1 ============================================================================
5 Readme file for the Controller Area Network Protocol Family (aka Socket CAN)
9 1 Overview / What is Socket CAN
11 2 Motivation / Why using the socket API
16 3.3 network security issues (capabilities)
17 3.4 network problem notifications
19 4 How to use Socket CAN
20 4.1 RAW protocol sockets with can_filters (SOCK_RAW)
21 4.1.1 RAW socket option CAN_RAW_FILTER
22 4.1.2 RAW socket option CAN_RAW_ERR_FILTER
23 4.1.3 RAW socket option CAN_RAW_LOOPBACK
24 4.1.4 RAW socket option CAN_RAW_RECV_OWN_MSGS
25 4.2 Broadcast Manager protocol sockets (SOCK_DGRAM)
26 4.3 connected transport protocols (SOCK_SEQPACKET)
27 4.4 unconnected transport protocols (SOCK_DGRAM)
29 5 Socket CAN core module
30 5.1 can.ko module params
32 5.3 writing own CAN protocol modules
37 6.3 CAN controller hardware filters
38 6.4 currently supported CAN hardware
43 ============================================================================
45 1. Overview / What is Socket CAN
46 --------------------------------
48 The socketcan package is an implementation of CAN protocols
49 (Controller Area Network) for Linux. CAN is a networking technology
50 which has wide-spread use in automation, embedded devices, and
51 automotive fields. While there have been other CAN implementations
52 for Linux based on character devices, Socket CAN uses the Berkeley
53 socket API, the Linux network stack and implements the CAN device
54 drivers as network interfaces. The CAN socket API has been designed
55 as similar as possible to the TCP/IP protocols to allow programmers,
56 familiar with network programming, to easily learn how to use CAN
59 2. Motivation / Why using the socket API
60 ----------------------------------------
62 There have been CAN implementations for Linux before Socket CAN so the
63 question arises, why we have started another project. Most existing
64 implementations come as a device driver for some CAN hardware, they
65 are based on character devices and provide comparatively little
66 functionality. Usually, there is only a hardware-specific device
67 driver which provides a character device interface to send and
68 receive raw CAN frames, directly to/from the controller hardware.
69 Queueing of frames and higher-level transport protocols like ISO-TP
70 have to be implemented in user space applications. Also, most
71 character-device implementations support only one single process to
72 open the device at a time, similar to a serial interface. Exchanging
73 the CAN controller requires employment of another device driver and
74 often the need for adaption of large parts of the application to the
77 Socket CAN was designed to overcome all of these limitations. A new
78 protocol family has been implemented which provides a socket interface
79 to user space applications and which builds upon the Linux network
80 layer, so to use all of the provided queueing functionality. Device
81 drivers for CAN controller hardware register itself with the Linux
82 network layer as a network device, so that CAN frames from the
83 controller can be passed up to the network layer and on to the CAN
84 protocol family module and also vice-versa. Also, the protocol family
85 module provides an API for transport protocol modules to register, so
86 that any number of transport protocols can be loaded or unloaded
87 dynamically. In fact, the can core module alone does not provide any
88 protocol and can not be used without loading at least one additional
89 protocol module. Multiple sockets can be opened at the same time,
90 on different or the same protocol module and they can listen/send
91 frames on different or the same CAN IDs. Several sockets listening on
92 the same interface for frames with the same CAN ID are all passed the
93 same received matching CAN frames. An application wishing to
94 communicate using a specific transport protocol, e.g. ISO-TP, just
95 selects that protocol when opening the socket, and then can read and
96 write application data byte streams, without having to deal with
99 Similar functionality visible from user-space could be provided by a
100 character decive, too, but this would lead to a technically inelegant
101 solution for a couple of reasons:
103 * Intricate usage. Instead of passing a protocol argument to
104 socket(2) and using bind(2) to select a CAN interface and CAN ID, an
105 application would have to do all these operations using ioctl(2)s.
107 * Code duplication. A character device cannot make use of the Linux
108 network queueing code, so all that code would have to be duplicated
111 * Abstraction. In most existing character-device implementations, the
112 hardware-specific device driver for a CAN controller directly
113 provides the character device for the application to work with.
