Bug 203619

We replaced the file “hid-logitech-dj.c” from the kernel 5.2-rc1 with the one from the kernel 5.1. After that, the Logitech mouse has worked again.

Can you please reboot into the non working kernel, move and click the mouse a couple of times and then do:

dmesg > dmesg.log

After this run:

sudo evemu-record

This gives a list of devices the kernel has found, please copy and paste that list here.

If your mouse is in the list, please type the number of your mouse and then move it around and click. Check if you see events being printed when moving / clicking. Please also report here which number(s) you've tried and if you see events from the mouse for those numbers.

Created attachment 282939 [details]
journalctl -b | grep -iE "kernel|udev" with usb logitech receiver error

Can also reproduce not working logitech usb receiver in 5.2-r1, but for me it appears differently:

13:12:12 systemd-udevd[221]: 1-8:1.0: Worker [247] processing SEQNUM=2452 is taking a long time
13:12:12 systemd-udevd[221]: 1-8:1.1: Worker [228] processing SEQNUM=2453 is taking a long time
13:12:12 systemd-udevd[221]: 1-8:1.2: Worker [375] processing SEQNUM=2454 is taking a long time
13:13:55 kernel: usb 1-8: USB disconnect, device number 2
13:14:12 systemd-udevd[221]: 1-8:1.2: Worker [375] processing SEQNUM=2454 killed
13:14:12 kernel: hid-generic 0003:046D:C52B.0001: usb_submit_urb(ctrl) failed: -19
13:14:12 kernel: input: Logitech USB Receiver as /devices/pci0000:00/0000:00:01.3/0000:03:00.0/usb1/1-8/1-8:1.0/0003:046D:C52B.0001/input/input13
13:14:12 systemd-udevd[221]: 1-8:1.1: Worker [228] processing SEQNUM=2453 killed
13:14:12 systemd-udevd[221]: 1-8:1.0: Worker [247] processing SEQNUM=2452 killed
13:14:12 systemd-udevd[221]: Worker [228] terminated by signal 9 (KILL)
13:14:12 systemd-udevd[221]: 1-8:1.1: Worker [228] failed
13:14:12 systemd-udevd[221]: Worker [375] terminated by signal 9 (KILL)
13:14:12 systemd-udevd[221]: 1-8:1.2: Worker [375] failed

There is some odds in behaviour:
1) Unplugging before boot and plugging after makes it work
2) Using usb3.0 port works after error above, but usb2 extender in usb3.0 port not

Created attachment 282943 [details]
Logitech Mouse Failure PPC

File includes dmesg.log and evemu-record output.  My mouse was listed as number 8, no movements were detected.

Allan

@Dmitry:
You seem to have a different mouse and a different receiver: yours is having the Product ID C52B, which is a Unifying receiver, when Allan has a C52F, which is an entirely different one.

Please open a separate bug so we can accurately track the progresses.

@Allan:
I'll need some raw debugging of your device. As far as I can tell, the enumeration went fine, ("HID++ 4.1 device connected." appeared in the logs), so there must be a mismatch in the configuration we send to the mouse.

Can you run "sudo hid-recorder /dev/hidraw*", move the mouse, click on it, then hit Ctrl-C. For hid-recorder, you can find it at https://gitlab.freedesktop.org/libevdev/hid-tools.

@Allen: I've just received another bug-report of a user with a 0xc52f receiver not working. I have the feeling that this may be a mouse-only receiver and that it needs some special handling, like we already have for 27 MHz mouse-only receiver.

That and/or its mouse HID input report may be different.

If you don't have time to build and run hid-recorder, can you at least run these 2 simple commands and attach the generated output here?  :

1:

lsusb -v -d 0x046d:c52f > lsusb.log

2: As root:

pushd /sys/kernel/debug/hid/
for i in *; do cat $i/rdesc > ~/$i.rdesc; done
popd

And then attach the generated lsusb.log and *.rdesc files here?

Thanks.

@Benjamin: Do you have a 0xc52f receiver? and have you tested with this ?

> @Benjamin: Do you have a 0xc52f receiver? and have you tested with this ?

I was sure I did, but now that I looked for, I can't get my hands on one of those. So I must only have tested the series on the 0xc534, which might solve most of the issues.

