aboutsummaryrefslogtreecommitdiff
path: root/fs/ubifs/journal.c
diff options
context:
space:
mode:
Diffstat (limited to 'fs/ubifs/journal.c')
-rw-r--r--fs/ubifs/journal.c106
1 files changed, 62 insertions, 44 deletions
diff --git a/fs/ubifs/journal.c b/fs/ubifs/journal.c
index a11ca0958a2..0e045e75abd 100644
--- a/fs/ubifs/journal.c
+++ b/fs/ubifs/journal.c
@@ -114,7 +114,7 @@ static inline void zero_trun_node_unused(struct ubifs_trun_node *trun)
*/
static int reserve_space(struct ubifs_info *c, int jhead, int len)
{
- int err = 0, err1, retries = 0, avail, lnum, offs, free, squeeze;
+ int err = 0, err1, retries = 0, avail, lnum, offs, squeeze;
struct ubifs_wbuf *wbuf = &c->jheads[jhead].wbuf;
/*
@@ -122,11 +122,12 @@ static int reserve_space(struct ubifs_info *c, int jhead, int len)
* better to try to allocate space at the ends of eraseblocks. This is
* what the squeeze parameter does.
*/
+ ubifs_assert(!c->ro_media && !c->ro_mount);
squeeze = (jhead == BASEHD);
again:
mutex_lock_nested(&wbuf->io_mutex, wbuf->jhead);
- if (c->ro_media) {
+ if (c->ro_error) {
err = -EROFS;
goto out_unlock;
}
@@ -139,16 +140,9 @@ again:
* Write buffer wasn't seek'ed or there is no enough space - look for an
* LEB with some empty space.
*/
- lnum = ubifs_find_free_space(c, len, &free, squeeze);
- if (lnum >= 0) {
- /* Found an LEB, add it to the journal head */
- offs = c->leb_size - free;
- err = ubifs_add_bud_to_log(c, jhead, lnum, offs);
- if (err)
- goto out_return;
- /* A new bud was successfully allocated and added to the log */
+ lnum = ubifs_find_free_space(c, len, &offs, squeeze);
+ if (lnum >= 0)
goto out;
- }
err = lnum;
if (err != -ENOSPC)
@@ -159,7 +153,7 @@ again:
* some. But the write-buffer mutex has to be unlocked because
* GC also takes it.
*/
- dbg_jnl("no free space jhead %d, run GC", jhead);
+ dbg_jnl("no free space in jhead %s, run GC", dbg_jhead(jhead));
mutex_unlock(&wbuf->io_mutex);
lnum = ubifs_garbage_collect(c, 0);
@@ -174,7 +168,8 @@ again:
* because we dropped @wbuf->io_mutex, so try once
* again.
*/
- dbg_jnl("GC couldn't make a free LEB for jhead %d", jhead);
+ dbg_jnl("GC couldn't make a free LEB for jhead %s",
+ dbg_jhead(jhead));
if (retries++ < 2) {
dbg_jnl("retry (%d)", retries);
goto again;
@@ -185,7 +180,7 @@ again:
}
mutex_lock_nested(&wbuf->io_mutex, wbuf->jhead);
- dbg_jnl("got LEB %d for jhead %d", lnum, jhead);
+ dbg_jnl("got LEB %d for jhead %s", lnum, dbg_jhead(jhead));
avail = c->leb_size - wbuf->offs - wbuf->used;
if (wbuf->lnum != -1 && avail >= len) {
@@ -202,13 +197,24 @@ again:
return 0;
}
- err = ubifs_add_bud_to_log(c, jhead, lnum, 0);
- if (err)
- goto out_return;
offs = 0;
out:
- err = ubifs_wbuf_seek_nolock(wbuf, lnum, offs, wbuf->dtype);
+ /*
+ * Make sure we synchronize the write-buffer before we add the new bud
+ * to the log. Otherwise we may have a power cut after the log
+ * reference node for the last bud (@lnum) is written but before the
+ * write-buffer data are written to the next-to-last bud
+ * (@wbuf->lnum). And the effect would be that the recovery would see
+ * that there is corruption in the next-to-last bud.
+ */
+ err = ubifs_wbuf_sync_nolock(wbuf);
+ if (err)
+ goto out_return;
+ err = ubifs_add_bud_to_log(c, jhead, lnum, offs);
+ if (err)
+ goto out_return;
+ err = ubifs_wbuf_seek_nolock(wbuf, lnum, offs);
if (err)
goto out_unlock;
@@ -256,7 +262,8 @@ static int write_node(struct ubifs_info *c, int jhead, void *node, int len,
*lnum = c->jheads[jhead].wbuf.lnum;
*offs = c->jheads[jhead].wbuf.offs + c->jheads[jhead].wbuf.used;
- dbg_jnl("jhead %d, LEB %d:%d, len %d", jhead, *lnum, *offs, len);
+ dbg_jnl("jhead %s, LEB %d:%d, len %d",
+ dbg_jhead(jhead), *lnum, *offs, len);
ubifs_prepare_node(c, node, len, 0);
return ubifs_wbuf_write_nolock(wbuf, node, len);
@@ -286,7 +293,8 @@ static int write_head(struct ubifs_info *c, int jhead, void *buf, int len,
*lnum = c->jheads[jhead].wbuf.lnum;
*offs = c->jheads[jhead].wbuf.offs + c->jheads[jhead].wbuf.used;
- dbg_jnl("jhead %d, LEB %d:%d, len %d", jhead, *lnum, *offs, len);
+ dbg_jnl("jhead %s, LEB %d:%d, len %d",
+ dbg_jhead(jhead), *lnum, *offs, len);
err = ubifs_wbuf_write_nolock(wbuf, buf, len);
if (err)
@@ -377,11 +385,9 @@ out:
if (err == -ENOSPC) {
/* This are some budgeting problems, print useful information */
down_write(&c->commit_sem);
- spin_lock(&c->space_lock);
- dbg_dump_stack();
- dbg_dump_budg(c);
- spin_unlock(&c->space_lock);
- dbg_dump_lprops(c);
+ dump_stack();
+ ubifs_dump_budg(c, &c->bi);
+ ubifs_dump_lprops(c);
cmt_retries = dbg_check_lprops(c);
up_write(&c->commit_sem);
}
@@ -463,8 +469,8 @@ static void pack_inode(struct ubifs_info *c, struct ubifs_ino_node *ino,
ino->ctime_nsec = cpu_to_le32(inode->i_ctime.tv_nsec);
ino->mtime_sec = cpu_to_le64(inode->i_mtime.tv_sec);
ino->mtime_nsec = cpu_to_le32(inode->i_mtime.tv_nsec);
- ino->uid = cpu_to_le32(inode->i_uid);
- ino->gid = cpu_to_le32(inode->i_gid);
+ ino->uid = cpu_to_le32(i_uid_read(inode));
+ ino->gid = cpu_to_le32(i_gid_read(inode));
ino->mode = cpu_to_le32(inode->i_mode);
ino->flags = cpu_to_le32(ui->flags);
ino->size = cpu_to_le64(ui->ui_size);
@@ -663,6 +669,7 @@ out_free:
out_release:
release_head(c, BASEHD);
+ kfree(dent);
out_ro:
ubifs_ro_mode(c, err);
if (last_reference)
@@ -687,17 +694,26 @@ int ubifs_jnl_write_data(struct ubifs_info *c, const struct inode *inode,
{
struct ubifs_data_node *data;
int err, lnum, offs, compr_type, out_len;
- int dlen = UBIFS_DATA_NODE_SZ + UBIFS_BLOCK_SIZE * WORST_COMPR_FACTOR;
+ int dlen = COMPRESSED_DATA_NODE_BUF_SZ, allocated = 1;
struct ubifs_inode *ui = ubifs_inode(inode);
- dbg_jnl("ino %lu, blk %u, len %d, key %s",
- (unsigned long)key_inum(c, key), key_block(c, key), len,
- DBGKEY(key));
+ dbg_jnlk(key, "ino %lu, blk %u, len %d, key ",
+ (unsigned long)key_inum(c, key), key_block(c, key), len);
ubifs_assert(len <= UBIFS_BLOCK_SIZE);
- data = kmalloc(dlen, GFP_NOFS);
- if (!data)
- return -ENOMEM;
+ data = kmalloc(dlen, GFP_NOFS | __GFP_NOWARN);
+ if (!data) {
+ /*
+ * Fall-back to the write reserve buffer. Note, we might be
+ * currently on the memory reclaim path, when the kernel is
+ * trying to free some memory by writing out dirty pages. The
+ * write reserve buffer helps us to guarantee that we are
+ * always able to write the data.
