Difference between EVP_CipherInit and EVP_CipherInit_ex

Fixes #10455

Reviewed-by: Paul Dale <paul.dale@oracle.com>
Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org>
(Merged from https://github.com/openssl/openssl/pull/10550)

diff --git a/doc/man3/EVP_EncryptInit.pod b/doc/man3/EVP_EncryptInit.pod
index 1ca3e7478eb3da7a4b676b12b4a1882adede6daf..50a6cc8702eacbf030b60916200c2c1dff31b532 100644 (file)
--- a/doc/man3/EVP_EncryptInit.pod
+++ b/doc/man3/EVP_EncryptInit.pod
@@ -667,6 +667,15 @@ EVP_EncryptFinal_ex(), EVP_DecryptInit_ex(), EVP_DecryptFinal_ex(),
 EVP_CipherInit_ex() and EVP_CipherFinal_ex() because they can reuse an
 existing context without allocating and freeing it up on each call.
 
+There are some differences between functions EVP_CipherInit() and
+EVP_CipherInit_ex(), significant in some circumstances. EVP_CipherInit() fills
+the passed context object with zeros.  As a consequence, EVP_CipherInit() does
+not allow step-by-step initialization of the ctx when the I<key> and I<iv> are
+passed in separate calls. It also means that the flags set for the CTX are
+removed, and it is especially important for the
+B<EVP_CIPHER_CTX_FLAG_WRAP_ALLOW> flag treated specially in
+EVP_CipherInit_ex().
+
 EVP_get_cipherbynid(), and EVP_get_cipherbyobj() are implemented as macros.
 
 =head1 BUGS