Merge branch 'for-linus' of git://git.kernel.dk/linux-block

* 'for-linus' of git://git.kernel.dk/linux-block:
  drbd: we should write meta data updates with FLUSH FUA
  drbd: fix limit define, we support 1 PiByte now
  drbd: when receive times out on meta socket, also check last receive time on data socket
  drbd: account bitmap IO during resync as resync-(related-)-io
  drbd: don't cond_resched_lock with IRQs disabled
  drbd: add missing spinlock to bitmap receive
  drbd: Use the correct max_bio_size when creating resync requests
  cfq-iosched: make code consistent
  cfq-iosched: fix a rcu warning

6 files changed, 39 insertions, 25 deletions

diff --git a/block/cfq-iosched.c b/block/cfq-iosched.c
index f3799432676..ae21919f15e 100644
--- a/block/cfq-iosched.c
+++ b/block/cfq-iosched.c
@@ -2773,11 +2773,14 @@ static void __cfq_exit_single_io_context(struct cfq_data *cfqd,
 	smp_wmb();
 	cic->key = cfqd_dead_key(cfqd);
 
+	rcu_read_lock();
 	if (rcu_dereference(ioc->ioc_data) == cic) {
+		rcu_read_unlock();
 		spin_lock(&ioc->lock);
 		rcu_assign_pointer(ioc->ioc_data, NULL);
 		spin_unlock(&ioc->lock);
-	}
+	} else
+		rcu_read_unlock();
 
 	if (cic->cfqq[BLK_RW_ASYNC]) {
 		cfq_exit_cfqq(cfqd, cic->cfqq[BLK_RW_ASYNC]);
@@ -3084,7 +3087,8 @@ cfq_drop_dead_cic(struct cfq_data *cfqd, struct io_context *ioc,
 
 	spin_lock_irqsave(&ioc->lock, flags);
 
-	BUG_ON(ioc->ioc_data == cic);
+	BUG_ON(rcu_dereference_check(ioc->ioc_data,
+		lockdep_is_held(&ioc->lock)) == cic);
 
 	radix_tree_delete(&ioc->radix_root, cfqd->cic_index);
 	hlist_del_rcu(&cic->cic_list);
diff --git a/drivers/block/drbd/drbd_actlog.c b/drivers/block/drbd/drbd_actlog.c
index 09ef9a878ef..cf0e63dd97d 100644
--- a/drivers/block/drbd/drbd_actlog.c
+++ b/drivers/block/drbd/drbd_actlog.c
@@ -79,7 +79,7 @@ static int _drbd_md_sync_page_io(struct drbd_conf *mdev,
 	md_io.error = 0;
 
 	if ((rw & WRITE) && !test_bit(MD_NO_FUA, &mdev->flags))
-		rw |= REQ_FUA;
+		rw |= REQ_FUA | REQ_FLUSH;
 	rw |= REQ_SYNC;
 
 	bio = bio_alloc(GFP_NOIO, 1);
diff --git a/drivers/block/drbd/drbd_bitmap.c b/drivers/block/drbd/drbd_bitmap.c
index f440a02dfdb..7b976296b56 100644
--- a/drivers/block/drbd/drbd_bitmap.c
+++ b/drivers/block/drbd/drbd_bitmap.c
@@ -112,9 +112,6 @@ struct drbd_bitmap {
 	struct task_struct *bm_task;
 };
 
-static int __bm_change_bits_to(struct drbd_conf *mdev, const unsigned long s,
-			       unsigned long e, int val, const enum km_type km);
-
 #define bm_print_lock_info(m) __bm_print_lock_info(m, __func__)
 static void __bm_print_lock_info(struct drbd_conf *mdev, const char *func)
 {
@@ -994,6 +991,9 @@ static void bm_page_io_async(struct bm_aio_ctx *ctx, int page_nr, int rw) __must
 		bio_endio(bio, -EIO);
 	} else {
 		submit_bio(rw, bio);
+		/* this should not count as user activity and cause the
+		 * resync to throttle -- see drbd_rs_should_slow_down(). */
+		atomic_add(len >> 9, &mdev->rs_sect_ev);
 	}
 }
 
