Minki/linux
1
0
Fork 1
mirror of https://github.com/torvalds/linux.git synced 2024-11-27 14:41:39 +00:00
Code Issues Packages Projects Releases Wiki Activity
a9280fed38
linux/fs/reiserfs/journal.c
Frederic Weisbecker 48f6ba5e69 kill-the-bkl/reiserfs: fix reiserfs lock to cpu_add_remove_lock dependency 
While creating the reiserfs workqueue during the journal
initialization, we are holding the reiserfs lock, but
create_workqueue() also holds the cpu_add_remove_lock, creating
then the following dependency:

- reiserfs lock -> cpu_add_remove_lock

But we also have the following existing dependencies:

- mm->mmap_sem -> reiserfs lock
- cpu_add_remove_lock -> cpu_hotplug.lock -> slub_lock -> sysfs_mutex

The merged dependency chain then becomes:

- mm->mmap_sem -> reiserfs lock -> cpu_add_remove_lock ->
	cpu_hotplug.lock -> slub_lock -> sysfs_mutex

But when we fill a dir entry in sysfs_readir(), we are holding the
sysfs_mutex and we also might fault while copying the directory entry
to the user, leading to the following dependency:

- sysfs_mutex -> mm->mmap_sem

The end result is then a lock inversion between sysfs_mutex and
mm->mmap_sem, as reported in the following lockdep warning:

[ INFO: possible circular locking dependency detected ]
2.6.31-07095-g25a3912 #4
-------------------------------------------------------
udevadm/790 is trying to acquire lock:
 (&mm->mmap_sem){++++++}, at: [<c1098942>] might_fault+0x72/0xc0

but task is already holding lock:
 (sysfs_mutex){+.+.+.}, at: [<c110813c>] sysfs_readdir+0x7c/0x260

which lock already depends on the new lock.

the existing dependency chain (in reverse order) is:

-> #5 (sysfs_mutex){+.+.+.}:
      [...]

-> #4 (slub_lock){+++++.}:
      [...]

-> #3 (cpu_hotplug.lock){+.+.+.}:
      [...]

-> #2 (cpu_add_remove_lock){+.+.+.}:
      [...]

-> #1 (&REISERFS_SB(s)->lock){+.+.+.}:
      [...]

-> #0 (&mm->mmap_sem){++++++}:
      [...]

This can be fixed by relaxing the reiserfs lock while creating the
workqueue.
This is fine to relax the lock here, we just keep it around to pass
through reiserfs lock checks and for paranoid reasons.

Reported-by: Alexander Beregalov <a.beregalov@gmail.com>
Tested-by: Alexander Beregalov <a.beregalov@gmail.com>
Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jeff Mahoney <jeffm@suse.com>
Cc: Chris Mason <chris.mason@oracle.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Alexander Beregalov <a.beregalov@gmail.com>
Cc: Laurent Riffard <laurent.riffard@free.fr>
2009-10-05 16:31:37 +02:00

