MEDIUM: cache: Avoid going over duplicates lists too often

The secondary entry counter cannot be updated without going over all the
items of a duplicates list periodically. In order to avoid doing it too
often and to impact the cache's performances, a timestamp is added to
the cache_entry. It will store the timestamp (with second precision) of
the last iteration over the list (actually the last call of the
clear_expired_duplicates function). This way, this function will not be
called more than once per second for a given duplicates list.
This commit is contained in:
Remi Tricot-Le Breton 2020-12-10 17:58:42 +01:00 committed by William Lallemand
parent 65904e4f07
commit 73be796462

View File

@ -119,6 +119,7 @@ struct cache_entry {
unsigned int secondary_key_signature; /* Bitfield of the HTTP headers that should be used
* to build secondary keys for this cache entry. */
unsigned int secondary_entries_count; /* Should only be filled in the last entry of a list of dup entries */
unsigned int last_clear_ts; /* Timestamp of the last call to clear_expired_duplicates. */
unsigned int etag_length; /* Length of the ETag value (if one was found in the response). */
unsigned int etag_offset; /* Offset of the ETag value in the data buffer. */
@ -211,6 +212,37 @@ struct cache_entry *secondary_entry_exist(struct cache *cache, struct cache_entr
}
/*
* Remove all expired entries from a list of duplicates.
* Return the number of alive entries in the list and sets dup_tail to the
* current last item of the list.
*/
static unsigned int clear_expired_duplicates(struct eb32_node **dup_tail)
{
unsigned int entry_count = 0;
struct cache_entry *entry = NULL;
struct eb32_node *prev = *dup_tail;
struct eb32_node *tail = NULL;
while (prev) {
entry = container_of(prev, struct cache_entry, eb);
prev = eb32_prev_dup(prev);
if (entry->expire <= now.tv_sec) {
eb32_delete(&entry->eb);
entry->eb.key = 0;
}
else {
if (!tail)
tail = &entry->eb;
++entry_count;
}
}
*dup_tail = tail;
return entry_count;
}
/*
* This function inserts a cache_entry in the cache's ebtree. In case of
@ -227,6 +259,7 @@ static struct eb32_node *insert_entry(struct cache *cache, struct cache_entry *n
struct eb32_node *prev = NULL;
struct cache_entry *entry = NULL;
unsigned int entry_count = 0;
unsigned int last_clear_ts = now.tv_sec;
struct eb32_node *node = eb32_insert(&cache->entries, &new_entry->eb);
@ -241,18 +274,37 @@ static struct eb32_node *insert_entry(struct cache *cache, struct cache_entry *n
* number of entries in the list. */
entry = container_of(prev, struct cache_entry, eb);
entry_count = entry->secondary_entries_count;
last_clear_ts = entry->last_clear_ts;
if (entry_count >= SECONDARY_ENTRY_MAX_COUNT) {
/* Too many entries for this primary key, delete
* the newly inserted one. */
entry = container_of(prev, struct cache_entry, eb);
eb32_delete(node);
node->key = 0;
return NULL;
/* Some entries of the duplicate list might be expired so
* we will iterate over all the items in order to free some
* space. In order to avoid going over the same list too
* often, we first check the timestamp of the last check
* performed. */
if (last_clear_ts == now.tv_sec) {
/* Too many entries for this primary key, clear the
* one that was inserted. */
eb32_delete(node);
node->key = 0;
return NULL;
}
entry_count = clear_expired_duplicates(&prev);
if (entry_count >= SECONDARY_ENTRY_MAX_COUNT) {
/* Still too many entries for this primary key, delete
* the newly inserted one. */
entry = container_of(prev, struct cache_entry, eb);
entry->last_clear_ts = now.tv_sec;
eb32_delete(node);
node->key = 0;
return NULL;
}
}
}
new_entry->secondary_entries_count = entry_count + 1;
new_entry->last_clear_ts = last_clear_ts;
return node;
}