RE: FIPS error on Apache httpd v2.4.3, OpenSSL 1.0.1c and fips-2.0.1

2012-08-27 Thread Ruiyuan Jiang
Hi, Cassie

I followed your post. I tried to recompile Apache with the recommendation that 
you given. I tried to rename Redhat's libcrypto and libssl to something else 
then Apache complains about LDAP library missing in the configure phase. I then 
tried using LDFLAGS for "configure" in Apache but no success. I also tried 
LD_LIBRARY_PATH to specify /usr/local/ssl/lib but also no luck. What else the 
trick that I can use? Thanks.

Ryan

-Original Message-
From: owner-openssl-us...@openssl.org [mailto:owner-openssl-us...@openssl.org] 
On Behalf Of Helms, Cassie
Sent: Thursday, August 23, 2012 3:26 PM
To: openssl-users@openssl.org
Subject: RE: FIPS error on Apache httpd v2.4.3, OpenSSL 1.0.1c and fips-2.0.1

Ryan,
A previous thread, "fingerprint does not match on FIPS_mode_set when FIPS + 
openssl is 
dynamically linked into build", might be of some use to you. As a first step, 
you may
want to use ldd on your executable to make sure libcrypto.so/a points to 1.0.1c 
and not
some other version of openssl. The thread has more information on this issue.

Cassie
__
OpenSSL Project http://www.openssl.org
User Support Mailing Listopenssl-users@openssl.org
Automated List Manager   majord...@openssl.org



This message (including any attachments) is intended
solely for the specific individual(s) or entity(ies) named
above, and may contain legally privileged and
confidential information. If you are not the intended 
recipient, please notify the sender immediately by 
replying to this message and then delete it.
Any disclosure, copying, or distribution of this message,
or the taking of any action based on it, by other than the
intended recipient, is strictly prohibited.
__
OpenSSL Project http://www.openssl.org
User Support Mailing Listopenssl-users@openssl.org
Automated List Manager   majord...@openssl.org


FIPS enabled OpenSSL v1.0.1c

2012-08-27 Thread Ruiyuan Jiang
Hi, 

When I tried to start Apache(v2.4.3) with FIPS enabled OpenSSL v1.0.1c on RHEL 
v6.3, I was prompted for the pass phrase which is normal. After I typed in 
correct pass phrase, I got a message:

Apache: mod_ssl:Error: Pass phrase incorrect (5 more retries permitted).

When I ctrl-c to exist, I got another message:

Apache:mod_ssl:Error: Private key not found.

Which is not correct since the private key is there. The key and certificate 
was generated by older version of FIPS disabled OpenSSL. I copied the key and 
certificate from older version of web server for the new version web server to 
use. Once I disabled FIPS in the Apache configuration file, I typed in the same 
pass phrase and I can start httpd v2.4.3. Is this something from Apache or 
OpenSSL side for the wrong pass phrase prompt? What else do I need to do or 
check? Thanks.

Ryan Jiang



This message (including any attachments) is intended
solely for the specific individual(s) or entity(ies) named
above, and may contain legally privileged and
confidential information. If you are not the intended 
recipient, please notify the sender immediately by 
replying to this message and then delete it.
Any disclosure, copying, or distribution of this message,
or the taking of any action based on it, by other than the
intended recipient, is strictly prohibited.
__
OpenSSL Project http://www.openssl.org
User Support Mailing Listopenssl-users@openssl.org
Automated List Manager   majord...@openssl.org


Re: Openssh error - Solaris 10 SPARC Platform

2012-08-27 Thread Gaiseric Vandal
Sounds like an LD_LIBRARY_PATH issue.  In /etc/default/profile you may
wabt to make sure that /usr/local/lib is set before /usr/lib in
LD_LIBRARY_PATH.  Solaris should include its own vers of ssh and ssl
which will account for the conflict.



On 08/23/12 19:06, Roberto Ballan wrote:
> Hi,
> I have a Sun Solaris 10 Server that I have upgraded the Openssh
> version 5.3p1 to version 6.0p1.
> I have upgraded openssl from version  0.9.81 to version 1.0.1c.
> When I check the ssh log file I see the error.
> Please can you provide some help how to resolve this issue ?
> ==
> [ Aug 23 16:19:03 Executing start method ("/lib/svc/method/sshd start") ]
> starting /usr/local/sbin/sshd... *OpenSSL version mismatch. Built
> against 1000103f, you have 10af*
> /lib/svc/method/sshd: Error 255 starting /usr/local/sbin/sshd... bailing.
> [ Aug 23 16:19:03 Method "start" exited with status 255 ]
> 
> Thanks.
> Roberto Ballan
> email: roberto.bal...@macys.com
> 




Re: Convert symmetrically encrypted content to base64

2012-08-27 Thread Bjoern Schiessle
On Fri, 24 Aug 2012 15:54:50 -0400 Dave Thompson wrote:
> Note OpenSSL's RSA privatekey *includes* publickey.
> RSA publickey is n,e and naive privatekey is n,d, 
> but OpenSSL privatekey is CRT form with n,d,e,p,q + more.
> There is no need to transmit the publickey separately, 
>  
>  [..]
> 
>
> Tiny aside: BIO_new_mem_buf will do the strlen() for you 
> if you pass -1 for length. Just a convenience.
> 
> [..]
>
> If PEM_read_* returns null (or nearly any other OpenSSL 
> routine returns a failure indication), look at the error queue.
> http://www.openssl.org/support/faq.html#PROG6
> and #PROG7 also if you don't get readable error.
> 
> If they didn't, look very carefully at your PEM data. 
> Commandline can do this: openssl asn1parse -in myprivkey.pem 
> and/or: openssal rsa -in myprivkey.pem -text


Thanks for your hints. After a lot of testing I figured out
that my functions pem2key() and key2pem() works fine. The problem is
that I lose some characters (e.g. '+' gets replaced by spaces) while
sending the key over the network. But I think this problem don't belong
to the mailing list. ;-)

Thanks a lot!
Björn


-- 
Björn Schießle 
www: http://schiessle.org 
gnupg key: 0x0x2378A753E2BF04F6 
fingerprint: 244F CEB0 CB09 9524 B21F B896 2378 A753 E2BF 04F6
__
OpenSSL Project http://www.openssl.org
User Support Mailing Listopenssl-users@openssl.org
Automated List Manager   majord...@openssl.org


Why key file in *client* certificate situation?

2012-08-27 Thread Charles Mills
I'm just trying to understand the SSL protocol -- this is not an alleged bug
or an "issue."

In OpenSSL s_client, or for that matter, in my client test program, an
attempt to use a *client* certificate fails unless I also specify -key or
call SSL_CTX_use_PrivateKey_file(). 

Why? What role does the private key play with a *client* certificate? My
understanding -- admittedly perhaps flawed -- is that the role of a client
certificate is solely to authenticate the client. Isn't that role complete
with just a CA-signed certificate? There's no encryption based on the client
certificate, right? So what role does the key play? If none, why does
OpenSSL fail without it?

Thanks,
Charles 


__
OpenSSL Project http://www.openssl.org
User Support Mailing Listopenssl-users@openssl.org
Automated List Manager   majord...@openssl.org


Re: Why key file in *client* certificate situation?

2012-08-27 Thread Jakob Bohm

On 8/27/2012 3:46 PM, Charles Mills wrote:

I'm just trying to understand the SSL protocol -- this is not an 
alleged bug


or an "issue."


In OpenSSL s_client, or for that matter, in my client test program, an

attempt to use a *client* certificate fails unless I also specify -key or

call SSL_CTX_use_PrivateKey_file().


Why? What role does the private key play with a *client* certificate? My

understanding -- admittedly perhaps flawed -- is that the role of a client

certificate is solely to authenticate the client. Isn't that role complete

with just a CA-signed certificate? There's no encryption based on the 
client


certificate, right? So what role does the key play? If none, why does

OpenSSL fail without it?



Basic principle: A certificate is not secret, it is a public statement
by a CA that a public key matches a private key belonging to a certain
person or other entity.

So just sending the client *certificate* to the server would prove
nothing and is not useful as authentication.  Just as handing someone
a (paper) business card doesn't prove it is *your* business card.

Signing some part of the SSL exchange with the clients private key and
sending along the certificate to tell the server what the public key is
and as proof of what identity is proven by the signature does prove a
lot.  So that is what SSL does.  And that is why an SSL client needs
the private key of the client certificate (if any).

Enjoy

Jakob

--
Jakob Bohm, CIO, Partner, WiseMo A/S.  http://www.wisemo.com
Transformervej 29, 2730 Herlev, Denmark.  Direct +45 31 13 16 10
This public discussion message is non-binding and may contain errors.
WiseMo - Remote Service Management for PCs, Phones and Embedded
__
OpenSSL Project http://www.openssl.org
User Support Mailing Listopenssl-users@openssl.org
Automated List Manager   majord...@openssl.org


RE: Why key file in *client* certificate situation?

2012-08-27 Thread Charles Mills
Thanks. I think I get it.

Charles

-Original Message-
From: owner-openssl-us...@openssl.org
[mailto:owner-openssl-us...@openssl.org] On Behalf Of Jakob Bohm
Sent: Monday, August 27, 2012 10:19 AM
To: openssl-users@openssl.org
Subject: Re: Why key file in *client* certificate situation?

On 8/27/2012 3:46 PM, Charles Mills wrote:

> I'm just trying to understand the SSL protocol -- this is not an 
> alleged bug
>
> or an "issue."
>
>
> In OpenSSL s_client, or for that matter, in my client test program, an
>
> attempt to use a *client* certificate fails unless I also specify -key 
> or
>
> call SSL_CTX_use_PrivateKey_file().
>
>
> Why? What role does the private key play with a *client* certificate? 
> My
>
> understanding -- admittedly perhaps flawed -- is that the role of a 
> client
>
> certificate is solely to authenticate the client. Isn't that role 
> complete
>
> with just a CA-signed certificate? There's no encryption based on the 
> client
>
> certificate, right? So what role does the key play? If none, why does
>
> OpenSSL fail without it?
>
>
Basic principle: A certificate is not secret, it is a public statement by a
CA that a public key matches a private key belonging to a certain person or
other entity.

So just sending the client *certificate* to the server would prove nothing
and is not useful as authentication.  Just as handing someone a (paper)
business card doesn't prove it is *your* business card.

Signing some part of the SSL exchange with the clients private key and
sending along the certificate to tell the server what the public key is and
as proof of what identity is proven by the signature does prove a lot.  So
that is what SSL does.  And that is why an SSL client needs the private key
of the client certificate (if any).

__
OpenSSL Project http://www.openssl.org
User Support Mailing Listopenssl-users@openssl.org
Automated List Manager   majord...@openssl.org


openssl smime verify fails in ASN1_CHECK_TLEN but asn1parse is ok?

2012-08-27 Thread GWu
Hello,

I'm trying to verify an email signature using openssl.

I've saved the complete mail to a file named mail.eml, then I'm using
openssl to verify:

openssl smime -inform SMIME -CAfile all.pem -verify -in mail.eml

which gives an error:

2674688:error:0D0680A8:asn1 encoding routines:ASN1_CHECK_TLEN:wrong
tag:tasn_dec.c:1319:
2674688:error:0D07803A:asn1 encoding routines:ASN1_ITEM_EX_D2I:nested
asn1 error:tasn_dec.c:381:Type=X509_SIG
2674688:error:21071069:PKCS7 routines:PKCS7_signatureVerify:signature
failure:pk7_doit.c:1120:
2674688:error:21075069:PKCS7 routines:PKCS7_verify:signature
failure:pk7_smime.c:410:

Now I've also saved the base64 encoded signature only from the mime
part into a separate file named mail_sig.txt to analyse further:

openssl asn1parse -i -in mail_sig.txt

which runs ok (at least there is no error shown) and prints out a nice
tree, which at least to me does not show any errors (I'm a novice in
this...).

I'm using OpenSSL Version 1.0.1 14 Mar 2012

Any advice what may be wrong with the signature or how verify it in
openssl? If needed, I can upload the e-mail or complete asn1parse
output to pastebin or similar.

Thanks for any help,
 Georg

PS: I have to note that the complete story is longer: the email
signature verifies fine in MS Outlook, but not in Thunderbird, so I
tried to analyse this behavior and stumbled over the fact that openssl
also has troubles verifying it. And Thunderbird most probably uses
openssl libraries internally, so I've ended up here.
__
OpenSSL Project http://www.openssl.org
User Support Mailing Listopenssl-users@openssl.org
Automated List Manager   majord...@openssl.org


Re: openssl smime verify fails in ASN1_CHECK_TLEN but asn1parse is ok?

2012-08-27 Thread Dr. Stephen Henson
On Mon, Aug 27, 2012, GWu wrote:

> Hello,
> 
> I'm trying to verify an email signature using openssl.
> 
> I've saved the complete mail to a file named mail.eml, then I'm using
> openssl to verify:
> 
> openssl smime -inform SMIME -CAfile all.pem -verify -in mail.eml
> 
> which gives an error:
> 
> 2674688:error:0D0680A8:asn1 encoding routines:ASN1_CHECK_TLEN:wrong
> tag:tasn_dec.c:1319:
> 2674688:error:0D07803A:asn1 encoding routines:ASN1_ITEM_EX_D2I:nested
> asn1 error:tasn_dec.c:381:Type=X509_SIG
> 2674688:error:21071069:PKCS7 routines:PKCS7_signatureVerify:signature
> failure:pk7_doit.c:1120:
> 2674688:error:21075069:PKCS7 routines:PKCS7_verify:signature
> failure:pk7_smime.c:410:
> 
> Now I've also saved the base64 encoded signature only from the mime
> part into a separate file named mail_sig.txt to analyse further:
> 
> openssl asn1parse -i -in mail_sig.txt
> 
> which runs ok (at least there is no error shown) and prints out a nice
> tree, which at least to me does not show any errors (I'm a novice in
> this...).
> 
> I'm using OpenSSL Version 1.0.1 14 Mar 2012
> 
> Any advice what may be wrong with the signature or how verify it in
> openssl? If needed, I can upload the e-mail or complete asn1parse
> output to pastebin or similar.
> 

It sounds like the signature is malformed. That wouldn't cause problems with
asn1parse but would when OpenSSL tried to verify it.

I'd need to see the email to be sure though.

> 
> PS: I have to note that the complete story is longer: the email
> signature verifies fine in MS Outlook, but not in Thunderbird, so I
> tried to analyse this behavior and stumbled over the fact that openssl
> also has troubles verifying it. And Thunderbird most probably uses
> openssl libraries internally, so I've ended up here.

Thunderbird doesn't use OpenSSL it uses its own cryptographic library called
NSS.

Steve.
--
Dr Stephen N. Henson. OpenSSL project core developer.
Commercial tech support now available see: http://www.openssl.org
__
OpenSSL Project http://www.openssl.org
User Support Mailing Listopenssl-users@openssl.org
Automated List Manager   majord...@openssl.org


Re: openssl smime verify fails in ASN1_CHECK_TLEN but asn1parse is ok?

2012-08-27 Thread GWu
On Mon, Aug 27, 2012 at 9:27 PM, Dr. Stephen Henson wrote:
> On Mon, Aug 27, 2012, GWu wrote:
>> [...]
>> openssl smime -inform SMIME -CAfile all.pem -verify -in mail.eml
>> which gives an error:
>> [...]
>
> It sounds like the signature is malformed. That wouldn't cause problems with
> asn1parse but would when OpenSSL tried to verify it.
>
> I'd need to see the email to be sure though.

The email is available at
http://www.buergerkarte.at/mvnforum/mvnforum/viewthread_thread,272#1180
(German language forum, but the email - or it's significant parts
respectively - is easily visble).

>> PS: [...] And Thunderbird most probably uses
>> openssl libraries internally, so I've ended up here.
>
> Thunderbird doesn't use OpenSSL it uses its own cryptographic library called
> NSS.

