On Tue, May 27, 2008 at 5:41 PM, Mart Raudsepp mart.raudsepp@artecdesign.ee wrote:
I don't quite remember what was wrong in my testings before. I just updated to HEAD, compiled and it worked as shown at http://www.pastebin.ca/1031398
it's got the alix1c memory settings in it from my port :-) The files from alix1c are the same.
Oh? I thought we got something based on the memory chip specs. Someone will have to work on making these optimal at some point anyway. I do have memtest errors in two specific memory addresses, which haven't surfaced with a known good coreboot-v2 based firmware, but it's still chugging along here with it.
I have no grafix from dbe62 at present. Don't know why.
How about coreinfo as payload with libpayload configured to Geode Video support? That's the only video output I've gotten as well. lxfb doesn't work with 2.6.25 (with any BIOS firmware) for me, and I haven't gotten to try 2.6.24 with v3 yet. Launching X with native driver (xf86-video-geode is the minimal that can work with coreboot) leads to kernel oopses as mentioned many times before, possibly involving mfpgt (or some other ordering of those letters :). Regular console obviously doesn't work, as there is no VGA BIOS running, and it seems there's quite some resistance to get that added in too anyway. All the other problems I've noted before (thread titled "DBE62 status") were noticed over a ssh session, with kernel spitting to serial console (where the oopses were visible).
Regards, Mart Raudsepp