Re: Oops while modprobing phy fixed module

[Gabriel C <nix.or.die@googlemail.com>]

Gabriel C wrote:
> Satyam Sharma wrote:
>   
>> Hi Gabriel,
>>   
>>     
>
> Hi Satyam ,
>
>   
>> On 7/14/07, Gabriel C <nix.or.die@googlemail.com> wrote:
>>   
>>     
>>> Hi,
>>>
>>> doing a modprobe fixed the driver segfaults and I get this Oops:
>>>
>>>
>>> Jul 14 13:43:30 lara [  157.952915] Fixed PHY: Registered new driver
>>> Jul 14 13:43:30 lara [  157.953010] Device 'fixed@100:1' does not have a
>>> release() function, it is broken and must be fixed.
>>> Jul 14 13:43:30 lara [  157.953019] WARNING: at drivers/base/core.c:107
>>> device_release()
>>> Jul 14 13:43:30 lara [  157.953032]  [<c01cf949>] kobject_cleanup+0x3d/0x54
>>> Jul 14 13:43:30 lara [  157.953050]  [<c01cf960>] kobject_release+0x0/0x8
>>> Jul 14 13:43:30 lara [  157.953060]  [<c01d0530>] kref_put+0x60/0x6d
>>> Jul 14 13:43:30 lara [  157.953068]  [<c022f31c>] device_del+0x1f3/0x215
>>> Jul 14 13:43:30 lara [  157.953083]  [<e8c8f2d5>]
>>> fixed_mdio_register_device+0x1e2/0x20d [fixed]
>>> Jul 14 13:43:30 lara [  157.953108]  [<e8c8701b>] fixed_init+0x1b/0x2f
>>> [fixed]
>>> Jul 14 13:43:30 lara [  157.953119]  [<c01376ab>]
>>> sys_init_module+0x1686/0x175a
>>> Jul 14 13:43:30 lara [  157.953133]  [<c01601e3>] do_sync_read+0x0/0x10a
>>> Jul 14 13:43:30 lara [  157.953180]  [<c0103dee>]
>>> sysenter_past_esp+0x5f/0x85
>>> Jul 14 13:43:30 lara [  157.953195]  [<c0300000>]
>>> packet_setsockopt+0x1c9/0x2f3
>>> Jul 14 13:43:30 lara [  157.953232]  =======================
>>>     
>>>       
>> This looks unrelated to the oops itself, but still something that
>> needs to be fixed, of course. [ Maintainers added to Cc: ]
>>
>>   
>>     
>>> Jul 14 13:43:30 lara [  157.953261] BUG: unable to handle kernel paging
>>> request at virtual address 43b7a800
>>>     
>>>       
>> 43b7a800 looks suspicious, it could have been a valid kernel
>> address, if only for what looks like a single-bit flip.
>>
>>   
>>     
>>> Jul 14 13:43:30 lara [  157.953273]  printing eip:
>>> Jul 14 13:43:30 lara [  157.953278] c015d269
>>> Jul 14 13:43:30 lara [  157.953283] *pde = 00000000
>>> Jul 14 13:43:30 lara [  157.953293] Oops: 0000 [#1]
>>> Jul 14 13:43:30 lara [  157.953301] PREEMPT SMP
>>> Jul 14 13:43:30 lara [  157.953309] Modules linked in: fixed pc87360
>>> hwmon_vid i2c_isa eeprom adm1021 uhci_hcd sr_mod shpchp pci_hotplug
>>> ohci_hcd iTCO_wdt iTCO_vendor_support intel_agp i82860_edac i2c_i801
>>> ehci_hcd usbcore edac_mc cdrom agpgart 3c59x mii ext4dev jbd2 capability
>>> commoncap loop lp parport_pc parport
>>> Jul 14 13:43:30 lara [  157.953386] CPU:    3
>>> Jul 14 13:43:30 lara [  157.953387] EIP:    0060:[<c015d269>]    Not
>>> tainted VLI
>>> Jul 14 13:43:30 lara [  157.953391] EFLAGS: 00210006   (2.6.22-g8d9107e8 #7)
>>> Jul 14 13:43:30 lara [  157.953404] EIP is at kmem_cache_zalloc+0x75/0x89
>>> Jul 14 13:43:30 lara [  157.953412] eax: 00000000   ebx: 00200282   ecx:
>>> c14c8760   edx: 43b7a800
>>> Jul 14 13:43:30 lara [  157.953423] esi: e7f75840   edi: c17937b8   ebp:
>>> 000000d0   esp: db1ced98
>>> Jul 14 13:43:30 lara [  157.953434] ds: 007b   es: 007b   fs: 00d8  gs:
>>> 0033  ss: 0068
>>> Jul 14 13:43:30 lara [  157.953444] Process modprobe (pid: 2164,
