Re: [Tigervnc-devel] vncviewer password prompt text mode instead of window
On Wed, 4 May 2011 21:57:44 + Cook, Jason jason.c...@wesd.org wrote: This question has been asked before in 2009: Jeff Kowalczyk Thu, 17 Sep 2009 14:20:27 -0700 (Begin quote) Tightvnc-1.3.10 prompted for a password in the terminal from which it was started. When using a -via ssh tunnel, both password prompts conveniently occurred in the same terminal. Now that I have switched to tigervnc-1.0, it raises a password prompt window in X. Can we have an option to use the terminal password prompt? (End quote) I second this request. The current X based prompt is not keyboard accessible. It forces a user to click in the input box before typing the password. If directing the password prompt back to the terminal is not technically feasible, is it possible to have the X input automatically selected without requiring a mouse click? This would provide a similar convenience. The new viewer I'm writing should have better focus behaviour, but asking on stdin/stdout should also be possible. Easiest way is probably by adding another option for this behaviour. I'm afraid I don't have any time to look at this issue now, but please file a bug in the tracker so that this request doesn't get lost. Rgds -- Pierre OssmanOpenSource-based Thin Client Technology System Developer Telephone: +46-13-21 46 00 Cendio ABWeb: http://www.cendio.com A: Because it messes up the order in which people normally read text. Q: Why is top-posting such a bad thing? signature.asc Description: PGP signature -- WhatsUp Gold - Download Free Network Management Software The most intuitive, comprehensive, and cost-effective network management toolset available today. Delivers lowest initial acquisition cost and overall TCO of any competing solution. http://p.sf.net/sfu/whatsupgold-sd___ Tigervnc-devel mailing list Tigervnc-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/tigervnc-devel
Re: [Tigervnc-devel] potential vulnerability in TLS secType?
Hello Brian, thanks for notification about this issue, you are right that password might been sent over network without proper validation of the X.509 certs. Can you please test if attached patch solves this issue? It is Martin's patch with little modification (result is rdr::U32 instead of int). Regards, Adam On 05/05/2011 08:43 AM, Martin Koegler wrote: On Wed, May 04, 2011 at 10:51:06PM -0400, Brian Hinz wrote: I think that I just stumbled onto a possible security vulnerability in CSecurityTLS. It seems as though CSecurityTLS::processMsg returns true before the handshake has completed (possibly due to the if (is.readU8() == 0) test on line 174). In the case of secTypes like x509plain, the user is prompted for a username and password (meaning the client is processing phase 2 of the security stack) before the certificate has been verified. I noticed this while testing a known bad certificate - presumably this means that the username password are sent in the clear since the TLS handshake never completes. Thanks for spotting this. The server should send 0, if he can't start the handshake - so it is still possible to sent the error messages back to the client. The problem is, that a malicus server could send 0 and then authentification succeded. It is caused by the fact, the I reused the common code for returing the error message. A solution can be, that the error fetching code from CConnection::processSecurityResultMsg is inlined in CSecurityTLS::processMsg, like the following (untested): diff --git a/common/rfb/CSecurityTLS.cxx b/common/rfb/CSecurityTLS.cxx index 6028792..17ed93c 100644 --- a/common/rfb/CSecurityTLS.cxx +++ b/common/rfb/CSecurityTLS.cxx @@ -171,8 +171,15 @@ bool CSecurityTLS::processMsg(CConnection* cc) if (!is-checkNoWait(1)) return false; -if (is-readU8() == 0) - return true; +if (is-readU8() == 0) { + int result = is-readU32(); + CharArray reason; + if (result == secResultFailed || result == secResultTooMany) +reason.buf = is-readString(); + else +reason.buf = strDup(Authentication failure (protocol error)); + throw AuthFailureException(reason.buf); +} if (gnutls_init(session, GNUTLS_CLIENT) != GNUTLS_E_SUCCESS) throw AuthFailureException(gnutls_init failed); Index: common/rfb/CSecurityTLS.cxx === --- common/rfb/CSecurityTLS.cxx (revision 4399) +++ common/rfb/CSecurityTLS.cxx (working copy) @@ -171,8 +171,15 @@ if (!is-checkNoWait(1)) return false; -if (is-readU8() == 0) - return true; +if (is-readU8() == 0) { + rdr::U32 result = is-readU32(); + CharArray reason; + if (result == secResultFailed || result == secResultTooMany) +reason.buf = is-readString(); + else +reason.buf = strDup(Authentication failure (protocol error)); + throw AuthFailureException(reason.buf); +} if (gnutls_init(session, GNUTLS_CLIENT) != GNUTLS_E_SUCCESS) throw AuthFailureException(gnutls_init failed); -- WhatsUp Gold - Download Free Network Management Software The most intuitive, comprehensive, and cost-effective network management toolset available today. Delivers lowest initial acquisition cost and overall TCO of any competing solution. http://p.sf.net/sfu/whatsupgold-sd___ Tigervnc-devel mailing list Tigervnc-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/tigervnc-devel