@@ -1256,7 +1256,7 @@ unsigned long _drbd_bm_find_next_zero(struct drbd_conf *mdev, unsigned long bm_f
  * expected to be called for only a few bits (e - s about BITS_PER_LONG).
  * Must hold bitmap lock already. */
 static int __bm_change_bits_to(struct drbd_conf *mdev, const unsigned long s,
-	unsigned long e, int val, const enum km_type km)
+	unsigned long e, int val)
 {
 	struct drbd_bitmap *b = mdev->bitmap;
 	unsigned long *p_addr = NULL;
@@ -1274,14 +1274,14 @@ static int __bm_change_bits_to(struct drbd_conf *mdev, const unsigned long s,
 		unsigned int page_nr = bm_bit_to_page_idx(b, bitnr);
 		if (page_nr != last_page_nr) {
 			if (p_addr)
-				__bm_unmap(p_addr, km);
+				__bm_unmap(p_addr, KM_IRQ1);
 			if (c < 0)
 				bm_set_page_lazy_writeout(b->bm_pages[last_page_nr]);
 			else if (c > 0)
 				bm_set_page_need_writeout(b->bm_pages[last_page_nr]);
 			changed_total += c;
 			c = 0;
-			p_addr = __bm_map_pidx(b, page_nr, km);
+			p_addr = __bm_map_pidx(b, page_nr, KM_IRQ1);
 			last_page_nr = page_nr;
 		}
 		if (val)
@@ -1290,7 +1290,7 @@ static int __bm_change_bits_to(struct drbd_conf *mdev, const unsigned long s,
 			c -= (0 != __test_and_clear_bit_le(bitnr & BITS_PER_PAGE_MASK, p_addr));
 	}
 	if (p_addr)
-		__bm_unmap(p_addr, km);
+		__bm_unmap(p_addr, KM_IRQ1);
 	if (c < 0)
 		bm_set_page_lazy_writeout(b->bm_pages[last_page_nr]);
 	else if (c > 0)
@@ -1318,7 +1318,7 @@ static int bm_change_bits_to(struct drbd_conf *mdev, const unsigned long s,
 	if ((val ? BM_DONT_SET : BM_DONT_CLEAR) & b->bm_flags)
 		bm_print_lock_info(mdev);
 
-	c = __bm_change_bits_to(mdev, s, e, val, KM_IRQ1);
+	c = __bm_change_bits_to(mdev, s, e, val);
 
 	spin_unlock_irqrestore(&b->bm_lock, flags);
 	return c;
@@ -1343,16 +1343,17 @@ static inline void bm_set_full_words_within_one_page(struct drbd_bitmap *b,
 {
 	int i;
 	int bits;
-	unsigned long *paddr = kmap_atomic(b->bm_pages[page_nr], KM_USER0);
+	unsigned long *paddr = kmap_atomic(b->bm_pages[page_nr], KM_IRQ1);
 	for (i = first_word; i < last_word; i++) {
 		bits = hweight_long(paddr[i]);
 		paddr[i] = ~0UL;
 		b->bm_set += BITS_PER_LONG - bits;
 	}
-	kunmap_atomic(paddr, KM_USER0);
+	kunmap_atomic(paddr, KM_IRQ1);
 }
 
-/* Same thing as drbd_bm_set_bits, but without taking the spin_lock_irqsave.
+/* Same thing as drbd_bm_set_bits,
+ * but more efficient for a large bit range.
  * You must first drbd_bm_lock().
  * Can be called to set the whole bitmap in one go.
  * Sets bits from s to e _inclusive_. */
@@ -1366,6 +1367,7 @@ void _drbd_bm_set_bits(struct drbd_conf *mdev, const unsigned long s, const unsi
 	 * Do not use memset, because we must account for changes,
 	 * so we need to loop over the words with hweight() anyways.
 	 */
+	struct drbd_bitmap *b = mdev->bitmap;
 	unsigned long sl = ALIGN(s,BITS_PER_LONG);
 	unsigned long el = (e+1) & ~((unsigned long)BITS_PER_LONG-1);
 	int first_page;
@@ -1376,15 +1378,19 @@ void _drbd_bm_set_bits(struct drbd_conf *mdev, const unsigned long s, const unsi
 
 	if (e - s <= 3*BITS_PER_LONG) {
 		/* don't bother; el and sl may even be wrong. */
-		__bm_change_bits_to(mdev, s, e, 1, KM_USER0);
+		spin_lock_irq(&b->bm_lock);
+		__bm_change_bits_to(mdev, s, e, 1);
+		spin_unlock_irq(&b->bm_lock);
 		return;
 	}
 
 	/* difference is large enough that we can trust sl and el */
 
+	spin_lock_irq(&b->bm_lock);
+
 	/* bits filling the current long */
 	if (sl)
-		__bm_change_bits_to(mdev, s, sl-1, 1, KM_USER0);
+		__bm_change_bits_to(mdev, s, sl-1, 1);
 
 	first_page = sl >> (3 + PAGE_SHIFT);
 	last_page = el >> (3 + PAGE_SHIFT);
@@ -1397,8 +1403,10 @@ void _drbd_bm_set_bits(struct drbd_conf *mdev, const unsigned long s, const unsi
 	/* first and full pages, unless first page == last page */
 	for (page_nr = first_page; page_nr < last_page; page_nr++) {
 		bm_set_full_words_within_one_page(mdev->bitmap, page_nr, first_word, last_word);
+		spin_unlock_irq(&b->bm_lock);
 		cond_resched();
 		first_word = 0;
+		spin_lock_irq(&b->bm_lock);
 	}
 
 	/* last page (respectively only page, for first page == last page) */
@@ -1411,7 +1419,8 @@ void _drbd_bm_set_bits(struct drbd_conf *mdev, const unsigned long s, const unsi
 	 * it would trigger an assert in __bm_change_bits_to()
 	 */
 	if (el <= e)
-		__bm_change_bits_to(mdev, el, e, 1, KM_USER0);
+		__bm_change_bits_to(mdev, el, e, 1);
+	spin_unlock_irq(&b->bm_lock);
 }
 
 /* returns bit state
diff --git a/drivers/block/drbd/drbd_receiver.c b/drivers/block/drbd/drbd_receiver.c
index 25d32c5aa50..43beaca5317 100644
--- a/drivers/block/drbd/drbd_receiver.c
+++ b/drivers/block/drbd/drbd_receiver.c
@@ -4602,6 +4602,11 @@ int drbd_asender(struct drbd_thread *thi)
 			dev_err(DEV, "meta connection shut down by peer.\n");
 			goto reconnect;
 		} else if (rv == -EAGAIN) {
+			/* If the data socket received something meanwhile,
+			 * that is good enough: peer is still alive. */
+			if (time_after(mdev->last_received,
+				jiffies - mdev->meta.socket->sk->sk_rcvtimeo))
+				continue;
 			if (ping_timeout_active) {
 				dev_err(DEV, "PingAck did not arrive in time.\n");
 				goto reconnect;
@@ -4637,6 +4642,7 @@ int drbd_asender(struct drbd_thread *thi)
 				goto reconnect;
 		}
 		if (received == expect) {
+			mdev->last_received = jiffies;
 			D_ASSERT(cmd != NULL);
 			if (!cmd->process(mdev, h))
 				goto reconnect;
diff --git a/drivers/block/drbd/drbd_worker.c b/drivers/block/drbd/drbd_worker.c
index 4d76b06b6b2..4d3e6f6213b 100644
--- a/drivers/block/drbd/drbd_worker.c
+++ b/drivers/block/drbd/drbd_worker.c
@@ -536,12 +536,7 @@ static int w_make_resync_request(struct drbd_conf *mdev,
 		return 1;
 	}
 
-	/* starting with drbd 8.3.8, we can handle multi-bio EEs,
-	 * if it should be necessary */
-	max_bio_size =
-		mdev->agreed_pro_version < 94 ? queue_max_hw_sectors(mdev->rq_queue) << 9 :
-		mdev->agreed_pro_version < 95 ?	DRBD_MAX_SIZE_H80_PACKET : DRBD_MAX_BIO_SIZE;
-
+	max_bio_size = queue_max_hw_sectors(mdev->rq_queue) << 9;
 	number = drbd_rs_number_requests(mdev);
 	if (number == 0)
 		goto requeue;
diff --git a/include/linux/drbd_limits.h b/include/linux/drbd_limits.h
index 246f576c981..447c3675238 100644
--- a/include/linux/drbd_limits.h
+++ b/include/linux/drbd_limits.h
@@ -117,10 +117,10 @@
 /* drbdsetup XY resize -d Z
  * you are free to reduce the device size to nothing, if you want to.
  * the upper limit with 64bit kernel, enough ram and flexible meta data
- * is 16 TB, currently. */
+ * is 1 PiB, currently. */
 /* DRBD_MAX_SECTORS */
 #define DRBD_DISK_SIZE_SECT_MIN  0
-#define DRBD_DISK_SIZE_SECT_MAX  (16 * (2LLU << 30))
+#define DRBD_DISK_SIZE_SECT_MAX  (1 * (2LLU << 40))
 #define DRBD_DISK_SIZE_SECT_DEF  0 /* = disabled = no user size... */
 
 #define DRBD_ON_IO_ERROR_DEF EP_PASS_ON