Bug 262093 - ixl(4): RX packet errors on Intel X710 after 12.2 to 12.3-RELEASE upgrade
Summary: ixl(4): RX packet errors on Intel X710 after 12.2 to 12.3-RELEASE upgrade
Status: Open
Alias: None
Product: Base System
Classification: Unclassified
Component: kern (show other bugs)
Version: 12.3-RELEASE
Hardware: amd64 Any
: --- Affects Only Me
Assignee: freebsd-net (Nobody)
URL:
Keywords: needs-qa
Depends on:
Blocks:
 
Reported: 2022-02-21 13:40 UTC by Cristian Cardoso
Modified: 2023-06-20 13:58 UTC (History)
8 users (show)

See Also:
koobs: maintainer-feedback? (kbowling)
koobs: mfc-stable13?
koobs: mfc-stable12?


Attachments
dmesg.boot (25.65 KB, text/plain)
2022-02-22 11:14 UTC, Cristian Cardoso
no flags Details
pciconf (48.85 KB, text/plain)
2022-02-22 11:15 UTC, Cristian Cardoso
no flags Details
rc-conf (2.19 KB, text/plain)
2022-02-22 11:23 UTC, Cristian Cardoso
no flags Details

Note You need to log in before you can comment on or make changes to this bug.
Description Cristian Cardoso 2022-02-21 13:40:08 UTC
After updating FreeBSD 12.2-RELEASE to version 12.3-RELEASE I detected an error increment in the Intel X710 interface on my router

I'm using the driver provided by the system natively IXL.

I already changed the SFP+ modules and optical cable, the error problem persists.

Error increment issue started after reboot to complete version upgrade.

# ifconfig -v ixl0
ixl0: flags=8843<UP,BROADCAST,RUNNING,SIMPLEX,MULTICAST> metric 0 mtu 1500
        options=e500bb<RXCSUM,TXCSUM,VLAN_MTU,VLAN_HWTAGGING,JUMBO_MTU,VLAN_HWCSUM,VLAN_HWFILTER,VLAN_HWTSO,RXCSUM_IPV6,TXCSUM_IPV6>
        ether f8:f2:1e:7c:52:60
        media: Ethernet autoselect (10Gbase-SR <full-duplex>)
        status: active
        nd6 options=21<PERFORMNUD,AUTO_LINKLOCAL>
        plugged: SFP/SFP+/SFP28 10G Base-SR (LC)
        vendor: Intel Corp PN: FTLX8574D3BCV-IT SN: N52E8V6 DATE: 2016-05-02
        module temperature: 34.32 C Voltage: 3.36 Volts
        RX: 0.54 mW (-2.66 dBm) TX: 0.57 mW (-2.44 dBm)


ixl0@pci0:130:0:0:      class=0x020000 card=0x00068086 chip=0x15728086 rev=0x02 hdr=0x00
    vendor     = 'Intel Corporation'
    device     = 'Ethernet Controller X710 for 10GbE SFP+'
    class      = network
    subclass   = ethernet
ixl1@pci0:130:0:1:      class=0x020000 card=0x00008086 chip=0x15728086 rev=0x02 hdr=0x00
    vendor     = 'Intel Corporation'
    device     = 'Ethernet Controller X710 for 10GbE SFP+'
    class      = network
    subclass   = ethernet
Comment 1 Cristian Cardoso 2022-02-21 13:43:17 UTC
# netstat -w 2 -q 5 -I ixl0
            input           ixl0           output
   packets  errs idrops      bytes    packets  errs      bytes colls
    310759     0     0   91640961     600107     0  769746920     0
    333019     1     0   91494425     640776     0  831636870     0
    336276     2     0   88746682     661940     0  866460583     0
    330508     2     0   96604908     635893     0  823980828     0
    230993     0     0   58178558     458670     0  596894902     0
Comment 2 Kubilay Kocak freebsd_committer freebsd_triage 2022-02-22 00:42:22 UTC
Thank you for your report Christian. Could you please include additional information including:

- /var/run/dmesg.boot output (as an attachment)
- pciconf -lv output (as an attachment)
- /etc/rc.conf network configuration (sanitized if necessary)

Additionally, are you able to test a 13.0-RELEASE or 14.0-CURRENT snapshot or live image?
Comment 3 Krzysztof Galazka 2022-02-22 10:16:19 UTC
Could you, please, add also sysctl counters? (e.g. sysctl dev.ixl.0.mac)

Most interesting are those:
dev.ixl.0.mac.checksum_errors: 0
dev.ixl.0.mac.rx_jabber: 0
dev.ixl.0.mac.rx_oversized: 0
dev.ixl.0.mac.rx_fragmented: 0
dev.ixl.0.mac.rx_undersize: 0
dev.ixl.0.mac.rx_length_errors: 0
dev.ixl.0.mac.remote_faults: 0
dev.ixl.0.mac.local_faults: 0
dev.ixl.0.mac.illegal_bytes: 0
dev.ixl.0.mac.crc_errors: 0
dev.ixl.0.mac.rx_discards: 0
Comment 4 Cristian Cardoso 2022-02-22 11:12:02 UTC
# sysctl dev.ixl.0.mac
dev.ixl.0.mac.xoff_recvd: 0
dev.ixl.0.mac.xoff_txd: 0
dev.ixl.0.mac.xon_recvd: 0
dev.ixl.0.mac.xon_txd: 0
dev.ixl.0.mac.tx_frames_big: 0
dev.ixl.0.mac.tx_frames_1024_1522: 31774677076
dev.ixl.0.mac.tx_frames_512_1023: 462780575
dev.ixl.0.mac.tx_frames_256_511: 578986010
dev.ixl.0.mac.tx_frames_128_255: 597835366
dev.ixl.0.mac.tx_frames_65_127: 3523090578
dev.ixl.0.mac.tx_frames_64: 1365035895
dev.ixl.0.mac.checksum_errors: 201977
dev.ixl.0.mac.rx_jabber: 0
dev.ixl.0.mac.rx_oversized: 0
dev.ixl.0.mac.rx_fragmented: 1
dev.ixl.0.mac.rx_undersize: 0
dev.ixl.0.mac.rx_frames_big: 0
dev.ixl.0.mac.rx_frames_1024_1522: 3392606266
dev.ixl.0.mac.rx_frames_512_1023: 448559537
dev.ixl.0.mac.rx_frames_256_511: 565210746
dev.ixl.0.mac.rx_frames_128_255: 854644724
dev.ixl.0.mac.rx_frames_65_127: 11385726957
dev.ixl.0.mac.rx_frames_64: 5486712279
dev.ixl.0.mac.rx_length_errors: 1
dev.ixl.0.mac.remote_faults: 2
dev.ixl.0.mac.local_faults: 2
dev.ixl.0.mac.illegal_bytes: 5
dev.ixl.0.mac.crc_errors: 51
dev.ixl.0.mac.bcast_pkts_txd: 2
dev.ixl.0.mac.mcast_pkts_txd: 96001
dev.ixl.0.mac.ucast_pkts_txd: 38302309491
dev.ixl.0.mac.good_octets_txd: 47708270163324
dev.ixl.0.mac.rx_discards: 0
dev.ixl.0.mac.bcast_pkts_rcvd: 793906
dev.ixl.0.mac.mcast_pkts_rcvd: 163342
dev.ixl.0.mac.ucast_pkts_rcvd: 22132503261
dev.ixl.0.mac.good_octets_rcvd: 6838145932076
Comment 5 Cristian Cardoso 2022-02-22 11:14:03 UTC
Created attachment 232004 [details]
dmesg.boot
Comment 6 Cristian Cardoso 2022-02-22 11:15:32 UTC
Created attachment 232005 [details]
pciconf
Comment 7 Cristian Cardoso 2022-02-22 11:23:19 UTC
Created attachment 232006 [details]
rc-conf
Comment 8 Cristian Cardoso 2022-02-22 11:24:17 UTC
(In reply to Kubilay Kocak from comment #2)
I inserted the files as requested.
Thank you for your help.
Comment 9 Krzysztof Galazka 2022-03-14 14:13:37 UTC
Thanks Cristian! I think, what you observed may be a side effect of this change:

https://reviews.freebsd.org/D27639

It was decided to include L3/L4 checksum errors in the error counter. Could you, please, confirm if number of errors reported by netstat is the same as value of dev.ixl.0.mac.checksum_errors sysctl?

Thanks,
Krzysiek
Comment 10 Cristian Cardoso 2022-03-14 14:20:03 UTC
(In reply to Krzysztof Galazka from comment #9)
Hi

[root@r002 ~]# sysctl dev.ixl.0.mac | grep checksum_errors
dev.ixl.0.mac.checksum_errors: 5552778

[root@r002 ~]# netstat -w 2 -q 5 -I ixl0
            input           ixl0           output
   packets  errs idrops      bytes    packets  errs      bytes colls
    339138     1     0  121852802     635035     0  815630324     0
    335658     2     0  114983153     642778     0  825897319     0
    275845     0     0   92767355     519756     0  674856436     0
    335447     2     0  124827607     651042     0  838541951     0
    360714     0     0  141314326     666179     0  858674839     0

The errors are usually not the same in number, the checksum_errors counter is much higher since the machine was booted.
Comment 11 Cristian Cardoso 2022-05-12 04:16:17 UTC
I upgraded to version 13.0-RELEASE-p11, as Kubilay had suggested, but unfortunately the checksum errors continue.
Comment 12 Peter Vanek 2022-09-12 14:23:36 UTC
Hello Krzysztof,

I had discussion with multiple engineers if this enhancement of UDP traffic with disabled checksum (value 0) should be reported into input error counter.

Per my initial understanding, checksum = 0 means do not check;

Cam you comment more on motivation behind that?

Peter Vanek
Comment 13 wooyu 2023-06-20 12:36:32 UTC
I solved it by updating my x710's firmware to 12. 
Perhaps your x710 firmware is outdated
Comment 14 Cristian Cardoso 2023-06-20 13:58:00 UTC
(In reply to wooyu from comment #13)
Hi wooyu, at the time I reported this bug there was no firmware update available for this intel card, so I didn't have much to do other than file a bug report. I had the bug occurring until version 13.0 and then the equipment where I ran baremetal FreeBSD was discontinued, so I no longer have a way to validate what you said.