Re: init_FIQ

On Thu, Jun 03, 2004 at 09:28:55AM +0900, A Matsui wrote:
> I got the reply that the CPU supports high vector as Mr. King and
> Mr. Plvarez(I'm not sure the spelling. Sorry for the wrong spelling.)
> said. But the CPU allocates original registers from 0xFFFF0000, then
> some settings should be done before using high vector.
> 
> 1. Set 1 to V bit(bit 13) of the CP15. (I think this has been done
>    already.)
> 
> 2. Allocate the logical address area from 0xFFFF0000-0xFFFF0FFF to
>    physical address where the vector tabel is assined to using MMU.
> 
> 3. Set to VIC(Vector Interrupt Controller) where is allocated to
>    0xFFFFF000(VICbase).

I _really_ do not understand what your problem is.  The kernel supports
high vectors and they *WORK*.  We do all the correct setup in the right
order across all CPUs that support it.

As I've already said, if you're seeing crashes while _READING_ a memory
location which should exist, then that is _NOT_ a high vector problem.
The possibility is that you either overwrote the mapping for 0xffff0000,
or have disabled access to that memory in some way.

Unfortunately, without seeing the modifications you've done to the kernel
there's no way for anyone here to diagnose your problem.  However, since
you're working on open source, there's nothing to prevent you from posting
a patch showing all your modifications to date.

So please get off the high vector whinge train and start thinking about
what's going on.  Otherwise I'm just going to ignore your future postings
on this subject.

-------------------------------------------------------------------
Subscription options: http://lists.arm.linux.org.uk/mailman/listinfo/linux-arm-kernel
FAQ:       http://www.arm.linux.org.uk/armlinux/mlfaq.php
Etiquette: http://www.arm.linux.org.uk/armlinux/mletiquette.php