[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Is GDB broken on ia64?

I have gotten the latest flightgear which now has a feature, a
scripting language called nasal. Apparently I'm the first to try
to run this on ia64. It compiles OK but I get a segfault during startup.
It occurs right after reporting that it is starting nasal and the stack
trace agrees. So I have the subroutine where it happens. But it is
a small routine that appears to be correct and the it is being
passed bad data. When I look at the back trace, the first three
entries look legit but higher up, the back trace has call sequences
that are not possible. ie, reports b called a but the code in b never
calls a
Also, I caught it on a breakpoint entering the routine and then
stepped three times and got the seg fault. But the call parameters
printed out at the breakpoint are completely different from the
call parameters printed after the seg fault.  If the call parameters
where printed correctly at the breakpoint, it should not have
executed the line that seg faulted.

I have trouble stepping thru C code. It seems to jump around
in a manner inconsistent with teh code. It steps thru assembly
just fine.
One other clue on the backtrace: there is a warning
warning: Can't fetch instructions for slot numbers greater than 2.
Using slot 0 instead

Richard Harke

Reply to: