On Sun, Aug 28, 2011 at 07:54:37PM -0400, Ted Ts'o wrote:

Has anyone else noticed failures with xfstests #74?  What I'm finding
is that if compile the fstest program so that it is statically linked,
I can't get it to fail:

I've put up an my KVM-based xfstests appliance here:

ftp://ftp.kernel.org/pub/linux/kernel/people/tytso/V-xfstests.tar.gz

Hopefully this will be useful for other people who are trying to do QA
work for ext4.  I find it's much more convenient to test out a kernel
by running "kvm-xfstests smoke", or "kvm-xfstests full".

BTW, I've found that the xfstests 74 test failure is very timing
dependent.  It doesn't fail when I use another system which has faster
disks than dm-crypt'ed laptop drives, and where writeback mode is
enabled with lots of memory available for caching purposes.  It seems
to require memory pressure and slower disks.  I should check and see
whether it involves ext4_writepage() getting invoked out of the direct
reclaim path...

	     	     	    	      	  	  - Ted