[JFFS2] Prevent ino cache removal for inodes in use
[GitHub/mt8127/android_kernel_alcatel_ttab.git] / fs / jffs2 / build.c
CommitLineData
1da177e4
LT
1/*
2 * JFFS2 -- Journalling Flash File System, Version 2.
3 *
4 * Copyright (C) 2001-2003 Red Hat, Inc.
5 *
6 * Created by David Woodhouse <dwmw2@infradead.org>
7 *
8 * For licensing information, see the file 'LICENCE' in this directory.
9 *
10 * $Id: build.c,v 1.69 2004/12/16 20:22:18 dmarlin Exp $
11 *
12 */
13
14#include <linux/kernel.h>
15#include <linux/sched.h>
16#include <linux/slab.h>
17#include <linux/vmalloc.h>
18#include <linux/mtd/mtd.h>
19#include "nodelist.h"
20
21static void jffs2_build_remove_unlinked_inode(struct jffs2_sb_info *, struct jffs2_inode_cache *, struct jffs2_full_dirent **);
22
23static inline struct jffs2_inode_cache *
24first_inode_chain(int *i, struct jffs2_sb_info *c)
25{
26 for (; *i < INOCACHE_HASHSIZE; (*i)++) {
27 if (c->inocache_list[*i])
28 return c->inocache_list[*i];
29 }
30 return NULL;
31}
32
33static inline struct jffs2_inode_cache *
34next_inode(int *i, struct jffs2_inode_cache *ic, struct jffs2_sb_info *c)
35{
36 /* More in this chain? */
37 if (ic->next)
38 return ic->next;
39 (*i)++;
40 return first_inode_chain(i, c);
41}
42
43#define for_each_inode(i, c, ic) \
44 for (i = 0, ic = first_inode_chain(&i, (c)); \
45 ic; \
46 ic = next_inode(&i, ic, (c)))
47
48
49static inline void jffs2_build_inode_pass1(struct jffs2_sb_info *c, struct jffs2_inode_cache *ic)
50{
51 struct jffs2_full_dirent *fd;
52
53 D1(printk(KERN_DEBUG "jffs2_build_inode building directory inode #%u\n", ic->ino));
54
55 /* For each child, increase nlink */
56 for(fd = ic->scan_dents; fd; fd = fd->next) {
57 struct jffs2_inode_cache *child_ic;
58 if (!fd->ino)
59 continue;
60
61 /* XXX: Can get high latency here with huge directories */
62
63 child_ic = jffs2_get_ino_cache(c, fd->ino);
64 if (!child_ic) {
65 printk(KERN_NOTICE "Eep. Child \"%s\" (ino #%u) of dir ino #%u doesn't exist!\n",
66 fd->name, fd->ino, ic->ino);
67 jffs2_mark_node_obsolete(c, fd->raw);
68 continue;
69 }
70
71 if (child_ic->nlink++ && fd->type == DT_DIR) {
72 printk(KERN_NOTICE "Child dir \"%s\" (ino #%u) of dir ino #%u appears to be a hard link\n", fd->name, fd->ino, ic->ino);
73 if (fd->ino == 1 && ic->ino == 1) {
74 printk(KERN_NOTICE "This is mostly harmless, and probably caused by creating a JFFS2 image\n");
75 printk(KERN_NOTICE "using a buggy version of mkfs.jffs2. Use at least v1.17.\n");
76 }
77 /* What do we do about it? */
78 }
79 D1(printk(KERN_DEBUG "Increased nlink for child \"%s\" (ino #%u)\n", fd->name, fd->ino));
80 /* Can't free them. We might need them in pass 2 */
81 }
82}
83
84/* Scan plan:
85 - Scan physical nodes. Build map of inodes/dirents. Allocate inocaches as we go
86 - Scan directory tree from top down, setting nlink in inocaches
87 - Scan inocaches for inodes with nlink==0
88*/
89static int jffs2_build_filesystem(struct jffs2_sb_info *c)
90{
91 int ret;
92 int i;
93 struct jffs2_inode_cache *ic;
94 struct jffs2_full_dirent *fd;
95 struct jffs2_full_dirent *dead_fds = NULL;
96
97 /* First, scan the medium and build all the inode caches with
98 lists of physical nodes */
99
100 c->flags |= JFFS2_SB_FLAG_MOUNTING;
101 ret = jffs2_scan_medium(c);
102 if (ret)
103 goto exit;
104
105 D1(printk(KERN_DEBUG "Scanned flash completely\n"));
106 D2(jffs2_dump_block_lists(c));
107
108 /* Now scan the directory tree, increasing nlink according to every dirent found. */
109 for_each_inode(i, c, ic) {
110 D1(printk(KERN_DEBUG "Pass 1: ino #%u\n", ic->ino));
111
112 D1(BUG_ON(ic->ino > c->highest_ino));
113
114 if (ic->scan_dents) {
115 jffs2_build_inode_pass1(c, ic);
116 cond_resched();
117 }
118 }
119 c->flags &= ~JFFS2_SB_FLAG_MOUNTING;
120
121 D1(printk(KERN_DEBUG "Pass 1 complete\n"));
122
123 /* Next, scan for inodes with nlink == 0 and remove them. If
124 they were directories, then decrement the nlink of their
125 children too, and repeat the scan. As that's going to be
126 a fairly uncommon occurrence, it's not so evil to do it this
127 way. Recursion bad. */
128 D1(printk(KERN_DEBUG "Pass 2 starting\n"));
129
130 for_each_inode(i, c, ic) {
131 D1(printk(KERN_DEBUG "Pass 2: ino #%u, nlink %d, ic %p, nodes %p\n", ic->ino, ic->nlink, ic, ic->nodes));
132 if (ic->nlink)
133 continue;
134
135 jffs2_build_remove_unlinked_inode(c, ic, &dead_fds);
136 cond_resched();
137 }
138
139 D1(printk(KERN_DEBUG "Pass 2a starting\n"));
140
141 while (dead_fds) {
142 fd = dead_fds;
143 dead_fds = fd->next;
144
145 ic = jffs2_get_ino_cache(c, fd->ino);
146 D1(printk(KERN_DEBUG "Removing dead_fd ino #%u (\"%s\"), ic at %p\n", fd->ino, fd->name, ic));
147
148 if (ic)
149 jffs2_build_remove_unlinked_inode(c, ic, &dead_fds);
150 jffs2_free_full_dirent(fd);
151 }
152
153 D1(printk(KERN_DEBUG "Pass 2 complete\n"));
154
155 /* Finally, we can scan again and free the dirent structs */
156 for_each_inode(i, c, ic) {
157 D1(printk(KERN_DEBUG "Pass 3: ino #%u, ic %p, nodes %p\n", ic->ino, ic, ic->nodes));
158
159 while(ic->scan_dents) {
160 fd = ic->scan_dents;
161 ic->scan_dents = fd->next;
162 jffs2_free_full_dirent(fd);
163 }
164 ic->scan_dents = NULL;
165 cond_resched();
166 }
167 D1(printk(KERN_DEBUG "Pass 3 complete\n"));
168 D2(jffs2_dump_block_lists(c));
169
170 /* Rotate the lists by some number to ensure wear levelling */
171 jffs2_rotate_lists(c);
172
173 ret = 0;
174
175exit:
176 if (ret) {
177 for_each_inode(i, c, ic) {
178 while(ic->scan_dents) {
179 fd = ic->scan_dents;
180 ic->scan_dents = fd->next;
181 jffs2_free_full_dirent(fd);
182 }
183 }
184 }
185
186 return ret;
187}
188
189static void jffs2_build_remove_unlinked_inode(struct jffs2_sb_info *c, struct jffs2_inode_cache *ic, struct jffs2_full_dirent **dead_fds)
190{
191 struct jffs2_raw_node_ref *raw;
192 struct jffs2_full_dirent *fd;
193
194 D1(printk(KERN_DEBUG "JFFS2: Removing ino #%u with nlink == zero.\n", ic->ino));
195
196 raw = ic->nodes;
197 while (raw != (void *)ic) {
198 struct jffs2_raw_node_ref *next = raw->next_in_ino;
199 D1(printk(KERN_DEBUG "obsoleting node at 0x%08x\n", ref_offset(raw)));
200 jffs2_mark_node_obsolete(c, raw);
