>>>>> On Sat, 8 Sep 2001 01:36:38 +0200, Ralf Baechle <email@example.com> said:
>> I make a change for do_signal() to check failure of setup_frame()
>> and continue processing pending signals. It seems work for me.
>> Here is the patch. Any comments are welcome.
ralf> Nice test case. Thanks. I decied for a differnet fix attached
I think your fix is good for Coprocessor Unusable exception. And same
fix required for Trap or Breakpoint exception. (Am I right?)
But, does not a debugger confused by skipping the instruction which
cause Trap or Breakpoint exception? (I do not know much about
communication between kernel and debugger...)