summaryrefslogtreecommitdiffstats
path: root/fs/jbd2/journal.c
diff options
context:
space:
mode:
Diffstat (limited to 'fs/jbd2/journal.c')
-rw-r--r--fs/jbd2/journal.c280
1 files changed, 206 insertions, 74 deletions
diff --git a/fs/jbd2/journal.c b/fs/jbd2/journal.c
index ac0d027595d..538417c1fdb 100644
--- a/fs/jbd2/journal.c
+++ b/fs/jbd2/journal.c
@@ -39,15 +39,18 @@
#include <linux/seq_file.h>
#include <linux/math64.h>
#include <linux/hash.h>
+#include <linux/log2.h>
+#include <linux/vmalloc.h>
+#include <linux/backing-dev.h>
+#include <linux/bitops.h>
#define CREATE_TRACE_POINTS
#include <trace/events/jbd2.h>
#include <asm/uaccess.h>
#include <asm/page.h>
+#include <asm/system.h>
-EXPORT_SYMBOL(jbd2_journal_start);
-EXPORT_SYMBOL(jbd2_journal_restart);
EXPORT_SYMBOL(jbd2_journal_extend);
EXPORT_SYMBOL(jbd2_journal_stop);
EXPORT_SYMBOL(jbd2_journal_lock_updates);
@@ -93,6 +96,7 @@ EXPORT_SYMBOL(jbd2_journal_begin_ordered_truncate);
static int journal_convert_superblock_v1(journal_t *, journal_superblock_t *);
static void __journal_abort_soft (journal_t *journal, int errno);
+static int jbd2_journal_create_slab(size_t slab_size);
/*
* Helper function used to manage commit timeouts
@@ -140,7 +144,7 @@ static int kjournald2(void *arg)
/*
* And now, wait forever for commit wakeup events.
*/
- spin_lock(&journal->j_state_lock);
+ write_lock(&journal->j_state_lock);
loop:
if (journal->j_flags & JBD2_UNMOUNT)
@@ -151,10 +155,10 @@ loop:
if (journal->j_commit_sequence != journal->j_commit_request) {
jbd_debug(1, "OK, requests differ\n");
- spin_unlock(&journal->j_state_lock);
+ write_unlock(&journal->j_state_lock);
del_timer_sync(&journal->j_commit_timer);
jbd2_journal_commit_transaction(journal);
- spin_lock(&journal->j_state_lock);
+ write_lock(&journal->j_state_lock);
goto loop;
}
@@ -166,9 +170,9 @@ loop:
* be already stopped.
*/
jbd_debug(1, "Now suspending kjournald2\n");
- spin_unlock(&journal->j_state_lock);
+ write_unlock(&journal->j_state_lock);
refrigerator();
- spin_lock(&journal->j_state_lock);
+ write_lock(&journal->j_state_lock);
} else {
/*
* We assume on resume that commits are already there,
@@ -188,9 +192,9 @@ loop:
if (journal->j_flags & JBD2_UNMOUNT)
should_sleep = 0;
if (should_sleep) {
- spin_unlock(&journal->j_state_lock);
+ write_unlock(&journal->j_state_lock);
schedule();
- spin_lock(&journal->j_state_lock);
+ write_lock(&journal->j_state_lock);
}
finish_wait(&journal->j_wait_commit, &wait);
}
@@ -208,7 +212,7 @@ loop:
goto loop;
end_loop:
- spin_unlock(&journal->j_state_lock);
+ write_unlock(&journal->j_state_lock);
del_timer_sync(&journal->j_commit_timer);
journal->j_task = NULL;
wake_up(&journal->j_wait_done_commit);
@@ -231,16 +235,16 @@ static int jbd2_journal_start_thread(journal_t *journal)
static void journal_kill_thread(journal_t *journal)
{
- spin_lock(&journal->j_state_lock);
+ write_lock(&journal->j_state_lock);
journal->j_flags |= JBD2_UNMOUNT;
while (journal->j_task) {
wake_up(&journal->j_wait_commit);
- spin_unlock(&journal->j_state_lock);
+ write_unlock(&journal->j_state_lock);
wait_event(journal->j_wait_done_commit, journal->j_task == NULL);
- spin_lock(&journal->j_state_lock);
+ write_lock(&journal->j_state_lock);
}
- spin_unlock(&journal->j_state_lock);
+ write_unlock(&journal->j_state_lock);
}
/*
@@ -294,7 +298,6 @@ int jbd2_journal_write_metadata_buffer(transaction_t *transaction,
struct page *new_page;
unsigned int new_offset;
struct buffer_head *bh_in = jh2bh(jh_in);
- struct jbd2_buffer_trigger_type *triggers;
journal_t *journal = transaction->t_journal;
/*
@@ -308,7 +311,17 @@ int jbd2_journal_write_metadata_buffer(transaction_t *transaction,
*/
J_ASSERT_BH(bh_in, buffer_jbddirty(bh_in));
- new_bh = alloc_buffer_head(GFP_NOFS|__GFP_NOFAIL);
+retry_alloc:
+ new_bh = alloc_buffer_head(GFP_NOFS);
+ if (!new_bh) {
+ /*
+ * Failure is not an option, but __GFP_NOFAIL is going
+ * away; so we retry ourselves here.
+ */
+ congestion_wait(BLK_RW_ASYNC, HZ/50);
+ goto retry_alloc;
+ }
+
/* keep subsequent assertions sane */
new_bh->b_state = 0;
init_buffer(new_bh, NULL, NULL);
@@ -325,21 +338,21 @@ repeat:
done_copy_out = 1;
new_page = virt_to_page(jh_in->b_frozen_data);
new_offset = offset_in_page(jh_in->b_frozen_data);
- triggers = jh_in->b_frozen_triggers;
} else {
new_page = jh2bh(jh_in)->b_page;
new_offset = offset_in_page(jh2bh(jh_in)->b_data);
- triggers = jh_in->b_triggers;
}
mapped_data = kmap_atomic(new_page, KM_USER0);
/*
- * Fire any commit trigger. Do this before checking for escaping,
- * as the trigger may modify the magic offset. If a copy-out
- * happens afterwards, it will have the correct data in the buffer.
+ * Fire data frozen trigger if data already wasn't frozen. Do this
+ * before checking for escaping, as the trigger may modify the magic
+ * offset. If a copy-out happens afterwards, it will have the correct
+ * data in the buffer.
*/
- jbd2_buffer_commit_trigger(jh_in, mapped_data + new_offset,
- triggers);
+ if (!done_copy_out)
+ jbd2_buffer_frozen_trigger(jh_in, mapped_data + new_offset,
+ jh_in->b_triggers);
/*
* Check for escaping
@@ -440,7 +453,7 @@ int __jbd2_log_space_left(journal_t *journal)
{
int left = journal->j_free;
- assert_spin_locked(&journal->j_state_lock);
+ /* assert_spin_locked(&journal->j_state_lock); */
/*
* Be pessimistic here about the number of those free blocks which
@@ -467,7 +480,7 @@ int __jbd2_log_start_commit(journal_t *journal, tid_t target)
*/
if (!tid_geq(journal->j_commit_request, target)) {
/*
- * We want a new commit: OK, mark the request and wakup the
+ * We want a new commit: OK, mark the request and wakeup the
* commit thread. We do _not_ do the commit ourselves.
*/
@@ -485,9 +498,9 @@ int jbd2_log_start_commit(journal_t *journal, tid_t tid)
{
int ret;
- spin_lock(&journal->j_state_lock);
+ write_lock(&journal->j_state_lock);
ret = __jbd2_log_start_commit(journal, tid);
- spin_unlock(&journal->j_state_lock);
+ write_unlock(&journal->j_state_lock);
return ret;
}
@@ -506,7 +519,7 @@ int jbd2_journal_force_commit_nested(journal_t *journal)
transaction_t *transaction = NULL;
tid_t tid;
- spin_lock(&journal->j_state_lock);
+ read_lock(&journal->j_state_lock);
if (journal->j_running_transaction && !current->journal_info) {
transaction = journal->j_running_transaction;
__jbd2_log_start_commit(journal, transaction->t_tid);
@@ -514,12 +527,12 @@ int jbd2_journal_force_commit_nested(journal_t *journal)
transaction = journal->j_committing_transaction;
if (!transaction) {
- spin_unlock(&journal->j_state_lock);
+ read_unlock(&journal->j_state_lock);
return 0; /* Nothing to retry */
}
tid = transaction->t_tid;
- spin_unlock(&journal->j_state_lock);
+ read_unlock(&journal->j_state_lock);
jbd2_log_wait_commit(journal, tid);
return 1;
}
@@ -533,7 +546,7 @@ int jbd2_journal_start_commit(journal_t *journal, tid_t *ptid)
{
int ret = 0;
- spin_lock(&journal->j_state_lock);
+ write_lock(&journal->j_state_lock);
if (journal->j_running_transaction) {
tid_t tid = journal->j_running_transaction->t_tid;
@@ -552,7 +565,7 @@ int jbd2_journal_start_commit(journal_t *journal, tid_t *ptid)
*ptid = journal->j_committing_transaction->t_tid;
ret = 1;
}
- spin_unlock(&journal->j_state_lock);
+ write_unlock(&journal->j_state_lock);
return ret;
}
@@ -564,26 +577,24 @@ int jbd2_log_wait_commit(journal_t *journal, tid_t tid)
{
int err = 0;
+ read_lock(&journal->j_state_lock);
#ifdef CONFIG_JBD2_DEBUG
- spin_lock(&journal->j_state_lock);
if (!tid_geq(journal->j_commit_request, tid)) {
printk(KERN_EMERG
"%s: error: j_commit_request=%d, tid=%d\n",
__func__, journal->j_commit_request, tid);
}
- spin_unlock(&journal->j_state_lock);
#endif
- spin_lock(&journal->j_state_lock);
while (tid_gt(tid, journal->j_commit_sequence)) {
jbd_debug(1, "JBD: want %d, j_commit_sequence=%d\n",
tid, journal->j_commit_sequence);
wake_up(&journal->j_wait_commit);
- spin_unlock(&journal->j_state_lock);
+ read_unlock(&journal->j_state_lock);
wait_event(journal->j_wait_done_commit,
!tid_gt(tid, journal->j_commit_sequence));
- spin_lock(&journal->j_state_lock);
+ read_lock(&journal->j_state_lock);
}
- spin_unlock(&journal->j_state_lock);
+ read_unlock(&journal->j_state_lock);
if (unlikely(is_journal_aborted(journal))) {
printk(KERN_EMERG "journal commit I/O error\n");
@@ -600,7 +611,7 @@ int jbd2_journal_next_log_block(journal_t *journal, unsigned long long *retp)
{
unsigned long blocknr;
- spin_lock(&journal->j_state_lock);
+ write_lock(&journal->j_state_lock);
J_ASSERT(journal->j_free > 1);
blocknr = journal->j_head;
@@ -608,7 +619,7 @@ int jbd2_journal_next_log_block(journal_t *journal, unsigned long long *retp)
journal->j_free--;
if (journal->j_head == journal->j_last)
journal->j_head = journal->j_first;
- spin_unlock(&journal->j_state_lock);
+ write_unlock(&journal->j_state_lock);
return jbd2_journal_bmap(journal, blocknr, retp);
}
@@ -828,7 +839,7 @@ static journal_t * journal_init_common (void)
mutex_init(&journal->j_checkpoint_mutex);
spin_lock_init(&journal->j_revoke_lock);
spin_lock_init(&journal->j_list_lock);
- spin_lock_init(&journal->j_state_lock);
+ rwlock_init(&journal->j_state_lock);
journal->j_commit_interval = (HZ * JBD2_DEFAULT_MAX_COMMIT_AGE);
journal->j_min_batch_time = 0;
@@ -1094,14 +1105,14 @@ void jbd2_journal_update_superblock(journal_t *journal, int wait)
set_buffer_uptodate(bh);
}
- spin_lock(&journal->j_state_lock);
+ read_lock(&journal->j_state_lock);
jbd_debug(1,"JBD: updating superblock (start %ld, seq %d, errno %d)\n",
journal->j_tail, journal->j_tail_sequence, journal->j_errno);
sb->s_sequence = cpu_to_be32(journal->j_tail_sequence);
sb->s_start = cpu_to_be32(journal->j_tail);
sb->s_errno = cpu_to_be32(journal->j_errno);
- spin_unlock(&journal->j_state_lock);
+ read_unlock(&journal->j_state_lock);
BUFFER_TRACE(bh, "marking dirty");
mark_buffer_dirty(bh);
@@ -1115,19 +1126,19 @@ void jbd2_journal_update_superblock(journal_t *journal, int wait)
set_buffer_uptodate(bh);
}
} else
- ll_rw_block(SWRITE, 1, &bh);
+ write_dirty_buffer(bh, WRITE);
out:
/* If we have just flushed the log (by marking s_start==0), then
* any future commit will have to be careful to update the
* superblock again to re-record the true start of the log. */
- spin_lock(&journal->j_state_lock);
+ write_lock(&journal->j_state_lock);
if (sb->s_start)
journal->j_flags &= ~JBD2_FLUSHED;
else
journal->j_flags |= JBD2_FLUSHED;
- spin_unlock(&journal->j_state_lock);
+ write_unlock(&journal->j_state_lock);
}
/*
@@ -1248,6 +1259,13 @@ int jbd2_journal_load(journal_t *journal)
}
}
+ /*
+ * Create a slab for this blocksize
+ */
+ err = jbd2_journal_create_slab(be32_to_cpu(sb->s_blocksize));
+ if (err)
+ return err;
+
/* Let the recovery code check whether it needs to recover any
* data from the journal. */
if (jbd2_journal_recover(journal))
@@ -1355,6 +1373,10 @@ int jbd2_journal_check_used_features (journal_t *journal, unsigned long compat,
if (!compat && !ro && !incompat)
return 1;
+ /* Load journal superblock if it is not loaded yet. */
+ if (journal->j_format_version == 0 &&
+ journal_get_superblock(journal) != 0)
+ return 0;
if (journal->j_format_version == 1)
return 0;
@@ -1382,13 +1404,9 @@ int jbd2_journal_check_used_features (journal_t *journal, unsigned long compat,
int jbd2_journal_check_available_features (journal_t *journal, unsigned long compat,
unsigned long ro, unsigned long incompat)
{
- journal_superblock_t *sb;
-
if (!compat && !ro && !incompat)
return 1;
- sb = journal->j_superblock;
-
/* We can support any known requested features iff the
* superblock is in version 2. Otherwise we fail to support any
* extended sb features. */
@@ -1536,7 +1554,7 @@ int jbd2_journal_flush(journal_t *journal)
transaction_t *transaction = NULL;
unsigned long old_tail;
- spin_lock(&journal->j_state_lock);
+ write_lock(&journal->j_state_lock);
/* Force everything buffered to the log... */
if (journal->j_running_transaction) {
@@ -1549,10 +1567,10 @@ int jbd2_journal_flush(journal_t *journal)
if (transaction) {
tid_t tid = transaction->t_tid;
- spin_unlock(&journal->j_state_lock);
+ write_unlock(&journal->j_state_lock);
jbd2_log_wait_commit(journal, tid);
} else {
- spin_unlock(&journal->j_state_lock);
+ write_unlock(&journal->j_state_lock);
}
/* ...and flush everything in the log out to disk. */
@@ -1576,12 +1594,12 @@ int jbd2_journal_flush(journal_t *journal)
* the magic code for a fully-recovered superblock. Any future
* commits of data to the journal will restore the current
* s_start value. */
- spin_lock(&journal->j_state_lock);
+ write_lock(&journal->j_state_lock);
old_tail = journal->j_tail;
journal->j_tail = 0;
- spin_unlock(&journal->j_state_lock);
+ write_unlock(&journal->j_state_lock);
jbd2_journal_update_superblock(journal, 1);
- spin_lock(&journal->j_state_lock);
+ write_lock(&journal->j_state_lock);
journal->j_tail = old_tail;
J_ASSERT(!journal->j_running_transaction);
@@ -1589,7 +1607,7 @@ int jbd2_journal_flush(journal_t *journal)
J_ASSERT(!journal->j_checkpoint_transactions);
J_ASSERT(journal->j_head == journal->j_tail);
J_ASSERT(journal->j_tail_sequence == journal->j_transaction_sequence);
- spin_unlock(&journal->j_state_lock);
+ write_unlock(&journal->j_state_lock);
return 0;
}
@@ -1608,7 +1626,6 @@ int jbd2_journal_flush(journal_t *journal)
int jbd2_journal_wipe(journal_t *journal, int write)
{
- journal_superblock_t *sb;
int err = 0;
J_ASSERT (!(journal->j_flags & JBD2_LOADED));
@@ -1617,8 +1634,6 @@ int jbd2_journal_wipe(journal_t *journal, int write)
if (err)
return err;
- sb = journal->j_superblock;
-
if (!journal->j_tail)
goto no_recovery;
@@ -1656,12 +1671,12 @@ void __jbd2_journal_abort_hard(journal_t *journal)
printk(KERN_ERR "Aborting journal on device %s.\n",
journal->j_devname);
- spin_lock(&journal->j_state_lock);
+ write_lock(&journal->j_state_lock);
journal->j_flags |= JBD2_ABORT;
transaction = journal->j_running_transaction;
if (transaction)
__jbd2_log_start_commit(journal, transaction->t_tid);
- spin_unlock(&journal->j_state_lock);
+ write_unlock(&journal->j_state_lock);
}
/* Soft abort: record the abort error status in the journal superblock,
@@ -1746,12 +1761,12 @@ int jbd2_journal_errno(journal_t *journal)
{
int err;
- spin_lock(&journal->j_state_lock);
+ read_lock(&journal->j_state_lock);
if (journal->j_flags & JBD2_ABORT)
err = -EROFS;
else
err = journal->j_errno;
- spin_unlock(&journal->j_state_lock);
+ read_unlock(&journal->j_state_lock);
return err;
}
@@ -1766,12 +1781,12 @@ int jbd2_journal_clear_err(journal_t *journal)
{
int err = 0;
- spin_lock(&journal->j_state_lock);
+ write_lock(&journal->j_state_lock);
if (journal->j_flags & JBD2_ABORT)
err = -EROFS;
else
journal->j_errno = 0;
- spin_unlock(&journal->j_state_lock);
+ write_unlock(&journal->j_state_lock);
return err;
}
@@ -1784,10 +1799,10 @@ int jbd2_journal_clear_err(journal_t *journal)
*/
void jbd2_journal_ack_err(journal_t *journal)
{
- spin_lock(&journal->j_state_lock);
+ write_lock(&journal->j_state_lock);
if (journal->j_errno)
journal->j_flags |= JBD2_ACK_ERR;
- spin_unlock(&journal->j_state_lock);
+ write_unlock(&journal->j_state_lock);
}
int jbd2_journal_blocks_per_page(struct inode *inode)
@@ -1807,6 +1822,127 @@ size_t journal_tag_bytes(journal_t *journal)
}
/*
+ * JBD memory management
+ *
+ * These functions are used to allocate block-sized chunks of memory
+ * used for making copies of buffer_head data. Very often it will be
+ * page-sized chunks of data, but sometimes it will be in
+ * sub-page-size chunks. (For example, 16k pages on Power systems
+ * with a 4k block file system.) For blocks smaller than a page, we
+ * use a SLAB allocator. There are slab caches for each block size,
+ * which are allocated at mount time, if necessary, and we only free
+ * (all of) the slab caches when/if the jbd2 module is unloaded. For
+ * this reason we don't need to a mutex to protect access to
+ * jbd2_slab[] allocating or releasing memory; only in
+ * jbd2_journal_create_slab().
