Bug 203789

Summary:	Multiple mt76u_process_rx_entry: 68 callbacks suppressed found with 5.2-rc2 kernel
Product:	Drivers	Reporter:	Stuart Foster (smf-linux)
Component:	network-wireless	Assignee:	drivers_network-wireless (drivers_network-wireless)
Status:	RESOLVED PATCH_ALREADY_AVAILABLE
Severity:	normal	CC:	bugzilla.kernel.org, greg, lorenzo.bianconi83
Priority:	P1
Hardware:	All
OS:	Linux
Kernel Version:	linux 5.2-rc3	Subsystem:
Regression:	No	Bisected commit-id:
Attachments:	Downloaded patch

Description Stuart Foster 2019-06-03 11:45:34 UTC

This device was found to be passing only a small amount of data.
Examining dmesg found the following. 

[    3.017898] usbcore: registered new interface driver mt76x0u

Multiple entries:

[  259.411503] mt76u_process_rx_entry: 68 callbacks suppressed
[  259.411525] usb 1-1: rx data too big 2044
[  259.517895] usb 1-1: rx data too big 2044
[  259.621517] usb 1-1: rx data too big 2044
[  259.622193] usb 1-1: rx data too big 2044
[  259.730383] usb 1-1: rx data too big 2044
[  259.835789] usb 1-1: rx data too big 2044
[  259.838003] usb 1-1: rx data too big 2044
[  260.040689] usb 1-1: rx data too big 2044
[  260.150328] usb 1-1: rx data too big 2044
[  260.354527] usb 1-1: rx data too big 2044

Comment 1 Lorenzo Bianconi 2019-06-03 16:30:34 UTC

could you please try this series?

https://patchwork.kernel.org/cover/10969843/

Comment 2 Stuart Foster 2019-06-04 07:26:14 UTC

(In reply to Lorenzo Bianconi from comment #1)
> could you please try this series?
> 
> https://patchwork.kernel.org/cover/10969843/

Thank you that patch has resolved my issue.

Comment 3 Lorenzo Bianconi 2019-06-04 09:08:35 UTC

(In reply to Stuart Foster from comment #2)
> (In reply to Lorenzo Bianconi from comment #1)
> > could you please try this series?
> > 
> > https://patchwork.kernel.org/cover/10969843/
> 
> Thank you that patch has resolved my issue.

Thx a lot for testing. Could you please add your 'tested-by' to the series?

Comment 4 Stuart Foster 2019-06-04 19:00:01 UTC

Yes I will (Just as soon as I get my registration completed with Patchwork).

Comment 5 Stuart Foster 2019-06-09 07:49:21 UTC

Still trying !!

Comment 6 Stuart Foster 2019-07-02 23:05:43 UTC

Given up, are you going to add the patch to 5.2 ?

Comment 7 David 2019-07-12 16:19:23 UTC

It seems that this bug made it into 5.2: on Arch Linux with 5.2 my dmesg output was littered with usb 4-3: rx data too big 2044 and wifi performance was abysmal.
From my point of view, this is a regression, as downgrading to 5.1.16 makes it go away.
Unfortunately, I'm not sure how to test the above patch, though with a bit of guidance I might be able to.

Comment 8 Stuart Foster 2019-07-13 12:18:04 UTC

Created attachment 283665 [details]
Downloaded patch

This is the patch refered to in the bug. Are you able to build your kernel (5.2) from source ?

Comment 9 David 2019-07-14 09:47:25 UTC

I managed to build 5.2 with the patch and am happy to report that it fixes all issues.
Thank you!

Comment 10 Stuart Foster 2019-07-14 11:21:00 UTC

I am please, hopefully the fix will get into linux 5.2 eventually.

Comment 11 Stuart Foster 2019-07-17 17:59:06 UTC

It looks like this and other changes are scheduled for linux 5.3.

Comment 12 Stuart Foster 2019-07-22 10:48:24 UTC

Just done a quick test on linux 5.3-rc1 and I can confirm that the changes are present and are working ok.