Re: tcnative CVE-2015-4000 (Logjam)
On 15/06/2015 16:29, Arthur Ramsey wrote: > I haven't that would have helped a while back. Did that exist back when > Heartbleed was discovered? No. I put it together when I was trying to do the release builds post heartbleed. Mark > > On 06/15/2015 10:28 AM, Mark Thomas wrote: >> On 15/06/2015 16:22, Arthur Ramsey wrote: >> >> >> Agreed, I probably wouldn't trust some random binary :P, but documentation is fairly lacking on building tcnative for windows x64 and it requires commercial software. >> Have you seen this? >> >> http://wiki.apache.org/tomcat/BuildTcNativeWin >> >> Mark >> >> - >> To unsubscribe, e-mail: users-unsubscr...@tomcat.apache.org >> For additional commands, e-mail: users-h...@tomcat.apache.org >> > - To unsubscribe, e-mail: users-unsubscr...@tomcat.apache.org For additional commands, e-mail: users-h...@tomcat.apache.org
Re: tcnative CVE-2015-4000 (Logjam)
I haven't that would have helped a while back. Did that exist back when Heartbleed was discovered? On 06/15/2015 10:28 AM, Mark Thomas wrote: On 15/06/2015 16:22, Arthur Ramsey wrote: Agreed, I probably wouldn't trust some random binary :P, but documentation is fairly lacking on building tcnative for windows x64 and it requires commercial software. Have you seen this? http://wiki.apache.org/tomcat/BuildTcNativeWin Mark - To unsubscribe, e-mail: users-unsubscr...@tomcat.apache.org For additional commands, e-mail: users-h...@tomcat.apache.org -- Arthur Ramsey Systems Administrator Mediture arthur_ram...@mediture.com 952.400.0323 This e-mail and any attachments may contain CONFIDENTIAL information, including PROTECTED HEALTH INFORMATION. If you are not the intended recipient, any use or disclosure of this information is STRICTLY PROHIBITED; you are requested to delete this e-mail and any attachments, notify the sender immediately, and notify the Mediture Privacy Officer at privacyoffi...@mediture.com. - To unsubscribe, e-mail: users-unsubscr...@tomcat.apache.org For additional commands, e-mail: users-h...@tomcat.apache.org
Re: tcnative CVE-2015-4000 (Logjam)
On 15/06/2015 16:22, Arthur Ramsey wrote: >> Agreed, I probably wouldn't trust some random binary :P, but >> documentation is fairly lacking on building tcnative for windows x64 and >> it requires commercial software. Have you seen this? http://wiki.apache.org/tomcat/BuildTcNativeWin Mark - To unsubscribe, e-mail: users-unsubscr...@tomcat.apache.org For additional commands, e-mail: users-h...@tomcat.apache.org
Re: tcnative CVE-2015-4000 (Logjam)
On 06/15/2015 09:50 AM, Christopher Schultz wrote: -BEGIN PGP SIGNED MESSAGE- Hash: SHA256 Arthur, On 6/13/15 8:42 PM, Arthur Ramsey wrote: I have working binaries for Linux x64 and Windows x64 if anyone needs them. Thanks for offering, but: 1. Anyone running Linux should be able to trivially build their own. True, but there may be other locked into an older version of tomcat like me, which requires some modifications to the latest 1.1.x. Here's the patch. jni/native/include/ssl_private.h 85,87c85 < #define SSL_PROTOCOL_TLSV1_1(1<<3) < #define SSL_PROTOCOL_TLSV1_2(1<<4) < #define SSL_PROTOCOL_ALL (SSL_PROTOCOL_SSLV2|SSL_PROTOCOL_SSLV3|SSL_PROTOCOL_TLSV1|SSL_PROTOCOL_TLSV1_1|SSL_PROTOCOL_TLSV1_2) --- #define SSL_PROTOCOL_ALL (SSL_PROTOCOL_SSLV2|SSL_PROTOCOL_SSLV3|SSL_PROTOCOL_TLSV1) jni/native/src/sslcontext.c 71,89c71 < if (protocol == SSL_PROTOCOL_TLSV1_2) { < #ifdef SSL_OP_NO_TLSv1_2 < if (mode == SSL_MODE_CLIENT) < ctx = SSL_CTX_new(TLSv1_2_client_method()); < else if (mode == SSL_MODE_SERVER) < ctx = SSL_CTX_new(TLSv1_2_server_method()); < else < ctx = SSL_CTX_new(TLSv1_2_method()); < #endif < } else if (protocol == SSL_PROTOCOL_TLSV1_1) { < #ifdef SSL_OP_NO_TLSv1_1 < if (mode == SSL_MODE_CLIENT) < ctx = SSL_CTX_new(TLSv1_1_client_method()); < else if (mode == SSL_MODE_SERVER) < ctx = SSL_CTX_new(TLSv1_1_server_method()); < else < ctx = SSL_CTX_new(TLSv1_1_method()); < #endif < } else if (protocol == SSL_PROTOCOL_TLSV1) { --- if (protocol == SSL_PROTOCOL_TLSV1) { 163,170d144 < #ifdef SSL_OP_NO_TLSv1_1 < if (!(protocol & SSL_PROTOCOL_TLSV1_1)) < SSL_CTX_set_options(c->ctx, SSL_OP_NO_TLSv1_1); < #endif < #ifdef SSL_OP_NO_TLSv1_2 < if (!