1 ==================== 1 ====================
2 DMA Engine API Guide 2 DMA Engine API Guide
3 ==================== 3 ====================
4 4
5 Vinod Koul <vinod dot koul at intel.com> 5 Vinod Koul <vinod dot koul at intel.com>
6 6
7 .. note:: For DMA Engine usage in async_tx ple 7 .. note:: For DMA Engine usage in async_tx please see:
8 ``Documentation/crypto/async-tx-api. 8 ``Documentation/crypto/async-tx-api.rst``
9 9
10 10
11 Below is a guide to device driver writers on h 11 Below is a guide to device driver writers on how to use the Slave-DMA API of the
12 DMA Engine. This is applicable only for slave 12 DMA Engine. This is applicable only for slave DMA usage only.
13 13
14 DMA usage 14 DMA usage
15 ========= 15 =========
16 16
17 The slave DMA usage consists of following step 17 The slave DMA usage consists of following steps:
18 18
19 - Allocate a DMA slave channel 19 - Allocate a DMA slave channel
20 20
21 - Set slave and controller specific parameters 21 - Set slave and controller specific parameters
22 22
23 - Get a descriptor for transaction 23 - Get a descriptor for transaction
24 24
25 - Submit the transaction 25 - Submit the transaction
26 26
27 - Issue pending requests and wait for callback 27 - Issue pending requests and wait for callback notification
28 28
29 The details of these operations are: 29 The details of these operations are:
30 30
31 1. Allocate a DMA slave channel 31 1. Allocate a DMA slave channel
32 32
33 Channel allocation is slightly different in 33 Channel allocation is slightly different in the slave DMA context,
34 client drivers typically need a channel fro 34 client drivers typically need a channel from a particular DMA
35 controller only and even in some cases a sp 35 controller only and even in some cases a specific channel is desired.
36 To request a channel dma_request_chan() API 36 To request a channel dma_request_chan() API is used.
37 37
38 Interface: 38 Interface:
39 39
40 .. code-block:: c 40 .. code-block:: c
41 41
42 struct dma_chan *dma_request_chan(struct 42 struct dma_chan *dma_request_chan(struct device *dev, const char *name);
43 43
44 Which will find and return the ``name`` DMA 44 Which will find and return the ``name`` DMA channel associated with the 'dev'
45 device. The association is done via DT, ACP 45 device. The association is done via DT, ACPI or board file based
46 dma_slave_map matching table. 46 dma_slave_map matching table.
47 47
48 A channel allocated via this interface is e 48 A channel allocated via this interface is exclusive to the caller,
49 until dma_release_channel() is called. 49 until dma_release_channel() is called.
50 50
51 2. Set slave and controller specific parameter 51 2. Set slave and controller specific parameters
52 52
53 Next step is always to pass some specific i 53 Next step is always to pass some specific information to the DMA
54 driver. Most of the generic information whi 54 driver. Most of the generic information which a slave DMA can use
55 is in struct dma_slave_config. This allows 55 is in struct dma_slave_config. This allows the clients to specify
56 DMA direction, DMA addresses, bus widths, D 56 DMA direction, DMA addresses, bus widths, DMA burst lengths etc
57 for the peripheral. 57 for the peripheral.
58 58
59 If some DMA controllers have more parameter 59 If some DMA controllers have more parameters to be sent then they
60 should try to embed struct dma_slave_config 60 should try to embed struct dma_slave_config in their controller
61 specific structure. That gives flexibility 61 specific structure. That gives flexibility to client to pass more
62 parameters, if required. 62 parameters, if required.
63 63
64 Interface: 64 Interface:
65 65
66 .. code-block:: c 66 .. code-block:: c
67 67
68 int dmaengine_slave_config(struct dma_ch 68 int dmaengine_slave_config(struct dma_chan *chan,
69 struct dma_slave_confi 69 struct dma_slave_config *config)
70 70
71 Please see the dma_slave_config structure d 71 Please see the dma_slave_config structure definition in dmaengine.h
72 for a detailed explanation of the struct me 72 for a detailed explanation of the struct members. Please note
73 that the 'direction' member will be going a 73 that the 'direction' member will be going away as it duplicates the
74 direction given in the prepare call. 74 direction given in the prepare call.
75 75
76 3. Get a descriptor for transaction 76 3. Get a descriptor for transaction
77 77
78 For slave usage the various modes of slave t 78 For slave usage the various modes of slave transfers supported by the
79 DMA-engine are: 79 DMA-engine are:
80 80
81 - slave_sg: DMA a list of scatter gather buf 81 - slave_sg: DMA a list of scatter gather buffers from/to a peripheral
82 82
83 - peripheral_dma_vec: DMA an array of scatte <<
84 peripheral. Similar to slave_sg, but uses <<
85 structures instead of a scatterlist. <<
86 <<
87 - dma_cyclic: Perform a cyclic DMA operation 83 - dma_cyclic: Perform a cyclic DMA operation from/to a peripheral till the
88 operation is explicitly stopped. 84 operation is explicitly stopped.
89 85
90 - interleaved_dma: This is common to Slave a 86 - interleaved_dma: This is common to Slave as well as M2M clients. For slave
91 address of devices' fifo could be already 87 address of devices' fifo could be already known to the driver.
92 Various types of operations could be expre 88 Various types of operations could be expressed by setting
93 appropriate values to the 'dma_interleaved 89 appropriate values to the 'dma_interleaved_template' members. Cyclic
94 interleaved DMA transfers are also possibl 90 interleaved DMA transfers are also possible if supported by the channel by
95 setting the DMA_PREP_REPEAT transfer flag. 91 setting the DMA_PREP_REPEAT transfer flag.
96 92
97 A non-NULL return of this transfer API repre 93 A non-NULL return of this transfer API represents a "descriptor" for
98 the given transaction. 94 the given transaction.
99 95
100 Interface: 96 Interface:
101 97
102 .. code-block:: c 98 .. code-block:: c
103 99
104 struct dma_async_tx_descriptor *dmaengine 100 struct dma_async_tx_descriptor *dmaengine_prep_slave_sg(
105 struct dma_chan *chan, struct 101 struct dma_chan *chan, struct scatterlist *sgl,
106 unsigned int sg_len, enum dma_ 102 unsigned int sg_len, enum dma_data_direction direction,
107 unsigned long flags); 103 unsigned long flags);
108 104
109 struct dma_async_tx_descriptor *dmaengine <<
110 struct dma_chan *chan, const s <<
111 size_t nents, enum dma_data_di <<
112 unsigned long flags); <<
113 <<
114 struct dma_async_tx_descriptor *dmaengine 105 struct dma_async_tx_descriptor *dmaengine_prep_dma_cyclic(
115 struct dma_chan *chan, dma_add 106 struct dma_chan *chan, dma_addr_t buf_addr, size_t buf_len,
116 size_t period_len, enum dma_da 107 size_t period_len, enum dma_data_direction direction);
117 108
118 struct dma_async_tx_descriptor *dmaengine 109 struct dma_async_tx_descriptor *dmaengine_prep_interleaved_dma(
119 struct dma_chan *chan, struct 110 struct dma_chan *chan, struct dma_interleaved_template *xt,
120 unsigned long flags); 111 unsigned long flags);
121 112
122 The peripheral driver is expected to have ma 113 The peripheral driver is expected to have mapped the scatterlist for
123 the DMA operation prior to calling dmaengine 114 the DMA operation prior to calling dmaengine_prep_slave_sg(), and must
124 keep the scatterlist mapped until the DMA op 115 keep the scatterlist mapped until the DMA operation has completed.
125 The scatterlist must be mapped using the DMA 116 The scatterlist must be mapped using the DMA struct device.
126 If a mapping needs to be synchronized later, 117 If a mapping needs to be synchronized later, dma_sync_*_for_*() must be
127 called using the DMA struct device, too. 118 called using the DMA struct device, too.
128 So, normal setup should look like this: 119 So, normal setup should look like this:
129 120
130 .. code-block:: c 121 .. code-block:: c
131 122
132 struct device *dma_dev = dmaengine_get_dm !! 123 nr_sg = dma_map_sg(chan->device->dev, sgl, sg_len);
133 <<
134 nr_sg = dma_map_sg(dma_dev, sgl, sg_len); <<
135 if (nr_sg == 0) 124 if (nr_sg == 0)
136 /* error */ 125 /* error */
137 126
138 desc = dmaengine_prep_slave_sg(chan, s 127 desc = dmaengine_prep_slave_sg(chan, sgl, nr_sg, direction, flags);
139 128
140 Once a descriptor has been obtained, the cal 129 Once a descriptor has been obtained, the callback information can be
141 added and the descriptor must then be submit 130 added and the descriptor must then be submitted. Some DMA engine
142 drivers may hold a spinlock between a succes 131 drivers may hold a spinlock between a successful preparation and
143 submission so it is important that these two 132 submission so it is important that these two operations are closely
144 paired. 133 paired.
145 134
146 .. note:: 135 .. note::
147 136
148 Although the async_tx API specifies that 137 Although the async_tx API specifies that completion callback
149 routines cannot submit any new operations 138 routines cannot submit any new operations, this is not the
150 case for slave/cyclic DMA. 139 case for slave/cyclic DMA.
151 140
152 For slave DMA, the subsequent transaction 141 For slave DMA, the subsequent transaction may not be available
153 for submission prior to callback function 142 for submission prior to callback function being invoked, so
154 slave DMA callbacks are permitted to prep 143 slave DMA callbacks are permitted to prepare and submit a new
155 transaction. 144 transaction.
156 145
157 For cyclic DMA, a callback function may w 146 For cyclic DMA, a callback function may wish to terminate the
158 DMA via dmaengine_terminate_async(). 147 DMA via dmaengine_terminate_async().
159 148
160 Therefore, it is important that DMA engin 149 Therefore, it is important that DMA engine drivers drop any
161 locks before calling the callback functio 150 locks before calling the callback function which may cause a
162 deadlock. 151 deadlock.
163 152
164 Note that callbacks will always be invoke 153 Note that callbacks will always be invoked from the DMA
165 engines tasklet, never from interrupt con 154 engines tasklet, never from interrupt context.
166 155
167 **Optional: per descriptor metadata** 156 **Optional: per descriptor metadata**
168 157
169 DMAengine provides two ways for metadata sup 158 DMAengine provides two ways for metadata support.
170 159
171 DESC_METADATA_CLIENT 160 DESC_METADATA_CLIENT
172 161
173 The metadata buffer is allocated/provided 162 The metadata buffer is allocated/provided by the client driver and it is
174 attached to the descriptor. 163 attached to the descriptor.
175 164
176 .. code-block:: c 165 .. code-block:: c
177 166
178 int dmaengine_desc_attach_metadata(struct 167 int dmaengine_desc_attach_metadata(struct dma_async_tx_descriptor *desc,
179 void *data, 168 void *data, size_t len);
180 169
181 DESC_METADATA_ENGINE 170 DESC_METADATA_ENGINE
182 171
183 The metadata buffer is allocated/managed b 172 The metadata buffer is allocated/managed by the DMA driver. The client
184 driver can ask for the pointer, maximum si 173 driver can ask for the pointer, maximum size and the currently used size of
185 the metadata and can directly update or re 174 the metadata and can directly update or read it.
186 175
187 Because the DMA driver manages the memory !! 176 Becasue the DMA driver manages the memory area containing the metadata,
188 clients must make sure that they do not tr 177 clients must make sure that they do not try to access or get the pointer
189 after their transfer completion callback h 178 after their transfer completion callback has run for the descriptor.
190 If no completion callback has been defined 179 If no completion callback has been defined for the transfer, then the
191 metadata must not be accessed after issue_ 180 metadata must not be accessed after issue_pending.
192 In other words: if the aim is to read back 181 In other words: if the aim is to read back metadata after the transfer is
193 completed, then the client must use comple 182 completed, then the client must use completion callback.
194 183
195 .. code-block:: c 184 .. code-block:: c
196 185
197 void *dmaengine_desc_get_metadata_ptr(str 186 void *dmaengine_desc_get_metadata_ptr(struct dma_async_tx_descriptor *desc,
198 size_t *payload_len, size_t *m 187 size_t *payload_len, size_t *max_len);
199 188
200 int dmaengine_desc_set_metadata_len(struc 189 int dmaengine_desc_set_metadata_len(struct dma_async_tx_descriptor *desc,
201 size_t payload_len); 190 size_t payload_len);
202 191
203 Client drivers can query if a given mode is 192 Client drivers can query if a given mode is supported with:
204 193
205 .. code-block:: c 194 .. code-block:: c
206 195
207 bool dmaengine_is_metadata_mode_supported 196 bool dmaengine_is_metadata_mode_supported(struct dma_chan *chan,
208 enum dma_desc_metadata_mode mo 197 enum dma_desc_metadata_mode mode);
209 198
210 Depending on the used mode client drivers mu 199 Depending on the used mode client drivers must follow different flow.
211 200
212 DESC_METADATA_CLIENT 201 DESC_METADATA_CLIENT
213 202
214 - DMA_MEM_TO_DEV / DEV_MEM_TO_MEM: 203 - DMA_MEM_TO_DEV / DEV_MEM_TO_MEM:
215 204
216 1. prepare the descriptor (dmaengine_pre 205 1. prepare the descriptor (dmaengine_prep_*)
217 construct the metadata in the client' 206 construct the metadata in the client's buffer
218 2. use dmaengine_desc_attach_metadata() 207 2. use dmaengine_desc_attach_metadata() to attach the buffer to the
219 descriptor 208 descriptor
220 3. submit the transfer 209 3. submit the transfer
221 210
222 - DMA_DEV_TO_MEM: 211 - DMA_DEV_TO_MEM:
223 212
224 1. prepare the descriptor (dmaengine_pre 213 1. prepare the descriptor (dmaengine_prep_*)
225 2. use dmaengine_desc_attach_metadata() 214 2. use dmaengine_desc_attach_metadata() to attach the buffer to the
226 descriptor 215 descriptor
227 3. submit the transfer 216 3. submit the transfer
228 4. when the transfer is completed, the m 217 4. when the transfer is completed, the metadata should be available in the
229 attached buffer 218 attached buffer
230 219
231 DESC_METADATA_ENGINE 220 DESC_METADATA_ENGINE
232 221
233 - DMA_MEM_TO_DEV / DEV_MEM_TO_MEM: 222 - DMA_MEM_TO_DEV / DEV_MEM_TO_MEM:
234 223
235 1. prepare the descriptor (dmaengine_pre 224 1. prepare the descriptor (dmaengine_prep_*)
236 2. use dmaengine_desc_get_metadata_ptr() 225 2. use dmaengine_desc_get_metadata_ptr() to get the pointer to the
237 engine's metadata area 226 engine's metadata area
238 3. update the metadata at the pointer 227 3. update the metadata at the pointer
239 4. use dmaengine_desc_set_metadata_len() 228 4. use dmaengine_desc_set_metadata_len() to tell the DMA engine the
240 amount of data the client has placed 229 amount of data the client has placed into the metadata buffer
241 5. submit the transfer 230 5. submit the transfer
242 231
243 - DMA_DEV_TO_MEM: 232 - DMA_DEV_TO_MEM:
244 233
245 1. prepare the descriptor (dmaengine_pre 234 1. prepare the descriptor (dmaengine_prep_*)
246 2. submit the transfer 235 2. submit the transfer
247 3. on transfer completion, use dmaengine 236 3. on transfer completion, use dmaengine_desc_get_metadata_ptr() to get
248 the pointer to the engine's metadata 237 the pointer to the engine's metadata area
249 4. read out the metadata from the pointe 238 4. read out the metadata from the pointer
250 239
251 .. note:: 240 .. note::
252 241
253 When DESC_METADATA_ENGINE mode is used th 242 When DESC_METADATA_ENGINE mode is used the metadata area for the descriptor
254 is no longer valid after the transfer has 243 is no longer valid after the transfer has been completed (valid up to the
255 point when the completion callback return 244 point when the completion callback returns if used).
256 245
257 Mixed use of DESC_METADATA_CLIENT / DESC_ 246 Mixed use of DESC_METADATA_CLIENT / DESC_METADATA_ENGINE is not allowed,
258 client drivers must use either of the mod 247 client drivers must use either of the modes per descriptor.
259 248
260 4. Submit the transaction 249 4. Submit the transaction
261 250
262 Once the descriptor has been prepared and t 251 Once the descriptor has been prepared and the callback information
263 added, it must be placed on the DMA engine 252 added, it must be placed on the DMA engine drivers pending queue.
264 253
265 Interface: 254 Interface:
266 255
267 .. code-block:: c 256 .. code-block:: c
268 257
269 dma_cookie_t dmaengine_submit(struct dma 258 dma_cookie_t dmaengine_submit(struct dma_async_tx_descriptor *desc)
270 259
271 This returns a cookie can be used to check 260 This returns a cookie can be used to check the progress of DMA engine
272 activity via other DMA engine calls not cov 261 activity via other DMA engine calls not covered in this document.
273 262
274 dmaengine_submit() will not start the DMA o 263 dmaengine_submit() will not start the DMA operation, it merely adds
275 it to the pending queue. For this, see step 264 it to the pending queue. For this, see step 5, dma_async_issue_pending.
276 265
277 .. note:: 266 .. note::
278 267
279 After calling ``dmaengine_submit()`` the 268 After calling ``dmaengine_submit()`` the submitted transfer descriptor
280 (``struct dma_async_tx_descriptor``) bel 269 (``struct dma_async_tx_descriptor``) belongs to the DMA engine.
281 Consequently, the client must consider i 270 Consequently, the client must consider invalid the pointer to that
282 descriptor. 271 descriptor.
283 272
284 5. Issue pending DMA requests and wait for cal 273 5. Issue pending DMA requests and wait for callback notification
285 274
286 The transactions in the pending queue can b 275 The transactions in the pending queue can be activated by calling the
287 issue_pending API. If channel is idle then 276 issue_pending API. If channel is idle then the first transaction in
288 queue is started and subsequent ones queued 277 queue is started and subsequent ones queued up.
289 278
290 On completion of each DMA operation, the ne 279 On completion of each DMA operation, the next in queue is started and
291 a tasklet triggered. The tasklet will then 280 a tasklet triggered. The tasklet will then call the client driver
292 completion callback routine for notificatio 281 completion callback routine for notification, if set.
293 282
294 Interface: 283 Interface:
295 284
296 .. code-block:: c 285 .. code-block:: c
297 286
298 void dma_async_issue_pending(struct dma_ 287 void dma_async_issue_pending(struct dma_chan *chan);
299 288
300 Further APIs 289 Further APIs
301 ------------ 290 ------------
302 291
303 1. Terminate APIs 292 1. Terminate APIs
304 293
305 .. code-block:: c 294 .. code-block:: c
306 295
307 int dmaengine_terminate_sync(struct dma_ 296 int dmaengine_terminate_sync(struct dma_chan *chan)
308 int dmaengine_terminate_async(struct dma 297 int dmaengine_terminate_async(struct dma_chan *chan)
309 int dmaengine_terminate_all(struct dma_c 298 int dmaengine_terminate_all(struct dma_chan *chan) /* DEPRECATED */
310 299
311 This causes all activity for the DMA channe 300 This causes all activity for the DMA channel to be stopped, and may
312 discard data in the DMA FIFO which hasn't b 301 discard data in the DMA FIFO which hasn't been fully transferred.
313 No callback functions will be called for an 302 No callback functions will be called for any incomplete transfers.
314 303
315 Two variants of this function are available 304 Two variants of this function are available.
316 305
317 dmaengine_terminate_async() might not wait 306 dmaengine_terminate_async() might not wait until the DMA has been fully
318 stopped or until any running complete callb 307 stopped or until any running complete callbacks have finished. But it is
319 possible to call dmaengine_terminate_async( 308 possible to call dmaengine_terminate_async() from atomic context or from
320 within a complete callback. dmaengine_synch 309 within a complete callback. dmaengine_synchronize() must be called before it
321 is safe to free the memory accessed by the 310 is safe to free the memory accessed by the DMA transfer or free resources
322 accessed from within the complete callback. 311 accessed from within the complete callback.
323 312
324 dmaengine_terminate_sync() will wait for th 313 dmaengine_terminate_sync() will wait for the transfer and any running
325 complete callbacks to finish before it retu 314 complete callbacks to finish before it returns. But the function must not be
326 called from atomic context or from within a 315 called from atomic context or from within a complete callback.
327 316
328 dmaengine_terminate_all() is deprecated and 317 dmaengine_terminate_all() is deprecated and should not be used in new code.
329 318
330 2. Pause API 319 2. Pause API
331 320
332 .. code-block:: c 321 .. code-block:: c
333 322
334 int dmaengine_pause(struct dma_chan *cha 323 int dmaengine_pause(struct dma_chan *chan)
335 324
336 This pauses activity on the DMA channel wit 325 This pauses activity on the DMA channel without data loss.
337 326
338 3. Resume API 327 3. Resume API
339 328
340 .. code-block:: c 329 .. code-block:: c
341 330
342 int dmaengine_resume(struct dma_chan *c 331 int dmaengine_resume(struct dma_chan *chan)
343 332
344 Resume a previously paused DMA channel. It 333 Resume a previously paused DMA channel. It is invalid to resume a
345 channel which is not currently paused. 334 channel which is not currently paused.
346 335
347 4. Check Txn complete 336 4. Check Txn complete
348 337
349 .. code-block:: c 338 .. code-block:: c
350 339
351 enum dma_status dma_async_is_tx_complete 340 enum dma_status dma_async_is_tx_complete(struct dma_chan *chan,
352 dma_cookie_t cookie, dma_cooki 341 dma_cookie_t cookie, dma_cookie_t *last, dma_cookie_t *used)
353 342
354 This can be used to check the status of the 343 This can be used to check the status of the channel. Please see
355 the documentation in include/linux/dmaengin 344 the documentation in include/linux/dmaengine.h for a more complete
356 description of this API. 345 description of this API.
357 346
358 This can be used in conjunction with dma_as 347 This can be used in conjunction with dma_async_is_complete() and
359 the cookie returned from dmaengine_submit() 348 the cookie returned from dmaengine_submit() to check for
360 completion of a specific DMA transaction. 349 completion of a specific DMA transaction.
361 350
362 .. note:: 351 .. note::
363 352
364 Not all DMA engine drivers can return re 353 Not all DMA engine drivers can return reliable information for
365 a running DMA channel. It is recommended 354 a running DMA channel. It is recommended that DMA engine users
366 pause or stop (via dmaengine_terminate_a 355 pause or stop (via dmaengine_terminate_all()) the channel before
367 using this API. 356 using this API.
368 357
369 5. Synchronize termination API 358 5. Synchronize termination API
370 359
371 .. code-block:: c 360 .. code-block:: c
372 361
373 void dmaengine_synchronize(struct dma_ch 362 void dmaengine_synchronize(struct dma_chan *chan)
374 363
375 Synchronize the termination of the DMA chan 364 Synchronize the termination of the DMA channel to the current context.
376 365
377 This function should be used after dmaengin 366 This function should be used after dmaengine_terminate_async() to synchronize
378 the termination of the DMA channel to the c 367 the termination of the DMA channel to the current context. The function will
379 wait for the transfer and any running compl 368 wait for the transfer and any running complete callbacks to finish before it
380 returns. 369 returns.
381 370
382 If dmaengine_terminate_async() is used to s 371 If dmaengine_terminate_async() is used to stop the DMA channel this function
383 must be called before it is safe to free me 372 must be called before it is safe to free memory accessed by previously
384 submitted descriptors or to free any resour 373 submitted descriptors or to free any resources accessed within the complete
385 callback of previously submitted descriptor 374 callback of previously submitted descriptors.
386 375
387 The behavior of this function is undefined 376 The behavior of this function is undefined if dma_async_issue_pending() has
388 been called between dmaengine_terminate_asy 377 been called between dmaengine_terminate_async() and this function.
Linux® is a registered trademark of Linus Torvalds in the United States and other countries.
TOMOYO® is a registered trademark of NTT DATA CORPORATION.