bbs/deb-mbse
Archived
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Fixed mbtask ping problem

Browse Source
This commit is contained in:
Michiel Broek 2002-05-04 16:05:48 +00:00
parent e8b82ff59c
commit f8e9d28fda
4 changed files with 146 additions and 107 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

4
ChangeLog
View File

@ -4790,6 +4790,10 @@ v0.33.20 10-Feb-2002
hopefully prevent that mbcico will use too much bandwidth on hopefully prevent that mbcico will use too much bandwidth on
TCP/IP trafic. TCP/IP trafic.
Splitted no TCP settings in no IBN, no IFC and no ITN. Splitted no TCP settings in no IBN, no IFC and no ITN.
Improved the ping tests, better errorlogging and suppresion of
icmp replies that are not for mbtask. Better protected against
flood pings. This should fixes the problem that the internet
seemed down while it was available.
mbcico: mbcico:
Fixed binkp driver to accept incoming unprotected sessions. Fixed binkp driver to accept incoming unprotected sessions.

9
mbtask/mbtask.c
View File

@ -1248,14 +1248,7 @@ int main(int argc, char **argv)
signal(i, SIG_IGN); signal(i, SIG_IGN);
} }
/* init_pingsocket();
* Create the ping socket while we are still root.
*/
if ((ping_isocket = socket(PF_INET, SOCK_RAW, IPPROTO_ICMP)) == -1) {
perror("");
printf("socket init failed, is mbtask not installed setuid root?\n");
exit(1);
}
/* /*
* mbtask is setuid root, drop privileges to user mbse. * mbtask is setuid root, drop privileges to user mbse.

