On Sun, May 04, 2008 at 12:34:42PM +0200, Peter Stuge wrote:
We are very, very close.
Dunno..
I'd even say we have arrived.
Strongly disagree.
Let's just say more work needs to be done. Polishing is really really really important. We want to be better than the proprietary alternative - and that means that all parts of a supported board need to work, and work properly.
With the right payload configuration,
Nono, "arrived" means there are zero reservations.
Agreed.
I spent a good hour creating a v3 that did nothing at all on my alix1c. Turned out to be a VSA blob formatting problem. It still can not use the CF.
I'm not seeing all these issues?
Neither buildrom nor v3 actually include VSA blob and payload.elf when I run make. I forget where the zerofill option was, but that wasn't done either.
.... I don't know what the problem is you are seeing, but buildrom (v3) works just fine for me. It generates an alix.1c image with filo payload and VSA blob, which boots perfectly from CF (with the dongle). All built on Ubuntu Gutsy 32 bit.
Config and generated image at
http://ward.vandewege.net/coreboot/alix.1c-v3-working/
Toolchain issues? Your CF boots under the proprietary bios right? Are there multiple alix.1c revisions?
Oh and stuffing everything into build/ feels so unusual. I go looking for lar in all the wrong places (util) but that's just me.
I'm with you there :) It's weird and a bit annoying.
these systems work fine (unless you want to use those pesky ISDN cards with horrible drivers). To be honest, I'd say we sidestep the ISDN issue completely by using any other miniPCI card for demos.
Yes definately. But please keep in mind that the product is not done until it has been so polished that users mistake it for a mirror. ;)
100% agreed. By that reasoning, I don't think we have a single board that is truly 'finished'.
Thanks, Ward.