5 files changed, 68 insertions, 54 deletions

diff --git a/fs/jbd/checkpoint.c b/fs/jbd/checkpoint.c
index 5a97e34..014a51f 100644
--- a/fs/jbd/checkpoint.c
+++ b/fs/jbd/checkpoint.c
@@ -204,7 +204,7 @@ __flush_batch(journal_t *journal, struct buffer_head **bhs, int *batch_count)
 	int i;
 
 	spin_unlock(&journal->j_list_lock);
-	ll_rw_block(WRITE, *batch_count, bhs);
+	ll_rw_block(SWRITE, *batch_count, bhs);
 	spin_lock(&journal->j_list_lock);
 	for (i = 0; i < *batch_count; i++) {
 		struct buffer_head *bh = bhs[i];
diff --git a/fs/jbd/commit.c b/fs/jbd/commit.c
index dac720c..2a3e310 100644
--- a/fs/jbd/commit.c
+++ b/fs/jbd/commit.c
@@ -358,7 +358,7 @@ write_out_data:
 					jbd_debug(2, "submit %d writes\n",
 							bufs);
 					spin_unlock(&journal->j_list_lock);
-					ll_rw_block(WRITE, bufs, wbuf);
+					ll_rw_block(SWRITE, bufs, wbuf);
 					journal_brelse_array(wbuf, bufs);
 					bufs = 0;
 					goto write_out_data;
@@ -381,7 +381,7 @@ write_out_data:
 
 	if (bufs) {
 		spin_unlock(&journal->j_list_lock);
-		ll_rw_block(WRITE, bufs, wbuf);
+		ll_rw_block(SWRITE, bufs, wbuf);
 		journal_brelse_array(wbuf, bufs);
 		spin_lock(&journal->j_list_lock);
 	}
@@ -720,11 +720,17 @@ wait_for_iobuf:
 	J_ASSERT(commit_transaction->t_log_list == NULL);
 
 restart_loop:
+	/*
+	 * As there are other places (journal_unmap_buffer()) adding buffers
+	 * to this list we have to be careful and hold the j_list_lock.
+	 */
+	spin_lock(&journal->j_list_lock);
 	while (commit_transaction->t_forget) {
 		transaction_t *cp_transaction;
 		struct buffer_head *bh;
 
 		jh = commit_transaction->t_forget;
+		spin_unlock(&journal->j_list_lock);
 		bh = jh2bh(jh);
 		jbd_lock_bh_state(bh);
 		J_ASSERT_JH(jh,	jh->b_transaction == commit_transaction ||
@@ -792,9 +798,25 @@ restart_loop:
 			journal_remove_journal_head(bh);  /* needs a brelse */
 			release_buffer_page(bh);
 		}
+		cond_resched_lock(&journal->j_list_lock);
+	}
+	spin_unlock(&journal->j_list_lock);
+	/*
+	 * This is a bit sleazy.  We borrow j_list_lock to protect
+	 * journal->j_committing_transaction in __journal_remove_checkpoint.
+	 * Really, __journal_remove_checkpoint should be using j_state_lock but
+	 * it's a bit hassle to hold that across __journal_remove_checkpoint
+	 */
+	spin_lock(&journal->j_state_lock);
+	spin_lock(&journal->j_list_lock);
+	/*
+	 * Now recheck if some buffers did not get attached to the transaction
+	 * while the lock was dropped...
+	 */
+	if (commit_transaction->t_forget) {
 		spin_unlock(&journal->j_list_lock);
-		if (cond_resched())
-			goto restart_loop;
+		spin_unlock(&journal->j_state_lock);
+		goto restart_loop;
 	}
 
 	/* Done with this transaction! */
@@ -803,14 +825,6 @@ restart_loop:
 
 	J_ASSERT(commit_transaction->t_state == T_COMMIT);
 