I wasn't aware of that, thanks for the clarification.

kind regards,
 Georg
__
OpenSSL Project http://www.openssl.org
User Support Mailing Listopenssl-users@openssl.org
Automated List Manager   majord...@openssl.org


Re: openssl smime verify fails in ASN1_CHECK_TLEN but asn1parse is ok?

2012-08-27 Thread Dr. Stephen Henson
On Mon, Aug 27, 2012, GWu wrote:

> On Mon, Aug 27, 2012 at 9:27 PM, Dr. Stephen Henson wrote:
> > On Mon, Aug 27, 2012, GWu wrote:
> >> [...]
> >> openssl smime -inform SMIME -CAfile all.pem -verify -in mail.eml
> >> which gives an error:
> >> [...]
> >
> > It sounds like the signature is malformed. That wouldn't cause problems with
> > asn1parse but would when OpenSSL tried to verify it.
> >
> > I'd need to see the email to be sure though.
> 
> The email is available at
> http://www.buergerkarte.at/mvnforum/mvnforum/viewthread_thread,272#1180
> (German language forum, but the email - or it's significant parts
> respectively - is easily visble).
> 

Well I'm not surprised Thunderbird and OpenSSL has problems with that. The
signature is malformed. It should contain the digest enclosed in an ASN1
structure called DigestInfo but it isn't: it just contains the raw digest.

Steve.
--
Dr Stephen N. Henson. OpenSSL project core developer.
Commercial tech support now available see: http://www.openssl.org
__
OpenSSL Project http://www.openssl.org
User Support Mailing Listopenssl-users@openssl.org
Automated List Manager   majord...@openssl.org


FW: OpenSSL on beagleboard

2012-08-27 Thread Paulo Roberto

Can no one help me? Isn't there a way of specifying the local the openssl is 
installed?
I need very much to make it works.

Thanks everybody.

From: bad_boy_...@hotmail.com
To: openssl-users@openssl.org
Subject: OpenSSL on beagleboard
Date: Thu, 23 Aug 2012 22:06:59 -0300





Hello, I am using the package libssl-dev on ubuntu in my beagleboard xm, and I 
have to run two C algorithms using the openSSL library..
Although I can't compile using the command: gcc test.c -lssl -o test. It seems 
the compiler isn't recognizing the "-lssl" command.
Does someone know how to solve this?
Do I have to set some path, or something like that?

Thanks Guys.

  

RE: OpenSSL on beagleboard

2012-08-27 Thread Paulo Roberto

Thanks for helping jeff, but it haven't worked yet.
I searched my libssl.so in my /usr/lib and I didn't find.
Does someone have any idea? I have installed the libssl-dev, libssl0.9.8.

Thanks for helping.

> Date: Thu, 23 Aug 2012 21:18:37 -0400
> Subject: Re: OpenSSL on beagleboard
> From: noloa...@gmail.com
> To: openssl-users@openssl.org
> 
> On Thu, Aug 23, 2012 at 9:06 PM, Paulo Roberto  
> wrote:
> > Hello, I am using the package libssl-dev on ubuntu in my beagleboard xm, and
> > I have to run two C algorithms using the openSSL library..
> > Although I can't compile using the command: gcc test.c -lssl -o test. It
> > seems the compiler isn't recognizing the "-lssl" command.
> > Does someone know how to solve this?
> > Do I have to set some path, or something like that?
> You specify linker commands (such as libraries) at the very end of the
> compiler drive command. From the g++ man pages (around line 25):
> "...the placement of the -l option is significant."
> 
> gcc test.c -o test -lssl
> 
> You might also want to add -Wl,-Bstatic unless you want to do the
> shared object thing.
> 
> Jeff
> __
> OpenSSL Project http://www.openssl.org
> User Support Mailing Listopenssl-users@openssl.org
> Automated List Manager   majord...@openssl.org
  

RE: OpenSSL on beagleboard

2012-08-27 Thread Paulo Roberto

When I use the command gcc teste.c -lssl -o teste: 

