Sunday, 13th March 2005

We build this city on rock and roll
— Jefferson Starship (circa 1978), San Francisco.

Got a certain board booting semi-properly last night. At least it listens on a serial boot loader console and takes commands. I can also build the same firmware the others were building, so I think my toolchain is working ok (although I am currently rebuilding a powerpc GNU toolchain as the GNU ld in ecos gnutools is horribly broken for me – no, bfd, it’s ok to have a “huge negative offset” if it’s your fscking boot vector…it won’t let me objcopy ELF binaries in to binary images). Inadvertly discovered a Redhat 6.2 box I’d forgotten about at home (we have enough old machines kicking about here) which was still handing out DHCP addresses to boxes upstairs and devkits. Moved all the DHCP lease serving over to another Debian box.

Read bits of last month’s LJ finally. I was annoyed at the time to see they’d also covered Netlink at the same time that we just did in LU&D – but meh, it’s popular. Read rml’s article on D-BUS and found it useful, need to play with that some more since I’m meant to be doing this pythony thing at the moment. Almost got chance to read a bit of LDD3 again, maybe tonight.

My eyes are playing up a lot at the moment, and I’m feeling rundown on occasion, but I just need to sort stuff out, I hope. Like, getting in to a regular routine, getting up in the morning, etc. Will try to set a really noisey alarm for the morning and get up super early so I can get work done before it’s too late. I’m too lazy generally, and I am trying to fix that now. My new year’s resolution was to do something about it, I am really trying. Today I must get somewhere with this board port and finish some of the uC stuff that’s been dragging out. Hopefully I’ll sort out visiting those folks in Sunnyvale tomorrow.

What happened on the ppc405 bi_recs front? Well, I’m in a mail thread about it right now. I have started to shove stuff in to my tree, but I’m lazy and need to figure stuff out too. The plan is to have this fake OF tree BLOB sitting somewhere in memory (shoved there by u-boot) and tell Linux about it through bd_info->bi_fakeoftree. Then I need to get this unflattened a la the ppc64 tree and tell the kernel to make it available as the OF tree (userspace visable via /proc/device-tree) so modules can find out stuff from it.


Comments are closed.