X-Git-Url: https://git.saurik.com/redis.git/blobdiff_plain/7dcc10b65e0075fccc90d93bac5b078baefdbb07..27fc5bf582b12a47e7497bc6f7a38a6d033dcb7c:/src/db.c

diff --git a/src/db.c b/src/db.c
index dc9ca8c5..e65106a5 100644
--- a/src/db.c
+++ b/src/db.c
@@ -10,28 +10,6 @@ void SlotToKeyDel(robj *key);
  * C-level DB API
  *----------------------------------------------------------------------------*/
 
-/* Important notes on lookup and disk store.
- *
- * When disk store is enabled on lookup we can have different cases.
- *
- * a) The key is in memory:
- *    - If the key is not in IO_SAVEINPROG state we can access it.
- *      As if it's just IO_SAVE this means we have the key in the IO queue
- *      but can't be accessed by the IO thread (it requires to be
- *      translated into an IO Job by the cache cron function.)
- *    - If the key is in IO_SAVEINPROG we can't touch the key and have
- *      to blocking wait completion of operations.
- * b) The key is not in memory:
- *    - If it's marked as non existing on disk as well (negative cache)
- *      we don't need to perform the disk access.
- *    - if the key MAY EXIST, but is not in memory, and it is marked as IO_SAVE
- *      then the key can only be a deleted one. As IO_SAVE keys are never
- *      evicted (dirty state), so the only possibility is that key was deleted.
- *    - if the key MAY EXIST we need to blocking load it.
- *      We check that the key is not in IO_SAVEINPROG state before accessing
- *      the disk object. If it is in this state, we wait.
- */
-
 robj *lookupKey(redisDb *db, robj *key) {
     dictEntry *de = dictFind(db->dict,key->ptr);
     if (de) {
@@ -40,19 +18,24 @@ robj *lookupKey(redisDb *db, robj *key) {
         /* Update the access time for the aging algorithm.
          * Don't do it if we have a saving child, as this will trigger
          * a copy on write madness. */
-        if (server.bgsavechildpid == -1 && server.bgrewritechildpid == -1)
+        if (server.rdb_child_pid == -1 && server.aof_child_pid == -1)
             val->lru = server.lruclock;
-        server.stat_keyspace_hits++;
         return val;
     } else {
-        server.stat_keyspace_misses++;
         return NULL;
     }
 }
 
 robj *lookupKeyRead(redisDb *db, robj *key) {
+    robj *val;
+
     expireIfNeeded(db,key);
-    return lookupKey(db,key);
+    val = lookupKey(db,key);
+    if (val == NULL)
+        server.stat_keyspace_misses++;
+    else
+        server.stat_keyspace_hits++;
+    return val;
 }
 
 robj *lookupKeyWrite(redisDb *db, robj *key) {
@@ -81,7 +64,6 @@ void dbAdd(redisDb *db, robj *key, robj *val) {
     int retval = dictAdd(db->dict, copy, val);
 
     redisAssertWithInfo(NULL,key,retval == REDIS_OK);
-    if (server.cluster_enabled) SlotToKeyAdd(key);
  }
 
 /* Overwrite an existing key with a new value. Incrementing the reference
@@ -110,7 +92,7 @@ void setKey(redisDb *db, robj *key, robj *val) {
     }
     incrRefCount(val);
     removeExpire(db,key);
-    touchWatchedKey(db,key);
+    signalModifiedKey(db,key);
 }
 
 int dbExists(redisDb *db, robj *key) {
@@ -149,15 +131,12 @@ int dbDelete(redisDb *db, robj *key) {
      * the key, because it is shared with the main dictionary. */
     if (dictSize(db->expires) > 0) dictDelete(db->expires,key->ptr);
     if (dictDelete(db->dict,key->ptr) == DICT_OK) {
-        if (server.cluster_enabled) SlotToKeyDel(key);
         return 1;
     } else {
         return 0;
     }
 }
 
-/* Empty the whole database.
- * If diskstore is enabled this function will just flush the in-memory cache. */
 long long emptyDb() {
     int j;
     long long removed = 0;
@@ -210,15 +189,15 @@ void flushallCommand(redisClient *c) {
     signalFlushedDb(-1);
     server.dirty += emptyDb();
     addReply(c,shared.ok);
-    if (server.bgsavechildpid != -1) {
-        kill(server.bgsavechildpid,SIGKILL);
-        rdbRemoveTempFile(server.bgsavechildpid);
+    if (server.rdb_child_pid != -1) {
+        kill(server.rdb_child_pid,SIGKILL);
+        rdbRemoveTempFile(server.rdb_child_pid);
     }
     if (server.saveparamslen > 0) {
         /* Normally rdbSave() will reset dirty, but we don't want this here
          * as otherwise FLUSHALL will not be replicated nor put into the AOF. */
         int saved_dirty = server.dirty;
-        rdbSave(server.dbfilename);
+        rdbSave(server.rdb_filename);
         server.dirty = saved_dirty;
     }
     server.dirty++;
@@ -249,10 +228,6 @@ void existsCommand(redisClient *c) {
 void selectCommand(redisClient *c) {
     int id = atoi(c->argv[1]->ptr);
 
