summaryrefslogtreecommitdiff
path: root/Documentation/filesystems/porting
blob: d44511e2082807937c1233c851da742297dd0e4e (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
Changes since 2.5.0:

---
[recommended]

New helpers: sb_bread(), sb_getblk(), sb_find_get_block(), set_bh(),
	sb_set_blocksize() and sb_min_blocksize().

Use them.

(sb_find_get_block() replaces 2.4's get_hash_table())

---
[recommended]

New methods: ->alloc_inode() and ->destroy_inode().

Remove inode->u.foo_inode_i
Declare
	struct foo_inode_info {
		/* fs-private stuff */
		struct inode vfs_inode;
	};
	static inline struct foo_inode_info *FOO_I(struct inode *inode)
	{
		return list_entry(inode, struct foo_inode_info, vfs_inode);
	}

Use FOO_I(inode) instead of &inode->u.foo_inode_i;

Add foo_alloc_inode() and foo_destroy_inode() - the former should allocate
foo_inode_info and return the address of ->vfs_inode, the latter should free
FOO_I(inode) (see in-tree filesystems for examples).

Make them ->alloc_inode and ->destroy_inode in your super_operations.

Keep in mind that now you need explicit initialization of private data
typically between calling iget_locked() and unlocking the inode.

At some point that will become mandatory.

---
[mandatory]

Change of file_system_type method (->read_super to ->get_sb)

->read_super() is no more.  Ditto for DECLARE_FSTYPE and DECLARE_FSTYPE_DEV.

Turn your foo_read_super() into a function that would return 0 in case of
success and negative number in case of error (-EINVAL unless you have more
informative error value to report).  Call it foo_fill_super().  Now declare

int foo_get_sb(struct file_system_type *fs_type,
	int flags, const char *dev_name, void *data, struct vfsmount *mnt)
{
	return get_sb_bdev(fs_type, flags, dev_name, data, foo_fill_super,
			   mnt);
}

(or similar with s/bdev/nodev/ or s/bdev/single/, depending on the kind of
filesystem).

Replace DECLARE_FSTYPE... with explicit initializer and have ->get_sb set as
foo_get_sb.

---
[mandatory]

Locking change: ->s_vfs_rename_sem is taken only by cross-directory renames.
Most likely there is no need to change anything, but if you relied on
global exclusion between renames for some internal purpose - you need to
change your internal locking.  Otherwise exclusion warranties remain the
same (i.e. parents and victim are locked, etc.).

---
[informational]

Now we have the exclusion between ->lookup() and directory removal (by
->rmdir() and ->rename()).  If you used to need that exclusion and do
it by internal locking (most of filesystems couldn't care less) - you
can relax your locking.

---
[mandatory]

->lookup(), ->truncate(), ->create(), ->unlink(), ->mknod(), ->mkdir(),
->rmdir(), ->link(), ->lseek(), ->symlink(), ->rename()
and ->readdir() are called without BKL now.  Grab it on entry, drop upon return
- that will guarantee the same locking you used to have.  If your method or its
parts do not need BKL - better yet, now you can shift lock_kernel() and
unlock_kernel() so that they would protect exactly what needs to be
protected.

---
[mandatory]

BKL is also moved from around sb operations.  ->write_super() Is now called 
without BKL held.  BKL should have been shifted into individual fs sb_op
functions.  If you don't need it, remove it.  

---
[informational]

check for ->link() target not being a directory is done by callers.  Feel
free to drop it...

---
[informational]

->link() callers hold ->i_mutex on the object we are linking to.  Some of your
problems might be over...

---
[mandatory]

new file_system_type method - kill_sb(superblock).  If you are converting
an existing filesystem, set it according to ->fs_flags:
	FS_REQUIRES_DEV		-	kill_block_super
	FS_LITTER		-	kill_litter_super
	neither			-	kill_anon_super
FS_LITTER is gone - just remove it from fs_flags.

---
[mandatory]

	FS_SINGLE is gone (actually, that had happened back when ->get_sb()
went in - and hadn't been documented ;-/).  Just remove it from fs_flags
(and see ->get_sb() entry for other actions).

---
[mandatory]

->setattr() is called without BKL now.  Caller _always_ holds ->i_mutex, so
watch for ->i_mutex-grabbing code that might be used by your ->setattr().
Callers of notify_change() need ->i_mutex now.

---
[recommended]

New super_block field "struct export_operations *s_export_op" for
explicit support for exporting, e.g. via NFS.  The structure is fully
documented at its declaration in include/linux/fs.h, and in
Documentation/filesystems/nfs/Exporting.

Briefly it allows for the definition of decode_fh and encode_fh operations
to encode and decode filehandles, and allows the filesystem to use
a standard helper function for decode_fh, and provide file-system specific
support for this helper, particularly get_parent.

It is planned that this will be required for exporting once the code
settles down a bit.

[mandatory]

s_export_op is now required for exporting a filesystem.
isofs, ext2, ext3, resierfs, fat
can be used as examples of very different filesystems.

---
[mandatory]

iget4() and the read_inode2 callback have been superseded by iget5_locked()
which has the following prototype,

    struct inode *iget5_locked(struct super_block *sb, unsigned long ino,
				int (*test)(struct inode *, void *),
				int (*set)(struct inode *, void *),
				void *data);

'test' is an additional function that can be used when the inode
number is not sufficient to identify the actual file object. 'set'
should be a non-blocking function that initializes those parts of a
newly created inode to allow the test function to succeed. 'data' is
passed as an opaque value to both test and set functions.

When the inode has been created by iget5_locked(), it will be returned with the
I_NEW flag set and will still be locked.  The filesystem then needs to finalize
the initialization. Once the inode is initialized it must be unlocked by
calling unlock_new_inode().

