freebsd-skq/libexec/tftpd/tftpd.c

843 lines
20 KiB
C
Raw Normal View History

1994-05-27 12:39:25 +00:00
/*
* 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. Neither the name of the University nor the names of its contributors
1994-05-27 12:39:25 +00:00
* 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[] =
1994-05-27 12:39:25 +00:00
"@(#) Copyright (c) 1983, 1993\n\
The Regents of the University of California. All rights reserved.\n";
#endif /* not lint */
#ifndef lint
#if 0
1994-05-27 12:39:25 +00:00
static char sccsid[] = "@(#)tftpd.c 8.1 (Berkeley) 6/4/93";
#endif
1994-05-27 12:39:25 +00:00
#endif /* not lint */
Go ahead and merge the work edwin@ on tftpd into the tree. It is a lot better than what's in the tree now. Edwin tested it at a prior employer, but can't test it today. I've found that it works a lot better with the various uboot versions that I've used in my embedded work. Here's the pkg-descr from the port that describes the changes: It all started when we got some new routers, which told me the following when trying to upload configuration or download images from it: The TFTP server doesn't support the blocksize option. My curiousity was triggered, it took me some reading of RFCs and other documentation to find out what was possible and what could be done. Was plain TFTP very simple in its handshake, TFTP with options was kind of messy because of its backwards capability: The first packet returned could either be an acknowledgement of options, or the first data packet. Going through the source code of src/libexec/tftpd and going through the code of src/usr.bin/tftp showed that there was a lot of duplicate code, and the addition of options would only increase the amount of duplicate code. After all, both the client and the server can act as a sender and receiver. At the end, it ended up with a nearly complete rewrite of the tftp client and server. It has been tested against the following TFTP clients and servers: - Itself (yay!) - The standard FreeBSD tftp client and server - The Fedora Core 6 tftp client and server - Cisco router tftp client - Extreme Networks tftp client It supports the following RFCs: RFC1350 - THE TFTP PROTOCOL (REVISION 2) RFC2347 - TFTP Option Extension RFC2348 - TFTP Blocksize Option RFC2349 - TFTP Timeout Interval and Transfer Size Options RFC3617 - Uniform Resource Identifier (URI) Scheme and Applicability Statement for the Trivial File Transfer Protocol (TFTP) It supports the following unofficial TFTP Options as described at http://www.compuphase.com/tftp.htm: blksize2 - Block size restricted to powers of 2, excluding protocol headers rollover - Block counter roll-over (roll back to zero or to one) From the tftp program point of view the following things are changed: - New commands: "blocksize", "blocksize2", "rollover" and "options" - Development features: "debug" and "packetdrop" If you try this tftp/tftpd implementation, please let me know if it works (or doesn't work) and against which implementaion so I can get a list of confirmed working systems. Author: Edwin Groothuis <edwin@FreeBSD.org>
2010-05-04 06:19:19 +00:00
#include <sys/cdefs.h>
__FBSDID("$FreeBSD$");
1994-05-27 12:39:25 +00:00
/*
* 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 <netinet/in.h>
#include <arpa/tftp.h>
#include <ctype.h>
#include <errno.h>
#include <fcntl.h>
#include <netdb.h>
#include <pwd.h>
#include <stdint.h>
1994-05-27 12:39:25 +00:00
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <syslog.h>
#include <unistd.h>
Go ahead and merge the work edwin@ on tftpd into the tree. It is a lot better than what's in the tree now. Edwin tested it at a prior employer, but can't test it today. I've found that it works a lot better with the various uboot versions that I've used in my embedded work. Here's the pkg-descr from the port that describes the changes: It all started when we got some new routers, which told me the following when trying to upload configuration or download images from it: The TFTP server doesn't support the blocksize option. My curiousity was triggered, it took me some reading of RFCs and other documentation to find out what was possible and what could be done. Was plain TFTP very simple in its handshake, TFTP with options was kind of messy because of its backwards capability: The first packet returned could either be an acknowledgement of options, or the first data packet. Going through the source code of src/libexec/tftpd and going through the code of src/usr.bin/tftp showed that there was a lot of duplicate code, and the addition of options would only increase the amount of duplicate code. After all, both the client and the server can act as a sender and receiver. At the end, it ended up with a nearly complete rewrite of the tftp client and server. It has been tested against the following TFTP clients and servers: - Itself (yay!) - The standard FreeBSD tftp client and server - The Fedora Core 6 tftp client and server - Cisco router tftp client - Extreme Networks tftp client It supports the following RFCs: RFC1350 - THE TFTP PROTOCOL (REVISION 2) RFC2347 - TFTP Option Extension RFC2348 - TFTP Blocksize Option RFC2349 - TFTP Timeout Interval and Transfer Size Options RFC3617 - Uniform Resource Identifier (URI) Scheme and Applicability Statement for the Trivial File Transfer Protocol (TFTP) It supports the following unofficial TFTP Options as described at http://www.compuphase.com/tftp.htm: blksize2 - Block size restricted to powers of 2, excluding protocol headers rollover - Block counter roll-over (roll back to zero or to one) From the tftp program point of view the following things are changed: - New commands: "blocksize", "blocksize2", "rollover" and "options" - Development features: "debug" and "packetdrop" If you try this tftp/tftpd implementation, please let me know if it works (or doesn't work) and against which implementaion so I can get a list of confirmed working systems. Author: Edwin Groothuis <edwin@FreeBSD.org>
2010-05-04 06:19:19 +00:00
#include "tftp-file.h"
#include "tftp-io.h"
#include "tftp-utils.h"
#include "tftp-transfer.h"
#include "tftp-options.h"
1994-05-27 12:39:25 +00:00
#ifdef LIBWRAP
#include <tcpd.h>
#endif
Go ahead and merge the work edwin@ on tftpd into the tree. It is a lot better than what's in the tree now. Edwin tested it at a prior employer, but can't test it today. I've found that it works a lot better with the various uboot versions that I've used in my embedded work. Here's the pkg-descr from the port that describes the changes: It all started when we got some new routers, which told me the following when trying to upload configuration or download images from it: The TFTP server doesn't support the blocksize option. My curiousity was triggered, it took me some reading of RFCs and other documentation to find out what was possible and what could be done. Was plain TFTP very simple in its handshake, TFTP with options was kind of messy because of its backwards capability: The first packet returned could either be an acknowledgement of options, or the first data packet. Going through the source code of src/libexec/tftpd and going through the code of src/usr.bin/tftp showed that there was a lot of duplicate code, and the addition of options would only increase the amount of duplicate code. After all, both the client and the server can act as a sender and receiver. At the end, it ended up with a nearly complete rewrite of the tftp client and server. It has been tested against the following TFTP clients and servers: - Itself (yay!) - The standard FreeBSD tftp client and server - The Fedora Core 6 tftp client and server - Cisco router tftp client - Extreme Networks tftp client It supports the following RFCs: RFC1350 - THE TFTP PROTOCOL (REVISION 2) RFC2347 - TFTP Option Extension RFC2348 - TFTP Blocksize Option RFC2349 - TFTP Timeout Interval and Transfer Size Options RFC3617 - Uniform Resource Identifier (URI) Scheme and Applicability Statement for the Trivial File Transfer Protocol (TFTP) It supports the following unofficial TFTP Options as described at http://www.compuphase.com/tftp.htm: blksize2 - Block size restricted to powers of 2, excluding protocol headers rollover - Block counter roll-over (roll back to zero or to one) From the tftp program point of view the following things are changed: - New commands: "blocksize", "blocksize2", "rollover" and "options" - Development features: "debug" and "packetdrop" If you try this tftp/tftpd implementation, please let me know if it works (or doesn't work) and against which implementaion so I can get a list of confirmed working systems. Author: Edwin Groothuis <edwin@FreeBSD.org>
2010-05-04 06:19:19 +00:00
static void tftp_wrq(int peer, char *, ssize_t);
static void tftp_rrq(int peer, char *, ssize_t);
1994-05-27 12:39:25 +00:00
/*
* 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 {
const char *name;
1994-05-27 12:39:25 +00:00
int len;
} dirs[MAXDIRS+1];
static int suppress_naks;
static int logging;
static int ipchroot;
static int create_new = 0;
2010-09-24 10:40:17 +00:00
static const char *newfile_format = "%Y%m%d";
static int increase_name = 0;
Go ahead and merge the work edwin@ on tftpd into the tree. It is a lot better than what's in the tree now. Edwin tested it at a prior employer, but can't test it today. I've found that it works a lot better with the various uboot versions that I've used in my embedded work. Here's the pkg-descr from the port that describes the changes: It all started when we got some new routers, which told me the following when trying to upload configuration or download images from it: The TFTP server doesn't support the blocksize option. My curiousity was triggered, it took me some reading of RFCs and other documentation to find out what was possible and what could be done. Was plain TFTP very simple in its handshake, TFTP with options was kind of messy because of its backwards capability: The first packet returned could either be an acknowledgement of options, or the first data packet. Going through the source code of src/libexec/tftpd and going through the code of src/usr.bin/tftp showed that there was a lot of duplicate code, and the addition of options would only increase the amount of duplicate code. After all, both the client and the server can act as a sender and receiver. At the end, it ended up with a nearly complete rewrite of the tftp client and server. It has been tested against the following TFTP clients and servers: - Itself (yay!) - The standard FreeBSD tftp client and server - The Fedora Core 6 tftp client and server - Cisco router tftp client - Extreme Networks tftp client It supports the following RFCs: RFC1350 - THE TFTP PROTOCOL (REVISION 2) RFC2347 - TFTP Option Extension RFC2348 - TFTP Blocksize Option RFC2349 - TFTP Timeout Interval and Transfer Size Options RFC3617 - Uniform Resource Identifier (URI) Scheme and Applicability Statement for the Trivial File Transfer Protocol (TFTP) It supports the following unofficial TFTP Options as described at http://www.compuphase.com/tftp.htm: blksize2 - Block size restricted to powers of 2, excluding protocol headers rollover - Block counter roll-over (roll back to zero or to one) From the tftp program point of view the following things are changed: - New commands: "blocksize", "blocksize2", "rollover" and "options" - Development features: "debug" and "packetdrop" If you try this tftp/tftpd implementation, please let me know if it works (or doesn't work) and against which implementaion so I can get a list of confirmed working systems. Author: Edwin Groothuis <edwin@FreeBSD.org>
2010-05-04 06:19:19 +00:00
static mode_t mask = S_IWGRP | S_IWOTH;
1994-05-27 12:39:25 +00:00
Go ahead and merge the work edwin@ on tftpd into the tree. It is a lot better than what's in the tree now. Edwin tested it at a prior employer, but can't test it today. I've found that it works a lot better with the various uboot versions that I've used in my embedded work. Here's the pkg-descr from the port that describes the changes: It all started when we got some new routers, which told me the following when trying to upload configuration or download images from it: The TFTP server doesn't support the blocksize option. My curiousity was triggered, it took me some reading of RFCs and other documentation to find out what was possible and what could be done. Was plain TFTP very simple in its handshake, TFTP with options was kind of messy because of its backwards capability: The first packet returned could either be an acknowledgement of options, or the first data packet. Going through the source code of src/libexec/tftpd and going through the code of src/usr.bin/tftp showed that there was a lot of duplicate code, and the addition of options would only increase the amount of duplicate code. After all, both the client and the server can act as a sender and receiver. At the end, it ended up with a nearly complete rewrite of the tftp client and server. It has been tested against the following TFTP clients and servers: - Itself (yay!) - The standard FreeBSD tftp client and server - The Fedora Core 6 tftp client and server - Cisco router tftp client - Extreme Networks tftp client It supports the following RFCs: RFC1350 - THE TFTP PROTOCOL (REVISION 2) RFC2347 - TFTP Option Extension RFC2348 - TFTP Blocksize Option RFC2349 - TFTP Timeout Interval and Transfer Size Options RFC3617 - Uniform Resource Identifier (URI) Scheme and Applicability Statement for the Trivial File Transfer Protocol (TFTP) It supports the following unofficial TFTP Options as described at http://www.compuphase.com/tftp.htm: blksize2 - Block size restricted to powers of 2, excluding protocol headers rollover - Block counter roll-over (roll back to zero or to one) From the tftp program point of view the following things are changed: - New commands: "blocksize", "blocksize2", "rollover" and "options" - Development features: "debug" and "packetdrop" If you try this tftp/tftpd implementation, please let me know if it works (or doesn't work) and against which implementaion so I can get a list of confirmed working systems. Author: Edwin Groothuis <edwin@FreeBSD.org>
2010-05-04 06:19:19 +00:00
struct formats;
static void tftp_recvfile(int peer, const char *mode);
static void tftp_xmitfile(int peer, const char *mode);
static int validate_access(int peer, char **, int);
static char peername[NI_MAXHOST];
static FILE *file;
static struct formats {
Go ahead and merge the work edwin@ on tftpd into the tree. It is a lot better than what's in the tree now. Edwin tested it at a prior employer, but can't test it today. I've found that it works a lot better with the various uboot versions that I've used in my embedded work. Here's the pkg-descr from the port that describes the changes: It all started when we got some new routers, which told me the following when trying to upload configuration or download images from it: The TFTP server doesn't support the blocksize option. My curiousity was triggered, it took me some reading of RFCs and other documentation to find out what was possible and what could be done. Was plain TFTP very simple in its handshake, TFTP with options was kind of messy because of its backwards capability: The first packet returned could either be an acknowledgement of options, or the first data packet. Going through the source code of src/libexec/tftpd and going through the code of src/usr.bin/tftp showed that there was a lot of duplicate code, and the addition of options would only increase the amount of duplicate code. After all, both the client and the server can act as a sender and receiver. At the end, it ended up with a nearly complete rewrite of the tftp client and server. It has been tested against the following TFTP clients and servers: - Itself (yay!) - The standard FreeBSD tftp client and server - The Fedora Core 6 tftp client and server - Cisco router tftp client - Extreme Networks tftp client It supports the following RFCs: RFC1350 - THE TFTP PROTOCOL (REVISION 2) RFC2347 - TFTP Option Extension RFC2348 - TFTP Blocksize Option RFC2349 - TFTP Timeout Interval and Transfer Size Options RFC3617 - Uniform Resource Identifier (URI) Scheme and Applicability Statement for the Trivial File Transfer Protocol (TFTP) It supports the following unofficial TFTP Options as described at http://www.compuphase.com/tftp.htm: blksize2 - Block size restricted to powers of 2, excluding protocol headers rollover - Block counter roll-over (roll back to zero or to one) From the tftp program point of view the following things are changed: - New commands: "blocksize", "blocksize2", "rollover" and "options" - Development features: "debug" and "packetdrop" If you try this tftp/tftpd implementation, please let me know if it works (or doesn't work) and against which implementaion so I can get a list of confirmed working systems. Author: Edwin Groothuis <edwin@FreeBSD.org>
2010-05-04 06:19:19 +00:00
const char *f_mode;
int f_convert;
} formats[] = {
{ "netascii", 1 },
{ "octet", 0 },
{ NULL, 0 }
};
1994-05-27 12:39:25 +00:00
int
main(int argc, char *argv[])
1994-05-27 12:39:25 +00:00
{
struct tftphdr *tp;
Go ahead and merge the work edwin@ on tftpd into the tree. It is a lot better than what's in the tree now. Edwin tested it at a prior employer, but can't test it today. I've found that it works a lot better with the various uboot versions that I've used in my embedded work. Here's the pkg-descr from the port that describes the changes: It all started when we got some new routers, which told me the following when trying to upload configuration or download images from it: The TFTP server doesn't support the blocksize option. My curiousity was triggered, it took me some reading of RFCs and other documentation to find out what was possible and what could be done. Was plain TFTP very simple in its handshake, TFTP with options was kind of messy because of its backwards capability: The first packet returned could either be an acknowledgement of options, or the first data packet. Going through the source code of src/libexec/tftpd and going through the code of src/usr.bin/tftp showed that there was a lot of duplicate code, and the addition of options would only increase the amount of duplicate code. After all, both the client and the server can act as a sender and receiver. At the end, it ended up with a nearly complete rewrite of the tftp client and server. It has been tested against the following TFTP clients and servers: - Itself (yay!) - The standard FreeBSD tftp client and server - The Fedora Core 6 tftp client and server - Cisco router tftp client - Extreme Networks tftp client It supports the following RFCs: RFC1350 - THE TFTP PROTOCOL (REVISION 2) RFC2347 - TFTP Option Extension RFC2348 - TFTP Blocksize Option RFC2349 - TFTP Timeout Interval and Transfer Size Options RFC3617 - Uniform Resource Identifier (URI) Scheme and Applicability Statement for the Trivial File Transfer Protocol (TFTP) It supports the following unofficial TFTP Options as described at http://www.compuphase.com/tftp.htm: blksize2 - Block size restricted to powers of 2, excluding protocol headers rollover - Block counter roll-over (roll back to zero or to one) From the tftp program point of view the following things are changed: - New commands: "blocksize", "blocksize2", "rollover" and "options" - Development features: "debug" and "packetdrop" If you try this tftp/tftpd implementation, please let me know if it works (or doesn't work) and against which implementaion so I can get a list of confirmed working systems. Author: Edwin Groothuis <edwin@FreeBSD.org>
2010-05-04 06:19:19 +00:00
int peer;
socklen_t peerlen, len;
ssize_t n;
int ch;
char *chroot_dir = NULL;
struct passwd *nobody;
const char *chuser = "nobody";
char recvbuffer[MAXPKTSIZE];
int allow_ro = 1, allow_wo = 1;
1994-05-27 12:39:25 +00:00
tzset(); /* syslog in localtime */
Go ahead and merge the work edwin@ on tftpd into the tree. It is a lot better than what's in the tree now. Edwin tested it at a prior employer, but can't test it today. I've found that it works a lot better with the various uboot versions that I've used in my embedded work. Here's the pkg-descr from the port that describes the changes: It all started when we got some new routers, which told me the following when trying to upload configuration or download images from it: The TFTP server doesn't support the blocksize option. My curiousity was triggered, it took me some reading of RFCs and other documentation to find out what was possible and what could be done. Was plain TFTP very simple in its handshake, TFTP with options was kind of messy because of its backwards capability: The first packet returned could either be an acknowledgement of options, or the first data packet. Going through the source code of src/libexec/tftpd and going through the code of src/usr.bin/tftp showed that there was a lot of duplicate code, and the addition of options would only increase the amount of duplicate code. After all, both the client and the server can act as a sender and receiver. At the end, it ended up with a nearly complete rewrite of the tftp client and server. It has been tested against the following TFTP clients and servers: - Itself (yay!) - The standard FreeBSD tftp client and server - The Fedora Core 6 tftp client and server - Cisco router tftp client - Extreme Networks tftp client It supports the following RFCs: RFC1350 - THE TFTP PROTOCOL (REVISION 2) RFC2347 - TFTP Option Extension RFC2348 - TFTP Blocksize Option RFC2349 - TFTP Timeout Interval and Transfer Size Options RFC3617 - Uniform Resource Identifier (URI) Scheme and Applicability Statement for the Trivial File Transfer Protocol (TFTP) It supports the following unofficial TFTP Options as described at http://www.compuphase.com/tftp.htm: blksize2 - Block size restricted to powers of 2, excluding protocol headers rollover - Block counter roll-over (roll back to zero or to one) From the tftp program point of view the following things are changed: - New commands: "blocksize", "blocksize2", "rollover" and "options" - Development features: "debug" and "packetdrop" If you try this tftp/tftpd implementation, please let me know if it works (or doesn't work) and against which implementaion so I can get a list of confirmed working systems. Author: Edwin Groothuis <edwin@FreeBSD.org>
2010-05-04 06:19:19 +00:00
acting_as_client = 0;
