// TODO: add support for byte-per-byte opcode search
R_API RList *r_core_asm_strsearch(RCore *core, const char *input, ut64 from, ut64 to) {
RCoreAsmHit *hit;
RAsmOp op;
RList *hits;
ut64 at, toff = core->offset;
ut8 *buf;
char *tok, *tokens[1024], *code = NULL, *ptr;
int idx, tidx = 0, ret, len;
int tokcount, matchcount;
if (!*input)
return NULL;
if (core->blocksize<=OPSZ) {
eprintf ("error: block size too small\n");
return NULL;
}
if (!(buf = (ut8 *)malloc (core->blocksize)))
return NULL;
if (!(ptr = strdup (input))) {
free (buf);
return NULL;
}
if (!(hits = r_core_asm_hit_list_new ())) {
free (buf);
free (ptr);
return NULL;
}
tokens[0] = NULL;
for (tokcount=0; tokcount<(sizeof (tokens) / sizeof (char*)) - 1; tokcount++) {
tok = strtok (tokcount? NULL: ptr, ",");
if (tok == NULL)
break;
tokens[tokcount] = r_str_trim_head_tail (tok);
}
tokens[tokcount] = NULL;
r_cons_break (NULL, NULL);
for (at = from, matchcount = 0; at < to; at += core->blocksize-OPSZ) {
if (r_cons_singleton ()->breaked)
break;
ret = r_io_read_at (core->io, at, buf, core->blocksize);
if (ret != core->blocksize)
break;
idx = 0, matchcount = 0;
while (idx<core->blocksize) {
r_asm_set_pc (core->assembler, at+idx);
op.buf_asm[0] = 0;
op.buf_hex[0] = 0;
if (!(len = r_asm_disassemble (core->assembler, &op, buf+idx, core->blocksize-idx))) {
idx = (matchcount)? tidx+1: idx+1;
matchcount = 0;
continue;
}
if (tokens[matchcount] && strstr (op.buf_asm, tokens[matchcount])) {
code = r_str_concatf (code, "%s", op.buf_asm);
if (matchcount == tokcount-1) {
if (tokcount == 1)
tidx = idx;
if (!(hit = r_core_asm_hit_new ())) {
r_list_purge (hits);
free (hits);
hits = NULL;
goto beach;
}
hit->addr = at+tidx;
hit->len = idx+len-tidx;
if (hit->len == -1) {
r_core_asm_hit_free (hit);
goto beach;
}
hit->code = strdup (code);
r_list_append (hits, hit);
R_FREE (code);
matchcount = 0;
idx = tidx+1;
} else if (matchcount == 0) {
tidx = idx;
matchcount++;
idx += len;
} else {
matchcount++;
idx += len;
}
} else {
idx = matchcount? tidx+1: idx+1;
R_FREE (code);
matchcount = 0;
}
}
}
r_asm_set_pc (core->assembler, toff);
beach:
free (buf);
free (ptr);
free (code);
return hits;
}
// TODO: add support for byte-per-byte opcode search
R_API RList *r_core_asm_strsearch(RCore *core, const char *input, ut64 from, ut64 to, int maxhits, int regexp) {
RCoreAsmHit *hit;
RAsmOp op;
RList *hits;
ut64 at, toff = core->offset;
ut8 *buf;
int align = core->search->align;
RRegex* rx = NULL;
char *tok, *tokens[1024], *code = NULL, *ptr;
int idx, tidx = 0, ret, len;
int tokcount, matchcount, count = 0;
int matches = 0;
if (!*input)
return NULL;
if (core->blocksize <= OPSZ) {
eprintf ("error: block size too small\n");
return NULL;
}
if (!(buf = (ut8 *)calloc (core->blocksize, 1)))
return NULL;
if (!(ptr = strdup (input))) {
free (buf);
return NULL;
}
if (!(hits = r_core_asm_hit_list_new ())) {
free (buf);
free (ptr);
return NULL;
}
tokens[0] = NULL;
for (tokcount=0; tokcount<(sizeof (tokens) / sizeof (char*)) - 1; tokcount++) {
tok = strtok (tokcount? NULL: ptr, ";");
if (!tok)
break;
