aboutsummaryrefslogtreecommitdiffstats
path: root/fs
diff options
context:
space:
mode:
Diffstat (limited to 'fs')
-rw-r--r--fs/btrfs/file-item.c2
-rw-r--r--fs/ceph/caps.c2
-rw-r--r--fs/nfs/blocklayout/blocklayout.c2
-rw-r--r--fs/nfs/blocklayout/rpc_pipefs.c14
-rw-r--r--fs/nfs/delegation.c25
-rw-r--r--fs/nfs/delegation.h1
-rw-r--r--fs/nfs/dir.c1
-rw-r--r--fs/nfs/direct.c1
-rw-r--r--fs/nfs/filelayout/filelayout.c3
-rw-r--r--fs/nfs/inode.c2
-rw-r--r--fs/nfs/netns.h1
-rw-r--r--fs/nfs/nfs4proc.c95
-rw-r--r--fs/nfs/write.c2
-rw-r--r--fs/notify/fsnotify.c36
-rw-r--r--fs/notify/fsnotify.h4
-rw-r--r--fs/notify/inode_mark.c8
-rw-r--r--fs/notify/mark.c36
-rw-r--r--fs/notify/vfsmount_mark.c8
-rw-r--r--fs/ocfs2/cluster/tcp.c2
-rw-r--r--fs/overlayfs/readdir.c2
-rw-r--r--fs/xfs/xfs_bmap_util.c72
-rw-r--r--fs/xfs/xfs_itable.c250
-rw-r--r--fs/xfs/xfs_itable.h16
23 files changed, 296 insertions, 289 deletions
diff --git a/fs/btrfs/file-item.c b/fs/btrfs/file-item.c
index 783a94355efd..84a2d1868271 100644
--- a/fs/btrfs/file-item.c
+++ b/fs/btrfs/file-item.c
@@ -413,7 +413,7 @@ int btrfs_lookup_csums_range(struct btrfs_root *root, u64 start, u64 end,
413 ret = 0; 413 ret = 0;
414fail: 414fail:
415 while (ret < 0 && !list_empty(&tmplist)) { 415 while (ret < 0 && !list_empty(&tmplist)) {
416 sums = list_entry(&tmplist, struct btrfs_ordered_sum, list); 416 sums = list_entry(tmplist.next, struct btrfs_ordered_sum, list);
417 list_del(&sums->list); 417 list_del(&sums->list);
418 kfree(sums); 418 kfree(sums);
419 } 419 }
diff --git a/fs/ceph/caps.c b/fs/ceph/caps.c
index 659f2ea9e6f7..cefca661464b 100644
--- a/fs/ceph/caps.c
+++ b/fs/ceph/caps.c
@@ -2638,7 +2638,7 @@ static void handle_cap_flush_ack(struct inode *inode, u64 flush_tid,
2638 2638
2639 for (i = 0; i < CEPH_CAP_BITS; i++) 2639 for (i = 0; i < CEPH_CAP_BITS; i++)
2640 if ((dirty & (1 << i)) && 2640 if ((dirty & (1 << i)) &&
2641 flush_tid == ci->i_cap_flush_tid[i]) 2641 (u16)flush_tid == ci->i_cap_flush_tid[i])
2642 cleaned |= 1 << i; 2642 cleaned |= 1 << i;
2643 2643
2644 dout("handle_cap_flush_ack inode %p mds%d seq %d on %s cleaned %s," 2644 dout("handle_cap_flush_ack inode %p mds%d seq %d on %s cleaned %s,"
diff --git a/fs/nfs/blocklayout/blocklayout.c b/fs/nfs/blocklayout/blocklayout.c
index 5228f201d3d5..4f46f7a05289 100644
--- a/fs/nfs/blocklayout/blocklayout.c
+++ b/fs/nfs/blocklayout/blocklayout.c
@@ -378,7 +378,7 @@ bl_write_pagelist(struct nfs_pgio_header *header, int sync)
378 loff_t offset = header->args.offset; 378 loff_t offset = header->args.offset;
379 size_t count = header->args.count; 379 size_t count = header->args.count;
380 struct page **pages = header->args.pages; 380 struct page **pages = header->args.pages;
381 int pg_index = pg_index = header->args.pgbase >> PAGE_CACHE_SHIFT; 381 int pg_index = header->args.pgbase >> PAGE_CACHE_SHIFT;
382 unsigned int pg_len; 382 unsigned int pg_len;
383 struct blk_plug plug; 383 struct blk_plug plug;
384 int i; 384 int i;
diff --git a/fs/nfs/blocklayout/rpc_pipefs.c b/fs/nfs/blocklayout/rpc_pipefs.c
index e966c023b1b7..acbf9ca4018c 100644
--- a/fs/nfs/blocklayout/rpc_pipefs.c
+++ b/fs/nfs/blocklayout/rpc_pipefs.c
@@ -65,17 +65,18 @@ bl_resolve_deviceid(struct nfs_server *server, struct pnfs_block_volume *b,
65 65
66 dprintk("%s CREATING PIPEFS MESSAGE\n", __func__); 66 dprintk("%s CREATING PIPEFS MESSAGE\n", __func__);
67 67
68 mutex_lock(&nn->bl_mutex);
68 bl_pipe_msg.bl_wq = &nn->bl_wq; 69 bl_pipe_msg.bl_wq = &nn->bl_wq;
69 70
70 b->simple.len += 4; /* single volume */ 71 b->simple.len += 4; /* single volume */
71 if (b->simple.len > PAGE_SIZE) 72 if (b->simple.len > PAGE_SIZE)
72 return -EIO; 73 goto out_unlock;
73 74
74 memset(msg, 0, sizeof(*msg)); 75 memset(msg, 0, sizeof(*msg));
75 msg->len = sizeof(*bl_msg) + b->simple.len; 76 msg->len = sizeof(*bl_msg) + b->simple.len;
76 msg->data = kzalloc(msg->len, gfp_mask); 77 msg->data = kzalloc(msg->len, gfp_mask);
77 if (!msg->data) 78 if (!msg->data)
78 goto out; 79 goto out_free_data;
79 80
80 bl_msg = msg->data; 81 bl_msg = msg->data;
81 bl_msg->type = BL_DEVICE_MOUNT, 82 bl_msg->type = BL_DEVICE_MOUNT,
@@ -87,7 +88,7 @@ bl_resolve_deviceid(struct nfs_server *server, struct pnfs_block_volume *b,
87 rc = rpc_queue_upcall(nn->bl_device_pipe, msg); 88 rc = rpc_queue_upcall(nn->bl_device_pipe, msg);
88 if (rc < 0) { 89 if (rc < 0) {
89 remove_wait_queue(&nn->bl_wq, &wq); 90 remove_wait_queue(&nn->bl_wq, &wq);
90 goto out; 91 goto out_free_data;
91 } 92 }
92 93
93 set_current_state(TASK_UNINTERRUPTIBLE); 94 set_current_state(TASK_UNINTERRUPTIBLE);
@@ -97,12 +98,14 @@ bl_resolve_deviceid(struct nfs_server *server, struct pnfs_block_volume *b,
97 if (reply->status != BL_DEVICE_REQUEST_PROC) { 98 if (reply->status != BL_DEVICE_REQUEST_PROC) {
98 printk(KERN_WARNING "%s failed to decode device: %d\n", 99 printk(KERN_WARNING "%s failed to decode device: %d\n",
99 __func__, reply->status); 100 __func__, reply->status);
100 goto out; 101 goto out_free_data;
101 } 102 }
102 103
103 dev = MKDEV(reply->major, reply->minor); 104 dev = MKDEV(reply->major, reply->minor);
104out: 105out_free_data:
105 kfree(msg->data); 106 kfree(msg->data);
107out_unlock:
108 mutex_unlock(&nn->bl_mutex);
106 return dev; 109 return dev;
107} 110}
108 111
@@ -232,6 +235,7 @@ static int nfs4blocklayout_net_init(struct net *net)
232 struct nfs_net *nn = net_generic(net, nfs_net_id); 235 struct nfs_net *nn = net_generic(net, nfs_net_id);
233 struct dentry *dentry; 236 struct dentry *dentry;
234 237
