09632ab08c
(about 31 MB - 32 MB). Submitted (partially) by: Pascal Hofstee <daeron@wit401305.student.utwente.nl>
703 lines
16 KiB
C
703 lines
16 KiB
C
/*
|
|
* Copyright (c) 1983, 1993
|
|
* The Regents of the University of California. All rights reserved.
|
|
*
|
|
* 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.
|
|
* 3. All advertising materials mentioning features or use of this software
|
|
* must display the following acknowledgement:
|
|
* This product includes software developed by the University of
|
|
* California, Berkeley and its contributors.
|
|
* 4. Neither the name of the University 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 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.
|
|
*/
|
|
|
|
#ifndef lint
|
|
static const char copyright[] =
|
|
"@(#) Copyright (c) 1983, 1993\n\
|
|
The Regents of the University of California. All rights reserved.\n";
|
|
#endif /* not lint */
|
|
|
|
#ifndef lint
|
|
#if 0
|
|
static char sccsid[] = "@(#)tftpd.c 8.1 (Berkeley) 6/4/93";
|
|
#endif
|
|
static const char rcsid[] =
|
|
"$FreeBSD$";
|
|
#endif /* not lint */
|
|
|
|
/*
|
|
* Trivial file transfer protocol server.
|
|
*
|
|
* This version includes many modifications by Jim Guyton
|
|
* <guyton@rand-unix>.
|
|
*/
|
|
|
|
#include <sys/param.h>
|
|
#include <sys/ioctl.h>
|
|
#include <sys/stat.h>
|
|
#include <sys/socket.h>
|
|
#include <sys/types.h>
|
|
|
|
#include <netinet/in.h>
|
|
#include <arpa/tftp.h>
|
|
#include <arpa/inet.h>
|
|
|
|
#include <ctype.h>
|
|
#include <errno.h>
|
|
#include <fcntl.h>
|
|
#include <libutil.h>
|
|
#include <netdb.h>
|
|
#include <pwd.h>
|
|
#include <setjmp.h>
|
|
#include <signal.h>
|
|
#include <stdio.h>
|
|
#include <stdlib.h>
|
|
#include <string.h>
|
|
#include <syslog.h>
|
|
#include <unistd.h>
|
|
|
|
#include "tftpsubs.h"
|
|
|
|
#define TIMEOUT 5
|
|
|
|
int peer;
|
|
int rexmtval = TIMEOUT;
|
|
int maxtimeout = 5*TIMEOUT;
|
|
|
|
#define PKTSIZE SEGSIZE+4
|
|
char buf[PKTSIZE];
|
|
char ackbuf[PKTSIZE];
|
|
struct sockaddr_in from;
|
|
int fromlen;
|
|
|
|
void tftp __P((struct tftphdr *, int));
|
|
|
|
/*
|
|
* Null-terminated directory prefix list for absolute pathname requests and
|
|
* search list for relative pathname requests.
|
|
*
|
|
* MAXDIRS should be at least as large as the number of arguments that
|
|
* inetd allows (currently 20).
