1 files changed, 174 insertions, 93 deletions

diff --git a/fs/jbd/checkpoint.c b/fs/jbd/checkpoint.c
index e6265a0b56b..08c03044abd 100644
--- a/fs/jbd/checkpoint.c
+++ b/fs/jbd/checkpoint.c
@@ -1,6 +1,6 @@
 /*
- * linux/fs/checkpoint.c
- * 
+ * linux/fs/jbd/checkpoint.c
+ *
  * Written by Stephen C. Tweedie <sct@redhat.com>, 1999
  *
  * Copyright 1999 Red Hat Software --- All Rights Reserved
@@ -9,8 +9,8 @@
  * the terms of the GNU General Public License, version 2, or at your
  * option, any later version, incorporated herein by reference.
  *
- * Checkpoint routines for the generic filesystem journaling code.  
- * Part of the ext2fs journaling system.  
+ * Checkpoint routines for the generic filesystem journaling code.
+ * Part of the ext2fs journaling system.
  *
  * Checkpointing is the process of ensuring that a section of the log is
  * committed fully to disk, so that that portion of the log can be
@@ -22,18 +22,17 @@
 #include <linux/jbd.h>
 #include <linux/errno.h>
 #include <linux/slab.h>
+#include <linux/blkdev.h>
+#include <trace/events/jbd.h>
 
 /*
  * Unlink a buffer from a transaction checkpoint list.
  *
  * Called with j_list_lock held.
  */
-
-static void __buffer_unlink_first(struct journal_head *jh)
+static inline void __buffer_unlink_first(struct journal_head *jh)
 {
-	transaction_t *transaction;
-
-	transaction = jh->b_cp_transaction;
+	transaction_t *transaction = jh->b_cp_transaction;
 
 	jh->b_cpnext->b_cpprev = jh->b_cpprev;
 	jh->b_cpprev->b_cpnext = jh->b_cpnext;
@@ -49,12 +48,9 @@ static void __buffer_unlink_first(struct journal_head *jh)
  *
  * Called with j_list_lock held.
  */
-
 static inline void __buffer_unlink(struct journal_head *jh)
 {
-	transaction_t *transaction;
-
-	transaction = jh->b_cp_transaction;
+	transaction_t *transaction = jh->b_cp_transaction;
 
 	__buffer_unlink_first(jh);
 	if (transaction->t_checkpoint_io_list == jh) {
@@ -69,12 +65,10 @@ static inline void __buffer_unlink(struct journal_head *jh)
  *
  * Called with j_list_lock held
  */
-
 static inline void __buffer_relink_io(struct journal_head *jh)
 {
-	transaction_t *transaction;
+	transaction_t *transaction = jh->b_cp_transaction;
 
-	transaction = jh->b_cp_transaction;
 	__buffer_unlink_first(jh);
 
 	if (!transaction->t_checkpoint_io_list) {
@@ -101,11 +95,16 @@ static int __try_to_free_cp_buf(struct journal_head *jh)
 	int ret = 0;
 	struct buffer_head *bh = jh2bh(jh);
 
-	if (jh->b_jlist == BJ_None && !buffer_locked(bh) && !buffer_dirty(bh)) {
+	if (jh->b_jlist == BJ_None && !buffer_locked(bh) &&
+	    !buffer_dirty(bh) && !buffer_write_io_error(bh)) {
+		/*
+		 * Get our reference so that bh cannot be freed before
+		 * we unlock it
+		 */
+		get_bh(bh);
 		JBUFFER_TRACE(jh, "remove from checkpoint list");
 		ret = __journal_remove_checkpoint(jh) + 1;
 		jbd_unlock_bh_state(bh);
-		journal_remove_journal_head(bh);
 		BUFFER_TRACE(bh, "release");
 		__brelse(bh);
 	} else {
@@ -122,7 +121,7 @@ static int __try_to_free_cp_buf(struct journal_head *jh)
  */
 void __log_wait_for_space(journal_t *journal)
 {
-	int nblocks;
+	int nblocks, space_left;
 	assert_spin_locked(&journal->j_state_lock);
 
 	nblocks = jbd_space_needed(journal);
@@ -130,20 +129,52 @@ void __log_wait_for_space(journal_t *journal)
 		if (journal->j_flags & JFS_ABORT)
 			return;
 		spin_unlock(&journal->j_state_lock);
-		down(&journal->j_checkpoint_sem);
+		mutex_lock(&journal->j_checkpoint_mutex);
 
