# Port of https://github.com/akkartik/crenshaw/blob/master/tutor2.1.pas # which corresponds to the section "single digits" in https://compilers.iecc.com/crenshaw/tutor2.txt # except that we support hex digits. # # To run: # $ ./subx translate init.linux 0*.subx apps/crenshaw2-1.subx -o apps/crenshaw2-1 # $ echo '3' |./subx run apps/crenshaw2-1 # Expected output: # # syscall(exit, 3) # bb/copy-to-ebx 3/imm32 # b8/copy-to-eax 1/imm32/exit # cd/syscall 0x80/imm8 # # To run the generated output: # $ echo '3' |./subx run apps/crenshaw2-1 > z1.subx # $ ./subx translate init.linux z1.subx -o z1 # $ ./subx run z1 # $ echo $? # 3 # # Stdin must contain just a single hex digit. Other input will print an error: # $ echo 'xyz' |./subx run apps/crenshaw2-1 # Error: integer expected # # Names in this file sometimes follow Crenshaw's original rather than my usual # naming conventions. == code # instruction effective address register displacement immediate # . op subop mod rm32 base index scale r32 # . 1-3 bytes 3 bits 2 bits 3 bits 3 bits 3 bits 2 bits 2 bits 0/1/2/4 bytes 0/1/2/4 bytes Entry: # run tests if necessary, call 'compile' if not # . prologue 89/copy 3/mod/direct 5/rm32/ebp . . . 4/r32/esp . . # copy esp to ebp # initialize heap # . Heap = new-segment(Heap-size) # . . push args 68/push Heap/imm32 68/push Heap-size/imm32 # . . call e8/call new-segment/disp32 # . . discard args 81 0/subop/add 3/mod/direct 4/rm32/esp . . . . . 8/imm32 # add to esp # - if argc > 1 and argv[1] == "test", then return run_tests() # if (argc <= 1) goto run-main 81 7/subop/compare 1/mod/*+disp8 5/rm32/ebp . . . . 0/disp8 1/imm32 # compare *ebp 7e/jump-if-lesser-or-equal $run-main/disp8 # if (!kernel-string-equal?(argv[1], "test")) goto run-main # . eax = kernel-string-equal?(argv[1], "test") # . . push args 68/push "test"/imm32 ff 6/subop/push 1/mod/*+disp8 5/rm32/ebp . . . . 8/disp8 . # push *(ebp+8) # . . call e8/call kernel-string-equal?/disp32 # . . discard args 81 0/subop/add 3/mod/direct 4/rm32/esp . . . . . 8/imm32 # add to esp # . if (eax == 0) goto run-main 3d/compare-eax-and 0/imm32 74/jump-if-equal $run-main/disp8 # run-tests() e8/call run-tests/disp32 # syscall(exit, *Num-test-failures) 8b/copy 0/mod/indirect 5/rm32/.disp32 . . 3/r32/ebx Num-test-failures/disp32 # copy *Num-test-failures to ebx eb/jump $main:end/disp8 $run-main: # - otherwise read a program from stdin and emit its translation to stdout # var ed/eax : exit-descriptor 81 5/subop/subtract 3/mod/direct 4/rm32/esp . . . . . 8/imm32 # subtract from esp 89/copy 3/mod/direct 0/rm32/eax . . . 4/r32/esp . . # copy esp to eax # configure ed to really exit() # . ed->target = 0 c7 0/subop/copy 0/mod/direct 0/rm32/eax . . . . . 0/imm32 # copy to *eax # compile(Stdin, 1/stdout, 2/stderr, ed) # . . push args 50/push-eax/ed 68/push 2/imm32/stderr 68/push 1/imm32/stdout 68/push Stdin/imm32 # . . call e8/call compile/disp32 # . . discard args 81 0/subop/add 3/mod/direct 4/rm32/esp . . . . . 0x10/imm32 # add to esp # syscall(exit, 0) bb/copy-to-ebx 0/imm32 $main:end: b8/copy-to-eax 1/imm32/exit cd/syscall 0x80/imm8 # the main entry point compile: # in : (address buffered-file), out : fd or (address stream), err : fd or (address stream), ed : (address exit-descriptor) # . prologue 55/push-ebp 89/copy 3/mod/direct 5/rm32/ebp . . . 4/r32/esp . . # copy esp to ebp # . save registers 50/push-eax 51/push-ecx # prime the pump # . Look = get-char(in) # . . push args ff 6/subop/push 1/mod/*+disp8 5/rm32/ebp . . . . 8/disp8 . # push *(ebp+8) # . . call e8/call get-char/disp32 # . . discard args 81 0/subop/add 3/mod/direct 4/rm32/esp . . . . . 4/imm32 # add to esp # var num/ecx : (address stream) on the stack # Numbers can be 32 bits or 8 hex bytes long. One of them will be in 'Look', so we need space for 7 bytes. # Sizing the stream just right buys us overflow-handling for free inside 'get-num'. # Add 12 bytes for 'read', 'write' and 'length' fields, for a total of 19 bytes, or 0x13 in hex. # The stack pointer is no longer aligned, so dump_stack() can be misleading past this point. 81 5/subop/subtract 3/mod/direct 4/rm32/esp . . . . . 0x13/imm32 # subtract from esp 89/copy 3/mod/direct 1/rm32/ecx . . . 4/r32/esp . . # copy esp to ecx # initialize the stream # . num->length = 7 c7 0/subop/copy 1/mod/*+disp8 1/rm32/ecx . . . . 8/disp8 7/imm32 # copy to *(ecx+8) # . clear-stream(num) # . . push args 51/push-ecx # . . call e8/call clear-stream/disp32 # . . discard args 81 0/subop/add 3/mod/direct 4/rm32/esp . . . . . 4/imm32 # add to esp # read a digit from 'in' into 'num' # . get-num(in, num, err, ed) # . . push args ff 6/subop/push 1/mod/*+disp8 5/rm32/ebp . . . . 0x14/disp8 . # push *(ebp+20) ff 6/subop/push 1/mod/*+disp8 5/rm32/ebp . . . . 0x10/disp8 . # push *(ebp+16) 51/push-ecx/num ff 6/subop/push 1/mod/*+disp8 5/rm32/ebp . . . . 8/disp8 . # push *(ebp+8) # . . call e8/call get-num/disp32 # . . discard args 81 0/subop/add 3/mod/direct 4/rm32/esp . . . . . 0x10/imm32 # add to esp # render 'num' into the following template on 'out': # bb/copy-to-ebx _num_ # b8/copy-to-eax 1/imm32/exit # cd/syscall 0x80/imm8 # # . write(out, "bb/copy-to-ebx ") # . . push args 68/push "bb/copy-to-ebx "/imm32 ff 6/subop/push 1/mod/*+disp8 5/rm32/ebp . . . . 0xc/disp8 . # push *(ebp+12) # . . call e8/call write/disp32 # . . discard args 81 0/subop/add 3/mod/direct 4/rm32/esp . . . . . 8/imm32 # add to esp # . write-stream(out, num) # . . push args 51/push-ecx/num ff 6/subop/push 1/mod/*+disp8 5/rm32/ebp . . . . 0xc/disp8 . # push *(ebp+12) # . . call e8/call write-stream/disp32 # . . discard args 81 0/subop/add 3/mod/direct 4/rm32/esp . . . . . 8/imm32 # add to esp # . write(out, Newline) # . . push args 68/push Newline/imm32 ff 6/subop/push 1/mod/*+disp8 5/rm32/ebp . . . . 0xc/disp8 . # push *(ebp+12) # . . call e8/call write/disp32 # . . discard args 81 0/subop/add 3/mod/direct 4/rm32/esp . . . . . 8/imm32 # add to esp # . write(out, "b8/copy-to-eax 1/imm32/exit\n") # . . push args 68/push "b8/copy-to-eax 1/imm32/exit\n"/imm32 ff 6/subop/push 1/mod/*+disp8 5/rm32/ebp . . . . 0xc/disp8 . # push *(ebp+12) # . . call e8/call write/disp32 # . . discard args 81 0/subop/add 3/mod/direct 4/rm32/esp . . . . . 8/imm32 # add to esp # . write(out, "cd/syscall 0x80/imm8\n") # . . push args 68/push "cd/syscall 0x80/imm8\n"/imm32 ff 6/subop/push 1/mod/*+disp8 5/rm32/ebp . . . . 