Created attachment 283063 [details]
M280 mouse. Moved left/right, double clicked left/right buttons, clicked wheel, scrolled wheel

Created attachment 283067 [details]
M280 mouse lsusb -v -d

Created attachment 283073 [details]
M280-rdesc-zip

Kostiantyn, thank you for the logs.

So as I suspected the c52f receiver is a mouse-only receiver and as such is different then the c534 nano receiver. So the addition of the c52f product-id to the logitech-dj driver when support for the c534 receiver was added, was a mistake and for 5.2 we are going to remove the c52f product it, reverting to the old 5.1 situation.

We do want to eventually support the c52f receiver, and the logs will be very helpful for that. Are you willing to build your own kernel with a patch added to test c52f support for us?

Created attachment 283087 [details]
[PATCH] HID: logitech-dj: Fix 064d:c52f receiver support

If you feel up to building your own kernel (see your distributions documentation), here is a patch which I believe should fix the support for the c52f receiver.

Created attachment 283089 [details]
M325-M280-M235-receivers

I've noticed that there's a different model suffix for M325 unifying receiver and M280 and M235 non-unifying. DJ vs DJR.
I could attach logs for M235 non-unifying receiver if need be.

Hello Hans,

Many thanks for your patch. I successfully compiled the RC3 of kernel 5.2 with your new patch for the AmigaOne X1000 and X5000 yesterday.

Download: http://www.xenosoft.de/linux-image-5.2-rc3-2_X1000_X5000.tar.gz

@Allan
Please test this kernel if you’re back from your work.

Thanks to all for your work!

Cheers,
Christian

I just tested the patch on a brand new M280 with the C52F, and I can confirm that Hans' patch works as expected.

I'll post this on the LKML and push it upstream right away.

Using the new release candidate 4 (5.2RC4) by Christian my X5000 boots to my second graphics card (Firepro) and Logitech mouse works fine using the included patch.

(In reply to Allan Cairns from comment #16)
> Using the new release candidate 4 (5.2RC4) by Christian my X5000 boots to my
> second graphics card (Firepro) and Logitech mouse works fine using the
> included patch.

Thank you for testing.

Since you have confirmed that this is fixed I will close this bug now.

Thank you. M280 started working after 5.2 RC5 kernel upgrade. Good news, everyone.
But there's still an issue. Solaar won't detect it. Solaar is a thirparty tool that manages Unifying receiver pairing and other things. It works with stock Ubuntu kernel 5.0.x and it was working for me in 5.1.x builds. Should I reopen this bug, create a new one here or in Solaar project on github?

(In reply to Kostiantyn Miklevskyi from comment #18)
> Thank you. M280 started working after 5.2 RC5 kernel upgrade. Good news,
> everyone.
> But there's still an issue. Solaar won't detect it. Solaar is a thirparty
> tool that manages Unifying receiver pairing and other things. It works with
> stock Ubuntu kernel 5.0.x and it was working for me in 5.1.x builds. Should
> I reopen this bug, create a new one here or in Solaar project on github?

Solaar should still be able to access the devices through the hidraw interface of the receiver (which is unchanged) just fine, so this is a bit weird. This likely has something to do with how solaar detects the devices.

Can you please file a bug against solaar upstream for this?

Please report back with a link to the solaar issue here, then I will add myself to the Cc of the solaar issue.

All
It appears the original issues have reappeared with the recent pre-releases of Kernel 5.3.  Christian Zigotzky has narrowed down the issues to HID changes between 5.3-alpha and release candidates "OK, we definitely know that the issue is somewhere in the latest HID updates".  Christian provided a test kernel without the latest HID update which work fine.  I am not sure if these issues are related to the previous.

Allan

A patch fixing thr 5.3 regression has been queued for merging into 5.3 here:
https://git.kernel.org/pub/scm/linux/kernel/git/hid/hid.git/log/?h=for-5.3/upstream-fixes

Specifically this patch fixes the regression:
https://git.kernel.org/pub/scm/linux/kernel/git/hid/hid.git/commit/?id=6fb08f1a5f7e5cdde1ce00104788e602f4299b99