linux.git/tools/testing/selftests/hid/tests/test_tablet.py, branch for-next

selftests/hid: add Huion Kamvas Pro 19 tests

2024-05-07T13:39:43+00:00

This tablets gets a lot of things wrong:
- the secondary button is reported through Secondary Tip Switch
- the third button is reported through Invert

We need to add some out of proximity intermediate state when moving
back and forth with the eraser mode as it can only be triggered by
physically returning the pen, meaning that the tolerated transitions
can never happen.

Link: https://lore.kernel.org/r/20240410-bpf_sources-v1-15-a8bf16033ef8@kernel.org
Reviewed-by: Peter Hutterer 
Signed-off-by: Benjamin Tissoires

selftests/hid: tablets: also check for XP-Pen offset correction

2024-05-07T13:39:39+00:00

The values are taken from the HID-BPF file.
Basically we are recomputing the array provided there.

Link: https://lore.kernel.org/r/20240410-bpf_sources-v1-14-a8bf16033ef8@kernel.org
Reviewed-by: Peter Hutterer 
Signed-off-by: Benjamin Tissoires

selftests/hid: tablets: add a couple of XP-PEN tablets

2024-05-07T13:39:34+00:00

Those tablets don't need special initialization, but are reporting
the events with the wrong usages:
- tip switch is used when the eraser should be used
- eraser is used instead of the secondary barrel switch

Add tests for those so we don't regress in the future.

Currently we set x/y tilt to 0 to not trigger the bpf program
compensate_coordinates_by_tilt()

Link: https://lore.kernel.org/r/20240410-bpf_sources-v1-13-a8bf16033ef8@kernel.org
Reviewed-by: Peter Hutterer 
Signed-off-by: Benjamin Tissoires

selftests/hid: tablets: reduce the number of pen state

2024-05-07T13:39:30+00:00

All the *_WITH*BUTTON states were almost identical except for the
button itself.

I need to add a new device with a third button, and adding a bunch of
states is going to be quite cumbersome.

So convert the `button` parameter of PenState as a boolean, and store
which button is the target as an argument to all functions that need it.

Link: https://lore.kernel.org/r/20240410-bpf_sources-v1-12-a8bf16033ef8@kernel.org
Reviewed-by: Peter Hutterer 
Signed-off-by: Benjamin Tissoires

selftests/hid: fix failing tablet button tests

2023-12-07T13:55:00+00:00

An overlook from commit 74452d6329be ("selftests/hid: tablets: add
variants of states with buttons"), where I don't use the Enum...

Fixes: 74452d6329be ("selftests/hid: tablets: add variants of states with buttons")
Acked-by: Jiri Kosina 
Link: https://lore.kernel.org/r/20231207-b4-wip-selftests-v1-1-c4e13fe04a70@kernel.org
Signed-off-by: Benjamin Tissoires

selftests/hid: fix mypy complains

2023-12-07T08:52:04+00:00

No code change, only typing information added/ignored

Acked-by: Jiri Kosina 
Link: https://lore.kernel.org/r/20231206-wip-selftests-v2-14-c0350c2f5986@kernel.org
Signed-off-by: Benjamin Tissoires

selftests/hid: tablets: be stricter for some transitions

2023-12-07T08:52:04+00:00

To accommodate for legacy devices, we rely on the last state of a
transition to be valid:
for example when we test PEN_IS_OUT_OF_RANGE to PEN_IS_IN_CONTACT,
any "normal" device that reports an InRange bit would insert a
PEN_IS_IN_RANGE state between the 2.

This is of course valid, but this solution prevents to detect false
releases emitted by some firmware:
when pressing an "eraser mode" button, they might send an extra
PEN_IS_OUT_OF_RANGE that we may want to filter.

So define 2 sets of transitions: one that is the ideal behavior, and
one that is OK, it won't break user space, but we have serious doubts
if we are doing the right thing. And depending on the test, either
ask only for valid transitions, or tolerate weird ones.

Reviewed-by: Peter Hutterer 
Acked-by: Jiri Kosina 
Link: https://lore.kernel.org/r/20231206-wip-selftests-v2-13-c0350c2f5986@kernel.org
Signed-off-by: Benjamin Tissoires

selftests/hid: tablets: add a secondary barrel switch test

2023-12-07T08:52:04+00:00

Some tablets report 2 barrel switches. We better test those too.

Use the same transistions description from the primary button tests.

Reviewed-by: Peter Hutterer 
Acked-by: Jiri Kosina 
Link: https://lore.kernel.org/r/20231206-wip-selftests-v2-12-c0350c2f5986@kernel.org
Signed-off-by: Benjamin Tissoires

selftests/hid: tablets: convert the primary button tests

2023-12-07T08:52:04+00:00

We get more descriptive in what we are doing, and also get more
information of what is actually being tested. Instead of having a non
exhaustive button changes that are semi-randomly done, we can describe
all the states we want to test.

Reviewed-by: Peter Hutterer 
Acked-by: Jiri Kosina 
Link: https://lore.kernel.org/r/20231206-wip-selftests-v2-11-c0350c2f5986@kernel.org
Signed-off-by: Benjamin Tissoires

selftests/hid: tablets: add variants of states with buttons

2023-12-07T08:52:04+00:00

Turns out that there are transitions that are unlikely to happen:
for example, having both the tip switch and a button being changed
at the same time (in the same report) would require either a very talented
and precise user or a very bad hardware with a very low sampling rate.

So instead of manually building the button test by hand and forgetting
about some cases, let's reuse the state machine and transitions we have.

This patch only adds the states and the valid transitions. The actual
tests will be replaced later.

Reviewed-by: Peter Hutterer 
Acked-by: Jiri Kosina 
Link: https://lore.kernel.org/r/20231206-wip-selftests-v2-10-c0350c2f5986@kernel.org
Signed-off-by: Benjamin Tissoires