Darren Steven (D.Steven@its.amc.edu.au) wrote:
: At 01:02 PM 11/09/95 -0600, you wrote:
: >
: >This is my first question to the group; please have mercy.
: >
: >I have installed NetBSD-1.0 and have created a new kernel as per the FAQ.
: ....
: >I guess my questions are:
: >
: >Do I need to use the new kernel with loadbsd to boot it ?
: 
: yes, sort of (see later)
: >
: >If so, how do I get it to the AmigaDOS partition ?
: >
: >If not, how do I get it loaded ?
: >
: >Is there a FAQ that addresses these issues that I should download ?
: >
: >Thanks for your patience...  Jack
: >
: Hi Jack,
: 
: I had the same problem (I also use NetBSD 1.0, from Meeting Pearls II CD).
: What you need to do is use BFF file system (available at all good NetBSD
: stores), which will allow you to mount your NetBSD partitions under
: AmigaDOS. You can then copy from the BSD partition to ADOS to allow 'loadbsd
: yourkernel' to work.
: 
: I have heard that BFFS should be only used read-only (don't write to it). I
: don't know if its true, but its better to be safe than sorry.

1.3 was read/write, but only knew about pre-4.4bsd filesystems; so it would
corrupt the NetBSD filesystems when trying to write. It succeeded to
read most of NetBSD filesystems, however; I always used it to loadbsd the
kernel.

1.4 is always forced readonly. I'm not sure it is guaranteed to read NetBSD
filesystems correctly (if the new features are used), but at least it
can't corrupt them.

I heard a rumour that 1.5 will support 4.4BSD == NetBSD filesystems
correctly read/write. Honestly, in this case I really don't know who is
working on it.

: An alternative/additional method (to avoid a complete halt/reboot
: ados/copy/start BSD) is to reboot within NetBSD. There is a device called
: /dev/reboot, and a reboot with the kernel of your choice can be accompished by:

No, it is called /dev/reload
: 
: cp /netbsd /dev/reboot 

cp /netbsd /dev/reload
: 
: This restarts BSD directly using that kernel image. /netbsd is your new
: kernel, which is copied to /netbsd. 

But this only works if the kernel hasn't changed too much in between;
especially the location of the kernel load support inside locore.s.

: 
: I don't know when it is safe to do this, as it seems fairly abrupt compared

It is, at least should be. The reload device initiates the reboot()
kernel call, which unmounts the file systems.

Regards,
	Ignatios Souvatzis