- stunnel-users - stunnel.org

Problem with openssl engine
by Nick Sayer 27 Feb '10

27 Feb '10

I'm trying to use a smart card to hold on to the private key for my server. Since access to the smart card is restricted to one process at a time (not sure why it can't be shared, but that's a whole separate discussion), I'd like to use stunnel as the machine-wide SSL proxy. The problem is that with the pkcs11 engine configuration set up, it fails. 2010.02.27 00:15:03 LOG5[91580:0]: imaps accepted connection from 71.141.64.2:32934 2010.02.27 00:15:03 LOG7[91580:0]: SSL state (accept): before/accept initialization 2010.02.27 00:15:03 LOG7[91580:0]: SSL state (accept): SSLv3 read client hello A 2010.02.27 00:15:03 LOG7[91580:0]: SSL state (accept): SSLv3 write server hello A 2010.02.27 00:15:03 LOG7[91580:0]: SSL state (accept): SSLv3 write certificate A 2010.02.27 00:15:03 LOG7[91580:0]: SSL state (accept): SSLv3 write server done A 2010.02.27 00:15:03 LOG7[91580:0]: SSL state (accept): SSLv3 flush data 2010.02.27 00:15:03 LOG7[91580:0]: SSL state (accept): SSLv3 read client key exchange A 2010.02.27 00:15:03 LOG7[91580:0]: SSL alert (write): fatal: bad record mac 2010.02.27 00:15:03 LOG3[91580:0]: SSL_accept: 1408F119: error:1408F119:SSL routines:SSL3_GET_RECORD:decryption failed or bad record mac 2010.02.27 00:15:03 LOG5[91580:0]: Connection reset: 0 bytes sent to SSL, 0 bytes sent to socket 2010.02.27 00:15:03 LOG7[91576:0]: Cleaning up the signal pipe 2010.02.27 00:15:03 LOG7[91576:0]: Process 91580 finished with code 0 (0 left) ^C2010.02.27 00:15:42 LOG3[91576:0]: Received signal 2; terminating This happens *whether or not* I use the private key on the smart card itself. If I load the private key from a disk file, it still happens - even if there is no engineNum statement in the service configuration block. The only way to make it not happen is to comment out the engineCtrl lines. engine=dynamic engineCtrl=SO_PATH:/usr/local/lib/engines/engine_pkcs11.so engineCtrl=ID:pkcs11 engineCtrl=LIST_ADD:1 engineCtrl=LOAD engineCtrl=MODULE_PATH:/usr/local/lib/pkcs11/opensc-pkcs11.so engineCtrl=INIT engineCtrl=PIN:12345678 [imaps] accept = 4993 connect = 143 engineNum = 1 key = id_10 ;cert = id_10 cert = /www.crt It's also worth noting that stunnel seems unable to pull the cert out of the card - that cert=id_10 line should work (works with openssl s_server). This stunnel was compiled last night from the FreeBSD ports tree with WITH_STUNNEL_SSL_ENGINE turned on. stunnel 4.29 on i386-portbld-freebsd8.0 with OpenSSL 0.9.8k 25 Mar 2009 Threading:PTHREAD SSL:ENGINE Sockets:POLL,IPv6 Global options debug = 5 pid = /var/tmp/run/stunnel/stunnel.pid RNDbytes = 64 RNDfile = /dev/urandom RNDoverwrite = yes Service-level options cert = /usr/local/etc/stunnel/stunnel.pem ciphers = AES:ALL:!aNULL:!eNULL:+RC4:@STRENGTH key = /usr/local/etc/stunnel/stunnel.pem session = 300 seconds stack = 65536 bytes sslVersion = SSLv3 for client, all for server TIMEOUTbusy = 300 seconds TIMEOUTclose = 60 seconds TIMEOUTconnect = 10 seconds TIMEOUTidle = 43200 seconds verify = none Note that the same card in the same system can successfully provide ssl server sockets with Java using the Sun PKCS11 JCE provider configured to talk to the opensc PKCS11 module. So there's nothing wrong with the card.

