1 /*
2 * Copyright (C) 2016 CNEX Labs
3 * Initial release: Javier Gonzalez <javier@cnexlabs.com>
4 *
5 * Based upon the circular ringbuffer.
6 *
7 * This program is free software; you can redistribute it and/or
8 * modify it under the terms of the GNU General Public License version
9 * 2 as published by the Free Software Foundation.
10 *
11 * This program is distributed in the hope that it will be useful, but
12 * WITHOUT ANY WARRANTY; without even the implied warranty of
13 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
14 * General Public License for more details.
15 *
16 * pblk-rb.c - pblk's write buffer
17 */
18
19 #include <linux/circ_buf.h>
20
21 #include "pblk.h"
22
23 static DECLARE_RWSEM(pblk_rb_lock);
24
pblk_rb_data_free(struct pblk_rb * rb)25 void pblk_rb_data_free(struct pblk_rb *rb)
26 {
27 struct pblk_rb_pages *p, *t;
28
29 down_write(&pblk_rb_lock);
30 list_for_each_entry_safe(p, t, &rb->pages, list) {
31 free_pages((unsigned long)page_address(p->pages), p->order);
32 list_del(&p->list);
33 kfree(p);
34 }
35 up_write(&pblk_rb_lock);
36 }
37
38 /*
39 * Initialize ring buffer. The data and metadata buffers must be previously
40 * allocated and their size must be a power of two
41 * (Documentation/core-api/circular-buffers.rst)
42 */
pblk_rb_init(struct pblk_rb * rb,struct pblk_rb_entry * rb_entry_base,unsigned int power_size,unsigned int power_seg_sz)43 int pblk_rb_init(struct pblk_rb *rb, struct pblk_rb_entry *rb_entry_base,
44 unsigned int power_size, unsigned int power_seg_sz)
45 {
46 struct pblk *pblk = container_of(rb, struct pblk, rwb);
47 unsigned int init_entry = 0;
48 unsigned int alloc_order = power_size;
49 unsigned int max_order = MAX_ORDER - 1;
50 unsigned int order, iter;
51
52 down_write(&pblk_rb_lock);
53 rb->entries = rb_entry_base;
54 rb->seg_size = (1 << power_seg_sz);
55 rb->nr_entries = (1 << power_size);
56 rb->mem = rb->subm = rb->sync = rb->l2p_update = 0;
57 rb->flush_point = EMPTY_ENTRY;
58
59 spin_lock_init(&rb->w_lock);
60 spin_lock_init(&rb->s_lock);
61
62 INIT_LIST_HEAD(&rb->pages);
63
64 if (alloc_order >= max_order) {
65 order = max_order;
66 iter = (1 << (alloc_order - max_order));
67 } else {
68 order = alloc_order;
69 iter = 1;
70 }
71
72 do {
73 struct pblk_rb_entry *entry;
74 struct pblk_rb_pages *page_set;
75 void *kaddr;
76 unsigned long set_size;
77 int i;
78
79 page_set = kmalloc(sizeof(struct pblk_rb_pages), GFP_KERNEL);
80 if (!page_set) {
81 up_write(&pblk_rb_lock);
82 return -ENOMEM;
83 }
84
85 page_set->order = order;
86 page_set->pages = alloc_pages(GFP_KERNEL, order);
87 if (!page_set->pages) {
88 kfree(page_set);
89 pblk_rb_data_free(rb);
90 up_write(&pblk_rb_lock);
91 return -ENOMEM;
92 }
93 kaddr = page_address(page_set->pages);
94
95 entry = &rb->entries[init_entry];
96 entry->data = kaddr;
97 entry->cacheline = pblk_cacheline_to_addr(init_entry++);
98 entry->w_ctx.flags = PBLK_WRITABLE_ENTRY;
99
100 set_size = (1 << order);
101 for (i = 1; i < set_size; i++) {
102 entry = &rb->entries[init_entry];
103 entry->cacheline = pblk_cacheline_to_addr(init_entry++);
104 entry->data = kaddr + (i * rb->seg_size);
105 entry->w_ctx.flags = PBLK_WRITABLE_ENTRY;
106 bio_list_init(&entry->w_ctx.bios);
107 }
108
109 list_add_tail(&page_set->list, &rb->pages);
110 iter--;
111 } while (iter > 0);
112 up_write(&pblk_rb_lock);
113
114 #ifdef CONFIG_NVM_PBLK_DEBUG
115 atomic_set(&rb->inflight_flush_point, 0);
116 #endif
117
118 /*
119 * Initialize rate-limiter, which controls access to the write buffer
120 * but user and GC I/O
121 */
122 pblk_rl_init(&pblk->rl, rb->nr_entries);
123
124 return 0;
125 }
126
127 /*
128 * pblk_rb_calculate_size -- calculate the size of the write buffer
