>>>>> On Mon, 10 Sep 2001 11:44:02 +0900 (JST), Atsushi Nemoto <nemoto@toshiba-tops.co.jp> said: ralf> Nice test case. Thanks. I decied for a differnet fix attached ralf> below. nemoto> I think your fix is good for Coprocessor Unusable exception. nemoto> And same fix required for Trap or Breakpoint exception. (Am I nemoto> right?) nemoto> But, does not a debugger confused by skipping the instruction nemoto> which cause Trap or Breakpoint exception? (I do not know much nemoto> about communication between kernel and debugger...) I tried same fix for Trap exception (I inserted compute_return_epc() before force_sig(SIGTRAP, current) line in do_tr()). With this fix, gdb did not work correctly. So we should take another fix (at least for Trap exception) ? --- Atsushi Nemoto