238 mutex_init(&nn->bl_mutex);
235 init_waitqueue_head(&nn->bl_wq); 239 init_waitqueue_head(&nn->bl_wq);
236 nn->bl_device_pipe = rpc_mkpipe_data(&bl_upcall_ops, 0); 240 nn->bl_device_pipe = rpc_mkpipe_data(&bl_upcall_ops, 0);
237 if (IS_ERR(nn->bl_device_pipe)) 241 if (IS_ERR(nn->bl_device_pipe))
diff --git a/fs/nfs/delegation.c b/fs/nfs/delegation.c
index 5853f53db732..7f3f60641344 100644
--- a/fs/nfs/delegation.c
+++ b/fs/nfs/delegation.c
@@ -125,6 +125,8 @@ again:
125 continue; 125 continue;
126 if (!test_bit(NFS_DELEGATED_STATE, &state->flags)) 126 if (!test_bit(NFS_DELEGATED_STATE, &state->flags))
127 continue; 127 continue;
128 if (!nfs4_valid_open_stateid(state))
129 continue;
128 if (!nfs4_stateid_match(&state->stateid, stateid)) 130 if (!nfs4_stateid_match(&state->stateid, stateid))
129 continue; 131 continue;
130 get_nfs_open_context(ctx); 132 get_nfs_open_context(ctx);
@@ -193,7 +195,11 @@ static int nfs_do_return_delegation(struct inode *inode, struct nfs_delegation *
193{ 195{
194 int res = 0; 196 int res = 0;
195 197
196 res = nfs4_proc_delegreturn(inode, delegation->cred, &delegation->stateid, issync); 198 if (!test_bit(NFS_DELEGATION_REVOKED, &delegation->flags))
199 res = nfs4_proc_delegreturn(inode,
200 delegation->cred,
201 &delegation->stateid,
202 issync);
197 nfs_free_delegation(delegation); 203 nfs_free_delegation(delegation);
198 return res; 204 return res;
199} 205}
@@ -380,11 +386,13 @@ static int nfs_end_delegation_return(struct inode *inode, struct nfs_delegation
380{ 386{
381 struct nfs_client *clp = NFS_SERVER(inode)->nfs_client; 387 struct nfs_client *clp = NFS_SERVER(inode)->nfs_client;
382 struct nfs_inode *nfsi = NFS_I(inode); 388 struct nfs_inode *nfsi = NFS_I(inode);
383 int err; 389 int err = 0;
384 390
385 if (delegation == NULL) 391 if (delegation == NULL)
386 return 0; 392 return 0;
387 do { 393 do {
394 if (test_bit(NFS_DELEGATION_REVOKED, &delegation->flags))
395 break;
388 err = nfs_delegation_claim_opens(inode, &delegation->stateid); 396 err = nfs_delegation_claim_opens(inode, &delegation->stateid);
389 if (!issync || err != -EAGAIN) 397 if (!issync || err != -EAGAIN)
390 break; 398 break;
@@ -605,10 +613,23 @@ static void nfs_client_mark_return_unused_delegation_types(struct nfs_client *cl
605 rcu_read_unlock(); 613 rcu_read_unlock();
606} 614}
607 615
616static void nfs_revoke_delegation(struct inode *inode)
617{
618 struct nfs_delegation *delegation;
619 rcu_read_lock();
620 delegation = rcu_dereference(NFS_I(inode)->delegation);
621 if (delegation != NULL) {
622 set_bit(NFS_DELEGATION_REVOKED, &delegation->flags);
623 nfs_mark_return_delegation(NFS_SERVER(inode), delegation);
624 }
625 rcu_read_unlock();
626}
627
608void nfs_remove_bad_delegation(struct inode *inode) 628void nfs_remove_bad_delegation(struct inode *inode)
609{ 629{
610 struct nfs_delegation *delegation; 630 struct nfs_delegation *delegation;
611 631
632 nfs_revoke_delegation(inode);
612 delegation = nfs_inode_detach_delegation(inode); 633 delegation = nfs_inode_detach_delegation(inode);
613 if (delegation) { 634 if (delegation) {
614 nfs_inode_find_state_and_recover(inode, &delegation->stateid); 635 nfs_inode_find_state_and_recover(inode, &delegation->stateid);
diff --git a/fs/nfs/delegation.h b/fs/nfs/delegation.h
index 5c1cce39297f..e3c20a3ccc93 100644
--- a/fs/nfs/delegation.h
+++ b/fs/nfs/delegation.h
@@ -31,6 +31,7 @@ enum {
31 NFS_DELEGATION_RETURN_IF_CLOSED, 31 NFS_DELEGATION_RETURN_IF_CLOSED,
32 NFS_DELEGATION_REFERENCED, 32 NFS_DELEGATION_REFERENCED,
33 NFS_DELEGATION_RETURNING, 33 NFS_DELEGATION_RETURNING,
34 NFS_DELEGATION_REVOKED,
34}; 35};
35 36
36int nfs_inode_set_delegation(struct inode *inode, struct rpc_cred *cred, struct nfs_openres *res); 37int nfs_inode_set_delegation(struct inode *inode, struct rpc_cred *cred, struct nfs_openres *res);
diff --git a/fs/nfs/dir.c b/fs/nfs/dir.c
index 06e8cfcbb670..6e62155abf26 100644
--- a/fs/nfs/dir.c
+++ b/fs/nfs/dir.c
@@ -1527,6 +1527,7 @@ int nfs_atomic_open(struct inode *dir, struct dentry *dentry,
1527 case -ENOENT: 1527 case -ENOENT:
1528 d_drop(dentry); 1528 d_drop(dentry);
1529 d_add(dentry, NULL); 1529 d_add(dentry, NULL);
1530 nfs_set_verifier(dentry, nfs_save_change_attribute(dir));
1530 break; 1531 break;
1531 case -EISDIR: 1532 case -EISDIR:
1532 case -ENOTDIR: 1533 case -ENOTDIR:
diff --git a/fs/nfs/direct.c b/fs/nfs/direct.c
index 20cffc830468..10bf07280f4a 100644
--- a/fs/nfs/direct.c
+++ b/fs/nfs/direct.c
@@ -266,6 +266,7 @@ static void nfs_direct_req_free(struct kref *kref)
266{ 266{
267 struct nfs_direct_req *dreq = container_of(kref, struct nfs_direct_req, kref); 267 struct nfs_direct_req *dreq = container_of(kref, struct nfs_direct_req, kref);
268 268
269 nfs_free_pnfs_ds_cinfo(&dreq->ds_cinfo);
269 if (dreq->l_ctx != NULL) 270 if (dreq->l_ctx != NULL)
270 nfs_put_lock_context(dreq->l_ctx); 271 nfs_put_lock_context(dreq->l_ctx);
271 if (dreq->ctx != NULL) 272 if (dreq->ctx != NULL)
diff --git a/fs/nfs/filelayout/filelayout.c b/fs/nfs/filelayout/filelayout.c
index 46fab1cb455a..7afb52f6a25a 100644
--- a/fs/nfs/filelayout/filelayout.c
+++ b/fs/nfs/filelayout/filelayout.c
@@ -145,9 +145,6 @@ static int filelayout_async_handle_error(struct rpc_task *task,
145 case -NFS4ERR_DELEG_REVOKED: 145 case -NFS4ERR_DELEG_REVOKED:
146 case -NFS4ERR_ADMIN_REVOKED: 146 case -NFS4ERR_ADMIN_REVOKED:
147 case -NFS4ERR_BAD_STATEID: 147 case -NFS4ERR_BAD_STATEID:
148 if (state == NULL)
149 break;
150 nfs_remove_bad_delegation(state->inode);