129 */
pblk_rb_calculate_size(unsigned int nr_entries)130 unsigned int pblk_rb_calculate_size(unsigned int nr_entries)
131 {
132 /* Alloc a write buffer that can at least fit 128 entries */
133 return (1 << max(get_count_order(nr_entries), 7));
134 }
135
pblk_rb_entries_ref(struct pblk_rb * rb)136 void *pblk_rb_entries_ref(struct pblk_rb *rb)
137 {
138 return rb->entries;
139 }
140
clean_wctx(struct pblk_w_ctx * w_ctx)141 static void clean_wctx(struct pblk_w_ctx *w_ctx)
142 {
143 int flags;
144
145 flags = READ_ONCE(w_ctx->flags);
146 WARN_ONCE(!(flags & PBLK_SUBMITTED_ENTRY),
147 "pblk: overwriting unsubmitted data\n");
148
149 /* Release flags on context. Protect from writes and reads */
150 smp_store_release(&w_ctx->flags, PBLK_WRITABLE_ENTRY);
151 pblk_ppa_set_empty(&w_ctx->ppa);
152 w_ctx->lba = ADDR_EMPTY;
153 }
154
155 #define pblk_rb_ring_count(head, tail, size) CIRC_CNT(head, tail, size)
156 #define pblk_rb_ring_space(rb, head, tail, size) \
157 (CIRC_SPACE(head, tail, size))
158
159 /*
160 * Buffer space is calculated with respect to the back pointer signaling
161 * synchronized entries to the media.
162 */
pblk_rb_space(struct pblk_rb * rb)163 static unsigned int pblk_rb_space(struct pblk_rb *rb)
164 {
165 unsigned int mem = READ_ONCE(rb->mem);
166 unsigned int sync = READ_ONCE(rb->sync);
167
168 return pblk_rb_ring_space(rb, mem, sync, rb->nr_entries);
169 }
170
171 /*
172 * Buffer count is calculated with respect to the submission entry signaling the
173 * entries that are available to send to the media
174 */
pblk_rb_read_count(struct pblk_rb * rb)175 unsigned int pblk_rb_read_count(struct pblk_rb *rb)
176 {
177 unsigned int mem = READ_ONCE(rb->mem);
178 unsigned int subm = READ_ONCE(rb->subm);
179
180 return pblk_rb_ring_count(mem, subm, rb->nr_entries);
181 }
182
pblk_rb_sync_count(struct pblk_rb * rb)183 unsigned int pblk_rb_sync_count(struct pblk_rb *rb)
184 {
185 unsigned int mem = READ_ONCE(rb->mem);
186 unsigned int sync = READ_ONCE(rb->sync);
187
188 return pblk_rb_ring_count(mem, sync, rb->nr_entries);
189 }
190
pblk_rb_read_commit(struct pblk_rb * rb,unsigned int nr_entries)191 unsigned int pblk_rb_read_commit(struct pblk_rb *rb, unsigned int nr_entries)
192 {
193 unsigned int subm;
194
195 subm = READ_ONCE(rb->subm);
196 /* Commit read means updating submission pointer */
197 smp_store_release(&rb->subm,
198 (subm + nr_entries) & (rb->nr_entries - 1));
199
200 return subm;
201 }
202
__pblk_rb_update_l2p(struct pblk_rb * rb,unsigned int to_update)203 static int __pblk_rb_update_l2p(struct pblk_rb *rb, unsigned int to_update)
204 {
205 struct pblk *pblk = container_of(rb, struct pblk, rwb);
206 struct pblk_line *line;
207 struct pblk_rb_entry *entry;
208 struct pblk_w_ctx *w_ctx;
209 unsigned int user_io = 0, gc_io = 0;
210 unsigned int i;
211 int flags;
212
213 for (i = 0; i < to_update; i++) {
214 entry = &rb->entries[rb->l2p_update];
215 w_ctx = &entry->w_ctx;
216
217 flags = READ_ONCE(entry->w_ctx.flags);
218 if (flags & PBLK_IOTYPE_USER)
219 user_io++;
220 else if (flags & PBLK_IOTYPE_GC)
221 gc_io++;
222 else
223 WARN(1, "pblk: unknown IO type\n");
224
225 pblk_update_map_dev(pblk, w_ctx->lba, w_ctx->ppa,
226 entry->cacheline);
227
228 line = &pblk->lines[pblk_ppa_to_line(w_ctx->ppa)];
229 kref_put(&line->ref, pblk_line_put);
230 clean_wctx(w_ctx);
231 rb->l2p_update = (rb->l2p_update + 1) & (rb->nr_entries - 1);
232 }
233
234 pblk_rl_out(&pblk->rl, user_io, gc_io);
235
236 return 0;
237 }
238
239 /*
240 * When we move the l2p_update pointer, we update the l2p table - lookups will
241 * point to the physical address instead of to the cacheline in the write buffer
242 * from this moment on.
243 */
pblk_rb_update_l2p(struct pblk_rb * rb,unsigned int nr_entries,unsigned int mem,unsigned int sync)244 static int pblk_rb_update_l2p(struct pblk_rb *rb, unsigned int nr_entries,
245 unsigned int mem, unsigned int sync)
246 {
247 unsigned int space, count;
248 int ret = 0;
249
250 lockdep_assert_held(&rb->w_lock);
251
252 /* Update l2p only as buffer entries are being overwritten */
253 space = pblk_rb_ring_space(rb, mem, rb->l2p_update, rb->nr_entries);
254 if (space > nr_entries)
255 goto out;
256
257 count = nr_entries - space;
258 /* l2p_update used exclusively under rb->w_lock */
259 ret = __pblk_rb_update_l2p(rb, count);
260
261 out:
262 return ret;
263 }
264
265 /*
266 * Update the l2p entry for all sectors stored on the write buffer. This means
267 * that all future lookups to the l2p table will point to a device address, not
268 * to the cacheline in the write buffer.
269 */
pblk_rb_sync_l2p(struct pblk_rb * rb)270 void pblk_rb_sync_l2p(struct pblk_rb *rb)
271 {
272 unsigned int sync;
273 unsigned int to_update;
274
275 spin_lock(&rb->w_lock);
276
277 /* Protect from reads and writes */
278 sync = smp_load_acquire(&rb->sync);
279
280 to_update = pblk_rb_ring_count(sync, rb->l2p_update, rb->nr_entries);
281 __pblk_rb_update_l2p(rb, to_update);
282
283 spin_unlock(&rb->w_lock);
284 }
285
286 /*
287 * Write @nr_entries to ring buffer from @data buffer if there is enough space.
288 * Typically, 4KB data chunks coming from a bio will be copied to the ring
289 * buffer, thus the write will fail if not all incoming data can be copied.
290 *
291 */
__pblk_rb_write_entry(struct pblk_rb * rb,void * data,struct pblk_w_ctx w_ctx,struct pblk_rb_entry * entry)292 static void __pblk_rb_write_entry(struct pblk_rb *rb, void *data,
293 struct pblk_w_ctx w_ctx,
294 struct pblk_rb_entry *entry)
295 {
296 memcpy(entry->data, data, rb->seg_size);
297
298 entry->w_ctx.lba = w_ctx.lba;
299 entry->w_ctx.ppa = w_ctx.ppa;
300 }
301
pblk_rb_write_entry_user(struct pblk_rb * rb,void * data,struct pblk_w_ctx w_ctx,unsigned int ring_pos)302 void pblk_rb_write_entry_user(struct pblk_rb *rb, void *data,
303 struct pblk_w_ctx w_ctx, unsigned int ring_pos)
304 {
305 struct pblk *pblk = container_of(rb, struct pblk, rwb);
306 struct pblk_rb_entry *entry;
307 int flags;
308
309 entry = &rb->entries[ring_pos];
310 flags = READ_ONCE(entry->w_ctx.flags);
311 #ifdef CONFIG_NVM_PBLK_DEBUG
312 /* Caller must guarantee that the entry is free */
313 BUG_ON(!(flags & PBLK_WRITABLE_ENTRY));
314 #endif
315
316 __pblk_rb_write_entry(rb, data, w_ctx, entry);
317
318 pblk_update_map_cache(pblk, w_ctx.lba, entry->cacheline);
319 flags = w_ctx.flags | PBLK_WRITTEN_DATA;
320
321 /* Release flags on write context. Protect from writes */
322 smp_store_release(&entry->w_ctx.flags, flags);
323 }
324
pblk_rb_write_entry_gc(struct pblk_rb * rb,void * data,struct pblk_w_ctx w_ctx,struct pblk_line * line,u64 paddr,unsigned int ring_pos)325 void pblk_rb_write_entry_gc(struct pblk_rb *rb, void *data,
326 struct pblk_w_ctx w_ctx, struct pblk_line *line,
327 u64 paddr, unsigned int ring_pos)
328 {
329 struct pblk *pblk = container_of(rb, struct pblk, rwb);
