libpcap:pcap_breakloop()导致内存泄漏

Ian*_*anC 5 c memory-leaks pthreads libpcap

在使用Linux pthreadlibpcap时,我注意到使用时有些奇怪的行为pcap_breakloop我的目标是: 打开一个新线程,该线程将运行pcap_loop并处理捕获的数据包,而主线程将执行其他操作。收到信号(SIGINT)或退出时,将设置一个全局变量,并且主线程将发出指令pcap_breakloop以结束第二个线程,然后完成自身。

这个想法可能仍然需要一些工作,但是我注意到的奇怪行为是:

当我们创建一个bpf_program结构和编译pcap_compile,以进一步使用它作为在捕获过滤器,一旦我们退出使用循环pcap_breakloop有8个字节(包括内存泄漏pcap_freecodepcap_close被调用)。即使不使用线程也是如此(就像在修改过的sniffex.c libpcap示例中所示)。没有过滤器(没有pcap_compile调用或bpf_program)的相同代码退出而没有内存泄漏

libpcap会发生这种内存泄漏,就像我在其他库上看到的某些内存泄漏一样gtk吗?图书馆以后会处理吗?还是我看不到我的代码有问题?

下面是代码示例:使用pthread和修改后的sniffex示例的示例pcap_breakloop(由于代码很大,但由于实际差异很小,所以我还包括了diff原始sniffex和修改后的示例):

pthreads示例:

#include <pcap.h>
#include <stdio.h>
#include <pthread.h>
#include <signal.h>

int CLOSE_PROGRAM = 0;
int CAPTURE_SETUP_SUCCESS = 0;
pcap_t *handler;    //pcap_t handler and filter structure are globals so we can free them from other
struct bpf_program fp;  //functions

//Prototypes
void INT_Handler(int signum);   //signal handler
void capture_loop(unsigned char *args, const struct pcap_pkthdr *header, const unsigned char *packet);  //pcap_loop function
void *thread_entrypoint(void *data);    //the entry point for the second thread

void INT_Handler(int signum){
    printf("Interrupt Signal Received\n");
    CLOSE_PROGRAM = 1;
}

void capture_loop(unsigned char *args, const struct pcap_pkthdr *header, const unsigned char *packet){
    printf("Captured Packet\n");
    //Process packet
}

void *thread_entrypoint(void *data){
    //We can't cancel the thread before we free libpcap stuff
    pthread_setcancelstate(PTHREAD_CANCEL_DISABLE, NULL);

    char error[PCAP_ERRBUF_SIZE];

    char filter_exp[] = "ip";
    bpf_u_int32 mask;
    bpf_u_int32 net;

    char *dev = pcap_lookupdev(error);
    if(NULL == dev){
        printf("Failed to look up device!\n");
        return NULL;
    }
    if(pcap_lookupnet(dev, &net, &mask, error) == -1){
        printf("Failed to look up netmask: %s\n", error);
        net = 0;
        mask = 0;
    }

    printf("Opening live capture on device %s\n", dev);
    handler = pcap_open_live(dev, 1518, 0, 1000, error);
    if(NULL == handler){
        printf("Failed to open live capture!\n");
        return NULL;
    }

    if(pcap_compile(handler, &fp, filter_exp, 0, net?net:PCAP_NETMASK_UNKNOWN) == -1){
        printf("Failed to compiler filter!\n");
        return NULL;
    }
    if(pcap_setfilter(handler, &fp) == -1){
        printf("Failed to install filter!\n");
        return NULL;
    }

    CAPTURE_SETUP_SUCCESS = 1;
    pcap_loop(handler, 0, capture_loop, NULL);

    pcap_freecode(&fp);
    pcap_close(handler);

    printf("Capture end\n");
    return NULL;
}

int main(int argc, char **argv){
    pthread_t capture_thread;

    pthread_create(&capture_thread, NULL, &thread_entrypoint, NULL);

    signal(SIGINT, &INT_Handler);

    while(0 == CLOSE_PROGRAM){
        //Wait for SIGINT
    }

    if(1 == CAPTURE_SETUP_SUCCESS){
        pcap_breakloop(handler);
    }
    pthread_join(capture_thread, NULL);

    return 0;
}
Run Code Online (Sandbox Code Playgroud)

修改后的sniffex示例:

/*
 * sniffex.c
 *
 * Sniffer example of TCP/IP packet capture using libpcap.
 * 
 * Version 0.1.1 (2005-07-05)
 * Copyright (c) 2005 The Tcpdump Group
 *
 * This software is intended to be used as a practical example and 
 * demonstration of the libpcap library; available at:
 * http://www.tcpdump.org/
 *
 ****************************************************************************
 *
 * This software is a modification of Tim Carstens' "sniffer.c"
 * demonstration source code, released as follows:
 * 
 * sniffer.c
 * Copyright (c) 2002 Tim Carstens
 * 2002-01-07
 * Demonstration of using libpcap
 * timcarst -at- yahoo -dot- com
 * 
 * "sniffer.c" is distributed under these terms:
 * 
 * Redistribution and use in source and binary forms, with or without
 * modification, are permitted provided that the following conditions
 * are met:
 * 1. Redistributions of source code must retain the above copyright
 *    notice, this list of conditions and the following disclaimer.
 * 2. 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.
 * 4. The name "Tim Carstens" may not be used to endorse or promote
 *    products derived from this software without prior written permission
 *
 * THIS SOFTWARE IS PROVIDED BY THE REGENTS 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 REGENTS 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.
 * <end of "sniffer.c" terms>
 *
 * This software, "sniffex.c", is a derivative work of "sniffer.c" and is
 * covered by the following terms:
 *
 * Redistribution and use in source and binary forms, with or without
 * modification, are permitted provided that the following conditions
 * are met:
 * 1. Because this is a derivative work, you must comply with the "sniffer.c"
 *    terms reproduced above.
 * 2. Redistributions of source code must retain the Tcpdump Group copyright
 *    notice at the top of this source file, this list of conditions and the
 *    following disclaimer.
 * 3. 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.
 * 4. The names "tcpdump" or "libpcap" may not be used to endorse or promote
 *    products derived from this software without prior written permission.
 *
 * THERE IS ABSOLUTELY NO WARRANTY FOR THIS PROGRAM.
 * BECAUSE THE PROGRAM IS LICENSED FREE OF CHARGE, THERE IS NO WARRANTY
 * FOR THE PROGRAM, TO THE EXTENT PERMITTED BY APPLICABLE LAW.  EXCEPT WHEN
 * OTHERWISE STATED IN WRITING THE COPYRIGHT HOLDERS AND/OR OTHER PARTIES
 * PROVIDE THE PROGRAM "AS IS" WITHOUT WARRANTY OF ANY KIND, EITHER EXPRESSED
 * OR IMPLIED, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF
 * MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE.  THE ENTIRE RISK AS
 * TO THE QUALITY AND PERFORMANCE OF THE PROGRAM IS WITH YOU.  SHOULD THE
 * PROGRAM PROVE DEFECTIVE, YOU ASSUME THE COST OF ALL NECESSARY SERVICING,
 * REPAIR OR CORRECTION.
 * 
 * IN NO EVENT UNLESS REQUIRED BY APPLICABLE LAW OR AGREED TO IN WRITING
 * WILL ANY COPYRIGHT HOLDER, OR ANY OTHER PARTY WHO MAY MODIFY AND/OR
 * REDISTRIBUTE THE PROGRAM AS PERMITTED ABOVE, BE LIABLE TO YOU FOR DAMAGES,
 * INCLUDING ANY GENERAL, SPECIAL, INCIDENTAL OR CONSEQUENTIAL DAMAGES ARISING
 * OUT OF THE USE OR INABILITY TO USE THE PROGRAM (INCLUDING BUT NOT LIMITED
 * TO LOSS OF DATA OR DATA BEING RENDERED INACCURATE OR LOSSES SUSTAINED BY
 * YOU OR THIRD PARTIES OR A FAILURE OF THE PROGRAM TO OPERATE WITH ANY OTHER
 * PROGRAMS), EVEN IF SUCH HOLDER OR OTHER PARTY HAS BEEN ADVISED OF THE
 * POSSIBILITY OF SUCH DAMAGES.
 * <end of "sniffex.c" terms>
 * 
 ****************************************************************************
 *
 * Below is an excerpt from an email from Guy Harris on the tcpdump-workers
 * mail list when someone asked, "How do I get the length of the TCP
 * payload?" Guy Harris' slightly snipped response (edited by him to
 * speak of the IPv4 header length and TCP data offset without referring
 * to bitfield structure members) is reproduced below:
 * 
 * The Ethernet size is always 14 bytes.
 * 
 * <snip>...</snip>
 *
 * In fact, you *MUST* assume the Ethernet header is 14 bytes, *and*, if 
 * you're using structures, you must use structures where the members 
 * always have the same size on all platforms, because the sizes of the 
 * fields in Ethernet - and IP, and TCP, and... - headers are defined by 
 * the protocol specification, not by the way a particular platform's C 
 * compiler works.)
 *
 * The IP header size, in bytes, is the value of the IP header length,
 * as extracted from the "ip_vhl" field of "struct sniff_ip" with
 * the "IP_HL()" macro, times 4 ("times 4" because it's in units of
 * 4-byte words).  If that value is less than 20 - i.e., if the value
 * extracted with "IP_HL()" is less than 5 - you have a malformed
 * IP datagram.
 *
 * The TCP header size, in bytes, is the value of the TCP data offset,
 * as extracted from the "th_offx2" field of "struct sniff_tcp" with
 * the "TH_OFF()" macro, times 4 (for the same reason - 4-byte words).
 * If that value is less than 20 - i.e., if the value extracted with
 * "TH_OFF()" is less than 5 - you have a malformed TCP segment.
 *
 * So, to find the IP header in an Ethernet packet, look 14 bytes after 
 * the beginning of the packet data.  To find the TCP header, look 
 * "IP_HL(ip)*4" bytes after the beginning of the IP header.  To find the
 * TCP payload, look "TH_OFF(tcp)*4" bytes after the beginning of the TCP
 * header.
 * 
 * To find out how much payload there is:
 *
 * Take the IP *total* length field - "ip_len" in "struct sniff_ip" 
 * - and, first, check whether it's less than "IP_HL(ip)*4" (after
 * you've checked whether "IP_HL(ip)" is >= 5).  If it is, you have
 * a malformed IP datagram.
 *
 * Otherwise, subtract "IP_HL(ip)*4" from it; that gives you the length
 * of the TCP segment, including the TCP header.  If that's less than
 * "TH_OFF(tcp)*4" (after you've checked whether "TH_OFF(tcp)" is >= 5),
 * you have a malformed TCP segment.
 *
 * Otherwise, subtract "TH_OFF(tcp)*4" from it; that gives you the
 * length of the TCP payload.
 *
 * Note that you also need to make sure that you don't go past the end 
 * of the captured data in the packet - you might, for example, have a 
 * 15-byte Ethernet packet that claims to contain an IP datagram, but if 
 * it's 15 bytes, it has only one byte of Ethernet payload, which is too 
 * small for an IP header.  The length of the captured data is given in 
 * the "caplen" field in the "struct pcap_pkthdr"; it might be less than 
 * the length of the packet, if you're capturing with a snapshot length 
 * other than a value >= the maximum packet size.
 * <end of response>
 * 
 ****************************************************************************
 * 
 * Example compiler command-line for GCC:
 *   gcc -Wall -o sniffex sniffex.c -lpcap
 * 
 ****************************************************************************
 *
 * Code Comments
 *
 * This section contains additional information and explanations regarding
 * comments in the source code. It serves as documentaion and rationale
 * for why the code is written as it is without hindering readability, as it
 * might if it were placed along with the actual code inline. References in
 * the code appear as footnote notation (e.g. [1]).
 *
 * 1. Ethernet headers are always exactly 14 bytes, so we define this
 * explicitly with "#define". Since some compilers might pad structures to a
 * multiple of 4 bytes - some versions of GCC for ARM may do this -
 * "sizeof (struct sniff_ethernet)" isn't used.
 * 
 * 2. Check the link-layer type of the device that's being opened to make
 * sure it's Ethernet, since that's all we handle in this example. Other
 * link-layer types may have different length headers (see [1]).
 *
 * 3. This is the filter expression that tells libpcap which packets we're
 * interested in (i.e. which packets to capture). Since this source example
 * focuses on IP and TCP, we use the expression "ip", so we know we'll only
 * encounter IP packets. The capture filter syntax, along with some
 * examples, is documented in the tcpdump man page under "expression."
 * Below are a few simple examples:
 *
 * Expression           Description
 * ----------           -----------
 * ip                   Capture all IP packets.
 * tcp                  Capture only TCP packets.
 * tcp port 80          Capture only TCP packets with a port equal to 80.
 * ip host 10.1.2.3     Capture all IP packets to or from host 10.1.2.3.
 *
 ****************************************************************************
 *
 */

