Merged ECP memory usage optimizations
diff --git a/CMakeLists.txt b/CMakeLists.txt
index 7b90993..257f6df 100644
--- a/CMakeLists.txt
+++ b/CMakeLists.txt
@@ -3,14 +3,25 @@
 
 enable_testing()
 
+string(REGEX MATCH "clang" CMAKE_COMPILER_IS_CLANG "${CMAKE_C_COMPILER}")
+
 if(CMAKE_COMPILER_IS_GNUCC)
   set(CMAKE_C_FLAGS "${CMAKE_C_FLAGS} -O2 -Wall -Wextra -W -Wdeclaration-after-statement")
   set(CMAKE_C_FLAGS_DEBUG "-g3 -O0")
   set(CMAKE_C_FLAGS_COVERAGE "-g3 -O0 -fprofile-arcs -ftest-coverage -lgcov")
-  set(CMAKE_C_FLAGS_CHECK "${CMAKE_C_FLAGS} -Wlogical-op -Wwrite-strings")
+  set(CMAKE_C_FLAGS_CHECK "${CMAKE_C_FLAGS} -Werror -Wlogical-op -Wwrite-strings")
   set(CMAKE_C_FLAGS_CHECKFULL "${CMAKE_C_FLAGS_CHECK} -Wcast-qual")
 endif(CMAKE_COMPILER_IS_GNUCC)
 
+if(CMAKE_COMPILER_IS_CLANG)
+  set(CMAKE_C_FLAGS "${CMAKE_C_FLAGS} -O2 -Wall -Wextra -W -Wdeclaration-after-statement")
+  set(CMAKE_C_FLAGS_DEBUG "-g3 -O0")
+  set(CMAKE_C_FLAGS_CHECK "${CMAKE_C_FLAGS} -Werror -Wpointer-arith -Wwrite-strings -Wdocumentation")
+endif(CMAKE_COMPILER_IS_CLANG)
+
+set(CMAKE_BUILD_TYPE ${CMAKE_BUILD_TYPE}
+    CACHE STRING "Choose the type of build: None Debug Release Coverage Check CheckFull"
+    FORCE)
 if(CMAKE_BUILD_TYPE STREQUAL "Coverage")
   if(CMAKE_COMPILER_IS_GNUCC)
     set(CMAKE_SHARED_LINKER_FLAGS "-fprofile-arcs -ftest-coverage")
diff --git a/ChangeLog b/ChangeLog
index cd3732f..030d720 100644
--- a/ChangeLog
+++ b/ChangeLog
@@ -8,6 +8,8 @@
    * Support for Curve25519
    * Support for ECDH-RSA and ECDH-ECDSA key exchanges and ciphersuites
    * Support for IPv6 in the NET module
+   * AES-NI support for AES, AES-GCM and AES key scheduling
+   * SSL Pthread-based server example added (ssl_pthread_server)
 
 Changes
    * gen_prime() speedup
@@ -17,6 +19,7 @@
    * More constant-time checks in the RSA module
    * Split off curves from ecp.c into ecp_curves.c
    * Curves are now stored fully in ROM
+   * Memory usage optimizations in ECP module
 
 Bugfix
    * Fixed bug in mpi_set_bit() on platforms where t_uint is wider than int
@@ -28,6 +31,10 @@
    * Memory leak in benchmark application
    * Fixed x509_crt_parse_path() bug on Windows platforms
 
+Security
+   * Possible remotely-triggered out-of-bounds memory access fixed (found by
+     TrustInSoft)
+
 = PolarSSL 1.3.2 released on 2013-11-04
 Features
    * PK tests added to test framework
diff --git a/include/polarssl/aes.h b/include/polarssl/aes.h
index 1b93e2a..9513a71 100644
--- a/include/polarssl/aes.h
+++ b/include/polarssl/aes.h
@@ -38,6 +38,7 @@
 #include <inttypes.h>
 #endif
 
+/* padlock.c and aesni.c rely on these values! */
 #define AES_ENCRYPT     1
 #define AES_DECRYPT     0
 
@@ -54,6 +55,11 @@
 
 /**
  * \brief          AES context structure
+ *
+ * \note           buf is able to hold 32 extra bytes, which can be used:
+ *                 - for alignment purposes if VIA padlock is used, and/or
+ *                 - to simplify key expansion in the 256-bit case by
+ *                 generating an extra round key
  */
 typedef struct
 {
diff --git a/include/polarssl/aesni.h b/include/polarssl/aesni.h
new file mode 100644
index 0000000..92b23cd
--- /dev/null
+++ b/include/polarssl/aesni.h
@@ -0,0 +1,107 @@
+/**
+ * \file aesni.h
+ *
+ * \brief AES-NI for hardware AES acceleration on some Intel processors
+ *
+ *  Copyright (C) 2013, Brainspark B.V.
+ *
+ *  This file is part of PolarSSL (http://www.polarssl.org)
+ *  Lead Maintainer: Paul Bakker <polarssl_maintainer at polarssl.org>
+ *
+ *  All rights reserved.
+ *
+ *  This program is free software; you can redistribute it and/or modify
+ *  it under the terms of the GNU General Public License as published by
+ *  the Free Software Foundation; either version 2 of the License, or
+ *  (at your option) any later version.
+ *
+ *  This program is distributed in the hope that it will be useful,
+ *  but WITHOUT ANY WARRANTY; without even the implied warranty of
+ *  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ *  GNU General Public License for more details.
+ *
+ *  You should have received a copy of the GNU General Public License along
+ *  with this program; if not, write to the Free Software Foundation, Inc.,
+ *  51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+ */
+#ifndef POLARSSL_AESNI_H
+#define POLARSSL_AESNI_H
+
+#include "aes.h"
+
+#define POLARSSL_AESNI_AES      0x02000000u
+#define POLARSSL_AESNI_CLMUL    0x00000002u
+
+#if defined(POLARSSL_HAVE_ASM) && defined(__GNUC__) &&  \
+    ( defined(__amd64__) || defined(__x86_64__) )   &&  \
+    ! defined(POLARSSL_HAVE_X86_64)
+#define POLARSSL_HAVE_X86_64
+#endif
+
+#if defined(POLARSSL_HAVE_X86_64)
+
+/**
+ * \brief          AES-NI features detection routine
+ *
+ * \param what     The feature to detect
+ *                 (POLARSSL_AESNI_AES or POLARSSL_AESNI_CLMUL)
+ *
+ * \return         1 if CPU has support for the feature, 0 otherwise
+ */
+int aesni_supports( unsigned int what );
+
+/**
+ * \brief          AES-NI AES-ECB block en(de)cryption
+ *
+ * \param ctx      AES context
+ * \param mode     AES_ENCRYPT or AES_DECRYPT
+ * \param input    16-byte input block
+ * \param output   16-byte output block
+ *
+ * \return         0 on success (cannot fail)
+ */
+int aesni_crypt_ecb( aes_context *ctx,
+                     int mode,
+                     const unsigned char input[16],
+                     unsigned char output[16] );
+
+/**
+ * \brief          GCM multiplication: c = a * b in GF(2^128)
+ *
+ * \param c        Result
+ * \param a        First operand
+ * \param b        Second operand
+ *
+ * \note           Both operands and result are bit strings interpreted as
+ *                 elements of GF(2^128) as per the GCM spec.
+ */
+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
+ *
+ * \param invkey    Round keys for the equivalent inverse cipher
+ * \param fwdkey    Original round keys (for encryption)
+ * \param nr        Number of rounds (that is, number of round keys minus one)
+ */
+void aesni_inverse_key( unsigned char *invkey,
+                        const unsigned char *fwdkey, int nr );
+
+/**
+ * \brief           Perform key expansion (for encryption)
+ *
+ * \param rk        Destination buffer where the round keys are written
+ * \param key       Encryption key
+ * \param bits      Key size in bits (must be 128, 192 or 256)
+ *
+ * \return          0 if successful, or POLARSSL_ERR_AES_INVALID_KEY_LENGTH
+ */
+int aesni_setkey_enc( unsigned char *rk,
+                      const unsigned char *key,
+                      size_t bits );
+
+#endif /* POLARSSL_HAVE_X86_64 */
+
+#endif /* POLARSSL_AESNI_H */
diff --git a/include/polarssl/config.h b/include/polarssl/config.h
index 16643dc..dfd107a 100644
--- a/include/polarssl/config.h
+++ b/include/polarssl/config.h
@@ -875,6 +875,20 @@
  */
 
 /**
+ * \def POLARSSL_AESNI_C
+ *
+ * Enable AES-NI support on x86-64.
+ *
+ * Module:  library/aesni.c
+ * Caller:  library/aes.c
+ *
+ * Requires: POLARSSL_HAVE_ASM
+ *
+ * This modules adds support for the AES-NI instructions on x86-64
+ */
+#define POLARSSL_AESNI_C
+
+/**
  * \def POLARSSL_AES_C
  *
  * Enable the AES block cipher.
@@ -1414,6 +1428,8 @@
  * Module:  library/padlock.c
  * Caller:  library/aes.c
  *
+ * Requires: POLARSSL_HAVE_ASM
+ *
  * This modules adds support for the VIA PadLock on x86.
  */
 #define POLARSSL_PADLOCK_C
@@ -1903,6 +1919,10 @@
 /*
  * Sanity checks on defines and dependencies
  */
+#if defined(POLARSSL_AESNI_C) && !defined(POLARSSL_HAVE_ASM)
+#error "POLARSSL_AESNI_C defined, but not all prerequisites"
+#endif
+
 #if defined(POLARSSL_CERTS_C) && !defined(POLARSSL_PEM_PARSE_C)
 #error "POLARSSL_CERTS_C defined, but not all prerequisites"
 #endif
@@ -2013,6 +2033,10 @@
 #error "POLARSSL_MEMORY_BUFFER_ALLOC_C defined, but not all prerequisites"
 #endif
 
+#if defined(POLARSSL_PADLOCK_C) && !defined(POLARSSL_HAVE_ASM)
+#error "POLARSSL_PADLOCK_C defined, but not all prerequisites"
+#endif
+
 #if defined(POLARSSL_PBKDF2_C) && !defined(POLARSSL_MD_C)
 #error "POLARSSL_PBKDF2_C defined, but not all prerequisites"
 #endif
diff --git a/include/polarssl/ecdh.h b/include/polarssl/ecdh.h
index d20f9bf..3c4800b 100644
--- a/include/polarssl/ecdh.h
+++ b/include/polarssl/ecdh.h
@@ -148,7 +148,7 @@
  *
  * \param ctx       ECDH constext to set
  * \param key       EC key to use
- * \param ours      Is it our key (1) or the peer's key (0) ?
+ * \param side      Is it our key (1) or the peer's key (0) ?
  *
  * \return          0 if successful, or an POLARSSL_ERR_ECP_XXX error code
  */
diff --git a/include/polarssl/ecp.h b/include/polarssl/ecp.h
index 23351a7..1e9e737 100644
--- a/include/polarssl/ecp.h
+++ b/include/polarssl/ecp.h
@@ -237,7 +237,7 @@
 /**
  * \brief           Get curve information from a TLS NamedCurve value
  *
- * \param grp_id    A POLARSSL_ECP_DP_XXX value
+ * \param tls_id    A POLARSSL_ECP_DP_XXX value
  *
  * \return          The associated curve information or NULL
  */
diff --git a/include/polarssl/entropy.h b/include/polarssl/entropy.h
index 235b773..7e65b78 100644
--- a/include/polarssl/entropy.h
+++ b/include/polarssl/entropy.h
@@ -81,7 +81,8 @@
  * \return          0 if no critical failures occurred,
  *                  POLARSSL_ERR_ENTROPY_SOURCE_FAILED otherwise
  */
-typedef int (*f_source_ptr)(void *, unsigned char *, size_t, size_t *);
+typedef int (*f_source_ptr)(void *data, unsigned char *output, size_t len,
+                            size_t *olen);
 
 /**
  * \brief           Entropy source state
diff --git a/include/polarssl/gcm.h b/include/polarssl/gcm.h
index e4267c6..c2829a0 100644
--- a/include/polarssl/gcm.h
+++ b/include/polarssl/gcm.h
@@ -201,7 +201,7 @@
 /**
  * \brief           Free a GCM context and underlying cipher sub-context
  *
- * \param ctx
+ * \param ctx       GCM context to free
  */
 void gcm_free( gcm_context *ctx );
 
diff --git a/include/polarssl/padlock.h b/include/polarssl/padlock.h
index 8df93c0..545120f 100644
--- a/include/polarssl/padlock.h
+++ b/include/polarssl/padlock.h
@@ -59,7 +59,7 @@
 /**
  * \brief          PadLock detection routine
  *
- * \param          The feature to detect
+ * \param feature  The feature to detect
  *
  * \return         1 if CPU has support for the feature, 0 otherwise
  */
diff --git a/include/polarssl/pk.h b/include/polarssl/pk.h
index 958672b..013d973 100644
--- a/include/polarssl/pk.h
+++ b/include/polarssl/pk.h
@@ -453,14 +453,14 @@
  *                        return value to determine where you should start
  *                        using the buffer
  *
- * \param key       private to write away
+ * \param ctx       private to write away
  * \param buf       buffer to write to
  * \param size      size of the buffer
  *
  * \return          length of data written if successful, or a specific
  *                  error code
  */
-int pk_write_key_der( pk_context *pk, unsigned char *buf, size_t size );
+int pk_write_key_der( pk_context *ctx, unsigned char *buf, size_t size );
 
 /**
  * \brief           Write a public key to a SubjectPublicKeyInfo DER structure
@@ -468,37 +468,37 @@
  *                        return value to determine where you should start
  *                        using the buffer
  *
- * \param key       public key to write away
+ * \param ctx       public key to write away
  * \param buf       buffer to write to
  * \param size      size of the buffer
  *
  * \return          length of data written if successful, or a specific
  *                  error code
  */
-int pk_write_pubkey_der( pk_context *key, unsigned char *buf, size_t size );
+int pk_write_pubkey_der( pk_context *ctx, unsigned char *buf, size_t size );
 
 #if defined(POLARSSL_PEM_WRITE_C)
 /**
  * \brief           Write a public key to a PEM string
  *
- * \param key       public key to write away
+ * \param ctx       public key to write away
  * \param buf       buffer to write to
  * \param size      size of the buffer
  *
  * \return          0 successful, or a specific error code
  */
-int pk_write_pubkey_pem( pk_context *key, unsigned char *buf, size_t size );
+int pk_write_pubkey_pem( pk_context *ctx, unsigned char *buf, size_t size );
 
 /**
  * \brief           Write a private key to a PKCS#1 or SEC1 PEM string
  *
- * \param key       private to write away
+ * \param ctx       private to write away
  * \param buf       buffer to write to
  * \param size      size of the buffer
  *
  * \return          0 successful, or a specific error code
  */
-int pk_write_key_pem( pk_context *key, unsigned char *buf, size_t size );
+int pk_write_key_pem( pk_context *ctx, unsigned char *buf, size_t size );
 #endif /* POLARSSL_PEM_WRITE_C */
 #endif /* POLARSSL_PK_WRITE_C */
 
diff --git a/include/polarssl/pkcs12.h b/include/polarssl/pkcs12.h
index 51bea3d..4bd5018 100644
--- a/include/polarssl/pkcs12.h
+++ b/include/polarssl/pkcs12.h
@@ -38,9 +38,9 @@
 #define POLARSSL_ERR_PKCS12_PBE_INVALID_FORMAT             -0x1E80  /**< PBE ASN.1 data not as expected. */
 #define POLARSSL_ERR_PKCS12_PASSWORD_MISMATCH              -0x1E00  /**< Given private key password does not allow for correct decryption. */
 
-#define PKCS12_DERIVE_KEY       1   /*< encryption/decryption key */
-#define PKCS12_DERIVE_IV        2   /*< initialization vector     */
-#define PKCS12_DERIVE_MAC_KEY   3   /*< integrity / MAC key       */
+#define PKCS12_DERIVE_KEY       1   /**< encryption/decryption key */
+#define PKCS12_DERIVE_IV        2   /**< initialization vector     */
+#define PKCS12_DERIVE_MAC_KEY   3   /**< integrity / MAC key       */
 
 #define PKCS12_PBE_DECRYPT      0
 #define PKCS12_PBE_ENCRYPT      1
diff --git a/include/polarssl/rsa.h b/include/polarssl/rsa.h
index e7b6191..504dde2 100644
--- a/include/polarssl/rsa.h
+++ b/include/polarssl/rsa.h
@@ -526,7 +526,6 @@
 
 /**
  * \brief          Perform a PKCS#1 v2.1 PSS verification (RSASSA-PSS-VERIFY)
- * \brief          Do a public RSA and check the message digest
  *
  * \param ctx      points to an RSA public key
  * \param f_rng    RNG function (Only needed for RSA_PRIVATE)
diff --git a/include/polarssl/ssl.h b/include/polarssl/ssl.h
index 4b0c5f8..7e668f9 100644
--- a/include/polarssl/ssl.h
+++ b/include/polarssl/ssl.h
@@ -237,8 +237,8 @@
 #endif /* !POLARSSL_CONFIG_OPTIONS */
 
 /*
- * Allow an extra 512 bytes for the record header
- * and encryption overhead (counter + MAC + padding)
+ * Allow an extra 301 bytes for the record header
+ * and encryption overhead: counter (8) + header (5) + MAC (32) + padding (256)
  * and allow for a maximum of 1024 of compression expansion if
  * enabled.
  */
@@ -248,9 +248,9 @@
 #define SSL_COMPRESSION_ADD             0
 #endif
 
-#define SSL_BUFFER_LEN (SSL_MAX_CONTENT_LEN + SSL_COMPRESSION_ADD + 512)
+#define SSL_BUFFER_LEN (SSL_MAX_CONTENT_LEN + SSL_COMPRESSION_ADD + 301)
 
-#define SSL_EMPTY_RENEGOTIATION_INFO    0xFF   /**< renegotiation info ext */ 
+#define SSL_EMPTY_RENEGOTIATION_INFO    0xFF   /**< renegotiation info ext */
 
 /*
  * Supported Signature and Hash algorithms (For TLS 1.2)
diff --git a/include/polarssl/ssl_ciphersuites.h b/include/polarssl/ssl_ciphersuites.h
index 3b13ef3..28a5855 100644
--- a/include/polarssl/ssl_ciphersuites.h
+++ b/include/polarssl/ssl_ciphersuites.h
@@ -226,7 +226,7 @@
 
 typedef struct _ssl_ciphersuite_t ssl_ciphersuite_t;
 
-#define POLARSSL_CIPHERSUITE_WEAK   0x01    /*<! Weak ciphersuite flag      */
+#define POLARSSL_CIPHERSUITE_WEAK   0x01    /**< Weak ciphersuite flag      */
 
