Security Concern: Unsigned Windows Executable
Dear GnuPG Team, I challenged myself to verify all software that I download on my new machine is verified and signed. Sadly, Win-GnuPG let me down. Heres why. Most software was distributed as a signed .exe file (using the Windows signed .EXE format). Some was not signed, but available via an https connection, allowing me to verify the originating source. And some, such as Gygwin, WinGnuPG, and sha1sum, required I already have GnuPG or sha1sum already installed to verify the .sig. Of course, this creates a bootstrapping problem for several reasons: 1) These .exe's aren't signed windows .exe's, 2) They aren't available via https (and thus can't ensure there isnt' a man-in-the middle), and 3) Even if I had sha1sum, I'd have to use http and not https to download the .sig file, allowing for the man-in-the-middle to deliver a checksum matching his hacked version. Using GnuPG to verify downloads does nothing, if I can't verify that GnuPG itself isn't valid. Now yes, you'll say "You're running Windows XP, that's your problem". Yes, yes, this is true. However, it still leaves the issue... why isn't an HTTPS download or a Signed Windows .EXE available, so that users can have confidence in what is downloaded from the GnuPG project? Regards, Doug Bateman P.S. Please CC: me on the reply if possible. ___ Gnupg-users mailing list Gnupg-users@gnupg.org http://lists.gnupg.org/mailman/listinfo/gnupg-users
Re: S/MIME certs and GPG
Sorry for the multiple emails. I was having a slow connection. Frustration = multiple clicks. Regards, Sauman ___ Gnupg-users mailing list Gnupg-users@gnupg.org http://lists.gnupg.org/mailman/listinfo/gnupg-users
Running GPG from a PHP Script under Windows
I'm trying to run gpg in a php script under Windows using Apache. It works fine using a batch file but using the same code using exec calls fails. Any suggestions ? -- View this message in context: http://www.nabble.com/Running-GPG-from-a-PHP-Script-under-Windows-tp23822809p23822809.html Sent from the GnuPG - User mailing list archive at Nabble.com. ___ Gnupg-users mailing list Gnupg-users@gnupg.org http://lists.gnupg.org/mailman/listinfo/gnupg-users
S/MIME certs and GPG
Hi, I'm having problems using S/MIME certificates in GPG. After importing my keys into gpg(they appear in Kleopatra when I'm done), signing and encrypting with S/MIME is unsuccessful. Using Kleopatra's system tray Clipboard function to sign with the S/MIME cert gives me the error: *Signing Failed: General Error* And using Kleopatra's Clipboard function to encrypt with the S/MIME cert gives the following error: *Encryption Failed: Not Found* I'm using Gpg4Win 1.9.16 BETA and the certs are issued by my CA(Windows Server 2008, Exchange Server). So far (after much effort) I have managed to export the certs(private and public keys) via Internet Explorer (.pfx files) and using OpenSSL, convert the bundle into .pem and extract the key, issuer cert and user cert. The method I go through to get the cert into gpg is as follows: gpgsm --call-protect-tool --p12-import --store charliekey.p12 gpgsm --import (after which I copy+paste the details from the pem file from Bag Attributes to ---END CERTIFICATE--- for the issuer cert and then repeated for the user's cert) However, the secret key does not show in --list-secret-keys. It shows however if I do this: gpgsm --import charliesign.pfx Output: gpgsm: gpg-protect-tool: 1240 bytes of 3DES encrypted text gpgsm: gpg-protect-tool: 2728 bytes of RC2 encrypted text gpgsm: gpg-protect-tool: processing certBag gpgsm: gpg-protect-tool: processing certBag gpgsm: gpg-protect-tool: keygrip: 8069846C970B7CC3FADEBE2B7995400817694359 gpgsm: gpg-protect-tool: AllowSetForegroundWindow(2188) failed: Access is denied. gpgsm: gpgsm: gpg-protect-tool: AllowSetForegroundWindow(3116) failed: Access is denied. gpgsm: gpgsm: gpg-protect-tool: NOTE: you should run 'diskperf -y' to enable the disk statistics gpgsm: gpg-protect-tool: NOTE: you should run 'diskperf -y' to enable the disk statistics gpgsm: gpg-protect-tool: NOTE: you should run 'diskperf -y' to enable the disk statistics gpgsm: gpg-protect-tool: NOTE: you should run 'diskperf -y' to enable the disk statistics gpgsm: gpg-protect-tool: secret key stored as `C:\Documents and Settings\user\Application Data\gnupg\private-keys-v1.d\8069846C970B7CC3FADEBE2B7995400817694359.key' gpgsm: total number processed: 2 gpgsm: unchanged: 2 As can be seen in the output, the secret key has been stored, and shows when I type gpgsm --list-secret-keys. Importing the pfx works in xp, but does not work in Vista. Does anyone know why my secret key can't be imported? I might have been doing something wrong. The command I use to extract the secret key is: openssl pkcs12 -in charliebundle.pem -export -out charliekey.p12 -nocerts -nodes Also, I have read that after importing the certs I have to make them trusted. How do I do so? I cannot find the file ~/.gnupg/trustlist.txt mentioned in these tutorials. Any help is much appreciated! Thanks! Regards, Sauman ___ Gnupg-users mailing list Gnupg-users@gnupg.org http://lists.gnupg.org/mailman/listinfo/gnupg-users
S/MIME certs and GPG
Hi, I'm having problems using S/MIME certificates in GPG. After importing my keys into gpg(they appear in Kleopatra when I'm done), signing and encrypting with S/MIME is unsuccessful. Using Kleopatra's system tray Clipboard function to sign with the S/MIME cert gives me the error: *Signing Failed: General Error* And using Kleopatra's Clipboard function to encrypt with the S/MIME cert gives the following error: *Encryption Failed: Not Found* I'm using Gpg4Win 1.9.16 BETA and the certs are issued by my CA(Windows Server 2008, Exchange Server). So far (after much effort) I have managed to export the certs(private and public keys) via Internet Explorer (.pfx files) and using OpenSSL, convert the bundle into .pem and extract the key, issuer cert and user cert. The method I go through to get the cert into gpg is as follows: gpgsm --call-protect-tool --p12-import --store charliekey.p12 gpgsm --import (after which I copy+paste the details from the pem file from Bag Attributes to ---END CERTIFICATE--- for the issuer cert and then repeated for the user's cert) However, the secret key does not show in --list-secret-keys. It shows however if I do this: gpgsm --import charliesign.pfx Output: gpgsm: gpg-protect-tool: 1240 bytes of 3DES encrypted text gpgsm: gpg-protect-tool: 2728 bytes of RC2 encrypted text gpgsm: gpg-protect-tool: processing certBag gpgsm: gpg-protect-tool: processing certBag gpgsm: gpg-protect-tool: keygrip: 8069846C970B7CC3FADEBE2B7995400817694359 gpgsm: gpg-protect-tool: AllowSetForegroundWindow(2188) failed: Access is denied. gpgsm: gpgsm: gpg-protect-tool: AllowSetForegroundWindow(3116) failed: Access is denied. gpgsm: gpgsm: gpg-protect-tool: NOTE: you should run 'diskperf -y' to enable the disk statistics gpgsm: gpg-protect-tool: NOTE: you should run 'diskperf -y' to enable the disk statistics gpgsm: gpg-protect-tool: NOTE: you should run 'diskperf -y' to enable the disk statistics gpgsm: gpg-protect-tool: NOTE: you should run 'diskperf -y' to enable the disk statistics gpgsm: gpg-protect-tool: secret key stored as `C:\Documents and Settings\user\Application Data\gnupg\private-keys-v1.d\8069846C970B7CC3FADEBE2B7995400817694359.key' gpgsm: total number processed: 2 gpgsm: unchanged: 2 As can be seen in the output, the secret key has been stored, and shows when I type gpgsm --list-secret-keys. Importing the pfx works in xp, but does not work in Vista. Does anyone know why my secret key can't be imported? I might have been doing something wrong. The command I use to extract the secret key is: openssl pkcs12 -in charliebundle.pem -export -out charliekey.p12 -nocerts -nodes Also, I have read that after importing the certs I have to make them trusted. How do I do so? I cannot find the file ~/.gnupg/trustlist.txt mentioned in these tutorials. Any help is much appreciated! Thanks! Regards, Sauman ___ Gnupg-users mailing list Gnupg-users@gnupg.org http://lists.gnupg.org/mailman/listinfo/gnupg-users
Re: Security Concern: Unsigned Windows Executable
