Re: A few issues while playing with virt68k
[email protected] (Michael van Elst) Sat, 16 Mar 2024 06:07:31 -0000 (UTC)
| Newsgroups | gmane.os.netbsd.ports.m68k |
|---|---|
| Organization | Serpens User Group |
| Message-ID | <[email protected]> |
[email protected] (John Klos) writes: > -I If a label cannot be read from disk request the default one from > the kernel. Implies -r. My method to write a disklabel from the host system is: $rdbedit -F -t "$image" \ | disklabel -w -F -f /dev/fd/0 "$image" hardfile "$product" $rdbedit dumps a disktab(4) stanza (named 'hardfile') and disklabel -f /dev/fd/0 reads that stanza from stdin. The example generates a label for the host. If you need a label for the target, you need to pass things like machine type, endianess or the exact label sector number. >> -append ?root=ld0a? >> On my qemu command line and it boots right up without prompting: >Right now I'm using GPT wedges, so I'll give it another try with >disklabel. Perhaps a properly disklabeled disk will boot automatically. This is yet another ad-hoc method to specify the boot device. - strip non-digits from the end - match remaining prefix against a device name. You can use something like 'dk1', but for the volatile dk unit numbers it's of limited use. MD code _could_ just set the variable "bootspec" from the command line argument and MI code would do the right thing. evbarm code does that (after trying to identify a device itself).