/* Write the buffer (possibly composed of multiple blocks) into the specified
* fd. If a short write or any other error happens -1 is returned,
* otherwise the number of bytes written is returned. */
ssize_t aofRewriteBufferWrite(int fd) {
listNode *ln;
listIter li;
ssize_t count = 0;
listRewind(server.aof_rewrite_buf_blocks,&li);
while((ln = listNext(&li))) {
aofrwblock *block = listNodeValue(ln);
ssize_t nwritten;
if (block->used) {
nwritten = write(fd,block->buf,block->used);
if (nwritten != block->used) {
if (nwritten == 0) errno = EIO;
return -1;
}
count += nwritten;
}
}
return count;
}
//监视机制触发见touchWatchedKey决定是否触发REDIS_DIRTY_CAS 取消事物函数更具watch的键是否有触发REDIS_DIRTY_CAS来决定是否继续执行事物中的命令，见execCommand
//取消watch见unwatchAllKeys
void touchWatchedKey(redisDb *db, robj *key) {
list *clients;
listIter li;
listNode *ln;
// 字典为空，没有任何键被监视
if (dictSize(db->watched_keys) == 0) return;
// 获取所有监视这个键的客户端
clients = dictFetchValue(db->watched_keys, key);
if (!clients) return;
/* Mark all the clients watching this key as REDIS_DIRTY_CAS */
/* Check if we are already watching for this key */
// 遍历所有客户端，打开他们的 REDIS_DIRTY_CAS 标识
listRewind(clients,&li);
while((ln = listNext(&li))) {
redisClient *c = listNodeValue(ln);
c->flags |= REDIS_DIRTY_CAS;
}
}
// WATCH 某个 KEY
void watchForKey(redisClient *c, robj *key) {
list *clients = NULL;
listIter li;
listNode *ln;
watchedKey *wk;
/* Check if we are already watching for this key */
// 所有被 WATCHED 的 KEY 都被放在 redisClient.watched_keys 链表中
// 遍历这个链表，查看这个 KEY 是否已经处于监视状态（WATCHED）
listRewind(c->watched_keys,&li);
while((ln = listNext(&li))) {
wk = listNodeValue(ln);
if (wk->db == c->db && equalStringObjects(key,wk->key))
return; /* Key already watched */
}
/* This key is not already watched in this DB. Let's add it */
// 如果 KEY 还没有被 WATCH 过，那么对它进行 WATCH
clients = dictFetchValue(c->db->watched_keys,key);
if (!clients) {
// 如果 clients 链表不存在
// 说明这个客户端是第一个监视这个 DB 的这个 KEY 的客户端
// 那么 clients 创建链表，并将它添加到 c->db->watched_keys 字典中
clients = listCreate();
dictAdd(c->db->watched_keys,key,clients);
incrRefCount(key);
}
// 将客户端添加到 clients 链表
listAddNodeTail(clients,c);
/* Add the new key to the lits of keys watched by this client */
// 除了 c->db->watched_keys 之外
// 还要将被 WATCH 的 KEY 添加到 c->watched_keys
wk = zmalloc(sizeof(*wk));
wk->key = key;
wk->db = c->db;
incrRefCount(key);
listAddNodeTail(c->watched_keys,wk);
}
void _msg_dump(const char *file, int line, struct msg *msg, int level, int begin)
{
struct mbuf *mbuf;
listIter *iter;
listNode *node;
uint8_t *p, *q;
long int len;
if (log_loggable(level) == 0) {
return;
}
_log(file, line, 0, "msg dump id %"PRIu64" request %d len %"PRIu32" type %d "
"(err %d) kind %d result %d mbuf_count %u keys_count %u",
msg->id, msg->request, msg->mlen,
msg->type, msg->err, msg->kind,
msg->result, listLength(msg->data),
msg->keys == NULL?0:array_n(msg->keys));
iter = listGetIterator(msg->data, AL_START_HEAD);
ASSERT(iter != NULL);
while((node = listNext(iter)) != NULL) {
mbuf = listNodeValue(node);
if (begin) {
p = mbuf->start;
} else {
p = mbuf->pos;
}
q = mbuf->last;
len = q - p;
_log(file, line, 0, "mbuf [%p] with %ld bytes of data, pos %p last %p",
p, len, mbuf->pos, mbuf->last);
_log_hexdump(file, line, p, len, NULL);
}
listReleaseIterator(iter);
}
void lindexCommand(redisClient *c) {
robj *o = lookupKeyReadOrReply(c,c->argv[1],shared.nullbulk);
if (o == NULL || checkType(c,o,REDIS_LIST)) return;
long index;
robj *value = NULL;
if ((getLongFromObjectOrReply(c, c->argv[2], &index, NULL) != REDIS_OK))
return;
if (o->encoding == REDIS_ENCODING_ZIPLIST) {
unsigned char *p;
unsigned char *vstr;
unsigned int vlen;
long long vlong;
p = ziplistIndex(o->ptr,index);
if (ziplistGet(p,&vstr,&vlen,&vlong)) {
if (vstr) {
value = createStringObject((char*)vstr,vlen);
} else {
value = createStringObjectFromLongLong(vlong);
}
addReplyBulk(c,value);
decrRefCount(value);
} else {
addReply(c,shared.nullbulk);
}
} else if (o->encoding == REDIS_ENCODING_LINKEDLIST) {
listNode *ln = listIndex(o->ptr,index);
if (ln != NULL) {
value = listNodeValue(ln);
addReplyBulk(c,value);
} else {
addReply(c,shared.nullbulk);
}
} else {
redisPanic("Unknown list encoding");
}
}
/*
* 取消所有该客户端监视的 key
* 对事务状态的清除由调用者执行
*
* T = O(N^2)
*/
void unwatchAllKeys(redisClient *c)
{
listIter li;
listNode *ln;
// 没有键被 watch ，直接返回
if (listLength(c->watched_keys) == 0) return;
// 从客户端以及 DB 中删除所有监视 key 和客户端的资料
// O(N^2)
listRewind(c->watched_keys,&li);
while((ln = listNext(&li)))
{
list *clients;
watchedKey *wk;
/* Lookup the watched key -> clients list and remove the client
* from the list */
// 取出 watchedKey 结构
wk = listNodeValue(ln);
// 删除 db 中的客户端信息, O(1)
clients = dictFetchValue(wk->db->watched_keys, wk->key);
redisAssertWithInfo(c,NULL,clients != NULL);
// O(N)
listDelNode(clients,listSearchKey(clients,c));
/* Kill the entry at all if this was the only client */
if (listLength(clients) == 0)
dictDelete(wk->db->watched_keys, wk->key);
/* Remove this watched key from the client->watched list */
// 将 key 从客户端的监视列表中删除, O(1)
listDelNode(c->watched_keys,ln);
decrRefCount(wk->key);
zfree(wk);
}
}
/* result buffer aggregation is taken from scripting.c */
sds fakeClientResultBuffer(redisClient *c) {
sds reply;
if (listLength(c->reply) == 0 && c->bufpos < REDIS_REPLY_CHUNK_BYTES) {
/* This is a fast path for the common case of a reply inside the
* client static buffer. Don't create an SDS string but just use
* the client buffer directly. */
c->buf[c->bufpos] = '\0';
reply = c->buf;
c->bufpos = 0;
} else {
reply = sdsnewlen(c->buf, c->bufpos);
c->bufpos = 0;
while (listLength(c->reply)) {
robj *o = listNodeValue(listFirst(c->reply));
reply = sdscatlen(reply, o->ptr, sdslen(o->ptr));
listDelNode(c->reply, listFirst(c->reply));
}
}
return reply;
}
/* Unwatch all the keys watched by this client. To clean the EXEC dirty
* flag is up to the caller.
*
* 取消客户端对所有键的监视。
*
* 清除客户端事务状态的任务由调用者执行。
*/
void unwatchAllKeys(redisClient *c) {
listIter li;
listNode *ln;
// 没有键被监视，直接返回
if (listLength(c->watched_keys) == 0) return;
// 遍历链表中所有被客户端监视的键
listRewind(c->watched_keys,&li);
while((ln = listNext(&li))) {
list *clients;
watchedKey *wk;
/* Lookup the watched key -> clients list and remove the client
* from the list */
// 从数据库的 watched_keys 字典的 key 键中
// 删除链表里包含的客户端节点
wk = listNodeValue(ln);
// 取出客户端链表
clients = dictFetchValue(wk->db->watched_keys, wk->key);
redisAssertWithInfo(c, NULL, clients != NULL);
// 删除链表中的客户端节点
listDelNode(clients,listSearchKey(clients,c));
/* Kill the entry at all if this was the only client */
// 如果链表已经被清空，那么删除这个键
if (listLength(clients) == 0)
dictDelete(wk->db->watched_keys, wk->key);
/* Remove this watched key from the client->watched list */
// 从链表中移除 key 节点
listDelNode(c->watched_keys,ln);
decrRefCount(wk->key);
zfree(wk);
}
}
struct client *createClient(int fd, char *ip, int port, struct protocol *p)
{
struct client *c;
struct listNode *node;
node = listFirst(FreeClients);
if (node == NULL) {
c = wmalloc(sizeof(*c));
} else {
c = listNodeValue(node);
removeListNode(FreeClients, node);
appendToListTail(Clients, c);
}
if (c == NULL)
return NULL;
c->clifd = fd;
c->ip = wstrNew(ip);
c->port = port;
c->protocol = p;
c->conns = createList();
listSetFree(c->conns, (void (*)(void*))connDealloc);
c->req_buf = msgCreate(Server.mbuf_size);
c->is_outer = 1;
c->should_close = 0;
c->valid = 1;
c->pending = NULL;
c->client_data = NULL;
c->notify = NULL;
c->last_io = Server.cron_time;
c->name = wstrEmpty();
createEvent(WorkerProcess->center, c->clifd, EVENT_READABLE,
handleRequest, c);
getStatVal(StatTotalClient)++;
return c;
}
ssize_t aofRewriteBufferWrite(int fd){
/*write cotents in AOF rewrite buffer blocks into given file*/
/*return -1 for short count or write error, else return bytes written*/
listIter *li;
listNode *ln;
aofrwblock *block;
int nwritten, nbytes=0;
li = listGetIterator(server.aof_rewrite_buf_blocks, AL_START_HEAD);
while((ln = listNext(li)) != NULL){
block = listNodeValue(ln);
if((nwritten = write(fd, block->buf, block->used)) == -1){
return -1;
}
else if((nwritten == 0) && (block.used != 0)){
errno = EIO;
return 0;
}
nbytes += nwritten;
}
listReleaseIterator(li);
return nbytes;
}
/* Watch for the specified key */
void watchForKey(redisClient *c, robj *key) {
list *clients = NULL;
listIter li;
listNode *ln;
watchedKey *wk;
// 是否已经监视该键值
/* Check if we are already watching for this key */
listRewind(c->watched_keys,&li);
while((ln = listNext(&li))) {
wk = listNodeValue(ln);
if (wk->db == c->db && equalStringObjects(key,wk->key))
return; /* Key already watched */
}
// 获取监视该键值的客户端链表
/* This key is not already watched in this DB. Let's add it */
clients = dictFetchValue(c->db->watched_keys,key);
// 如果不存在链表，需要新建一个
if (!clients) {
clients = listCreate();
dictAdd(c->db->watched_keys,key,clients);
incrRefCount(key);
}
// 尾插法。将客户端添加到链表尾部
listAddNodeTail(clients,c);
// 将监视键添加到 redisClient.watched_keys 的尾部
/* Add the new key to the list of keys watched by this client */
wk = zmalloc(sizeof(*wk));
wk->key = key;
wk->db = c->db;
incrRefCount(key);
listAddNodeTail(c->watched_keys,wk);
}
/* Return entry or NULL at the current position of the iterator. */
robj *listTypeGet(listTypeEntry *entry) {
listTypeIterator *li = entry->li;
robj *value = NULL;
if (li->encoding == REDIS_ENCODING_ZIPLIST) {
unsigned char *vstr;
unsigned int vlen;
long long vlong;
redisAssert(entry->zi != NULL);
if (ziplistGet(entry->zi,&vstr,&vlen,&vlong)) {
if (vstr) {
value = createStringObject((char*)vstr,vlen);
} else {
value = createStringObjectFromLongLong(vlong);
}
}
} else if (li->encoding == REDIS_ENCODING_LINKEDLIST) {
redisAssert(entry->ln != NULL);
value = listNodeValue(entry->ln);
incrRefCount(value);
} else {
redisPanic("Unknown list encoding");
}
return value;
}
// 让client监视所有的指定的key
void watchForKey(client *c, robj *key) {
list *clients = NULL;
listIter li;
listNode *ln;
watchedKey *wk;
/* Check if we are already watching for this key */
listRewind(c->watched_keys,&li);
// 遍历客户端监视的键的链表，检查是否已经监视了指定的键
while((ln = listNext(&li))) {
wk = listNodeValue(ln);
// 如果键已经被监视，则直接返回
if (wk->db == c->db && equalStringObjects(key,wk->key))
return; /* Key already watched */
}
/* This key is not already watched in this DB. Let's add it */
// 如果数据库中该键没有被client监视则添加它
clients = dictFetchValue(c->db->watched_keys,key);
// 没有被client监视
if (!clients) {
// 创建一个空链表
clients = listCreate();
// 值是被client监控的key，键是client，添加到数据库的watched_keys字典中
dictAdd(c->db->watched_keys,key,clients);
incrRefCount(key);
}
// 将当前client添加到监视该key的client链表的尾部
listAddNodeTail(clients,c);
/* Add the new key to the list of keys watched by this client */
// 将新的被监视的key和与该key关联的数据库加入到客户端的watched_keys中
wk = zmalloc(sizeof(*wk));
wk->key = key;
wk->db = c->db;
incrRefCount(key);
listAddNodeTail(c->watched_keys,wk);
}
void _masterProcessFinishedIO() {
sds key = NULL;
sds content = NULL;
/* For each IO worker */
int tid = 0;
/* Polling all io thread */
for(tid=0;tid<CCACHE_NUM_BIO_THREADS;tid++) {
while(bioGetResult(tid,&key,&content))
{
master_numjob++;
objSds *value = dictFetchValue(master_cache,key);
if(content == NULL) {
value->ptr = HTTP_NOT_FOUND->ptr;
}
else {
value->ptr = content;
master_total_mem += sdslen(content);
}
value->state = OBJSDS_OK;
listIter li;
listNode *ln;
cacheEntry *ce;
/* notify all cache entries waiting for this key */
listRewind(value->waiting_entries,&li);
/* for each waiting ce */
while ((ln = listNext(&li)) != NULL){
/* unwatch client */
ce = listNodeValue(ln);
ce->val = value->ptr;
/* notify all clients waiting for this entry */
cacheSendMessage(ce->mycache,ce,CACHE_REPLY_NEW);
printf("Cache in Worker %.2lf \n", (double)(clock()));
}
}
}
}
robj *listTypePop(robj *subject, int where) {
robj *value = NULL;
if (subject->encoding == REDIS_ENCODING_ZIPLIST) {
unsigned char *p;
unsigned char *vstr;
unsigned int vlen;
long long vlong;
int pos = (where == REDIS_HEAD) ? 0 : -1;
p = ziplistIndex(subject->ptr,pos);
if (ziplistGet(p,&vstr,&vlen,&vlong)) {
if (vstr) {
value = createStringObject((char*)vstr,vlen);
} else {
value = createStringObjectFromLongLong(vlong);
}
/* We only need to delete an element when it exists */
subject->ptr = ziplistDelete(subject->ptr,&p);
}
} else if (subject->encoding == REDIS_ENCODING_LINKEDLIST) {
list *list = subject->ptr;
listNode *ln;
if (where == REDIS_HEAD) {
ln = listFirst(list);
} else {
ln = listLast(list);
}
if (ln != NULL) {
value = listNodeValue(ln);
incrRefCount(value);
listDelNode(list,ln);
}
} else {
redisPanic("Unknown list encoding");
}
return value;
}
/* Watch for the specified key
*
* 让客户端 c 监视给定的键 key
*/
void watchForKey(redisClient *c, robj *key) {
list *clients = NULL;
listIter li;
listNode *ln;
watchedKey *wk;
/* Check if we are already watching for this key */
// 检查 key 是否已经保存在 watched_keys 链表中，
// 如果是的话，直接返回
listRewind(c->watched_keys,&li);
while((ln = listNext(&li))) {
wk = listNodeValue(ln);
if (wk->db == c->db && equalStringObjects(key,wk->key))
return; /* Key already watched */
}
// 键不存在于 watched_keys ，添加它
// 以下是一个 key 不存在于字典的例子：
// before :
// {
// 'key-1' : [c1, c2, c3],
// 'key-2' : [c1, c2],
// }
// after c-10086 WATCH key-1 and key-3:
// {
// 'key-1' : [c1, c2, c3, c-10086],
// 'key-2' : [c1, c2],
// 'key-3' : [c-10086]
// }
/* This key is not already watched in this DB. Let's add it */
// 检查 key 是否存在于数据库的 watched_keys 字典中
clients = dictFetchValue(c->db->watched_keys,key);
// 如果不存在的话，添加它
if (!clients) {
// 值为链表
clients = listCreate();
// 关联键值对到字典
dictAdd(c->db->watched_keys,key,clients);
incrRefCount(key);
}
// 将客户端添加到链表的末尾
listAddNodeTail(clients,c);
/* Add the new key to the list of keys watched by this client */
// 将新 watchedKey 结构添加到客户端 watched_keys 链表的表尾
// 以下是一个添加 watchedKey 结构的例子
// before:
// [
// {
// 'key': 'key-1',
// 'db' : 0
// }
// ]
// after client watch key-123321 in db 0:
// [
// {
// 'key': 'key-1',
// 'db' : 0
// }
// ,
// {
// 'key': 'key-123321',
// 'db': 0
// }
// ]
wk = zmalloc(sizeof(*wk));
wk->key = key;
wk->db = c->db;
incrRefCount(key);
listAddNodeTail(c->watched_keys,wk);
}
void sendReplyToClientWritev(aeEventLoop *el, int fd, void *privdata, int mask)
{
redisClient *c = privdata;
int nwritten = 0, totwritten = 0, objlen, willwrite;
robj *o;
struct iovec iov[REDIS_WRITEV_IOVEC_COUNT];
int offset, ion = 0;
REDIS_NOTUSED(el);
REDIS_NOTUSED(mask);
listNode *node;
while (listLength(c->reply)) {
offset = c->sentlen;
ion = 0;
willwrite = 0;
/* fill-in the iov[] array */
for(node = listFirst(c->reply); node; node = listNextNode(node)) {
o = listNodeValue(node);
objlen = sdslen(o->ptr);
if (totwritten + objlen - offset > REDIS_MAX_WRITE_PER_EVENT)
break;
if(ion == REDIS_WRITEV_IOVEC_COUNT)
break; /* no more iovecs */
iov[ion].iov_base = ((char*)o->ptr) + offset;
iov[ion].iov_len = objlen - offset;
willwrite += objlen - offset;
offset = 0; /* just for the first item */
ion++;
}
if(willwrite == 0)
break;
/* write all collected blocks at once */
if((nwritten = writev(fd, iov, ion)) < 0) {
if (errno != EAGAIN) {
redisLog(REDIS_VERBOSE,
"Error writing to client: %s", strerror(errno));
freeClient(c);
return;
}
break;
}
totwritten += nwritten;
offset = c->sentlen;
/* remove written robjs from c->reply */
while (nwritten && listLength(c->reply)) {
o = listNodeValue(listFirst(c->reply));
objlen = sdslen(o->ptr);
if(nwritten >= objlen - offset) {
listDelNode(c->reply, listFirst(c->reply));
nwritten -= objlen - offset;
c->sentlen = 0;
} else {
/* partial write */
c->sentlen += nwritten;
break;
}
offset = 0;
}
}
if (totwritten > 0)
c->lastinteraction = time(NULL);
if (listLength(c->reply) == 0) {
c->sentlen = 0;
aeDeleteFileEvent(server.el,c->fd,AE_WRITABLE);
}
}
void sendReplyToClient(aeEventLoop *el, int fd, void *privdata, int mask) {
redisClient *c = privdata;
int nwritten = 0, totwritten = 0, objlen;
robj *o;
REDIS_NOTUSED(el);
REDIS_NOTUSED(mask);
/* Use writev() if we have enough buffers to send */
if (!server.glueoutputbuf &&
listLength(c->reply) > REDIS_WRITEV_THRESHOLD &&
!(c->flags & REDIS_MASTER))
{
sendReplyToClientWritev(el, fd, privdata, mask);
return;
}
while(listLength(c->reply)) {
if (server.glueoutputbuf && listLength(c->reply) > 1)
glueReplyBuffersIfNeeded(c);
o = listNodeValue(listFirst(c->reply));
objlen = sdslen(o->ptr);
if (objlen == 0) {
listDelNode(c->reply,listFirst(c->reply));
continue;
}
if (c->flags & REDIS_MASTER) {
/* Don't reply to a master */
nwritten = objlen - c->sentlen;
} else {
nwritten = write(fd, ((char*)o->ptr)+c->sentlen, objlen - c->sentlen);
if (nwritten <= 0) break;
}
c->sentlen += nwritten;
totwritten += nwritten;
/* If we fully sent the object on head go to the next one */
if (c->sentlen == objlen) {
listDelNode(c->reply,listFirst(c->reply));
c->sentlen = 0;
}
/* Note that we avoid to send more thank REDIS_MAX_WRITE_PER_EVENT
* bytes, in a single threaded server it's a good idea to serve
* other clients as well, even if a very large request comes from
* super fast link that is always able to accept data (in real world
* scenario think about 'KEYS *' against the loopback interfae) */
if (totwritten > REDIS_MAX_WRITE_PER_EVENT) break;
}
if (nwritten == -1) {
if (errno == EAGAIN) {
nwritten = 0;
} else {
redisLog(REDIS_VERBOSE,
"Error writing to client: %s", strerror(errno));
freeClient(c);
return;
}
}
if (totwritten > 0) c->lastinteraction = time(NULL);
if (listLength(c->reply) == 0) {
c->sentlen = 0;
aeDeleteFileEvent(server.el,c->fd,AE_WRITABLE);
}
}
if (c->bufpos == 0 && listLength(c->reply) == 0 &&
(c->replstate == REDIS_REPL_NONE ||
c->replstate == REDIS_REPL_ONLINE) &&
aeCreateFileEvent(server.el, c->fd, AE_WRITABLE,
sendReplyToClient, c) == AE_ERR) return REDIS_ERR;
return REDIS_OK;
}
/* Create a duplicate of the last object in the reply list when
* it is not exclusively owned by the reply list. */
robj *dupLastObjectIfNeeded(list *reply) {
robj *new, *cur;
listNode *ln;
redisAssert(listLength(reply) > 0);
ln = listLast(reply);
cur = listNodeValue(ln);
if (cur->refcount > 1) {
new = dupStringObject(cur);
decrRefCount(cur);
listNodeValue(ln) = new;
}
return listNodeValue(ln);
}
/* -----------------------------------------------------------------------------
* Low level functions to add more data to output buffers.
