make it so that from/to can be missaligned as it can happen (the geli

regression manages to do it)...  We use a packed struct to coerce
gcc/clang into producing unaligned loads (there is not packed pointer
attribute, otherwise this would be easier)...

use _storeu_ and _loadu_ when using the structure is overkill...

be better at using types properly...  Since we allocate our own key
schedule and make sure it's aligned, use the __m128i type in various
arguments to functions...

clang ignores __aligned on prototypes and gcc errors on them, leave them
in comments to document that these function arguments are require to be
aligned...

about all that changes is movdqa -> movdqu from reading the diff of the
disassembly output...

Noticed by:	symbolics at gmx.com
MFC after:	3 days

1 files changed, 20 insertions, 19 deletions

diff --git a/sys/crypto/aesni/aesni.h b/sys/crypto/aesni/aesni.h
index 17ca9c5..99edd42 100644
--- a/sys/crypto/aesni/aesni.h
+++ b/sys/crypto/aesni/aesni.h
@@ -71,29 +71,30 @@ struct aesni_session {
 /*
  * Internal functions, implemented in assembler.
  */
-void aesni_set_enckey(const uint8_t *userkey, uint8_t *encrypt_schedule,
-    int number_of_rounds);
-void aesni_set_deckey(const uint8_t *encrypt_schedule,
-    uint8_t *decrypt_schedule, int number_of_rounds);
+void aesni_set_enckey(const uint8_t *userkey,
+    uint8_t *encrypt_schedule /*__aligned(16)*/, int number_of_rounds);
+void aesni_set_deckey(const uint8_t *encrypt_schedule /*__aligned(16)*/,
+    uint8_t *decrypt_schedule /*__aligned(16)*/, int number_of_rounds);
 
 /*
  * Slightly more public interfaces.
  */
-void aesni_encrypt_cbc(int rounds, const void *key_schedule, size_t len,
-    const uint8_t *from, uint8_t *to, const uint8_t iv[AES_BLOCK_LEN]);
-void aesni_decrypt_cbc(int rounds, const void *key_schedule, size_t len,
-    uint8_t *buf, const uint8_t iv[AES_BLOCK_LEN]);
-void aesni_encrypt_ecb(int rounds, const void *key_schedule, size_t len,
-    const uint8_t from[AES_BLOCK_LEN], uint8_t to[AES_BLOCK_LEN]);
-void aesni_decrypt_ecb(int rounds, const void *key_schedule, size_t len,
-    const uint8_t from[AES_BLOCK_LEN], uint8_t to[AES_BLOCK_LEN]);
-
-void aesni_encrypt_xts(int rounds, const void *data_schedule,
-    const void *tweak_schedule, size_t len, const uint8_t *from, uint8_t *to,
-    const uint8_t iv[AES_BLOCK_LEN]);
-void aesni_decrypt_xts(int rounds, const void *data_schedule,
-    const void *tweak_schedule, size_t len, const uint8_t *from, uint8_t *to,
+void aesni_encrypt_cbc(int rounds, const void *key_schedule /*__aligned(16)*/,
+    size_t len, const uint8_t *from, uint8_t *to,
     const uint8_t iv[AES_BLOCK_LEN]);
+void aesni_decrypt_cbc(int rounds, const void *key_schedule /*__aligned(16)*/,
+    size_t len, uint8_t *buf, const uint8_t iv[AES_BLOCK_LEN]);
+void aesni_encrypt_ecb(int rounds, const void *key_schedule /*__aligned(16)*/,
+    size_t len, const uint8_t *from, uint8_t *to);
+void aesni_decrypt_ecb(int rounds, const void *key_schedule /*__aligned(16)*/,
+    size_t len, const uint8_t *from, uint8_t *to);
+
+void aesni_encrypt_xts(int rounds, const void *data_schedule /*__aligned(16)*/,
+    const void *tweak_schedule /*__aligned(16)*/, size_t len,
+    const uint8_t *from, uint8_t *to, const uint8_t iv[AES_BLOCK_LEN]);
+void aesni_decrypt_xts(int rounds, const void *data_schedule /*__aligned(16)*/,
+    const void *tweak_schedule /*__aligned(16)*/, size_t len,
+    const uint8_t *from, uint8_t *to, const uint8_t iv[AES_BLOCK_LEN]);
 
 int aesni_cipher_setup(struct aesni_session *ses,
     struct cryptoini *encini);
@@ -103,4 +104,4 @@ int aesni_cipher_process(struct aesni_session *ses,
 uint8_t *aesni_cipher_alloc(struct cryptodesc *enccrd, struct cryptop *crp,
     int *allocated);
 
-#endif
+#endif /* _AESNI_H_ */