[PATCH] D12709: [PATCH] [sanitizers] [msan] Enable MSAN for aarch64

Renato Golin via llvm-commits llvm-commits at lists.llvm.org
Wed Sep 9 04:01:17 PDT 2015


rengolin added a reviewer: t.p.northover.
rengolin added a comment.

Adding Tim as a reviewer for the Darwin part of the XFAILs.


================
Comment at: lib/msan/msan_interceptors.cc:254
@@ +253,3 @@
+  register uptr r8 asm("x8");
+  sret = reinterpret_cast<__sanitizer_mallinfo*>(r8);
+#endif
----------------
implement this in asm and it will work on both gcc and clang.

================
Comment at: test/msan/chained_origin_limits.cc:65
@@ -64,1 +64,3 @@
+//
+// XFAIL: aarch64
 
----------------
I think this is "arm64" for Darwin. It'd be good if someone could test this in Darwin and make sure we need the same XFAILs, and if so, we should add their arch-name, too.

================
Comment at: test/msan/strlen_of_shadow.cc:25
@@ +24,3 @@
+  unsigned long vma = (unsigned long)__builtin_frame_address(0);
+  vma = 64 - __builtin_clzll(vma);
+
----------------
Can't you re-use the one you have above? Maybe this should even be in compiler-rt?

================
Comment at: test/sanitizer_common/TestCases/Linux/ptrace.cc:82
@@ +81,3 @@
+      printf("%x\n", fpregs.fpsr);
+#endif
+
----------------
#endif // (__aarch64__)


http://reviews.llvm.org/D12709





More information about the llvm-commits mailing list