1 0

stunnel-4.15-2.el5.1 zombie when xinetd
by tatiana philippova 25 Feb '10

25 Feb '10

Hi All, I have a problem with stunnel in xinetd mode. It cannot connect to specified localhost:port - always getting child stunnel defunct as a result.. 4789 ? Ss 0:00 \_ stunnel /etc/stunnel/stunnel.conf 4790 ? Z 0:00 \_ [stunnel] <defunct> I'm trying use rsync via stunnel (!ssh for rsync does not suits me in this particular case, but I need encrypt backup traffic anyway.. so have chosen stunnel) client config: cert = /etc/stunnel/stunnel_client_cert.pem client = yes pid = /var/run/stunnel.pid debug = 7 output = /var/log/stunnel.log sslVersion = SSLv2 [ssync] accept = 873 connect = myhost.somewhere.net:273 server config: cert = /etc/stunnel/certs/host.pem key = /etc/stunnel/certs/private/cakey.pem client = no pid = /var/run/stunnel.pid debug = 7 #foreground = yes output = /var/log/stunel.log connect = 873 xinetd config: # default: off # description: SSL wrapper for rsyncd service ssync { disable = no socket_type = stream wait = no user = root server = /usr/sbin/stunnel server_args = /etc/stunnel/stunnel.conf log_on_success += HOST DURATION log_on_failure += HOST only_from = 192.168.5.15 127.0.0.1 } on server side /etc/services: ssync 273/tcp # rsync over stunnel # ps axf | grep rsyn 15637 ? S 0:00 /usr/bin/rsync -4 --daemon --no-detach # netstat -an --program | grep 273 tcp 0 0 0.0.0.0:273 0.0.0.0:* LISTEN 4751/xinetd # netstat -an --program | grep 873 tcp 0 0 127.0.0.1:873 0.0.0.0:* LISTEN 15637/rsync In case of server stunnel running as standalone - everything works perfectly well, but I need it working as xinetd service rather than standalone Could somebody kindly advise what I'm doing wrong? soft on server side: stunnel-4.15-2.el5.1, openssl-0.9.8e-7.el5 OS CentOS release 5.3 (Final), kernel 2.6.18-128.1.6.el5 on client side: stunnel-4.30-1.i386.rpm (compiled as rpmbuild -ta stunnel-4.30.tar.gz) openssl-0.9.8g-9.14.fc9.i686 OS Fedora release 9 (Sulphur) kernel 2.6.28 Many thanks in advance Taphy

1 0

How do you configure Stunnel Ver 4.31 for TLSv1
by David M. Douglass 25 Feb '10

25 Feb '10

I am using the Windows Binary of Version 4.31, and an external email utility that sends SMTP outgoing email. I can configure for specific ports (no problem). and I can send SSMTP SSLv3 with success. That is GOOD ! I am trying to configure Stunnel for sending TLSv1 without success. I understand (via the Version command and screen) that Stunnel defaults to SSLv3. How do I change that to TLSv1 ?? I found the parameter : sslVersion = version select version of SSL protocol Allowed options: all, SSLv2, SSLv3, TLSv1 This appears to be a "SERVICE LEVEL" option, and I have tried to use the line: sslVersion = TLSv1 in the [SSMTP] group of options (along with accept = and connect = ), but get an error message in the log file, which is shown next: 2010.02.25 10:35:44 LOG3[2696:2740]: SSL_connect: 1408F10B: error:1408F10B:SSL routines:SSL3_GET_RECORD:wrong version number I have also tried to put the sslVersion = line in the general configuration area, but generate the same error message. The outbound email does, in fact, fail. Can anyone suggest what I am doing wrong ?? Or offer a solution for configuring Stunnel (V4.31) for TLSv1 ?? David M. Douglass david(a)az-douglass.net Home (480) 839-2629 Cell (602) 908-9092

1 0

How to use stunnel
by Gabriel kamega 24 Feb '10

24 Feb '10

Hello sir, Thank you for my subscription. This is the issue: I want to sent automaticly email via smtp.gmail.com server, And I think I need stunnel program. My questions are : *How to use stunnel ?* ; *How to configure the stunnel.conf file ?* I have dowloaded stunnel.exe version 431, but I can't download the * stunnel.pem* file. on the net I see that I have to install with this command : stunnel -install is that all ? on WXP Professionnel SP2 Regards merci

