About Kernel Documentation Linux Kernel Contact Linux Resources Linux Blog

Documentation / DMA-API.txt




Custom Search

Based on kernel version 3.2. Page generated on 2012-01-05 23:28 EST.

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

Information is copyright its respective author. All material is available from the Linux Kernel Source distributed under a GPL License. This page is provided as a free service by mjmwired.net.