[U-Boot] [PATCH] drivers/ddr/fsl: Dual-license DDR driver

[Marek Vasut]

On 02/13/2018 08:33 PM, York Sun wrote:
> On 02/13/2018 11:16 AM, Marek Vasut wrote:
>> On 02/13/2018 07:32 PM, York Sun wrote:
>>> On 02/13/2018 09:38 AM, Marek Vasut wrote:
>>>> On 02/13/2018 05:30 PM, York Sun wrote:
>>>>> On 02/13/2018 04:49 AM, Wolfgang Denk wrote:
>>>>>> Dear York,
>>>>>>
>>>>>> In message <VI1PR04MB20785EF7D2578E39C048EE219AF70 at VI1PR04MB2078.eurprd04.prod.outlook.com> you wrote:
>>>>>>>
>>>>>>> Nobody said anything. Some addresses bounced. And most changes made out
>>>>>>> people outside Freescale/NXP are minor changes, except twice the files
>>>>>>> were moved during U-Boot structure change. What options do I have?
>>>>>>
>>>>>> Ask all people who contributed to that code for their explicit
>>>>>> permission.  Legally it is a huge difference between actively
>>>>>> confirming approval and not reacting at all.
>>>>>>
>>>>>
>>>>>
>>>>> All,
>>>>>
>>>>> If you haven't responded, please give your explicit approval to change
>>>>> Freescale DDR driver to dual-license so it can be re-used by other
>>>>> project(s) with BSD license. Here is the list I compiled from the git
>>>>> history. All commits made by Freescale/NXP employees are removed from
>>>>> this list.
>>>>
>>>> [...]
>>>>
>>>>> cd84b1f - Marek Vasut, marek.vasut at gmail.com, 6 years ago : GCC4.6:
>>>>> Squash warnings in ddr[123]_dimm_params.c
>>>>
>>>> I do NOT approve.
>>>>
>>>> My previous experience with dual-licensed code was with wpa-supplicant.
>>>> A certain company manufacturing handhelds took it, modified it and was
>>>> selling the binary. While we were porting Linux onto the device, we
>>>> asked for the modifications to get the WiFi operational in the Linux port.
>>>>
>>>> What we got from this company was "it's BSD licensed, go away". Were the
>>>> code GPL, they would be legally obliged to provide the changes, but it
>>>> was BSD, so the company in question could make profit and the community
>>>> lost.
>>>>
>>>> This was a prime example of how BSD license is harmful to software
>>>> freedom and how the community lost because of the BSD license. I do not
>>>> want to see this happening ever again and I like GPL for that very much.
>>>>
>>>
>>> Marek,
>>>
>>> Please allow me to try to convince you.
>>> Git log shows you have one commit cd84b1f which fixed the compiling
>>> warning for GCC 4.6 on three debug messages. I appreciate your fix.
>>>
>>> This driver is for Freescale/NXP DDR controllers, specifically designed
>>> on Freescale/NXP SoCs. We spent tremendous effort to make it robust.
>>> This driver is useful to initialize DDR for the platforms. While we are
>>> moving the platform initialization to ATF (Arm Trusted Firmware), or
>>> other pre-bootloader code (such as NXP's implementation of ATF), this
>>> driver can be reused to provide the same level of hardware support. As
>>> you may know, ATF uses BSD-3 license (some files have GPL/BSD dual
>>> licnese). Your approval will make our life easier without having to
>>> rewrite the entire driver from scratch.
>>
>> So what is in it for me ?
> 
> You may have the flexibility to use ATF or other pre-bootloader software
> _if_ we successfully upstream this driver to ATF project.

It also allows you to just distribute binaries of the ATF without
releasing the source.

>> If the code remains GPL, I can ask NXP for changes to the driver if I
>> have the binary which contains this code.
>>
>> If the code gets re-licensed to dual GPL/BSD, I assume in certain cases,
>> NXP will choose BSD and will not be obliged to provide the changes.
> 
> Guess who makes substantial changes to the hardware driver? The people
> with extensive knowledge of the hardware design. It's not our interest
> to hide our design from any users.
>
>> I don't see any benefit for me, any way I look at it, I'm either even or
>> loose .
> 
> If we don't find a way to reuse this driver, I will have to write a new
> driver. It's not easy to keep two different drivers in sync. So _this_
> driver will probably be left behind. I don't think that's in anyone's
> interest.
> 
>>
>> Why can't you use the code under the current (GPL) license anyway ?
>>
> 
> Do you think the GPL driver can be added to ATF project? I don't think
> so. So it is a matter of we either can have it in ATF, or we can't.

Well, it seems this patch was applied to U-Boot master anyway [1], even
though there are concerns and ongoing discussion ... so I lost anyway.

I am _extremely_ disappointed !

[1] https://patchwork.ozlabs.org/patch/872169/

-- 
Best regards,
Marek Vasut