|
|
*/
|
|
#define MAXDIRS 20
|
|
static struct dirlist {
|
|
char *name;
|
|
int len;
|
|
} dirs[MAXDIRS+1];
|
|
static int suppress_naks;
|
|
static int logging;
|
|
static int ipchroot;
|
|
|
|
static char *errtomsg __P((int));
|
|
static void nak __P((int));
|
|
|
|
int
|
|
main(argc, argv)
|
|
int argc;
|
|
char *argv[];
|
|
{
|
|
register struct tftphdr *tp;
|
|
register int n;
|
|
int ch, on;
|
|
struct sockaddr_in sin;
|
|
char *chroot_dir = NULL;
|
|
struct passwd *nobody;
|
|
char *chuser = "nobody";
|
|
|
|
openlog("tftpd", LOG_PID | LOG_NDELAY, LOG_FTP);
|
|
while ((ch = getopt(argc, argv, "cClns:u:")) != -1) {
|
|
switch (ch) {
|
|
case 'c':
|
|
ipchroot = 1;
|
|
break;
|
|
case 'C':
|
|
ipchroot = 2;
|
|
break;
|
|
case 'l':
|
|
logging = 1;
|
|
break;
|
|
case 'n':
|
|
suppress_naks = 1;
|
|
break;
|
|
case 's':
|
|
chroot_dir = optarg;
|
|
break;
|
|
case 'u':
|
|
chuser = optarg;
|
|
break;
|
|
default:
|
|
syslog(LOG_WARNING, "ignoring unknown option -%c", ch);
|
|
}
|
|
}
|
|
if (optind < argc) {
|
|
struct dirlist *dirp;
|
|
|
|
/* Get list of directory prefixes. Skip relative pathnames. */
|
|
for (dirp = dirs; optind < argc && dirp < &dirs[MAXDIRS];
|
|
optind++) {
|
|
if (argv[optind][0] == '/') {
|
|
dirp->name = argv[optind];
|
|
dirp->len = strlen(dirp->name);
|
|
dirp++;
|
|
}
|
|
}
|
|
}
|
|
else if (chroot_dir) {
|
|
dirs->name = "/";
|
|
dirs->len = 1;
|
|
}
|
|
if (ipchroot && chroot_dir == NULL) {
|
|
syslog(LOG_ERR, "-c requires -s");
|
|
exit(1);
|
|
}
|
|
|
|
on = 1;
|
|
if (ioctl(0, FIONBIO, &on) < 0) {
|
|
syslog(LOG_ERR, "ioctl(FIONBIO): %m");
|
|
exit(1);
|
|
}
|
|
fromlen = sizeof (from);
|
|
n = recvfrom(0, buf, sizeof (buf), 0,
|
|
(struct sockaddr *)&from, &fromlen);
|
|
if (n < 0) {
|
|
syslog(LOG_ERR, "recvfrom: %m");
|
|
exit(1);
|
|
}
|
|
/*
|
|
* Now that we have read the message out of the UDP
|
|
* socket, we fork and exit. Thus, inetd will go back
|
|
* to listening to the tftp port, and the next request
|
|
* to come in will start up a new instance of tftpd.
|
|
*
|
|
* We do this so that inetd can run tftpd in "wait" mode.
|
|
* The problem with tftpd running in "nowait" mode is that
|
|
* inetd may get one or more successful "selects" on the
|
|
* tftp port before we do our receive, so more than one
|
|
* instance of tftpd may be started up. Worse, if tftpd
|
|
* break before doing the above "recvfrom", inetd would
|
|
* spawn endless instances, clogging the system.
|
|
*/
|
|
{
|
|
int pid;
|
|
int i, j;
|
|
|
|
for (i = 1; i < 20; i++) {
|
|
pid = fork();
|
|
if (pid < 0) {
|
|
sleep(i);
|
|
/*
|
|
* flush out to most recently sent request.
|
|
*
|
|
* This may drop some request, but those
|
|
* will be resent by the clients when
|
|
* they timeout. The positive effect of
|
|
* this flush is to (try to) prevent more
|
|
* than one tftpd being started up to service
|
|
* a single request from a single client.
|
|
*/
|
|
j = sizeof from;
|
|
i = recvfrom(0, buf, sizeof (buf), 0,
|
|
(struct sockaddr *)&from, &j);
|
|
if (i > 0) {
|
|
n = i;
|
|
fromlen = j;
|
|
}
|
|
} else {
|
|
break;
|
|
}
|
|
}
|
|
if (pid < 0) {
|
|
syslog(LOG_ERR, "fork: %m");
|
|
exit(1);
|
|
} else if (pid != 0) {
|
|
exit(0);
|
|
}
|
|
}
|
|
|
|
/*
|
|
* Since we exit here, we should do that only after the above
|
|
* recvfrom to keep inetd from constantly forking should there
|
|
* be a problem. See the above comment about system clogging.