0xc/disp8 . # push *(ebp+12) # . . call e8/call write/disp32 # . . discard args 81 0/subop/add 3/mod/direct 4/rm32/esp . . . . . 8/imm32 # add to esp $compile:end: # . restore registers 59/pop-to-ecx 58/pop-to-eax # . epilogue 89/copy 3/mod/direct 4/rm32/esp . . . 5/r32/ebp . . # copy ebp to esp 5d/pop-to-ebp c3/return # Read a single digit into 'out'. Abort if there are none, or if there is no space in 'out'. # Input comes from the global variable 'Look', and we leave the next byte from # 'in' into it on exit. get-num: # in : (address buffered-file), out : (address stream), err : fd or (address stream), ed : (address exit-descriptor) # pseudocode: # if (!is-digit?(Look)) expected(ed, err, "integer") # if out->write >= out->length # write(err, "Error: too many digits in number\n") # stop(ed, 1) # out->data[out->write] = LSB(Look) # ++out->write # Look = get-char(in) # # registers: # in: esi # out: edi # out->write: ecx (cached copy; need to keep in sync) # out->length: edx # temporaries: eax, ebx # We can't allocate Look to a register because it gets written implicitly in # get-char in each iteration of the loop. (Thereby demonstrating that it's # not the right interface for us. But we'll keep it just to follow Crenshaw.) # # . prologue 55/push-ebp 89/copy 3/mod/direct 5/rm32/ebp . . . 4/r32/esp . . # copy esp to ebp # - if (is-digit?(Look)) expected(ed, err, "integer") # . eax = is-digit?(Look) # . . push args ff 6/subop/push 0/mod/indirect 5/rm32/.disp32 . . . Look/disp32 . # push *Look # . . call e8/call is-digit?/disp32 # . . discard args 81 0/subop/add 3/mod/direct 4/rm32/esp . . . . . 4/imm32 # add to esp # . if (eax == 0) 3d/compare-eax-and 0/imm32 75/jump-if-not-equal $get-num:main/disp8 # . expected(ed, err, "integer") # . . push args 68/push "integer"/imm32 ff 6/subop/push 1/mod/*+disp8 5/rm32/ebp . . . . 0x10/disp8 . # push *(ebp+16) ff 6/subop/push 1/mod/*+disp8 5/rm32/ebp . . . . 0x14/disp8 . # push *(ebp+20) # . . call e8/call expected/disp32 # never returns # . . discard args 81 0/subop/add 3/mod/direct 4/rm32/esp . . . . . 0xc/imm32 # add to esp $get-num:main: # - otherwise read a digit # . save registers 50/push-eax 51/push-ecx 52/push-edx 53/push-ebx 56/push-esi 57/push-edi # read necessary variables to registers # esi = in 8b/copy 1/mod/*+disp8 5/rm32/ebp . . . 6/r32/esi 8/disp8 . # copy *(ebp+8) to esi # edi = out 8b/copy 1/mod/*+disp8 5/rm32/ebp . . . 7/r32/edi 0xc/disp8 . # copy *(ebp+12) to edi # ecx = out->write 8b/copy 0/mod/indirect 7/rm32/edi . . . 1/r32/ecx . . # copy *edi to ecx # edx = out->length 8b/copy 1/mod/*+disp8 7/rm32/edi . . . 2/r32/edx 8/disp8 . # copy *(edi+8) to edx # if (out->write >= out->length) error 39/compare 3/mod/direct 2/rm32/edx . . . 1/r32/ecx . . # compare edx with ecx 7d/jump-if-lesser $get-num:stage2/disp8 # . error(ed, err, msg) # TODO: show full number # . . push args 68/push "get-num: too many digits in number"/imm32 ff 6/subop/push 1/mod/*+disp8 5/rm32/ebp . . . . 0x10/disp8 . # push *(ebp+16) ff 6/subop/push 1/mod/*+disp8 5/rm32/ebp . . . . 0x14/disp8 . # push *(ebp+20) # . . call e8/call error/disp32 # never returns # . . discard args 81 0/subop/add 3/mod/direct 4/rm32/esp . . . . . 