[lldb] Deflake TestTsanBasic.py

[dvyukov]

Summary

The test flaked on bots:
http://green.lab.llvm.org/green/job/lldb-cmake/38666/
The test expects that tsan will detect a single race
with concurrent memory accesses. TSan doesn't do this reliably.
Run 100 iterations of the racing threads, which should
make the race much more likely to be detected.

[melver]

Would using __tsan_testonly_barrier be appropriate here? If it's not meant to use any tricks to achieve the race then I guess retrying a bunch of times is fine.

[dvyukov]

In D114444#3148963, @melver wrote:

Would using __tsan_testonly_barrier be appropriate here? If it's not meant to use any tricks to achieve the race then I guess retrying a bunch of times is fine.

I did not want to leak this internal interface from tsan tests.

lldb/test/API/functionalities/tsan/basic/main.c

 
 void *f2(void *p) {
     pointer[0] = 'y'; // thread2 line
     return NULL;
 }
 
 int main (int argc, char const *argv[])
 {
+    for (int i = 0; i < 100; i++) {
         pointer = (char *)malloc(10); // malloc line
 
         pthread_t t1, t2;
         pthread_create(&t1, NULL, f1, NULL);
         pthread_create(&t2, NULL, f2, NULL);
 
         pthread_join(t1, NULL);
         pthread_join(t2, NULL);
 
+        free(pointer);
+    }
     return 0;
 }