4379 lines
124 KiB
C
Raw Blame History

/*
** Write ahead logging implementation copyright Chris Mason 2000
**
** The background commits make this code very interelated, and
** overly complex. I need to rethink things a bit....The major players:
**
** journal_begin -- call with the number of blocks you expect to log.
** If the current transaction is too
** old, it will block until the current transaction is
** finished, and then start a new one.
** Usually, your transaction will get joined in with
** previous ones for speed.
**
** journal_join -- same as journal_begin, but won't block on the current
** transaction regardless of age. Don't ever call
** this. Ever. There are only two places it should be
** called from, and they are both inside this file.
**
** journal_mark_dirty -- adds blocks into this transaction. clears any flags
** that might make them get sent to disk
** and then marks them BH_JDirty. Puts the buffer head
** into the current transaction hash.
**
** journal_end -- if the current transaction is batchable, it does nothing
** otherwise, it could do an async/synchronous commit, or
** a full flush of all log and real blocks in the
** transaction.
**
** flush_old_commits -- if the current transaction is too old, it is ended and
** commit blocks are sent to disk. Forces commit blocks
** to disk for all backgrounded commits that have been
** around too long.
** -- Note, if you call this as an immediate flush from
** from within kupdate, it will ignore the immediate flag
*/
#include <linux/time.h>
#include <linux/semaphore.h>
#include <linux/vmalloc.h>
#include <linux/reiserfs_fs.h>
#include <linux/kernel.h>
#include <linux/errno.h>
#include <linux/fcntl.h>
#include <linux/stat.h>
#include <linux/string.h>
#include <linux/smp_lock.h>
#include <linux/buffer_head.h>
#include <linux/workqueue.h>
#include <linux/writeback.h>
#include <linux/blkdev.h>
#include <linux/backing-dev.h>
#include <linux/uaccess.h>
#include <asm/system.h>
/* gets a struct reiserfs_journal_list * from a list head */
#define JOURNAL_LIST_ENTRY(h) (list_entry((h), struct reiserfs_journal_list, \
j_list))
#define JOURNAL_WORK_ENTRY(h) (list_entry((h), struct reiserfs_journal_list, \
j_working_list))
/* the number of mounted filesystems. This is used to decide when to
** start and kill the commit workqueue
*/
static int reiserfs_mounted_fs_count;
static struct workqueue_struct *commit_wq;
#define JOURNAL_TRANS_HALF 1018 /* must be correct to keep the desc and commit
structs at 4k */
#define BUFNR 64 /*read ahead */
/* cnode stat bits. Move these into reiserfs_fs.h */
#define BLOCK_FREED 2 /* this block was freed, and can't be written. */
#define BLOCK_FREED_HOLDER 3 /* this block was freed during this transaction, and can't be written */
#define BLOCK_NEEDS_FLUSH 4 /* used in flush_journal_list */
#define BLOCK_DIRTIED 5
/* journal list state bits */
#define LIST_TOUCHED 1
#define LIST_DIRTY 2
#define LIST_COMMIT_PENDING 4 /* someone will commit this list */
/* flags for do_journal_end */
#define FLUSH_ALL 1 /* flush commit and real blocks */
#define COMMIT_NOW 2 /* end and commit this transaction */
#define WAIT 4 /* wait for the log blocks to hit the disk */
static int do_journal_end(struct reiserfs_transaction_handle *,
struct super_block *, unsigned long nblocks,
int flags);
static int flush_journal_list(struct super_block *s,
struct reiserfs_journal_list *jl, int flushall);
static int flush_commit_list(struct super_block *s,
struct reiserfs_journal_list *jl, int flushall);
static int can_dirty(struct reiserfs_journal_cnode *cn);
static int journal_join(struct reiserfs_transaction_handle *th,
struct super_block *sb, unsigned long nblocks);
static int release_journal_dev(struct super_block *super,
struct reiserfs_journal *journal);
static int dirty_one_transaction(struct super_block *s,
struct reiserfs_journal_list *jl);
static void flush_async_commits(struct work_struct *work);
static void queue_log_writer(struct super_block *s);
/* values for join in do_journal_begin_r */
enum {
JBEGIN_REG = 0, /* regular journal begin */
JBEGIN_JOIN = 1, /* join the running transaction if at all possible */
JBEGIN_ABORT = 2, /* called from cleanup code, ignores aborted flag */
};
static int do_journal_begin_r(struct reiserfs_transaction_handle *th,
struct super_block *sb,
unsigned long nblocks, int join);
static void init_journal_hash(struct super_block *sb)
{
struct reiserfs_journal *journal = SB_JOURNAL(sb);
memset(journal->j_hash_table, 0,
JOURNAL_HASH_SIZE * sizeof(struct reiserfs_journal_cnode *));
}
/*
** clears BH_Dirty and sticks the buffer on the clean list. Called because I can't allow refile_buffer to
** make schedule happen after I've freed a block. Look at remove_from_transaction and journal_mark_freed for
** more details.
*/
static int reiserfs_clean_and_file_buffer(struct buffer_head *bh)
{
if (bh) {
clear_buffer_dirty(bh);
clear_buffer_journal_test(bh);
}
return 0;
}
static void disable_barrier(struct super_block *s)
{
REISERFS_SB(s)->s_mount_opt &= ~(1 << REISERFS_BARRIER_FLUSH);
printk("reiserfs: disabling flush barriers on %s\n",
reiserfs_bdevname(s));
}
static struct reiserfs_bitmap_node *allocate_bitmap_node(struct super_block
*sb)
{
struct reiserfs_bitmap_node *bn;
static int id;
bn = kmalloc(sizeof(struct reiserfs_bitmap_node), GFP_NOFS);
if (!bn) {
return NULL;
}
bn->data = kzalloc(sb->s_blocksize, GFP_NOFS);
if (!bn->data) {
kfree(bn);
return NULL;
}
bn->id = id++;
INIT_LIST_HEAD(&bn->list);
return bn;
}
static struct reiserfs_bitmap_node *get_bitmap_node(struct super_block *sb)
{
struct reiserfs_journal *journal = SB_JOURNAL(sb);
struct reiserfs_bitmap_node *bn = NULL;
struct list_head *entry = journal->j_bitmap_nodes.next;
journal->j_used_bitmap_nodes++;
repeat:
if (entry != &journal->j_bitmap_nodes) {
bn = list_entry(entry, struct reiserfs_bitmap_node, list);
list_del(entry);
memset(bn->data, 0, sb->s_blocksize);
journal->j_free_bitmap_nodes--;
return bn;
}
bn = allocate_bitmap_node(sb);
if (!bn) {
yield();
goto repeat;
}
return bn;
}
static inline void free_bitmap_node(struct super_block *sb,
struct reiserfs_bitmap_node *bn)
{
struct reiserfs_journal *journal = SB_JOURNAL(sb);
journal->j_used_bitmap_nodes--;
if (journal->j_free_bitmap_nodes > REISERFS_MAX_BITMAP_NODES) {
kfree(bn->data);
kfree(bn);
} else {
list_add(&bn->list, &journal->j_bitmap_nodes);
journal->j_free_bitmap_nodes++;
}
}
static void allocate_bitmap_nodes(struct super_block *sb)
{
int i;
struct reiserfs_journal *journal = SB_JOURNAL(sb);
struct reiserfs_bitmap_node *bn = NULL;
for (i = 0; i < REISERFS_MIN_BITMAP_NODES; i++) {
bn = allocate_bitmap_node(sb);
if (bn) {
list_add(&bn->list, &journal->j_bitmap_nodes);
journal->j_free_bitmap_nodes++;
} else {
break; /* this is ok, we'll try again when more are needed */
}
}
}
static int set_bit_in_list_bitmap(struct super_block *sb,
b_blocknr_t block,
struct reiserfs_list_bitmap *jb)
{
unsigned int bmap_nr = block / (sb->s_blocksize << 3);
unsigned int bit_nr = block % (sb->s_blocksize << 3);
if (!jb->bitmaps[bmap_nr]) {
jb->bitmaps[bmap_nr] = get_bitmap_node(sb);
}
set_bit(bit_nr, (unsigned long *)jb->bitmaps[bmap_nr]->data);
return 0;
}
static void cleanup_bitmap_list(struct super_block *sb,
struct reiserfs_list_bitmap *jb)
{
int i;
if (jb->bitmaps == NULL)
return;
for (i = 0; i < reiserfs_bmap_count(sb); i++) {
if (jb->bitmaps[i]) {
free_bitmap_node(sb, jb->bitmaps[i]);
jb->bitmaps[i] = NULL;
}
}
}
/*
** only call this on FS unmount.
*/
static int free_list_bitmaps(struct super_block *sb,
struct reiserfs_list_bitmap *jb_array)
{
int i;
struct reiserfs_list_bitmap *jb;
for (i = 0; i < JOURNAL_NUM_BITMAPS; i++) {
jb = jb_array + i;
jb->journal_list = NULL;
cleanup_bitmap_list(sb, jb);
vfree(jb->bitmaps);
jb->bitmaps = NULL;
}
return 0;
}
static int free_bitmap_nodes(struct super_block *sb)
{
struct reiserfs_journal *journal = SB_JOURNAL(sb);
struct list_head *next = journal->j_bitmap_nodes.next;
struct reiserfs_bitmap_node *bn;
while (next != &journal->j_bitmap_nodes) {
bn = list_entry(next, struct reiserfs_bitmap_node, list);
list_del(next);
kfree(bn->data);
kfree(bn);
next = journal->j_bitmap_nodes.next;
journal->j_free_bitmap_nodes--;
}
return 0;
}
/*
** get memory for JOURNAL_NUM_BITMAPS worth of bitmaps.
** jb_array is the array to be filled in.
*/
int reiserfs_allocate_list_bitmaps(struct super_block *sb,
struct reiserfs_list_bitmap *jb_array,
unsigned int bmap_nr)
{
int i;
int failed = 0;
struct reiserfs_list_bitmap *jb;
int mem = bmap_nr * sizeof(struct reiserfs_bitmap_node *);
for (i = 0; i < JOURNAL_NUM_BITMAPS; i++) {
jb = jb_array + i;
jb->journal_list = NULL;
jb->bitmaps = vmalloc(mem);
if (!jb->bitmaps) {
reiserfs_warning(sb, "clm-2000", "unable to "
"allocate bitmaps for journal lists");
failed = 1;
break;
}
memset(jb->bitmaps, 0, mem);
}
if (failed) {
free_list_bitmaps(sb, jb_array);
return -1;
}
return 0;
}
/*
** find an available list bitmap. If you can't find one, flush a commit list
** and try again
*/
static struct reiserfs_list_bitmap *get_list_bitmap(struct super_block *sb,
struct reiserfs_journal_list
*jl)
{
int i, j;
struct reiserfs_journal *journal = SB_JOURNAL(sb);
struct reiserfs_list_bitmap *jb = NULL;
for (j = 0; j < (JOURNAL_NUM_BITMAPS * 3); j++) {
i = journal->j_list_bitmap_index;
journal->j_list_bitmap_index = (i + 1) % JOURNAL_NUM_BITMAPS;
jb = journal->j_list_bitmap + i;
if (journal->j_list_bitmap[i].journal_list) {
flush_commit_list(sb,
journal->j_list_bitmap[i].
journal_list, 1);
if (!journal->j_list_bitmap[i].journal_list) {
break;
}
} else {
break;
}
}
if (jb->journal_list) { /* double check to make sure if flushed correctly */
return NULL;
}
jb->journal_list = jl;
return jb;
}
/*
** allocates a new chunk of X nodes, and links them all together as a list.
** Uses the cnode->next and cnode->prev pointers
** returns NULL on failure
*/
static struct reiserfs_journal_cnode *allocate_cnodes(int num_cnodes)
{
struct reiserfs_journal_cnode *head;
int i;
if (num_cnodes <= 0) {
return NULL;
}
head = vmalloc(num_cnodes * sizeof(struct reiserfs_journal_cnode));
if (!head) {
return NULL;
}
memset(head, 0, num_cnodes * sizeof(struct reiserfs_journal_cnode));
head[0].prev = NULL;
head[0].next = head + 1;
for (i = 1; i < num_cnodes; i++) {
head[i].prev = head + (i - 1);
head[i].next = head + (i + 1); /* if last one, overwrite it after the if */
}
head[num_cnodes - 1].next = NULL;
return head;
}
/*
** pulls a cnode off the free list, or returns NULL on failure
*/
static struct reiserfs_journal_cnode *get_cnode(struct super_block *sb)
{
struct reiserfs_journal_cnode *cn;
struct reiserfs_journal *journal = SB_JOURNAL(sb);
reiserfs_check_lock_depth(sb, "get_cnode");
if (journal->j_cnode_free <= 0) {
return NULL;
}
journal->j_cnode_used++;
journal->j_cnode_free--;
cn = journal->j_cnode_free_list;
if (!cn) {
return cn;
}
if (cn->next) {
cn->next->prev = NULL;
}
journal->j_cnode_free_list = cn->next;
memset(cn, 0, sizeof(struct reiserfs_journal_cnode));
return cn;
}
/*
** returns a cnode to the free list
*/
static void free_cnode(struct super_block *sb,
struct reiserfs_journal_cnode *cn)
{
struct reiserfs_journal *journal = SB_JOURNAL(sb);
reiserfs_check_lock_depth(sb, "free_cnode");
journal->j_cnode_used--;
journal->j_cnode_free++;
/* memset(cn, 0, sizeof(struct reiserfs_journal_cnode)) ; */
cn->next = journal->j_cnode_free_list;
if (journal->j_cnode_free_list) {
journal->j_cnode_free_list->prev = cn;
}
cn->prev = NULL; /* not needed with the memset, but I might kill the memset, and forget to do this */
journal->j_cnode_free_list = cn;
}
static void clear_prepared_bits(struct buffer_head *bh)
{
clear_buffer_journal_prepared(bh);
clear_buffer_journal_restore_dirty(bh);
}
/* return a cnode with same dev, block number and size in table, or null if not found */
static inline struct reiserfs_journal_cnode *get_journal_hash_dev(struct
super_block
*sb,
struct
reiserfs_journal_cnode
**table,
long bl)
{
struct reiserfs_journal_cnode *cn;
cn = journal_hash(table, sb, bl);
while (cn) {
if (cn->blocknr == bl && cn->sb == sb)
return cn;
cn = cn->hnext;
}
return (struct reiserfs_journal_cnode *)0;
}
/*
** this actually means 'can this block be reallocated yet?'. If you set search_all, a block can only be allocated
** if it is not in the current transaction, was not freed by the current transaction, and has no chance of ever
** being overwritten by a replay after crashing.
**
** If you don't set search_all, a block can only be allocated if it is not in the current transaction. Since deleting
** a block removes it from the current transaction, this case should never happen. If you don't set search_all, make
** sure you never write the block without logging it.
**
** next_zero_bit is a suggestion about the next block to try for find_forward.
** when bl is rejected because it is set in a journal list bitmap, we search
** for the next zero bit in the bitmap that rejected bl. Then, we return that
** through next_zero_bit for find_forward to try.
**
** Just because we return something in next_zero_bit does not mean we won't
** reject it on the next call to reiserfs_in_journal
**
*/
int reiserfs_in_journal(struct super_block *sb,
unsigned int bmap_nr, int bit_nr, int search_all,
b_blocknr_t * next_zero_bit)
{
struct reiserfs_journal *journal = SB_JOURNAL(sb);
struct reiserfs_journal_cnode *cn;
struct reiserfs_list_bitmap *jb;
int i;
unsigned long bl;
*next_zero_bit = 0; /* always start this at zero. */
PROC_INFO_INC(sb, journal.in_journal);
/* If we aren't doing a search_all, this is a metablock, and it will be logged before use.
** if we crash before the transaction that freed it commits, this transaction won't
** have committed either, and the block will never be written
*/
if (search_all) {
for (i = 0; i < JOURNAL_NUM_BITMAPS; i++) {
PROC_INFO_INC(sb, journal.in_journal_bitmap);
jb = journal->j_list_bitmap + i;
if (jb->journal_list && jb->bitmaps[bmap_nr] &&
test_bit(bit_nr,
(unsigned long *)jb->bitmaps[bmap_nr]->
data)) {
*next_zero_bit =
find_next_zero_bit((unsigned long *)
(jb->bitmaps[bmap_nr]->
data),
sb->s_blocksize << 3,
bit_nr + 1);
return 1;
}
}
}
bl = bmap_nr * (sb->s_blocksize << 3) + bit_nr;
/* is it in any old transactions? */
if (search_all
&& (cn =
get_journal_hash_dev(sb, journal->j_list_hash_table, bl))) {
return 1;
}
/* is it in the current transaction. This should never happen */
if ((cn = get_journal_hash_dev(sb, journal->j_hash_table, bl))) {
BUG();
return 1;
}
PROC_INFO_INC(sb, journal.in_journal_reusable);
/* safe for reuse */
return 0;
}
/* insert cn into table
*/
static inline void insert_journal_hash(struct reiserfs_journal_cnode **table,
struct reiserfs_journal_cnode *cn)
{
struct reiserfs_journal_cnode *cn_orig;
cn_orig = journal_hash(table, cn->sb, cn->blocknr);
cn->hnext = cn_orig;
cn->hprev = NULL;
if (cn_orig) {
cn_orig->hprev = cn;
}
journal_hash(table, cn->sb, cn->blocknr) = cn;
}
/* lock the current transaction */
static inline void lock_journal(struct super_block *sb)
{
PROC_INFO_INC(sb, journal.lock_journal);
reiserfs_mutex_lock_safe(&SB_JOURNAL(sb)->j_mutex, sb);
}
/* unlock the current transaction */
static inline void unlock_journal(struct super_block *sb)
{
mutex_unlock(&SB_JOURNAL(sb)->j_mutex);
}
static inline void get_journal_list(struct reiserfs_journal_list *jl)
{
jl->j_refcount++;
}
static inline void put_journal_list(struct super_block *s,
struct reiserfs_journal_list *jl)
{
if (jl->j_refcount < 1) {
reiserfs_panic(s, "journal-2", "trans id %u, refcount at %d",
jl->j_trans_id, jl->j_refcount);
}
if (--jl->j_refcount == 0)
kfree(jl);
}
/*
** this used to be much more involved, and I'm keeping it just in case things get ugly again.
** it gets called by flush_commit_list, and cleans up any data stored about blocks freed during a
** transaction.
*/
static void cleanup_freed_for_journal_list(struct super_block *sb,
struct reiserfs_journal_list *jl)
{
struct reiserfs_list_bitmap *jb = jl->j_list_bitmap;
if (jb) {
cleanup_bitmap_list(sb, jb);
}
jl->j_list_bitmap->journal_list = NULL;
jl->j_list_bitmap = NULL;
}
static int journal_list_still_alive(struct super_block *s,
unsigned int trans_id)
{
struct reiserfs_journal *journal = SB_JOURNAL(s);
struct list_head *entry = &journal->j_journal_list;
struct reiserfs_journal_list *jl;
if (!list_empty(entry)) {
jl = JOURNAL_LIST_ENTRY(entry->next);
if (jl->j_trans_id <= trans_id) {
return 1;
}
}
return 0;
}
/*
* If page->mapping was null, we failed to truncate this page for
* some reason. Most likely because it was truncated after being
* logged via data=journal.
*
* This does a check to see if the buffer belongs to one of these
* lost pages before doing the final put_bh. If page->mapping was
* null, it tries to free buffers on the page, which should make the
* final page_cache_release drop the page from the lru.
*/
static void release_buffer_page(struct buffer_head *bh)
{
struct page *page = bh->b_page;
if (!page->mapping && trylock_page(page)) {
page_cache_get(page);
put_bh(bh);
if (!page->mapping)
try_to_free_buffers(page);
unlock_page(page);
page_cache_release(page);
} else {
put_bh(bh);
}
}
static void reiserfs_end_buffer_io_sync(struct buffer_head *bh, int uptodate)
{
char b[BDEVNAME_SIZE];
if (buffer_journaled(bh)) {
reiserfs_warning(NULL, "clm-2084",
"pinned buffer %lu:%s sent to disk",
bh->b_blocknr, bdevname(bh->b_bdev, b));
}
if (uptodate)
set_buffer_uptodate(bh);
else
clear_buffer_uptodate(bh);
unlock_buffer(bh);
release_buffer_page(bh);
}
static void reiserfs_end_ordered_io(struct buffer_head *bh, int uptodate)
{
if (uptodate)
set_buffer_uptodate(bh);
else
clear_buffer_uptodate(bh);
unlock_buffer(bh);
put_bh(bh);
}
static void submit_logged_buffer(struct buffer_head *bh)
{
get_bh(bh);
bh->b_end_io = reiserfs_end_buffer_io_sync;
clear_buffer_journal_new(bh);
clear_buffer_dirty(bh);
if (!test_clear_buffer_journal_test(bh))
BUG();
if (!buffer_uptodate(bh))
BUG();
submit_bh(WRITE, bh);
}
static void submit_ordered_buffer(struct buffer_head *bh)
{
get_bh(bh);
bh->b_end_io = reiserfs_end_ordered_io;
clear_buffer_dirty(bh);
if (!buffer_uptodate(bh))
BUG();
submit_bh(WRITE, bh);
}
static int submit_barrier_buffer(struct buffer_head *bh)
{
get_bh(bh);
bh->b_end_io = reiserfs_end_ordered_io;
clear_buffer_dirty(bh);
if (!buffer_uptodate(bh))
BUG();
return submit_bh(WRITE_BARRIER, bh);
}
static void check_barrier_completion(struct super_block *s,
struct buffer_head *bh)
{
if (buffer_eopnotsupp(bh)) {
clear_buffer_eopnotsupp(bh);
disable_barrier(s);
set_buffer_uptodate(bh);
set_buffer_dirty(bh);
reiserfs_write_unlock(s);
sync_dirty_buffer(bh);
reiserfs_write_lock(s);
}
}
#define CHUNK_SIZE 32
struct buffer_chunk {
struct buffer_head *bh[CHUNK_SIZE];
int nr;
};
static void write_chunk(struct buffer_chunk *chunk)
{
int i;
get_fs_excl();
for (i = 0; i < chunk->nr; i++) {
submit_logged_buffer(chunk->bh[i]);
}
chunk->nr = 0;
put_fs_excl();
}
static void write_ordered_chunk(struct buffer_chunk *chunk)
{
int i;
get_fs_excl();
for (i = 0; i < chunk->nr; i++) {
submit_ordered_buffer(chunk->bh[i]);
}
chunk->nr = 0;
put_fs_excl();
}
static int add_to_chunk(struct buffer_chunk *chunk, struct buffer_head *bh,
spinlock_t * lock, void (fn) (struct buffer_chunk *))
{
int ret = 0;
BUG_ON(chunk->nr >= CHUNK_SIZE);
chunk->bh[chunk->nr++] = bh;
if (chunk->nr >= CHUNK_SIZE) {
ret = 1;
if (lock)
spin_unlock(lock);
fn(chunk);
if (lock)
spin_lock(lock);
}
return ret;
}
static atomic_t nr_reiserfs_jh = ATOMIC_INIT(0);
static struct reiserfs_jh *alloc_jh(void)
{
struct reiserfs_jh *jh;
while (1) {
jh = kmalloc(sizeof(*jh), GFP_NOFS);
if (jh) {
atomic_inc(&nr_reiserfs_jh);
return jh;
}
yield();
}
}
/*
* we want to free the jh when the buffer has been written
* and waited on
*/
void reiserfs_free_jh(struct buffer_head *bh)
{
struct reiserfs_jh *jh;
jh = bh->b_private;
if (jh) {
bh->b_private = NULL;
jh->bh = NULL;
list_del_init(&jh->list);
kfree(jh);
if (atomic_read(&nr_reiserfs_jh) <= 0)
BUG();
atomic_dec(&nr_reiserfs_jh);
put_bh(bh);
}
}
static inline int __add_jh(struct reiserfs_journal *j, struct buffer_head *bh,
int tail)
{
struct reiserfs_jh *jh;
if (bh->b_private) {
spin_lock(&j->j_dirty_buffers_lock);
if (!bh->b_private) {
spin_unlock(&j->j_dirty_buffers_lock);
goto no_jh;
}
jh = bh->b_private;
list_del_init(&jh->list);
} else {
no_jh:
get_bh(bh);
jh = alloc_jh();
spin_lock(&j->j_dirty_buffers_lock);
/* buffer must be locked for __add_jh, should be able to have
* two adds at the same time
*/
BUG_ON(bh->b_private);
jh->bh = bh;
bh->b_private = jh;
}
jh->jl = j->j_current_jl;
if (tail)
list_add_tail(&jh->list, &jh->jl->j_tail_bh_list);
else {
list_add_tail(&jh->list, &jh->jl->j_bh_list);
}
spin_unlock(&j->j_dirty_buffers_lock);
return 0;
}
int reiserfs_add_tail_list(struct inode *inode, struct buffer_head *bh)
{
return __add_jh(SB_JOURNAL(inode->i_sb), bh, 1);
}
int reiserfs_add_ordered_list(struct inode *inode, struct buffer_head *bh)
{
return __add_jh(SB_JOURNAL(inode->i_sb), bh, 0);
}
#define JH_ENTRY(l) list_entry((l), struct reiserfs_jh, list)
static int write_ordered_buffers(spinlock_t * lock,
struct reiserfs_journal *j,
struct reiserfs_journal_list *jl,
struct list_head *list)
{
struct buffer_head *bh;
struct reiserfs_jh *jh;
int ret = j->j_errno;
struct buffer_chunk chunk;
struct list_head tmp;
INIT_LIST_HEAD(&tmp);
chunk.nr = 0;
spin_lock(lock);
while (!list_empty(list)) {
jh = JH_ENTRY(list->next);
bh = jh->bh;
get_bh(bh);
if (!trylock_buffer(bh)) {
if (!buffer_dirty(bh)) {
list_move(&jh->list, &tmp);
goto loop_next;
}
spin_unlock(lock);
if (chunk.nr)
write_ordered_chunk(&chunk);
wait_on_buffer(bh);
cond_resched();
spin_lock(lock);
goto loop_next;
}
/* in theory, dirty non-uptodate buffers should never get here,
* but the upper layer io error paths still have a few quirks.
* Handle them here as gracefully as we can
*/
if (!buffer_uptodate(bh) && buffer_dirty(bh)) {
clear_buffer_dirty(bh);
ret = -EIO;
}
if (buffer_dirty(bh)) {
list_move(&jh->list, &tmp);
add_to_chunk(&chunk, bh, lock, write_ordered_chunk);
} else {
reiserfs_free_jh(bh);
unlock_buffer(bh);
}
loop_next:
put_bh(bh);
cond_resched_lock(lock);
}
if (chunk.nr) {
spin_unlock(lock);
write_ordered_chunk(&chunk);
spin_lock(lock);
}
while (!list_empty(&tmp)) {
jh = JH_ENTRY(tmp.prev);
bh = jh->bh;
get_bh(bh);
reiserfs_free_jh(bh);
if (buffer_locked(bh)) {
spin_unlock(lock);
wait_on_buffer(bh);
spin_lock(lock);
}
if (!buffer_uptodate(bh)) {
ret = -EIO;
}
/* ugly interaction with invalidatepage here.
* reiserfs_invalidate_page will pin any buffer that has a valid
* journal head from an older transaction. If someone else sets
* our buffer dirty after we write it in the first loop, and
* then someone truncates the page away, nobody will ever write
* the buffer. We're safe if we write the page one last time
* after freeing the journal header.
*/
if (buffer_dirty(bh) && unlikely(bh->b_page->mapping == NULL)) {
spin_unlock(lock);
ll_rw_block(WRITE, 1, &bh);
spin_lock(lock);
}
put_bh(bh);
cond_resched_lock(lock);
}
spin_unlock(lock);
return ret;
}
static int flush_older_commits(struct super_block *s,
struct reiserfs_journal_list *jl)
{
struct reiserfs_journal *journal = SB_JOURNAL(s);
struct reiserfs_journal_list *other_jl;
struct reiserfs_journal_list *first_jl;
struct list_head *entry;
unsigned int trans_id = jl->j_trans_id;
unsigned int other_trans_id;
unsigned int first_trans_id;
find_first:
/*
* first we walk backwards to find the oldest uncommitted transation
*/
first_jl = jl;
entry = jl->j_list.prev;
while (1) {
other_jl = JOURNAL_LIST_ENTRY(entry);
if (entry == &journal->j_journal_list ||
atomic_read(&other_jl->j_older_commits_done))
break;
first_jl = other_jl;
entry = other_jl->j_list.prev;
}
/* if we didn't find any older uncommitted transactions, return now */
if (first_jl == jl) {
return 0;
}
first_trans_id = first_jl->j_trans_id;
entry = &first_jl->j_list;
while (1) {
other_jl = JOURNAL_LIST_ENTRY(entry);
other_trans_id = other_jl->j_trans_id;
if (other_trans_id < trans_id) {
if (atomic_read(&other_jl->j_commit_left) != 0) {
flush_commit_list(s, other_jl, 0);
/* list we were called with is gone, return */
if (!journal_list_still_alive(s, trans_id))
return 1;
/* the one we just flushed is gone, this means all
* older lists are also gone, so first_jl is no longer
* valid either. Go back to the beginning.
*/
if (!journal_list_still_alive
(s, other_trans_id)) {
goto find_first;
}
}
entry = entry->next;
if (entry == &journal->j_journal_list)
return 0;
} else {
return 0;
}
}
return 0;
}
static int reiserfs_async_progress_wait(struct super_block *s)
{
DEFINE_WAIT(wait);
struct reiserfs_journal *j = SB_JOURNAL(s);
if (atomic_read(&j->j_async_throttle)) {
reiserfs_write_unlock(s);
congestion_wait(BLK_RW_ASYNC, HZ / 10);
reiserfs_write_lock(s);
}
return 0;
}
/*
** if this journal list still has commit blocks unflushed, send them to disk.
**
** log areas must be flushed in order (transaction 2 can't commit before transaction 1)
** Before the commit block can by written, every other log block must be safely on disk
**
*/
static int flush_commit_list(struct super_block *s,
struct reiserfs_journal_list *jl, int flushall)
{
int i;
b_blocknr_t bn;
struct buffer_head *tbh = NULL;
unsigned int trans_id = jl->j_trans_id;
struct reiserfs_journal *journal = SB_JOURNAL(s);
int barrier = 0;
int retval = 0;
int write_len;
reiserfs_check_lock_depth(s, "flush_commit_list");
if (atomic_read(&jl->j_older_commits_done)) {
return 0;
}
get_fs_excl();
/* before we can put our commit blocks on disk, we have to make sure everyone older than
** us is on disk too
*/
BUG_ON(jl->j_len <= 0);
BUG_ON(trans_id == journal->j_trans_id);
get_journal_list(jl);
if (flushall) {
if (flush_older_commits(s, jl) == 1) {
/* list disappeared during flush_older_commits. return */
goto put_jl;
}
}
/* make sure nobody is trying to flush this one at the same time */
reiserfs_mutex_lock_safe(&jl->j_commit_mutex, s);
if (!journal_list_still_alive(s, trans_id)) {
mutex_unlock(&jl->j_commit_mutex);
goto put_jl;
}
BUG_ON(jl->j_trans_id == 0);
/* this commit is done, exit */
if (atomic_read(&(jl->j_commit_left)) <= 0) {
if (flushall) {
atomic_set(&(jl->j_older_commits_done), 1);
}
mutex_unlock(&jl->j_commit_mutex);
goto put_jl;
}
if (!list_empty(&jl->j_bh_list)) {
int ret;
/*
* We might sleep in numerous places inside
* write_ordered_buffers. Relax the write lock.
*/
reiserfs_write_unlock(s);
ret = write_ordered_buffers(&journal->j_dirty_buffers_lock,
journal, jl, &jl->j_bh_list);
if (ret < 0 && retval == 0)
retval = ret;
reiserfs_write_lock(s);
}
BUG_ON(!list_empty(&jl->j_bh_list));
/*
* for the description block and all the log blocks, submit any buffers
* that haven't already reached the disk. Try to write at least 256
* log blocks. later on, we will only wait on blocks that correspond
* to this transaction, but while we're unplugging we might as well
* get a chunk of data on there.
*/
atomic_inc(&journal->j_async_throttle);
write_len = jl->j_len + 1;
if (write_len < 256)
write_len = 256;
for (i = 0 ; i < write_len ; i++) {
bn = SB_ONDISK_JOURNAL_1st_BLOCK(s) + (jl->j_start + i) %
SB_ONDISK_JOURNAL_SIZE(s);
tbh = journal_find_get_block(s, bn);
if (tbh) {
if (buffer_dirty(tbh)) {
reiserfs_write_unlock(s);
ll_rw_block(WRITE, 1, &tbh);
reiserfs_write_lock(s);
}
put_bh(tbh) ;
}
}
atomic_dec(&journal->j_async_throttle);
/* We're skipping the commit if there's an error */
if (retval || reiserfs_is_journal_aborted(journal))
barrier = 0;
/* wait on everything written so far before writing the commit
* if we are in barrier mode, send the commit down now
*/
barrier = reiserfs_barrier_flush(s);
if (barrier) {
int ret;
lock_buffer(jl->j_commit_bh);
ret = submit_barrier_buffer(jl->j_commit_bh);
if (ret == -EOPNOTSUPP) {
set_buffer_uptodate(jl->j_commit_bh);
disable_barrier(s);
barrier = 0;
}
}
for (i = 0; i < (jl->j_len + 1); i++) {
bn = SB_ONDISK_JOURNAL_1st_BLOCK(s) +
(jl->j_start + i) % SB_ONDISK_JOURNAL_SIZE(s);
tbh = journal_find_get_block(s, bn);
reiserfs_write_unlock(s);
wait_on_buffer(tbh);
reiserfs_write_lock(s);
// since we're using ll_rw_blk above, it might have skipped over
// a locked buffer. Double check here
//
/* redundant, sync_dirty_buffer() checks */
if (buffer_dirty(tbh)) {
reiserfs_write_unlock(s);
sync_dirty_buffer(tbh);
reiserfs_write_lock(s);
}
if (unlikely(!buffer_uptodate(tbh))) {
#ifdef CONFIG_REISERFS_CHECK
reiserfs_warning(s, "journal-601",
"buffer write failed");
#endif
retval = -EIO;
}
put_bh(tbh); /* once for journal_find_get_block */
put_bh(tbh); /* once due to original getblk in do_journal_end */
atomic_dec(&(jl->j_commit_left));
}
BUG_ON(atomic_read(&(jl->j_commit_left)) != 1);
if (!barrier) {
/* If there was a write error in the journal - we can't commit
* this transaction - it will be invalid and, if successful,
* will just end up propagating the write error out to
* the file system. */
if (likely(!retval && !reiserfs_is_journal_aborted (journal))) {
if (buffer_dirty(jl->j_commit_bh))
BUG();
mark_buffer_dirty(jl->j_commit_bh) ;
reiserfs_write_unlock(s);
sync_dirty_buffer(jl->j_commit_bh) ;
reiserfs_write_lock(s);
}
} else {
reiserfs_write_unlock(s);
wait_on_buffer(jl->j_commit_bh);
reiserfs_write_lock(s);
}
check_barrier_completion(s, jl->j_commit_bh);
/* If there was a write error in the journal - we can't commit this
* transaction - it will be invalid and, if successful, will just end
* up propagating the write error out to the filesystem. */
if (unlikely(!buffer_uptodate(jl->j_commit_bh))) {
#ifdef CONFIG_REISERFS_CHECK
reiserfs_warning(s, "journal-615", "buffer write failed");
#endif
retval = -EIO;
}
bforget(jl->j_commit_bh);
if (journal->j_last_commit_id != 0 &&
(jl->j_trans_id - journal->j_last_commit_id) != 1) {
reiserfs_warning(s, "clm-2200", "last commit %lu, current %lu",
journal->j_last_commit_id, jl->j_trans_id);
}
journal->j_last_commit_id = jl->j_trans_id;
/* now, every commit block is on the disk. It is safe to allow blocks freed during this transaction to be reallocated */
cleanup_freed_for_journal_list(s, jl);
retval = retval ? retval : journal->j_errno;
/* mark the metadata dirty */
if (!retval)
dirty_one_transaction(s, jl);
atomic_dec(&(jl->j_commit_left));
if (flushall) {
atomic_set(&(jl->j_older_commits_done), 1);
}
mutex_unlock(&jl->j_commit_mutex);
put_jl:
put_journal_list(s, jl);
if (retval)
reiserfs_abort(s, retval, "Journal write error in %s",
__func__);
put_fs_excl();
return retval;
}
/*
** flush_journal_list frequently needs to find a newer transaction for a given block. This does that, or
** returns NULL if it can't find anything
*/
static struct reiserfs_journal_list *find_newer_jl_for_cn(struct
reiserfs_journal_cnode
*cn)
{
struct super_block *sb = cn->sb;
b_blocknr_t blocknr = cn->blocknr;
cn = cn->hprev;
while (cn) {
if (cn->sb == sb && cn->blocknr == blocknr && cn->jlist) {
return cn->jlist;
}
cn = cn->hprev;
}
return NULL;
}
static int newer_jl_done(struct reiserfs_journal_cnode *cn)
{
struct super_block *sb = cn->sb;
b_blocknr_t blocknr = cn->blocknr;
cn = cn->hprev;
while (cn) {
if (cn->sb == sb && cn->blocknr == blocknr && cn->jlist &&
atomic_read(&cn->jlist->j_commit_left) != 0)
return 0;
cn = cn->hprev;
}
return 1;
}
static void remove_journal_hash(struct super_block *,
struct reiserfs_journal_cnode **,
struct reiserfs_journal_list *, unsigned long,
int);
/*
** once all the real blocks have been flushed, it is safe to remove them from the
** journal list for this transaction. Aside from freeing the cnode, this also allows the
** block to be reallocated for data blocks if it had been deleted.
*/
static void remove_all_from_journal_list(struct super_block *sb,
struct reiserfs_journal_list *jl,
int debug)
{
struct reiserfs_journal *journal = SB_JOURNAL(sb);
struct reiserfs_journal_cnode *cn, *last;
cn = jl->j_realblock;
/* which is better, to lock once around the whole loop, or
** to lock for each call to remove_journal_hash?
*/
while (cn) {
if (cn->blocknr != 0) {
if (debug) {
reiserfs_warning(sb, "reiserfs-2201",
"block %u, bh is %d, state %ld",
cn->blocknr, cn->bh ? 1 : 0,
cn->state);
}
cn->state = 0;
remove_journal_hash(sb, journal->j_list_hash_table,
jl, cn->blocknr, 1);
}
last = cn;
cn = cn->next;
free_cnode(sb, last);
}
jl->j_realblock = NULL;
}
/*
** if this timestamp is greater than the timestamp we wrote last to the header block, write it to the header block.
** once this is done, I can safely say the log area for this transaction won't ever be replayed, and I can start
** releasing blocks in this transaction for reuse as data blocks.
** called by flush_journal_list, before it calls remove_all_from_journal_list
**
*/
static int _update_journal_header_block(struct super_block *sb,
unsigned long offset,
unsigned int trans_id)
{
struct reiserfs_journal_header *jh;
struct reiserfs_journal *journal = SB_JOURNAL(sb);
if (reiserfs_is_journal_aborted(journal))
return -EIO;
if (trans_id >= journal->j_last_flush_trans_id) {
if (buffer_locked((journal->j_header_bh))) {
reiserfs_write_unlock(sb);
wait_on_buffer((journal->j_header_bh));
reiserfs_write_lock(sb);
if (unlikely(!buffer_uptodate(journal->j_header_bh))) {
#ifdef CONFIG_REISERFS_CHECK
reiserfs_warning(sb, "journal-699",
"buffer write failed");
#endif
return -EIO;
}
}
journal->j_last_flush_trans_id = trans_id;
journal->j_first_unflushed_offset = offset;
jh = (struct reiserfs_journal_header *)(journal->j_header_bh->
b_data);
jh->j_last_flush_trans_id = cpu_to_le32(trans_id);
jh->j_first_unflushed_offset = cpu_to_le32(offset);
jh->j_mount_id = cpu_to_le32(journal->j_mount_id);
if (reiserfs_barrier_flush(sb)) {
int ret;
lock_buffer(journal->j_header_bh);
ret = submit_barrier_buffer(journal->j_header_bh);
if (ret == -EOPNOTSUPP) {
set_buffer_uptodate(journal->j_header_bh);
disable_barrier(sb);
goto sync;
}
reiserfs_write_unlock(sb);
wait_on_buffer(journal->j_header_bh);
reiserfs_write_lock(sb);
check_barrier_completion(sb, journal->j_header_bh);
} else {
sync:
set_buffer_dirty(journal->j_header_bh);
reiserfs_write_unlock(sb);
sync_dirty_buffer(journal->j_header_bh);
reiserfs_write_lock(sb);
}
if (!buffer_uptodate(journal->j_header_bh)) {
reiserfs_warning(sb, "journal-837",
"IO error during journal replay");
return -EIO;
}
}
return 0;
}
static int update_journal_header_block(struct super_block *sb,
unsigned long offset,
unsigned int trans_id)
{
return _update_journal_header_block(sb, offset, trans_id);
}
/*
** flush any and all journal lists older than you are
** can only be called from flush_journal_list
*/
static int flush_older_journal_lists(struct super_block *sb,
struct reiserfs_journal_list *jl)
{
struct list_head *entry;
struct reiserfs_journal_list *other_jl;
struct reiserfs_journal *journal = SB_JOURNAL(sb);
unsigned int trans_id = jl->j_trans_id;
/* we know we are the only ones flushing things, no extra race
* protection is required.
*/
restart:
entry = journal->j_journal_list.next;
/* Did we wrap? */
if (entry == &journal->j_journal_list)
return 0;
other_jl = JOURNAL_LIST_ENTRY(entry);
if (other_jl->j_trans_id < trans_id) {
BUG_ON(other_jl->j_refcount <= 0);
/* do not flush all */
flush_journal_list(sb, other_jl, 0);
/* other_jl is now deleted from the list */
goto restart;
}
return 0;
}
static void del_from_work_list(struct super_block *s,
struct reiserfs_journal_list *jl)
{
struct reiserfs_journal *journal = SB_JOURNAL(s);
if (!list_empty(&jl->j_working_list)) {
list_del_init(&jl->j_working_list);
journal->j_num_work_lists--;
}
}
/* flush a journal list, both commit and real blocks
**
** always set flushall to 1, unless you are calling from inside
** flush_journal_list
**
** IMPORTANT. This can only be called while there are no journal writers,
** and the journal is locked. That means it can only be called from
** do_journal_end, or by journal_release
*/
static int flush_journal_list(struct super_block *s,
struct reiserfs_journal_list *jl, int flushall)
{
struct reiserfs_journal_list *pjl;
struct reiserfs_journal_cnode *cn, *last;
int count;
int was_jwait = 0;
int was_dirty = 0;
struct buffer_head *saved_bh;
unsigned long j_len_saved = jl->j_len;
struct reiserfs_journal *journal = SB_JOURNAL(s);
int err = 0;
BUG_ON(j_len_saved <= 0);
if (atomic_read(&journal->j_wcount) != 0) {
reiserfs_warning(s, "clm-2048", "called with wcount %d",
atomic_read(&journal->j_wcount));
}
BUG_ON(jl->j_trans_id == 0);
/* if flushall == 0, the lock is already held */
if (flushall) {
reiserfs_mutex_lock_safe(&journal->j_flush_mutex, s);
} else if (mutex_trylock(&journal->j_flush_mutex)) {
BUG();
}
count = 0;
if (j_len_saved > journal->j_trans_max) {
reiserfs_panic(s, "journal-715", "length is %lu, trans id %lu",
j_len_saved, jl->j_trans_id);
return 0;
}
get_fs_excl();
/* if all the work is already done, get out of here */
if (atomic_read(&(jl->j_nonzerolen)) <= 0 &&
atomic_read(&(jl->j_commit_left)) <= 0) {
goto flush_older_and_return;
}
/* start by putting the commit list on disk. This will also flush
** the commit lists of any olders transactions
*/
flush_commit_list(s, jl, 1);
if (!(jl->j_state & LIST_DIRTY)
&& !reiserfs_is_journal_aborted(journal))
BUG();
/* are we done now? */
if (atomic_read(&(jl->j_nonzerolen)) <= 0 &&
atomic_read(&(jl->j_commit_left)) <= 0) {
goto flush_older_and_return;
}
/* loop through each cnode, see if we need to write it,
** or wait on a more recent transaction, or just ignore it
*/
if (atomic_read(&(journal->j_wcount)) != 0) {
reiserfs_panic(s, "journal-844", "journal list is flushing, "
"wcount is not 0");
}
cn = jl->j_realblock;
while (cn) {
was_jwait = 0;
was_dirty = 0;
saved_bh = NULL;
/* blocknr of 0 is no longer in the hash, ignore it */
if (cn->blocknr == 0) {
goto free_cnode;
}
/* This transaction failed commit. Don't write out to the disk */
if (!(jl->j_state & LIST_DIRTY))
goto free_cnode;
pjl = find_newer_jl_for_cn(cn);
/* the order is important here. We check pjl to make sure we
** don't clear BH_JDirty_wait if we aren't the one writing this
** block to disk
*/
if (!pjl && cn->bh) {
saved_bh = cn->bh;
/* we do this to make sure nobody releases the buffer while
** we are working with it
*/
get_bh(saved_bh);
if (buffer_journal_dirty(saved_bh)) {
BUG_ON(!can_dirty(cn));
was_jwait = 1;
was_dirty = 1;
} else if (can_dirty(cn)) {
/* everything with !pjl && jwait should be writable */
BUG();
}
}
/* if someone has this block in a newer transaction, just make
** sure they are committed, and don't try writing it to disk
*/
if (pjl) {
if (atomic_read(&pjl->j_commit_left))
flush_commit_list(s, pjl, 1);
goto free_cnode;
}
/* bh == NULL when the block got to disk on its own, OR,
** the block got freed in a future transaction
*/
if (saved_bh == NULL) {
goto free_cnode;
}
/* this should never happen. kupdate_one_transaction has this list
** locked while it works, so we should never see a buffer here that
** is not marked JDirty_wait
*/
if ((!was_jwait) && !buffer_locked(saved_bh)) {
reiserfs_warning(s, "journal-813",
"BAD! buffer %llu %cdirty %cjwait, "
"not in a newer tranasction",
(unsigned long long)saved_bh->
b_blocknr, was_dirty ? ' ' : '!',
was_jwait ? ' ' : '!');
}
if (was_dirty) {
/* we inc again because saved_bh gets decremented at free_cnode */
get_bh(saved_bh);
set_bit(BLOCK_NEEDS_FLUSH, &cn->state);
lock_buffer(saved_bh);
BUG_ON(cn->blocknr != saved_bh->b_blocknr);
if (buffer_dirty(saved_bh))
submit_logged_buffer(saved_bh);
else
unlock_buffer(saved_bh);
count++;
} else {
reiserfs_warning(s, "clm-2082",
"Unable to flush buffer %llu in %s",
(unsigned long long)saved_bh->
b_blocknr, __func__);
}
free_cnode:
last = cn;
cn = cn->next;
if (saved_bh) {
/* we incremented this to keep others from taking the buffer head away */
put_bh(saved_bh);
if (atomic_read(&(saved_bh->b_count)) < 0) {
reiserfs_warning(s, "journal-945",
"saved_bh->b_count < 0");
}
}
}
if (count > 0) {
cn = jl->j_realblock;
while (cn) {
if (test_bit(BLOCK_NEEDS_FLUSH, &cn->state)) {
if (!cn->bh) {
reiserfs_panic(s, "journal-1011",
"cn->bh is NULL");
}
reiserfs_write_unlock(s);
wait_on_buffer(cn->bh);
reiserfs_write_lock(s);
if (!cn->bh) {
reiserfs_panic(s, "journal-1012",
"cn->bh is NULL");
}
if (unlikely(!buffer_uptodate(cn->bh))) {
#ifdef CONFIG_REISERFS_CHECK
reiserfs_warning(s, "journal-949",
"buffer write failed");
#endif
err = -EIO;
}
/* note, we must clear the JDirty_wait bit after the up to date
** check, otherwise we race against our flushpage routine
*/
BUG_ON(!test_clear_buffer_journal_dirty
(cn->bh));
/* drop one ref for us */
put_bh(cn->bh);
/* drop one ref for journal_mark_dirty */
release_buffer_page(cn->bh);
}
cn = cn->next;
}
}
if (err)
reiserfs_abort(s, -EIO,
"Write error while pushing transaction to disk in %s",
__func__);
flush_older_and_return:
/* before we can update the journal header block, we _must_ flush all
** real blocks from all older transactions to disk. This is because
** once the header block is updated, this transaction will not be
** replayed after a crash
*/
if (flushall) {
flush_older_journal_lists(s, jl);
}
err = journal->j_errno;
/* before we can remove everything from the hash tables for this
** transaction, we must make sure it can never be replayed
**
** since we are only called from do_journal_end, we know for sure there
** are no allocations going on while we are flushing journal lists. So,
** we only need to update the journal header block for the last list
** being flushed
*/
if (!err && flushall) {
err =
update_journal_header_block(s,
(jl->j_start + jl->j_len +
2) % SB_ONDISK_JOURNAL_SIZE(s),
jl->j_trans_id);
if (err)
reiserfs_abort(s, -EIO,
"Write error while updating journal header in %s",
__func__);
}
remove_all_from_journal_list(s, jl, 0);
list_del_init(&jl->j_list);
journal->j_num_lists--;
del_from_work_list(s, jl);
if (journal->j_last_flush_id != 0 &&
(jl->j_trans_id - journal->j_last_flush_id) != 1) {
reiserfs_warning(s, "clm-2201", "last flush %lu, current %lu",
journal->j_last_flush_id, jl->j_trans_id);
}
journal->j_last_flush_id = jl->j_trans_id;
/* not strictly required since we are freeing the list, but it should
* help find code using dead lists later on
*/
jl->j_len = 0;
atomic_set(&(jl->j_nonzerolen), 0);
jl->j_start = 0;
jl->j_realblock = NULL;
jl->j_commit_bh = NULL;
jl->j_trans_id = 0;
jl->j_state = 0;
put_journal_list(s, jl);
if (flushall)
mutex_unlock(&journal->j_flush_mutex);
put_fs_excl();
return err;
}
static int test_transaction(struct super_block *s,
struct reiserfs_journal_list *jl)
{
struct reiserfs_journal_cnode *cn;
if (jl->j_len == 0 || atomic_read(&jl->j_nonzerolen) == 0)
return 1;
cn = jl->j_realblock;
while (cn) {
/* if the blocknr == 0, this has been cleared from the hash,
** skip it
*/
if (cn->blocknr == 0) {
goto next;
}
if (cn->bh && !newer_jl_done(cn))
return 0;
next:
cn = cn->next;
cond_resched();
}
return 0;
}
static int write_one_transaction(struct super_block *s,
struct reiserfs_journal_list *jl,
struct buffer_chunk *chunk)
{
struct reiserfs_journal_cnode *cn;
int ret = 0;
jl->j_state |= LIST_TOUCHED;
del_from_work_list(s, jl);
if (jl->j_len == 0 || atomic_read(&jl->j_nonzerolen) == 0) {
return 0;
}
cn = jl->j_realblock;
while (cn) {
/* if the blocknr == 0, this has been cleared from the hash,
** skip it
*/
if (cn->blocknr == 0) {
goto next;
}
if (cn->bh && can_dirty(cn) && buffer_dirty(cn->bh)) {
struct buffer_head *tmp_bh;
/* we can race against journal_mark_freed when we try
* to lock_buffer(cn->bh), so we have to inc the buffer
* count, and recheck things after locking
*/
tmp_bh = cn->bh;
get_bh(tmp_bh);
lock_buffer(tmp_bh);
if (cn->bh && can_dirty(cn) && buffer_dirty(tmp_bh)) {
if (!buffer_journal_dirty(tmp_bh) ||
buffer_journal_prepared(tmp_bh))
BUG();
add_to_chunk(chunk, tmp_bh, NULL, write_chunk);
ret++;
} else {
/* note, cn->bh might be null now */
unlock_buffer(tmp_bh);
}
put_bh(tmp_bh);
}
next:
cn = cn->next;
cond_resched();
}
return ret;
}
/* used by flush_commit_list */
static int dirty_one_transaction(struct super_block *s,
struct reiserfs_journal_list *jl)
{
struct reiserfs_journal_cnode *cn;
struct reiserfs_journal_list *pjl;
int ret = 0;
jl->j_state |= LIST_DIRTY;
cn = jl->j_realblock;
while (cn) {
/* look for a more recent transaction that logged this
** buffer. Only the most recent transaction with a buffer in
** it is allowed to send that buffer to disk
*/
pjl = find_newer_jl_for_cn(cn);
if (!pjl && cn->blocknr && cn->bh
&& buffer_journal_dirty(cn->bh)) {
BUG_ON(!can_dirty(cn));
/* if the buffer is prepared, it will either be logged
* or restored. If restored, we need to make sure
* it actually gets marked dirty
*/
clear_buffer_journal_new(cn->bh);
if (buffer_journal_prepared(cn->bh)) {
set_buffer_journal_restore_dirty(cn->bh);
} else {
set_buffer_journal_test(cn->bh);
mark_buffer_dirty(cn->bh);
}
}
cn = cn->next;
}
return ret;
}
static int kupdate_transactions(struct super_block *s,
struct reiserfs_journal_list *jl,
struct reiserfs_journal_list **next_jl,
unsigned int *next_trans_id,
int num_blocks, int num_trans)
{
int ret = 0;
int written = 0;
int transactions_flushed = 0;
unsigned int orig_trans_id = jl->j_trans_id;
struct buffer_chunk chunk;
struct list_head *entry;
struct reiserfs_journal *journal = SB_JOURNAL(s);
chunk.nr = 0;
reiserfs_mutex_lock_safe(&journal->j_flush_mutex, s);
if (!journal_list_still_alive(s, orig_trans_id)) {
goto done;
}
/* we've got j_flush_mutex held, nobody is going to delete any
* of these lists out from underneath us
*/
while ((num_trans && transactions_flushed < num_trans) ||
(!num_trans && written < num_blocks)) {
if (jl->j_len == 0 || (jl->j_state & LIST_TOUCHED) ||
atomic_read(&jl->j_commit_left)
|| !(jl->j_state & LIST_DIRTY)) {
del_from_work_list(s, jl);
break;
}
ret = write_one_transaction(s, jl, &chunk);
if (ret < 0)
goto done;
transactions_flushed++;
written += ret;
entry = jl->j_list.next;
/* did we wrap? */
if (entry == &journal->j_journal_list) {
break;
}
jl = JOURNAL_LIST_ENTRY(entry);
/* don't bother with older transactions */
if (jl->j_trans_id <= orig_trans_id)
break;
}
if (chunk.nr) {
write_chunk(&chunk);
}
done:
mutex_unlock(&journal->j_flush_mutex);
return ret;
}
/* for o_sync and fsync heavy applications, they tend to use
** all the journa list slots with tiny transactions. These
** trigger lots and lots of calls to update the header block, which
** adds seeks and slows things down.
**
** This function tries to clear out a large chunk of the journal lists
** at once, which makes everything faster since only the newest journal
** list updates the header block
*/
static int flush_used_journal_lists(struct super_block *s,
struct reiserfs_journal_list *jl)
{
unsigned long len = 0;
unsigned long cur_len;
int ret;
int i;
int limit = 256;
struct reiserfs_journal_list *tjl;
struct reiserfs_journal_list *flush_jl;
unsigned int trans_id;
struct reiserfs_journal *journal = SB_JOURNAL(s);
flush_jl = tjl = jl;
/* in data logging mode, try harder to flush a lot of blocks */
if (reiserfs_data_log(s))
limit = 1024;
/* flush for 256 transactions or limit blocks, whichever comes first */
for (i = 0; i < 256 && len < limit; i++) {
if (atomic_read(&tjl->j_commit_left) ||
tjl->j_trans_id < jl->j_trans_id) {
break;
}
cur_len = atomic_read(&tjl->j_nonzerolen);
if (cur_len > 0) {
tjl->j_state &= ~LIST_TOUCHED;
}
len += cur_len;
flush_jl = tjl;
if (tjl->j_list.next == &journal->j_journal_list)
break;
tjl = JOURNAL_LIST_ENTRY(tjl->j_list.next);
}
/* try to find a group of blocks we can flush across all the
** transactions, but only bother if we've actually spanned
** across multiple lists
*/
if (flush_jl != jl) {
ret = kupdate_transactions(s, jl, &tjl, &trans_id, len, i);
}
flush_journal_list(s, flush_jl, 1);
return 0;
}
/*
** removes any nodes in table with name block and dev as bh.
** only touchs the hnext and hprev pointers.
*/
void remove_journal_hash(struct super_block *sb,
struct reiserfs_journal_cnode **table,
struct reiserfs_journal_list *jl,
unsigned long block, int remove_freed)
{
struct reiserfs_journal_cnode *cur;
struct reiserfs_journal_cnode **head;
head = &(journal_hash(table, sb, block));
if (!head) {
return;
}
cur = *head;
while (cur) {
if (cur->blocknr == block && cur->sb == sb
&& (jl == NULL || jl == cur->jlist)
&& (!test_bit(BLOCK_FREED, &cur->state) || remove_freed)) {
if (cur->hnext) {
cur->hnext->hprev = cur->hprev;
}
if (cur->hprev) {
cur->hprev->hnext = cur->hnext;
} else {
*head = cur->hnext;
}
cur->blocknr = 0;
cur->sb = NULL;
cur->state = 0;
if (cur->bh && cur->jlist) /* anybody who clears the cur->bh will also dec the nonzerolen */
atomic_dec(&(cur->jlist->j_nonzerolen));
cur->bh = NULL;
cur->jlist = NULL;
}
cur = cur->hnext;
}
}
static void free_journal_ram(struct super_block *sb)
{
struct reiserfs_journal *journal = SB_JOURNAL(sb);
kfree(journal->j_current_jl);
journal->j_num_lists--;
vfree(journal->j_cnode_free_orig);
free_list_bitmaps(sb, journal->j_list_bitmap);
free_bitmap_nodes(sb); /* must be after free_list_bitmaps */
if (journal->j_header_bh) {
brelse(journal->j_header_bh);
}
/* j_header_bh is on the journal dev, make sure not to release the journal
* dev until we brelse j_header_bh
*/
release_journal_dev(sb, journal);
vfree(journal);
}
/*
** call on unmount. Only set error to 1 if you haven't made your way out
** of read_super() yet. Any other caller must keep error at 0.
*/
static int do_journal_release(struct reiserfs_transaction_handle *th,
struct super_block *sb, int error)
{
struct reiserfs_transaction_handle myth;
int flushed = 0;
struct reiserfs_journal *journal = SB_JOURNAL(sb);
/* we only want to flush out transactions if we were called with error == 0
*/
if (!error && !(sb->s_flags & MS_RDONLY)) {
/* end the current trans */
BUG_ON(!th->t_trans_id);
do_journal_end(th, sb, 10, FLUSH_ALL);
/* make sure something gets logged to force our way into the flush code */
if (!journal_join(&myth, sb, 1)) {
reiserfs_prepare_for_journal(sb,
SB_BUFFER_WITH_SB(sb),
1);
journal_mark_dirty(&myth, sb,
SB_BUFFER_WITH_SB(sb));
do_journal_end(&myth, sb, 1, FLUSH_ALL);
flushed = 1;
}
}
/* this also catches errors during the do_journal_end above */
if (!error && reiserfs_is_journal_aborted(journal)) {
memset(&myth, 0, sizeof(myth));
if (!journal_join_abort(&myth, sb, 1)) {
