def parse_txt(mnemo, attrib, txt, symbol_pool=None, gen_label_index=0):
if symbol_pool is None:
symbol_pool = asmbloc.asm_symbol_pool()
lines_text = []
lines_data = []
lines_bss = []
C_NEXT = asmbloc.asm_constraint.c_next
C_TO = asmbloc.asm_constraint.c_to
lines = lines_text
# parse each line
for line in txt.split('\n'):
# empty
if re.match(r'\s*$', line):
continue
# comment
if re.match(r'\s*;\S*', line):
continue
# labels to forget
r = re.match(r'\s*\.LF[BE]\d\s*:', line)
if r:
continue
# label beginning with .L
r = re.match(r'\s*(\.L\S+)\s*:', line)
if r:
l = r.groups()[0]
l = symbol_pool.getby_name_create(l)
lines.append(l)
continue
# directive
if re.match(r'\s*\.', line):
r = re.match(r'\s*\.(\S+)', line)
directive = r.groups()[0]
if directive == 'text':
lines = lines_text
continue
if directive == 'data':
lines = lines_data
continue
if directive == 'bss':
lines = lines_bss
continue
if directive in ['string', 'ascii']:
# XXX HACK
line = line.replace(r'\n', '\n').replace(r'\r', '\r')
raw = line[line.find(r'"') + 1:line.rfind(r"'")]
raw = raw.decode('string_escape')
if directive == 'string':
raw += "\x00"
lines.append(asmbloc.asm_raw(raw))
continue
if directive == 'ustring':
# XXX HACK
line = line.replace(r'\n', '\n').replace(r'\r', '\r')
raw = line[line.find(r'"') + 1:line.rfind(r"'")] + "\x00"
raw = raw.decode('string_escape')
raw = "".join([string + '\x00' for string in raw])
lines.append(asmbloc.asm_raw(raw))
continue
if directive in declarator:
data_raw = line[r.end():].split(' ', 1)[1]
data_raw = data_raw.split(',')
size = declarator[directive]
data_int = []
# parser
base_expr = gen_base_expr()[2]
my_var_parser = parse_ast(lambda x: m2_expr.ExprId(x, size),
lambda x:
m2_expr.ExprInt_fromsize(size, x))
base_expr.setParseAction(my_var_parser)
for b in data_raw:
b = b.strip()
x = base_expr.parseString(b)[0]
data_int.append(x.canonize())
raw = data_int
x = asmbloc.asm_raw(raw)
x.element_size = size
lines.append(x)
continue
if directive == 'comm':
# TODO
continue
if directive == 'split': # custom command
x = asmbloc.asm_raw()
x.split = True
lines.append(x)
continue
if directive == 'dontsplit': # custom command
lines.append(asmbloc.asm_raw())
continue
if directive == "align":
align_value = int(line[r.end():])
lines.append(DirectiveAlign(align_value))
continue
if directive in ['file', 'intel_syntax', 'globl', 'local',
'type', 'size', 'align', 'ident', 'section']:
continue
if directive[0:4] == 'cfi_':
continue
raise ValueError("unknown directive %s" % str(directive))
# label
r = re.match(r'\s*(\S+)\s*:', line)
if r:
l = r.groups()[0]
l = symbol_pool.getby_name_create(l)
lines.append(l)
continue
# code
if ';' in line:
line = line[:line.find(';')]
line = line.strip(' ').strip('\t')
instr = mnemo.fromstring(line, attrib)
# replace orphan asm_label with labels from symbol_pool
replace_orphan_labels(instr, symbol_pool)
if instr.dstflow():
instr.dstflow2label(symbol_pool)
lines.append(instr)
asmbloc.log_asmbloc.info("___pre asm oki___")
# make blocs
blocs_sections = []
bloc_num = 0
b = None
for lines in [lines_text, lines_data, lines_bss]:
state = 0
i = 0
blocs = []
blocs_sections.append(blocs)
bloc_to_nlink = None
block_may_link = False
while i < len(lines):
# no current bloc
if state == 0:
if not isinstance(lines[i], asmbloc.asm_label):
l = guess_next_new_label(symbol_pool)
lines[i:i] = [l]
else:
l = lines[i]
b = asmbloc.asm_bloc(l, alignment=mnemo.alignment)
b.bloc_num = bloc_num
bloc_num += 1
blocs.append(b)
state = 1
