On Thu, Aug 28, 2014 at 02:29:46PM +0200, Michal Suchanek wrote: > On 28 August 2014 13:57, Luc Verhaegen <[email protected]> wrote: > > On Thu, Aug 28, 2014 at 10:07:46AM +0300, Simos Xenitellis wrote: > >> It is indeed bad blood. Let's see a few: > >> > >> 1. "*Allwinner <http://linux-sunxi.org/Allwinner> does not actively > >> participate in or support this community. In fact, it is violating the > >> GPLv2 license <http://linux-sunxi.org/GPL_Violations> in several ways and > >> has so far not shown willingness to resolve this."* > >> Source: http://linux-sunxi.org/Main_Page (it is in bold, on the first page > >> of the Wiki!) > > > > Allwinner never has participated or supported this community so far. It > > This is simply not true. > > Allwinner has provided devboards and probably some docs and sources > upon request.
Facts please. I know Oliver got sent a 50usd A33 tablet just now, even though he got told he would get an a23 tablet shortly before allwinner joined linaro. What else? > Of course, there is much left to be desired. Still this sentence alone > is false. Again, facts. We only get SDK sources (for some decreasing value thereof) from device makers. Which of the docs that we have came from allwinner directly, and didn't come from some other source? > As you might be aware they depend on the suppliers of the IP to > provide the drivers or documentation. In the industry GPL violation is > the standard. Often it is resolved not by opening the driver but by > making a properly licensed glue. As has been seen with Mali this can > somewhat help with porting as you can port the glue to newer kernel > and reuse the blob but it does not provide much advantage for writing > proper drivers. So by enforcing GPL you might get something that > adheres to the letter of copyright laws and licenses but does not help > you at all. Allwinner has that source. Allwinner also integrated that code into the kernel and u-boot trees. Board makers which get those SDKs don't get that code. Allwinner seems to be the point where the violation happens. Have you ever seen complete documentation? Have you ever seen code that tells you the whole story? You need both, but code gets you further, as you can keep it going. Going from incomplete/incorrect documentation alone very rarely leads to anything useful. When you have both, and they are contradictory, 9 out of 10 times, the code will be correct. Actually, in the RadeonHD project, ATI was forced to accept an open source driver and to give us docs. They gave us twice 500pages of "this register, this bit" and then told us "good luck!". If we hadn't also been able to disassemble ATOMBios and see how things fit together, and what information in the docs was contradictory with actual code, we wouldn't have brought up the hw in the record time we did, and ATI would've been able to tell AMD in september 2007: "We told you so, this open source stuff doesn't work". > You could argue that they could get IP for which sanely licensed > driver or documentation is available but then again if you do have > cost-effective providers of such IP please point them out. > > I can't tell much about the AW's willingness (or lack of) to resolve > the GPL issues but they are certainly not in the position in which > resolving the issues is simple. I believe that Allwinner can and will resolve this. I also know for a fact that GPLing some existing GPL violating code is not always possible. But there is a more than satisfactory way around that one. Luc Verhaegen. -- You received this message because you are subscribed to the Google Groups "linux-sunxi" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. For more options, visit https://groups.google.com/d/optout.
