示例#1
0
文件: map.c 项目: 0xroot/radare2
int main () {
	int ret;
	RIODesc *fd;
	char buf[1024];
	struct r_io_t *io;

	io = r_io_new();

	r_io_plugin_list(io);
	//fd = r_io_open(io, "/bin/ls", R_IO_READ, 0);
	fd = r_io_open(io, "dbg:///bin/ls", R_IO_READ, 0);
	r_io_set_fd(io, fd);

	//r_io_map_add(io, fd, R_IO_READ, 0, 0xf00000, 0xffff);
	r_io_map_add(io, fd->fd, R_IO_READ, 0x8048000, 0, 0xffff);

	memset(buf, 0, 1024);
	//ret = r_io_read_at(io, 0xf00000, buf, 1024);
//	ret = r_io_seek(io, 0x8048000, R_IO_SEEK_SET);
//	printf("seek = 0x%"PFMT64x"\n", ret);
	ret = r_io_read_at (io, 0, (ut8*)buf, 64);
	//ret = r_io_read_at(io, 0x8048000, buf, 64);
	printf("%d = %02x %02x %02x %02x\n", ret, buf[0], buf[1], buf[2], buf[3]);
	r_io_free(io);

	return 0;
}
示例#2
0
文件: desc.c 项目: Jo-Con-El/radare2
R_API ut64 r_io_desc_size(RIO *io, RIODesc *desc){
	RIODesc *old = NULL;
    	ut64 sz = -1;
	
	if (desc && io->fd != desc){
		old = io->fd;
		r_io_set_fd(io, desc);
	}
        
	if (desc) sz = r_io_size(io);
	
	if(old){
		r_io_set_fd(io, old);
	}
	return sz;
}
示例#3
0
文件: io.c 项目: ericfode/radare2
R_API int r_core_write_at(RCore *core, ut64 addr, const ut8 *buf, int size) {
	int ret;
	if (!core->io || !core->file || size<1)
		return R_FALSE;
	ret = r_io_set_fd (core->io, core->file->fd);
	if (ret != -1) {
		ret = r_io_write_at (core->io, addr, buf, size);
		if (addr >= core->offset && addr <= core->offset+core->blocksize)
			r_core_block_read (core, 0);
	}
	return (ret==-1)? R_FALSE: R_TRUE;
}
示例#4
0
文件: io.c 项目: adamharder/radare2
R_API int r_core_read_at(RCore *core, ut64 addr, ut8 *buf, int size) {
	int ret;
	if (!core->io || !core->file || size<1)
		return R_FALSE;
#if 0
	r_io_set_fd (core->io, core->file->fd); // XXX ignore ret? -- ultra slow method.. inverse resolution of io plugin brbrb
	ret = r_io_read_at (core->io, addr, buf, size);
	if (addr>=core->offset && addr<=core->offset+core->blocksize)
		r_core_block_read (core, 0);
#else
	r_io_set_fd (core->io, core->file->fd); // XXX ignore ret? -- ultra slow method.. inverse resolution of io plugin brbrb
	//ret = r_io_read_at (core->io, addr, buf, size);
	r_io_seek (core->io, addr, R_IO_SEEK_SET);
	ret = r_io_read (core->io, buf, size);
	if (ret != size) {
		if (ret>=size || ret<0) ret = 0;
		memset (buf+ret, 0xff, size-ret);
	}
	if (addr>=core->offset && addr<=core->offset+core->blocksize)
		r_core_block_read (core, 0);
#endif
	return (ret==size); //UT64_MAX);
}
示例#5
0
文件: io.c 项目: adamharder/radare2
R_API int r_core_block_read(RCore *core, int next) {
	ut64 off;
	if (core->file == NULL) {
		memset (core->block, 0xff, core->blocksize);
		return -1;
	}
	r_io_set_fd (core->io, core->file->fd);
	off = r_io_seek (core->io, core->offset+((next)?core->blocksize:0), R_IO_SEEK_SET);
	if (off == UT64_MAX) {
		memset (core->block, 0xff, core->blocksize);
		return -1;
	}
	return (int)r_io_read (core->io, core->block, core->blocksize);
}
示例#6
0
文件: io.c 项目: ericfode/radare2
R_API int r_core_read_at(RCore *core, ut64 addr, ut8 *buf, int size) {
	int ret;
	if (!core->io || !core->file || size<1)
		return R_FALSE;
	r_io_set_fd (core->io, core->file->fd); // XXX ignore ret? -- ultra slow method.. inverse resolution of io plugin brbrb
	ret = r_io_read_at (core->io, addr, buf, size);
	if (ret != size) {
		if (ret<size && ret>0)
			memset (buf+ret, 0xff, size-ret);
		else	memset (buf, 0xff, size);
	}
	if (addr>=core->offset && addr<=core->offset+core->blocksize)
		r_core_block_read (core, 0);
	return (ret!=UT64_MAX);
}
示例#7
0
文件: io.c 项目: pixilla/radare2
R_API RIODesc *r_io_open(struct r_io_t *io, const char *file, int flags, int mode) {
	RIODesc *desc = NULL;
	int fd = -2;
	char *uri = strdup (file);
	struct r_io_plugin_t *plugin;
	if (!io) return NULL;
	for (;;) {
		plugin = r_io_plugin_resolve (io, uri);
		if (plugin && plugin->open) {
			desc = plugin->open (io, uri, flags, mode);
			if (io->redirect) {
				free ((void *)uri);
				uri = strdup (io->redirect);
				r_io_redirect (io, NULL);
				continue;
			}
			if (desc != NULL) {
				r_io_desc_add (io, desc);
				fd = desc->fd;
				if (fd != -1)
					r_io_plugin_open (io, fd, plugin);
				if (desc != io->fd)
					io->plugin = plugin;
			}
		}
		break;
	}
	if (fd == -2) {
#if __WINDOWS__
		if (flags & R_IO_WRITE) {
			fd = open (file, O_BINARY | 1);
			if (fd == -1)
				creat (file, O_BINARY);
			fd = open (file, O_BINARY | 1);
		} else fd = open (file, O_BINARY);
#else
		fd = open (file, (flags&R_IO_WRITE)?O_RDWR:O_RDONLY, mode);
#endif
	}
	if (fd >= 0) {
		if (desc == NULL)
			desc = r_io_desc_new (io->plugin, fd, file, flags, mode, NULL);
		r_io_desc_add (io, desc);
		r_io_set_fd (io, desc);
	}
	free ((void *)uri);
	return desc;
}
示例#8
0
文件: io.c 项目: Rogunix/radare2
// TODO: remove int fd here???
R_API int r_io_close(RIO *io, RIODesc *fd) {
	if (io == NULL || fd == NULL)
		return -1;
	int nfd = fd->fd;
	if (r_io_set_fd (io, fd)) {
		RIODesc *desc = r_io_desc_get (io, fd->fd);
		if (desc) {
			r_io_map_del (io, fd->fd);
			r_io_plugin_close (io, fd->fd, io->plugin);
			if (io->plugin && io->plugin->close)
				return io->plugin->close (desc);
			r_io_desc_del (io, desc->fd);
		}
	}
	io->fd = NULL; // unset current fd
	return close (nfd);
}
示例#9
0
文件: io.c 项目: Rogunix/radare2
R_API RIODesc *r_io_open(RIO *io, const char *file, int flags, int mode) {
	RIODesc *desc = __getioplugin (io, file, flags, mode);
	int fd;
	if (desc) {
		fd = desc->fd;
	} else {
		fd = __io_posix_open (io, file, flags, mode);
		if (fd>=0)
			desc = r_io_desc_new (io->plugin,
				fd, file, flags, mode, NULL);
	}
	if (fd >= 0) {
		r_io_desc_add (io, desc);
		r_io_set_fd (io, desc);
	}
	return desc;
}
示例#10
0
文件: cmd_write.c 项目: rydnr/radare2
/* TODO: simplify using r_write */
static int cmd_write(void *data, const char *input) {
    ut64 off;
    ut8 *buf;
    const char *arg;
    int wseek, i, size, len = strlen (input);
    char *tmp, *str, *ostr;
    RCore *core = (RCore *)data;
#define WSEEK(x,y) if(wseek)r_core_seek_delta(x,y)
    wseek = r_config_get_i (core->config, "cfg.wseek");
    str = ostr = strdup (input+1);

