On Sun, Jul 31, 2022 at 02:16:34PM -0400, Theodore Ts'o wrote: > I was just doing a last test of ext4 merged with linux-next before the > merge window opened, and I noticed generic/471 is now failing. After > some more investigation it's failing for xfs and ext4, with the same > problem: > > --- tests/generic/471.out 2022-07-31 00:02:23.000000000 -0400 > +++ /results/xfs/results-4k/generic/471.out.bad 2022-07-31 14:11:47.045330411 0 > @@ -2,12 +2,10 @@ > pwrite: Resource temporarily unavailable > wrote 8388608/8388608 bytes at offset 0 > XXX Bytes, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec) > -RWF_NOWAIT time is within limits. > +pwrite: Resource temporarily unavailable > +(standard_in) 1: syntax error > +RWF_NOWAIT took seconds > ... > > I haven't had a chance to bisect this yet, and for a day or two --- so > I figured I would ask --- is this a known issue? Might have something to do with this set of changes: https://lore.kernel.org/io-uring/c737af00-e879-fe01-380c-ba95b555f423@xxxxxxxxx/ as it adds new places in the VFS that check for IOCB_NOWAIT. Pretty sure it triggers -EAGAIN on timestamp updates now (in file_modified() calls), which is probably what is happening here.... Cheers, Dave. -- Dave Chinner david@xxxxxxxxxxxxx