(protocol & SSL_PROTOCOL_TLSV1_2)) < SSL_CTX_set_options(c->ctx, SSL_OP_NO_TLSv1_2); < #endif 2. Anyone wanting Arthur's binaries should be absolutely sure they trust him. (No offense to Arthur, but accepting binaries from someone on a mailing list is always a big of a red flag.) :) Agreed, I probably wouldn't trust some random binary :P, but documentation is fairly lacking on building tcnative for windows x64 and it requires commercial software. I've built it before and had some trouble building it from the branch. Any estimate on when 1.1.34 binaries will be released? It should still work with newer versions of tomcat 7 providing the SSLProtocol is set to TLSv1? SSLProtocol should be set to "TLSv1+TLSv1.1+TLSv1.2", but on certain older versions of Tomcat 7 this won't work. I can't remember exactly when we implemented this, and then there was a bug in the implementation, etc., so you should just upgrade to 7.0.current which definitely works as documented. Yeah, I use 7.0.55 and "TLSv1+TLSv1.1+TLSv1.2" only enables TLSv1. We should update tomcat, but we are migrating to Wildfly anyway. The Windows binary has SSLv2 and SSLv3 disabled at compile time. Good, though lots and lots of people still need to support SSLv3. - -chris On 6/13/2015 3:30 PM, Arthur Ramsey wrote: Building the latest from svn branch 1.1.x seems to work. I had to do some modifications to get TLSv1.1 and TLSv1.2 when using |SSLProtocol="all" |because I'm using tomcat 7.0.55. Thanks for the help, Arthur On 6/11/2015 3:34 PM, Arthur Ramsey wrote: On 06/11/2015 02:35 PM, Christopher Schultz wrote: Arthur, On 6/11/15 2:14 PM, Arthur Ramsey wrote: Is anyone aware of a way to mitigate the Logjam attack with tomcat 7 and java 7? Disable DHE_EXPORT on the server? I believe I have, but Qualys SSL Server Test still fails me on the Logjam check. SSLCipherSuite="ECDHE-RSA-AES128-GCM-SHA256:ECDHE-ECDSA-AES128-GCM- SHA256:ECDHE-RSA-AES256-GCM-SHA384:ECDHE-ECDSA-AES256-GCM-SHA384:DHE-RSA - -AES128-GCM-SHA256:DHE-DSS-AES128-GCM-SHA256:kEDH+AESGCM:ECDHE-RSA-AES12 8-SHA256:ECDHE-ECDSA-AES128-SHA256:ECDHE-RSA-AES128-SHA:ECDHE-ECDSA-AES1 28-SHA:ECDHE-RSA-AES256-SHA384:ECDHE-ECDSA-AES256-SHA384:ECDHE-RSA-AES25 6-SHA:ECDHE-ECDSA-AES256-SHA:DHE-RSA-AES128-SHA256:DHE-RSA-AES128-SHA:DH E-DSS-AES128-SHA256:DHE-RSA-AES256-SHA256:DHE-DSS-AES256-SHA:DHE-RSA-AES 256-SHA:!aNULL:!eNULL:!EXPORT:!DES:!RC4:!3DES:!MD5:!PSK I use tcnative and openssl-1.0.2a both compiled from source in production today, but I would be open to JSSE too. I believe I need Java 8 to mitigate CVE-2015-4000 with JSSE. Why? See http://stackoverflow.com/questions/30352105/how-to-set-custom-dh-gr oup-in-java-sslengine-to-prevent-logjam-attack I don't see anyway to use a unique 2048-bit or greater DH group with tcnative currently. I believe you are correct; there is a bug in BZ: https://bz.apache.org/bugzilla/show_bug.cgi?id=56108 It looks like 1.1.34 will have this feature. You can build the current trunk of the 1.1 branch and probably be okay. Thanks, I'll give it a try. Scary to use in production, but it may be my be
Re: tcnative CVE-2015-4000 (Logjam)
-BEGIN PGP SIGNED MESSAGE- Hash: SHA256 Arthur, On 6/13/15 8:42 PM, Arthur Ramsey wrote: > I have working binaries for Linux x64 and Windows x64 if anyone > needs them. Thanks for offering, but: 1. Anyone running Linux should be able to trivially build their own. 2. Anyone wanting Arthur's binaries should be absolutely sure they trust him. (No offense to Arthur, but accepting binaries from someone on a mailing list is always a big of a red flag.) :) > It should still work with newer versions of tomcat 7 providing the > SSLProtocol is set to TLSv1? SSLProtocol should be set to "TLSv1+TLSv1.1+TLSv1.2", but on certain older versions of Tomcat 7 this won't work. I can't remember exactly when we implemented this, and then there was a bug in the implementation, etc., so you should just upgrade to 7.0.current which definitely works as documented. > The Windows binary has SSLv2 and SSLv3 disabled at compile time. Good, though lots and lots of people still need to support SSLv3. - -chris > > On 6/13/2015 3:30 PM, Arthur Ramsey wrote: >> Building the latest from svn branch 1.1.x seems to work. I had >> to do some modifications to get TLSv1.1 and TLSv1.2 when using >> |SSLProtocol="all" |because I'm using tomcat 7.0.55. >> >> Thanks for the help, Arthur >> >> On 6/11/2015 3:34 PM, Arthur Ramsey wrote: >>> On 06/11/2015 02:35 PM, Christopher Schultz wrote: > Arthur, > > On 6/11/15 2:14 PM, Arthur Ramsey wrote: >> Is anyone aware of a way to mitigate the Logjam attack >> with tomcat 7 and java 7? > Disable DHE_EXPORT on the server? I believe I have, but Qualys SSL Server Test still fails me on the Logjam check. SSLCipherSuite="ECDHE-RSA-AES128-GCM-SHA256:ECDHE-ECDSA-AES128-GCM- SHA256:ECDHE-RSA-AES256-GCM-SHA384:ECDHE-ECDSA-AES256-GCM-SHA384:DHE-RSA - -AES128-GCM-SHA256:DHE-DSS-AES128-GCM-SHA256:kEDH+AESGCM:ECDHE-RSA-AES12 8-SHA256:ECDHE-ECDSA-AES128-SHA256:ECDHE-RSA-AES128-SHA:ECDHE-ECDSA-AES1 28-SHA:ECDHE-RSA-AES256-SHA384:ECDHE-ECDSA-AES256-SHA384:ECDHE-RSA-AES25 6-SHA:ECDHE-ECDSA-AES256-SHA:DHE-RSA-AES128-SHA256:DHE-RSA-AES128-SHA:DH E-DSS-AES128-SHA256:DHE-RSA-AES256-SHA256:DHE-DSS-AES256-SHA:DHE-RSA-AES 256-SHA:!aNULL:!eNULL:!EXPORT:!DES:!RC4:!3DES:!MD5:!PSK > >> I use tcnative and openssl-1.0.2a both compiled from >> source in production today, but I would be open to JSSE >> too. I believe I need Java 8 to mitigate CVE-2015-4000 >> with JSSE. > Why? See http://stackoverflow.com/questions/30352105/how-to-set-custom-dh-gr oup-in-java-sslengine-to-prevent-logjam-attack > >> I don't see anyway to use a unique 2048-bit or greater DH >> group with tcnative currently. > I believe you are correct; there is a bug in BZ: > https://bz.apache.org/bugzilla/show_bug.cgi?id=56108 > > It looks like 1.1.34 will have this feature. You can build the > current trunk of the 1.1 branch and probably be okay. Thanks, I'll give it a try. Scary to use in production, but it may be my best answer. > >> I'm not sure if there is anything I can do at compile >> time. I'd rather not change the cipher suites as I want >> to maintain browser support. > You should disable EXPORT certificates no matter what. Or were you > talking about the DH parameters? I was talking about DH parameters. > >> My server configuration passed the Qualys SSL Server Test >> with flying colors until Logjam, so I would be worried >> about regressions on other security fixes if I used >> JSSE. > -chris --- - -- To unsubscribe, e-mail: users-unsubscr...@tomcat.apache.org For additional commands, e-mail: users-h...@tomcat.apache.org >>> >>> Thanks, Arthur >> > > > -BEGIN PGP SIGNATURE- Comment: GPGTools - http://gpgtools.org iQIcBAEBCAAGBQJVfuZHAAoJEBzwKT+lPKRYrkAP/0LsRarD3oy3Gu/eT1qRazKq XDZD2ebdpdMnEdaYgNu8xv1T26PerAZSpHStcd+UN+fcjxTeXuIxcRVIHBJ+8Ctu QMZc1Xb3dVUt8f2717m+zd0ACEA2Uzkcl1TDzkrmvSGpEE8/iOltskxnaRxV7nAm ojBp8ksdxfdIrKUviC+lLOqBZfE23BQdL5BA09KqFEPMoMWcDc28aUNAB3Heh05A NEWVgD1WFXV7XLQggv869wF453+vberSpCfDP64UxdwPLpsSH/B0hPfuJg1ap86t g5cUnEmcATIxEtnbkh+kBwjDqvyLzPSG1BUvvZNVUiMCdGjs+WGUX6Indfj4+Fpx nMAMRfbLqR4jJ3HWNkZ70iRzLsmd2XTaVFklgqHGCScjrrqtAQCIq1nGY0ro5MIq dXoJkaQuV0Vw4767r4JuhWewt+OEhSsvrmcMyVrFQtamkVt+Pngr3ufwk+HuVjhB l2+54YVSHAs7wj310vuz8ymI2rQ07AGLEXAHkVKiIB/VoQOVoQjLkMNHTB4Mda0T w+9f9d0op4SOdLhqDyeJd8Hn8cAaEVxk88vFaOo6orFVHGs0BsS1R3PWrx4xj//q IXgJ8Cp0Y1sbZ/8kPtpyLsmg0250BmBj1icRla3EZiVidbVRSlnPQPT+KEAWc8hZ qzgsCk2dvMHi8FFQvK1Y =b2wW -END PGP SIGNATURE- - To unsubscribe, e-mail: users-unsubscr...@tomcat.apache.org For additional commands, e-mail: users-h...@tomcat.apache.org
Re: tcnative CVE-2015-4000 (Logjam)
I have working binaries for Linux x64 and Windows x64 if anyone needs them. It should still work with newer versions of tomcat 7 providing the SSLProtocol is set to TLSv1? The Windows binary has SSLv2 and SSLv3 disabled at compile time. On 6/13/2015 3:30 PM, Arthur Ramsey wrote: Building the latest from svn branch 1.1.x seems to work. I had to do some modifications to get TLSv1.1 and TLSv1.2 when using |SSLProtocol="all" |because I'm using tomcat 7.0.55. Thanks for the help, Arthur On 6/11/2015 3:34 PM, Arthur Ramsey wrote: On 06/11/2015 02:35 PM, Christopher Schultz wrote: -BEGIN PGP SIGNED MESSAGE- Hash: SHA256 Arthur, On 6/11/15 2:14 PM, Arthur Ramsey wrote: Is anyone aware of a way to mitigate the Logjam attack with tomcat 7 and java 7? Disable DHE_EXPORT on the server? I believe I have, but Qualys SSL Server Test still fails me on the Logjam check. SSLCipherSuite="ECDHE-RSA-AES128-GCM-SHA256:ECDHE-ECDSA-AES128-GCM-SHA256:ECDHE-RSA-AES256-GCM-SHA384:ECDHE-ECDSA-AES256-GCM-SHA384:DHE-RSA-AES128-GCM-SHA256:DHE-DSS-AES128-GCM-SHA256:kEDH+AESGCM:ECDHE-RSA-AES128-SHA256:ECDHE-ECDSA-AES128-SHA256:ECDHE-RSA-AES128-SHA:ECDHE-ECDSA-AES128-SHA:ECDHE-RSA-AES256-SHA384:ECDHE-ECDSA-AES256-SHA384:ECDHE-RSA-AES256-SHA:ECDHE-ECDSA-AES256-SHA:DHE-RSA-AES128-SHA256:DHE-RSA-AES128-SHA:DHE-DSS-AES128-SHA256:DHE-RSA-AES256-SHA256:DHE-DSS-AES256-SHA:DHE-RSA-AES256-SHA:!aNULL:!eNULL:!EXPORT:!DES:!RC4:!3DES:!MD5:!PSK I use tcnative and openssl-1.0.2a both compiled from source in production today, but I would be open to JSSE too. I believe I need Java 8 to mitigate CVE-2015-4000 with JSSE. Why? See http://stackoverflow.com/questions/30352105/how-to-set-custom-dh-group-in-java-sslengine-to-prevent-logjam-attack I don't see anyway to use a unique 2048-bit or greater DH group with tcnative currently. I believe you are correct; there is a bug in BZ: https://bz.apache.org/bugzilla/show_bug.cgi?id=56108 It looks like 1.1.34 will have this feature. You can build the current trunk of the 1.1 branch and probably be okay. Thanks, I'll give it a try. Scary to use in production, but it may be my best answer. I'm not sure if there is anything I can do at compile time. I'd rather not change the cipher suites as I want to maintain browser support. You should disable EXPORT certificates no matter what. Or were you talking about the DH parameters? I was talking about DH parameters. My server configuration passed the Qualys SSL Server Test with flying colors until Logjam, so I would be worried about regressions on other security fixes if I used JSSE. - -chris -BEGIN PGP SIGNATURE- Comment: GPGTools - http://gpgtools.org iQIcBAEBCAAGBQJVeeL0AAoJEBzwKT+lPKRYkJQQAIyplWF0F65zvlzuQTrsFYPC Ioh+w4ddwalB1OFzaxGjnulwN9eO91iudqFiZyFpZnh7jV8GOJCQVO5whbBIXvQm l4RUispklWXNh2ClFfkW2YoXwfZPBhk4um5oVo2KHN7wf3F9AhvA/oz3Ecm2WUdg lQ7q4+WapZknWS6YdxwMzG7Jl7k6gGgnhfe6SmtEYMDKE8ktTcyAjpHX+NhXXC+e iCiZ0+DH1lYmUIHdVJu2FIgdui0CVecArJ9ufniiIpbYOnjWFxu+IZGlBuTgoAHg 8Lu7koGDOOnagSdJ6DNJeEyniRVPA61zcKRIqB1IWJJzgZVIpo8/wF4r9jGFIH3b x+3cqqSiDLppHar48ENIGbqYRCwybRCiJu3SvKLJ/zRs51ybxKbSXOondPWqIRD/ rbLQN6Z/2nQUeSp7A7iKGQj1CqFSDp5IFBqwvP4A9xWFbqCbwOWUfKhgM8UrToLN DRbtjdpGZvA0lJqxmR9nKWn9K9nNRcViI2wlcDOB22RFjz2S+fUToylf8utUJbW0 MJ5GdqnPYMp3r0NajnWaY8z1POneaqnLHnW5xnhLA2UgDBoClUA2Xe/UmU+ngUT3 OOJDb52+Xr3V+JvsDuK6cgoHTM7X+2i3+75acigwMyPYO34hA1uanVhx7XTvheqA XkCixeOIXgynHCDcWYDc =Lycq -END PGP SIGNATURE- - To unsubscribe, e-mail: users-unsubscr...@tomcat.apache.org For additional commands, e-mail: users-h...@tomcat.apache.org Thanks, Arthur
Re: tcnative CVE-2015-4000 (Logjam)
Building the latest from svn branch 1.1.x seems to work. I had to do some modifications to get TLSv1.1 and TLSv1.2 when using |SSLProtocol="all" |because I'm using tomcat 7.0.55. Thanks for the help, Arthur On 6/11/2015 3:34 PM, Arthur Ramsey wrote: On 06/11/2015 02:35 PM, Christopher Schultz wrote: -BEGIN PGP SIGNED MESSAGE- Hash: SHA256 Arthur, On 6/11/15 2:14 PM, Arthur Ramsey wrote: Is anyone aware of a way to mitigate the Logjam attack with tomcat 7 and java 7? Disable DHE_EXPORT on the server? I believe I have, but Qualys SSL Server Test still fails me on the Logjam check. SSLCipherSuite="ECDHE-RSA-AES128-GCM-SHA256:ECDHE-ECDSA-AES128-GCM-SHA256:ECDHE-RSA-AES256-GCM-SHA384:ECDHE-ECDSA-AES256-GCM-SHA384:DHE-RSA-AES128-GCM-SHA256:DHE-DSS-AES128-GCM-SHA256:kEDH+AESGCM:ECDHE-RSA-AES128-SHA256:ECDHE-ECDSA-AES128-SHA256:ECDHE-RSA-AES128-SHA:ECDHE-ECDSA-AES128-SHA:ECDHE-RSA-AES256-SHA384:ECDHE-ECDSA-AES256-SHA384:ECDHE-RSA-AES256-SHA:ECDHE-ECDSA-AES256-SHA:DHE-RSA-AES128-SHA256:DHE-RSA-AES128-SHA:DHE-DSS-AES128-SHA256:DHE-RSA-AES256-SHA256:DHE-DSS-AES256-SHA:DHE-RSA-AES256-SHA:!aNULL:!eNULL:!EXPORT:!DES:!RC4:!3DES:!MD5:!PSK I use tcnative and openssl-1.0.2a both compiled from source in production today, but I would be open to JSSE too. I believe I need Java 8 to mitigate CVE-2015-4000 with JSSE. Why? See http://stackoverflow.com/questions/30352105/how-to-set-custom-dh-group-in-java-sslengine-to-prevent-logjam-attack I don't see anyway to use a unique 2048-bit or greater DH group with tcnative currently. I believe you are correct; there is a bug in BZ: https://bz.apache.org/bugzilla/show_bug.cgi?id=56108 It looks like 1.1.34 will have this feature. You can build the current trunk of the 1.1 branch and probably be okay. Thanks, I'll give it a try. Scary to use in production, but