Error:
ubuntu@omap:~/arquivos$ gcc rsa.c -lssl -o teste
/tmp/ccyvrO2i.o: In function `main':
rsa.c:(.text+0x8): undefined reference to `BN_new'
rsa.c:(.text+0xe): undefined reference to `BN_new'
rsa.c:(.text+0x14): undefined reference to `BN_new'
rsa.c:(.text+0x1a): undefined reference to `BN_new'
rsa.c:(.text+0x20): undefined reference to `BN_new'
/tmp/ccyvrO2i.o:rsa.c:(.text+0x26): more undefined references to `BN_new' follow
/tmp/ccyvrO2i.o: In function `main':
rsa.c:(.text+0x5c): undefined reference to `BN_set_word'
rsa.c:(.text+0x66): undefined reference to `BN_set_word'
rsa.c:(.text+0x70): undefined reference to `BN_set_word'
rsa.c:(.text+0x98): undefined reference to `BN_generate_prime'
rsa.c:(.text+0xba): undefined reference to `BN_generate_prime'
rsa.c:(.text+0xbe): undefined reference to `BN_CTX_new'
rsa.c:(.text+0xca): undefined reference to `BN_mul'
rsa.c:(.text+0xd4): undefined reference to `BN_sub'
rsa.c:(.text+0xe2): undefined reference to `BN_bn2dec'
rsa.c:(.text+0xf6): undefined reference to `BN_sub'
rsa.c:(.text+0xfa): undefined reference to `BN_CTX_new'
rsa.c:(.text+0x106): undefined reference to `BN_mul'
rsa.c:(.text+0x114): undefined reference to `BN_bn2dec'
rsa.c:(.text+0x11c): undefined reference to `BN_bn2dec'
rsa.c:(.text+0x124): undefined reference to `BN_bn2dec'
rsa.c:(.text+0x140): undefined reference to `BN_rand'
rsa.c:(.text+0x1b8): undefined reference to `BN_mul'
rsa.c:(.text+0x1c6): undefined reference to `BN_bn2dec'
rsa.c:(.text+0x1de): undefined reference to `BN_bn2dec'
rsa.c:(.text+0x1f6): undefined reference to `BN_bn2dec'
rsa.c:(.text+0x204): undefined reference to `BN_CTX_new'
rsa.c:(.text+0x216): undefined reference to `BN_div'
rsa.c:(.text+0x224): undefined reference to `BN_bn2dec'
rsa.c:(.text+0x236): undefined reference to `BN_cmp'
rsa.c:(.text+0x24a): undefined reference to `BN_bn2dec'
rsa.c:(.text+0x252): undefined reference to `BN_bn2dec'
rsa.c:(.text+0x25a): undefined reference to `BN_bn2dec'
rsa.c:(.text+0x262): undefined reference to `BN_bn2dec'
rsa.c:(.text+0x26a): undefined reference to `BN_bn2dec'
collect2: ld returned 1 exit status

Another attempt: 

ubuntu@omap:~/arquivos$ gcc rsa.c -Wl, -lssl, -lcrypto -o teste
/usr/bin/ld: cannot find : No such file or directory
/usr/bin/ld: cannot find -lssl,
collect2: ld returned 1 exit status

Any idea? 

I haven't found the libssl.so in my directory /usr/lib.

Thanks everybody.


> Date: Fri, 24 Aug 2012 09:45:15 +0100
> Subject: Re: OpenSSL on beagleboard
> From: b...@links.org
> To: openssl-users@openssl.org
> 
> On Fri, Aug 24, 2012 at 2:18 AM, Jeffrey Walton  wrote:
> > On Thu, Aug 23, 2012 at 9:06 PM, Paulo Roberto  
> > wrote:
> >> Hello, I am using the package libssl-dev on ubuntu in my beagleboard xm, 
> >> and
> >> I have to run two C algorithms using the openSSL library..
> >> Although I can't compile using the command: gcc test.c -lssl -o test. It
> >> seems the compiler isn't recognizing the "-lssl" command.
> 
> You really need to show the error, I doubt it is "not recognizing" it.
> 
> >> Does someone know how to solve this?
> >> Do I have to set some path, or something like that?
> 
> You might do, use -L for this.
> 
> > You specify linker commands (such as libraries) at the very end of the
> > compiler drive command. From the g++ man pages (around line 25):
> > "...the placement of the -l option is significant."
> 
> Significant relative to .o and other libraries, not to the output
> file, so this should make no difference.
> 
> >
> > gcc test.c -o test -lssl
> >
> > You might also want to add -Wl,-Bstatic unless you want to do the
> > shared object thing.
> >
> > Jeff
> > __
> > OpenSSL Project http://www.openssl.org
> > User Support Mailing Listopenssl-users@openssl.org
> > Automated List Manager   majord...@openssl.org
> __
> OpenSSL Project http://www.openssl.org
> User Support Mailing Listopenssl-users@openssl.org
> Automated List Manager   majord...@openssl.org
  

