diff --git a/include/polarssl/aesni.h b/include/polarssl/aesni.h
index 808ff4d49..92b23cd6b 100644
--- a/include/polarssl/aesni.h
+++ b/include/polarssl/aesni.h
@@ -74,12 +74,10 @@ int aesni_crypt_ecb( aes_context *ctx,
  *
  * \note           Both operands and result are bit strings interpreted as
  *                 elements of GF(2^128) as per the GCM spec.
- *
- * \return         0 on success (cannot fail)
  */
-int aesni_gcm_mult( unsigned char c[16],
-                    const unsigned char a[16],
-                    const unsigned char b[16] );
+void aesni_gcm_mult( unsigned char c[16],
+                     const unsigned char a[16],
+                     const unsigned char b[16] );
 
 /**
  * \brief           Compute decryption round keys from encryption round keys
diff --git a/library/aesni.c b/library/aesni.c
index aa1f4b1c4..b5858dc35 100644
--- a/library/aesni.c
+++ b/library/aesni.c
@@ -107,9 +107,9 @@ int aesni_crypt_ecb( aes_context *ctx,
  * GCM multiplication: c = a times b in GF(2^128)
  * Based on [CLMUL-WP] algorithms 1 (with equation 27) and 5.
  */
-int aesni_gcm_mult( unsigned char c[16],
-                    const unsigned char a[16],
-                    const unsigned char b[16] )
+void aesni_gcm_mult( unsigned char c[16],
+                     const unsigned char a[16],
+                     const unsigned char b[16] )
 {
     unsigned char aa[16], bb[16], cc[16];
     size_t i;
diff --git a/library/gcm.c b/library/gcm.c
index 894c5d3eb..8950360f6 100644
--- a/library/gcm.c
+++ b/library/gcm.c
@@ -198,7 +198,7 @@ static void gcm_mult( gcm_context *ctx, const unsigned char x[16],
         PUT_UINT32_BE( ctx->HL[8] >> 32, h,  8 );
         PUT_UINT32_BE( ctx->HL[8],       h, 12 );
 
-        (void) aesni_gcm_mult( output, x, h );
+        aesni_gcm_mult( output, x, h );
         return;
     }
 #endif