Hi Willy,
Le 13/08/2018 à 09:28, Willy Tarreau a écrit :
Hi Cyril,
On Sun, Aug 12, 2018 at 10:49:13PM +0200, Cyril Bonté wrote:
2. hlua_socket_settimeout() initializes rto/wto values, maybe it should also
compute the rex/wex values :
socket->s->req.rex = tick_add_ifset(now_ms, tmout);
socket->s->req.wex = tick_add_ifset(now_ms, tmout);
socket->s->res.rex = tick_add_ifset(now_ms, tmout);
socket->s->res.wex = tick_add_ifset(now_ms, tmout);
You're absolutely right, it's indeed incorrect to set {r,w}to without
updating the expire date. It only works when the timeout is extended,
but not when it's shortened.
3. It may require to wake up the task if a new timeout is set after a first
one was already set (in your case the task doesn't wake up after 2 secondes
because a first timeout was set to 3 seconds) :
task_wakeup(socket->s->task, TASK_WOKEN_OTHER);
Normally it should not be needed, however it needs to be requeued to
take care of the fact that the timeout might have been shortened :
socket->s->task->expire = tick_add_ifset(now_ms, tmout);
task_queue(socket->s->task);
At least, it seems to fix the issue but before sending a patch, I want to be
sure that's how we should fix this.
If you can just validate that the two lines above are enough, that would
be great. Avoiding to call process_stream() just to update a timeout is
desirable. If for any reason it doesn't work or seems more complicated,
then the wakeup as you proposed should indeed work.
Thanks for the feedback. I've made some tests withe the lines of code
you suggested and it looks to work fine. I'll prepare the patches tomorrow.
--
Cyril Bonté