 /**
  * \brief   This structure is used for storing ciphersuite information
diff --git a/include/polarssl/x509_crt.h b/include/polarssl/x509_crt.h
index 4e7bbb7..ee8f9e6 100644
--- a/include/polarssl/x509_crt.h
+++ b/include/polarssl/x509_crt.h
@@ -371,7 +371,7 @@
  *                  (e.g. POLARSSL_MD_SHA1)
  *
  * \param ctx       CRT context to use
- * \param md_ald    MD algorithm to use
+ * \param md_alg    MD algorithm to use
  */
 void x509write_crt_set_md_alg( x509write_cert *ctx, md_type_t md_alg );
 
@@ -467,7 +467,7 @@
  *                        return value to determine where you should start
  *                        using the buffer
  *
- * \param crt       certificate to write away
+ * \param ctx       certificate to write away
  * \param buf       buffer to write to
  * \param size      size of the buffer
  * \param f_rng     RNG function (for signature, see note)
@@ -489,7 +489,7 @@
 /**
  * \brief           Write a built up certificate to a X509 PEM string
  *
- * \param crt       certificate to write away
+ * \param ctx       certificate to write away
  * \param buf       buffer to write to
  * \param size      size of the buffer
  * \param f_rng     RNG function (for signature, see note)
diff --git a/library/CMakeLists.txt b/library/CMakeLists.txt
index 411c07d..1a5efde 100644
--- a/library/CMakeLists.txt
+++ b/library/CMakeLists.txt
@@ -2,6 +2,7 @@
 
 set(src
      aes.c
+     aesni.c
      arc4.c
      asn1parse.c
      asn1write.c
diff --git a/library/Makefile b/library/Makefile
index a9c86f8..3f80162 100644
--- a/library/Makefile
+++ b/library/Makefile
@@ -34,7 +34,8 @@
 LDFLAGS += -lws2_32
 endif
 
-OBJS=	aes.o		arc4.o		asn1parse.o		\
+OBJS=	aes.o		aesni.o		arc4.o			\
+		asn1parse.o								\
 		asn1write.o base64.o	bignum.o		\
 		blowfish.o	camellia.o					\
 		certs.o		cipher.o	cipher_wrap.o	\
diff --git a/library/aes.c b/library/aes.c
index a3835ce..69da3f6 100644
--- a/library/aes.c
+++ b/library/aes.c
@@ -37,6 +37,9 @@
 #if defined(POLARSSL_PADLOCK_C)
 #include "polarssl/padlock.h"
 #endif
+#if defined(POLARSSL_AESNI_C)
+#include "polarssl/aesni.h"
+#endif
 
 #if !defined(POLARSSL_AES_ALT)
 
@@ -480,6 +483,11 @@
 #endif
     ctx->rk = RK = ctx->buf;
 
+#if defined(POLARSSL_AESNI_C) && defined(POLARSSL_HAVE_X86_64)
+    if( aesni_supports( POLARSSL_AESNI_AES ) )
+        return( aesni_setkey_enc( (unsigned char *) ctx->rk, key, keysize ) );
+#endif
+
     for( i = 0; i < (keysize >> 5); i++ )
     {
         GET_UINT32_LE( RK[i], key, i << 2 );
@@ -588,6 +596,15 @@
     if( ret != 0 )
         return( ret );
 
+#if defined(POLARSSL_AESNI_C) && defined(POLARSSL_HAVE_X86_64)
+    if( aesni_supports( POLARSSL_AESNI_AES ) )
+    {
+        aesni_inverse_key( (unsigned char *) ctx->rk,
+                           (const unsigned char *) cty.rk, ctx->nr );
+        goto done;
+    }
+#endif
+
     SK = cty.rk + cty.nr * 4;
 
     *RK++ = *SK++;
@@ -611,6 +628,9 @@
     *RK++ = *SK++;
     *RK++ = *SK++;
 
+#if defined(POLARSSL_AESNI_C) && defined(POLARSSL_HAVE_X86_64)
+done:
+#endif
     memset( &cty, 0, sizeof( aes_context ) );
 
     return( 0 );
@@ -673,6 +693,11 @@
     int i;
     uint32_t *RK, X0, X1, X2, X3, Y0, Y1, Y2, Y3;
 
+#if defined(POLARSSL_AESNI_C) && defined(POLARSSL_HAVE_X86_64)
+    if( aesni_supports( POLARSSL_AESNI_AES ) )
+        return( aesni_crypt_ecb( ctx, mode, input, output ) );
+#endif
+
 #if defined(POLARSSL_PADLOCK_C) && defined(POLARSSL_HAVE_X86)
     if( aes_padlock_ace )
     {
diff --git a/library/aesni.c b/library/aesni.c
new file mode 100644
index 0000000..aa054cb
--- /dev/null
+++ b/library/aesni.c
@@ -0,0 +1,433 @@
+/*
+ *  AES-NI support functions
+ *
+ *  Copyright (C) 2013, Brainspark B.V.
+ *
+ *  This file is part of PolarSSL (http://www.polarssl.org)
+ *  Lead Maintainer: Paul Bakker <polarssl_maintainer at polarssl.org>
+ *
+ *  All rights reserved.
+ *
+ *  This program is free software; you can redistribute it and/or modify
+ *  it under the terms of the GNU General Public License as published by
+ *  the Free Software Foundation; either version 2 of the License, or
+ *  (at your option) any later version.
+ *
+ *  This program is distributed in the hope that it will be useful,
+ *  but WITHOUT ANY WARRANTY; without even the implied warranty of
+ *  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ *  GNU General Public License for more details.
+ *
+ *  You should have received a copy of the GNU General Public License along
+ *  with this program; if not, write to the Free Software Foundation, Inc.,
+ *  51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+ */
+
+/*
+ * [AES-WP] http://software.intel.com/en-us/articles/intel-advanced-encryption-standard-aes-instructions-set
+ * [CLMUL-WP] http://software.intel.com/en-us/articles/intel-carry-less-multiplication-instruction-and-its-usage-for-computing-the-gcm-mode/
+ */
+
+#include "polarssl/config.h"
+
+#if defined(POLARSSL_AESNI_C)
+
+#include "polarssl/aesni.h"
+#include <stdio.h>
+
+#if defined(POLARSSL_HAVE_X86_64)
+
+/*
+ * AES-NI support detection routine
+ */
+int aesni_supports( unsigned int what )
+{
+    static int done = 0;
+    static unsigned int c = 0;
+
+    if( ! done )
+    {
+        asm( "movl  $1, %%eax   \n"
+             "cpuid             \n"
+             : "=c" (c)
+             :
+             : "eax", "ebx", "edx" );
+        done = 1;
+    }
+
+    return( ( c & what ) != 0 );
+}
+
+/*
+ * AES-NI AES-ECB block en(de)cryption
+ */
+int aesni_crypt_ecb( aes_context *ctx,
+                     int mode,
+                     const unsigned char input[16],
+                     unsigned char output[16] )
+{
+    asm( "movdqu    (%3), %%xmm0    \n" // load input
+         "movdqu    (%1), %%xmm1    \n" // load round key 0
+         "pxor      %%xmm1, %%xmm0  \n" // round 0
+         "addq      $16, %1         \n" // point to next round key
+         "subl      $1, %0          \n" // normal rounds = nr - 1
+         "test      %2, %2          \n" // mode?
+         "jz        2f              \n" // 0 = decrypt
+
+         "1:                        \n" // encryption loop
+         "movdqu    (%1), %%xmm1    \n" // load round key
+         "aesenc    %%xmm1, %%xmm0  \n" // do round
+         "addq      $16, %1         \n" // point to next round key
+         "subl      $1, %0          \n" // loop
+         "jnz       1b              \n"
+         "movdqu    (%1), %%xmm1    \n" // load round key
+         "aesenclast %%xmm1, %%xmm0 \n" // last round
+         "jmp       3f              \n"
+
+         "2:                        \n" // decryption loop
+         "movdqu    (%1), %%xmm1    \n"
+         "aesdec    %%xmm1, %%xmm0  \n"
+         "addq      $16, %1         \n"
+         "subl      $1, %0          \n"
+         "jnz       2b              \n"
+         "movdqu    (%1), %%xmm1    \n" // load round key
+         "aesdeclast %%xmm1, %%xmm0 \n" // last round
+
+         "3:                        \n"
+         "movdqu    %%xmm0, (%4)    \n" // export output
+         :
+         : "r" (ctx->nr), "r" (ctx->rk), "r" (mode), "r" (input), "r" (output)
+         : "memory", "cc", "xmm0", "xmm1" );
+
+
+    return( 0 );
+}
+
+/*
+ * GCM multiplication: c = a times b in GF(2^128)
+ * Based on [CLMUL-WP] algorithms 1 (with equation 27) and 5.
+ */
+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;
+
+    /* The inputs are in big-endian order, so byte-reverse them */
+    for( i = 0; i < 16; i++ )
+    {
+        aa[i] = a[15 - i];
+        bb[i] = b[15 - i];
+    }
+
+    asm( "movdqu (%0), %%xmm0               \n" // a1:a0
+         "movdqu (%1), %%xmm1               \n" // b1:b0
+
+         /*
+          * Caryless multiplication xmm2:xmm1 = xmm0 * xmm1
+          * using [CLMUL-WP] algorithm 1 (p. 13).
+          */
+         "movdqa %%xmm1, %%xmm2             \n" // copy of b1:b0
+         "movdqa %%xmm1, %%xmm3             \n" // same
+         "movdqa %%xmm1, %%xmm4             \n" // same
+         "pclmulqdq $0x00, %%xmm0, %%xmm1   \n" // a0*b0 = c1:c0
+         "pclmulqdq $0x11, %%xmm0, %%xmm2   \n" // a1*b1 = d1:d0
+         "pclmulqdq $0x10, %%xmm0, %%xmm3   \n" // a0*b1 = e1:e0
+         "pclmulqdq $0x01, %%xmm0, %%xmm4   \n" // a1*b0 = f1:f0
+         "pxor %%xmm3, %%xmm4               \n" // e1+f1:e0+f0
+         "movdqa %%xmm4, %%xmm3             \n" // same
+         "psrldq $8, %%xmm4                 \n" // 0:e1+f1
+         "pslldq $8, %%xmm3                 \n" // e0+f0:0
+         "pxor %%xmm4, %%xmm2               \n" // d1:d0+e1+f1
+         "pxor %%xmm3, %%xmm1               \n" // c1+e0+f1:c0
+
+         /*
+          * Now shift the result one bit to the left,
+          * taking advantage of [CLMUL-WP] eq 27 (p. 20)
+          */
+         "movdqa %%xmm1, %%xmm3             \n" // r1:r0
+         "movdqa %%xmm2, %%xmm4             \n" // r3:r2
+         "psllq $1, %%xmm1                  \n" // r1<<1:r0<<1
+         "psllq $1, %%xmm2                  \n" // r3<<1:r2<<1
+         "psrlq $63, %%xmm3                 \n" // r1>>63:r0>>63
+         "psrlq $63, %%xmm4                 \n" // r3>>63:r2>>63
+         "movdqa %%xmm3, %%xmm5             \n" // r1>>63:r0>>63
+         "pslldq $8, %%xmm3                 \n" // r0>>63:0
+         "pslldq $8, %%xmm4                 \n" // r2>>63:0
+         "psrldq $8, %%xmm5                 \n" // 0:r1>>63
+         "por %%xmm3, %%xmm1                \n" // r1<<1|r0>>63:r0<<1
+         "por %%xmm4, %%xmm2                \n" // r3<<1|r2>>62:r2<<1
+         "por %%xmm5, %%xmm2                \n" // r3<<1|r2>>62:r2<<1|r1>>63
+
+         /*
+          * Now reduce modulo the GCM polynomial x^128 + x^7 + x^2 + x + 1
+          * using [CLMUL-WP] algorithm 5 (p. 20).
+          * Currently xmm2:xmm1 holds x3:x2:x1:x0 (already shifted).
+          */
+         /* Step 2 (1) */
+         "movdqa %%xmm1, %%xmm3             \n" // x1:x0
+         "movdqa %%xmm1, %%xmm4             \n" // same
+         "movdqa %%xmm1, %%xmm5             \n" // same
+         "psllq $63, %%xmm3                 \n" // x1<<63:x0<<63 = stuff:a
+         "psllq $62, %%xmm4                 \n" // x1<<62:x0<<62 = stuff:b
+         "psllq $57, %%xmm5                 \n" // x1<<57:x0<<57 = stuff:c
+
+         /* Step 2 (2) */
+         "pxor %%xmm4, %%xmm3               \n" // stuff:a+b
+         "pxor %%xmm5, %%xmm3               \n" // stuff:a+b+c
+         "pslldq $8, %%xmm3                 \n" // a+b+c:0
+         "pxor %%xmm3, %%xmm1               \n" // x1+a+b+c:x0 = d:x0
+
+         /* Steps 3 and 4 */
+         "movdqa %%xmm1,%%xmm0              \n" // d:x0
+         "movdqa %%xmm1,%%xmm4              \n" // same
+         "movdqa %%xmm1,%%xmm5              \n" // same
+         "psrlq $1, %%xmm0                  \n" // e1:x0>>1 = e1:e0'
+         "psrlq $2, %%xmm4                  \n" // f1:x0>>2 = f1:f0'
+         "psrlq $7, %%xmm5                  \n" // g1:x0>>7 = g1:g0'
+         "pxor %%xmm4, %%xmm0               \n" // e1+f1:e0'+f0'
+         "pxor %%xmm5, %%xmm0               \n" // e1+f1+g1:e0'+f0'+g0'
+         // e0'+f0'+g0' is almost e0+f0+g0, except for some missing
+         // bits carried from d. Now get those bits back in.
+         "movdqa %%xmm1,%%xmm3              \n" // d:x0
+         "movdqa %%xmm1,%%xmm4              \n" // same
+         "movdqa %%xmm1,%%xmm5              \n" // same
+         "psllq $63, %%xmm3                 \n" // d<<63:stuff
+         "psllq $62, %%xmm4                 \n" // d<<62:stuff
+         "psllq $57, %%xmm5                 \n" // d<<57:stuff
+         "pxor %%xmm4, %%xmm3               \n" // d<<63+d<<62:stuff
+         "pxor %%xmm5, %%xmm3               \n" // missing bits of d:stuff
+         "psrldq $8, %%xmm3                 \n" // 0:missing bits of d
+         "pxor %%xmm3, %%xmm0               \n" // e1+f1+g1:e0+f0+g0
+         "pxor %%xmm1, %%xmm0               \n" // h1:h0
+         "pxor %%xmm2, %%xmm0               \n" // x3+h1:x2+h0
+
+         "movdqu %%xmm0, (%2)               \n" // done
+         :
+         : "r" (aa), "r" (bb), "r" (cc)
+         : "memory", "cc", "xmm0", "xmm1", "xmm2", "xmm3", "xmm4", "xmm5" );
+
+    /* Now byte-reverse the outputs */
+    for( i = 0; i < 16; i++ )
+        c[i] = cc[15 - i];
+
+    return;
+}
+
+/*
+ * Compute decryption round keys from encryption round keys
+ */
+void aesni_inverse_key( unsigned char *invkey,
+                        const unsigned char *fwdkey, int nr )
+{
+    unsigned char *ik = invkey;
+    const unsigned char *fk = fwdkey + 16 * nr;
+
+    memcpy( ik, fk, 16 );
+
+    for( fk -= 16, ik += 16; fk > fwdkey; fk -= 16, ik += 16 )
+        asm( "movdqu (%0), %%xmm0       \n"
+             "aesimc %%xmm0, %%xmm0     \n"
+             "movdqu %%xmm0, (%1)       \n"
+             :
+             : "r" (fk), "r" (ik)
+             : "memory", "xmm0" );
+
+    memcpy( ik, fk, 16 );
+}
+
+/*
+ * Key expansion, 128-bit case
+ */
+static void aesni_setkey_enc_128( unsigned char *rk,
+                                  const unsigned char *key )
+{
+    asm( "movdqu (%1), %%xmm0               \n" // copy the original key
+         "movdqu %%xmm0, (%0)               \n" // as round key 0
+         "jmp 2f                            \n" // skip auxiliary routine
+
+         /*
+          * Finish generating the next round key.
+          *
+          * On entry xmm0 is r3:r2:r1:r0 and xmm1 is X:stuff:stuff:stuff
+          * with X = rot( sub( r3 ) ) ^ RCON.
+          *
+          * On exit, xmm0 is r7:r6:r5:r4
+          * with r4 = X + r0, r5 = r4 + r1, r6 = r5 + r2, r7 = r6 + r3
+          * and those are written to the round key buffer.
+          */
+         "1:                                \n"
+         "pshufd $0xff, %%xmm1, %%xmm1      \n" // X:X:X:X
+         "pxor %%xmm0, %%xmm1               \n" // X+r3:X+r2:X+r1:r4
+         "pslldq $4, %%xmm0                 \n" // r2:r1:r0:0
+         "pxor %%xmm0, %%xmm1               \n" // X+r3+r2:X+r2+r1:r5:r4
+         "pslldq $4, %%xmm0                 \n" // etc
+         "pxor %%xmm0, %%xmm1               \n"
+         "pslldq $4, %%xmm0                 \n"
+         "pxor %%xmm1, %%xmm0               \n" // update xmm0 for next time!
+         "add $16, %0                       \n" // point to next round key
+         "movdqu %%xmm0, (%0)               \n" // write it
+         "ret                               \n"
+
+         /* Main "loop" */
+         "2:                                    \n"
+         "aeskeygenassist $0x01, %%xmm0, %%xmm1 \ncall 1b   \n"
+         "aeskeygenassist $0x02, %%xmm0, %%xmm1 \ncall 1b   \n"
+         "aeskeygenassist $0x04, %%xmm0, %%xmm1 \ncall 1b   \n"
+         "aeskeygenassist $0x08, %%xmm0, %%xmm1 \ncall 1b   \n"
+         "aeskeygenassist $0x10, %%xmm0, %%xmm1 \ncall 1b   \n"
+         "aeskeygenassist $0x20, %%xmm0, %%xmm1 \ncall 1b   \n"
+         "aeskeygenassist $0x40, %%xmm0, %%xmm1 \ncall 1b   \n"
+         "aeskeygenassist $0x80, %%xmm0, %%xmm1 \ncall 1b   \n"
+         "aeskeygenassist $0x1B, %%xmm0, %%xmm1 \ncall 1b   \n"
+         "aeskeygenassist $0x36, %%xmm0, %%xmm1 \ncall 1b   \n"
+         :
+         : "r" (rk), "r" (key)
+         : "memory", "cc", "0" );
+}
+
+/*
+ * Key expansion, 192-bit case
+ */
+static void aesni_setkey_enc_192( unsigned char *rk,
+                                  const unsigned char *key )
+{
+    asm( "movdqu (%1), %%xmm0   \n" // copy original round key
+         "movdqu %%xmm0, (%0)   \n"
+         "add $16, %0           \n"
+         "movq 16(%1), %%xmm1   \n"
+         "movq %%xmm1, (%0)     \n"
+         "add $8, %0            \n"
+         "jmp 2f                \n" // skip auxiliary routine
+
+         /*
+          * Finish generating the next 6 quarter-keys.
+          *
+          * On entry xmm0 is r3:r2:r1:r0, xmm1 is stuff:stuff:r5:r4
+          * and xmm2 is stuff:stuff:X:stuff with X = rot( sub( r3 ) ) ^ RCON.
+          *
+          * On exit, xmm0 is r9:r8:r7:r6 and xmm1 is stuff:stuff:r11:r10
+          * and those are written to the round key buffer.
+          */
+         "1:                            \n"
+         "pshufd $0x55, %%xmm2, %%xmm2  \n" // X:X:X:X
+         "pxor %%xmm0, %%xmm2           \n" // X+r3:X+r2:X+r1:r4
+         "pslldq $4, %%xmm0             \n" // etc
+         "pxor %%xmm0, %%xmm2           \n"
+         "pslldq $4, %%xmm0             \n"
+         "pxor %%xmm0, %%xmm2           \n"
+         "pslldq $4, %%xmm0             \n"
+         "pxor %%xmm2, %%xmm0           \n" // update xmm0 = r9:r8:r7:r6
+         "movdqu %%xmm0, (%0)           \n"
+         "add $16, %0                   \n"
+         "pshufd $0xff, %%xmm0, %%xmm2  \n" // r9:r9:r9:r9
+         "pxor %%xmm1, %%xmm2           \n" // stuff:stuff:r9+r5:r10
+         "pslldq $4, %%xmm1             \n" // r2:r1:r0:0
+         "pxor %%xmm2, %%xmm1           \n" // update xmm1 = stuff:stuff:r11:r10
+         "movq %%xmm1, (%0)             \n"
+         "add $8, %0                    \n"
+         "ret                           \n"
+
+         "2:                                    \n"
+         "aeskeygenassist $0x01, %%xmm1, %%xmm2 \ncall 1b   \n"
+         "aeskeygenassist $0x02, %%xmm1, %%xmm2 \ncall 1b   \n"
+         "aeskeygenassist $0x04, %%xmm1, %%xmm2 \ncall 1b   \n"
+         "aeskeygenassist $0x08, %%xmm1, %%xmm2 \ncall 1b   \n"
+         "aeskeygenassist $0x10, %%xmm1, %%xmm2 \ncall 1b   \n"
+         "aeskeygenassist $0x20, %%xmm1, %%xmm2 \ncall 1b   \n"
+         "aeskeygenassist $0x40, %%xmm1, %%xmm2 \ncall 1b   \n"
+         "aeskeygenassist $0x80, %%xmm1, %%xmm2 \ncall 1b   \n"
+
+         :
+         : "r" (rk), "r" (key)
+         : "memory", "cc", "0" );
+}
+
+/*
+ * Key expansion, 256-bit case
+ */
+static void aesni_setkey_enc_256( unsigned char *rk,
+                                  const unsigned char *key )
+{
+    asm( "movdqu (%1), %%xmm0           \n"
+         "movdqu %%xmm0, (%0)           \n"
+         "add $16, %0                   \n"
+         "movdqu 16(%1), %%xmm1         \n"
+         "movdqu %%xmm1, (%0)           \n"
+         "jmp 2f                        \n" // skip auxiliary routine
+
+         /*
+          * Finish generating the next two round keys.
+          *
+          * On entry xmm0 is r3:r2:r1:r0, xmm1 is r7:r6:r5:r4 and
+          * xmm2 is X:stuff:stuff:stuff with X = rot( sub( r7 )) ^ RCON
+          *
+          * On exit, xmm0 is r11:r10:r9:r8 and xmm1 is r15:r14:r13:r12
+          * and those have been written to the output buffer.
+          */
+         "1:                                \n"
+         "pshufd $0xff, %%xmm2, %%xmm2      \n"
+         "pxor %%xmm0, %%xmm2               \n"
+         "pslldq $4, %%xmm0                 \n"
+         "pxor %%xmm0, %%xmm2               \n"
+         "pslldq $4, %%xmm0                 \n"
+         "pxor %%xmm0, %%xmm2               \n"
+         "pslldq $4, %%xmm0                 \n"
+         "pxor %%xmm2, %%xmm0               \n"
+         "add $16, %0                       \n"
+         "movdqu %%xmm0, (%0)               \n"
+
+         /* Set xmm2 to stuff:Y:stuff:stuff with Y = subword( r11 )
+          * and proceed to generate next round key from there */
+         "aeskeygenassist $0, %%xmm0, %%xmm2\n"
+         "pshufd $0xaa, %%xmm2, %%xmm2      \n"
+         "pxor %%xmm1, %%xmm2               \n"
+         "pslldq $4, %%xmm1                 \n"
+         "pxor %%xmm1, %%xmm2               \n"
+         "pslldq $4, %%xmm1                 \n"
+         "pxor %%xmm1, %%xmm2               \n"
+         "pslldq $4, %%xmm1                 \n"
+         "pxor %%xmm2, %%xmm1               \n"
+         "add $16, %0                       \n"
+         "movdqu %%xmm1, (%0)               \n"
+         "ret                               \n"
+
+         /*
+          * Main "loop" - Generating one more key than necessary,
+          * see definition of aes_context.buf
+          */
+         "2:                                    \n"
+         "aeskeygenassist $0x01, %%xmm1, %%xmm2 \ncall 1b   \n"
+         "aeskeygenassist $0x02, %%xmm1, %%xmm2 \ncall 1b   \n"
+         "aeskeygenassist $0x04, %%xmm1, %%xmm2 \ncall 1b   \n"
+         "aeskeygenassist $0x08, %%xmm1, %%xmm2 \ncall 1b   \n"
+         "aeskeygenassist $0x10, %%xmm1, %%xmm2 \ncall 1b   \n"
+         "aeskeygenassist $0x20, %%xmm1, %%xmm2 \ncall 1b   \n"
+         "aeskeygenassist $0x40, %%xmm1, %%xmm2 \ncall 1b   \n"
+         :
+         : "r" (rk), "r" (key)
+         : "memory", "cc", "0" );
+}
+
+/*
+ * Key expansion, wrapper
+ */
+int aesni_setkey_enc( unsigned char *rk,
+                      const unsigned char *key,
+                      size_t bits )
+{
+    switch( bits )
+    {
+        case 128: aesni_setkey_enc_128( rk, key ); break;
+        case 192: aesni_setkey_enc_192( rk, key ); break;
+        case 256: aesni_setkey_enc_256( rk, key ); break;
+        default : return( POLARSSL_ERR_AES_INVALID_KEY_LENGTH );
+    }
+
+    return( 0 );
+}
+
+#endif /* POLARSSL_HAVE_X86_64 */
+
+#endif /* POLARSSL_AESNI_C */
diff --git a/library/ecp.c b/library/ecp.c
index 3adab15..8b34bf3 100644
--- a/library/ecp.c
+++ b/library/ecp.c
@@ -1870,7 +1870,7 @@
             mul_count != mul_c_prev )
         {
             if( verbose != 0 )
-                printf( "failed (%zu)\n", i );
+                printf( "failed (%u)\n", (unsigned int) i );
 