151 case -NFS4ERR_OPENMODE: 148 case -NFS4ERR_OPENMODE:
152 if (state == NULL) 149 if (state == NULL)
153 break; 150 break;
diff --git a/fs/nfs/inode.c b/fs/nfs/inode.c
index 6388a59f2add..00689a8a85e4 100644
--- a/fs/nfs/inode.c
+++ b/fs/nfs/inode.c
@@ -626,7 +626,7 @@ int nfs_getattr(struct vfsmount *mnt, struct dentry *dentry, struct kstat *stat)
626{ 626{
627 struct inode *inode = dentry->d_inode; 627 struct inode *inode = dentry->d_inode;
628 int need_atime = NFS_I(inode)->cache_validity & NFS_INO_INVALID_ATIME; 628 int need_atime = NFS_I(inode)->cache_validity & NFS_INO_INVALID_ATIME;
629 int err; 629 int err = 0;
630 630
631 trace_nfs_getattr_enter(inode); 631 trace_nfs_getattr_enter(inode);
632 /* Flush out writes to the server in order to update c/mtime. */ 632 /* Flush out writes to the server in order to update c/mtime. */
diff --git a/fs/nfs/netns.h b/fs/nfs/netns.h
index ef221fb8a183..f0e06e4acbef 100644
--- a/fs/nfs/netns.h
+++ b/fs/nfs/netns.h
@@ -19,6 +19,7 @@ struct nfs_net {
19 struct rpc_pipe *bl_device_pipe; 19 struct rpc_pipe *bl_device_pipe;
20 struct bl_dev_msg bl_mount_reply; 20 struct bl_dev_msg bl_mount_reply;
21 wait_queue_head_t bl_wq; 21 wait_queue_head_t bl_wq;
22 struct mutex bl_mutex;
22 struct list_head nfs_client_list; 23 struct list_head nfs_client_list;
23 struct list_head nfs_volume_list; 24 struct list_head nfs_volume_list;
24#if IS_ENABLED(CONFIG_NFS_V4) 25#if IS_ENABLED(CONFIG_NFS_V4)
diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c
index 405bd95c1f58..69dc20a743f9 100644
--- a/fs/nfs/nfs4proc.c
+++ b/fs/nfs/nfs4proc.c
@@ -370,11 +370,6 @@ static int nfs4_handle_exception(struct nfs_server *server, int errorcode, struc
370 case -NFS4ERR_DELEG_REVOKED: 370 case -NFS4ERR_DELEG_REVOKED:
371 case -NFS4ERR_ADMIN_REVOKED: 371 case -NFS4ERR_ADMIN_REVOKED:
372 case -NFS4ERR_BAD_STATEID: 372 case -NFS4ERR_BAD_STATEID:
373 if (inode != NULL && nfs4_have_delegation(inode, FMODE_READ)) {
374 nfs_remove_bad_delegation(inode);
375 exception->retry = 1;
376 break;
377 }
378 if (state == NULL) 373 if (state == NULL)
379 break; 374 break;
380 ret = nfs4_schedule_stateid_recovery(server, state); 375 ret = nfs4_schedule_stateid_recovery(server, state);
@@ -1654,7 +1649,7 @@ static int nfs4_handle_delegation_recall_error(struct nfs_server *server, struct
1654 nfs_inode_find_state_and_recover(state->inode, 1649 nfs_inode_find_state_and_recover(state->inode,
1655 stateid); 1650 stateid);
1656 nfs4_schedule_stateid_recovery(server, state); 1651 nfs4_schedule_stateid_recovery(server, state);
1657 return 0; 1652 return -EAGAIN;
1658 case -NFS4ERR_DELAY: 1653 case -NFS4ERR_DELAY:
1659 case -NFS4ERR_GRACE: 1654 case -NFS4ERR_GRACE:
1660 set_bit(NFS_DELEGATED_STATE, &state->flags); 1655 set_bit(NFS_DELEGATED_STATE, &state->flags);
@@ -2109,46 +2104,60 @@ static int nfs4_open_expired(struct nfs4_state_owner *sp, struct nfs4_state *sta
2109 return ret; 2104 return ret;
2110} 2105}
2111 2106
2107static void nfs_finish_clear_delegation_stateid(struct nfs4_state *state)
2108{
2109 nfs_remove_bad_delegation(state->inode);
2110 write_seqlock(&state->seqlock);
2111 nfs4_stateid_copy(&state->stateid, &state->open_stateid);
2112 write_sequnlock(&state->seqlock);
2113 clear_bit(NFS_DELEGATED_STATE, &state->flags);
2114}
2115
2116static void nfs40_clear_delegation_stateid(struct nfs4_state *state)
2117{
2118 if (rcu_access_pointer(NFS_I(state->inode)->delegation) != NULL)
2119 nfs_finish_clear_delegation_stateid(state);
2120}
2121
2122static int nfs40_open_expired(struct nfs4_state_owner *sp, struct nfs4_state *state)
2123{
2124 /* NFSv4.0 doesn't allow for delegation recovery on open expire */
2125 nfs40_clear_delegation_stateid(state);
2126 return nfs4_open_expired(sp, state);
2127}
2128
2112#if defined(CONFIG_NFS_V4_1) 2129#if defined(CONFIG_NFS_V4_1)
2113static void nfs41_clear_delegation_stateid(struct nfs4_state *state) 2130static void nfs41_check_delegation_stateid(struct nfs4_state *state)
2114{ 2131{
2115 struct nfs_server *server = NFS_SERVER(state->inode); 2132 struct nfs_server *server = NFS_SERVER(state->inode);
2116 nfs4_stateid *stateid = &state->stateid; 2133 nfs4_stateid stateid;
2117 struct nfs_delegation *delegation; 2134 struct nfs_delegation *delegation;
2118 struct rpc_cred *cred = NULL; 2135 struct rpc_cred *cred;
2119 int status = -NFS4ERR_BAD_STATEID; 2136 int status;
2120
2121 /* If a state reset has been done, test_stateid is unneeded */
2122 if (test_bit(NFS_DELEGATED_STATE, &state->flags) == 0)
2123 return;
2124 2137
2125 /* Get the delegation credential for use by test/free_stateid */ 2138 /* Get the delegation credential for use by test/free_stateid */
2126 rcu_read_lock(); 2139 rcu_read_lock();
2127 delegation = rcu_dereference(NFS_I(state->inode)->delegation); 2140 delegation = rcu_dereference(NFS_I(state->inode)->delegation);
2128 if (delegation != NULL && 2141 if (delegation == NULL) {
2129 nfs4_stateid_match(&delegation->stateid, stateid)) {
2130 cred = get_rpccred(delegation->cred);
2131 rcu_read_unlock();
2132 status = nfs41_test_stateid(server, stateid, cred);
2133 trace_nfs4_test_delegation_stateid(state, NULL, status);
2134 } else
2135 rcu_read_unlock(); 2142 rcu_read_unlock();
2143 return;
2144 }
2145
2146 nfs4_stateid_copy(&stateid, &delegation->stateid);
2147 cred = get_rpccred(delegation->cred);
2148 rcu_read_unlock();
2149 status = nfs41_test_stateid(server, &stateid, cred);
2150 trace_nfs4_test_delegation_stateid(state, NULL, status);
2136 2151
2137 if (status != NFS_OK) { 2152 if (status != NFS_OK) {