Go ahead and merge the work edwin@ on tftpd into the tree. It is a lot better than what's in the tree now. Edwin tested it at a prior employer, but can't test it today. I've found that it works a lot better with the various uboot versions that I've used in my embedded work. Here's the pkg-descr from the port that describes the changes: It all started when we got some new routers, which told me the following when trying to upload configuration or download images from it: The TFTP server doesn't support the blocksize option. My curiousity was triggered, it took me some reading of RFCs and other documentation to find out what was possible and what could be done. Was plain TFTP very simple in its handshake, TFTP with options was kind of messy because of its backwards capability: The first packet returned could either be an acknowledgement of options, or the first data packet. Going through the source code of src/libexec/tftpd and going through the code of src/usr.bin/tftp showed that there was a lot of duplicate code, and the addition of options would only increase the amount of duplicate code. After all, both the client and the server can act as a sender and receiver. At the end, it ended up with a nearly complete rewrite of the tftp client and server. It has been tested against the following TFTP clients and servers: - Itself (yay!) - The standard FreeBSD tftp client and server - The Fedora Core 6 tftp client and server - Cisco router tftp client - Extreme Networks tftp client It supports the following RFCs: RFC1350 - THE TFTP PROTOCOL (REVISION 2) RFC2347 - TFTP Option Extension RFC2348 - TFTP Blocksize Option RFC2349 - TFTP Timeout Interval and Transfer Size Options RFC3617 - Uniform Resource Identifier (URI) Scheme and Applicability Statement for the Trivial File Transfer Protocol (TFTP) It supports the following unofficial TFTP Options as described at http://www.compuphase.com/tftp.htm: blksize2 - Block size restricted to powers of 2, excluding protocol headers rollover - Block counter roll-over (roll back to zero or to one) From the tftp program point of view the following things are changed: - New commands: "blocksize", "blocksize2", "rollover" and "options" - Development features: "debug" and "packetdrop" If you try this tftp/tftpd implementation, please let me know if it works (or doesn't work) and against which implementaion so I can get a list of confirmed working systems. Author: Edwin Groothuis <edwin@FreeBSD.org>
2010-05-04 06:19:19 +00:00
tftp_openlog("tftpd", LOG_PID | LOG_NDELAY, LOG_FTP);
while ((ch = getopt(argc, argv, "cCd:F:lnoOp:s:u:U:wW")) != -1) {
1994-05-27 12:39:25 +00:00
switch (ch) {
case 'c':
ipchroot = 1;
break;
case 'C':
ipchroot = 2;
break;
Go ahead and merge the work edwin@ on tftpd into the tree. It is a lot better than what's in the tree now. Edwin tested it at a prior employer, but can't test it today. I've found that it works a lot better with the various uboot versions that I've used in my embedded work. Here's the pkg-descr from the port that describes the changes: It all started when we got some new routers, which told me the following when trying to upload configuration or download images from it: The TFTP server doesn't support the blocksize option. My curiousity was triggered, it took me some reading of RFCs and other documentation to find out what was possible and what could be done. Was plain TFTP very simple in its handshake, TFTP with options was kind of messy because of its backwards capability: The first packet returned could either be an acknowledgement of options, or the first data packet. Going through the source code of src/libexec/tftpd and going through the code of src/usr.bin/tftp showed that there was a lot of duplicate code, and the addition of options would only increase the amount of duplicate code. After all, both the client and the server can act as a sender and receiver. At the end, it ended up with a nearly complete rewrite of the tftp client and server. It has been tested against the following TFTP clients and servers: - Itself (yay!) - The standard FreeBSD tftp client and server - The Fedora Core 6 tftp client and server - Cisco router tftp client - Extreme Networks tftp client It supports the following RFCs: RFC1350 - THE TFTP PROTOCOL (REVISION 2) RFC2347 - TFTP Option Extension RFC2348 - TFTP Blocksize Option RFC2349 - TFTP Timeout Interval and Transfer Size Options RFC3617 - Uniform Resource Identifier (URI) Scheme and Applicability Statement for the Trivial File Transfer Protocol (TFTP) It supports the following unofficial TFTP Options as described at http://www.compuphase.com/tftp.htm: blksize2 - Block size restricted to powers of 2, excluding protocol headers rollover - Block counter roll-over (roll back to zero or to one) From the tftp program point of view the following things are changed: - New commands: "blocksize", "blocksize2", "rollover" and "options" - Development features: "debug" and "packetdrop" If you try this tftp/tftpd implementation, please let me know if it works (or doesn't work) and against which implementaion so I can get a list of confirmed working systems. Author: Edwin Groothuis <edwin@FreeBSD.org>
2010-05-04 06:19:19 +00:00
case 'd':
if (atoi(optarg) != 0)
debug += atoi(optarg);
else
debug |= debug_finds(optarg);
break;
case 'F':
newfile_format = optarg;
break;
1994-05-27 12:39:25 +00:00
case 'l':
logging = 1;
break;
case 'n':
suppress_naks = 1;
break;
Go ahead and merge the work edwin@ on tftpd into the tree. It is a lot better than what's in the tree now. Edwin tested it at a prior employer, but can't test it today. I've found that it works a lot better with the various uboot versions that I've used in my embedded work. Here's the pkg-descr from the port that describes the changes: It all started when we got some new routers, which told me the following when trying to upload configuration or download images from it: The TFTP server doesn't support the blocksize option. My curiousity was triggered, it took me some reading of RFCs and other documentation to find out what was possible and what could be done. Was plain TFTP very simple in its handshake, TFTP with options was kind of messy because of its backwards capability: The first packet returned could either be an acknowledgement of options, or the first data packet. Going through the source code of src/libexec/tftpd and going through the code of src/usr.bin/tftp showed that there was a lot of duplicate code, and the addition of options would only increase the amount of duplicate code. After all, both the client and the server can act as a sender and receiver. At the end, it ended up with a nearly complete rewrite of the tftp client and server. It has been tested against the following TFTP clients and servers: - Itself (yay!) - The standard FreeBSD tftp client and server - The Fedora Core 6 tftp client and server - Cisco router tftp client - Extreme Networks tftp client It supports the following RFCs: RFC1350 - THE TFTP PROTOCOL (REVISION 2) RFC2347 - TFTP Option Extension RFC2348 - TFTP Blocksize Option RFC2349 - TFTP Timeout Interval and Transfer Size Options RFC3617 - Uniform Resource Identifier (URI) Scheme and Applicability Statement for the Trivial File Transfer Protocol (TFTP) It supports the following unofficial TFTP Options as described at http://www.compuphase.com/tftp.htm: blksize2 - Block size restricted to powers of 2, excluding protocol headers rollover - Block counter roll-over (roll back to zero or to one) From the tftp program point of view the following things are changed: - New commands: "blocksize", "blocksize2", "rollover" and "options" - Development features: "debug" and "packetdrop" If you try this tftp/tftpd implementation, please let me know if it works (or doesn't work) and against which implementaion so I can get a list of confirmed working systems. Author: Edwin Groothuis <edwin@FreeBSD.org>
2010-05-04 06:19:19 +00:00
case 'o':
options_rfc_enabled = 0;
break;
case 'O':
options_extra_enabled = 0;
break;
case 'p':
packetdroppercentage = atoi(optarg);
tftp_log(LOG_INFO,
"Randomly dropping %d out of 100 packets",
packetdroppercentage);
break;
case 's':
chroot_dir = optarg;
break;
case 'u':
chuser = optarg;
break;
case 'U':
mask = strtol(optarg, NULL, 0);
break;
case 'w':
create_new = 1;
break;
case 'W':
create_new = 1;
increase_name = 1;
break;
1994-05-27 12:39:25 +00:00
default:
Go ahead and merge the work edwin@ on tftpd into the tree. It is a lot better than what's in the tree now. Edwin tested it at a prior employer, but can't test it today. I've found that it works a lot better with the various uboot versions that I've used in my embedded work. Here's the pkg-descr from the port that describes the changes: It all started when we got some new routers, which told me the following when trying to upload configuration or download images from it: The TFTP server doesn't support the blocksize option. My curiousity was triggered, it took me some reading of RFCs and other documentation to find out what was possible and what could be done. Was plain TFTP very simple in its handshake, TFTP with options was kind of messy because of its backwards capability: The first packet returned could either be an acknowledgement of options, or the first data packet. Going through the source code of src/libexec/tftpd and going through the code of src/usr.bin/tftp showed that there was a lot of duplicate code, and the addition of options would only increase the amount of duplicate code. After all, both the client and the server can act as a sender and receiver. At the end, it ended up with a nearly complete rewrite of the tftp client and server. It has been tested against the following TFTP clients and servers: - Itself (yay!) - The standard FreeBSD tftp client and server - The Fedora Core 6 tftp client and server - Cisco router tftp client - Extreme Networks tftp client It supports the following RFCs: RFC1350 - THE TFTP PROTOCOL (REVISION 2) RFC2347 - TFTP Option Extension RFC2348 - TFTP Blocksize Option RFC2349 - TFTP Timeout Interval and Transfer Size Options RFC3617 - Uniform Resource Identifier (URI) Scheme and Applicability Statement for the Trivial File Transfer Protocol (TFTP) It supports the following unofficial TFTP Options as described at http://www.compuphase.com/tftp.htm: blksize2 - Block size restricted to powers of 2, excluding protocol headers rollover - Block counter roll-over (roll back to zero or to one) From the tftp program point of view the following things are changed: - New commands: "blocksize", "blocksize2", "rollover" and "options" - Development features: "debug" and "packetdrop" If you try this tftp/tftpd implementation, please let me know if it works (or doesn't work) and against which implementaion so I can get a list of confirmed working systems. Author: Edwin Groothuis <edwin@FreeBSD.org>
2010-05-04 06:19:19 +00:00
tftp_log(LOG_WARNING,
"ignoring unknown option -%c", ch);
1994-05-27 12:39:25 +00:00
}
}
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 > 0 && chroot_dir == NULL) {
Go ahead and merge the work edwin@ on tftpd into the tree. It is a lot better than what's in the tree now. Edwin tested it at a prior employer, but can't test it today. I've found that it works a lot better with the various uboot versions that I've used in my embedded work. Here's the pkg-descr from the port that describes the changes: It all started when we got some new routers, which told me the following when trying to upload configuration or download images from it: The TFTP server doesn't support the blocksize option. My curiousity was triggered, it took me some reading of RFCs and other documentation to find out what was possible and what could be done. Was plain TFTP very simple in its handshake, TFTP with options was kind of messy because of its backwards capability: The first packet returned could either be an acknowledgement of options, or the first data packet. Going through the source code of src/libexec/tftpd and going through the code of src/usr.bin/tftp showed that there was a lot of duplicate code, and the addition of options would only increase the amount of duplicate code. After all, both the client and the server can act as a sender and receiver. At the end, it ended up with a nearly complete rewrite of the tftp client and server. It has been tested against the following TFTP clients and servers: - Itself (yay!) - The standard FreeBSD tftp client and server - The Fedora Core 6 tftp client and server - Cisco router tftp client - Extreme Networks tftp client It supports the following RFCs: RFC1350 - THE TFTP PROTOCOL (REVISION 2) RFC2347 - TFTP Option Extension RFC2348 - TFTP Blocksize Option RFC2349 - TFTP Timeout Interval and Transfer Size Options RFC3617 - Uniform Resource Identifier (URI) Scheme and Applicability Statement for the Trivial File Transfer Protocol (TFTP) It supports the following unofficial TFTP Options as described at http://www.compuphase.com/tftp.htm: blksize2 - Block size restricted to powers of 2, excluding protocol headers rollover - Block counter roll-over (roll back to zero or to one) From the tftp program point of view the following things are changed: - New commands: "blocksize", "blocksize2", "rollover" and "options" - Development features: "debug" and "packetdrop" If you try this tftp/tftpd implementation, please let me know if it works (or doesn't work) and against which implementaion so I can get a list of confirmed working systems. Author: Edwin Groothuis <edwin@FreeBSD.org>
2010-05-04 06:19:19 +00:00
tftp_log(LOG_ERR, "-c requires -s");
exit(1);
}
1994-05-27 12:39:25 +00:00
umask(mask);
Go ahead and merge the work edwin@ on tftpd into the tree. It is a lot better than what's in the tree now. Edwin tested it at a prior employer, but can't test it today. I've found that it works a lot better with the various uboot versions that I've used in my embedded work. Here's the pkg-descr from the port that describes the changes: It all started when we got some new routers, which told me the following when trying to upload configuration or download images from it: The TFTP server doesn't support the blocksize option. My curiousity was triggered, it took me some reading of RFCs and other documentation to find out what was possible and what could be done. Was plain TFTP very simple in its handshake, TFTP with options was kind of messy because of its backwards capability: The first packet returned could either be an acknowledgement of options, or the first data packet. Going through the source code of src/libexec/tftpd and going through the code of src/usr.bin/tftp showed that there was a lot of duplicate code, and the addition of options would only increase the amount of duplicate code. After all, both the client and the server can act as a sender and receiver. At the end, it ended up with a nearly complete rewrite of the tftp client and server. It has been tested against the following TFTP clients and servers: - Itself (yay!) - The standard FreeBSD tftp client and server - The Fedora Core 6 tftp client and server - Cisco router tftp client - Extreme Networks tftp client It supports the following RFCs: RFC1350 - THE TFTP PROTOCOL (REVISION 2) RFC2347 - TFTP Option Extension RFC2348 - TFTP Blocksize Option RFC2349 - TFTP Timeout Interval and Transfer Size Options RFC3617 - Uniform Resource Identifier (URI) Scheme and Applicability Statement for the Trivial File Transfer Protocol (TFTP) It supports the following unofficial TFTP Options as described at http://www.compuphase.com/tftp.htm: blksize2 - Block size restricted to powers of 2, excluding protocol headers rollover - Block counter roll-over (roll back to zero or to one) From the tftp program point of view the following things are changed: - New commands: "blocksize", "blocksize2", "rollover" and "options" - Development features: "debug" and "packetdrop" If you try this tftp/tftpd implementation, please let me know if it works (or doesn't work) and against which implementaion so I can get a list of confirmed working systems. Author: Edwin Groothuis <edwin@FreeBSD.org>
2010-05-04 06:19:19 +00:00
{
int on = 1;
if (ioctl(0, FIONBIO, &on) < 0) {
tftp_log(LOG_ERR, "ioctl(FIONBIO): %s", strerror(errno));
exit(1);
}
1994-05-27 12:39:25 +00:00
}
Go ahead and merge the work edwin@ on tftpd into the tree. It is a lot better than what's in the tree now. Edwin tested it at a prior employer, but can't test it today. I've found that it works a lot better with the various uboot versions that I've used in my embedded work. Here's the pkg-descr from the port that describes the changes: It all started when we got some new routers, which told me the following when trying to upload configuration or download images from it: The TFTP server doesn't support the blocksize option. My curiousity was triggered, it took me some reading of RFCs and other documentation to find out what was possible and what could be done. Was plain TFTP very simple in its handshake, TFTP with options was kind of messy because of its backwards capability: The first packet returned could either be an acknowledgement of options, or the first data packet. Going through the source code of src/libexec/tftpd and going through the code of src/usr.bin/tftp showed that there was a lot of duplicate code, and the addition of options would only increase the amount of duplicate code. After all, both the client and the server can act as a sender and receiver. At the end, it ended up with a nearly complete rewrite of the tftp client and server. It has been tested against the following TFTP clients and servers: - Itself (yay!) - The standard FreeBSD tftp client and server - The Fedora Core 6 tftp client and server - Cisco router tftp client - Extreme Networks tftp client It supports the following RFCs: RFC1350 - THE TFTP PROTOCOL (REVISION 2) RFC2347 - TFTP Option Extension RFC2348 - TFTP Blocksize Option RFC2349 - TFTP Timeout Interval and Transfer Size Options RFC3617 - Uniform Resource Identifier (URI) Scheme and Applicability Statement for the Trivial File Transfer Protocol (TFTP) It supports the following unofficial TFTP Options as described at http://www.compuphase.com/tftp.htm: blksize2 - Block size restricted to powers of 2, excluding protocol headers rollover - Block counter roll-over (roll back to zero or to one) From the tftp program point of view the following things are changed: - New commands: "blocksize", "blocksize2", "rollover" and "options" - Development features: "debug" and "packetdrop" If you try this tftp/tftpd implementation, please let me know if it works (or doesn't work) and against which implementaion so I can get a list of confirmed working systems. Author: Edwin Groothuis <edwin@FreeBSD.org>
2010-05-04 06:19:19 +00:00
/* Find out who we are talking to and what we are going to do */
peerlen = sizeof(peer_sock);
n = recvfrom(0, recvbuffer, MAXPKTSIZE, 0,
(struct sockaddr *)&peer_sock, &peerlen);
1994-05-27 12:39:25 +00:00
if (n < 0) {
Go ahead and merge the work edwin@ on tftpd into the tree. It is a lot better than what's in the tree now. Edwin tested it at a prior employer, but can't test it today. I've found that it works a lot better with the various uboot versions that I've used in my embedded work. Here's the pkg-descr from the port that describes the changes: It all started when we got some new routers, which told me the following when trying to upload configuration or download images from it: The TFTP server doesn't support the blocksize option. My curiousity was triggered, it took me some reading of RFCs and other documentation to find out what was possible and what could be done. Was plain TFTP very simple in its handshake, TFTP with options was kind of messy because of its backwards capability: The first packet returned could either be an acknowledgement of options, or the first data packet. Going through the source code of src/libexec/tftpd and going through the code of src/usr.bin/tftp showed that there was a lot of duplicate code, and the addition of options would only increase the amount of duplicate code. After all, both the client and the server can act as a sender and receiver. At the end, it ended up with a nearly complete rewrite of the tftp client and server. It has been tested against the following TFTP clients and servers: - Itself (yay!) - The standard FreeBSD tftp client and server - The Fedora Core 6 tftp client and server - Cisco router tftp client - Extreme Networks tftp client It supports the following RFCs: RFC1350 - THE TFTP PROTOCOL (REVISION 2) RFC2347 - TFTP Option Extension RFC2348 - TFTP Blocksize Option RFC2349 - TFTP Timeout Interval and Transfer Size Options RFC3617 - Uniform Resource Identifier (URI) Scheme and Applicability Statement for the Trivial File Transfer Protocol (TFTP) It supports the following unofficial TFTP Options as described at http://www.compuphase.com/tftp.htm: blksize2 - Block size restricted to powers of 2, excluding protocol headers rollover - Block counter roll-over (roll back to zero or to one) From the tftp program point of view the following things are changed: - New commands: "blocksize", "blocksize2", "rollover" and "options" - Development features: "debug" and "packetdrop" If you try this tftp/tftpd implementation, please let me know if it works (or doesn't work) and against which implementaion so I can get a list of confirmed working systems. Author: Edwin Groothuis <edwin@FreeBSD.org>
2010-05-04 06:19:19 +00:00
tftp_log(LOG_ERR, "recvfrom: %s", strerror(errno));
1994-05-27 12:39:25 +00:00
exit(1);
}
Go ahead and merge the work edwin@ on tftpd into the tree. It is a lot better than what's in the tree now. Edwin tested it at a prior employer, but can't test it today. I've found that it works a lot better with the various uboot versions that I've used in my embedded work. Here's the pkg-descr from the port that describes the changes: It all started when we got some new routers, which told me the following when trying to upload configuration or download images from it: The TFTP server doesn't support the blocksize option. My curiousity was triggered, it took me some reading of RFCs and other documentation to find out what was possible and what could be done. Was plain TFTP very simple in its handshake, TFTP with options was kind of messy because of its backwards capability: The first packet returned could either be an acknowledgement of options, or the first data packet. Going through the source code of src/libexec/tftpd and going through the code of src/usr.bin/tftp showed that there was a lot of duplicate code, and the addition of options would only increase the amount of duplicate code. After all, both the client and the server can act as a sender and receiver. At the end, it ended up with a nearly complete rewrite of the tftp client and server. It has been tested against the following TFTP clients and servers: - Itself (yay!) - The standard FreeBSD tftp client and server - The Fedora Core 6 tftp client and server - Cisco router tftp client - Extreme Networks tftp client It supports the following RFCs: RFC1350 - THE TFTP PROTOCOL (REVISION 2) RFC2347 - TFTP Option Extension RFC2348 - TFTP Blocksize Option RFC2349 - TFTP Timeout Interval and Transfer Size Options RFC3617 - Uniform Resource Identifier (URI) Scheme and Applicability Statement for the Trivial File Transfer Protocol (TFTP) It supports the following unofficial TFTP Options as described at http://www.compuphase.com/tftp.htm: blksize2 - Block size restricted to powers of 2, excluding protocol headers rollover - Block counter roll-over (roll back to zero or to one) From the tftp program point of view the following things are changed: - New commands: "blocksize", "blocksize2", "rollover" and "options" - Development features: "debug" and "packetdrop" If you try this tftp/tftpd implementation, please let me know if it works (or doesn't work) and against which implementaion so I can get a list of confirmed working systems. Author: Edwin Groothuis <edwin@FreeBSD.org>
2010-05-04 06:19:19 +00:00
getnameinfo((struct sockaddr *)&peer_sock, peer_sock.ss_len,
peername, sizeof(peername), NULL, 0, NI_NUMERICHOST);
1994-05-27 12:39:25 +00:00
/*
* 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 i, pid;
1994-05-27 12:39:25 +00:00
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.
