Subject: Re: Do LKMs work *at*all* on powerpc platforms?
To: Todd Vierling <tv@wasabisystems.com>
From: Matt Thomas <matt@3am-software.com>
List: port-macppc
Date: 08/15/2000 17:48:32
At 07:17 PM 8/15/2000 -0400, Todd Vierling wrote:
>On Tue, 15 Aug 2000, Bill Studenmund wrote:
>
>: Well, gabriel rosenkoetter <gr@eclipsed.net> has had problems. He made
>: prototypes of all kernel functions with __attribute__ ((long-call)), and
>: made an lkm. He was still left with truncated references to memcpy - the
>: module still had it as an external symbol, and had used the too-small
>: branch option.
>
>If it's used as a command-line option, it should affect all references (so
>long as the implementation is written properly).

Digital UNIX had a similar problem and changed the linker so that it could
generate small "intermediate jumps" to bridge the displacement problem.
--
Matt Thomas               Internet:   matt@3am-software.com
3am Software Foundry      WWW URL:    http://www.3am-software.com/bio/matt/
Cupertino, CA             Disclaimer: I avow all knowledge of this message