Re: [ath9k-devel] BUG kmalloc-8192: Poison overwritten

[Luis R. Rodriguez]

On Thu, Dec 3, 2009 at 1:36 PM, Justin P. Mattock
<justinmattock@xxxxxxxxx> wrote:
> On 12/02/09 15:16, Luis R. Rodriguez wrote:
>>
>> On Wed, Dec 2, 2009 at 2:27 PM, Justin Mattock<justinmattock@xxxxxxxxx>
>> Âwrote:
>>>
>>> I have no idea what this is:
>>> (first time I've seen this);
>>>
>>>
>>> [ Â 13.055592]
>>> =============================================================================
>>> [ Â 13.055865] BUG kmalloc-8192: Poison overwritten
>>> [ Â 13.056046]
>>> -----------------------------------------------------------------------------
>>
>> I believe this means we write to an area in memory after we freed an skb.
>>
>>> [ Â 13.056366] INFO: 0xffff88003290a100-0xffff88003290a143. First byte
>>> 0x50 instead of 0x6b
>>> [ Â 13.056366] INFO: Allocated in ath_rxbuf_alloc+0x25/0x94 [ath]
>>> age=8898 cpu=0 pid=892
>>> [ Â 13.056366] INFO: Freed in skb_release_data+0xc1/0xc6 age=99 cpu=0
>>> pid=0
>>> [ Â 13.056366] INFO: Slab 0xffffea0000b0f9c0 objects=3 used=0
>>> fp=0xffff88003290c090 flags=0x4000000000004082
>>> [ Â 13.056366] INFO: Object 0xffff88003290a048 @offset=8264
>>> fp=0xffff880032908000
>>> [ Â 13.056366]
>>> [ Â 13.056366] Bytes b4 0xffff88003290a038: Âb9 9e fb ff 00 00 00 00
>>> 5a 5a 5a 5a 5a 5a 5a 5a Â.ÃÃ....ZZZZZZZZ
>>
>> Can you reproduce this? We had a similar report from Miles Lanes I
>> have been trying to followup on for a while now. The suggested patches
>> to try are:
>>
>>
>> http://bombadil.infradead.org/~mcgrof/patches/ath9k/2009/11-19/70-mac80211-total-ampdu-tx-info.patch
>>
>> http://bombadil.infradead.org/~mcgrof/patches/ath9k/2009/11-19/71-ath9k-remove-tx_info_priv.patch
>>
>> Â Luis
>>
>
> as anticipated I had a feeling reproducing
> this would be a bit of a challeng.
>
> so far I was taking into account the actions that
> I did yesterday to trigger this, but nothing i.e.
>
> reboot 30 times(until fsck engages), and shutdown/start 30 times
> with nothing of the sort.
>
> If you have any ideas on how to reproduce this it would
> be greatly appreciated(don't want to throw in the patches, until
> I have something to go by).
>
> on the other hand I could throw in the patches, and wait and see,
> but would rather have more hard evidence that the patches do fix something.

I think waiting to be able to reproduce is best, and appreciate your
patience on this. Maybe Miles Lanes also cannot reproduce easily and
that may explain why he hasn't replied.

  Luis
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/