1 | :orphan: |
---|
2 | |
---|
3 | |
---|
4 | |
---|
5 | .. COMMENT: %**end of header |
---|
6 | |
---|
7 | .. COMMENT: COPYRIGHT (c) 1989-2013. |
---|
8 | |
---|
9 | .. COMMENT: On-Line Applications Research Corporation (OAR). |
---|
10 | |
---|
11 | .. COMMENT: All rights reserved. |
---|
12 | |
---|
13 | .. COMMENT: Master file for the network Supplement |
---|
14 | |
---|
15 | .. COMMENT: COPYRIGHT (c) 1988-2002. |
---|
16 | |
---|
17 | .. COMMENT: On-Line Applications Research Corporation (OAR). |
---|
18 | |
---|
19 | .. COMMENT: All rights reserved. |
---|
20 | |
---|
21 | .. COMMENT: The following determines which set of the tables and figures we will use. |
---|
22 | |
---|
23 | .. COMMENT: We default to ASCII but if available TeX or HTML versions will |
---|
24 | |
---|
25 | .. COMMENT: be used instead. |
---|
26 | |
---|
27 | .. COMMENT: @clear use-html |
---|
28 | |
---|
29 | .. COMMENT: @clear use-tex |
---|
30 | |
---|
31 | .. COMMENT: The following variable says to use texinfo or html for the two column |
---|
32 | |
---|
33 | .. COMMENT: texinfo tables. For somethings the format does not look good in html. |
---|
34 | |
---|
35 | .. COMMENT: With our adjustment to the left column in TeX, it nearly always looks |
---|
36 | |
---|
37 | .. COMMENT: good printed. |
---|
38 | |
---|
39 | .. COMMENT: Custom whitespace adjustments. We could fiddle a bit more. |
---|
40 | |
---|
41 | .. COMMENT: Title Page Stuff |
---|
42 | |
---|
43 | .. COMMENT: I don't really like having a short title page. -joel |
---|
44 | |
---|
45 | .. COMMENT: @shorttitlepage RTEMS Network Supplement |
---|
46 | |
---|
47 | ======================== |
---|
48 | RTEMS Network Supplement |
---|
49 | ======================== |
---|
50 | |
---|
51 | .. COMMENT: COPYRIGHT (c) 1988-2015. |
---|
52 | |
---|
53 | .. COMMENT: On-Line Applications Research Corporation (OAR). |
---|
54 | |
---|
55 | .. COMMENT: All rights reserved. |
---|
56 | |
---|
57 | .. COMMENT: The following puts a space somewhere on an otherwise empty page so we |
---|
58 | |
---|
59 | .. COMMENT: can force the copyright description onto a left hand page. |
---|
60 | |
---|
61 | COPYRIGHT © 1988 - 2015. |
---|
62 | |
---|
63 | On-Line Applications Research Corporation (OAR). |
---|
64 | |
---|
65 | The authors have used their best efforts in preparing |
---|
66 | this material. These efforts include the development, research, |
---|
67 | and testing of the theories and programs to determine their |
---|
68 | effectiveness. No warranty of any kind, expressed or implied, |
---|
69 | with regard to the software or the material contained in this |
---|
70 | document is provided. No liability arising out of the |
---|
71 | application or use of any product described in this document is |
---|
72 | assumed. The authors reserve the right to revise this material |
---|
73 | and to make changes from time to time in the content hereof |
---|
74 | without obligation to notify anyone of such revision or changes. |
---|
75 | |
---|
76 | The RTEMS Project is hosted at http://www.rtems.org. Any |
---|
77 | inquiries concerning RTEMS, its related support components, or its |
---|
78 | documentation should be directed to the Community Project hosted athttp://www.rtems.org. |
---|
79 | |
---|
80 | Any inquiries for commercial services including training, support, custom |
---|
81 | development, application development assistance should be directed tohttp://www.rtems.com. |
---|
82 | |
---|
83 | .. COMMENT: This prevents a black box from being printed on "overflow" lines. |
---|
84 | |
---|
85 | .. COMMENT: The alternative is to rework a sentence to avoid this problem. |
---|
86 | |
---|
87 | RTEMS TCP/IP Networking Supplement |
---|
88 | ################################## |
---|
89 | |
---|
90 | .. COMMENT: COPYRIGHT (c) 1989-2011. |
---|
91 | |
---|
92 | .. COMMENT: On-Line Applications Research Corporation (OAR). |
---|
93 | |
---|
94 | .. COMMENT: All rights reserved. |
---|
95 | |
---|
96 | Preface |
---|
97 | ####### |
---|
98 | |
---|
99 | This document describes the RTEMS specific parts of the FreeBSD TCP/IP |
---|
100 | stack. Much of this documentation was written by Eric Norum |
---|
101 | (eric@skatter.usask.ca) |
---|
102 | of the Saskatchewan Accelerator Laboratory |
---|
103 | who also ported the FreeBSD TCP/IP stack to RTEMS. |
---|
104 | |
---|
105 | The following is a list of resources which should be useful in trying |
---|
106 | to understand Ethernet: |
---|
107 | |
---|
108 | - *Charles Spurgeonâs Ethernet Web Site* |
---|
109 | "This site provides extensive information about Ethernet |
---|
110 | (IEEE 802.3) local area network (LAN) technology. Including |
---|
111 | the original 10 Megabit per second (Mbps) system, the 100 Mbps |
---|
112 | Fast Ethernet system (802.3u), and the Gigabit Ethernet system (802.3z)." |
---|
113 | The URL is: |
---|
114 | (http://www.ethermanage.com/ethernet/ethernet.html) |
---|
115 | |
---|
116 | - *TCP/IP Illustrated, Volume 1 : The Protocols* by |
---|
117 | by W. Richard Stevens (ISBN: 0201633469) |
---|
118 | This book provides detailed introduction to TCP/IP and includes diagnostic |
---|
119 | programs which are publicly available. |
---|
120 | |
---|
121 | - *TCP/IP Illustrated, Volume 2 : The Implementation* by W. Richard |
---|
122 | Stevens and Gary Wright (ISBN: 020163354X) |
---|
123 | This book focuses on implementation issues regarding TCP/IP. The |
---|
124 | treat for RTEMS users is that the implementation covered is the BSD |
---|
125 | stack with most of the source code described in detail. |
---|
126 | |
---|
127 | - *UNIX Network Programming, Volume 1 : 2nd Edition* by W. Richard |
---|
128 | Stevens (ISBN: 0-13-490012-X) |
---|
129 | This book describes how to write basic TCP/IP applications, again with primary |
---|
130 | focus on the BSD stack. |
---|
131 | |
---|
132 | .. COMMENT: Written by Eric Norum |
---|
133 | |
---|
134 | .. COMMENT: COPYRIGHT (c) 1988-2002. |
---|
135 | |
---|
136 | .. COMMENT: On-Line Applications Research Corporation (OAR). |
---|
137 | |
---|
138 | .. COMMENT: All rights reserved. |
---|
139 | |
---|
140 | Network Task Structure and Data Flow |
---|
141 | #################################### |
---|
142 | |
---|
143 | A schematic diagram of the tasks and message *mbuf* queues in a |
---|
144 | simple RTEMS networking application is shown in the following |
---|
145 | figure: |
---|
146 | |
---|
147 | .. image:: images/networkflow.jpg |
---|
148 | |
---|
149 | |
---|
150 | The transmit task for each network interface is normally blocked waiting |
---|
151 | for a packet to arrive in the transmit queue. Once a packet arrives, the |
---|
152 | transmit task may block waiting for an event from the transmit interrupt |
---|
153 | handler. The transmit interrupt handler sends an RTEMS event to the transmit |
---|
154 | task to indicate that transmit hardware resources have become available. |
---|
155 | |
---|
156 | The receive task for each network interface is normally blocked waiting |
---|
157 | for an event from the receive interrupt handler. When this event is received |
---|
158 | the receive task reads the packet and forwards it to the network stack |
---|
159 | for subsequent processing by the network task. |
---|
160 | |
---|
161 | The network task processes incoming packets and takes care of |
---|
162 | timed operations such as handling TCP timeouts and |
---|
163 | aging and removing routing table entries. |
---|
164 | |
---|
165 | The âNetwork codeâ contains routines which may run in the context of |
---|
166 | the user application tasks, the interface receive task or the network task. |
---|
167 | A network semaphore ensures that |
---|
168 | the data structures manipulated by the network code remain consistent. |
---|
169 | |
---|
170 | .. COMMENT: Written by Eric Norum |
---|
171 | |
---|
172 | .. COMMENT: COPYRIGHT (c) 1988-2002. |
---|
173 | |
---|
174 | .. COMMENT: On-Line Applications Research Corporation (OAR). |
---|
175 | |
---|
176 | .. COMMENT: All rights reserved. |
---|
177 | |
---|
178 | Networking Driver |
---|
179 | ################# |
---|
180 | |
---|
181 | Introduction |
---|
182 | ============ |
---|
183 | |
---|
184 | This chapter is intended to provide an introduction to the |
---|
185 | procedure for writing RTEMS network device drivers. |
---|
186 | The example code is taken from the âGeneric 68360â network device |
---|
187 | driver. The source code for this driver is located in the``c/src/lib/libbsp/m68k/gen68360/network`` directory in the RTEMS |
---|
188 | source code distribution. Having a copy of this driver at |
---|
189 | hand when reading the following notes will help significantly. |
---|
190 | |
---|
191 | Learn about the network device |
---|
192 | ============================== |
---|
193 | |
---|
194 | Before starting to write the network driver become completely |
---|
195 | familiar with the programmerâs view of the device. |
---|
196 | The following points list some of the details of the |
---|
197 | device that must be understood before a driver can be written. |
---|
198 | |
---|
199 | - Does the device use DMA to transfer packets to and from |
---|
200 | memory or does the processor have to |
---|
201 | copy packets to and from memory on the device? |
---|
202 | |
---|
203 | - If the device uses DMA, is it capable of forming a single |
---|
204 | outgoing packet from multiple fragments scattered in separate |
---|
205 | memory buffers? |
---|
206 | |
---|
207 | - If the device uses DMA, is it capable of chaining multiple |
---|
208 | outgoing packets, or does each outgoing packet require |
---|
209 | intervention by the driver? |
---|
210 | |
---|
211 | - Does the device automatically pad short frames to the minimum |
---|
212 | 64 bytes or does the driver have to supply the padding? |
---|
213 | |
---|
214 | - Does the device automatically retry a transmission on detection |
---|
215 | of a collision? |
---|
216 | |
---|
217 | - If the device uses DMA, is it capable of buffering multiple |
---|
218 | packets to memory, or does the receiver have to be restarted |
---|
219 | after the arrival of each packet? |
---|
220 | |
---|
221 | - How are packets that are too short, too long, or received with |
---|
222 | CRC errors handled? Does the device automatically continue |
---|
223 | reception or does the driver have to intervene? |
---|
224 | |
---|
225 | - How is the device Ethernet address set? How is the device |
---|
226 | programmed to accept or reject broadcast and multicast packets? |
---|
227 | |
---|
228 | - What interrupts does the device generate? Does it generate an |
---|
229 | interrupt for each incoming packet, or only for packets received |
---|
230 | without error? Does it generate an interrupt for each packet |
---|
231 | transmitted, or only when the transmit queue is empty? What |
---|
232 | happens when a transmit error is detected? |
---|
233 | |
---|
234 | In addition, some controllers have specific questions regarding |
---|
235 | board specific configuration. For example, the SONIC Ethernet |
---|
236 | controller has a very configurable data bus interface. It can |
---|
237 | even be configured for sixteen and thirty-two bit data buses. This |
---|
238 | type of information should be obtained from the board vendor. |
---|
239 | |
---|
240 | Understand the network scheduling conventions |
---|
241 | ============================================= |
---|
242 | |
---|
243 | When writing code for the driver transmit and receive tasks, |
---|
244 | take care to follow the network scheduling conventions. All tasks |
---|
245 | which are associated with networking share various |
---|
246 | data structures and resources. To ensure the consistency |
---|
247 | of these structures the tasks |
---|
248 | execute only when they hold the network semaphore (``rtems_bsdnet_semaphore``). |
---|
249 | The transmit and receive tasks must abide by this protocol. Be very |
---|
250 | careful to avoid âdeadly embracesâ with the other network tasks. |
---|
