Convert sa to ptr and memory leakage fix.

[kibnakamoto]

When I compiled using make, I got the following error:

[  1%] Building C object src/CMakeFiles/Crypto.dir/src_main/sadb_routine_inmemory.template.c.o
/home/kibnakamoto/workspace/NASA/CryptoLib/src/src_main/sadb_routine_inmemory.template.c: In function ‘sadb_get_sa_from_spi’:
/home/kibnakamoto/workspace/NASA/CryptoLib/src/src_main/sadb_routine_inmemory.template.c:384:12: error: the comparison will always evaluate as ‘false’ for the address of ‘sa’ will never be NULL [-Werror=address]
  384 |     if (sa == NULL)
      |            ^~
/home/kibnakamoto/workspace/NASA/CryptoLib/src/src_main/sadb_routine_inmemory.template.c:41:30: note: ‘sa’ declared here
   41 | static SecurityAssociation_t sa[NUM_SA];
      |                              ^~
/home/kibnakamoto/workspace/NASA/CryptoLib/src/src_main/sadb_routine_inmemory.template.c: In function ‘sadb_get_operational_sa_from_gvcid’:
/home/kibnakamoto/workspace/NASA/CryptoLib/src/src_main/sadb_routine_inmemory.template.c:419:12: error: the comparison will always evaluate as ‘false’ for the address of ‘sa’ will never be NULL [-Werror=address]
  419 |     if (sa == NULL)
      |            ^~
/home/kibnakamoto/workspace/NASA/CryptoLib/src/src_main/sadb_routine_inmemory.template.c:41:30: note: ‘sa’ declared here
   41 | static SecurityAssociation_t sa[NUM_SA];
      |                              ^~

To fix it, I changed the sa array into a pointer initialized with calloc, I freed it in the sadb_close() function.

This solves it because if sa is an array, comparing to NULL always returns false because array != NULL. So I converted the sa to pointer.

The second fix is:
I ran valgrind and saw that there was 14 bytes of memory definitely lost. This was caused by sa[12].arsn definition which was completely wrong to begin with.

Line 311:

sa[12].arsn = (uint8_t*) calloc(1, sa[1].arsn_len * sizeof(uint8_t)); 

To my understanding there are a few problems with this line, sa[1].arsn_len = 2, while sa[12].arsn_len = 0
I think this line was meant to be

sa[12].arsn = (uint8_t*) calloc(1, sa[12].arsn_len * sizeof(uint8_t)); 

but either way, this line causes a memory leakage that can be solved by not setting it to anything (so that sa[12].arsn = NULL).

Therefore, considering that sa[12].arsn_len = 0, then sa[12].arsn = NULL. This was correctly applied to sa[8] but not here.

Running tests...
Test project /home/kibnakamoto/workspace/NASA/CryptoLib
    Start 1: UT_TC_APPLY
1/8 Test #1: UT_TC_APPLY ......................   Passed    0.59 sec
    Start 2: UT_TC_PROCESS
2/8 Test #2: UT_TC_PROCESS ....................   Passed    0.32 sec
    Start 3: UT_CRYPTO_CONFIG
3/8 Test #3: UT_CRYPTO_CONFIG .................   Passed    0.00 sec
    Start 4: UT_CRYPTO
4/8 Test #4: UT_CRYPTO ........................   Passed    0.13 sec
    Start 5: UT_CRYPTO_AOS
5/8 Test #5: UT_CRYPTO_AOS ....................   Passed    0.00 sec
    Start 6: UT_CRYPTO_MC
6/8 Test #6: UT_CRYPTO_MC .....................   Passed    0.04 sec
    Start 7: UT_TM_APPLY
7/8 Test #7: UT_TM_APPLY ......................   Passed    0.28 sec
    Start 8: UT_TM_PROCESS
8/8 Test #8: UT_TM_PROCESS ....................   Passed    0.27 sec

100% tests passed, 0 tests failed out of 8

Total Test time (real) =   1.64 sec

All tests passed.

[kibnakamoto]

The memory leak fix solved around 30k byte leak.

[rjbrown2]

This may be a larger problem than with just this instance. The recent change to the SA from pointer to array could be causing false positives in unit tests as well. @jlucas9

[dccutrig]

Let's discuss tomorrow 10/19

[kibnakamoto]

@dccutrig
That works.
Thank you.

[rjbrown2]

Moved to an issue:
#211
Until licensing can occur, or this is fixed. It cannot currently be remedied through a PR with an external contributor.