+ */
+ allocated = 0;
+ mutex_lock(&c->write_reserve_mutex);
+ data = c->write_reserve_buf;
+ }
data->ch.node_type = UBIFS_DATA_NODE;
key_write(c, key, &data->key);
@@ -733,7 +749,10 @@ int ubifs_jnl_write_data(struct ubifs_info *c, const struct inode *inode,
goto out_ro;
finish_reservation(c);
- kfree(data);
+ if (!allocated)
+ mutex_unlock(&c->write_reserve_mutex);
+ else
+ kfree(data);
return 0;
out_release:
@@ -742,7 +761,10 @@ out_ro:
ubifs_ro_mode(c, err);
finish_reservation(c);
out_free:
- kfree(data);
+ if (!allocated)
+ mutex_unlock(&c->write_reserve_mutex);
+ else
+ kfree(data);
return err;
}
@@ -911,10 +933,8 @@ int ubifs_jnl_rename(struct ubifs_info *c, const struct inode *old_dir,
int move = (old_dir != new_dir);
struct ubifs_inode *uninitialized_var(new_ui);
- dbg_jnl("dent '%.*s' in dir ino %lu to dent '%.*s' in dir ino %lu",
- old_dentry->d_name.len, old_dentry->d_name.name,
- old_dir->i_ino, new_dentry->d_name.len,
- new_dentry->d_name.name, new_dir->i_ino);
+ dbg_jnl("dent '%pd' in dir ino %lu to dent '%pd' in dir ino %lu",
+ old_dentry, old_dir->i_ino, new_dentry, new_dir->i_ino);
ubifs_assert(ubifs_inode(old_dir)->data_len == 0);
ubifs_assert(ubifs_inode(new_dir)->data_len == 0);
ubifs_assert(mutex_is_locked(&ubifs_inode(old_dir)->ui_mutex));
@@ -1154,7 +1174,7 @@ int ubifs_jnl_truncate(struct ubifs_info *c, const struct inode *inode,
dn = (void *)trun + UBIFS_TRUN_NODE_SZ;
blk = new_size >> UBIFS_BLOCK_SHIFT;
data_key_init(c, &key, inum, blk);
- dbg_jnl("last block key %s", DBGKEY(&key));
+ dbg_jnlk(&key, "last block key ");
err = ubifs_tnc_lookup(c, &key, dn);
if (err == -ENOENT)
dlen = 0; /* Not found (so it is a hole) */
@@ -1245,7 +1265,6 @@ out_free:
return err;
}
-#ifdef CONFIG_UBIFS_FS_XATTR
/**
* ubifs_jnl_delete_xattr - delete an extended attribute.
@@ -1366,7 +1385,7 @@ out_ro:
* @host: host inode
*
* This function writes the updated version of an extended attribute inode and
- * the host inode tho the journal (to the base head). The host inode is written
+ * the host inode to the journal (to the base head). The host inode is written
* after the extended attribute inode in order to guarantee that the extended
* attribute will be flushed when the inode is synchronized by 'fsync()' and
* consequently, the write-buffer is synchronized. This function returns zero
@@ -1440,4 +1459,3 @@ out_free:
return err;
}
-#endif /* CONFIG_UBIFS_FS_XATTR */
generated by cgit v1.2.3-70-g09d2 (git 2.46.0) at 2025-12-07 04:51:13 +0000