From vincent.torri at gmail.com  Sat Oct  1 07:34:18 2011
From: vincent.torri at gmail.com (Vincent Torri)
Date: Sat, 1 Oct 2011 07:34:18 +0200
Subject: libgpg-error : code-from-errno.h not correctly generated when
	cross-compiling to Windows on Linux
In-Reply-To: <87hb3ugjr7.fsf@vigenere.g10code.de>
References: <CAMq1adoRb_aOXCzrDPGY9utk--qaq1zaanJqZq8ptb93cUvAXA@mail.gmail.com>
	<87hb3ugjr7.fsf@vigenere.g10code.de>
Message-ID: <CAMq1adq6v=m+zVhs5qQ6Tf3ZyNt_rRwa2Nf+oPrZ+HJBrKZM6w@mail.gmail.com>

On Fri, Sep 30, 2011 at 4:58 PM, Werner Koch <wk at gnupg.org> wrote:

> On Fri, 30 Sep 2011 10:02, vincent.torri at gmail.com said:
>
> > configuration: ./configure --prefix=/opt/efl --host=x86_64-w64-mingw32
> > --disable-nls --disable-static
>

I actually have the same problem when cross compiling to win32 :

configuration :  ./configure --prefix=/opt/efl --host=i686-w64-mingw32
--disable-nls --disable-static

exactly the same problem

Vincent Torri
-------------- next part --------------
An HTML attachment was scrubbed...
URL: </pipermail/attachments/20111001/3ff29492/attachment.htm>

From wk at gnupg.org  Sun Oct  2 11:04:38 2011
From: wk at gnupg.org (Werner Koch)
Date: Sun, 02 Oct 2011 11:04:38 +0200
Subject: libgpg-error : code-from-errno.h not correctly generated when
	cross-compiling to Windows on Linux
In-Reply-To: <CAMq1adqX8d-U46CY5dm7Qjwn0=37S8H18kvmSr_km02E_Q1Mvw@mail.gmail.com>
	(Vincent Torri's message of "Fri, 30 Sep 2011 20:43:00 +0200")
References: <CAMq1adoRb_aOXCzrDPGY9utk--qaq1zaanJqZq8ptb93cUvAXA@mail.gmail.com>
	<87hb3ugjr7.fsf@vigenere.g10code.de>
	<CAMq1adqX8d-U46CY5dm7Qjwn0=37S8H18kvmSr_km02E_Q1Mvw@mail.gmail.com>
Message-ID: <878vp3hihl.fsf@vigenere.g10code.de>

On Fri, 30 Sep 2011 20:43, vincent.torri at gmail.com said:
> Actually, I need libgcrypt for crypto in gnutls. So it means that,
> currently, it's useless to try to compile libgcrypt for win64 ?

Don't use it.  We do not have an RNG for w64.


Shalom-Salam,

   Werner


-- 
Die Gedanken sind frei.  Ausnahmen regelt ein Bundesgesetz.



From vincent.torri at gmail.com  Sun Oct  2 23:32:42 2011
From: vincent.torri at gmail.com (Vincent Torri)
Date: Sun, 2 Oct 2011 23:32:42 +0200
Subject: libgpg-error : code-from-errno.h not correctly generated when
	cross-compiling to Windows on Linux
In-Reply-To: <CAMq1adq6v=m+zVhs5qQ6Tf3ZyNt_rRwa2Nf+oPrZ+HJBrKZM6w@mail.gmail.com>
References: <CAMq1adoRb_aOXCzrDPGY9utk--qaq1zaanJqZq8ptb93cUvAXA@mail.gmail.com>
	<87hb3ugjr7.fsf@vigenere.g10code.de>
	<CAMq1adq6v=m+zVhs5qQ6Tf3ZyNt_rRwa2Nf+oPrZ+HJBrKZM6w@mail.gmail.com>
Message-ID: <CAMq1adoBDyWL6zBX2zUX=m2nDi3vhAKB+1MjnqgjBEzXy5PqsA@mail.gmail.com>

On Sat, Oct 1, 2011 at 7:34 AM, Vincent Torri <vincent.torri at gmail.com>wrote:

>
>
> On Fri, Sep 30, 2011 at 4:58 PM, Werner Koch <wk at gnupg.org> wrote:
>
>> On Fri, 30 Sep 2011 10:02, vincent.torri at gmail.com said:
>>
>> > configuration: ./configure --prefix=/opt/efl --host=x86_64-w64-mingw32
>> > --disable-nls --disable-static
>>
>
> I actually have the same problem when cross compiling to win32 :
>
> configuration :  ./configure --prefix=/opt/efl --host=i686-w64-mingw32
> --disable-nls --disable-static
>
> exactly the same problem
>

Just lookin at the output of ./mkerrcodes, I get:

-2147168224    GPG_ERR_EPROTO
1    GPG_ERR_EPERM
2    GPG_ERR_ENOENT
etc...

