On Fri, Apr 20, 2012 at 10:58:19AM +0200, Martin Husemann wrote: > Just to be sure I understand this correctly: > - old binaries do not set the flag and continue to work
Yes. > - new binaries compiled against current sources need to be aware of this > behaviour and explicitly enable/disable it accordingly Yes, if the flag is set, the filesystem must behave appropriately. > - the test source code has not been updated to deal with the ttl cahcing > yet, so is modified to disable it In fact the test code is begging for trouble (but perhaps this is intended): it just sets all valid PUFFS kernel flags without taking care of what they do. It is just like if you called open(2) with O_MASK as the flags argument. -- Emmanuel Dreyfus m...@netbsd.org
