Hi you all,

I'm getting a bit desillusionated...  I can't get NetBSD 1.4 working on my
RiscPC/ARM710.... I'm still running 1.3.2 now, and still experimenting to
find the fault (hacking in progresss :-)), but structured, but can't find
a real clue other than that there must be a problem with the actual
switching of the TTB... some oddities? somebody at ARM any HELP ??? 

Is there some documentation on the differences between ARM710a/ARM610 and
ARM710b ??? Is there a ARM710 rev. b ??

NetBSD 1.4 seems to crash directly after the setttb() function in the
second bootstrap, no mather what kernel boot-flag I set, no mather what
memory configuration...replacing the ARM610/ARM710a with a StrongARM has
proven to be a solution... but buying a StrongARM to solve the problem is
quite unsatisfactury and a bit problematic for me, far too expensive since
I'm still a PhD student.

I contacted Mark about the bug, but since he looks quite busy now and for
the academic interest.. i'm happyly hacking...

My main questions are now :
- does anybody know how to change border colour in just one/two
instructions ? (HW address?) to see how many instructions DO get fetched
before it dies..  it never get's to the printf() after the setttb :-) 

- could it be an undocumented `feature' ??... the ARM610 / ARM710a doc's
example of setting up the MMU is with the MMU __DISABLED__ ... this might
give a clue to the problem. But why does the first bootstrap succeed? Did
the actual physical page change on 1.3.2 on the second bootstrap? I can't
see any evidence in the 1.3.2 kernel sources that it's checked for, nor in
the 1.4; was it coincidence?

My ARM710's ID is : CPUID = 41047100... Could somebody with a ARM710 mail
me theirs? It helps in the search for clues!

Cheers,

Reinoud