114 This is at least very unusual in Unix systems, for both, char and
115 block devices. For example you don't have a character device for a
116 certain UART of a serial interface, a certain sound chip in your
117 computer, a SCSI or IDE controller providing access to your hard
118 disk or tape streamer device. Instead, you have abstraction layers
119 which provide a unified character or block device interface to the
120 application on the one hand, and a interface for hardware-specific
121 device drivers on the other hand. These abstractions are provided
122 by subsystems like the tty layer, the audio subsystem or the SCSI
123 and IDE subsystems for the devices mentioned above.
125 The easiest way to implement a CAN device driver is as a character
126 without such a (complete) abstraction layer, as is done by most
127 existing drivers. The right way, however, would be to add such a
128 layer with all the functionality like registering for certain CAN
129 IDs, supporting several open file descriptors and (de)multplexing
130 CAN frames between them, (sophisticated) queueing of CAN frames, and
131 providing an API for device driver to register with. However, then
132 it would be no more difficult, or may be even easier, to use the
133 networking framework provided by the Linux kernel, and this is what
136 The use of the networking framework of the Linux kernel is just the
137 natural and most appropriate way to implement CAN for Linux.
139 3. Socket CAN concept
140 ---------------------
142 As described in chapter 2 it is the main goal of Socket CAN to
143 provide a socket interface to user space applications which builds
144 upon the Linux networklayer. In opposite to the commonly known
145 TCP/IP and ethernet networking the CAN bus is a broadcast-only(!)
146 medium that has no MAC-layer adressing like ethernet. The CAN-identifier
147 (can_id) is used for arbitration on the CAN-bus. Therefore the CAN-IDs
148 have to be choosen unique on the bus. When designing a CAN-ECU
149 network the CAN-IDs are mapped to be sent by a specific ECU.
150 For this reason a CAN-ID can be treatened best as a kind of source address.
154 The network transparent access of multiple applications leads to the
155 problem that different applications may be interrested in the same
156 CAN-IDs from the same CAN network interface. The Socket CAN core
157 module - which implements the protocol family CAN - provides several
158 high efficient receive lists for this reason. If e.g. a user space
159 application opens a CAN RAW socket, the raw protocol module itself
160 requests the (range of) CAN-IDs from the Socket CAN core that are
161 requested by the user. The subscription and unsubscription of
162 CAN-IDs can be done for specific CAN interfaces or for all(!) known
163 CAN interfaces with the can_rx_(un)register() functions provided to
164 CAN protocol modules by the SocketCAN core (see chapter 5).
165 To optimize the CPU usage at runtime the receive lists are split up
166 into several specific lists per device that match the requested
167 filter complexity for a given use-case.
171 As known from other networking concepts the data exchanging
172 applications may run on the same or different nodes without any
173 change (except if the according addressing information):
175 ___ ___ ___ _______ ___
176 | _ | | _ | | _ | | _ _ | | _ |
177 ||A|| ||B|| ||C|| ||A| |B|| ||C||
178 |___| |___| |___| |_______| |___|
180 -----------------(1)- CAN bus -(2)---------------
182 To ensure that application A receives the same information in the
183 expample (2) as it would receive in example (1) there is need for
184 some kind of local loopback on the appropriate node.
186 The Linux network devices (by default) just can handle the
187 transmission and receiption of media dependend frames. Due to the
188 arbritration on the CAN bus the transmission of a low prio CAN-ID
189 may be delayed from the receipition of a high prio CAN frame. To
190 reflect the correct* traffic on the node the loopback of the sent
191 data has to be performed right after a successful transmission. If
192 the CAN network interface is not capable to perform the loopback for
193 some reason the SocketCAN core can do this task as a fallback solution.
194 See chapter 6.2 for details (recommended).
196 The loopback functionality is enabled by default to reflect standard
197 networking behaviour for CAN applications. Due to some requests from
198 the RT-SocketCAN group the loopback optionally may be disabled for each
199 seperate socket. See sockopts from the CAN RAW sockets in chapter 4.1 .
201 * = you really like to have this when you're running analyser tools
202 like 'candump' or 'cansniffer' on the (same) node.
204 3.3 network security issues (capabilities)
206 The Controller Area Network is a local field bus transmitting only
207 broadcast messages without any routing and security concepts.
208 In the majority of cases the user application has to deal with
209 raw CAN frames. Therefore it might be reasonable NOT to restrict
210 the CAN access only to the user root, as known from other networks.
211 Since the currently implemented CAN_RAW and CAN_BCM sockets can only
212 send and receive frames to/from CAN interfaces it does not affect
213 security of others networks to allow all users to access the CAN.
214 To enable non-root users to access CAN_RAW and CAN_BCM protocol
215 sockets the Kconfig options CAN_RAW_USER and/or CAN_BCM_USER may be
216 selected at kernel compile time.
218 3.4 network problem notifications
220 The use of the CAN bus may lead to several problems on the physical
221 and media access control layer. Detecting and logging of these lower
222 layer problems is a vital requirement for CAN users to identify
223 hardware issues on the physical transceiver layer as well as
224 arbitration problems and error frames caused by the different
225 ECUs. The occurance of detected errors are important for diagnosis
226 and have to be logged together with the exact timestamp. For this
227 reason the CAN interface driver can generate so called Error Frames
228 that can optionally be passed to the user application on the same
229 way like other CAN frames. Whenever an error on the physical layer
230 or the MAC layer is detected (e.g. by the CAN controller) the driver
231 creates an appropriate error frame. Error frames can be requested by
232 the user application using the common CAN filter mechanisms. Inside
233 this filter definition the (interrested) type of errors may be
234 selected. The receiption of error frames is disabled by default.
236 4. How to use Socket CAN
237 ------------------------
239 Like TCP/IP, you first need to open a socket for communicating over a
240 CAN network. Since Socket CAN implements a new protocol family, you
241 need to pass PF_CAN as the first argument to the socket(2) system
242 call. Currently, there are two CAN protocols to choose from, the raw
243 socket protocol and the broadcast manager (BCM). So to open a socket,
246 s = socket(PF_CAN, SOCK_RAW, CAN_RAW);
250 s = socket(PF_CAN, SOCK_DGRAM, CAN_BCM);
252 respectively. After the successful creation of the socket, you would
253 normally use the bind(2) system call to bind the socket to a CAN
254 interface (which is different to TCP/IP due to different addressing
255 - see chapter 3). After binding (CAN_RAW) or connecting (CAN_BCM)
256 the socket, you can read(2) and write(2) from/to the socket or use
257 send(2), sendto(2), sendmsg(2) and the recv* counterpart operations
258 on the socket as usual. There are also CAN specific socket options
261 The basic CAN frame structure and the sockaddr structure are defined
262 in include/linux/can.h:
265 canid_t can_id; /* 32 bit CAN_ID + EFF/RTR/ERR flags */
266 __u8 can_dlc; /* data length code: 0 .. 8 */
267 __u8 data[8] __attribute__((aligned(8)));
270 The alignment of the (linear) payload data[] to a 64bit boundary
271 allows the user to define own structs and unions to easily access the
272 CAN payload. There is no given byteorder on the CAN bus by
273 default. A read(2) system call on a CAN_RAW socket transfers a
274 struct can_frame to the user space.
276 The sockaddr_can structure has an interface index analogue to the
277 PF_PACKET socket, that also binds to a specific interface:
279 struct sockaddr_can {
280 sa_family_t can_family;
283 struct { canid_t rx_id, tx_id; } tp16;
284 struct { canid_t rx_id, tx_id; } tp20;
285 struct { canid_t rx_id, tx_id; } mcnet;
286 struct { canid_t rx_id, tx_id; } isotp;
287 struct { int lcu, type; } bap;
291 To determine the interface index the an appropriate ioctl() has to
292 be used (example for CAN_RAW sockets without error checking):
295 struct sockaddr_can addr;
298 s = socket(PF_CAN, SOCK_RAW, CAN_RAW);
300 strcpy(ifr.ifr_name, "can0" );
301 ioctl(s, SIOCGIFINDEX, &ifr);
303 addr.can_family = AF_CAN;
304 addr.can_ifindex = ifr.ifr_ifindex;
306 bind(s, (struct sockaddr *)&addr, sizeof(addr));
310 To bind a socket to all(!) CAN interfaces the interface index might
311 be 0 (zero). In this case the socket receives CAN frames from every
312 enabled CAN interface. To determine the originating CAN interface
313 the system call recvfrom(2) may be used instead of read(2). To send
314 on a socket that is bound to 'any' interface sendto(2) is needed to
315 specify the outgoing interface.
