blob: e865279cec5855818d83eb281fcce06ee0510040 [file] [log] [blame]
Linus Torvalds1da177e2005-04-16 15:20:36 -07001 Dynamic DMA mapping using the generic device
2 ============================================
3
4 James E.J. Bottomley <James.Bottomley@HansenPartnership.com>
5
6This document describes the DMA API. For a more gentle introduction
FUJITA Tomonorif5a69f42010-03-10 15:23:43 -08007of the API (and actual examples) see
FUJITA Tomonori51e73642010-03-10 15:23:44 -08008Documentation/DMA-API-HOWTO.txt.
Linus Torvalds1da177e2005-04-16 15:20:36 -07009
FUJITA Tomonorif5a69f42010-03-10 15:23:43 -080010This API is split into two pieces. Part I describes the API. Part II
11describes the extensions to the API for supporting non-consistent
12memory machines. Unless you know that your driver absolutely has to
13support non-consistent platforms (this is usually only legacy
14platforms) you should only use the API described in part I.
Linus Torvalds1da177e2005-04-16 15:20:36 -070015
FUJITA Tomonorif5a69f42010-03-10 15:23:43 -080016Part I - dma_ API
Linus Torvalds1da177e2005-04-16 15:20:36 -070017-------------------------------------
18
Linus Torvalds1da177e2005-04-16 15:20:36 -070019To get the dma_ API, you must #include <linux/dma-mapping.h>
20
21
22Part Ia - Using large dma-coherent buffers
23------------------------------------------
24
25void *
26dma_alloc_coherent(struct device *dev, size_t size,
Randy Dunlapa12e2c62007-07-31 00:38:17 -070027 dma_addr_t *dma_handle, gfp_t flag)
Linus Torvalds1da177e2005-04-16 15:20:36 -070028
29Consistent memory is memory for which a write by either the device or
30the processor can immediately be read by the processor or device
David Brownell21440d32006-04-01 10:21:52 -080031without having to worry about caching effects. (You may however need
32to make sure to flush the processor's write buffers before telling
33devices to read that memory.)
Linus Torvalds1da177e2005-04-16 15:20:36 -070034
35This routine allocates a region of <size> bytes of consistent memory.
Randy Dunlapa12e2c62007-07-31 00:38:17 -070036It also returns a <dma_handle> which may be cast to an unsigned
Linus Torvalds1da177e2005-04-16 15:20:36 -070037integer the same width as the bus and used as the physical address
38base of the region.
39
40Returns: a pointer to the allocated region (in the processor's virtual
41address space) or NULL if the allocation failed.
42
43Note: consistent memory can be expensive on some platforms, and the
44minimum allocation length may be as big as a page, so you should
45consolidate your requests for consistent memory as much as possible.
46The simplest way to do that is to use the dma_pool calls (see below).
47
48The flag parameter (dma_alloc_coherent only) allows the caller to
49specify the GFP_ flags (see kmalloc) for the allocation (the
Randy Dunlapa12e2c62007-07-31 00:38:17 -070050implementation may choose to ignore flags that affect the location of
FUJITA Tomonorif5a69f42010-03-10 15:23:43 -080051the returned memory, like GFP_DMA).
Linus Torvalds1da177e2005-04-16 15:20:36 -070052
Andrew Morton842fa692011-11-02 13:39:33 -070053void *
54dma_zalloc_coherent(struct device *dev, size_t size,
55 dma_addr_t *dma_handle, gfp_t flag)
56
57Wraps dma_alloc_coherent() and also zeroes the returned memory if the
58allocation attempt succeeded.
59
Linus Torvalds1da177e2005-04-16 15:20:36 -070060void
Randy Dunlapa12e2c62007-07-31 00:38:17 -070061dma_free_coherent(struct device *dev, size_t size, void *cpu_addr,
Linus Torvalds1da177e2005-04-16 15:20:36 -070062 dma_addr_t dma_handle)
Linus Torvalds1da177e2005-04-16 15:20:36 -070063
64Free the region of consistent memory you previously allocated. dev,
65size and dma_handle must all be the same as those passed into the
66consistent allocate. cpu_addr must be the virtual address returned by
Randy Dunlapa12e2c62007-07-31 00:38:17 -070067the consistent allocate.
Linus Torvalds1da177e2005-04-16 15:20:36 -070068
David Brownellaa248862007-08-10 13:10:27 -070069Note that unlike their sibling allocation calls, these routines
70may only be called with IRQs enabled.
