[PATCH] D72011: [sanitizers][windows] Global/LocalAlloc interception and tests

[Matthew G McGovern via Phabricator via llvm-commits]

mcgov abandoned this revision.
mcgov marked an inline comment as done.
mcgov added a subscriber: iakronqu.
mcgov added inline comments.


================
Comment at: compiler-rt/lib/asan/asan_malloc_win.cpp:95
 ALLOCATION_FUNCTION_ATTRIBUTE
-size_t _msize_base(void *ptr) {
-  return _msize(ptr);
-}
+size_t _msize_base(void *ptr) { return _msize(ptr); }
 
----------------
rnk wrote:
> Most of the changes in this file are from clang-format. I don't mind a few, but this is too many. Can you revert the non-functional whitespace changes far from the code you are editing? You can use `git checkout -p` to speed this up.
> 
> I also encourage you to run run `check-sanitizer` on Linux. The ASan codebase has its own linter script that doesn't run on Windows, and sometimes disagrees with clang-format.
I can do all this when I reactivate the review. I'm going to abandon this one for now and return to the Global/Local Alloc work after a few other changes are submitted.

We have a set of patches that will fix 
a) the cmake issues causing both the asan and asan_dynamic libs to be built with MT.
b) Interceptor issues preventing us from supporting _DEBUG runtimes
c) remove dependence on HeapValidate for control-flow in these interceptors since it throws debug breakpoints under a debugger when the result is false (and that function also throws assertions with M[TD]d).

After some discussion with @iakronqu I'm going to pull this review for now and add these interceptors back until after these other patches are submitted.


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D72011/new/

https://reviews.llvm.org/D72011