BTW: I want to know the advantage about instead of the ic->ic_mgtstart with ic_mgtq in new main CVS and bsc branch?
On Fri, 2005-07-22 at 18:07, Sebastian Weitzel wrote:
> > I can only use my Netgear WT311 (IIRC) under 2.4.29 at the moment, and
> > without the attached patch, get occasional oopses. This is because
> > Win4Lin's vnet adapter dereferences the skb value. Feel free to tell me
> > this is wrong if I should be getting the vnet adapter to return
> > immediately on !skb.
> > - (*dev->hard_start_xmit)(NULL, dev);
> > + (*dev->hard_start_xmit)(skb, dev);
> This is definetly not ok. Please note that management frames are
> handled differently from normal data frames. If hard_start_xmit is
> called with skb reference only for data frames. Management frames
> should go to ic_mgtq.
> If there is a bug, then it is most probably in ath_start(). I dont
> understand was vnet has to do with madwifi? If it interferes with
> madwifi devices, it could be that vnet is doing wrong.
When I set MERGE_VNET_IFNAME=ath0 in /etc/default/merge (telling Win4Lin
to use ath0), hard_start_xmit calls mkia_vnetint_outgoing
/* Outgoing Packet Handler */
mkia_vnetint_outgoing(struct sk_buff *skb, MKI_DEV_T *dev)
struct vnetint_ifdev *vd;
struct vnetint_pcb *vi;
int (*real_xmit)() = 0;
unsigned short protocol;
* We can't rely on skb->protocol to hold a valid protocol
* type so go look in the ethernet header.
eth = (ETH_HDR_T *)skb->data;
protocol = eth->type;
As you can see the skb is dereferenced, resulting in the oops.
Enumerate the requirements.
Consider the interdependencies.
Calculate the probabilities.
SF.Net email is sponsored by: Discover Easy Linux Migration Strategies
from IBM. Find simple to follow Roadmaps, straightforward articles,
informative Webcasts and more! Get everything you need to get up to
speed, fast. http://ads.osdn.com/?ad_id=7477&alloc_id=16492&op=click
Madwifi-users mailing list