 		/*
 		 * Test again, another process may have checkpointed while we
-		 * were waiting for the checkpoint lock
+		 * were waiting for the checkpoint lock. If there are no
+		 * transactions ready to be checkpointed, try to recover
+		 * journal space by calling cleanup_journal_tail(), and if
+		 * that doesn't work, by waiting for the currently committing
+		 * transaction to complete.  If there is absolutely no way
+		 * to make progress, this is either a BUG or corrupted
+		 * filesystem, so abort the journal and leave a stack
+		 * trace for forensic evidence.
 		 */
 		spin_lock(&journal->j_state_lock);
+		spin_lock(&journal->j_list_lock);
 		nblocks = jbd_space_needed(journal);
-		if (__log_space_left(journal) < nblocks) {
+		space_left = __log_space_left(journal);
+		if (space_left < nblocks) {
+			int chkpt = journal->j_checkpoint_transactions != NULL;
+			tid_t tid = 0;
+
+			if (journal->j_committing_transaction)
+				tid = journal->j_committing_transaction->t_tid;
+			spin_unlock(&journal->j_list_lock);
 			spin_unlock(&journal->j_state_lock);
-			log_do_checkpoint(journal);
+			if (chkpt) {
+				log_do_checkpoint(journal);
+			} else if (cleanup_journal_tail(journal) == 0) {
+				/* We were able to recover space; yay! */
+				;
+			} else if (tid) {
+				log_wait_commit(journal, tid);
+			} else {
+				printk(KERN_ERR "%s: needed %d blocks and "
+				       "only had %d space available\n",
+				       __func__, nblocks, space_left);
+				printk(KERN_ERR "%s: no way to get more "
+				       "journal space\n", __func__);
+				WARN_ON(1);
+				journal_abort(journal, 0);
+			}
 			spin_lock(&journal->j_state_lock);
+		} else {
+			spin_unlock(&journal->j_list_lock);
 		}
-		up(&journal->j_checkpoint_sem);
+		mutex_unlock(&journal->j_checkpoint_mutex);
 	}
 }
 
@@ -153,6 +184,7 @@ void __log_wait_for_space(journal_t *journal)
  * jbd_unlock_bh_state().
  */
 static void jbd_sync_bh(journal_t *journal, struct buffer_head *bh)
+	__releases(journal->j_list_lock)
 {
 	get_bh(bh);
 	spin_unlock(&journal->j_list_lock);
@@ -167,22 +199,25 @@ static void jbd_sync_bh(journal_t *journal, struct buffer_head *bh)
  * buffers. Note that we take the buffers in the opposite ordering
  * from the one in which they were submitted for IO.
  *
+ * Return 0 on success, and return <0 if some buffers have failed
+ * to be written out.
+ *
  * Called with j_list_lock held.
  */
-
-static void __wait_cp_io(journal_t *journal, transaction_t *transaction)
+static int __wait_cp_io(journal_t *journal, transaction_t *transaction)
 {
 	struct journal_head *jh;
 	struct buffer_head *bh;
 	tid_t this_tid;
 	int released = 0;
+	int ret = 0;
 
 	this_tid = transaction->t_tid;
 restart:
-	/* Didn't somebody clean up the transaction in the meanwhile */
+	/* Did somebody clean up the transaction in the meanwhile? */
 	if (journal->j_checkpoint_transactions != transaction ||
-		transaction->t_tid != this_tid)
-		return;
+			transaction->t_tid != this_tid)
+		return ret;
 	while (!released && transaction->t_checkpoint_io_list) {
 		jh = transaction->t_checkpoint_io_list;
 		bh = jh2bh(jh);
@@ -191,8 +226,8 @@ restart:
 			spin_lock(&journal->j_list_lock);
 			goto restart;
 		}
+		get_bh(bh);
 		if (buffer_locked(bh)) {
-			atomic_inc(&bh->b_count);
 			spin_unlock(&journal->j_list_lock);
 			jbd_unlock_bh_state(bh);
 			wait_on_buffer(bh);
@@ -202,13 +237,19 @@ restart:
 			spin_lock(&journal->j_list_lock);
 			goto restart;
 		}
+		if (unlikely(buffer_write_io_error(bh)))
+			ret = -EIO;
+
 		/*
 		 * Now in whatever state the buffer currently is, we know that
 		 * it has been written out and so we can drop it from the list
 		 */
 		released = __journal_remove_checkpoint(jh);
 		jbd_unlock_bh_state(bh);
+		__brelse(bh);
 	}
+
+	return ret;
 }
 
 #define NR_BATCH	64
@@ -217,8 +258,13 @@ static void
 __flush_batch(journal_t *journal, struct buffer_head **bhs, int *batch_count)
 {
 	int i;
+	struct blk_plug plug;
+
+	blk_start_plug(&plug);
+	for (i = 0; i < *batch_count; i++)
+		write_dirty_buffer(bhs[i], WRITE_SYNC);
+	blk_finish_plug(&plug);
 
-	ll_rw_block(SWRITE, *batch_count, bhs);
 	for (i = 0; i < *batch_count; i++) {
 		struct buffer_head *bh = bhs[i];
 		clear_buffer_jwrite(bh);
@@ -232,7 +278,8 @@ __flush_batch(journal_t *journal, struct buffer_head **bhs, int *batch_count)
  * Try to flush one buffer from the checkpoint list to disk.
  *
  * Return 1 if something happened which requires us to abort the current
- * scan of the checkpoint list.  
+ * scan of the checkpoint list.  Return <0 if the buffer has failed to
+ * be written out.
  *
  * Called with j_list_lock held and drops it if 1 is returned
  * Called under jbd_lock_bh_state(jh2bh(jh)), and drops it
@@ -250,10 +297,9 @@ static int __process_buffer(journal_t *journal, struct journal_head *jh,
 		wait_on_buffer(bh);
 		/* the journal_head may have gone by now */
 		BUFFER_TRACE(bh, "brelse");
-		put_bh(bh);
+		__brelse(bh);
 		ret = 1;
-	}
-	else if (jh->b_transaction != NULL) {
+	} else if (jh->b_transaction != NULL) {
 		transaction_t *t = jh->b_transaction;
 		tid_t tid = t->t_tid;
 
@@ -262,24 +308,24 @@ static int __process_buffer(journal_t *journal, struct journal_head *jh,
 		log_start_commit(journal, tid);
 		log_wait_commit(journal, tid);
 		ret = 1;
-	}
-	else if (!buffer_dirty(bh)) {
+	} else if (!buffer_dirty(bh)) {
+		ret = 1;
+		if (unlikely(buffer_write_io_error(bh)))
+			ret = -EIO;
+		get_bh(bh);
 		J_ASSERT_JH(jh, !buffer_jbddirty(bh));
 		BUFFER_TRACE(bh, "remove from checkpoint");
 		__journal_remove_checkpoint(jh);
 		spin_unlock(&journal->j_list_lock);
 		jbd_unlock_bh_state(bh);
-		journal_remove_journal_head(bh);
-		put_bh(bh);
-		ret = 1;
-	}
-	else {
+		__brelse(bh);
+	} else {
 		/*
 		 * Important: we are about to write the buffer, and
 		 * possibly block, while still holding the journal lock.
 		 * We cannot afford to let the transaction logic start
 		 * messing around with this buffer before we write it to
-		 * disk, as that would break recoverability.  
+		 * disk, as that would break recoverability.
 		 */
 		BUFFER_TRACE(bh, "queue");
 		get_bh(bh);
@@ -302,8 +348,9 @@ static int __process_buffer(journal_t *journal, struct journal_head *jh,
  * Perform an actual checkpoint. We take the first transaction on the
  * list of transactions to be checkpointed and send all its buffers
  * to disk. We submit larger chunks of data at once.
- * 
+ *
  * The journal should be locked before calling this function.
+ * Called with j_checkpoint_mutex held.
  */
 int log_do_checkpoint(journal_t *journal)
 {
@@ -313,12 +360,13 @@ int log_do_checkpoint(journal_t *journal)
 
 	jbd_debug(1, "Start checkpoint\n");
 
-	/* 
+	/*
 	 * First thing: if there are any transactions in the log which
 	 * don't need checkpointing, just eliminate them from the
-	 * journal straight away.  
+	 * journal straight away.
 	 */
 	result = cleanup_journal_tail(journal);
+	trace_jbd_checkpoint(journal, result);
 	jbd_debug(1, "cleanup_journal_tail returned %d\n", result);
 	if (result <= 0)
 		return result;
@@ -327,6 +375,7 @@ int log_do_checkpoint(journal_t *journal)
 	 * OK, we need to start writing disk blocks.  Take one transaction
 	 * and write it.
 	 */
+	result = 0;
 	spin_lock(&journal->j_list_lock);
 	if (!journal->j_checkpoint_transactions)
 		goto out;
@@ -338,12 +387,12 @@ restart:
 	 * done (maybe it's a new transaction, but it fell at the same
 	 * address).
 	 */
- 	if (journal->j_checkpoint_transactions == transaction &&
+	if (journal->j_checkpoint_transactions == transaction &&
 			transaction->t_tid == this_tid) {
 		int batch_count = 0;
 		struct buffer_head *bhs[NR_BATCH];
 		struct journal_head *jh;
-		int retry = 0;
+		int retry = 0, err;
 
 		while (!retry && transaction->t_checkpoint_list) {
 			struct buffer_head *bh;
@@ -355,10 +404,11 @@ restart:
 				retry = 1;
 				break;
 			}
-			retry = __process_buffer(journal, jh, bhs,
-						&batch_count);
-			if (!retry &&
-			    lock_need_resched(&journal->j_list_lock)) {
+			retry = __process_buffer(journal, jh, bhs,&batch_count);
+			if (retry < 0 && !result)
+				result = retry;
+			if (!retry && (need_resched() ||
+				spin_needbreak(&journal->j_list_lock))) {
 				spin_unlock(&journal->j_list_lock);
 				retry = 1;
 				break;
@@ -379,16 +429,20 @@ restart:
 		}
 		/*
 		 * Now we have cleaned up the first transaction's checkpoint
-		 * list.  Let's clean up the second one.
+		 * list. Let's clean up the second one
 		 */
-		__wait_cp_io(journal, transaction);
+		err = __wait_cp_io(journal, transaction);
+		if (!result)
+			result = err;
 	}
 out:
 	spin_unlock(&journal->j_list_lock);
-	result = cleanup_journal_tail(journal);
 	if (result < 0)
-		return result;
-	return 0;
+		journal_abort(journal, result);
+	else
+		result = cleanup_journal_tail(journal);
+
+	return (result < 0) ? result : 0;
 }
 
 /*
@@ -396,31 +450,34 @@ out:
  * we have already got rid of any since the last update of the log tail
  * in the journal superblock.  If so, we can instantly roll the
  * superblock forward to remove those transactions from the log.
- * 
+ *
  * Return <0 on error, 0 on success, 1 if there was nothing to clean up.
- * 
- * Called with the journal lock held.
  *
  * This is the only part of the journaling code which really needs to be
  * aware of transaction aborts.  Checkpointing involves writing to the
  * main filesystem area rather than to the journal, so it can proceed
- * even in abort state, but we must not update the journal superblock if
- * we have an abort error outstanding.
+ * even in abort state, but we must not update the super block if
+ * checkpointing may have failed.  Otherwise, we would lose some metadata
+ * buffers which should be written-back to the filesystem.
  */
 
 int cleanup_journal_tail(journal_t *journal)
 {
 	transaction_t * transaction;
 	tid_t		first_tid;
-	unsigned long	blocknr, freed;
+	unsigned int	blocknr, freed;
 
-	/* OK, work out the oldest transaction remaining in the log, and
-	 * the log block it starts at. 
-	 * 
+	if (is_journal_aborted(journal))
+		return 1;
+
+	/*
+	 * OK, work out the oldest transaction remaining in the log, and
+	 * the log block it starts at.
+	 *
 	 * If the log is now empty, we need to work out which is the
 	 * next transaction ID we will write, and where it will
-	 * start. */
-
+	 * start.
+	 */
 	spin_lock(&journal->j_state_lock);
 	spin_lock(&journal->j_list_lock);
 	transaction = journal->j_checkpoint_transactions;
@@ -446,7 +503,24 @@ int cleanup_journal_tail(journal_t *journal)
 		spin_unlock(&journal->j_state_lock);
 		return 1;
 	}
+	spin_unlock(&journal->j_state_lock);
 
