David Miller wrote:
Joel it seems that you should be able to add "nohz=off" to the
kernel boot command line as a workaround for this bug.
David,
I have tried to boot a 2.6.30 with nohz=off. Kernel boots but bbc
module doesn't work anymore :
Jun 22 11:59:25 tchebytchev kernel: i2c-0: Regs at 000007fe7e00002e, 8
devices, own a0, clock 10.
Jun 22 11:59:25 tchebytchev kernel: bbc_i2c: probe of f007f00c failed
with error -22
When I've tried to investigate, kernel oopses :
Unable to handle kernel NULL pointer dereference
tsk->{mm,active_mm}->context = 0000000000000811
tsk->{mm,active_mm}->pgd = fffff8007b9e2000
\|/ ____ \|/
"@'/ .. \`@"
/_| \__/ |_\
\__U_/
modprobe(1852): Oops [#1]
TSTATE: 0000004411001601 TPC: 00000000005d7104 TNPC: 00000000005d7108 Y:
00000000 Not tainted
TPC: <atomic_add+0x4/0x48>
g0: 0000000000000400 g1: 0000000000000001 g2: 00000000008f9000 g3:
fffff8007c4e0910
g4: fffff8007d710fc0 g5: fffff8007f460000 g6: fffff8007b9ec000 g7:
0000000000000000
o0: 0000000000000001 o1: 0000000000000010 o2: 0000000000000001 o3:
0000000000550c90
o4: 0000000000000000 o5: 00000000000002d0 sp: fffff8007b9ef041 ret_pc:
000000000047f638
RPC: <kthread_stop+0x18/0x100>
l0: 0000000000550c90 l1: 0000000000550c64 l2: 0000000000000010 l3:
00000000008c8400
l4: fffff8007fd8be60 l5: 00000000008f96c0 l6: 0000000000967b98 l7:
0000000000967f98
i0: 0000000000000000 i1: fffff8007e380b40 i2: 0000000000000001 i3:
fffff8007e380b40
i4: 0000000000968398 i5: 0000000000000001 i6: fffff8007b9ef101 i7:
000000001008e6ec
I7: <bbc_envctrl_cleanup+0xc/0x120 [bbc]>
Disabling lock debugging due to kernel taint
Caller[000000001008e6ec]: bbc_envctrl_cleanup+0xc/0x120 [bbc]
Caller[000000001008f620]: bbc_i2c_remove+0x8/0x58 [bbc]
Caller[00000000006f287c]: of_platform_device_remove+0x1c/0x40
Caller[000000000062b2e4]: __device_release_driver+0x64/0xa0
Caller[000000000062b3cc]: driver_detach+0xac/0xc0
Caller[000000000062a428]: bus_remove_driver+0x88/0xc0
Caller[0000000000493e78]: SyS_delete_module+0x178/0x240
Caller[0000000000406214]: linux_sparc_syscall32+0x34/0x40
Caller[0000000000014464]: 0x14464
Instruction DUMP: 01000000 01000000 94102001 <c2024000> 8e004008
cfe25001 80a04007 12400004 01000000
Root tchebytchev:[/var/log] >
I have rebuild the _same_ kernel with qla2xxx in module and external
firmware. I have checked that both module and firmware are in my initrd.
2.6.30 hangs when it load qla2xxx module with "NMI watchdog on CPU1" and
"NMI watchdog on CPU0". If I retry with nohz=off option, it boots.
For information, my initrd :
Root tchebytchev:[/boot] > mkinitramfs -o initrd-2.6.30.img -v 2.6.30
...
Copying module directory kernel/drivers/scsi
Adding module /lib/modules/2.6.30/kernel/drivers/scsi/scsi_tgt.ko
Adding module
/lib/modules/2.6.30/kernel/drivers/scsi/scsi_transport_iscsi.ko
Adding module /lib/modules/2.6.30/kernel/drivers/scsi/libiscsi.ko
Adding module /lib/modules/2.6.30/kernel/drivers/scsi/libiscsi_tcp.ko
Adding module /lib/modules/2.6.30/kernel/drivers/cdrom/cdrom.ko
Adding module /lib/modules/2.6.30/kernel/drivers/scsi/sr_mod.ko
Adding module /lib/modules/2.6.30/kernel/drivers/scsi/scsi_transport_sas.ko
Adding module /lib/modules/2.6.30/kernel/drivers/scsi/libsas/libsas.ko
Adding module /lib/modules/2.6.30/kernel/drivers/scsi/iscsi_tcp.ko
Adding module /lib/modules/2.6.30/kernel/drivers/scsi/qla2xxx/qla2xxx.ko
Adding binary /lib/udev/firmware.agent
Adding binary /lib/firmware/ql2200_fw.bin
Adding firmware ql2200_fw.bin
...
Regards,
JKB
--
To unsubscribe from this list: send the line "unsubscribe sparclinux" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