2138 /* Free the stateid unless the server explicitly 2153 /* Free the stateid unless the server explicitly
2139 * informs us the stateid is unrecognized. */ 2154 * informs us the stateid is unrecognized. */
2140 if (status != -NFS4ERR_BAD_STATEID) 2155 if (status != -NFS4ERR_BAD_STATEID)
2141 nfs41_free_stateid(server, stateid, cred); 2156 nfs41_free_stateid(server, &stateid, cred);
2142 nfs_remove_bad_delegation(state->inode); 2157 nfs_finish_clear_delegation_stateid(state);
2143
2144 write_seqlock(&state->seqlock);
2145 nfs4_stateid_copy(&state->stateid, &state->open_stateid);
2146 write_sequnlock(&state->seqlock);
2147 clear_bit(NFS_DELEGATED_STATE, &state->flags);
2148 } 2158 }
2149 2159
2150 if (cred != NULL) 2160 put_rpccred(cred);
2151 put_rpccred(cred);
2152} 2161}
2153 2162
2154/** 2163/**
@@ -2192,7 +2201,7 @@ static int nfs41_open_expired(struct nfs4_state_owner *sp, struct nfs4_state *st
2192{ 2201{
2193 int status; 2202 int status;
2194 2203
2195 nfs41_clear_delegation_stateid(state); 2204 nfs41_check_delegation_stateid(state);
2196 status = nfs41_check_open_stateid(state); 2205 status = nfs41_check_open_stateid(state);
2197 if (status != NFS_OK) 2206 if (status != NFS_OK)
2198 status = nfs4_open_expired(sp, state); 2207 status = nfs4_open_expired(sp, state);
@@ -2231,19 +2240,8 @@ static int _nfs4_open_and_get_state(struct nfs4_opendata *opendata,
2231 seq = raw_seqcount_begin(&sp->so_reclaim_seqcount); 2240 seq = raw_seqcount_begin(&sp->so_reclaim_seqcount);
2232 2241
2233 ret = _nfs4_proc_open(opendata); 2242 ret = _nfs4_proc_open(opendata);
2234 if (ret != 0) { 2243 if (ret != 0)
2235 if (ret == -ENOENT) {
2236 dentry = opendata->dentry;
2237 if (dentry->d_inode)
2238 d_delete(dentry);
2239 else if (d_unhashed(dentry))
2240 d_add(dentry, NULL);
2241
2242 nfs_set_verifier(dentry,
2243 nfs_save_change_attribute(opendata->dir->d_inode));
2244 }
2245 goto out; 2244 goto out;
2246 }
2247 2245
2248 state = nfs4_opendata_to_nfs4_state(opendata); 2246 state = nfs4_opendata_to_nfs4_state(opendata);
2249 ret = PTR_ERR(state); 2247 ret = PTR_ERR(state);
@@ -4841,9 +4839,6 @@ nfs4_async_handle_error(struct rpc_task *task, const struct nfs_server *server,
4841 case -NFS4ERR_DELEG_REVOKED: 4839 case -NFS4ERR_DELEG_REVOKED:
4842 case -NFS4ERR_ADMIN_REVOKED: 4840 case -NFS4ERR_ADMIN_REVOKED:
4843 case -NFS4ERR_BAD_STATEID: 4841 case -NFS4ERR_BAD_STATEID:
4844 if (state == NULL)
4845 break;
4846 nfs_remove_bad_delegation(state->inode);
4847 case -NFS4ERR_OPENMODE: 4842 case -NFS4ERR_OPENMODE:
4848 if (state == NULL) 4843 if (state == NULL)
4849 break; 4844 break;
@@ -8341,7 +8336,7 @@ static const struct nfs4_state_recovery_ops nfs41_reboot_recovery_ops = {
8341static const struct nfs4_state_recovery_ops nfs40_nograce_recovery_ops = { 8336static const struct nfs4_state_recovery_ops nfs40_nograce_recovery_ops = {
8342 .owner_flag_bit = NFS_OWNER_RECLAIM_NOGRACE, 8337 .owner_flag_bit = NFS_OWNER_RECLAIM_NOGRACE,
8343 .state_flag_bit = NFS_STATE_RECLAIM_NOGRACE, 8338 .state_flag_bit = NFS_STATE_RECLAIM_NOGRACE,
8344 .recover_open = nfs4_open_expired, 8339 .recover_open = nfs40_open_expired,
8345 .recover_lock = nfs4_lock_expired, 8340 .recover_lock = nfs4_lock_expired,
8346 .establish_clid = nfs4_init_clientid, 8341 .establish_clid = nfs4_init_clientid,
8347}; 8342};
@@ -8408,8 +8403,7 @@ static const struct nfs4_minor_version_ops nfs_v4_1_minor_ops = {
8408 | NFS_CAP_CHANGE_ATTR 8403 | NFS_CAP_CHANGE_ATTR
8409 | NFS_CAP_POSIX_LOCK 8404 | NFS_CAP_POSIX_LOCK
8410 | NFS_CAP_STATEID_NFSV41 8405 | NFS_CAP_STATEID_NFSV41
8411 | NFS_CAP_ATOMIC_OPEN_V1 8406 | NFS_CAP_ATOMIC_OPEN_V1,
8412 | NFS_CAP_SEEK,
8413 .init_client = nfs41_init_client, 8407 .init_client = nfs41_init_client,
8414 .shutdown_client = nfs41_shutdown_client, 8408 .shutdown_client = nfs41_shutdown_client,
8415 .match_stateid = nfs41_match_stateid, 8409 .match_stateid = nfs41_match_stateid,
@@ -8431,7 +8425,8 @@ static const struct nfs4_minor_version_ops nfs_v4_2_minor_ops = {
8431 | NFS_CAP_CHANGE_ATTR 8425 | NFS_CAP_CHANGE_ATTR
8432 | NFS_CAP_POSIX_LOCK 8426 | NFS_CAP_POSIX_LOCK
8433 | NFS_CAP_STATEID_NFSV41 8427 | NFS_CAP_STATEID_NFSV41
8434 | NFS_CAP_ATOMIC_OPEN_V1, 8428 | NFS_CAP_ATOMIC_OPEN_V1
8429 | NFS_CAP_SEEK,
8435 .init_client = nfs41_init_client, 8430 .init_client = nfs41_init_client,
8436 .shutdown_client = nfs41_shutdown_client, 8431 .shutdown_client = nfs41_shutdown_client,
8437 .match_stateid = nfs41_match_stateid, 8432 .match_stateid = nfs41_match_stateid,
diff --git a/fs/nfs/write.c b/fs/nfs/write.c
index 12493846a2d3..f83b02dc9166 100644
--- a/fs/nfs/write.c
+++ b/fs/nfs/write.c
@@ -715,8 +715,6 @@ static void nfs_inode_remove_request(struct nfs_page *req)
715 715
716 if (test_and_clear_bit(PG_INODE_REF, &req->wb_flags)) 716 if (test_and_clear_bit(PG_INODE_REF, &req->wb_flags))
717 nfs_release_request(req); 717 nfs_release_request(req);
718 else
719 WARN_ON_ONCE(1);
720} 718}
721 719
722static void 720static void
diff --git a/fs/notify/fsnotify.c b/fs/notify/fsnotify.c
index 9d3e9c50066a..89326acd4561 100644
--- a/fs/notify/fsnotify.c
+++ b/fs/notify/fsnotify.c
@@ -229,8 +229,16 @@ int fsnotify(struct inode *to_tell, __u32 mask, void *data, int data_is,
229 &fsnotify_mark_srcu); 229 &fsnotify_mark_srcu);
230 } 230 }
231 231
232 /*
233 * We need to merge inode & vfsmount mark lists so that inode mark
234 * ignore masks are properly reflected for mount mark notifications.
235 * That's why this traversal is so complicated...
