From 5b1016e62f74c53e0330403025954c8d95384c03 Mon Sep 17 00:00:00 2001
From: Kent Overstreet <kmo@daterainc.com>
Date: Wed, 19 Mar 2014 17:49:37 -0700
Subject: bcache: Fix a bug when detaching

After detaching a backing device from a cache set, a bit wasn't getting
reset meaning the second detach wouldn't work correctly.

Signed-off-by: Kent Overstreet <kmo@daterainc.com>

diff --git a/drivers/md/bcache/super.c b/drivers/md/bcache/super.c
index 926ded8..1ea9fa2 100644
--- a/drivers/md/bcache/super.c
+++ b/drivers/md/bcache/super.c
@@ -927,6 +927,7 @@ static void cached_dev_detach_finish(struct work_struct *w)
 	list_move(&dc->list, &uncached_devices);
 
 	clear_bit(BCACHE_DEV_DETACHING, &dc->disk.flags);
+	clear_bit(BCACHE_DEV_UNLINK_DONE, &dc->disk.flags);
 
 	mutex_unlock(&bch_register_lock);
 
@@ -1405,9 +1406,11 @@ static void cache_set_flush(struct closure *cl)
 		if (ca->alloc_thread)
 			kthread_stop(ca->alloc_thread);
 
-	cancel_delayed_work_sync(&c->journal.work);
-	/* flush last journal entry if needed */
-	c->journal.work.work.func(&c->journal.work.work);
+	if (c->journal.cur) {
+		cancel_delayed_work_sync(&c->journal.work);
+		/* flush last journal entry if needed */
+		c->journal.work.work.func(&c->journal.work.work);
+	}
 
 	closure_return(cl);
 }
-- 
cgit v0.10.2


From 9aa61a992acceeec0d1de2cd99938421498659d5 Mon Sep 17 00:00:00 2001
From: Kent Overstreet <kmo@daterainc.com>
Date: Thu, 10 Apr 2014 17:58:49 -0700
Subject: bcache: Fix a journal replay bug

journal replay wansn't validating pointers with bch_extent_invalid() before
derefing, fixed

Signed-off-by: Kent Overstreet <kmo@daterainc.com>

diff --git a/drivers/md/bcache/extents.c b/drivers/md/bcache/extents.c
index 3a0de4c..243de0bf 100644
--- a/drivers/md/bcache/extents.c
+++ b/drivers/md/bcache/extents.c
@@ -474,9 +474,8 @@ out:
 	return false;
 }
 
-static bool bch_extent_invalid(struct btree_keys *bk, const struct bkey *k)
+bool __bch_extent_invalid(struct cache_set *c, const struct bkey *k)
 {
-	struct btree *b = container_of(bk, struct btree, keys);
 	char buf[80];
 
 	if (!KEY_SIZE(k))
@@ -485,16 +484,22 @@ static bool bch_extent_invalid(struct btree_keys *bk, const struct bkey *k)
 	if (KEY_SIZE(k) > KEY_OFFSET(k))
 		goto bad;
 
-	if (__ptr_invalid(b->c, k))
+	if (__ptr_invalid(c, k))
 		goto bad;
 
 	return false;
 bad:
 	bch_extent_to_text(buf, sizeof(buf), k);
-	cache_bug(b->c, "spotted extent %s: %s", buf, bch_ptr_status(b->c, k));
+	cache_bug(c, "spotted extent %s: %s", buf, bch_ptr_status(c, k));
 	return true;
 }
 
+static bool bch_extent_invalid(struct btree_keys *bk, const struct bkey *k)
+{
+	struct btree *b = container_of(bk, struct btree, keys);
+	return __bch_extent_invalid(b->c, k);
+}
+
 static bool bch_extent_bad_expensive(struct btree *b, const struct bkey *k,
 				     unsigned ptr)
 {
diff --git a/drivers/md/bcache/extents.h b/drivers/md/bcache/extents.h
index e4e2340..e2ed540 100644
--- a/drivers/md/bcache/extents.h
+++ b/drivers/md/bcache/extents.h
@@ -9,5 +9,6 @@ struct cache_set;
 
 void bch_extent_to_text(char *, size_t, const struct bkey *);
 bool __bch_btree_ptr_invalid(struct cache_set *, const struct bkey *);
+bool __bch_extent_invalid(struct cache_set *, const struct bkey *);
 
 #endif /* _BCACHE_EXTENTS_H */
diff --git a/drivers/md/bcache/journal.c b/drivers/md/bcache/journal.c
index 59e8202..363b881 100644
--- a/drivers/md/bcache/journal.c
+++ b/drivers/md/bcache/journal.c
@@ -7,6 +7,7 @@
 #include "bcache.h"
 #include "btree.h"
 #include "debug.h"
+#include "extents.h"
 
 #include <trace/events/bcache.h>
 
@@ -291,15 +292,16 @@ void bch_journal_mark(struct cache_set *c, struct list_head *list)
 
 		for (k = i->j.start;
 		     k < bset_bkey_last(&i->j);
-		     k = bkey_next(k)) {
-			unsigned j;
+		     k = bkey_next(k))
+			if (!__bch_extent_invalid(c, k)) {
+				unsigned j;
 
-			for (j = 0; j < KEY_PTRS(k); j++)
-				if (ptr_available(c, k, j))
-					atomic_inc(&PTR_BUCKET(c, k, j)->pin);
+				for (j = 0; j < KEY_PTRS(k); j++)
+					if (ptr_available(c, k, j))
+						atomic_inc(&PTR_BUCKET(c, k, j)->pin);
 
-			bch_initial_mark_key(c, 0, k);
-		}
+				bch_initial_mark_key(c, 0, k);
+			}
 	}
 }
 
-- 
cgit v0.10.2


From dbd810ab678d262d3772d29b65844d7b20dc47bc Mon Sep 17 00:00:00 2001
From: Surbhi Palande <sap@daterainc.com>
Date: Thu, 10 Apr 2014 16:09:51 -0700
Subject: bcache: Fix to remove the rcu_sched stalls.

while loop was executing infinitely.
This fix ends the while loop gracefully.

Signed-off-by: Surbhi Palande <sap@daterainc.com>
Signed-off-by: Kent Overstreet <kmo@daterainc.com>

diff --git a/drivers/md/bcache/journal.c b/drivers/md/bcache/journal.c
index 363b881..ead001c 100644
--- a/drivers/md/bcache/journal.c
+++ b/drivers/md/bcache/journal.c
@@ -194,7 +194,8 @@ int bch_journal_read(struct cache_set *c, struct list_head *list)
 			continue;
 bsearch:
 		/* Binary search */
-		m = r = find_next_bit(bitmap, ca->sb.njournal_buckets, l + 1);
+		m = l;
+		r = find_next_bit(bitmap, ca->sb.njournal_buckets, l + 1);
 		pr_debug("starting binary search, l %u r %u", l, r);
 
 		while (l + 1 < r) {
-- 
cgit v0.10.2


From 8b326d3a2a76912dfed2f0ab937d59fae9512ca2 Mon Sep 17 00:00:00 2001
From: Slava Pestov <sp@daterainc.com>
Date: Mon, 21 Apr 2014 18:22:35 -0700
Subject: bcache allocator: send discards with correct size


diff --git a/drivers/md/bcache/alloc.c b/drivers/md/bcache/alloc.c
index 443d03f..8eeab72 100644
--- a/drivers/md/bcache/alloc.c
+++ b/drivers/md/bcache/alloc.c
@@ -331,7 +331,7 @@ static int bch_allocator_thread(void *arg)
 				mutex_unlock(&ca->set->bucket_lock);
 				blkdev_issue_discard(ca->bdev,
 					bucket_to_sector(ca->set, bucket),
-					ca->sb.block_size, GFP_KERNEL, 0);
+					ca->sb.bucket_size, GFP_KERNEL, 0);
 				mutex_lock(&ca->set->bucket_lock);
 			}
 
-- 
cgit v0.10.2


From e5112201c1285841f8b565ece5d6ae7e0d7947a2 Mon Sep 17 00:00:00 2001
From: Slava Pestov <sp@daterainc.com>
Date: Tue, 29 Apr 2014 15:39:27 -0700
Subject: bcache: fix lockdep warnings on shutdown


diff --git a/drivers/md/bcache/super.c b/drivers/md/bcache/super.c
index 1ea9fa2..09573c2 100644
--- a/drivers/md/bcache/super.c
+++ b/drivers/md/bcache/super.c
@@ -1214,7 +1214,9 @@ void bch_flash_dev_release(struct kobject *kobj)
 static void flash_dev_free(struct closure *cl)
 {
 	struct bcache_device *d = container_of(cl, struct bcache_device, cl);
+	mutex_lock(&bch_register_lock);
 	bcache_device_free(d);
+	mutex_unlock(&bch_register_lock);
 	kobject_put(&d->kobj);
 }
 
@@ -1222,7 +1224,9 @@ static void flash_dev_flush(struct closure *cl)
 {
 	struct bcache_device *d = container_of(cl, struct bcache_device, cl);
 
+	mutex_lock(&bch_register_lock);
 	bcache_device_unlink(d);
+	mutex_unlock(&bch_register_lock);
 	kobject_del(&d->kobj);
 	continue_at(cl, flash_dev_free, system_wq);
 }
-- 
cgit v0.10.2


From a664d0f05a2ec02c8f042db536d84d15d6e19e81 Mon Sep 17 00:00:00 2001
From: Slava Pestov <sp@daterainc.com>
Date: Tue, 20 May 2014 12:20:28 -0700
Subject: bcache: fix crash on shutdown in passthrough mode

We never started the writeback thread in this case, so don't stop it.

diff --git a/drivers/md/bcache/super.c b/drivers/md/bcache/super.c
index 09573c2..6ceaec3 100644
--- a/drivers/md/bcache/super.c
+++ b/drivers/md/bcache/super.c
@@ -1071,7 +1071,8 @@ static void cached_dev_free(struct closure *cl)
 	struct cached_dev *dc = container_of(cl, struct cached_dev, disk.cl);
 
 	cancel_delayed_work_sync(&dc->writeback_rate_update);
-	kthread_stop(dc->writeback_thread);
+	if (!IS_ERR_OR_NULL(dc->writeback_thread))
+		kthread_stop(dc->writeback_thread);
 
 	mutex_lock(&bch_register_lock);
 
-- 
cgit v0.10.2


From c5aa4a3157b55bdca18dd2a9d9f43314470b6d32 Mon Sep 17 00:00:00 2001
From: Slava Pestov <sp@daterainc.com>
Date: Mon, 21 Apr 2014 18:23:12 -0700
Subject: bcache: wait for buckets when allocating new btree root

Tested:
- sometimes bcache_tier test would hang on startup with a failure
  to allocate the btree root -- no longer seeing this

Signed-off-by: Kent Overstreet <kmo@daterainc.com>

diff --git a/drivers/md/bcache/btree.c b/drivers/md/bcache/btree.c
index 7347b61..9dd9f1c 100644
--- a/drivers/md/bcache/btree.c
+++ b/drivers/md/bcache/btree.c
@@ -1060,15 +1060,15 @@ static void btree_node_free(struct btree *b)
 	mutex_unlock(&b->c->bucket_lock);
 }
 
-struct btree *bch_btree_node_alloc(struct cache_set *c, struct btree_op *op,
-				   int level)
+struct btree *__bch_btree_node_alloc(struct cache_set *c, struct btree_op *op,
+				     int level, bool wait)
 {
 	BKEY_PADDED(key) k;
 	struct btree *b = ERR_PTR(-EAGAIN);
 
 	mutex_lock(&c->bucket_lock);
 retry:
-	if (__bch_bucket_alloc_set(c, RESERVE_BTREE, &k.key, 1, op != NULL))
+	if (__bch_bucket_alloc_set(c, RESERVE_BTREE, &k.key, 1, wait))
 		goto err;
 
 	bkey_put(c, &k.key);
@@ -1100,6 +1100,12 @@ err:
 	return b;
 }
 
+static struct btree *bch_btree_node_alloc(struct cache_set *c,
+					  struct btree_op *op, int level)
+{
+	return __bch_btree_node_alloc(c, op, level, op != NULL);
+}
+
 static struct btree *btree_node_alloc_replacement(struct btree *b,
 						  struct btree_op *op)
 {
diff --git a/drivers/md/bcache/btree.h b/drivers/md/bcache/btree.h
index 91dfa5e..0044182 100644
--- a/drivers/md/bcache/btree.h
+++ b/drivers/md/bcache/btree.h
@@ -242,7 +242,8 @@ void __bch_btree_node_write(struct btree *, struct closure *);
 void bch_btree_node_write(struct btree *, struct closure *);
 
 void bch_btree_set_root(struct btree *);
-struct btree *bch_btree_node_alloc(struct cache_set *, struct btree_op *, int);
+struct btree *__bch_btree_node_alloc(struct cache_set *, struct btree_op *,
+				     int, bool);
 struct btree *bch_btree_node_get(struct cache_set *, struct btree_op *,
 				 struct bkey *, int, bool);
 
diff --git a/drivers/md/bcache/super.c b/drivers/md/bcache/super.c
index 6ceaec3..3b043a0 100644
--- a/drivers/md/bcache/super.c
+++ b/drivers/md/bcache/super.c
@@ -1669,7 +1669,7 @@ static void run_cache_set(struct cache_set *c)
 			goto err;
 
 		err = "cannot allocate new btree root";
-		c->root = bch_btree_node_alloc(c, NULL, 0);
+		c->root = __bch_btree_node_alloc(c, NULL, 0, true);
 		if (IS_ERR_OR_NULL(c->root))
 			goto err;
 
-- 
cgit v0.10.2


From 9e5c353510b26500bd6b8309823ac9ef2837b761 Mon Sep 17 00:00:00 2001
From: Slava Pestov <sp@daterainc.com>
Date: Thu, 1 May 2014 13:48:57 -0700
Subject: bcache: fix uninterruptible sleep in writeback thread

There were two issues here:

- writeback thread did not start until the device first became dirty
- writeback thread used uninterruptible sleep once running

Without this patch I see kernel warnings printed and a load average of
1.52 after booting my test VM. With this patch the warnings are gone and
the load average is near 0.00 as expected.

Signed-off-by: Kent Overstreet <kmo@daterainc.com>

diff --git a/drivers/md/bcache/super.c b/drivers/md/bcache/super.c
index 3b043a0..00cc425 100644
--- a/drivers/md/bcache/super.c
+++ b/drivers/md/bcache/super.c
@@ -1042,6 +1042,9 @@ int bch_cached_dev_attach(struct cached_dev *dc, struct cache_set *c)
 	 */
 	atomic_set(&dc->count, 1);
 
+	if (bch_cached_dev_writeback_start(dc))
+		return -ENOMEM;
+
 	if (BDEV_STATE(&dc->sb) == BDEV_STATE_DIRTY) {
 		bch_sectors_dirty_init(dc);
 		atomic_set(&dc->has_dirty, 1);
diff --git a/drivers/md/bcache/writeback.c b/drivers/md/bcache/writeback.c
index f4300e4..f1986bc 100644
--- a/drivers/md/bcache/writeback.c
+++ b/drivers/md/bcache/writeback.c
@@ -239,7 +239,7 @@ static void read_dirty(struct cached_dev *dc)
 		if (KEY_START(&w->key) != dc->last_read ||
 		    jiffies_to_msecs(delay) > 50)
 			while (!kthread_should_stop() && delay)
-				delay = schedule_timeout_uninterruptible(delay);
+				delay = schedule_timeout_interruptible(delay);
 
 		dc->last_read	= KEY_OFFSET(&w->key);
 
@@ -436,7 +436,7 @@ static int bch_writeback_thread(void *arg)
 			while (delay &&
 			       !kthread_should_stop() &&
 			       !test_bit(BCACHE_DEV_DETACHING, &dc->disk.flags))
-				delay = schedule_timeout_uninterruptible(delay);
+				delay = schedule_timeout_interruptible(delay);
 		}
 	}
 
@@ -478,7 +478,7 @@ void bch_sectors_dirty_init(struct cached_dev *dc)
 	dc->disk.sectors_dirty_last = bcache_dev_sectors_dirty(&dc->disk);
 }
 
-int bch_cached_dev_writeback_init(struct cached_dev *dc)
+void bch_cached_dev_writeback_init(struct cached_dev *dc)
 {
 	sema_init(&dc->in_flight, 64);
 	init_rwsem(&dc->writeback_lock);
@@ -494,14 +494,20 @@ int bch_cached_dev_writeback_init(struct cached_dev *dc)
 	dc->writeback_rate_d_term	= 30;
 	dc->writeback_rate_p_term_inverse = 6000;
 
+	INIT_DELAYED_WORK(&dc->writeback_rate_update, update_writeback_rate);
+}
+
+int bch_cached_dev_writeback_start(struct cached_dev *dc)
+{
 	dc->writeback_thread = kthread_create(bch_writeback_thread, dc,
 					      "bcache_writeback");
 	if (IS_ERR(dc->writeback_thread))
 		return PTR_ERR(dc->writeback_thread);
 
-	INIT_DELAYED_WORK(&dc->writeback_rate_update, update_writeback_rate);
 	schedule_delayed_work(&dc->writeback_rate_update,
 			      dc->writeback_rate_update_seconds * HZ);
 
+	bch_writeback_queue(dc);
+
 	return 0;
 }
diff --git a/drivers/md/bcache/writeback.h b/drivers/md/bcache/writeback.h
index e2f8598..0a9dab1 100644
--- a/drivers/md/bcache/writeback.h
+++ b/drivers/md/bcache/writeback.h
@@ -85,6 +85,7 @@ static inline void bch_writeback_add(struct cached_dev *dc)
 void bcache_dev_sectors_dirty_add(struct cache_set *, unsigned, uint64_t, int);
 
 void bch_sectors_dirty_init(struct cached_dev *dc);
-int bch_cached_dev_writeback_init(struct cached_dev *);
+void bch_cached_dev_writeback_init(struct cached_dev *);
+int bch_cached_dev_writeback_start(struct cached_dev *);
 
 #endif
-- 
cgit v0.10.2


From bcf090e0040e30f8409e6a535a01e6473afb096f Mon Sep 17 00:00:00 2001
From: Kent Overstreet <kmo@daterainc.com>
Date: Mon, 19 May 2014 08:57:55 -0700
Subject: bcache: Make sure to pass GFP_WAIT to mempool_alloc()

this was very wrong - mempool_alloc() only guarantees success with GFP_WAIT.
bcache uses GFP_NOWAIT in various other places where we have a fallback,
circuits must've gotten crossed when writing this code or something.