Doug Bateman wrote: > I challenged myself to verify all software that I download on my new > machine is verified and signed. Sadly, Win-GnuPG let me down. Heres why. What's Win-GnuPG? Are you referring to the windows installer build of GnuPG from http://www.gnupg.org/download/ as such? It's just GnuPG. > Most software was distributed as a signed .exe file (using the Windows > signed .EXE format). Some was not signed, but available via an https > connection, allowing me to verify the originating source. And some, > such as Gygwin, WinGnuPG, and sha1sum, required I already have GnuPG or > sha1sum already installed to verify the .sig. Of course, this creates a > bootstrapping problem for several reasons: 1) These .exe's aren't signed > windows .exe's, 2) They aren't available via https (and thus can't > ensure there isnt' a man-in-the middle), and 3) Even if I had sha1sum, > I'd have to use http and not https to download the .sig file, allowing > for the man-in-the-middle to deliver a checksum matching his hacked version. > > Using GnuPG to verify downloads does nothing, if I can't verify that > GnuPG itself isn't valid. I believe the Windows signed .EXE format is X.509 cert based and as such isn't going to help much if the signing certificate doesn't chain back to Windows set of root certs. COTS products will probably invest the money to implement this, it's unlikely for F/OSS. It also assumes the Microsoft technology to create Authenticode signatures is available to F/OSS developers. Your MITM scenarios leave out the crucial step of your attacker also needing to possess Werner Koch's signing key. The .SIG is not just a checksum, it is a digital signature. The verification looks like this: $ gpg -v gnupg-w32cli-1.4.9.exe.sig gpg: assuming signed data in `gnupg-w32cli-1.4.9.exe' gpg: Signature made 03/26/08 12:51:54 using RSA key ID 1CE0C630 gpg: using PGP trust model gpg: Good signature from "Werner Koch (dist sig) " gpg: binary signature, digest algorithm SHA1 Your #3 comment is confusing. There is no .SIG to download if verifying with sha1sum. You run sha1sum against the file you wish to verify and compare the program output with the published value. Are you proposing some MITM attack of a replaced installer executable with an /identical/ SHA-1 value? sha1sum and md5sum are widely available as source. If you're so committed to this verified and signed thing that you're unwilling to trust anything, you probably should look into building some things of your own. sha1sum is available as source and/or windows executable along with the respective digital signatures from ftp://ftp.gnupg.org/gcrypt/binary/ Sooner or later you have to establish a base trust. OH! Maybe you could use an eval version of PGP to verify the cryptographic signature on the GnuPG installer. Of course that probably hinges on its installer being a Windows signed-executable right? ;-) Links discussed in this message: Installer: ftp://ftp.gnupg.org/gcrypt/binary/gnupg-w32cli-1.4.9.exe Installer signature ftp://ftp.gnupg.org/gcrypt/binary/gnupg-w32cli-1.4.9.exe.sig SHA-1 checksum for Installer c2efad983dfe50e6d8007257bad2c76604be389a gnupg-w32cli-1.4.9.exe > P.S. Please CC: me on the reply if possible. -- John P. Clizbe Inet:John (a) Mozilla-Enigmail.org You can't spell fiasco without SCO. hkp://keyserver.gingerbear.net or mailto:pgp-public-k...@gingerbear.net?subject=help Q:"Just how do the residents of Haiku, Hawai'i hold conversations?" A:"An odd melody / island voices on the winds / surplus of vowels" signature.asc Description: OpenPGP digital signature ___ Gnupg-users mailing list Gnupg-users@gnupg.org http://lists.gnupg.org/mailman/listinfo/gnupg-users
Re: Avoid pinentry-gtk-2 when using console!
On Sun, 31 May 2009 07:49, rog...@sdf.lonestar.org said:
> if {environmental variable is set to console/gtk/qt3}
> use the specified pinentry flavor
You can easily implement this with a little pinentry wrapper script and
using the PINENTRY_USER_DATA envvar which is passed all the way from gpg
to Pinentry.
> I'm guessing, the current solution is to assume the user is a dumb X
> user. ;-)
Definitely not. Pinentry pops up and grabs the keyboard for a good
reasons: This makes it much harder to preset a faked Pinentry prompt and
sniff the Passphrase entered by the user. The curses version can't do
that and thus the default is to use an X window if XDISPLAY is set. If
you fear faked popup windows you may modify pinentry to show a custom
image.
I am using gpg-agent for many years now and do almost all my work in
xterms and Emacs. It does not bother me if Pinentry popups due to
background jobs every hour or so.
Salam-Shalom,
Werner
--
Die Gedanken sind frei. Auschnahme regelt ein Bundeschgesetz.
