Date:   Fri, 14 Jun 2019 16:25:07 +0200
From:   Nicolas Saenz Julienne <nsaenzjulienne@...e.de>
To:     Benjamin Tissoires <benjamin.tissoires@...hat.com>, wbauer1@...net
Cc:     Jiri Kosina <jikos@...nel.org>,
        Dmitry Torokhov <dmitry.torokhov@...il.com>,
        "open list:HID CORE LAYER" <linux-input@...r.kernel.org>,
        lkml <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH v2] HID: input: fix a4tech horizontal wheel custom usage

On Fri, 2019-06-14 at 15:36 +0200, Benjamin Tissoires wrote:
> Hi Wolfgang,
> 
> On Thu, Jun 13, 2019 at 1:49 PM Wolfgang Bauer <wbauer@....at> wrote:
> > On Tuesday, 11. Juni 2019, 16:42:37 Benjamin Tissoires wrote:
> > > On Tue, Jun 11, 2019 at 2:13 PM Nicolas Saenz Julienne
> > > 
> > > <nsaenzjulienne@...e.de> wrote:
> > > > NOTE: I CC'd Wolfgang as he's the one who can test this.
> > > 
> > > I'll wait for Wolfram to confirm that the patch works before pushing then.
> > 
> > My name is Wolfgang, not Wolfram... ;-)
> 
> ouch, sorry for that (I am more used to talk to the I2C maintainer apparently)
> 
> > But never mind.
> > 
> > I tested the patch meanwhile on top of kernel 5.2.rc4, where the mouse wheel
> > actually worked.
> 
> Actually, I am a little bit lost here.
>
> The patch mentions a fix of c01908a14bf73, which is in 5.1 final.
> So if your mouse works in 5.2.rc4, I am not sure how
> HID-a4tech-fix-horizontal-scrolling.patch could break it.
> 
> Could you be slightly more specific in what "works" and what doesn't?

Hi Benjamin,

First of all here's the descriptor:
0x05, 0x01, /*  Usage Page (Desktop),               */
0x09, 0x02, /*  Usage (Mouse),                      */
0xA1, 0x01, /*  Collection (Application),           */
0x09, 0x01, /*      Usage (Pointer),                */
0xA1, 0x00, /*      Collection (Physical),          */
0x05, 0x09, /*          Usage Page (Button),        */
0x19, 0x01, /*          Usage Minimum (01h),        */
0x29, 0x08, /*          Usage Maximum (08h),        */
0x15, 0x00, /*          Logical Minimum (0),        */
0x25, 0x01, /*          Logical Maximum (1),        */
0x75, 0x01, /*          Report Size (1),            */
0x95, 0x08, /*          Report Count (8),           */
0x81, 0x02, /*          Input (Variable),           */
0x05, 0x01, /*          Usage Page (Desktop),       */
0x09, 0x30, /*          Usage (X),                  */
0x09, 0x31, /*          Usage (Y),                  */
0x09, 0x38, /*          Usage (Wheel),              */
0x09, 0xB8, /*          Usage (B8h),                */
0x15, 0x81, /*          Logical Minimum (-127),     */
0x25, 0x7F, /*          Logical Maximum (127),      */
0x75, 0x08, /*          Report Size (8),            */
0x95, 0x04, /*          Report Count (4),           */
0x81, 0x06, /*          Input (Variable, Relative), */
0xC0,       /*      End Collection,                 */
0xC0        /*  End Collection


Sorry for the confusion, I'll try to explain the situation:

In v5.2-rc4 without "HID-a4tech-fix-horizontal-scrolling.patch" the vertical
wheel works out of luck as it's mapped to REL_WHEEL_HIGH_RES, which hid-a4tech
ignores and lets hid-input process, the horizontal wheel is broken. On top of
that Usage(0xB8) is also ignored by both hid-a4tech and hid-input as it isn't
mapped to anything.

There are two distinct bugs here:
  - High resolution wheel processing in hid-a4tech not being implemented,
    breaking horizontal wheels.
  - hid-a4tech not taking care of Usage(0xB8) correctly as it depended on it
    being mapped to "Rel.Misc". That behaviour changed in v5.1 with "HID:
    input: add mapping for "Toggle Display" key".

Once high resolution wheel reports are fixed and handled in hid-a4tech's custom
event, the mouse breaks as it's the processing of Usage(0xB8) that triggers the
input_events, which is being ignored.

You'll probably ask how come we didn't see this when
"HID-a4tech-fix-horizontal-scrolling.patch" was merged. It's due to the fact it
was tested on an older kernel, v5.0.15, that didn't contain "HID: input: add
mapping for "Toggle Display" key"[1].

So that's why I added that specific fix tag. For LTS kernels, it is possible
that "Toggle Display" support was back-ported but not the high resolution
wheels support.

Hope it made things more clear.
Regards,
Nicolas

[1] 
https://lkml.kernel.org/lkml/nycvar.YFH.7.76.1906010028440.1962@cbobk.fhfr.pm/T/

> 
> Do we have the report descriptors available somewhere?
> And if not, could you run hid-recorder from
> https://gitlab.freedesktop.org/libevdev/hid-tools and attach the logs
> when you move the horizontal wheel?
> 
> Cheers,
> Benjamin
> 
> > As the patch didn't apply cleanly (it's obviously based upon
> > 
https://git.kernel.org/pub/scm/linux/kernel/git/next/linux-next.git/commit/?id=abf82e8f7e9af40a49e3d905187c662a43c96c8f
, called
> > "HID-
> > a4tech-fix-horizontal-scrolling.patch" below), I added that patch as well.
> > 
> > My results:
> > kernel 5.2.rc4 works
> > kernel 5.2.rc4 + HID-a4tech-fix-horizontal-scrolling.patch is broken
> > kernel 5.2.rc4 + HID-a4tech-fix-horizontal-scrolling.patch +
> > HID-input-fix-a4tech-horizontal-wheel-custom-usage.patch (i.e. this patch)
> > works again
> > 
> > kernel 5.2.rc4 + HID-input-fix-a4tech-horizontal-wheel-custom-usage.patch
> > works as well.
> > 
> > So AFAICT this patch seems to be fine.
> > 
> > For completeness, this is my mouse as listed by lsusb:
> > Bus 003 Device 002: ID 09da:000a A4Tech Co., Ltd. Optical Mouse Opto 510D /
> > OP-620D
> > 
> > Kind Regards,
> > Wolfgang
> > 


Download attachment "signature.asc" of type "application/pgp-signature" (489 bytes)

Powered by blists - more mailing lists