btrfs: stop reserving excessive space for block group item insertions

Space for block group item insertions, necessary after allocating a new
block group, is reserved in the delayed refs block reserve. Currently we
do this by incrementing the transaction handle's delayed_ref_updates
counter and then calling btrfs_update_delayed_refs_rsv(), which will
increase the size of the delayed refs block reserve by an amount that
corresponds to the same amount we use for delayed refs, given by
btrfs_calc_delayed_ref_bytes().

That is an excessive amount because it corresponds to the amount of space
needed to insert one item in a btree (btrfs_calc_insert_metadata_size())
times 2 when the free space tree feature is enabled. All we need is an
amount as given by btrfs_calc_insert_metadata_size(), since we only need to
insert a block group item in the extent tree (or block group tree if this
feature is enabled). By using btrfs_calc_insert_metadata_size() we will
need to reserve 2 times less space when using the free space tree, putting
less pressure on space reservation.

So use helpers to reserve and release space for block group item
insertions that use btrfs_calc_insert_metadata_size() for calculation of
the space.

Signed-off-by: Filipe Manana <fdmanana@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>

diff --git a/fs/btrfs/block-group.c b/fs/btrfs/block-group.c
index 9d17b0580fbf9cf5c0e26870ccf8cb91494dccb8..6e5dc68ff661f45d6992869024117b345e62e329 100644 (file)
--- a/fs/btrfs/block-group.c
+++ b/fs/btrfs/block-group.c
@@ -2709,7 +2709,7 @@ void btrfs_create_pending_block_groups(struct btrfs_trans_handle *trans)
 
                /* Already aborted the transaction if it failed. */
 next:
-               btrfs_delayed_refs_rsv_release(fs_info, 1, 0);
+               btrfs_dec_delayed_refs_rsv_bg_inserts(fs_info);
                list_del_init(&block_group->bg_list);
                clear_bit(BLOCK_GROUP_FLAG_NEW, &block_group->runtime_flags);
        }
@@ -2819,8 +2819,7 @@ struct btrfs_block_group *btrfs_make_block_group(struct btrfs_trans_handle *tran
 #endif
 
        list_add_tail(&cache->bg_list, &trans->new_bgs);
-       trans->delayed_ref_updates++;
-       btrfs_update_delayed_refs_rsv(trans);
+       btrfs_inc_delayed_refs_rsv_bg_inserts(fs_info);
 
        set_avail_alloc_bits(fs_info, type);
        return cache;

diff --git a/fs/btrfs/delayed-ref.c b/fs/btrfs/delayed-ref.c
index de843479591349153344b4a7026a46126a202702..9223934d95f4724cbbc4e726306049d444112ee1 100644 (file)
--- a/fs/btrfs/delayed-ref.c
+++ b/fs/btrfs/delayed-ref.c
@@ -125,6 +125,41 @@ void btrfs_update_delayed_refs_rsv(struct btrfs_trans_handle *trans)
        trans->delayed_ref_csum_deletions = 0;
 }
 
+/*
+ * Adjust the size of the delayed refs block reserve for 1 block group item
+ * insertion, used after allocating a block group.
+ */
+void btrfs_inc_delayed_refs_rsv_bg_inserts(struct btrfs_fs_info *fs_info)
+{
+       struct btrfs_block_rsv *delayed_rsv = &fs_info->delayed_refs_rsv;
+
+       spin_lock(&delayed_rsv->lock);
+       /*
+        * Inserting a block group item does not require changing the free space
+        * tree, only the extent tree or the block group tree, so this is all we
+        * need.
+        */
+       delayed_rsv->size += btrfs_calc_insert_metadata_size(fs_info, 1);
+       delayed_rsv->full = false;
+       spin_unlock(&delayed_rsv->lock);
+}
+
+/*
+ * Adjust the size of the delayed refs block reserve to release space for 1
+ * block group item insertion.
+ */
+void btrfs_dec_delayed_refs_rsv_bg_inserts(struct btrfs_fs_info *fs_info)
+{
+       struct btrfs_block_rsv *delayed_rsv = &fs_info->delayed_refs_rsv;
+       const u64 num_bytes = btrfs_calc_insert_metadata_size(fs_info, 1);
+       u64 released;
+
+       released = btrfs_block_rsv_release(fs_info, delayed_rsv, num_bytes, NULL);
+       if (released > 0)
+               trace_btrfs_space_reservation(fs_info, "delayed_refs_rsv",
+                                             0, released, 0);
+}
+
 /*
  * Adjust the size of the delayed refs block reserve for 1 block group item
  * update.

diff --git a/fs/btrfs/delayed-ref.h b/fs/btrfs/delayed-ref.h
index 250e8742a136545e1f3f1b2e4ee5c4fcc2237701..62d679d40f4f91c77123f9a69a80c7fe25fa34a3 100644 (file)
--- a/fs/btrfs/delayed-ref.h
+++ b/fs/btrfs/delayed-ref.h
@@ -436,6 +436,8 @@ int btrfs_check_delayed_seq(struct btrfs_fs_info *fs_info, u64 seq);
 
 void btrfs_delayed_refs_rsv_release(struct btrfs_fs_info *fs_info, int nr_refs, int nr_csums);
 void btrfs_update_delayed_refs_rsv(struct btrfs_trans_handle *trans);
+void btrfs_inc_delayed_refs_rsv_bg_inserts(struct btrfs_fs_info *fs_info);
+void btrfs_dec_delayed_refs_rsv_bg_inserts(struct btrfs_fs_info *fs_info);
 void btrfs_inc_delayed_refs_rsv_bg_updates(struct btrfs_fs_info *fs_info);
 void btrfs_dec_delayed_refs_rsv_bg_updates(struct btrfs_fs_info *fs_info);
 int btrfs_delayed_refs_rsv_refill(struct btrfs_fs_info *fs_info,

diff --git a/fs/btrfs/transaction.c b/fs/btrfs/transaction.c
index 77f8175d464ba0a50b244defd72c0cfa7d00f8a5..3aa59cfa4ab013f91a4c5d8f1035b04d6ef65b5c 100644 (file)
--- a/fs/btrfs/transaction.c
+++ b/fs/btrfs/transaction.c
@@ -2135,7 +2135,7 @@ static void btrfs_cleanup_pending_block_groups(struct btrfs_trans_handle *trans)
        struct btrfs_block_group *block_group, *tmp;
 
        list_for_each_entry_safe(block_group, tmp, &trans->new_bgs, bg_list) {
-               btrfs_delayed_refs_rsv_release(fs_info, 1, 0);
+               btrfs_dec_delayed_refs_rsv_bg_inserts(fs_info);
                list_del_init(&block_group->bg_list);
        }
 }