Hi,

browsing around in McKusick's Red Book, I was once again annoyed about the
'traditional' approach that the *BSD ffs takes towards modern disk
hardware.

So I decided to give the Log-Structured Filesystem (LFS) a try... whipped
out some spare MO media, formatted it, partitioned it as A/UX USR, said
'newlfs -L /dev/rsd2g' and -- freeze. Ahh, the -current SCSI code... sbc in
this case. Pressing the debugger switch had no effect at all, so I rebooted
with an older (mid-May, ncrscsi) kernel.

I tried newlfs(8) again, this time with '-N' (tell me what you would do,
but don't do it yet), and it told me it didn't find a disklabel for my
disk.
Now 'disklabel sd2' reports some data, obviously faked by the kernel - does
this mean that newlfs(8) goes directly to the raw disk? Has anyone out
there given the LFS a try?



	hauke

---
"It's never straight up and down"     (DEVO)