* -------------------------------------------------------------------------- */
int _addReplyToBuffer(redisClient *c, char *s, size_t len) {
size_t available = sizeof(c->buf)-c->bufpos;
/* This command implements SCAN, HSCAN and SSCAN commands.
* If object 'o' is passed, then it must be a Hash or Set object, otherwise
* if 'o' is NULL the command will operate on the dictionary associated with
* the current database.
*
* When 'o' is not NULL the function assumes that the first argument in
* the client arguments vector is a key so it skips it before iterating
* in order to parse options.
*
* In the case of a Hash object the function returns both the field and value
* of every element on the Hash. */
void scanGenericCommand(client *c, robj *o, unsigned long cursor) {
int i, j;
list *keys = listCreate();
listNode *node, *nextnode;
long count = 10;
sds pat = NULL;
int patlen = 0, use_pattern = 0;
dict *ht;
/* Object must be NULL (to iterate keys names), or the type of the object
* must be Set, Sorted Set, or Hash. */
serverAssert(o == NULL || o->type == OBJ_SET || o->type == OBJ_HASH ||
o->type == OBJ_ZSET);
/* Set i to the first option argument. The previous one is the cursor. */
i = (o == NULL) ? 2 : 3; /* Skip the key argument if needed. */
/* Step 1: Parse options. */
while (i < c->argc) {
j = c->argc - i;
if (!strcasecmp(c->argv[i]->ptr, "count") && j >= 2) {
if (getLongFromObjectOrReply(c, c->argv[i+1], &count, NULL)
!= C_OK)
{
goto cleanup;
}
if (count < 1) {
addReply(c,shared.syntaxerr);
goto cleanup;
}
i += 2;
} else if (!strcasecmp(c->argv[i]->ptr, "match") && j >= 2) {
pat = c->argv[i+1]->ptr;
patlen = sdslen(pat);
/* The pattern always matches if it is exactly "*", so it is
* equivalent to disabling it. */
use_pattern = !(pat[0] == '*' && patlen == 1);
i += 2;
} else {
addReply(c,shared.syntaxerr);
goto cleanup;
}
}
/* Step 2: Iterate the collection.
*
* Note that if the object is encoded with a ziplist, intset, or any other
* representation that is not a hash table, we are sure that it is also
* composed of a small number of elements. So to avoid taking state we
* just return everything inside the object in a single call, setting the
* cursor to zero to signal the end of the iteration. */
/* Handle the case of a hash table. */
ht = NULL;
if (o == NULL) {
ht = c->db->dict;
} else if (o->type == OBJ_SET && o->encoding == OBJ_ENCODING_HT) {
ht = o->ptr;
} else if (o->type == OBJ_HASH && o->encoding == OBJ_ENCODING_HT) {
ht = o->ptr;
count *= 2; /* We return key / value for this type. */
} else if (o->type == OBJ_ZSET && o->encoding == OBJ_ENCODING_SKIPLIST) {
zset *zs = o->ptr;
ht = zs->dict;
count *= 2; /* We return key / value for this type. */
}
if (ht) {
void *privdata[2];
/* We set the max number of iterations to ten times the specified
* COUNT, so if the hash table is in a pathological state (very
* sparsely populated) we avoid to block too much time at the cost
* of returning no or very few elements. */
long maxiterations = count*10;
/* We pass two pointers to the callback: the list to which it will
* add new elements, and the object containing the dictionary so that
* it is possible to fetch more data in a type-dependent way. */
privdata[0] = keys;
privdata[1] = o;
do {
cursor = dictScan(ht, cursor, scanCallback, privdata);
} while (cursor &&
maxiterations-- &&
listLength(keys) < (unsigned long)count);
} else if (o->type == OBJ_SET) {
int pos = 0;
int64_t ll;
while(intsetGet(o->ptr,pos++,&ll))
listAddNodeTail(keys,createStringObjectFromLongLong(ll));
cursor = 0;
} else if (o->type == OBJ_HASH || o->type == OBJ_ZSET) {
unsigned char *p = ziplistIndex(o->ptr,0);
unsigned char *vstr;
unsigned int vlen;
long long vll;
while(p) {
ziplistGet(p,&vstr,&vlen,&vll);
listAddNodeTail(keys,
(vstr != NULL) ? createStringObject((char*)vstr,vlen) :
createStringObjectFromLongLong(vll));
p = ziplistNext(o->ptr,p);
}
cursor = 0;
} else {
serverPanic("Not handled encoding in SCAN.");
}
/* Step 3: Filter elements. */
node = listFirst(keys);
while (node) {
robj *kobj = listNodeValue(node);
nextnode = listNextNode(node);
int filter = 0;
/* Filter element if it does not match the pattern. */
if (!filter && use_pattern) {
if (sdsEncodedObject(kobj)) {
if (!stringmatchlen(pat, patlen, kobj->ptr, sdslen(kobj->ptr), 0))
filter = 1;
} else {
char buf[LONG_STR_SIZE];
int len;
serverAssert(kobj->encoding == OBJ_ENCODING_INT);
len = ll2string(buf,sizeof(buf),(long)kobj->ptr);
if (!stringmatchlen(pat, patlen, buf, len, 0)) filter = 1;
}
}
/* Filter element if it is an expired key. */
if (!filter && o == NULL && expireIfNeeded(c->db, kobj)) filter = 1;
/* Remove the element and its associted value if needed. */
if (filter) {
decrRefCount(kobj);
listDelNode(keys, node);
}
/* If this is a hash or a sorted set, we have a flat list of
* key-value elements, so if this element was filtered, remove the
* value, or skip it if it was not filtered: we only match keys. */
if (o && (o->type == OBJ_ZSET || o->type == OBJ_HASH)) {
node = nextnode;
nextnode = listNextNode(node);
if (filter) {
kobj = listNodeValue(node);
decrRefCount(kobj);
listDelNode(keys, node);
}
}
node = nextnode;
}
/* Step 4: Reply to the client. */
addReplyMultiBulkLen(c, 2);
addReplyBulkLongLong(c,cursor);
addReplyMultiBulkLen(c, listLength(keys));
while ((node = listFirst(keys)) != NULL) {
robj *kobj = listNodeValue(node);
addReplyBulk(c, kobj);
decrRefCount(kobj);
listDelNode(keys, node);
}
cleanup:
listSetFreeMethod(keys,decrRefCountVoid);
listRelease(keys);
}
/* Write a sequence of commands able to fully rebuild the dataset into
* "filename". Used both by REWRITEAOF and BGREWRITEAOF. */
int rewriteAppendOnlyFile(char *filename) {
dictIterator *di = NULL;
dictEntry *de;
FILE *fp;
char tmpfile[256];
int j;
time_t now = time(NULL);
/* Note that we have to use a different temp name here compared to the
* one used by rewriteAppendOnlyFileBackground() function. */
snprintf(tmpfile,256,"temp-rewriteaof-%d.aof", (int) getpid());
fp = fopen(tmpfile,"w");
if (!fp) {
redisLog(REDIS_WARNING, "Failed rewriting the append only file: %s", strerror(errno));
return REDIS_ERR;
}
for (j = 0; j < server.dbnum; j++) {
char selectcmd[] = "*2\r\n$6\r\nSELECT\r\n";
redisDb *db = server.db+j;
dict *d = db->dict;
if (dictSize(d) == 0) continue;
di = dictGetSafeIterator(d);
if (!di) {
fclose(fp);
return REDIS_ERR;
}
/* SELECT the new DB */
if (fwrite(selectcmd,sizeof(selectcmd)-1,1,fp) == 0) goto werr;
if (fwriteBulkLongLong(fp,j) == 0) goto werr;
/* Iterate this DB writing every entry */
while((de = dictNext(di)) != NULL) {
sds keystr = dictGetEntryKey(de);
robj key, *o;
time_t expiretime;
int swapped;
keystr = dictGetEntryKey(de);
o = dictGetEntryVal(de);
initStaticStringObject(key,keystr);
/* If the value for this key is swapped, load a preview in memory.
* We use a "swapped" flag to remember if we need to free the
* value object instead to just increment the ref count anyway
* in order to avoid copy-on-write of pages if we are forked() */
if (!server.vm_enabled || o->storage == REDIS_VM_MEMORY ||
o->storage == REDIS_VM_SWAPPING) {
swapped = 0;
} else {
o = vmPreviewObject(o);
swapped = 1;
}
expiretime = getExpire(db,&key);
/* Save the key and associated value */
if (o->type == REDIS_STRING) {
/* Emit a SET command */
char cmd[]="*3\r\n$3\r\nSET\r\n";
if (fwrite(cmd,sizeof(cmd)-1,1,fp) == 0) goto werr;
/* Key and value */
if (fwriteBulkObject(fp,&key) == 0) goto werr;
if (fwriteBulkObject(fp,o) == 0) goto werr;
} else if (o->type == REDIS_LIST) {
/* Emit the RPUSHes needed to rebuild the list */
char cmd[]="*3\r\n$5\r\nRPUSH\r\n";
if (o->encoding == REDIS_ENCODING_ZIPLIST) {
unsigned char *zl = o->ptr;
unsigned char *p = ziplistIndex(zl,0);
unsigned char *vstr;
unsigned int vlen;
long long vlong;
while(ziplistGet(p,&vstr,&vlen,&vlong)) {
if (fwrite(cmd,sizeof(cmd)-1,1,fp) == 0) goto werr;
if (fwriteBulkObject(fp,&key) == 0) goto werr;
if (vstr) {
if (fwriteBulkString(fp,(char*)vstr,vlen) == 0)
goto werr;
} else {
if (fwriteBulkLongLong(fp,vlong) == 0)
goto werr;
}
p = ziplistNext(zl,p);
}
} else if (o->encoding == REDIS_ENCODING_LINKEDLIST) {
list *list = o->ptr;
listNode *ln;
listIter li;
listRewind(list,&li);
while((ln = listNext(&li))) {
robj *eleobj = listNodeValue(ln);
if (fwrite(cmd,sizeof(cmd)-1,1,fp) == 0) goto werr;
if (fwriteBulkObject(fp,&key) == 0) goto werr;
if (fwriteBulkObject(fp,eleobj) == 0) goto werr;
}
} else {
redisPanic("Unknown list encoding");
}
} else if (o->type == REDIS_SET) {
char cmd[]="*3\r\n$4\r\nSADD\r\n";
/* Emit the SADDs needed to rebuild the set */
if (o->encoding == REDIS_ENCODING_INTSET) {
int ii = 0;
int64_t llval;
while(intsetGet(o->ptr,ii++,&llval)) {
if (fwrite(cmd,sizeof(cmd)-1,1,fp) == 0) goto werr;
if (fwriteBulkObject(fp,&key) == 0) goto werr;
if (fwriteBulkLongLong(fp,llval) == 0) goto werr;
}
} else if (o->encoding == REDIS_ENCODING_HT) {
dictIterator *di = dictGetIterator(o->ptr);
dictEntry *de;
while((de = dictNext(di)) != NULL) {
robj *eleobj = dictGetEntryKey(de);
if (fwrite(cmd,sizeof(cmd)-1,1,fp) == 0) goto werr;
if (fwriteBulkObject(fp,&key) == 0) goto werr;
if (fwriteBulkObject(fp,eleobj) == 0) goto werr;
}
dictReleaseIterator(di);
} else {
redisPanic("Unknown set encoding");
}
} else if (o->type == REDIS_ZSET) {
/* Emit the ZADDs needed to rebuild the sorted set */
char cmd[]="*4\r\n$4\r\nZADD\r\n";
if (o->encoding == REDIS_ENCODING_ZIPLIST) {
unsigned char *zl = o->ptr;
unsigned char *eptr, *sptr;
unsigned char *vstr;
unsigned int vlen;
long long vll;
double score;
eptr = ziplistIndex(zl,0);
redisAssert(eptr != NULL);
sptr = ziplistNext(zl,eptr);
redisAssert(sptr != NULL);
while (eptr != NULL) {
redisAssert(ziplistGet(eptr,&vstr,&vlen,&vll));
score = zzlGetScore(sptr);
if (fwrite(cmd,sizeof(cmd)-1,1,fp) == 0) goto werr;
if (fwriteBulkObject(fp,&key) == 0) goto werr;
if (fwriteBulkDouble(fp,score) == 0) goto werr;
if (vstr != NULL) {
if (fwriteBulkString(fp,(char*)vstr,vlen) == 0)
goto werr;
} else {
if (fwriteBulkLongLong(fp,vll) == 0)
goto werr;
}
zzlNext(zl,&eptr,&sptr);
}
} else if (o->encoding == REDIS_ENCODING_SKIPLIST) {
zset *zs = o->ptr;
dictIterator *di = dictGetIterator(zs->dict);
dictEntry *de;
while((de = dictNext(di)) != NULL) {
robj *eleobj = dictGetEntryKey(de);
double *score = dictGetEntryVal(de);
if (fwrite(cmd,sizeof(cmd)-1,1,fp) == 0) goto werr;
if (fwriteBulkObject(fp,&key) == 0) goto werr;
if (fwriteBulkDouble(fp,*score) == 0) goto werr;
if (fwriteBulkObject(fp,eleobj) == 0) goto werr;
}
dictReleaseIterator(di);
} else {
redisPanic("Unknown sorted set encoding");
}
} else if (o->type == REDIS_HASH) {
char cmd[]="*4\r\n$4\r\nHSET\r\n";
/* Emit the HSETs needed to rebuild the hash */
if (o->encoding == REDIS_ENCODING_ZIPMAP) {
unsigned char *p = zipmapRewind(o->ptr);
unsigned char *field, *val;
unsigned int flen, vlen;
while((p = zipmapNext(p,&field,&flen,&val,&vlen)) != NULL) {
if (fwrite(cmd,sizeof(cmd)-1,1,fp) == 0) goto werr;
if (fwriteBulkObject(fp,&key) == 0) goto werr;
if (fwriteBulkString(fp,(char*)field,flen) == 0)
goto werr;
if (fwriteBulkString(fp,(char*)val,vlen) == 0)
goto werr;
}
} else {
dictIterator *di = dictGetIterator(o->ptr);
dictEntry *de;
while((de = dictNext(di)) != NULL) {
robj *field = dictGetEntryKey(de);
robj *val = dictGetEntryVal(de);
if (fwrite(cmd,sizeof(cmd)-1,1,fp) == 0) goto werr;
if (fwriteBulkObject(fp,&key) == 0) goto werr;
if (fwriteBulkObject(fp,field) == 0) goto werr;
if (fwriteBulkObject(fp,val) == 0) goto werr;
}
dictReleaseIterator(di);
}
} else {
redisPanic("Unknown object type");
}
/* Save the expire time */
if (expiretime != -1) {
char cmd[]="*3\r\n$8\r\nEXPIREAT\r\n";
/* If this key is already expired skip it */
if (expiretime < now) continue;
if (fwrite(cmd,sizeof(cmd)-1,1,fp) == 0) goto werr;
if (fwriteBulkObject(fp,&key) == 0) goto werr;
if (fwriteBulkLongLong(fp,expiretime) == 0) goto werr;
}
if (swapped) decrRefCount(o);
}
dictReleaseIterator(di);
}
/* Make sure data will not remain on the OS's output buffers */
fflush(fp);
aof_fsync(fileno(fp));
fclose(fp);
/* Use RENAME to make sure the DB file is changed atomically only
* if the generate DB file is ok. */
if (rename(tmpfile,filename) == -1) {
redisLog(REDIS_WARNING,"Error moving temp append only file on the final destination: %s", strerror(errno));
unlink(tmpfile);
return REDIS_ERR;
}
redisLog(REDIS_NOTICE,"SYNC append only file rewrite performed");
return REDIS_OK;
werr:
fclose(fp);
unlink(tmpfile);
redisLog(REDIS_WARNING,"Write error writing append only file on disk: %s", strerror(errno));
if (di) dictReleaseIterator(di);
return REDIS_ERR;
}
int luaRedisGenericCommand(lua_State *lua, int raise_error) {
int j, argc = lua_gettop(lua);
struct redisCommand *cmd;
redisClient *c = server.lua_client;
sds reply;
/* Cached across calls. */
static robj **argv = NULL;
static int argv_size = 0;
static robj *cached_objects[LUA_CMD_OBJCACHE_SIZE];
static size_t cached_objects_len[LUA_CMD_OBJCACHE_SIZE];
static int inuse = 0; /* Recursive calls detection. */
/* By using Lua debug hooks it is possible to trigger a recursive call
* to luaRedisGenericCommand(), which normally should never happen.
* To make this function reentrant is futile and makes it slower, but
* we should at least detect such a misuse, and abort. */
if (inuse) {
char *recursion_warning =
"luaRedisGenericCommand() recursive call detected. "
"Are you doing funny stuff with Lua debug hooks?";
redisLog(REDIS_WARNING,"%s",recursion_warning);
luaPushError(lua,recursion_warning);
return 1;
}
inuse++;
/* Require at least one argument */
if (argc == 0) {
luaPushError(lua,
"Please specify at least one argument for redis.call()");
inuse--;
return 1;
}
/* Build the arguments vector */
if (argv_size < argc) {
argv = zrealloc(argv,sizeof(robj*)*argc);
argv_size = argc;
}
for (j = 0; j < argc; j++) {
char *obj_s;
size_t obj_len;
char dbuf[64];
if (lua_type(lua,j+1) == LUA_TNUMBER) {
/* We can't use lua_tolstring() for number -> string conversion
* since Lua uses a format specifier that loses precision. */
lua_Number num = lua_tonumber(lua,j+1);
obj_len = snprintf(dbuf,sizeof(dbuf),"%.17g",(double)num);
obj_s = dbuf;
} else {
obj_s = (char*)lua_tolstring(lua,j+1,&obj_len);
if (obj_s == NULL) break; /* Not a string. */
}
/* Try to use a cached object. */
if (j < LUA_CMD_OBJCACHE_SIZE && cached_objects[j] &&
cached_objects_len[j] >= obj_len)
{
char *s = cached_objects[j]->ptr;
struct sdshdr *sh = (void*)(s-(sizeof(struct sdshdr)));
argv[j] = cached_objects[j];
cached_objects[j] = NULL;
memcpy(s,obj_s,obj_len+1);
sh->free += sh->len - obj_len;
sh->len = obj_len;
} else {
argv[j] = createStringObject(obj_s, obj_len);
}
}
/* Check if one of the arguments passed by the Lua script
* is not a string or an integer (lua_isstring() return true for
* integers as well). */
if (j != argc) {
j--;
while (j >= 0) {
decrRefCount(argv[j]);
j--;
}
luaPushError(lua,
"Lua redis() command arguments must be strings or integers");
inuse--;
return 1;
}
/* Setup our fake client for command execution */
c->argv = argv;
c->argc = argc;
/* Command lookup */
cmd = lookupCommand(argv[0]->ptr);
if (!cmd || ((cmd->arity > 0 && cmd->arity != argc) ||
(argc < -cmd->arity)))
{
if (cmd)
luaPushError(lua,
"Wrong number of args calling Redis command From Lua script");
else
luaPushError(lua,"Unknown Redis command called from Lua script");
goto cleanup;
}
c->cmd = cmd;
/* There are commands that are not allowed inside scripts. */
if (cmd->flags & REDIS_CMD_NOSCRIPT) {
luaPushError(lua, "This Redis command is not allowed from scripts");
goto cleanup;
}
/* Write commands are forbidden against read-only slaves, or if a
* command marked as non-deterministic was already called in the context
* of this script. */
if (cmd->flags & REDIS_CMD_WRITE) {
if (server.lua_random_dirty) {
luaPushError(lua,
"Write commands not allowed after non deterministic commands");
goto cleanup;
} else if (server.masterhost && server.repl_slave_ro &&
!server.loading &&
!(server.lua_caller->flags & REDIS_MASTER))
{
luaPushError(lua, shared.roslaveerr->ptr);
goto cleanup;
} else if (server.stop_writes_on_bgsave_err &&
server.saveparamslen > 0 &&
server.lastbgsave_status == REDIS_ERR)
{
luaPushError(lua, shared.bgsaveerr->ptr);
goto cleanup;
}
}
/* If we reached the memory limit configured via maxmemory, commands that
* could enlarge the memory usage are not allowed, but only if this is the
* first write in the context of this script, otherwise we can't stop
* in the middle. */
if (server.maxmemory && server.lua_write_dirty == 0 &&
(cmd->flags & REDIS_CMD_DENYOOM))
{
if (freeMemoryIfNeeded() == REDIS_ERR) {
luaPushError(lua, shared.oomerr->ptr);
goto cleanup;
}
}
if (cmd->flags & REDIS_CMD_RANDOM) server.lua_random_dirty = 1;
if (cmd->flags & REDIS_CMD_WRITE) server.lua_write_dirty = 1;
/* If this is a Redis Cluster node, we need to make sure Lua is not
* trying to access non-local keys, with the exception of commands
* received from our master. */
if (server.cluster_enabled && !(server.lua_caller->flags & REDIS_MASTER)) {
/* Duplicate relevant flags in the lua client. */
c->flags &= ~(REDIS_READONLY|REDIS_ASKING);
c->flags |= server.lua_caller->flags & (REDIS_READONLY|REDIS_ASKING);
if (getNodeByQuery(c,c->cmd,c->argv,c->argc,NULL,NULL) !=
server.cluster->myself)
{
luaPushError(lua,
"Lua script attempted to access a non local key in a "
"cluster node");
goto cleanup;
}
}
/* Run the command */
call(c,REDIS_CALL_SLOWLOG | REDIS_CALL_STATS);
/* Convert the result of the Redis command into a suitable Lua type.
* The first thing we need is to create a single string from the client
* output buffers. */
if (listLength(c->reply) == 0 && c->bufpos < REDIS_REPLY_CHUNK_BYTES) {
/* This is a fast path for the common case of a reply inside the
* client static buffer. Don't create an SDS string but just use
* the client buffer directly. */
c->buf[c->bufpos] = '\0';
reply = c->buf;
c->bufpos = 0;
} else {
reply = sdsnewlen(c->buf,c->bufpos);
c->bufpos = 0;
while(listLength(c->reply)) {
robj *o = listNodeValue(listFirst(c->reply));
reply = sdscatlen(reply,o->ptr,sdslen(o->ptr));
listDelNode(c->reply,listFirst(c->reply));
}
}
if (raise_error && reply[0] != '-') raise_error = 0;
redisProtocolToLuaType(lua,reply);
/* Sort the output array if needed, assuming it is a non-null multi bulk
* reply as expected. */
if ((cmd->flags & REDIS_CMD_SORT_FOR_SCRIPT) &&
(reply[0] == '*' && reply[1] != '-')) {
luaSortArray(lua);
}
if (reply != c->buf) sdsfree(reply);
c->reply_bytes = 0;
cleanup:
/* Clean up. Command code may have changed argv/argc so we use the
* argv/argc of the client instead of the local variables. */
for (j = 0; j < c->argc; j++) {
robj *o = c->argv[j];
/* Try to cache the object in the cached_objects array.
* The object must be small, SDS-encoded, and with refcount = 1
* (we must be the only owner) for us to cache it. */
if (j < LUA_CMD_OBJCACHE_SIZE &&
o->refcount == 1 &&
(o->encoding == REDIS_ENCODING_RAW ||
o->encoding == REDIS_ENCODING_EMBSTR) &&
sdslen(o->ptr) <= LUA_CMD_OBJCACHE_MAX_LEN)
{
struct sdshdr *sh = (void*)(((char*)(o->ptr))-(sizeof(struct sdshdr)));
if (cached_objects[j]) decrRefCount(cached_objects[j]);
cached_objects[j] = o;
cached_objects_len[j] = sh->free + sh->len;
} else {
decrRefCount(o);
}
}
if (c->argv != argv) {
zfree(c->argv);
argv = NULL;
argv_size = 0;
}
if (raise_error) {
/* If we are here we should have an error in the stack, in the
* form of a table with an "err" field. Extract the string to
* return the plain error. */
lua_pushstring(lua,"err");
lua_gettable(lua,-2);
inuse--;
return lua_error(lua);
}
inuse--;
return 1;
}
/* *
* slotscheck
* */
void
slotscheckCommand(redisClient *c) {
sds bug = NULL;
int i;
for (i = 0; i < HASH_SLOTS_SIZE && bug == NULL; i ++) {
dict *d = c->db->hash_slots[i];
if (dictSize(d) == 0) {
continue;
}
list *l = listCreate();
listSetFreeMethod(l, decrRefCountVoid);
unsigned long cursor = 0;
do {
cursor = dictScan(d, cursor, slotsScanSdsKeyCallback, l);
while (1) {
listNode *head = listFirst(l);
if (head == NULL) {
break;
}
robj *key = listNodeValue(head);
if (lookupKey(c->db, key) == NULL) {
if (bug == NULL) {
bug = sdsdup(key->ptr);
}
}
listDelNode(l, head);
}
} while (cursor != 0 && bug == NULL);
listRelease(l);
}
if (bug != NULL) {
addReplyErrorFormat(c, "step 1, miss = '%s'", bug);
sdsfree(bug);
return;
}
do {
dict *d = c->db->dict;
if (dictSize(d) == 0) {
break;
}
list *l = listCreate();
listSetFreeMethod(l, decrRefCountVoid);
unsigned long cursor = 0;
do {
cursor = dictScan(d, cursor, slotsScanSdsKeyCallback, l);
while (1) {
listNode *head = listFirst(l);
if (head == NULL) {
break;
}
robj *key = listNodeValue(head);
int slot = slots_num(key->ptr, NULL);
if (dictFind(c->db->hash_slots[slot], key->ptr) == NULL) {
if (bug == NULL) {
bug = sdsdup(key->ptr);
}
}
listDelNode(l, head);
}
} while (cursor != 0 && bug == NULL);
listRelease(l);
} while (0);
if (bug != NULL) {
addReplyErrorFormat(c, "step 2, miss = '%s'", bug);
sdsfree(bug);
return;
}
addReply(c, shared.ok);
}
int luaRedisGenericCommand(lua_State *lua, int raise_error) {
int j, argc = lua_gettop(lua);
struct redisCommand *cmd;
robj **argv;
redisClient *c = server.lua_client;
sds reply;
/* Require at least one argument */
if (argc == 0) {
luaPushError(lua,
"Please specify at least one argument for redis.call()");
return 1;
}
/* Build the arguments vector */
argv = zmalloc(sizeof(robj*)*argc);
for (j = 0; j < argc; j++) {
if (!lua_isstring(lua,j+1)) break;
argv[j] = createStringObject((char*)lua_tostring(lua,j+1),
lua_strlen(lua,j+1));
}
/* Check if one of the arguments passed by the Lua script
* is not a string or an integer (lua_isstring() return true for
* integers as well). */
if (j != argc) {
j--;
while (j >= 0) {
decrRefCount(argv[j]);
j--;
}
zfree(argv);
luaPushError(lua,
"Lua redis() command arguments must be strings or integers");
return 1;
}
/* Setup our fake client for command execution */
c->argv = argv;
c->argc = argc;
/* Command lookup */
cmd = lookupCommand(argv[0]->ptr);
if (!cmd || ((cmd->arity > 0 && cmd->arity != argc) ||
(argc < -cmd->arity)))
{
if (cmd)
luaPushError(lua,
"Wrong number of args calling Redis command From Lua script");
else
luaPushError(lua,"Unknown Redis command called from Lua script");
goto cleanup;
}
/* There are commands that are not allowed inside scripts. */
if (cmd->flags & REDIS_CMD_NOSCRIPT) {
luaPushError(lua, "This Redis command is not allowed from scripts");
goto cleanup;
}
/* Write commands are forbidden against read-only slaves, or if a
* command marked as non-deterministic was already called in the context
* of this script. */
if (cmd->flags & REDIS_CMD_WRITE) {
if (server.lua_random_dirty) {
luaPushError(lua,
"Write commands not allowed after non deterministic commands");
goto cleanup;
} else if (server.masterhost && server.repl_slave_ro &&
!server.loading &&
!(server.lua_caller->flags & REDIS_MASTER))
{
luaPushError(lua, shared.roslaveerr->ptr);
goto cleanup;
} else if (server.stop_writes_on_bgsave_err &&
server.saveparamslen > 0 &&
server.lastbgsave_status == REDIS_ERR)
{
luaPushError(lua, shared.bgsaveerr->ptr);
goto cleanup;
}
}
/* If we reached the memory limit configured via maxmemory, commands that
* could enlarge the memory usage are not allowed, but only if this is the
* first write in the context of this script, otherwise we can't stop
* in the middle. */
if (server.maxmemory && server.lua_write_dirty == 0 &&
(cmd->flags & REDIS_CMD_DENYOOM))
{
if (freeMemoryIfNeeded() == REDIS_ERR) {
luaPushError(lua, shared.oomerr->ptr);
goto cleanup;
}
}
if (cmd->flags & REDIS_CMD_RANDOM) server.lua_random_dirty = 1;
if (cmd->flags & REDIS_CMD_WRITE) server.lua_write_dirty = 1;
/* Run the command */
c->cmd = cmd;
call(c,REDIS_CALL_SLOWLOG | REDIS_CALL_STATS);
/* Convert the result of the Redis command into a suitable Lua type.
* The first thing we need is to create a single string from the client
* output buffers. */
reply = sdsempty();
if (c->bufpos) {
reply = sdscatlen(reply,c->buf,c->bufpos);
c->bufpos = 0;
}
while(listLength(c->reply)) {
robj *o = listNodeValue(listFirst(c->reply));
reply = sdscatlen(reply,o->ptr,sdslen(o->ptr));
listDelNode(c->reply,listFirst(c->reply));
}
if (raise_error && reply[0] != '-') raise_error = 0;
redisProtocolToLuaType(lua,reply);
/* Sort the output array if needed, assuming it is a non-null multi bulk
* reply as expected. */
if ((cmd->flags & REDIS_CMD_SORT_FOR_SCRIPT) &&
(reply[0] == '*' && reply[1] != '-')) {
luaSortArray(lua);
}
sdsfree(reply);
c->reply_bytes = 0;
cleanup:
/* Clean up. Command code may have changed argv/argc so we use the
* argv/argc of the client instead of the local variables. */
for (j = 0; j < c->argc; j++)
decrRefCount(c->argv[j]);
zfree(c->argv);
if (raise_error) {
/* If we are here we should have an error in the stack, in the
* form of a table with an "err" field. Extract the string to
* return the plain error. */
lua_pushstring(lua,"err");
lua_gettable(lua,-2);
return lua_error(lua);
}
return 1;
}
int freeMemoryIfNeeded(void) {
size_t mem_reported, mem_used, mem_tofree, mem_freed;
int slaves = listLength(server.slaves);
mstime_t latency, eviction_latency;
long long delta;
/* Check if we are over the memory usage limit. If we are not, no need
* to subtract the slaves output buffers. We can just return ASAP. */
mem_reported = zmalloc_used_memory();
if (mem_reported <= server.maxmemory) return C_OK;
/* Remove the size of slaves output buffers and AOF buffer from the
* count of used memory. */
mem_used = mem_reported;
if (slaves) {
listIter li;
listNode *ln;
listRewind(server.slaves,&li);
while((ln = listNext(&li))) {
client *slave = listNodeValue(ln);
unsigned long obuf_bytes = getClientOutputBufferMemoryUsage(slave);
if (obuf_bytes > mem_used)
mem_used = 0;
else
mem_used -= obuf_bytes;
}
}
if (server.aof_state != AOF_OFF) {
mem_used -= sdslen(server.aof_buf);
mem_used -= aofRewriteBufferSize();
}
/* Check if we are still over the memory limit. */
if (mem_used <= server.maxmemory) return C_OK;
/* Compute how much memory we need to free. */
mem_tofree = mem_used - server.maxmemory;
mem_freed = 0;
if (server.maxmemory_policy == MAXMEMORY_NO_EVICTION)
goto cant_free; /* We need to free memory, but policy forbids. */
latencyStartMonitor(latency);
while (mem_freed < mem_tofree) {
int j, k, i, keys_freed = 0;
static int next_db = 0;
sds bestkey = NULL;
int bestdbid;
redisDb *db;
dict *dict;
dictEntry *de;
if (server.maxmemory_policy == MAXMEMORY_ALLKEYS_LRU ||
server.maxmemory_policy == MAXMEMORY_VOLATILE_LRU)
{
struct evictionPoolEntry *pool = EvictionPoolLRU;
while(bestkey == NULL) {
unsigned long total_keys = 0, keys;
/* We don't want to make local-db choices when expiring keys,
* so to start populate the eviction pool sampling keys from
* every DB. */
for (i = 0; i < server.dbnum; i++) {
db = server.db+i;
dict = (server.maxmemory_policy == MAXMEMORY_ALLKEYS_LRU) ?
db->dict : db->expires;
if ((keys = dictSize(dict)) != 0) {
evictionPoolPopulate(i, dict, db->dict, pool);
total_keys += keys;
}
}
if (!total_keys) break; /* No keys to evict. */
/* Go backward from best to worst element to evict. */
for (k = EVPOOL_SIZE-1; k >= 0; k--) {
if (pool[k].key == NULL) continue;
bestdbid = pool[k].dbid;
if (server.maxmemory_policy == MAXMEMORY_ALLKEYS_LRU) {
de = dictFind(server.db[pool[k].dbid].dict,
pool[k].key);
} else {
de = dictFind(server.db[pool[k].dbid].expires,
pool[k].key);
}
/* Remove the entry from the pool. */
if (pool[k].key != pool[k].cached)
sdsfree(pool[k].key);
pool[k].key = NULL;
pool[k].idle = 0;
/* If the key exists, is our pick. Otherwise it is
* a ghost and we need to try the next element. */
if (de) {
bestkey = dictGetKey(de);
break;
} else {
/* Ghost... Iterate again. */
}
}
}
}
/* volatile-random and allkeys-random policy */
else if (server.maxmemory_policy == MAXMEMORY_ALLKEYS_RANDOM ||
server.maxmemory_policy == MAXMEMORY_VOLATILE_RANDOM)
{
/* When evicting a random key, we try to evict a key for
* each DB, so we use the static 'next_db' variable to
* incrementally visit all DBs. */
for (i = 0; i < server.dbnum; i++) {
j = (++next_db) % server.dbnum;
db = server.db+j;
dict = (server.maxmemory_policy == MAXMEMORY_ALLKEYS_RANDOM) ?
db->dict : db->expires;
if (dictSize(dict) != 0) {
de = dictGetRandomKey(dict);
bestkey = dictGetKey(de);
bestdbid = j;
break;
}
}
}
/* volatile-ttl */
else if (server.maxmemory_policy == MAXMEMORY_VOLATILE_TTL) {
long bestttl = 0; /* Initialized to avoid warning. */
/* In this policy we scan a single DB per iteration (visiting
* a different DB per call), expiring the key with the smallest
* TTL among the few sampled.
*
* Note that this algorithm makes local-DB choices, and should
* use a pool and code more similr to the one used in the
* LRU eviction policies in the future. */
for (i = 0; i < server.dbnum; i++) {
j = (++next_db) % server.dbnum;
db = server.db+j;
dict = db->expires;
if (dictSize(dict) != 0) {
for (k = 0; k < server.maxmemory_samples; k++) {
sds thiskey;
long thisttl;
de = dictGetRandomKey(dict);
thiskey = dictGetKey(de);
thisttl = (long) dictGetVal(de);
/* Keys expiring sooner (smaller unix timestamp) are
* better candidates for deletion */
if (bestkey == NULL || thisttl < bestttl) {
bestkey = thiskey;
bestttl = thisttl;
bestdbid = j;
}
}
}
}
}
/* Finally remove the selected key. */
if (bestkey) {
db = server.db+bestdbid;
robj *keyobj = createStringObject(bestkey,sdslen(bestkey));
propagateExpire(db,keyobj,server.lazyfree_lazy_eviction);
/* We compute the amount of memory freed by db*Delete() alone.
* It is possible that actually the memory needed to propagate
* the DEL in AOF and replication link is greater than the one
* we are freeing removing the key, but we can't account for
* that otherwise we would never exit the loop.
*
* AOF and Output buffer memory will be freed eventually so
* we only care about memory used by the key space. */
delta = (long long) zmalloc_used_memory();
latencyStartMonitor(eviction_latency);
if (server.lazyfree_lazy_eviction)
dbAsyncDelete(db,keyobj);
else
dbSyncDelete(db,keyobj);
latencyEndMonitor(eviction_latency);
latencyAddSampleIfNeeded("eviction-del",eviction_latency);
latencyRemoveNestedEvent(latency,eviction_latency);
delta -= (long long) zmalloc_used_memory();
mem_freed += delta;
server.stat_evictedkeys++;
notifyKeyspaceEvent(NOTIFY_EVICTED, "evicted",
keyobj, db->id);
decrRefCount(keyobj);
keys_freed++;
/* When the memory to free starts to be big enough, we may
* start spending so much time here that is impossible to
* deliver data to the slaves fast enough, so we force the
* transmission here inside the loop. */
if (slaves) flushSlavesOutputBuffers();
}
if (!keys_freed) {
latencyEndMonitor(latency);
latencyAddSampleIfNeeded("eviction-cycle",latency);
goto cant_free; /* nothing to free... */
}
}
latencyEndMonitor(latency);
latencyAddSampleIfNeeded("eviction-cycle",latency);
return C_OK;
cant_free:
/* We are here if we are not able to reclaim memory. There is only one
* last thing we can try: check if the lazyfree thread has jobs in queue
* and wait... */
while(bioPendingJobsOfType(BIO_LAZY_FREE)) {
if (((mem_reported - zmalloc_used_memory()) + mem_freed) >= mem_tofree)
break;
usleep(1000);
}
return C_ERR;
}
/* *
* do migrate mutli key-value(s) for {slotsmgrt/slotsmgrtone}with tag commands
* return value:
* -1 - error happens
* >=0 - # of success migration
* */
static int
slotsmgrttag_command(redisClient *c, sds host, sds port, int timeout, robj *key) {
int taglen;
void *tag = slots_tag(key->ptr, &taglen);
if (tag == NULL) {
return slotsmgrtone_command(c, host, port, timeout, key);
}
int fd = slotsmgrt_get_socket(c, host, port, timeout);
if (fd == -1) {
return -1;
}
list *l = listCreate();
listSetFreeMethod(l, decrRefCountVoid);
do {
uint32_t crc;
int slot = slots_num(key->ptr, &crc);
dict *d = c->db->hash_slots[slot];
long long cursor = 0;
void *args[] = {l, tag, &taglen, (void *)(long)crc};
do {
cursor = dictScan(d, cursor, slotsScanSdsKeyTagCallback, args);
} while (cursor != 0);
} while (0);
int max = listLength(l);
if (max == 0) {
listRelease(l);
return 0;
}
robj **keys = zmalloc(sizeof(robj *) * max);
robj **vals = zmalloc(sizeof(robj *) * max);
int n = 0;
for (int i = 0; i < max; i ++) {
listNode *head = listFirst(l);
robj *key = listNodeValue(head);
robj *val = lookupKeyWrite(c->db, key);
if (val != NULL) {
keys[n] = key;
vals[n] = val;
n ++;
incrRefCount(key);
}
listDelNode(l, head);
}
int ret = 0;
if (n != 0) {
if (slotsmgrt(c, host, port, fd, c->db->id, timeout, keys, vals, n) != 0) {
slotsmgrt_close_socket(host, port);
ret = -1;
} else {
slotsremove(c, keys, n, 1);
ret = n;
}
}
listRelease(l);
for (int i = 0; i < n; i ++) {
decrRefCount(keys[i]);
}
zfree(keys);
zfree(vals);
return ret;
}
/* return array of sockets that are ready for read or write
depending on the mask for each socket */
static int aeApiPoll(aeEventLoop *eventLoop, struct timeval *tvp) {
aeApiState *state = (aeApiState *)eventLoop->apidata;
aeSockState *sockstate;
ULONG j;
int numevents = 0;
ULONG numComplete = 0;
int rc;
int mswait = (tvp == NULL) ? 100 : (tvp->tv_sec * 1000) + (tvp->tv_usec / 1000);
if (pGetQueuedCompletionStatusEx != NULL) {
/* first get an array of completion notifications */
rc = pGetQueuedCompletionStatusEx(state->iocp,
state->entries,
MAX_COMPLETE_PER_POLL,
&numComplete,
mswait,
FALSE);
} else {
/* need to get one at a time. Use first array element */
rc = GetQueuedCompletionStatus(state->iocp,
&state->entries[0].dwNumberOfBytesTransferred,
&state->entries[0].lpCompletionKey,
&state->entries[0].lpOverlapped,
mswait);
if (!rc && state->entries[0].lpOverlapped == NULL) {
// timeout. Return.
