belka/elna
1
0
Fork 0
Code Issues 5 Pull Requests Releases Activity

Fix strings search looping

Browse Source
This commit is contained in:
Eugen Wissner 2025-05-04 23:49:39 +02:00
parent 0a0bc4e1f2
commit df1c0486c5
Signed by: belka
GPG Key ID: A27FDC1E8EE902C0
4 changed files with 243 additions and 350 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

5
boot/common-boot.s
View File

@ -425,6 +425,9 @@ _memcpy:
addi a0, a0, 1
addi a1, a1, 1
addi a2, a2, -1
j .Lmemcpy_loop
.Lmemcpy_end:
mv a0, t0
@ -473,6 +476,8 @@ _strings_index:
beqz a0, .Lstrings_index_end
.Lstrings_index_next:
# Advance the pointer, reduce the length.
lw a2, (s2)
addi s2, s2, 4
add s2, s2, a2
addi s1, s1, -1

39
boot/definitions.inc
View File

@ -19,34 +19,35 @@
.equ TOKEN_END, 13
.equ TOKEN_TYPE, 14
.equ TOKEN_RECORD, 15
.equ TOKEN_RECORD, 16
.equ TOKEN_UNION, 16
.equ TOKEN_TRUE, 17
.equ TOKEN_FASE, 18
.equ TOKEN_FALSE, 18
.equ TOKEN_NIL, 19
.equ TOKEN_XOR, 20
.equ TOKEN_OR, 21
.equ TOKEN_RETURN, 22
.equ TOKEN_CAST, 23
.equ TOKEN_DEFER, 24
.equ TOKEN_GOTO, 24
.equ TOKEN_CASE, 25
.equ TOKEN_OF, 26
.equ TOKEN_IDENTIFIER, 27
# The constant should match the character index in the byte_keywords string.
.equ TOKEN_AND, 27
.equ TOKEN_DOT, 28
.equ TOKEN_COMMA, 29
.equ TOKEN_COLON, 30
.equ TOKEN_SEMICOLON, 31
.equ TOKEN_LEFT_PAREN, 32
.equ TOKEN_RIGHT_PAREN, 33
.equ TOKEN_LEFT_BRACKET, 34
.equ TOKEN_RIGHT_BRACKET, 35
.equ TOKEN_HAT, 36
.equ TOKEN_EQUALS, 37
.equ TOKEN_PLUS, 38
.equ TOKEN_MINUS, 39
.equ TOKEN_ASTERISK, 40
.equ TOKEN_AT, 41
.equ TOKEN_AND, 28
.equ TOKEN_DOT, 29
.equ TOKEN_COMMA, 30
.equ TOKEN_COLON, 31
.equ TOKEN_SEMICOLON, 32
.equ TOKEN_LEFT_PAREN, 33
.equ TOKEN_RIGHT_PAREN, 34
.equ TOKEN_LEFT_BRACKET, 35
.equ TOKEN_RIGHT_BRACKET, 36
.equ TOKEN_HAT, 37
.equ TOKEN_EQUALS, 38
.equ TOKEN_PLUS, 39
.equ TOKEN_MINUS, 40
.equ TOKEN_ASTERISK, 41
.equ TOKEN_AT, 42
.equ TOKEN_ASSIGN, 42
.equ TOKEN_ASSIGN, 43

484
boot/stage1.s
View File

@ -53,10 +53,12 @@ asm_type_object: .ascii ", @object\n"
asm_restore_parameters:
.ascii "lw a0, 60(sp)\nlw a1, 56(sp)\nlw a2, 52(sp)\nlw a3, 48(sp)\nlw a4, 44(sp)\nlw a5, 40(sp)\n"
.equ ASM_RESTORE_PARAMETERS_SIZE, . - asm_restore_parameters
asm_preserve_parameters:
.ascii "sw a0, 84(sp)\nsw a1, 80(sp)\nsw a2, 76(sp)\nsw a2, 76(sp)\nsw a3, 72(sp)\nsw a4, 68(sp)\nsw a5, 64(sp)\n"
.equ ASM_PRESERVE_PARAMETERS_SIZE, . - asm_preserve_parameters
.section .bss
.type source_code, @object
.size source_code, SOURCE_BUFFER_SIZE
source_code: .zero SOURCE_BUFFER_SIZE
.section .text
@ -65,10 +67,10 @@ source_code: .zero SOURCE_BUFFER_SIZE
.type _compile_import, @function
_compile_import:
# Prologue.
addi sp, sp, -16
sw ra, 12(sp)
sw s0, 8(sp)
addi s0, sp, 16
addi sp, sp, -24
sw ra, 20(sp)
sw s0, 16(sp)
addi s0, sp, 24
.Lcompile_import_loop:
call _skip_comment
@ -89,9 +91,9 @@ _compile_import:
.Lcompile_import_end:
# Epilogue.
lw ra, 12(sp)
lw s0, 8(sp)
addi sp, sp, 16
lw ra, 20(sp)
lw s0, 16(sp)
addi sp, sp, 24
ret
.type _build_binary_expression, @function
@ -107,9 +109,9 @@ _build_binary_expression:
call _skip_spaces
mv a0, s1
addi a1, sp, 16
addi a1, sp, 12
call _tokenize_next
lw t0, 16(sp)
lw t0, 12(sp)
li t1, TOKEN_AND
beq t0, t1, .L_build_binary_expression_and
@ -146,11 +148,6 @@ _build_binary_expression:
j .Lbuild_binary_expression_end
.L_build_binary_expression_and:
/* DEBUG
addi a0, s1, 0
li a1, 4
call _write_error */
mv s1, a0 # Skip &.
li a0, 1
call _build_expression
@ -385,25 +382,29 @@ _compile_identifier_expression:
.type _build_expression, @function
_build_expression:
# Prologue.
addi sp, sp, -40
sw ra, 36(sp)
sw s0, 32(sp)
addi s0, sp, 40
addi sp, sp, -48
sw ra, 44(sp)
sw s0, 40(sp)
addi s0, sp, 48
addi a0, a0, '0' # Make the register number to a character.
sw a0, 28(sp) # And save it.
sw a0, 36(sp) # And save it.
mv a0, s1
addi a1, sp, 24
call _tokenize_next
call _skip_spaces
call _read_token
sw s1, 24(sp)
sw a0, 20(sp)
sw s1, 32(sp)
sw a0, 28(sp)
lbu a0, (s1)
li t0, '-'
lw a0, 24(sp)
li t0, TOKEN_MINUS
beq a0, t0, .Lbuild_expression_negate
lbu a0, (s1)
li t0, '@'
li t0, TOKEN_AT
beq a0, t0, .Lbuild_expression_address
lbu a0, (s1)
@ -414,8 +415,8 @@ _build_expression:
li t0, '_'
beq a0, t0, .Lbuild_expression_call
lw a0, 20(sp)
lw a1, 28(sp)
lw a0, 28(sp)
lw a1, 36(sp)
call _compile_identifier_expression
j .Lbuild_expression_advance
@ -432,7 +433,7 @@ _build_expression:
j .Lbuild_expression_advance
.Lbuild_expression_address:
lw t1, 28(sp)
lw t1, 36(sp)
li t0, 0x20 # _
sw t0, 16(sp)
li t0, 0x2c707320 # _sp,
@ -449,11 +450,11 @@ _build_expression:
addi s1, s1, 1 # Skip @.
call _skip_spaces
call _read_token
sw s1, 24(sp)
sw a0, 20(sp)
sw s1, 32(sp)
sw a0, 28(sp)
lw a0, 24(sp)
lw a1, 20(sp)
lw a0, 32(sp)
lw a1, 28(sp)
addi a0, a0, 4 # Skip the "loca" variable prefix.
addi a1, a1, -4 # Skip the "loca" variable prefix.
call _write_out
@ -464,8 +465,8 @@ _build_expression:
j .Lbuild_expression_advance
.Lbuild_expression_call:
lw a0, 24(sp)
lw a1, 20(sp)
lw a0, 32(sp)
lw a1, 28(sp)
add s1, s1, a1
addi s1, s1, 1
call _compile_call
@ -473,7 +474,7 @@ _build_expression:
j .Lbuild_expression_end
.Lbuild_expression_literal:
lw t1, 28(sp)
lw t1, 36(sp)
li t0, 0x00202c00 # \0,_
or t0, t0, t1
sw t0, 16(sp)
@ -483,8 +484,8 @@ _build_expression:
li a1, 7
call _write_out
lw a0, 24(sp)
lw a1, 20(sp)
lw a0, 32(sp)
lw a1, 28(sp)
call _write_out
li a0, '\n'
@ -493,14 +494,14 @@ _build_expression:
j .Lbuild_expression_advance
.Lbuild_expression_advance:
lw a0, 20(sp)
lw a0, 28(sp)
add s1, s1, a0
.Lbuild_expression_end:
# Epilogue.
lw ra, 36(sp)
lw s0, 32(sp)
addi sp, sp, 40
lw ra, 44(sp)
lw s0, 40(sp)
addi sp, sp, 48
ret
# Compiles an lvalue.
@ -595,34 +596,20 @@ _compile_identifier:
sw s0, 24(sp)
addi s0, sp, 32
call _read_token
# Save the pointer to the identifier and its length on the stack.
sw s1, 20(sp)
sw a0, 16(sp)
mv a0, s1
addi a1, sp, 12
call _tokenize_next
addi a1, sp, 0
call _tokenize_next
mv s1, a0
add s1, s1, a0
call _skip_spaces
call _read_token
lw t0, 0(sp)
# Save the pointer and the length of the token following the identifier.
sw s1, 12(sp)
sw a0, 8(sp)
li t1, TOKEN_ASSIGN
beq t0, t1, .Lcompile_identifier_assign
add s1, s1, a0 # Skip that token.
call _skip_spaces
li t0, 0x3d3a # :=
sw t0, 4(sp)
lw a0, 12(sp)
lw a1, 8(sp)
addi a2, sp, 4
call _token_compare
beqz a0, .Lcompile_identifier_assign
lw t0, 12(sp)
lbu t0, (t0)
li t1, 0x28 # (
li t1, TOKEN_LEFT_PAREN
beq t0, t1, .Lcompile_identifier_call
j .Lcompile_identifier_end
@ -917,10 +904,10 @@ _skip_comment:
.type _compile_procedure_section, @function
_compile_procedure_section:
# Prologue.
addi sp, sp, -16
sw ra, 12(sp)
sw s0, 8(sp)
addi s0, sp, 16
addi sp, sp, -24
sw ra, 20(sp)
sw s0, 16(sp)
addi s0, sp, 24
.Lcompile_procedure_section_loop:
call _skip_spaces
@ -928,10 +915,10 @@ _compile_procedure_section:
call _skip_spaces
mv a0, s1
addi a1, sp, 0
addi a1, sp, 4
call _tokenize_next
li t0, TOKEN_PROC
lw t1, 0(sp)
lw t1, 4(sp)
bne t0, t1, .Lcompile_procedure_section_end
call _compile_procedure
@ -940,18 +927,18 @@ _compile_procedure_section:
.Lcompile_procedure_section_end:
# Epilogue.
lw ra, 12(sp)
lw s0, 8(sp)
addi sp, sp, 16
lw ra, 20(sp)
lw s0, 16(sp)
addi sp, sp, 24
ret
.type _compile_module_declaration, @function
_compile_module_declaration:
# Prologue.
addi sp, sp, -16
sw ra, 12(sp)
sw s0, 8(sp)
addi s0, sp, 16
addi sp, sp, -24
sw ra, 20(sp)
sw s0, 16(sp)
addi s0, sp, 24
la a0, global_start
li a1, GLOBAL_START_SIZE
@ -960,32 +947,32 @@ _compile_module_declaration:
# Skip "program".
call _skip_comment
mv a0, s1
addi a1, sp, 0
addi a1, sp, 4
call _tokenize_next
mv s1, a0
# Epilogue.
lw ra, 12(sp)
lw s0, 8(sp)
addi sp, sp, 16
lw ra, 20(sp)
lw s0, 16(sp)
addi sp, sp, 24
ret
.type _compile_constant_section, @function
_compile_constant_section:
# Prologue.
addi sp, sp, -16
sw ra, 12(sp)
sw s0, 8(sp)
addi s0, sp, 16
addi sp, sp, -24
sw ra, 20(sp)
sw s0, 16(sp)
addi s0, sp, 24
call _skip_comment
call _skip_spaces
mv a0, s1
addi a1, sp, 0
addi a1, sp, 4
call _tokenize_next
li t0, TOKEN_CONST
lw t1, 0(sp)
lw t1, 4(sp)
bne t0, t1, .Lcompile_constant_section_end
mv s1, a0
@ -1004,21 +991,21 @@ _compile_constant_section:
.Lcompile_constant_section_end:
# Epilogue.
lw ra, 12(sp)
lw s0, 8(sp)
addi sp, sp, 16
lw ra, 20(sp)
lw s0, 16(sp)
addi sp, sp, 24
ret
.type _compile_constant, @function
_compile_constant:
# Prologue.
addi sp, sp, -16
sw ra, 12(sp)
sw s0, 8(sp)
addi s0, sp, 16
addi sp, sp, -24
sw ra, 20(sp)
sw s0, 16(sp)
addi s0, sp, 24
mv a0, s1
addi a1, sp, 0
addi a1, sp, 4
call _tokenize_next
sub a1, a0, s1 # The identifier end from _tokenize_next should be in a0.
@ -1027,7 +1014,7 @@ _compile_constant:
call _write_out
mv a0, s1
addi a1, sp, 0
addi a1, sp, 4
call _tokenize_next
mv s1, a0 # Skip the assignment sign.
@ -1052,24 +1039,24 @@ _compile_constant:
call _put_char
# Epilogue.
lw ra, 12(sp)
lw s0, 8(sp)
addi sp, sp, 16
lw ra, 20(sp)
lw s0, 16(sp)
addi sp, sp, 24
ret
.type _compile_variable_section, @function
_compile_variable_section:
# Prologue.
addi sp, sp, -16
sw ra, 12(sp)
sw s0, 8(sp)
addi s0, sp, 16
addi sp, sp, -24
sw ra, 20(sp)
sw s0, 16(sp)
addi s0, sp, 24
mv a0, s1
addi a1, sp, 0
addi a1, sp, 4
call _tokenize_next
li t0, TOKEN_VAR
lw t1, 0(sp)
lw t1, 4(sp)
bne t0, t1, .Lcompile_variable_section_end
mv s1, a0
@ -1088,24 +1075,24 @@ _compile_variable_section:
.Lcompile_variable_section_end:
# Epilogue.
lw ra, 12(sp)
lw s0, 8(sp)
addi sp, sp, 16
lw ra, 20(sp)
lw s0, 16(sp)
addi sp, sp, 24
ret
.type _compile_variable, @function
_compile_variable:
# Prologue.
addi sp, sp, -40
sw ra, 36(sp)
sw s0, 32(sp)
addi s0, sp, 40
addi sp, sp, -48
sw ra, 44(sp)
sw s0, 40(sp)
addi s0, sp, 48
call _read_token
# Save the identifier on the stack since it should emitted multiple times.
sw s1, 28(sp)
sw a0, 24(sp)
sw s1, 36(sp)
sw a0, 32(sp)
add s1, s1, a0
call _skip_spaces
@ -1117,8 +1104,8 @@ _compile_variable:
call _read_token
# Save the array size on the stack since it has to be emitted multiple times.
sw s1, 20(sp)
sw a0, 16(sp)
sw s1, 28(sp)
sw a0, 24(sp)
add s1, s1, a0
call _skip_spaces
@ -1133,8 +1120,8 @@ _compile_variable:
li a1, ASM_TYPE_SIZE
call _write_out
lw a0, 28(sp)
lw a1, 24(sp)
lw a0, 36(sp)
lw a1, 32(sp)
call _write_out
la a0, asm_type_object
@ -1142,29 +1129,29 @@ _compile_variable:
call _write_out
# identifier: .zero size
lw a0, 28(sp)
lw a1, 24(sp)
lw a0, 36(sp)
lw a1, 32(sp)
call _write_out
li t0, 0x206f7265 # ero_
sw t0, 12(sp)
sw t0, 20(sp)
li t0, 0x7a2e203a # : .z
sw t0, 8(sp)
addi a0, sp, 8
sw t0, 16(sp)
addi a0, sp, 16
li a1, 8
call _write_out
lw a0, 20(sp)
lw a1, 16(sp)
lw a0, 28(sp)
lw a1, 24(sp)
call _write_out
li a0, '\n'
call _put_char
# Epilogue.
lw ra, 36(sp)
lw s0, 32(sp)
addi sp, sp, 40
lw ra, 44(sp)
lw s0, 40(sp)
addi sp, sp, 48
ret
.type _compile_procedure, @function
@ -1175,11 +1162,12 @@ _compile_procedure:
sw s0, 24(sp)
addi s0, sp, 32
addi s1, s1, 5 # Skip proc_
call _read_token
sw s1, 20(sp)
sw a0, 16(sp)
add s1, s1, a0
mv a0, s1
addi a1, sp, 12
call _tokenize_next # Skip proc.
addi a1, sp, 12
call _tokenize_next
mv s1, a0
# .type identifier, @function
la a0, asm_type
@ -1204,29 +1192,16 @@ _compile_procedure:
li a1, 2
call _write_out
call _skip_spaces
addi s1, s1, 1 # Skip opening argument paren.
call _skip_spaces
addi s1, s1, 1 # Skip closing argument paren.
li t0, 0x6e # n
sw t0, 12(sp)
li t0, 0x69676562 # begi
sw t0, 8(sp)
# Skip all declarations until we find the "begin" keyword, denoting the
# beginning of the procedure body.
.Lcompile_procedure_begin:
call _skip_spaces
call _read_token
mv a1, a0
mv a0, s1
addi a2, sp, 8
add s1, s1, a1
call _token_compare
bnez a0, .Lcompile_procedure_begin
addi a1, sp, 4
call _tokenize_next
mv s1, a0
lw t0, 4(sp)
li t1, TOKEN_BEGIN
bne t0, t1, .Lcompile_procedure_begin
# Generate the procedure prologue with a predefined stack size.
la a0, prologue
@ -1234,56 +1209,8 @@ _compile_procedure:
call _write_out
# Save passed arguments on the stack.
li t0, 0x0a29 # )\n
sw t0, 12(sp)
li t0, 0x70732834 # 4(sp
sw t0, 8(sp)
li t0, 0x38202c30 # 0, 8
sw t0, 4(sp)
li t0, 0x61207773 # sw a
sw t0, 0(sp)
addi a0, sp, 0
li a1, 14
call _write_out
li t0, '0'
sb t0, 8(sp)
li t0, 0x38202c31 # 1, 8
sw t0, 4(sp)
addi a0, sp, 0
li a1, 14
call _write_out
li t0, '6'
sb t0, 8(sp)
li t0, 0x37202c32 # 2, 7
sw t0, 4(sp)
addi a0, sp, 0
li a1, 14
call _write_out
li t0, '2'
sb t0, 8(sp)
li t0, 0x37202c33 # 3, 7
sw t0, 4(sp)
addi a0, sp, 0
li a1, 14
call _write_out
li t0, '8'
sb t0, 8(sp)
li t0, 0x36202c34 # 4, 6
sw t0, 4(sp)
addi a0, sp, 0
li a1, 14
call _write_out
li t0, '4'
sb t0, 8(sp)
li t0, 0x36202c35 # 5, 6
sw t0, 4(sp)
addi a0, sp, 0
li a1, 14
la a0, asm_preserve_parameters
li a1, ASM_PRESERVE_PARAMETERS_SIZE
call _write_out
# Generate the body of the procedure.
@ -1318,48 +1245,6 @@ _compile_procedure:
addi sp, sp, 32
ret
# Compares two string, which of one has a length, the other one is null-terminated.
#
# a0 - The address of the token string.
# a1 - The length of the string in a0.
# a2 - The address of the null-terminated string.
#
# If the strings match sets a0 to 0, otherwise sets it to 1.
.type _token_compare, @function
_token_compare:
addi t0, a0, 0
addi t1, a1, 0
addi t2, a2, 0
.Ltoken_compare_loop:
lbu t3, (t2)
# Will only be 0 if the current character in the null terminated string is \0 and the remaining length of the
# another string is 0.
or t4, t3, t1
beqz t4, .Ltoken_compare_equal
beqz t1, .Ltoken_compare_not_equal
beqz t3, .Ltoken_compare_not_equal
lbu t4, (t0)
bne t3, t4, .Ltoken_compare_not_equal
addi t0, t0, 1
addi t1, t1, -1
addi t2, t2, 1
j .Ltoken_compare_loop
.Ltoken_compare_not_equal:
li a0, 1
j .Ltoken_compare_end
.Ltoken_compare_equal:
li a0, 0
.Ltoken_compare_end:
ret
.type _compile_goto, @function
_compile_goto:
# Prologue.
@ -1475,30 +1360,30 @@ _compile_if:
sw s0, 24(sp)
addi s0, sp, 32
addi s1, s1, 2 # Skip the if.
call _skip_spaces
mv a0, s1
addi a1, sp, 0
call _tokenize_next
mv s1, a0 # Skip the if.
call _build_binary_expression
call _skip_spaces
addi s1, s1, 4 # Skip the then.
# if end marker.
li t0, 0x00646e65 # end\0
sw t0, 20(sp)
mv a0, s1
addi a1, sp, 0
call _tokenize_next
mv s1, a0 # Skip the then.
# Label prefix.
li t0, 0x66694c2e # .Lif
sw t0, 16(sp)
sw t0, 20(sp)
li t0, 0x202c3061 # a0,_
sw t0, 12(sp)
sw t0, 16(sp)
li t0, 0x207a7165 # eqz_
sw t0, 8(sp)
sw t0, 12(sp)
li t0, 0x62626262 # bbbb
sb t0, 7(sp)
sb t0, 11(sp)
addi a0, sp, 7
addi a0, sp, 11
li a1, 13
call _write_out
@ -1510,15 +1395,13 @@ _compile_if:
call _put_char
.Lcompile_if_loop:
call _skip_spaces
call _read_token
mv a1, a0
mv a0, s1
addi a2, sp, 20
call _token_compare
addi a1, sp, 0
call _tokenize_next
beqz a0, .Lcompile_if_end
lw t0, 0(sp)
li t1, TOKEN_END
beq t0, t1, .Lcompile_if_end
call _read_line
li a1, 1
@ -1527,8 +1410,10 @@ _compile_if:
j .Lcompile_if_loop
.Lcompile_if_end:
mv s1, a0 # Skip the end with newline. a0 is set by the last call to _tokenize_next.
# Write the label prefix.
addi a0, sp, 16
addi a0, sp, 20
li a1, 4
call _write_out
@ -1537,14 +1422,13 @@ _compile_if:
call _printi
# Finalize the label.
li t0, 0x0a3a0a3a # :\n:\n
sh t0, 12(sp)
addi a0, sp, 12
li t0, 0x0a3a # :\n:\n
sh t0, 16(sp)
addi a0, sp, 16
li a1, 2
call _write_out
addi s2, s2, 1 # Increment the label counter.
addi s1, s1, 4 # Skip the end with newline.
# Epilogue.
lw ra, 28(sp)
@ -1554,8 +1438,6 @@ _compile_if:
# Parameters:
# a0 - Line length.
# a1 - Whether the section header was already emitted. If not it should be
# emitted before any code is written.
#
# Returns 1 in a0 if the parsed line contained a text section element such a
# procedure or the program entry point. Otherwise sets a0 to 0.
@ -1569,9 +1451,6 @@ _compile_statement:
# Preserve passed arguments.
sw a0, 20(sp)
sw a1, 16(sp)
call _skip_comment
mv a0, s1
lw a1, 20(sp)
@ -1591,25 +1470,31 @@ _compile_statement:
call _memcmp
beqz a0, .Lcompile_statement_goto
li t0, 0x75746572 # retu
sw t0, 12(sp)
call _skip_comment
/* DEBUG
mv a0, s1
addi a1, sp, 12
li a2, 4
call _memcmp
beqz a0, .Lcompile_statement_return
li a1, 4
call _write_error
mv a0, s1
li a1, 4
call _write_error
*/
li t0, 0x6669 # if
sw t0, 12(sp)
mv a0, s1
addi a1, sp, 12
li a2, 2
call _memcmp
beqz a0, .Lcompile_statement_if
addi a1, sp, 0
call _tokenize_next
lw t0, 0(sp)
li t1, TOKEN_RETURN
beq t0, t1, .Lcompile_statement_return
li t1, TOKEN_IF
beq t0, t1, .Lcompile_statement_if
li t1, TOKEN_DOT
beq t0, t1, .Lcompile_statement_label
lbu t0, (s1)
li t1, '.'
beq t0, t1, .Lcompile_statement_label
li t1, '_'
beq t0, t1, .Lcompile_statement_identifier
@ -1688,30 +1573,25 @@ _compile_entry_point:
# Generate the body of the procedure.
.Lcompile_entry_point_body:
call _skip_spaces
call _read_line
sw a0, 12(sp)
li t0, 0x2e646e65 # end
sw t0, 8(sp)
mv a0, s1
addi a1, sp, 8
li a2, 4
call _memcmp
addi a1, sp, 4
call _tokenize_next
beqz a0, .Lcompile_entry_point_end
lw t0, 4(sp)
li t1, TOKEN_END
beq t0, t1, .Lcompile_entry_point_end
lw a0, 12(sp)
call _compile_statement
j .Lcompile_entry_point_body
.Lcompile_entry_point_end:
mv s1, a0 # Skip end. a0 is set by the last _tokenize_next call.
la a0, asm_exit
li a1, ASM_EXIT_SIZE
call _write_out
addi s1, s1, 4 # Skip end.
call _skip_spaces # Read the possible new line at the end of the file.
# Epilogue.
lw ra, 4(sp)
lw s0, 0(sp)

65
boot/tokenizer.s
View File

@ -38,7 +38,6 @@
.equ CLASS_COUNT, 20
.type classification, @object
.size classification, 128
classification:
.byte CLASS_EOF # 00 NUL
.byte CLASS_INVALID # 01 SOH
@ -172,7 +171,7 @@ classification:
#
# Textual keywords in the language.
#
.equ KEYWORDS_COUNT, 21
.equ KEYWORDS_COUNT, TOKEN_IDENTIFIER - 1
.type keywords, @object
keywords:
@ -222,8 +221,8 @@ keywords:
.ascii "return"
.word 4
.ascii "cast"
.word 5
.ascii "defer"
.word 4
.ascii "goto"
.word 4
.ascii "case"
.word 2
@ -251,13 +250,12 @@ byte_keywords: .ascii "&.,:;()[]^=+-*@"
# handles each action.
#
.type transitions, @object
.size transitions, 14 * CLASS_COUNT # state count * CLASS_COUNT
transitions:
# Invalid Digit Alpha Space : = ( )
# * _ Single Hex 0 x NUL .
# - " or ' > <
.word 0x00ff, 0x0103, 0x0102, 0x0300, 0x0101, 0x06ff, 0x0106, 0x06ff
.word 0x06ff, 0x0102, 0x06ff, 0x0102, 0x010c, 0x0102, 0x00ff, 0x0108
.word 0x06ff, 0x0102, 0x06ff, 0x0102, 0x010c, 0x0102, 0x00ff, 0x06ff
.word 0x0105, 0x0110, 0x0104, 0x0107 # 0x00 Start
.word 0x02ff, 0x02ff, 0x02ff, 0x02ff, 0x02ff, 0x07ff, 0x02ff, 0x02ff
@ -280,9 +278,9 @@ transitions:
.word 0x06ff, 0x06ff, 0x06ff, 0x06ff, 0x06ff, 0x06ff, 0x06ff, 0x06ff
.word 0x06ff, 0x06ff, 0x04ff, 0x06ff # 0x05 Minus
.word 0x02ff, 0x02ff, 0x02ff, 0x02ff, 0x02ff, 0x02ff, 0x02ff, 0x02ff
.word 0x0109, 0x02ff, 0x02ff, 0x02ff, 0x02ff, 0x02ff, 0x02ff, 0x02ff
.word 0x02ff, 0x02ff, 0x02ff, 0x02ff # 0x06 Left paren
.word 0x06ff, 0x06ff, 0x06ff, 0x06ff, 0x06ff, 0x06ff, 0x06ff, 0x06ff
.word 0x0109, 0x06ff, 0x06ff, 0x06ff, 0x06ff, 0x06ff, 0x06ff, 0x06ff
.word 0x06ff, 0x06ff, 0x06ff, 0x06ff # 0x06 Left paren
.word 0x02ff, 0x02ff, 0x02ff, 0x02ff, 0x02ff, 0x02ff, 0x02ff, 0x02ff
.word 0x02ff, 0x02ff, 0x02ff, 0x02ff, 0x02ff, 0x02ff, 0x02ff, 0x02ff
@ -399,6 +397,10 @@ _classify_identifier:
la a1, keywords
call _strings_index
bnez a0, .Lclassify_identifier_end
li a0, TOKEN_IDENTIFIER
.Lclassify_identifier_end:
# Epilogue.
lw ra, 12(sp)
lw s0, 8(sp)
@ -426,7 +428,7 @@ _classify_single:
la a1, byte_keywords
sub a0, a0, a1
addi a0, a0, 27
addi a0, a0, TOKEN_IDENTIFIER + 1
# Epilogue.
lw ra, 12(sp)
@ -466,16 +468,17 @@ _classify_composite:
.type _tokenize_next, @function
_tokenize_next:
# Prologue.
addi sp, sp, -24
sw ra, 20(sp)
sw s0, 16(sp)
addi s0, sp, 24
addi sp, sp, -32
sw ra, 28(sp)
sw s0, 24(sp)
addi s0, sp, 32
sw s1, 12(sp) # Preserve s1 used for current source text position.
sw s1, 20(sp) # Preserve s1 used for current source text position.
mv s1, a0
sw a0, 4(sp) # Keeps a pointer to the beginning of a token.
sw a0, 12(sp) # Keeps a pointer to the beginning of a token.
# 4(sp) and 8(sp) are reserved for the kind and length of the token if needed.
sw s2, 8(sp) # Preserve s2 containing the current state.
sw s2, 16(sp) # Preserve s2 containing the current state.
li s2, 0x00 # Initial, start state.
sw a1, 0(sp)
@ -529,9 +532,9 @@ _tokenize_next:
.Ltokenize_next_skip:
addi s1, s1, 1
lw t0, 4(sp)
lw t0, 12(sp)
addi t0, t0, 1
sw t0, 4(sp)
sw t0, 12(sp)
j .Ltokenize_next_loop
@ -553,16 +556,20 @@ _tokenize_next:
.Ltokenize_next_identifier:
# An identifier can be a textual keyword.
# Check the kind of the token and write it into the output parameter.
lw a1, 4(sp)
lw a1, 12(sp)
sub a0, s1, a1
sw a0, 8(sp)
call _classify_identifier
lw a1, 0(sp)
sw a0, (a1)
sw a0, 4(sp)
lw a0, 0(sp)
addi a1, sp, 4
li a2, 12
call _memcpy
j .Ltokenize_next_end
.Ltokenize_next_single:
lw a0, 4(sp)
lw a0, 12(sp)
addi s1, a0, 1
lbu a0, (a0)
call _classify_single
@ -573,7 +580,7 @@ _tokenize_next:
.Ltokenize_next_composite:
addi s1, s1, 1
lw a1, 4(sp)
lw a1, 12(sp)
sub a0, s1, a1
call _classify_composite
lw a1, 0(sp)
@ -585,11 +592,11 @@ _tokenize_next:
mv a0, s1 # Return the advanced text pointer.
# Restore saved registers.
lw s1, 12(sp)
lw s2, 8(sp)
lw s1, 20(sp)
lw s2, 16(sp)
# Epilogue.
lw ra, 20(sp)
lw s0, 16(sp)
addi sp, sp, 24
lw ra, 28(sp)
lw s0, 24(sp)
addi sp, sp, 32
ret