blk-merge.c 10.7 KB
Newer Older
1 2 3 4 5 6 7 8 9 10 11
/*
 * Functions related to segment and merge handling
 */
#include <linux/kernel.h>
#include <linux/module.h>
#include <linux/bio.h>
#include <linux/blkdev.h>
#include <linux/scatterlist.h>

#include "blk.h"

12
static unsigned int __blk_recalc_rq_segments(struct request_queue *q,
13
					     struct bio *bio)
14 15 16
{
	unsigned int phys_size;
	struct bio_vec *bv, *bvprv = NULL;
17 18
	int cluster, i, high, highprv = 1;
	unsigned int seg_size, nr_phys_segs;
19
	struct bio *fbio, *bbio;
20

21 22
	if (!bio)
		return 0;
23

24
	fbio = bio;
25
	cluster = test_bit(QUEUE_FLAG_CLUSTER, &q->queue_flags);
Mikulas Patocka's avatar
Mikulas Patocka committed
26 27
	seg_size = 0;
	phys_size = nr_phys_segs = 0;
28 29 30 31 32 33 34
	for_each_bio(bio) {
		bio_for_each_segment(bv, bio, i) {
			/*
			 * the trick here is making sure that a high page is
			 * never considered part of another segment, since that
			 * might change with the bounce page.
			 */
35
			high = page_to_pfn(bv->bv_page) > queue_bounce_pfn(q);
36
			if (high || highprv)
37
				goto new_segment;
38
			if (cluster) {
39 40
				if (seg_size + bv->bv_len
				    > queue_max_segment_size(q))
41 42 43 44 45
					goto new_segment;
				if (!BIOVEC_PHYS_MERGEABLE(bvprv, bv))
					goto new_segment;
				if (!BIOVEC_SEG_BOUNDARY(q, bvprv, bv))
					goto new_segment;
46

47 48 49 50
				seg_size += bv->bv_len;
				bvprv = bv;
				continue;
			}
51
new_segment:
52 53 54
			if (nr_phys_segs == 1 && seg_size >
			    fbio->bi_seg_front_size)
				fbio->bi_seg_front_size = seg_size;
55

56 57 58 59 60
			nr_phys_segs++;
			bvprv = bv;
			seg_size = bv->bv_len;
			highprv = high;
		}
61
		bbio = bio;
62 63
	}

64 65 66 67
	if (nr_phys_segs == 1 && seg_size > fbio->bi_seg_front_size)
		fbio->bi_seg_front_size = seg_size;
	if (seg_size > bbio->bi_seg_back_size)
		bbio->bi_seg_back_size = seg_size;
68 69 70 71 72 73

	return nr_phys_segs;
}

void blk_recalc_rq_segments(struct request *rq)
{
74
	rq->nr_phys_segments = __blk_recalc_rq_segments(rq->q, rq->bio);
75 76 77 78 79
}

void blk_recount_segments(struct request_queue *q, struct bio *bio)
{
	struct bio *nxt = bio->bi_next;
80

81
	bio->bi_next = NULL;
82
	bio->bi_phys_segments = __blk_recalc_rq_segments(q, bio);
83 84 85 86 87 88 89 90
	bio->bi_next = nxt;
	bio->bi_flags |= (1 << BIO_SEG_VALID);
}
EXPORT_SYMBOL(blk_recount_segments);

static int blk_phys_contig_segment(struct request_queue *q, struct bio *bio,
				   struct bio *nxt)
{
91
	if (!test_bit(QUEUE_FLAG_CLUSTER, &q->queue_flags))
92 93
		return 0;

94
	if (bio->bi_seg_back_size + nxt->bi_seg_front_size >
95
	    queue_max_segment_size(q))
96 97
		return 0;

98 99 100 101 102 103
	if (!bio_has_data(bio))
		return 1;

	if (!BIOVEC_PHYS_MERGEABLE(__BVEC_END(bio), __BVEC_START(nxt)))
		return 0;