             ret = 1;
             goto cleanup;
@@ -1907,7 +1907,7 @@
             mul_count != mul_c_prev )
         {
             if( verbose != 0 )
-                printf( "failed (%zu)\n", i );
+                printf( "failed (%u)\n", (unsigned int) i );
 
             ret = 1;
             goto cleanup;
diff --git a/library/gcm.c b/library/gcm.c
index a9e18c8..8950360 100644
--- a/library/gcm.c
+++ b/library/gcm.c
@@ -22,15 +22,27 @@
  *  with this program; if not, write to the Free Software Foundation, Inc.,
  *  51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
  */
+
 /*
- *  http://csrc.nist.gov/publications/nistpubs/800-38D/SP-800-38D.pdf
+ * http://csrc.nist.gov/publications/nistpubs/800-38D/SP-800-38D.pdf
+ *
+ * See also:
+ * [MGV] http://csrc.nist.gov/groups/ST/toolkit/BCM/documents/proposedmodes/gcm/gcm-revised-spec.pdf
+ *
+ * We use the algorithm described as Shoup's method with 4-bit tables in
+ * [MGV] 4.1, pp. 12-13, to enhance speed without using too much memory.
  */
+
 #include "polarssl/config.h"
 
 #if defined(POLARSSL_GCM_C)
 
 #include "polarssl/gcm.h"
 
+#if defined(POLARSSL_AESNI_C)
+#include "polarssl/aesni.h"
+#endif
+
 /*
  * 32-bit integer manipulation macros (big endian)
  */
@@ -54,6 +66,14 @@
 }
 #endif
 
+/*
+ * Precompute small multiples of H, that is set
+ *      HH[i] || HL[i] = H times i,
+ * where i is seen as a field element as in [MGV], ie high-order bits
+ * correspond to low powers of P. The result is stored in the same way, that
+ * is the high-order bit of HH corresponds to P^0 and the low-order bit of HL
+ * corresponds to P^127.
+ */
 static int gcm_gen_table( gcm_context *ctx )
 {
     int ret, i, j;
@@ -66,9 +86,7 @@
     if( ( ret = cipher_update( &ctx->cipher_ctx, h, 16, h, &olen ) ) != 0 )
         return( ret );
 
-    ctx->HH[0] = 0;
-    ctx->HL[0] = 0;
-
+    /* pack h as two 64-bits ints, big-endian */
     GET_UINT32_BE( hi, h,  0  );
     GET_UINT32_BE( lo, h,  4  );
     vh = (uint64_t) hi << 32 | lo;
@@ -77,9 +95,20 @@
     GET_UINT32_BE( lo, h,  12 );
     vl = (uint64_t) hi << 32 | lo;
 
+    /* 8 = 1000 corresponds to 1 in GF(2^128) */
     ctx->HL[8] = vl;
     ctx->HH[8] = vh;
 
+#if defined(POLARSSL_AESNI_C) && defined(POLARSSL_HAVE_X86_64)
+    /* With CLMUL support, we need only h, not the rest of the table */
+    if( aesni_supports( POLARSSL_AESNI_CLMUL ) )
+        return( 0 );
+#endif
+
+    /* 0 corresponds to 0 in GF(2^128) */
+    ctx->HH[0] = 0;
+    ctx->HL[0] = 0;
+
     for( i = 4; i > 0; i >>= 1 )
     {
         uint32_t T = ( vl & 1 ) * 0xe1000000U;
@@ -135,6 +164,11 @@
     return( 0 );
 }
 
+/*
+ * Shoup's method for multiplication use this table with
+ *      last4[x] = x times P^128
+ * where x and last4[x] are seen as elements of GF(2^128) as in [MGV]
+ */
 static const uint64_t last4[16] =
 {
     0x0000, 0x1c20, 0x3840, 0x2460,
@@ -143,6 +177,10 @@
     0x9180, 0x8da0, 0xa9c0, 0xb5e0
 };
 
+/*
+ * Sets output to x times H using the precomputed tables.
+ * x and output are seen as elements of GF(2^128) as in [MGV].
+ */
 static void gcm_mult( gcm_context *ctx, const unsigned char x[16],
                       unsigned char output[16] )
 {
@@ -151,6 +189,20 @@
     unsigned char lo, hi, rem;
     uint64_t zh, zl;
 
+#if defined(POLARSSL_AESNI_C) && defined(POLARSSL_HAVE_X86_64)
+    if( aesni_supports( POLARSSL_AESNI_CLMUL ) ) {
+        unsigned char h[16];
+
+        PUT_UINT32_BE( ctx->HH[8] >> 32, h,  0 );
+        PUT_UINT32_BE( ctx->HH[8],       h,  4 );
+        PUT_UINT32_BE( ctx->HL[8] >> 32, h,  8 );
+        PUT_UINT32_BE( ctx->HL[8],       h, 12 );
+
+        aesni_gcm_mult( output, x, h );
+        return;
+    }
+#endif
+
     memset( z, 0x00, 16 );
 
     lo = x[15] & 0xf;
diff --git a/library/net.c b/library/net.c
index fd5fd12..67080a2 100644
--- a/library/net.c
+++ b/library/net.c
@@ -40,11 +40,13 @@
 #include <winsock2.h>
 #include <windows.h>
 
+#if defined(_MSC_VER)
 #if defined(_WIN32_WCE)
 #pragma comment( lib, "ws2.lib" )
 #else
 #pragma comment( lib, "ws2_32.lib" )
 #endif
+#endif /* _MSC_VER */
 
 #define read(fd,buf,len)        recv(fd,(char*)buf,(int) len,0)
 #define write(fd,buf,len)       send(fd,(char*)buf,(int) len,0)
diff --git a/library/padlock.c b/library/padlock.c
index a7b4c0c..6e94532 100644
--- a/library/padlock.c
+++ b/library/padlock.c
@@ -43,7 +43,7 @@
 int padlock_supports( int feature )
 {
     static int flags = -1;
-    int ebx, edx;
+    int ebx = 0, edx = 0;
 
     if( flags == -1 )
     {
@@ -76,7 +76,7 @@
                        const unsigned char input[16],
                        unsigned char output[16] )
 {
-    int ebx;
+    int ebx = 0;
     uint32_t *rk;
     uint32_t *blk;
     uint32_t *ctrl;
@@ -117,7 +117,7 @@
                        const unsigned char *input,
                        unsigned char *output )
 {
-    int ebx;
+    int ebx = 0;
     size_t count;
     uint32_t *rk;
     uint32_t *iw;
diff --git a/library/ssl_tls.c b/library/ssl_tls.c
index 75ba907..6ea2821 100644
--- a/library/ssl_tls.c
+++ b/library/ssl_tls.c
@@ -1610,6 +1610,21 @@
             size_t pad_count = 0, real_count = 1;
             size_t padding_idx = ssl->in_msglen - padlen - 1;
 
+            /*
+             * Padding is guaranteed to be incorrect if:
+             *   1. padlen - 1 > ssl->in_msglen
+             *
+             *   2. ssl->in_msglen + padlen >
+             *        SSL_MAX_CONTENT_LEN + 256 (max padding)
+             *
+             * In both cases we reset padding_idx to a safe value (0) to
+             * prevent out-of-buffer reads.
+             */
+            correct &= ( ssl->in_msglen >= padlen - 1 );
+            correct &= ( ssl->in_msglen + padlen <= SSL_MAX_CONTENT_LEN + 256 );
+
+            padding_idx *= correct;
+
             for( i = 1; i <= 256; i++ )
             {
                 real_count &= ( i <= padlen );
diff --git a/programs/Makefile b/programs/Makefile
index 6cee0f1..9238ae8 100644
--- a/programs/Makefile
+++ b/programs/Makefile
@@ -49,6 +49,10 @@
 APPS +=	test/o_p_test
 endif
 
+ifdef PTHREAD
+APPS +=	ssl/ssl_pthread_server
+endif
+
 .SILENT:
 
 all: $(APPS)
@@ -189,6 +193,10 @@
 	echo   "  CC    ssl/ssl_fork_server.c"
 	$(CC) $(CFLAGS) $(OFLAGS) ssl/ssl_fork_server.c   $(LDFLAGS) -o $@
 
+ssl/ssl_pthread_server: ssl/ssl_pthread_server.c ../library/libpolarssl.a
+	echo   "  CC    ssl/ssl_pthread_server.c"
+	$(CC) $(CFLAGS) $(OFLAGS) ssl/ssl_pthread_server.c   $(LDFLAGS) -o $@ -lpthread
+
 ssl/ssl_mail_client: ssl/ssl_mail_client.c ../library/libpolarssl.a
 	echo   "  CC    ssl/ssl_mail_client.c"
 	$(CC) $(CFLAGS) $(OFLAGS) ssl/ssl_mail_client.c   $(LDFLAGS) -o $@
diff --git a/programs/pkey/ecdsa.c b/programs/pkey/ecdsa.c
index 3cba7e8..3dccfeb 100644
--- a/programs/pkey/ecdsa.c
+++ b/programs/pkey/ecdsa.c
@@ -162,7 +162,7 @@
         printf( " failed\n  ! ecdsa_genkey returned %d\n", ret );
         goto exit;
     }
-    printf( " ok (signature length = %zu)\n", sig_len );
+    printf( " ok (signature length = %u)\n", (unsigned int) sig_len );
 
     dump_buf( "  + Hash: ", hash, sizeof hash );
     dump_buf( "  + Signature: ", sig, sig_len );
diff --git a/programs/ssl/CMakeLists.txt b/programs/ssl/CMakeLists.txt
index a117b58..e0cfba0 100644
--- a/programs/ssl/CMakeLists.txt
+++ b/programs/ssl/CMakeLists.txt
@@ -1,7 +1,18 @@
+set(THREADS_USE_PTHREADS_WIN32 true)
+find_package(Threads)
+
 set(libs
      polarssl
 )
 
+set(targets
+     ssl_client1
+     ssl_client2
+     ssl_server
+     ssl_fork_server
+     ssl_mail_client
+)
+
 if(USE_PKCS11_HELPER_LIBRARY)
     set(libs ${libs} pkcs11-helper)
 endif(USE_PKCS11_HELPER_LIBRARY)
@@ -28,6 +39,12 @@
 add_executable(ssl_mail_client ssl_mail_client.c)
 target_link_libraries(ssl_mail_client ${libs})
 
-install(TARGETS ssl_client1 ssl_client2 ssl_server ssl_fork_server ssl_mail_client
+if(THREADS_FOUND)
+    add_executable(ssl_pthread_server ssl_pthread_server.c)
+    target_link_libraries(ssl_pthread_server ${libs} ${CMAKE_THREAD_LIBS_INIT})
+    set(targets ${targets} ssl_pthread_server)
+endif(THREADS_FOUND)
+
+install(TARGETS ${targets}
         DESTINATION "bin"
         PERMISSIONS OWNER_READ OWNER_WRITE OWNER_EXECUTE GROUP_READ GROUP_EXECUTE WORLD_READ WORLD_EXECUTE)
diff --git a/programs/ssl/ssl_mail_client.c b/programs/ssl/ssl_mail_client.c
index a3e9f2b..6abc073 100644
--- a/programs/ssl/ssl_mail_client.c
+++ b/programs/ssl/ssl_mail_client.c
@@ -42,11 +42,13 @@
 #include <winsock2.h>
 #include <windows.h>
 
+#if defined(_MSC_VER)
 #if defined(_WIN32_WCE)
 #pragma comment( lib, "ws2.lib" )
 #else
 #pragma comment( lib, "ws2_32.lib" )
 #endif
+#endif /* _MSC_VER */
 #endif
 