reiserfs_prepare_for_journal(sb,
SB_BUFFER_WITH_SB(sb),
1);
journal_mark_dirty(&myth, sb,
SB_BUFFER_WITH_SB(sb));
do_journal_end(&myth, sb, 1, FLUSH_ALL);
}
}
reiserfs_mounted_fs_count--;
/* wait for all commits to finish */
cancel_delayed_work(&SB_JOURNAL(sb)->j_work);
/*
* We must release the write lock here because
* the workqueue job (flush_async_commit) needs this lock
*/
reiserfs_write_unlock(sb);
flush_workqueue(commit_wq);
if (!reiserfs_mounted_fs_count) {
destroy_workqueue(commit_wq);
commit_wq = NULL;
}
reiserfs_write_lock(sb);
free_journal_ram(sb);
return 0;
}
/*
** call on unmount. flush all journal trans, release all alloc'd ram
*/
int journal_release(struct reiserfs_transaction_handle *th,
struct super_block *sb)
{
return do_journal_release(th, sb, 0);
}
/*
** only call from an error condition inside reiserfs_read_super!
*/
int journal_release_error(struct reiserfs_transaction_handle *th,
struct super_block *sb)
{
return do_journal_release(th, sb, 1);
}
/* compares description block with commit block. returns 1 if they differ, 0 if they are the same */
static int journal_compare_desc_commit(struct super_block *sb,
struct reiserfs_journal_desc *desc,
struct reiserfs_journal_commit *commit)
{
if (get_commit_trans_id(commit) != get_desc_trans_id(desc) ||
get_commit_trans_len(commit) != get_desc_trans_len(desc) ||
get_commit_trans_len(commit) > SB_JOURNAL(sb)->j_trans_max ||
get_commit_trans_len(commit) <= 0) {
return 1;
}
return 0;
}
/* returns 0 if it did not find a description block
** returns -1 if it found a corrupt commit block
** returns 1 if both desc and commit were valid
*/
static int journal_transaction_is_valid(struct super_block *sb,
struct buffer_head *d_bh,
unsigned int *oldest_invalid_trans_id,
unsigned long *newest_mount_id)
{
struct reiserfs_journal_desc *desc;
struct reiserfs_journal_commit *commit;
struct buffer_head *c_bh;
unsigned long offset;
if (!d_bh)
return 0;
desc = (struct reiserfs_journal_desc *)d_bh->b_data;
if (get_desc_trans_len(desc) > 0
&& !memcmp(get_journal_desc_magic(d_bh), JOURNAL_DESC_MAGIC, 8)) {
if (oldest_invalid_trans_id && *oldest_invalid_trans_id
&& get_desc_trans_id(desc) > *oldest_invalid_trans_id) {
reiserfs_debug(sb, REISERFS_DEBUG_CODE,
"journal-986: transaction "
"is valid returning because trans_id %d is greater than "
"oldest_invalid %lu",
get_desc_trans_id(desc),
*oldest_invalid_trans_id);
return 0;
}
if (newest_mount_id
&& *newest_mount_id > get_desc_mount_id(desc)) {
reiserfs_debug(sb, REISERFS_DEBUG_CODE,
"journal-1087: transaction "
"is valid returning because mount_id %d is less than "
"newest_mount_id %lu",
get_desc_mount_id(desc),
*newest_mount_id);
return -1;
}
if (get_desc_trans_len(desc) > SB_JOURNAL(sb)->j_trans_max) {
reiserfs_warning(sb, "journal-2018",
"Bad transaction length %d "
"encountered, ignoring transaction",
get_desc_trans_len(desc));
return -1;
}
offset = d_bh->b_blocknr - SB_ONDISK_JOURNAL_1st_BLOCK(sb);
/* ok, we have a journal description block, lets see if the transaction was valid */
c_bh =
journal_bread(sb,
SB_ONDISK_JOURNAL_1st_BLOCK(sb) +
((offset + get_desc_trans_len(desc) +
1) % SB_ONDISK_JOURNAL_SIZE(sb)));
if (!c_bh)
return 0;
commit = (struct reiserfs_journal_commit *)c_bh->b_data;
if (journal_compare_desc_commit(sb, desc, commit)) {
reiserfs_debug(sb, REISERFS_DEBUG_CODE,
"journal_transaction_is_valid, commit offset %ld had bad "
"time %d or length %d",
c_bh->b_blocknr -
SB_ONDISK_JOURNAL_1st_BLOCK(sb),
get_commit_trans_id(commit),
get_commit_trans_len(commit));
brelse(c_bh);
if (oldest_invalid_trans_id) {
*oldest_invalid_trans_id =
get_desc_trans_id(desc);
reiserfs_debug(sb, REISERFS_DEBUG_CODE,
"journal-1004: "
"transaction_is_valid setting oldest invalid trans_id "
"to %d",
get_desc_trans_id(desc));
}
return -1;
}
brelse(c_bh);
reiserfs_debug(sb, REISERFS_DEBUG_CODE,
"journal-1006: found valid "
"transaction start offset %llu, len %d id %d",
d_bh->b_blocknr -
SB_ONDISK_JOURNAL_1st_BLOCK(sb),
get_desc_trans_len(desc),
get_desc_trans_id(desc));
return 1;
} else {
return 0;
}
}
static void brelse_array(struct buffer_head **heads, int num)
{
int i;
for (i = 0; i < num; i++) {
brelse(heads[i]);
}
}
/*
** given the start, and values for the oldest acceptable transactions,
** this either reads in a replays a transaction, or returns because the transaction
** is invalid, or too old.
*/
static int journal_read_transaction(struct super_block *sb,
unsigned long cur_dblock,
unsigned long oldest_start,
unsigned int oldest_trans_id,
unsigned long newest_mount_id)
{
struct reiserfs_journal *journal = SB_JOURNAL(sb);
struct reiserfs_journal_desc *desc;
struct reiserfs_journal_commit *commit;
unsigned int trans_id = 0;
struct buffer_head *c_bh;
struct buffer_head *d_bh;
struct buffer_head **log_blocks = NULL;
struct buffer_head **real_blocks = NULL;
unsigned int trans_offset;
int i;
int trans_half;
d_bh = journal_bread(sb, cur_dblock);
if (!d_bh)
return 1;
desc = (struct reiserfs_journal_desc *)d_bh->b_data;
trans_offset = d_bh->b_blocknr - SB_ONDISK_JOURNAL_1st_BLOCK(sb);
reiserfs_debug(sb, REISERFS_DEBUG_CODE, "journal-1037: "
"journal_read_transaction, offset %llu, len %d mount_id %d",
d_bh->b_blocknr - SB_ONDISK_JOURNAL_1st_BLOCK(sb),
get_desc_trans_len(desc), get_desc_mount_id(desc));
if (get_desc_trans_id(desc) < oldest_trans_id) {
reiserfs_debug(sb, REISERFS_DEBUG_CODE, "journal-1039: "
"journal_read_trans skipping because %lu is too old",
cur_dblock -
SB_ONDISK_JOURNAL_1st_BLOCK(sb));
brelse(d_bh);
return 1;
}
if (get_desc_mount_id(desc) != newest_mount_id) {
reiserfs_debug(sb, REISERFS_DEBUG_CODE, "journal-1146: "
"journal_read_trans skipping because %d is != "
"newest_mount_id %lu", get_desc_mount_id(desc),
newest_mount_id);
brelse(d_bh);
return 1;
}
c_bh = journal_bread(sb, SB_ONDISK_JOURNAL_1st_BLOCK(sb) +
((trans_offset + get_desc_trans_len(desc) + 1) %
SB_ONDISK_JOURNAL_SIZE(sb)));
if (!c_bh) {
brelse(d_bh);
return 1;
}
commit = (struct reiserfs_journal_commit *)c_bh->b_data;
if (journal_compare_desc_commit(sb, desc, commit)) {
reiserfs_debug(sb, REISERFS_DEBUG_CODE,
"journal_read_transaction, "
"commit offset %llu had bad time %d or length %d",
c_bh->b_blocknr -
SB_ONDISK_JOURNAL_1st_BLOCK(sb),
get_commit_trans_id(commit),
get_commit_trans_len(commit));
brelse(c_bh);
brelse(d_bh);
return 1;
}
trans_id = get_desc_trans_id(desc);
/* now we know we've got a good transaction, and it was inside the valid time ranges */
log_blocks = kmalloc(get_desc_trans_len(desc) *
sizeof(struct buffer_head *), GFP_NOFS);
real_blocks = kmalloc(get_desc_trans_len(desc) *
sizeof(struct buffer_head *), GFP_NOFS);
if (!log_blocks || !real_blocks) {
brelse(c_bh);
brelse(d_bh);
kfree(log_blocks);
kfree(real_blocks);
reiserfs_warning(sb, "journal-1169",
"kmalloc failed, unable to mount FS");
return -1;
}
/* get all the buffer heads */
trans_half = journal_trans_half(sb->s_blocksize);
for (i = 0; i < get_desc_trans_len(desc); i++) {
log_blocks[i] =
journal_getblk(sb,
SB_ONDISK_JOURNAL_1st_BLOCK(sb) +
(trans_offset + 1 +
i) % SB_ONDISK_JOURNAL_SIZE(sb));
if (i < trans_half) {
real_blocks[i] =
sb_getblk(sb,
le32_to_cpu(desc->j_realblock[i]));
} else {
real_blocks[i] =
sb_getblk(sb,
le32_to_cpu(commit->
j_realblock[i - trans_half]));
}
if (real_blocks[i]->b_blocknr > SB_BLOCK_COUNT(sb)) {
reiserfs_warning(sb, "journal-1207",
"REPLAY FAILURE fsck required! "
"Block to replay is outside of "
"filesystem");
goto abort_replay;
}
/* make sure we don't try to replay onto log or reserved area */
if (is_block_in_log_or_reserved_area
(sb, real_blocks[i]->b_blocknr)) {
reiserfs_warning(sb, "journal-1204",
"REPLAY FAILURE fsck required! "
"Trying to replay onto a log block");
abort_replay:
brelse_array(log_blocks, i);
brelse_array(real_blocks, i);
brelse(c_bh);
brelse(d_bh);
kfree(log_blocks);
kfree(real_blocks);
return -1;
}
}
/* read in the log blocks, memcpy to the corresponding real block */
ll_rw_block(READ, get_desc_trans_len(desc), log_blocks);
for (i = 0; i < get_desc_trans_len(desc); i++) {
reiserfs_write_unlock(sb);
wait_on_buffer(log_blocks[i]);
reiserfs_write_lock(sb);
if (!buffer_uptodate(log_blocks[i])) {
reiserfs_warning(sb, "journal-1212",
"REPLAY FAILURE fsck required! "
"buffer write failed");
brelse_array(log_blocks + i,
get_desc_trans_len(desc) - i);
brelse_array(real_blocks, get_desc_trans_len(desc));
brelse(c_bh);
brelse(d_bh);
kfree(log_blocks);
kfree(real_blocks);
return -1;
}
memcpy(real_blocks[i]->b_data, log_blocks[i]->b_data,
real_blocks[i]->b_size);
set_buffer_uptodate(real_blocks[i]);
brelse(log_blocks[i]);
}
/* flush out the real blocks */
for (i = 0; i < get_desc_trans_len(desc); i++) {
set_buffer_dirty(real_blocks[i]);
ll_rw_block(SWRITE, 1, real_blocks + i);
}
for (i = 0; i < get_desc_trans_len(desc); i++) {
wait_on_buffer(real_blocks[i]);
if (!buffer_uptodate(real_blocks[i])) {
reiserfs_warning(sb, "journal-1226",
"REPLAY FAILURE, fsck required! "
"buffer write failed");
brelse_array(real_blocks + i,
get_desc_trans_len(desc) - i);
brelse(c_bh);
brelse(d_bh);
kfree(log_blocks);
kfree(real_blocks);
return -1;
}
brelse(real_blocks[i]);
}
cur_dblock =
SB_ONDISK_JOURNAL_1st_BLOCK(sb) +
((trans_offset + get_desc_trans_len(desc) +
2) % SB_ONDISK_JOURNAL_SIZE(sb));
reiserfs_debug(sb, REISERFS_DEBUG_CODE,
"journal-1095: setting journal " "start to offset %ld",
cur_dblock - SB_ONDISK_JOURNAL_1st_BLOCK(sb));
/* init starting values for the first transaction, in case this is the last transaction to be replayed. */
journal->j_start = cur_dblock - SB_ONDISK_JOURNAL_1st_BLOCK(sb);
journal->j_last_flush_trans_id = trans_id;
journal->j_trans_id = trans_id + 1;
/* check for trans_id overflow */
if (journal->j_trans_id == 0)
journal->j_trans_id = 10;
brelse(c_bh);
brelse(d_bh);
kfree(log_blocks);
kfree(real_blocks);
return 0;
}
/* This function reads blocks starting from block and to max_block of bufsize
size (but no more than BUFNR blocks at a time). This proved to improve
mounting speed on self-rebuilding raid5 arrays at least.
Right now it is only used from journal code. But later we might use it
from other places.
Note: Do not use journal_getblk/sb_getblk functions here! */
static struct buffer_head *reiserfs_breada(struct block_device *dev,
b_blocknr_t block, int bufsize,
b_blocknr_t max_block)
{
struct buffer_head *bhlist[BUFNR];
unsigned int blocks = BUFNR;
struct buffer_head *bh;
int i, j;
bh = __getblk(dev, block, bufsize);
if (buffer_uptodate(bh))
return (bh);
if (block + BUFNR > max_block) {
blocks = max_block - block;
}
bhlist[0] = bh;
j = 1;
for (i = 1; i < blocks; i++) {
bh = __getblk(dev, block + i, bufsize);
if (buffer_uptodate(bh)) {
brelse(bh);
break;
} else
bhlist[j++] = bh;
}
ll_rw_block(READ, j, bhlist);
for (i = 1; i < j; i++)
brelse(bhlist[i]);
bh = bhlist[0];
wait_on_buffer(bh);
if (buffer_uptodate(bh))
return bh;
brelse(bh);
return NULL;
}
/*
** read and replay the log
** on a clean unmount, the journal header's next unflushed pointer will be to an invalid
** transaction. This tests that before finding all the transactions in the log, which makes normal mount times fast.
**
** After a crash, this starts with the next unflushed transaction, and replays until it finds one too old, or invalid.
**
** On exit, it sets things up so the first transaction will work correctly.
*/
static int journal_read(struct super_block *sb)
{
struct reiserfs_journal *journal = SB_JOURNAL(sb);
struct reiserfs_journal_desc *desc;
unsigned int oldest_trans_id = 0;
unsigned int oldest_invalid_trans_id = 0;
time_t start;
unsigned long oldest_start = 0;
unsigned long cur_dblock = 0;
unsigned long newest_mount_id = 9;
struct buffer_head *d_bh;
struct reiserfs_journal_header *jh;
int valid_journal_header = 0;
int replay_count = 0;
int continue_replay = 1;
int ret;
char b[BDEVNAME_SIZE];
cur_dblock = SB_ONDISK_JOURNAL_1st_BLOCK(sb);
reiserfs_info(sb, "checking transaction log (%s)\n",
bdevname(journal->j_dev_bd, b));
start = get_seconds();
/* step 1, read in the journal header block. Check the transaction it says
** is the first unflushed, and if that transaction is not valid,
** replay is done
*/
journal->j_header_bh = journal_bread(sb,
SB_ONDISK_JOURNAL_1st_BLOCK(sb)
+ SB_ONDISK_JOURNAL_SIZE(sb));
if (!journal->j_header_bh) {
return 1;
}
jh = (struct reiserfs_journal_header *)(journal->j_header_bh->b_data);
if (le32_to_cpu(jh->j_first_unflushed_offset) <
SB_ONDISK_JOURNAL_SIZE(sb)
&& le32_to_cpu(jh->j_last_flush_trans_id) > 0) {
oldest_start =
SB_ONDISK_JOURNAL_1st_BLOCK(sb) +
le32_to_cpu(jh->j_first_unflushed_offset);
oldest_trans_id = le32_to_cpu(jh->j_last_flush_trans_id) + 1;
newest_mount_id = le32_to_cpu(jh->j_mount_id);
reiserfs_debug(sb, REISERFS_DEBUG_CODE,
"journal-1153: found in "
"header: first_unflushed_offset %d, last_flushed_trans_id "
"%lu", le32_to_cpu(jh->j_first_unflushed_offset),
le32_to_cpu(jh->j_last_flush_trans_id));
valid_journal_header = 1;
/* now, we try to read the first unflushed offset. If it is not valid,
** there is nothing more we can do, and it makes no sense to read
** through the whole log.
*/
d_bh =
journal_bread(sb,
SB_ONDISK_JOURNAL_1st_BLOCK(sb) +
le32_to_cpu(jh->j_first_unflushed_offset));
ret = journal_transaction_is_valid(sb, d_bh, NULL, NULL);
if (!ret) {
continue_replay = 0;
}
brelse(d_bh);
goto start_log_replay;
}
if (continue_replay && bdev_read_only(sb->s_bdev)) {
reiserfs_warning(sb, "clm-2076",
"device is readonly, unable to replay log");
return -1;
}
/* ok, there are transactions that need to be replayed. start with the first log block, find
** all the valid transactions, and pick out the oldest.
*/
while (continue_replay
&& cur_dblock <
(SB_ONDISK_JOURNAL_1st_BLOCK(sb) +
SB_ONDISK_JOURNAL_SIZE(sb))) {
/* Note that it is required for blocksize of primary fs device and journal
device to be the same */
d_bh =
reiserfs_breada(journal->j_dev_bd, cur_dblock,
sb->s_blocksize,
SB_ONDISK_JOURNAL_1st_BLOCK(sb) +
SB_ONDISK_JOURNAL_SIZE(sb));
ret =
journal_transaction_is_valid(sb, d_bh,
&oldest_invalid_trans_id,
&newest_mount_id);
if (ret == 1) {
desc = (struct reiserfs_journal_desc *)d_bh->b_data;
if (oldest_start == 0) { /* init all oldest_ values */
oldest_trans_id = get_desc_trans_id(desc);
oldest_start = d_bh->b_blocknr;
newest_mount_id = get_desc_mount_id(desc);
reiserfs_debug(sb, REISERFS_DEBUG_CODE,
"journal-1179: Setting "
"oldest_start to offset %llu, trans_id %lu",
oldest_start -
SB_ONDISK_JOURNAL_1st_BLOCK
(sb), oldest_trans_id);
} else if (oldest_trans_id > get_desc_trans_id(desc)) {
/* one we just read was older */
oldest_trans_id = get_desc_trans_id(desc);
oldest_start = d_bh->b_blocknr;
reiserfs_debug(sb, REISERFS_DEBUG_CODE,
"journal-1180: Resetting "
"oldest_start to offset %lu, trans_id %lu",
oldest_start -
SB_ONDISK_JOURNAL_1st_BLOCK
(sb), oldest_trans_id);
}
if (newest_mount_id < get_desc_mount_id(desc)) {
newest_mount_id = get_desc_mount_id(desc);
reiserfs_debug(sb, REISERFS_DEBUG_CODE,
"journal-1299: Setting "
"newest_mount_id to %d",
get_desc_mount_id(desc));
}
cur_dblock += get_desc_trans_len(desc) + 2;
} else {
cur_dblock++;
}
brelse(d_bh);
}
start_log_replay:
cur_dblock = oldest_start;
if (oldest_trans_id) {
reiserfs_debug(sb, REISERFS_DEBUG_CODE,
"journal-1206: Starting replay "
"from offset %llu, trans_id %lu",
cur_dblock - SB_ONDISK_JOURNAL_1st_BLOCK(sb),
oldest_trans_id);
}
replay_count = 0;
while (continue_replay && oldest_trans_id > 0) {
ret =
journal_read_transaction(sb, cur_dblock, oldest_start,
oldest_trans_id, newest_mount_id);
if (ret < 0) {
return ret;
} else if (ret != 0) {
break;
}
cur_dblock =
SB_ONDISK_JOURNAL_1st_BLOCK(sb) + journal->j_start;
replay_count++;
if (cur_dblock == oldest_start)
break;
}
if (oldest_trans_id == 0) {
reiserfs_debug(sb, REISERFS_DEBUG_CODE,
"journal-1225: No valid " "transactions found");
}
/* j_start does not get set correctly if we don't replay any transactions.
** if we had a valid journal_header, set j_start to the first unflushed transaction value,
** copy the trans_id from the header
*/
if (valid_journal_header && replay_count == 0) {
journal->j_start = le32_to_cpu(jh->j_first_unflushed_offset);
journal->j_trans_id =
le32_to_cpu(jh->j_last_flush_trans_id) + 1;
/* check for trans_id overflow */
if (journal->j_trans_id == 0)
journal->j_trans_id = 10;
journal->j_last_flush_trans_id =
le32_to_cpu(jh->j_last_flush_trans_id);
journal->j_mount_id = le32_to_cpu(jh->j_mount_id) + 1;
} else {
journal->j_mount_id = newest_mount_id + 1;
}
reiserfs_debug(sb, REISERFS_DEBUG_CODE, "journal-1299: Setting "
"newest_mount_id to %lu", journal->j_mount_id);
journal->j_first_unflushed_offset = journal->j_start;
if (replay_count > 0) {
reiserfs_info(sb,
"replayed %d transactions in %lu seconds\n",
replay_count, get_seconds() - start);
}
if (!bdev_read_only(sb->s_bdev) &&
_update_journal_header_block(sb, journal->j_start,
journal->j_last_flush_trans_id)) {
/* replay failed, caller must call free_journal_ram and abort
** the mount
*/
return -1;
}
return 0;
}
static struct reiserfs_journal_list *alloc_journal_list(struct super_block *s)
{
struct reiserfs_journal_list *jl;
jl = kzalloc(sizeof(struct reiserfs_journal_list),
GFP_NOFS | __GFP_NOFAIL);
INIT_LIST_HEAD(&jl->j_list);
INIT_LIST_HEAD(&jl->j_working_list);
INIT_LIST_HEAD(&jl->j_tail_bh_list);
INIT_LIST_HEAD(&jl->j_bh_list);
mutex_init(&jl->j_commit_mutex);
SB_JOURNAL(s)->j_num_lists++;
get_journal_list(jl);
return jl;
}
static void journal_list_init(struct super_block *sb)
{
SB_JOURNAL(sb)->j_current_jl = alloc_journal_list(sb);
}
static int release_journal_dev(struct super_block *super,
struct reiserfs_journal *journal)
{
int result;
result = 0;
if (journal->j_dev_bd != NULL) {
if (journal->j_dev_bd->bd_dev != super->s_dev)
bd_release(journal->j_dev_bd);
result = blkdev_put(journal->j_dev_bd, journal->j_dev_mode);
journal->j_dev_bd = NULL;
}
if (result != 0) {
reiserfs_warning(super, "sh-457",
"Cannot release journal device: %i", result);
}
return result;
}
static int journal_init_dev(struct super_block *super,
struct reiserfs_journal *journal,
const char *jdev_name)
{
int result;
dev_t jdev;
fmode_t blkdev_mode = FMODE_READ | FMODE_WRITE;
char b[BDEVNAME_SIZE];
result = 0;
journal->j_dev_bd = NULL;
jdev = SB_ONDISK_JOURNAL_DEVICE(super) ?
new_decode_dev(SB_ONDISK_JOURNAL_DEVICE(super)) : super->s_dev;
if (bdev_read_only(super->s_bdev))
blkdev_mode = FMODE_READ;
/* there is no "jdev" option and journal is on separate device */
if ((!jdev_name || !jdev_name[0])) {
journal->j_dev_bd = open_by_devnum(jdev, blkdev_mode);
journal->j_dev_mode = blkdev_mode;
if (IS_ERR(journal->j_dev_bd)) {
result = PTR_ERR(journal->j_dev_bd);
journal->j_dev_bd = NULL;
reiserfs_warning(super, "sh-458",
"cannot init journal device '%s': %i",
__bdevname(jdev, b), result);
return result;
} else if (jdev != super->s_dev) {
result = bd_claim(journal->j_dev_bd, journal);
if (result) {
blkdev_put(journal->j_dev_bd, blkdev_mode);
return result;
}
set_blocksize(journal->j_dev_bd, super->s_blocksize);
}
return 0;
}
journal->j_dev_mode = blkdev_mode;
journal->j_dev_bd = open_bdev_exclusive(jdev_name,
blkdev_mode, journal);
if (IS_ERR(journal->j_dev_bd)) {
result = PTR_ERR(journal->j_dev_bd);
journal->j_dev_bd = NULL;
reiserfs_warning(super,
"journal_init_dev: Cannot open '%s': %i",
jdev_name, result);
return result;
}
set_blocksize(journal->j_dev_bd, super->s_blocksize);
reiserfs_info(super,
"journal_init_dev: journal device: %s\n",
bdevname(journal->j_dev_bd, b));
return 0;
}
/**
* When creating/tuning a file system user can assign some
* journal params within boundaries which depend on the ratio
* blocksize/standard_blocksize.
*
* For blocks >= standard_blocksize transaction size should
* be not less then JOURNAL_TRANS_MIN_DEFAULT, and not more
* then JOURNAL_TRANS_MAX_DEFAULT.
*
* For blocks < standard_blocksize these boundaries should be
* decreased proportionally.
*/
#define REISERFS_STANDARD_BLKSIZE (4096)
static int check_advise_trans_params(struct super_block *sb,
struct reiserfs_journal *journal)
{
if (journal->j_trans_max) {
/* Non-default journal params.
Do sanity check for them. */
int ratio = 1;
if (sb->s_blocksize < REISERFS_STANDARD_BLKSIZE)
ratio = REISERFS_STANDARD_BLKSIZE / sb->s_blocksize;
if (journal->j_trans_max > JOURNAL_TRANS_MAX_DEFAULT / ratio ||
journal->j_trans_max < JOURNAL_TRANS_MIN_DEFAULT / ratio ||
SB_ONDISK_JOURNAL_SIZE(sb) / journal->j_trans_max <
JOURNAL_MIN_RATIO) {
reiserfs_warning(sb, "sh-462",
"bad transaction max size (%u). "
"FSCK?", journal->j_trans_max);
return 1;
}
if (journal->j_max_batch != (journal->j_trans_max) *
JOURNAL_MAX_BATCH_DEFAULT/JOURNAL_TRANS_MAX_DEFAULT) {
reiserfs_warning(sb, "sh-463",
"bad transaction max batch (%u). "
"FSCK?", journal->j_max_batch);
return 1;
}
} else {
/* Default journal params.
The file system was created by old version
of mkreiserfs, so some fields contain zeros,
and we need to advise proper values for them */
if (sb->s_blocksize != REISERFS_STANDARD_BLKSIZE) {
reiserfs_warning(sb, "sh-464", "bad blocksize (%u)",
sb->s_blocksize);
return 1;
}
journal->j_trans_max = JOURNAL_TRANS_MAX_DEFAULT;
journal->j_max_batch = JOURNAL_MAX_BATCH_DEFAULT;
journal->j_max_commit_age = JOURNAL_MAX_COMMIT_AGE;
}
return 0;
}
/*
** must be called once on fs mount. calls journal_read for you
*/
int journal_init(struct super_block *sb, const char *j_dev_name,
int old_format, unsigned int commit_max_age)
{
int num_cnodes = SB_ONDISK_JOURNAL_SIZE(sb) * 2;
struct buffer_head *bhjh;
struct reiserfs_super_block *rs;
struct reiserfs_journal_header *jh;
struct reiserfs_journal *journal;
struct reiserfs_journal_list *jl;
char b[BDEVNAME_SIZE];
journal = SB_JOURNAL(sb) = vmalloc(sizeof(struct reiserfs_journal));
if (!journal) {
reiserfs_warning(sb, "journal-1256",
"unable to get memory for journal structure");
return 1;
}
memset(journal, 0, sizeof(struct reiserfs_journal));
INIT_LIST_HEAD(&journal->j_bitmap_nodes);
INIT_LIST_HEAD(&journal->j_prealloc_list);
INIT_LIST_HEAD(&journal->j_working_list);
INIT_LIST_HEAD(&journal->j_journal_list);
journal->j_persistent_trans = 0;
if (reiserfs_allocate_list_bitmaps(sb,
journal->j_list_bitmap,
reiserfs_bmap_count(sb)))
goto free_and_return;
allocate_bitmap_nodes(sb);
/* reserved for journal area support */
SB_JOURNAL_1st_RESERVED_BLOCK(sb) = (old_format ?
REISERFS_OLD_DISK_OFFSET_IN_BYTES
/ sb->s_blocksize +
reiserfs_bmap_count(sb) +
1 :
REISERFS_DISK_OFFSET_IN_BYTES /
sb->s_blocksize + 2);
/* Sanity check to see is the standard journal fitting withing first bitmap
(actual for small blocksizes) */
if (!SB_ONDISK_JOURNAL_DEVICE(sb) &&
(SB_JOURNAL_1st_RESERVED_BLOCK(sb) +
SB_ONDISK_JOURNAL_SIZE(sb) > sb->s_blocksize * 8)) {
reiserfs_warning(sb, "journal-1393",
"journal does not fit for area addressed "
"by first of bitmap blocks. It starts at "
"%u and its size is %u. Block size %ld",
SB_JOURNAL_1st_RESERVED_BLOCK(sb),
SB_ONDISK_JOURNAL_SIZE(sb),
sb->s_blocksize);
goto free_and_return;
}
/*
* We need to unlock here to avoid creating the following
* dependency:
* reiserfs_lock -> sysfs_mutex
* Because the reiserfs mmap path creates the following dependency:
* mm->mmap -> reiserfs_lock, hence we have
* mm->mmap -> reiserfs_lock ->sysfs_mutex
* This would ends up in a circular dependency with sysfs readdir path
* which does sysfs_mutex -> mm->mmap_sem
* This is fine because the reiserfs lock is useless in mount path,
* at least until we call journal_begin. We keep it for paranoid
* reasons.
*/
reiserfs_write_unlock(sb);
if (journal_init_dev(sb, journal, j_dev_name) != 0) {
reiserfs_write_lock(sb);
reiserfs_warning(sb, "sh-462",
"unable to initialize jornal device");
goto free_and_return;
}
reiserfs_write_lock(sb);
rs = SB_DISK_SUPER_BLOCK(sb);
/* read journal header */
bhjh = journal_bread(sb,
SB_ONDISK_JOURNAL_1st_BLOCK(sb) +
SB_ONDISK_JOURNAL_SIZE(sb));
if (!bhjh) {
reiserfs_warning(sb, "sh-459",
"unable to read journal header");
goto free_and_return;
}
jh = (struct reiserfs_journal_header *)(bhjh->b_data);
/* make sure that journal matches to the super block */
if (is_reiserfs_jr(rs)
&& (le32_to_cpu(jh->jh_journal.jp_journal_magic) !=
sb_jp_journal_magic(rs))) {
reiserfs_warning(sb, "sh-460",
"journal header magic %x (device %s) does "
"not match to magic found in super block %x",
jh->jh_journal.jp_journal_magic,
bdevname(journal->j_dev_bd, b),
sb_jp_journal_magic(rs));
brelse(bhjh);
goto free_and_return;
}
journal->j_trans_max = le32_to_cpu(jh->jh_journal.jp_journal_trans_max);
journal->j_max_batch = le32_to_cpu(jh->jh_journal.jp_journal_max_batch);
journal->j_max_commit_age =
le32_to_cpu(jh->jh_journal.jp_journal_max_commit_age);
journal->j_max_trans_age = JOURNAL_MAX_TRANS_AGE;
if (check_advise_trans_params(sb, journal) != 0)
goto free_and_return;
journal->j_default_max_commit_age = journal->j_max_commit_age;
if (commit_max_age != 0) {
journal->j_max_commit_age = commit_max_age;
journal->j_max_trans_age = commit_max_age;
}
reiserfs_info(sb, "journal params: device %s, size %u, "
"journal first block %u, max trans len %u, max batch %u, "
"max commit age %u, max trans age %u\n",
bdevname(journal->j_dev_bd, b),
SB_ONDISK_JOURNAL_SIZE(sb),
SB_ONDISK_JOURNAL_1st_BLOCK(sb),
journal->j_trans_max,
journal->j_max_batch,
journal->j_max_commit_age, journal->j_max_trans_age);
brelse(bhjh);
journal->j_list_bitmap_index = 0;
journal_list_init(sb);
memset(journal->j_list_hash_table, 0,
JOURNAL_HASH_SIZE * sizeof(struct reiserfs_journal_cnode *));
INIT_LIST_HEAD(&journal->j_dirty_buffers);
spin_lock_init(&journal->j_dirty_buffers_lock);
journal->j_start = 0;
journal->j_len = 0;
journal->j_len_alloc = 0;
atomic_set(&(journal->j_wcount), 0);
atomic_set(&(journal->j_async_throttle), 0);
journal->j_bcount = 0;
journal->j_trans_start_time = 0;
journal->j_last = NULL;
journal->j_first = NULL;
init_waitqueue_head(&(journal->j_join_wait));
mutex_init(&journal->j_mutex);
mutex_init(&journal->j_flush_mutex);
journal->j_trans_id = 10;
journal->j_mount_id = 10;
journal->j_state = 0;
atomic_set(&(journal->j_jlock), 0);
journal->j_cnode_free_list = allocate_cnodes(num_cnodes);
journal->j_cnode_free_orig = journal->j_cnode_free_list;
journal->j_cnode_free = journal->j_cnode_free_list ? num_cnodes : 0;
journal->j_cnode_used = 0;
journal->j_must_wait = 0;
if (journal->j_cnode_free == 0) {
reiserfs_warning(sb, "journal-2004", "Journal cnode memory "
"allocation failed (%ld bytes). Journal is "
"too large for available memory. Usually "
"this is due to a journal that is too large.",
sizeof (struct reiserfs_journal_cnode) * num_cnodes);
goto free_and_return;
}
init_journal_hash(sb);
jl = journal->j_current_jl;
jl->j_list_bitmap = get_list_bitmap(sb, jl);
if (!jl->j_list_bitmap) {
reiserfs_warning(sb, "journal-2005",
"get_list_bitmap failed for journal list 0");
goto free_and_return;
}
if (journal_read(sb) < 0) {
reiserfs_warning(sb, "reiserfs-2006",
"Replay Failure, unable to mount");
goto free_and_return;
}
reiserfs_mounted_fs_count++;
if (reiserfs_mounted_fs_count <= 1) {
reiserfs_write_unlock(sb);
commit_wq = create_workqueue("reiserfs");
reiserfs_write_lock(sb);
}
INIT_DELAYED_WORK(&journal->j_work, flush_async_commits);
journal->j_work_sb = sb;
return 0;
free_and_return:
free_journal_ram(sb);
return 1;
}
/*
** test for a polite end of the current transaction. Used by file_write, and should
** be used by delete to make sure they don't write more than can fit inside a single
** transaction
*/
int journal_transaction_should_end(struct reiserfs_transaction_handle *th,
int new_alloc)
{
struct reiserfs_journal *journal = SB_JOURNAL(th->t_super);
time_t now = get_seconds();
/* cannot restart while nested */
BUG_ON(!th->t_trans_id);
if (th->t_refcount > 1)
return 0;
if (journal->j_must_wait > 0 ||
(journal->j_len_alloc + new_alloc) >= journal->j_max_batch ||
atomic_read(&(journal->j_jlock)) ||
(now - journal->j_trans_start_time) > journal->j_max_trans_age ||
journal->j_cnode_free < (journal->j_trans_max * 3)) {
return 1;
}
/* protected by the BKL here */
journal->j_len_alloc += new_alloc;
th->t_blocks_allocated += new_alloc ;
return 0;
}
/* this must be called inside a transaction, and requires the
** kernel_lock to be held
*/
void reiserfs_block_writes(struct reiserfs_transaction_handle *th)
{
struct reiserfs_journal *journal = SB_JOURNAL(th->t_super);
BUG_ON(!th->t_trans_id);
journal->j_must_wait = 1;
set_bit(J_WRITERS_BLOCKED, &journal->j_state);
return;
}
/* this must be called without a transaction started, and does not
** require BKL
*/
void reiserfs_allow_writes(struct super_block *s)
{
struct reiserfs_journal *journal = SB_JOURNAL(s);
clear_bit(J_WRITERS_BLOCKED, &journal->j_state);
wake_up(&journal->j_join_wait);
}
/* this must be called without a transaction started, and does not
** require BKL
*/
void reiserfs_wait_on_write_block(struct super_block *s)
{
struct reiserfs_journal *journal = SB_JOURNAL(s);
wait_event(journal->j_join_wait,
!test_bit(J_WRITERS_BLOCKED, &journal->j_state));
}
static void queue_log_writer(struct super_block *s)
{
wait_queue_t wait;
struct reiserfs_journal *journal = SB_JOURNAL(s);
set_bit(J_WRITERS_QUEUED, &journal->j_state);
/*
* we don't want to use wait_event here because
* we only want to wait once.
*/
init_waitqueue_entry(&wait, current);
add_wait_queue(&journal->j_join_wait, &wait);
set_current_state(TASK_UNINTERRUPTIBLE);
if (test_bit(J_WRITERS_QUEUED, &journal->j_state)) {
reiserfs_write_unlock(s);
schedule();
reiserfs_write_lock(s);
}
__set_current_state(TASK_RUNNING);
remove_wait_queue(&journal->j_join_wait, &wait);
}
static void wake_queued_writers(struct super_block *s)
{
struct reiserfs_journal *journal = SB_JOURNAL(s);
if (test_and_clear_bit(J_WRITERS_QUEUED, &journal->j_state))
wake_up(&journal->j_join_wait);
}
static void let_transaction_grow(struct super_block *sb, unsigned int trans_id)
{
struct reiserfs_journal *journal = SB_JOURNAL(sb);
unsigned long bcount = journal->j_bcount;
while (1) {
reiserfs_write_unlock(sb);
schedule_timeout_uninterruptible(1);
reiserfs_write_lock(sb);
journal->j_current_jl->j_state |= LIST_COMMIT_PENDING;
while ((atomic_read(&journal->j_wcount) > 0 ||
atomic_read(&journal->j_jlock)) &&
journal->j_trans_id == trans_id) {
queue_log_writer(sb);
}
if (journal->j_trans_id != trans_id)
break;
if (bcount == journal->j_bcount)
break;
bcount = journal->j_bcount;
}
}
/* join == true if you must join an existing transaction.
** join == false if you can deal with waiting for others to finish
**
** this will block until the transaction is joinable. send the number of blocks you
** expect to use in nblocks.
*/
static int do_journal_begin_r(struct reiserfs_transaction_handle *th,
struct super_block *sb, unsigned long nblocks,
int join)
{
time_t now = get_seconds();
unsigned int old_trans_id;
struct reiserfs_journal *journal = SB_JOURNAL(sb);
struct reiserfs_transaction_handle myth;
int sched_count = 0;
int retval;
reiserfs_check_lock_depth(sb, "journal_begin");
BUG_ON(nblocks > journal->j_trans_max);
PROC_INFO_INC(sb, journal.journal_being);
/* set here for journal_join */
th->t_refcount = 1;
th->t_super = sb;
relock:
lock_journal(sb);
if (join != JBEGIN_ABORT && reiserfs_is_journal_aborted(journal)) {
unlock_journal(sb);
retval = journal->j_errno;
goto out_fail;
}
journal->j_bcount++;
if (test_bit(J_WRITERS_BLOCKED, &journal->j_state)) {
unlock_journal(sb);
reiserfs_write_unlock(sb);
reiserfs_wait_on_write_block(sb);
reiserfs_write_lock(sb);
PROC_INFO_INC(sb, journal.journal_relock_writers);
goto relock;
}
now = get_seconds();
/* if there is no room in the journal OR
** if this transaction is too old, and we weren't called joinable, wait for it to finish before beginning
** we don't sleep if there aren't other writers
*/
if ((!join && journal->j_must_wait > 0) ||
(!join
&& (journal->j_len_alloc + nblocks + 2) >= journal->j_max_batch)
|| (!join && atomic_read(&journal->j_wcount) > 0
&& journal->j_trans_start_time > 0
&& (now - journal->j_trans_start_time) >
journal->j_max_trans_age) || (!join
&& atomic_read(&journal->j_jlock))
|| (!join && journal->j_cnode_free < (journal->j_trans_max * 3))) {
old_trans_id = journal->j_trans_id;
unlock_journal(sb); /* allow others to finish this transaction */
if (!join && (journal->j_len_alloc + nblocks + 2) >=
journal->j_max_batch &&
((journal->j_len + nblocks + 2) * 100) <
(journal->j_len_alloc * 75)) {
if (atomic_read(&journal->j_wcount) > 10) {
sched_count++;
queue_log_writer(sb);
goto relock;
}
}
/* don't mess with joining the transaction if all we have to do is
* wait for someone else to do a commit
*/
if (atomic_read(&journal->j_jlock)) {
while (journal->j_trans_id == old_trans_id &&
atomic_read(&journal->j_jlock)) {
queue_log_writer(sb);
}
goto relock;
}
retval = journal_join(&myth, sb, 1);
if (retval)
goto out_fail;
/* someone might have ended the transaction while we joined */
if (old_trans_id != journal->j_trans_id) {
retval = do_journal_end(&myth, sb, 1, 0);
} else {
retval = do_journal_end(&myth, sb, 1, COMMIT_NOW);
}
if (retval)
goto out_fail;
PROC_INFO_INC(sb, journal.journal_relock_wcount);
goto relock;
}
/* we are the first writer, set trans_id */
if (journal->j_trans_start_time == 0) {
journal->j_trans_start_time = get_seconds();
}
atomic_inc(&(journal->j_wcount));
journal->j_len_alloc += nblocks;
th->t_blocks_logged = 0;
th->t_blocks_allocated = nblocks;
th->t_trans_id = journal->j_trans_id;
unlock_journal(sb);
INIT_LIST_HEAD(&th->t_list);
get_fs_excl();
return 0;
out_fail:
