On Sunday 13 July 2008, Ian Campbell wrote:
> > > What about retasking the new virtio udebs into virtuali{s,z}ion (or
> > > just virt) and adding them there?
> >
> > That could work, but I suspect it could result in a dependency mess.
> > You can either check /lib/modules/<ver>/modules.dep to see how bad it
> > is or just try it. If these modules do not depend on anything much
> > then this could be a good option.
>
> The xen-{blk,net}front.ko modules currently don't depend on any others
> so it should be fairly clean. What do you think?
I think the patch you have now is fine and this option has my vote. It is
IMO by far the simplest solution. As the modules don't exist for the 486
flavor, it does not add any overhead for regular images.
> I guess I'd need to add this udeb to the pkglist for the i386 netboot.
Why would we add it to the normal netboot? Is it going to be used there?
If not, we don't want to do that. Instead we should create a "subclass" in
pkg-lists and then set that "subclass" as TYPE in the config, comparable
to what has been done for the gtk variant of netboot.
> I've gone with EXTRANAME=netboot/xen. I decided that since non-PAE
> support is gone upsteam using e.g. xenpae would be redundant in the
> long run, so might as well drop it now.
Fine by me.
> The kernel-wedge patch is now trivial as mentioned above. I've included
> the base-installer (on top of #490542, called bigmem.patch) and
> linux-kernel-di...patch and the main installer.patch again.
I've not had any feedback on my patch yet. I'll try pinging a few people
on Monday, but I won't delay much longer.
> I will comment out netboot-xen from MEDIUM_SUPPORTED until the udebs
> are uploaded.
Fine.
> I've used the "set" trick which was recently removed from
> base-installer again. I could duplicate the echo 686 or add an ||
> 686-bigmem check to the 686 test if you would prefer.
Yes, that's what it was there for: to provide multiple levels of
fallbacks.
As far as I'm concerned you can start committing the changes, but you
should probably allow a few days for feedback from others.
Cheers,
FJP
Attachment:
signature.asc
Description: This is a digitally signed message part.