2 * JFFS -- Journaling Flash File System, Linux implementation.
4 * Copyright (C) 1999, 2000 Axis Communications, Inc.
6 * Created by Finn Hakansson <finn@axis.com>.
8 * This is free software; you can redistribute it and/or modify it
9 * under the terms of the GNU General Public License as published by
10 * the Free Software Foundation; either version 2 of the License, or
11 * (at your option) any later version.
13 * $Id: intrep.c,v 1.102 2001/09/23 23:28:36 dwmw2 Exp $
15 * Ported to Linux 2.3.x and MTD:
16 * Copyright (C) 2000 Alexander Larsson (alex@cendio.se), Cendio Systems AB
20 /* This file contains the code for the internal structure of the
21 Journaling Flash File System, JFFS. */
26 * memcpy_to_flash() and memcpy_from_flash() functions.
28 * Implementation of hard links.
30 * Organize the source code in a better way. Against the VFS we could
31 * have jffs_ext.c, and against the block device jffs_int.c.
32 * A better file-internal organization too.
34 * A better checksum algorithm.
36 * Consider endianness stuff. ntohl() etc.
38 * Are we handling the atime, mtime, ctime members of the inode right?
40 * Remove some duplicated code. Take a look at jffs_write_node() and
41 * jffs_rewrite_data() for instance.
43 * Implement more meaning of the nlink member in various data structures.
44 * nlink could be used in conjunction with hard links for instance.
46 * Better memory management. Allocate data structures in larger chunks
49 * If too much meta data is stored, a garbage collect should be issued.
50 * We have experienced problems with too much meta data with for instance
53 * Improve the calls to jffs_ioctl(). We would like to retrieve more
54 * information to be able to debug (or to supervise) JFFS during run-time.
58 #include <linux/config.h>
59 #include <linux/types.h>
60 #include <linux/slab.h>
61 #include <linux/jffs.h>
63 #include <linux/stat.h>
64 #include <linux/pagemap.h>
65 #include <asm/semaphore.h>
66 #include <asm/byteorder.h>
67 #include <linux/smp_lock.h>
68 #include <linux/time.h>
69 #include <linux/ctype.h>
74 long no_jffs_node = 0;
75 static long no_jffs_file = 0;
76 #if defined(JFFS_MEMORY_DEBUG) && JFFS_MEMORY_DEBUG
77 long no_jffs_control = 0;
78 long no_jffs_raw_inode = 0;
79 long no_jffs_node_ref = 0;
81 long no_jffs_fmcontrol = 0;
86 static int jffs_scan_flash(struct jffs_control *c);
87 static int jffs_update_file(struct jffs_file *f, struct jffs_node *node);
88 static int jffs_build_file(struct jffs_file *f);
89 static int jffs_free_file(struct jffs_file *f);
90 static int jffs_free_node_list(struct jffs_file *f);
91 static int jffs_garbage_collect_now(struct jffs_control *c);
92 static int jffs_insert_file_into_hash(struct jffs_file *f);
93 static int jffs_remove_redundant_nodes(struct jffs_file *f);
95 /* Is there enough space on the flash? */
96 static inline int JFFS_ENOUGH_SPACE(struct jffs_control *c, __u32 space)
98 struct jffs_fmcontrol *fmc = c->fmc;
101 if ((fmc->flash_size - (fmc->used_size + fmc->dirty_size))
102 >= fmc->min_free_size + space) {
105 if (fmc->dirty_size < fmc->sector_size)
108 if (jffs_garbage_collect_now(c)) {
109 D1(printk("JFFS_ENOUGH_SPACE: jffs_garbage_collect_now() failed.\n"));
115 #if CONFIG_JFFS_FS_VERBOSE > 0
117 flash_read_u8(struct mtd_info *mtd, loff_t from)
123 res = MTD_READ(mtd, from, 1, &retlen, &ret);
125 printk("Didn't read a byte in flash_read_u8(). Returned %d\n", res);
133 jffs_hexdump(struct mtd_info *mtd, loff_t pos, int size)
141 printk("%ld:", (long) pos);
142 for (j = 0; j < 16; j++) {
143 line[j] = flash_read_u8(mtd, pos++);
145 for (i = 0; i < j; i++) {
147 printk(" %.2x", line[i] & 0xff);
150 printk("%.2x", line[i] & 0xff);
154 /* Print empty space */
155 for (; i < 16; i++) {
165 for (i = 0; i < j; i++) {
166 if (isgraph(line[i])) {
167 printk("%c", line[i]);
178 /* Print the contents of a node. */
180 jffs_print_node(struct jffs_node *n)
182 D(printk("jffs_node: 0x%p\n", n));
184 D(printk(" 0x%08x, /* version */\n", n->version));
185 D(printk(" 0x%08x, /* data_offset */\n", n->data_offset));
186 D(printk(" 0x%08x, /* data_size */\n", n->data_size));
187 D(printk(" 0x%08x, /* removed_size */\n", n->removed_size));
188 D(printk(" 0x%08x, /* fm_offset */\n", n->fm_offset));
189 D(printk(" 0x%02x, /* name_size */\n", n->name_size));
190 D(printk(" 0x%p, /* fm, fm->offset: %u */\n",
191 n->fm, (n->fm ? n->fm->offset : 0)));
192 D(printk(" 0x%p, /* version_prev */\n", n->version_prev));
193 D(printk(" 0x%p, /* version_next */\n", n->version_next));
194 D(printk(" 0x%p, /* range_prev */\n", n->range_prev));
195 D(printk(" 0x%p, /* range_next */\n", n->range_next));
201 /* Print the contents of a raw inode. */
203 jffs_print_raw_inode(struct jffs_raw_inode *raw_inode)
205 D(printk("jffs_raw_inode: inode number: %u\n", raw_inode->ino));
207 D(printk(" 0x%08x, /* magic */\n", raw_inode->magic));
208 D(printk(" 0x%08x, /* ino */\n", raw_inode->ino));
209 D(printk(" 0x%08x, /* pino */\n", raw_inode->pino));
210 D(printk(" 0x%08x, /* version */\n", raw_inode->version));
211 D(printk(" 0x%08x, /* mode */\n", raw_inode->mode));
212 D(printk(" 0x%04x, /* uid */\n", raw_inode->uid));
213 D(printk(" 0x%04x, /* gid */\n", raw_inode->gid));
214 D(printk(" 0x%08x, /* atime */\n", raw_inode->atime));
215 D(printk(" 0x%08x, /* mtime */\n", raw_inode->mtime));
216 D(printk(" 0x%08x, /* ctime */\n", raw_inode->ctime));
217 D(printk(" 0x%08x, /* offset */\n", raw_inode->offset));
218 D(printk(" 0x%08x, /* dsize */\n", raw_inode->dsize));
219 D(printk(" 0x%08x, /* rsize */\n", raw_inode->rsize));
220 D(printk(" 0x%02x, /* nsize */\n", raw_inode->nsize));
221 D(printk(" 0x%02x, /* nlink */\n", raw_inode->nlink));
222 D(printk(" 0x%02x, /* spare */\n",
224 D(printk(" %u, /* rename */\n",
226 D(printk(" %u, /* deleted */\n",
227 raw_inode->deleted));
228 D(printk(" 0x%02x, /* accurate */\n",
229 raw_inode->accurate));
230 D(printk(" 0x%08x, /* dchksum */\n", raw_inode->dchksum));
231 D(printk(" 0x%04x, /* nchksum */\n", raw_inode->nchksum));
232 D(printk(" 0x%04x, /* chksum */\n", raw_inode->chksum));
236 #define flash_safe_acquire(arg)
237 #define flash_safe_release(arg)
241 flash_safe_read(struct mtd_info *mtd, loff_t from,
242 u_char *buf, size_t count)
247 D3(printk(KERN_NOTICE "flash_safe_read(%p, %08x, %p, %08x)\n",
248 mtd, (unsigned int) from, buf, count));
250 res = MTD_READ(mtd, from, count, &retlen, buf);
251 if (retlen != count) {
252 panic("Didn't read all bytes in flash_safe_read(). Returned %d\n", res);
254 return res?res:retlen;
259 flash_read_u32(struct mtd_info *mtd, loff_t from)
265 res = MTD_READ(mtd, from, 4, &retlen, (unsigned char *)&ret);
267 printk("Didn't read all bytes in flash_read_u32(). Returned %d\n", res);
276 flash_safe_write(struct mtd_info *mtd, loff_t to,
277 const u_char *buf, size_t count)
282 D3(printk(KERN_NOTICE "flash_safe_write(%p, %08x, %p, %08x)\n",
283 mtd, (unsigned int) to, buf, count));
285 res = MTD_WRITE(mtd, to, count, &retlen, buf);
286 if (retlen != count) {
287 printk("Didn't write all bytes in flash_safe_write(). Returned %d\n", res);
289 return res?res:retlen;
294 flash_safe_writev(struct mtd_info *mtd, const struct kvec *vecs,
295 unsigned long iovec_cnt, loff_t to)
297 size_t retlen, retlen_a;
301 D3(printk(KERN_NOTICE "flash_safe_writev(%p, %08x, %p)\n",
302 mtd, (unsigned int) to, vecs));
305 res = MTD_WRITEV(mtd, vecs, iovec_cnt, to, &retlen);
306 return res ? res : retlen;
308 /* Not implemented writev. Repeatedly use write - on the not so
309 unreasonable assumption that the mtd driver doesn't care how
310 many write cycles we use. */
314 for (i=0; !res && i<iovec_cnt; i++) {
315 res = MTD_WRITE(mtd, to, vecs[i].iov_len, &retlen_a, vecs[i].iov_base);
316 if (retlen_a != vecs[i].iov_len) {
317 printk("Didn't write all bytes in flash_safe_writev(). Returned %d\n", res);
318 if (i != iovec_cnt-1)
321 /* If res is non-zero, retlen_a is undefined, but we don't
322 care because in that case it's not going to be
328 return res?res:retlen;
333 flash_memset(struct mtd_info *mtd, loff_t to,
334 const u_char c, size_t size)
336 static unsigned char pattern[64];
339 /* fill up pattern */
341 for(i = 0; i < 64; i++)
344 /* write as many 64-byte chunks as we can */
347 flash_safe_write(mtd, to, pattern, 64);
355 flash_safe_write(mtd, to, pattern, size);
362 intrep_erase_callback(struct erase_info *done)
364 wait_queue_head_t *wait_q;
366 wait_q = (wait_queue_head_t *)done->priv;
373 flash_erase_region(struct mtd_info *mtd, loff_t start,
376 struct erase_info *erase;
377 DECLARE_WAITQUEUE(wait, current);
378 wait_queue_head_t wait_q;
380 erase = kmalloc(sizeof(struct erase_info), GFP_KERNEL);
384 init_waitqueue_head(&wait_q);
387 erase->callback = intrep_erase_callback;
390 erase->priv = (u_long)&wait_q;
392 /* FIXME: Use TASK_INTERRUPTIBLE and deal with being interrupted */
393 set_current_state(TASK_UNINTERRUPTIBLE);
394 add_wait_queue(&wait_q, &wait);
396 if (MTD_ERASE(mtd, erase) < 0) {
397 set_current_state(TASK_RUNNING);
398 remove_wait_queue(&wait_q, &wait);
401 printk(KERN_WARNING "flash: erase of region [0x%lx, 0x%lx] "
402 "totally failed\n", (long)start, (long)start + size);
407 schedule(); /* Wait for flash to finish. */
408 remove_wait_queue(&wait_q, &wait);
415 /* This routine calculates checksums in JFFS. */
417 jffs_checksum(const void *data, int size)
420 __u8 *ptr = (__u8 *)data;
424 D3(printk(", result: 0x%08x\n", sum));
430 jffs_checksum_flash(struct mtd_info *mtd, loff_t start, int size, __u32 *result)
437 /* Allocate read buffer */
438 read_buf = (__u8 *) kmalloc (sizeof(__u8) * 4096, GFP_KERNEL);
440 printk(KERN_NOTICE "kmalloc failed in jffs_checksum_flash()\n");
443 /* Loop until checksum done */
445 /* Get amount of data to read */
451 /* Perform flash read */
452 D3(printk(KERN_NOTICE "jffs_checksum_flash\n"));
453 flash_safe_read(mtd, ptr, &read_buf[0], length);
455 /* Compute checksum */
456 for (i=0; i < length ; i++)
459 /* Update pointer and size */
464 /* Free read buffer */
468 D3(printk("checksum result: 0x%08x\n", sum));
473 static __inline__ void jffs_fm_write_lock(struct jffs_fmcontrol *fmc)
475 // down(&fmc->wlock);
478 static __inline__ void jffs_fm_write_unlock(struct jffs_fmcontrol *fmc)
484 /* Create and initialize a new struct jffs_file. */
485 static struct jffs_file *
486 jffs_create_file(struct jffs_control *c,
487 const struct jffs_raw_inode *raw_inode)
491 if (!(f = (struct jffs_file *)kmalloc(sizeof(struct jffs_file),
493 D(printk("jffs_create_file(): Failed!\n"));
497 memset(f, 0, sizeof(struct jffs_file));
498 f->ino = raw_inode->ino;
499 f->pino = raw_inode->pino;
500 f->nlink = raw_inode->nlink;
501 f->deleted = raw_inode->deleted;
508 /* Build a control block for the file system. */
509 static struct jffs_control *
510 jffs_create_control(struct super_block *sb)
512 struct jffs_control *c;
513 register int s = sizeof(struct jffs_control);
517 D2(printk("jffs_create_control()\n"));
519 if (!(c = (struct jffs_control *)kmalloc(s, GFP_KERNEL))) {
522 DJM(no_jffs_control++);
525 c->hash_len = JFFS_HASH_SIZE;
526 s = sizeof(struct list_head) * c->hash_len;
527 if (!(c->hash = (struct list_head *)kmalloc(s, GFP_KERNEL))) {
531 for (i = 0; i < c->hash_len; i++)
532 INIT_LIST_HEAD(&c->hash[i]);
533 if (!(c->fmc = jffs_build_begin(c, MINOR(sb->s_dev)))) {
536 c->next_ino = JFFS_MIN_INO + 1;
537 c->delete_list = (struct jffs_delete_list *) 0;
544 DJM(no_jffs_control--);
545 D(t = t ? t : "c->hash");
547 D(t = t ? t : "control");
548 D(printk("jffs_create_control(): Allocation failed: (%s)\n", t));
549 return (struct jffs_control *)0;
553 /* Clean up all data structures associated with the file system. */
555 jffs_cleanup_control(struct jffs_control *c)
557 D2(printk("jffs_cleanup_control()\n"));
560 D(printk("jffs_cleanup_control(): c == NULL !!!\n"));
564 while (c->delete_list) {
565 struct jffs_delete_list *delete_list_element;
566 delete_list_element = c->delete_list;
567 c->delete_list = c->delete_list->next;
568 kfree(delete_list_element);
571 /* Free all files and nodes. */
573 jffs_foreach_file(c, jffs_free_node_list);
574 jffs_foreach_file(c, jffs_free_file);
578 jffs_cleanup_fmcontrol(c->fmc);
580 DJM(no_jffs_control--);
581 D3(printk("jffs_cleanup_control(): Leaving...\n"));
585 /* This function adds a virtual root node to the in-RAM representation.
586 Called by jffs_build_fs(). */
588 jffs_add_virtual_root(struct jffs_control *c)
590 struct jffs_file *root;
591 struct jffs_node *node;
593 D2(printk("jffs_add_virtual_root(): "
594 "Creating a virtual root directory.\n"));
596 if (!(root = (struct jffs_file *)kmalloc(sizeof(struct jffs_file),
601 if (!(node = jffs_alloc_node())) {
607 memset(node, 0, sizeof(struct jffs_node));
608 node->ino = JFFS_MIN_INO;
609 memset(root, 0, sizeof(struct jffs_file));
610 root->ino = JFFS_MIN_INO;
611 root->mode = S_IFDIR | S_IRWXU | S_IRGRP
612 | S_IXGRP | S_IROTH | S_IXOTH;
613 root->atime = root->mtime = root->ctime = get_seconds();
616 root->version_head = root->version_tail = node;
617 jffs_insert_file_into_hash(root);
622 /* This is where the file system is built and initialized. */
624 jffs_build_fs(struct super_block *sb)
626 struct jffs_control *c;
629 D2(printk("jffs_build_fs()\n"));
631 if (!(c = jffs_create_control(sb))) {
636 if ((err = jffs_scan_flash(c)) < 0) {
638 /* scan_flash() wants us to try once more. A flipping
639 bits sector was detect in the middle of the scan flash.
640 Clean up old allocated memory before going in.
642 D1(printk("jffs_build_fs: Cleaning up all control structures,"
643 " reallocating them and trying mount again.\n"));
644 jffs_cleanup_control(c);
645 if (!(c = jffs_create_control(sb))) {
651 if ((err = jffs_scan_flash(c)) < 0) {
652 goto jffs_build_fs_fail;
655 goto jffs_build_fs_fail;
659 /* Add a virtual root node if no one exists. */
660 if (!jffs_find_file(c, JFFS_MIN_INO)) {
661 if ((err = jffs_add_virtual_root(c)) < 0) {
662 goto jffs_build_fs_fail;
666 while (c->delete_list) {
668 struct jffs_delete_list *delete_list_element;
670 if ((f = jffs_find_file(c, c->delete_list->ino))) {
673 delete_list_element = c->delete_list;
674 c->delete_list = c->delete_list->next;
675 kfree(delete_list_element);
678 /* Remove deleted nodes. */
679 if ((err = jffs_foreach_file(c, jffs_possibly_delete_file)) < 0) {
680 printk(KERN_ERR "JFFS: Failed to remove deleted nodes.\n");
681 goto jffs_build_fs_fail;
683 /* Remove redundant nodes. (We are not interested in the
684 return value in this case.) */
685 jffs_foreach_file(c, jffs_remove_redundant_nodes);
686 /* Try to build a tree from all the nodes. */
687 if ((err = jffs_foreach_file(c, jffs_insert_file_into_tree)) < 0) {
688 printk("JFFS: Failed to build tree.\n");
689 goto jffs_build_fs_fail;
691 /* Compute the sizes of all files in the filesystem. Adjust if
693 if ((err = jffs_foreach_file(c, jffs_build_file)) < 0) {
694 printk("JFFS: Failed to build file system.\n");
695 goto jffs_build_fs_fail;
697 sb->s_fs_info = (void *)c;
700 D1(jffs_print_hash_table(c));
701 D1(jffs_print_tree(c->root, 0));
706 jffs_cleanup_control(c);
708 } /* jffs_build_fs() */
712 This checks for sectors that were being erased in their previous
713 lifetimes and for some reason or the other (power fail etc.),
714 the erase cycles never completed.
715 As the flash array would have reverted back to read status,
716 these sectors are detected by the symptom of the "flipping bits",
717 i.e. bits being read back differently from the same location in
718 flash if read multiple times.
719 The only solution to this is to re-erase the entire
721 Unfortunately detecting "flipping bits" is not a simple exercise
722 as a bit may be read back at 1 or 0 depending on the alignment
723 of the stars in the universe.
724 The level of confidence is in direct proportion to the number of
725 scans done. By power fail testing I (Vipin) have been able to
726 proove that reading twice is not enough.
727 Maybe 4 times? Change NUM_REREADS to a higher number if you want
728 a (even) higher degree of confidence in your mount process.
729 A higher number would of course slow down your mount.
731 static int check_partly_erased_sectors(struct jffs_fmcontrol *fmc){
733 #define NUM_REREADS 4 /* see note above */
734 #define READ_AHEAD_BYTES 4096 /* must be a multiple of 4,
735 usually set to kernel page size */
746 loff_t end = fmc->flash_size;
749 /* Allocate read buffers */
750 read_buf1 = (__u8 *) kmalloc (sizeof(__u8) * READ_AHEAD_BYTES, GFP_KERNEL);
754 read_buf2 = (__u8 *) kmalloc (sizeof(__u8) * READ_AHEAD_BYTES, GFP_KERNEL);
763 D1(printk("check_partly_erased_sector():checking sector which contains"
764 " offset 0x%x for flipping bits..\n", (__u32)pos));
766 retlen = flash_safe_read(fmc->mtd, pos,
767 &read_buf1[0], READ_AHEAD_BYTES);
770 for(cnt = 0; cnt < NUM_REREADS; cnt++){
771 (void)flash_safe_read(fmc->mtd, pos,
772 &read_buf2[0], READ_AHEAD_BYTES);
774 for (i=0 ; i < retlen ; i+=4) {
775 /* buffers MUST match, double word for word! */
776 if(*((__u32 *) &read_buf1[i]) !=
777 *((__u32 *) &read_buf2[i])
779 /* flipping bits detected, time to erase sector */
780 /* This will help us log some statistics etc. */
781 D1(printk("Flipping bits detected in re-read round:%i of %i\n",
783 D1(printk("check_partly_erased_sectors:flipping bits detected"
784 " @offset:0x%x(0x%x!=0x%x)\n",
785 (__u32)pos+i, *((__u32 *) &read_buf1[i]),
786 *((__u32 *) &read_buf2[i])));
788 /* calculate start of present sector */
789 offset = (((__u32)pos+i)/(__u32)fmc->sector_size) * (__u32)fmc->sector_size;
791 D1(printk("check_partly_erased_sector():erasing sector starting 0x%x.\n",
794 if (flash_erase_region(fmc->mtd,
795 offset, fmc->sector_size) < 0) {
796 printk(KERN_ERR "JFFS: Erase of flash failed. "
797 "offset = %u, erase_size = %d\n",
798 offset , fmc->sector_size);
804 D1(printk("JFFS: Erase of flash sector @0x%x successful.\n",
806 /* skip ahead to the next sector */
807 pos = (((__u32)pos+i)/(__u32)fmc->sector_size) * (__u32)fmc->sector_size;
808 pos += fmc->sector_size;
814 pos += READ_AHEAD_BYTES;
821 D2(printk("check_partly_erased_sector():Done checking all sectors till offset 0x%x for flipping bits.\n",
826 }/* end check_partly_erased_sectors() */
830 /* Scan the whole flash memory in order to find all nodes in the
833 jffs_scan_flash(struct jffs_control *c)
835 char name[JFFS_MAX_NAME_LEN + 2];
836 struct jffs_raw_inode raw_inode;
837 struct jffs_node *node = NULL;
838 struct jffs_fmcontrol *fmc = c->fmc;
846 loff_t end = fmc->flash_size;
851 __u32 free_chunk_size1;
852 __u32 free_chunk_size2;
855 #define NUMFREEALLOWED 2 /* 2 chunks of at least erase size space allowed */
856 int num_free_space = 0; /* Flag err if more than TWO
857 free blocks found. This is NOT allowed
858 by the current jffs design.
860 int num_free_spc_not_accp = 0; /* For debugging purposed keep count
861 of how much free space was rejected and
865 D1(printk("jffs_scan_flash(): start pos = 0x%lx, end = 0x%lx\n",
866 (long)pos, (long)end));
868 flash_safe_acquire(fmc->mtd);
871 check and make sure that any sector does not suffer
872 from the "partly erased, bit flipping syndrome" (TM Vipin :)
873 If so, offending sectors will be erased.
875 if(check_partly_erased_sectors(fmc) < 0){
877 flash_safe_release(fmc->mtd);
878 return -EIO; /* bad, bad, bad error. Cannot continue.*/
881 /* Allocate read buffer */
882 read_buf = (__u8 *) kmalloc (sizeof(__u8) * 4096, GFP_KERNEL);
884 flash_safe_release(fmc->mtd);
888 /* Start the scan. */
892 /* Remember the position from where we started this scan. */
895 switch (flash_read_u32(fmc->mtd, pos)) {
896 case JFFS_EMPTY_BITMASK:
897 /* We have found 0xffffffff at this position. We have to
898 scan the rest of the flash till the end or till
899 something else than 0xffffffff is found.
900 Keep going till we do not find JFFS_EMPTY_BITMASK
903 D1(printk("jffs_scan_flash(): 0xffffffff at pos 0x%lx.\n",
908 len = end - pos < 4096 ? end - pos : 4096;
910 retlen = flash_safe_read(fmc->mtd, pos,
915 for (i=0 ; i < retlen ; i+=4, pos += 4) {
916 if(*((__u32 *) &read_buf[i]) !=
926 D1(printk("jffs_scan_flash():0xffffffff ended at pos 0x%lx.\n",
929 /* If some free space ends in the middle of a sector,
930 treat it as dirty rather than clean.
931 This is to handle the case where one thread
932 allocated space for a node, but didn't get to
933 actually _write_ it before power was lost, leaving
934 a gap in the log. Shifting all node writes into
935 a single kernel thread will fix the original problem.
937 if ((__u32) pos % fmc->sector_size) {
938 /* If there was free space in previous
939 sectors, don't mark that dirty too -
940 only from the beginning of this sector
944 test_start = pos & ~(fmc->sector_size-1); /* end of last sector */
946 if (start < test_start) {
948 /* free space started in the previous sector! */
950 if((num_free_space < NUMFREEALLOWED) &&
951 ((unsigned int)(test_start - start) >= fmc->sector_size)){
954 Count it in if we are still under NUMFREEALLOWED *and* it is
955 at least 1 erase sector in length. This will keep us from
956 picking any little ole' space as "free".
959 D1(printk("Reducing end of free space to 0x%x from 0x%x\n",
960 (unsigned int)test_start, (unsigned int)pos));
962 D1(printk("Free space accepted: Starting 0x%x for 0x%x bytes\n",
963 (unsigned int) start,
964 (unsigned int)(test_start - start)));
966 /* below, space from "start" to "pos" will be marked dirty. */
969 /* Being in here means that we have found at least an entire
970 erase sector size of free space ending on a sector boundary.
