| Re: [hatari-devel] Re: Magic doesn't work with Hatari [ First tests] |
[ Thread Index |
Date Index
| More lists.tuxfamily.org/hatari-devel Archives
]
Hi all,
I've done a few tests with Magic (I've found it somewhere, I'll just use
it for some tests, then I remove it (of course :))
Falcon test :
----------------
It tests for the FPU addresses, doesn't find it and continues until a
test for the Microwire.
It loops forever here :
$00c792 : 4a78 8922 tst.w $ffff8922.w
$00c796 : 66fa bne.s $c792
As there's no microwire in the Falcon, I've tried to remove it from
IoMemTabFalcon.c
Then, it continues, to the message I sent before (Thanks to Stefan for
the translation):
Offscreen driver not found
Starting MCMD
Halting system
Do anybody knows this MCMD ?
Thomas, should we remove the microwire addresses from the
IoMemTabFalcon.c as I did (There's really no microwire in the Falcon) ?
I know you did a lot of tests on the bus errors.
Actually, it looks like :
// { 0xff8922, SIZE_WORD, IoMem_VoidRead, IoMem_VoidWrite
}, /* Microwire data */
// { 0xff8924, SIZE_WORD, IoMem_VoidRead, IoMem_VoidWrite
}, /* Microwire mask */
To notice: Magic tests a lot of things :
M68000 Bus Error reading at address $f00039.
M68000 Bus Error reading at address $f00039. // IDE Controller
(Falcon, ST-Book, IDE cards) (Alternate status register)
M68000 Bus Error reading at address $fffa46. // FPU coprocessor
M68000 Bus Error writing at address $ff8924. // Microwire access
after I removed it from IoMemTabFalcon
M68000 Bus Error writing at address $ff8e0d. // VME Bus system
control (Mste, TT only)
....
Exception 2 (0) at e02ce2 -> e02ce6!
A-Trap a000 at e010a0 (0x1a7ee20)
A-Trap a00e at e010c0 (0x1a7ee40)
A-Trap a000 at fa022c (0x1c1dfac)
Exception 2 (0) at e01854 -> e01836!
Exception 2 (0) at e01854 -> e01836!
PMOVE 00000000,TC PC=00004DAE // MMU in action
PMOVE 00000000,TT0 PC=00004DB6 // MMU in action
PMOVE 00000000,TT1 PC=00004DBE // MMU in action
Exception 2 (0) at c6ce -> c6d4!
Exception 2 (0) at c77c -> c7ce!
Exception 2 (0) at 505e -> 5064!
Illegal instruction: 43bf at 0000CA48 -> 0000CA5A
PMOVE 8000000200000700,CRP PC=000051D4
PMOVE 80F04445,TC PC=000051DC
PMOVE 017E8107,TT0 PC=000051E4
PMOVE 807E8507,TT1 PC=000051EC
A-Trap a000 at fa022c (0x1c1dfac)
Exception 2 (0) at 6788 -> 67c6!
Exception 2 (0) at 67de -> 67fc!
Another problem :
Magic loads a program into the DSP, but it corrupts it (it does an
illegal instruction at address $51 (which is generally a jmp $51
instruction).
But when they load the program to the DSP memory, they send a non
instruction opcode, which generates a lot on ILLEGAL INSTRUCTIONS.
If we manage to let magic run one day, I'll have to change this (remove
this message, or display it only once).
TT TEST:
-----------
I don't know TT a lot, I've booted hatari with 32 Mhz, M68030 (without
cycle exact and prefetch), Tos 2.05 and tos 2.06 fr, 4 Mo of Ram.
I get 2 bombs, then 20 bombs.
The console displays :
GEMDOS HDD emulation, C: <->
/media/0470BC8670BC7FCA/Jeux/Atari/DiskDur.FAL/UTILITIE.S/DESKTOP/MAGIC/magic504.crk/MAGICUPD.
M68000 Bus Error reading at address $fffa40. // FPU
Exception 2 (0) at e00408 -> e0040e!
M68000 Bus Error reading at address $ff8c83. // Zilog 8530 SCC
(MSTe/TT/F030) What is this ????????
Exception 2 (0) at e020c8 -> e02140!
M68000 Bus Error reading at address $ff8a00. // Blitter (not
present on TT)
Exception 2 (0) at e0119c -> e0119e!
M68000 Bus Error writing at address $fffc3b. // Realtime
clock ??????
Exception 2 (0) at e013dc -> e01408!
A-Trap a000 at fa022c (0x1c1cfac)
PMOVE 00000000,TC PC=00004DAE
PMOVE 00000000,TT0 PC=00004DB6
PMOVE 00000000,TT1 PC=00004DBE
M68000 Bus Error reading at address $fffa46. // FPU
Exception 2 (0) at c6ce -> c6d4!
M68000 Bus Error reading at address $ff82c0. // Videl (Falcon only)
Exception 2 (0) at c72c -> c746!
M68000 Bus Error writing at address $ff8c14. // SCC-DMA (TT Only!)
Exception 2 (0) at 9ec2 -> 89e0!
Illegal instruction: 00e0 at 00000108 -> 000089E0
Then it freezes.
For the latest memory test, I've got the following infos :
$FF8C12|word |Residue Data register (Low-Word)
|R (TT)
$FF8C15|byte |Control register BIT 7 6 . . . . 1
0|R/W (TT)
| |Bus error ----------------------------' | | ||
| |Byte count zero ------------------------' | ||
| |Enable -------------------------------------------' ||
| |DMA Direction (1 - out to port) --------------------'|
It seems that Magic is really close to the hardware at the beginning (it
does a lot of tests).
Any suggestion here ?
I don't know if I can do more things on it for now.
Regards
Laurent
Le 29/01/2012 21:51, Stefan Meier a écrit :
Hi,
the lines mean:
Offscreen driver not found
Starting MCMD
Halting system
Stefan