0xc/imm32 # add to esp $get-num:stage2: # out->data[out->write] = LSB(Look) 8d/copy-address 1/mod/*+disp8 4/rm32/sib 7/base/edi 1/index/ecx . 3/r32/ebx 0xc/disp8 . # copy edi+ecx+12 to ebx 8b/copy 0/mod/indirect 5/rm32/.disp32 . . 0/r32/eax Look/disp32 . # copy *Look to eax 88/copy-byte 0/mod/indirect 3/rm32/ebx . . . 0/r32/AL . . # copy byte at AL to *ebx # ++out->write 41/increment-ecx # Look = get-char(in) # . . push args 56/push-esi # . . call e8/call get-char/disp32 # . . discard args 81 0/subop/add 3/mod/direct 4/rm32/esp . . . . . 4/imm32 # add to esp $get-num:loop-end: # persist necessary variables from registers 89/copy 0/mod/indirect 7/rm32/edi . . . 1/r32/ecx . . # copy ecx to *edi $get-num:end: # . restore registers 5f/pop-to-edi 5e/pop-to-esi 5b/pop-to-ebx 5a/pop-to-edx 59/pop-to-ecx 58/pop-to-eax # . epilogue 89/copy 3/mod/direct 4/rm32/esp . . . 5/r32/ebp . . # copy ebp to esp 5d/pop-to-ebp c3/return test-get-num-reads-single-digit: # - check that get-num returns first character if it's a digit # This test uses exit-descriptors. Use ebp for setting up local variables. 55/push-ebp 89/copy 3/mod/direct 5/rm32/ebp . . . 4/r32/esp . . # copy esp to ebp # clear all streams # . clear-stream(_test-stream) # . . push args 68/push _test-stream/imm32 # . . call e8/call clear-stream/disp32 # . . discard args 81 0/subop/add 3/mod/direct 4/rm32/esp . . . . . 4/imm32 # add to esp # . clear-stream(_test-buffered-file+4) # . . push args b8/copy-to-eax _test-buffered-file/imm32 05/add-to-eax 4/imm32 50/push-eax # . . call e8/call clear-stream/disp32 # . . discard args 81 0/subop/add 3/mod/direct 4/rm32/esp . . . . . 4/imm32 # add to esp # . clear-stream(_test-output-stream) # . . push args 68/push _test-output-stream/imm32 # . . call e8/call clear-stream/disp32 # . . discard args 81 0/subop/add 3/mod/direct 4/rm32/esp . . . . . 4/imm32 # add to esp # . clear-stream(_test-error-stream) # . . push args 68/push _test-error-stream/imm32 # . . call e8/call clear-stream/disp32 # . . discard args 81 0/subop/add 3/mod/direct 4/rm32/esp . . . . . 4/imm32 # add to esp # initialize 'in' # . write(_test-stream, "3") # . . push args 68/push "3"/imm32 68/push _test-stream/imm32 # . . call e8/call write/disp32 # . . discard args 81 0/subop/add 3/mod/direct 4/rm32/esp . . . . . 8/imm32 # add to esp # initialize exit-descriptor 'ed' for the call to 'get-num' below # . var ed/eax : exit-descriptor 81 5/subop/subtract 3/mod/direct 4/rm32/esp . . . . . 8/imm32 # subtract from esp 89/copy 3/mod/direct 0/rm32/eax . . . 4/r32/esp . . # copy esp to eax # . tailor-exit-descriptor(ed, 16) # . . push args 68/push 0x10/imm32/nbytes-of-args-for-get-num 50/push-eax/ed # . . call e8/call tailor-exit-descriptor/disp32 # . . discard args 81 0/subop/add 3/mod/direct 4/rm32/esp . . . . . 8/imm32 # add to esp # prime the pump # . get-char(_test-buffered-file) # . . push args 68/push _test-buffered-file/imm32 # . . call e8/call get-char/disp32 # . . discard args 81 0/subop/add 3/mod/direct 4/rm32/esp . . . . . 4/imm32 # add to esp # get-num(in, out, err, ed) # . . push args 50/push-eax/ed 68/push _test-error-stream/imm32 68/push _test-output-stream/imm32 68/push _test-buffered-file/imm32 # . . call e8/call get-num/disp32 # registers except esp may be clobbered at this point # . . discard args 81 0/subop/add 3/mod/direct 4/rm32/esp . . . . . 