251 | A number of routines are provided to make it easier for the network |
---|
252 | driver code to conform to the network task scheduling conventions. |
---|
253 | |
---|
254 | - ``void rtems_bsdnet_semaphore_release(void)`` |
---|
255 | This function releases the network semaphore. |
---|
256 | The network driver tasks must call this function immediately before |
---|
257 | making any blocking RTEMS request. |
---|
258 | |
---|
259 | - ``void rtems_bsdnet_semaphore_obtain(void)`` |
---|
260 | This function obtains the network semaphore. |
---|
261 | If a network driver task has released the network semaphore to allow other |
---|
262 | network-related tasks to run while the task blocks, then this function must |
---|
263 | be called to reobtain the semaphore immediately after the return from the |
---|
264 | blocking RTEMS request. |
---|
265 | |
---|
266 | - ``rtems_bsdnet_event_receive(rtems_event_set, rtems_option, rtems_interval, rtems_event_set \*)`` |
---|
267 | The network driver task should call this function when it wishes to wait |
---|
268 | for an event. This function releases the network semaphore, |
---|
269 | calls ``rtems_event_receive`` to wait for the specified event |
---|
270 | or events and reobtains the semaphore. |
---|
271 | The value returned is the value returned by the ``rtems_event_receive``. |
---|
272 | |
---|
273 | Network Driver Makefile |
---|
274 | ======================= |
---|
275 | |
---|
276 | Network drivers are considered part of the BSD network package and as such |
---|
277 | are to be compiled with the appropriate flags. This can be accomplished by |
---|
278 | adding ``-D__INSIDE_RTEMS_BSD_TCPIP_STACK__`` to the ``command line``. |
---|
279 | If the driver is inside the RTEMS source tree or is built using the |
---|
280 | RTEMS application Makefiles, then adding the following line accomplishes |
---|
281 | this: |
---|
282 | .. code:: c |
---|
283 | |
---|
284 | DEFINES += -D__INSIDE_RTEMS_BSD_TCPIP_STACK__ |
---|
285 | |
---|
286 | This is equivalent to the following list of definitions. Early versions |
---|
287 | of the RTEMS BSD network stack required that all of these be defined. |
---|
288 | |
---|
289 | .. code:: c |
---|
290 | |
---|
291 | -D_COMPILING_BSD_KERNEL_ -DKERNEL -DINET -DNFS \\ |
---|
292 | -DDIAGNOSTIC -DBOOTP_COMPAT |
---|
293 | |
---|
294 | Defining these macros tells the network header files that the driver |
---|
295 | is to be compiled with extended visibility into the network stack. This |
---|
296 | is in sharp contrast to applications that simply use the network stack. |
---|
297 | Applications do not require this level of visibility and should stick |
---|
298 | to the portable application level API. |
---|
299 | |
---|
300 | As a direct result of being logically internal to the network stack, |
---|
301 | network drivers use the BSD memory allocation routines This means, |
---|
302 | for example, that malloc takes three arguments. See the SONIC |
---|
303 | device driver (``c/src/lib/libchip/network/sonic.c``) for an example |
---|
304 | of this. Because of this, network drivers should not include``<stdlib.h>``. Doing so will result in conflicting definitions |
---|
305 | of ``malloc()``. |
---|
306 | |
---|
307 | *Application level* code including network servers such as the FTP |
---|
308 | daemon are *not* part of the BSD kernel network code and should not be |
---|
309 | compiled with the BSD network flags. They should include``<stdlib.h>`` and not define the network stack visibility |
---|
310 | macros. |
---|
311 | |
---|
312 | Write the Driver Attach Function |
---|
313 | ================================ |
---|
314 | |
---|
315 | The driver attach function is responsible for configuring the driver |
---|
316 | and making the connection between the network stack |
---|
317 | and the driver. |
---|
318 | |
---|
319 | Driver attach functions take a pointer to an``rtems_bsdnet_ifconfig`` structure as their only argument. |
---|
320 | and set the driver parameters based on the |
---|
321 | values in this structure. If an entry in the configuration |
---|
322 | structure is zero the attach function chooses an |
---|
323 | appropriate default value for that parameter. |
---|
324 | |
---|
325 | The driver should then set up several fields in the ifnet structure |
---|
326 | in the device-dependent data structure supplied and maintained by the driver: |
---|
327 | |
---|
328 | ``ifp->if_softc`` |
---|
329 | Pointer to the device-dependent data. The first entry |
---|
330 | in the device-dependent data structure must be an ``arpcom`` |
---|
331 | structure. |
---|
332 | |
---|
333 | ``ifp->if_name`` |
---|
334 | The name of the device. The network stack uses this string |
---|
335 | and the device number for device name lookups. The device name should |
---|
336 | be obtained from the ``name`` entry in the configuration structure. |
---|
337 | |
---|
338 | ``ifp->if_unit`` |
---|
339 | The device number. The network stack uses this number and the |
---|
340 | device name for device name lookups. For example, if``ifp->if_name`` is â``scc``â and ``ifp->if_unit`` is â``1``â, |
---|
341 | the full device name would be â``scc1``â. The unit number should be |
---|
342 | obtained from the ânameâ entry in the configuration structure. |
---|
343 | |
---|
344 | ``ifp->if_mtu`` |
---|
345 | The maximum transmission unit for the device. For Ethernet |
---|
346 | devices this value should almost always be 1500. |
---|
347 | |
---|
348 | ``ifp->if_flags`` |
---|
349 | The device flags. Ethernet devices should set the flags |
---|
350 | to ``IFF_BROADCAST|IFF_SIMPLEX``, indicating that the |
---|
351 | device can broadcast packets to multiple destinations |
---|
352 | and does not receive and transmit at the same time. |
---|
353 | |
---|
354 | ``ifp->if_snd.ifq_maxlen`` |
---|
355 | The maximum length of the queue of packets waiting to be |
---|
356 | sent to the driver. This is normally set to ``ifqmaxlen``. |
---|
357 | |
---|
358 | ``ifp->if_init`` |
---|
359 | The address of the driver initialization function. |
---|
360 | |
---|
361 | ``ifp->if_start`` |
---|
362 | The address of the driver start function. |
---|
363 | |
---|
364 | ``ifp->if_ioctl`` |
---|
365 | The address of the driver ioctl function. |
---|
366 | |
---|
367 | ``ifp->if_output`` |
---|
368 | The address of the output function. Ethernet devices |
---|
369 | should set this to ``ether_output``. |
---|
370 | |
---|
371 | RTEMS provides a function to parse the driver name in the |
---|
372 | configuration structure into a device name and unit number. |
---|
373 | .. code:: c |
---|
374 | |
---|
375 | int rtems_bsdnet_parse_driver_name ( |
---|
376 | const struct rtems_bsdnet_ifconfig \*config, |
---|
377 | char \**namep |
---|
378 | ); |
---|
379 | |
---|
380 | The function takes two arguments; a pointer to the configuration |
---|
381 | structure and a pointer to a pointer to a character. The function |
---|
382 | parses the configuration name entry, allocates memory for the driver |
---|
383 | name, places the driver name in this memory, sets the second argument |
---|
384 | to point to the name and returns the unit number. |
---|
385 | On error, a message is printed and -1 is returned. |
---|
386 | |
---|
387 | Once the attach function has set up the above entries it must link the |
---|
388 | driver data structure onto the list of devices by |
---|
389 | calling ``if_attach``. Ethernet devices should then |
---|
390 | call ``ether_ifattach``. Both functions take a pointer to the |
---|
391 | deviceâs ``ifnet`` structure as their only argument. |
---|
392 | |
---|
393 | The attach function should return a non-zero value to indicate that |
---|
394 | the driver has been successfully configured and attached. |
---|
395 | |
---|
396 | Write the Driver Start Function. |
---|
397 | ================================ |
---|
398 | |
---|
399 | This function is called each time the network stack wants to start the |
---|
400 | transmitter. This occures whenever the network stack adds a packet |
---|
401 | to a deviceâs send queue and the ``IFF_OACTIVE`` bit in the |
---|
402 | deviceâs ``if_flags`` is not set. |
---|
403 | |
---|
404 | For many devices this function need only set the ``IFF_OACTIVE`` bit in the``if_flags`` and send an event to the transmit task |
---|
405 | indicating that a packet is in the driver transmit queue. |
---|
406 | |
---|
407 | Write the Driver Initialization Function. |
---|
408 | ========================================= |
---|
409 | |
---|
410 | This function should initialize the device, attach to interrupt handler, |
---|
411 | and start the driver transmit and receive tasks. The function |
---|
412 | .. code:: c |
---|
413 | |
---|
414 | rtems_id |
---|
415 | rtems_bsdnet_newproc (char \*name, |
---|
416 | int stacksize, |
---|
417 | void(\*entry)(void \*), |
---|
418 | void \*arg); |
---|
419 | |
---|
420 | should be used to start the driver tasks. |
---|
421 | |
---|
422 | Note that the network stack may call the driver initialization function more |
---|
423 | than once. |
---|
424 | Make sure multiple versions of the receive and transmit tasks are not accidentally |
---|
425 | started. |
---|
426 | |
---|
427 | Write the Driver Transmit Task |
---|
428 | ============================== |
---|
429 | |
---|
430 | This task is reponsible for removing packets from the driver send queue and sending them to the device. The task should block waiting for an event from the |
---|
431 | driver start function indicating that packets are waiting to be transmitted. |
---|
432 | When the transmit task has drained the driver send queue the task should clear |
---|
433 | the ``IFF_OACTIVE`` bit in ``if_flags`` and block until another outgoing |
---|
434 | packet is queued. |
---|
435 | |
---|
436 | Write the Driver Receive Task |
---|
437 | ============================= |
---|
438 | |
---|
439 | This task should block until a packet arrives from the device. If the |
---|
440 | device is an Ethernet interface the function ``ether_input`` should be called |
---|
441 | to forward the packet to the network stack. The arguments to ``ether_input`` |
---|
442 | are a pointer to the interface data structure, a pointer to the ethernet |
---|
443 | header and a pointer to an mbuf containing the packet itself. |
---|
444 | |
---|
445 | Write the Driver Interrupt Handler |
---|
446 | ================================== |
---|
447 | |
---|
448 | A typical interrupt handler will do nothing more than the hardware |
---|
449 | manipulation required to acknowledge the interrupt and send an RTEMS event |
---|
450 | to wake up the driver receive or transmit task waiting for the event. |
---|
451 | Network interface interrupt handlers must not make any calls to other |
---|
452 | network routines. |
---|
453 | |
---|
454 | Write the Driver IOCTL Function |
---|
455 | =============================== |
---|
456 | |
---|
457 | This function handles ioctl requests directed at the device. The ioctl |
---|
458 | commands which must be handled are: |
---|
459 | |
---|
460 | ``SIOCGIFADDR`` |
---|
461 | |
---|
462 | ``SIOCSIFADDR`` |
---|
463 | |
---|
464 | If the device is an Ethernet interface these |
---|
465 | commands should be passed on to ``ether_ioctl``. |
---|
466 | |
---|
467 | ``SIOCSIFFLAGS`` |
---|
468 | |
---|
469 | This command should be used to start or stop the device, |
---|
470 | depending on the state of the interface ``IFF_UP`` and``IFF_RUNNING`` bits in ``if_flags``: |
---|
471 | |
---|
472 | ``IFF_RUNNING`` |
---|
473 | |
---|
474 | Stop the device. |
---|
475 | |
---|
476 | ``IFF_UP`` |
---|
477 | |
---|
478 | Start the device. |
---|
479 | |
---|
480 | ``IFF_UP|IFF_RUNNING`` |
---|
481 | |
---|
482 | Stop then start the device. |
---|
483 | |
---|
484 | ``0`` |
---|
485 | |
---|
486 | Do nothing. |
---|
487 | |
---|
488 | Write the Driver Statistic-Printing Function |
---|
489 | ============================================ |
---|
490 | |
---|
491 | This function should print the values of any statistic/diagnostic |
---|
492 | counters the network driver may use. The driver ioctl function should call |
---|
493 | the statistic-printing function when the ioctl command is``SIO_RTEMS_SHOW_STATS``. |
---|
494 | |
---|
495 | .. COMMENT: Written by Eric Norum |
---|
496 | |
---|
497 | .. COMMENT: COPYRIGHT (c) 1988-2002. |
---|
498 | |
---|
499 | .. COMMENT: On-Line Applications Research Corporation (OAR). |
---|
500 | |
---|
501 | .. COMMENT: All rights reserved. |
---|
502 | |
---|
503 | Using Networking in an RTEMS Application |
---|
504 | ######################################## |
---|
505 | |
---|
506 | Makefile changes |
---|
507 | ================ |
---|
508 | |
---|
509 | Including the required managers |
---|
510 | ------------------------------- |
---|
511 | |
---|
512 | The FreeBSD networking code requires several RTEMS managers |
---|
513 | in the application: |
---|
514 | .. code:: c |
---|
515 | |
---|
516 | MANAGERS = io event semaphore |
---|
517 | |
---|
518 | Increasing the size of the heap |
---|
519 | ------------------------------- |
---|
520 | |
---|
521 | The networking tasks allocate a lot of memory. For most applications |
---|
522 | the heap should be at least 256 kbytes. |
---|
523 | The amount of memory set aside for the heap can be adjusted by setting |
---|
524 | the ``CFLAGS_LD`` definition as shown below: |
---|
525 | .. code:: c |
---|
526 | |
---|
527 | CFLAGS_LD += -Wl,--defsym -Wl,HeapSize=0x80000 |
---|
528 | |
---|
529 | This sets aside 512 kbytes of memory for the heap. |
---|
530 | |
---|
531 | System Configuration |
---|
532 | ==================== |
---|
533 | |
---|
534 | The networking tasks allocate some RTEMS objects. These |
---|
535 | must be accounted for in the application configuration table. The following |
---|
536 | lists the requirements. |
---|
537 | |
---|
538 | *TASKS* |
---|
539 | One network task plus a receive and transmit task for each device. |
---|
540 | |
---|
541 | *SEMAPHORES* |
---|
542 | One network semaphore plus one syslog mutex semaphore if the application uses |
---|
543 | openlog/syslog. |
---|
544 | |
---|
545 | *EVENTS* |
---|
546 | The network stack uses ``RTEMS_EVENT_24`` and ``RTEMS_EVENT_25``. |
---|
547 | This has no effect on the application configuration, but |
---|
548 | application tasks which call the network functions should not |
---|
549 | use these events for other purposes. |
---|
550 | |
---|
551 | Initialization |
---|
552 | ============== |
---|
553 | |
---|
554 | Additional include files |
---|
555 | ------------------------ |
---|
556 | |
---|
557 | The source file which declares the network configuration |
---|
558 | structures and calls the network initialization function must include |
---|
559 | .. code:: c |
---|
560 | |
---|
561 | #include <rtems/rtems_bsdnet.h> |
---|
562 | |
---|
563 | Network Configuration |
---|
564 | --------------------- |
---|
565 | |
---|
566 | The network configuration is specified by declaring |
---|
567 | and initializing the ``rtems_bsdnet_config`` |
---|
568 | structure. |
---|
569 | .. code:: c |
---|
570 | |
---|
571 | struct rtems_bsdnet_config { |
---|
572 | /* |
---|
573 | * This entry points to the head of the ifconfig chain. |
---|
574 | \*/ |
---|
575 | struct rtems_bsdnet_ifconfig \*ifconfig; |
---|
576 | /* |
---|
577 | * This entry should be rtems_bsdnet_do_bootp if BOOTP |
---|
578 | * is being used to configure the network, and NULL |
---|
579 | * if BOOTP is not being used. |
---|
580 | \*/ |
---|
581 | void (\*bootp)(void); |
---|
582 | /* |
---|
583 | * The remaining items can be initialized to 0, in |
---|
584 | * which case the default value will be used. |
---|
585 | \*/ |
---|
586 | rtems_task_priority network_task_priority; /* 100 \*/ |
---|
587 | unsigned long mbuf_bytecount; /* 64 kbytes \*/ |
---|
588 | unsigned long mbuf_cluster_bytecount; /* 128 kbytes \*/ |
---|
589 | char \*hostname; /* BOOTP \*/ |
---|
590 | char \*domainname; /* BOOTP \*/ |
---|
591 | char \*gateway; /* BOOTP \*/ |
---|
592 | char \*log_host; /* BOOTP \*/ |
---|
593 | char \*name_server[3]; /* BOOTP \*/ |
---|
594 | char \*ntp_server[3]; /* BOOTP \*/ |
---|
595 | unsigned long sb_efficiency; /* 2 \*/ |
---|
596 | /* UDP TX: 9216 bytes \*/ |
---|
597 | unsigned long udp_tx_buf_size; |
---|
598 | /* UDP RX: 40 * (1024 + sizeof(struct sockaddr_in)) \*/ |
---|
599 | unsigned long udp_rx_buf_size; |
---|
600 | /* TCP TX: 16 * 1024 bytes \*/ |
---|
601 | unsigned long tcp_tx_buf_size; |
---|
602 | /* TCP TX: 16 * 1024 bytes \*/ |
---|
603 | unsigned long tcp_rx_buf_size; |
---|
604 | /* Default Network Tasks CPU Affinity \*/ |
---|
605 | #ifdef RTEMS_SMP |
---|
606 | const cpu_set_t \*network_task_cpuset; |
---|
607 | size_t network_task_cpuset_size; |
---|
608 | #endif |
---|
609 | }; |
---|
610 | |
---|
611 | The structure entries are described in the following table. |
---|
612 | If your application uses BOOTP/DHCP to obtain network configuration |
---|
613 | information and if you are happy with the default values described |
---|
614 | below, you need to provide only the first two entries in this structure. |
---|
615 | |
---|
616 | ``struct rtems_bsdnet_ifconfig \*ifconfig`` |
---|
617 | |
---|
618 | A pointer to the first configuration structure of the first network |
---|
619 | device. This structure is described in the following section. |
---|
620 | You must provide a value for this entry since there is no default value for it. |
---|
621 | |
---|
622 | ``void (\*bootp)(void)`` |
---|
623 | |
---|
624 | This entry should be set to ``rtems_bsdnet_do_bootp`` if your |
---|
625 | application by default uses the BOOTP/DHCP client protocol to obtain |
---|
626 | network configuration information. It should be set to ``NULL`` if |
---|
627 | your application does not use BOOTP/DHCP. |
---|
628 | You can also use ``rtems_bsdnet_do_bootp_rootfs`` to have a set of |
---|
629 | standard files created with the information return by the BOOTP/DHCP |
---|
630 | protocol. The IP address is added to :file:`/etc/hosts` with the host |
---|
631 | name and domain returned. If no host name or domain is returned``me.mydomain`` is used. The BOOTP/DHCP serverâs address is also |
---|
632 | added to :file:`/etc/hosts`. The domain name server listed in the |
---|
633 | BOOTP/DHCP information are added to :file:`/etc/resolv.conf`. A``search`` record is also added if a domain is returned. The files |
---|
634 | are created if they do not exist. |
---|
635 | The default ``rtems_bsdnet_do_bootp`` and``rtems_bsdnet_do_bootp_rootfs`` handlers will loop for-ever |
---|
636 | waiting for a BOOTP/DHCP server to respond. If an error is detected |
---|
637 | such as not valid interface or valid hardware address the target will |
---|
638 | reboot allowing any hardware reset to correct itself. |
---|
639 | You can provide your own custom handler which allows you to perform |
---|
640 | an initialization that meets your specific system requirements. For |
---|
641 | example you could try BOOTP/DHCP then enter a configuration tool if no |
---|
642 | server is found allowing the user to switch to a static configuration. |
---|
643 | |
---|
644 | ``int network_task_priority`` |
---|
645 | The priority at which the network task and network device |
---|
646 | receive and transmit tasks will run. |
---|
647 | If a value of 0 is specified the tasks will run at priority 100. |
---|
648 | |
---|
649 | ``unsigned long mbuf_bytecount`` |
---|
650 | The number of bytes to allocate from the heap for use as mbufs. |
---|
651 | If a value of 0 is specified, 64 kbytes will be allocated. |
---|
652 | |
---|
653 | ``unsigned long mbuf_cluster_bytecount`` |
---|
654 | The number of bytes to allocate from the heap for use as mbuf clusters. |
---|
655 | If a value of 0 is specified, 128 kbytes will be allocated. |
---|
656 | |
---|
657 | ``char \*hostname`` |
---|
658 | The host name of the system. |
---|
659 | If this, or any of the following, entries are ``NULL`` the value |
---|
660 | may be obtained from a BOOTP/DHCP server. |
---|
661 | |
---|
662 | ``char \*domainname`` |
---|
663 | The name of the Internet domain to which the system belongs. |
---|
664 | |
---|
665 | ``char \*gateway`` |
---|
666 | The Internet host number of the network gateway machine, |
---|
667 | specified in âdotted decimalâ (``129.128.4.1``) form. |
---|
668 | |
---|
669 | ``char \*log_host`` |
---|
670 | The Internet host number of the machine to which ``syslog`` messages |
---|
671 | will be sent. |
---|
672 | |
---|
673 | ``char \*name_server[3]`` |
---|
674 | The Internet host numbers of up to three machines to be used as |
---|
675 | Internet Domain Name Servers. |
---|
676 | |
---|
677 | ``char \*ntp_server[3]`` |
---|
678 | The Internet host numbers of up to three machines to be used as |
---|
679 | Network Time Protocol (NTP) Servers. |
---|
680 | |
---|
681 | ``unsigned long sb_efficiency`` |
---|
682 | This is the first of five configuration parameters related to |
---|
683 | the amount of memory each socket may consume for buffers. The |
---|
684 | TCP/IP stack reserves buffers (e.g. mbufs) for each open socket. The |
---|
685 | TCP/IP stack has different limits for the transmit and receive |
---|
686 | buffers associated with each TCP and UDP socket. By tuning these |
---|
687 | parameters, the application developer can make trade-offs between |
---|
688 | memory consumption and performance. The default parameters favor |
---|
689 | performance over memory consumption. Seehttp://www.rtems.org/ml/rtems-users/2004/february/msg00200.html |
---|
690 | for more details but note that after the RTEMS 4.8 release series, |
---|
691 | the sb_efficiency default was changed from ``8`` to ``2``. |
---|
692 | The user should also be aware of the ``SO_SNDBUF`` and ``SO_RCVBUF`` |
---|
693 | IO control operations. These can be used to specify the |
---|
694 | send and receive buffer sizes for a specific socket. There |
---|
695 | is no standard IO control to change the ``sb_efficiency`` factor. |
---|
696 | The ``sb_efficiency`` parameter is a buffering factor used |
---|
697 | in the implementation of the TCP/IP stack. The default is ``2`` |
---|
698 | which indicates double buffering. When allocating memory for each |
---|
699 | socket, this number is multiplied by the buffer sizes for that socket. |
---|
700 | |
---|
701 | ``unsigned long udp_tx_buf_size`` |
---|
702 | |
---|
703 | This configuration parameter specifies the maximum amount of |
---|
704 | buffer memory which may be used for UDP sockets to transmit |
---|
705 | with. The default size is 9216 bytes which corresponds to |
---|
706 | the maximum datagram size. |
---|
707 | |
---|
708 | ``unsigned long udp_rx_buf_size`` |
---|
709 | |
---|
710 | This configuration parameter specifies the maximum amount of |
---|
711 | buffer memory which may be used for UDP sockets to receive |
---|
712 | into. The default size is the following length in bytes: |
---|
713 | |
---|
714 | .. code:: c |
---|
715 | |
---|
716 | 40 * (1024 + sizeof(struct sockaddr_in) |
---|
717 | |
---|
718 | ``unsigned long tcp_tx_buf_size`` |
---|
719 | |
---|
720 | This configuration parameter specifies the maximum amount of |
---|
721 | buffer memory which may be used for TCP sockets to transmit |
---|
722 | with. The default size is sixteen kilobytes. |
---|
723 | |
---|
724 | ``unsigned long tcp_rx_buf_size`` |
---|
725 | |
---|
726 | This configuration parameter specifies the maximum amount of |
---|
727 | buffer memory which may be used for TCP sockets to receive |
---|
728 | into. The default size is sixteen kilobytes. |
---|
729 | |
---|
730 | ``const cpu_set_t \*network_task_cpuset`` |
---|
731 | |
---|
732 | This configuration parameter specifies the CPU affinity of the |
---|
733 | network task. If set to ``0`` the network task can be scheduled on |
---|
734 | any CPU. Only available in SMP configurations. |
---|
735 | |
---|
736 | ``size_t network_task_cpuset_size`` |
---|
737 | |
---|
738 | This configuration parameter specifies the size of the``network_task_cpuset`` used. Only available in SMP configurations. |
---|
739 | |
---|
740 | In addition, the following fields in the ``rtems_bsdnet_ifconfig`` |
---|
741 | are of interest. |
---|
742 | |
---|
743 | *int port* |
---|
744 | The I/O port number (ex: 0x240) on which the external Ethernet |
---|
745 | can be accessed. |
---|
746 | |
---|
747 | *int irno* |
---|
748 | The interrupt number of the external Ethernet controller. |
---|
749 | |
---|
750 | *int bpar* |
---|
751 | The address of the shared memory on the external Ethernet controller. |
---|
752 | |
---|
753 | Network device configuration |
---|
754 | ---------------------------- |
---|
755 | |
---|
756 | Network devices are specified and configured by declaring and initializing a``struct rtems_bsdnet_ifconfig`` structure for each network device. |
---|
757 | |
---|
758 | The structure entries are described in the following table. An application |
---|
759 | which uses a single network interface, gets network configuration information |
---|
760 | from a BOOTP/DHCP server, and uses the default values for all driver |
---|
761 | parameters needs to initialize only the first two entries in the |
---|
762 | structure. |
---|
763 | |
---|
764 | ``char \*name`` |
---|
765 | The full name of the network device. This name consists of the |
---|
766 | driver name and the unit number (e.g. ``"scc1"``). |
---|
767 | The ``bsp.h`` include file usually defines RTEMS_BSP_NETWORK_DRIVER_NAME as |
---|
768 | the name of the primary (or only) network driver. |
---|
769 | |
---|
770 | ``int (\*attach)(struct rtems_bsdnet_ifconfig \*conf)`` |
---|
771 | The address of the driver ``attach`` function. The network |
---|
772 | initialization function calls this function to configure the driver and |
---|
773 | attach it to the network stack. |
---|
774 | The ``bsp.h`` include file usually defines RTEMS_BSP_NETWORK_DRIVER_ATTACH as |
---|
775 | the name of the attach function of the primary (or only) network driver. |
---|
776 | |
---|
777 | ``struct rtems_bsdnet_ifconfig \*next`` |
---|
778 | A pointer to the network device configuration structure for the next network |
---|
779 | interface, or ``NULL`` if this is the configuration structure of the |
---|
780 | last network interface. |
---|
781 | |
---|
782 | ``char \*ip_address`` |
---|
783 | The Internet address of the device, |
---|
784 | specified in âdotted decimalâ (``129.128.4.2``) form, or ``NULL`` |
---|
785 | if the device configuration information is being obtained from a |
---|
786 | BOOTP/DHCP server. |
---|
787 | |
---|
788 | ``char \*ip_netmask`` |
---|
789 | The Internet inetwork mask of the device, |
---|
790 | specified in âdotted decimalâ (``255.255.255.0``) form, or ``NULL`` |
---|
791 | if the device configuration information is being obtained from a |
---|
792 | BOOTP/DHCP server. |
---|
793 | |
---|
794 | ``void \*hardware_address`` |
---|
795 | The hardware address of the device, or ``NULL`` if the driver is |
---|
796 | to obtain the hardware address in some other way (usually by reading |
---|
797 | it from the device or from the bootstrap ROM). |
---|
798 | |
---|
799 | ``int ignore_broadcast`` |
---|
800 | Zero if the device is to accept broadcast packets, non-zero if the device |
---|
801 | is to ignore broadcast packets. |
---|
802 | |
---|
803 | ``int mtu`` |
---|
804 | The maximum transmission unit of the device, or zero if the driver |
---|
805 | is to choose a default value (typically 1500 for Ethernet devices). |
---|
806 | |
---|
807 | ``int rbuf_count`` |
---|
808 | The number of receive buffers to use, or zero if the driver is to |
---|
809 | choose a default value |
---|
810 | |
---|
811 | ``int xbuf_count`` |
---|
812 | The number of transmit buffers to use, or zero if the driver is to |
---|
813 | choose a default value |
---|
814 | Keep in mind that some network devices may use 4 or more |
---|
815 | transmit descriptors for a single transmit buffer. |
---|
816 | |
---|
817 | A complete network configuration specification can be as simple as the one |
---|
818 | shown in the following example. |
---|
819 | This configuration uses a single network interface, gets |
---|
820 | network configuration information |
---|
821 | from a BOOTP/DHCP server, and uses the default values for all driver |
---|
822 | parameters. |
---|
823 | .. code:: c |
---|
824 | |
---|
825 | static struct rtems_bsdnet_ifconfig netdriver_config = { |
---|
826 | RTEMS_BSP_NETWORK_DRIVER_NAME, |
---|
827 | RTEMS_BSP_NETWORK_DRIVER_ATTACH |
---|
828 | }; |
---|
829 | struct rtems_bsdnet_config rtems_bsdnet_config = { |
---|
830 | &netdriver_config, |
---|
831 | rtems_bsdnet_do_bootp, |
---|
832 | }; |
---|
833 | |
---|
834 | Network initialization |
---|
835 | ---------------------- |
---|
836 | |
---|
837 | The networking tasks must be started before any network I/O operations |
---|
838 | can be performed. This is done by calling: |
---|
839 | |
---|
840 | .. code:: c |
---|
841 | |
---|
842 | rtems_bsdnet_initialize_network (); |
---|
843 | |
---|
844 | This function is declared in ``rtems/rtems_bsdnet.h``. |
---|
845 | t returns 0 on success and -1 on failure with an error code |
---|
846 | in ``errno``. It is not possible to undo the effects of |
---|
847 | a partial initialization, though, so the function can be |
---|
848 | called only once irregardless of the return code. Consequently, |
---|
849 | if the condition for the failure can be corrected, the |
---|
850 | system must be reset to permit another network initialization |
---|
851 | attempt. |
---|
852 | |
---|
853 | Application Programming Interface |
---|
854 | ================================= |
---|
855 | |
---|
856 | The RTEMS network package provides almost a complete set of BSD network |
---|
857 | services. The network functions work like their BSD counterparts |
---|
858 | with the following exceptions: |
---|
859 | |
---|
860 | - A given socket can be read or written by only one task at a time. |
---|
861 | |
---|
862 | - The ``select`` function only works for file descriptors associated |
---|
863 | with sockets. |
---|
864 | |
---|
865 | - You must call ``openlog`` before calling any of the ``syslog`` functions. |
---|
866 | |
---|
867 | - *Some of the network functions are not thread-safe.* |
---|
868 | For example the following functions return a pointer to a static |
---|
869 | buffer which remains valid only until the next call: |
---|
870 | |
---|
871 | ``gethostbyaddr`` |
---|
872 | |
---|
873 | ``gethostbyname`` |
---|
874 | |
---|
875 | ``inet_ntoa`` |
---|
876 | |
---|
877 | (``inet_ntop`` is thread-safe, though). |
---|
878 | |
---|
879 | - The RTEMS network package gathers statistics. |
---|
880 | |
---|
881 | - Addition of a mechanism to "tap onto" an interface |
---|
882 | and monitor every packet received and transmitted. |
---|
883 | |
---|
884 | - Addition of ``SO_SNDWAKEUP`` and ``SO_RCVWAKEUP`` socket options. |
---|
885 | |
---|
886 | Some of the new features are discussed in more detail in the following |
---|
887 | sections. |
---|
888 | |
---|
889 | Network Statistics |
---|
890 | ------------------ |
---|
891 | |
---|
892 | There are a number of functions to print statistics gathered by |
---|
893 | the network stack. |
---|
894 | These function are declared in ``rtems/rtems_bsdnet.h``. |
---|
895 | |
---|
896 | ``rtems_bsdnet_show_if_stats`` |
---|
897 | Display statistics gathered by network interfaces. |
---|
898 | |
---|
899 | ``rtems_bsdnet_show_ip_stats`` |
---|
900 | Display IP packet statistics. |
---|
901 | |
---|
902 | ``rtems_bsdnet_show_icmp_stats`` |
---|
903 | Display ICMP packet statistics. |
---|
904 | |
---|
905 | ``rtems_bsdnet_show_tcp_stats`` |
---|
906 | Display TCP packet statistics. |
---|
907 | |
---|
908 | ``rtems_bsdnet_show_udp_stats`` |
---|
909 | Display UDP packet statistics. |
---|
910 | |
---|
911 | ``rtems_bsdnet_show_mbuf_stats`` |
---|
912 | Display mbuf statistics. |
---|
913 | |
---|
914 | ``rtems_bsdnet_show_inet_routes`` |
---|
915 | Display the routing table. |
---|
916 | |
---|
917 | Tapping Into an Interface |
---|
918 | ------------------------- |
---|
919 | |
---|
920 | RTEMS add two new ioctls to the BSD networking code: |
---|
921 | SIOCSIFTAP and SIOCGIFTAP. These may be used to set and get a*tap function*. The tap function will be called for every |
---|
922 | Ethernet packet received by the interface. |
---|
923 | |
---|
924 | These are called like other interface ioctls, such as SIOCSIFADDR. |
---|
925 | When setting the tap function with SIOCSIFTAP, set the ifr_tap field |
---|
926 | of the ifreq struct to the tap function. When retrieving the tap |
---|
927 | function with SIOCGIFTAP, the current tap function will be returned in |
---|
928 | the ifr_tap field. To stop tapping packets, call SIOCSIFTAP with a |
---|
929 | ifr_tap field of 0. |
---|
930 | |
---|
931 | The tap function is called like this: |
---|
932 | .. code:: c |
---|
933 | |
---|
934 | int tap (struct ifnet \*, struct ether_header \*, struct mbuf \*) |
---|
935 | |
---|
936 | The tap function should return 1 if the packet was fully handled, in |
---|
937 | which case the caller will simply discard the mbuf. The tap function |
---|
938 | should return 0 if the packet should be passed up to the higher |
---|
939 | networking layers. |
---|
940 | |
---|
941 | The tap function is called with the network semaphore locked. It must |
---|
942 | not make any calls on the application levels of the networking level |
---|
943 | itself. It is safe to call other non-networking RTEMS functions. |
---|
944 | |
---|
945 | Socket Options |
---|
946 | -------------- |
---|
947 | |
---|
948 | RTEMS adds two new ``SOL_SOCKET`` level options for ``setsockopt`` and``getsockopt``: ``SO_SNDWAKEUP`` and ``SO_RCVWAKEUP``. For both, the |
---|
949 | option value should point to a sockwakeup structure. The sockwakeup |
---|
950 | structure has the following fields: |
---|
951 | .. code:: c |
---|
952 | |
---|
953 | void (\*sw_pfn) (struct socket \*, caddr_t); |
---|
954 | caddr_t sw_arg; |
---|
955 | |
---|
956 | These options are used to set a callback function to be called when, for |
---|
957 | example, there is |
---|
958 | data available from the socket (``SO_RCVWAKEUP``) and when there is space |
---|
959 | available to accept data written to the socket (``SO_SNDWAKEUP``). |
---|
960 | |
---|
961 | If ``setsockopt`` is called with the ``SO_RCVWAKEUP`` option, and the``sw_pfn`` field is not zero, then when there is data |
---|
962 | available to be read from |
---|
963 | the socket, the function pointed to by the ``sw_pfn`` field will be |
---|
964 | called. A pointer to the socket structure will be passed as the first |
---|
965 | argument to the function. The ``sw_arg`` field set by the``SO_RCVWAKEUP`` call will be passed as the second argument to the function. |
---|
966 | |
---|
967 | If ``setsockopt`` is called with the ``SO_SNDWAKEUP`` |
---|
968 | function, and the ``sw_pfn`` field is not zero, then when |
---|
969 | there is space available to accept data written to the socket, |
---|
970 | the function pointed to by the ``sw_pfn`` field |
---|
971 | will be called. The arguments passed to the function will be as with``SO_SNDWAKEUP``. |
---|
972 | |
---|
973 | When the function is called, the network semaphore will be locked and |
---|
974 | the callback function runs in the context of the networking task. |
---|
975 | The function must be careful not to call any networking functions. It |
---|
976 | is OK to call an RTEMS function; for example, it is OK to send an |
---|
977 | RTEMS event. |
---|
978 | |
---|
979 | The purpose of these callback functions is to permit a more efficient |
---|
980 | alternative to the select call when dealing with a large number of |
---|
981 | sockets. |
---|
982 | |
---|
983 | The callbacks are called by the same criteria that the select |
---|
984 | function uses for indicating "ready" sockets. In Stevens *Unix |
---|
985 | Network Programming* on page 153-154 in the section "Under what Conditions |
---|
986 | Is a Descriptor Ready?" you will find the definitive list of conditions |
---|
987 | for readable and writable that also determine when the functions are |
