mpir/build.vc10/g2y.py
gladman de924e3d62 1. Add Jason's fantastic new mul assembler to Windows
2. remove bug from Windows core2 popcount assemmbler
3. remove bug from Python AT&T to Intel converter
2010-12-17 14:03:48 +00:00

576 lines
16 KiB
Python

'''
---------------------------------------------------------------------------
Copyright (c) 1998-2008, Brian Gladman, Worcester, UK. All rights reserved.
LICENSE TERMS
The redistribution and use of this software (with or without changes)
is allowed without the payment of fees or royalties provided that:
1. source code distributions include the above copyright notice, this
list of conditions and the following disclaimer;
2. binary distributions include the above copyright notice, this list
of conditions and the following disclaimer in their documentation;
3. the name of the copyright holder is not used to endorse products
built using this software without specific written permission.
DISCLAIMER
This software is provided 'as is' with no explicit or implied warranties
in respect of its properties, including, but not limited to, correctness
and/or fitness for purpose.
---------------------------------------------------------------------------
'''
from __future__ import print_function
import string, re, os, sys, shutil
debug = False
dot_labels = True
remove_comment_lines = True
is_linux = False
# translate 64 bit registers to 8 bit form
r08 = { 'rax' : 'al', 'rbx' : 'bl', 'rcx' : 'cl', 'rdx' : 'dl',
'rsi' : 'sil', 'rdi' : 'dil', 'rbp' : 'bpl',
'r8' : 'r8b', 'r9' : 'r9b', 'r10' : 'r10b', 'r11' : 'r11b',
'r12' : 'r12b', 'r13' : 'r13b', 'r14' : 'r14b', 'r15' : 'r15b' }
# translate 64 bit registers to 32 bit form
r32 = { 'rax' : 'eax', 'rbx' : 'ebx', 'rcx' : 'ecx', 'rdx' : 'edx',
'rsi' : 'esi', 'rdi' : 'edi', 'rbp' : 'ebp',
'r8' : 'r8d', 'r9' : 'r9d', 'r10' : 'r10d', 'r11' : 'r11d',
'r12' : 'r12d', 'r13' : 'r13d', 'r14' : 'r14d', 'r15' : 'r15d' }
# regular expression for registers
r_q = r'(?:r[abcd]x|r[sd]i|r[bsi]p)|(?:r8|r9|r1[0-5])|' # 64 bit
r_d = r'(?:e[abcd]x|e[sd]i|e[bsi]p|r[89]d|r1[0-5]d)|' # 32 bit
r_w = r'(?:[abcd]x|[sd]i|[bsi]p|r[89]w|r1[0-5]w)|' # 16 bit
r_b = r'(?:[abcd]l|[ds]il|[bsi]pl|r[89]b|r1[0-5]b)|' # 8 bit
r_x = r'(?:x?mm1[0-5]|x?mm\d)|(?:mmx1[0-5]|mmx\d)|(?:st\([0-7]\))'
p_r1 = r'(?:\s*%(?P<reg1>' + r_b + r_w + r_d + r_q + r_x + r'))'
p_r2 = r'(?:\s*%(?P<reg2>' + r_b + r_w + r_d + r_q + r_x + r'))'
p_r3 = r'(?:\s*%(?P<reg3>' + r_b + r_w + r_d + r_q + r_x + r'))'
# regular expression for immediate (numeric, not symbolic)
p_im = r'\s+\`?\$\'?(?P<imm>[\+\-]?[0-9]+|0x[a-zA-Z0-9]+)'
# regular expressions for labels
p_l1 = r'(?P<lab1>[a-zA-Z$_][a-zA-Z0-9\$_]*|\.[0-9]+)'
p_l2 = r'L\((?P<lab2>[a-zA-Z0-9\$_]+)\)'
p_l3 = r'(?P<lab3>[a-zA-Z$_][a-zA-Z0-9\$_]*|\.[0-9]+)'
p_l4 = r'L\((?P<lab4>[a-zA-Z0-9\$_]+)\)'
p_lr = r'\s*(?:' + p_l2 + r'|' + p_l1 + r')'
p_la = r'\s*(?:' + p_l2 + r'|' + p_l1 + r')\s*(:)'
p_ri = p_lr + r'\(\%rip\)'
p_34 = r'\s*(?:' + p_l4 + r'|' + p_l3 + r')'
p_jt = r'\s+\.long\s+' + p_lr + r'\s*\-\s*' + p_34
# regular expression for dis(r1, r2, mul) forms
p_di = r'(?:(?:\s*(?P<dis>[-+]?[0-9\+\-]*)(?=\())?' + '|' + p_lr + ')' # numeric displacement
p_mu = r'(?:\s*,\s*(?P<mul>[1248]))?\s*(?=\))' # multiplier (1, 2, 4, 8)
p_t1 = p_di + r'\s*\(' + p_r1 + r'(?:\s*\)|\s*,' + p_r2 + p_mu + r'\s*\))'
# regular expression for instructions
p_in = r'\s*(?P<ins>[a-zA-Z][a-zA-Z0-9]*)'
p_def = r'define\s*\(\s*\`([a-zA-Z_][a-zA-Z0-9_]*)\'\s*\,\s*\`' + p_r1 + '\'\s*\)'
m_f1 = re.compile(p_in + p_r3 + r'\s*,' + p_t1)
m_f2 = re.compile(p_in + p_t1 + r'\s*,' + p_r3)
m_f3 = re.compile(p_in + p_t1)
m_f4 = re.compile(p_in + p_im + r'\s*,' + p_t1)
m_f5 = re.compile(p_in + p_im + r'\s*,' + p_r1)
m_f6 = re.compile(p_in + p_r1 + r'\s*,' + p_r2)
m_f7 = re.compile(p_in + p_r1)
m_g7 = re.compile(p_in + r'\s+\*' + p_r1)
m_f8 = re.compile(p_in + p_im)
m_f9 = re.compile(p_in + p_lr)
m_fa = re.compile(p_in + '(?:' + p_im + '|' + p_r1 + r')\s*,' + p_r2 + r'\s*,' + p_r3)
m_la = re.compile(p_la)
m_jt = re.compile(p_jt)
m_ri = re.compile(p_in + p_ri)
r_mac = re.compile(r"^\s*(?:define|DEFINE)\s*\(\s*`"
"([A-Z$_][A-Z0-9$_]*)'\s*,\s*`\s*$")
r_mnd = re.compile(r"^\s*'\s*\)\s*$")
r_dlr = re.compile(r'\$([0-9]+)')
r_def = re.compile(p_def)
r1 = r'\s*(\%{0,1}[0-9]+){0,1}\s*'
r2 = r'(?:,\s*(\%{0,1}[0-9]+))?\s*'
r_mrz = re.compile(r'([A-Z$_][A-Z0-9$_]*)\s*$')
r_mrf = re.compile(r'([A-Z$_][A-Z0-9$_]*)\s*\(' + r1 + r2 +
r2 + r2 + r2 + r2 + r2 + r2 + r2 + r2 + r'\)')
def lab_ref(v, labels, macros, mac_name):
if mac_name and v in macros[mac_name][3]:
return '%%{0}'.format(macros[mac_name][3][v])
elif dot_labels:
return '.{0}'.format(labels.index(v))
else:
return 'L_{0}'.format(re.sub('\$', '%', v))
def pass_one(code):
labels = []
for l in code:
m = re.match(r'(\s*)(?:C|#|;|dnl)\s*(.*)', l)
if m:
v = list(m.groups())
continue
m = m_la.match(l)
if m:
lb = m.group(1) if m.group(1) else m.group(2)
if lb:
labels += [lb]
return labels
def pass_two(code, labels):
mac_name = ''
mac_dict = {}
lab_idx = 0
for ln in range(len(code)):
l = code[ln]
m = r_mac.search(l)
if m and not mac_name:
mac_name = m.group(1)
mac_dict[mac_name] = [ln, 0, 0, {}]
continue
m = r_mnd.search(l)
if m and mac_name:
mac_dict[mac_name][1] = ln
mac_name = ''
lab_idx = 0
continue
if mac_name:
m = r_dlr.findall(l)
if m:
mn = max(int(i) for i in m)
if mn > mac_dict[mac_name][2]:
mac_dict[mac_name][2] = mn
m = m_la.match(l)
if m:
lb = m.group(1) if m.group(1) else m.group(2)
if lb in labels:
lab_idx += 1
mac_dict[mac_name][3][lb] = lab_idx