return 0;
} else {
// check if more completions are ready
int lrc = 1;
rc = 1;
numComplete = 1;
while (numComplete < MAX_COMPLETE_PER_POLL) {
lrc = GetQueuedCompletionStatus(state->iocp,
&state->entries[numComplete].dwNumberOfBytesTransferred,
&state->entries[numComplete].lpCompletionKey,
&state->entries[numComplete].lpOverlapped,
0);
if (lrc) {
numComplete++;
} else {
if (state->entries[numComplete].lpOverlapped == NULL) break;
}
}
}
}
if (rc && numComplete > 0) {
LPOVERLAPPED_ENTRY entry = state->entries;
for (j = 0; j < numComplete && numevents < state->setsize; j++, entry++) {
/* the competion key is the socket */
int rfd = (int)entry->lpCompletionKey;
sockstate = aeGetExistingSockState(state, rfd);
if (sockstate != NULL) {
if ((sockstate->masks & LISTEN_SOCK) && entry->lpOverlapped != NULL) {
/* need to set event for listening */
aacceptreq *areq = (aacceptreq *)entry->lpOverlapped;
areq->next = sockstate->reqs;
sockstate->reqs = areq;
sockstate->masks &= ~ACCEPT_PENDING;
if (sockstate->masks & AE_READABLE) {
eventLoop->fired[numevents].fd = rfd;
eventLoop->fired[numevents].mask = AE_READABLE;
numevents++;
}
} else if (sockstate->masks & CONNECT_PENDING) {
/* check if connect complete */
if (entry->lpOverlapped == &sockstate->ov_read) {
sockstate->masks &= ~CONNECT_PENDING;
/* enable read and write events for this connection */
aeApiAddEvent(eventLoop, rfd, sockstate->masks);
}
} else {
int matched = 0;
/* check if event is read complete (may be 0 length read) */
if (entry->lpOverlapped == &sockstate->ov_read) {
matched = 1;
sockstate->masks &= ~READ_QUEUED;
if (sockstate->masks & AE_READABLE) {
eventLoop->fired[numevents].fd = rfd;
eventLoop->fired[numevents].mask = AE_READABLE;
numevents++;
}
} else if (sockstate->wreqs > 0 && entry->lpOverlapped != NULL) {
/* should be write complete. Get results */
asendreq *areq = (asendreq *)entry->lpOverlapped;
matched = removeMatchFromList(&sockstate->wreqlist, areq);
if (matched) {
/* call write complete callback so buffers can be freed */
if (areq->proc != NULL) {
DWORD written = 0;
DWORD flags;
WSAGetOverlappedResult(rfd, &areq->ov, &written, FALSE, &flags);
areq->proc(areq->eventLoop, rfd, &areq->req, (int)written);
}
sockstate->wreqs--;
zfree(areq);
/* if no active write requests, set ready to write */
if (sockstate->wreqs == 0 && sockstate->masks & AE_WRITABLE) {
eventLoop->fired[numevents].fd = rfd;
eventLoop->fired[numevents].mask = AE_WRITABLE;
numevents++;
}
}
}
if (matched == 0) {
/* redisLog */printf("Sec:%lld Unknown complete (closed) on %d\n", gettimeofdaysecs(NULL), rfd);
sockstate = NULL;
}
}
} else {
// no match for active connection.
// Try the closing list.
list *socklist = &(state->closing);
listNode *node;
node = listFirst(socklist);
while (node != NULL) {
sockstate = (aeSockState *)listNodeValue(node);
if (sockstate->fd == rfd) {
if (sockstate->masks & CONNECT_PENDING) {
/* check if connect complete */
if (entry->lpOverlapped == &sockstate->ov_read) {
sockstate->masks &= ~CONNECT_PENDING;
}
} else if (entry->lpOverlapped == &sockstate->ov_read) {
// read complete
sockstate->masks &= ~READ_QUEUED;
} else {
// check pending writes
asendreq *areq = (asendreq *)entry->lpOverlapped;
if (removeMatchFromList(&sockstate->wreqlist, areq)) {
sockstate->wreqs--;
zfree(areq);
}
}
if (sockstate->wreqs == 0 &&
(sockstate->masks & (CONNECT_PENDING | READ_QUEUED | SOCKET_ATTACHED)) == 0) {
if ((sockstate->masks & CLOSE_PENDING) != 0) {
close(rfd);
sockstate->masks &= ~(CLOSE_PENDING);
}
// safe to delete sockstate
aeDelSockState(state, sockstate);
}
break;
}
node = listNextNode(node);
}
}
}
}
return numevents;
}
int staticFileCall(struct conn *c, void *arg)
{
wstr path = arg;
struct staticFileData *static_data;
off_t len;
time_t m_time;
int ret;
struct stat stat;
static_data = c->app_private_data;
static_data->fd = open(path, O_RDONLY);
if (static_data->fd == -1) {
wheatLog(WHEAT_VERBOSE, "open file failed: %s", strerror(errno));
goto failed404;
}
ret = lstat(path, &stat);
if (ret == -1)
goto failed404;
if (!S_ISREG(stat.st_mode)) {
close(static_data->fd);
static_data->fd = -1;
if (S_ISDIR(stat.st_mode) && DirectoryIndex) {
struct listNode *node;
struct listIterator *iter;
wstr last;
char append_path[255];
iter = listGetIterator(DirectoryIndex, START_HEAD);
while ((node = listNext(iter)) != NULL) {
last = listNodeValue(node);
snprintf(append_path, 255, "%s/%s", path, last);
static_data->fd = open(append_path, O_RDONLY);
if (static_data->fd != -1) {
break;
}
}
}
if (static_data->fd == -1) {
wheatLog(WHEAT_VERBOSE, "open file failed: %s", strerror(errno));
goto failed404;
}
}
if (AllowExtensions && static_data->extension &&
!dictFetchValue(AllowExtensions, static_data->extension)) {
goto failed404;
}
ret = getFileSizeAndMtime(static_data->fd , &len, &m_time);
if (ret == WHEAT_WRONG) {
wheatLog(WHEAT_VERBOSE, "open file failed: %s", strerror(errno));
goto failed404;
}
if (len > MaxFileSize) {
wheatLog(WHEAT_NOTICE, "file exceed max limit %d", len);
wheatLog(WHEAT_VERBOSE, "open file failed: %s", strerror(errno));
goto failed404;
}
wstr modified = dictFetchValue(httpGetReqHeaders(c), IfModifiedSince);
if (modified != NULL) {
char buf[wstrlen(modified)];
memcpy(buf, modified, wstrlen(modified));
time_t client_m_time = fromHttpDate(buf);
if (m_time <= client_m_time) {
fillResInfo(c, 304, "Not Modified");
ret = fillResHeaders(c, 0, 0);
if (ret == -1)
goto failed;
ret = httpSendHeaders(c);
if (ret == -1)
goto failed;
return WHEAT_OK;
}
}
fillResInfo(c, 200, "OK");
ret = fillResHeaders(c, len, m_time);
if (ret == -1) {
wheatLog(WHEAT_WARNING, "fill Res Headers failes: %s", strerror(errno));
goto failed;
}
ret = httpSendHeaders(c);
if (ret == -1) {
wheatLog(WHEAT_WARNING, "static file send headers failed: %s", strerror(errno));
goto failed;
}
ret = sendClientFile(c, static_data->fd , len);
if (ret == WHEAT_WRONG) {
wheatLog(WHEAT_WARNING, "send static file failed: %s", strerror(errno));
goto failed;
}
return WHEAT_OK;
failed404:
sendResponse404(c);
return WHEAT_OK;
failed:
setClientClose(c);
return WHEAT_OK;
}
int
main(int argc, char *argv[])
{
unsigned char *zl, *p;
unsigned char *entry;
unsigned int elen;
long long value;
/* If an argument is given, use it as the random seed. */
if (argc == 2)
srand(atoi(argv[1]));
zl = createIntList();
ziplistRepr(zl);
zl = createList();
ziplistRepr(zl);
pop(zl,ZIPLIST_TAIL);
ziplistRepr(zl);
pop(zl,ZIPLIST_HEAD);
ziplistRepr(zl);
pop(zl,ZIPLIST_TAIL);
ziplistRepr(zl);
pop(zl,ZIPLIST_TAIL);
ziplistRepr(zl);
printf("Get element at index 3:\n");
{
zl = createList();
p = ziplistIndex(zl, 3);
if (!ziplistGet(p, &entry, &elen, &value)) {
printf("ERROR: Could not access index 3\n");
return 1;
}
if (entry) {
if (elen && fwrite(entry,elen,1,stdout) == 0) perror("fwrite");
printf("\n");
} else {
printf("%lld\n", value);
}
printf("\n");
}
printf("Get element at index 4 (out of range):\n");
{
zl = createList();
p = ziplistIndex(zl, 4);
if (p == NULL) {
printf("No entry\n");
} else {
printf("ERROR: Out of range index should return NULL, returned offset: %ld\n", p-zl);
return 1;
}
printf("\n");
}
printf("Get element at index -1 (last element):\n");
{
zl = createList();
p = ziplistIndex(zl, -1);
if (!ziplistGet(p, &entry, &elen, &value)) {
printf("ERROR: Could not access index -1\n");
return 1;
}
if (entry) {
if (elen && fwrite(entry,elen,1,stdout) == 0) perror("fwrite");
printf("\n");
} else {
printf("%lld\n", value);
}
printf("\n");
}
printf("Get element at index -4 (first element):\n");
{
zl = createList();
p = ziplistIndex(zl, -4);
if (!ziplistGet(p, &entry, &elen, &value)) {
printf("ERROR: Could not access index -4\n");
return 1;
}
if (entry) {
if (elen && fwrite(entry,elen,1,stdout) == 0) perror("fwrite");
printf("\n");
} else {
printf("%lld\n", value);
}
printf("\n");
}
printf("Get element at index -5 (reverse out of range):\n");
{
zl = createList();
p = ziplistIndex(zl, -5);
if (p == NULL) {
printf("No entry\n");
} else {
printf("ERROR: Out of range index should return NULL, returned offset: %ld\n", p-zl);
return 1;
}
printf("\n");
}
