Focus -Wno-over-aligned to just 32-bit x86 Android.

I've even found the code that's making this happen, just don't know why.
I've added a test to assert that it's safe to assume malloc() is 8-byte aligned.

BUG=skia:

GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2662

CQ_INCLUDE_TRYBOTS=master.client.skia.android:Test-Android-Clang-NexusPlayer-CPU-Moorefield-x86-Release-GN_Android-Trybot

Change-Id: If8a2898ab3a77571622eb125c97f676e029b902c
Reviewed-on: https://skia-review.googlesource.com/2662
Commit-Queue: Ben Wagner <bungeman@google.com>
Reviewed-by: Ben Wagner <bungeman@google.com>

gn/BUILD.gn

@@ -83,8 +83,11 @@ config("default") {
         "-Wno-unknown-warning-option",  # Let older Clangs ignore newer Clangs' warnings.
       ]
 
-      # High priority to fix!
-      cflags += [ "-Wno-over-aligned" ]
+      if (is_android && target_cpu == "x86") {
+        # Clang seems to think new/malloc will only be 4-byte aligned on x86 Android.
+        # We're pretty sure it's actually 8-byte alignment.
+        cflags += [ "-Wno-over-aligned" ]
+      }
 
       cflags += [
         "-Wno-cast-align",

tests/OverAlignedTest.cpp

@@ -0,0 +1,22 @@
+/*
+ * Copyright 2016 Google Inc.
+ *
+ * Use of this source code is governed by a BSD-style license that can be
+ * found in the LICENSE file.
+ */
+
+#include "Test.h"
+#include "SkRandom.h"
+
+// Clang seems to think only 32-bit alignment is guaranteed on 32-bit x86 Android.
+// See https://reviews.llvm.org/D8357
+// This is why we have disabled -Wover-aligned there (we allocate 8-byte aligned structs in Ganesh).
+DEF_TEST(OverAligned, r) {
+    SkRandom rand;
+    // Let's test that assertion.  We think it really should be providing 8-byte alignment.
+    for (int i = 0; i < 1000; i++) {
+        void* p = malloc(rand.nextRangeU(0,100));
+        REPORTER_ASSERT(r, SkIsAlign8(p));
+        free(p);
+    }
+}