Apple-2-SW/prog8
1
0
Fork 0
mirror of https://github.com/irmen/prog8.git synced 2025-11-02 13:16:07 +00:00
Code Issues Projects Releases Wiki Activity

fix word multiplication to not clobber r0 and r1 anymore

Browse Source 
This was causing corruption in certain programs such as the cx16/amiga example.
The problem was introduced in 9.4 with the new multiply_words routine
This commit is contained in:
Irmen de Jong
2023-09-02 12:02:07 +02:00
parent 3b786c819d
commit e97303c226
5 changed files with 65 additions and 45 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
codeGenCpu6502/src/prog8/codegen/cpu6502/assignment/AssignmentAsmGen.kt
View File

@@ -543,7 +543,7 @@ internal class AssignmentAsmGen(private val program: PtProgram,
return true return true
} }
in WordDatatypes -> { in WordDatatypes -> {
asmgen.assignWordOperandsToAYAndVar(expr.right, expr.left, "cx16.r0") asmgen.assignWordOperandsToAYAndVar(expr.right, expr.left, "math.multiply_words.multiplier")
asmgen.out(" jsr math.multiply_words") asmgen.out(" jsr math.multiply_words")
assignRegisterpairWord(target, RegisterOrPair.AY) assignRegisterpairWord(target, RegisterOrPair.AY)
return true return true
@@ -567,7 +567,7 @@ internal class AssignmentAsmGen(private val program: PtProgram,
asmgen.out(" jsr math.mul_word_${value}") asmgen.out(" jsr math.mul_word_${value}")
} }
else { else {
asmgen.assignWordOperandsToAYAndVar(expr.right, expr.left, "cx16.r0") asmgen.assignWordOperandsToAYAndVar(expr.right, expr.left, "math.multiply_words.multiplier")
asmgen.out(" jsr math.multiply_words") asmgen.out(" jsr math.multiply_words")
} }
assignRegisterpairWord(target, RegisterOrPair.AY) assignRegisterpairWord(target, RegisterOrPair.AY)

51
codeGenCpu6502/src/prog8/codegen/cpu6502/assignment/AugmentableAssignmentAsmGen.kt
View File

@@ -1335,9 +1335,9 @@ internal class AugmentableAssignmentAsmGen(private val program: PtProgram,
} else { } else {
asmgen.out(""" asmgen.out("""
lda $name lda $name
sta cx16.r0 sta math.multiply_words.multiplier
lda $name+1 lda $name+1
sta cx16.r0+1 sta math.multiply_words.multiplier+1
lda #<$value lda #<$value
ldy #>$value ldy #>$value
jsr math.multiply_words jsr math.multiply_words
@@ -1786,15 +1786,15 @@ internal class AugmentableAssignmentAsmGen(private val program: PtProgram,
} }
"*" -> { "*" -> {
if(valueDt==DataType.UBYTE) { if(valueDt==DataType.UBYTE) {
asmgen.out(" lda $otherName | sta cx16.r0") asmgen.out(" lda $otherName | sta math.multiply_words.multiplier")
if(asmgen.isTargetCpu(CpuType.CPU65c02)) if(asmgen.isTargetCpu(CpuType.CPU65c02))
asmgen.out(" stz cx16.r0+1") asmgen.out(" stz math.multiply_words.multiplier+1")
else else
asmgen.out(" lda #0 | sta cx16.r0+1") asmgen.out(" lda #0 | sta math.multiply_words.multiplier+1")
} else { } else {
asmgen.out(" lda $otherName") asmgen.out(" lda $otherName")
asmgen.signExtendAYlsb(valueDt) asmgen.signExtendAYlsb(valueDt)
asmgen.out(" sta cx16.r0 | sty cx16.r0+1") asmgen.out(" sta math.multiply_words.multiplier | sty math.multiply_words.multiplier+1")
} }
asmgen.out(""" asmgen.out("""
lda $name lda $name
@@ -1930,31 +1930,16 @@ internal class AugmentableAssignmentAsmGen(private val program: PtProgram,
"+" -> asmgen.out(" lda $name | clc | adc $otherName | sta $name | lda $name+1 | adc $otherName+1 | sta $name+1") "+" -> asmgen.out(" lda $name | clc | adc $otherName | sta $name | lda $name+1 | adc $otherName+1 | sta $name+1")
"-" -> asmgen.out(" lda $name | sec | sbc $otherName | sta $name | lda $name+1 | sbc $otherName+1 | sta $name+1") "-" -> asmgen.out(" lda $name | sec | sbc $otherName | sta $name | lda $name+1 | sbc $otherName+1 | sta $name+1")
"*" -> { "*" -> {
if(otherName=="cx16.r0") asmgen.out("""
asmgen.out(""" lda $otherName
lda $name ldy $otherName+1
ldy $name+1 sta math.multiply_words.multiplier
jsr math.multiply_words sty math.multiply_words.multiplier+1
sta $name lda $name
sty $name+1""") ldy $name+1
else if(name=="cx16.r0") jsr math.multiply_words
asmgen.out(""" sta $name
lda $otherName sty $name+1""")
ldy $otherName+1
jsr math.multiply_words
sta $name
sty $name+1""")
else
asmgen.out("""
lda $otherName
ldy $otherName+1
sta cx16.r0
sty cx16.r0+1
lda $name
ldy $name+1
jsr math.multiply_words
sta $name
sty $name+1""")
} }
"/" -> { "/" -> {
if(dt==DataType.WORD) { if(dt==DataType.WORD) {
@@ -2135,8 +2120,8 @@ internal class AugmentableAssignmentAsmGen(private val program: PtProgram,
private fun inplacemodificationWordWithValue(name: String, dt: DataType, operator: String, value: PtExpression) { private fun inplacemodificationWordWithValue(name: String, dt: DataType, operator: String, value: PtExpression) {
fun multiplyVarByWordInAY() { fun multiplyVarByWordInAY() {
asmgen.out(""" asmgen.out("""
sta cx16.r0 sta math.multiply_words.multiplier
sty cx16.r0+1 sty math.multiply_words.multiplier+1
lda $name lda $name
ldy $name+1 ldy $name+1
jsr math.multiply_words jsr math.multiply_words

