1 .. SPDX-License-Identifier: GPL-2.0 1 .. SPDX-License-Identifier: GPL-2.0
2 2
3 ======= 3 =======
4 HID-BPF 4 HID-BPF
5 ======= 5 =======
6 6
7 HID is a standard protocol for input devices b 7 HID is a standard protocol for input devices but some devices may require
8 custom tweaks, traditionally done with a kerne 8 custom tweaks, traditionally done with a kernel driver fix. Using the eBPF
9 capabilities instead speeds up development and 9 capabilities instead speeds up development and adds new capabilities to the
10 existing HID interfaces. 10 existing HID interfaces.
11 11
12 .. contents:: 12 .. contents::
13 :local: 13 :local:
14 :depth: 2 14 :depth: 2
15 15
16 16
17 When (and why) to use HID-BPF 17 When (and why) to use HID-BPF
18 ============================= 18 =============================
19 19
20 There are several use cases when using HID-BPF 20 There are several use cases when using HID-BPF is better
21 than standard kernel driver fix: 21 than standard kernel driver fix:
22 22
23 Dead zone of a joystick 23 Dead zone of a joystick
24 ----------------------- 24 -----------------------
25 25
26 Assuming you have a joystick that is getting o 26 Assuming you have a joystick that is getting older, it is common to see it
27 wobbling around its neutral point. This is usu 27 wobbling around its neutral point. This is usually filtered at the application
28 level by adding a *dead zone* for this specifi 28 level by adding a *dead zone* for this specific axis.
29 29
30 With HID-BPF, we can apply this filtering in t 30 With HID-BPF, we can apply this filtering in the kernel directly so userspace
31 does not get woken up when nothing else is hap 31 does not get woken up when nothing else is happening on the input controller.
32 32
33 Of course, given that this dead zone is specif 33 Of course, given that this dead zone is specific to an individual device, we
34 can not create a generic fix for all of the sa 34 can not create a generic fix for all of the same joysticks. Adding a custom
35 kernel API for this (e.g. by adding a sysfs en 35 kernel API for this (e.g. by adding a sysfs entry) does not guarantee this new
36 kernel API will be broadly adopted and maintai 36 kernel API will be broadly adopted and maintained.
37 37
38 HID-BPF allows the userspace program to load t 38 HID-BPF allows the userspace program to load the program itself, ensuring we
39 only load the custom API when we have a user. 39 only load the custom API when we have a user.
40 40
41 Simple fixup of report descriptor 41 Simple fixup of report descriptor
42 --------------------------------- 42 ---------------------------------
43 43
44 In the HID tree, half of the drivers only fix 44 In the HID tree, half of the drivers only fix one key or one byte
45 in the report descriptor. These fixes all requ 45 in the report descriptor. These fixes all require a kernel patch and the
46 subsequent shepherding into a release, a long 46 subsequent shepherding into a release, a long and painful process for users.
47 47
48 We can reduce this burden by providing an eBPF 48 We can reduce this burden by providing an eBPF program instead. Once such a
49 program has been verified by the user, we can 49 program has been verified by the user, we can embed the source code into the
50 kernel tree and ship the eBPF program and load 50 kernel tree and ship the eBPF program and load it directly instead of loading
51 a specific kernel module for it. 51 a specific kernel module for it.
52 52
53 Note: distribution of eBPF programs and their 53 Note: distribution of eBPF programs and their inclusion in the kernel is not
54 yet fully implemented 54 yet fully implemented
55 55
56 Add a new feature that requires a new kernel A 56 Add a new feature that requires a new kernel API
57 ---------------------------------------------- 57 ------------------------------------------------
58 58
59 An example for such a feature are the Universa 59 An example for such a feature are the Universal Stylus Interface (USI) pens.
60 Basically, USI pens require a new kernel API b 60 Basically, USI pens require a new kernel API because there are new
61 channels of communication that our HID and inp 61 channels of communication that our HID and input stack do not support.
62 Instead of using hidraw or creating new sysfs 62 Instead of using hidraw or creating new sysfs entries or ioctls, we can rely
63 on eBPF to have the kernel API controlled by t 63 on eBPF to have the kernel API controlled by the consumer and to not
64 impact the performances by waking up userspace 64 impact the performances by waking up userspace every time there is an
65 event. 65 event.
66 66
67 Morph a device into something else and control 67 Morph a device into something else and control that from userspace
68 ---------------------------------------------- 68 ------------------------------------------------------------------
69 69
70 The kernel has a relatively static mapping of 70 The kernel has a relatively static mapping of HID items to evdev bits.
