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

Re: problem with /etc/kernel scripts

On Thu, 2011-03-31 at 10:59 +0200, Harald Dunkel wrote:
> Hash: SHA1
> On 03/21/11 14:22, Ben Hutchings wrote:
> > On Mon, 2011-03-21 at 06:17 +0100, Harald Dunkel wrote:
> >>
> >> Do you think that dkms should drop its /etc/kernel/postinst.d/dkms
> >> script? AFAICS this script might be called before the headers are
> >> installed.
> > 
> > It presumably should not rebuild modules unless both the image and
> > header packages are installed.  Depending on the order those packages
> > are installed, either hook may trigger this.  Also, it may be useful to
> > warn the user when only the image package is upgraded.
> > 
> Doesn't this mean that the modules are built twice?


> I don't see that the kernel itself would be necessary to build a kernel
> module.

It isn't.  However it is perhaps useful for dkms to warn you that your
OOT module will not be available for the new kernel.

> Of course I understand that this is highly complex. Maybe it would help to
> introduce some scheme to register event handlers to run after the installation
> of all packages has been completed?

We had an extensive discussion defining a policy for the kernel and
initramfs builder hooks only last year.  It's a little late to comment


Ben Hutchings
Once a job is fouled up, anything done to improve it makes it worse.

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

Reply to: