Jason --- Now instead of four in the eights place / you’ve got three, ‘Cause you added one / (That is to say, eight) to the two, / But you can’t take seven from three, / So you look at the sixty-fours.... On Thu, Oct 20, 2022 at 7:48 AM Jason Gerecke <killertofu@xxxxxxxxx> wrote: > > On Wed, Oct 19, 2022 at 8:40 AM Benjamin Tissoires > <benjamin.tissoires@xxxxxxxxxx> wrote: > > > > On Wed, Oct 19, 2022 at 5:18 PM Gerecke, Jason <killertofu@xxxxxxxxx> wrote: > > > > > > From: Jason Gerecke <jason.gerecke@xxxxxxxxx> > > > > > > "Digitizer" is a generic usage that may be used by various devices but > > > which is particularly used by non-display pen tablets. This patch adds the > > > usage to the list of values matched by the IS_INPUT_APPLICATION() macro > > > that determines if an input device should be allocated or not. > > > > > > Signed-off-by: Jason Gerecke <jason.gerecke@xxxxxxxxx> > > > Reviewed-by: Ping Cheng <ping.cheng@xxxxxxxxx> > > > --- > > > include/linux/hid.h | 2 +- > > > 1 file changed, 1 insertion(+), 1 deletion(-) > > > > > > diff --git a/include/linux/hid.h b/include/linux/hid.h > > > index 4363a63b9775..07803e144d98 100644 > > > --- a/include/linux/hid.h > > > +++ b/include/linux/hid.h > > > @@ -883,7 +883,7 @@ static inline bool hid_is_usb(struct hid_device *hdev) > > > /* We ignore a few input applications that are not widely used */ > > > #define IS_INPUT_APPLICATION(a) \ > > > (((a >= HID_UP_GENDESK) && (a <= HID_GD_MULTIAXIS)) \ > > > - || ((a >= HID_DG_PEN) && (a <= HID_DG_WHITEBOARD)) \ > > > > FWIW, this has always been problematic, and I am pretty sure this is > > breaking existing devices. > > > > Have you been running the hid-tools testsuite to see if there were any > > regressions? > > > > Cheers, > > Benjamin > > > > I was slightly worried that this usage might have been explicitly > excluded for some compatibility reason, but I didn't see anything in > the commit history that said that. I also had a hard time convincing > myself that allocating an input device for a weird device where it is > unnecessary would cause too much trouble. > > I didn't see any regressions when running the hid-tools testsuite. The > output from the 5.15.74 kernel with / without the patch applied is > almost identical (there are numerous test failures in test_tablet.py > even in the unpatched case; maybe I should re-run with Linus's latest > master instead?) > > Jason > (Apologies for this doubled message, Benjamin -- I didn't "reply all" on my previous attempt...) I re-ran the tests with the Linus's latest 6.1-rc2 and have good results with / without the patch there. The test_tablet.py failures I previously saw no longer occur, so seem to be specific to stable. There are still two tests that fail regardless of if the patch is applied or not, but the patch doesn't seem to introduce any new failures of its own. Jason
