On Wed, Feb 17, 2021 at 10:34:39PM +0000, David Howells wrote: > Matthew Wilcox <willy@xxxxxxxxxxxxx> wrote: > > > We're defeating the ondemand_readahead() algorithm here. Let's suppose > > userspace is doing 64kB reads, the filesystem is OrangeFS which only > > wants to do 4MB reads, the page cache is initially empty and there's > > only one thread doing a sequential read. ondemand_readahead() calls > > get_init_ra_size() which tells it to allocate 128kB and set the async > > marker at 64kB. Then orangefs calls readahead_expand() to allocate the > > remainder of the 4MB. After the app has read the first 64kB, it comes > > back to read the next 64kB, sees the readahead marker and tries to trigger > > the next batch of readahead, but it's already present, so it does nothing > > (see page_cache_ra_unbounded() for what happens with pages present). > > It sounds like Christoph is right on the right track and the vm needs to ask > the filesystem (and by extension, the cache) before doing the allocation and > before setting the trigger flag. Then we don't need to call back into the vm > to expand the readahead. Doesn't work. You could read my reply to Christoph, or try to figure out how to get rid of https://evilpiepirate.org/git/bcachefs.git/tree/fs/bcachefs/fs-io.c#n742 for yourself. > Also, there's Steve's request to try and keep at least two requests in flight > for CIFS/SMB at the same time to consider. That's not relevant to this problem.
