Installing on zx6000

Marcel Moolenaar marcel at xcllnt.net
Mon Jul 2 15:37:49 UTC 2012


On Jul 2, 2012, at 5:41 AM, Anton Shterenlikht wrote:
> 
> Hit [Enter] to boot immediately, or any other key for command prompt.
> Booting [/boot/kernel/kernel]...
> Entering /boot/kernel/kernel at 0x9ffc000000010500...
> ***********************************************************
> * ROM Version : 02.31
> * ROM Date    : 03/11/2004
> * BMC Version :  01.53
> ***********************************************************

I can possibly explain this one, having bitten by it myself:

Starting with loader 3.0, we support having the kernel linked
against what I call "Pre-Boot Virtual Memory" rather than
having it linked against a 1:1 mapping. This to support the
SGI Altix 350.

There's a bug or omission in the kernel with respect to this.
The kernel increases "kernend" to accomodate some internal
(dynamic) structure. Since "kernend" is in the PBVM, it is
possible for it to cross a page boundary. If that happens,
the kernel will effectively write to unmapped memory. The
kernel needs to make sure that if it increases the size of
the PBVM, there's a mapping in the page table for it.

If this is indeed the bug, then all you need to do is pre-
load some module. It doesn't matter which one. This has the
effect that "kernend" moves into the next 64K page and you
don't hit the bug.


> uhub1: 2 ports with 2 removable, self powered
> uhub0: 3 ports with 3 removable, self powered
> uhub2: 5 ports with 5 removable, self powered
> da0 at mpt0 bus 0 scbus1 target 0 lun 0
> ***********************************************************
> * ROM Version : 02.31
> * ROM Date    : 03/11/2004
> * BMC Version :  01.53
> ***********************************************************

This one could be related to the USB stack. If it is, then it
means you have a CD in the DVD drive and all you need to do
is eject the disc. If you don't have a disc in the CD drive
then it must be something else, and I don't know what it could
be then.

I've tried to understand what is causing the USB problem, but
so far no luck. I think it started with the new USB stack and
my best guess at this time is that it has something to do with
DMA.

HTH,

-- 
Marcel Moolenaar
marcel at xcllnt.net




More information about the freebsd-ia64 mailing list