-	/*
-	 * This is a bit sleazy.  We borrow j_list_lock to protect
-	 * journal->j_committing_transaction in __journal_remove_checkpoint.
-	 * Really, __jornal_remove_checkpoint should be using j_state_lock but
-	 * it's a bit hassle to hold that across __journal_remove_checkpoint
-	 */
-	spin_lock(&journal->j_state_lock);
-	spin_lock(&journal->j_list_lock);
 	commit_transaction->t_state = T_FINISHED;
 	J_ASSERT(commit_transaction == journal->j_committing_transaction);
 	journal->j_commit_sequence = commit_transaction->t_tid;
diff --git a/fs/jbd/journal.c b/fs/jbd/journal.c
index 5e7b439..7ae2c4f 100644
--- a/fs/jbd/journal.c
+++ b/fs/jbd/journal.c
@@ -65,7 +65,6 @@ EXPORT_SYMBOL(journal_set_features);
 EXPORT_SYMBOL(journal_create);
 EXPORT_SYMBOL(journal_load);
 EXPORT_SYMBOL(journal_destroy);
-EXPORT_SYMBOL(journal_recover);
 EXPORT_SYMBOL(journal_update_superblock);
 EXPORT_SYMBOL(journal_abort);
 EXPORT_SYMBOL(journal_errno);
@@ -81,6 +80,7 @@ EXPORT_SYMBOL(journal_try_to_free_buffers);
 EXPORT_SYMBOL(journal_force_commit);
 
 static int journal_convert_superblock_v1(journal_t *, journal_superblock_t *);
+static void __journal_abort_soft (journal_t *journal, int errno);
 
 /*
  * Helper function used to manage commit timeouts
@@ -93,16 +93,6 @@ static void commit_timeout(unsigned long __data)
 	wake_up_process(p);
 }
 
-/* Static check for data structure consistency.  There's no code
- * invoked --- we'll just get a linker failure if things aren't right.
- */
-void __journal_internal_check(void)
-{
-	extern void journal_bad_superblock_size(void);
-	if (sizeof(struct journal_superblock_s) != 1024)
-		journal_bad_superblock_size();
-}
-
 /*
  * kjournald: The main thread function used to manage a logging device
  * journal.
@@ -119,16 +109,12 @@ void __journal_internal_check(void)
  *    known as checkpointing, and this thread is responsible for that job.
  */
 
-journal_t *current_journal;		// AKPM: debug
-
-int kjournald(void *arg)
+static int kjournald(void *arg)
 {
 	journal_t *journal = (journal_t *) arg;
 	transaction_t *transaction;
 	struct timer_list timer;
 
-	current_journal = journal;
-
 	daemonize("kjournald");
 
 	/* Set up an interval timer which can be used to trigger a
@@ -193,6 +179,8 @@ loop:
 		if (transaction && time_after_eq(jiffies,
 						transaction->t_expires))
 			should_sleep = 0;
+		if (journal->j_flags & JFS_UNMOUNT)
+ 			should_sleep = 0;
 		if (should_sleep) {
 			spin_unlock(&journal->j_state_lock);
 			schedule();
@@ -969,7 +957,7 @@ void journal_update_superblock(journal_t *journal, int wait)
 	if (wait)
 		sync_dirty_buffer(bh);
 	else
-		ll_rw_block(WRITE, 1, &bh);
+		ll_rw_block(SWRITE, 1, &bh);
 
 out:
 	/* If we have just flushed the log (by marking s_start==0), then
@@ -1439,7 +1427,7 @@ int journal_wipe(journal_t *journal, int write)
  * device this journal is present.
  */
 
-const char *journal_dev_name(journal_t *journal, char *buffer)
+static const char *journal_dev_name(journal_t *journal, char *buffer)
 {
 	struct block_device *bdev;
 
@@ -1485,7 +1473,7 @@ void __journal_abort_hard(journal_t *journal)
 
 /* Soft abort: record the abort error status in the journal superblock,
  * but don't do any other IO. */
-void __journal_abort_soft (journal_t *journal, int errno)
+static void __journal_abort_soft (journal_t *journal, int errno)
 {
 	if (journal->j_flags & JFS_ABORT)
 		return;
@@ -1880,7 +1868,7 @@ EXPORT_SYMBOL(journal_enable_debug);
 
 static struct proc_dir_entry *proc_jbd_debug;
 
-int read_jbd_debug(char *page, char **start, off_t off,
+static int read_jbd_debug(char *page, char **start, off_t off,
 			  int count, int *eof, void *data)
 {
 	int ret;
@@ -1890,7 +1878,7 @@ int read_jbd_debug(char *page, char **start, off_t off,
 	return ret;
 }
 
-int write_jbd_debug(struct file *file, const char __user *buffer,
+static int write_jbd_debug(struct file *file, const char __user *buffer,
 			   unsigned long count, void *data)
 {
 	char buf[32];
@@ -1979,6 +1967,14 @@ static int __init journal_init(void)
 {
 	int ret;
 
+/* Static check for data structure consistency.  There's no code
+ * invoked --- we'll just get a linker failure if things aren't right.
+ */
+	extern void journal_bad_superblock_size(void);
+	if (sizeof(struct journal_superblock_s) != 1024)
+		journal_bad_superblock_size();
+
+
 	ret = journal_init_caches();
 	if (ret != 0)
 		journal_destroy_caches();
diff --git a/fs/jbd/revoke.c b/fs/jbd/revoke.c
index d327a59..a561441 100644
--- a/fs/jbd/revoke.c
+++ b/fs/jbd/revoke.c
@@ -116,7 +116,8 @@ static inline int hash(journal_t *journal, unsigned long block)
 		(block << (hash_shift - 12))) & (table->hash_size - 1);
 }
 
-int insert_revoke_hash(journal_t *journal, unsigned long blocknr, tid_t seq)
+static int insert_revoke_hash(journal_t *journal, unsigned long blocknr,
+			      tid_t seq)
 {
 	struct list_head *hash_list;
 	struct jbd_revoke_record_s *record;
@@ -613,7 +614,7 @@ static void flush_descriptor(journal_t *journal,
 	set_buffer_jwrite(bh);
 	BUFFER_TRACE(bh, "write");
 	set_buffer_dirty(bh);
-	ll_rw_block(WRITE, 1, &bh);
+	ll_rw_block(SWRITE, 1, &bh);
 }
 #endif
 
diff --git a/fs/jbd/transaction.c b/fs/jbd/transaction.c
index 77b7662..c6ec66f 100644
--- a/fs/jbd/transaction.c
+++ b/fs/jbd/transaction.c
@@ -490,23 +490,21 @@ void journal_unlock_updates (journal_t *journal)
  */
 static void jbd_unexpected_dirty_buffer(struct journal_head *jh)
 {
-	struct buffer_head *bh = jh2bh(jh);
 	int jlist;
 
-	if (buffer_dirty(bh)) {
-		/* If this buffer is one which might reasonably be dirty
-		 * --- ie. data, or not part of this journal --- then
-		 * we're OK to leave it alone, but otherwise we need to
-		 * move the dirty bit to the journal's own internal
-		 * JBDDirty bit. */
-		jlist = jh->b_jlist;
-
-		if (jlist == BJ_Metadata || jlist == BJ_Reserved || 
-		    jlist == BJ_Shadow || jlist == BJ_Forget) {
-			if (test_clear_buffer_dirty(jh2bh(jh))) {
-				set_bit(BH_JBDDirty, &jh2bh(jh)->b_state);
-			}
-		}
+	/* If this buffer is one which might reasonably be dirty
+	 * --- ie. data, or not part of this journal --- then
+	 * we're OK to leave it alone, but otherwise we need to
+	 * move the dirty bit to the journal's own internal
+	 * JBDDirty bit. */
+	jlist = jh->b_jlist;
+
+	if (jlist == BJ_Metadata || jlist == BJ_Reserved ||
+	    jlist == BJ_Shadow || jlist == BJ_Forget) {
+		struct buffer_head *bh = jh2bh(jh);
+
+		if (test_clear_buffer_dirty(bh))
+			set_buffer_jbddirty(bh);
 	}
 }
 
@@ -574,9 +572,14 @@ repeat:
 			if (jh->b_next_transaction)
 				J_ASSERT_JH(jh, jh->b_next_transaction ==
 							transaction);
-			JBUFFER_TRACE(jh, "Unexpected dirty buffer");
-			jbd_unexpected_dirty_buffer(jh);
- 		}
+		}
+		/*
+		 * In any case we need to clean the dirty flag and we must
+		 * do it under the buffer lock to be sure we don't race
+		 * with running write-out.
+		 */
+		JBUFFER_TRACE(jh, "Unexpected dirty buffer");
+		jbd_unexpected_dirty_buffer(jh);
  	}
 
 	unlock_buffer(bh);