Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mason/linux-btrfs

Pull second set of btrfs updates from Chris Mason:
 "The most important changes here are from Josef, fixing a btrfs
  regression in 3.14 that can cause corruptions in the extent allocation
  tree when snapshots are in use.

  Josef also fixed some deadlocks in send/recv and other assorted races
  when balance is running"

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mason/linux-btrfs: (23 commits)
  Btrfs: fix compile warnings on on avr32 platform
  btrfs: allow mounting btrfs subvolumes with different ro/rw options
  btrfs: export global block reserve size as space_info
  btrfs: fix crash in remount(thread_pool=) case
  Btrfs: abort the transaction when we don't find our extent ref
  Btrfs: fix EINVAL checks in btrfs_clone
  Btrfs: fix unlock in __start_delalloc_inodes()
  Btrfs: scrub raid56 stripes in the right way
  Btrfs: don't compress for a small write
  Btrfs: more efficient io tree navigation on wait_extent_bit
  Btrfs: send, build path string only once in send_hole
  btrfs: filter invalid arg for btrfs resize
  Btrfs: send, fix data corruption due to incorrect hole detection
  Btrfs: kmalloc() doesn't return an ERR_PTR
  Btrfs: fix snapshot vs nocow writting
  btrfs: Change the expanding write sequence to fix snapshot related bug.
  btrfs: make device scan less noisy
  btrfs: fix lockdep warning with reclaim lock inversion
  Btrfs: hold the commit_root_sem when getting the commit root during send
  Btrfs: remove transaction from send
  ...

1 files changed, 89 insertions, 19 deletions

diff --git a/fs/btrfs/scrub.c b/fs/btrfs/scrub.c
index 93e6d7172844..0be77993378e 100644
--- a/fs/btrfs/scrub.c
+++ b/fs/btrfs/scrub.c
@@ -2235,6 +2235,47 @@ behind_scrub_pages:
         return 0;
 }
 
+/*
+ * Given a physical address, this will calculate it's
+ * logical offset. if this is a parity stripe, it will return
+ * the most left data stripe's logical offset.
+ *
+ * return 0 if it is a data stripe, 1 means parity stripe.
+ */
+static int get_raid56_logic_offset(u64 physical, int num,
+                                   struct map_lookup *map, u64 *offset)
+{
+        int i;
+        int j = 0;
+        u64 stripe_nr;
+        u64 last_offset;
+        int stripe_index;
+        int rot;
+
+        last_offset = (physical - map->stripes[num].physical) *
+                      nr_data_stripes(map);
+        *offset = last_offset;
+        for (i = 0; i < nr_data_stripes(map); i++) {
+                *offset = last_offset + i * map->stripe_len;
+
+                stripe_nr = *offset;
+                do_div(stripe_nr, map->stripe_len);
+                do_div(stripe_nr, nr_data_stripes(map));
+
+                /* Work out the disk rotation on this stripe-set */
+                rot = do_div(stripe_nr, map->num_stripes);
+                /* calculate which stripe this data locates */
+                rot += i;
+                stripe_index = rot % map->num_stripes;
+                if (stripe_index == num)
+                        return 0;
+                if (stripe_index < num)
+                        j++;
+        }
+        *offset = last_offset + j * map->stripe_len;
+        return 1;
+}
+
 static noinline_for_stack int scrub_stripe(struct scrub_ctx *sctx,
                                            struct map_lookup *map,
                                            struct btrfs_device *scrub_dev,
@@ -2256,6 +2297,7 @@ static noinline_for_stack int scrub_stripe(struct scrub_ctx *sctx,
         u64 physical;
         u64 logical;
         u64 logic_end;
+        u64 physical_end;
         u64 generation;
         int mirror_num;
         struct reada_control *reada1;
@@ -2269,16 +2311,10 @@ static noinline_for_stack int scrub_stripe(struct scrub_ctx *sctx,
         u64 extent_len;
         struct btrfs_device *extent_dev;
         int extent_mirror_num;
-        int stop_loop;
-
-        if (map->type & (BTRFS_BLOCK_GROUP_RAID5 |
-                         BTRFS_BLOCK_GROUP_RAID6)) {
-                if (num >= nr_data_stripes(map)) {
-                        return 0;
-                }
-        }
+        int stop_loop = 0;
 
