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

Re: [PATCH/RFC] Re: Kernel selection on i386 for Xen compatibility



On Wed, 2008-05-07 at 23:36 +0200, Frans Pop wrote:
> On Wednesday 07 May 2008, Ian Campbell wrote:
> > On Wed, 2008-05-07 at 15:35 +0200, Frans Pop wrote:
> > > Please submit the final patch as a wishlist BR against base-installer
> > > with a link to this discussion in the mailing list archives.
> >
> > Done: http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=480054
> 
> Thanks. Given how close we are to Beta2, I think committing this will have 
> to wait until after.

That's fine with me.

> On Wednesday 07 May 2008, Ian Campbell wrote:
> > I assume that d-i won't be switching to a 2.6.25 kernel until after the
> > next beta, is that right?
> 
> Correct.
> 
> > Shall I file a wishlist bug now for the creation of the -bigmem kernel
> > udebs when the switch happens or shall I wait until it is being planned?
> 
> If you can create a patch, then feel free to file the BR. If not, there's 
> not much point.

Well, my current patch is simply:
        Index: kernel-versions
        ===================================================================
        --- kernel-versions	(revision 53021)
        +++ kernel-versions	(working copy)
        @@ -1,2 +1,3 @@
         # arch   version  flavour       installedname        suffix build-depends
        -i386     2.6.24-1 486           2.6.24-1-486         -      linux-image-2.6.24-1-486
        +i386     2.6.25-1 486           2.6.25-1-486         -      linux-image-2.6.25-1-486
        +i386     2.6.25-1 686-bigmem    2.6.25-1-686-bigmem  -      linux-image-2.6.25-1-686-bigmem
(plus a hacked upgrade to kernel wedge for 2.6.25)

I could file the equivalent change but without the 24->25 bit I suppose
or I could generate what I know the patch will look like once the 24->25
change has been made.

Probably I should just wait till after beta2 and keep an eye out for the
early stages of the change.

> For a potential patch. It could be worth trying if just creating a symlink 
> 686-bigmem -> 486 in the modules directory will work. If it does that would 
> reduce future maintenance overhead.

It certainly won't work to try loading 486 modules into a 686-bigmem
kernel but I don't think that is what you meant so I'm not sure I
understand -- where are you proposing I use a symlink?.

> Of course it won't work if you need modules we currently don't include.

Everything seems fine so far 
-- 
Ian Campbell

A gentleman is a man who wouldn't hit a lady with his hat on.
		-- Evan Esar
		[ And why not?  For why does she have his hat on?  Ed.]

Attachment: signature.asc
Description: This is a digitally signed message part


Reply to: