The following reply was made to PR bin/32414; it has been noted by GNATS.

From: christos@zoulas.com (Christos Zoulas)
To: gnats-bugs@netbsd.org, gnats-admin@netbsd.org,
	netbsd-bugs@netbsd.org, ijk@speakeasy.net
Cc: 
Subject: Re: bin/32414
Date: Sun, 23 Apr 2006 20:01:09 -0400

 On Apr 23,  6:50pm, ijk@speakeasy.net (Jeff Ito) wrote:
 -- Subject: Re: bin/32414
 
 | The following reply was made to PR bin/32414; it has been noted by GNATS.
 | 
 | From: Jeff Ito <ijk@speakeasy.net>
 | To: Christos Zoulas <christos@zoulas.com>
 | Cc: gnats-bugs@netbsd.org
 | Subject: Re: bin/32414
 | Date: Sun, 23 Apr 2006 14:47:18 -0400
 | 
 |  > |  Running a kernel with CRYPTO_DEBUG results in the following.
 |  > |  
 |  > |  OpenSSL> speed -engine cryptodev aes
 |  > |  SIOCSESSION violates kernel parameters 22
 |  > |  SIOCSESSION violates kernel parameters 22
 |  > |  SIOCSESSION violates kernel parameters 22
 |  > |  SIOCSESSION violates kernel parameters 22
 |  > |  SIOCSESSION violates kernel parameters 22
 |  > 
 |  > Can you try playing with the sysctls?
 |  > 
 |  > kern.usercrypto = 1
 |  > kern.userasymcrypto = 1
 |  > kern.cryptodevallowsoft = 1
 |  > 
 |   
 |  Is there a particular combination of these variables you would like to
 |  see?  So far testing various combinations has caused the "...violates
 |  kernel parameters..." message to go away, however the benchmarks do
 |  not seem to indicate any leveraging of the cryto card.
 
 Can you add some debugging printfs in crypto_newsession() and see where
 it fails?
 
 christos