71
Linus Torvalds1da177e2005-04-16 15:20:36 -070072
73Part Ib - Using small dma-coherent buffers
74------------------------------------------
75
76To get this part of the dma_ API, you must #include <linux/dmapool.h>
77
78Many drivers need lots of small dma-coherent memory regions for DMA
79descriptors or I/O buffers. Rather than allocating in units of a page
80or more using dma_alloc_coherent(), you can use DMA pools. These work
Randy Dunlapa12e2c62007-07-31 00:38:17 -070081much like a struct kmem_cache, except that they use the dma-coherent allocator,
Linus Torvalds1da177e2005-04-16 15:20:36 -070082not __get_free_pages(). Also, they understand common hardware constraints
Randy Dunlapa12e2c62007-07-31 00:38:17 -070083for alignment, like queue heads needing to be aligned on N-byte boundaries.
Linus Torvalds1da177e2005-04-16 15:20:36 -070084
85
86 struct dma_pool *
87 dma_pool_create(const char *name, struct device *dev,
88 size_t size, size_t align, size_t alloc);
89
Linus Torvalds1da177e2005-04-16 15:20:36 -070090The pool create() routines initialize a pool of dma-coherent buffers
91for use with a given device. It must be called in a context which
92can sleep.
93
Christoph Lametere18b8902006-12-06 20:33:20 -080094The "name" is for diagnostics (like a struct kmem_cache name); dev and size
Linus Torvalds1da177e2005-04-16 15:20:36 -070095are like what you'd pass to dma_alloc_coherent(). The device's hardware
96alignment requirement for this type of data is "align" (which is expressed
97in bytes, and must be a power of two). If your device has no boundary
98crossing restrictions, pass 0 for alloc; passing 4096 says memory allocated
99from this pool must not cross 4KByte boundaries.
100
101
Randy Dunlapa12e2c62007-07-31 00:38:17 -0700102 void *dma_pool_alloc(struct dma_pool *pool, gfp_t gfp_flags,
Linus Torvalds1da177e2005-04-16 15:20:36 -0700103 dma_addr_t *dma_handle);
104
Linus Torvalds1da177e2005-04-16 15:20:36 -0700105This allocates memory from the pool; the returned memory will meet the size
106and alignment requirements specified at creation time. Pass GFP_ATOMIC to
Randy Dunlapa12e2c62007-07-31 00:38:17 -0700107prevent blocking, or if it's permitted (not in_interrupt, not holding SMP locks),
Linus Torvalds1da177e2005-04-16 15:20:36 -0700108pass GFP_KERNEL to allow blocking. Like dma_alloc_coherent(), this returns
109two values: an address usable by the cpu, and the dma address usable by the
110pool's device.
111
112
113 void dma_pool_free(struct dma_pool *pool, void *vaddr,
114 dma_addr_t addr);
115
Linus Torvalds1da177e2005-04-16 15:20:36 -0700116This puts memory back into the pool. The pool is what was passed to
Randy Dunlapa12e2c62007-07-31 00:38:17 -0700117the pool allocation routine; the cpu (vaddr) and dma addresses are what
Linus Torvalds1da177e2005-04-16 15:20:36 -0700118were returned when that routine allocated the memory being freed.
119
120
121 void dma_pool_destroy(struct dma_pool *pool);
122
Linus Torvalds1da177e2005-04-16 15:20:36 -0700123The pool destroy() routines free the resources of the pool. They must be
124called in a context which can sleep. Make sure you've freed all allocated
125memory back to the pool before you destroy it.
126
127
128Part Ic - DMA addressing limitations
129------------------------------------
130
131int
132dma_supported(struct device *dev, u64 mask)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700133
134Checks to see if the device can support DMA to the memory described by
135mask.
136
137Returns: 1 if it can and 0 if it can't.
138
139Notes: This routine merely tests to see if the mask is possible. It
140won't change the current mask settings. It is more intended as an
141internal API for use by the platform than an external API for use by
142driver writers.
143
144int
Russell King4aa806b2013-06-26 13:49:44 +0100145dma_set_mask_and_coherent(struct device *dev, u64 mask)
146
147Checks to see if the mask is possible and updates the device
148streaming and coherent DMA mask parameters if it is.
149
150Returns: 0 if successful and a negative error if not.
151
152int
Linus Torvalds1da177e2005-04-16 15:20:36 -0700153dma_set_mask(struct device *dev, u64 mask)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700154
155Checks to see if the mask is possible and updates the device
156parameters if it is.
157
158Returns: 0 if successful and a negative error if not.
159
FUJITA Tomonori6a1961f2010-03-10 15:23:39 -0800160int
161dma_set_coherent_mask(struct device *dev, u64 mask)
FUJITA Tomonori6a1961f2010-03-10 15:23:39 -0800162
163Checks to see if the mask is possible and updates the device
164parameters if it is.
165
166Returns: 0 if successful and a negative error if not.
167
Linus Torvalds1da177e2005-04-16 15:20:36 -0700168u64
169dma_get_required_mask(struct device *dev)
170
John Keller175add12008-11-24 16:47:17 -0600171This API returns the mask that the platform requires to
172operate efficiently. Usually this means the returned mask
Linus Torvalds1da177e2005-04-16 15:20:36 -0700173is the minimum required to cover all of memory. Examining the
174required mask gives drivers with variable descriptor sizes the
175opportunity to use smaller descriptors as necessary.
176
177Requesting the required mask does not alter the current mask. If you
John Keller175add12008-11-24 16:47:17 -0600178wish to take advantage of it, you should issue a dma_set_mask()
179call to set the mask to the value returned.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700180
181
182Part Id - Streaming DMA mappings
183--------------------------------
184
185dma_addr_t
186dma_map_single(struct device *dev, void *cpu_addr, size_t size,
187 enum dma_data_direction direction)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700188
189Maps a piece of processor virtual memory so it can be accessed by the
190device and returns the physical handle of the memory.
191
192The direction for both api's may be converted freely by casting.
193However the dma_ API uses a strongly typed enumerator for its
194direction:
195
FUJITA Tomonorif5a69f42010-03-10 15:23:43 -0800196DMA_NONE no direction (used for debugging)
197DMA_TO_DEVICE data is going from the memory to the device
198DMA_FROM_DEVICE data is coming from the device to the memory
199DMA_BIDIRECTIONAL direction isn't known
Linus Torvalds1da177e2005-04-16 15:20:36 -0700200
201Notes: Not all memory regions in a machine can be mapped by this
202API. Further, regions that appear to be physically contiguous in
203kernel virtual space may not be contiguous as physical memory. Since
204this API does not provide any scatter/gather capability, it will fail
Randy Dunlapa12e2c62007-07-31 00:38:17 -0700205if the user tries to map a non-physically contiguous piece of memory.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700206For this reason, it is recommended that memory mapped by this API be
Randy Dunlapa12e2c62007-07-31 00:38:17 -0700207obtained only from sources which guarantee it to be physically contiguous
Linus Torvalds1da177e2005-04-16 15:20:36 -0700208(like kmalloc).
209
210Further, the physical address of the memory must be within the
211dma_mask of the device (the dma_mask represents a bit mask of the
Randy Dunlapa12e2c62007-07-31 00:38:17 -0700212addressable region for the device. I.e., if the physical address of
Linus Torvalds1da177e2005-04-16 15:20:36 -0700213the memory anded with the dma_mask is still equal to the physical
214address, then the device can perform DMA to the memory). In order to
215ensure that the memory allocated by kmalloc is within the dma_mask,
Randy Dunlapa12e2c62007-07-31 00:38:17 -0700216the driver may specify various platform-dependent flags to restrict
Linus Torvalds1da177e2005-04-16 15:20:36 -0700217the physical memory range of the allocation (e.g. on x86, GFP_DMA
218guarantees to be within the first 16Mb of available physical memory,
219as required by ISA devices).
220
221Note also that the above constraints on physical contiguity and
222dma_mask may not apply if the platform has an IOMMU (a device which
223supplies a physical to virtual mapping between the I/O memory bus and
224the device). However, to be portable, device driver writers may *not*
225assume that such an IOMMU exists.
226
227Warnings: Memory coherency operates at a granularity called the cache
228line width. In order for memory mapped by this API to operate
229correctly, the mapped region must begin exactly on a cache line
230boundary and end exactly on one (to prevent two separately mapped
231regions from sharing a single cache line). Since the cache line size
232may not be known at compile time, the API will not enforce this
233requirement. Therefore, it is recommended that driver writers who
234don't take special care to determine the cache line size at run time
235only map virtual regions that begin and end on page boundaries (which
236are guaranteed also to be cache line boundaries).
237
238DMA_TO_DEVICE synchronisation must be done after the last modification
239of the memory region by the software and before it is handed off to
Randy Dunlapa12e2c62007-07-31 00:38:17 -0700240the driver. Once this primitive is used, memory covered by this
241primitive should be treated as read-only by the device. If the device
Linus Torvalds1da177e2005-04-16 15:20:36 -0700242may write to it at any point, it should be DMA_BIDIRECTIONAL (see
243below).
244
245DMA_FROM_DEVICE synchronisation must be done before the driver
246accesses data that may be changed by the device. This memory should
Randy Dunlapa12e2c62007-07-31 00:38:17 -0700247be treated as read-only by the driver. If the driver needs to write
Linus Torvalds1da177e2005-04-16 15:20:36 -0700248to it at any point, it should be DMA_BIDIRECTIONAL (see below).
249
250DMA_BIDIRECTIONAL requires special handling: it means that the driver
251isn't sure if the memory was modified before being handed off to the
252device and also isn't sure if the device will also modify it. Thus,
253you must always sync bidirectional memory twice: once before the
254memory is handed off to the device (to make sure all memory changes
255are flushed from the processor) and once before the data may be
256accessed after being used by the device (to make sure any processor
Randy Dunlapa12e2c62007-07-31 00:38:17 -0700257cache lines are updated with data that the device may have changed).
Linus Torvalds1da177e2005-04-16 15:20:36 -0700258
259void
260dma_unmap_single(struct device *dev, dma_addr_t dma_addr, size_t size,
261 enum dma_data_direction direction)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700262
263Unmaps the region previously mapped. All the parameters passed in
264must be identical to those passed in (and returned) by the mapping
265API.
266
267dma_addr_t
268dma_map_page(struct device *dev, struct page *page,
269 unsigned long offset, size_t size,
270 enum dma_data_direction direction)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700271void
272dma_unmap_page(struct device *dev, dma_addr_t dma_address, size_t size,
273 enum dma_data_direction direction)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700274
275API for mapping and unmapping for pages. All the notes and warnings
276for the other mapping APIs apply here. Also, although the <offset>
277and <size> parameters are provided to do partial page mapping, it is
278recommended that you never use these unless you really know what the
279cache width is.
280
281int
FUJITA Tomonori8d8bb392008-07-25 19:44:49 -0700282dma_mapping_error(struct device *dev, dma_addr_t dma_addr)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700283
Linus Torvalds1da177e2005-04-16 15:20:36 -0700284In some circumstances dma_map_single and dma_map_page will fail to create
285a mapping. A driver can check for these errors by testing the returned
Randy Dunlapa12e2c62007-07-31 00:38:17 -0700286dma address with dma_mapping_error(). A non-zero return value means the mapping
287could not be created and the driver should take appropriate action (e.g.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700288reduce current DMA mapping usage or delay and try again later).
289
David Brownell21440d32006-04-01 10:21:52 -0800290 int
291 dma_map_sg(struct device *dev, struct scatterlist *sg,
292 int nents, enum dma_data_direction direction)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700293
Randy Dunlapa12e2c62007-07-31 00:38:17 -0700294Returns: the number of physical segments mapped (this may be shorter
FUJITA Tomonori1d678f32008-12-01 13:14:01 -0800295than <nents> passed in if some elements of the scatter/gather list are
296physically or virtually adjacent and an IOMMU maps them with a single
297entry).
Linus Torvalds1da177e2005-04-16 15:20:36 -0700298
299Please note that the sg cannot be mapped again if it has been mapped once.
300The mapping process is allowed to destroy information in the sg.
301
302As with the other mapping interfaces, dma_map_sg can fail. When it
303does, 0 is returned and a driver must take appropriate action. It is
304critical that the driver do something, in the case of a block driver
305aborting the request or even oopsing is better than doing nothing and
306corrupting the filesystem.
307
David Brownell21440d32006-04-01 10:21:52 -0800308With scatterlists, you use the resulting mapping like this:
309
310 int i, count = dma_map_sg(dev, sglist, nents, direction);
311 struct scatterlist *sg;
312
FUJITA Tomonori79eb0142008-09-18 09:35:28 -0700313 for_each_sg(sglist, sg, count, i) {
David Brownell21440d32006-04-01 10:21:52 -0800314 hw_address[i] = sg_dma_address(sg);
315 hw_len[i] = sg_dma_len(sg);
316 }
317
318where nents is the number of entries in the sglist.
319
320The implementation is free to merge several consecutive sglist entries
321into one (e.g. with an IOMMU, or if several pages just happen to be
322physically contiguous) and returns the actual number of sg entries it
323mapped them to. On failure 0, is returned.
324
325Then you should loop count times (note: this can be less than nents times)
326and use sg_dma_address() and sg_dma_len() macros where you previously
327accessed sg->address and sg->length as shown above.
328
329 void
330 dma_unmap_sg(struct device *dev, struct scatterlist *sg,
331 int nhwentries, enum dma_data_direction direction)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700332
Randy Dunlapa12e2c62007-07-31 00:38:17 -0700333Unmap the previously mapped scatter/gather list. All the parameters
Linus Torvalds1da177e2005-04-16 15:20:36 -0700334must be the same as those and passed in to the scatter/gather mapping
335API.
336
337Note: <nents> must be the number you passed in, *not* the number of
338physical entries returned.
339
FUJITA Tomonori9705ef72010-03-10 15:23:17 -0800340void
341dma_sync_single_for_cpu(struct device *dev, dma_addr_t dma_handle, size_t size,
342 enum dma_data_direction direction)
343void
FUJITA Tomonori9705ef72010-03-10 15:23:17 -0800344dma_sync_single_for_device(struct device *dev, dma_addr_t dma_handle, size_t size,
345 enum dma_data_direction direction)
346void
FUJITA Tomonori9705ef72010-03-10 15:23:17 -0800347dma_sync_sg_for_cpu(struct device *dev, struct scatterlist *sg, int nelems,
348 enum dma_data_direction direction)
349void
FUJITA Tomonori9705ef72010-03-10 15:23:17 -0800350dma_sync_sg_for_device(struct device *dev, struct scatterlist *sg, int nelems,
351 enum dma_data_direction direction)
FUJITA Tomonori9705ef72010-03-10 15:23:17 -0800352
353Synchronise a single contiguous or scatter/gather mapping for the cpu
354and device. With the sync_sg API, all the parameters must be the same
355as those passed into the single mapping API. With the sync_single API,
356you can use dma_handle and size parameters that aren't identical to
357those passed into the single mapping API to do a partial sync.
358
359Notes: You must do this:
360
361- Before reading values that have been written by DMA from the device
362 (use the DMA_FROM_DEVICE direction)
363- After writing values that will be written to the device using DMA
364 (use the DMA_TO_DEVICE) direction
365- before *and* after handing memory to the device if the memory is
366 DMA_BIDIRECTIONAL
367
368See also dma_map_single().
369
Arthur Kepnera75b0a22008-04-29 01:00:31 -0700370dma_addr_t
371dma_map_single_attrs(struct device *dev, void *cpu_addr, size_t size,
372 enum dma_data_direction dir,
373 struct dma_attrs *attrs)
374
375void
376dma_unmap_single_attrs(struct device *dev, dma_addr_t dma_addr,
377 size_t size, enum dma_data_direction dir,
378 struct dma_attrs *attrs)
379
380int
381dma_map_sg_attrs(struct device *dev, struct scatterlist *sgl,
382 int nents, enum dma_data_direction dir,
383 struct dma_attrs *attrs)
384
385void
386dma_unmap_sg_attrs(struct device *dev, struct scatterlist *sgl,
387 int nents, enum dma_data_direction dir,
388 struct dma_attrs *attrs)
389
390The four functions above are just like the counterpart functions
391without the _attrs suffixes, except that they pass an optional
392struct dma_attrs*.
393
394struct dma_attrs encapsulates a set of "dma attributes". For the
395definition of struct dma_attrs see linux/dma-attrs.h.
396
397The interpretation of dma attributes is architecture-specific, and
398each attribute should be documented in Documentation/DMA-attributes.txt.
399
400If struct dma_attrs* is NULL, the semantics of each of these
401functions is identical to those of the corresponding function
402without the _attrs suffix. As a result dma_map_single_attrs()
403can generally replace dma_map_single(), etc.
404
405As an example of the use of the *_attrs functions, here's how
406you could pass an attribute DMA_ATTR_FOO when mapping memory
407for DMA:
408
409#include <linux/dma-attrs.h>
410/* DMA_ATTR_FOO should be defined in linux/dma-attrs.h and
411 * documented in Documentation/DMA-attributes.txt */
412...
413
414 DEFINE_DMA_ATTRS(attrs);
415 dma_set_attr(DMA_ATTR_FOO, &attrs);
416 ....
417 n = dma_map_sg_attrs(dev, sg, nents, DMA_TO_DEVICE, &attr);
418 ....
419
420Architectures that care about DMA_ATTR_FOO would check for its
421presence in their implementations of the mapping and unmapping
422routines, e.g.:
423
424void whizco_dma_map_sg_attrs(struct device *dev, dma_addr_t dma_addr,
425 size_t size, enum dma_data_direction dir,
426 struct dma_attrs *attrs)
427{
428 ....
429 int foo = dma_get_attr(DMA_ATTR_FOO, attrs);
430 ....
431 if (foo)
432 /* twizzle the frobnozzle */
433 ....
434
Linus Torvalds1da177e2005-04-16 15:20:36 -0700435
436Part II - Advanced dma_ usage
437-----------------------------
438
FUJITA Tomonorif5a69f42010-03-10 15:23:43 -0800439Warning: These pieces of the DMA API should not be used in the
440majority of cases, since they cater for unlikely corner cases that
441don't belong in usual drivers.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700442
443If you don't understand how cache line coherency works between a
444processor and an I/O device, you should not be using this part of the
445API at all.
446
447void *
448dma_alloc_noncoherent(struct device *dev, size_t size,
Randy Dunlapa12e2c62007-07-31 00:38:17 -0700449 dma_addr_t *dma_handle, gfp_t flag)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700450
451Identical to dma_alloc_coherent() except that the platform will
452choose to return either consistent or non-consistent memory as it sees
453fit. By using this API, you are guaranteeing to the platform that you
454have all the correct and necessary sync points for this memory in the
455driver should it choose to return non-consistent memory.
456
457Note: where the platform can return consistent memory, it will
458guarantee that the sync points become nops.
459
460Warning: Handling non-consistent memory is a real pain. You should
461only ever use this API if you positively know your driver will be
462required to work on one of the rare (usually non-PCI) architectures
463that simply cannot make consistent memory.
464
465void
466dma_free_noncoherent(struct device *dev, size_t size, void *cpu_addr,
467 dma_addr_t dma_handle)
468
Randy Dunlapa12e2c62007-07-31 00:38:17 -0700469Free memory allocated by the nonconsistent API. All parameters must
Linus Torvalds1da177e2005-04-16 15:20:36 -0700470be identical to those passed in (and returned by
471dma_alloc_noncoherent()).
472
473int
Linus Torvalds1da177e2005-04-16 15:20:36 -0700474dma_get_cache_alignment(void)
475
Randy Dunlapa12e2c62007-07-31 00:38:17 -0700476Returns the processor cache alignment. This is the absolute minimum
Linus Torvalds1da177e2005-04-16 15:20:36 -0700477alignment *and* width that you must observe when either mapping
478memory or doing partial flushes.
479
480Notes: This API may return a number *larger* than the actual cache
481line, but it will guarantee that one or more cache lines fit exactly
482into the width returned by this call. It will also always be a power
Randy Dunlapa12e2c62007-07-31 00:38:17 -0700483of two for easy alignment.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700484
485void
Ralf Baechled3fa72e2006-12-06 20:38:56 -0800486dma_cache_sync(struct device *dev, void *vaddr, size_t size,
Linus Torvalds1da177e2005-04-16 15:20:36 -0700487 enum dma_data_direction direction)
488
489Do a partial sync of memory that was allocated by
490dma_alloc_noncoherent(), starting at virtual address vaddr and
491continuing on for size. Again, you *must* observe the cache line
492boundaries when doing this.
493
494int
495dma_declare_coherent_memory(struct device *dev, dma_addr_t bus_addr,
496 dma_addr_t device_addr, size_t size, int
497 flags)
498
Linus Torvalds1da177e2005-04-16 15:20:36 -0700499Declare region of memory to be handed out by dma_alloc_coherent when
500it's asked for coherent memory for this device.
501
502bus_addr is the physical address to which the memory is currently
503assigned in the bus responding region (this will be used by the
Randy Dunlapa12e2c62007-07-31 00:38:17 -0700504platform to perform the mapping).
Linus Torvalds1da177e2005-04-16 15:20:36 -0700505
506device_addr is the physical address the device needs to be programmed
507with actually to address this memory (this will be handed out as the
Randy Dunlapa12e2c62007-07-31 00:38:17 -0700508dma_addr_t in dma_alloc_coherent()).
Linus Torvalds1da177e2005-04-16 15:20:36 -0700509
510size is the size of the area (must be multiples of PAGE_SIZE).
511
Randy Dunlapa12e2c62007-07-31 00:38:17 -0700512flags can be or'd together and are:
Linus Torvalds1da177e2005-04-16 15:20:36 -0700513
514DMA_MEMORY_MAP - request that the memory returned from
Matt LaPlante4ae0edc2006-11-30 04:58:40 +0100515dma_alloc_coherent() be directly writable.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700516
517DMA_MEMORY_IO - request that the memory returned from
518dma_alloc_coherent() be addressable using read/write/memcpy_toio etc.
519
Randy Dunlapa12e2c62007-07-31 00:38:17 -0700520One or both of these flags must be present.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700521
522DMA_MEMORY_INCLUDES_CHILDREN - make the declared memory be allocated by
523dma_alloc_coherent of any child devices of this one (for memory residing
524on a bridge).
525
526DMA_MEMORY_EXCLUSIVE - only allocate memory from the declared regions.
527Do not allow dma_alloc_coherent() to fall back to system memory when
528it's out of memory in the declared region.
529
530The return value will be either DMA_MEMORY_MAP or DMA_MEMORY_IO and
531must correspond to a passed in flag (i.e. no returning DMA_MEMORY_IO
532if only DMA_MEMORY_MAP were passed in) for success or zero for
533failure.
534
535Note, for DMA_MEMORY_IO returns, all subsequent memory returned by
536dma_alloc_coherent() may no longer be accessed directly, but instead
537must be accessed using the correct bus functions. If your driver
538isn't prepared to handle this contingency, it should not specify
539DMA_MEMORY_IO in the input flags.
540
541As a simplification for the platforms, only *one* such region of
542memory may be declared per device.
543
544For reasons of efficiency, most platforms choose to track the declared
545region only at the granularity of a page. For smaller allocations,
546you should use the dma_pool() API.
547
548void
549dma_release_declared_memory(struct device *dev)
550
551Remove the memory region previously declared from the system. This
552API performs *no* in-use checking for this region and will return
553unconditionally having removed all the required structures. It is the
Randy Dunlapa12e2c62007-07-31 00:38:17 -0700554driver's job to ensure that no parts of this memory region are
Linus Torvalds1da177e2005-04-16 15:20:36 -0700555currently in use.
556
557void *
558dma_mark_declared_memory_occupied(struct device *dev,
559 dma_addr_t device_addr, size_t size)
560
561This is used to occupy specific regions of the declared space
562(dma_alloc_coherent() will hand out the first free region it finds).
563
Randy Dunlapa12e2c62007-07-31 00:38:17 -0700564device_addr is the *device* address of the region requested.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700565
Randy Dunlapa12e2c62007-07-31 00:38:17 -0700566size is the size (and should be a page-sized multiple).
Linus Torvalds1da177e2005-04-16 15:20:36 -0700567
568The return value will be either a pointer to the processor virtual
569address of the memory, or an error (via PTR_ERR()) if any part of the
570region is occupied.
Joerg Roedel187f9c32009-01-09 16:28:07 +0100571
572Part III - Debug drivers use of the DMA-API
573-------------------------------------------
574
575The DMA-API as described above as some constraints. DMA addresses must be
576released with the corresponding function with the same size for example. With
577the advent of hardware IOMMUs it becomes more and more important that drivers
578do not violate those constraints. In the worst case such a violation can
579result in data corruption up to destroyed filesystems.
580
581To debug drivers and find bugs in the usage of the DMA-API checking code can
582be compiled into the kernel which will tell the developer about those
583violations. If your architecture supports it you can select the "Enable
584debugging of DMA-API usage" option in your kernel configuration. Enabling this
585option has a performance impact. Do not enable it in production kernels.
586
587If you boot the resulting kernel will contain code which does some bookkeeping
588about what DMA memory was allocated for which device. If this code detects an
589error it prints a warning message with some details into your kernel log. An
590example warning message may look like this:
591
592------------[ cut here ]------------
593WARNING: at /data2/repos/linux-2.6-iommu/lib/dma-debug.c:448
594 check_unmap+0x203/0x490()
595Hardware name:
596forcedeth 0000:00:08.0: DMA-API: device driver frees DMA memory with wrong
597 function [device address=0x00000000640444be] [size=66 bytes] [mapped as
598single] [unmapped as page]
599Modules linked in: nfsd exportfs bridge stp llc r8169
600Pid: 0, comm: swapper Tainted: G W 2.6.28-dmatest-09289-g8bb99c0 #1
601Call Trace:
602 <IRQ> [<ffffffff80240b22>] warn_slowpath+0xf2/0x130
603 [<ffffffff80647b70>] _spin_unlock+0x10/0x30
604 [<ffffffff80537e75>] usb_hcd_link_urb_to_ep+0x75/0xc0
605 [<ffffffff80647c22>] _spin_unlock_irqrestore+0x12/0x40
606 [<ffffffff8055347f>] ohci_urb_enqueue+0x19f/0x7c0
607 [<ffffffff80252f96>] queue_work+0x56/0x60
608 [<ffffffff80237e10>] enqueue_task_fair+0x20/0x50
609 [<ffffffff80539279>] usb_hcd_submit_urb+0x379/0xbc0
610 [<ffffffff803b78c3>] cpumask_next_and+0x23/0x40
611 [<ffffffff80235177>] find_busiest_group+0x207/0x8a0
612 [<ffffffff8064784f>] _spin_lock_irqsave+0x1f/0x50
613 [<ffffffff803c7ea3>] check_unmap+0x203/0x490
614 [<ffffffff803c8259>] debug_dma_unmap_page+0x49/0x50
615 [<ffffffff80485f26>] nv_tx_done_optimized+0xc6/0x2c0
616 [<ffffffff80486c13>] nv_nic_irq_optimized+0x73/0x2b0
617 [<ffffffff8026df84>] handle_IRQ_event+0x34/0x70
618 [<ffffffff8026ffe9>] handle_edge_irq+0xc9/0x150
619 [<ffffffff8020e3ab>] do_IRQ+0xcb/0x1c0
620 [<ffffffff8020c093>] ret_from_intr+0x0/0xa
621 <EOI> <4>---[ end trace f6435a98e2a38c0e ]---
622
623The driver developer can find the driver and the device including a stacktrace
624of the DMA-API call which caused this warning.
