Re: Latest git broken with latest Stable kernel 2.6.36.1

[Jarod W. <ja...@wi...>]

On Nov 26, 2010, at 1:08 AM, VDR User wrote:

> Debian testing, gcc 4.4.5-8, lirc git nov.25 2010.  This is from dmesg:
> 
> lirc_dev: IR Remote Control driver registered, major 254
> BUG: using smp_processor_id() in preemptible [00000000] code: modprobe/1226
> caller is init_timing_params+0x26/0xdc [lirc_serial]
> Pid: 1226, comm: modprobe Tainted: P            2.6.36.1.amd64-x2.112510.1 #1
> Call Trace:
> [<c116b795>] ? debug_smp_processor_id+0xa1/0xb4
> [<f8325371>] ? init_timing_params+0x26/0xdc [lirc_serial]
> [<f83290f9>] ? lirc_serial_init_module+0xf9/0x2d1 [lirc_serial]
> [<f8329000>] ? lirc_serial_init_module+0x0/0x2d1 [lirc_serial]
> [<c1001151>] ? do_one_initcall+0x68/0x106
> [<c10460d1>] ? sys_init_module+0x77/0x196
> [<c1002610>] ? sysenter_do_call+0x12/0x26
> lirc_serial: auto-detected active low receiver
> lirc_serial lirc_serial.0: lirc_dev: driver lirc_serial registered at minor = 0
> lirc_serial $Revision: 5.108 $ registered
> 
> Then after I load VDR and press a button:
> 
> divide error: 0000 [#1] PREEMPT SMP
> last sysfs file: /sys/devices/virtual/vtconsole/vtcon0/uevent
> Modules linked in: lirc_serial lirc_dev
> 
> Pid: 1230, comm: lircd Tainted: P
> 2.6.36.1.amd64-x2.112510.1 #1 MS-7309/MS-7309
> EIP: 0060:[<f8317b6c>] EFLAGS: 00010246 CPU: 1
> EIP is at lirc_dev_fop_read+0xb6/0x276 [lirc_dev]
> EAX: 00000004 EBX: f663d500 ECX: c138a448 EDX: 00000000
> ESI: bfcaa8bc EDI: fffffff4 EBP: f6a311b0 ESP: f68e1f34
> DS: 007b ES: 007b FS: 00d8 GS: 0033 SS: 0068
> Process lircd (pid: 1230, ti=f68e0000 task=f6a311b0 task.ti=f68e0000)
> Stack:
> f6a6cc7c 00000010 f663d570 00000004 f68d1940 f73e3b74 bfcaa8bc 00000000
> <0> f6a311b0 c1021477 00000000 00000000 00000000 f68d1940 bfcaa8bc f8317ab6
> <0> f68e1f9c c1072cb6 f68e1f9c c1017899 f68d1940 fffffff7 bfcaa948 f68e0000
> Call Trace:
> [<c1021477>] ? default_wake_function+0x0/0x8
> [<f8317ab6>] ? lirc_dev_fop_read+0x0/0x276 [lirc_dev]
> [<c1072cb6>] ? vfs_read+0x70/0x9f
> [<c1017899>] ? do_page_fault+0x1eb/0x24e
> [<c1072d78>] ? sys_read+0x3c/0x60
> [<c1002610>] ? sysenter_do_call+0x12/0x26
> Code: d7 92 f7 c8 85 c0 74 0b 31 f6 66 bf 00 fe e9 bf 01 00 00 83 7b
> 68 00 75 0c 31 f6 bf ed ff ff ff e9 a4 01 00 00 31 d2 8b 44 24 0c <f7>
> b3 88 00 00 00 85 d2 74 0c 31 f6 bf ea ff ff ff e9 88 01 00
> EIP: [<f8317b6c>] lirc_dev_fop_read+0xb6/0x276 [lirc_dev] SS:ESP 0068:f68e1f34
> ---[ end trace 0644bbe5a97a4206 ]---
> 
> Anyone have a patch or know how to fix this?

Not yet, but I've finally had a moment to look into it. I've reproduced
the divide error problem here, now attempting to isolate the cause.

-- 
Jarod Wilson
ja...@wi...