330 struct pblk_rb_entry *entry;
331 int flags;
332
333 entry = &rb->entries[ring_pos];
334 flags = READ_ONCE(entry->w_ctx.flags);
335 #ifdef CONFIG_NVM_PBLK_DEBUG
336 /* Caller must guarantee that the entry is free */
337 BUG_ON(!(flags & PBLK_WRITABLE_ENTRY));
338 #endif
339
340 __pblk_rb_write_entry(rb, data, w_ctx, entry);
341
342 if (!pblk_update_map_gc(pblk, w_ctx.lba, entry->cacheline, line, paddr))
343 entry->w_ctx.lba = ADDR_EMPTY;
344
345 flags = w_ctx.flags | PBLK_WRITTEN_DATA;
346
347 /* Release flags on write context. Protect from writes */
348 smp_store_release(&entry->w_ctx.flags, flags);
349 }
350
pblk_rb_flush_point_set(struct pblk_rb * rb,struct bio * bio,unsigned int pos)351 static int pblk_rb_flush_point_set(struct pblk_rb *rb, struct bio *bio,
352 unsigned int pos)
353 {
354 struct pblk_rb_entry *entry;
355 unsigned int sync, flush_point;
356
357 pblk_rb_sync_init(rb, NULL);
358 sync = READ_ONCE(rb->sync);
359
360 if (pos == sync) {
361 pblk_rb_sync_end(rb, NULL);
362 return 0;
363 }
364
365 #ifdef CONFIG_NVM_PBLK_DEBUG
366 atomic_inc(&rb->inflight_flush_point);
367 #endif
368
369 flush_point = (pos == 0) ? (rb->nr_entries - 1) : (pos - 1);
370 entry = &rb->entries[flush_point];
371
372 /* Protect flush points */
373 smp_store_release(&rb->flush_point, flush_point);
374
375 if (bio)
376 bio_list_add(&entry->w_ctx.bios, bio);
377
378 pblk_rb_sync_end(rb, NULL);
379
380 return bio ? 1 : 0;
381 }
382
__pblk_rb_may_write(struct pblk_rb * rb,unsigned int nr_entries,unsigned int * pos)383 static int __pblk_rb_may_write(struct pblk_rb *rb, unsigned int nr_entries,
384 unsigned int *pos)
385 {
386 unsigned int mem;
387 unsigned int sync;
388
389 sync = READ_ONCE(rb->sync);
390 mem = READ_ONCE(rb->mem);
391
392 if (pblk_rb_ring_space(rb, mem, sync, rb->nr_entries) < nr_entries)
393 return 0;
394
395 if (pblk_rb_update_l2p(rb, nr_entries, mem, sync))
396 return 0;
397
398 *pos = mem;
399
400 return 1;
401 }
402
pblk_rb_may_write(struct pblk_rb * rb,unsigned int nr_entries,unsigned int * pos)403 static int pblk_rb_may_write(struct pblk_rb *rb, unsigned int nr_entries,
404 unsigned int *pos)
405 {
406 if (!__pblk_rb_may_write(rb, nr_entries, pos))
407 return 0;
408
409 /* Protect from read count */
410 smp_store_release(&rb->mem, (*pos + nr_entries) & (rb->nr_entries - 1));
411 return 1;
412 }
413
pblk_rb_flush(struct pblk_rb * rb)414 void pblk_rb_flush(struct pblk_rb *rb)
415 {
416 struct pblk *pblk = container_of(rb, struct pblk, rwb);
417 unsigned int mem = READ_ONCE(rb->mem);
418
419 if (pblk_rb_flush_point_set(rb, NULL, mem))
420 return;
421
422 pblk_write_kick(pblk);
423 }
424
pblk_rb_may_write_flush(struct pblk_rb * rb,unsigned int nr_entries,unsigned int * pos,struct bio * bio,int * io_ret)425 static int pblk_rb_may_write_flush(struct pblk_rb *rb, unsigned int nr_entries,
426 unsigned int *pos, struct bio *bio,
427 int *io_ret)
428 {
429 unsigned int mem;
430
431 if (!__pblk_rb_may_write(rb, nr_entries, pos))
432 return 0;
433
434 mem = (*pos + nr_entries) & (rb->nr_entries - 1);
435 *io_ret = NVM_IO_DONE;
436
437 if (bio->bi_opf & REQ_PREFLUSH) {