236 */
232 while (inode_node || vfsmount_node) { 237 while (inode_node || vfsmount_node) {
233 inode_group = vfsmount_group = NULL; 238 inode_group = NULL;
239 inode_mark = NULL;
240 vfsmount_group = NULL;
241 vfsmount_mark = NULL;
234 242
235 if (inode_node) { 243 if (inode_node) {
236 inode_mark = hlist_entry(srcu_dereference(inode_node, &fsnotify_mark_srcu), 244 inode_mark = hlist_entry(srcu_dereference(inode_node, &fsnotify_mark_srcu),
@@ -244,21 +252,19 @@ int fsnotify(struct inode *to_tell, __u32 mask, void *data, int data_is,
244 vfsmount_group = vfsmount_mark->group; 252 vfsmount_group = vfsmount_mark->group;
245 } 253 }
246 254
247 if (inode_group > vfsmount_group) { 255 if (inode_group && vfsmount_group) {
248 /* handle inode */ 256 int cmp = fsnotify_compare_groups(inode_group,
249 ret = send_to_group(to_tell, inode_mark, NULL, mask, 257 vfsmount_group);
250 data, data_is, cookie, file_name); 258 if (cmp > 0) {
251 /* we didn't use the vfsmount_mark */ 259 inode_group = NULL;
252 vfsmount_group = NULL; 260 inode_mark = NULL;
253 } else if (vfsmount_group > inode_group) { 261 } else if (cmp < 0) {
254 ret = send_to_group(to_tell, NULL, vfsmount_mark, mask, 262 vfsmount_group = NULL;
255 data, data_is, cookie, file_name); 263 vfsmount_mark = NULL;
256 inode_group = NULL; 264 }
257 } else {
258 ret = send_to_group(to_tell, inode_mark, vfsmount_mark,
259 mask, data, data_is, cookie,
260 file_name);
261 } 265 }
266 ret = send_to_group(to_tell, inode_mark, vfsmount_mark, mask,
267 data, data_is, cookie, file_name);
262 268
263 if (ret && (mask & ALL_FSNOTIFY_PERM_EVENTS)) 269 if (ret && (mask & ALL_FSNOTIFY_PERM_EVENTS))
264 goto out; 270 goto out;
diff --git a/fs/notify/fsnotify.h b/fs/notify/fsnotify.h
index 9c0898c4cfe1..3b68b0ae0a97 100644
--- a/fs/notify/fsnotify.h
+++ b/fs/notify/fsnotify.h
@@ -12,6 +12,10 @@ extern void fsnotify_flush_notify(struct fsnotify_group *group);
12/* protects reads of inode and vfsmount marks list */ 12/* protects reads of inode and vfsmount marks list */
13extern struct srcu_struct fsnotify_mark_srcu; 13extern struct srcu_struct fsnotify_mark_srcu;
14 14
15/* compare two groups for sorting of marks lists */
16extern int fsnotify_compare_groups(struct fsnotify_group *a,
17 struct fsnotify_group *b);
18
15extern void fsnotify_set_inode_mark_mask_locked(struct fsnotify_mark *fsn_mark, 19extern void fsnotify_set_inode_mark_mask_locked(struct fsnotify_mark *fsn_mark,
16 __u32 mask); 20 __u32 mask);
17/* add a mark to an inode */ 21/* add a mark to an inode */
diff --git a/fs/notify/inode_mark.c b/fs/notify/inode_mark.c
index e8497144b323..dfbf5447eea4 100644
--- a/fs/notify/inode_mark.c
+++ b/fs/notify/inode_mark.c
@@ -194,6 +194,7 @@ int fsnotify_add_inode_mark(struct fsnotify_mark *mark,
194{ 194{
195 struct fsnotify_mark *lmark, *last = NULL; 195 struct fsnotify_mark *lmark, *last = NULL;
196 int ret = 0; 196 int ret = 0;
197 int cmp;
197 198
198 mark->flags |= FSNOTIFY_MARK_FLAG_INODE; 199 mark->flags |= FSNOTIFY_MARK_FLAG_INODE;
199 200
@@ -219,11 +220,8 @@ int fsnotify_add_inode_mark(struct fsnotify_mark *mark,
219 goto out; 220 goto out;
220 } 221 }
221 222
222 if (mark->group->priority < lmark->group->priority) 223 cmp = fsnotify_compare_groups(lmark->group, mark->group);
223 continue; 224 if (cmp < 0)
224
225 if ((mark->group->priority == lmark->group->priority) &&
226 (mark->group < lmark->group))
227 continue; 225 continue;
228 226
229 hlist_add_before_rcu(&mark->i.i_list, &lmark->i.i_list); 227 hlist_add_before_rcu(&mark->i.i_list, &lmark->i.i_list);
diff --git a/fs/notify/mark.c b/fs/notify/mark.c
index d90deaa08e78..34c38fabf514 100644
--- a/fs/notify/mark.c
+++ b/fs/notify/mark.c
@@ -210,6 +210,42 @@ void fsnotify_set_mark_ignored_mask_locked(struct fsnotify_mark *mark, __u32 mas
210} 210}
211 211
212/* 212/*
213 * Sorting function for lists of fsnotify marks.
214 *
215 * Fanotify supports different notification classes (reflected as priority of
216 * notification group). Events shall be passed to notification groups in
217 * decreasing priority order. To achieve this marks in notification lists for
218 * inodes and vfsmounts are sorted so that priorities of corresponding groups
219 * are descending.
220 *
221 * Furthermore correct handling of the ignore mask requires processing inode
222 * and vfsmount marks of each group together. Using the group address as
223 * further sort criterion provides a unique sorting order and thus we can
224 * merge inode and vfsmount lists of marks in linear time and find groups
225 * present in both lists.
226 *
227 * A return value of 1 signifies that b has priority over a.
228 * A return value of 0 signifies that the two marks have to be handled together.
229 * A return value of -1 signifies that a has priority over b.
230 */
231int fsnotify_compare_groups(struct fsnotify_group *a, struct fsnotify_group *b)
232{
233 if (a == b)
234 return 0;
235 if (!a)
236 return 1;
237 if (!b)
238 return -1;
239 if (a->priority < b->priority)
240 return 1;
241 if (a->priority > b->priority)
242 return -1;
243 if (a < b)
244 return 1;
245 return -1;
246}
247
248/*
213 * Attach an initialized mark to a given group and fs object. 249 * Attach an initialized mark to a given group and fs object.
214 * These marks may be used for the fsnotify backend to determine which 250 * These marks may be used for the fsnotify backend to determine which
215 * event types should be delivered to which group. 251 * event types should be delivered to which group.
diff --git a/fs/notify/vfsmount_mark.c b/fs/notify/vfsmount_mark.c
index ac851e8376b1..faefa72a11eb 100644
--- a/fs/notify/vfsmount_mark.c
+++ b/fs/notify/vfsmount_mark.c
@@ -153,6 +153,7 @@ int fsnotify_add_vfsmount_mark(struct fsnotify_mark *mark,
153 struct mount *m = real_mount(mnt); 153 struct mount *m = real_mount(mnt);
154 struct fsnotify_mark *lmark, *last = NULL; 154 struct fsnotify_mark *lmark, *last = NULL;
155 int ret = 0; 155 int ret = 0;
156 int cmp;
156 157
157 mark->flags |= FSNOTIFY_MARK_FLAG_VFSMOUNT; 158 mark->flags |= FSNOTIFY_MARK_FLAG_VFSMOUNT;
158 159
@@ -178,11 +179,8 @@ int fsnotify_add_vfsmount_mark(struct fsnotify_mark *mark,
178 goto out; 179 goto out;
179 } 180 }
180 181
181 if (mark->group->priority < lmark->group->priority) 182 cmp = fsnotify_compare_groups(lmark->group, mark->group);
182 continue; 183 if (cmp < 0)
183
184 if ((mark->group->priority == lmark->group->priority) &&
185 (mark->group < lmark->group))
186 continue; 184 continue;
187 185
188 hlist_add_before_rcu(&mark->m.m_list, &lmark->m.m_list); 186 hlist_add_before_rcu(&mark->m.m_list, &lmark->m.m_list);
diff --git a/fs/ocfs2/cluster/tcp.c b/fs/ocfs2/cluster/tcp.c
index 97de0fbd9f78..a96044004064 100644
--- a/fs/ocfs2/cluster/tcp.c
+++ b/fs/ocfs2/cluster/tcp.c
@@ -925,7 +925,7 @@ static int o2net_send_tcp_msg(struct socket *sock, struct kvec *vec,
925 size_t veclen, size_t total) 925 size_t veclen, size_t total)
926{ 926{
927 int ret; 927 int ret;
928 struct msghdr msg; 928 struct msghdr msg = {.msg_flags = 0,};
929 929
930 if (sock == NULL) { 930 if (sock == NULL) {
931 ret = -EINVAL; 931 ret = -EINVAL;
diff --git a/fs/overlayfs/readdir.c b/fs/overlayfs/readdir.c
index 4e9d7c1fea52..2a7ef4f8e2a6 100644
--- a/fs/overlayfs/readdir.c
+++ b/fs/overlayfs/readdir.c
@@ -168,7 +168,7 @@ static void ovl_cache_put(struct ovl_dir_file *od, struct dentry *dentry)
168{ 168{
169 struct ovl_dir_cache *cache = od->cache; 169 struct ovl_dir_cache *cache = od->cache;
170 170
171 list_del(&od->cursor.l_node); 171 list_del_init(&od->cursor.l_node);
172 WARN_ON(cache->refcount <= 0); 172 WARN_ON(cache->refcount <= 0);
173 cache->refcount--; 173 cache->refcount--;
174 if (!cache->refcount) { 174 if (!cache->refcount) {
diff --git a/fs/xfs/xfs_bmap_util.c b/fs/xfs/xfs_bmap_util.c
index 92e8f99a5857..281002689d64 100644
--- a/fs/xfs/xfs_bmap_util.c
+++ b/fs/xfs/xfs_bmap_util.c
@@ -1338,7 +1338,10 @@ xfs_free_file_space(
1338 goto out; 1338 goto out;
1339} 1339}
1340 1340
1341 1341/*
1342 * Preallocate and zero a range of a file. This mechanism has the allocation
1343 * semantics of fallocate and in addition converts data in the range to zeroes.