 #include "polarssl/base64.h"
diff --git a/programs/ssl/ssl_pthread_server.c b/programs/ssl/ssl_pthread_server.c
new file mode 100644
index 0000000..7e1b52e
--- /dev/null
+++ b/programs/ssl/ssl_pthread_server.c
@@ -0,0 +1,491 @@
+/*
+ *  SSL server demonstration program using pthread for handling multiple
+ *  clients.
+ *
+ *  Copyright (C) 2006-2013, Brainspark B.V.
+ *
+ *  This file is part of PolarSSL (http://www.polarssl.org)
+ *  Lead Maintainer: Paul Bakker <polarssl_maintainer at polarssl.org>
+ *
+ *  All rights reserved.
+ *
+ *  This program is free software; you can redistribute it and/or modify
+ *  it under the terms of the GNU General Public License as published by
+ *  the Free Software Foundation; either version 2 of the License, or
+ *  (at your option) any later version.
+ *
+ *  This program is distributed in the hope that it will be useful,
+ *  but WITHOUT ANY WARRANTY; without even the implied warranty of
+ *  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ *  GNU General Public License for more details.
+ *
+ *  You should have received a copy of the GNU General Public License along
+ *  with this program; if not, write to the Free Software Foundation, Inc.,
+ *  51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+ */
+
+#include "polarssl/config.h"
+
+#if defined(_WIN32)
+#include <windows.h>
+#endif
+
+#include <string.h>
+#include <stdlib.h>
+#include <stdio.h>
+
+#include "polarssl/entropy.h"
+#include "polarssl/ctr_drbg.h"
+#include "polarssl/certs.h"
+#include "polarssl/x509.h"
+#include "polarssl/ssl.h"
+#include "polarssl/net.h"
+#include "polarssl/error.h"
+
+#if defined(POLARSSL_SSL_CACHE_C)
+#include "polarssl/ssl_cache.h"
+#endif
+
+#if defined(POLARSSL_MEMORY_BUFFER_ALLOC_C)
+#include "polarssl/memory.h"
+#endif
+
+#if !defined(POLARSSL_BIGNUM_C) || !defined(POLARSSL_CERTS_C) ||            \
+    !defined(POLARSSL_ENTROPY_C) || !defined(POLARSSL_SSL_TLS_C) ||         \
+    !defined(POLARSSL_SSL_SRV_C) || !defined(POLARSSL_NET_C) ||             \
+    !defined(POLARSSL_RSA_C) || !defined(POLARSSL_CTR_DRBG_C) ||            \
+    !defined(POLARSSL_X509_CRT_PARSE_C) ||                                  \
+    !defined(POLARSSL_THREADING_C) || !defined(POLARSSL_THREADING_PTHREAD)
+int main( int argc, char *argv[] )
+{
+    ((void) argc);
+    ((void) argv);
+
+    printf("POLARSSL_BIGNUM_C and/or POLARSSL_CERTS_C and/or POLARSSL_ENTROPY_C "
+           "and/or POLARSSL_SSL_TLS_C and/or POLARSSL_SSL_SRV_C and/or "
+           "POLARSSL_NET_C and/or POLARSSL_RSA_C and/or "
+           "POLARSSL_CTR_DRBG_C and/or POLARSSL_X509_CRT_PARSE_C and/or "
+           "POLARSSL_THREADING_C and/or POLARSSL_THREADING_PTHREAD "
+           "not defined.\n");
+    return( 0 );
+}
+#else
+
+#define HTTP_RESPONSE \
+    "HTTP/1.0 200 OK\r\nContent-Type: text/html\r\n\r\n" \
+    "<h2>PolarSSL Test Server</h2>\r\n" \
+    "<p>Successful connection using: %s</p>\r\n"
+
+#define DEBUG_LEVEL 0
+
+threading_mutex_t debug_mutex;
+
+static void my_mutexed_debug( void *ctx, int level, const char *str )
+{
+    polarssl_mutex_lock( &debug_mutex );
+    if( level < DEBUG_LEVEL )
+    {
+        fprintf( (FILE *) ctx, "%s", str );
+        fflush(  (FILE *) ctx  );
+    }
+    polarssl_mutex_unlock( &debug_mutex );
+}
+
+typedef struct {
+    int client_fd;
+    int thread_complete;
+    entropy_context *entropy;
+#if defined(POLARSSL_SSL_CACHE_C)
+    ssl_cache_context *cache;
+#endif
+    x509_crt *ca_chain;
+    x509_crt *server_cert;
+    pk_context *server_key;
+} thread_info_t;
+
+typedef struct {
+    int active;
+    thread_info_t   data;
+    pthread_t       thread;
+} pthread_info_t;
+
+#define MAX_NUM_THREADS 5
+
+static thread_info_t    base_info;
+static pthread_info_t   threads[MAX_NUM_THREADS];
+
+static void *handle_ssl_connection( void *data )
+{
+    int ret, len;
+    thread_info_t *thread_info = (thread_info_t *) data;
+    int client_fd = thread_info->client_fd;
+    int thread_id = (int) pthread_self();
+    unsigned char buf[1024];
+    char pers[50];
+    ssl_context ssl;
+    ctr_drbg_context ctr_drbg;
+
+    snprintf( pers, sizeof(pers), "SSL Pthread Thread %d", thread_id );
+    printf( "  [ #%d ]  Client FD %d\n", thread_id, client_fd );
+    printf( "  [ #%d ]  Seeding the random number generator...\n", thread_id );
+
+    /* entropy_func() is thread-safe if POLARSSL_THREADING_C is set
+     */
+    if( ( ret = ctr_drbg_init( &ctr_drbg, entropy_func, thread_info->entropy,
+                               (const unsigned char *) pers,
+                               strlen( pers ) ) ) != 0 )
+    {
+        printf( "  [ #%d ]  failed: ctr_drbg_init returned -0x%04x\n",
+                thread_id, -ret );
+        goto thread_exit;
+    }
+
+    printf( "  [ #%d ]  ok\n", thread_id );
+
+    /*
+     * 4. Setup stuff
+     */
+    printf( "  [ #%d ]  Setting up the SSL data....\n", thread_id );
+
+    if( ( ret = ssl_init( &ssl ) ) != 0 )
+    {
+        printf( "  [ #%d ]  failed: ssl_init returned -0x%04x\n",
+                thread_id, -ret );
+        goto thread_exit;
+    }
+
+    ssl_set_endpoint( &ssl, SSL_IS_SERVER );
+    ssl_set_authmode( &ssl, SSL_VERIFY_NONE );
+
+    ssl_set_rng( &ssl, ctr_drbg_random, &ctr_drbg );
+    ssl_set_dbg( &ssl, my_mutexed_debug, stdout );
+
+    /* ssl_cache_get() and ssl_cache_set() are thread-safe if
+     * POLARSSL_THREADING_C is set.
+     */
+#if defined(POLARSSL_SSL_CACHE_C)
+    ssl_set_session_cache( &ssl, ssl_cache_get, thread_info->cache,
+                                 ssl_cache_set, thread_info->cache );
+#endif
+
+    ssl_set_ca_chain( &ssl, thread_info->ca_chain, NULL, NULL );
+    ssl_set_own_cert( &ssl, thread_info->server_cert, thread_info->server_key );
+
+    printf( "  [ #%d ]  ok\n", thread_id );
+
+    ssl_set_bio( &ssl, net_recv, &client_fd,
+                       net_send, &client_fd );
+
+    printf( "  [ #%d ]  ok\n", thread_id );
+
+    /*
+     * 5. Handshake
+     */
+    printf( "  [ #%d ]  Performing the SSL/TLS handshake\n", thread_id );
+
+    while( ( ret = ssl_handshake( &ssl ) ) != 0 )
+    {
+        if( ret != POLARSSL_ERR_NET_WANT_READ && ret != POLARSSL_ERR_NET_WANT_WRITE )
+        {
+            printf( "  [ #%d ]  failed: ssl_handshake returned -0x%04x\n",
+                    thread_id, -ret );
+            goto thread_exit;
+        }
+    }
+
+    printf( "  [ #%d ]  ok\n", thread_id );
+
+    /*
+     * 6. Read the HTTP Request
+     */
+    printf( "  [ #%d ]  < Read from client\n", thread_id );
+
+    do
+    {
+        len = sizeof( buf ) - 1;
+        memset( buf, 0, sizeof( buf ) );
+        ret = ssl_read( &ssl, buf, len );
+
+        if( ret == POLARSSL_ERR_NET_WANT_READ || ret == POLARSSL_ERR_NET_WANT_WRITE )
+            continue;
+
+        if( ret <= 0 )
+        {
+            switch( ret )
+            {
+                case POLARSSL_ERR_SSL_PEER_CLOSE_NOTIFY:
+                    printf( "  [ #%d ]  connection was closed gracefully\n",
+                            thread_id );
+                    goto thread_exit;
+
+                case POLARSSL_ERR_NET_CONN_RESET:
+                    printf( "  [ #%d ]  connection was reset by peer\n",
+                            thread_id );
+                    goto thread_exit;
+
+                default:
+                    printf( "  [ #%d ]  ssl_read returned -0x%04x\n",
+                            thread_id, -ret );
+                    goto thread_exit;
+            }
+
+            break;
+        }
+
+        len = ret;
+        printf( "  [ #%d ]  %d bytes read\n=====\n%s\n=====\n",
+                thread_id, len, (char *) buf );
+
+        if( ret > 0 )
+            break;
+    }
+    while( 1 );
+
+    /*
+     * 7. Write the 200 Response
+     */
+    printf( "  [ #%d ]  > Write to client:\n", thread_id );
+
+    len = sprintf( (char *) buf, HTTP_RESPONSE,
+                   ssl_get_ciphersuite( &ssl ) );
+
+    while( ( ret = ssl_write( &ssl, buf, len ) ) <= 0 )
+    {
+        if( ret == POLARSSL_ERR_NET_CONN_RESET )
+        {
+            printf( "  [ #%d ]  failed: peer closed the connection\n",
+                    thread_id );
+            goto thread_exit;
+        }
+
+        if( ret != POLARSSL_ERR_NET_WANT_READ && ret != POLARSSL_ERR_NET_WANT_WRITE )
+        {
+            printf( "  [ #%d ]  failed: ssl_write returned -0x%04x\n",
+                    thread_id, ret );
+            goto thread_exit;
+        }
+    }
+
+    len = ret;
+    printf( "  [ #%d ]  %d bytes written\n=====\n%s\n=====\n",
+            thread_id, len, (char *) buf );
+
+    ret = 0;
+
+thread_exit:
+
+#ifdef POLARSSL_ERROR_C
+    if( ret != 0 )
+    {
+        char error_buf[100];
+        polarssl_strerror( ret, error_buf, 100 );
+        printf("  [ #%d ]  Last error was: -0x%04x - %s\n\n",
+               thread_id, -ret, error_buf );
+    }
+#endif
+
+    net_close( client_fd );
+    ssl_free( &ssl );
+
+    thread_info->thread_complete = 1;
+
+    return( NULL );
+}
+
+static int thread_create( int client_fd )
+{
+    int ret, i;
+
+    /*
+     * Find in-active or finished thread slot
+     */
+    for( i = 0; i < MAX_NUM_THREADS; i++ )
+    {
+        if( threads[i].active == 0 )
+            break;
+
+        if( threads[i].data.thread_complete == 1 )
+        {
+            printf( "  [ main ]  Cleaning up thread %d\n", i );
+            pthread_join(threads[i].thread, NULL );
+            memset( &threads[i], 0, sizeof(pthread_info_t) );
+            break;
+        }
+    }
+
+    if( i == MAX_NUM_THREADS )
+        return( -1 );
+
+    /*
+     * Fill thread-info for thread
+     */
+    memcpy( &threads[i].data, &base_info, sizeof(base_info) );
+    threads[i].active = 1;
+    threads[i].data.client_fd = client_fd;
+
+    if( ( ret = pthread_create( &threads[i].thread, NULL, handle_ssl_connection,                                &threads[i].data ) ) != 0 )
+    {
+        return( ret );
+    }
+
+    return( 0 );
+}
+
+int main( int argc, char *argv[] )
+{
+    int ret;
+    int listen_fd;
+    int client_fd = -1;
+
+    entropy_context entropy;
+    x509_crt srvcert;
+    pk_context pkey;
+#if defined(POLARSSL_MEMORY_BUFFER_ALLOC_C)
+    unsigned char alloc_buf[100000];
+#endif
+#if defined(POLARSSL_SSL_CACHE_C)
+    ssl_cache_context cache;
+#endif
+
+    ((void) argc);
+    ((void) argv);
+
+#if defined(POLARSSL_MEMORY_BUFFER_ALLOC_C)
+    memory_buffer_alloc_init( alloc_buf, sizeof(alloc_buf) );
+#endif
+
+#if defined(POLARSSL_SSL_CACHE_C)
+    ssl_cache_init( &cache );
+    base_info.cache = &cache;
+#endif
+
+    memset( threads, 0, sizeof(threads) );
+
+    polarssl_mutex_init( &debug_mutex );
+
+    /*
+     * We use only a single entropy source that is used in all the threads.
+     */
+    entropy_init( &entropy );
+    base_info.entropy = &entropy;
+
+    /*
+     * 1. Load the certificates and private RSA key
+     */
+    printf( "\n  . Loading the server cert. and key..." );
+    fflush( stdout );
+
+    x509_crt_init( &srvcert );
+
+    /*
+     * This demonstration program uses embedded test certificates.
+     * Instead, you may want to use x509_crt_parse_file() to read the
+     * server and CA certificates, as well as pk_parse_keyfile().
+     */
+    ret = x509_crt_parse( &srvcert, (const unsigned char *) test_srv_crt,
+                          strlen( test_srv_crt ) );
+    if( ret != 0 )
+    {
+        printf( " failed\n  !  x509_crt_parse returned %d\n\n", ret );
+        goto exit;
+    }
+
+    ret = x509_crt_parse( &srvcert, (const unsigned char *) test_ca_list,
+                          strlen( test_ca_list ) );
+    if( ret != 0 )
+    {
+        printf( " failed\n  !  x509_crt_parse returned %d\n\n", ret );
+        goto exit;
+    }
+
+    pk_init( &pkey );
+    ret =  pk_parse_key( &pkey, (const unsigned char *) test_srv_key,
+                         strlen( test_srv_key ), NULL, 0 );
+    if( ret != 0 )
+    {
+        printf( " failed\n  !  pk_parse_key returned %d\n\n", ret );
+        goto exit;
+    }
+
+    base_info.ca_chain = srvcert.next;
+    base_info.server_cert = &srvcert;
+    base_info.server_key = &pkey;
+
+    printf( " ok\n" );
+
+    /*
+     * 2. Setup the listening TCP socket
+     */
+    printf( "  . Bind on https://localhost:4433/ ..." );
+    fflush( stdout );
+
+    if( ( ret = net_bind( &listen_fd, NULL, 4433 ) ) != 0 )
+    {
+        printf( " failed\n  ! net_bind returned %d\n\n", ret );
+        goto exit;
+    }
+
+    printf( " ok\n" );
+
+reset:
+#ifdef POLARSSL_ERROR_C
+    if( ret != 0 )
+    {
+        char error_buf[100];
+        polarssl_strerror( ret, error_buf, 100 );
+        printf( "  [ main ]  Last error was: -0x%04x - %s\n", -ret, error_buf );
+    }
+#endif
+
+    /*
+     * 3. Wait until a client connects
+     */
+    client_fd = -1;
+
+    printf( "  [ main ]  Waiting for a remote connection\n" );
+
+    if( ( ret = net_accept( listen_fd, &client_fd, NULL ) ) != 0 )
+    {
+        printf( "  [ main ] failed: net_accept returned -0x%04x\n", ret );
+        goto exit;
+    }
+
+    printf( "  [ main ]  ok\n" );
+    printf( "  [ main ]  Creating a new thread\n" );
+
+    if( ( ret = thread_create( client_fd ) ) != 0 )
+    {
+        printf( "  [ main ]  failed: thread_create returned %d\n", ret );
+        net_close( client_fd );
+        goto reset;
+    }
+
+    ret = 0;
+    goto reset;
+
+exit:
+    x509_crt_free( &srvcert );
+    pk_free( &pkey );
+#if defined(POLARSSL_SSL_CACHE_C)
+    ssl_cache_free( &cache );
+#endif
+    entropy_free( &entropy );
+
+    polarssl_mutex_free( &debug_mutex );
+
+#if defined(POLARSSL_MEMORY_BUFFER_ALLOC_C)
+    memory_buffer_alloc_free();
+#endif
+
+#if defined(_WIN32)
+    printf( "  Press Enter to exit this program.\n" );
+    fflush( stdout ); getchar();
+#endif
+
+    return( ret );
+}
+
+#endif /* POLARSSL_BIGNUM_C && POLARSSL_CERTS_C && POLARSSL_ENTROPY_C &&
+          POLARSSL_SSL_TLS_C && POLARSSL_SSL_SRV_C && POLARSSL_NET_C &&
+          POLARSSL_RSA_C && POLARSSL_CTR_DRBG_C && POLARSSL_THREADING_C &&
+          POLARSSL_THREADING_PTHREAD */
diff --git a/tests/suites/test_suite_x509write.function b/tests/suites/test_suite_x509write.function
index 8b57f8c..ca097e7 100644
--- a/tests/suites/test_suite_x509write.function
+++ b/tests/suites/test_suite_x509write.function
@@ -21,7 +21,7 @@
     int ret;
     size_t olen = 0, pem_len = 0;
     FILE *f;
-    char *subject_name = "C=NL,O=PolarSSL,CN=PolarSSL Server 1";
+    const char *subject_name = "C=NL,O=PolarSSL,CN=PolarSSL Server 1";
     rnd_pseudo_info rnd_info;
 
