Michael Holzheu wrote:
anderson(a)redhat.com wrote on 08/14/2006 07:52:40 PM:
>
> One thing that I noticed, and that I've added to your patch, is this:
>
> diff -r1.264 defs.h
> 2410c2410
> < #define IS_VMALLOC_ADDR(X) (vt->vmalloc_start && (ulong)(X) >=
> vt->vmalloc_start)
> ---
> > #define IS_VMALLOC_ADDR(X) s390x_IS_VMALLOC_ADDR(X)
>
> It seems that s390x_IS_VMALLOC_ADDR() was created but never
> put in place for IS_VMALLOC_ADDR() to use.
>
> Let me know if that's not correct.
>
At least that is not wrong. But I noticed that we do not need
s390(x)_IS_VMALLOC_ADDR() at all.
In memory.c/vm_init(), we set vt->vmalloc_start to:
vt->vmalloc_start = machdep->vmalloc_start();
which leads to a call of s390(x)_vmalloc_start().
Therefore, we can use that for IS_VMALLOC_ADDR() like other architectures
do. I attached a patch, which removes s390(x)_IS_VMALLOC_ADDR().
Michael
Ok fine -- all checked in...
Thanks,
Dave