Re: [PATCH v3 00/18] HID: logitech-hidpp: G920 force-feedback rework and driver cleanups
From: Jiri Kosina
Date: Fri Nov 20 2026 - 07:14:43 EST
On Fri, 20 Nov 2026, Dmitry Torokhov wrote:
> Per Jiri Kosina's request, this 18-patch series for v7.5 is split out
> from v2 and applies on top of:
> [PATCH v3 00/14] HID: logitech-hidpp: G920 force-feedback and battery fixes
>
> Note that v2 patch 22/33 ("HID: logitech-hidpp: shut down G920 force
> feedback on input unregister") has been dropped in favor of "Input:
> destroy force-feedback device on input unregistration" in input core.
>
> The first group of patches improves transport handling and error
> propagation: keeping connect work disabled during probe until
> hid_connect() completes, preserving transport error codes in
> __hidpp_send_report(), and translating HID++ protocol errors into
> standard negative Linux error codes in hidpp_send_message_sync().
>
> The G920 force-feedback rework moves capability discovery back into
> hidpp_probe() while device I/O is active (excluding generic USB PID
> force feedback for validated HID++ devices), separates device-level
> wheel settings from the input FF queue lifetime, moves the G920 range
> sysfs attribute to driver dev_groups using sysfs_emit(), executes range
> stores, effect uploads, and effect erases synchronously, indexes FF
> slot arrays directly by effect ID, and cleans up FF helpers.
>
> The final patches modernize the driver by switching multi-byte protocol
> fields to unaligned accessors, storing device serials as byte arrays,
> using guard(mutex), __free(kfree), alloc_ordered_workqueue(),
> devm_mutex_init(), and devm_device_add_group(), and removing unused
> kfifo and wtp_data fields and redundant helpers.
>
> Thanks.
>
> Signed-off-by: Dmitry Torokhov <dmitry.torokhov@xxxxxxxxx>
> ---
> Changes in v3:
> - Split out from v2 as a separate 18-patch series for v7.5 per Jiri
> Kosina, applying on top of the v7.4 fixes series.
> - Dropped "HID: logitech-hidpp: shut down G920 force feedback on input
> unregister" (v2 patch 22/33) in favor of "Input: destroy
> force-feedback device on input unregistration" in input core.
> - Collected Tested-by tags from Gideon Heinz across all patches in this
> series.
> - Link to v2: https://patch.msgid.link/20261110-hid-logitech-hidpp-v2-0-30743e91d83b@xxxxxxxxx
Applied to hid.git#for-7.5/logitech.
Thanks,
--
Jiri Kosina
SUSE Labs