|
|
*/
|
|
if (chroot_dir) {
|
|
if (ipchroot) {
|
|
char *tempchroot;
|
|
struct stat sb;
|
|
int statret;
|
|
|
|
tempchroot = inet_ntoa(from.sin_addr);
|
|
asprintf(&tempchroot, "%s/%s", chroot_dir, tempchroot);
|
|
statret = stat(tempchroot, &sb);
|
|
if ((sb.st_mode & S_IFDIR) &&
|
|
(statret == 0 || (statret == -1 && ipchroot == 1)))
|
|
chroot_dir = tempchroot;
|
|
}
|
|
/* Must get this before chroot because /etc might go away */
|
|
if ((nobody = getpwnam(chuser)) == NULL) {
|
|
syslog(LOG_ERR, "%s: no such user", chuser);
|
|
exit(1);
|
|
}
|
|
if (chroot(chroot_dir)) {
|
|
syslog(LOG_ERR, "chroot: %s: %m", chroot_dir);
|
|
exit(1);
|
|
}
|
|
chdir( "/" );
|
|
setuid(nobody->pw_uid);
|
|
}
|
|
|
|
from.sin_family = AF_INET;
|
|
alarm(0);
|
|
close(0);
|
|
close(1);
|
|
peer = socket(AF_INET, SOCK_DGRAM, 0);
|
|
if (peer < 0) {
|
|
syslog(LOG_ERR, "socket: %m");
|
|
exit(1);
|
|
}
|
|
memset(&sin, 0, sizeof(sin));
|
|
sin.sin_family = AF_INET;
|
|
if (bind(peer, (struct sockaddr *)&sin, sizeof (sin)) < 0) {
|
|
syslog(LOG_ERR, "bind: %m");
|
|
exit(1);
|
|
}
|
|
if (connect(peer, (struct sockaddr *)&from, sizeof(from)) < 0) {
|
|
syslog(LOG_ERR, "connect: %m");
|
|
exit(1);
|
|
}
|
|
tp = (struct tftphdr *)buf;
|
|
tp->th_opcode = ntohs(tp->th_opcode);
|
|
if (tp->th_opcode == RRQ || tp->th_opcode == WRQ)
|
|
tftp(tp, n);
|
|
exit(1);
|
|
}
|
|
|
|
struct formats;
|
|
int validate_access __P((char **, int));
|
|
void xmitfile __P((struct formats *));
|
|
void recvfile __P((struct formats *));
|
|
|
|
struct formats {
|
|
char *f_mode;
|
|
int (*f_validate) __P((char **, int));
|
|
void (*f_send) __P((struct formats *));
|
|
void (*f_recv) __P((struct formats *));
|
|
int f_convert;
|
|
} formats[] = {
|
|
{ "netascii", validate_access, xmitfile, recvfile, 1 },
|
|
{ "octet", validate_access, xmitfile, recvfile, 0 },
|
|
#ifdef notdef
|
|
{ "mail", validate_user, sendmail, recvmail, 1 },
|
|
#endif
|
|
{ 0 }
|
|
};
|
|
|
|
/*
|
|
* Handle initial connection protocol.
|
|
*/
|
|
void
|
|
tftp(tp, size)
|
|
struct tftphdr *tp;
|
|
int size;
|
|
{
|
|
register char *cp;
|
|
int first = 1, ecode;
|
|
register struct formats *pf;
|
|
char *filename, *mode;
|
|
|
|
filename = cp = tp->th_stuff;
|
|
again:
|
|
while (cp < buf + size) {
|
|
if (*cp == '\0')
|
|
break;
|
|
cp++;
|
|
}
|
|
if (*cp != '\0') {
|
|
nak(EBADOP);
|
|
exit(1);
|
|
}
|
|
if (first) {
|
|
mode = ++cp;
|
|
first = 0;
|
|
goto again;
|
|
}
|
|
for (cp = mode; *cp; cp++)
|
|
if (isupper(*cp))
|
|
*cp = tolower(*cp);
|
|
for (pf = formats; pf->f_mode; pf++)
|
|
if (strcmp(pf->f_mode, mode) == 0)
|
|
break;
|
|
if (pf->f_mode == 0) {
|
|
nak(EBADOP);
|
|
exit(1);
|
|
}
|
|
ecode = (*pf->f_validate)(&filename, tp->th_opcode);
|
|
if (logging) {
|
|
char host[MAXHOSTNAMELEN];
|
|
|
|
realhostname(host, sizeof(host) - 1, &from.sin_addr);
|
|
host[sizeof(host) - 1] = '\0';
|
|
syslog(LOG_INFO, "%s: %s request for %s: %s", host,
|
|
tp->th_opcode == WRQ ? "write" : "read",
|
|
filename, errtomsg(ecode));
|
|
}
|
|
if (ecode) {
|
|
/*
|
|
* Avoid storms of naks to a RRQ broadcast for a relative
|
|
* bootfile pathname from a diskless Sun.
