Re: -current on 7600-type machine

["Marco Trillo" <marcotrillo%gmail.com@localhost>]

Hi,

On 9/2/08, John Klos <john%ziaspace.com@localhost> wrote:
> Has anyone tried current (from today) on a 7600-type machine? For me, the
> kernel boots to where it configures mc0, then this:
>
> mc is not split
> mc0 at obio0 offset 0x11000trap: kernel ISI by 0x7c3142a4 (SRR1
> 0x40001032), lr: 0x484
> 778
> panic: trap
> Begin traceback...
> 0x006df670: at trap+0x100
> 0x006df700: kernel ISI trap by _prop_array_pool+0x7bcb1ae0:
> srr1=0x40001032
>              r1=0x6df7c0 cr=0x33a33e33 xer=0xe000be6f ctr=0x7c3142a6
> 0x006df7c0: at mc_attach+0x80

I think I found where the problem is.

Could you try the following patch for src/sys/arch/macppc/dev/if_mc.c ?

--- if_mc.c.orig        2008-09-02 01:33:10.000000000 +0200
+++ if_mc.c     2008-09-02 01:32:54.000000000 +0200
@@ -125,8 +125,9 @@

        sc->sc_txdma = mapiodev(reg[2], reg[3]);
        sc->sc_rxdma = mapiodev(reg[4], reg[5]);
+       sc->sc_regt = ca->ca_tag;
        bus_space_map(sc->sc_regt, reg[0], reg[1], 0, &sc->sc_regh);
-                                       /* XXX sc_regt is uninitialized */
+
        sc->sc_tail = 0;
        sc->sc_txdmacmd = dbdma_alloc(sizeof(dbdma_command_t) * 2);
        sc->sc_rxdmacmd = (void *)dbdma_alloc(sizeof(dbdma_command_t) * 8);


(I hope that this fixes it and I do not add more noise...)

Regards,
Marco.