[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: powermac boot-floppy patches

On Fri, Jun 01, 2001 at 03:07:25PM -0800, Ethan Benson wrote:
> On Fri, Jun 01, 2001 at 10:09:52AM -0500, Stephen R Marenka wrote:
> > On Fri, Jun 01, 2001 at 12:25:15AM -0800, Ethan Benson wrote:
> > > On Thu, May 31, 2001 at 09:02:02PM -0500, Stephen R Marenka wrote:
> > > 
> > > i changed around the tests some.. i think this needs to be done
> > > differently anyway.  
> > 
> > I think I'll go back to putting the checks in
> > check_for_native_partition, then we can keep them in a loop (if they
> > want) until they get it right. The only problem is it's called for each
> > disk being partitioned. But hey, worst case they get multiple warnings.
> hmm.  well i don't want to enforce creation of a bootstrap partition
> on every disk, its only necessary for the one containing the root
> partition (since thats the only one that will be bootable anyway).  

I understand that, the problem with check_for_native_partition is that
it is called after any disk is partitioned, so I'm not sure there's a
good way to check for root and Apple_Bootstrap on the same disk. That
may have to be checked when root is mounted. That's why I abandoned
check_for_native partition originally. What I was thinking about was
checking for the existence of an Apple_Bootstrap partition on any disk
after any disk is formatted. If none exists, then raise an alarm
(complete with continue anyway, let me be an idiot).

> there should probably be a `i know its wrong, let me be an idiot
> anyway' switch to bypass the after partitioning check.  we should
> leave the tests in run_yaboot() as they are to ensure it will fail on
> bad partitioning, there are no exceptions there.  

Yes, that's what I was thinking too.

> > I was thinking about breaking it out by model, but if you don't think
> > it's worth it, then I happily won't do it.
> that would be fine if you broke it down by model, if you want to do
> the work...  testing all models to see which can support what is a lot
> of work. 

I was thinking more on relying on existing documentation and mail
archives. I only have a 7500 and a 7600 available for testing.

> > I re-worked your readlink code. readlink doesn't append a null or
> > include the initial /. I just sorta hacked it in, so feel free to blow
> > it away. I don't get what the strrchr bit is about, but maybe more
> > coffee will help.
> i think it takes the symlink and strips all leading directorys from
> it, so a symlink to boot/vmlinux-foo becomes a symlink to simply
> vmlinux-foo.  

Oh right. Since my patch added a preceding /, you might not want that
bit. Although, I'm not still not clear why you'd want that symlink.
Doesn't it depend on your current working directory? What am I missing?

> i am not seeing any change to that right now though.. you sure you
> commited that change?

My screen still says bootconfig.c new version 1.98. Hmmmm. Maybe I'll do
a fresh checkout to see what's going on.

Stephen R. Marenka     If life's not fun, you're not doing it right!

Reply to: