diff --git a/ChangeLog b/ChangeLog
index cd3732f..51fdcc3 100644
--- a/ChangeLog
+++ b/ChangeLog
@@ -8,6 +8,7 @@
    * 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
 
 Changes
    * gen_prime() speedup
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 d98bdb3..283e294 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
@@ -1902,6 +1918,10 @@
 /*
  * 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
@@ -2012,6 +2032,10 @@
 #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/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..0286a7f 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 +628,7 @@
     *RK++ = *SK++;
     *RK++ = *SK++;
 
+done:
     memset( &cty, 0, sizeof( aes_context ) );
 
     return( 0 );
@@ -673,6 +691,11 @@
     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..b5858dc
--- /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( 0 );
+}
+
+/*
+ * 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
+ */
+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
+ */
+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
+ */
+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/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/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