In general, T4 runs much faster, this goes not just for PKCS11 provider,
but for pure java provider as well.
My run is focused on the settings of disabled mechanism section. Based
on the run results, message digests excluded, performance numbers are
similar between the default one and the special one which doesn't
disable any mechanisms.
Valerie
On 04/29/11 01:38 PM, Frances Ho wrote:
So what were the results from your run on the T4? performance wise.
On 4/29/2011 11:14 AM, Valerie (Yu-Ching) Peng wrote:
Right, good point. I'll update the bugster w/ evaluation on
CKM_SSL3_KEY_AND_MAC_DERIVE and CKM_TLS_KEY_AND_MAC_DERIVE.
True, conceptually this should be done for every release. Although,
realistically, I don't think we'll need to adjust this again for jdk8
given that the remaining Solaris bugs have already been closed. But
still, I'll add a sub-CR for jdk8, so we can check and cross it off
later (not really worthy of a P2 for jdk 8 though).
Thanks,
Valerie
On 04/29/11 05:45 AM, Sean Mullan wrote:
This looks fine. However, the bug report never mentions why it is ok
to remove the CKM_SSL3_KEY_AND_MAC_DERIVE and
CKM_TLS_KEY_AND_MAC_DERIVE mechanisms. I think it should be updated.
Also, isn't this the kind of thing we should periodically review each
release? If so, maybe we should create a sub-CR for JDK 8, keep that
open, and re-evaluate later.
--Sean
On 4/28/11 8:59 PM, Valerie (Yu-Ching) Peng wrote:
Sean,
Could you please review this PKCS11 configuration file change before
next Monday?
http://cr.openjdk.java.net/~valeriep/7036252/webrev.00/
I have run the regression tests against the new config file and
things look fine.
More information can be found in the bugster record.
Thanks,
Valerie