memset(th, 0, sizeof(*th));
/* Re-set th->t_super, so we can properly keep track of how many
* persistent transactions there are. We need to do this so if this
* call is part of a failed restart_transaction, we can free it later */
th->t_super = sb;
return retval;
}
struct reiserfs_transaction_handle *reiserfs_persistent_transaction(struct
super_block
*s,
int nblocks)
{
int ret;
struct reiserfs_transaction_handle *th;
/* if we're nesting into an existing transaction. It will be
** persistent on its own
*/
if (reiserfs_transaction_running(s)) {
th = current->journal_info;
th->t_refcount++;
BUG_ON(th->t_refcount < 2);
return th;
}
th = kmalloc(sizeof(struct reiserfs_transaction_handle), GFP_NOFS);
if (!th)
return NULL;
ret = journal_begin(th, s, nblocks);
if (ret) {
kfree(th);
return NULL;
}
SB_JOURNAL(s)->j_persistent_trans++;
return th;
}
int reiserfs_end_persistent_transaction(struct reiserfs_transaction_handle *th)
{
struct super_block *s = th->t_super;
int ret = 0;
if (th->t_trans_id)
ret = journal_end(th, th->t_super, th->t_blocks_allocated);
else
ret = -EIO;
if (th->t_refcount == 0) {
SB_JOURNAL(s)->j_persistent_trans--;
kfree(th);
}
return ret;
}
static int journal_join(struct reiserfs_transaction_handle *th,
struct super_block *sb, unsigned long nblocks)
{
struct reiserfs_transaction_handle *cur_th = current->journal_info;
/* this keeps do_journal_end from NULLing out the current->journal_info
** pointer
*/
th->t_handle_save = cur_th;
BUG_ON(cur_th && cur_th->t_refcount > 1);
return do_journal_begin_r(th, sb, nblocks, JBEGIN_JOIN);
}
int journal_join_abort(struct reiserfs_transaction_handle *th,
struct super_block *sb, unsigned long nblocks)
{
struct reiserfs_transaction_handle *cur_th = current->journal_info;
/* this keeps do_journal_end from NULLing out the current->journal_info
** pointer
*/
th->t_handle_save = cur_th;
BUG_ON(cur_th && cur_th->t_refcount > 1);
return do_journal_begin_r(th, sb, nblocks, JBEGIN_ABORT);
}
int journal_begin(struct reiserfs_transaction_handle *th,
struct super_block *sb, unsigned long nblocks)
{
struct reiserfs_transaction_handle *cur_th = current->journal_info;
int ret;
th->t_handle_save = NULL;
if (cur_th) {
/* we are nesting into the current transaction */
if (cur_th->t_super == sb) {
BUG_ON(!cur_th->t_refcount);
cur_th->t_refcount++;
memcpy(th, cur_th, sizeof(*th));
if (th->t_refcount <= 1)
reiserfs_warning(sb, "reiserfs-2005",
"BAD: refcount <= 1, but "
"journal_info != 0");
return 0;
} else {
/* we've ended up with a handle from a different filesystem.
** save it and restore on journal_end. This should never
** really happen...
*/
reiserfs_warning(sb, "clm-2100",
"nesting info a different FS");
th->t_handle_save = current->journal_info;
current->journal_info = th;
}
} else {
current->journal_info = th;
}
ret = do_journal_begin_r(th, sb, nblocks, JBEGIN_REG);
BUG_ON(current->journal_info != th);
/* I guess this boils down to being the reciprocal of clm-2100 above.
* If do_journal_begin_r fails, we need to put it back, since journal_end
* won't be called to do it. */
if (ret)
current->journal_info = th->t_handle_save;
else
BUG_ON(!th->t_refcount);
return ret;
}
/*
** puts bh into the current transaction. If it was already there, reorders removes the
** old pointers from the hash, and puts new ones in (to make sure replay happen in the right order).
**
** if it was dirty, cleans and files onto the clean list. I can't let it be dirty again until the
** transaction is committed.
**
** if j_len, is bigger than j_len_alloc, it pushes j_len_alloc to 10 + j_len.
*/
int journal_mark_dirty(struct reiserfs_transaction_handle *th,
struct super_block *sb, struct buffer_head *bh)
{
struct reiserfs_journal *journal = SB_JOURNAL(sb);
struct reiserfs_journal_cnode *cn = NULL;
int count_already_incd = 0;
int prepared = 0;
BUG_ON(!th->t_trans_id);
PROC_INFO_INC(sb, journal.mark_dirty);
if (th->t_trans_id != journal->j_trans_id) {
reiserfs_panic(th->t_super, "journal-1577",
"handle trans id %ld != current trans id %ld",
th->t_trans_id, journal->j_trans_id);
}
sb->s_dirt = 1;
prepared = test_clear_buffer_journal_prepared(bh);
clear_buffer_journal_restore_dirty(bh);
/* already in this transaction, we are done */
if (buffer_journaled(bh)) {
PROC_INFO_INC(sb, journal.mark_dirty_already);
return 0;
}
/* this must be turned into a panic instead of a warning. We can't allow
** a dirty or journal_dirty or locked buffer to be logged, as some changes
** could get to disk too early. NOT GOOD.
*/
if (!prepared || buffer_dirty(bh)) {
reiserfs_warning(sb, "journal-1777",
"buffer %llu bad state "
"%cPREPARED %cLOCKED %cDIRTY %cJDIRTY_WAIT",
(unsigned long long)bh->b_blocknr,
prepared ? ' ' : '!',
buffer_locked(bh) ? ' ' : '!',
buffer_dirty(bh) ? ' ' : '!',
buffer_journal_dirty(bh) ? ' ' : '!');
}
if (atomic_read(&(journal->j_wcount)) <= 0) {
reiserfs_warning(sb, "journal-1409",
"returning because j_wcount was %d",
atomic_read(&(journal->j_wcount)));
return 1;
}
/* this error means I've screwed up, and we've overflowed the transaction.
** Nothing can be done here, except make the FS readonly or panic.
*/
if (journal->j_len >= journal->j_trans_max) {
reiserfs_panic(th->t_super, "journal-1413",
"j_len (%lu) is too big",
journal->j_len);
}
if (buffer_journal_dirty(bh)) {
count_already_incd = 1;
PROC_INFO_INC(sb, journal.mark_dirty_notjournal);
clear_buffer_journal_dirty(bh);
}
if (journal->j_len > journal->j_len_alloc) {
journal->j_len_alloc = journal->j_len + JOURNAL_PER_BALANCE_CNT;
}
set_buffer_journaled(bh);
/* now put this guy on the end */
if (!cn) {
cn = get_cnode(sb);
if (!cn) {
reiserfs_panic(sb, "journal-4", "get_cnode failed!");
}
if (th->t_blocks_logged == th->t_blocks_allocated) {
th->t_blocks_allocated += JOURNAL_PER_BALANCE_CNT;
journal->j_len_alloc += JOURNAL_PER_BALANCE_CNT;
}
th->t_blocks_logged++;
journal->j_len++;
cn->bh = bh;
cn->blocknr = bh->b_blocknr;
cn->sb = sb;
cn->jlist = NULL;
insert_journal_hash(journal->j_hash_table, cn);
if (!count_already_incd) {
get_bh(bh);
}
}
cn->next = NULL;
cn->prev = journal->j_last;
cn->bh = bh;
if (journal->j_last) {
journal->j_last->next = cn;
journal->j_last = cn;
} else {
journal->j_first = cn;
journal->j_last = cn;
}
return 0;
}
int journal_end(struct reiserfs_transaction_handle *th,
struct super_block *sb, unsigned long nblocks)
{
if (!current->journal_info && th->t_refcount > 1)
reiserfs_warning(sb, "REISER-NESTING",
"th NULL, refcount %d", th->t_refcount);
if (!th->t_trans_id) {
WARN_ON(1);
return -EIO;
}
th->t_refcount--;
if (th->t_refcount > 0) {
struct reiserfs_transaction_handle *cur_th =
current->journal_info;
/* we aren't allowed to close a nested transaction on a different
** filesystem from the one in the task struct
*/
BUG_ON(cur_th->t_super != th->t_super);
if (th != cur_th) {
memcpy(current->journal_info, th, sizeof(*th));
th->t_trans_id = 0;
}
return 0;
} else {
return do_journal_end(th, sb, nblocks, 0);
}
}
/* removes from the current transaction, relsing and descrementing any counters.
** also files the removed buffer directly onto the clean list
**
** called by journal_mark_freed when a block has been deleted
**
** returns 1 if it cleaned and relsed the buffer. 0 otherwise
*/
static int remove_from_transaction(struct super_block *sb,
b_blocknr_t blocknr, int already_cleaned)
{
struct buffer_head *bh;
struct reiserfs_journal_cnode *cn;
struct reiserfs_journal *journal = SB_JOURNAL(sb);
int ret = 0;
cn = get_journal_hash_dev(sb, journal->j_hash_table, blocknr);
if (!cn || !cn->bh) {
return ret;
}
bh = cn->bh;
if (cn->prev) {
cn->prev->next = cn->next;
}
if (cn->next) {
cn->next->prev = cn->prev;
}
if (cn == journal->j_first) {
journal->j_first = cn->next;
}
if (cn == journal->j_last) {
journal->j_last = cn->prev;
}
if (bh)
remove_journal_hash(sb, journal->j_hash_table, NULL,
bh->b_blocknr, 0);
clear_buffer_journaled(bh); /* don't log this one */
if (!already_cleaned) {
clear_buffer_journal_dirty(bh);
clear_buffer_dirty(bh);
clear_buffer_journal_test(bh);
put_bh(bh);
if (atomic_read(&(bh->b_count)) < 0) {
reiserfs_warning(sb, "journal-1752",
"b_count < 0");
}
ret = 1;
}
journal->j_len--;
journal->j_len_alloc--;
free_cnode(sb, cn);
return ret;
}
/*
** for any cnode in a journal list, it can only be dirtied of all the
** transactions that include it are committed to disk.
** this checks through each transaction, and returns 1 if you are allowed to dirty,
** and 0 if you aren't
**
** it is called by dirty_journal_list, which is called after flush_commit_list has gotten all the log
** blocks for a given transaction on disk
**
*/
static int can_dirty(struct reiserfs_journal_cnode *cn)
{
struct super_block *sb = cn->sb;
b_blocknr_t blocknr = cn->blocknr;
struct reiserfs_journal_cnode *cur = cn->hprev;
int can_dirty = 1;
/* first test hprev. These are all newer than cn, so any node here
** with the same block number and dev means this node can't be sent
** to disk right now.
*/
while (cur && can_dirty) {
if (cur->jlist && cur->bh && cur->blocknr && cur->sb == sb &&
cur->blocknr == blocknr) {
can_dirty = 0;
}
cur = cur->hprev;
}
/* then test hnext. These are all older than cn. As long as they
** are committed to the log, it is safe to write cn to disk
*/
cur = cn->hnext;
while (cur && can_dirty) {
if (cur->jlist && cur->jlist->j_len > 0 &&
atomic_read(&(cur->jlist->j_commit_left)) > 0 && cur->bh &&
cur->blocknr && cur->sb == sb && cur->blocknr == blocknr) {
can_dirty = 0;
}
cur = cur->hnext;
}
return can_dirty;
}
/* syncs the commit blocks, but does not force the real buffers to disk
** will wait until the current transaction is done/committed before returning
*/
int journal_end_sync(struct reiserfs_transaction_handle *th,
struct super_block *sb, unsigned long nblocks)
{
struct reiserfs_journal *journal = SB_JOURNAL(sb);
BUG_ON(!th->t_trans_id);
/* you can sync while nested, very, very bad */
BUG_ON(th->t_refcount > 1);
if (journal->j_len == 0) {
reiserfs_prepare_for_journal(sb, SB_BUFFER_WITH_SB(sb),
1);
journal_mark_dirty(th, sb, SB_BUFFER_WITH_SB(sb));
}
return do_journal_end(th, sb, nblocks, COMMIT_NOW | WAIT);
}
/*
** writeback the pending async commits to disk
*/
static void flush_async_commits(struct work_struct *work)
{
struct reiserfs_journal *journal =
container_of(work, struct reiserfs_journal, j_work.work);
struct super_block *sb = journal->j_work_sb;
struct reiserfs_journal_list *jl;
struct list_head *entry;
reiserfs_write_lock(sb);
if (!list_empty(&journal->j_journal_list)) {
/* last entry is the youngest, commit it and you get everything */
entry = journal->j_journal_list.prev;
jl = JOURNAL_LIST_ENTRY(entry);
flush_commit_list(sb, jl, 1);
}
reiserfs_write_unlock(sb);
}
/*
** flushes any old transactions to disk
** ends the current transaction if it is too old
*/
int reiserfs_flush_old_commits(struct super_block *sb)
{
time_t now;
struct reiserfs_transaction_handle th;
struct reiserfs_journal *journal = SB_JOURNAL(sb);
now = get_seconds();
/* safety check so we don't flush while we are replaying the log during
* mount
*/
if (list_empty(&journal->j_journal_list)) {
return 0;
}
/* check the current transaction. If there are no writers, and it is
* too old, finish it, and force the commit blocks to disk
*/
if (atomic_read(&journal->j_wcount) <= 0 &&
journal->j_trans_start_time > 0 &&
journal->j_len > 0 &&
(now - journal->j_trans_start_time) > journal->j_max_trans_age) {
if (!journal_join(&th, sb, 1)) {
reiserfs_prepare_for_journal(sb,
SB_BUFFER_WITH_SB(sb),
1);
journal_mark_dirty(&th, sb,
SB_BUFFER_WITH_SB(sb));
/* we're only being called from kreiserfsd, it makes no sense to do
** an async commit so that kreiserfsd can do it later
*/
do_journal_end(&th, sb, 1, COMMIT_NOW | WAIT);
}
}
return sb->s_dirt;
}
/*
** returns 0 if do_journal_end should return right away, returns 1 if do_journal_end should finish the commit
**
** if the current transaction is too old, but still has writers, this will wait on j_join_wait until all
** the writers are done. By the time it wakes up, the transaction it was called has already ended, so it just
** flushes the commit list and returns 0.
**
** Won't batch when flush or commit_now is set. Also won't batch when others are waiting on j_join_wait.
**
** Note, we can't allow the journal_end to proceed while there are still writers in the log.
*/
static int check_journal_end(struct reiserfs_transaction_handle *th,
struct super_block *sb, unsigned long nblocks,
int flags)
{
time_t now;
int flush = flags & FLUSH_ALL;
int commit_now = flags & COMMIT_NOW;
int wait_on_commit = flags & WAIT;
struct reiserfs_journal_list *jl;
struct reiserfs_journal *journal = SB_JOURNAL(sb);
BUG_ON(!th->t_trans_id);
if (th->t_trans_id != journal->j_trans_id) {
reiserfs_panic(th->t_super, "journal-1577",
"handle trans id %ld != current trans id %ld",
th->t_trans_id, journal->j_trans_id);
}
journal->j_len_alloc -= (th->t_blocks_allocated - th->t_blocks_logged);
if (atomic_read(&(journal->j_wcount)) > 0) { /* <= 0 is allowed. unmounting might not call begin */
atomic_dec(&(journal->j_wcount));
}
/* BUG, deal with case where j_len is 0, but people previously freed blocks need to be released
** will be dealt with by next transaction that actually writes something, but should be taken
** care of in this trans
*/
BUG_ON(journal->j_len == 0);
/* if wcount > 0, and we are called to with flush or commit_now,
** we wait on j_join_wait. We will wake up when the last writer has
** finished the transaction, and started it on its way to the disk.
** Then, we flush the commit or journal list, and just return 0
** because the rest of journal end was already done for this transaction.
*/
if (atomic_read(&(journal->j_wcount)) > 0) {
if (flush || commit_now) {
unsigned trans_id;
jl = journal->j_current_jl;
trans_id = jl->j_trans_id;
if (wait_on_commit)
jl->j_state |= LIST_COMMIT_PENDING;
atomic_set(&(journal->j_jlock), 1);
if (flush) {
journal->j_next_full_flush = 1;
}
unlock_journal(sb);
/* sleep while the current transaction is still j_jlocked */
while (journal->j_trans_id == trans_id) {
if (atomic_read(&journal->j_jlock)) {
queue_log_writer(sb);
} else {
lock_journal(sb);
if (journal->j_trans_id == trans_id) {
atomic_set(&(journal->j_jlock),
1);
}
unlock_journal(sb);
}
}
BUG_ON(journal->j_trans_id == trans_id);
if (commit_now
&& journal_list_still_alive(sb, trans_id)
&& wait_on_commit) {
flush_commit_list(sb, jl, 1);
}
return 0;
}
unlock_journal(sb);
return 0;
}
/* deal with old transactions where we are the last writers */
now = get_seconds();
if ((now - journal->j_trans_start_time) > journal->j_max_trans_age) {
commit_now = 1;
journal->j_next_async_flush = 1;
}
/* don't batch when someone is waiting on j_join_wait */
/* don't batch when syncing the commit or flushing the whole trans */
if (!(journal->j_must_wait > 0) && !(atomic_read(&(journal->j_jlock)))
&& !flush && !commit_now && (journal->j_len < journal->j_max_batch)
&& journal->j_len_alloc < journal->j_max_batch
&& journal->j_cnode_free > (journal->j_trans_max * 3)) {
journal->j_bcount++;
unlock_journal(sb);
return 0;
}
if (journal->j_start > SB_ONDISK_JOURNAL_SIZE(sb)) {
reiserfs_panic(sb, "journal-003",
"j_start (%ld) is too high",
journal->j_start);
}
return 1;
}
/*
** Does all the work that makes deleting blocks safe.
** when deleting a block mark BH_JNew, just remove it from the current transaction, clean it's buffer_head and move on.
**
** otherwise:
** set a bit for the block in the journal bitmap. That will prevent it from being allocated for unformatted nodes
** before this transaction has finished.
**
** mark any cnodes for this block as BLOCK_FREED, and clear their bh pointers. That will prevent any old transactions with
** this block from trying to flush to the real location. Since we aren't removing the cnode from the journal_list_hash,
** the block can't be reallocated yet.
**
** Then remove it from the current transaction, decrementing any counters and filing it on the clean list.
*/
int journal_mark_freed(struct reiserfs_transaction_handle *th,
struct super_block *sb, b_blocknr_t blocknr)
{
struct reiserfs_journal *journal = SB_JOURNAL(sb);
struct reiserfs_journal_cnode *cn = NULL;
struct buffer_head *bh = NULL;
struct reiserfs_list_bitmap *jb = NULL;
int cleaned = 0;
BUG_ON(!th->t_trans_id);
cn = get_journal_hash_dev(sb, journal->j_hash_table, blocknr);
if (cn && cn->bh) {
bh = cn->bh;
get_bh(bh);
}
/* if it is journal new, we just remove it from this transaction */
if (bh && buffer_journal_new(bh)) {
clear_buffer_journal_new(bh);
clear_prepared_bits(bh);
reiserfs_clean_and_file_buffer(bh);
cleaned = remove_from_transaction(sb, blocknr, cleaned);
} else {
/* set the bit for this block in the journal bitmap for this transaction */
