sparc64/144867: panic: trap: memory address not aligned booting
Sunfire 280R
Marius Strobl
marius at alchemy.franken.de
Tue Mar 30 17:33:40 UTC 2010
On Mon, Mar 29, 2010 at 11:13:47PM +0100, Craig Butler wrote:
>
>
> On Mon, 2010-03-29 at 22:00 +0000, Rob Farmer wrote:
> > The following reply was made to PR sparc64/144867; it has been noted by GNATS.
> >
> > cas0: <Sun Cassini+ Gigabit Ethernet> mem 0x200000-0x3fffff at device
> > 1.0 on pci1
> > panic: trap: memory address not aligned
> > cpuid = 0
> > KDB: enter: panic
> > [ thread pid 0 tid 100000 ]
> > Stopped at kdb_enter+0x80: ta %xcc, 1
> > db> bt
> > Tracing pid 0 tid 100000 td 0xc0a0e330
> > panic() at panic+0x198
> > trap() at trap+0x220
> > -- memory address not aligned sfar=0x7fd00200001 sfsr=0x1d007d %o7=0xc016f1c8 --
> > cas_pci_attach() at cas_pci_attach+0xb1c
> > device_attach() at device_attach+0x68
> > bus_generic_attach() at bus_generic_attach+0x10
> > ofw_pcibus_attach() at ofw_pcibus_attach+0x2e4
> > device_attach() at device_attach+0x68
> > bus_generic_attach() at bus_generic_attach+0x10
> > schizo_attach() at schizo_attach+0xe04
> > device_attach() at device_attach+0x68
> > bus_generic_new_pass() at bus_generic_new_pass+0x130
> > bus_generic_new_pass() at bus_generic_new_pass+0xbc
> > bus_set_pass() at bus_set_pass+0xc4
> > configure() at configure+0x4
> > mi_startup() at mi_startup+0xf0
> > btext() at btext+0x2c
> > db>
> >
> > --
> > Rob Farmer
> > >
> > > Marius
> > >
>
> cassini memory address trap.... same as on the b100s in the B1600
> chassis I think -- I bet if you compile a kernel without device cas
> it'll get past that bit -- problem then is no network...
Yes, at this point it appears to be the same problem as seen with
the on-board NIC of b100.
>
> Marius- I'll have a look at setting up a netboot server and look at
> getting terminal and blade access for you this weekend if your still
> interested.
That would be helpful.
> I hope to help in anyway possible but me head exploded
> after trying to debug the panic :)
>
> That is if they are related... IIRC you said something to do with the
> way the BAR was getting assigned.
>
Treating BARs containing 0 as invalid was another issue preventing
the on-board NICs of b100 from working but that has been fixed and
now works as seen with other hardware.
Marius
More information about the freebsd-sparc64
mailing list