Re: [PATCH 5.5 000/150] 5.5.7-stable review

From: Andre Tomt
Date: Fri Feb 28 2020 - 09:38:09 EST

On 28.02.2020 13:22, Greg Kroah-Hartman wrote:
On Fri, Feb 28, 2020 at 01:06:00PM +0100, Andre Tomt wrote:
On 27.02.2020 14:35, Greg Kroah-Hartman wrote:
This is the start of the stable review cycle for the 5.5.7 release.
There are 150 patches in this series, all will be posted as a response
to this one. If anyone has any issues with these being applied, please
let me know.

Responses should be made by Sat, 29 Feb 2020 13:21:24 +0000.
Anything received after that time might be too late.

There is something going on with USB in this release. My AMD X570 board is
constantly having ports stop working, while a older AMD X399 board seems
fine (maybe, there is an ATEN USB extender involved on the X570 system)

I've only had time to do very rudimentary debugging, but reverting all usb
and xhci related patches seems to have solved it, eg:

<snip>

I might be able to narrow it down in a day or two.

Narrowing it down would be good, try sticking with either the hub or
xhci patches. And, 'git bisect' might make it easier.

It seems to be caused by "USB: hub: Don't record a connect-change event during reset-resume". Running 5.5.7-rc1 with only that patch reverted works fine.

Also happens with the X399. And only when the ATEN UCE260 USB extender is involved. Connecting a device is the problem, if it is connected at boot, things seem fine, until you connect something new. Then the port dies completely, dead even if you disconnect the extender and plug something in directly - until system is rebooted (perhaps driver reload would work too - have not tested.)

Aw well, the problem does seem a little narrower than I expected at first :)

There's no errors in the kernel log at all. The last thing logged is that whatever device was on the other end got disconnected.

Also, does Linus's current tree show the same problems for you?

Yes, it does

Unfortunately, email deliverability will be terrible for me for a few days (there was a mishap), so if you could loop in the author/maintainers for me that would be good.