71 It cannot decide to dynamically transform a gi 71 It cannot decide to dynamically transform a given device into something else
72 as it does not have the required context and a 72 as it does not have the required context and any such transformation cannot be
73 undone (or even discovered) by userspace. 73 undone (or even discovered) by userspace.
74 74
75 However, some devices are useless with that st 75 However, some devices are useless with that static way of defining devices. For
76 example, the Microsoft Surface Dial is a pushb 76 example, the Microsoft Surface Dial is a pushbutton with haptic feedback that
77 is barely usable as of today. 77 is barely usable as of today.
78 78
79 With eBPF, userspace can morph that device int 79 With eBPF, userspace can morph that device into a mouse, and convert the dial
80 events into wheel events. Also, the userspace 80 events into wheel events. Also, the userspace program can set/unset the haptic
81 feedback depending on the context. For example 81 feedback depending on the context. For example, if a menu is visible on the
82 screen we likely need to have a haptic click e 82 screen we likely need to have a haptic click every 15 degrees. But when
83 scrolling in a web page the user experience is 83 scrolling in a web page the user experience is better when the device emits
84 events at the highest resolution. 84 events at the highest resolution.
85 85
86 Firewall 86 Firewall
87 -------- 87 --------
88 88
89 What if we want to prevent other users to acce 89 What if we want to prevent other users to access a specific feature of a
90 device? (think a possibly broken firmware upda 90 device? (think a possibly broken firmware update entry point)
91 91
92 With eBPF, we can intercept any HID command em 92 With eBPF, we can intercept any HID command emitted to the device and
93 validate it or not. 93 validate it or not.
94 94
95 This also allows to sync the state between the 95 This also allows to sync the state between the userspace and the
96 kernel/bpf program because we can intercept an 96 kernel/bpf program because we can intercept any incoming command.
97 97
98 Tracing 98 Tracing
99 ------- 99 -------
100 100
101 The last usage is tracing events and all the f 101 The last usage is tracing events and all the fun we can do we BPF to summarize
102 and analyze events. 102 and analyze events.
103 103
104 Right now, tracing relies on hidraw. It works 104 Right now, tracing relies on hidraw. It works well except for a couple
105 of issues: 105 of issues:
106 106
107 1. if the driver doesn't export a hidraw node, 107 1. if the driver doesn't export a hidraw node, we can't trace anything
108 (eBPF will be a "god-mode" there, so this m 108 (eBPF will be a "god-mode" there, so this may raise some eyebrows)
109 2. hidraw doesn't catch other processes' reque 109 2. hidraw doesn't catch other processes' requests to the device, which
110 means that we have cases where we need to a 110 means that we have cases where we need to add printks to the kernel
111 to understand what is happening. 111 to understand what is happening.
112 112
113 High-level view of HID-BPF 113 High-level view of HID-BPF
114 ========================== 114 ==========================
115 115
116 The main idea behind HID-BPF is that it works 116 The main idea behind HID-BPF is that it works at an array of bytes level.
117 Thus, all of the parsing of the HID report and 117 Thus, all of the parsing of the HID report and the HID report descriptor
118 must be implemented in the userspace component 118 must be implemented in the userspace component that loads the eBPF
119 program. 119 program.
120 120
121 For example, in the dead zone joystick from ab 121 For example, in the dead zone joystick from above, knowing which fields
122 in the data stream needs to be set to ``0`` ne 122 in the data stream needs to be set to ``0`` needs to be computed by userspace.
123 123
124 A corollary of this is that HID-BPF doesn't kn 124 A corollary of this is that HID-BPF doesn't know about the other subsystems
125 available in the kernel. *You can not directly 125 available in the kernel. *You can not directly emit input event through the
126 input API from eBPF*. 126 input API from eBPF*.
127 127
128 When a BPF program needs to emit input events, 128 When a BPF program needs to emit input events, it needs to talk with the HID
129 protocol, and rely on the HID kernel processin 129 protocol, and rely on the HID kernel processing to translate the HID data into
130 input events. 130 input events.
131 131
132 In-tree HID-BPF programs and ``udev-hid-bpf`` 132 In-tree HID-BPF programs and ``udev-hid-bpf``
133 ============================================= 133 =============================================
134 134
135 Official device fixes are shipped in the kerne 135 Official device fixes are shipped in the kernel tree as source in the
136 ``drivers/hid/bpf/progs`` directory. This allo 136 ``drivers/hid/bpf/progs`` directory. This allows to add selftests to them in
137 ``tools/testing/selftests/hid``. 137 ``tools/testing/selftests/hid``.
138 138
139 However, the compilation of these objects is n 139 However, the compilation of these objects is not part of a regular kernel compilation
140 given that they need an external tool to be lo 140 given that they need an external tool to be loaded. This tool is currently
141 `udev-hid-bpf <https://libevdev.pages.freedesk 141 `udev-hid-bpf <https://libevdev.pages.freedesktop.org/udev-hid-bpf/index.html>`_.
142 142
143 For convenience, that external repository dupl 143 For convenience, that external repository duplicates the files from here in
144 ``drivers/hid/bpf/progs`` into its own ``src/b 144 ``drivers/hid/bpf/progs`` into its own ``src/bpf/stable`` directory. This allows
145 distributions to not have to pull the entire k 145 distributions to not have to pull the entire kernel source tree to ship and package
146 those HID-BPF fixes. ``udev-hid-bpf`` also has 146 those HID-BPF fixes. ``udev-hid-bpf`` also has capabilities of handling multiple
147 objects files depending on the kernel the user 147 objects files depending on the kernel the user is running.
148 148
149 Available types of programs 149 Available types of programs
150 =========================== 150 ===========================
151 151
152 HID-BPF is built "on top" of BPF, meaning that 152 HID-BPF is built "on top" of BPF, meaning that we use bpf struct_ops method to
153 declare our programs. 153 declare our programs.
154 154
155 HID-BPF has the following attachment types ava 155 HID-BPF has the following attachment types available:
156 156
157 1. event processing/filtering with ``SEC("stru 157 1. event processing/filtering with ``SEC("struct_ops/hid_device_event")`` in libbpf
158 2. actions coming from userspace with ``SEC("s 158 2. actions coming from userspace with ``SEC("syscall")`` in libbpf
159 3. change of the report descriptor with ``SEC( 159 3. change of the report descriptor with ``SEC("struct_ops/hid_rdesc_fixup")`` or
160 ``SEC("struct_ops.s/hid_rdesc_fixup")`` in 160 ``SEC("struct_ops.s/hid_rdesc_fixup")`` in libbpf
161 161
162 A ``hid_device_event`` is calling a BPF progra 162 A ``hid_device_event`` is calling a BPF program when an event is received from
163 the device. Thus we are in IRQ context and can 163 the device. Thus we are in IRQ context and can act on the data or notify userspace.
164 And given that we are in IRQ context, we can n 164 And given that we are in IRQ context, we can not talk back to the device.
165 165
166 A ``syscall`` means that userspace called the 166 A ``syscall`` means that userspace called the syscall ``BPF_PROG_RUN`` facility.
167 This time, we can do any operations allowed by 167 This time, we can do any operations allowed by HID-BPF, and talking to the device is
168 allowed. 168 allowed.
169 169
170 Last, ``hid_rdesc_fixup`` is different from th 170 Last, ``hid_rdesc_fixup`` is different from the others as there can be only one
171 BPF program of this type. This is called on `` 171 BPF program of this type. This is called on ``probe`` from the driver and allows to
172 change the report descriptor from the BPF prog 172 change the report descriptor from the BPF program. Once a ``hid_rdesc_fixup``
173 program has been loaded, it is not possible to 173 program has been loaded, it is not possible to overwrite it unless the program which
174 inserted it allows us by pinning the program a 174 inserted it allows us by pinning the program and closing all of its fds pointing to it.
175 175
176 Note that ``hid_rdesc_fixup`` can be declared 176 Note that ``hid_rdesc_fixup`` can be declared as sleepable (``SEC("struct_ops.s/hid_rdesc_fixup")``).
