Re: add hash signature as part of the stream on which this hash is based on

2012-11-20 Thread Michael Zintakis


My guess is that if you could write a hash working the way you say, it 
would be vulnerable to all sorts of cryptographic attacks: give up!
I have indeed given up and found other ways to incorporate the hash 
while verifying the stream integrity.


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


EVP Padding size

2012-11-20 Thread coderl
Hi,

How can I set the EVP padding size to 32byte instead of 16?

I currently have the problem, using aes 256bit, that if I encrypt a 20 or 30
size block it will only encrypt 16bytes instead of the whole part.


Thanks.



--
View this message in context: 
http://openssl.6102.n7.nabble.com/EVP-Padding-size-tp42413.html
Sent from the OpenSSL - User mailing list archive at Nabble.com.
__
OpenSSL Project http://www.openssl.org
User Support Mailing Listopenssl-users@openssl.org
Automated List Manager   majord...@openssl.org


Re: Regrading reducing the size of libcrypto.a

2012-11-20 Thread John Doe
From: jeetendra gangele gangele...@gmail.com

 Please guide me How can i minimize the size since I need only selected
 algorithms.

Can one strip a .a the file...?
Mine goes from 3.3M down to 1.8M.
But I have no idea if it will still be working...

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


Re: Regrading reducing the size of libcrypto.a

2012-11-20 Thread Jakob Bohm

On 11/20/2012 3:40 PM, jeetendra gangele wrote:

Hi,
I need to use only AES,SHA-2,EC-DH,EC-DSA.
I configure for only those ciphertext and compile the code still I am
getting 3.3MB library.
How can I minimize it?.
When I build with all the ciphertext enable that time also I am
getting 3.3 MB size.

Please guide me How can i minimize the size since I need only selected
algorithms.


Are you sure you want to reduce the size of the .a file?

Since an .a file is a collection or .o files only some of which will
be included in any given program, it is usually more interesting to reduce
the amount of unused code ending up in your (real or sample) program.

What I generally do is to find a linker option which will tell me which
.o files got included, how much the contribute to the size of the program
and (for better linkers) why those .o files were included.  Then start
looking for ways to

1. Avoid .o files which are not really used (but the linker thinks so
  because of a badly placed reference in a really used .o file)

2. Avoid large .o files where only a small portion of that .o file is
needed (the rest is linked in only because someone put it in the same
.c file as the needed stuff)

I have found that libcrypt is particularly bad in both areas:

1. Many common or required functions contain code such as
   if (ourObject-somepointer) /* Rarely true */
  SomeOtherObject_Free(ourObject-somepointer);

  Or even worse
 pAlg = SearchTableBySomeId(n); // Pulls in all the algs in table

2. Functions pulled in by logic such as the first bad example above are
often in the same .o file as the full implementation of all
functionality and needs of SomeOtherObject, which tends to be much
more than what is needed by the (really unused) _Free(ptr) method.

A lot of the bloat in libcrypt.a comes from the ASN.1 object handling
code, from functions that take an algorithm or object id rather than
a pointer to an algorithm/type descriptor, and from functions that
are coded to fallback to some specific method if NULL is passed in
(thus forcing the fallback to be linked in even if never called with
NULL).

My own attempts to reduce this were not as successful as I had hoped.

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: Regrading reducing the size of libcrypto.a

2012-11-20 Thread Jakob Bohm

On 11/20/2012 3:53 PM, John Doe wrote:

From: jeetendra gangele gangele...@gmail.com


Please guide me How can i minimize the size since I need only selected
algorithms.


Can one strip a .a the file...?
Mine goes from 3.3M down to 1.8M.
But I have no idea if it will still be working...


Properly stripping an .a file simply strips the debug information from
all the included .o files, making it harder to debug your program while
only providing the same savings as stripping the program after linking
it.

Improperly stripping an .a file also strips the symbols needed for
linking, which will be quite obvious as your program would no longer link.

Except for that incident on Solaris earlier this year, the size of .a
files is rarely a real problem.


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: Regrading reducing the size of libcrypto.a

2012-11-20 Thread jeetendra gangele
No, it gives me .a file then i convert it to .so file.
Since it was saying .so not supported yet for arm-linux evn in configure script.
i stripped .so file and i could reduce it to 2.5 MB but still its a
big one only.

I am compiling with only few ciphers algorithms like sha256,aes,EC-DH
and EC-DSA.
Interested thing is then I configure with all the size is 3.5MB and
when i configure with only above 5.
In that case also same size,it should not compile code for other ciphers right?


Thanks
Jeetendra
On 20 November 2012 20:23, John Doe jd...@yahoo.com wrote:
 From: jeetendra gangele gangele...@gmail.com

 Please guide me How can i minimize the size since I need only selected
 algorithms.

 Can one strip a .a the file...?
 Mine goes from 3.3M down to 1.8M.
 But I have no idea if it will still be working...

 JD
 __
 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: Regrading reducing the size of libcrypto.a

2012-11-20 Thread Jakob Bohm

On 11/20/2012 4:21 PM, jeetendra gangele wrote:

No, it gives me .a file then i convert it to .so file.
Since it was saying .so not supported yet for arm-linux evn in 
configure script.

i stripped .so file and i could reduce it to 2.5 MB but still its a
big one only.


Strange, why would it say such a thing?

Some incompatible assembler code?

Anyone knows?


I am compiling with only few ciphers algorithms like sha256,aes,EC-DH
and EC-DSA.
Interested thing is then I configure with all the size is 3.5MB and
when i configure with only above 5.
In that case also same size,it should not compile code for other 
ciphers right?




As I tried to explain in my first reply, most of the size of libcrypt
is not the ciphers,but all the ASN.1 stuff, such as code to handle
X.509 certificates, CRLs, PKCS#7messages, PKCS#12 and PKCS#8 key
files, etc. etc.

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


Does OpenSSL 0.9.7 support SHA256 Digest Algorithm

2012-11-20 Thread Wu, Hong-Tao (Aaron, HPSW-RD-SH)
Hi All,

In our product, we are still using OpenSSL 0.9.7d (on Windows platform) for 
certificate validation. Recently we suffered an issue about certificates based 
SHA256, and the following error is thrown:

ldap_bind: Can't contact LDAP server (-1) 
   additional info: error:0D0890A1:asn1 encoding 
routines:ASN1_verify:unknown message digest algorithm 

My question is:

1) Does OpenSSL 0.9.7 support SHA256 Digest Algorithm? 
2) If answer is NO for point 1, since which version does OpenSSL support SHA 
Digest Algorithm?

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


Newsreader (was: Question about exporting user certificate files to .pfx)

2012-11-20 Thread Jeffrey Walton
On Tue, Nov 20, 2012 at 11:56 AM,  jw72...@verizon.net wrote:
  it would be swell if I could get my email address removed from the list
 without removing my subscription too. This way I could just use my
 newsreader to get the messages without having my email box cluttered every
 day with this high-volume list.  Is such a think outside the realm of
 conceivable possibilities for anyone here? Who here after all is in
 control (for lack of a better word) of this list?  :)Thanks.
Google Groups mirrors this list. No need for a subscription/
http://groups.google.com/group/mailing.openssl.users.
__
OpenSSL Project http://www.openssl.org
User Support Mailing Listopenssl-users@openssl.org
Automated List Manager   majord...@openssl.org


RE: Newsreader (was: Question about exporting user certificate files to .pfx)

2012-11-20 Thread John A. Wallace
So can you get that group of messages sent to a newsreader?


 -Original Message-
 From: owner-openssl-us...@openssl.org [mailto:owner-openssl-
 us...@openssl.org] On Behalf Of Jeffrey Walton
 Sent: Tuesday, November 20, 2012 11:38 AM
 To: openssl-users@openssl.org
 Cc: st...@openssl.org
 Subject: Newsreader (was: Question about exporting user certificate
 files to .pfx)
 
 On Tue, Nov 20, 2012 at 11:56 AM,  jw72...@verizon.net wrote:
   it would be swell if I could get my email address removed from the
  list without removing my subscription too. This way I could just use
  my newsreader to get the messages without having my email box
  cluttered every day with this high-volume list.  Is such a think
  outside the realm of conceivable possibilities for anyone here? Who
 here after all is in
  control (for lack of a better word) of this list?  :)Thanks.
 Google Groups mirrors this list. No need for a subscription/
 http://groups.google.com/group/mailing.openssl.users.
 __
 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


Undefined reference to 'FIPS_text_start()'

2012-11-20 Thread Santhosh Kokala
Hi,
I am trying to build an application with the FIPS Object module. I followed the 
build instructions mentioned in FIPS User Guide 2.0.

