3DO ZONE Forums

The Only R.E.A.L. 3DO Experience.
It is currently Sun May 28, 2017 4:11 am

All times are UTC [ DST ]




Post new topic Reply to topic  [ 8 posts ] 
Author Message
PostPosted: Tue Mar 24, 2015 6:11 pm 
Offline
Jaguar MOD
User avatar

Joined: Mon Apr 01, 2013 7:09 am
Posts: 820
I am going to start a new thread for this as it deserves one.

A new build of VASM for the Jrisc is here. It's farther along now but still in it's infancy. I will be able to test it this weekend. It will be up and available for download by tonight or tomorrow.

Quote:
Ok... I did it. The last days I implemented a whole new syntax module
for MadMac compatibility.

The new source snapshot will be available in 11 hours.

Build the new vasm for Jaguar with MadMac syntax like this:
make CPU=jagrisc SYNTAX=madmac

Please report any problems which are left.

What I didn't find out:

1. When you have a .text section and switch into absolute mode with .org,
how do you return to .text in MadMac? With another .text directive?

2. What does .gpumain do? I left it out.

Regards,

--
Frank Wille


http://sun.hasenbraten.de/vasm/index.php?view=relsrc

I already talked to him about the .gpu main directive. The other one I couldn't help with.

_________________
What came after the Jaguar was the PS1 which for all it's greatness, ushered in corporate development and with it the bleached, repetitive, bland titles which for the most part we're still playing today. - David Wightman


Top
 Profile  
 
PostPosted: Wed Mar 25, 2015 6:13 pm 
Offline
Jaguar MOD
User avatar

Joined: Mon Apr 01, 2013 7:09 am
Posts: 820
Here is the 32bit Windows build of VASM_madmac for anyone who wants to try it out.

Attachment:
vasmjagrisc_madmac.zip [161.7 KiB]
Downloaded 54 times

_________________
What came after the Jaguar was the PS1 which for all it's greatness, ushered in corporate development and with it the bleached, repetitive, bland titles which for the most part we're still playing today. - David Wightman


Top
 Profile  
 
 Post subject: Re: JRISC VASM
PostPosted: Wed Mar 25, 2015 6:36 pm 
Offline
Jaguar MOD
User avatar

Joined: Mon Apr 01, 2013 7:09 am
Posts: 820
I actually had some time this morning so I couldn't control my curiousity and gave the new madmac backend a spin.

Once again I am using the test asm output file from the Jaguars only risc gcc.

Below that will be the the results of VASM_madmacs first testrun on it.

Here is the test source:

Code:
;GCC for Atari Jaguar GPU/DSP (Jun 12 1995) (C)1994-95 Brainstorm
   MACRO   _RTS
   load   (ST),TMP
   jump   T,(TMP)
   addqt   #4,ST   ;rts
   ENDM
_test_start::
   .GPU
   .ORG   $F03000
ST   .REGEQU   r18
FP   .REGEQU   r17
TMP   .REGEQU   r16
GT   .CCDEF   $15
gcc2_compiled_for_madmac:
   ;(.DATA)
   .LONG
_global::
   .DC.L   5
   ;(.TEXT)
   .EVEN
_GetTest1::
   subqt   #4,ST
   store   FP,(ST)
   move   ST,FP   ;link
   subqt   #8,ST
   move   FP,r1   ;movsi   FP->r1
   move   FP,r2   ;movsi   FP->r2
   subqt   #4,r2   ;isubqtsi3   r2-#4->r2
   move   r2,r3   ;movsi   r2->r3
   store   r0,(r3)   ;movsi   r0->(r3)
   move   FP,r0   ;movsi   FP->r0
   move   FP,r4   ;movsi   FP->r4
   subqt   #8,r4   ;isubqtsi3   r4-#8->r4
   move   r4,r0   ;movsi   r4->r0
   moveq   #0,r6   ;movsi   #0->r6
   store   r6,(r0)   ;movsi   r6->(r0)
   move   FP,r0   ;movsi   FP->r0
   move   FP,r4   ;movsi   FP->r4
   subqt   #4,r4   ;isubqtsi3   r4-#4->r4
   move   r4,r0   ;movsi   r4->r0
   load   (r0),r4   ;movsi   (r0)->r4
   moveq   #9,r0   ;movsi   #9->r0
   cmp   r4,r0   ;rcmpsi   r4,r0
   movei   #L2,TMP
   jump   GT,(TMP)
   nop      ;jgt   L2
   move   FP,r0   ;movsi   FP->r0
   move   FP,r4   ;movsi   FP->r4
   subqt   #8,r4   ;isubqtsi3   r4-#8->r4
   move   r4,r0   ;movsi   r4->r0
   movei   #100,r6   ;movsi   #100->r6
   store   r6,(r0)   ;movsi   r6->(r0)
   movei   #L3,TMP
   jump   T,(TMP)
   nop      ;jt   L3
   .EVEN
L2:
   move   FP,r0   ;movsi   FP->r0
   move   FP,r4   ;movsi   FP->r4
   subqt   #4,r4   ;isubqtsi3   r4-#4->r4
   move   r4,r0   ;movsi   r4->r0
   load   (r0),r4   ;movsi   (r0)->r4
   moveq   #9,r0   ;movsi   #9->r0
   cmp   r4,r0   ;rcmpsi   r4,r0
   movei   #L4,TMP
   jump   GT,(TMP)
   nop      ;jgt   L4
   move   FP,r0   ;movsi   FP->r0
   move   FP,r4   ;movsi   FP->r4
   subqt   #8,r4   ;isubqtsi3   r4-#8->r4
   move   r4,r0   ;movsi   r4->r0
   moveq   #10,r6   ;movsi   #10->r6
   store   r6,(r0)   ;movsi   r6->(r0)
   movei   #L5,TMP
   jump   T,(TMP)
   nop      ;jt   L5
   .EVEN
L4:
   move   FP,r0   ;movsi   FP->r0
   move   FP,r4   ;movsi   FP->r4
   subqt   #4,r4   ;isubqtsi3   r4-#4->r4
   move   r4,r0   ;movsi   r4->r0
   load   (r0),r4   ;movsi   (r0)->r4
   cmpq   #0,r4   ;tstsi   r4
   movei   #L6,TMP
   jump   EQ,(TMP)
   nop      ;jeq   L6
   move   FP,r0   ;movsi   FP->r0
   move   FP,r4   ;movsi   FP->r4
   subqt   #8,r4   ;isubqtsi3   r4-#8->r4
   move   r4,r0   ;movsi   r4->r0
   moveq   #1,r6   ;movsi   #1->r6
   store   r6,(r0)   ;movsi   r6->(r0)
L6:
L5:
L3:
   move   FP,r0   ;movsi   FP->r0
   move   FP,r4   ;movsi   FP->r4
   subqt   #8,r4   ;isubqtsi3   r4-#8->r4
   move   r4,r5   ;movsi   r4->r5
   load   (r5),r0   ;movsi   (r5)->r0
   movei   #L1,TMP
   jump   T,(TMP)
   nop      ;jt   L1
   .EVEN
L1:
   move   FP,ST
   load   (ST),FP
   addqt   #4,ST   ;unlk
   _RTS
   .EVEN
_GetTest2::
   subqt   #4,ST
   store   FP,(ST)
   move   ST,FP   ;link
   subqt   #8,ST
   move   FP,r1   ;movsi   FP->r1
   move   FP,r2   ;movsi   FP->r2
   subqt   #4,r2   ;isubqtsi3   r2-#4->r2
