btrfs: account ticket size at add/delete time

Instead of iterating all pending tickets on the normal/priority list to
sum their total size the cost can be amortized across ticket addition/
removal. This turns O(n) + O(m) (where n is the size of the normal list
and m of the priority list) into O(1). This will mostly have effect in
workloads that experience heavy flushing.

Signed-off-by: Nikolay Borisov <nborisov@suse.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>

diff --git a/fs/btrfs/space-info.c b/fs/btrfs/space-info.c
index 9cb511d8cd9da695dfa5fd90a87448fa6037321b..8b0fe053a25dadc5bc836f58a65e1c607d48340e 100644 (file)
--- a/fs/btrfs/space-info.c
+++ b/fs/btrfs/space-info.c
@@ -389,6 +389,8 @@ again:
                                                              space_info,
                                                              ticket->bytes);
                        list_del_init(&ticket->list);
+                       ASSERT(space_info->reclaim_size >= ticket->bytes);
+                       space_info->reclaim_size -= ticket->bytes;
                        ticket->bytes = 0;
                        space_info->tickets_id++;
                        wake_up(&ticket->wait);
@@ -784,16 +786,12 @@ static inline u64
 btrfs_calc_reclaim_metadata_size(struct btrfs_fs_info *fs_info,
                                 struct btrfs_space_info *space_info)
 {
-       struct reserve_ticket *ticket;
        u64 used;
        u64 avail;
        u64 expected;
-       u64 to_reclaim = 0;
+       u64 to_reclaim = space_info->reclaim_size;
 
-       list_for_each_entry(ticket, &space_info->tickets, list)
-               to_reclaim += ticket->bytes;
-       list_for_each_entry(ticket, &space_info->priority_tickets, list)
-               to_reclaim += ticket->bytes;
+       lockdep_assert_held(&space_info->lock);
 
        avail = calc_available_free_space(fs_info, space_info,
                                          BTRFS_RESERVE_FLUSH_ALL);
@@ -1192,8 +1190,10 @@ static int __reserve_metadata_bytes(struct btrfs_fs_info *fs_info,
         * the list and we will do our own flushing further down.
         */
        if (ret && flush != BTRFS_RESERVE_NO_FLUSH) {
+               ASSERT(space_info->reclaim_size >= 0);
                ticket.bytes = orig_bytes;
                ticket.error = 0;
+               space_info->reclaim_size += ticket.bytes;
                init_waitqueue_head(&ticket.wait);
                if (flush == BTRFS_RESERVE_FLUSH_ALL) {
                        list_add_tail(&ticket.list, &space_info->tickets);

diff --git a/fs/btrfs/space-info.h b/fs/btrfs/space-info.h
index 24514cd2c6c1f5ee6215c8c75a3fbb8100fb4717..0a5001ef1481712e401d39f75651c411decb7ead 100644 (file)
--- a/fs/btrfs/space-info.h
+++ b/fs/btrfs/space-info.h
@@ -54,6 +54,13 @@ struct btrfs_space_info {
        struct list_head ro_bgs;
        struct list_head priority_tickets;
        struct list_head tickets;
+
+       /*
+        * Size of space that needs to be reclaimed in order to satisfy pending
+        * tickets
+        */
+       u64 reclaim_size;
+
        /*
         * tickets_id just indicates the next ticket will be handled, so note
         * it's not stored per ticket.