FIPS Object Module:

./config
Make
make install

Open SSL:

./config fips
make depend
Make

My Application:

export FIPSLD_CC=g++
export CC=gcc
export CXX=/usr/local/ssl/fips-2.0/bin/fipsld


Build Errors:


/tmp/cczHeW7i.o: In function `FINGERPRINT_premain()':
/usr/local/ssl/fips-2.0/bin/../lib/fips_premain.c:103: undefined reference
to `FIPS_text_start()'
/usr/local/ssl/fips-2.0/bin/../lib/fips_premain.c:116: undefined reference
to `FIPS_incore_fingerprint(unsigned char*, unsigned int)'



So my question is, How to fix these errors?

Santhosh



Re: Undefined reference to 'FIPS_text_start()'

2012-11-20 Thread Jeffrey Walton
On Tue, Nov 20, 2012 at 6:16 PM, Santhosh Kokala
santhosh.kok...@riverbed.com wrote:
 Hi,
 I am trying to build an application with the FIPS Object module. I followed
 the build instructions mentioned in FIPS User Guide 2.0.

 FIPS Object Module:

 ./config
 Make
 make install

 Open SSL:

 ./config fips
 make depend
 Make

 My Application:

 export FIPSLD_CC=g++
 export CC=gcc
 export CXX=/usr/local/ssl/fips-2.0/bin/fipsld

 Build Errors:


 /tmp/cczHeW7i.o: In function `FINGERPRINT_premain()':
 /usr/local/ssl/fips-2.0/bin/../lib/fips_premain.c:103: undefined reference
 to `FIPS_text_start()'
 /usr/local/ssl/fips-2.0/bin/../lib/fips_premain.c:116: undefined reference
 to `FIPS_incore_fingerprint(unsigned char*, unsigned int)'
Add fipscanister.o to the link phase/

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


RE: Does OpenSSL 0.9.7 support SHA256 Digest Algorithm

2012-11-20 Thread Dave Thompson
 From: owner-openssl-us...@openssl.org On Behalf Of Wu, Hong-Tao (Aaron,
HPSW-RD-SH)
 Sent: Tuesday, 20 November, 2012 10:42

 In our product, we are still using OpenSSL 0.9.7d (on Windows 
 platform) for certificate validation. Recently we suffered an 
 issue about certificates based SHA256, and the following 
 error is thrown:
 
 ldap_bind: Can't contact LDAP server (-1) 
additional info: error:0D0890A1:asn1 encoding 
 routines:ASN1_verify:unknown message digest algorithm 
 
 My question is:
 
 1) Does OpenSSL 0.9.7 support SHA256 Digest Algorithm? 

On one system I still have 0.9.7m which does have SHA-2, 
with no mention in CHANGES. However, the 0.9.8* CHANGES 
describe it as added by 0.9.8(original) and not in 0.9.7h. 
Thus it was apparently backported somewhere 7h =7m.

 2) If answer is NO for point 1, since which version does 
 OpenSSL support SHA Digest Algorithm?
 
(You mean SHA-256, or more broadly SHA-2. SHA-1 has been in 
OpenSSL much longer, and the algorithm originally released 
as SHA was defective and withdrawn before use.)

Some 0.9.7 and all? 0.9.8 support SHA-2 as pure digests, 
and signatures using SHA-2 with RSA. Due to API limitation 
they can't do signatures using SHA-2 with ECDSA, and AIR 
don't have the DSA sizes for SHA-2 added by FIPS 186-3.


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


Unable to compile fips_hmac.c file

2012-11-20 Thread Santhosh Kokala

I am trying to compile tips_hmac.c to verify the FIPS build. But I see errors. 
Can you please let me know what needs to be done to fix this?

Errors:
gcc -c fips_hmac.c -I/usr/local/ssl/include
FIPSLD_CC=gcc /usr/local/ssl/fips-2.0/bin/fipsld -o fips_hmac fips_hmac.o 
/usr/local/ssl/fips-2.0/lib/fipscanister.o /usr/local/ssl/lib/libcrypto.a
/usr/local/ssl/lib/libcrypto.a(dso_dlfcn.o): In function `dlfcn_globallookup':
dso_dlfcn.c:(.text+0x2d): undefined reference to `dlopen'
dso_dlfcn.c:(.text+0x43): undefined reference to `dlsym'
dso_dlfcn.c:(.text+0x4d): undefined reference to `dlclose'
/usr/local/ssl/lib/libcrypto.a(dso_dlfcn.o): In function `dlfcn_pathbyaddr':
dso_dlfcn.c:(.text+0x8f): undefined reference to `dladdr'
dso_dlfcn.c:(.text+0xe9): undefined reference to `dlerror'
/usr/local/ssl/lib/libcrypto.a(dso_dlfcn.o): In function `dlfcn_bind_func':
dso_dlfcn.c:(.text+0x451): undefined reference to `dlsym'
dso_dlfcn.c:(.text+0x528): undefined reference to `dlerror'
/usr/local/ssl/lib/libcrypto.a(dso_dlfcn.o): In function `dlfcn_bind_var':
dso_dlfcn.c:(.text+0x5b1): undefined reference to `dlsym'
dso_dlfcn.c:(.text+0x688): undefined reference to `dlerror'
/usr/local/ssl/lib/libcrypto.a(dso_dlfcn.o): In function `dlfcn_unload':
dso_dlfcn.c:(.text+0x6f3): undefined reference to `dlclose'
/usr/local/ssl/lib/libcrypto.a(dso_dlfcn.o): In function `dlfcn_load':
dso_dlfcn.c:(.text+0x7d7): undefined reference to `dlopen'
dso_dlfcn.c:(.text+0x848): undefined reference to `dlclose'
dso_dlfcn.c:(.text+0x88d): undefined reference to `dlerror'
collect2: ld returned 1 exit status
make: *** [fips_hmac] Error 1

Steps I followed to build FIPS Object Module, Openssl and my application 
(fips_hma.c)

FIPS Object Module:

./config
make
make install

OpenSSL

./config fips
Make depend
Make
Make install


Openssl Version Output:

/usr/local/ssl/bin/openssl version -a
OpenSSL 1.0.1c-fips 10 May 2012
built on: Mon Oct 22 16:20:18 PDT 2012
platform: linux-elf
options:  bn(64,32) rc4(8x,mmx) des(ptr,risc1,16,long) idea(int) blowfish(idx)
compiler: gcc -DOPENSSL_THREADS -D_REENTRANT -DDSO_DLFCN -DHAVE_DLFCN_H 
-Wa,--noexecstack -DL_ENDIAN -DTERMIO -O3 -fomit-frame-pointer -Wall 
-DOPENSSL_BN_ASM_PART_WORDS -DOPENSSL_IA32_SSE2 -DOPENSSL_BN_ASM_MONT 
-DOPENSSL_BN_ASM_GF2m -I/usr/local/ssl/fips-2.0/include -DSHA1_ASM -DSHA256_ASM 
-DSHA512_ASM -DMD5_ASM -DRMD160_ASM -DAES_ASM -DVPAES_ASM -DWHIRLPOOL_ASM 
-DGHASH_ASM
OPENSSLDIR: /usr/local/ssl



My Application I used Makefile give in the FIPS User Guide 2.0

Makefile:

CC = gcc
OPENSSLDIR = /usr/local/ssl
LIBCRYPTO = $(OPENSSLDIR)/lib/libcrypto.a
INCLUDES = -I$(OPENSSLDIR)/include
CMD = fips_hmac
OBJS = $(CMD).o /usr/local/ssl/fips-2.0/lib/fipscanister.o

$(CMD): $(OBJS)
FIPSLD_CC=$(CC) /usr/local/ssl/fips-2.0/bin/fipsld -o $(CMD) $(OBJS) 
$(LIBCRYPTO)

$(OBJS): $(CMD).c
$(CC) -c $(CMD).c $(INCLUDES)
clean:
rm $(OBJS)





RE: Does OpenSSL 0.9.7 support SHA256 Digest Algorithm

2012-11-20 Thread Wu, Hong-Tao (Aaron, HPSW-RD-SH)
Thanks a lot, Dave. 