1344 */
1342int 1345int
1343xfs_zero_file_space( 1346xfs_zero_file_space(
1344 struct xfs_inode *ip, 1347 struct xfs_inode *ip,
@@ -1346,65 +1349,30 @@ xfs_zero_file_space(
1346 xfs_off_t len) 1349 xfs_off_t len)
1347{ 1350{
1348 struct xfs_mount *mp = ip->i_mount; 1351 struct xfs_mount *mp = ip->i_mount;
1349 uint granularity; 1352 uint blksize;
1350 xfs_off_t start_boundary;
1351 xfs_off_t end_boundary;
1352 int error; 1353 int error;
1353 1354
1354 trace_xfs_zero_file_space(ip); 1355 trace_xfs_zero_file_space(ip);
1355 1356
1356 granularity = max_t(uint, 1 << mp->m_sb.sb_blocklog, PAGE_CACHE_SIZE); 1357 blksize = 1 << mp->m_sb.sb_blocklog;
1357 1358
1358 /* 1359 /*
1359 * Round the range of extents we are going to convert inwards. If the 1360 * Punch a hole and prealloc the range. We use hole punch rather than
1360 * offset is aligned, then it doesn't get changed so we zero from the 1361 * unwritten extent conversion for two reasons:
1361 * start of the block offset points to. 1362 *
1363 * 1.) Hole punch handles partial block zeroing for us.
1364 *
1365 * 2.) If prealloc returns ENOSPC, the file range is still zero-valued
1366 * by virtue of the hole punch.
1362 */ 1367 */
1363 start_boundary = round_up(offset, granularity); 1368 error = xfs_free_file_space(ip, offset, len);
1364 end_boundary = round_down(offset + len, granularity); 1369 if (error)
1365 1370 goto out;
1366 ASSERT(start_boundary >= offset);
1367 ASSERT(end_boundary <= offset + len);
1368
1369 if (start_boundary < end_boundary - 1) {
1370 /*
1371 * Writeback the range to ensure any inode size updates due to
1372 * appending writes make it to disk (otherwise we could just
1373 * punch out the delalloc blocks).
1374 */
1375 error = filemap_write_and_wait_range(VFS_I(ip)->i_mapping,
1376 start_boundary, end_boundary - 1);
1377 if (error)
1378 goto out;
1379 truncate_pagecache_range(VFS_I(ip), start_boundary,
1380 end_boundary - 1);
1381
1382 /* convert the blocks */
1383 error = xfs_alloc_file_space(ip, start_boundary,
1384 end_boundary - start_boundary - 1,
1385 XFS_BMAPI_PREALLOC | XFS_BMAPI_CONVERT);
1386 if (error)
1387 goto out;
1388
1389 /* We've handled the interior of the range, now for the edges */
1390 if (start_boundary != offset) {
1391 error = xfs_iozero(ip, offset, start_boundary - offset);
1392 if (error)
1393 goto out;
1394 }
1395
1396 if (end_boundary != offset + len)
1397 error = xfs_iozero(ip, end_boundary,
1398 offset + len - end_boundary);
1399
1400 } else {
1401 /*
1402 * It's either a sub-granularity range or the range spanned lies
1403 * partially across two adjacent blocks.
1404 */
1405 error = xfs_iozero(ip, offset, len);
1406 }
1407 1371
1372 error = xfs_alloc_file_space(ip, round_down(offset, blksize),
1373 round_up(offset + len, blksize) -
1374 round_down(offset, blksize),
1375 XFS_BMAPI_PREALLOC);
1408out: 1376out:
1409 return error; 1377 return error;
1410 1378
diff --git a/fs/xfs/xfs_itable.c b/fs/xfs/xfs_itable.c
index f1deb961a296..894924a5129b 100644
--- a/fs/xfs/xfs_itable.c
+++ b/fs/xfs/xfs_itable.c
@@ -236,8 +236,10 @@ xfs_bulkstat_grab_ichunk(
236 XFS_WANT_CORRUPTED_RETURN(stat == 1); 236 XFS_WANT_CORRUPTED_RETURN(stat == 1);
237 237
238 /* Check if the record contains the inode in request */ 238 /* Check if the record contains the inode in request */
239 if (irec->ir_startino + XFS_INODES_PER_CHUNK <= agino) 239 if (irec->ir_startino + XFS_INODES_PER_CHUNK <= agino) {
240 return -EINVAL; 240 *icount = 0;
241 return 0;
242 }
241 243
242 idx = agino - irec->ir_startino + 1; 244 idx = agino - irec->ir_startino + 1;
243 if (idx < XFS_INODES_PER_CHUNK && 245 if (idx < XFS_INODES_PER_CHUNK &&
@@ -262,75 +264,76 @@ xfs_bulkstat_grab_ichunk(
262 264
263#define XFS_BULKSTAT_UBLEFT(ubleft) ((ubleft) >= statstruct_size) 265#define XFS_BULKSTAT_UBLEFT(ubleft) ((ubleft) >= statstruct_size)
264 266
267struct xfs_bulkstat_agichunk {
268 char __user **ac_ubuffer;/* pointer into user's buffer */
269 int ac_ubleft; /* bytes left in user's buffer */
270 int ac_ubelem; /* spaces used in user's buffer */
271};
272
265/* 273/*
266 * Process inodes in chunk with a pointer to a formatter function 274 * Process inodes in chunk with a pointer to a formatter function
267 * that will iget the inode and fill in the appropriate structure. 275 * that will iget the inode and fill in the appropriate structure.
268 */ 276 */
269int 277static int
270xfs_bulkstat_ag_ichunk( 278xfs_bulkstat_ag_ichunk(
271 struct xfs_mount *mp, 279 struct xfs_mount *mp,
272 xfs_agnumber_t agno, 280 xfs_agnumber_t agno,
273 struct xfs_inobt_rec_incore *irbp, 281 struct xfs_inobt_rec_incore *irbp,
274 bulkstat_one_pf formatter, 282 bulkstat_one_pf formatter,
275 size_t statstruct_size, 283 size_t statstruct_size,
276 struct xfs_bulkstat_agichunk *acp) 284 struct xfs_bulkstat_agichunk *acp,
285 xfs_agino_t *last_agino)
277{ 286{
278 xfs_ino_t lastino = acp->ac_lastino;
279 char __user **ubufp = acp->ac_ubuffer; 287 char __user **ubufp = acp->ac_ubuffer;
280 int ubleft = acp->ac_ubleft; 288 int chunkidx;
281 int ubelem = acp->ac_ubelem;
282 int chunkidx, clustidx;
283 int error = 0; 289 int error = 0;
284 xfs_agino_t agino; 290 xfs_agino_t agino = irbp->ir_startino;
285 291
286 for (agino = irbp->ir_startino, chunkidx = clustidx = 0; 292 for (chunkidx = 0; chunkidx < XFS_INODES_PER_CHUNK;
287 XFS_BULKSTAT_UBLEFT(ubleft) && 293 chunkidx++, agino++) {
288 irbp->ir_freecount < XFS_INODES_PER_CHUNK; 294 int fmterror;
289 chunkidx++, clustidx++, agino++) {
290 int fmterror; /* bulkstat formatter result */
291 int ubused; 295 int ubused;
292 xfs_ino_t ino = XFS_AGINO_TO_INO(mp, agno, agino);
293 296
294 ASSERT(chunkidx < XFS_INODES_PER_CHUNK); 297 /* inode won't fit in buffer, we are done */
298 if (acp->ac_ubleft < statstruct_size)
299 break;
295 300
296 /* Skip if this inode is free */ 301 /* Skip if this inode is free */
297 if (XFS_INOBT_MASK(chunkidx) & irbp->ir_free) { 302 if (XFS_INOBT_MASK(chunkidx) & irbp->ir_free)
298 lastino = ino;
299 continue; 303 continue;
300 }
301
302 /*
303 * Count used inodes as free so we can tell when the
304 * chunk is used up.