0x10/imm32 # add to esp # check-ints-equal(*_test-output-stream->data, '3', msg) # . . push args 68/push "F - test-get-num-reads-single-digit"/imm32 68/push 0x33/imm32 b8/copy-to-eax _test-output-stream/imm32 ff 6/subop/push 1/mod/*+disp8 0/rm32/eax . . . . 0xc/disp8 . # push *(eax+12) # . . call e8/call check-ints-equal/disp32 # . . discard args 81 0/subop/add 3/mod/direct 4/rm32/esp . . . . . 0xc/imm32 # add to esp # . reclaim locals 81 0/subop/add 3/mod/direct 4/rm32/esp . . . . . 8/imm32 # add to esp 5d/pop-to-ebp c3/return test-get-num-aborts-on-non-digit-in-Look: # - check that get-num returns first character if it's a digit # This test uses exit-descriptors. Use ebp for setting up local variables. 55/push-ebp 89/copy 3/mod/direct 5/rm32/ebp . . . 4/r32/esp . . # copy esp to ebp # clear all streams # . clear-stream(_test-stream) # . . push args 68/push _test-stream/imm32 # . . call e8/call clear-stream/disp32 # . . discard args 81 0/subop/add 3/mod/direct 4/rm32/esp . . . . . 4/imm32 # add to esp # . clear-stream(_test-buffered-file+4) # . . push args b8/copy-to-eax _test-buffered-file/imm32 05/add-to-eax 4/imm32 50/push-eax # . . call e8/call clear-stream/disp32 # . . discard args 81 0/subop/add 3/mod/direct 4/rm32/esp . . . . . 4/imm32 # add to esp # . clear-stream(_test-output-stream) # . . push args 68/push _test-output-stream/imm32 # . . call e8/call clear-stream/disp32 # . . discard args 81 0/subop/add 3/mod/direct 4/rm32/esp . . . . . 4/imm32 # add to esp # . clear-stream(_test-error-stream) # . . push args 68/push _test-error-stream/imm32 # . . call e8/call clear-stream/disp32 # . . discard args 81 0/subop/add 3/mod/direct 4/rm32/esp . . . . . 4/imm32 # add to esp # initialize 'in' # . write(_test-stream, "3") # . . push args 68/push "3"/imm32 68/push _test-stream/imm32 # . . call e8/call write/disp32 # . . discard args 81 0/subop/add 3/mod/direct 4/rm32/esp . . . . . 8/imm32 # add to esp # initialize exit-descriptor 'ed' for the call to 'get-num' below # . var ed/eax : (address exit-descriptor) 81 5/subop/subtract 3/mod/direct 4/rm32/esp . . . . . 8/imm32 # subtract from esp 89/copy 3/mod/direct 0/rm32/eax . . . 4/r32/esp . . # copy esp to eax # . tailor-exit-descriptor(ed, 16) # . . push args 68/push 0x10/imm32/nbytes-of-args-for-get-num 50/push-eax/ed # . . call e8/call tailor-exit-descriptor/disp32 # . . discard args 81 0/subop/add 3/mod/direct 4/rm32/esp . . . . . 8/imm32 # add to esp # *don't* prime the pump # get-num(in, out, err, ed) # . . push args 50/push-eax/ed 68/push _test-error-stream/imm32 68/push _test-output-stream/imm32 68/push _test-buffered-file/imm32 # . . call e8/call get-num/disp32 # registers except esp may be clobbered at this point # . . discard args 81 0/subop/add 3/mod/direct 4/rm32/esp . . . . . 0x10/imm32 # add to esp # check that get-num tried to call exit(1) # . check-ints-equal(ed->value, 2, msg) # i.e. stop was called with value 1 # . . push args 68/push "F - test-get-num-aborts-on-non-digit-in-Look"/imm32 68/push 2/imm32 # . . push ed->value ff 6/subop/push 1/mod/*+disp8 0/rm32/eax . . . . 