So it seems that  there is a problem with the fact that GPG_ERR_EPROTO is
negative (i think). So integer size overflow.

I changed mkerrcodes.c so that it displays an unsigned int. The 2 changes
lines are:

  unsigned int i;
...
    printf ("%u\t%s\n", err_table[i].err, err_table[i].err_sym);

and it works now.

regards

Vincent
-------------- next part --------------
An HTML attachment was scrubbed...
URL: </pipermail/attachments/20111002/eb458ce4/attachment.htm>

From vincent.torri at gmail.com  Mon Oct  3 05:02:35 2011
From: vincent.torri at gmail.com (Vincent Torri)
Date: Mon, 3 Oct 2011 05:02:35 +0200
Subject: libgpg-error : code-from-errno.h not correctly generated when
	cross-compiling to Windows on Linux
In-Reply-To: <CAMq1adoBDyWL6zBX2zUX=m2nDi3vhAKB+1MjnqgjBEzXy5PqsA@mail.gmail.com>
References: <CAMq1adoRb_aOXCzrDPGY9utk--qaq1zaanJqZq8ptb93cUvAXA@mail.gmail.com>
	<87hb3ugjr7.fsf@vigenere.g10code.de>
	<CAMq1adq6v=m+zVhs5qQ6Tf3ZyNt_rRwa2Nf+oPrZ+HJBrKZM6w@mail.gmail.com>
	<CAMq1adoBDyWL6zBX2zUX=m2nDi3vhAKB+1MjnqgjBEzXy5PqsA@mail.gmail.com>
Message-ID: <CAMq1adqjd62D0DQ+=dJ00reS-e5TvHLc7KgJm-JLgKxb2NQWFQ@mail.gmail.com>

On Sun, Oct 2, 2011 at 11:32 PM, Vincent Torri <vincent.torri at gmail.com>wrote:

>
>
> On Sat, Oct 1, 2011 at 7:34 AM, Vincent Torri <vincent.torri at gmail.com>wrote:
>
>>
>>
>> On Fri, Sep 30, 2011 at 4:58 PM, Werner Koch <wk at gnupg.org> wrote:
>>
>>> On Fri, 30 Sep 2011 10:02, vincent.torri at gmail.com said:
>>>
>>> > configuration: ./configure --prefix=/opt/efl --host=x86_64-w64-mingw32
>>> > --disable-nls --disable-static
>>>
>>
>> I actually have the same problem when cross compiling to win32 :
>>
>> configuration :  ./configure --prefix=/opt/efl --host=i686-w64-mingw32
>> --disable-nls --disable-static
>>
>> exactly the same problem
>>
>
> Just lookin at the output of ./mkerrcodes, I get:
>
> -2147168224    GPG_ERR_EPROTO
> 1    GPG_ERR_EPERM
> 2    GPG_ERR_ENOENT
> etc...
>
> So it seems that  there is a problem with the fact that GPG_ERR_EPROTO is
> negative (i think). So integer size overflow.
>
> I changed mkerrcodes.c so that it displays an unsigned int. The 2 changes
> lines are:
>
>   unsigned int i;
>

that change is not necessary, of course, but as i >= 0 anyway...

Vincent Torri


> ...
>     printf ("%u\t%s\n", err_table[i].err, err_table[i].err_sym);
>
> and it works now.
>
> regards
>
> Vincent
>
>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: </pipermail/attachments/20111003/9547425a/attachment.htm>

From cmouse at youzen.ext.b2.fi  Sun Oct  2 15:39:06 2011
From: cmouse at youzen.ext.b2.fi (Aki Tuomi)
Date: Sun, 2 Oct 2011 16:39:06 +0300
Subject: libgpg-error : code-from-errno.h not correctly generated when
	cross-compiling to Windows on Linux
In-Reply-To: <878vp3hihl.fsf@vigenere.g10code.de>
References: <CAMq1adoRb_aOXCzrDPGY9utk--qaq1zaanJqZq8ptb93cUvAXA@mail.gmail.com>
	<87hb3ugjr7.fsf@vigenere.g10code.de>
	<CAMq1adqX8d-U46CY5dm7Qjwn0=37S8H18kvmSr_km02E_Q1Mvw@mail.gmail.com>
	<878vp3hihl.fsf@vigenere.g10code.de>
Message-ID: <20111002133906.GA12006@pi.ip.fi>