305 */
306 irbp->ir_freecount++;
307 304
308 /* Get the inode and fill in a single buffer */ 305 /* Get the inode and fill in a single buffer */
309 ubused = statstruct_size; 306 ubused = statstruct_size;
310 error = formatter(mp, ino, *ubufp, ubleft, &ubused, &fmterror); 307 error = formatter(mp, XFS_AGINO_TO_INO(mp, agno, agino),
311 if (fmterror == BULKSTAT_RV_NOTHING) { 308 *ubufp, acp->ac_ubleft, &ubused, &fmterror);
312 if (error && error != -ENOENT && error != -EINVAL) { 309
313 ubleft = 0; 310 if (fmterror == BULKSTAT_RV_GIVEUP ||
314 break; 311 (error && error != -ENOENT && error != -EINVAL)) {
315 } 312 acp->ac_ubleft = 0;
316 lastino = ino;
317 continue;
318 }
319 if (fmterror == BULKSTAT_RV_GIVEUP) {
320 ubleft = 0;
321 ASSERT(error); 313 ASSERT(error);
322 break; 314 break;
323 } 315 }
324 if (*ubufp) 316
325 *ubufp += ubused; 317 /* be careful not to leak error if at end of chunk */
326 ubleft -= ubused; 318 if (fmterror == BULKSTAT_RV_NOTHING || error) {
327 ubelem++; 319 error = 0;
328 lastino = ino; 320 continue;
321 }
322
323 *ubufp += ubused;
324 acp->ac_ubleft -= ubused;
325 acp->ac_ubelem++;
329 } 326 }
330 327
331 acp->ac_lastino = lastino; 328 /*
332 acp->ac_ubleft = ubleft; 329 * Post-update *last_agino. At this point, agino will always point one
333 acp->ac_ubelem = ubelem; 330 * inode past the last inode we processed successfully. Hence we
331 * substract that inode when setting the *last_agino cursor so that we
332 * return the correct cookie to userspace. On the next bulkstat call,
333 * the inode under the lastino cookie will be skipped as we have already
334 * processed it here.
335 */
336 *last_agino = agino - 1;
334 337
335 return error; 338 return error;
336} 339}
@@ -353,45 +356,33 @@ xfs_bulkstat(
353 xfs_agino_t agino; /* inode # in allocation group */ 356 xfs_agino_t agino; /* inode # in allocation group */
354 xfs_agnumber_t agno; /* allocation group number */ 357 xfs_agnumber_t agno; /* allocation group number */
355 xfs_btree_cur_t *cur; /* btree cursor for ialloc btree */ 358 xfs_btree_cur_t *cur; /* btree cursor for ialloc btree */
356 int end_of_ag; /* set if we've seen the ag end */
357 int error; /* error code */
358 int fmterror;/* bulkstat formatter result */
359 int i; /* loop index */
360 int icount; /* count of inodes good in irbuf */
361 size_t irbsize; /* size of irec buffer in bytes */ 359 size_t irbsize; /* size of irec buffer in bytes */
362 xfs_ino_t ino; /* inode number (filesystem) */
363 xfs_inobt_rec_incore_t *irbp; /* current irec buffer pointer */
364 xfs_inobt_rec_incore_t *irbuf; /* start of irec buffer */ 360 xfs_inobt_rec_incore_t *irbuf; /* start of irec buffer */
365 xfs_inobt_rec_incore_t *irbufend; /* end of good irec buffer entries */
366 xfs_ino_t lastino; /* last inode number returned */
367 int nirbuf; /* size of irbuf */ 361 int nirbuf; /* size of irbuf */
368 int rval; /* return value error code */
369 int tmp; /* result value from btree calls */
370 int ubcount; /* size of user's buffer */ 362 int ubcount; /* size of user's buffer */
371 int ubleft; /* bytes left in user's buffer */ 363 struct xfs_bulkstat_agichunk ac;
372 char __user *ubufp; /* pointer into user's buffer */ 364 int error = 0;
373 int ubelem; /* spaces used in user's buffer */
374 365
375 /* 366 /*
376 * Get the last inode value, see if there's nothing to do. 367 * Get the last inode value, see if there's nothing to do.
377 */ 368 */
378 ino = (xfs_ino_t)*lastinop; 369 agno = XFS_INO_TO_AGNO(mp, *lastinop);
379 lastino = ino; 370 agino = XFS_INO_TO_AGINO(mp, *lastinop);
380 agno = XFS_INO_TO_AGNO(mp, ino);
381 agino = XFS_INO_TO_AGINO(mp, ino);
382 if (agno >= mp->m_sb.sb_agcount || 371 if (agno >= mp->m_sb.sb_agcount ||
383 ino != XFS_AGINO_TO_INO(mp, agno, agino)) { 372 *lastinop != XFS_AGINO_TO_INO(mp, agno, agino)) {
384 *done = 1; 373 *done = 1;
385 *ubcountp = 0; 374 *ubcountp = 0;
386 return 0; 375 return 0;
387 } 376 }
388 377
389 ubcount = *ubcountp; /* statstruct's */ 378 ubcount = *ubcountp; /* statstruct's */
390 ubleft = ubcount * statstruct_size; /* bytes */ 379 ac.ac_ubuffer = &ubuffer;
391 *ubcountp = ubelem = 0; 380 ac.ac_ubleft = ubcount * statstruct_size; /* bytes */;
381 ac.ac_ubelem = 0;
382
383 *ubcountp = 0;
392 *done = 0; 384 *done = 0;
393 fmterror = 0; 385
394 ubufp = ubuffer;
395 irbuf = kmem_zalloc_greedy(&irbsize, PAGE_SIZE, PAGE_SIZE * 4); 386 irbuf = kmem_zalloc_greedy(&irbsize, PAGE_SIZE, PAGE_SIZE * 4);
396 if (!irbuf) 387 if (!irbuf)
397 return -ENOMEM; 388 return -ENOMEM;
@@ -402,9 +393,13 @@ xfs_bulkstat(
402 * Loop over the allocation groups, starting from the last 393 * Loop over the allocation groups, starting from the last
403 * inode returned; 0 means start of the allocation group. 394 * inode returned; 0 means start of the allocation group.