#define APP_NAME        "sniffex"
#define APP_DESC        "Sniffer example using libpcap"
#define APP_COPYRIGHT   "Copyright (c) 2005 The Tcpdump Group"
#define APP_DISCLAIMER  "THERE IS ABSOLUTELY NO WARRANTY FOR THIS PROGRAM."

#include <pcap.h>
#include <stdio.h>
#include <string.h>
#include <stdlib.h>
#include <ctype.h>
#include <errno.h>
#include <sys/types.h>
#include <sys/socket.h>
#include <netinet/in.h>
#include <arpa/inet.h>

pcap_t *handle;             /* packet capture handle */

/* default snap length (maximum bytes per packet to capture) */
#define SNAP_LEN 1518

/* ethernet headers are always exactly 14 bytes [1] */
#define SIZE_ETHERNET 14

/* Ethernet addresses are 6 bytes */
#define ETHER_ADDR_LEN  6

/* Ethernet header */
struct sniff_ethernet {
        u_char  ether_dhost[ETHER_ADDR_LEN];    /* destination host address */
        u_char  ether_shost[ETHER_ADDR_LEN];    /* source host address */
        u_short ether_type;                     /* IP? ARP? RARP? etc */
};

/* IP header */
struct sniff_ip {
        u_char  ip_vhl;                 /* version << 4 | header length >> 2 */
        u_char  ip_tos;                 /* type of service */
        u_short ip_len;                 /* total length */
        u_short ip_id;                  /* identification */
        u_short ip_off;                 /* fragment offset field */
        #define IP_RF 0x8000            /* reserved fragment flag */
        #define IP_DF 0x4000            /* dont fragment flag */
        #define IP_MF 0x2000            /* more fragments flag */
        #define IP_OFFMASK 0x1fff       /* mask for fragmenting bits */
        u_char  ip_ttl;                 /* time to live */
        u_char  ip_p;                   /* protocol */
        u_short ip_sum;                 /* checksum */
        struct  in_addr ip_src,ip_dst;  /* source and dest address */
};
#define IP_HL(ip)               (((ip)->ip_vhl) & 0x0f)
#define IP_V(ip)                (((ip)->ip_vhl) >> 4)

/* TCP header */
typedef u_int tcp_seq;

struct sniff_tcp {
        u_short th_sport;               /* source port */
        u_short th_dport;               /* destination port */
        tcp_seq th_seq;                 /* sequence number */
        tcp_seq th_ack;                 /* acknowledgement number */
        u_char  th_offx2;               /* data offset, rsvd */
#define TH_OFF(th)      (((th)->th_offx2 & 0xf0) >> 4)
        u_char  th_flags;
        #define TH_FIN  0x01
        #define TH_SYN  0x02
        #define TH_RST  0x04
        #define TH_PUSH 0x08
        #define TH_ACK  0x10
        #define TH_URG  0x20
        #define TH_ECE  0x40
        #define TH_CWR  0x80
        #define TH_FLAGS        (TH_FIN|TH_SYN|TH_RST|TH_ACK|TH_URG|TH_ECE|TH_CWR)
        u_short th_win;                 /* window */
        u_short th_sum;                 /* checksum */
        u_short th_urp;                 /* urgent pointer */
};

void
got_packet(u_char *args, const struct pcap_pkthdr *header, const u_char *packet);

void
print_payload(const u_char *payload, int len);

void
print_hex_ascii_line(const u_char *payload, int len, int offset);

void
print_app_banner(void);

void
print_app_usage(void);

/*
 * app name/banner
 */
void
print_app_banner(void)
{

    printf("%s - %s\n", APP_NAME, APP_DESC);
    printf("%s\n", APP_COPYRIGHT);
    printf("%s\n", APP_DISCLAIMER);
    printf("\n");

return;
}

/*
 * print help text
 */
void
print_app_usage(void)
{

    printf("Usage: %s [interface]\n", APP_NAME);
    printf("\n");
    printf("Options:\n");
    printf("    interface    Listen on <interface> for packets.\n");
    printf("\n");

return;
}

/*
 * print data in rows of 16 bytes: offset   hex   ascii
 *
 * 00000   47 45 54 20 2f 20 48 54  54 50 2f 31 2e 31 0d 0a   GET / HTTP/1.1..
 */
void
print_hex_ascii_line(const u_char *payload, int len, int offset)
{

    int i;
    int gap;
    const u_char *ch;

    /* offset */
    printf("%05d   ", offset);

    /* hex */
    ch = payload;
    for(i = 0; i < len; i++) {
        printf("%02x ", *ch);
        ch++;
        /* print extra space after 8th byte for visual aid */
        if (i == 7)
            printf(" ");
    }
    /* print space to handle line less than 8 bytes */
    if (len < 8)
        printf(" ");

    /* fill hex gap with spaces if not full line */
    if (len < 16) {
        gap = 16 - len;
        for (i = 0; i < gap; i++) {
            printf("   ");
        }
    }
    printf("   ");

    /* ascii (if printable) */
    ch = payload;
    for(i = 0; i < len; i++) {
        if (isprint(*ch))
            printf("%c", *ch);
        else
            printf(".");
        ch++;
    }

    printf("\n");

return;
}

/*
 * print packet payload data (avoid printing binary data)
 */
void
print_payload(const u_char *payload, int len)
{

    int len_rem = len;
    int line_width = 16;            /* number of bytes per line */
    int line_len;
    int offset = 0;                 /* zero-based offset counter */
    const u_char *ch = payload;

    if (len <= 0)
        return;

    /* data fits on one line */
    if (len <= line_width) {
        print_hex_ascii_line(ch, len, offset);
        return;
    }

    /* data spans multiple lines */
    for ( ;; ) {
        /* compute current line length */
        line_len = line_width % len_rem;
        /* print line */
        print_hex_ascii_line(ch, line_len, offset);
        /* compute total remaining */
        len_rem = len_rem - line_len;
        /* shift pointer to remaining bytes to print */
        ch = ch + line_len;
        /* add offset */
        offset = offset + line_width;
        /* check if we have line width chars or less */
        if (len_rem <= line_width) {
            /* print last line and get out */
            print_hex_ascii_line(ch, len_rem, offset);
            break;
        }
    }

return;
}

/*
 * dissect/print packet
 */
void
got_packet(u_char *args, const struct pcap_pkthdr *header, const u_char *packet)
{

    pcap_breakloop(handle);

    static int count = 1;                   /* packet counter */

    /* declare pointers to packet headers */
    const struct sniff_ethernet *ethernet;  /* The ethernet header [1] */
    const struct sniff_ip *ip;              /* The IP header */
    const struct sniff_tcp *tcp;            /* The TCP header */
    const char *payload;                    /* Packet payload */

    int size_ip;
    int size_tcp;
    int size_payload;

    printf("\nPacket number %d:\n", count);
    count++;

    /* define ethernet header */
    ethernet = (struct sniff_ethernet*)(packet);

    /* define/compute ip header offset */
    ip = (struct sniff_ip*)(packet + SIZE_ETHERNET);
    size_ip = IP_HL(ip)*4;
    if (size_ip < 20) {
        printf("   * Invalid IP header length: %u bytes\n", size_ip);
        return;
    

Ian*_*anC 4

看来这确实是一个错误libpcap。根据他们的 GitHub 页面上的错误报告pcap_compile导致了一些内存泄漏,与我在代码中遇到的情况非常相似。

==1505== still reachable: 8 bytes in 1 blocks
Run Code Online (Sandbox Code Playgroud)

1 个块中正好有 8 个字节,就像我的valgrind报告中那样。有评论称最新的master分支版本已经修复了该bug。

目前,Ubuntu 的存储库有libpcap 1.7.4-2,而 GitHub 上的最新版本libpcap 1.8.1显然是。我使用虚拟机来测试代码,因为我需要valgrind以 root 身份运行才能使libpcap实时捕获工作。我清除了当前安装的libpcap库,克隆了最新版本,从源代码构建了它,并将其符号链接为旧版本,/usr/lib以便在具有最新版本的虚拟机上再次测试它。

运行相同的代码(多线程示例和修改后的 sniffex 代码)会产生绝对为零的报告

它显然与它本身有关,flex而不是pcap它本身,但如果有人遇到同样神秘的内存泄漏,并且因它而烦恼,请下载最新pcap版本,它应该被修复。

感谢@Kamiccolo 指出它可能是一个错误!:)