Signed-off-by: Kent Overstreet <kmo@daterainc.com>

diff --git a/drivers/md/bcache/btree.c b/drivers/md/bcache/btree.c
index 9dd9f1c..e538d45 100644
--- a/drivers/md/bcache/btree.c
+++ b/drivers/md/bcache/btree.c
@@ -202,7 +202,7 @@ void bch_btree_node_read_done(struct btree *b)
 	struct bset *i = btree_bset_first(b);
 	struct btree_iter *iter;
 
-	iter = mempool_alloc(b->c->fill_iter, GFP_NOWAIT);
+	iter = mempool_alloc(b->c->fill_iter, GFP_NOIO);
 	iter->size = b->c->sb.bucket_size / b->c->sb.block_size;
 	iter->used = 0;
 
-- 
cgit v0.10.2


From 501d52a90cbe652b41336c206ff0e95799d5a9b5 Mon Sep 17 00:00:00 2001
From: Kent Overstreet <kmo@daterainc.com>
Date: Mon, 19 May 2014 08:55:40 -0700
Subject: bcache: Allocate bounce buffers with GFP_NOWAIT

There's no point in blocking on these allocations, since our fallback paths will
probably go faster than blocking.

Change-Id: I733ca202c25cb36bde02607a0a60552229a4241c

diff --git a/drivers/md/bcache/bset.c b/drivers/md/bcache/bset.c
index 5454164..646fe85 100644
--- a/drivers/md/bcache/bset.c
+++ b/drivers/md/bcache/bset.c
@@ -1182,7 +1182,7 @@ static void __btree_sort(struct btree_keys *b, struct btree_iter *iter,
 {
 	uint64_t start_time;
 	bool used_mempool = false;
-	struct bset *out = (void *) __get_free_pages(__GFP_NOWARN|GFP_NOIO,
+	struct bset *out = (void *) __get_free_pages(__GFP_NOWARN|GFP_NOWAIT,
 						     order);
 	if (!out) {
 		struct page *outp;
diff --git a/drivers/md/bcache/btree.c b/drivers/md/bcache/btree.c
index e538d45..39c7f5b 100644
--- a/drivers/md/bcache/btree.c
+++ b/drivers/md/bcache/btree.c
@@ -421,7 +421,7 @@ static void do_btree_node_write(struct btree *b)
 	SET_PTR_OFFSET(&k.key, 0, PTR_OFFSET(&k.key, 0) +
 		       bset_sector_offset(&b->keys, i));
 
-	if (!bio_alloc_pages(b->bio, GFP_NOIO)) {
+	if (!bio_alloc_pages(b->bio, __GFP_NOWARN|GFP_NOWAIT)) {
 		int j;
 		struct bio_vec *bv;
 		void *base = (void *) ((unsigned long) i & ~(PAGE_SIZE - 1));
-- 
cgit v0.10.2


From 8e0948080670f6330229718b15a6a1a011d441ce Mon Sep 17 00:00:00 2001
From: Slava Pestov <sp@daterainc.com>
Date: Mon, 30 Jun 2014 22:31:20 -0700
Subject: bcache: fix typo in bch_bkey_equal_header

Signed-off-by: Kent Overstreet <kmo@daterainc.com>

diff --git a/drivers/md/bcache/bset.h b/drivers/md/bcache/bset.h
index 5f6728d..ae96462 100644
--- a/drivers/md/bcache/bset.h
+++ b/drivers/md/bcache/bset.h
@@ -453,7 +453,7 @@ static inline bool bch_bkey_equal_header(const struct bkey *l,
 {
 	return (KEY_DIRTY(l) == KEY_DIRTY(r) &&
 		KEY_PTRS(l) == KEY_PTRS(r) &&
-		KEY_CSUM(l) == KEY_CSUM(l));
+		KEY_CSUM(l) == KEY_CSUM(r));
 }
 
 /* Keylists */
-- 
cgit v0.10.2


From 60ae81eee86dd7a520db8c1e3d702b49fc0418b5 Mon Sep 17 00:00:00 2001
From: Slava Pestov <sp@daterainc.com>
Date: Thu, 22 May 2014 12:14:24 -0700
Subject: bcache: bcache_write tracepoint was crashing

Signed-off-by: Kent Overstreet <kmo@daterainc.com>

diff --git a/drivers/md/bcache/request.c b/drivers/md/bcache/request.c
index 15fff4f..62e6e98 100644
--- a/drivers/md/bcache/request.c
+++ b/drivers/md/bcache/request.c
@@ -311,7 +311,8 @@ void bch_data_insert(struct closure *cl)
 {
 	struct data_insert_op *op = container_of(cl, struct data_insert_op, cl);
 
-	trace_bcache_write(op->bio, op->writeback, op->bypass);
+	trace_bcache_write(op->c, op->inode, op->bio,
+			   op->writeback, op->bypass);
 
 	bch_keylist_init(&op->insert_keys);
 	bio_get(op->bio);
diff --git a/include/trace/events/bcache.h b/include/trace/events/bcache.h
index c9c3c04..6778e41 100644
--- a/include/trace/events/bcache.h
+++ b/include/trace/events/bcache.h
@@ -148,11 +148,13 @@ TRACE_EVENT(bcache_read,
 );
 
 TRACE_EVENT(bcache_write,
-	TP_PROTO(struct bio *bio, bool writeback, bool bypass),
-	TP_ARGS(bio, writeback, bypass),
+	TP_PROTO(struct cache_set *c, u64 inode, struct bio *bio,
+		bool writeback, bool bypass),
+	TP_ARGS(c, inode, bio, writeback, bypass),
 
