1559906 - fipstest: use CKM_TLS12_MASTER_KEY_DERIVE instead of vendor specific mechanism

Status: RESOLVED FIXED

(NSS :: Tools, defect, P1)

Description

[Daiki Ueno [:ueno]]

Attachment: nss-fipstest-tls12prf.patch

The fipstest command currently only exercises CKM_NSS_TLS_MASTER_KEY_DERIVE_SHA256. It should switch to using the standard CKM_TLS12_MASTER_KEY_DERIVE mechanism.

Patch from Tomas Mraz.

Comment 1

[Robert Relyea]

Comment on attachment 9072625 [details] [diff] [review]
nss-fipstest-tls12prf.patch

Review of attachment 9072625 [details] [diff] [review]:
-----------------------------------------------------------------

yes, this is a good fix. It also lets us cavs tests more than just SHA256.

::: cmd/fipstest/fipstest.c
@@ +6728,5 @@
>                  if (buf[7] == '0') {
>                      master_mech.mechanism = CKM_TLS_MASTER_KEY_DERIVE;
>                      key_block_mech.mechanism = CKM_TLS_KEY_AND_MAC_DERIVE;
> +                    master_mech.ulParameterLen = sizeof(CK_SSL3_MASTER_KEY_DERIVE_PARAMS);
> +                    key_block_mech.ulParameterLen = sizeof(CK_SSL3_KEY_MAT_PARAMS);

perhaps a comment here that says: :CK_SSL3_MASTER_KEY_DERIVE_PARAMS is a subset of CK_TLS12_MASTER_KEY_DERIVE_PARAMS and CK_SSL3_KEY_MAT_PARAMS is a subset of CK_TLS12_KEY_MAT_PARAMS. The latter params as and extra prfHashMechanism field at the end.

Not required for checking.

Comment 2

[Daiki Ueno [:ueno]]

Pushed after adding the comment:
https://hg.mozilla.org/projects/nss/rev/6ef49fe67d6227a1d290da5537ec0dade379a15a

Thank you for the review!