On Thursday 30 July 2009 04:10:32 Larry Finger wrote:
> Francesco Gringoli wrote:
> > 
> > is this stuff related in some way to the error triggered by the
> > opensource firmware? What happens with your patch?
> 
> Using the same test with tcpperf running on one console and a flood
> ping on a second, the system lasted 7 minutes before meta->skb was
> NULL. The data are:
> 
> b43: meta data: skb              (null)
>                 dmaaddr          000000004fa820bc
>                 is_last_fragment 1
> b43: ring data: nr_slots       256
>                 used_slots     205
>                 current_slot   105
>                 index          1
>                 tx             1
>                 max_used_slots 256
> 
> Notice that the DMA queue is not even close to full - there are about
> 50 slots left or enough space for 25 more packets.

Yeah but it _was_ full earlier (max_used_slots 256).
I don't say that this is triggering the issue, however.

skb=NULL somehow smells like a double-free caused by a double-report
or something like that.

> This error was generated with the DMA queue full patch applied.



-- 
Greetings, Michael.
_______________________________________________
Bcm43xx-dev mailing list
[email protected]
https://lists.berlios.de/mailman/listinfo/bcm43xx-dev

Reply via email to