*/
Go ahead and merge the work edwin@ on tftpd into the tree. It is a lot better than what's in the tree now. Edwin tested it at a prior employer, but can't test it today. I've found that it works a lot better with the various uboot versions that I've used in my embedded work. Here's the pkg-descr from the port that describes the changes: It all started when we got some new routers, which told me the following when trying to upload configuration or download images from it: The TFTP server doesn't support the blocksize option. My curiousity was triggered, it took me some reading of RFCs and other documentation to find out what was possible and what could be done. Was plain TFTP very simple in its handshake, TFTP with options was kind of messy because of its backwards capability: The first packet returned could either be an acknowledgement of options, or the first data packet. Going through the source code of src/libexec/tftpd and going through the code of src/usr.bin/tftp showed that there was a lot of duplicate code, and the addition of options would only increase the amount of duplicate code. After all, both the client and the server can act as a sender and receiver. At the end, it ended up with a nearly complete rewrite of the tftp client and server. It has been tested against the following TFTP clients and servers: - Itself (yay!) - The standard FreeBSD tftp client and server - The Fedora Core 6 tftp client and server - Cisco router tftp client - Extreme Networks tftp client It supports the following RFCs: RFC1350 - THE TFTP PROTOCOL (REVISION 2) RFC2347 - TFTP Option Extension RFC2348 - TFTP Blocksize Option RFC2349 - TFTP Timeout Interval and Transfer Size Options RFC3617 - Uniform Resource Identifier (URI) Scheme and Applicability Statement for the Trivial File Transfer Protocol (TFTP) It supports the following unofficial TFTP Options as described at http://www.compuphase.com/tftp.htm: blksize2 - Block size restricted to powers of 2, excluding protocol headers rollover - Block counter roll-over (roll back to zero or to one) From the tftp program point of view the following things are changed: - New commands: "blocksize", "blocksize2", "rollover" and "options" - Development features: "debug" and "packetdrop" If you try this tftp/tftpd implementation, please let me know if it works (or doesn't work) and against which implementaion so I can get a list of confirmed working systems. Author: Edwin Groothuis <edwin@FreeBSD.org>
2010-05-04 06:19:19 +00:00
peerlen = sizeof peer_sock;
i = recvfrom(0, recvbuffer, MAXPKTSIZE, 0,
(struct sockaddr *)&peer_sock, &peerlen);
1994-05-27 12:39:25 +00:00
if (i > 0) {
n = i;
}
} else {
break;
}
}
if (pid < 0) {
Go ahead and merge the work edwin@ on tftpd into the tree. It is a lot better than what's in the tree now. Edwin tested it at a prior employer, but can't test it today. I've found that it works a lot better with the various uboot versions that I've used in my embedded work. Here's the pkg-descr from the port that describes the changes: It all started when we got some new routers, which told me the following when trying to upload configuration or download images from it: The TFTP server doesn't support the blocksize option. My curiousity was triggered, it took me some reading of RFCs and other documentation to find out what was possible and what could be done. Was plain TFTP very simple in its handshake, TFTP with options was kind of messy because of its backwards capability: The first packet returned could either be an acknowledgement of options, or the first data packet. Going through the source code of src/libexec/tftpd and going through the code of src/usr.bin/tftp showed that there was a lot of duplicate code, and the addition of options would only increase the amount of duplicate code. After all, both the client and the server can act as a sender and receiver. At the end, it ended up with a nearly complete rewrite of the tftp client and server. It has been tested against the following TFTP clients and servers: - Itself (yay!) - The standard FreeBSD tftp client and server - The Fedora Core 6 tftp client and server - Cisco router tftp client - Extreme Networks tftp client It supports the following RFCs: RFC1350 - THE TFTP PROTOCOL (REVISION 2) RFC2347 - TFTP Option Extension RFC2348 - TFTP Blocksize Option RFC2349 - TFTP Timeout Interval and Transfer Size Options RFC3617 - Uniform Resource Identifier (URI) Scheme and Applicability Statement for the Trivial File Transfer Protocol (TFTP) It supports the following unofficial TFTP Options as described at http://www.compuphase.com/tftp.htm: blksize2 - Block size restricted to powers of 2, excluding protocol headers rollover - Block counter roll-over (roll back to zero or to one) From the tftp program point of view the following things are changed: - New commands: "blocksize", "blocksize2", "rollover" and "options" - Development features: "debug" and "packetdrop" If you try this tftp/tftpd implementation, please let me know if it works (or doesn't work) and against which implementaion so I can get a list of confirmed working systems. Author: Edwin Groothuis <edwin@FreeBSD.org>
2010-05-04 06:19:19 +00:00
tftp_log(LOG_ERR, "fork: %s", strerror(errno));
1994-05-27 12:39:25 +00:00
exit(1);
} else if (pid != 0) {
exit(0);
}
}
#ifdef LIBWRAP
Go ahead and merge the work edwin@ on tftpd into the tree. It is a lot better than what's in the tree now. Edwin tested it at a prior employer, but can't test it today. I've found that it works a lot better with the various uboot versions that I've used in my embedded work. Here's the pkg-descr from the port that describes the changes: It all started when we got some new routers, which told me the following when trying to upload configuration or download images from it: The TFTP server doesn't support the blocksize option. My curiousity was triggered, it took me some reading of RFCs and other documentation to find out what was possible and what could be done. Was plain TFTP very simple in its handshake, TFTP with options was kind of messy because of its backwards capability: The first packet returned could either be an acknowledgement of options, or the first data packet. Going through the source code of src/libexec/tftpd and going through the code of src/usr.bin/tftp showed that there was a lot of duplicate code, and the addition of options would only increase the amount of duplicate code. After all, both the client and the server can act as a sender and receiver. At the end, it ended up with a nearly complete rewrite of the tftp client and server. It has been tested against the following TFTP clients and servers: - Itself (yay!) - The standard FreeBSD tftp client and server - The Fedora Core 6 tftp client and server - Cisco router tftp client - Extreme Networks tftp client It supports the following RFCs: RFC1350 - THE TFTP PROTOCOL (REVISION 2) RFC2347 - TFTP Option Extension RFC2348 - TFTP Blocksize Option RFC2349 - TFTP Timeout Interval and Transfer Size Options RFC3617 - Uniform Resource Identifier (URI) Scheme and Applicability Statement for the Trivial File Transfer Protocol (TFTP) It supports the following unofficial TFTP Options as described at http://www.compuphase.com/tftp.htm: blksize2 - Block size restricted to powers of 2, excluding protocol headers rollover - Block counter roll-over (roll back to zero or to one) From the tftp program point of view the following things are changed: - New commands: "blocksize", "blocksize2", "rollover" and "options" - Development features: "debug" and "packetdrop" If you try this tftp/tftpd implementation, please let me know if it works (or doesn't work) and against which implementaion so I can get a list of confirmed working systems. Author: Edwin Groothuis <edwin@FreeBSD.org>
2010-05-04 06:19:19 +00:00
/*
* See if the client is allowed to talk to me.
* (This needs to be done before the chroot())
*/
{
struct request_info req;
request_init(&req, RQ_CLIENT_ADDR, peername, 0);
request_set(&req, RQ_DAEMON, "tftpd", 0);
if (hosts_access(&req) == 0) {
if (debug&DEBUG_ACCESS)
tftp_log(LOG_WARNING,
"Access denied by 'tftpd' entry "
"in /etc/hosts.allow");
/*
* Full access might be disabled, but maybe the
* client is allowed to do read-only access.
*/
request_set(&req, RQ_DAEMON, "tftpd-ro", 0);
allow_ro = hosts_access(&req);
request_set(&req, RQ_DAEMON, "tftpd-wo", 0);
allow_wo = hosts_access(&req);
if (allow_ro == 0 && allow_wo == 0) {
tftp_log(LOG_WARNING,
"Unauthorized access from %s", peername);
exit(1);
}
if (debug&DEBUG_ACCESS) {
if (allow_ro)
tftp_log(LOG_WARNING,
"But allowed readonly access "
"via 'tftpd-ro' entry");
if (allow_wo)
tftp_log(LOG_WARNING,
"But allowed writeonly access "
"via 'tftpd-wo' entry");
}
} else
if (debug&DEBUG_ACCESS)
tftp_log(LOG_WARNING,
"Full access allowed"
"in /etc/hosts.allow");
}
#endif
Go ahead and merge the work edwin@ on tftpd into the tree. It is a lot better than what's in the tree now. Edwin tested it at a prior employer, but can't test it today. I've found that it works a lot better with the various uboot versions that I've used in my embedded work. Here's the pkg-descr from the port that describes the changes: It all started when we got some new routers, which told me the following when trying to upload configuration or download images from it: The TFTP server doesn't support the blocksize option. My curiousity was triggered, it took me some reading of RFCs and other documentation to find out what was possible and what could be done. Was plain TFTP very simple in its handshake, TFTP with options was kind of messy because of its backwards capability: The first packet returned could either be an acknowledgement of options, or the first data packet. Going through the source code of src/libexec/tftpd and going through the code of src/usr.bin/tftp showed that there was a lot of duplicate code, and the addition of options would only increase the amount of duplicate code. After all, both the client and the server can act as a sender and receiver. At the end, it ended up with a nearly complete rewrite of the tftp client and server. It has been tested against the following TFTP clients and servers: - Itself (yay!) - The standard FreeBSD tftp client and server - The Fedora Core 6 tftp client and server - Cisco router tftp client - Extreme Networks tftp client It supports the following RFCs: RFC1350 - THE TFTP PROTOCOL (REVISION 2) RFC2347 - TFTP Option Extension RFC2348 - TFTP Blocksize Option RFC2349 - TFTP Timeout Interval and Transfer Size Options RFC3617 - Uniform Resource Identifier (URI) Scheme and Applicability Statement for the Trivial File Transfer Protocol (TFTP) It supports the following unofficial TFTP Options as described at http://www.compuphase.com/tftp.htm: blksize2 - Block size restricted to powers of 2, excluding protocol headers rollover - Block counter roll-over (roll back to zero or to one) From the tftp program point of view the following things are changed: - New commands: "blocksize", "blocksize2", "rollover" and "options" - Development features: "debug" and "packetdrop" If you try this tftp/tftpd implementation, please let me know if it works (or doesn't work) and against which implementaion so I can get a list of confirmed working systems. Author: Edwin Groothuis <edwin@FreeBSD.org>
2010-05-04 06:19:19 +00:00
/*
* 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 > 0) {
char *tempchroot;
struct stat sb;
int statret;
struct sockaddr_storage ss;
char hbuf[NI_MAXHOST];
Go ahead and merge the work edwin@ on tftpd into the tree. It is a lot better than what's in the tree now. Edwin tested it at a prior employer, but can't test it today. I've found that it works a lot better with the various uboot versions that I've used in my embedded work. Here's the pkg-descr from the port that describes the changes: It all started when we got some new routers, which told me the following when trying to upload configuration or download images from it: The TFTP server doesn't support the blocksize option. My curiousity was triggered, it took me some reading of RFCs and other documentation to find out what was possible and what could be done. Was plain TFTP very simple in its handshake, TFTP with options was kind of messy because of its backwards capability: The first packet returned could either be an acknowledgement of options, or the first data packet. Going through the source code of src/libexec/tftpd and going through the code of src/usr.bin/tftp showed that there was a lot of duplicate code, and the addition of options would only increase the amount of duplicate code. After all, both the client and the server can act as a sender and receiver. At the end, it ended up with a nearly complete rewrite of the tftp client and server. It has been tested against the following TFTP clients and servers: - Itself (yay!) - The standard FreeBSD tftp client and server - The Fedora Core 6 tftp client and server - Cisco router tftp client - Extreme Networks tftp client It supports the following RFCs: RFC1350 - THE TFTP PROTOCOL (REVISION 2) RFC2347 - TFTP Option Extension RFC2348 - TFTP Blocksize Option RFC2349 - TFTP Timeout Interval and Transfer Size Options RFC3617 - Uniform Resource Identifier (URI) Scheme and Applicability Statement for the Trivial File Transfer Protocol (TFTP) It supports the following unofficial TFTP Options as described at http://www.compuphase.com/tftp.htm: blksize2 - Block size restricted to powers of 2, excluding protocol headers rollover - Block counter roll-over (roll back to zero or to one) From the tftp program point of view the following things are changed: - New commands: "blocksize", "blocksize2", "rollover" and "options" - Development features: "debug" and "packetdrop" If you try this tftp/tftpd implementation, please let me know if it works (or doesn't work) and against which implementaion so I can get a list of confirmed working systems. Author: Edwin Groothuis <edwin@FreeBSD.org>
2010-05-04 06:19:19 +00:00
statret = -1;
memcpy(&ss, &peer_sock, peer_sock.ss_len);
unmappedaddr((struct sockaddr_in6 *)&ss);
getnameinfo((struct sockaddr *)&ss, ss.ss_len,
hbuf, sizeof(hbuf), NULL, 0,
NI_NUMERICHOST);
asprintf(&tempchroot, "%s/%s", chroot_dir, hbuf);
if (ipchroot == 2)
statret = stat(tempchroot, &sb);
if (ipchroot == 1 ||
(statret == 0 && (sb.st_mode & S_IFDIR)))
chroot_dir = tempchroot;
}
/* Must get this before chroot because /etc might go away */
if ((nobody = getpwnam(chuser)) == NULL) {
Go ahead and merge the work edwin@ on tftpd into the tree. It is a lot better than what's in the tree now. Edwin tested it at a prior employer, but can't test it today. I've found that it works a lot better with the various uboot versions that I've used in my embedded work. Here's the pkg-descr from the port that describes the changes: It all started when we got some new routers, which told me the following when trying to upload configuration or download images from it: The TFTP server doesn't support the blocksize option. My curiousity was triggered, it took me some reading of RFCs and other documentation to find out what was possible and what could be done. Was plain TFTP very simple in its handshake, TFTP with options was kind of messy because of its backwards capability: The first packet returned could either be an acknowledgement of options, or the first data packet. Going through the source code of src/libexec/tftpd and going through the code of src/usr.bin/tftp showed that there was a lot of duplicate code, and the addition of options would only increase the amount of duplicate code. After all, both the client and the server can act as a sender and receiver. At the end, it ended up with a nearly complete rewrite of the tftp client and server. It has been tested against the following TFTP clients and servers: - Itself (yay!) - The standard FreeBSD tftp client and server - The Fedora Core 6 tftp client and server - Cisco router tftp client - Extreme Networks tftp client It supports the following RFCs: RFC1350 - THE TFTP PROTOCOL (REVISION 2) RFC2347 - TFTP Option Extension RFC2348 - TFTP Blocksize Option RFC2349 - TFTP Timeout Interval and Transfer Size Options RFC3617 - Uniform Resource Identifier (URI) Scheme and Applicability Statement for the Trivial File Transfer Protocol (TFTP) It supports the following unofficial TFTP Options as described at http://www.compuphase.com/tftp.htm: blksize2 - Block size restricted to powers of 2, excluding protocol headers rollover - Block counter roll-over (roll back to zero or to one) From the tftp program point of view the following things are changed: - New commands: "blocksize", "blocksize2", "rollover" and "options" - Development features: "debug" and "packetdrop" If you try this tftp/tftpd implementation, please let me know if it works (or doesn't work) and against which implementaion so I can get a list of confirmed working systems. Author: Edwin Groothuis <edwin@FreeBSD.org>
2010-05-04 06:19:19 +00:00
tftp_log(LOG_ERR, "%s: no such user", chuser);
exit(1);
}
if (chroot(chroot_dir)) {
Go ahead and merge the work edwin@ on tftpd into the tree. It is a lot better than what's in the tree now. Edwin tested it at a prior employer, but can't test it today. I've found that it works a lot better with the various uboot versions that I've used in my embedded work. Here's the pkg-descr from the port that describes the changes: It all started when we got some new routers, which told me the following when trying to upload configuration or download images from it: The TFTP server doesn't support the blocksize option. My curiousity was triggered, it took me some reading of RFCs and other documentation to find out what was possible and what could be done. Was plain TFTP very simple in its handshake, TFTP with options was kind of messy because of its backwards capability: The first packet returned could either be an acknowledgement of options, or the first data packet. Going through the source code of src/libexec/tftpd and going through the code of src/usr.bin/tftp showed that there was a lot of duplicate code, and the addition of options would only increase the amount of duplicate code. After all, both the client and the server can act as a sender and receiver. At the end, it ended up with a nearly complete rewrite of the tftp client and server. It has been tested against the following TFTP clients and servers: - Itself (yay!) - The standard FreeBSD tftp client and server - The Fedora Core 6 tftp client and server - Cisco router tftp client - Extreme Networks tftp client It supports the following RFCs: RFC1350 - THE TFTP PROTOCOL (REVISION 2) RFC2347 - TFTP Option Extension RFC2348 - TFTP Blocksize Option RFC2349 - TFTP Timeout Interval and Transfer Size Options RFC3617 - Uniform Resource Identifier (URI) Scheme and Applicability Statement for the Trivial File Transfer Protocol (TFTP) It supports the following unofficial TFTP Options as described at http://www.compuphase.com/tftp.htm: blksize2 - Block size restricted to powers of 2, excluding protocol headers rollover - Block counter roll-over (roll back to zero or to one) From the tftp program point of view the following things are changed: - New commands: "blocksize", "blocksize2", "rollover" and "options" - Development features: "debug" and "packetdrop" If you try this tftp/tftpd implementation, please let me know if it works (or doesn't work) and against which implementaion so I can get a list of confirmed working systems. Author: Edwin Groothuis <edwin@FreeBSD.org>
2010-05-04 06:19:19 +00:00
tftp_log(LOG_ERR, "chroot: %s: %s",
chroot_dir, strerror(errno));
exit(1);
}
chdir("/");
setgroups(1, &nobody->pw_gid);
if (setuid(nobody->pw_uid) != 0) {
tftp_log(LOG_ERR, "setuid failed");
exit(1);
}
}
Go ahead and merge the work edwin@ on tftpd into the tree. It is a lot better than what's in the tree now. Edwin tested it at a prior employer, but can't test it today. I've found that it works a lot better with the various uboot versions that I've used in my embedded work. Here's the pkg-descr from the port that describes the changes: It all started when we got some new routers, which told me the following when trying to upload configuration or download images from it: The TFTP server doesn't support the blocksize option. My curiousity was triggered, it took me some reading of RFCs and other documentation to find out what was possible and what could be done. Was plain TFTP very simple in its handshake, TFTP with options was kind of messy because of its backwards capability: The first packet returned could either be an acknowledgement of options, or the first data packet. Going through the source code of src/libexec/tftpd and going through the code of src/usr.bin/tftp showed that there was a lot of duplicate code, and the addition of options would only increase the amount of duplicate code. After all, both the client and the server can act as a sender and receiver. At the end, it ended up with a nearly complete rewrite of the tftp client and server. It has been tested against the following TFTP clients and servers: - Itself (yay!) - The standard FreeBSD tftp client and server - The Fedora Core 6 tftp client and server - Cisco router tftp client - Extreme Networks tftp client It supports the following RFCs: RFC1350 - THE TFTP PROTOCOL (REVISION 2) RFC2347 - TFTP Option Extension RFC2348 - TFTP Blocksize Option RFC2349 - TFTP Timeout Interval and Transfer Size Options RFC3617 - Uniform Resource Identifier (URI) Scheme and Applicability Statement for the Trivial File Transfer Protocol (TFTP) It supports the following unofficial TFTP Options as described at http://www.compuphase.com/tftp.htm: blksize2 - Block size restricted to powers of 2, excluding protocol headers rollover - Block counter roll-over (roll back to zero or to one) From the tftp program point of view the following things are changed: - New commands: "blocksize", "blocksize2", "rollover" and "options" - Development features: "debug" and "packetdrop" If you try this tftp/tftpd implementation, please let me know if it works (or doesn't work) and against which implementaion so I can get a list of confirmed working systems. Author: Edwin Groothuis <edwin@FreeBSD.org>
2010-05-04 06:19:19 +00:00
len = sizeof(me_sock);
if (getsockname(0, (struct sockaddr *)&me_sock, &len) == 0) {
switch (me_sock.ss_family) {
case AF_INET:
