Btrfs: rework how we reserve metadata bytes

With multi-threaded writes we were getting ENOSPC early because somebody would
come in, start flushing delalloc because they couldn't make their reservation,
and in the meantime other threads would come in and use the space that was
getting freed up, so when the original thread went to check to see if they had
space they didn't and they'd return ENOSPC.  So instead if we have some free
space but not enough for our reservation, take the reservation and then start
doing the flushing.  The only time we don't take reservations is when we've
already overcommitted our space, that way we don't have people who come late to
the party way overcommitting ourselves.  This also moves all of the retrying and
flushing code into reserve_metdata_bytes so it's all uniform.  This keeps my
fs_mark test from returning -ENOSPC as soon as it starts and actually lets me
fill up the disk.  Thanks,

Signed-off-by: Josef Bacik <josef@redhat.com>

1 files changed, 3 insertions, 11 deletions

diff --git a/fs/btrfs/relocation.c b/fs/btrfs/relocation.c
index b37d723b9d4a..39adb68a653f 100644
--- a/fs/btrfs/relocation.c
+++ b/fs/btrfs/relocation.c
@@ -178,8 +178,6 @@ struct reloc_control {
         u64 search_start;
         u64 extents_found;
 
-        int block_rsv_retries;
-
         unsigned int stage:8;
         unsigned int create_reloc_tree:1;
         unsigned int merge_reloc_tree:1;
@@ -2133,7 +2131,6 @@ int prepare_to_merge(struct reloc_control *rc, int err)
         LIST_HEAD(reloc_roots);
         u64 num_bytes = 0;
         int ret;
-        int retries = 0;
 
         mutex_lock(&root->fs_info->trans_mutex);
         rc->merging_rsv_size += root->nodesize * (BTRFS_MAX_LEVEL - 1) * 2;
@@ -2143,7 +2140,7 @@ again:
         if (!err) {
                 num_bytes = rc->merging_rsv_size;
                 ret = btrfs_block_rsv_add(NULL, root, rc->block_rsv,
-                                          num_bytes, &retries);
+                                          num_bytes);
                 if (ret)
                         err = ret;
         }
@@ -2155,7 +2152,6 @@ again:
                         btrfs_end_transaction(trans, rc->extent_root);
                         btrfs_block_rsv_release(rc->extent_root,
                                                 rc->block_rsv, num_bytes);
-                        retries = 0;
                         goto again;
                 }
         }
@@ -2405,15 +2401,13 @@ static int reserve_metadata_space(struct btrfs_trans_handle *trans,
         num_bytes = calcu_metadata_size(rc, node, 1) * 2;
 
         trans->block_rsv = rc->block_rsv;
-        ret = btrfs_block_rsv_add(trans, root, rc->block_rsv, num_bytes,
-                                  &rc->block_rsv_retries);
+        ret = btrfs_block_rsv_add(trans, root, rc->block_rsv, num_bytes);
         if (ret) {
                 if (ret == -EAGAIN)
                         rc->commit_transaction = 1;
                 return ret;
         }
 
-        rc->block_rsv_retries = 0;
         return 0;
 }
 
@@ -3554,8 +3548,7 @@ int prepare_to_relocate(struct reloc_control *rc)
          * is no reservation in transaction handle.
          */
         ret = btrfs_block_rsv_add(NULL, rc->extent_root, rc->block_rsv,
-                                  rc->extent_root->nodesize * 256,
-                                  &rc->block_rsv_retries);
+                                  rc->extent_root->nodesize * 256);
         if (ret)
                 return ret;
 
@@ -3567,7 +3560,6 @@ int prepare_to_relocate(struct reloc_control *rc)
         rc->extents_found = 0;
         rc->nodes_relocated = 0;
         rc->merging_rsv_size = 0;
-        rc->block_rsv_retries = 0;
 
         rc->create_reloc_tree = 1;
         set_reloc_control(rc);