___
Gnupg-users mailing list
Gnupg-users@gnupg.org
http://lists.gnupg.org/mailman/listinfo/gnupg-users
Re: Security Concern: Unsigned Windows Executable
John Clizbe wrote: > Your #3 comment is confusing. There is no .SIG to download if verifying > with sha1sum. You run sha1sum against the file you wish to verify and > compare the program output with the published value. > > Are you proposing some MITM attack of a replaced installer executable > with an /identical/ SHA-1 value? Alternately, he could be implying an active MitM attack, where the attacker is intercepting both the downloaded hash value (replacing it with the trojaned version's hash value) and the application itself (replacing it with a trojaned version). That said, if you're presently being targeted by people who are capable of intercepting and modifying your network traffic in realtime, neither GnuPG nor Authenticode signatures can help you. You need professional help: lawyers and security geeks will help you an awful lot more than HTTPS or Authenticode. > sha1sum and md5sum are widely available as source. If you're so > committed to this verified and signed thing that you're unwilling to > trust anything, you probably should look into building some things of > your own. Insert mandatory "reflections on trusting trust" reference here. The sentiment of "I must build it from source if I'm going to trust it" is great, but then you have to ask questions about your compiler, your system libraries, etc., until you're left hand-hacking Assembly instructions for a low transistor count CPU you've personally lithographed yourself from your own personal design. ___ Gnupg-users mailing list Gnupg-users@gnupg.org http://lists.gnupg.org/mailman/listinfo/gnupg-users
Re: Running GPG from a PHP Script under Windows
On 06/01/2009 05:38 PM, jnhemley wrote: > I'm trying to run gpg in a php script under Windows using Apache. It works > fine using a batch file but using the same code using exec calls fails. Any > suggestions ? This sounds more like a question for php folks than gpg folks to me. I recommend you try asking on one of the many php forums. And if you want to get a helpful answer, you should probably consider including more detail. For example, show the code you are using exactly, the output it produces under the different circumstances, and the exact error codes and messages produced during a failure. hth, --dkg signature.asc Description: OpenPGP digital signature ___ Gnupg-users mailing list Gnupg-users@gnupg.org http://lists.gnupg.org/mailman/listinfo/gnupg-users
Changing the expiration date after the key has expired
Hi, I just wondered if it was possible to postpone the expiration date after it has been set and/or after the deadline has been reached. 2 years ago, I created a personal key and set the expiration to 2y, so it has now expired. It looks like I can change the expiration date (of the primary key) in gnupg software "using edit-key". My subkey expired too, so I have generated a new one. I've tried to export the result and put it on the mit keyserver but it failed. According to the message I've read, it was because my userids wer signed by two keys (which is more or less wrong : I've checked and they are signed twice by the same key, but at different dates). What is strange is I've tried another keyserver and it worked (without removing the expired signature). But, well, the real problem is that now, even if my new subkey has been imported successfully, the primary key on the keyserver still has the old expiration date set - i.e. the primary key has expired : do you know if I can update the key on the keyserver so that it is aware of the new expiration date ? Thanks, Vincent ___ Gnupg-users mailing list Gnupg-users@gnupg.org http://lists.gnupg.org/mailman/listinfo/gnupg-users
Re: Security Concern: Unsigned Windows Executable
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 Robert J. Hansen wrote: | Insert mandatory "reflections on trusting trust" reference here. | | The sentiment of "I must build it from source if I'm going to trust it" | is great, but then you have to ask questions about your compiler, your | system libraries, etc., until you're left hand-hacking Assembly | instructions for a low transistor count CPU you've personally | lithographed yourself from your own personal design. | Let's say I did all that. But do I trust the guy who looked over my shoulder to be sure I did not make a mistake in my own personal design? And if I believe, in principle, in automatically proving programs (or hardware, their equivalent) correct, do I trust the program that does that? And the rules given that program that the program to be verified is to meet? We get into the very problem Rene Descartes was stuck in until he came up with "Cogito, ergo sum." Which I do not think was a solution at all. - -- ~ .~. Jean-David Beyer Registered Linux User 85642. ~ /V\ PGP-Key: 9A2FC99A Registered Machine 241939. ~ /( )\ Shrewsbury, New Jerseyhttp://counter.li.org ~ ^^-^^ 08:50:01 up 69 days, 15:04, 3 users, load average: 4.06, 4.24, 4.31 -BEGIN PGP SIGNATURE- Version: GnuPG v1.4.5 (GNU/Linux) Comment: Using GnuPG with CentOS - http://enigmail.mozdev.org iD8DBQFKJSFOPtu2XpovyZoRAmheAKC7PlUg4LWQsz9HdbP09cXdu/mIHwCcDrYG X15Zb0CWZ1SbmpgFl+JibYs= =NdyX -END PGP SIGNATURE- ___ Gnupg-users mailing list Gnupg-users@gnupg.org http://lists.gnupg.org/mailman/listinfo/gnupg-users
Re: Changing the expiration date after the key has expired
On 06/02/2009 10:14 AM, Vincent Panel wrote: > I just wondered if it was possible to postpone the expiration date > after it has been set and/or after the deadline has been reached. yes, this is possible. Assming you're talking about 56B55C11, it looks like you've successfully done so. > I've tried to export the result and put it on the mit keyserver but it > failed. According to the message I've read, it was because my userids > wer signed by two keys (which is more or less wrong : I've checked and > they are signed twice by the same key, but at different dates). It's actually self-signed three times by the same key: * the original self-signature * the new self-signature with the updated expiration * a third self-signature which moves the "primary User ID" flag from one UID to another. If pgp.mit.edu rejected the key, that's a bug in that keyserver. I just tried pulling this key from pgp.mit.edu and from pool.sks-keyservers.net, and found that pgp.mit.edu only had the first two self-sigs on each UID, while pool.sks-keyservers.net had all three. then i tried pushing the full key (with all three self-sigs) back to pgp.mit.edu. After that, pgp.mit.edu returned all three self-sigs. So it seems there was a buggy propagation in there, but i might have just fixed it manually for this specific key. (the explicit steps described above were: umask 077 mkdir yohonet yohonet/mit yohonet/sks GNUPGHOME=yohonet/mit gpg --keyserver pgp.mit.edu --recv 56B55C11 GNUPGHOME=yohonet/sks gpg --keyserver pool.sks-keyservers.net --recv 56B55C11 GNUPGHOME=yohonet/sks gpg --list-sigs 56B55C11 GNUPGHOME=yohonet/mit gpg --list-sigs 56B55C11 GNUPGHOME=yohonet/sks gpg --keyserver pgp.mit.edu --send 56B55C11 GNUPGHOME=yohonet/mit gpg --keyserver pgp.mit.edu --recv 56B55C11 GNUPGHOME=yohonet/mit gpg --list-sigs 56B55C11 ) I'd be interested in seeing the error output you got from sending the key to pgp.mit.edu. When i sent the full key back to pgp.mit.edu, i got no error message at all, just the expected line from gpg: gpg: sending key 56B55C11 to hkp server pgp.mit.edu > What > is strange is I've tried another keyserver and it worked (without > removing the expired signature). It's probably a good idea to use the other keyserver then, and avoid pgp.mit.edu. > But, well, the real problem is that now, even if my new subkey has > been imported successfully, the primary key on the keyserver still has > the old expiration date set - i.e. the primary key has expired : do > you know if I can update the key on the keyserver so that it is aware > of the new expiration date ? this is already done. the old self-signature with the old expiration date will persist forever, but the new self-sig has a more recent creation date, and RFC-compliant OpenPGP implementations will respect it. --dkg signature.asc Description: OpenPGP digital signature ___ Gnupg-users mailing list Gnupg-users@gnupg.org http://lists.gnupg.org/mailman/listinfo/gnupg-users
Re: Changing the expiration date after the key has expired
On Jun 2, 2009, at 10:14 AM, Vincent Panel wrote:
Hi,
I just wondered if it was possible to postpone the expiration date
after it has been set and/or after the deadline has been reached.
Yes, you can.
2 years ago, I created a personal key and set the expiration to 2y, so
it has now expired. It looks like I can change the expiration date (of
the primary key) in gnupg software "using edit-key". My subkey expired
too, so I have generated a new one.
This is fine. Note that you can also change the expiration date of a
subkey by selecting it ("key X") and then doing the usual "expire".
I've tried to export the result and put it on the mit keyserver but it
failed. According to the message I've read, it was because my userids
wer signed by two keys (which is more or less wrong : I've checked and
they are signed twice by the same key, but at different dates). What
is strange is I've tried another keyserver and it worked (without
removing the expired signature).
This is a problem with the MIT keyserver. It runs an older keyserver
software package, that is not fully compatible with standard keys (the
software was written before the standard was finalized).
But, well, the real problem is that now, even if my new subkey has
been imported successfully, the primary key on the keyserver still has
the old expiration date set - i.e. the primary key has expired : do
you know if I can update the key on the keyserver so that it is aware
of the new expiration date ?
Your best bet is simply to not use the MIT keyserver. The other
keyservers (try "pool.sks-keyservers.net") should handle this case
without any problems.
David
___
Gnupg-users mailing list
Gnupg-users@gnupg.org
http://lists.gnupg.org/mailman/listinfo/gnupg-users
Re: Security Concern: Unsigned Windows Executable
On 06/02/2009 08:55 AM, Jean-David Beyer wrote: > obert J. Hansen wrote: > > | you're left hand-hacking Assembly > | instructions for a low transistor count CPU you've personally > | lithographed yourself from your own personal design. > > We get into the very problem Rene Descartes was stuck in until he came up > with "Cogito, ergo sum." Which I do not think was a solution at all. guys, with all due respect, the original poster was not asking for a philosophical digression. he was asking how he could practically identify the provenance of the copy of gpg he was hoping to use. It's neat to point out how each layer of trust rests on another one, but we should be giving practical advice which helps the OP push the leaps of faith necessary to run gpg back by a few levels. John Clizbe has offered one practical choice (see if PGP Corp. offers a demo version with a signed executable). Another choice would be to use a local, trusted GNU/Linux or *BSD installation to verify Werner's signature on the package (e.g. put it on a USB stick) and then transfer the package back to the windows machine for installation. A third way (if you don't currently have a local trusted free OS installation) would be to reboot the machine with a liveCD (if you can find a satisfactory trust path to a LiveCD) or with something like wubi [0] which itself might offer a signed windows installer (i haven't checked). You can use wubi or the liveCD to verify Werner's signature on the packages, and then transfer them back to the windows machine to install. hth, --dkg [0] http://wubi-installer.org/ signature.asc Description: OpenPGP digital signature ___ Gnupg-users mailing list Gnupg-users@gnupg.org http://lists.gnupg.org/mailman/listinfo/gnupg-users
AUTO: Richard Hamilton is out of the office (returning 06/08/2009)
I am out of the office until 06/08/2009. I am out of the office until June 8th 2009. If this is a production problem, please call the solution center at 918-573-2336 or email Bob Olson at robert.ol...@williams.com. I will have limited mail and cell phone access. Note: This is an automated response to your message "Re: Running GPG from a PHP Script under Windows" sent on 6/2/09 9:58:59. This is the only notification you will receive while this person is away.___ Gnupg-users mailing list Gnupg-users@gnupg.org http://lists.gnupg.org/mailman/listinfo/gnupg-users
Re: Avoid pinentry-gtk-2 when using console!
On Tue, 2009-06-02 at 12:31 +0200, Werner Koch wrote:
> On Sun, 31 May 2009 07:49, rog...@sdf.lonestar.org said:
>
> > if {environmental variable is set to console/gtk/qt3}
> > use the specified pinentry flavor
>
> You can easily implement this with a little pinentry wrapper script and
> using the PINENTRY_USER_DATA envvar which is passed all the way from gpg
> to Pinentry.
Again, still sounds like a hack as (I could have done this here). It's
the reason for posting this issue to this list (since others have the
same issue on the Internet).
> > I'm guessing, the current solution is to assume the user is a dumb X
> > user. ;-)
>
> Definitely not. Pinentry pops up and grabs the keyboard for a good
> reasons: This makes it much harder to preset a faked Pinentry prompt and
> sniff the Passphrase entered by the user. The curses version can't do
> that and thus the default is to use an X window if XDISPLAY is set. If
> you fear faked popup windows you may modify pinentry to show a custom
> image.
Think it's paranoia unless one is on a public network or is being
aggressively sought after all the time. If this is a issue, it sounds
more sensible for the administrator to use a compile time flag
(or .gnupg/option statement or environmental variable) which seeks to
make gpg/pinentry usage stricter.
Of course, then you run into a problem with users having access to their
$HOME/.gnupg option versus an /etc/gnupg file preventing writing for
enabling such a feature. Hence, a compile time option being better.
> I am using gpg-agent for many years now and do almost all my work in
> xterms and Emacs. It does not bother me if Pinentry popups due to
> background jobs every hour or so.
This is what drove me up the wall with Evolution. Granted, it enhances
security if you're always entering the pin, but quickly hinders if a
user rarely uses gpg/pgp.
(Granted, I find X useful and prefer still strongly prefer the console.
Just don't try forcing the X windows down my throat like Windows
does. ;-)
In summary -- from info gathered from this thread -- there is no coded
solution besides hacking the current files with a script that will
permit the user to use the terminal /usr/bin/pinentry
or /usr/bin/pinentry-curses while within X. (Except unsetting the X
display variable which which then would cause all X apps to fail when
starting from the terminal.)
--
Roger
http://rogerx.freeshell.org
___
Gnupg-users mailing list
Gnupg-users@gnupg.org
http://lists.gnupg.org/mailman/listinfo/gnupg-users
Re: Security Concern: Unsigned Windows Executable
Daniel Kahn Gillmor wrote: > guys, with all due respect, the original poster was not asking for a > philosophical digression. he was asking how he could practically > identify the provenance of the copy of gpg he was hoping to use. John Clizbe answered, "[i]f you're so committed to this verified and signed thing that you're unwilling to trust anything, you should probably look into building some things of your own." My remark was a very serious warning: if the OP is so committed, my "philosophical digression" is what lies at the bottom of that rabbit-hole. > John Clizbe has offered one practical choice (see if PGP Corp. offers a > demo version with a signed executable). Active MitM assumes that you have an attacker who is technically skilled and highly motivated. It is ludicrous to think that an attacker skilled enough to do active MitM and motivated enough to go after you directly would for some reason be constrained to play within the carefully defined box the crypto community has created. Rule number one of successful attacks: get outside the box. If the OP is seriously concerned that there's an active MitM attack going on against him, he needs get off the internet and obtain the professional services he needs to end the threat. No, I'm not kidding. ___ Gnupg-users mailing list Gnupg-users@gnupg.org http://lists.gnupg.org/mailman/listinfo/gnupg-users
Re: Security Concern: Unsigned Windows Executable
On 06/02/2009 07:02 PM, Robert J. Hansen wrote: > my "philosophical > digression" is what lies at the bottom of that rabbit-hole. I understand (and i very much appreciate the warning), but sometimes it's useful to go a little way into the rabbit hole instead of all the way to the bottom, no? otherwise, why bother with crypto software at all, built as it is upon a teetering edifice of trust-in-others? > Active MitM assumes that you have an attacker who is technically skilled > and highly motivated. I beg to differ. In today's wireless network, active MitM can be done by a moderately-skilled hacker on a lark, or by an unskilled user who can follow directions: http://www.ex-parrot.com/~pete/upside-down-ternet.html it's a small step from there to a script kiddie's bot on a WAP replacing everything that looks like a windows executable or installer crossing the network with a program of his own choosing (even a signed one!). Even checking sha1sums from a web page would defeat this basic attack though, which is why i think it's reasonable for the OP to ask his question. The OP wanted to to know how to make a few more checks than zero, forcing any possible attacker to be marginally more clever than the hypothesized bot above. > It is ludicrous to think that an attacker skilled > enough to do active MitM and motivated enough to go after you directly > would for some reason be constrained to play within the carefully > defined box the crypto community has created. Rule number one of > successful attacks: get outside the box. But that's exactly the OPs point: "the box" on windows is a "signed executable", whatever that is. Since gpg is distributed outside of that framework, he's concerned that an attacker could exploit it. Fortunately, we can offer the OP some other techniques to put things back "in the box" of secured communications -- he still has to trust our recommendations, but he can research those as well and make his own decisions. > If the OP is seriously concerned that there's an active MitM attack > going on against him, The MitM attack doesn't need to be "against him", it just needs to be going on. Do you cryptographically check the integrity of software you downoload? Do you check the host keys of machines you ssh into? these are both reasonable actions, based on a concern that there may be an active MitM attack. > he needs get off the internet and obtain the > professional services he needs to end the threat. The OP may not have the resources to do what you're suggesting, or may want to see what other options are available before resorting to such lengths. Practical approaches exist, and while they are imperfect, they do have an effect against some non-zero subset of real-world threats. let's not overstate their capabilities (the phrase "100% secure" is meaningless), but let's offer practical approaches even as we warn of their limitations. Regards, --dkg signature.asc Description: OpenPGP digital signature ___ Gnupg-users mailing list Gnupg-users@gnupg.org http://lists.gnupg.org/mailman/listinfo/gnupg-users
Re: Security Concern: Unsigned Windows Executable