438 struct pblk *pblk = container_of(rb, struct pblk, rwb);
439
440 atomic64_inc(&pblk->nr_flush);
441 if (pblk_rb_flush_point_set(&pblk->rwb, bio, mem))
442 *io_ret = NVM_IO_OK;
443 }
444
445 /* Protect from read count */
446 smp_store_release(&rb->mem, mem);
447
448 return 1;
449 }
450
451 /*
452 * Atomically check that (i) there is space on the write buffer for the
453 * incoming I/O, and (ii) the current I/O type has enough budget in the write
454 * buffer (rate-limiter).
455 */
pblk_rb_may_write_user(struct pblk_rb * rb,struct bio * bio,unsigned int nr_entries,unsigned int * pos)456 int pblk_rb_may_write_user(struct pblk_rb *rb, struct bio *bio,
457 unsigned int nr_entries, unsigned int *pos)
458 {
459 struct pblk *pblk = container_of(rb, struct pblk, rwb);
460 int io_ret;
461
462 spin_lock(&rb->w_lock);
463 io_ret = pblk_rl_user_may_insert(&pblk->rl, nr_entries);
464 if (io_ret) {
465 spin_unlock(&rb->w_lock);
466 return io_ret;
467 }
468
469 if (!pblk_rb_may_write_flush(rb, nr_entries, pos, bio, &io_ret)) {
470 spin_unlock(&rb->w_lock);
471 return NVM_IO_REQUEUE;
472 }
473
474 pblk_rl_user_in(&pblk->rl, nr_entries);
475 spin_unlock(&rb->w_lock);
476
477 return io_ret;
478 }
479
480 /*
481 * Look at pblk_rb_may_write_user comment
482 */
pblk_rb_may_write_gc(struct pblk_rb * rb,unsigned int nr_entries,unsigned int * pos)483 int pblk_rb_may_write_gc(struct pblk_rb *rb, unsigned int nr_entries,
484 unsigned int *pos)
485 {
486 struct pblk *pblk = container_of(rb, struct pblk, rwb);
487
488 spin_lock(&rb->w_lock);
489 if (!pblk_rl_gc_may_insert(&pblk->rl, nr_entries)) {
490 spin_unlock(&rb->w_lock);
491 return 0;
492 }
493
494 if (!pblk_rb_may_write(rb, nr_entries, pos)) {
495 spin_unlock(&rb->w_lock);
496 return 0;
497 }
498
499 pblk_rl_gc_in(&pblk->rl, nr_entries);
500 spin_unlock(&rb->w_lock);
501
502 return 1;
503 }
504
505 /*
506 * Read available entries on rb and add them to the given bio. To avoid a memory
507 * copy, a page reference to the write buffer is used to be added to the bio.
508 *
509 * This function is used by the write thread to form the write bio that will
510 * persist data on the write buffer to the media.
511 */
pblk_rb_read_to_bio(struct pblk_rb * rb,struct nvm_rq * rqd,unsigned int pos,unsigned int nr_entries,unsigned int count)512 unsigned int pblk_rb_read_to_bio(struct pblk_rb *rb, struct nvm_rq *rqd,
513 unsigned int pos, unsigned int nr_entries,
514 unsigned int count)
515 {
516 struct pblk *pblk = container_of(rb, struct pblk, rwb);
517 struct request_queue *q = pblk->dev->q;
518 struct pblk_c_ctx *c_ctx = nvm_rq_to_pdu(rqd);
519 struct bio *bio = rqd->bio;
520 struct pblk_rb_entry *entry;
521 struct page *page;
522 unsigned int pad = 0, to_read = nr_entries;
523 unsigned int i;
524 int flags;
525
526 if (count < nr_entries) {
527 pad = nr_entries - count;
528 to_read = count;
529 }
530
531 c_ctx->sentry = pos;
532 c_ctx->nr_valid = to_read;
533 c_ctx->nr_padded = pad;
534
535 for (i = 0; i < to_read; i++) {
536 entry = &rb->entries[pos];
537
538 /* A write has been allowed into the buffer, but data is still
539 * being copied to it. It is ok to busy wait.
