Merge tag 'v3.3-rc7' into sched/core

Merge reason: merge back final fixes, prepare for the merge window.

Signed-off-by: Ingo Molnar <mingo@elte.hu>

7 files changed, 77 insertions, 32 deletions

diff --git a/drivers/md/dm-flakey.c b/drivers/md/dm-flakey.c
index 9fb18c14782..b280c433e4a 100644
--- a/drivers/md/dm-flakey.c
+++ b/drivers/md/dm-flakey.c
@@ -323,7 +323,7 @@ static int flakey_end_io(struct dm_target *ti, struct bio *bio,
          * Corrupt successful READs while in down state.
          * If flags were specified, only corrupt those that match.
          */
-        if (!error && bio_submitted_while_down &&
+        if (fc->corrupt_bio_byte && !error && bio_submitted_while_down &&
             (bio_data_dir(bio) == READ) && (fc->corrupt_bio_rw == READ) &&
             all_corrupt_bio_flags_match(bio, fc))
                 corrupt_bio_data(bio, fc);
diff --git a/drivers/md/dm-io.c b/drivers/md/dm-io.c
index ad2eba40e31..ea5dd289fe2 100644
--- a/drivers/md/dm-io.c
+++ b/drivers/md/dm-io.c
@@ -296,6 +296,8 @@ static void do_region(int rw, unsigned region, struct dm_io_region *where,
         unsigned offset;
         unsigned num_bvecs;
         sector_t remaining = where->count;
+        struct request_queue *q = bdev_get_queue(where->bdev);
+        sector_t discard_sectors;
 
         /*
          * where->count may be zero if rw holds a flush and we need to
@@ -305,9 +307,12 @@ static void do_region(int rw, unsigned region, struct dm_io_region *where,
                 /*
                  * Allocate a suitably sized-bio.
                  */
-                num_bvecs = dm_sector_div_up(remaining,
-                                             (PAGE_SIZE >> SECTOR_SHIFT));
-                num_bvecs = min_t(int, bio_get_nr_vecs(where->bdev), num_bvecs);
+                if (rw & REQ_DISCARD)
+                        num_bvecs = 1;
+                else
+                        num_bvecs = min_t(int, bio_get_nr_vecs(where->bdev),
+                                          dm_sector_div_up(remaining, (PAGE_SIZE >> SECTOR_SHIFT)));
+
                 bio = bio_alloc_bioset(GFP_NOIO, num_bvecs, io->client->bios);
                 bio->bi_sector = where->sector + (where->count - remaining);
                 bio->bi_bdev = where->bdev;
@@ -315,10 +320,14 @@ static void do_region(int rw, unsigned region, struct dm_io_region *where,
                 bio->bi_destructor = dm_bio_destructor;
                 store_io_and_region_in_bio(bio, io, region);
 
-                /*
-                 * Try and add as many pages as possible.
-                 */
-                while (remaining) {
+                if (rw & REQ_DISCARD) {
+                        discard_sectors = min_t(sector_t, q->limits.max_discard_sectors, remaining);
+                        bio->bi_size = discard_sectors << SECTOR_SHIFT;
+                        remaining -= discard_sectors;
+                } else while (remaining) {
+                        /*
+                         * Try and add as many pages as possible.
+                         */
                         dp->get_page(dp, &page, &len, &offset);
                         len = min(len, to_bytes(remaining));
                         if (!bio_add_page(bio, page, len, offset))
diff --git a/drivers/md/dm-ioctl.c b/drivers/md/dm-ioctl.c
index 31c2dc25886..1ce84ed0b76 100644
--- a/drivers/md/dm-ioctl.c
+++ b/drivers/md/dm-ioctl.c
@@ -1437,7 +1437,7 @@ static int target_message(struct dm_ioctl *param, size_t param_size)
 
         if (!argc) {
                 DMWARN("Empty message received.");
-                goto out;
+                goto out_argv;
         }
 
         table = dm_get_live_table(md);
diff --git a/drivers/md/dm-raid.c b/drivers/md/dm-raid.c
index 86cb7e5d83d..787022c1818 100644
--- a/drivers/md/dm-raid.c
+++ b/drivers/md/dm-raid.c
@@ -668,7 +668,14 @@ static int super_load(struct md_rdev *rdev, struct md_rdev *refdev)
                 return ret;
 
         sb = page_address(rdev->sb_page);
-        if (sb->magic != cpu_to_le32(DM_RAID_MAGIC)) {
+
+        /*
+         * Two cases that we want to write new superblocks and rebuild:
+         * 1) New device (no matching magic number)
+         * 2) Device specified for rebuild (!In_sync w/ offset == 0)
+         */
+        if ((sb->magic != cpu_to_le32(DM_RAID_MAGIC)) ||
+            (!test_bit(In_sync, &rdev->flags) && !rdev->recovery_offset)) {
                 super_sync(rdev->mddev, rdev);
 
