1263310 - Lots of things in cmd/bltest/blapitest.c aren't run

Status: NEW

(NSS :: Test, defect, P5)

Description

[Jed Davis [:jld] ⟨⏰|UTC-7⟩ ⟦he/him⟧]

Some of these might overlap, so I'm just going to throw them into one bug:

* None of the hash function "restart" variants are called; there's a commented-out /*params->hash.restart = PR_TRUE;*/ on line 3066, and there's also this inside a large block of code that's unreached for other reasons:

[.]    4084         if (is_hashCipher(cipherInfo->mode))
       4085             cipherInfo->params.hash.restart =
[.]    4086                 bltest.options[opt_Restart].activated;

* Nothing related to ECDSA is reached.  It's all inside #ifndef NSS_DISABLE_ECC, but I haven't defined that and the code is being compiled, but not run.

* Nothing guarded by (cipherInfo->repetitionsToPerfom != 0) is reached.

* In the else case of each of those "repetitionsToPerfom" (sic) tests, there's a while (! (TIMETOFINISH())) loop whose body is also never executed; i.e., TIMETOFINISH() was apparently already true when entering the loop.  Perhaps we're always telling it to run for 0 seconds?  Example:

       2438         if (cipherInfo->repetitionsToPerfom != 0) {
[.]    2439             TIMESTART();
    (10 lines elided)
[#]    2450             int opsBetweenChecks = 0;
[.#]   2451             TIMEMARK(cipherInfo->seconds);
[#]    2452             while (! (TIMETOFINISH())) {
       2453                 int j = 0;
[.]    2454                 for (;j < opsBetweenChecks;j++) {
[.]    2455                     (*cipherInfo->cipher.symmkeyCipher)(

* We aren't using the RSAPopulate command.

* We're only using the SelfTest ("BLAPI self-test") command in combination with other commands.  This means that the bottom ~400 lines of main() are completely unreached.  This includes the block from the first item.

Comment 1

[Anna Weine]

Attachment: Bug 1748245 - Modification of the cipher script and the list of the algorithms it uses

Comment 2

[Anna Weine]

We aren't using the RSAPopulate command.

It seems that RSAPopulate is now used.

Comment 3

[Anna Weine]

• Nothing guarded by (cipherInfo->repetitionsToPerfom != 0) is reached.

This also seems to be solved.

Comment 4

[Phabricator Automation]

Comment on attachment 9257153 [details]
Bug 1748245 - Modification of the cipher script and the list of the algorithms it uses

Revision D134866 was moved to bug 1748245. Setting attachment 9257153 [details] to obsolete.