[PATCH] Hauppauge HVR-1975 support

[PATCH] Hauppauge HVR-1975 support

[Bernhard Rosenkränzer]

Hi,
Hauppauge HVR-1975 is a USB DVB receiver box,
http://www.hauppauge.co.uk/site/products/data_hvr1900.html

It is currently not supported by v4l; Hauppauge provides a patch for
kernel 3.19 at http://www.hauppauge.com/site/support/linux.html

As expected, the patch doesn't work with more recent kernels, so I've
ported it (verified to work on 4.11.8). Due to the size of the patch,
I've uploaded my patch to
http://lindev.ch/hauppauge-hvr-1975.patch

While it works well, there's a potential license problem in one of the files:
>From drivers/media/dvb-frontend/silg.c:

/* MODULE_LICENSE("Proprietary"); */
/* GPL discussion for silg not finished. Set to GPL for internal usage only. */
/* The module uses GPL functions and is rejected by the kernel build if the */
/* license is set to 'Proprietary'. */
MODULE_LICENSE("GPL");

I'm not a lawyer, but my understanding is that by Hauppauge actually
releasing that file to the public (and it being so clearly a derivate
of GPL code that they even have to acknowledge it), their claim that
it is anything but GPL is null and void - but we may have to make
sure.

ttyl
bero

Re: [PATCH] Hauppauge HVR-1975 support

[Steven Toth]

(Resending)

Bernhard, thank you for sharing.

Mauro,

I've reviewed this patch, it has a host of problems.

Ignoring the fact it contains patches to all sorts of different cards
(saa7164, CX231xx, PVR-USB2)... the patch also contains materials that
I suspect Silicon Labs would consider proprietary and confidential,
its definitely derived works from proprietary SILABS drivers.

Proceed with caution.

- Steve

-- 
Steven Toth - Kernel Labs
http://www.kernellabs.com

On Mon, Jul 3, 2017 at 5:57 AM, Bernhard Rosenkränzer
<bernhard.rosenkranzer@linaro.org> wrote:
> Hi,
> Hauppauge HVR-1975 is a USB DVB receiver box,
> http://www.hauppauge.co.uk/site/products/data_hvr1900.html
>
> It is currently not supported by v4l; Hauppauge provides a patch for
> kernel 3.19 at http://www.hauppauge.com/site/support/linux.html
>
> As expected, the patch doesn't work with more recent kernels, so I've
> ported it (verified to work on 4.11.8). Due to the size of the patch,
> I've uploaded my patch to
> http://lindev.ch/hauppauge-hvr-1975.patch
>
> While it works well, there's a potential license problem in one of the files:
> From drivers/media/dvb-frontend/silg.c:
>
> /* MODULE_LICENSE("Proprietary"); */
> /* GPL discussion for silg not finished. Set to GPL for internal usage only. */
> /* The module uses GPL functions and is rejected by the kernel build if the */
> /* license is set to 'Proprietary'. */
> MODULE_LICENSE("GPL");
>
> I'm not a lawyer, but my understanding is that by Hauppauge actually
> releasing that file to the public (and it being so clearly a derivate
> of GPL code that they even have to acknowledge it), their claim that
> it is anything but GPL is null and void - but we may have to make
> sure.
>
> ttyl
> bero

Re: [PATCH] Hauppauge HVR-1975 support

[Bernhard Rosenkränzer]

On 3 July 2017 at 14:56, Steven Toth <stoth@kernellabs.com> wrote:
> Bernhard, thank you for sharing.
>
> Mauro,
>
> I've reviewed this patch, it has a host of problems.

Yes - it's a 1:1 forward port of the patch Hauppauge released for 3.19
(apparently with the goal to support as many of their devices as
possible).

> the patch also contains materials that I
> suspect Silicon Labs would consider proprietary and confidential, its
> definitely derived works from proprietary SILABS drivers.

Does anyone know for sure what the legal situation is when a HW
manufacturer releases a patch (as Hauppauge did) that is clearly
derived from GPL code yet at the same time derived from non-free code?
My interpretation is that by putting it out, they've released a GPL
derived work, which they can legally do only if they agree to comply
with the GPL, therefore any other license notice would be void.
But as I pointed out before I'm not a lawyer...

ttyl
bero

Re: [PATCH] Hauppauge HVR-1975 support

[Steven Toth]

> Yes - it's a 1:1 forward port of the patch Hauppauge released for 3.19
> (apparently with the goal to support as many of their devices as
> possible).

Agreed.

>
>> the patch also contains materials that I
>> suspect Silicon Labs would consider proprietary and confidential, its
>> definitely derived works from proprietary SILABS drivers.
>
> Does anyone know for sure what the legal situation is when a HW
> manufacturer releases a patch (as Hauppauge did) that is clearly
> derived from GPL code yet at the same time derived from non-free code?
> My interpretation is that by putting it out, they've released a GPL
> derived work, which they can legally do only if they agree to comply
> with the GPL, therefore any other license notice would be void.
> But as I pointed out before I'm not a lawyer...

You've raised a valid question, I don't know the answer. Others might.

I'm not a lawyer either, but if Hauppauge are not careful then they
may be violating an NDA, especially as the patch doesn't appear to
come with a sign-off, and leans very heavily on intellectual property
of Silicon Labs. I think in its current format the patch probably
wouldn't be acceptable for merge unless Hauppauge themselves provide a
sign-off.

Side note: obviously the fact it's such a large patch would require it
to be split into patches to each sub-system/card, but that's largely
beside the point of my larger concern.

Perhaps Hauppauge have legal approval to derive GPL drivers from
proprietary ,aterials, in which case I'm just making noise and a
sign-off will be soon to follow.

I'll reach out to them and ask.

-- 
Steven Toth - Kernel Labs
http://www.kernellabs.com