i += 1
if bloc_to_nlink:
bloc_to_nlink.addto(asmbloc.asm_constraint(b.label,
C_NEXT))
bloc_to_nlink = None
# in bloc
elif state == 1:
if isinstance(lines[i], asmbloc.asm_raw):
if hasattr(lines[i], 'split'):
state = 0
block_may_link = False
i += 1
else:
state = 1
block_may_link = True
b.addline(lines[i])
i += 1
elif isinstance(lines[i], DirectiveAlign):
b.alignment = lines[i].alignment
i += 1
# asmbloc.asm_label
elif isinstance(lines[i], asmbloc.asm_label):
if block_may_link:
b.addto(
asmbloc.asm_constraint(lines[i], C_NEXT))
block_may_link = False
state = 0
# instruction
else:
b.addline(lines[i])
if lines[i].dstflow():
for x in lines[i].getdstflow(symbol_pool):
if not isinstance(x, m2_expr.ExprId):
continue
if x in mnemo.regs.all_regs_ids:
continue
b.addto(asmbloc.asm_constraint(x, C_TO))
# TODO XXX redo this really
if not lines[i].breakflow() and i + 1 < len(lines):
if isinstance(lines[i + 1], asmbloc.asm_label):
l = lines[i + 1]
else:
l = guess_next_new_label(symbol_pool)
lines[i + 1:i + 1] = [l]
else:
state = 0
if lines[i].splitflow():
bloc_to_nlink = b
if not lines[i].breakflow() or lines[i].splitflow():
block_may_link = True
else:
block_may_link = False
i += 1
for block in blocs_sections[0]:
asmbloc.log_asmbloc.info(block)
return blocs_sections, symbol_pool
def parse_txt(mnemo, attrib, txt, symbol_pool=None):
"""Parse an assembly listing. Returns a couple (blocks, symbol_pool), where
blocks is a list of asm_bloc and symbol_pool the associated asm_symbol_pool
@mnemo: architecture used
@attrib: architecture attribute
@txt: assembly listing
@symbol_pool: (optional) the asm_symbol_pool instance used to handle labels
of the listing
"""
if symbol_pool is None:
symbol_pool = asmbloc.asm_symbol_pool()
C_NEXT = asmbloc.asm_constraint.c_next
C_TO = asmbloc.asm_constraint.c_to
lines = []
# parse each line
for line in txt.split('\n'):
# empty
if EMPTY_RE.match(line):
continue
# comment
if COMMENT_RE.match(line):
continue
# labels to forget
if FORGET_LABEL_RE.match(line):
continue
# label beginning with .L
match_re = LABEL_RE.match(line)
if match_re:
label_name = match_re.group(1)
label = symbol_pool.getby_name_create(label_name)
lines.append(label)
continue
# directive
if DIRECTIVE_START_RE.match(line):
match_re = DIRECTIVE_RE.match(line)
directive = match_re.group(1)
if directive in ['text', 'data', 'bss']:
continue
if directive in ['string', 'ascii']:
# XXX HACK
line = line.replace(r'\n', '\n').replace(r'\r', '\r')
raw = line[line.find(r'"') + 1:line.rfind(r'"')]
raw = raw.decode('string_escape')
if directive == 'string':
raw += "\x00"
lines.append(asmbloc.asm_raw(raw))
continue
if directive == 'ustring':
# XXX HACK
line = line.replace(r'\n', '\n').replace(r'\r', '\r')
raw = line[line.find(r'"') + 1:line.rfind(r'"')] + "\x00"
raw = raw.decode('string_escape')
raw = "".join([string + '\x00' for string in raw])
lines.append(asmbloc.asm_raw(raw))
continue
if directive in declarator:
data_raw = line[match_re.end():].split(' ', 1)[1]
data_raw = data_raw.split(',')
size = declarator[directive]
expr_list = []
# parser
base_expr = gen_base_expr()[2]
my_var_parser = parse_ast(lambda x: m2_expr.ExprId(x, size),
lambda x: m2_expr.ExprInt(x, size))
base_expr.setParseAction(my_var_parser)
for element in data_raw:
element = element.strip()
element_expr = base_expr.parseString(element)[0]
expr_list.append(element_expr.canonize())
raw_data = asmbloc.asm_raw(expr_list)
raw_data.element_size = size
lines.append(raw_data)
continue
if directive == 'comm':
# TODO
continue
if directive == 'split': # custom command
lines.append(DirectiveSplit())
continue
if directive == 'dontsplit': # custom command
lines.append(DirectiveDontSplit())
continue
if directive == "align":
align_value = int(line[match_re.end():], 0)
lines.append(DirectiveAlign(align_value))
continue
if directive in [
'file', 'intel_syntax', 'globl', 'local', 'type', 'size',
'align', 'ident', 'section'
]:
continue
if directive[0:4] == 'cfi_':
continue
raise ValueError("unknown directive %s" % str(directive))
# label
match_re = LABEL_RE.match(line)
if match_re:
label_name = match_re.group(1)
label = symbol_pool.getby_name_create(label_name)
lines.append(label)
continue
# code
if ';' in line:
line = line[:line.find(';')]
line = line.strip(' ').strip('\t')
instr = mnemo.fromstring(line, attrib)
# replace orphan asm_label with labels from symbol_pool
replace_orphan_labels(instr, symbol_pool)
if instr.dstflow():
instr.dstflow2label(symbol_pool)
lines.append(instr)
asmbloc.log_asmbloc.info("___pre asm oki___")
# make blocks
cur_block = None
state = STATE_NO_BLOC
i = 0
blocks = []
block_to_nlink = None
block_may_link = False
delayslot = 0
while i < len(lines):
if delayslot:
delayslot -= 1
if delayslot == 0:
state = STATE_NO_BLOC
line = lines[i]
# no current block
if state == STATE_NO_BLOC:
if isinstance(line, DirectiveDontSplit):
block_to_nlink = cur_block
i += 1
continue
elif isinstance(line, DirectiveSplit):
block_to_nlink = None
i += 1
continue
elif not isinstance(line, asmbloc.asm_label):
# First line must be a label. If it's not the case, generate
# it.
label = guess_next_new_label(symbol_pool)
cur_block = asmbloc.asm_bloc(label, alignment=mnemo.alignment)
else:
cur_block = asmbloc.asm_bloc(line, alignment=mnemo.alignment)
i += 1
# Generate the current bloc
blocks.append(cur_block)
state = STATE_IN_BLOC
if block_to_nlink:
block_to_nlink.addto(
asmbloc.asm_constraint(cur_block.label, C_NEXT))
block_to_nlink = None
continue
# in block
elif state == STATE_IN_BLOC:
if isinstance(line, DirectiveSplit):
state = STATE_NO_BLOC
block_to_nlink = None
elif isinstance(line, DirectiveDontSplit):
state = STATE_NO_BLOC
block_to_nlink = cur_block
elif isinstance(line, DirectiveAlign):
cur_block.alignment = line.alignment
elif isinstance(line, asmbloc.asm_raw):
cur_block.addline(line)
block_to_nlink = cur_block
elif isinstance(line, asmbloc.asm_label):
if block_to_nlink:
cur_block.addto(asmbloc.asm_constraint(line, C_NEXT))
block_to_nlink = None
state = STATE_NO_BLOC
continue
# instruction
elif isinstance(line, instruction):
cur_block.addline(line)
block_to_nlink = cur_block
if not line.breakflow():
i += 1
continue
if delayslot:
raise RuntimeError("Cannot have breakflow in delayslot")
if line.dstflow():
for dst in line.getdstflow(symbol_pool):
if not isinstance(dst, m2_expr.ExprId):
continue
if dst in mnemo.regs.all_regs_ids:
continue
cur_block.addto(asmbloc.asm_constraint(dst.name, C_TO))
if not line.splitflow():
block_to_nlink = None
delayslot = line.delayslot + 1
else:
raise RuntimeError("unknown class %s" % line.__class__)
i += 1
for block in blocks:
# Fix multiple constraints
block.fix_constraints()
# Log block
asmbloc.log_asmbloc.info(block)
return blocks, symbol_pool
def parse_txt(mnemo, attrib, txt, symbol_pool=None):
"""Parse an assembly listing. Returns a couple (blocks, symbol_pool), where
blocks is a list of asm_bloc and symbol_pool the associated asm_symbol_pool
@mnemo: architecture used
