WineHQ Bugzilla – Bug 57524 – Commit c9592bae7f475c1b208de0a72ed29e94e3017094 breaks VKB Gladiator HIDRAW support

Bugzilla

Last modified: 2024-12-13 21:36:36 UTC

Commit c9592bae7f475c1b208de0a72ed29e94e3017094 breaks VKB Gladiator HIDRAW support

Bug 57524 - Commit c9592bae7f475c1b208de0a72ed29e94e3017094 breaks VKB Gladiator HIDRAW support


Summary:	Commit c9592bae7f475c1b208de0a72ed29e94e3017094 breaks VKB Gladiator HIDRAW s...

Status:	CLOSED FIXED

AppDB:	Show Apps affected by this bug
Product:	Wine
Classification:	Unclassified
Component:	-unknown
Version:	10.0-rc1
Hardware:	x86-64 Linux

Importance:	P2 normal
Target Milestone:	---
Assigned To:	Mr. Bugs

URL:
Keywords:	regression

Depends on:
Blocks:
	Show dependency tree

Reported:	2024-12-10 10:13 UTC by T E Dixon
Modified:	2024-12-13 21:36 UTC (History)
CC List:	1 user (show)

See Also:
Regression SHA1:	c9592bae7f475c1b208de0a72ed29e94e3017094
Fixed by SHA1:	a70ac0960de3645bb12ea4ca24ccdf4d99d6867e
Distribution:	Gentoo
Staged patchset:

Attachments
Combined output from running 'wine control' before and after c9592bae7 (27.73 KB, text/plain) 2024-12-10 10:13 UTC, T E Dixon	Details
Output with WINEDEBUG=+hid,+dinput from commit c9592bae7f475c1b208de0a72ed29e94e3017094 (203.31 KB, text/plain) 2024-12-10 11:07 UTC, T E Dixon	Details
possible fix (1.86 KB, patch) 2024-12-10 16:10 UTC, Rémi Bernon	Details \| Diff
View All Add an attachment (back traces, logs, proposed patch, testcase, etc.)

Note You need to log in before you can comment on or make changes to this bug.

Description T E Dixon 2024-12-10 10:13:14 UTC

Created attachment 77574 [details]
Combined output from running 'wine control' before and after c9592bae7

I tested 10.0rc1 and found that neither of my VKB Gladiator sticks, nor my Fanatec pedals, would show up in wine control or in games. Upon further investigation it seems that commit c9592bae7f475c1b208de0a72ed29e94e3017094 is causing the issue.

Interestingly, if I unplug the sticks and wait a while before reopening wine control, the Fanatec pedals will show up. But if the sticks are plugged in, no controllers show up.

I've attached the console output from running wine control with the sticks and pedals connected, from builds of wine-staging as at commit 9b1381fc785a3ff8c3cc82e7339ebe29d12003ed (the commit immediately before, where the sticks are both visible and so are the pedals), and as at commit c9592bae7f475c1b208de0a72ed29e94e3017094.

The issue seems to be related to using hidraw, as disabling it will allow all devices to show up.

As an aside, I was testing 10.0rc1 originally because in 9.22 the sticks will stop working after a random period of time (between 10 minutes and multiple hours). I wouldn't be surprised if this was somehow related but I'm not sure what I can do to debug that issue.

Comment 1 Rémi Bernon 2024-12-10 10:40:45 UTC

Thanks for the report, could you make a log with WINEDEBUG=+hid,+dinput?

Comment 2 T E Dixon 2024-12-10 11:07:05 UTC

Created attachment 77575 [details]
Output with WINEDEBUG=+hid,+dinput from commit c9592bae7f475c1b208de0a72ed29e94e3017094

Here it is. Would you like the same from commit 9b1381fc785a3ff8c3cc82e7339ebe29d12003ed as well? I'll have to rebuild it first.

Comment 3 Rémi Bernon 2024-12-10 16:10:10 UTC

Created attachment 77577 [details]
possible fix

This probably helps? I missed that we could receive reports before the device has been started from the NT point of view, also accessing the device desc probably needed to be done within the CS.

Comment 4 T E Dixon 2024-12-11 02:31:23 UTC

(In reply to Rémi Bernon from comment #3)
> Created attachment 77577 [details]
> possible fix
> 
> This probably helps? I missed that we could receive reports before the
> device has been started from the NT point of view, also accessing the device
> desc probably needed to be done within the CS.

I've applied this patch to 10.0rc1, and it seems to resolve my issue. Thanks!

Any suggestions on how to debug the random disconnect issue that may or may not be related? It happens after a random time period and seems to require a restart of wineserver to resolve. I'll make another bug entry for that, of course.

Comment 5 Rémi Bernon 2024-12-11 07:39:41 UTC

Hmm, hard to say for sure. We usually rely on traces but for long runs it can become a bit unwieldy. Something like WINEDEBUG=+hid,+plugplay,+dinput,+rawinput maybe could tell. You can add +server if you want to trace wineserver requests, this will be absolutely massive though.

Comment 6 T E Dixon 2024-12-11 09:17:41 UTC

(In reply to Rémi Bernon from comment #5)
> Hmm, hard to say for sure. We usually rely on traces but for long runs it
> can become a bit unwieldy. Something like
> WINEDEBUG=+hid,+plugplay,+dinput,+rawinput maybe could tell. You can add
> +server if you want to trace wineserver requests, this will be absolutely
> massive though.

Maybe I'll get lucky and be able to reproduce it without a game running. Thanks, I'll see what I can do with that, but it'll likely take me some time.

Comment 7 Rémi Bernon 2024-12-12 10:29:24 UTC

Should be fixed with a70ac0960de3645bb12ea4ca24ccdf4d99d6867e.

Comment 8 T E Dixon 2024-12-12 10:46:28 UTC

(In reply to Rémi Bernon from comment #7)
> Should be fixed with a70ac0960de3645bb12ea4ca24ccdf4d99d6867e.

A brief test on my end suggests this is the case.

Comment 9 Alexandre Julliard 2024-12-13 21:36:36 UTC

Closing bugs fixed in 10.0-rc2.