1 0

Access control/TCP wrappers
by Craig Watkinson 23 Feb '10

23 Feb '10

Hi, We're using stunnel to provide a secure interface to an old server that doesn't support HTTPS natively. I'd like to implement some access control so that connections are only supported from specific IP addresses. I am using v4.27 of stunnel that I downloaded from HPs website, and am running it from inittab to ensure it is always running. Unfortunately I don't think it's compiled with libwrap. Should I see libwrap listed when I run ldd against the binary (see below for output)? I think it's possible to run stunnel from inetd. Could I wrapper it here? Is the following entry correct? stunnel stream tcp nowait root /usr/lbin/tcpd /opt/iexpress/stunnel/bin/stunnel stunnel I think this would work, but I'm concerned that if stunnel was to crash or be killed that there would be nothing restarting it if we ran it from inetd. Any advice much appreciated Craig ------------------------------------- # ./stunnel -version stunnel 4.27 on ia64-hp-hpux11.23 with OpenSSL 0.9.7m 23 Feb 2007 Threading:PTHREAD SSL:ENGINE Sockets:POLL,IPv6 Global options debug = 5 pid = /opt/iexpress/stunnel/var/run/stunnel/stunnel.pid RNDbytes = 64 RNDfile = /dev/urandom RNDoverwrite = yes Service-level options cert = /opt/iexpress/stunnel/etc/stunnel/stunnel.pem ciphers = ALL:!aNULL:!eNULL+RC4:@STRENGTH key = /opt/iexpress/stunnel/etc/stunnel/stunnel.pem session = 300 seconds stack = 65536 bytes sslVersion = SSLv3 for client, all for server TIMEOUTbusy = 300 seconds TIMEOUTclose = 60 seconds TIMEOUTconnect = 10 seconds TIMEOUTidle = 43200 seconds verify = none # ldd ./stunnel libdl.so.1 => /usr/lib/hpux32/libdl.so.1 libnsl.so.1 => /usr/lib/hpux32/libnsl.so.1 libpthread.so.1 => /usr/lib/hpux32/libpthread.so.1 libunwind.so.1 => /usr/lib/hpux32/libunwind.so.1 libc.so.1 => /usr/lib/hpux32/libc.so.1 libxti.so.1 => /usr/lib/hpux32/libxti.so.1 libuca.so.1 => /usr/lib/hpux32/libuca.so.1 libdl.so.1 => /usr/lib/hpux32/libdl.so.1 _________________________________________________________________ Do you have a story that started on Hotmail? Tell us now http://clk.atdmt.com/UKM/go/195013117/direct/01/

2 2

4.26 and commercial certificate - Now working.
by Editor (Kevin) 20 Feb '10

20 Feb '10

Hi: Thanks. Everything is working now. > Message: 1 > Date: Wed, 17 Feb 2010 07:11:16 -0800 (PST) > From: Leandro Avila <leandro.avila(a)ymail.com> > Subject: Re: [stunnel-users] Small challenge with version 4.26 and a > commercial certificate > To: stunnel-users(a)mirt.net > Message-ID: <544640.49520.qm(a)web111011.mail.gq1.yahoo.com> > Content-Type: text/plain; charset=us-ascii > > Hi Kevin, > > I think there is ambiguous information about this. > The man page states that the private key should be unencrypted > However, the changelog states that the support for pass phrases was fixed in v4.20 > > I'll do some testing see what I can find. Or maybe someone can shed some light on the issue. > > Best regards > > ----------------- > Leandro Avila

1 0

Re: [stunnel-users] Make reports an error on Aix with stunnel-4.30
by Michal Trojnara 19 Feb '10

19 Feb '10

Derek Nicholls wrote: > gcc -g -O2 -Wall -Wshadow -Wcast-align -Wpointer-arith > -I/opt/freeware/include -o stunnel file.o client.o log.o options.o > protocol.o network.o resolver.o ssl.o ctx.o verify.o sthreads.o stunnel.o > pty.o libwrap.o -lz -ldl -L/opt/freeware/lib -lssl -lcrypto > ld: 0711-317 ERROR: Undefined symbol: .SSL_CTX_set_info_callback > ld: 0711-317 ERROR: Undefined symbol: .SSL_CTX_sess_set_new_cb > ld: 0711-317 ERROR: Undefined symbol: .SSL_CTX_sess_set_get_cb > ld: 0711-317 ERROR: Undefined symbol: .SSL_CTX_sess_set_remove_cb > ld: 0711-317 ERROR: Undefined symbol: .BIO_set_flags > ld: 0711-345 Use the -bloadmap or -bnoquiet option to obtain more > information. > collect2: ld returned 8 exit status > make: 1254-004 The error code from the last command is 1. What it means is that the linker has found on its path another version of libssl, older than your headers in /opt/freeware/include. Mike

1 0

Cert Chain Question
by Craig Kelley 18 Feb '10

18 Feb '10

I've been attempting to include an intermediate chain for my stunnel setup. First, I previously used an entrust-signed certificate with stunnel just fine, but now I've purchased one from godaddy ($190 for 3 certs for 5 years!). The only problem is that the server has multiple certificates to install. Under Apache, I solved it with this: SSLCertificateFile /etc/httpd/conf/ssl.crt/server.crt SSLCertificateKeyFile /etc/httpd/conf/ssl.key/server.key SSLCertificateChainFile /etc/httpd/conf/ssl.crt/godaddy.crt Which works just fine. With stunnel I attempted this configuration: cert = /etc/stunnel/server.crt key = /etc/stunnel/server.key CAfile = /etc/stunnel/godaddy.crt All those files are identical to the Apache configuration. Stunnel starts up, but clients loudly complain that the certificate is not valid. If I examine the certificate in Thunderbird (I use stunnel for IMAPS and POP3S), it correctly identifies the cert as being from GoDaddy and that it will expire in 2015. But for some reason, the chain to its root server is broken. What am I doing wrong? -- Craig Kelley http://inconnu.islug.org/~ink finger same server for PGP block

2 2

Small challenge with version 4.26 and a commercial certificate
by editor 17 Feb '10

17 Feb '10

Hi: I decided to spend the money and get a commercial certificate from Thawte. It was not a bad price. I have installed in the Sun host and while it works perfectly for Apache but I can not get the file to work for stunnel. The idea is to allow mobile users to access their mail without annoying certificate warnings. This is what happens: # /usr/local/bin/stunnel & [1] 13704 # Enter PEM pass phrase: 2010.02.14 05:32:46 LOG7[13704:1]: Snagged 64 random bytes from /export/home/kgreene/.rnd 2010.02.14 05:32:46 LOG7[13704:1]: Wrote 1024 new random bytes to /export/home/kgreene/.rnd 2010.02.14 05:32:46 LOG7[13704:1]: RAND_status claims sufficient entropy for the PRNG 2010.02.14 05:32:46 LOG7[13704:1]: PRNG seeded successfully 2010.02.14 05:32:46 LOG7[13704:1]: Certificate: /usr/local/etc/stunnel/stunnel.pem 2010.02.14 05:32:46 LOG7[13704:1]: Certificate loaded 2010.02.14 05:32:46 LOG7[13704:1]: Key file: /usr/local/etc/stunnel/stunnel.pem 2010.02.14 05:32:46 LOG3[13704:1]: error stack: 140B3009 : error:140B3009:SSL routines:SSL_CTX_use_RSAPrivateKey_file:PEM lib 2010.02.14 05:32:46 LOG3[13704:1]: error stack: 906A068 : error:0906A068:PEM routines:PEM_do_header:bad password read 2010.02.14 05:32:46 LOG3[13704:1]: SSL_CTX_use_RSAPrivateKey_file: 906406D: error:0906406D:PEM routines:PEM_def_callback:problems getting password [1]+ Exit 1 /usr/local/bin/stunnel # It never pauses to let me enter the PEM pass phase. As in instructed in the man pages, I created the pem file by merging the private key and the certificate from Thawte. This is the version statement: # /usr/local/bin/stunnel -version stunnel 4.26 on sparc-sun-solaris2.9 with OpenSSL 0.9.8l 5 Nov 2009 Threading:PTHREAD SSL:ENGINE Sockets:POLL,IPv6 Auth:LIBWRAP Global options debug = 5 pid = /usr/local/var/run/stunnel/stunnel.pid RNDbytes = 64 RNDfile = /dev/urandom RNDoverwrite = yes Service-level options cert = /usr/local/etc/stunnel/stunnel.pem ciphers = AES:ALL:!aNULL:!eNULL:+RC4:@STRENGTH key = /usr/local/etc/stunnel/stunnel.pem session = 300 seconds stack = 65536 bytes sslVersion = SSLv3 for client, all for server TIMEOUTbusy = 300 seconds TIMEOUTclose = 60 seconds TIMEOUTconnect = 10 seconds TIMEOUTidle = 43200 seconds verify = none ------------------------------- Any ideas would be great. Thanks. Kevin