         nstripes = length;
+        physical = map->stripes[num].physical;
         offset = 0;
         do_div(nstripes, map->stripe_len);
         if (map->type & BTRFS_BLOCK_GROUP_RAID0) {
@@ -2296,6 +2332,11 @@ static noinline_for_stack int scrub_stripe(struct scrub_ctx *sctx,
         } else if (map->type & BTRFS_BLOCK_GROUP_DUP) {
                 increment = map->stripe_len;
                 mirror_num = num % map->num_stripes + 1;
+        } else if (map->type & (BTRFS_BLOCK_GROUP_RAID5 |
+                                BTRFS_BLOCK_GROUP_RAID6)) {
+                get_raid56_logic_offset(physical, num, map, &offset);
+                increment = map->stripe_len * nr_data_stripes(map);
+                mirror_num = 1;
         } else {
                 increment = map->stripe_len;
                 mirror_num = 1;
@@ -2319,7 +2360,15 @@ static noinline_for_stack int scrub_stripe(struct scrub_ctx *sctx,
          * to not hold off transaction commits
          */
         logical = base + offset;
-
+        physical_end = physical + nstripes * map->stripe_len;
+        if (map->type & (BTRFS_BLOCK_GROUP_RAID5 |
+                         BTRFS_BLOCK_GROUP_RAID6)) {
+                get_raid56_logic_offset(physical_end, num,
+                                        map, &logic_end);
+                logic_end += base;
+        } else {
+                logic_end = logical + increment * nstripes;
+        }
         wait_event(sctx->list_wait,
                    atomic_read(&sctx->bios_in_flight) == 0);
         scrub_blocked_if_needed(fs_info);
@@ -2328,7 +2377,7 @@ static noinline_for_stack int scrub_stripe(struct scrub_ctx *sctx,
         key_start.objectid = logical;
         key_start.type = BTRFS_EXTENT_ITEM_KEY;
         key_start.offset = (u64)0;
-        key_end.objectid = base + offset + nstripes * increment;
+        key_end.objectid = logic_end;
         key_end.type = BTRFS_METADATA_ITEM_KEY;
         key_end.offset = (u64)-1;
         reada1 = btrfs_reada_add(root, &key_start, &key_end);
@@ -2338,7 +2387,7 @@ static noinline_for_stack int scrub_stripe(struct scrub_ctx *sctx,
         key_start.offset = logical;
         key_end.objectid = BTRFS_EXTENT_CSUM_OBJECTID;
         key_end.type = BTRFS_EXTENT_CSUM_KEY;
-        key_end.offset = base + offset + nstripes * increment;
+        key_end.offset = logic_end;
         reada2 = btrfs_reada_add(csum_root, &key_start, &key_end);
 
         if (!IS_ERR(reada1))
@@ -2356,11 +2405,17 @@ static noinline_for_stack int scrub_stripe(struct scrub_ctx *sctx,
         /*
          * now find all extents for each stripe and scrub them
          */
-        logical = base + offset;
-        physical = map->stripes[num].physical;
-        logic_end = logical + increment * nstripes;
         ret = 0;
-        while (logical < logic_end) {
+        while (physical < physical_end) {
+                /* for raid56, we skip parity stripe */
+                if (map->type & (BTRFS_BLOCK_GROUP_RAID5 |
+                                BTRFS_BLOCK_GROUP_RAID6)) {
+                        ret = get_raid56_logic_offset(physical, num,
+                                        map, &logical);
+                        logical += base;
+                        if (ret)
+                                goto skip;
+                }
                 /*
                  * canceled?
                  */
@@ -2504,15 +2559,29 @@ again:
                         scrub_free_csums(sctx);
                         if (extent_logical + extent_len <
                             key.objectid + bytes) {
-                                logical += increment;
-                                physical += map->stripe_len;
-
+                                if (map->type & (BTRFS_BLOCK_GROUP_RAID5 |
+                                        BTRFS_BLOCK_GROUP_RAID6)) {
+                                        /*
+                                         * loop until we find next data stripe
+                                         * or we have finished all stripes.
+                                         */
+                                        do {
+                                                physical += map->stripe_len;
+                                                ret = get_raid56_logic_offset(
+                                                                physical, num,
+                                                                map, &logical);
+                                                logical += base;
+                                        } while (physical < physical_end && ret);
+                                } else {
+                                        physical += map->stripe_len;
+                                        logical += increment;
+                                }
                                 if (logical < key.objectid + bytes) {
                                         cond_resched();
                                         goto again;
                                 }
 
-                                if (logical >= logic_end) {
+                                if (physical >= physical_end) {
                                         stop_loop = 1;
                                         break;
                                 }
@@ -2521,6 +2590,7 @@ next:
                         path->slots[0]++;
                 }
                 btrfs_release_path(path);
+skip:
                 logical += increment;
                 physical += map->stripe_len;
                 spin_lock(&sctx->stat_lock);