[kaffe] DArwin6 JIT3

Dalibor Topic robilad at kaffe.org
Sun Feb 20 11:52:21 PST 2005


Timothy Stack wrote:
> 
> On Feb 20, 2005, at 5:21 AM, Dalibor Topic wrote:
> 
> As for why stuff doesn't work, have you guys tried reverting my last 
> changes to
> sysdepCallMethod and the ppc trampoline?  I probably screwed it up wrt 
> darwin
> when I was getting linux to work.

Hi Tim!

thanks for the quick reply. I've actually got the problem with 
ppc-linux-jit3 now, too. jitBasic dies as soon as the generated code is 
invoked in the first test :(

In order to get familiar with the jit3/ppc setup i thought about porting 
the mnemonicizer setup to x86. Does that souind like a useful idea?

>> Now I just need to figure out which switch to toggle on to see the 
>> generated code :)
> 
> 
> I almost always look at gcc's disassembly of it.  So you want to:
> 
>   # put a breakpoint on the jitter and start the test
>   (gdb) break translate
>   (gdb) run
>   # let the jitter run
>   (gdb) finish
>   # load the xdebugging symbols
>   (gdb) xdb

ah! xdebugging works with ppc?

>   # disassemble the native code.
>   (gdb) disas meth.ncode    # or whatever the structure is nowadays
> 
> It doesn't always work quite right unfortunately, but it should get you 
> pretty far.

thanks :) any ppc-assmebler reference you could recommend?

cheers,
dalibor topic





More information about the kaffe mailing list