Re: [PATCH -next 4/5] net: sched: remove AT INGRESS/EGRESS

[Florian Westphal <fw@strlen.de>]

Daniel Borkmann <daniel@iogearbox.net> wrote:
> On 05/04/2015 08:48 PM, Florian Westphal wrote:
> >use single marker to propagate location.
> >tc_at_ingress 1: ingress, 0 is egress.
> >
> >The new flag is set/unset in sch_ingress instead of the core.
> >We will also no longer set skb->tc_verd to AT_EGRESS in the xmit
> >handler.
> >
> >Signed-off-by: Florian Westphal <fw@strlen.de>
> ...
> >diff --git a/include/linux/skbuff.h b/include/linux/skbuff.h
> >index 911d84e..d794077 100644
> >--- a/include/linux/skbuff.h
> >+++ b/include/linux/skbuff.h
> >@@ -488,6 +488,7 @@ static inline u32 skb_mstamp_us_delta(const struct skb_mstamp *t1,
> >   *	@queue_mapping: Queue mapping for multiqueue devices
> >   *	@xmit_more: More SKBs are pending for this queue
> >   *	@tc_from_ingress: skb is processed during rx, not transmit
> >+ *	@tc_at_ingress: skb is processed during rx, not transmit
> 
> Minor nit:
> 
> I think both comments needs to be a bit more clear. I.e. tc_at_ingress
> tells the *current* location whether we are invoked from ingress or
> egress. And, tc_from_ingress explains the *previous* location, whether
> the skbs has been mirred from ingress path or egress path.

Agree, the reason is that I tried to go with single tc_ingress:1 but it
turned out to be not as trivial as I thought.

I'll fix it up in v2, I'll wait a bit more though to give others
a chance to comment.

Thanks for reviewing!