hiredis/read.c
michael-grunder 7bef042c03 Use string2ll from Redis
This commit pulls string2ll from Redis (with permission from Antirez)
as strtoll is 2-3x slower and even worse vs the original version in
hiredis that didn't check for overflow at all.

By using string2ll there is almost no measurable performance impact
of overflow detection even in integer parsing heavy workloads (e.g.
INCRBY commands).
2018-05-20 10:44:19 -07:00

599 lines
17 KiB
C

/*
* Copyright (c) 2009-2011, Salvatore Sanfilippo <antirez at gmail dot com>
* Copyright (c) 2010-2011, Pieter Noordhuis <pcnoordhuis at gmail dot com>
*
* All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions are met:
*
* * Redistributions of source code must retain the above copyright notice,
* this list of conditions and the following disclaimer.
* * Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
* * Neither the name of Redis nor the names of its contributors may be used
* to endorse or promote products derived from this software without
* specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
* AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
* ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE
* LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
* CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
* SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
* INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
* CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
* ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
* POSSIBILITY OF SUCH DAMAGE.
*/
#include "fmacros.h"
#include <string.h>
#include <stdlib.h>
#ifndef _MSC_VER
#include <unistd.h>
#endif
#include <assert.h>
#include <errno.h>
#include <ctype.h>
#include <limits.h>
#include "read.h"
#include "sds.h"
static void __redisReaderSetError(redisReader *r, int type, const char *str) {
size_t len;
if (r->reply != NULL && r->fn && r->fn->freeObject) {
r->fn->freeObject(r->reply);
r->reply = NULL;
}
/* Clear input buffer on errors. */
sdsfree(r->buf);
r->buf = NULL;
r->pos = r->len = 0;
/* Reset task stack. */
r->ridx = -1;
/* Set error. */
r->err = type;
len = strlen(str);
len = len < (sizeof(r->errstr)-1) ? len : (sizeof(r->errstr)-1);
memcpy(r->errstr,str,len);
r->errstr[len] = '\0';
}
static size_t chrtos(char *buf, size_t size, char byte) {
size_t len = 0;
switch(byte) {
case '\\':
case '"':
len = snprintf(buf,size,"\"\\%c\"",byte);
break;
case '\n': len = snprintf(buf,size,"\"\\n\""); break;
case '\r': len = snprintf(buf,size,"\"\\r\""); break;
case '\t': len = snprintf(buf,size,"\"\\t\""); break;
case '\a': len = snprintf(buf,size,"\"\\a\""); break;
case '\b': len = snprintf(buf,size,"\"\\b\""); break;
default:
if (isprint(byte))
len = snprintf(buf,size,"\"%c\"",byte);
else
len = snprintf(buf,size,"\"\\x%02x\"",(unsigned char)byte);
break;
}
return len;
}
static void __redisReaderSetErrorProtocolByte(redisReader *r, char byte) {
char cbuf[8], sbuf[128];
chrtos(cbuf,sizeof(cbuf),byte);
snprintf(sbuf,sizeof(sbuf),
"Protocol error, got %s as reply type byte", cbuf);
__redisReaderSetError(r,REDIS_ERR_PROTOCOL,sbuf);
}
static void __redisReaderSetErrorOOM(redisReader *r) {
__redisReaderSetError(r,REDIS_ERR_OOM,"Out of memory");
}
static char *readBytes(redisReader *r, unsigned int bytes) {
char *p;
if (r->len-r->pos >= bytes) {
p = r->buf+r->pos;
r->pos += bytes;
return p;
}
return NULL;
}
/* Find pointer to \r\n. */
static char *seekNewline(char *s, size_t len) {
int pos = 0;
int _len = len-1;
/* Position should be < len-1 because the character at "pos" should be
* followed by a \n. Note that strchr cannot be used because it doesn't
* allow to search a limited length and the buffer that is being searched
* might not have a trailing NULL character. */
while (pos < _len) {
while(pos < _len && s[pos] != '\r') pos++;
if (pos==_len) {
/* Not found. */
return NULL;
} else {
if (s[pos+1] == '\n') {
/* Found. */
return s+pos;
} else {
/* Continue searching. */
pos++;
}
}
}
return NULL;
}
/* Convert a string into a long long. Returns REDIS_OK if the string could be
* parsed into a (non-overflowing) long long, REDIS_ERR otherwise. The value
* will be set to the parsed value when appropriate.
*
* Note that this function demands that the string strictly represents
* a long long: no spaces or other characters before or after the string
* representing the number are accepted, nor zeroes at the start if not
* for the string "0" representing the zero number.
*
* Because of its strictness, it is safe to use this function to check if
* you can convert a string into a long long, and obtain back the string
* from the number without any loss in the string representation. */
int string2ll(const char *s, size_t slen, long long *value) {
const char *p = s;
size_t plen = 0;
int negative = 0;
unsigned long long v;
if (plen == slen)
return REDIS_ERR;
/* Special case: first and only digit is 0. */
if (slen == 1 && p[0] == '0') {
if (value != NULL) *value = 0;
return REDIS_OK;
}
if (p[0] == '-') {
negative = 1;
p++; plen++;
/* Abort on only a negative sign. */
if (plen == slen)
return REDIS_ERR;
}
/* First digit should be 1-9, otherwise the string should just be 0. */
if (p[0] >= '1' && p[0] <= '9') {
v = p[0]-'0';
p++; plen++;
} else if (p[0] == '0' && slen == 1) {
*value = 0;
return REDIS_OK;
} else {
return REDIS_ERR;
}
while (plen < slen && p[0] >= '0' && p[0] <= '9') {
if (v > (ULLONG_MAX / 10)) /* Overflow. */
return REDIS_ERR;
v *= 10;
if (v > (ULLONG_MAX - (p[0]-'0'))) /* Overflow. */
return REDIS_ERR;
v += p[0]-'0';
p++; plen++;
}
/* Return if not all bytes were used. */
if (plen < slen)
return REDIS_ERR;
if (negative) {
if (v > ((unsigned long long)(-(LLONG_MIN+1))+1)) /* Overflow. */
return REDIS_ERR;
if (value != NULL) *value = -v;
} else {
if (v > LLONG_MAX) /* Overflow. */
return REDIS_ERR;
if (value != NULL) *value = v;
}
return REDIS_OK;
}
static char *readLine(redisReader *r, int *_len) {
char *p, *s;
int len;
p = r->buf+r->pos;
s = seekNewline(p,(r->len-r->pos));
if (s != NULL) {
len = s-(r->buf+r->pos);
r->pos += len+2; /* skip \r\n */
if (_len) *_len = len;
return p;
}
return NULL;
}
static void moveToNextTask(redisReader *r) {
redisReadTask *cur, *prv;
while (r->ridx >= 0) {
/* Return a.s.a.p. when the stack is now empty. */
if (r->ridx == 0) {
r->ridx--;
return;
}
cur = &(r->rstack[r->ridx]);
prv = &(r->rstack[r->ridx-1]);
assert(prv->type == REDIS_REPLY_ARRAY);
if (cur->idx == prv->elements-1) {
r->ridx--;
} else {
/* Reset the type because the next item can be anything */
assert(cur->idx < prv->elements);
cur->type = -1;
cur->elements = -1;
cur->idx++;
return;
}
}
}
static int processLineItem(redisReader *r) {
redisReadTask *cur = &(r->rstack[r->ridx]);
void *obj;
char *p;
int len;
if ((p = readLine(r,&len)) != NULL) {
if (cur->type == REDIS_REPLY_INTEGER) {
if (r->fn && r->fn->createInteger) {
long long v;
if (string2ll(p, len, &v) == REDIS_ERR) {
__redisReaderSetError(r,REDIS_ERR_PROTOCOL,
"Bad integer value");
return REDIS_ERR;
}
obj = r->fn->createInteger(cur,v);
} else {
obj = (void*)REDIS_REPLY_INTEGER;
}
} else {
/* Type will be error or status. */
if (r->fn && r->fn->createString)
obj = r->fn->createString(cur,p,len);
else
obj = (void*)(size_t)(cur->type);
}
if (obj == NULL) {
__redisReaderSetErrorOOM(r);
return REDIS_ERR;