return mac_dict
def addr(dd, labels, macros, mac_name):
pfx = '['
ss = ']'
if dd['lab1'] or dd['lab2']:
lb = dd['lab1'] if dd['lab1'] else dd['lab2']
if lb in labels:
ss = lab_ref(lb, labels, macros, mac_name) + ']'
pfx = '[rel '
if dd['reg1'] == 'rip':
return pfx + ss
ss = '+' + ss
elif dd['dis']:
ss = (dd['dis'] if not dd['dis'] or dd['dis'][0] in '+-'
else '+' + dd['dis']) + ']'
ss = ('' if not dd['mul'] else '*' + dd['mul']) + ss
if dd['reg2']:
ss = '+' + dd['reg2'] + ss
if dd['reg1']:
ss = dd['reg1'] + ss
return pfx + ss
def pass_three(code, labels, macros, level):
lo = []
mac_name = ''
for l in code:
m = re.match(r'(\s*)(?:C|#|;|dnl)\s*(.*)', l)
if m:
v = list(m.groups())
if not remove_comment_lines:
lo += ['{0[0]}; {0[1]}'.format(v)]
continue
m = re.search(r'(.*)\s+#(.*)$', l)
sfx = ''
if m:
if m.group(1):
l = m.group(1) + ' '
else:
continue
if m.group(2):
sfx = '\t;' + m.group(2)
# labels
lp = ''
m = m_la.match(l)
if m:
lb = m.group(1) if m.group(1) else m.group(2)
l = l[ l.find(':') + 1 : -1 ]
if mac_name:
if lb in macros[mac_name][3]:
ii = macros[mac_name][3][lb]
else:
print('internal error')
else:
ii = labels.index(lb)
if lb:
lb = re.sub('\$', '%', lb)
if not l:
if mac_name:
lo += ['\n%%{0}:'.format(ii)]
elif dot_labels:
lo += ['\n.{0}:'.format(ii)]
else:
lo += ['\nL_{0}:'.format(lb)]
continue
else:
if mac_name:
lp = '\n%%{0}:'.format(ii)
elif dot_labels:
lp = '\n.{0}:'.format(ii)
else:
lp = '\nL_{0}:'.format(lb)
if not re.search(r'(\S+)', l):
if lp:
lo += [lp]
continue
elif not lp:
lp = '\n'
# three operand instructions
m = m_fa.search(l)
if m:
d = m.groupdict()
if debug:
print(l, end = '')
e = d['imm'] if d['imm'] else d['reg1']
lo += [lp + '\t{0[ins]:7s} {0[reg3]}, {0[reg2]}, {1}'.format(d, e)]
continue
# ins reg, dis(reg, reg, off)
m = m_f1.search(l)
if m:
if debug:
print(l, end = '')
d = m.groupdict()
s = addr(d, labels, macros, mac_name)
lo += [lp + '\t{0[ins]:7s} {1}, {0[reg3]}'.format(d, s)]
continue
# ins dis(reg, reg, off), reg
m = m_f2.search(l)
if m:
if debug:
print(l, end = '')
d = m.groupdict()
s = addr(d, labels, macros, mac_name)
lo += [lp + '\t{0[ins]:7s} {0[reg3]}, {1}'.format(d, s)]
continue
# ins dis(reg, reg, off)
m = m_f3.search(l)
if m:
if debug:
print(l, end = '')
d = m.groupdict()
s = addr(d, labels, macros, mac_name)
lo += [lp + '\t{0[ins]:7s} {1}'.format(d, s)]
continue
# ins imm, dis(reg, reg, off)
m = m_f4.search(l)
if m:
if debug:
print(l, end = '')
d = m.groupdict()
s = addr(d, labels, macros, mac_name)
lo += [lp + '\t{0[ins]:7s} {1}, {0[imm]}'.format(d, s)]
continue
# ins imm, reg
m = m_f5.search(l)
if m:
d = m.groupdict()
if debug:
print(l, end = '')
lo += [lp + '\t{0[ins]:7s} {0[reg1]}, {0[imm]}'.format(d)]
continue
# ins reg, reg
m = m_f6.search(l)
if m:
d = m.groupdict()
if debug:
print(l, end = '')
lo += [lp + '\t{0[ins]:7s} {0[reg2]}, {0[reg1]}'.format(d)]
continue
# ins reg
m = m_f7.search(l)
if m:
d = m.groupdict()
if debug:
print(l, end = '')
lo += [lp + '\t{0[ins]:7s} {0[reg1]}'.format(d)]
continue
# ins *reg
m = m_g7.search(l)
if m:
d = m.groupdict()
if debug:
print(l, end = '')
lo += [lp + '\t{0[ins]:7s} {0[reg1]}'.format(d)]
continue
# ins imm
m = m_f8.search(l)
if m:
d = m.groupdict()
if debug:
print(l, end = '')
lo += [lp + '\t{0[ins]:7s} {0[imm]}'.format(d)]
continue
# jump table
m = m_jt.search(l)
if m:
d = m.groupdict()
lb0 = d['lab1'] if d['lab1'] else d['lab2']
lb1 = d['lab3'] if d['lab3'] else d['lab4']
if lb0 in labels and lb1 in labels:
if debug:
print(l, end = '')
st1 = lab_ref(lb0, labels, macros, mac_name)
st2 = lab_ref(lb1, labels, macros, mac_name)
lo += [lp + '\tdd ' + st1 + ' - ' + st2]
continue
# jump label
m = m_f9.search(l)
if m:
d = m.groupdict()
lb = d['lab1'] if d['lab1'] else d['lab2']
if lb and lb in labels:
if debug:
print(l, end = '')
ss = lab_ref(lb, labels, macros, mac_name)
lo += [lp + '\t{0[ins]:7s} {1}'.format(d, ss)]
continue
m = re.search(r'\s*\.byte\s+((?:0x|0X)[0-9a-fA-F]+|[0-9]+)\s*', l)
if m:
v = list(m.groups())
lo += [lp + '\tdb {0[0]}'.format(v)]
continue
# macro definitions
m = r_mac.search(l)
if m:
mac_name = m.group(1)
if mac_name in macros:
lo += [lp + '%macro ' + mac_name.lower() + ' '
+ str(macros[mac_name][2])]
continue
m = r_mnd.search(l)
if m and mac_name:
mac_name = ''
lab_ofs = 0
lo += [lp + '%endmacro']
continue
m = r_dlr.findall(l)
if m:
l = re.sub(r'\$([0-9]+)', r'%\1', l)
m = re.search('PROLOGUE\(([a-zA-Z$_][a-zA-Z0-9$_]*)\)', l)
if m:
if is_linux:
lo += [lp + '\tGLOBAL_FUNC {0}'.format(m.group(1))]
else:
lo += [lp + '\tWIN64_GCC_PROC {0}'.format(m.group(1))]
continue
m = re.search('EPILOGUE\(\)', l)
if m:
if is_linux:
lo += [lp + '\tend']
else:
lo += [lp + '\tWIN64_GCC_END']
continue
# macro calls
m = r_mrz.search(l)
if not m:
m = r_mrf.search(l)
if m:
if m.group(1).lower() == 'align':
fs = '\talign {0}' if is_linux else '\txalign {0}'
lo += [lp + fs.format(m.group(2))]
continue
elif m.group(1) in macros:
lp += '\t{0:7s}'.format(m.group(1).lower())
i = 2
while i <= m.lastindex and m.group(i):
lp += ' {0}'.format(m.group(i))
i += 1
lo += [lp]
continue
elif m.group(1) and (m.lastindex == 1 or not m.group(2)):
lo += [lp + '\t{0}'.format(m.group(1))]
continue
if mac_name:
m = re.search(r'\s*([^%]+)%([0-9]+)\s*', l)
if m and m.lastindex == 2 and int(m.group(2)) <= macros[mac_name][2]:
lo += [lp + '\t{0}%{1}'.format(m.group(1).lower(),m.group(2))]
continue
# ins
m = re.search(p_in + r'\s+(.*)', l)
if m:
v = list(m.groups())
if debug:
print(l, end = '')
if len(v) == 1 or len(v) == 2 and v[1] == '':
if v[0] in macros:
