Re: [exim] data-phase efficiency

Philip Hazel wrote:
> On Fri, 20 May 2005, Philip Hazel wrote:
>
> > On Fri, 20 May 2005, Ian FREISLICH wrote:
> >
> > > You can do much the same with select(). Just replace my use of
> > > poll() with appropriate select() magic. The purpose of this code
> > > snippet was not to demonstrate the use of poll, but non-blocking
> > > IO. When you write to an fd, it may block and you have no idea how
> > > much you can write before it will block. If it's set to O_NONBLOCK,
> > > as soon as poll()/select() says you can write without blocking, you
> > > write() and write() returns at the point it would block with the
> > > number of bytes actually written.
> >
> > OK. Point taken.
>
> However, when I came to look at possibly doing something to the code, I
> realized that it is complicated by the TLS case. I am not sure if it is
> even possible to do any kind of blocking check when TLS is in use. So I
> have left the code as it is. Well, almost. While reading it I noticed a
> bug: when a buffer is only partly written by one write() call, Exim was
> re-applying the same timeout to the next write(). This seems wrong, as
> it would allow a very slow acceptor to delay for a very long time. So I
> fixed that.

Oh. I would think that is right. Any data accepted by the remote
host should reset the timeout. You can't expect your whole buffer
to be transferred in a single packet, can you?

Ian

--
Ian Freislich