aboutsummaryrefslogtreecommitdiffstats
path: root/fs
diff options
context:
space:
mode:
Diffstat (limited to 'fs')
-rw-r--r--fs/9p/v9fs.c4
-rw-r--r--fs/9p/v9fs_vfs.h6
-rw-r--r--fs/9p/vfs_addr.c5
-rw-r--r--fs/9p/vfs_dir.c60
-rw-r--r--fs/9p/vfs_file.c93
-rw-r--r--fs/9p/vfs_inode.c39
-rw-r--r--fs/9p/vfs_super.c6
-rw-r--r--fs/Kconfig607
-rw-r--r--fs/Kconfig.binfmt22
-rw-r--r--fs/binfmt_elf.c31
-rw-r--r--fs/binfmt_elf_fdpic.c19
-rw-r--r--fs/buffer.c3
-rw-r--r--fs/cifs/Kconfig142
-rw-r--r--fs/cifs/file.c4
-rw-r--r--fs/exec.c7
-rw-r--r--fs/ext2/Kconfig55
-rw-r--r--fs/ext3/Kconfig67
-rw-r--r--fs/ext3/balloc.c3
-rw-r--r--fs/ext3/dir.c30
-rw-r--r--fs/ext3/inode.c7
-rw-r--r--fs/ext3/resize.c3
-rw-r--r--fs/ext3/super.c16
-rw-r--r--fs/ext4/Kconfig79
-rw-r--r--fs/fuse/file.c5
-rw-r--r--fs/fuse/fuse_i.h5
-rw-r--r--fs/fuse/inode.c3
-rw-r--r--fs/hfsplus/extents.c3
-rw-r--r--fs/hfsplus/inode.c2
-rw-r--r--fs/jbd/Kconfig30
-rw-r--r--fs/jbd/commit.c10
-rw-r--r--fs/jbd/transaction.c16
-rw-r--r--fs/jbd2/Kconfig33
-rw-r--r--fs/jffs2/Kconfig188
-rw-r--r--fs/jffs2/compr.c4
-rw-r--r--fs/jffs2/dir.c2
-rw-r--r--fs/jffs2/erase.c4
-rw-r--r--fs/jffs2/fs.c6
-rw-r--r--fs/jffs2/nodemgmt.c4
-rw-r--r--fs/jffs2/wbuf.c5
-rw-r--r--fs/nfs/callback.c19
-rw-r--r--fs/nfs/dir.c2
-rw-r--r--fs/nfs/nfs4proc.c6
-rw-r--r--fs/nfs/super.c4
-rw-r--r--fs/ntfs/file.c4
-rw-r--r--fs/proc/array.c8
-rw-r--r--fs/proc/proc_misc.c125
-rw-r--r--fs/proc/vmcore.c5
-rw-r--r--fs/ramfs/file-nommu.c4
-rw-r--r--fs/ramfs/inode.c1
-rw-r--r--fs/seq_file.c29
-rw-r--r--fs/ubifs/budget.c26
-rw-r--r--fs/ubifs/compress.c2
-rw-r--r--fs/ubifs/debug.c79
-rw-r--r--fs/ubifs/debug.h6
-rw-r--r--fs/ubifs/file.c260
-rw-r--r--fs/ubifs/find.c4
-rw-r--r--fs/ubifs/gc.c90
-rw-r--r--fs/ubifs/io.c12
-rw-r--r--fs/ubifs/key.h22
-rw-r--r--fs/ubifs/lprops.c34
-rw-r--r--fs/ubifs/lpt.c48
-rw-r--r--fs/ubifs/lpt_commit.c187
-rw-r--r--fs/ubifs/misc.h27
-rw-r--r--fs/ubifs/scan.c2
-rw-r--r--fs/ubifs/super.c109
-rw-r--r--fs/ubifs/tnc.c345
-rw-r--r--fs/ubifs/tnc_misc.c4
-rw-r--r--fs/ubifs/ubifs-media.h1
-rw-r--r--fs/ubifs/ubifs.h85
-rw-r--r--fs/ubifs/xattr.c2
70 files changed, 2217 insertions, 963 deletions
diff --git a/fs/9p/v9fs.c b/fs/9p/v9fs.c
index c061c3f18e7..24eb01087b6 100644
--- a/fs/9p/v9fs.c
+++ b/fs/9p/v9fs.c
@@ -30,8 +30,8 @@
30#include <linux/parser.h> 30#include <linux/parser.h>
31#include <linux/idr.h> 31#include <linux/idr.h>
32#include <net/9p/9p.h> 32#include <net/9p/9p.h>
33#include <net/9p/transport.h>
34#include <net/9p/client.h> 33#include <net/9p/client.h>
34#include <net/9p/transport.h>
35#include "v9fs.h" 35#include "v9fs.h"
36#include "v9fs_vfs.h" 36#include "v9fs_vfs.h"
37 37
@@ -234,7 +234,7 @@ struct p9_fid *v9fs_session_init(struct v9fs_session_info *v9ses,
234 if (!v9ses->clnt->dotu) 234 if (!v9ses->clnt->dotu)
235 v9ses->flags &= ~V9FS_EXTENDED; 235 v9ses->flags &= ~V9FS_EXTENDED;
236 236
237 v9ses->maxdata = v9ses->clnt->msize; 237 v9ses->maxdata = v9ses->clnt->msize - P9_IOHDRSZ;
238 238
239 /* for legacy mode, fall back to V9FS_ACCESS_ANY */ 239 /* for legacy mode, fall back to V9FS_ACCESS_ANY */
240 if (!v9fs_extended(v9ses) && 240 if (!v9fs_extended(v9ses) &&
diff --git a/fs/9p/v9fs_vfs.h b/fs/9p/v9fs_vfs.h
index 57997fa14e6..c295ba786ed 100644
--- a/fs/9p/v9fs_vfs.h
+++ b/fs/9p/v9fs_vfs.h
@@ -46,9 +46,11 @@ extern struct dentry_operations v9fs_cached_dentry_operations;
46 46
47struct inode *v9fs_get_inode(struct super_block *sb, int mode); 47struct inode *v9fs_get_inode(struct super_block *sb, int mode);
48ino_t v9fs_qid2ino(struct p9_qid *qid); 48ino_t v9fs_qid2ino(struct p9_qid *qid);
49void v9fs_stat2inode(struct p9_stat *, struct inode *, struct super_block *); 49void v9fs_stat2inode(struct p9_wstat *, struct inode *, struct super_block *);
50int v9fs_dir_release(struct inode *inode, struct file *filp); 50int v9fs_dir_release(struct inode *inode, struct file *filp);
51int v9fs_file_open(struct inode *inode, struct file *file); 51int v9fs_file_open(struct inode *inode, struct file *file);
52void v9fs_inode2stat(struct inode *inode, struct p9_stat *stat); 52void v9fs_inode2stat(struct inode *inode, struct p9_wstat *stat);
53void v9fs_dentry_release(struct dentry *); 53void v9fs_dentry_release(struct dentry *);
54int v9fs_uflags2omode(int uflags, int extended); 54int v9fs_uflags2omode(int uflags, int extended);
55
56ssize_t v9fs_file_readn(struct file *, char *, char __user *, u32, u64);
diff --git a/fs/9p/vfs_addr.c b/fs/9p/vfs_addr.c
index 97d3aed5798..6fcb1e7095c 100644
--- a/fs/9p/vfs_addr.c
+++ b/fs/9p/vfs_addr.c
@@ -38,7 +38,6 @@
38 38
39#include "v9fs.h" 39#include "v9fs.h"
40#include "v9fs_vfs.h" 40#include "v9fs_vfs.h"
41#include "fid.h"
42 41
43/** 42/**
44 * v9fs_vfs_readpage - read an entire page in from 9P 43 * v9fs_vfs_readpage - read an entire page in from 9P
@@ -53,14 +52,12 @@ static int v9fs_vfs_readpage(struct file *filp, struct page *page)
53 int retval; 52 int retval;
54 loff_t offset; 53 loff_t offset;
55 char *buffer; 54 char *buffer;
56 struct p9_fid *fid;
57 55
58 P9_DPRINTK(P9_DEBUG_VFS, "\n"); 56 P9_DPRINTK(P9_DEBUG_VFS, "\n");
59 fid = filp->private_data;
60 buffer = kmap(page); 57 buffer = kmap(page);
61 offset = page_offset(page); 58 offset = page_offset(page);
62 59
63 retval = p9_client_readn(fid, buffer, offset, PAGE_CACHE_SIZE); 60 retval = v9fs_file_readn(filp, buffer, NULL, offset, PAGE_CACHE_SIZE);
64 if (retval < 0) 61 if (retval < 0)
65 goto done; 62 goto done;
66 63
diff --git a/fs/9p/vfs_dir.c b/fs/9p/vfs_dir.c
index e298fe19409..873cd31baa4 100644
--- a/fs/9p/vfs_dir.c
+++ b/fs/9p/vfs_dir.c
@@ -45,7 +45,7 @@
45 * 45 *
46 */ 46 */
47 47
48static inline int dt_type(struct p9_stat *mistat) 48static inline int dt_type(struct p9_wstat *mistat)
49{ 49{
50 unsigned long perm = mistat->mode; 50 unsigned long perm = mistat->mode;
51 int rettype = DT_REG; 51 int rettype = DT_REG;
@@ -69,32 +69,58 @@ static inline int dt_type(struct p9_stat *mistat)
69static int v9fs_dir_readdir(struct file *filp, void *dirent, filldir_t filldir) 69static int v9fs_dir_readdir(struct file *filp, void *dirent, filldir_t filldir)
70{ 70{
71 int over; 71 int over;
72 struct p9_wstat st;
73 int err;
72 struct p9_fid *fid; 74 struct p9_fid *fid;
73 struct v9fs_session_info *v9ses; 75 int buflen;
74 struct inode *inode; 76 char *statbuf;
75 struct p9_stat *st; 77 int n, i = 0;
76 78
77 P9_DPRINTK(P9_DEBUG_VFS, "name %s\n", filp->f_path.dentry->d_name.name); 79 P9_DPRINTK(P9_DEBUG_VFS, "name %s\n", filp->f_path.dentry->d_name.name);
78 inode = filp->f_path.dentry->d_inode;
79 v9ses = v9fs_inode2v9ses(inode);
80 fid = filp->private_data; 80 fid = filp->private_data;
81 while ((st = p9_client_dirread(fid, filp->f_pos)) != NULL) {
82 if (IS_ERR(st))
83 return PTR_ERR(st);
84 81
85 over = filldir(dirent, st->name.str, st->name.len, filp->f_pos, 82 buflen = fid->clnt->msize - P9_IOHDRSZ;
86 v9fs_qid2ino(&st->qid), dt_type(st)); 83 statbuf = kmalloc(buflen, GFP_KERNEL);
84 if (!statbuf)
85 return -ENOMEM;
87 86
88 if (over) 87 while (1) {
88 err = v9fs_file_readn(filp, statbuf, NULL, buflen,
89 fid->rdir_fpos);
90 if (err <= 0)
89 break; 91 break;
90 92
91 filp->f_pos += st->size; 93 n = err;
92 kfree(st); 94 while (i < n) {
93 st = NULL; 95 err = p9stat_read(statbuf + i, buflen-i, &st,
96 fid->clnt->dotu);
97 if (err) {
98 P9_DPRINTK(P9_DEBUG_VFS, "returned %d\n", err);
99 err = -EIO;
100 p9stat_free(&st);
101 goto free_and_exit;
102 }
103
104 i += st.size+2;
105 fid->rdir_fpos += st.size+2;
106
107 over = filldir(dirent, st.name, strlen(st.name),
108 filp->f_pos, v9fs_qid2ino(&st.qid), dt_type(&st));
109
110 filp->f_pos += st.size+2;
111
112 p9stat_free(&st);
113
114 if (over) {
115 err = 0;
116 goto free_and_exit;
117 }
118 }
94 } 119 }
95 120
96 kfree(st); 121free_and_exit:
97 return 0; 122 kfree(statbuf);
123 return err;
98} 124}
99 125
100 126
diff --git a/fs/9p/vfs_file.c b/fs/9p/vfs_file.c
index 52944d2249a..041c5269228 100644
--- a/fs/9p/vfs_file.c
+++ b/fs/9p/vfs_file.c
@@ -120,23 +120,72 @@ static int v9fs_file_lock(struct file *filp, int cmd, struct file_lock *fl)
120} 120}
121 121
122/** 122/**
123 * v9fs_file_read - read from a file 123 * v9fs_file_readn - read from a file
124 * @filp: file pointer to read 124 * @filp: file pointer to read
125 * @data: data buffer to read data into 125 * @data: data buffer to read data into
126 * @udata: user data buffer to read data into
126 * @count: size of buffer 127 * @count: size of buffer
127 * @offset: offset at which to read data 128 * @offset: offset at which to read data
128 * 129 *
129 */ 130 */
131
132ssize_t
133v9fs_file_readn(struct file *filp, char *data, char __user *udata, u32 count,
134 u64 offset)
135{
136 int n, total;
137 struct p9_fid *fid = filp->private_data;
138
139 P9_DPRINTK(P9_DEBUG_VFS, "fid %d offset %llu count %d\n", fid->fid,
140 (long long unsigned) offset, count);
141
142 n = 0;
143 total = 0;
144 do {
145 n = p9_client_read(fid, data, udata, offset, count);
146 if (n <= 0)
147 break;
148
149 if (data)
150 data += n;
151 if (udata)
152 udata += n;
153
154 offset += n;
155 count -= n;
156 total += n;
157 } while (count > 0 && n == (fid->clnt->msize - P9_IOHDRSZ));
158
159 if (n < 0)
160 total = n;
161
162 return total;
163}
164
165/**
166 * v9fs_file_read - read from a file
167 * @filp: file pointer to read
168 * @udata: user data buffer to read data into
169 * @count: size of buffer
170 * @offset: offset at which to read data
171 *
172 */
173
130static ssize_t 174static ssize_t
131v9fs_file_read(struct file *filp, char __user * data, size_t count, 175v9fs_file_read(struct file *filp, char __user *udata, size_t count,
132 loff_t * offset) 176 loff_t * offset)
133{ 177{
134 int ret; 178 int ret;
135 struct p9_fid *fid; 179 struct p9_fid *fid;
136 180
137 P9_DPRINTK(P9_DEBUG_VFS, "\n"); 181 P9_DPRINTK(P9_DEBUG_VFS, "count %d offset %lld\n", count, *offset);
138 fid = filp->private_data; 182 fid = filp->private_data;
139 ret = p9_client_uread(fid, data, *offset, count); 183
184 if (count > (fid->clnt->msize - P9_IOHDRSZ))
185 ret = v9fs_file_readn(filp, NULL, udata, count, *offset);
186 else
187 ret = p9_client_read(fid, NULL, udata, *offset, count);
188
140 if (ret > 0) 189 if (ret > 0)
141 *offset += ret; 190 *offset += ret;
142 191
@@ -156,19 +205,38 @@ static ssize_t
156v9fs_file_write(struct file *filp, const char __user * data, 205v9fs_file_write(struct file *filp, const char __user * data,
157 size_t count, loff_t * offset) 206 size_t count, loff_t * offset)
158{ 207{
159 int ret; 208 int n, rsize, total = 0;
160 struct p9_fid *fid; 209 struct p9_fid *fid;
210 struct p9_client *clnt;
161 struct inode *inode = filp->f_path.dentry->d_inode; 211 struct inode *inode = filp->f_path.dentry->d_inode;
212 int origin = *offset;
162 213
163 P9_DPRINTK(P9_DEBUG_VFS, "data %p count %d offset %x\n", data, 214 P9_DPRINTK(P9_DEBUG_VFS, "data %p count %d offset %x\n", data,
164 (int)count, (int)*offset); 215 (int)count, (int)*offset);
165 216
166 fid = filp->private_data; 217 fid = filp->private_data;
167 ret = p9_client_uwrite(fid, data, *offset, count); 218 clnt = fid->clnt;
168 if (ret > 0) { 219
169 invalidate_inode_pages2_range(inode->i_mapping, *offset, 220 rsize = fid->iounit;
170 *offset+ret); 221 if (!rsize || rsize > clnt->msize-P9_IOHDRSZ)
171 *offset += ret; 222 rsize = clnt->msize - P9_IOHDRSZ;
223
224 do {
225 if (count < rsize)
226 rsize = count;
227
228 n = p9_client_write(fid, NULL, data+total, *offset+total,
229 rsize);
230 if (n <= 0)
231 break;
232 count -= n;
233 total += n;
234 } while (count > 0);
235
236 if (total > 0) {
237 invalidate_inode_pages2_range(inode->i_mapping, origin,
238 origin+total);
239 *offset += total;
172 } 240 }
173 241
174 if (*offset > inode->i_size) { 242 if (*offset > inode->i_size) {
@@ -176,7 +244,10 @@ v9fs_file_write(struct file *filp, const char __user * data,
176 inode->i_blocks = (inode->i_size + 512 - 1) >> 9; 244 inode->i_blocks = (inode->i_size + 512 - 1) >> 9;
177 } 245 }
178 246
179 return ret; 247 if (n < 0)
248 return n;
249
250 return total;
180} 251}
181 252
182static const struct file_operations v9fs_cached_file_operations = { 253static const struct file_operations v9fs_cached_file_operations = {
diff --git a/fs/9p/vfs_inode.c b/fs/9p/vfs_inode.c
index e83aa5ebe86..8314d3f43b7 100644
--- a/fs/9p/vfs_inode.c
+++ b/fs/9p/vfs_inode.c
@@ -334,7 +334,7 @@ v9fs_inode_from_fid(struct v9fs_session_info *v9ses, struct p9_fid *fid,
334{ 334{
335 int err, umode; 335 int err, umode;
336 struct inode *ret; 336 struct inode *ret;
337 struct p9_stat *st; 337 struct p9_wstat *st;
338 338
339 ret = NULL; 339 ret = NULL;
340 st = p9_client_stat(fid); 340 st = p9_client_stat(fid);
@@ -417,6 +417,8 @@ v9fs_create(struct v9fs_session_info *v9ses, struct inode *dir,
417 struct p9_fid *dfid, *ofid, *fid; 417 struct p9_fid *dfid, *ofid, *fid;
418 struct inode *inode; 418 struct inode *inode;
419 419
420 P9_DPRINTK(P9_DEBUG_VFS, "name %s\n", dentry->d_name.name);
421
420 err = 0; 422 err = 0;
421 ofid = NULL; 423 ofid = NULL;
422 fid = NULL; 424 fid = NULL;
@@ -424,6 +426,7 @@ v9fs_create(struct v9fs_session_info *v9ses, struct inode *dir,
424 dfid = v9fs_fid_clone(dentry->d_parent); 426 dfid = v9fs_fid_clone(dentry->d_parent);
425 if (IS_ERR(dfid)) { 427 if (IS_ERR(dfid)) {
426 err = PTR_ERR(dfid); 428 err = PTR_ERR(dfid);
429 P9_DPRINTK(P9_DEBUG_VFS, "fid clone failed %d\n", err);
427 dfid = NULL; 430 dfid = NULL;
428 goto error; 431 goto error;
429 } 432 }
@@ -432,18 +435,22 @@ v9fs_create(struct v9fs_session_info *v9ses, struct inode *dir,
432 ofid = p9_client_walk(dfid, 0, NULL, 1); 435 ofid = p9_client_walk(dfid, 0, NULL, 1);
433 if (IS_ERR(ofid)) { 436 if (IS_ERR(ofid)) {
434 err = PTR_ERR(ofid); 437 err = PTR_ERR(ofid);
438 P9_DPRINTK(P9_DEBUG_VFS, "p9_client_walk failed %d\n", err);
435 ofid = NULL; 439 ofid = NULL;
436 goto error; 440 goto error;
437 } 441 }
438 442
439 err = p9_client_fcreate(ofid, name, perm, mode, extension); 443 err = p9_client_fcreate(ofid, name, perm, mode, extension);
440 if (err < 0) 444 if (err < 0) {
445 P9_DPRINTK(P9_DEBUG_VFS, "p9_client_fcreate failed %d\n", err);
441 goto error; 446 goto error;
447 }
442 448
443 /* now walk from the parent so we can get unopened fid */ 449 /* now walk from the parent so we can get unopened fid */
444 fid = p9_client_walk(dfid, 1, &name, 0); 450 fid = p9_client_walk(dfid, 1, &name, 0);
445 if (IS_ERR(fid)) { 451 if (IS_ERR(fid)) {
446 err = PTR_ERR(fid); 452 err = PTR_ERR(fid);
453 P9_DPRINTK(P9_DEBUG_VFS, "p9_client_walk failed %d\n", err);
447 fid = NULL; 454 fid = NULL;
448 goto error; 455 goto error;
449 } else 456 } else
@@ -453,6 +460,7 @@ v9fs_create(struct v9fs_session_info *v9ses, struct inode *dir,
453 inode = v9fs_inode_from_fid(v9ses, fid, dir->i_sb); 460 inode = v9fs_inode_from_fid(v9ses, fid, dir->i_sb);
454 if (IS_ERR(inode)) { 461 if (IS_ERR(inode)) {
455 err = PTR_ERR(inode); 462 err = PTR_ERR(inode);
463 P9_DPRINTK(P9_DEBUG_VFS, "inode creation failed %d\n", err);
456 goto error; 464 goto error;
457 } 465 }
458 466
@@ -734,7 +742,7 @@ v9fs_vfs_getattr(struct vfsmount *mnt, struct dentry *dentry,
734 int err; 742 int err;
735 struct v9fs_session_info *v9ses; 743 struct v9fs_session_info *v9ses;
736 struct p9_fid *fid; 744 struct p9_fid *fid;
737 struct p9_stat *st; 745 struct p9_wstat *st;
738 746
739 P9_DPRINTK(P9_DEBUG_VFS, "dentry: %p\n", dentry); 747 P9_DPRINTK(P9_DEBUG_VFS, "dentry: %p\n", dentry);
740 err = -EPERM; 748 err = -EPERM;
@@ -815,10 +823,9 @@ static int v9fs_vfs_setattr(struct dentry *dentry, struct iattr *iattr)
815 */ 823 */
816 824
817void 825void
818v9fs_stat2inode(struct p9_stat *stat, struct inode *inode, 826v9fs_stat2inode(struct p9_wstat *stat, struct inode *inode,
819 struct super_block *sb) 827 struct super_block *sb)
820{ 828{
821 int n;
822 char ext[32]; 829 char ext[32];
823 struct v9fs_session_info *v9ses = sb->s_fs_info; 830 struct v9fs_session_info *v9ses = sb->s_fs_info;
824 831
@@ -842,11 +849,7 @@ v9fs_stat2inode(struct p9_stat *stat, struct inode *inode,
842 int major = -1; 849 int major = -1;
843 int minor = -1; 850 int minor = -1;
844 851
845 n = stat->extension.len; 852 strncpy(ext, stat->extension, sizeof(ext));
846 if (n > sizeof(ext)-1)
847 n = sizeof(ext)-1;
848 memmove(ext, stat->extension.str, n);
849 ext[n] = 0;
850 sscanf(ext, "%c %u %u", &type, &major, &minor); 853 sscanf(ext, "%c %u %u", &type, &major, &minor);
851 switch (type) { 854 switch (type) {
852 case 'c': 855 case 'c':
@@ -857,10 +860,11 @@ v9fs_stat2inode(struct p9_stat *stat, struct inode *inode,
857 break; 860 break;
858 default: 861 default:
859 P9_DPRINTK(P9_DEBUG_ERROR, 862 P9_DPRINTK(P9_DEBUG_ERROR,
860 "Unknown special type %c (%.*s)\n", type, 863 "Unknown special type %c %s\n", type,
861 stat->extension.len, stat->extension.str); 864 stat->extension);
862 }; 865 };
863 inode->i_rdev = MKDEV(major, minor); 866 inode->i_rdev = MKDEV(major, minor);
867 init_special_inode(inode, inode->i_mode, inode->i_rdev);
864 } else 868 } else
865 inode->i_rdev = 0; 869 inode->i_rdev = 0;
866 870
@@ -904,7 +908,7 @@ static int v9fs_readlink(struct dentry *dentry, char *buffer, int buflen)
904 908
905 struct v9fs_session_info *v9ses; 909 struct v9fs_session_info *v9ses;
906 struct p9_fid *fid; 910 struct p9_fid *fid;
907 struct p9_stat *st; 911 struct p9_wstat *st;
908 912
909 P9_DPRINTK(P9_DEBUG_VFS, " %s\n", dentry->d_name.name); 913 P9_DPRINTK(P9_DEBUG_VFS, " %s\n", dentry->d_name.name);
910 retval = -EPERM; 914 retval = -EPERM;
@@ -926,15 +930,10 @@ static int v9fs_readlink(struct dentry *dentry, char *buffer, int buflen)
926 } 930 }
927 931
928 /* copy extension buffer into buffer */ 932 /* copy extension buffer into buffer */
929 if (st->extension.len < buflen) 933 strncpy(buffer, st->extension, buflen);
930 buflen = st->extension.len + 1;
931
932 memmove(buffer, st->extension.str, buflen - 1);
933 buffer[buflen-1] = 0;
934 934
935 P9_DPRINTK(P9_DEBUG_VFS, 935 P9_DPRINTK(P9_DEBUG_VFS,
936 "%s -> %.*s (%s)\n", dentry->d_name.name, st->extension.len, 936 "%s -> %s (%s)\n", dentry->d_name.name, st->extension, buffer);
937 st->extension.str, buffer);
938 937
939 retval = buflen; 938 retval = buflen;
940 939
diff --git a/fs/9p/vfs_super.c b/fs/9p/vfs_super.c
index bf59c396049..d6cb1a0ca72 100644
--- a/fs/9p/vfs_super.c
+++ b/fs/9p/vfs_super.c
@@ -111,7 +111,7 @@ static int v9fs_get_sb(struct file_system_type *fs_type, int flags,
111 struct inode *inode = NULL; 111 struct inode *inode = NULL;
112 struct dentry *root = NULL; 112 struct dentry *root = NULL;
113 struct v9fs_session_info *v9ses = NULL; 113 struct v9fs_session_info *v9ses = NULL;
114 struct p9_stat *st = NULL; 114 struct p9_wstat *st = NULL;
115 int mode = S_IRWXUGO | S_ISVTX; 115 int mode = S_IRWXUGO | S_ISVTX;
116 uid_t uid = current->fsuid; 116 uid_t uid = current->fsuid;
117 gid_t gid = current->fsgid; 117 gid_t gid = current->fsgid;
@@ -161,10 +161,14 @@ static int v9fs_get_sb(struct file_system_type *fs_type, int flags,
161 161
162 sb->s_root = root; 162 sb->s_root = root;
163 root->d_inode->i_ino = v9fs_qid2ino(&st->qid); 163 root->d_inode->i_ino = v9fs_qid2ino(&st->qid);
164
164 v9fs_stat2inode(st, root->d_inode, sb); 165 v9fs_stat2inode(st, root->d_inode, sb);
166
165 v9fs_fid_add(root, fid); 167 v9fs_fid_add(root, fid);
168 p9stat_free(st);
166 kfree(st); 169 kfree(st);
167 170
171P9_DPRINTK(P9_DEBUG_VFS, " return simple set mount\n");
168 return simple_set_mnt(mnt, sb); 172 return simple_set_mnt(mnt, sb);
169 173
170release_sb: 174release_sb:
diff --git a/fs/Kconfig b/fs/Kconfig
index d0a1174fb51..e46297f020c 100644
--- a/fs/Kconfig
+++ b/fs/Kconfig
@@ -6,61 +6,9 @@ menu "File systems"
6 6
7if BLOCK 7if BLOCK
8 8
9config EXT2_FS 9source "fs/ext2/Kconfig"
10 tristate "Second extended fs support" 10source "fs/ext3/Kconfig"
11 help 11source "fs/ext4/Kconfig"
12 Ext2 is a standard Linux file system for hard disks.
13
14 To compile this file system support as a module, choose M here: the
15 module will be called ext2.
16
17 If unsure, say Y.
18
19config EXT2_FS_XATTR
20 bool "Ext2 extended attributes"
21 depends on EXT2_FS
22 help
23 Extended attributes are name:value pairs associated with inodes by
24 the kernel or by users (see the attr(5) manual page, or visit
25 <http://acl.bestbits.at/> for details).
26
27 If unsure, say N.
28
29config EXT2_FS_POSIX_ACL
30 bool "Ext2 POSIX Access Control Lists"
31 depends on EXT2_FS_XATTR
32 select FS_POSIX_ACL
33 help
34 Posix Access Control Lists (ACLs) support permissions for users and
35 groups beyond the owner/group/world scheme.
36
37 To learn more about Access Control Lists, visit the Posix ACLs for
38 Linux website <http://acl.bestbits.at/>.
39
40 If you don't know what Access Control Lists are, say N
41
42config EXT2_FS_SECURITY
43 bool "Ext2 Security Labels"
44 depends on EXT2_FS_XATTR
45 help
46 Security labels support alternative access control models
47 implemented by security modules like SELinux. This option
48 enables an extended attribute handler for file security
49 labels in the ext2 filesystem.
50
51 If you are not using a security module that requires using
52 extended attributes for file security labels, say N.
53
54config EXT2_FS_XIP
55 bool "Ext2 execute in place support"
56 depends on EXT2_FS && MMU
57 help
58 Execute in place can be used on memory-backed block devices. If you
59 enable this option, you can select to mount block devices which are
60 capable of this feature without using the page cache.
61
62 If you do not use a block device that is capable of using this,
63 or if unsure, say N.
64 12
65config FS_XIP 13config FS_XIP
66# execute in place 14# execute in place
@@ -68,218 +16,8 @@ config FS_XIP
68 depends on EXT2_FS_XIP 16 depends on EXT2_FS_XIP
69 default y 17 default y
70 18
71config EXT3_FS 19source "fs/jbd/Kconfig"
72 tristate "Ext3 journalling file system support" 20source "fs/jbd2/Kconfig"
73 select JBD
74 help
75 This is the journalling version of the Second extended file system
76 (often called ext3), the de facto standard Linux file system
77 (method to organize files on a storage device) for hard disks.
78
79 The journalling code included in this driver means you do not have
80 to run e2fsck (file system checker) on your file systems after a
81 crash. The journal keeps track of any changes that were being made
82 at the time the system crashed, and can ensure that your file system
83 is consistent without the need for a lengthy check.
84
85 Other than adding the journal to the file system, the on-disk format
86 of ext3 is identical to ext2. It is possible to freely switch
87 between using the ext3 driver and the ext2 driver, as long as the
88 file system has been cleanly unmounted, or e2fsck is run on the file
89 system.
90
91 To add a journal on an existing ext2 file system or change the
92 behavior of ext3 file systems, you can use the tune2fs utility ("man
93 tune2fs"). To modify attributes of files and directories on ext3
94 file systems, use chattr ("man chattr"). You need to be using
95 e2fsprogs version 1.20 or later in order to create ext3 journals
96 (available at <http://sourceforge.net/projects/e2fsprogs/>).
97
98 To compile this file system support as a module, choose M here: the
99 module will be called ext3.
100
101config EXT3_FS_XATTR
102 bool "Ext3 extended attributes"
103 depends on EXT3_FS
104 default y
105 help
106 Extended attributes are name:value pairs associated with inodes by
107 the kernel or by users (see the attr(5) manual page, or visit
108 <http://acl.bestbits.at/> for details).
109
110 If unsure, say N.
111
112 You need this for POSIX ACL support on ext3.
113
114config EXT3_FS_POSIX_ACL
115 bool "Ext3 POSIX Access Control Lists"
116 depends on EXT3_FS_XATTR
117 select FS_POSIX_ACL
118 help
119 Posix Access Control Lists (ACLs) support permissions for users and
120 groups beyond the owner/group/world scheme.
121
122 To learn more about Access Control Lists, visit the Posix ACLs for
123 Linux website <http://acl.bestbits.at/>.
124
125 If you don't know what Access Control Lists are, say N
126
127config EXT3_FS_SECURITY
128 bool "Ext3 Security Labels"
129 depends on EXT3_FS_XATTR
130 help
131 Security labels support alternative access control models
132 implemented by security modules like SELinux. This option
133 enables an extended attribute handler for file security
134 labels in the ext3 filesystem.
135
136 If you are not using a security module that requires using
137 extended attributes for file security labels, say N.
138
139config EXT4_FS
140 tristate "The Extended 4 (ext4) filesystem"
141 select JBD2
142 select CRC16
143 help
144 This is the next generation of the ext3 filesystem.
145
146 Unlike the change from ext2 filesystem to ext3 filesystem,
147 the on-disk format of ext4 is not forwards compatible with
148 ext3; it is based on extent maps and it supports 48-bit
149 physical block numbers. The ext4 filesystem also supports delayed
150 allocation, persistent preallocation, high resolution time stamps,
151 and a number of other features to improve performance and speed
152 up fsck time. For more information, please see the web pages at
153 http://ext4.wiki.kernel.org.
154
155 The ext4 filesystem will support mounting an ext3
156 filesystem; while there will be some performance gains from
157 the delayed allocation and inode table readahead, the best
158 performance gains will require enabling ext4 features in the
159 filesystem, or formating a new filesystem as an ext4
160 filesystem initially.
161
162 To compile this file system support as a module, choose M here. The
163 module will be called ext4.
164
165 If unsure, say N.
166
167config EXT4DEV_COMPAT
168 bool "Enable ext4dev compatibility"
169 depends on EXT4_FS
170 help
171 Starting with 2.6.28, the name of the ext4 filesystem was
172 renamed from ext4dev to ext4. Unfortunately there are some
173 legacy userspace programs (such as klibc's fstype) have
174 "ext4dev" hardcoded.
175
176 To enable backwards compatibility so that systems that are
177 still expecting to mount ext4 filesystems using ext4dev,
178 chose Y here. This feature will go away by 2.6.31, so
179 please arrange to get your userspace programs fixed!
180
181config EXT4_FS_XATTR
182 bool "Ext4 extended attributes"
183 depends on EXT4_FS
184 default y
185 help
186 Extended attributes are name:value pairs associated with inodes by
187 the kernel or by users (see the attr(5) manual page, or visit
188 <http://acl.bestbits.at/> for details).
189
190 If unsure, say N.
191
192 You need this for POSIX ACL support on ext4.
193
194config EXT4_FS_POSIX_ACL
195 bool "Ext4 POSIX Access Control Lists"
196 depends on EXT4_FS_XATTR
197 select FS_POSIX_ACL
198 help
199 POSIX Access Control Lists (ACLs) support permissions for users and
200 groups beyond the owner/group/world scheme.
201
202 To learn more about Access Control Lists, visit the POSIX ACLs for
203 Linux website <http://acl.bestbits.at/>.
204
205 If you don't know what Access Control Lists are, say N
206
207config EXT4_FS_SECURITY
208 bool "Ext4 Security Labels"
209 depends on EXT4_FS_XATTR
210 help
211 Security labels support alternative access control models
212 implemented by security modules like SELinux. This option
213 enables an extended attribute handler for file security
214 labels in the ext4 filesystem.
215
216 If you are not using a security module that requires using
217 extended attributes for file security labels, say N.
218
219config JBD
220 tristate
221 help
222 This is a generic journalling layer for block devices. It is
223 currently used by the ext3 file system, but it could also be
224 used to add journal support to other file systems or block
225 devices such as RAID or LVM.
226
227 If you are using the ext3 file system, you need to say Y here.
228 If you are not using ext3 then you will probably want to say N.
229
230 To compile this device as a module, choose M here: the module will be
231 called jbd. If you are compiling ext3 into the kernel, you
232 cannot compile this code as a module.
233
234config JBD_DEBUG
235 bool "JBD (ext3) debugging support"
236 depends on JBD && DEBUG_FS
237 help
238 If you are using the ext3 journaled file system (or potentially any
239 other file system/device using JBD), this option allows you to
240 enable debugging output while the system is running, in order to
241 help track down any problems you are having. By default the
242 debugging output will be turned off.
243
244 If you select Y here, then you will be able to turn on debugging
245 with "echo N > /sys/kernel/debug/jbd/jbd-debug", where N is a
246 number between 1 and 5, the higher the number, the more debugging
247 output is generated. To turn debugging off again, do
248 "echo 0 > /sys/kernel/debug/jbd/jbd-debug".
249
250config JBD2
251 tristate
252 select CRC32
253 help
254 This is a generic journaling layer for block devices that support
255 both 32-bit and 64-bit block numbers. It is currently used by
256 the ext4 and OCFS2 filesystems, but it could also be used to add
257 journal support to other file systems or block devices such
258 as RAID or LVM.
259
260 If you are using ext4 or OCFS2, you need to say Y here.
261 If you are not using ext4 or OCFS2 then you will
262 probably want to say N.
263
264 To compile this device as a module, choose M here. The module will be
265 called jbd2. If you are compiling ext4 or OCFS2 into the kernel,
266 you cannot compile this code as a module.
267
268config JBD2_DEBUG
269 bool "JBD2 (ext4) debugging support"
270 depends on JBD2 && DEBUG_FS
271 help
272 If you are using the ext4 journaled file system (or
273 potentially any other filesystem/device using JBD2), this option
274 allows you to enable debugging output while the system is running,
275 in order to help track down any problems you are having.
276 By default, the debugging output will be turned off.
277
278 If you select Y here, then you will be able to turn on debugging
279 with "echo N > /sys/kernel/debug/jbd2/jbd2-debug", where N is a
280 number between 1 and 5. The higher the number, the more debugging
281 output is generated. To turn debugging off again, do
282 "echo 0 > /sys/kernel/debug/jbd2/jbd2-debug".
283 21
284config FS_MBCACHE 22config FS_MBCACHE
285# Meta block cache for Extended Attributes (ext2/ext3/ext4) 23# Meta block cache for Extended Attributes (ext2/ext3/ext4)
@@ -665,7 +403,7 @@ config AUTOFS4_FS
665 N here. 403 N here.
666 404
667config FUSE_FS 405config FUSE_FS
668 tristate "Filesystem in Userspace support" 406 tristate "FUSE (Filesystem in Userspace) support"
669 help 407 help
670 With FUSE it is possible to implement a fully functional filesystem 408 With FUSE it is possible to implement a fully functional filesystem
671 in a userspace program. 409 in a userspace program.
@@ -1168,195 +906,7 @@ config EFS_FS
1168 To compile the EFS file system support as a module, choose M here: the 906 To compile the EFS file system support as a module, choose M here: the
1169 module will be called efs. 907 module will be called efs.
1170 908
1171config JFFS2_FS 909source "fs/jffs2/Kconfig"
1172 tristate "Journalling Flash File System v2 (JFFS2) support"
1173 select CRC32
1174 depends on MTD
1175 help
1176 JFFS2 is the second generation of the Journalling Flash File System
1177 for use on diskless embedded devices. It provides improved wear
1178 levelling, compression and support for hard links. You cannot use
1179 this on normal block devices, only on 'MTD' devices.
1180
1181 Further information on the design and implementation of JFFS2 is
1182 available at <http://sources.redhat.com/jffs2/>.
1183
1184config JFFS2_FS_DEBUG
1185 int "JFFS2 debugging verbosity (0 = quiet, 2 = noisy)"
1186 depends on JFFS2_FS
1187 default "0"
1188 help
1189 This controls the amount of debugging messages produced by the JFFS2
1190 code. Set it to zero for use in production systems. For evaluation,
1191 testing and debugging, it's advisable to set it to one. This will
1192 enable a few assertions and will print debugging messages at the
1193 KERN_DEBUG loglevel, where they won't normally be visible. Level 2
1194 is unlikely to be useful - it enables extra debugging in certain
1195 areas which at one point needed debugging, but when the bugs were
1196 located and fixed, the detailed messages were relegated to level 2.
1197
1198 If reporting bugs, please try to have available a full dump of the
1199 messages at debug level 1 while the misbehaviour was occurring.
1200
1201config JFFS2_FS_WRITEBUFFER
1202 bool "JFFS2 write-buffering support"
1203 depends on JFFS2_FS
1204 default y
1205 help
1206 This enables the write-buffering support in JFFS2.
1207
1208 This functionality is required to support JFFS2 on the following
1209 types of flash devices:
1210 - NAND flash
1211 - NOR flash with transparent ECC
1212 - DataFlash
1213
1214config JFFS2_FS_WBUF_VERIFY
1215 bool "Verify JFFS2 write-buffer reads"
1216 depends on JFFS2_FS_WRITEBUFFER
1217 default n
1218 help
1219 This causes JFFS2 to read back every page written through the
1220 write-buffer, and check for errors.
1221
1222config JFFS2_SUMMARY
1223 bool "JFFS2 summary support (EXPERIMENTAL)"
1224 depends on JFFS2_FS && EXPERIMENTAL
1225 default n
1226 help
1227 This feature makes it possible to use summary information
1228 for faster filesystem mount.
1229
1230 The summary information can be inserted into a filesystem image
1231 by the utility 'sumtool'.
1232
1233 If unsure, say 'N'.
1234
1235config JFFS2_FS_XATTR
1236 bool "JFFS2 XATTR support (EXPERIMENTAL)"
1237 depends on JFFS2_FS && EXPERIMENTAL
1238 default n
1239 help
1240 Extended attributes are name:value pairs associated with inodes by
1241 the kernel or by users (see the attr(5) manual page, or visit
1242 <http://acl.bestbits.at/> for details).
1243
1244 If unsure, say N.
1245
1246config JFFS2_FS_POSIX_ACL
1247 bool "JFFS2 POSIX Access Control Lists"
1248 depends on JFFS2_FS_XATTR
1249 default y
1250 select FS_POSIX_ACL
1251 help
1252 Posix Access Control Lists (ACLs) support permissions for users and
1253 groups beyond the owner/group/world scheme.
1254
1255 To learn more about Access Control Lists, visit the Posix ACLs for
1256 Linux website <http://acl.bestbits.at/>.
1257
1258 If you don't know what Access Control Lists are, say N
1259
1260config JFFS2_FS_SECURITY
1261 bool "JFFS2 Security Labels"
1262 depends on JFFS2_FS_XATTR
1263 default y
1264 help
1265 Security labels support alternative access control models
1266 implemented by security modules like SELinux. This option
1267 enables an extended attribute handler for file security
1268 labels in the jffs2 filesystem.
1269
1270 If you are not using a security module that requires using
1271 extended attributes for file security labels, say N.
1272
1273config JFFS2_COMPRESSION_OPTIONS
1274 bool "Advanced compression options for JFFS2"
1275 depends on JFFS2_FS
1276 default n
1277 help
1278 Enabling this option allows you to explicitly choose which
1279 compression modules, if any, are enabled in JFFS2. Removing
1280 compressors can mean you cannot read existing file systems,
1281 and enabling experimental compressors can mean that you
1282 write a file system which cannot be read by a standard kernel.
1283
1284 If unsure, you should _definitely_ say 'N'.
1285
1286config JFFS2_ZLIB
1287 bool "JFFS2 ZLIB compression support" if JFFS2_COMPRESSION_OPTIONS
1288 select ZLIB_INFLATE
1289 select ZLIB_DEFLATE
1290 depends on JFFS2_FS
1291 default y
1292 help
1293 Zlib is designed to be a free, general-purpose, legally unencumbered,
1294 lossless data-compression library for use on virtually any computer
1295 hardware and operating system. See <http://www.gzip.org/zlib/> for
1296 further information.
1297
1298 Say 'Y' if unsure.
1299
1300config JFFS2_LZO
1301 bool "JFFS2 LZO compression support" if JFFS2_COMPRESSION_OPTIONS
1302 select LZO_COMPRESS
1303 select LZO_DECOMPRESS
1304 depends on JFFS2_FS
1305 default n
1306 help
1307 minilzo-based compression. Generally works better than Zlib.
1308
1309 This feature was added in July, 2007. Say 'N' if you need
1310 compatibility with older bootloaders or kernels.
1311
1312config JFFS2_RTIME
1313 bool "JFFS2 RTIME compression support" if JFFS2_COMPRESSION_OPTIONS
1314 depends on JFFS2_FS
1315 default y
1316 help
1317 Rtime does manage to recompress already-compressed data. Say 'Y' if unsure.
1318
1319config JFFS2_RUBIN
1320 bool "JFFS2 RUBIN compression support" if JFFS2_COMPRESSION_OPTIONS
1321 depends on JFFS2_FS
1322 default n
1323 help
1324 RUBINMIPS and DYNRUBIN compressors. Say 'N' if unsure.
1325
1326choice
1327 prompt "JFFS2 default compression mode" if JFFS2_COMPRESSION_OPTIONS
1328 default JFFS2_CMODE_PRIORITY
1329 depends on JFFS2_FS
1330 help
1331 You can set here the default compression mode of JFFS2 from
1332 the available compression modes. Don't touch if unsure.
1333
1334config JFFS2_CMODE_NONE
1335 bool "no compression"
1336 help
1337 Uses no compression.
1338
1339config JFFS2_CMODE_PRIORITY
1340 bool "priority"
1341 help
1342 Tries the compressors in a predefined order and chooses the first
1343 successful one.
1344
1345config JFFS2_CMODE_SIZE
1346 bool "size (EXPERIMENTAL)"
1347 help
1348 Tries all compressors and chooses the one which has the smallest
1349 result.
1350
1351config JFFS2_CMODE_FAVOURLZO
1352 bool "Favour LZO"
1353 help
1354 Tries all compressors and chooses the one which has the smallest
1355 result but gives some preference to LZO (which has faster
1356 decompression) at the expense of size.
1357
1358endchoice
1359
1360# UBIFS File system configuration 910# UBIFS File system configuration
1361source "fs/ubifs/Kconfig" 911source "fs/ubifs/Kconfig"
1362 912
@@ -1913,148 +1463,7 @@ config SMB_NLS_REMOTE
1913 1463
1914 smbmount from samba 2.2.0 or later supports this. 1464 smbmount from samba 2.2.0 or later supports this.
1915 1465
1916config CIFS 1466source "fs/cifs/Kconfig"
1917 tristate "CIFS support (advanced network filesystem, SMBFS successor)"
1918 depends on INET
1919 select NLS
1920 help
1921 This is the client VFS module for the Common Internet File System
1922 (CIFS) protocol which is the successor to the Server Message Block
1923 (SMB) protocol, the native file sharing mechanism for most early
1924 PC operating systems. The CIFS protocol is fully supported by
1925 file servers such as Windows 2000 (including Windows 2003, NT 4
1926 and Windows XP) as well by Samba (which provides excellent CIFS
1927 server support for Linux and many other operating systems). Limited
1928 support for OS/2 and Windows ME and similar servers is provided as
1929 well.
1930
1931 The cifs module provides an advanced network file system
1932 client for mounting to CIFS compliant servers. It includes
1933 support for DFS (hierarchical name space), secure per-user
1934 session establishment via Kerberos or NTLM or NTLMv2,
1935 safe distributed caching (oplock), optional packet
1936 signing, Unicode and other internationalization improvements.
1937 If you need to mount to Samba or Windows from this machine, say Y.
1938
1939config CIFS_STATS
1940 bool "CIFS statistics"
1941 depends on CIFS
1942 help
1943 Enabling this option will cause statistics for each server share
1944 mounted by the cifs client to be displayed in /proc/fs/cifs/Stats
1945
1946config CIFS_STATS2
1947 bool "Extended statistics"
1948 depends on CIFS_STATS
1949 help
1950 Enabling this option will allow more detailed statistics on SMB
1951 request timing to be displayed in /proc/fs/cifs/DebugData and also
1952 allow optional logging of slow responses to dmesg (depending on the
1953 value of /proc/fs/cifs/cifsFYI, see fs/cifs/README for more details).
1954 These additional statistics may have a minor effect on performance
1955 and memory utilization.
1956
1957 Unless you are a developer or are doing network performance analysis
1958 or tuning, say N.
1959
1960config CIFS_WEAK_PW_HASH
1961 bool "Support legacy servers which use weaker LANMAN security"
1962 depends on CIFS
1963 help
1964 Modern CIFS servers including Samba and most Windows versions
1965 (since 1997) support stronger NTLM (and even NTLMv2 and Kerberos)
1966 security mechanisms. These hash the password more securely
1967 than the mechanisms used in the older LANMAN version of the
1968 SMB protocol but LANMAN based authentication is needed to
1969 establish sessions with some old SMB servers.
1970
1971 Enabling this option allows the cifs module to mount to older
1972 LANMAN based servers such as OS/2 and Windows 95, but such
1973 mounts may be less secure than mounts using NTLM or more recent
1974 security mechanisms if you are on a public network. Unless you
1975 have a need to access old SMB servers (and are on a private
1976 network) you probably want to say N. Even if this support
1977 is enabled in the kernel build, LANMAN authentication will not be
1978 used automatically. At runtime LANMAN mounts are disabled but
1979 can be set to required (or optional) either in
1980 /proc/fs/cifs (see fs/cifs/README for more detail) or via an
1981 option on the mount command. This support is disabled by
1982 default in order to reduce the possibility of a downgrade
1983 attack.
1984
1985 If unsure, say N.
1986
1987config CIFS_UPCALL
1988 bool "Kerberos/SPNEGO advanced session setup"
1989 depends on CIFS && KEYS
1990 help
1991 Enables an upcall mechanism for CIFS which accesses
1992 userspace helper utilities to provide SPNEGO packaged (RFC 4178)
1993 Kerberos tickets which are needed to mount to certain secure servers
1994 (for which more secure Kerberos authentication is required). If
1995 unsure, say N.
1996
1997config CIFS_XATTR
1998 bool "CIFS extended attributes"
1999 depends on CIFS
2000 help
2001 Extended attributes are name:value pairs associated with inodes by
2002 the kernel or by users (see the attr(5) manual page, or visit
2003 <http://acl.bestbits.at/> for details). CIFS maps the name of
2004 extended attributes beginning with the user namespace prefix
2005 to SMB/CIFS EAs. EAs are stored on Windows servers without the
2006 user namespace prefix, but their names are seen by Linux cifs clients
2007 prefaced by the user namespace prefix. The system namespace
2008 (used by some filesystems to store ACLs) is not supported at
2009 this time.
2010
2011 If unsure, say N.
2012
2013config CIFS_POSIX
2014 bool "CIFS POSIX Extensions"
2015 depends on CIFS_XATTR
2016 help
2017 Enabling this option will cause the cifs client to attempt to
2018 negotiate a newer dialect with servers, such as Samba 3.0.5
2019 or later, that optionally can handle more POSIX like (rather
2020 than Windows like) file behavior. It also enables
2021 support for POSIX ACLs (getfacl and setfacl) to servers
2022 (such as Samba 3.10 and later) which can negotiate
2023 CIFS POSIX ACL support. If unsure, say N.
2024
2025config CIFS_DEBUG2
2026 bool "Enable additional CIFS debugging routines"
2027 depends on CIFS
2028 help
2029 Enabling this option adds a few more debugging routines
2030 to the cifs code which slightly increases the size of
2031 the cifs module and can cause additional logging of debug
2032 messages in some error paths, slowing performance. This
2033 option can be turned off unless you are debugging
2034 cifs problems. If unsure, say N.
2035
2036config CIFS_EXPERIMENTAL
2037 bool "CIFS Experimental Features (EXPERIMENTAL)"
2038 depends on CIFS && EXPERIMENTAL
2039 help
2040 Enables cifs features under testing. These features are
2041 experimental and currently include DFS support and directory
2042 change notification ie fcntl(F_DNOTIFY), as well as the upcall
2043 mechanism which will be used for Kerberos session negotiation
2044 and uid remapping. Some of these features also may depend on
2045 setting a value of 1 to the pseudo-file /proc/fs/cifs/Experimental
2046 (which is disabled by default). See the file fs/cifs/README
2047 for more details. If unsure, say N.
2048
2049config CIFS_DFS_UPCALL
2050 bool "DFS feature support (EXPERIMENTAL)"
2051 depends on CIFS_EXPERIMENTAL
2052 depends on KEYS
2053 help
2054 Enables an upcall mechanism for CIFS which contacts userspace
2055 helper utilities to provide server name resolution (host names to
2056 IP addresses) which is needed for implicit mounts of DFS junction
2057 points. If unsure, say N.
2058 1467
2059config NCP_FS 1468config NCP_FS
2060 tristate "NCP file system support (to mount NetWare volumes)" 1469 tristate "NCP file system support (to mount NetWare volumes)"
diff --git a/fs/Kconfig.binfmt b/fs/Kconfig.binfmt
index 801db134181..ce9fb3fbfae 100644
--- a/fs/Kconfig.binfmt
+++ b/fs/Kconfig.binfmt
@@ -40,6 +40,28 @@ config BINFMT_ELF_FDPIC
40 40
41 It is also possible to run FDPIC ELF binaries on MMU linux also. 41 It is also possible to run FDPIC ELF binaries on MMU linux also.
42 42
43config CORE_DUMP_DEFAULT_ELF_HEADERS
44 bool "Write ELF core dumps with partial segments"
45 default n
46 depends on BINFMT_ELF
47 help
48 ELF core dump files describe each memory mapping of the crashed
49 process, and can contain or omit the memory contents of each one.
50 The contents of an unmodified text mapping are omitted by default.
51
52 For an unmodified text mapping of an ELF object, including just
53 the first page of the file in a core dump makes it possible to
54 identify the build ID bits in the file, without paying the i/o
55 cost and disk space to dump all the text. However, versions of
56 GDB before 6.7 are confused by ELF core dump files in this format.
57
58 The core dump behavior can be controlled per process using
59 the /proc/PID/coredump_filter pseudo-file; this setting is
60 inherited. See Documentation/filesystems/proc.txt for details.
61
62 This config option changes the default setting of coredump_filter
63 seen at boot time. If unsure, say N.
64
43config BINFMT_FLAT 65config BINFMT_FLAT
44 bool "Kernel support for flat binaries" 66 bool "Kernel support for flat binaries"
45 depends on !MMU && (!FRV || BROKEN) 67 depends on !MMU && (!FRV || BROKEN)
diff --git a/fs/binfmt_elf.c b/fs/binfmt_elf.c
index c76afa26edf..8fcfa398d35 100644
--- a/fs/binfmt_elf.c
+++ b/fs/binfmt_elf.c
@@ -1156,16 +1156,24 @@ static int dump_seek(struct file *file, loff_t off)
1156static unsigned long vma_dump_size(struct vm_area_struct *vma, 1156static unsigned long vma_dump_size(struct vm_area_struct *vma,
1157 unsigned long mm_flags) 1157 unsigned long mm_flags)
1158{ 1158{
1159#define FILTER(type) (mm_flags & (1UL << MMF_DUMP_##type))
1160
1159 /* The vma can be set up to tell us the answer directly. */ 1161 /* The vma can be set up to tell us the answer directly. */
1160 if (vma->vm_flags & VM_ALWAYSDUMP) 1162 if (vma->vm_flags & VM_ALWAYSDUMP)
1161 goto whole; 1163 goto whole;
1162 1164
1165 /* Hugetlb memory check */
1166 if (vma->vm_flags & VM_HUGETLB) {
1167 if ((vma->vm_flags & VM_SHARED) && FILTER(HUGETLB_SHARED))
1168 goto whole;
1169 if (!(vma->vm_flags & VM_SHARED) && FILTER(HUGETLB_PRIVATE))
1170 goto whole;
1171 }
1172
1163 /* Do not dump I/O mapped devices or special mappings */ 1173 /* Do not dump I/O mapped devices or special mappings */
1164 if (vma->vm_flags & (VM_IO | VM_RESERVED)) 1174 if (vma->vm_flags & (VM_IO | VM_RESERVED))
1165 return 0; 1175 return 0;
1166 1176
1167#define FILTER(type) (mm_flags & (1UL << MMF_DUMP_##type))
1168
1169 /* By default, dump shared memory if mapped from an anonymous file. */ 1177 /* By default, dump shared memory if mapped from an anonymous file. */
1170 if (vma->vm_flags & VM_SHARED) { 1178 if (vma->vm_flags & VM_SHARED) {
1171 if (vma->vm_file->f_path.dentry->d_inode->i_nlink == 0 ? 1179 if (vma->vm_file->f_path.dentry->d_inode->i_nlink == 0 ?
@@ -1333,20 +1341,15 @@ static void fill_prstatus(struct elf_prstatus *prstatus,
1333 prstatus->pr_pgrp = task_pgrp_vnr(p); 1341 prstatus->pr_pgrp = task_pgrp_vnr(p);
1334 prstatus->pr_sid = task_session_vnr(p); 1342 prstatus->pr_sid = task_session_vnr(p);
1335 if (thread_group_leader(p)) { 1343 if (thread_group_leader(p)) {
1344 struct task_cputime cputime;
1345
1336 /* 1346 /*
1337 * This is the record for the group leader. Add in the 1347 * This is the record for the group leader. It shows the
1338 * cumulative times of previous dead threads. This total 1348 * group-wide total, not its individual thread total.
1339 * won't include the time of each live thread whose state
1340 * is included in the core dump. The final total reported
1341 * to our parent process when it calls wait4 will include
1342 * those sums as well as the little bit more time it takes
1343 * this and each other thread to finish dying after the
1344 * core dump synchronization phase.
1345 */ 1349 */
1346 cputime_to_timeval(cputime_add(p->utime, p->signal->utime), 1350 thread_group_cputime(p, &cputime);
1347 &prstatus->pr_utime); 1351 cputime_to_timeval(cputime.utime, &prstatus->pr_utime);
1348 cputime_to_timeval(cputime_add(p->stime, p->signal->stime), 1352 cputime_to_timeval(cputime.stime, &prstatus->pr_stime);
1349 &prstatus->pr_stime);
1350 } else { 1353 } else {
1351 cputime_to_timeval(p->utime, &prstatus->pr_utime); 1354 cputime_to_timeval(p->utime, &prstatus->pr_utime);
1352 cputime_to_timeval(p->stime, &prstatus->pr_stime); 1355 cputime_to_timeval(p->stime, &prstatus->pr_stime);
diff --git a/fs/binfmt_elf_fdpic.c b/fs/binfmt_elf_fdpic.c
index 0e8367c5462..5b5424cb339 100644
--- a/fs/binfmt_elf_fdpic.c
+++ b/fs/binfmt_elf_fdpic.c
@@ -1390,20 +1390,15 @@ static void fill_prstatus(struct elf_prstatus *prstatus,
1390 prstatus->pr_pgrp = task_pgrp_vnr(p); 1390 prstatus->pr_pgrp = task_pgrp_vnr(p);
1391 prstatus->pr_sid = task_session_vnr(p); 1391 prstatus->pr_sid = task_session_vnr(p);
1392 if (thread_group_leader(p)) { 1392 if (thread_group_leader(p)) {
1393 struct task_cputime cputime;
1394
1393 /* 1395 /*
1394 * This is the record for the group leader. Add in the 1396 * This is the record for the group leader. It shows the
1395 * cumulative times of previous dead threads. This total 1397 * group-wide total, not its individual thread total.
1396 * won't include the time of each live thread whose state
1397 * is included in the core dump. The final total reported
1398 * to our parent process when it calls wait4 will include
1399 * those sums as well as the little bit more time it takes
1400 * this and each other thread to finish dying after the
1401 * core dump synchronization phase.
1402 */ 1398 */
1403 cputime_to_timeval(cputime_add(p->utime, p->signal->utime), 1399 thread_group_cputime(p, &cputime);
1404 &prstatus->pr_utime); 1400 cputime_to_timeval(cputime.utime, &prstatus->pr_utime);
1405 cputime_to_timeval(cputime_add(p->stime, p->signal->stime), 1401 cputime_to_timeval(cputime.stime, &prstatus->pr_stime);
1406 &prstatus->pr_stime);
1407 } else { 1402 } else {
1408 cputime_to_timeval(p->utime, &prstatus->pr_utime); 1403 cputime_to_timeval(p->utime, &prstatus->pr_utime);
1409 cputime_to_timeval(p->stime, &prstatus->pr_stime); 1404 cputime_to_timeval(p->stime, &prstatus->pr_stime);
diff --git a/fs/buffer.c b/fs/buffer.c
index ac78d4c19b3..6569fda5cfe 100644
--- a/fs/buffer.c
+++ b/fs/buffer.c
@@ -76,8 +76,7 @@ EXPORT_SYMBOL(__lock_buffer);
76 76
77void unlock_buffer(struct buffer_head *bh) 77void unlock_buffer(struct buffer_head *bh)
78{ 78{
79 smp_mb__before_clear_bit(); 79 clear_bit_unlock(BH_Lock, &bh->b_state);
80 clear_buffer_locked(bh);
81 smp_mb__after_clear_bit(); 80 smp_mb__after_clear_bit();
82 wake_up_bit(&bh->b_state, BH_Lock); 81 wake_up_bit(&bh->b_state, BH_Lock);
83} 82}
diff --git a/fs/cifs/Kconfig b/fs/cifs/Kconfig
new file mode 100644
index 00000000000..341a98965bd
--- /dev/null
+++ b/fs/cifs/Kconfig
@@ -0,0 +1,142 @@
1config CIFS
2 tristate "CIFS support (advanced network filesystem, SMBFS successor)"
3 depends on INET
4 select NLS
5 help
6 This is the client VFS module for the Common Internet File System
7 (CIFS) protocol which is the successor to the Server Message Block
8 (SMB) protocol, the native file sharing mechanism for most early
9 PC operating systems. The CIFS protocol is fully supported by
10 file servers such as Windows 2000 (including Windows 2003, NT 4
11 and Windows XP) as well by Samba (which provides excellent CIFS
12 server support for Linux and many other operating systems). Limited
13 support for OS/2 and Windows ME and similar servers is provided as
14 well.
15
16 The cifs module provides an advanced network file system
17 client for mounting to CIFS compliant servers. It includes
18 support for DFS (hierarchical name space), secure per-user
19 session establishment via Kerberos or NTLM or NTLMv2,
20 safe distributed caching (oplock), optional packet
21 signing, Unicode and other internationalization improvements.
22 If you need to mount to Samba or Windows from this machine, say Y.
23
24config CIFS_STATS
25 bool "CIFS statistics"
26 depends on CIFS
27 help
28 Enabling this option will cause statistics for each server share
29 mounted by the cifs client to be displayed in /proc/fs/cifs/Stats
30
31config CIFS_STATS2
32 bool "Extended statistics"
33 depends on CIFS_STATS
34 help
35 Enabling this option will allow more detailed statistics on SMB
36 request timing to be displayed in /proc/fs/cifs/DebugData and also
37 allow optional logging of slow responses to dmesg (depending on the
38 value of /proc/fs/cifs/cifsFYI, see fs/cifs/README for more details).
39 These additional statistics may have a minor effect on performance
40 and memory utilization.
41
42 Unless you are a developer or are doing network performance analysis
43 or tuning, say N.
44
45config CIFS_WEAK_PW_HASH
46 bool "Support legacy servers which use weaker LANMAN security"
47 depends on CIFS
48 help
49 Modern CIFS servers including Samba and most Windows versions
50 (since 1997) support stronger NTLM (and even NTLMv2 and Kerberos)
51 security mechanisms. These hash the password more securely
52 than the mechanisms used in the older LANMAN version of the
53 SMB protocol but LANMAN based authentication is needed to
54 establish sessions with some old SMB servers.
55
56 Enabling this option allows the cifs module to mount to older
57 LANMAN based servers such as OS/2 and Windows 95, but such
58 mounts may be less secure than mounts using NTLM or more recent
59 security mechanisms if you are on a public network. Unless you
60 have a need to access old SMB servers (and are on a private
61 network) you probably want to say N. Even if this support
62 is enabled in the kernel build, LANMAN authentication will not be
63 used automatically. At runtime LANMAN mounts are disabled but
64 can be set to required (or optional) either in
65 /proc/fs/cifs (see fs/cifs/README for more detail) or via an
66 option on the mount command. This support is disabled by
67 default in order to reduce the possibility of a downgrade
68 attack.
69
70 If unsure, say N.
71
72config CIFS_UPCALL
73 bool "Kerberos/SPNEGO advanced session setup"
74 depends on CIFS && KEYS
75 help
76 Enables an upcall mechanism for CIFS which accesses
77 userspace helper utilities to provide SPNEGO packaged (RFC 4178)
78 Kerberos tickets which are needed to mount to certain secure servers
79 (for which more secure Kerberos authentication is required). If
80 unsure, say N.
81
82config CIFS_XATTR
83 bool "CIFS extended attributes"
84 depends on CIFS
85 help
86 Extended attributes are name:value pairs associated with inodes by
87 the kernel or by users (see the attr(5) manual page, or visit
88 <http://acl.bestbits.at/> for details). CIFS maps the name of
89 extended attributes beginning with the user namespace prefix
90 to SMB/CIFS EAs. EAs are stored on Windows servers without the
91 user namespace prefix, but their names are seen by Linux cifs clients
92 prefaced by the user namespace prefix. The system namespace
93 (used by some filesystems to store ACLs) is not supported at
94 this time.
95
96 If unsure, say N.
97
98config CIFS_POSIX
99 bool "CIFS POSIX Extensions"
100 depends on CIFS_XATTR
101 help
102 Enabling this option will cause the cifs client to attempt to
103 negotiate a newer dialect with servers, such as Samba 3.0.5
104 or later, that optionally can handle more POSIX like (rather
105 than Windows like) file behavior. It also enables
106 support for POSIX ACLs (getfacl and setfacl) to servers
107 (such as Samba 3.10 and later) which can negotiate
108 CIFS POSIX ACL support. If unsure, say N.
109
110config CIFS_DEBUG2
111 bool "Enable additional CIFS debugging routines"
112 depends on CIFS
113 help
114 Enabling this option adds a few more debugging routines
115 to the cifs code which slightly increases the size of
116 the cifs module and can cause additional logging of debug
117 messages in some error paths, slowing performance. This
118 option can be turned off unless you are debugging
119 cifs problems. If unsure, say N.
120
121config CIFS_EXPERIMENTAL
122 bool "CIFS Experimental Features (EXPERIMENTAL)"
123 depends on CIFS && EXPERIMENTAL
124 help
125 Enables cifs features under testing. These features are
126 experimental and currently include DFS support and directory
127 change notification ie fcntl(F_DNOTIFY), as well as the upcall
128 mechanism which will be used for Kerberos session negotiation
129 and uid remapping. Some of these features also may depend on
130 setting a value of 1 to the pseudo-file /proc/fs/cifs/Experimental
131 (which is disabled by default). See the file fs/cifs/README
132 for more details. If unsure, say N.
133
134config CIFS_DFS_UPCALL
135 bool "DFS feature support (EXPERIMENTAL)"
136 depends on CIFS_EXPERIMENTAL
137 depends on KEYS
138 help
139 Enables an upcall mechanism for CIFS which contacts userspace
140 helper utilities to provide server name resolution (host names to
141 IP addresses) which is needed for implicit mounts of DFS junction
142 points. If unsure, say N.
diff --git a/fs/cifs/file.c b/fs/cifs/file.c
index c4a8a060512..62d8bd8f14c 100644
--- a/fs/cifs/file.c
+++ b/fs/cifs/file.c
@@ -1791,7 +1791,7 @@ static void cifs_copy_cache_pages(struct address_space *mapping,
1791 SetPageUptodate(page); 1791 SetPageUptodate(page);
1792 unlock_page(page); 1792 unlock_page(page);
1793 if (!pagevec_add(plru_pvec, page)) 1793 if (!pagevec_add(plru_pvec, page))
1794 __pagevec_lru_add(plru_pvec); 1794 __pagevec_lru_add_file(plru_pvec);
1795 data += PAGE_CACHE_SIZE; 1795 data += PAGE_CACHE_SIZE;
1796 } 1796 }
1797 return; 1797 return;
@@ -1925,7 +1925,7 @@ static int cifs_readpages(struct file *file, struct address_space *mapping,
1925 bytes_read = 0; 1925 bytes_read = 0;
1926 } 1926 }
1927 1927
1928 pagevec_lru_add(&lru_pvec); 1928 pagevec_lru_add_file(&lru_pvec);
1929 1929
1930/* need to free smb_read_data buf before exit */ 1930/* need to free smb_read_data buf before exit */
1931 if (smb_read_data) { 1931 if (smb_read_data) {
diff --git a/fs/exec.c b/fs/exec.c
index a41e7902ed0..4e834f16d9d 100644
--- a/fs/exec.c
+++ b/fs/exec.c
@@ -1386,7 +1386,7 @@ EXPORT_SYMBOL(set_binfmt);
1386 * name into corename, which must have space for at least 1386 * name into corename, which must have space for at least
1387 * CORENAME_MAX_SIZE bytes plus one byte for the zero terminator. 1387 * CORENAME_MAX_SIZE bytes plus one byte for the zero terminator.
1388 */ 1388 */
1389static int format_corename(char *corename, int nr_threads, long signr) 1389static int format_corename(char *corename, long signr)
1390{ 1390{
1391 const char *pat_ptr = core_pattern; 1391 const char *pat_ptr = core_pattern;
1392 int ispipe = (*pat_ptr == '|'); 1392 int ispipe = (*pat_ptr == '|');
@@ -1493,8 +1493,7 @@ static int format_corename(char *corename, int nr_threads, long signr)
1493 * If core_pattern does not include a %p (as is the default) 1493 * If core_pattern does not include a %p (as is the default)
1494 * and core_uses_pid is set, then .%pid will be appended to 1494 * and core_uses_pid is set, then .%pid will be appended to
1495 * the filename. Do not do this for piped commands. */ 1495 * the filename. Do not do this for piped commands. */
1496 if (!ispipe && !pid_in_pattern 1496 if (!ispipe && !pid_in_pattern && core_uses_pid) {
1497 && (core_uses_pid || nr_threads)) {
1498 rc = snprintf(out_ptr, out_end - out_ptr, 1497 rc = snprintf(out_ptr, out_end - out_ptr,
1499 ".%d", task_tgid_vnr(current)); 1498 ".%d", task_tgid_vnr(current));
1500 if (rc > out_end - out_ptr) 1499 if (rc > out_end - out_ptr)
@@ -1757,7 +1756,7 @@ int do_coredump(long signr, int exit_code, struct pt_regs * regs)
1757 * uses lock_kernel() 1756 * uses lock_kernel()
1758 */ 1757 */
1759 lock_kernel(); 1758 lock_kernel();
1760 ispipe = format_corename(corename, retval, signr); 1759 ispipe = format_corename(corename, signr);
1761 unlock_kernel(); 1760 unlock_kernel();
1762 /* 1761 /*
1763 * Don't bother to check the RLIMIT_CORE value if core_pattern points 1762 * Don't bother to check the RLIMIT_CORE value if core_pattern points
diff --git a/fs/ext2/Kconfig b/fs/ext2/Kconfig
new file mode 100644
index 00000000000..14a6780fd03
--- /dev/null
+++ b/fs/ext2/Kconfig
@@ -0,0 +1,55 @@
1config EXT2_FS
2 tristate "Second extended fs support"
3 help
4 Ext2 is a standard Linux file system for hard disks.
5
6 To compile this file system support as a module, choose M here: the
7 module will be called ext2.
8
9 If unsure, say Y.
10
11config EXT2_FS_XATTR
12 bool "Ext2 extended attributes"
13 depends on EXT2_FS
14 help
15 Extended attributes are name:value pairs associated with inodes by
16 the kernel or by users (see the attr(5) manual page, or visit
17 <http://acl.bestbits.at/> for details).
18
19 If unsure, say N.
20
21config EXT2_FS_POSIX_ACL
22 bool "Ext2 POSIX Access Control Lists"
23 depends on EXT2_FS_XATTR
24 select FS_POSIX_ACL
25 help
26 Posix Access Control Lists (ACLs) support permissions for users and
27 groups beyond the owner/group/world scheme.
28
29 To learn more about Access Control Lists, visit the Posix ACLs for
30 Linux website <http://acl.bestbits.at/>.
31
32 If you don't know what Access Control Lists are, say N
33
34config EXT2_FS_SECURITY
35 bool "Ext2 Security Labels"
36 depends on EXT2_FS_XATTR
37 help
38 Security labels support alternative access control models
39 implemented by security modules like SELinux. This option
40 enables an extended attribute handler for file security
41 labels in the ext2 filesystem.
42
43 If you are not using a security module that requires using
44 extended attributes for file security labels, say N.
45
46config EXT2_FS_XIP
47 bool "Ext2 execute in place support"
48 depends on EXT2_FS && MMU
49 help
50 Execute in place can be used on memory-backed block devices. If you
51 enable this option, you can select to mount block devices which are
52 capable of this feature without using the page cache.
53
54 If you do not use a block device that is capable of using this,
55 or if unsure, say N.
diff --git a/fs/ext3/Kconfig b/fs/ext3/Kconfig
new file mode 100644
index 00000000000..8e0cfe44b0f
--- /dev/null
+++ b/fs/ext3/Kconfig
@@ -0,0 +1,67 @@
1config EXT3_FS
2 tristate "Ext3 journalling file system support"
3 select JBD
4 help
5 This is the journalling version of the Second extended file system
6 (often called ext3), the de facto standard Linux file system
7 (method to organize files on a storage device) for hard disks.
8
9 The journalling code included in this driver means you do not have
10 to run e2fsck (file system checker) on your file systems after a
11 crash. The journal keeps track of any changes that were being made
12 at the time the system crashed, and can ensure that your file system
13 is consistent without the need for a lengthy check.
14
15 Other than adding the journal to the file system, the on-disk format
16 of ext3 is identical to ext2. It is possible to freely switch
17 between using the ext3 driver and the ext2 driver, as long as the
18 file system has been cleanly unmounted, or e2fsck is run on the file
19 system.
20
21 To add a journal on an existing ext2 file system or change the
22 behavior of ext3 file systems, you can use the tune2fs utility ("man
23 tune2fs"). To modify attributes of files and directories on ext3
24 file systems, use chattr ("man chattr"). You need to be using
25 e2fsprogs version 1.20 or later in order to create ext3 journals
26 (available at <http://sourceforge.net/projects/e2fsprogs/>).
27
28 To compile this file system support as a module, choose M here: the
29 module will be called ext3.
30
31config EXT3_FS_XATTR
32 bool "Ext3 extended attributes"
33 depends on EXT3_FS
34 default y
35 help
36 Extended attributes are name:value pairs associated with inodes by
37 the kernel or by users (see the attr(5) manual page, or visit
38 <http://acl.bestbits.at/> for details).
39
40 If unsure, say N.
41
42 You need this for POSIX ACL support on ext3.
43
44config EXT3_FS_POSIX_ACL
45 bool "Ext3 POSIX Access Control Lists"
46 depends on EXT3_FS_XATTR
47 select FS_POSIX_ACL
48 help
49 Posix Access Control Lists (ACLs) support permissions for users and
50 groups beyond the owner/group/world scheme.
51
52 To learn more about Access Control Lists, visit the Posix ACLs for
53 Linux website <http://acl.bestbits.at/>.
54
55 If you don't know what Access Control Lists are, say N
56
57config EXT3_FS_SECURITY
58 bool "Ext3 Security Labels"
59 depends on EXT3_FS_XATTR
60 help
61 Security labels support alternative access control models
62 implemented by security modules like SELinux. This option
63 enables an extended attribute handler for file security
64 labels in the ext3 filesystem.
65
66 If you are not using a security module that requires using
67 extended attributes for file security labels, say N.
diff --git a/fs/ext3/balloc.c b/fs/ext3/balloc.c
index 92fd0338a6e..f5b57a2ca35 100644
--- a/fs/ext3/balloc.c
+++ b/fs/ext3/balloc.c
@@ -1547,6 +1547,7 @@ retry_alloc:
1547 * turn off reservation for this allocation 1547 * turn off reservation for this allocation
1548 */ 1548 */
1549 if (my_rsv && (free_blocks < windowsz) 1549 if (my_rsv && (free_blocks < windowsz)
1550 && (free_blocks > 0)
1550 && (rsv_is_empty(&my_rsv->rsv_window))) 1551 && (rsv_is_empty(&my_rsv->rsv_window)))
1551 my_rsv = NULL; 1552 my_rsv = NULL;
1552 1553
@@ -1585,7 +1586,7 @@ retry_alloc:
1585 * free blocks is less than half of the reservation 1586 * free blocks is less than half of the reservation
1586 * window size. 1587 * window size.
1587 */ 1588 */
1588 if (free_blocks <= (windowsz/2)) 1589 if (my_rsv && (free_blocks <= (windowsz/2)))
1589 continue; 1590 continue;
1590 1591
1591 brelse(bitmap_bh); 1592 brelse(bitmap_bh);
diff --git a/fs/ext3/dir.c b/fs/ext3/dir.c
index 2eea96ec78e..4c82531ea0a 100644
--- a/fs/ext3/dir.c
+++ b/fs/ext3/dir.c
@@ -102,6 +102,7 @@ static int ext3_readdir(struct file * filp,
102 int err; 102 int err;
103 struct inode *inode = filp->f_path.dentry->d_inode; 103 struct inode *inode = filp->f_path.dentry->d_inode;
104 int ret = 0; 104 int ret = 0;
105 int dir_has_error = 0;
105 106
106 sb = inode->i_sb; 107 sb = inode->i_sb;
107 108
@@ -148,9 +149,12 @@ static int ext3_readdir(struct file * filp,
148 * of recovering data when there's a bad sector 149 * of recovering data when there's a bad sector
149 */ 150 */
150 if (!bh) { 151 if (!bh) {
151 ext3_error (sb, "ext3_readdir", 152 if (!dir_has_error) {
152 "directory #%lu contains a hole at offset %lu", 153 ext3_error(sb, __func__, "directory #%lu "
153 inode->i_ino, (unsigned long)filp->f_pos); 154 "contains a hole at offset %lld",
155 inode->i_ino, filp->f_pos);
156 dir_has_error = 1;
157 }
154 /* corrupt size? Maybe no more blocks to read */ 158 /* corrupt size? Maybe no more blocks to read */
155 if (filp->f_pos > inode->i_blocks << 9) 159 if (filp->f_pos > inode->i_blocks << 9)
156 break; 160 break;
@@ -410,7 +414,7 @@ static int call_filldir(struct file * filp, void * dirent,
410 get_dtype(sb, fname->file_type)); 414 get_dtype(sb, fname->file_type));
411 if (error) { 415 if (error) {
412 filp->f_pos = curr_pos; 416 filp->f_pos = curr_pos;
413 info->extra_fname = fname->next; 417 info->extra_fname = fname;
414 return error; 418 return error;
415 } 419 }
416 fname = fname->next; 420 fname = fname->next;
@@ -449,11 +453,21 @@ static int ext3_dx_readdir(struct file * filp,
449 * If there are any leftover names on the hash collision 453 * If there are any leftover names on the hash collision
450 * chain, return them first. 454 * chain, return them first.
451 */ 455 */
452 if (info->extra_fname && 456 if (info->extra_fname) {
453 call_filldir(filp, dirent, filldir, info->extra_fname)) 457 if (call_filldir(filp, dirent, filldir, info->extra_fname))
454 goto finished; 458 goto finished;
455 459
456 if (!info->curr_node) 460 info->extra_fname = NULL;
461 info->curr_node = rb_next(info->curr_node);
462 if (!info->curr_node) {
463 if (info->next_hash == ~0) {
464 filp->f_pos = EXT3_HTREE_EOF;
465 goto finished;
466 }
467 info->curr_hash = info->next_hash;
468 info->curr_minor_hash = 0;
469 }
470 } else if (!info->curr_node)
457 info->curr_node = rb_first(&info->root); 471 info->curr_node = rb_first(&info->root);
458 472
459 while (1) { 473 while (1) {
diff --git a/fs/ext3/inode.c b/fs/ext3/inode.c
index ebfec4d0148..f8424ad8997 100644
--- a/fs/ext3/inode.c
+++ b/fs/ext3/inode.c
@@ -1186,6 +1186,13 @@ write_begin_failed:
1186 ext3_journal_stop(handle); 1186 ext3_journal_stop(handle);
1187 unlock_page(page); 1187 unlock_page(page);
1188 page_cache_release(page); 1188 page_cache_release(page);
1189 /*
1190 * block_write_begin may have instantiated a few blocks
1191 * outside i_size. Trim these off again. Don't need
1192 * i_size_read because we hold i_mutex.
1193 */
1194 if (pos + len > inode->i_size)
1195 vmtruncate(inode, inode->i_size);
1189 } 1196 }
1190 if (ret == -ENOSPC && ext3_should_retry_alloc(inode->i_sb, &retries)) 1197 if (ret == -ENOSPC && ext3_should_retry_alloc(inode->i_sb, &retries))
1191 goto retry; 1198 goto retry;
diff --git a/fs/ext3/resize.c b/fs/ext3/resize.c
index 77278e947e9..78fdf383637 100644
--- a/fs/ext3/resize.c
+++ b/fs/ext3/resize.c
@@ -790,7 +790,8 @@ int ext3_group_add(struct super_block *sb, struct ext3_new_group_data *input)
790 790
791 if (reserved_gdb || gdb_off == 0) { 791 if (reserved_gdb || gdb_off == 0) {
792 if (!EXT3_HAS_COMPAT_FEATURE(sb, 792 if (!EXT3_HAS_COMPAT_FEATURE(sb,
793 EXT3_FEATURE_COMPAT_RESIZE_INODE)){ 793 EXT3_FEATURE_COMPAT_RESIZE_INODE)
794 || !le16_to_cpu(es->s_reserved_gdt_blocks)) {
794 ext3_warning(sb, __func__, 795 ext3_warning(sb, __func__,
795 "No reserved GDT blocks, can't resize"); 796 "No reserved GDT blocks, can't resize");
796 return -EPERM; 797 return -EPERM;
diff --git a/fs/ext3/super.c b/fs/ext3/super.c
index 399a96a6c55..3a260af5544 100644
--- a/fs/ext3/super.c
+++ b/fs/ext3/super.c
@@ -625,6 +625,9 @@ static int ext3_show_options(struct seq_file *seq, struct vfsmount *vfs)
625 else if (test_opt(sb, DATA_FLAGS) == EXT3_MOUNT_WRITEBACK_DATA) 625 else if (test_opt(sb, DATA_FLAGS) == EXT3_MOUNT_WRITEBACK_DATA)
626 seq_puts(seq, ",data=writeback"); 626 seq_puts(seq, ",data=writeback");
627 627
628 if (test_opt(sb, DATA_ERR_ABORT))
629 seq_puts(seq, ",data_err=abort");
630
628 ext3_show_quota_options(seq, sb); 631 ext3_show_quota_options(seq, sb);
629 632
630 return 0; 633 return 0;
@@ -754,6 +757,7 @@ enum {
754 Opt_reservation, Opt_noreservation, Opt_noload, Opt_nobh, Opt_bh, 757 Opt_reservation, Opt_noreservation, Opt_noload, Opt_nobh, Opt_bh,
755 Opt_commit, Opt_journal_update, Opt_journal_inum, Opt_journal_dev, 758 Opt_commit, Opt_journal_update, Opt_journal_inum, Opt_journal_dev,
756 Opt_abort, Opt_data_journal, Opt_data_ordered, Opt_data_writeback, 759 Opt_abort, Opt_data_journal, Opt_data_ordered, Opt_data_writeback,
760 Opt_data_err_abort, Opt_data_err_ignore,
757 Opt_usrjquota, Opt_grpjquota, Opt_offusrjquota, Opt_offgrpjquota, 761 Opt_usrjquota, Opt_grpjquota, Opt_offusrjquota, Opt_offgrpjquota,
758 Opt_jqfmt_vfsold, Opt_jqfmt_vfsv0, Opt_quota, Opt_noquota, 762 Opt_jqfmt_vfsold, Opt_jqfmt_vfsv0, Opt_quota, Opt_noquota,
759 Opt_ignore, Opt_barrier, Opt_err, Opt_resize, Opt_usrquota, 763 Opt_ignore, Opt_barrier, Opt_err, Opt_resize, Opt_usrquota,
@@ -796,6 +800,8 @@ static const match_table_t tokens = {
796 {Opt_data_journal, "data=journal"}, 800 {Opt_data_journal, "data=journal"},
797 {Opt_data_ordered, "data=ordered"}, 801 {Opt_data_ordered, "data=ordered"},
798 {Opt_data_writeback, "data=writeback"}, 802 {Opt_data_writeback, "data=writeback"},
803 {Opt_data_err_abort, "data_err=abort"},
804 {Opt_data_err_ignore, "data_err=ignore"},
799 {Opt_offusrjquota, "usrjquota="}, 805 {Opt_offusrjquota, "usrjquota="},
800 {Opt_usrjquota, "usrjquota=%s"}, 806 {Opt_usrjquota, "usrjquota=%s"},
801 {Opt_offgrpjquota, "grpjquota="}, 807 {Opt_offgrpjquota, "grpjquota="},
@@ -1011,6 +1017,12 @@ static int parse_options (char *options, struct super_block *sb,
1011 sbi->s_mount_opt |= data_opt; 1017 sbi->s_mount_opt |= data_opt;
1012 } 1018 }
1013 break; 1019 break;
1020 case Opt_data_err_abort:
1021 set_opt(sbi->s_mount_opt, DATA_ERR_ABORT);
1022 break;
1023 case Opt_data_err_ignore:
1024 clear_opt(sbi->s_mount_opt, DATA_ERR_ABORT);
1025 break;
1014#ifdef CONFIG_QUOTA 1026#ifdef CONFIG_QUOTA
1015 case Opt_usrjquota: 1027 case Opt_usrjquota:
1016 qtype = USRQUOTA; 1028 qtype = USRQUOTA;
@@ -1986,6 +1998,10 @@ static void ext3_init_journal_params(struct super_block *sb, journal_t *journal)
1986 journal->j_flags |= JFS_BARRIER; 1998 journal->j_flags |= JFS_BARRIER;
1987 else 1999 else
1988 journal->j_flags &= ~JFS_BARRIER; 2000 journal->j_flags &= ~JFS_BARRIER;
2001 if (test_opt(sb, DATA_ERR_ABORT))
2002 journal->j_flags |= JFS_ABORT_ON_SYNCDATA_ERR;
2003 else
2004 journal->j_flags &= ~JFS_ABORT_ON_SYNCDATA_ERR;
1989 spin_unlock(&journal->j_state_lock); 2005 spin_unlock(&journal->j_state_lock);
1990} 2006}
1991 2007
diff --git a/fs/ext4/Kconfig b/fs/ext4/Kconfig
new file mode 100644
index 00000000000..7505482a08f
--- /dev/null
+++ b/fs/ext4/Kconfig
@@ -0,0 +1,79 @@
1config EXT4_FS
2 tristate "The Extended 4 (ext4) filesystem"
3 select JBD2
4 select CRC16
5 help
6 This is the next generation of the ext3 filesystem.
7
8 Unlike the change from ext2 filesystem to ext3 filesystem,
9 the on-disk format of ext4 is not forwards compatible with
10 ext3; it is based on extent maps and it supports 48-bit
11 physical block numbers. The ext4 filesystem also supports delayed
12 allocation, persistent preallocation, high resolution time stamps,
13 and a number of other features to improve performance and speed
14 up fsck time. For more information, please see the web pages at
15 http://ext4.wiki.kernel.org.
16
17 The ext4 filesystem will support mounting an ext3
18 filesystem; while there will be some performance gains from
19 the delayed allocation and inode table readahead, the best
20 performance gains will require enabling ext4 features in the
21 filesystem, or formating a new filesystem as an ext4
22 filesystem initially.
23
24 To compile this file system support as a module, choose M here. The
25 module will be called ext4.
26
27 If unsure, say N.
28
29config EXT4DEV_COMPAT
30 bool "Enable ext4dev compatibility"
31 depends on EXT4_FS
32 help
33 Starting with 2.6.28, the name of the ext4 filesystem was
34 renamed from ext4dev to ext4. Unfortunately there are some
35 legacy userspace programs (such as klibc's fstype) have
36 "ext4dev" hardcoded.
37
38 To enable backwards compatibility so that systems that are
39 still expecting to mount ext4 filesystems using ext4dev,
40 chose Y here. This feature will go away by 2.6.31, so
41 please arrange to get your userspace programs fixed!
42
43config EXT4_FS_XATTR
44 bool "Ext4 extended attributes"
45 depends on EXT4_FS
46 default y
47 help
48 Extended attributes are name:value pairs associated with inodes by
49 the kernel or by users (see the attr(5) manual page, or visit
50 <http://acl.bestbits.at/> for details).
51
52 If unsure, say N.
53
54 You need this for POSIX ACL support on ext4.
55
56config EXT4_FS_POSIX_ACL
57 bool "Ext4 POSIX Access Control Lists"
58 depends on EXT4_FS_XATTR
59 select FS_POSIX_ACL
60 help
61 POSIX Access Control Lists (ACLs) support permissions for users and
62 groups beyond the owner/group/world scheme.
63
64 To learn more about Access Control Lists, visit the POSIX ACLs for
65 Linux website <http://acl.bestbits.at/>.
66
67 If you don't know what Access Control Lists are, say N
68
69config EXT4_FS_SECURITY
70 bool "Ext4 Security Labels"
71 depends on EXT4_FS_XATTR
72 help
73 Security labels support alternative access control models
74 implemented by security modules like SELinux. This option
75 enables an extended attribute handler for file security
76 labels in the ext4 filesystem.
77
78 If you are not using a security module that requires using
79 extended attributes for file security labels, say N.
diff --git a/fs/fuse/file.c b/fs/fuse/file.c
index 2bada6bbc31..34930a964b8 100644
--- a/fs/fuse/file.c
+++ b/fs/fuse/file.c
@@ -101,6 +101,8 @@ void fuse_finish_open(struct inode *inode, struct file *file,
101 file->f_op = &fuse_direct_io_file_operations; 101 file->f_op = &fuse_direct_io_file_operations;
102 if (!(outarg->open_flags & FOPEN_KEEP_CACHE)) 102 if (!(outarg->open_flags & FOPEN_KEEP_CACHE))
103 invalidate_inode_pages2(inode->i_mapping); 103 invalidate_inode_pages2(inode->i_mapping);
104 if (outarg->open_flags & FOPEN_NONSEEKABLE)
105 nonseekable_open(inode, file);
104 ff->fh = outarg->fh; 106 ff->fh = outarg->fh;
105 file->private_data = fuse_file_get(ff); 107 file->private_data = fuse_file_get(ff);
106} 108}
@@ -1448,6 +1450,9 @@ static loff_t fuse_file_llseek(struct file *file, loff_t offset, int origin)
1448 mutex_lock(&inode->i_mutex); 1450 mutex_lock(&inode->i_mutex);
1449 switch (origin) { 1451 switch (origin) {
1450 case SEEK_END: 1452 case SEEK_END:
1453 retval = fuse_update_attributes(inode, NULL, file, NULL);
1454 if (retval)
1455 return retval;
1451 offset += i_size_read(inode); 1456 offset += i_size_read(inode);
1452 break; 1457 break;
1453 case SEEK_CUR: 1458 case SEEK_CUR:
diff --git a/fs/fuse/fuse_i.h b/fs/fuse/fuse_i.h
index 3a876076bdd..35accfdd747 100644
--- a/fs/fuse/fuse_i.h
+++ b/fs/fuse/fuse_i.h
@@ -6,6 +6,9 @@
6 See the file COPYING. 6 See the file COPYING.
7*/ 7*/
8 8
9#ifndef _FS_FUSE_I_H
10#define _FS_FUSE_I_H
11
9#include <linux/fuse.h> 12#include <linux/fuse.h>
10#include <linux/fs.h> 13#include <linux/fs.h>
11#include <linux/mount.h> 14#include <linux/mount.h>
@@ -655,3 +658,5 @@ void fuse_set_nowrite(struct inode *inode);
655void fuse_release_nowrite(struct inode *inode); 658void fuse_release_nowrite(struct inode *inode);
656 659
657u64 fuse_get_attr_version(struct fuse_conn *fc); 660u64 fuse_get_attr_version(struct fuse_conn *fc);
661
662#endif /* _FS_FUSE_I_H */
diff --git a/fs/fuse/inode.c b/fs/fuse/inode.c
index 6a84388cacf..54b1f0e1ef5 100644
--- a/fs/fuse/inode.c
+++ b/fs/fuse/inode.c
@@ -865,7 +865,7 @@ static int fuse_fill_super(struct super_block *sb, void *data, int silent)
865 if (is_bdev) { 865 if (is_bdev) {
866 fc->destroy_req = fuse_request_alloc(); 866 fc->destroy_req = fuse_request_alloc();
867 if (!fc->destroy_req) 867 if (!fc->destroy_req)
868 goto err_put_root; 868 goto err_free_init_req;
869 } 869 }
870 870
871 mutex_lock(&fuse_mutex); 871 mutex_lock(&fuse_mutex);
@@ -895,6 +895,7 @@ static int fuse_fill_super(struct super_block *sb, void *data, int silent)
895 895
896 err_unlock: 896 err_unlock:
897 mutex_unlock(&fuse_mutex); 897 mutex_unlock(&fuse_mutex);
898 err_free_init_req:
898 fuse_request_free(init_req); 899 fuse_request_free(init_req);
899 err_put_root: 900 err_put_root:
900 dput(root_dentry); 901 dput(root_dentry);
diff --git a/fs/hfsplus/extents.c b/fs/hfsplus/extents.c
index fec8f61227f..0022eec63cd 100644
--- a/fs/hfsplus/extents.c
+++ b/fs/hfsplus/extents.c
@@ -199,6 +199,9 @@ int hfsplus_get_block(struct inode *inode, sector_t iblock,
199 goto done; 199 goto done;
200 } 200 }
201 201
202 if (inode->i_ino == HFSPLUS_EXT_CNID)
203 return -EIO;
204
202 mutex_lock(&HFSPLUS_I(inode).extents_lock); 205 mutex_lock(&HFSPLUS_I(inode).extents_lock);
203 res = hfsplus_ext_read_extent(inode, ablock); 206 res = hfsplus_ext_read_extent(inode, ablock);
204 if (!res) { 207 if (!res) {
diff --git a/fs/hfsplus/inode.c b/fs/hfsplus/inode.c
index b085d64a2b6..963be644297 100644
--- a/fs/hfsplus/inode.c
+++ b/fs/hfsplus/inode.c
@@ -254,6 +254,8 @@ static int hfsplus_file_open(struct inode *inode, struct file *file)
254{ 254{
255 if (HFSPLUS_IS_RSRC(inode)) 255 if (HFSPLUS_IS_RSRC(inode))
256 inode = HFSPLUS_I(inode).rsrc_inode; 256 inode = HFSPLUS_I(inode).rsrc_inode;
257 if (!(file->f_flags & O_LARGEFILE) && i_size_read(inode) > MAX_NON_LFS)
258 return -EOVERFLOW;
257 atomic_inc(&HFSPLUS_I(inode).opencnt); 259 atomic_inc(&HFSPLUS_I(inode).opencnt);
258 return 0; 260 return 0;
259} 261}
diff --git a/fs/jbd/Kconfig b/fs/jbd/Kconfig
new file mode 100644
index 00000000000..4e28beeed15
--- /dev/null
+++ b/fs/jbd/Kconfig
@@ -0,0 +1,30 @@
1config JBD
2 tristate
3 help
4 This is a generic journalling layer for block devices. It is
5 currently used by the ext3 file system, but it could also be
6 used to add journal support to other file systems or block
7 devices such as RAID or LVM.
8
9 If you are using the ext3 file system, you need to say Y here.
10 If you are not using ext3 then you will probably want to say N.
11
12 To compile this device as a module, choose M here: the module will be
13 called jbd. If you are compiling ext3 into the kernel, you
14 cannot compile this code as a module.
15
16config JBD_DEBUG
17 bool "JBD (ext3) debugging support"
18 depends on JBD && DEBUG_FS
19 help
20 If you are using the ext3 journaled file system (or potentially any
21 other file system/device using JBD), this option allows you to
22 enable debugging output while the system is running, in order to
23 help track down any problems you are having. By default the
24 debugging output will be turned off.
25
26 If you select Y here, then you will be able to turn on debugging
27 with "echo N > /sys/kernel/debug/jbd/jbd-debug", where N is a
28 number between 1 and 5, the higher the number, the more debugging
29 output is generated. To turn debugging off again, do
30 "echo 0 > /sys/kernel/debug/jbd/jbd-debug".
diff --git a/fs/jbd/commit.c b/fs/jbd/commit.c
index ae08c057e75..25719d902c5 100644
--- a/fs/jbd/commit.c
+++ b/fs/jbd/commit.c
@@ -482,6 +482,8 @@ void journal_commit_transaction(journal_t *journal)
482 printk(KERN_WARNING 482 printk(KERN_WARNING
483 "JBD: Detected IO errors while flushing file data " 483 "JBD: Detected IO errors while flushing file data "
484 "on %s\n", bdevname(journal->j_fs_dev, b)); 484 "on %s\n", bdevname(journal->j_fs_dev, b));
485 if (journal->j_flags & JFS_ABORT_ON_SYNCDATA_ERR)
486 journal_abort(journal, err);
485 err = 0; 487 err = 0;
486 } 488 }
487 489
@@ -518,9 +520,10 @@ void journal_commit_transaction(journal_t *journal)
518 jh = commit_transaction->t_buffers; 520 jh = commit_transaction->t_buffers;
519 521
520 /* If we're in abort mode, we just un-journal the buffer and 522 /* If we're in abort mode, we just un-journal the buffer and
521 release it for background writing. */ 523 release it. */
522 524
523 if (is_journal_aborted(journal)) { 525 if (is_journal_aborted(journal)) {
526 clear_buffer_jbddirty(jh2bh(jh));
524 JBUFFER_TRACE(jh, "journal is aborting: refile"); 527 JBUFFER_TRACE(jh, "journal is aborting: refile");
525 journal_refile_buffer(journal, jh); 528 journal_refile_buffer(journal, jh);
526 /* If that was the last one, we need to clean up 529 /* If that was the last one, we need to clean up
@@ -762,6 +765,9 @@ wait_for_iobuf:
762 /* AKPM: bforget here */ 765 /* AKPM: bforget here */
763 } 766 }
764 767
768 if (err)
769 journal_abort(journal, err);
770
765 jbd_debug(3, "JBD: commit phase 6\n"); 771 jbd_debug(3, "JBD: commit phase 6\n");
766 772
767 if (journal_write_commit_record(journal, commit_transaction)) 773 if (journal_write_commit_record(journal, commit_transaction))
@@ -852,6 +858,8 @@ restart_loop:
852 if (buffer_jbddirty(bh)) { 858 if (buffer_jbddirty(bh)) {
853 JBUFFER_TRACE(jh, "add to new checkpointing trans"); 859 JBUFFER_TRACE(jh, "add to new checkpointing trans");
854 __journal_insert_checkpoint(jh, commit_transaction); 860 __journal_insert_checkpoint(jh, commit_transaction);
861 if (is_journal_aborted(journal))
862 clear_buffer_jbddirty(bh);
855 JBUFFER_TRACE(jh, "refile for checkpoint writeback"); 863 JBUFFER_TRACE(jh, "refile for checkpoint writeback");
856 __journal_refile_buffer(jh); 864 __journal_refile_buffer(jh);
857 jbd_unlock_bh_state(bh); 865 jbd_unlock_bh_state(bh);
diff --git a/fs/jbd/transaction.c b/fs/jbd/transaction.c
index 0540ca27a44..d15cd6e7251 100644
--- a/fs/jbd/transaction.c
+++ b/fs/jbd/transaction.c
@@ -954,9 +954,10 @@ int journal_dirty_data(handle_t *handle, struct buffer_head *bh)
954 journal_t *journal = handle->h_transaction->t_journal; 954 journal_t *journal = handle->h_transaction->t_journal;
955 int need_brelse = 0; 955 int need_brelse = 0;
956 struct journal_head *jh; 956 struct journal_head *jh;
957 int ret = 0;
957 958
958 if (is_handle_aborted(handle)) 959 if (is_handle_aborted(handle))
959 return 0; 960 return ret;
960 961
961 jh = journal_add_journal_head(bh); 962 jh = journal_add_journal_head(bh);
962 JBUFFER_TRACE(jh, "entry"); 963 JBUFFER_TRACE(jh, "entry");
@@ -1067,7 +1068,16 @@ int journal_dirty_data(handle_t *handle, struct buffer_head *bh)
1067 time if it is redirtied */ 1068 time if it is redirtied */
1068 } 1069 }
1069 1070
1070 /* journal_clean_data_list() may have got there first */ 1071 /*
1072 * We cannot remove the buffer with io error from the
1073 * committing transaction, because otherwise it would
1074 * miss the error and the commit would not abort.
1075 */
1076 if (unlikely(!buffer_uptodate(bh))) {
1077 ret = -EIO;
1078 goto no_journal;
1079 }
1080
1071 if (jh->b_transaction != NULL) { 1081 if (jh->b_transaction != NULL) {
1072 JBUFFER_TRACE(jh, "unfile from commit"); 1082 JBUFFER_TRACE(jh, "unfile from commit");
1073 __journal_temp_unlink_buffer(jh); 1083 __journal_temp_unlink_buffer(jh);
@@ -1108,7 +1118,7 @@ no_journal:
1108 } 1118 }
1109 JBUFFER_TRACE(jh, "exit"); 1119 JBUFFER_TRACE(jh, "exit");
1110 journal_put_journal_head(jh); 1120 journal_put_journal_head(jh);
1111 return 0; 1121 return ret;
1112} 1122}
1113 1123
1114/** 1124/**
diff --git a/fs/jbd2/Kconfig b/fs/jbd2/Kconfig
new file mode 100644
index 00000000000..f32f346f4b0
--- /dev/null
+++ b/fs/jbd2/Kconfig
@@ -0,0 +1,33 @@
1config JBD2
2 tristate
3 select CRC32
4 help
5 This is a generic journaling layer for block devices that support
6 both 32-bit and 64-bit block numbers. It is currently used by
7 the ext4 and OCFS2 filesystems, but it could also be used to add
8 journal support to other file systems or block devices such
9 as RAID or LVM.
10
11 If you are using ext4 or OCFS2, you need to say Y here.
12 If you are not using ext4 or OCFS2 then you will
13 probably want to say N.
14
15 To compile this device as a module, choose M here. The module will be
16 called jbd2. If you are compiling ext4 or OCFS2 into the kernel,
17 you cannot compile this code as a module.
