[HWASan] symbolize: use buildid index for locals.

[kstoimenov]

compiler-rt/lib/hwasan/scripts/hwasan_symbolize

261–263

Consider splitting this into multiple lines?

[fmayer]

oformat

[fmayer]

format

[fmayer]

format2

[eugenis]

LGTM

compiler-rt/lib/hwasan/scripts/hwasan_symbolize

260–261

Please update the comment.

[fmayer]

comment

compiler-rt/lib/hwasan/scripts/hwasan_symbolize

class Symbolizer:
   def __process_source_path(self, file_name):
     for path_to_cut in self.__paths_to_cut:
       file_name = re.sub(".*" + path_to_cut, "", file_name)
     file_name = re.sub(".*hwasan_[a-z_]*.(cc|h):[0-9]*", "[hwasan_rtl]", file_name)
     file_name = re.sub(".*asan_[a-z_]*.(cc|h):[0-9]*", "[asan_rtl]", file_name)
     file_name = re.sub(".*crtstuff.c:0", "???:0", file_name)
     return file_name
 
-  def __process_binary_name(self, name, buildid=None):
+  def __process_binary_name(self, name, buildid):
     if name.startswith('/'):
       name = name[1:]
     if buildid is not None and buildid in self.__index:
       return self.__index[buildid]
 
     for p in self.__binary_prefixes:
       full_path = os.path.join(p, name)
       if os.path.exists(full_path):
         return full_path
       apex_prefix = "apex/com.android."
       if name.startswith(apex_prefix):
         full_path = os.path.join(p, "apex/com.google.android." + name[len(apex_prefix):])
         if os.path.exists(full_path):
           return full_path
     # Try stripping extra path components as the last resort.
     for p in self.__binary_prefixes:
       full_path = os.path.join(p, os.path.basename(name))
       if os.path.exists(full_path):
         return full_path
     if name not in self.__warnings:
       print("Could not find symbols for", name, file=sys.stderr)
       self.__warnings.add(name)
     return None
 
-  def iter_locals(self, binary, addr):
+  def iter_locals(self, binary, addr, buildid):
     self.__open_pipe()
     p = self.__pipe
-    binary = self.__process_binary_name(binary)
+    binary = self.__process_binary_name(binary, buildid)
     if not binary:
       return
     self.__write("FRAME %s %s" % (binary, addr))
     try:
       while True:
         function_name = self.__read()
         local_name = self.__read()
         file_line = self.__read()
for p in self.__binary_prefixes:
       for dname, _, fnames in os.walk(p):
         for fn in fnames:
           filename = os.path.join(dname, fn)
           bid = get_buildid(filename)
           if bid is not None:
             self.__index[bid] = filename
 
 def symbolize_line(line, symbolizer_path):
-  #0 0x7f6e35cf2e45  (/blah/foo.so+0x11fe45)
-  match = re.match(r'^(.*?)#([0-9]+)( *)(0x[0-9a-f]*) *\((.*)\+(0x[0-9a-f]+)\)(?:\s*\(BuildId: ([0-9a-f]+)\))?', line, re.UNICODE)
+  #0 0x7f6e35cf2e45  (/blah/foo.so+0x11fe45) (BuildId: 4abce4cd41ea5c2f34753297b7e774d9)
+  match = re.match(r'^(.*?)#([0-9]+)( *)(0x[0-9a-f]*) *\((.*)\+(0x[0-9a-f]+)\)'
+                   r'(?:\s*\(BuildId: ([0-9a-f]+)\))?', line, re.UNICODE)
   if match:
     frameno = match.group(2)
     binary = match.group(5)
     addr = int(match.group(6), 16)
     buildid = match.group(7)
 
     frames = list(symbolizer.iter_call_stack(binary, buildid, addr))
 
if match:
     last_access_tag = int(match.group(2), 16)
 
 def process_stack_history(line, symbolizer, ignore_tags=False):
   if last_access_address is None or last_access_tag is None:
     return
   if re.match(r'Previously allocated frames:', line, re.UNICODE):
     return True
   pc_mask = (1 << 48) - 1
   fp_mask = (1 << 20) - 1
-  # record_addr:0x1234ABCD record:0x1234ABCD (/path/to/binary+0x1234ABCD)
-  match = re.match(r'^(.*?)record_addr:(0x[0-9a-f]+) +record:(0x[0-9a-f]+) +\((.*)\+(0x[0-9a-f]+)\)', line, re.UNICODE)
+  # record_addr:0x1234ABCD record:0x1234ABCD (/path/to/binary+0x1234ABCD) (BuildId: 4abce4cd41ea5c2f34753297b7e774d9)
+  match = re.match(r'^(.*?)record_addr:(0x[0-9a-f]+) +record:(0x[0-9a-f]+) +\((.*)\+(0x[0-9a-f]+)\)'
+                   r'(?:\s*\(BuildId: ([0-9a-f]+)\))?', line, re.UNICODE)
   if match:
     record_addr = int(match.group(2), 16)
     record = int(match.group(3), 16)
     binary = match.group(4)
     addr = int(match.group(5), 16)
+    buildid = match.group(6)
     base_tag = (record_addr >> 3) & 0xFF
     fp = (record >> 48) << 4
     pc = record & pc_mask
 
-    for local in symbolizer.iter_locals(binary, addr):
+    for local in symbolizer.iter_locals(binary, addr, buildid):
       frame_offset = local[3]
       size = local[4]
       if frame_offset is None or size is None:
         continue
       obj_offset = (last_access_address - fp - frame_offset) & fp_mask
       if obj_offset >= size:
         continue
       tag_offset = local[5]