177 177
178 178
179 Developer API: 179 Developer API:
180 ============== 180 ==============
181 181
182 Available ``struct_ops`` for HID-BPF: 182 Available ``struct_ops`` for HID-BPF:
183 ------------------------------------- 183 -------------------------------------
184 184
185 .. kernel-doc:: include/linux/hid_bpf.h 185 .. kernel-doc:: include/linux/hid_bpf.h
186 :identifiers: hid_bpf_ops 186 :identifiers: hid_bpf_ops
187 187
188 188
189 User API data structures available in programs 189 User API data structures available in programs:
190 ---------------------------------------------- 190 -----------------------------------------------
191 191
192 .. kernel-doc:: include/linux/hid_bpf.h 192 .. kernel-doc:: include/linux/hid_bpf.h
193 :identifiers: hid_bpf_ctx 193 :identifiers: hid_bpf_ctx
194 194
195 Available API that can be used in all HID-BPF 195 Available API that can be used in all HID-BPF struct_ops programs:
196 ---------------------------------------------- 196 ------------------------------------------------------------------
197 197
198 .. kernel-doc:: drivers/hid/bpf/hid_bpf_dispat 198 .. kernel-doc:: drivers/hid/bpf/hid_bpf_dispatch.c
199 :identifiers: hid_bpf_get_data 199 :identifiers: hid_bpf_get_data
200 200
201 Available API that can be used in syscall HID- 201 Available API that can be used in syscall HID-BPF programs or in sleepable HID-BPF struct_ops programs:
202 ---------------------------------------------- 202 -------------------------------------------------------------------------------------------------------
203 203
204 .. kernel-doc:: drivers/hid/bpf/hid_bpf_dispat 204 .. kernel-doc:: drivers/hid/bpf/hid_bpf_dispatch.c
205 :identifiers: hid_bpf_hw_request hid_bpf_hw 205 :identifiers: hid_bpf_hw_request hid_bpf_hw_output_report hid_bpf_input_report hid_bpf_try_input_report hid_bpf_allocate_context hid_bpf_release_context
206 206
207 General overview of a HID-BPF program 207 General overview of a HID-BPF program
208 ===================================== 208 =====================================
209 209
210 Accessing the data attached to the context 210 Accessing the data attached to the context
211 ------------------------------------------ 211 ------------------------------------------
212 212
213 The ``struct hid_bpf_ctx`` doesn't export the 213 The ``struct hid_bpf_ctx`` doesn't export the ``data`` fields directly and to access
214 it, a bpf program needs to first call :c:func: 214 it, a bpf program needs to first call :c:func:`hid_bpf_get_data`.
215 215
216 ``offset`` can be any integer, but ``size`` ne 216 ``offset`` can be any integer, but ``size`` needs to be constant, known at compile
217 time. 217 time.
218 218
219 This allows the following: 219 This allows the following:
220 220
221 1. for a given device, if we know that the rep 221 1. for a given device, if we know that the report length will always be of a certain value,
222 we can request the ``data`` pointer to poin 222 we can request the ``data`` pointer to point at the full report length.
223 223
224 The kernel will ensure we are using a corre 224 The kernel will ensure we are using a correct size and offset and eBPF will ensure
225 the code will not attempt to read or write 225 the code will not attempt to read or write outside of the boundaries::
226 226
227 __u8 *data = hid_bpf_get_data(ctx, 0 /* o 227 __u8 *data = hid_bpf_get_data(ctx, 0 /* offset */, 256 /* size */);
228 228
229 if (!data) 229 if (!data)
230 return 0; /* ensure data is correct, 230 return 0; /* ensure data is correct, now the verifier knows we
231 * have 256 bytes available 231 * have 256 bytes available */
232 232
233 bpf_printk("hello world: %02x %02x %02x", 233 bpf_printk("hello world: %02x %02x %02x", data[0], data[128], data[255]);
234 234
235 2. if the report length is variable, but we kn 235 2. if the report length is variable, but we know the value of ``X`` is always a 16-bit
236 integer, we can then have a pointer to that 236 integer, we can then have a pointer to that value only::
237 237
238 __u16 *x = hid_bpf_get_data(ctx, offset, 238 __u16 *x = hid_bpf_get_data(ctx, offset, sizeof(*x));
239 239
240 if (!x) 240 if (!x)
241 return 0; /* something went wrong */ 241 return 0; /* something went wrong */
242 242
243 *x += 1; /* increment X by one */ 243 *x += 1; /* increment X by one */
244 244
245 Effect of a HID-BPF program 245 Effect of a HID-BPF program
246 --------------------------- 246 ---------------------------
247 247
248 For all HID-BPF attachment types except for :c 248 For all HID-BPF attachment types except for :c:func:`hid_rdesc_fixup`, several eBPF
249 programs can be attached to the same device. I 249 programs can be attached to the same device. If a HID-BPF struct_ops has a
250 :c:func:`hid_rdesc_fixup` while another is alr 250 :c:func:`hid_rdesc_fixup` while another is already attached to the device, the
251 kernel will return `-EINVAL` when attaching th 251 kernel will return `-EINVAL` when attaching the struct_ops.
252 252
253 Unless ``BPF_F_BEFORE`` is added to the flags 253 Unless ``BPF_F_BEFORE`` is added to the flags while attaching the program, the new
254 program is appended at the end of the list. 254 program is appended at the end of the list.
255 ``BPF_F_BEFORE`` will insert the new program a 255 ``BPF_F_BEFORE`` will insert the new program at the beginning of the list which is
256 useful for e.g. tracing where we need to get t 256 useful for e.g. tracing where we need to get the unprocessed events from the device.
257 257
258 Note that if there are multiple programs using 258 Note that if there are multiple programs using the ``BPF_F_BEFORE`` flag,
259 only the most recently loaded one is actually 259 only the most recently loaded one is actually the first in the list.
