Re: [LKP] [drm/radeon] BUG: unable to handle kernel paging request at 03200000

[Oded Gabbay <oded.gabbay@amd.com>]

On 12/18/2014 10:26 AM, Huang Ying wrote:
> On Thu, 2014-12-18 at 10:20 +0200, Oded Gabbay wrote:
>> On 12/18/2014 10:17 AM, Huang Ying wrote:
>>> FYI, we noticed the below changes on
>>>
>>> commit e28740ece34d314002b1ddfa14e8fb7c7b909489 ("drm/radeon: Add radeon <--> amdkfd interface")
>>>
>>>
>>> +------------------------------------------+------------+------------+
>>> |                                          | 1c0a46255f | e28740ece3 |
>>> +------------------------------------------+------------+------------+
>>> | boot_successes                           | 0          | 0          |
>>> | boot_failures                            | 10         | 13         |
>>> | BUG:kernel_boot_hang                     | 9          | 9          |
>>> | INFO:suspicious_RCU_usage                | 1          | 3          |
>>> | backtrace:do_fork                        | 1          | 3          |
>>> | BUG:unable_to_handle_kernel              | 0          | 3          |
>>> | Oops                                     | 0          | 3          |
>>> | Kernel_panic-not_syncing:Fatal_exception | 0          | 3          |
>>> | backtrace:radeon_init                    | 0          | 3          |
>>> | backtrace:kernel_init_freeable           | 0          | 3          |
>>> | BUG:kernel_boot_crashed                  | 0          | 1          |
>>> +------------------------------------------+------------+------------+
>>>
>>>
>>> [    5.407193] Linux agpgart interface v0.103
>>> [    5.428469] [drm] Initialized drm 1.1.0 20060810
>>> [    5.438456] [drm] radeon kernel modesetting enabled.
>>> [    5.439741] BUG: unable to handle kernel paging request at 03200000
>>> [    5.440063] IP: [<03200000>] 0x3200000
>>> [    5.440063] *pde = 00000000
>>> [    5.440063] Oops: 0000 [#1] SMP
>>> [    5.440063] CPU: 0 PID: 1 Comm: swapper/0 Not tainted 3.18.0-g603ba7e #14
>>> [    5.440063] task: 42e08000 ti: 42e02000 task.ti: 42e02000
>>> [    5.440063] EIP: 0060:[<03200000>] EFLAGS: 00010246 CPU: 0
>>> [    5.440063] EIP is at 0x3200000
>>> [    5.440063] EAX: 00000001 EBX: 00000000 ECX: 4738b460 EDX: 45735a64
>>> [    5.440063] ESI: 4662d0dd EDI: 401fafa0 EBP: 42e03f20 ESP: 42e03f18
>>> [    5.440063]  DS: 007b ES: 007b FS: 00d8 GS: 0000 SS: 0068
>>> [    5.440063] CR0: 8005003b CR2: 03200000 CR3: 067f0000 CR4: 00000690
>>> [    5.440063] Stack:
>>> [    5.440063]  44bf19b1 00000000 42e03f28 4662d20b 42e03f8c 465b6e5d 00000000 00000000
>>> [    5.440063]  00000000 00000000 00000000 00000000 000000ea 00000000 00000000 00000101
>>> [    5.440063]  45f2dfba 00000010 00000012 00000200 47fdcf54 47fdcf54 00000006 00000291
>>> [    5.440063] Call Trace:
>>> [    5.440063]  [<44bf19b1>] ? radeon_kfd_init+0x67/0xab
>>> [    5.440063]  [<4662d20b>] radeon_init+0x12e/0x15c
>>> [    5.440063]  [<465b6e5d>] do_one_initcall+0x25f/0x41e
>>> [    5.440063]  [<465b7291>] kernel_init_freeable+0x275/0x423
>>> [    5.440063]  [<455b4e52>] kernel_init+0x1c/0x293
>>> [    5.440063]  [<455df181>] ret_from_kernel_thread+0x21/0x30
>>> [    5.440063]  [<455b4e36>] ? rest_init+0x33e/0x33e
>>> [    5.440063] Code:  Bad EIP value.
>>> [    5.440063] EIP: [<03200000>] 0x3200000 SS:ESP 0068:42e03f18
>>> [    5.440063] CR2: 0000000003200000
>>> [    5.440063] ---[ end trace 1959b295b01c0c14 ]---
>>> [    5.440063] Kernel panic - not syncing: Fatal exception
>>>
>>>
>>> Thanks,
>>> Huang, Ying
>>>
>>>
>>>
>>>
>>> _______________________________________________
>>> LKP mailing list
>>> LKP@linux.intel.com
>>>
>>
>> Is this a 32-bit kernel ?
>> Is amdkfd compiled ?
>
> The kconfig file is attached in the original email.  From it, it is
> 32-bit kernel.  I don't find AMDKFD in config file.  Could you help me
> to check it?
>
> Best Regards,
> Huang, Ying
>
>
> --
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at  http://www.tux.org/lkml/
>
OK, I found the culprit.
If you define 32-bit kernel, No modules and random kernel base, than the 
symbol_request function doesn't return NULL if the symbol doesn't exists, which 
causes the rest of the code to have undefined results.

I prepared a patch to fix it and hopefully it will be in -rc2. I will sent it to 
dri-devel as usual.

	Oded