971 Keep track of free spaces accepted.
975 num_free_spc_not_accp++;
976 D1(printk("Free space (#%i) found but *Not* accepted: Starting"
977 " 0x%x for 0x%x bytes\n",
978 num_free_spc_not_accp, (unsigned int)start,
979 (unsigned int)((unsigned int)(pos & ~(fmc->sector_size-1)) - (unsigned int)start)));
984 if((((__u32)(pos - start)) != 0)){
986 D1(printk("Dirty space: Starting 0x%x for 0x%x bytes\n",
987 (unsigned int) start, (unsigned int) (pos - start)));
988 jffs_fmalloced(fmc, (__u32) start,
989 (__u32) (pos - start), NULL);
991 /* "Flipping bits" detected. This means that our scan for them
992 did not catch this offset. See check_partly_erased_sectors() for
996 D1(printk("jffs_scan_flash():wants to allocate dirty flash "
997 "space for 0 bytes.\n"));
998 D1(printk("jffs_scan_flash(): Flipping bits! We will free "
999 "all allocated memory, erase this sector and remount\n"));
1001 /* calculate start of present sector */
1002 offset = (((__u32)pos)/(__u32)fmc->sector_size) * (__u32)fmc->sector_size;
1004 D1(printk("jffs_scan_flash():erasing sector starting 0x%x.\n",
1007 if (flash_erase_region(fmc->mtd,
1008 offset, fmc->sector_size) < 0) {
1009 printk(KERN_ERR "JFFS: Erase of flash failed. "
1010 "offset = %u, erase_size = %d\n",
1011 offset , fmc->sector_size);
1013 flash_safe_release(fmc->mtd);
1015 return -1; /* bad, bad, bad! */
1018 flash_safe_release(fmc->mtd);
1021 return -EAGAIN; /* erased offending sector. Try mount one more time please. */
1024 /* Being in here means that we have found free space that ends on an erase sector
1026 Count it in if we are still under NUMFREEALLOWED *and* it is at least 1 erase
1027 sector in length. This will keep us from picking any little ole' space as "free".
1029 if((num_free_space < NUMFREEALLOWED) &&
1030 ((unsigned int)(pos - start) >= fmc->sector_size)){
1031 /* We really don't do anything to mark space as free, except *not*
1032 mark it dirty and just advance the "pos" location pointer.
1033 It will automatically be picked up as free space.
1036 D1(printk("Free space accepted: Starting 0x%x for 0x%x bytes\n",
1037 (unsigned int) start, (unsigned int) (pos - start)));
1039 num_free_spc_not_accp++;
1040 D1(printk("Free space (#%i) found but *Not* accepted: Starting "
1041 "0x%x for 0x%x bytes\n", num_free_spc_not_accp,
1042 (unsigned int) start,
1043 (unsigned int) (pos - start)));
1045 /* Mark this space as dirty. We already have our free space. */
1046 D1(printk("Dirty space: Starting 0x%x for 0x%x bytes\n",
1047 (unsigned int) start, (unsigned int) (pos - start)));
1048 jffs_fmalloced(fmc, (__u32) start,
1049 (__u32) (pos - start), NULL);
1053 if(num_free_space > NUMFREEALLOWED){
1054 printk(KERN_WARNING "jffs_scan_flash(): Found free space "
1055 "number %i. Only %i free space is allowed.\n",
1056 num_free_space, NUMFREEALLOWED);
1060 case JFFS_DIRTY_BITMASK:
1061 /* We have found 0x00000000 at this position. Scan as far
1062 as possible to find out how much is dirty. */
1063 D1(printk("jffs_scan_flash(): 0x00000000 at pos 0x%lx.\n",
1066 && JFFS_DIRTY_BITMASK == flash_read_u32(fmc->mtd, pos);
1068 D1(printk("jffs_scan_flash(): 0x00 ended at "
1069 "pos 0x%lx.\n", (long)pos));
1070 jffs_fmalloced(fmc, (__u32) start,
1071 (__u32) (pos - start), NULL);
1074 case JFFS_MAGIC_BITMASK:
1075 /* We have probably found a new raw inode. */
1080 /* We're f*cked. This is not solved yet. We have
1081 to scan for the magic pattern. */
1082 D1(printk("*************** Dirty flash memory or "
1084 "hexdump(pos = 0x%lx, len = 128):\n",
1086 D1(jffs_hexdump(fmc->mtd, pos, 128));
1088 for (pos += 4; pos < end; pos += 4) {
1089 switch (flash_read_u32(fmc->mtd, pos)) {
1090 case JFFS_MAGIC_BITMASK:
1091 case JFFS_EMPTY_BITMASK:
1092 /* handle these in the main switch() loop */
1101 /* First, mark as dirty the region
1102 which really does contain crap. */
1103 jffs_fmalloced(fmc, (__u32) start,
1104 (__u32) (pos - start),
1110 /* We have found the beginning of an inode. Create a
1111 node for it unless there already is one available. */
1113 if (!(node = jffs_alloc_node())) {
1114 /* Free read buffer */
1117 /* Release the flash device */
1118 flash_safe_release(fmc->mtd);
1122 DJM(no_jffs_node++);
1125 /* Read the next raw inode. */
1127 flash_safe_read(fmc->mtd, pos, (u_char *) &raw_inode,
1128 sizeof(struct jffs_raw_inode));
1130 /* When we compute the checksum for the inode, we never
1131 count the 'accurate' or the 'checksum' fields. */
1132 tmp_accurate = raw_inode.accurate;
1133 tmp_chksum = raw_inode.chksum;
1134 raw_inode.accurate = 0;
1135 raw_inode.chksum = 0;
1136 checksum = jffs_checksum(&raw_inode,
1137 sizeof(struct jffs_raw_inode));
1138 raw_inode.accurate = tmp_accurate;
1139 raw_inode.chksum = tmp_chksum;
1141 D3(printk("*** We have found this raw inode at pos 0x%lx "
1142 "on the flash:\n", (long)pos));
1143 D3(jffs_print_raw_inode(&raw_inode));
1145 if (checksum != raw_inode.chksum) {
1146 D1(printk("jffs_scan_flash(): Bad checksum: "
1148 "raw_inode.chksum = %u\n",
1149 checksum, raw_inode.chksum));
1150 pos += sizeof(struct jffs_raw_inode);
1151 jffs_fmalloced(fmc, (__u32) start,
1152 (__u32) (pos - start), NULL);
1153 /* Reuse this unused struct jffs_node. */
1157 /* Check the raw inode read so far. Start with the
1158 maximum length of the filename. */
1159 if (raw_inode.nsize > JFFS_MAX_NAME_LEN) {
1160 printk(KERN_WARNING "jffs_scan_flash: Found a "
1161 "JFFS node with name too large\n");
1165 if (raw_inode.rename && raw_inode.dsize != sizeof(__u32)) {
1166 printk(KERN_WARNING "jffs_scan_flash: Found a "
1167 "rename node with dsize %u.\n",
1169 jffs_print_raw_inode(&raw_inode);
1173 /* The node's data segment should not exceed a
1175 if (raw_inode.dsize > fmc->max_chunk_size) {
1176 printk(KERN_WARNING "jffs_scan_flash: Found a "
1177 "JFFS node with dsize (0x%x) > max_chunk_size (0x%x)\n",
1178 raw_inode.dsize, fmc->max_chunk_size);
1182 pos += sizeof(struct jffs_raw_inode);
1184 /* This shouldn't be necessary because a node that
1185 violates the flash boundaries shouldn't be written
1186 in the first place. */
1191 /* Read the name. */
1193 if (raw_inode.nsize) {
1194 flash_safe_read(fmc->mtd, pos, name, raw_inode.nsize);
1195 name[raw_inode.nsize] = '\0';
1196 pos += raw_inode.nsize
1197 + JFFS_GET_PAD_BYTES(raw_inode.nsize);
1198 D3(printk("name == \"%s\"\n", name));
1199 checksum = jffs_checksum(name, raw_inode.nsize);
1200 if (checksum != raw_inode.nchksum) {
1201 D1(printk("jffs_scan_flash(): Bad checksum: "
1203 "raw_inode.nchksum = %u\n",
1204 checksum, raw_inode.nchksum));
1205 jffs_fmalloced(fmc, (__u32) start,
1206 (__u32) (pos - start), NULL);
1207 /* Reuse this unused struct jffs_node. */
1215 /* Read the data, if it exists, in order to be sure it
1216 matches the checksum. */
1217 if (raw_inode.dsize) {
1218 if (raw_inode.rename) {
1219 deleted_file = flash_read_u32(fmc->mtd, pos);
1221 if (jffs_checksum_flash(fmc->mtd, pos, raw_inode.dsize, &checksum)) {
1222 printk("jffs_checksum_flash() failed to calculate a checksum\n");
1223 jffs_fmalloced(fmc, (__u32) start,
1224 (__u32) (pos - start), NULL);
1225 /* Reuse this unused struct jffs_node. */
1228 pos += raw_inode.dsize
1229 + JFFS_GET_PAD_BYTES(raw_inode.dsize);
1231 if (checksum != raw_inode.dchksum) {
1232 D1(printk("jffs_scan_flash(): Bad checksum: "
1234 "raw_inode.dchksum = %u\n",
1235 checksum, raw_inode.dchksum));
1236 jffs_fmalloced(fmc, (__u32) start,
1237 (__u32) (pos - start), NULL);
1238 /* Reuse this unused struct jffs_node. */
1245 /* Remember the highest inode number in the whole file
1246 system. This information will be used when assigning
1247 new files new inode numbers. */
1248 if (c->next_ino <= raw_inode.ino) {
1249 c->next_ino = raw_inode.ino + 1;
1252 if (raw_inode.accurate) {
1254 node->data_offset = raw_inode.offset;
1255 node->data_size = raw_inode.dsize;
1256 node->removed_size = raw_inode.rsize;
1257 /* Compute the offset to the actual data in the
1260 = sizeof(struct jffs_raw_inode)
1262 + JFFS_GET_PAD_BYTES(raw_inode.nsize);
1263 node->fm = jffs_fmalloced(fmc, (__u32) start,
1264 (__u32) (pos - start),
1267 D(printk("jffs_scan_flash(): !node->fm\n"));
1268 jffs_free_node(node);
1269 DJM(no_jffs_node--);
1271 /* Free read buffer */
1274 /* Release the flash device */
1275 flash_safe_release(fmc->mtd);
1279 if ((err = jffs_insert_node(c, NULL, &raw_inode,
1281 printk("JFFS: Failed to handle raw inode. "
1282 "(err = %d)\n", err);
1285 if (raw_inode.rename) {
1286 struct jffs_delete_list *dl
1287 = (struct jffs_delete_list *)
1288 kmalloc(sizeof(struct jffs_delete_list),
1291 D(printk("jffs_scan_flash: !dl\n"));
1292 jffs_free_node(node);
1293 DJM(no_jffs_node--);
1295 /* Release the flash device */
1296 flash_safe_release(fmc->flash_part);
1298 /* Free read buffer */
1303 dl->ino = deleted_file;
1304 dl->next = c->delete_list;
1305 c->delete_list = dl;
1306 node->data_size = 0;
1308 D3(jffs_print_node(node));
1309 node = NULL; /* Don't free the node! */
1312 jffs_fmalloced(fmc, (__u32) start,
1313 (__u32) (pos - start), NULL);
1314 D3(printk("jffs_scan_flash(): Just found an obsolete "
1315 "raw_inode. Continuing the scan...\n"));
1316 /* Reuse this unused struct jffs_node. */
1321 jffs_free_node(node);
1322 DJM(no_jffs_node--);
1324 jffs_build_end(fmc);
1326 /* Free read buffer */
1329 if(!num_free_space){
1330 printk(KERN_WARNING "jffs_scan_flash(): Did not find even a single "
1331 "chunk of free space. This is BAD!\n");
1335 D3(printk("jffs_scan_flash(): Leaving...\n"));
1336 flash_safe_release(fmc->mtd);
1338 /* This is to trap the "free size accounting screwed error. */
1339 free_chunk_size1 = jffs_free_size1(fmc);
1340 free_chunk_size2 = jffs_free_size2(fmc);
1342 if (free_chunk_size1 + free_chunk_size2 != fmc->free_size) {
1344 printk(KERN_WARNING "jffs_scan_falsh():Free size accounting screwed\n");
1345 printk(KERN_WARNING "jfffs_scan_flash():free_chunk_size1 == 0x%x, "
1346 "free_chunk_size2 == 0x%x, fmc->free_size == 0x%x\n",
1347 free_chunk_size1, free_chunk_size2, fmc->free_size);
1349 return -1; /* Do NOT mount f/s so that we can inspect what happened.
