Re: All PPC ports broken since 5.99.61

[Frank Wille <frank%phoenix.owl.de@localhost>]

On Mon, 30 Jan 2012 16:34:33 -0800
Matt Thomas <matt%3am-software.com@localhost> wrote:

> > They hang early, directly after cpu_startup(), probably in sysctl_init
> > ().
> > [...]
> > I guess it has something to do with the major UVM/kmem/vmem changes?
> 
> Yep.  An else clause was missing.  It's now fixed.

Thanks. Now the kernel works until it tries to mount the filesystem.

[...]
boot device: wd0
root on wd0a dumps on wd0b
root file system type: ext2fs
warning: lookup /dev/console: error 13
exec /sbin/init: error 13
[...]

That's a bit strange, because I have a working ffs filesystem, and with
a recent kernel it insists on ext2fs? This is on sandpoint.

On ofppc (Pegasos2) it is even worse. The problem with the filesystem
is similar (/dev/console should exist), but is followed by a crash:

[...]
root file system type: ffs
warning: no /dev/console
trap: kernel read DSI trap @ 0xa00011c8 by 0x3d5dec (DSISR 0x40000000, err=14), 
lr 0x3d5dd8
Press a key to panic

FYI, 0xa and 0xb are used for the kernel segments under ofppc.
Here is the stack frame:

vpanic+0x21c
panic+0x4c
trap+0x124
kernel DSI read trap @ 0xa00011c8 by pool_cache_get_paddr+0x4c: srr1=0x9032
r1=0xa00abaa0 cr=0x28284084 xer=0x20000000 ctr=0x164a4c dsisr=0x40000000
copyright+0x126a4
kmem_intr_alloc+0x88
exec_elf32_makecmds+0x148
check_exec+0x240
execve1+0x164
start_init+0x2b8
setfunc_trampoline+0x8

I fear this is another problem caused by the recent uvm/kmem changes...

-- 
Frank Wille