@attrib: architecture attribute
@txt: assembly listing
@symbol_pool: (optional) the asm_symbol_pool instance used to handle labels
of the listing
"""
if symbol_pool is None:
symbol_pool = asmbloc.asm_symbol_pool()
C_NEXT = asmbloc.asm_constraint.c_next
C_TO = asmbloc.asm_constraint.c_to
lines = []
# parse each line
for line in txt.split('\n'):
# empty
if EMPTY_RE.match(line):
continue
# comment
if COMMENT_RE.match(line):
continue
# labels to forget
if FORGET_LABEL_RE.match(line):
continue
# label beginning with .L
match_re = LABEL_RE.match(line)
if match_re:
label_name = match_re.group(1)
label = symbol_pool.getby_name_create(label_name)
lines.append(label)
continue
# directive
if DIRECTIVE_START_RE.match(line):
match_re = DIRECTIVE_RE.match(line)
directive = match_re.group(1)
if directive in ['text', 'data', 'bss']:
continue
if directive in ['string', 'ascii']:
# XXX HACK
line = line.replace(r'\n', '\n').replace(r'\r', '\r')
raw = line[line.find(r'"') + 1:line.rfind(r'"')]
raw = raw.decode('string_escape')
if directive == 'string':
raw += "\x00"
lines.append(asmbloc.asm_raw(raw))
continue
if directive == 'ustring':
# XXX HACK
line = line.replace(r'\n', '\n').replace(r'\r', '\r')
raw = line[line.find(r'"') + 1:line.rfind(r'"')] + "\x00"
raw = raw.decode('string_escape')
raw = "".join([string + '\x00' for string in raw])
lines.append(asmbloc.asm_raw(raw))
continue
if directive in declarator:
data_raw = line[match_re.end():].split(' ', 1)[1]
data_raw = data_raw.split(',')
size = declarator[directive]
expr_list = []
# parser
base_expr = gen_base_expr()[2]
my_var_parser = parse_ast(lambda x: m2_expr.ExprId(x, size),
lambda x:
m2_expr.ExprInt(x, size))
base_expr.setParseAction(my_var_parser)
for element in data_raw:
element = element.strip()
element_expr = base_expr.parseString(element)[0]
expr_list.append(element_expr.canonize())
raw_data = asmbloc.asm_raw(expr_list)
raw_data.element_size = size
lines.append(raw_data)
continue
if directive == 'comm':
# TODO
continue
if directive == 'split': # custom command
lines.append(DirectiveSplit())
continue
if directive == 'dontsplit': # custom command
lines.append(DirectiveDontSplit())
continue
if directive == "align":
align_value = int(line[match_re.end():], 0)
lines.append(DirectiveAlign(align_value))
continue
if directive in ['file', 'intel_syntax', 'globl', 'local',
'type', 'size', 'align', 'ident', 'section']:
continue
if directive[0:4] == 'cfi_':
continue
raise ValueError("unknown directive %s" % str(directive))
# label
match_re = LABEL_RE.match(line)
if match_re:
label_name = match_re.group(1)
label = symbol_pool.getby_name_create(label_name)
lines.append(label)
continue
# code
if ';' in line:
line = line[:line.find(';')]
line = line.strip(' ').strip('\t')
instr = mnemo.fromstring(line, attrib)
# replace orphan asm_label with labels from symbol_pool
replace_orphan_labels(instr, symbol_pool)
if instr.dstflow():
instr.dstflow2label(symbol_pool)
lines.append(instr)
asmbloc.log_asmbloc.info("___pre asm oki___")
# make blocks
cur_block = None
state = STATE_NO_BLOC
i = 0
blocks = asmbloc.AsmCFG()
block_to_nlink = None
delayslot = 0
while i < len(lines):
if delayslot:
delayslot -= 1
if delayslot == 0:
state = STATE_NO_BLOC
line = lines[i]
# no current block
if state == STATE_NO_BLOC:
if isinstance(line, DirectiveDontSplit):
block_to_nlink = cur_block
i += 1
continue
elif isinstance(line, DirectiveSplit):
block_to_nlink = None
i += 1
continue
elif not isinstance(line, asmbloc.asm_label):
