From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <gorcunov@gmail.com>
Received: from mail-lf1-f65.google.com (mail-lf1-f65.google.com
 [209.85.167.65])
 (using TLSv1.2 with cipher ECDHE-RSA-AES128-GCM-SHA256 (128/128 bits))
 (No client certificate requested)
 by dev.tarantool.org (Postfix) with ESMTPS id DA97946970E
 for <tarantool-patches@dev.tarantool.org>;
 Fri, 20 Dec 2019 17:50:47 +0300 (MSK)
Received: by mail-lf1-f65.google.com with SMTP id 15so7265516lfr.2
 for <tarantool-patches@dev.tarantool.org>;
 Fri, 20 Dec 2019 06:50:47 -0800 (PST)
Date: Fri, 20 Dec 2019 17:50:44 +0300
From: Cyrill Gorcunov <gorcunov@gmail.com>
Message-ID: <20191220145044.GA18623@uranus>
References: <20191217125420.20881-1-gorcunov@gmail.com>
 <20191217125420.20881-2-gorcunov@gmail.com>
 <20191220074847.GF30445@atlas>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20191220074847.GF30445@atlas>
Subject: Re: [Tarantool-patches] [PATCH v6 1/4] coio: Export helpers and
 provide coio_read_fd_timeout
List-Id: Tarantool development patches <tarantool-patches.dev.tarantool.org>
List-Unsubscribe: <https://lists.tarantool.org/mailman/options/tarantool-patches>, 
 <mailto:tarantool-patches-request@dev.tarantool.org?subject=unsubscribe>
List-Archive: <https://lists.tarantool.org/pipermail/tarantool-patches/>
List-Post: <mailto:tarantool-patches@dev.tarantool.org>
List-Help: <mailto:tarantool-patches-request@dev.tarantool.org?subject=help>
List-Subscribe: <https://lists.tarantool.org/mailman/listinfo/tarantool-patches>, 
 <mailto:tarantool-patches-request@dev.tarantool.org?subject=subscribe>
To: Konstantin Osipov <kostja.osipov@gmail.com>
Cc: tml <tarantool-patches@dev.tarantool.org>

On Fri, Dec 20, 2019 at 10:48:47AM +0300, Konstantin Osipov wrote:
> * Cyrill Gorcunov <gorcunov@gmail.com> [19/12/17 15:57]:
> > There is no reason to hide functions. In particular
> > we will use coio_write_fd_timeout and coio_read_fd_timeout
> > for popen.
> 
> AFAIU the only difference between the new function and
> coio_read_timeout() is that the new function works without 
> struct coio* object, but uses coio_wait(), which creates/
> destroys this object on demand.

Yes. This allows us to reuse it instead of code duplication.

> 
> Could you provide a rationale for this? coio_wait has to call 
> EPOLL_CTL_ADD/EPOLL_CTL_DEL on every wait, so it tripples the 
> number of syscalls per wait. 

But how else we can do that? libev uses poll internally to find
out if there some "event" in the event queue (timer event, io
event and such). Could you please elaborate which idea you have?

Indeed every time we start io on fd, the libev puts this fd into
watchee list to be notified if there some event to handle. I must
admit I don't see an other way.

> 
> On the other hand I realize that it's not super important
> for popen IO, but still I don't understand *why* you need this.