   move   r2,r3   ;movsi   r2->r3
   store   r0,(r3)   ;movsi   r0->(r3)
   move   FP,r0   ;movsi   FP->r0
   move   FP,r4   ;movsi   FP->r4
   subqt   #8,r4   ;isubqtsi3   r4-#8->r4
   move   r4,r0   ;movsi   r4->r0
   moveq   #0,r6   ;movsi   #0->r6
   store   r6,(r0)   ;movsi   r6->(r0)
   move   FP,r0   ;movsi   FP->r0
   move   FP,r4   ;movsi   FP->r4
   subqt   #4,r4   ;isubqtsi3   r4-#4->r4
   move   r4,r5   ;movsi   r4->r5
   load   (r5),r0   ;movsi   (r5)->r0
   moveq   #9,r4   ;movsi   #9->r4
   cmp   r4,r0   ;cmpsi   r4,r0
   movei   #L18,TMP
   jump   GT,(TMP)
   nop      ;jgt   L18
   moveq   #1,r4   ;movsi   #1->r4
   cmp   r4,r0   ;cmpsi   r4,r0
   movei   #L18,TMP
   jump   MI,(TMP)
   nop      ;jlt   L18
   movei   #L9,TMP
   jump   T,(TMP)
   nop      ;jt   L9
   .EVEN
L9:
L10:
L11:
L12:
L13:
L14:
L15:
L16:
L17:
   move   FP,r0   ;movsi   FP->r0
   move   FP,r4   ;movsi   FP->r4
   subqt   #8,r4   ;isubqtsi3   r4-#8->r4
   move   r4,r0   ;movsi   r4->r0
   moveq   #1,r6   ;movsi   #1->r6
   store   r6,(r0)   ;movsi   r6->(r0)
   movei   #L8,TMP
   jump   T,(TMP)
   nop      ;jt   L8
   .EVEN
L18:
   move   FP,r0   ;movsi   FP->r0
   move   FP,r4   ;movsi   FP->r4
   subqt   #8,r4   ;isubqtsi3   r4-#8->r4
   move   r4,r0   ;movsi   r4->r0
   moveq   #10,r6   ;movsi   #10->r6
   store   r6,(r0)   ;movsi   r6->(r0)
L8:
   move   FP,r0   ;movsi   FP->r0
   move   FP,r4   ;movsi   FP->r4
   subqt   #8,r4   ;isubqtsi3   r4-#8->r4
   move   r4,r5   ;movsi   r4->r5
   load   (r5),r0   ;movsi   (r5)->r0
   movei   #L7,TMP
   jump   T,(TMP)
   nop      ;jt   L7
   .EVEN
L7:
   move   FP,ST
   load   (ST),FP
   addqt   #4,ST   ;unlk
   _RTS
   .EVEN
_Test3::
   subqt   #4,ST
   store   FP,(ST)
   move   ST,FP   ;link
   subqt   #32,ST
   subqt   #8,ST
   move   ST,r14
   store   r19,(ST)
   store   r20,(r14+1)
   store   r21,(r14+2)
   store   r22,(r14+3)
   store   r23,(r14+4)
   store   r24,(r14+5)
   move   FP,r19   ;movsi   FP->r19
   move   FP,r20   ;movsi   FP->r20
   subqt   #4,r20   ;isubqtsi3   r20-#4->r20
   move   r20,r21   ;movsi   r20->r21
   store   r0,(r21)   ;movsi   r0->(r21)
   move   FP,r0   ;movsi   FP->r0
   move   FP,r1   ;movsi   FP->r1
   subqt   #8,r1   ;isubqtsi3   r1-#8->r1
   move   r1,r0   ;movsi   r1->r0
   moveq   #0,r3   ;movsi   #0->r3
   store   r3,(r0)   ;movsi   r3->(r0)
   move   FP,r0   ;movsi   FP->r0
   move   FP,r1   ;movsi   FP->r1
   subqt   #16,r1   ;isubqtsi3   r1-#16->r1
   move   r1,r0   ;movsi   r1->r0
   moveq   #0,r3   ;movsi   #0->r3
   store   r3,(r0)   ;movsi   r3->(r0)
L21:
   move   FP,r0   ;movsi   FP->r0
   move   FP,r1   ;movsi   FP->r1
   subqt   #8,r1   ;isubqtsi3   r1-#8->r1
   move   r1,r0   ;movsi   r1->r0
   move   FP,r1   ;movsi   FP->r1
   move   FP,r2   ;movsi   FP->r2
   subqt   #4,r2   ;isubqtsi3   r2-#4->r2
   move   r2,r1   ;movsi   r2->r1
   load   (r0),r0   ;movsi   (r0)->r0
   load   (r1),r1   ;movsi   (r1)->r1
   cmp   r1,r0   ;cmpsi   r1,r0
   movei   #L24,TMP
   jump   MI,(TMP)
   nop      ;jlt   L24
   movei   #L22,TMP
   jump   T,(TMP)
   nop      ;jt   L22
   .EVEN
L24:
   move   FP,r0   ;movsi   FP->r0
   move   FP,r1   ;movsi   FP->r1
   subqt   #12,r1   ;isubqtsi3   r1-#12->r1
   move   r1,r0   ;movsi   r1->r0
   move   FP,r1   ;movsi   FP->r1
   move   FP,r2   ;movsi   FP->r2
   subqt   #8,r2   ;isubqtsi3   r2-#8->r2
   move   r2,r1   ;movsi   r2->r1
   moveq   #1,r2   ;movsi   #1->r2
   load   (r1),r1   ;movsi   (r1)->r1
   neg   r1   ;negsi2   r1->r1
   sha   r1,r2   ;iashlsi3   r2<<r1->r2
   neg   r1   ;negsi2   r1->r1
   store   r2,(r0)   ;movsi   r2->(r0)
   move   FP,r0   ;movsi   FP->r0
   move   FP,r1   ;movsi   FP->r1
   subqt   #16,r1   ;isubqtsi3   r1-#16->r1
   move   r1,r22   ;movsi   r1->r22
   move   ST,r0   ;movsi   ST->r0
   move   FP,r1   ;movsi   FP->r1
   move   FP,r0   ;movsi   FP->r0
   subqt   #8,r0   ;isubqtsi3   r0-#8->r0
   move   r0,r1   ;movsi   r0->r1
   load   (r1),r0   ;movsi   (r1)->r0
   movei   #_GetTest1,r23   ;movsi   #_GetTest1->r23
   move   PC,TMP
   subqt   #4,ST
   addqt   #10,TMP
   jump   T,(r23)
   store   TMP,(ST)   ;call   r23->r0
   move   r0,r23   ;movsi   r0->r23
   move   ST,r0   ;movsi   ST->r0
   move   FP,r1   ;movsi   FP->r1
   move   FP,r0   ;movsi   FP->r0
   subqt   #12,r0   ;isubqtsi3   r0-#12->r0
   move   r0,r1   ;movsi   r0->r1
   load   (r1),r0   ;movsi   (r1)->r0
   movei   #_GetTest2,r24   ;movsi   #_GetTest2->r24
   move   PC,TMP
   subqt   #4,ST
   addqt   #10,TMP
   jump   T,(r24)
   store   TMP,(ST)   ;call   r24->r0
   move   FP,r1   ;movsi   FP->r1
   move   FP,r2   ;movsi   FP->r2
   subqt   #16,r2   ;isubqtsi3   r2-#16->r2
   move   r2,r1   ;movsi   r2->r1
   add   r23,r0   ;iaddsi3   r23+r0->r0
   load   (r1),r1   ;movsi   (r1)->r1
   add   r1,r0   ;iaddsi3   r1+r0->r0
   store   r0,(r22)   ;movsi   r0->(r22)
L23:
   move   FP,r0   ;movsi   FP->r0
   move   FP,r1   ;movsi   FP->r1
   subqt   #8,r1   ;isubqtsi3   r1-#8->r1
   move   r1,r0   ;movsi   r1->r0
   move   FP,r1   ;movsi   FP->r1
   move   FP,r2   ;movsi   FP->r2
   subqt   #8,r2   ;isubqtsi3   r2-#8->r2
   move   r2,r0   ;movsi   r2->r0
   move   FP,r1   ;movsi   FP->r1
   move   FP,r2   ;movsi   FP->r2
   subqt   #8,r2   ;isubqtsi3   r2-#8->r2
   move   r2,r1   ;movsi   r2->r1
   load   (r1),r2   ;movsi   (r1)->r2
   move   r2,r1   ;movsi   r2->r1
   addqt   #1,r1   ;iaddqtsi3   #1+r1->r1
   move   r1,r2   ;movsi   r1->r2
   store   r2,(r0)   ;movsi   r2->(r0)
   movei   #L21,TMP
   jump   T,(TMP)
   nop      ;jt   L21
   .EVEN
L22:
   move   FP,r0   ;movsi   FP->r0
   move   FP,r1   ;movsi   FP->r1
   subqt   #16,r1   ;isubqtsi3   r1-#16->r1
   move   r1,r2   ;movsi   r1->r2
   load   (r2),r0   ;movsi   (r2)->r0
   movei   #L20,TMP
   jump   T,(TMP)
   nop      ;jt   L20
   .EVEN
L20:
   move   ST,r14
   load   (ST),r19
   load   (r14+1),r20
   load   (r14+2),r21
   load   (r14+3),r22
   load   (r14+4),r23
   load   (r14+5),r24
   move   FP,ST
   load   (ST),FP
   addqt   #4,ST   ;unlk
   _RTS
   .LONG
   .68000
_test_end::
_test_size   .EQU   *-_test_start
   .GLOBL   _test_size
   .IF   _test_size>$1000
   .PRINT   "Code size (",/l/x _test_size,") is over $1000"
   .FAIL
   .ENDIF