|
|
*/
|
|
if (suppress_naks && *filename != '/' && ecode == ENOTFOUND)
|
|
exit(0);
|
|
nak(ecode);
|
|
exit(1);
|
|
}
|
|
if (tp->th_opcode == WRQ)
|
|
(*pf->f_recv)(pf);
|
|
else
|
|
(*pf->f_send)(pf);
|
|
exit(0);
|
|
}
|
|
|
|
|
|
FILE *file;
|
|
|
|
/*
|
|
* Validate file access. Since we
|
|
* have no uid or gid, for now require
|
|
* file to exist and be publicly
|
|
* readable/writable.
|
|
* If we were invoked with arguments
|
|
* from inetd then the file must also be
|
|
* in one of the given directory prefixes.
|
|
* Note also, full path name must be
|
|
* given as we have no login directory.
|
|
*/
|
|
int
|
|
validate_access(filep, mode)
|
|
char **filep;
|
|
int mode;
|
|
{
|
|
struct stat stbuf;
|
|
int fd;
|
|
struct dirlist *dirp;
|
|
static char pathname[MAXPATHLEN];
|
|
char *filename = *filep;
|
|
|
|
/*
|
|
* Prevent tricksters from getting around the directory restrictions
|
|
*/
|
|
if (strstr(filename, "/../"))
|
|
return (EACCESS);
|
|
|
|
if (*filename == '/') {
|
|
/*
|
|
* Allow the request if it's in one of the approved locations.
|
|
* Special case: check the null prefix ("/") by looking
|
|
* for length = 1 and relying on the arg. processing that
|
|
* it's a /.
|
|
*/
|
|
for (dirp = dirs; dirp->name != NULL; dirp++) {
|
|
if (dirp->len == 1 ||
|
|
(!strncmp(filename, dirp->name, dirp->len) &&
|
|
filename[dirp->len] == '/'))
|
|
break;
|
|
}
|
|
/* If directory list is empty, allow access to any file */
|
|
if (dirp->name == NULL && dirp != dirs)
|
|
return (EACCESS);
|
|
if (stat(filename, &stbuf) < 0)
|
|
return (errno == ENOENT ? ENOTFOUND : EACCESS);
|
|
if ((stbuf.st_mode & S_IFMT) != S_IFREG)
|
|
return (ENOTFOUND);
|
|
if (mode == RRQ) {
|
|
if ((stbuf.st_mode & S_IROTH) == 0)
|
|
return (EACCESS);
|
|
} else {
|
|
if ((stbuf.st_mode & S_IWOTH) == 0)
|
|
return (EACCESS);
|
|
}
|
|
} else {
|
|
int err;
|
|
|
|
/*
|
|
* Relative file name: search the approved locations for it.
|
|
* Don't allow write requests that avoid directory
|
|
* restrictions.
|
|
*/
|
|
|
|
if (!strncmp(filename, "../", 3))
|
|
return (EACCESS);
|
|
|
|
/*
|
|
* If the file exists in one of the directories and isn't
|
|
* readable, continue looking. However, change the error code
|
|
* to give an indication that the file exists.
|
|
*/
|
|
err = ENOTFOUND;
|
|
for (dirp = dirs; dirp->name != NULL; dirp++) {
|
|
snprintf(pathname, sizeof(pathname), "%s/%s",
|
|
dirp->name, filename);
|
|
if (stat(pathname, &stbuf) == 0 &&
|
|
(stbuf.st_mode & S_IFMT) == S_IFREG) {
|
|
if ((stbuf.st_mode & S_IROTH) != 0) {
|
|
break;
|
|
}
|
|
err = EACCESS;
|
|
}
|
|
}
|
|
if (dirp->name == NULL)
|
|
return (err);
|
|
*filep = filename = pathname;
|
|
}
|
|
fd = open(filename, mode == RRQ ? O_RDONLY : O_WRONLY|O_TRUNC);
|
|
if (fd < 0)
|
|
return (errno + 100);
|
|
file = fdopen(fd, (mode == RRQ)? "r":"w");
|
|
if (file == NULL) {
|
|
return errno+100;
|
|
}
|
|
return (0);
|
|
}
|
|
|
|
int timeout;
|
|
jmp_buf timeoutbuf;
|
|
|
|
void
|
|
timer()
|
|
{
|
|
|
|
timeout += rexmtval;
|
|
if (timeout >= maxtimeout)
|
|
exit(1);
|
|
longjmp(timeoutbuf, 1);
|
|
}
|
|
|
|
/*
|
|
* Send the requested file.
|
|
*/
|
|
void
|
|
xmitfile(pf)
|
|
struct formats *pf;
|
|
{
|
|
struct tftphdr *dp, *r_init();
|
|
register struct tftphdr *ap; /* ack packet */
|
|
register int size, n;
|
|
volatile unsigned short block;
|
|
|
|
signal(SIGALRM, timer);
|
|
dp = r_init();
|
|
ap = (struct tftphdr *)ackbuf;
|
|
block = 1;
|
|
do {
|
|
size = readit(file, &dp, pf->f_convert);
|
|
if (size < 0) {
|
|
nak(errno + 100);
|
|
goto abort;
|
|
}
|
|
dp->th_opcode = htons((u_short)DATA);
|
|
dp->th_block = htons((u_short)block);
|
|
timeout = 0;
|
|
(void)setjmp(timeoutbuf);
|
|
|
|
send_data:
|
|
if (send(peer, dp, size + 4, 0) != size + 4) {
|
|
syslog(LOG_ERR, "write: %m");
|
|
goto abort;
|
|
}
|
|
read_ahead(file, pf->f_convert);
|
|
for ( ; ; ) {
|
|
alarm(rexmtval); /* read the ack */
|
|
n = recv(peer, ackbuf, sizeof (ackbuf), 0);
|
|
alarm(0);
|
|
if (n < 0) {
|
|
syslog(LOG_ERR, "read: %m");
|
|
goto abort;
|
|
}
|
|
ap->th_opcode = ntohs((u_short)ap->th_opcode);
|
|
ap->th_block = ntohs((u_short)ap->th_block);
|
|
|
|
if (ap->th_opcode == ERROR)
|
|
goto abort;
|
|
|
|
if (ap->th_opcode == ACK) {
|
|
if (ap->th_block == block)
|
|
break;
|
|
/* Re-synchronize with the other side */
|
|
(void) synchnet(peer);
|
|
if (ap->th_block == (block -1))
|
|
goto send_data;
|
|
}
|
|
|
|
}
|
|
block++;
|
|
} while (size == SEGSIZE);
|
|
abort:
|
|
(void) fclose(file);
|
|
}
|
|
|
|
void
|
|
justquit()
|
|
{
|
|
exit(0);
|
|
}
|
|
|
|
|
|
/*
|
|
* Receive a file.
|
|
*/
|
|
void
|
|
recvfile(pf)
|
|
struct formats *pf;
|
|
{
|
|
struct tftphdr *dp, *w_init();
|
|
register struct tftphdr *ap; /* ack buffer */
|
|
register int n, size;
|
|
volatile unsigned short block;
|
|
|
|
signal(SIGALRM, timer);
|
|
dp = w_init();
|
|
ap = (struct tftphdr *)ackbuf;
|
|
block = 0;
|
|
do {
|
|
timeout = 0;
|
|
ap->th_opcode = htons((u_short)ACK);
|
|
ap->th_block = htons((u_short)block);
|
|
block++;
|
|
(void) setjmp(timeoutbuf);
|
|
send_ack:
|
|
if (send(peer, ackbuf, 4, 0) != 4) {
|
|
syslog(LOG_ERR, "write: %m");
|
|
goto abort;
|
|
}
|
|
write_behind(file, pf->f_convert);
|
|
for ( ; ; ) {
|
|
alarm(rexmtval);
|
|
n = recv(peer, dp, PKTSIZE, 0);
|
|
alarm(0);
|
|
if (n < 0) { /* really? */
|
|
syslog(LOG_ERR, "read: %m");
|
|
goto abort;
|
|
}
|
|
dp->th_opcode = ntohs((u_short)dp->th_opcode);
|
|
dp->th_block = ntohs((u_short)dp->th_block);
|
|
if (dp->th_opcode == ERROR)
|
|
goto abort;
|
|
if (dp->th_opcode == DATA) {
|
|
if (dp->th_block == block) {
|
|
break; /* normal */
|
|
}
|
|
/* Re-synchronize with the other side */
|
|
(void) synchnet(peer);
|
|
