Kai> The isdn_ppp_skb_push function discarded packets without
Kai> sufficient headroom, no it reallocs headroom instead, which
Kai> makes more sense.
Well, the error message in previous kernels also made sense because it
indicates a real problem that should never happen (upper layers must
never pass an skb with insufficient headroom to a net_device). The
re-allocation is a work-around (which is reasonnable to apply only
until we know better),
Kai> So the "isdn_ppp_skb_push: under %d %d" bug, which some
Kai> people saw when using MPPP, is fixed.
but it does not fix the problem itself. I think that with the
reallocation work-around, it is even more important to leave the
error message such that users will continue to generate bug-reports
until the real problem is fixed.
Kai> There remains one question: What has changed lately, which
Kai> might cause packets without sufficient headroom being
Kai> delivered from the network-layer to the device driver? This
Kai> could affect performance on e.g. ethernet cards as well, I
Kai> suppose.
Don't know. In order to track this down, I think it is essential
that we leave an error message in there. Otherwise, the patch will
just fix the symptoms an maybe people will cease to be aware
of that severe problem.
Henner
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/