Go ahead and merge the work edwin@ on tftpd into the tree. It is a lot better than what's in the tree now. Edwin tested it at a prior employer, but can't test it today. I've found that it works a lot better with the various uboot versions that I've used in my embedded work. Here's the pkg-descr from the port that describes the changes: It all started when we got some new routers, which told me the following when trying to upload configuration or download images from it: The TFTP server doesn't support the blocksize option. My curiousity was triggered, it took me some reading of RFCs and other documentation to find out what was possible and what could be done. Was plain TFTP very simple in its handshake, TFTP with options was kind of messy because of its backwards capability: The first packet returned could either be an acknowledgement of options, or the first data packet. Going through the source code of src/libexec/tftpd and going through the code of src/usr.bin/tftp showed that there was a lot of duplicate code, and the addition of options would only increase the amount of duplicate code. After all, both the client and the server can act as a sender and receiver. At the end, it ended up with a nearly complete rewrite of the tftp client and server. It has been tested against the following TFTP clients and servers: - Itself (yay!) - The standard FreeBSD tftp client and server - The Fedora Core 6 tftp client and server - Cisco router tftp client - Extreme Networks tftp client It supports the following RFCs: RFC1350 - THE TFTP PROTOCOL (REVISION 2) RFC2347 - TFTP Option Extension RFC2348 - TFTP Blocksize Option RFC2349 - TFTP Timeout Interval and Transfer Size Options RFC3617 - Uniform Resource Identifier (URI) Scheme and Applicability Statement for the Trivial File Transfer Protocol (TFTP) It supports the following unofficial TFTP Options as described at http://www.compuphase.com/tftp.htm: blksize2 - Block size restricted to powers of 2, excluding protocol headers rollover - Block counter roll-over (roll back to zero or to one) From the tftp program point of view the following things are changed: - New commands: "blocksize", "blocksize2", "rollover" and "options" - Development features: "debug" and "packetdrop" If you try this tftp/tftpd implementation, please let me know if it works (or doesn't work) and against which implementaion so I can get a list of confirmed working systems. Author: Edwin Groothuis <edwin@FreeBSD.org>
2010-05-04 06:19:19 +00:00
((struct sockaddr_in *)&me_sock)->sin_port = 0;
break;
case AF_INET6:
Go ahead and merge the work edwin@ on tftpd into the tree. It is a lot better than what's in the tree now. Edwin tested it at a prior employer, but can't test it today. I've found that it works a lot better with the various uboot versions that I've used in my embedded work. Here's the pkg-descr from the port that describes the changes: It all started when we got some new routers, which told me the following when trying to upload configuration or download images from it: The TFTP server doesn't support the blocksize option. My curiousity was triggered, it took me some reading of RFCs and other documentation to find out what was possible and what could be done. Was plain TFTP very simple in its handshake, TFTP with options was kind of messy because of its backwards capability: The first packet returned could either be an acknowledgement of options, or the first data packet. Going through the source code of src/libexec/tftpd and going through the code of src/usr.bin/tftp showed that there was a lot of duplicate code, and the addition of options would only increase the amount of duplicate code. After all, both the client and the server can act as a sender and receiver. At the end, it ended up with a nearly complete rewrite of the tftp client and server. It has been tested against the following TFTP clients and servers: - Itself (yay!) - The standard FreeBSD tftp client and server - The Fedora Core 6 tftp client and server - Cisco router tftp client - Extreme Networks tftp client It supports the following RFCs: RFC1350 - THE TFTP PROTOCOL (REVISION 2) RFC2347 - TFTP Option Extension RFC2348 - TFTP Blocksize Option RFC2349 - TFTP Timeout Interval and Transfer Size Options RFC3617 - Uniform Resource Identifier (URI) Scheme and Applicability Statement for the Trivial File Transfer Protocol (TFTP) It supports the following unofficial TFTP Options as described at http://www.compuphase.com/tftp.htm: blksize2 - Block size restricted to powers of 2, excluding protocol headers rollover - Block counter roll-over (roll back to zero or to one) From the tftp program point of view the following things are changed: - New commands: "blocksize", "blocksize2", "rollover" and "options" - Development features: "debug" and "packetdrop" If you try this tftp/tftpd implementation, please let me know if it works (or doesn't work) and against which implementaion so I can get a list of confirmed working systems. Author: Edwin Groothuis <edwin@FreeBSD.org>
2010-05-04 06:19:19 +00:00
((struct sockaddr_in6 *)&me_sock)->sin6_port = 0;
break;
default:
/* unsupported */
break;
}
} else {
Go ahead and merge the work edwin@ on tftpd into the tree. It is a lot better than what's in the tree now. Edwin tested it at a prior employer, but can't test it today. I've found that it works a lot better with the various uboot versions that I've used in my embedded work. Here's the pkg-descr from the port that describes the changes: It all started when we got some new routers, which told me the following when trying to upload configuration or download images from it: The TFTP server doesn't support the blocksize option. My curiousity was triggered, it took me some reading of RFCs and other documentation to find out what was possible and what could be done. Was plain TFTP very simple in its handshake, TFTP with options was kind of messy because of its backwards capability: The first packet returned could either be an acknowledgement of options, or the first data packet. Going through the source code of src/libexec/tftpd and going through the code of src/usr.bin/tftp showed that there was a lot of duplicate code, and the addition of options would only increase the amount of duplicate code. After all, both the client and the server can act as a sender and receiver. At the end, it ended up with a nearly complete rewrite of the tftp client and server. It has been tested against the following TFTP clients and servers: - Itself (yay!) - The standard FreeBSD tftp client and server - The Fedora Core 6 tftp client and server - Cisco router tftp client - Extreme Networks tftp client It supports the following RFCs: RFC1350 - THE TFTP PROTOCOL (REVISION 2) RFC2347 - TFTP Option Extension RFC2348 - TFTP Blocksize Option RFC2349 - TFTP Timeout Interval and Transfer Size Options RFC3617 - Uniform Resource Identifier (URI) Scheme and Applicability Statement for the Trivial File Transfer Protocol (TFTP) It supports the following unofficial TFTP Options as described at http://www.compuphase.com/tftp.htm: blksize2 - Block size restricted to powers of 2, excluding protocol headers rollover - Block counter roll-over (roll back to zero or to one) From the tftp program point of view the following things are changed: - New commands: "blocksize", "blocksize2", "rollover" and "options" - Development features: "debug" and "packetdrop" If you try this tftp/tftpd implementation, please let me know if it works (or doesn't work) and against which implementaion so I can get a list of confirmed working systems. Author: Edwin Groothuis <edwin@FreeBSD.org>
2010-05-04 06:19:19 +00:00
memset(&me_sock, 0, sizeof(me_sock));
me_sock.ss_family = peer_sock.ss_family;
me_sock.ss_len = peer_sock.ss_len;
}
1994-05-27 12:39:25 +00:00
close(0);
close(1);
Go ahead and merge the work edwin@ on tftpd into the tree. It is a lot better than what's in the tree now. Edwin tested it at a prior employer, but can't test it today. I've found that it works a lot better with the various uboot versions that I've used in my embedded work. Here's the pkg-descr from the port that describes the changes: It all started when we got some new routers, which told me the following when trying to upload configuration or download images from it: The TFTP server doesn't support the blocksize option. My curiousity was triggered, it took me some reading of RFCs and other documentation to find out what was possible and what could be done. Was plain TFTP very simple in its handshake, TFTP with options was kind of messy because of its backwards capability: The first packet returned could either be an acknowledgement of options, or the first data packet. Going through the source code of src/libexec/tftpd and going through the code of src/usr.bin/tftp showed that there was a lot of duplicate code, and the addition of options would only increase the amount of duplicate code. After all, both the client and the server can act as a sender and receiver. At the end, it ended up with a nearly complete rewrite of the tftp client and server. It has been tested against the following TFTP clients and servers: - Itself (yay!) - The standard FreeBSD tftp client and server - The Fedora Core 6 tftp client and server - Cisco router tftp client - Extreme Networks tftp client It supports the following RFCs: RFC1350 - THE TFTP PROTOCOL (REVISION 2) RFC2347 - TFTP Option Extension RFC2348 - TFTP Blocksize Option RFC2349 - TFTP Timeout Interval and Transfer Size Options RFC3617 - Uniform Resource Identifier (URI) Scheme and Applicability Statement for the Trivial File Transfer Protocol (TFTP) It supports the following unofficial TFTP Options as described at http://www.compuphase.com/tftp.htm: blksize2 - Block size restricted to powers of 2, excluding protocol headers rollover - Block counter roll-over (roll back to zero or to one) From the tftp program point of view the following things are changed: - New commands: "blocksize", "blocksize2", "rollover" and "options" - Development features: "debug" and "packetdrop" If you try this tftp/tftpd implementation, please let me know if it works (or doesn't work) and against which implementaion so I can get a list of confirmed working systems. Author: Edwin Groothuis <edwin@FreeBSD.org>
2010-05-04 06:19:19 +00:00
peer = socket(peer_sock.ss_family, SOCK_DGRAM, 0);
1994-05-27 12:39:25 +00:00
if (peer < 0) {
Go ahead and merge the work edwin@ on tftpd into the tree. It is a lot better than what's in the tree now. Edwin tested it at a prior employer, but can't test it today. I've found that it works a lot better with the various uboot versions that I've used in my embedded work. Here's the pkg-descr from the port that describes the changes: It all started when we got some new routers, which told me the following when trying to upload configuration or download images from it: The TFTP server doesn't support the blocksize option. My curiousity was triggered, it took me some reading of RFCs and other documentation to find out what was possible and what could be done. Was plain TFTP very simple in its handshake, TFTP with options was kind of messy because of its backwards capability: The first packet returned could either be an acknowledgement of options, or the first data packet. Going through the source code of src/libexec/tftpd and going through the code of src/usr.bin/tftp showed that there was a lot of duplicate code, and the addition of options would only increase the amount of duplicate code. After all, both the client and the server can act as a sender and receiver. At the end, it ended up with a nearly complete rewrite of the tftp client and server. It has been tested against the following TFTP clients and servers: - Itself (yay!) - The standard FreeBSD tftp client and server - The Fedora Core 6 tftp client and server - Cisco router tftp client - Extreme Networks tftp client It supports the following RFCs: RFC1350 - THE TFTP PROTOCOL (REVISION 2) RFC2347 - TFTP Option Extension RFC2348 - TFTP Blocksize Option RFC2349 - TFTP Timeout Interval and Transfer Size Options RFC3617 - Uniform Resource Identifier (URI) Scheme and Applicability Statement for the Trivial File Transfer Protocol (TFTP) It supports the following unofficial TFTP Options as described at http://www.compuphase.com/tftp.htm: blksize2 - Block size restricted to powers of 2, excluding protocol headers rollover - Block counter roll-over (roll back to zero or to one) From the tftp program point of view the following things are changed: - New commands: "blocksize", "blocksize2", "rollover" and "options" - Development features: "debug" and "packetdrop" If you try this tftp/tftpd implementation, please let me know if it works (or doesn't work) and against which implementaion so I can get a list of confirmed working systems. Author: Edwin Groothuis <edwin@FreeBSD.org>
2010-05-04 06:19:19 +00:00
tftp_log(LOG_ERR, "socket: %s", strerror(errno));
1994-05-27 12:39:25 +00:00
exit(1);
}
Go ahead and merge the work edwin@ on tftpd into the tree. It is a lot better than what's in the tree now. Edwin tested it at a prior employer, but can't test it today. I've found that it works a lot better with the various uboot versions that I've used in my embedded work. Here's the pkg-descr from the port that describes the changes: It all started when we got some new routers, which told me the following when trying to upload configuration or download images from it: The TFTP server doesn't support the blocksize option. My curiousity was triggered, it took me some reading of RFCs and other documentation to find out what was possible and what could be done. Was plain TFTP very simple in its handshake, TFTP with options was kind of messy because of its backwards capability: The first packet returned could either be an acknowledgement of options, or the first data packet. Going through the source code of src/libexec/tftpd and going through the code of src/usr.bin/tftp showed that there was a lot of duplicate code, and the addition of options would only increase the amount of duplicate code. After all, both the client and the server can act as a sender and receiver. At the end, it ended up with a nearly complete rewrite of the tftp client and server. It has been tested against the following TFTP clients and servers: - Itself (yay!) - The standard FreeBSD tftp client and server - The Fedora Core 6 tftp client and server - Cisco router tftp client - Extreme Networks tftp client It supports the following RFCs: RFC1350 - THE TFTP PROTOCOL (REVISION 2) RFC2347 - TFTP Option Extension RFC2348 - TFTP Blocksize Option RFC2349 - TFTP Timeout Interval and Transfer Size Options RFC3617 - Uniform Resource Identifier (URI) Scheme and Applicability Statement for the Trivial File Transfer Protocol (TFTP) It supports the following unofficial TFTP Options as described at http://www.compuphase.com/tftp.htm: blksize2 - Block size restricted to powers of 2, excluding protocol headers rollover - Block counter roll-over (roll back to zero or to one) From the tftp program point of view the following things are changed: - New commands: "blocksize", "blocksize2", "rollover" and "options" - Development features: "debug" and "packetdrop" If you try this tftp/tftpd implementation, please let me know if it works (or doesn't work) and against which implementaion so I can get a list of confirmed working systems. Author: Edwin Groothuis <edwin@FreeBSD.org>
2010-05-04 06:19:19 +00:00
if (bind(peer, (struct sockaddr *)&me_sock, me_sock.ss_len) < 0) {
tftp_log(LOG_ERR, "bind: %s", strerror(errno));
1994-05-27 12:39:25 +00:00
exit(1);
}
Go ahead and merge the work edwin@ on tftpd into the tree. It is a lot better than what's in the tree now. Edwin tested it at a prior employer, but can't test it today. I've found that it works a lot better with the various uboot versions that I've used in my embedded work. Here's the pkg-descr from the port that describes the changes: It all started when we got some new routers, which told me the following when trying to upload configuration or download images from it: The TFTP server doesn't support the blocksize option. My curiousity was triggered, it took me some reading of RFCs and other documentation to find out what was possible and what could be done. Was plain TFTP very simple in its handshake, TFTP with options was kind of messy because of its backwards capability: The first packet returned could either be an acknowledgement of options, or the first data packet. Going through the source code of src/libexec/tftpd and going through the code of src/usr.bin/tftp showed that there was a lot of duplicate code, and the addition of options would only increase the amount of duplicate code. After all, both the client and the server can act as a sender and receiver. At the end, it ended up with a nearly complete rewrite of the tftp client and server. It has been tested against the following TFTP clients and servers: - Itself (yay!) - The standard FreeBSD tftp client and server - The Fedora Core 6 tftp client and server - Cisco router tftp client - Extreme Networks tftp client It supports the following RFCs: RFC1350 - THE TFTP PROTOCOL (REVISION 2) RFC2347 - TFTP Option Extension RFC2348 - TFTP Blocksize Option RFC2349 - TFTP Timeout Interval and Transfer Size Options RFC3617 - Uniform Resource Identifier (URI) Scheme and Applicability Statement for the Trivial File Transfer Protocol (TFTP) It supports the following unofficial TFTP Options as described at http://www.compuphase.com/tftp.htm: blksize2 - Block size restricted to powers of 2, excluding protocol headers rollover - Block counter roll-over (roll back to zero or to one) From the tftp program point of view the following things are changed: - New commands: "blocksize", "blocksize2", "rollover" and "options" - Development features: "debug" and "packetdrop" If you try this tftp/tftpd implementation, please let me know if it works (or doesn't work) and against which implementaion so I can get a list of confirmed working systems. Author: Edwin Groothuis <edwin@FreeBSD.org>
2010-05-04 06:19:19 +00:00
tp = (struct tftphdr *)recvbuffer;
1994-05-27 12:39:25 +00:00
tp->th_opcode = ntohs(tp->th_opcode);
Go ahead and merge the work edwin@ on tftpd into the tree. It is a lot better than what's in the tree now. Edwin tested it at a prior employer, but can't test it today. I've found that it works a lot better with the various uboot versions that I've used in my embedded work. Here's the pkg-descr from the port that describes the changes: It all started when we got some new routers, which told me the following when trying to upload configuration or download images from it: The TFTP server doesn't support the blocksize option. My curiousity was triggered, it took me some reading of RFCs and other documentation to find out what was possible and what could be done. Was plain TFTP very simple in its handshake, TFTP with options was kind of messy because of its backwards capability: The first packet returned could either be an acknowledgement of options, or the first data packet. Going through the source code of src/libexec/tftpd and going through the code of src/usr.bin/tftp showed that there was a lot of duplicate code, and the addition of options would only increase the amount of duplicate code. After all, both the client and the server can act as a sender and receiver. At the end, it ended up with a nearly complete rewrite of the tftp client and server. It has been tested against the following TFTP clients and servers: - Itself (yay!) - The standard FreeBSD tftp client and server - The Fedora Core 6 tftp client and server - Cisco router tftp client - Extreme Networks tftp client It supports the following RFCs: RFC1350 - THE TFTP PROTOCOL (REVISION 2) RFC2347 - TFTP Option Extension RFC2348 - TFTP Blocksize Option RFC2349 - TFTP Timeout Interval and Transfer Size Options RFC3617 - Uniform Resource Identifier (URI) Scheme and Applicability Statement for the Trivial File Transfer Protocol (TFTP) It supports the following unofficial TFTP Options as described at http://www.compuphase.com/tftp.htm: blksize2 - Block size restricted to powers of 2, excluding protocol headers rollover - Block counter roll-over (roll back to zero or to one) From the tftp program point of view the following things are changed: - New commands: "blocksize", "blocksize2", "rollover" and "options" - Development features: "debug" and "packetdrop" If you try this tftp/tftpd implementation, please let me know if it works (or doesn't work) and against which implementaion so I can get a list of confirmed working systems. Author: Edwin Groothuis <edwin@FreeBSD.org>
2010-05-04 06:19:19 +00:00
if (tp->th_opcode == RRQ) {
if (allow_ro)
tftp_rrq(peer, tp->th_stuff, n - 1);
else {
tftp_log(LOG_WARNING,
"%s read access denied", peername);
exit(1);
}
}
if (tp->th_opcode == WRQ) {
if (allow_wo)
tftp_wrq(peer, tp->th_stuff, n - 1);
else {
tftp_log(LOG_WARNING,
"%s write access denied", peername);
exit(1);
}
}
1994-05-27 12:39:25 +00:00
exit(1);
}
static void
reduce_path(char *fn)
{
char *slash, *ptr;
/* Reduce all "/+./" to "/" (just in case we've got "/./../" later */
while ((slash = strstr(fn, "/./")) != NULL) {
for (ptr = slash; ptr > fn && ptr[-1] == '/'; ptr--)
;
slash += 2;
while (*slash)
*++ptr = *++slash;
}
/* Now reduce all "/something/+../" to "/" */
while ((slash = strstr(fn, "/../")) != NULL) {
if (slash == fn)
break;
for (ptr = slash; ptr > fn && ptr[-1] == '/'; ptr--)
;
for (ptr--; ptr >= fn; ptr--)
if (*ptr == '/')
break;
if (ptr < fn)
break;
slash += 3;
while (*slash)
*++ptr = *++slash;
}
}
Go ahead and merge the work edwin@ on tftpd into the tree. It is a lot better than what's in the tree now. Edwin tested it at a prior employer, but can't test it today. I've found that it works a lot better with the various uboot versions that I've used in my embedded work. Here's the pkg-descr from the port that describes the changes: It all started when we got some new routers, which told me the following when trying to upload configuration or download images from it: The TFTP server doesn't support the blocksize option. My curiousity was triggered, it took me some reading of RFCs and other documentation to find out what was possible and what could be done. Was plain TFTP very simple in its handshake, TFTP with options was kind of messy because of its backwards capability: The first packet returned could either be an acknowledgement of options, or the first data packet. Going through the source code of src/libexec/tftpd and going through the code of src/usr.bin/tftp showed that there was a lot of duplicate code, and the addition of options would only increase the amount of duplicate code. After all, both the client and the server can act as a sender and receiver. At the end, it ended up with a nearly complete rewrite of the tftp client and server. It has been tested against the following TFTP clients and servers: - Itself (yay!) - The standard FreeBSD tftp client and server - The Fedora Core 6 tftp client and server - Cisco router tftp client - Extreme Networks tftp client It supports the following RFCs: RFC1350 - THE TFTP PROTOCOL (REVISION 2) RFC2347 - TFTP Option Extension RFC2348 - TFTP Blocksize Option RFC2349 - TFTP Timeout Interval and Transfer Size Options RFC3617 - Uniform Resource Identifier (URI) Scheme and Applicability Statement for the Trivial File Transfer Protocol (TFTP) It supports the following unofficial TFTP Options as described at http://www.compuphase.com/tftp.htm: blksize2 - Block size restricted to powers of 2, excluding protocol headers rollover - Block counter roll-over (roll back to zero or to one) From the tftp program point of view the following things are changed: - New commands: "blocksize", "blocksize2", "rollover" and "options" - Development features: "debug" and "packetdrop" If you try this tftp/tftpd implementation, please let me know if it works (or doesn't work) and against which implementaion so I can get a list of confirmed working systems. Author: Edwin Groothuis <edwin@FreeBSD.org>
2010-05-04 06:19:19 +00:00
static char *
parse_header(int peer, char *recvbuffer, ssize_t size,
char **filename, char **mode)
{
char *cp;
int i;
struct formats *pf;
1994-05-27 12:39:25 +00:00
Go ahead and merge the work edwin@ on tftpd into the tree. It is a lot better than what's in the tree now. Edwin tested it at a prior employer, but can't test it today. I've found that it works a lot better with the various uboot versions that I've used in my embedded work. Here's the pkg-descr from the port that describes the changes: It all started when we got some new routers, which told me the following when trying to upload configuration or download images from it: The TFTP server doesn't support the blocksize option. My curiousity was triggered, it took me some reading of RFCs and other documentation to find out what was possible and what could be done. Was plain TFTP very simple in its handshake, TFTP with options was kind of messy because of its backwards capability: The first packet returned could either be an acknowledgement of options, or the first data packet. Going through the source code of src/libexec/tftpd and going through the code of src/usr.bin/tftp showed that there was a lot of duplicate code, and the addition of options would only increase the amount of duplicate code. After all, both the client and the server can act as a sender and receiver. At the end, it ended up with a nearly complete rewrite of the tftp client and server. It has been tested against the following TFTP clients and servers: - Itself (yay!) - The standard FreeBSD tftp client and server - The Fedora Core 6 tftp client and server - Cisco router tftp client - Extreme Networks tftp client It supports the following RFCs: RFC1350 - THE TFTP PROTOCOL (REVISION 2) RFC2347 - TFTP Option Extension RFC2348 - TFTP Blocksize Option RFC2349 - TFTP Timeout Interval and Transfer Size Options RFC3617 - Uniform Resource Identifier (URI) Scheme and Applicability Statement for the Trivial File Transfer Protocol (TFTP) It supports the following unofficial TFTP Options as described at http://www.compuphase.com/tftp.htm: blksize2 - Block size restricted to powers of 2, excluding protocol headers rollover - Block counter roll-over (roll back to zero or to one) From the tftp program point of view the following things are changed: - New commands: "blocksize", "blocksize2", "rollover" and "options" - Development features: "debug" and "packetdrop" If you try this tftp/tftpd implementation, please let me know if it works (or doesn't work) and against which implementaion so I can get a list of confirmed working systems. Author: Edwin Groothuis <edwin@FreeBSD.org>
2010-05-04 06:19:19 +00:00
*mode = NULL;
cp = recvbuffer;
1994-05-27 12:39:25 +00:00
Go ahead and merge the work edwin@ on tftpd into the tree. It is a lot better than what's in the tree now. Edwin tested it at a prior employer, but can't test it today. I've found that it works a lot better with the various uboot versions that I've used in my embedded work. Here's the pkg-descr from the port that describes the changes: It all started when we got some new routers, which told me the following when trying to upload configuration or download images from it: The TFTP server doesn't support the blocksize option. My curiousity was triggered, it took me some reading of RFCs and other documentation to find out what was possible and what could be done. Was plain TFTP very simple in its handshake, TFTP with options was kind of messy because of its backwards capability: The first packet returned could either be an acknowledgement of options, or the first data packet. Going through the source code of src/libexec/tftpd and going through the code of src/usr.bin/tftp showed that there was a lot of duplicate code, and the addition of options would only increase the amount of duplicate code. After all, both the client and the server can act as a sender and receiver. At the end, it ended up with a nearly complete rewrite of the tftp client and server. It has been tested against the following TFTP clients and servers: - Itself (yay!) - The standard FreeBSD tftp client and server - The Fedora Core 6 tftp client and server - Cisco router tftp client - Extreme Networks tftp client It supports the following RFCs: RFC1350 - THE TFTP PROTOCOL (REVISION 2) RFC2347 - TFTP Option Extension RFC2348 - TFTP Blocksize Option RFC2349 - TFTP Timeout Interval and Transfer Size Options RFC3617 - Uniform Resource Identifier (URI) Scheme and Applicability Statement for the Trivial File Transfer Protocol (TFTP) It supports the following unofficial TFTP Options as described at http://www.compuphase.com/tftp.htm: blksize2 - Block size restricted to powers of 2, excluding protocol headers rollover - Block counter roll-over (roll back to zero or to one) From the tftp program point of view the following things are changed: - New commands: "blocksize", "blocksize2", "rollover" and "options" - Development features: "debug" and "packetdrop" If you try this tftp/tftpd implementation, please let me know if it works (or doesn't work) and against which implementaion so I can get a list of confirmed working systems. Author: Edwin Groothuis <edwin@FreeBSD.org>
2010-05-04 06:19:19 +00:00
i = get_field(peer, recvbuffer, size);
if (i >= PATH_MAX) {
tftp_log(LOG_ERR, "Bad option - filename too long");
send_error(peer, EBADOP);
exit(1);
}
*filename = recvbuffer;
tftp_log(LOG_INFO, "Filename: '%s'", *filename);
cp += i;
Go ahead and merge the work edwin@ on tftpd into the tree. It is a lot better than what's in the tree now. Edwin tested it at a prior employer, but can't test it today. I've found that it works a lot better with the various uboot versions that I've used in my embedded work. Here's the pkg-descr from the port that describes the changes: It all started when we got some new routers, which told me the following when trying to upload configuration or download images from it: The TFTP server doesn't support the blocksize option. My curiousity was triggered, it took me some reading of RFCs and other documentation to find out what was possible and what could be done. Was plain TFTP very simple in its handshake, TFTP with options was kind of messy because of its backwards capability: The first packet returned could either be an acknowledgement of options, or the first data packet. Going through the source code of src/libexec/tftpd and going through the code of src/usr.bin/tftp showed that there was a lot of duplicate code, and the addition of options would only increase the amount of duplicate code. After all, both the client and the server can act as a sender and receiver. At the end, it ended up with a nearly complete rewrite of the tftp client and server. It has been tested against the following TFTP clients and servers: - Itself (yay!) - The standard FreeBSD tftp client and server - The Fedora Core 6 tftp client and server - Cisco router tftp client - Extreme Networks tftp client It supports the following RFCs: RFC1350 - THE TFTP PROTOCOL (REVISION 2) RFC2347 - TFTP Option Extension RFC2348 - TFTP Blocksize Option RFC2349 - TFTP Timeout Interval and Transfer Size Options RFC3617 - Uniform Resource Identifier (URI) Scheme and Applicability Statement for the Trivial File Transfer Protocol (TFTP) It supports the following unofficial TFTP Options as described at http://www.compuphase.com/tftp.htm: blksize2 - Block size restricted to powers of 2, excluding protocol headers rollover - Block counter roll-over (roll back to zero or to one) From the tftp program point of view the following things are changed: - New commands: "blocksize", "blocksize2", "rollover" and "options" - Development features: "debug" and "packetdrop" If you try this tftp/tftpd implementation, please let me know if it works (or doesn't work) and against which implementaion so I can get a list of confirmed working systems. Author: Edwin Groothuis <edwin@FreeBSD.org>
2010-05-04 06:19:19 +00:00
i = get_field(peer, cp, size);
*mode = cp;
cp += i;
/* Find the file transfer mode */
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 == NULL) {
tftp_log(LOG_ERR,
"Bad option - Unknown transfer mode (%s)", *mode);
send_error(peer, EBADOP);
exit(1);
}
tftp_log(LOG_INFO, "Mode: '%s'", *mode);
return (cp + 1);
}
1994-05-27 12:39:25 +00:00
/*
Go ahead and merge the work edwin@ on tftpd into the tree. It is a lot better than what's in the tree now. Edwin tested it at a prior employer, but can't test it today. I've found that it works a lot better with the various uboot versions that I've used in my embedded work. Here's the pkg-descr from the port that describes the changes: It all started when we got some new routers, which told me the following when trying to upload configuration or download images from it: The TFTP server doesn't support the blocksize option. My curiousity was triggered, it took me some reading of RFCs and other documentation to find out what was possible and what could be done. Was plain TFTP very simple in its handshake, TFTP with options was kind of messy because of its backwards capability: The first packet returned could either be an acknowledgement of options, or the first data packet. Going through the source code of src/libexec/tftpd and going through the code of src/usr.bin/tftp showed that there was a lot of duplicate code, and the addition of options would only increase the amount of duplicate code. After all, both the client and the server can act as a sender and receiver. At the end, it ended up with a nearly complete rewrite of the tftp client and server. It has been tested against the following TFTP clients and servers: - Itself (yay!) - The standard FreeBSD tftp client and server - The Fedora Core 6 tftp client and server - Cisco router tftp client - Extreme Networks tftp client It supports the following RFCs: RFC1350 - THE TFTP PROTOCOL (REVISION 2) RFC2347 - TFTP Option Extension RFC2348 - TFTP Blocksize Option RFC2349 - TFTP Timeout Interval and Transfer Size Options RFC3617 - Uniform Resource Identifier (URI) Scheme and Applicability Statement for the Trivial File Transfer Protocol (TFTP) It supports the following unofficial TFTP Options as described at http://www.compuphase.com/tftp.htm: blksize2 - Block size restricted to powers of 2, excluding protocol headers rollover - Block counter roll-over (roll back to zero or to one) From the tftp program point of view the following things are changed: - New commands: "blocksize", "blocksize2", "rollover" and "options" - Development features: "debug" and "packetdrop" If you try this tftp/tftpd implementation, please let me know if it works (or doesn't work) and against which implementaion so I can get a list of confirmed working systems. Author: Edwin Groothuis <edwin@FreeBSD.org>
2010-05-04 06:19:19 +00:00
* WRQ - receive a file from the client
1994-05-27 12:39:25 +00:00
*/
void
Go ahead and merge the work edwin@ on tftpd into the tree. It is a lot better than what's in the tree now. Edwin tested it at a prior employer, but can't test it today. I've found that it works a lot better with the various uboot versions that I've used in my embedded work. Here's the pkg-descr from the port that describes the changes: It all started when we got some new routers, which told me the following when trying to upload configuration or download images from it: The TFTP server doesn't support the blocksize option. My curiousity was triggered, it took me some reading of RFCs and other documentation to find out what was possible and what could be done. Was plain TFTP very simple in its handshake, TFTP with options was kind of messy because of its backwards capability: The first packet returned could either be an acknowledgement of options, or the first data packet. Going through the source code of src/libexec/tftpd and going through the code of src/usr.bin/tftp showed that there was a lot of duplicate code, and the addition of options would only increase the amount of duplicate code. After all, both the client and the server can act as a sender and receiver. At the end, it ended up with a nearly complete rewrite of the tftp client and server. It has been tested against the following TFTP clients and servers: - Itself (yay!) - The standard FreeBSD tftp client and server - The Fedora Core 6 tftp client and server - Cisco router tftp client - Extreme Networks tftp client It supports the following RFCs: RFC1350 - THE TFTP PROTOCOL (REVISION 2) RFC2347 - TFTP Option Extension RFC2348 - TFTP Blocksize Option RFC2349 - TFTP Timeout Interval and Transfer Size Options RFC3617 - Uniform Resource Identifier (URI) Scheme and Applicability Statement for the Trivial File Transfer Protocol (TFTP) It supports the following unofficial TFTP Options as described at http://www.compuphase.com/tftp.htm: blksize2 - Block size restricted to powers of 2, excluding protocol headers rollover - Block counter roll-over (roll back to zero or to one) From the tftp program point of view the following things are changed: - New commands: "blocksize", "blocksize2", "rollover" and "options" - Development features: "debug" and "packetdrop" If you try this tftp/tftpd implementation, please let me know if it works (or doesn't work) and against which implementaion so I can get a list of confirmed working systems. Author: Edwin Groothuis <edwin@FreeBSD.org>
2010-05-04 06:19:19 +00:00
tftp_wrq(int peer, char *recvbuffer, ssize_t size)
1994-05-27 12:39:25 +00:00
{
char *cp;
Go ahead and merge the work edwin@ on tftpd into the tree. It is a lot better than what's in the tree now. Edwin tested it at a prior employer, but can't test it today. I've found that it works a lot better with the various uboot versions that I've used in my embedded work. Here's the pkg-descr from the port that describes the changes: It all started when we got some new routers, which told me the following when trying to upload configuration or download images from it: The TFTP server doesn't support the blocksize option. My curiousity was triggered, it took me some reading of RFCs and other documentation to find out what was possible and what could be done. Was plain TFTP very simple in its handshake, TFTP with options was kind of messy because of its backwards capability: The first packet returned could either be an acknowledgement of options, or the first data packet. Going through the source code of src/libexec/tftpd and going through the code of src/usr.bin/tftp showed that there was a lot of duplicate code, and the addition of options would only increase the amount of duplicate code. After all, both the client and the server can act as a sender and receiver. At the end, it ended up with a nearly complete rewrite of the tftp client and server. It has been tested against the following TFTP clients and servers: - Itself (yay!) - The standard FreeBSD tftp client and server - The Fedora Core 6 tftp client and server - Cisco router tftp client - Extreme Networks tftp client It supports the following RFCs: RFC1350 - THE TFTP PROTOCOL (REVISION 2) RFC2347 - TFTP Option Extension RFC2348 - TFTP Blocksize Option RFC2349 - TFTP Timeout Interval and Transfer Size Options RFC3617 - Uniform Resource Identifier (URI) Scheme and Applicability Statement for the Trivial File Transfer Protocol (TFTP) It supports the following unofficial TFTP Options as described at http://www.compuphase.com/tftp.htm: blksize2 - Block size restricted to powers of 2, excluding protocol headers rollover - Block counter roll-over (roll back to zero or to one) From the tftp program point of view the following things are changed: - New commands: "blocksize", "blocksize2", "rollover" and "options" - Development features: "debug" and "packetdrop" If you try this tftp/tftpd implementation, please let me know if it works (or doesn't work) and against which implementaion so I can get a list of confirmed working systems. Author: Edwin Groothuis <edwin@FreeBSD.org>
2010-05-04 06:19:19 +00:00
int has_options = 0, ecode;
char *filename, *mode;
char fnbuf[PATH_MAX];
1994-05-27 12:39:25 +00:00
Go ahead and merge the work edwin@ on tftpd into the tree. It is a lot better than what's in the tree now. Edwin tested it at a prior employer, but can't test it today. I've found that it works a lot better with the various uboot versions that I've used in my embedded work. Here's the pkg-descr from the port that describes the changes: It all started when we got some new routers, which told me the following when trying to upload configuration or download images from it: The TFTP server doesn't support the blocksize option. My curiousity was triggered, it took me some reading of RFCs and other documentation to find out what was possible and what could be done. Was plain TFTP very simple in its handshake, TFTP with options was kind of messy because of its backwards capability: The first packet returned could either be an acknowledgement of options, or the first data packet. Going through the source code of src/libexec/tftpd and going through the code of src/usr.bin/tftp showed that there was a lot of duplicate code, and the addition of options would only increase the amount of duplicate code. After all, both the client and the server can act as a sender and receiver. At the end, it ended up with a nearly complete rewrite of the tftp client and server. It has been tested against the following TFTP clients and servers: - Itself (yay!) - The standard FreeBSD tftp client and server - The Fedora Core 6 tftp client and server - Cisco router tftp client - Extreme Networks tftp client It supports the following RFCs: RFC1350 - THE TFTP PROTOCOL (REVISION 2) RFC2347 - TFTP Option Extension RFC2348 - TFTP Blocksize Option RFC2349 - TFTP Timeout Interval and Transfer Size Options RFC3617 - Uniform Resource Identifier (URI) Scheme and Applicability Statement for the Trivial File Transfer Protocol (TFTP) It supports the following unofficial TFTP Options as described at http://www.compuphase.com/tftp.htm: blksize2 - Block size restricted to powers of 2, excluding protocol headers rollover - Block counter roll-over (roll back to zero or to one) From the tftp program point of view the following things are changed: - New commands: "blocksize", "blocksize2", "rollover" and "options" - Development features: "debug" and "packetdrop" If you try this tftp/tftpd implementation, please let me know if it works (or doesn't work) and against which implementaion so I can get a list of confirmed working systems. Author: Edwin Groothuis <edwin@FreeBSD.org>
2010-05-04 06:19:19 +00:00
cp = parse_header(peer, recvbuffer, size, &filename, &mode);
size -= (cp - recvbuffer) + 1;
strcpy(fnbuf, filename);
reduce_path(fnbuf);
filename = fnbuf;
Go ahead and merge the work edwin@ on tftpd into the tree. It is a lot better than what's in the tree now. Edwin tested it at a prior employer, but can't test it today. I've found that it works a lot better with the various uboot versions that I've used in my embedded work. Here's the pkg-descr from the port that describes the changes: It all started when we got some new routers, which told me the following when trying to upload configuration or download images from it: The TFTP server doesn't support the blocksize option. My curiousity was triggered, it took me some reading of RFCs and other documentation to find out what was possible and what could be done. Was plain TFTP very simple in its handshake, TFTP with options was kind of messy because of its backwards capability: The first packet returned could either be an acknowledgement of options, or the first data packet. Going through the source code of src/libexec/tftpd and going through the code of src/usr.bin/tftp showed that there was a lot of duplicate code, and the addition of options would only increase the amount of duplicate code. After all, both the client and the server can act as a sender and receiver. At the end, it ended up with a nearly complete rewrite of the tftp client and server. It has been tested against the following TFTP clients and servers: - Itself (yay!) - The standard FreeBSD tftp client and server - The Fedora Core 6 tftp client and server - Cisco router tftp client - Extreme Networks tftp client It supports the following RFCs: RFC1350 - THE TFTP PROTOCOL (REVISION 2) RFC2347 - TFTP Option Extension RFC2348 - TFTP Blocksize Option RFC2349 - TFTP Timeout Interval and Transfer Size Options RFC3617 - Uniform Resource Identifier (URI) Scheme and Applicability Statement for the Trivial File Transfer Protocol (TFTP) It supports the following unofficial TFTP Options as described at http://www.compuphase.com/tftp.htm: blksize2 - Block size restricted to powers of 2, excluding protocol headers rollover - Block counter roll-over (roll back to zero or to one) From the tftp program point of view the following things are changed: - New commands: "blocksize", "blocksize2", "rollover" and "options" - Development features: "debug" and "packetdrop" If you try this tftp/tftpd implementation, please let me know if it works (or doesn't work) and against which implementaion so I can get a list of confirmed working systems. Author: Edwin Groothuis <edwin@FreeBSD.org>
2010-05-04 06:19:19 +00:00
if (size > 0) {
if (options_rfc_enabled)
has_options = !parse_options(peer, cp, size);
else
tftp_log(LOG_INFO, "Options found but not enabled");
1994-05-27 12:39:25 +00:00
}
