stunnel-users August 2006

stunnel-users@stunnel.org

13 participants
16 discussions

stunnel 4.16 released
by Michal Trojnara 01 Sep '06

01 Sep '06

A new version of stunnel has been released. Version 4.16, 2006.08.31, urgency: MEDIUM: * New features sponsored by Hewlett-Packard - A new global option to control engine: engineCtrl = <command>[:<parameter>] - A new service-level option to select engine to read private key: engineNum = <engine number> - OCSP support: ocsp = <URL> * New features - A new option to select version of SSL protocol: sslVersion = all|SSLv2|SSLv3|TLSv1 - Visual Studio vc.mak by David Gillingham <dgillingham(a)gmail.com>. - OS2 support by Paul Smedley (http://smedley.info) * Bugfixes - An ordinary user can install stunnel again. - Compilation problem with --enable-dh fixed. - Some minor compilation warnings fixed. - Service-level CRL cert store implemented. - GPF on protocol negotiations fixed. - Problem detecting addrinfo() on Tru64 fixed. - Default group is now detected by configure script. - Check for maximum number of defined services added. - OpenSSL_add_all_algorithms() added to SSL initialization. - configure script sections reordered to detect pthread library funcions. - RFC 2487 autodetection improved (thx to Hans Werner Strube). High resolution s_poll_wait() not currently supported by UCONTEXT threading. - More precise description of cert directory file names (thx to Muhammad Muquit). * Other changes - Maximum number of services increased from 64 to 256 when poll() is used. Homepage: http://stunnel.mirt.net/ Download: ftp://stunnel.mirt.net/stunnel/ sha1 hash for stunnel-4.16.tar.gz file: 6772e0c7f26c2596564ba66978597db8cd229a72 Best regards, Mike

2 1

Windows simultaneous sockets
by Alex 26 Aug '06

26 Aug '06

Guys, Bit of a newbie question here - but very import to me at least. We are looking to secure a web server that does no currently support SSL. One reason for this is that we want to be able to run the server on multiple hosts and doing so means sharing the sessions etc :-( So - I am interested if anyone has tested the stuffing out of stunnel on windows. IE seeing how well it could handle several hundred sockets simultaneously. Not only from a resource point of view, but from a scaling (n, n**2 etc) point of view. Any input at all will be very welcome. Soon I will be starting some test of my own using JMeter to load test. If anyone is interested in the results I will be happy to post. Cheers AJ -- www.deployview.com www.nerds-central.com www.project-network.com

1 0

stunnel and AJP
by Jones Scott - sjones 25 Aug '06

25 Aug '06

Has anyone used stunnel to secure AJP traffic from apache to jboss *************************************************************************** The information contained in this communication is confidential, is intended only for the use of the recipient named above, and may be legally privileged. If the reader of this message is not the intended recipient, you are hereby notified that any dissemination, distribution or copying of this communication is strictly prohibited. If you have received this communication in error, please resend this communication to the sender and delete the original message or any copy of it from your computer system. Thank You. ****************************************************************************

2 1

stunnel 4-15 through proxy server
by Ingvar Tjøstheim 25 Aug '06

25 Aug '06

Hello I am on a network at work behind a firewall and a proxy server. I can connect to my friends linux computer with putty using proxy configuration and port 443. My friend set up his SSH server on port 443, of course. Yesterday, back home, I installed the following on my Windows2000 computer: Win32OpenSSL-v0.9.8b stunnel ver 4-14 FreeProxy version 3.92 This is the stunnel config file for my home computer: client=no verify=0 [wow] accept = 443 connect = 127.0.0.1:3456 TIMEOUTclose = 0 I set up my router to forward port 443 to this computer. I also registered a DynDNS domain and set up a client to update it. Then, still at home, I prepared another computer which I brought to work today. I installed the following on an XP machine: Win32OpenSSL-v0.9.8b stunnel ver 4-14 (which I upgraded to 4.15 today) SockScap_sc32r238 I set up SockScap to connect to 127.0.0.1 on port 2345 and used it to tunnel a program to the internet through my LAN and this Win2K computer. This was working fine when I was at home. My network at home has no proxy server, and I changed stunnel on my XP computer once at work to version 4-15. Then I added the two protocol lines in the config file as you see below. This is the config file on my XP computer at work: verify = 0 client = yes [wow] accept = 2345 protocol = connect protocolHost = www-proxy.DOMAINNAME.no:80 connect = MYDOMAINNAME.dnsalias.com:443 TIMEOUTclose = 0 Now, the funny thing is that it doesn't work. I used to have a SSH server on my computer at home (which is not running now), and I could connect to it from work. I know the setup could work if I just did things right. I know the FreeProxy and SockScap stuff works, cause it did at home. stunnel and the network is the only difference. Still I get these messages on the XP machine running stunnel as client at work when the SockScap program connects to it: 2006.08.25 10:02:29 LOG5[1280:2940]: stunnel 4.15 on x86-pc-mingw32-gnu with OpenSSL 0.9.7i 14 Oct 2005 2006.08.25 10:02:29 LOG5[1280:2940]: Threading:WIN32 SSL:ENGINE Sockets:SELECT,IPv6 2006.08.25 10:02:29 LOG5[1280:2156]: No limit detected for the number of clients 2006.08.25 10:02:59 LOG5[1280:664]: wow connected from 127.0.0.1:3723 2006.08.25 10:03:05 LOG5[1280:972]: wow connected from 127.0.0.1:3725 2006.08.25 10:03:09 LOG5[1280:664]: Connection reset: 0 bytes sent to SSL, 0 bytes sent to socket 2006.08.25 10:03:15 LOG5[1280:972]: Connection reset: 0 bytes sent to SSL, 0 bytes sent to socket ...and it doesn't work. The program I try using through the tunnel never manages to connect. Anyone have a clue? Thank you very much!

1 0

Re: [stunnel-users] safestring and safename insecure
by Hans Werner Strube 24 Aug '06

24 Aug '06

Michal Trojnara wrote: > Assuming on some systems isprint(x) != isprint(x&0xff): any idea how to > exploit this "insecure" code? > > BTW: At least on Linux and Mac OS X isprint() returns 0 in both cases. The same holds for Solaris. I did not actually encounter an error but just got the idea that something may be insecure here by looking at the lines. After all, isprint etc. are often table-based macros without range-checking the argument. And "man isprint" clearly says that the character is unsigned: (Solaris:) The macro isascii() is defined on all integer values; the rest are defined only where the argument is an int, the value of which is representable as an unsigned char, or EOF, (Linux:) These functions check whether c, which must have the value of an unsigned char or EOF, falls into a certain character class according to the current locale.

1 0

safestring and safename insecure
by Hans Werner Strube 24 Aug '06

24 Aug '06

The macros safestring and safename in common.h are insecure if called for a character with high bit set, because (int)*p will sign-extend the byte on machines where char is signed. Possible patch (here for safestring; for safename correspondingly with isprint -> isalnum): do {char *p; for(p=(s); *p; p++) if(!isprint((int)*(unsigned char *)p)) *p='.';} while(0)

2 1

Re: [stunnel-users] accept is not allowed in inetd mode
by Tomasz Feliksiak 24 Aug '06

24 Aug '06

> I'm afraid I won't be able to tell you what's wrong with your > /etc/stunnel/smtps.conf without your /etc/stunnel/smtps.conf file. > 8-) > I guess you didn't define any service section. > > BTW. Please make sure to read the fine manual. There's also a Polish > version. > http://stunnel.mirt.net/static/stunnel.pl.html > > Greetings from Poland, > Mike You were right. I only define ports but dont uncomment protocol names in [] (like [pop3s]). [pop3s] accept = 995 connect = 110 ;[imaps] ;accept = 993 ;connect = 143 [ssmtp] accept = 465 connect = 25 Now it looks like this and it's working (earlier i have two configs, one per service). Poor thing about this is that it don't log anything before you dont uncomment this [] :P - hope no one other make this mistake. When I uncomment everything else was simple cause I have log :) and know what's going on. There was a problem with binding to ports cause I try to run it by inetd but I find your post at google to not run it by inetd. I change stunnel-inetd-4.15-1 to stunnel-standalone-4.15-1 (PLD Rulez :P) and problems gone. Is possible to run stunnel by inetd??? Thanks for help Michal. Stunnel is great :) Ps. I read the manual, english version. Sad thing is that I dont find out where I make mistake and when I read Polish wersion I do. Good thing is that it will be help for others who make this simple mistake. -- Pozdrawiam Tomasz Feliksiak vonom(a)intertel.net.pl

2 2

Stunnel Crash: Gmail SMTP over TLS
by Shatadal Ghosh 23 Aug '06

23 Aug '06

Hi, I am using stunnel 4.15 on Windows XP SP1 with Mozilla Thunderbird as the e-mail client. I do not have openSSL installed locally. I do have the libssl32.dll and libeay32.dll in the same directory as stunnel.exe (C:\Program Files\stunnel). I am running stunnel as a service. Stunnel crashes when I use it to send e-mail (SMTP via TLS) via gmail. According to the e-mail client configuration page on the gmail website http://mail.google.com/support/bin/answer.py?answer=13287&topic=1555 I tried to use SMTP over TLS to connect to smtp.gmail.com:587 My gmail-smtps block in stunnel.conf was ; SMTP service, listens on localhost:250 [gmail-smtps] protocol=smtp accept=localhost:250 connect=smtp.gmail.com:587 On trying to send an e-mail message via the above configuration stunnel crashed. The log file is as follows 2006.06.14 03:17:15 LOG5[3956:3512]: stunnel 4.15 on x86-pc-mingw32-gnu with OpenSSL 0.9.7i 14 Oct 2005 2006.06.14 03:17:15 LOG5[3956:3512]: Threading:WIN32 SSL:ENGINE Sockets:SELECT,IPv6 2006.06.14 03:17:15 LOG5[3956:3600]: No limit detected for the number of clients 2006.06.14 03:17:15 LOG7[3956:3600]: FD 172 in non-blocking mode 2006.06.14 03:17:15 LOG7[3956:3600]: SO_REUSEADDR option set on accept socket 2006.06.14 03:17:15 LOG7[3956:3600]: gmail-pop3s bound to 127.0.0.1:1100 2006.06.14 03:17:15 LOG7[3956:3600]: FD 180 in non-blocking mode 2006.06.14 03:17:15 LOG7[3956:3600]: SO_REUSEADDR option set on accept socket 2006.06.14 03:17:15 LOG7[3956:3600]: gmail-smtps bound to 127.0.0.1:250 ....# info about other mail connections 2006.06.14 03:17:26 LOG7[3956:3600]: gmail-smtps accepted FD=236 from 127.0.0.1:3665 2006.06.14 03:17:26 LOG7[3956:3600]: Creating a new thread 2006.06.14 03:17:26 LOG7[3956:3600]: New thread created 2006.06.14 03:17:26 LOG7[3956:2812]: gmail-smtps started 2006.06.14 03:17:26 LOG7[3956:2812]: FD 236 in non-blocking mode 2006.06.14 03:17:26 LOG7[3956:2812]: TCP_NODELAY option set on local socket 2006.06.14 03:17:26 LOG5[3956:2812]: gmail-smtps connected from 127.0.0.1:3665 2006.06.14 03:17:26 LOG7[3956:2812]: FD 268 in non-blocking mode 2006.06.14 03:17:26 LOG7[3956:2812]: gmail-smtps connecting 64.233.167.111:587 2006.06.14 03:17:26 LOG7[3956:2812]: connect_wait: waiting 10 seconds 2006.06.14 03:17:26 LOG7[3956:2812]: connect_wait: connected 2006.06.14 03:17:26 LOG7[3956:2812]: Remote FD=268 initialized 2006.06.14 03:17:26 LOG7[3956:2812]: TCP_NODELAY option set on remote socket 2006.06.14 03:17:26 LOG5[3956:2812]: Negotiations for smtp (client side) started 2006.06.14 03:17:26 LOG7[3956:2812]: <- 220 mx.gmail.com ESMTP w66sm450524pyw 2006.06.14 03:17:26 LOG7[3956:2812]: -> 220 mx.gmail.com ESMTP w66sm450524pyw <logfile ends> What mistake am I making? Fortunately it seems that gmail supports SMTP over SSL too (smtp.gmail.com:465) and I am using that. However I am also facing this problem with another account which unfortunately allows SMTP only over TLS. Thanks, Shatadal. --- avast! Antivirus: Outbound message clean. Virus Database (VPS): 0624-1, 06/13/2006 Tested on: 6/14/2006 6:22:04 AM avast! - copyright (c) 2000-2006 ALWIL Software. http://www.avast.com _________________________________________________________________ Is your PC infected? Get a FREE online computer virus scan from McAfee® Security. http://clinic.mcafee.com/clinic/ibuy/campaign.asp?cid=3963

2 1

Re: [stunnel-users] RFC 2487 test in smtp_server not reliable
by Hans Werner Strube 23 Aug '06

23 Aug '06

Michal Trojnara wrote: > On 2006-08-22, at 10:55, Hans Werner Strube wrote: > > Instead of waiting, a better method would > > be to have a finer timeout resolution than 1 sec in > > s_poll_wait(...,...); > > setting the timeout to 1 (sec) would be too embarrassing for TLS > > connections. But this would require major changes in the sources. > > I'm going to do it the hard way and add another parameter to > s_poll_wait() function. > > IMHO the timeout should be between 100ms and 500ms. What do you think? I have not tested the minimum required. However, it should be as short as possible, since for a client using TLS, the full delay will always occur on starting a connection. It should be documented somewhere how this new constant could be configured in compilation.

1 0

RFC 2487 test in smtp_server not reliable
by Hans Werner Strube 23 Aug '06

23 Aug '06

In protocol.c, function smtp_server(), a test for plain SSL or STARTTLS usage (RFC 2487) is based on the availability of data from the client (c->local_rfd.fd). This works well in fast LANs but often fails for WAN connections through slow V90 modems. Especially with new Mozilla-based clients (also KMail and Evolution reported by other users) under Windows XP SP2 the test for data fails almost always, incorrectly resulting in an RFC 2487 handshake with error "Unknown client EHLO". This depends to some extent on the client: Outlook Express mostly works, Mozilla 1.7.x often worked after initial failure, SeaMonkey 1.x always fails. This seems to be a timing problem. I tried inserting a "usleep(50000);" before the switch(s_poll_wait(...)), which made SSL work with SeaMonkey (but often not for the first mail-sending attempt). Unfortunately, in Solaris 7 (where I tested stunnel), usleep is not yet thread-safe, maybe nanosleep(...,...) is preferable. Instead of waiting, a better method would be to have a finer timeout resolution than 1 sec in s_poll_wait(...,...); setting the timeout to 1 (sec) would be too embarrassing for TLS connections. But this would require major changes in the sources.

2 1

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

2004

stunnel-users August 2006