[Tigervnc-devel] [ tigervnc-Feature Request Tracker-3297845 ] vncviewer multiple fullscreen
Feature Request Tracker item #3297845, was opened at 2011-05-05 12:38 Message generated for change (Tracker Item Submitted) made by timtatanka You can respond by visiting: https://sourceforge.net/tracker/?func=detailatid=1126849aid=3297845group_id=254363 Please note that this message will contain a full copy of the comment thread, including the initial issue submission, for this request, not just the latest update. Category: Windows version Group: trunk Status: Open Resolution: None Priority: 5 Private: No Submitted By: tatanka (timtatanka) Assigned to: Peter Åstrand (astrand) Summary: vncviewer multiple fullscreen Initial Comment: An option to let the windows vncviewer be used in fullscreen mode over multiple/all screens. Now the viewer is only able to be in fullscreen mode on 1 of the screens. -- You can respond by visiting: https://sourceforge.net/tracker/?func=detailatid=1126849aid=3297845group_id=254363 -- WhatsUp Gold - Download Free Network Management Software The most intuitive, comprehensive, and cost-effective network management toolset available today. Delivers lowest initial acquisition cost and overall TCO of any competing solution. http://p.sf.net/sfu/whatsupgold-sd ___ Tigervnc-devel mailing list Tigervnc-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/tigervnc-devel
Re: [Tigervnc-devel] potential vulnerability in TLS secType?
Wouldn't this (also untested) work as well, and have the advantage of relying on gnutls to verify that the handshake was completed? diff -Nr -C 6 rfb.unix/CSecurityTLS.cxx.bak rfb.unix/CSecurityTLS.cxx *** rfb.unix/CSecurityTLS.cxx.bak 2011-05-05 06:54:11.018963720 -0400 --- rfb.unix/CSecurityTLS.cxx 2011-05-05 06:55:24.826533250 -0400 *** *** 168,182 initGlobal(); if (!session) { if (!is-checkNoWait(1)) return false; - if (is-readU8() == 0) - return true; - if (gnutls_init(session, GNUTLS_CLIENT) != GNUTLS_E_SUCCESS) throw AuthFailureException(gnutls_init failed); if (gnutls_set_default_priority(session) != GNUTLS_E_SUCCESS) throw AuthFailureException(gnutls_set_default_priority failed); --- 168,179 *** *** 197,208 --- 194,208 if (err != GNUTLS_E_SUCCESS) { vlog.error(TLS Handshake failed: %s\n, gnutls_strerror (err)); shutdown(false); throw AuthFailureException(TLS Handshake failed); } + if (is-readU8() == 0) + return true; + checkSession(); cc-setStreams(fis = new rdr::TLSInStream(is, session), fos = new rdr::TLSOutStream(os, session)); return true; On Thu, May 5, 2011 at 2:43 AM, Martin Koegler mkoeg...@auto.tuwien.ac.atwrote: On Wed, May 04, 2011 at 10:51:06PM -0400, Brian Hinz wrote: I think that I just stumbled onto a possible security vulnerability in CSecurityTLS. It seems as though CSecurityTLS::processMsg returns true before the handshake has completed (possibly due to the if (is.readU8() == 0) test on line 174). In the case of secTypes like x509plain, the user is prompted for a username and password (meaning the client is processing phase 2 of the security stack) before the certificate has been verified. I noticed this while testing a known bad certificate - presumably this means that the username password are sent in the clear since the TLS handshake never completes. Thanks for spotting this. The server should send 0, if he can't start the handshake - so it is still possible to sent the error messages back to the client. The problem is, that a malicus server could send 0 and then authentification succeded. It is caused by the fact, the I reused the common code for returing the error message. A solution can be, that the error fetching code from CConnection::processSecurityResultMsg is inlined in CSecurityTLS::processMsg, like the following (untested): diff --git a/common/rfb/CSecurityTLS.cxx b/common/rfb/CSecurityTLS.cxx index 6028792..17ed93c 100644 --- a/common/rfb/CSecurityTLS.cxx +++ b/common/rfb/CSecurityTLS.cxx @@ -171,8 +171,15 @@ bool CSecurityTLS::processMsg(CConnection* cc) if (!is-checkNoWait(1)) return false; -if (is-readU8() == 0) - return true; +if (is-readU8() == 0) { + int result = is-readU32(); + CharArray reason; + if (result == secResultFailed || result == secResultTooMany) +reason.buf = is-readString(); + else +reason.buf = strDup(Authentication failure (protocol error)); + throw AuthFailureException(reason.buf); +} if (gnutls_init(session, GNUTLS_CLIENT) != GNUTLS_E_SUCCESS) throw AuthFailureException(gnutls_init failed); Regards, Martin Kögler -- WhatsUp Gold - Download Free Network Management Software The most intuitive, comprehensive, and cost-effective network management toolset available today. Delivers lowest initial acquisition cost and overall TCO of any competing solution. http://p.sf.net/sfu/whatsupgold-sd ___ Tigervnc-devel mailing list Tigervnc-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/tigervnc-devel -- WhatsUp Gold - Download Free Network Management Software The most intuitive, comprehensive, and cost-effective network management toolset available today. Delivers lowest initial acquisition cost and overall TCO of any competing solution. http://p.sf.net/sfu/whatsupgold-sd___ Tigervnc-devel mailing list Tigervnc-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/tigervnc-devel
Re: [Tigervnc-devel] potential vulnerability in TLS secType?
