[Bug 238730] r349985 on ppc64 IBM 970MP PowerMac G5 possible panic in bus_generic_new_pass

[bugzilla-noreply at freebsd.org]

https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=238730

Dennis Clarke <dclarke at blastwave.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
            Summary|r349241 on ppc64 IBM 970    |r349985 on ppc64 IBM 970MP
                   |instant panic in            |PowerMac G5 possible panic
                   |vm_mem_init(0)              |in bus_generic_new_pass

--- Comment #11 from Dennis Clarke <dclarke at blastwave.org> ---
(In reply to Mark Millard from comment #10)


I did try to boot -sv without "usefdt=1" and was given a new type of
panic so this may be progress. 

.
.
.
vgapci0: Boot video device
bge0: <Broadcom BCM5714 B3, ASIC rev. 0x008003> mem
0xfa530000-0xfa53ffff,0xfa520000-0xfa52ffff irq 66 at device 4.0 on pci3
ofw_pci mapdev: start fa530000, len 65536
panic: pci_get_vendor failed for pcib1 on bus ofwbus0, error = 2
cpuid = 0
time = 1
KDB: stack backtrace:
0xe000000000007f00: at .kdb_backtrace+0x5c
0xe000000000008030: at .vpanic+0x1b4
0xe0000000000080f0: at .panic+0x38
0xe000000000008180: at .bge_attach+0x2e08
0xe0000000000082c0: at .device_attach+0x420
0xe0000000000083a0: at .device_probe_and_attach+0xb4
0xe000000000008430: at .bus_generic_new_pass+0x12c
0xe0000000000084c0: at .bus_generic_new_pass+0x114
0xe000000000008550: at .bus_generic_new_pass+0x114
0xe0000000000085e0: at .bus_generic_new_pass+0x114
0xe000000000008670: at .bus_generic_new_pass+0x114
0xe000000000008700: at .bus_generic_new_pass+0x114
0xe000000000008790: at .bus_generic_new_pass+0x114
0xe000000000008820: at .bus_set_pass+0xc4
0xe0000000000088b0: at .root_bus_configure+0x1c
0xe000000000008930: at .configure+0x14
0xe0000000000089b0: at .mi_startup+0x1f8
0xe000000000008a50: at btext+0xc4
KDB: enter: panic
[ thread pid 0 tid 100000 ]
Stopped at      .kdb_enter+0x60:       ld      r2, r1, 0x28
db>

Now this is interesting and I am making a wild guess that it may be 
related to the bge0 abd bge1 interface re-order symptom which is not
really a problem at all. At least not in a practical sense as I don't
care which interface gets the lowest number mac address from the pair
of interfaces available. However I see what looks like a loop there in
bus_generic_new_pass which goes down a fark rabbit hole. Just a guess.

So this is what happens if and only if I do not suggest "usefdt=1" to 
the boot loader as that results in an instant panic.  See picture that
I attached yesterday. 

If I do nothing other than suggest "boot -sv" then we get the above
panic and that seems to be a better place to look. 

It may be useful to go hunting into sys/powerpc/powermac/macgpio.c and
drop a pile of printf's and just see where this bug actually is. 

Also I want to change the title of this bug. 

Dennis

-- 
You are receiving this mail because:
You are the assignee for the bug.