On Fri, Mar 20, 2020 at 03:10:41PM +0100, Christoph Hellwig wrote: > Btw, while you touch this code: > > It seems like btrfs_dio_private.subio_endio is rather pointless, > as it is always set to one function for reads and otherwise never > set. De-virtualizing this call could help making the code a little > faster and easier to understand. Doh, one of the later patches actually kills it. Time for more coffee..
