Blame view

block/blk-settings.c 25.7 KB
86db1e297   Jens Axboe   block: continue l...
1
2
3
4
5
6
7
8
9
  /*
   * Functions related to setting various queue properties from drivers
   */
  #include <linux/kernel.h>
  #include <linux/module.h>
  #include <linux/init.h>
  #include <linux/bio.h>
  #include <linux/blkdev.h>
  #include <linux/bootmem.h>	/* for max_pfn/max_low_pfn */
70dd5bf3b   Martin K. Petersen   block: Stack opti...
10
  #include <linux/gcd.h>
2cda2728a   Martin K. Petersen   block: Fix overru...
11
  #include <linux/lcm.h>
ad5ebd2fa   Randy Dunlap   block: jiffies fixes
12
  #include <linux/jiffies.h>
5a0e3ad6a   Tejun Heo   include cleanup: ...
13
  #include <linux/gfp.h>
86db1e297   Jens Axboe   block: continue l...
14
15
  
  #include "blk.h"
6728cb0e6   Jens Axboe   block: make core ...
16
  unsigned long blk_max_low_pfn;
86db1e297   Jens Axboe   block: continue l...
17
  EXPORT_SYMBOL(blk_max_low_pfn);
6728cb0e6   Jens Axboe   block: make core ...
18
19
  
  unsigned long blk_max_pfn;
86db1e297   Jens Axboe   block: continue l...
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
  
  /**
   * blk_queue_prep_rq - set a prepare_request function for queue
   * @q:		queue
   * @pfn:	prepare_request function
   *
   * It's possible for a queue to register a prepare_request callback which
   * is invoked before the request is handed to the request_fn. The goal of
   * the function is to prepare a request for I/O, it can be used to build a
   * cdb from the request data for instance.
   *
   */
  void blk_queue_prep_rq(struct request_queue *q, prep_rq_fn *pfn)
  {
  	q->prep_rq_fn = pfn;
  }
86db1e297   Jens Axboe   block: continue l...
36
37
38
  EXPORT_SYMBOL(blk_queue_prep_rq);
  
  /**
28018c242   James Bottomley   block: implement ...
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
   * blk_queue_unprep_rq - set an unprepare_request function for queue
   * @q:		queue
   * @ufn:	unprepare_request function
   *
   * It's possible for a queue to register an unprepare_request callback
   * which is invoked before the request is finally completed. The goal
   * of the function is to deallocate any data that was allocated in the
   * prepare_request callback.
   *
   */
  void blk_queue_unprep_rq(struct request_queue *q, unprep_rq_fn *ufn)
  {
  	q->unprep_rq_fn = ufn;
  }
  EXPORT_SYMBOL(blk_queue_unprep_rq);
  
  /**
86db1e297   Jens Axboe   block: continue l...
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
   * blk_queue_merge_bvec - set a merge_bvec function for queue
   * @q:		queue
   * @mbfn:	merge_bvec_fn
   *
   * Usually queues have static limitations on the max sectors or segments that
   * we can put in a request. Stacking drivers may have some settings that
   * are dynamic, and thus we have to query the queue whether it is ok to
   * add a new bio_vec to a bio at a given offset or not. If the block device
   * has such limitations, it needs to register a merge_bvec_fn to control
   * the size of bio's sent to it. Note that a block device *must* allow a
   * single page to be added to an empty bio. The block device driver may want
   * to use the bio_split() function to deal with these bio's. By default
   * no merge_bvec_fn is defined for a queue, and only the fixed limits are
   * honored.
   */
  void blk_queue_merge_bvec(struct request_queue *q, merge_bvec_fn *mbfn)
  {
  	q->merge_bvec_fn = mbfn;
  }
86db1e297   Jens Axboe   block: continue l...
75
76
77
78
79
80
  EXPORT_SYMBOL(blk_queue_merge_bvec);
  
  void blk_queue_softirq_done(struct request_queue *q, softirq_done_fn *fn)
  {
  	q->softirq_done_fn = fn;
  }
86db1e297   Jens Axboe   block: continue l...
81
  EXPORT_SYMBOL(blk_queue_softirq_done);
242f9dcb8   Jens Axboe   block: unify requ...
82
83
84
85
86
87
88
89
90
91
92
  void blk_queue_rq_timeout(struct request_queue *q, unsigned int timeout)
  {
  	q->rq_timeout = timeout;
  }
  EXPORT_SYMBOL_GPL(blk_queue_rq_timeout);
  
  void blk_queue_rq_timed_out(struct request_queue *q, rq_timed_out_fn *fn)
  {
  	q->rq_timed_out_fn = fn;
  }
  EXPORT_SYMBOL_GPL(blk_queue_rq_timed_out);
ef9e3facd   Kiyoshi Ueda   block: add lld bu...
93
94
95
96
97
  void blk_queue_lld_busy(struct request_queue *q, lld_busy_fn *fn)
  {
  	q->lld_busy_fn = fn;
  }
  EXPORT_SYMBOL_GPL(blk_queue_lld_busy);
86db1e297   Jens Axboe   block: continue l...
98
  /**
e475bba2f   Martin K. Petersen   block: Introduce ...
99
   * blk_set_default_limits - reset limits to default values
f740f5ca0   Randy Dunlap   Fix kernel-doc pa...
100
   * @lim:  the queue_limits structure to reset
e475bba2f   Martin K. Petersen   block: Introduce ...
101
102
   *
   * Description:
b1bd055d3   Martin K. Petersen   block: Introduce ...
103
   *   Returns a queue_limit struct to its default state.
e475bba2f   Martin K. Petersen   block: Introduce ...
104
105
106
   */
  void blk_set_default_limits(struct queue_limits *lim)
  {
8a78362c4   Martin K. Petersen   block: Consolidat...
107
  	lim->max_segments = BLK_MAX_SEGMENTS;
13f05c8d8   Martin K. Petersen   block/scsi: Provi...
108
  	lim->max_integrity_segments = 0;
e475bba2f   Martin K. Petersen   block: Introduce ...
109
  	lim->seg_boundary_mask = BLK_SEG_BOUNDARY_MASK;
eb28d31bc   Martin K. Petersen   block: Add BLK_ p...
110
  	lim->max_segment_size = BLK_MAX_SEGMENT_SIZE;
b1bd055d3   Martin K. Petersen   block: Introduce ...
111
  	lim->max_sectors = lim->max_hw_sectors = BLK_SAFE_MAX_SECTORS;
86b372814   Martin K. Petersen   block: Expose dis...
112
113
114
115
  	lim->max_discard_sectors = 0;
  	lim->discard_granularity = 0;
  	lim->discard_alignment = 0;
  	lim->discard_misaligned = 0;
b1bd055d3   Martin K. Petersen   block: Introduce ...
116
  	lim->discard_zeroes_data = 0;
e475bba2f   Martin K. Petersen   block: Introduce ...
117
  	lim->logical_block_size = lim->physical_block_size = lim->io_min = 512;
3a02c8e81   Martin K. Petersen   block: Fix bounce...
118
  	lim->bounce_pfn = (unsigned long)(BLK_BOUNCE_ANY >> PAGE_SHIFT);
e475bba2f   Martin K. Petersen   block: Introduce ...
119
120
121
  	lim->alignment_offset = 0;
  	lim->io_opt = 0;
  	lim->misaligned = 0;
e692cb668   Martin K. Petersen   block: Deprecate ...
122
  	lim->cluster = 1;
e475bba2f   Martin K. Petersen   block: Introduce ...
123
124
125
126
  }
  EXPORT_SYMBOL(blk_set_default_limits);
  
  /**
b1bd055d3   Martin K. Petersen   block: Introduce ...
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
   * blk_set_stacking_limits - set default limits for stacking devices
   * @lim:  the queue_limits structure to reset
   *
   * Description:
   *   Returns a queue_limit struct to its default state. Should be used
   *   by stacking drivers like DM that have no internal limits.
   */
  void blk_set_stacking_limits(struct queue_limits *lim)
  {
  	blk_set_default_limits(lim);
  
  	/* Inherit limits from component devices */
  	lim->discard_zeroes_data = 1;
  	lim->max_segments = USHRT_MAX;
  	lim->max_hw_sectors = UINT_MAX;
  
  	lim->max_sectors = BLK_DEF_MAX_SECTORS;
  }
  EXPORT_SYMBOL(blk_set_stacking_limits);
  
  /**
86db1e297   Jens Axboe   block: continue l...
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
   * blk_queue_make_request - define an alternate make_request function for a device
   * @q:  the request queue for the device to be affected
   * @mfn: the alternate make_request function
   *
   * Description:
   *    The normal way for &struct bios to be passed to a device
   *    driver is for them to be collected into requests on a request
   *    queue, and then to allow the device driver to select requests
   *    off that queue when it is ready.  This works well for many block
   *    devices. However some block devices (typically virtual devices
   *    such as md or lvm) do not benefit from the processing on the
   *    request queue, and are served best by having the requests passed
   *    directly to them.  This can be achieved by providing a function
   *    to blk_queue_make_request().
   *
   * Caveat:
   *    The driver that does this *must* be able to deal appropriately
   *    with buffers in "highmemory". This can be accomplished by either calling
   *    __bio_kmap_atomic() to get a temporary kernel mapping, or by calling
   *    blk_queue_bounce() to create a buffer in normal memory.
   **/
6728cb0e6   Jens Axboe   block: make core ...
169
  void blk_queue_make_request(struct request_queue *q, make_request_fn *mfn)
86db1e297   Jens Axboe   block: continue l...
170
171
172
173
174
  {
  	/*
  	 * set defaults
  	 */
  	q->nr_requests = BLKDEV_MAX_RQ;
0e435ac26   Milan Broz   block: fix settin...
175

86db1e297   Jens Axboe   block: continue l...
176
  	q->make_request_fn = mfn;
86db1e297   Jens Axboe   block: continue l...
177
178
179
  	blk_queue_dma_alignment(q, 511);
  	blk_queue_congestion_threshold(q);
  	q->nr_batching = BLK_BATCH_REQ;
e475bba2f   Martin K. Petersen   block: Introduce ...
180
  	blk_set_default_limits(&q->limits);
86db1e297   Jens Axboe   block: continue l...
181
182
183
184
185
  	/*
  	 * by default assume old behaviour and bounce for any highmem page
  	 */
  	blk_queue_bounce_limit(q, BLK_BOUNCE_HIGH);
  }
86db1e297   Jens Axboe   block: continue l...
186
187
188
189
  EXPORT_SYMBOL(blk_queue_make_request);
  
  /**
   * blk_queue_bounce_limit - set bounce buffer limit for queue
cd0aca2d5   Tejun Heo   block: fix queue ...
190
191
   * @q: the request queue for the device
   * @dma_mask: the maximum address the device can handle
86db1e297   Jens Axboe   block: continue l...
192
193
194
195
196
   *
   * Description:
   *    Different hardware can have different requirements as to what pages
   *    it can do I/O directly to. A low level driver can call
   *    blk_queue_bounce_limit to have lower memory pages allocated as bounce
cd0aca2d5   Tejun Heo   block: fix queue ...
197
   *    buffers for doing I/O to pages residing above @dma_mask.
86db1e297   Jens Axboe   block: continue l...
198
   **/
cd0aca2d5   Tejun Heo   block: fix queue ...
199
  void blk_queue_bounce_limit(struct request_queue *q, u64 dma_mask)
86db1e297   Jens Axboe   block: continue l...
200
  {
cd0aca2d5   Tejun Heo   block: fix queue ...
201
  	unsigned long b_pfn = dma_mask >> PAGE_SHIFT;
86db1e297   Jens Axboe   block: continue l...
202
203
204
205
  	int dma = 0;
  
  	q->bounce_gfp = GFP_NOIO;
  #if BITS_PER_LONG == 64
cd0aca2d5   Tejun Heo   block: fix queue ...
206
207
208
209
210
211
  	/*
  	 * Assume anything <= 4GB can be handled by IOMMU.  Actually
  	 * some IOMMUs can handle everything, but I don't know of a
  	 * way to test this here.
  	 */
  	if (b_pfn < (min_t(u64, 0xffffffffUL, BLK_BOUNCE_HIGH) >> PAGE_SHIFT))
86db1e297   Jens Axboe   block: continue l...
212
  		dma = 1;
efb012b36   Malahal Naineni   block: set the bo...
213
  	q->limits.bounce_pfn = max(max_low_pfn, b_pfn);
86db1e297   Jens Axboe   block: continue l...
214
  #else
6728cb0e6   Jens Axboe   block: make core ...
215
  	if (b_pfn < blk_max_low_pfn)
86db1e297   Jens Axboe   block: continue l...
216
  		dma = 1;
c49825fac   Malahal Naineni   block: set the bo...
217
  	q->limits.bounce_pfn = b_pfn;
260a67a9e   Jens Axboe   block: revert bad...
218
  #endif
86db1e297   Jens Axboe   block: continue l...
219
220
221
  	if (dma) {
  		init_emergency_isa_pool();
  		q->bounce_gfp = GFP_NOIO | GFP_DMA;
260a67a9e   Jens Axboe   block: revert bad...
222
  		q->limits.bounce_pfn = b_pfn;
86db1e297   Jens Axboe   block: continue l...
223
224
  	}
  }
86db1e297   Jens Axboe   block: continue l...
225
226
227
  EXPORT_SYMBOL(blk_queue_bounce_limit);
  
  /**
72d4cd9f3   Mike Snitzer   block: max hardwa...
228
229
   * blk_limits_max_hw_sectors - set hard and soft limit of max sectors for request
   * @limits: the queue limits
2800aac11   Martin K. Petersen   block: Update blk...
230
   * @max_hw_sectors:  max hardware sectors in the usual 512b unit
86db1e297   Jens Axboe   block: continue l...
231
232
   *
   * Description:
2800aac11   Martin K. Petersen   block: Update blk...
233
234
235
236
237
238
239
240
241
   *    Enables a low level driver to set a hard upper limit,
   *    max_hw_sectors, on the size of requests.  max_hw_sectors is set by
   *    the device driver based upon the combined capabilities of I/O
   *    controller and storage device.
   *
   *    max_sectors is a soft limit imposed by the block layer for
   *    filesystem type requests.  This value can be overridden on a
   *    per-device basis in /sys/block/<device>/queue/max_sectors_kb.
   *    The soft limit can not exceed max_hw_sectors.
86db1e297   Jens Axboe   block: continue l...
242
   **/
72d4cd9f3   Mike Snitzer   block: max hardwa...
243
  void blk_limits_max_hw_sectors(struct queue_limits *limits, unsigned int max_hw_sectors)
86db1e297   Jens Axboe   block: continue l...
244
  {
2800aac11   Martin K. Petersen   block: Update blk...
245
246
  	if ((max_hw_sectors << 9) < PAGE_CACHE_SIZE) {
  		max_hw_sectors = 1 << (PAGE_CACHE_SHIFT - 9);
24c03d47d   Harvey Harrison   block: remove rem...
247
248
  		printk(KERN_INFO "%s: set to minimum %d
  ",
2800aac11   Martin K. Petersen   block: Update blk...
249
  		       __func__, max_hw_sectors);
86db1e297   Jens Axboe   block: continue l...
250
  	}
72d4cd9f3   Mike Snitzer   block: max hardwa...
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
  	limits->max_hw_sectors = max_hw_sectors;
  	limits->max_sectors = min_t(unsigned int, max_hw_sectors,
  				    BLK_DEF_MAX_SECTORS);
  }
  EXPORT_SYMBOL(blk_limits_max_hw_sectors);
  
  /**
   * blk_queue_max_hw_sectors - set max sectors for a request for this queue
   * @q:  the request queue for the device
   * @max_hw_sectors:  max hardware sectors in the usual 512b unit
   *
   * Description:
   *    See description for blk_limits_max_hw_sectors().
   **/
  void blk_queue_max_hw_sectors(struct request_queue *q, unsigned int max_hw_sectors)
  {
  	blk_limits_max_hw_sectors(&q->limits, max_hw_sectors);
86db1e297   Jens Axboe   block: continue l...
268
  }
086fa5ff0   Martin K. Petersen   block: Rename blk...
269
  EXPORT_SYMBOL(blk_queue_max_hw_sectors);
86db1e297   Jens Axboe   block: continue l...
270
271
  
  /**
67efc9258   Christoph Hellwig   block: allow larg...
272
273
   * blk_queue_max_discard_sectors - set max sectors for a single discard
   * @q:  the request queue for the device
c7ebf0657   Randy Dunlap   blk-settings: fix...
274
   * @max_discard_sectors: maximum number of sectors to discard
67efc9258   Christoph Hellwig   block: allow larg...
275
276
277
278
279
280
281
282
283
   **/
  void blk_queue_max_discard_sectors(struct request_queue *q,
  		unsigned int max_discard_sectors)
  {
  	q->limits.max_discard_sectors = max_discard_sectors;
  }
  EXPORT_SYMBOL(blk_queue_max_discard_sectors);
  
  /**
8a78362c4   Martin K. Petersen   block: Consolidat...
284
   * blk_queue_max_segments - set max hw segments for a request for this queue
86db1e297   Jens Axboe   block: continue l...
285
286
287
288
289
   * @q:  the request queue for the device
   * @max_segments:  max number of segments
   *
   * Description:
   *    Enables a low level driver to set an upper limit on the number of
8a78362c4   Martin K. Petersen   block: Consolidat...
290
   *    hw data segments in a request.
86db1e297   Jens Axboe   block: continue l...
291
   **/
8a78362c4   Martin K. Petersen   block: Consolidat...
292
  void blk_queue_max_segments(struct request_queue *q, unsigned short max_segments)
86db1e297   Jens Axboe   block: continue l...
293
294
295
  {
  	if (!max_segments) {
  		max_segments = 1;
24c03d47d   Harvey Harrison   block: remove rem...
296
297
298
  		printk(KERN_INFO "%s: set to minimum %d
  ",
  		       __func__, max_segments);
86db1e297   Jens Axboe   block: continue l...
299
  	}
8a78362c4   Martin K. Petersen   block: Consolidat...
300
  	q->limits.max_segments = max_segments;
86db1e297   Jens Axboe   block: continue l...
301
  }
8a78362c4   Martin K. Petersen   block: Consolidat...
302
  EXPORT_SYMBOL(blk_queue_max_segments);
86db1e297   Jens Axboe   block: continue l...
303
304
305
306
307
308
309
310
311
312
313
314
315
316
  
  /**
   * blk_queue_max_segment_size - set max segment size for blk_rq_map_sg
   * @q:  the request queue for the device
   * @max_size:  max size of segment in bytes
   *
   * Description:
   *    Enables a low level driver to set an upper limit on the size of a
   *    coalesced segment
   **/
  void blk_queue_max_segment_size(struct request_queue *q, unsigned int max_size)
  {
  	if (max_size < PAGE_CACHE_SIZE) {
  		max_size = PAGE_CACHE_SIZE;
24c03d47d   Harvey Harrison   block: remove rem...
317
318
319
  		printk(KERN_INFO "%s: set to minimum %d
  ",
  		       __func__, max_size);
86db1e297   Jens Axboe   block: continue l...
320
  	}
025146e13   Martin K. Petersen   block: Move queue...
321
  	q->limits.max_segment_size = max_size;
86db1e297   Jens Axboe   block: continue l...
322
  }
86db1e297   Jens Axboe   block: continue l...
323
324
325
  EXPORT_SYMBOL(blk_queue_max_segment_size);
  
  /**
e1defc4ff   Martin K. Petersen   block: Do away wi...
326
   * blk_queue_logical_block_size - set logical block size for the queue
86db1e297   Jens Axboe   block: continue l...
327
   * @q:  the request queue for the device
e1defc4ff   Martin K. Petersen   block: Do away wi...
328
   * @size:  the logical block size, in bytes
86db1e297   Jens Axboe   block: continue l...
329
330
   *
   * Description:
e1defc4ff   Martin K. Petersen   block: Do away wi...
331
332
333
   *   This should be set to the lowest possible block size that the
   *   storage device can address.  The default of 512 covers most
   *   hardware.
86db1e297   Jens Axboe   block: continue l...
334
   **/
e1defc4ff   Martin K. Petersen   block: Do away wi...
335
  void blk_queue_logical_block_size(struct request_queue *q, unsigned short size)
86db1e297   Jens Axboe   block: continue l...
336
  {
025146e13   Martin K. Petersen   block: Move queue...
337
  	q->limits.logical_block_size = size;
c72758f33   Martin K. Petersen   block: Export I/O...
338
339
340
341
342
343
  
  	if (q->limits.physical_block_size < size)
  		q->limits.physical_block_size = size;
  
  	if (q->limits.io_min < q->limits.physical_block_size)
  		q->limits.io_min = q->limits.physical_block_size;
86db1e297   Jens Axboe   block: continue l...
344
  }
e1defc4ff   Martin K. Petersen   block: Do away wi...
345
  EXPORT_SYMBOL(blk_queue_logical_block_size);
86db1e297   Jens Axboe   block: continue l...
346

c72758f33   Martin K. Petersen   block: Export I/O...
347
348
349
350
351
352
353
354
355
356
  /**
   * blk_queue_physical_block_size - set physical block size for the queue
   * @q:  the request queue for the device
   * @size:  the physical block size, in bytes
   *
   * Description:
   *   This should be set to the lowest possible sector size that the
   *   hardware can operate on without reverting to read-modify-write
   *   operations.
   */
892b6f90d   Martin K. Petersen   block: Ensure phy...
357
  void blk_queue_physical_block_size(struct request_queue *q, unsigned int size)
c72758f33   Martin K. Petersen   block: Export I/O...
358
359
360
361
362
363
364
365
366
367
368
369
370
371
  {
  	q->limits.physical_block_size = size;
  
  	if (q->limits.physical_block_size < q->limits.logical_block_size)
  		q->limits.physical_block_size = q->limits.logical_block_size;
  
  	if (q->limits.io_min < q->limits.physical_block_size)
  		q->limits.io_min = q->limits.physical_block_size;
  }
  EXPORT_SYMBOL(blk_queue_physical_block_size);
  
  /**
   * blk_queue_alignment_offset - set physical block alignment offset
   * @q:	the request queue for the device
8ebf97560   Randy Dunlap   block: fix kernel...
372
   * @offset: alignment offset in bytes
c72758f33   Martin K. Petersen   block: Export I/O...
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
   *
   * Description:
   *   Some devices are naturally misaligned to compensate for things like
   *   the legacy DOS partition table 63-sector offset.  Low-level drivers
   *   should call this function for devices whose first sector is not
   *   naturally aligned.
   */
  void blk_queue_alignment_offset(struct request_queue *q, unsigned int offset)
  {
  	q->limits.alignment_offset =
  		offset & (q->limits.physical_block_size - 1);
  	q->limits.misaligned = 0;
  }
  EXPORT_SYMBOL(blk_queue_alignment_offset);
  
  /**
7c958e326   Martin K. Petersen   block: Add a wrap...
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
   * blk_limits_io_min - set minimum request size for a device
   * @limits: the queue limits
   * @min:  smallest I/O size in bytes
   *
   * Description:
   *   Some devices have an internal block size bigger than the reported
   *   hardware sector size.  This function can be used to signal the
   *   smallest I/O the device can perform without incurring a performance
   *   penalty.
   */
  void blk_limits_io_min(struct queue_limits *limits, unsigned int min)
  {
  	limits->io_min = min;
  
  	if (limits->io_min < limits->logical_block_size)
  		limits->io_min = limits->logical_block_size;
  
  	if (limits->io_min < limits->physical_block_size)
  		limits->io_min = limits->physical_block_size;
  }
  EXPORT_SYMBOL(blk_limits_io_min);
  
  /**
c72758f33   Martin K. Petersen   block: Export I/O...
412
413
   * blk_queue_io_min - set minimum request size for the queue
   * @q:	the request queue for the device
8ebf97560   Randy Dunlap   block: fix kernel...
414
   * @min:  smallest I/O size in bytes
c72758f33   Martin K. Petersen   block: Export I/O...
415
416
   *
   * Description:
7e5f5fb09   Martin K. Petersen   block: Update top...
417
418
419
420
421
422
423
   *   Storage devices may report a granularity or preferred minimum I/O
   *   size which is the smallest request the device can perform without
   *   incurring a performance penalty.  For disk drives this is often the
   *   physical block size.  For RAID arrays it is often the stripe chunk
   *   size.  A properly aligned multiple of minimum_io_size is the
   *   preferred request size for workloads where a high number of I/O
   *   operations is desired.
c72758f33   Martin K. Petersen   block: Export I/O...
424
425
426
   */
  void blk_queue_io_min(struct request_queue *q, unsigned int min)
  {
7c958e326   Martin K. Petersen   block: Add a wrap...
427
  	blk_limits_io_min(&q->limits, min);
c72758f33   Martin K. Petersen   block: Export I/O...
428
429
430
431
  }
  EXPORT_SYMBOL(blk_queue_io_min);
  
  /**
3c5820c74   Martin K. Petersen   block: Optimal I/...
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
   * blk_limits_io_opt - set optimal request size for a device
   * @limits: the queue limits
   * @opt:  smallest I/O size in bytes
   *
   * Description:
   *   Storage devices may report an optimal I/O size, which is the
   *   device's preferred unit for sustained I/O.  This is rarely reported
   *   for disk drives.  For RAID arrays it is usually the stripe width or
   *   the internal track size.  A properly aligned multiple of
   *   optimal_io_size is the preferred request size for workloads where
   *   sustained throughput is desired.
   */
  void blk_limits_io_opt(struct queue_limits *limits, unsigned int opt)
  {
  	limits->io_opt = opt;
  }
  EXPORT_SYMBOL(blk_limits_io_opt);
  
  /**
c72758f33   Martin K. Petersen   block: Export I/O...
451
452
   * blk_queue_io_opt - set optimal request size for the queue
   * @q:	the request queue for the device
8ebf97560   Randy Dunlap   block: fix kernel...
453
   * @opt:  optimal request size in bytes
c72758f33   Martin K. Petersen   block: Export I/O...
454
455
   *
   * Description:
7e5f5fb09   Martin K. Petersen   block: Update top...
456
457
458
459
460
461
   *   Storage devices may report an optimal I/O size, which is the
   *   device's preferred unit for sustained I/O.  This is rarely reported
   *   for disk drives.  For RAID arrays it is usually the stripe width or
   *   the internal track size.  A properly aligned multiple of
   *   optimal_io_size is the preferred request size for workloads where
   *   sustained throughput is desired.
c72758f33   Martin K. Petersen   block: Export I/O...
462
463
464
   */
  void blk_queue_io_opt(struct request_queue *q, unsigned int opt)
  {
3c5820c74   Martin K. Petersen   block: Optimal I/...
465
  	blk_limits_io_opt(&q->limits, opt);
c72758f33   Martin K. Petersen   block: Export I/O...
466
467
  }
  EXPORT_SYMBOL(blk_queue_io_opt);
86db1e297   Jens Axboe   block: continue l...
468
469
470
471
472
473
474
  /**
   * blk_queue_stack_limits - inherit underlying queue limits for stacked drivers
   * @t:	the stacking driver (top)
   * @b:  the underlying device (bottom)
   **/
  void blk_queue_stack_limits(struct request_queue *t, struct request_queue *b)
  {
fef246672   Martin K. Petersen   block: Make blk_q...
475
  	blk_stack_limits(&t->limits, &b->limits, 0);
86db1e297   Jens Axboe   block: continue l...
476
  }
86db1e297   Jens Axboe   block: continue l...
477
478
479
  EXPORT_SYMBOL(blk_queue_stack_limits);
  
  /**
c72758f33   Martin K. Petersen   block: Export I/O...
480
   * blk_stack_limits - adjust queue_limits for stacked devices
81744ee44   Martin K. Petersen   block: Fix incorr...
481
482
   * @t:	the stacking driver limits (top device)
   * @b:  the underlying queue limits (bottom, component device)
e03a72e13   Martin K. Petersen   block: Stop using...
483
   * @start:  first data sector within component device
c72758f33   Martin K. Petersen   block: Export I/O...
484
485
   *
   * Description:
81744ee44   Martin K. Petersen   block: Fix incorr...
486
487
488
489
490
491
492
493
494
495
496
497
498
   *    This function is used by stacking drivers like MD and DM to ensure
   *    that all component devices have compatible block sizes and
   *    alignments.  The stacking driver must provide a queue_limits
   *    struct (top) and then iteratively call the stacking function for
   *    all component (bottom) devices.  The stacking function will
   *    attempt to combine the values and ensure proper alignment.
   *
   *    Returns 0 if the top and bottom queue_limits are compatible.  The
   *    top device's block sizes and alignment offsets may be adjusted to
   *    ensure alignment with the bottom device. If no compatible sizes
   *    and alignments exist, -1 is returned and the resulting top
   *    queue_limits will have the misaligned flag set to indicate that
   *    the alignment_offset is undefined.
c72758f33   Martin K. Petersen   block: Export I/O...
499
500
   */
  int blk_stack_limits(struct queue_limits *t, struct queue_limits *b,
e03a72e13   Martin K. Petersen   block: Stop using...
501
  		     sector_t start)
c72758f33   Martin K. Petersen   block: Export I/O...
502
  {
e03a72e13   Martin K. Petersen   block: Stop using...
503
  	unsigned int top, bottom, alignment, ret = 0;
86b372814   Martin K. Petersen   block: Expose dis...
504

c72758f33   Martin K. Petersen   block: Export I/O...
505
506
  	t->max_sectors = min_not_zero(t->max_sectors, b->max_sectors);
  	t->max_hw_sectors = min_not_zero(t->max_hw_sectors, b->max_hw_sectors);
77634f33d   Martin K. Petersen   block: Add missin...
507
  	t->bounce_pfn = min_not_zero(t->bounce_pfn, b->bounce_pfn);
c72758f33   Martin K. Petersen   block: Export I/O...
508
509
510
  
  	t->seg_boundary_mask = min_not_zero(t->seg_boundary_mask,
  					    b->seg_boundary_mask);
8a78362c4   Martin K. Petersen   block: Consolidat...
511
  	t->max_segments = min_not_zero(t->max_segments, b->max_segments);
13f05c8d8   Martin K. Petersen   block/scsi: Provi...
512
513
  	t->max_integrity_segments = min_not_zero(t->max_integrity_segments,
  						 b->max_integrity_segments);
c72758f33   Martin K. Petersen   block: Export I/O...
514
515
516
  
  	t->max_segment_size = min_not_zero(t->max_segment_size,
  					   b->max_segment_size);
fe0b393f2   Martin K. Petersen   block: Correct ha...
517
  	t->misaligned |= b->misaligned;
e03a72e13   Martin K. Petersen   block: Stop using...
518
  	alignment = queue_limit_alignment_offset(b, start);
9504e0864   Martin K. Petersen   block: Fix topolo...
519