1350 Mounting this screwed up f/s will screw us up anyway.
1354 return 0; /* as far as we are concerned, we are happy! */
1355 } /* jffs_scan_flash() */
1358 /* Insert any kind of node into the file system. Take care of data
1359 insertions and deletions. Also remove redundant information. The
1360 memory allocated for the `name' is regarded as "given away" in the
1361 caller's perspective. */
1363 jffs_insert_node(struct jffs_control *c, struct jffs_file *f,
1364 const struct jffs_raw_inode *raw_inode,
1365 const char *name, struct jffs_node *node)
1367 int update_name = 0;
1368 int insert_into_tree = 0;
1370 D2(printk("jffs_insert_node(): ino = %u, version = %u, "
1371 "name = \"%s\", deleted = %d\n",
1372 raw_inode->ino, raw_inode->version,
1373 ((name && *name) ? name : ""), raw_inode->deleted));
1375 /* If there doesn't exist an associated jffs_file, then
1376 create, initialize and insert one into the file system. */
1377 if (!f && !(f = jffs_find_file(c, raw_inode->ino))) {
1378 if (!(f = jffs_create_file(c, raw_inode))) {
1381 jffs_insert_file_into_hash(f);
1382 insert_into_tree = 1;
1384 node->ino = raw_inode->ino;
1385 node->version = raw_inode->version;
1386 node->data_size = raw_inode->dsize;
1387 node->fm_offset = sizeof(struct jffs_raw_inode) + raw_inode->nsize
1388 + JFFS_GET_PAD_BYTES(raw_inode->nsize);
1389 node->name_size = raw_inode->nsize;
1391 /* Now insert the node at the correct position into the file's
1393 if (!f->version_head) {
1394 /* This is the first node. */
1395 f->version_head = node;
1396 f->version_tail = node;
1397 node->version_prev = NULL;
1398 node->version_next = NULL;
1399 f->highest_version = node->version;
1401 f->mode = raw_inode->mode;
1402 f->uid = raw_inode->uid;
1403 f->gid = raw_inode->gid;
1404 f->atime = raw_inode->atime;
1405 f->mtime = raw_inode->mtime;
1406 f->ctime = raw_inode->ctime;
1408 else if ((f->highest_version < node->version)
1409 || (node->version == 0)) {
1410 /* Insert at the end of the list. I.e. this node is the
1411 newest one so far. */
1412 node->version_prev = f->version_tail;
1413 node->version_next = NULL;
1414 f->version_tail->version_next = node;
1415 f->version_tail = node;
1416 f->highest_version = node->version;
1418 f->pino = raw_inode->pino;
1419 f->mode = raw_inode->mode;
1420 f->uid = raw_inode->uid;
1421 f->gid = raw_inode->gid;
1422 f->atime = raw_inode->atime;
1423 f->mtime = raw_inode->mtime;
1424 f->ctime = raw_inode->ctime;
1426 else if (f->version_head->version > node->version) {
1427 /* Insert at the bottom of the list. */
1428 node->version_prev = NULL;
1429 node->version_next = f->version_head;
1430 f->version_head->version_prev = node;
1431 f->version_head = node;
1437 struct jffs_node *n;
1439 /* Search for the insertion position starting from
1440 the tail (newest node). */
1441 for (n = f->version_tail; n; n = n->version_prev) {
1442 if (n->version < node->version) {
1443 node->version_prev = n;
1444 node->version_next = n->version_next;
1445 node->version_next->version_prev = node;
1446 n->version_next = node;
1458 /* Deletion is irreversible. If any 'deleted' node is ever
1459 written, the file is deleted */
1460 if (raw_inode->deleted)
1461 f->deleted = raw_inode->deleted;
1463 /* Perhaps update the name. */
1464 if (raw_inode->nsize && update_name && name && *name && (name != f->name)) {
1469 if (!(f->name = (char *) kmalloc(raw_inode->nsize + 1,
1474 memcpy(f->name, name, raw_inode->nsize);
1475 f->name[raw_inode->nsize] = '\0';
1476 f->nsize = raw_inode->nsize;
1477 D3(printk("jffs_insert_node(): Updated the name of "
1478 "the file to \"%s\".\n", name));
1481 if (!c->building_fs) {
1482 D3(printk("jffs_insert_node(): ---------------------------"
1483 "------------------------------------------- 1\n"));
1484 if (insert_into_tree) {
1485 jffs_insert_file_into_tree(f);
1487 /* Once upon a time, we would call jffs_possibly_delete_file()
1488 here. That causes an oops if someone's still got the file
1489 open, so now we only do it in jffs_delete_inode()
1492 if (node->data_size || node->removed_size) {
1493 jffs_update_file(f, node);
1495 jffs_remove_redundant_nodes(f);
1497 jffs_garbage_collect_trigger(c);
1499 D3(printk("jffs_insert_node(): ---------------------------"
1500 "------------------------------------------- 2\n"));
1504 } /* jffs_insert_node() */
1507 /* Unlink a jffs_node from the version list it is in. */
1509 jffs_unlink_node_from_version_list(struct jffs_file *f,
1510 struct jffs_node *node)
1512 if (node->version_prev) {
1513 node->version_prev->version_next = node->version_next;
1515 f->version_head = node->version_next;
1517 if (node->version_next) {
1518 node->version_next->version_prev = node->version_prev;
1520 f->version_tail = node->version_prev;
1525 /* Unlink a jffs_node from the range list it is in. */
1527 jffs_unlink_node_from_range_list(struct jffs_file *f, struct jffs_node *node)
1529 if (node->range_prev) {
1530 node->range_prev->range_next = node->range_next;
1533 f->range_head = node->range_next;
1535 if (node->range_next) {
1536 node->range_next->range_prev = node->range_prev;
1539 f->range_tail = node->range_prev;
1544 /* Function used by jffs_remove_redundant_nodes() below. This function
1545 classifies what kind of information a node adds to a file. */
1547 jffs_classify_node(struct jffs_node *node)
1549 __u8 mod_type = JFFS_MODIFY_INODE;
1551 if (node->name_size) {
1552 mod_type |= JFFS_MODIFY_NAME;
1554 if (node->data_size || node->removed_size) {
1555 mod_type |= JFFS_MODIFY_DATA;
1561 /* Remove redundant nodes from a file. Mark the on-flash memory
1564 jffs_remove_redundant_nodes(struct jffs_file *f)
1566 struct jffs_node *newest_node;
1567 struct jffs_node *cur;
1568 struct jffs_node *prev;
1571 __u8 node_with_name_later = 0;
1573 if (!(newest_node = f->version_tail)) {
1577 /* What does the `newest_node' modify? */
1578 newest_type = jffs_classify_node(newest_node);
1579 node_with_name_later = newest_type & JFFS_MODIFY_NAME;
1581 D3(printk("jffs_remove_redundant_nodes(): ino: %u, name: \"%s\", "
1582 "newest_type: %u\n", f->ino, (f->name ? f->name : ""),
1585 /* Traverse the file's nodes and determine which of them that are
1586 superfluous. Yeah, this might look very complex at first
1587 glance but it is actually very simple. */
1588 for (cur = newest_node->version_prev; cur; cur = prev) {
1589 prev = cur->version_prev;
1590 mod_type = jffs_classify_node(cur);
1591 if ((mod_type <= JFFS_MODIFY_INODE)
1592 || ((newest_type & JFFS_MODIFY_NAME)
1594 <= (JFFS_MODIFY_INODE + JFFS_MODIFY_NAME)))
1595 || (cur->data_size == 0 && cur->removed_size
1596 && !cur->version_prev && node_with_name_later)) {
1597 /* Yes, this node is redundant. Remove it. */
1598 D2(printk("jffs_remove_redundant_nodes(): "
1599 "Removing node: ino: %u, version: %u, "
1600 "mod_type: %u\n", cur->ino, cur->version,
1602 jffs_unlink_node_from_version_list(f, cur);
1603 jffs_fmfree(f->c->fmc, cur->fm, cur);
1604 jffs_free_node(cur);
1605 DJM(no_jffs_node--);
1608 node_with_name_later |= (mod_type & JFFS_MODIFY_NAME);
1616 /* Insert a file into the hash table. */
1618 jffs_insert_file_into_hash(struct jffs_file *f)
1620 int i = f->ino % f->c->hash_len;
1622 D3(printk("jffs_insert_file_into_hash(): f->ino: %u\n", f->ino));
1624 list_add(&f->hash, &f->c->hash[i]);
1629 /* Insert a file into the file system tree. */
1631 jffs_insert_file_into_tree(struct jffs_file *f)
1633 struct jffs_file *parent;
1635 D3(printk("jffs_insert_file_into_tree(): name: \"%s\"\n",
1636 (f->name ? f->name : "")));
1638 if (!(parent = jffs_find_file(f->c, f->pino))) {
1642 f->sibling_prev = NULL;
1643 f->sibling_next = NULL;
1647 D1(printk("jffs_insert_file_into_tree(): Found "
1648 "inode with no parent and pino == %u\n",
1654 f->sibling_next = parent->children;
1655 if (f->sibling_next) {
1656 f->sibling_next->sibling_prev = f;
1658 f->sibling_prev = NULL;
1659 parent->children = f;
1664 /* Remove a file from the hash table. */
1666 jffs_unlink_file_from_hash(struct jffs_file *f)
1668 D3(printk("jffs_unlink_file_from_hash(): f: 0x%p, "
1669 "ino %u\n", f, f->ino));
1676 /* Just remove the file from the parent's children. Don't free
1679 jffs_unlink_file_from_tree(struct jffs_file *f)
1681 D3(printk("jffs_unlink_file_from_tree(): ino: %d, pino: %d, name: "
1682 "\"%s\"\n", f->ino, f->pino, (f->name ? f->name : "")));
1684 if (f->sibling_prev) {
1685 f->sibling_prev->sibling_next = f->sibling_next;
1687 else if (f->parent) {
1688 D3(printk("f->parent=%p\n", f->parent));
1689 f->parent->children = f->sibling_next;
1691 if (f->sibling_next) {
1692 f->sibling_next->sibling_prev = f->sibling_prev;
1698 /* Find a file with its inode number. */
1700 jffs_find_file(struct jffs_control *c, __u32 ino)
1702 struct jffs_file *f;
1703 int i = ino % c->hash_len;
1704 struct list_head *tmp;
1706 D3(printk("jffs_find_file(): ino: %u\n", ino));
1708 for (tmp = c->hash[i].next; tmp != &c->hash[i]; tmp = tmp->next) {
1709 f = list_entry(tmp, struct jffs_file, hash);
1712 D3(printk("jffs_find_file(): Found file with ino "
1713 "%u. (name: \"%s\")\n",
1714 ino, (f->name ? f->name : ""));
1718 D3(printk("jffs_find_file(): Didn't find file "
1719 "with ino %u.\n", ino);
1725 /* Find a file in a directory. We are comparing the names. */
1727 jffs_find_child(struct jffs_file *dir, const char *name, int len)
1729 struct jffs_file *f;
1731 D3(printk("jffs_find_child()\n"));
1733 for (f = dir->children; f; f = f->sibling_next) {
1734 if (!f->deleted && f->name
1735 && !strncmp(f->name, name, len)
1736 && f->name[len] == '\0') {
1742 printk("jffs_find_child(): Found \"%s\".\n", f->name);
1745 char *copy = (char *) kmalloc(len + 1, GFP_KERNEL);
1747 memcpy(copy, name, len);
1750 printk("jffs_find_child(): Didn't find the file \"%s\".\n",
1751 (copy ? copy : ""));
1761 /* Write a raw inode that takes up a certain amount of space in the flash
1762 memory. At the end of the flash device, there is often space that is
1763 impossible to use. At these times we want to mark this space as not
1764 used. In the cases when the amount of space is greater or equal than
1765 a struct jffs_raw_inode, we write a "dummy node" that takes up this
1766 space. The space after the raw inode, if it exists, is left as it is.
1767 Since this space after the raw inode contains JFFS_EMPTY_BITMASK bytes,
1768 we can compute the checksum of it; we don't have to manipulate it any
1771 If the space left on the device is less than the size of a struct
1772 jffs_raw_inode, this space is filled with JFFS_DIRTY_BITMASK bytes.
