This commit is contained in:
Irmen de Jong 2021-01-02 17:50:08 +01:00
parent edd3a22848
commit baf9dfb46c
4 changed files with 248 additions and 62 deletions

View File

@ -2,6 +2,8 @@
TODO
====
- add syslib.rdtim16() function to wrap the RDTIM to read only the lower 16 bits of the jiffyclock - fix examples.
- move all str* builtin functions to a str library module
- detect variables that are written but never read - mark those as unused too and remove them, such as uword unused = memory("unused222", 20) - also remove the memory slab allocation
- hoist all variable declarations up to the subroutine scope *before* even the constant folding takes place (to avoid undefined symbol errors when referring to a variable from another nested scope in the subroutine)
- make it possible to use cpu opcodes such as 'nop' as variable names by prefixing all asm vars with something such as '_'

View File

@ -6,42 +6,116 @@
main {
sub start() {
ubyte instr_opcode=parse_mnemonic("lda", 3)
if_cc
txt.print("not found")
else
txt.print_ubhex(instr_opcode, 1)
str[16] addr_modes = [ "Imp", "Acc", "Imm", "Zp", "ZpX", "ZpY", "Rel", "Abs", "AbsX", "AbsY", "Ind", "IzX", "IzY", "Zpr", "Izp", "IaX" ]
txt.lowercase()
txt.print("\nAssembler.\n\n")
txt.print("Addressing mode ids:\n")
ubyte mode_id = 0
uword mode
for mode in addr_modes {
mode_id++
txt.print_ub(mode_id)
txt.chrout('.')
txt.chrout(' ')
txt.print(mode)
txt.chrout('\n')
}
txt.chrout('\n')
instr_opcode=parse_mnemonic("rts", 1)
if_cc
txt.print("not found")
else
txt.print_ubhex(instr_opcode, 1)
;user_input()
benchmark()
; test_stack.test()
}
sub benchmark() {
str[20] instructions = ["lda", "ldx", "ldy", "jsr", "bcs", "rts", "lda", "ora", "and", "eor", "wai", "nop", "wai", "nop", "wai", "nop", "wai", "nop", "wai", "nop"]
ubyte[20] modes = [3, 4, 8, 8, 7, 1, 12, 13, 5, 4, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1]
uword valid = 0
const uword iterations = 40000 / len(instructions)
const uword amount = iterations * len(instructions)
txt.print("Benchmark.\nMatching ")
txt.print_uw(amount)
txt.print(" mnemonics")
uword current_time = 0
c64.SETTIM(0,0,0)
uword total = 0
repeat iterations {
if lsb(total)==0
txt.chrout('.')
ubyte idx
for idx in 0 to len(instructions)-1 {
ubyte instr_opcode=parse_mnemonic(instructions[idx], modes[idx])
if_cs
valid++
total++
}
}
; read clock back
%asm {{
phx
jsr c64.RDTIM
sta current_time
stx current_time+1
plx
}}
txt.print("\nDone.\nValid: ")
txt.print_uw(valid)
txt.print("\ninvalid: ")
txt.print_uw(amount-valid)
txt.print("\ntotal: ")
txt.print_uw(total)
txt.print("\nSeconds:")
uword secs = current_time / 60
current_time = (current_time - secs*60)*1000/60
txt.print_uw(secs)
txt.chrout('.')
if current_time<10
txt.chrout('0')
if current_time<100
txt.chrout('0')
txt.print_uw(current_time)
txt.chrout('\n')
}
instr_opcode=parse_mnemonic("rts", 0)
if_cc
txt.print("not found")
else
txt.print_ubhex(instr_opcode, 1)
txt.chrout('\n')
sub user_input() {
repeat {
str input_mnem = "????????"
str input_am = "????"
ubyte input_length = 0
while input_length<5 {
txt.print("\nmnemonic,addrmode? ")
input_length = txt.input_chars(input_mnem)
}
txt.chrout('\n')
; if not instr_opcodes {
; txt.print("invalid instruction\n")
; } else {
; ubyte opcode
; if @(instr_opcodes) {
; ; instruction has only a single addressing mode and opcode
; opcode = @(instr_opcodes+1)
; } else {
; ; instruction has multiple addressing modes and possible opcodes
; opcode = @(instr_opcodes+addr_mode)
; }
; txt.print_ubhex(opcode,1)
; }
uword addrmode_str = strfind(input_mnem, ',')
if addrmode_str {
;@(addrmode) = 0
;addrmode++
ubyte addr_mode = conv.str2ubyte(addrmode_str+1)
ubyte instr_opcode=parse_mnemonic(input_mnem, addr_mode)
if_cc
txt.print("-> invalid\n")
else {
txt.print("-> opcode: ")
txt.print_ubhex(instr_opcode, 1)
txt.chrout('\n')
}
} else {
txt.print("??\n")
}
test_stack.test()
}
}
inline asmsub parse_mnemonic(uword mnemonic_ptr @AY, ubyte addr_mode @X) clobbers(X) -> ubyte @A, ubyte @Pc {

View File

@ -1,3 +1,4 @@
from collections import Counter
from enum import IntEnum
@ -20,7 +21,7 @@ class AddrMode(IntEnum):
IaX = 16
Instructions = (
AllInstructions = [
(0x00, "brk", AddrMode.Imp),
(0x01, "ora", AddrMode.IzX),
(0x02, "nop", AddrMode.Imm),
@ -277,7 +278,12 @@ Instructions = (
(0xfd, "sbc", AddrMode.AbsX),
(0xfe, "inc", AddrMode.AbsX),
(0xff, "bbs7", AddrMode.Zpr)
)
]
# NOP is weird, it is all over the place.
# For the 'common' immediate NOP, keep only the $EA opcode (this was the original NOP on the 6502)
Instructions = [ins for ins in AllInstructions if ins[1] != "nop"] + [(0xea, "nop", AddrMode.Imp)]
InstructionsByName = {}
for ins in Instructions:
@ -329,30 +335,82 @@ for instr in sorted(InstructionsByName.items()):
print()
mnemonics = list(sorted(set(ins[1] for ins in Instructions)))
def determine_mnemonics():
mnemonics = list(sorted(set(ins[1] for ins in Instructions)))
# opcodes histogram (ordered by occurrence) (in kernal + basic roms of the c64):
opcode_occurrences = [
(32, 839), (133, 502), (165, 488), (0, 429), (208, 426), (169, 390), (76, 324), (240, 322), (2, 314), (160, 245),
(96, 228), (3, 201), (1, 191), (255, 186), (144, 182), (170, 175), (162, 169), (177, 165), (104, 159), (164, 158),
(132, 157), (201, 156), (72, 151), (141, 150), (200, 146), (173, 144), (166, 139), (176, 139), (16, 138),
(134, 138), (73, 127), (24, 119), (101, 113), (69, 109), (13, 107), (34, 104), (145, 103), (4, 102), (168, 101),
(221, 98), (230, 93), (48, 91), (189, 87), (41, 86), (6, 86), (9, 86), (8, 85), (79, 85), (138, 80), (10, 80),
(7, 79), (185, 77), (56, 75), (44, 75), (78, 74), (105, 73), (5, 73), (174, 73), (220, 71), (198, 69), (232, 69),
(36, 69), (202, 67), (152, 67), (95, 67), (100, 65), (102, 65), (247, 65), (188, 64), (136, 64), (84, 64),
(122, 62), (128, 61), (80, 61), (186, 60), (82, 59), (97, 58), (15, 57), (70, 57), (229, 56), (19, 55), (40, 54),
(183, 54), (65, 54), (233, 53), (180, 53), (12, 53), (171, 53), (197, 53), (83, 52), (248, 52), (112, 51),
(237, 51), (89, 50), (11, 50), (158, 50), (74, 49), (224, 48), (20, 47), (238, 47), (108, 46), (234, 46),
(251, 46), (254, 46), (184, 45), (14, 44), (163, 44), (226, 43), (211, 43), (88, 43), (98, 42), (17, 42),
(153, 42), (243, 41), (228, 41), (99, 41), (253, 41), (209, 41), (187, 39), (123, 39), (67, 39), (196, 38),
(68, 38), (35, 38), (172, 38), (175, 38), (161, 38), (85, 38), (191, 37), (113, 37), (182, 37), (151, 37),
(71, 36), (181, 35), (214, 35), (121, 35), (157, 35), (178, 35), (77, 35), (42, 34), (212, 33), (18, 33),
(127, 33), (241, 33), (21, 33), (249, 32), (23, 31), (245, 30), (142, 30), (55, 29), (140, 29), (46, 29),
(192, 29), (179, 29), (252, 29), (115, 29), (22, 29), (43, 28), (215, 28), (45, 28), (246, 28), (38, 28),
(86, 27), (225, 27), (25, 26), (239, 26), (58, 26), (167, 26), (147, 26), (217, 26), (149, 25), (30, 25),
(206, 25), (28, 24), (47, 24), (37, 24), (155, 24), (129, 23), (148, 23), (111, 23), (29, 23), (39, 23),
(51, 22), (193, 22), (236, 22), (120, 22), (64, 22), (204, 21), (210, 21), (244, 21), (52, 21), (66, 21),
(114, 20), (250, 20), (106, 20), (93, 19), (199, 19), (218, 19), (154, 19), (205, 19), (50, 19), (159, 19),