On Sun, Oct 02, 2011 at 11:04:38AM +0200, Werner Koch wrote:
> On Fri, 30 Sep 2011 20:43, vincent.torri at gmail.com said:
> > Actually, I need libgcrypt for crypto in gnutls. So it means that,
> > currently, it's useless to try to compile libgcrypt for win64 ?
> 
> Don't use it.  We do not have an RNG for w64.
> 
> 
> Shalom-Salam,
> 
>    Werner
> 
> 
> -- 
> Die Gedanken sind frei.  Ausnahmen regelt ein Bundesgesetz.
>

I thought windows has a PRNG suited for cryptography bundled. Why can't you
use that? 

Aki Tuomi 
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 197 bytes
Desc: Digital signature
URL: </pipermail/attachments/20111002/fcef208e/attachment.pgp>

From wk at gnupg.org  Fri Oct  7 10:35:50 2011
From: wk at gnupg.org (Werner Koch)
Date: Fri, 07 Oct 2011 10:35:50 +0200
Subject: make check errors
In-Reply-To: <1CCAF9A3-64DD-416F-965F-B32F4A885174@1fehr.ch> (Markus Fehr's
	message of "Fri, 30 Sep 2011 00:36:24 +0200")
References: <1CCAF9A3-64DD-416F-965F-B32F4A885174@1fehr.ch>
Message-ID: <874nzlb3mh.fsf@vigenere.g10code.de>

On Fri, 30 Sep 2011 00:36, mailnov at 1fehr.ch said:

> After I compiled netatalk 2.2, I noticed an error message "symbol
> gcry_cipher_setkey, version GCRYPT_1.2 not defined in file
> libgcrypt.so.11 with link time reference". I then ran make check and
> found errors as well. The output looks as fo

netatalk tries to pick up a a version of Libgcrypt older than 1.4.2 -
check with ldd which libraries are loaded at runtime.


Shalom-Salam,

   Werner

-- 
Die Gedanken sind frei.  Ausnahmen regelt ein Bundesgesetz.



From ametzler at downhill.at.eu.org  Sun Oct  9 16:23:14 2011
From: ametzler at downhill.at.eu.org (Andreas Metzler)
Date: Sun, 9 Oct 2011 16:23:14 +0200
Subject: libgcrypt11 1.5.0 version segfaults with NSS/PAM LDAP (AES-NI)
Message-ID: <20111009142314.GB7633@downhill.g.la>

Hello,
this is http://bugs.debian.org/643336 reported by Marc Dequ?nes, the
crash was not easily reproducible with gnutls-cli. The respective
system features AES-NI support.

cu andreas

--------------------------------------------
Package: libgcrypt11
Version: 1.5.0-3
Severity: important


Coin,

I'm using:
  - libgnutls26  2.12.10-2
  - libldap-2.4-2  2.4.25-3
  - libnss-ldap  264-2.2

After an upgrade a mere "id <user>" lead to the following segfault:
#0  0xb72011cd in do_aesni_enc_aligned (
    a=0xb723a1b8 "\001K\257\"x\246\235\063\035Q\200\020\066C\351\232gC\303\321Q\232\264\362?x\253\t\245\021\275]\036\362\r\316?\274\022\023\032\307\305G\210\252\b\016\225\027\353\026wq\232\317r\200\206\004",
<incomplete sequence \343>, b=0xbfb28ad8
"(\335%\267p\213\262\277\004{\343\t", ctx=0xbfb288e8) at
rijndael.c:710
#1  do_aesni (ctx=0xbfb288e8, decrypt_flag=0, bx=0xbfb28ad8
"(\335%\267p\213\262\277\004{\343\t",
    ax=0xb723a1b8 "\001K\257\"x\246\235\063\035Q\200\020\066C\351\232gC\303\321Q\232\264\362?x\253\t\245\021\275]\036\362\r\316?\274\022\023\032\307\305G\210\252\b\016\225\027\353\026wq\232\317r\200\206\004",
<incomplete sequence \343>) at rijndael.c:1132
#2  0xb72014c6 in rijndael_encrypt (context=0xbfb288e8, b=0xbfb28ad8
"(\335%\267p\213\262\277\004{\343\t",
    a=0xb723a1b8 "\001K\257\"x\246\235\063\035Q\200\020\066C\351\232gC\303\321Q\232\264\362?x\253\t\245\021\275]\036\362\r\316?\274\022\023\032\307\305G\210\252\b\016\225\027\353\026wq\232\317r\200\206\004",
<incomplete sequence \343>) at rijndael.c:1155
#3  0xb7201aa8 in selftest_basic_128 () at rijndael.c:1660
#4  0xb7201657 in selftest () at rijndael.c:1749
#5  do_setkey (keylen=16, key=0x9ff3fa8
"\035\224<n\372KWy1\355\344y\260\332\064\031\030", ctx=0xa001e90)
    at rijndael.c:209
#6  rijndael_setkey (context=0xa001e90, key=0x9ff3fa8
"\035\224<n\372KWy1\355\344y\260\332\064\031\030",
    keylen=16) at rijndael.c:444
#7  0xb71e5ae7 in cipher_setkey (c=0xa001e10, key=<optimized out>,
keylen=16) at cipher.c:896
#8  0xb71dbe04 in gcry_cipher_setkey (hd=0xa001e10, key=0x9ff3fa8,
keylen=16) at visibility.c:521
#9  0xb72faceb in wrap_gcry_cipher_setkey (ctx=0xa001e10,
key=0x9ff3fa8, keysize=16) at cipher.c:115
#10 0xb727ea92 in _gnutls_cipher_init (handle=0x9ff485c,
cipher=GNUTLS_CIPHER_AES_128_CBC, key=0x9ff4854,
    iv=0x9ff484c) at gnutls_cipher_int.c:71
#11 0xb7289d05 in _gnutls_init_record_state (read=1, state=0x9ff4844,
params=<optimized out>)
    at gnutls_constate.c:299
#12 0xb728a2c1 in _gnutls_epoch_set_keys (session=0x9ff70c0, epoch=1)
at gnutls_constate.c:431
#13 0xb728a997 in _gnutls_write_connection_state_init
(session=0x9ff70c0) at gnutls_constate.c:602
#14 0xb7272fdd in _gnutls_send_handshake_final (session=0x9ff70c0,
init=1) at gnutls_handshake.c:2888
#15 0xb72765e0 in _gnutls_handshake_common (session=0x9ff70c0) at
gnutls_handshake.c:3121
#16 0xb7277fad in gnutls_handshake (session=0x9ff70c0) at
gnutls_handshake.c:2690
#17 0xb74bfdac in tlsg_session_accept (session=0x9ff70a8) at tls_g.c:472
#18 0xb74bcd39 in ldap_int_tls_connect (ld=0x9e3ca30, conn=<optimized
out>) at tls2.c:358
#19 0xb74bd917 in ldap_int_tls_start (ld=0x9e3ca30, conn=0x9e3c980,
srv=0x0) at tls2.c:825
#20 0xb74bdc79 in ldap_install_tls (ld=0x9e3ca30) at tls2.c:897
#21 0xb74dbc1d in ?? () from /lib/libnss_ldap.so.2
#22 0xb74dc251 in ?? () from /lib/libnss_ldap.so.2
#23 0xb74dcaa5 in ?? () from /lib/libnss_ldap.so.2
#24 0xb74dcdf1 in ?? () from /lib/libnss_ldap.so.2
#25 0xb74dd570 in _nss_ldap_getpwnam_r () from /lib/libnss_ldap.so.2
#26 0xb773e495 in getpwnam_r () from /lib/i386-linux-gnu/i686/cmov/libc.so.6
#27 0xb773deff in getpwnam () from /lib/i386-linux-gnu/i686/cmov/libc.so.6
#28 0x08049594 in ?? ()
#29 0xb76bee46 in __libc_start_main () from
/lib/i386-linux-gnu/i686/cmov/libc.so.6
#30 0x08049b49 in ?? ()
Backtrace stopped: Not enough registers or memory available to unwind further

Reverting to 1.4.6-9 fixed the problem.

Regards.


-- System Information:
Debian Release: wheezy/sid
  APT prefers unstable
  APT policy: (500, 'unstable'), (500, 'testing')
Architecture: i386 (i686)

Kernel: Linux 3.0.0-1-686-pae (SMP w/8 CPU cores)
--------------------------------------------


From mailnov at 1fehr.ch  Sun Oct  9 21:30:21 2011
From: mailnov at 1fehr.ch (Markus Fehr)
Date: Sun, 9 Oct 2011 21:30:21 +0200
Subject: make check errors
In-Reply-To: <874nzlb3mh.fsf@vigenere.g10code.de>
References: <1CCAF9A3-64DD-416F-965F-B32F4A885174@1fehr.ch>
	<874nzlb3mh.fsf@vigenere.g10code.de>
Message-ID: <106CACD2-4D11-466E-8F0E-961051706909@1fehr.ch>

Perfect! Problem solved: ldd pointed out that it used runtime libraries located in /lib while libgpg-error and libgcrypt were installed to /usr/local/lib

Thanks,
Markus

On 7. Oct 2011, at 10:35, Werner Koch wrote:

> On Fri, 30 Sep 2011 00:36, mailnov at 1fehr.ch said:
> 
>> After I compiled netatalk 2.2, I noticed an error message "symbol
>> gcry_cipher_setkey, version GCRYPT_1.2 not defined in file
>> libgcrypt.so.11 with link time reference". I then ran make check and
>> found errors as well. The output looks as fo
> 
> netatalk tries to pick up a a version of Libgcrypt older than 1.4.2 -
> check with ldd which libraries are loaded at runtime.
> 
> 
> Shalom-Salam,
> 
>   Werner
> 
> -- 
> Die Gedanken sind frei.  Ausnahmen regelt ein Bundesgesetz.
> 



From wk at gnupg.org  Tue Oct 11 10:01:22 2011
From: wk at gnupg.org (Werner Koch)
Date: Tue, 11 Oct 2011 10:01:22 +0200
Subject: libgcrypt11 1.5.0 version segfaults with NSS/PAM LDAP (AES-NI)
In-Reply-To: <20111009142314.GB7633@downhill.g.la> (Andreas Metzler's message
	of "Sun, 9 Oct 2011 16:23:14 +0200")
References: <20111009142314.GB7633@downhill.g.la>
Message-ID: <87k48c7y99.fsf@vigenere.g10code.de>

On Sun,  9 Oct 2011 16:23, ametzler at downhill.at.eu.org said:

> this is http://bugs.debian.org/643336 reported by Marc Dequ?nes, the
> crash was not easily reproducible with gnutls-cli. The respective
> system features AES-NI support.

Please try to break this down to a little test program without LDAP and
without GNUTLS.


Shalom-Salam,

   Werner

-- 
Die Gedanken sind frei.  Ausnahmen regelt ein Bundesgesetz.



From kurt.w.heberlein at hp.com  Fri Oct 21 18:57:05 2011
From: kurt.w.heberlein at hp.com (Heberlein, Kurt William)
Date: Fri, 21 Oct 2011 17:57:05 +0100
Subject: FIPS-186 CAVP verification tests
Message-ID: <EA7CA6F14D28E947BC54C01BDDECDFAA77CE6874FA@GVW0670EXC.americas.hpqcorp.net>

Hi,
   Does anyone have a good set of  tests for running against the RSAVS test vectors?   I have been successful writing my own to do the Key Generation tests (validating the result against the openssl tests against the same vectors).  The signature generation test, likewise seems to function ok, but I  cannot seem to get the signature verification tests to succeed.   The verification always fails, even when I know it should succeed.    Thanks in advance.


Kurt Heberlein
Senior Staff Software Engineer
3PAR R&D
   HP Storage Works Division
   www.hp.com<http://www.hp.com/>

   (510) 668 9441

    kurt dot w dot heberlein at-sign hp dot com


   [cid:image001.png at 01CC8FE8.8D433B80]


-------------- next part --------------
An HTML attachment was scrubbed...
URL: </pipermail/attachments/20111021/79c6a05a/attachment-0001.htm>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: image001.png
Type: image/png
Size: 11657 bytes
Desc: image001.png
URL: </pipermail/attachments/20111021/79c6a05a/attachment-0001.png>

From wk at gnupg.org  Sat Oct 22 07:05:21 2011
From: wk at gnupg.org (Werner Koch)
Date: Sat, 22 Oct 2011 07:05:21 +0200
Subject: FIPS-186 CAVP verification tests
In-Reply-To: <EA7CA6F14D28E947BC54C01BDDECDFAA77CE6874FA@GVW0670EXC.americas.hpqcorp.net>
	(Kurt William Heberlein's message of "Fri, 21 Oct 2011 17:57:05
	+0100")
References: <EA7CA6F14D28E947BC54C01BDDECDFAA77CE6874FA@GVW0670EXC.americas.hpqcorp.net>
Message-ID: <87zkgtvcoe.fsf@vigenere.g10code.de>

On Fri, 21 Oct 2011 18:57, kurt.w.heberlein at hp.com said:

>    Does anyone have a good set of tests for running against the RSAVS
>    test vectors?  I have been successful writing my own to do the Key

Libgcrypt comes with support for the CAVS tests; see tests/cavs_tests.sh
.  AFAIK the test vectors are not redistributable.


Salam-Shalom,

   Werner


-- 
Die Gedanken sind frei.  Ausnahmen regelt ein Bundesgesetz.