search for: d12001

...ttps://android-review.googlesource.com/#/c/167420/ I asked somewhere else, but, will this give you a reasonably looking thread state without resorting to pthread_get/specific calls? Will this be a unique and non-redundant state for each thread? > 4. use the emutls work http://reviews.llvm.org/D12001 I only suggested this because Joerg/Dan mentioned it. I haven't looked at that patch in detail. From what I'm seeing, options 1 and 3 above may very well suit us. > TESTS! > Currently, about 2/3 tests for tsan fail/flake on android+aarch64. > They aren't xfailed, because so...

On Wed, Aug 19, 2015 at 1:15 PM, Renato Golin <renato.golin at linaro.org> wrote: > On 19 August 2015 at 11:29, Dmitry Vyukov <dvyukov at google.com> wrote: >> Wait, this change is not submitted yet, right? Or you mean mailing of >> this change in bad shape? > > Right. > > Jason has submitted high quality patches before, so this is in no way > a reprimand

On Tue, Aug 18, 2015 at 10:53 PM, Renato Golin <renato.golin at linaro.org> wrote: > On 18 August 2015 at 19:57, Dmitry Vyukov <dvyukov at google.com> wrote: >> Hi Renato, >> >> What exactly hack/hacks do you mean? > > Hi Dmitry, > > The sanitizer code seems to be more accepting to adding complicated > pre-processor logic, as well as a number of