commit 8ecfb0be1f42f3f88bc7117830ed054dd9ebe8d0
parent 0acd25e23eb849d8570612c2eb44d864e6ad04eb
Author: Brian Swetland <swetland@frotz.net>
Date: Mon, 2 Mar 2020 13:10:55 -0800
new disassmbler
- based on rv32 table driven disassembler
- library-ized
Diffstat:
| M | Makefile | | | 15 | ++++++++++++--- |
| A | src/mkinstab.c | | | 38 | ++++++++++++++++++++++++++++++++++++++ |
| A | src/r5d.c | | | 94 | +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ |
| D | src/r5dis.c | | | 206 | ------------------------------------------------------------------------------- |
| A | src/risc5.h | | | 4 | ++++ |
| A | src/risc5dis.c | | | 80 | +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ |
| A | src/risc5ins.txt | | | 33 | +++++++++++++++++++++++++++++++++ |
7 files changed, 261 insertions(+), 209 deletions(-)
diff --git a/Makefile b/Makefile
@@ -1,5 +1,5 @@
-all: bin/tlc bin/fs bin/r5dis
+all: bin/tlc bin/fs bin/r5d bin/mkinstab
clean:
rm -rf bin out
@@ -15,6 +15,15 @@ bin/fs: src/fs.c src/fs.h
@mkdir -p bin
$(CC) -o $@ $(CFLAGS) src/fs.c
-bin/r5dis: src/r5dis.c
+bin/r5d: src/r5d.c src/risc5dis.c src/risc5.h out/risc5ins.h
@mkdir -p bin
- $(CC) -o $@ $(CFLAGS) src/r5dis.c
+ $(CC) -o $@ $(CFLAGS) src/r5d.c src/risc5dis.c src/risc5.h
+
+bin/mkinstab: src/mkinstab.c
+ @mkdir -p bin
+ $(CC) -o $@ $(CFLAGS) src/mkinstab.c
+
+out/risc5ins.h: src/risc5ins.txt bin/mkinstab
+ @mkdir -p out
+ bin/mkinstab < src/risc5ins.txt > $@
+
diff --git a/src/mkinstab.c b/src/mkinstab.c
@@ -0,0 +1,38 @@
+// Copyright 2019, Brian Swetland <swetland@frotz.net>
+// Licensed under the Apache License, Version 2.0.
+
+#include <stdio.h>
+#include <stdint.h>
+#include <string.h>
+#include <ctype.h>
+
+int main(int argc, char** argv) {
+ char line[128];
+ while (fgets(line, sizeof(line), stdin) != NULL) {
+ unsigned end = strlen(line);
+ while (end > 0) {
+ end--;
+ if (!isspace(line[end])) break;
+ line[end] = 0;
+ }
+ if ((line[0] == 0) || (line[0] == '#') ||
+ isspace(line[0]) || (end < 34)) {
+ continue;
+ }
+ uint32_t mask = 0, bits = 0;
+ for (unsigned n = 0; n < 32; n++) {
+ uint32_t bit = 1U << (31 - n);
+ switch (line[n]) {
+ case '1':
+ mask |= bit;
+ bits |= bit;
+ break;
+ case '0':
+ mask |= bit;
+ break;
+ }
+ }
+ printf("{ 0x%08x, 0x%08x, \"%s\" },\n", mask, bits, line + 33);
+ }
+ return 0;
+}
diff --git a/src/r5d.c b/src/r5d.c
@@ -0,0 +1,94 @@
+// Copyright 2020, Brian Swetland <swetland@frotz.net>
+// Licensed under the Apache License, Version 2.0.
+
+#include <stdio.h>
+#include <stdlib.h>
+#include <stdint.h>
+#include <unistd.h>
+#include <fcntl.h>
+#include <string.h>
+
+#include "risc5.h"
+
+void disasm(uint32_t pc, uint32_t ins) {
+ char buf[256];
+ risc5dis(pc, ins, buf);
+ printf("%08x: %08x %s\n", pc, ins, buf);
+}
+
+const char* readname(int fd) {
+ static char name[64];
+ for (int n = 0; n < 64; n++) {
+ if (read(fd, name + n, 1) != 1) exit(1);
+ if (name[n] == 0) return name;
+ }
+ exit(1);
+ return NULL;
+}
+
+uint32_t readint(int fd) {
+ uint32_t n;
+ if (read(fd, &n, 4) != 4) exit(1);
+ return n;
+}
+
+uint32_t readbyte(int fd) {
+ uint8_t n;
+ if (read(fd, &n, 1) != 1) exit(1);
+ return n;
+}
+
+uint32_t dishdr(int fd) {
+ const char* name = readname(fd);
+ uint32_t key = readint(fd);
+ uint32_t cls = readbyte(fd);
+ uint32_t size = readint(fd);
+ printf("[ name='%s', key=%08x, class=%02x, size=%u ]\n",
+ name, key, cls, size);
+ printf("[ imports:");
+ for (;;) {
+ name = readname(fd);
+ if (name[0] == 0) break;
+ printf(" %s(%08x)", name, readint(fd));
+ }
+ printf(" ]\n");
+ // type descrs?
+ uint32_t n = readint(fd) / 4;
+ printf("[ typedesc=%u", n);
+ while (n > 0) { readint(fd); n--; }
+ // data?
+ n = readint(fd);
+ printf(", data=%08x", n);
+ // stringdata
+ n = readint(fd);
+ printf(", stringdata=%u", n);
+ while (n > 0) { readbyte(fd); n--; }
+ // instructions
+ n = readint(fd);
+ printf(", instructions=%u ]\n", n);
+ return n;
+}
+
+int main(int argc, char** argv) {
+ int fd;
+ if (argc != 2) return -1;
+ if ((fd = open(argv[1], O_RDONLY)) < 0) return -1;
+
+ uint32_t count;
+ count = strlen(argv[1]);
+ if ((count > 5) && (!strcmp(argv[1] + count - 4, ".rsc"))) {
+ count = dishdr(fd);
+ } else {
+ count = 0xffffffff;
+ }
+
+ uint32_t ins;
+ uint32_t pc = 0;
+ while (count > 0) {
+ count--;
+ if (read(fd, &ins, sizeof(ins)) != sizeof(ins)) break;
+ disasm(pc, ins);
+ pc += 4;
+ }
+ return 0;
+}
diff --git a/src/r5dis.c b/src/r5dis.c
@@ -1,206 +0,0 @@
-// Copyright 2020, Brian Swetland <swetland@frotz.net>
-// Licensed under the Apache License, Version 2.0.
-
-#include <stdio.h>
-#include <stdlib.h>
-#include <stdint.h>
-#include <unistd.h>
-#include <fcntl.h>
-#include <string.h>
-
-static const char* opname[16] = {
- "MOV", "LSL", "ASR", "ROR", "AND", "ANN", "IOR", "XOR",
- "ADD", "SUB", "MUL", "DIV", "FAD", "FSB", "FML", "FDV",
-};
-const char* OP(uint32_t n) { return opname[n & 15]; }
-
-static const char* ccname[16] = {
- "MI", "EQ", "CS", "VS", "LS", "LT", "LE", "",
- "PL", "NE", "CC", "VC", "HI", "GE", "GT", "NV",
-};
-const char* CC(uint32_t n) { return ccname[n & 15]; }
-
-static const char* regname[16] = {
- "R0", "R1", "R2", "R3", "R4", "R5", "R6", "R7",
- "R8", "R9", "R10", "R11", "MT", "SB", "SP", "LR",
-};
-const char* R(uint32_t n) { return regname[n & 15]; }
-
-void disasm(uint32_t pc, uint32_t ins) {
- uint32_t t = ins >> 28;
- uint32_t a = (ins >> 24) & 15;
- uint32_t b = (ins >> 20) & 15;
- uint32_t o = (ins >> 16) & 15;
- uint32_t c = ins & 15;
- uint32_t i16 = ins & 0xffff;
- uint32_t i20 = ins & 0xfffff;
- uint32_t i24 = ins & 0xffffff;
- if (i16 & 0x8000) i16 |= 0xffff0000;
- if (i20 & 0x80000) i20 |= 0xfff00000;
- if (i24 & 0x800000) i24 |= 0xff000000;
-
- printf("%08x: %08x ", pc, ins);
- switch (t) {
- // --uv
- case 0b0000:
- case 0b0001:
- case 0b0010:
- case 0b0011:
- if (o == 0) { // MOV ignores Rb
- switch (t & 3) {
- case 0b00:
- case 0b01:
- printf("MOV %s, %s\n", R(a), R(c));
- break;
- case 0b10:
- printf("MOV %s, H\n", R(a));
- break;
- case 0b11:
- printf("MOV %s, NZCF\n", R(a));
- break;
- }
- } else {
- const char* op = OP(o);
- switch (o) { // u-bit modifiers
- case 8: if (t & 2) op = "ADC"; break;
- case 9: if (t & 2) op = "SBC"; break;
- case 10: if (t & 2) op = "UMUL"; break;
- }
- printf("%s %s, %s, %s\n", op, R(a), R(b), R(c));
- }
- break;
- case 0b0100:
- case 0b0101:
- case 0b0110:
- case 0b0111:
- if (o == 0) {
- printf("MOV %s, %d\n", R(a), (t & 1) ? (i16 << 16) : i16);
- } else {
- const char* op = OP(o);
- switch (o) { // u-bit modifiers
- case 8: if (t & 2) op = "ADC"; break;
- case 9: if (t & 2) op = "SBC"; break;
- case 10: if (t & 2) op = "UMUL"; break;
- }
- printf("%s %s, %s, %d\n", op, R(a), R(b), i16);
- }
- break;
- case 0b1000:
- printf("LW %s, [%s, %d]\n", R(a), R(b), i20);
- break;
- case 0b1001:
- printf("LB %s, [%s, %d]\n", R(a), R(b), i20);
- break;
- case 0b1010:
- printf("SW %s, [%s, %d]\n", R(a), R(b), i20);
- break;
- case 0b1011:
- printf("SB %s, [%s, %d]\n", R(a), R(b), i20);
- break;
- case 0b1100:
- switch ((ins >> 4) & 15) {
- case 0b0000:
- printf("B%s %s\n", CC(a), R(c));
- break;
- case 0b0001:
- printf("RTI %s\n", R(c));
- break;
- case 0b0010:
- printf("%s\n", ins & 1 ? "STI" : "CLI");
- break;
- default:
- printf("??? %08x\n", ins);
- break;
- }
- break;
- case 0b1101:
- printf("BL%s %s\n", CC(a), R(c));
- break;
- case 0b1110:
- printf("B%s %d\n", CC(a), i24);
- break;
- case 0b1111:
- printf("BL%s %d\n", CC(a), i24);
- break;
- default:
- printf("??? %08x\n", ins);
- break;
- }
-}
-
-const char* readname(int fd) {
- static char name[64];
- for (int n = 0; n < 64; n++) {
- if (read(fd, name + n, 1) != 1) exit(1);
- if (name[n] == 0) return name;
- }
- exit(1);
- return NULL;
-}
-
-uint32_t readint(int fd) {
- uint32_t n;
- if (read(fd, &n, 4) != 4) exit(1);
- return n;
-}
-
-uint32_t readbyte(int fd) {
- uint8_t n;
- if (read(fd, &n, 1) != 1) exit(1);
- return n;
-}
-
-uint32_t dishdr(int fd) {
- const char* name = readname(fd);
- uint32_t key = readint(fd);
- uint32_t cls = readbyte(fd);
- uint32_t size = readint(fd);
- printf("[ name='%s', key=%08x, class=%02x, size=%u ]\n",
- name, key, cls, size);
- printf("[ imports:");
- for (;;) {
- name = readname(fd);
- if (name[0] == 0) break;
- printf(" %s(%08x)", name, readint(fd));
- }
- printf(" ]\n");
- // type descrs?
- uint32_t n = readint(fd) / 4;
- printf("[ typedesc=%u", n);
- while (n > 0) { readint(fd); n--; }
- // data?
- n = readint(fd);
- printf(", data=%08x", n);
- // stringdata
- n = readint(fd);
- printf(", stringdata=%u", n);
- while (n > 0) { readbyte(fd); n--; }
- // instructions
- n = readint(fd);
- printf(", instructions=%u ]\n", n);
- return n;
-}
-
-int main(int argc, char** argv) {
- int fd;
- if (argc != 2) return -1;
- if ((fd = open(argv[1], O_RDONLY)) < 0) return -1;
-
- uint32_t count;
- count = strlen(argv[1]);
- if ((count > 5) && (!strcmp(argv[1] + count - 4, ".rsc"))) {
- count = dishdr(fd);
- } else {
- count = 0xffffffff;
- }
-
- uint32_t ins;
- uint32_t pc = 0;
- while (count > 0) {
- count--;
- if (read(fd, &ins, sizeof(ins)) != sizeof(ins)) break;
- disasm(pc, ins);
- pc += 4;
- }
- return 0;
-}
diff --git a/src/risc5.h b/src/risc5.h
@@ -0,0 +1,4 @@
+
+#include <stdint.h>
+
+void risc5dis(uint32_t pc, uint32_t ins, char *out);
diff --git a/src/risc5dis.c b/src/risc5dis.c
@@ -0,0 +1,80 @@
+// Copyright 2020, Brian Swetland <swetland@frotz.net>
+// Licensed under the Apache License, Version 2.0.
+
+#include <string.h>
+#include <stdio.h>
+
+#include "risc5.h"
+
+static char *append_str(char *buf, const char *s) {
+ while (*s) *buf++ = *s++;
+ return buf;
+}
+
+static char *append_i32(char *buf, int32_t n) {
+ return buf + sprintf(buf, "%d", n);
+}
+
+static char *append_u32(char *buf, int32_t n) {
+ return buf + sprintf(buf, "0x%x", n);
+}
+
+static const char* regname[16] = {
+ "r0", "r1", "r2", "r3", "r4", "r5", "r6", "r7",
+ "r8", "r9", "r10", "r11", "mt", "sb", "sp", "lr",
+};
+#define R(n) regname[(n) & 15]
+
+static const char* opname[16] = {
+ "mov", "lsl", "asr", "ror", "and", "ann", "ior", "xor",
+ "add", "sub", "mul", "div", "fad", "fsb", "fml", "fdv",
+};
+#define OP(n) opname[(n) & 15]
+
+static const char* ccname[16] = {
+ "mi", "eq", "cs", "vs", "ls", "lt", "le", "",
+ "pl", "ne", "cc", "vc", "hi", "ge", "gt", "nv",
+};
+#define CC(n) ccname[(n) & 15]
+
+typedef struct {
+ uint32_t mask;
+ uint32_t bits;
+ const char* fmt;
+} ins_t;
+
+static ins_t instab[] = {
+#include "../out/risc5ins.h"
+};
+
+void risc5dis(uint32_t pc, uint32_t ins, char *out) {
+ unsigned n = 0;
+ while ((ins & instab[n].mask) != instab[n].bits) n++;
+ const char* fmt = instab[n].fmt;
+ char x;
+
+ unsigned mo = ins & 0xFFFFF;
+ unsigned bo = ins & 0xFFFFFF;
+ if (mo & 0x80000) mo |= 0xFFF00000;
+ if (bo & 0x800000) bo |= 0xFF000000;
+
+ while ((x = *fmt++) != 0) {
+ if (x != '%') {
+ *out++ = x;
+ continue;
+ }
+ switch (*fmt++) {
+ case 'C': out = append_str(out, CC((ins >> 24) & 15)); break;
+ case 'a': out = append_str(out, R((ins >> 24) & 15)); break;
+ case 'b': out = append_str(out, R((ins >> 20) & 15)); break;
+ case 'o': out = append_str(out, OP((ins >> 16) & 15)); break;
+ case 'c': out = append_str(out, R(ins & 15)); break;
+ case 'n': out = append_i32(out, ins & 0xFFFF); break;
+ case 's': out = append_u32(out, (ins & 0xFFFF) | 0xFFFF0000); break;
+ case 'N': out = append_u32(out, (ins & 0xFFFF) << 16); break;
+ case 'm': out = append_i32(out, mo); break;
+ case 'B': out = append_u32(out, pc + 4 + (bo << 2)); break;
+ }
+ }
+ *out = 0;
+}
diff --git a/src/risc5ins.txt b/src/risc5ins.txt
@@ -0,0 +1,33 @@
+# Copyright 2020, Brian Swetland <swetland@frotz.net>
+# Licensed under the Apache License, Version 2.0.
+
+0000aaaa----0000------------cccc mov %a, c
+0010aaaa----0000---------------- mov %a, h
+0011aaaa----0000---------------- mov %a, nzcv
+0010aaaabbbb1000------------cccc adc %a, %b, %c
+0010aaaabbbb1001------------cccc sbc %a, %b, %c
+0010aaaabbbb1010------------cccc uml %a, %b, %c
+0000aaaabbbboooo------------cccc %o %a, %b, %c
+0100aaaa----0000nnnnnnnnnnnnnnnn mov %a, %n
+0101aaaa----0000nnnnnnnnnnnnnnnn mov %a, %s
+0100aaaabbbboooonnnnnnnnnnnnnnnn %o %a, %b, %n
+0101aaaabbbboooossssssssssssssss %o %a, %b, %s
+0110aaaa----0000NNNNNNNNNNNNNNNN mhi %a, %N
+0111aaaa----0000NNNNNNNNNNNNNNNN mhi %a, %N
+0110aaaabbbb1000nnnnnnnnnnnnnnnn adc %a, %b, %n
+0111aaaabbbb1000ssssssssssssssss adc %a, %b, %s
+0110aaaabbbb1001nnnnnnnnnnnnnnnn sbc %a, %b, %n
+0111aaaabbbb1001ssssssssssssssss sbc %a, %b, %s
+0110aaaabbbb1010nnnnnnnnnnnnnnnn uml %a, %b, %n
+0111aaaabbbb1010ssssssssssssssss uml %%a, %b, %s
+1000aaaabbbbmmmmmmmmmmmmmmmmmmmm lw %a, [%b, %m]
+1001aaaabbbbmmmmmmmmmmmmmmmmmmmm lb %a, [%b, %m]
+1010aaaabbbbmmmmmmmmmmmmmmmmmmmm sw %a, [%b, %m]
+1011aaaabbbbmmmmmmmmmmmmmmmmmmmm sb %a, [%b, %m]
+11000111-------------------1---- rti
+11001111------------------1----0 sti
+11001111------------------1----1 cli
+1100CCCC--------------------cccc b%C %c
+1101CCCC--------------------cccc bl%C %c
+1110CCCCBBBBBBBBBBBBBBBBBBBBBBBB b%C %B
+1111CCCCBBBBBBBBBBBBBBBBBBBBBBBB bl%C %B
