Documentation/DMA-API.txt

   1                Dynamic DMA mapping using the generic device
   2                ============================================
   3
   4         James E.J. Bottomley <James.Bottomley@HansenPartnership.com>
   5
   6 This document describes the DMA API.  For a more gentle introduction
   7 of the API (and actual examples), see Documentation/DMA-API-HOWTO.txt.
   8
   9 This API is split into two pieces.  Part I describes the basic API.
  10 Part II describes extensions for supporting non-consistent memory
  11 machines.  Unless you know that your driver absolutely has to support
  12 non-consistent platforms (this is usually only legacy platforms) you
  13 should only use the API described in part I.
  14
  15 Part I - dma_ API
  16 -------------------------------------
  17
  18 To get the dma_ API, you must #include <linux/dma-mapping.h>.  This
  19 provides dma_addr_t and the interfaces described below.
  20
  21 A dma_addr_t can hold any valid DMA address for the platform.  It can be
  22 given to a device to use as a DMA source or target.  A CPU cannot reference
  23 a dma_addr_t directly because there may be translation between its physical
  24 address space and the DMA address space.
  25
  26 Part Ia - Using large DMA-coherent buffers
  27 ------------------------------------------
  28
  29 void *
  30 dma_alloc_coherent(struct device *dev, size_t size,
  31                              dma_addr_t *dma_handle, gfp_t flag)
  32
  33 Consistent memory is memory for which a write by either the device or
  34 the processor can immediately be read by the processor or device
  35 without having to worry about caching effects.  (You may however need
  36 to make sure to flush the processor's write buffers before telling
  37 devices to read that memory.)
  38
  39 This routine allocates a region of <size> bytes of consistent memory.
  40
  41 It returns a pointer to the allocated region (in the processor's virtual
  42 address space) or NULL if the allocation failed.
  43
  44 It also returns a <dma_handle> which may be cast to an unsigned integer the
  45 same width as the bus and given to the device as the DMA address base of
  46 the region.
  47
  48 Note: consistent memory can be expensive on some platforms, and the
  49 minimum allocation length may be as big as a page, so you should
  50 consolidate your requests for consistent memory as much as possible.
  51 The simplest way to do that is to use the dma_pool calls (see below).
  52
  53 The flag parameter (dma_alloc_coherent() only) allows the caller to
  54 specify the GFP_ flags (see kmalloc()) for the allocation (the
  55 implementation may choose to ignore flags that affect the location of
  56 the returned memory, like GFP_DMA).
  57
  58 void *
  59 dma_zalloc_coherent(struct device *dev, size_t size,
  60                              dma_addr_t *dma_handle, gfp_t flag)
  61
  62 Wraps dma_alloc_coherent() and also zeroes the returned memory if the
  63 allocation attempt succeeded.
  64
  65 void
  66 dma_free_coherent(struct device *dev, size_t size, void *cpu_addr,
  67                            dma_addr_t dma_handle)
  68
  69 Free a region of consistent memory you previously allocated.  dev,
  70 size and dma_handle must all be the same as those passed into
  71 dma_alloc_coherent().  cpu_addr must be the virtual address returned by
  72 the dma_alloc_coherent().
  73
  74 Note that unlike their sibling allocation calls, these routines
  75 may only be called with IRQs enabled.
  76
  77
  78 Part Ib - Using small DMA-coherent buffers
  79 ------------------------------------------
  80
  81 To get this part of the dma_ API, you must #include <linux/dmapool.h>
  82
  83 Many drivers need lots of small DMA-coherent memory regions for DMA
  84 descriptors or I/O buffers.  Rather than allocating in units of a page
  85 or more using dma_alloc_coherent(), you can use DMA pools.  These work
  86 much like a struct kmem_cache, except that they use the DMA-coherent allocator,
  87 not __get_free_pages().  Also, they understand common hardware constraints
  88 for alignment, like queue heads needing to be aligned on N-byte boundaries.
  89
  90
  91         struct dma_pool *
  92         dma_pool_create(const char *name, struct device *dev,
  93                         size_t size, size_t align, size_t alloc);
  94
  95 dma_pool_create() initializes a pool of DMA-coherent buffers
  96 for use with a given device.  It must be called in a context which
  97 can sleep.
  98
  99 The "name" is for diagnostics (like a struct kmem_cache name); dev and size
 100 are like what you'd pass to dma_alloc_coherent().  The device's hardware
 101 alignment requirement for this type of data is "align" (which is expressed
 102 in bytes, and must be a power of two).  If your device has no boundary
 103 crossing restrictions, pass 0 for alloc; passing 4096 says memory allocated
 104 from this pool must not cross 4KByte boundaries.
 105
 106
 107         void *dma_pool_alloc(struct dma_pool *pool, gfp_t gfp_flags,
 108                         dma_addr_t *dma_handle);
 109
 110 This allocates memory from the pool; the returned memory will meet the
 111 size and alignment requirements specified at creation time.  Pass
 112 GFP_ATOMIC to prevent blocking, or if it's permitted (not
 113 in_interrupt, not holding SMP locks), pass GFP_KERNEL to allow
 114 blocking.  Like dma_alloc_coherent(), this returns two values:  an
 115 address usable by the CPU, and the DMA address usable by the pool's
 116 device.
 117
 118
 119         void dma_pool_free(struct dma_pool *pool, void *vaddr,
 120                         dma_addr_t addr);
 121
 122 This puts memory back into the pool.  The pool is what was passed to
 123 dma_pool_alloc(); the CPU (vaddr) and DMA addresses are what
 124 were returned when that routine allocated the memory being freed.
 125
 126
 127         void dma_pool_destroy(struct dma_pool *pool);
 128
 129 dma_pool_destroy() frees the resources of the pool.  It must be
 130 called in a context which can sleep.  Make sure you've freed all allocated
 131 memory back to the pool before you destroy it.
 132
 133
 134 Part Ic - DMA addressing limitations
 135 ------------------------------------
 136
 137 int
 138 dma_supported(struct device *dev, u64 mask)
 139
 140 Checks to see if the device can support DMA to the memory described by
 141 mask.
 142
 143 Returns: 1 if it can and 0 if it can't.
 144
 145 Notes: This routine merely tests to see if the mask is possible.  It
 146 won't change the current mask settings.  It is more intended as an
 147 internal API for use by the platform than an external API for use by
 148 driver writers.
 149
 150 int
 151 dma_set_mask_and_coherent(struct device *dev, u64 mask)
 152
 153 Checks to see if the mask is possible and updates the device
 154 streaming and coherent DMA mask parameters if it is.
 155
 156 Returns: 0 if successful and a negative error if not.
 157
 158 int
 159 dma_set_mask(struct device *dev, u64 mask)
 160
 161 Checks to see if the mask is possible and updates the device
 162 parameters if it is.
 163
 164 Returns: 0 if successful and a negative error if not.
 165
 166 int
 167 dma_set_coherent_mask(struct device *dev, u64 mask)
 168
 169 Checks to see if the mask is possible and updates the device
 170 parameters if it is.
 171
 172 Returns: 0 if successful and a negative error if not.
 173
 174 u64
 175 dma_get_required_mask(struct device *dev)
 176
 177 This API returns the mask that the platform requires to
 178 operate efficiently.  Usually this means the returned mask
 179 is the minimum required to cover all of memory.  Examining the
 180 required mask gives drivers with variable descriptor sizes the
 181 opportunity to use smaller descriptors as necessary.
 182
 183 Requesting the required mask does not alter the current mask.  If you
 184 wish to take advantage of it, you should issue a dma_set_mask()
 185 call to set the mask to the value returned.
 186
 187
 188 Part Id - Streaming DMA mappings
 189 --------------------------------
 190
 191 dma_addr_t
 192 dma_map_single(struct device *dev, void *cpu_addr, size_t size,
 193                       enum dma_data_direction direction)
 194
 195 Maps a piece of processor virtual memory so it can be accessed by the
 196 device and returns the DMA address of the memory.
 197
 198 The direction for both APIs may be converted freely by casting.
 199 However the dma_ API uses a strongly typed enumerator for its
 200 direction:
 201
 202 DMA_NONE                no direction (used for debugging)
 203 DMA_TO_DEVICE           data is going from the memory to the device
 204 DMA_FROM_DEVICE         data is coming from the device to the memory
 205 DMA_BIDIRECTIONAL       direction isn't known
 206
 207 Notes:  Not all memory regions in a machine can be mapped by this API.
 208 Further, contiguous kernel virtual space may not be contiguous as
 209 physical memory.  Since this API does not provide any scatter/gather
 210 capability, it will fail if the user tries to map a non-physically
 211 contiguous piece of memory.  For this reason, memory to be mapped by
 212 this API should be obtained from sources which guarantee it to be
 213 physically contiguous (like kmalloc).
 214
 215 Further, the DMA address of the memory must be within the
 216 dma_mask of the device (the dma_mask is a bit mask of the
 217 addressable region for the device, i.e., if the DMA address of
 218 the memory ANDed with the dma_mask is still equal to the DMA
 219 address, then the device can perform DMA to the memory).  To
 220 ensure that the memory allocated by kmalloc is within the dma_mask,
 221 the driver may specify various platform-dependent flags to restrict
 222 the DMA address range of the allocation (e.g., on x86, GFP_DMA
 223 guarantees to be within the first 16MB of available DMA addresses,
 224 as required by ISA devices).
 225
 226 Note also that the above constraints on physical contiguity and
 227 dma_mask may not apply if the platform has an IOMMU (a device which
 228 maps an I/O DMA address to a physical memory address).  However, to be
 229 portable, device driver writers may *not* assume that such an IOMMU
 230 exists.
 231
 232 Warnings:  Memory coherency operates at a granularity called the cache
 233 line width.  In order for memory mapped by this API to operate
 234 correctly, the mapped region must begin exactly on a cache line
 235 boundary and end exactly on one (to prevent two separately mapped
 236 regions from sharing a single cache line).  Since the cache line size
 237 may not be known at compile time, the API will not enforce this
 238 requirement.  Therefore, it is recommended that driver writers who
 239 don't take special care to determine the cache line size at run time
 240 only map virtual regions that begin and end on page boundaries (which
 241 are guaranteed also to be cache line boundaries).
 242
 243 DMA_TO_DEVICE synchronisation must be done after the last modification
 244 of the memory region by the software and before it is handed off to
 245 the driver.  Once this primitive is used, memory covered by this
 246 primitive should be treated as read-only by the device.  If the device
 247 may write to it at any point, it should be DMA_BIDIRECTIONAL (see
 248 below).
 249
 250 DMA_FROM_DEVICE synchronisation must be done before the driver
 251 accesses data that may be changed by the device.  This memory should
 252 be treated as read-only by the driver.  If the driver needs to write
 253 to it at any point, it should be DMA_BIDIRECTIONAL (see below).
 254
 255 DMA_BIDIRECTIONAL requires special handling: it means that the driver
 256 isn't sure if the memory was modified before being handed off to the
 257 device and also isn't sure if the device will also modify it.  Thus,
 258 you must always sync bidirectional memory twice: once before the
 259 memory is handed off to the device (to make sure all memory changes
 260 are flushed from the processor) and once before the data may be
 261 accessed after being used by the device (to make sure any processor
 262 cache lines are updated with data that the device may have changed).