18
19config JBD2_DEBUG
20 bool "JBD2 (ext4) debugging support"
21 depends on JBD2 && DEBUG_FS
22 help
23 If you are using the ext4 journaled file system (or
24 potentially any other filesystem/device using JBD2), this option
25 allows you to enable debugging output while the system is running,
26 in order to help track down any problems you are having.
27 By default, the debugging output will be turned off.
28
29 If you select Y here, then you will be able to turn on debugging
30 with "echo N > /sys/kernel/debug/jbd2/jbd2-debug", where N is a
31 number between 1 and 5. The higher the number, the more debugging
32 output is generated. To turn debugging off again, do
33 "echo 0 > /sys/kernel/debug/jbd2/jbd2-debug".
diff --git a/fs/jffs2/Kconfig b/fs/jffs2/Kconfig
new file mode 100644
index 00000000000..6ae169cd8fa
--- /dev/null
+++ b/fs/jffs2/Kconfig
@@ -0,0 +1,188 @@
1config JFFS2_FS
2 tristate "Journalling Flash File System v2 (JFFS2) support"
3 select CRC32
4 depends on MTD
5 help
6 JFFS2 is the second generation of the Journalling Flash File System
7 for use on diskless embedded devices. It provides improved wear
8 levelling, compression and support for hard links. You cannot use
9 this on normal block devices, only on 'MTD' devices.
10
11 Further information on the design and implementation of JFFS2 is
12 available at <http://sources.redhat.com/jffs2/>.
13
14config JFFS2_FS_DEBUG
15 int "JFFS2 debugging verbosity (0 = quiet, 2 = noisy)"
16 depends on JFFS2_FS
17 default "0"
18 help
19 This controls the amount of debugging messages produced by the JFFS2
20 code. Set it to zero for use in production systems. For evaluation,
21 testing and debugging, it's advisable to set it to one. This will
22 enable a few assertions and will print debugging messages at the
23 KERN_DEBUG loglevel, where they won't normally be visible. Level 2
24 is unlikely to be useful - it enables extra debugging in certain
25 areas which at one point needed debugging, but when the bugs were
26 located and fixed, the detailed messages were relegated to level 2.
27
28 If reporting bugs, please try to have available a full dump of the
29 messages at debug level 1 while the misbehaviour was occurring.
30
31config JFFS2_FS_WRITEBUFFER
32 bool "JFFS2 write-buffering support"
33 depends on JFFS2_FS
34 default y
35 help
36 This enables the write-buffering support in JFFS2.
37
38 This functionality is required to support JFFS2 on the following
39 types of flash devices:
40 - NAND flash
41 - NOR flash with transparent ECC
42 - DataFlash
43
44config JFFS2_FS_WBUF_VERIFY
45 bool "Verify JFFS2 write-buffer reads"
46 depends on JFFS2_FS_WRITEBUFFER
47 default n
48 help
49 This causes JFFS2 to read back every page written through the
50 write-buffer, and check for errors.
51
52config JFFS2_SUMMARY
53 bool "JFFS2 summary support (EXPERIMENTAL)"
54 depends on JFFS2_FS && EXPERIMENTAL
55 default n
56 help
57 This feature makes it possible to use summary information
58 for faster filesystem mount.
59
60 The summary information can be inserted into a filesystem image
61 by the utility 'sumtool'.
62
63 If unsure, say 'N'.
64
65config JFFS2_FS_XATTR
66 bool "JFFS2 XATTR support (EXPERIMENTAL)"
67 depends on JFFS2_FS && EXPERIMENTAL
68 default n
69 help
70 Extended attributes are name:value pairs associated with inodes by
71 the kernel or by users (see the attr(5) manual page, or visit
72 <http://acl.bestbits.at/> for details).
73
74 If unsure, say N.
75
76config JFFS2_FS_POSIX_ACL
77 bool "JFFS2 POSIX Access Control Lists"
78 depends on JFFS2_FS_XATTR
79 default y
80 select FS_POSIX_ACL
81 help
82 Posix Access Control Lists (ACLs) support permissions for users and
83 groups beyond the owner/group/world scheme.
84
85 To learn more about Access Control Lists, visit the Posix ACLs for
86 Linux website <http://acl.bestbits.at/>.
87
88 If you don't know what Access Control Lists are, say N
89
90config JFFS2_FS_SECURITY
91 bool "JFFS2 Security Labels"
92 depends on JFFS2_FS_XATTR
93 default y
94 help
95 Security labels support alternative access control models
96 implemented by security modules like SELinux. This option
97 enables an extended attribute handler for file security
98 labels in the jffs2 filesystem.
99
100 If you are not using a security module that requires using
101 extended attributes for file security labels, say N.
102
103config JFFS2_COMPRESSION_OPTIONS
104 bool "Advanced compression options for JFFS2"
105 depends on JFFS2_FS
106 default n
107 help
108 Enabling this option allows you to explicitly choose which
109 compression modules, if any, are enabled in JFFS2. Removing
110 compressors can mean you cannot read existing file systems,
111 and enabling experimental compressors can mean that you
112 write a file system which cannot be read by a standard kernel.
113
114 If unsure, you should _definitely_ say 'N'.
115
116config JFFS2_ZLIB
117 bool "JFFS2 ZLIB compression support" if JFFS2_COMPRESSION_OPTIONS
118 select ZLIB_INFLATE
119 select ZLIB_DEFLATE
120 depends on JFFS2_FS
121 default y
122 help
123 Zlib is designed to be a free, general-purpose, legally unencumbered,
124 lossless data-compression library for use on virtually any computer
125 hardware and operating system. See <http://www.gzip.org/zlib/> for
126 further information.
127
128 Say 'Y' if unsure.
129
130config JFFS2_LZO
131 bool "JFFS2 LZO compression support" if JFFS2_COMPRESSION_OPTIONS
132 select LZO_COMPRESS
133 select LZO_DECOMPRESS
134 depends on JFFS2_FS
135 default n
136 help
137 minilzo-based compression. Generally works better than Zlib.
138
139 This feature was added in July, 2007. Say 'N' if you need
140 compatibility with older bootloaders or kernels.
141
142config JFFS2_RTIME
143 bool "JFFS2 RTIME compression support" if JFFS2_COMPRESSION_OPTIONS
144 depends on JFFS2_FS
145 default y
146 help
147 Rtime does manage to recompress already-compressed data. Say 'Y' if unsure.
148
149config JFFS2_RUBIN
150 bool "JFFS2 RUBIN compression support" if JFFS2_COMPRESSION_OPTIONS
151 depends on JFFS2_FS
152 default n
153 help
154 RUBINMIPS and DYNRUBIN compressors. Say 'N' if unsure.
155
156choice
157 prompt "JFFS2 default compression mode" if JFFS2_COMPRESSION_OPTIONS
158 default JFFS2_CMODE_PRIORITY
159 depends on JFFS2_FS
160 help
161 You can set here the default compression mode of JFFS2 from
162 the available compression modes. Don't touch if unsure.
163
164config JFFS2_CMODE_NONE
165 bool "no compression"
166 help
167 Uses no compression.
168
169config JFFS2_CMODE_PRIORITY
170 bool "priority"
171 help
172 Tries the compressors in a predefined order and chooses the first
173 successful one.
174
175config JFFS2_CMODE_SIZE
176 bool "size (EXPERIMENTAL)"
177 help
178 Tries all compressors and chooses the one which has the smallest
179 result.
180
181config JFFS2_CMODE_FAVOURLZO
182 bool "Favour LZO"
183 help
184 Tries all compressors and chooses the one which has the smallest
185 result but gives some preference to LZO (which has faster
186 decompression) at the expense of size.
187
188endchoice
diff --git a/fs/jffs2/compr.c b/fs/jffs2/compr.c
index 86739ee53b3..f25e70c1b51 100644
--- a/fs/jffs2/compr.c
+++ b/fs/jffs2/compr.c
@@ -53,8 +53,8 @@ static int jffs2_is_best_compression(struct jffs2_compressor *this,
53} 53}
54 54
55/* jffs2_compress: 55/* jffs2_compress:
56 * @data: Pointer to uncompressed data 56 * @data_in: Pointer to uncompressed data
57 * @cdata: Pointer to returned pointer to buffer for compressed data 57 * @cpage_out: Pointer to returned pointer to buffer for compressed data
58 * @datalen: On entry, holds the amount of data available for compression. 58 * @datalen: On entry, holds the amount of data available for compression.
59 * On exit, expected to hold the amount of data actually compressed. 59 * On exit, expected to hold the amount of data actually compressed.
60 * @cdatalen: On entry, holds the amount of space available for compressed 60 * @cdatalen: On entry, holds the amount of space available for compressed
diff --git a/fs/jffs2/dir.c b/fs/jffs2/dir.c
index cd219ef5525..b1aaae823a5 100644
--- a/fs/jffs2/dir.c
+++ b/fs/jffs2/dir.c
@@ -311,7 +311,7 @@ static int jffs2_symlink (struct inode *dir_i, struct dentry *dentry, const char
311 /* FIXME: If you care. We'd need to use frags for the target 311 /* FIXME: If you care. We'd need to use frags for the target
312 if it grows much more than this */ 312 if it grows much more than this */
313 if (targetlen > 254) 313 if (targetlen > 254)
314 return -EINVAL; 314 return -ENAMETOOLONG;
315 315
316 ri = jffs2_alloc_raw_inode(); 316 ri = jffs2_alloc_raw_inode();
317 317
diff --git a/fs/jffs2/erase.c b/fs/jffs2/erase.c
index dddb2a6c9e2..259461b910a 100644
--- a/fs/jffs2/erase.c
+++ b/fs/jffs2/erase.c
@@ -68,7 +68,7 @@ static void jffs2_erase_block(struct jffs2_sb_info *c,
68 instr->len = c->sector_size; 68 instr->len = c->sector_size;
69 instr->callback = jffs2_erase_callback; 69 instr->callback = jffs2_erase_callback;
70 instr->priv = (unsigned long)(&instr[1]); 70 instr->priv = (unsigned long)(&instr[1]);
71 instr->fail_addr = 0xffffffff; 71 instr->fail_addr = MTD_FAIL_ADDR_UNKNOWN;
72 72
73 ((struct erase_priv_struct *)instr->priv)->jeb = jeb; 73 ((struct erase_priv_struct *)instr->priv)->jeb = jeb;
74 ((struct erase_priv_struct *)instr->priv)->c = c; 74 ((struct erase_priv_struct *)instr->priv)->c = c;
@@ -175,7 +175,7 @@ static void jffs2_erase_failed(struct jffs2_sb_info *c, struct jffs2_eraseblock
175{ 175{
176 /* For NAND, if the failure did not occur at the device level for a 176 /* For NAND, if the failure did not occur at the device level for a
177 specific physical page, don't bother updating the bad block table. */ 177 specific physical page, don't bother updating the bad block table. */
178 if (jffs2_cleanmarker_oob(c) && (bad_offset != 0xffffffff)) { 178 if (jffs2_cleanmarker_oob(c) && (bad_offset != MTD_FAIL_ADDR_UNKNOWN)) {
179 /* We had a device-level failure to erase. Let's see if we've 179 /* We had a device-level failure to erase. Let's see if we've
180 failed too many times. */ 180 failed too many times. */
181 if (!jffs2_write_nand_badblock(c, jeb, bad_offset)) { 181 if (!jffs2_write_nand_badblock(c, jeb, bad_offset)) {
diff --git a/fs/jffs2/fs.c b/fs/jffs2/fs.c
index 086c4383022..249305d65d5 100644
--- a/fs/jffs2/fs.c
+++ b/fs/jffs2/fs.c
@@ -207,6 +207,8 @@ int jffs2_statfs(struct dentry *dentry, struct kstatfs *buf)
207 buf->f_files = 0; 207 buf->f_files = 0;
208 buf->f_ffree = 0; 208 buf->f_ffree = 0;
209 buf->f_namelen = JFFS2_MAX_NAME_LEN; 209 buf->f_namelen = JFFS2_MAX_NAME_LEN;
210 buf->f_fsid.val[0] = JFFS2_SUPER_MAGIC;
211 buf->f_fsid.val[1] = c->mtd->index;
210 212
211 spin_lock(&c->erase_completion_lock); 213 spin_lock(&c->erase_completion_lock);
212 avail = c->dirty_size + c->free_size; 214 avail = c->dirty_size + c->free_size;
@@ -440,14 +442,14 @@ struct inode *jffs2_new_inode (struct inode *dir_i, int mode, struct jffs2_raw_i
440 442
441 memset(ri, 0, sizeof(*ri)); 443 memset(ri, 0, sizeof(*ri));
442 /* Set OS-specific defaults for new inodes */ 444 /* Set OS-specific defaults for new inodes */
443 ri->uid = cpu_to_je16(current->fsuid); 445 ri->uid = cpu_to_je16(current_fsuid());
444 446
445 if (dir_i->i_mode & S_ISGID) { 447 if (dir_i->i_mode & S_ISGID) {
446 ri->gid = cpu_to_je16(dir_i->i_gid); 448 ri->gid = cpu_to_je16(dir_i->i_gid);
447 if (S_ISDIR(mode)) 449 if (S_ISDIR(mode))
448 mode |= S_ISGID; 450 mode |= S_ISGID;
449 } else { 451 } else {
450 ri->gid = cpu_to_je16(current->fsgid); 452 ri->gid = cpu_to_je16(current_fsgid());
451 } 453 }
452 454
453 /* POSIX ACLs have to be processed now, at least partly. 455 /* POSIX ACLs have to be processed now, at least partly.
diff --git a/fs/jffs2/nodemgmt.c b/fs/jffs2/nodemgmt.c
index a9bf9603c1b..0875b60b4bf 100644
--- a/fs/jffs2/nodemgmt.c
+++ b/fs/jffs2/nodemgmt.c
@@ -261,6 +261,10 @@ static int jffs2_find_nextblock(struct jffs2_sb_info *c)
261 261
262 jffs2_sum_reset_collected(c->summary); /* reset collected summary */ 262 jffs2_sum_reset_collected(c->summary); /* reset collected summary */
263 263
264 /* adjust write buffer offset, else we get a non contiguous write bug */
265 if (!(c->wbuf_ofs % c->sector_size) && !c->wbuf_len)
266 c->wbuf_ofs = 0xffffffff;
267
264 D1(printk(KERN_DEBUG "jffs2_find_nextblock(): new nextblock = 0x%08x\n", c->nextblock->offset)); 268 D1(printk(KERN_DEBUG "jffs2_find_nextblock(): new nextblock = 0x%08x\n", c->nextblock->offset));
265 269
266 return 0; 270 return 0;
diff --git a/fs/jffs2/wbuf.c b/fs/jffs2/wbuf.c
index 0e78b00035e..d9a721e6db7 100644
--- a/fs/jffs2/wbuf.c
+++ b/fs/jffs2/wbuf.c
@@ -679,10 +679,7 @@ static int __jffs2_flush_wbuf(struct jffs2_sb_info *c, int pad)
679 679
680 memset(c->wbuf,0xff,c->wbuf_pagesize); 680 memset(c->wbuf,0xff,c->wbuf_pagesize);
681 /* adjust write buffer offset, else we get a non contiguous write bug */ 681 /* adjust write buffer offset, else we get a non contiguous write bug */
682 if (SECTOR_ADDR(c->wbuf_ofs) == SECTOR_ADDR(c->wbuf_ofs+c->wbuf_pagesize)) 682 c->wbuf_ofs += c->wbuf_pagesize;
683 c->wbuf_ofs += c->wbuf_pagesize;
684 else
685 c->wbuf_ofs = 0xffffffff;
686 c->wbuf_len = 0; 683 c->wbuf_len = 0;
687 return 0; 684 return 0;
688} 685}
diff --git a/fs/nfs/callback.c b/fs/nfs/callback.c
index 6a09760c596..c2e9cfd9e5a 100644
--- a/fs/nfs/callback.c
+++ b/fs/nfs/callback.c
@@ -40,6 +40,16 @@ unsigned short nfs_callback_tcpport;
40static const int nfs_set_port_min = 0; 40static const int nfs_set_port_min = 0;
41static const int nfs_set_port_max = 65535; 41static const int nfs_set_port_max = 65535;
42 42
43/*
44 * If the kernel has IPv6 support available, always listen for
45 * both AF_INET and AF_INET6 requests.
46 */
47#if defined(CONFIG_IPV6) || defined(CONFIG_IPV6_MODULE)
48static const sa_family_t nfs_callback_family = AF_INET6;
49#else
50static const sa_family_t nfs_callback_family = AF_INET;
51#endif
52
43static int param_set_port(const char *val, struct kernel_param *kp) 53static int param_set_port(const char *val, struct kernel_param *kp)
44{ 54{
45 char *endp; 55 char *endp;
@@ -106,7 +116,7 @@ int nfs_callback_up(void)
106 if (nfs_callback_info.users++ || nfs_callback_info.task != NULL) 116 if (nfs_callback_info.users++ || nfs_callback_info.task != NULL)
107 goto out; 117 goto out;
108 serv = svc_create(&nfs4_callback_program, NFS4_CALLBACK_BUFSIZE, 118 serv = svc_create(&nfs4_callback_program, NFS4_CALLBACK_BUFSIZE,
109 AF_INET, NULL); 119 nfs_callback_family, NULL);
110 ret = -ENOMEM; 120 ret = -ENOMEM;
111 if (!serv) 121 if (!serv)
112 goto out_err; 122 goto out_err;
@@ -116,7 +126,8 @@ int nfs_callback_up(void)
116 if (ret <= 0) 126 if (ret <= 0)
117 goto out_err; 127 goto out_err;
118 nfs_callback_tcpport = ret; 128 nfs_callback_tcpport = ret;
119 dprintk("Callback port = 0x%x\n", nfs_callback_tcpport); 129 dprintk("NFS: Callback listener port = %u (af %u)\n",
130 nfs_callback_tcpport, nfs_callback_family);
120 131
121 nfs_callback_info.rqst = svc_prepare_thread(serv, &serv->sv_pools[0]); 132 nfs_callback_info.rqst = svc_prepare_thread(serv, &serv->sv_pools[0]);
122 if (IS_ERR(nfs_callback_info.rqst)) { 133 if (IS_ERR(nfs_callback_info.rqst)) {
@@ -149,8 +160,8 @@ out:
149 mutex_unlock(&nfs_callback_mutex); 160 mutex_unlock(&nfs_callback_mutex);
150 return ret; 161 return ret;
151out_err: 162out_err:
152 dprintk("Couldn't create callback socket or server thread; err = %d\n", 163 dprintk("NFS: Couldn't create callback socket or server thread; "
153 ret); 164 "err = %d\n", ret);
154 nfs_callback_info.users--; 165 nfs_callback_info.users--;
155 goto out; 166 goto out;
156} 167}
diff --git a/fs/nfs/dir.c b/fs/nfs/dir.c
index 2ab70d46ecb..efdba2e802d 100644
--- a/fs/nfs/dir.c
+++ b/fs/nfs/dir.c
@@ -1517,7 +1517,7 @@ static int nfs_symlink(struct inode *dir, struct dentry *dentry, const char *sym
1517 if (!add_to_page_cache(page, dentry->d_inode->i_mapping, 0, 1517 if (!add_to_page_cache(page, dentry->d_inode->i_mapping, 0,
1518 GFP_KERNEL)) { 1518 GFP_KERNEL)) {
1519 pagevec_add(&lru_pvec, page); 1519 pagevec_add(&lru_pvec, page);
1520 pagevec_lru_add(&lru_pvec); 1520 pagevec_lru_add_file(&lru_pvec);
1521 SetPageUptodate(page); 1521 SetPageUptodate(page);
1522 unlock_page(page); 1522 unlock_page(page);
1523 } else 1523 } else
diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c
index c910413eaec..83e700a2b0c 100644
--- a/fs/nfs/nfs4proc.c
+++ b/fs/nfs/nfs4proc.c
@@ -1659,8 +1659,10 @@ nfs4_proc_setattr(struct dentry *dentry, struct nfs_fattr *fattr,
1659 struct nfs_open_context *ctx; 1659 struct nfs_open_context *ctx;
1660 1660
1661 ctx = nfs_file_open_context(sattr->ia_file); 1661 ctx = nfs_file_open_context(sattr->ia_file);
1662 cred = ctx->cred; 1662 if (ctx) {
1663 state = ctx->state; 1663 cred = ctx->cred;
1664 state = ctx->state;
1665 }
1664 } 1666 }
1665 1667
1666 status = nfs4_do_setattr(inode, cred, fattr, sattr, state); 1668 status = nfs4_do_setattr(inode, cred, fattr, sattr, state);
diff --git a/fs/nfs/super.c b/fs/nfs/super.c
index 8b28b95c9e4..a3b0061dfd4 100644
--- a/fs/nfs/super.c
+++ b/fs/nfs/super.c
@@ -2459,7 +2459,7 @@ static int nfs4_xdev_get_sb(struct file_system_type *fs_type, int flags,
2459 compare_super = NULL; 2459 compare_super = NULL;
2460 2460
2461 /* Get a superblock - note that we may end up sharing one that already exists */ 2461 /* Get a superblock - note that we may end up sharing one that already exists */
2462 s = sget(&nfs_fs_type, compare_super, nfs_set_super, &sb_mntdata); 2462 s = sget(&nfs4_fs_type, compare_super, nfs_set_super, &sb_mntdata);
2463 if (IS_ERR(s)) { 2463 if (IS_ERR(s)) {
2464 error = PTR_ERR(s); 2464 error = PTR_ERR(s);
2465 goto out_err_nosb; 2465 goto out_err_nosb;
@@ -2544,7 +2544,7 @@ static int nfs4_referral_get_sb(struct file_system_type *fs_type, int flags,
2544 compare_super = NULL; 2544 compare_super = NULL;
2545 2545
2546 /* Get a superblock - note that we may end up sharing one that already exists */ 2546 /* Get a superblock - note that we may end up sharing one that already exists */
2547 s = sget(&nfs_fs_type, compare_super, nfs_set_super, &sb_mntdata); 2547 s = sget(&nfs4_fs_type, compare_super, nfs_set_super, &sb_mntdata);
2548 if (IS_ERR(s)) { 2548 if (IS_ERR(s)) {
2549 error = PTR_ERR(s); 2549 error = PTR_ERR(s);
2550 goto out_err_nosb; 2550 goto out_err_nosb;
diff --git a/fs/ntfs/file.c b/fs/ntfs/file.c
index d020866d423..3140a4429af 100644
--- a/fs/ntfs/file.c
+++ b/fs/ntfs/file.c
@@ -439,7 +439,7 @@ static inline int __ntfs_grab_cache_pages(struct address_space *mapping,
439 pages[nr] = *cached_page; 439 pages[nr] = *cached_page;
440 page_cache_get(*cached_page); 440 page_cache_get(*cached_page);
441 if (unlikely(!pagevec_add(lru_pvec, *cached_page))) 441 if (unlikely(!pagevec_add(lru_pvec, *cached_page)))
442 __pagevec_lru_add(lru_pvec); 442 __pagevec_lru_add_file(lru_pvec);
443 *cached_page = NULL; 443 *cached_page = NULL;
444 } 444 }
445 index++; 445 index++;
@@ -2084,7 +2084,7 @@ err_out:
2084 OSYNC_METADATA|OSYNC_DATA); 2084 OSYNC_METADATA|OSYNC_DATA);
2085 } 2085 }
2086 } 2086 }
2087 pagevec_lru_add(&lru_pvec); 2087 pagevec_lru_add_file(&lru_pvec);
2088 ntfs_debug("Done. Returning %s (written 0x%lx, status %li).", 2088 ntfs_debug("Done. Returning %s (written 0x%lx, status %li).",
2089 written ? "written" : "status", (unsigned long)written, 2089 written ? "written" : "status", (unsigned long)written,
2090 (long)status); 2090 (long)status);
diff --git a/fs/proc/array.c b/fs/proc/array.c
index f4bc0e78953..bb9f4b05703 100644
--- a/fs/proc/array.c
+++ b/fs/proc/array.c
@@ -388,20 +388,20 @@ static int do_task_stat(struct seq_file *m, struct pid_namespace *ns,
388 388
389 /* add up live thread stats at the group level */ 389 /* add up live thread stats at the group level */
390 if (whole) { 390 if (whole) {
391 struct task_cputime cputime;
391 struct task_struct *t = task; 392 struct task_struct *t = task;
392 do { 393 do {
393 min_flt += t->min_flt; 394 min_flt += t->min_flt;
394 maj_flt += t->maj_flt; 395 maj_flt += t->maj_flt;
395 utime = cputime_add(utime, task_utime(t));
396 stime = cputime_add(stime, task_stime(t));
397 gtime = cputime_add(gtime, task_gtime(t)); 396 gtime = cputime_add(gtime, task_gtime(t));
398 t = next_thread(t); 397 t = next_thread(t);
399 } while (t != task); 398 } while (t != task);
400 399
401 min_flt += sig->min_flt; 400 min_flt += sig->min_flt;
402 maj_flt += sig->maj_flt; 401 maj_flt += sig->maj_flt;
403 utime = cputime_add(utime, sig->utime); 402 thread_group_cputime(task, &cputime);
404 stime = cputime_add(stime, sig->stime); 403 utime = cputime.utime;
404 stime = cputime.stime;
405 gtime = cputime_add(gtime, sig->gtime); 405 gtime = cputime_add(gtime, sig->gtime);
406 } 406 }
407 407
diff --git a/fs/proc/proc_misc.c b/fs/proc/proc_misc.c
index 59ea42e1ef0..7ea52c79b2d 100644
--- a/fs/proc/proc_misc.c
+++ b/fs/proc/proc_misc.c
@@ -30,6 +30,7 @@
30#include <linux/mm.h> 30#include <linux/mm.h>
31#include <linux/mmzone.h> 31#include <linux/mmzone.h>
32#include <linux/pagemap.h> 32#include <linux/pagemap.h>
33#include <linux/irq.h>
33#include <linux/interrupt.h> 34#include <linux/interrupt.h>
34#include <linux/swap.h> 35#include <linux/swap.h>
35#include <linux/slab.h> 36#include <linux/slab.h>
@@ -136,6 +137,8 @@ static int meminfo_read_proc(char *page, char **start, off_t off,
136 unsigned long allowed; 137 unsigned long allowed;
137 struct vmalloc_info vmi; 138 struct vmalloc_info vmi;
138 long cached; 139 long cached;
140 unsigned long pages[NR_LRU_LISTS];
141 int lru;
139 142
140/* 143/*
141 * display in kilobytes. 144 * display in kilobytes.
@@ -154,51 +157,70 @@ static int meminfo_read_proc(char *page, char **start, off_t off,
154 157
155 get_vmalloc_info(&vmi); 158 get_vmalloc_info(&vmi);
156 159
160 for (lru = LRU_BASE; lru < NR_LRU_LISTS; lru++)
161 pages[lru] = global_page_state(NR_LRU_BASE + lru);
162
157 /* 163 /*
158 * Tagged format, for easy grepping and expansion. 164 * Tagged format, for easy grepping and expansion.
