sm-discuss AT lists.ibiblio.org
Subject: Public SourceMage Discussion List
List archive
- From: neuron <vlmarek AT volny.cz>
- To: sm-discuss AT lists.ibiblio.org
- Subject: Re: [SM-Discuss] linux-new call
- Date: Tue, 31 May 2005 08:10:29 +0200
> > 2) I know what you mean about the kernel modules not compiling against
> > the right
> > kernel this is a problem that is greater than the linux spell, the only
> > thing
> > that can be done is go to each spell that compiles an external kernel
> > module
> > and make it use installed_version linux to compile against, I'll be
> > working on
> > converting the kernels and mobile sections over the summer to use the
> > version
> > returned by sorcery instead of uname (which is what most of the spells
> > use)
> >
> > I know your frustration with this problem but this is the only real
> > solution
> > for fixing it. And *nothing* can be done in the linux spell to fix this.
>
> I'll make a shameless plug for a small piece of code i wrote for an iso
> generation experiment to intercept the uname system call (attached). If
> interested you could just roll this into a spell, and then in the module
> spells just preload the library. Then anything and everything that uses
> uname, either through the uname command or the system call will get our
> custom set values.
Sounds a bit scary, but seems to do the job !
> > >I can see two possibilities, either remember somewhere that the modules
> > >should be cast later ( install queue ? ) or force their compilation to
> > >use different install directory. In either case, I'm not sure how to do
> > >that :)
> >
> > The second part of this is what my solution is basically...
> >
> > >The second thing is that once the module is recompilled, it dispells old
> > >version, which makes old kernel un-usable. If I am going to reboot to
> > >older kernel, I have to recast theese modules again. I would like to
> > >have the modules stay on the system for each kernel I compiled them.
> >
> > A solution would be to have sorcery ignore dispel in /lib/modules
> Those spells need to simply add a protected file.
Maybe I'm blind, but this is allready working feature ? I haven't found
anything near 'protected' or 'filter' in libapi.
Thank you
--
neuron
Attachment:
pgpCMfjY7pSiL.pgp
Description: PGP signature
-
[SM-Discuss] linux-new call,
Eric Sandall, 05/30/2005
- <Possible follow-up(s)>
-
RE: [SM-Discuss] linux-new call,
David Michael Leo Brown Jr., 05/30/2005
- Re: [SM-Discuss] linux-new call, Duane Malcolm, 05/30/2005
- Re: [SM-Discuss] linux-new call, neuron, 05/30/2005
-
Re: [SM-Discuss] linux-new call,
David Michael Leo Brown Jr., 05/30/2005
- Re: [SM-Discuss] linux-new call, Sergey A. Lipnevich, 05/30/2005
-
Re: [SM-Discuss] linux-new call,
David Michael Leo Brown Jr., 05/30/2005
-
Re: [SM-Discuss] linux-new call,
Andrew, 05/30/2005
-
Re: [SM-Discuss] linux-new call,
neuron, 05/31/2005
-
Re: [SM-Discuss] linux-new call,
Andrew, 05/31/2005
- Re: [SM-Discuss] linux-new call, neuron, 05/31/2005
-
Re: [SM-Discuss] linux-new call,
Andrew, 05/31/2005
-
Re: [SM-Discuss] linux-new call,
neuron, 05/31/2005
- Re: [SM-Discuss] linux-new call, neuron, 05/31/2005
-
Re: [SM-Discuss] linux-new call,
Andrew, 05/30/2005
-
Re: [SM-Discuss] linux-new call,
David Michael Leo Brown Jr., 05/30/2005
-
Re: [SM-Discuss] linux-new call,
Sergey A. Lipnevich, 05/30/2005
- Re: [SM-Discuss] linux-new call, neuron, 05/31/2005
-
Re: [SM-Discuss] linux-new call,
Sergey A. Lipnevich, 05/30/2005
-
Re: [SM-Discuss] linux-new call,
David Michael Leo Brown Jr., 05/31/2005
- Re: [SM-Discuss] linux-new call, Seth Alan Woolley, 05/31/2005
- Re: [SM-Discuss] linux-new call, Sergey A. Lipnevich, 05/31/2005
- Re: [SM-Discuss] linux-new call, David Michael Leo Brown Jr., 05/31/2005
Archive powered by MHonArc 2.6.24.