-    if (server.cluster_enabled && id != 0) {
-        addReplyError(c,"SELECT is not allowed in cluster mode");
-        return;
-    }
     if (selectDb(c,id) == REDIS_ERR) {
         addReplyError(c,"invalid DB index");
     } else {
@@ -280,7 +255,7 @@ void keysCommand(redisClient *c) {
     unsigned long numkeys = 0;
     void *replylen = addDeferredMultiBulkLength(c);
 
-    di = dictGetIterator(c->db->dict);
+    di = dictGetSafeIterator(c->db->dict);
     allkeys = (pattern[0] == '*' && pattern[1] == '\0');
     while((de = dictNext(di)) != NULL) {
         sds key = dictGetKey(de);
@@ -328,8 +303,22 @@ void typeCommand(redisClient *c) {
 }
 
 void shutdownCommand(redisClient *c) {
-    if (prepareForShutdown() == REDIS_OK)
-        exit(0);
+    int flags = 0;
+
+    if (c->argc > 2) {
+        addReply(c,shared.syntaxerr);
+        return;
+    } else if (c->argc == 2) {
+        if (!strcasecmp(c->argv[1]->ptr,"nosave")) {
+            flags |= REDIS_SHUTDOWN_NOSAVE;
+        } else if (!strcasecmp(c->argv[1]->ptr,"save")) {
+            flags |= REDIS_SHUTDOWN_SAVE;
+        } else {
+            addReply(c,shared.syntaxerr);
+            return;
+        }
+    }
+    if (prepareForShutdown(flags) == REDIS_OK) exit(0);
     addReplyError(c,"Errors trying to SHUTDOWN. Check logs.");
 }
 
@@ -379,11 +368,6 @@ void moveCommand(redisClient *c) {
     redisDb *src, *dst;
     int srcid;
 
-    if (server.cluster_enabled) {
-        addReplyError(c,"MOVE is not allowed in cluster mode");
-        return;
-    }
-
     /* Obtain source and target DB pointers */
     src = c->db;
     srcid = c->db->id;
@@ -469,11 +453,12 @@ long long getExpire(redisDb *db, robj *key) {
 void propagateExpire(redisDb *db, robj *key) {
     robj *argv[2];
 
-    argv[0] = createStringObject("DEL",3);
+    argv[0] = shared.del;
     argv[1] = key;
-    incrRefCount(key);
+    incrRefCount(argv[0]);
+    incrRefCount(argv[1]);
 
-    if (server.appendonly)
+    if (server.aof_state != REDIS_AOF_OFF)
         feedAppendOnlyFile(server.delCommand,db->id,argv,2);
     if (listLength(server.slaves))
         replicationFeedSlaves(server.slaves,db->id,argv,2);
@@ -498,7 +483,7 @@ int expireIfNeeded(redisDb *db, robj *key) {
      * that is, 0 if we think the key should be still valid, 1 if
      * we think the key is expired at this time. */
     if (server.masterhost != NULL) {
-        return time(NULL) > when;
+        return mstime() > when;
     }
 
     /* Return when this key has not expired */
@@ -514,28 +499,23 @@ int expireIfNeeded(redisDb *db, robj *key) {
  * Expires Commands
  *----------------------------------------------------------------------------*/
 
-void expireGenericCommand(redisClient *c, long long offset) {
+/* This is the generic command implementation for EXPIRE, PEXPIRE, EXPIREAT
+ * and PEXPIREAT. Because the commad second argument may be relative or absolute
+ * the "basetime" argument is used to signal what the base time is (either 0
+ * for *AT variants of the command, or the current time for relative expires).
+ *
+ * unit is either UNIT_SECONDS or UNIT_MILLISECONDS, and is only used for
+ * the argv[2] parameter. The basetime is always specified in milliesconds. */
+void expireGenericCommand(redisClient *c, long long basetime, int unit) {
     dictEntry *de;
     robj *key = c->argv[1], *param = c->argv[2];
-    long long milliseconds;
-    int time_in_seconds = 1;
+    long long when; /* unix time in milliseconds when the key will expire. */
 
