Re: panic: Duplicate free of item 0xc2580a00 from zone 0xc103e9a0(Mbuf)

Robert Watson wrote:
> On Thu, 24 Feb 2005, Peter Holm wrote:
>>With GENERIC HEAD from Feb 23 06:23 UTC + mpsafe_vfs = 1 I got:
>>
>>Slab at 0xc2580fa8, freei 10 = 0.  panic: Duplicate free of item
>>0xc2580a00 from zone 0xc103e9a0(Mbuf) 
> 
> rl_encap() appears to be pretty broken in the event that m_defrag() has to
> allocate a new mbuf on the head of the chain.  Specifcally, rl_encap() 
> may allocate a new head to the chain, but the caller won't use that new
> head, instead using the old head (which may have been free'd, or otherwise
> munged).  This might also explain other crashes or panics involving rl
> interfaces.  I'm surprised this lasted as long as it apparently has in the
> wild (since 2003?).

 From src/sys/pci/if_rl.c:
"The RealTek 8139 PCI NIC redefines the meaning of 'low end.'  This is 
probably the worst PCI ethernet controller ever made, with the possible 
exception of the FEAST chip made by SMC.  The 8139 supports bus-master 
DMA, but it has a terrible interface that nullifies any performance 
gains that bus-master DMA usually offers.
..."

I would bet that (in)famous comment coupled with everyone recommending 
fxp or em adapters[1] causes most who experience problems to abandon rl 
adapters at the first hint of trouble.  In that type of an environment, 
it is not overly surprising to me that a bug could last this long.

Jon

[1] I use fxp and em adapters and I've never had a problem with them. 
As such, I can't argue with the recommendation.