317 Reading CAN frames from a bound CAN_RAW socket (see above) consists
318 of reading a struct can_frame:
320 struct can_frame frame;
322 nbytes = read(s, &frame, sizeof(struct can_frame));
325 perror("can raw socket read");
329 /* paraniod check ... */
330 if (nbytes < sizeof(struct can_frame)) {
331 fprintf(stderr, "read: incomplete CAN frame\n");
335 /* do something with the received CAN frame */
337 Writing CAN frames can be done analogue with the write(2) system call:
339 nbytes = write(s, &frame, sizeof(struct can_frame));
341 When the CAN interface is bound to 'any' existing CAN interface
342 (addr.can_ifindex = 0) it is recommended to use recvfrom(2) if the
343 information about the originating CAN interface is needed:
345 struct sockaddr_can addr;
347 socklen_t len = sizeof(addr);
348 struct can_frame frame;
350 nbytes = recvfrom(s, &frame, sizeof(struct can_frame),
351 0, (struct sockaddr*)&addr, &len);
353 /* get interface name of the received CAN frame */
354 ifr.ifr_ifindex = addr.can_ifindex;
355 ioctl(s, SIOCGIFNAME, &ifr);
356 printf("Received a CAN frame from interface %s", ifr.ifr_name);
358 To write CAN frames on sockets bound to 'any' CAN interface the
359 outgoing interface has to be defined certainly.
361 strcpy(ifr.ifr_name, "can0");
362 ioctl(s, SIOCGIFINDEX, &ifr);
363 addr.can_ifindex = ifr.ifr_ifindex;
364 addr.can_family = AF_CAN;
366 nbytes = sendto(s, &frame, sizeof(struct can_frame),
367 0, (struct sockaddr*)&addr, sizeof(addr));
369 4.1 RAW protocol sockets with can_filters (SOCK_RAW)
371 Using CAN_RAW sockets is extensively comparable to the commonly
372 known access to CAN character devices. To meet the new possibilities
373 provided by the multi user SocketCAN approach, some reasonable
374 defaults are set at RAW socket bindung time:
376 - The filters are set to exactly one filter receiving everything
377 - The socket only receives valid data frames (=> no error frames)
378 - The loopback of sent CAN frames is enabled (see chapter 3.2)
379 - The socket does not receive it's own sent frames (in loopback mode)
381 These default settings may be changed before or after binding the socket.
382 To use the referenced definitions of the socket options for CAN_RAW
383 sockets include linux/can/raw.h .
385 4.1.1 RAW socket option CAN_RAW_FILTER
387 The receiption of CAN frames using CAN_RAW sockets can be controlled
388 by defining 0 .. n filters with the CAN_RAW_FILTER socket option.
390 The CAN filter structure is defined in include/linux/can.h:
397 A filter matches, when
399 <received_can_id> & mask == can_id & mask
401 which is analogue to known CAN controllers hardware filter semantics.
402 The filter can be inverted in this semantic, when the CAN_INV_FILTER
403 bit is set in can_id element of the can_filter structure. In
404 opposite to CAN controller hardware filters the user may set 0 .. n
405 receive filters for each open socket separately:
407 struct can_filter rfilter[2];
409 rfilter[0].can_id = 0x123;
410 rfilter[0].can_mask = CAN_SFF_MASK;
411 rfilter[1].can_id = 0x200;
412 rfilter[1].can_mask = 0x700;
414 setsockopt(s, SOL_CAN_RAW, CAN_RAW_FILTER, &rfilter, sizeof(rfilter));
416 To disable the receiption of CAN frames on the selected CAN_RAW socket:
418 setsockopt(s, SOL_CAN_RAW, CAN_RAW_FILTER, NULL, 0);
420 To set the filters to zero filters is quite obsolete as not readed
421 data causes the raw socket to discard the received CAN frames. But
422 having this 'send only' use-case we may remove the receive list in the
423 Kernel to save a little (really a very little!) CPU usage.
425 4.1.2 RAW socket option CAN_RAW_ERR_FILTER
427 As described in chapter 3.4 the CAN interface driver can generate so
428 called Error Frames that can optionally be passed to the user
429 application on the same way like other CAN frames. The possible
430 errors are devided into different error classes that may be filtered
431 using the appropriate error mask. To register for every possible
432 error condition CAN_ERR_MASK can be used as value for the error mask.
433 The values for the error mask are defined in linux/can/error.h .
435 can_err_mask_t err_mask = ( CAN_ERR_TX_TIMEOUT | CAN_ERR_BUSOFF );
437 setsockopt(s, SOL_CAN_RAW, CAN_RAW_ERR_FILTER,
438 &err_mask, sizeof(err_mask));
440 4.1.3 RAW socket option CAN_RAW_LOOPBACK
442 To meet multi user needs the local loopback is enabled by default
443 (see chapter 3.2 for details). But in some embedded use-cases
444 (e.g. when only one application uses the CAN bus) this loopback
445 functionality can be disabled (separately for each socket):
447 int loopback = 0; /* 0 = disabled, 1 = enabled (default) */
449 setsockopt(s, SOL_CAN_RAW, CAN_RAW_LOOPBACK, &loopback, sizeof(loopback));
451 4.1.4 RAW socket option CAN_RAW_RECV_OWN_MSGS
453 When the local loopback is enabled, all the sent CAN frames are
454 looped back to the open CAN sockets that registered for the CAN
455 frames' CAN-ID on this given interface to meet the multi user
456 needs. The receiption of the CAN frames on the same socket that was
457 sending the CAN frame is assumed to be unwanted and therefore
458 disabled by default. This default behaviour may be changed on
461 int set_recv_own_msgs = 1; /* 0 = disabled (default), 1 = enabled */
463 setsockopt(s, SOL_CAN_RAW, CAN_RAW_RECV_OWN_MSGS,
464 &recv_own_msgs, sizeof(recv_own_msgs));
466 4.2 Broadcast Manager protocol sockets (SOCK_DGRAM)
467 4.3 connected transport protocols (SOCK_SEQPACKET)
468 4.4 unconnected transport protocols (SOCK_DGRAM)
471 5. Socket CAN core module
472 -------------------------
474 The Socket CAN core module implements the protocol family
475 PF_CAN. CAN protocol modules are loaded by the core module at
476 runtime. The core module provides an interface for CAN protocol
477 modules to subscribe needed CAN IDs (see chapter 3.1).
479 5.1 can.ko module params
481 - stats_timer: To calculate the Socket CAN core statistics
482 (e.g. current/maximum frames per second) this 1 second timer is
483 invoked at can.ko module start time by default. This timer can be
484 disabled giving stattimer=0 on the module comandline.
486 - debug: When the Kconfig option CONFIG_CAN_DEBUG_CORE is set at
487 compile time, the debug output code is compiled into the module.
488 debug = 0x01 => print general debug information
489 debug = 0x02 => print content of processed CAN frames
490 debug = 0x04 => print content of processed socket buffers
492 It is possible or have ORed values e.g. 3 or 7 for an output off
493 all available debug information. Using 0x02 and 0x04 may flood
494 your kernel log - so be careful.
498 As described in chapter 3.1 the Socket CAN core uses several filter
499 lists to deliver received CAN frames to CAN protocol modules. These
500 receive lists, their filters and the count of filter matches can be
501 checked in the appropriate receive list. All entries contain the
502 device and a protocol module identifier:
504 foo@bar:~$ cat /proc/net/can/rcvlist_all
506 receive list 'rx_all':
510 device can_id can_mask function userdata matches ident
511 vcan0 000 00000000 f88e6370 f6c6f400 0 raw
514 In this example an application requests any CAN traffic from vcan0.
516 rcvlist_all - list for unfiltered entries (no filter operations)
517 rcvlist_eff - list for single extended frame (EFF) entries
518 rcvlist_err - list for error frames masks
519 rcvlist_fil - list for mask/value filters
520 rcvlist_inv - list for mask/value filters (inverse semantic)
521 rcvlist_sff - list for single standard frame (SFF) entries
523 Additional procfs files in /proc/net/can
525 stats - Socket CAN core statistics (rx/tx frames, match ratios, ...)
526 reset_stats - manual statistic reset
527 version - prints the Socket CAN core version and the ABI version
529 5.3 writing own CAN protocol modules
531 To implement a new protocol in the protocol family PF_CAN a new
532 protocol has to be defined in include/linux/can.h .
533 The prototypes and definitions to use the Socket CAN core can be
534 accessed by including include/linux/can/core.h .
