Hi. I've upgraded to Fedora Core 4 and as a result compiled mame with
gcc4, with usual optimisations (-o3 -march=prescott). Unfortunately,
some drivers segfault after 'press any key' screen: segas32, model2. How
to investigate it deeper? Greets.
___
Xmam
I would compile a debug version of xmame and run it through gdb
to narrow it down to where it's dying.
gdb xmame.x11
run -rompath /path/to/games gamename
[get it to crash]
bt
[copy resulting output to show where it died]
To compile the debug version:
1. probably don't want any aggressive CFLAGS
F.J. McCloud wrote:
> I would compile a debug version of xmame and run it through gdb
> to narrow it down to where it's dying.
> gdb xmame.x11
> run -rompath /path/to/games gamename
>
> [get it to crash]
> bt
> [copy resulting output to show where it died]
>
> To compile the debug version:
> 1.
segas32 ga2
#0 0x083ed9ce in ?? ()
#1 0x in ?? ()
model2 vf2
(gdb) bt
#0 0x in ?? ()
#1 0x080af0d5 in ?? ()
#2 0x in ?? ()
___
Xmame mailing list
Xmame@toybox.twisted.org.uk
http://toybox.twisted.org.uk/mailman/listinfo/xmam
Julian Sikorski wrote:
> segas32 ga2
> #0 0x083ed9ce in ?? ()
> #1 0x in ?? ()
>
> model2 vf2
> (gdb) bt
> #0 0x in ?? ()
> #1 0x080af0d5 in ?? ()
> #2 0x in ?? ()
Did you enable -ggdb in your CFLAGS? If so, you also need to enabled
the LD option labeled "For debuggi
Lawrence Gold wrote:
> Julian Sikorski wrote:
>
>>segas32 ga2
>>#0 0x083ed9ce in ?? ()
>>#1 0x in ?? ()
>>
>>model2 vf2
>>(gdb) bt
>>#0 0x in ?? ()
>>#1 0x080af0d5 in ?? ()
>>#2 0x in ?? ()
>
>
> Did you enable -ggdb in your CFLAGS? If so, you also need to enabled
>
Julian Sikorski wrote:
> Lawrence Gold wrote:
>
>>Julian Sikorski wrote:
>>
>>
>>>segas32 ga2
>>>#0 0x083ed9ce in ?? ()
>>>#1 0x in ?? ()
>>>
>>>model2 vf2
>>>(gdb) bt
>>>#0 0x in ?? ()
>>>#1 0x080af0d5 in ?? ()
>>>#2 0x in ?? ()
>>
>>
>>Did you enable -ggdb in your CF
Julian Sikorski wrote:
>
> No difference. I'm attaching the makefile I use.
Strange, your makefile looks fine. I'll try building with GCC 4.0 here
and reproduce the problem in GDB. Could you name some of the games that
are giving you trouble?
Thanks.
___
Lawrence Gold wrote:
Julian Sikorski wrote:
No difference. I'm attaching the makefile I use.
Strange, your makefile looks fine. I'll try building with GCC 4.0 here
and reproduce the problem in GDB. Could you name some of the games that
are giving you trouble?
Thanks.
Lawrence,
RedHa
It crashes with arescue. I get the following message when it crashes:
Unhandled OpCode found : at 00fd
running it inside gdb it works fine. :(
On 6/19/05, Hans de Goede <[EMAIL PROTECTED]> wrote:
>
>
> Lawrence Gold wrote:
> > Julian Sikorski wrote:
> >
> >>No difference. I'm attach
I managed to get a back trace with desert:
(gdb) bt
#0 0x in ?? ()
#1 0x080be41f in cpuintrf_push_context (cpunum=2) at src/cpuintrf.c:790
#2 0x080c23d2 in cpunum_empty_event_queue (cpu_and_inputline=0)
at src/cpuint.c:195
#3 0x080ce126 in mame_timer_set_global_time (newbase=
Lawrence Gold wrote:
> Julian Sikorski wrote:
>
>>No difference. I'm attaching the makefile I use.
>
>
> Strange, your makefile looks fine. I'll try building with GCC 4.0 here
> and reproduce the problem in GDB. Could you name some of the games that
> are giving you trouble?
>
> Thanks.
>
>
Hans de Goede wrote:
>
>
> Lawrence Gold wrote:
>
>> Julian Sikorski wrote:
>>
>>> No difference. I'm attaching the makefile I use.
