[kaffe] Re: Kaffe GC performance, Boehm GC issues

[Kiyo Inaba]

I wrote:
>>I tested boehm with three systems.
>>1) i386, gcc version 3.2 20020903, Red Hat Linux 8.0 3.2-7
>>2) i386, gcc version 3.2.3 20030502, FC5
>>3) i386, gcc version 4.1.1 20060525, FC5
>>
>>If I use CVS head, and boehm-gc 6.8, they all fail. Testing HWA generates
>>nothing :-< Could you guess any reason why it happens?
>
>This report was wrong. The case no.1 works fine for boehm-gc. The
>previous result was based on much older version of boehm-gc.

So, I go back to FC5 and tested with both kaffe-gc and boehm-gc on gdb.

------------------------- kaffe-gc --------------------------
(gdb) r HelloWorldApp
Starting program: /home/inaba/src/kaffe-snap-071220-linux/kaffe/kaffe/kaffe-bin HelloWorldApp
Reading symbols from shared object read from target memory...done.
Loaded system supplied DSO at 0xa02000
[Thread debugging using libthread_db enabled]
[New Thread -1209026880 (LWP 22333)]
[New Thread -1221637216 (LWP 22336)]
[New Thread -1221903456 (LWP 22337)]
Hello World!
[Thread -1221903456 (zombie) exited]
[Thread -1221637216 (LWP 22336) exited]

Program exited normally.
----------------------- kaffe-gc end ------------------------

------------------------- boehm-gc --------------------------
(gdb) r HelloWorldApp
Starting program: /home/inaba/src/kaffe-snap-071220-linux-gc/kaffe/kaffe/kaffe-bin HelloWorldApp
Reading symbols from shared object read from target memory...done.
Loaded system supplied DSO at 0xa02000
[Thread debugging using libthread_db enabled]
[New Thread -1208199488 (LWP 22327)]

Program received signal SIGSEGV, Segmentation fault.
[Switching to Thread -1208199488 (LWP 22327)]
engine_callMethod (call=0xbfcab270)
    at ../../../../kaffe-snap-071220/config/i386/sysdepCallMethod.h:46
46          ((void (*)(void))(call->function))();
(gdb) c
Continuing.

Program terminated with signal SIGSEGV, Segmentation fault.
The program no longer exists.
----------------------- boehm-gc end ------------------------

In kaffe-gc case, is it true that we no more have intentional seg fault
to check stack boundary? In boehm-gc case, we may loose signal handler
to treat seg fault properly. It may case why boehm-gc does not work
on (my) FC5...

Kiyo