RE: OpenSSL on beagleboard

2012-08-27 Thread Dave Thompson
>From: owner-openssl-us...@openssl.org On Behalf Of Paulo Roberto
>Sent: Monday, 27 August, 2012 18:37

>Can no one help me? Isn't there a way of specifying the local 
>the openssl is installed?

You mean "location" i.e. in the file system? As far as I know 
packages on most Linuxes, including ubuntu, install into system 
standard locations, maybe something like /usr/lib or /opt/ssl .
"System" locations vary for different Unixes but are usually the same 
for different libraries on the same Unix, so you might be better off 
asking people who know about ubuntu in particular.

If it is in a standard location, your gcc (assuming your gcc is 
the correct package of gcc for your Unix) should find it. 
If you have some kind of nonstandard perhaps experimental package, 
and you know where it is installed, -L/path/to/libdir is the gcc way 
to specify additional library directories. (The -Ldir must occur 
before the/any -llib that's lower-ell that needs it.) 

You always need to make sure you *compile* with headers that are 
compatible with the library(ies) you *link* with; if you are linking 
with libraries in a nonstandard location you may need -I/path/to/incdir 
that's upper-eye (for gcc) to find the corresponding headers. I'm not sure 
if -I must precede the sourcefile(s) but I put it there for clarity.

And if you have a non-package build, such as one you did yourself 
from source, both of these (-L and -I) are much more likely.

>From: bad_boy_...@hotmail.com
>Date: Thu, 23 Aug 2012 22:06:59 -0300

>Hello, I am using the package libssl-dev on ubuntu in my beagleboard xm, 
>and I have to run two C algorithms using the openSSL library..
>   Although I can't compile using the command: gcc test.c -lssl -o
test. 
>It seems the compiler isn't recognizing the "-lssl" command.
>Does someone know how to solve this?
>Do I have to set some path, or something like that?

You haven't obeyed the response Friday asking for the actual error.
There are many possible errors and exactly which one you have may 
point to very different causes and solutions. 


__
OpenSSL Project http://www.openssl.org
User Support Mailing Listopenssl-users@openssl.org
Automated List Manager   majord...@openssl.org


Re: Generation ECDHE parameters

2012-08-27 Thread Varma Dantuluri
Thanks Steve for the response. That was very useful information.

Thanks
Varma

On Thu, Aug 23, 2012 at 6:05 AM, Dr. Stephen Henson wrote:

> On Wed, Aug 22, 2012, Varma Dantuluri wrote:
>
> > Hi
> >
> > We are in the process of adding support for ECDSA-ECDHE cipher suites and
> > hence ECDSA certificates to our server.
> >
> > Right now, the server does the following:
> >
> > 1) Assign the ECDSA certificate to the SSL_CTX.
> > 2) Set the callback for ECDH parameter generation using
> > SSL_CTX_set_tmp_ecdh_callback.
> >
> > In ssl3_send_server_key_exchange, when this callback is called, the value
> > of 'keylength' parameter is always either 512 or 1024. Shouldnt
> 'keylength'
> > have the curve name or id in the case of ECDH? Are we doing something
> wrong
> > here?
> >
>
> No, it's a limitation in some versions of OpenSSL. You basically have to
> pick
> a curve you think the peer will support, P-256 is usually a safe choice. If
> the peer doesn't support it then ECDHE will be disabled. You might as as
> well
> set the curve using SSL_CTX_set_tmp_ecdh instead of the callback.
>
> This is fixed in the development version of OpenSSL: for that you can just
> set
> it to automatically use the right curve based on client and server
> preferences.
>
> Steve.
> --
> Dr Stephen N. Henson. OpenSSL project core developer.
> Commercial tech support now available see: http://www.openssl.org
> __
> OpenSSL Project http://www.openssl.org
> User Support Mailing Listopenssl-users@openssl.org
> Automated List Manager   majord...@openssl.org
>