---|
988 | called. |
---|
989 | |
---|
990 | When the number of received bytes equals or exceeds the socket receive |
---|
991 | buffer "low water mark" (default 1 byte) you get a readable callback. If |
---|
992 | there are 100 bytes in the receive buffer and you only read 1, you will |
---|
993 | not immediately get another callback. However, you will get another |
---|
994 | callback after you read the remaining 99 bytes and at least 1 more byte |
---|
995 | arrives. Using a non-blocking socket you should probably read until it |
---|
996 | produces error EWOULDBLOCK and then allow the readable callback to tell |
---|
997 | you when more data has arrived. (Condition 1.a.) |
---|
998 | |
---|
999 | For sending, when the socket is connected and the free space becomes at |
---|
1000 | or above the "low water mark" for the send buffer (default 4096 bytes) |
---|
1001 | you will receive a writable callback. You donât get continuous callbacks |
---|
1002 | if you donât write anything. Using a non-blocking write socket, you can |
---|
1003 | then call write until it returns a value less than the amount of data |
---|
1004 | requested to be sent or it produces error EWOULDBLOCK (indicating buffer |
---|
1005 | full and no longer writable). When this happens you can |
---|
1006 | try the write again, but it is often better to go do other things and |
---|
1007 | let the writable callback tell you when space is available to send |
---|
1008 | again. You only get a writable callback when the free space transitions |
---|
1009 | to above the "low water mark" and not every time you |
---|
1010 | write to a non-full send buffer. (Condition 2.a.) |
---|
1011 | |
---|
1012 | The remaining conditions enumerated by Stevens handle the fact that |
---|
1013 | sockets become readable and/or writable when connects, disconnects and |
---|
1014 | errors occur, not just when data is received or sent. For example, when |
---|
1015 | a server "listening" socket becomes readable it indicates that a client |
---|
1016 | has connected and accept can be called without blocking, not that |
---|
1017 | network data was received (Condition 1.c). |
---|
1018 | |
---|
1019 | Adding an IP Alias |
---|
1020 | ------------------ |
---|
1021 | |
---|
1022 | The following code snippet adds an IP alias: |
---|
1023 | .. code:: c |
---|
1024 | |
---|
1025 | void addAlias(const char \*pName, const char \*pAddr, const char \*pMask) |
---|
1026 | { |
---|
1027 | struct ifaliasreq aliasreq; |
---|
1028 | struct sockaddr_in \*in; |
---|
1029 | /* initialize alias request \*/ |
---|
1030 | memset(&aliasreq, 0, sizeof(aliasreq)); |
---|
1031 | sprintf(aliasreq.ifra_name, pName); |
---|
1032 | /* initialize alias address \*/ |
---|
1033 | in = (struct sockaddr_in \*)&aliasreq.ifra_addr; |
---|
1034 | in->sin_family = AF_INET; |
---|
1035 | in->sin_len = sizeof(aliasreq.ifra_addr); |
---|
1036 | in->sin_addr.s_addr = inet_addr(pAddr); |
---|
1037 | /* initialize alias mask \*/ |
---|
1038 | in = (struct sockaddr_in \*)&aliasreq.ifra_mask; |
---|
1039 | in->sin_family = AF_INET; |
---|
1040 | in->sin_len = sizeof(aliasreq.ifra_mask); |
---|
1041 | in->sin_addr.s_addr = inet_addr(pMask); |
---|
1042 | /* call to setup the alias \*/ |
---|
1043 | rtems_bsdnet_ifconfig(pName, SIOCAIFADDR, &aliasreq); |
---|
1044 | } |
---|
1045 | |
---|
1046 | Thanks to `Mike Seirs <mailto:mikes@poliac.com>`_ for this example |
---|
1047 | code. |
---|
1048 | |
---|
1049 | Adding a Default Route |
---|
1050 | ---------------------- |
---|
1051 | |
---|
1052 | The function provided in this section is functionally equivalent to |
---|
1053 | the command ``route add default gw yyy.yyy.yyy.yyy``: |
---|
1054 | .. code:: c |
---|
1055 | |
---|
1056 | void mon_ifconfig(int argc, char \*argv[], unsigned32 command_arg, |
---|
1057 | bool verbose) |
---|
1058 | { |
---|
1059 | struct sockaddr_in ipaddr; |
---|
1060 | struct sockaddr_in dstaddr; |
---|
1061 | struct sockaddr_in netmask; |
---|
1062 | struct sockaddr_in broadcast; |
---|
1063 | char \*iface; |
---|
1064 | int f_ip = 0; |
---|
1065 | int f_ptp = 0; |
---|
1066 | int f_netmask = 0; |
---|
1067 | int f_up = 0; |
---|
1068 | int f_down = 0; |
---|
1069 | int f_bcast = 0; |
---|
1070 | int cur_idx; |
---|
1071 | int rc; |
---|
1072 | int flags; |
---|
1073 | bzero((void*) &ipaddr, sizeof(ipaddr)); |
---|
1074 | bzero((void*) &dstaddr, sizeof(dstaddr)); |
---|
1075 | bzero((void*) &netmask, sizeof(netmask)); |
---|
1076 | bzero((void*) &broadcast, sizeof(broadcast)); |
---|
1077 | ipaddr.sin_len = sizeof(ipaddr); |
---|
1078 | ipaddr.sin_family = AF_INET; |
---|
1079 | dstaddr.sin_len = sizeof(dstaddr); |
---|
1080 | dstaddr.sin_family = AF_INET; |
---|
1081 | netmask.sin_len = sizeof(netmask); |
---|
1082 | netmask.sin_family = AF_INET; |
---|
1083 | broadcast.sin_len = sizeof(broadcast); |
---|
1084 | broadcast.sin_family = AF_INET; |
---|
1085 | cur_idx = 0; |
---|
1086 | if (argc <= 1) { |
---|
1087 | /* display all interfaces \*/ |
---|
1088 | iface = NULL; |
---|
1089 | cur_idx += 1; |
---|
1090 | } else { |
---|
1091 | iface = argv[1]; |
---|
1092 | if (isdigit(\*argv[2])) { |
---|
1093 | if (inet_pton(AF_INET, argv[2], &ipaddr.sin_addr) < 0) { |
---|
1094 | printf("bad ip address: %s\\n", argv[2]); |
---|
1095 | return; |
---|
1096 | } |
---|
1097 | f_ip = 1; |
---|
1098 | cur_idx += 3; |
---|
1099 | } else { |
---|
1100 | cur_idx += 2; |
---|
1101 | } |
---|
1102 | } |
---|
1103 | if ((f_down !=0) && (f_ip != 0)) { |
---|
1104 | f_up = 1; |
---|
1105 | } |
---|
1106 | while(argc > cur_idx) { |
---|
1107 | if (strcmp(argv[cur_idx], "up") == 0) { |
---|
1108 | f_up = 1; |
---|
1109 | if (f_down != 0) { |
---|
1110 | printf("Can't make interface up and down\\n"); |
---|
1111 | } |
---|
1112 | } else if(strcmp(argv[cur_idx], "down") == 0) { |
---|
1113 | f_down = 1; |
---|
1114 | if (f_up != 0) { |
---|
1115 | printf("Can't make interface up and down\\n"); |
---|
1116 | } |
---|
1117 | } else if(strcmp(argv[cur_idx], "netmask") == 0) { |
---|
1118 | if ((cur_idx + 1) >= argc) { |
---|
1119 | printf("No netmask address\\n"); |
---|
1120 | return; |
---|
1121 | } |
---|
1122 | if (inet_pton(AF_INET, argv[cur_idx+1], &netmask.sin_addr) < 0) { |
---|
1123 | printf("bad netmask: %s\\n", argv[cur_idx]); |
---|
1124 | return; |
---|
1125 | } |
---|
1126 | f_netmask = 1; |
---|
1127 | cur_idx += 1; |
---|
1128 | } else if(strcmp(argv[cur_idx], "broadcast") == 0) { |
---|
1129 | if ((cur_idx + 1) >= argc) { |
---|
1130 | printf("No broadcast address\\n"); |
---|
1131 | return; |
---|
1132 | } |
---|
1133 | if (inet_pton(AF_INET, argv[cur_idx+1], &broadcast.sin_addr) < 0) { |
---|
1134 | printf("bad broadcast: %s\\n", argv[cur_idx]); |
---|
1135 | return; |
---|
1136 | } |
---|
1137 | f_bcast = 1; |
---|
1138 | cur_idx += 1; |
---|
1139 | } else if(strcmp(argv[cur_idx], "pointopoint") == 0) { |
---|
1140 | if ((cur_idx + 1) >= argc) { |
---|
1141 | printf("No pointopoint address\\n"); |
---|
1142 | return; |
---|
1143 | } |
---|
1144 | if (inet_pton(AF_INET, argv[cur_idx+1], &dstaddr.sin_addr) < 0) { |
---|
1145 | printf("bad pointopoint: %s\\n", argv[cur_idx]); |
---|
1146 | return; |
---|
1147 | } |
---|
1148 | f_ptp = 1; |
---|
1149 | cur_idx += 1; |
---|
1150 | } else { |
---|
1151 | printf("Bad parameter: %s\\n", argv[cur_idx]); |
---|
1152 | return; |
---|
1153 | } |
---|
1154 | cur_idx += 1; |
---|
1155 | } |
---|
1156 | printf("ifconfig "); |
---|
1157 | if (iface != NULL) { |
---|
1158 | printf("%s ", iface); |
---|
1159 | if (f_ip != 0) { |
---|
1160 | char str[256]; |
---|
1161 | inet_ntop(AF_INET, &ipaddr.sin_addr, str, 256); |
---|
1162 | printf("%s ", str); |
---|
1163 | } |
---|
1164 | if (f_netmask != 0) { |
---|
1165 | char str[256]; |
---|
1166 | inet_ntop(AF_INET, &netmask.sin_addr, str, 256); |
---|
1167 | printf("netmask %s ", str); |
---|
1168 | } |
---|
1169 | if (f_bcast != 0) { |
---|
1170 | char str[256]; |
---|
1171 | inet_ntop(AF_INET, &broadcast.sin_addr, str, 256); |
---|
1172 | printf("broadcast %s ", str); |
---|
1173 | } |
---|
1174 | if (f_ptp != 0) { |
---|
1175 | char str[256]; |
---|
1176 | inet_ntop(AF_INET, &dstaddr.sin_addr, str, 256); |
---|
1177 | printf("pointopoint %s ", str); |
---|
1178 | } |
---|
1179 | if (f_up != 0) { |
---|
1180 | printf("up\\n"); |
---|
1181 | } else if (f_down != 0) { |
---|
1182 | printf("down\\n"); |
---|
1183 | } else { |
---|
1184 | printf("\\n"); |
---|
1185 | } |
---|
1186 | } |
---|
1187 | if ((iface == NULL) \|| ((f_ip == 0) && (f_down == 0) && (f_up == 0))) { |
---|
1188 | rtems_bsdnet_show_if_stats(); |
---|
1189 | return; |
---|
1190 | } |
---|
1191 | flags = 0; |
---|
1192 | if (f_netmask) { |
---|
1193 | rc = rtems_bsdnet_ifconfig(iface, SIOCSIFNETMASK, &netmask); |
---|
1194 | if (rc < 0) { |
---|
1195 | printf("Could not set netmask: %s\\n", strerror(errno)); |
---|
1196 | return; |
---|
1197 | } |
---|
1198 | } |
---|
1199 | if (f_bcast) { |
---|
1200 | rc = rtems_bsdnet_ifconfig(iface, SIOCSIFBRDADDR, &broadcast); |
---|
1201 | if (rc < 0) { |
---|
1202 | printf("Could not set broadcast: %s\\n", strerror(errno)); |
---|
1203 | return; |
---|
1204 | } |
---|
1205 | } |
---|
1206 | if (f_ptp) { |
---|
1207 | rc = rtems_bsdnet_ifconfig(iface, SIOCSIFDSTADDR, &dstaddr); |
---|
1208 | if (rc < 0) { |
---|
1209 | printf("Could not set destination address: %s\\n", strerror(errno)); |
---|
1210 | return; |
---|
1211 | } |
---|
1212 | flags \|= IFF_POINTOPOINT; |
---|
1213 | } |
---|
1214 | /* This must come _after_ setting the netmask, broadcast addresses \*/ |
---|
1215 | if (f_ip) { |
---|
1216 | rc = rtems_bsdnet_ifconfig(iface, SIOCSIFADDR, &ipaddr); |
---|
1217 | if (rc < 0) { |
---|
1218 | printf("Could not set IP address: %s\\n", strerror(errno)); |
---|
1219 | return; |
---|
1220 | } |
---|
1221 | } |
---|
1222 | if (f_up != 0) { |
---|
1223 | flags \|= IFF_UP; |
---|
1224 | } |
---|
1225 | if (f_down != 0) { |
---|
1226 | printf("Warning: taking interfaces down is not supported\\n"); |
---|
1227 | } |
---|
1228 | rc = rtems_bsdnet_ifconfig(iface, SIOCSIFFLAGS, &flags); |
---|
1229 | if (rc < 0) { |
---|
1230 | printf("Could not set interface flags: %s\\n", strerror(errno)); |
---|
1231 | return; |
---|
1232 | } |
---|
1233 | } |
---|
1234 | void mon_route(int argc, char \*argv[], unsigned32 command_arg, |
---|
1235 | bool verbose) |
---|
1236 | { |
---|
1237 | int cmd; |
---|
1238 | struct sockaddr_in dst; |
---|
1239 | struct sockaddr_in gw; |
---|
1240 | struct sockaddr_in netmask; |
---|
1241 | int f_host; |
---|
1242 | int f_gw = 0; |
---|
1243 | int cur_idx; |
---|
1244 | int flags; |
---|
1245 | int rc; |
---|
1246 | memset(&dst, 0, sizeof(dst)); |
---|
1247 | memset(&gw, 0, sizeof(gw)); |
---|
1248 | memset(&netmask, 0, sizeof(netmask)); |
---|
1249 | dst.sin_len = sizeof(dst); |
---|
1250 | dst.sin_family = AF_INET; |
---|
1251 | dst.sin_addr.s_addr = inet_addr("0.0.0.0"); |
---|
1252 | gw.sin_len = sizeof(gw); |
---|
1253 | gw.sin_family = AF_INET; |
---|
1254 | gw.sin_addr.s_addr = inet_addr("0.0.0.0"); |
---|
1255 | netmask.sin_len = sizeof(netmask); |
---|
1256 | netmask.sin_family = AF_INET; |
---|
1257 | netmask.sin_addr.s_addr = inet_addr("255.255.255.0"); |
---|
1258 | if (argc < 2) { |
---|
1259 | rtems_bsdnet_show_inet_routes(); |
---|
1260 | return; |
---|
1261 | } |
---|
1262 | if (strcmp(argv[1], "add") == 0) { |
---|
1263 | cmd = RTM_ADD; |
---|
1264 | } else if (strcmp(argv[1], "del") == 0) { |
---|
1265 | cmd = RTM_DELETE; |
---|
1266 | } else { |
---|
1267 | printf("invalid command: %s\\n", argv[1]); |
---|
1268 | printf("\\tit should be 'add' or 'del'\\n"); |
---|
1269 | return; |
---|
1270 | } |
---|
1271 | if (argc < 3) { |
---|
1272 | printf("not enough arguments\\n"); |
---|
1273 | return; |
---|
1274 | } |
---|
1275 | if (strcmp(argv[2], "-host") == 0) { |
---|
1276 | f_host = 1; |
---|
1277 | } else if (strcmp(argv[2], "-net") == 0) { |
---|
1278 | f_host = 0; |
---|
1279 | } else { |
---|
1280 | printf("Invalid type: %s\\n", argv[1]); |
---|
