input/hid subsystem rebase to v6.3

Bugzilla: https://bugzilla.redhat.com/show_bug.cgi?id=2166610
Tested: with the hid-tools test suite and some hardware

This is the usual HID rebase to v6.3, with a little bit of salt from the input tree as well.

Some commits were purposely left aside:

• 4f1f391869ee HID: sony: remove DualShock4 support.
  • We don't enable HID_PLAYSTATION in centos-stream-9, so we would strip out support for DS4 controllers with this one
  • The commit has been included in the backport then reverted to not tempt anybody to backport it
• 78abf1b52055 platform/surface: aggregator: Enforce use of target-ID enum in device ID macros and b09ee1cd5991 platform/surface: aggregator: Rename top-level request functions to avoid ambiguities
  • I left aside the platform changes, and I'm not sure we have customers with surface hardware. We can always cherry-pick those later on request

And some are taken directly from v6.4-rcx because they are going to be backported into stable and are fixing actual bugs:

• 199eec814224 HID: hidpp: terminate retry loop on success
• 3a379450a153 HID: logitech-hidpp: Handle timeout differently from busy
• abf0bd9472c2 HID: wacom: avoid integer overflow in wacom_intuos_inout()

Besides that, nothing jumps out AFAICT.

Signed-off-by: Benjamin Tissoires benjamin.tissoires@redhat.com