 263
 264 void
 265 dma_unmap_single(struct device *dev, dma_addr_t dma_addr, size_t size,
 266                  enum dma_data_direction direction)
 267
 268 Unmaps the region previously mapped.  All the parameters passed in
 269 must be identical to those passed in (and returned) by the mapping
 270 API.
 271
 272 dma_addr_t
 273 dma_map_page(struct device *dev, struct page *page,
 274                     unsigned long offset, size_t size,
 275                     enum dma_data_direction direction)
 276 void
 277 dma_unmap_page(struct device *dev, dma_addr_t dma_address, size_t size,
 278                enum dma_data_direction direction)
 279
 280 API for mapping and unmapping for pages.  All the notes and warnings
 281 for the other mapping APIs apply here.  Also, although the <offset>
 282 and <size> parameters are provided to do partial page mapping, it is
 283 recommended that you never use these unless you really know what the
 284 cache width is.
 285
 286 int
 287 dma_mapping_error(struct device *dev, dma_addr_t dma_addr)
 288
 289 In some circumstances dma_map_single() and dma_map_page() will fail to create
 290 a mapping. A driver can check for these errors by testing the returned
 291 DMA address with dma_mapping_error(). A non-zero return value means the mapping
 292 could not be created and the driver should take appropriate action (e.g.
 293 reduce current DMA mapping usage or delay and try again later).
 294
 295         int
 296         dma_map_sg(struct device *dev, struct scatterlist *sg,
 297                 int nents, enum dma_data_direction direction)
 298
 299 Returns: the number of DMA address segments mapped (this may be shorter
 300 than <nents> passed in if some elements of the scatter/gather list are
 301 physically or virtually adjacent and an IOMMU maps them with a single
 302 entry).
 303
 304 Please note that the sg cannot be mapped again if it has been mapped once.
 305 The mapping process is allowed to destroy information in the sg.
 306
 307 As with the other mapping interfaces, dma_map_sg() can fail. When it
 308 does, 0 is returned and a driver must take appropriate action. It is
 309 critical that the driver do something, in the case of a block driver
 310 aborting the request or even oopsing is better than doing nothing and
 311 corrupting the filesystem.
 312
 313 With scatterlists, you use the resulting mapping like this:
 314
 315         int i, count = dma_map_sg(dev, sglist, nents, direction);
 316         struct scatterlist *sg;
 317
 318         for_each_sg(sglist, sg, count, i) {
 319                 hw_address[i] = sg_dma_address(sg);
 320                 hw_len[i] = sg_dma_len(sg);
 321         }
 322
 323 where nents is the number of entries in the sglist.
 324
 325 The implementation is free to merge several consecutive sglist entries
 326 into one (e.g. with an IOMMU, or if several pages just happen to be
 327 physically contiguous) and returns the actual number of sg entries it
 328 mapped them to. On failure 0, is returned.
 329
 330 Then you should loop count times (note: this can be less than nents times)
 331 and use sg_dma_address() and sg_dma_len() macros where you previously
 332 accessed sg->address and sg->length as shown above.
 333
 334         void
 335         dma_unmap_sg(struct device *dev, struct scatterlist *sg,
 336                 int nhwentries, enum dma_data_direction direction)
 337
 338 Unmap the previously mapped scatter/gather list.  All the parameters
 339 must be the same as those and passed in to the scatter/gather mapping
 340 API.
 341
 342 Note: <nents> must be the number you passed in, *not* the number of
 343 DMA address entries returned.
 344
 345 void
 346 dma_sync_single_for_cpu(struct device *dev, dma_addr_t dma_handle, size_t size,
 347                         enum dma_data_direction direction)
 348 void
 349 dma_sync_single_for_device(struct device *dev, dma_addr_t dma_handle, size_t size,
 350                            enum dma_data_direction direction)
 351 void
 352 dma_sync_sg_for_cpu(struct device *dev, struct scatterlist *sg, int nelems,
 353                     enum dma_data_direction direction)
 354 void
 355 dma_sync_sg_for_device(struct device *dev, struct scatterlist *sg, int nelems,
 356                        enum dma_data_direction direction)
 357
 358 Synchronise a single contiguous or scatter/gather mapping for the CPU
 359 and device. With the sync_sg API, all the parameters must be the same
 360 as those passed into the single mapping API. With the sync_single API,
 361 you can use dma_handle and size parameters that aren't identical to
 362 those passed into the single mapping API to do a partial sync.
