On 30/06/08 13:18 -0400, Joseph Smith wrote:
On Mon, 30 Jun 2008 18:28:37 +0200, Ward Vandewege ward@gnu.org wrote:
On Mon, Jun 30, 2008 at 10:18:47AM -0400, Joseph Smith wrote:
Maybe it is just me, but it seems like lately there have been a lot of posts to the mailing list about, coreboot not supporting ANY Intel chipsets/cpu's. Is there rumors spreading around the mill??? If so we
need
to nip this in the bud, this is probably discouraging people that want
to
get involved with coreboot.
Well - look at it from the other perspective. Intel's being difficult.
The
more obvious that becomes, the more they will be encouraged to change their ways and become more friendly and open.
I don't think spreading the word that coreboot does not support anything Intel is a positive way to accomplish anything. It just frustrates and discourages people willing to get involved. We have to face the facts here, Intel may never get involved with coreboot, so the next best thing is to get as many developers involved willing to take the time to fill in the missing pieces of code. Thinking ouside of the box. This is what I did for the i830, It might have taken me a little longer, but in the long run I did it without any help from Intel (except for one public datasheet).
Honesty is the best (and only) practice here. Firmware development is hard - exceedingly hard, and thats with copious amounts of documentation. Without documentation, it becomes nearly impossible. Somebody look at the Barcelona memory initialization code and tell me they could have figured it out on their own. I'm thinking not.
So we have to be blatantly obvious about what our shortcomings are. You can believe all you want in the power of open source, but the truth of the matter is that 90% of all the people who come on this list to ask for a particular port are not interested in doing the work themselves, and the precious few that are willing to do the work need to be told right up front what they are facing. They need to know what NDAs they need to secure, what code distribution rights they need to ask for, and just how difficult this work really is, especially in the cpu/* and northbridge/* directories.
Thats not to say that people shouldn't be encouraged to ask the right questions - if enough people ask a vendor for datasheets, it might have a positive effect. But acting as though we are not developing from behind the eight ball is going to give everybody the wrong impression, and it will result in even more upset feelings then if we just told them right off the bat that they are asking for something that probably isn't going to happen.
And just in case this sounds like I'm bashing the competition, know that my own company is in the same boat. We have yet to release the datasheet for the SB600, and until we do, I'm going to be the first to tell the honest truth if somebody asks for it.
Jordan