Actually, when you look at the scsi_ofpath() function that's currently
in the "ofpath" script, here is how the other options were coded --
aic7xxx)
HOST_LIST="$(for i in `find /proc/device-tree -name compatible` ; do
lgrep "$i" "^ADPT" "^pci900[45]" "^pciclass,01000" ;
done)"
(...)What I added (the sbp2 option) was very similar. Doesn't the lgrep() function in ofpath return on the first successful match? The particular file in the device tree that I needed to match was this one -- me: $ cat -v [EMAIL PROTECTED]/[EMAIL PROTECTED]/compatible pci106b,[EMAIL PROTECTED],[EMAIL PROTECTED],0c0010^@
me: $ hexdump -c /proc/device-tree/[EMAIL PROTECTED]/[EMAIL PROTECTED]/compatible 0000000 p c i 1 0 6 b , 3 0 \0 p c i 1 0 0000010 6 b , 3 0 \0 p c i c l a s s , 0 0000020 c 0 0 1 0 \0
I didn't know that the null delimiters worked with (busybox) grep, but that's better left to the experts (i.e. why I didn't anchor in my attempt).
By the way, I did get my system fixed/booted up, what I did was to re-run the entire installation, suspending (entering shell) the installer just after it unpacked the yaboot stuff, but just prior to it trying to install in /dev/sda2 ... hand-edited ofpath, umounted some of the target fs'es (including /proc), chroot'ed to /target, "mount -a", then rebuilt initrd, and removed/remade the initrd.img symlink. Exited back to the installer, told it to install yaboot, finished up and this system is working great. (The time before that, I told the installer to update the kernel-image to a newer one, and in that case I had a hard time making a new initrd ... so the final time I did this sequence, I left the kernel at the same level that the installer image was running.) Thanks to those to posted the good hints (Joerg in 288010); regards, Larry.
Ethan wrote:
are you sure that 0c0010 is not specific to your particular hardware? I suspect this patch would work nowhere else.
-- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