+	/*
+	 * We need to make sure that any blocks that were recently written out
+	 * --- perhaps by log_do_checkpoint() --- are flushed out before we
+	 * drop the transactions from the journal. Similarly we need to be sure
+	 * superblock makes it to disk before next transaction starts reusing
+	 * freed space (otherwise we could replay some blocks of the new
+	 * transaction thinking they belong to the old one). So we use
+	 * WRITE_FLUSH_FUA. It's unlikely this will be necessary, especially
+	 * with an appropriately sized journal, but we need this to guarantee
+	 * correctness.  Fortunately cleanup_journal_tail() doesn't get called
+	 * all that often.
+	 */
+	journal_update_sb_log_tail(journal, first_tid, blocknr,
+				   WRITE_FLUSH_FUA);
+
+	spin_lock(&journal->j_state_lock);
 	/* OK, update the superblock to recover the freed space.
 	 * Physical blocks come first: have we wrapped beyond the end of
 	 * the log?  */
@@ -454,17 +528,16 @@ int cleanup_journal_tail(journal_t *journal)
 	if (blocknr < journal->j_tail)
 		freed = freed + journal->j_last - journal->j_first;
 
+	trace_jbd_cleanup_journal_tail(journal, first_tid, blocknr, freed);
 	jbd_debug(1,
-		  "Cleaning journal tail from %d to %d (offset %lu), "
-		  "freeing %lu\n",
+		  "Cleaning journal tail from %d to %d (offset %u), "
+		  "freeing %u\n",
 		  journal->j_tail_sequence, first_tid, blocknr, freed);
 
 	journal->j_free += freed;
 	journal->j_tail_sequence = first_tid;
 	journal->j_tail = blocknr;
 	spin_unlock(&journal->j_state_lock);
-	if (!(journal->j_flags & JFS_ABORT))
-		journal_update_superblock(journal, 1);
 	return 0;
 }
 
@@ -474,11 +547,11 @@ int cleanup_journal_tail(journal_t *journal)
 /*
  * journal_clean_one_cp_list
  *
- * Find all the written-back checkpoint buffers in the given list and release them.
+ * Find all the written-back checkpoint buffers in the given list and release
+ * them.
  *
- * Called with the journal locked.
  * Called with j_list_lock held.
- * Returns number of bufers reaped (for debug)
+ * Returns number of buffers reaped (for debug)
  */
 
 static int journal_clean_one_cp_list(struct journal_head *jh, int *released)
@@ -491,7 +564,7 @@ static int journal_clean_one_cp_list(struct journal_head *jh, int *released)
 	if (!jh)
 		return 0;
 
- 	last_jh = jh->b_cpprev;
+	last_jh = jh->b_cpprev;
 	do {
 		jh = next_jh;
 		next_jh = jh->b_cpnext;
@@ -507,9 +580,10 @@ static int journal_clean_one_cp_list(struct journal_head *jh, int *released)
 			}
 		}
 		/*
-		 * This function only frees up some memory if possible so we
-		 * dont have an obligation to finish processing. Bail out if
-		 * preemption requested:
+		 * This function only frees up some memory
+		 * if possible so we dont have an obligation
+		 * to finish processing. Bail out if preemption
+		 * requested:
 		 */
 		if (need_resched())
 			return freed;
@@ -531,7 +605,8 @@ static int journal_clean_one_cp_list(struct journal_head *jh, int *released)
 int __journal_clean_checkpoint_list(journal_t *journal)
 {
 	transaction_t *transaction, *last_transaction, *next_transaction;
-	int ret = 0, released;
+	int ret = 0;
+	int released;
 
 	transaction = journal->j_checkpoint_transactions;
 	if (!transaction)
@@ -544,6 +619,11 @@ int __journal_clean_checkpoint_list(journal_t *journal)
 		next_transaction = transaction->t_cpnext;
 		ret += journal_clean_one_cp_list(transaction->
 				t_checkpoint_list, &released);
+		/*
+		 * This function only frees up some memory if possible so we
+		 * dont have an obligation to finish processing. Bail out if
+		 * preemption requested:
+		 */
 		if (need_resched())
 			goto out;
 		if (released)
@@ -562,7 +642,7 @@ out:
 	return ret;
 }
 
