[caiman-discuss] Code review for bug 434

[Anil Gulecha]

On 1/30/08, Dave Miner <Dave.Miner at sun.com> wrote:
> Anil Gulecha wrote:
> > Hi,
> >
> > I've modified the live-fs-root file in the slim_prototype repo to be
> > able to boot usb without checking for /.liveusb. This will allow the
> > same minirot to be used for both cd and usb.
> >
> > The bug is filed at http://defect.opensolaris.org/bz/show_bug.cgi?id=434
> > and code review at http://cr.opensolaris.org/~anilg/001_fsroot/
> >
>
> 191: One thing that I think has always been wrong here is that it may
> potentially find a different UFS file system than the one it's expecting
> on some other USB device.  I seem to recall having made this happen with
> my external USB hard drive plugged in during boot from a stick.  I think
> we should look for a signature right after mount (perhaps the presence
> of solaris.zlib) in order to ensure we found the right device.
>

This should not be hard to implement.. a simple [ -f
/.cdrom/solaris.zlib].. I'll try this out and add to the next review.

> 229,232: I know this was here previously, but we really should be using
> the device output from lofiadm -a rather than hard-coding these.
 This too.

>
> 235: We should just set a variable rather than touching in the file
> system.  There's nothing else that references this outside this script.
>

Actually I make use of these in the session persistace during shutdown
to check for liveUSB mode. I was having trouble getting $LIVEUSB to
work.. it wasnt being permanently set system wide.

So, I have a question.. if I set $LIVEUSB=1 and export it.. will this
variable be available all around during the session?

Anil