+ */
+#define JBD2_MAX_SLABS 8
+static struct kmem_cache *jbd2_slab[JBD2_MAX_SLABS];
+static DECLARE_MUTEX(jbd2_slab_create_sem);
+
+static const char *jbd2_slab_names[JBD2_MAX_SLABS] = {
+ "jbd2_1k", "jbd2_2k", "jbd2_4k", "jbd2_8k",
+ "jbd2_16k", "jbd2_32k", "jbd2_64k", "jbd2_128k"
+};
+
+
+static void jbd2_journal_destroy_slabs(void)
+{
+ int i;
+
+ for (i = 0; i < JBD2_MAX_SLABS; i++) {
+ if (jbd2_slab[i])
+ kmem_cache_destroy(jbd2_slab[i]);
+ jbd2_slab[i] = NULL;
+ }
+}
+
+static int jbd2_journal_create_slab(size_t size)
+{
+ int i = order_base_2(size) - 10;
+ size_t slab_size;
+
+ if (size == PAGE_SIZE)
+ return 0;
+
+ if (i >= JBD2_MAX_SLABS)
+ return -EINVAL;
+
+ if (unlikely(i < 0))
+ i = 0;
+ down(&jbd2_slab_create_sem);
+ if (jbd2_slab[i]) {
+ up(&jbd2_slab_create_sem);
+ return 0; /* Already created */
+ }
+
+ slab_size = 1 << (i+10);
+ jbd2_slab[i] = kmem_cache_create(jbd2_slab_names[i], slab_size,
+ slab_size, 0, NULL);
+ up(&jbd2_slab_create_sem);
+ if (!jbd2_slab[i]) {
+ printk(KERN_EMERG "JBD2: no memory for jbd2_slab cache\n");
+ return -ENOMEM;
+ }
+ return 0;
+}
+
+static struct kmem_cache *get_slab(size_t size)
+{
+ int i = order_base_2(size) - 10;
+
+ BUG_ON(i >= JBD2_MAX_SLABS);
+ if (unlikely(i < 0))
+ i = 0;
+ BUG_ON(jbd2_slab[i] == NULL);
+ return jbd2_slab[i];
+}
+
+void *jbd2_alloc(size_t size, gfp_t flags)
+{
+ void *ptr;
+
+ BUG_ON(size & (size-1)); /* Must be a power of 2 */
+
+ flags |= __GFP_REPEAT;
+ if (size == PAGE_SIZE)
+ ptr = (void *)__get_free_pages(flags, 0);
+ else if (size > PAGE_SIZE) {
+ int order = get_order(size);
+
+ if (order < 3)
+ ptr = (void *)__get_free_pages(flags, order);
+ else
+ ptr = vmalloc(size);
+ } else
+ ptr = kmem_cache_alloc(get_slab(size), flags);
+
+ /* Check alignment; SLUB has gotten this wrong in the past,
+ * and this can lead to user data corruption! */
+ BUG_ON(((unsigned long) ptr) & (size-1));
+
+ return ptr;
+}
+
+void jbd2_free(void *ptr, size_t size)
+{
+ if (size == PAGE_SIZE) {
+ free_pages((unsigned long)ptr, 0);
+ return;
+ }
+ if (size > PAGE_SIZE) {
+ int order = get_order(size);
+
+ if (order < 3)
+ free_pages((unsigned long)ptr, order);
+ else
+ vfree(ptr);
+ return;
+ }
+ kmem_cache_free(get_slab(size), ptr);
+};
+
+/*
* Journal_head storage management
*/
static struct kmem_cache *jbd2_journal_head_cache;
@@ -2071,14 +2207,12 @@ void jbd2_journal_init_jbd_inode(struct jbd2_inode *jinode, struct inode *inode)
void jbd2_journal_release_jbd_inode(journal_t *journal,
struct jbd2_inode *jinode)
{
- int writeout = 0;
-
if (!journal)
return;
restart:
spin_lock(&journal->j_list_lock);
/* Is commit writing out inode - we have to wait */
- if (jinode->i_flags & JI_COMMIT_RUNNING) {
+ if (test_bit(__JI_COMMIT_RUNNING, &jinode->i_flags)) {
wait_queue_head_t *wq;
DEFINE_WAIT_BIT(wait, &jinode->i_flags, __JI_COMMIT_RUNNING);
wq = bit_waitqueue(&jinode->i_flags, __JI_COMMIT_RUNNING);
@@ -2089,9 +2223,6 @@ restart:
goto restart;
}
- /* Do we need to wait for data writeback? */
- if (journal->j_committing_transaction == jinode->i_transaction)
- writeout = 1;
if (jinode->i_transaction) {
list_del(&jinode->i_list);
jinode->i_transaction = NULL;
@@ -2204,6 +2335,7 @@ static void jbd2_journal_destroy_caches(void)
jbd2_journal_destroy_revoke_caches();
jbd2_journal_destroy_jbd2_journal_head_cache();
jbd2_journal_destroy_handle_cache();
+ jbd2_journal_destroy_slabs();
}
static int __init journal_init(void)
generated by cgit v1.2.3-70-g09d2 (git 2.46.0) at 2024-11-24 20:57:55 +0000