Daniel Kahn Gillmor wrote: > I beg to differ. In today's wireless network, active MitM can be done > by a moderately-skilled hacker on a lark, or by an unskilled user who > can follow directions: > > http://www.ex-parrot.com/~pete/upside-down-ternet.html There is a big difference between hack-in-a-box stunts like that and serious attacks by people intent on succeeding. > Even checking sha1sums from a web page would defeat this basic attack > though, which is why i think it's reasonable for the OP to ask his question. You can't have it both ways. You can't say, "it's really easy to do active MitM, you just need to follow these basic instructions," and then say, "but an attacker wouldn't be able to change sha1sums." No, of course they'd be able to: if you're assuming the attacker can inject whatever they like into the data stream, then you have to assume the attacker will use that capability intelligently. > The OP wanted to to know how to make a few more checks than zero, > forcing any possible attacker to be marginally more clever than the > hypothesized bot above. And now you're arguing my point for me: there is a big difference between hack-in-a-box stunts and serious attempts to subvert your system. > But that's exactly the OPs point: "the box" on windows is a "signed > executable", whatever that is. Since gpg is distributed outside of that > framework, he's concerned that an attacker could exploit it. You are not understanding the metaphor; that may be my own fault. "The box" refers to the popular phrase, "think outside the box." > let's not overstate their capabilities (the phrase "100% secure" is > meaningless), but let's offer practical approaches even as we warn of > their limitations. I do not see that what you are presenting is practical. The presence of a serious attacker who can subvert your traffic in ways of the attacker's choosing is a massive game-changer. ___ Gnupg-users mailing list Gnupg-users@gnupg.org http://lists.gnupg.org/mailman/listinfo/gnupg-users
Re: Avoid pinentry-gtk-2 when using console!
On Tue, 2 Jun 2009 22:43, rog...@sdf.lonestar.org said: > Again, still sounds like a hack as (I could have done this here). It's Or write your own pinentry; it is a separate package for a reason. > Think it's paranoia unless one is on a public network or is being > aggressively sought after all the time. If this is a issue, it sounds Depends on how you work; I have heard that some users have Javascript and popup windows enabled in their browser and also read mail with a GUI tool. Malicious webpages could easily take advantage of that and popup a PInentry like window. > (or .gnupg/option statement or environmental variable) which seeks to > make gpg/pinentry usage stricter. Right; that is why a standard solution would not work. > (Granted, I find X useful and prefer still strongly prefer the console. > Just don't try forcing the X windows down my throat like Windows No need to. There is the curses pinentry and the gpg-agent options --keep-tty --keep-display Ignore requests to change the current tty or X window system's DISPLAY variable respectively. This is useful to lock the pinentry to pop up at the tty or display you started the agent. Originally added to support running gpg-agent in screen(1). Still you will have the annoyance/security_feature to switch to the other tty. Salam-Shalom, Werner -- Die Gedanken sind frei. Auschnahme regelt ein Bundeschgesetz. ___ Gnupg-users mailing list Gnupg-users@gnupg.org http://lists.gnupg.org/mailman/listinfo/gnupg-users
Re: Avoid pinentry-gtk-2 when using console!
On Wed, 2009-06-03 at 08:02 +0200, Werner Koch wrote: > On Tue, 2 Jun 2009 22:43, rog...@sdf.lonestar.org said: > > (or .gnupg/option statement or environmental variable) which seeks to > > make gpg/pinentry usage stricter. > > Right; that is why a standard solution would not work. > > > (Granted, I find X useful and prefer still strongly prefer the console. > > Just don't try forcing the X windows down my throat like Windows > > No need to. There is the curses pinentry and the gpg-agent options > > --keep-tty > --keep-display > Ignore requests to change the current tty or X window system's > DISPLAY variable respectively. This is useful to lock the > pinentry to pop up at the tty or display you started the agent. > > Originally added to support running gpg-agent in screen(1). Still you > will have the annoyance/security_feature to switch to the other tty. > > > Salam-Shalom, > >Werner Ah. Bingo. So really I should go file a (nagging) bug for Evolution PIM client (nagging) me to enter a passphrase each time I start the Evolution application. Ten to one, it's using a gpg keyring for encrypting the email account passwords. (ie. standard POP & IMAP which are sent unecrypted anyways!) Looks like I need to go back, re-enable the gpg-agent compile time option, ensure it's configured to my liking per your above notes concerning "keep-*" options. Gather accurate data & go bug them! ;-) -- Roger http://rogerx.freeshell.org signature.asc Description: This is a digitally signed message part ___ Gnupg-users mailing list Gnupg-users@gnupg.org http://lists.gnupg.org/mailman/listinfo/gnupg-users
