Re: Rusty's module talk at the Kernel Summit

[Keith Owens <kaos@ocs.com.au>]

On Wed, 3 Jul 2002 08:53:03 -0700, 
"Adam J. Richter" <adam@yggdrasil.com> wrote:
>	I don't know enough about the formats of these tables right now
>to really understand the best way to handle them, but I suspect that
>the simplest approach might be a mechanism where copy_*_user and the like
>could generate assembler that does a .pushsection to a different section
>depending on the current section, so you could have "__ex_table" and
>".init.__ex_table", etc.

Unfortunately there is no way to get the current section name from
code.  I looked for one when I was trying to solve the __devexit
dangling pointer problem.

>	Come to think of it, if the core kernel's .text.init pages could
>later be vmalloc'ed for module .text section, then I think you may have
>found a potential kernel bug.

AFAICT that cannot occur.  The freed kernel .text.init pages are used
for kmalloc, not vmalloc.