I'm aware that the default installation of the Solaris 10 OS provides a
PKCS#11-based OpenSSL implementation. I'm trying to take some metrics to
figure out how much more efficient certain processes are with the PKCS
engine. How would I disable the PKCS engine on an UltraSPARC T1 processor?
-Chris
--
View this message in context: http://www.nabble.com/Disabling-the-PKCS--11-on-Solaris-10-tp18587438p18587438.html
Sent from the OpenSSL - User mailing list archive at Nabble.com.
______________________________________________________________________
OpenSSL Project http://www.openssl.org
User Support Mailing List openss...@openssl.org
Automated List Manager majo...@openssl.org
>I'm aware that the default installation of the Solaris 10 OS provides a
>PKCS#11-based OpenSSL implementation. I'm trying to take some metrics to
>figure out how much more efficient certain processes are with the PKCS
>engine. How would I disable the PKCS engine on an UltraSPARC T1 processor?
an application must explicitly initialize the engine so that it can
use it. The only app that uses the engine by default on S10 is Apache. You
can also use "-engine pkcs11" with openssl(1) command. J.
--
Jan Pechanec
>Jan,
>
>Thanks for the information! How can I explicitly initialize the PKCS11
>engine when writing in C? I've looked at examples, but I think I'm missing a
>step because when I verify or sign using OpenSSL, I'm seeing no improvement
>in performance.
see how openssl does that in apps/ subdirectory in the OpenSSL
tarball. You can see what is going on by trussing your app with
-u"libcrypto::" and -u"libpkcs11::". J.
>When I ran the same test with the pkcs chip initialized ("speed sha1 -engine
>pkcs11") the results were:
>
> The 'numbers' are in 1000s of bytes per second processed.
> type 16 bytes 64 bytes 256 bytes 1024 bytes 8192
>bytes
> sha1 408.02k 1508.91k 5448.41k 15804.53k
>34058.17k
>
>This leads me to believe that while the pkcs11 might not improve the
>performance of SHA-1. Would this be a fair assessment, or am I missing
>something?
T1 has ncp(7d) which accelerates RSA/DSA/DH only, see "cryptoadm
list -m". What you see is soft token. Given the inherent overhead of the
crypto framework the soft token is slower than OpenSSL.
in OpenSolaris, there is a project that mechanisms that are not
implemented in hw will stay in the soft token.
if you want acceleration for symmetric crypto and digests you must
use T2 or SCA-6000 (which is significantly slower than n2cp that is shipped
with Niagara2).
> in OpenSolaris, there is a project that mechanisms that are not
>implemented in hw will stay in the soft token.
of course, I meant "will stay in OpenSSL"
>I see what you mean, however, when I get the processor info with the command
>"uname -a" it returns:
>
> SunOS t5200tx 5.10 Generic_120011-14 sun4v sparc
>SUNW,SPARC-Enterprise-T5220
>
>I was working under the impression that this is a T2. Was I misinformed?
no, it's T2. You mentioned earlier that you had T1.
it stays in soft token, I vaguely remember a bug there.
check with "kstat -m n2cp" before and after. If the counters don't
change it stays in the soft token. You can also try other mechs, AES, for
example. BTW you should use "-elapsed" with the -engine options; if it goes
to the hw then the final numbers would be wrong without that option.
for Solaris specific crypto stuff I suggest to use
crypto-...@opensolaris.org.
J.