Adam, Martin's patch seems to work (my suggestion did not). -brian On Thu, May 5, 2011 at 5:45 AM, Adam Tkac at...@redhat.com wrote: Hello Brian, thanks for notification about this issue, you are right that password might been sent over network without proper validation of the X.509 certs. Can you please test if attached patch solves this issue? It is Martin's patch with little modification (result is rdr::U32 instead of int). Regards, Adam On 05/05/2011 08:43 AM, Martin Koegler wrote: On Wed, May 04, 2011 at 10:51:06PM -0400, Brian Hinz wrote: I think that I just stumbled onto a possible security vulnerability in CSecurityTLS. It seems as though CSecurityTLS::processMsg returns true before the handshake has completed (possibly due to the if (is.readU8() == 0) test on line 174). In the case of secTypes like x509plain, the user is prompted for a username and password (meaning the client is processing phase 2 of the security stack) before the certificate has been verified. I noticed this while testing a known bad certificate - presumably this means that the username password are sent in the clear since the TLS handshake never completes. Thanks for spotting this. The server should send 0, if he can't start the handshake - so it is still possible to sent the error messages back to the client. The problem is, that a malicus server could send 0 and then authentification succeded. It is caused by the fact, the I reused the common code for returing the error message. A solution can be, that the error fetching code from CConnection::processSecurityResultMsg is inlined in CSecurityTLS::processMsg, like the following (untested): diff --git a/common/rfb/CSecurityTLS.cxx b/common/rfb/CSecurityTLS.cxx index 6028792..17ed93c 100644 --- a/common/rfb/CSecurityTLS.cxx +++ b/common/rfb/CSecurityTLS.cxx @@ -171,8 +171,15 @@ bool CSecurityTLS::processMsg(CConnection* cc) if (!is-checkNoWait(1)) return false; -if (is-readU8() == 0) - return true; +if (is-readU8() == 0) { + int result = is-readU32(); + CharArray reason; + if (result == secResultFailed || result == secResultTooMany) +reason.buf = is-readString(); + else +reason.buf = strDup(Authentication failure (protocol error)); + throw AuthFailureException(reason.buf); +} if (gnutls_init(session, GNUTLS_CLIENT) != GNUTLS_E_SUCCESS) throw AuthFailureException(gnutls_init failed); -- WhatsUp Gold - Download Free Network Management Software The most intuitive, comprehensive, and cost-effective network management toolset available today. Delivers lowest initial acquisition cost and overall TCO of any competing solution. http://p.sf.net/sfu/whatsupgold-sd___ Tigervnc-devel mailing list Tigervnc-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/tigervnc-devel
Re: [Tigervnc-devel] potential vulnerability in TLS secType?
On Thu, May 05, 2011 at 07:01:49AM -0400, Brian Hinz wrote: Wouldn't this (also untested) work as well, and have the advantage of relying on gnutls to verify that the handshake was completed? diff -Nr -C 6 rfb.unix/CSecurityTLS.cxx.bak rfb.unix/CSecurityTLS.cxx *** rfb.unix/CSecurityTLS.cxx.bak 2011-05-05 06:54:11.018963720 -0400 --- rfb.unix/CSecurityTLS.cxx 2011-05-05 06:55:24.826533250 -0400 *** *** 168,182 initGlobal(); if (!session) { if (!is-checkNoWait(1)) return false; - if (is-readU8() == 0) - return true; - if (gnutls_init(session, GNUTLS_CLIENT) != GNUTLS_E_SUCCESS) throw AuthFailureException(gnutls_init failed); if (gnutls_set_default_priority(session) != GNUTLS_E_SUCCESS) throw AuthFailureException(gnutls_set_default_priority failed); --- 168,179 *** *** 197,208 --- 194,208 if (err != GNUTLS_E_SUCCESS) { vlog.error(TLS Handshake failed: %s\n, gnutls_strerror (err)); shutdown(false); throw AuthFailureException(TLS Handshake failed); } + if (is-readU8() == 0) + return true; + checkSession(); cc-setStreams(fis = new rdr::TLSInStream(is, session), fos = new rdr::TLSOutStream(os, session)); return true; The readU8 is a flag to signal by the server to the client, that it can't start the handshake and therefore has to sent the error as plaintext back (eg because it can't read the certificate). Your are proposing a different protocol, which don't allow to return an error message before the handshake completed. Regards, Martin K?gler -- WhatsUp Gold - Download Free Network Management Software The most intuitive, comprehensive, and cost-effective network management toolset available today. Delivers lowest initial acquisition cost and overall TCO of any competing solution. http://p.sf.net/sfu/whatsupgold-sd___ Tigervnc-devel mailing list Tigervnc-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/tigervnc-devel
[Tigervnc-devel] Report on TigerVNC autorepeat issue.
I have finally tested DRC's new build. The autorepeat functionality is working now from a windows client to a linux server. It does seem to take a bit longer for the second character to kick in but I have been working without autorepeat in VI for so long now that it is a very welcome fix. Thats for getting this issue resolved. -- Robert Goley FOSS Implementation Specialist Toll Free: (800) 338-4984 Local: (770) 479-7933 Fax: (770) 479-4076 www.openrda.com America's only Free Open Source fund accounting software company. -- WhatsUp Gold - Download Free Network Management Software The most intuitive, comprehensive, and cost-effective network management toolset available today. Delivers lowest initial acquisition cost and overall TCO of any competing solution. http://p.sf.net/sfu/whatsupgold-sd___ Tigervnc-devel mailing list Tigervnc-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/tigervnc-devel
[Tigervnc-devel] [ tigervnc-Bug Tracker-3297994 ] password focus behaviour
Bug Tracker item #3297994, was opened at 2011-05-05 10:31 Message generated for change (Tracker Item Submitted) made by cookja You can respond by visiting: https://sourceforge.net/tracker/?func=detailatid=1126848aid=3297994group_id=254363 Please note that this message will contain a full copy of the comment thread, including the initial issue submission, for this request, not just the latest update. Category: UN*X version Group: trunk Status: Open Resolution: None Priority: 5 Private: No Submitted By: derelict (cookja) Assigned to: Adam Tkac (atkac) Summary: password focus behaviour Initial Comment: The password prompt is not keyboard accessible. It forces a user to click in the input box before typing the password. Correct behavior would be that the input field of the prompt window is automatically selected and that the TAB key toggles between fields. -- You can respond by visiting: https://sourceforge.net/tracker/?func=detailatid=1126848aid=3297994group_id=254363 -- WhatsUp Gold - Download Free Network Management Software The most intuitive, comprehensive, and cost-effective network management toolset available today. Delivers lowest initial acquisition cost and overall TCO of any competing solution. http://p.sf.net/sfu/whatsupgold-sd ___ Tigervnc-devel mailing list Tigervnc-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/tigervnc-devel
[Tigervnc-devel] [ tigervnc-Feature Request Tracker-3297998 ] password prompt on stdin/stdout
Feature Request Tracker item #3297998, was opened at 2011-05-05 10:44 Message generated for change (Tracker Item Submitted) made by cookja You can respond by visiting: https://sourceforge.net/tracker/?func=detailatid=1126849aid=3297998group_id=254363 Please note that this message will contain a full copy of the comment thread, including the initial issue submission, for this request, not just the latest update. Category: UN*X version Group: trunk Status: Open Resolution: None Priority: 5 Private: No Submitted By: derelict (cookja) Assigned to: Adam Tkac (atkac) Summary: password prompt on stdin/stdout Initial Comment: Option to prompt for password on the same terminal from which it was launched. -- You can respond by visiting: https://sourceforge.net/tracker/?func=detailatid=1126849aid=3297998group_id=254363 -- WhatsUp Gold - Download Free Network Management Software The most intuitive, comprehensive, and cost-effective network management toolset available today. Delivers lowest initial acquisition cost and overall TCO of any competing solution. http://p.sf.net/sfu/whatsupgold-sd ___ Tigervnc-devel mailing list Tigervnc-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/tigervnc-devel