printf("Iterate list from 0 to end:\n");
{
zl = createList();
p = ziplistIndex(zl, 0);
while (ziplistGet(p, &entry, &elen, &value)) {
printf("Entry: ");
if (entry) {
if (elen && fwrite(entry,elen,1,stdout) == 0) perror("fwrite");
} else {
printf("%lld", value);
}
p = ziplistNext(zl,p);
printf("\n");
}
printf("\n");
}
printf("Iterate list from 1 to end:\n");
{
zl = createList();
p = ziplistIndex(zl, 1);
while (ziplistGet(p, &entry, &elen, &value)) {
printf("Entry: ");
if (entry) {
if (elen && fwrite(entry,elen,1,stdout) == 0) perror("fwrite");
} else {
printf("%lld", value);
}
p = ziplistNext(zl,p);
printf("\n");
}
printf("\n");
}
printf("Iterate list from 2 to end:\n");
{
zl = createList();
p = ziplistIndex(zl, 2);
while (ziplistGet(p, &entry, &elen, &value)) {
printf("Entry: ");
if (entry) {
if (elen && fwrite(entry,elen,1,stdout) == 0) perror("fwrite");
} else {
printf("%lld", value);
}
p = ziplistNext(zl,p);
printf("\n");
}
printf("\n");
}
printf("Iterate starting out of range:\n");
{
zl = createList();
p = ziplistIndex(zl, 4);
if (!ziplistGet(p, &entry, &elen, &value)) {
printf("No entry\n");
} else {
printf("ERROR\n");
}
printf("\n");
}
printf("Iterate from back to front:\n");
{
zl = createList();
p = ziplistIndex(zl, -1);
while (ziplistGet(p, &entry, &elen, &value)) {
printf("Entry: ");
if (entry) {
if (elen && fwrite(entry,elen,1,stdout) == 0) perror("fwrite");
} else {
printf("%lld", value);
}
p = ziplistPrev(zl,p);
printf("\n");
}
printf("\n");
}
printf("Iterate from back to front, deleting all items:\n");
{
zl = createList();
p = ziplistIndex(zl, -1);
while (ziplistGet(p, &entry, &elen, &value)) {
printf("Entry: ");
if (entry) {
if (elen && fwrite(entry,elen,1,stdout) == 0) perror("fwrite");
} else {
printf("%lld", value);
}
zl = ziplistDelete(zl,&p);
p = ziplistPrev(zl,p);
printf("\n");
}
printf("\n");
}
printf("Delete inclusive range 0,0:\n");
{
zl = createList();
zl = ziplistDeleteRange(zl, 0, 1);
ziplistRepr(zl);
}
printf("Delete inclusive range 0,1:\n");
{
zl = createList();
zl = ziplistDeleteRange(zl, 0, 2);
ziplistRepr(zl);
}
printf("Delete inclusive range 1,2:\n");
{
zl = createList();
zl = ziplistDeleteRange(zl, 1, 2);
ziplistRepr(zl);
}
printf("Delete with start index out of range:\n");
{
zl = createList();
zl = ziplistDeleteRange(zl, 5, 1);
ziplistRepr(zl);
}
printf("Delete with num overflow:\n");
{
zl = createList();
zl = ziplistDeleteRange(zl, 1, 5);
ziplistRepr(zl);
}
printf("Delete foo while iterating:\n");
{
zl = createList();
p = ziplistIndex(zl,0);
while (ziplistGet(p,&entry,&elen,&value)) {
if (entry && strncmp("foo",(char*)entry,elen) == 0) {
printf("Delete foo\n");
zl = ziplistDelete(zl,&p);
} else {
printf("Entry: ");
if (entry) {
if (elen && fwrite(entry,elen,1,stdout) == 0)
perror("fwrite");
} else {
printf("%lld",value);
}
p = ziplistNext(zl,p);
printf("\n");
}
}
printf("\n");
ziplistRepr(zl);
}
printf("Regression test for >255 byte strings:\n");
{
char v1[257],v2[257];
memset(v1,'x',256);
memset(v2,'y',256);
zl = ziplistNew();
zl = ziplistPush(zl,(unsigned char*)v1,strlen(v1),ZIPLIST_TAIL);
zl = ziplistPush(zl,(unsigned char*)v2,strlen(v2),ZIPLIST_TAIL);
/* Pop values again and compare their value. */
p = ziplistIndex(zl,0);
assert(ziplistGet(p,&entry,&elen,&value));
assert(strncmp(v1,(char*)entry,elen) == 0);
p = ziplistIndex(zl,1);
assert(ziplistGet(p,&entry,&elen,&value));
assert(strncmp(v2,(char*)entry,elen) == 0);
printf("SUCCESS\n\n");
}
printf("Regression test deleting next to last entries:\n");
{
char v[3][257];
zlentry e[3];
int i;
for (i = 0; i < (sizeof(v)/sizeof(v[0])); i++) {
memset(v[i], 'a' + i, sizeof(v[0]));
}
v[0][256] = '\0';
v[1][ 1] = '\0';
v[2][256] = '\0';
zl = ziplistNew();
for (i = 0; i < (sizeof(v)/sizeof(v[0])); i++) {
zl = ziplistPush(zl, (unsigned char *) v[i], strlen(v[i]), ZIPLIST_TAIL);
}
verify(zl, e);
assert(e[0].prevrawlensize == 1);
assert(e[1].prevrawlensize == 5);
assert(e[2].prevrawlensize == 1);
/* Deleting entry 1 will increase `prevrawlensize` for entry 2 */
unsigned char *p = e[1].p;
zl = ziplistDelete(zl, &p);
verify(zl, e);
assert(e[0].prevrawlensize == 1);
assert(e[1].prevrawlensize == 5);
printf("SUCCESS\n\n");
}
printf("Create long list and check indices:\n");
{
zl = ziplistNew();
char buf[32];
int i,len;
for (i = 0; i < 1000; i++) {
len = sprintf(buf,"%d",i);
zl = ziplistPush(zl,(unsigned char*)buf,len,ZIPLIST_TAIL);
}
for (i = 0; i < 1000; i++) {
p = ziplistIndex(zl,i);
assert(ziplistGet(p,NULL,NULL,&value));
assert(i == value);
p = ziplistIndex(zl,-i-1);
assert(ziplistGet(p,NULL,NULL,&value));
assert(999-i == value);
}
printf("SUCCESS\n\n");
}
printf("Compare strings with ziplist entries:\n");
{
zl = createList();
p = ziplistIndex(zl,0);
if (!ziplistCompare(p,(unsigned char*)"hello",5)) {
printf("ERROR: not \"hello\"\n");
return 1;
}
if (ziplistCompare(p,(unsigned char*)"hella",5)) {
printf("ERROR: \"hella\"\n");
return 1;
}
p = ziplistIndex(zl,3);
if (!ziplistCompare(p,(unsigned char*)"1024",4)) {
printf("ERROR: not \"1024\"\n");
return 1;
}
if (ziplistCompare(p,(unsigned char*)"1025",4)) {
printf("ERROR: \"1025\"\n");
return 1;
}
printf("SUCCESS\n\n");
}
printf("Stress with random payloads of different encoding:\n");
{
int i,j,len,where;
unsigned char *p;
char buf[1024];
int buflen;
list *ref;
listNode *refnode;
/* Hold temp vars from ziplist */
unsigned char *sstr;
unsigned int slen;
long long sval;
for (i = 0; i < 20000; i++) {
zl = ziplistNew();
ref = listCreate();
listSetFreeMethod(ref, sdsfree);
len = rand() % 256;
/* Create lists */
for (j = 0; j < len; j++) {
where = (rand() & 1) ? ZIPLIST_HEAD : ZIPLIST_TAIL;
if (rand() % 2) {
buflen = randstring(buf,1,sizeof(buf)-1);
} else {
switch(rand() % 3) {
case 0:
buflen = sprintf(buf,"%lld",(0LL + rand()) >> 20);
break;
case 1:
buflen = sprintf(buf,"%lld",(0LL + rand()));
break;
case 2:
buflen = sprintf(buf,"%lld",(0LL + rand()) << 20);
break;
default:
assert(NULL);
}
}
/* Add to ziplist */
zl = ziplistPush(zl, (unsigned char*)buf, buflen, where);
/* Add to reference list */
if (where == ZIPLIST_HEAD) {
listAddNodeHead(ref,sdsnewlen(buf, buflen));
} else if (where == ZIPLIST_TAIL) {
listAddNodeTail(ref,sdsnewlen(buf, buflen));
} else {
assert(NULL);
}
}
assert(listLength(ref) == ziplistLen(zl));
for (j = 0; j < len; j++) {
/* Naive way to get elements, but similar to the stresser
* executed from the Tcl test suite. */
p = ziplistIndex(zl,j);
refnode = listIndex(ref,j);
assert(ziplistGet(p,&sstr,&slen,&sval));
if (sstr == NULL) {
buflen = sprintf(buf,"%lld",sval);
} else {
buflen = slen;
memcpy(buf,sstr,buflen);
buf[buflen] = '\0';
}
assert(memcmp(buf,listNodeValue(refnode),buflen) == 0);
}
zfree(zl);
listRelease(ref);
}
printf("SUCCESS\n\n");
}
printf("Stress with variable ziplist size:\n");
{
stress(ZIPLIST_HEAD,100000,16384,256);
stress(ZIPLIST_TAIL,100000,16384,256);
}
return 0;
}
static int send_msg_to_all(check_unit *cunit, struct msg *msg)
{
int ret;
thread_data *cdata = cunit->cdata;
redis_group *trgroup = cdata->trgroup;
redis_node *trnode;
struct msg *msg_same;
listNode *lnode;
struct mbuf *mbuf;
if (cunit == NULL || msg == NULL) {
return RMT_ERROR;
}
msg_same = msg_get(msg->mb, msg->request, msg->kind);
if (msg_same == NULL) {
log_error("ERROR: msg clone failed.");
msg_put(msg);
msg_free(msg);
msg = NULL;
return RMT_ERROR;
}
lnode = listFirst(msg->data);
while (lnode) {
mbuf = listNodeValue(lnode);
lnode = lnode->next;
ret = msg_append_full(msg_same, mbuf->pos, mbuf_length(mbuf));
if (ret != RMT_OK) {
log_error("ERROR: out of memory.");
msg_put(msg_same);
msg_free(msg_same);
msg = NULL;
return RMT_ERROR;
}
}
msg_same->ptr = msg->ptr;
msg_same->resp_check = msg->resp_check;
ret = prepare_send_msg(cunit->srnode, msg, cunit->srnode);
if (ret != RMT_OK) {
msg_put(msg);
msg_free(msg);
msg = NULL;
msg_put(msg_same);
msg_free(msg_same);
return RMT_ERROR;
}
msg = NULL;
trnode = trgroup->get_backend_node(trgroup, (uint8_t *)cunit->key, (uint32_t)sdslen(cunit->key));
if(prepare_send_msg(trnode, msg_same, trnode) != RMT_OK){
msg_put(msg_same);
msg_free(msg_same);
return RMT_ERROR;
}
return RMT_OK;
}
