On Thu, 2024-08-22 at 15:32 +0200, Miklos Szeredi wrote: > On Thu, 22 Aug 2024 at 15:14, Miklos Szeredi <miklos@xxxxxxxxxx> > wrote: > > > > > What I don't understand is how this results in the -EIO that Jürg > > > > reported. > > > > > > I'm not really familiar with this code but it seems `folio_end_read()` > > > uses xor to update the `PG_uptodate` flag. So if it was already set, it > > > will incorrectly clear the `PG_uptodate` set, which I guess triggers > > > the issue. > > Untested patch attached. Could you please try this? I can no longer reproduce the issue with this patch. So at least with regards to my test case, it looks good to me. Cheers, Jürg