1281 | printf("\\tit should be '-host' or '-net'\\n"); |
---|
1282 | return; |
---|
1283 | } |
---|
1284 | if (argc < 4) { |
---|
1285 | printf("not enough arguments\\n"); |
---|
1286 | return; |
---|
1287 | } |
---|
1288 | inet_pton(AF_INET, argv[3], &dst.sin_addr); |
---|
1289 | cur_idx = 4; |
---|
1290 | while(cur_idx < argc) { |
---|
1291 | if (strcmp(argv[cur_idx], "gw") == 0) { |
---|
1292 | if ((cur_idx +1) >= argc) { |
---|
1293 | printf("no gateway address\\n"); |
---|
1294 | return; |
---|
1295 | } |
---|
1296 | f_gw = 1; |
---|
1297 | inet_pton(AF_INET, argv[cur_idx + 1], &gw.sin_addr); |
---|
1298 | cur_idx += 1; |
---|
1299 | } else if(strcmp(argv[cur_idx], "netmask") == 0) { |
---|
1300 | if ((cur_idx +1) >= argc) { |
---|
1301 | printf("no netmask address\\n"); |
---|
1302 | return; |
---|
1303 | } |
---|
1304 | f_gw = 1; |
---|
1305 | inet_pton(AF_INET, argv[cur_idx + 1], &netmask.sin_addr); |
---|
1306 | cur_idx += 1; |
---|
1307 | } else { |
---|
1308 | printf("Unknown argument\\n"); |
---|
1309 | return; |
---|
1310 | } |
---|
1311 | cur_idx += 1; |
---|
1312 | } |
---|
1313 | flags = RTF_STATIC; |
---|
1314 | if (f_gw != 0) { |
---|
1315 | flags \|= RTF_GATEWAY; |
---|
1316 | } |
---|
1317 | if (f_host != 0) { |
---|
1318 | flags \|= RTF_HOST; |
---|
1319 | } |
---|
1320 | rc = rtems_bsdnet_rtrequest(cmd, &dst, &gw, &netmask, flags, NULL); |
---|
1321 | if (rc < 0) { |
---|
1322 | printf("Error adding route\\n"); |
---|
1323 | } |
---|
1324 | } |
---|
1325 | |
---|
1326 | Thanks to `Jay Monkman <mailto:jtm@smoothmsmoothie.com>`_ for this example |
---|
1327 | code. |
---|
1328 | |
---|
1329 | Time Synchronization Using NTP |
---|
1330 | ------------------------------ |
---|
1331 | |
---|
1332 | .. code:: c |
---|
1333 | |
---|
1334 | int rtems_bsdnet_synchronize_ntp (int interval, rtems_task_priority priority); |
---|
1335 | |
---|
1336 | If the interval argument is 0 the routine synchronizes the RTEMS time-of-day |
---|
1337 | clock with the first NTP server in the rtems_bsdnet_ntpserve array and |
---|
1338 | returns. The priority argument is ignored. |
---|
1339 | |
---|
1340 | If the interval argument is greater than 0, the routine also starts an |
---|
1341 | RTEMS task at the specified priority and polls the NTP server every |
---|
1342 | âintervalâ seconds. NOTE: This mode of operation has not yet been |
---|
1343 | implemented. |
---|
1344 | |
---|
1345 | On successful synchronization of the RTEMS time-of-day clock the routine |
---|
1346 | returns 0. If an error occurs a message is printed and the routine returns -1 |
---|
1347 | with an error code in errno. |
---|
1348 | There is no timeout â if there is no response from an NTP server the |
---|
1349 | routine will wait forever. |
---|
1350 | |
---|
1351 | .. COMMENT: Written by Eric Norum |
---|
1352 | |
---|
1353 | .. COMMENT: COPYRIGHT (c) 1988-2002. |
---|
1354 | |
---|
1355 | .. COMMENT: On-Line Applications Research Corporation (OAR). |
---|
1356 | |
---|
1357 | .. COMMENT: All rights reserved. |
---|
1358 | |
---|
1359 | Testing the Driver |
---|
1360 | ################## |
---|
1361 | |
---|
1362 | Preliminary Setup |
---|
1363 | ================= |
---|
1364 | |
---|
1365 | The network used to test the driver should include at least: |
---|
1366 | |
---|
1367 | - The hardware on which the driver is to run. |
---|
1368 | It makes testing much easier if you can run a debugger to control |
---|
1369 | the operation of the target machine. |
---|
1370 | |
---|
1371 | - An Ethernet network analyzer or a workstation with an |
---|
1372 | âEthernet snoopâ program such as ``ethersnoop`` or``tcpdump``. |
---|
1373 | |
---|
1374 | - A workstation. |
---|
1375 | |
---|
1376 | During early debug, you should consider putting the target, workstation, |
---|
1377 | and snooper on a small network by themselves. This offers a few |
---|
1378 | advantages: |
---|
1379 | |
---|
1380 | - There is less traffic to look at on the snooper and for the target |
---|
1381 | to process while bringing the driver up. |
---|
1382 | |
---|
1383 | - Any serious errors will impact only your small network not a building |
---|
1384 | or campus network. You want to avoid causing any unnecessary problems. |
---|
1385 | |
---|
1386 | - Test traffic is easier to repeatably generate. |
---|
1387 | |
---|
1388 | - Performance measurements are not impacted by other systems on |
---|
1389 | the network. |
---|
1390 | |
---|
1391 | Debug Output |
---|
1392 | ============ |
---|
1393 | |
---|
1394 | There are a number of sources of debug output that can be enabled |
---|
1395 | to aid in tracing the behavior of the network stack. The following |
---|
1396 | is a list of them: |
---|
1397 | |
---|
1398 | - mbuf activity |
---|
1399 | There are commented out calls to ``printf`` in the file``sys/mbuf.h`` in the network stack code. Uncommenting |
---|
1400 | these lines results in output when mbufâs are allocated |
---|
1401 | and freed. This is very useful for finding memory leaks. |
---|
1402 | |
---|
1403 | - TX and RX queuing |
---|
1404 | There are commented out calls to ``printf`` in the file``net/if.h`` in the network stack code. Uncommenting |
---|
1405 | these lines results in output when packets are placed |
---|
1406 | on or removed from one of the transmit or receive packet |
---|
1407 | queues. These queues can be viewed as the boundary line |
---|
1408 | between a device driver and the network stack. If the |
---|
1409 | network stack is enqueuing packets to be transmitted that |
---|
1410 | the device driver is not dequeuing, then that is indicative |
---|
1411 | of a problem in the transmit side of the device driver. |
---|
1412 | Conversely, if the device driver is enqueueing packets |
---|
1413 | as it receives them (via a call to ``ether_input``) and |
---|
1414 | they are not being dequeued by the network stack, |
---|
1415 | then there is a problem. This situation would likely indicate |
---|
1416 | that the network server task is not running. |
---|
1417 | |
---|
1418 | - TCP state transitions |
---|
1419 | In the unlikely event that one would actually want to see |
---|
1420 | TCP state transitions, the ``TCPDEBUG`` macro can be defined |
---|
1421 | in the file ``opt_tcpdebug.h``. This results in the routine``tcp_trace()`` being called by the network stack and |
---|
1422 | the state transitions logged into the ``tcp_debug`` data |
---|
1423 | structure. If the variable ``tcpconsdebug`` in the file``netinet/tcp_debug.c`` is set to 1, then the state transitions |
---|
1424 | will also be printed to the console. |
---|
1425 | |
---|
1426 | Monitor Commands |
---|
1427 | ================ |
---|
1428 | |
---|
1429 | There are a number of command available in the shell / monitor |
---|
1430 | to aid in tracing the behavior of the network stack. The following |
---|
1431 | is a list of them: |
---|
1432 | |
---|
1433 | - ``inet`` |
---|
1434 | This command shows the current routing information for the TCP/IP stack. Following is an |
---|
1435 | example showing the output of this command. |
---|
1436 | |
---|
1437 | .. code:: c |
---|
1438 | |
---|
1439 | Destination Gateway/Mask/Hw Flags Refs Use Expire Interface |
---|
1440 | 10.0.0.0 255.0.0.0 U 0 0 17 smc1 |
---|
1441 | 127.0.0.1 127.0.0.1 UH 0 0 0 lo0 |
---|
1442 | |
---|
1443 | In this example, there is only one network interface with an IP address of 10.8.1.1. This |
---|
1444 | link is currently not up. |
---|
1445 | Two routes that are shown are the default routes for the Ethernet interface (10.0.0.0) and the |
---|
1446 | loopback interface (127.0.0.1). |
---|
1447 | Since the stack comes from BSD, this command is very similar to the netstat command. For more |
---|
1448 | details on the network routing please look the following |
---|
1449 | URL: (http://www.freebsd.org/doc/en_US.ISO8859-1/books/handbook/network-routing.html) |
---|
1450 | For a quick reference to the flags, see the table below: |
---|
1451 | |
---|
1452 | â``U``â |
---|
1453 | Up: The route is active. |
---|
1454 | |
---|
1455 | â``H``â |
---|
1456 | Host: The route destination is a single host. |
---|
1457 | |
---|
1458 | â``G``â |
---|
1459 | Gateway: Send anything for this destination on to this remote system, which |
---|
1460 | will figure out from there where to send it. |
---|
1461 | |
---|
1462 | â``S``â |
---|
1463 | Static: This route was configured manually, not automatically generated by the |
---|
1464 | system. |
---|
1465 | |
---|
1466 | â``C``â |
---|
1467 | Clone: Generates a new route based upon this route for machines we connect |
---|
1468 | to. This type of route is normally used for local networks. |
---|
1469 | |
---|
1470 | â``W``â |
---|
1471 | WasCloned: Indicated a route that was auto-configured based upon a local area |
---|
1472 | network (Clone) route. |
---|
1473 | |
---|
1474 | â``L``â |
---|
1475 | Link: Route involves references to Ethernet hardware. |
---|
1476 | |
---|
1477 | - ``mbuf`` |
---|
1478 | |
---|
1479 | This command shows the current MBUF statistics. An example of the command is shown below: |
---|
1480 | |
---|
1481 | .. code:: c |
---|
1482 | |
---|
1483 | ************ MBUF STATISTICS \************ |
---|
1484 | mbufs:4096 clusters: 256 free: 241 |
---|
1485 | drops: 0 waits: 0 drains: 0 |
---|
1486 | free:4080 data:16 header:0 socket:0 |
---|
1487 | pcb:0 rtable:0 htable:0 atable:0 |
---|
1488 | soname:0 soopts:0 ftable:0 rights:0 |
---|
1489 | ifaddr:0 control:0 oobdata:0 |
---|
1490 | |
---|
1491 | - ``if`` |
---|
1492 | |
---|
1493 | This command shows the current statistics for your Ethernet driver as long as the ioctl hook``SIO_RTEMS_SHOW_STATS`` has been implemented. Below is an example: |
---|
1494 | |
---|
1495 | .. code:: c |
---|
1496 | |
---|
1497 | ************ INTERFACE STATISTICS \************ |
---|
1498 | \***** smc1 \***** |
---|
1499 | Ethernet Address: 00:12:76:43:34:25 |
---|
1500 | Address:10.8.1.1 Broadcast Address:10.255.255.255 Net mask:255.0.0.0 |
---|
1501 | Flags: Up Broadcast Running Simplex |
---|
1502 | Send queue limit:50 length:0 Dropped:0 |
---|
1503 | SMC91C111 RTEMS driver A0.01 11/03/2002 Ian Caddy (ianc@microsol.iinet.net.au) |
---|
1504 | Rx Interrupts:0 Not First:0 Not Last:0 |
---|
1505 | Giant:0 Runt:0 Non-octet:0 |
---|
1506 | Bad CRC:0 Overrun:0 Collision:0 |
---|
1507 | Tx Interrupts:2 Deferred:0 Missed Hearbeat:0 |
---|
1508 | No Carrier:0 Retransmit Limit:0 Late Collision:0 |
---|
1509 | Underrun:0 Raw output wait:0 Coalesced:0 |
---|
1510 | Coalesce failed:0 Retries:0 |
---|
1511 | \***** lo0 \***** |
---|
1512 | Address:127.0.0.1 Net mask:255.0.0.0 |
---|
1513 | Flags: Up Loopback Running Multicast |
---|
1514 | Send queue limit:50 length:0 Dropped:0 |
---|
1515 | |
---|
1516 | - ``ip`` |
---|
1517 | This command show the IP statistics for the currently configured interfaces. |
---|
1518 | |
---|
1519 | - ``icmp`` |
---|
1520 | This command show the ICMP statistics for the currently configured interfaces. |
---|
1521 | |
---|
1522 | - ``tcp`` |
---|
1523 | This command show the TCP statistics for the currently configured interfaces. |
---|
1524 | |
---|
1525 | - ``udp`` |
---|
1526 | This command show the UDP statistics for the currently configured interfaces. |
---|
1527 | |
---|
1528 | Driver basic operation |
---|
1529 | ====================== |
---|
1530 | |
---|
1531 | The network demonstration program ``netdemo`` may be used for these tests. |
---|
1532 | |
---|
1533 | - Edit ``networkconfig.h`` to reflect the values for your network. |
---|
1534 | |
---|
1535 | - Start with ``RTEMS_USE_BOOTP`` not defined. |
---|
1536 | |
---|
1537 | - Edit ``networkconfig.h`` to configure the driver |
---|
1538 | with an |
---|
1539 | explicit Ethernet and Internet address and with reception of |
---|
1540 | broadcast packets disabled: |
---|
1541 | Verify that the program continues to run once the driver has been attached. |
---|
1542 | |
---|
1543 | - Issue a â``u``â command to send UDP |
---|
1544 | packets to the âdiscardâ port. |
---|
1545 | Verify that the packets appear on the network. |
---|
1546 | |
---|