201 raw = next;
202 }
203
204 if (ic->scan_dents) {
205 int whinged = 0;
206 D1(printk(KERN_DEBUG "Inode #%u was a directory which may have children...\n", ic->ino));
207
208 while(ic->scan_dents) {
209 struct jffs2_inode_cache *child_ic;
210
211 fd = ic->scan_dents;
212 ic->scan_dents = fd->next;
213
214 if (!fd->ino) {
215 /* It's a deletion dirent. Ignore it */
216 D1(printk(KERN_DEBUG "Child \"%s\" is a deletion dirent, skipping...\n", fd->name));
217 jffs2_free_full_dirent(fd);
218 continue;
219 }
220 if (!whinged) {
221 whinged = 1;
222 printk(KERN_NOTICE "Inode #%u was a directory with children - removing those too...\n", ic->ino);
223 }
224
225 D1(printk(KERN_DEBUG "Removing child \"%s\", ino #%u\n",
226 fd->name, fd->ino));
227
228 child_ic = jffs2_get_ino_cache(c, fd->ino);
229 if (!child_ic) {
230 printk(KERN_NOTICE "Cannot remove child \"%s\", ino #%u, because it doesn't exist\n", fd->name, fd->ino);
231 jffs2_free_full_dirent(fd);
232 continue;
233 }
234
235 /* Reduce nlink of the child. If it's now zero, stick it on the
236 dead_fds list to be cleaned up later. Else just free the fd */
237
238 child_ic->nlink--;
239
240 if (!child_ic->nlink) {
241 D1(printk(KERN_DEBUG "Inode #%u (\"%s\") has now got zero nlink. Adding to dead_fds list.\n",
242 fd->ino, fd->name));
243 fd->next = *dead_fds;
244 *dead_fds = fd;
245 } else {
246 D1(printk(KERN_DEBUG "Inode #%u (\"%s\") has now got nlink %d. Ignoring.\n",
247 fd->ino, fd->name, child_ic->nlink));
248 jffs2_free_full_dirent(fd);
249 }
250 }
251 }
252
253 /*
254 We don't delete the inocache from the hash list and free it yet.
255 The erase code will do that, when all the nodes are completely gone.
256 */
257}
258
259static void jffs2_calc_trigger_levels(struct jffs2_sb_info *c)
260{
261 uint32_t size;
262
263 /* Deletion should almost _always_ be allowed. We're fairly
264 buggered once we stop allowing people to delete stuff
265 because there's not enough free space... */
266 c->resv_blocks_deletion = 2;
267
268 /* Be conservative about how much space we need before we allow writes.
269 On top of that which is required for deletia, require an extra 2%
270 of the medium to be available, for overhead caused by nodes being
271 split across blocks, etc. */
272
273 size = c->flash_size / 50; /* 2% of flash size */
274 size += c->nr_blocks * 100; /* And 100 bytes per eraseblock */
275 size += c->sector_size - 1; /* ... and round up */
276
277 c->resv_blocks_write = c->resv_blocks_deletion + (size / c->sector_size);
278
279 /* When do we let the GC thread run in the background */
280
281 c->resv_blocks_gctrigger = c->resv_blocks_write + 1;
282
283 /* When do we allow garbage collection to merge nodes to make
284 long-term progress at the expense of short-term space exhaustion? */
285 c->resv_blocks_gcmerge = c->resv_blocks_deletion + 1;
286
287 /* When do we allow garbage collection to eat from bad blocks rather
288 than actually making progress? */
289 c->resv_blocks_gcbad = 0;//c->resv_blocks_deletion + 2;
290