if (dp->th_block == (block-1))
|
|
goto send_ack; /* rexmit */
|
|
}
|
|
}
|
|
/* size = write(file, dp->th_data, n - 4); */
|
|
size = writeit(file, &dp, n - 4, pf->f_convert);
|
|
if (size != (n-4)) { /* ahem */
|
|
if (size < 0) nak(errno + 100);
|
|
else nak(ENOSPACE);
|
|
goto abort;
|
|
}
|
|
} while (size == SEGSIZE);
|
|
write_behind(file, pf->f_convert);
|
|
(void) fclose(file); /* close data file */
|
|
|
|
ap->th_opcode = htons((u_short)ACK); /* send the "final" ack */
|
|
ap->th_block = htons((u_short)(block));
|
|
(void) send(peer, ackbuf, 4, 0);
|
|
|
|
signal(SIGALRM, justquit); /* just quit on timeout */
|
|
alarm(rexmtval);
|
|
n = recv(peer, buf, sizeof (buf), 0); /* normally times out and quits */
|
|
alarm(0);
|
|
if (n >= 4 && /* if read some data */
|
|
dp->th_opcode == DATA && /* and got a data block */
|
|
block == dp->th_block) { /* then my last ack was lost */
|
|
(void) send(peer, ackbuf, 4, 0); /* resend final ack */
|
|
}
|
|
abort:
|
|
return;
|
|
}
|
|
|
|
struct errmsg {
|
|
int e_code;
|
|
char *e_msg;
|
|
} errmsgs[] = {
|
|
{ EUNDEF, "Undefined error code" },
|
|
{ ENOTFOUND, "File not found" },
|
|
{ EACCESS, "Access violation" },
|
|
{ ENOSPACE, "Disk full or allocation exceeded" },
|
|
{ EBADOP, "Illegal TFTP operation" },
|
|
{ EBADID, "Unknown transfer ID" },
|
|
{ EEXISTS, "File already exists" },
|
|
{ ENOUSER, "No such user" },
|
|
{ -1, 0 }
|
|
};
|
|
|
|
static char *
|
|
errtomsg(error)
|
|
int error;
|
|
{
|
|
static char buf[20];
|
|
register struct errmsg *pe;
|
|
if (error == 0)
|
|
return "success";
|
|
for (pe = errmsgs; pe->e_code >= 0; pe++)
|
|
if (pe->e_code == error)
|
|
return pe->e_msg;
|
|
snprintf(buf, sizeof(buf), "error %d", error);
|
|
return buf;
|
|
}
|
|
|
|
/*
|
|
* Send a nak packet (error message).
|
|
* Error code passed in is one of the
|
|
* standard TFTP codes, or a UNIX errno
|
|
* offset by 100.
|
|
*/
|
|
static void
|
|
nak(error)
|
|
int error;
|
|
{
|
|
register struct tftphdr *tp;
|
|
int length;
|
|
register struct errmsg *pe;
|
|
|
|
tp = (struct tftphdr *)buf;
|
|
tp->th_opcode = htons((u_short)ERROR);
|
|
tp->th_code = htons((u_short)error);
|
|
for (pe = errmsgs; pe->e_code >= 0; pe++)
|
|
if (pe->e_code == error)
|
|
break;
|
|
if (pe->e_code < 0) {
|
|
pe->e_msg = strerror(error - 100);
|
|
tp->th_code = EUNDEF; /* set 'undef' errorcode */
|
|
}
|
|
strcpy(tp->th_msg, pe->e_msg);
|
|
length = strlen(pe->e_msg);
|
|
tp->th_msg[length] = '\0';
|
|
length += 5;
|
|
if (send(peer, buf, length, 0) != length)
|
|
syslog(LOG_ERR, "nak: %m");
|
|
}
|