=EF=BB=BF =20 =20 =20

Hi,

thanks for your suggestion. Turns out, when i unplug the vlan interface from the bridge and put the ip address on the vlan interface, as you suggested, things start to work, e.g. arp resolves.

as soon as i put the ip and the vlan interface back on the bridge, things stop again. so where does this lead me? the problem is not in the vlan handling, but on the bridge?

I started playing with the net.link.bridge sysctls and indeed, when i set

# sysctl net.link.bridge.inherit_mac=3D1

and then recreate the bridge

# ifconfig bridge0 deletem ixl3.15 deletem vnet0.1
# ifconfig bridge0 addm ixl3.15 addm vnet0.1

with the ip address still on bridge0 and ixl3, ixl3.15 and bridge0 all sharing the same mac address, arp starts resolving. but only for requests sent from the bridge0 interface. inside of the jail things still don't work (as the vnet interface again has another mac address).

# ifconfig ixl3
ixl3: flags=3D28963<UP,BROADCAST,RUNNING,PROMISC,SIMPLEX,MULTICAST> metric 0 mtu 1500
        options=3D4a500b9<RXCSUM,VLAN_MTU,VLAN_HWTAGGING,JUMBO_MTU,VLAN_HWCSUM,VL= AN_HWFILTER,VLAN_HWTSO,RXCSUM_IPV6,NOMAP>
        ether a4:bf:01:76:ef:9d
=         media: Ethernet autoselect (= 10Gbase-SR <full-duplex>)
        status: active
        nd6 options=3D29<PERFORM= NUD,IFDISABLED,AUTO_LINKLOCAL>
# ifconfig ixl3.15
ixl3.15: flags=3D8943<UP,BROADCAST,RUNNING,PROMISC,SIMPLEX,MULTICAST> metric 0 mtu 1500
        options=3D4200001<RXCSUM= ,RXCSUM_IPV6,NOMAP>
        ether a4:bf:01:76:ef:9d
=         groups: vlan
        vlan: 15 vlanproto: 802.1q v= lanpcp: 0 parent interface: ixl3
        media: Ethernet autoselect (= 10Gbase-SR <full-duplex>)
        status: active
        nd6 options=3D29<PERFORM= NUD,IFDISABLED,AUTO_LINKLOCAL>
# ifconfig bridge0
bridge0: flags=3D8843<UP,BROADCAST,RUNNING,SIMPLEX,MULTICAST> metric 0 mtu 1500
        ether a4:bf:01:76:ef:9d
=         inet 192.168.55.20 netmask 0= xffffff00 broadcast 192.168.55.255
        id 00:00:00:00:00:00 priori= ty 32768 hellotime 2 fwddelay 15
        maxage 20 holdcnt 6 proto r= stp maxaddr 2000 timeout 1200
        root id 00:00:00:00:00:00 p= riority 32768 ifcost 0 port 0
        member: vnet0.1 flags=3D143<LEARNING,DISCOVER,AUTOEDGE,AUTOPTP>
           &nbs= p;    ifmaxaddr 0 port 9 priority 128 path cost 2000
        member: ixl3.15 flags=3D143<LEARNING,DISCOVER,AUTOEDGE,AUTOPTP>
           &nbs= p;    ifmaxaddr 0 port 8 priority 128 path cost 2000
        groups: bridge
        nd6 options=3D9<PERFORMN= UD,IFDISABLED>
# ping 192.168.55.1
PING 192.168.55.1 (192.168.55.1): 56 data bytes
^C
--- 192.168.55.1 ping statistics ---
2 packets transmitted, 0 packets received, 100.0% packet loss

[! yes, the host does not answer on ICMP, but that is to be expected !]

# arp -an
? (192.168.55.20) at a4:bf:01:76:ef:9d on bridge0 permanent [bridge]
? (192.168.55.1) at b8:27:eb:47:8f:43 on bridge0 expires in 1197 seconds [bridge]
[...]

[! into the jail !]

JAIL # ifconfig epair0b
epair0b: flags=3D8863<UP,BROADCAST,RUNNING,SIMPLEX,MULTICAST> metric 0 mtu 1500
        options=3D8<VLAN_MTU>=
        ether ac:16:2d:bd:b7:34
=         hwaddr 02:51:73:d1:33:0b         inet 192.168.55.10 netmask 0= xffffff00 broadcast 192.168.55.255
        inet6 fe80::ae16:2dff:febd:= b734%epair0b prefixlen 64 scopeid 0x2
        groups: epair
        media: Ethernet 10Gbase-T (= 10Gbase-T <full-duplex>)
        status: active
        nd6 options=3D21<PERFORM= NUD,AUTO_LINKLOCAL>
JAIL # ping 192.168.55.1
PING 192.168.55.1 (192.168.55.1): 56 data bytes
^C
--- 192.168.55.1 ping statistics ---
1 packets transmitted, 0 packets received, 100.0% packet loss
JAIL # arp -an
? (192.168.55.10) at ac:16:2d:bd:b7:34 on epair0b permanent [ethernet]
? (192.168.55.1) at (incomplete) on epair0b expired [ethernet]

I conclude that there must be some mac address filtering going on in the data path, whether its on ixl or the bridge.

In dmesg I also see:

>> bridge0: can't disable some capabilities on ixl3.15: 0x400

but as of /usr/src/sys/net/if.h:233 this maps to IFCAP_LRO which afaik should not have any influence on L2 filtering.

Have to say, I'm out of ideas again. Never had something like this. So far just 'throwing interfaces on a bridge' worked in the past. Any ideas where to look next?

Thanks a lot & best

Daniel

On 1/25/24 08:22, Zhenlei Huang wrote:

On Jan 23, 2024, at 11:03 PM, Daniel <freebsd-net@c0decafe.de> wrote:

Hi List,

just recently I discovered a problem with the ixl(4) driver. Hopefully someone here can help me. my setup is as follows:

Network ----- ixl3 interface ----- ixl3.15 vlan interface ----- bridge0 ----- vnet0.1 to jail

the problem now is that the jail can send data out (arp requests), i do see the responses on the ixl3 interface of the host, but they never make their way up to the ixl3.15 vlan interface (even though they are tagged correctly). To rule out that my config or the network is the cruel pit i did test the same setup with a cheap usb-ethernet adapter and there everything works as expected. I'm on FreeBSD 13.2-RELEASE-p8 and I did test both, the in kernel driver and the driver from ports intel-ixl-kmod-1.13.4_1.

I would encourage you to do plain VLAN tests, i.e. plug ixl3.15 out from bridge0

```

# ifconfig bridge0 deletem ixl3.15

# ifconfig bridge0 inet 192.168.55.20/24 delete ### to prevent confusion

# ifconfig ixl3.15 inet 192.168.55.x/24

# ping -c1 192.168.55.1

```

Here is a bit of information on my environment:

# uname -a
FreeBSD mimir 13.2-RELEASE-p8 FreeBSD 13.2-RELEASE-p8 GENERIC amd64

# pciconf -lBbcevV pci0:25:0:3
ixl3@pci0:25:0:3:       class=3D= 0x020000 rev=3D0x09 hdr=3D0x00 vendor=3D0x8086 device=3D0x37d3 subvendor=3D0x8086 subdevice=3D0x35d5
    vendor     =3D 'Intel= Corporation'
    device     =3D 'Ether= net Connection X722 for 10GbE SFP+'
    class      =3D n= etwork
    subclass   =3D ethernet
    bar   [10] =3D type Prefetchabl= e Memory, range 64, base 0xb0000000, size 16777216, enabled
    bar   [1c] =3D type Prefetchabl= e Memory, range 64, base 0xb5000000, size 32768, enabled
    cap 01[40] =3D powerspec 3 supports= D0 D3 current D0
    cap 05[50] =3D MSI supports 1 message, 64= bit, vector masks
    cap 11[70] =3D MSI-X supports 129 message= s, enabled
          =        Table in map 0x1c[0x0], PBA in map 0x1c[0x1000]
    cap 10[a0] =3D PCI-Express 2 endpoint max= data 256(512) FLR RO
          =        max read 512
          =        link x1(x1) speed 2.5(2.5) ASPM disabled(L0s/L1)
    cap 03[e0] =3D VPD
    ecap 0001[100] =3D AER 2 0 fatal 0 non-fa= tal 1 corrected
    ecap 0003[140] =3D Serial 1 9aef76ffff01b= fa4
    ecap 000e[150] =3D ARI 1
    ecap 0010[160] =3D SR-IOV 1 IOV disabled,= Memory Space disabled, ARI disabled
          =            0 VFs configure= d out of 32 supported
          =            First VF RID Of= fset 0x006d, VF RID Stride 0x0001
          =            VF Device ID 0x= 37cd
          =            Page Sizes: 409= 6 (enabled), 8192, 65536, 262144, 1048576, 4194304
    ecap 0017[1a0] =3D TPH Requester 1
    ecap 000d[1b0] =3D ACS 1 Source Validatio= n unavailable, Translation Blocking unavailable
          =            P2P Req Redirec= t unavailable, P2P Cmpl Redirect unavailable
          =            P2P Upstream Fo= rwarding unavailable, P2P Egress Control unavailable
          =            P2P Direct Tran= slated unavailable, Enhanced Capability unavailable
PCI-e errors =3D Correctable Error Detected
          =        Unsupported Request Detected
     Corrected =3D Advisory Non-Fatal Er= ror
    VPD ident =3D 'Example VPD'

