Bug 1171677 - USB-C external hub randomly stops working
USB-C external hub randomly stops working
Status: REOPENED
Classification: openSUSE
Product: openSUSE Distribution
Classification: openSUSE
Component: Kernel
Leap 15.2
x86-64 Other
: P3 - Medium : Normal (vote)
: ---
Assigned To: Oliver Neukum
E-mail List
:
Depends on:
Blocks:
  Show dependency treegraph
 
Reported: 2020-05-14 13:58 UTC by Ivan Murray-Smith
Modified: 2021-03-03 16:07 UTC (History)
8 users (show)

See Also:
Found By: ---
Services Priority:
Business Priority:
Blocker: ---
Marketing QA Status: ---
IT Deployment: ---
oneukum: needinfo? (ivanleo)
ivanleo: needinfo?


Attachments
dmesg output after network failure (498.38 KB, text/plain)
2020-05-14 13:58 UTC, Ivan Murray-Smith
Details
hwinfo (580.35 KB, text/plain)
2020-05-14 13:59 UTC, Ivan Murray-Smith
Details
Re-connection failure screenshot (946.52 KB, image/png)
2020-05-14 13:59 UTC, Ivan Murray-Smith
Details
dmesg after unplugging and reconnecting (498.75 KB, text/plain)
2020-05-19 08:11 UTC, Ivan Murray-Smith
Details
lsmod before connecting the usb-c hub (7.48 KB, text/plain)
2020-12-07 12:04 UTC, Ivan Murray-Smith
Details
lsmod after connecting the usb-c hub (7.48 KB, text/plain)
2020-12-07 12:04 UTC, Ivan Murray-Smith
Details
dmesg before connecting the usb-c hub (107.06 KB, text/plain)
2020-12-07 12:04 UTC, Ivan Murray-Smith
Details
dmesg after connecting the usb-c hub (115.53 KB, text/plain)
2020-12-07 12:05 UTC, Ivan Murray-Smith
Details
dmesg before connecting the usb-c hub (74.06 KB, text/plain)
2020-12-08 19:42 UTC, Ivan Murray-Smith
Details
dmesg after connecting the usb-c hub (82.99 KB, text/plain)
2020-12-08 19:43 UTC, Ivan Murray-Smith
Details
lsmod before connecting the usb-c hub (7.35 KB, text/plain)
2020-12-08 19:43 UTC, Ivan Murray-Smith
Details
lsmod after connecting the usb-c hub (7.52 KB, text/plain)
2020-12-08 19:44 UTC, Ivan Murray-Smith
Details
usb_devices after booting with "usbcore.autosuspend=-1" kernel parameter (5.30 KB, text/plain)
2021-01-25 09:15 UTC, Ivan Murray-Smith
Details

Note You need to log in before you can comment on or make changes to this bug.
Description Ivan Murray-Smith 2020-05-14 13:58:45 UTC
Created attachment 837784 [details]
dmesg output after network failure

Network works for a day or so, then although there is an active connection, there is simply no connectivity, I can't even get to my router at 192.168.1.1 and chrome gives an error message saying there's no network access.

Ping to any address says network is unreachable.

If I tell network manager to disconnect and reconnect, it fails with the error shown in the attached screenshot, I'm also attaching the dmesg and hwinfo outpus.
Comment 1 Ivan Murray-Smith 2020-05-14 13:59:07 UTC
Created attachment 837785 [details]
hwinfo
Comment 2 Ivan Murray-Smith 2020-05-14 13:59:38 UTC
Created attachment 837786 [details]
Re-connection failure screenshot
Comment 3 Ivan Murray-Smith 2020-05-14 13:59:58 UTC
Also the only workaround so far is to reboot.
Comment 4 Ivan Murray-Smith 2020-05-16 19:58:19 UTC
I've since established it doesn't seem to be a networking issue as such but rather a problem with the USB system, the ethernet card is built into a little external box that connects via USB C.

Normally you could reset things by just unplugging it and plugging it in again, however when the network goes does this happens:

The eth lights stay on
The mouse, plugged into the same external box, keeps working fine

However if I unplug the USB C connector and reconnect it, everything connected via the external box is dead.
Comment 6 Ivan Murray-Smith 2020-05-19 08:11:47 UTC
Created attachment 837957 [details]
dmesg after unplugging and reconnecting
Comment 7 Lubos Kocman 2020-06-08 11:34:42 UTC
Switching from GNOME desktop to kernel.
Comment 8 Oliver Neukum 2020-12-07 11:31:11 UTC
The machine shows bus errors:

[71334.334181] pcieport 0000:00:1d.6: AER: Corrected error received: 0000:00:1d.6


and thermal reports.
Comment 9 Ivan Murray-Smith 2020-12-07 11:33:04 UTC
Still broken on 5.3.18-lp152.54-default

Is there any test script or anything I can run to help diagnose this?
Comment 10 Oliver Neukum 2020-12-07 11:36:28 UTC
(In reply to Ivan Murray-Smith from comment #9)
> Still broken on 5.3.18-lp152.54-default
> 
> Is there any test script or anything I can run to help diagnose this?

We can try the dynamic debugging feature for teh modules involved to enable more logging. We need to pinpoint what to switch on. Please provide "dmesg" and "lsmod" before and after plugging the hub in.
Comment 11 Ivan Murray-Smith 2020-12-07 12:04:12 UTC
Created attachment 844171 [details]
lsmod before connecting the usb-c hub
Comment 12 Ivan Murray-Smith 2020-12-07 12:04:35 UTC
Created attachment 844172 [details]
lsmod after connecting the usb-c hub
Comment 13 Ivan Murray-Smith 2020-12-07 12:04:54 UTC
Created attachment 844173 [details]
dmesg before connecting the usb-c hub
Comment 14 Ivan Murray-Smith 2020-12-07 12:05:17 UTC
Created attachment 844174 [details]
dmesg after connecting the usb-c hub
Comment 15 Ivan Murray-Smith 2020-12-08 19:42:50 UTC
Created attachment 844287 [details]
dmesg before connecting the usb-c hub
Comment 16 Ivan Murray-Smith 2020-12-08 19:43:16 UTC
Created attachment 844288 [details]
dmesg after connecting the usb-c hub
Comment 17 Ivan Murray-Smith 2020-12-08 19:43:42 UTC
Created attachment 844289 [details]
lsmod before connecting the usb-c hub
Comment 18 Ivan Murray-Smith 2020-12-08 19:44:08 UTC
Created attachment 844290 [details]
lsmod after connecting the usb-c hub
Comment 19 Ivan Murray-Smith 2020-12-08 19:47:43 UTC
Re-run on 5.3.18-lp152.57-default, as far as I can see plugging the hub in loads these additional modules:

Module                  Size  Used by
cdc_ether              24576  0
usbnet                 53248  1 cdc_ether
r8152                  77824  0
mii                    16384  2 usbnet,r8152

The dmesg messages are the ones in the "after" log from [  126.109687] onwards.

Please let me know what else you need.
Comment 20 Oliver Neukum 2021-01-13 10:28:02 UTC
This may be a duplicate. Please test with "usbcore.autosuspend=-1" on the kernel command line to verify the connection.
Comment 21 Ivan Murray-Smith 2021-01-24 15:13:29 UTC
I've rebooted with that kernel option and so far (5 minutes or so) so good. Is there anything I can do other than wait and see if the problem re-occurs?
Comment 22 Oliver Neukum 2021-01-25 09:02:25 UTC
(In reply to Ivan Murray-Smith from comment #21)
> I've rebooted with that kernel option and so far (5 minutes or so) so good.
> Is there anything I can do other than wait and see if the problem re-occurs?

Please provide the output of "usb-devices" with the option on the kernel command line.
Comment 23 Ivan Murray-Smith 2021-01-25 09:15:26 UTC
Created attachment 845473 [details]
usb_devices after booting with "usbcore.autosuspend=-1" kernel parameter
Comment 24 Ivan Murray-Smith 2021-01-29 13:17:09 UTC
Well nobody's replied but the problem has not occurred all week, so I think the kernel parameter has proven to be an effective workaround.
Comment 25 Oliver Neukum 2021-02-17 10:49:34 UTC
This looks like it is fixed by the combination of
a08c0d309d8c078d22717d815cf9853f6f2c07bd
80fd850b31f09263ad175b2f640d5c5c6f76ed41

They, however, need a bit more infrastructure
Comment 26 Oliver Neukum 2021-02-23 15:01:02 UTC
Please try the kernel at

https://build.opensuse.org/project/monitor/home:oneukum:bsc1171677_20210223a

without the work around on the kernel command line.
Comment 27 Ivan Murray-Smith 2021-02-23 15:13:24 UTC
The "Download package" link at https://build.opensuse.org/package/show/home:oneukum:bsc1171677_20210223a/kernel-default takes me to a page that says "No data for home:oneukum:bsc1171677_20210223a / kernel-default"

Should I try again in a few hours?
Comment 28 Oliver Neukum 2021-02-24 09:02:54 UTC
You can now find the binaries at:
https://download.opensuse.org/repositories/home:/oneukum:/bsc1171677_20210223a/standard/x86_64/

Sorry for the problem
Comment 29 Ivan Murray-Smith 2021-03-03 16:07:42 UTC
I've tried but I get this error message: https://photos.app.goo.gl/EdUQtVBujBVMJSNC8

My /boot/efi/EFI/opensuse/boot.csv has one line as follows (not sure if this is relevant):

shim.efi,opensuse-secureboot

Not sure if that is relevant? Also all partitions are fully encrypted, but I can't see why that would make any difference.