Summary: | mlx5_core0 ERR | ||
---|---|---|---|
Product: | Base System | Reporter: | Michael <michael.adm> |
Component: | kern | Assignee: | freebsd-net (Nobody) <net> |
Status: | Open --- | ||
Severity: | Affects Many People | CC: | hselasky, kib, lwhsu |
Priority: | --- | ||
Version: | 12.1-STABLE | ||
Hardware: | amd64 | ||
OS: | Any |
Description
Michael
2019-12-26 10:19:40 UTC
This error should go away if you upgrade the firmware I believe. --HPS (In reply to Hans Petter Selasky from comment #1) https://support.hpe.com/hpsc/swd/public/detail?swItemId=MTX_2b1b10070f40416cbc3f997ff0#tab4 Version: 1.0.0.11(20 Dec 2019) FW version 14.26.1040 This is the newest firmware I have already installed. (In reply to Hans Petter Selasky from comment #1) I suspect the error would not go away. Am I right that the device is in pass-through mode on Hyper-V, and FreeBSD runs inside a VM ? If yes, then most likely hypervisor (rightfully) filters out unknown capabilities. (In reply to Konstantin Belousov from comment #3) Yes, FreeBSD is truly virtualized in Hyper-V. Without this, I would go broke on servers and electricity. Well, maybe the specifics of the vendor are not scary that they are filtered out, but I would like to hear the opinion "does this pose a problem?" or I can ignore this, especially this message hn0: mce0 SIOCGRSSKEY failed: 22 By the way, there are no errors in virtualized Windows systems. Later I will look at virtualized Linux for such errors. A commit references this bug: Author: hselasky Date: Mon Dec 30 09:22:53 UTC 2019 New revision: 356201 URL: https://svnweb.freebsd.org/changeset/base/356201 Log: Fix spelling. PR: 242891 MFC after: 1 week Sponsored by: Mellanox Technologies Changes: head/sys/dev/hyperv/netvsc/if_hn.c Maybe this can clarify the problem. Messages of mlx5 in the Windows system log: HPE Ethernet 10/25Gb 2-port 640FLR-SFP28 Adapter #2 has got: vendor_id 15b3 device_id 1016 subvendor_id 1590 subsystem_id 00d3 HW revision 80 FW version 14.26.1040 port type ETH HPE Ethernet 10/25Gb 2-port 640FLR-SFP28 Adapter #2: Zero Touch RoCE: Some of the required capabilities are not supported by FW (Required 7 Supported 5). HPE Ethernet 10/25Gb 2-port 640FLR-SFP28 Adapter #2: Zero Touch RoCE is supported. Current configuration is: Slow restart 0 TX window 0 Adp Retrans 0. HPE Ethernet 10/25Gb 2-port 640FLR-SFP28 Adapter #2: Zero Touch RoCE: New configuration is set: Slow restart 1 TX window 0 Adp Retrans 1. HPE Ethernet 10/25Gb 2-port 640FLR-SFP28 Adapter #2 device detects that the link is up, and has initiated a normal operation. HPE Ethernet 10/25Gb 2-port 640FLR-SFP28 Adapter #2: is currently running: GUID: 0015:5dff:fd00:f100 MAC: 00-15-5d-00-f1-00 HPE Ethernet 10/25Gb 2-port 640FLR-SFP28 Adapter #2: configuration updated: value ZeroTouchRoceCfgMask changed from 5 to 7. HPE Ethernet 10/25Gb 2-port 640FLR-SFP28 Adapter #2: is currently running: Driver Version: 2.30.21713.0 Firmware Version: 14.26.1040 PSID number: HP_2690110034 Eventid 46; Hyper-V-Netvsc; Message: NDK function message not received from VF adapter "\DEVICE\{CC634541-26A7-4DF0-B2D1-A4025E8351FE}". In Hyper-V log message are: Eventid 102; Microsoft-Windows-Hyper-V-VmSwitch; Message: Networking driver in VM1 is loaded and the protocol version is negotiated to the most recent version (Virtual machine ID 9B4A1702-666D-40BE-8E83-C280326BF300). A commit references this bug: Author: hselasky Date: Mon Jan 6 09:50:44 UTC 2020 New revision: 356410 URL: https://svnweb.freebsd.org/changeset/base/356410 Log: MFC r356201: Fix spelling. PR: 242891 Sponsored by: Mellanox Technologies Changes: _U stable/12/ stable/12/sys/dev/hyperv/netvsc/if_hn.c A commit references this bug: Author: hselasky Date: Mon Jan 6 09:51:22 UTC 2020 New revision: 356411 URL: https://svnweb.freebsd.org/changeset/base/356411 Log: MFC r356201: Fix spelling. PR: 242891 Sponsored by: Mellanox Technologies Changes: _U stable/11/ stable/11/sys/dev/hyperv/netvsc/if_hn.c A commit references this bug: Author: hselasky Date: Mon Jan 6 09:51:55 UTC 2020 New revision: 356412 URL: https://svnweb.freebsd.org/changeset/base/356412 Log: MFC r356201: Fix spelling. PR: 242891 Sponsored by: Mellanox Technologies Changes: _U stable/10/ stable/10/sys/dev/hyperv/netvsc/if_hn.c I did not find a more informative output of messages from different Linux. Everyone has a system log as some sort of chaotic nonsense. CentOS v8 garbage collection of messages when adding a network interface with SR-IOV enabled: /var/log/messages : nm-dispatcher[40509]: req:1 'dhcp4-change' [eth0]: new request (3 scripts) nm-dispatcher[40509]: req:1 'dhcp4-change' [eth0]: start running ordered scripts... systemd-udevd[40520]: link_config: autonegotiation is unset or enabled, the speed and duplex are not writable. NetworkManager[996]: <info> [1578320647.9579] manager: (eth1): new Ethernet device (/org/freedesktop/NetworkManager/Devices/5) NetworkManager[996]: <info> [1578320647.9795] device (eth1): state change: unmanaged -> unavailable (reason 'managed', sys-iface-state: 'external') kernel: IPv6: ADDRCONF(NETDEV_UP): eth1: link is not ready kernel: IPv6: ADDRCONF(NETDEV_UP): eth1: link is not ready NetworkManager[996]: <info> [1578320647.9954] device (eth1): state change: unavailable -> disconnected (reason 'none', sys-iface-state: 'managed') kernel: IPv6: ADDRCONF(NETDEV_UP): eth1: link is not ready kernel: IPv6: ADDRCONF(NETDEV_CHANGE): eth1: link becomes ready NetworkManager[996]: <info> [1578320648.1653] device (eth1): carrier: link connected kernel: hv_netvsc cc91b9de-cf8f-4729-8e30-bc4cf0e3f604 eth1: VF slot 2 added kernel: hv_vmbus: registering driver hv_pci kernel: hv_pci c2f22477-5994-443d-ae4b-7012efabe867: PCI VMBus probing: Using version 0x10002 kernel: hv_pci c2f22477-5994-443d-ae4b-7012efabe867: PCI host bridge to bus ae4b:00 kernel: pci_bus ae4b:00: root bus resource [mem 0xfe0000000-0xfe00fffff window] kernel: pci ae4b:00:02.0: enabling Extended Tags kernel: pci ae4b:00:02.0: 0.000 Gb/s available PCIe bandwidth, limited by Unknown speed x0 link at ae4b:00:02.0 (capable of 63.008 Gb/s with 8 GT/s x8 link) kernel: pci ae4b:00:02.0: BAR 0: assigned [mem 0xfe0000000-0xfe00fffff 64bit pref] systemd[1]: Starting Initialize the iWARP/InfiniBand/RDMA stack in the kernel... kernel: mlx5_core ae4b:00:02.0: firmware version: 14.26.1040 kernel: mlx5_core ae4b:00:02.0: handle_hca_cap:537:(pid 12693): log_max_qp value in current profile is 18, changing it to HCA capability limit (12) kernel: mlx5_core ae4b:00:02.0: MLX5E: StrdRq(0) RqSz(1024) StrdSz(64) RxCqeCmprss(0) kernel: hv_netvsc cc91b9de-cf8f-4729-8e30-bc4cf0e3f604 eth1: VF registering: eth2 kernel: mlx5_core ae4b:00:02.0 eth2: joined to eth1 NetworkManager[996]: <info> [1578320649.0723] manager: (eth2): new Ethernet device (/org/freedesktop/NetworkManager/Devices/6) NetworkManager[996]: <info> [1578320649.0726] device (eth2): enslaved to non-master-type device eth1; ignoring kernel: mlx5_ib: Mellanox Connect-IB Infiniband driver v5.0-0 systemd-udevd[40538]: link_config: autonegotiation is unset or enabled, the speed and duplex are not writable. kernel: mlx5_core ae4b:00:02.0 enP44619s2: renamed from eth2 NetworkManager[996]: <info> [1578320649.1427] device (eth2): interface index 6 renamed iface from 'eth2' to 'enP44619s2' NetworkManager[996]: <info> [1578320649.1431] device (enP44619s2): enslaved to non-master-type device eth1; ignoring NetworkManager[996]: <info> [1578320649.1543] device (enP44619s2): state change: unmanaged -> unavailable (reason 'managed', sys-iface-state: 'external') systemd[1]: Created slice system-rdma\x2dload\x2dmodules.slice. systemd[1]: Starting Load RDMA modules from /etc/rdma/modules/rdma.conf... systemd[1]: Starting Load RDMA modules from /etc/rdma/modules/infiniband.conf... systemd[1]: Starting Load RDMA modules from /etc/rdma/modules/roce.conf... systemd[1]: Starting RDMA Node Description Daemon... systemd[1]: Started RDMA Node Description Daemon. systemd[1]: Started Load RDMA modules from /etc/rdma/modules/roce.conf. kernel: Loading