236
mbtask/ping.c
View File

@ -150,134 +150,147 @@ struct sockaddr_in to;
*/ */
int ping_send(struct in_addr addr) int ping_send(struct in_addr addr)
{ {
int len; int len, sentlen;
int isock;
#ifdef __linux__ #ifdef __linux__
struct icmp_filter f; struct icmp_filter f;
#else #else
struct protoent *pe; struct protoent *pe;
int SOL_IP; int SOL_IP;
#endif #endif
unsigned long sum; unsigned long sum;
unsigned short *ptr; unsigned short *ptr;
#ifndef __linux__ #ifndef __linux__
if (!(pe = getprotobyname("ip"))) { if (!(pe = getprotobyname("ip"))) {
tasklog('?', "icmp ping: getprotobyname() failed: %s", strerror(errno)); tasklog('?', "icmp ping: getprotobyname() failed: %s", strerror(errno));
return -1; return -1;
} }
SOL_IP = pe->p_proto; SOL_IP = pe->p_proto;
#endif #endif
isock = ping_isocket; p_sequence++;
p_sequence = 1; id = (unsigned short)get_rand16(); /* randomize a ping id */
id = (unsigned short)get_rand16(); /* randomize a ping id */
#ifdef __linux__ #ifdef __linux__
/* Fancy ICMP filering -- only on Linux (as far is I know) */ /* Fancy ICMP filering -- only on Linux (as far is I know) */
/* In fact, there should be macros for treating icmp_filter, but I haven't found them in Linux 2.2.15. /* In fact, there should be macros for treating icmp_filter, but I haven't found them in Linux 2.2.15.
* So, set it manually and unportable ;-) */ * So, set it manually and unportable ;-) */
/* This filter lets ECHO_REPLY (0), DEST_UNREACH(3) and TIME_EXCEEDED(11) pass. */ /* This filter lets ECHO_REPLY (0), DEST_UNREACH(3) and TIME_EXCEEDED(11) pass. */
/* !(0000 1000 0000 1001) = 0xff ff f7 f6 */ /* !(0000 1000 0000 1001) = 0xff ff f7 f6 */
f.data=0xfffff7f6; f.data=0xfffff7f6;
if (setsockopt(isock, SOL_RAW, ICMP_FILTER, &f, sizeof(f)) == -1) { if (setsockopt(ping_isocket, SOL_RAW, ICMP_FILTER, &f, sizeof(f)) == -1) {
if (icmp_errs < ICMP_MAX_ERRS) if (icmp_errs < ICMP_MAX_ERRS)
tasklog('?', "$icmp ping: setsockopt() failed %d", isock); tasklog('?', "$icmp ping: setsockopt() failed %d", ping_isocket);
return -1; return -1;
} }
#endif #endif
icmpd.icmp_type = ICMP_ECHO; icmpd.icmp_type = ICMP_ECHO;
icmpd.icmp_code = 0; icmpd.icmp_code = 0;
icmpd.icmp_cksum = 0; icmpd.icmp_cksum = 0;
icmpd.icmp_id = htons((short)id); icmpd.icmp_id = htons((short)id);
icmpd.icmp_seq = htons(p_sequence); icmpd.icmp_seq = htons(p_sequence);
/* Checksumming - Algorithm taken from nmap. Thanks... */ /* Checksumming - Algorithm taken from nmap. Thanks... */
ptr = (unsigned short *)&icmpd; ptr = (unsigned short *)&icmpd;
sum = 0; sum = 0;
for (len = 0; len < 4; len++) { for (len = 0; len < 4; len++) {
sum += *ptr++; sum += *ptr++;
} }
sum = (sum >> 16) + (sum & 0xffff); sum = (sum >> 16) + (sum & 0xffff);
sum += (sum >> 16); sum += (sum >> 16);
icmpd.icmp_cksum = ~sum; icmpd.icmp_cksum = ~sum;
memset(&to, 0, sizeof(to)); memset(&to, 0, sizeof(to));
to.sin_family = AF_INET; to.sin_family = AF_INET;
to.sin_port = 0; to.sin_port = 0;
to.sin_addr = addr; to.sin_addr = addr;
SET_SOCKA_LEN4(to); SET_SOCKA_LEN4(to);
if (sendto(isock, &icmpd, ICMP4_ECHO_LEN, 0, (struct sockaddr *)&to, sizeof(to)) == -1) {
return -2; sentlen = sendto(ping_isocket, &icmpd, ICMP4_ECHO_LEN, 0, (struct sockaddr *)&to, sizeof(to));
} if (sentlen != ICMP4_ECHO_LEN) {
return 0; tasklog('+', "ping: sent %d octets, ret %d", ICMP4_ECHO_LEN, sentlen);
return -2;
}
return 0;
} }
/*
* 0 = reply received Ok.
* -1 = reply packet not for us, this is Ok.
* -2 = destination unreachable.
* -3 = poll/select error.
* -4 = time exceeded.
* -5 = wrong packetlen received.
* -6 = no data received, this is Ok.
* -7 = icmp parameter problem.
*/
int ping_receive(struct in_addr addr) int ping_receive(struct in_addr addr)
{ {
char buf[1024]; char buf[1024];
int isock; int len;
int len; struct sockaddr_in ffrom;
struct sockaddr_in ffrom; struct icmphdr icmpp;
struct icmphdr icmpp; struct iphdr iph;
struct iphdr iph; socklen_t sl;
socklen_t sl; struct pollfd pfd;
struct pollfd pfd;
isock = ping_isocket; pfd.fd = ping_isocket;
pfd.events = POLLIN;
pfd.fd = isock; /*
pfd.events = POLLIN; * 10 mSec is enough, this function is called at regular intervals.
/* */
* 10 mSec is enough, this function is called at regular intervals. if (poll(&pfd, 1, 10) < 0) {
*/ if (icmp_errs < ICMP_MAX_ERRS)
if (poll(&pfd, 1, 10) < 0) { tasklog('?', "$poll/select failed");
if (icmp_errs < ICMP_MAX_ERRS) return -3;
tasklog('?', "$poll/select failed"); }
return -3;
}
if (pfd.revents & POLLIN || pfd.revents & POLLERR || pfd.revents & POLLHUP || pfd.revents & POLLNVAL) { if (pfd.revents & POLLIN || pfd.revents & POLLERR || pfd.revents & POLLHUP || pfd.revents & POLLNVAL) {
sl = sizeof(ffrom); sl = sizeof(ffrom);
if ((len = recvfrom(isock, &buf, sizeof(buf), 0,(struct sockaddr *)&ffrom, &sl)) != -1) { if ((len = recvfrom(ping_isocket, &buf, sizeof(buf)-1, 0,(struct sockaddr *)&ffrom, &sl)) != -1) {
if (len > sizeof(struct iphdr)) { if (len > sizeof(struct iphdr)) {
memcpy(&iph, buf, sizeof(iph)); memcpy(&iph, buf, sizeof(iph));
if (len - iph.ip_hl * 4 >= ICMP_BASEHDR_LEN) { if (len - iph.ip_hl * 4 >= ICMP_BASEHDR_LEN) {
memcpy(&icmpp, ((unsigned long int *)buf)+iph.ip_hl, sizeof(icmpp)); memcpy(&icmpp, ((unsigned long int *)buf)+iph.ip_hl, sizeof(icmpp));
if (iph.ip_saddr == addr.s_addr && if (iph.ip_saddr == addr.s_addr && icmpp.icmp_type == ICMP_ECHOREPLY &&
icmpp.icmp_type == ICMP_ECHOREPLY && ntohs(icmpp.icmp_id) == id && ntohs(icmpp.icmp_seq) == p_sequence) {
ntohs(icmpp.icmp_id) == id && // tasklog('-', "ping: valid reply received, id %d, sequence %d", ntohs(icmpp.icmp_id), ntohs(icmpp.icmp_seq));
ntohs(icmpp.icmp_seq) <= p_sequence) { return 0;
return 0; } else {
} else { /* No regular echo reply. Maybe an error? */
/* No regular echo reply. Maybe an error? */ if (icmp4_errcmp((char *)&icmpd, ICMP4_ECHO_LEN, &to.sin_addr, buf, len, ICMP_DEST_UNREACH))
if (icmp4_errcmp((char *)&icmpd, ICMP4_ECHO_LEN, return -2;
&to.sin_addr, buf, len, ICMP_DEST_UNREACH) || if (icmp4_errcmp((char *)&icmpd, ICMP4_ECHO_LEN, &to.sin_addr, buf, len, ICMP_TIME_EXCEEDED))
icmp4_errcmp((char *)&icmpd, ICMP4_ECHO_LEN, return -4;
&to.sin_addr, buf, len, ICMP_TIME_EXCEEDED)) { if (icmp4_errcmp((char *)&icmpd, ICMP4_ECHO_LEN, &to.sin_addr, buf, len, ICMP_PARAMETERPROB))
return -4; return -7;
} /*
} * No fatal problem, the return code will be -1 caused by other
} * icmp trafic on the network (packets not for us).
} */
} else { return -1;
return -5; /* error */ }
} }
}
} else {
return -5; /* error */
} }
return -6; /* no answer */ }
return -6; /* no answer */
} }
void check_ping(void) void check_ping(void)
{ {
int rc; int rc;
/* /*
* If the previous pingstat is still P_SENT, then we now consider it a timeout. * If the previous pingstat is still P_SENT, then we now consider it a timeout.
@ -358,12 +371,16 @@ void state_ping(void)
switch (pingstate) { switch (pingstate) {
case P_NONE: pingresult[pingnr] = TRUE; case P_NONE: pingresult[pingnr] = TRUE;
break; break;
case P_SENT: rc = ping_receive(paddr); case P_SENT: /*
* Quickly eat all packets not for us, we only want our
* packets and empty results (packet still underway).
*/
while ((rc = ping_receive(paddr)) == -1);
if (!rc) { if (!rc) {
pingstate = P_OK; pingstate = P_OK;
pingresult[pingnr] = TRUE; pingresult[pingnr] = TRUE;
} else { } else {
if (rc != -6) // if (rc != -6)
tasklog('p', "ping recv %s id=%d rc=%d", pingaddress, id, rc); tasklog('p', "ping recv %s id=%d rc=%d", pingaddress, id, rc);
} }
break; break;
@ -371,3 +388,28 @@ void state_ping(void)
} }
/*
* Create the ping socket, called from main() durig init as root.
*/
void init_pingsocket(void)
{
if ((ping_isocket = socket(PF_INET, SOCK_RAW, IPPROTO_ICMP)) == -1) {
if (errno == EPERM) {
fprintf(stderr, "socket init failed, mbtask not installed setuid root\n");
} else {
fprintf(stderr, "socket init failed\n");
}
exit(1);
}
/*
* If someone's messing with us, bail.
* It would be nice to issue an error message, but to where?
*/
if (ping_isocket == STDIN_FILENO || ping_isocket == STDOUT_FILENO || ping_isocket == STDERR_FILENO) {
exit(255);
}
}

4
mbtask/ping.h
View File

@ -1,4 +1,4 @@
/* $Id */ /* $Id$ */
#ifndef _PING_H #ifndef _PING_H
#define _PING_H #define _PING_H
@ -17,6 +17,6 @@ typedef enum {P_INIT, P_SENT, P_FAIL, P_OK, P_ERROR, P_NONE} PINGSTATE;
void check_ping(void); void check_ping(void);
void state_ping(void); void state_ping(void);
void init_pingsocket(void);
#endif #endif