 363
 364 Notes:  You must do this:
 365
 366 - Before reading values that have been written by DMA from the device
 367   (use the DMA_FROM_DEVICE direction)
 368 - After writing values that will be written to the device using DMA
 369   (use the DMA_TO_DEVICE) direction
 370 - before *and* after handing memory to the device if the memory is
 371   DMA_BIDIRECTIONAL
 372
 373 See also dma_map_single().
 374
 375 dma_addr_t
 376 dma_map_single_attrs(struct device *dev, void *cpu_addr, size_t size,
 377                      enum dma_data_direction dir,
 378                      struct dma_attrs *attrs)
 379
 380 void
 381 dma_unmap_single_attrs(struct device *dev, dma_addr_t dma_addr,
 382                        size_t size, enum dma_data_direction dir,
 383                        struct dma_attrs *attrs)
 384
 385 int
 386 dma_map_sg_attrs(struct device *dev, struct scatterlist *sgl,
 387                  int nents, enum dma_data_direction dir,
 388                  struct dma_attrs *attrs)
 389
 390 void
 391 dma_unmap_sg_attrs(struct device *dev, struct scatterlist *sgl,
 392                    int nents, enum dma_data_direction dir,
 393                    struct dma_attrs *attrs)
 394
 395 The four functions above are just like the counterpart functions
 396 without the _attrs suffixes, except that they pass an optional
 397 struct dma_attrs*.
 398
 399 struct dma_attrs encapsulates a set of "DMA attributes". For the
 400 definition of struct dma_attrs see linux/dma-attrs.h.
 401
 402 The interpretation of DMA attributes is architecture-specific, and
 403 each attribute should be documented in Documentation/DMA-attributes.txt.
 404
 405 If struct dma_attrs* is NULL, the semantics of each of these
 406 functions is identical to those of the corresponding function
 407 without the _attrs suffix. As a result dma_map_single_attrs()
 408 can generally replace dma_map_single(), etc.
 409
 410 As an example of the use of the *_attrs functions, here's how
 411 you could pass an attribute DMA_ATTR_FOO when mapping memory
 412 for DMA:
 413
 414 #include <linux/dma-attrs.h>
 415 /* DMA_ATTR_FOO should be defined in linux/dma-attrs.h and
 416  * documented in Documentation/DMA-attributes.txt */
 417 ...
 418
 419         DEFINE_DMA_ATTRS(attrs);
 420         dma_set_attr(DMA_ATTR_FOO, &attrs);
 421         ....
 422         n = dma_map_sg_attrs(dev, sg, nents, DMA_TO_DEVICE, &attr);
 423         ....
 424
 425 Architectures that care about DMA_ATTR_FOO would check for its
 426 presence in their implementations of the mapping and unmapping
 427 routines, e.g.:
 428
 429 void whizco_dma_map_sg_attrs(struct device *dev, dma_addr_t dma_addr,
 430                              size_t size, enum dma_data_direction dir,
 431                              struct dma_attrs *attrs)
 432 {
 433         ....
 434         int foo =  dma_get_attr(DMA_ATTR_FOO, attrs);
 435         ....
 436         if (foo)
 437                 /* twizzle the frobnozzle */
 438         ....
 439
 440
 441 Part II - Advanced dma_ usage
 442 -----------------------------
 443
 444 Warning: These pieces of the DMA API should not be used in the
 445 majority of cases, since they cater for unlikely corner cases that
 446 don't belong in usual drivers.
 447
 448 If you don't understand how cache line coherency works between a
 449 processor and an I/O device, you should not be using this part of the
 450 API at all.
 451
 452 void *
 453 dma_alloc_noncoherent(struct device *dev, size_t size,
 454                                dma_addr_t *dma_handle, gfp_t flag)
 455
 456 Identical to dma_alloc_coherent() except that the platform will
 457 choose to return either consistent or non-consistent memory as it sees
 458 fit.  By using this API, you are guaranteeing to the platform that you
 459 have all the correct and necessary sync points for this memory in the
 460 driver should it choose to return non-consistent memory.
 461
 462 Note: where the platform can return consistent memory, it will
 463 guarantee that the sync points become nops.
 464
 465 Warning:  Handling non-consistent memory is a real pain.  You should
 466 only use this API if you positively know your driver will be
 467 required to work on one of the rare (usually non-PCI) architectures
 468 that simply cannot make consistent memory.
 469
 470 void
 471 dma_free_noncoherent(struct device *dev, size_t size, void *cpu_addr,
 472                               dma_addr_t dma_handle)
 473
 474 Free memory allocated by the nonconsistent API.  All parameters must
 475 be identical to those passed in (and returned by
 476 dma_alloc_noncoherent()).
 477
 478 int
 479 dma_get_cache_alignment(void)
 480
 481 Returns the processor cache alignment.  This is the absolute minimum
 482 alignment *and* width that you must observe when either mapping
 483 memory or doing partial flushes.
 484
 485 Notes: This API may return a number *larger* than the actual cache
 486 line, but it will guarantee that one or more cache lines fit exactly
 487 into the width returned by this call.  It will also always be a power
 488 of two for easy alignment.
 489
 490 void
 491 dma_cache_sync(struct device *dev, void *vaddr, size_t size,
 492                enum dma_data_direction direction)
 493
 494 Do a partial sync of memory that was allocated by
 495 dma_alloc_noncoherent(), starting at virtual address vaddr and
 496 continuing on for size.  Again, you *must* observe the cache line
 497 boundaries when doing this.
 498
 499 int
 500 dma_declare_coherent_memory(struct device *dev, phys_addr_t phys_addr,
 501                             dma_addr_t device_addr, size_t size, int
 502                             flags)
 503
 504 Declare region of memory to be handed out by dma_alloc_coherent() when
 505 it's asked for coherent memory for this device.
 506
 507 phys_addr is the CPU physical address to which the memory is currently
 508 assigned (this will be ioremapped so the CPU can access the region).
 509
 510 device_addr is the DMA address the device needs to be programmed
 511 with to actually address this memory (this will be handed out as the
 512 dma_addr_t in dma_alloc_coherent()).
 513
 514 size is the size of the area (must be multiples of PAGE_SIZE).
 515
 516 flags can be ORed together and are:
 517
 518 DMA_MEMORY_MAP - request that the memory returned from
 519 dma_alloc_coherent() be directly writable.
 520
 521 DMA_MEMORY_IO - request that the memory returned from
 522 dma_alloc_coherent() be addressable using read()/write()/memcpy_toio() etc.
 523
 524 One or both of these flags must be present.
 525
 526 DMA_MEMORY_INCLUDES_CHILDREN - make the declared memory be allocated by
 527 dma_alloc_coherent of any child devices of this one (for memory residing
 528 on a bridge).
 529
 530 DMA_MEMORY_EXCLUSIVE - only allocate memory from the declared regions.
 531 Do not allow dma_alloc_coherent() to fall back to system memory when
 532 it's out of memory in the declared region.
 533
 534 The return value will be either DMA_MEMORY_MAP or DMA_MEMORY_IO and
 535 must correspond to a passed in flag (i.e. no returning DMA_MEMORY_IO
 536 if only DMA_MEMORY_MAP were passed in) for success or zero for
 537 failure.
 538
 539 Note, for DMA_MEMORY_IO returns, all subsequent memory returned by
 540 dma_alloc_coherent() may no longer be accessed directly, but instead
 541 must be accessed using the correct bus functions.  If your driver
 542 isn't prepared to handle this contingency, it should not specify
 543 DMA_MEMORY_IO in the input flags.
 544
 545 As a simplification for the platforms, only *one* such region of
 546 memory may be declared per device.
