Bug#365730: additional info
This bug is still present in the daily image from 5/3/06 (closed bug 865859)
The crash occurs even when there is no usb device present. The console log
of an attempt to install the image to the same Xserve G5 box from an external
firewire drive ended in the following:
.
.
.
requesting IPIs ...
IPIs requested...
smp_core99_kick_cpu
smp_core99_kick_cpu done
Processor 1 found.
Brought up 2 CPUs
core99_setup_cpu 0 done
Unable to handle kernel paging request for data at address 0xc00000007c4d7480
Faulting instruction address: 0xc00000000002cc80
Oops: Kernel access of bad area, sig: 11 [#1]
SMP NR_CPUS=32 NUMA POWERMAC
Modules linked in:
NIP: C00000000002CC80 LR: C00000000002AAFC CTR: C0000000000467D0
REGS: c00000007ff5b3b0 TRAP: 0300 Not tainted (2.6.16-1-powerpc64)
MSR: 9000000000001032 <ME,IR,DR> CR: 28000082 XER: 000FFFFF
DAR: C00000007C4D7480, DSISR: 0000000040010000
TASK = c00000007ff47820[1] 'swapper' THREAD: c00000007ff58000 CPU: 1
GPR00: 000001E137135EA0 C00000007FF5B630 C000000000439230 00000000004D7480
GPR04: 0000000000000194 F09B89AF5001CFA8 0000000000000000 0000000000000000
GPR08: 0000F09B89AF5000 000001E137135EA0 0000000000000008 C00000007C4D7480
GPR12: 9000000000009032 C00000000037C100 000000000000007D 00000000FFFB6EF6
GPR16: 0000000000000000 C00000000043C8B0 0000000000000000 0000000000012C40
GPR20: 0000000000A00000 000000000001B186 0000000000000000 0000004601187920
GPR24: 000000000001E1B1 D000000000000000 0000000000000000 000000000003FFFF
GPR28: 0000000F09B89AE9 F09B89AF5001CFA8 0000000000000000 00000001FE4C0791
NIP [C00000000002CC80] .native_hpte_updatepp+0x68/0x290
LR [C00000000002AAFC] .htab_call_hpte_updatepp+0x4/0x18
Call Trace:
[C00000007FF5B630] [C000000000393918] 0xc000000000393918 (unreliable)
[C00000007FF5B6D0] [C00000000002AAFC] .htab_call_hpte_updatepp+0x4/0x18
[C00000007FF5B7D0] [C00000000002A78C] .hash_page+0x13c/0x158
[C00000007FF5B840] [C000000000004734] .do_hash_page+0x34/0x40
--- Exception: 301 at .touch_cache+0x90/0x104
LR = .measure_one+0xec/0x174
[C00000007FF5BB30] [C00000000004A960] .measure_one+0xb0/0x174 (unreliable)
[C00000007FF5BBF0] [C00000000004BEB8] .build_sched_domains+0xbc0/0x1014
[C00000007FF5BE70] [C00000000032E2A4] .sched_init_smp+0x34/0x48
[C00000007FF5BF00] [C000000000009408] .init+0x1bc/0x3d8
[C00000007FF5BF90] [C000000000021518] .kernel_thread+0x4c/0x68
Instruction dump:
7cbd2b78 7d290214 78a04dc2 e9290010 e96b0000 7c094878 7d6b1a14 79283e24
419e0008 61080004 39400008 7c2004ac <7d2058a8> 7d205378 7c0059ad 40a2fff4
Note that this installation attempt was identical to that in Bug#365844.
except that a different Xserve G5 box was used. However, the two Xserves
are the same model, purchased at the same time and exist as part of the
same Xserve cluster.
The only difference between the two machines is that the USB ports in this
one have been used, whereas the USB ports in the box from Bug#365844 have
never been touched.
Note also that this box and all it's ports work fine under Mac OS X
Comparing the console log of Bug#365844 shows a difference that
seems odd. In Bug#365844, the detection of the pmu is represented
with 'via-pmu: Server Mode is enabled'. The box on which the install
crashes instead shows 'via-pmu: Server Mode is disabled'. There
also differences in the reported size of nvram, although this may
be normal.
Clearly there is something different about the two boxes, despite
the fact that they *should* be identical.
Reply to: