Merge branch 'for-linus' of master.kernel.org:/home/rmk/linux-2.6-arm

[GitHub/mt8127/android_kernel_alcatel_ttab.git] / drivers / block / drbd / drbd_receiver.c

diff --git a/drivers/block/drbd/drbd_receiver.c b/drivers/block/drbd/drbd_receiver.c
index a56b107e01eb48364e2ed3e47832d172c3d68ba0..fe1564c7d8b6e835643939d18f3aba2bc9e6b60d 100644 (file)
--- a/drivers/block/drbd/drbd_receiver.c
+++ b/drivers/block/drbd/drbd_receiver.c
@@ -912,6 +912,7 @@ retry:
        drbd_send_state(mdev);
        clear_bit(USE_DEGR_WFC_T, &mdev->flags);
        clear_bit(RESIZE_PENDING, &mdev->flags);
+       mod_timer(&mdev->request_timer, jiffies + HZ); /* just start it here. */
 
        return 1;
 
@@ -1073,6 +1074,16 @@ void drbd_bump_write_ordering(struct drbd_conf *mdev, enum write_ordering_e wo)
  * @mdev:      DRBD device.
  * @e:         epoch entry
  * @rw:                flag field, see bio->bi_rw
+ *
+ * May spread the pages to multiple bios,
+ * depending on bio_add_page restrictions.
+ *
+ * Returns 0 if all bios have been submitted,
+ * -ENOMEM if we could not allocate enough bios,
+ * -ENOSPC (any better suggestion?) if we have not been able to bio_add_page a
+ *  single page to an empty bio (which should never happen and likely indicates
+ *  that the lower level IO stack is in some way broken). This has been observed
+ *  on certain Xen deployments.
  */
 /* TODO allocate from our own bio_set. */
 int drbd_submit_ee(struct drbd_conf *mdev, struct drbd_epoch_entry *e,
@@ -1085,6 +1096,7 @@ int drbd_submit_ee(struct drbd_conf *mdev, struct drbd_epoch_entry *e,
        unsigned ds = e->size;
        unsigned n_bios = 0;
        unsigned nr_pages = (ds + PAGE_SIZE -1) >> PAGE_SHIFT;
+       int err = -ENOMEM;
 
        /* In most cases, we will only need one bio.  But in case the lower
         * level restrictions happen to be different at this offset on this
@@ -1110,8 +1122,17 @@ next_bio:
        page_chain_for_each(page) {
                unsigned len = min_t(unsigned, ds, PAGE_SIZE);
                if (!bio_add_page(bio, page, len, 0)) {
-                       /* a single page must always be possible! */
-                       BUG_ON(bio->bi_vcnt == 0);
+                       /* A single page must always be possible!
+                        * But in case it fails anyways,
+                        * we deal with it, and complain (below). */
+                       if (bio->bi_vcnt == 0) {
+                               dev_err(DEV,
+                                       "bio_add_page failed for len=%u, "
+                                       "bi_vcnt=0 (bi_sector=%llu)\n",
+                                       len, (unsigned long long)bio->bi_sector);
+                               err = -ENOSPC;
+                               goto fail;
+                       }
                        goto next_bio;
                }
                ds -= len;
@@ -1137,7 +1158,7 @@ fail:
                bios = bios->bi_next;
                bio_put(bio);
        }
-       return -ENOMEM;
+       return err;
 }
 
 static int receive_Barrier(struct drbd_conf *mdev, enum drbd_packets cmd, unsigned int data_size)
@@ -1247,7 +1268,8 @@ read_in_block(struct drbd_conf *mdev, u64 id, sector_t sector, int data_size) __
        /* even though we trust out peer,
         * we sometimes have to double check. */
        if (sector + (data_size>>9) > capacity) {
-               dev_err(DEV, "capacity: %llus < sector: %llus + size: %u\n",
+               dev_err(DEV, "request from peer beyond end of local disk: "
+                       "capacity: %llus < sector: %llus + size: %u\n",
                        (unsigned long long)capacity,
                        (unsigned long long)sector, data_size);
                return NULL;
@@ -1436,9 +1458,8 @@ static int recv_resync_read(struct drbd_conf *mdev, sector_t sector, int data_si
        if (drbd_submit_ee(mdev, e, WRITE, DRBD_FAULT_RS_WR) == 0)
                return true;
 
-       /* drbd_submit_ee currently fails for one reason only:
-        * not being able to allocate enough bios.
-        * Is dropping the connection going to help? */
+       /* don't care for the reason here */
+       dev_err(DEV, "submit failed, triggering re-connect\n");
        spin_lock_irq(&mdev->req_lock);
        list_del(&e->w.list);
        spin_unlock_irq(&mdev->req_lock);
@@ -1837,9 +1858,8 @@ static int receive_Data(struct drbd_conf *mdev, enum drbd_packets cmd, unsigned
        if (drbd_submit_ee(mdev, e, rw, DRBD_FAULT_DT_WR) == 0)
                return true;
 
-       /* drbd_submit_ee currently fails for one reason only:
-        * not being able to allocate enough bios.
-        * Is dropping the connection going to help? */
+       /* don't care for the reason here */
+       dev_err(DEV, "submit failed, triggering re-connect\n");
        spin_lock_irq(&mdev->req_lock);
        list_del(&e->w.list);
        hlist_del_init(&e->colision);
@@ -1848,9 +1868,7 @@ static int receive_Data(struct drbd_conf *mdev, enum drbd_packets cmd, unsigned
                drbd_al_complete_io(mdev, e->sector);
 
 out_interrupted:
-       /* yes, the epoch_size now is imbalanced.
-        * but we drop the connection anyways, so we don't have a chance to
-        * receive a barrier... atomic_inc(&mdev->epoch_size); */
+       drbd_may_finish_epoch(mdev, e->epoch, EV_PUT + EV_CLEANUP);
        put_ldev(mdev);
        drbd_free_ee(mdev, e);
        return false;
@@ -2096,9 +2114,8 @@ submit:
        if (drbd_submit_ee(mdev, e, READ, fault_type) == 0)
                return true;
 
-       /* drbd_submit_ee currently fails for one reason only:
-        * not being able to allocate enough bios.
-        * Is dropping the connection going to help? */
+       /* don't care for the reason here */
+       dev_err(DEV, "submit failed, triggering re-connect\n");
        spin_lock_irq(&mdev->req_lock);
        list_del(&e->w.list);
        spin_unlock_irq(&mdev->req_lock);
@@ -3806,6 +3823,8 @@ static void drbd_disconnect(struct drbd_conf *mdev)
        atomic_set(&mdev->rs_pending_cnt, 0);
        wake_up(&mdev->misc_wait);
 
+       del_timer(&mdev->request_timer);
+
        /* make sure syncer is stopped and w_resume_next_sg queued */
        del_timer_sync(&mdev->resync_timer);
        resync_timer_fn((unsigned long)mdev);