diff --git a/newbrt/brt-cachetable-wrappers.c b/newbrt/brt-cachetable-wrappers.c
index 64f738b9721a8e34f60826d53779339f6936d484..f75f11f202044676cbe8f2e11cabf357aeff5127 100644
--- a/newbrt/brt-cachetable-wrappers.c
+++ b/newbrt/brt-cachetable-wrappers.c
@@ -128,9 +128,11 @@ toku_pin_brtnode(
const PIVOT_BOUNDS bounds,
BRTNODE_FETCH_EXTRA bfe,
BOOL apply_ancestor_messages, // this BOOL is probably temporary, for #3972, once we know how range query estimates work, will revisit this
- BRTNODE *node_p)
+ BRTNODE *node_p,
+ BOOL* msgs_applied)
{
void *node_v;
+ *msgs_applied = FALSE;
int r = toku_cachetable_get_and_pin_nonblocking(
brt->cf,
blocknum,
@@ -146,7 +148,7 @@ toku_pin_brtnode(
if (r==0) {
BRTNODE node = node_v;
if (apply_ancestor_messages) {
- maybe_apply_ancestors_messages_to_node(brt, node, ancestors, bounds);
+ maybe_apply_ancestors_messages_to_node(brt, node, ancestors, bounds, msgs_applied);
}
*node_p = node;
// printf("%*sPin %ld\n", 8-node->height, "", blocknum.b);
@@ -183,7 +185,8 @@ toku_pin_brtnode_holding_lock(
);
assert(r==0);
BRTNODE node = node_v;
- if (apply_ancestor_messages) maybe_apply_ancestors_messages_to_node(brt, node, ancestors, bounds);
+ BOOL msgs_applied;
+ if (apply_ancestor_messages) maybe_apply_ancestors_messages_to_node(brt, node, ancestors, bounds, &msgs_applied);
*node_p = node;
}
diff --git a/newbrt/brt-cachetable-wrappers.h b/newbrt/brt-cachetable-wrappers.h
index 847359a41f0c34bcfa1fcef5e9e6e40d6a17d808..8f96c093d7f1016ba1175fc8fcdc1057b9e70c61 100644
--- a/newbrt/brt-cachetable-wrappers.h
+++ b/newbrt/brt-cachetable-wrappers.h
@@ -72,7 +72,8 @@ toku_pin_brtnode(
const PIVOT_BOUNDS pbounds,
BRTNODE_FETCH_EXTRA bfe,
BOOL apply_ancestor_messages, // this BOOL is probably temporary, for #3972, once we know how range query estimates work, will revisit this
- BRTNODE *node_p
+ BRTNODE *node_p,
+ BOOL* msgs_applied
) __attribute__((__warn_unused_result__));
/**
diff --git a/newbrt/brt-internal.h b/newbrt/brt-internal.h
index 2845d5df456492cc31756a01f895f2795eed83a5..42efdd7155da7cc2b9f5e20842ce3e5e68c9ff8d 100644
--- a/newbrt/brt-internal.h
+++ b/newbrt/brt-internal.h
@@ -706,7 +706,7 @@ struct pivot_bounds {
};
// FIXME needs toku prefix
-void maybe_apply_ancestors_messages_to_node (BRT t, BRTNODE node, ANCESTORS ancestors, struct pivot_bounds const * const bounds);
+void maybe_apply_ancestors_messages_to_node (BRT t, BRTNODE node, ANCESTORS ancestors, struct pivot_bounds const * const bounds, BOOL* msgs_applied);
int
toku_brt_search_which_child(
diff --git a/newbrt/brt.c b/newbrt/brt.c
index 50125ebcf56ad274f431267f100b1e3d8541c8f0..0ae9f846c0b385a8ff7aae2451aa305a110d7058 100644
--- a/newbrt/brt.c
+++ b/newbrt/brt.c
@@ -4777,7 +4777,8 @@ bnc_apply_messages_to_basement_node(
BASEMENTNODE bn, // where to apply messages
BRTNODE ancestor, // the ancestor node where we can find messages to apply
int childnum, // which child buffer of ancestor contains messages we want
- struct pivot_bounds const * const bounds // contains pivot key bounds of this basement node
+ struct pivot_bounds const * const bounds, // contains pivot key bounds of this basement node
+ BOOL* msgs_applied
)
{
int r;
@@ -4830,6 +4831,7 @@ bnc_apply_messages_to_basement_node(
// Apply the messages in MSN order.
