diff options
| author | unknown <heikki@hundin.mysql.fi> | 2004-10-27 15:45:10 +0300 |
|---|---|---|
| committer | unknown <heikki@hundin.mysql.fi> | 2004-10-27 15:45:10 +0300 |
| commit | f41bba8c6156a7adf4c67dfa75e16112767a5d3c (patch) | |
| tree | be92bb32c7a2c9e1f63fd51e686e90bea155b26b /innobase/buf | |
| parent | 8b44be63c5279ebfd5fb4f7c8d209c2adb2d8697 (diff) | |
| download | mariadb-git-f41bba8c6156a7adf4c67dfa75e16112767a5d3c.tar.gz | |
btr0sea.c, buf0lru.c, buf0buf.c, ha0ha.c, hash0hash.h, ha0ha.h, buf0buf.h:
Link adaptive hash index entries to the buffer page, so that we can remove them quickly without knowing the record structure on that page; this was requested by Marko for the compact InnoDB table format; note that the adaptive hash index memory overhead grows by 67 %, maybe we have to tune this later somehow
innobase/include/buf0buf.h:
Link adaptive hash index entries to the buffer page, so that we can remove them quickly without knowing the record structure on that page; this was requested by Marko for the compact InnoDB table format; note that the adaptive hash index overhead grows by 67 %, maybe we have to tune this later somehow
innobase/include/ha0ha.h:
Link adaptive hash index entries to the buffer page, so that we can remove them quickly without knowing the record structure on that page; this was requested by Marko for the compact InnoDB table format; note that the adaptive hash index overhead grows by 67 %, maybe we have to tune this later somehow
innobase/include/hash0hash.h:
Link adaptive hash index entries to the buffer page, so that we can remove them quickly without knowing the record structure on that page; this was requested by Marko for the compact InnoDB table format; note that the adaptive hash index overhead grows by 67 %, maybe we have to tune this later somehow
innobase/ha/ha0ha.c:
Link adaptive hash index entries to the buffer page, so that we can remove them quickly without knowing the record structure on that page; this was requested by Marko for the compact InnoDB table format; note that the adaptive hash index overhead grows by 67 %, maybe we have to tune this later somehow
innobase/buf/buf0buf.c:
Link adaptive hash index entries to the buffer page, so that we can remove them quickly without knowing the record structure on that page; this was requested by Marko for the compact InnoDB table format; note that the adaptive hash index overhead grows by 67 %, maybe we have to tune this later somehow
innobase/buf/buf0lru.c:
Link adaptive hash index entries to the buffer page, so that we can remove them quickly without knowing the record structure on that page; this was requested by Marko for the compact InnoDB table format; note that the adaptive hash index overhead grows by 67 %, maybe we have to tune this later somehow
innobase/btr/btr0sea.c:
Link adaptive hash index entries to the buffer page, so that we can remove them quickly without knowing the record structure on that page; this was requested by Marko for the compact InnoDB table format; note that the adaptive hash index overhead grows by 67 %, maybe we have to tune this later somehow
Diffstat (limited to 'innobase/buf')
| -rw-r--r-- | innobase/buf/buf0buf.c | 1 | ||||
| -rw-r--r-- | innobase/buf/buf0lru.c | 1 |
2 files changed, 2 insertions, 0 deletions
diff --git a/innobase/buf/buf0buf.c b/innobase/buf/buf0buf.c index 5ec8998473d..bee322a1631 100644 --- a/innobase/buf/buf0buf.c +++ b/innobase/buf/buf0buf.c @@ -465,6 +465,7 @@ buf_block_init( block->in_LRU_list = FALSE; block->n_pointers = 0; + block->hash_nodes = NULL; rw_lock_create(&(block->lock)); ut_ad(rw_lock_validate(&(block->lock))); diff --git a/innobase/buf/buf0lru.c b/innobase/buf/buf0lru.c index 796311f0157..21dd0e304eb 100644 --- a/innobase/buf/buf0lru.c +++ b/innobase/buf/buf0lru.c @@ -789,6 +789,7 @@ buf_LRU_block_free_non_file_page( || (block->state == BUF_BLOCK_READY_FOR_USE)); ut_a(block->n_pointers == 0); + ut_a(block->hash_nodes == NULL); ut_a(!block->in_free_list); block->state = BUF_BLOCK_NOT_USED; |