The filesystem is responsible for setting (and possibly testing) i_ino
when appropriate. There is also a simpler iget_locked function that
just takes the superblock and inode number as arguments and does the
test and set for you.

e.g.
	inode = iget_locked(sb, ino);
	if (inode->i_state & I_NEW) {
		err = read_inode_from_disk(inode);
		if (err < 0) {
			iget_failed(inode);
			return err;
		}
		unlock_new_inode(inode);
	}

Note that if the process of setting up a new inode fails, then iget_failed()
should be called on the inode to render it dead, and an appropriate error
should be passed back to the caller.

---
[recommended]

->getattr() finally getting used.  See instances in nfs, minix, etc.

---
[mandatory]

->revalidate() is gone.  If your filesystem had it - provide ->getattr()
and let it call whatever you had as ->revlidate() + (for symlinks that
had ->revalidate()) add calls in ->follow_link()/->readlink().

---
[mandatory]

->d_parent changes are not protected by BKL anymore.  Read access is safe
if at least one of the following is true:
	* filesystem has no cross-directory rename()
	* dcache_lock is held
	* we know that parent had been locked (e.g. we are looking at
->d_parent of ->lookup() argument).
	* we are called from ->rename().
	* the child's ->d_lock is held
Audit your code and add locking if needed.  Notice that any place that is
not protected by the conditions above is risky even in the old tree - you
had been relying on BKL and that's prone to screwups.  Old tree had quite
a few holes of that kind - unprotected access to ->d_parent leading to
anything from oops to silent memory corruption.

---
[mandatory]

	FS_NOMOUNT is gone.  If you use it - just set MS_NOUSER in flags
(see rootfs for one kind of solution and bdev/socket/pipe for another).

---
[recommended]

	Use bdev_read_only(bdev) instead of is_read_only(kdev).  The latter
is still alive, but only because of the mess in drivers/s390/block/dasd.c.
As soon as it gets fixed is_read_only() will die.

---
[mandatory]

->permission() is called without BKL now. Grab it on entry, drop upon
return - that will guarantee the same locking you used to have.  If
your method or its parts do not need BKL - better yet, now you can
shift lock_kernel() and unlock_kernel() so that they would protect
exactly what needs to be protected.

---
[mandatory]

->statfs() is now called without BKL held.  BKL should have been
shifted into individual fs sb_op functions where it's not clear that
it's safe to remove it.  If you don't need it, remove it.

---
[mandatory]

	is_read_only() is gone; use bdev_read_only() instead.

---
[mandatory]

	destroy_buffers() is gone; use invalidate_bdev().

---
[mandatory]

	fsync_dev() is gone; use fsync_bdev().  NOTE: lvm breakage is
deliberate; as soon as struct block_device * is propagated in a reasonable
way by that code fixing will become trivial; until then nothing can be
done.

[mandatory]

	block truncatation on error exit from ->write_begin, and ->direct_IO
moved from generic methods (block_write_begin, cont_write_begin,
nobh_write_begin, blockdev_direct_IO*) to callers.  Take a look at
ext2_write_failed and callers for an example.

[mandatory]

	->truncate is going away.  The whole truncate sequence needs to be
implemented in ->setattr, which is now mandatory for filesystems
implementing on-disk size changes.  Start with a copy of the old inode_setattr
and vmtruncate, and the reorder the vmtruncate + foofs_vmtruncate sequence to
be in order of zeroing blocks using block_truncate_page or similar helpers,
size update and on finally on-disk truncation which should not fail.
inode_change_ok now includes the size checks for ATTR_SIZE and must be called
in the beginning of ->setattr unconditionally.

[mandatory]

	->clear_inode() and ->delete_inode() are gone; ->evict_inode() should
be used instead.  It gets called whenever the inode is evicted, whether it has
remaining links or not.  Caller does *not* evict the pagecache or inode-associated
metadata buffers; getting rid of those is responsibility of method, as it had
been for ->delete_inode().
	->drop_inode() returns int now; it's called on final iput() with inode_lock
held and it returns true if filesystems wants the inode to be dropped.  As before,
generic_drop_inode() is still the default and it's been updated appropriately.
generic_delete_inode() is also alive and it consists simply of return 1.  Note that
all actual eviction work is done by caller after ->drop_inode() returns.
	clear_inode() is gone; use end_writeback() instead.  As before, it must
be called exactly once on each call of ->evict_inode() (as it used to be for
each call of ->delete_inode()).  Unlike before, if you are using inode-associated
metadata buffers (i.e. mark_buffer_dirty_inode()), it's your responsibility to
call invalidate_inode_buffers() before end_writeback().
	No async writeback (and thus no calls of ->write_inode()) will happen
after end_writeback() returns, so actions that should not overlap with ->write_inode()
(e.g. freeing on-disk inode if i_nlink is 0) ought to be done after that call.

	NOTE: checking i_nlink in the beginning of ->write_inode() and bailing out
if it's zero is not *and* *never* *had* *been* enough.  Final unlink() and iput()
may happen while the inode is in the middle of ->write_inode(); e.g. if you blindly
free the on-disk inode, you may end up doing that while ->write_inode() is writing
to it.

---
[mandatory]

	.d_delete() now only advises the dcache as to whether or not to cache
unreferenced dentries, and is now only called when the dentry refcount goes to
0. Even on 0 refcount transition, it must be able to tolerate being called 0,
1, or more times (eg. constant, idempotent).

---
[mandatory]

	.d_compare() calling convention and locking rules are significantly
changed. Read updated documentation in Documentation/filesystems/vfs.txt (and
look at examples of other filesystems) for guidance.