     memset( &rnd_info, 0x2a, sizeof( rnd_pseudo_info ) );
diff --git a/visualc/VS2010/PolarSSL.vcxproj b/visualc/VS2010/PolarSSL.vcxproj
index 33f0031..4c067be 100644
--- a/visualc/VS2010/PolarSSL.vcxproj
+++ b/visualc/VS2010/PolarSSL.vcxproj
@@ -143,6 +143,7 @@
   </ItemDefinitionGroup>

   <ItemGroup>

     <ClInclude Include="..\..\include\polarssl\aes.h" />

+    <ClInclude Include="..\..\include\polarssl\aesni.h" />
     <ClInclude Include="..\..\include\polarssl\arc4.h" />

     <ClInclude Include="..\..\include\polarssl\asn1.h" />

     <ClInclude Include="..\..\include\polarssl\asn1write.h" />

@@ -201,6 +202,7 @@
   </ItemGroup>

   <ItemGroup>

     <ClCompile Include="..\..\library\aes.c" />

+    <ClCompile Include="..\..\library\aesni.c" />
     <ClCompile Include="..\..\library\arc4.c" />

     <ClCompile Include="..\..\library\asn1parse.c" />

     <ClCompile Include="..\..\library\asn1write.c" />

diff --git a/visualc/VS6/polarssl.dsp b/visualc/VS6/polarssl.dsp
index c40b75b..b6a82c0 100644
--- a/visualc/VS6/polarssl.dsp
+++ b/visualc/VS6/polarssl.dsp
@@ -89,6 +89,10 @@
 # End Source File

 # Begin Source File

 

+SOURCE=..\..\library\aesni.c
+# End Source File
+# Begin Source File
+
 SOURCE=..\..\library\arc4.c

 # End Source File

 # Begin Source File

@@ -349,6 +353,10 @@
 # End Source File

 # Begin Source File

 

+SOURCE=..\..\include\polarssl\aesni.h
+# End Source File
+# Begin Source File
+
 SOURCE=..\..\include\polarssl\arc4.h

 # End Source File

 # Begin Source File