Note: I didn't compare the GCM code with the NIST spec.

DEA-1
Nit: The webrev should have been generated with webrev -O so the link to 
6260053 works outside of Sun

DEA-2 kcf_cryptoadm.c
The new AES mechanism string (CKM_AES_GCM) will need to be added at the top of 
kcf_cryptoadm.c (lines 47 and 126) if my kcf.conf changes are putback before 
your changes:
http://dan.drydog.com/reviews/6414175-kcfconf/usr/src/uts/common/crypto/core/kcf_cryptoadm.c.html

DEA-3 modes.c crypto_free_mode_ctx()
If crypto_free_mode_ctx() is called frequently (my guess it's not_, consider 
replacing the ever-growing cascading if with a switch statement on 
(common_ctx->cc_flags & (ECB_MODE|CBC_MODE|CTR_MODE|CCM_MODE|GCM_MODE))

DEA-4 aes.c
This:
 463         if (((aes_ctx->ac_flags & CTR_MODE) == 0) &&
 464             ((aes_ctx->ac_flags & CCM_MODE) == 0) &&
 465             ((aes_ctx->ac_flags & GCM_MODE) == 0) &&
Could be replaced with this:
        if (((aes_ctx->ac_flags & (CTR_MODE|CCM_MODE|GCM_MODE) == 0) &&

Also lines 581-583 and lines 1003-1005

DEA-5 aes.c
 774         if (((aes_ctx->ac_flags & CCM_MODE) == 0) &&
 775             ((aes_ctx->ac_flags & GCM_MODE) == 0)) {
COuld be replaced with:
        if (((aes_ctx->ac_flags & (CCM_MODE|GCM_MODE) == 0) &&

DEA-6 gcm.c
You're using ntohll()--yeah!

DEA-7 gcm.c
Nit: lines 299-300: if statement indented 1 tab too much.
--
This message posted from opensolaris.org

Reply via email to