Comment 8 for bug 1815808

Revision history for this message
Gene Olson (linutic) wrote : Re: [Bug 1815808] Re: Thunderbolt Ethernet high-traffic exception
  • usb.failed Edit (58.5 KiB, application/octet-stream; name="usb.failed")
  • usb.home Edit (93.3 KiB, application/octet-stream; name="usb.home")
  • usb.work Edit (89.6 KiB, application/octet-stream; name="usb.work")

I tried what you suggested. Ethtool reported no changes in configuration
when I turned off receive offload. It reported substantial changes when I
turned off transmit offload. After turning off transmit offload, transmit
rate dropped from 930 Mb/s to abut 660 Mb/s, although I did not see a big
CPU hit. Receive speeds were not affected by turning off receive offload.

gene-pc:~$ sudo ethtool --offload eth0 rx off

gene-pc:~$ sudo ethtool --offload eth0 tx off
Actual changes:
tx-checksumming: off
tx-checksum-ipv4: off
tx-checksum-ipv6: off
tcp-segmentation-offload: off
tx-tcp-segmentation: off [requested on]
tx-tcp6-segmentation: off [requested on]
gene-pc:~$

Unfortunately, after doing this, I have had great difficulty reproducing
the crash, and I do not know why. Here's what changed: I did a routine apt
update, but I did not get a new kernel. I installed ethtool, and did the
command you suggested. After doing that, the system refused to crash
either transmit or receive. So I rebooted, this time did not run ethtool,
and the system again refused to crash. Previously I could reproduce the
crash in about 30 seconds. Now I cannot reproduce the crash after over an
hour. An hour of transmit, and an hour of receive. And an hour
bidirectional. I do not know why.

One problem here is that I test at home and at work, with the same
computer, but with a different TB-10 dock, monitor, keyboard, mouse.
Previous to my bug report, I had never seen the problem at home. The bug
report I sent was from work. Since receiving your request for information,
I cannot readily reproduce the problem either at home or at work. Except
once. I did see a crash at home when *I unplugged/plugged in a keyboard
while the test was going on*. I have attached three lsusb outputs:

   - usb.work →At work, system healthy.
   - usb.home → At home, system healthy.
   - usb.failed → At home, after the crash where I moved the keyboard

What else can I do to help diagnose the problem?

On Thu, Feb 14, 2019 at 11:05 PM Anthony Wong <email address hidden>
wrote:

> ** Changed in: linux (Ubuntu)
> Status: Confirmed => Incomplete
>
> --
> You received this bug notification because you are subscribed to the bug
> report.
> https://bugs.launchpad.net/bugs/1815808
>
> Title:
> Thunderbolt Ethernet high-traffic exception
>
> Status in linux package in Ubuntu:
> Incomplete
>
> Bug description:
> Running iperf3 on Dell XP-15 with TB-10 Thunderbolt dock causes a
> kernel exception and ethernet failure after about 30 seconds.
>
> Easy to reproduce. Connect to another machine via Gig-E and run
> iperf3 in either direction until the problem occurs. Should run 900+
> Mb in either direction. On one occasion I also lost the USB ports.
> However the display continues to work.
>
> <on my-machine>
> # uname -a
> Linux hauser 4.15.0-45-generic #48-Ubuntu SMP Tue Jan 29 16:28:13 UTC
> 2019 x86_64 x86_64 x86_64 GNU/Linux
> # apt install iperf3
> # iperf3 -s
>
> <other machine>
> # apt install iperf3
> # iperf3 -c my-machine -t 1000 # send to my machine
> # iperf3 -c my-machine -t 1000 -R # receive data from my machine
>
> When the bitrate drops to zero, look in syslog for the exception.
>
> Note: Until recently the TB-10 Ethernet would not receive full speed.
> (prior to -43 kernel) Did not see this problem before the Ethernet
> performance increased.
>
> ProblemType: Bug
> DistroRelease: Ubuntu 18.04
> Package: linux-image-4.15.0-45-generic 4.15.0-45.48
> ProcVersionSignature: Ubuntu 4.15.0-45.48-generic 4.15.18
> Uname: Linux 4.15.0-45-generic x86_64
> NonfreeKernelModules: nvidia_modeset nvidia
> ApportVersion: 2.20.9-0ubuntu7.5
> Architecture: amd64
> AudioDevicesInUse:
> USER PID ACCESS COMMAND
> /dev/snd/controlC0: gene 1994 F.... pulseaudio
> CurrentDesktop: ubuntu:GNOME
> Date: Wed Feb 13 13:54:28 2019
> HibernationDevice: RESUME=UUID=d15ecb23-8691-42a1-9909-a56f934b536c
> InstallationDate: Installed on 2018-11-19 (86 days ago)
> InstallationMedia: Ubuntu 18.04.1 LTS "Bionic Beaver" - Release amd64
> (20180725)
> MachineType: Dell Inc. XPS 15 9570
> ProcFB: 0 inteldrmfb
> ProcKernelCmdLine: BOOT_IMAGE=/boot/vmlinuz-4.15.0-45-generic
> root=UUID=a73e2d3e-5ce3-4130-917d-d116c1316153 ro net.ifnames=0
> mem_sleep_default=deep
> RelatedPackageVersions:
> linux-restricted-modules-4.15.0-45-generic N/A
> linux-backports-modules-4.15.0-45-generic N/A
> linux-firmware 1.173.3
> SourcePackage: linux
> UpgradeStatus: No upgrade log present (probably fresh install)
> dmi.bios.date: 12/25/2018
> dmi.bios.vendor: Dell Inc.
> dmi.bios.version: 1.7.0
> dmi.board.name: 02MJVY
> dmi.board.vendor: Dell Inc.
> dmi.board.version: A00
> dmi.chassis.type: 10
> dmi.chassis.vendor: Dell Inc.
> dmi.modalias:
> dmi:bvnDellInc.:bvr1.7.0:bd12/25/2018:svnDellInc.:pnXPS159570:pvr:rvnDellInc.:rn02MJVY:rvrA00:cvnDellInc.:ct10:cvr:
> dmi.product.family: XPS
> dmi.product.name: XPS 15 9570
> dmi.sys.vendor: Dell Inc.
>
> To manage notifications about this bug go to:
> https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1815808/+subscriptions
>