On Thursday 11 August 2005 12:04, Väisänen Teemu wrote:
> Thanks Mikael!
>
> This is still a problem:
> > >>Problem is, that if I save .config -file in menuconfig/xconfig etc.,
> > >>lines which caused errors before, disappear. Without ARCH=um
> > >>-parameter they work fine.
> >
> > I _think_ that's because en uml guest kernel doesn't use those symbols
> > in the kernel and therefore they are not needed in the kernel.
>
> OK. So can I just compile kernel after
> # make xconfig ARCH=um
> with
> # make linux ARCH=um
> even there are those 'trying to assign nonexisting symbol' -messages?
You'd better remove .config, and then use "make defconfig ARCH=um", and then 
"make xconfig ARCH=um". The problem is that it is trying to read the 
configuration from /boot/config saved conf, which is a configuration for i386 
kernels, not for UML kernels.

> This gives me linux -file that is used to start UML?
It should do, but it'll have a strange configuration, and that would possibly 
cause many problems.
> Problem is also a program that 'needs' UML, because it's not very
> stable. Program also needs to patch the kernel. So if I try to to edit
> .config -files with ARCH=um, important lines (which program's patch
> adds to kernel's .config) disappear...

Well, the program's patch probably doesn't work very well for UML. Kernel 
modules which work for i386 may or may not work on UML, even if they are 
source modules. It depends on what they do.

In particular: the .config file is checked against Kconfig files, which 
specifiy possible configuration options. The program patch adds to i386 
Kconfig files some new option, which it doesn't add to Uml kconfig files 
(arch/um/Kconfig*), so make xconfig ARCH=um recognizes them as invalid.

In some cases, you simply need to copy the changes the program patch does to 
i386 Kconfig files to Uml Kconfig files (see 
Documentation/kbuild/kconfig-language for some clues). I'm not saying that's 
trivial for the casual user.

> A way to solve this, might be, is to compile UML-kernel, start UML and
> try to compile patched kernel there. What do you think about that
> solution? Is it possible?
No, it doesn't smell like it's useful.
> Could there be easier solutions? 
-- 
Inform me of my mistakes, so I can keep imitating Homer Simpson's "Doh!".
Paolo Giarrusso, aka Blaisorblade (Skype ID "PaoloGiarrusso", ICQ 215621894)
http://www.user-mode-linux.org/~blaisorblade


___________________________________
Yahoo! Messenger: chiamate gratuite in tutto il mondo
http://it.beta.messenger.yahoo.com



-------------------------------------------------------
SF.Net email is Sponsored by the Better Software Conference & EXPO
September 19-22, 2005 * San Francisco, CA * Development Lifecycle Practices
Agile & Plan-Driven Development * Managing Projects & Teams * Testing & QA
Security * Process Improvement & Measurement * http://www.sqe.com/bsce5sf
_______________________________________________
User-mode-linux-user mailing list
User-mode-linux-user@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/user-mode-linux-user

Reply via email to