Merge tag 'block-5.6-2020-02-05' of git://git.kernel.dk/linux-block

Pull more block updates from Jens Axboe:
 "Some later arrivals, but all fixes at this point:

   - bcache fix series (Coly)

   - Series of BFQ fixes (Paolo)

   - NVMe pull request from Keith with a few minor NVMe fixes

   - Various little tweaks"

* tag 'block-5.6-2020-02-05' of git://git.kernel.dk/linux-block: (23 commits)
  nvmet: update AEN list and array at one place
  nvmet: Fix controller use after free
  nvmet: Fix error print message at nvmet_install_queue function
  brd: check and limit max_part par
  nvme-pci: remove nvmeq->tags
  nvmet: fix dsm failure when payload does not match sgl descriptor
  nvmet: Pass lockdep expression to RCU lists
  block, bfq: clarify the goal of bfq_split_bfqq()
  block, bfq: get a ref to a group when adding it to a service tree
  block, bfq: remove ifdefs from around gets/puts of bfq groups
  block, bfq: extend incomplete name of field on_st
  block, bfq: get extra ref to prevent a queue from being freed during a group move
  block, bfq: do not insert oom queue into position tree
  block, bfq: do not plug I/O for bfq_queues with no proc refs
  bcache: check return value of prio_read()
  bcache: fix incorrect data type usage in btree_flush_write()
  bcache: add readahead cache policy options via sysfs interface
  bcache: explicity type cast in bset_bkey_last()
  bcache: fix memory corruption in bch_cache_accounting_clear()
  xen/blkfront: limit allocated memory size to actual use case
  ...

7 files changed, 64 insertions, 15 deletions

diff --git a/drivers/md/bcache/bcache.h b/drivers/md/bcache/bcache.h
index adf26a21fcd1..74a9849ea164 100644
--- a/drivers/md/bcache/bcache.h
+++ b/drivers/md/bcache/bcache.h
@@ -330,6 +330,9 @@ struct cached_dev {
 	 */
 	atomic_t		has_dirty;
 
+#define BCH_CACHE_READA_ALL		0
+#define BCH_CACHE_READA_META_ONLY	1
+	unsigned int		cache_readahead_policy;
 	struct bch_ratelimit	writeback_rate;
 	struct delayed_work	writeback_rate_update;
 
diff --git a/drivers/md/bcache/bset.h b/drivers/md/bcache/bset.h
index c71365e7c1fa..a50dcfda656f 100644
--- a/drivers/md/bcache/bset.h
+++ b/drivers/md/bcache/bset.h
@@ -397,7 +397,8 @@ void bch_btree_keys_stats(struct btree_keys *b, struct bset_stats *state);
 
 /* Bkey utility code */
 
-#define bset_bkey_last(i)	bkey_idx((struct bkey *) (i)->d, (i)->keys)
+#define bset_bkey_last(i)	bkey_idx((struct bkey *) (i)->d, \
+					 (unsigned int)(i)->keys)
 
 static inline struct bkey *bset_bkey_idx(struct bset *i, unsigned int idx)
 {
diff --git a/drivers/md/bcache/journal.c b/drivers/md/bcache/journal.c
index 33ddc5269e8d..6730820780b0 100644
--- a/drivers/md/bcache/journal.c
+++ b/drivers/md/bcache/journal.c
@@ -422,7 +422,8 @@ err:
 static void btree_flush_write(struct cache_set *c)
 {
 	struct btree *b, *t, *btree_nodes[BTREE_FLUSH_NR];
-	unsigned int i, nr, ref_nr;
+	unsigned int i, nr;
+	int ref_nr;
 	atomic_t *fifo_front_p, *now_fifo_front_p;
 	size_t mask;
 
diff --git a/drivers/md/bcache/request.c b/drivers/md/bcache/request.c
index 73478a91a342..820d8402a1dc 100644
--- a/drivers/md/bcache/request.c
+++ b/drivers/md/bcache/request.c
@@ -379,13 +379,20 @@ static bool check_should_bypass(struct cached_dev *dc, struct bio *bio)
 		goto skip;
 
 	/*
-	 * Flag for bypass if the IO is for read-ahead or background,
-	 * unless the read-ahead request is for metadata
+	 * If the bio is for read-ahead or background IO, bypass it or
+	 * not depends on the following situations,
+	 * - If the IO is for meta data, always cache it and no bypass
+	 * - If the IO is not meta data, check dc->cache_reada_policy,
+	 *      BCH_CACHE_READA_ALL: cache it and not bypass
+	 *      BCH_CACHE_READA_META_ONLY: not cache it and bypass
+	 * That is, read-ahead request for metadata always get cached
 	 * (eg, for gfs2 or xfs).
 	 */
-	if (bio->bi_opf & (REQ_RAHEAD|REQ_BACKGROUND) &&
-	    !(bio->bi_opf & (REQ_META|REQ_PRIO)))
-		goto skip;
+	if ((bio->bi_opf & (REQ_RAHEAD|REQ_BACKGROUND))) {
+		if (!(bio->bi_opf & (REQ_META|REQ_PRIO)) &&
+		    (dc->cache_readahead_policy != BCH_CACHE_READA_ALL))
+			goto skip;
+	}
 
 	if (bio->bi_iter.bi_sector & (c->sb.block_size - 1) ||
 	    bio_sectors(bio) & (c->sb.block_size - 1)) {
diff --git a/drivers/md/bcache/stats.c b/drivers/md/bcache/stats.c
index ba1c93791d8d..503aafe188dc 100644
--- a/drivers/md/bcache/stats.c
+++ b/drivers/md/bcache/stats.c
@@ -109,9 +109,13 @@ int bch_cache_accounting_add_kobjs(struct cache_accounting *acc,
 