 	TP_STRUCT__entry(
-		__field(dev_t,		dev			)
+		__array(char,		uuid,	16		)
+		__field(u64,		inode			)
 		__field(sector_t,	sector			)
 		__field(unsigned int,	nr_sector		)
 		__array(char,		rwbs,	6		)
@@ -161,7 +163,8 @@ TRACE_EVENT(bcache_write,
 	),
 
 	TP_fast_assign(
-		__entry->dev		= bio->bi_bdev->bd_dev;
+		memcpy(__entry->uuid, c->sb.set_uuid, 16);
+		__entry->inode		= inode;
 		__entry->sector		= bio->bi_iter.bi_sector;
 		__entry->nr_sector	= bio->bi_iter.bi_size >> 9;
 		blk_fill_rwbs(__entry->rwbs, bio->bi_rw, bio->bi_iter.bi_size);
@@ -169,8 +172,8 @@ TRACE_EVENT(bcache_write,
 		__entry->bypass = bypass;
 	),
 
-	TP_printk("%d,%d  %s %llu + %u hit %u bypass %u",
-		  MAJOR(__entry->dev), MINOR(__entry->dev),
+	TP_printk("%pU inode %llu  %s %llu + %u hit %u bypass %u",
+		  __entry->uuid, __entry->inode,
 		  __entry->rwbs, (unsigned long long)__entry->sector,
 		  __entry->nr_sector, __entry->writeback, __entry->bypass)
 );
-- 
cgit v0.10.2


From 913dc33fb2720fb5f979011664294137ddd8b13b Mon Sep 17 00:00:00 2001
From: Slava Pestov <sp@daterainc.com>
Date: Fri, 23 May 2014 11:18:35 -0700
Subject: bcache: fix crash in bcache_btree_node_alloc_fail tracepoint

'b' was NULL.

Change-Id: Icac0fd04afa2d23f213d96d51afd53374e6dd0c0

diff --git a/drivers/md/bcache/btree.c b/drivers/md/bcache/btree.c
index 39c7f5b..f823785 100644
--- a/drivers/md/bcache/btree.c
+++ b/drivers/md/bcache/btree.c
@@ -1096,7 +1096,7 @@ err_free:
 err:
 	mutex_unlock(&c->bucket_lock);
 
-	trace_bcache_btree_node_alloc_fail(b);
+	trace_bcache_btree_node_alloc_fail(c);
 	return b;
 }
 
diff --git a/include/trace/events/bcache.h b/include/trace/events/bcache.h
index 6778e41..981acf7 100644
--- a/include/trace/events/bcache.h
+++ b/include/trace/events/bcache.h
@@ -261,9 +261,9 @@ DEFINE_EVENT(btree_node, bcache_btree_node_alloc,
 	TP_ARGS(b)
 );
 
-DEFINE_EVENT(btree_node, bcache_btree_node_alloc_fail,
-	TP_PROTO(struct btree *b),
-	TP_ARGS(b)
+DEFINE_EVENT(cache_set, bcache_btree_node_alloc_fail,
+	TP_PROTO(struct cache_set *c),
+	TP_ARGS(c)
 );
 
 DEFINE_EVENT(btree_node, bcache_btree_node_free,
-- 
cgit v0.10.2


From 6b708de64adb6dc8319e7aeac922b46904fbeeec Mon Sep 17 00:00:00 2001
From: Kent Overstreet <kmo@daterainc.com>
Date: Mon, 2 Jun 2014 15:39:44 -0700
Subject: bcache: Fix an infinite loop in journal replay

When running with multiple cache devices, if one of the devices has a completely
empty journal but we'd already found some journal entries on a previosu device
we'd go into an infinite loop.

Change-Id: I1dcdc0d738192746de28f40e8b08825b0dea5e2b
Signed-off-by: Kent Overstreet <kmo@daterainc.com>

diff --git a/drivers/md/bcache/journal.c b/drivers/md/bcache/journal.c
index ead001c..fe080ad 100644
--- a/drivers/md/bcache/journal.c
+++ b/drivers/md/bcache/journal.c
@@ -190,9 +190,12 @@ int bch_journal_read(struct cache_set *c, struct list_head *list)
 			if (read_bucket(l))
 				goto bsearch;
 
-		if (list_empty(list))
+		/* no journal entries on this device? */
+		if (l == ca->sb.njournal_buckets)
 			continue;
 bsearch:
+		BUG_ON(list_empty(list));
+
 		/* Binary search */
 		m = l;
 		r = find_next_bit(bitmap, ca->sb.njournal_buckets, l + 1);
-- 
cgit v0.10.2


From 400ffaa2acd72274e2c7293a9724382383bebf3e Mon Sep 17 00:00:00 2001
From: Slava Pestov <sp@daterainc.com>
Date: Sat, 12 Jul 2014 21:53:11 -0700
Subject: bcache: fix use-after-free in btree_gc_coalesce()

If we goto out_nocoalesce after we free new_nodes[0], we end up freeing
new_nodes[0] again. This was generating a lockdep warning. The fix is
to set new_nodes[0] to NULL, since the out_nocoalesce path safely
ignores NULL entries in the new_nodes array.

This regression was introduced in 2d7f9531.

Change-Id: I76564d7257800583214376b4bacf236cda90c89c

diff --git a/drivers/md/bcache/btree.c b/drivers/md/bcache/btree.c
index f823785..776583f 100644
--- a/drivers/md/bcache/btree.c
+++ b/drivers/md/bcache/btree.c
@@ -1409,6 +1409,7 @@ static int btree_gc_coalesce(struct btree *b, struct btree_op *op,
 	BUG_ON(btree_bset_first(new_nodes[0])->keys);
 	btree_node_free(new_nodes[0]);
 	rw_unlock(true, new_nodes[0]);
+	new_nodes[0] = NULL;
 
 	for (i = 0; i < nodes; i++) {
 		if (__bch_keylist_realloc(&keylist, bkey_u64s(&r[i].b->key)))
-- 
cgit v0.10.2


From d83353b319d47ef8cce82467da6a25c2d558253f Mon Sep 17 00:00:00 2001
From: Kent Overstreet <kmo@daterainc.com>
Date: Wed, 11 Jun 2014 19:44:49 -0700
Subject: bcache: Fix more early shutdown bugs

Signed-off-by: Kent Overstreet <kmo@daterainc.com>

diff --git a/drivers/md/bcache/super.c b/drivers/md/bcache/super.c
index 00cc425..29dd1e8 100644
--- a/drivers/md/bcache/super.c
+++ b/drivers/md/bcache/super.c
@@ -733,8 +733,6 @@ static void bcache_device_detach(struct bcache_device *d)
 static void bcache_device_attach(struct bcache_device *d, struct cache_set *c,
 				 unsigned id)
 {
-	BUG_ON(test_bit(CACHE_SET_STOPPING, &c->flags));
-
 	d->id = id;
 	d->c = c;
 	c->devices[id] = d;
@@ -1771,6 +1769,7 @@ found:
 		pr_debug("set version = %llu", c->sb.version);
 	}
 
+	kobject_get(&ca->kobj);
 	ca->set = c;
 	ca->set->cache[ca->sb.nr_this_dev] = ca;
 	c->cache_by_alloc[c->caches_loaded++] = ca;
@@ -1888,10 +1887,12 @@ static void register_cache(struct cache_sb *sb, struct page *sb_page,
 		goto err;
 
 	pr_info("registered cache device %s", bdevname(bdev, name));
+out:
+	kobject_put(&ca->kobj);
 	return;
 err:
 	pr_notice("error opening %s: %s", bdevname(bdev, name), err);
-	kobject_put(&ca->kobj);
+	goto out;
 }
 
 /* Global interfaces/init */
-- 
cgit v0.10.2


From bf0c55c986540483c34ca640f2eef4c3314388b1 Mon Sep 17 00:00:00 2001
From: Slava Pestov <sp@daterainc.com>
Date: Fri, 11 Jul 2014 12:17:41 -0700
Subject: bcache: fix crash with incomplete cache set

Change-Id: I6abde52afe917633480caaf4e2518f42a816d886

diff --git a/drivers/md/bcache/bcache.h b/drivers/md/bcache/bcache.h
index d2ebcf3..04f7bc2 100644
--- a/drivers/md/bcache/bcache.h
+++ b/drivers/md/bcache/bcache.h
@@ -477,9 +477,13 @@ struct gc_stat {
  * CACHE_SET_STOPPING always gets set first when we're closing down a cache set;
  * we'll continue to run normally for awhile with CACHE_SET_STOPPING set (i.e.
  * flushing dirty data).
+ *
+ * CACHE_SET_RUNNING means all cache devices have been registered and journal
+ * replay is complete.
  */
 #define CACHE_SET_UNREGISTERING		0
 #define	CACHE_SET_STOPPING		1
+#define	CACHE_SET_RUNNING		2
 
 struct cache_set {
 	struct closure		cl;
diff --git a/drivers/md/bcache/super.c b/drivers/md/bcache/super.c
index 29dd1e8..72fbaf7 100644
--- a/drivers/md/bcache/super.c
+++ b/drivers/md/bcache/super.c
@@ -1284,6 +1284,9 @@ int bch_flash_dev_create(struct cache_set *c, uint64_t size)
 	if (test_bit(CACHE_SET_STOPPING, &c->flags))
 		return -EINTR;
 
+	if (!test_bit(CACHE_SET_RUNNING, &c->flags))
+		return -EPERM;
+
 	u = uuid_find_empty(c);
 	if (!u) {
 		pr_err("Can't create volume, no room for UUID");
@@ -1706,6 +1709,7 @@ static void run_cache_set(struct cache_set *c)
 
 	flash_devs_run(c);
 
+	set_bit(CACHE_SET_RUNNING, &c->flags);
 	return;
 err:
 	closure_sync(&cl);
-- 
cgit v0.10.2


From c9a78332b42cbdcdd386a95192a716b67d1711a4 Mon Sep 17 00:00:00 2001
From: Slava Pestov <sp@daterainc.com>
Date: Thu, 19 Jun 2014 15:05:59 -0700
Subject: bcache: fix memory corruption in init error path

If register_cache_set() failed, we would touch ca->set after
it had already been freed. Also, fix an assertion to catch
this.

Change-Id: I748e5f5b223e2d9b2602075dec2f997cced2394d

diff --git a/drivers/md/bcache/super.c b/drivers/md/bcache/super.c
index 72fbaf7..12ad381 100644
--- a/drivers/md/bcache/super.c
+++ b/drivers/md/bcache/super.c
@@ -1356,8 +1356,11 @@ static void cache_set_free(struct closure *cl)
 	bch_journal_free(c);
 
 	for_each_cache(ca, c, i)
-		if (ca)
+		if (ca) {
+			ca->set = NULL;
+			c->cache[ca->sb.nr_this_dev] = NULL;
 			kobject_put(&ca->kobj);
+		}
 
 	bch_bset_sort_state_free(&c->sort);
 	free_pages((unsigned long) c->uuids, ilog2(bucket_pages(c)));
@@ -1794,8 +1797,10 @@ void bch_cache_release(struct kobject *kobj)
 	struct cache *ca = container_of(kobj, struct cache, kobj);
 	unsigned i;
 
-	if (ca->set)
+	if (ca->set) {
+		BUG_ON(ca->set->cache[ca->sb.nr_this_dev] != ca);
 		ca->set->cache[ca->sb.nr_this_dev] = NULL;
+	}
 
 	bio_split_pool_free(&ca->bio_split_hook);
 
@@ -1858,7 +1863,7 @@ static int cache_alloc(struct cache_sb *sb, struct cache *ca)
 }
 
 static void register_cache(struct cache_sb *sb, struct page *sb_page,
-				  struct block_device *bdev, struct cache *ca)
+				struct block_device *bdev, struct cache *ca)
 {
 	char name[BDEVNAME_SIZE];
 	const char *err = "cannot allocate memory";
-- 
cgit v0.10.2


From 2452cc89063a2a6890368f185c4b6d7d8802179e Mon Sep 17 00:00:00 2001
From: Slava Pestov <sp@daterainc.com>
Date: Sat, 12 Jul 2014 00:22:53 -0700
Subject: bcache: try to set b->parent properly

bcache_flash_dev.ktest would reliably crash with 8k and 16k bucket size
before; now it passes.