(194, 19), (49, 19), (190, 19), (103, 18), (216, 18), (213, 18), (107, 18), (131, 18), (63, 18), (94, 18),
(91, 17), (242, 17), (109, 17), (53, 16), (227, 16), (139, 16), (31, 16), (75, 16), (60, 16), (195, 15),
(231, 15), (62, 15), (59, 15), (87, 14), (207, 14), (27, 14), (90, 14), (110, 13), (223, 13), (57, 13),
(118, 12), (26, 12), (203, 12), (81, 12), (156, 12), (54, 12), (235, 12), (146, 11), (135, 11), (126, 11),
(150, 11), (130, 11), (143, 10), (61, 10), (219, 10), (124, 9), (222, 9), (125, 9), (119, 7), (137, 7),
(33, 7), (117, 5), (92, 4), (116, 3)
]
cnt = Counter()
for opcode, amount in opcode_occurrences:
cnt[AllInstructions[opcode][1]] += amount
cnt["nop"] = 13
cnt["tsb"] = 13
four_letter_mnemonics = list(sorted([ins[1] for ins in AllInstructions if len(ins[1])>3]))
for ins4 in four_letter_mnemonics:
del cnt[ins4]
cnt[ins4] = 1
mnem2 = [c[0] for c in cnt.most_common()]
if len(mnem2)!=len(mnemonics):
raise ValueError("mnem count mismatch")
return mnem2
def first_letters(mnem):
return list(sorted(set(m[0] for m in mnemonics)))
mnemonics = determine_mnemonics()
def first_letters():
firstletters = {m[0]: 0 for m in mnemonics}
return firstletters.keys()
def second_letters(firstletter):
return list(sorted(set(m[1] for m in mnemonics if m[0] == firstletter)))
secondletters = {m[1]: 0 for m in mnemonics if m[0] == firstletter}
return secondletters.keys()
def third_letters(firstletter, secondletter):
return list(sorted(set(m[2] for m in mnemonics if m[0] == firstletter and m[1] == secondletter)))
thirdletters = {m[2]: 0 for m in mnemonics if m[0] == firstletter and m[1] == secondletter}
return thirdletters.keys()
def fourth_letters(firstletter, secondletter, thirdletter):
longmnem = [m for m in mnemonics if len(m) > 3]
return list(
sorted(set(m[3] for m in longmnem if m[0] == firstletter and m[1] == secondletter and m[2] == thirdletter)))
fourthletters = {m[3]: 0 for m in longmnem if m[0] == firstletter and m[1] == secondletter and m[2] == thirdletter}
return fourthletters.keys()
def make_tree():
tree = {}
for first in first_letters(mnemonics):
for first in first_letters():
tree[first] = {
secondletter: {
thirdletter: {
@ -379,15 +437,19 @@ parse_mnemonic_asm .proc
sty P8ZP_SCRATCH_W1+1
ldy #0
lda (P8ZP_SCRATCH_W1),y
and #$7f ; lowercase
pha
iny
lda (P8ZP_SCRATCH_W1),y
and #$7f ; lowercase
pha
iny
lda (P8ZP_SCRATCH_W1),y
and #$7f ; lowercase
pha
iny
lda (P8ZP_SCRATCH_W1),y
and #$7f ; lowercase
sta cx16.r4
pla
tay
@ -415,11 +477,27 @@ parse_mnemonic_asm .proc
bne _not_found
iny
lda (P8ZP_SCRATCH_W2),y
bra _valid
bra _checkvalid
_found_multi
ldy cx16.r15 ; addressing mode index
lda (P8ZP_SCRATCH_W2),y
_checkvalid
bne _valid
; check if it is valid BRK $00
ldy #0
lda (P8ZP_SCRATCH_W1),y
cmp #'b'
bne _not_found
iny
lda (P8ZP_SCRATCH_W1),y
cmp #'r'
bne _not_found
iny
lda (P8ZP_SCRATCH_W1),y
cmp #'k'
bne _not_found
lda #0
_valid sec
rts
@ -442,6 +520,7 @@ for first in tree:
print(" cpy #'%s'" % third)
print(" bne _not_%s%s%s" % (first, second, third))
if tree[first][second][third]:
# TODO OPTIMIZE CODE FOR THE FOUR LETTER MNEMONICS (the fourth letter is always a digit 0-9)
for fourth in tree[first][second][third]:
print(" pha")
print(" lda _opcode_fourth_letter")

