a2audit/audit/notes.org

• Common bytes in FP and INT D0 ROMS
• TODOs and Links to things to test
  • csa2: question on HGR behavior with enhanced 128K Apple //e
  • Check whether c8-cf sticks after reading c3xx but c3 is only active due to the whole c3 ROM switch
• Language card operation
• Floating bus notes
• Printing disassembly notes
• Auxmem tests
• Softswitch tests
• Auxmem with mame
• 80 column details
• apple2e.rom

Common bytes in FP and INT D0 ROMS

D17B 53 D1B3 54 D267 54 D273 54 D307 20 D32A 20 D337 20 D347 20 D51C 00 D57C F0 D665 20 D667 D6 D694 85 D69B FF D6DC B1 D6F7 84 D729 4C D72B D7 D747 F0 D74B D7 D768 85 D76E 65

TODOs and Links to things to test

TODO csa2: question on HGR behavior with enhanced 128K Apple //e

TODO Check whether c8-cf sticks after reading c3xx but c3 is only active due to the whole c3 ROM switch

Language card operation

Bank 2	Bank 1		Action	RAM	Alt Bank 2	Alt Bank 1
C080	C088	R/W	WRTCOUNT=0	READ ENABLE	C084	C08C
C081	C089	R	WRTCOUNT++	READ DISABLE	C085	C08D
		W	WRTCOUNT=0	READ DISABLE
C082	C08A	R/W	WRTCOUNT=0	READ DISABLE	C086	C08E
C083	C08B	R	WRTCOUNT++	READ ENABLE	C087	C08F
		W	WRTCOUNT=0	READ ENABLE

Writing to high RAM is enabled when WRTCOUNT >= 2 Source: UtA2e, pg 5-24.

Floating bus notes

17030 cycles for repeat: 2*5*13*131

UtAIIe: 5-40: Reading video data from a program UtAII: 3-16: Switching screen modes in timed loops

My plan:

• 1. Check that the floating bus works at all
• 1. Save the address at each tick
  • Write the screen full of the lower byte of its address (0x00 -> 0xFF, so we can use 0 for timing)
  • Synch to a run of zeros, then delay to 000000000 counter.
  • Read 17030 values in a fixed-cycle, prime-relative-to-17030 cycles loop, and write to RAM
  • Write the screen full of the upper byte of its address (+ 0x80 if the lower byte is 0, so we can check 0x00->0xFF bytes)
  • Read the upper bytes
  • Check that they all match - should be able to add cycle count, then mod 17030 to find actual cycle number.

17030 * 2 = 34060 (0x850C) If we end at 0xC000, we have to start at 0x3af4. Perhaps we do lower bytes first, then upper? Then we can start at 0x7d7a

Printing disassembly notes

$3A - PCL $3B - PCH Call PCADJ to update with length of currently-pointed-to instruction, then save A to PCL, Y to PCH.

Disassemble an instruction: FE5E (LIST): Set counter to 20, call disassembler 20 times:

FE5E: 20 75 FE 920 LIST JSR A1PC ;MOVE A1 (2 BYTES) TO FE61: A9 14 921 LDA #$14 ; PC IF SPEC'D AND FE63: 48 922 LIST2 PHA ; DISEMBLE 20 INSTRS FE64: 20 D0 F8 923 JSR INSTDSP FE67: 20 53 F9 924 JSR PCADJ ;ADJUST PC EACH INSTR FE6A: 85 3A 925 STA PCL FE6C: 84 3B 926 STY PCH FE6E: 68 927 PLA FE6F: 38 928 SEC FE70: E9 01 929 SBC #$01 ;NEXT OF 20 INSTRS FE72: D0 EF 930 BNE LIST2 FE74: 60 931 RTS FE75: 8A 932 A1PC TXA ;IF USER SPEC'D ADR FE76: F0 07 933 BEQ A1PCRTS ; COPY FROM A1 TO PC FE78: B5 3C 934 A1PCLP LDA A1L,X FE7A: 95 3A 935 STA PCL,X FE7C: CA 936 DEX FE7D: 10 F9 937 BPL A1PCLP FE7F: 60 938 A1PCRTS RTS

