> +    /* For threads other than main thread, wait for timewarp. */
> +    if (ovsthread_id_self()) {
> +        timewarp_wait();
> +    }
> +

Why do we need the special case for the main thread btw?  Does it
really matter if it wakes up twice?

> +    if (timewarp_enabled) {
> +        /* Changes 'timewarp_seq' to wake up all waiting thread. */
> +        seq_change(timewarp_seq);
> +    }

I'd expect that timewarp would always be enabled here, or if it isn't,
incrementing the seq won't hurt anything, may as well ditch the if
statement.  Also I don't think the comment is adding anything that
isn't obvious from the code.

If you don't disagree with my previous comments, I'll just fold them
in and merge.

Acked-by: Ethan Jackson <et...@nicira.com>



>      unixctl_command_reply(conn, "warped");
>  }
>
>  void
>  timeval_dummy_register(void)
>  {
> +    timewarp_enabled = true;
>      unixctl_command_register("time/stop", "", 0, 0, timeval_stop_cb, NULL);
>      unixctl_command_register("time/warp", "MSECS", 1, 1,
>                               timeval_warp_cb, NULL);
> diff --git a/lib/timeval.h b/lib/timeval.h
> index 99b3af0..1bbfd5c 100644
> --- a/lib/timeval.h
> +++ b/lib/timeval.h
> @@ -69,6 +69,8 @@ int get_cpu_usage(void);
>
>  long long int time_boot_msec(void);
>
> +void timewarp_wait(void);
> +
>  #ifdef  __cplusplus
>  }
>  #endif
> --
> 1.7.9.5
>
> _______________________________________________
> dev mailing list
> dev@openvswitch.org
> http://openvswitch.org/mailman/listinfo/dev
_______________________________________________
dev mailing list
dev@openvswitch.org
http://openvswitch.org/mailman/listinfo/dev

Reply via email to