1773 No raw inode is written this time. */
1775 jffs_write_dummy_node(struct jffs_control *c, struct jffs_fm *dirty_fm)
1777 struct jffs_fmcontrol *fmc = c->fmc;
1780 D1(printk("jffs_write_dummy_node(): dirty_fm->offset = 0x%08x, "
1781 "dirty_fm->size = %u\n",
1782 dirty_fm->offset, dirty_fm->size));
1784 if (dirty_fm->size >= sizeof(struct jffs_raw_inode)) {
1785 struct jffs_raw_inode raw_inode;
1786 memset(&raw_inode, 0, sizeof(struct jffs_raw_inode));
1787 raw_inode.magic = JFFS_MAGIC_BITMASK;
1788 raw_inode.dsize = dirty_fm->size
1789 - sizeof(struct jffs_raw_inode);
1790 raw_inode.dchksum = raw_inode.dsize * 0xff;
1792 = jffs_checksum(&raw_inode, sizeof(struct jffs_raw_inode));
1794 if ((err = flash_safe_write(fmc->mtd,
1796 (u_char *)&raw_inode,
1797 sizeof(struct jffs_raw_inode)))
1799 printk(KERN_ERR "JFFS: jffs_write_dummy_node: "
1800 "flash_safe_write failed!\n");
1805 flash_safe_acquire(fmc->mtd);
1806 flash_memset(fmc->mtd, dirty_fm->offset, 0, dirty_fm->size);
1807 flash_safe_release(fmc->mtd);
1810 D3(printk("jffs_write_dummy_node(): Leaving...\n"));
1815 /* Write a raw inode, possibly its name and possibly some data. */
1817 jffs_write_node(struct jffs_control *c, struct jffs_node *node,
1818 struct jffs_raw_inode *raw_inode,
1819 const char *name, const unsigned char *data,
1821 struct jffs_file *f)
1823 struct jffs_fmcontrol *fmc = c->fmc;
1825 struct kvec node_iovec[4];
1826 unsigned long iovec_cnt;
1832 __u32 total_name_size = raw_inode->nsize
1833 + JFFS_GET_PAD_BYTES(raw_inode->nsize);
1834 __u32 total_data_size = raw_inode->dsize
1835 + JFFS_GET_PAD_BYTES(raw_inode->dsize);
1836 __u32 total_size = sizeof(struct jffs_raw_inode)
1837 + total_name_size + total_data_size;
1839 /* If this node isn't something that will eventually let
1840 GC free even more space, then don't allow it unless
1841 there's at least max_chunk_size space still available
1844 slack = fmc->max_chunk_size;
1847 /* Fire the retrorockets and shoot the fruiton torpedoes, sir! */
1850 printk("jffs_write_node(): node == NULL\n");
1853 ASSERT(if (raw_inode && raw_inode->nsize && !name) {
1854 printk("*** jffs_write_node(): nsize = %u but name == NULL\n",
1859 D1(printk("jffs_write_node(): filename = \"%s\", ino = %u, "
1860 "total_size = %u\n",
1861 (name ? name : ""), raw_inode->ino,
1864 jffs_fm_write_lock(fmc);
1871 /* Deadlocks suck. */
1872 while(fmc->free_size < fmc->min_free_size + total_size + slack) {
1873 jffs_fm_write_unlock(fmc);
1874 if (!JFFS_ENOUGH_SPACE(c, total_size + slack))
1876 jffs_fm_write_lock(fmc);
1879 /* First try to allocate some flash memory. */
1880 err = jffs_fmalloc(fmc, total_size, node, &fm);
1882 if (err == -ENOSPC) {
1883 /* Just out of space. GC and try again */
1884 if (fmc->dirty_size < fmc->sector_size) {
1885 D(printk("jffs_write_node(): jffs_fmalloc(0x%p, %u) "
1886 "failed, no dirty space to GC\n", fmc,
1891 D1(printk(KERN_INFO "jffs_write_node(): Calling jffs_garbage_collect_now()\n"));
1892 jffs_fm_write_unlock(fmc);
1893 if ((err = jffs_garbage_collect_now(c))) {
1894 D(printk("jffs_write_node(): jffs_garbage_collect_now() failed\n"));
1897 jffs_fm_write_lock(fmc);
1902 jffs_fm_write_unlock(fmc);
1904 D(printk("jffs_write_node(): jffs_fmalloc(0x%p, %u) "
1905 "failed!\n", fmc, total_size));
1910 /* The jffs_fm struct that we got is not good enough.
1911 Make that space dirty and try again */
1912 if ((err = jffs_write_dummy_node(c, fm)) < 0) {
1915 jffs_fm_write_unlock(fmc);
1916 D(printk("jffs_write_node(): "
1917 "jffs_write_dummy_node(): Failed!\n"));
1925 ASSERT(if (fm->nodes == 0) {
1926 printk(KERN_ERR "jffs_write_node(): fm->nodes == 0\n");
1929 pos = node->fm->offset;
1931 /* Increment the version number here. We can't let the caller
1932 set it beforehand, because we might have had to do GC on a node
1933 of this file - and we'd end up reusing version numbers.
1936 raw_inode->version = f->highest_version + 1;
1937 D1(printk (KERN_NOTICE "jffs_write_node(): setting version of %s to %d\n", f->name, raw_inode->version));
1939 /* if the file was deleted, set the deleted bit in the raw inode */
1941 raw_inode->deleted = 1;
1944 /* Compute the checksum for the data and name chunks. */
1945 raw_inode->dchksum = jffs_checksum(data, raw_inode->dsize);
1946 raw_inode->nchksum = jffs_checksum(name, raw_inode->nsize);
1948 /* The checksum is calculated without the chksum and accurate
1949 fields so set them to zero first. */
1950 raw_inode->accurate = 0;
1951 raw_inode->chksum = 0;
1952 raw_inode->chksum = jffs_checksum(raw_inode,
1953 sizeof(struct jffs_raw_inode));
1954 raw_inode->accurate = 0xff;
1956 D3(printk("jffs_write_node(): About to write this raw inode to the "
1957 "flash at pos 0x%lx:\n", (long)pos));
1958 D3(jffs_print_raw_inode(raw_inode));
1960 /* The actual raw JFFS node */
1961 node_iovec[0].iov_base = (void *) raw_inode;
1962 node_iovec[0].iov_len = (size_t) sizeof(struct jffs_raw_inode);
1965 /* Get name and size if there is one */
1966 if (raw_inode->nsize) {
1967 node_iovec[iovec_cnt].iov_base = (void *) name;
1968 node_iovec[iovec_cnt].iov_len = (size_t) raw_inode->nsize;
1971 if (JFFS_GET_PAD_BYTES(raw_inode->nsize)) {
1972 static char allff[3]={255,255,255};
1973 /* Add some extra padding if necessary */
1974 node_iovec[iovec_cnt].iov_base = allff;
1975 node_iovec[iovec_cnt].iov_len =
1976 JFFS_GET_PAD_BYTES(raw_inode->nsize);
1981 /* Get data and size if there is any */
1982 if (raw_inode->dsize) {
1983 node_iovec[iovec_cnt].iov_base = (void *) data;
1984 node_iovec[iovec_cnt].iov_len = (size_t) raw_inode->dsize;
1986 /* No need to pad this because we're not actually putting
1991 if ((err = flash_safe_writev(fmc->mtd, node_iovec, iovec_cnt,
1993 jffs_fmfree_partly(fmc, fm, 0);
1994 jffs_fm_write_unlock(fmc);
1995 printk(KERN_ERR "JFFS: jffs_write_node: Failed to write, "
1996 "requested %i, wrote %i\n", total_size, err);
1999 if (raw_inode->deleted)
2002 jffs_fm_write_unlock(fmc);
2003 D3(printk("jffs_write_node(): Leaving...\n"));
2004 return raw_inode->dsize;
2005 } /* jffs_write_node() */
2008 /* Read data from the node and write it to the buffer. 'node_offset'
2009 is how much we have read from this particular node before and which
2010 shouldn't be read again. 'max_size' is how much space there is in
2013 jffs_get_node_data(struct jffs_file *f, struct jffs_node *node,
2014 unsigned char *buf,__u32 node_offset, __u32 max_size)
2016 struct jffs_fmcontrol *fmc = f->c->fmc;
2017 __u32 pos = node->fm->offset + node->fm_offset + node_offset;
2018 __u32 avail = node->data_size - node_offset;
2021 D2(printk(" jffs_get_node_data(): file: \"%s\", ino: %u, "
2022 "version: %u, node_offset: %u\n",
2023 f->name, node->ino, node->version, node_offset));
2025 r = min(avail, max_size);
2026 D3(printk(KERN_NOTICE "jffs_get_node_data\n"));
2027 flash_safe_read(fmc->mtd, pos, buf, r);
2029 D3(printk(" jffs_get_node_data(): Read %u byte%s.\n",
2030 r, (r == 1 ? "" : "s")));
2036 /* Read data from the file's nodes. Write the data to the buffer
2037 'buf'. 'read_offset' tells how much data we should skip. */
2039 jffs_read_data(struct jffs_file *f, unsigned char *buf, __u32 read_offset,
2042 struct jffs_node *node;
2043 __u32 read_data = 0; /* Total amount of read data. */
2044 __u32 node_offset = 0;
2045 __u32 pos = 0; /* Number of bytes traversed. */
2047 D2(printk("jffs_read_data(): file = \"%s\", read_offset = %d, "
2049 (f->name ? f->name : ""), read_offset, size));
2051 if (read_offset >= f->size) {
2052 D(printk(" f->size: %d\n", f->size));
2056 /* First find the node to read data from. */
2057 node = f->range_head;
2058 while (pos <= read_offset) {
2059 node_offset = read_offset - pos;
2060 if (node_offset >= node->data_size) {
2061 pos += node->data_size;
2062 node = node->range_next;
2069 /* "Cats are living proof that not everything in nature
2071 - Garrison Keilor ('97) */
2073 /* Fill the buffer. */
2074 while (node && (read_data < size)) {
2077 /* This node does not refer to real data. */
2078 r = min(size - read_data,
2079 node->data_size - node_offset);
2080 memset(&buf[read_data], 0, r);
2082 else if ((r = jffs_get_node_data(f, node, &buf[read_data],
2084 size - read_data)) < 0) {
2089 node = node->range_next;
2091 D3(printk(" jffs_read_data(): Read %u bytes.\n", read_data));
2096 /* Used for traversing all nodes in the hash table. */
2098 jffs_foreach_file(struct jffs_control *c, int (*func)(struct jffs_file *))
2104 for (pos = 0; pos < c->hash_len; pos++) {
2105 struct list_head *p, *next;
2106 for (p = c->hash[pos].next; p != &c->hash[pos]; p = next) {
2107 /* We need a reference to the next file in the
2108 list because `func' might remove the current
2111 r = func(list_entry(p, struct jffs_file, hash));
2122 /* Free all nodes associated with a file. */
2124 jffs_free_node_list(struct jffs_file *f)
2126 struct jffs_node *node;
2127 struct jffs_node *p;
2129 D3(printk("jffs_free_node_list(): f #%u, \"%s\"\n",
2130 f->ino, (f->name ? f->name : "")));
2131 node = f->version_head;
2134 node = node->version_next;
2136 DJM(no_jffs_node--);
2142 /* Free a file and its name. */
2144 jffs_free_file(struct jffs_file *f)
2146 D3(printk("jffs_free_file: f #%u, \"%s\"\n",
2147 f->ino, (f->name ? f->name : "")));
2159 jffs_get_file_count(void)
2161 return no_jffs_file;
2164 /* See if a file is deleted. If so, mark that file's nodes as obsolete. */
2166 jffs_possibly_delete_file(struct jffs_file *f)
2168 struct jffs_node *n;
2170 D3(printk("jffs_possibly_delete_file(): ino: %u\n",
2174 printk(KERN_ERR "jffs_possibly_delete_file(): f == NULL\n");
2179 /* First try to remove all older versions. Commence with
2181 for (n = f->version_head; n; n = n->version_next) {
2185 if (jffs_fmfree(f->c->fmc, n->fm, n) < 0) {
2189 /* Unlink the file from the filesystem. */
2190 if (!f->c->building_fs) {
2191 jffs_unlink_file_from_tree(f);
2193 jffs_unlink_file_from_hash(f);
2194 jffs_free_node_list(f);
2201 /* Used in conjunction with jffs_foreach_file() to count the number
2202 of files in the file system. */
2204 jffs_file_count(struct jffs_file *f)
2210 /* Build up a file's range list from scratch by going through the
2213 jffs_build_file(struct jffs_file *f)
2215 struct jffs_node *n;
2217 D3(printk("jffs_build_file(): ino: %u, name: \"%s\"\n",
2218 f->ino, (f->name ? f->name : "")));
2220 for (n = f->version_head; n; n = n->version_next) {
2221 jffs_update_file(f, n);
2227 /* Remove an amount of data from a file. If this amount of data is
2228 zero, that could mean that a node should be split in two parts.
