
- if (flags & ~(RENAME_NOREPLACE | RENAME_EXCHANGE))
+ if (flags & ~(RENAME_NOREPLACE | RENAME_EXCHANGE | RENAME_WHITEOUT))
return -EINVAL;
/* if we are exchanging files, we need to set i_mode of both files */
@@ -751,6 +756,7 @@ xfs_setattr_size(
int error;
uint lock_flags = 0;
uint commit_flags = 0;
+ bool did_zeroing = false;
trace_xfs_setattr(ip);
@@ -765,6 +771,7 @@ xfs_setattr_size(
return error;
ASSERT(xfs_isilocked(ip, XFS_IOLOCK_EXCL));
+ ASSERT(xfs_isilocked(ip, XFS_MMAPLOCK_EXCL));
ASSERT(S_ISREG(ip->i_d.di_mode));
ASSERT((iattr->ia_valid & (ATTR_UID|ATTR_GID|ATTR_ATIME|ATTR_ATIME_SET|
ATTR_MTIME_SET|ATTR_KILL_PRIV|ATTR_TIMES_SET)) == 0);
@@ -794,20 +801,16 @@ xfs_setattr_size(
return error;
/*
- * Now we can make the changes. Before we join the inode to the
- * transaction, take care of the part of the truncation that must be
- * done without the inode lock. This needs to be done before joining
- * the inode to the transaction, because the inode cannot be unlocked
- * once it is a part of the transaction.
+ * File data changes must be complete before we start the transaction to
+ * modify the inode. This needs to be done before joining the inode to
+ * the transaction because the inode cannot be unlocked once it is a
+ * part of the transaction.
+ *
+ * Start with zeroing any data block beyond EOF that we may expose on
+ * file extension.
*/
if (newsize > oldsize) {
- /*
- * Do the first part of growing a file: zero any data in the
- * last block that is beyond the old EOF. We need to do this
- * before the inode is joined to the transaction to modify
- * i_size.
- */
- error = xfs_zero_eof(ip, newsize, oldsize);
+ error = xfs_zero_eof(ip, newsize, oldsize, &did_zeroing);
if (error)
return error;
}
@@ -817,75 +820,42 @@ xfs_setattr_size(
* any previous writes that are beyond the on disk EOF and the new
* EOF that have not been written out need to be written here. If we
* do not write the data out, we expose ourselves to the null files
- * problem.
- *
- * Only flush from the on disk size to the smaller of the in memory
- * file size or the new size as that's the range we really care about
- * here and prevents waiting for other data not within the range we
- * care about here.
+ * problem. Note that this includes any block zeroing we did above;
+ * otherwise those blocks may not be zeroed after a crash.
*/
- if (oldsize != ip->i_d.di_size && newsize > ip->i_d.di_size) {
+ if (newsize > ip->i_d.di_size &&
+ (oldsize != ip->i_d.di_size || did_zeroing)) {
error = filemap_write_and_wait_range(VFS_I(ip)->i_mapping,
ip->i_d.di_size, newsize);
if (error)
return error;
}
- /*
- * Wait for all direct I/O to complete.
- */
+ /* Now wait for all direct I/O to complete. */
inode_dio_wait(inode);
/*
- * Do all the page cache truncate work outside the transaction context
- * as the "lock" order is page lock->log space reservation. i.e.
- * locking pages inside the transaction can ABBA deadlock with
- * writeback. We have to do the VFS inode size update before we truncate
- * the pagecache, however, to avoid racing with page faults beyond the
- * new EOF they are not serialised against truncate operations except by
- * page locks and size updates.
+ * We've already locked out new page faults, so now we can safely remove
+ * pages from the page cache knowing they won't get refaulted until we
+ * drop the XFS_MMAP_EXCL lock after the extent manipulations are
+ * complete. The truncate_setsize() call also cleans partial EOF page
+ * PTEs on extending truncates and hence ensures sub-page block size
+ * filesystems are correctly handled, too.
*
- * Hence we are in a situation where a truncate can fail with ENOMEM
- * from xfs_trans_reserve(), but having already truncated the in-memory
- * version of the file (i.e. made user visible changes). There's not
- * much we can do about this, except to hope that the caller sees ENOMEM
- * and retries the truncate operation.
+ * We have to do all the page cache truncate work outside the
+ * transaction context as the "lock" order is page lock->log space
+ * reservation as defined by extent allocation in the writeback path.
