[gentoo-dev] Help on adapting cman init scripts to kernels with things built in instead of modules

[gentoo-dev] Help on adapting cman init scripts to kernels with things built in instead of modules

[Pacho Ramos]

[-- Attachment #1: Type: text/plain, Size: 476 bytes --]

Hello

Looks like cman stabilization (that is needed to stabilize newer lvm2,
that is needed to stabilize newer udev...) is blocked by its init.d
script wanting to load modules even on kernels without modules:
https://bugs.gentoo.org/show_bug.cgi?id=442512#c5

Arch team people think that this should be handled before but... how
should it be handled? Do you know about some similar case that I could
look for possible solutions to it?

Thanks a lot for your help

[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 198 bytes --]

Re: [gentoo-dev] Help on adapting cman init scripts to kernels with things built in instead of modules

[Tony "Chainsaw" Vroon]

[-- Attachment #1: Type: text/plain, Size: 314 bytes --]

On Sun, 2012-12-02 at 23:10 +0100, Pacho Ramos wrote:
> Arch team people think that this should be handled before but... how
> should it be handled? 

I agree with the arch teams here. You can do something as mundane as:
if [ -e /proc/modules ]; then
    COMPLICATED MODULE MADNESS
fi

Regards,
Tony V.

[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 198 bytes --]

Re: [gentoo-dev] Help on adapting cman init scripts to kernels with things built in instead of modules

[Peter Stuge]

[-- Attachment #1: Type: text/plain, Size: 1222 bytes --]

Pacho Ramos wrote:
> Looks like cman stabilization (that is needed to stabilize newer lvm2,
> that is needed to stabilize newer udev...) is blocked by its init.d
> script wanting to load modules even on kernels without modules:
> https://bugs.gentoo.org/show_bug.cgi?id=442512#c5
> 
> Arch team people think that this should be handled before but... how
> should it be handled? Do you know about some similar case that I could
> look for possible solutions to it?

Don't know about similar cases.

Anyway, the init script seems to only have one real problem; it tries
to unconditionally load "dlm" as a module.

If configfs has been compiled-in, no module will be loaded.
If configfs has been mounted, the script doesn't try to mount it.

I don't know what the dlm module does, but the init script would have
to probe for it at runtime and only try to load a module when the
module isn't already available.

Of course it might be nice to also add some nicer error messages to
the init script. And probably there should be some pkg_postinst (or?)
einfo about how cman requires certain things enabled in the kernel.

There are several of those in the tree to look at for inspiration.


//Peter

[-- Attachment #2: Type: application/pgp-signature, Size: 190 bytes --]