Merge branch 'x86/core' into x86/xsave

4 files changed, 181 insertions, 111 deletions

diff --git a/fs/jbd/commit.c b/fs/jbd/commit.c
index 5a8ca61498c..2eccbfaa1d4 100644
--- a/fs/jbd/commit.c
+++ b/fs/jbd/commit.c
@@ -36,7 +36,7 @@ static void journal_end_buffer_io_sync(struct buffer_head *bh, int uptodate)
 
 /*
  * When an ext3-ordered file is truncated, it is possible that many pages are
- * not sucessfully freed, because they are attached to a committing transaction.
+ * not successfully freed, because they are attached to a committing transaction.
  * After the transaction commits, these pages are left on the LRU, with no
  * ->mapping, and with attached buffers.  These pages are trivially reclaimable
  * by the VM, but their apparent absence upsets the VM accounting, and it makes
@@ -45,8 +45,8 @@ static void journal_end_buffer_io_sync(struct buffer_head *bh, int uptodate)
  * So here, we have a buffer which has just come off the forget list.  Look to
  * see if we can strip all buffers from the backing page.
  *
- * Called under lock_journal(), and possibly under journal_datalist_lock.  The
- * caller provided us with a ref against the buffer, and we drop that here.
+ * Called under journal->j_list_lock.  The caller provided us with a ref
+ * against the buffer, and we drop that here.
  */
 static void release_buffer_page(struct buffer_head *bh)
 {
@@ -78,6 +78,19 @@ nope:
 }
 
 /*
+ * Decrement reference counter for data buffer. If it has been marked
+ * 'BH_Freed', release it and the page to which it belongs if possible.
+ */
+static void release_data_buffer(struct buffer_head *bh)
+{
+	if (buffer_freed(bh)) {
+		clear_buffer_freed(bh);
+		release_buffer_page(bh);
+	} else
+		put_bh(bh);
+}
+
+/*
  * Try to acquire jbd_lock_bh_state() against the buffer, when j_list_lock is
  * held.  For ranking reasons we must trylock.  If we lose, schedule away and
  * return 0.  j_list_lock is dropped in this case.
@@ -172,7 +185,7 @@ static void journal_do_submit_data(struct buffer_head **wbuf, int bufs)
 /*
  *  Submit all the data buffers to disk
  */
-static void journal_submit_data_buffers(journal_t *journal,
+static int journal_submit_data_buffers(journal_t *journal,
 				transaction_t *commit_transaction)
 {
 	struct journal_head *jh;
@@ -180,6 +193,7 @@ static void journal_submit_data_buffers(journal_t *journal,
 	int locked;
 	int bufs = 0;
 	struct buffer_head **wbuf = journal->j_wbuf;
+	int err = 0;
 
 	/*
 	 * Whenever we unlock the journal and sleep, things can get added
@@ -231,7 +245,7 @@ write_out_data:
 			if (locked)
 				unlock_buffer(bh);
 			BUFFER_TRACE(bh, "already cleaned up");
-			put_bh(bh);
+			release_data_buffer(bh);
 			continue;
 		}
 		if (locked && test_clear_buffer_dirty(bh)) {
@@ -253,15 +267,17 @@ write_out_data:
 			put_bh(bh);
 		} else {
 			BUFFER_TRACE(bh, "writeout complete: unfile");
+			if (unlikely(!buffer_uptodate(bh)))
+				err = -EIO;
 			__journal_unfile_buffer(jh);
 			jbd_unlock_bh_state(bh);
 			if (locked)
 				unlock_buffer(bh);
 			journal_remove_journal_head(bh);
-			/* Once for our safety reference, once for
+			/* One for our safety reference, other for
 			 * journal_remove_journal_head() */
 			put_bh(bh);
-			put_bh(bh);
+			release_data_buffer(bh);
 		}
 
 		if (need_resched() || spin_needbreak(&journal->j_list_lock)) {
@@ -271,6 +287,8 @@ write_out_data:
 	}
 	spin_unlock(&journal->j_list_lock);
 	journal_do_submit_data(wbuf, bufs);
+
+	return err;
 }
 
 /*
@@ -410,8 +428,7 @@ void journal_commit_transaction(journal_t *journal)
 	 * Now start flushing things to disk, in the order they appear
 	 * on the transaction lists.  Data blocks go first.
 	 */
-	err = 0;
-	journal_submit_data_buffers(journal, commit_transaction);
+	err = journal_submit_data_buffers(journal, commit_transaction);
 