260 260
261 ``SEC("struct_ops/hid_device_event")`` 261 ``SEC("struct_ops/hid_device_event")``
262 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 262 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
263 263
264 Whenever a matching event is raised, the eBPF 264 Whenever a matching event is raised, the eBPF programs are called one after the other
265 and are working on the same data buffer. 265 and are working on the same data buffer.
266 266
267 If a program changes the data associated with 267 If a program changes the data associated with the context, the next one will see
268 the modified data but it will have *no* idea o 268 the modified data but it will have *no* idea of what the original data was.
269 269
270 Once all the programs are run and return ``0`` 270 Once all the programs are run and return ``0`` or a positive value, the rest of the
271 HID stack will work on the modified data, with 271 HID stack will work on the modified data, with the ``size`` field of the last hid_bpf_ctx
272 being the new size of the input stream of data 272 being the new size of the input stream of data.
273 273
274 A BPF program returning a negative error disca 274 A BPF program returning a negative error discards the event, i.e. this event will not be
275 processed by the HID stack. Clients (hidraw, i 275 processed by the HID stack. Clients (hidraw, input, LEDs) will **not** see this event.
276 276
277 ``SEC("syscall")`` 277 ``SEC("syscall")``
278 ~~~~~~~~~~~~~~~~~~ 278 ~~~~~~~~~~~~~~~~~~
279 279
280 ``syscall`` are not attached to a given device 280 ``syscall`` are not attached to a given device. To tell which device we are working
281 with, userspace needs to refer to the device b 281 with, userspace needs to refer to the device by its unique system id (the last 4 numbers
282 in the sysfs path: ``/sys/bus/hid/devices/xxxx 282 in the sysfs path: ``/sys/bus/hid/devices/xxxx:yyyy:zzzz:0000``).
283 283
284 To retrieve a context associated with the devi 284 To retrieve a context associated with the device, the program must call
285 hid_bpf_allocate_context() and must release it 285 hid_bpf_allocate_context() and must release it with hid_bpf_release_context()
286 before returning. 286 before returning.
287 Once the context is retrieved, one can also re 287 Once the context is retrieved, one can also request a pointer to kernel memory with
288 hid_bpf_get_data(). This memory is big enough 288 hid_bpf_get_data(). This memory is big enough to support all input/output/feature
289 reports of the given device. 289 reports of the given device.
290 290
291 ``SEC("struct_ops/hid_rdesc_fixup")`` 291 ``SEC("struct_ops/hid_rdesc_fixup")``
292 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 292 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
293 293
294 The ``hid_rdesc_fixup`` program works in a sim 294 The ``hid_rdesc_fixup`` program works in a similar manner to ``.report_fixup``
295 of ``struct hid_driver``. 295 of ``struct hid_driver``.
296 296
297 When the device is probed, the kernel sets the 297 When the device is probed, the kernel sets the data buffer of the context with the
298 content of the report descriptor. The memory a 298 content of the report descriptor. The memory associated with that buffer is
299 ``HID_MAX_DESCRIPTOR_SIZE`` (currently 4kB). 299 ``HID_MAX_DESCRIPTOR_SIZE`` (currently 4kB).
300 300
301 The eBPF program can modify the data buffer at 301 The eBPF program can modify the data buffer at-will and the kernel uses the
302 modified content and size as the report descri 302 modified content and size as the report descriptor.
303 303
304 Whenever a struct_ops containing a ``SEC("stru 304 Whenever a struct_ops containing a ``SEC("struct_ops/hid_rdesc_fixup")`` program
305 is attached (if no program was attached before 305 is attached (if no program was attached before), the kernel immediately disconnects
306 the HID device and does a reprobe. 306 the HID device and does a reprobe.
307 307
308 In the same way, when this struct_ops is detac 308 In the same way, when this struct_ops is detached, the kernel issues a disconnect
309 on the device. 309 on the device.
310 310
311 There is no ``detach`` facility in HID-BPF. De 311 There is no ``detach`` facility in HID-BPF. Detaching a program happens when
312 all the user space file descriptors pointing a 312 all the user space file descriptors pointing at a HID-BPF struct_ops link are closed.
313 Thus, if we need to replace a report descripto 313 Thus, if we need to replace a report descriptor fixup, some cooperation is
314 required from the owner of the original report 314 required from the owner of the original report descriptor fixup.
315 The previous owner will likely pin the struct_ 315 The previous owner will likely pin the struct_ops link in the bpffs, and we can then
316 replace it through normal bpf operations. 316 replace it through normal bpf operations.
317 317
318 Attaching a bpf program to a device 318 Attaching a bpf program to a device
319 =================================== 319 ===================================
320 320
321 We now use standard struct_ops attachment thro 321 We now use standard struct_ops attachment through ``bpf_map__attach_struct_ops()``.
322 But given that we need to attach a struct_ops 322 But given that we need to attach a struct_ops to a dedicated HID device, the caller
323 must set ``hid_id`` in the struct_ops map befo 323 must set ``hid_id`` in the struct_ops map before loading the program in the kernel.
324 324
325 ``hid_id`` is the unique system ID of the HID 325 ``hid_id`` is the unique system ID of the HID device (the last 4 numbers in the
326 sysfs path: ``/sys/bus/hid/devices/xxxx:yyyy:z 326 sysfs path: ``/sys/bus/hid/devices/xxxx:yyyy:zzzz:0000``)
327 327
328 One can also set ``flags``, which is of type ` 328 One can also set ``flags``, which is of type ``enum hid_bpf_attach_flags``.
329 329
330 We can not rely on hidraw to bind a BPF progra 330 We can not rely on hidraw to bind a BPF program to a HID device. hidraw is an
331 artefact of the processing of the HID device, 331 artefact of the processing of the HID device, and is not stable. Some drivers
332 even disable it, so that removes the tracing c 332 even disable it, so that removes the tracing capabilities on those devices
333 (where it is interesting to get the non-hidraw 333 (where it is interesting to get the non-hidraw traces).
334 334
335 On the other hand, the ``hid_id`` is stable fo 335 On the other hand, the ``hid_id`` is stable for the entire life of the HID device,
336 even if we change its report descriptor. 336 even if we change its report descriptor.
337 337
338 Given that hidraw is not stable when the devic 338 Given that hidraw is not stable when the device disconnects/reconnects, we recommend
339 accessing the current report descriptor of the 339 accessing the current report descriptor of the device through the sysfs.
340 This is available at ``/sys/bus/hid/devices/BU 340 This is available at ``/sys/bus/hid/devices/BUS:VID:PID.000N/report_descriptor`` as a
341 binary stream. 341 binary stream.
342 342
343 Parsing the report descriptor is the responsib 343 Parsing the report descriptor is the responsibility of the BPF programmer or the userspace
344 component that loads the eBPF program. 344 component that loads the eBPF program.
345 345
346 An (almost) complete example of a BPF enhanced 346 An (almost) complete example of a BPF enhanced HID device
347 ============================================== 347 =========================================================
348 348
349 *Foreword: for most parts, this could be imple 349 *Foreword: for most parts, this could be implemented as a kernel driver*
350 350
351 Let's imagine we have a new tablet device that 351 Let's imagine we have a new tablet device that has some haptic capabilities
352 to simulate the surface the user is scratching 352 to simulate the surface the user is scratching on. This device would also have
353 a specific 3 positions switch to toggle betwee 353 a specific 3 positions switch to toggle between *pencil on paper*, *cray on a wall*
354 and *brush on a painting canvas*. To make thin 354 and *brush on a painting canvas*. To make things even better, we can control the
355 physical position of the switch through a feat 355 physical position of the switch through a feature report.
356 356
357 And of course, the switch is relying on some u 357 And of course, the switch is relying on some userspace component to control the
358 haptic feature of the device itself. 358 haptic feature of the device itself.
359 359
360 Filtering events 360 Filtering events
361 ---------------- 361 ----------------
362 362
363 The first step consists in filtering events fr 363 The first step consists in filtering events from the device. Given that the switch
364 position is actually reported in the flow of t 364 position is actually reported in the flow of the pen events, using hidraw to implement
365 that filtering would mean that we wake up user 365 that filtering would mean that we wake up userspace for every single event.
366 366
367 This is OK for libinput, but having an externa 367 This is OK for libinput, but having an external library that is just interested in
368 one byte in the report is less than ideal. 368 one byte in the report is less than ideal.
369 369
370 For that, we can create a basic skeleton for o 370 For that, we can create a basic skeleton for our BPF program::
371 371
372 #include "vmlinux.h" 372 #include "vmlinux.h"
373 #include <bpf/bpf_helpers.h> 373 #include <bpf/bpf_helpers.h>
374 #include <bpf/bpf_tracing.h> 374 #include <bpf/bpf_tracing.h>
375 375
376 /* HID programs need to be GPL */ 376 /* HID programs need to be GPL */
377 char _license[] SEC("license") = "GPL"; 377 char _license[] SEC("license") = "GPL";
378 378
379 /* HID-BPF kfunc API definitions */ 379 /* HID-BPF kfunc API definitions */
380 extern __u8 *hid_bpf_get_data(struct hid_bpf 380 extern __u8 *hid_bpf_get_data(struct hid_bpf_ctx *ctx,
381 unsigned int off 381 unsigned int offset,
382 const size_t __s 382 const size_t __sz) __ksym;
383 383
384 struct { 384 struct {
385 __uint(type, BPF_MAP_TYPE_RINGBUF); 385 __uint(type, BPF_MAP_TYPE_RINGBUF);
386 __uint(max_entries, 4096 * 64); 386 __uint(max_entries, 4096 * 64);
387 } ringbuf SEC(".maps"); 387 } ringbuf SEC(".maps");
388 388
389 __u8 current_value = 0; 389 __u8 current_value = 0;
390 390
391 SEC("struct_ops/hid_device_event") 391 SEC("struct_ops/hid_device_event")
392 int BPF_PROG(filter_switch, struct hid_bpf_c 392 int BPF_PROG(filter_switch, struct hid_bpf_ctx *hid_ctx)
393 { 393 {
394 __u8 *data = hid_bpf_get_data(hid_ctx, 394 __u8 *data = hid_bpf_get_data(hid_ctx, 0 /* offset */, 192 /* size */);
395 __u8 *buf; 395 __u8 *buf;
396 396
397 if (!data) 397 if (!data)
398 return 0; /* EPERM check */ 398 return 0; /* EPERM check */
399 399
400 if (current_value != data[152]) { 400 if (current_value != data[152]) {
401 buf = bpf_ringbuf_reserve(&rin 401 buf = bpf_ringbuf_reserve(&ringbuf, 1, 0);
402 if (!buf) 402 if (!buf)
403 return 0; 403 return 0;
404 404
405 *buf = data[152]; 405 *buf = data[152];
406 406
407 bpf_ringbuf_commit(buf, 0); 407 bpf_ringbuf_commit(buf, 0);
408 408
409 current_value = data[152]; 409 current_value = data[152];
410 } 410 }
411 411
412 return 0; 412 return 0;
413 } 413 }
414 414
415 SEC(".struct_ops.link") 415 SEC(".struct_ops.link")
416 struct hid_bpf_ops haptic_tablet = { 416 struct hid_bpf_ops haptic_tablet = {
417 .hid_device_event = (void *)filter_swi 417 .hid_device_event = (void *)filter_switch,
418 }; 418 };
419 419
420 420
421 To attach ``haptic_tablet``, userspace needs t 421 To attach ``haptic_tablet``, userspace needs to set ``hid_id`` first::
422 422
423 static int attach_filter(struct hid *hid_ske 423 static int attach_filter(struct hid *hid_skel, int hid_id)
424 { 424 {
425 int err, link_fd; 425 int err, link_fd;
426 426
427 hid_skel->struct_ops.haptic_tablet->hi 427 hid_skel->struct_ops.haptic_tablet->hid_id = hid_id;
428 err = hid__load(skel); 428 err = hid__load(skel);
429 if (err) 429 if (err)
430 return err; 430 return err;
431 431
432 link_fd = bpf_map__attach_struct_ops(h 432 link_fd = bpf_map__attach_struct_ops(hid_skel->maps.haptic_tablet);
433 if (!link_fd) { 433 if (!link_fd) {
434 fprintf(stderr, "can not attac 434 fprintf(stderr, "can not attach HID-BPF program: %m\n");
435 return -1; 435 return -1;
436 } 436 }
437 437
438 return link_fd; /* the fd of the creat 438 return link_fd; /* the fd of the created bpf_link */
439 } 439 }
440 440
441 Our userspace program can now listen to notifi 441 Our userspace program can now listen to notifications on the ring buffer, and
442 is awaken only when the value changes. 442 is awaken only when the value changes.
443 443
444 When the userspace program doesn't need to lis 444 When the userspace program doesn't need to listen to events anymore, it can just
445 close the returned bpf link from :c:func:`atta 445 close the returned bpf link from :c:func:`attach_filter`, which will tell the kernel to
446 detach the program from the HID device. 446 detach the program from the HID device.
447 447
448 Of course, in other use cases, the userspace p 448 Of course, in other use cases, the userspace program can also pin the fd to the
449 BPF filesystem through a call to :c:func:`bpf_ 449 BPF filesystem through a call to :c:func:`bpf_obj_pin`, as with any bpf_link.
450 450
451 Controlling the device 451 Controlling the device
452 ---------------------- 452 ----------------------
453 453
454 To be able to change the haptic feedback from 454 To be able to change the haptic feedback from the tablet, the userspace program
455 needs to emit a feature report on the device i 455 needs to emit a feature report on the device itself.
456 456
457 Instead of using hidraw for that, we can creat 457 Instead of using hidraw for that, we can create a ``SEC("syscall")`` program
458 that talks to the device:: 458 that talks to the device::
459 459
460 /* some more HID-BPF kfunc API definitions * 460 /* some more HID-BPF kfunc API definitions */
461 extern struct hid_bpf_ctx *hid_bpf_allocate_ 461 extern struct hid_bpf_ctx *hid_bpf_allocate_context(unsigned int hid_id) __ksym;
462 extern void hid_bpf_release_context(struct h 462 extern void hid_bpf_release_context(struct hid_bpf_ctx *ctx) __ksym;
463 extern int hid_bpf_hw_request(struct hid_bpf 463 extern int hid_bpf_hw_request(struct hid_bpf_ctx *ctx,
464 __u8* data, 464 __u8* data,
465 size_t len, 465 size_t len,
466 enum hid_report_ 466 enum hid_report_type type,
467 enum hid_class_r 467 enum hid_class_request reqtype) __ksym;
468 468
469 469
470 struct hid_send_haptics_args { 470 struct hid_send_haptics_args {
471 /* data needs to come at offset 0 so w 471 /* data needs to come at offset 0 so we can do a memcpy into it */
472 __u8 data[10]; 472 __u8 data[10];
473 unsigned int hid; 473 unsigned int hid;
474 }; 474 };
475 475
476 SEC("syscall") 476 SEC("syscall")
477 int send_haptic(struct hid_send_haptics_args 477 int send_haptic(struct hid_send_haptics_args *args)
478 { 478 {
479 struct hid_bpf_ctx *ctx; 479 struct hid_bpf_ctx *ctx;
480 int ret = 0; 480 int ret = 0;
481 481
482 ctx = hid_bpf_allocate_context(args->h 482 ctx = hid_bpf_allocate_context(args->hid);
483 if (!ctx) 483 if (!ctx)
484 return 0; /* EPERM check */ 484 return 0; /* EPERM check */
485 485
486 ret = hid_bpf_hw_request(ctx, 486 ret = hid_bpf_hw_request(ctx,
487 args->data, 487 args->data,
488 10, 488 10,
489 HID_FEATURE_R 489 HID_FEATURE_REPORT,
490 HID_REQ_SET_R 490 HID_REQ_SET_REPORT);
491 491
492 hid_bpf_release_context(ctx); 492 hid_bpf_release_context(ctx);
493 493
494 return ret; 494 return ret;
495 } 495 }
496 496
497 And then userspace needs to call that program 497 And then userspace needs to call that program directly::
498 498
499 static int set_haptic(struct hid *hid_skel, 499 static int set_haptic(struct hid *hid_skel, int hid_id, __u8 haptic_value)
500 { 500 {
501 int err, prog_fd; 501 int err, prog_fd;
502 int ret = -1; 502 int ret = -1;
503 struct hid_send_haptics_args args = { 503 struct hid_send_haptics_args args = {
504 .hid = hid_id, 504 .hid = hid_id,
505 }; 505 };
506 DECLARE_LIBBPF_OPTS(bpf_test_run_opts, 506 DECLARE_LIBBPF_OPTS(bpf_test_run_opts, tattrs,
507 .ctx_in = &args, 507 .ctx_in = &args,
508 .ctx_size_in = sizeof(args), 508 .ctx_size_in = sizeof(args),
509 ); 509 );
510 510
511 args.data[0] = 0x02; /* report ID of t 511 args.data[0] = 0x02; /* report ID of the feature on our device */
512 args.data[1] = haptic_value; 512 args.data[1] = haptic_value;
513 513
514 prog_fd = bpf_program__fd(hid_skel->pr 514 prog_fd = bpf_program__fd(hid_skel->progs.set_haptic);
515 515
516 err = bpf_prog_test_run_opts(prog_fd, 516 err = bpf_prog_test_run_opts(prog_fd, &tattrs);
517 return err; 517 return err;
518 } 518 }
519 519
520 Now our userspace program is aware of the hapt 520 Now our userspace program is aware of the haptic state and can control it. The
521 program could make this state further availabl 521 program could make this state further available to other userspace programs
522 (e.g. via a DBus API). 522 (e.g. via a DBus API).
523 523
524 The interesting bit here is that we did not cr 524 The interesting bit here is that we did not created a new kernel API for this.
525 Which means that if there is a bug in our impl 525 Which means that if there is a bug in our implementation, we can change the
526 interface with the kernel at-will, because the 526 interface with the kernel at-will, because the userspace application is
527 responsible for its own usage. 527 responsible for its own usage.
Linux® is a registered trademark of Linus Torvalds in the United States and other countries.
TOMOYO® is a registered trademark of NTT DATA CORPORATION.