for (int i = 0; i < buffer_size; ++i) {
+ *msgs_applied = TRUE;
const struct fifo_entry *entry = toku_fifo_get_entry(bnc->buffer, offsets[i]);
do_brt_leaf_put_cmd(t, leafnode, bn, ancestor, childnum, entry);
}
@@ -4838,12 +4840,13 @@ bnc_apply_messages_to_basement_node(
} else if (stale_lbi == stale_ube) {
// No stale messages to apply, we just apply fresh messages.
struct iterate_do_brt_leaf_put_cmd_extra iter_extra = { .t = t, .leafnode = leafnode, .bn = bn, .ancestor = ancestor, .childnum = childnum };
-
+ if (fresh_ube - fresh_lbi > 0) *msgs_applied = TRUE;
r = toku_omt_iterate_on_range(bnc->fresh_message_tree, fresh_lbi, fresh_ube, iterate_do_brt_leaf_put_cmd, &iter_extra);
assert_zero(r);
} else if (fresh_lbi == fresh_ube) {
// No fresh messages to apply, we just apply stale messages.
+ if (stale_ube - stale_lbi > 0) *msgs_applied = TRUE;
struct iterate_do_brt_leaf_put_cmd_extra iter_extra = { .t = t, .leafnode = leafnode, .bn = bn, .ancestor = ancestor, .childnum = childnum };
r = toku_omt_iterate_on_range(bnc->stale_message_tree, stale_lbi, stale_ube, iterate_do_brt_leaf_put_cmd, &iter_extra);
@@ -4867,6 +4870,7 @@ bnc_apply_messages_to_basement_node(
// Iterate over both lists, applying the smaller (in (key, msn)
// order) message at each step
while (stale_i < stale_ube && fresh_i < fresh_ube) {
+ *msgs_applied = TRUE;
const long stale_offset = (long) stale_v;
const long fresh_offset = (long) fresh_v;
int c = toku_fifo_entry_key_msn_cmp(&extra, &stale_offset, &fresh_offset);
@@ -4944,7 +4948,7 @@ bnc_apply_messages_to_basement_node(
}
void
-maybe_apply_ancestors_messages_to_node (BRT t, BRTNODE node, ANCESTORS ancestors, struct pivot_bounds const * const bounds)
+maybe_apply_ancestors_messages_to_node (BRT t, BRTNODE node, ANCESTORS ancestors, struct pivot_bounds const * const bounds, BOOL* msgs_applied)
// Effect:
// Bring a leaf node up-to-date according to all the messages in the ancestors.
// If the leaf node is already up-to-date then do nothing.
@@ -4975,7 +4979,8 @@ maybe_apply_ancestors_messages_to_node (BRT t, BRTNODE node, ANCESTORS ancestors
curr_bn,
curr_ancestors->node,
curr_ancestors->childnum,
- &curr_bounds
+ &curr_bounds,
+ msgs_applied
);
// We don't want to check this ancestor node again if the
// next time we query it, the msn hasn't changed.