81744ee44   Martin K. Petersen   block: Fix incorr...
520
521
522
  	/* Bottom device has different alignment.  Check that it is
  	 * compatible with the current top alignment.
  	 */
9504e0864   Martin K. Petersen   block: Fix topolo...
523
524
525
526
  	if (t->alignment_offset != alignment) {
  
  		top = max(t->physical_block_size, t->io_min)
  			+ t->alignment_offset;
81744ee44   Martin K. Petersen   block: Fix incorr...
527
  		bottom = max(b->physical_block_size, b->io_min) + alignment;
9504e0864   Martin K. Petersen   block: Fix topolo...
528

81744ee44   Martin K. Petersen   block: Fix incorr...
529
  		/* Verify that top and bottom intervals line up */
fe0b393f2   Martin K. Petersen   block: Correct ha...
530
  		if (max(top, bottom) & (min(top, bottom) - 1)) {
9504e0864   Martin K. Petersen   block: Fix topolo...
531
  			t->misaligned = 1;
fe0b393f2   Martin K. Petersen   block: Correct ha...
532
533
  			ret = -1;
  		}
9504e0864   Martin K. Petersen   block: Fix topolo...
534
  	}
c72758f33   Martin K. Petersen   block: Export I/O...
535
536
537
538
539
540
541
  	t->logical_block_size = max(t->logical_block_size,
  				    b->logical_block_size);
  
  	t->physical_block_size = max(t->physical_block_size,
  				     b->physical_block_size);
  
  	t->io_min = max(t->io_min, b->io_min);
9504e0864   Martin K. Petersen   block: Fix topolo...
542
  	t->io_opt = lcm(t->io_opt, b->io_opt);
e692cb668   Martin K. Petersen   block: Deprecate ...
543
  	t->cluster &= b->cluster;
98262f276   Martin K. Petersen   block: Allow devi...
544
  	t->discard_zeroes_data &= b->discard_zeroes_data;
c72758f33   Martin K. Petersen   block: Export I/O...
545

81744ee44   Martin K. Petersen   block: Fix incorr...
546
  	/* Physical block size a multiple of the logical block size? */
9504e0864   Martin K. Petersen   block: Fix topolo...
547
548
  	if (t->physical_block_size & (t->logical_block_size - 1)) {
  		t->physical_block_size = t->logical_block_size;
c72758f33   Martin K. Petersen   block: Export I/O...
549
  		t->misaligned = 1;
fe0b393f2   Martin K. Petersen   block: Correct ha...
550
  		ret = -1;
86b372814   Martin K. Petersen   block: Expose dis...
551
  	}
81744ee44   Martin K. Petersen   block: Fix incorr...
552
  	/* Minimum I/O a multiple of the physical block size? */
9504e0864   Martin K. Petersen   block: Fix topolo...
553
554
555
  	if (t->io_min & (t->physical_block_size - 1)) {
  		t->io_min = t->physical_block_size;
  		t->misaligned = 1;
fe0b393f2   Martin K. Petersen   block: Correct ha...
556
  		ret = -1;
c72758f33   Martin K. Petersen   block: Export I/O...
557
  	}
81744ee44   Martin K. Petersen   block: Fix incorr...
558
  	/* Optimal I/O a multiple of the physical block size? */
9504e0864   Martin K. Petersen   block: Fix topolo...
559
560
561
  	if (t->io_opt & (t->physical_block_size - 1)) {
  		t->io_opt = 0;
  		t->misaligned = 1;
fe0b393f2   Martin K. Petersen   block: Correct ha...
562
  		ret = -1;
9504e0864   Martin K. Petersen   block: Fix topolo...
563
  	}
c72758f33   Martin K. Petersen   block: Export I/O...
564

81744ee44   Martin K. Petersen   block: Fix incorr...
565
  	/* Find lowest common alignment_offset */
9504e0864   Martin K. Petersen   block: Fix topolo...
566
567
  	t->alignment_offset = lcm(t->alignment_offset, alignment)
  		& (max(t->physical_block_size, t->io_min) - 1);
86b372814   Martin K. Petersen   block: Expose dis...
568

81744ee44   Martin K. Petersen   block: Fix incorr...
569
  	/* Verify that new alignment_offset is on a logical block boundary */
fe0b393f2   Martin K. Petersen   block: Correct ha...
570
  	if (t->alignment_offset & (t->logical_block_size - 1)) {
c72758f33   Martin K. Petersen   block: Export I/O...
571
  		t->misaligned = 1;
fe0b393f2   Martin K. Petersen   block: Correct ha...
572
573
  		ret = -1;
  	}
c72758f33   Martin K. Petersen   block: Export I/O...
574

9504e0864   Martin K. Petersen   block: Fix topolo...
575
576
  	/* Discard alignment and granularity */
  	if (b->discard_granularity) {
e03a72e13   Martin K. Petersen   block: Stop using...
577
  		alignment = queue_limit_discard_alignment(b, start);
9504e0864   Martin K. Petersen   block: Fix topolo...
578
579
580
581
582
  
  		if (t->discard_granularity != 0 &&
  		    t->discard_alignment != alignment) {
  			top = t->discard_granularity + t->discard_alignment;
  			bottom = b->discard_granularity + alignment;
70dd5bf3b   Martin K. Petersen   block: Stack opti...
583

9504e0864   Martin K. Petersen   block: Fix topolo...
584
585
586
587
  			/* Verify that top and bottom intervals line up */
  			if (max(top, bottom) & (min(top, bottom) - 1))
  				t->discard_misaligned = 1;
  		}
81744ee44   Martin K. Petersen   block: Fix incorr...
588
589
  		t->max_discard_sectors = min_not_zero(t->max_discard_sectors,
  						      b->max_discard_sectors);
9504e0864   Martin K. Petersen   block: Fix topolo...
590
591
592
593
594
  		t->discard_granularity = max(t->discard_granularity,
  					     b->discard_granularity);
  		t->discard_alignment = lcm(t->discard_alignment, alignment) &
  			(t->discard_granularity - 1);
  	}
70dd5bf3b   Martin K. Petersen   block: Stack opti...
595

fe0b393f2   Martin K. Petersen   block: Correct ha...
596
  	return ret;
c72758f33   Martin K. Petersen   block: Export I/O...
597
  }
5d85d3247   Mike Snitzer   block: export blk...
598
  EXPORT_SYMBOL(blk_stack_limits);
c72758f33   Martin K. Petersen   block: Export I/O...
599
600
  
  /**
17be8c245   Martin K. Petersen   block: bdev_stack...
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
   * bdev_stack_limits - adjust queue limits for stacked drivers
   * @t:	the stacking driver limits (top device)
   * @bdev:  the component block_device (bottom)
   * @start:  first data sector within component device
   *
   * Description:
   *    Merges queue limits for a top device and a block_device.  Returns
   *    0 if alignment didn't change.  Returns -1 if adding the bottom
   *    device caused misalignment.
   */
  int bdev_stack_limits(struct queue_limits *t, struct block_device *bdev,
  		      sector_t start)
  {
  	struct request_queue *bq = bdev_get_queue(bdev);
  
  	start += get_start_sect(bdev);
e03a72e13   Martin K. Petersen   block: Stop using...
617
  	return blk_stack_limits(t, &bq->limits, start);
17be8c245   Martin K. Petersen   block: bdev_stack...
618
619
620
621
  }
  EXPORT_SYMBOL(bdev_stack_limits);
  
  /**
c72758f33   Martin K. Petersen   block: Export I/O...
622
   * disk_stack_limits - adjust queue limits for stacked drivers
77634f33d   Martin K. Petersen   block: Add missin...
623
   * @disk:  MD/DM gendisk (top)
c72758f33   Martin K. Petersen   block: Export I/O...
624
625
626
627
   * @bdev:  the underlying block device (bottom)
   * @offset:  offset to beginning of data within component device
   *
   * Description:
e03a72e13   Martin K. Petersen   block: Stop using...
628
629
   *    Merges the limits for a top level gendisk and a bottom level
   *    block_device.
c72758f33   Martin K. Petersen   block: Export I/O...
630
631
632
633
634
   */
  void disk_stack_limits(struct gendisk *disk, struct block_device *bdev,
  		       sector_t offset)
  {
  	struct request_queue *t = disk->queue;
c72758f33   Martin K. Petersen   block: Export I/O...
635

e03a72e13   Martin K. Petersen   block: Stop using...
636
  	if (bdev_stack_limits(&t->limits, bdev, offset >> 9) < 0) {
c72758f33   Martin K. Petersen   block: Export I/O...
637
638
639
640
641
642
643
644
645
  		char top[BDEVNAME_SIZE], bottom[BDEVNAME_SIZE];
  
  		disk_name(disk, 0, top);
  		bdevname(bdev, bottom);
  
  		printk(KERN_NOTICE "%s: Warning: Device %s is misaligned
  ",
  		       top, bottom);
  	}
c72758f33   Martin K. Petersen   block: Export I/O...
646
647
648
649
  }
  EXPORT_SYMBOL(disk_stack_limits);
  
  /**
e3790c7d4   Tejun Heo   block: separate o...
650
651
652
653
   * blk_queue_dma_pad - set pad mask
   * @q:     the request queue for the device
   * @mask:  pad mask
   *
27f8221af   FUJITA Tomonori   block: add blk_qu...
654
   * Set dma pad mask.
e3790c7d4   Tejun Heo   block: separate o...
655
   *
27f8221af   FUJITA Tomonori   block: add blk_qu...
656
657
   * Appending pad buffer to a request modifies the last entry of a
   * scatter list such that it includes the pad buffer.
e3790c7d4   Tejun Heo   block: separate o...
658
659
660
661
662
663
664
665
   **/
  void blk_queue_dma_pad(struct request_queue *q, unsigned int mask)
  {
  	q->dma_pad_mask = mask;
  }
  EXPORT_SYMBOL(blk_queue_dma_pad);
  
  /**
27f8221af   FUJITA Tomonori   block: add blk_qu...
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
   * blk_queue_update_dma_pad - update pad mask
   * @q:     the request queue for the device
   * @mask:  pad mask
   *
   * Update dma pad mask.
   *
   * Appending pad buffer to a request modifies the last entry of a
   * scatter list such that it includes the pad buffer.
   **/
  void blk_queue_update_dma_pad(struct request_queue *q, unsigned int mask)
  {
  	if (mask > q->dma_pad_mask)
  		q->dma_pad_mask = mask;
  }
  EXPORT_SYMBOL(blk_queue_update_dma_pad);
  
  /**
86db1e297   Jens Axboe   block: continue l...
683
   * blk_queue_dma_drain - Set up a drain buffer for excess dma.
86db1e297   Jens Axboe   block: continue l...
684
   * @q:  the request queue for the device
2fb98e841   Tejun Heo   block: implement ...
685
   * @dma_drain_needed: fn which returns non-zero if drain is necessary
86db1e297   Jens Axboe   block: continue l...
686
687
688
689
690
691
692
693
694
695
696
697
   * @buf:	physically contiguous buffer
   * @size:	size of the buffer in bytes
   *
   * Some devices have excess DMA problems and can't simply discard (or
   * zero fill) the unwanted piece of the transfer.  They have to have a
   * real area of memory to transfer it into.  The use case for this is
   * ATAPI devices in DMA mode.  If the packet command causes a transfer
   * bigger than the transfer size some HBAs will lock up if there
   * aren't DMA elements to contain the excess transfer.  What this API
   * does is adjust the queue so that the buf is always appended
   * silently to the scatterlist.
   *
8a78362c4   Martin K. Petersen   block: Consolidat...
698
699
700
701
   * Note: This routine adjusts max_hw_segments to make room for appending
   * the drain buffer.  If you call blk_queue_max_segments() after calling
   * this routine, you must set the limit to one fewer than your device
   * can support otherwise there won't be room for the drain buffer.
86db1e297   Jens Axboe   block: continue l...
702
   */
448da4d26   Harvey Harrison   block: remove ext...
703
  int blk_queue_dma_drain(struct request_queue *q,
2fb98e841   Tejun Heo   block: implement ...
704
705
  			       dma_drain_needed_fn *dma_drain_needed,
  			       void *buf, unsigned int size)
86db1e297   Jens Axboe   block: continue l...
706
  {
8a78362c4   Martin K. Petersen   block: Consolidat...
707
  	if (queue_max_segments(q) < 2)
86db1e297   Jens Axboe   block: continue l...
708
709
  		return -EINVAL;
  	/* make room for appending the drain */
8a78362c4   Martin K. Petersen   block: Consolidat...
710
  	blk_queue_max_segments(q, queue_max_segments(q) - 1);
2fb98e841   Tejun Heo   block: implement ...
711
  	q->dma_drain_needed = dma_drain_needed;
86db1e297   Jens Axboe   block: continue l...
712
713
714
715
716
  	q->dma_drain_buffer = buf;
  	q->dma_drain_size = size;
  
  	return 0;
  }
86db1e297   Jens Axboe   block: continue l...
717
718
719
720
721
722
723
724
725
726
727
  EXPORT_SYMBOL_GPL(blk_queue_dma_drain);
  
  /**
   * blk_queue_segment_boundary - set boundary rules for segment merging
   * @q:  the request queue for the device
   * @mask:  the memory boundary mask
   **/
  void blk_queue_segment_boundary(struct request_queue *q, unsigned long mask)
  {
  	if (mask < PAGE_CACHE_SIZE - 1) {
  		mask = PAGE_CACHE_SIZE - 1;
24c03d47d   Harvey Harrison   block: remove rem...
728
729
730
  		printk(KERN_INFO "%s: set to minimum %lx
  ",
  		       __func__, mask);
86db1e297   Jens Axboe   block: continue l...
731
  	}
025146e13   Martin K. Petersen   block: Move queue...
732
  	q->limits.seg_boundary_mask = mask;
86db1e297   Jens Axboe   block: continue l...
733
  }
86db1e297   Jens Axboe   block: continue l...
734
735
736
737
738
739
740
741
  EXPORT_SYMBOL(blk_queue_segment_boundary);
  
  /**
   * blk_queue_dma_alignment - set dma length and memory alignment
   * @q:     the request queue for the device
   * @mask:  alignment mask
   *
   * description:
710027a48   Randy Dunlap   Add some block/ s...
742
   *    set required memory and length alignment for direct dma transactions.
8feb4d20b   Alan Cox   pata_artop: typo
743
   *    this is used when building direct io requests for the queue.
86db1e297   Jens Axboe   block: continue l...
744
745
746
747
748
749
   *
   **/
  void blk_queue_dma_alignment(struct request_queue *q, int mask)
  {
  	q->dma_alignment = mask;
  }
86db1e297   Jens Axboe   block: continue l...
750
751
752
753
754
755
756
757
  EXPORT_SYMBOL(blk_queue_dma_alignment);
  
  /**
   * blk_queue_update_dma_alignment - update dma length and memory alignment
   * @q:     the request queue for the device
   * @mask:  alignment mask
   *
   * description:
710027a48   Randy Dunlap   Add some block/ s...
758
   *    update required memory and length alignment for direct dma transactions.
86db1e297   Jens Axboe   block: continue l...
759
760
761
762
763
764
765
766
767
768
769
770
771
772
   *    If the requested alignment is larger than the current alignment, then
   *    the current queue alignment is updated to the new value, otherwise it
   *    is left alone.  The design of this is to allow multiple objects
   *    (driver, device, transport etc) to set their respective
   *    alignments without having them interfere.
   *
   **/
  void blk_queue_update_dma_alignment(struct request_queue *q, int mask)
  {
  	BUG_ON(mask > PAGE_SIZE);
  
  	if (mask > q->dma_alignment)
  		q->dma_alignment = mask;
  }
86db1e297   Jens Axboe   block: continue l...
773
  EXPORT_SYMBOL(blk_queue_update_dma_alignment);
4913efe45   Tejun Heo   block: deprecate ...
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
  /**
   * blk_queue_flush - configure queue's cache flush capability
   * @q:		the request queue for the device
   * @flush:	0, REQ_FLUSH or REQ_FLUSH | REQ_FUA
   *
   * Tell block layer cache flush capability of @q.  If it supports
   * flushing, REQ_FLUSH should be set.  If it supports bypassing
   * write cache for individual writes, REQ_FUA should be set.
   */
  void blk_queue_flush(struct request_queue *q, unsigned int flush)
  {
  	WARN_ON_ONCE(flush & ~(REQ_FLUSH | REQ_FUA));
  
  	if (WARN_ON_ONCE(!(flush & REQ_FLUSH) && (flush & REQ_FUA)))
  		flush &= ~REQ_FUA;
  
  	q->flush_flags = flush & (REQ_FLUSH | REQ_FUA);
  }
  EXPORT_SYMBOL_GPL(blk_queue_flush);
f38769309   shaohua.li@intel.com   block: add a non-...
793
794
795
796
797
  void blk_queue_flush_queueable(struct request_queue *q, bool queueable)
  {
  	q->flush_not_queueable = !queueable;
  }
  EXPORT_SYMBOL_GPL(blk_queue_flush_queueable);
aeb3d3a81   Harvey Harrison   block: kmalloc ar...
798
  static int __init blk_settings_init(void)
86db1e297   Jens Axboe   block: continue l...
799
800
801
802
803
804
  {
  	blk_max_low_pfn = max_low_pfn - 1;
  	blk_max_pfn = max_pfn - 1;
  	return 0;
  }
  subsys_initcall(blk_settings_init);