Received: (qmail 92819 invoked by uid 501); 14 Sep 2000 00:08:57 -0000
Message-Id: <20000914000857.92818.qmail@locus.apache.org>
Date: 14 Sep 2000 00:08:57 -0000
From: Mike Abbott <mja@sgi.com>
Reply-To: mja@sgi.com
To: submit@bugz.apache.org
Subject: 2.0a6 doesn't disable Nagle algorithm
X-Send-Pr-Version: 3.110

>Number:         6531
>Category:       general
>Synopsis:       2.0a6 doesn't disable Nagle algorithm
>Confidential:   no
>Severity:       non-critical
>Priority:       medium
>Responsible:    apache
>State:          closed
>Quarter:        
>Keywords:       
>Date-Required:
>Class:          sw-bug
>Submitter-Id:   apache
>Arrival-Date:   Wed Sep 13 17:10:00 PDT 2000
>Closed-Date:    Sat Oct 07 21:03:55 PDT 2000
>Last-Modified:  Sat Oct 07 21:03:55 PDT 2000
>Originator:     mja@sgi.com
>Release:        2.0a6
>Organization:
>Environment:
Irix 6.5
>Description:
Apache/2.0a6 doesn't disable the Nagle algorithm on listening
sockets like Apache/1.3 does.  It is still disabled on accepted sockets
(returned by accept()) but not listening sockets (passed to accept()).
Compare make_sock() in 1.3 and 2.0.

This affects performance at least on Irix, where setsockopt(TCP_NODELAY)
is not necessary on the accepted socket if the listening socket has
TCP_NODELAY set.
>How-To-Repeat:

>Fix:
Restore the call to ap_sock_disable_nagle() to make_sock() ala 1.3 --
except that ap_sock_disable_nagle() is part of the MPM API (for some MPMs),
so maybe instead make every MPM do it after calling ap_listen_open() or
ap_setup_listeners().
>Release-Note:
>Audit-Trail:
State-Changed-From-To: open-closed
State-Changed-By: abagchi
State-Changed-When: Sat Oct  7 21:03:55 PDT 2000
State-Changed-Why:
If DISABLE_NAGLE_INHERITED is defined when building the
server, then the listening socket will disable_nagle,
otherwise the accepted socket must disable nagle itself.
>Unformatted:
 [In order for any reply to be added to the PR database, you need]
 [to include <apbugs@Apache.Org> in the Cc line and make sure the]
 [subject line starts with the report component and number, with ]
 [or without any 'Re:' prefixes (such as "general/1098:" or      ]
 ["Re: general/1098:").  If the subject doesn't match this       ]
 [pattern, your message will be misfiled and ignored.  The       ]
 ["apbugs" address is not added to the Cc line of messages from  ]
 [the database automatically because of the potential for mail   ]
 [loops.  If you do not include this Cc, your reply may be ig-   ]
 [nored unless you are responding to an explicit request from a  ]
 [developer.  Reply only with text; DO NOT SEND ATTACHMENTS!     ]