Re: [PATCH 0/4] HID: implement new transport-driver callbacks

From: David Herrmann <dh.herrmann@gmail.com>
To: Jiri Kosina <jkosina@suse.cz>
Cc: Frank Praznik <frank.praznik@oh.rr.com>,
	"open list:HID CORE LAYER" <linux-input@vger.kernel.org>,
	Benjamin Tissoires <benjamin.tissoires@gmail.com>
Subject: Re: [PATCH 0/4] HID: implement new transport-driver callbacks
Date: Wed, 29 Jan 2014 09:37:38 +0100	[thread overview]
Message-ID: <CANq1E4S8sqsWa-y5nQFUr6EaO0ct6HgqzFRBJ2KOLxZReujV4Q@mail.gmail.com> (raw)
In-Reply-To: <alpine.LNX.2.00.1401282153141.22492@pobox.suse.cz>

Hi

On Tue, Jan 28, 2014 at 9:53 PM, Jiri Kosina <jkosina@suse.cz> wrote:
> On Wed, 22 Jan 2014, David Herrmann wrote:
>
>> > These patches are originally the work of David Herrmann who suggested that I
>> > update and submit them as their functionality is required for some pending
>> > patches to the hid-sony driver.
>> >
>> > These patches implement the SET/GET_REPORT and raw intr OUTPUT requests for all
>> > transport drivers.  It adds two callbacks to the hid_ll_driver struct:
>> >
>> > int (*raw_request)(struct hid_device *hdev, unsigned char reportnum,
>> >                    __u8 *buf, size_t len, unsigned char rtype, int reqtype);
>> >
>> > int (*output_report)(struct hid_device *hdev, __u8 *buf, size_t len);
>> >
>> > along with the necessary support fuctions in the USBHID, and HIDP drivers.
>> >
>> > UHID is not converted yet.
>>
>> Thanks for picking it up. As background, people should read my HID
>> summary which originally was part of this series:
>>   http://cgit.freedesktop.org/~dvdhrm/linux/tree/Documentation/hid/hid-transport.txt?h=hid&id=86c08bb28302bb31dcd3b9aaf22b222f890397e0
>>
>> Our current hid_output_raw_report() callbacks are implemented
>> differently in the USBHID, HIDP, I2CHID and UHID backends (I even
>> think they're all mutually different). We cannot easily change these
>> as drivers actually depend on the backends to do it differently.
>> Therefore, I proposed the raw_request() and output_report() functions.
>> raw_request() is basically the same as request() but takes a raw
>> buffer instead of an hid_report. output_report() is what HIDP
>> currently does with hid_output_raw_report() and sends the report as
>> asynchronous intr report.
>>
>> The plan should be to use request(), raw_request() and output_report()
>> exclusively and carefully port drivers to use them. Once we're done,
>> we can remove hid_output_raw_report() (and any other legacy). This
>> should guarantee, that drivers can choose between ctrl-SET_REPORT and
>> intr-OUTPUT_REPORT messages without depending on the underlying
>> backend to choose the right one.
>
> I haven't finished reviewing the patchset yet, but anyway -- David, can I
> consider your e-mail as a potential Acked-by:?

Yes, definitely. It's basically my 1-year old patch split into 3.

Thanks
David