iSCSI transport class v2.0-870. kernel: mlx5_core ae4b:00:02.0 enP44619s2: Link up kernel: hv_netvsc cc91b9de-cf8f-4729-8e30-bc4cf0e3f604 eth1: Data path switched to VF: enP44619s2 NetworkManager[996]: <info> [1578320649.2483] device (enP44619s2): carrier: link connected NetworkManager[996]: <info> [1578320649.2562] device (enP44619s2): enslaved to non-master-type device eth1; ignoring NetworkManager[996]: <info> [1578320649.2564] device (enP44619s2): enslaved to non-master-type device eth1; ignoring NetworkManager[996]: <info> [1578320649.2565] device (enP44619s2): state change: unavailable -> disconnected (reason 'none', sys-iface-state: 'managed') systemd-modules-load[40640]: Inserted module 'ib_ipoib' systemd-modules-load[40640]: Inserted module 'ib_umad' NetworkManager[996]: <info> [1578320649.3233] device (enP44619s2): enslaved to non-master-type device eth1; ignoring systemd[1]: Started Load RDMA modules from /etc/rdma/modules/infiniband.conf. kernel: iscsi: registered transport (iser) systemd-modules-load[40639]: Inserted module 'ib_iser' systemd-modules-load[40639]: Inserted module 'rdma_ucm' kernel: RPC: Registered rdma transport module. kernel: RPC: Registered rdma backchannel transport module. systemd-modules-load[40639]: Inserted module 'rpcrdma' systemd[1]: Started Load RDMA modules from /etc/rdma/modules/rdma.conf. systemd[1]: Reached target RDMA Hardware. kernel: Rounding down aligned max_sectors from 4294967295 to 4294967288 kernel: db_root: cannot open: /etc/target kernel: ib_srpt MAD registration failed for mlx5_0-1. kernel: ib_srpt srpt_add_one(mlx5_0) failed. systemd[1]: Started Initialize the iWARP/InfiniBand/RDMA stack in the kernel. ModemManager[876]: <info> Couldn't check support for device '/sys/devices/LNXSYSTM:00/LNXSYBUS:00/ACPI0004:00/VMBUS:00/c2f22477-5994-443d-ae4b-7012efabe867/pciae4b:00/ae4b:00:02.0': not supported by any plugin I also see this on 12.1-R, while 12.0-R works. Can you test -stable? I've tested 12-STABLE (r362180) which doesn't work, while -CURRENT (r362186) works. They have similar dmesg: FreeBSD 12.1-STABLE r362180 GENERIC root@antest:~ # dmesg -a | grep mlx mlx5en: Mellanox Ethernet driver 3.5.2 (September 2019) mlx5_core0: <mlx5_core> at device 2.0 on pci1 mlx5_core0: ERR: mlx5_init_once:937:(pid 0): Unable to find vendor specific capabilities mlx5_core0: WARN: mlx5_fwdump_prep:76:(pid 0): Failed to find vendor-specific capability, error 2 mce0: ERR: mlx5e_ioctl:3303:(pid 0): tso6 disabled due to -txcsum6. mce0: ERR: mlx5e_ioctl:3323:(pid 0): enable txcsum6 first. root@antest:~ # dmesg -a | grep mce0 mce0: Ethernet address: 00:0d:3a:8a:b6:bd mce0: link state changed to DOWN mce0: ERR: mlx5e_ioctl:3303:(pid 0): tso6 disabled due to -txcsum6. mce0: ERR: mlx5e_ioctl:3323:(pid 0): enable txcsum6 first. mce0: link state changed to UP hn0: mce0 SIOCGIFRSSKEY failed: 22 Starting Network: lo0 hn0 mce0. mce0: flags=8a43<UP,BROADCAST,RUNNING,ALLMULTI,SIMPLEX,MULTICAST> metric 0 mtu 1500 FreeBSD 13.0-CURRENT r362186 GENERIC lwhsu@antest:~ > dmesg -a | grep mlx mlx5en: Mellanox Ethernet driver 3.5.2 (September 2019) mlx5_core0: <mlx5_core> at device 2.0 on pci1 mlx5_core0: ERR: mlx5_init_once:938:(pid 0): Unable to find vendor specific capabilities mlx5_core0: WARN: mlx5_fwdump_prep:76:(pid 0): Failed to find vendor-specific capability, error 2 mce0: ERR: mlx5e_ioctl:3341:(pid 0): tso6 disabled due to -txcsum6. mce0: ERR: mlx5e_ioctl:3367:(pid 0): enable txcsum6 first. lwhsu@antest:~ > dmesg -a | grep mce0 mce0: Ethernet address: 00:0d:3a:8a:b6:bd mce0: link state changed to DOWN mce0: ERR: mlx5e_ioctl:3341:(pid 0): tso6 disabled due to -txcsum6. mce0: ERR: mlx5e_ioctl:3367:(pid 0): enable txcsum6 first. mce0: link state changed to UP hn0: mce0 SIOCGIFRSSKEY failed: 22 Starting Network: lo0 hn0 mce0. mce0: flags=8a43<UP,BROADCAST,RUNNING,ALLMULTI,SIMPLEX,MULTICAST> metric 0 mtu 1500 I'm testing these on Azure with manually enabling accelerated networking: https://microsoft.github.io/AzureTipsAndTricks/blog/tip226.html |