Subject: Re: Pegasos port freezes at uvm_km_alloc1
To: Frank Wille <>
From: Matt Thomas <>
List: port-powerpc
Date: 12/07/2004 13:51:19
At 01:13 PM 12/7/2004, Frank Wille wrote:
>Yesterday I started to port NetBSD to the Pegasos (II) hardware,
>based on the ofppc code.
>Status: I get the copyright-message on screen, UVM is initialized,
>but the system freezes at the end of uvm_km_alloc1(), while trying
>to clear the first page of kernel memory with memset().

Are the L1/L2/L3 caches enabled?  The heavy use dcbz with caches
disabled might cause that.

>I have experimented with different segments. Starting with the
>default USER_SR=11,KERNEL_SR=13,KERNEL2_SR=14 over KERNEL_SR=14,
>KERNEL2_SR=15(also tried 14) to USER_SR=5,KERNEL_SR=6,KERNEL2_SR=7.
>Always the same effect.

I would expect 5,6,7 to work fine.  What bats are you initializing?
I'm using 7, 8, and b on a private port.

>The memory layout is (according to OpenFirmware):
>00000000-10000000   256MB RAM
>80000000-c0000000   1st PCI region
>c0000000-e0000000   2nd PCI region (vga)
>According to PMAPDEBUG:
>vm_physmem[0]: 01800000-0fbfe000
>vm_physmem[1]: 0053d000-01000000
>vm_physmem[2]: 00240000-00400000
>The kernel was loaded to 0x400000-0x53d000.
>What could I check next? Any help is appreciated.
>BTW, is it possible to disable KERNEL2_SR? Why is it needed?
> From powerpc/oea/pmap.c line 3174 it looks like not-defining the
>macro would not set up the segment, but then there is also a
>direct reference to KERNEL2_SR in line 2024... :|

I don't think that not defining KERNEL2_SR will help.

Matt Thomas                     email:
3am Software Foundry              www:
Cupertino, CA              disclaimer: I avow all knowledge of this message.