drbd: Track all IO requests on the TL, not writes only

With that the drbd_fail_pending_reads() function becomes obsolete.
Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
Signed-off-by: Lars Ellenberg <lars.ellenberg@linbit.com>

drivers/block/drbd/drbd_main.c

@@ -401,6 +401,8 @@ void tl_clear(struct drbd_conf *mdev)
 	/* ensure bit indicating barrier is required is clear */
 	clear_bit(CREATE_BARRIER, &mdev->flags);
 
+	memset(mdev->app_reads_hash, 0, APP_R_HSIZE*sizeof(void *));
+
 	spin_unlock_irq(&mdev->req_lock);
 }
 

drivers/block/drbd/drbd_receiver.c

@@ -3666,41 +3666,6 @@ static void drbdd(struct drbd_conf *mdev)
 	}
 }
 
-static void drbd_fail_pending_reads(struct drbd_conf *mdev)
-{
-	struct hlist_head *slot;
-	struct hlist_node *pos;
-	struct hlist_node *tmp;
-	struct drbd_request *req;
-	int i;
-
-	/*
-	 * Application READ requests
-	 */
-	spin_lock_irq(&mdev->req_lock);
-	for (i = 0; i < APP_R_HSIZE; i++) {
-		slot = mdev->app_reads_hash+i;
-		hlist_for_each_entry_safe(req, pos, tmp, slot, colision) {
-			/* it may (but should not any longer!)
-			 * be on the work queue; if that assert triggers,
-			 * we need to also grab the
-			 * spin_lock_irq(&mdev->data.work.q_lock);
-			 * and list_del_init here. */
-			D_ASSERT(list_empty(&req->w.list));
-			/* It would be nice to complete outside of spinlock.
-			 * But this is easier for now. */
-			_req_mod(req, connection_lost_while_pending);
-		}
-	}
-	for (i = 0; i < APP_R_HSIZE; i++)
-		if (!hlist_empty(mdev->app_reads_hash+i))
-			dev_warn(DEV, "ASSERT FAILED: app_reads_hash[%d].first: "
-				"%p, should be NULL\n", i, mdev->app_reads_hash[i].first);
-
-	memset(mdev->app_reads_hash, 0, APP_R_HSIZE*sizeof(void *));
-	spin_unlock_irq(&mdev->req_lock);
-}
-
 void drbd_flush_workqueue(struct drbd_conf *mdev)
 {
 	struct drbd_wq_barrier barr;
@@ -3770,8 +3735,6 @@ static void drbd_disconnect(struct drbd_conf *mdev)
 	if (!mdev->state.susp)
 		tl_clear(mdev);
 
-	drbd_fail_pending_reads(mdev);
-
 	dev_info(DEV, "Connection closed\n");
 
 	drbd_md_sync(mdev);

drivers/block/drbd/drbd_req.c

@@ -59,17 +59,19 @@ static void _drbd_end_io_acct(struct drbd_conf *mdev, struct drbd_request *req)
 static void _req_is_done(struct drbd_conf *mdev, struct drbd_request *req, const int rw)
 {
 	const unsigned long s = req->rq_state;
+
+	/* remove it from the transfer log.
+	 * well, only if it had been there in the first
+	 * place... if it had not (local only or conflicting
+	 * and never sent), it should still be "empty" as
+	 * initialized in drbd_req_new(), so we can list_del() it
+	 * here unconditionally */
+	list_del(&req->tl_requests);
+
 	/* if it was a write, we may have to set the corresponding
 	 * bit(s) out-of-sync first. If it had a local part, we need to
 	 * release the reference to the activity log. */
 	if (rw == WRITE) {
-		/* remove it from the transfer log.
-		 * well, only if it had been there in the first
-		 * place... if it had not (local only or conflicting
-		 * and never sent), it should still be "empty" as
-		 * initialized in drbd_req_new(), so we can list_del() it
-		 * here unconditionally */
-		list_del(&req->tl_requests);
 		/* Set out-of-sync unless both OK flags are set
 		 * (local only or remote failed).
 		 * Other places where we set out-of-sync:
@@ -517,8 +519,6 @@ void __req_mod(struct drbd_request *req, enum drbd_req_event what,
 		D_ASSERT(test_bit(CREATE_BARRIER, &mdev->flags) == 0);
 
 		req->epoch = mdev->newest_tle->br_number;
-		list_add_tail(&req->tl_requests,
-				&mdev->newest_tle->requests);
 
 		/* increment size of current epoch */
 		mdev->newest_tle->n_writes++;
@@ -634,6 +634,9 @@ void __req_mod(struct drbd_request *req, enum drbd_req_event what,
 		break;
 
 	case barrier_acked:
+		if (!(req->rq_state & RQ_WRITE))
+			break;
+
 		if (req->rq_state & RQ_NET_PENDING) {
 			/* barrier came in before all requests have been acked.
 			 * this is bad, because if the connection is lost now,
@@ -892,6 +895,9 @@ static int drbd_make_request_common(struct drbd_conf *mdev, struct bio *bio)
 		remote = 0;
 	}
 
+
+	list_add_tail(&req->tl_requests, &mdev->newest_tle->requests);
+
 	/* NOTE remote first: to get the concurrent write detection right,
 	 * we must register the request before start of local IO.  */
 	if (remote) {

drivers/block/drbd/drbd_req.h

@@ -183,6 +183,9 @@ enum drbd_req_state_bits {
 
 	/* keep this last, its for the RQ_NET_MASK */
 	__RQ_NET_MAX,
+
+	/* Set when this is a write, clear for a read */
+	__RQ_WRITE,
 };
 
 #define RQ_LOCAL_PENDING   (1UL << __RQ_LOCAL_PENDING)
@@ -201,6 +204,8 @@ enum drbd_req_state_bits {
 /* 0x1f8 */
 #define RQ_NET_MASK        (((1UL << __RQ_NET_MAX)-1) & ~RQ_LOCAL_MASK)
 
+#define RQ_WRITE           (1UL << __RQ_WRITE)
+
 /* epoch entries */
 static inline
 struct hlist_head *ee_hash_slot(struct drbd_conf *mdev, sector_t sector)
@@ -253,7 +258,7 @@ static inline struct drbd_request *drbd_req_new(struct drbd_conf *mdev,
 	if (likely(req)) {
 		bio = bio_clone(bio_src, GFP_NOIO); /* XXX cannot fail?? */
 
-		req->rq_state    = 0;
+		req->rq_state    = bio_data_dir(bio_src) == WRITE ? RQ_WRITE : 0;
 		req->mdev        = mdev;
 		req->master_bio  = bio_src;
 		req->private_bio = bio;