summaryrefslogtreecommitdiff
diff options
context:
space:
mode:
authorKent Overstreet <kmo@daterainc.com>2014-03-17 22:13:26 (GMT)
committerKent Overstreet <kmo@daterainc.com>2014-03-18 19:22:33 (GMT)
commit487dded86ea065317aea121bec8f1816f2f235c9 (patch)
tree3d333832a9b133156985af197d1bd2952b033c6b
parent0bd143fd800055b1db756693289bbebdb93f2a73 (diff)
downloadlinux-487dded86ea065317aea121bec8f1816f2f235c9.tar.xz
bcache: Fix another bug recovering from unclean shutdown
The on disk bucket gens are allowed to be out of date, when we reuse buckets that didn't have any live data in them. To deal with this, the initial gc has to update the bucket gen when we find a pointer gen newer than the bucket's gen. Unfortunately we weren't doing this for pointers in the journal that we're about to replay. Signed-off-by: Kent Overstreet <kmo@daterainc.com>
-rw-r--r--drivers/md/bcache/btree.c82
-rw-r--r--drivers/md/bcache/btree.h2
-rw-r--r--drivers/md/bcache/journal.c17
3 files changed, 36 insertions, 65 deletions
diff --git a/drivers/md/bcache/btree.c b/drivers/md/bcache/btree.c
index 2d4a864..5f587ce 100644
--- a/drivers/md/bcache/btree.c
+++ b/drivers/md/bcache/btree.c
@@ -1126,7 +1126,8 @@ static int btree_check_reserve(struct btree *b, struct btree_op *op)
/* Garbage collection */
-uint8_t __bch_btree_mark_key(struct cache_set *c, int level, struct bkey *k)
+static uint8_t __bch_btree_mark_key(struct cache_set *c, int level,
+ struct bkey *k)
{
uint8_t stale = 0;
unsigned i;
@@ -1177,6 +1178,26 @@ uint8_t __bch_btree_mark_key(struct cache_set *c, int level, struct bkey *k)
#define btree_mark_key(b, k) __bch_btree_mark_key(b->c, b->level, k)
+void bch_initial_mark_key(struct cache_set *c, int level, struct bkey *k)
+{
+ unsigned i;
+
+ for (i = 0; i < KEY_PTRS(k); i++)
+ if (ptr_available(c, k, i) &&
+ !ptr_stale(c, k, i)) {
+ struct bucket *b = PTR_BUCKET(c, k, i);
+
+ b->gen = PTR_GEN(k, i);
+
+ if (level && bkey_cmp(k, &ZERO_KEY))
+ b->prio = BTREE_PRIO;
+ else if (!level && b->prio == BTREE_PRIO)
+ b->prio = INITIAL_PRIO;
+ }
+
+ __bch_btree_mark_key(c, level, k);
+}
+
static bool btree_gc_mark_node(struct btree *b, struct gc_stat *gc)
{
uint8_t stale = 0;
@@ -1511,6 +1532,8 @@ static int bch_btree_gc_root(struct btree *b, struct btree_op *op,
}
}
+ __bch_btree_mark_key(b->c, b->level + 1, &b->key);
+
if (b->level) {
ret = btree_gc_recurse(b, op, writes, gc);
if (ret)
@@ -1561,11 +1584,6 @@ size_t bch_btree_gc_finish(struct cache_set *c)
c->gc_mark_valid = 1;
c->need_gc = 0;
- if (c->root)
- for (i = 0; i < KEY_PTRS(&c->root->key); i++)
- SET_GC_MARK(PTR_BUCKET(c, &c->root->key, i),
- GC_MARK_METADATA);
-
for (i = 0; i < KEY_PTRS(&c->uuid_bucket); i++)
SET_GC_MARK(PTR_BUCKET(c, &c->uuid_bucket, i),
GC_MARK_METADATA);
@@ -1705,36 +1723,16 @@ int bch_gc_thread_start(struct cache_set *c)
/* Initial partial gc */
-static int bch_btree_check_recurse(struct btree *b, struct btree_op *op,
- unsigned long **seen)
+static int bch_btree_check_recurse(struct btree *b, struct btree_op *op)
{
int ret = 0;
- unsigned i;
struct bkey *k, *p = NULL;
- struct bucket *g;
struct btree_iter iter;
- for_each_key_filter(&b->keys, k, &iter, bch_ptr_invalid) {
- for (i = 0; i < KEY_PTRS(k); i++) {
- if (!ptr_available(b->c, k, i))
- continue;
-
- g = PTR_BUCKET(b->c, k, i);
-
- if (!__test_and_set_bit(PTR_BUCKET_NR(b->c, k, i),
- seen[PTR_DEV(k, i)]) ||
- !ptr_stale(b->c, k, i)) {
- g->gen = PTR_GEN(k, i);
-
- if (b->level && bkey_cmp(k, &ZERO_KEY))
- g->prio = BTREE_PRIO;
- else if (!b->level && g->prio == BTREE_PRIO)
- g->prio = INITIAL_PRIO;
- }
- }
+ for_each_key_filter(&b->keys, k, &iter, bch_ptr_invalid)
+ bch_initial_mark_key(b->c, b->level, k);
- btree_mark_key(b, k);
- }
+ bch_initial_mark_key(b->c, b->level + 1, &b->key);
if (b->level) {
bch_btree_iter_init(&b->keys, &iter, NULL);
@@ -1746,40 +1744,22 @@ static int bch_btree_check_recurse(struct btree *b, struct btree_op *op,
btree_node_prefetch(b->c, k, b->level - 1);
if (p)
- ret = btree(check_recurse, p, b, op, seen);
+ ret = btree(check_recurse, p, b, op);
p = k;
} while (p && !ret);
}
- return 0;
+ return ret;
}
int bch_btree_check(struct cache_set *c)
{
- int ret = -ENOMEM;
- unsigned i;
- unsigned long *seen[MAX_CACHES_PER_SET];
struct btree_op op;
- memset(seen, 0, sizeof(seen));
bch_btree_op_init(&op, SHRT_MAX);
- for (i = 0; c->cache[i]; i++) {
- size_t n = DIV_ROUND_UP(c->cache[i]->sb.nbuckets, 8);
- seen[i] = kmalloc(n, GFP_KERNEL);
- if (!seen[i])
- goto err;
-
- /* Disables the seen array until prio_read() uses it too */
- memset(seen[i], 0xFF, n);
- }
-
- ret = btree_root(check_recurse, c, &op, seen);
-err:
- for (i = 0; i < MAX_CACHES_PER_SET; i++)
- kfree(seen[i]);
- return ret;
+ return btree_root(check_recurse, c, &op);
}
/* Btree insertion */
diff --git a/drivers/md/bcache/btree.h b/drivers/md/bcache/btree.h
index af065e9..def9dc4 100644
--- a/drivers/md/bcache/btree.h
+++ b/drivers/md/bcache/btree.h
@@ -251,7 +251,7 @@ int bch_gc_thread_start(struct cache_set *);
size_t bch_btree_gc_finish(struct cache_set *);
void bch_moving_gc(struct cache_set *);
int bch_btree_check(struct cache_set *);
-uint8_t __bch_btree_mark_key(struct cache_set *, int, struct bkey *);
+void bch_initial_mark_key(struct cache_set *, int, struct bkey *);
static inline void wake_up_gc(struct cache_set *c)
{
diff --git a/drivers/md/bcache/journal.c b/drivers/md/bcache/journal.c
index 4152a91..cf8e093 100644
--- a/drivers/md/bcache/journal.c
+++ b/drivers/md/bcache/journal.c
@@ -293,21 +293,12 @@ void bch_journal_mark(struct cache_set *c, struct list_head *list)
k < bset_bkey_last(&i->j);
k = bkey_next(k)) {
unsigned j;
- struct bucket *g;
- for (j = 0; j < KEY_PTRS(k); j++) {
- if (!ptr_available(c, k, j))
- continue;
+ for (j = 0; j < KEY_PTRS(k); j++)
+ if (ptr_available(c, k, j))
+ atomic_inc(&PTR_BUCKET(c, k, j)->pin);
- g = PTR_BUCKET(c, k, j);
- atomic_inc(&g->pin);
-
- if (g->prio == BTREE_PRIO &&
- !ptr_stale(c, k, j))
- g->prio = INITIAL_PRIO;
- }
-
- __bch_btree_mark_key(c, 0, k);
+ bch_initial_mark_key(c, 0, k);
}
}
}