2229 We remove or change the appropriate nodes in the lists.
2231 Starting offset of area to be removed is node->data_offset,
2232 and the length of the area is in node->removed_size. */
2234 jffs_delete_data(struct jffs_file *f, struct jffs_node *node)
2236 struct jffs_node *n;
2237 __u32 offset = node->data_offset;
2238 __u32 remove_size = node->removed_size;
2240 D3(printk("jffs_delete_data(): offset = %u, remove_size = %u\n",
2241 offset, remove_size));
2243 if (remove_size == 0
2245 && f->range_tail->data_offset + f->range_tail->data_size
2247 /* A simple append; nothing to remove or no node to split. */
2251 /* Find the node where we should begin the removal. */
2252 for (n = f->range_head; n; n = n->range_next) {
2253 if (n->data_offset + n->data_size > offset) {
2258 /* If there's no data in the file there's no data to
2263 if (n->data_offset > offset) {
2264 /* XXX: Not implemented yet. */
2265 printk(KERN_WARNING "JFFS: An unexpected situation "
2266 "occurred in jffs_delete_data.\n");
2268 else if (n->data_offset < offset) {
2269 /* See if the node has to be split into two parts. */
2270 if (n->data_offset + n->data_size > offset + remove_size) {
2272 struct jffs_node *new_node;
2273 D3(printk("jffs_delete_data(): Split node with "
2274 "version number %u.\n", n->version));
2276 if (!(new_node = jffs_alloc_node())) {
2277 D(printk("jffs_delete_data(): -ENOMEM\n"));
2280 DJM(no_jffs_node++);
2282 new_node->ino = n->ino;
2283 new_node->version = n->version;
2284 new_node->data_offset = offset;
2285 new_node->data_size = n->data_size - (remove_size + (offset - n->data_offset));
2286 new_node->fm_offset = n->fm_offset + (remove_size + (offset - n->data_offset));
2287 new_node->name_size = n->name_size;
2288 new_node->fm = n->fm;
2289 new_node->version_prev = n;
2290 new_node->version_next = n->version_next;
2291 if (new_node->version_next) {
2292 new_node->version_next->version_prev
2296 f->version_tail = new_node;
2298 n->version_next = new_node;
2299 new_node->range_prev = n;
2300 new_node->range_next = n->range_next;
2301 if (new_node->range_next) {
2302 new_node->range_next->range_prev = new_node;
2305 f->range_tail = new_node;
2307 /* A very interesting can of worms. */
2308 n->range_next = new_node;
2309 n->data_size = offset - n->data_offset;
2311 jffs_add_node(new_node);
2313 D1(printk(KERN_WARNING "jffs_delete_data(): Splitting an empty node (file hold).\n!"));
2314 D1(printk(KERN_WARNING "FIXME: Did dwmw2 do the right thing here?\n"));
2316 n = new_node->range_next;
2320 /* No. No need to split the node. Just remove
2321 the end of the node. */
2322 int r = min(n->data_offset + n->data_size
2323 - offset, remove_size);
2330 /* Remove as many nodes as necessary. */
2331 while (n && remove_size) {
2332 if (n->data_size <= remove_size) {
2333 struct jffs_node *p = n;
2334 remove_size -= n->data_size;
2336 D3(printk("jffs_delete_data(): Removing node: "
2337 "ino: %u, version: %u%s\n",
2339 (p->fm ? "" : " (virtual)")));
2341 jffs_fmfree(f->c->fmc, p->fm, p);
2343 jffs_unlink_node_from_range_list(f, p);
2344 jffs_unlink_node_from_version_list(f, p);
2346 DJM(no_jffs_node--);
2349 n->data_size -= remove_size;
2350 n->fm_offset += remove_size;
2351 n->data_offset -= (node->removed_size - remove_size);
2357 /* Adjust the following nodes' information about offsets etc. */
2358 while (n && node->removed_size) {
2359 n->data_offset -= node->removed_size;
2363 if (node->removed_size > (f->size - node->data_offset)) {
2364 /* It's possible that the removed_size is in fact
2365 * greater than the amount of data we actually thought
2366 * were present in the first place - some of the nodes
2367 * which this node originally obsoleted may already have
2368 * been deleted from the flash by subsequent garbage
2371 * If this is the case, don't let f->size go negative.
2372 * Bad things would happen :)
2374 f->size = node->data_offset;
2376 f->size -= node->removed_size;
2378 D3(printk("jffs_delete_data(): f->size = %d\n", f->size));
2380 } /* jffs_delete_data() */
2383 /* Insert some data into a file. Prior to the call to this function,
2384 jffs_delete_data should be called. */
2386 jffs_insert_data(struct jffs_file *f, struct jffs_node *node)
2388 D3(printk("jffs_insert_data(): node->data_offset = %u, "
2389 "node->data_size = %u, f->size = %u\n",
2390 node->data_offset, node->data_size, f->size));
2392 /* Find the position where we should insert data. */
2394 if (node->data_offset == f->size) {
2395 /* A simple append. This is the most common operation. */
2396 node->range_next = NULL;
2397 node->range_prev = f->range_tail;
2398 if (node->range_prev) {
2399 node->range_prev->range_next = node;
2401 f->range_tail = node;
2402 f->size += node->data_size;
2403 if (!f->range_head) {
2404 f->range_head = node;
2407 else if (node->data_offset < f->size) {
2408 /* Trying to insert data into the middle of the file. This
2409 means no problem because jffs_delete_data() has already
2410 prepared the range list for us. */
2411 struct jffs_node *n;
2413 /* Find the correct place for the insertion and then insert
2415 for (n = f->range_head; n; n = n->range_next) {
2416 D2(printk("Cool stuff's happening!\n"));
2418 if (n->data_offset == node->data_offset) {
2419 node->range_prev = n->range_prev;
2420 if (node->range_prev) {
2421 node->range_prev->range_next = node;
2424 f->range_head = node;
2426 node->range_next = n;
2427 n->range_prev = node;
2430 ASSERT(else if (n->data_offset + n->data_size >
2431 node->data_offset) {
2432 printk(KERN_ERR "jffs_insert_data(): "
2433 "Couldn't find a place to insert "
2439 /* Adjust later nodes' offsets etc. */
2440 n = node->range_next;
2442 n->data_offset += node->data_size;
2445 f->size += node->data_size;
2447 else if (node->data_offset > f->size) {
2448 /* Okay. This is tricky. This means that we want to insert
2449 data at a place that is beyond the limits of the file as
2450 it is constructed right now. This is actually a common
2451 event that for instance could occur during the mounting
2452 of the file system if a large file have been truncated,
2453 rewritten and then only partially garbage collected. */
2455 struct jffs_node *n;
2457 /* We need a place holder for the data that is missing in
2458 front of this insertion. This "virtual node" will not
2459 be associated with any space on the flash device. */
2460 struct jffs_node *virtual_node;
2461 if (!(virtual_node = jffs_alloc_node())) {
2465 D(printk("jffs_insert_data: Inserting a virtual node.\n"));
2466 D(printk(" node->data_offset = %u\n", node->data_offset));
2467 D(printk(" f->size = %u\n", f->size));
2469 virtual_node->ino = node->ino;
2470 virtual_node->version = node->version;
2471 virtual_node->removed_size = 0;
2472 virtual_node->fm_offset = 0;
2473 virtual_node->name_size = 0;
2474 virtual_node->fm = NULL; /* This is a virtual data holder. */
2475 virtual_node->version_prev = NULL;
2476 virtual_node->version_next = NULL;
2477 virtual_node->range_next = NULL;
2479 /* Are there any data at all in the file yet? */
2480 if (f->range_head) {
2481 virtual_node->data_offset
2482 = f->range_tail->data_offset
2483 + f->range_tail->data_size;
2484 virtual_node->data_size
2485 = node->data_offset - virtual_node->data_offset;
2486 virtual_node->range_prev = f->range_tail;
2487 f->range_tail->range_next = virtual_node;
2490 virtual_node->data_offset = 0;
2491 virtual_node->data_size = node->data_offset;
2492 virtual_node->range_prev = NULL;
2493 f->range_head = virtual_node;
2496 f->range_tail = virtual_node;
2497 f->size += virtual_node->data_size;
2499 /* Insert this virtual node in the version list as well. */
2500 for (n = f->version_head; n ; n = n->version_next) {
2501 if (n->version == virtual_node->version) {
2502 virtual_node->version_prev = n->version_prev;
2503 n->version_prev = virtual_node;
2504 if (virtual_node->version_prev) {
2505 virtual_node->version_prev
2506 ->version_next = virtual_node;
2509 f->version_head = virtual_node;
2511 virtual_node->version_next = n;
2516 D(jffs_print_node(virtual_node));
2518 /* Make a new try to insert the node. */
2522 D3(printk("jffs_insert_data(): f->size = %d\n", f->size));
2527 /* A new node (with data) has been added to the file and now the range
2528 list has to be modified. */
2530 jffs_update_file(struct jffs_file *f, struct jffs_node *node)
2534 D3(printk("jffs_update_file(): ino: %u, version: %u\n",
2535 f->ino, node->version));
2537 if (node->data_size == 0) {
2538 if (node->removed_size == 0) {
2539 /* data_offset == X */
2540 /* data_size == 0 */
2541 /* remove_size == 0 */
2544 /* data_offset == X */
2545 /* data_size == 0 */
2546 /* remove_size != 0 */
2547 if ((err = jffs_delete_data(f, node)) < 0) {
2553 /* data_offset == X */
2554 /* data_size != 0 */
2555 /* remove_size == Y */
2556 if ((err = jffs_delete_data(f, node)) < 0) {
2559 if ((err = jffs_insert_data(f, node)) < 0) {
2566 /* Print the contents of a file. */
2569 jffs_print_file(struct jffs_file *f)
2572 D(printk("jffs_file: 0x%p\n", f));
2574 D(printk(" 0x%08x, /* ino */\n", f->ino));
2575 D(printk(" 0x%08x, /* pino */\n", f->pino));
2576 D(printk(" 0x%08x, /* mode */\n", f->mode));
2577 D(printk(" 0x%04x, /* uid */\n", f->uid));
2578 D(printk(" 0x%04x, /* gid */\n", f->gid));
2579 D(printk(" 0x%08x, /* atime */\n", f->atime));
2580 D(printk(" 0x%08x, /* mtime */\n", f->mtime));
2581 D(printk(" 0x%08x, /* ctime */\n", f->ctime));
2582 D(printk(" 0x%02x, /* nsize */\n", f->nsize));
2583 D(printk(" 0x%02x, /* nlink */\n", f->nlink));
2584 D(printk(" 0x%02x, /* deleted */\n", f->deleted));
2585 D(printk(" \"%s\", ", (f->name ? f->name : "")));
2586 D(for (i = strlen(f->name ? f->name : ""); i < 8; ++i) {
2589 D(printk("/* name */\n"));
2590 D(printk(" 0x%08x, /* size */\n", f->size));
2591 D(printk(" 0x%08x, /* highest_version */\n",
2592 f->highest_version));
2593 D(printk(" 0x%p, /* c */\n", f->c));
2594 D(printk(" 0x%p, /* parent */\n", f->parent));
2595 D(printk(" 0x%p, /* children */\n", f->children));
2596 D(printk(" 0x%p, /* sibling_prev */\n", f->sibling_prev));
2597 D(printk(" 0x%p, /* sibling_next */\n", f->sibling_next));
2598 D(printk(" 0x%p, /* hash_prev */\n", f->hash.prev));
2599 D(printk(" 0x%p, /* hash_next */\n", f->hash.next));
2600 D(printk(" 0x%p, /* range_head */\n", f->range_head));
2601 D(printk(" 0x%p, /* range_tail */\n", f->range_tail));
2602 D(printk(" 0x%p, /* version_head */\n", f->version_head));
2603 D(printk(" 0x%p, /* version_tail */\n", f->version_tail));
2610 jffs_print_hash_table(struct jffs_control *c)
2614 printk("JFFS: Dumping the file system's hash table...\n");
2615 for (i = 0; i < c->hash_len; i++) {
2616 struct list_head *p;
2617 for (p = c->hash[i].next; p != &c->hash[i]; p = p->next) {
2618 struct jffs_file *f=list_entry(p,struct jffs_file,hash);
2619 printk("*** c->hash[%u]: \"%s\" "
2620 "(ino: %u, pino: %u)\n",
2621 i, (f->name ? f->name : ""),
2629 jffs_print_tree(struct jffs_file *first_file, int indent)
2631 struct jffs_file *f;
2639 if (!(space = (char *) kmalloc(indent + 1, GFP_KERNEL))) {
2640 printk("jffs_print_tree(): Out of memory!\n");
2644 memset(space, ' ', indent);
2645 space[indent] = '\0';
2647 for (f = first_file; f; f = f->sibling_next) {
2648 dir = S_ISDIR(f->mode);
2649 printk("%s%s%s (ino: %u, highest_version: %u, size: %u)\n",
2650 space, (f->name ? f->name : ""), (dir ? "/" : ""),
2651 f->ino, f->highest_version, f->size);
2653 jffs_print_tree(f->children, indent + 2);
2661 #if defined(JFFS_MEMORY_DEBUG) && JFFS_MEMORY_DEBUG
2663 jffs_print_memory_allocation_statistics(void)
2665 static long printout;