                 set_bit(FirstUse, &rdev->flags);
@@ -745,11 +752,8 @@ static int super_init_validation(struct mddev *mddev, struct md_rdev *rdev)
          */
         rdev_for_each(r, t, mddev) {
                 if (!test_bit(In_sync, &r->flags)) {
-                        if (!test_bit(FirstUse, &r->flags))
-                                DMERR("Superblock area of "
-                                      "rebuild device %d should have been "
-                                      "cleared.", r->raid_disk);
-                        set_bit(FirstUse, &r->flags);
+                        DMINFO("Device %d specified for rebuild: "
+                               "Clearing superblock", r->raid_disk);
                         rebuilds++;
                 } else if (test_bit(FirstUse, &r->flags))
                         new_devs++;
@@ -971,6 +975,7 @@ static int raid_ctr(struct dm_target *ti, unsigned argc, char **argv)
 
         INIT_WORK(&rs->md.event_work, do_table_event);
         ti->private = rs;
+        ti->num_flush_requests = 1;
 
         mutex_lock(&rs->md.reconfig_mutex);
         ret = md_run(&rs->md);
diff --git a/drivers/md/dm-thin-metadata.c b/drivers/md/dm-thin-metadata.c
index 59c4f0446ff..237571af77f 100644
--- a/drivers/md/dm-thin-metadata.c
+++ b/drivers/md/dm-thin-metadata.c
@@ -385,6 +385,7 @@ static int init_pmd(struct dm_pool_metadata *pmd,
                 data_sm = dm_sm_disk_create(tm, nr_blocks);
                 if (IS_ERR(data_sm)) {
                         DMERR("sm_disk_create failed");
+                        dm_tm_unlock(tm, sblock);
                         r = PTR_ERR(data_sm);
                         goto bad;
                 }
@@ -789,6 +790,11 @@ int dm_pool_metadata_close(struct dm_pool_metadata *pmd)
         return 0;
 }
 
+/*
+ * __open_device: Returns @td corresponding to device with id @dev,
+ * creating it if @create is set and incrementing @td->open_count.
+ * On failure, @td is undefined.
+ */
 static int __open_device(struct dm_pool_metadata *pmd,
                          dm_thin_id dev, int create,
                          struct dm_thin_device **td)
@@ -799,10 +805,16 @@ static int __open_device(struct dm_pool_metadata *pmd,
         struct disk_device_details details_le;
 
         /*
-         * Check the device isn't already open.
+         * If the device is already open, return it.
          */
         list_for_each_entry(td2, &pmd->thin_devices, list)
                 if (td2->id == dev) {
+                        /*
+                         * May not create an already-open device.
+                         */
+                        if (create)
+                                return -EEXIST;
+
                         td2->open_count++;
                         *td = td2;
                         return 0;
@@ -817,6 +829,9 @@ static int __open_device(struct dm_pool_metadata *pmd,
                 if (r != -ENODATA || !create)
                         return r;
 
+                /*
+                 * Create new device.
+                 */
                 changed = 1;
                 details_le.mapped_blocks = 0;
                 details_le.transaction_id = cpu_to_le64(pmd->trans_id);
@@ -882,12 +897,10 @@ static int __create_thin(struct dm_pool_metadata *pmd,
 
         r = __open_device(pmd, dev, 1, &td);
         if (r) {
-                __close_device(td);
                 dm_btree_remove(&pmd->tl_info, pmd->root, &key, &pmd->root);
                 dm_btree_del(&pmd->bl_info, dev_root);
                 return r;
         }
-        td->changed = 1;
         __close_device(td);
 
         return r;
@@ -967,14 +980,14 @@ static int __create_snap(struct dm_pool_metadata *pmd,
                 goto bad;
 
         r = __set_snapshot_details(pmd, td, origin, pmd->time);
+        __close_device(td);
+
         if (r)
                 goto bad;
 
-        __close_device(td);
         return 0;
 
 bad:
-        __close_device(td);
         dm_btree_remove(&pmd->tl_info, pmd->root, &key, &pmd->root);
         dm_btree_remove(&pmd->details_info, pmd->details_root,
                         &key, &pmd->details_root);
@@ -1211,6 +1224,8 @@ static int __remove(struct dm_thin_device *td, dm_block_t block)
         if (r)
                 return r;
 
+        td->mapped_blocks--;
+        td->changed = 1;
         pmd->need_commit = 1;
 
         return 0;
diff --git a/drivers/md/raid1.c b/drivers/md/raid1.c
index a368db2431a..a0b225eb4ac 100644
--- a/drivers/md/raid1.c
+++ b/drivers/md/raid1.c
@@ -624,7 +624,7 @@ int md_raid1_congested(struct mddev *mddev, int bits)
                 return 1;
 
         rcu_read_lock();
-        for (i = 0; i < conf->raid_disks; i++) {
+        for (i = 0; i < conf->raid_disks * 2; i++) {
                 struct md_rdev *rdev = rcu_dereference(conf->mirrors[i].rdev);
                 if (rdev && !test_bit(Faulty, &rdev->flags)) {
                         struct request_queue *q = bdev_get_queue(rdev->bdev);
diff --git a/drivers/md/raid10.c b/drivers/md/raid10.c
index 6e8aa213f0d..58c44d6453a 100644
--- a/drivers/md/raid10.c
+++ b/drivers/md/raid10.c
@@ -67,6 +67,7 @@ static int max_queued_requests = 1024;
 
 static void allow_barrier(struct r10conf *conf);
 static void lower_barrier(struct r10conf *conf);
+static int enough(struct r10conf *conf, int ignore);
 
 static void * r10bio_pool_alloc(gfp_t gfp_flags, void *data)
 {
@@ -347,6 +348,19 @@ static void raid10_end_read_request(struct bio *bio, int error)
                  * wait for the 'master' bio.
                  */
                 set_bit(R10BIO_Uptodate, &r10_bio->state);
+        } else {
+                /* If all other devices that store this block have
+                 * failed, we want to return the error upwards rather
+                 * than fail the last device.  Here we redefine
+                 * "uptodate" to mean "Don't want to retry"
+                 */
+                unsigned long flags;
+                spin_lock_irqsave(&conf->device_lock, flags);
+                if (!enough(conf, rdev->raid_disk))
+                        uptodate = 1;
+                spin_unlock_irqrestore(&conf->device_lock, flags);
+        }
+        if (uptodate) {
                 raid_end_bio_io(r10_bio);
                 rdev_dec_pending(rdev, conf->mddev);
         } else {
@@ -2052,6 +2066,7 @@ static void fix_read_error(struct r10conf *conf, struct mddev *mddev, struct r10
                        "md/raid10:%s: %s: Failing raid device\n",
                        mdname(mddev), b);
                 md_error(mddev, conf->mirrors[d].rdev);
+                r10_bio->devs[r10_bio->read_slot].bio = IO_BLOCKED;
                 return;
         }
 
@@ -2105,8 +2120,11 @@ static void fix_read_error(struct r10conf *conf, struct mddev *mddev, struct r10
                                     rdev,
                                     r10_bio->devs[r10_bio->read_slot].addr
                                     + sect,
-                                    s, 0))
+                                    s, 0)) {
                                 md_error(mddev, rdev);
+                                r10_bio->devs[r10_bio->read_slot].bio
+                                        = IO_BLOCKED;
+                        }
                         break;
                 }
 
@@ -2299,17 +2317,20 @@ static void handle_read_error(struct mddev *mddev, struct r10bio *r10_bio)
          * This is all done synchronously while the array is
          * frozen.
          */
+        bio = r10_bio->devs[slot].bio;
+        bdevname(bio->bi_bdev, b);
+        bio_put(bio);
+        r10_bio->devs[slot].bio = NULL;
+
         if (mddev->ro == 0) {
                 freeze_array(conf);
                 fix_read_error(conf, mddev, r10_bio);
                 unfreeze_array(conf);
-        }
+        } else
+                r10_bio->devs[slot].bio = IO_BLOCKED;
+
         rdev_dec_pending(rdev, mddev);
 
-        bio = r10_bio->devs[slot].bio;
-        bdevname(bio->bi_bdev, b);
-        r10_bio->devs[slot].bio =
-                mddev->ro ? IO_BLOCKED : NULL;
 read_more:
         rdev = read_balance(conf, r10_bio, &max_sectors);
         if (rdev == NULL) {
@@ -2318,13 +2339,10 @@ read_more:
                        mdname(mddev), b,
                        (unsigned long long)r10_bio->sector);
                 raid_end_bio_io(r10_bio);
-                bio_put(bio);
                 return;
         }
 
         do_sync = (r10_bio->master_bio->bi_rw & REQ_SYNC);
-        if (bio)
-                bio_put(bio);
         slot = r10_bio->read_slot;
         printk_ratelimited(
                 KERN_ERR
@@ -2360,7 +2378,6 @@ read_more:
                         mbio->bi_phys_segments++;
                 spin_unlock_irq(&conf->device_lock);
                 generic_make_request(bio);
-                bio = NULL;
 
                 r10_bio = mempool_alloc(conf->r10bio_pool,
                                         GFP_NOIO);
@@ -3243,7 +3260,6 @@ static int run(struct mddev *mddev)
                         disk->rdev = rdev;
                 }
 
-                disk->rdev = rdev;
                 disk_stack_limits(mddev->gendisk, rdev->bdev,
                                   rdev->data_offset << 9);
                 /* as we don't honour merge_bvec_fn, we must never risk