From mboxrd@z Thu Jan 1 00:00:00 1970 From: Chris Mason Subject: Re: btrfs: warn_slowpath in clean_tree_block and others Date: Wed, 25 Feb 2009 14:26:43 -0500 Message-ID: <1235590003.32346.61.camel@think.oraclecorp.com> References: <20090224230249.GA6859@vlad.carfax.org.uk> <1235586986.32346.48.camel@think.oraclecorp.com> <89ed0c690902251050g1e6dd23ay5d5426adb7086018@mail.gmail.com> Mime-Version: 1.0 Content-Type: text/plain To: "Mitch Harder (aka DontPanic)" , linux-btrfs Return-path: In-Reply-To: <89ed0c690902251050g1e6dd23ay5d5426adb7086018@mail.gmail.com> List-ID: [ resend with the list cc'd ] On Wed, 2009-02-25 at 12:50 -0600, Mitch Harder (aka DontPanic) wrote: > I'll try to test that out. > > I had just noticed that some of my kernel configuration settings (not > sure which ones) seem to affect the clean_tree_block warnings I've > been getting, and one of my customizations is usually to configure the > kernel for a single processor. > I'll push out a patch tonight that fixes things, the code to test for a locked buffer is just broken on UP. For now, the patch below will do: diff --git a/fs/btrfs/locking.c b/fs/btrfs/locking.c index 85506c4..4513ecf 100644 --- a/fs/btrfs/locking.c +++ b/fs/btrfs/locking.c @@ -222,6 +222,5 @@ int btrfs_tree_unlock(struct extent_buffer *eb) int btrfs_tree_locked(struct extent_buffer *eb) { - return test_bit(EXTENT_BUFFER_BLOCKING, &eb->bflags) || - spin_is_locked(&eb->lock); + return 1; }