TO DO LIST
- official machine number DONE Machine ID = 597
- configuration option
- default config for slug that enables all the required drivers
so someone can grab kernel.org, do a 'make nslu2_defconfig && make zImage' and have a bootable kernel :)
[22:56] <dsaxena> so back to my original reason for coming by tonight...where's the board port for going upstream?
[22:56] <dsaxena> :)
[22:56] <jacques> what needs to be done ?
[22:56] <dsaxena> or does stock 2.6.9 just work on slug?
[22:57] <dsaxena> would be nice to have:
[22:57] <dsaxena> - official machine number
[22:57] <dsaxena> - configuration option
[22:57] <dsaxena> - default config for slug that enables all the required drivers
[22:58] <dsaxena> so someone can grab kernel.org, do a 'make nslu2_defconfig && make zImage' and have a bootable kernel :)
[22:58] <jacques> thanks for the todo - we definitely want to do that
[23:00] <jacques> for the official machine number to work, we would have to modify redboot
[23:00] * ka6sox-away puts it in the wiki
[23:00] <jacques> thanks ka6sox-away
[23:00] <dsaxena> jacques: yeah
[23:00] <dsaxena> but should be a 1 line change
[23:01] <dyoung> Oh yeah, thats what I've been meaning to ask. so the stock redboot is passing machine id 245. Do we need to change that to 597 (nslu2) at some point?
[23:01] <dsaxena> yes
[23:02] <jacques> so we already have a machine number! cool check that one off the list :-D
[23:02] <dyoung> Will booting break if someone takes a changed redboot and tries to boot a stock linksys/kernel ?
[23:02] <jacques> yep
[23:02] <dyoung> (thank mrakes, he did that a while back)
[23:02] <dsaxena> hmm..that's a problem
[23:02] <jacques> the machine number the bootloader sets must match the kernel binary
[23:03] <dsaxena> actually, let's rewind back a bit
[23:03] <dsaxena> does the nslu2 use the same pci irq routing, gpio layout for i2c, etc as IXDP425? [23:03] * dsaxena not sure if folks would know that
[23:04] * dsaxena grabs openslug
[23:04] <jacques> there are some ppl here who are very familiar with the gpio layout
[23:04] <dyoung> the GPIO mapping is a bit differnt for power, etc.
[23:04] <dyoung> LED's what have you.
[23:04] <dsaxena> s/6/5
[23:04] <dyoung> kas11 wrote a really pretty driver for the GPIO bits.
[23:05] <jacques> this is always the problem when a manufacturer uses a dev platform machine number for a production device
[23:05] <jacques> had same problem with OSK
[23:05] <ka6sox-away> dsaxena, http://www.nslu2-linux.org/wiki/Info/GPIOConnections
[23:06] <jacques> I'm trying to think of all the things we had to change in 2.6.9 - gpio mapping, memory map, ...
[23:07] <dsaxena> yeah, different then dev board. the dev board has 4 pci irqs instead of 3