 	/*
 	 * Wait for all previously submitted IO to complete.
@@ -426,10 +443,21 @@ void journal_commit_transaction(journal_t *journal)
 		if (buffer_locked(bh)) {
 			spin_unlock(&journal->j_list_lock);
 			wait_on_buffer(bh);
-			if (unlikely(!buffer_uptodate(bh)))
-				err = -EIO;
 			spin_lock(&journal->j_list_lock);
 		}
+		if (unlikely(!buffer_uptodate(bh))) {
+			if (TestSetPageLocked(bh->b_page)) {
+				spin_unlock(&journal->j_list_lock);
+				lock_page(bh->b_page);
+				spin_lock(&journal->j_list_lock);
+			}
+			if (bh->b_page->mapping)
+				set_bit(AS_EIO, &bh->b_page->mapping->flags);
+
+			unlock_page(bh->b_page);
+			SetPageError(bh->b_page);
+			err = -EIO;
+		}
 		if (!inverted_lock(journal, bh)) {
 			put_bh(bh);
 			spin_lock(&journal->j_list_lock);
@@ -443,17 +471,21 @@ void journal_commit_transaction(journal_t *journal)
 		} else {
 			jbd_unlock_bh_state(bh);
 		}
-		put_bh(bh);
+		release_data_buffer(bh);
 		cond_resched_lock(&journal->j_list_lock);
 	}
 	spin_unlock(&journal->j_list_lock);
 
-	if (err)
-		journal_abort(journal, err);
+	if (err) {
+		char b[BDEVNAME_SIZE];
 
-	journal_write_revoke_records(journal, commit_transaction);
+		printk(KERN_WARNING
+			"JBD: Detected IO errors while flushing file data "
+			"on %s\n", bdevname(journal->j_fs_dev, b));
+		err = 0;
+	}
 
-	jbd_debug(3, "JBD: commit phase 2\n");
+	journal_write_revoke_records(journal, commit_transaction);
 
 	/*
 	 * If we found any dirty or locked buffers, then we should have
diff --git a/fs/jbd/journal.c b/fs/jbd/journal.c
index b99c3b3654c..aa7143a8349 100644
--- a/fs/jbd/journal.c
+++ b/fs/jbd/journal.c
@@ -68,7 +68,6 @@ EXPORT_SYMBOL(journal_set_features);
 EXPORT_SYMBOL(journal_create);
 EXPORT_SYMBOL(journal_load);
 EXPORT_SYMBOL(journal_destroy);
-EXPORT_SYMBOL(journal_update_superblock);
 EXPORT_SYMBOL(journal_abort);
 EXPORT_SYMBOL(journal_errno);
 EXPORT_SYMBOL(journal_ack_err);
@@ -1636,9 +1635,10 @@ static int journal_init_journal_head_cache(void)
 
 static void journal_destroy_journal_head_cache(void)
 {
-	J_ASSERT(journal_head_cache != NULL);
-	kmem_cache_destroy(journal_head_cache);
-	journal_head_cache = NULL;
+	if (journal_head_cache) {
+		kmem_cache_destroy(journal_head_cache);
+		journal_head_cache = NULL;
+	}
 }
 
 /*
diff --git a/fs/jbd/revoke.c b/fs/jbd/revoke.c
index 1bb43e987f4..c7bd649bbbd 100644
--- a/fs/jbd/revoke.c
+++ b/fs/jbd/revoke.c
@@ -166,138 +166,123 @@ static struct jbd_revoke_record_s *find_revoke_record(journal_t *journal,
 	return NULL;
 }
 
+void journal_destroy_revoke_caches(void)
+{
+	if (revoke_record_cache) {
+		kmem_cache_destroy(revoke_record_cache);
+		revoke_record_cache = NULL;
+	}
+	if (revoke_table_cache) {
+		kmem_cache_destroy(revoke_table_cache);
+		revoke_table_cache = NULL;
+	}
+}
+
 int __init journal_init_revoke_caches(void)
 {
+	J_ASSERT(!revoke_record_cache);
+	J_ASSERT(!revoke_table_cache);
+
 	revoke_record_cache = kmem_cache_create("revoke_record",
 					   sizeof(struct jbd_revoke_record_s),
 					   0,
 					   SLAB_HWCACHE_ALIGN|SLAB_TEMPORARY,
 					   NULL);
 	if (!revoke_record_cache)
-		return -ENOMEM;
+		goto record_cache_failure;
 
 	revoke_table_cache = kmem_cache_create("revoke_table",
 					   sizeof(struct jbd_revoke_table_s),
 					   0, SLAB_TEMPORARY, NULL);
-	if (!revoke_table_cache) {
-		kmem_cache_destroy(revoke_record_cache);
-		revoke_record_cache = NULL;
-		return -ENOMEM;
-	}
+	if (!revoke_table_cache)
+		goto table_cache_failure;
+
 	return 0;
-}
 
-void journal_destroy_revoke_caches(void)
-{
-	kmem_cache_destroy(revoke_record_cache);
-	revoke_record_cache = NULL;
-	kmem_cache_destroy(revoke_table_cache);
-	revoke_table_cache = NULL;
+table_cache_failure:
+	journal_destroy_revoke_caches();
+record_cache_failure:
+	return -ENOMEM;
 }
 