-/* 
+/*
  * journal_remove_checkpoint: called after a buffer has been committed
  * to disk (either by being write-back flushed to disk, or being
  * committed to the log).
@@ -576,8 +656,8 @@ out:
  * checkpoint lists.
  *
  * The function returns 1 if it frees the transaction, 0 otherwise.
+ * The function can free jh and bh.
  *
- * This function is called with the journal locked.
  * This function is called with j_list_lock held.
  * This function is called with jbd_lock_bh_state(jh2bh(jh))
  */
@@ -596,27 +676,26 @@ int __journal_remove_checkpoint(struct journal_head *jh)
 	}
 	journal = transaction->t_journal;
 
+	JBUFFER_TRACE(jh, "removing from transaction");
 	__buffer_unlink(jh);
 	jh->b_cp_transaction = NULL;
+	journal_put_journal_head(jh);
 
 	if (transaction->t_checkpoint_list != NULL ||
 	    transaction->t_checkpoint_io_list != NULL)
 		goto out;
-	JBUFFER_TRACE(jh, "transaction has no more buffers");
 
 	/*
 	 * There is one special case to worry about: if we have just pulled the
-	 * buffer off a committing transaction's forget list, then even if the
-	 * checkpoint list is empty, the transaction obviously cannot be
-	 * dropped!
+	 * buffer off a running or committing transaction's checkpoing list,
+	 * then even if the checkpoint list is empty, the transaction obviously
+	 * cannot be dropped!
 	 *
-	 * The locking here around j_committing_transaction is a bit sleazy.
+	 * The locking here around t_state is a bit sleazy.
 	 * See the comment at the end of journal_commit_transaction().
 	 */
-	if (transaction == journal->j_committing_transaction) {
-		JBUFFER_TRACE(jh, "belongs to committing transaction");
+	if (transaction->t_state != T_FINISHED)
 		goto out;
-	}
 
 	/* OK, that was the last buffer for the transaction: we can now
 	   safely remove this transaction from the log */
@@ -628,7 +707,6 @@ int __journal_remove_checkpoint(struct journal_head *jh)
 	wake_up(&journal->j_wait_logspace);
 	ret = 1;
 out:
-	JBUFFER_TRACE(jh, "exit");
 	return ret;
 }
 
@@ -640,13 +718,15 @@ out:
  * Called with the journal locked.
  * Called with j_list_lock held.
  */
-void __journal_insert_checkpoint(struct journal_head *jh, 
+void __journal_insert_checkpoint(struct journal_head *jh,
 			       transaction_t *transaction)
 {
 	JBUFFER_TRACE(jh, "entry");
 	J_ASSERT_JH(jh, buffer_dirty(jh2bh(jh)) || buffer_jbddirty(jh2bh(jh)));
 	J_ASSERT_JH(jh, jh->b_cp_transaction == NULL);
 
+	/* Get reference for checkpointing transaction */
+	journal_grab_journal_head(jh2bh(jh));
 	jh->b_cp_transaction = transaction;
 
 	if (!transaction->t_checkpoint_list) {
@@ -662,7 +742,7 @@ void __journal_insert_checkpoint(struct journal_head *jh,
 
 /*
  * We've finished with this transaction structure: adios...
- * 
+ *
  * The transaction must have no links except for the checkpoint by this
  * point.
  *
@@ -696,6 +776,7 @@ void __journal_drop_transaction(journal_t *journal, transaction_t *transaction)
 	J_ASSERT(journal->j_committing_transaction != transaction);
 	J_ASSERT(journal->j_running_transaction != transaction);
 
+	trace_jbd_drop_transaction(journal, transaction);
 	jbd_debug(1, "Dropping transaction %d, all done\n", transaction->t_tid);
 	kfree(transaction);
 }