}
/* Set reply if this is the root object. */
if (r->ridx == 0) r->reply = obj;
moveToNextTask(r);
return REDIS_OK;
}
return REDIS_ERR;
}
static int processBulkItem(redisReader *r) {
redisReadTask *cur = &(r->rstack[r->ridx]);
void *obj = NULL;
char *p, *s;
long long len;
unsigned long bytelen;
int success = 0;
p = r->buf+r->pos;
s = seekNewline(p,r->len-r->pos);
if (s != NULL) {
p = r->buf+r->pos;
bytelen = s-(r->buf+r->pos)+2; /* include \r\n */
if (string2ll(p, bytelen - 2, &len) == REDIS_ERR) {
__redisReaderSetError(r,REDIS_ERR_PROTOCOL,
"Bad bulk string length");
return REDIS_ERR;
}
if (len < -1 || (LLONG_MAX > SIZE_MAX && len > (long long)SIZE_MAX)) {
__redisReaderSetError(r,REDIS_ERR_PROTOCOL,
"Bulk string length out of range");
return REDIS_ERR;
}
if (len == -1) {
/* The nil object can always be created. */
if (r->fn && r->fn->createNil)
obj = r->fn->createNil(cur);
else
obj = (void*)REDIS_REPLY_NIL;
success = 1;
} else {
/* Only continue when the buffer contains the entire bulk item. */
bytelen += len+2; /* include \r\n */
if (r->pos+bytelen <= r->len) {
if (r->fn && r->fn->createString)
obj = r->fn->createString(cur,s+2,len);
else
obj = (void*)REDIS_REPLY_STRING;
success = 1;
}
}
/* Proceed when obj was created. */
if (success) {
if (obj == NULL) {
__redisReaderSetErrorOOM(r);
return REDIS_ERR;
}
r->pos += bytelen;
/* Set reply if this is the root object. */
if (r->ridx == 0) r->reply = obj;
moveToNextTask(r);
return REDIS_OK;
}
}
return REDIS_ERR;
}
static int processMultiBulkItem(redisReader *r) {
redisReadTask *cur = &(r->rstack[r->ridx]);
void *obj;
char *p;
long long elements;
int root = 0, len;
/* Set error for nested multi bulks with depth > 7 */
if (r->ridx == 8) {
__redisReaderSetError(r,REDIS_ERR_PROTOCOL,
"No support for nested multi bulk replies with depth > 7");
return REDIS_ERR;
}
if ((p = readLine(r,&len)) != NULL) {
if (string2ll(p, len, &elements) == REDIS_ERR) {
__redisReaderSetError(r,REDIS_ERR_PROTOCOL,
"Bad multi-bulk length");
return REDIS_ERR;
}
root = (r->ridx == 0);
if (elements < -1 || elements > INT_MAX) {
__redisReaderSetError(r,REDIS_ERR_PROTOCOL,
"Multi-bulk length out of range");
return REDIS_ERR;
}
if (elements == -1) {
if (r->fn && r->fn->createNil)
obj = r->fn->createNil(cur);
else
obj = (void*)REDIS_REPLY_NIL;
if (obj == NULL) {
__redisReaderSetErrorOOM(r);
return REDIS_ERR;
}
moveToNextTask(r);
} else {
if (r->fn && r->fn->createArray)
obj = r->fn->createArray(cur,elements);
else
obj = (void*)REDIS_REPLY_ARRAY;
if (obj == NULL) {
__redisReaderSetErrorOOM(r);
return REDIS_ERR;
}
/* Modify task stack when there are more than 0 elements. */
if (elements > 0) {
cur->elements = elements;
cur->obj = obj;
r->ridx++;
r->rstack[r->ridx].type = -1;
r->rstack[r->ridx].elements = -1;
r->rstack[r->ridx].idx = 0;
r->rstack[r->ridx].obj = NULL;
r->rstack[r->ridx].parent = cur;
r->rstack[r->ridx].privdata = r->privdata;
} else {
moveToNextTask(r);
}
}
/* Set reply if this is the root object. */
if (root) r->reply = obj;
return REDIS_OK;
}
return REDIS_ERR;
}
static int processItem(redisReader *r) {