-/* Initialise the revoke table for a given journal to a given size. */
-
-int journal_init_revoke(journal_t *journal, int hash_size)
+static struct jbd_revoke_table_s *journal_init_revoke_table(int hash_size)
 {
-	int shift, tmp;
+	int shift = 0;
+	int tmp = hash_size;
+	struct jbd_revoke_table_s *table;
 
-	J_ASSERT (journal->j_revoke_table[0] == NULL);
+	table = kmem_cache_alloc(revoke_table_cache, GFP_KERNEL);
+	if (!table)
+		goto out;
 
-	shift = 0;
-	tmp = hash_size;
 	while((tmp >>= 1UL) != 0UL)
 		shift++;
 
-	journal->j_revoke_table[0] = kmem_cache_alloc(revoke_table_cache, GFP_KERNEL);
-	if (!journal->j_revoke_table[0])
-		return -ENOMEM;
-	journal->j_revoke = journal->j_revoke_table[0];
-
-	/* Check that the hash_size is a power of two */
-	J_ASSERT(is_power_of_2(hash_size));
-
-	journal->j_revoke->hash_size = hash_size;
-
-	journal->j_revoke->hash_shift = shift;
-
-	journal->j_revoke->hash_table =
+	table->hash_size = hash_size;
+	table->hash_shift = shift;
+	table->hash_table =
 		kmalloc(hash_size * sizeof(struct list_head), GFP_KERNEL);
-	if (!journal->j_revoke->hash_table) {
-		kmem_cache_free(revoke_table_cache, journal->j_revoke_table[0]);
-		journal->j_revoke = NULL;
-		return -ENOMEM;
+	if (!table->hash_table) {
+		kmem_cache_free(revoke_table_cache, table);
+		table = NULL;
+		goto out;
 	}
 
 	for (tmp = 0; tmp < hash_size; tmp++)
-		INIT_LIST_HEAD(&journal->j_revoke->hash_table[tmp]);
+		INIT_LIST_HEAD(&table->hash_table[tmp]);
 
-	journal->j_revoke_table[1] = kmem_cache_alloc(revoke_table_cache, GFP_KERNEL);
-	if (!journal->j_revoke_table[1]) {
-		kfree(journal->j_revoke_table[0]->hash_table);
-		kmem_cache_free(revoke_table_cache, journal->j_revoke_table[0]);
-		return -ENOMEM;
+out:
+	return table;
+}
+
+static void journal_destroy_revoke_table(struct jbd_revoke_table_s *table)
+{
+	int i;
+	struct list_head *hash_list;
+
+	for (i = 0; i < table->hash_size; i++) {
+		hash_list = &table->hash_table[i];
+		J_ASSERT(list_empty(hash_list));
 	}
 
-	journal->j_revoke = journal->j_revoke_table[1];
+	kfree(table->hash_table);
+	kmem_cache_free(revoke_table_cache, table);
+}
 
-	/* Check that the hash_size is a power of two */
+/* Initialise the revoke table for a given journal to a given size. */
+int journal_init_revoke(journal_t *journal, int hash_size)
+{
+	J_ASSERT(journal->j_revoke_table[0] == NULL);
 	J_ASSERT(is_power_of_2(hash_size));
 
-	journal->j_revoke->hash_size = hash_size;
+	journal->j_revoke_table[0] = journal_init_revoke_table(hash_size);
+	if (!journal->j_revoke_table[0])
+		goto fail0;
 
-	journal->j_revoke->hash_shift = shift;
+	journal->j_revoke_table[1] = journal_init_revoke_table(hash_size);
+	if (!journal->j_revoke_table[1])
+		goto fail1;
 
-	journal->j_revoke->hash_table =
-		kmalloc(hash_size * sizeof(struct list_head), GFP_KERNEL);
-	if (!journal->j_revoke->hash_table) {
-		kfree(journal->j_revoke_table[0]->hash_table);
-		kmem_cache_free(revoke_table_cache, journal->j_revoke_table[0]);
-		kmem_cache_free(revoke_table_cache, journal->j_revoke_table[1]);
-		journal->j_revoke = NULL;
-		return -ENOMEM;
-	}
-
-	for (tmp = 0; tmp < hash_size; tmp++)
-		INIT_LIST_HEAD(&journal->j_revoke->hash_table[tmp]);
+	journal->j_revoke = journal->j_revoke_table[1];
 
 	spin_lock_init(&journal->j_revoke_lock);
 
 	return 0;
-}
 
-/* Destoy a journal's revoke table.  The table must already be empty! */
+fail1:
+	journal_destroy_revoke_table(journal->j_revoke_table[0]);
+fail0:
+	return -ENOMEM;
+}
 