So it means that SHA256 is only technically supported from 0.9.7h. If I am 
using 0.9.7d, it is not included, right?

Best Regards,
Aaron

-Original Message-
From: owner-openssl-us...@openssl.org [mailto:owner-openssl-us...@openssl.org] 
On Behalf Of Dave Thompson
Sent: Wednesday, November 21, 2012 10:01 AM
To: openssl-users@openssl.org
Subject: RE: Does OpenSSL 0.9.7 support SHA256 Digest Algorithm

 From: owner-openssl-us...@openssl.org On Behalf Of Wu, Hong-Tao 
 (Aaron,
HPSW-RD-SH)
 Sent: Tuesday, 20 November, 2012 10:42

 In our product, we are still using OpenSSL 0.9.7d (on Windows
 platform) for certificate validation. Recently we suffered an issue 
 about certificates based SHA256, and the following error is thrown:
 
 ldap_bind: Can't contact LDAP server (-1) 
additional info: error:0D0890A1:asn1 encoding 
 routines:ASN1_verify:unknown message digest algorithm
 
 My question is:
 
 1) Does OpenSSL 0.9.7 support SHA256 Digest Algorithm? 

On one system I still have 0.9.7m which does have SHA-2, with no mention in 
CHANGES. However, the 0.9.8* CHANGES describe it as added by 0.9.8(original) 
and not in 0.9.7h. 
Thus it was apparently backported somewhere 7h =7m.

 2) If answer is NO for point 1, since which version does OpenSSL 
 support SHA Digest Algorithm?
 
(You mean SHA-256, or more broadly SHA-2. SHA-1 has been in OpenSSL much 
longer, and the algorithm originally released as SHA was defective and 
withdrawn before use.)

Some 0.9.7 and all? 0.9.8 support SHA-2 as pure digests, and signatures using 
SHA-2 with RSA. Due to API limitation they can't do signatures using SHA-2 with 
ECDSA, and AIR don't have the DSA sizes for SHA-2 added by FIPS 186-3.


__
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: This is one for the Pros: cert is not privkey

2012-11-20 Thread Dave Thompson
From: owner-openssl-us...@openssl.org On Behalf Of Peter Parker
Sent: Tuesday, 20 November, 2012 20:59

Subject: This is one for the Pros

Not really. This is pretty basic. 

I've been trying to generate a public/private key pair after 
generating the certificates, but OpenSSL keeps giving me an error. 
The commands and the error are below. Thanks.

No you're not; you're generating a CA keypair and cert (directly),  
then an application keypair, then an application cert (via CSR).
Which is the (well, a) correct sequence, for one entity.

Commands
#openssl req -new -x509 -extensions v3_ca -days 365 -keyout caKey.pem 
-passout pass:test -out caCert.crt -batch
#openssl genrsa -out application.pem -passout pass:test -des3 1028

1028 is an unusual size for an RSA key; most folks use power-of-2 
based values like 1024 1536 2048. 1024 is presently rather marginal 
for security; for example, NIST has it deprecated since the end of 
2010, and prohibited after the end of 2013, for US government use.

#openssl req -new -key application.pem -passin pass:test -out
application.csr -batch

A second req -new -batch generates a CSR with the same DN ...

#openssl x509 -req -days 365 -in application.csr -CA caCert.crt
-CAcreateserial 
-CAkey caKey.pem -passin:test -out test-key.pem -extensions ssl_cert

... thus this creates a CA-signed cert which appears to be self-signed, 
and will not chain correctly with OpenSSL. If the ssl_cert section of your 
config file (which doesn't exist in the distro file) includes AKI, other 
software that chains primarily by AKI may work, but this is still incorrect.

This puts the cert in a file named test-key.pem, which is a misleading name.

#openssl rsa -in test-key.pem -passin pass:test -out pub-key.pem -outform
PEM -pubout

And therefore this command, which is not the last one, fails because 
you told it to read the privatekey from a file which is a certificate.
application.pem is your privatekey.

#openssl rsautl -encrypt -inkey pub-key.pem -pubin -in testfile.txt -out
eFile.ssl

Error
unable to load Private Key (I receive this after the last command)

Not last.

