On Fri, 20 Nov 2026, Dmitry Torokhov wrote:
Per Jiri Kosina's request, the v2 series has been split into this
14-patch bug-fix series targeting v7.4 and a separate 18-patch rework
and cleanup series targeting v7.5. The open-evdev-across-unbind
force-feedback teardown use-after-free (v2 patch 22/33) is now handled
in input core ("Input: destroy force-feedback device on input
unregistration") and has been dropped.
This series starts from Gideon Heinz's report of Logitech G920
force-feedback initialization timing out in v7.3 and addresses several
pre-existing bugs and lifetime races in hid-logitech-hidpp.
The force-feedback and teardown fixes temporarily enable device I/O
around g920_get_config() in hidpp_input_configured(), switch
hidpp_ff_queue_work() to GFP_ATOMIC since callbacks run under the input
event spinlock, drain the G920 FF workqueue before closing the HID
transport on input close, synchronously disable driver work items and
disconnect input devices with device I/O enabled before stopping the
transport on removal and probe failure, avoid storing an ERR_PTR in
battery->ps on registration failure, avoid NULL pointer subtraction in
the HID++ 1.0 consumer keys report fixup, and fix an off-by-one in G920
slot tracking that made the 63rd user effect in slot 64 unusable and
leaked it across reopen.
The transport and battery fixes update hidpp_send_message_sync() to stop
retrying timeouts and transport errors, initialize response buffers
before publishing send_receive_buf to eliminate races with concurrent or
spurious packets, fix HID++ 1.0 battery mileage event parsing and state
updates, properly propagate negative errors in ADC measurement (0x1f20)
queries, mark devices online on unified battery (0x1004) events while
gating 0x1000/0x1004 dispatch, fix K750 solar battery feature lookup and
reconnect handling, and expose and correctly map the charge_type
property for the battery voltage (0x1001) feature.
Note that device removal teardown still has limitations due to how HID
core's hid_device_io_start() and hid_device_io_stop() (via
driver_input_lock) gate both driver raw command/response I/O and
subdriver event forwarding, disabling I/O before .remove() and
hid_disconnect(). For instance, HIDPP_QUIRK_DELAYED_INIT devm-managed
input devices are still unregistered after hid_hw_stop(), and
hidpp_connect_event() publishes hidpp->input before input registration
completes. Resolving these teardown edge cases cleanly will require
follow-up work in HID core and driver input registration.
Thanks.
Signed-off-by: Dmitry Torokhov <dmitry.torokhov@gmail.com>
---
Changes in v3:
- Split out the 14 bug fixes into this series for v7.4 per Jiri Kosina,
moving the 18 rework and cleanup patches into a separate series for
v7.5 and dropping v2 patch 22/33 in favor of input core teardown.
- Collected Tested-by tags from Gideon Heinz on all unchanged patches in
this series (omitted from patches 12 and 14 which were updated after
v2 testing).
- Gated event dispatch in hidpp_raw_hidpp_event() on
HIDPP_CAPABILITY_UNIFIED_BATTERY so 0x1000 and 0x1004 events are
routed to their respective handlers instead of cross-corrupting
battery state.
- In hidpp20_battery_map_status_voltage(), set charge_type to unknown
when status is unknown, to none when discharging, full, or not
charging, and to standard, fast, or trickle only when charging.
- Reset charge_type to unknown on wireless disconnect in
hidpp_connect_event().
- Link to v2: https://patch.msgid.link/20261110-hid-logitech-hidpp-v2-0-30743e91d83b@gmail.com (local)
Applied to hid.git#for-7.4/upstream-fixes.
Thanks,
--
Jiri Kosina
SUSE Labs