# ifconfig
[...]
ixl3: flags=3D8963<UP,BROADCAST,RUNNING,PROMISC,SIMPLEX,MULTICA= ST> metric 0 mtu 1500
        options=3D4a500b9<RXCSUM,VLAN_MTU,VLAN_HWTAGGING,JUMBO_MTU,VLAN_HWCSUM,VL= AN_HWFILTER,VLAN_HWTSO,RXCSUM_IPV6,NOMAP>
        ether a4:bf:01:76= :ef:9d
        media: Ethernet a= utoselect (10Gbase-SR <full-duplex>)
        status: active         nd6 options=3D29<PERFORMNUD,IFDISABLED,AUTO_LINKLOCAL>
ixl3.15: flags=3D8942<BROADCAST,RUNNING,PROMISC,SIMPLEX,MULTICAST&= gt; metric 0 mtu 1500
        options=3D4200001= <RXCSUM,RXCSUM_IPV6,NOMAP>
        ether a4:bf:01:76= :ef:9d
        groups: vlan
        vlan: 15 vlanprot= o: 802.1q vlanpcp: 0 parent interface: ixl3
        media: Ethernet a= utoselect (10Gbase-SR <full-duplex>)
        status: active         nd6 options=3D29<PERFORMNUD,IFDISABLED,AUTO_LINKLOCAL>
bridge0: flags=3D8843<UP,BROADCAST,RUNNING,SIMPLEX,MULTICAST> metric 0 mtu 1500
        ether 58:9c:fc:10= :dd:05
        inet 192.168.55.2= 0 netmask 0xffffff00 broadcast 192.168.55.255
        id 00:00:00:00:00= :00 priority 32768 hellotime 2 fwddelay 15
        maxage 20 holdcnt= 6 proto rstp maxaddr 2000 timeout 1200
        root id 00:00:00:= 00:00:00 priority 32768 ifcost 0 port 0
        member: vnet0.1 flags=3D143<LEARNING,DISCOVER,AUTOEDGE,AUTOPTP>
          =       ifmaxaddr 0 port 9 priority 128 path cost 2000
        member: ixl3.15 flags=3D143<LEARNING,DISCOVER,AUTOEDGE,AUTOPTP>
          =       ifmaxaddr 0 port 8 priority 128 path cost 55
        groups: bridge         nd6 options=3D9&l= t;PERFORMNUD,IFDISABLED>
[...]

# cat /etc/rc.conf
[...]
ifconfig_ixl3=3D"up"
vlans_ixl3=3D"15"
cloned_interfaces=3D"bridge0"
ifconfig_bridge0=3D"addm ixl3.15 up"
[...]

# dmesg | grep ixl
ixl0: <Intel(R) Ethernet Connection X722 for 10GBASE-T - 2.3.3-k> mem 0xb3000000-0xb3ffffff,0xb5018000-0xb501ffff at device 0.0 numa-domain 0 on pci6
ixl0: fw 3.1.55727 api 1.5 nvm 3.31 etid 80000d32 oem 1.262.0
ixl0: PF-ID[0]: VFs 32, MSI-X 129, VF MSI-X 5, QPs 384, MDIO shared
ixl0: Using 1024 TX descriptors and 1024 RX descriptors
ixl0: Using 12 RX queues 12 TX queues
ixl0: Using MSI-X interrupts with 13 vectors
= ixl0: Ethernet address: a4:bf:01:76:ef:9a
ixl0: Allocating 16 queues for PF LAN VSI; 12 queues active
ixl0: SR-IOV ready
ixl0: netmap queues/slots: TX 12/1024, RX 12/1024
ixl1: <Intel(R) Ethernet Connection X722 for 10GBASE-T - 2.3.3-k> mem 0xb2000000-0xb2ffffff,0xb5010000-0xb5017fff at device 0.1 numa-domain 0 on pci6
ixl1: fw 3.1.55727 api 1.5 nvm 3.31 etid 80000d32 oem 1.262.0
ixl1: PF-ID[1]: VFs 32, MSI-X 129, VF MSI-X 5, QPs 384, MDIO shared
ixl1: Using 1024 TX descriptors and 1024 RX descriptors
ixl1: Using 12 RX queues 12 TX queues
ixl1: Using MSI-X interrupts with 13 vectors
= ixl1: Ethernet address: a4:bf:01:76:ef:9b
ixl1: Allocating 16 queues for PF LAN VSI; 12 queues active
ixl1: SR-IOV ready
ixl1: netmap queues/slots: TX 12/1024, RX 12/1024
ixl2: <Intel(R) Ethernet Connection X722 for 10GbE SFP+ - 2.3.3-k> mem 0xb1000000-0xb1ffffff,0xb5008000-0xb500ffff at device 0.2 numa-domain 0 on pci6
ixl2: fw 3.1.55727 api 1.5 nvm 3.31 etid 80000d32 oem 1.262.0
ixl2: PF-ID[2]: VFs 32, MSI-X 129, VF MSI-X 5, QPs 384, I2C
ixl2: Using 1024 TX descriptors and 1024 RX descriptors
ixl2: Using 12 RX queues 12 TX queues
ixl2: Using MSI-X interrupts with 13 vectors
= ixl2: Ethernet address: a4:bf:01:76:ef:9c
ixl2: Allocating 16 queues for PF LAN VSI; 12 queues active
ixl2: ixl_set_link: Error getting phy capabilities -7, aq error: 5
ixl2: SR-IOV ready
ixl2: netmap queues/slots: TX 12/1024, RX 12/1024
ixl3: <Intel(R) Ethernet Connection X722 for 10GbE SFP+ - 2.3.3-k> mem 0xb0000000-0xb0ffffff,0xb5000000-0xb5007fff at device 0.3 numa-domain 0 on pci6
ixl3: fw 3.1.55727 api 1.5 nvm 3.31 etid 80000d32 oem 1.262.0
ixl3: PF-ID[3]: VFs 32, MSI-X 129, VF MSI-X 5, QPs 384, I2C
ixl3: Using 1024 TX descriptors and 1024 RX descriptors
ixl3: Using 12 RX queues 12 TX queues
ixl3: Using MSI-X interrupts with 13 vectors
= ixl3: Ethernet address: a4:bf:01:76:ef:9d
ixl3: Allocating 16 queues for PF LAN VSI; 12 queues active
ixl3: ixl_set_link: Error getting phy capabilities -7, aq error: 5
ixl3: SR-IOV ready
ixl3: netmap queues/slots: TX 12/1024, RX 12/1024
ixl2: Link is up, 10 Gbps Full Duplex, Requested FEC: None, Negotiated FEC: None, Autoneg: False, Flow Control: None
ixl2: link state changed to UP
ixl3: Link is up, 10 Gbps Full Duplex, Requested FEC: None, Negotiated FEC: None, Autoneg: False, Flow Control: None
ixl3: link state changed to UP
bridge0: can't disable some capabilities on ixl3.15: 0x400
ixl3: promiscuous mode enabled
ixl3.15: promiscuous mode enabled

from my packet traces:

# tcpdump -vvv -i ixl3

May you please add the option -e to tcpdump, so that the link-level header can be printed out.

```

# tcpdump -nvei ixl3

```

tcpdump: listening on ixl3, link-type EN10MB (Ethernet), capture size 262144 bytes
[...]
13:36:20.155843 ARP, Ethernet (len 6), IPv4 (len 4), Request who-has 192.168.55.1 tell 192.168.55.10, length 28
13:36:20.156285 ARP, Ethernet (len 6), IPv4 (len 4), Reply 192.168.55.1 is-at b8:27:eb:47:8f:43 (oui Unknown), length 46
13:36:21.169003 ARP, Ethernet (len 6), IPv4 (len 4), Request who-has 192.168.55.1 tell 192.168.55.10, length 28
13:36:21.169538 ARP, Ethernet (len 6), IPv4 (len 4), Reply 192.168.55.1 is-at b8:27:eb:47:8f:43 (oui Unknown), length 46

Here the answer can be see, its tagged with 802.1q tag 15

# tcpdump -vvv -i ixl3.15
tcpdump: listening on ixl3.15, link-type EN10MB (Ethernet), capture size 262144 bytes
14:14:37.255429 ARP, Ethernet (len 6), IPv4 (len 4), Request who-has 192.168.55.1 tell 192.168.55.10, length 28
14:14:42.263475 ARP, Ethernet (len 6), IPv4 (len 4), Request who-has 192.168.55.1 tell 192.168.55.10, length 28
14:15:02.556311 ARP, Ethernet (len 6), IPv4 (len 4), Request who-has 192.168.55.1 tell 192.168.55.10, length 28
14:15:07.557644 ARP, Ethernet (len 6), IPv4 (len 4), Request who-has 192.168.55.1 tell 192.168.55.10, length 28

The answer cannot be seen on the VLAN interface ):

I hope the list can help me out here, as I am lost.

Thanks & best

Daniel

Best regards,

Zhenlei

=20