md/bcache/alloc.c

7  * Allocation in bcache is done in terms of buckets:
17  * of buckets on disk, with a pointer to them in the journal header.
25  * batch this up: We fill up the free_inc list with freshly invalidated buckets,
26  * call prio_write(), and when prio_write() finishes we pull buckets off the
31  * smaller freelist, and buckets on that list are always ready to be used.
36  * There is another freelist, because sometimes we have buckets that we know
38  * priorities to be rewritten. These come from freed btree nodes and buckets
40  * them (because they were overwritten). That's the unused list - buckets on the
57  * buckets are ready.
59  * invalidate_buckets_(lru|fifo)() find buckets that are available to be
119  * Background allocation thread: scans for buckets to be invalidated,
121  * then optionally issues discard commands to the newly free buckets, then puts
146 		trace_bcache_invalidate(ca, b - ca->buckets);  in __bch_invalidate_one_bucket()
157 	fifo_push(&ca->free_inc, b - ca->buckets);  in bch_invalidate_one_bucket()
161  * Determines what order we're going to reuse buckets, smallest bucket_prio()
226 		b = ca->buckets + ca->fifo_last_bucket++;  in invalidate_buckets_fifo()
252 		b = ca->buckets + n;  in invalidate_buckets_random()
325 		 * First, we pull buckets off of the unused and free_inc lists,  in bch_allocator_thread()
349 		 * We've run out of free buckets, we need to find some buckets  in bch_allocator_thread()
445 	b = ca->buckets + r;  in bch_bucket_alloc()
507 	k->ptr[0] = MAKE_PTR(ca->buckets[b].gen,  in __bch_bucket_alloc_set()
541  * We keep multiple buckets open for writes, and try to segregate different
552  * dirty sectors mixed with dirty sectors of cached device, such buckets will
559  * data to the same buckets it'd get invalidated at the same time.
683 	 * k takes refcounts on the buckets it points to until it's inserted  in bch_alloc_sectors()