104
	/*
105
	 * bio and nxt are contiguous in memory; check if the queue allows
106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126
	 * these two to be merged into one
	 */
	if (BIO_SEG_BOUNDARY(q, bio, nxt))
		return 1;

	return 0;
}

/*
 * map a request to scatterlist, return number of sg entries setup. Caller
 * must make sure sg can hold rq->nr_phys_segments entries
 */
int blk_rq_map_sg(struct request_queue *q, struct request *rq,
		  struct scatterlist *sglist)
{
	struct bio_vec *bvec, *bvprv;
	struct req_iterator iter;
	struct scatterlist *sg;
	int nsegs, cluster;

	nsegs = 0;
127
	cluster = test_bit(QUEUE_FLAG_CLUSTER, &q->queue_flags);
128 129 130 131 132 133 134 135 136 137

	/*
	 * for each bio in rq
	 */
	bvprv = NULL;
	sg = NULL;
	rq_for_each_segment(bvec, rq, iter) {
		int nbytes = bvec->bv_len;

		if (bvprv && cluster) {
138
			if (sg->length + nbytes > queue_max_segment_size(q))
139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171
				goto new_segment;

			if (!BIOVEC_PHYS_MERGEABLE(bvprv, bvec))
				goto new_segment;
			if (!BIOVEC_SEG_BOUNDARY(q, bvprv, bvec))
				goto new_segment;

			sg->length += nbytes;
		} else {
new_segment:
			if (!sg)
				sg = sglist;
			else {
				/*
				 * If the driver previously mapped a shorter
				 * list, we could see a termination bit
				 * prematurely unless it fully inits the sg
				 * table on each mapping. We KNOW that there
				 * must be more entries here or the driver
				 * would be buggy, so force clear the
				 * termination bit to avoid doing a full
				 * sg_init_table() in drivers for each command.
				 */
				sg->page_link &= ~0x02;
				sg = sg_next(sg);
			}

			sg_set_page(sg, bvec->bv_page, nbytes, bvec->bv_offset);
			nsegs++;
		}
		bvprv = bvec;
	} /* segments in rq */

172 173

	if (unlikely(rq->cmd_flags & REQ_COPY_USER) &&
174 175 176
	    (blk_rq_bytes(rq) & q->dma_pad_mask)) {
		unsigned int pad_len =
			(q->dma_pad_mask & ~blk_rq_bytes(rq)) + 1;
177 178 179 180 181

		sg->length += pad_len;
		rq->extra_len += pad_len;
	}

182
	if (q->dma_drain_size && q->dma_drain_needed(rq)) {
183 184 185
		if (rq->cmd_flags & REQ_RW)
			memset(q->dma_drain_buffer, 0, q->dma_drain_size);

186 187 188 189 190 191 192
		sg->page_link &= ~0x02;
		sg = sg_next(sg);
		sg_set_page(sg, virt_to_page(q->dma_drain_buffer),
			    q->dma_drain_size,
			    ((unsigned long)q->dma_drain_buffer) &
			    (PAGE_SIZE - 1));
		nsegs++;
193
		rq->extra_len += q->dma_drain_size;
194 195 196 197 198 199 200 201 202 203 204 205 206 207 208
	}

	if (sg)
		sg_mark_end(sg);

	return nsegs;
}
EXPORT_SYMBOL(blk_rq_map_sg);

static inline int ll_new_hw_segment(struct request_queue *q,
				    struct request *req,
				    struct bio *bio)
{
	int nr_phys_segs = bio_phys_segments(q, bio);

209
	if (req->nr_phys_segments + nr_phys_segs > queue_max_segments(q)) {
210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228
		req->cmd_flags |= REQ_NOMERGE;
		if (req == q->last_merge)
			q->last_merge = NULL;
		return 0;
	}

	/*
	 * This will form the start of a new hw segment.  Bump both
	 * counters.
	 */
	req->nr_phys_segments += nr_phys_segs;
	return 1;
}

int ll_back_merge_fn(struct request_queue *q, struct request *req,
		     struct bio *bio)
{
	unsigned short max_sectors;