 void bch_cache_accounting_clear(struct cache_accounting *acc)
 {
-	memset(&acc->total.cache_hits,
-	       0,
-	       sizeof(struct cache_stats));
+	acc->total.cache_hits = 0;
+	acc->total.cache_misses = 0;
+	acc->total.cache_bypass_hits = 0;
+	acc->total.cache_bypass_misses = 0;
+	acc->total.cache_readaheads = 0;
+	acc->total.cache_miss_collisions = 0;
+	acc->total.sectors_bypassed = 0;
 }
 
 void bch_cache_accounting_destroy(struct cache_accounting *acc)
diff --git a/drivers/md/bcache/super.c b/drivers/md/bcache/super.c
index 3dea1d5acd5c..2749daf09724 100644
--- a/drivers/md/bcache/super.c
+++ b/drivers/md/bcache/super.c
@@ -609,12 +609,13 @@ int bch_prio_write(struct cache *ca, bool wait)
 	return 0;
 }
 
-static void prio_read(struct cache *ca, uint64_t bucket)
+static int prio_read(struct cache *ca, uint64_t bucket)
 {
 	struct prio_set *p = ca->disk_buckets;
 	struct bucket_disk *d = p->data + prios_per_bucket(ca), *end = d;
 	struct bucket *b;
 	unsigned int bucket_nr = 0;
+	int ret = -EIO;
 
 	for (b = ca->buckets;
 	     b < ca->buckets + ca->sb.nbuckets;
@@ -627,11 +628,15 @@ static void prio_read(struct cache *ca, uint64_t bucket)
 			prio_io(ca, bucket, REQ_OP_READ, 0);
 
 			if (p->csum !=
-			    bch_crc64(&p->magic, bucket_bytes(ca) - 8))
+			    bch_crc64(&p->magic, bucket_bytes(ca) - 8)) {
 				pr_warn("bad csum reading priorities");
+				goto out;
+			}
 
-			if (p->magic != pset_magic(&ca->sb))
+			if (p->magic != pset_magic(&ca->sb)) {
 				pr_warn("bad magic reading priorities");
+				goto out;
+			}
 
 			bucket = p->next_bucket;
 			d = p->data;
@@ -640,6 +645,10 @@ static void prio_read(struct cache *ca, uint64_t bucket)
 		b->prio = le16_to_cpu(d->prio);
 		b->gen = b->last_gc = d->gen;
 	}
+
+	ret = 0;
+out:
+	return ret;
 }
 
 /* Bcache device */
@@ -1873,8 +1882,10 @@ static int run_cache_set(struct cache_set *c)
 		j = &list_entry(journal.prev, struct journal_replay, list)->j;
 
 		err = "IO error reading priorities";
-		for_each_cache(ca, c, i)
-			prio_read(ca, j->prio_bucket[ca->sb.nr_this_dev]);
+		for_each_cache(ca, c, i) {
+			if (prio_read(ca, j->prio_bucket[ca->sb.nr_this_dev]))
+				goto err;
+		}
 
 		/*
 		 * If prio_read() fails it'll call cache_set_error and we'll
diff --git a/drivers/md/bcache/sysfs.c b/drivers/md/bcache/sysfs.c
index 733e2ddf3c78..3470fae4eabc 100644
--- a/drivers/md/bcache/sysfs.c
+++ b/drivers/md/bcache/sysfs.c
@@ -27,6 +27,12 @@ static const char * const bch_cache_modes[] = {
 	NULL
 };
 
+static const char * const bch_reada_cache_policies[] = {
+	"all",
+	"meta-only",
+	NULL
+};
+
 /* Default is 0 ("auto") */
 static const char * const bch_stop_on_failure_modes[] = {
 	"auto",
@@ -100,6 +106,7 @@ rw_attribute(congested_write_threshold_us);
 rw_attribute(sequential_cutoff);
 rw_attribute(data_csum);
 rw_attribute(cache_mode);
+rw_attribute(readahead_cache_policy);
 rw_attribute(stop_when_cache_set_failed);
 rw_attribute(writeback_metadata);
 rw_attribute(writeback_running);
@@ -168,6 +175,11 @@ SHOW(__bch_cached_dev)
 					       bch_cache_modes,
 					       BDEV_CACHE_MODE(&dc->sb));
 
+	if (attr == &sysfs_readahead_cache_policy)
+		return bch_snprint_string_list(buf, PAGE_SIZE,
+					      bch_reada_cache_policies,
+					      dc->cache_readahead_policy);
+
 	if (attr == &sysfs_stop_when_cache_set_failed)
 		return bch_snprint_string_list(buf, PAGE_SIZE,
 					       bch_stop_on_failure_modes,
@@ -353,6 +365,15 @@ STORE(__cached_dev)
 		}
 	}
 
+	if (attr == &sysfs_readahead_cache_policy) {
+		v = __sysfs_match_string(bch_reada_cache_policies, -1, buf);
+		if (v < 0)
+			return v;
+
+		if ((unsigned int) v != dc->cache_readahead_policy)
+			dc->cache_readahead_policy = v;
+	}
+
 	if (attr == &sysfs_stop_when_cache_set_failed) {
 		v = __sysfs_match_string(bch_stop_on_failure_modes, -1, buf);
 		if (v < 0)
@@ -467,6 +488,7 @@ static struct attribute *bch_cached_dev_files[] = {
 	&sysfs_data_csum,
 #endif
 	&sysfs_cache_mode,
+	&sysfs_readahead_cache_policy,
 	&sysfs_stop_when_cache_set_failed,
 	&sysfs_writeback_metadata,
 	&sysfs_writeback_running,