# First line must be a label. If it's not the case, generate
# it.
label = guess_next_new_label(symbol_pool)
cur_block = asmbloc.asm_bloc(label, alignment=mnemo.alignment)
else:
cur_block = asmbloc.asm_bloc(line, alignment=mnemo.alignment)
i += 1
# Generate the current bloc
blocks.add_node(cur_block)
state = STATE_IN_BLOC
if block_to_nlink:
block_to_nlink.addto(
asmbloc.asm_constraint(cur_block.label,
C_NEXT))
block_to_nlink = None
continue
# in block
elif state == STATE_IN_BLOC:
if isinstance(line, DirectiveSplit):
state = STATE_NO_BLOC
block_to_nlink = None
elif isinstance(line, DirectiveDontSplit):
state = STATE_NO_BLOC
block_to_nlink = cur_block
elif isinstance(line, DirectiveAlign):
cur_block.alignment = line.alignment
elif isinstance(line, asmbloc.asm_raw):
cur_block.addline(line)
block_to_nlink = cur_block
elif isinstance(line, asmbloc.asm_label):
if block_to_nlink:
cur_block.addto(
asmbloc.asm_constraint(line, C_NEXT))
block_to_nlink = None
state = STATE_NO_BLOC
continue
# instruction
elif isinstance(line, instruction):
cur_block.addline(line)
block_to_nlink = cur_block
if not line.breakflow():
i += 1
continue
if delayslot:
raise RuntimeError("Cannot have breakflow in delayslot")
if line.dstflow():
for dst in line.getdstflow(symbol_pool):
if not isinstance(dst, m2_expr.ExprId):
continue
if dst in mnemo.regs.all_regs_ids:
continue
cur_block.addto(asmbloc.asm_constraint(dst.name, C_TO))
if not line.splitflow():
block_to_nlink = None
delayslot = line.delayslot + 1
else:
raise RuntimeError("unknown class %s" % line.__class__)
i += 1
for block in blocks:
# Fix multiple constraints
block.fix_constraints()
# Log block
asmbloc.log_asmbloc.info(block)
return blocks, symbol_pool
def parse_txt(mnemo, attrib, txt, symbol_pool=None, gen_label_index=0):
if symbol_pool is None:
symbol_pool = asmbloc.asm_symbol_pool()
lines_text = []
lines_data = []
lines_bss = []
C_NEXT = asmbloc.asm_constraint.c_next
C_TO = asmbloc.asm_constraint.c_to
lines = lines_text
# parse each line
for line in txt.split('\n'):
# empty
if re.match(r'\s*$', line):
continue
# comment
if re.match(r'\s*;\S*', line):
continue
# labels to forget
r = re.match(r'\s*\.LF[BE]\d\s*:', line)
if r:
continue
# label beginning with .L
r = re.match(r'\s*(\.L\S+)\s*:', line)
if r:
l = r.groups()[0]
l = symbol_pool.getby_name_create(l)
lines.append(l)
continue
# directive
if re.match(r'\s*\.', line):
r = re.match(r'\s*\.(\S+)', line)
directive = r.groups()[0]
if directive == 'text':
lines = lines_text
continue
if directive == 'data':
lines = lines_data
continue
if directive == 'bss':
lines = lines_bss
continue
if directive in ['string', 'ascii']:
# XXX HACK
line = line.replace(r'\n', '\n').replace(r'\r', '\r')
raw = line[line.find(r'"') + 1:line.rfind(r"'")]
raw = raw.decode('string_escape')
if directive == 'string':
raw += "\x00"
lines.append(asmbloc.asm_raw(raw))
continue
if directive == 'ustring':
# XXX HACK
line = line.replace(r'\n', '\n').replace(r'\r', '\r')
raw = line[line.find(r'"') + 1:line.rfind(r"'")] + "\x00"
raw = raw.decode('string_escape')
raw = "".join([string + '\x00' for string in raw])
lines.append(asmbloc.asm_raw(raw))
continue
if directive in declarator:
data_raw = line[r.end():].split(' ', 1)[1]