Here are the results:

Code:
C:\vasm\test>vasmjagrisc_madmac -o test.o test.s
vasm 1.7c (c) in 2002-2015 Volker Barthelmann
vasm Jaguar RISC cpu backend 0.4 (c) 2014-2015 Frank Wille
vasm madmac syntax module 0.1 (c) 2015 Frank Wille
vasm test output module 1.0 (c) 2002 Volker Barthelmann

error 2 in line 335 of "test.s": unknown mnemonic <.68000>
>   .68000

error 31 in line 339 of "test.s": expression must be constant
>   .IF   _test_size>$1000

error 54: non-relocatable expression in equate <_test_size>


My limited understanding is that everything past the .68000 directive is all m68k asm code. For a first run I think this is very very impressive. Have reported these bugs already.

_________________
What came after the Jaguar was the PS1 which for all it's greatness, ushered in corporate development and with it the bleached, repetitive, bland titles which for the most part we're still playing today. - David Wightman


Top
 Profile  
 
PostPosted: Sat Mar 28, 2015 2:18 am 
Offline
Jaguar MOD
User avatar

Joined: Mon Apr 01, 2013 7:09 am
Posts: 820
Quote:


> error 2 in line 335 of "test.s": unknown mnemonic <.68000>
> > .68000

Hmm. From this context I doubt that I can get away with just ignoring
.68000, although there are no M68k instructions in the source.

Seems like .68000 also switches back from absolute assembling (to the
fixed address $f03000) to the .text section. This makes the label
_test_end being placed in .text again.


> error 31 in line 339 of "test.s": expression must be constant
> > .IF _test_size>$1000

It's probably because of:

_test_size .EQU *-_test_start

The difference between two labels is not constant in vasm, as there
may always be optimizations in a later pass. But the parser needs a
constant for conditional assembly in the first pass.

I will look into that.


Regards,

--
Frank Wille

_________________
What came after the Jaguar was the PS1 which for all it's greatness, ushered in corporate development and with it the bleached, repetitive, bland titles which for the most part we're still playing today. - David Wightman


Top
 Profile  
 
PostPosted: Sat Mar 28, 2015 2:19 am 
Offline
Jaguar MOD
User avatar

Joined: Mon Apr 01, 2013 7:09 am
Posts: 820
Quote:
> error 2 in line 335 of "test.s": unknown mnemonic <.68000>
> > .68000
>
> error 31 in line 339 of "test.s": expression must be constant
> > .IF _test_size>$1000

Ok, the reason was that vasm handles ORG differently than madmac/smac.

Usually ORG starts a separate section with a fixed base address
defined by ORG. But in the Jaguar/Madmac case it is more like a RORG,
relocating the labels inside a still active text section.

And .68000 seems to work like ending such a RORG-block.

I wanted to keep the usual function of ORG to make the madmac syntax
module also usable for other CPUs, so I implemented the new option
"-rorg", which changes the function of .ORG and .68000 in the way you
need it.

Your test source assembles now. New snapshot available in 11 hours.

frank@sun ../../vasmjagrisc_madmac -rorg -Faout gcc.s
vasm 1.7c © in 2002-2015 Volker Barthelmann
vasm Jaguar RISC cpu backend 0.4 © 2014-2015 Frank Wille
vasm madmac syntax module 0.2 © 2015 Frank Wille
vasm a.out output module 0.6 © 2008-2015 Frank Wille

.text(acrx2): 640 bytes


Regards,

--
Frank Wille

_________________
What came after the Jaguar was the PS1 which for all it's greatness, ushered in corporate development and with it the bleached, repetitive, bland titles which for the most part we're still playing today. - David Wightman


Top
 Profile  
 
PostPosted: Sat Mar 28, 2015 11:15 pm 
Offline
Jaguar MOD
User avatar

Joined: Mon Apr 01, 2013 7:09 am
Posts: 820
Code:
C:\vasm\test>vasmjagrisc_madmac -rorg -Faout test.s
vasm 1.7c (c) in 2002-2015 Volker Barthelmann
vasm Jaguar RISC cpu backend 0.4 (c) 2014-2015 Frank Wille
vasm madmac syntax module 0.2 (c) 2015 Frank Wille
vasm a.out output module 0.6 (c) 2008-2015 Frank Wille

.text(acrx2):            640 byte

C:\vasm\test>


Attachment:
File comment: Win32 build of vasm_madmac w/syntax module 0.2
vasmjagrisc_madmac.zip [162.32 KiB]
Downloaded 46 times

_________________
What came after the Jaguar was the PS1 which for all it's greatness, ushered in corporate development and with it the bleached, repetitive, bland titles which for the most part we're still playing today. - David Wightman


Top
 Profile  
 
PostPosted: Tue Mar 31, 2015 3:40 am 
Offline
Jaguar MOD
User avatar

Joined: Mon Apr 01, 2013 7:09 am
Posts: 820
Did a quick vasm test on gorf's source file for his jrisc gcc experiments.

Here is the source:

Code:
   MACRO   _RTS
   load   (ST),TMP
   jump   T,(TMP)
   addqt   #4,ST   ;rts
   ENDM

_jag_start::
   .GPU
   .ORG   $F03000
ST   .REGEQU   r31
TMP   .REGEQU   r16
GT   .CCDEF   $15
LE      .CCDEF  $1C

gcc2_compiled_for_madmac:
   ;(.DATA)
   .LONG

_byte_line::
   .DC.L   320
   ;(.TEXT)
   .EVEN

_start::
        movei   #$F04000,ST             ; stack pointer = $F04000
        moveq   #0,r0                   ; r0 = x1
        moveq   #0,r1                   ; r1 = y1
        movei   #_video_buffer,r3       ; r3 = pointer to video_buffer
        movei   #319,r2                 ; r2 = x2
        movei   #1703936,r6             ; r6 = start of video_buffer