    switch (*input) {
    case 'p':
        if (input[1]==' ' && input[2]) {
            r_core_patch (core, input+2);
        } else {
            eprintf ("Usage: wp [rapatch-file]\n"
                     "TODO: rapatch format documentation here\n");
        }
        break;
    case 'r':
        off = r_num_math (core->num, input+1);
        len = (int)off;
        if (len>0) {
            buf = malloc (len);
            if (buf != NULL) {
                r_num_irand ();
                for (i=0; i<len; i++)
                    buf[i] = r_num_rand (256);
                r_core_write_at (core, core->offset, buf, len);
                WSEEK (core, len);
                free (buf);
            } else eprintf ("Cannot allocate %d bytes\n", len);
        }
        break;
    case 'A':
        switch (input[1]) {
        case ' ':
            if (input[2] && input[3]==' ') {
                r_asm_set_pc (core->assembler, core->offset);
                eprintf ("modify (%c)=%s\n", input[2], input+4);
                len = r_asm_modify (core->assembler, core->block, input[2],
                                    r_num_math (core->num, input+4));
                eprintf ("len=%d\n", len);
                if (len>0) {
                    r_core_write_at (core, core->offset, core->block, len);
                    WSEEK (core, len);
                } else eprintf ("r_asm_modify = %d\n", len);
            } else eprintf ("Usage: wA [type] [value]\n");
            break;
        case '?':
        default:
            r_cons_printf ("Usage: wA [type] [value]\n"
                           "Types:\n"
                           " r   raw write value\n"
                           " v   set value (taking care of current address)\n"
                           " d   destination register\n"
                           " 0   1st src register\n"
                           " 1   2nd src register\n"
                           "Example: wA r 0 # e800000000\n");
            break;
        }
        break;
    case 'c':
        switch (input[1]) {
        case 'i':
            r_io_cache_commit (core->io);
            r_core_block_read (core, 0);
            break;
        case 'r':
            r_io_cache_reset (core->io, R_TRUE);
            /* Before loading the core block we have to make sure that if
             * the cache wrote past the original EOF these changes are no
             * longer displayed. */
            memset (core->block, 0xff, core->blocksize);
            r_core_block_read (core, 0);
            break;
        case '-':
            if (input[2]=='*') {
                r_io_cache_reset (core->io, R_TRUE);
            } else if (input[2]==' ') {
                char *p = strchr (input+3, ' ');
                ut64 to, from = core->offset;
                if (p) {
                    *p = 0;
                    from = r_num_math (core->num, input+3);
                    to = r_num_math (core->num, input+3);
                    if (to<from) {
                        eprintf ("Invalid range (from>to)\n");
                        return 0;
                    }
                } else {
                    from = r_num_math (core->num, input+3);
                    to = from + core->blocksize;
                }
                r_io_cache_invalidate (core->io, from, to);
            } else {
                eprintf ("Invalidate write cache at 0x%08"PFMT64x"\n", core->offset);
                r_io_cache_invalidate (core->io, core->offset, core->offset+core->blocksize);
            }
            /* See 'r' above. */
            memset (core->block, 0xff, core->blocksize);
            r_core_block_read (core, 0);
            break;
        case '?':
            r_cons_printf (
                "Usage: wc[ir*?]\n"
                " wc           list all write changes\n"
                " wc- [a] [b]  remove write op at curseek or given addr\n"
                " wc*          \"\" in radare commands\n"
                " wcr          reset all write changes in cache\n"
                " wci          commit write cache\n"
                "NOTE: Requires 'e io.cache=true'\n");
            break;
        case '*':
            r_io_cache_list (core->io, R_TRUE);
            break;
        case '\0':
            r_io_cache_list (core->io, R_FALSE);
            break;
        }
        break;
    case ' ':
        /* write string */
        len = r_str_escape (str);
        r_core_write_at (core, core->offset, (const ut8*)str, len);
#if 0
        r_io_set_fd (core->io, core->file->fd);
        r_io_write_at (core->io, core->offset, (const ut8*)str, len);
#endif
        WSEEK (core, len);
        r_core_block_read (core, 0);
        break;
    case 't':
        if (*str != ' ') {
            eprintf ("Usage: wt file [size]\n");
        } else {
            tmp = strchr (str+1, ' ');
            if (tmp) {
                st64 sz = (st64) r_num_math (core->num, tmp+1);
                *tmp = 0;
                if (sz<1) eprintf ("Invalid length\n");
                else r_core_dump (core, str+1, core->offset, (ut64)sz);
            } else r_file_dump (str+1, core->block, core->blocksize);
        }
        break;
    case 'T':
        eprintf ("TODO: wT // why?\n");
        break;
    case 'f':
        arg = (const char *)(input+((input[1]==' ')?2:1));
        if ((buf = (ut8*) r_file_slurp (arg, &size))) {
            r_io_set_fd (core->io, core->file->fd);
            r_io_write_at (core->io, core->offset, buf, size);
            WSEEK (core, size);
            free(buf);
            r_core_block_read (core, 0);
        } else eprintf ("Cannot open file '%s'\n", arg);
        break;
    case 'F':
        arg = (const char *)(input+((input[1]==' ')?2:1));