jb = journal->j_current_jl->j_list_bitmap;
if (!jb) {
reiserfs_panic(sb, "journal-1702",
"journal_list_bitmap is NULL");
}
set_bit_in_list_bitmap(sb, blocknr, jb);
/* Note, the entire while loop is not allowed to schedule. */
if (bh) {
clear_prepared_bits(bh);
reiserfs_clean_and_file_buffer(bh);
}
cleaned = remove_from_transaction(sb, blocknr, cleaned);
/* find all older transactions with this block, make sure they don't try to write it out */
cn = get_journal_hash_dev(sb, journal->j_list_hash_table,
blocknr);
while (cn) {
if (sb == cn->sb && blocknr == cn->blocknr) {
set_bit(BLOCK_FREED, &cn->state);
if (cn->bh) {
if (!cleaned) {
/* remove_from_transaction will brelse the buffer if it was
** in the current trans
*/
clear_buffer_journal_dirty(cn->
bh);
clear_buffer_dirty(cn->bh);
clear_buffer_journal_test(cn->
bh);
cleaned = 1;
put_bh(cn->bh);
if (atomic_read
(&(cn->bh->b_count)) < 0) {
reiserfs_warning(sb,
"journal-2138",
"cn->bh->b_count < 0");
}
}
if (cn->jlist) { /* since we are clearing the bh, we MUST dec nonzerolen */
atomic_dec(&
(cn->jlist->
j_nonzerolen));
}
cn->bh = NULL;
}
}
cn = cn->hnext;
}
}
if (bh)
release_buffer_page(bh); /* get_hash grabs the buffer */
return 0;
}
void reiserfs_update_inode_transaction(struct inode *inode)
{
struct reiserfs_journal *journal = SB_JOURNAL(inode->i_sb);
REISERFS_I(inode)->i_jl = journal->j_current_jl;
REISERFS_I(inode)->i_trans_id = journal->j_trans_id;
}
/*
* returns -1 on error, 0 if no commits/barriers were done and 1
* if a transaction was actually committed and the barrier was done
*/
static int __commit_trans_jl(struct inode *inode, unsigned long id,
struct reiserfs_journal_list *jl)
{
struct reiserfs_transaction_handle th;
struct super_block *sb = inode->i_sb;
struct reiserfs_journal *journal = SB_JOURNAL(sb);
int ret = 0;
/* is it from the current transaction, or from an unknown transaction? */
if (id == journal->j_trans_id) {
jl = journal->j_current_jl;
/* try to let other writers come in and grow this transaction */
let_transaction_grow(sb, id);
if (journal->j_trans_id != id) {
goto flush_commit_only;
}
ret = journal_begin(&th, sb, 1);
if (ret)
return ret;
/* someone might have ended this transaction while we joined */
if (journal->j_trans_id != id) {
reiserfs_prepare_for_journal(sb, SB_BUFFER_WITH_SB(sb),
1);
journal_mark_dirty(&th, sb, SB_BUFFER_WITH_SB(sb));
ret = journal_end(&th, sb, 1);
goto flush_commit_only;
}
ret = journal_end_sync(&th, sb, 1);
if (!ret)
ret = 1;
} else {
/* this gets tricky, we have to make sure the journal list in
* the inode still exists. We know the list is still around
* if we've got a larger transaction id than the oldest list
*/
flush_commit_only:
if (journal_list_still_alive(inode->i_sb, id)) {
/*
* we only set ret to 1 when we know for sure
* the barrier hasn't been started yet on the commit
* block.
*/
if (atomic_read(&jl->j_commit_left) > 1)
ret = 1;
flush_commit_list(sb, jl, 1);
if (journal->j_errno)
ret = journal->j_errno;
}
}
/* otherwise the list is gone, and long since committed */
return ret;
}
int reiserfs_commit_for_inode(struct inode *inode)
{
unsigned int id = REISERFS_I(inode)->i_trans_id;
struct reiserfs_journal_list *jl = REISERFS_I(inode)->i_jl;
/* for the whole inode, assume unset id means it was
* changed in the current transaction. More conservative
*/
if (!id || !jl) {
reiserfs_update_inode_transaction(inode);
id = REISERFS_I(inode)->i_trans_id;
/* jl will be updated in __commit_trans_jl */
}
return __commit_trans_jl(inode, id, jl);
}
void reiserfs_restore_prepared_buffer(struct super_block *sb,
struct buffer_head *bh)
{
struct reiserfs_journal *journal = SB_JOURNAL(sb);
PROC_INFO_INC(sb, journal.restore_prepared);
if (!bh) {
return;
}
if (test_clear_buffer_journal_restore_dirty(bh) &&
buffer_journal_dirty(bh)) {
struct reiserfs_journal_cnode *cn;
cn = get_journal_hash_dev(sb,
journal->j_list_hash_table,
bh->b_blocknr);
if (cn && can_dirty(cn)) {
set_buffer_journal_test(bh);
mark_buffer_dirty(bh);
}
}
clear_buffer_journal_prepared(bh);
}
extern struct tree_balance *cur_tb;
/*
** before we can change a metadata block, we have to make sure it won't
** be written to disk while we are altering it. So, we must:
** clean it
** wait on it.
**
*/
int reiserfs_prepare_for_journal(struct super_block *sb,
struct buffer_head *bh, int wait)
{
PROC_INFO_INC(sb, journal.prepare);
if (!trylock_buffer(bh)) {
if (!wait)
return 0;
lock_buffer(bh);
}
set_buffer_journal_prepared(bh);
if (test_clear_buffer_dirty(bh) && buffer_journal_dirty(bh)) {
clear_buffer_journal_test(bh);
set_buffer_journal_restore_dirty(bh);
}
unlock_buffer(bh);
return 1;
}
static void flush_old_journal_lists(struct super_block *s)
{
struct reiserfs_journal *journal = SB_JOURNAL(s);
struct reiserfs_journal_list *jl;
struct list_head *entry;
time_t now = get_seconds();
while (!list_empty(&journal->j_journal_list)) {
entry = journal->j_journal_list.next;
jl = JOURNAL_LIST_ENTRY(entry);
/* this check should always be run, to send old lists to disk */
if (jl->j_timestamp < (now - (JOURNAL_MAX_TRANS_AGE * 4)) &&
atomic_read(&jl->j_commit_left) == 0 &&
test_transaction(s, jl)) {
flush_used_journal_lists(s, jl);
} else {
break;
}
}
}
/*
** long and ugly. If flush, will not return until all commit
** blocks and all real buffers in the trans are on disk.
** If no_async, won't return until all commit blocks are on disk.
**
** keep reading, there are comments as you go along
**
** If the journal is aborted, we just clean up. Things like flushing
** journal lists, etc just won't happen.
*/
static int do_journal_end(struct reiserfs_transaction_handle *th,
struct super_block *sb, unsigned long nblocks,
int flags)
{
struct reiserfs_journal *journal = SB_JOURNAL(sb);
struct reiserfs_journal_cnode *cn, *next, *jl_cn;
struct reiserfs_journal_cnode *last_cn = NULL;
struct reiserfs_journal_desc *desc;
struct reiserfs_journal_commit *commit;
struct buffer_head *c_bh; /* commit bh */
struct buffer_head *d_bh; /* desc bh */
int cur_write_start = 0; /* start index of current log write */
int old_start;
int i;
int flush;
int wait_on_commit;
struct reiserfs_journal_list *jl, *temp_jl;
struct list_head *entry, *safe;
unsigned long jindex;
unsigned int commit_trans_id;
int trans_half;
BUG_ON(th->t_refcount > 1);
BUG_ON(!th->t_trans_id);
/* protect flush_older_commits from doing mistakes if the
transaction ID counter gets overflowed. */
if (th->t_trans_id == ~0U)
flags |= FLUSH_ALL | COMMIT_NOW | WAIT;
flush = flags & FLUSH_ALL;
wait_on_commit = flags & WAIT;
put_fs_excl();
current->journal_info = th->t_handle_save;
reiserfs_check_lock_depth(sb, "journal end");
if (journal->j_len == 0) {
reiserfs_prepare_for_journal(sb, SB_BUFFER_WITH_SB(sb),
1);
journal_mark_dirty(th, sb, SB_BUFFER_WITH_SB(sb));
}
lock_journal(sb);
if (journal->j_next_full_flush) {
flags |= FLUSH_ALL;
flush = 1;
}
if (journal->j_next_async_flush) {
flags |= COMMIT_NOW | WAIT;
wait_on_commit = 1;
}
/* check_journal_end locks the journal, and unlocks if it does not return 1
** it tells us if we should continue with the journal_end, or just return
*/
if (!check_journal_end(th, sb, nblocks, flags)) {
sb->s_dirt = 1;
wake_queued_writers(sb);
reiserfs_async_progress_wait(sb);
goto out;
}
/* check_journal_end might set these, check again */
if (journal->j_next_full_flush) {
flush = 1;
}
/*
** j must wait means we have to flush the log blocks, and the real blocks for
** this transaction
*/
if (journal->j_must_wait > 0) {
flush = 1;
}
#ifdef REISERFS_PREALLOCATE
/* quota ops might need to nest, setup the journal_info pointer for them
* and raise the refcount so that it is > 0. */
current->journal_info = th;
th->t_refcount++;
reiserfs_discard_all_prealloc(th); /* it should not involve new blocks into
* the transaction */
th->t_refcount--;
current->journal_info = th->t_handle_save;
#endif
/* setup description block */
d_bh =
journal_getblk(sb,
SB_ONDISK_JOURNAL_1st_BLOCK(sb) +
journal->j_start);
set_buffer_uptodate(d_bh);
desc = (struct reiserfs_journal_desc *)(d_bh)->b_data;
memset(d_bh->b_data, 0, d_bh->b_size);
memcpy(get_journal_desc_magic(d_bh), JOURNAL_DESC_MAGIC, 8);
set_desc_trans_id(desc, journal->j_trans_id);
/* setup commit block. Don't write (keep it clean too) this one until after everyone else is written */
c_bh = journal_getblk(sb, SB_ONDISK_JOURNAL_1st_BLOCK(sb) +
((journal->j_start + journal->j_len +
1) % SB_ONDISK_JOURNAL_SIZE(sb)));
commit = (struct reiserfs_journal_commit *)c_bh->b_data;
memset(c_bh->b_data, 0, c_bh->b_size);
set_commit_trans_id(commit, journal->j_trans_id);
set_buffer_uptodate(c_bh);
/* init this journal list */
jl = journal->j_current_jl;
/* we lock the commit before doing anything because
* we want to make sure nobody tries to run flush_commit_list until
* the new transaction is fully setup, and we've already flushed the
* ordered bh list
*/
reiserfs_mutex_lock_safe(&jl->j_commit_mutex, sb);
/* save the transaction id in case we need to commit it later */
commit_trans_id = jl->j_trans_id;
atomic_set(&jl->j_older_commits_done, 0);
jl->j_trans_id = journal->j_trans_id;
jl->j_timestamp = journal->j_trans_start_time;
jl->j_commit_bh = c_bh;
jl->j_start = journal->j_start;
jl->j_len = journal->j_len;
atomic_set(&jl->j_nonzerolen, journal->j_len);
atomic_set(&jl->j_commit_left, journal->j_len + 2);
jl->j_realblock = NULL;
/* The ENTIRE FOR LOOP MUST not cause schedule to occur.
** for each real block, add it to the journal list hash,
** copy into real block index array in the commit or desc block
*/
trans_half = journal_trans_half(sb->s_blocksize);
for (i = 0, cn = journal->j_first; cn; cn = cn->next, i++) {
if (buffer_journaled(cn->bh)) {
jl_cn = get_cnode(sb);
if (!jl_cn) {
reiserfs_panic(sb, "journal-1676",
"get_cnode returned NULL");
}
if (i == 0) {
jl->j_realblock = jl_cn;
}
jl_cn->prev = last_cn;
jl_cn->next = NULL;
if (last_cn) {
last_cn->next = jl_cn;
}
last_cn = jl_cn;
/* make sure the block we are trying to log is not a block
of journal or reserved area */
if (is_block_in_log_or_reserved_area
(sb, cn->bh->b_blocknr)) {
reiserfs_panic(sb, "journal-2332",
"Trying to log block %lu, "
"which is a log block",
cn->bh->b_blocknr);
}
jl_cn->blocknr = cn->bh->b_blocknr;
jl_cn->state = 0;
jl_cn->sb = sb;
jl_cn->bh = cn->bh;
jl_cn->jlist = jl;
insert_journal_hash(journal->j_list_hash_table, jl_cn);
if (i < trans_half) {
desc->j_realblock[i] =
cpu_to_le32(cn->bh->b_blocknr);
} else {
commit->j_realblock[i - trans_half] =
cpu_to_le32(cn->bh->b_blocknr);
}
} else {
i--;
}
}
set_desc_trans_len(desc, journal->j_len);
set_desc_mount_id(desc, journal->j_mount_id);
set_desc_trans_id(desc, journal->j_trans_id);
set_commit_trans_len(commit, journal->j_len);
/* special check in case all buffers in the journal were marked for not logging */
BUG_ON(journal->j_len == 0);
/* we're about to dirty all the log blocks, mark the description block
* dirty now too. Don't mark the commit block dirty until all the
* others are on disk
*/
mark_buffer_dirty(d_bh);
/* first data block is j_start + 1, so add one to cur_write_start wherever you use it */
cur_write_start = journal->j_start;
cn = journal->j_first;
jindex = 1; /* start at one so we don't get the desc again */
while (cn) {
clear_buffer_journal_new(cn->bh);
/* copy all the real blocks into log area. dirty log blocks */
if (buffer_journaled(cn->bh)) {
struct buffer_head *tmp_bh;
char *addr;
struct page *page;
tmp_bh =
journal_getblk(sb,
SB_ONDISK_JOURNAL_1st_BLOCK(sb) +
((cur_write_start +
jindex) %
SB_ONDISK_JOURNAL_SIZE(sb)));
set_buffer_uptodate(tmp_bh);
page = cn->bh->b_page;
addr = kmap(page);
memcpy(tmp_bh->b_data,
addr + offset_in_page(cn->bh->b_data),
cn->bh->b_size);
kunmap(page);
mark_buffer_dirty(tmp_bh);
jindex++;
set_buffer_journal_dirty(cn->bh);
clear_buffer_journaled(cn->bh);
} else {
/* JDirty cleared sometime during transaction. don't log this one */
reiserfs_warning(sb, "journal-2048",
"BAD, buffer in journal hash, "
"but not JDirty!");
brelse(cn->bh);
}
next = cn->next;
free_cnode(sb, cn);
cn = next;
reiserfs_write_unlock(sb);
cond_resched();
reiserfs_write_lock(sb);
}
/* we are done with both the c_bh and d_bh, but
** c_bh must be written after all other commit blocks,
** so we dirty/relse c_bh in flush_commit_list, with commit_left <= 1.
*/
journal->j_current_jl = alloc_journal_list(sb);
/* now it is safe to insert this transaction on the main list */
list_add_tail(&jl->j_list, &journal->j_journal_list);
list_add_tail(&jl->j_working_list, &journal->j_working_list);
journal->j_num_work_lists++;
/* reset journal values for the next transaction */
old_start = journal->j_start;
journal->j_start =
(journal->j_start + journal->j_len +
2) % SB_ONDISK_JOURNAL_SIZE(sb);
atomic_set(&(journal->j_wcount), 0);
journal->j_bcount = 0;
journal->j_last = NULL;
journal->j_first = NULL;
journal->j_len = 0;
journal->j_trans_start_time = 0;
/* check for trans_id overflow */
if (++journal->j_trans_id == 0)
journal->j_trans_id = 10;
journal->j_current_jl->j_trans_id = journal->j_trans_id;
journal->j_must_wait = 0;
journal->j_len_alloc = 0;
journal->j_next_full_flush = 0;
journal->j_next_async_flush = 0;
init_journal_hash(sb);
// make sure reiserfs_add_jh sees the new current_jl before we
// write out the tails
smp_mb();
/* tail conversion targets have to hit the disk before we end the
* transaction. Otherwise a later transaction might repack the tail
* before this transaction commits, leaving the data block unflushed and
* clean, if we crash before the later transaction commits, the data block
* is lost.
*/
if (!list_empty(&jl->j_tail_bh_list)) {
reiserfs_write_unlock(sb);
write_ordered_buffers(&journal->j_dirty_buffers_lock,
journal, jl, &jl->j_tail_bh_list);
reiserfs_write_lock(sb);
}
BUG_ON(!list_empty(&jl->j_tail_bh_list));
mutex_unlock(&jl->j_commit_mutex);
/* honor the flush wishes from the caller, simple commits can
** be done outside the journal lock, they are done below
**
** if we don't flush the commit list right now, we put it into
** the work queue so the people waiting on the async progress work
** queue don't wait for this proc to flush journal lists and such.
*/
if (flush) {
flush_commit_list(sb, jl, 1);
flush_journal_list(sb, jl, 1);
} else if (!(jl->j_state & LIST_COMMIT_PENDING))
queue_delayed_work(commit_wq, &journal->j_work, HZ / 10);
/* if the next transaction has any chance of wrapping, flush
** transactions that might get overwritten. If any journal lists are very
** old flush them as well.
*/
first_jl:
list_for_each_safe(entry, safe, &journal->j_journal_list) {
temp_jl = JOURNAL_LIST_ENTRY(entry);
if (journal->j_start <= temp_jl->j_start) {
if ((journal->j_start + journal->j_trans_max + 1) >=
temp_jl->j_start) {
flush_used_journal_lists(sb, temp_jl);
goto first_jl;
} else if ((journal->j_start +
journal->j_trans_max + 1) <
SB_ONDISK_JOURNAL_SIZE(sb)) {
/* if we don't cross into the next transaction and we don't
* wrap, there is no way we can overlap any later transactions
* break now
*/
break;
}
} else if ((journal->j_start +
journal->j_trans_max + 1) >
SB_ONDISK_JOURNAL_SIZE(sb)) {
if (((journal->j_start + journal->j_trans_max + 1) %
SB_ONDISK_JOURNAL_SIZE(sb)) >=
temp_jl->j_start) {
flush_used_journal_lists(sb, temp_jl);
goto first_jl;
} else {
/* we don't overlap anything from out start to the end of the
* log, and our wrapped portion doesn't overlap anything at
* the start of the log. We can break
*/
break;
}
}
}
flush_old_journal_lists(sb);
journal->j_current_jl->j_list_bitmap =
get_list_bitmap(sb, journal->j_current_jl);
if (!(journal->j_current_jl->j_list_bitmap)) {
reiserfs_panic(sb, "journal-1996",
"could not get a list bitmap");
}
atomic_set(&(journal->j_jlock), 0);
unlock_journal(sb);
/* wake up any body waiting to join. */
clear_bit(J_WRITERS_QUEUED, &journal->j_state);
wake_up(&(journal->j_join_wait));
if (!flush && wait_on_commit &&
journal_list_still_alive(sb, commit_trans_id)) {
flush_commit_list(sb, jl, 1);
}
out:
reiserfs_check_lock_depth(sb, "journal end2");
memset(th, 0, sizeof(*th));
/* Re-set th->t_super, so we can properly keep track of how many
* persistent transactions there are. We need to do this so if this
* call is part of a failed restart_transaction, we can free it later */
th->t_super = sb;
return journal->j_errno;
}
/* Send the file system read only and refuse new transactions */
void reiserfs_abort_journal(struct super_block *sb, int errno)
{
struct reiserfs_journal *journal = SB_JOURNAL(sb);
if (test_bit(J_ABORTED, &journal->j_state))
return;
if (!journal->j_errno)
journal->j_errno = errno;
sb->s_flags |= MS_RDONLY;
set_bit(J_ABORTED, &journal->j_state);
#ifdef CONFIG_REISERFS_CHECK
dump_stack();
#endif
}
Reference in New Issue View Git Blame Copy Permalink