RE: OpenSSL on beagleboard

2012-08-27 Thread Dave Thompson
>From: owner-openssl-us...@openssl.org On Behalf Of Paulo Roberto
>Sent: Monday, 27 August, 2012 20:21

Okay, this time you did post the error.

>When I use the command gcc teste.c -lssl -o teste: 
>/tmp/ccyvrO2i.o: In function `main':
>rsa.c:(.text+0x8): undefined reference to `BN_new'


BN_* are in libcrypto not libssl.

>Another attempt: 

>ubuntu@omap:~/arquivos$ gcc rsa.c -Wl, -lssl, -lcrypto -o teste
>/usr/bin/ld: cannot find : No such file or directory
>/usr/bin/ld: cannot find -lssl,
>collect2: ld returned 1 exit status

That's closer, but you've mixed up two different things.
-Wl,x,y -- comma(s) and NO space -- passes "opaque" linker 
arguments, like -Wl,-Bstatic in a previous response.
-lxxx is not opaque (driver knows it) so don't use -Wl .
You do want -lssl -lcrypto (space but no comma).

>Any idea? 

If you need both SSL and "low-level" routines like BN_*,
use -lssl -lcrypto . If only need low-level, just -lcrypto .
If you need both depending on your linker order may matter, 
in which case -lssl should be first.

>I haven't found the libssl.so in my directory /usr/lib.

Then that's (probably) not the right location on your Unix/distro. 
If it is dynamic, on Linux at least, ldd should work. Do 
  ldd `which openssl`
and look to see where libssl.* and libcrypto.* are.


__
OpenSSL Project http://www.openssl.org
User Support Mailing Listopenssl-users@openssl.org
Automated List Manager   majord...@openssl.org


Re: openssl smime verify fails in ASN1_CHECK_TLEN but asn1parse is ok?

2012-08-27 Thread GWu
On Mon, Aug 27, 2012 at 10:50 PM, Dr. Stephen Henson wrote:
> On Mon, Aug 27, 2012, GWu wrote:
>> The email is available at
>> http://www.buergerkarte.at/mvnforum/mvnforum/viewthread_thread,272#1180
>> (German language forum, but the email - or it's significant parts
>> respectively - is easily visble).
>>
>
> Well I'm not surprised Thunderbird and OpenSSL has problems with that. The
> signature is malformed. It should contain the digest enclosed in an ASN1
> structure called DigestInfo but it isn't: it just contains the raw digest.

Can you give me a hint where to find DigestInfo or the offending raw
digest in the result of asn1parse? I can spot messageDigest, which is
shown as:

 3957:d=7  hl=2 l=   9 prim:OBJECT:messageDigest
 3968:d=7  hl=2 l=  22 cons:SET
 3970:d=8  hl=2 l=  20 prim: OCTET STRING  [HEX
DUMP]:38BA6AE720F09EFFB46BC8859293743DD13EDBF0

But this looks very similar in a message which verifies successfully.
The asn1parse output of another, successfully verified signature also
does not contain "DigestInfo".

Is DigestInfo a structure inside messageDigest in asn1parse output? If
yes, is there a way to display it in structured/readable form? Or did
you mean that the content inside of messageDigest is not encoded
properly?

Thanks for any advice and please excuse my beginner's questions, I'm
trying to get a grip on these things ...
__
OpenSSL Project http://www.openssl.org
User Support Mailing Listopenssl-users@openssl.org
Automated List Manager   majord...@openssl.org