+/* Destroy a journal's revoke table.  The table must already be empty! */
 void journal_destroy_revoke(journal_t *journal)
 {
-	struct jbd_revoke_table_s *table;
-	struct list_head *hash_list;
-	int i;
-
-	table = journal->j_revoke_table[0];
-	if (!table)
-		return;
-
-	for (i=0; i<table->hash_size; i++) {
-		hash_list = &table->hash_table[i];
-		J_ASSERT (list_empty(hash_list));
-	}
-
-	kfree(table->hash_table);
-	kmem_cache_free(revoke_table_cache, table);
-	journal->j_revoke = NULL;
-
-	table = journal->j_revoke_table[1];
-	if (!table)
-		return;
-
-	for (i=0; i<table->hash_size; i++) {
-		hash_list = &table->hash_table[i];
-		J_ASSERT (list_empty(hash_list));
-	}
-
-	kfree(table->hash_table);
-	kmem_cache_free(revoke_table_cache, table);
 	journal->j_revoke = NULL;
+	if (journal->j_revoke_table[0])
+		journal_destroy_revoke_table(journal->j_revoke_table[0]);
+	if (journal->j_revoke_table[1])
+		journal_destroy_revoke_table(journal->j_revoke_table[1]);
 }
 
 
diff --git a/fs/jbd/transaction.c b/fs/jbd/transaction.c
index 67ff2024c23..8dee3200750 100644
--- a/fs/jbd/transaction.c
+++ b/fs/jbd/transaction.c
@@ -1648,12 +1648,42 @@ out:
 	return;
 }
 
+/*
+ * journal_try_to_free_buffers() could race with journal_commit_transaction()
+ * The latter might still hold the a count on buffers when inspecting
+ * them on t_syncdata_list or t_locked_list.
+ *
+ * journal_try_to_free_buffers() will call this function to
+ * wait for the current transaction to finish syncing data buffers, before
+ * tryinf to free that buffer.
+ *
+ * Called with journal->j_state_lock held.
+ */
+static void journal_wait_for_transaction_sync_data(journal_t *journal)
+{
+	transaction_t *transaction = NULL;
+	tid_t tid;
+
+	spin_lock(&journal->j_state_lock);
+	transaction = journal->j_committing_transaction;
+
+	if (!transaction) {
+		spin_unlock(&journal->j_state_lock);
+		return;
+	}
+
+	tid = transaction->t_tid;
+	spin_unlock(&journal->j_state_lock);
+	log_wait_commit(journal, tid);
+}
 
 /**
  * int journal_try_to_free_buffers() - try to free page buffers.
  * @journal: journal for operation
  * @page: to try and free
- * @unused_gfp_mask: unused
+ * @gfp_mask: we use the mask to detect how hard should we try to release
+ * buffers. If __GFP_WAIT and __GFP_FS is set, we wait for commit code to
+ * release the buffers.
  *
  *
  * For all the buffers on this page,
@@ -1682,9 +1712,11 @@ out:
  * journal_try_to_free_buffer() is changing its state.  But that
  * cannot happen because we never reallocate freed data as metadata
  * while the data is part of a transaction.  Yes?
+ *
+ * Return 0 on failure, 1 on success
  */
 int journal_try_to_free_buffers(journal_t *journal,
-				struct page *page, gfp_t unused_gfp_mask)
+				struct page *page, gfp_t gfp_mask)
 {
 	struct buffer_head *head;
 	struct buffer_head *bh;
@@ -1713,7 +1745,28 @@ int journal_try_to_free_buffers(journal_t *journal,
 		if (buffer_jbd(bh))
 			goto busy;
 	} while ((bh = bh->b_this_page) != head);
+
 	ret = try_to_free_buffers(page);
+
+	/*
+	 * There are a number of places where journal_try_to_free_buffers()
+	 * could race with journal_commit_transaction(), the later still
+	 * holds the reference to the buffers to free while processing them.
+	 * try_to_free_buffers() failed to free those buffers. Some of the
+	 * caller of releasepage() request page buffers to be dropped, otherwise
+	 * treat the fail-to-free as errors (such as generic_file_direct_IO())
+	 *
+	 * So, if the caller of try_to_release_page() wants the synchronous
+	 * behaviour(i.e make sure buffers are dropped upon return),
+	 * let's wait for the current transaction to finish flush of
+	 * dirty data buffers, then try to free those buffers again,
+	 * with the journal locked.
+	 */
+	if (ret == 0 && (gfp_mask & __GFP_WAIT) && (gfp_mask & __GFP_FS)) {
+		journal_wait_for_transaction_sync_data(journal);
+		ret = try_to_free_buffers(page);
+	}
+
 busy:
 	return ret;
 }