[Bug 1137347] New: e1000e connects at 10Mb/s only
http://bugzilla.opensuse.org/show_bug.cgi?id=1137347 Bug ID: 1137347 Summary: e1000e connects at 10Mb/s only Classification: openSUSE Product: openSUSE Distribution Version: Leap 15.0 Hardware: Other OS: Other Status: NEW Severity: Normal Priority: P5 - None Component: Kernel Assignee: kernel-maintainers@forge.provo.novell.com Reporter: aeszter@gwdg.de QA Contact: qa-bugs@suse.de Found By: --- Blocker: --- Symptoms: on machines using the e1000e driver, ethtool reports 10Mb/s speed after booting. Pings are lossy (ping -f from the affected machine will show a few dropped packets per second), and ifconfig/ethtool show RX error accumulation. Calling ethtool -r <device> fixes this (i.e. 1000Mb/s are reported, pings aren't lost anymore, errors stop accumulating). If the network cable is unplugged for a few seconds, the connection will be back to 10Mb/s. Symptoms are similar to https://bugzilla.redhat.com/show_bug.cgi?id=1627816. The commit identified in comment 37 there is also present in Leap's kernel tree (92ca535, citing Bug #1087204), but I don't know how to determine the package version from there. I have a bunch of machines with 4.12.14-lp150.12.48 that do exhibit the problem and a single one with 4.12.14-lp150.12.22 that is working fine. -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.opensuse.org/show_bug.cgi?id=1137347
http://bugzilla.opensuse.org/show_bug.cgi?id=1137347#c2
Ansgar Esztermann
Could you try to blacklist mei_me module and see whether the behavior differs?
Tried that today, but no change observed. -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.opensuse.org/show_bug.cgi?id=1137347
http://bugzilla.opensuse.org/show_bug.cgi?id=1137347#c3
--- Comment #3 from Takashi Iwai
http://bugzilla.opensuse.org/show_bug.cgi?id=1137347
http://bugzilla.opensuse.org/show_bug.cgi?id=1137347#c4
--- Comment #4 from Ansgar Esztermann
http://bugzilla.opensuse.org/show_bug.cgi?id=1137347
http://bugzilla.opensuse.org/show_bug.cgi?id=1137347#c5
Takashi Iwai
http://bugzilla.opensuse.org/show_bug.cgi?id=1137347
http://bugzilla.opensuse.org/show_bug.cgi?id=1137347#c6
--- Comment #6 from Benjamin Poirier
Symptoms: on machines using the e1000e driver, ethtool reports 10Mb/s speed after booting. Pings are lossy (ping -f from the affected machine will show a few dropped packets per second), and ifconfig/ethtool show RX error accumulation.
Which error counters are increasing exactly? -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.opensuse.org/show_bug.cgi?id=1137347
http://bugzilla.opensuse.org/show_bug.cgi?id=1137347#c7
--- Comment #7 from Ansgar Esztermann
Sounds like it might be hardware-related.
It might be helpful to know the exact chipset revision on each adater and whether it is affected or not. The driver prints a line like this at probe time: e1000e 0000:00:19.0 eth0: MAC: 11, PHY: 12, PBA No: 1000FF-0FF Please check if each of your hosts are identical in this regard (MAC, PHY and PBA). The affected hosts are all the same model, bought at the same time, so they are indeed identical:
e1000e 0000:00:1f.6 eth0: MAC: 12, PHY: 12, PBA No: FFFFFF-0FF The one host not affected is an older model, so this could explain the difference.
I would also guess that this is related to the link partner (the switch or whatever device the host is connected to). How many different link partners have you tried?
Two switch models, but both are HP (E3800, 2848).
Please look at the following and provide some examples when things don't work/do work: 1) Which link settings does `ethtool` show on the host?
attachments follow
2) What does the driver print ("e1000e: lom0 NIC Link is Up 1000 Mbps Full Duplex, Flow Control: Rx/Tx")?
working: e1000e: eth0 NIC Link is Up 1000 Mbps Full Duplex, Flow Control: None not working: e1000e: eth0 NIC Link is Up 10 Mbps Full Duplex, Flow Control: None
3) Which link settings does the partner show? Do they match the host?
No, they don't. The switch shows 1000 Mbps Full Duplex in both cases, only MDI vs MDIX changes. Very puzzling.
I would say that does not affect link establishment so the difference in behavior that you see stems from something else.
As mentioned above, the machine is a slightly older version, so the difference may be in the hardware. Boot messages are unavailable right now, but I'll supply them later.
(In reply to Ansgar Esztermann from comment #0)
Symptoms: on machines using the e1000e driver, ethtool reports 10Mb/s speed after booting. Pings are lossy (ping -f from the affected machine will show a few dropped packets per second), and ifconfig/ethtool show RX error accumulation.
Which error counters are increasing exactly?
The first number in the "RX errors" line shown by ifconfig; however, I cannot reproduce that now. -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.opensuse.org/show_bug.cgi?id=1137347
http://bugzilla.opensuse.org/show_bug.cgi?id=1137347#c8
--- Comment #8 from Ansgar Esztermann
http://bugzilla.opensuse.org/show_bug.cgi?id=1137347
http://bugzilla.opensuse.org/show_bug.cgi?id=1137347#c9
--- Comment #9 from Ansgar Esztermann
http://bugzilla.opensuse.org/show_bug.cgi?id=1137347
Björn Voigt
participants (2)
-
bugzilla_noreply@novell.com
-
bugzilla_noreply@suse.com