159 */ 165 */
160 len = sprintf(page, 166 len = sprintf(page,
161 "MemTotal: %8lu kB\n" 167 "MemTotal: %8lu kB\n"
162 "MemFree: %8lu kB\n" 168 "MemFree: %8lu kB\n"
163 "Buffers: %8lu kB\n" 169 "Buffers: %8lu kB\n"
164 "Cached: %8lu kB\n" 170 "Cached: %8lu kB\n"
165 "SwapCached: %8lu kB\n" 171 "SwapCached: %8lu kB\n"
166 "Active: %8lu kB\n" 172 "Active: %8lu kB\n"
167 "Inactive: %8lu kB\n" 173 "Inactive: %8lu kB\n"
174 "Active(anon): %8lu kB\n"
175 "Inactive(anon): %8lu kB\n"
176 "Active(file): %8lu kB\n"
177 "Inactive(file): %8lu kB\n"
178#ifdef CONFIG_UNEVICTABLE_LRU
179 "Unevictable: %8lu kB\n"
180 "Mlocked: %8lu kB\n"
181#endif
168#ifdef CONFIG_HIGHMEM 182#ifdef CONFIG_HIGHMEM
169 "HighTotal: %8lu kB\n" 183 "HighTotal: %8lu kB\n"
170 "HighFree: %8lu kB\n" 184 "HighFree: %8lu kB\n"
171 "LowTotal: %8lu kB\n" 185 "LowTotal: %8lu kB\n"
172 "LowFree: %8lu kB\n" 186 "LowFree: %8lu kB\n"
173#endif 187#endif
174 "SwapTotal: %8lu kB\n" 188 "SwapTotal: %8lu kB\n"
175 "SwapFree: %8lu kB\n" 189 "SwapFree: %8lu kB\n"
176 "Dirty: %8lu kB\n" 190 "Dirty: %8lu kB\n"
177 "Writeback: %8lu kB\n" 191 "Writeback: %8lu kB\n"
178 "AnonPages: %8lu kB\n" 192 "AnonPages: %8lu kB\n"
179 "Mapped: %8lu kB\n" 193 "Mapped: %8lu kB\n"
180 "Slab: %8lu kB\n" 194 "Slab: %8lu kB\n"
181 "SReclaimable: %8lu kB\n" 195 "SReclaimable: %8lu kB\n"
182 "SUnreclaim: %8lu kB\n" 196 "SUnreclaim: %8lu kB\n"
183 "PageTables: %8lu kB\n" 197 "PageTables: %8lu kB\n"
184#ifdef CONFIG_QUICKLIST 198#ifdef CONFIG_QUICKLIST
185 "Quicklists: %8lu kB\n" 199 "Quicklists: %8lu kB\n"
186#endif 200#endif
187 "NFS_Unstable: %8lu kB\n" 201 "NFS_Unstable: %8lu kB\n"
188 "Bounce: %8lu kB\n" 202 "Bounce: %8lu kB\n"
189 "WritebackTmp: %8lu kB\n" 203 "WritebackTmp: %8lu kB\n"
190 "CommitLimit: %8lu kB\n" 204 "CommitLimit: %8lu kB\n"
191 "Committed_AS: %8lu kB\n" 205 "Committed_AS: %8lu kB\n"
192 "VmallocTotal: %8lu kB\n" 206 "VmallocTotal: %8lu kB\n"
193 "VmallocUsed: %8lu kB\n" 207 "VmallocUsed: %8lu kB\n"
194 "VmallocChunk: %8lu kB\n", 208 "VmallocChunk: %8lu kB\n",
195 K(i.totalram), 209 K(i.totalram),
196 K(i.freeram), 210 K(i.freeram),
197 K(i.bufferram), 211 K(i.bufferram),
198 K(cached), 212 K(cached),
199 K(total_swapcache_pages), 213 K(total_swapcache_pages),
200 K(global_page_state(NR_ACTIVE)), 214 K(pages[LRU_ACTIVE_ANON] + pages[LRU_ACTIVE_FILE]),
201 K(global_page_state(NR_INACTIVE)), 215 K(pages[LRU_INACTIVE_ANON] + pages[LRU_INACTIVE_FILE]),
216 K(pages[LRU_ACTIVE_ANON]),
217 K(pages[LRU_INACTIVE_ANON]),
218 K(pages[LRU_ACTIVE_FILE]),
219 K(pages[LRU_INACTIVE_FILE]),
220#ifdef CONFIG_UNEVICTABLE_LRU
221 K(pages[LRU_UNEVICTABLE]),
222 K(global_page_state(NR_MLOCK)),
223#endif
202#ifdef CONFIG_HIGHMEM 224#ifdef CONFIG_HIGHMEM
203 K(i.totalhigh), 225 K(i.totalhigh),
204 K(i.freehigh), 226 K(i.freehigh),
@@ -500,17 +522,13 @@ static const struct file_operations proc_vmalloc_operations = {
500 522
501static int show_stat(struct seq_file *p, void *v) 523static int show_stat(struct seq_file *p, void *v)
502{ 524{
503 int i; 525 int i, j;
504 unsigned long jif; 526 unsigned long jif;
505 cputime64_t user, nice, system, idle, iowait, irq, softirq, steal; 527 cputime64_t user, nice, system, idle, iowait, irq, softirq, steal;
506 cputime64_t guest; 528 cputime64_t guest;
507 u64 sum = 0; 529 u64 sum = 0;
508 struct timespec boottime; 530 struct timespec boottime;
509 unsigned int *per_irq_sum; 531 unsigned int per_irq_sum;
510
511 per_irq_sum = kzalloc(sizeof(unsigned int)*NR_IRQS, GFP_KERNEL);
512 if (!per_irq_sum)
513 return -ENOMEM;
514 532
515 user = nice = system = idle = iowait = 533 user = nice = system = idle = iowait =
516 irq = softirq = steal = cputime64_zero; 534 irq = softirq = steal = cputime64_zero;
@@ -519,8 +537,6 @@ static int show_stat(struct seq_file *p, void *v)
519 jif = boottime.tv_sec; 537 jif = boottime.tv_sec;
520 538
521 for_each_possible_cpu(i) { 539 for_each_possible_cpu(i) {
522 int j;
523
524 user = cputime64_add(user, kstat_cpu(i).cpustat.user); 540 user = cputime64_add(user, kstat_cpu(i).cpustat.user);
525 nice = cputime64_add(nice, kstat_cpu(i).cpustat.nice); 541 nice = cputime64_add(nice, kstat_cpu(i).cpustat.nice);
526 system = cputime64_add(system, kstat_cpu(i).cpustat.system); 542 system = cputime64_add(system, kstat_cpu(i).cpustat.system);
@@ -530,11 +546,10 @@ static int show_stat(struct seq_file *p, void *v)
530 softirq = cputime64_add(softirq, kstat_cpu(i).cpustat.softirq); 546 softirq = cputime64_add(softirq, kstat_cpu(i).cpustat.softirq);
531 steal = cputime64_add(steal, kstat_cpu(i).cpustat.steal); 547 steal = cputime64_add(steal, kstat_cpu(i).cpustat.steal);
532 guest = cputime64_add(guest, kstat_cpu(i).cpustat.guest); 548 guest = cputime64_add(guest, kstat_cpu(i).cpustat.guest);
533 for (j = 0; j < NR_IRQS; j++) { 549
534 unsigned int temp = kstat_cpu(i).irqs[j]; 550 for_each_irq_nr(j)
535 sum += temp; 551 sum += kstat_irqs_cpu(j, i);
536 per_irq_sum[j] += temp; 552
537 }
538 sum += arch_irq_stat_cpu(i); 553 sum += arch_irq_stat_cpu(i);
539 } 554 }
540 sum += arch_irq_stat(); 555 sum += arch_irq_stat();
@@ -576,8 +591,15 @@ static int show_stat(struct seq_file *p, void *v)
576 } 591 }
577 seq_printf(p, "intr %llu", (unsigned long long)sum); 592 seq_printf(p, "intr %llu", (unsigned long long)sum);
578 593
579 for (i = 0; i < NR_IRQS; i++) 594 /* sum again ? it could be updated? */
580 seq_printf(p, " %u", per_irq_sum[i]); 595 for_each_irq_nr(j) {
596 per_irq_sum = 0;
597
598 for_each_possible_cpu(i)
599 per_irq_sum += kstat_irqs_cpu(j, i);
600
601 seq_printf(p, " %u", per_irq_sum);
602 }
581 603
582 seq_printf(p, 604 seq_printf(p,
583 "\nctxt %llu\n" 605 "\nctxt %llu\n"
@@ -591,7 +613,6 @@ static int show_stat(struct seq_file *p, void *v)
591 nr_running(), 613 nr_running(),
592 nr_iowait()); 614 nr_iowait());
593 615
594 kfree(per_irq_sum);
595 return 0; 616 return 0;
596} 617}
597 618
@@ -630,15 +651,14 @@ static const struct file_operations proc_stat_operations = {
630 */ 651 */
631static void *int_seq_start(struct seq_file *f, loff_t *pos) 652static void *int_seq_start(struct seq_file *f, loff_t *pos)
632{ 653{
633 return (*pos <= NR_IRQS) ? pos : NULL; 654 return (*pos <= nr_irqs) ? pos : NULL;
634} 655}
635 656
657
636static void *int_seq_next(struct seq_file *f, void *v, loff_t *pos) 658static void *int_seq_next(struct seq_file *f, void *v, loff_t *pos)
637{ 659{
638 (*pos)++; 660 (*pos)++;
639 if (*pos > NR_IRQS) 661 return (*pos <= nr_irqs) ? pos : NULL;
640 return NULL;
641 return pos;
642} 662}
643 663
644static void int_seq_stop(struct seq_file *f, void *v) 664static void int_seq_stop(struct seq_file *f, void *v)
@@ -646,7 +666,6 @@ static void int_seq_stop(struct seq_file *f, void *v)
646 /* Nothing to do */ 666 /* Nothing to do */
647} 667}
648 668
649
650static const struct seq_operations int_seq_ops = { 669static const struct seq_operations int_seq_ops = {
651 .start = int_seq_start, 670 .start = int_seq_start,
652 .next = int_seq_next, 671 .next = int_seq_next,
diff --git a/fs/proc/vmcore.c b/fs/proc/vmcore.c
index 841368b87a2..cd9ca67f841 100644
--- a/fs/proc/vmcore.c
+++ b/fs/proc/vmcore.c
@@ -32,9 +32,6 @@ static size_t elfcorebuf_sz;
32/* Total size of vmcore file. */ 32/* Total size of vmcore file. */
33static u64 vmcore_size; 33static u64 vmcore_size;
34 34
35/* Stores the physical address of elf header of crash image. */
36unsigned long long elfcorehdr_addr = ELFCORE_ADDR_MAX;
37
38struct proc_dir_entry *proc_vmcore = NULL; 35struct proc_dir_entry *proc_vmcore = NULL;
39 36
40/* Reads a page from the oldmem device from given offset. */ 37/* Reads a page from the oldmem device from given offset. */
@@ -647,7 +644,7 @@ static int __init vmcore_init(void)
647 int rc = 0; 644 int rc = 0;
648 645
649 /* If elfcorehdr= has been passed in cmdline, then capture the dump.*/ 646 /* If elfcorehdr= has been passed in cmdline, then capture the dump.*/
650 if (!(elfcorehdr_addr < ELFCORE_ADDR_MAX)) 647 if (!(is_vmcore_usable()))
651 return rc; 648 return rc;
652 rc = parse_crash_elf_headers(); 649 rc = parse_crash_elf_headers();
653 if (rc) { 650 if (rc) {
diff --git a/fs/ramfs/file-nommu.c b/fs/ramfs/file-nommu.c
index 5145cb9125a..76acdbc3461 100644
--- a/fs/ramfs/file-nommu.c
+++ b/fs/ramfs/file-nommu.c
@@ -112,12 +112,12 @@ int ramfs_nommu_expand_for_mapping(struct inode *inode, size_t newsize)
112 goto add_error; 112 goto add_error;
113 113
114 if (!pagevec_add(&lru_pvec, page)) 114 if (!pagevec_add(&lru_pvec, page))
115 __pagevec_lru_add(&lru_pvec); 115 __pagevec_lru_add_file(&lru_pvec);
116 116
117 unlock_page(page); 117 unlock_page(page);
118 } 118 }
119 119
120 pagevec_lru_add(&lru_pvec); 120 pagevec_lru_add_file(&lru_pvec);
121 return 0; 121 return 0;
122 122
123 fsize_exceeded: 123 fsize_exceeded:
diff --git a/fs/ramfs/inode.c b/fs/ramfs/inode.c
index b13123424e4..f031d1c925f 100644
--- a/fs/ramfs/inode.c
+++ b/fs/ramfs/inode.c
@@ -61,6 +61,7 @@ struct inode *ramfs_get_inode(struct super_block *sb, int mode, dev_t dev)
61 inode->i_mapping->a_ops = &ramfs_aops; 61 inode->i_mapping->a_ops = &ramfs_aops;
62 inode->i_mapping->backing_dev_info = &ramfs_backing_dev_info; 62 inode->i_mapping->backing_dev_info = &ramfs_backing_dev_info;
63 mapping_set_gfp_mask(inode->i_mapping, GFP_HIGHUSER); 63 mapping_set_gfp_mask(inode->i_mapping, GFP_HIGHUSER);
64 mapping_set_unevictable(inode->i_mapping);
64 inode->i_atime = inode->i_mtime = inode->i_ctime = CURRENT_TIME; 65 inode->i_atime = inode->i_mtime = inode->i_ctime = CURRENT_TIME;
65 switch (mode & S_IFMT) { 66 switch (mode & S_IFMT) {
66 default: 67 default:
diff --git a/fs/seq_file.c b/fs/seq_file.c
index bd20f7f5a93..eba2eabcd2b 100644
--- a/fs/seq_file.c
+++ b/fs/seq_file.c
@@ -452,17 +452,34 @@ int seq_dentry(struct seq_file *m, struct dentry *dentry, char *esc)
452 452
453int seq_bitmap(struct seq_file *m, unsigned long *bits, unsigned int nr_bits) 453int seq_bitmap(struct seq_file *m, unsigned long *bits, unsigned int nr_bits)
454{ 454{
455 size_t len = bitmap_scnprintf_len(nr_bits); 455 if (m->count < m->size) {
456 int len = bitmap_scnprintf(m->buf + m->count,
457 m->size - m->count, bits, nr_bits);
458 if (m->count + len < m->size) {
459 m->count += len;
460 return 0;
461 }
462 }
463 m->count = m->size;
464 return -1;
465}
466EXPORT_SYMBOL(seq_bitmap);
456 467
457 if (m->count + len < m->size) { 468int seq_bitmap_list(struct seq_file *m, unsigned long *bits,
458 bitmap_scnprintf(m->buf + m->count, m->size - m->count, 469 unsigned int nr_bits)
459 bits, nr_bits); 470{
460 m->count += len; 471 if (m->count < m->size) {
461 return 0; 472 int len = bitmap_scnlistprintf(m->buf + m->count,
473 m->size - m->count, bits, nr_bits);
474 if (m->count + len < m->size) {
475 m->count += len;
476 return 0;
477 }
462 } 478 }
463 m->count = m->size; 479 m->count = m->size;
464 return -1; 480 return -1;
465} 481}
482EXPORT_SYMBOL(seq_bitmap_list);
466 483
467static void *single_start(struct seq_file *p, loff_t *pos) 484static void *single_start(struct seq_file *p, loff_t *pos)
468{ 485{
diff --git a/fs/ubifs/budget.c b/fs/ubifs/budget.c
index 73db464cd08..1a4973e1066 100644
--- a/fs/ubifs/budget.c
+++ b/fs/ubifs/budget.c
@@ -414,19 +414,21 @@ static int do_budget_space(struct ubifs_info *c)
414 * @c->lst.empty_lebs + @c->freeable_cnt + @c->idx_gc_cnt - 414 * @c->lst.empty_lebs + @c->freeable_cnt + @c->idx_gc_cnt -
415 * @c->lst.taken_empty_lebs 415 * @c->lst.taken_empty_lebs
416 * 416 *
417 * @empty_lebs are available because they are empty. @freeable_cnt are 417 * @c->lst.empty_lebs are available because they are empty.
418 * available because they contain only free and dirty space and the 418 * @c->freeable_cnt are available because they contain only free and
419 * index allocation always occurs after wbufs are synch'ed. 419 * dirty space, @c->idx_gc_cnt are available because they are index
420 * @idx_gc_cnt are available because they are index LEBs that have been 420 * LEBs that have been garbage collected and are awaiting the commit
421 * garbage collected (including trivial GC) and are awaiting the commit 421 * before they can be used. And the in-the-gaps method will grab these
422 * before they can be unmapped - note that the in-the-gaps method will 422 * if it needs them. @c->lst.taken_empty_lebs are empty LEBs that have
423 * grab these if it needs them. @taken_empty_lebs are empty_lebs that 423 * already been allocated for some purpose.
424 * have already been allocated for some purpose (also includes those
425 * LEBs on the @idx_gc list).
426 * 424 *
427 * Note, @taken_empty_lebs may temporarily be higher by one because of 425 * Note, @c->idx_gc_cnt is included to both @c->lst.empty_lebs (because
428 * the way we serialize LEB allocations and budgeting. See a comment in 426 * these LEBs are empty) and to @c->lst.taken_empty_lebs (because they
429 * 'ubifs_find_free_space()'. 427 * are taken until after the commit).
428 *
429 * Note, @c->lst.taken_empty_lebs may temporarily be higher by one
430 * because of the way we serialize LEB allocations and budgeting. See a
431 * comment in 'ubifs_find_free_space()'.
430 */ 432 */
431 lebs = c->lst.empty_lebs + c->freeable_cnt + c->idx_gc_cnt - 433 lebs = c->lst.empty_lebs + c->freeable_cnt + c->idx_gc_cnt -
432 c->lst.taken_empty_lebs; 434 c->lst.taken_empty_lebs;
diff --git a/fs/ubifs/compress.c b/fs/ubifs/compress.c
index 5bb51dac3c1..a0ada596b17 100644
--- a/fs/ubifs/compress.c
+++ b/fs/ubifs/compress.c
@@ -91,8 +91,6 @@ struct ubifs_compressor *ubifs_compressors[UBIFS_COMPR_TYPES_CNT];
91 * 91 *
92 * Note, if the input buffer was not compressed, it is copied to the output 92 * Note, if the input buffer was not compressed, it is copied to the output
93 * buffer and %UBIFS_COMPR_NONE is returned in @compr_type. 93 * buffer and %UBIFS_COMPR_NONE is returned in @compr_type.
94 *
95 * This functions returns %0 on success or a negative error code on failure.
96 */ 94 */
97void ubifs_compress(const void *in_buf, int in_len, void *out_buf, int *out_len, 95void ubifs_compress(const void *in_buf, int in_len, void *out_buf, int *out_len,
98 int *compr_type) 96 int *compr_type)
diff --git a/fs/ubifs/debug.c b/fs/ubifs/debug.c
index d7f7645779f..7186400750e 100644
--- a/fs/ubifs/debug.c
+++ b/fs/ubifs/debug.c
@@ -222,30 +222,38 @@ void dbg_dump_inode(const struct ubifs_info *c, const struct inode *inode)
222{ 222{
223 const struct ubifs_inode *ui = ubifs_inode(inode); 223 const struct ubifs_inode *ui = ubifs_inode(inode);
224 224
225 printk(KERN_DEBUG "inode %lu\n", inode->i_ino); 225 printk(KERN_DEBUG "Dump in-memory inode:");
226 printk(KERN_DEBUG "size %llu\n", 226 printk(KERN_DEBUG "\tinode %lu\n", inode->i_ino);
227 printk(KERN_DEBUG "\tsize %llu\n",
227 (unsigned long long)i_size_read(inode)); 228 (unsigned long long)i_size_read(inode));
228 printk(KERN_DEBUG "nlink %u\n", inode->i_nlink); 229 printk(KERN_DEBUG "\tnlink %u\n", inode->i_nlink);
229 printk(KERN_DEBUG "uid %u\n", (unsigned int)inode->i_uid); 230 printk(KERN_DEBUG "\tuid %u\n", (unsigned int)inode->i_uid);
230 printk(KERN_DEBUG "gid %u\n", (unsigned int)inode->i_gid); 231 printk(KERN_DEBUG "\tgid %u\n", (unsigned int)inode->i_gid);
231 printk(KERN_DEBUG "atime %u.%u\n", 232 printk(KERN_DEBUG "\tatime %u.%u\n",
232 (unsigned int)inode->i_atime.tv_sec, 233 (unsigned int)inode->i_atime.tv_sec,
233 (unsigned int)inode->i_atime.tv_nsec); 234 (unsigned int)inode->i_atime.tv_nsec);
234 printk(KERN_DEBUG "mtime %u.%u\n", 235 printk(KERN_DEBUG "\tmtime %u.%u\n",
235 (unsigned int)inode->i_mtime.tv_sec, 236 (unsigned int)inode->i_mtime.tv_sec,
236 (unsigned int)inode->i_mtime.tv_nsec); 237 (unsigned int)inode->i_mtime.tv_nsec);
237 printk(KERN_DEBUG "ctime %u.%u\n", 238 printk(KERN_DEBUG "\tctime %u.%u\n",
238 (unsigned int)inode->i_ctime.tv_sec, 239 (unsigned int)inode->i_ctime.tv_sec,
239 (unsigned int)inode->i_ctime.tv_nsec); 240 (unsigned int)inode->i_ctime.tv_nsec);
240 printk(KERN_DEBUG "creat_sqnum %llu\n", ui->creat_sqnum); 241 printk(KERN_DEBUG "\tcreat_sqnum %llu\n", ui->creat_sqnum);
241 printk(KERN_DEBUG "xattr_size %u\n", ui->xattr_size); 242 printk(KERN_DEBUG "\txattr_size %u\n", ui->xattr_size);
242 printk(KERN_DEBUG "xattr_cnt %u\n", ui->xattr_cnt); 243 printk(KERN_DEBUG "\txattr_cnt %u\n", ui->xattr_cnt);
243 printk(KERN_DEBUG "xattr_names %u\n", ui->xattr_names); 244 printk(KERN_DEBUG "\txattr_names %u\n", ui->xattr_names);
244 printk(KERN_DEBUG "dirty %u\n", ui->dirty); 245 printk(KERN_DEBUG "\tdirty %u\n", ui->dirty);
245 printk(KERN_DEBUG "xattr %u\n", ui->xattr); 246 printk(KERN_DEBUG "\txattr %u\n", ui->xattr);
246 printk(KERN_DEBUG "flags %d\n", ui->flags); 247 printk(KERN_DEBUG "\tbulk_read %u\n", ui->xattr);
247 printk(KERN_DEBUG "compr_type %d\n", ui->compr_type); 248 printk(KERN_DEBUG "\tsynced_i_size %llu\n",
248 printk(KERN_DEBUG "data_len %d\n", ui->data_len); 249 (unsigned long long)ui->synced_i_size);
250 printk(KERN_DEBUG "\tui_size %llu\n",
251 (unsigned long long)ui->ui_size);
252 printk(KERN_DEBUG "\tflags %d\n", ui->flags);
253 printk(KERN_DEBUG "\tcompr_type %d\n", ui->compr_type);
254 printk(KERN_DEBUG "\tlast_page_read %lu\n", ui->last_page_read);
255 printk(KERN_DEBUG "\tread_in_a_row %lu\n", ui->read_in_a_row);
256 printk(KERN_DEBUG "\tdata_len %d\n", ui->data_len);
249} 257}
250 258
251void dbg_dump_node(const struct ubifs_info *c, const void *node) 259void dbg_dump_node(const struct ubifs_info *c, const void *node)
@@ -647,6 +655,43 @@ void dbg_dump_lprops(struct ubifs_info *c)
647 } 655 }
648} 656}
649 657
658void dbg_dump_lpt_info(struct ubifs_info *c)
659{
660 int i;
661
662 spin_lock(&dbg_lock);
663 printk(KERN_DEBUG "\tlpt_sz: %lld\n", c->lpt_sz);
664 printk(KERN_DEBUG "\tpnode_sz: %d\n", c->pnode_sz);
665 printk(KERN_DEBUG "\tnnode_sz: %d\n", c->nnode_sz);
666 printk(KERN_DEBUG "\tltab_sz: %d\n", c->ltab_sz);
667 printk(KERN_DEBUG "\tlsave_sz: %d\n", c->lsave_sz);
668 printk(KERN_DEBUG "\tbig_lpt: %d\n", c->big_lpt);
669 printk(KERN_DEBUG "\tlpt_hght: %d\n", c->lpt_hght);
670 printk(KERN_DEBUG "\tpnode_cnt: %d\n", c->pnode_cnt);
671 printk(KERN_DEBUG "\tnnode_cnt: %d\n", c->nnode_cnt);
672 printk(KERN_DEBUG "\tdirty_pn_cnt: %d\n", c->dirty_pn_cnt);
673 printk(KERN_DEBUG "\tdirty_nn_cnt: %d\n", c->dirty_nn_cnt);
674 printk(KERN_DEBUG "\tlsave_cnt: %d\n", c->lsave_cnt);
675 printk(KERN_DEBUG "\tspace_bits: %d\n", c->space_bits);
676 printk(KERN_DEBUG "\tlpt_lnum_bits: %d\n", c->lpt_lnum_bits);
677 printk(KERN_DEBUG "\tlpt_offs_bits: %d\n", c->lpt_offs_bits);
678 printk(KERN_DEBUG "\tlpt_spc_bits: %d\n", c->lpt_spc_bits);
679 printk(KERN_DEBUG "\tpcnt_bits: %d\n", c->pcnt_bits);
680 printk(KERN_DEBUG "\tlnum_bits: %d\n", c->lnum_bits);
681 printk(KERN_DEBUG "\tLPT root is at %d:%d\n", c->lpt_lnum, c->lpt_offs);
682 printk(KERN_DEBUG "\tLPT head is at %d:%d\n",
683 c->nhead_lnum, c->nhead_offs);
684 printk(KERN_DEBUG "\tLPT ltab is at %d:%d\n", c->ltab_lnum, c->ltab_offs);
685 if (c->big_lpt)
686 printk(KERN_DEBUG "\tLPT lsave is at %d:%d\n",
687 c->lsave_lnum, c->lsave_offs);
688 for (i = 0; i < c->lpt_lebs; i++)
689 printk(KERN_DEBUG "\tLPT LEB %d free %d dirty %d tgc %d "
690 "cmt %d\n", i + c->lpt_first, c->ltab[i].free,
691 c->ltab[i].dirty, c->ltab[i].tgc, c->ltab[i].cmt);
692 spin_unlock(&dbg_lock);
693}
694
650void dbg_dump_leb(const struct ubifs_info *c, int lnum) 695void dbg_dump_leb(const struct ubifs_info *c, int lnum)
651{ 696{
652 struct ubifs_scan_leb *sleb; 697 struct ubifs_scan_leb *sleb;
diff --git a/fs/ubifs/debug.h b/fs/ubifs/debug.h
index 50315fc5718..33d6b95071e 100644
--- a/fs/ubifs/debug.h
+++ b/fs/ubifs/debug.h
@@ -224,6 +224,7 @@ void dbg_dump_lstats(const struct ubifs_lp_stats *lst);
224void dbg_dump_budg(struct ubifs_info *c); 224void dbg_dump_budg(struct ubifs_info *c);
225void dbg_dump_lprop(const struct ubifs_info *c, const struct ubifs_lprops *lp); 225void dbg_dump_lprop(const struct ubifs_info *c, const struct ubifs_lprops *lp);
226void dbg_dump_lprops(struct ubifs_info *c); 226void dbg_dump_lprops(struct ubifs_info *c);
227void dbg_dump_lpt_info(struct ubifs_info *c);
227void dbg_dump_leb(const struct ubifs_info *c, int lnum); 228void dbg_dump_leb(const struct ubifs_info *c, int lnum);
228void dbg_dump_znode(const struct ubifs_info *c, 229void dbg_dump_znode(const struct ubifs_info *c,
229 const struct ubifs_znode *znode); 230 const struct ubifs_znode *znode);
@@ -249,6 +250,8 @@ int dbg_old_index_check_init(struct ubifs_info *c, struct ubifs_zbranch *zroot);
249int dbg_check_old_index(struct ubifs_info *c, struct ubifs_zbranch *zroot); 250int dbg_check_old_index(struct ubifs_info *c, struct ubifs_zbranch *zroot);
250int dbg_check_cats(struct ubifs_info *c); 251int dbg_check_cats(struct ubifs_info *c);
251int dbg_check_ltab(struct ubifs_info *c); 252int dbg_check_ltab(struct ubifs_info *c);
253int dbg_chk_lpt_free_spc(struct ubifs_info *c);
254int dbg_chk_lpt_sz(struct ubifs_info *c, int action, int len);
252int dbg_check_synced_i_size(struct inode *inode); 255int dbg_check_synced_i_size(struct inode *inode);
253int dbg_check_dir_size(struct ubifs_info *c, const struct inode *dir); 256int dbg_check_dir_size(struct ubifs_info *c, const struct inode *dir);
254int dbg_check_tnc(struct ubifs_info *c, int extra); 257int dbg_check_tnc(struct ubifs_info *c, int extra);
@@ -367,6 +370,7 @@ static inline int dbg_change(struct ubi_volume_desc *desc, int lnum,
367#define dbg_dump_budg(c) ({}) 370#define dbg_dump_budg(c) ({})
368#define dbg_dump_lprop(c, lp) ({}) 371#define dbg_dump_lprop(c, lp) ({})
369#define dbg_dump_lprops(c) ({}) 372#define dbg_dump_lprops(c) ({})
373#define dbg_dump_lpt_info(c) ({})
370#define dbg_dump_leb(c, lnum) ({}) 374#define dbg_dump_leb(c, lnum) ({})
371#define dbg_dump_znode(c, znode) ({}) 375#define dbg_dump_znode(c, znode) ({})
372#define dbg_dump_heap(c, heap, cat) ({}) 376#define dbg_dump_heap(c, heap, cat) ({})
@@ -379,6 +383,8 @@ static inline int dbg_change(struct ubi_volume_desc *desc, int lnum,
379#define dbg_check_old_index(c, zroot) 0 383#define dbg_check_old_index(c, zroot) 0
380#define dbg_check_cats(c) 0 384#define dbg_check_cats(c) 0
381#define dbg_check_ltab(c) 0 385#define dbg_check_ltab(c) 0
386#define dbg_chk_lpt_free_spc(c) 0
387#define dbg_chk_lpt_sz(c, action, len) 0
382#define dbg_check_synced_i_size(inode) 0 388#define dbg_check_synced_i_size(inode) 0
383#define dbg_check_dir_size(c, dir) 0 389#define dbg_check_dir_size(c, dir) 0
384#define dbg_check_tnc(c, x) 0 390#define dbg_check_tnc(c, x) 0
diff --git a/fs/ubifs/file.c b/fs/ubifs/file.c
index 3d698e2022b..51cf511d44d 100644
--- a/fs/ubifs/file.c
+++ b/fs/ubifs/file.c
@@ -147,6 +147,12 @@ static int do_readpage(struct page *page)
147 err = ret; 147 err = ret;
148 if (err != -ENOENT) 148 if (err != -ENOENT)
149 break; 149 break;
150 } else if (block + 1 == beyond) {
151 int dlen = le32_to_cpu(dn->size);
152 int ilen = i_size & (UBIFS_BLOCK_SIZE - 1);
153
154 if (ilen && ilen < dlen)
155 memset(addr + ilen, 0, dlen - ilen);
150 } 156 }
151 } 157 }
152 if (++i >= UBIFS_BLOCKS_PER_PAGE) 158 if (++i >= UBIFS_BLOCKS_PER_PAGE)
@@ -577,8 +583,262 @@ out:
577 return copied; 583 return copied;
578} 584}
579 585
586/**
587 * populate_page - copy data nodes into a page for bulk-read.
588 * @c: UBIFS file-system description object
589 * @page: page
590 * @bu: bulk-read information
591 * @n: next zbranch slot
592 *
593 * This function returns %0 on success and a negative error code on failure.
594 */
595static int populate_page(struct ubifs_info *c, struct page *page,
596 struct bu_info *bu, int *n)
597{
598 int i = 0, nn = *n, offs = bu->zbranch[0].offs, hole = 0, read = 0;
599 struct inode *inode = page->mapping->host;
600 loff_t i_size = i_size_read(inode);
601 unsigned int page_block;
602 void *addr, *zaddr;
603 pgoff_t end_index;
604
605 dbg_gen("ino %lu, pg %lu, i_size %lld, flags %#lx",
606 inode->i_ino, page->index, i_size, page->flags);
607
608 addr = zaddr = kmap(page);
609
610 end_index = (i_size - 1) >> PAGE_CACHE_SHIFT;
611 if (!i_size || page->index > end_index) {
612 hole = 1;
613 memset(addr, 0, PAGE_CACHE_SIZE);
614 goto out_hole;
615 }
616
617 page_block = page->index << UBIFS_BLOCKS_PER_PAGE_SHIFT;
618 while (1) {
619 int err, len, out_len, dlen;
620
621 if (nn >= bu->cnt) {
622 hole = 1;
623 memset(addr, 0, UBIFS_BLOCK_SIZE);
624 } else if (key_block(c, &bu->zbranch[nn].key) == page_block) {
625 struct ubifs_data_node *dn;
626
627 dn = bu->buf + (bu->zbranch[nn].offs - offs);
628
629 ubifs_assert(dn->ch.sqnum >
630 ubifs_inode(inode)->creat_sqnum);
631
632 len = le32_to_cpu(dn->size);
633 if (len <= 0 || len > UBIFS_BLOCK_SIZE)
634 goto out_err;
635
636 dlen = le32_to_cpu(dn->ch.len) - UBIFS_DATA_NODE_SZ;
637 out_len = UBIFS_BLOCK_SIZE;
638 err = ubifs_decompress(&dn->data, dlen, addr, &out_len,
639 le16_to_cpu(dn->compr_type));
640 if (err || len != out_len)
641 goto out_err;
642
643 if (len < UBIFS_BLOCK_SIZE)
644 memset(addr + len, 0, UBIFS_BLOCK_SIZE - len);
645
646 nn += 1;
647 read = (i << UBIFS_BLOCK_SHIFT) + len;
648 } else if (key_block(c, &bu->zbranch[nn].key) < page_block) {
649 nn += 1;
650 continue;
651 } else {
652 hole = 1;
653 memset(addr, 0, UBIFS_BLOCK_SIZE);
654 }
655 if (++i >= UBIFS_BLOCKS_PER_PAGE)
656 break;
657 addr += UBIFS_BLOCK_SIZE;
658 page_block += 1;
659 }
660
661 if (end_index == page->index) {
662 int len = i_size & (PAGE_CACHE_SIZE - 1);
663
664 if (len && len < read)
665 memset(zaddr + len, 0, read - len);
666 }
667
668out_hole:
669 if (hole) {
670 SetPageChecked(page);
671 dbg_gen("hole");
672 }
673
674 SetPageUptodate(page);
675 ClearPageError(page);
676 flush_dcache_page(page);
677 kunmap(page);
678 *n = nn;
679 return 0;
680
681out_err:
682 ClearPageUptodate(page);
683 SetPageError(page);
684 flush_dcache_page(page);
685 kunmap(page);
686 ubifs_err("bad data node (block %u, inode %lu)",
687 page_block, inode->i_ino);
688 return -EINVAL;
689}
690
691/**
692 * ubifs_do_bulk_read - do bulk-read.
693 * @c: UBIFS file-system description object
694 * @page1: first page
695 *
696 * This function returns %1 if the bulk-read is done, otherwise %0 is returned.
697 */
698static int ubifs_do_bulk_read(struct ubifs_info *c, struct page *page1)
699{
700 pgoff_t offset = page1->index, end_index;
701 struct address_space *mapping = page1->mapping;
702 struct inode *inode = mapping->host;
703 struct ubifs_inode *ui = ubifs_inode(inode);
704 struct bu_info *bu;
705 int err, page_idx, page_cnt, ret = 0, n = 0;
706 loff_t isize;
707
708 bu = kmalloc(sizeof(struct bu_info), GFP_NOFS);
709 if (!bu)
710 return 0;
711
712 bu->buf_len = c->bulk_read_buf_size;
713 bu->buf = kmalloc(bu->buf_len, GFP_NOFS);
714 if (!bu->buf)
715 goto out_free;
716
717 data_key_init(c, &bu->key, inode->i_ino,
718 offset << UBIFS_BLOCKS_PER_PAGE_SHIFT);
719
720 err = ubifs_tnc_get_bu_keys(c, bu);
721 if (err)
722 goto out_warn;
723
724 if (bu->eof) {
725 /* Turn off bulk-read at the end of the file */
726 ui->read_in_a_row = 1;
727 ui->bulk_read = 0;
728 }
729
730 page_cnt = bu->blk_cnt >> UBIFS_BLOCKS_PER_PAGE_SHIFT;
731 if (!page_cnt) {
732 /*
733 * This happens when there are multiple blocks per page and the
734 * blocks for the first page we are looking for, are not
735 * together. If all the pages were like this, bulk-read would
736 * reduce performance, so we turn it off for a while.
737 */
738 ui->read_in_a_row = 0;
739 ui->bulk_read = 0;
740 goto out_free;
741 }
742
743 if (bu->cnt) {
744 err = ubifs_tnc_bulk_read(c, bu);
745 if (err)
746 goto out_warn;
747 }
748
749 err = populate_page(c, page1, bu, &n);
750 if (err)
751 goto out_warn;
752
753 unlock_page(page1);
754 ret = 1;
755
756 isize = i_size_read(inode);
757 if (isize == 0)
758 goto out_free;
759 end_index = ((isize - 1) >> PAGE_CACHE_SHIFT);
760
761 for (page_idx = 1; page_idx < page_cnt; page_idx++) {
762 pgoff_t page_offset = offset + page_idx;
763 struct page *page;
764
765 if (page_offset > end_index)
766 break;
767 page = find_or_create_page(mapping, page_offset,
768 GFP_NOFS | __GFP_COLD);
769 if (!page)
770 break;
771 if (!PageUptodate(page))
772 err = populate_page(c, page, bu, &n);
773 unlock_page(page);
774 page_cache_release(page);
775 if (err)
776 break;
777 }
778
779 ui->last_page_read = offset + page_idx - 1;
780
781out_free:
782 kfree(bu->buf);
783 kfree(bu);
784 return ret;
785
786out_warn:
787 ubifs_warn("ignoring error %d and skipping bulk-read", err);
788 goto out_free;
789}
790
791/**
792 * ubifs_bulk_read - determine whether to bulk-read and, if so, do it.
793 * @page: page from which to start bulk-read.
794 *
795 * Some flash media are capable of reading sequentially at faster rates. UBIFS
796 * bulk-read facility is designed to take advantage of that, by reading in one
797 * go consecutive data nodes that are also located consecutively in the same
798 * LEB. This function returns %1 if a bulk-read is done and %0 otherwise.
799 */
800static int ubifs_bulk_read(struct page *page)
801{
802 struct inode *inode = page->mapping->host;
803 struct ubifs_info *c = inode->i_sb->s_fs_info;
804 struct ubifs_inode *ui = ubifs_inode(inode);
805 pgoff_t index = page->index, last_page_read = ui->last_page_read;
806 int ret = 0;
807
808 ui->last_page_read = index;
809
810 if (!c->bulk_read)
811 return 0;
812 /*
813 * Bulk-read is protected by ui_mutex, but it is an optimization, so
814 * don't bother if we cannot lock the mutex.
815 */
816 if (!mutex_trylock(&ui->ui_mutex))
817 return 0;
818 if (index != last_page_read + 1) {
819 /* Turn off bulk-read if we stop reading sequentially */
820 ui->read_in_a_row = 1;
821 if (ui->bulk_read)
822 ui->bulk_read = 0;
823 goto out_unlock;
824 }
825 if (!ui->bulk_read) {
826 ui->read_in_a_row += 1;
827 if (ui->read_in_a_row < 3)
828 goto out_unlock;
829 /* Three reads in a row, so switch on bulk-read */
830 ui->bulk_read = 1;
831 }
832 ret = ubifs_do_bulk_read(c, page);
833out_unlock:
834 mutex_unlock(&ui->ui_mutex);
835 return ret;
836}
837
580static int ubifs_readpage(struct file *file, struct page *page) 838static int ubifs_readpage(struct file *file, struct page *page)
581{ 839{
840 if (ubifs_bulk_read(page))
841 return 0;
582 do_readpage(page); 842 do_readpage(page);
583 unlock_page(page); 843 unlock_page(page);
584 return 0; 844 return 0;
diff --git a/fs/ubifs/find.c b/fs/ubifs/find.c
index 47814cde240..717d79c97c5 100644
--- a/fs/ubifs/find.c
+++ b/fs/ubifs/find.c
@@ -901,11 +901,11 @@ static int get_idx_gc_leb(struct ubifs_info *c)
901 * it is needed now for this commit. 901 * it is needed now for this commit.
902 */ 902 */
903 lp = ubifs_lpt_lookup_dirty(c, lnum); 903 lp = ubifs_lpt_lookup_dirty(c, lnum);
904 if (unlikely(IS_ERR(lp))) 904 if (IS_ERR(lp))
905 return PTR_ERR(lp); 905 return PTR_ERR(lp);
906 lp = ubifs_change_lp(c, lp, LPROPS_NC, LPROPS_NC, 906 lp = ubifs_change_lp(c, lp, LPROPS_NC, LPROPS_NC,
907 lp->flags | LPROPS_INDEX, -1); 907 lp->flags | LPROPS_INDEX, -1);
908 if (unlikely(IS_ERR(lp))) 908 if (IS_ERR(lp))
909 return PTR_ERR(lp); 909 return PTR_ERR(lp);
910 dbg_find("LEB %d, dirty %d and free %d flags %#x", 910 dbg_find("LEB %d, dirty %d and free %d flags %#x",
911 lp->lnum, lp->dirty, lp->free, lp->flags); 911 lp->lnum, lp->dirty, lp->free, lp->flags);
diff --git a/fs/ubifs/gc.c b/fs/ubifs/gc.c
index 02aba36fe3d..0bef6501d58 100644
--- a/fs/ubifs/gc.c
+++ b/fs/ubifs/gc.c
@@ -96,6 +96,48 @@ static int switch_gc_head(struct ubifs_info *c)
96} 96}
97 97
98/** 98/**
99 * joinup - bring data nodes for an inode together.
100 * @c: UBIFS file-system description object
101 * @sleb: describes scanned LEB
102 * @inum: inode number
103 * @blk: block number
104 * @data: list to which to add data nodes
105 *
106 * This function looks at the first few nodes in the scanned LEB @sleb and adds
107 * them to @data if they are data nodes from @inum and have a larger block
108 * number than @blk. This function returns %0 on success and a negative error
109 * code on failure.
110 */
111static int joinup(struct ubifs_info *c, struct ubifs_scan_leb *sleb, ino_t inum,
112 unsigned int blk, struct list_head *data)
113{
114 int err, cnt = 6, lnum = sleb->lnum, offs;
115 struct ubifs_scan_node *snod, *tmp;
116 union ubifs_key *key;
117
118 list_for_each_entry_safe(snod, tmp, &sleb->nodes, list) {
119 key = &snod->key;
120 if (key_inum(c, key) == inum &&
121 key_type(c, key) == UBIFS_DATA_KEY &&
122 key_block(c, key) > blk) {
123 offs = snod->offs;
124 err = ubifs_tnc_has_node(c, key, 0, lnum, offs, 0);
125 if (err < 0)
126 return err;
127 list_del(&snod->list);
128 if (err) {
129 list_add_tail(&snod->list, data);
130 blk = key_block(c, key);
131 } else
132 kfree(snod);
133 cnt = 6;
134 } else if (--cnt == 0)
135 break;
136 }
137 return 0;
138}
139
140/**
99 * move_nodes - move nodes. 141 * move_nodes - move nodes.
100 * @c: UBIFS file-system description object 142 * @c: UBIFS file-system description object
101 * @sleb: describes nodes to move 143 * @sleb: describes nodes to move
@@ -116,16 +158,21 @@ static int switch_gc_head(struct ubifs_info *c)
116static int move_nodes(struct ubifs_info *c, struct ubifs_scan_leb *sleb) 158static int move_nodes(struct ubifs_info *c, struct ubifs_scan_leb *sleb)
117{ 159{
118 struct ubifs_scan_node *snod, *tmp; 160 struct ubifs_scan_node *snod, *tmp;
119 struct list_head large, medium, small; 161 struct list_head data, large, medium, small;
120 struct ubifs_wbuf *wbuf = &c->jheads[GCHD].wbuf; 162 struct ubifs_wbuf *wbuf = &c->jheads[GCHD].wbuf;
121 int avail, err, min = INT_MAX; 163 int avail, err, min = INT_MAX;
164 unsigned int blk = 0;
165 ino_t inum = 0;
122 166
167 INIT_LIST_HEAD(&data);
123 INIT_LIST_HEAD(&large); 168 INIT_LIST_HEAD(&large);
124 INIT_LIST_HEAD(&medium); 169 INIT_LIST_HEAD(&medium);
125 INIT_LIST_HEAD(&small); 170 INIT_LIST_HEAD(&small);
126 171
127 list_for_each_entry_safe(snod, tmp, &sleb->nodes, list) { 172 while (!list_empty(&sleb->nodes)) {
128 struct list_head *lst; 173 struct list_head *lst = sleb->nodes.next;
174
175 snod = list_entry(lst, struct ubifs_scan_node, list);
129 176
130 ubifs_assert(snod->type != UBIFS_IDX_NODE); 177 ubifs_assert(snod->type != UBIFS_IDX_NODE);
131 ubifs_assert(snod->type != UBIFS_REF_NODE); 178 ubifs_assert(snod->type != UBIFS_REF_NODE);
@@ -136,7 +183,6 @@ static int move_nodes(struct ubifs_info *c, struct ubifs_scan_leb *sleb)
136 if (err < 0) 183 if (err < 0)
137 goto out; 184 goto out;
138 185
139 lst = &snod->list;
140 list_del(lst); 186 list_del(lst);
141 if (!err) { 187 if (!err) {
142 /* The node is obsolete, remove it from the list */ 188 /* The node is obsolete, remove it from the list */
@@ -145,15 +191,30 @@ static int move_nodes(struct ubifs_info *c, struct ubifs_scan_leb *sleb)
145 } 191 }
146 192
147 /* 193 /*
148 * Sort the list of nodes so that large nodes go first, and 194 * Sort the list of nodes so that data nodes go first, large
149 * small nodes go last. 195 * nodes go second, and small nodes go last.
150 */ 196 */
151 if (snod->len > MEDIUM_NODE_WM) 197 if (key_type(c, &snod->key) == UBIFS_DATA_KEY) {
152 list_add(lst, &large); 198 if (inum != key_inum(c, &snod->key)) {
199 if (inum) {
200 /*
201 * Try to move data nodes from the same
202 * inode together.
203 */
204 err = joinup(c, sleb, inum, blk, &data);
205 if (err)
206 goto out;
207 }
208 inum = key_inum(c, &snod->key);
209 blk = key_block(c, &snod->key);
210 }
211 list_add_tail(lst, &data);
212 } else if (snod->len > MEDIUM_NODE_WM)
213 list_add_tail(lst, &large);
153 else if (snod->len > SMALL_NODE_WM) 214 else if (snod->len > SMALL_NODE_WM)
154 list_add(lst, &medium); 215 list_add_tail(lst, &medium);
155 else 216 else
156 list_add(lst, &small); 217 list_add_tail(lst, &small);
157 218
158 /* And find the smallest node */ 219 /* And find the smallest node */
159 if (snod->len < min) 220 if (snod->len < min)
@@ -164,6 +225,7 @@ static int move_nodes(struct ubifs_info *c, struct ubifs_scan_leb *sleb)
164 * Join the tree lists so that we'd have one roughly sorted list 225 * Join the tree lists so that we'd have one roughly sorted list
165 * ('large' will be the head of the joined list). 226 * ('large' will be the head of the joined list).
166 */ 227 */
228 list_splice(&data, &large);
167 list_splice(&medium, large.prev); 229 list_splice(&medium, large.prev);
168 list_splice(&small, large.prev); 230 list_splice(&small, large.prev);
169 231
@@ -653,7 +715,7 @@ int ubifs_gc_start_commit(struct ubifs_info *c)
653 */ 715 */
654 while (1) { 716 while (1) {
655 lp = ubifs_fast_find_freeable(c); 717 lp = ubifs_fast_find_freeable(c);
656 if (unlikely(IS_ERR(lp))) { 718 if (IS_ERR(lp)) {
657 err = PTR_ERR(lp); 719 err = PTR_ERR(lp);
658 goto out; 720 goto out;
659 } 721 }
@@ -665,7 +727,7 @@ int ubifs_gc_start_commit(struct ubifs_info *c)
665 if (err) 727 if (err)
666 goto out; 728 goto out;
667 lp = ubifs_change_lp(c, lp, c->leb_size, 0, lp->flags, 0); 729 lp = ubifs_change_lp(c, lp, c->leb_size, 0, lp->flags, 0);
668 if (unlikely(IS_ERR(lp))) { 730 if (IS_ERR(lp)) {
669 err = PTR_ERR(lp); 731 err = PTR_ERR(lp);
670 goto out; 732 goto out;
671 } 733 }
@@ -680,7 +742,7 @@ int ubifs_gc_start_commit(struct ubifs_info *c)
680 /* Record index freeable LEBs for unmapping after commit */ 742 /* Record index freeable LEBs for unmapping after commit */
681 while (1) { 743 while (1) {
682 lp = ubifs_fast_find_frdi_idx(c); 744 lp = ubifs_fast_find_frdi_idx(c);
683 if (unlikely(IS_ERR(lp))) { 745 if (IS_ERR(lp)) {
684 err = PTR_ERR(lp); 746 err = PTR_ERR(lp);
685 goto out; 747 goto out;
686 } 748 }
@@ -696,7 +758,7 @@ int ubifs_gc_start_commit(struct ubifs_info *c)
696 /* Don't release the LEB until after the next commit */ 758 /* Don't release the LEB until after the next commit */
697 flags = (lp->flags | LPROPS_TAKEN) ^ LPROPS_INDEX; 759 flags = (lp->flags | LPROPS_TAKEN) ^ LPROPS_INDEX;
698 lp = ubifs_change_lp(c, lp, c->leb_size, 0, flags, 1); 760 lp = ubifs_change_lp(c, lp, c->leb_size, 0, flags, 1);
699 if (unlikely(IS_ERR(lp))) { 761 if (IS_ERR(lp)) {
700 err = PTR_ERR(lp); 762 err = PTR_ERR(lp);
701 kfree(idx_gc); 763 kfree(idx_gc);
702 goto out; 764 goto out;
diff --git a/fs/ubifs/io.c b/fs/ubifs/io.c
index 054363f2b20..01682713af6 100644
--- a/fs/ubifs/io.c
+++ b/fs/ubifs/io.c
@@ -62,6 +62,7 @@ void ubifs_ro_mode(struct ubifs_info *c, int err)
62{ 62{
63 if (!c->ro_media) { 63 if (!c->ro_media) {
64 c->ro_media = 1; 64 c->ro_media = 1;
65 c->no_chk_data_crc = 0;
65 ubifs_warn("switched to read-only mode, error %d", err); 66 ubifs_warn("switched to read-only mode, error %d", err);
66 dbg_dump_stack(); 67 dbg_dump_stack();
67 } 68 }
@@ -74,6 +75,7 @@ void ubifs_ro_mode(struct ubifs_info *c, int err)
74 * @lnum: logical eraseblock number 75 * @lnum: logical eraseblock number
75 * @offs: offset within the logical eraseblock 76 * @offs: offset within the logical eraseblock
76 * @quiet: print no messages 77 * @quiet: print no messages
78 * @chk_crc: indicates whether to always check the CRC
77 * 79 *
78 * This function checks node magic number and CRC checksum. This function also 80 * This function checks node magic number and CRC checksum. This function also
79 * validates node length to prevent UBIFS from becoming crazy when an attacker 81 * validates node length to prevent UBIFS from becoming crazy when an attacker
@@ -85,7 +87,7 @@ void ubifs_ro_mode(struct ubifs_info *c, int err)
85 * or magic. 87 * or magic.
86 */ 88 */
87int ubifs_check_node(const struct ubifs_info *c, const void *buf, int lnum, 89int ubifs_check_node(const struct ubifs_info *c, const void *buf, int lnum,
88 int offs, int quiet) 90 int offs, int quiet, int chk_crc)
89{ 91{
90 int err = -EINVAL, type, node_len; 92 int err = -EINVAL, type, node_len;
91 uint32_t crc, node_crc, magic; 93 uint32_t crc, node_crc, magic;
@@ -121,6 +123,10 @@ int ubifs_check_node(const struct ubifs_info *c, const void *buf, int lnum,
121 node_len > c->ranges[type].max_len) 123 node_len > c->ranges[type].max_len)
122 goto out_len; 124 goto out_len;
123 125
126 if (!chk_crc && type == UBIFS_DATA_NODE && !c->always_chk_crc)
127 if (c->no_chk_data_crc)
128 return 0;
129
124 crc = crc32(UBIFS_CRC32_INIT, buf + 8, node_len - 8); 130 crc = crc32(UBIFS_CRC32_INIT, buf + 8, node_len - 8);
125 node_crc = le32_to_cpu(ch->crc); 131 node_crc = le32_to_cpu(ch->crc);
126 if (crc != node_crc) { 132 if (crc != node_crc) {
@@ -722,7 +728,7 @@ int ubifs_read_node_wbuf(struct ubifs_wbuf *wbuf, void *buf, int type, int len,
722 goto out; 728 goto out;
723 } 729 }
724 730
725 err = ubifs_check_node(c, buf, lnum, offs, 0); 731 err = ubifs_check_node(c, buf, lnum, offs, 0, 0);
726 if (err) { 732 if (err) {
727 ubifs_err("expected node type %d", type); 733 ubifs_err("expected node type %d", type);
728 return err; 734 return err;
@@ -781,7 +787,7 @@ int ubifs_read_node(const struct ubifs_info *c, void *buf, int type, int len,
781 goto out; 787 goto out;
782 } 788 }
783 789
784 err = ubifs_check_node(c, buf, lnum, offs, 0); 790 err = ubifs_check_node(c, buf, lnum, offs, 0, 0);
785 if (err) { 791 if (err) {
786 ubifs_err("expected node type %d", type); 792 ubifs_err("expected node type %d", type);
787 return err; 793 return err;
diff --git a/fs/ubifs/key.h b/fs/ubifs/key.h
index 8f747600754..9ee65086f62 100644
--- a/fs/ubifs/key.h
+++ b/fs/ubifs/key.h
@@ -484,7 +484,7 @@ static inline void key_copy(const struct ubifs_info *c,
484 * @key2: the second key to compare 484 * @key2: the second key to compare
485 * 485 *
486 * This function compares 2 keys and returns %-1 if @key1 is less than 486 * This function compares 2 keys and returns %-1 if @key1 is less than
487 * @key2, 0 if the keys are equivalent and %1 if @key1 is greater than @key2. 487 * @key2, %0 if the keys are equivalent and %1 if @key1 is greater than @key2.
488 */ 488 */
489static inline int keys_cmp(const struct ubifs_info *c, 489static inline int keys_cmp(const struct ubifs_info *c,
490 const union ubifs_key *key1, 490 const union ubifs_key *key1,
@@ -503,6 +503,26 @@ static inline int keys_cmp(const struct ubifs_info *c,
503} 503}
504 504
505/** 505/**
506 * keys_eq - determine if keys are equivalent.
507 * @c: UBIFS file-system description object
508 * @key1: the first key to compare
509 * @key2: the second key to compare
510 *
511 * This function compares 2 keys and returns %1 if @key1 is equal to @key2 and
512 * %0 if not.
513 */
514static inline int keys_eq(const struct ubifs_info *c,
515 const union ubifs_key *key1,
516 const union ubifs_key *key2)
517{
518 if (key1->u32[0] != key2->u32[0])
519 return 0;
520 if (key1->u32[1] != key2->u32[1])
521 return 0;
522 return 1;
523}
524
525/**
506 * is_hash_key - is a key vulnerable to hash collisions. 526 * is_hash_key - is a key vulnerable to hash collisions.
507 * @c: UBIFS file-system description object 527 * @c: UBIFS file-system description object
508 * @key: key 528 * @key: key
diff --git a/fs/ubifs/lprops.c b/fs/ubifs/lprops.c
index 2ba93da71b6..f27176e9b70 100644
--- a/fs/ubifs/lprops.c
+++ b/fs/ubifs/lprops.c
@@ -125,6 +125,7 @@ static void adjust_lpt_heap(struct ubifs_info *c, struct ubifs_lpt_heap *heap,
125 } 125 }
126 } 126 }
127 } 127 }
128
128 /* Not greater than parent, so compare to children */ 129 /* Not greater than parent, so compare to children */
129 while (1) { 130 while (1) {
130 /* Compare to left child */ 131 /* Compare to left child */
@@ -460,18 +461,6 @@ static void change_category(struct ubifs_info *c, struct ubifs_lprops *lprops)
460} 461}
461 462
462/** 463/**
463 * ubifs_get_lprops - get reference to LEB properties.
464 * @c: the UBIFS file-system description object
465 *
466 * This function locks lprops. Lprops have to be unlocked by
467 * 'ubifs_release_lprops()'.
468 */
469void ubifs_get_lprops(struct ubifs_info *c)
470{
471 mutex_lock(&c->lp_mutex);
472}
473
474/**
475 * calc_dark - calculate LEB dark space size. 464 * calc_dark - calculate LEB dark space size.
476 * @c: the UBIFS file-system description object 465 * @c: the UBIFS file-system description object
477 * @spc: amount of free and dirty space in the LEB 466 * @spc: amount of free and dirty space in the LEB
@@ -576,7 +565,6 @@ const struct ubifs_lprops *ubifs_change_lp(struct ubifs_info *c,
576 ubifs_assert(!(lprops->free & 7) && !(lprops->dirty & 7)); 565 ubifs_assert(!(lprops->free & 7) && !(lprops->dirty & 7));
577 566
578 spin_lock(&c->space_lock); 567 spin_lock(&c->space_lock);
579
580 if ((lprops->flags & LPROPS_TAKEN) && lprops->free == c->leb_size) 568 if ((lprops->flags & LPROPS_TAKEN) && lprops->free == c->leb_size)
581 c->lst.taken_empty_lebs -= 1; 569 c->lst.taken_empty_lebs -= 1;
582 570
@@ -637,31 +625,12 @@ const struct ubifs_lprops *ubifs_change_lp(struct ubifs_info *c,
637 c->lst.taken_empty_lebs += 1; 625 c->lst.taken_empty_lebs += 1;
638 626
639 change_category(c, lprops); 627 change_category(c, lprops);
640
641 c->idx_gc_cnt += idx_gc_cnt; 628 c->idx_gc_cnt += idx_gc_cnt;
642
643 spin_unlock(&c->space_lock); 629 spin_unlock(&c->space_lock);
644
645 return lprops; 630 return lprops;
646} 631}
647 632
648/** 633/**
649 * ubifs_release_lprops - release lprops lock.
650 * @c: the UBIFS file-system description object
651 *
652 * This function has to be called after each 'ubifs_get_lprops()' call to
653 * unlock lprops.
654 */
655void ubifs_release_lprops(struct ubifs_info *c)
656{
657 ubifs_assert(mutex_is_locked(&c->lp_mutex));
658 ubifs_assert(c->lst.empty_lebs >= 0 &&
659 c->lst.empty_lebs <= c->main_lebs);
660
661 mutex_unlock(&c->lp_mutex);
662}
663
664/**
665 * ubifs_get_lp_stats - get lprops statistics. 634 * ubifs_get_lp_stats - get lprops statistics.
666 * @c: UBIFS file-system description object 635 * @c: UBIFS file-system description object
667 * @st: return statistics 636 * @st: return statistics
@@ -1262,7 +1231,6 @@ static int scan_check_cb(struct ubifs_info *c,
1262 } 1231 }
1263 1232
1264 ubifs_scan_destroy(sleb); 1233 ubifs_scan_destroy(sleb);
1265
1266 return LPT_SCAN_CONTINUE; 1234 return LPT_SCAN_CONTINUE;
1267 1235
1268out_print: 1236out_print:
diff --git a/fs/ubifs/lpt.c b/fs/ubifs/lpt.c
index 9ff2463177e..db8bd0e518b 100644
--- a/fs/ubifs/lpt.c
+++ b/fs/ubifs/lpt.c
@@ -109,7 +109,8 @@ static void do_calc_lpt_geom(struct ubifs_info *c)
109 c->lpt_sz = (long long)c->pnode_cnt * c->pnode_sz; 109 c->lpt_sz = (long long)c->pnode_cnt * c->pnode_sz;
110 c->lpt_sz += (long long)c->nnode_cnt * c->nnode_sz; 110 c->lpt_sz += (long long)c->nnode_cnt * c->nnode_sz;
111 c->lpt_sz += c->ltab_sz; 111 c->lpt_sz += c->ltab_sz;
112 c->lpt_sz += c->lsave_sz; 112 if (c->big_lpt)
113 c->lpt_sz += c->lsave_sz;
113 114
114 /* Add wastage */ 115 /* Add wastage */
115 sz = c->lpt_sz; 116 sz = c->lpt_sz;
@@ -287,25 +288,56 @@ uint32_t ubifs_unpack_bits(uint8_t **addr, int *pos, int nrbits)
287 const int k = 32 - nrbits; 288 const int k = 32 - nrbits;
288 uint8_t *p = *addr; 289 uint8_t *p = *addr;
289 int b = *pos; 290 int b = *pos;
290 uint32_t val; 291 uint32_t uninitialized_var(val);
292 const int bytes = (nrbits + b + 7) >> 3;
291 293
292 ubifs_assert(nrbits > 0); 294 ubifs_assert(nrbits > 0);
293 ubifs_assert(nrbits <= 32); 295 ubifs_assert(nrbits <= 32);
294 ubifs_assert(*pos >= 0); 296 ubifs_assert(*pos >= 0);
295 ubifs_assert(*pos < 8); 297 ubifs_assert(*pos < 8);
296 if (b) { 298 if (b) {
297 val = p[1] | ((uint32_t)p[2] << 8) | ((uint32_t)p[3] << 16) | 299 switch (bytes) {
298 ((uint32_t)p[4] << 24); 300 case 2:
301 val = p[1];
302 break;
303 case 3:
304 val = p[1] | ((uint32_t)p[2] << 8);
305 break;
306 case 4:
307 val = p[1] | ((uint32_t)p[2] << 8) |
308 ((uint32_t)p[3] << 16);
309 break;
310 case 5:
311 val = p[1] | ((uint32_t)p[2] << 8) |
312 ((uint32_t)p[3] << 16) |
313 ((uint32_t)p[4] << 24);
314 }
299 val <<= (8 - b); 315 val <<= (8 - b);
300 val |= *p >> b; 316 val |= *p >> b;
301 nrbits += b; 317 nrbits += b;
302 } else 318 } else {
303 val = p[0] | ((uint32_t)p[1] << 8) | ((uint32_t)p[2] << 16) | 319 switch (bytes) {
304 ((uint32_t)p[3] << 24); 320 case 1:
321 val = p[0];
322 break;
323 case 2:
324 val = p[0] | ((uint32_t)p[1] << 8);
325 break;
326 case 3:
327 val = p[0] | ((uint32_t)p[1] << 8) |
328 ((uint32_t)p[2] << 16);
329 break;
330 case 4:
331 val = p[0] | ((uint32_t)p[1] << 8) |
332 ((uint32_t)p[2] << 16) |
333 ((uint32_t)p[3] << 24);
334 break;
335 }
336 }
305 val <<= k; 337 val <<= k;
306 val >>= k; 338 val >>= k;
307 b = nrbits & 7; 339 b = nrbits & 7;
308 p += nrbits / 8; 340 p += nrbits >> 3;
309 *addr = p; 341 *addr = p;
310 *pos = b; 342 *pos = b;
311 ubifs_assert((val >> nrbits) == 0 || nrbits - b == 32); 343 ubifs_assert((val >> nrbits) == 0 || nrbits - b == 32);
diff --git a/fs/ubifs/lpt_commit.c b/fs/ubifs/lpt_commit.c
index 5f0b83e20af..eed5a0025d6 100644
--- a/fs/ubifs/lpt_commit.c
+++ b/fs/ubifs/lpt_commit.c
@@ -177,8 +177,6 @@ static int alloc_lpt_leb(struct ubifs_info *c, int *lnum)
177 return 0; 177 return 0;
178 } 178 }
179 } 179 }
180 dbg_err("last LEB %d", *lnum);
181 dump_stack();
182 return -ENOSPC; 180 return -ENOSPC;
183} 181}
184 182
@@ -193,6 +191,9 @@ static int layout_cnodes(struct ubifs_info *c)
193 int lnum, offs, len, alen, done_lsave, done_ltab, err; 191 int lnum, offs, len, alen, done_lsave, done_ltab, err;
194 struct ubifs_cnode *cnode; 192 struct ubifs_cnode *cnode;
195 193
194 err = dbg_chk_lpt_sz(c, 0, 0);
195 if (err)
196 return err;
196 cnode = c->lpt_cnext; 197 cnode = c->lpt_cnext;
197 if (!cnode) 198 if (!cnode)
198 return 0; 199 return 0;
@@ -206,6 +207,7 @@ static int layout_cnodes(struct ubifs_info *c)
206 c->lsave_lnum = lnum; 207 c->lsave_lnum = lnum;
207 c->lsave_offs = offs; 208 c->lsave_offs = offs;
208 offs += c->lsave_sz; 209 offs += c->lsave_sz;
210 dbg_chk_lpt_sz(c, 1, c->lsave_sz);
209 } 211 }
210 212
211 if (offs + c->ltab_sz <= c->leb_size) { 213 if (offs + c->ltab_sz <= c->leb_size) {
@@ -213,6 +215,7 @@ static int layout_cnodes(struct ubifs_info *c)
213 c->ltab_lnum = lnum; 215 c->ltab_lnum = lnum;
214 c->ltab_offs = offs; 216 c->ltab_offs = offs;
215 offs += c->ltab_sz; 217 offs += c->ltab_sz;
218 dbg_chk_lpt_sz(c, 1, c->ltab_sz);
216 } 219 }
217 220
218 do { 221 do {
@@ -226,9 +229,10 @@ static int layout_cnodes(struct ubifs_info *c)
226 while (offs + len > c->leb_size) { 229 while (offs + len > c->leb_size) {
227 alen = ALIGN(offs, c->min_io_size); 230 alen = ALIGN(offs, c->min_io_size);
228 upd_ltab(c, lnum, c->leb_size - alen, alen - offs); 231 upd_ltab(c, lnum, c->leb_size - alen, alen - offs);
232 dbg_chk_lpt_sz(c, 2, alen - offs);
229 err = alloc_lpt_leb(c, &lnum); 233 err = alloc_lpt_leb(c, &lnum);
230 if (err) 234 if (err)
231 return err; 235 goto no_space;
232 offs = 0; 236 offs = 0;
233 ubifs_assert(lnum >= c->lpt_first && 237 ubifs_assert(lnum >= c->lpt_first &&
234 lnum <= c->lpt_last); 238 lnum <= c->lpt_last);
@@ -238,6 +242,7 @@ static int layout_cnodes(struct ubifs_info *c)
238 c->lsave_lnum = lnum; 242 c->lsave_lnum = lnum;
239 c->lsave_offs = offs; 243 c->lsave_offs = offs;
240 offs += c->lsave_sz; 244 offs += c->lsave_sz;
245 dbg_chk_lpt_sz(c, 1, c->lsave_sz);
241 continue; 246 continue;
242 } 247 }
243 if (!done_ltab) { 248 if (!done_ltab) {
@@ -245,6 +250,7 @@ static int layout_cnodes(struct ubifs_info *c)
245 c->ltab_lnum = lnum; 250 c->ltab_lnum = lnum;
246 c->ltab_offs = offs; 251 c->ltab_offs = offs;
247 offs += c->ltab_sz; 252 offs += c->ltab_sz;
253 dbg_chk_lpt_sz(c, 1, c->ltab_sz);
248 continue; 254 continue;
249 } 255 }
250 break; 256 break;
@@ -257,6 +263,7 @@ static int layout_cnodes(struct ubifs_info *c)
257 c->lpt_offs = offs; 263 c->lpt_offs = offs;
258 } 264 }
259 offs += len; 265 offs += len;
266 dbg_chk_lpt_sz(c, 1, len);
260 cnode = cnode->cnext; 267 cnode = cnode->cnext;
261 } while (cnode && cnode != c->lpt_cnext); 268 } while (cnode && cnode != c->lpt_cnext);
262 269
@@ -265,9 +272,10 @@ static int layout_cnodes(struct ubifs_info *c)
265 if (offs + c->lsave_sz > c->leb_size) { 272 if (offs + c->lsave_sz > c->leb_size) {
266 alen = ALIGN(offs, c->min_io_size); 273 alen = ALIGN(offs, c->min_io_size);
267 upd_ltab(c, lnum, c->leb_size - alen, alen - offs); 274 upd_ltab(c, lnum, c->leb_size - alen, alen - offs);
275 dbg_chk_lpt_sz(c, 2, alen - offs);
268 err = alloc_lpt_leb(c, &lnum); 276 err = alloc_lpt_leb(c, &lnum);
269 if (err) 277 if (err)
270 return err; 278 goto no_space;
271 offs = 0; 279 offs = 0;
272 ubifs_assert(lnum >= c->lpt_first && 280 ubifs_assert(lnum >= c->lpt_first &&
273 lnum <= c->lpt_last); 281 lnum <= c->lpt_last);
@@ -276,6 +284,7 @@ static int layout_cnodes(struct ubifs_info *c)
276 c->lsave_lnum = lnum; 284 c->lsave_lnum = lnum;
277 c->lsave_offs = offs; 285 c->lsave_offs = offs;
278 offs += c->lsave_sz; 286 offs += c->lsave_sz;
287 dbg_chk_lpt_sz(c, 1, c->lsave_sz);
279 } 288 }
280 289
281 /* Make sure to place LPT's own lprops table */ 290 /* Make sure to place LPT's own lprops table */
@@ -283,9 +292,10 @@ static int layout_cnodes(struct ubifs_info *c)
283 if (offs + c->ltab_sz > c->leb_size) { 292 if (offs + c->ltab_sz > c->leb_size) {
284 alen = ALIGN(offs, c->min_io_size); 293 alen = ALIGN(offs, c->min_io_size);
285 upd_ltab(c, lnum, c->leb_size - alen, alen - offs); 294 upd_ltab(c, lnum, c->leb_size - alen, alen - offs);
295 dbg_chk_lpt_sz(c, 2, alen - offs);
286 err = alloc_lpt_leb(c, &lnum); 296 err = alloc_lpt_leb(c, &lnum);
287 if (err) 297 if (err)
288 return err; 298 goto no_space;
289 offs = 0; 299 offs = 0;
290 ubifs_assert(lnum >= c->lpt_first && 300 ubifs_assert(lnum >= c->lpt_first &&
291 lnum <= c->lpt_last); 301 lnum <= c->lpt_last);
@@ -294,11 +304,23 @@ static int layout_cnodes(struct ubifs_info *c)
294 c->ltab_lnum = lnum; 304 c->ltab_lnum = lnum;
295 c->ltab_offs = offs; 305 c->ltab_offs = offs;
296 offs += c->ltab_sz; 306 offs += c->ltab_sz;
307 dbg_chk_lpt_sz(c, 1, c->ltab_sz);
297 } 308 }
298 309
299 alen = ALIGN(offs, c->min_io_size); 310 alen = ALIGN(offs, c->min_io_size);
300 upd_ltab(c, lnum, c->leb_size - alen, alen - offs); 311 upd_ltab(c, lnum, c->leb_size - alen, alen - offs);
312 dbg_chk_lpt_sz(c, 4, alen - offs);
313 err = dbg_chk_lpt_sz(c, 3, alen);
314 if (err)
315 return err;
301 return 0; 316 return 0;
317
318no_space:
319 ubifs_err("LPT out of space");
320 dbg_err("LPT out of space at LEB %d:%d needing %d, done_ltab %d, "
321 "done_lsave %d", lnum, offs, len, done_ltab, done_lsave);
322 dbg_dump_lpt_info(c);
323 return err;
302} 324}
303 325
304/** 326/**
@@ -333,8 +355,6 @@ static int realloc_lpt_leb(struct ubifs_info *c, int *lnum)
333 *lnum = i + c->lpt_first; 355 *lnum = i + c->lpt_first;
334 return 0; 356 return 0;
335 } 357 }
336 dbg_err("last LEB %d", *lnum);
337 dump_stack();
338 return -ENOSPC; 358 return -ENOSPC;
339} 359}
340 360
@@ -369,12 +389,14 @@ static int write_cnodes(struct ubifs_info *c)
369 done_lsave = 1; 389 done_lsave = 1;
370 ubifs_pack_lsave(c, buf + offs, c->lsave); 390 ubifs_pack_lsave(c, buf + offs, c->lsave);
371 offs += c->lsave_sz; 391 offs += c->lsave_sz;
392 dbg_chk_lpt_sz(c, 1, c->lsave_sz);
372 } 393 }
373 394
374 if (offs + c->ltab_sz <= c->leb_size) { 395 if (offs + c->ltab_sz <= c->leb_size) {
375 done_ltab = 1; 396 done_ltab = 1;
376 ubifs_pack_ltab(c, buf + offs, c->ltab_cmt); 397 ubifs_pack_ltab(c, buf + offs, c->ltab_cmt);
377 offs += c->ltab_sz; 398 offs += c->ltab_sz;
399 dbg_chk_lpt_sz(c, 1, c->ltab_sz);
378 } 400 }
379 401
380 /* Loop for each cnode */ 402 /* Loop for each cnode */
@@ -392,10 +414,12 @@ static int write_cnodes(struct ubifs_info *c)
392 alen, UBI_SHORTTERM); 414 alen, UBI_SHORTTERM);
393 if (err) 415 if (err)
394 return err; 416 return err;
417 dbg_chk_lpt_sz(c, 4, alen - wlen);
395 } 418 }
419 dbg_chk_lpt_sz(c, 2, 0);
396 err = realloc_lpt_leb(c, &lnum); 420 err = realloc_lpt_leb(c, &lnum);
397 if (err) 421 if (err)
398 return err; 422 goto no_space;
399 offs = 0; 423 offs = 0;
400 from = 0; 424 from = 0;
401 ubifs_assert(lnum >= c->lpt_first && 425 ubifs_assert(lnum >= c->lpt_first &&
@@ -408,12 +432,14 @@ static int write_cnodes(struct ubifs_info *c)
408 done_lsave = 1; 432 done_lsave = 1;
409 ubifs_pack_lsave(c, buf + offs, c->lsave); 433 ubifs_pack_lsave(c, buf + offs, c->lsave);
410 offs += c->lsave_sz; 434 offs += c->lsave_sz;
435 dbg_chk_lpt_sz(c, 1, c->lsave_sz);
411 continue; 436 continue;
412 } 437 }
413 if (!done_ltab) { 438 if (!done_ltab) {
414 done_ltab = 1; 439 done_ltab = 1;
415 ubifs_pack_ltab(c, buf + offs, c->ltab_cmt); 440 ubifs_pack_ltab(c, buf + offs, c->ltab_cmt);
416 offs += c->ltab_sz; 441 offs += c->ltab_sz;
442 dbg_chk_lpt_sz(c, 1, c->ltab_sz);
417 continue; 443 continue;
418 } 444 }
419 break; 445 break;
@@ -435,6 +461,7 @@ static int write_cnodes(struct ubifs_info *c)
435 clear_bit(COW_ZNODE, &cnode->flags); 461 clear_bit(COW_ZNODE, &cnode->flags);
436 smp_mb__after_clear_bit(); 462 smp_mb__after_clear_bit();
437 offs += len; 463 offs += len;
464 dbg_chk_lpt_sz(c, 1, len);
438 cnode = cnode->cnext; 465 cnode = cnode->cnext;
439 } while (cnode && cnode != c->lpt_cnext); 466 } while (cnode && cnode != c->lpt_cnext);
440 467
@@ -448,9 +475,10 @@ static int write_cnodes(struct ubifs_info *c)
448 UBI_SHORTTERM); 475 UBI_SHORTTERM);
449 if (err) 476 if (err)
450 return err; 477 return err;
478 dbg_chk_lpt_sz(c, 2, alen - wlen);
451 err = realloc_lpt_leb(c, &lnum); 479 err = realloc_lpt_leb(c, &lnum);
452 if (err) 480 if (err)
453 return err; 481 goto no_space;
454 offs = 0; 482 offs = 0;
455 ubifs_assert(lnum >= c->lpt_first && 483 ubifs_assert(lnum >= c->lpt_first &&
456 lnum <= c->lpt_last); 484 lnum <= c->lpt_last);
@@ -461,6 +489,7 @@ static int write_cnodes(struct ubifs_info *c)
461 done_lsave = 1; 489 done_lsave = 1;
462 ubifs_pack_lsave(c, buf + offs, c->lsave); 490 ubifs_pack_lsave(c, buf + offs, c->lsave);
463 offs += c->lsave_sz; 491 offs += c->lsave_sz;
492 dbg_chk_lpt_sz(c, 1, c->lsave_sz);
464 } 493 }
465 494
466 /* Make sure to place LPT's own lprops table */ 495 /* Make sure to place LPT's own lprops table */
@@ -473,9 +502,10 @@ static int write_cnodes(struct ubifs_info *c)
473 UBI_SHORTTERM); 502 UBI_SHORTTERM);
474 if (err) 503 if (err)
475 return err; 504 return err;
505 dbg_chk_lpt_sz(c, 2, alen - wlen);
476 err = realloc_lpt_leb(c, &lnum); 506 err = realloc_lpt_leb(c, &lnum);
477 if (err) 507 if (err)
478 return err; 508 goto no_space;
479 offs = 0; 509 offs = 0;
480 ubifs_assert(lnum >= c->lpt_first && 510 ubifs_assert(lnum >= c->lpt_first &&
481 lnum <= c->lpt_last); 511 lnum <= c->lpt_last);
@@ -486,6 +516,7 @@ static int write_cnodes(struct ubifs_info *c)
486 done_ltab = 1; 516 done_ltab = 1;
487 ubifs_pack_ltab(c, buf + offs, c->ltab_cmt); 517 ubifs_pack_ltab(c, buf + offs, c->ltab_cmt);
488 offs += c->ltab_sz; 518 offs += c->ltab_sz;
519 dbg_chk_lpt_sz(c, 1, c->ltab_sz);
489 } 520 }
490 521
491 /* Write remaining data in buffer */ 522 /* Write remaining data in buffer */
@@ -495,6 +526,12 @@ static int write_cnodes(struct ubifs_info *c)
495 err = ubifs_leb_write(c, lnum, buf + from, from, alen, UBI_SHORTTERM); 526 err = ubifs_leb_write(c, lnum, buf + from, from, alen, UBI_SHORTTERM);
496 if (err) 527 if (err)
497 return err; 528 return err;
529
530 dbg_chk_lpt_sz(c, 4, alen - wlen);
531 err = dbg_chk_lpt_sz(c, 3, ALIGN(offs, c->min_io_size));
532 if (err)
533 return err;
534
498 c->nhead_lnum = lnum; 535 c->nhead_lnum = lnum;
499 c->nhead_offs = ALIGN(offs, c->min_io_size); 536 c->nhead_offs = ALIGN(offs, c->min_io_size);
500 537
@@ -503,7 +540,15 @@ static int write_cnodes(struct ubifs_info *c)
503 dbg_lp("LPT ltab is at %d:%d", c->ltab_lnum, c->ltab_offs); 540 dbg_lp("LPT ltab is at %d:%d", c->ltab_lnum, c->ltab_offs);
504 if (c->big_lpt) 541 if (c->big_lpt)
505 dbg_lp("LPT lsave is at %d:%d", c->lsave_lnum, c->lsave_offs); 542 dbg_lp("LPT lsave is at %d:%d", c->lsave_lnum, c->lsave_offs);
543
506 return 0; 544 return 0;
545
546no_space:
547 ubifs_err("LPT out of space mismatch");
548 dbg_err("LPT out of space mismatch at LEB %d:%d needing %d, done_ltab "
549 "%d, done_lsave %d", lnum, offs, len, done_ltab, done_lsave);
550 dbg_dump_lpt_info(c);
551 return err;
507} 552}
508 553
509/** 554/**
@@ -1044,6 +1089,8 @@ static int is_a_node(struct ubifs_info *c, uint8_t *buf, int len)
1044 int pos = 0, node_type, node_len; 1089 int pos = 0, node_type, node_len;
1045 uint16_t crc, calc_crc; 1090 uint16_t crc, calc_crc;
1046 1091
1092 if (len < UBIFS_LPT_CRC_BYTES + (UBIFS_LPT_TYPE_BITS + 7) / 8)
1093 return 0;
1047 node_type = ubifs_unpack_bits(&addr, &pos, UBIFS_LPT_TYPE_BITS); 1094 node_type = ubifs_unpack_bits(&addr, &pos, UBIFS_LPT_TYPE_BITS);
1048 if (node_type == UBIFS_LPT_NOT_A_NODE) 1095 if (node_type == UBIFS_LPT_NOT_A_NODE)
1049 return 0; 1096 return 0;
@@ -1156,6 +1203,9 @@ int ubifs_lpt_start_commit(struct ubifs_info *c)
1156 dbg_lp(""); 1203 dbg_lp("");
1157 1204
1158 mutex_lock(&c->lp_mutex); 1205 mutex_lock(&c->lp_mutex);
1206 err = dbg_chk_lpt_free_spc(c);
1207 if (err)
1208 goto out;
1159 err = dbg_check_ltab(c); 1209 err = dbg_check_ltab(c);
1160 if (err) 1210 if (err)
1161 goto out; 1211 goto out;
@@ -1645,4 +1695,121 @@ int dbg_check_ltab(struct ubifs_info *c)
1645 return 0; 1695 return 0;
1646} 1696}
1647 1697
1698/**
1699 * dbg_chk_lpt_free_spc - check LPT free space is enough to write entire LPT.
1700 * @c: the UBIFS file-system description object
1701 *
1702 * This function returns %0 on success and a negative error code on failure.
1703 */
1704int dbg_chk_lpt_free_spc(struct ubifs_info *c)
1705{
1706 long long free = 0;
1707 int i;
1708
1709 for (i = 0; i < c->lpt_lebs; i++) {
1710 if (c->ltab[i].tgc || c->ltab[i].cmt)
1711 continue;
1712 if (i + c->lpt_first == c->nhead_lnum)
1713 free += c->leb_size - c->nhead_offs;
1714 else if (c->ltab[i].free == c->leb_size)
1715 free += c->leb_size;
1716 }
1717 if (free < c->lpt_sz) {
1718 dbg_err("LPT space error: free %lld lpt_sz %lld",
1719 free, c->lpt_sz);
1720 dbg_dump_lpt_info(c);
1721 return -EINVAL;
1722 }
1723 return 0;
1724}
1725
1726/**
1727 * dbg_chk_lpt_sz - check LPT does not write more than LPT size.
1728 * @c: the UBIFS file-system description object
1729 * @action: action
1730 * @len: length written
1731 *
1732 * This function returns %0 on success and a negative error code on failure.
1733 */
1734int dbg_chk_lpt_sz(struct ubifs_info *c, int action, int len)
1735{
1736 long long chk_lpt_sz, lpt_sz;
1737 int err = 0;
1738
1739 switch (action) {
1740 case 0:
1741 c->chk_lpt_sz = 0;
1742 c->chk_lpt_sz2 = 0;
1743 c->chk_lpt_lebs = 0;
1744 c->chk_lpt_wastage = 0;
1745 if (c->dirty_pn_cnt > c->pnode_cnt) {
1746 dbg_err("dirty pnodes %d exceed max %d",
1747 c->dirty_pn_cnt, c->pnode_cnt);
1748 err = -EINVAL;
1749 }
1750 if (c->dirty_nn_cnt > c->nnode_cnt) {
1751 dbg_err("dirty nnodes %d exceed max %d",
1752 c->dirty_nn_cnt, c->nnode_cnt);
1753 err = -EINVAL;
1754 }
1755 return err;
1756 case 1:
1757 c->chk_lpt_sz += len;
1758 return 0;
1759 case 2:
1760 c->chk_lpt_sz += len;
1761 c->chk_lpt_wastage += len;
1762 c->chk_lpt_lebs += 1;
1763 return 0;
1764 case 3:
1765 chk_lpt_sz = c->leb_size;
1766 chk_lpt_sz *= c->chk_lpt_lebs;
1767 chk_lpt_sz += len - c->nhead_offs;
1768 if (c->chk_lpt_sz != chk_lpt_sz) {
1769 dbg_err("LPT wrote %lld but space used was %lld",
1770 c->chk_lpt_sz, chk_lpt_sz);
1771 err = -EINVAL;
1772 }
1773 if (c->chk_lpt_sz > c->lpt_sz) {
1774 dbg_err("LPT wrote %lld but lpt_sz is %lld",
1775 c->chk_lpt_sz, c->lpt_sz);
1776 err = -EINVAL;
1777 }
1778 if (c->chk_lpt_sz2 && c->chk_lpt_sz != c->chk_lpt_sz2) {
1779 dbg_err("LPT layout size %lld but wrote %lld",
1780 c->chk_lpt_sz, c->chk_lpt_sz2);
1781 err = -EINVAL;
1782 }
1783 if (c->chk_lpt_sz2 && c->new_nhead_offs != len) {
1784 dbg_err("LPT new nhead offs: expected %d was %d",
1785 c->new_nhead_offs, len);
1786 err = -EINVAL;
1787 }
1788 lpt_sz = (long long)c->pnode_cnt * c->pnode_sz;
1789 lpt_sz += (long long)c->nnode_cnt * c->nnode_sz;
1790 lpt_sz += c->ltab_sz;
1791 if (c->big_lpt)
1792 lpt_sz += c->lsave_sz;
1793 if (c->chk_lpt_sz - c->chk_lpt_wastage > lpt_sz) {
1794 dbg_err("LPT chk_lpt_sz %lld + waste %lld exceeds %lld",
1795 c->chk_lpt_sz, c->chk_lpt_wastage, lpt_sz);
1796 err = -EINVAL;
1797 }
1798 if (err)
1799 dbg_dump_lpt_info(c);
1800 c->chk_lpt_sz2 = c->chk_lpt_sz;
1801 c->chk_lpt_sz = 0;
1802 c->chk_lpt_wastage = 0;
1803 c->chk_lpt_lebs = 0;
1804 c->new_nhead_offs = len;
1805 return err;
1806 case 4:
1807 c->chk_lpt_sz += len;
1808 c->chk_lpt_wastage += len;
1809 return 0;
1810 default:
1811 return -EINVAL;
1812 }
1813}
1814
1648#endif /* CONFIG_UBIFS_FS_DEBUG */ 1815#endif /* CONFIG_UBIFS_FS_DEBUG */
diff --git a/fs/ubifs/misc.h b/fs/ubifs/misc.h
index 4c12a9215d7..4fa81d867e4 100644
--- a/fs/ubifs/misc.h
+++ b/fs/ubifs/misc.h
@@ -310,4 +310,31 @@ static inline int ubifs_tnc_lookup(struct ubifs_info *c,
310 return ubifs_tnc_locate(c, key, node, NULL, NULL); 310 return ubifs_tnc_locate(c, key, node, NULL, NULL);
311} 311}
312 312
313/**
314 * ubifs_get_lprops - get reference to LEB properties.
315 * @c: the UBIFS file-system description object
316 *
317 * This function locks lprops. Lprops have to be unlocked by
318 * 'ubifs_release_lprops()'.
319 */
320static inline void ubifs_get_lprops(struct ubifs_info *c)
321{
322 mutex_lock(&c->lp_mutex);
323}
324
325/**
326 * ubifs_release_lprops - release lprops lock.
327 * @c: the UBIFS file-system description object
328 *
329 * This function has to be called after each 'ubifs_get_lprops()' call to
330 * unlock lprops.
331 */
332static inline void ubifs_release_lprops(struct ubifs_info *c)
333{
334 ubifs_assert(mutex_is_locked(&c->lp_mutex));
335 ubifs_assert(c->lst.empty_lebs >= 0 &&
336 c->lst.empty_lebs <= c->main_lebs);
337 mutex_unlock(&c->lp_mutex);
338}
339
313#endif /* __UBIFS_MISC_H__ */ 340#endif /* __UBIFS_MISC_H__ */
diff --git a/fs/ubifs/scan.c b/fs/ubifs/scan.c
index acf5c5fffc6..0ed82479b44 100644
--- a/fs/ubifs/scan.c
+++ b/fs/ubifs/scan.c
@@ -87,7 +87,7 @@ int ubifs_scan_a_node(const struct ubifs_info *c, void *buf, int len, int lnum,
87 87
88 dbg_scan("scanning %s", dbg_ntype(ch->node_type)); 88 dbg_scan("scanning %s", dbg_ntype(ch->node_type));
89 89
90 if (ubifs_check_node(c, buf, lnum, offs, quiet)) 90 if (ubifs_check_node(c, buf, lnum, offs, quiet, 1))
91 return SCANNED_A_CORRUPT_NODE; 91 return SCANNED_A_CORRUPT_NODE;
92 92
93 if (ch->node_type == UBIFS_PAD_NODE) { 93 if (ch->node_type == UBIFS_PAD_NODE) {
diff --git a/fs/ubifs/super.c b/fs/ubifs/super.c
index 9a9220333b3..8780efbf40a 100644
--- a/fs/ubifs/super.c
+++ b/fs/ubifs/super.c
@@ -401,6 +401,16 @@ static int ubifs_show_options(struct seq_file *s, struct vfsmount *mnt)
401 else if (c->mount_opts.unmount_mode == 1) 401 else if (c->mount_opts.unmount_mode == 1)
402 seq_printf(s, ",norm_unmount"); 402 seq_printf(s, ",norm_unmount");
403 403
404 if (c->mount_opts.bulk_read == 2)
405 seq_printf(s, ",bulk_read");
406 else if (c->mount_opts.bulk_read == 1)
407 seq_printf(s, ",no_bulk_read");
408
409 if (c->mount_opts.chk_data_crc == 2)
410 seq_printf(s, ",chk_data_crc");
411 else if (c->mount_opts.chk_data_crc == 1)
412 seq_printf(s, ",no_chk_data_crc");
413
404 return 0; 414 return 0;
405} 415}
406 416
@@ -408,13 +418,26 @@ static int ubifs_sync_fs(struct super_block *sb, int wait)
408{ 418{
409 struct ubifs_info *c = sb->s_fs_info; 419 struct ubifs_info *c = sb->s_fs_info;
410 int i, ret = 0, err; 420 int i, ret = 0, err;
421 long long bud_bytes;
411 422
412 if (c->jheads) 423 if (c->jheads) {
413 for (i = 0; i < c->jhead_cnt; i++) { 424 for (i = 0; i < c->jhead_cnt; i++) {
414 err = ubifs_wbuf_sync(&c->jheads[i].wbuf); 425 err = ubifs_wbuf_sync(&c->jheads[i].wbuf);
415 if (err && !ret) 426 if (err && !ret)
416 ret = err; 427 ret = err;
417 } 428 }
429
430 /* Commit the journal unless it has too little data */
431 spin_lock(&c->buds_lock);
432 bud_bytes = c->bud_bytes;
433 spin_unlock(&c->buds_lock);
434 if (bud_bytes > c->leb_size) {
435 err = ubifs_run_commit(c);
436 if (err)
437 return err;
438 }
439 }
440
418 /* 441 /*
419 * We ought to call sync for c->ubi but it does not have one. If it had 442 * We ought to call sync for c->ubi but it does not have one. If it had
420 * it would in turn call mtd->sync, however mtd operations are 443 * it would in turn call mtd->sync, however mtd operations are
@@ -538,6 +561,18 @@ static int init_constants_early(struct ubifs_info *c)
538 * calculations when reporting free space. 561 * calculations when reporting free space.
539 */ 562 */
540 c->leb_overhead = c->leb_size % UBIFS_MAX_DATA_NODE_SZ; 563 c->leb_overhead = c->leb_size % UBIFS_MAX_DATA_NODE_SZ;
564 /* Buffer size for bulk-reads */
565 c->bulk_read_buf_size = UBIFS_MAX_BULK_READ * UBIFS_MAX_DATA_NODE_SZ;
566 if (c->bulk_read_buf_size > c->leb_size)
567 c->bulk_read_buf_size = c->leb_size;
568 if (c->bulk_read_buf_size > 128 * 1024) {
569 /* Check if we can kmalloc more than 128KiB */
570 void *try = kmalloc(c->bulk_read_buf_size, GFP_KERNEL);
571
572 kfree(try);
573 if (!try)
574 c->bulk_read_buf_size = 128 * 1024;
575 }
541 return 0; 576 return 0;
542} 577}
543 578
@@ -840,17 +875,29 @@ static int check_volume_empty(struct ubifs_info *c)
840 * 875 *
841 * Opt_fast_unmount: do not run a journal commit before un-mounting 876 * Opt_fast_unmount: do not run a journal commit before un-mounting
842 * Opt_norm_unmount: run a journal commit before un-mounting 877 * Opt_norm_unmount: run a journal commit before un-mounting
878 * Opt_bulk_read: enable bulk-reads
879 * Opt_no_bulk_read: disable bulk-reads
880 * Opt_chk_data_crc: check CRCs when reading data nodes
881 * Opt_no_chk_data_crc: do not check CRCs when reading data nodes
843 * Opt_err: just end of array marker 882 * Opt_err: just end of array marker
844 */ 883 */
845enum { 884enum {
846 Opt_fast_unmount, 885 Opt_fast_unmount,
847 Opt_norm_unmount, 886 Opt_norm_unmount,
887 Opt_bulk_read,
888 Opt_no_bulk_read,
889 Opt_chk_data_crc,
890 Opt_no_chk_data_crc,
848 Opt_err, 891 Opt_err,
849}; 892};
850 893
851static const match_table_t tokens = { 894static const match_table_t tokens = {
852 {Opt_fast_unmount, "fast_unmount"}, 895 {Opt_fast_unmount, "fast_unmount"},
853 {Opt_norm_unmount, "norm_unmount"}, 896 {Opt_norm_unmount, "norm_unmount"},
897 {Opt_bulk_read, "bulk_read"},
898 {Opt_no_bulk_read, "no_bulk_read"},
899 {Opt_chk_data_crc, "chk_data_crc"},
900 {Opt_no_chk_data_crc, "no_chk_data_crc"},
854 {Opt_err, NULL}, 901 {Opt_err, NULL},
855}; 902};
856 903
@@ -888,6 +935,22 @@ static int ubifs_parse_options(struct ubifs_info *c, char *options,
888 c->mount_opts.unmount_mode = 1; 935 c->mount_opts.unmount_mode = 1;
889 c->fast_unmount = 0; 936 c->fast_unmount = 0;
890 break; 937 break;
938 case Opt_bulk_read:
939 c->mount_opts.bulk_read = 2;
940 c->bulk_read = 1;
941 break;
942 case Opt_no_bulk_read:
943 c->mount_opts.bulk_read = 1;
944 c->bulk_read = 0;
945 break;
946 case Opt_chk_data_crc:
947 c->mount_opts.chk_data_crc = 2;
948 c->no_chk_data_crc = 0;
949 break;
950 case Opt_no_chk_data_crc:
951 c->mount_opts.chk_data_crc = 1;
952 c->no_chk_data_crc = 1;
953 break;
891 default: 954 default:
892 ubifs_err("unrecognized mount option \"%s\" " 955 ubifs_err("unrecognized mount option \"%s\" "
893 "or missing value", p); 956 "or missing value", p);
@@ -996,6 +1059,8 @@ static int mount_ubifs(struct ubifs_info *c)
996 goto out_free; 1059 goto out_free;
997 } 1060 }
998 1061
1062 c->always_chk_crc = 1;
1063
999 err = ubifs_read_superblock(c); 1064 err = ubifs_read_superblock(c);
1000 if (err) 1065 if (err)
1001 goto out_free; 1066 goto out_free;
@@ -1032,8 +1097,6 @@ static int mount_ubifs(struct ubifs_info *c)
1032 1097
1033 /* Create background thread */ 1098 /* Create background thread */
1034 c->bgt = kthread_create(ubifs_bg_thread, c, c->bgt_name); 1099 c->bgt = kthread_create(ubifs_bg_thread, c, c->bgt_name);
1035 if (!c->bgt)
1036 c->bgt = ERR_PTR(-EINVAL);
1037 if (IS_ERR(c->bgt)) { 1100 if (IS_ERR(c->bgt)) {
1038 err = PTR_ERR(c->bgt); 1101 err = PTR_ERR(c->bgt);
1039 c->bgt = NULL; 1102 c->bgt = NULL;
@@ -1139,24 +1202,28 @@ static int mount_ubifs(struct ubifs_info *c)
1139 if (err) 1202 if (err)
1140 goto out_infos; 1203 goto out_infos;
1141 1204
1205 c->always_chk_crc = 0;
1206
1142 ubifs_msg("mounted UBI device %d, volume %d, name \"%s\"", 1207 ubifs_msg("mounted UBI device %d, volume %d, name \"%s\"",
1143 c->vi.ubi_num, c->vi.vol_id, c->vi.name); 1208 c->vi.ubi_num, c->vi.vol_id, c->vi.name);
1144 if (mounted_read_only) 1209 if (mounted_read_only)
1145 ubifs_msg("mounted read-only"); 1210 ubifs_msg("mounted read-only");
1146 x = (long long)c->main_lebs * c->leb_size; 1211 x = (long long)c->main_lebs * c->leb_size;
1147 ubifs_msg("file system size: %lld bytes (%lld KiB, %lld MiB, %d LEBs)", 1212 ubifs_msg("file system size: %lld bytes (%lld KiB, %lld MiB, %d "
1148 x, x >> 10, x >> 20, c->main_lebs); 1213 "LEBs)", x, x >> 10, x >> 20, c->main_lebs);
1149 x = (long long)c->log_lebs * c->leb_size + c->max_bud_bytes; 1214 x = (long long)c->log_lebs * c->leb_size + c->max_bud_bytes;
1150 ubifs_msg("journal size: %lld bytes (%lld KiB, %lld MiB, %d LEBs)", 1215 ubifs_msg("journal size: %lld bytes (%lld KiB, %lld MiB, %d "
1151 x, x >> 10, x >> 20, c->log_lebs + c->max_bud_cnt); 1216 "LEBs)", x, x >> 10, x >> 20, c->log_lebs + c->max_bud_cnt);
1152 ubifs_msg("default compressor: %s", ubifs_compr_name(c->default_compr)); 1217 ubifs_msg("media format: %d (latest is %d)",
1153 ubifs_msg("media format %d, latest format %d",
1154 c->fmt_version, UBIFS_FORMAT_VERSION); 1218 c->fmt_version, UBIFS_FORMAT_VERSION);
1219 ubifs_msg("default compressor: %s", ubifs_compr_name(c->default_compr));
1220 ubifs_msg("reserved for root: %llu bytes (%llu KiB)",
1221 c->report_rp_size, c->report_rp_size >> 10);
1155 1222
1156 dbg_msg("compiled on: " __DATE__ " at " __TIME__); 1223 dbg_msg("compiled on: " __DATE__ " at " __TIME__);
1157 dbg_msg("min. I/O unit size: %d bytes", c->min_io_size); 1224 dbg_msg("min. I/O unit size: %d bytes", c->min_io_size);
1158 dbg_msg("LEB size: %d bytes (%d KiB)", 1225 dbg_msg("LEB size: %d bytes (%d KiB)",
1159 c->leb_size, c->leb_size / 1024); 1226 c->leb_size, c->leb_size >> 10);
1160 dbg_msg("data journal heads: %d", 1227 dbg_msg("data journal heads: %d",
1161 c->jhead_cnt - NONDATA_JHEADS_CNT); 1228 c->jhead_cnt - NONDATA_JHEADS_CNT);
1162 dbg_msg("UUID: %02X%02X%02X%02X-%02X%02X" 1229 dbg_msg("UUID: %02X%02X%02X%02X-%02X%02X"
@@ -1282,6 +1349,7 @@ static int ubifs_remount_rw(struct ubifs_info *c)
1282 1349
1283 mutex_lock(&c->umount_mutex); 1350 mutex_lock(&c->umount_mutex);
1284 c->remounting_rw = 1; 1351 c->remounting_rw = 1;
1352 c->always_chk_crc = 1;
1285 1353
1286 /* Check for enough free space */ 1354 /* Check for enough free space */
1287 if (ubifs_calc_available(c, c->min_idx_lebs) <= 0) { 1355 if (ubifs_calc_available(c, c->min_idx_lebs) <= 0) {
@@ -1345,20 +1413,20 @@ static int ubifs_remount_rw(struct ubifs_info *c)
1345 1413
1346 /* Create background thread */ 1414 /* Create background thread */
1347 c->bgt = kthread_create(ubifs_bg_thread, c, c->bgt_name); 1415 c->bgt = kthread_create(ubifs_bg_thread, c, c->bgt_name);
1348 if (!c->bgt)
1349 c->bgt = ERR_PTR(-EINVAL);
1350 if (IS_ERR(c->bgt)) { 1416 if (IS_ERR(c->bgt)) {
1351 err = PTR_ERR(c->bgt); 1417 err = PTR_ERR(c->bgt);
1352 c->bgt = NULL; 1418 c->bgt = NULL;
1353 ubifs_err("cannot spawn \"%s\", error %d", 1419 ubifs_err("cannot spawn \"%s\", error %d",
1354 c->bgt_name, err); 1420 c->bgt_name, err);
1355 return err; 1421 goto out;
1356 } 1422 }
1357 wake_up_process(c->bgt); 1423 wake_up_process(c->bgt);
1358 1424
1359 c->orph_buf = vmalloc(c->leb_size); 1425 c->orph_buf = vmalloc(c->leb_size);
1360 if (!c->orph_buf) 1426 if (!c->orph_buf) {
1361 return -ENOMEM; 1427 err = -ENOMEM;
1428 goto out;
1429 }
1362 1430
1363 /* Check for enough log space */ 1431 /* Check for enough log space */
1364 lnum = c->lhead_lnum + 1; 1432 lnum = c->lhead_lnum + 1;
@@ -1385,6 +1453,7 @@ static int ubifs_remount_rw(struct ubifs_info *c)
1385 dbg_gen("re-mounted read-write"); 1453 dbg_gen("re-mounted read-write");
1386 c->vfs_sb->s_flags &= ~MS_RDONLY; 1454 c->vfs_sb->s_flags &= ~MS_RDONLY;
1387 c->remounting_rw = 0; 1455 c->remounting_rw = 0;
1456 c->always_chk_crc = 0;
1388 mutex_unlock(&c->umount_mutex); 1457 mutex_unlock(&c->umount_mutex);
1389 return 0; 1458 return 0;
1390 1459
@@ -1400,6 +1469,7 @@ out:
1400 c->ileb_buf = NULL; 1469 c->ileb_buf = NULL;
1401 ubifs_lpt_free(c, 1); 1470 ubifs_lpt_free(c, 1);
1402 c->remounting_rw = 0; 1471 c->remounting_rw = 0;
1472 c->always_chk_crc = 0;
1403 mutex_unlock(&c->umount_mutex); 1473 mutex_unlock(&c->umount_mutex);
1404 return err; 1474 return err;
1405} 1475}
@@ -1408,12 +1478,9 @@ out:
1408 * commit_on_unmount - commit the journal when un-mounting. 1478 * commit_on_unmount - commit the journal when un-mounting.
1409 * @c: UBIFS file-system description object 1479 * @c: UBIFS file-system description object
1410 * 1480 *
1411 * This function is called during un-mounting and it commits the journal unless 1481 * This function is called during un-mounting and re-mounting, and it commits
1412 * the "fast unmount" mode is enabled. It also avoids committing the journal if 1482 * the journal unless the "fast unmount" mode is enabled. It also avoids
1413 * it contains too few data. 1483 * committing the journal if it contains too few data.
1414 *
1415 * Sometimes recovery requires the journal to be committed at least once, and
1416 * this function takes care about this.
1417 */ 1484 */
1418static void commit_on_unmount(struct ubifs_info *c) 1485static void commit_on_unmount(struct ubifs_info *c)
1419{ 1486{
diff --git a/fs/ubifs/tnc.c b/fs/ubifs/tnc.c
index 7634c597088..d27fd918b9c 100644
--- a/fs/ubifs/tnc.c
+++ b/fs/ubifs/tnc.c
@@ -284,7 +284,7 @@ static struct ubifs_znode *dirty_cow_znode(struct ubifs_info *c,
284 } 284 }
285 285
286 zn = copy_znode(c, znode); 286 zn = copy_znode(c, znode);
287 if (unlikely(IS_ERR(zn))) 287 if (IS_ERR(zn))
288 return zn; 288 return zn;
289 289
290 if (zbr->len) { 290 if (zbr->len) {
@@ -470,6 +470,10 @@ static int try_read_node(const struct ubifs_info *c, void *buf, int type,
470 if (node_len != len) 470 if (node_len != len)
471 return 0; 471 return 0;
472 472
473 if (type == UBIFS_DATA_NODE && !c->always_chk_crc)
474 if (c->no_chk_data_crc)
475 return 0;
476
473 crc = crc32(UBIFS_CRC32_INIT, buf + 8, node_len - 8); 477 crc = crc32(UBIFS_CRC32_INIT, buf + 8, node_len - 8);
474 node_crc = le32_to_cpu(ch->crc); 478 node_crc = le32_to_cpu(ch->crc);
475 if (crc != node_crc) 479 if (crc != node_crc)
@@ -1128,7 +1132,7 @@ static struct ubifs_znode *dirty_cow_bottom_up(struct ubifs_info *c,
1128 ubifs_assert(znode == c->zroot.znode); 1132 ubifs_assert(znode == c->zroot.znode);
1129 znode = dirty_cow_znode(c, &c->zroot); 1133 znode = dirty_cow_znode(c, &c->zroot);
1130 } 1134 }
1131 if (unlikely(IS_ERR(znode)) || !p) 1135 if (IS_ERR(znode) || !p)
1132 break; 1136 break;
1133 ubifs_assert(path[p - 1] >= 0); 1137 ubifs_assert(path[p - 1] >= 0);
1134 ubifs_assert(path[p - 1] < znode->child_cnt); 1138 ubifs_assert(path[p - 1] < znode->child_cnt);
@@ -1492,6 +1496,289 @@ out:
1492} 1496}
1493 1497
1494/** 1498/**
1499 * ubifs_tnc_get_bu_keys - lookup keys for bulk-read.
1500 * @c: UBIFS file-system description object
1501 * @bu: bulk-read parameters and results
1502 *
1503 * Lookup consecutive data node keys for the same inode that reside
1504 * consecutively in the same LEB.
1505 */
1506int ubifs_tnc_get_bu_keys(struct ubifs_info *c, struct bu_info *bu)
1507{
1508 int n, err = 0, lnum = -1, uninitialized_var(offs);
1509 int uninitialized_var(len);
1510 unsigned int block = key_block(c, &bu->key);
1511 struct ubifs_znode *znode;
1512
1513 bu->cnt = 0;
1514 bu->blk_cnt = 0;
1515 bu->eof = 0;
1516
1517 mutex_lock(&c->tnc_mutex);
1518 /* Find first key */
1519 err = ubifs_lookup_level0(c, &bu->key, &znode, &n);
1520 if (err < 0)
1521 goto out;
1522 if (err) {
1523 /* Key found */
1524 len = znode->zbranch[n].len;
1525 /* The buffer must be big enough for at least 1 node */
1526 if (len > bu->buf_len) {
1527 err = -EINVAL;
1528 goto out;
1529 }
1530 /* Add this key */
1531 bu->zbranch[bu->cnt++] = znode->zbranch[n];
1532 bu->blk_cnt += 1;
1533 lnum = znode->zbranch[n].lnum;
1534 offs = ALIGN(znode->zbranch[n].offs + len, 8);
1535 }
1536 while (1) {
1537 struct ubifs_zbranch *zbr;
1538 union ubifs_key *key;
1539 unsigned int next_block;
1540
1541 /* Find next key */
1542 err = tnc_next(c, &znode, &n);
1543 if (err)
1544 goto out;
1545 zbr = &znode->zbranch[n];
1546 key = &zbr->key;
1547 /* See if there is another data key for this file */
1548 if (key_inum(c, key) != key_inum(c, &bu->key) ||
1549 key_type(c, key) != UBIFS_DATA_KEY) {
1550 err = -ENOENT;
1551 goto out;
1552 }
1553 if (lnum < 0) {
1554 /* First key found */
1555 lnum = zbr->lnum;
1556 offs = ALIGN(zbr->offs + zbr->len, 8);
1557 len = zbr->len;
1558 if (len > bu->buf_len) {
1559 err = -EINVAL;
1560 goto out;
1561 }
1562 } else {
1563 /*
1564 * The data nodes must be in consecutive positions in
1565 * the same LEB.
1566 */
1567 if (zbr->lnum != lnum || zbr->offs != offs)
1568 goto out;
1569 offs += ALIGN(zbr->len, 8);
1570 len = ALIGN(len, 8) + zbr->len;
1571 /* Must not exceed buffer length */
1572 if (len > bu->buf_len)
1573 goto out;
1574 }
1575 /* Allow for holes */
1576 next_block = key_block(c, key);
1577 bu->blk_cnt += (next_block - block - 1);
1578 if (bu->blk_cnt >= UBIFS_MAX_BULK_READ)
1579 goto out;
1580 block = next_block;
1581 /* Add this key */
1582 bu->zbranch[bu->cnt++] = *zbr;
1583 bu->blk_cnt += 1;
1584 /* See if we have room for more */
1585 if (bu->cnt >= UBIFS_MAX_BULK_READ)
1586 goto out;
1587 if (bu->blk_cnt >= UBIFS_MAX_BULK_READ)
1588 goto out;
1589 }
1590out:
1591 if (err == -ENOENT) {
1592 bu->eof = 1;
1593 err = 0;
1594 }
1595 bu->gc_seq = c->gc_seq;
1596 mutex_unlock(&c->tnc_mutex);
1597 if (err)
1598 return err;
1599 /*
1600 * An enormous hole could cause bulk-read to encompass too many
1601 * page cache pages, so limit the number here.
1602 */
1603 if (bu->blk_cnt > UBIFS_MAX_BULK_READ)
1604 bu->blk_cnt = UBIFS_MAX_BULK_READ;
1605 /*
1606 * Ensure that bulk-read covers a whole number of page cache
1607 * pages.
1608 */
1609 if (UBIFS_BLOCKS_PER_PAGE == 1 ||
1610 !(bu->blk_cnt & (UBIFS_BLOCKS_PER_PAGE - 1)))
1611 return 0;
1612 if (bu->eof) {
1613 /* At the end of file we can round up */
1614 bu->blk_cnt += UBIFS_BLOCKS_PER_PAGE - 1;
1615 return 0;
1616 }
1617 /* Exclude data nodes that do not make up a whole page cache page */
1618 block = key_block(c, &bu->key) + bu->blk_cnt;
1619 block &= ~(UBIFS_BLOCKS_PER_PAGE - 1);
1620 while (bu->cnt) {
1621 if (key_block(c, &bu->zbranch[bu->cnt - 1].key) < block)
1622 break;
1623 bu->cnt -= 1;
1624 }
1625 return 0;
1626}
1627
1628/**
1629 * read_wbuf - bulk-read from a LEB with a wbuf.
1630 * @wbuf: wbuf that may overlap the read
1631 * @buf: buffer into which to read
1632 * @len: read length
1633 * @lnum: LEB number from which to read
1634 * @offs: offset from which to read
1635 *
1636 * This functions returns %0 on success or a negative error code on failure.
1637 */
1638static int read_wbuf(struct ubifs_wbuf *wbuf, void *buf, int len, int lnum,
1639 int offs)
1640{
1641 const struct ubifs_info *c = wbuf->c;
1642 int rlen, overlap;
1643
1644 dbg_io("LEB %d:%d, length %d", lnum, offs, len);
1645 ubifs_assert(wbuf && lnum >= 0 && lnum < c->leb_cnt && offs >= 0);
1646 ubifs_assert(!(offs & 7) && offs < c->leb_size);
1647 ubifs_assert(offs + len <= c->leb_size);
1648
1649 spin_lock(&wbuf->lock);
1650 overlap = (lnum == wbuf->lnum && offs + len > wbuf->offs);
1651 if (!overlap) {
1652 /* We may safely unlock the write-buffer and read the data */
1653 spin_unlock(&wbuf->lock);
1654 return ubi_read(c->ubi, lnum, buf, offs, len);
1655 }
1656
1657 /* Don't read under wbuf */
1658 rlen = wbuf->offs - offs;
1659 if (rlen < 0)
1660 rlen = 0;
1661
1662 /* Copy the rest from the write-buffer */
1663 memcpy(buf + rlen, wbuf->buf + offs + rlen - wbuf->offs, len - rlen);
1664 spin_unlock(&wbuf->lock);
1665
1666 if (rlen > 0)
1667 /* Read everything that goes before write-buffer */
1668 return ubi_read(c->ubi, lnum, buf, offs, rlen);
1669
1670 return 0;
1671}
1672
1673/**
1674 * validate_data_node - validate data nodes for bulk-read.
1675 * @c: UBIFS file-system description object
1676 * @buf: buffer containing data node to validate
1677 * @zbr: zbranch of data node to validate
1678 *
1679 * This functions returns %0 on success or a negative error code on failure.
1680 */
1681static int validate_data_node(struct ubifs_info *c, void *buf,
1682 struct ubifs_zbranch *zbr)
1683{
1684 union ubifs_key key1;
1685 struct ubifs_ch *ch = buf;
1686 int err, len;
1687
1688 if (ch->node_type != UBIFS_DATA_NODE) {
1689 ubifs_err("bad node type (%d but expected %d)",
1690 ch->node_type, UBIFS_DATA_NODE);
1691 goto out_err;
1692 }
1693
1694 err = ubifs_check_node(c, buf, zbr->lnum, zbr->offs, 0, 0);
1695 if (err) {
1696 ubifs_err("expected node type %d", UBIFS_DATA_NODE);
1697 goto out;
1698 }
1699
1700 len = le32_to_cpu(ch->len);
1701 if (len != zbr->len) {
1702 ubifs_err("bad node length %d, expected %d", len, zbr->len);
1703 goto out_err;
1704 }
1705
1706 /* Make sure the key of the read node is correct */
1707 key_read(c, buf + UBIFS_KEY_OFFSET, &key1);
1708 if (!keys_eq(c, &zbr->key, &key1)) {
1709 ubifs_err("bad key in node at LEB %d:%d",
1710 zbr->lnum, zbr->offs);
1711 dbg_tnc("looked for key %s found node's key %s",
1712 DBGKEY(&zbr->key), DBGKEY1(&key1));
1713 goto out_err;
1714 }
1715
1716 return 0;
1717
1718out_err:
1719 err = -EINVAL;
1720out:
1721 ubifs_err("bad node at LEB %d:%d", zbr->lnum, zbr->offs);
1722 dbg_dump_node(c, buf);
1723 dbg_dump_stack();
1724 return err;
1725}
1726
1727/**
1728 * ubifs_tnc_bulk_read - read a number of data nodes in one go.
1729 * @c: UBIFS file-system description object
1730 * @bu: bulk-read parameters and results
1731 *
1732 * This functions reads and validates the data nodes that were identified by the
1733 * 'ubifs_tnc_get_bu_keys()' function. This functions returns %0 on success,
1734 * -EAGAIN to indicate a race with GC, or another negative error code on
1735 * failure.
1736 */
1737int ubifs_tnc_bulk_read(struct ubifs_info *c, struct bu_info *bu)
1738{
1739 int lnum = bu->zbranch[0].lnum, offs = bu->zbranch[0].offs, len, err, i;
1740 struct ubifs_wbuf *wbuf;
1741 void *buf;
1742
1743 len = bu->zbranch[bu->cnt - 1].offs;
1744 len += bu->zbranch[bu->cnt - 1].len - offs;
1745 if (len > bu->buf_len) {
1746 ubifs_err("buffer too small %d vs %d", bu->buf_len, len);
1747 return -EINVAL;
1748 }
1749
1750 /* Do the read */
1751 wbuf = ubifs_get_wbuf(c, lnum);
1752 if (wbuf)
1753 err = read_wbuf(wbuf, bu->buf, len, lnum, offs);
1754 else
1755 err = ubi_read(c->ubi, lnum, bu->buf, offs, len);
1756
1757 /* Check for a race with GC */
1758 if (maybe_leb_gced(c, lnum, bu->gc_seq))
1759 return -EAGAIN;
1760
1761 if (err && err != -EBADMSG) {
1762 ubifs_err("failed to read from LEB %d:%d, error %d",
1763 lnum, offs, err);
1764 dbg_dump_stack();
1765 dbg_tnc("key %s", DBGKEY(&bu->key));
1766 return err;
1767 }
1768
1769 /* Validate the nodes read */
1770 buf = bu->buf;
1771 for (i = 0; i < bu->cnt; i++) {
1772 err = validate_data_node(c, buf, &bu->zbranch[i]);
1773 if (err)
1774 return err;
1775 buf = buf + ALIGN(bu->zbranch[i].len, 8);
1776 }
1777
1778 return 0;
1779}
1780
1781/**
1495 * do_lookup_nm- look up a "hashed" node. 1782 * do_lookup_nm- look up a "hashed" node.
1496 * @c: UBIFS file-system description object 1783 * @c: UBIFS file-system description object
1497 * @key: node key to lookup 1784 * @key: node key to lookup
@@ -1675,7 +1962,7 @@ static int tnc_insert(struct ubifs_info *c, struct ubifs_znode *znode,
1675{ 1962{
1676 struct ubifs_znode *zn, *zi, *zp; 1963 struct ubifs_znode *zn, *zi, *zp;
1677 int i, keep, move, appending = 0; 1964 int i, keep, move, appending = 0;
1678 union ubifs_key *key = &zbr->key; 1965 union ubifs_key *key = &zbr->key, *key1;
1679 1966
1680 ubifs_assert(n >= 0 && n <= c->fanout); 1967 ubifs_assert(n >= 0 && n <= c->fanout);
1681 1968
@@ -1716,20 +2003,33 @@ again:
1716 zn->level = znode->level; 2003 zn->level = znode->level;
1717 2004
1718 /* Decide where to split */ 2005 /* Decide where to split */
1719 if (znode->level == 0 && n == c->fanout && 2006 if (znode->level == 0 && key_type(c, key) == UBIFS_DATA_KEY) {
1720 key_type(c, key) == UBIFS_DATA_KEY) { 2007 /* Try not to split consecutive data keys */
1721 union ubifs_key *key1; 2008 if (n == c->fanout) {
1722 2009 key1 = &znode->zbranch[n - 1].key;
1723 /* 2010 if (key_inum(c, key1) == key_inum(c, key) &&
1724 * If this is an inode which is being appended - do not split 2011 key_type(c, key1) == UBIFS_DATA_KEY)
1725 * it because no other zbranches can be inserted between 2012 appending = 1;
1726 * zbranches of consecutive data nodes anyway. 2013 } else
1727 */ 2014 goto check_split;
1728 key1 = &znode->zbranch[n - 1].key; 2015 } else if (appending && n != c->fanout) {
1729 if (key_inum(c, key1) == key_inum(c, key) && 2016 /* Try not to split consecutive data keys */
1730 key_type(c, key1) == UBIFS_DATA_KEY && 2017 appending = 0;
1731 key_block(c, key1) == key_block(c, key) - 1) 2018check_split:
1732 appending = 1; 2019 if (n >= (c->fanout + 1) / 2) {
2020 key1 = &znode->zbranch[0].key;
2021 if (key_inum(c, key1) == key_inum(c, key) &&
2022 key_type(c, key1) == UBIFS_DATA_KEY) {
2023 key1 = &znode->zbranch[n].key;
2024 if (key_inum(c, key1) != key_inum(c, key) ||
2025 key_type(c, key1) != UBIFS_DATA_KEY) {
2026 keep = n;
2027 move = c->fanout - keep;
2028 zi = znode;
2029 goto do_split;
2030 }
2031 }
2032 }
1733 } 2033 }
1734 2034
1735 if (appending) { 2035 if (appending) {
@@ -1759,6 +2059,8 @@ again:
1759 zbr->znode->parent = zn; 2059 zbr->znode->parent = zn;
1760 } 2060 }
1761 2061
2062do_split:
2063
1762 __set_bit(DIRTY_ZNODE, &zn->flags); 2064 __set_bit(DIRTY_ZNODE, &zn->flags);
1763 atomic_long_inc(&c->dirty_zn_cnt); 2065 atomic_long_inc(&c->dirty_zn_cnt);
1764 2066
@@ -1785,14 +2087,11 @@ again:
1785 2087
1786 /* Insert new znode (produced by spitting) into the parent */ 2088 /* Insert new znode (produced by spitting) into the parent */
1787 if (zp) { 2089 if (zp) {
1788 i = n; 2090 if (n == 0 && zi == znode && znode->iip == 0)
2091 correct_parent_keys(c, znode);
2092
1789 /* Locate insertion point */ 2093 /* Locate insertion point */
1790 n = znode->iip + 1; 2094 n = znode->iip + 1;
1791 if (appending && n != c->fanout)
1792 appending = 0;
1793
1794 if (i == 0 && zi == znode && znode->iip == 0)
1795 correct_parent_keys(c, znode);
1796 2095
1797 /* Tail recursion */ 2096 /* Tail recursion */
1798 zbr->key = zn->zbranch[0].key; 2097 zbr->key = zn->zbranch[0].key;
diff --git a/fs/ubifs/tnc_misc.c b/fs/ubifs/tnc_misc.c
index a25c1cc1f8d..b48db999903 100644
--- a/fs/ubifs/tnc_misc.c
+++ b/fs/ubifs/tnc_misc.c
@@ -480,8 +480,8 @@ int ubifs_tnc_read_node(struct ubifs_info *c, struct ubifs_zbranch *zbr,
480 } 480 }
481 481
482 /* Make sure the key of the read node is correct */ 482 /* Make sure the key of the read node is correct */
483 key_read(c, key, &key1); 483 key_read(c, node + UBIFS_KEY_OFFSET, &key1);
484 if (memcmp(node + UBIFS_KEY_OFFSET, &key1, c->key_len)) { 484 if (!keys_eq(c, key, &key1)) {
485 ubifs_err("bad key in node at LEB %d:%d", 485 ubifs_err("bad key in node at LEB %d:%d",
486 zbr->lnum, zbr->offs); 486 zbr->lnum, zbr->offs);
487 dbg_tnc("looked for key %s found node's key %s", 487 dbg_tnc("looked for key %s found node's key %s",
diff --git a/fs/ubifs/ubifs-media.h b/fs/ubifs/ubifs-media.h
index a9ecbd9af20..0b378042a3a 100644
--- a/fs/ubifs/ubifs-media.h
+++ b/fs/ubifs/ubifs-media.h
@@ -75,7 +75,6 @@
75 */ 75 */
76#define UBIFS_BLOCK_SIZE 4096 76#define UBIFS_BLOCK_SIZE 4096
77#define UBIFS_BLOCK_SHIFT 12 77#define UBIFS_BLOCK_SHIFT 12
78#define UBIFS_BLOCK_MASK 0x00000FFF
79 78
80/* UBIFS padding byte pattern (must not be first or last byte of node magic) */ 79/* UBIFS padding byte pattern (must not be first or last byte of node magic) */
81#define UBIFS_PADDING_BYTE 0xCE 80#define UBIFS_PADDING_BYTE 0xCE
diff --git a/fs/ubifs/ubifs.h b/fs/ubifs/ubifs.h
index 17c620b93ee..a7bd32fa15b 100644
--- a/fs/ubifs/ubifs.h
+++ b/fs/ubifs/ubifs.h
@@ -142,6 +142,9 @@
142/* Maximum expected tree height for use by bottom_up_buf */ 142/* Maximum expected tree height for use by bottom_up_buf */
143#define BOTTOM_UP_HEIGHT 64 143#define BOTTOM_UP_HEIGHT 64
144 144
145/* Maximum number of data nodes to bulk-read */
146#define UBIFS_MAX_BULK_READ 32
147
145/* 148/*
146 * Lockdep classes for UBIFS inode @ui_mutex. 149 * Lockdep classes for UBIFS inode @ui_mutex.
147 */ 150 */
@@ -328,9 +331,10 @@ struct ubifs_gced_idx_leb {
328 * this inode 331 * this inode
329 * @dirty: non-zero if the inode is dirty 332 * @dirty: non-zero if the inode is dirty
330 * @xattr: non-zero if this is an extended attribute inode 333 * @xattr: non-zero if this is an extended attribute inode
334 * @bulk_read: non-zero if bulk-read should be used
331 * @ui_mutex: serializes inode write-back with the rest of VFS operations, 335 * @ui_mutex: serializes inode write-back with the rest of VFS operations,
332 * serializes "clean <-> dirty" state changes, protects @dirty, 336 * serializes "clean <-> dirty" state changes, serializes bulk-read,
333 * @ui_size, and @xattr_size 337 * protects @dirty, @bulk_read, @ui_size, and @xattr_size
334 * @ui_lock: protects @synced_i_size 338 * @ui_lock: protects @synced_i_size
335 * @synced_i_size: synchronized size of inode, i.e. the value of inode size 339 * @synced_i_size: synchronized size of inode, i.e. the value of inode size
336 * currently stored on the flash; used only for regular file 340 * currently stored on the flash; used only for regular file
@@ -338,6 +342,8 @@ struct ubifs_gced_idx_leb {
338 * @ui_size: inode size used by UBIFS when writing to flash 342 * @ui_size: inode size used by UBIFS when writing to flash
339 * @flags: inode flags (@UBIFS_COMPR_FL, etc) 343 * @flags: inode flags (@UBIFS_COMPR_FL, etc)
340 * @compr_type: default compression type used for this inode 344 * @compr_type: default compression type used for this inode
345 * @last_page_read: page number of last page read (for bulk read)
346 * @read_in_a_row: number of consecutive pages read in a row (for bulk read)
341 * @data_len: length of the data attached to the inode 347 * @data_len: length of the data attached to the inode
342 * @data: inode's data 348 * @data: inode's data
343 * 349 *
@@ -379,12 +385,15 @@ struct ubifs_inode {
379 unsigned int xattr_names; 385 unsigned int xattr_names;
380 unsigned int dirty:1; 386 unsigned int dirty:1;
381 unsigned int xattr:1; 387 unsigned int xattr:1;
388 unsigned int bulk_read:1;
382 struct mutex ui_mutex; 389 struct mutex ui_mutex;
383 spinlock_t ui_lock; 390 spinlock_t ui_lock;
384 loff_t synced_i_size; 391 loff_t synced_i_size;
385 loff_t ui_size; 392 loff_t ui_size;
386 int flags; 393 int flags;
387 int compr_type; 394 int compr_type;
395 pgoff_t last_page_read;
396 pgoff_t read_in_a_row;
388 int data_len; 397 int data_len;
389 void *data; 398 void *data;
390}; 399};
@@ -698,8 +707,8 @@ struct ubifs_jhead {
698 * struct ubifs_zbranch - key/coordinate/length branch stored in znodes. 707 * struct ubifs_zbranch - key/coordinate/length branch stored in znodes.
699 * @key: key 708 * @key: key
700 * @znode: znode address in memory 709 * @znode: znode address in memory
701 * @lnum: LEB number of the indexing node 710 * @lnum: LEB number of the target node (indexing node or data node)
702 * @offs: offset of the indexing node within @lnum 711 * @offs: target node offset within @lnum
703 * @len: target node length 712 * @len: target node length
704 */ 713 */
705struct ubifs_zbranch { 714struct ubifs_zbranch {
@@ -744,6 +753,28 @@ struct ubifs_znode {
744}; 753};
745 754
746/** 755/**
756 * struct bu_info - bulk-read information
757 * @key: first data node key
758 * @zbranch: zbranches of data nodes to bulk read
759 * @buf: buffer to read into
760 * @buf_len: buffer length
761 * @gc_seq: GC sequence number to detect races with GC
762 * @cnt: number of data nodes for bulk read
763 * @blk_cnt: number of data blocks including holes
764 * @oef: end of file reached
765 */
766struct bu_info {
767 union ubifs_key key;
768 struct ubifs_zbranch zbranch[UBIFS_MAX_BULK_READ];
769 void *buf;
770 int buf_len;
771 int gc_seq;
772 int cnt;
773 int blk_cnt;
774 int eof;
775};
776
777/**
747 * struct ubifs_node_range - node length range description data structure. 778 * struct ubifs_node_range - node length range description data structure.
748 * @len: fixed node length 779 * @len: fixed node length
749 * @min_len: minimum possible node length 780 * @min_len: minimum possible node length
@@ -862,9 +893,13 @@ struct ubifs_orphan {
862/** 893/**
863 * struct ubifs_mount_opts - UBIFS-specific mount options information. 894 * struct ubifs_mount_opts - UBIFS-specific mount options information.
864 * @unmount_mode: selected unmount mode (%0 default, %1 normal, %2 fast) 895 * @unmount_mode: selected unmount mode (%0 default, %1 normal, %2 fast)
896 * @bulk_read: enable bulk-reads
897 * @chk_data_crc: check CRCs when reading data nodes
865 */ 898 */
866struct ubifs_mount_opts { 899struct ubifs_mount_opts {
867 unsigned int unmount_mode:2; 900 unsigned int unmount_mode:2;
901 unsigned int bulk_read:2;
902 unsigned int chk_data_crc:2;
868}; 903};
869 904
870/** 905/**
@@ -905,13 +940,12 @@ struct ubifs_mount_opts {
905 * @cmt_state: commit state 940 * @cmt_state: commit state
906 * @cs_lock: commit state lock 941 * @cs_lock: commit state lock
907 * @cmt_wq: wait queue to sleep on if the log is full and a commit is running 942 * @cmt_wq: wait queue to sleep on if the log is full and a commit is running
943 *
908 * @fast_unmount: do not run journal commit before un-mounting 944 * @fast_unmount: do not run journal commit before un-mounting
909 * @big_lpt: flag that LPT is too big to write whole during commit 945 * @big_lpt: flag that LPT is too big to write whole during commit
910 * @check_lpt_free: flag that indicates LPT GC may be needed 946 * @no_chk_data_crc: do not check CRCs when reading data nodes (except during
911 * @nospace: non-zero if the file-system does not have flash space (used as 947 * recovery)
912 * optimization) 948 * @bulk_read: enable bulk-reads
913 * @nospace_rp: the same as @nospace, but additionally means that even reserved
914 * pool is full
915 * 949 *
916 * @tnc_mutex: protects the Tree Node Cache (TNC), @zroot, @cnext, @enext, and 950 * @tnc_mutex: protects the Tree Node Cache (TNC), @zroot, @cnext, @enext, and
917 * @calc_idx_sz 951 * @calc_idx_sz
@@ -935,6 +969,7 @@ struct ubifs_mount_opts {
935 * @mst_node: master node 969 * @mst_node: master node
936 * @mst_offs: offset of valid master node 970 * @mst_offs: offset of valid master node
937 * @mst_mutex: protects the master node area, @mst_node, and @mst_offs 971 * @mst_mutex: protects the master node area, @mst_node, and @mst_offs
972 * @bulk_read_buf_size: buffer size for bulk-reads
938 * 973 *
939 * @log_lebs: number of logical eraseblocks in the log 974 * @log_lebs: number of logical eraseblocks in the log
940 * @log_bytes: log size in bytes 975 * @log_bytes: log size in bytes
@@ -977,12 +1012,17 @@ struct ubifs_mount_opts {
977 * but which still have to be taken into account because 1012 * but which still have to be taken into account because
978 * the index has not been committed so far 1013 * the index has not been committed so far
979 * @space_lock: protects @budg_idx_growth, @budg_data_growth, @budg_dd_growth, 1014 * @space_lock: protects @budg_idx_growth, @budg_data_growth, @budg_dd_growth,
980 * @budg_uncommited_idx, @min_idx_lebs, @old_idx_sz, and @lst; 1015 * @budg_uncommited_idx, @min_idx_lebs, @old_idx_sz, @lst,
1016 * @nospace, and @nospace_rp;
981 * @min_idx_lebs: minimum number of LEBs required for the index 1017 * @min_idx_lebs: minimum number of LEBs required for the index
982 * @old_idx_sz: size of index on flash 1018 * @old_idx_sz: size of index on flash
983 * @calc_idx_sz: temporary variable which is used to calculate new index size 1019 * @calc_idx_sz: temporary variable which is used to calculate new index size
984 * (contains accurate new index size at end of TNC commit start) 1020 * (contains accurate new index size at end of TNC commit start)
985 * @lst: lprops statistics 1021 * @lst: lprops statistics
1022 * @nospace: non-zero if the file-system does not have flash space (used as
1023 * optimization)
1024 * @nospace_rp: the same as @nospace, but additionally means that even reserved
1025 * pool is full
986 * 1026 *
987 * @page_budget: budget for a page 1027 * @page_budget: budget for a page
988 * @inode_budget: budget for an inode 1028 * @inode_budget: budget for an inode
@@ -1061,6 +1101,7 @@ struct ubifs_mount_opts {
1061 * @lpt_drty_flgs: dirty flags for LPT special nodes e.g. ltab 1101 * @lpt_drty_flgs: dirty flags for LPT special nodes e.g. ltab
1062 * @dirty_nn_cnt: number of dirty nnodes 1102 * @dirty_nn_cnt: number of dirty nnodes
1063 * @dirty_pn_cnt: number of dirty pnodes 1103 * @dirty_pn_cnt: number of dirty pnodes
1104 * @check_lpt_free: flag that indicates LPT GC may be needed
1064 * @lpt_sz: LPT size 1105 * @lpt_sz: LPT size
1065 * @lpt_nod_buf: buffer for an on-flash nnode or pnode 1106 * @lpt_nod_buf: buffer for an on-flash nnode or pnode
1066 * @lpt_buf: buffer of LEB size used by LPT 1107 * @lpt_buf: buffer of LEB size used by LPT
@@ -1102,6 +1143,7 @@ struct ubifs_mount_opts {
1102 * @rcvrd_mst_node: recovered master node to write when mounting ro to rw 1143 * @rcvrd_mst_node: recovered master node to write when mounting ro to rw
1103 * @size_tree: inode size information for recovery 1144 * @size_tree: inode size information for recovery
1104 * @remounting_rw: set while remounting from ro to rw (sb flags have MS_RDONLY) 1145 * @remounting_rw: set while remounting from ro to rw (sb flags have MS_RDONLY)
1146 * @always_chk_crc: always check CRCs (while mounting and remounting rw)
1105 * @mount_opts: UBIFS-specific mount options 1147 * @mount_opts: UBIFS-specific mount options
1106 * 1148 *
1107 * @dbg_buf: a buffer of LEB size used for debugging purposes 1149 * @dbg_buf: a buffer of LEB size used for debugging purposes
@@ -1146,11 +1188,11 @@ struct ubifs_info {
1146 int cmt_state; 1188 int cmt_state;
1147 spinlock_t cs_lock; 1189 spinlock_t cs_lock;
1148 wait_queue_head_t cmt_wq; 1190 wait_queue_head_t cmt_wq;
1191
1149 unsigned int fast_unmount:1; 1192 unsigned int fast_unmount:1;
1150 unsigned int big_lpt:1; 1193 unsigned int big_lpt:1;
1151 unsigned int check_lpt_free:1; 1194 unsigned int no_chk_data_crc:1;
1152 unsigned int nospace:1; 1195 unsigned int bulk_read:1;
1153 unsigned int nospace_rp:1;
1154 1196
1155 struct mutex tnc_mutex; 1197 struct mutex tnc_mutex;
1156 struct ubifs_zbranch zroot; 1198 struct ubifs_zbranch zroot;
@@ -1175,6 +1217,7 @@ struct ubifs_info {
1175 struct ubifs_mst_node *mst_node; 1217 struct ubifs_mst_node *mst_node;
1176 int mst_offs; 1218 int mst_offs;
1177 struct mutex mst_mutex; 1219 struct mutex mst_mutex;
1220 int bulk_read_buf_size;
1178 1221
1179 int log_lebs; 1222 int log_lebs;
1180 long long log_bytes; 1223 long long log_bytes;
@@ -1218,6 +1261,8 @@ struct ubifs_info {
1218 unsigned long long old_idx_sz; 1261 unsigned long long old_idx_sz;
1219 unsigned long long calc_idx_sz; 1262 unsigned long long calc_idx_sz;
1220 struct ubifs_lp_stats lst; 1263 struct ubifs_lp_stats lst;
1264 unsigned int nospace:1;
1265 unsigned int nospace_rp:1;
1221 1266
1222 int page_budget; 1267 int page_budget;
1223 int inode_budget; 1268 int inode_budget;
@@ -1294,6 +1339,7 @@ struct ubifs_info {
1294 int lpt_drty_flgs; 1339 int lpt_drty_flgs;
1295 int dirty_nn_cnt; 1340 int dirty_nn_cnt;
1296 int dirty_pn_cnt; 1341 int dirty_pn_cnt;
1342 int check_lpt_free;
1297 long long lpt_sz; 1343 long long lpt_sz;
1298 void *lpt_nod_buf; 1344 void *lpt_nod_buf;
1299 void *lpt_buf; 1345 void *lpt_buf;
@@ -1335,6 +1381,7 @@ struct ubifs_info {
1335 struct ubifs_mst_node *rcvrd_mst_node; 1381 struct ubifs_mst_node *rcvrd_mst_node;
1336 struct rb_root size_tree; 1382 struct rb_root size_tree;
1337 int remounting_rw; 1383 int remounting_rw;
1384 int always_chk_crc;
1338 struct ubifs_mount_opts mount_opts; 1385 struct ubifs_mount_opts mount_opts;
1339 1386
1340#ifdef CONFIG_UBIFS_FS_DEBUG 1387#ifdef CONFIG_UBIFS_FS_DEBUG
@@ -1347,6 +1394,12 @@ struct ubifs_info {
1347 unsigned long fail_timeout; 1394 unsigned long fail_timeout;
1348 unsigned int fail_cnt; 1395 unsigned int fail_cnt;
1349 unsigned int fail_cnt_max; 1396 unsigned int fail_cnt_max;
1397 long long chk_lpt_sz;
1398 long long chk_lpt_sz2;
1399 long long chk_lpt_wastage;
1400 int chk_lpt_lebs;
1401 int new_nhead_lnum;
1402 int new_nhead_offs;
1350#endif 1403#endif
1351}; 1404};
1352 1405
@@ -1377,7 +1430,7 @@ int ubifs_read_node_wbuf(struct ubifs_wbuf *wbuf, void *buf, int type, int len,
1377int ubifs_write_node(struct ubifs_info *c, void *node, int len, int lnum, 1430int ubifs_write_node(struct ubifs_info *c, void *node, int len, int lnum,
1378 int offs, int dtype); 1431 int offs, int dtype);
1379int ubifs_check_node(const struct ubifs_info *c, const void *buf, int lnum, 1432int ubifs_check_node(const struct ubifs_info *c, const void *buf, int lnum,
1380 int offs, int quiet); 1433 int offs, int quiet, int chk_crc);
1381void ubifs_prepare_node(struct ubifs_info *c, void *buf, int len, int pad); 1434void ubifs_prepare_node(struct ubifs_info *c, void *buf, int len, int pad);
1382void ubifs_prep_grp_node(struct ubifs_info *c, void *node, int len, int last); 1435void ubifs_prep_grp_node(struct ubifs_info *c, void *node, int len, int last);
1383int ubifs_io_init(struct ubifs_info *c); 1436int ubifs_io_init(struct ubifs_info *c);
@@ -1490,6 +1543,8 @@ void destroy_old_idx(struct ubifs_info *c);
1490int is_idx_node_in_tnc(struct ubifs_info *c, union ubifs_key *key, int level, 1543int is_idx_node_in_tnc(struct ubifs_info *c, union ubifs_key *key, int level,
1491 int lnum, int offs); 1544 int lnum, int offs);
1492int insert_old_idx_znode(struct ubifs_info *c, struct ubifs_znode *znode); 1545int insert_old_idx_znode(struct ubifs_info *c, struct ubifs_znode *znode);
1546int ubifs_tnc_get_bu_keys(struct ubifs_info *c, struct bu_info *bu);
1547int ubifs_tnc_bulk_read(struct ubifs_info *c, struct bu_info *bu);
1493 1548
1494/* tnc_misc.c */ 1549/* tnc_misc.c */
1495struct ubifs_znode *ubifs_tnc_levelorder_next(struct ubifs_znode *zr, 1550struct ubifs_znode *ubifs_tnc_levelorder_next(struct ubifs_znode *zr,
@@ -1586,12 +1641,10 @@ int ubifs_lpt_post_commit(struct ubifs_info *c);
1586void ubifs_lpt_free(struct ubifs_info *c, int wr_only); 1641void ubifs_lpt_free(struct ubifs_info *c, int wr_only);
1587 1642
1588/* lprops.c */ 1643/* lprops.c */
1589void ubifs_get_lprops(struct ubifs_info *c);
1590const struct ubifs_lprops *ubifs_change_lp(struct ubifs_info *c, 1644const struct ubifs_lprops *ubifs_change_lp(struct ubifs_info *c,
1591 const struct ubifs_lprops *lp, 1645 const struct ubifs_lprops *lp,
1592 int free, int dirty, int flags, 1646 int free, int dirty, int flags,
1593 int idx_gc_cnt); 1647 int idx_gc_cnt);
1594void ubifs_release_lprops(struct ubifs_info *c);
1595void ubifs_get_lp_stats(struct ubifs_info *c, struct ubifs_lp_stats *stats); 1648void ubifs_get_lp_stats(struct ubifs_info *c, struct ubifs_lp_stats *stats);
1596void ubifs_add_to_cat(struct ubifs_info *c, struct ubifs_lprops *lprops, 1649void ubifs_add_to_cat(struct ubifs_info *c, struct ubifs_lprops *lprops,
1597 int cat); 1650 int cat);
diff --git a/fs/ubifs/xattr.c b/fs/ubifs/xattr.c
index 649bec78b64..cfd31e229c8 100644
--- a/fs/ubifs/xattr.c
+++ b/fs/ubifs/xattr.c
@@ -446,7 +446,7 @@ ssize_t ubifs_listxattr(struct dentry *dentry, char *buffer, size_t size)
446 int type; 446 int type;
447 447
448 xent = ubifs_tnc_next_ent(c, &key, &nm); 448 xent = ubifs_tnc_next_ent(c, &key, &nm);
449 if (unlikely(IS_ERR(xent))) { 449 if (IS_ERR(xent)) {
450 err = PTR_ERR(xent); 450 err = PTR_ERR(xent);
451 break; 451 break;
452 } 452 }
generated by cgit v1.2.2 (git 2.25.0) at 2026-01-03 06:58:46 -0500