        if ((buf = r_file_slurp_hexpairs (arg, &size))) {
            r_io_set_fd (core->io, core->file->fd);
            r_io_write_at (core->io, core->offset, buf, size);
            WSEEK (core, size);
            free (buf);
            r_core_block_read (core, 0);
        } else eprintf ("Cannot open file '%s'\n", arg);
        break;
    case 'w':
        str++;
        len = (len-1)<<1;
        if (len>0) tmp = malloc (len+1);
        else tmp = NULL;
        if (tmp) {
            for (i=0; i<len; i++) {
                if (i%2) tmp[i] = 0;
                else tmp[i] = str[i>>1];
            }
            str = tmp;
            r_io_set_fd (core->io, core->file->fd);
            r_io_write_at (core->io, core->offset, (const ut8*)str, len);
            WSEEK (core, len);
            r_core_block_read (core, 0);
            free (tmp);
        } else eprintf ("Cannot malloc %d\n", len);
        break;
    case 'x':
    {
        int b, len = strlen (input);
        ut8 *buf = malloc (len+1);
        len = r_hex_str2bin (input+1, buf);
        if (len != 0) {
            if (len<0) len = -len+1;
            b = core->block[len]&0xf;
            b |= (buf[len]&0xf0);
            buf[len] = b;
            r_core_write_at (core, core->offset, buf, len);
            WSEEK (core, len);
            r_core_block_read (core, 0);
        } else eprintf ("Error: invalid hexpair string\n");
        free (buf);
    }
    break;
    case 'a':
        switch (input[1]) {
        case 'o':
            if (input[2] == ' ')
                r_core_hack (core, input+3);
            else r_core_hack_help (core);
            break;
        case ' ':
        case '*':
        {   const char *file = input[1]=='*'? input+2: input+1;
            RAsmCode *acode;
            r_asm_set_pc (core->assembler, core->offset);
            acode = r_asm_massemble (core->assembler, file);
            if (acode) {
                if (input[1]=='*') {
                    r_cons_printf ("wx %s\n", acode->buf_hex);
                } else {
                    if (r_config_get_i (core->config, "scr.prompt"))
                        eprintf ("Written %d bytes (%s)=wx %s\n", acode->len, input+1, acode->buf_hex);
                    r_core_write_at (core, core->offset, acode->buf, acode->len);
                    WSEEK (core, acode->len);
                    r_core_block_read (core, 0);
                }
                r_asm_code_free (acode);
            }
        }
        break;
        case 'f':
            if ((input[2]==' '||input[2]=='*')) {
                const char *file = input[2]=='*'? input+4: input+3;
                RAsmCode *acode;
                r_asm_set_pc (core->assembler, core->offset);
                acode = r_asm_assemble_file (core->assembler, file);
                if (acode) {
                    if (input[2]=='*') {
                        r_cons_printf ("wx %s\n", acode->buf_hex);
                    } else {
                        if (r_config_get_i (core->config, "scr.prompt"))
                            eprintf ("Written %d bytes (%s)=wx %s\n", acode->len, input+1, acode->buf_hex);
                        r_core_write_at (core, core->offset, acode->buf, acode->len);
                        WSEEK (core, acode->len);
                        r_core_block_read (core, 0);
                    }
                    r_asm_code_free (acode);
                } else eprintf ("Cannot assemble file\n");
            } else eprintf ("Wrong argument\n");
            break;
        default:
            eprintf ("Usage: wa[of*] [arg]\n"
                     " wa nop           : write nopcode using asm.arch and asm.bits\n"
                     " wa* mov eax, 33  : show 'wx' op with hexpair bytes of sassembled opcode\n"
                     " \"wa nop;nop\"     : assemble more than one instruction (note the quotes)\n"
                     " waf foo.asm      : assemble file and write bytes\n"
                     " wao nop          : convert current opcode into nops\n"
                     " wao?             : show help for assembler operation on current opcode (hack)\n");
            break;
        }
        break;
    case 'b':
    {
        int len = strlen (input);
        ut8 *buf = malloc (len+1);
        if (buf) {
            len = r_hex_str2bin (input+1, buf);
            if (len > 0) {
                r_mem_copyloop (core->block, buf, core->blocksize, len);
                r_core_write_at (core, core->offset, core->block, core->blocksize);
                WSEEK (core, core->blocksize);
                r_core_block_read (core, 0);
            } else eprintf ("Wrong argument\n");
        } else eprintf ("Cannot malloc %d\n", len+1);
    }
    break;
    case 'm':
        size = r_hex_str2bin (input+1, (ut8*)str);
        switch (input[1]) {
        case '\0':
            eprintf ("Current write mask: TODO\n");
            // TODO
            break;
        case '?':
            break;
        case '-':
            r_io_set_write_mask(core->io, 0, 0);
            eprintf ("Write mask disabled\n");
            break;
        case ' ':
            if (size>0) {
                r_io_set_fd (core->io, core->file->fd);
                r_io_set_write_mask (core->io, (const ut8*)str, size);
                WSEEK (core, size);
                eprintf ("Write mask set to '");
                for (i=0; i<size; i++)
                    eprintf ("%02x", str[i]);
                eprintf ("'\n");
            } else eprintf ("Invalid string\n");
            break;
        }
        break;
    case 'v':
    {
        int type = 0;
        ut8 addr1;
        ut16 addr2;
        ut32 addr4, addr4_;
        ut64 addr8;

