Commit graph

13 commits

Author SHA1 Message Date
Benjamin Tissoires
8d355b56f2 selftests/hid: fix kfunc inclusions with newer bpftool
bpftool now embeds the kfuncs definitions directly in the generated
vmlinux.h

This is great, but because the selftests dir might be compiled with
HID_BPF disabled, we have no guarantees to be able to compile the
sources with the generated kfuncs.

If we have the kfuncs, because we have the `__not_used` hack, the newly
defined kfuncs do not match the ones from vmlinux.h and things go wrong.

Prevent vmlinux.h to define its kfuncs and also add the missing `__weak`
symbols for our custom kfuncs definitions

Link: https://patch.msgid.link/20241128-fix-new-bpftool-v1-1-c9abdf94a719@kernel.org
Signed-off-by: Benjamin Tissoires <bentiss@kernel.org>
2024-11-29 15:27:04 +01:00
Linus Torvalds
919464deec Revert "HID: bpf: allow write access to quirks field in struct hid_device"
This reverts commit 6fd47effe9, and the related self-test update
commit e14e0eaeb0 ("selftests/hid: add test for assigning a given
device to hid-generic").

It results in things like the scroll wheel on Logitech mice not working
after a reboot due to the kernel being confused about the state of the
high-resolution mode.

Quoting Benjamin Tissoires:
 "The idea of 6fd47effe9 was to be able to call hid_bpf_rdesc_fixup()
  once per reprobe of the device.

  However, because the bpf filter can now change the quirk value, the
  call had to be moved before the driver gets bound (which was
  previously ensuring the unicity of the call).

  The net effect is that now, in the case hid-generic gets loaded first
  and then the specific driver gets loaded once the disk is available,
  the value of ->quirks is not reset, but kept to the value that was set
  by hid-generic (HID_QUIRK_INPUT_PER_APP).

  Once hid-logitech-hidpp kicks in, that quirk is now set, which creates
  two inputs for the single mouse: one keyboard for fancy shortcuts, and
  one mouse node.

  However, hid-logitech-hidpp expects only one input node to be attached
  (it stores it into hidpp->input), and when a wheel event is received,
  because there is some processing with high-resolution wheel events,
  the wheel event is injected into hidpp->input.

  And of course, when HID_QUIRK_INPUT_PER_APP is set, hidpp->input gets
  the keyboard node, which doesn't have wheel event type, and the events
  are ignored"

Reported-and-bisected-by: Mike Galbraith <efault@gmx.de>
Link: https://lore.kernel.org/all/CAHk-=wiUkQM3uheit2cNM0Y0OOY5qqspJgC8LkmOkJ2p2LDxcw@mail.gmail.com/
Acked-by: Benjamin Tissoires <bentiss@kernel.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2024-11-25 09:21:47 -08:00
Benjamin Tissoires
e14e0eaeb0 selftests/hid: add test for assigning a given device to hid-generic
We use a well known VID/PID on a driver that doesn't need to talk to
the device, ensures we created the device against the target driver,
then load our program and ensure we have unbound to this driver and use
hid-generic instead.

Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Acked-by: Shuah Khan <skhan@linuxfoundation.org>
Link: https://patch.msgid.link/20241001-hid-bpf-hid-generic-v3-9-2ef1019468df@kernel.org
Signed-off-by: Benjamin Tissoires <bentiss@kernel.org>
2024-10-04 16:10:49 +02:00
Benjamin Tissoires
ff9fbcafba selftests/hid: fix bpf_wq new API
Since commit f56f4d541e ("bpf: helpers: fix bpf_wq_set_callback_impl
signature"), the API for bpf_wq changed a bit.

We need to update the selftests/hid code to reflect that or the
bpf program will not load.

Link: https://patch.msgid.link/20240723-fix-6-11-bpf-v1-1-b9d770346784@kernel.org
Signed-off-by: Benjamin Tissoires <bentiss@kernel.org>
2024-07-24 18:27:21 +02:00
Benjamin Tissoires
fcdf830ea6 selftests/hid: ensure CKI can compile our new tests on old kernels
In the same way than commit ae7487d112 ("selftests/hid: ensure we can
compile the tests on kernels pre-6.3") we should expose struct hid_bpf_ops
when it's not available in vmlinux.h.

So unexpose an eventual struct hid_bpf_ops, include vmlinux.h, and
re-export struct hid_bpf_ops.

Fixes: d7696738d6 ("selftests/hid: convert the hid_bpf selftests with struct_ops")
Reported-by: kernel test robot <lkp@intel.com>
Closes: https://lore.kernel.org/r/202406270328.bscLN1IF-lkp@intel.com/
Link: https://patch.msgid.link/20240701-fix-cki-v2-1-20564e2e1393@kernel.org
Signed-off-by: Benjamin Tissoires <bentiss@kernel.org>
2024-07-01 14:48:03 +02:00
Benjamin Tissoires
62f2e1a096 selftests/hid: add another test for injecting an event from an event hook
Similar to test_multiply_events_wq: we receive one event and inject a
new one. But given that this time we are already in the event hook, we
can use hid_bpf_try_input_report() directly as this function will not
sleep.

Note that the injected event gets processed before the original one this
way.

Link: https://patch.msgid.link/20240626-hid_hw_req_bpf-v2-12-cfd60fb6c79f@kernel.org
Acked-by: Jiri Kosina <jkosina@suse.com>
Signed-off-by: Benjamin Tissoires <bentiss@kernel.org>
2024-06-27 11:00:49 +02:00
Benjamin Tissoires
fe8d561db3 selftests/hid: add wq test for hid_bpf_input_report()
Now that bpf_wq is available, we can write a test with it. Having
hid_bpf_input_report() waiting for the device means that we can
directly call it, and we get that event when the device is ready.

Link: https://patch.msgid.link/20240626-hid_hw_req_bpf-v2-10-cfd60fb6c79f@kernel.org
Acked-by: Jiri Kosina <jkosina@suse.com>
Signed-off-by: Benjamin Tissoires <bentiss@kernel.org>
2024-06-27 11:00:48 +02:00
Benjamin Tissoires
d7696738d6 selftests/hid: convert the hid_bpf selftests with struct_ops
We drop the need for the attach() bpf syscall, but we need to set up
the hid_id field before calling __load().

The .bpf.c part is mechanical: we create one struct_ops per HID-BPF
program, as all the tests are for one program at a time.

Link: https://lore.kernel.org/r/20240608-hid_bpf_struct_ops-v3-4-6ac6ade58329@kernel.org
Signed-off-by: Benjamin Tissoires <bentiss@kernel.org>
2024-06-14 11:20:19 +02:00
Benjamin Tissoires
2c0e8ced7d selftests/hid: add tests for hid_bpf_input_report
Usual way of testing, we call the function and ensures we receive
the event

Link: https://lore.kernel.org/r/20240315-b4-hid-bpf-new-funcs-v4-6-079c282469d3@kernel.org
Signed-off-by: Benjamin Tissoires <bentiss@kernel.org>
2024-04-10 16:38:37 +02:00
Benjamin Tissoires
db624e82c5 selftests/hid: Add test for hid_bpf_hw_output_report
This time we need to ensure uhid receives it, thus the new mutex and
condition.

Link: https://lore.kernel.org/r/20240315-b4-hid-bpf-new-funcs-v4-4-079c282469d3@kernel.org
Signed-off-by: Benjamin Tissoires <bentiss@kernel.org>
2024-04-10 16:37:23 +02:00
Benjamin Tissoires
ae7487d112 selftests/hid: ensure we can compile the tests on kernels pre-6.3
For the hid-bpf tests to compile, we need to have the definition of
struct hid_bpf_ctx. This definition is an internal one from the kernel
and it is supposed to be defined in the generated vmlinux.h.

This vmlinux.h header is generated based on the currently running kernel
or if the kernel was already compiled in the tree. If you just compile
the selftests without compiling the kernel beforehand and you are running
on a 6.2 kernel, you'll end up with a vmlinux.h without the hid_bpf_ctx
definition.

Use the clever trick from tools/testing/selftests/bpf/progs/bpf_iter.h
to force the definition of that symbol in case we don't find it in the
BTF and also add __attribute__((preserve_access_index)) to further
support CO-RE functionality for these tests.

Signed-off-by: Justin Stitt <justinstitt@google.com>
Tested-by: Nick Desaulniers <ndesaulniers@google.com> # Build
Tested-by: Justin Stitt <justinstitt@google.com>
Link: https://lore.kernel.org/r/20230825-wip-selftests-v3-1-639963c54109@kernel.org
Signed-off-by: Benjamin Tissoires <bentiss@kernel.org>
2023-10-11 14:02:54 +02:00
Benjamin Tissoires
4f7153cf46 selftests/hid: add tests for bpf_hid_hw_request
Add tests for the newly implemented function.
We test here only the GET_REPORT part because the other calls are pure
HID protocol and won't infer the result of the test of the bpf hook.

Signed-off-by: Benjamin Tissoires <benjamin.tissoires@redhat.com>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>
2022-11-15 16:28:31 +01:00
Benjamin Tissoires
dbb60c8a26 selftests: add tests for the HID-bpf initial implementation
The tests are pretty basic:
- create a virtual uhid device that no userspace will like (to not mess
  up the running system)
- attach a BPF prog to it
- open the matching hidraw node
- inject one event and check:
  * that the BPF program can do something on the event stream
  * can modify the event stream
- add another test where we attach/detach BPF programs to see if we get
  errors

Note: the Makefile is extracted from selftests/bpf so we can rebuild
the libbpf and bpftool components from the current kernel tree without
relying on system installed components.

Signed-off-by: Benjamin Tissoires <benjamin.tissoires@redhat.com>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>
2022-11-15 16:28:29 +01:00