Making CSSValue Pool thread local

build/sanitizers/lsan_suppressions.cc

Index: build/sanitizers/lsan_suppressions.cc
diff --git a/build/sanitizers/lsan_suppressions.cc b/build/sanitizers/lsan_suppressions.cc
index afe4c356e9c2064d04457a8b92fee2d33baa4f99..72cb4c40197857f75faa89783b28b22a526db07e 100644
--- a/build/sanitizers/lsan_suppressions.cc
+++ b/build/sanitizers/lsan_suppressions.cc
@@ -21,6 +21,11 @@ char kLSanDefaultSuppressions[] =
 
 // False positives in libfontconfig. http://crbug.com/39050
 "leak:libfontconfig\n"
+// eglibc-2.19/string creates false positive leak errors because of the same
+// reason as crbug.com/39050. The leak error stack trace, when unwind on malloc,
+// includes a call to libfontconfig. But the default stack trace is too short
+// in leak sanitizer bot to make the libfontconfig suppression works.
+"leak:eglibc-2.19/string\n"

[Evgeniy Stepanov, 2016/04/20 20:53:26]

This would break on any different version of glibc.
Also, this effectively disables all leak detection for memory allocated in
strdup, which we would really like to avoid.

Do you have the full stack trace (with fast_unwind_on_malloc=0)?

How do I reproduce it? Would running webkit_unit_tests with asan+lsan do it?

[Evgeniy Stepanov, 2016/04/20 21:13:07]

OK, I've got it. I'll fix ASan/LSan upstream so that this stacktrace would look
like
1. strdup
2. /path/to/libfontconfig.so.1

Then the existing suppression would apply.

In the mean time, please change this suppression to "leak:__strdup" and file a
bug to remove it. Make this bug blocked by crbug.com/604993.

 
 // Leaks in Nvidia's libGL.
 "leak:libGL.so\n"