3 2

nessus crashes stunnel
by Roger Cruz 12 Feb '10

12 Feb '10

We ran Nessus on a Linux system and found that stunnel crashes when the weak certificate suite is enabled. GDB shows the stack trace below and it appears to be in the SSL library. My search on this mailing list led me to this message that appears to report a similar problem http://mirt.net/pipermail/stunnel-users/2008-January/001830.html Am I to understand that the problem is really with OpenSSL and that's what needs to be upgraded? If so, do we know what version will have the fixed problem? I'm going to post in the OpenSSL forum as well, but I wanted to start here since from our perspective, Stunnel is the one crashing. Thank you Roger Cruz 2010.02.05 12:49:11 LOG7[13524:3086718672]: Cleaning up the signal pipe 2010.02.05 12:49:11 LOG6[13524:3086718672]: Child process 13575 finished with code 0 Program received signal SIGSEGV, Segmentation fault. [Switching to Thread 0xb7fa7b90 (LWP 13574)] 0x007b24a7 in krb5_is_referral_realm () from /usr/lib/libkrb5.so.3 (gdb) bt #0 0x007b24a7 in krb5_is_referral_realm () from /usr/lib/libkrb5.so.3 #1 0x0089d338 in kssl_keytab_is_available () from /lib/libssl.so.6 #2 0x008803b1 in ssl3_choose_cipher () from /lib/libssl.so.6 #3 0x0087b2a2 in ssl3_get_client_hello () from /lib/libssl.so.6 #4 0x0087bc85 in ssl3_accept () from /lib/libssl.so.6 #5 0x0089109a in SSL_accept () from /lib/libssl.so.6 #6 0x00884d0d in ssl23_get_client_hello () from /lib/libssl.so.6 #7 0x0088554b in ssl23_accept () from /lib/libssl.so.6 #8 0x0089109a in SSL_accept () from /lib/libssl.so.6 #9 0x003628e2 in ?? () from /usr/sbin/stunnel #10 0x00363acd in ?? () from /usr/sbin/stunnel #11 0x003645ba in ?? () from /usr/sbin/stunnel #12 0x003646a8 in client () from /usr/sbin/stunnel #13 0x0025f45b in start_thread () from /lib/libpthread.so.0 #14 0x00448e5e in clone () from /lib/libc.so.6 (gdb) quit [root@p20xen1 current_hq]# stunnel -version stunnel 4.15 on i686-redhat-linux-gnu with OpenSSL 0.9.8b 04 May 2006 Threading:PTHREAD SSL:ENGINE Sockets:POLL,IPv6 Auth:LIBWRAP Global options debug = 5 pid = /var/run/stunnel.pid RNDbytes = 64 RNDfile = /dev/urandom RNDoverwrite = yes Service-level options cert = /etc/stunnel/stunnel.pem ciphers = ALL:!ADH:+RC4:@STRENGTH key = /etc/stunnel/stunnel.pem session = 300 seconds TIMEOUTbusy = 300 seconds TIMEOUTclose = 60 seconds TIMEOUTconnect = 10 seconds TIMEOUTidle = 43200 seconds verify = none

1 0