        switch (input[1]) {
        case '?':
            r_cons_printf ("Usage: wv[size] [value]    # write value of given size\n"
                           "  wv1 234      # write one byte with this value\n"
                           "  wv 0x834002  # write dword with this value\n"
                           "Supported sizes are: 1, 2, 4, 8\n");
            return 0;
        case '1':
            type = 1;
            break;
        case '2':
            type = 2;
            break;
        case '4':
            type = 4;
            break;
        case '8':
            type = 8;
            break;
        }
        off = r_num_math (core->num, input+2);
        r_io_set_fd (core->io, core->file->fd);
        r_io_seek (core->io, core->offset, R_IO_SEEK_SET);
        if (type == 0)
            type = (off&UT64_32U)? 8: 4;
        switch (type) {
        case 1:
            addr1 = (ut8)off;
            r_io_write (core->io, (const ut8 *)&addr1, 1);
            WSEEK (core, 1);
            break;
        case 2:
            addr2 = (ut16)off;
            r_io_write (core->io, (const ut8 *)&addr2, 2);
            WSEEK (core, 2);
            break;
        case 4:
            addr4_ = (ut32)off;
            //drop_endian((ut8*)&addr4_, (ut8*)&addr4, 4); /* addr4_ = addr4 */
            //endian_memcpy((ut8*)&addr4, (ut8*)&addr4_, 4); /* addr4 = addr4_ */
            memcpy ((ut8*)&addr4, (ut8*)&addr4_, 4); // XXX needs endian here too
            r_io_write (core->io, (const ut8 *)&addr4, 4);
            WSEEK (core, 4);
            break;
        case 8:
            /* 8 byte addr */
            memcpy ((ut8*)&addr8, (ut8*)&off, 8); // XXX needs endian here
            //	endian_memcpy((ut8*)&addr8, (ut8*)&off, 8);
            r_io_write (core->io, (const ut8 *)&addr8, 8);
            WSEEK (core, 8);
            break;
        }
        r_core_block_read (core, 0);
    }
    break;
    case 'o':
        switch (input[1]) {
        case 'a':
        case 's':
        case 'A':
        case 'x':
        case 'r':
        case 'l':
        case 'm':
        case 'd':
        case 'o':
        case 'w':
            if (input[2]!=' ') {
                r_cons_printf ("Usage: 'wo%c 00 11 22'\n", input[1]);
                return 0;
            }
        case '2':
        case '4':
            r_core_write_op (core, input+3, input[1]);
            r_core_block_read (core, 0);
            break;
        case 'n':
            r_core_write_op (core, "ff", 'x');
            r_core_block_read (core, 0);
            break;
        case '\0':
        case '?':
        default:
            r_cons_printf (
                "Usage: wo[asmdxoArl24] [hexpairs] @ addr[:bsize]\n"
                "Example:\n"
                "  wox 0x90   ; xor cur block with 0x90\n"
                "  wox 90     ; xor cur block with 0x90\n"
                "  wox 0x0203 ; xor cur block with 0203\n"
                "  woa 02 03  ; add [0203][0203][...] to curblk\n"
                "Supported operations:\n"
                "  wow  ==  write looped value (alias for 'wb')\n"
                "  woa  +=  addition\n"
                "  wos  -=  substraction\n"
                "  wom  *=  multiply\n"
                "  wod  /=  divide\n"
                "  wox  ^=  xor\n"
                "  woo  |=  or\n"
                "  woA  &=  and\n"
                "  wor  >>= shift right\n"
                "  wol  <<= shift left\n"
                "  wo2  2=  2 byte endian swap\n"
                "  wo4  4=  4 byte endian swap\n"
            );
            break;
        }
        break;
    default:
    case '?':
        if (core->oobi) {
            eprintf ("Writing oobi buffer!\n");
            r_io_set_fd (core->io, core->file->fd);
            r_io_write (core->io, core->oobi, core->oobi_len);
            WSEEK (core, core->oobi_len);
            r_core_block_read (core, 0);
        } else r_cons_printf (
                "Usage: w[x] [str] [<file] [<<EOF] [@addr]\n"
                " w foobar     write string 'foobar'\n"
                " wr 10        write 10 random bytes\n"
                " ww foobar    write wide string 'f\\x00o\\x00o\\x00b\\x00a\\x00r\\x00'\n"
                " wa push ebp  write opcode, separated by ';' (use '\"' around the command)\n"
                " waf file     assemble file and write bytes\n"
                " wA r 0       alter/modify opcode at current seek (see wA?)\n"
                " wb 010203    fill current block with cyclic hexpairs\n"
                " wc[ir*?]     write cache commit/reset/list\n"
                " wx 9090      write two intel nops\n"
                " wv eip+34    write 32-64 bit value\n"
                " wo? hex      write in block with operation. 'wo?' fmi\n"
                " wm f0ff      set binary mask hexpair to be used as cyclic write mask\n"
                " wf file      write contents of file at current offset\n"
                " wF file      write contents of hexpairs file here\n"
                " wt file [sz] write to file (from current seek, blocksize or sz bytes)\n"
                " wp file      apply radare patch file. See wp? fmi\n");
        //TODO: add support for offset+seek
        // " wf file o s ; write contents of file from optional offset 'o' and size 's'.\n"
        break;
    }
示例#11
0
/* TODO: simplify using r_write */
static int cmd_write(void *data, const char *input) {
	ut64 off;
	ut8 *buf;
	const char *arg;
	int wseek, i, size, len = strlen (input);
	char *tmp, *str, *ostr;
	RCore *core = (RCore *)data;
	#define WSEEK(x,y) if (wseek)r_core_seek_delta (x,y)
	wseek = r_config_get_i (core->config, "cfg.wseek");
	str = ostr = strdup (input+1);