View File

@ -7,30 +7,61 @@
main {
sub start () {
uword uw
ubyte ub
float ff
uword current_time
uword secs
const uword test_value = 1261
uw = funcuw()
ub = funcub()
ff = funcff()
}
current_time = test_value
secs = current_time / 60
current_time = (current_time - secs*60)*1000/60
txt.print_uw(secs)
txt.chrout('.')
if current_time<10
txt.chrout('0')
if current_time<100
txt.chrout('0')
txt.print_uw(current_time)
txt.chrout('\n')
sub funcuw() -> uword {
uword a=1111
uword b=2222
return a+b
}
current_time = test_value
secs = current_time / 60
current_time = current_time - secs*60
current_time *= 1000
current_time /= 60
txt.print_uw(secs)
txt.chrout('.')
if current_time<10
txt.chrout('0')
if current_time<100
txt.chrout('0')
txt.print_uw(current_time)
txt.chrout('\n')
sub funcub() -> ubyte {
ubyte a=11
ubyte b=22
return a+b
}
sub funcff() -> float {
float a=1111.1
float b=2222.2
return a+b
; found = strfind("irmen de jong", ' ')
; txt.print_uwhex(found, 1)
; txt.chrout('\n')
; found = strfind(" irmen-de-jong", ' ')
; txt.print_uwhex(found, 1)
; txt.chrout('\n')
; found = strfind("irmen-de-jong ", ' ')
; txt.print_uwhex(found, 1)
; txt.chrout('\n')
; found = strfind("irmen-de-jong", ' ')
; txt.print_uwhex(found, 1)
; txt.chrout('\n')
; found = strfinds("irmen de jong", "de")
; txt.print_uwhex(found, 1)
; txt.chrout('\n')
; found = strfinds("irmen de jong", "irmen")
; txt.print_uwhex(found, 1)
; txt.chrout('\n')
; found = strfinds("irmen de jong", "jong")
; txt.print_uwhex(found, 1)
; txt.chrout('\n')
; found = strfinds("irmen de jong", "de456")
; txt.print_uwhex(found, 1)
; txt.chrout('\n')
}
}