redisReadTask *cur = &(r->rstack[r->ridx]);
char *p;
/* check if we need to read type */
if (cur->type < 0) {
if ((p = readBytes(r,1)) != NULL) {
switch (p[0]) {
case '-':
cur->type = REDIS_REPLY_ERROR;
break;
case '+':
cur->type = REDIS_REPLY_STATUS;
break;
case ':':
cur->type = REDIS_REPLY_INTEGER;
break;
case '$':
cur->type = REDIS_REPLY_STRING;
break;
case '*':
cur->type = REDIS_REPLY_ARRAY;
break;
default:
__redisReaderSetErrorProtocolByte(r,*p);
return REDIS_ERR;
}
} else {
/* could not consume 1 byte */
return REDIS_ERR;
}
}
/* process typed item */
switch(cur->type) {
case REDIS_REPLY_ERROR:
case REDIS_REPLY_STATUS:
case REDIS_REPLY_INTEGER:
return processLineItem(r);
case REDIS_REPLY_STRING:
return processBulkItem(r);
case REDIS_REPLY_ARRAY:
return processMultiBulkItem(r);
default:
assert(NULL);
return REDIS_ERR; /* Avoid warning. */
}
}
redisReader *redisReaderCreateWithFunctions(redisReplyObjectFunctions *fn) {
redisReader *r;
r = calloc(1,sizeof(redisReader));
if (r == NULL)
return NULL;
r->fn = fn;
r->buf = sdsempty();
r->maxbuf = REDIS_READER_MAX_BUF;
if (r->buf == NULL) {
free(r);
return NULL;
}
r->ridx = -1;
return r;
}
void redisReaderFree(redisReader *r) {
if (r == NULL)
return;
if (r->reply != NULL && r->fn && r->fn->freeObject)
r->fn->freeObject(r->reply);
sdsfree(r->buf);
free(r);
}
int redisReaderFeed(redisReader *r, const char *buf, size_t len) {
sds newbuf;
/* Return early when this reader is in an erroneous state. */
if (r->err)
return REDIS_ERR;
/* Copy the provided buffer. */
if (buf != NULL && len >= 1) {
/* Destroy internal buffer when it is empty and is quite large. */
if (r->len == 0 && r->maxbuf != 0 && sdsavail(r->buf) > r->maxbuf) {
sdsfree(r->buf);
r->buf = sdsempty();
r->pos = 0;
/* r->buf should not be NULL since we just free'd a larger one. */
assert(r->buf != NULL);
}
newbuf = sdscatlen(r->buf,buf,len);
if (newbuf == NULL) {
__redisReaderSetErrorOOM(r);
return REDIS_ERR;
}
r->buf = newbuf;
r->len = sdslen(r->buf);
}
return REDIS_OK;
}
int redisReaderGetReply(redisReader *r, void **reply) {
/* Default target pointer to NULL. */
if (reply != NULL)
*reply = NULL;
/* Return early when this reader is in an erroneous state. */
if (r->err)
return REDIS_ERR;
/* When the buffer is empty, there will never be a reply. */
if (r->len == 0)
return REDIS_OK;
/* Set first item to process when the stack is empty. */
if (r->ridx == -1) {
r->rstack[0].type = -1;
r->rstack[0].elements = -1;
r->rstack[0].idx = -1;
r->rstack[0].obj = NULL;
r->rstack[0].parent = NULL;
r->rstack[0].privdata = r->privdata;
r->ridx = 0;
}
/* Process items in reply. */
while (r->ridx >= 0)
if (processItem(r) != REDIS_OK)
break;
/* Return ASAP when an error occurred. */
if (r->err)
return REDIS_ERR;
/* Discard part of the buffer when we've consumed at least 1k, to avoid
* doing unnecessary calls to memmove() in sds.c. */
if (r->pos >= 1024) {
sdsrange(r->buf,r->pos,-1);
r->pos = 0;
r->len = sdslen(r->buf);
}
/* Emit a reply when there is one. */
if (r->ridx == -1) {
if (reply != NULL)
*reply = r->reply;
r->reply = NULL;
}
return REDIS_OK;
}