Re: [PATCH RESEND] HID: input: do not run GET_REPORT unless there's a Resolution Multiplier

[Peter Hutterer <peter.hutterer@who-t.net>]

On Fri, May 15, 2020 at 09:48:18AM +0200, Benjamin Tissoires wrote:
> Hi Peter,
> 
> On Fri, May 15, 2020 at 12:49 AM Peter Hutterer
> <peter.hutterer@who-t.net> wrote:
> >
> > hid-multitouch currently runs GET_REPORT for Contact Max and again to
> > retrieve the Win8 blob. If both are within the same report, the
> > Resolution Multiplier code calls GET_FEATURE again and this time,
> > possibly due to timing, it causes the ILITEK-TP device interpret the
> > GET_FEATURE as an instruction to change the mode and effectively stop
> > the device from functioning as expected.
> >
> > Notably: the device doesn't even have a Resolution Multiplier so it
> > shouldn't be affected by any of this at all.
> >
> > Fix this by making sure we only execute GET_REPORT if there is
> > a Resolution Multiplier in the respective report.
> >
> > Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
> > Tested-by: Wen He <wen.he_1@nxp.com>
> > ---
> > Same patch as before, but this time with diff.noprefix set to false again.
> > Too bad that setting messes up format-patch :( Apologies for the broken
> > one.
> 
> Thanks for the quick respin. I was about to apply it, and then I
> realized that something was off (see inlined)
> 
> >
> >  drivers/hid/hid-input.c | 22 ++++++++++++----------
> >  1 file changed, 12 insertions(+), 10 deletions(-)
> >
> > diff --git a/drivers/hid/hid-input.c b/drivers/hid/hid-input.c
> > index dea9cc65bf80..a54824d451bf 100644
> > --- a/drivers/hid/hid-input.c
> > +++ b/drivers/hid/hid-input.c
> > @@ -1560,21 +1560,12 @@ static bool __hidinput_change_resolution_multipliers(struct hid_device *hid,
> >  {
> >         struct hid_usage *usage;
> >         bool update_needed = false;
> > +       bool get_report_completed = false;
> >         int i, j;
> >
> >         if (report->maxfield == 0)
> >                 return false;
> >
> > -       /*
> > -        * If we have more than one feature within this report we
> > -        * need to fill in the bits from the others before we can
> > -        * overwrite the ones for the Resolution Multiplier.
> > -        */
> > -       if (report->maxfield > 1) {
> > -               hid_hw_request(hid, report, HID_REQ_GET_REPORT);
> > -               hid_hw_wait(hid);
> > -       }
> > -
> >         for (i = 0; i < report->maxfield; i++) {
> >                 __s32 value = use_logical_max ?
> >                               report->field[i]->logical_maximum :
> > @@ -1593,6 +1584,17 @@ static bool __hidinput_change_resolution_multipliers(struct hid_device *hid,
> >                         if (usage->hid != HID_GD_RESOLUTION_MULTIPLIER)
> >                                 continue;
> >
> > +                       /*
> > +                        * If we have more than one feature within this report we
> > +                        * need to fill in the bits from the others before we can
> > +                        * overwrite the ones for the Resolution Multiplier.
> > +                        */
> > +                       if (!get_report_completed && report->maxfield > 1) {
> > +                               hid_hw_request(hid, report, HID_REQ_GET_REPORT);
> 
> I think here we said that the reading of this particular feature was
> mandatory by Microsoft, but what if a device doesn't like it.
> I wonder if we should not guard this against HID_QUIRK_NO_INIT_REPORTS
> too, in the event we need to quirk a particular device.

just to clarify: "I wonder if" means "please add this" here? :)

tbh I don't see how a device could function if one cannot read the report
with the RM - Windows reads and sets it unconditionally so that device would
break under Windows. Which, presumably, is motivation enough for a vendor to
fix it.

I'm not even sure there are devices where this is ever triggered now, having
two unrelated features in the same report seems a bit of a niche case.
We can easily add the check but whether it'll ever be needed is doubtful.

Cheers,
   Peter