Change-Id: Ib542232235e39298c3a7548fe52b645cabb823d1

diff --git a/drivers/md/bcache/btree.c b/drivers/md/bcache/btree.c
index 776583f..00cde40 100644
--- a/drivers/md/bcache/btree.c
+++ b/drivers/md/bcache/btree.c
@@ -117,9 +117,9 @@
 ({									\
 	int _r, l = (b)->level - 1;					\
 	bool _w = l <= (op)->lock;					\
-	struct btree *_child = bch_btree_node_get((b)->c, op, key, l, _w);\
+	struct btree *_child = bch_btree_node_get((b)->c, op, key, l,	\
+						  _w, b);		\
 	if (!IS_ERR(_child)) {						\
-		_child->parent = (b);					\
 		_r = bch_btree_ ## fn(_child, op, ##__VA_ARGS__);	\
 		rw_unlock(_w, _child);					\
 	} else								\
@@ -142,7 +142,6 @@
 		rw_lock(_w, _b, _b->level);				\
 		if (_b == (c)->root &&					\
 		    _w == insert_lock(op, _b)) {			\
-			_b->parent = NULL;				\
 			_r = bch_btree_ ## fn(_b, op, ##__VA_ARGS__);	\
 		}							\
 		rw_unlock(_w, _b);					\
@@ -967,7 +966,8 @@ err:
  * level and op->lock.
  */
 struct btree *bch_btree_node_get(struct cache_set *c, struct btree_op *op,
-				 struct bkey *k, int level, bool write)
+				 struct bkey *k, int level, bool write,
+				 struct btree *parent)
 {
 	int i = 0;
 	struct btree *b;
@@ -1002,6 +1002,7 @@ retry:
 		BUG_ON(b->level != level);
 	}
 
+	b->parent = parent;
 	b->accessed = 1;
 
 	for (; i <= b->keys.nsets && b->keys.set[i].size; i++) {
@@ -1022,15 +1023,16 @@ retry:
 	return b;
 }
 
-static void btree_node_prefetch(struct cache_set *c, struct bkey *k, int level)
+static void btree_node_prefetch(struct btree *parent, struct bkey *k)
 {
 	struct btree *b;
 
-	mutex_lock(&c->bucket_lock);
-	b = mca_alloc(c, NULL, k, level);
-	mutex_unlock(&c->bucket_lock);
+	mutex_lock(&parent->c->bucket_lock);
+	b = mca_alloc(parent->c, NULL, k, parent->level - 1);
+	mutex_unlock(&parent->c->bucket_lock);
 
 	if (!IS_ERR_OR_NULL(b)) {
+		b->parent = parent;
 		bch_btree_node_read(b);
 		rw_unlock(true, b);
 	}
@@ -1061,7 +1063,8 @@ static void btree_node_free(struct btree *b)
 }
 
 struct btree *__bch_btree_node_alloc(struct cache_set *c, struct btree_op *op,
-				     int level, bool wait)
+				     int level, bool wait,
+				     struct btree *parent)
 {
 	BKEY_PADDED(key) k;
 	struct btree *b = ERR_PTR(-EAGAIN);
@@ -1085,6 +1088,7 @@ retry:
 	}
 
 	b->accessed = 1;
+	b->parent = parent;
 	bch_bset_init_next(&b->keys, b->keys.set->data, bset_magic(&b->c->sb));
 
 	mutex_unlock(&c->bucket_lock);
@@ -1101,15 +1105,16 @@ err:
 }
 
 static struct btree *bch_btree_node_alloc(struct cache_set *c,
-					  struct btree_op *op, int level)
+					  struct btree_op *op, int level,
+					  struct btree *parent)
 {
-	return __bch_btree_node_alloc(c, op, level, op != NULL);
+	return __bch_btree_node_alloc(c, op, level, op != NULL, parent);
 }
 
 static struct btree *btree_node_alloc_replacement(struct btree *b,
 						  struct btree_op *op)
 {
-	struct btree *n = bch_btree_node_alloc(b->c, op, b->level);
+	struct btree *n = bch_btree_node_alloc(b->c, op, b->level, b->parent);
 	if (!IS_ERR_OR_NULL(n)) {
 		mutex_lock(&n->write_lock);
 		bch_btree_sort_into(&b->keys, &n->keys, &b->c->sort);
@@ -1523,7 +1528,7 @@ static int btree_gc_recurse(struct btree *b, struct btree_op *op,
 		k = bch_btree_iter_next_filter(&iter, &b->keys, bch_ptr_bad);
 		if (k) {
 			r->b = bch_btree_node_get(b->c, op, k, b->level - 1,
-						  true);
+						  true, b);
 			if (IS_ERR(r->b)) {
 				ret = PTR_ERR(r->b);
 				break;
@@ -1818,7 +1823,7 @@ static int bch_btree_check_recurse(struct btree *b, struct btree_op *op)
 			k = bch_btree_iter_next_filter(&iter, &b->keys,
 						       bch_ptr_bad);
 			if (k)
-				btree_node_prefetch(b->c, k, b->level - 1);
+				btree_node_prefetch(b, k);
 
 			if (p)
 				ret = btree(check_recurse, p, b, op);
@@ -1983,12 +1988,12 @@ static int btree_split(struct btree *b, struct btree_op *op,
 
 		trace_bcache_btree_node_split(b, btree_bset_first(n1)->keys);
 
-		n2 = bch_btree_node_alloc(b->c, op, b->level);
+		n2 = bch_btree_node_alloc(b->c, op, b->level, b->parent);
 		if (IS_ERR(n2))
 			goto err_free1;
 
 		if (!b->parent) {
-			n3 = bch_btree_node_alloc(b->c, op, b->level + 1);
+			n3 = bch_btree_node_alloc(b->c, op, b->level + 1, NULL);
 			if (IS_ERR(n3))
 				goto err_free2;
 		}
diff --git a/drivers/md/bcache/btree.h b/drivers/md/bcache/btree.h
index 0044182..5c391fa 100644
--- a/drivers/md/bcache/btree.h
+++ b/drivers/md/bcache/btree.h
@@ -243,9 +243,9 @@ void bch_btree_node_write(struct btree *, struct closure *);
 