291 /* If there's less than this amount of dirty space, don't bother
292 trying to GC to make more space. It'll be a fruitless task */
293 c->nospc_dirty_size = c->sector_size + (c->flash_size / 100);
294
295 D1(printk(KERN_DEBUG "JFFS2 trigger levels (size %d KiB, block size %d KiB, %d blocks)\n",
296 c->flash_size / 1024, c->sector_size / 1024, c->nr_blocks));
297 D1(printk(KERN_DEBUG "Blocks required to allow deletion: %d (%d KiB)\n",
298 c->resv_blocks_deletion, c->resv_blocks_deletion*c->sector_size/1024));
299 D1(printk(KERN_DEBUG "Blocks required to allow writes: %d (%d KiB)\n",
300 c->resv_blocks_write, c->resv_blocks_write*c->sector_size/1024));
301 D1(printk(KERN_DEBUG "Blocks required to quiesce GC thread: %d (%d KiB)\n",
302 c->resv_blocks_gctrigger, c->resv_blocks_gctrigger*c->sector_size/1024));
303 D1(printk(KERN_DEBUG "Blocks required to allow GC merges: %d (%d KiB)\n",
304 c->resv_blocks_gcmerge, c->resv_blocks_gcmerge*c->sector_size/1024));
305 D1(printk(KERN_DEBUG "Blocks required to GC bad blocks: %d (%d KiB)\n",
306 c->resv_blocks_gcbad, c->resv_blocks_gcbad*c->sector_size/1024));
307 D1(printk(KERN_DEBUG "Amount of dirty space required to GC: %d bytes\n",
308 c->nospc_dirty_size));
309}
310
311int jffs2_do_mount_fs(struct jffs2_sb_info *c)
312{
313 int i;
314
315 c->free_size = c->flash_size;
316 c->nr_blocks = c->flash_size / c->sector_size;
317 if (c->mtd->flags & MTD_NO_VIRTBLOCKS)
318 c->blocks = vmalloc(sizeof(struct jffs2_eraseblock) * c->nr_blocks);
319 else
320 c->blocks = kmalloc(sizeof(struct jffs2_eraseblock) * c->nr_blocks, GFP_KERNEL);
321 if (!c->blocks)
322 return -ENOMEM;
323 for (i=0; i<c->nr_blocks; i++) {
324 INIT_LIST_HEAD(&c->blocks[i].list);
325 c->blocks[i].offset = i * c->sector_size;
326 c->blocks[i].free_size = c->sector_size;
327 c->blocks[i].dirty_size = 0;
328 c->blocks[i].wasted_size = 0;
329 c->blocks[i].unchecked_size = 0;
330 c->blocks[i].used_size = 0;
331 c->blocks[i].first_node = NULL;
332 c->blocks[i].last_node = NULL;
333 c->blocks[i].bad_count = 0;
334 }
335
336 init_MUTEX(&c->alloc_sem);
337 init_MUTEX(&c->erase_free_sem);
338 init_waitqueue_head(&c->erase_wait);
339 init_waitqueue_head(&c->inocache_wq);
340 spin_lock_init(&c->erase_completion_lock);
341 spin_lock_init(&c->inocache_lock);
342
343 INIT_LIST_HEAD(&c->clean_list);
344 INIT_LIST_HEAD(&c->very_dirty_list);
345 INIT_LIST_HEAD(&c->dirty_list);
346 INIT_LIST_HEAD(&c->erasable_list);
347 INIT_LIST_HEAD(&c->erasing_list);
348 INIT_LIST_HEAD(&c->erase_pending_list);
349 INIT_LIST_HEAD(&c->erasable_pending_wbuf_list);
350 INIT_LIST_HEAD(&c->erase_complete_list);
351 INIT_LIST_HEAD(&c->free_list);
352 INIT_LIST_HEAD(&c->bad_list);
353 INIT_LIST_HEAD(&c->bad_used_list);
354 c->highest_ino = 1;
355
356 if (jffs2_build_filesystem(c)) {
357 D1(printk(KERN_DEBUG "build_fs failed\n"));
358 jffs2_free_ino_caches(c);
359 jffs2_free_raw_node_refs(c);
360 if (c->mtd->flags & MTD_NO_VIRTBLOCKS) {
361 vfree(c->blocks);
362 } else {
363 kfree(c->blocks);
364 }
365 return -EIO;
366 }
367
368 jffs2_calc_trigger_levels(c);
369
370 return 0;
371}