[PATCH 0/1] MR11319: Draft: winebus: Add g29 to devices with g920 usage page

[Drountian Silen (＠drountian)]

This fixes the Logitech G29 brake not being correctly detected in "Exo Rally Championship", I'm not _certain_ that it works or breaks things in other games, but I think it's likely to be correct. The G29 and G920 are closely related devices, the G29 was merely sold as the "PlayStation compatible" version of the G920. Note that this patch only applies to PS3 mode, not PS4 mode, which has a different device id. PS3 mode seems to be the preferred mode for use on Linux due to FFB (force feedback) compatibility, and all of the wheel's features seem to function in PS3 mode, so I have not taken it further. I have only tested this patch with the out-of-tree `new-lg4ff` drivers generally recommended across the internet. These support a few other features in addition to the in-tree driver, but I have no reason to believe axis-assignment is different with the in-tree `lg4ff` drivers. I _can_ say the problem (where the brake axis was simply not recognized by the game) _was_ certainly observable 
 with the in-tree drivers.

I also think this is likely to be the correct fix because I can see the `RZ` axis as the axis reported by `evtest` when I depress the brake, so it seems consistently reported as `RZ` in both Linux (under evdev), and in the game's expectations for DirectInput. Without this fix, it gets passed through as the `RX` DirectInput axis, which is simply not what the game expects.

Stylistically, adding an override for a specific device like this in the driver feels bad. That said, the appropriate usage page override already existed here, and I'm merely adding one more (closely related) hardware device to the page. That said, if reviewers have a better idea about _how_ I should implement this override, I'm happy to make any changes the maintainers would like to see. I would also consider extracting the `||` product id comparison to a variable to make the code a little easier to read, and obviate the extra `()` but I don't know what the project policy on that kind of style is, and I was unable to find a comprehensive style guide in my initial search (admittedly I didn't have time for an in-depth investigation). Once the maintainers are happy with this tiny patch and I have a better idea of how to do this best, I'm happy to remove draft status.

Let me know what folks think!

-- 
https://gitlab.winehq.org/wine/wine/-/merge_requests/11319

[PATCH 1/1] Add g29 to devices with g920 usage page

[Drountian Silen]

From: Drountian Silen <[e-mail]>

---
 dlls/winebus.sys/bus_sdl.c | 7 ++++---
 1 file changed, 4 insertions(+), 3 deletions(-)

diff --git a/dlls/winebus.sys/bus_sdl.c b/dlls/winebus.sys/bus_sdl.c
index 7da1a883f1a..f52ff665d12 100644
--- a/dlls/winebus.sys/bus_sdl.c
+++ b/dlls/winebus.sys/bus_sdl.c
@@ -254,8 +254,8 @@ static const USAGE_AND_PAGE g920_absolute_usages[] =
 {
     {.UsagePage = HID_USAGE_PAGE_GENERIC, .Usage = HID_USAGE_GENERIC_X},  /* wheel */
     {.UsagePage = HID_USAGE_PAGE_GENERIC, .Usage = HID_USAGE_GENERIC_Y},  /* accelerator */
-    {.UsagePage = HID_USAGE_PAGE_GENERIC, .Usage = HID_USAGE_GENERIC_Z},  /* brake */
-    {.UsagePage = HID_USAGE_PAGE_GENERIC, .Usage = HID_USAGE_GENERIC_RZ}, /* clutch */
+    {.UsagePage = HID_USAGE_PAGE_GENERIC, .Usage = HID_USAGE_GENERIC_Z},  /* brake (clutch on g29) */
+    {.UsagePage = HID_USAGE_PAGE_GENERIC, .Usage = HID_USAGE_GENERIC_RZ}, /* clutch (brake on g29) */
 };
 static const USAGE_AND_PAGE absolute_axis_usages[] =
 {
@@ -283,7 +283,8 @@ static const USAGE_AND_PAGE relative_axis_usages[] =
 
 static int get_absolute_usages(const struct device_desc *desc, const USAGE_AND_PAGE **absolute_usages)
 {
-    if (desc->vid == 0x046d && desc->pid == 0xc262)
+    // This also works for the g29 (PID 0xc24f)
+    if (desc->vid == 0x046d && (desc->pid == 0xc262 || desc->pid == 0xc24f))
     {
         *absolute_usages = g920_absolute_usages;
         return ARRAY_SIZE(g920_absolute_usages);
-- 
GitLab

https://gitlab.winehq.org/wine/wine/-/merge_requests/11319