saulteafarmer/XenonRecomp
1
0
Fork 0
mirror of https://github.com/hedge-dev/XenonRecomp.git synced 2025-06-23 16:05:30 +00:00
Code Issues Packages Projects Releases Wiki Activity

missing 3 functions, but otherwise correct

Browse Source
This commit is contained in:
JillianTo 2025-03-09 17:56:37 -04:00
parent 21f1a81aa3
commit 2365f4d697
1 changed files with 45 additions and 49 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

92
parser.py
View File

@ -75,11 +75,17 @@ in_text = False
# Mark if we should end parsing # Mark if we should end parsing
end_parse = False end_parse = False
# Initialize address of last bctr instruction to 0
bctr_addr = '00000000'
# Initialize address of last blr instruction to 0 # Initialize address of last blr instruction to 0
blr_addr = '00000000' blr_addr = '00000000'
# Initialize address of last bctr instruction to 0 # Initialize address of last 'End of function' comment to 0
bctr_addr = '00000000' eof_addr = '00000000'
# Initialize address of last restgprlr instruction to 0
restgprlr_addr = '00000000'
# Initialize address of last padding to 0 # Initialize address of last padding to 0
pad_addr = 0 pad_addr = 0
@ -103,8 +109,8 @@ with open(ida_html, 'r') as file:
# If this is not the first function being added # If this is not the first function being added
if num_functs > 0: if num_functs > 0:
# If last address had padding, then this function was already added # If last address had padding or restgprlr instruction, then this function was already added
if curr_addr_int-4 == pad_addr: if curr_addr_int-4 == pad_addr or curr_addr_int-4 == restgprlr_addr:
# Set function type for start address # Set function type for start address
functs[num_functs-1][3] = 'sub' functs[num_functs-1][3] = 'sub'
else: else:
@ -130,59 +136,31 @@ with open(ida_html, 'r') as file:
curr_funct = functs[num_functs-1] curr_funct = functs[num_functs-1]
# If previous address was a blr instruction # If previous address was a blr instruction
if curr_addr_int-4 == int(blr_addr, 16): if curr_addr_int-4 == int(blr_addr, 16):
# If last added function is a subroutine and has no nested functions # If previous address had an 'End of function' comment or if there was a bctr with the comment
if curr_funct[3] == 'sub' and not curr_funct[2]: if blr_addr == eof_addr or bctr_addr == eof_addr:
xref_idx = line.find('XREF: sub_') # Find a XREF pointing to a .text address
# If XREF is a subroutine xref_idx = line.find('XREF: .text:')
if xref_idx > -1: if xref_idx > -1:
xref = line[xref_idx+10:xref_idx+18] underscore_idx = line.find('_', xref_idx)
# If the XREF is equivalent to the last function's start address if underscore_idx > -1:
if int(xref, 16) == curr_funct[0]: xref = line[underscore_idx+1:underscore_idx+9]
# Store as nested function in latest function
functs[num_functs-1][2].append(xref)
# If not, add this address as a new function
else: else:
add_function(curr_addr_int, curr_addr_int, 'loc') xref = line[xref_idx+12:xref_idx+20]
# If not, add this address as new function
else: else:
xref = None
# Couldn't find XREF pointing to .text address or the XREF is after this address
if xref == None or int(xref, 16) > curr_addr_int:
# Add as new function
add_function(curr_addr_int, curr_addr_int, 'loc') add_function(curr_addr_int, curr_addr_int, 'loc')
# If last added function is not a subroutine or has nested functions:
else:
# Check for XREF to subroutine
xref_idx = line.find('XREF: sub_')
if xref_idx > -1:
xref = line[xref_idx+10:xref_idx+18]
# If not found, check for XREF to .text address
else:
xref_idx = line.find('XREF: .text:')
if xref_idx > -1:
underscore_idx = line.find('_', xref_idx)
# If referencing sub_, loc_, etc.
if underscore_idx > -1:
xref = line[underscore_idx+1:underscore_idx+9]
# Else, there's only the address after .text
else:
xref = line[xref_idx+12:xref_idx+20]
else:
xref = '-1'
# If XREF points to subroutine or .text address before current address
if int(xref, 16) < curr_addr_int:
# Store as nested function
functs[num_functs-1][2].append(xref)
# If not, add this address as new funciton
else:
add_function(curr_addr_int, curr_addr_int, 'loc')
# If not, store as nested function in latest function
else: else:
# Find address of function that references this # Find address of function that references this
xref_idx = line.find('XREF: sub_') xref_idx = line.find('CODE XREF: sub_')
# If it was found # If it was found
if xref_idx > -1: if xref_idx > -1:
# Store as nested function in latest function # Store as nested function in latest function
functs[num_functs-1][2].append(line[xref_idx+10:xref_idx+18]) functs[num_functs-1][2].append(line[xref_idx+15:xref_idx+23])
# Check if this line is padding # Check if this line is padding
elif num_functs > 0 and re.search('<span class="c[0-9]*">\.long </span><span class="c[0-9]*">0$', line): elif num_functs > 0 and re.search('<span class="c[0-9]*">\.long </span><span class="c[0-9]*">0$', line):
@ -196,9 +174,27 @@ with open(ida_html, 'r') as file:
pad_addr = curr_addr_int pad_addr = curr_addr_int
# Check for blr instruction # Check for blr instruction
elif re.search('<span class="c[0-9]*">blr', line): elif re.search('<span class="c[0-9]*">blr$', line):
blr_addr = curr_addr blr_addr = curr_addr
# Check for 'End of function' comment
elif re.search('End of function ', line):
eof_addr = curr_addr
# Check for bctr instruction
elif re.search('<span class="c[0-9]*">bctr$', line):
bctr_addr = curr_addr
# Check for restgprlr instruction
elif re.search('<span class="c[0-9]*">b </span><span class="c[0-9]*">__restgprlr_[0-9][0-9]$', line):
# Convert current address to integer
curr_addr_int = int(curr_addr, 16)
# Add a new function at the line after restgprlr instruction, and end the current function at this address
add_function(curr_addr_int+4, curr_addr_int, None)
restgprlr_addr = curr_addr_int
# If not in .text # If not in .text
else: else:
# If .text section header found # If .text section header found