Marco d'Itri wrote: > IIRC 2.4 kernels used c0d0p1 and 2.6 kernels used the other format or > something like this, so who initially contributed the script choose the > old format since it would be more compatible. > As an extra data point, I have a SuSE system installed a few years ago > with static c0d0p1-like devices.
I've checked 2.4 and it uses disc/part0 devices with devfs for this device. > Does d-i really really need another script to generate disc/part > devices? They are hard to manage (creation requires locking) and I'd > rather not spend time trying to support an highly deprecated interfaces. I don't understand, since raid-devfs.sh shows every indication of creating devices in disc/part format: case "$NAME" in *p*) LONG_NAME=disc$(get_dev_number ${NAME%%p*})/part${NAME##*p} ;; *) LONG_NAME=disc$(get_dev_number $NAME)/disc ;; esac It just doesn't seem to be working in d-i. -- see shy jo
signature.asc
Description: Digital signature