Bug 217640
Summary: | 10G SFP+ work at 3.41 Gbits/sec on multiple hardware | ||
---|---|---|---|
Product: | Networking | Reporter: | alpha_one_x86 |
Component: | Other | Assignee: | Stephen Hemminger (stephen) |
Status: | REOPENED --- | ||
Severity: | high | CC: | bagasdotme |
Priority: | P3 | ||
Hardware: | All | ||
OS: | Linux | ||
Kernel Version: | Subsystem: | ||
Regression: | No | Bisected commit-id: | |
Attachments: |
dmesg
lspci |
Description
alpha_one_x86
2023-07-06 23:23:02 UTC
(In reply to alpha_one_x86 from comment #0) > Hi, > Since I upgrade linux kernel from 5.10.137 to 6.1.38 all my interface of > macchiatobin work at 100M (90 Mbits/sec detected by iperf) > ethtool detect the link at correct speed (10G for SFP+ and 1Gbps for > ethernet) > What can be the regression? Where is your dmesg? And lspci? Also, and the most important, can you do bisection to find the culprit? Please refer to Documentation/admin-guide/bug-bisect.rst for how. Created attachment 304557 [details]
dmesg
Created attachment 304558 [details]
lspci
I doing bisect, but will take me a time The problem was fix on reboot, then very hard to reproduce if I need wait month to bug If that can help with your bisect, mine which is still on 5.15.109 has never exhibited this problem. I expected to upgrade to 6.1 soon, if I encounter it I'll update this entry. I use eth2, eth1, eth3, or SPF+ SFP and ethernet, and you? Any one is able to do speedtest at 10G via SFP+? I have unstable rate around 3.26 Gbits/sec. Strangely I have same speed limit (3.41Gbits/sec very stable) into core Intel Core i5-12400 with Intel Corporation Ethernet Controller X710/Broadcom Inc. and subsidiaries NetXtreme II BCM57810 in my usage (with fresh install all worked at 10G with iperf3) Then I reopen this bug, now I have more time to diag it, but how detect the problem? |