Go ahead and merge the work edwin@ on tftpd into the tree. It is a lot better than what's in the tree now. Edwin tested it at a prior employer, but can't test it today. I've found that it works a lot better with the various uboot versions that I've used in my embedded work. Here's the pkg-descr from the port that describes the changes: It all started when we got some new routers, which told me the following when trying to upload configuration or download images from it: The TFTP server doesn't support the blocksize option. My curiousity was triggered, it took me some reading of RFCs and other documentation to find out what was possible and what could be done. Was plain TFTP very simple in its handshake, TFTP with options was kind of messy because of its backwards capability: The first packet returned could either be an acknowledgement of options, or the first data packet. Going through the source code of src/libexec/tftpd and going through the code of src/usr.bin/tftp showed that there was a lot of duplicate code, and the addition of options would only increase the amount of duplicate code. After all, both the client and the server can act as a sender and receiver. At the end, it ended up with a nearly complete rewrite of the tftp client and server. It has been tested against the following TFTP clients and servers: - Itself (yay!) - The standard FreeBSD tftp client and server - The Fedora Core 6 tftp client and server - Cisco router tftp client - Extreme Networks tftp client It supports the following RFCs: RFC1350 - THE TFTP PROTOCOL (REVISION 2) RFC2347 - TFTP Option Extension RFC2348 - TFTP Blocksize Option RFC2349 - TFTP Timeout Interval and Transfer Size Options RFC3617 - Uniform Resource Identifier (URI) Scheme and Applicability Statement for the Trivial File Transfer Protocol (TFTP) It supports the following unofficial TFTP Options as described at http://www.compuphase.com/tftp.htm: blksize2 - Block size restricted to powers of 2, excluding protocol headers rollover - Block counter roll-over (roll back to zero or to one) From the tftp program point of view the following things are changed: - New commands: "blocksize", "blocksize2", "rollover" and "options" - Development features: "debug" and "packetdrop" If you try this tftp/tftpd implementation, please let me know if it works (or doesn't work) and against which implementaion so I can get a list of confirmed working systems. Author: Edwin Groothuis <edwin@FreeBSD.org>
2010-05-04 06:19:19 +00:00
ecode = validate_access(peer, &filename, WRQ);
if (ecode == 0) {
if (has_options)
send_oack(peer);
else
send_ack(peer, 0);
1994-05-27 12:39:25 +00:00
}
Go ahead and merge the work edwin@ on tftpd into the tree. It is a lot better than what's in the tree now. Edwin tested it at a prior employer, but can't test it today. I've found that it works a lot better with the various uboot versions that I've used in my embedded work. Here's the pkg-descr from the port that describes the changes: It all started when we got some new routers, which told me the following when trying to upload configuration or download images from it: The TFTP server doesn't support the blocksize option. My curiousity was triggered, it took me some reading of RFCs and other documentation to find out what was possible and what could be done. Was plain TFTP very simple in its handshake, TFTP with options was kind of messy because of its backwards capability: The first packet returned could either be an acknowledgement of options, or the first data packet. Going through the source code of src/libexec/tftpd and going through the code of src/usr.bin/tftp showed that there was a lot of duplicate code, and the addition of options would only increase the amount of duplicate code. After all, both the client and the server can act as a sender and receiver. At the end, it ended up with a nearly complete rewrite of the tftp client and server. It has been tested against the following TFTP clients and servers: - Itself (yay!) - The standard FreeBSD tftp client and server - The Fedora Core 6 tftp client and server - Cisco router tftp client - Extreme Networks tftp client It supports the following RFCs: RFC1350 - THE TFTP PROTOCOL (REVISION 2) RFC2347 - TFTP Option Extension RFC2348 - TFTP Blocksize Option RFC2349 - TFTP Timeout Interval and Transfer Size Options RFC3617 - Uniform Resource Identifier (URI) Scheme and Applicability Statement for the Trivial File Transfer Protocol (TFTP) It supports the following unofficial TFTP Options as described at http://www.compuphase.com/tftp.htm: blksize2 - Block size restricted to powers of 2, excluding protocol headers rollover - Block counter roll-over (roll back to zero or to one) From the tftp program point of view the following things are changed: - New commands: "blocksize", "blocksize2", "rollover" and "options" - Development features: "debug" and "packetdrop" If you try this tftp/tftpd implementation, please let me know if it works (or doesn't work) and against which implementaion so I can get a list of confirmed working systems. Author: Edwin Groothuis <edwin@FreeBSD.org>
2010-05-04 06:19:19 +00:00
if (logging) {
tftp_log(LOG_INFO, "%s: write request for %s: %s", peername,
filename, errtomsg(ecode));
}
Go ahead and merge the work edwin@ on tftpd into the tree. It is a lot better than what's in the tree now. Edwin tested it at a prior employer, but can't test it today. I've found that it works a lot better with the various uboot versions that I've used in my embedded work. Here's the pkg-descr from the port that describes the changes: It all started when we got some new routers, which told me the following when trying to upload configuration or download images from it: The TFTP server doesn't support the blocksize option. My curiousity was triggered, it took me some reading of RFCs and other documentation to find out what was possible and what could be done. Was plain TFTP very simple in its handshake, TFTP with options was kind of messy because of its backwards capability: The first packet returned could either be an acknowledgement of options, or the first data packet. Going through the source code of src/libexec/tftpd and going through the code of src/usr.bin/tftp showed that there was a lot of duplicate code, and the addition of options would only increase the amount of duplicate code. After all, both the client and the server can act as a sender and receiver. At the end, it ended up with a nearly complete rewrite of the tftp client and server. It has been tested against the following TFTP clients and servers: - Itself (yay!) - The standard FreeBSD tftp client and server - The Fedora Core 6 tftp client and server - Cisco router tftp client - Extreme Networks tftp client It supports the following RFCs: RFC1350 - THE TFTP PROTOCOL (REVISION 2) RFC2347 - TFTP Option Extension RFC2348 - TFTP Blocksize Option RFC2349 - TFTP Timeout Interval and Transfer Size Options RFC3617 - Uniform Resource Identifier (URI) Scheme and Applicability Statement for the Trivial File Transfer Protocol (TFTP) It supports the following unofficial TFTP Options as described at http://www.compuphase.com/tftp.htm: blksize2 - Block size restricted to powers of 2, excluding protocol headers rollover - Block counter roll-over (roll back to zero or to one) From the tftp program point of view the following things are changed: - New commands: "blocksize", "blocksize2", "rollover" and "options" - Development features: "debug" and "packetdrop" If you try this tftp/tftpd implementation, please let me know if it works (or doesn't work) and against which implementaion so I can get a list of confirmed working systems. Author: Edwin Groothuis <edwin@FreeBSD.org>
2010-05-04 06:19:19 +00:00
tftp_recvfile(peer, mode);
exit(0);
}
/*
* RRQ - send a file to the client
*/
void
tftp_rrq(int peer, char *recvbuffer, ssize_t size)
{
char *cp;
int has_options = 0, ecode;
char *filename, *mode;
char fnbuf[PATH_MAX];
cp = parse_header(peer, recvbuffer, size, &filename, &mode);
size -= (cp - recvbuffer) + 1;
strcpy(fnbuf, filename);
reduce_path(fnbuf);
filename = fnbuf;
if (size > 0) {
if (options_rfc_enabled)
has_options = !parse_options(peer, cp, size);
else
Go ahead and merge the work edwin@ on tftpd into the tree. It is a lot better than what's in the tree now. Edwin tested it at a prior employer, but can't test it today. I've found that it works a lot better with the various uboot versions that I've used in my embedded work. Here's the pkg-descr from the port that describes the changes: It all started when we got some new routers, which told me the following when trying to upload configuration or download images from it: The TFTP server doesn't support the blocksize option. My curiousity was triggered, it took me some reading of RFCs and other documentation to find out what was possible and what could be done. Was plain TFTP very simple in its handshake, TFTP with options was kind of messy because of its backwards capability: The first packet returned could either be an acknowledgement of options, or the first data packet. Going through the source code of src/libexec/tftpd and going through the code of src/usr.bin/tftp showed that there was a lot of duplicate code, and the addition of options would only increase the amount of duplicate code. After all, both the client and the server can act as a sender and receiver. At the end, it ended up with a nearly complete rewrite of the tftp client and server. It has been tested against the following TFTP clients and servers: - Itself (yay!) - The standard FreeBSD tftp client and server - The Fedora Core 6 tftp client and server - Cisco router tftp client - Extreme Networks tftp client It supports the following RFCs: RFC1350 - THE TFTP PROTOCOL (REVISION 2) RFC2347 - TFTP Option Extension RFC2348 - TFTP Blocksize Option RFC2349 - TFTP Timeout Interval and Transfer Size Options RFC3617 - Uniform Resource Identifier (URI) Scheme and Applicability Statement for the Trivial File Transfer Protocol (TFTP) It supports the following unofficial TFTP Options as described at http://www.compuphase.com/tftp.htm: blksize2 - Block size restricted to powers of 2, excluding protocol headers rollover - Block counter roll-over (roll back to zero or to one) From the tftp program point of view the following things are changed: - New commands: "blocksize", "blocksize2", "rollover" and "options" - Development features: "debug" and "packetdrop" If you try this tftp/tftpd implementation, please let me know if it works (or doesn't work) and against which implementaion so I can get a list of confirmed working systems. Author: Edwin Groothuis <edwin@FreeBSD.org>
2010-05-04 06:19:19 +00:00
tftp_log(LOG_INFO, "Options found but not enabled");
}
Go ahead and merge the work edwin@ on tftpd into the tree. It is a lot better than what's in the tree now. Edwin tested it at a prior employer, but can't test it today. I've found that it works a lot better with the various uboot versions that I've used in my embedded work. Here's the pkg-descr from the port that describes the changes: It all started when we got some new routers, which told me the following when trying to upload configuration or download images from it: The TFTP server doesn't support the blocksize option. My curiousity was triggered, it took me some reading of RFCs and other documentation to find out what was possible and what could be done. Was plain TFTP very simple in its handshake, TFTP with options was kind of messy because of its backwards capability: The first packet returned could either be an acknowledgement of options, or the first data packet. Going through the source code of src/libexec/tftpd and going through the code of src/usr.bin/tftp showed that there was a lot of duplicate code, and the addition of options would only increase the amount of duplicate code. After all, both the client and the server can act as a sender and receiver. At the end, it ended up with a nearly complete rewrite of the tftp client and server. It has been tested against the following TFTP clients and servers: - Itself (yay!) - The standard FreeBSD tftp client and server - The Fedora Core 6 tftp client and server - Cisco router tftp client - Extreme Networks tftp client It supports the following RFCs: RFC1350 - THE TFTP PROTOCOL (REVISION 2) RFC2347 - TFTP Option Extension RFC2348 - TFTP Blocksize Option RFC2349 - TFTP Timeout Interval and Transfer Size Options RFC3617 - Uniform Resource Identifier (URI) Scheme and Applicability Statement for the Trivial File Transfer Protocol (TFTP) It supports the following unofficial TFTP Options as described at http://www.compuphase.com/tftp.htm: blksize2 - Block size restricted to powers of 2, excluding protocol headers rollover - Block counter roll-over (roll back to zero or to one) From the tftp program point of view the following things are changed: - New commands: "blocksize", "blocksize2", "rollover" and "options" - Development features: "debug" and "packetdrop" If you try this tftp/tftpd implementation, please let me know if it works (or doesn't work) and against which implementaion so I can get a list of confirmed working systems. Author: Edwin Groothuis <edwin@FreeBSD.org>
2010-05-04 06:19:19 +00:00
ecode = validate_access(peer, &filename, RRQ);
if (ecode == 0) {
if (has_options) {
int n;
char lrecvbuffer[MAXPKTSIZE];
struct tftphdr *rp = (struct tftphdr *)lrecvbuffer;
Go ahead and merge the work edwin@ on tftpd into the tree. It is a lot better than what's in the tree now. Edwin tested it at a prior employer, but can't test it today. I've found that it works a lot better with the various uboot versions that I've used in my embedded work. Here's the pkg-descr from the port that describes the changes: It all started when we got some new routers, which told me the following when trying to upload configuration or download images from it: The TFTP server doesn't support the blocksize option. My curiousity was triggered, it took me some reading of RFCs and other documentation to find out what was possible and what could be done. Was plain TFTP very simple in its handshake, TFTP with options was kind of messy because of its backwards capability: The first packet returned could either be an acknowledgement of options, or the first data packet. Going through the source code of src/libexec/tftpd and going through the code of src/usr.bin/tftp showed that there was a lot of duplicate code, and the addition of options would only increase the amount of duplicate code. After all, both the client and the server can act as a sender and receiver. At the end, it ended up with a nearly complete rewrite of the tftp client and server. It has been tested against the following TFTP clients and servers: - Itself (yay!) - The standard FreeBSD tftp client and server - The Fedora Core 6 tftp client and server - Cisco router tftp client - Extreme Networks tftp client It supports the following RFCs: RFC1350 - THE TFTP PROTOCOL (REVISION 2) RFC2347 - TFTP Option Extension RFC2348 - TFTP Blocksize Option RFC2349 - TFTP Timeout Interval and Transfer Size Options RFC3617 - Uniform Resource Identifier (URI) Scheme and Applicability Statement for the Trivial File Transfer Protocol (TFTP) It supports the following unofficial TFTP Options as described at http://www.compuphase.com/tftp.htm: blksize2 - Block size restricted to powers of 2, excluding protocol headers rollover - Block counter roll-over (roll back to zero or to one) From the tftp program point of view the following things are changed: - New commands: "blocksize", "blocksize2", "rollover" and "options" - Development features: "debug" and "packetdrop" If you try this tftp/tftpd implementation, please let me know if it works (or doesn't work) and against which implementaion so I can get a list of confirmed working systems. Author: Edwin Groothuis <edwin@FreeBSD.org>
2010-05-04 06:19:19 +00:00
send_oack(peer);
n = receive_packet(peer, lrecvbuffer, MAXPKTSIZE,
NULL, timeoutpacket);
if (n < 0) {
if (debug&DEBUG_SIMPLE)
tftp_log(LOG_DEBUG, "Aborting: %s",
rp_strerror(n));
return;
}
if (rp->th_opcode != ACK) {
if (debug&DEBUG_SIMPLE)
tftp_log(LOG_DEBUG,
"Expected ACK, got %s on OACK",
packettype(rp->th_opcode));
return;
}
}
1994-05-27 12:39:25 +00:00
}
Go ahead and merge the work edwin@ on tftpd into the tree. It is a lot better than what's in the tree now. Edwin tested it at a prior employer, but can't test it today. I've found that it works a lot better with the various uboot versions that I've used in my embedded work. Here's the pkg-descr from the port that describes the changes: It all started when we got some new routers, which told me the following when trying to upload configuration or download images from it: The TFTP server doesn't support the blocksize option. My curiousity was triggered, it took me some reading of RFCs and other documentation to find out what was possible and what could be done. Was plain TFTP very simple in its handshake, TFTP with options was kind of messy because of its backwards capability: The first packet returned could either be an acknowledgement of options, or the first data packet. Going through the source code of src/libexec/tftpd and going through the code of src/usr.bin/tftp showed that there was a lot of duplicate code, and the addition of options would only increase the amount of duplicate code. After all, both the client and the server can act as a sender and receiver. At the end, it ended up with a nearly complete rewrite of the tftp client and server. It has been tested against the following TFTP clients and servers: - Itself (yay!) - The standard FreeBSD tftp client and server - The Fedora Core 6 tftp client and server - Cisco router tftp client - Extreme Networks tftp client It supports the following RFCs: RFC1350 - THE TFTP PROTOCOL (REVISION 2) RFC2347 - TFTP Option Extension RFC2348 - TFTP Blocksize Option RFC2349 - TFTP Timeout Interval and Transfer Size Options RFC3617 - Uniform Resource Identifier (URI) Scheme and Applicability Statement for the Trivial File Transfer Protocol (TFTP) It supports the following unofficial TFTP Options as described at http://www.compuphase.com/tftp.htm: blksize2 - Block size restricted to powers of 2, excluding protocol headers rollover - Block counter roll-over (roll back to zero or to one) From the tftp program point of view the following things are changed: - New commands: "blocksize", "blocksize2", "rollover" and "options" - Development features: "debug" and "packetdrop" If you try this tftp/tftpd implementation, please let me know if it works (or doesn't work) and against which implementaion so I can get a list of confirmed working systems. Author: Edwin Groothuis <edwin@FreeBSD.org>
2010-05-04 06:19:19 +00:00
if (logging)
tftp_log(LOG_INFO, "%s: read request for %s: %s", peername,
filename, errtomsg(ecode));
1994-05-27 12:39:25 +00:00
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);
Go ahead and merge the work edwin@ on tftpd into the tree. It is a lot better than what's in the tree now. Edwin tested it at a prior employer, but can't test it today. I've found that it works a lot better with the various uboot versions that I've used in my embedded work. Here's the pkg-descr from the port that describes the changes: It all started when we got some new routers, which told me the following when trying to upload configuration or download images from it: The TFTP server doesn't support the blocksize option. My curiousity was triggered, it took me some reading of RFCs and other documentation to find out what was possible and what could be done. Was plain TFTP very simple in its handshake, TFTP with options was kind of messy because of its backwards capability: The first packet returned could either be an acknowledgement of options, or the first data packet. Going through the source code of src/libexec/tftpd and going through the code of src/usr.bin/tftp showed that there was a lot of duplicate code, and the addition of options would only increase the amount of duplicate code. After all, both the client and the server can act as a sender and receiver. At the end, it ended up with a nearly complete rewrite of the tftp client and server. It has been tested against the following TFTP clients and servers: - Itself (yay!) - The standard FreeBSD tftp client and server - The Fedora Core 6 tftp client and server - Cisco router tftp client - Extreme Networks tftp client It supports the following RFCs: RFC1350 - THE TFTP PROTOCOL (REVISION 2) RFC2347 - TFTP Option Extension RFC2348 - TFTP Blocksize Option RFC2349 - TFTP Timeout Interval and Transfer Size Options RFC3617 - Uniform Resource Identifier (URI) Scheme and Applicability Statement for the Trivial File Transfer Protocol (TFTP) It supports the following unofficial TFTP Options as described at http://www.compuphase.com/tftp.htm: blksize2 - Block size restricted to powers of 2, excluding protocol headers rollover - Block counter roll-over (roll back to zero or to one) From the tftp program point of view the following things are changed: - New commands: "blocksize", "blocksize2", "rollover" and "options" - Development features: "debug" and "packetdrop" If you try this tftp/tftpd implementation, please let me know if it works (or doesn't work) and against which implementaion so I can get a list of confirmed working systems. Author: Edwin Groothuis <edwin@FreeBSD.org>
2010-05-04 06:19:19 +00:00
send_error(peer, ecode);
1994-05-27 12:39:25 +00:00
exit(1);
}
Go ahead and merge the work edwin@ on tftpd into the tree. It is a lot better than what's in the tree now. Edwin tested it at a prior employer, but can't test it today. I've found that it works a lot better with the various uboot versions that I've used in my embedded work. Here's the pkg-descr from the port that describes the changes: It all started when we got some new routers, which told me the following when trying to upload configuration or download images from it: The TFTP server doesn't support the blocksize option. My curiousity was triggered, it took me some reading of RFCs and other documentation to find out what was possible and what could be done. Was plain TFTP very simple in its handshake, TFTP with options was kind of messy because of its backwards capability: The first packet returned could either be an acknowledgement of options, or the first data packet. Going through the source code of src/libexec/tftpd and going through the code of src/usr.bin/tftp showed that there was a lot of duplicate code, and the addition of options would only increase the amount of duplicate code. After all, both the client and the server can act as a sender and receiver. At the end, it ended up with a nearly complete rewrite of the tftp client and server. It has been tested against the following TFTP clients and servers: - Itself (yay!) - The standard FreeBSD tftp client and server - The Fedora Core 6 tftp client and server - Cisco router tftp client - Extreme Networks tftp client It supports the following RFCs: RFC1350 - THE TFTP PROTOCOL (REVISION 2) RFC2347 - TFTP Option Extension RFC2348 - TFTP Blocksize Option RFC2349 - TFTP Timeout Interval and Transfer Size Options RFC3617 - Uniform Resource Identifier (URI) Scheme and Applicability Statement for the Trivial File Transfer Protocol (TFTP) It supports the following unofficial TFTP Options as described at http://www.compuphase.com/tftp.htm: blksize2 - Block size restricted to powers of 2, excluding protocol headers rollover - Block counter roll-over (roll back to zero or to one) From the tftp program point of view the following things are changed: - New commands: "blocksize", "blocksize2", "rollover" and "options" - Development features: "debug" and "packetdrop" If you try this tftp/tftpd implementation, please let me know if it works (or doesn't work) and against which implementaion so I can get a list of confirmed working systems. Author: Edwin Groothuis <edwin@FreeBSD.org>
2010-05-04 06:19:19 +00:00
tftp_xmitfile(peer, mode);
1994-05-27 12:39:25 +00:00
}
/*
* Find the next value for YYYYMMDD.nn when the file to be written should
* be unique. Due to the limitations of nn, we will fail if nn reaches 100.
