On Wed, Jan 04, 2017 at 12:21:51AM +0100, Michał Mirosław wrote: > On Tue, Jan 03, 2017 at 04:32:17PM -0500, David Miller wrote: > > From: Michał Mirosław <mirq-li...@rere.qmqm.pl> > > Date: Tue, 3 Jan 2017 21:52:33 +0100 (CET) > > > > > Dear NetDevs > > > > > > This series removes an abuse of VLAN CFI bit in Linux networking stack. > > > Currently Linux always clears the bit on outgoing traffic and presents > > > it cleared to userspace (even via AF_PACKET/tcpdump when hw-accelerated). > > > > > > This uses a new vlan_present bit in struct skbuff, and removes an > > > assumption > > > that vlan_proto != 0 when VLAN tag is present. > > > > > > As I can't test most of the driver changes, please look at them carefully. > > > > > > The series is supposed to be bisect-friendly and that requires temporary > > > insertion of #define VLAN_TAG_PRESENT in BPF code to be able to split > > > JIT changes per architecture. > > > > > > Best Regards, > > > Michał Mirosław > > > > > > v2: rebase onto net-next > > > > This patch series is really way too large. > > > > You're going to have to find a way to combine related changes, or submit > > this as a series of logical sets, one at a time. > > The dependency graph is really sparse: main patch 25 depends on all previous, > 19-23 all depend only on 18 and 26 depends on 25. > > That's it. So the question is: how would it be easier for you to manage?
Ah, I forgot about patches 1-2 that most depend on... So if you get those two early, then most others can be split. Best Regards, Michał Mirosław