535 Additionally to functions that register the CAN protocol and the
536 CAN device notifier chain there are functions to subscribe CAN
537 frames received by CAN interfaces and to send CAN frames:
539 can_rx_register - subscribe CAN frames from a specific interface
540 can_rx_unregister - unsubscribe CAN frames from a specific interface
541 can_send - transmit a CAN frame (optional with local loopback)
543 For details see the kerneldoc documentation in net/can/af_can.c or
544 the source code of net/can/raw.c or net/can/bcm.c .
546 6. CAN network drivers
547 ----------------------
549 Writing a CAN network device driver is much easier than writing a
550 CAN character device driver. Analogue to other know network device
551 drivers you mainly have to deal with:
553 - TX: Put the CAN frame from the socket buffer to the CAN controller.
554 - RX: Put the CAN frame from the CAN controller to the socket buffer.
556 See e.g. at Documentation/networking/netdevices.txt . The differences
557 for writing CAN network device driver are described below:
561 dev->type = ARPHRD_CAN; /* the netdevice hardware type */
562 dev->flags = IFF_NOARP; /* CAN has no arp */
564 dev->mtu = sizeof(struct can_frame);
566 The struct can_frame is the payload of each socket buffer in the
567 protocol family PF_CAN.
571 As described in chapter 3.2 the CAN network device driver should
572 support a local loopback functionality. If so the driver flag
573 IFF_LOOPBACK has to be set to omit the PF_CAN core to perform the
574 loopback as fallback solution:
576 dev->flags = (IFF_NOARP | IFF_LOOPBACK);
578 6.3 CAN controller hardware filters
580 To reduce the interrupt load on deep embedded systems some CAN
581 controllers support the filtering of CAN IDs or ranges of CAN IDs.
582 These hardware filter capabilities vary from controller to
583 controller and have to be identified as not feasible in a multi-user
584 networking approach. The use of the very controller specific
585 hardware filters could make sense in a very dedicated use-case, as a
586 filter on driver level would affect all users in the multi-user
587 system. The high efficient filter sets inside the PF_CAN core allow
588 to set different multiple filters for each socket separately.
589 Therefore the use of hardware filters goes to the category 'handmade
590 tuning on deep embedded systems'. The author is running a MPC603e
591 @133MHz with four SJA1000 CAN controllers from 2002 under heavy bus
592 load without any problems ...
594 6.4 currently supported CAN hardware (May 2007)
596 On the project website http://developer.berlios.de/projects/socketcan
597 there are different drivers available:
599 vcan: Virtual CAN interface driver (if no real hardware is available)
600 sja1000: Philips SJA1000 CAN controller (recommended)
601 i82527: Intel i82527 CAN controller
602 mscan: Motorola/Freescale CAN controller (e.g. inside SOC MPC5200)
603 slcan: For a bunch of CAN adaptors that are attached via a
604 serial line ASCII protocol (for serial / USB adaptors)
606 Additionally the different CAN adaptors (ISA/PCI/PCMCIA/USB/Parport)
607 from PEAK Systemtechnik support the CAN netdevice driver modell
608 since Linux driver v6.0: http://www.peak-system.com/linux/index.htm
610 Please check the Mailing Lists on the berlios OSS project website.
614 The configuration interface for CAN network drivers is still an open
615 issue that has not been finalized in the socketcan project. Also the
616 idea of having a library module (candev.ko) that holds functions
617 that are needed by all CAN netdevices is not ready to ship.
618 Your contribution is welcome.
623 Oliver Hartkopp (PF_CAN core, filters, drivers, bcm)
624 Urs Thuermann (PF_CAN core, kernel integration, socket interfaces, raw, vcan)
625 Jan Kizka (RT-SocketCAN core, Socket-API reconciliation)
626 Wolfgang Grandegger (RT-SocketCAN core & drivers, Raw Socket-API reviews)
627 Robert Schwebel (design reviews, PTXdist integration)
628 Marc Kleine-Budde (design reviews, Kernel 2.6 cleanups, drivers)
629 Benedikt Spranger (reviews)
630 Thomas Gleixner (LKML reviews, coding style, posting hints)
631 Andrey Volkov (kernel subtree structure, ioctls, mscan driver)
632 Matthias Brukner (first SJA1000 CAN netdevice implementation Q2/2003)
633 Klaus Hitschler (PEAK driver integration)
634 Uwe Koppe (CAN netdevices with PF_PACKET approach)
635 Michael Schulze (driver layer loopback requirement, RT CAN drivers review)