On Thu, 26 May 2011, Hans-Peter Nilsson wrote: > Good, a way forward (well, for me) even though the cure is > suspiciously incidental. Thanks for your patience.
JFTR, setting CRIS_CC0_REGNUM to fixed just moves the assert to trig at the same place in the v10 multilib, where CRIS_MOF_REGNUM is available and allocatable. With regards to regclasses it looks just like CRIS_CC0_REGNUM, so no big surprise. I'll have to debug this myself and throw a patch at you - or suggest documentation updates to rationalise the change to add a separate subclass for CRIS_SRP_REGNUM! brgds, H-P
