On Tue, Jun 10, 2008 at 11:27:39AM -0400, erik quanstrom wrote: > > yes, but if you need the keyboard to enter stuff during boot time, like > > the authid, or the path to the filesystem, then it's not possible > > because I think those choices happen before the point where those > > commands can be issued, right? Unless there is some place other than > > cpurc where I could set those commands? > > yes, if that is the case, you're in a pickle. > > however, if you have a place for local nvram and you make sure > your plan9.ini file (or pxe version thereof) specifies the path to > the fs, you can still boot to a native usb kbd. > > if you can't resolve one of those issues, you will likely need to add > some code to the init process and link usb/usbd and usb/kb into > the kernel. > > - or - if you have working ethernet, the cec patches that i submitted > a year or so ago will allow you to enter very early boot information > using cec(8). i also have a prototype 9load that will also interact > via cec. > > an older version of cec is part of the distribution. my current > version is in /n/sources/contrib/quanstro/src/cec. i'd be happy to > put the kernel patches somewhere if folks would find it useful.
Well in my case I am not stuck anymore since I could deal with the nvram stuff with a usbohci-free kernel. And yeah, I can go back to the regular kernel if I set the fs path in plan9.ini without the query for it. I had intially mailed so that if the behaviour qualified as a bug, it would be reported, and if someone ever stumbled upon the same thing he could find the quick fix (remove usbohci) if he/she cared enough to digg in 9fans archives. Now your solutions are archived as well if needed. ☺ Cheers, Mathieu.