tokens[tokcount] = r_str_trim_head_tail (tok);
}
tokens[tokcount] = NULL;
r_cons_break (NULL, NULL);
for (at = from, matchcount = 0; at < to; at += core->blocksize-OPSZ) {
matches = 0;
if (r_cons_singleton ()->breaked)
break;
ret = r_io_read_at (core->io, at, buf, core->blocksize);
if (ret != core->blocksize)
break;
idx = 0, matchcount = 0;
while (idx < core->blocksize) {
ut64 addr = at + idx;
r_asm_set_pc (core->assembler, addr);
op.buf_asm[0] = 0;
op.buf_hex[0] = 0;
if (!(len = r_asm_disassemble (core->assembler, &op, buf+idx, core->blocksize-idx))) {
idx = (matchcount)? tidx+1: idx+1;
matchcount = 0;
continue;
}
matches = true;
if (!strcmp (op.buf_asm, "unaligned"))
matches = false;
if (!strcmp (op.buf_asm, "invalid"))
matches = false;
if (matches && tokens[matchcount]) {
if (!regexp) matches = strstr(op.buf_asm, tokens[matchcount]) != NULL;
else {
rx = r_regex_new (tokens[matchcount], "");
matches = r_regex_exec (rx, op.buf_asm, 0, 0, 0) == 0;
r_regex_free (rx);
}
}
if (align && align>1) {
if (addr % align) {
matches = false;
}
}
if (matches) {
code = r_str_concatf (code, "%s; ", op.buf_asm);
if (matchcount == tokcount-1) {
if (tokcount == 1)
tidx = idx;
if (!(hit = r_core_asm_hit_new ())) {
r_list_purge (hits);
free (hits);
hits = NULL;
goto beach;
}
hit->addr = addr;
hit->len = idx + len - tidx;
if (hit->len == -1) {
r_core_asm_hit_free (hit);
goto beach;
}
code[strlen (code)-2] = 0;
hit->code = strdup (code);
r_list_append (hits, hit);
R_FREE (code);
matchcount = 0;
idx = tidx+1;
if (maxhits) {
count ++;
if (count >= maxhits) {
//eprintf ("Error: search.maxhits reached\n");
goto beach;
}
}
} else if (matchcount == 0) {
tidx = idx;
matchcount++;
idx += len;
} else {
matchcount++;
idx += len;
}
} else {
idx = matchcount? tidx+1: idx+1;
R_FREE (code);
matchcount = 0;
}
}
at += OPSZ;
}
r_asm_set_pc (core->assembler, toff);
beach:
free (buf);
free (ptr);
free (code);
return hits;
}
// TODO: add support for byte-per-byte opcode search
R_API RList *r_core_asm_strsearch(RCore *core, const char *input, ut64 from, ut64 to, int maxhits, int regexp, int everyByte, int mode) {
RCoreAsmHit *hit;
RAsmOp op;
RList *hits;
ut64 at, toff = core->offset;
ut8 *buf;
int align = core->search->align;
RRegex* rx = NULL;
char *tok, *tokens[1024], *code = NULL, *ptr;
int idx, tidx = 0, len = 0;
int tokcount, matchcount, count = 0;
int matches = 0;
const int addrbytes = core->io->addrbytes;
if (!input || !*input) {
return NULL;
}
ut64 usrimm = r_num_math (core->num, input + 1);
if (core->blocksize < 8) {
eprintf ("error: block size too small\n");
return NULL;
}
if (!(buf = (ut8 *)calloc (core->blocksize, 1))) {
return NULL;
}
if (!(ptr = strdup (input))) {
free (buf);
return NULL;
}
if (!(hits = r_core_asm_hit_list_new ())) {
free (buf);
free (ptr);
return NULL;
}
tokens[0] = NULL;
for (tokcount = 0; tokcount < R_ARRAY_SIZE (tokens) - 1; tokcount++) {
tok = strtok (tokcount? NULL: ptr, ";");
if (!tok) {
break;
}
tokens[tokcount] = r_str_trim_head_tail (tok);
}
tokens[tokcount] = NULL;
r_cons_break_push (NULL, NULL);
char *opst = NULL;
for (at = from, matchcount = 0; at < to; at += core->blocksize) {
if (r_cons_is_breaked ()) {
break;
}
if (!r_io_is_valid_offset (core->io, at, 0)) {
break;
}
(void)r_io_read_at (core->io, at, buf, core->blocksize);
idx = 0, matchcount = 0;
while (addrbytes * (idx + 1) <= core->blocksize) {
ut64 addr = at + idx;
if (addr >= to) {
break;
}
r_asm_set_pc (core->assembler, addr);
if (mode == 'i') {
RAnalOp analop = {0};
if (r_anal_op (core->anal, &analop, addr, buf + idx, 15, 0) < 1) {
idx ++; // TODO: honor mininstrsz
continue;
}
if (analop.val == usrimm) {
if (!(hit = r_core_asm_hit_new ())) {
r_list_purge (hits);
R_FREE (hits);
goto beach;
}
hit->addr = addr;
hit->len = analop.size; // idx + len - tidx;
if (hit->len == -1) {
r_core_asm_hit_free (hit);
goto beach;
}
r_asm_disassemble (core->assembler, &op, buf + addrbytes * idx,
core->blocksize - addrbytes * idx);
hit->code = r_str_newf (r_strbuf_get (&op.buf_asm));
idx = (matchcount)? tidx + 1: idx + 1;
matchcount = 0;
r_list_append (hits, hit);
continue;
}
r_anal_op_fini (&analop);
idx ++; // TODO: honor mininstrsz
continue;
} else if (mode == 'e') {
RAnalOp analop = {0};
if (r_anal_op (core->anal, &analop, addr, buf + idx, 15, R_ANAL_OP_MASK_ESIL) < 1) {
idx ++; // TODO: honor mininstrsz
continue;
}
//opsz = analop.size;
opst = strdup (r_strbuf_get (&analop.esil));
r_anal_op_fini (&analop);
} else {
if (!(len = r_asm_disassemble (
core->assembler, &op,
buf + addrbytes * idx,
core->blocksize - addrbytes * idx))) {
idx = (matchcount)? tidx + 1: idx + 1;
matchcount = 0;
continue;
}
//opsz = op.size;
opst = strdup (r_strbuf_get (&op.buf_asm));
}
if (opst) {
matches = strcmp (opst, "invalid") && strcmp (opst, "unaligned");
}
if (matches && tokens[matchcount]) {
if (!regexp) {
matches = strstr (opst, tokens[matchcount]) != NULL;
} else {
rx = r_regex_new (tokens[matchcount], "");
if (r_regex_comp (rx, tokens[matchcount], R_REGEX_EXTENDED|R_REGEX_NOSUB) == 0) {
matches = r_regex_exec (rx, opst, 0, 0, 0) == 0;
}
r_regex_free (rx);
}
}
if (align && align > 1) {
if (addr % align) {
matches = false;
}
}
if (matches) {
code = r_str_appendf (code, "%s; ", opst);
if (matchcount == tokcount - 1) {
if (tokcount == 1) {
tidx = idx;
}
if (!(hit = r_core_asm_hit_new ())) {
r_list_purge (hits);
R_FREE (hits);
goto beach;
}
hit->addr = addr;
hit->len = idx + len - tidx;
if (hit->len == -1) {
r_core_asm_hit_free (hit);
goto beach;
}
code[strlen (code) - 2] = 0;
hit->code = strdup (code);
r_list_append (hits, hit);
R_FREE (code);
matchcount = 0;
idx = tidx + 1;
if (maxhits) {
count++;
if (count >= maxhits) {
//eprintf ("Error: search.maxhits reached\n");
goto beach;
}
}
} else if (!matchcount) {
tidx = idx;
matchcount++;
idx += len;
} else {
matchcount++;
idx += len;
}
} else {
if (everyByte) {
idx = matchcount? tidx + 1: idx + 1;
} else {
idx += R_MAX (1, len);
}
R_FREE (code);
matchcount = 0;
}
R_FREE (opst);
}
}
r_cons_break_pop ();
r_asm_set_pc (core->assembler, toff);
beach:
free (buf);
free (ptr);
free (code);
R_FREE (opst);
r_cons_break_pop ();
return hits;
}
