This is the mail archive of the gcc-patches@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

RE: [patch/hppa-linux] MD_FALLBACK_FRAME_STATE_FOR definition


>>  > For the libgcj part, don't you also want a patch to configure.host
>>  > that sets can_unwind_signal=yes for hppa?
>>
>> Good point.  I wonder how Randolph got it to work...
>
>hrm, i dunno, it seems to work ;-)
>
>ok, let me turn that on and retest.

Run a simple single threaded application that causes and catches a lot of NPEs  under strace, you should see the SIGSEGVs in the system call trace.  Noting of course that "throw new NullPointerException();" does not count as it will not generate a SIGSEGV.
 
You should also try to create a case where the SIGSEGV happens in the first instruction of a function.  This is an edge case that tripped my initial MIPS32 code.  This would probably have to be a leaf method that does not do much so that the compiler can omit the function prolog.  I don't actually know if you can have a function in hppa that will SIGSEGV in the first instruction, but it seems plausable.
 
David Daney
 

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]