Bug in C_CloseAllSessions()

[Anthony Scarpino]

Roy Arends wrote:
> I encountered a bug with the Metaslot. It seems that
> meta_release_slot_session (and others), used by C_CloseAllSessions,
> is causing a nasty segmentation fault when trying to close a certain
> amount of idle sessions. This is circumvented by closing individual
> sessions one by one.
> 
> Please find code that replicates the issue here: 
> http://download.nominet.org.uk/hsm-tools/hsm-tools.tar.gz (The file
> is hsm-solaris-bug.c).
> 
> Run it as ./hsm-solaris-bug -p <pin> -s <slot> -t <sessions> label
> 
> Where label is a key in some token. slot is the slot number (defaults
> to 0) and pin is the user pin for the token in the slot. If
> <sessions> is high, (in some configuration 101), there will be a
> segmentation fault.
> 
> I've blogged about the performance of the SCA6000 here: 
> http://blog.nominet.org.uk/tech/2008/06/02/40k-signatures-second-on-fips-140-2-level-3-hardware/
> 

Do you have a stack trace of this problem?.. I've been able to run your 
program, but the core dump I am seeing is occurring during C_Finalize 
involving objects.. I didn't see anything with sessions.. I also had to 
use 1000 sessions..101 worked fine..

Also, to verify, during this test you are using the SCA6000 as a 
keystore correct?  You set the
METASLOT_OBJECTSTORE_TOKEN environment variable or set it in 'cryptoadm 
metaslot'?

thanks

Tony