Bug 1182304

Summary: Integrated USB3 port / hub only works with AC adapter connected
Product: [openSUSE] openSUSE Tumbleweed Reporter: robert spitzenpfeil <rs.opensuse>
Component: KernelAssignee: openSUSE Kernel Bugs <kernel-bugs>
Status: NEW --- QA Contact: E-mail List <qa-bugs>
Severity: Normal    
Priority: P5 - None CC: mbenes, nsaenzjulienne, oneukum, rs.opensuse, tiwai
Version: CurrentFlags: mbenes: needinfo? (oneukum)
Target Milestone: ---   
Hardware: x86-64   
OS: openSUSE Tumbleweed   
Found By: --- Services Priority:
Business Priority: Blocker: ---
Marketing QA Status: --- IT Deployment: ---
Attachments: hwinfo
lspci -vvvx

Description robert spitzenpfeil 2021-02-16 07:57:58 UTC
Kernel: 5.10.12-1-default

Yet again I'm having issues with my laptops integrated USB3 port / hub.

One of the USB-3 + USB-C ports only works with the AC adapter plugged in. Pulling the plug with devices connected leads to stuckage.

Same hardware as in: https://bugzilla.opensuse.org/show_bug.cgi?id=1167146
Comment 1 robert spitzenpfeil 2021-02-16 10:26:56 UTC
After plugging the power in, syslog reads:

[Date] [hostname] kernel: usb usb3: root hub lost power or was reset
[Date] [hostname] kernel: usb usb4: root hub lost power or was reset
Comment 2 Takashi Iwai 2021-02-16 10:48:59 UTC
Something for Oliver and Nicolas.
Comment 3 robert spitzenpfeil 2021-02-16 14:09:18 UTC
Just to be clear, the USB-C port ist affected & this adapter from the other bug report is the only device I have for testing.

Without AC power, the USB type A port doesn't work out of the box, neither with usb flash drives nor a mouse.

With AC power everything appears to be fine.
Comment 4 robert spitzenpfeil 2021-02-17 08:58:08 UTC
Just to rule out grossly broken hardware, I've tested the following without the AC adapter connected:

* All ports are available at boot time and can boot from USB sticks
* Running W10 from USB-stick, all ports are available and work
Comment 5 robert spitzenpfeil 2021-02-17 17:47:26 UTC
I've tested with an old 4.15.11-pmagic64 kernel, no problems whatsoever.
Comment 6 robert spitzenpfeil 2021-02-17 18:23:33 UTC
5.10.1-pmagic (64bit) is usable as well.

One of the USB3 type A ports is a bit picky about which usb-stick it likes, but at least there is power. The USB-C works too.
Comment 7 robert spitzenpfeil 2021-02-22 19:24:41 UTC
Created attachment 846381 [details]
Comment 8 robert spitzenpfeil 2021-03-08 16:59:24 UTC
Still an issue with 5.11.2-1-default (TW 20210307)
Comment 9 Oliver Neukum 2021-03-10 10:46:13 UTC
This looks like a bad interaction between the hub driver and ACPI. Can you please test


on the kernel command line?
Comment 10 robert spitzenpfeil 2021-03-10 19:02:37 UTC
Very nice, Very nice!

This has helped a lot!
Comment 11 Oliver Neukum 2021-03-11 10:30:48 UTC
(In reply to robert spitzenpfeil from comment #10)
> Very nice, Very nice!
> This has helped a lot!

At a cost to your battery charge. This confirms that either ACPI is acting up or the HC ends up in teh wrong D state.

What exactly does "helped a lot" mean? Do you see this issue, only rarer?
And please attach the output of "lspci -vvvx" (as root) to this report.
Comment 12 robert spitzenpfeil 2021-03-11 18:47:53 UTC
Well, maybe I should've said that it appears to be working now.

I don't really care much about battery charge, this is a desktop replacement device. But it still really sucks if you need a working usb port and the power is out.
Comment 13 Miroslav Beneš 2022-02-25 17:20:16 UTC
Robert, any chance this has improved with the latest TW kernel? Without disabling the usb autosuspend.
Comment 14 robert spitzenpfeil 2022-02-25 18:25:58 UTC
Thanks for reminding me of this. A few quick tests are looking good.

I will do some more testing tomorrow.
Comment 15 robert spitzenpfeil 2022-02-25 18:28:40 UTC
Well, I still have usbcore.autosuspend=-1 in the kernel cmd line.

I will try without it this weekend.
Comment 16 robert spitzenpfeil 2022-03-01 12:33:32 UTC
I've checked right now. The problem is still the same.

Without an AC adapter connected, "usbcore.autosuspend=-1" is required to keep the USB3 ports functional.
Comment 17 Miroslav Beneš 2022-03-11 11:50:24 UTC
Ok, could you attach the output of "lspci -vvvx" (as root) so that Oliver could take a proper look, please?
Comment 18 robert spitzenpfeil 2022-03-11 19:19:39 UTC
Created attachment 856946 [details]
lspci -vvvx
Comment 19 Miroslav Beneš 2022-03-18 16:22:32 UTC
Oliver, could you take a look when you find time, please?