229
	if (unlikely(req->cmd_type == REQ_TYPE_BLOCK_PC))
230
		max_sectors = queue_max_hw_sectors(q);
231
	else
232
		max_sectors = queue_max_sectors(q);
233

234
	if (blk_rq_sectors(req) + bio_sectors(bio) > max_sectors) {
235 236 237 238 239
		req->cmd_flags |= REQ_NOMERGE;
		if (req == q->last_merge)
			q->last_merge = NULL;
		return 0;
	}
240
	if (!bio_flagged(req->biotail, BIO_SEG_VALID))
241
		blk_recount_segments(q, req->biotail);
242
	if (!bio_flagged(bio, BIO_SEG_VALID))
243 244 245 246 247
		blk_recount_segments(q, bio);

	return ll_new_hw_segment(q, req, bio);
}

248
int ll_front_merge_fn(struct request_queue *q, struct request *req,
249 250 251 252
		      struct bio *bio)
{
	unsigned short max_sectors;

253
	if (unlikely(req->cmd_type == REQ_TYPE_BLOCK_PC))
254
		max_sectors = queue_max_hw_sectors(q);
255
	else
256
		max_sectors = queue_max_sectors(q);
257 258


259
	if (blk_rq_sectors(req) + bio_sectors(bio) > max_sectors) {
260 261 262 263 264
		req->cmd_flags |= REQ_NOMERGE;
		if (req == q->last_merge)
			q->last_merge = NULL;
		return 0;
	}
265
	if (!bio_flagged(bio, BIO_SEG_VALID))
266
		blk_recount_segments(q, bio);
267
	if (!bio_flagged(req->bio, BIO_SEG_VALID))
268 269 270 271 272 273 274 275 276
		blk_recount_segments(q, req->bio);

	return ll_new_hw_segment(q, req, bio);
}

static int ll_merge_requests_fn(struct request_queue *q, struct request *req,
				struct request *next)
{
	int total_phys_segments;
277 278
	unsigned int seg_size =
		req->biotail->bi_seg_back_size + next->bio->bi_seg_front_size;
279 280 281 282 283 284 285 286 287 288 289

	/*
	 * First check if the either of the requests are re-queued
	 * requests.  Can't merge them if they are.
	 */
	if (req->special || next->special)
		return 0;

	/*
	 * Will it become too large?
	 */
290
	if ((blk_rq_sectors(req) + blk_rq_sectors(next)) > queue_max_sectors(q))
291 292 293
		return 0;

	total_phys_segments = req->nr_phys_segments + next->nr_phys_segments;
294 295 296 297 298
	if (blk_phys_contig_segment(q, req->biotail, next->bio)) {
		if (req->nr_phys_segments == 1)
			req->bio->bi_seg_front_size = seg_size;
		if (next->nr_phys_segments == 1)
			next->biotail->bi_seg_back_size = seg_size;
299
		total_phys_segments--;
300
	}
301

302
	if (total_phys_segments > queue_max_segments(q))
303 304 305 306 307 308 309
		return 0;

	/* Merge is OK... */
	req->nr_phys_segments = total_phys_segments;
	return 1;
}

310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339
/**
 * blk_rq_set_mixed_merge - mark a request as mixed merge
 * @rq: request to mark as mixed merge
 *
 * Description:
 *     @rq is about to be mixed merged.  Make sure the attributes
 *     which can be mixed are set in each bio and mark @rq as mixed
 *     merged.
 */
void blk_rq_set_mixed_merge(struct request *rq)
{
	unsigned int ff = rq->cmd_flags & REQ_FAILFAST_MASK;
	struct bio *bio;

	if (rq->cmd_flags & REQ_MIXED_MERGE)
		return;