+ * Hence a truncate can fail with ENOMEM from xfs_trans_reserve(), but
+ * having already truncated the in-memory version of the file (i.e. made
+ * user visible changes). There's not much we can do about this, except
+ * to hope that the caller sees ENOMEM and retries the truncate
+ * operation.
*/
error = block_truncate_page(inode->i_mapping, newsize, xfs_get_blocks);
if (error)
return error;
truncate_setsize(inode, newsize);
- /*
- * The "we can't serialise against page faults" pain gets worse.
- *
- * If the file is mapped then we have to clean the page at the old EOF
- * when extending the file. Extending the file can expose changes the
- * underlying page mapping (e.g. from beyond EOF to a hole or
- * unwritten), and so on the next attempt to write to that page we need
- * to remap it for write. i.e. we need .page_mkwrite() to be called.
- * Hence we need to clean the page to clean the pte and so a new write
- * fault will be triggered appropriately.
- *
- * If we do it before we change the inode size, then we can race with a
- * page fault that maps the page with exactly the same problem. If we do
- * it after we change the file size, then a new page fault can come in
- * and allocate space before we've run the rest of the truncate
- * transaction. That's kinda grotesque, but it's better than have data
- * over a hole, and so that's the lesser evil that has been chosen here.
- *
- * The real solution, however, is to have some mechanism for locking out
- * page faults while a truncate is in progress.
- */
- if (newsize > oldsize && mapping_mapped(VFS_I(ip)->i_mapping)) {
- error = filemap_write_and_wait_range(
- VFS_I(ip)->i_mapping,
- round_down(oldsize, PAGE_CACHE_SIZE),
- round_up(oldsize, PAGE_CACHE_SIZE) - 1);
- if (error)
- return error;
- }
-
tp = xfs_trans_alloc(mp, XFS_TRANS_SETATTR_SIZE);
error = xfs_trans_reserve(tp, &M_RES(mp)->tr_itruncate, 0, 0);
if (error)
@@ -984,8 +954,12 @@ xfs_vn_setattr(
xfs_ilock(ip, iolock);
error = xfs_break_layouts(dentry->d_inode, &iolock);
- if (!error)
+ if (!error) {
+ xfs_ilock(ip, XFS_MMAPLOCK_EXCL);
+ iolock |= XFS_MMAPLOCK_EXCL;
+
error = xfs_setattr_size(ip, iattr);
+ }
xfs_iunlock(ip, iolock);
} else {
error = xfs_setattr_nonsize(ip, iattr, 0);
@@ -1236,16 +1210,12 @@ xfs_diflags_to_iflags(
}
/*
- * Initialize the Linux inode, set up the operation vectors and
- * unlock the inode.
- *
- * When reading existing inodes from disk this is called directly
- * from xfs_iget, when creating a new inode it is called from
- * xfs_ialloc after setting up the inode.
+ * Initialize the Linux inode and set up the operation vectors.
*
- * We are always called with an uninitialised linux inode here.
- * We need to initialise the necessary fields and take a reference
- * on it.
+ * When reading existing inodes from disk this is called directly from xfs_iget,
+ * when creating a new inode it is called from xfs_ialloc after setting up the
+ * inode. These callers have different criteria for clearing XFS_INEW, so leave
+ * it up to the caller to deal with unlocking the inode appropriately.
*/
void
xfs_setup_inode(
@@ -1332,9 +1302,4 @@ xfs_setup_inode(
inode_has_no_xattr(inode);
cache_no_acl(inode);
}
-
- xfs_iflags_clear(ip, XFS_INEW);
- barrier();
-
- unlock_new_inode(inode);
}
diff --git a/fs/xfs/xfs_iops.h b/fs/xfs/xfs_iops.h
index ea7a98e9cb70..a0f84abb0d09 100644
--- a/fs/xfs/xfs_iops.h
+++ b/fs/xfs/xfs_iops.h
@@ -25,8 +25,6 @@ extern const struct file_operations xfs_dir_file_operations;
extern ssize_t xfs_vn_listxattr(struct dentry *, char *data, size_t size);
-extern void xfs_setup_inode(struct xfs_inode *);
-
/*
* Internal setattr interfaces.