	switch (*input) {
	case 'h':
		{
		char *p = strchr (input, ' ');
		if (p) {
			while (*p==' ') p++;
			p = r_file_path (p);
			if (p) {
				r_cons_printf ("%s\n", p);
				free (p);
			}
		}
		}
		break;
	case 'e':
		{
		ut64 addr = 0, len = 0, b_size = 0;
		st64 dist = 0;
		ut8* bytes = NULL;
		int cmd_suc = R_FALSE;
		char *input_shadow = NULL, *p = NULL;

		switch (input[1]) {
			case 'n':
				if (input[2] == ' ') {
					len = *input ? r_num_math (core->num, input+3) : 0;
					if (len > 0){
						ut64 cur_off = core->offset;
						cmd_suc = r_core_extend_at (core, core->offset, len);
						core->offset = cur_off;
						r_core_block_read (core, 0);
					}
				}
				break;
			case 'N':
				if (input[2] == ' ') {
					input += 3;
					while (*input && *input == ' ') input++;
					addr = r_num_math (core->num, input);
					while (*input && *input != ' ') input++;
					input++;
					len = *input ? r_num_math (core->num, input) : 0;
					if (len > 0){
						ut64 cur_off = core->offset;
						cmd_suc = r_core_extend_at (core, addr, len);
						cmd_suc = r_core_seek (core, cur_off, 1);
						core->offset = addr;
						r_core_block_read (core, 0);
					}
				}
				break;
			case 'x':
				if (input[2] == ' ') {
					input+=2;
					len = *input ? strlen (input) : 0;
					bytes = len > 1? malloc (len+1) : NULL;
					len = bytes ? r_hex_str2bin (input, bytes) : 0;
					if (len > 0) {
						ut64 cur_off = core->offset;
						cmd_suc = r_core_extend_at (core, cur_off, len);
						if (cmd_suc) {
							r_core_write_at (core, cur_off, bytes, len);
						}
						core->offset = cur_off;
						r_core_block_read (core, 0);
					}
					free (bytes);
				}
				break;
			case 'X':
				if (input[2] == ' ') {
					addr = r_num_math (core->num, input+3);
					input += 3;
					while (*input && *input != ' ') input++;
					input++;
					len = *input ? strlen (input) : 0;
					bytes = len > 1? malloc (len+1) : NULL;
					len = bytes ? r_hex_str2bin (input, bytes) : 0;
					if (len > 0) {
						//ut64 cur_off = core->offset;
						cmd_suc = r_core_extend_at (core, addr, len);
						if (cmd_suc) {
							r_core_write_at (core, addr, bytes, len);
						}
						core->offset = addr;
						r_core_block_read (core, 0);
					}
					free (bytes);
				}
				break;
			case 's':
				input +=  3;
				while (*input && *input == ' ') input++;
				len = strlen (input);
				input_shadow = len > 0? malloc (len+1): 0;

				// since the distance can be negative,
				// the r_num_math will perform an unwanted operation
				// the solution is to tokenize the string :/
				if (input_shadow) {
					strncpy (input_shadow, input, len+1);
					p = strtok (input_shadow, " ");
					addr = p && *p ? r_num_math (core->num, p) : 0;

					p = strtok (NULL, " ");
					dist = p && *p ? r_num_math (core->num, p) : 0;

					p = strtok (NULL, " ");
					b_size = p && *p ? r_num_math (core->num, p) : 0;
					if (dist != 0){
						r_core_shift_block (core, addr, b_size, dist);
						r_core_seek (core, addr, 1);
						cmd_suc = R_TRUE;
					}
				}
				free (input_shadow);
				break;
			case '?':
			default:
				cmd_suc = R_FALSE;
		}


