Re: [PATCH] Pipeline prefetch build-time control

From: Kinkie <gkinkie_at_gmail.com>
Date: Tue, 14 May 2013 23:47:23 +0200

On Sun, May 12, 2013 at 6:00 AM, Amos Jeffries <squid3_at_treenet.co.nz> wrote:
> The attached patch adds a built-time setting SQUID_PIPELINE_MAX to replace
> the magic numbers previously used for limiting the amount of prefetch Squid
> does when pipeline_prefeth config directive is enabled. The default remains
> at 2 unless altered by using -DSQUID_PIPELINE_MAX=n in CXXFLAGS and this
> adds some documentation about why, and what can be done to improve
> pipelining in future.

On the matter of the TODO:
- it's up tro HTTP/1.0 clients not to send pipelined requests. If a
client behaves like HTTP/1.1, and it's harmless, why deny them the
pleasure?
- it'd probably be best to implement the blocking on NTLM/NEGOTIATE.
However, I doubt that a client would send a type 3 message until it
has received the type 2 back from squid. This is no guarantee and is
probably hard to troubleshoot if it happens, but I suspect that it
won't bite us in the wild. It'd have bitten us by now otherwise.

+1

--
    /kinkie
Received on Tue May 14 2013 - 21:47:30 MDT

This archive was generated by hypermail 2.2.0 : Wed May 15 2013 - 12:00:17 MDT