bug#32452: 26.1; gnutls_try_handshake maxes out cpu retrying when server is a bit busy

[Lars Ingebrigtsen]

Noam Postavsky <npostavs@gmail.com> writes:

> +  enum { MAX_DELAY_NS = 100 * 1000 * 1000 }; // Max 100ms delay.
> +  struct timespec delay = { 0, 1000 };
>    do
>      {
>        ret = gnutls_handshake (state);
> @@ -558,7 +560,9 @@ gnutls_try_handshake (struct Lisp_Process *proc)
>      }
>    while (ret < 0
>        && gnutls_error_is_fatal (ret) == 0
> -      && ! non_blocking);
> +      && ! non_blocking
> +         && (nanosleep (&delay, NULL),
> +             delay.tv_nsec = min (delay.tv_nsec * 2, MAX_DELAY_NS)));

This code has changed somewhat since the proposed patch:

  while ((ret = gnutls_handshake (state)) < 0)
    {
      do
        ret = gnutls_handshake (state);
      while (ret == GNUTLS_E_INTERRUPTED);

      if (0 <= ret || emacs_gnutls_handle_error (state, ret) == 0
          || non_blocking)
        break;
      maybe_quit ();
    }

But perhaps adding some sleep here would be a good idea, anyway.  On the
other hand, 100ms seems way too long -- that's a noticeable slow-down in
network set-up.

-- 
(domestic pets only, the antidote for overdose, milk.)
   bloggy blog: http://lars.ingebrigtsen.no