Re: free tg3 ethernet driver

[Davi Leal]

Ali Gunduz wrote:
> Davi Leal wrote:
> > I propose we cancel and close the donation-pledge-group 58, [1], because
> > of Alan Cox tell us it would be an almost impossible task to work out.
>
> I don't doubt Alan Cox' technical opinion, but still I see no reason
> to cancel my pledge. In worst case, it raises awareness of the
> non-free status of a current linux module and may increase pressure on
> Broadcom just a little more.

Now I see you are right. My proposal was a very bad one.

> I can't tell anyone else whether to cancel their pledged money or not,
> but I personally continue to pledge that I would happily pay $100 to
> the developer(s) (even if it's Broadcom in case they apply to the
> pledge system) who provide a free driver for tg3 devices.
>
> Please let me know if there is a misunderstanding on my part.

My proposal was a bad one.

> Also, the crucial part here is what's non-free in the overall-GPL'ed
> driver. In case, it is a relatively small part it may still be
> possible after all. An example of such a development is the openHAL to
> replace (directly hardware-interacting) binary-only HAL part of the
> otherwise gpl'ed atheros wireless drivers. (for details:
> http://madwifi.org/wiki/About/OpenHAL?redirectedfrom=OpenHAL)
>
> I don't know if Alan Cox made his opinion about rewriting the _whole_
> driver with considering exactly what part of tg3 is non-free or as a
> general example of device driver reverse engineerings.

I have updated the pledge [1] with more of your information. As you know, you 
can delete, add or modify anything of such information.

 [1] http://gnuherds.org/pledges?id=58

> > Also, the crucial part here is what's non-free in the overall-GPL'ed
> > driver.
>
> I sent a second email about exactly what part is considered non-free
> to FSF 4 days ago since I didn't get any response to my first email I
> sent 6 weeks ago. I still haven't got a response. It would be very
> useful, if we can make this clear. Do you know anyone in FSF that
> could accelerate this process?

The non-free part is said to be a firmware blob [2]. You can just ask at the 
linux kernel mailing list.

 [2] http://www.uwsg.indiana.edu/hypermail/linux/kernel/0810.1/0316.html

Of course, reverse engineer such firmware is not in the "FSF High Priority 
list" [3]. There are others more important thing to do for general good. 
However, IMHO, the GNU Herds pledge tool is a good one to try to get done 
something of your personal-or-group interest.

 [3] http://www.fsf.org/campaigns/priority.html