@@ -5186,6 +5191,7 @@ brt_node_maybe_prefetch(BRT brt, BRTNODE node, int childnum, BRT_CURSOR brtcurso
struct unlock_brtnode_extra {
BRT brt;
BRTNODE node;
+ BOOL msgs_applied;
};
// When this is called, the cachetable lock is held
static void
@@ -5194,7 +5200,13 @@ unlock_brtnode_fun (void *v) {
BRT brt = x->brt;
BRTNODE node = x->node;
// CT lock is held
- int r = toku_cachetable_unpin_ct_prelocked_no_flush(brt->cf, node->thisnodename, node->fullhash, (enum cachetable_dirty) node->dirty, make_brtnode_pair_attr(node));
+ int r = toku_cachetable_unpin_ct_prelocked_no_flush(
+ brt->cf,
+ node->thisnodename,
+ node->fullhash,
+ (enum cachetable_dirty) node->dirty,
+ x->msgs_applied ? make_brtnode_pair_attr(node) : make_invalid_pair_attr()
+ );
assert(r==0);
}
@@ -5221,24 +5233,24 @@ brt_search_child(BRT brt, BRTNODE node, int childnum, brt_search_t *search, BRT_
brtcursor->right_is_pos_infty,
brtcursor->disable_prefetching
);
+ BOOL msgs_applied = FALSE;
{
int rr = toku_pin_brtnode(brt, childblocknum, fullhash,
unlockers,
&next_ancestors, bounds,
&bfe,
TRUE,
- &childnode);
+ &childnode,
+ &msgs_applied);
if (rr==TOKUDB_TRY_AGAIN) return rr;
assert(rr==0);
}
- struct unlock_brtnode_extra unlock_extra = {brt,childnode};
+ struct unlock_brtnode_extra unlock_extra = {brt,childnode,msgs_applied};
struct unlockers next_unlockers = {TRUE, unlock_brtnode_fun, (void*)&unlock_extra, unlockers};
int r = brt_search_node(brt, childnode, search, bfe.child_to_read, getf, getf_v, doprefetch, brtcursor, &next_unlockers, &next_ancestors, bounds, can_bulk_fetch);
if (r!=TOKUDB_TRY_AGAIN) {
- // Even if r is reactive, we want to handle the maybe reactive child.
-
#if TOKU_DO_PREFETCH
// maybe prefetch the next child
if (r == 0 && node->height == 1) {
@@ -5247,7 +5259,12 @@ brt_search_child(BRT brt, BRTNODE node, int childnum, brt_search_t *search, BRT_
#endif
assert(next_unlockers.locked);
- toku_unpin_brtnode_read_only(brt, childnode); // unpin the childnode before handling the reactive child (because that may make the childnode disappear.)
+ if (msgs_applied) {
+ toku_unpin_brtnode(brt, childnode);
+ }
+ else {
+ toku_unpin_brtnode_read_only(brt, childnode);
+ }
} else {
// try again.
@@ -5258,7 +5275,12 @@ brt_search_child(BRT brt, BRTNODE node, int childnum, brt_search_t *search, BRT_
// some piece of a node that it needed was not in memory. In this case,
// the node was not unpinned, so we unpin it here
if (next_unlockers.locked) {
- toku_unpin_brtnode_read_only(brt, childnode);
+ if (msgs_applied) {
+ toku_unpin_brtnode(brt, childnode);
+ }
+ else {
+ toku_unpin_brtnode_read_only(brt, childnode);
+ }
}
}
@@ -5557,7 +5579,7 @@ try_again:
//uint tree_height = node->height + 1; // How high is the tree? This is the height of the root node plus one (leaf is at height 0).
- struct unlock_brtnode_extra unlock_extra = {brt,node};
+ struct unlock_brtnode_extra unlock_extra = {brt,node,FALSE};
struct unlockers unlockers = {TRUE, unlock_brtnode_fun, (void*)&unlock_extra, (UNLOCKERS)NULL};
{
@@ -6061,11 +6083,13 @@ toku_brt_keyrange_internal (BRT brt, BRTNODE node,
BLOCKNUM childblocknum = BP_BLOCKNUM(node, child_number);
u_int32_t fullhash = compute_child_fullhash(brt->cf, node, child_number);
BRTNODE childnode;
- r = toku_pin_brtnode(brt, childblocknum, fullhash, unlockers, &next_ancestors, bounds, bfe, FALSE, &childnode);
+ BOOL msgs_applied = FALSE;
+ r = toku_pin_brtnode(brt, childblocknum, fullhash, unlockers, &next_ancestors, bounds, bfe, FALSE, &childnode, &msgs_applied);
+ assert(!msgs_applied);
if (r != TOKUDB_TRY_AGAIN) {
assert(r == 0);
- struct unlock_brtnode_extra unlock_extra = {brt,childnode};
+ struct unlock_brtnode_extra unlock_extra = {brt,childnode,FALSE};
struct unlockers next_unlockers = {TRUE, unlock_brtnode_fun, (void*)&unlock_extra, unlockers};
const struct pivot_bounds next_bounds = next_pivot_keys(node, child_number, bounds);
@@ -6119,7 +6143,7 @@ toku_brt_keyrange (BRT brt, DBT *key, u_int64_t *less_p, u_int64_t *equal_p, u_i
toku_brtheader_release_treelock(brt->h);
}
- struct unlock_brtnode_extra unlock_extra = {brt,node};
+ struct unlock_brtnode_extra unlock_extra = {brt,node,FALSE};
struct unlockers unlockers = {TRUE, unlock_brtnode_fun, (void*)&unlock_extra, (UNLOCKERS)NULL};
{
diff --git a/newbrt/tests/orthopush-flush.c b/newbrt/tests/orthopush-flush.c
index 13f887c785f3b9ad9ab96722641e683d6fd6e3ec..61378d63c411cdd37b1fea5c9f2dd7ab6ca1bc50 100644
--- a/newbrt/tests/orthopush-flush.c
+++ b/newbrt/tests/orthopush-flush.c
@@ -620,7 +620,8 @@ flush_to_leaf(BRT t, bool make_leaf_up_to_date, bool use_flush) {
parentnode->max_msn_applied_to_node_on_disk = max_parent_msn;
struct ancestors ancestors = { .node = parentnode, .childnum = 0, .next = NULL };
const struct pivot_bounds infinite_bounds = { .lower_bound_exclusive = NULL, .upper_bound_inclusive = NULL };
- maybe_apply_ancestors_messages_to_node(t, child, &ancestors, &infinite_bounds);
+ BOOL* msgs_applied;
+ maybe_apply_ancestors_messages_to_node(t, child, &ancestors, &infinite_bounds, &msgs_applied);
FIFO_ITERATE(parent_bnc->buffer, key, keylen, val, vallen, type, msn, xids, is_fresh,
{
@@ -840,7 +841,8 @@ flush_to_leaf_with_keyrange(BRT t, bool make_leaf_up_to_date) {
parentnode->max_msn_applied_to_node_on_disk = max_parent_msn;
struct ancestors ancestors = { .node = parentnode, .childnum = 0, .next = NULL };
const struct pivot_bounds bounds = { .lower_bound_exclusive = NULL, .upper_bound_inclusive = kv_pair_malloc(childkeys[7].data, childkeys[7].size, NULL, 0) };
- maybe_apply_ancestors_messages_to_node(t, child, &ancestors, &bounds);
+ BOOL msgs_applied;
+ maybe_apply_ancestors_messages_to_node(t, child, &ancestors, &bounds, &msgs_applied);
FIFO_ITERATE(parent_bnc->buffer, key, keylen, val, vallen, type, msn, xids, is_fresh,
{
@@ -1024,7 +1026,8 @@ compare_apply_and_flush(BRT t, bool make_leaf_up_to_date) {
parentnode->max_msn_applied_to_node_on_disk = max_parent_msn;
struct ancestors ancestors = { .node = parentnode, .childnum = 0, .next = NULL };
const struct pivot_bounds infinite_bounds = { .lower_bound_exclusive = NULL, .upper_bound_inclusive = NULL };
- maybe_apply_ancestors_messages_to_node(t, child2, &ancestors, &infinite_bounds);
+ BOOL msgs_applied;
+ maybe_apply_ancestors_messages_to_node(t, child2, &ancestors, &infinite_bounds, &msgs_applied);
FIFO_ITERATE(parent_bnc->buffer, key, keylen, val, vallen, type, msn, xids, is_fresh,
{