Version 4 of fixes for uaccess fault handling on 68030, tested on 68030 and 68040. Patch 1 has additional exception table entries as well as a final NOP added - evidently, in some cases the fault PC points more than a single instruction past the moves that caused the fault. Patch 2 is unchanged from RFC v2. These patches would still benefit from testing on 68060 and Coldfire. Looking at access_error060, I would not expect the 68060 to behave much different from the 68040 and in particular, addition of exception table entries to not cause any regression. Less confident for Coldfire. Cheers, Michael