2666 printk("________ Memory printout #%ld ________\n", ++printout);
2667 printk("no_jffs_file = %ld\n", no_jffs_file);
2668 printk("no_jffs_node = %ld\n", no_jffs_node);
2669 printk("no_jffs_control = %ld\n", no_jffs_control);
2670 printk("no_jffs_raw_inode = %ld\n", no_jffs_raw_inode);
2671 printk("no_jffs_node_ref = %ld\n", no_jffs_node_ref);
2672 printk("no_jffs_fm = %ld\n", no_jffs_fm);
2673 printk("no_jffs_fmcontrol = %ld\n", no_jffs_fmcontrol);
2674 printk("no_hash = %ld\n", no_hash);
2675 printk("no_name = %ld\n", no_name);
2681 /* Rewrite `size' bytes, and begin at `node'. */
2683 jffs_rewrite_data(struct jffs_file *f, struct jffs_node *node, __u32 size)
2685 struct jffs_control *c = f->c;
2686 struct jffs_fmcontrol *fmc = c->fmc;
2687 struct jffs_raw_inode raw_inode;
2688 struct jffs_node *new_node;
2692 __u32 total_name_size;
2693 __u32 total_data_size;
2697 D1(printk("***jffs_rewrite_data(): node: %u, name: \"%s\", size: %u\n",
2698 f->ino, (f->name ? f->name : "(null)"), size));
2700 /* Create and initialize the new node. */
2701 if (!(new_node = jffs_alloc_node())) {
2702 D(printk("jffs_rewrite_data(): "
2703 "Failed to allocate node.\n"));
2706 DJM(no_jffs_node++);
2707 new_node->data_offset = node->data_offset;
2708 new_node->removed_size = size;
2709 total_name_size = JFFS_PAD(f->nsize);
2710 total_data_size = JFFS_PAD(size);
2711 total_size = sizeof(struct jffs_raw_inode)
2712 + total_name_size + total_data_size;
2713 new_node->fm_offset = sizeof(struct jffs_raw_inode)
2717 jffs_fm_write_lock(fmc);
2720 if ((err = jffs_fmalloc(fmc, total_size, new_node, &fm)) < 0) {
2721 DJM(no_jffs_node--);
2722 jffs_fm_write_unlock(fmc);
2723 D(printk("jffs_rewrite_data(): Failed to allocate fm.\n"));
2724 jffs_free_node(new_node);
2727 else if (!fm->nodes) {
2728 /* The jffs_fm struct that we got is not big enough. */
2729 /* This should never happen, because we deal with this case
2730 in jffs_garbage_collect_next().*/
2731 printk(KERN_WARNING "jffs_rewrite_data(): Allocated node is too small (%d bytes of %d)\n", fm->size, total_size);
2732 if ((err = jffs_write_dummy_node(c, fm)) < 0) {
2733 D(printk("jffs_rewrite_data(): "
2734 "jffs_write_dummy_node() Failed!\n"));
2739 jffs_fm_write_unlock(fmc);
2746 /* Initialize the raw inode. */
2747 raw_inode.magic = JFFS_MAGIC_BITMASK;
2748 raw_inode.ino = f->ino;
2749 raw_inode.pino = f->pino;
2750 raw_inode.version = f->highest_version + 1;
2751 raw_inode.mode = f->mode;
2752 raw_inode.uid = f->uid;
2753 raw_inode.gid = f->gid;
2754 raw_inode.atime = f->atime;
2755 raw_inode.mtime = f->mtime;
2756 raw_inode.ctime = f->ctime;
2757 raw_inode.offset = node->data_offset;
2758 raw_inode.dsize = size;
2759 raw_inode.rsize = size;
2760 raw_inode.nsize = f->nsize;
2761 raw_inode.nlink = f->nlink;
2762 raw_inode.spare = 0;
2763 raw_inode.rename = 0;
2764 raw_inode.deleted = f->deleted;
2765 raw_inode.accurate = 0xff;
2766 raw_inode.dchksum = 0;
2767 raw_inode.nchksum = 0;
2769 pos = new_node->fm->offset;
2770 pos_dchksum = pos +JFFS_RAW_INODE_DCHKSUM_OFFSET;
2772 D3(printk("jffs_rewrite_data(): Writing this raw inode "
2773 "to pos 0x%ul.\n", pos));
2774 D3(jffs_print_raw_inode(&raw_inode));
2776 if ((err = flash_safe_write(fmc->mtd, pos,
2777 (u_char *) &raw_inode,
2778 sizeof(struct jffs_raw_inode)
2780 - sizeof(__u16) - sizeof(__u16))) < 0) {
2781 jffs_fmfree_partly(fmc, fm,
2782 total_name_size + total_data_size);
2783 jffs_fm_write_unlock(fmc);
2784 printk(KERN_ERR "JFFS: jffs_rewrite_data: Write error during "
2785 "rewrite. (raw inode)\n");
2786 printk(KERN_ERR "JFFS: jffs_rewrite_data: Now retrying "
2787 "rewrite. (raw inode)\n");
2790 pos += sizeof(struct jffs_raw_inode);
2792 /* Write the name to the flash memory. */
2794 D3(printk("jffs_rewrite_data(): Writing name \"%s\" to "
2795 "pos 0x%ul.\n", f->name, (unsigned int) pos));
2796 if ((err = flash_safe_write(fmc->mtd, pos,
2799 jffs_fmfree_partly(fmc, fm, total_data_size);
2800 jffs_fm_write_unlock(fmc);
2801 printk(KERN_ERR "JFFS: jffs_rewrite_data: Write "
2802 "error during rewrite. (name)\n");
2803 printk(KERN_ERR "JFFS: jffs_rewrite_data: Now retrying "
2804 "rewrite. (name)\n");
2807 pos += total_name_size;
2808 raw_inode.nchksum = jffs_checksum(f->name, f->nsize);
2811 /* Write the data. */
2814 unsigned char *page;
2815 __u32 offset = node->data_offset;
2817 if (!(page = (unsigned char *)__get_free_page(GFP_KERNEL))) {
2818 jffs_fmfree_partly(fmc, fm, 0);
2823 __u32 s = min(size, (__u32)PAGE_SIZE);
2824 if ((r = jffs_read_data(f, (char *)page,
2826 free_page((unsigned long)page);
2827 jffs_fmfree_partly(fmc, fm, 0);
2828 jffs_fm_write_unlock(fmc);
2829 printk(KERN_ERR "JFFS: jffs_rewrite_data: "
2831 "failed! (r = %d)\n", r);
2834 if ((err = flash_safe_write(fmc->mtd,
2835 pos, page, r)) < 0) {
2836 free_page((unsigned long)page);
2837 jffs_fmfree_partly(fmc, fm, 0);
2838 jffs_fm_write_unlock(fmc);
2839 printk(KERN_ERR "JFFS: jffs_rewrite_data: "
2840 "Write error during rewrite. "
2847 raw_inode.dchksum += jffs_checksum(page, r);
2850 free_page((unsigned long)page);
2853 raw_inode.accurate = 0;
2854 raw_inode.chksum = jffs_checksum(&raw_inode,
2855 sizeof(struct jffs_raw_inode)
2858 /* Add the checksum. */
2860 = flash_safe_write(fmc->mtd, pos_dchksum,
2862 &raw_inode)[JFFS_RAW_INODE_DCHKSUM_OFFSET],
2863 sizeof(__u32) + sizeof(__u16)
2864 + sizeof(__u16))) < 0) {
2865 jffs_fmfree_partly(fmc, fm, 0);
2866 jffs_fm_write_unlock(fmc);
2867 printk(KERN_ERR "JFFS: jffs_rewrite_data: Write error during "
2868 "rewrite. (checksum)\n");
2872 /* Now make the file system aware of the newly written node. */
2873 jffs_insert_node(c, f, &raw_inode, f->name, new_node);
2874 jffs_fm_write_unlock(fmc);
2876 D3(printk("jffs_rewrite_data(): Leaving...\n"));
2878 } /* jffs_rewrite_data() */
2881 /* jffs_garbage_collect_next implements one step in the garbage collect
2882 process and is often called multiple times at each occasion of a
2886 jffs_garbage_collect_next(struct jffs_control *c)
2888 struct jffs_fmcontrol *fmc = c->fmc;
2889 struct jffs_node *node;
2890 struct jffs_file *f;
2894 __u32 total_name_size;
2895 __u32 extra_available;
2897 __u32 free_chunk_size1 = jffs_free_size1(fmc);
2898 D2(__u32 free_chunk_size2 = jffs_free_size2(fmc));
2900 /* Get the oldest node in the flash. */
2901 node = jffs_get_oldest_node(fmc);
2903 printk(KERN_ERR "JFFS: jffs_garbage_collect_next: "
2904 "No oldest node found!\n");
2906 goto jffs_garbage_collect_next_end;
2911 /* Find its corresponding file too. */
2912 f = jffs_find_file(c, node->ino);
2915 printk (KERN_ERR "JFFS: jffs_garbage_collect_next: "
2916 "No file to garbage collect! "
2917 "(ino = 0x%08x)\n", node->ino);
2918 /* FIXME: Free the offending node and recover. */
2920 goto jffs_garbage_collect_next_end;
2923 /* We always write out the name. Theoretically, we don't need
2924 to, but for now it's easier - because otherwise we'd have
2925 to keep track of how many times the current name exists on
2926 the flash and make sure it never reaches zero.
2928 The current approach means that would be possible to cause
2929 the GC to end up eating its tail by writing lots of nodes
2930 with no name for it to garbage-collect. Hence the change in
2931 inode.c to write names with _every_ node.
2933 It sucks, but it _should_ work.
2935 total_name_size = JFFS_PAD(f->nsize);
2937 D1(printk("jffs_garbage_collect_next(): \"%s\", "
2938 "ino: %u, version: %u, location 0x%x, dsize %u\n",
2939 (f->name ? f->name : ""), node->ino, node->version,
2940 node->fm->offset, node->data_size));
2942 /* Compute how many data it's possible to rewrite at the moment. */
2943 data_size = f->size - node->data_offset;
2945 /* And from that, the total size of the chunk we want to write */
2946 size = sizeof(struct jffs_raw_inode) + total_name_size
2947 + data_size + JFFS_GET_PAD_BYTES(data_size);
2949 /* If that's more than max_chunk_size, reduce it accordingly */
2950 if (size > fmc->max_chunk_size) {
2951 size = fmc->max_chunk_size;
2952 data_size = size - sizeof(struct jffs_raw_inode)
2956 /* If we're asking to take up more space than free_chunk_size1
2957 but we _could_ fit in it, shrink accordingly.
2959 if (size > free_chunk_size1) {
2961 if (free_chunk_size1 <
2962 (sizeof(struct jffs_raw_inode) + total_name_size + BLOCK_SIZE)){
2963 /* The space left is too small to be of any
2965 struct jffs_fm *dirty_fm
2966 = jffs_fmalloced(fmc,
2967 fmc->tail->offset + fmc->tail->size,
2968 free_chunk_size1, NULL);
2970 printk(KERN_ERR "JFFS: "
2971 "jffs_garbage_collect_next: "
2972 "Failed to allocate `dirty' "
2975 goto jffs_garbage_collect_next_end;
2977 D1(printk("Dirtying end of flash - too small\n"));
2978 jffs_write_dummy_node(c, dirty_fm);
2980 goto jffs_garbage_collect_next_end;
2982 D1(printk("Reducing size of new node from %d to %d to avoid "
2983 " exceeding free_chunk_size1\n",
2984 size, free_chunk_size1));
2986 size = free_chunk_size1;
2987 data_size = size - sizeof(struct jffs_raw_inode)
2992 /* Calculate the amount of space needed to hold the nodes
2993 which are remaining in the tail */
2994 space_needed = fmc->min_free_size - (node->fm->offset % fmc->sector_size);
2996 /* From that, calculate how much 'extra' space we can use to
2997 increase the size of the node we're writing from the size
2998 of the node we're obsoleting
3000 if (space_needed > fmc->free_size) {
3001 /* If we've gone below min_free_size for some reason,
3002 don't fuck up. This is why we have
3003 min_free_size > sector_size. Whinge about it though,
3004 just so I can convince myself my maths is right.
3006 D1(printk(KERN_WARNING "jffs_garbage_collect_next(): "
3007 "space_needed %d exceeded free_size %d\n",
3008 space_needed, fmc->free_size));
3009 extra_available = 0;
3011 extra_available = fmc->free_size - space_needed;
3014 /* Check that we don't use up any more 'extra' space than
3016 if (size > JFFS_PAD(node->data_size) + total_name_size +
3017 sizeof(struct jffs_raw_inode) + extra_available) {
3018 D1(printk("Reducing size of new node from %d to %ld to avoid "
3019 "catching our tail\n", size,
3020 (long) (JFFS_PAD(node->data_size) + JFFS_PAD(node->name_size) +
3021 sizeof(struct jffs_raw_inode) + extra_available)));
3022 D1(printk("space_needed = %d, extra_available = %d\n",
3023 space_needed, extra_available));
3025 size = JFFS_PAD(node->data_size) + total_name_size +
3026 sizeof(struct jffs_raw_inode) + extra_available;
3027 data_size = size - sizeof(struct jffs_raw_inode)
3031 D2(printk(" total_name_size: %u\n", total_name_size));
3032 D2(printk(" data_size: %u\n", data_size));
3033 D2(printk(" size: %u\n", size));
3034 D2(printk(" f->nsize: %u\n", f->nsize));
3035 D2(printk(" f->size: %u\n", f->size));
3036 D2(printk(" node->data_offset: %u\n", node->data_offset));
3037 D2(printk(" free_chunk_size1: %u\n", free_chunk_size1));
3038 D2(printk(" free_chunk_size2: %u\n", free_chunk_size2));
3039 D2(printk(" node->fm->offset: 0x%08x\n", node->fm->offset));
3041 if ((err = jffs_rewrite_data(f, node, data_size))) {
3042 printk(KERN_WARNING "jffs_rewrite_data() failed: %d\n", err);
3046 jffs_garbage_collect_next_end:
3047 D3(printk("jffs_garbage_collect_next: Leaving...\n"));
3049 } /* jffs_garbage_collect_next */
3052 /* If an obsolete node is partly going to be erased due to garbage
3053 collection, the part that isn't going to be erased must be filled
3054 with zeroes so that the scan of the flash will work smoothly next
3055 time. (The data in the file could for instance be a JFFS image
3056 which could cause enormous confusion during a scan of the flash
3057 device if we didn't do this.)
