Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
MIME-Version: 1.0
References: <2262737.ElGaqSPkdT@kreacher>
In-Reply-To: <2262737.ElGaqSPkdT@kreacher>
From:   "Rafael J. Wysocki" <rafael@kernel.org>
Date:   Wed, 7 Dec 2022 09:58:22 +0100
Message-ID: <CAJZ5v0jBo-_XnN2m0jeVdeTi7kjr6C3OSzc1NEJgav0srD0JGQ@mail.gmail.com>
Subject: Re: [Regression] Logitech BT mouse unusable after commit 532223c8ac57
 (still in 6.1-rc8)
To:     Bastien Nocera <hadess@hadess.net>, Jiri Kosina <jikos@kernel.org>
Cc:     =?UTF-8?Q?Filipe_La=C3=ADns?= <lains@riseup.net>,
        Benjamin Tissoires <benjamin.tissoires@redhat.com>,
        linux-input@vger.kernel.org, LKML <linux-kernel@vger.kernel.org>,
        "Rafael J. Wysocki" <rafael@kernel.org>,
        Thorsten Leemhuis <regressions@leemhuis.info>
Content-Type: text/plain; charset="UTF-8"
Precedence: bulk

On Tue, Dec 6, 2022 at 3:58 PM Rafael J. Wysocki <rjw@rjwysocki.net> wrote:
>
> Bastien, Jiri,
>
> Commit 532223c8ac57 ("HID: logitech-hidpp: Enable HID++ for all the Logitech
> Bluetooth devices") caused my Logitech Bluetooth mouse to become unusable.
>
> Appended is the change I need to make it work again (note that adding the
> device ID to unhandled_hidpp_devices[] doesn't help, so there must be some
> significant enough difference in how the two cases are handled in the stack).
>
> Here's what I get in the log without the patch below:
>
> [   36.710574] Bluetooth: HIDP (Human Interface Emulation) ver 1.2
> [   36.710592] Bluetooth: HIDP socket layer initialized
> [   36.724644] hid-generic 0005:046D:B016.0001: unknown main item tag 0x0
> [   36.725860] input: Bluetooth Mouse M336/M337/M535 Mouse as /devices/pci0000:00/0000:00:14.0/usb1/1-7/1-7:1.0/bluetooth/hci0/hci0:1/0005:046D:B016.0001/input/input14
> [   36.728036] input: Bluetooth Mouse M336/M337/M535 Consumer Control as /devices/pci0000:00/0000:00:14.0/usb1/1-7/1-7:1.0/bluetooth/hci0/hci0:1/0005:046D:B016.0001/input/input15
> [   36.728823] input: Bluetooth Mouse M336/M337/M535 Keyboard as /devices/pci0000:00/0000:00:14.0/usb1/1-7/1-7:1.0/bluetooth/hci0/hci0:1/0005:046D:B016.0001/input/input18
> [   36.731550] hid-generic 0005:046D:B016.0001: input,hidraw0: BLUETOOTH HID v12.03 Mouse [Bluetooth Mouse M336/M337/M535] on 9c:b6:d0:96:8e:c8
> [   36.833039] logitech-hidpp-device 0005:046D:B016.0001: unknown main item tag 0x0
> [   36.999064] logitech-hidpp-device 0005:046D:B016.0001: Device not connected
>
> and here's what I get with it:
>
> [   43.642546] Bluetooth: HIDP (Human Interface Emulation) ver 1.2
> [   43.642559] Bluetooth: HIDP socket layer initialized
> [   43.652898] hid-generic 0005:046D:B016.0001: unknown main item tag 0x0
> [   43.653833] input: Bluetooth Mouse M336/M337/M535 Mouse as /devices/pci0000:00/0000:00:14.0/usb1/1-7/1-7:1.0/bluetooth/hci0/hci0:1/0005:046D:B016.0001/input/input14
> [   43.655025] input: Bluetooth Mouse M336/M337/M535 Consumer Control as /devices/pci0000:00/0000:00:14.0/usb1/1-7/1-7:1.0/bluetooth/hci0/hci0:1/0005:046D:B016.0001/input/input15
> [   43.655400] input: Bluetooth Mouse M336/M337/M535 Keyboard as /devices/pci0000:00/0000:00:14.0/usb1/1-7/1-7:1.0/bluetooth/hci0/hci0:1/0005:046D:B016.0001/input/input18
> [   43.657521] hid-generic 0005:046D:B016.0001: input,hidraw0: BLUETOOTH HID v12.03 Mouse [Bluetooth Mouse M336/M337/M535] on 9c:b6:d0:96:8e:c8
>
> The only difference seems to be that in the former case the logitech-hidpp
> driver tries to bind to the device, but I guess that is expected.  However,
> when the device ID is added to unhandled_hidpp_devices[], the messages look
> exactly like in the "good" case, but the mouse still doesn't work.

Here's what happens.

- The logitech-hidpp driver is modular and is not present initially,
so hid-generic probes first (successfully).

- logitech-hidpp is loaded which causes the device to be reporobed due
to __hid_bus_driver_added().

- Because the ->match() callback in hid-generic returns 0 for the
device now, it is unbound from the device.

- The probing of logitech-hidpp fails (due to an error in ->probe()).

- The probing of hid-generic fails due to the failing ->match().

So in order for unhandled_hidpp_devices[] in logitech-hidpp to work
with hid-generic, __check_hid_generic() needs to be amended with a
change to run the ->match() callback of the "specific" driver being
tried.

I have a working patch for this that will be sent shortly.

Thanks!