lo += [lp + '\t{0}'.format(v[0].lower())]
else:
lo += [lp + '\t{0[0]}'.format(v)]
continue
elif v[0] == 'C':
lo += [lp + ';\t{0}'.format(v[1])]
continue
m = re.search(r'include\(.+config.m4.+\)', l)
if m:
if is_linux:
lo += [lp + "%include 'yasm_mac.inc'"]
else:
lo += [lp + "%include '"
+ ''.join(['..\\'] * level) + "yasm_mac.inc'"]
continue
m = re.search(r'\s*(ret\b.*)', l)
if m:
lo += [lp + '{0}'.format(l.rstrip(string.whitespace))]
continue
m = re.search(r'\s*(\S+)', l)
if m:
if len(l):
lo += [lp + '{0} ; < not translated >'
.format(l.rstrip(string.whitespace))]
else:
lo += [lp]
elif lp:
lo += [lp]
return lo + ['\n']
def form_path(p):
n = str.rfind(p, '\\', 0, -1)
if n != -1:
if not os.path.exists(p[ : n + 1]):
form_path(p[ : n + 1])
if p[-1] == '\\' and not os.path.exists(p):
os.mkdir(p)
def conv_lines(code, l):
labels = pass_one(code)
macros = pass_two(code, labels)
code = pass_three(code, labels, macros, l)
return (labels, macros, code)
def conv_file(f_in, f_out, l):
f = open(f_in, 'r')
code = f.readlines()
f.close()
labels, macros, code = conv_lines(code, l)
f = open(f_out, 'w')
f.writelines(code)
f.close()
def conv_dirs(s, d, l):
fd = os.listdir(s)
for f in fd:
sp = os.path.join(s,f)
dp = os.path.join(d,f)
if os.path.isdir(sp):
conv_dirs(sp, dp, l + 1)
elif os.path.isfile(sp):
if sp != dp and os.path.exists(dp):
continue
fn = os.path.basename(f)
x = os.path.splitext(fn)
if x[1] == '.asm':
form_path(dp)
print("translating '{0}'".format(sp))
f = open(sp, 'r')
code = f.readlines()
f.close()
if sp == dp:
rp = os.path.join(s, x[0] + ('.as' if is_linux else '.old'))
os.rename(sp, rp)
labels, macros, code = conv_lines(code, l)
f = open(dp, 'w')
f.writelines(code)
f.close()
elif False:
form_path(dp)
shutil.copyfile(sp, dp)
if len(sys.argv) == 1:
cd = os.getcwd() # if run in the build.vc10 directory
if cd.endswith('build.vc10'):
cd1 = cd + '\\..\\mpn\\x86_64' # the GAS assembler directory
cd2 = cd + '\\..\\mpn\\x86_64w' # the YASM (Windows) assembler directory
elif cd.endswith('x86_64'): # if run in the GAS assembler directory
cd1 = cd
cd2 = cd + '\\..\\x86_64w'
elif cd.endswith('x86_64w'): # if run in the YASM assembler directory
cd2 = cd
cd1 = cd + '\\..\\x86_64'
else:
print("cannot locate assembler source directory")
sys.exit(-1)
elif len(sys.argv) == 3:
cd1 = sys.argv[1]
cd2 = sys.argv[2]
else:
print("invalid program location or input")
sys.exit(-1)
if cd1 and os.path.exists(cd1):
if os.path.isdir(cd1):
if os.path.exists(cd2) and os.path.isdir(cd2):
print("warning: output directory '{0}' already exists".format(cd2))
if cd1 != cd2:
conv_dirs(cd1, cd2, 0) # convert format from GAS to YASM
elif os.path.isfile(cd1):
if not os.path.exists(cd2):
conv_file(cd1, cd2, 0)
else:
print('{0} is not a file or directory'.format(cd1))