3058 There are two phases in this procedure: First, the clearing of
3059 the name and data parts of the node. Second, possibly also clearing
3060 a part of the raw inode as well. If the box is power cycled during
3061 the first phase, only the checksum of this node-to-be-cleared-at-
3062 the-end will be wrong. If the box is power cycled during, or after,
3063 the clearing of the raw inode, the information like the length of
3064 the name and data parts are zeroed. The next time the box is
3065 powered up, the scanning algorithm manages this faulty data too
3068 - The checksum is invalid and thus the raw inode must be discarded
3070 - If the lengths of the data part or the name part are zeroed, the
3071 scanning just continues after the raw inode. But after the inode
3072 the scanning procedure just finds zeroes which is the same as
3075 So, in the end, this could never fail. :-) Even if it does fail,
3076 the scanning algorithm should manage that too. */
3079 jffs_clear_end_of_node(struct jffs_control *c, __u32 erase_size)
3082 struct jffs_fmcontrol *fmc = c->fmc;
3085 __u32 zero_offset_data;
3086 __u32 zero_size_data;
3087 __u32 cutting_raw_inode = 0;
3089 if (!(fm = jffs_cut_node(fmc, erase_size))) {
3090 D3(printk("jffs_clear_end_of_node(): fm == NULL\n"));
3094 /* Where and how much shall we clear? */
3095 zero_offset = fmc->head->offset + erase_size;
3096 zero_size = fm->offset + fm->size - zero_offset;
3098 /* Do we have to clear the raw_inode explicitly? */
3099 if (fm->size - zero_size < sizeof(struct jffs_raw_inode)) {
3100 cutting_raw_inode = sizeof(struct jffs_raw_inode)
3101 - (fm->size - zero_size);
3104 /* First, clear the name and data fields. */
3105 zero_offset_data = zero_offset + cutting_raw_inode;
3106 zero_size_data = zero_size - cutting_raw_inode;
3107 flash_safe_acquire(fmc->mtd);
3108 flash_memset(fmc->mtd, zero_offset_data, 0, zero_size_data);
3109 flash_safe_release(fmc->mtd);
3111 /* Should we clear a part of the raw inode? */
3112 if (cutting_raw_inode) {
3113 /* I guess it is ok to clear the raw inode in this order. */
3114 flash_safe_acquire(fmc->mtd);
3115 flash_memset(fmc->mtd, zero_offset, 0,
3117 flash_safe_release(fmc->mtd);
3121 } /* jffs_clear_end_of_node() */
3123 /* Try to erase as much as possible of the dirt in the flash memory. */
3125 jffs_try_to_erase(struct jffs_control *c)
3127 struct jffs_fmcontrol *fmc = c->fmc;
3132 D3(printk("jffs_try_to_erase()\n"));
3134 erase_size = jffs_erasable_size(fmc);
3136 D2(printk("jffs_try_to_erase(): erase_size = %ld\n", erase_size));
3138 if (erase_size == 0) {
3141 else if (erase_size < 0) {
3142 printk(KERN_ERR "JFFS: jffs_try_to_erase: "
3143 "jffs_erasable_size returned %ld.\n", erase_size);
3147 if ((err = jffs_clear_end_of_node(c, erase_size)) < 0) {
3148 printk(KERN_ERR "JFFS: jffs_try_to_erase: "
3149 "Clearing of node failed.\n");
3153 offset = fmc->head->offset;
3155 /* Now, let's try to do the erase. */
3156 if ((err = flash_erase_region(fmc->mtd,
3157 offset, erase_size)) < 0) {
3158 printk(KERN_ERR "JFFS: Erase of flash failed. "
3159 "offset = %u, erase_size = %ld\n",
3160 offset, erase_size);
3161 /* XXX: Here we should allocate this area as dirty
3162 with jffs_fmalloced or something similar. Now
3163 we just report the error. */
3168 /* Check if the erased sectors really got erased. */
3173 pos = (__u32)flash_get_direct_pointer(to_kdev_t(c->sb->s_dev), offset);
3174 end = pos + erase_size;
3176 D2(printk("JFFS: Checking erased sector(s)...\n"));
3178 flash_safe_acquire(fmc->mtd);
3180 for (; pos < end; pos += 4) {
3181 if (*(__u32 *)pos != JFFS_EMPTY_BITMASK) {
3182 printk("JFFS: Erase failed! pos = 0x%lx\n",
3184 jffs_hexdump(fmc->mtd, pos,
3185 jffs_min(256, end - pos));
3191 flash_safe_release(fmc->mtd);
3194 D2(printk("JFFS: Erase succeeded.\n"));
3197 /* XXX: Here we should allocate the memory
3198 with jffs_fmalloced() in order to prevent
3199 JFFS from using this area accidentally. */
3205 /* Update the flash memory data structures. */
3206 jffs_sync_erase(fmc, erase_size);
3212 /* There are different criteria that should trigger a garbage collect:
3214 1. There is too much dirt in the memory.
3215 2. The free space is becoming small.
3216 3. There are many versions of a node.
3218 The garbage collect should always be done in a manner that guarantees
3219 that future garbage collects cannot be locked. E.g. Rewritten chunks
3220 should not be too large (span more than one sector in the flash memory
3221 for exemple). Of course there is a limit on how intelligent this garbage
3222 collection can be. */
3226 jffs_garbage_collect_now(struct jffs_control *c)
3228 struct jffs_fmcontrol *fmc = c->fmc;
3232 D2(printk("***jffs_garbage_collect_now(): fmc->dirty_size = %u, fmc->free_size = 0x%x\n, fcs1=0x%x, fcs2=0x%x",
3233 fmc->dirty_size, fmc->free_size, jffs_free_size1(fmc), jffs_free_size2(fmc)));
3234 D2(jffs_print_fmcontrol(fmc));
3236 // down(&fmc->gclock);
3238 /* If it is possible to garbage collect, do so. */
3240 while (erased == 0) {
3241 D1(printk("***jffs_garbage_collect_now(): round #%u, "
3242 "fmc->dirty_size = %u\n", i++, fmc->dirty_size));
3243 D2(jffs_print_fmcontrol(fmc));
3245 if ((erased = jffs_try_to_erase(c)) < 0) {
3246 printk(KERN_WARNING "JFFS: Error in "
3247 "garbage collector.\n");
3254 if (fmc->free_size == 0) {
3256 printk(KERN_ERR "jffs_garbage_collect_now(): free_size == 0. This is BAD.\n");
3261 if (fmc->dirty_size < fmc->sector_size) {
3262 /* Actually, we _may_ have been able to free some,
3263 * if there are many overlapping nodes which aren't
3264 * actually marked dirty because they still have
3265 * some valid data in each.
3271 /* Let's dare to make a garbage collect. */
3272 if ((result = jffs_garbage_collect_next(c)) < 0) {
3273 printk(KERN_ERR "JFFS: Something "
3274 "has gone seriously wrong "
3275 "with a garbage collect.\n");
3279 D1(printk(" jffs_garbage_collect_now(): erased: %ld\n", erased));
3280 DJM(jffs_print_memory_allocation_statistics());
3284 // up(&fmc->gclock);
3286 D3(printk(" jffs_garbage_collect_now(): Leaving...\n"));
3288 printk("jffs_g_c_now(): erased = %ld\n", erased);
3289 jffs_print_fmcontrol(fmc);
3292 if (!erased && !result)
3296 } /* jffs_garbage_collect_now() */
3299 /* Determine if it is reasonable to start garbage collection.
3300 We start a gc pass if either:
3301 - The number of free bytes < MIN_FREE_BYTES && at least one
3303 - The number of dirty bytes > MAX_DIRTY_BYTES
3305 static inline int thread_should_wake (struct jffs_control *c)
3307 D1(printk (KERN_NOTICE "thread_should_wake(): free=%d, dirty=%d, blocksize=%d.\n",
3308 c->fmc->free_size, c->fmc->dirty_size, c->fmc->sector_size));
3310 /* If there's not enough dirty space to free a block, there's no point. */
3311 if (c->fmc->dirty_size < c->fmc->sector_size) {
3312 D2(printk(KERN_NOTICE "thread_should_wake(): Not waking. Insufficient dirty space\n"));
3316 /* If there is too much RAM used by the various structures, GC */
3317 if (jffs_get_node_inuse() > (c->fmc->used_size/c->fmc->max_chunk_size * 5 + jffs_get_file_count() * 2 + 50)) {
3318 /* FIXME: Provide proof that this test can be satisfied. We
3319 don't want a filesystem doing endless GC just because this
3320 condition cannot ever be false.
3322 D2(printk(KERN_NOTICE "thread_should_wake(): Waking due to number of nodes\n"));
3326 /* If there are fewer free bytes than the threshold, GC */
3327 if (c->fmc->free_size < c->gc_minfree_threshold) {
3328 D2(printk(KERN_NOTICE "thread_should_wake(): Waking due to insufficent free space\n"));
3331 /* If there are more dirty bytes than the threshold, GC */
3332 if (c->fmc->dirty_size > c->gc_maxdirty_threshold) {
3333 D2(printk(KERN_NOTICE "thread_should_wake(): Waking due to excessive dirty space\n"));
3336 /* FIXME: What about the "There are many versions of a node" condition? */
3342 void jffs_garbage_collect_trigger(struct jffs_control *c)
3344 /* NOTE: We rely on the fact that we have the BKL here.
3345 * Otherwise, the gc_task could go away between the check
3346 * and the wake_up_process()
3348 if (c->gc_task && thread_should_wake(c))
3349 send_sig(SIGHUP, c->gc_task, 1);
3353 /* Kernel threads take (void *) as arguments. Thus we pass
3354 the jffs_control data as a (void *) and then cast it. */
3356 jffs_garbage_collect_thread(void *ptr)
3358 struct jffs_control *c = (struct jffs_control *) ptr;
3359 struct jffs_fmcontrol *fmc = c->fmc;
3364 daemonize("jffs_gcd");
3366 c->gc_task = current;
3369 init_completion(&c->gc_thread_comp); /* barrier */
3370 spin_lock_irq(¤t->sighand->siglock);
3371 siginitsetinv (¤t->blocked, sigmask(SIGHUP) | sigmask(SIGKILL) | sigmask(SIGSTOP) | sigmask(SIGCONT));
3372 recalc_sigpending();
3373 spin_unlock_irq(¤t->sighand->siglock);
3375 D1(printk (KERN_NOTICE "jffs_garbage_collect_thread(): Starting infinite loop.\n"));
3379 /* See if we need to start gc. If we don't, go to sleep.
3381 Current implementation is a BAD THING(tm). If we try
3382 to unmount the FS, the unmount operation will sleep waiting
3383 for this thread to exit. We need to arrange to send it a
3384 sig before the umount process sleeps.
3387 if (!thread_should_wake(c))
3388 set_current_state (TASK_INTERRUPTIBLE);
3390 schedule(); /* Yes, we do this even if we want to go
3391 on immediately - we're a low priority
3394 /* Put_super will send a SIGKILL and then wait on the sem.
3396 while (signal_pending(current)) {
3398 unsigned long signr = 0;
3400 spin_lock_irq(¤t->sighand->siglock);
3401 signr = dequeue_signal(current, ¤t->blocked, &info);
3402 spin_unlock_irq(¤t->sighand->siglock);
3406 D1(printk("jffs_garbage_collect_thread(): SIGSTOP received.\n"));
3407 set_current_state(TASK_STOPPED);
3412 D1(printk("jffs_garbage_collect_thread(): SIGKILL received.\n"));
3414 complete_and_exit(&c->gc_thread_comp, 0);
3419 D1(printk (KERN_NOTICE "jffs_garbage_collect_thread(): collecting.\n"));
3421 D3(printk (KERN_NOTICE "g_c_thread(): down biglock\n"));
3422 down(&fmc->biglock);
3424 D1(printk("***jffs_garbage_collect_thread(): round #%u, "
3425 "fmc->dirty_size = %u\n", i++, fmc->dirty_size));
3426 D2(jffs_print_fmcontrol(fmc));
3428 if ((erased = jffs_try_to_erase(c)) < 0) {
3429 printk(KERN_WARNING "JFFS: Error in "
3430 "garbage collector: %ld.\n", erased);
3436 if (fmc->free_size == 0) {
3437 /* Argh. Might as well commit suicide. */
3438 printk(KERN_ERR "jffs_garbage_collect_thread(): free_size == 0. This is BAD.\n");
3439 send_sig(SIGQUIT, c->gc_task, 1);
3444 /* Let's dare to make a garbage collect. */
3445 if ((result = jffs_garbage_collect_next(c)) < 0) {
3446 printk(KERN_ERR "JFFS: Something "
3447 "has gone seriously wrong "
3448 "with a garbage collect: %d\n", result);
3452 D3(printk (KERN_NOTICE "g_c_thread(): up biglock\n"));
3455 } /* jffs_garbage_collect_thread() */