>>> ti=db1ce000 task=de78ac20 task.ti=db1ce000)
>>> Jul 14 13:43:30 lara [  157.953450] Stack: c014cfd4 c01cf2f7 43b7a800
>>> c03ae384 c036d990 db150690 c17937b8 c17937b8
>>> Jul 14 13:43:30 lara [  157.953470]        c019752a 00000002 41ed0000
>>> e643bb60 c036d990 db150690 c036d990 e643bb60
>>> Jul 14 13:43:30 lara [  157.953489]        c01979b1 c0197712 db1cede4
>>> 00000000 e643bb60 c036d990 00000000 c03b8e98
>>> Jul 14 13:43:30 lara [  157.953513] Call Trace:
>>> Jul 14 13:43:30 lara [  157.953518]  [<c014cfd4>] kstrdup+0x27/0x47
>>> Jul 14 13:43:30 lara [  157.953530]  [<c01cf2f7>]
>>> ida_get_new_above+0xe6/0x166
>>> Jul 14 13:43:30 lara [  157.953551]  [<c019752a>] sysfs_new_dirent+0x3d/0xdc
>>> Jul 14 13:43:30 lara [  157.953574]  [<c01979b1>] create_dir+0x1e/0x8c
>>> Jul 14 13:43:30 lara [  157.953588]  [<c0197712>]
>>> sysfs_addrm_finish+0x13/0x1d8
>>> Jul 14 13:43:30 lara [  157.953609]  [<c0197a90>]
>>> sysfs_create_subdir+0x13/0x16
>>> Jul 14 13:43:30 lara [  157.953622]  [<c0198dba>]
>>> sysfs_create_group+0x25/0xe7
>>> Jul 14 13:43:30 lara [  157.953642]  [<c0233616>] device_pm_add+0x38/0x72
>>> Jul 14 13:43:30 lara [  157.953655]  [<c022f61d>] device_add+0x230/0x3d8
>>> Jul 14 13:43:30 lara [  157.953678]  [<e8c8f26d>]
>>> fixed_mdio_register_device+0x17a/0x20d [fixed]
>>> Jul 14 13:43:30 lara [  157.953706]  [<e8c8702c>] fixed_init+0x2c/0x2f
>>> [fixed]
>>> Jul 14 13:43:30 lara [  157.953718]  [<c01376ab>]
>>> sys_init_module+0x1686/0x175a
>>> Jul 14 13:43:30 lara [  157.953734]  [<c01601e3>] do_sync_read+0x0/0x10a
>>> Jul 14 13:43:30 lara [  157.953797]  [<c0103dee>]
>>> sysenter_past_esp+0x5f/0x85
>>> Jul 14 13:43:30 lara [  157.953816]  [<c0300000>]
>>> packet_setsockopt+0x1c9/0x2f3
>>> Jul 14 13:43:30 lara [  157.953835]  =======================
>>> Jul 14 13:43:30 lara [  157.953841] Code: 08 00 74 2f 8b 56 08 31 c0 89
>>> d1 c1 e9 02 8b 7c 24 08 f3 ab f6 c2 02 74 02 66 ab f6 c2 01 74 01 aa eb
>>> 10 0f b7 41 0a 8b 54 24 08 <8b> 04 82 89 41 0c eb c8 8b 44 24 08 83 c4
>>> 10 5b 5e 5f 5d c3 57
>>>     
>>>       
>> Thanks to Randy's wonderful decodecode script, we know this is
>> mov    (%edx,%eax,4),%eax
>>
>>   
>>     
>>> Jul 14 13:43:30 lara [  157.953943] EIP: [<c015d269>]
>>> kmem_cache_zalloc+0x75/0x89 SS:ESP 0068:db1ced98
>>>     
>>>       
>> I think this is slab_alloc() inlined from SLUB's kmem_cache_zalloc():
>>
>> 	page = s->cpu_slab[smp_processor_id()];
>> 	if (unlikely(!page || !page->lockless_freelist ||
>> 			(node != -1 && page_to_nid(page) != node)))
>>
>> 		object = __slab_alloc(s, gfpflags, node, addr, page);
>>
>> 	else {
>> 		object = page->lockless_freelist;
>> 		page->lockless_freelist = object[page->offset];	<=== Oops
>> 	}
>>
>> which shouldn't happen unless you have memory errors. I'd suggest
>> to check with memtest86+.
>>   
>>     
>
> I did so again ( did a 24h memtest run like 2 weeks ago with no errors )
> for 12 hours now an no errors / no ECC errors , so I think my RAM is ok.
>
> That box is using RDRAM PC800 ECC Rambus.
>
> Do you want me to try reproduce this Oops with a full debug enabled kernel ?
>   

Also I switched to SLAB ( with the same config ) and no Oops anymore
just the driver warning left.


Gabriel