data_raw = data_raw.split(',')
size = declarator[directive]
data_int = []
# parser
base_expr = gen_base_expr()[2]
my_var_parser = parse_ast(
lambda x: m2_expr.ExprId(x, size),
lambda x: m2_expr.ExprInt_fromsize(size, x))
base_expr.setParseAction(my_var_parser)
for b in data_raw:
b = b.strip()
x = base_expr.parseString(b)[0]
data_int.append(x.canonize())
raw = data_int
x = asmbloc.asm_raw(raw)
x.element_size = size
lines.append(x)
continue
if directive == 'comm':
# TODO
continue
if directive == 'split': # custom command
x = asmbloc.asm_raw()
x.split = True
lines.append(x)
continue
if directive == 'dontsplit': # custom command
lines.append(asmbloc.asm_raw())
continue
if directive == "align":
align_value = int(line[r.end():])
lines.append(DirectiveAlign(align_value))
continue
if directive in [
'file', 'intel_syntax', 'globl', 'local', 'type', 'size',
'align', 'ident', 'section'
]:
continue
if directive[0:4] == 'cfi_':
continue
raise ValueError("unknown directive %s" % str(directive))
# label
r = re.match(r'\s*(\S+)\s*:', line)
if r:
l = r.groups()[0]
l = symbol_pool.getby_name_create(l)
lines.append(l)
continue
# code
if ';' in line:
line = line[:line.find(';')]
line = line.strip(' ').strip('\t')
instr = mnemo.fromstring(line, attrib)
# replace orphan asm_label with labels from symbol_pool
replace_orphan_labels(instr, symbol_pool)
if instr.dstflow():
instr.dstflow2label(symbol_pool)
lines.append(instr)
asmbloc.log_asmbloc.info("___pre asm oki___")
# make blocs
blocs_sections = []
bloc_num = 0
b = None
for lines in [lines_text, lines_data, lines_bss]:
state = 0
i = 0
blocs = []
blocs_sections.append(blocs)
bloc_to_nlink = None
block_may_link = False
while i < len(lines):
# no current bloc
if state == 0:
if not isinstance(lines[i], asmbloc.asm_label):
l = guess_next_new_label(symbol_pool)
lines[i:i] = [l]
else:
l = lines[i]
b = asmbloc.asm_bloc(l, alignment=mnemo.alignment)
b.bloc_num = bloc_num
bloc_num += 1
blocs.append(b)
state = 1
i += 1
if bloc_to_nlink:
bloc_to_nlink.addto(
asmbloc.asm_constraint(b.label, C_NEXT))
bloc_to_nlink = None
# in bloc
elif state == 1:
if isinstance(lines[i], asmbloc.asm_raw):
if hasattr(lines[i], 'split'):
state = 0
block_may_link = False
i += 1
else:
state = 1
block_may_link = True
b.addline(lines[i])
i += 1
elif isinstance(lines[i], DirectiveAlign):
b.alignment = lines[i].alignment
i += 1
# asmbloc.asm_label
elif isinstance(lines[i], asmbloc.asm_label):
if block_may_link:
b.addto(asmbloc.asm_constraint(lines[i], C_NEXT))
block_may_link = False
state = 0
# instruction
else:
b.addline(lines[i])
if lines[i].dstflow():
for x in lines[i].getdstflow(symbol_pool):
if not isinstance(x, m2_expr.ExprId):
continue
if x in mnemo.regs.all_regs_ids:
continue
b.addto(asmbloc.asm_constraint(x, C_TO))
# TODO XXX redo this really
if not lines[i].breakflow() and i + 1 < len(lines):
if isinstance(lines[i + 1], asmbloc.asm_label):
l = lines[i + 1]
else:
l = guess_next_new_label(symbol_pool)
lines[i + 1:i + 1] = [l]
else:
state = 0
if lines[i].splitflow():
bloc_to_nlink = b
if not lines[i].breakflow() or lines[i].splitflow():
block_may_link = True
else:
block_may_link = False
i += 1
for block in blocs_sections[0]:
asmbloc.log_asmbloc.info(block)
return blocs_sections, symbol_pool
