24.10.2020 19:54, Tom Lane wrote:
> I'm forced to the conclusion that there's something wrong with
> ECPG's emulation of pthread_once ...
>
> ... and now that I look at it, it seems just as obvious what
> is wrong there:
>
> void
> win32_pthread_once(volatile pthread_once_t *once, void (*fn) (void))
> {
> if (!*once)
> {
> pthread_mutex_lock(&win32_pthread_once_lock);
> if (!*once)
> {
> *once = true;
> fn();
> }
> pthread_mutex_unlock(&win32_pthread_once_lock);
> }
> }
>
> We should not set *once until AFTER we execute fn().
> Otherwise, other threads passing through pthread_once()
> will mistakenly fall through, expecting the initialization
> to be done already.
>
> (So in this view, adding a sleep just before fn() would
> make the failure more reproducible.)
Yes, adding "pg_usleep(1000L);" just before fn() leads to 15 of 100
tests failed (without the delay more than 100 iterations could pass
successfully).
And the reverse test construction:
fn();
pg_usleep(1000L);
*once = true;
Makes all the tests (I ran 30x100 iterations) pass just fine.
Thank you for looking into this!
Best regards,
Alexander