AW: svn commit: r1787606 - /httpd/httpd/trunk/server/core.c

[Plüm , Rüdiger , Vodafone Group]

C2 General

> -----Ursprüngliche Nachricht-----
> Von: Christophe JAILLET <christophe.jail...@wanadoo.fr>
> Gesendet: Samstag, 4. Mai 2019 09:20
> An: dev@httpd.apache.org
> Betreff: Re: svn commit: r1787606 - /httpd/httpd/trunk/server/core.c
> 
> Hi,
> 
> In my perpetual effort to synch 2.4 and trunk, I have stopped r1787606
> as a potential candidate.
> 
> However, I don't understand the rational with this patch and wonder if
> it should be backported, left as-is in trunk only or removed from trunk.
> 
> Why should we avoid timeout and TCP_NODELAY settings for slave
> connection?

Because to my understanding these are only 'logical' connections that share the
masters TCP connection.

> Is it already set some other way, (i.e APR_TCP_NODELAY_INHERITED???)

IMHO on the master connection.

> 
> 
> This patch also makes the 'if (c->master == NULL) {' block at the end of
> this function unclear.
> Now, we know that c->master != NULL and the logic could be simplified.

I agree that the c->master == NULL could be removed. I guess this could then be
backported together with r1787606.

Regards

Rüdiger

> 
> 
> Any opinion?
> 
> CJ
> 
> 
> Le 19/03/2017 à 11:33, ic...@apache.org a écrit :
> > Author: icing
> > Date: Sun Mar 19 10:33:43 2017
> > New Revision: 1787606
> >
> > URL: http://svn.apache.org/viewvc?rev=1787606&view=rev
> > Log:
> > On the trunk:
> >
> > core: avoid socket timeout settings etc. on slave connections.
> >
> >
> > Modified:
> >      httpd/httpd/trunk/server/core.c
> >
> > Modified: httpd/httpd/trunk/server/core.c
> > URL:
> http://svn.apache.org/viewvc/httpd/httpd/trunk/server/core.c?rev=1787606
> &r1=1787605&r2=1787606&view=diff
> >
> ========================================================================
> ======
> > --- httpd/httpd/trunk/server/core.c (original)
> > +++ httpd/httpd/trunk/server/core.c Sun Mar 19 10:33:43 2017
> > @@ -5280,9 +5280,14 @@ static conn_rec *core_create_conn(apr_po
> >
> >   static int core_pre_connection(conn_rec *c, void *csd)
> >   {
> > -    core_net_rec *net = apr_palloc(c->pool, sizeof(*net));
> > +    core_net_rec *net;
> >       apr_status_t rv;
> >
> > +    if (c->master) {
> > +        return DONE;
> > +    }
> > +
> > +    net = apr_palloc(c->pool, sizeof(*net));
> >       /* The Nagle algorithm says that we should delay sending partial
> >        * packets in hopes of getting more data.  We don't want to do
> >        * this; we are not telnet.  There are bad interactions between
> >
> >
> >