On Mon, Jun 26, 2017 at 06:30:34PM +0300, Nikolay Borisov wrote:
>
>
> On 26.06.2017 17:42, Nikolay Borisov wrote:
> > With this patch applied pahole stats look like:
> >
> > /* size: 840, cachelines: 14, members: 40 */
> > /* sum members: 833, holes: 1, sum holes: 7 */
> > /* bit holes: 1, sum bit holes: 28 bits */
> > /* last cacheline: 8 bytes */
> >
> > No functional changes.
> >
> > Signed-off-by: Nikolay Borisov <nbori...@suse.com>
> > ---
> > fs/btrfs/ctree.h | 14 +++++++-------
> > 1 file changed, 7 insertions(+), 7 deletions(-)
> >
> > diff --git a/fs/btrfs/ctree.h b/fs/btrfs/ctree.h
> > index cdd3775e930b..bdd06bbeb9aa 100644
> > --- a/fs/btrfs/ctree.h
> > +++ b/fs/btrfs/ctree.h
> > @@ -586,6 +586,11 @@ struct btrfs_block_group_cache {
> > unsigned int iref:1;
> > unsigned int has_caching_ctl:1;
> > unsigned int removed:1;
> > + /*
> > + * Does the block group need to be added to the free space tree?
> > + * Protected by free_space_lock.
> > + */
> > + unsigned int needs_free_space:1;
> Upon closer inspection of memory-barriers.txt I'm not confident in this
> change. This puts fields protected by different locks in the same
> bitfield which can lead to corrupted values.
Alternatively, you can switch the flags to unsigned long and access them
using set_bit/test_bit, they're atomic regarding updates.
--
To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