* Besides, that is four updates per hour on a file, which is kind of
* execessive anyway.
*/
static int
find_next_name(char *filename, int *fd)
{
int i;
time_t tval;
size_t len;
struct tm lt;
char yyyymmdd[MAXPATHLEN];
char newname[MAXPATHLEN];
/* Create the YYYYMMDD part of the filename */
time(&tval);
lt = *localtime(&tval);
len = strftime(yyyymmdd, sizeof(yyyymmdd), newfile_format, &lt);
if (len == 0) {
syslog(LOG_WARNING,
"Filename suffix too long (%d characters maximum)",
MAXPATHLEN);
return (EACCESS);
}
/* Make sure the new filename is not too long */
if (strlen(filename) > MAXPATHLEN - len - 5) {
syslog(LOG_WARNING,
Go ahead and merge the work edwin@ on tftpd into the tree. It is a lot better than what's in the tree now. Edwin tested it at a prior employer, but can't test it today. I've found that it works a lot better with the various uboot versions that I've used in my embedded work. Here's the pkg-descr from the port that describes the changes: It all started when we got some new routers, which told me the following when trying to upload configuration or download images from it: The TFTP server doesn't support the blocksize option. My curiousity was triggered, it took me some reading of RFCs and other documentation to find out what was possible and what could be done. Was plain TFTP very simple in its handshake, TFTP with options was kind of messy because of its backwards capability: The first packet returned could either be an acknowledgement of options, or the first data packet. Going through the source code of src/libexec/tftpd and going through the code of src/usr.bin/tftp showed that there was a lot of duplicate code, and the addition of options would only increase the amount of duplicate code. After all, both the client and the server can act as a sender and receiver. At the end, it ended up with a nearly complete rewrite of the tftp client and server. It has been tested against the following TFTP clients and servers: - Itself (yay!) - The standard FreeBSD tftp client and server - The Fedora Core 6 tftp client and server - Cisco router tftp client - Extreme Networks tftp client It supports the following RFCs: RFC1350 - THE TFTP PROTOCOL (REVISION 2) RFC2347 - TFTP Option Extension RFC2348 - TFTP Blocksize Option RFC2349 - TFTP Timeout Interval and Transfer Size Options RFC3617 - Uniform Resource Identifier (URI) Scheme and Applicability Statement for the Trivial File Transfer Protocol (TFTP) It supports the following unofficial TFTP Options as described at http://www.compuphase.com/tftp.htm: blksize2 - Block size restricted to powers of 2, excluding protocol headers rollover - Block counter roll-over (roll back to zero or to one) From the tftp program point of view the following things are changed: - New commands: "blocksize", "blocksize2", "rollover" and "options" - Development features: "debug" and "packetdrop" If you try this tftp/tftpd implementation, please let me know if it works (or doesn't work) and against which implementaion so I can get a list of confirmed working systems. Author: Edwin Groothuis <edwin@FreeBSD.org>
2010-05-04 06:19:19 +00:00
"Filename too long (%zd characters, %zd maximum)",
strlen(filename), MAXPATHLEN - len - 5);
return (EACCESS);
}
/* Find the first file which doesn't exist */
for (i = 0; i < 100; i++) {
sprintf(newname, "%s.%s.%02d", filename, yyyymmdd, i);
*fd = open(newname,
O_WRONLY | O_CREAT | O_EXCL,
S_IRUSR | S_IWUSR | S_IRGRP |
S_IWGRP | S_IROTH | S_IWOTH);
if (*fd > 0)
return 0;
}
return (EEXIST);
}
1994-05-27 12:39:25 +00:00
/*
* 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
Go ahead and merge the work edwin@ on tftpd into the tree. It is a lot better than what's in the tree now. Edwin tested it at a prior employer, but can't test it today. I've found that it works a lot better with the various uboot versions that I've used in my embedded work. Here's the pkg-descr from the port that describes the changes: It all started when we got some new routers, which told me the following when trying to upload configuration or download images from it: The TFTP server doesn't support the blocksize option. My curiousity was triggered, it took me some reading of RFCs and other documentation to find out what was possible and what could be done. Was plain TFTP very simple in its handshake, TFTP with options was kind of messy because of its backwards capability: The first packet returned could either be an acknowledgement of options, or the first data packet. Going through the source code of src/libexec/tftpd and going through the code of src/usr.bin/tftp showed that there was a lot of duplicate code, and the addition of options would only increase the amount of duplicate code. After all, both the client and the server can act as a sender and receiver. At the end, it ended up with a nearly complete rewrite of the tftp client and server. It has been tested against the following TFTP clients and servers: - Itself (yay!) - The standard FreeBSD tftp client and server - The Fedora Core 6 tftp client and server - Cisco router tftp client - Extreme Networks tftp client It supports the following RFCs: RFC1350 - THE TFTP PROTOCOL (REVISION 2) RFC2347 - TFTP Option Extension RFC2348 - TFTP Blocksize Option RFC2349 - TFTP Timeout Interval and Transfer Size Options RFC3617 - Uniform Resource Identifier (URI) Scheme and Applicability Statement for the Trivial File Transfer Protocol (TFTP) It supports the following unofficial TFTP Options as described at http://www.compuphase.com/tftp.htm: blksize2 - Block size restricted to powers of 2, excluding protocol headers rollover - Block counter roll-over (roll back to zero or to one) From the tftp program point of view the following things are changed: - New commands: "blocksize", "blocksize2", "rollover" and "options" - Development features: "debug" and "packetdrop" If you try this tftp/tftpd implementation, please let me know if it works (or doesn't work) and against which implementaion so I can get a list of confirmed working systems. Author: Edwin Groothuis <edwin@FreeBSD.org>
2010-05-04 06:19:19 +00:00
validate_access(int peer, char **filep, int mode)
1994-05-27 12:39:25 +00:00
{
struct stat stbuf;
int fd;
int error;
1994-05-27 12:39:25 +00:00
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
1994-05-27 12:39:25 +00:00
* restrictions.
*/
if (!strncmp(filename, "../", 3))
1994-05-27 12:39:25 +00:00
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);
1994-05-27 12:39:25 +00:00
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)
*filep = filename = pathname;
else if (mode == RRQ)
1994-05-27 12:39:25 +00:00
return (err);
}
Go ahead and merge the work edwin@ on tftpd into the tree. It is a lot better than what's in the tree now. Edwin tested it at a prior employer, but can't test it today. I've found that it works a lot better with the various uboot versions that I've used in my embedded work. Here's the pkg-descr from the port that describes the changes: It all started when we got some new routers, which told me the following when trying to upload configuration or download images from it: The TFTP server doesn't support the blocksize option. My curiousity was triggered, it took me some reading of RFCs and other documentation to find out what was possible and what could be done. Was plain TFTP very simple in its handshake, TFTP with options was kind of messy because of its backwards capability: The first packet returned could either be an acknowledgement of options, or the first data packet. Going through the source code of src/libexec/tftpd and going through the code of src/usr.bin/tftp showed that there was a lot of duplicate code, and the addition of options would only increase the amount of duplicate code. After all, both the client and the server can act as a sender and receiver. At the end, it ended up with a nearly complete rewrite of the tftp client and server. It has been tested against the following TFTP clients and servers: - Itself (yay!) - The standard FreeBSD tftp client and server - The Fedora Core 6 tftp client and server - Cisco router tftp client - Extreme Networks tftp client It supports the following RFCs: RFC1350 - THE TFTP PROTOCOL (REVISION 2) RFC2347 - TFTP Option Extension RFC2348 - TFTP Blocksize Option RFC2349 - TFTP Timeout Interval and Transfer Size Options RFC3617 - Uniform Resource Identifier (URI) Scheme and Applicability Statement for the Trivial File Transfer Protocol (TFTP) It supports the following unofficial TFTP Options as described at http://www.compuphase.com/tftp.htm: blksize2 - Block size restricted to powers of 2, excluding protocol headers rollover - Block counter roll-over (roll back to zero or to one) From the tftp program point of view the following things are changed: - New commands: "blocksize", "blocksize2", "rollover" and "options" - Development features: "debug" and "packetdrop" If you try this tftp/tftpd implementation, please let me know if it works (or doesn't work) and against which implementaion so I can get a list of confirmed working systems. Author: Edwin Groothuis <edwin@FreeBSD.org>
2010-05-04 06:19:19 +00:00
/*
* This option is handled here because it (might) require(s) the
* size of the file.
*/
option_tsize(peer, NULL, mode, &stbuf);
if (mode == RRQ)
fd = open(filename, O_RDONLY);
else {
if (create_new) {
if (increase_name) {
error = find_next_name(filename, &fd);
if (error > 0)
return (error + 100);
} else
fd = open(filename,
O_WRONLY | O_TRUNC | O_CREAT,
S_IRUSR | S_IWUSR | S_IRGRP |
S_IWGRP | S_IROTH | S_IWOTH );
} else
fd = open(filename, O_WRONLY | O_TRUNC);
}
1994-05-27 12:39:25 +00:00
if (fd < 0)
return (errno + 100);
file = fdopen(fd, (mode == RRQ)? "r":"w");
if (file == NULL) {
close(fd);
return (errno + 100);
1994-05-27 12:39:25 +00:00
}
return (0);
}
Go ahead and merge the work edwin@ on tftpd into the tree. It is a lot better than what's in the tree now. Edwin tested it at a prior employer, but can't test it today. I've found that it works a lot better with the various uboot versions that I've used in my embedded work. Here's the pkg-descr from the port that describes the changes: It all started when we got some new routers, which told me the following when trying to upload configuration or download images from it: The TFTP server doesn't support the blocksize option. My curiousity was triggered, it took me some reading of RFCs and other documentation to find out what was possible and what could be done. Was plain TFTP very simple in its handshake, TFTP with options was kind of messy because of its backwards capability: The first packet returned could either be an acknowledgement of options, or the first data packet. Going through the source code of src/libexec/tftpd and going through the code of src/usr.bin/tftp showed that there was a lot of duplicate code, and the addition of options would only increase the amount of duplicate code. After all, both the client and the server can act as a sender and receiver. At the end, it ended up with a nearly complete rewrite of the tftp client and server. It has been tested against the following TFTP clients and servers: - Itself (yay!) - The standard FreeBSD tftp client and server - The Fedora Core 6 tftp client and server - Cisco router tftp client - Extreme Networks tftp client It supports the following RFCs: RFC1350 - THE TFTP PROTOCOL (REVISION 2) RFC2347 - TFTP Option Extension RFC2348 - TFTP Blocksize Option RFC2349 - TFTP Timeout Interval and Transfer Size Options RFC3617 - Uniform Resource Identifier (URI) Scheme and Applicability Statement for the Trivial File Transfer Protocol (TFTP) It supports the following unofficial TFTP Options as described at http://www.compuphase.com/tftp.htm: blksize2 - Block size restricted to powers of 2, excluding protocol headers rollover - Block counter roll-over (roll back to zero or to one) From the tftp program point of view the following things are changed: - New commands: "blocksize", "blocksize2", "rollover" and "options" - Development features: "debug" and "packetdrop" If you try this tftp/tftpd implementation, please let me know if it works (or doesn't work) and against which implementaion so I can get a list of confirmed working systems. Author: Edwin Groothuis <edwin@FreeBSD.org>
2010-05-04 06:19:19 +00:00
static void
tftp_xmitfile(int peer, const char *mode)
1994-05-27 12:39:25 +00:00
{
Go ahead and merge the work edwin@ on tftpd into the tree. It is a lot better than what's in the tree now. Edwin tested it at a prior employer, but can't test it today. I've found that it works a lot better with the various uboot versions that I've used in my embedded work. Here's the pkg-descr from the port that describes the changes: It all started when we got some new routers, which told me the following when trying to upload configuration or download images from it: The TFTP server doesn't support the blocksize option. My curiousity was triggered, it took me some reading of RFCs and other documentation to find out what was possible and what could be done. Was plain TFTP very simple in its handshake, TFTP with options was kind of messy because of its backwards capability: The first packet returned could either be an acknowledgement of options, or the first data packet. Going through the source code of src/libexec/tftpd and going through the code of src/usr.bin/tftp showed that there was a lot of duplicate code, and the addition of options would only increase the amount of duplicate code. After all, both the client and the server can act as a sender and receiver. At the end, it ended up with a nearly complete rewrite of the tftp client and server. It has been tested against the following TFTP clients and servers: - Itself (yay!) - The standard FreeBSD tftp client and server - The Fedora Core 6 tftp client and server - Cisco router tftp client - Extreme Networks tftp client It supports the following RFCs: RFC1350 - THE TFTP PROTOCOL (REVISION 2) RFC2347 - TFTP Option Extension RFC2348 - TFTP Blocksize Option RFC2349 - TFTP Timeout Interval and Transfer Size Options RFC3617 - Uniform Resource Identifier (URI) Scheme and Applicability Statement for the Trivial File Transfer Protocol (TFTP) It supports the following unofficial TFTP Options as described at http://www.compuphase.com/tftp.htm: blksize2 - Block size restricted to powers of 2, excluding protocol headers rollover - Block counter roll-over (roll back to zero or to one) From the tftp program point of view the following things are changed: - New commands: "blocksize", "blocksize2", "rollover" and "options" - Development features: "debug" and "packetdrop" If you try this tftp/tftpd implementation, please let me know if it works (or doesn't work) and against which implementaion so I can get a list of confirmed working systems. Author: Edwin Groothuis <edwin@FreeBSD.org>
2010-05-04 06:19:19 +00:00
uint16_t block;
time_t now;
struct tftp_stats ts;
1994-05-27 12:39:25 +00:00
Go ahead and merge the work edwin@ on tftpd into the tree. It is a lot better than what's in the tree now. Edwin tested it at a prior employer, but can't test it today. I've found that it works a lot better with the various uboot versions that I've used in my embedded work. Here's the pkg-descr from the port that describes the changes: It all started when we got some new routers, which told me the following when trying to upload configuration or download images from it: The TFTP server doesn't support the blocksize option. My curiousity was triggered, it took me some reading of RFCs and other documentation to find out what was possible and what could be done. Was plain TFTP very simple in its handshake, TFTP with options was kind of messy because of its backwards capability: The first packet returned could either be an acknowledgement of options, or the first data packet. Going through the source code of src/libexec/tftpd and going through the code of src/usr.bin/tftp showed that there was a lot of duplicate code, and the addition of options would only increase the amount of duplicate code. After all, both the client and the server can act as a sender and receiver. At the end, it ended up with a nearly complete rewrite of the tftp client and server. It has been tested against the following TFTP clients and servers: - Itself (yay!) - The standard FreeBSD tftp client and server - The Fedora Core 6 tftp client and server - Cisco router tftp client - Extreme Networks tftp client It supports the following RFCs: RFC1350 - THE TFTP PROTOCOL (REVISION 2) RFC2347 - TFTP Option Extension RFC2348 - TFTP Blocksize Option RFC2349 - TFTP Timeout Interval and Transfer Size Options RFC3617 - Uniform Resource Identifier (URI) Scheme and Applicability Statement for the Trivial File Transfer Protocol (TFTP) It supports the following unofficial TFTP Options as described at http://www.compuphase.com/tftp.htm: blksize2 - Block size restricted to powers of 2, excluding protocol headers rollover - Block counter roll-over (roll back to zero or to one) From the tftp program point of view the following things are changed: - New commands: "blocksize", "blocksize2", "rollover" and "options" - Development features: "debug" and "packetdrop" If you try this tftp/tftpd implementation, please let me know if it works (or doesn't work) and against which implementaion so I can get a list of confirmed working systems. Author: Edwin Groothuis <edwin@FreeBSD.org>
2010-05-04 06:19:19 +00:00
now = time(NULL);
if (debug&DEBUG_SIMPLE)
tftp_log(LOG_DEBUG, "Transmitting file");
1994-05-27 12:39:25 +00:00
Go ahead and merge the work edwin@ on tftpd into the tree. It is a lot better than what's in the tree now. Edwin tested it at a prior employer, but can't test it today. I've found that it works a lot better with the various uboot versions that I've used in my embedded work. Here's the pkg-descr from the port that describes the changes: It all started when we got some new routers, which told me the following when trying to upload configuration or download images from it: The TFTP server doesn't support the blocksize option. My curiousity was triggered, it took me some reading of RFCs and other documentation to find out what was possible and what could be done. Was plain TFTP very simple in its handshake, TFTP with options was kind of messy because of its backwards capability: The first packet returned could either be an acknowledgement of options, or the first data packet. Going through the source code of src/libexec/tftpd and going through the code of src/usr.bin/tftp showed that there was a lot of duplicate code, and the addition of options would only increase the amount of duplicate code. After all, both the client and the server can act as a sender and receiver. At the end, it ended up with a nearly complete rewrite of the tftp client and server. It has been tested against the following TFTP clients and servers: - Itself (yay!) - The standard FreeBSD tftp client and server - The Fedora Core 6 tftp client and server - Cisco router tftp client - Extreme Networks tftp client It supports the following RFCs: RFC1350 - THE TFTP PROTOCOL (REVISION 2) RFC2347 - TFTP Option Extension RFC2348 - TFTP Blocksize Option RFC2349 - TFTP Timeout Interval and Transfer Size Options RFC3617 - Uniform Resource Identifier (URI) Scheme and Applicability Statement for the Trivial File Transfer Protocol (TFTP) It supports the following unofficial TFTP Options as described at http://www.compuphase.com/tftp.htm: blksize2 - Block size restricted to powers of 2, excluding protocol headers rollover - Block counter roll-over (roll back to zero or to one) From the tftp program point of view the following things are changed: - New commands: "blocksize", "blocksize2", "rollover" and "options" - Development features: "debug" and "packetdrop" If you try this tftp/tftpd implementation, please let me know if it works (or doesn't work) and against which implementaion so I can get a list of confirmed working systems. Author: Edwin Groothuis <edwin@FreeBSD.org>
2010-05-04 06:19:19 +00:00
read_init(0, file, mode);
block = 1;
tftp_send(peer, &block, &ts);
read_close();
if (debug&DEBUG_SIMPLE)
tftp_log(LOG_INFO, "Sent %jd bytes in %jd seconds",
(intmax_t)ts.amount, (intmax_t)time(NULL) - now);
1994-05-27 12:39:25 +00:00
}