1547 | - Issue a â``s``â command to print the network and driver statistics. |
---|
1548 | |
---|
1549 | - On a workstation, add a static route to the target system. |
---|
1550 | |
---|
1551 | - On that same workstation try to âpingâ the target system. |
---|
1552 | Verify that the ICMP echo request and reply packets appear on the net. |
---|
1553 | |
---|
1554 | - Remove the static route to the target system. |
---|
1555 | Modify ``networkconfig.h`` to attach the driver |
---|
1556 | with reception of broadcast packets enabled. |
---|
1557 | Try to âpingâ the target system again. |
---|
1558 | Verify that ARP request/reply and ICMP echo request/reply packets appear |
---|
1559 | on the net. |
---|
1560 | |
---|
1561 | - Issue a â``t``â command to send TCP |
---|
1562 | packets to the âdiscardâ port. |
---|
1563 | Verify that the packets appear on the network. |
---|
1564 | |
---|
1565 | - Issue a â``s``â command to print the network and driver statistics. |
---|
1566 | |
---|
1567 | - Verify that you can telnet to ports 24742 |
---|
1568 | and 24743 on the target system from one or more |
---|
1569 | workstations on your network. |
---|
1570 | |
---|
1571 | BOOTP/DHCP operation |
---|
1572 | ==================== |
---|
1573 | |
---|
1574 | Set up a BOOTP/DHCP server on the network. |
---|
1575 | Set define ``RTEMS USE_BOOT`` in ``networkconfig.h``. |
---|
1576 | Run the ``netdemo`` test program. |
---|
1577 | Verify that the target system configures itself from the BOOTP/DHCP server and |
---|
1578 | that all the above tests succeed. |
---|
1579 | |
---|
1580 | Stress Tests |
---|
1581 | ============ |
---|
1582 | |
---|
1583 | Once the driver passes the tests described in the previous section it should |
---|
1584 | be subjected to conditions which exercise it more |
---|
1585 | thoroughly and which test its error handling routines. |
---|
1586 | |
---|
1587 | Giant packets |
---|
1588 | ------------- |
---|
1589 | |
---|
1590 | - Recompile the driver with ``MAXIMUM_FRAME_SIZE`` set to |
---|
1591 | a smaller value, say 514. |
---|
1592 | |
---|
1593 | - âPingâ the driver from another workstation and verify |
---|
1594 | that frames larger than 514 bytes are correctly rejected. |
---|
1595 | |
---|
1596 | - Recompile the driver with ``MAXIMUM_FRAME_SIZE`` restored to 1518. |
---|
1597 | |
---|
1598 | Resource Exhaustion |
---|
1599 | ------------------- |
---|
1600 | |
---|
1601 | - Edit ``networkconfig.h`` |
---|
1602 | so that the driver is configured with just two receive and transmit descriptors. |
---|
1603 | |
---|
1604 | - Compile and run the ``netdemo`` program. |
---|
1605 | |
---|
1606 | - Verify that the program operates properly and that you can |
---|
1607 | still telnet to both the ports. |
---|
1608 | |
---|
1609 | - Display the driver statistics (Console â``s``â command or telnet |
---|
1610 | âcontrol-Gâ character) and verify that: |
---|
1611 | |
---|
1612 | # The number of transmit interrupts is non-zero. |
---|
1613 | This indicates that all transmit descriptors have been in use at some time. |
---|
1614 | |
---|
1615 | # The number of missed packets is non-zero. |
---|
1616 | This indicates that all receive descriptors have been in use at some time. |
---|
1617 | |
---|
1618 | Cable Faults |
---|
1619 | ------------ |
---|
1620 | |
---|
1621 | - Run the ``netdemo`` program. |
---|
1622 | |
---|
1623 | - Issue a â``u``â console command to make the target machine transmit |
---|
1624 | a bunch of UDP packets. |
---|
1625 | |
---|
1626 | - While the packets are being transmitted, disconnect and reconnect the |
---|
1627 | network cable. |
---|
1628 | |
---|
1629 | - Display the network statistics and verify that the driver has |
---|
1630 | detected the loss of carrier. |
---|
1631 | |
---|
1632 | - Verify that you can still telnet to both ports on the target machine. |
---|
1633 | |
---|
1634 | Throughput |
---|
1635 | ---------- |
---|
1636 | |
---|
1637 | Run the ``ttcp`` network benchmark program. |
---|
1638 | Transfer large amounts of data (100âs of megabytes) to and from the target |
---|
1639 | system. |
---|
1640 | |
---|
1641 | The procedure for testing throughput from a host to an RTEMS target |
---|
1642 | is as follows: |
---|
1643 | |
---|
1644 | # Download and start the ttcp program on the Target. |
---|
1645 | |
---|
1646 | # In response to the ``ttcp`` prompt, enter ``-s -r``. The |
---|
1647 | meaning of these flags is described in the ``ttcp.1`` manual page |
---|
1648 | found in the ``ttcp_orig`` subdirectory. |
---|
1649 | |
---|
1650 | # On the host run ``ttcp -s -t <<insert the hostname or IP address of the Target here>>`` |
---|
1651 | |
---|
1652 | The procedure for testing throughput from an RTEMS target |
---|
1653 | to a Host is as follows: |
---|
1654 | |
---|
1655 | # On the host run ``ttcp -s -r``. |
---|
1656 | |
---|
1657 | # Download and start the ttcp program on the Target. |
---|
1658 | |
---|
1659 | # In response to the ``ttcp`` prompt, enter ``-s -t <<insert the hostname or IP address of the Target here>>``. You need to type the |
---|
1660 | IP address of the host unless your Target is talking to your Domain Name |
---|
1661 | Server. |
---|
1662 | |
---|
1663 | To change the number of buffers, the buffer size, etc. you just add the |
---|
1664 | extra flags to the ``-t`` machine as specified in the ``ttcp.1`` |
---|
1665 | manual page found in the ``ttcp_orig`` subdirectory. |
---|
1666 | |
---|
1667 | .. COMMENT: Text Written by Jake Janovetz |
---|
1668 | |
---|
1669 | .. COMMENT: COPYRIGHT (c) 1988-2002. |
---|
1670 | |
---|
1671 | .. COMMENT: On-Line Applications Research Corporation (OAR). |
---|
1672 | |
---|
1673 | .. COMMENT: All rights reserved. |
---|
1674 | |
---|
1675 | Network Servers |
---|
1676 | ############### |
---|
1677 | |
---|
1678 | RTEMS FTP Daemon |
---|
1679 | ================ |
---|
1680 | |
---|
1681 | The RTEMS FTPD is a complete file transfer protocol (FTP) daemon |
---|
1682 | which can store, retrieve, and manipulate files on the local |
---|
1683 | filesystem. In addition, the RTEMS FTPD provides âhooksâ |
---|
1684 | which are actions performed on received data. Hooks are useful |
---|
1685 | in situations where a destination file is not necessarily |
---|
1686 | appropriate or in cases when a formal device driver has not yet |
---|
1687 | been implemented. |
---|
1688 | |
---|
1689 | This server was implemented and documented by Jake Janovetz |
---|
1690 | (janovetz@tempest.ece.uiuc.edu). |
---|
1691 | |
---|
1692 | Configuration Parameters |
---|
1693 | ------------------------ |
---|
1694 | |
---|
1695 | The configuration structure for FTPD is as follows: |
---|
1696 | .. code:: c |
---|
1697 | |
---|
1698 | struct rtems_ftpd_configuration |
---|
1699 | { |
---|
1700 | rtems_task_priority priority; /* FTPD task priority \*/ |
---|
1701 | unsigned long max_hook_filesize; /* Maximum buffersize \*/ |
---|
1702 | /* for hooks \*/ |
---|
1703 | int port; /* Well-known port \*/ |
---|
1704 | struct rtems_ftpd_hook \*hooks; /* List of hooks \*/ |
---|
1705 | }; |
---|
1706 | |
---|
1707 | The FTPD task priority is specified with ``priority``. Because |
---|
1708 | hooks are not saved as files, the received data is placed in an |
---|
1709 | allocated buffer. ``max_hook_filesize`` specifies the maximum |
---|
1710 | size of this buffer. Finally, ``hooks`` is a pointer to the |
---|
1711 | configured hooks structure. |
---|
1712 | |
---|
1713 | Initializing FTPD (Starting the daemon) |
---|
1714 | --------------------------------------- |
---|
1715 | |
---|
1716 | Starting FTPD is done with a call to ``rtems_initialize_ftpd()``. |
---|
1717 | The configuration structure must be provided in the application |
---|
1718 | source code. Example hooks structure and configuration structure |
---|
1719 | folllow. |
---|
1720 | .. code:: c |
---|
1721 | |
---|
1722 | struct rtems_ftpd_hook ftp_hooks[] = |
---|
1723 | { |
---|
1724 | {"untar", Untar_FromMemory}, |
---|
1725 | {NULL, NULL} |
---|
1726 | }; |
---|
1727 | struct rtems_ftpd_configuration rtems_ftpd_configuration = |
---|
1728 | { |
---|
1729 | 40, /* FTPD task priority \*/ |
---|
1730 | 512*1024, /* Maximum hook 'file' size \*/ |
---|
1731 | 0, /* Use default port \*/ |
---|
1732 | ftp_hooks /* Local ftp hooks \*/ |
---|
1733 | }; |
---|
1734 | |
---|
1735 | Specifying 0 for the well-known port causes FTPD to use the |
---|
1736 | UNIX standard FTPD port (21). |
---|
1737 | |
---|
1738 | Using Hooks |
---|
1739 | ----------- |
---|
1740 | |
---|
1741 | In the example above, one hook was installed. The hook causes |
---|
1742 | FTPD to call the function ``Untar_FromMemory`` when the |
---|
1743 | user sends data to the file ``untar``. The prototype for |
---|
1744 | the ``untar`` hook (and hooks, in general) is: |
---|
1745 | .. code:: c |
---|
1746 | |
---|
1747 | int Untar_FromMemory(unsigned char \*tar_buf, unsigned long size); |
---|
1748 | |
---|
1749 | An example FTP transcript which exercises this hook is: |
---|
1750 | .. code:: c |
---|
1751 | |
---|
1752 | 220 RTEMS FTP server (Version 1.0-JWJ) ready. |
---|
1753 | Name (dcomm0:janovetz): John Galt |
---|
1754 | 230 User logged in. |
---|
1755 | Remote system type is RTEMS. |
---|
1756 | ftp> bin |
---|
1757 | 200 Type set to I. |
---|
1758 | ftp> dir |
---|
1759 | 200 PORT command successful. |
---|
1760 | 150 ASCII data connection for LIST. |
---|
1761 | drwxrwx--x 0 0 268 dev |
---|
1762 | drwxrwx--x 0 0 0 TFTP |
---|
1763 | 226 Transfer complete. |
---|
1764 | ftp> put html.tar untar |
---|
1765 | local: html.tar remote: untar |
---|
1766 | 200 PORT command successful. |
---|
1767 | 150 BINARY data connection. |
---|
1768 | 210 File transferred successfully. |
---|
1769 | 471040 bytes sent in 0.48 secs (9.6e+02 Kbytes/sec) |
---|
1770 | ftp> dir |
---|
1771 | 200 PORT command successful. |
---|
1772 | 150 ASCII data connection for LIST. |
---|
1773 | drwxrwx--x 0 0 268 dev |
---|
1774 | drwxrwx--x 0 0 0 TFTP |
---|
1775 | drwxrwx--x 0 0 3484 public_html |
---|
1776 | 226 Transfer complete. |
---|
1777 | ftp> quit |
---|
1778 | 221 Goodbye. |
---|
1779 | |
---|
1780 | .. COMMENT: RTEMS Remote Debugger Server Specifications |
---|
1781 | |
---|
1782 | .. COMMENT: Written by: Emmanuel Raguet <raguet@crf.canon.fr> |
---|
1783 | |
---|
1784 | DEC 21140 Driver |
---|
1785 | ################ |
---|
1786 | |
---|
1787 | DEC 21240 Driver Introduction |
---|
1788 | ============================= |
---|
1789 | |
---|
1790 | .. COMMENT: XXX add back in cross reference to list of boards. |
---|
1791 | |
---|
1792 | One aim of our project is to port RTEMS on a standard PowerPC platform. |
---|
1793 | To achieve it, we have chosen a Motorola MCP750 board. This board includes |
---|
1794 | an Ethernet controller based on a DEC21140 chip. Because RTEMS has a |
---|
1795 | TCP/IP stack, we will |
---|
1796 | have to develop the DEC21140 related ethernet driver for the PowerPC port of |
---|
1797 | RTEMS. As this controller is able to support 100Mbps network and as there is |
---|
1798 | a lot of PCI card using this DEC chip, we have decided to first |
---|
1799 | implement this driver on an Intel PC386 target to provide a solution for using |
---|
1800 | RTEMS on PC with the 100Mbps network and then to port this code on PowerPC in |
---|
1801 | a second phase. |
---|
1802 | |
---|
1803 | The aim of this document is to give some PCI board generalities and |
---|
1804 | to explain the software architecture of the RTEMS driver. Finally, we will see |
---|
1805 | what will be done for ChorusOs and Netboot environment . |
---|
1806 | |
---|
1807 | Document Revision History |
---|
1808 | ========================= |
---|
1809 | |
---|
1810 | *Current release*: |
---|
1811 | |
---|
1812 | - Current applicable release is 1.0. |
---|
1813 | |
---|
1814 | *Existing releases*: |
---|
1815 | |
---|
1816 | - 1.0 : Released the 10/02/98. First version of this document. |
---|
1817 | |
---|
1818 | - 0.1 : First draft of this document |
---|
1819 | |
---|
1820 | *Planned releases*: |
---|
1821 | |
---|
1822 | - None planned today. |
---|
1823 | |
---|
1824 | DEC21140 PCI Board Generalities |
---|
1825 | =============================== |
---|
1826 | |
---|
1827 | .. COMMENT: XXX add crossreference to PCI Register Figure |
---|
1828 | |
---|
1829 | This chapter describes rapidely the PCI interface of this Ethernet controller. |
---|
1830 | The board we have chosen for our PC386 implementation is a D-Link DFE-500TX. |
---|
1831 | This is a dual-speed 10/100Mbps Ethernet PCI adapter with a DEC21140AF chip. |
---|
1832 | Like other PCI devices, this board has a PCI deviceâs header containing some |
---|
1833 | required configuration registers, as shown in the PCI Register Figure. |
---|
1834 | By reading |
---|
1835 | or writing these registers, a driver can obtain information about the type of |
---|
1836 | the board, the interrupt it uses, the mapping of the chip specific registers, ... |
---|
1837 | |
---|
1838 | On Intel target, the chip specific registers can be accessed via 2 |
---|
1839 | methods : I/O port access or PCI address mapped access. We have chosen to implement |
---|
1840 | the PCI address access to obtain compatible source code to the port the driver |
---|
1841 | on a PowerPC target. |
---|
1842 | |
---|
1843 | .. COMMENT: PCI Device's Configuration Header Space Format |
---|
1844 | |
---|
1845 | |
---|
1846 | .. image:: images/PCIreg.jpg |
---|
1847 | |
---|
1848 | |
---|
1849 | .. COMMENT: XXX add crossreference to PCI Register Figure |
---|
1850 | |
---|
1851 | On RTEMS, a PCI API exists. We have used it to configure the board. After initializing |
---|
1852 | this PCI module via the ``pci_initialize()`` function, we try to detect |
---|
1853 | the DEC21140 based ethernet board. This board is characterized by its Vendor |
---|
1854 | ID (0x1011) and its Device ID (0x0009). We give these arguments to the``pcib_find_by_deviceid`` |
---|
1855 | function which returns , if the device is present, a pointer to the configuration |
---|
1856 | header space (see PCI Registers Fgure). Once this operation performed, |
---|
1857 | the driver |
---|
1858 | is able to extract the information it needs to configure the board internal |
---|
1859 | registers, like the interrupt line, the base address,... The board internal |
---|
1860 | registers will not be detailled here. You can find them in *DIGITAL |
---|
1861 | Semiconductor 21140A PCI Fast Ethernet LAN Controller |
---|
1862 | - Hardware Reference Manual*. |
---|
1863 | |
---|
1864 | .. COMMENT: fix citation |
---|
1865 | |
---|
1866 | RTEMS Driver Software Architecture |
---|
1867 | ================================== |
---|
1868 | |
---|
1869 | In this chapter will see the initialization phase, how the controller uses the |
---|
1870 | host memory and the 2 threads launched at the initialization time. |
---|
1871 | |
---|
1872 | Initialization phase |
---|
1873 | -------------------- |
---|
1874 | |
---|
1875 | The DEC21140 Ethernet driver keeps the same software architecture than the other |
---|
1876 | RTEMS ethernet drivers. The only API the programmer can use is the ``rtems_dec21140_driver_attach````(struct rtems_bsdnet_ifconfig \*config)`` function which |
---|
1877 | detects the board and initializes the associated data structure (with registers |
---|
1878 | base address, entry points to low-level initialization function,...), if the |
---|
1879 | board is found. |
---|
1880 | |
---|
1881 | Once the attach function executed, the driver initializes the DEC |
---|
1882 | chip. Then the driver connects an interrupt handler to the interrupt line driven |
---|
1883 | by the Ethernet controller (the only interrupt which will be treated is the |
---|
1884 | receive interrupt) and launches 2 threads : a receiver thread and a transmitter |
---|
1885 | thread. Then the driver waits for incoming frame to give to the protocol stack |
---|
1886 | or outcoming frame to send on the physical link. |
---|
1887 | |
---|
1888 | Memory Buffer |
---|
1889 | ------------- |
---|
1890 | |
---|
1891 | .. COMMENT: XXX add cross reference to Problem |
---|
1892 | |
---|
1893 | This DEC chip uses the host memory to store the incoming Ethernet frames and |
---|
1894 | the descriptor of these frames. We have chosen to use 7 receive buffers and |
---|
1895 | 1 transmit buffer to optimize memory allocation due to cache and paging problem |
---|
1896 | that will be explained in the section *Encountered Problems*. |
---|
1897 | |
---|
1898 | To reference these buffers to the DEC chip we use a buffer descriptors |
---|
1899 | ring. The descriptor structure is defined in the Buffer Descriptor Figure. |
---|
1900 | Each descriptor |
---|
1901 | can reference one or two memory buffers. We choose to use only one buffer of |
---|
1902 | 1520 bytes per descriptor. |
---|
1903 | |
---|
1904 | The difference between a receive and a transmit buffer descriptor |
---|
1905 | is located in the status and control bits fields. We do not give details here, |
---|
1906 | please refer to the \[DEC21140 Hardware Manual]. |
---|
1907 | |
---|
1908 | .. COMMENT: Buffer Descriptor |
---|
1909 | |
---|
1910 | |
---|
1911 | .. image:: images/recvbd.jpg |
---|
1912 | |
---|
1913 | |
---|
1914 | Receiver Thread |
---|
1915 | --------------- |
---|
1916 | |
---|
1917 | This thread is event driven. Each time a DEC PCI board interrupt occurs, the |
---|
1918 | handler checks if this is a receive interrupt and send an event âreceptionâ |
---|
1919 | to the receiver thread which looks into the entire buffer descriptors ring the |
---|
1920 | ones that contain a valid incoming frame (bit OWN=0 means descriptor belongs |
---|
1921 | to host processor). Each valid incoming ethernet frame is sent to the protocol |
---|
1922 | stack and the buffer descriptor is given back to the DEC board (the host processor |
---|
1923 | reset bit OWN, which means descriptor belongs to 21140). |
---|
1924 | |
---|
1925 | Transmitter Thread |
---|
1926 | ------------------ |
---|
1927 | |
---|
1928 | This thread is also event driven. Each time an Ethernet frame is put in the |
---|
1929 | transmit queue, an event is sent to the transmit thread, which empty the queue |
---|
1930 | by sending each outcoming frame. Because we use only one transmit buffer, we |
---|
1931 | are sure that the frame is well-sent before sending the next. |
---|
1932 | |
---|
1933 | Encountered Problems |
---|
1934 | ==================== |
---|
1935 | |
---|
1936 | On Intel PC386 target, we were faced with a problem of memory cache management. |
---|
1937 | Because the DEC chip uses the host memory to store the incoming frame and because |
---|
1938 | the DEC21140 configuration registers are mapped into the PCI address space, |
---|
1939 | we must ensure that the data read (or written) by the host processor are the |
---|
1940 | ones written (or read) by the DEC21140 device in the host memory and not old |
---|
1941 | data stored in the cache memory. Therefore, we had to provide a way to manage |
---|
1942 | the cache. This module is described in the document *RTEMS |
---|
1943 | Cache Management For Intel*. On Intel, the |
---|
1944 | memory region cache management is available only if the paging unit is enabled. |
---|
1945 | We have used this paging mechanism, with 4Kb page. All the buffers allocated |
---|
1946 | to store the incoming or outcoming frames, buffer descriptor and also the PCI |
---|
1947 | address space of the DEC board are located in a memory space with cache disable. |
---|
1948 | |
---|
1949 | Concerning the buffers and their descriptors, we have tried to optimize |
---|
1950 | the memory space in term of allocated page. One buffer has 1520 bytes, one descriptor |
---|
1951 | has 16 bytes. We have 7 receive buffers and 1 transmit buffer, and for each, |
---|
1952 | 1 descriptor : (7+1)*(1520+16) = 12288 bytes = 12Kb = 3 entire pages. This |
---|
1953 | allows not to lose too much memory or not to disable cache memory for a page |
---|
1954 | which contains other data than buffer, which could decrease performance. |
---|
1955 | |
---|
1956 | ChorusOs DEC Driver |
---|
1957 | =================== |
---|
1958 | |
---|
1959 | Because ChorusOs is used in several Canon CRF projects, we must provide such |
---|
1960 | a driver on this OS to ensure compatibility between the RTEMS and ChorusOs developments. |
---|
1961 | On ChorusOs, a DEC driver source code already exists but only for a PowerPC |
---|
1962 | target. We plan to port this code (which uses ChorusOs API) on Intel target. |
---|
1963 | This will allow us to have homogeneous developments. Moreover, the port of the |
---|
1964 | development performed with ChorusOs environment to RTEMS environment will be |
---|
1965 | easier for the developers. |
---|
1966 | |
---|
1967 | Netboot DEC driver |
---|
1968 | ================== |
---|
1969 | |
---|
1970 | We use Netboot tool to load our development from a server to the target via |
---|
1971 | an ethernet network. Currently, this tool does not support the DEC board. We |
---|
1972 | plan to port the DEC driver for the Netboot tool. |
---|
1973 | |
---|
1974 | But concerning the port of the DEC driver into Netboot, we are faced |
---|
1975 | with a problem : in RTEMS environment, the DEC driver is interrupt or event |
---|
1976 | driven, in Netboot environment, it must be used in polling mode. It means that |
---|
1977 | we will have to re-write some mechanisms of this driver. |
---|
1978 | |
---|
1979 | List of Ethernet cards using the DEC chip |
---|
1980 | ========================================= |
---|
1981 | |
---|
1982 | Many Ethernet adapter cards use the Tulip chip. Here is a non exhaustive list |
---|
1983 | of adapters which support this driver : |
---|
1984 | |
---|
1985 | - Accton EtherDuo PCI. |
---|
1986 | |
---|
1987 | - Accton EN1207 All three media types supported. |
---|
1988 | |
---|
1989 | - Adaptec ANA6911/TX 21140-AC. |
---|
1990 | |
---|
1991 | - Cogent EM110 21140-A with DP83840 N-Way MII transceiver. |
---|
1992 | |
---|
1993 | - Cogent EM400 EM100 with 4 21140 100mbps-only ports + PCI Bridge. |
---|
1994 | |
---|
1995 | - Danpex EN-9400P3. |
---|
1996 | |
---|
1997 | - D-Link DFE500-Tx 21140-A with DP83840 transceiver. |
---|
1998 | |
---|
1999 | - Kingston EtherX KNE100TX 21140AE. |
---|
2000 | |
---|
2001 | - Netgear FX310 TX 10/100 21140AE. |
---|
2002 | |
---|
2003 | - SMC EtherPower10/100 With DEC21140 and 68836 SYM transceiver. |
---|
2004 | |
---|
2005 | - SMC EtherPower10/100 With DEC21140-AC and DP83840 MII transceiver. |
---|
2006 | Note: The EtherPower II uses the EPIC chip, which requires a different driver. |
---|
2007 | |
---|
2008 | - Surecom EP-320X DEC 21140. |
---|
2009 | |
---|
2010 | - Thomas Conrad TC5048. |
---|
2011 | |
---|
2012 | - Znyx ZX345 21140-A, usually with the DP83840 N-Way MII transciever. Some ZX345 |
---|
2013 | cards made in 1996 have an ICS 1890 transciver instead. |
---|
2014 | |
---|
2015 | - ZNYX ZX348 Two 21140-A chips using ICS 1890 transcievers and either a 21052 |
---|
2016 | or 21152 bridge. Early versions used National 83840 transcievers, but later |
---|
2017 | versions are depopulated ZX346 boards. |
---|
2018 | |
---|
2019 | - ZNYX ZX351 21140 chip with a Broadcom 100BaseT4 transciever. |
---|
2020 | |
---|
2021 | Our DEC driver has not been tested with all these cards, only with the D-Link |
---|
2022 | DFE500-TX. |
---|
2023 | |
---|
2024 | - *[DEC21140 Hardware Manual] DIGITAL, *DIGITAL |
---|
2025 | Semiconductor 21140A PCI Fast Ethernet LAN Controller - Hardware |
---|
2026 | Reference Manual**. |
---|
2027 | |
---|
2028 | - *[99.TA.0021.M.ER]Emmanuel Raguet,*RTEMS Cache Management For Intel**. |
---|
2029 | |
---|
2030 | Command and Variable Index |
---|
2031 | ########################## |
---|
2032 | |
---|
2033 | There are currently no Command and Variable Index entries. |
---|
2034 | |
---|
2035 | .. COMMENT: @printindex fn |
---|
2036 | |
---|
2037 | Concept Index |
---|
2038 | ############# |
---|
2039 | |
---|
2040 | There are currently no Concept Index entries. |
---|
2041 | |
---|
2042 | .. COMMENT: @printindex cp |
---|