		if (cmd_suc == R_FALSE) {
			r_cons_printf ("|Usage: write extend\n"
			"wen <num>               insert num null bytes at current offset\n"
			"wex <hex_bytes>         insert bytes at current offset\n"
			"weN <addr> <len>        insert bytes at address\n"
			"weX <addr> <hex_bytes>  insert bytes at address\n"
			"wes <addr>  <dist> <block_size>   shift a blocksize left or write in the editor\n"
			);
		}
		}
		break;
	case 'p':
		if (input[1]=='-' || (input[1]==' '&&input[2]=='-')) {
			const char *tmpfile = ".tmp";
			char *out = r_core_editor (core, NULL);
			if (out) {
				// XXX hacky .. patch should support str, not only file
				r_file_dump (tmpfile, (ut8*)out, strlen (out));
				r_core_patch (core, tmpfile);
				r_file_rm (tmpfile);
				free (out);
			}
		} else {
			if (input[1]==' ' && input[2]) {
				r_core_patch (core, input+2);
			} else {
				eprintf ("Usage: wp [-|r2patch-file]\n"
			         "TODO: rapatch format documentation here\n");
			}
		}
		break;
	case 'u':
		// TODO: implement it in an API RCore.write_unified_hexpatch() is ETOOLONG
		if (input[1]==' ') {
			char *data = r_file_slurp (input+2, NULL);
			if (data) {
				char sign = ' ';
				int line = 0, offs = 0, hexa = 0;
				int newline = 1;
				for (i=0; data[i]; i++) {
					switch (data[i]) {
					case '+':
						if (newline)
							sign = 1;
						break;
					case '-':
						if (newline) {
							sign = 0;
							offs = i + ((data[i+1]==' ')?2:1);
						}
						break;
					case ' ':
						data[i] = 0;
						if (sign) {
							if (!line) line = i+1;
							else
							if (!hexa) hexa = i+1;
						}
						break;
					case '\r':
						break;
					case '\n':
						newline = 1;
						if (sign == -1) {
							offs = 0;
							line = 0;
							hexa = 0;
						} else if (sign) {
							if (offs && hexa) {
								r_cons_printf ("wx %s @ %s\n", data+hexa, data+offs);
							} else eprintf ("food\n");
							offs = 0;
							line = 0;
						} else hexa = 0;
						sign = -1;
						continue;
					}
					newline = 0;
				}
				free (data);
			}
		} else {
			eprintf ("|Usage: wu [unified-diff-patch]    # see 'cu'\n");
		}
		break;
	case 'r':
		off = r_num_math (core->num, input+1);
		len = (int)off;
		if (len>0) {
			buf = malloc (len);
			if (buf != NULL) {
				r_num_irand ();
				for (i=0; i<len; i++)
					buf[i] = r_num_rand (256);
				r_core_write_at (core, core->offset, buf, len);
				WSEEK (core, len);
				free (buf);
			} else eprintf ("Cannot allocate %d bytes\n", len);
		}
		break;
	case 'A':
		switch (input[1]) {
		case ' ':
			if (input[2] && input[3]==' ') {
				r_asm_set_pc (core->assembler, core->offset);
				eprintf ("modify (%c)=%s\n", input[2], input+4);
				len = r_asm_modify (core->assembler, core->block, input[2],
					r_num_math (core->num, input+4));
				eprintf ("len=%d\n", len);
				if (len>0) {
					r_core_write_at (core, core->offset, core->block, len);
					WSEEK (core, len);
				} else eprintf ("r_asm_modify = %d\n", len);
			} else eprintf ("Usage: wA [type] [value]\n");
			break;
		case '?':
		default:
			r_cons_printf ("|Usage: wA [type] [value]\n"
			"|Types:\n"
			"| r   raw write value\n"
			"| v   set value (taking care of current address)\n"
			"| d   destination register\n"
			"| 0   1st src register\n"
			"| 1   2nd src register\n"
			"|Example: wA r 0 # e800000000\n");
			break;
		}
		break;
	case 'c':
		switch (input[1]) {
		case 'i':
			r_io_cache_commit (core->io);
			r_core_block_read (core, 0);
			break;
		case 'r':
			r_io_cache_reset (core->io, R_TRUE);
			/* Before loading the core block we have to make sure that if
			 * the cache wrote past the original EOF these changes are no
			 * longer displayed. */
			memset (core->block, 0xff, core->blocksize);
			r_core_block_read (core, 0);
			break;
		case '-':
			if (input[2]=='*') {
				r_io_cache_reset (core->io, R_TRUE);
			} else if (input[2]==' ') {
				char *p = strchr (input+3, ' ');
				ut64 to, from = core->offset;
				if (p) {
					*p = 0;
					from = r_num_math (core->num, input+3);
					to = r_num_math (core->num, input+3);
					if (to<from) {
						eprintf ("Invalid range (from>to)\n");
						return 0;
					}
				} else {
					from = r_num_math (core->num, input+3);
					to = from + core->blocksize;
				}
				r_io_cache_invalidate (core->io, from, to);
			} else {
				eprintf ("Invalidate write cache at 0x%08"PFMT64x"\n", core->offset);
				r_io_cache_invalidate (core->io, core->offset, core->offset+core->blocksize);
			}
			/* See 'r' above. */
			memset (core->block, 0xff, core->blocksize);
			r_core_block_read (core, 0);
			break;
		case '?':
			r_cons_printf (
			"|Usage: wc[ir*?]\n"
			"| wc           list all write changes\n"
			"| wc- [a] [b]  remove write op at curseek or given addr\n"
			"| wc*          \"\" in radare commands\n"
			"| wcr          reset all write changes in cache\n"
			"| wci          commit write cache\n"
			"|NOTE: Requires 'e io.cache=true'\n");
			break;
		case '*':
			r_io_cache_list (core->io, R_TRUE);
			break;
		case '\0':
			if (!r_config_get_i (core->config, "io.cache"))
				eprintf ("[warning] e io.cache must be true\n");
			r_io_cache_list (core->io, R_FALSE);
			break;
		}
		break;
	case ' ':
		/* write string */
		len = r_str_unescape (str);
		r_core_write_at (core, core->offset, (const ut8*)str, len);
#if 0
		r_io_set_fd (core->io, core->file->fd);
		r_io_write_at (core->io, core->offset, (const ut8*)str, len);
#endif
		WSEEK (core, len);
		r_core_block_read (core, 0);
		break;
	case 't':
		if (*str != ' ') {
			eprintf ("Usage: wt file [size]\n");
		} else {
			tmp = strchr (str+1, ' ');
			if (tmp) {
				st64 sz = (st64) r_num_math (core->num, tmp+1);
				*tmp = 0;
				if (sz<1) eprintf ("Invalid length\n");
				else r_core_dump (core, str+1, core->offset, (ut64)sz);
			} else r_file_dump (str+1, core->block, core->blocksize);
		}
		break;
	case 'T':
		eprintf ("TODO: wT // why?\n");
		break;
	case 'f':
		arg = (const char *)(input+((input[1]==' ')?2:1));
		if (!strcmp (arg, "-")) {
			char *out = r_core_editor (core, NULL);
			if (out) {
				r_io_write_at (core->io, core->offset,
					(ut8*)out, strlen (out));
				free (out);
			}
		} else
		if ((buf = (ut8*) r_file_slurp (arg, &size))) {
			r_io_set_fd (core->io, core->file->fd);
			r_io_write_at (core->io, core->offset, buf, size);
			WSEEK (core, size);
			free (buf);
			r_core_block_read (core, 0);
		} else eprintf ("Cannot open file '%s'\n", arg);
		break;
	case 'F':
		arg = (const char *)(input+((input[1]==' ')?2:1));
		if (!strcmp (arg, "-")) {
			int len;
			ut8 *out;
			char *in = r_core_editor (core, NULL);
			if (in) {
				out = (ut8 *)strdup (in);
				if (out) {
					len = r_hex_str2bin (in, out);
					if (len>0)
						r_io_write_at (core->io, core->offset, out, len);
					free (out);
				}
				free (in);
			}
		} else
		if ((buf = r_file_slurp_hexpairs (arg, &size))) {
			r_io_set_fd (core->io, core->file->fd);
			r_io_write_at (core->io, core->offset, buf, size);
			WSEEK (core, size);
			free (buf);
			r_core_block_read (core, 0);
		} else eprintf ("Cannot open file '%s'\n", arg);
		break;
	case 'w':
		str++;
		len = (len-1)<<1;
		if (len>0) tmp = malloc (len+1);
		else tmp = NULL;
		if (tmp) {
			for (i=0; i<len; i++) {
				if (i%2) tmp[i] = 0;
				else tmp[i] = str[i>>1];
			}
			str = tmp;
			r_io_set_fd (core->io, core->file->fd);
			r_io_write_at (core->io, core->offset, (const ut8*)str, len);
			WSEEK (core, len);
			r_core_block_read (core, 0);
			free (tmp);
		} else eprintf ("Cannot malloc %d\n", len);
		break;
	case 'x':
		{
		int b, len = strlen (input);
		ut8 *buf = malloc (len+1);
		len = r_hex_str2bin (input+1, buf);
		if (len != 0) {
			if (len<0) len = -len+1;
			if (len<core->blocksize) {
				b = core->block[len]&0xf;
				b |= (buf[len]&0xf0);
			} else b = buf[len];
			buf[len] = b;
			r_core_write_at (core, core->offset, buf, len);
			WSEEK (core, len);
			r_core_block_read (core, 0);
		} else eprintf ("Error: invalid hexpair string\n");
		free (buf);
		}
		break;
	case 'a':
		switch (input[1]) {
		case 'o':
			if (input[2] == ' ')
				r_core_hack (core, input+3);
			else r_core_hack_help (core);
			break;
		case ' ':
		case '*':
			{ const char *file = input[1]=='*'? input+2: input+1;
			RAsmCode *acode;
			r_asm_set_pc (core->assembler, core->offset);
			acode = r_asm_massemble (core->assembler, file);
			if (acode) {
				if (input[1]=='*') {
					r_cons_printf ("wx %s\n", acode->buf_hex);
				} else {
					if (r_config_get_i (core->config, "scr.prompt"))
						eprintf ("Written %d bytes (%s) = wx %s\n", acode->len, input+2, acode->buf_hex);
					r_core_write_at (core, core->offset, acode->buf, acode->len);
					WSEEK (core, acode->len);
					r_core_block_read (core, 0);
				}
				r_asm_code_free (acode);
			}
			} break;
		case 'f':
			if ((input[2]==' '||input[2]=='*')) {
				const char *file = input[2]=='*'? input+4: input+3;
				RAsmCode *acode;
				r_asm_set_pc (core->assembler, core->offset);
				acode = r_asm_assemble_file (core->assembler, file);
				if (acode) {
					if (input[2]=='*') {
						r_cons_printf ("wx %s\n", acode->buf_hex);
					} else {
						if (r_config_get_i (core->config, "scr.prompt"))
						eprintf ("Written %d bytes (%s)=wx %s\n", acode->len, input+1, acode->buf_hex);
						r_core_write_at (core, core->offset, acode->buf, acode->len);
						WSEEK (core, acode->len);
						r_core_block_read (core, 0);
					}
					r_asm_code_free (acode);
				} else eprintf ("Cannot assemble file\n");
			} else eprintf ("Wrong argument\n");
			break;
		default:
			r_cons_printf ("|Usage: wa[of*] [arg]\n"
				"| wa nop           : write nopcode using asm.arch and asm.bits\n"
				"| wa* mov eax, 33  : show 'wx' op with hexpair bytes of sassembled opcode\n"
				"| \"wa nop;nop\"     : assemble more than one instruction (note the quotes)\n"
				"| waf foo.asm      : assemble file and write bytes\n"
				"| wao nop          : convert current opcode into nops\n"
				"| wao?             : show help for assembler operation on current opcode (hack)\n");
			break;
		}
		break;
	case 'b':
		{
		int len = strlen (input);
		ut8 *buf = malloc (len+1);
		if (buf) {
			len = r_hex_str2bin (input+1, buf);
			if (len > 0) {
				r_mem_copyloop (core->block, buf, core->blocksize, len);
				r_core_write_at (core, core->offset, core->block, core->blocksize);
				WSEEK (core, core->blocksize);
				r_core_block_read (core, 0);
			} else eprintf ("Wrong argument\n");
			free (buf);
		} else eprintf ("Cannot malloc %d\n", len+1);
		}
		break;
	case 'm':
		size = r_hex_str2bin (input+1, (ut8*)str);
		switch (input[1]) {
		case '\0':
			eprintf ("Current write mask: TODO\n");
			// TODO
			break;
		case '?':
			break;
		case '-':
			r_io_set_write_mask (core->io, 0, 0);
			eprintf ("Write mask disabled\n");
			break;
		case ' ':
			if (size>0) {
				r_io_set_fd (core->io, core->file->fd);
				r_io_set_write_mask (core->io, (const ut8*)str, size);
				WSEEK (core, size);
				eprintf ("Write mask set to '");
				for (i=0; i<size; i++)
					eprintf ("%02x", str[i]);
				eprintf ("'\n");
			} else eprintf ("Invalid string\n");
			break;
		}
		break;
	case 'v':
		{
			int type = 0;
			ut8 addr1;
			ut16 addr2;
			ut32 addr4, addr4_;
			ut64 addr8;

			switch (input[1]) {
			case '?':
				r_cons_printf ("|Usage: wv[size] [value]    # write value of given size\n"
					"|  wv1 234      # write one byte with this value\n"
					"|  wv 0x834002  # write dword with this value\n"
					"|Supported sizes are: 1, 2, 4, 8\n");
				return 0;
			case '1': type = 1; break;
			case '2': type = 2; break;
			case '4': type = 4; break;
			case '8': type = 8; break;
			}
			off = r_num_math (core->num, input+2);
			r_io_set_fd (core->io, core->file->fd);
			r_io_seek (core->io, core->offset, R_IO_SEEK_SET);
			if (type == 0)
				type = (off&UT64_32U)? 8: 4;
			switch (type) {
			case 1:
				addr1 = (ut8)off;
				r_io_write (core->io, (const ut8 *)&addr1, 1);
				WSEEK (core, 1);
				break;
			case 2:
				addr2 = (ut16)off;
				r_io_write (core->io, (const ut8 *)&addr2, 2);
				WSEEK (core, 2);
				break;
			case 4:
				addr4_ = (ut32)off;
				//drop_endian((ut8*)&addr4_, (ut8*)&addr4, 4); /* addr4_ = addr4 */
				//endian_memcpy((ut8*)&addr4, (ut8*)&addr4_, 4); /* addr4 = addr4_ */
				memcpy ((ut8*)&addr4, (ut8*)&addr4_, 4); // XXX needs endian here too
				r_io_write (core->io, (const ut8 *)&addr4, 4);
				WSEEK (core, 4);
				break;
			case 8:
				/* 8 byte addr */
				memcpy ((ut8*)&addr8, (ut8*)&off, 8); // XXX needs endian here
			//	endian_memcpy((ut8*)&addr8, (ut8*)&off, 8);
				r_io_write (core->io, (const ut8 *)&addr8, 8);
				WSEEK (core, 8);
				break;
			}
			r_core_block_read (core, 0);
		}
		break;
	case 'o':
		switch (input[1]) {
			case 'a':
			case 's':
			case 'e':
			case 'A':
			case 'x':
			case 'r':
			case 'l':
			case 'm':
			case 'd':
			case 'o':
			case 'w':
				if (input[2]!=' ') {
					if (input[1]=='e') r_cons_printf (
						"Usage: 'woe from-to step'\n");
					else r_cons_printf (
						"Usage: 'wo%c 00 11 22'\n", input[1]);
					free (ostr);
					return 0;
				}
			case '2':
			case '4':
				r_core_write_op (core, input+3, input[1]);
				r_core_block_read (core, 0);
				break;
			case 'R':
				r_core_cmd0 (core, "wr $b");
				break;
			case 'n':
				r_core_write_op (core, "ff", 'x');
				r_core_block_read (core, 0);
				break;
			case '\0':
			case '?':
			default:
				r_cons_printf (
					"|Usage: wo[asmdxoArl24] [hexpairs] @ addr[:bsize]\n"
					"|Example:\n"
					"|  wox 0x90   ; xor cur block with 0x90\n"
					"|  wox 90     ; xor cur block with 0x90\n"
					"|  wox 0x0203 ; xor cur block with 0203\n"
					"|  woa 02 03  ; add [0203][0203][...] to curblk\n"
					"|  woe 02 03  \n"
					"|Supported operations:\n"
					"|  wow  ==  write looped value (alias for 'wb')\n"
					"|  woa  +=  addition\n"
					"|  wos  -=  substraction\n"
					"|  wom  *=  multiply\n"
					"|  wod  /=  divide\n"
					"|  wox  ^=  xor\n"
					"|  woo  |=  or\n"
					"|  woA  &=  and\n"
					"|  woR  random bytes (alias for 'wr $b'\n"
					"|  wor  >>= shift right\n"
					"|  wol  <<= shift left\n"
					"|  wo2  2=  2 byte endian swap\n"
					"|  wo4  4=  4 byte endian swap\n"
						);
				break;
		}
		break;
	case 's':
		{
			ut8 ulen;
			len = r_str_unescape (str+1);
			if (len>255) {
				eprintf ("Too large\n");
			} else {
				ulen = (ut8)len;
				r_core_write_at (core, core->offset, &ulen, 1);
				r_core_write_at (core, core->offset+1, (const ut8*)str+1, len);
				WSEEK (core, len);
				r_core_block_read (core, 0);
			}
		}
		break;
	default:
	case '?':
		if (core->oobi) {
			eprintf ("Writing oobi buffer!\n");
			r_io_set_fd (core->io, core->file->fd);
			r_io_write (core->io, core->oobi, core->oobi_len);
			WSEEK (core, core->oobi_len);
			r_core_block_read (core, 0);
		} else r_cons_printf (
			"|Usage: w[x] [str] [<file] [<<EOF] [@addr]\n"
			"| w foobar     write string 'foobar'\n"
			"| wh r2        whereis/which shell command\n"
			"| wr 10        write 10 random bytes\n"
			"| ww foobar    write wide string 'f\\x00o\\x00o\\x00b\\x00a\\x00r\\x00'\n"
			"| wa push ebp  write opcode, separated by ';' (use '\"' around the command)\n"
			"| waf file     assemble file and write bytes\n"
			"| wA r 0       alter/modify opcode at current seek (see wA?)\n"
			"| wb 010203    fill current block with cyclic hexpairs\n"
			"| wc[ir*?]     write cache undo/commit/reset/list (io.cache)\n"
			"| wx 9090      write two intel nops\n"
			"| wv eip+34    write 32-64 bit value\n"
			"| wo? hex      write in block with operation. 'wo?' fmi\n"
			"| wm f0ff      set binary mask hexpair to be used as cyclic write mask\n"
			"| ws pstring   write 1 byte for length and then the string\n"
			"| wf -|file    write contents of file at current offset\n"
			"| wF -|file    write contents of hexpairs file here\n"
			"| wp -|file    apply radare patch file. See wp? fmi\n"
			"| wt file [sz] write to file (from current seek, blocksize or sz bytes)\n"
			);
			//TODO: add support for offset+seek
			// " wf file o s ; write contents of file from optional offset 'o' and size 's'.\n"
		break;
	}