404 */ 395 */
405 rval = 0; 396 while (agno < mp->m_sb.sb_agcount) {
406 while (XFS_BULKSTAT_UBLEFT(ubleft) && agno < mp->m_sb.sb_agcount) { 397 struct xfs_inobt_rec_incore *irbp = irbuf;
407 cond_resched(); 398 struct xfs_inobt_rec_incore *irbufend = irbuf + nirbuf;
399 bool end_of_ag = false;
400 int icount = 0;
401 int stat;
402
408 error = xfs_ialloc_read_agi(mp, NULL, agno, &agbp); 403 error = xfs_ialloc_read_agi(mp, NULL, agno, &agbp);
409 if (error) 404 if (error)
410 break; 405 break;
@@ -414,10 +409,6 @@ xfs_bulkstat(
414 */ 409 */
415 cur = xfs_inobt_init_cursor(mp, NULL, agbp, agno, 410 cur = xfs_inobt_init_cursor(mp, NULL, agbp, agno,
416 XFS_BTNUM_INO); 411 XFS_BTNUM_INO);
417 irbp = irbuf;
418 irbufend = irbuf + nirbuf;
419 end_of_ag = 0;
420 icount = 0;
421 if (agino > 0) { 412 if (agino > 0) {
422 /* 413 /*
423 * In the middle of an allocation group, we need to get 414 * In the middle of an allocation group, we need to get
@@ -427,22 +418,23 @@ xfs_bulkstat(
427 418
428 error = xfs_bulkstat_grab_ichunk(cur, agino, &icount, &r); 419 error = xfs_bulkstat_grab_ichunk(cur, agino, &icount, &r);
429 if (error) 420 if (error)
430 break; 421 goto del_cursor;
431 if (icount) { 422 if (icount) {
432 irbp->ir_startino = r.ir_startino; 423 irbp->ir_startino = r.ir_startino;
433 irbp->ir_freecount = r.ir_freecount; 424 irbp->ir_freecount = r.ir_freecount;
434 irbp->ir_free = r.ir_free; 425 irbp->ir_free = r.ir_free;
435 irbp++; 426 irbp++;
436 agino = r.ir_startino + XFS_INODES_PER_CHUNK;
437 } 427 }
438 /* Increment to the next record */ 428 /* Increment to the next record */
439 error = xfs_btree_increment(cur, 0, &tmp); 429 error = xfs_btree_increment(cur, 0, &stat);
440 } else { 430 } else {
441 /* Start of ag. Lookup the first inode chunk */ 431 /* Start of ag. Lookup the first inode chunk */
442 error = xfs_inobt_lookup(cur, 0, XFS_LOOKUP_GE, &tmp); 432 error = xfs_inobt_lookup(cur, 0, XFS_LOOKUP_GE, &stat);
433 }
434 if (error || stat == 0) {
435 end_of_ag = true;
436 goto del_cursor;
443 } 437 }
444 if (error)
445 break;
446 438
447 /* 439 /*
448 * Loop through inode btree records in this ag, 440 * Loop through inode btree records in this ag,
@@ -451,10 +443,10 @@ xfs_bulkstat(
451 while (irbp < irbufend && icount < ubcount) { 443 while (irbp < irbufend && icount < ubcount) {
452 struct xfs_inobt_rec_incore r; 444 struct xfs_inobt_rec_incore r;
453 445
454 error = xfs_inobt_get_rec(cur, &r, &i); 446 error = xfs_inobt_get_rec(cur, &r, &stat);
455 if (error || i == 0) { 447 if (error || stat == 0) {
456 end_of_ag = 1; 448 end_of_ag = true;
457 break; 449 goto del_cursor;
458 } 450 }
459 451
460 /* 452 /*
@@ -469,77 +461,79 @@ xfs_bulkstat(
469 irbp++; 461 irbp++;
470 icount += XFS_INODES_PER_CHUNK - r.ir_freecount; 462 icount += XFS_INODES_PER_CHUNK - r.ir_freecount;
471 } 463 }
472 /* 464 error = xfs_btree_increment(cur, 0, &stat);
473 * Set agino to after this chunk and bump the cursor. 465 if (error || stat == 0) {
474 */ 466 end_of_ag = true;
475 agino = r.ir_startino + XFS_INODES_PER_CHUNK; 467 goto del_cursor;
476 error = xfs_btree_increment(cur, 0, &tmp); 468 }
477 cond_resched(); 469 cond_resched();
478 } 470 }
471
479 /* 472 /*
480 * Drop the btree buffers and the agi buffer. 473 * Drop the btree buffers and the agi buffer as we can't hold any
481 * We can't hold any of the locks these represent 474 * of the locks these represent when calling iget. If there is a
482 * when calling iget. 475 * pending error, then we are done.
483 */ 476 */
477del_cursor:
484 xfs_btree_del_cursor(cur, XFS_BTREE_NOERROR); 478 xfs_btree_del_cursor(cur, XFS_BTREE_NOERROR);
485 xfs_buf_relse(agbp); 479 xfs_buf_relse(agbp);
480 if (error)
481 break;
486 /* 482 /*
487 * Now format all the good inodes into the user's buffer. 483 * Now format all the good inodes into the user's buffer. The
484 * call to xfs_bulkstat_ag_ichunk() sets up the agino pointer
485 * for the next loop iteration.
488 */ 486 */
489 irbufend = irbp; 487 irbufend = irbp;
490 for (irbp = irbuf; 488 for (irbp = irbuf;
491 irbp < irbufend && XFS_BULKSTAT_UBLEFT(ubleft); irbp++) { 489 irbp < irbufend && ac.ac_ubleft >= statstruct_size;
492 struct xfs_bulkstat_agichunk ac; 490 irbp++) {
493
494 ac.ac_lastino = lastino;
495 ac.ac_ubuffer = &ubuffer;
496 ac.ac_ubleft = ubleft;
497 ac.ac_ubelem = ubelem;
498 error = xfs_bulkstat_ag_ichunk(mp, agno, irbp, 491 error = xfs_bulkstat_ag_ichunk(mp, agno, irbp,
499 formatter, statstruct_size, &ac); 492 formatter, statstruct_size, &ac,
493 &agino);
500 if (error) 494 if (error)
501 rval = error; 495 break;
502
503 lastino = ac.ac_lastino;
504 ubleft = ac.ac_ubleft;
505 ubelem = ac.ac_ubelem;
506 496
507 cond_resched(); 497 cond_resched();
508 } 498 }
499
509 /* 500 /*
510 * Set up for the next loop iteration. 501 * If we've run out of space or had a formatting error, we
502 * are now done
511 */ 503 */
512 if (XFS_BULKSTAT_UBLEFT(ubleft)) { 504 if (ac.ac_ubleft < statstruct_size || error)
513 if (end_of_ag) {
514 agno++;
515 agino = 0;
516 } else
517 agino = XFS_INO_TO_AGINO(mp, lastino);
518 } else
519 break; 505 break;
506
507 if (end_of_ag) {
508 agno++;
509 agino = 0;
510 }
520 } 511 }
521 /* 512 /*
522 * Done, we're either out of filesystem or space to put the data. 513 * Done, we're either out of filesystem or space to put the data.
523 */ 514 */
524 kmem_free(irbuf); 515 kmem_free(irbuf);
525 *ubcountp = ubelem; 516 *ubcountp = ac.ac_ubelem;
517
526 /* 518 /*
527 * Found some inodes, return them now and return the error next time. 519 * We found some inodes, so clear the error status and return them.
520 * The lastino pointer will point directly at the inode that triggered
521 * any error that occurred, so on the next call the error will be
522 * triggered again and propagated to userspace as there will be no
523 * formatted inodes in the buffer.
528 */ 524 */
529 if (ubelem) 525 if (ac.ac_ubelem)
530 rval = 0; 526 error = 0;
531 if (agno >= mp->m_sb.sb_agcount) { 527
532 /* 528 /*
533 * If we ran out of filesystem, mark lastino as off 529 * If we ran out of filesystem, lastino will point off the end of
534 * the end of the filesystem, so the next call 530 * the filesystem so the next call will return immediately.
535 * will return immediately. 531 */
536 */ 532 *lastinop = XFS_AGINO_TO_INO(mp, agno, agino);
537 *lastinop = (xfs_ino_t)XFS_AGINO_TO_INO(mp, agno, 0); 533 if (agno >= mp->m_sb.sb_agcount)
538 *done = 1; 534 *done = 1;
539 } else
540 *lastinop = (xfs_ino_t)lastino;
541 535
542 return rval; 536 return error;
543} 537}
544 538
545int 539int
diff --git a/fs/xfs/xfs_itable.h b/fs/xfs/xfs_itable.h
index aaed08022eb9..6ea8b3912fa4 100644
--- a/fs/xfs/xfs_itable.h
+++ b/fs/xfs/xfs_itable.h
@@ -30,22 +30,6 @@ typedef int (*bulkstat_one_pf)(struct xfs_mount *mp,
30 int *ubused, 30 int *ubused,
31 int *stat); 31 int *stat);
32 32
33struct xfs_bulkstat_agichunk {
34 xfs_ino_t ac_lastino; /* last inode returned */
35 char __user **ac_ubuffer;/* pointer into user's buffer */
36 int ac_ubleft; /* bytes left in user's buffer */
37 int ac_ubelem; /* spaces used in user's buffer */
38};
39
40int
41xfs_bulkstat_ag_ichunk(
42 struct xfs_mount *mp,
43 xfs_agnumber_t agno,
44 struct xfs_inobt_rec_incore *irbp,
45 bulkstat_one_pf formatter,
46 size_t statstruct_size,
47 struct xfs_bulkstat_agichunk *acp);
48
49/* 33/*
50 * Values for stat return value. 34 * Values for stat return value.
51 */ 35 */