Go ahead and merge the work edwin@ on tftpd into the tree. It is a lot better than what's in the tree now. Edwin tested it at a prior employer, but can't test it today. I've found that it works a lot better with the various uboot versions that I've used in my embedded work. Here's the pkg-descr from the port that describes the changes: It all started when we got some new routers, which told me the following when trying to upload configuration or download images from it: The TFTP server doesn't support the blocksize option. My curiousity was triggered, it took me some reading of RFCs and other documentation to find out what was possible and what could be done. Was plain TFTP very simple in its handshake, TFTP with options was kind of messy because of its backwards capability: The first packet returned could either be an acknowledgement of options, or the first data packet. Going through the source code of src/libexec/tftpd and going through the code of src/usr.bin/tftp showed that there was a lot of duplicate code, and the addition of options would only increase the amount of duplicate code. After all, both the client and the server can act as a sender and receiver. At the end, it ended up with a nearly complete rewrite of the tftp client and server. It has been tested against the following TFTP clients and servers: - Itself (yay!) - The standard FreeBSD tftp client and server - The Fedora Core 6 tftp client and server - Cisco router tftp client - Extreme Networks tftp client It supports the following RFCs: RFC1350 - THE TFTP PROTOCOL (REVISION 2) RFC2347 - TFTP Option Extension RFC2348 - TFTP Blocksize Option RFC2349 - TFTP Timeout Interval and Transfer Size Options RFC3617 - Uniform Resource Identifier (URI) Scheme and Applicability Statement for the Trivial File Transfer Protocol (TFTP) It supports the following unofficial TFTP Options as described at http://www.compuphase.com/tftp.htm: blksize2 - Block size restricted to powers of 2, excluding protocol headers rollover - Block counter roll-over (roll back to zero or to one) From the tftp program point of view the following things are changed: - New commands: "blocksize", "blocksize2", "rollover" and "options" - Development features: "debug" and "packetdrop" If you try this tftp/tftpd implementation, please let me know if it works (or doesn't work) and against which implementaion so I can get a list of confirmed working systems. Author: Edwin Groothuis <edwin@FreeBSD.org>
2010-05-04 06:19:19 +00:00
static void
tftp_recvfile(int peer, const char *mode)
1994-05-27 12:39:25 +00:00
{
Go ahead and merge the work edwin@ on tftpd into the tree. It is a lot better than what's in the tree now. Edwin tested it at a prior employer, but can't test it today. I've found that it works a lot better with the various uboot versions that I've used in my embedded work. Here's the pkg-descr from the port that describes the changes: It all started when we got some new routers, which told me the following when trying to upload configuration or download images from it: The TFTP server doesn't support the blocksize option. My curiousity was triggered, it took me some reading of RFCs and other documentation to find out what was possible and what could be done. Was plain TFTP very simple in its handshake, TFTP with options was kind of messy because of its backwards capability: The first packet returned could either be an acknowledgement of options, or the first data packet. Going through the source code of src/libexec/tftpd and going through the code of src/usr.bin/tftp showed that there was a lot of duplicate code, and the addition of options would only increase the amount of duplicate code. After all, both the client and the server can act as a sender and receiver. At the end, it ended up with a nearly complete rewrite of the tftp client and server. It has been tested against the following TFTP clients and servers: - Itself (yay!) - The standard FreeBSD tftp client and server - The Fedora Core 6 tftp client and server - Cisco router tftp client - Extreme Networks tftp client It supports the following RFCs: RFC1350 - THE TFTP PROTOCOL (REVISION 2) RFC2347 - TFTP Option Extension RFC2348 - TFTP Blocksize Option RFC2349 - TFTP Timeout Interval and Transfer Size Options RFC3617 - Uniform Resource Identifier (URI) Scheme and Applicability Statement for the Trivial File Transfer Protocol (TFTP) It supports the following unofficial TFTP Options as described at http://www.compuphase.com/tftp.htm: blksize2 - Block size restricted to powers of 2, excluding protocol headers rollover - Block counter roll-over (roll back to zero or to one) From the tftp program point of view the following things are changed: - New commands: "blocksize", "blocksize2", "rollover" and "options" - Development features: "debug" and "packetdrop" If you try this tftp/tftpd implementation, please let me know if it works (or doesn't work) and against which implementaion so I can get a list of confirmed working systems. Author: Edwin Groothuis <edwin@FreeBSD.org>
2010-05-04 06:19:19 +00:00
uint16_t block;
struct timeval now1, now2;
struct tftp_stats ts;
1994-05-27 12:39:25 +00:00
Go ahead and merge the work edwin@ on tftpd into the tree. It is a lot better than what's in the tree now. Edwin tested it at a prior employer, but can't test it today. I've found that it works a lot better with the various uboot versions that I've used in my embedded work. Here's the pkg-descr from the port that describes the changes: It all started when we got some new routers, which told me the following when trying to upload configuration or download images from it: The TFTP server doesn't support the blocksize option. My curiousity was triggered, it took me some reading of RFCs and other documentation to find out what was possible and what could be done. Was plain TFTP very simple in its handshake, TFTP with options was kind of messy because of its backwards capability: The first packet returned could either be an acknowledgement of options, or the first data packet. Going through the source code of src/libexec/tftpd and going through the code of src/usr.bin/tftp showed that there was a lot of duplicate code, and the addition of options would only increase the amount of duplicate code. After all, both the client and the server can act as a sender and receiver. At the end, it ended up with a nearly complete rewrite of the tftp client and server. It has been tested against the following TFTP clients and servers: - Itself (yay!) - The standard FreeBSD tftp client and server - The Fedora Core 6 tftp client and server - Cisco router tftp client - Extreme Networks tftp client It supports the following RFCs: RFC1350 - THE TFTP PROTOCOL (REVISION 2) RFC2347 - TFTP Option Extension RFC2348 - TFTP Blocksize Option RFC2349 - TFTP Timeout Interval and Transfer Size Options RFC3617 - Uniform Resource Identifier (URI) Scheme and Applicability Statement for the Trivial File Transfer Protocol (TFTP) It supports the following unofficial TFTP Options as described at http://www.compuphase.com/tftp.htm: blksize2 - Block size restricted to powers of 2, excluding protocol headers rollover - Block counter roll-over (roll back to zero or to one) From the tftp program point of view the following things are changed: - New commands: "blocksize", "blocksize2", "rollover" and "options" - Development features: "debug" and "packetdrop" If you try this tftp/tftpd implementation, please let me know if it works (or doesn't work) and against which implementaion so I can get a list of confirmed working systems. Author: Edwin Groothuis <edwin@FreeBSD.org>
2010-05-04 06:19:19 +00:00
gettimeofday(&now1, NULL);
if (debug&DEBUG_SIMPLE)
tftp_log(LOG_DEBUG, "Receiving file");
write_init(0, file, mode);
1994-05-27 12:39:25 +00:00
block = 0;
Go ahead and merge the work edwin@ on tftpd into the tree. It is a lot better than what's in the tree now. Edwin tested it at a prior employer, but can't test it today. I've found that it works a lot better with the various uboot versions that I've used in my embedded work. Here's the pkg-descr from the port that describes the changes: It all started when we got some new routers, which told me the following when trying to upload configuration or download images from it: The TFTP server doesn't support the blocksize option. My curiousity was triggered, it took me some reading of RFCs and other documentation to find out what was possible and what could be done. Was plain TFTP very simple in its handshake, TFTP with options was kind of messy because of its backwards capability: The first packet returned could either be an acknowledgement of options, or the first data packet. Going through the source code of src/libexec/tftpd and going through the code of src/usr.bin/tftp showed that there was a lot of duplicate code, and the addition of options would only increase the amount of duplicate code. After all, both the client and the server can act as a sender and receiver. At the end, it ended up with a nearly complete rewrite of the tftp client and server. It has been tested against the following TFTP clients and servers: - Itself (yay!) - The standard FreeBSD tftp client and server - The Fedora Core 6 tftp client and server - Cisco router tftp client - Extreme Networks tftp client It supports the following RFCs: RFC1350 - THE TFTP PROTOCOL (REVISION 2) RFC2347 - TFTP Option Extension RFC2348 - TFTP Blocksize Option RFC2349 - TFTP Timeout Interval and Transfer Size Options RFC3617 - Uniform Resource Identifier (URI) Scheme and Applicability Statement for the Trivial File Transfer Protocol (TFTP) It supports the following unofficial TFTP Options as described at http://www.compuphase.com/tftp.htm: blksize2 - Block size restricted to powers of 2, excluding protocol headers rollover - Block counter roll-over (roll back to zero or to one) From the tftp program point of view the following things are changed: - New commands: "blocksize", "blocksize2", "rollover" and "options" - Development features: "debug" and "packetdrop" If you try this tftp/tftpd implementation, please let me know if it works (or doesn't work) and against which implementaion so I can get a list of confirmed working systems. Author: Edwin Groothuis <edwin@FreeBSD.org>
2010-05-04 06:19:19 +00:00
tftp_receive(peer, &block, &ts, NULL, 0);
1994-05-27 12:39:25 +00:00
Go ahead and merge the work edwin@ on tftpd into the tree. It is a lot better than what's in the tree now. Edwin tested it at a prior employer, but can't test it today. I've found that it works a lot better with the various uboot versions that I've used in my embedded work. Here's the pkg-descr from the port that describes the changes: It all started when we got some new routers, which told me the following when trying to upload configuration or download images from it: The TFTP server doesn't support the blocksize option. My curiousity was triggered, it took me some reading of RFCs and other documentation to find out what was possible and what could be done. Was plain TFTP very simple in its handshake, TFTP with options was kind of messy because of its backwards capability: The first packet returned could either be an acknowledgement of options, or the first data packet. Going through the source code of src/libexec/tftpd and going through the code of src/usr.bin/tftp showed that there was a lot of duplicate code, and the addition of options would only increase the amount of duplicate code. After all, both the client and the server can act as a sender and receiver. At the end, it ended up with a nearly complete rewrite of the tftp client and server. It has been tested against the following TFTP clients and servers: - Itself (yay!) - The standard FreeBSD tftp client and server - The Fedora Core 6 tftp client and server - Cisco router tftp client - Extreme Networks tftp client It supports the following RFCs: RFC1350 - THE TFTP PROTOCOL (REVISION 2) RFC2347 - TFTP Option Extension RFC2348 - TFTP Blocksize Option RFC2349 - TFTP Timeout Interval and Transfer Size Options RFC3617 - Uniform Resource Identifier (URI) Scheme and Applicability Statement for the Trivial File Transfer Protocol (TFTP) It supports the following unofficial TFTP Options as described at http://www.compuphase.com/tftp.htm: blksize2 - Block size restricted to powers of 2, excluding protocol headers rollover - Block counter roll-over (roll back to zero or to one) From the tftp program point of view the following things are changed: - New commands: "blocksize", "blocksize2", "rollover" and "options" - Development features: "debug" and "packetdrop" If you try this tftp/tftpd implementation, please let me know if it works (or doesn't work) and against which implementaion so I can get a list of confirmed working systems. Author: Edwin Groothuis <edwin@FreeBSD.org>
2010-05-04 06:19:19 +00:00
write_close();
2010-09-24 10:40:17 +00:00
gettimeofday(&now2, NULL);
1994-05-27 12:39:25 +00:00
Go ahead and merge the work edwin@ on tftpd into the tree. It is a lot better than what's in the tree now. Edwin tested it at a prior employer, but can't test it today. I've found that it works a lot better with the various uboot versions that I've used in my embedded work. Here's the pkg-descr from the port that describes the changes: It all started when we got some new routers, which told me the following when trying to upload configuration or download images from it: The TFTP server doesn't support the blocksize option. My curiousity was triggered, it took me some reading of RFCs and other documentation to find out what was possible and what could be done. Was plain TFTP very simple in its handshake, TFTP with options was kind of messy because of its backwards capability: The first packet returned could either be an acknowledgement of options, or the first data packet. Going through the source code of src/libexec/tftpd and going through the code of src/usr.bin/tftp showed that there was a lot of duplicate code, and the addition of options would only increase the amount of duplicate code. After all, both the client and the server can act as a sender and receiver. At the end, it ended up with a nearly complete rewrite of the tftp client and server. It has been tested against the following TFTP clients and servers: - Itself (yay!) - The standard FreeBSD tftp client and server - The Fedora Core 6 tftp client and server - Cisco router tftp client - Extreme Networks tftp client It supports the following RFCs: RFC1350 - THE TFTP PROTOCOL (REVISION 2) RFC2347 - TFTP Option Extension RFC2348 - TFTP Blocksize Option RFC2349 - TFTP Timeout Interval and Transfer Size Options RFC3617 - Uniform Resource Identifier (URI) Scheme and Applicability Statement for the Trivial File Transfer Protocol (TFTP) It supports the following unofficial TFTP Options as described at http://www.compuphase.com/tftp.htm: blksize2 - Block size restricted to powers of 2, excluding protocol headers rollover - Block counter roll-over (roll back to zero or to one) From the tftp program point of view the following things are changed: - New commands: "blocksize", "blocksize2", "rollover" and "options" - Development features: "debug" and "packetdrop" If you try this tftp/tftpd implementation, please let me know if it works (or doesn't work) and against which implementaion so I can get a list of confirmed working systems. Author: Edwin Groothuis <edwin@FreeBSD.org>
2010-05-04 06:19:19 +00:00
if (debug&DEBUG_SIMPLE) {
double f;
if (now1.tv_usec > now2.tv_usec) {
now2.tv_usec += 1000000;
now2.tv_sec--;
}
1994-05-27 12:39:25 +00:00
Go ahead and merge the work edwin@ on tftpd into the tree. It is a lot better than what's in the tree now. Edwin tested it at a prior employer, but can't test it today. I've found that it works a lot better with the various uboot versions that I've used in my embedded work. Here's the pkg-descr from the port that describes the changes: It all started when we got some new routers, which told me the following when trying to upload configuration or download images from it: The TFTP server doesn't support the blocksize option. My curiousity was triggered, it took me some reading of RFCs and other documentation to find out what was possible and what could be done. Was plain TFTP very simple in its handshake, TFTP with options was kind of messy because of its backwards capability: The first packet returned could either be an acknowledgement of options, or the first data packet. Going through the source code of src/libexec/tftpd and going through the code of src/usr.bin/tftp showed that there was a lot of duplicate code, and the addition of options would only increase the amount of duplicate code. After all, both the client and the server can act as a sender and receiver. At the end, it ended up with a nearly complete rewrite of the tftp client and server. It has been tested against the following TFTP clients and servers: - Itself (yay!) - The standard FreeBSD tftp client and server - The Fedora Core 6 tftp client and server - Cisco router tftp client - Extreme Networks tftp client It supports the following RFCs: RFC1350 - THE TFTP PROTOCOL (REVISION 2) RFC2347 - TFTP Option Extension RFC2348 - TFTP Blocksize Option RFC2349 - TFTP Timeout Interval and Transfer Size Options RFC3617 - Uniform Resource Identifier (URI) Scheme and Applicability Statement for the Trivial File Transfer Protocol (TFTP) It supports the following unofficial TFTP Options as described at http://www.compuphase.com/tftp.htm: blksize2 - Block size restricted to powers of 2, excluding protocol headers rollover - Block counter roll-over (roll back to zero or to one) From the tftp program point of view the following things are changed: - New commands: "blocksize", "blocksize2", "rollover" and "options" - Development features: "debug" and "packetdrop" If you try this tftp/tftpd implementation, please let me know if it works (or doesn't work) and against which implementaion so I can get a list of confirmed working systems. Author: Edwin Groothuis <edwin@FreeBSD.org>
2010-05-04 06:19:19 +00:00
f = now2.tv_sec - now1.tv_sec +
(now2.tv_usec - now1.tv_usec) / 100000.0;
tftp_log(LOG_INFO,
"Download of %jd bytes in %d blocks completed after %0.1f seconds\n",
(intmax_t)ts.amount, block, f);
1994-05-27 12:39:25 +00:00
}
Go ahead and merge the work edwin@ on tftpd into the tree. It is a lot better than what's in the tree now. Edwin tested it at a prior employer, but can't test it today. I've found that it works a lot better with the various uboot versions that I've used in my embedded work. Here's the pkg-descr from the port that describes the changes: It all started when we got some new routers, which told me the following when trying to upload configuration or download images from it: The TFTP server doesn't support the blocksize option. My curiousity was triggered, it took me some reading of RFCs and other documentation to find out what was possible and what could be done. Was plain TFTP very simple in its handshake, TFTP with options was kind of messy because of its backwards capability: The first packet returned could either be an acknowledgement of options, or the first data packet. Going through the source code of src/libexec/tftpd and going through the code of src/usr.bin/tftp showed that there was a lot of duplicate code, and the addition of options would only increase the amount of duplicate code. After all, both the client and the server can act as a sender and receiver. At the end, it ended up with a nearly complete rewrite of the tftp client and server. It has been tested against the following TFTP clients and servers: - Itself (yay!) - The standard FreeBSD tftp client and server - The Fedora Core 6 tftp client and server - Cisco router tftp client - Extreme Networks tftp client It supports the following RFCs: RFC1350 - THE TFTP PROTOCOL (REVISION 2) RFC2347 - TFTP Option Extension RFC2348 - TFTP Blocksize Option RFC2349 - TFTP Timeout Interval and Transfer Size Options RFC3617 - Uniform Resource Identifier (URI) Scheme and Applicability Statement for the Trivial File Transfer Protocol (TFTP) It supports the following unofficial TFTP Options as described at http://www.compuphase.com/tftp.htm: blksize2 - Block size restricted to powers of 2, excluding protocol headers rollover - Block counter roll-over (roll back to zero or to one) From the tftp program point of view the following things are changed: - New commands: "blocksize", "blocksize2", "rollover" and "options" - Development features: "debug" and "packetdrop" If you try this tftp/tftpd implementation, please let me know if it works (or doesn't work) and against which implementaion so I can get a list of confirmed working systems. Author: Edwin Groothuis <edwin@FreeBSD.org>
2010-05-04 06:19:19 +00:00
return;
}