Comment 78 for bug 120434

Revision history for this message
Alexander Hixon (ahixon) wrote :

Hi Micheal,

Nice work with the packaging! :)
Some comments:

> sudo r5u87x-loader
> reboot
> cheese

Rebooting is not necessary.
If you read README, you will note that it's suggested you run it as r5u87x-loader --reload instead.

I should probably also note two other things:

r5u87x has already been packaged before, so you can always want to have a peek at the existing dscs and stuff. You might be able to contact wvengen to see if he still has them (see earlier in this bug report).

Secondly, and most importantly; we do not have copyright on the firmware. This means it is impossible to redistribute this legally (comment #66) without permission (including in the PPA). As such you won't be able to include this in universe, from my understanding. Multiverse perhaps. Even then, we would also need some mechanism for users to know that they need package xyz from multiverse. What would you (or any Canonical folk) suggest?

There are two options re firmware:
a) re-implement the firmware from scratch - can be done, in fact, we've successfully decompiled the microcode to assembler. It'd be probably insanely time consuming to redo it all, though.
b) see if there's anything in the click-through EULA's provided with the drivers that allow us to redistribute it.
c) get permission from the copyright owner(s) - AFAIK this is Ricoh, except several people (including myself) have attempted to get in contact with them and failed. Sales has never contacted me, ever, even after being assured they would. If somebody in the right timezone is willing to give them a ring...
d) create a tool that extracts the firmware from Windows firmware files, however this is probably the least ideal solution. There is code that assists in this process, however it's not quite automated at the moment.

Ideally, I'd still be like to be able to include something based off the old r5u870 module in the kernel based on gspca. However, I really just don't have the time that I'd like to be able to develop something like this. If someone is feeling adventurous, a free beer is up for grabs. :) This would also mean we can support WDM based cameras, too - which is a limitation with the current approach; we only work with UVC-based cameras (and then again, not completely correctly, either - controls are not accurately recognised - yay non standards).

Cheers,
Alex Hixon