lkml.org 
[lkml]   [1996]   [May]   [29]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
From
SubjectRe: alias wanted for binfmt_java
Date
In article <19960512144854.16808.qmail@Mail.UTexas.EDU>,
lilo <TaRDiS@mail.utexas.edu> wrote:
>
> Check the next release of the kernel (pre-2.0.1), available from All Fine
> Kernel Resellers Now<tm>. :)
>

This is really crazy that such auxilliary information should be needed
at all. A module should be self-contained; even if that would mean
making them .zip files with a description and a code section. That
way a module can be registered with kerneld without needing a bogus
configuration file.

Why .zip? The .zip format is better than .tar.gz when it comes to
extracting individual members, since each member is compressed
individually. It should not be hard to modify insmod and kerneld to
be able to extract the .o file from the kernel module.

Otherwise, it could just be kept in a constant string or array in the
module.

Something else I keep repeating I'd like to see is compatibility
between the module and kernel command line argument passing
convention, i.e. I'd like the module insertion system to be able to
call *_setup the same way the kernel does. It seems to be necessary
in order for drivers to support both in a coherent manner. Currently
a lot of devices are really bizarre when they need parameters passed
in module mode, and in addition, the whole concept of a completely
different user interface for module parameters is IMHO sick.

-hpa


\
 
 \ /
  Last update: 2005-03-22 13:37    [W:0.067 / U:2.284 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site