-    if (getLongLongFromObjectOrReply(c, param, &milliseconds, NULL) != REDIS_OK)
+    if (getLongLongFromObjectOrReply(c, param, &when, NULL) != REDIS_OK)
         return;
 
-    /* If no "ms" argument was passed the time is in second, so we need
-     * to multilpy it by 1000 */
-    if (c->argc == 4) {
-        char *arg = c->argv[3]->ptr;
-
-        if (tolower(arg[0]) != 'm' || tolower(arg[1]) != 's' || arg[2]) {
-            addReply(c,shared.syntaxerr);
-            return;
-        }
-        time_in_seconds = 0; /* "ms" argument passed. */
-    }
-    if (time_in_seconds) milliseconds *= 1000;
-    milliseconds -= offset;
+    if (unit == UNIT_SECONDS) when *= 1000;
+    when += basetime;
 
     de = dictFind(c->db->dict,key->ptr);
     if (de == NULL) {
@@ -548,7 +528,7 @@ void expireGenericCommand(redisClient *c, long long offset) {
      *
      * Instead we take the other branch of the IF statement setting an expire
      * (possibly in the past) and wait for an explicit DEL from the master. */
-    if (milliseconds <= 0 && !server.loading && !server.masterhost) {
+    if (when <= mstime() && !server.loading && !server.masterhost) {
         robj *aux;
 
         redisAssertWithInfo(c,key,dbDelete(c->db,key));
@@ -562,7 +542,6 @@ void expireGenericCommand(redisClient *c, long long offset) {
         addReply(c, shared.cone);
         return;
     } else {
-        long long when = mstime()+milliseconds;
         setExpire(c->db,key,when);
         addReply(c,shared.cone);
         signalModifiedKey(c->db,key);
@@ -572,14 +551,22 @@ void expireGenericCommand(redisClient *c, long long offset) {
 }
 
 void expireCommand(redisClient *c) {
-    expireGenericCommand(c,0);
+    expireGenericCommand(c,mstime(),UNIT_SECONDS);
 }
 
 void expireatCommand(redisClient *c) {
-    expireGenericCommand(c,mstime());
+    expireGenericCommand(c,0,UNIT_SECONDS);
 }
 
-void ttlCommand(redisClient *c) {
+void pexpireCommand(redisClient *c) {
+    expireGenericCommand(c,mstime(),UNIT_MILLISECONDS);
+}
+
+void pexpireatCommand(redisClient *c) {
+    expireGenericCommand(c,0,UNIT_MILLISECONDS);
+}
+
+void ttlGenericCommand(redisClient *c, int output_ms) {
     long long expire, ttl = -1;
 
     expire = getExpire(c->db,c->argv[1]);
@@ -590,10 +577,18 @@ void ttlCommand(redisClient *c) {
     if (ttl == -1) {
         addReplyLongLong(c,-1);
     } else {
-        addReplyLongLong(c,(ttl+500)/1000);
+        addReplyLongLong(c,output_ms ? ttl : ((ttl+500)/1000));
     }
 }
 
+void ttlCommand(redisClient *c) {
+    ttlGenericCommand(c, 0);
+}
+
+void pttlCommand(redisClient *c) {
+    ttlGenericCommand(c, 1);
+}
+
 void persistCommand(redisClient *c) {
     dictEntry *de;
 
@@ -682,35 +677,3 @@ int *zunionInterGetKeys(struct redisCommand *cmd,robj **argv, int argc, int *num
     *numkeys = num;
     return keys;
 }
-
-/* Slot to Key API. This is used by Redis Cluster in order to obtain in
- * a fast way a key that belongs to a specified hash slot. This is useful
- * while rehashing the cluster. */
-void SlotToKeyAdd(robj *key) {
-    unsigned int hashslot = keyHashSlot(key->ptr,sdslen(key->ptr));
-
-    zslInsert(server.cluster.slots_to_keys,hashslot,key);
-    incrRefCount(key);
-}
-
-void SlotToKeyDel(robj *key) {
-    unsigned int hashslot = keyHashSlot(key->ptr,sdslen(key->ptr));
-
-    zslDelete(server.cluster.slots_to_keys,hashslot,key);
-}
-
-unsigned int GetKeysInSlot(unsigned int hashslot, robj **keys, unsigned int count) {
-    zskiplistNode *n;
-    zrangespec range;
-    int j = 0;
-
-    range.min = range.max = hashslot;
-    range.minex = range.maxex = 0;
-    
-    n = zslFirstInRange(server.cluster.slots_to_keys, range);
-    while(n && n->score == hashslot && count--) {
-        keys[j++] = n->obj;
-        n = n->level[0].forward;
-    }
-    return j;
-}