625
626Per default only the first error will result in a warning message. All other
627errors will only silently counted. This limitation exist to prevent the code
628from flooding your kernel log. To support debugging a device driver this can
629be disabled via debugfs. See the debugfs interface documentation below for
630details.
631
632The debugfs directory for the DMA-API debugging code is called dma-api/. In
633this directory the following files can currently be found:
634
635 dma-api/all_errors This file contains a numeric value. If this
636 value is not equal to zero the debugging code
637 will print a warning for every error it finds
Matt LaPlante19f59462009-04-27 15:06:31 +0200638 into the kernel log. Be careful with this
639 option, as it can easily flood your logs.
Joerg Roedel187f9c32009-01-09 16:28:07 +0100640
641 dma-api/disabled This read-only file contains the character 'Y'
642 if the debugging code is disabled. This can
643 happen when it runs out of memory or if it was
644 disabled at boot time
645
646 dma-api/error_count This file is read-only and shows the total
647 numbers of errors found.
648
649 dma-api/num_errors The number in this file shows how many
650 warnings will be printed to the kernel log
651 before it stops. This number is initialized to
652 one at system boot and be set by writing into
653 this file
654
655 dma-api/min_free_entries
656 This read-only file can be read to get the
657 minimum number of free dma_debug_entries the
658 allocator has ever seen. If this value goes
659 down to zero the code will disable itself
660 because it is not longer reliable.
661
662 dma-api/num_free_entries
663 The current number of free dma_debug_entries
664 in the allocator.
665
Joerg Roedel016ea682009-05-22 21:57:23 +0200666 dma-api/driver-filter
667 You can write a name of a driver into this file
668 to limit the debug output to requests from that
669 particular driver. Write an empty string to
670 that file to disable the filter and see
671 all errors again.
672
Joerg Roedel187f9c32009-01-09 16:28:07 +0100673If you have this code compiled into your kernel it will be enabled by default.
674If you want to boot without the bookkeeping anyway you can provide
675'dma_debug=off' as a boot parameter. This will disable DMA-API debugging.
676Notice that you can not enable it again at runtime. You have to reboot to do
677so.
678
Joerg Roedel016ea682009-05-22 21:57:23 +0200679If you want to see debug messages only for a special device driver you can
680specify the dma_debug_driver=<drivername> parameter. This will enable the
681driver filter at boot time. The debug code will only print errors for that
682driver afterwards. This filter can be disabled or changed later using debugfs.
683
Joerg Roedel187f9c32009-01-09 16:28:07 +0100684When the code disables itself at runtime this is most likely because it ran
685out of dma_debug_entries. These entries are preallocated at boot. The number
686of preallocated entries is defined per architecture. If it is too low for you
687boot with 'dma_debug_entries=<your_desired_number>' to overwrite the
688architectural default.
Shuah Khan6c9c6d62012-10-08 11:08:06 -0600689
690void debug_dmap_mapping_error(struct device *dev, dma_addr_t dma_addr);
691
692dma-debug interface debug_dma_mapping_error() to debug drivers that fail
693to check dma mapping errors on addresses returned by dma_map_single() and
694dma_map_page() interfaces. This interface clears a flag set by
695debug_dma_map_page() to indicate that dma_mapping_error() has been called by
696the driver. When driver does unmap, debug_dma_unmap() checks the flag and if
697this flag is still set, prints warning message that includes call trace that
698leads up to the unmap. This interface can be called from dma_mapping_error()
699routines to enable dma mapping error check debugging.
700