17
compiler/res/prog8lib/math.asm
View File

@@ -56,9 +56,12 @@ _multiplier = P8ZP_SCRATCH_REG
multiply_words .proc multiply_words .proc
; -- multiply two 16-bit words into a 32-bit result (signed and unsigned) ; -- multiply two 16-bit words into a 32-bit result (signed and unsigned)
; input: A/Y = first 16-bit number, cx16.R0 = second 16-bit number ; input: A/Y = first 16-bit number, multiply_words.multiplier = second 16-bit number
; output: multiply_words.result == cx16.R0:R1, 4-bytes/32-bits product, LSB order (low-to-high) low 16 bits also in AY. ; output: multiply_words.result, 4-bytes/32-bits product, LSB order (low-to-high) low 16 bits also in AY.
; TODO: should not use R0 and R1 at all !!! result needs 4 consecutive bytes, so it can't be in zeropage at all...
; NOTE: the result (which includes the multiplier parameter on entry) is a 4-byte array.
; this routine could be faster if we could stick that into zeropage,
; but there currently is no way to use 4 consecutive bytes in ZP (without disabling irq and saving/restoring them)...
; mult62.a ; mult62.a
; based on Dr Jefyll, http://forum.6502.org/viewtopic.php?f=9&t=689&start=0#p19958 ; based on Dr Jefyll, http://forum.6502.org/viewtopic.php?f=9&t=689&start=0#p19958
@@ -73,9 +76,8 @@ multiply_words .proc
; Average cycles: ; Average cycles:
; 93 bytes ; 93 bytes
_multiplicand = P8ZP_SCRATCH_W1 ; 2 bytes _multiplicand = P8ZP_SCRATCH_W2 ; 2 bytes
_multiplier = cx16.r0 ; 2 bytes multiplier = result
result = cx16.r0 ; 4 bytes (note: shares memory with multiplier) so is r0 and ALSO r1.
; 16 bit x 16 bit unsigned multiply, 32 bit result ; 16 bit x 16 bit unsigned multiply, 32 bit result
; ;
@@ -175,6 +177,9 @@ _inner_loop2
lda result lda result
ldy result+1 ldy result+1
rts rts
result .byte 0,0,0,0
.pend .pend

6
docs/source/todo.rst
View File

@@ -1,10 +1,8 @@
TODO TODO
==== ====
- fix on c64 target: examples/cube3d-float (broken since 9.3 with the evalstack removal) it works on x16 target, oddly enough.
More detailed and simpler code for this problem in floatproblem64.p8
- fix: amiga example with noopt draws wrong lines, caused by "2x faster word multiplication routine" because it trashes r0 and r1 now
multiply_words in math.asm needs fixing.
- fix: test all other things with noopt once again! (examples/c64 are all ok)
- fix: search for TODO("swap operand order") - fix: search for TODO("swap operand order")
- optimize: search for TODO optimize: don't use scratch var - optimize: search for TODO optimize: don't use scratch var
- prefix prog8 subroutines with p8s_ instead of p8_ to not let them clash with variables in the asm? - prefix prog8 subroutines with p8s_ instead of p8_ to not let them clash with variables in the asm?

32
floatproblem64.p8 Normal file
View File

@@ -0,0 +1,32 @@
%import textio
%import floats
%zeropage dontuse
main {
sub start() {
float value1 = -0.8
float value2 = 0.3
float two = 2.0
float result = value1*two + value2*two ; TODO FIX: invalid result on c64, ok when the *two is removed or expression is split (it's not caused by pushFAC1/popFAC1)
floats.print_f(result)
txt.nl()
txt.print("-1 was expected\n\n") ; on C64: -1.1 is printed :(
result = value2*two + value1*two ; swapped operands around, now it's suddenly fine on C64...
floats.print_f(result)
txt.nl()
txt.print("-1 was expected\n\n") ; on C64: correct value is printed
value1 = 0.8
value2 = 0.3
result = value1*two + value2*two
floats.print_f(result)
txt.nl()
txt.print("2.2 was expected\n\n") ; on C64: correct value is printed
repeat {
}
}
}