PUFFS race on parent mtime update

[Emmanuel Dreyfus <manu%netbsd.org@localhost>]

Hi

When running regression tests on FUSE filesystems, I hit a i
race condition: when I create a node in a directory, the 
parent directory mtime/ctime seems to be  updated asynchronously,
and some tests that check it report an error because it has not 
been updated at test time.

Here is a PoC:
stat -f '%m' . ; mkfifo x; stat -f  '%m' . ; sleep 1; stat -f  '%m' .

This gives the result:
1409757799
1409757799
1409757818

Which means that when mkfifo returns, the parent directory
mtime is not yet updated. If I wait one second I get it alrgiht.

I have trouble to understand where parent directory ctime/mtime is
updated; Anyone can tell me how this should work?

-- 
Emmanuel Dreyfus
manu%netbsd.org@localhost