Commit bf81de76 authored by Benjamin Tissoires's avatar Benjamin Tissoires

selftests: hid: import hid-tools usb-crash tests

These tests have been developed in the hid-tools[0] tree for a while.
Now that we have  a proper selftests/hid kernel entry and that the tests
are more reliable, it is time to directly include those in the kernel
tree.

This one gets skipped when run by vmtest.sh as we currently need to test
against actual kernel modules (.ko), not built-in to fetch the list
of supported devices.

[0] https://gitlab.freedesktop.org/libevdev/hid-toolsSigned-off-by: default avatarBenjamin Tissoires <benjamin.tissoires@redhat.com>
parent a4ee40b6
......@@ -14,6 +14,7 @@ TEST_PROGS += hid-mouse.sh
TEST_PROGS += hid-multitouch.sh
TEST_PROGS += hid-sony.sh
TEST_PROGS += hid-tablet.sh
TEST_PROGS += hid-usb_crash.sh
TEST_PROGS += hid-wacom.sh
CXX ?= $(CROSS_COMPILE)g++
......
#!/bin/sh
# SPDX-License-Identifier: GPL-2.0
# Runs tests for the HID subsystem
export TARGET=test_usb_crash.py
bash ./run-hid-tools-tests.sh
#!/bin/env python3
# SPDX-License-Identifier: GPL-2.0
# -*- coding: utf-8 -*-
#
# Copyright (c) 2021 Benjamin Tissoires <benjamin.tissoires@gmail.com>
# Copyright (c) 2021 Red Hat, Inc.
#
# This is to ensure we don't crash when emulating USB devices
from . import base
import pytest
import logging
logger = logging.getLogger("hidtools.test.usb")
class USBDev(base.UHIDTestDevice):
# fmt: off
report_descriptor = [
0x05, 0x01, # .Usage Page (Generic Desktop) 0
0x09, 0x02, # .Usage (Mouse) 2
0xa1, 0x01, # .Collection (Application) 4
0x09, 0x02, # ..Usage (Mouse) 6
0xa1, 0x02, # ..Collection (Logical) 8
0x09, 0x01, # ...Usage (Pointer) 10
0xa1, 0x00, # ...Collection (Physical) 12
0x05, 0x09, # ....Usage Page (Button) 14
0x19, 0x01, # ....Usage Minimum (1) 16
0x29, 0x03, # ....Usage Maximum (3) 18
0x15, 0x00, # ....Logical Minimum (0) 20
0x25, 0x01, # ....Logical Maximum (1) 22
0x75, 0x01, # ....Report Size (1) 24
0x95, 0x03, # ....Report Count (3) 26
0x81, 0x02, # ....Input (Data,Var,Abs) 28
0x75, 0x05, # ....Report Size (5) 30
0x95, 0x01, # ....Report Count (1) 32
0x81, 0x03, # ....Input (Cnst,Var,Abs) 34
0x05, 0x01, # ....Usage Page (Generic Desktop) 36
0x09, 0x30, # ....Usage (X) 38
0x09, 0x31, # ....Usage (Y) 40
0x15, 0x81, # ....Logical Minimum (-127) 42
0x25, 0x7f, # ....Logical Maximum (127) 44
0x75, 0x08, # ....Report Size (8) 46
0x95, 0x02, # ....Report Count (2) 48
0x81, 0x06, # ....Input (Data,Var,Rel) 50
0xc0, # ...End Collection 52
0xc0, # ..End Collection 53
0xc0, # .End Collection 54
]
# fmt: on
def __init__(self, name=None, input_info=None):
super().__init__(
name, "Mouse", input_info=input_info, rdesc=USBDev.report_descriptor
)
# skip witing for udev events, it's likely that the report
# descriptor is wrong
def is_ready(self):
return True
# we don't have an evdev node here, so paper over
# the checks
def get_evdev(self, application=None):
return "OK"
class TestUSBDevice(base.BaseTestCase.TestUhid):
"""
Test class to test if an emulated USB device crashes
the kernel.
"""
# conftest.py is generating the following fixture:
#
# @pytest.fixture(params=[('modulename', 1, 2)])
# def usbVidPid(self, request):
# return request.param
@pytest.fixture()
def new_uhdev(self, usbVidPid, request):
self.module, self.vid, self.pid = usbVidPid
self._load_kernel_module(None, self.module)
return USBDev(input_info=(3, self.vid, self.pid))
def test_creation(self):
"""
inject the USB dev through uhid and immediately see if there is a crash:
uhid can create a USB device with the BUS_USB bus, and some
drivers assume that they can then access USB related structures
when they are actually provided a uhid device. This leads to
a crash because those access result in a segmentation fault.
The kernel should not crash on any (random) user space correct
use of its API. So run through all available modules and declared
devices to see if we can generate a uhid device without a crash.
The test is empty as the fixture `check_taint` is doing the job (and
honestly, when the kernel crashes, the whole machine freezes).
"""
assert True
Markdown is supported
0%
or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment