On Mon, Jan 24, 2011 at 04:06:56AM -0500, Christoph Hellwig wrote: > On Wed, Jan 19, 2011 at 03:29:59PM +1100, Dave Chinner wrote: > > + * Also, for non-realtime files, limit the extent size hint to > > + * half the size of the AGs in the filesystem so alignment > > + * doesn't result in extents larger than an AG. > > What do we do about the last AG, which potentionally is smaller than the > others? Not sure - I didn't actually consider that case because I forgot about it. At first glance it doesn't matter because it won't get selected because the length of the AG is less than the requested size unless it is the best match. If it is the best match, we can't do extsize alignment anyway, so it's the same case as any other allocation without a freespace extent large enough for alignment. I'll do a bit more analysis, but I think we're OK here. Cheers, Dave. -- Dave Chinner david@xxxxxxxxxxxxx _______________________________________________ xfs mailing list xfs@xxxxxxxxxxx http://oss.sgi.com/mailman/listinfo/xfs