it may be my best answer. I'm not sure if there is anything I can do at compile time. I'd rather not change the cipher suites as I want to maintain browser support. You should disable EXPORT certificates no matter what. Or were you talking about the DH parameters? I was talking about DH parameters. My server configuration passed the Qualys SSL Server Test with flying colors until Logjam, so I would be worried about regressions on other security fixes if I used JSSE. - -chris -BEGIN PGP SIGNATURE- Comment: GPGTools - http://gpgtools.org iQIcBAEBCAAGBQJVeeL0AAoJEBzwKT+lPKRYkJQQAIyplWF0F65zvlzuQTrsFYPC Ioh+w4ddwalB1OFzaxGjnulwN9eO91iudqFiZyFpZnh7jV8GOJCQVO5whbBIXvQm l4RUispklWXNh2ClFfkW2YoXwfZPBhk4um5oVo2KHN7wf3F9AhvA/oz3Ecm2WUdg lQ7q4+WapZknWS6YdxwMzG7Jl7k6gGgnhfe6SmtEYMDKE8ktTcyAjpHX+NhXXC+e iCiZ0+DH1lYmUIHdVJu2FIgdui0CVecArJ9ufniiIpbYOnjWFxu+IZGlBuTgoAHg 8Lu7koGDOOnagSdJ6DNJeEyniRVPA61zcKRIqB1IWJJzgZVIpo8/wF4r9jGFIH3b x+3cqqSiDLppHar48ENIGbqYRCwybRCiJu3SvKLJ/zRs51ybxKbSXOondPWqIRD/ rbLQN6Z/2nQUeSp7A7iKGQj1CqFSDp5IFBqwvP4A9xWFbqCbwOWUfKhgM8UrToLN DRbtjdpGZvA0lJqxmR9nKWn9K9nNRcViI2wlcDOB22RFjz2S+fUToylf8utUJbW0 MJ5GdqnPYMp3r0NajnWaY8z1POneaqnLHnW5xnhLA2UgDBoClUA2Xe/UmU+ngUT3 OOJDb52+Xr3V+JvsDuK6cgoHTM7X+2i3+75acigwMyPYO34hA1uanVhx7XTvheqA XkCixeOIXgynHCDcWYDc =Lycq -END PGP SIGNATURE- - To unsubscribe, e-mail: users-unsubscr...@tomcat.apache.org For additional commands, e-mail: users-h...@tomcat.apache.org Thanks, Arthur
Re: tcnative CVE-2015-4000 (Logjam)
-BEGIN PGP SIGNED MESSAGE- Hash: SHA256 Rainer, On 6/12/15 6:32 AM, Rainer Jung wrote: > With existing 1.1.33 you can choose your cipher suite, so that > non-DHE ciphers come first and set SSLHonorCipherOrder such that > the client chooses the first matching cipher and DHE will likely > not be used, only by client who do not support a cipher to the left > of DHE in your cipher list. A slight correction: the *server* chooses the cipher suite to be used, not the client. > Note that old Java versions as clients (6, maybe 7 depending on > patch level?) have a problem with DHE keys longer than 768 or 1024 > bits (depending on JVM details). So by mitigating Logjam you might > run into compatibility issues with those. +1 > It would be interesting to know, what details SSLLabs tell you, > e.g. if they say you are vulnerable to the export downgrade attack > (really bad), or "just" to your DH params should be longer. You can > use the OpenSSL commandline client in version 1.0.2 to check, what > param length a handshake results in: > > openssl s_client -connect www.example.com:443 -cipher "EDH" | \ > grep "Server Temp Key" > > See: https://www.openssl.org/blog/ +1 - -chris -BEGIN PGP SIGNATURE- Comment: GPGTools - http://gpgtools.org iQIcBAEBCAAGBQJVetd7AAoJEBzwKT+lPKRYN7AQAIWyRymVO3NYefp/tdMU/9Kf 2uTnWgmL9j7iI9EeF8RwKNNBQBLWxJItEipsn44z6Cx16yh+ZbbI0ePKvSE3UBlQ 9lJfgRtHNfOLkUZQ0NWgl+cSAs8dfctN5Qpv9kSetO/IylRQE35uMN3UubNzieoo qpS3ub28bstMVD7ATmgG7/Cyhap2IVbVVQ4/EiuaxuZkrE1Yp+JujJFJ1kktbync rWC3EvYfQm2cThFXhwZQlewOqysvNkFh4wKLQf+SuVrVqBdrZ5CjrfkqfsrFqhRo pORL+q60Ik+7vu6Cymb1GCgFU6nnb/NCe5yZ07jzcYg1ebmFuOL/cginrfzeirsU CwZf/7XOblJToYLNGP/G33lmREPc4h/QOfnvcakjznkeKMRB6ijFEvcYTh5EOPfd IaNCnAqhv+zD7R4W00QfMZRricUfrzhHlwGSoLrU49ct+wwbZXfqW8N2mQRz11Bx LdsOVp2mitFvCFq0rf/88ZER+ub12NVYWiuJERtpV4mS2r3Hkck2wnj5pYIeLtti 9gl/8E8dNF5tuE/XnLreynHkEiUZov5KLszIihj5tgSbEmQkcr17RtkhnbTYFHq8 PsakYpaxactc8nBXvoi7Ev25VtOFUJzbG+jtQsJSscaE4dF4RnfruliBfTuLVzAh /XqCtf1Q2y/9LW6EbRb4 =si8C -END PGP SIGNATURE- - To unsubscribe, e-mail: users-unsubscr...@tomcat.apache.org For additional commands, e-mail: users-h...@tomcat.apache.org
Re: tcnative CVE-2015-4000 (Logjam)
Am 12.06.2015 um 04:01 schrieb Christopher Schultz: -BEGIN PGP SIGNED MESSAGE- Hash: SHA256 Arthur, On 6/11/15 4:34 PM, Arthur Ramsey wrote: On 06/11/2015 02:35 PM, Christopher Schultz wrote: Arthur, On 6/11/15 2:14 PM, Arthur Ramsey wrote: Is anyone aware of a way to mitigate the Logjam attack with tomcat 7 and java 7? Disable DHE_EXPORT on the server? I believe I have, but Qualys SSL Server Test still fails me on the Logjam check. SSLCipherSuite="ECDHE-RSA-AES128-GCM-SHA256:ECDHE-ECDSA-AES128-GCM-SH A256:ECDHE-RSA-AES256-GCM-SHA384:ECDHE-ECDSA-AES256-GCM-SHA384:DHE-RSA-A ES128-GCM-SHA256:DHE-DSS-AES128-GCM-SHA256:kEDH+AESGCM:ECDHE-RSA-AES128- SHA256:ECDHE-ECDSA-AES128-SHA256:ECDHE-RSA-AES128-SHA:ECDHE-ECDSA-AES128 - -SHA:ECDHE-RSA-AES256-SHA384:ECDHE-ECDSA-AES256-SHA384:ECDHE-RSA-AES256- SHA:ECDHE-ECDSA-AES256-SHA:DHE-RSA-AES128-SHA256:DHE-RSA-AES128-SHA:DHE- DSS-AES128-SHA256:DHE-RSA-AES256-SHA256:DHE-DSS-AES256-SHA:DHE-RSA-AES25 6-SHA:!aNULL:!eNULL:!EXPORT:!DES:!RC4:!3DES:!MD5:!PSK You also have DHE-* ciphers in there, which is probably the problem. Remove those and I think Qualys will be happier. Really, who is using DHE in the first place? I use tcnative and openssl-1.0.2a both compiled from source in production today, but I would be open to JSSE too. I believe I need Java 8 to mitigate CVE-2015-4000 with JSSE. Why? See http://stackoverflow.com/questions/30352105/how-to-set-custom-dh-grou p-in-java-sslengine-to-prevent-logjam-attack Understood. I thought you just wanted to remove the EXPORT and DHE ciphers in general. Increasing the number of bits in the DH parameters will in fact require an upgrade. I don't see anyway to use a unique 2048-bit or greater DH group with tcnative currently. I believe you are correct; there is a bug in BZ: https://bz.apache.org/bugzilla/show_bug.cgi?id=56108 It looks like 1.1.34 will have this feature. You can build the current trunk of the 1.1 branch and probably be okay. Thanks, I'll give it a try. Scary to use in production, but it may be my best answer. I'm not sure if there is anything I can do at compile time. I'd rather not change the cipher suites as I want to maintain browser support. You should disable EXPORT certificates no matter what. Or were you talking about the DH parameters? I was talking about DH parameters. My server configuration passed the Qualys SSL Server Test with flying colors until Logjam, so I would be worried about regressions on other security fixes if I used JSSE. -chris - -chris There's two parts under Logjam: - a downgrade attack that makes the real attack very feasible. The downgrade only works if client and derver support the export ciphers (it is not necessary that they are the preferred ciphers) and the attacker is an active man-in-the-middle, ie. she can observe and change the communication. In this case the encryption can be forced to use a 512 Bit key and is relatively easy to break. To mitigate the downgrade attack, it should be posible to just disable export ciphers on the server side, which is doable per configuration. - in addition for non-export ciphers, key length of 768 bits and 1024 bits are assumed to be atackable depending on the computing ressources tha attacker has at her hand. 768 is expected to be breakable using academic computing ressources, 1024 bits using national computing resources. To mitigate this, one should use longer keys. I think that is not possible with current tcnative 1.1.33. Only the head of 1.1 has code to allow that. This code would - use a longer key automatically, if the key in the server certificate is longer. E.g. a 2048 bit RSA key would lead to using also a 2048 bit DHE key automatically. This 2048 DH params are standard DH params but should nevertheless be safe due to their length. - allows to add custom DH params to the certificate file to choose completely custom DH params. With existing 1.1.33 you can choose your cipher suite, so that non-DHE ciphers come first and set SSLHonorCipherOrder such that the client chooses the first matching cipher and DHE will likely not be used, only by client who do not support a cipher to the left of DHE in your cipher list. Note that old Java versions as clients (6, maybe 7 depending on patch level?) have a problem with DHE keys longer than 768 or 1024 bits (depending on JVM details). So by mitigating Logjam you might run into compatibility issues with those. It would be interesting to know, what details SSLLabs tell you, e.g. if they say you are vulnerable to the export downgrade attack (really bad), or "just" to your DH params should be longer. You can use the OpenSSL commandline client in version 1.0.2 to check, what param length a handshake results in: openssl s_client -connect www.example.com:443 -cipher "EDH" | \ grep "Server Temp Key" See: https://www.openssl.org/blog/ Regards, Rainer ---
Re: tcnative CVE-2015-4000 (Logjam)
-BEGIN PGP SIGNED MESSAGE- Hash: SHA256 Arthur, On 6/11/15 4:34 PM, Arthur Ramsey wrote: > On 06/11/2015 02:35 PM, Christopher Schultz wrote: Arthur, > > On 6/11/15 2:14 PM, Arthur Ramsey wrote: Is anyone aware of a way to mitigate the Logjam attack with tomcat 7 and java 7? > Disable DHE_EXPORT on the server? >> I believe I have, but Qualys SSL Server Test still fails me on >> the Logjam check. > >> SSLCipherSuite="ECDHE-RSA-AES128-GCM-SHA256:ECDHE-ECDSA-AES128-GCM-SH A256:ECDHE-RSA-AES256-GCM-SHA384:ECDHE-ECDSA-AES256-GCM-SHA384:DHE-RSA-A ES128-GCM-SHA256:DHE-DSS-AES128-GCM-SHA256:kEDH+AESGCM:ECDHE-RSA-AES128- SHA256:ECDHE-ECDSA-AES128-SHA256:ECDHE-RSA-AES128-SHA:ECDHE-ECDSA-AES128 - -SHA:ECDHE-RSA-AES256-SHA384:ECDHE-ECDSA-AES256-SHA384:ECDHE-RSA-AES256- SHA:ECDHE-ECDSA-AES256-SHA:DHE-RSA-AES128-SHA256:DHE-RSA-AES128-SHA:DHE- DSS-AES128-SHA256:DHE-RSA-AES256-SHA256:DHE-DSS-AES256-SHA:DHE-RSA-AES25 6-SHA:!aNULL:!eNULL:!EXPORT:!DES:!RC4:!3DES:!MD5:!PSK You >> also have DHE-* ciphers in there, which is probably the problem. Remove those and I think Qualys will be happier. Really, who is using DHE in the first place? I use tcnative and openssl-1.0.2a both compiled from source in production today, but I would be open to JSSE too. I believe I need Java 8 to mitigate CVE-2015-4000 with JSSE. > Why? >> See >> http://stackoverflow.com/questions/30352105/how-to-set-custom-dh-grou p-in-java-sslengine-to-prevent-logjam-attack Understood. >> I thought you just wanted to remove the EXPORT and DHE ciphers in general. Increasing the number of bits in the DH parameters will in fact require an upgrade. I don't see anyway to use a unique 2048-bit or greater DH group with tcnative currently. > I believe you are correct; there is a bug in BZ: > https://bz.apache.org/bugzilla/show_bug.cgi?id=56108 > > It looks like 1.1.34 will have this feature. You can build the > current trunk of the 1.1 branch and probably be okay. >> Thanks, I'll give it a try. Scary to use in production, but it >> may be my best answer. > I'm not sure if there is anything I can do at compile time. I'd rather not change the cipher suites as I want to maintain browser support. > You should disable EXPORT certificates no matter what. Or were you > talking about the DH parameters? >> I was talking about DH parameters. > My server configuration passed the Qualys SSL Server Test with flying colors until Logjam, so I would be worried about regressions on other security fixes if I used JSSE. > -chris - -chris -BEGIN PGP SIGNATURE- Comment: GPGTools - http://gpgtools.org iQIcBAEBCAAGBQJVej2CAAoJEBzwKT+lPKRYyPIQAIJvUDQE7KP7Zmlx5JkQhmJP GYPT7TdElrY1cUgThCkswN2zKSK8q000BSjggO+s64iChEUn+FeQW9BRBRXXdhXX 0p13g9F+SwbMskrJxVYVK9dAjeamlND2OyIo6WS5u6LHc+a2z5fT1p6Qvu5n4HbD 4iy+rBwHZuW9RBQgg8R3w7iT1ioGzRIu6gexXZzc23l1JUi9lnr1CiJETrISK3Dh 4159fpVR3yZHPiukckyfRGR3KghhgFyJO7WXO5xBIOvyR/Pmg0ltBlDyi7U1cJQx ZxX3Rq/kLUy5SFDVTYDf8cJ8pNXbG8hUfnGGWUtxar05Xb7zIHESjajUZuwIGFkg D7z7IaWPSE4w6Y1Mwb2oBcVgQCALhUXwqULsoo09YRuspYB5cKSGi6FuN2GiGObO MDeB+r/qZJJM9sPsu3pvehHlsuktE5nux30XUAxmm9xUZryVfFPnj9XSCE43rT33 0m7VqVCGzRTQGqjEMdiOg0HX94SPJR8vAWyfnyrsQL4iIlytWJpOQAhFy+p/4llp P+1zIn0XTDwGGLpOZVsfJIWnuhQQCROK5rDdgXz0fAWx7ET8tLbepGdwVvDMUc4i Pasbly7JtCbdg1Vr1aX1UPWp3W0irfAMZx3zfz0qrvoK0voWfBenOzuHThwcJMqk hIIKb20oTJkEvGsmtzXX =eetm -END PGP SIGNATURE- - To unsubscribe, e-mail: users-unsubscr...@tomcat.apache.org For additional commands, e-mail: users-h...@tomcat.apache.org
Re: tcnative CVE-2015-4000 (Logjam)
On 06/11/2015 02:35 PM, Christopher Schultz wrote: -BEGIN PGP SIGNED MESSAGE- Hash: SHA256 Arthur, On 6/11/15 2:14 PM, Arthur Ramsey wrote: Is anyone aware of a way to mitigate the Logjam attack with tomcat 7 and java 7? Disable DHE_EXPORT on the server? I believe I have, but Qualys SSL Server Test still fails me on the Logjam check. SSLCipherSuite="ECDHE-RSA-AES128-GCM-SHA256:ECDHE-ECDSA-AES128-GCM-SHA256:ECDHE-RSA-AES256-GCM-SHA384:ECDHE-ECDSA-AES256-GCM-SHA384:DHE-RSA-AES128-GCM-SHA256:DHE-DSS-AES128-GCM-SHA256:kEDH+AESGCM:ECDHE-RSA-AES128-SHA256:ECDHE-ECDSA-AES128-SHA256:ECDHE-RSA-AES128-SHA:ECDHE-ECDSA-AES128-SHA:ECDHE-RSA-AES256-SHA384:ECDHE-ECDSA-AES256-SHA384:ECDHE-RSA-AES256-SHA:ECDHE-ECDSA-AES256-SHA:DHE-RSA-AES128-SHA256:DHE-RSA-AES128-SHA:DHE-DSS-AES128-SHA256:DHE-RSA-AES256-SHA256:DHE-DSS-AES256-SHA:DHE-RSA-AES256-SHA:!aNULL:!eNULL:!EXPORT:!DES:!RC4:!3DES:!MD5:!PSK I use tcnative and openssl-1.0.2a both compiled from source in production today, but I would be open to JSSE too. I believe I need Java 8 to mitigate CVE-2015-4000 with JSSE. Why? See http://stackoverflow.com/questions/30352105/how-to-set-custom-dh-group-in-java-sslengine-to-prevent-logjam-attack I don't see anyway to use a unique 2048-bit or greater DH group with tcnative currently. I believe you are correct; there is a bug in BZ: https://bz.apache.org/bugzilla/show_bug.cgi?id=56108 It looks like 1.1.34 will have this feature. You can build the current trunk of the 1.1 branch and probably be okay. Thanks, I'll give it a try. Scary to use in production, but it may be my best answer. I'm not sure if there is anything I can do at compile time. I'd rather not change the cipher suites as I want to maintain browser support. You should disable EXPORT certificates no matter what. Or were you talking about the DH parameters? I was talking about DH parameters. My server configuration passed the Qualys SSL Server Test with flying colors until Logjam, so I would be worried about regressions on other security fixes if I used JSSE. - -chris -BEGIN PGP SIGNATURE- Comment: GPGTools - http://gpgtools.org iQIcBAEBCAAGBQJVeeL0AAoJEBzwKT+lPKRYkJQQAIyplWF0F65zvlzuQTrsFYPC Ioh+w4ddwalB1OFzaxGjnulwN9eO91iudqFiZyFpZnh7jV8GOJCQVO5whbBIXvQm l4RUispklWXNh2ClFfkW2YoXwfZPBhk4um5oVo2KHN7wf3F9AhvA/oz3Ecm2WUdg lQ7q4+WapZknWS6YdxwMzG7Jl7k6gGgnhfe6SmtEYMDKE8ktTcyAjpHX+NhXXC+e iCiZ0+DH1lYmUIHdVJu2FIgdui0CVecArJ9ufniiIpbYOnjWFxu+IZGlBuTgoAHg 8Lu7koGDOOnagSdJ6DNJeEyniRVPA61zcKRIqB1IWJJzgZVIpo8/wF4r9jGFIH3b x+3cqqSiDLppHar48ENIGbqYRCwybRCiJu3SvKLJ/zRs51ybxKbSXOondPWqIRD/ rbLQN6Z/2nQUeSp7A7iKGQj1CqFSDp5IFBqwvP4A9xWFbqCbwOWUfKhgM8UrToLN DRbtjdpGZvA0lJqxmR9nKWn9K9nNRcViI2wlcDOB22RFjz2S+fUToylf8utUJbW0 MJ5GdqnPYMp3r0NajnWaY8z1POneaqnLHnW5xnhLA2UgDBoClUA2Xe/UmU+ngUT3 OOJDb52+Xr3V+JvsDuK6cgoHTM7X+2i3+75acigwMyPYO34hA1uanVhx7XTvheqA XkCixeOIXgynHCDcWYDc =Lycq -END PGP SIGNATURE- - To unsubscribe, e-mail: users-unsubscr...@tomcat.apache.org For additional commands, e-mail: users-h...@tomcat.apache.org Thanks, Arthur - To unsubscribe, e-mail: users-unsubscr...@tomcat.apache.org For additional commands, e-mail: users-h...@tomcat.apache.org
Re: tcnative CVE-2015-4000 (Logjam)
-BEGIN PGP SIGNED MESSAGE- Hash: SHA256 Arthur, On 6/11/15 2:14 PM, Arthur Ramsey wrote: > Is anyone aware of a way to mitigate the Logjam attack with tomcat > 7 and java 7? Disable DHE_EXPORT on the server? > I use tcnative and openssl-1.0.2a both compiled from source in > production today, but I would be open to JSSE too. I believe I > need Java 8 to mitigate CVE-2015-4000 with JSSE. Why? > I don't see anyway to use a unique 2048-bit or greater DH group > with tcnative currently. I believe you are correct; there is a bug in BZ: https://bz.apache.org/bugzilla/show_bug.cgi?id=56108 It looks like 1.1.34 will have this feature. You can build the current trunk of the 1.1 branch and probably be okay. > I'm not sure if there is anything I can do at compile time. I'd > rather not change the cipher suites as I want to maintain browser > support. You should disable EXPORT certificates no matter what. Or were you talking about the DH parameters? > My server configuration passed the Qualys SSL Server Test with > flying colors until Logjam, so I would be worried about regressions > on other security fixes if I used JSSE. - -chris -BEGIN PGP SIGNATURE- Comment: GPGTools - http://gpgtools.org iQIcBAEBCAAGBQJVeeL0AAoJEBzwKT+lPKRYkJQQAIyplWF0F65zvlzuQTrsFYPC Ioh+w4ddwalB1OFzaxGjnulwN9eO91iudqFiZyFpZnh7jV8GOJCQVO5whbBIXvQm l4RUispklWXNh2ClFfkW2YoXwfZPBhk4um5oVo2KHN7wf3F9AhvA/oz3Ecm2WUdg lQ7q4+WapZknWS6YdxwMzG7Jl7k6gGgnhfe6SmtEYMDKE8ktTcyAjpHX+NhXXC+e iCiZ0+DH1lYmUIHdVJu2FIgdui0CVecArJ9ufniiIpbYOnjWFxu+IZGlBuTgoAHg 8Lu7koGDOOnagSdJ6DNJeEyniRVPA61zcKRIqB1IWJJzgZVIpo8/wF4r9jGFIH3b x+3cqqSiDLppHar48ENIGbqYRCwybRCiJu3SvKLJ/zRs51ybxKbSXOondPWqIRD/ rbLQN6Z/2nQUeSp7A7iKGQj1CqFSDp5IFBqwvP4A9xWFbqCbwOWUfKhgM8UrToLN DRbtjdpGZvA0lJqxmR9nKWn9K9nNRcViI2wlcDOB22RFjz2S+fUToylf8utUJbW0 MJ5GdqnPYMp3r0NajnWaY8z1POneaqnLHnW5xnhLA2UgDBoClUA2Xe/UmU+ngUT3 OOJDb52+Xr3V+JvsDuK6cgoHTM7X+2i3+75acigwMyPYO34hA1uanVhx7XTvheqA XkCixeOIXgynHCDcWYDc =Lycq -END PGP SIGNATURE- - To unsubscribe, e-mail: users-unsubscr...@tomcat.apache.org For additional commands, e-mail: users-h...@tomcat.apache.org