Hello Jakub,

>> So, i ran out of steam. Is here something wrong with MMU
>> initialization/activation sequence or similar?
JJ> As Martin already noted, it is hard to say. It may be just not ready for
JJ> ARMv5 or you introduced some problem by the initial page table changes.

The last sounds to be very logical, especially after studying of a few code 
pieces
from Linux kernel. Flushing of tables is prepended by flushing of instruction 
and
data caches, synchronized with memory barriers... And this all depends on 
particular
ARMv? architecture.

JJ> One thing which comes to mind is that on ARMv4, we maintain the
JJ> exception table at the high virtual address 0xffff0000 (we don't want to
JJ> map page 0 in the kernel). I don't remember if this is compatible with
JJ> ARMv5 or not.

Hm, Atmel says (doc6221.pdf) that address-block at 0xffff0000 is reserved.
Neither high vectors nor VINITHI is mentioned there. In Technical Reference
Manual (arm926ej-s, rev. r0p5) both are present.

JJ> Let us know if our suggestions were of any help. If not, we shall see
JJ> what to do then.

I have sent a patch to Martin, which allows running HelenOS on ARMv5 simulated
by QEMU. I will be trying to do a little bit debug work with the patch, too.

JJ> As always, I am curious about your motives to port HelenOS to a new ARM
JJ> device. Can you share a little bit with us?

I'm working as an application developer for linux-based devices now and i would
like to go one step back from "application", even to get deeper into "embedded".
I was looking around for an "thinner" OS than Linux in order to port it
on my arm-board with the goal to learn more about OS-Architecture, CPU-Hardware,
other tools  - just for fun (c) Linus :).

Thank you for reply.

-- 
Best regards,
 Alexander


_______________________________________________
HelenOS-devel mailing list
[email protected]
http://lists.modry.cz/cgi-bin/listinfo/helenos-devel

Reply via email to