>>
>>
>>
>> Strange, your makefile looks fine. I'll try building with GCC 4.0 here
>> and reproduce the problem in GDB. Could you name some of the games that
>> a
I've been busy debugging this. The problem is that src/cpu/v60/v60.c
gets miscompiled by gcc4. If you compile that single file with gcc3 and
the rest with gcc4 ga2 works fine.
I don't know howto fix this, can someone forward this to mamedev, maybe
someone there knows what the tricky parts of v
On Jun 21, 2005, at 6:51 AM, Hans de Goede wrote:
I've been busy debugging this. The problem is that src/cpu/v60/
v60.c gets miscompiled by gcc4. If you compile that single file
with gcc3 and the rest with gcc4 ga2 works fine.
I don't know howto fix this, can someone forward this to mamedev,
Lawrence Gold wrote:
On Jun 21, 2005, at 6:51 AM, Hans de Goede wrote:
I've been busy debugging this. The problem is that src/cpu/v60/ v60.c
gets miscompiled by gcc4. If you compile that single file with gcc3
and the rest with gcc4 ga2 works fine.
I don't know howto fix this, can someone
On Jun 21, 2005, at 9:25 AM, Hans de Goede wrote:
Are you using mess-CVS it always crashes for me with mess-CVS with
both gcc4 and gcc3.
I was using plain old 0.97. I'll try CVS tonight, but I imagine it'll
crash.
___
Xmame mailing list
Xmame@t
Hans de Goede wrote:
> I've been busy debugging this. The problem is that src/cpu/v60/v60.c
> gets miscompiled by gcc4. If you compile that single file with gcc3 and
> the rest with gcc4 ga2 works fine.
>
> I don't know howto fix this, can someone forward this to mamedev, maybe
> someone there kno
Julian Sikorski wrote:
> Hans de Goede wrote:
>
>>I've been busy debugging this. The problem is that src/cpu/v60/v60.c
>>gets miscompiled by gcc4. If you compile that single file with gcc3 and
>>the rest with gcc4 ga2 works fine.
>>
>>I don't know howto fix this, can someone forward this to mamede
I have compiled 0.97u2 with gcc 3.2.3, and now both model1 and segas32
drivers work correctly. Guess we'll have to wait till mamedev start
supporting gcc4, and this will probably happen when it wil become more
mature. Greets.
___
Xmame mailing list
Xmame
Julian Sikorski wrote:
> I have compiled 0.97u2 with gcc 3.2.3, and now both model1 and segas32
> drivers work correctly. Guess we'll have to wait till mamedev start
> supporting gcc4, and this will probably happen when it wil become more
> mature. Greets.
It could well be a bug in GCC 4.0 too. Tr
Lawrence Gold wrote:
> Julian Sikorski wrote:
>
>>I have compiled 0.97u2 with gcc 3.2.3, and now both model1 and segas32
>>drivers work correctly. Guess we'll have to wait till mamedev start
>>supporting gcc4, and this will probably happen when it wil become more
>>mature. Greets.
>
>
> It could
Julian Sikorski wrote:
I have compiled 0.97u2 with gcc 3.2.3, and now both model1 and segas32
drivers work correctly. Guess we'll have to wait till mamedev start
supporting gcc4, and this will probably happen when it wil become more
mature. Greets.
_
Hans de Goede wrote:
>
>
> Julian Sikorski wrote:
>
>> I have compiled 0.97u2 with gcc 3.2.3, and now both model1 and segas32
>> drivers work correctly. Guess we'll have to wait till mamedev start
>> supporting gcc4, and this will probably happen when it wil become more
>> mature. Greets.
>>
>>
Hmm,
Can you list a couple of model1 games, currently I've only investiged
sega2 games failing with gcc4, this happens because of a miscompile of
the v60 cpucore, dunno howto fix it though.
Do model1 game also use the v60 cpu?
Regards,
Hans
Julian Sikorski wrote:
Hans de Goede wrote:
Yes, Sega Model 1 uses v60 cpu. As for the games, there are four of
them: vf, vr, swa and wingwar. For a more complete list, go to
http://www.system16.com
___
Xmame mailing list
Xmame@toybox.twisted.org.uk
http://toybox.twisted.org.uk/mailman/listinfo/xm
26 matches
Mail list logo