Subject: Re: Alternative approach for interface events
To: Jason Thorpe <thorpej@shagadelic.org>
From: Jonathan Stone <jonathan@dsg.stanford.edu>
List: tech-net
Date: 09/20/2004 19:42:50
[.... again, replies should probably go to tech-kern...]
In message <E47413F0-0B50-11D9-AD6D-000A957650EC@shagadelic.org>,
Jason Thorpe writes:
>On Sep 20, 2004, at 1:21 PM, Peter Postma wrote:
>
>> Ok, here's a port of FreeBSD's eventhandler (and cleaned up).
>
>Aside from whether or not I like this API (I have to study it some
>more), here are some comments on the patch:
``Me too'', on needing to study more. I'm pretty sure there are bugs
in the kcont usages, even apart from the restrictions I noted earlier.
>1. eventhandler_init() is unnecessary. Statically initialize
>eventhandler_lists and eventhandler_slock instead.
Uh... would that restrict dynamically-loaded kernel modules which want
to use this API? What about building a dynamic list of all known
eventhanlder consumers, for management-api and debugging purposes?
[...]