*/
diff --git a/fs/xfs/xfs_itable.c b/fs/xfs/xfs_itable.c
index 82e314258f73..80429891dc9b 100644
--- a/fs/xfs/xfs_itable.c
+++ b/fs/xfs/xfs_itable.c
@@ -229,7 +229,7 @@ xfs_bulkstat_grab_ichunk(
error = xfs_inobt_get_rec(cur, irec, &stat);
if (error)
return error;
- XFS_WANT_CORRUPTED_RETURN(stat == 1);
+ XFS_WANT_CORRUPTED_RETURN(cur->bc_mp, stat == 1);
/* Check if the record contains the inode in request */
if (irec->ir_startino + XFS_INODES_PER_CHUNK <= agino) {
diff --git a/fs/xfs/xfs_pnfs.c b/fs/xfs/xfs_pnfs.c
index 4b33ef112400..365dd57ea760 100644
--- a/fs/xfs/xfs_pnfs.c
+++ b/fs/xfs/xfs_pnfs.c
@@ -300,8 +300,10 @@ xfs_fs_commit_blocks(
tp = xfs_trans_alloc(mp, XFS_TRANS_SETATTR_NOT_SIZE);
error = xfs_trans_reserve(tp, &M_RES(mp)->tr_ichange, 0, 0);
- if (error)
+ if (error) {
+ xfs_trans_cancel(tp, 0);
goto out_drop_iolock;
+ }
xfs_ilock(ip, XFS_ILOCK_EXCL);
xfs_trans_ijoin(tp, ip, XFS_ILOCK_EXCL);
diff --git a/fs/xfs/xfs_qm.c b/fs/xfs/xfs_qm.c
index 53cc2aaf8d2b..5538468c7f63 100644
--- a/fs/xfs/xfs_qm.c
+++ b/fs/xfs/xfs_qm.c
@@ -719,6 +719,7 @@ xfs_qm_qino_alloc(
xfs_trans_t *tp;
int error;
int committed;
+ bool need_alloc = true;
*ip = NULL;
/*
@@ -747,6 +748,7 @@ xfs_qm_qino_alloc(
return error;
mp->m_sb.sb_gquotino = NULLFSINO;
mp->m_sb.sb_pquotino = NULLFSINO;
+ need_alloc = false;
}
}
@@ -758,7 +760,7 @@ xfs_qm_qino_alloc(
return error;
}
- if (!*ip) {
+ if (need_alloc) {
error = xfs_dir_ialloc(&tp, NULL, S_IFREG, 1, 0, 0, 1, ip,
&committed);
if (error) {
@@ -794,11 +796,14 @@ xfs_qm_qino_alloc(
spin_unlock(&mp->m_sb_lock);
xfs_log_sb(tp);
- if ((error = xfs_trans_commit(tp, XFS_TRANS_RELEASE_LOG_RES))) {
+ error = xfs_trans_commit(tp, XFS_TRANS_RELEASE_LOG_RES);
+ if (error) {
+ ASSERT(XFS_FORCED_SHUTDOWN(mp));
xfs_alert(mp, "%s failed (error %d)!", __func__, error);
- return error;
}
- return 0;
+ if (need_alloc)
+ xfs_finish_inode_setup(*ip);
+ return error;
}
@@ -836,6 +841,11 @@ xfs_qm_reset_dqcounts(
*/
xfs_dqcheck(mp, ddq, id+j, type, XFS_QMOPT_DQREPAIR,
"xfs_quotacheck");
+ /*
+ * Reset type in case we are reusing group quota file for
+ * project quotas or vice versa
+ */
+ ddq->d_flags = type;
ddq->d_bcount = 0;
ddq->d_icount = 0;
ddq->d_rtbcount = 0;
diff --git a/fs/xfs/xfs_super.c b/fs/xfs/xfs_super.c
index 194291381252..8782b36517b5 100644
--- a/fs/xfs/xfs_super.c
+++ b/fs/xfs/xfs_super.c
@@ -109,8 +109,6 @@ static struct xfs_kobj xfs_dbg_kobj; /* global debug sysfs attrs */
#define MNTOPT_GQUOTANOENF "gqnoenforce"/* group quota limit enforcement */
#define MNTOPT_PQUOTANOENF "pqnoenforce"/* project quota limit enforcement */
#define MNTOPT_QUOTANOENF "qnoenforce" /* same as uqnoenforce */
-#define MNTOPT_DELAYLOG "delaylog" /* Delayed logging enabled */
-#define MNTOPT_NODELAYLOG "nodelaylog" /* Delayed logging disabled */
#define MNTOPT_DISCARD "discard" /* Discard unused blocks */
