On Wed, 2007-08-22 at 13:39 -0700, David Miller wrote:
From: Tom \spot\ Callaway [EMAIL PROTECTED]
Date: Wed, 22 Aug 2007 16:31:58 -0400
Got this oops with the attached config very quickly after init, threw me
back to the prom. Firmware is fully updated on the box, latest SILO too.
Starting udev: OOPS: Bogus kernel PC [] in fault handler
OOPS: RPC [00477aa4]
OOPS: Fault was to vaddr[f7f22000]
Something jumped to address zero.
This happened at address 0x477aa4, find out where that is.
Also, if you're using gcc-4.2.x for kernel builds, don't.
There is a miscompile, that others have run into, which
I have tracked down and am trying to build a test case
for so the gcc folks can look at and hopefully fix it.
I'm not. Fedora is still on gcc 4.1.
System.map says:
004778bc t rcu_start_batch
00477928 t __rcu_process_callbacks
00477b98 t rcu_process_callbacks
00477bd4 t rcu_barrier_callback
Not sure how useful that is. I accidentally rebooted into this kernel,
and it booted all the way to login this time, but I noticed the
following in dmesg:
ldc.c:v1.0 (June 25, 2007)
ldc: Domaining disabled.
NET: Registered protocol family 16
VIO: Adding device channel-devices
VIO: Adding device vldc-port-0-0
VIO: Adding device vldc-port-0-1
VIO: Adding device vldc-port-0-2
VIO: Adding device vldc-port-1-0
VIO: Adding device vldc-port-3-0
VIO: Adding device vldc-port-3-8
VIO: Adding device ds-0
VIO: Adding device ds-1
VIO: Adding device ds-0
kobject_add failed for ds-0 with -EEXIST, don't try to register things
with the
same name in the same directory.
Call Trace:
[005536e8] kobject_shadow_add+0x1a4/0x1e8
[0055373c] kobject_add+0x10/0x1c
[005b60a8] device_add+0x88/0x588
[005b65bc] device_register+0x14/0x20
[0044fd58] vio_create_one+0x408/0x45c
[0044fdc8] vio_add+0x1c/0x24
[0043c340] mdesc_register_notifier+0x40/0x78
[007e2c00] vio_init+0x17c/0x19c
[007d6390] kernel_init+0x228/0x3c8
[0042783c] kernel_thread+0x38/0x48
[00685e74] rest_init+0x18/0x64
VIO: Could not register device ds-0, err=-17
One last item: This kernel loads the e1000 driver fine, but while it
detects the devices, it doesn't think there is link on the eth0 port (or
any of eth0-3 for that matter). There is definitely a good link there,
but it won't pull an IP over dhcp.
bash-3.1# dmesg |grep e1000
e1000: 0001:07:00.0: e1000_probe: (PCI Express:2.5Gb/s:Width x4)
00:14:4f:1d:9f:4e
e1000: eth0: e1000_probe: Intel(R) PRO/1000 Network Connection
e1000: 0001:07:00.1: e1000_probe: (PCI Express:2.5Gb/s:Width x4)
00:14:4f:1d:9f:4f
e1000: eth1: e1000_probe: Intel(R) PRO/1000 Network Connection
e1000: :04:00.0: e1000_probe: (PCI Express:2.5Gb/s:Width x4)
00:14:4f:1d:9f:4c
e1000: eth2: e1000_probe: Intel(R) PRO/1000 Network Connection
e1000: :04:00.1: e1000_probe: (PCI Express:2.5Gb/s:Width x4)
00:14:4f:1d:9f:4d
e1000: eth3: e1000_probe: Intel(R) PRO/1000 Network Connection
lspci says:
:04:00.0 Ethernet controller: Intel Corporation 82571EB Gigabit
Ethernet Controller (rev 06)
:04:00.1 Ethernet controller: Intel Corporation 82571EB Gigabit
Ethernet Controller (rev 06)
0001:07:00.0 Ethernet controller: Intel Corporation 82571EB Gigabit
Ethernet Controller (rev 06)
0001:07:00.1 Ethernet controller: Intel Corporation 82571EB Gigabit
Ethernet Controller (rev 06)
Is this the correct place to report that regression, or should I go to
LKML or Intel?
Thanks,
~spot
-
To unsubscribe from this list: send the line unsubscribe sparclinux in
the body of a message to [EMAIL PROTECTED]
More majordomo info at http://vger.kernel.org/majordomo-info.html