On Sat, Aug 22, 2020 at 03:24:14PM +0100, Matthew Wilcox wrote: > The case I was worrying about: > > fill a filesystem so that free space is very fragmented > readahead into a hole > hole is large, don't allocate an iop > writeback the page > don't have an iop, can't track the write count > > I'd be fine with choosing to allocate an iop later (and indeed I do that > as part of the THP work). But does this scenario make you think of any > other corner cases? Can't think of a corner case. And as said last time this comes up I think trying to allocate the iop as late and lazy as possible is probably a good thing. I just went for the dumb way because it was simpler and already a huge improvement over buffer heads.