The key thing that I am trying to do is to encrypt some files with the key 
that I generate. I do however want to use the public and private keys that 
I get out of the PEM file using the x509 (or the ca utilities).

If (any of) your files are larger than about 100 bytes (for 1024-bit RSA) 
you can't use raw RSA; even if they aren't, you can't interoperate with 
properly designed software that doesn't use raw RSA. The conventional 
approach is to encrypt the bulk data symmetrically with a nonce key (DEK) 
and PK (RSA) encrypt that DEK; there are numerous schemes that do this, 
but the one that OpenSSL supports directly is PKCS7/CMS/SMIME. (CMS is an 
updated Internet version of PKCS7, and SMIME is a simple wrapping of CMS.)

You can generate and use RSA keys without using certificates (and without 
using the req x509 ca utilities) IF you have a way to distribute them 
correctly -- that is, to make sure the enveloper always uses a correct 
publickey for the recipient and not a forged, tampered, or obsolete one.
(And similarly the verifier for a signer.) Most standard schemes do use 
X.509 certs for this purpose, because they are also standard.


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


RE: Does OpenSSL 0.9.7 support SHA256 Digest Algorithm

2012-11-20 Thread Dave Thompson
 From: Wu, Hong-Tao (Aaron, HPSW-RD-SH) [mailto:hongtao...@hp.com] 
 Sent: Tuesday, 20 November, 2012 22:05
 To: openssl-users@openssl.org
 Cc: dthomp...@prinpay.com
 Subject: RE: Does OpenSSL 0.9.7 support SHA256 Digest Algorithm
 
 Thanks a lot, Dave. 
 
 So it means that SHA256 is only technically supported from 
 0.9.7h. If I am using 0.9.7d, it is not included, right?
 
I don't have all the old versions to check, but according to CHANGES 
it is NOT in 0.9.7h which means it SHOULD NOT be in any earlier 0.9.7*, 
and 0.9.7d IS earlier than 0.9.7h so it SHOULD NOT be in 0.9.7d.

It IS in the code for 0.9.7m, therefore it was added sometime 
after 0.9.7h and before or at 0.9.7m. That's what 7h =7m means.

 Best Regards,
 Aaron
 
 -Original Message-
 From: owner-openssl-us...@openssl.org 
 [mailto:owner-openssl-us...@openssl.org] On Behalf Of Dave Thompson
 Sent: Wednesday, November 21, 2012 10:01 AM
 To: openssl-users@openssl.org
 Subject: RE: Does OpenSSL 0.9.7 support SHA256 Digest Algorithm
 
  From: owner-openssl-us...@openssl.org On Behalf Of Wu, Hong-Tao 
  (Aaron,
 HPSW-RD-SH)
  Sent: Tuesday, 20 November, 2012 10:42
 
  In our product, we are still using OpenSSL 0.9.7d (on Windows
  platform) for certificate validation. Recently we suffered an issue 
  about certificates based SHA256, and the following error is thrown:
  
  ldap_bind: Can't contact LDAP server (-1) 
 additional info: error:0D0890A1:asn1 encoding 
  routines:ASN1_verify:unknown message digest algorithm
  
  My question is:
  
  1) Does OpenSSL 0.9.7 support SHA256 Digest Algorithm? 
 
 On one system I still have 0.9.7m which does have SHA-2, with 
 no mention in CHANGES. However, the 0.9.8* CHANGES describe 
 it as added by 0.9.8(original) and not in 0.9.7h. 
 Thus it was apparently backported somewhere 7h =7m.
 
  2) If answer is NO for point 1, since which version does OpenSSL 
  support SHA Digest Algorithm?
  
 (You mean SHA-256, or more broadly SHA-2. SHA-1 has been in 
 OpenSSL much longer, and the algorithm originally released as 
 SHA was defective and withdrawn before use.)
 
 Some 0.9.7 and all? 0.9.8 support SHA-2 as pure digests, and 
 signatures using SHA-2 with RSA. Due to API limitation they 
 can't do signatures using SHA-2 with ECDSA, and AIR don't 
 have the DSA sizes for SHA-2 added by FIPS 186-3.
 
 
 __
 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