 void bch_btree_set_root(struct btree *);
 struct btree *__bch_btree_node_alloc(struct cache_set *, struct btree_op *,
-				     int, bool);
+				     int, bool, struct btree *);
 struct btree *bch_btree_node_get(struct cache_set *, struct btree_op *,
-				 struct bkey *, int, bool);
+				 struct bkey *, int, bool, struct btree *);
 
 int bch_btree_insert_check_key(struct btree *, struct btree_op *,
 			       struct bkey *);
diff --git a/drivers/md/bcache/super.c b/drivers/md/bcache/super.c
index 12ad381..b6114d6 100644
--- a/drivers/md/bcache/super.c
+++ b/drivers/md/bcache/super.c
@@ -1601,7 +1601,7 @@ static void run_cache_set(struct cache_set *c)
 			goto err;
 
 		err = "error reading btree root";
-		c->root = bch_btree_node_get(c, NULL, k, j->btree_level, true);
+		c->root = bch_btree_node_get(c, NULL, k, j->btree_level, true, NULL);
 		if (IS_ERR_OR_NULL(c->root))
 			goto err;
 
@@ -1676,7 +1676,7 @@ static void run_cache_set(struct cache_set *c)
 			goto err;
 
 		err = "cannot allocate new btree root";
-		c->root = __bch_btree_node_alloc(c, NULL, 0, true);
+		c->root = __bch_btree_node_alloc(c, NULL, 0, true, NULL);
 		if (IS_ERR_OR_NULL(c->root))
 			goto err;
 
-- 
cgit v0.10.2


From 5b25abade29616d42d60f9bd5e6a5ad07f7314e3 Mon Sep 17 00:00:00 2001
From: Surbhi Palande <sap@daterainc.com>
Date: Thu, 17 Apr 2014 12:07:04 -0700
Subject: bcache: Correct printing of btree_gc_max_duration_ms

time_stats::btree_gc_max_duration_mc is not bit shifted by 8

Fixes BUG #138

Change-Id: I44fc6e1d0579674016acc533f1a546b080e5371a
Signed-off-by: Surbhi Palande <sap@daterainc.com>

diff --git a/drivers/md/bcache/util.h b/drivers/md/bcache/util.h
index ac7d0d1..98df757 100644
--- a/drivers/md/bcache/util.h
+++ b/drivers/md/bcache/util.h
@@ -416,8 +416,8 @@ do {									\
 			  average_frequency,	frequency_units);	\
 	__print_time_stat(stats, name,					\
 			  average_duration,	duration_units);	\
-	__print_time_stat(stats, name,					\
-			  max_duration,		duration_units);	\
+	sysfs_print(name ## _ ##max_duration ## _ ## duration_units,	\
+			div_u64((stats)->max_duration, NSEC_PER_ ## duration_units));\
 									\
 	sysfs_print(name ## _last_ ## frequency_units, (stats)->last	\
 		    ? div_s64(local_clock() - (stats)->last,		\
-- 
cgit v0.10.2


From 789d21dbd9d8889e62c79ec19585fcc97e42ef07 Mon Sep 17 00:00:00 2001
From: Jianjian Huo <samuel.huo@gmail.com>
Date: Sun, 13 Jul 2014 09:08:59 -0700
Subject: bcache: add mutex lock for bch_is_open

Since bch_is_open will iterate linked list bch_cache_sets and
uncached_devices, it needs bch_register_lock.

Signed-off-by: Jianjian Huo <samuel.huo@gmail.com>

diff --git a/drivers/md/bcache/super.c b/drivers/md/bcache/super.c
index b6114d6..60e7513 100644
--- a/drivers/md/bcache/super.c
+++ b/drivers/md/bcache/super.c
@@ -1966,10 +1966,12 @@ static ssize_t register_bcache(struct kobject *k, struct kobj_attribute *attr,
 	if (IS_ERR(bdev)) {
 		if (bdev == ERR_PTR(-EBUSY)) {
 			bdev = lookup_bdev(strim(path));
+			mutex_lock(&bch_register_lock);
 			if (!IS_ERR(bdev) && bch_is_open(bdev))
 				err = "device already registered";
 			else
 				err = "device busy";
+			mutex_unlock(&bch_register_lock);
 		}
 		goto err;
 	}
-- 
cgit v0.10.2


From 0781c8748cf1ea2b0dcd966571103909528c4efa Mon Sep 17 00:00:00 2001
From: Kent Overstreet <kmo@daterainc.com>
Date: Mon, 7 Jul 2014 13:03:36 -0700
Subject: bcache: Drop unneeded blk_sync_queue() calls

this is needed for the queue/block device we created (it's done by
blk_cleanup_queue() which we do call) - but calling it for the block devices we
only opened is pointless.

Change-Id: I53dfded14ed15b9581d10ca8399d5e1b3abbf9f2

diff --git a/drivers/md/bcache/super.c b/drivers/md/bcache/super.c
index 60e7513..d4713d0 100644
--- a/drivers/md/bcache/super.c
+++ b/drivers/md/bcache/super.c
@@ -1084,12 +1084,8 @@ static void cached_dev_free(struct closure *cl)
 
 	mutex_unlock(&bch_register_lock);
 
-	if (!IS_ERR_OR_NULL(dc->bdev)) {
-		if (dc->bdev->bd_disk)
-			blk_sync_queue(bdev_get_queue(dc->bdev));
-
+	if (!IS_ERR_OR_NULL(dc->bdev))
 		blkdev_put(dc->bdev, FMODE_READ|FMODE_WRITE|FMODE_EXCL);
-	}
 
 	wake_up(&unregister_wait);
 
@@ -1817,10 +1813,8 @@ void bch_cache_release(struct kobject *kobj)
 	if (ca->sb_bio.bi_inline_vecs[0].bv_page)
 		put_page(ca->sb_bio.bi_io_vec[0].bv_page);
 
-	if (!IS_ERR_OR_NULL(ca->bdev)) {
-		blk_sync_queue(bdev_get_queue(ca->bdev));
+	if (!IS_ERR_OR_NULL(ca->bdev))
 		blkdev_put(ca->bdev, FMODE_READ|FMODE_WRITE|FMODE_EXCL);
-	}
 
 	kfree(ca);
 	module_put(THIS_MODULE);
-- 
cgit v0.10.2