F8D0: 20 82 F8 225 INSTDSP JSR INSDS1 ;GEN FMT, LEN BYTES F8D3: 48 226 PHA ;SAVE MNEMONIC TABLE INDEX F8D4: B1 3A 227 PRNTOP LDA (PCL),Y F8D6: 20 DA FD 228 JSR PRBYTE F8D9: A2 01 229 LDX #$01 ;PRINT 2 BLANKS {ACTUALLY JUST 1} F8DB: 20 4A F9 230 PRNTBL JSR PRBL2 F8DE: C4 2F 231 CPY LENGTH ;PRINT INST (1-3 BYTES) F8E0: C8 232 INY ;IN A 12 CHR FIELD F8E1: 90 F1 233 BCC PRNTOP F8E3: A2 03 234 LDX #$03 ;CHAR COUNT FOR MNEMONIC PRINT F8E5: C0 04 235 CPY #$04 F8E7: 90 F2 236 BCC PRNTBL F8E9: 68 237 PLA ;RECOVER MNEMONIC INDEX F8EA: A8 238 TAY F8EB: B9 C0 F9 239 LDA MNEML,Y F8EE: 85 2C 240 STA LMNEM ;FETCH 3-CHAR MNEMONIC F8F0: B9 00 FA 241 LDA MNEMR,Y ; (PACKED IN 2-BYTES) F8F3: 85 2D 242 STA RMNEM … …

F882: A6 3A 180 INSDS1 LDX PCL ;PRINT PCL,H {PRECEDED BY CR, FOLLOWED BY '-'} F884: A4 3B 181 LDY PCH F886: 20 96 FD 182 JSR PRYX2 F889: 20 48 F9 183 JSR PRBLNK ;FOLLOWED BY A BLANK {3 SPACES, FINISH WITH X=0} F88C: A1 3A 184 LDA (PCL,X) ;GET OP CODE F88E: A8 185 INSDS2 TAY … …

300: LDA #0 STA PCL LDA #3 STA PCH LDX #0 LDA (PCL,X) PHA JSR F88E (INSDS2) LDX #3 JSR F8EA JSR F953 STA PCL STY PCH RTS

Auxmem tests

0000-01ff Zero 0200-03ff Main 0400-07ff Text 0800-1fff Main 2000-3fff Hires 4000-bfff Main

Plan: Each test is a set of actual instructions, followed by a jsr, and a list of memory check data. Keep test locations in PCL,PCH. Check data:

• zero/main/text/hires main mem
• zero/main/text/hires aux mem
• zero page current/main/aux: $ff, $100
• main current/main/aux: $200, $3ff, $800, $1fff, $4000, $5fff, $bfff
• text: $400, $7ff
• hires: $2000, $3fff
• CXXX ROM: 1=c300-c3ff, 2=c100-c2ff+c400-c7ff, 4=c800-cffe

For each test:

• Test location is in PCL,PCH.
• JMP to the start of the test (we know where we're coming back to)
• end of test jsr's to CHECK

• CHECK:

  • try to increment all memory locations
  • pull address off the stack, so it knows where the check data is
  • compare with check data

  • if we get a problem: X=index of memory address in a table, Y=desired | high-bit-set=main/unset=aux, A=actual JMP to error routine:

    • restore normal memory configuration
    • save xya
    • disassemble from PCL,PCH to JMP instruction
    • print out memory address, current/main/aux, desired, actual
    • quit test

  • check rom

    • if we see a problem: Y = index of memory location X = ROM-value A = actual X==A if we wanted RAM JMP to error routine 2

      • restore normal memory configuration
      • save xya,carry
      • disassemble from PCL,PCH to JMP instruction

      • want ROM?

        • yes: print out memory address, desired, actual
        • no: print out memory address, non-desired value
      • quit test
• increment PCL,PCH past check data
• if it's pointing at a JMP, we're done.
• otherwise, loop

Softswitch tests

Just switch them on and off, and test each, one at a time.

Auxmem with mame

On entry into AUXMEMTESTS SP: 01FE returned to 02A0

80 column details

Lores80 color differences: UtA2e: 8-29

apple2e.rom

4000-40ff is zeros. I believe this corresponds to C000-C0ff