Commit 970e74d9 authored Apr 04, 2018 by David Sterba

btrfs: simplify waiting loop in btrfs_tree_lock



Currently, the number of readers and writers is checked and in case
there are any, wait and redo the locks. There's some duplication
before the branches go back to again label, eg. calling wait_event on
blocking_readers twice.

The sequence is transformed

loop:
* wait for readers
* wait for writers
* write_lock
* check readers, unlock and wait for readers, loop
* check writers, unlock and wait for writers, loop

The new sequence is not exactly the same due to the simplification, for
readers it's slightly faster. For the writers, original code does

* wait for writers
* (loop) wait for readers
*        wait for writers -- again

while the new goes directly to the reader check. This should behave the
same on a contended lock with multiple writers and readers, but can
reduce number of times we're waiting on something.

Reviewed-by: Johannes Thumshirn <jthumshirn@suse.de>
Signed-off-by: David Sterba <dsterba@suse.com>

parent 8bead258

fs/btrfs/locking.c

+2 −9

Original line number	Diff line number	Diff line
		@@ -237,16 +237,9 @@ again:
		wait_event(eb->read_lock_wq, atomic_read(&eb->blocking_readers) == 0);
		wait_event(eb->write_lock_wq, atomic_read(&eb->blocking_writers) == 0);
		write_lock(&eb->lock);
		if (atomic_read(&eb->blocking_readers)) {
		if (atomic_read(&eb->blocking_readers) \|\|
		atomic_read(&eb->blocking_writers)) {
		write_unlock(&eb->lock);
		wait_event(eb->read_lock_wq,
		atomic_read(&eb->blocking_readers) == 0);
		goto again;
		}
		if (atomic_read(&eb->blocking_writers)) {
		write_unlock(&eb->lock);
		wait_event(eb->write_lock_wq,
		atomic_read(&eb->blocking_writers) == 0);
		goto again;
		}
		WARN_ON(atomic_read(&eb->spinning_writers));

Admin message