Re: [PATCH 14/19] OpenRISC: Module support

[Arnd Bergmann <arnd@arndb.de>]

On Tuesday 21 June 2011, Valdis.Kletnieks@vt.edu wrote:
> On Sun, 19 Jun 2011 13:43:40 +0200, Jonas Bonn said:

> > diff --git a/arch/openrisc/kernel/module.c b/arch/openrisc/kernel/module.c
> > new file mode 100644
> > index 0000000..952b129
> > --- /dev/null
> > +++ b/arch/openrisc/kernel/module.c
> > @@ -0,0 +1,126 @@
> > +/*
> > + * OpenRISC module.c
> 
> Lose the filenames in the headers - you do this in multiple files.

Good point.
 
> Also, do you even *need* a module.c? There isn't much in here, will the main
> kernel/module.c do the work needed?  Or did you end up with one because the
> arch you cloned had one?

Right now you need a module.c in each architecture defining all these
functions. However, I think it would be cool if we didn't need this
all of them. In case of openrisc, the architecture should only have
to define a apply_relocate_add() function.

Jonas, could you add a default implementation for each of the functions
for include/linux/moduleloader.h, so that you only have to provide the
one function?

I think that an easy way to do that would be to add to kernel/module.c
code like:

#ifndef module_alloc
void *module_alloc(unsigned long size)
{
	return vmalloc(size);
}
#endif

Then each architecture that needs a special version can do
#define module_alloc(size) module_alloc(size)
in their respective asm/module.h

	Arnd