540 */
541 try:
542 flags = READ_ONCE(entry->w_ctx.flags);
543 if (!(flags & PBLK_WRITTEN_DATA)) {
544 io_schedule();
545 goto try;
546 }
547
548 page = virt_to_page(entry->data);
549 if (!page) {
550 pblk_err(pblk, "could not allocate write bio page\n");
551 flags &= ~PBLK_WRITTEN_DATA;
552 flags |= PBLK_SUBMITTED_ENTRY;
553 /* Release flags on context. Protect from writes */
554 smp_store_release(&entry->w_ctx.flags, flags);
555 return NVM_IO_ERR;
556 }
557
558 if (bio_add_pc_page(q, bio, page, rb->seg_size, 0) !=
559 rb->seg_size) {
560 pblk_err(pblk, "could not add page to write bio\n");
561 flags &= ~PBLK_WRITTEN_DATA;
562 flags |= PBLK_SUBMITTED_ENTRY;
563 /* Release flags on context. Protect from writes */
564 smp_store_release(&entry->w_ctx.flags, flags);
565 return NVM_IO_ERR;
566 }
567
568 flags &= ~PBLK_WRITTEN_DATA;
569 flags |= PBLK_SUBMITTED_ENTRY;
570
571 /* Release flags on context. Protect from writes */
572 smp_store_release(&entry->w_ctx.flags, flags);
573
574 pos = (pos + 1) & (rb->nr_entries - 1);
575 }
576
577 if (pad) {
578 if (pblk_bio_add_pages(pblk, bio, GFP_KERNEL, pad)) {
579 pblk_err(pblk, "could not pad page in write bio\n");
580 return NVM_IO_ERR;
581 }
582
583 if (pad < pblk->min_write_pgs)
584 atomic64_inc(&pblk->pad_dist[pad - 1]);
585 else
586 pblk_warn(pblk, "padding more than min. sectors\n");
587
588 atomic64_add(pad, &pblk->pad_wa);
589 }
590
591 #ifdef CONFIG_NVM_PBLK_DEBUG
592 atomic_long_add(pad, &pblk->padded_writes);
593 #endif
594
595 return NVM_IO_OK;
596 }
597
598 /*
599 * Copy to bio only if the lba matches the one on the given cache entry.
600 * Otherwise, it means that the entry has been overwritten, and the bio should
601 * be directed to disk.
602 */
pblk_rb_copy_to_bio(struct pblk_rb * rb,struct bio * bio,sector_t lba,struct ppa_addr ppa,int bio_iter,bool advanced_bio)603 int pblk_rb_copy_to_bio(struct pblk_rb *rb, struct bio *bio, sector_t lba,
604 struct ppa_addr ppa, int bio_iter, bool advanced_bio)
605 {
606 struct pblk *pblk = container_of(rb, struct pblk, rwb);
607 struct pblk_rb_entry *entry;
608 struct pblk_w_ctx *w_ctx;
609 struct ppa_addr l2p_ppa;
610 u64 pos = pblk_addr_to_cacheline(ppa);
611 void *data;
612 int flags;
613 int ret = 1;
614
615
616 #ifdef CONFIG_NVM_PBLK_DEBUG
617 /* Caller must ensure that the access will not cause an overflow */
618 BUG_ON(pos >= rb->nr_entries);
619 #endif
620 entry = &rb->entries[pos];
621 w_ctx = &entry->w_ctx;
622 flags = READ_ONCE(w_ctx->flags);
623
624 spin_lock(&rb->w_lock);
625 spin_lock(&pblk->trans_lock);
626 l2p_ppa = pblk_trans_map_get(pblk, lba);
627 spin_unlock(&pblk->trans_lock);
628
629 /* Check if the entry has been overwritten or is scheduled to be */
630 if (!pblk_ppa_comp(l2p_ppa, ppa) || w_ctx->lba != lba ||
631 flags & PBLK_WRITABLE_ENTRY) {
632 ret = 0;
633 goto out;
634 }
635
636 /* Only advance the bio if it hasn't been advanced already. If advanced,
637 * this bio is at least a partial bio (i.e., it has partially been
638 * filled with data from the cache). If part of the data resides on the
639 * media, we will read later on
640 */
641 if (unlikely(!advanced_bio))
642 bio_advance(bio, bio_iter * PBLK_EXPOSED_PAGE_SIZE);
643
644 data = bio_data(bio);
645 memcpy(data, entry->data, rb->seg_size);
646
647 out:
648 spin_unlock(&rb->w_lock);
649 return ret;
650 }
651
pblk_rb_w_ctx(struct pblk_rb * rb,unsigned int pos)652 struct pblk_w_ctx *pblk_rb_w_ctx(struct pblk_rb *rb, unsigned int pos)
653 {
654 unsigned int entry = pos & (rb->nr_entries - 1);
655
656 return &rb->entries[entry].w_ctx;
657 }
658
pblk_rb_sync_init(struct pblk_rb * rb,unsigned long * flags)659 unsigned int pblk_rb_sync_init(struct pblk_rb *rb, unsigned long *flags)
660 __acquires(&rb->s_lock)
661 {
662 if (flags)
663 spin_lock_irqsave(&rb->s_lock, *flags);
664 else
665 spin_lock_irq(&rb->s_lock);
666
667 return rb->sync;
668 }
669
pblk_rb_sync_end(struct pblk_rb * rb,unsigned long * flags)670 void pblk_rb_sync_end(struct pblk_rb *rb, unsigned long *flags)
671 __releases(&rb->s_lock)
672 {
673 lockdep_assert_held(&rb->s_lock);
674
675 if (flags)
676 spin_unlock_irqrestore(&rb->s_lock, *flags);
677 else
678 spin_unlock_irq(&rb->s_lock);
679 }
680
pblk_rb_sync_advance(struct pblk_rb * rb,unsigned int nr_entries)681 unsigned int pblk_rb_sync_advance(struct pblk_rb *rb, unsigned int nr_entries)
682 {
683 unsigned int sync, flush_point;
684 lockdep_assert_held(&rb->s_lock);
685
686 sync = READ_ONCE(rb->sync);
687 flush_point = READ_ONCE(rb->flush_point);
688
689 if (flush_point != EMPTY_ENTRY) {
690 unsigned int secs_to_flush;
691
692 secs_to_flush = pblk_rb_ring_count(flush_point, sync,
693 rb->nr_entries);
694 if (secs_to_flush < nr_entries) {
695 /* Protect flush points */
696 smp_store_release(&rb->flush_point, EMPTY_ENTRY);
697 }
698 }
699
700 sync = (sync + nr_entries) & (rb->nr_entries - 1);
701
702 /* Protect from counts */
703 smp_store_release(&rb->sync, sync);
704
705 return sync;
706 }
707
708 /* Calculate how many sectors to submit up to the current flush point. */
pblk_rb_flush_point_count(struct pblk_rb * rb)709 unsigned int pblk_rb_flush_point_count(struct pblk_rb *rb)
710 {
711 unsigned int subm, sync, flush_point;
712 unsigned int submitted, to_flush;
713
714 /* Protect flush points */
715 flush_point = smp_load_acquire(&rb->flush_point);
716 if (flush_point == EMPTY_ENTRY)
717 return 0;
718
719 /* Protect syncs */
720 sync = smp_load_acquire(&rb->sync);
721
722 subm = READ_ONCE(rb->subm);
723 submitted = pblk_rb_ring_count(subm, sync, rb->nr_entries);
724
725 /* The sync point itself counts as a sector to sync */
726 to_flush = pblk_rb_ring_count(flush_point, sync, rb->nr_entries) + 1;
727
728 return (submitted < to_flush) ? (to_flush - submitted) : 0;
729 }
730
731 /*
732 * Scan from the current position of the sync pointer to find the entry that
733 * corresponds to the given ppa. This is necessary since write requests can be
734 * completed out of order. The assumption is that the ppa is close to the sync
735 * pointer thus the search will not take long.
736 *
737 * The caller of this function must guarantee that the sync pointer will no
738 * reach the entry while it is using the metadata associated with it. With this
739 * assumption in mind, there is no need to take the sync lock.
740 */
pblk_rb_sync_scan_entry(struct pblk_rb * rb,struct ppa_addr * ppa)741 struct pblk_rb_entry *pblk_rb_sync_scan_entry(struct pblk_rb *rb,
742 struct ppa_addr *ppa)
743 {
744 unsigned int sync, subm, count;
745 unsigned int i;
746
747 sync = READ_ONCE(rb->sync);
748 subm = READ_ONCE(rb->subm);
749 count = pblk_rb_ring_count(subm, sync, rb->nr_entries);
750
751 for (i = 0; i < count; i++)
752 sync = (sync + 1) & (rb->nr_entries - 1);
753
754 return NULL;
755 }
756
pblk_rb_tear_down_check(struct pblk_rb * rb)757 int pblk_rb_tear_down_check(struct pblk_rb *rb)
758 {
759 struct pblk_rb_entry *entry;
760 int i;
761 int ret = 0;
762
763 spin_lock(&rb->w_lock);
764 spin_lock_irq(&rb->s_lock);
765
766 if ((rb->mem == rb->subm) && (rb->subm == rb->sync) &&
767 (rb->sync == rb->l2p_update) &&
768 (rb->flush_point == EMPTY_ENTRY)) {
769 goto out;
770 }
771
772 if (!rb->entries) {
773 ret = 1;
774 goto out;
775 }
776
777 for (i = 0; i < rb->nr_entries; i++) {
778 entry = &rb->entries[i];
779
780 if (!entry->data) {
781 ret = 1;
782 goto out;
783 }
784 }
785
786 out:
787 spin_unlock_irq(&rb->s_lock);
788 spin_unlock(&rb->w_lock);
789
790 return ret;
791 }
792
pblk_rb_wrap_pos(struct pblk_rb * rb,unsigned int pos)793 unsigned int pblk_rb_wrap_pos(struct pblk_rb *rb, unsigned int pos)
794 {
795 return (pos & (rb->nr_entries - 1));
796 }
797
pblk_rb_pos_oob(struct pblk_rb * rb,u64 pos)798 int pblk_rb_pos_oob(struct pblk_rb *rb, u64 pos)
799 {
800 return (pos >= rb->nr_entries);
801 }
802
pblk_rb_sysfs(struct pblk_rb * rb,char * buf)803 ssize_t pblk_rb_sysfs(struct pblk_rb *rb, char *buf)
804 {
805 struct pblk *pblk = container_of(rb, struct pblk, rwb);
806 struct pblk_c_ctx *c;
807 ssize_t offset;
808 int queued_entries = 0;
809
810 spin_lock_irq(&rb->s_lock);
811 list_for_each_entry(c, &pblk->compl_list, list)
812 queued_entries++;
813 spin_unlock_irq(&rb->s_lock);
814
815 if (rb->flush_point != EMPTY_ENTRY)
816 offset = scnprintf(buf, PAGE_SIZE,
817 "%u\t%u\t%u\t%u\t%u\t%u\t%u - %u/%u/%u - %d\n",
818 rb->nr_entries,
819 rb->mem,
820 rb->subm,
821 rb->sync,
822 rb->l2p_update,
823 #ifdef CONFIG_NVM_PBLK_DEBUG
824 atomic_read(&rb->inflight_flush_point),
825 #else
826 0,
827 #endif
828 rb->flush_point,
829 pblk_rb_read_count(rb),
830 pblk_rb_space(rb),
831 pblk_rb_flush_point_count(rb),
832 queued_entries);
833 else
834 offset = scnprintf(buf, PAGE_SIZE,
835 "%u\t%u\t%u\t%u\t%u\t%u\tNULL - %u/%u/%u - %d\n",
836 rb->nr_entries,
837 rb->mem,
838 rb->subm,
839 rb->sync,
840 rb->l2p_update,
841 #ifdef CONFIG_NVM_PBLK_DEBUG
842 atomic_read(&rb->inflight_flush_point),
843 #else
844 0,
845 #endif
846 pblk_rb_read_count(rb),
847 pblk_rb_space(rb),
848 pblk_rb_flush_point_count(rb),
849 queued_entries);
850
851 return offset;
852 }
853