4/disp8 . # push *(eax+4) # . . call e8/call check-ints-equal/disp32 # . . discard args 81 0/subop/add 3/mod/direct 4/rm32/esp . . . . . 0xc/imm32 # add to esp # . reclaim locals 81 0/subop/add 3/mod/direct 4/rm32/esp . . . . . 8/imm32 # add to esp 5d/pop-to-ebp c3/return ## helpers # write(f, "Error: "+s+" expected\n") then stop(ed, 1) expected: # ed : (address exit-descriptor), f : fd or (address stream), s : (address array byte) # . prologue 55/push-ebp 89/copy 3/mod/direct 5/rm32/ebp . . . 4/r32/esp . . # copy esp to ebp # write(f, "Error: ") # . . push args 68/push "Error: "/imm32 ff 6/subop/push 1/mod/*+disp8 5/rm32/ebp . . . . 0xc/disp8 . # push *(ebp+12) # . . call e8/call write/disp32 # . . discard args 81 0/subop/add 3/mod/direct 4/rm32/esp . . . . . 8/imm32 # add to esp # write(f, s) # . . push args ff 6/subop/push 1/mod/*+disp8 5/rm32/ebp . . . . 0x10/disp8 . # push *(ebp+16) ff 6/subop/push 1/mod/*+disp8 5/rm32/ebp . . . . 0xc/disp8 . # push *(ebp+12) # . . call e8/call write/disp32 # . . discard args 81 0/subop/add 3/mod/direct 4/rm32/esp . . . . . 8/imm32 # add to esp # write(f, " expected") # . . push args 68/push " expected\n"/imm32 ff 6/subop/push 1/mod/*+disp8 5/rm32/ebp . . . . 0xc/disp8 . # push *(ebp+12) # . . call e8/call write/disp32 # . . discard args 81 0/subop/add 3/mod/direct 4/rm32/esp . . . . . 8/imm32 # add to esp # stop(ed, 1) # . . push args 68/push 1/imm32 ff 6/subop/push 1/mod/*+disp8 5/rm32/ebp . . . . 8/disp8 . # push *(ebp+8) # . . call e8/call stop/disp32 # should never get past this point $expected:dead-end: # . epilogue 89/copy 3/mod/direct 4/rm32/esp . . . 5/r32/ebp . . # copy ebp to esp 5d/pop-to-ebp c3/return # read a byte from 'f', and save it in 'Look' get-char: # f : (address buffered-file) # . prologue 55/push-ebp 89/copy 3/mod/direct 5/rm32/ebp . . . 4/r32/esp . . # copy esp to ebp # . save registers 50/push-eax # eax = read-byte-buffered(f) # . . push args ff 6/subop/push 1/mod/*+disp8 5/rm32/ebp . . . . 8/disp8 . # push *(ebp+8) # . . call e8/call read-byte-buffered/disp32 # . . discard args 81 0/subop/add 3/mod/direct 4/rm32/esp . . . . . 4/imm32 # add to esp # save eax to Look 89/copy 0/mod/indirect 5/rm32/.disp32 . . 0/r32/eax Look/disp32 . # copy eax to *Look $get-char:end: # . restore registers 58/pop-to-eax # . epilogue 89/copy 3/mod/direct 4/rm32/esp . . . 5/r32/ebp . . # copy ebp to esp 5d/pop-to-ebp c3/return is-digit?: # c : int -> eax : boolean # . prologue 55/push-ebp 89/copy 3/mod/direct 5/rm32/ebp . . . 4/r32/esp . . # copy esp to ebp # eax = false b8/copy-to-eax 0/imm32 # if (c < '0') return false 81 7/subop/compare 1/mod/*+disp8 5/rm32/ebp . . . . 8/disp8 0x30/imm32 # compare *(ebp+8) 7c/jump-if-lesser $is-digit?:end/disp8 # if (c > '9') return false 81 7/subop/compare 1/mod/*+disp8 5/rm32/ebp . . . . 8/disp8 0x39/imm32 # compare *(ebp+8) 7f/jump-if-greater $is-digit?:end/disp8 # otherwise return true b8/copy-to-eax 1/imm32 $is-digit?:end: # . epilogue 89/copy 3/mod/direct 4/rm32/esp . . . 5/r32/ebp . . # copy ebp to esp 5d/pop-to-ebp c3/return == data Look: # (char with some extra padding) 0/imm32 # . . vim:nowrap:textwidth=0