Being honest with MODULE_LICENSE [LWN.net]

Being honest with MODULE_LICENSE

Posted Apr 28, 2004 22:26 UTC (Wed) by proski (subscriber, #104)
In reply to: Being honest with MODULE_LICENSE by rjamestaylor
Parent article: Being honest with MODULE_LICENSE

The free replacement is http://ndiswrapper.sourceforge.net/

But it looks like LinuxAnt has fixed the problem in the latest drivers. driverloader-1.66_k2.4.22_1.2188.nptl-1fdr.athlon.rpm.zip doesn't lie about its license.

$ modinfo driverloader.o 
kernel_version: 2.4.22-1.2188.nptl
using_checksums:1
author:         Copyright (C) 2003-2004 Linuxant inc.
description:    Linuxant DriverLoader for Wireless LAN devices
license:        see LICENSE file; Copyright (c)2003-2004 Linuxant inc.
kernel_version: 2.4.22-1.2188.nptl

(Log in to post comments)

Being honest with MODULE_LICENSE

Posted Apr 29, 2004 1:47 UTC (Thu) by farnz (guest, #17727) [Link]

DriverLoader has always been honest about its licence; LinuxAnt's "objection" to the taint for their modem driver is that it "confuses" customers if each module that the driver loads causes a taint message to appear. So, one taint message is OK, but several messages is trouble.

Personally, I feel LinuxAnt are being unnecessarily underhand; it would have been just as effective to ask the developer of modutils to provide a method to warn once for a sequence of module insertions done by modprobe. In other words, if I call "modprobe foo; modprobe bar", where both are tainted, I get two taint messages. If I call "modprobe --error-on-fail foo --no-error-on-fail bar baz biff" where all 4 modules are tainted, modprobe prints one taint message for all four modules, inserts foo or fails, and then tries to insert bar, baz and biff, not worrying if any of them fail to insert due to lack of hardware.

This fits what they claim to want to do, but avoids annoying kernel developers too much. It also ensures that oops output remains tainted, which the current solution doesn't.

Thank you for correction!

Posted Apr 29, 2004 2:09 UTC (Thu) by proski (subscriber, #104) [Link]

Sorry, I wrongly assumed that rjamestaylor checked the facts before uttering his tirade.

Thank you for reading!

Posted Apr 29, 2004 2:23 UTC (Thu) by rjamestaylor (guest, #339) [Link]

Eh, huh?

I noticed your previous post went off about "DriverLoader", which I didn't understand as I'm not familiar with it, but I never meant to refer to that package in the first place. In fact, the context of my "tirade" was Linuxant, not DriverLoader; nor do I mention DriverLoader or driverloader. Though I may have thrown you off with general descriptive language:

but I will definitely request a refund from Linuxant and remove the driver loader from my system.

In my non-kernel hacker mind, Linuxant is loading the Windows binary driver into Linux for me to use. That's what I meant. That there is a package/utility called DriverLoader...well, I assumed you would exercise reading comprehension skills before jumping to wild conclusions from an insignificant phrase that would have to be taken out of context not only of the post but of the very sentence in which the phrase exists.

Thanks for playing!

Posted Apr 29, 2004 2:26 UTC (Thu) by rjamestaylor (guest, #339) [Link]

(NB: that "retort" was meant as self-deprecating ironic humor to anyone who didn't get it :)

I got confused too initially

Posted Apr 29, 2004 2:34 UTC (Thu) by farnz (guest, #17727) [Link]

When I first heard about the LinuxAnt fuss, I also wrongly assumed that it was both DriverLoader and the Conexant modem stuff (having taken care to ensure that there are free drivers for all hardware in this machine). I still agree with rjamestaylor; LinuxAnt should be penalised by people who care about freedom of software for trying to bypass the usual rules, whether or not their intent was honourable.

I know this isn't Slashdot, and thus I'm probably preaching to the choir, but I don't feel that anything that helps Linux adoption is always beneficial. Part of the point of Linux for me is that it's free software in RMS's sense, and as such, people who for whatever reason do not wish to keep to the freedom side of things deserve penalising. The rules exist that allow you to co-exist adequately with those of us who don't buy devices that require binary-only drivers, and if you start bypassing them, I'm going to be unhappy.

My personal setup includes machines of 4 different processor architectures; I have both 64 and 32 bit systems, with both little and big endian byte orders. If I install free software on one, I know that it'll work on the others, provided I'm prepared to debug it. I like to have it made clear to me when I'm installing non-free software, so that I know that it may not work on any other machine I posess. LinuxAnt's trick has potential to confuse me if I'm not careful and rely on kernel tainting to spot non-free drivers, and for that reason I feel they are morally in the wrong.