        movei   #_Draw_Line,r5          ; r5 = the address of DrawLine fnc.
        store   r6,(r3)                 ; video_buffer points to video start
        moveq   #4,r4                   ; r4 = color
        movei   #239,r3                 ; r3 = y2
        move    PC,TMP                  ; this
        subqt   #4,ST                   ; is
        addqt   #10,TMP                 ; for
        jump    T,(r5)                  ; calling the routine
        store   TMP,(ST)                ; call   r5

        moveq   #6,r2                   ; r2 = pixel color
        movei   #100,r0                 ; r0 = x
        movei   #_Plot_Pixel_Fast,r3    ; r3 = get address of routine
        move    r0,r1                   ; r1 = y
   move   PC,TMP
   subqt   #4,ST
   addqt   #10,TMP
   jump   T,(r3)
   store   TMP,(ST)   ;call   r3

loop:
        jr      T,loop
   nop   ;FOREVER
   .EVEN

_Draw_Line::
        moveq   #0,r6                   ; clear r6 ?(error variable maybe)
        move    r2,r7                   ; r7 = x2
        movei   #_video_buffer,r5       ; r5 = address of video start
        sub     r0,r7                   ; r7 = dx (x2 - x1)
        move    r1,r2                   ; r2 = y1
        load    (r5),r5                 ; r5 = video pointer
        sub     r1,r3                   ; r3 = dy (y2 - y1)
        shlq    #6,r2                   ; (y1<<6)
        add     r5,r2                   ; + vid_start (r2 = vidpos pointer)
        shlq    #8,r1                   ; (y1<<8)
        add     r1,r2                   ; + vid_start (r2 = vidpos pointer) 
        move    r2,r5                   ; put vidpos pointer back in r5
        cmpq    #0,r7                   ; is dx negitive?
        jr      MI,dxp_xim              ; set dx = pos | x_inc = neg
        add     r0,r5                   ; add x1 to vidpos
        jr      T,dxm_xip               ; set dx = neg | x_inc = pos
        moveq   #1,r2                   ; this line sets x_inc = 1 (RISC)
   .EVEN

dxp_xim:
        moveq   #0,r2                   ; r2 = 0
        neg     r7                      ; dx = pos(was neg)
        subqt   #1,r2                   ; x_inc = -1

dxm_xip:
        cmpq    #0,r3                   ; is dy negitive?
        jr      MI,dyp_yim              ; set dy = pos | y_inc = neg
        nop                             ; pad instruction cache
        movei   #_byte_line,r0          ; r0 = points to y_inc
        jr      T,dyn_yip               ; set dy = neg | y_inc = pos
        load    (r0),r0                 ; this line sets r0 = y_inc (RISC)
   .EVEN

dyp_yim:
        movei   #_byte_line,r0          ; r0 point to y_inc
        load    (r0),r0                 ; r0 = y_inc
        neg     r3                      ; dy = pos (was neg)
        neg     r0                      ; y_inc = -320

dyn_yip:
        cmp     r7,r3                   ; is dy(r3) greater dx(r7)
        movei   #do_line_y,TMP          ; set up jump pointer
        jump    GT,(TMP)                ; if so do a y incremented line
        moveq   #0,r1                   ; cache padded instruction
        cmp     r7,r1                   ; is dy(r3) greater than index
        movei   #leave_draw,TMP         ; then we are done and
        jump    GT,(TMP)                ; and we can now
        nop                             ; leave
L31:
        add     r3,r6                   ; error += dy
        cmp     r6,r7                   ; if error>dx
        jr      GT,L32                  ; jump to
        storeb  r4,(r5)                 ; plot line point in video_buffer
        sub     r7,r6                   ; error -= dx
        add     r0,r5                   ; video_pointer += y_inc
L32:
        addqt   #1,r1                   ; index++
        cmp     r1,r7                   ; index(r1) < dx?
        jr      GT,L31                  ; no, then do next pixel
        add     r2,r5                   ; video_pointer += x_inc
        movei   #leave_draw,TMP         ; index is greater so
        jump    T,(TMP)                 ; lets get out
        nop                             ; here
        .EVEN                           

do_line_y:                               
        cmp     r3,r1                   ; index > dy?
        jr      GT,leave_draw           ; if so, later
   nop
L38:                                   
        add     r7,r6                   ; error += dy
        jr      LE,L39                  ; if error goes 0 or neg?
        storeb  r4,(r5)                 ; plot line point to video_buffer
        sub     r3,r6                   ; error -= dy
        add     r2,r5                   ; video_pointer += x_inc
L39:
        addqt   #1,r1                   ; index += 1
        cmp     r1,r3                   ; dx > index?
        jr      GT,L38                  ; if so, continue
        add     r0,r5                   ; video_pointer += y_inc
leave_draw:
   _RTS
   .EVEN

_Plot_Pixel_Fast::
        move    r1,r4                   ; r4 = y
        movei   #_video_buffer,r3       ; r3 = video buffer start address
        shlq    #8,r4                   ; r4 = (y<<8)
        shlq    #6,r1                   ; r1 = (y<<6)
        load    (r3),r3                 ; r3 points to video_start
        add     r1,r4                   ; (y<<8) + (y<<6) +
        add     r0,r3                   ; x +
        add     r4,r3                   ; video_start
        storeb  r2,(r3)                 ; plot the pixel with color in r2
        _RTS                            ; done

_video_buffer::   .DCB.B   8,0
   .LONG
   .68000

_jag_end::
_jag_size      .EQU    _jag_end - _jag_start
      .GLOBL  _jag_size
      .IF     _jag_size>$1000
      .PRINT  "Code size (",/l/x _jag_size,") is over $1000"
   .FAIL
   .ENDIF


VASM assembled it with no errors.

Code:
C:\vasm\test>vasmjagrisc_madmac -rorg -Faout -o jagcircle.o jagcircle.s
vasm 1.7c (c) in 2002-2015 Volker Barthelmann
vasm Jaguar RISC cpu backend 0.4 (c) 2014-2015 Frank Wille
vasm madmac syntax module 0.2 (c) 2015 Frank Wille
vasm a.out output module 0.6 (c) 2008-2015 Frank Wille

.text(acrx2):            280 byte

C:\vasm\test>

_________________
What came after the Jaguar was the PS1 which for all it's greatness, ushered in corporate development and with it the bleached, repetitive, bland titles which for the most part we're still playing today. - David Wightman


Top
 Profile  
 
PostPosted: Mon May 18, 2015 6:52 am 
Offline
Jaguar MOD
User avatar

Joined: Mon Apr 01, 2013 7:09 am
Posts: 820
Jaguar specific updates for vasm 1.7c

    New syntax module: madmac, which tries to be compatible to the Atari Madmac assembler syntax, used for 6502, 68000 and Jaguar RISC targets.
    jagrisc: Added absolute and relative 5 bit source operand relocations. Fixed relative JR relocations.
    jagrisc: Fixed (R14+Rn) addressing mode.
    jagrisc: New JRISC specific directives: GPU, DSP, REGEQU, EQUR, REGUNDEF, EQURUNDEF, CCDEF, CCUNDEF (madmac/smac compatible).
    elf-output: Add support for jagrisc (EM_JAGRISC = 0x9004). This is an unofficial definition done by the Jaguar scene.
    aout-output: Absolute ORG sections are appended to a .text section. Support for Jaguar, which uses MID 0.

The complete list of updates beyond that can be found here:

http://sun.hasenbraten.de/vasm/

_________________
What came after the Jaguar was the PS1 which for all it's greatness, ushered in corporate development and with it the bleached, repetitive, bland titles which for the most part we're still playing today. - David Wightman


Top
 Profile  
 
Display posts from previous:  Sort by  
Post new topic Reply to topic  [ 8 posts ] 

All times are UTC [ DST ]


Who is online

Users browsing this forum: No registered users and 1 guest


You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot post attachments in this forum

Search for:
Jump to:  
cron
Powered by phpBB® Forum Software © phpBB Group