#define MNTOPT_NODISCARD "nodiscard" /* Do not discard unused blocks */
@@ -361,28 +359,10 @@ xfs_parseargs(
} else if (!strcmp(this_char, MNTOPT_GQUOTANOENF)) {
mp->m_qflags |= (XFS_GQUOTA_ACCT | XFS_GQUOTA_ACTIVE);
mp->m_qflags &= ~XFS_GQUOTA_ENFD;
- } else if (!strcmp(this_char, MNTOPT_DELAYLOG)) {
- xfs_warn(mp,
- "delaylog is the default now, option is deprecated.");
- } else if (!strcmp(this_char, MNTOPT_NODELAYLOG)) {
- xfs_warn(mp,
- "nodelaylog support has been removed, option is deprecated.");
} else if (!strcmp(this_char, MNTOPT_DISCARD)) {
mp->m_flags |= XFS_MOUNT_DISCARD;
} else if (!strcmp(this_char, MNTOPT_NODISCARD)) {
mp->m_flags &= ~XFS_MOUNT_DISCARD;
- } else if (!strcmp(this_char, "ihashsize")) {
- xfs_warn(mp,
- "ihashsize no longer used, option is deprecated.");
- } else if (!strcmp(this_char, "osyncisdsync")) {
- xfs_warn(mp,
- "osyncisdsync has no effect, option is deprecated.");
- } else if (!strcmp(this_char, "osyncisosync")) {
- xfs_warn(mp,
- "osyncisosync has no effect, option is deprecated.");
- } else if (!strcmp(this_char, "irixsgid")) {
- xfs_warn(mp,
- "irixsgid is now a sysctl(2) variable, option is deprecated.");
} else {
xfs_warn(mp, "unknown mount option [%s].", this_char);
return -EINVAL;
@@ -986,6 +966,8 @@ xfs_fs_inode_init_once(
atomic_set(&ip->i_pincount, 0);
spin_lock_init(&ip->i_flags_lock);
+ mrlock_init(&ip->i_mmaplock, MRLOCK_ALLOW_EQUAL_PRI|MRLOCK_BARRIER,
+ "xfsino", ip->i_ino);
mrlock_init(&ip->i_lock, MRLOCK_ALLOW_EQUAL_PRI|MRLOCK_BARRIER,
"xfsino", ip->i_ino);
}
@@ -1564,6 +1546,7 @@ xfs_fs_put_super(
{
struct xfs_mount *mp = XFS_M(sb);
+ xfs_notice(mp, "Unmounting Filesystem");
xfs_filestream_unmount(mp);
xfs_unmountfs(mp);
diff --git a/fs/xfs/xfs_symlink.c b/fs/xfs/xfs_symlink.c
index 25791df6f638..3df411eadb86 100644
--- a/fs/xfs/xfs_symlink.c
+++ b/fs/xfs/xfs_symlink.c
@@ -177,7 +177,7 @@ xfs_symlink(
int pathlen;
struct xfs_bmap_free free_list;
xfs_fsblock_t first_block;
- bool unlock_dp_on_error = false;
+ bool unlock_dp_on_error = false;
uint cancel_flags;
int committed;
xfs_fileoff_t first_fsb;
@@ -221,7 +221,7 @@ xfs_symlink(
XFS_QMOPT_QUOTALL | XFS_QMOPT_INHERIT,
&udqp, &gdqp, &pdqp);
if (error)
- goto std_return;
+ return error;
tp = xfs_trans_alloc(mp, XFS_TRANS_SYMLINK);
cancel_flags = XFS_TRANS_RELEASE_LOG_RES;
@@ -241,7 +241,7 @@ xfs_symlink(
}
if (error) {
cancel_flags = 0;
- goto error_return;
+ goto out_trans_cancel;
}
xfs_ilock(dp, XFS_ILOCK_EXCL | XFS_ILOCK_PARENT);
@@ -252,7 +252,7 @@ xfs_symlink(
*/
if (dp->i_d.di_flags & XFS_DIFLAG_NOSYMLINKS) {
error = -EPERM;
- goto error_return;
+ goto out_trans_cancel;
}
/*
@@ -261,7 +261,7 @@ xfs_symlink(
error = xfs_trans_reserve_quota(tp, mp, udqp, gdqp,
pdqp, resblks, 1, 0);
if (error)
- goto error_return;
+ goto out_trans_cancel;
/*
* Check for ability to enter directory entry, if no space reserved.
@@ -269,7 +269,7 @@ xfs_symlink(
if (!resblks) {
error = xfs_dir_canenter(tp, dp, link_name);
if (error)
- goto error_return;
+ goto out_trans_cancel;
}
/*
* Initialize the bmap freelist prior to calling either
@@ -282,15 +282,14 @@ xfs_symlink(
*/
error = xfs_dir_ialloc(&tp, dp, S_IFLNK | (mode & ~S_IFMT), 1, 0,
prid, resblks > 0, &ip, NULL);
- if (error) {
- if (error == -ENOSPC)
- goto error_return;
- goto error1;
- }
+ if (error)
+ goto out_trans_cancel;
/*
- * An error after we've joined dp to the transaction will result in the
- * transaction cancel unlocking dp so don't do it explicitly in the
+ * Now we join the directory inode to the transaction. We do not do it
+ * earlier because xfs_dir_ialloc might commit the previous transaction
+ * (and release all the locks). An error from here on will result in
+ * the transaction cancel unlocking dp so don't do it explicitly in the
* error path.
*/
xfs_trans_ijoin(tp, dp, XFS_ILOCK_EXCL);
@@ -330,7 +329,7 @@ xfs_symlink(
XFS_BMAPI_METADATA, &first_block, resblks,
mval, &nmaps, &free_list);
if (error)
- goto error2;
+ goto out_bmap_cancel;
if (resblks)
resblks -= fs_blocks;
@@ -348,7 +347,7 @@ xfs_symlink(
BTOBB(byte_cnt), 0);
if (!bp) {
error = -ENOMEM;
- goto error2;
+ goto out_bmap_cancel;
}
bp->b_ops = &xfs_symlink_buf_ops;
@@ -378,7 +377,7 @@ xfs_symlink(
error = xfs_dir_createname(tp, dp, link_name, ip->i_ino,
&first_block, &free_list, resblks);
if (error)
- goto error2;
+ goto out_bmap_cancel;
xfs_trans_ichgtime(tp, dp, XFS_ICHGTIME_MOD | XFS_ICHGTIME_CHG);
xfs_trans_log_inode(tp, dp, XFS_ILOG_CORE);
@@ -392,10 +391,13 @@ xfs_symlink(
}
error = xfs_bmap_finish(&tp, &free_list, &committed);
- if (error) {
- goto error2;
- }
+ if (error)
+ goto out_bmap_cancel;
+
error = xfs_trans_commit(tp, XFS_TRANS_RELEASE_LOG_RES);
+ if (error)
+ goto out_release_inode;
+
xfs_qm_dqrele(udqp);
xfs_qm_dqrele(gdqp);
xfs_qm_dqrele(pdqp);
@@ -403,20 +405,28 @@ xfs_symlink(
*ipp = ip;
return 0;
- error2:
- IRELE(ip);
- error1:
+out_bmap_cancel:
xfs_bmap_cancel(&free_list);
cancel_flags |= XFS_TRANS_ABORT;
- error_return:
+out_trans_cancel:
xfs_trans_cancel(tp, cancel_flags);
+out_release_inode:
+ /*
+ * Wait until after the current transaction is aborted to finish the
+ * setup of the inode and release the inode. This prevents recursive
+ * transactions and deadlocks from xfs_inactive.
+ */
+ if (ip) {
+ xfs_finish_inode_setup(ip);
+ IRELE(ip);
+ }
+
xfs_qm_dqrele(udqp);
xfs_qm_dqrele(gdqp);
xfs_qm_dqrele(pdqp);
if (unlock_dp_on_error)
xfs_iunlock(dp, XFS_ILOCK_EXCL);
- std_return:
return error;
}
diff --git a/fs/xfs/xfs_trace.h b/fs/xfs/xfs_trace.h
index b5ac81eeb061..a302738d224d 100644
--- a/fs/xfs/xfs_trace.h
+++ b/fs/xfs/xfs_trace.h
@@ -685,6 +685,9 @@ DEFINE_INODE_EVENT(xfs_inode_set_eofblocks_tag);
DEFINE_INODE_EVENT(xfs_inode_clear_eofblocks_tag);
DEFINE_INODE_EVENT(xfs_inode_free_eofblocks_invalid);
+DEFINE_INODE_EVENT(xfs_filemap_fault);
+DEFINE_INODE_EVENT(xfs_filemap_page_mkwrite);
+
DECLARE_EVENT_CLASS(xfs_iref_class,
TP_PROTO(struct xfs_inode *ip, unsigned long caller_ip),
TP_ARGS(ip, caller_ip),