	/*
	 * @rq will no longer represent mixable attributes for all the
	 * contained bios.  It will just track those of the first one.
	 * Distributes the attributs to each bio.
	 */
	for (bio = rq->bio; bio; bio = bio->bi_next) {
		WARN_ON_ONCE((bio->bi_rw & REQ_FAILFAST_MASK) &&
			     (bio->bi_rw & REQ_FAILFAST_MASK) != ff);
		bio->bi_rw |= ff;
	}
	rq->cmd_flags |= REQ_MIXED_MERGE;
}

340 341 342 343 344 345 346
static void blk_account_io_merge(struct request *req)
{
	if (blk_do_io_stat(req)) {
		struct hd_struct *part;
		int cpu;

		cpu = part_stat_lock();
347
		part = disk_map_sector_rcu(req->rq_disk, blk_rq_pos(req));
348 349

		part_round_stats(cpu, part);
350
		part_dec_in_flight(part, rq_data_dir(req));
351 352 353 354 355

		part_stat_unlock();
	}
}

356 357 358 359 360 361 362 363 364 365 366 367
/*
 * Has to be called with the request spinlock acquired
 */
static int attempt_merge(struct request_queue *q, struct request *req,
			  struct request *next)
{
	if (!rq_mergeable(req) || !rq_mergeable(next))
		return 0;

	/*
	 * not contiguous
	 */
368
	if (blk_rq_pos(req) + blk_rq_sectors(req) != blk_rq_pos(next))
369 370 371 372 373 374 375
		return 0;

	if (rq_data_dir(req) != rq_data_dir(next)
	    || req->rq_disk != next->rq_disk
	    || next->special)
		return 0;

376 377 378
	if (blk_integrity_rq(req) != blk_integrity_rq(next))
		return 0;

379 380 381 382 383 384 385 386 387
	/*
	 * If we are allowed to merge, then append bio list
	 * from next to rq and release next. merge_requests_fn
	 * will have updated segment counts, update sector
	 * counts here.
	 */
	if (!ll_merge_requests_fn(q, req, next))
		return 0;

388 389 390 391 392 393 394 395 396 397 398 399 400
	/*
	 * If failfast settings disagree or any of the two is already
	 * a mixed merge, mark both as mixed before proceeding.  This
	 * makes sure that all involved bios have mixable attributes
	 * set properly.
	 */
	if ((req->cmd_flags | next->cmd_flags) & REQ_MIXED_MERGE ||
	    (req->cmd_flags & REQ_FAILFAST_MASK) !=
	    (next->cmd_flags & REQ_FAILFAST_MASK)) {
		blk_rq_set_mixed_merge(req);
		blk_rq_set_mixed_merge(next);
	}

401 402 403 404 405 406 407 408 409 410 411 412
	/*
	 * At this point we have either done a back merge
	 * or front merge. We need the smaller start_time of
	 * the merged requests to be the current request
	 * for accounting purposes.
	 */
	if (time_after(req->start_time, next->start_time))
		req->start_time = next->start_time;

	req->biotail->bi_next = next->bio;
	req->biotail = next->biotail;

413
	req->__data_len += blk_rq_bytes(next);
414 415 416

	elv_merge_requests(q, req, next);

417 418 419 420
	/*
	 * 'next' is going away, so update stats accordingly
	 */
	blk_account_io_merge(next);
421 422

	req->ioprio = ioprio_best(req->ioprio, next->ioprio);
423 424
	if (blk_rq_cpu_valid(next))
		req->cpu = next->cpu;
425

426 427
	/* owner-ship of bio passed from next to req */
	next->bio = NULL;
428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450
	__blk_put_request(q, next);
	return 1;
}

int attempt_back_merge(struct request_queue *q, struct request *rq)
{
	struct request *next = elv_latter_request(q, rq);

	if (next)
		return attempt_merge(q, rq, next);

	return 0;
}

int attempt_front_merge(struct request_queue *q, struct request *rq)
{
	struct request *prev = elv_former_request(q, rq);

	if (prev)
		return attempt_merge(q, prev, rq);

	return 0;
}