----------------------------------------
> Date: Tue, 6 Jan 2015 10:16:33 +0100
> From: ja...@redhat.com
> To: bernd.edlin...@hotmail.de
> CC: mikest...@comcast.net; hjl.to...@gmail.com; gcc-patches@gcc.gnu.org;
> dvyu...@google.com
> Subject: Re: [PATCH] Fix sporadic failure in
> g++.dg/tsan/aligned_vs_unaligned_race.C
>
> On Tue, Jan 06, 2015 at 10:08:22AM +0100, Bernd Edlinger wrote:
>> Hi Mike,
>>
>> after some hours of sleep I realized that your step function can do
>> something very interesting,
>> (which you already requested previously):
>>
>> That is: create a race condition that is _always_ at 100% missed by tsan:
>>
>> cat lib.c
>> /* { dg-do compile } */
>> /* { dg-options "-O2 -fno-sanitize=all" } */
>>
>> static volatile int serial = 0;
>>
>> void step (int i)
>> {
>> while (__atomic_load_n (&serial, __ATOMIC_ACQUIRE) != i - 1);
>> __atomic_store_n (&serial, i, __ATOMIC_RELEASE);
>> }
>
> Such busy waiting is not very CPU time friendly in the testsuite, especially
> if you have just a single HW thread.
> If libtsan is not fixed not to be so racy, perhaps instead of all the sleeps
> we could arrange (on x86_64-linux, which is the only target supporting tsan
> right now) to make sure the thread runs on the same core/hw thread as the
> initial thread using pthread_[gs]etaffinity_np/pthread_attr_setaffinity_np ?
> Does tsan then always report the races when the threads can't run
> concurrently?
>
> Jakub
I tried your suggestion now, and it seems to work. (on a 4-way core AMD64
laptop)
Would you prefer this over adding a sleep in Thread1, which I posted previously?
2015-01-06 Bernd Edlinger <bernd.edlin...@hotmail.de>
* g++.dg/tsan/aligned_vs_unaligned_race.C: Fixed sporadic failure.
Index: aligned_vs_unaligned_race.C
===================================================================
--- aligned_vs_unaligned_race.C (revision 219198)
+++ aligned_vs_unaligned_race.C (working copy)
@@ -1,5 +1,7 @@
/* { dg-shouldfail "tsan" } */
+/* { dg-additional-options "-pthread" } */
#include <pthread.h>
+#include <sched.h>
#include <stdio.h>
#include <stdint.h>
@@ -19,6 +21,12 @@ void *Thread2(void *x) {
}
int main() {
+ /* Run this test on a single CPU, to make it somewhat easier for tsan to
+ detect the race condition. */
+ cpu_set_t s;
+ CPU_ZERO(&s);
+ CPU_SET(0, &s);
+ pthread_setaffinity_np(pthread_self(), sizeof(s), &s);
pthread_t t[2];
pthread_create(&t[0], NULL, Thread1, NULL);
pthread_create(&t[1], NULL, Thread2, NULL);
Thanks,
Bernd.
