Subzero: implemented wrapper script to replace calls to calloc()

pydir/sz_driver.py

Index: pydir/sz_driver.py
diff --git a/pydir/sz_driver.py b/pydir/sz_driver.py
new file mode 100644
index 0000000000000000000000000000000000000000..7475b3bb070b56fa04fe744712281726552d2db0
--- /dev/null
+++ b/pydir/sz_driver.py
@@ -0,0 +1,88 @@
+import os
+import shutil
+import subprocess
+import sys
+import tempfile
+
+from utils import FindBaseNaCl, shellcmd
+
+def subsToMacros(subs, src):
+    macros = (

[Jim Stichnoth, 2016/07/14 20:42:11]

https://www.chromium.org/chromium-os/python-style-guidelines#TOC-Official-Sty...

says use an indent of 2 spaces, not 4.

Subzero has some "legacy" scripts that use 4 spaces, but for new files, it would
be nice to use 2.

[tlively, 2016/07/15 01:54:25]

Decided against this change to maintain consistency within directory.

+        '#include <stddef.h>\n'

[Jim Stichnoth, 2016/07/14 20:42:11]

It would be nice if you could remove all the '\n' and then maybe add them back,
in one shot, before "return macros".

[tlively, 2016/07/15 01:54:25]

Done.

+        '#ifdef __cplusplus\n'
+        'extern "C" {\n'
+        '#endif\n')
+    for func in subs:
+        args = [('{atype} a{num}').format(atype=atype, num=i) for
+                i, atype in enumerate(subs[func]['sig'][1:])]
+        macros += (
+            '{ftype} {name}({args});\n'
+            ).format(ftype=subs[func]['sig'][0],
+                     name=subs[func]['sub'],
+                     args=', '.join(args))
+        macros += (
+            '#define {func}(args...) ({sub}(args))\n'
+            ).format(func=func, sub=subs[func]['sub'])
+    macros += (
+        '#ifdef __cplusplus\n'
+        '}} // extern "C"\n'
+        '#endif\n'
+        '#line 1 "{src}"\n').format(src=src)
+    return macros
+
+def run(is_cpp):
+    """Passes its arguments directly to pnacl-clang.
+
+    If -fsanitize-address is specified, extra information is passed to
+    pnacl-clang to ensure that later instrumentation in pnacl-sz can be
+    performed. For example, clang automatically inlines many memory allocation
+    functions, so this script will redefine them at compile time to make sure
+    they can be correctly instrumented by pnacl-sz.
+    """
+    pnacl_root = FindBaseNaCl()
+    dummy_subs = {'calloc': {'sig': ['void *', 'size_t', 'size_t'],
+                             'sub': '__asan_dummy_calloc'},
+                  '_calloc': {'sig': ['void *', 'size_t', 'size_t'],
+                              'sub': '__asan_dummy_calloc'}}
+    subs_src = (
+        '{root}/toolchain_build/src/subzero/pydir/sz_clang_dummies.c'
+        ).format(root=pnacl_root)
+    clang = (
+        '{root}/toolchain/linux_x86/pnacl_newlib_raw/bin/pnacl-clang{pp}'
+        ).format(root=pnacl_root, pp='++' if is_cpp else '')
+    args = sys.argv
+    args[0] = clang
+    tmp_dir = ''
+    if '-fsanitize-address' in args:
+        args.remove('-fsanitize-address')
+        include_dirs = set()
+        tmp_dir = tempfile.mkdtemp()
+        for i, arg in enumerate(args[1:], 1):
+            if not os.path.isfile(arg):
+                continue
+            src = os.path.basename(arg)
+            ext = os.path.splitext(arg)[1]
+            if ext in ['.c', '.cc', '.cpp']:
+                include_dirs |= {os.path.dirname(arg)}
+                dest_name = os.path.join(tmp_dir, src)
+                with open(dest_name, 'w') as dest:
+                    dest.write(subsToMacros(dummy_subs, arg))
+                    with open(arg) as src:
+                        for line in src:
+                            dest.write(line)
+                args[i] = dest_name
+        # If linking (not single file compilation) then add dummy definitions
+        if not ('-o' in args and
+                ('-c' in args or '-S' in args or '-E' in args)):
+            args.append(subs_src)
+        for d in include_dirs:
+            args.append('-iquote {d}'.format(d=d))
+    err_code = 0
+    try:
+        shellcmd(args, echo=True)
+    except subprocess.CalledProcessError as e:
+        print e.output
+        err_code = e.returncode
+    if tmp_dir != '':
+        shutil.rmtree(tmp_dir)
+    exit(err_code)