 547
 548 For reasons of efficiency, most platforms choose to track the declared
 549 region only at the granularity of a page.  For smaller allocations,
 550 you should use the dma_pool() API.
 551
 552 void
 553 dma_release_declared_memory(struct device *dev)
 554
 555 Remove the memory region previously declared from the system.  This
 556 API performs *no* in-use checking for this region and will return
 557 unconditionally having removed all the required structures.  It is the
 558 driver's job to ensure that no parts of this memory region are
 559 currently in use.
 560
 561 void *
 562 dma_mark_declared_memory_occupied(struct device *dev,
 563                                   dma_addr_t device_addr, size_t size)
 564
 565 This is used to occupy specific regions of the declared space
 566 (dma_alloc_coherent() will hand out the first free region it finds).
 567
 568 device_addr is the *device* address of the region requested.
 569
 570 size is the size (and should be a page-sized multiple).
 571
 572 The return value will be either a pointer to the processor virtual
 573 address of the memory, or an error (via PTR_ERR()) if any part of the
 574 region is occupied.
 575
 576 Part III - Debug drivers use of the DMA-API
 577 -------------------------------------------
 578
 579 The DMA-API as described above has some constraints. DMA addresses must be
 580 released with the corresponding function with the same size for example. With
 581 the advent of hardware IOMMUs it becomes more and more important that drivers
 582 do not violate those constraints. In the worst case such a violation can
 583 result in data corruption up to destroyed filesystems.
 584
 585 To debug drivers and find bugs in the usage of the DMA-API checking code can
 586 be compiled into the kernel which will tell the developer about those
 587 violations. If your architecture supports it you can select the "Enable
 588 debugging of DMA-API usage" option in your kernel configuration. Enabling this
 589 option has a performance impact. Do not enable it in production kernels.
 590
 591 If you boot the resulting kernel will contain code which does some bookkeeping
 592 about what DMA memory was allocated for which device. If this code detects an
 593 error it prints a warning message with some details into your kernel log. An
 594 example warning message may look like this:
 595
 596 ------------[ cut here ]------------
 597 WARNING: at /data2/repos/linux-2.6-iommu/lib/dma-debug.c:448
 598         check_unmap+0x203/0x490()
 599 Hardware name:
 600 forcedeth 0000:00:08.0: DMA-API: device driver frees DMA memory with wrong
 601         function [device address=0x00000000640444be] [size=66 bytes] [mapped as
 602 single] [unmapped as page]
 603 Modules linked in: nfsd exportfs bridge stp llc r8169
 604 Pid: 0, comm: swapper Tainted: G        W  2.6.28-dmatest-09289-g8bb99c0 #1
 605 Call Trace:
 606  <IRQ>  [<ffffffff80240b22>] warn_slowpath+0xf2/0x130
 607  [<ffffffff80647b70>] _spin_unlock+0x10/0x30
 608  [<ffffffff80537e75>] usb_hcd_link_urb_to_ep+0x75/0xc0
 609  [<ffffffff80647c22>] _spin_unlock_irqrestore+0x12/0x40
 610  [<ffffffff8055347f>] ohci_urb_enqueue+0x19f/0x7c0
 611  [<ffffffff80252f96>] queue_work+0x56/0x60
 612  [<ffffffff80237e10>] enqueue_task_fair+0x20/0x50
 613  [<ffffffff80539279>] usb_hcd_submit_urb+0x379/0xbc0
 614  [<ffffffff803b78c3>] cpumask_next_and+0x23/0x40
 615  [<ffffffff80235177>] find_busiest_group+0x207/0x8a0
 616  [<ffffffff8064784f>] _spin_lock_irqsave+0x1f/0x50
 617  [<ffffffff803c7ea3>] check_unmap+0x203/0x490
 618  [<ffffffff803c8259>] debug_dma_unmap_page+0x49/0x50
 619  [<ffffffff80485f26>] nv_tx_done_optimized+0xc6/0x2c0
 620  [<ffffffff80486c13>] nv_nic_irq_optimized+0x73/0x2b0
 621  [<ffffffff8026df84>] handle_IRQ_event+0x34/0x70
 622  [<ffffffff8026ffe9>] handle_edge_irq+0xc9/0x150
 623  [<ffffffff8020e3ab>] do_IRQ+0xcb/0x1c0
 624  [<ffffffff8020c093>] ret_from_intr+0x0/0xa
 625  <EOI> <4>---[ end trace f6435a98e2a38c0e ]---
 626
 627 The driver developer can find the driver and the device including a stacktrace
 628 of the DMA-API call which caused this warning.
 629
 630 Per default only the first error will result in a warning message. All other
 631 errors will only silently counted. This limitation exist to prevent the code
 632 from flooding your kernel log. To support debugging a device driver this can
 633 be disabled via debugfs. See the debugfs interface documentation below for
 634 details.
 635
 636 The debugfs directory for the DMA-API debugging code is called dma-api/. In
 637 this directory the following files can currently be found:
 638
 639         dma-api/all_errors      This file contains a numeric value. If this
 640                                 value is not equal to zero the debugging code
 641                                 will print a warning for every error it finds
 642                                 into the kernel log. Be careful with this
 643                                 option, as it can easily flood your logs.
 644
 645         dma-api/disabled        This read-only file contains the character 'Y'
 646                                 if the debugging code is disabled. This can
 647                                 happen when it runs out of memory or if it was
 648                                 disabled at boot time
 649
 650         dma-api/error_count     This file is read-only and shows the total
 651                                 numbers of errors found.
 652
 653         dma-api/num_errors      The number in this file shows how many
 654                                 warnings will be printed to the kernel log
 655                                 before it stops. This number is initialized to
 656                                 one at system boot and be set by writing into
 657                                 this file
 658
 659         dma-api/min_free_entries
 660                                 This read-only file can be read to get the
 661                                 minimum number of free dma_debug_entries the
 662                                 allocator has ever seen. If this value goes
 663                                 down to zero the code will disable itself
 664                                 because it is not longer reliable.
 665
 666         dma-api/num_free_entries
 667                                 The current number of free dma_debug_entries
 668                                 in the allocator.
 669
 670         dma-api/driver-filter
 671                                 You can write a name of a driver into this file
 672                                 to limit the debug output to requests from that
 673                                 particular driver. Write an empty string to
 674                                 that file to disable the filter and see
 675                                 all errors again.
 676
 677 If you have this code compiled into your kernel it will be enabled by default.
 678 If you want to boot without the bookkeeping anyway you can provide
 679 'dma_debug=off' as a boot parameter. This will disable DMA-API debugging.
 680 Notice that you can not enable it again at runtime. You have to reboot to do
 681 so.
 682
 683 If you want to see debug messages only for a special device driver you can
 684 specify the dma_debug_driver=<drivername> parameter. This will enable the
 685 driver filter at boot time. The debug code will only print errors for that
 686 driver afterwards. This filter can be disabled or changed later using debugfs.
 687
 688 When the code disables itself at runtime this is most likely because it ran
 689 out of dma_debug_entries. These entries are preallocated at boot. The number
 690 of preallocated entries is defined per architecture. If it is too low for you
 691 boot with 'dma_debug_entries=<your_desired_number>' to overwrite the
 692 architectural default.
 693
 694 void debug_dmap_mapping_error(struct device *dev, dma_addr_t dma_addr);
 695
 696 dma-debug interface debug_dma_mapping_error() to debug drivers that fail
 697 to check DMA mapping errors on addresses returned by dma_map_single() and
 698 dma_map_page() interfaces. This interface clears a flag set by
 699 debug_dma_map_page() to indicate that dma_mapping_error() has been called by
 700 the driver. When driver does unmap, debug_dma_unmap() checks the flag and if
 701 this flag is still set, prints warning message that includes call trace that
 702 leads up to the unmap. This interface can be called from dma_mapping_error()
 703 routines to enable DMA mapping error check debugging.
 704