what? no NES support??!??!

Print Thread

Page 13 of 15

…

Re: what? no NES support??!??! etabeta78 #54543 10/02/09 07:09 AM
Joined: Dec 2005 Posts: 443 Heretical_One Senior Member
Heretical_One Senior Member Joined: Dec 2005 Posts: 443	Nice nes_mmc changes. The new format is good, and encourages a mapper system closer to being savestate friendly!

Re: what? no NES support??!??! Heretical_One #54546 10/02/09 07:26 AM
Joined: Jan 2006 Posts: 3,695 Likes: 1 Trondheim, Norway etabeta78 Very Senior Member
etabeta78 Very Senior Member Joined: Jan 2006 Posts: 3,695 Likes: 1 Trondheim, Norway	wait for the next update and you'll see it being even friendlier the final goal is to move all the mapper related variable to a mmc struct so that they can be saved properly with no effort. however, MMC1/MMC3/MMC5 makes life harder, so we are still not there

Re: what? no NES support??!??!

etabeta78 #54551 10/02/09 10:57 AM

Joined: Sep 2001

Posts: 535

Firewave

Senior Member

Firewave

Senior Member

Joined: Sep 2001

Posts: 535

I updated to GoodNES 3.14 and ran the test again. Results have been updated:
http://mess.redump.net/tools:mame_regtest:results:mess:nes

I should note, that I am running a 32-bit MinGW binary on Windows XP x64.
Here's a backtrace of the .unf crash using "Shadowgate (E) [U][!].unf". I had to compile it with GCC 4.4.0 as the official one will not show a proper backtrace on x64.

Code

Program received signal SIGSEGV, Segmentation fault.
0x0091d67f in read_byte_generic (space=0x125e1348, byteaddress=65532)
    at src/emu/memory.c:472
472                     result = (*handler->bankbaseptr)[byteoffset];
(gdb) bt full
#0  0x0091d67f in read_byte_generic (space=0x125e1348, byteaddress=65532)
    at src/emu/memory.c:472
        handler = (const handler_data *) 0x126110c0
        byteoffset = 8188
        entry = 4
        result = 0 '\0'
#1  0x009255cf in memory_read_byte_8le (space=0x125e1348, address=65532)
    at src/emu/memory.c:3901
No locals.
#2  0x00a61b10 in cpu_reset_m6502 (device=0xa4a1f75)
    at src/emu/cpu/m6502/m6502.c:190
        cpustate = (m6502_Regs *) 0x12bc1764
#3  0x0092ba14 in device_reset_cpu (device=0xa4a1f75) at src/emu/cpuexec.c:567
        classdata = (cpu_class_data *) 0x12bc17b8
        config = (const cpu_config *) 0xa4a1fdd
        reset = (cpu_reset_func) 0xa61ae6 <cpu_reset_m6502>
        line = 131074
#4  0x0093ece3 in device_reset (device=0xa4a1f75) at src/emu/devintrf.c:736
        reset = (device_reset_func) 0x92b979 <device_reset_cpu>
#5  0x00688b8e in machine_reset_nes (machine=0xa441efc)
    at src/mess/machine/nes.c:211
        state = (nes_state *) 0xa491ff4
#6  0x00929d6b in soft_reset (machine=0xa441efc, ptr=0x0, param=0)
    at src/emu/mame.c:1601
        mame = (mame_private *) 0xa461f68
        cb = (callback_item *) 0x0
#7  0x0092819c in mame_execute (options=0x7301e58) at src/emu/mame.c:342
        settingsloaded = 0
        driver = (const game_driver *) 0x1103780
        machine = (running_machine *) 0xa441efc
        mame = (mame_private *) 0xa461f68
        cb = (callback_item *) 0x7301e58
        gamename = (astring *) 0xa441f00
        exit_pending = 0
        error = 0
        firstgame = 0
        firstrun = 0
#8  0x009e4b6f in cli_execute (argc=7, argv=0x7251fe4, osd_options=0x123a420)
    at src/emu/clifront.c:172
        options = (core_options *) 0x7301e58
        gamename = (astring *) 0x72e1f00
        exename = (astring *) 0x72f1f00
        gamename_option = 0x7331f08 "nes"
        driver = (const game_driver *) 0x1103780
        result = -1
#9  0x00892753 in utf8_main (argc=7, argv=0x7251fe4)
    at src/osd/windows/winmain.c:259
        ext = 0x350fa90 ".map"
#10 0x00511293 in mame_main (argc=7, a_argv=0x3f4360)
    at src/osd/windows/main.c:74
        i = 7
        rc = 8
        utf8_argv = (char **) 0x7251fe4
        argv = (TCHAR **) 0x3f4280
        wenviron = (WCHAR **) 0x3f5790
        startupinfo = 2008856237
#11 0x00511318 in mess_cli_main (argc=7, argv=0x3f4360)
    at src/mess/osd/windows/messlib.c:25
No locals.
#12 0x0040135a in main (argc=7, argv=0x3f4360)
    at src/mess/osd/windows/climain.c:17
No locals.

Update:
The crash might be related to bad INIT_FAIL handling of cart loading as changing the logerror() in DEVICE_IMAGE_LOAD( nes_cart ) in machine/nes.c to fatalerror() will cause MESS to just exit on UNIF files.

Last edited by Firewave; 10/02/09 11:02 AM.

Re: what? no NES support??!??!

Firewave #54566 10/02/09 03:27 PM

Joined: Sep 2001

Posts: 535

Firewave

Senior Member

Firewave

Senior Member

Joined: Sep 2001

Posts: 535

Speaking of GCC 4.4. We still have some warnings in the MESS code with it:

Code

src/mess/machine/dc.c: In function 'dc_mess_g1_ctrl_w':
src/mess/machine/dc.c:707: warning: unknown conversion type character 'l' in format
src/mess/machine/dc.c:707: warning: format '%x' expects type 'unsigned int', but argument 4 has type 'UINT64'
src/mess/machine/dc.c:707: warning: unknown conversion type character 'l' in format
src/mess/machine/dc.c:707: warning: too many arguments for format

I know I can fix it using core_i64_hex_format(), but I am not sure what the second parameter would be.

Re: what? no NES support??!??! Firewave #54570 10/02/09 07:03 PM
Joined: May 2009 Posts: 2,234 Likes: 410 J Just Desserts Very Senior Member
Just Desserts Very Senior Member J Joined: May 2009 Posts: 2,234 Likes: 410	I was under the impression that GCC 4.4 is crap and not considered ready for prime time in MAME yet. Has this changed?

Re: what? no NES support??!??! Firewave #54571 10/02/09 07:03 PM
Joined: Jan 2006 Posts: 3,695 Likes: 1 Trondheim, Norway etabeta78 Very Senior Member
etabeta78 Very Senior Member Joined: Jan 2006 Posts: 3,695 Likes: 1 Trondheim, Norway	the GCC issue should be fixed now (I copied Aaron's solution for a similar problem in MAME last August) on the NES side, I cleaned up MMC1 and MMC3 and I'm adding some MMC3 clones. The clean up fixed a few Final Fantasy 3 scrolling glitches reported by Shoegazer on IRC (maybe also other small MMC3 problems, I dunno) and it allowed me to greatly simplify MMC1 code. Time for some daily snap Mapper 74 (previously, no chinese text was displayed in the menus) Mapper 12 (a brand new pirate mapper for MESS) Currently, I'm trying to fix two x-in-1 mappers which regressed at a certain point (mapper 45 and 49). I should anyway commit the current progresses later tonight

Re: what? no NES support??!??! etabeta78 #54572 10/02/09 07:35 PM
Joined: Apr 2004 Posts: 1,563 Likes: 12 J judge Very Senior Member
judge Very Senior Member J Joined: Apr 2004 Posts: 1,563 Likes: 12	Nice work

Re: what? no NES support??!??! Just Desserts #54573 10/02/09 07:44 PM
Joined: Mar 2001 Posts: 17,316 Likes: 280 USA R R. Belmont Very Senior Member
R. Belmont Very Senior Member R Joined: Mar 2001 Posts: 17,316 Likes: 280 USA	Originally Posted by Just Desserts I was under the impression that GCC 4.4 is crap and not considered ready for prime time in MAME yet. Has this changed? Only on 32-bit MinGW. MinGW64 and Mac/Linux are fine.

Re: what? no NES support??!??! R. Belmont #54578 10/03/09 12:10 AM
Joined: Feb 2004 Posts: 2,651 Likes: 371 Sydney, Australia Vas Crabb Very Senior Member
Vas Crabb Very Senior Member Joined: Feb 2004 Posts: 2,651 Likes: 371 Sydney, Australia	BS - GCC 4.4 is rubbish on any platform. The optimiser is buggy as hell, and don't even try to use it on PowerPC or SPARC.

Re: what? no NES support??!??! Vas Crabb #54580 10/03/09 12:18 AM
Joined: Mar 2001 Posts: 17,316 Likes: 280 USA R R. Belmont Very Senior Member
R. Belmont Very Senior Member R Joined: Mar 2001 Posts: 17,316 Likes: 280 USA	I dunno, I find it works much better than any compiler Apple ships, all of which continue to mangle machine/model1.c. And Red Hat's 4.4.1 is even nicer.

Page 13 of 15

…

Moderated by Bletch, R. Belmont, Richard Bannister

Link Copied to Clipboard

Home Page Forums MAME MAME what? no NES support??!??!