Address more CipherTest flakiness due to random IV values

In a recent change, we stopped generating random keys to prevent
flakiness in the BadPaddingException test cases. This change removes
the use of random initialization vector (IV) values. There still are
uses of SecureRandom where the API being tested requires it, but I
visually checked that none of the BadPaddingException cases fall in
this category.

Bug: 6873673

Change-Id: Ia4e53973da05377687faf327b64acd42a2db344a
1 file changed