[PATCH net-next 00/27] Remove VLAN CFI bit abuse
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 --- Michał Mirosław (27): net/vlan: introduce __vlan_hwaccel_clear_tag() helper net/vlan: introduce __vlan_hwaccel_copy_tag() helper ibmvnic: fix accelerated VLAN handling qlcnic: remove assumption that vlan_tci != 0 i40iw: remove use of VLAN_TAG_PRESENT cnic: remove use of VLAN_TAG_PRESENT gianfar: remove use of VLAN_TAG_PRESENT net/hyperv: remove use of VLAN_TAG_PRESENT cxgb4: use __vlan_hwaccel helpers benet: use __vlan_hwaccel helpers sky2: use __vlan_hwaccel helpers net/core: use __vlan_hwaccel helpers bridge: use __vlan_hwaccel helpers 8021q: use __vlan_hwaccel helpers ipv4/tunnel: use __vlan_hwaccel helpers nfnetlink/queue: use __vlan_hwaccel helpers OVS: remove assumptions about VLAN_TAG_PRESENT bit net/skbuff: add macros for VLAN_PRESENT bit net/bpf_jit: ARM: split VLAN_PRESENT bit handling from VLAN_TCI net/bpf_jit: MIPS: split VLAN_PRESENT bit handling from VLAN_TCI net/bpf_jit: PPC: split VLAN_PRESENT bit handling from VLAN_TCI net/bpf_jit: SPARC: split VLAN_PRESENT bit handling from VLAN_TCI net/bpf: split VLAN_PRESENT bit handling from VLAN_TCI bpf_test: prepare for VLAN_TAG_PRESENT removal net: remove VLAN_TAG_PRESENT net/hyperv: enable passing of VLAN.CFI bit net/vlan: remove unused #define HAVE_VLAN_GET_TAG Documentation/networking/openvswitch.txt | 14 -- arch/arm/net/bpf_jit_32.c| 17 --- arch/mips/net/bpf_jit.c | 17 +++ arch/powerpc/net/bpf_jit_comp.c | 14 +++--- arch/sparc/net/bpf_jit_comp.c| 14 +++--- drivers/infiniband/hw/cxgb4/cm.c | 2 +- drivers/infiniband/hw/i40iw/i40iw_cm.c | 8 ++-- drivers/net/ethernet/broadcom/cnic.c | 2 +- drivers/net/ethernet/emulex/benet/be_main.c | 4 +- drivers/net/ethernet/freescale/gianfar_ethtool.c | 8 ++-- drivers/net/ethernet/ibm/ibmvnic.c | 5 +- drivers/net/ethernet/marvell/sky2.c | 6 +-- drivers/net/ethernet/qlogic/qlcnic/qlcnic_io.c | 8 ++-- drivers/net/hyperv/hyperv_net.h | 2 +- drivers/net/hyperv/netvsc_drv.c | 14 +++--- drivers/net/hyperv/rndis_filter.c| 5 +- include/linux/if_vlan.h | 37 +++--- include/linux/skbuff.h | 10 +++- lib/test_bpf.c | 14 +++--- net/8021q/vlan_core.c| 2 +- net/bridge/br_netfilter_hooks.c | 14 +++--- net/bridge/br_private.h | 2 +- net/bridge/br_vlan.c | 6 +-- net/core/dev.c | 8 ++-- net/core/filter.c| 17 +++ net/core/skbuff.c| 2 +- net/ipv4/ip_tunnel_core.c| 2 +- net/netfilter/nfnetlink_queue.c | 5 +- net/openvswitch/actions.c| 13 ++--- net/openvswitch/flow.c | 4 +- net/openvswitch/flow.h | 4 +- net/openvswitch/flow_netlink.c | 61 net/sched/act_vlan.c | 2 +- 33 files changed, 170 insertions(+), 173 deletions(-) -- 2.10.2
Re: [PATCH net-next 00/27] Remove VLAN CFI bit abuse
On Tue, Dec 13, 2016 at 01:12:32AM +0100, Michał Mirosław wrote: > 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). [...] I just noticed net-next got closed few days ago. I'll resend after it opens again. Nevertheless, review is appreciated. Best Regards, Michał Mirosław
Re: [PATCH net-next 00/27] Remove VLAN CFI bit abuse
On Tue, 13 Dec 2016 01:12:32 +0100 (CET) Michał Mirosław wrote: > 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 I wonder if CFI can every validly be non-zero in the modern world, on Hyper-V. There are no token ring devices and that seems to be the only use case where CFI would be non-zero. Unless someone is planning to reuse it a a protocol bit which seems like a really bad idea. Maybe the right thing is to keep hard coded as zero and not start adding more untestable code conditions. My recommendation would be get rid of VLAN_TAG_PRESENT, but don't preserve CFI bit.
Re: [PATCH net-next 00/27] Remove VLAN CFI bit abuse
On Tue, Dec 13, 2016 at 05:16:26PM -0800, Stephen Hemminger wrote: > On Tue, 13 Dec 2016 01:12:32 +0100 (CET) > Michał Mirosław wrote: > > 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. > > I wonder if CFI can every validly be non-zero in the modern world, on Hyper-V. > There are no token ring devices and that seems to be the only use case where > CFI would > be non-zero. Unless someone is planning to reuse it a a protocol bit which > seems > like a really bad idea. > > Maybe the right thing is to keep hard coded as zero and not start adding > more untestable code conditions. > > My recommendation would be get rid of VLAN_TAG_PRESENT, but don't preserve > CFI bit. According to Wikipedia page [1] on 802.1Q, CFI bit got already changed to DEI (Drop eligible indicator) in 2011 revision of the IEEE standard. I can't verify this, though. Best Regards, Michał Mirosław [1] https://en.wikipedia.org/wiki/IEEE_802.1Q#Frame_format