chadmed/gentoo-asahi-releng
1
0
Fork 0
Code Issues Pull requests Actions Packages Projects Releases Wiki Activity
gentoo-asahi-releng/tools-hardened/desktop/configs/loop-AES-kernel-3.14.patch

9129 lines
286 KiB
Diff
Raw Normal View History

tools-hardened/desktop/: Updates kernel to 3.14.2-r1. This update includes changes to the README which instruct on how to apply the loop-AES patch along with updates to the kernel requirements. It also updates the kernel config to a 3.14.2-r1 compatible .config file, as well as updating the kernel-config variable to point towards the 3.14.2-r1 config file.
2014-05-06 19:22:32 +00:00
Before this patch can be applied to kernel, drivers/block/loop.c and
include/linux/loop.h source files must be removed:
rm -f drivers/block/loop.c include/linux/loop.h
diff -urN linux-3.14-noloop/drivers/block/Kconfig linux-3.14-AES/drivers/block/Kconfig
--- linux-3.14-noloop/drivers/block/Kconfig 2014-03-31 06:40:15.000000000 +0300
+++ linux-3.14-AES/drivers/block/Kconfig 2014-03-31 22:20:36.000000000 +0300
@@ -235,14 +235,6 @@
bits of, say, a sound file). This is also safe if the file resides
on a remote file server.
- There are several ways of encrypting disks. Some of these require
- kernel patches. The vanilla kernel offers the cryptoloop option
- and a Device Mapper target (which is superior, as it supports all
- file systems). If you want to use the cryptoloop, say Y to both
- LOOP and CRYPTOLOOP, and make sure you have a recent (version 2.12
- or later) version of util-linux. Additionally, be aware that
- the cryptoloop is not safe for storing journaled filesystems.
-
Note that this loop device has nothing to do with the loopback
device used for network connections from the machine to itself.
@@ -251,35 +243,40 @@
Most users will answer N here.
-config BLK_DEV_LOOP_MIN_COUNT
- int "Number of loop devices to pre-create at init time"
+config BLK_DEV_LOOP_AES
+ bool "AES encrypted loop device support"
depends on BLK_DEV_LOOP
- default 8
- help
- Static number of loop devices to be unconditionally pre-created
- at init time.
-
- This default value can be overwritten on the kernel command
- line or with module-parameter loop.max_loop.
-
- The historic default is 8. If a late 2011 version of losetup(8)
- is used, it can be set to 0, since needed loop devices can be
- dynamically allocated with the /dev/loop-control interface.
-
-config BLK_DEV_CRYPTOLOOP
- tristate "Cryptoloop Support"
- select CRYPTO
- select CRYPTO_CBC
+ ---help---
+ If you want to use AES encryption algorithm to encrypt loop
+ devices, say Y here. If you don't know what to do here, say N.
+
+config BLK_DEV_LOOP_KEYSCRUB
+ bool "loop encryption key scrubbing support"
depends on BLK_DEV_LOOP
---help---
- Say Y here if you want to be able to use the ciphers that are
- provided by the CryptoAPI as loop transformation. This might be
- used as hard disk encryption.
-
- WARNING: This device is not safe for journaled file systems like
- ext3 or Reiserfs. Please use the Device Mapper crypto module
- instead, which can be configured to be on-disk compatible with the
- cryptoloop device.
+ Loop encryption key scrubbing moves and inverts key bits in
+ kernel RAM so that the thin oxide which forms the storage
+ capacitor dielectric of DRAM cells is not permitted to develop
+ detectable property. For more info, see Peter Gutmann's paper:
+ http://www.cypherpunks.to/~peter/usenix01.pdf
+
+ Paranoid tinfoil hat crowd say Y here, everyone else say N.
+
+config BLK_DEV_LOOP_PADLOCK
+ bool "VIA padlock hardware AES support"
+ depends on BLK_DEV_LOOP && BLK_DEV_LOOP_AES && (X86 || X86_64)
+ ---help---
+ If you have VIA processor that supports padlock xcrypt instructions,
+ say Y here. If enabled, presence of VIA padlock instructions is detected
+ at run time, but code still works on non-padlock processors too.
+
+config BLK_DEV_LOOP_INTELAES
+ bool "Intel hardware AES support"
+ depends on BLK_DEV_LOOP && BLK_DEV_LOOP_AES && (X86 || X86_64)
+ ---help---
+ If you have a processor that supports Intel AES instructions,
+ say Y here. If enabled, presence of Intel AES instructions is detected
+ at run time, but code still works on older processors too.
source "drivers/block/drbd/Kconfig"
diff -urN linux-3.14-noloop/drivers/block/loop.c linux-3.14-AES/drivers/block/loop.c
--- linux-3.14-noloop/drivers/block/loop.c 1970-01-01 02:00:00.000000000 +0200
+++ linux-3.14-AES/drivers/block/loop.c 2014-03-31 22:20:36.000000000 +0300
@@ -0,0 +1,3278 @@
+/*
+ * linux/drivers/block/loop.c
+ *
+ * Written by Theodore Ts'o, 3/29/93
+ *
+ * Copyright 1993 by Theodore Ts'o. Redistribution of this file is
+ * permitted under the GNU General Public License.
+ *
+ * DES encryption plus some minor changes by Werner Almesberger, 30-MAY-1993
+ * more DES encryption plus IDEA encryption by Nicholas J. Leon, June 20, 1996
+ *
+ * Modularized and updated for 1.1.16 kernel - Mitch Dsouza 28th May 1994
+ * Adapted for 1.3.59 kernel - Andries Brouwer, 1 Feb 1996
+ *
+ * Fixed do_loop_request() re-entrancy - Vincent.Renardias@waw.com Mar 20, 1997
+ *
+ * Added devfs support - Richard Gooch <rgooch@atnf.csiro.au> 16-Jan-1998
+ *
+ * Handle sparse backing files correctly - Kenn Humborg, Jun 28, 1998
+ *
+ * Loadable modules and other fixes by AK, 1998
+ *
+ * Make real block number available to downstream transfer functions, enables
+ * CBC (and relatives) mode encryption requiring unique IVs per data block.
+ * Reed H. Petty, rhp@draper.net
+ *
+ * Maximum number of loop devices now dynamic via max_loop module parameter.
+ * Russell Kroll <rkroll@exploits.org> 19990701
+ *
+ * Maximum number of loop devices when compiled-in now selectable by passing
+ * max_loop=<1-255> to the kernel on boot.
+ * Erik I. Bols<6C>, <eriki@himolde.no>, Oct 31, 1999
+ *
+ * Completely rewrite request handling to be make_request_fn style and
+ * non blocking, pushing work to a helper thread. Lots of fixes from
+ * Al Viro too.
+ * Jens Axboe <axboe@suse.de>, Nov 2000
+ *
+ * Support up to 256 loop devices
+ * Heinz Mauelshagen <mge@sistina.com>, Feb 2002
+ *
+ * AES transfer added. IV is now passed as (512 byte) sector number.
+ * Jari Ruusu, May 18 2001
+ *
+ * External encryption module locking bug fixed.
+ * Ingo Rohloff <rohloff@in.tum.de>, June 21 2001
+ *
+ * Make device backed loop work with swap (pre-allocated buffers + queue rewrite).
+ * Jari Ruusu, September 2 2001
+ *
+ * Ported 'pre-allocated buffers + queue rewrite' to BIO for 2.5 kernels
+ * Ben Slusky <sluskyb@stwing.org>, March 1 2002
+ * Jari Ruusu, March 27 2002
+ *
+ * File backed code now uses file->f_op->read/write. Based on Andrew Morton's idea.
+ * Jari Ruusu, May 23 2002
+ *
+ * Exported hard sector size correctly, fixed file-backed-loop-on-tmpfs bug,
+ * plus many more enhancements and optimizations.
+ * Adam J. Richter <adam@yggdrasil.com>, Aug 2002
+ *
+ * Added support for removing offset from IV computations.
+ * Jari Ruusu, September 21 2003
+ *
+ * Added support for MD5 IV computation and multi-key operation.
+ * Jari Ruusu, October 8 2003
+ *
+ *
+ * Still To Fix:
+ * - Advisory locking is ignored here.
+ * - Should use an own CAP_* category instead of CAP_SYS_ADMIN
+ */
+
+#include <linux/version.h>
+#include <linux/module.h>
+#include <linux/moduleparam.h>
+#include <linux/sched.h>
+#include <linux/fs.h>
+#include <linux/file.h>
+#include <linux/bio.h>
+#include <linux/stat.h>
+#include <linux/errno.h>
+#include <linux/major.h>
+#include <linux/wait.h>
+#include <linux/blkdev.h>
+#include <linux/blkpg.h>
+#include <linux/init.h>
+#ifdef CONFIG_DEVFS_FS
+# include <linux/devfs_fs_kernel.h>
+#endif
+#include <linux/swap.h>
+#include <linux/slab.h>
+#include <linux/loop.h>
+#include <linux/suspend.h>
+#include <linux/writeback.h>
+#include <linux/buffer_head.h> /* for invalidate_bdev() */
+#include <linux/completion.h>
+#include <linux/kthread.h>
+#if defined(CONFIG_COMPAT) && defined(HAVE_COMPAT_IOCTL)
+# include <linux/compat.h>
+#endif
+#include <linux/mqueue.h>
+#include <linux/delay.h>
+#include <linux/spinlock.h>
+#include <linux/backing-dev.h>
+
+#include <asm/uaccess.h>
+#include <asm/byteorder.h>
+#if (defined(CONFIG_BLK_DEV_LOOP_PADLOCK) || defined(CONFIG_BLK_DEV_LOOP_INTELAES)) && (defined(CONFIG_X86) || defined(CONFIG_X86_64))
+# include <asm/processor.h>
+#endif
+#if defined(CONFIG_BLK_DEV_LOOP_INTELAES) && (defined(CONFIG_X86) || defined(CONFIG_X86_64))
+# include <asm/i387.h>
+#endif
+
+#if defined(CONFIG_X86) && !defined(CONFIG_X86_64)
+# define X86_ASM 1
+#endif
+#if defined(CONFIG_X86_64)
+# define AMD64_ASM 1
+#endif
+
+#include "../misc/aes.h"
+#include "../misc/md5.h"
+
+#if defined(CONFIG_COMPAT) && !defined(HAVE_COMPAT_IOCTL)
+# include <linux/ioctl32.h>
+# define IOCTL32_COMPATIBLE_PTR ((void*)0)
+#endif
+
+//#define LOOP_HAVE_CONGESTED_FN 1
+
+#define L_BIO_RW_AHEAD (REQ_RAHEAD)
+#define L_BIO_RW_NOIDLE (REQ_NOIDLE)
+#define L_BIO_RW_SYNCIO (REQ_SYNC)
+
+static int max_loop = 8;
+
+#ifdef MODULE
+module_param(max_loop, int, 0);
+MODULE_PARM_DESC(max_loop, "Maximum number of loop devices (1-256)");
+#else
+static int __init max_loop_setup(char *str)
+{
+ int y;
+
+ if (get_option(&str, &y) == 1)
+ max_loop = y;
+ return 1;
+}
+__setup("max_loop=", max_loop_setup);
+#endif
+
+static struct gendisk **disks;
+
+/*
+ * Transfer functions
+ */
+static int transfer_none(struct loop_device *lo, int cmd, char *raw_buf,
+ char *loop_buf, int size, sector_t real_block)
+{
+ /* this code is only called from file backed loop */
+ /* and that code expects this function to be no-op */
+
+ cond_resched();
+ return 0;
+}
+
+static int transfer_xor(struct loop_device *lo, int cmd, char *raw_buf,
+ char *loop_buf, int size, sector_t real_block)
+{
+ char *in, *out, *key;
+ int i, keysize;
+
+ if (cmd == READ) {
+ in = raw_buf;
+ out = loop_buf;
+ } else {
+ in = loop_buf;
+ out = raw_buf;
+ }
+
+ key = lo->lo_encrypt_key;
+ keysize = lo->lo_encrypt_key_size;
+ for (i = 0; i < size; i++)
+ *out++ = *in++ ^ key[(i & 511) % keysize];
+ cond_resched();
+ return 0;
+}
+
+static int xor_init(struct loop_device *lo, struct loop_info64 *info)
+{
+ if (info->lo_encrypt_key_size <= 0)
+ return -EINVAL;
+ return 0;
+}
+
+static struct loop_func_table none_funcs = {
+ .number = LO_CRYPT_NONE,
+ .transfer = transfer_none,
+};
+
+static struct loop_func_table xor_funcs = {
+ .number = LO_CRYPT_XOR,
+ .transfer = transfer_xor,
+ .init = xor_init,
+};
+
+#ifdef CONFIG_BLK_DEV_LOOP_AES
+#ifdef CONFIG_BLK_DEV_LOOP_KEYSCRUB
+# define KEY_ALLOC_COUNT 128
+#else
+# define KEY_ALLOC_COUNT 64
+#endif
+
+typedef struct {
+ aes_context *keyPtr[KEY_ALLOC_COUNT];
+ unsigned keyMask;
+#ifdef CONFIG_BLK_DEV_LOOP_KEYSCRUB
+ u_int32_t *partialMD5;
+ u_int32_t partialMD5buf[8];
+ rwlock_t rwlock;
+ unsigned reversed;
+ unsigned blocked;
+ struct timer_list timer;
+#else
+ u_int32_t partialMD5[4];
+#endif
+#if defined(CONFIG_BLK_DEV_LOOP_PADLOCK) && (defined(CONFIG_X86) || defined(CONFIG_X86_64))
+ u_int32_t padlock_cw_e;
+ u_int32_t padlock_cw_d;
+#endif
+} AESmultiKey;
+
+#if (defined(CONFIG_BLK_DEV_LOOP_PADLOCK) || defined(CONFIG_BLK_DEV_LOOP_INTELAES)) && (defined(CONFIG_X86) || defined(CONFIG_X86_64))
+/* This function allocates AES context structures at special address such */
+/* that returned address % 16 == 8 . That way expanded encryption and */
+/* decryption keys in AES context structure are always 16 byte aligned */
+static void *specialAligned_kmalloc(size_t size, unsigned int flags)
+{
+ void *pn, **ps;
+ pn = kmalloc(size + (16 + 8), flags);
+ if(!pn) return (void *)0;
+ ps = (void **)((((unsigned long)pn + 15) & ~((unsigned long)15)) + 8);
+ *(ps - 1) = pn;
+ return (void *)ps;
+}
+static void specialAligned_kfree(void *ps)
+{
+ if(ps) kfree(*((void **)ps - 1));
+}
+# define specialAligned_ctxSize ((sizeof(aes_context) + 15) & ~15)
+#else
+# define specialAligned_kmalloc kmalloc
+# define specialAligned_kfree kfree
+# define specialAligned_ctxSize sizeof(aes_context)
+#endif
+
+#ifdef CONFIG_BLK_DEV_LOOP_KEYSCRUB
+static void keyScrubWork(AESmultiKey *m)
+{
+ aes_context *a0, *a1;
+ u_int32_t *p;
+ int x, y, z;
+
+ z = m->keyMask + 1;
+ for(x = 0; x < z; x++) {
+ a0 = m->keyPtr[x];
+ a1 = m->keyPtr[x + z];
+ memcpy(a1, a0, sizeof(aes_context));
+ m->keyPtr[x] = a1;
+ m->keyPtr[x + z] = a0;
+ p = (u_int32_t *) a0;
+ y = sizeof(aes_context) / sizeof(u_int32_t);
+ while(y > 0) {
+ *p ^= 0xFFFFFFFF;
+ p++;
+ y--;
+ }
+ }
+
+ x = m->reversed; /* x is 0 or 4 */
+ m->reversed ^= 4;
+ y = m->reversed; /* y is 4 or 0 */
+ p = &m->partialMD5buf[x];
+ memcpy(&m->partialMD5buf[y], p, 16);
+ m->partialMD5 = &m->partialMD5buf[y];
+ p[0] ^= 0xFFFFFFFF;
+ p[1] ^= 0xFFFFFFFF;
+ p[2] ^= 0xFFFFFFFF;
+ p[3] ^= 0xFFFFFFFF;
+
+ /* try to flush dirty cache data to RAM */
+#if !defined(CONFIG_XEN) && (defined(CONFIG_X86_64) || (defined(CONFIG_X86) && !defined(CONFIG_M386) && !defined(CONFIG_CPU_386)))
+ __asm__ __volatile__ ("wbinvd": : :"memory");
+#else
+ mb();
+#endif
+}
+
+/* called only from loop thread process context */
+static void keyScrubThreadFn(AESmultiKey *m)
+{
+ write_lock(&m->rwlock);
+ if(!m->blocked) keyScrubWork(m);
+ write_unlock(&m->rwlock);
+}
+
+#if defined(NEW_TIMER_VOID_PTR_PARAM)
+# define KeyScrubTimerFnParamType void *
+#else
+# define KeyScrubTimerFnParamType unsigned long
+#endif
+
+static void keyScrubTimerFn(KeyScrubTimerFnParamType);
+
+static void keyScrubTimerInit(struct loop_device *lo)
+{
+ AESmultiKey *m;
+ unsigned long expire;
+
+ m = (AESmultiKey *)lo->key_data;
+ expire = jiffies + HZ;
+ init_timer(&m->timer);
+ m->timer.expires = expire;
+ m->timer.data = (KeyScrubTimerFnParamType)lo;
+ m->timer.function = keyScrubTimerFn;
+ add_timer(&m->timer);
+}
+
+/* called only from timer handler context */
+static void keyScrubTimerFn(KeyScrubTimerFnParamType d)
+{
+ struct loop_device *lo = (struct loop_device *)d;
+ extern void loop_add_keyscrub_fn(struct loop_device *, void (*)(void *), void *);
+
+ /* rw lock needs process context, so make loop thread do scrubbing */
+ loop_add_keyscrub_fn(lo, (void (*)(void*))keyScrubThreadFn, lo->key_data);
+ /* start timer again */
+ keyScrubTimerInit(lo);
+}
+#endif
+
+static AESmultiKey *allocMultiKey(void)
+{
+ AESmultiKey *m;
+ aes_context *a;
+ int x = 0, n;
+
+ m = (AESmultiKey *) kmalloc(sizeof(AESmultiKey), GFP_KERNEL);
+ if(!m) return 0;
+ memset(m, 0, sizeof(AESmultiKey));
+#ifdef CONFIG_BLK_DEV_LOOP_KEYSCRUB
+ m->partialMD5 = &m->partialMD5buf[0];
+ rwlock_init(&m->rwlock);
+ init_timer(&m->timer);
+ again:
+#endif
+
+ n = PAGE_SIZE / specialAligned_ctxSize;
+ if(!n) n = 1;
+
+ a = (aes_context *) specialAligned_kmalloc(specialAligned_ctxSize * n, GFP_KERNEL);
+ if(!a) {
+#ifdef CONFIG_BLK_DEV_LOOP_KEYSCRUB
+ if(x) specialAligned_kfree(m->keyPtr[0]);
+#endif
+ kfree(m);
+ return 0;
+ }
+
+ while((x < KEY_ALLOC_COUNT) && n) {
+ m->keyPtr[x] = a;
+ a = (aes_context *)((unsigned char *)a + specialAligned_ctxSize);
+ x++;
+ n--;
+ }
+#ifdef CONFIG_BLK_DEV_LOOP_KEYSCRUB
+ if(x < 2) goto again;
+#endif
+ return m;
+}
+
+static void clearAndFreeMultiKey(AESmultiKey *m)
+{
+ aes_context *a;
+ int x, n;
+
+#ifdef CONFIG_BLK_DEV_LOOP_KEYSCRUB
+ /* stop scrub timer. loop thread was killed earlier */
+ del_timer_sync(&m->timer);
+ /* make sure allocated keys are in original order */
+ if(m->reversed) keyScrubWork(m);
+#endif
+ n = PAGE_SIZE / specialAligned_ctxSize;
+ if(!n) n = 1;
+
+ x = 0;
+ while(x < KEY_ALLOC_COUNT) {
+ a = m->keyPtr[x];
+ if(!a) break;
+ memset(a, 0, specialAligned_ctxSize * n);
+ specialAligned_kfree(a);
+ x += n;
+ }
+
+ memset(m, 0, sizeof(AESmultiKey));
+ kfree(m);
+}
+
+static int multiKeySetup(struct loop_device *lo, unsigned char *k, int version3)
+{
+ AESmultiKey *m;
+ aes_context *a;
+ int x, y, n, err = 0;
+ union {
+ u_int32_t w[16];
+ unsigned char b[64];
+ } un;
+
+#if LINUX_VERSION_CODE >= 0x30600
+ if(!uid_eq(lo->lo_key_owner, current_uid()) && !capable(CAP_SYS_ADMIN))
+ return -EPERM;
+#elif LINUX_VERSION_CODE >= 0x2061c
+ if(lo->lo_key_owner != current_uid() && !capable(CAP_SYS_ADMIN))
+ return -EPERM;
+#else
+ if(lo->lo_key_owner != current->uid && !capable(CAP_SYS_ADMIN))
+ return -EPERM;
+#endif
+
+ m = (AESmultiKey *)lo->key_data;
+ if(!m) return -ENXIO;
+
+#ifdef CONFIG_BLK_DEV_LOOP_KEYSCRUB
+ /* temporarily prevent loop thread from messing with keys */
+ write_lock(&m->rwlock);
+ m->blocked = 1;
+ /* make sure allocated keys are in original order */
+ if(m->reversed) keyScrubWork(m);
+ write_unlock(&m->rwlock);
+#endif
+ n = PAGE_SIZE / specialAligned_ctxSize;
+ if(!n) n = 1;
+
+ x = 0;
+ while(x < KEY_ALLOC_COUNT) {
+ if(!m->keyPtr[x]) {
+ a = (aes_context *) specialAligned_kmalloc(specialAligned_ctxSize * n, GFP_KERNEL);
+ if(!a) {
+ err = -ENOMEM;
+ goto error_out;
+ }
+ y = x;
+ while((y < (x + n)) && (y < KEY_ALLOC_COUNT)) {
+ m->keyPtr[y] = a;
+ a = (aes_context *)((unsigned char *)a + specialAligned_ctxSize);
+ y++;
+ }
+ }
+#ifdef CONFIG_BLK_DEV_LOOP_KEYSCRUB
+ if(x >= 64) {
+ x++;
+ continue;
+ }
+#endif
+ if(copy_from_user(&un.b[0], k, 32)) {
+ err = -EFAULT;
+ goto error_out;
+ }
+ aes_set_key(m->keyPtr[x], &un.b[0], lo->lo_encrypt_key_size, 0);
+ k += 32;
+ x++;
+ }
+
+ m->partialMD5[0] = 0x67452301;
+ m->partialMD5[1] = 0xefcdab89;
+ m->partialMD5[2] = 0x98badcfe;
+ m->partialMD5[3] = 0x10325476;
+ if(version3) {
+ /* only first 128 bits of iv-key is used */
+ if(copy_from_user(&un.b[0], k, 16)) {
+ err = -EFAULT;
+ goto error_out;
+ }
+#if defined(__BIG_ENDIAN)
+ un.w[0] = cpu_to_le32(un.w[0]);
+ un.w[1] = cpu_to_le32(un.w[1]);
+ un.w[2] = cpu_to_le32(un.w[2]);
+ un.w[3] = cpu_to_le32(un.w[3]);
+#endif
+ memset(&un.b[16], 0, 48);
+ md5_transform_CPUbyteorder(&m->partialMD5[0], &un.w[0]);
+ lo->lo_flags |= 0x080000; /* multi-key-v3 (info exported to user space) */
+ }
+
+ m->keyMask = 0x3F; /* range 0...63 */
+ lo->lo_flags |= 0x100000; /* multi-key (info exported to user space) */
+ memset(&un.b[0], 0, 32);
+error_out:
+#ifdef CONFIG_BLK_DEV_LOOP_KEYSCRUB
+ /* re-enable loop thread key scrubbing */
+ write_lock(&m->rwlock);
+ m->blocked = 0;
+ write_unlock(&m->rwlock);
+#endif
+ return err;
+}
+
+static int keySetup_aes(struct loop_device *lo, struct loop_info64 *info)
+{
+ AESmultiKey *m;
+ union {
+ u_int32_t w[8]; /* needed for 4 byte alignment for b[] */
+ unsigned char b[32];
+ } un;
+
+ lo->key_data = m = allocMultiKey();
+ if(!m) return(-ENOMEM);
+ memcpy(&un.b[0], &info->lo_encrypt_key[0], 32);
+ aes_set_key(m->keyPtr[0], &un.b[0], info->lo_encrypt_key_size, 0);
+ memset(&info->lo_encrypt_key[0], 0, sizeof(info->lo_encrypt_key));
+ memset(&un.b[0], 0, 32);
+#if defined(CONFIG_BLK_DEV_LOOP_PADLOCK) && (defined(CONFIG_X86) || defined(CONFIG_X86_64))
+ switch(info->lo_encrypt_key_size) {
+ case 256: /* bits */
+ case 32: /* bytes */
+ /* 14 rounds, AES, software key gen, normal oper, encrypt, 256-bit key */
+ m->padlock_cw_e = 14 | (1<<7) | (2<<10);
+ /* 14 rounds, AES, software key gen, normal oper, decrypt, 256-bit key */
+ m->padlock_cw_d = 14 | (1<<7) | (1<<9) | (2<<10);
+ break;
+ case 192: /* bits */
+ case 24: /* bytes */
+ /* 12 rounds, AES, software key gen, normal oper, encrypt, 192-bit key */
+ m->padlock_cw_e = 12 | (1<<7) | (1<<10);
+ /* 12 rounds, AES, software key gen, normal oper, decrypt, 192-bit key */
+ m->padlock_cw_d = 12 | (1<<7) | (1<<9) | (1<<10);
+ break;
+ default:
+ /* 10 rounds, AES, software key gen, normal oper, encrypt, 128-bit key */
+ m->padlock_cw_e = 10 | (1<<7);
+ /* 10 rounds, AES, software key gen, normal oper, decrypt, 128-bit key */
+ m->padlock_cw_d = 10 | (1<<7) | (1<<9);
+ break;
+ }
+#endif
+#ifdef CONFIG_BLK_DEV_LOOP_KEYSCRUB
+ keyScrubTimerInit(lo);
+#endif
+ return(0);
+}
+
+static int keyClean_aes(struct loop_device *lo)
+{
+ if(lo->key_data) {
+ clearAndFreeMultiKey((AESmultiKey *)lo->key_data);
+ lo->key_data = 0;
+ }
+ return(0);
+}
+
+static int handleIoctl_aes(struct loop_device *lo, int cmd, unsigned long arg)
+{
+ int err;
+
+ switch (cmd) {
+ case LOOP_MULTI_KEY_SETUP:
+ err = multiKeySetup(lo, (unsigned char *)arg, 0);
+ break;
+ case LOOP_MULTI_KEY_SETUP_V3:
+ err = multiKeySetup(lo, (unsigned char *)arg, 1);
+ break;
+ default:
+ err = -EINVAL;
+ }
+ return err;
+}
+
+void loop_compute_sector_iv(sector_t devSect, u_int32_t *ivout)
+{
+ if(sizeof(sector_t) == 8) {
+ ivout[0] = cpu_to_le32(devSect);
+ ivout[1] = cpu_to_le32((u_int64_t)devSect>>32);
+ ivout[3] = ivout[2] = 0;
+ } else {
+ ivout[0] = cpu_to_le32(devSect);
+ ivout[3] = ivout[2] = ivout[1] = 0;
+ }
+}
+
+void loop_compute_md5_iv_v3(sector_t devSect, u_int32_t *ivout, u_int32_t *data)
+{
+ int x;
+#if defined(__BIG_ENDIAN)
+ int y, e;
+#endif
+ u_int32_t buf[16];
+
+#if defined(__BIG_ENDIAN)
+ y = 7;
+ e = 16;
+ do {
+ if (!y) {
+ e = 12;
+ /* md5_transform_CPUbyteorder wants data in CPU byte order */
+ /* devSect is already in CPU byte order -- no need to convert */
+ if(sizeof(sector_t) == 8) {
+ /* use only 56 bits of sector number */
+ buf[12] = devSect;
+ buf[13] = (((u_int64_t)devSect >> 32) & 0xFFFFFF) | 0x80000000;
+ } else {
+ /* 32 bits of sector number + 24 zero bits */
+ buf[12] = devSect;
+ buf[13] = 0x80000000;
+ }
+ /* 4024 bits == 31 * 128 bit plaintext blocks + 56 bits of sector number */
+ /* For version 3 on-disk format this really should be 4536 bits, but can't be */
+ /* changed without breaking compatibility. V3 uses MD5-with-wrong-length IV */
+ buf[14] = 4024;
+ buf[15] = 0;
+ }
+ x = 0;
+ do {
+ buf[x ] = cpu_to_le32(data[0]);
+ buf[x + 1] = cpu_to_le32(data[1]);
+ buf[x + 2] = cpu_to_le32(data[2]);
+ buf[x + 3] = cpu_to_le32(data[3]);
+ x += 4;
+ data += 4;
+ } while (x < e);
+ md5_transform_CPUbyteorder(&ivout[0], &buf[0]);
+ } while (--y >= 0);
+ ivout[0] = cpu_to_le32(ivout[0]);
+ ivout[1] = cpu_to_le32(ivout[1]);
+ ivout[2] = cpu_to_le32(ivout[2]);
+ ivout[3] = cpu_to_le32(ivout[3]);
+#else
+ x = 6;
+ do {
+ md5_transform_CPUbyteorder(&ivout[0], data);
+ data += 16;
+ } while (--x >= 0);
+ memcpy(buf, data, 48);
+ /* md5_transform_CPUbyteorder wants data in CPU byte order */
+ /* devSect is already in CPU byte order -- no need to convert */
+ if(sizeof(sector_t) == 8) {
+ /* use only 56 bits of sector number */
+ buf[12] = devSect;
+ buf[13] = (((u_int64_t)devSect >> 32) & 0xFFFFFF) | 0x80000000;
+ } else {
+ /* 32 bits of sector number + 24 zero bits */
+ buf[12] = devSect;
+ buf[13] = 0x80000000;
+ }
+ /* 4024 bits == 31 * 128 bit plaintext blocks + 56 bits of sector number */
+ /* For version 3 on-disk format this really should be 4536 bits, but can't be */
+ /* changed without breaking compatibility. V3 uses MD5-with-wrong-length IV */
+ buf[14] = 4024;
+ buf[15] = 0;
+ md5_transform_CPUbyteorder(&ivout[0], &buf[0]);
+#endif
+}
+
+/* this function exists for compatibility with old external cipher modules */
+void loop_compute_md5_iv(sector_t devSect, u_int32_t *ivout, u_int32_t *data)
+{
+ ivout[0] = 0x67452301;
+ ivout[1] = 0xefcdab89;
+ ivout[2] = 0x98badcfe;
+ ivout[3] = 0x10325476;
+ loop_compute_md5_iv_v3(devSect, ivout, data);
+}
+
+/* Some external modules do not know if md5_transform_CPUbyteorder() */
+/* is asmlinkage or not, so here is C language wrapper for them. */
+void md5_transform_CPUbyteorder_C(u_int32_t *hash, u_int32_t const *in)
+{
+ md5_transform_CPUbyteorder(hash, in);
+}
+
+#if defined(CONFIG_X86_64) && defined(AMD64_ASM)
+# define HAVE_MD5_2X_IMPLEMENTATION 1
+#endif
+#if defined(HAVE_MD5_2X_IMPLEMENTATION)
+/*
+ * This 2x code is currently only available on little endian AMD64
+ * This 2x code assumes little endian byte order
+ * Context A input data is at zero offset, context B at data + 512 bytes
+ * Context A ivout at zero offset, context B at ivout + 16 bytes
+ */
+void loop_compute_md5_iv_v3_2x(sector_t devSect, u_int32_t *ivout, u_int32_t *data)
+{
+ int x;
+ u_int32_t buf[2*16];
+
+ x = 6;
+ do {
+ md5_transform_CPUbyteorder_2x(&ivout[0], data, data + (512/4));
+ data += 16;
+ } while (--x >= 0);
+ memcpy(&buf[0], data, 48);
+ memcpy(&buf[16], data + (512/4), 48);
+ /* md5_transform_CPUbyteorder wants data in CPU byte order */
+ /* devSect is already in CPU byte order -- no need to convert */
+ if(sizeof(sector_t) == 8) {
+ /* use only 56 bits of sector number */
+ buf[12] = devSect;
+ buf[13] = (((u_int64_t)devSect >> 32) & 0xFFFFFF) | 0x80000000;
+ buf[16 + 12] = ++devSect;
+ buf[16 + 13] = (((u_int64_t)devSect >> 32) & 0xFFFFFF) | 0x80000000;
+ } else {
+ /* 32 bits of sector number + 24 zero bits */
+ buf[12] = devSect;
+ buf[16 + 13] = buf[13] = 0x80000000;
+ buf[16 + 12] = ++devSect;
+ }
+ /* 4024 bits == 31 * 128 bit plaintext blocks + 56 bits of sector number */
+ /* For version 3 on-disk format this really should be 4536 bits, but can't be */
+ /* changed without breaking compatibility. V3 uses MD5-with-wrong-length IV */
+ buf[16 + 14] = buf[14] = 4024;
+ buf[16 + 15] = buf[15] = 0;
+ md5_transform_CPUbyteorder_2x(&ivout[0], &buf[0], &buf[16]);
+}
+#endif /* defined(HAVE_MD5_2X_IMPLEMENTATION) */
+
+/*
+ * Special requirements for transfer functions:
+ * (1) Plaintext data (loop_buf) may change while it is being read.
+ * (2) On 2.2 and older kernels ciphertext buffer (raw_buf) may be doing
+ * writes to disk at any time, so it can't be used as temporary buffer.
+ */
+static int transfer_aes(struct loop_device *lo, int cmd, char *raw_buf,
+ char *loop_buf, int size, sector_t devSect)
+{
+ aes_context *a;
+ AESmultiKey *m;
+ int x;
+ unsigned y;
+ u_int64_t iv[4], *dip;
+
+ if(!size || (size & 511)) {
+ return -EINVAL;
+ }
+ m = (AESmultiKey *)lo->key_data;
+ y = m->keyMask;
+#ifdef CONFIG_BLK_DEV_LOOP_KEYSCRUB
+ read_lock(&m->rwlock);
+#endif
+ if(cmd == READ) {
+#if defined(HAVE_MD5_2X_IMPLEMENTATION)
+ /* if possible, use faster 2x MD5 implementation, currently AMD64 only (#6) */
+ while((size >= (2*512)) && y) {
+ /* multi-key mode, decrypt 2 sectors at a time */
+ a = m->keyPtr[((unsigned)devSect ) & y];
+ /* decrypt using fake all-zero IV, first sector */
+ memset(iv, 0, 16);
+ x = 15;
+ do {
+ memcpy(&iv[2], raw_buf, 16);
+ aes_decrypt(a, raw_buf, loop_buf);
+ *((u_int64_t *)(&loop_buf[0])) ^= iv[0];
+ *((u_int64_t *)(&loop_buf[8])) ^= iv[1];
+ raw_buf += 16;
+ loop_buf += 16;
+ memcpy(iv, raw_buf, 16);
+ aes_decrypt(a, raw_buf, loop_buf);
+ *((u_int64_t *)(&loop_buf[0])) ^= iv[2];
+ *((u_int64_t *)(&loop_buf[8])) ^= iv[3];
+ raw_buf += 16;
+ loop_buf += 16;
+ } while(--x >= 0);
+ a = m->keyPtr[((unsigned)devSect + 1) & y];
+ /* decrypt using fake all-zero IV, second sector */
+ memset(iv, 0, 16);
+ x = 15;
+ do {
+ memcpy(&iv[2], raw_buf, 16);
+ aes_decrypt(a, raw_buf, loop_buf);
+ *((u_int64_t *)(&loop_buf[0])) ^= iv[0];
+ *((u_int64_t *)(&loop_buf[8])) ^= iv[1];
+ raw_buf += 16;
+ loop_buf += 16;
+ memcpy(iv, raw_buf, 16);
+ aes_decrypt(a, raw_buf, loop_buf);
+ *((u_int64_t *)(&loop_buf[0])) ^= iv[2];
+ *((u_int64_t *)(&loop_buf[8])) ^= iv[3];
+ raw_buf += 16;
+ loop_buf += 16;
+ } while(--x >= 0);
+ /* compute correct IV */
+ memcpy(&iv[0], &m->partialMD5[0], 16);
+ memcpy(&iv[2], &m->partialMD5[0], 16);
+ loop_compute_md5_iv_v3_2x(devSect, (u_int32_t *)iv, (u_int32_t *)(loop_buf - 1008));
+ /* XOR with correct IV now */
+ *((u_int64_t *)(loop_buf - 1024)) ^= iv[0];
+ *((u_int64_t *)(loop_buf - 1016)) ^= iv[1];
+ *((u_int64_t *)(loop_buf - 512)) ^= iv[2];
+ *((u_int64_t *)(loop_buf - 504)) ^= iv[3];
+ size -= 2*512;
+ devSect += 2;
+ }
+#endif /* defined(HAVE_MD5_2X_IMPLEMENTATION) */
+ while(size) {
+ /* decrypt one sector at a time */
+ a = m->keyPtr[((unsigned)devSect) & y];
+ /* decrypt using fake all-zero IV */
+ memset(iv, 0, 16);
+ x = 15;
+ do {
+ memcpy(&iv[2], raw_buf, 16);
+ aes_decrypt(a, raw_buf, loop_buf);
+ *((u_int64_t *)(&loop_buf[0])) ^= iv[0];
+ *((u_int64_t *)(&loop_buf[8])) ^= iv[1];
+ raw_buf += 16;
+ loop_buf += 16;
+ memcpy(iv, raw_buf, 16);
+ aes_decrypt(a, raw_buf, loop_buf);
+ *((u_int64_t *)(&loop_buf[0])) ^= iv[2];
+ *((u_int64_t *)(&loop_buf[8])) ^= iv[3];
+ raw_buf += 16;
+ loop_buf += 16;
+ } while(--x >= 0);
+ if(y) {
+ /* multi-key mode, compute correct IV */
+ memcpy(iv, &m->partialMD5[0], 16);
+ loop_compute_md5_iv_v3(devSect, (u_int32_t *)iv, (u_int32_t *)(loop_buf - 496));
+ } else {
+ /* single-key mode, compute correct IV */
+ loop_compute_sector_iv(devSect, (u_int32_t *)iv);
+ }
+ /* XOR with correct IV now */
+ *((u_int64_t *)(loop_buf - 512)) ^= iv[0];
+ *((u_int64_t *)(loop_buf - 504)) ^= iv[1];
+ size -= 512;
+ devSect++;
+ }
+ } else {
+#if defined(HAVE_MD5_2X_IMPLEMENTATION) && (LINUX_VERSION_CODE >= 0x20400)
+ /* if possible, use faster 2x MD5 implementation, currently AMD64 only (#5) */
+ while((size >= (2*512)) && y) {
+ /* multi-key mode, encrypt 2 sectors at a time */
+ memcpy(raw_buf, loop_buf, 2*512);
+ memcpy(&iv[0], &m->partialMD5[0], 16);
+ memcpy(&iv[2], &m->partialMD5[0], 16);
+ loop_compute_md5_iv_v3_2x(devSect, (u_int32_t *)iv, (u_int32_t *)(&raw_buf[16]));
+ /* first sector */
+ a = m->keyPtr[((unsigned)devSect ) & y];
+ dip = &iv[0];
+ x = 15;
+ do {
+ *((u_int64_t *)(&raw_buf[0])) ^= dip[0];
+ *((u_int64_t *)(&raw_buf[8])) ^= dip[1];
+ aes_encrypt(a, raw_buf, raw_buf);
+ dip = (u_int64_t *)raw_buf;
+ raw_buf += 16;
+ *((u_int64_t *)(&raw_buf[0])) ^= dip[0];
+ *((u_int64_t *)(&raw_buf[8])) ^= dip[1];
+ aes_encrypt(a, raw_buf, raw_buf);
+ dip = (u_int64_t *)raw_buf;
+ raw_buf += 16;
+ } while(--x >= 0);
+ /* second sector */
+ a = m->keyPtr[((unsigned)devSect + 1) & y];
+ dip = &iv[2];
+ x = 15;
+ do {
+ *((u_int64_t *)(&raw_buf[0])) ^= dip[0];
+ *((u_int64_t *)(&raw_buf[8])) ^= dip[1];
+ aes_encrypt(a, raw_buf, raw_buf);
+ dip = (u_int64_t *)raw_buf;
+ raw_buf += 16;
+ *((u_int64_t *)(&raw_buf[0])) ^= dip[0];
+ *((u_int64_t *)(&raw_buf[8])) ^= dip[1];
+ aes_encrypt(a, raw_buf, raw_buf);
+ dip = (u_int64_t *)raw_buf;
+ raw_buf += 16;
+ } while(--x >= 0);
+ loop_buf += 2*512;
+ size -= 2*512;
+ devSect += 2;
+ }
+#endif /* defined(HAVE_MD5_2X_IMPLEMENTATION) && (LINUX_VERSION_CODE >= 0x20400) */
+ while(size) {
+ /* encrypt one sector at a time */
+ a = m->keyPtr[((unsigned)devSect) & y];
+ if(y) {
+ /* multi-key mode encrypt, linux 2.4 and newer */
+ memcpy(raw_buf, loop_buf, 512);
+ memcpy(iv, &m->partialMD5[0], 16);
+ loop_compute_md5_iv_v3(devSect, (u_int32_t *)iv, (u_int32_t *)(&raw_buf[16]));
+ dip = iv;
+ x = 15;
+ do {
+ *((u_int64_t *)(&raw_buf[0])) ^= dip[0];
+ *((u_int64_t *)(&raw_buf[8])) ^= dip[1];
+ aes_encrypt(a, raw_buf, raw_buf);
+ dip = (u_int64_t *)raw_buf;
+ raw_buf += 16;
+ *((u_int64_t *)(&raw_buf[0])) ^= dip[0];
+ *((u_int64_t *)(&raw_buf[8])) ^= dip[1];
+ aes_encrypt(a, raw_buf, raw_buf);
+ dip = (u_int64_t *)raw_buf;
+ raw_buf += 16;
+ } while(--x >= 0);
+ loop_buf += 512;
+ } else {
+ /* single-key mode encrypt */
+ loop_compute_sector_iv(devSect, (u_int32_t *)iv);
+ dip = iv;
+ x = 15;
+ do {
+ iv[2] = *((u_int64_t *)(&loop_buf[0])) ^ dip[0];
+ iv[3] = *((u_int64_t *)(&loop_buf[8])) ^ dip[1];
+ aes_encrypt(a, (unsigned char *)(&iv[2]), raw_buf);
+ dip = (u_int64_t *)raw_buf;
+ loop_buf += 16;
+ raw_buf += 16;
+ iv[2] = *((u_int64_t *)(&loop_buf[0])) ^ dip[0];
+ iv[3] = *((u_int64_t *)(&loop_buf[8])) ^ dip[1];
+ aes_encrypt(a, (unsigned char *)(&iv[2]), raw_buf);
+ dip = (u_int64_t *)raw_buf;
+ loop_buf += 16;
+ raw_buf += 16;
+ } while(--x >= 0);
+ }
+ size -= 512;
+ devSect++;
+ }
+ }
+#ifdef CONFIG_BLK_DEV_LOOP_KEYSCRUB
+ read_unlock(&m->rwlock);
+#endif
+ cond_resched();
+ return(0);
+}
+
+#if defined(CONFIG_BLK_DEV_LOOP_PADLOCK) && (defined(CONFIG_X86) || defined(CONFIG_X86_64))
+static __inline__ void padlock_flush_key_context(void)
+{
+ __asm__ __volatile__("pushf; popf" : : : "cc");
+}
+
+static __inline__ void padlock_rep_xcryptcbc(void *cw, void *k, void *s, void *d, void *iv, unsigned long cnt)
+{
+ __asm__ __volatile__(".byte 0xF3,0x0F,0xA7,0xD0"
+ : "+a" (iv), "+c" (cnt), "+S" (s), "+D" (d) /*output*/
+ : "b" (k), "d" (cw) /*input*/
+ : "cc", "memory" /*modified*/ );
+}
+
+typedef struct {
+#if defined(HAVE_MD5_2X_IMPLEMENTATION)
+ u_int64_t iv[2*2];
+#else
+ u_int64_t iv[2];
+#endif
+ u_int32_t cw[4];
+ u_int32_t dummy1[4];
+} Padlock_IV_CW;
+
+static int transfer_padlock_aes(struct loop_device *lo, int cmd, char *raw_buf,
+ char *loop_buf, int size, sector_t devSect)
+{
+ aes_context *a;
+ AESmultiKey *m;
+ unsigned y;
+ Padlock_IV_CW ivcwua;
+ Padlock_IV_CW *ivcw;
+
+ /* ivcw->iv and ivcw->cw must have 16 byte alignment */
+ ivcw = (Padlock_IV_CW *)(((unsigned long)&ivcwua + 15) & ~((unsigned long)15));
+ ivcw->cw[3] = ivcw->cw[2] = ivcw->cw[1] = 0;
+
+ if(!size || (size & 511) || (((unsigned long)raw_buf | (unsigned long)loop_buf) & 15)) {
+ return -EINVAL;
+ }
+ m = (AESmultiKey *)lo->key_data;
+ y = m->keyMask;
+#ifdef CONFIG_BLK_DEV_LOOP_KEYSCRUB
+ read_lock(&m->rwlock);
+#endif
+ if(cmd == READ) {
+ ivcw->cw[0] = m->padlock_cw_d;
+#if defined(HAVE_MD5_2X_IMPLEMENTATION)
+ /* if possible, use faster 2x MD5 implementation, currently AMD64 only (#4) */
+ while((size >= (2*512)) && y) {
+ /* decrypt using fake all-zero IV */
+ memset(&ivcw->iv[0], 0, 2*16);
+ a = m->keyPtr[((unsigned)devSect ) & y];
+ padlock_flush_key_context();
+ padlock_rep_xcryptcbc(&ivcw->cw[0], &a->aes_d_key[0], raw_buf, loop_buf, &ivcw->iv[0], 32);
+ a = m->keyPtr[((unsigned)devSect + 1) & y];
+ padlock_flush_key_context();
+ padlock_rep_xcryptcbc(&ivcw->cw[0], &a->aes_d_key[0], raw_buf + 512, loop_buf + 512, &ivcw->iv[2], 32);
+ /* compute correct IV */
+ memcpy(&ivcw->iv[0], &m->partialMD5[0], 16);
+ memcpy(&ivcw->iv[2], &m->partialMD5[0], 16);
+ loop_compute_md5_iv_v3_2x(devSect, (u_int32_t *)(&ivcw->iv[0]), (u_int32_t *)(&loop_buf[16]));
+ /* XOR with correct IV now */
+ *((u_int64_t *)(&loop_buf[0])) ^= ivcw->iv[0];
+ *((u_int64_t *)(&loop_buf[8])) ^= ivcw->iv[1];
+ *((u_int64_t *)(&loop_buf[512 + 0])) ^= ivcw->iv[2];
+ *((u_int64_t *)(&loop_buf[512 + 8])) ^= ivcw->iv[3];
+ size -= 2*512;
+ raw_buf += 2*512;
+ loop_buf += 2*512;
+ devSect += 2;
+ }
+#endif /* defined(HAVE_MD5_2X_IMPLEMENTATION) */
+ while(size) {
+ a = m->keyPtr[((unsigned)devSect) & y];
+ padlock_flush_key_context();
+ if(y) {
+ /* decrypt using fake all-zero IV */
+ memset(&ivcw->iv[0], 0, 16);
+ padlock_rep_xcryptcbc(&ivcw->cw[0], &a->aes_d_key[0], raw_buf, loop_buf, &ivcw->iv[0], 32);
+ /* compute correct IV */
+ memcpy(&ivcw->iv[0], &m->partialMD5[0], 16);
+ loop_compute_md5_iv_v3(devSect, (u_int32_t *)(&ivcw->iv[0]), (u_int32_t *)(&loop_buf[16]));
+ /* XOR with correct IV now */
+ *((u_int64_t *)(&loop_buf[ 0])) ^= ivcw->iv[0];
+ *((u_int64_t *)(&loop_buf[ 8])) ^= ivcw->iv[1];
+ } else {
+ loop_compute_sector_iv(devSect, (u_int32_t *)(&ivcw->iv[0]));
+ padlock_rep_xcryptcbc(&ivcw->cw[0], &a->aes_d_key[0], raw_buf, loop_buf, &ivcw->iv[0], 32);
+ }
+ size -= 512;
+ raw_buf += 512;
+ loop_buf += 512;
+ devSect++;
+ }
+ } else {
+ ivcw->cw[0] = m->padlock_cw_e;
+#if defined(HAVE_MD5_2X_IMPLEMENTATION)
+ /* if possible, use faster 2x MD5 implementation, currently AMD64 only (#3) */
+ while((size >= (2*512)) && y) {
+ memcpy(raw_buf, loop_buf, 2*512);
+ memcpy(&ivcw->iv[0], &m->partialMD5[0], 16);
+ memcpy(&ivcw->iv[2], &m->partialMD5[0], 16);
+ loop_compute_md5_iv_v3_2x(devSect, (u_int32_t *)(&ivcw->iv[0]), (u_int32_t *)(&raw_buf[16]));
+ a = m->keyPtr[((unsigned)devSect ) & y];
+ padlock_flush_key_context();
+ padlock_rep_xcryptcbc(&ivcw->cw[0], &a->aes_e_key[0], raw_buf, raw_buf, &ivcw->iv[0], 32);
+ a = m->keyPtr[((unsigned)devSect + 1) & y];
+ padlock_flush_key_context();
+ padlock_rep_xcryptcbc(&ivcw->cw[0], &a->aes_e_key[0], raw_buf + 512, raw_buf + 512, &ivcw->iv[2], 32);
+ size -= 2*512;
+ raw_buf += 2*512;
+ loop_buf += 2*512;
+ devSect += 2;
+ }
+#endif /* defined(HAVE_MD5_2X_IMPLEMENTATION) */
+ while(size) {
+ a = m->keyPtr[((unsigned)devSect) & y];
+ padlock_flush_key_context();
+ if(y) {
+ memcpy(raw_buf, loop_buf, 512);
+ memcpy(&ivcw->iv[0], &m->partialMD5[0], 16);
+ loop_compute_md5_iv_v3(devSect, (u_int32_t *)(&ivcw->iv[0]), (u_int32_t *)(&raw_buf[16]));
+ padlock_rep_xcryptcbc(&ivcw->cw[0], &a->aes_e_key[0], raw_buf, raw_buf, &ivcw->iv[0], 32);
+ } else {
+ loop_compute_sector_iv(devSect, (u_int32_t *)(&ivcw->iv[0]));
+ padlock_rep_xcryptcbc(&ivcw->cw[0], &a->aes_e_key[0], loop_buf, raw_buf, &ivcw->iv[0], 32);
+ }
+ size -= 512;
+ raw_buf += 512;
+ loop_buf += 512;
+ devSect++;
+ }
+ }
+#ifdef CONFIG_BLK_DEV_LOOP_KEYSCRUB
+ read_unlock(&m->rwlock);
+#endif
+ cond_resched();
+ return(0);
+}
+#endif
+
+#if defined(CONFIG_BLK_DEV_LOOP_INTELAES) && (defined(CONFIG_X86) || defined(CONFIG_X86_64))
+asmlinkage extern void intel_aes_cbc_encrypt(const aes_context *, void *src, void *dst, size_t len, void *iv);
+asmlinkage extern void intel_aes_cbc_decrypt(const aes_context *, void *src, void *dst, size_t len, void *iv);
+asmlinkage extern void intel_aes_cbc_enc_4x512(aes_context **, void *src, void *dst, void *iv);
+
+static int transfer_intel_aes(struct loop_device *lo, int cmd, char *raw_buf,
+ char *loop_buf, int size, sector_t devSect)
+{
+ aes_context *acpa[4];
+ AESmultiKey *m;
+ unsigned y;
+ u_int64_t ivua[(4*2)+2];
+ u_int64_t *iv;
+
+ /* make iv 16 byte aligned */
+ iv = (u_int64_t *)(((unsigned long)&ivua + 15) & ~((unsigned long)15));
+
+ if(!size || (size & 511) || (((unsigned long)raw_buf | (unsigned long)loop_buf) & 15)) {
+ return -EINVAL;
+ }
+ m = (AESmultiKey *)lo->key_data;
+ y = m->keyMask;
+#ifdef CONFIG_BLK_DEV_LOOP_KEYSCRUB
+ read_lock(&m->rwlock);
+#endif
+ kernel_fpu_begin(); /* intel_aes_* code uses xmm registers */
+ if(cmd == READ) {
+#if defined(HAVE_MD5_2X_IMPLEMENTATION)
+ /* if possible, use faster 2x MD5 implementation, currently AMD64 only (#2) */
+ while((size >= (2*512)) && y) {
+ acpa[0] = m->keyPtr[((unsigned)devSect ) & y];
+ acpa[1] = m->keyPtr[((unsigned)devSect + 1) & y];
+ /* decrypt using fake all-zero IV */
+ memset(iv, 0, 2*16);
+ intel_aes_cbc_decrypt(acpa[0], raw_buf, loop_buf, 512, &iv[0]);
+ intel_aes_cbc_decrypt(acpa[1], raw_buf + 512, loop_buf + 512, 512, &iv[2]);
+ /* compute correct IV, use 2x parallelized version */
+ memcpy(&iv[0], &m->partialMD5[0], 16);
+ memcpy(&iv[2], &m->partialMD5[0], 16);
+ loop_compute_md5_iv_v3_2x(devSect, (u_int32_t *)iv, (u_int32_t *)(&loop_buf[16]));
+ /* XOR with correct IV now */
+ *((u_int64_t *)(&loop_buf[0])) ^= iv[0];
+ *((u_int64_t *)(&loop_buf[8])) ^= iv[1];
+ *((u_int64_t *)(&loop_buf[512 + 0])) ^= iv[2];
+ *((u_int64_t *)(&loop_buf[512 + 8])) ^= iv[3];
+ size -= 2*512;
+ raw_buf += 2*512;
+ loop_buf += 2*512;
+ devSect += 2;
+ }
+#endif /* defined(HAVE_MD5_2X_IMPLEMENTATION) */
+ while(size) {
+ acpa[0] = m->keyPtr[((unsigned)devSect) & y];
+ if(y) {
+ /* decrypt using fake all-zero IV */
+ memset(iv, 0, 16);
+ intel_aes_cbc_decrypt(acpa[0], raw_buf, loop_buf, 512, iv);
+ /* compute correct IV */
+ memcpy(iv, &m->partialMD5[0], 16);
+ loop_compute_md5_iv_v3(devSect, (u_int32_t *)iv, (u_int32_t *)(&loop_buf[16]));
+ /* XOR with correct IV now */
+ *((u_int64_t *)(&loop_buf[0])) ^= iv[0];
+ *((u_int64_t *)(&loop_buf[8])) ^= iv[1];
+ } else {
+ loop_compute_sector_iv(devSect, (u_int32_t *)iv);
+ intel_aes_cbc_decrypt(acpa[0], raw_buf, loop_buf, 512, iv);
+ }
+ size -= 512;
+ raw_buf += 512;
+ loop_buf += 512;
+ devSect++;
+ }
+ } else {
+ /* if possible, use faster 4-chains at a time encrypt implementation (#1) */
+ while(size >= (4*512)) {
+ acpa[0] = m->keyPtr[((unsigned)devSect ) & y];
+ acpa[1] = m->keyPtr[((unsigned)devSect + 1) & y];
+ acpa[2] = m->keyPtr[((unsigned)devSect + 2) & y];
+ acpa[3] = m->keyPtr[((unsigned)devSect + 3) & y];
+ if(y) {
+ memcpy(raw_buf, loop_buf, 4*512);
+ memcpy(&iv[0], &m->partialMD5[0], 16);
+ memcpy(&iv[2], &m->partialMD5[0], 16);
+ memcpy(&iv[4], &m->partialMD5[0], 16);
+ memcpy(&iv[6], &m->partialMD5[0], 16);
+#if defined(HAVE_MD5_2X_IMPLEMENTATION)
+ /* use 2x parallelized version */
+ loop_compute_md5_iv_v3_2x(devSect, (u_int32_t *)(&iv[0]), (u_int32_t *)(&raw_buf[ 16]));
+ loop_compute_md5_iv_v3_2x(devSect + 2, (u_int32_t *)(&iv[4]), (u_int32_t *)(&raw_buf[0x400 + 16]));
+#else
+ loop_compute_md5_iv_v3(devSect, (u_int32_t *)(&iv[0]), (u_int32_t *)(&raw_buf[ 16]));
+ loop_compute_md5_iv_v3(devSect + 1, (u_int32_t *)(&iv[2]), (u_int32_t *)(&raw_buf[0x200 + 16]));
+ loop_compute_md5_iv_v3(devSect + 2, (u_int32_t *)(&iv[4]), (u_int32_t *)(&raw_buf[0x400 + 16]));
+ loop_compute_md5_iv_v3(devSect + 3, (u_int32_t *)(&iv[6]), (u_int32_t *)(&raw_buf[0x600 + 16]));
+#endif
+ intel_aes_cbc_enc_4x512(&acpa[0], raw_buf, raw_buf, iv);
+ } else {
+ loop_compute_sector_iv(devSect, (u_int32_t *)(&iv[0]));
+ loop_compute_sector_iv(devSect + 1, (u_int32_t *)(&iv[2]));
+ loop_compute_sector_iv(devSect + 2, (u_int32_t *)(&iv[4]));
+ loop_compute_sector_iv(devSect + 3, (u_int32_t *)(&iv[6]));
+ intel_aes_cbc_enc_4x512(&acpa[0], loop_buf, raw_buf, iv);
+ }
+ size -= 4*512;
+ raw_buf += 4*512;
+ loop_buf += 4*512;
+ devSect += 4;
+ }
+ /* encrypt the rest (if any) using slower 1-chain at a time implementation */
+ while(size) {
+ acpa[0] = m->keyPtr[((unsigned)devSect) & y];
+ if(y) {
+ memcpy(raw_buf, loop_buf, 512);
+ memcpy(iv, &m->partialMD5[0], 16);
+ loop_compute_md5_iv_v3(devSect, (u_int32_t *)iv, (u_int32_t *)(&raw_buf[16]));
+ intel_aes_cbc_encrypt(acpa[0], raw_buf, raw_buf, 512, iv);
+ } else {
+ loop_compute_sector_iv(devSect, (u_int32_t *)iv);
+ intel_aes_cbc_encrypt(acpa[0], loop_buf, raw_buf, 512, iv);
+ }
+ size -= 512;
+ raw_buf += 512;
+ loop_buf += 512;
+ devSect++;
+ }
+ }
+ kernel_fpu_end(); /* intel_aes_* code uses xmm registers */
+#ifdef CONFIG_BLK_DEV_LOOP_KEYSCRUB
+ read_unlock(&m->rwlock);
+#endif
+ cond_resched();
+ return(0);
+}
+#endif
+
+static struct loop_func_table funcs_aes = {
+ number: 16, /* 16 == AES */
+ transfer: transfer_aes,
+ init: keySetup_aes,
+ release: keyClean_aes,
+ ioctl: handleIoctl_aes
+};
+
+#if defined(CONFIG_BLK_DEV_LOOP_PADLOCK) && (defined(CONFIG_X86) || defined(CONFIG_X86_64))
+static struct loop_func_table funcs_padlock_aes = {
+ number: 16, /* 16 == AES */
+ transfer: transfer_padlock_aes,
+ init: keySetup_aes,
+ release: keyClean_aes,
+ ioctl: handleIoctl_aes
+};
+#endif
+
+#if defined(CONFIG_BLK_DEV_LOOP_INTELAES) && (defined(CONFIG_X86) || defined(CONFIG_X86_64))
+static struct loop_func_table funcs_intel_aes = {
+ number: 16, /* 16 == AES */
+ transfer: transfer_intel_aes,
+ init: keySetup_aes,
+ release: keyClean_aes,
+ ioctl: handleIoctl_aes
+};
+#endif
+
+#if defined(CONFIG_BLK_DEV_LOOP_PADLOCK) && (defined(CONFIG_X86) || defined(CONFIG_X86_64))
+static int CentaurHauls_ID_and_enabled_ACE(void)
+{
+ unsigned int eax = 0, ebx = 0, ecx = 0, edx = 0;
+
+ /* check for "CentaurHauls" ID string, and enabled ACE */
+ cpuid(0x00000000, &eax, &ebx, &ecx, &edx);
+ if((ebx == 0x746e6543) && (edx == 0x48727561) && (ecx == 0x736c7561)
+ && (cpuid_eax(0xC0000000) >= 0xC0000001)
+ && ((cpuid_edx(0xC0000001) & 0xC0) == 0xC0)) {
+ return 1; /* ACE enabled */
+ }
+ return 0;
+}
+#endif
+
+EXPORT_SYMBOL(loop_compute_sector_iv);
+EXPORT_SYMBOL(loop_compute_md5_iv_v3);
+EXPORT_SYMBOL(loop_compute_md5_iv);
+EXPORT_SYMBOL(md5_transform_CPUbyteorder_C);
+#endif /* CONFIG_BLK_DEV_LOOP_AES */
+
+/* xfer_funcs[0] is special - its release function is never called */
+static struct loop_func_table *xfer_funcs[MAX_LO_CRYPT] = {
+ &none_funcs,
+ &xor_funcs,
+#ifdef CONFIG_BLK_DEV_LOOP_AES
+ [LO_CRYPT_AES] = &funcs_aes,
+#endif
+};
+
+/*
+ * First number of 'lo_prealloc' is the default number of RAM pages
+ * to pre-allocate for each device backed loop. Every (configured)
+ * device backed loop pre-allocates this amount of RAM pages unless
+ * later 'lo_prealloc' numbers provide an override. 'lo_prealloc'
+ * overrides are defined in pairs: loop_index,number_of_pages
+ */
+static int lo_prealloc[9] = { 256, -1, 0, -1, 0, -1, 0, -1, 0 };
+#define LO_PREALLOC_MIN 4 /* minimum user defined pre-allocated RAM pages */
+#define LO_PREALLOC_MAX 4096 /* maximum user defined pre-allocated RAM pages */
+
+#ifdef MODULE
+static int dummy1;
+module_param_array(lo_prealloc, int, &dummy1, 0);
+MODULE_PARM_DESC(lo_prealloc, "Number of pre-allocated pages [,index,pages]...");
+#else
+static int __init lo_prealloc_setup(char *str)
+{
+ int x, y, z;
+
+ for (x = 0; x < (sizeof(lo_prealloc) / sizeof(int)); x++) {
+ z = get_option(&str, &y);
+ if (z > 0)
+ lo_prealloc[x] = y;
+ if (z < 2)
+ break;
+ }
+ return 1;
+}
+__setup("lo_prealloc=", lo_prealloc_setup);
+#endif
+
+/*
+ * First number of 'lo_threads' is the default number of helper threads to
+ * create for each device backed loop device. Every (configured) device
+ * backed loop device has this many threads unless later 'lo_threads'
+ * numbers provide an override. File backed loops always have 1 helper
+ * thread. 'lo_threads' overrides are defined in pairs: loop_index,threads
+ *
+ * This value is ignored on 2.6.18 and older kernels.
+ */
+static int lo_threads[9] = { 1, -1, 0, -1, 0, -1, 0, -1, 0 };
+#define LO_THREADS_MIN 1 /* minimum user defined thread count */
+#define LO_THREADS_MAX 4 /* maximum user defined thread count */
+
+#ifdef MODULE
+static int dummy2;
+module_param_array(lo_threads, int, &dummy2, 0);
+MODULE_PARM_DESC(lo_threads, "Number of threads per loop [,index,threads]...");
+#else
+static int __init lo_threads_setup(char *str)
+{
+ int x, y, z;
+
+ for (x = 0; x < (sizeof(lo_threads) / sizeof(int)); x++) {
+ z = get_option(&str, &y);
+ if (z > 0)
+ lo_threads[x] = y;
+ if (z < 2)
+ break;
+ }
+ return 1;
+}
+__setup("lo_threads=", lo_threads_setup);
+#endif
+
+/*
+ * This is loop helper thread nice value in range
+ * from 0 (low priority) to -20 (high priority).
+ */
+static int lo_nice = -1;
+
+#ifdef MODULE
+module_param(lo_nice, int, 0);
+MODULE_PARM_DESC(lo_nice, "Loop thread scheduler nice (0 ... -20)");
+#else
+static int __init lo_nice_setup(char *str)
+{
+ int y;
+
+ if (get_option(&str, &y) == 1)
+ lo_nice = y;
+ return 1;
+}
+__setup("lo_nice=", lo_nice_setup);
+#endif
+
+struct loop_bio_extension {
+ struct bio *bioext_merge;
+ struct loop_device *bioext_loop;
+ struct bio_vec *bioext_bi_io_vec_orig;
+ sector_t bioext_iv;
+ int bioext_index;
+ int bioext_size;
+ unsigned int bioext_bi_max_vecs_orig;
+#if LINUX_VERSION_CODE >= 0x30e00
+ unsigned int bioext_done_offset;
+#endif
+};
+
+#if LINUX_VERSION_CODE >= 0x30e00
+# define LOOP_COMPAT_BI_SECTOR bi_iter.bi_sector
+# define LOOP_COMPAT_BI_SIZE bi_iter.bi_size
+# define LOOP_COMPAT_BI_IDX bi_iter.bi_idx
+#else
+# define LOOP_COMPAT_BI_SECTOR bi_sector
+# define LOOP_COMPAT_BI_SIZE bi_size
+# define LOOP_COMPAT_BI_IDX bi_idx
+#endif
+
+static struct loop_device **loop_dev_ptr_arr;
+
+static void loop_prealloc_cleanup(struct loop_device *lo)
+{
+ struct bio *bio;
+ struct loop_bio_extension *extension;
+
+ while ((bio = lo->lo_bio_free0)) {
+ lo->lo_bio_free0 = bio->bi_next;
+ extension = bio->bi_private;
+ bio->bi_io_vec = extension->bioext_bi_io_vec_orig;
+ bio->bi_max_vecs = extension->bioext_bi_max_vecs_orig;
+ bio->bi_vcnt = 1;
+ __free_page(bio->bi_io_vec[0].bv_page);
+ kfree(extension);
+ bio->bi_next = NULL;
+ bio_put(bio);
+ }
+ while ((bio = lo->lo_bio_free1)) {
+ lo->lo_bio_free1 = bio->bi_next;
+ /* bi_flags bit 0 was used for other purpose */
+ clear_bit(0, &bio->bi_flags);
+ /* bi_size was used for other purpose */
+ bio->LOOP_COMPAT_BI_SIZE = 0;
+ /* bi_cnt was used for other purpose */
+ atomic_set(&bio->bi_cnt, 1);
+ bio->bi_next = NULL;
+ bio_put(bio);
+ }
+}
+
+static int loop_prealloc_init(struct loop_device *lo, int y)
+{
+ struct bio *bio;
+ struct loop_bio_extension *extension;
+ int x;
+
+ if(!y) {
+ y = lo_prealloc[0];
+ for (x = 1; x < (sizeof(lo_prealloc) / sizeof(int)); x += 2) {
+ if (lo_prealloc[x + 1] && (lo->lo_number == lo_prealloc[x])) {
+ y = lo_prealloc[x + 1];
+ break;
+ }
+ }
+ }
+ lo->lo_bio_flshMax = (y * 3) / 4;
+ lo->lo_bio_flshCnt = 0;
+
+ for (x = 0; x < y; x++) {
+ bio = bio_alloc(GFP_KERNEL, 1);
+ if (!bio) {
+ fail1:
+ loop_prealloc_cleanup(lo);
+ return 1;
+ }
+ bio->bi_io_vec[0].bv_page = alloc_page(GFP_KERNEL);
+ if (!bio->bi_io_vec[0].bv_page) {
+ fail2:
+ bio->bi_next = NULL;
+ bio_put(bio);
+ goto fail1;
+ }
+ memset(page_address(bio->bi_io_vec[0].bv_page), 0, PAGE_SIZE);
+ bio->bi_vcnt = 1;
+ extension = kmalloc(sizeof(struct loop_bio_extension), GFP_KERNEL);
+ if (!extension) {
+ __free_page(bio->bi_io_vec[0].bv_page);
+ goto fail2;
+ }
+ bio->bi_private = extension;
+ extension->bioext_bi_io_vec_orig = bio->bi_io_vec;
+ extension->bioext_bi_max_vecs_orig = bio->bi_max_vecs;
+ bio->bi_next = lo->lo_bio_free0;
+ lo->lo_bio_free0 = bio;
+
+ bio = bio_alloc(GFP_KERNEL, 1);
+ if (!bio)
+ goto fail1;
+ bio->bi_vcnt = 1;
+ bio->bi_next = lo->lo_bio_free1;
+ lo->lo_bio_free1 = bio;
+ }
+ return 0;
+}
+
+static void loop_add_queue_last(struct loop_device *lo, struct bio *bio, struct bio **q)
+{
+ unsigned long flags;
+
+ spin_lock_irqsave(&lo->lo_lock, flags);
+ if (*q) {
+ bio->bi_next = (*q)->bi_next;
+ (*q)->bi_next = bio;
+ } else {
+ bio->bi_next = bio;
+ }
+ *q = bio;
+ spin_unlock_irqrestore(&lo->lo_lock, flags);
+
+ if (waitqueue_active(&lo->lo_bio_wait))
+ wake_up_interruptible_all(&lo->lo_bio_wait);
+}
+
+static struct bio *loop_get_bio(struct loop_device *lo)
+{
+ struct bio *bio = NULL, *last;
+
+ spin_lock_irq(&lo->lo_lock);
+ if ((last = lo->lo_bio_que0)) {
+ bio = last->bi_next;
+ if (bio == last)
+ lo->lo_bio_que0 = NULL;
+ else
+ last->bi_next = bio->bi_next;
+ bio->bi_next = NULL;
+ }
+ spin_unlock_irq(&lo->lo_lock);
+ return bio;
+}
+
+static void loop_put_buffer(struct loop_device *lo, struct bio *b, int flist)
+{
+ unsigned long flags;
+
+ spin_lock_irqsave(&lo->lo_lock, flags);
+ if(!flist) {
+ b->bi_next = lo->lo_bio_free0;
+ lo->lo_bio_free0 = b;
+ } else {
+ b->bi_next = lo->lo_bio_free1;
+ lo->lo_bio_free1 = b;
+ }
+ spin_unlock_irqrestore(&lo->lo_lock, flags);
+
+ if (waitqueue_active(&lo->lo_buf_wait))
+ wake_up_all(&lo->lo_buf_wait);
+}
+
+static void loop_end_io_transfer(struct bio *bio, int err)
+{
+ struct loop_bio_extension *extension = bio->bi_private;
+ struct bio *merge = extension->bioext_merge;
+ struct loop_device *lo = extension->bioext_loop;
+ struct bio *origbio = merge->bi_private;
+
+ if (err) {
+ merge->LOOP_COMPAT_BI_SIZE = err; /* used as error code */
+ if(err == -EIO)
+ clear_bit(0, &merge->bi_flags);
+ printk(KERN_ERR "loop%d: loop_end_io_transfer err=%d bi_rw=0x%lx\n", lo->lo_number, err, bio->bi_rw);
+ }
+ if (bio_rw(bio) == WRITE) {
+ loop_put_buffer(lo, bio, 0);
+ if (!atomic_dec_and_test(&merge->bi_cnt)) {
+ return;
+ }
+ origbio->bi_next = NULL;
+ bio_endio(origbio, test_bit(0, &merge->bi_flags) ? (int)merge->LOOP_COMPAT_BI_SIZE : -EIO);
+ loop_put_buffer(lo, merge, 1);
+ if (atomic_dec_and_test(&lo->lo_pending))
+ wake_up_interruptible_all(&lo->lo_bio_wait);
+ } else {
+ loop_add_queue_last(lo, bio, &lo->lo_bio_que0);
+ }
+}
+
+static struct bio *loop_get_buffer(struct loop_device *lo, struct bio *orig_bio,
+ struct bio **merge_ptr, int *flushPtr, int origHasData, int *lastVePtr)
+{
+ struct bio *bio = NULL, *merge = *merge_ptr, *fbtst;
+ struct loop_bio_extension *extension;
+ int len, nzCnt, flsh = 0, firstVec, lastVec;
+
+ spin_lock_irq(&lo->lo_lock);
+ if (!merge) {
+ merge = lo->lo_bio_free1;
+ if (merge) {
+ lo->lo_bio_free1 = merge->bi_next;
+ }
+ }
+ if (merge) {
+ bio = lo->lo_bio_free0;
+ if (bio) {
+ lo->lo_bio_free0 = bio->bi_next;
+ }
+ }
+ fbtst = lo->lo_bio_free0;
+ if(!fbtst || !fbtst->bi_next) {
+ flsh = 1;
+ }
+ fbtst = lo->lo_bio_free1;
+ if(!fbtst || !fbtst->bi_next) {
+ flsh = 1;
+ }
+ spin_unlock_irq(&lo->lo_lock);
+
+ *flushPtr = flsh;
+
+ if (!(*merge_ptr) && merge) {
+ /*
+ * initialize "merge-bio" which is used as
+ * rendezvous point among multiple vecs
+ */
+ *merge_ptr = merge;
+ merge->LOOP_COMPAT_BI_SECTOR = orig_bio->LOOP_COMPAT_BI_SECTOR + lo->lo_offs_sec;
+ merge->LOOP_COMPAT_BI_SIZE = 0; /* used as error code */
+ set_bit(0, &merge->bi_flags);
+ merge->LOOP_COMPAT_BI_IDX = orig_bio->LOOP_COMPAT_BI_IDX;
+#if LINUX_VERSION_CODE >= 0x30e00
+ nzCnt = 1;
+ if(origHasData) {
+ /* compute total number of vecs that this driver must process */
+ unsigned int orIdx = orig_bio->LOOP_COMPAT_BI_IDX;
+ unsigned int orSiz = orig_bio->LOOP_COMPAT_BI_SIZE;
+ unsigned int orDon = orig_bio->bi_iter.bi_bvec_done;
+ unsigned int vecRem;
+ do {
+ vecRem = orig_bio->bi_io_vec[orIdx].bv_len - orDon;
+ if(orSiz <= vecRem)
+ break;
+ orSiz -= vecRem;
+ nzCnt++;
+ orIdx++;
+ orDon = 0;
+ } while(1);
+ }
+#else
+ nzCnt = orig_bio->bi_vcnt - orig_bio->LOOP_COMPAT_BI_IDX;
+ if(nzCnt < 1) nzCnt = 1;
+#endif
+ atomic_set(&merge->bi_cnt, nzCnt);
+ merge->bi_private = orig_bio;
+ }
+
+ if (!bio)
+ return NULL;
+
+ extension = bio->bi_private;
+ if(origHasData) {
+ firstVec = (merge->LOOP_COMPAT_BI_IDX == orig_bio->LOOP_COMPAT_BI_IDX) ? 1 : 0;
+#if LINUX_VERSION_CODE >= 0x30e00
+ lastVec = (orig_bio->LOOP_COMPAT_BI_SIZE <= (orig_bio->bi_io_vec[merge->LOOP_COMPAT_BI_IDX].bv_len - orig_bio->bi_iter.bi_bvec_done)) ? 1 : 0;
+#else
+ lastVec = (merge->LOOP_COMPAT_BI_IDX == (orig_bio->bi_vcnt - 1)) ? 1 : 0;
+#endif
+ } else {
+ firstVec = 1;
+ lastVec = 1;
+ }
+ *lastVePtr = lastVec;
+
+ /*
+ * initialize one page "buffer-bio"
+ */
+#if LINUX_VERSION_CODE >= 0x30700
+ bio_reset(bio);
+ bio->bi_private = extension;
+#else
+#if !defined(BIO_RESET_BITS)
+# define BIO_RESET_BITS BIO_POOL_OFFSET
+#endif
+ bio->bi_flags &= (~0UL << BIO_RESET_BITS);
+ bio->bi_flags |= (1 << BIO_UPTODATE);
+#endif
+ bio->LOOP_COMPAT_BI_SECTOR = merge->LOOP_COMPAT_BI_SECTOR;
+ bio->bi_next = NULL;
+ bio->bi_bdev = lo->lo_device;
+
+#if LINUX_VERSION_CODE < 0x30200
+ if(orig_bio->bi_flags & (1 << BIO_CPU_AFFINE)) {
+ bio->bi_comp_cpu = orig_bio->bi_comp_cpu;
+ bio->bi_flags |= (1 << BIO_CPU_AFFINE);
+ }
+#endif
+ /* read-ahead bit needs to be cleared to work around kernel bug */
+ /* that causes I/O errors on -EWOULDBLOCK I/O elevator failures */
+ bio->bi_rw = orig_bio->bi_rw & ~L_BIO_RW_AHEAD;
+
+ if(orig_bio->bi_rw & REQ_FLUSH) {
+ if(!firstVec) {
+ bio->bi_rw &= ~REQ_FLUSH;
+ } else {
+ *flushPtr = 1;
+ }
+ }
+ if(orig_bio->bi_rw & REQ_FUA) {
+ if(!lastVec) {
+ bio->bi_rw &= ~REQ_FUA;
+ } else {
+ *flushPtr = 1;
+ }
+ }
+ if(orig_bio->bi_rw & L_BIO_RW_SYNCIO) {
+ if(!lastVec) {
+ bio->bi_rw &= ~L_BIO_RW_SYNCIO;
+ } else {
+ *flushPtr = 1;
+ }
+ }
+ if(orig_bio->bi_rw & L_BIO_RW_NOIDLE) {
+ if(!lastVec) {
+ bio->bi_rw &= ~L_BIO_RW_NOIDLE;
+ }
+ }
+ if(flsh) {
+ bio->bi_rw |= L_BIO_RW_NOIDLE;
+ }
+
+ bio->LOOP_COMPAT_BI_IDX = 0;
+ bio->bi_phys_segments = 0;
+ if(origHasData) {
+ /* original bio has data */
+ bio->bi_io_vec = extension->bioext_bi_io_vec_orig;
+ bio->bi_max_vecs = extension->bioext_bi_max_vecs_orig;
+ bio->bi_vcnt = 1;
+#if LINUX_VERSION_CODE >= 0x30e00
+ len = orig_bio->bi_io_vec[merge->LOOP_COMPAT_BI_IDX].bv_len - orig_bio->bi_iter.bi_bvec_done;
+ if(len > orig_bio->LOOP_COMPAT_BI_SIZE)
+ len = orig_bio->LOOP_COMPAT_BI_SIZE;
+ bio->LOOP_COMPAT_BI_SIZE = len;
+ extension->bioext_done_offset = orig_bio->bi_iter.bi_bvec_done;
+#else
+ bio->LOOP_COMPAT_BI_SIZE = len = orig_bio->bi_io_vec[merge->LOOP_COMPAT_BI_IDX].bv_len;
+#endif
+ bio->bi_io_vec[0].bv_len = len;
+ bio->bi_io_vec[0].bv_offset = 0;
+ } else {
+ /* original bio does not have data */
+ bio->bi_io_vec = 0;
+ bio->bi_max_vecs = 0;
+ bio->bi_vcnt = 0;
+#if LINUX_VERSION_CODE >= 0x30e00
+ extension->bioext_done_offset = 0;
+#endif
+ bio->LOOP_COMPAT_BI_SIZE = len = 0;
+ }
+
+ bio->bi_seg_front_size = 0;
+ bio->bi_seg_back_size = 0;
+ bio->bi_end_io = loop_end_io_transfer;
+
+ /*
+ * initialize "buffer-bio" extension. This extension is
+ * permanently glued to above "buffer-bio" via bio->bi_private
+ */
+ extension->bioext_merge = merge;
+ extension->bioext_loop = lo;
+ extension->bioext_iv = merge->LOOP_COMPAT_BI_SECTOR - lo->lo_iv_remove;
+ extension->bioext_index = merge->LOOP_COMPAT_BI_IDX;
+ extension->bioext_size = len;
+
+ /*
+ * prepare "merge-bio" for next vec
+ */
+ merge->LOOP_COMPAT_BI_SECTOR += len >> 9;
+ merge->LOOP_COMPAT_BI_IDX++;
+
+#if LINUX_VERSION_CODE >= 0x30e00
+ orig_bio->LOOP_COMPAT_BI_SIZE -= len;
+ orig_bio->bi_iter.bi_bvec_done = 0;
+ bio->bi_iter.bi_bvec_done = 0;
+#endif
+ return bio;
+}
+
+static int figure_loop_size(struct loop_device *lo, struct block_device *bdev)
+{
+ loff_t size, offs;
+ sector_t x;
+ int err = 0;
+
+ size = i_size_read(lo->lo_backing_file->f_path.dentry->d_inode->i_mapping->host);
+ offs = lo->lo_offset;
+ if (!(lo->lo_flags & LO_FLAGS_DO_BMAP))
+ offs &= ~((loff_t)511);
+ if ((offs > 0) && (offs < size)) {
+ size -= offs;
+ } else {
+ if (offs)
+ err = -EINVAL;
+ lo->lo_offset = 0;
+ lo->lo_offs_sec = lo->lo_iv_remove = 0;
+ }
+ if ((lo->lo_sizelimit > 0) && (lo->lo_sizelimit <= size)) {
+ size = lo->lo_sizelimit;
+ } else {
+ if (lo->lo_sizelimit)
+ err = -EINVAL;
+ lo->lo_sizelimit = 0;
+ }
+ size >>= 9;
+
+ /*
+ * Unfortunately, if we want to do I/O on the device,
+ * the number of 512-byte sectors has to fit into a sector_t.
+ */
+ x = (sector_t)size;
+ if ((loff_t)x != size) {
+ err = -EFBIG;
+ size = 0;
+ }
+
+ set_capacity(disks[lo->lo_number], size); /* 512 byte units */
+ i_size_write(bdev->bd_inode, size << 9); /* byte units */
+ return err;
+}
+
+static inline int lo_do_transfer(struct loop_device *lo, int cmd, char *rbuf,
+ char *lbuf, int size, sector_t rblock)
+{
+ if (!lo->transfer)
+ return 0;
+
+ return lo->transfer(lo, cmd, rbuf, lbuf, size, rblock);
+}
+
+static int loop_file_io(struct file *file, char *buf, int size, loff_t *ppos, int w)
+{
+ mm_segment_t fs;
+ int x, y, z;
+
+ y = 0;
+ do {
+ z = size - y;
+ fs = get_fs();
+ set_fs(get_ds());
+ if (w) {
+ x = file->f_op->write(file, buf + y, z, ppos);
+ set_fs(fs);
+ } else {
+ x = file->f_op->read(file, buf + y, z, ppos);
+ set_fs(fs);
+ if (!x)
+ return 1;
+ }
+ if (x < 0) {
+ if ((x == -EAGAIN) || (x == -ENOMEM) || (x == -ERESTART) || (x == -EINTR)) {
+ set_current_state(TASK_INTERRUPTIBLE);
+ schedule_timeout(HZ / 2);
+ continue;
+ }
+ return 1;
+ }
+ y += x;
+ } while (y < size);
+ return 0;
+}
+
+static int do_bio_filebacked(struct loop_device *lo, struct bio *bio)
+{
+ loff_t pos;
+ struct file *file = lo->lo_backing_file;
+ char *data, *buf;
+ unsigned int size, len, more;
+ sector_t IV;
+ struct page *pg;
+#if LINUX_VERSION_CODE >= 0x30e00
+ unsigned int oln;
+#endif
+
+ if(!bio_has_data(bio))
+ return 0;
+
+ pos = ((loff_t) bio->LOOP_COMPAT_BI_SECTOR << 9) + lo->lo_offset;
+ buf = page_address(lo->lo_bio_free0->bi_io_vec[0].bv_page);
+ IV = bio->LOOP_COMPAT_BI_SECTOR;
+ if (!lo->lo_iv_remove)
+ IV += lo->lo_offs_sec;
+ do {
+ pg = bio->bi_io_vec[bio->LOOP_COMPAT_BI_IDX].bv_page;
+#if LINUX_VERSION_CODE >= 0x30e00
+ len = bio->bi_io_vec[bio->LOOP_COMPAT_BI_IDX].bv_len - bio->bi_iter.bi_bvec_done;
+ if(len > bio->LOOP_COMPAT_BI_SIZE)
+ len = bio->LOOP_COMPAT_BI_SIZE;
+ oln = len;
+ data = kmap(pg) + bio->bi_io_vec[bio->LOOP_COMPAT_BI_IDX].bv_offset + bio->bi_iter.bi_bvec_done;
+ bio->bi_iter.bi_bvec_done = 0;
+#else
+ len = bio->bi_io_vec[bio->LOOP_COMPAT_BI_IDX].bv_len;
+ data = kmap(pg) + bio->bi_io_vec[bio->LOOP_COMPAT_BI_IDX].bv_offset;
+#endif
+ while (len > 0) {
+ if (!lo->lo_encryption) {
+ /* this code relies that NONE transfer is a no-op */
+ buf = data;
+ }
+ size = PAGE_CACHE_SIZE;
+ if (size > len)
+ size = len;
+ if (bio_rw(bio) == WRITE) {
+ if (lo_do_transfer(lo, WRITE, buf, data, size, IV)) {
+ printk(KERN_ERR "loop%d: write transfer error, sector %llu\n", lo->lo_number, (unsigned long long)IV);
+ goto kunmap_and_out;
+ }
+ if (loop_file_io(file, buf, size, &pos, 1)) {
+ printk(KERN_ERR "loop%d: write i/o error, sector %llu\n", lo->lo_number, (unsigned long long)IV);
+ goto kunmap_and_out;
+ }
+ } else {
+ if (loop_file_io(file, buf, size, &pos, 0)) {
+ printk(KERN_ERR "loop%d: read i/o error, sector %llu\n", lo->lo_number, (unsigned long long)IV);
+ goto kunmap_and_out;
+ }
+ if (lo_do_transfer(lo, READ, buf, data, size, IV)) {
+ printk(KERN_ERR "loop%d: read transfer error, sector %llu\n", lo->lo_number, (unsigned long long)IV);
+ goto kunmap_and_out;
+ }
+ flush_dcache_page(pg);
+ }
+ data += size;
+ len -= size;
+ IV += size >> 9;
+ }
+ kunmap(pg);
+ bio->LOOP_COMPAT_BI_IDX++;
+#if LINUX_VERSION_CODE >= 0x30e00
+ more = ((bio->LOOP_COMPAT_BI_SIZE -= oln) > 0) ? 1 : 0;
+#else
+ more = (bio->LOOP_COMPAT_BI_IDX < bio->bi_vcnt) ? 1 : 0;
+#endif
+ } while (more);
+ return 0;
+
+kunmap_and_out:
+ kunmap(pg);
+ return -EIO;
+}
+
+static void loop_unplug_backingdev(struct request_queue *bq)
+{
+ struct blk_plug *plug = current->plug;
+ if(plug) {
+ /* A thread may sleep and wait for new buffers from previously submitted requests. */
+ /* Make sure requests are actually sent to backing device, and not just queued. */
+ struct bio_list *blistTmp = current->bio_list;
+ current->bio_list = NULL;
+ blk_finish_plug(plug); /* clears current->plug */
+ current->bio_list = blistTmp;
+ blk_start_plug(plug); /* sets current->plug */
+ }
+}
+
+#if LINUX_VERSION_CODE >= 0x30200
+static void loop_make_request_err(struct request_queue *q, struct bio *old_bio)
+#else
+static int loop_make_request_err(struct request_queue *q, struct bio *old_bio)
+#endif
+{
+ old_bio->bi_next = NULL;
+ bio_io_error(old_bio);
+#if LINUX_VERSION_CODE >= 0x30200
+ return;
+#else
+ return 0;
+#endif
+}
+
+#if LINUX_VERSION_CODE >= 0x30200
+static void loop_make_request_real(struct request_queue *q, struct bio *old_bio)
+#else
+static int loop_make_request_real(struct request_queue *q, struct bio *old_bio)
+#endif
+{
+ struct bio *new_bio, *merge;
+ struct loop_device *lo = q->queuedata;
+ struct loop_bio_extension *extension;
+ int rw = bio_rw(old_bio), y, x, flushFlag = 0, origHasData, lastVec = 0;
+ char *md;
+ wait_queue_t waitq;
+
+ set_current_state(TASK_RUNNING);
+ origHasData = bio_has_data(old_bio);
+ if (!lo)
+ goto out;
+ if ((rw == WRITE) && (lo->lo_flags & LO_FLAGS_READ_ONLY))
+ goto out;
+ atomic_inc(&lo->lo_pending);
+
+ /*
+ * file backed, queue for loop_thread to handle
+ */
+ if (lo->lo_flags & LO_FLAGS_DO_BMAP) {
+ loop_add_queue_last(lo, old_bio, &lo->lo_bio_que0);
+#if LINUX_VERSION_CODE >= 0x30200
+ return;
+#else
+ return 0;
+#endif
+ }
+
+ /*
+ * device backed, just remap bdev & sector for NONE transfer
+ */
+ if (!lo->lo_encryption) {
+ old_bio->LOOP_COMPAT_BI_SECTOR += lo->lo_offs_sec;
+ old_bio->bi_bdev = lo->lo_device;
+ generic_make_request(old_bio);
+ if (atomic_dec_and_test(&lo->lo_pending))
+ wake_up_interruptible_all(&lo->lo_bio_wait);
+#if LINUX_VERSION_CODE >= 0x30200
+ return;
+#else
+ return 0;
+#endif
+ }
+
+ /*
+ * device backed, start reads and writes now if buffer available
+ */
+ merge = NULL;
+ init_waitqueue_entry(&waitq, current);
+ try_next_old_bio_vec:
+ new_bio = loop_get_buffer(lo, old_bio, &merge, &flushFlag, origHasData, &lastVec);
+ if (!new_bio) {
+ /* wait for buffer to be freed, and try again */
+ spin_lock_irq(&lo->lo_lock);
+ lo->lo_bio_flshCnt = 0;
+ spin_unlock_irq(&lo->lo_lock);
+ loop_unplug_backingdev(lo->lo_backingQueue);
+ add_wait_queue(&lo->lo_buf_wait, &waitq);
+ for (;;) {
+ set_current_state(TASK_UNINTERRUPTIBLE);
+ x = 0;
+ spin_lock_irq(&lo->lo_lock);
+ if (!merge && lo->lo_bio_free1) {
+ /* don't sleep if merge bio is available */
+ x = 1;
+ }
+ if (merge && lo->lo_bio_free0) {
+ /* don't sleep if buffer bio is available */
+ x = 1;
+ }
+ spin_unlock_irq(&lo->lo_lock);
+ if (x)
+ break;
+ schedule();
+ }
+ set_current_state(TASK_RUNNING);
+ remove_wait_queue(&lo->lo_buf_wait, &waitq);
+ goto try_next_old_bio_vec;
+ }
+ if ((rw == WRITE) && origHasData) {
+ extension = new_bio->bi_private;
+ y = extension->bioext_index;
+#if LINUX_VERSION_CODE >= 0x30e00
+ md = kmap(old_bio->bi_io_vec[y].bv_page) + old_bio->bi_io_vec[y].bv_offset + extension->bioext_done_offset;
+#else
+ md = kmap(old_bio->bi_io_vec[y].bv_page) + old_bio->bi_io_vec[y].bv_offset;
+#endif
+ if (lo_do_transfer(lo, WRITE, page_address(new_bio->bi_io_vec[0].bv_page), md, extension->bioext_size, extension->bioext_iv)) {
+ clear_bit(0, &merge->bi_flags);
+ }
+ kunmap(old_bio->bi_io_vec[y].bv_page);
+ }
+
+ x = 0;
+ spin_lock_irq(&lo->lo_lock);
+ if((++lo->lo_bio_flshCnt >= lo->lo_bio_flshMax) || flushFlag) {
+ x = 1;
+ lo->lo_bio_flshCnt = 0;
+ new_bio->bi_rw |= L_BIO_RW_NOIDLE;
+ }
+ spin_unlock_irq(&lo->lo_lock);
+
+ /* A thread may sleep and wait for new buffers from previously submitted requests. */
+ /* Make sure requests are actually sent to backing device, and not just queued. */
+ {
+ struct bio_list *blistTmp = current->bio_list;
+ current->bio_list = NULL;
+ generic_make_request(new_bio);
+ current->bio_list = blistTmp;
+ }
+
+ if (x)
+ loop_unplug_backingdev(lo->lo_backingQueue);
+
+ /* other vecs may need processing too */
+ if (!lastVec)
+ goto try_next_old_bio_vec;
+#if LINUX_VERSION_CODE >= 0x30200
+ return;
+#else
+ return 0;
+#endif
+
+out:
+ old_bio->bi_next = NULL;
+ bio_io_error(old_bio);
+#if LINUX_VERSION_CODE >= 0x30200
+ return;
+#else
+ return 0;
+#endif
+}
+
+struct loop_switch_request {
+ struct file *file;
+ struct completion wait;
+};
+
+static void do_loop_switch(struct loop_device *lo, struct loop_switch_request *p)
+{
+ struct file *file = p->file;
+ struct file *old_file=lo->lo_backing_file;
+ struct address_space *mapping = file->f_path.dentry->d_inode->i_mapping;
+
+ /* This code runs on file backed loop only */
+ /* no need to worry about -1 old_gfp_mask */
+ mapping_set_gfp_mask(old_file->f_path.dentry->d_inode->i_mapping, lo->old_gfp_mask);
+ lo->lo_backing_file = file;
+ memset(lo->lo_file_name, 0, LO_NAME_SIZE);
+ lo->old_gfp_mask = mapping_gfp_mask(mapping);
+ mapping_set_gfp_mask(mapping, (lo->old_gfp_mask & ~(__GFP_IO|__GFP_FS)) | __GFP_HIGH);
+ complete(&p->wait);
+}
+
+/*
+ * worker thread that handles reads/writes to file backed loop devices,
+ * to avoid blocking in our make_request_fn. it also does loop decrypting
+ * on reads for block backed loop, as that is too heavy to do from
+ * b_end_io context where irqs may be disabled.
+ */
+static int loop_thread(void *data)
+{
+ struct loop_device *lo = data;
+ struct bio *bio, *xbio, *merge;
+ struct loop_bio_extension *extension;
+ int x = 0, y;
+ wait_queue_t waitq;
+ char *md;
+ static const struct rlimit loop_rlim_defaults[RLIM_NLIMITS] = INIT_RLIMITS;
+#ifdef CONFIG_BLK_DEV_LOOP_KEYSCRUB
+ void (*keyscrubFn)(void *) = 0;
+#endif
+
+ init_waitqueue_entry(&waitq, current);
+ memcpy(&current->signal->rlim[0], &loop_rlim_defaults[0], sizeof(current->signal->rlim));
+
+ /*
+ * loop can be used in an encrypted device,
+ * hence, it mustn't be stopped at all
+ * because it could be indirectly used during suspension
+ */
+ current->flags |= PF_NOFREEZE;
+ current->flags |= PF_LESS_THROTTLE;
+
+ if (lo_nice > 0)
+ lo_nice = 0;
+ if (lo_nice < -20)
+ lo_nice = -20;
+ set_user_nice(current, lo_nice);
+
+ atomic_inc(&lo->lo_pending);
+
+ /*
+ * up sem, we are running
+ */
+ complete(&lo->lo_done);
+
+ for (;;) {
+ add_wait_queue(&lo->lo_bio_wait, &waitq);
+ for (;;) {
+ set_current_state(TASK_INTERRUPTIBLE);
+ if (!atomic_read(&lo->lo_pending))
+ break;
+
+ x = 0;
+ spin_lock_irq(&lo->lo_lock);
+#ifdef CONFIG_BLK_DEV_LOOP_KEYSCRUB
+ if((keyscrubFn = lo->lo_keyscrub_fn) != 0) {
+ lo->lo_keyscrub_fn = 0;
+ x = 1;
+ }
+#endif
+ if (lo->lo_bio_que0) {
+ /* don't sleep if device backed READ needs processing */
+ /* don't sleep if file backed READ/WRITE needs processing */
+ x = 1;
+ }
+ spin_unlock_irq(&lo->lo_lock);
+ if (x)
+ break;
+
+ schedule();
+ }
+ set_current_state(TASK_RUNNING);
+ remove_wait_queue(&lo->lo_bio_wait, &waitq);
+
+#ifdef CONFIG_BLK_DEV_LOOP_KEYSCRUB
+ if(keyscrubFn) {
+ (*keyscrubFn)(lo->lo_keyscrub_ptr);
+ keyscrubFn = 0;
+ }
+#endif
+ /*
+ * could be woken because of tear-down, not because of
+ * pending work
+ */
+ if (!atomic_read(&lo->lo_pending))
+ break;
+
+ bio = loop_get_bio(lo);
+ if (!bio)
+ continue;
+
+ if (lo->lo_flags & LO_FLAGS_DO_BMAP) {
+ /* request is for file backed device */
+ if(unlikely(!bio->bi_bdev)) {
+ do_loop_switch(lo, bio->bi_private);
+ bio->bi_next = NULL;
+ bio_put(bio);
+ } else {
+ y = do_bio_filebacked(lo, bio);
+ bio->bi_next = NULL;
+ bio_endio(bio, y);
+ }
+ } else {
+ /* device backed read has completed, do decrypt now */
+ extension = bio->bi_private;
+ merge = extension->bioext_merge;
+ y = extension->bioext_index;
+ xbio = merge->bi_private;
+ if(extension->bioext_size) {
+#if LINUX_VERSION_CODE >= 0x30e00
+ md = kmap(xbio->bi_io_vec[y].bv_page) + xbio->bi_io_vec[y].bv_offset + extension->bioext_done_offset;
+#else
+ md = kmap(xbio->bi_io_vec[y].bv_page) + xbio->bi_io_vec[y].bv_offset;
+#endif
+ if (lo_do_transfer(lo, READ, page_address(bio->bi_io_vec[0].bv_page), md, extension->bioext_size, extension->bioext_iv)) {
+ clear_bit(0, &merge->bi_flags);
+ }
+ flush_dcache_page(xbio->bi_io_vec[y].bv_page);
+ kunmap(xbio->bi_io_vec[y].bv_page);
+ }
+ loop_put_buffer(lo, bio, 0);
+ if (!atomic_dec_and_test(&merge->bi_cnt))
+ continue;
+ xbio->bi_next = NULL;
+ bio_endio(xbio, test_bit(0, &merge->bi_flags) ? (int)merge->LOOP_COMPAT_BI_SIZE : -EIO);
+ loop_put_buffer(lo, merge, 1);
+ }
+
+ /*
+ * woken both for pending work and tear-down, lo_pending
+ * will hit zero then
+ */
+ if (atomic_dec_and_test(&lo->lo_pending))
+ break;
+ }
+
+ complete(&lo->lo_done);
+ return 0;
+}
+
+static void loop_set_softblksz(struct loop_device *lo, struct block_device *bdev)
+{
+ int bs, x;
+
+ if (lo->lo_device)
+ bs = block_size(lo->lo_device);
+ else
+ bs = PAGE_SIZE;
+ if (lo->lo_flags & LO_FLAGS_DO_BMAP) {
+ x = (int) bdev->bd_inode->i_size;
+ if ((bs == 8192) && (x & 0x1E00))
+ bs = 4096;
+ if ((bs == 4096) && (x & 0x0E00))
+ bs = 2048;
+ if ((bs == 2048) && (x & 0x0600))
+ bs = 1024;
+ if ((bs == 1024) && (x & 0x0200))
+ bs = 512;
+ }
+ set_blocksize(bdev, bs);
+}
+
+/*
+ * loop_change_fd switches the backing store of a loopback device to a
+ * new file. This is useful for operating system installers to free up the
+ * original file and in High Availability environments to switch to an
+ * alternative location for the content in case of server meltdown.
+ * This can only work if the loop device is used read-only, file backed,
+ * and if the new backing store is the same size and type as the old
+ * backing store.
+ */
+static int loop_change_fd(struct loop_device *lo, unsigned int arg)
+{
+ struct file *file, *old_file;
+ struct inode *inode;
+ struct loop_switch_request w;
+ struct bio *bio;
+ int error;
+
+ error = -EINVAL;
+ /* loop must be read-only */
+ if (!(lo->lo_flags & LO_FLAGS_READ_ONLY))
+ goto out;
+
+ /* loop must be file backed */
+ if (!(lo->lo_flags & LO_FLAGS_DO_BMAP))
+ goto out;
+
+ error = -EBADF;
+ file = fget(arg);
+ if (!file)
+ goto out;
+
+ inode = file->f_path.dentry->d_inode;
+ old_file = lo->lo_backing_file;
+
+ error = -EINVAL;
+ /* new backing store must be file backed */
+ if (!S_ISREG(inode->i_mode))
+ goto out_putf;
+
+ /* new backing store must support reads */
+ if (!file->f_op || !file->f_op->read)
+ goto out_putf;
+
+ /* new backing store must be same size as the old one */
+ if(i_size_read(inode) != i_size_read(old_file->f_path.dentry->d_inode))
+ goto out_putf;
+
+ /* loop must be in properly initialized state */
+ if(lo->lo_queue->make_request_fn != loop_make_request_real)
+ goto out_putf;
+
+ error = -ENOMEM;
+ bio = bio_alloc(GFP_KERNEL, 1);
+ if (!bio)
+ goto out_putf;
+
+ /* wait for loop thread to do the switch */
+ init_completion(&w.wait);
+ w.file = file;
+ bio->bi_private = &w;
+ bio->bi_bdev = NULL;
+ bio->bi_rw = 0;
+ loop_make_request_real(lo->lo_queue, bio);
+ wait_for_completion(&w.wait);
+
+ fput(old_file);
+ return 0;
+
+out_putf:
+ fput(file);
+out:
+ return error;
+}
+
+static int loop_get_threads_count(struct loop_device *lo)
+{
+ int x, y;
+
+ if (lo->lo_flags & LO_FLAGS_DO_BMAP) {
+ /* file backed has only 1 pre-allocated page, so limit to 1 helper thread */
+ return 1;
+ }
+
+ y = lo_threads[0];
+ for (x = 1; x < (sizeof(lo_threads) / sizeof(int)); x += 2) {
+ if (lo_threads[x + 1] && (lo->lo_number == lo_threads[x])) {
+ y = lo_threads[x + 1];
+ break;
+ }
+ }
+ return y;
+}
+
+#if defined(LOOP_HAVE_CONGESTED_FN)
+static int loop_congested(void *data, int bits)
+{
+ struct loop_device *lo = data;
+ struct bio *bio;
+ int ret = 0;
+ unsigned long flags;
+ const int cong = (1 << BDI_sync_congested) | (1 << BDI_async_congested);
+
+ if(lo && lo->lo_backingQueue) {
+ /* check if backing device is congested */
+ ret |= bdi_congested(&lo->lo_backingQueue->backing_dev_info, bits);
+ /* check if loop device is low on resources */
+ spin_lock_irqsave(&lo->lo_lock, flags);
+ bio = lo->lo_bio_free0;
+ if(!bio || !bio->bi_next) {
+ ret |= cong;
+ }
+ bio = lo->lo_bio_free1;
+ if(!bio || !bio->bi_next) {
+ ret |= cong;
+ }
+ spin_unlock_irqrestore(&lo->lo_lock, flags);
+ }
+ return (ret & bits);
+}
+#endif
+
+static int loop_set_fd(struct loop_device *lo, unsigned int ldom,
+ struct block_device *bdev, unsigned int arg)
+{
+ struct file *file;
+ struct inode *inode;
+ struct block_device *lo_device = NULL;
+ int lo_flags = 0;
+ int error;
+ int x, y;
+ struct task_struct *t[LO_THREADS_MAX];
+
+ error = -EBADF;
+ file = fget(arg);
+ if (!file)
+ goto out;
+
+ error = -EINVAL;
+ inode = file->f_path.dentry->d_inode;
+
+ if (!(file->f_mode & FMODE_WRITE))
+ lo_flags |= LO_FLAGS_READ_ONLY;
+
+ init_completion(&lo->lo_done);
+ spin_lock_init(&lo->lo_lock);
+ init_waitqueue_head(&lo->lo_bio_wait);
+ init_waitqueue_head(&lo->lo_buf_wait);
+ atomic_set(&lo->lo_pending, 0);
+#ifdef CONFIG_BLK_DEV_LOOP_KEYSCRUB
+ lo->lo_keyscrub_fn = 0;
+#endif
+ lo->lo_offset = lo->lo_sizelimit = 0;
+ lo->lo_offs_sec = lo->lo_iv_remove = 0;
+ lo->lo_encryption = NULL;
+ lo->lo_encrypt_key_size = 0;
+ lo->transfer = NULL;
+ lo->lo_crypt_name[0] = 0;
+ lo->lo_file_name[0] = 0;
+ lo->lo_init[1] = lo->lo_init[0] = 0;
+#if LINUX_VERSION_CODE >= 0x30600
+ lo->lo_key_owner = GLOBAL_ROOT_UID;
+#else
+ lo->lo_key_owner = 0;
+#endif
+ lo->ioctl = NULL;
+ lo->key_data = NULL;
+ lo->lo_bio_que0 = NULL;
+ lo->lo_bio_free1 = lo->lo_bio_free0 = NULL;
+ lo->lo_bio_flshMax = lo->lo_bio_flshCnt = 0;
+
+ if (S_ISBLK(inode->i_mode)) {
+ lo_device = inode->i_bdev;
+ if (lo_device == bdev) {
+ error = -EBUSY;
+ goto out_putf;
+ }
+ if (loop_prealloc_init(lo, 0)) {
+ error = -ENOMEM;
+ goto out_putf;
+ }
+ if (bdev_read_only(lo_device))
+ lo_flags |= LO_FLAGS_READ_ONLY;
+ else
+ filemap_fdatawrite(inode->i_mapping);
+ } else if (S_ISREG(inode->i_mode)) {
+ /*
+ * If we can't read - sorry. If we only can't write - well,
+ * it's going to be read-only.
+ */
+ if (!file->f_op || !file->f_op->read)
+ goto out_putf;
+
+ if (!file->f_op->write)
+ lo_flags |= LO_FLAGS_READ_ONLY;
+
+ lo_flags |= LO_FLAGS_DO_BMAP;
+ if (loop_prealloc_init(lo, 1)) {
+ error = -ENOMEM;
+ goto out_putf;
+ }
+ } else
+ goto out_putf;
+
+ get_file(file);
+
+ if (!(ldom & FMODE_WRITE))
+ lo_flags |= LO_FLAGS_READ_ONLY;
+
+ set_device_ro(bdev, (lo_flags & LO_FLAGS_READ_ONLY) != 0);
+
+ lo->lo_device = lo_device;
+ lo->lo_flags = lo_flags;
+ if(lo_flags & LO_FLAGS_READ_ONLY)
+ lo->lo_flags |= 0x200000; /* export to user space */
+ lo->lo_backing_file = file;
+ if (figure_loop_size(lo, bdev)) {
+ error = -EFBIG;
+ goto out_cleanup;
+ }
+
+ /*
+ * set queue make_request_fn, and add limits based on lower level
+ * device
+ */
+ blk_queue_make_request(lo->lo_queue, loop_make_request_err);
+ blk_queue_bounce_limit(lo->lo_queue, BLK_BOUNCE_ANY);
+ blk_queue_max_segment_size(lo->lo_queue, PAGE_CACHE_SIZE);
+ blk_queue_segment_boundary(lo->lo_queue, PAGE_CACHE_SIZE - 1);
+ blk_queue_max_segments(lo->lo_queue, BLK_MAX_SEGMENTS);
+ blk_queue_max_hw_sectors(lo->lo_queue, BLK_DEF_MAX_SECTORS);
+ lo->lo_queue->limits.cluster = 0;
+ blk_queue_flush(lo->lo_queue, 0);
+ lo->lo_backingQueue = 0;
+
+ /*
+ * we remap to a block device, make sure we correctly stack limits
+ */
+ if (S_ISBLK(inode->i_mode) && lo_device) {
+ struct request_queue *q = bdev_get_queue(lo_device);
+
+ blk_queue_logical_block_size(lo->lo_queue, queue_logical_block_size(q));
+ blk_queue_flush(lo->lo_queue, q->flush_flags & (REQ_FLUSH | REQ_FUA));
+ lo->lo_queue->limits.io_min = q->limits.io_min;
+ if(lo->lo_queue->limits.io_min > (BLK_MAX_SEGMENTS * PAGE_CACHE_SIZE))
+ lo->lo_queue->limits.io_min = (BLK_MAX_SEGMENTS * PAGE_CACHE_SIZE);
+ lo->lo_queue->limits.io_opt = q->limits.io_opt;
+ if(lo->lo_queue->limits.io_opt > (BLK_MAX_SEGMENTS * PAGE_CACHE_SIZE))
+ lo->lo_queue->limits.io_opt = (BLK_MAX_SEGMENTS * PAGE_CACHE_SIZE);
+ lo->lo_backingQueue = q;
+ }
+
+ if (lo_flags & LO_FLAGS_DO_BMAP) {
+ lo->old_gfp_mask = mapping_gfp_mask(inode->i_mapping);
+ mapping_set_gfp_mask(inode->i_mapping, (lo->old_gfp_mask & ~(__GFP_IO|__GFP_FS)) | __GFP_HIGH);
+ } else {
+ lo->old_gfp_mask = -1;
+ }
+
+ loop_set_softblksz(lo, bdev);
+
+ y = loop_get_threads_count(lo);
+ for(x = 0; x < y; x++) {
+ if(y > 1) {
+ t[x] = kthread_create(loop_thread, lo, "loop%d%c", lo->lo_number, x + 'a');
+ } else {
+ t[x] = kthread_create(loop_thread, lo, "loop%d", lo->lo_number);
+ }
+ if (IS_ERR(t[x])) {
+ error = PTR_ERR(t[x]);
+ while(--x >= 0) {
+ kthread_stop(t[x]);
+ }
+ goto out_mapping;
+ }
+ }
+ for(x = 0; x < y; x++) {
+ wake_up_process(t[x]);
+ wait_for_completion(&lo->lo_done);
+ }
+
+ fput(file);
+#if defined(LOOP_HAVE_CONGESTED_FN)
+ lo->lo_queue->backing_dev_info.congested_data = lo;
+ lo->lo_queue->backing_dev_info.congested_fn = loop_congested;
+#endif
+ wmb();
+ lo->lo_queue->queuedata = lo;
+ __module_get(THIS_MODULE);
+ return 0;
+
+ out_mapping:
+ if(lo->old_gfp_mask != -1)
+ mapping_set_gfp_mask(inode->i_mapping, lo->old_gfp_mask);
+ out_cleanup:
+ loop_prealloc_cleanup(lo);
+ fput(file);
+ out_putf:
+ fput(file);
+ out:
+ return error;
+}
+
+static int loop_release_xfer(struct loop_device *lo)
+{
+ int err = 0;
+ struct loop_func_table *xfer = lo->lo_encryption;
+
+ if (xfer) {
+ lo->transfer = NULL;
+ if (xfer->release)
+ err = xfer->release(lo);
+ lo->lo_encryption = NULL;
+ module_put(xfer->owner);
+ }
+ return err;
+}
+
+static int loop_init_xfer(struct loop_device *lo, struct loop_func_table *xfer, struct loop_info64 *i)
+{
+ int err = 0;
+
+ if (xfer) {
+ struct module *owner = xfer->owner;
+
+ if(!try_module_get(owner))
+ return -EINVAL;
+ if (xfer->init)
+ err = xfer->init(lo, i);
+ if (err)
+ module_put(owner);
+ else
+ lo->lo_encryption = xfer;
+ }
+ return err;
+}
+
+static int loop_clr_fd(struct loop_device *lo, struct block_device *bdev)
+{
+ struct file *filp = lo->lo_backing_file;
+ int gfp = lo->old_gfp_mask;
+ int bdocnt, x, y;
+
+ /* sync /dev/loop? device */
+ sync_blockdev(bdev);
+ /* sync backing /dev/hda? device */
+ sync_blockdev(lo->lo_device);
+
+ for(x = 0; x < 20; x++) {
+ spin_lock(&lo->lo_ioctl_spin);
+ bdocnt = lo->lo_refcnt;
+ spin_unlock(&lo->lo_ioctl_spin);
+ if(bdocnt == 1) break;
+ /* work around reference count race */
+ msleep(50);
+ }
+
+ if (bdocnt != 1) /* one for this fd being open */
+ return -EBUSY;
+ if (filp==NULL)
+ return -EINVAL;
+
+ lo->lo_queue->queuedata = NULL;
+ lo->lo_queue->make_request_fn = loop_make_request_err;
+ lo->lo_backingQueue = 0;
+ y = loop_get_threads_count(lo);
+ for(x = 0; x < y; x++) {
+ if (atomic_dec_and_test(&lo->lo_pending))
+ wake_up_interruptible_all(&lo->lo_bio_wait);
+ }
+ for(x = 0; x < y; x++) {
+ wait_for_completion(&lo->lo_done);
+ }
+ blk_queue_flush(lo->lo_queue, 0);
+ loop_prealloc_cleanup(lo);
+ lo->lo_backing_file = NULL;
+ loop_release_xfer(lo);
+ lo->transfer = NULL;
+ lo->ioctl = NULL;
+ lo->lo_device = NULL;
+ lo->lo_encryption = NULL;
+#ifdef CONFIG_BLK_DEV_LOOP_KEYSCRUB
+ lo->lo_keyscrub_fn = 0;
+#endif
+ lo->lo_offset = lo->lo_sizelimit = 0;
+ lo->lo_offs_sec = lo->lo_iv_remove = 0;
+ lo->lo_encrypt_key_size = 0;
+ lo->lo_flags = 0;
+ lo->lo_init[1] = lo->lo_init[0] = 0;
+#if LINUX_VERSION_CODE >= 0x30600
+ lo->lo_key_owner = GLOBAL_ROOT_UID;
+#else
+ lo->lo_key_owner = 0;
+#endif
+ lo->key_data = NULL;
+ memset(lo->lo_encrypt_key, 0, LO_KEY_SIZE);
+ memset(lo->lo_crypt_name, 0, LO_NAME_SIZE);
+ memset(lo->lo_file_name, 0, LO_NAME_SIZE);
+ invalidate_bdev(bdev);
+ set_capacity(disks[lo->lo_number], 0);
+ if (gfp != -1)
+ mapping_set_gfp_mask(filp->f_path.dentry->d_inode->i_mapping, gfp);
+ fput(filp);
+ module_put(THIS_MODULE);
+ return 0;
+}
+
+static int loop_set_status(struct loop_device *lo, struct block_device *bdev, struct loop_info64 *info)
+{
+ int err;
+ struct loop_func_table *xfer = NULL;
+#if LINUX_VERSION_CODE >= 0x30600
+ kuid_t uid = current_uid();
+
+ if (lo->lo_encrypt_key_size && !uid_eq(lo->lo_key_owner, uid) && !capable(CAP_SYS_ADMIN))
+ return -EPERM;
+#else
+ uid_t uid = current_uid();
+
+ if (lo->lo_encrypt_key_size && lo->lo_key_owner != uid && !capable(CAP_SYS_ADMIN))
+ return -EPERM;
+#endif
+ if ((unsigned int) info->lo_encrypt_key_size > LO_KEY_SIZE)
+ return -EINVAL;
+
+ err = loop_release_xfer(lo);
+ if (err)
+ return err;
+
+ if ((loff_t)info->lo_offset < 0) {
+ /* negative offset == remove offset from IV computations */
+ lo->lo_offset = -(info->lo_offset);
+ lo->lo_iv_remove = lo->lo_offset >> 9;
+ } else {
+ /* positive offset == include offset in IV computations */
+ lo->lo_offset = info->lo_offset;
+ lo->lo_iv_remove = 0;
+ }
+ lo->lo_offs_sec = lo->lo_offset >> 9;
+ lo->lo_sizelimit = info->lo_sizelimit;
+ err = figure_loop_size(lo, bdev);
+ if (err)
+ return err;
+ loop_set_softblksz(lo, bdev);
+
+ if (info->lo_encrypt_type) {
+ unsigned int type = info->lo_encrypt_type;
+
+ if (type >= MAX_LO_CRYPT)
+ return -EINVAL;
+ xfer = xfer_funcs[type];
+ if (xfer == NULL)
+ return -EINVAL;
+ } else if(!(lo->lo_flags & LO_FLAGS_DO_BMAP)) {
+ blk_queue_max_hw_sectors(lo->lo_queue, PAGE_CACHE_SIZE >> 9);
+ }
+ err = loop_init_xfer(lo, xfer, info);
+ if (err)
+ return err;
+
+ if (!xfer)
+ xfer = &none_funcs;
+ lo->transfer = xfer->transfer;
+ lo->ioctl = xfer->ioctl;
+
+ memcpy(lo->lo_file_name, info->lo_file_name, LO_NAME_SIZE);
+ memcpy(lo->lo_crypt_name, info->lo_crypt_name, LO_NAME_SIZE);
+ lo->lo_file_name[LO_NAME_SIZE-1] = 0;
+ lo->lo_crypt_name[LO_NAME_SIZE-1] = 0;
+ lo->lo_encrypt_key_size = info->lo_encrypt_key_size;
+ lo->lo_init[0] = info->lo_init[0];
+ lo->lo_init[1] = info->lo_init[1];
+ if (info->lo_encrypt_key_size) {
+ memcpy(lo->lo_encrypt_key, info->lo_encrypt_key,
+ info->lo_encrypt_key_size);
+ lo->lo_key_owner = uid;
+ }
+
+ lo->lo_queue->make_request_fn = loop_make_request_real;
+ return 0;
+}
+
+static int loop_get_status(struct loop_device *lo, struct loop_info64 *info)
+{
+ struct file *file = lo->lo_backing_file;
+ struct kstat stat;
+ int error;
+
+#if LINUX_VERSION_CODE >= 0x30900
+ error = vfs_getattr(&file->f_path, &stat);
+#else
+ error = vfs_getattr(file->f_vfsmnt, file->f_path.dentry, &stat);
+#endif
+ if (error)
+ return error;
+ memset(info, 0, sizeof(*info));
+ info->lo_number = lo->lo_number;
+ info->lo_device = huge_encode_dev(stat.dev);
+ info->lo_inode = stat.ino;
+ info->lo_rdevice = huge_encode_dev(lo->lo_device ? stat.rdev : stat.dev);
+ info->lo_offset = lo->lo_iv_remove ? -(lo->lo_offset) : lo->lo_offset;
+ info->lo_sizelimit = lo->lo_sizelimit;
+ info->lo_flags = lo->lo_flags;
+ memcpy(info->lo_file_name, lo->lo_file_name, LO_NAME_SIZE);
+ memcpy(info->lo_crypt_name, lo->lo_crypt_name, LO_NAME_SIZE);
+ info->lo_encrypt_type = lo->lo_encryption ? lo->lo_encryption->number : 0;
+ if (lo->lo_encrypt_key_size && capable(CAP_SYS_ADMIN)) {
+ info->lo_encrypt_key_size = lo->lo_encrypt_key_size;
+ memcpy(info->lo_encrypt_key, lo->lo_encrypt_key,
+ lo->lo_encrypt_key_size);
+ info->lo_init[0] = lo->lo_init[0];
+ info->lo_init[1] = lo->lo_init[1];
+ }
+ return 0;
+}
+
+static void
+loop_info64_from_old(const struct loop_info *info, struct loop_info64 *info64)
+{
+ memset(info64, 0, sizeof(*info64));
+ info64->lo_number = info->lo_number;
+ info64->lo_device = info->lo_device;
+ info64->lo_inode = info->lo_inode;
+ info64->lo_rdevice = info->lo_rdevice;
+ info64->lo_offset = info->lo_offset;
+ info64->lo_encrypt_type = info->lo_encrypt_type;
+ info64->lo_encrypt_key_size = info->lo_encrypt_key_size;
+ info64->lo_flags = info->lo_flags;
+ info64->lo_init[0] = info->lo_init[0];
+ info64->lo_init[1] = info->lo_init[1];
+ if (info->lo_encrypt_type == LO_CRYPT_CRYPTOAPI)
+ memcpy(info64->lo_crypt_name, info->lo_name, LO_NAME_SIZE);
+ else
+ memcpy(info64->lo_file_name, info->lo_name, LO_NAME_SIZE);
+ memcpy(info64->lo_encrypt_key, info->lo_encrypt_key, LO_KEY_SIZE);
+}
+
+static int
+loop_info64_to_old(struct loop_info64 *info64, struct loop_info *info)
+{
+ memset(info, 0, sizeof(*info));
+ info->lo_number = info64->lo_number;
+ info->lo_device = info64->lo_device;
+ info->lo_inode = info64->lo_inode;
+ info->lo_rdevice = info64->lo_rdevice;
+ info->lo_offset = info64->lo_offset;
+ info->lo_encrypt_type = info64->lo_encrypt_type;
+ info->lo_encrypt_key_size = info64->lo_encrypt_key_size;
+ info->lo_flags = info64->lo_flags;
+ info->lo_init[0] = info64->lo_init[0];
+ info->lo_init[1] = info64->lo_init[1];
+ if (info->lo_encrypt_type == LO_CRYPT_CRYPTOAPI)
+ memcpy(info->lo_name, info64->lo_crypt_name, LO_NAME_SIZE);
+ else
+ memcpy(info->lo_name, info64->lo_file_name, LO_NAME_SIZE);
+ memcpy(info->lo_encrypt_key, info64->lo_encrypt_key, LO_KEY_SIZE);
+
+ /* error in case values were truncated */
+ if (info->lo_device != info64->lo_device ||
+ info->lo_rdevice != info64->lo_rdevice ||
+ info->lo_inode != info64->lo_inode ||
+ info->lo_offset != info64->lo_offset ||
+ info64->lo_sizelimit)
+ return -EOVERFLOW;
+
+ return 0;
+}
+
+static int
+loop_set_status_old(struct loop_device *lo, struct block_device *bdev, const struct loop_info *arg)
+{
+ struct loop_info info;
+ struct loop_info64 info64;
+
+ if (copy_from_user(&info, arg, sizeof (struct loop_info)))
+ return -EFAULT;
+ loop_info64_from_old(&info, &info64);
+ memset(&info.lo_encrypt_key[0], 0, sizeof(info.lo_encrypt_key));
+ return loop_set_status(lo, bdev, &info64);
+}
+
+static int
+loop_set_status64(struct loop_device *lo, struct block_device *bdev, struct loop_info64 *arg)
+{
+ struct loop_info64 info64;
+
+ if (copy_from_user(&info64, arg, sizeof (struct loop_info64)))
+ return -EFAULT;
+ return loop_set_status(lo, bdev, &info64);
+}
+
+static int
+loop_get_status_old(struct loop_device *lo, struct loop_info *arg) {
+ struct loop_info info;
+ struct loop_info64 info64;
+ int err = 0;
+
+ if (!arg)
+ err = -EINVAL;
+ if (!err)
+ err = loop_get_status(lo, &info64);
+ if (!err)
+ err = loop_info64_to_old(&info64, &info);
+ if (!err && copy_to_user(arg, &info, sizeof(info)))
+ err = -EFAULT;
+
+ return err;
+}
+
+static int
+loop_get_status64(struct loop_device *lo, struct loop_info64 *arg) {
+ struct loop_info64 info64;
+ int err = 0;
+
+ if (!arg)
+ err = -EINVAL;
+ if (!err)
+ err = loop_get_status(lo, &info64);
+ if (!err && copy_to_user(arg, &info64, sizeof(info64)))
+ err = -EFAULT;
+
+ return err;
+}
+
+static int lo_ioctl(struct block_device *bdev, fmode_t ldom, unsigned int cmd, unsigned long arg)
+{
+ struct loop_device *lo = bdev->bd_disk->private_data;
+ int err;
+ wait_queue_t waitq;
+
+ /*
+ * mutual exclusion - lock
+ */
+ init_waitqueue_entry(&waitq, current);
+ add_wait_queue(&lo->lo_ioctl_wait, &waitq);
+ for (;;) {
+ set_current_state(TASK_UNINTERRUPTIBLE);
+ spin_lock(&lo->lo_ioctl_spin);
+ err = lo->lo_ioctl_busy;
+ if(!err) lo->lo_ioctl_busy = 1;
+ spin_unlock(&lo->lo_ioctl_spin);
+ if(!err) break;
+ schedule();
+ }
+ set_current_state(TASK_RUNNING);
+ remove_wait_queue(&lo->lo_ioctl_wait, &waitq);
+
+ /*
+ * LOOP_SET_FD can only be called when no device is attached.
+ * All other ioctls can only be called when a device is attached.
+ */
+ if (bdev->bd_disk->queue->queuedata != NULL) {
+ if (cmd == LOOP_SET_FD) {
+ err = -EBUSY;
+ goto out_err;
+ }
+ } else {
+ if (cmd != LOOP_SET_FD) {
+ err = -ENXIO;
+ goto out_err;
+ }
+ }
+
+ switch (cmd) {
+ case LOOP_SET_FD:
+ err = loop_set_fd(lo, ldom, bdev, arg);
+ break;
+ case LOOP_CHANGE_FD:
+ err = loop_change_fd(lo, arg);
+ break;
+ case LOOP_CLR_FD:
+ err = loop_clr_fd(lo, bdev);
+ break;
+ case LOOP_SET_STATUS:
+ err = loop_set_status_old(lo, bdev, (struct loop_info *) arg);
+ break;
+ case LOOP_GET_STATUS:
+ err = loop_get_status_old(lo, (struct loop_info *) arg);
+ break;
+ case LOOP_SET_STATUS64:
+ err = loop_set_status64(lo, bdev, (struct loop_info64 *) arg);
+ break;
+ case LOOP_GET_STATUS64:
+ err = loop_get_status64(lo, (struct loop_info64 *) arg);
+ break;
+ case LOOP_RECOMPUTE_DEV_SIZE:
+ err = figure_loop_size(lo, bdev);
+ break;
+ default:
+ err = lo->ioctl ? lo->ioctl(lo, cmd, arg) : -EINVAL;
+ }
+out_err:
+ /*
+ * mutual exclusion - unlock
+ */
+ spin_lock(&lo->lo_ioctl_spin);
+ lo->lo_ioctl_busy = 0;
+ spin_unlock(&lo->lo_ioctl_spin);
+ wake_up_all(&lo->lo_ioctl_wait);
+
+ return err;
+}
+
+#if defined(CONFIG_COMPAT) && defined(HAVE_COMPAT_IOCTL)
+struct loop_info32 {
+ compat_int_t lo_number; /* ioctl r/o */
+ compat_dev_t lo_device; /* ioctl r/o */
+ compat_ulong_t lo_inode; /* ioctl r/o */
+ compat_dev_t lo_rdevice; /* ioctl r/o */
+ compat_int_t lo_offset;
+ compat_int_t lo_encrypt_type;
+ compat_int_t lo_encrypt_key_size; /* ioctl w/o */
+ compat_int_t lo_flags; /* ioctl r/o */
+ char lo_name[LO_NAME_SIZE];
+ unsigned char lo_encrypt_key[LO_KEY_SIZE]; /* ioctl w/o */
+ compat_ulong_t lo_init[2];
+ char reserved[4];
+};
+
+static int lo_compat_ioctl(struct block_device *p1, fmode_t p2, unsigned int cmd, unsigned long arg)
+{
+ mm_segment_t old_fs = get_fs();
+ struct loop_info l;
+ struct loop_info32 *ul = (struct loop_info32 *)arg;
+ int err = -ENOIOCTLCMD;
+
+ switch (cmd) {
+ case LOOP_SET_FD:
+ case LOOP_CLR_FD:
+ case LOOP_SET_STATUS64:
+ case LOOP_GET_STATUS64:
+ case LOOP_CHANGE_FD:
+ case LOOP_MULTI_KEY_SETUP:
+ case LOOP_MULTI_KEY_SETUP_V3:
+ case LOOP_RECOMPUTE_DEV_SIZE:
+ err = lo_ioctl(p1, p2, cmd, arg);
+ break;
+ case LOOP_SET_STATUS:
+ memset(&l, 0, sizeof(l));
+ err = get_user(l.lo_number, &ul->lo_number);
+ err |= get_user(l.lo_device, &ul->lo_device);
+ err |= get_user(l.lo_inode, &ul->lo_inode);
+ err |= get_user(l.lo_rdevice, &ul->lo_rdevice);
+ err |= copy_from_user(&l.lo_offset, &ul->lo_offset,
+ 8 + (unsigned long)l.lo_init - (unsigned long)&l.lo_offset);
+ if (err) {
+ err = -EFAULT;
+ } else {
+ set_fs (KERNEL_DS);
+ err = lo_ioctl(p1, p2, cmd, (unsigned long)&l);
+ set_fs (old_fs);
+ }
+ memset(&l, 0, sizeof(l));
+ break;
+ case LOOP_GET_STATUS:
+ set_fs (KERNEL_DS);
+ err = lo_ioctl(p1, p2, cmd, (unsigned long)&l);
+ set_fs (old_fs);
+ if (!err) {
+ err = put_user(l.lo_number, &ul->lo_number);
+ err |= put_user(l.lo_device, &ul->lo_device);
+ err |= put_user(l.lo_inode, &ul->lo_inode);
+ err |= put_user(l.lo_rdevice, &ul->lo_rdevice);
+ err |= copy_to_user(&ul->lo_offset, &l.lo_offset,
+ (unsigned long)l.lo_init - (unsigned long)&l.lo_offset);
+ if (err)
+ err = -EFAULT;
+ }
+ memset(&l, 0, sizeof(l));
+ break;
+
+ }
+ return err;
+}
+#endif
+
+static int lo_open(struct block_device *bdev, fmode_t mode)
+{
+ struct loop_device *lo = bdev->bd_disk->private_data;
+
+ spin_lock(&lo->lo_ioctl_spin);
+ lo->lo_refcnt++;
+ spin_unlock(&lo->lo_ioctl_spin);
+ return 0;
+}
+
+#if LINUX_VERSION_CODE >= 0x30a00
+static void lo_release(struct gendisk *disk, fmode_t mode)
+#else
+static int lo_release(struct gendisk *disk, fmode_t mode)
+#endif
+{
+ struct loop_device *lo = disk->private_data;
+
+ spin_lock(&lo->lo_ioctl_spin);
+ lo->lo_refcnt--;
+ spin_unlock(&lo->lo_ioctl_spin);
+#if LINUX_VERSION_CODE < 0x30a00
+ return 0;
+#endif
+}
+
+static struct block_device_operations lo_fops = {
+ .owner = THIS_MODULE,
+ .open = lo_open,
+ .release = lo_release,
+ .ioctl = lo_ioctl,
+#if defined(CONFIG_COMPAT) && defined(HAVE_COMPAT_IOCTL)
+ .compat_ioctl = lo_compat_ioctl,
+#endif
+};
+
+/*
+ * And now the modules code and kernel interface.
+ */
+MODULE_LICENSE("GPL");
+MODULE_ALIAS_BLOCKDEV_MAJOR(LOOP_MAJOR);
+
+int loop_register_transfer(struct loop_func_table *funcs)
+{
+ unsigned int n = funcs->number;
+
+ if (n >= MAX_LO_CRYPT || xfer_funcs[n])
+ return -EINVAL;
+ xfer_funcs[n] = funcs;
+ return 0;
+}
+
+int loop_unregister_transfer(int number)
+{
+ unsigned int n = number;
+ struct loop_device *lo;
+ struct loop_func_table *xfer;
+ int x;
+
+ if (n == 0 || n >= MAX_LO_CRYPT || (xfer = xfer_funcs[n]) == NULL)
+ return -EINVAL;
+ xfer_funcs[n] = NULL;
+ for (x = 0; x < max_loop; x++) {
+ lo = loop_dev_ptr_arr[x];
+ if (!lo)
+ continue;
+ if (lo->lo_encryption == xfer)
+ loop_release_xfer(lo);
+ }
+ return 0;
+}
+
+EXPORT_SYMBOL(loop_register_transfer);
+EXPORT_SYMBOL(loop_unregister_transfer);
+
+int __init loop_init(void)
+{
+ int i;
+
+#ifdef CONFIG_BLK_DEV_LOOP_AES
+#if defined(CONFIG_BLK_DEV_LOOP_PADLOCK) && (defined(CONFIG_X86) || defined(CONFIG_X86_64))
+ if((boot_cpu_data.x86 >= 6) && CentaurHauls_ID_and_enabled_ACE()) {
+ xfer_funcs[LO_CRYPT_AES] = &funcs_padlock_aes;
+ printk(KERN_INFO "loop: padlock hardware AES enabled\n");
+ } else
+#endif
+#if defined(CONFIG_BLK_DEV_LOOP_INTELAES) && (defined(CONFIG_X86) || defined(CONFIG_X86_64))
+ if((boot_cpu_data.x86 >= 6) && ((cpuid_ecx(1) & 0x02000000) == 0x02000000)) {
+ xfer_funcs[LO_CRYPT_AES] = &funcs_intel_aes;
+ printk("loop: Intel hardware AES enabled\n");
+ } else
+#endif
+#endif
+ { } /* needed because of above else statements */
+
+ if ((max_loop < 1) || (max_loop > 256)) {
+ printk(KERN_WARNING "loop: invalid max_loop (must be between"
+ " 1 and 256), using default (8)\n");
+ max_loop = 8;
+ }
+
+ if (register_blkdev(LOOP_MAJOR, "loop"))
+ return -EIO;
+
+ loop_dev_ptr_arr = kmalloc(max_loop * sizeof(struct loop_device *), GFP_KERNEL);
+ if (!loop_dev_ptr_arr)
+ goto out_mem1;
+
+ disks = kmalloc(max_loop * sizeof(struct gendisk *), GFP_KERNEL);
+ if (!disks)
+ goto out_mem2;
+
+ for (i = 0; i < max_loop; i++) {
+ loop_dev_ptr_arr[i] = kmalloc(sizeof(struct loop_device), GFP_KERNEL);
+ if (!loop_dev_ptr_arr[i])
+ goto out_mem3;
+ }
+
+ for (i = 0; i < max_loop; i++) {
+ disks[i] = alloc_disk(1);
+ if (!disks[i])
+ goto out_mem4;
+ }
+
+ for (i = 0; i < max_loop; i++) {
+ disks[i]->queue = blk_alloc_queue(GFP_KERNEL);
+ if (!disks[i]->queue)
+ goto out_mem5;
+ disks[i]->queue->queuedata = NULL;
+ blk_queue_make_request(disks[i]->queue, loop_make_request_err);
+ }
+
+ for (i = 0; i < (sizeof(lo_prealloc) / sizeof(int)); i += 2) {
+ if (!lo_prealloc[i])
+ continue;
+ if (lo_prealloc[i] < LO_PREALLOC_MIN)
+ lo_prealloc[i] = LO_PREALLOC_MIN;
+ if (lo_prealloc[i] > LO_PREALLOC_MAX)
+ lo_prealloc[i] = LO_PREALLOC_MAX;
+ }
+ for (i = 0; i < (sizeof(lo_threads) / sizeof(int)); i += 2) {
+ if (!lo_threads[i])
+ continue;
+ if (lo_threads[i] < LO_THREADS_MIN)
+ lo_threads[i] = LO_THREADS_MIN;
+ if (lo_threads[i] > LO_THREADS_MAX)
+ lo_threads[i] = LO_THREADS_MAX;
+ }
+
+#if defined(IOCTL32_COMPATIBLE_PTR)
+ register_ioctl32_conversion(LOOP_MULTI_KEY_SETUP, IOCTL32_COMPATIBLE_PTR);
+ register_ioctl32_conversion(LOOP_MULTI_KEY_SETUP_V3, IOCTL32_COMPATIBLE_PTR);
+ register_ioctl32_conversion(LOOP_RECOMPUTE_DEV_SIZE, IOCTL32_COMPATIBLE_PTR);
+#endif
+
+#ifdef CONFIG_DEVFS_FS
+ devfs_mk_dir("loop");
+#endif
+
+ for (i = 0; i < max_loop; i++) {
+ struct loop_device *lo = loop_dev_ptr_arr[i];
+ struct gendisk *disk = disks[i];
+ memset(lo, 0, sizeof(struct loop_device));
+ lo->lo_number = i;
+ lo->lo_queue = disk->queue;
+ spin_lock_init(&lo->lo_ioctl_spin);
+ init_waitqueue_head(&lo->lo_ioctl_wait);
+ disk->major = LOOP_MAJOR;
+ disk->first_minor = i;
+ disk->fops = &lo_fops;
+ sprintf(disk->disk_name, "loop%d", i);
+#ifdef CONFIG_DEVFS_FS
+ sprintf(disk->devfs_name, "loop/%d", i);
+#endif
+ disk->private_data = lo;
+ add_disk(disk);
+ }
+
+#ifdef CONFIG_BLK_DEV_LOOP_AES
+#ifdef CONFIG_BLK_DEV_LOOP_KEYSCRUB
+ printk(KERN_INFO "loop: AES key scrubbing enabled\n");
+#endif
+#endif
+ printk(KERN_INFO "loop: loaded (max %d devices)\n", max_loop);
+ return 0;
+
+out_mem5:
+ while (i--)
+ blk_cleanup_queue(disks[i]->queue);
+ i = max_loop;
+out_mem4:
+ while (i--)
+ put_disk(disks[i]);
+ i = max_loop;
+out_mem3:
+ while (i--)
+ kfree(loop_dev_ptr_arr[i]);
+ kfree(disks);
+out_mem2:
+ kfree(loop_dev_ptr_arr);
+out_mem1:
+ unregister_blkdev(LOOP_MAJOR, "loop");
+ printk(KERN_ERR "loop: ran out of memory\n");
+ return -ENOMEM;
+}
+
+void loop_exit(void)
+{
+ int i;
+
+ for (i = 0; i < max_loop; i++) {
+ del_gendisk(disks[i]);
+ put_disk(disks[i]);
+ blk_cleanup_queue(loop_dev_ptr_arr[i]->lo_queue);
+ kfree(loop_dev_ptr_arr[i]);
+ }
+#ifdef CONFIG_DEVFS_FS
+ devfs_remove("loop");
+#endif
+ unregister_blkdev(LOOP_MAJOR, "loop");
+ kfree(disks);
+ kfree(loop_dev_ptr_arr);
+
+#if defined(IOCTL32_COMPATIBLE_PTR)
+ unregister_ioctl32_conversion(LOOP_MULTI_KEY_SETUP);
+ unregister_ioctl32_conversion(LOOP_MULTI_KEY_SETUP_V3);
+ unregister_ioctl32_conversion(LOOP_RECOMPUTE_DEV_SIZE);
+#endif
+}
+
+module_init(loop_init);
+module_exit(loop_exit);
+
+#ifdef CONFIG_BLK_DEV_LOOP_KEYSCRUB
+void loop_add_keyscrub_fn(struct loop_device *lo, void (*fn)(void *), void *ptr)
+{
+ lo->lo_keyscrub_ptr = ptr;
+ wmb();
+ lo->lo_keyscrub_fn = fn;
+ wake_up_interruptible(&lo->lo_bio_wait);
+}
+EXPORT_SYMBOL(loop_add_keyscrub_fn);
+#endif
diff -urN linux-3.14-noloop/drivers/misc/Makefile linux-3.14-AES/drivers/misc/Makefile
--- linux-3.14-noloop/drivers/misc/Makefile 2014-03-31 06:40:15.000000000 +0300
+++ linux-3.14-AES/drivers/misc/Makefile 2014-03-31 22:20:36.000000000 +0300
@@ -2,6 +2,33 @@
# Makefile for misc devices that really don't fit anywhere else.
#
+ifeq ($(CONFIG_BLK_DEV_LOOP_AES),y)
+AES_X86_ASM=n
+ifeq ($(CONFIG_X86),y)
+ifneq ($(CONFIG_X86_64),y)
+ AES_X86_ASM=y
+endif
+endif
+ifeq ($(AES_X86_ASM),y)
+ obj-y += aes-x86.o md5-x86.o crypto-ksym.o
+ AFLAGS_aes-x86.o := -DUSE_UNDERLINE=1
+ifeq ($(CONFIG_BLK_DEV_LOOP_INTELAES),y)
+ obj-y += aes-intel32.o
+endif
+else
+ifeq ($(CONFIG_X86_64),y)
+ obj-y += aes-amd64.o md5-amd64.o md5-2x-amd64.o crypto-ksym.o
+ AFLAGS_aes-amd64.o := -DUSE_UNDERLINE=1
+ifeq ($(CONFIG_BLK_DEV_LOOP_INTELAES),y)
+ obj-y += aes-intel64.o
+endif
+else
+ obj-y += aes.o md5.o crypto-ksym.o
+ CFLAGS_aes.o := -DDATA_ALWAYS_ALIGNED=1
+endif
+endif
+endif
+
obj-$(CONFIG_IBM_ASM) += ibmasm/
obj-$(CONFIG_AD525X_DPOT) += ad525x_dpot.o
obj-$(CONFIG_AD525X_DPOT_I2C) += ad525x_dpot-i2c.o
diff -urN linux-3.14-noloop/drivers/misc/aes-amd64.S linux-3.14-AES/drivers/misc/aes-amd64.S
--- linux-3.14-noloop/drivers/misc/aes-amd64.S 1970-01-01 02:00:00.000000000 +0200
+++ linux-3.14-AES/drivers/misc/aes-amd64.S 2014-03-31 22:20:36.000000000 +0300
@@ -0,0 +1,897 @@
+//
+// Copyright (c) 2001, Dr Brian Gladman <brg@gladman.uk.net>, Worcester, UK.
+// All rights reserved.
+//
+// TERMS
+//
+// Redistribution and use in source and binary forms, with or without
+// modification, are permitted subject to the following conditions:
+//
+// 1. Redistributions of source code must retain the above copyright
+// notice, this list of conditions and the following disclaimer.
+//
+// 2. Redistributions in binary form must reproduce the above copyright
+// notice, this list of conditions and the following disclaimer in the
+// documentation and/or other materials provided with the distribution.
+//
+// 3. The copyright holder's name must not be used to endorse or promote
+// any products derived from this software without his specific prior
+// written permission.
+//
+// This software is provided 'as is' with no express or implied warranties
+// of correctness or fitness for purpose.
+
+// Modified by Jari Ruusu, December 24 2001
+// - Converted syntax to GNU CPP/assembler syntax
+// - C programming interface converted back to "old" API
+// - Minor portability cleanups and speed optimizations
+
+// Modified by Jari Ruusu, April 11 2002
+// - Added above copyright and terms to resulting object code so that
+// binary distributions can avoid legal trouble
+
+// Modified by Jari Ruusu, June 12 2004
+// - Converted 32 bit x86 code to 64 bit AMD64 code
+// - Re-wrote encrypt and decrypt code from scratch
+
+// An AES (Rijndael) implementation for the AMD64. This version only
+// implements the standard AES block length (128 bits, 16 bytes). This code
+// does not preserve the rax, rcx, rdx, rsi, rdi or r8-r11 registers or the
+// artihmetic status flags. However, the rbx, rbp and r12-r15 registers are
+// preserved across calls.
+
+// void aes_set_key(aes_context *cx, const unsigned char key[], const int key_len, const int f)
+// void aes_encrypt(const aes_context *cx, const unsigned char in_blk[], unsigned char out_blk[])
+// void aes_decrypt(const aes_context *cx, const unsigned char in_blk[], unsigned char out_blk[])
+
+#if defined(USE_UNDERLINE)
+# define aes_set_key _aes_set_key
+# define aes_encrypt _aes_encrypt
+# define aes_decrypt _aes_decrypt
+#endif
+#if !defined(ALIGN64BYTES)
+# define ALIGN64BYTES 64
+#endif
+
+ .file "aes-amd64.S"
+ .globl aes_set_key
+ .globl aes_encrypt
+ .globl aes_decrypt
+
+ .section .rodata
+copyright:
+ .ascii " \000"
+ .ascii "Copyright (c) 2001, Dr Brian Gladman <brg@gladman.uk.net>, Worcester, UK.\000"
+ .ascii "All rights reserved.\000"
+ .ascii " \000"
+ .ascii "TERMS\000"
+ .ascii " \000"
+ .ascii " Redistribution and use in source and binary forms, with or without\000"
+ .ascii " modification, are permitted subject to the following conditions:\000"
+ .ascii " \000"
+ .ascii " 1. Redistributions of source code must retain the above copyright\000"
+ .ascii " notice, this list of conditions and the following disclaimer.\000"
+ .ascii " \000"
+ .ascii " 2. Redistributions in binary form must reproduce the above copyright\000"
+ .ascii " notice, this list of conditions and the following disclaimer in the\000"
+ .ascii " documentation and/or other materials provided with the distribution.\000"
+ .ascii " \000"
+ .ascii " 3. The copyright holder's name must not be used to endorse or promote\000"
+ .ascii " any products derived from this software without his specific prior\000"
+ .ascii " written permission.\000"
+ .ascii " \000"
+ .ascii " This software is provided 'as is' with no express or implied warranties\000"
+ .ascii " of correctness or fitness for purpose.\000"
+ .ascii " \000"
+
+#define tlen 1024 // length of each of 4 'xor' arrays (256 32-bit words)
+
+// offsets in context structure
+
+#define nkey 0 // key length, size 4
+#define nrnd 4 // number of rounds, size 4
+#define ekey 8 // encryption key schedule base address, size 256
+#define dkey 264 // decryption key schedule base address, size 256
+
+// This macro performs a forward encryption cycle. It is entered with
+// the first previous round column values in I1E, I2E, I3E and I4E and
+// exits with the final values OU1, OU2, OU3 and OU4 registers.
+
+#define fwd_rnd(p1,p2,I1E,I1B,I1H,I2E,I2B,I2H,I3E,I3B,I3R,I4E,I4B,I4R,OU1,OU2,OU3,OU4) \
+ movl p2(%rbp),OU1 ;\
+ movl p2+4(%rbp),OU2 ;\
+ movl p2+8(%rbp),OU3 ;\
+ movl p2+12(%rbp),OU4 ;\
+ movzbl I1B,%edi ;\
+ movzbl I2B,%esi ;\
+ movzbl I3B,%r8d ;\
+ movzbl I4B,%r13d ;\
+ shrl $8,I3E ;\
+ shrl $8,I4E ;\
+ xorl p1(,%rdi,4),OU1 ;\
+ xorl p1(,%rsi,4),OU2 ;\
+ xorl p1(,%r8,4),OU3 ;\
+ xorl p1(,%r13,4),OU4 ;\
+ movzbl I2H,%esi ;\
+ movzbl I3B,%r8d ;\
+ movzbl I4B,%r13d ;\
+ movzbl I1H,%edi ;\
+ shrl $8,I3E ;\
+ shrl $8,I4E ;\
+ xorl p1+tlen(,%rsi,4),OU1 ;\
+ xorl p1+tlen(,%r8,4),OU2 ;\
+ xorl p1+tlen(,%r13,4),OU3 ;\
+ xorl p1+tlen(,%rdi,4),OU4 ;\
+ shrl $16,I1E ;\
+ shrl $16,I2E ;\
+ movzbl I3B,%r8d ;\
+ movzbl I4B,%r13d ;\
+ movzbl I1B,%edi ;\
+ movzbl I2B,%esi ;\
+ xorl p1+2*tlen(,%r8,4),OU1 ;\
+ xorl p1+2*tlen(,%r13,4),OU2 ;\
+ xorl p1+2*tlen(,%rdi,4),OU3 ;\
+ xorl p1+2*tlen(,%rsi,4),OU4 ;\
+ shrl $8,I4E ;\
+ movzbl I1H,%edi ;\
+ movzbl I2H,%esi ;\
+ shrl $8,I3E ;\
+ xorl p1+3*tlen(,I4R,4),OU1 ;\
+ xorl p1+3*tlen(,%rdi,4),OU2 ;\
+ xorl p1+3*tlen(,%rsi,4),OU3 ;\
+ xorl p1+3*tlen(,I3R,4),OU4
+
+// This macro performs an inverse encryption cycle. It is entered with
+// the first previous round column values in I1E, I2E, I3E and I4E and
+// exits with the final values OU1, OU2, OU3 and OU4 registers.
+
+#define inv_rnd(p1,p2,I1E,I1B,I1R,I2E,I2B,I2R,I3E,I3B,I3H,I4E,I4B,I4H,OU1,OU2,OU3,OU4) \
+ movl p2+12(%rbp),OU4 ;\
+ movl p2+8(%rbp),OU3 ;\
+ movl p2+4(%rbp),OU2 ;\
+ movl p2(%rbp),OU1 ;\
+ movzbl I4B,%edi ;\
+ movzbl I3B,%esi ;\
+ movzbl I2B,%r8d ;\
+ movzbl I1B,%r13d ;\
+ shrl $8,I2E ;\
+ shrl $8,I1E ;\
+ xorl p1(,%rdi,4),OU4 ;\
+ xorl p1(,%rsi,4),OU3 ;\
+ xorl p1(,%r8,4),OU2 ;\
+ xorl p1(,%r13,4),OU1 ;\
+ movzbl I3H,%esi ;\
+ movzbl I2B,%r8d ;\
+ movzbl I1B,%r13d ;\
+ movzbl I4H,%edi ;\
+ shrl $8,I2E ;\
+ shrl $8,I1E ;\
+ xorl p1+tlen(,%rsi,4),OU4 ;\
+ xorl p1+tlen(,%r8,4),OU3 ;\
+ xorl p1+tlen(,%r13,4),OU2 ;\
+ xorl p1+tlen(,%rdi,4),OU1 ;\
+ shrl $16,I4E ;\
+ shrl $16,I3E ;\
+ movzbl I2B,%r8d ;\
+ movzbl I1B,%r13d ;\
+ movzbl I4B,%edi ;\
+ movzbl I3B,%esi ;\
+ xorl p1+2*tlen(,%r8,4),OU4 ;\
+ xorl p1+2*tlen(,%r13,4),OU3 ;\
+ xorl p1+2*tlen(,%rdi,4),OU2 ;\
+ xorl p1+2*tlen(,%rsi,4),OU1 ;\
+ shrl $8,I1E ;\
+ movzbl I4H,%edi ;\
+ movzbl I3H,%esi ;\
+ shrl $8,I2E ;\
+ xorl p1+3*tlen(,I1R,4),OU4 ;\
+ xorl p1+3*tlen(,%rdi,4),OU3 ;\
+ xorl p1+3*tlen(,%rsi,4),OU2 ;\
+ xorl p1+3*tlen(,I2R,4),OU1
+
+// AES (Rijndael) Encryption Subroutine
+
+// rdi = pointer to AES context
+// rsi = pointer to input plaintext bytes
+// rdx = pointer to output ciphertext bytes
+
+ .text
+ .align ALIGN64BYTES
+aes_encrypt:
+ movl (%rsi),%eax // read in plaintext
+ movl 4(%rsi),%ecx
+ movl 8(%rsi),%r10d
+ movl 12(%rsi),%r11d
+
+ pushq %rbp
+ leaq ekey+16(%rdi),%rbp // encryption key pointer
+ movq %rdx,%r9 // pointer to out block
+ movl nrnd(%rdi),%edx // number of rounds
+ pushq %rbx
+ pushq %r13
+ pushq %r14
+ pushq %r15
+
+ xorl -16(%rbp),%eax // xor in first round key
+ xorl -12(%rbp),%ecx
+ xorl -8(%rbp),%r10d
+ xorl -4(%rbp),%r11d
+
+ subl $10,%edx
+ je aes_15
+ addq $32,%rbp
+ subl $2,%edx
+ je aes_13
+ addq $32,%rbp
+
+ fwd_rnd(aes_ft_tab,-64,%eax,%al,%ah,%ecx,%cl,%ch,%r10d,%r10b,%r10,%r11d,%r11b,%r11,%ebx,%edx,%r14d,%r15d)
+ fwd_rnd(aes_ft_tab,-48,%ebx,%bl,%bh,%edx,%dl,%dh,%r14d,%r14b,%r14,%r15d,%r15b,%r15,%eax,%ecx,%r10d,%r11d)
+ jmp aes_13
+ .align ALIGN64BYTES
+aes_13: fwd_rnd(aes_ft_tab,-32,%eax,%al,%ah,%ecx,%cl,%ch,%r10d,%r10b,%r10,%r11d,%r11b,%r11,%ebx,%edx,%r14d,%r15d)
+ fwd_rnd(aes_ft_tab,-16,%ebx,%bl,%bh,%edx,%dl,%dh,%r14d,%r14b,%r14,%r15d,%r15b,%r15,%eax,%ecx,%r10d,%r11d)
+ jmp aes_15
+ .align ALIGN64BYTES
+aes_15: fwd_rnd(aes_ft_tab,0, %eax,%al,%ah,%ecx,%cl,%ch,%r10d,%r10b,%r10,%r11d,%r11b,%r11,%ebx,%edx,%r14d,%r15d)
+ fwd_rnd(aes_ft_tab,16, %ebx,%bl,%bh,%edx,%dl,%dh,%r14d,%r14b,%r14,%r15d,%r15b,%r15,%eax,%ecx,%r10d,%r11d)
+ fwd_rnd(aes_ft_tab,32, %eax,%al,%ah,%ecx,%cl,%ch,%r10d,%r10b,%r10,%r11d,%r11b,%r11,%ebx,%edx,%r14d,%r15d)
+ fwd_rnd(aes_ft_tab,48, %ebx,%bl,%bh,%edx,%dl,%dh,%r14d,%r14b,%r14,%r15d,%r15b,%r15,%eax,%ecx,%r10d,%r11d)
+ fwd_rnd(aes_ft_tab,64, %eax,%al,%ah,%ecx,%cl,%ch,%r10d,%r10b,%r10,%r11d,%r11b,%r11,%ebx,%edx,%r14d,%r15d)
+ fwd_rnd(aes_ft_tab,80, %ebx,%bl,%bh,%edx,%dl,%dh,%r14d,%r14b,%r14,%r15d,%r15b,%r15,%eax,%ecx,%r10d,%r11d)
+ fwd_rnd(aes_ft_tab,96, %eax,%al,%ah,%ecx,%cl,%ch,%r10d,%r10b,%r10,%r11d,%r11b,%r11,%ebx,%edx,%r14d,%r15d)
+ fwd_rnd(aes_ft_tab,112,%ebx,%bl,%bh,%edx,%dl,%dh,%r14d,%r14b,%r14,%r15d,%r15b,%r15,%eax,%ecx,%r10d,%r11d)
+ fwd_rnd(aes_ft_tab,128,%eax,%al,%ah,%ecx,%cl,%ch,%r10d,%r10b,%r10,%r11d,%r11b,%r11,%ebx,%edx,%r14d,%r15d)
+ fwd_rnd(aes_fl_tab,144,%ebx,%bl,%bh,%edx,%dl,%dh,%r14d,%r14b,%r14,%r15d,%r15b,%r15,%eax,%ecx,%r10d,%r11d)
+
+ popq %r15
+ popq %r14
+ popq %r13
+ popq %rbx
+ popq %rbp
+
+ movl %eax,(%r9) // move final values to the output array.
+ movl %ecx,4(%r9)
+ movl %r10d,8(%r9)
+ movl %r11d,12(%r9)
+ ret
+
+// AES (Rijndael) Decryption Subroutine
+
+// rdi = pointer to AES context
+// rsi = pointer to input ciphertext bytes
+// rdx = pointer to output plaintext bytes
+
+ .align ALIGN64BYTES
+aes_decrypt:
+ movl 12(%rsi),%eax // read in ciphertext
+ movl 8(%rsi),%ecx
+ movl 4(%rsi),%r10d
+ movl (%rsi),%r11d
+
+ pushq %rbp
+ leaq dkey+16(%rdi),%rbp // decryption key pointer
+ movq %rdx,%r9 // pointer to out block
+ movl nrnd(%rdi),%edx // number of rounds
+ pushq %rbx
+ pushq %r13
+ pushq %r14
+ pushq %r15
+
+ xorl -4(%rbp),%eax // xor in first round key
+ xorl -8(%rbp),%ecx
+ xorl -12(%rbp),%r10d
+ xorl -16(%rbp),%r11d
+
+ subl $10,%edx
+ je aes_25
+ addq $32,%rbp
+ subl $2,%edx
+ je aes_23
+ addq $32,%rbp
+
+ inv_rnd(aes_it_tab,-64,%r11d,%r11b,%r11,%r10d,%r10b,%r10,%ecx,%cl,%ch,%eax,%al,%ah,%r15d,%r14d,%edx,%ebx)
+ inv_rnd(aes_it_tab,-48,%r15d,%r15b,%r15,%r14d,%r14b,%r14,%edx,%dl,%dh,%ebx,%bl,%bh,%r11d,%r10d,%ecx,%eax)
+ jmp aes_23
+ .align ALIGN64BYTES
+aes_23: inv_rnd(aes_it_tab,-32,%r11d,%r11b,%r11,%r10d,%r10b,%r10,%ecx,%cl,%ch,%eax,%al,%ah,%r15d,%r14d,%edx,%ebx)
+ inv_rnd(aes_it_tab,-16,%r15d,%r15b,%r15,%r14d,%r14b,%r14,%edx,%dl,%dh,%ebx,%bl,%bh,%r11d,%r10d,%ecx,%eax)
+ jmp aes_25
+ .align ALIGN64BYTES
+aes_25: inv_rnd(aes_it_tab,0, %r11d,%r11b,%r11,%r10d,%r10b,%r10,%ecx,%cl,%ch,%eax,%al,%ah,%r15d,%r14d,%edx,%ebx)
+ inv_rnd(aes_it_tab,16, %r15d,%r15b,%r15,%r14d,%r14b,%r14,%edx,%dl,%dh,%ebx,%bl,%bh,%r11d,%r10d,%ecx,%eax)
+ inv_rnd(aes_it_tab,32, %r11d,%r11b,%r11,%r10d,%r10b,%r10,%ecx,%cl,%ch,%eax,%al,%ah,%r15d,%r14d,%edx,%ebx)
+ inv_rnd(aes_it_tab,48, %r15d,%r15b,%r15,%r14d,%r14b,%r14,%edx,%dl,%dh,%ebx,%bl,%bh,%r11d,%r10d,%ecx,%eax)
+ inv_rnd(aes_it_tab,64, %r11d,%r11b,%r11,%r10d,%r10b,%r10,%ecx,%cl,%ch,%eax,%al,%ah,%r15d,%r14d,%edx,%ebx)
+ inv_rnd(aes_it_tab,80, %r15d,%r15b,%r15,%r14d,%r14b,%r14,%edx,%dl,%dh,%ebx,%bl,%bh,%r11d,%r10d,%ecx,%eax)
+ inv_rnd(aes_it_tab,96, %r11d,%r11b,%r11,%r10d,%r10b,%r10,%ecx,%cl,%ch,%eax,%al,%ah,%r15d,%r14d,%edx,%ebx)
+ inv_rnd(aes_it_tab,112,%r15d,%r15b,%r15,%r14d,%r14b,%r14,%edx,%dl,%dh,%ebx,%bl,%bh,%r11d,%r10d,%ecx,%eax)
+ inv_rnd(aes_it_tab,128,%r11d,%r11b,%r11,%r10d,%r10b,%r10,%ecx,%cl,%ch,%eax,%al,%ah,%r15d,%r14d,%edx,%ebx)
+ inv_rnd(aes_il_tab,144,%r15d,%r15b,%r15,%r14d,%r14b,%r14,%edx,%dl,%dh,%ebx,%bl,%bh,%r11d,%r10d,%ecx,%eax)
+
+ popq %r15
+ popq %r14
+ popq %r13
+ popq %rbx
+ popq %rbp
+
+ movl %eax,12(%r9) // move final values to the output array.
+ movl %ecx,8(%r9)
+ movl %r10d,4(%r9)
+ movl %r11d,(%r9)
+ ret
+
+// AES (Rijndael) Key Schedule Subroutine
+
+// This macro performs a column mixing operation on an input 32-bit
+// word to give a 32-bit result. It uses each of the 4 bytes in the
+// the input column to index 4 different tables of 256 32-bit words
+// that are xored together to form the output value.
+
+#define mix_col(p1) \
+ movzbl %bl,%ecx ;\
+ movl p1(,%rcx,4),%eax ;\
+ movzbl %bh,%ecx ;\
+ ror $16,%ebx ;\
+ xorl p1+tlen(,%rcx,4),%eax ;\
+ movzbl %bl,%ecx ;\
+ xorl p1+2*tlen(,%rcx,4),%eax ;\
+ movzbl %bh,%ecx ;\
+ xorl p1+3*tlen(,%rcx,4),%eax
+
+// Key Schedule Macros
+
+#define ksc4(p1) \
+ rol $24,%ebx ;\
+ mix_col(aes_fl_tab) ;\
+ ror $8,%ebx ;\
+ xorl 4*p1+aes_rcon_tab,%eax ;\
+ xorl %eax,%esi ;\
+ xorl %esi,%ebp ;\
+ movl %esi,16*p1(%rdi) ;\
+ movl %ebp,16*p1+4(%rdi) ;\
+ xorl %ebp,%edx ;\
+ xorl %edx,%ebx ;\
+ movl %edx,16*p1+8(%rdi) ;\
+ movl %ebx,16*p1+12(%rdi)
+
+#define ksc6(p1) \
+ rol $24,%ebx ;\
+ mix_col(aes_fl_tab) ;\
+ ror $8,%ebx ;\
+ xorl 4*p1+aes_rcon_tab,%eax ;\
+ xorl 24*p1-24(%rdi),%eax ;\
+ movl %eax,24*p1(%rdi) ;\
+ xorl 24*p1-20(%rdi),%eax ;\
+ movl %eax,24*p1+4(%rdi) ;\
+ xorl %eax,%esi ;\
+ xorl %esi,%ebp ;\
+ movl %esi,24*p1+8(%rdi) ;\
+ movl %ebp,24*p1+12(%rdi) ;\
+ xorl %ebp,%edx ;\
+ xorl %edx,%ebx ;\
+ movl %edx,24*p1+16(%rdi) ;\
+ movl %ebx,24*p1+20(%rdi)
+
+#define ksc8(p1) \
+ rol $24,%ebx ;\
+ mix_col(aes_fl_tab) ;\
+ ror $8,%ebx ;\
+ xorl 4*p1+aes_rcon_tab,%eax ;\
+ xorl 32*p1-32(%rdi),%eax ;\
+ movl %eax,32*p1(%rdi) ;\
+ xorl 32*p1-28(%rdi),%eax ;\
+ movl %eax,32*p1+4(%rdi) ;\
+ xorl 32*p1-24(%rdi),%eax ;\
+ movl %eax,32*p1+8(%rdi) ;\
+ xorl 32*p1-20(%rdi),%eax ;\
+ movl %eax,32*p1+12(%rdi) ;\
+ pushq %rbx ;\
+ movl %eax,%ebx ;\
+ mix_col(aes_fl_tab) ;\
+ popq %rbx ;\
+ xorl %eax,%esi ;\
+ xorl %esi,%ebp ;\
+ movl %esi,32*p1+16(%rdi) ;\
+ movl %ebp,32*p1+20(%rdi) ;\
+ xorl %ebp,%edx ;\
+ xorl %edx,%ebx ;\
+ movl %edx,32*p1+24(%rdi) ;\
+ movl %ebx,32*p1+28(%rdi)
+
+// rdi = pointer to AES context
+// rsi = pointer to key bytes
+// rdx = key length, bytes or bits
+// rcx = ed_flag, 1=encrypt only, 0=both encrypt and decrypt
+
+ .align ALIGN64BYTES
+aes_set_key:
+ pushfq
+ pushq %rbp
+ pushq %rbx
+
+ movq %rcx,%r11 // ed_flg
+ movq %rdx,%rcx // key length
+ movq %rdi,%r10 // AES context
+
+ cmpl $128,%ecx
+ jb aes_30
+ shrl $3,%ecx
+aes_30: cmpl $32,%ecx
+ je aes_32
+ cmpl $24,%ecx
+ je aes_32
+ movl $16,%ecx
+aes_32: shrl $2,%ecx
+ movl %ecx,nkey(%r10)
+ leaq 6(%rcx),%rax // 10/12/14 for 4/6/8 32-bit key length
+ movl %eax,nrnd(%r10)
+ leaq ekey(%r10),%rdi // key position in AES context
+ cld
+ movl %ecx,%eax // save key length in eax
+ rep ; movsl // words in the key schedule
+ movl -4(%rsi),%ebx // put some values in registers
+ movl -8(%rsi),%edx // to allow faster code
+ movl -12(%rsi),%ebp
+ movl -16(%rsi),%esi
+
+ cmpl $4,%eax // jump on key size
+ je aes_36
+ cmpl $6,%eax
+ je aes_35
+
+ ksc8(0)
+ ksc8(1)
+ ksc8(2)
+ ksc8(3)
+ ksc8(4)
+ ksc8(5)
+ ksc8(6)
+ jmp aes_37
+aes_35: ksc6(0)
+ ksc6(1)
+ ksc6(2)
+ ksc6(3)
+ ksc6(4)
+ ksc6(5)
+ ksc6(6)
+ ksc6(7)
+ jmp aes_37
+aes_36: ksc4(0)
+ ksc4(1)
+ ksc4(2)
+ ksc4(3)
+ ksc4(4)
+ ksc4(5)
+ ksc4(6)
+ ksc4(7)
+ ksc4(8)
+ ksc4(9)
+aes_37: cmpl $0,%r11d // ed_flg
+ jne aes_39
+
+// compile decryption key schedule from encryption schedule - reverse
+// order and do mix_column operation on round keys except first and last
+
+ movl nrnd(%r10),%eax // kt = cx->d_key + nc * cx->Nrnd
+ shl $2,%rax
+ leaq dkey(%r10,%rax,4),%rdi
+ leaq ekey(%r10),%rsi // kf = cx->e_key
+
+ movsq // copy first round key (unmodified)
+ movsq
+ subq $32,%rdi
+ movl $1,%r9d
+aes_38: // do mix column on each column of
+ lodsl // each round key
+ movl %eax,%ebx
+ mix_col(aes_im_tab)
+ stosl
+ lodsl
+ movl %eax,%ebx
+ mix_col(aes_im_tab)
+ stosl
+ lodsl
+ movl %eax,%ebx
+ mix_col(aes_im_tab)
+ stosl
+ lodsl
+ movl %eax,%ebx
+ mix_col(aes_im_tab)
+ stosl
+ subq $32,%rdi
+
+ incl %r9d
+ cmpl nrnd(%r10),%r9d
+ jb aes_38
+
+ movsq // copy last round key (unmodified)
+ movsq
+aes_39: popq %rbx
+ popq %rbp
+ popfq
+ ret
+
+
+// finite field multiplies by {02}, {04} and {08}
+
+#define f2(x) ((x<<1)^(((x>>7)&1)*0x11b))
+#define f4(x) ((x<<2)^(((x>>6)&1)*0x11b)^(((x>>6)&2)*0x11b))
+#define f8(x) ((x<<3)^(((x>>5)&1)*0x11b)^(((x>>5)&2)*0x11b)^(((x>>5)&4)*0x11b))
+
+// finite field multiplies required in table generation
+
+#define f3(x) (f2(x) ^ x)
+#define f9(x) (f8(x) ^ x)
+#define fb(x) (f8(x) ^ f2(x) ^ x)
+#define fd(x) (f8(x) ^ f4(x) ^ x)
+#define fe(x) (f8(x) ^ f4(x) ^ f2(x))
+
+// These defines generate the forward table entries
+
+#define u0(x) ((f3(x) << 24) | (x << 16) | (x << 8) | f2(x))
+#define u1(x) ((x << 24) | (x << 16) | (f2(x) << 8) | f3(x))
+#define u2(x) ((x << 24) | (f2(x) << 16) | (f3(x) << 8) | x)
+#define u3(x) ((f2(x) << 24) | (f3(x) << 16) | (x << 8) | x)
+
+// These defines generate the inverse table entries
+
+#define v0(x) ((fb(x) << 24) | (fd(x) << 16) | (f9(x) << 8) | fe(x))
+#define v1(x) ((fd(x) << 24) | (f9(x) << 16) | (fe(x) << 8) | fb(x))
+#define v2(x) ((f9(x) << 24) | (fe(x) << 16) | (fb(x) << 8) | fd(x))
+#define v3(x) ((fe(x) << 24) | (fb(x) << 16) | (fd(x) << 8) | f9(x))
+
+// These defines generate entries for the last round tables
+
+#define w0(x) (x)
+#define w1(x) (x << 8)
+#define w2(x) (x << 16)
+#define w3(x) (x << 24)
+
+// macro to generate inverse mix column tables (needed for the key schedule)
+
+#define im_data0(p1) \
+ .long p1(0x00),p1(0x01),p1(0x02),p1(0x03),p1(0x04),p1(0x05),p1(0x06),p1(0x07) ;\
+ .long p1(0x08),p1(0x09),p1(0x0a),p1(0x0b),p1(0x0c),p1(0x0d),p1(0x0e),p1(0x0f) ;\
+ .long p1(0x10),p1(0x11),p1(0x12),p1(0x13),p1(0x14),p1(0x15),p1(0x16),p1(0x17) ;\
+ .long p1(0x18),p1(0x19),p1(0x1a),p1(0x1b),p1(0x1c),p1(0x1d),p1(0x1e),p1(0x1f)
+#define im_data1(p1) \
+ .long p1(0x20),p1(0x21),p1(0x22),p1(0x23),p1(0x24),p1(0x25),p1(0x26),p1(0x27) ;\
+ .long p1(0x28),p1(0x29),p1(0x2a),p1(0x2b),p1(0x2c),p1(0x2d),p1(0x2e),p1(0x2f) ;\
+ .long p1(0x30),p1(0x31),p1(0x32),p1(0x33),p1(0x34),p1(0x35),p1(0x36),p1(0x37) ;\
+ .long p1(0x38),p1(0x39),p1(0x3a),p1(0x3b),p1(0x3c),p1(0x3d),p1(0x3e),p1(0x3f)
+#define im_data2(p1) \
+ .long p1(0x40),p1(0x41),p1(0x42),p1(0x43),p1(0x44),p1(0x45),p1(0x46),p1(0x47) ;\
+ .long p1(0x48),p1(0x49),p1(0x4a),p1(0x4b),p1(0x4c),p1(0x4d),p1(0x4e),p1(0x4f) ;\
+ .long p1(0x50),p1(0x51),p1(0x52),p1(0x53),p1(0x54),p1(0x55),p1(0x56),p1(0x57) ;\
+ .long p1(0x58),p1(0x59),p1(0x5a),p1(0x5b),p1(0x5c),p1(0x5d),p1(0x5e),p1(0x5f)
+#define im_data3(p1) \
+ .long p1(0x60),p1(0x61),p1(0x62),p1(0x63),p1(0x64),p1(0x65),p1(0x66),p1(0x67) ;\
+ .long p1(0x68),p1(0x69),p1(0x6a),p1(0x6b),p1(0x6c),p1(0x6d),p1(0x6e),p1(0x6f) ;\
+ .long p1(0x70),p1(0x71),p1(0x72),p1(0x73),p1(0x74),p1(0x75),p1(0x76),p1(0x77) ;\
+ .long p1(0x78),p1(0x79),p1(0x7a),p1(0x7b),p1(0x7c),p1(0x7d),p1(0x7e),p1(0x7f)
+#define im_data4(p1) \
+ .long p1(0x80),p1(0x81),p1(0x82),p1(0x83),p1(0x84),p1(0x85),p1(0x86),p1(0x87) ;\
+ .long p1(0x88),p1(0x89),p1(0x8a),p1(0x8b),p1(0x8c),p1(0x8d),p1(0x8e),p1(0x8f) ;\
+ .long p1(0x90),p1(0x91),p1(0x92),p1(0x93),p1(0x94),p1(0x95),p1(0x96),p1(0x97) ;\
+ .long p1(0x98),p1(0x99),p1(0x9a),p1(0x9b),p1(0x9c),p1(0x9d),p1(0x9e),p1(0x9f)
+#define im_data5(p1) \
+ .long p1(0xa0),p1(0xa1),p1(0xa2),p1(0xa3),p1(0xa4),p1(0xa5),p1(0xa6),p1(0xa7) ;\
+ .long p1(0xa8),p1(0xa9),p1(0xaa),p1(0xab),p1(0xac),p1(0xad),p1(0xae),p1(0xaf) ;\
+ .long p1(0xb0),p1(0xb1),p1(0xb2),p1(0xb3),p1(0xb4),p1(0xb5),p1(0xb6),p1(0xb7) ;\
+ .long p1(0xb8),p1(0xb9),p1(0xba),p1(0xbb),p1(0xbc),p1(0xbd),p1(0xbe),p1(0xbf)
+#define im_data6(p1) \
+ .long p1(0xc0),p1(0xc1),p1(0xc2),p1(0xc3),p1(0xc4),p1(0xc5),p1(0xc6),p1(0xc7) ;\
+ .long p1(0xc8),p1(0xc9),p1(0xca),p1(0xcb),p1(0xcc),p1(0xcd),p1(0xce),p1(0xcf) ;\
+ .long p1(0xd0),p1(0xd1),p1(0xd2),p1(0xd3),p1(0xd4),p1(0xd5),p1(0xd6),p1(0xd7) ;\
+ .long p1(0xd8),p1(0xd9),p1(0xda),p1(0xdb),p1(0xdc),p1(0xdd),p1(0xde),p1(0xdf)
+#define im_data7(p1) \
+ .long p1(0xe0),p1(0xe1),p1(0xe2),p1(0xe3),p1(0xe4),p1(0xe5),p1(0xe6),p1(0xe7) ;\
+ .long p1(0xe8),p1(0xe9),p1(0xea),p1(0xeb),p1(0xec),p1(0xed),p1(0xee),p1(0xef) ;\
+ .long p1(0xf0),p1(0xf1),p1(0xf2),p1(0xf3),p1(0xf4),p1(0xf5),p1(0xf6),p1(0xf7) ;\
+ .long p1(0xf8),p1(0xf9),p1(0xfa),p1(0xfb),p1(0xfc),p1(0xfd),p1(0xfe),p1(0xff)
+
+// S-box data - 256 entries
+
+#define sb_data0(p1) \
+ .long p1(0x63),p1(0x7c),p1(0x77),p1(0x7b),p1(0xf2),p1(0x6b),p1(0x6f),p1(0xc5) ;\
+ .long p1(0x30),p1(0x01),p1(0x67),p1(0x2b),p1(0xfe),p1(0xd7),p1(0xab),p1(0x76) ;\
+ .long p1(0xca),p1(0x82),p1(0xc9),p1(0x7d),p1(0xfa),p1(0x59),p1(0x47),p1(0xf0) ;\
+ .long p1(0xad),p1(0xd4),p1(0xa2),p1(0xaf),p1(0x9c),p1(0xa4),p1(0x72),p1(0xc0)
+#define sb_data1(p1) \
+ .long p1(0xb7),p1(0xfd),p1(0x93),p1(0x26),p1(0x36),p1(0x3f),p1(0xf7),p1(0xcc) ;\
+ .long p1(0x34),p1(0xa5),p1(0xe5),p1(0xf1),p1(0x71),p1(0xd8),p1(0x31),p1(0x15) ;\
+ .long p1(0x04),p1(0xc7),p1(0x23),p1(0xc3),p1(0x18),p1(0x96),p1(0x05),p1(0x9a) ;\
+ .long p1(0x07),p1(0x12),p1(0x80),p1(0xe2),p1(0xeb),p1(0x27),p1(0xb2),p1(0x75)
+#define sb_data2(p1) \
+ .long p1(0x09),p1(0x83),p1(0x2c),p1(0x1a),p1(0x1b),p1(0x6e),p1(0x5a),p1(0xa0) ;\
+ .long p1(0x52),p1(0x3b),p1(0xd6),p1(0xb3),p1(0x29),p1(0xe3),p1(0x2f),p1(0x84) ;\
+ .long p1(0x53),p1(0xd1),p1(0x00),p1(0xed),p1(0x20),p1(0xfc),p1(0xb1),p1(0x5b) ;\
+ .long p1(0x6a),p1(0xcb),p1(0xbe),p1(0x39),p1(0x4a),p1(0x4c),p1(0x58),p1(0xcf)
+#define sb_data3(p1) \
+ .long p1(0xd0),p1(0xef),p1(0xaa),p1(0xfb),p1(0x43),p1(0x4d),p1(0x33),p1(0x85) ;\
+ .long p1(0x45),p1(0xf9),p1(0x02),p1(0x7f),p1(0x50),p1(0x3c),p1(0x9f),p1(0xa8) ;\
+ .long p1(0x51),p1(0xa3),p1(0x40),p1(0x8f),p1(0x92),p1(0x9d),p1(0x38),p1(0xf5) ;\
+ .long p1(0xbc),p1(0xb6),p1(0xda),p1(0x21),p1(0x10),p1(0xff),p1(0xf3),p1(0xd2)
+#define sb_data4(p1) \
+ .long p1(0xcd),p1(0x0c),p1(0x13),p1(0xec),p1(0x5f),p1(0x97),p1(0x44),p1(0x17) ;\
+ .long p1(0xc4),p1(0xa7),p1(0x7e),p1(0x3d),p1(0x64),p1(0x5d),p1(0x19),p1(0x73) ;\
+ .long p1(0x60),p1(0x81),p1(0x4f),p1(0xdc),p1(0x22),p1(0x2a),p1(0x90),p1(0x88) ;\
+ .long p1(0x46),p1(0xee),p1(0xb8),p1(0x14),p1(0xde),p1(0x5e),p1(0x0b),p1(0xdb)
+#define sb_data5(p1) \
+ .long p1(0xe0),p1(0x32),p1(0x3a),p1(0x0a),p1(0x49),p1(0x06),p1(0x24),p1(0x5c) ;\
+ .long p1(0xc2),p1(0xd3),p1(0xac),p1(0x62),p1(0x91),p1(0x95),p1(0xe4),p1(0x79) ;\
+ .long p1(0xe7),p1(0xc8),p1(0x37),p1(0x6d),p1(0x8d),p1(0xd5),p1(0x4e),p1(0xa9) ;\
+ .long p1(0x6c),p1(0x56),p1(0xf4),p1(0xea),p1(0x65),p1(0x7a),p1(0xae),p1(0x08)
+#define sb_data6(p1) \
+ .long p1(0xba),p1(0x78),p1(0x25),p1(0x2e),p1(0x1c),p1(0xa6),p1(0xb4),p1(0xc6) ;\
+ .long p1(0xe8),p1(0xdd),p1(0x74),p1(0x1f),p1(0x4b),p1(0xbd),p1(0x8b),p1(0x8a) ;\
+ .long p1(0x70),p1(0x3e),p1(0xb5),p1(0x66),p1(0x48),p1(0x03),p1(0xf6),p1(0x0e) ;\
+ .long p1(0x61),p1(0x35),p1(0x57),p1(0xb9),p1(0x86),p1(0xc1),p1(0x1d),p1(0x9e)
+#define sb_data7(p1) \
+ .long p1(0xe1),p1(0xf8),p1(0x98),p1(0x11),p1(0x69),p1(0xd9),p1(0x8e),p1(0x94) ;\
+ .long p1(0x9b),p1(0x1e),p1(0x87),p1(0xe9),p1(0xce),p1(0x55),p1(0x28),p1(0xdf) ;\
+ .long p1(0x8c),p1(0xa1),p1(0x89),p1(0x0d),p1(0xbf),p1(0xe6),p1(0x42),p1(0x68) ;\
+ .long p1(0x41),p1(0x99),p1(0x2d),p1(0x0f),p1(0xb0),p1(0x54),p1(0xbb),p1(0x16)
+
+// Inverse S-box data - 256 entries
+
+#define ib_data0(p1) \
+ .long p1(0x52),p1(0x09),p1(0x6a),p1(0xd5),p1(0x30),p1(0x36),p1(0xa5),p1(0x38) ;\
+ .long p1(0xbf),p1(0x40),p1(0xa3),p1(0x9e),p1(0x81),p1(0xf3),p1(0xd7),p1(0xfb) ;\
+ .long p1(0x7c),p1(0xe3),p1(0x39),p1(0x82),p1(0x9b),p1(0x2f),p1(0xff),p1(0x87) ;\
+ .long p1(0x34),p1(0x8e),p1(0x43),p1(0x44),p1(0xc4),p1(0xde),p1(0xe9),p1(0xcb)
+#define ib_data1(p1) \
+ .long p1(0x54),p1(0x7b),p1(0x94),p1(0x32),p1(0xa6),p1(0xc2),p1(0x23),p1(0x3d) ;\
+ .long p1(0xee),p1(0x4c),p1(0x95),p1(0x0b),p1(0x42),p1(0xfa),p1(0xc3),p1(0x4e) ;\
+ .long p1(0x08),p1(0x2e),p1(0xa1),p1(0x66),p1(0x28),p1(0xd9),p1(0x24),p1(0xb2) ;\
+ .long p1(0x76),p1(0x5b),p1(0xa2),p1(0x49),p1(0x6d),p1(0x8b),p1(0xd1),p1(0x25)
+#define ib_data2(p1) \
+ .long p1(0x72),p1(0xf8),p1(0xf6),p1(0x64),p1(0x86),p1(0x68),p1(0x98),p1(0x16) ;\
+ .long p1(0xd4),p1(0xa4),p1(0x5c),p1(0xcc),p1(0x5d),p1(0x65),p1(0xb6),p1(0x92) ;\
+ .long p1(0x6c),p1(0x70),p1(0x48),p1(0x50),p1(0xfd),p1(0xed),p1(0xb9),p1(0xda) ;\
+ .long p1(0x5e),p1(0x15),p1(0x46),p1(0x57),p1(0xa7),p1(0x8d),p1(0x9d),p1(0x84)
+#define ib_data3(p1) \
+ .long p1(0x90),p1(0xd8),p1(0xab),p1(0x00),p1(0x8c),p1(0xbc),p1(0xd3),p1(0x0a) ;\
+ .long p1(0xf7),p1(0xe4),p1(0x58),p1(0x05),p1(0xb8),p1(0xb3),p1(0x45),p1(0x06) ;\
+ .long p1(0xd0),p1(0x2c),p1(0x1e),p1(0x8f),p1(0xca),p1(0x3f),p1(0x0f),p1(0x02) ;\
+ .long p1(0xc1),p1(0xaf),p1(0xbd),p1(0x03),p1(0x01),p1(0x13),p1(0x8a),p1(0x6b)
+#define ib_data4(p1) \
+ .long p1(0x3a),p1(0x91),p1(0x11),p1(0x41),p1(0x4f),p1(0x67),p1(0xdc),p1(0xea) ;\
+ .long p1(0x97),p1(0xf2),p1(0xcf),p1(0xce),p1(0xf0),p1(0xb4),p1(0xe6),p1(0x73) ;\
+ .long p1(0x96),p1(0xac),p1(0x74),p1(0x22),p1(0xe7),p1(0xad),p1(0x35),p1(0x85) ;\
+ .long p1(0xe2),p1(0xf9),p1(0x37),p1(0xe8),p1(0x1c),p1(0x75),p1(0xdf),p1(0x6e)
+#define ib_data5(p1) \
+ .long p1(0x47),p1(0xf1),p1(0x1a),p1(0x71),p1(0x1d),p1(0x29),p1(0xc5),p1(0x89) ;\
+ .long p1(0x6f),p1(0xb7),p1(0x62),p1(0x0e),p1(0xaa),p1(0x18),p1(0xbe),p1(0x1b) ;\
+ .long p1(0xfc),p1(0x56),p1(0x3e),p1(0x4b),p1(0xc6),p1(0xd2),p1(0x79),p1(0x20) ;\
+ .long p1(0x9a),p1(0xdb),p1(0xc0),p1(0xfe),p1(0x78),p1(0xcd),p1(0x5a),p1(0xf4)
+#define ib_data6(p1) \
+ .long p1(0x1f),p1(0xdd),p1(0xa8),p1(0x33),p1(0x88),p1(0x07),p1(0xc7),p1(0x31) ;\
+ .long p1(0xb1),p1(0x12),p1(0x10),p1(0x59),p1(0x27),p1(0x80),p1(0xec),p1(0x5f) ;\
+ .long p1(0x60),p1(0x51),p1(0x7f),p1(0xa9),p1(0x19),p1(0xb5),p1(0x4a),p1(0x0d) ;\
+ .long p1(0x2d),p1(0xe5),p1(0x7a),p1(0x9f),p1(0x93),p1(0xc9),p1(0x9c),p1(0xef)
+#define ib_data7(p1) \
+ .long p1(0xa0),p1(0xe0),p1(0x3b),p1(0x4d),p1(0xae),p1(0x2a),p1(0xf5),p1(0xb0) ;\
+ .long p1(0xc8),p1(0xeb),p1(0xbb),p1(0x3c),p1(0x83),p1(0x53),p1(0x99),p1(0x61) ;\
+ .long p1(0x17),p1(0x2b),p1(0x04),p1(0x7e),p1(0xba),p1(0x77),p1(0xd6),p1(0x26) ;\
+ .long p1(0xe1),p1(0x69),p1(0x14),p1(0x63),p1(0x55),p1(0x21),p1(0x0c),p1(0x7d)
+
+// The rcon_table (needed for the key schedule)
+//
+// Here is original Dr Brian Gladman's source code:
+// _rcon_tab:
+// %assign x 1
+// %rep 29
+// dd x
+// %assign x f2(x)
+// %endrep
+//
+// Here is precomputed output (it's more portable this way):
+
+ .section .rodata
+ .align ALIGN64BYTES
+aes_rcon_tab:
+ .long 0x01,0x02,0x04,0x08,0x10,0x20,0x40,0x80
+ .long 0x1b,0x36,0x6c,0xd8,0xab,0x4d,0x9a,0x2f
+ .long 0x5e,0xbc,0x63,0xc6,0x97,0x35,0x6a,0xd4
+ .long 0xb3,0x7d,0xfa,0xef,0xc5
+
+// The forward xor tables
+
+ .align ALIGN64BYTES
+aes_ft_tab:
+ sb_data0(u0)
+ sb_data1(u0)
+ sb_data2(u0)
+ sb_data3(u0)
+ sb_data4(u0)
+ sb_data5(u0)
+ sb_data6(u0)
+ sb_data7(u0)
+
+ sb_data0(u1)
+ sb_data1(u1)
+ sb_data2(u1)
+ sb_data3(u1)
+ sb_data4(u1)
+ sb_data5(u1)
+ sb_data6(u1)
+ sb_data7(u1)
+
+ sb_data0(u2)
+ sb_data1(u2)
+ sb_data2(u2)
+ sb_data3(u2)
+ sb_data4(u2)
+ sb_data5(u2)
+ sb_data6(u2)
+ sb_data7(u2)
+
+ sb_data0(u3)
+ sb_data1(u3)
+ sb_data2(u3)
+ sb_data3(u3)
+ sb_data4(u3)
+ sb_data5(u3)
+ sb_data6(u3)
+ sb_data7(u3)
+
+ .align ALIGN64BYTES
+aes_fl_tab:
+ sb_data0(w0)
+ sb_data1(w0)
+ sb_data2(w0)
+ sb_data3(w0)
+ sb_data4(w0)
+ sb_data5(w0)
+ sb_data6(w0)
+ sb_data7(w0)
+
+ sb_data0(w1)
+ sb_data1(w1)
+ sb_data2(w1)
+ sb_data3(w1)
+ sb_data4(w1)
+ sb_data5(w1)
+ sb_data6(w1)
+ sb_data7(w1)
+
+ sb_data0(w2)
+ sb_data1(w2)
+ sb_data2(w2)
+ sb_data3(w2)
+ sb_data4(w2)
+ sb_data5(w2)
+ sb_data6(w2)
+ sb_data7(w2)
+
+ sb_data0(w3)
+ sb_data1(w3)
+ sb_data2(w3)
+ sb_data3(w3)
+ sb_data4(w3)
+ sb_data5(w3)
+ sb_data6(w3)
+ sb_data7(w3)
+
+// The inverse xor tables
+
+ .align ALIGN64BYTES
+aes_it_tab:
+ ib_data0(v0)
+ ib_data1(v0)
+ ib_data2(v0)
+ ib_data3(v0)
+ ib_data4(v0)
+ ib_data5(v0)
+ ib_data6(v0)
+ ib_data7(v0)
+
+ ib_data0(v1)
+ ib_data1(v1)
+ ib_data2(v1)
+ ib_data3(v1)
+ ib_data4(v1)
+ ib_data5(v1)
+ ib_data6(v1)
+ ib_data7(v1)
+
+ ib_data0(v2)
+ ib_data1(v2)
+ ib_data2(v2)
+ ib_data3(v2)
+ ib_data4(v2)
+ ib_data5(v2)
+ ib_data6(v2)
+ ib_data7(v2)
+
+ ib_data0(v3)
+ ib_data1(v3)
+ ib_data2(v3)
+ ib_data3(v3)
+ ib_data4(v3)
+ ib_data5(v3)
+ ib_data6(v3)
+ ib_data7(v3)
+
+ .align ALIGN64BYTES
+aes_il_tab:
+ ib_data0(w0)
+ ib_data1(w0)
+ ib_data2(w0)
+ ib_data3(w0)
+ ib_data4(w0)
+ ib_data5(w0)
+ ib_data6(w0)
+ ib_data7(w0)
+
+ ib_data0(w1)
+ ib_data1(w1)
+ ib_data2(w1)
+ ib_data3(w1)
+ ib_data4(w1)
+ ib_data5(w1)
+ ib_data6(w1)
+ ib_data7(w1)
+
+ ib_data0(w2)
+ ib_data1(w2)
+ ib_data2(w2)
+ ib_data3(w2)
+ ib_data4(w2)
+ ib_data5(w2)
+ ib_data6(w2)
+ ib_data7(w2)
+
+ ib_data0(w3)
+ ib_data1(w3)
+ ib_data2(w3)
+ ib_data3(w3)
+ ib_data4(w3)
+ ib_data5(w3)
+ ib_data6(w3)
+ ib_data7(w3)
+
+// The inverse mix column tables
+
+ .align ALIGN64BYTES
+aes_im_tab:
+ im_data0(v0)
+ im_data1(v0)
+ im_data2(v0)
+ im_data3(v0)
+ im_data4(v0)
+ im_data5(v0)
+ im_data6(v0)
+ im_data7(v0)
+
+ im_data0(v1)
+ im_data1(v1)
+ im_data2(v1)
+ im_data3(v1)
+ im_data4(v1)
+ im_data5(v1)
+ im_data6(v1)
+ im_data7(v1)
+
+ im_data0(v2)
+ im_data1(v2)
+ im_data2(v2)
+ im_data3(v2)
+ im_data4(v2)
+ im_data5(v2)
+ im_data6(v2)
+ im_data7(v2)
+
+ im_data0(v3)
+ im_data1(v3)
+ im_data2(v3)
+ im_data3(v3)
+ im_data4(v3)
+ im_data5(v3)
+ im_data6(v3)
+ im_data7(v3)
+
+#if defined(__ELF__) && defined(SECTION_NOTE_GNU_STACK)
+ .section .note.GNU-stack,"",@progbits
+#endif
diff -urN linux-3.14-noloop/drivers/misc/aes-intel32.S linux-3.14-AES/drivers/misc/aes-intel32.S
--- linux-3.14-noloop/drivers/misc/aes-intel32.S 1970-01-01 02:00:00.000000000 +0200
+++ linux-3.14-AES/drivers/misc/aes-intel32.S 2014-03-31 22:20:36.000000000 +0300
@@ -0,0 +1,655 @@
+/*
+ * Implement AES algorithm in Intel AES-NI instructions.
+ *
+ * The white paper of AES-NI instructions can be downloaded from:
+ * http://softwarecommunity.intel.com/isn/downloads/intelavx/AES-Instructions-Set_WP.pdf
+ *
+ * Copyright (C) 2008, Intel Corp.
+ * Author: Huang Ying <ying.huang@intel.com>
+ * Vinodh Gopal <vinodh.gopal@intel.com>
+ * Kahraman Akdemir
+ *
+ * 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.
+ */
+
+/*
+ * Modified by Jari Ruusu, October 2009
+ * - Adapted for loop-AES
+ */
+
+/*
+ * Modified by Jari Ruusu, March 2010
+ * - Added parallelized 4x512 CBC encrypt
+ */
+
+#if !defined(ALIGN64BYTES)
+# define ALIGN64BYTES 64
+#endif
+
+ .file "aes-intel32.S"
+ .globl intel_aes_cbc_encrypt
+ .globl intel_aes_cbc_decrypt
+ .globl intel_aes_cbc_enc_4x512
+ .text
+
+#define STATE1 %xmm0
+#define STATE2 %xmm4
+#define STATE3 %xmm5
+#define STATE STATE1
+#define IN1 %xmm1
+#define IN2 %xmm7
+#define IN3 %xmm6
+#define IN IN1
+#define KEY %xmm2
+#define IV %xmm3
+
+#define KEYP %edi
+#define INP %esi
+#define OUTP %edx
+#define LEN %ecx
+#define IVP %ebx
+#define NRND %eax
+#define TKEYP %ebp
+
+/*
+ * void intel_aes_cbc_encrypt(const aes_context *, void *src, void *dst, size_t len, void *iv)
+ *
+ * Stack after reg saves: 36(%esp) = void *iv
+ * 32(%esp) = size_t len
+ * 28(%esp) = void *dst
+ * 24(%esp) = void *src
+ * 20(%esp) = aes_context *
+ */
+ .align ALIGN64BYTES
+intel_aes_cbc_encrypt:
+ push %edi
+ push %esi
+ push %ebx
+ push %ebp
+ mov 20(%esp),KEYP
+ mov 24(%esp),INP
+ mov 28(%esp),OUTP
+ mov 32(%esp),LEN
+ mov 36(%esp),IVP
+ mov 4(KEYP), NRND
+ add $8, KEYP
+ movups (IVP), STATE # load iv as initial state
+.align 4
+.Lcbc_enc_loop:
+ movups (INP), IN # load input
+ pxor IN, STATE
+
+ movaps (KEYP), KEY # key
+ mov KEYP, TKEYP
+ pxor KEY, STATE # round 0
+ add $0x30, TKEYP
+ cmp $12, NRND
+ jb .Lenc128
+ lea 0x20(TKEYP), TKEYP
+ je .Lenc192
+ add $0x20, TKEYP
+ movaps -0x60(TKEYP), KEY
+ # aesenc KEY, STATE
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xc2
+ movaps -0x50(TKEYP), KEY
+ # aesenc KEY, STATE
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xc2
+.align 4
+.Lenc192:
+ movaps -0x40(TKEYP), KEY
+ # aesenc KEY, STATE
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xc2
+ movaps -0x30(TKEYP), KEY
+ # aesenc KEY, STATE
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xc2
+.align 4
+.Lenc128:
+ movaps -0x20(TKEYP), KEY
+ # aesenc KEY, STATE
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xc2
+ movaps -0x10(TKEYP), KEY
+ # aesenc KEY, STATE
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xc2
+ movaps (TKEYP), KEY
+ # aesenc KEY, STATE
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xc2
+ movaps 0x10(TKEYP), KEY
+ # aesenc KEY, STATE
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xc2
+ movaps 0x20(TKEYP), KEY
+ # aesenc KEY, STATE
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xc2
+ movaps 0x30(TKEYP), KEY
+ # aesenc KEY, STATE
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xc2
+ movaps 0x40(TKEYP), KEY
+ # aesenc KEY, STATE
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xc2
+ movaps 0x50(TKEYP), KEY
+ # aesenc KEY, STATE
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xc2
+ movaps 0x60(TKEYP), KEY
+ # aesenc KEY, STATE
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xc2
+ movaps 0x70(TKEYP), KEY
+ # aesenclast KEY, STATE # last round
+ .byte 0x66, 0x0f, 0x38, 0xdd, 0xc2
+
+ movups STATE, (OUTP) # store output
+ sub $16, LEN
+ add $16, INP
+ add $16, OUTP
+ cmp $16, LEN
+ jge .Lcbc_enc_loop
+ emms
+ pop %ebp
+ pop %ebx
+ pop %esi
+ pop %edi
+ ret
+
+/*
+ * void intel_aes_cbc_decrypt(const aes_context *, void *src, void *dst, size_t len, void *iv)
+ *
+ * Stack after reg saves: 36(%esp) = void *iv
+ * 32(%esp) = size_t len
+ * 28(%esp) = void *dst
+ * 24(%esp) = void *src
+ * 20(%esp) = aes_context *
+ */
+ .align ALIGN64BYTES
+intel_aes_cbc_decrypt:
+ push %edi
+ push %esi
+ push %ebx
+ push %ebp
+ mov 20(%esp),KEYP
+ mov 24(%esp),INP
+ mov 28(%esp),OUTP
+ mov 32(%esp),LEN
+ mov 36(%esp),IVP
+ mov 4(KEYP), NRND
+ add $264, KEYP
+ movups (IVP), IV
+ cmp $48, LEN
+ jb .Lcbc_dec_loop1
+.align 4
+.Lcbc_dec_loop3:
+ movups (INP), IN1
+ movaps IN1, STATE1
+ movups 0x10(INP), IN2
+ movaps IN2, STATE2
+ movups 0x20(INP), IN3
+ movaps IN3, STATE3
+
+ movaps (KEYP), KEY # key
+ mov KEYP, TKEYP
+ pxor KEY, STATE1 # round 0
+ pxor KEY, STATE2
+ pxor KEY, STATE3
+ add $0x30, TKEYP
+ cmp $12, NRND
+ jb .L4dec128
+ lea 0x20(TKEYP), TKEYP
+ je .L4dec192
+ add $0x20, TKEYP
+ movaps -0x60(TKEYP), KEY
+ # aesdec KEY, STATE1
+ .byte 0x66, 0x0f, 0x38, 0xde, 0xc2
+ # aesdec KEY, STATE2
+ .byte 0x66, 0x0f, 0x38, 0xde, 0xe2
+ # aesdec KEY, STATE3
+ .byte 0x66, 0x0f, 0x38, 0xde, 0xea
+ movaps -0x50(TKEYP), KEY
+ # aesdec KEY, STATE1
+ .byte 0x66, 0x0f, 0x38, 0xde, 0xc2
+ # aesdec KEY, STATE2
+ .byte 0x66, 0x0f, 0x38, 0xde, 0xe2
+ # aesdec KEY, STATE3
+ .byte 0x66, 0x0f, 0x38, 0xde, 0xea
+.align 4
+.L4dec192:
+ movaps -0x40(TKEYP), KEY
+ # aesdec KEY, STATE1
+ .byte 0x66, 0x0f, 0x38, 0xde, 0xc2
+ # aesdec KEY, STATE2
+ .byte 0x66, 0x0f, 0x38, 0xde, 0xe2
+ # aesdec KEY, STATE3
+ .byte 0x66, 0x0f, 0x38, 0xde, 0xea
+ movaps -0x30(TKEYP), KEY
+ # aesdec KEY, STATE1
+ .byte 0x66, 0x0f, 0x38, 0xde, 0xc2
+ # aesdec KEY, STATE2
+ .byte 0x66, 0x0f, 0x38, 0xde, 0xe2
+ # aesdec KEY, STATE3
+ .byte 0x66, 0x0f, 0x38, 0xde, 0xea
+.align 4
+.L4dec128:
+ movaps -0x20(TKEYP), KEY
+ # aesdec KEY, STATE1
+ .byte 0x66, 0x0f, 0x38, 0xde, 0xc2
+ # aesdec KEY, STATE2
+ .byte 0x66, 0x0f, 0x38, 0xde, 0xe2
+ # aesdec KEY, STATE3
+ .byte 0x66, 0x0f, 0x38, 0xde, 0xea
+ movaps -0x10(TKEYP), KEY
+ # aesdec KEY, STATE1
+ .byte 0x66, 0x0f, 0x38, 0xde, 0xc2
+ # aesdec KEY, STATE2
+ .byte 0x66, 0x0f, 0x38, 0xde, 0xe2
+ # aesdec KEY, STATE3
+ .byte 0x66, 0x0f, 0x38, 0xde, 0xea
+ movaps (TKEYP), KEY
+ # aesdec KEY, STATE1
+ .byte 0x66, 0x0f, 0x38, 0xde, 0xc2
+ # aesdec KEY, STATE2
+ .byte 0x66, 0x0f, 0x38, 0xde, 0xe2
+ # aesdec KEY, STATE3
+ .byte 0x66, 0x0f, 0x38, 0xde, 0xea
+ movaps 0x10(TKEYP), KEY
+ # aesdec KEY, STATE1
+ .byte 0x66, 0x0f, 0x38, 0xde, 0xc2
+ # aesdec KEY, STATE2
+ .byte 0x66, 0x0f, 0x38, 0xde, 0xe2
+ # aesdec KEY, STATE3
+ .byte 0x66, 0x0f, 0x38, 0xde, 0xea
+ movaps 0x20(TKEYP), KEY
+ # aesdec KEY, STATE1
+ .byte 0x66, 0x0f, 0x38, 0xde, 0xc2
+ # aesdec KEY, STATE2
+ .byte 0x66, 0x0f, 0x38, 0xde, 0xe2
+ # aesdec KEY, STATE3
+ .byte 0x66, 0x0f, 0x38, 0xde, 0xea
+ movaps 0x30(TKEYP), KEY
+ # aesdec KEY, STATE1
+ .byte 0x66, 0x0f, 0x38, 0xde, 0xc2
+ # aesdec KEY, STATE2
+ .byte 0x66, 0x0f, 0x38, 0xde, 0xe2
+ # aesdec KEY, STATE3
+ .byte 0x66, 0x0f, 0x38, 0xde, 0xea
+ movaps 0x40(TKEYP), KEY
+ # aesdec KEY, STATE1
+ .byte 0x66, 0x0f, 0x38, 0xde, 0xc2
+ # aesdec KEY, STATE2
+ .byte 0x66, 0x0f, 0x38, 0xde, 0xe2
+ # aesdec KEY, STATE3
+ .byte 0x66, 0x0f, 0x38, 0xde, 0xea
+ movaps 0x50(TKEYP), KEY
+ # aesdec KEY, STATE1
+ .byte 0x66, 0x0f, 0x38, 0xde, 0xc2
+ # aesdec KEY, STATE2
+ .byte 0x66, 0x0f, 0x38, 0xde, 0xe2
+ # aesdec KEY, STATE3
+ .byte 0x66, 0x0f, 0x38, 0xde, 0xea
+ movaps 0x60(TKEYP), KEY
+ # aesdec KEY, STATE1
+ .byte 0x66, 0x0f, 0x38, 0xde, 0xc2
+ # aesdec KEY, STATE2
+ .byte 0x66, 0x0f, 0x38, 0xde, 0xe2
+ # aesdec KEY, STATE3
+ .byte 0x66, 0x0f, 0x38, 0xde, 0xea
+ movaps 0x70(TKEYP), KEY
+ # aesdeclast KEY, STATE1 # last round
+ .byte 0x66, 0x0f, 0x38, 0xdf, 0xc2
+ # aesdeclast KEY, STATE2
+ .byte 0x66, 0x0f, 0x38, 0xdf, 0xe2
+ # aesdeclast KEY, STATE3
+ .byte 0x66, 0x0f, 0x38, 0xdf, 0xea
+
+ pxor IV, STATE1
+ pxor IN1, STATE2
+ pxor IN2, STATE3
+ movaps IN3, IV
+ movups STATE1, (OUTP)
+ movups STATE2, 0x10(OUTP)
+ movups STATE3, 0x20(OUTP)
+ sub $48, LEN
+ add $48, INP
+ add $48, OUTP
+ cmp $48, LEN
+ jge .Lcbc_dec_loop3
+ cmp $16, LEN
+ jb .Lcbc_dec_ret
+.align 4
+.Lcbc_dec_loop1:
+ movups (INP), IN
+ movaps IN, STATE
+
+ movaps (KEYP), KEY # key
+ mov KEYP, TKEYP
+ pxor KEY, STATE # round 0
+ add $0x30, TKEYP
+ cmp $12, NRND
+ jb .Ldec128
+ lea 0x20(TKEYP), TKEYP
+ je .Ldec192
+ add $0x20, TKEYP
+ movaps -0x60(TKEYP), KEY
+ # aesdec KEY, STATE
+ .byte 0x66, 0x0f, 0x38, 0xde, 0xc2
+ movaps -0x50(TKEYP), KEY
+ # aesdec KEY, STATE
+ .byte 0x66, 0x0f, 0x38, 0xde, 0xc2
+.align 4
+.Ldec192:
+ movaps -0x40(TKEYP), KEY
+ # aesdec KEY, STATE
+ .byte 0x66, 0x0f, 0x38, 0xde, 0xc2
+ movaps -0x30(TKEYP), KEY
+ # aesdec KEY, STATE
+ .byte 0x66, 0x0f, 0x38, 0xde, 0xc2
+.align 4
+.Ldec128:
+ movaps -0x20(TKEYP), KEY
+ # aesdec KEY, STATE
+ .byte 0x66, 0x0f, 0x38, 0xde, 0xc2
+ movaps -0x10(TKEYP), KEY
+ # aesdec KEY, STATE
+ .byte 0x66, 0x0f, 0x38, 0xde, 0xc2
+ movaps (TKEYP), KEY
+ # aesdec KEY, STATE
+ .byte 0x66, 0x0f, 0x38, 0xde, 0xc2
+ movaps 0x10(TKEYP), KEY
+ # aesdec KEY, STATE
+ .byte 0x66, 0x0f, 0x38, 0xde, 0xc2
+ movaps 0x20(TKEYP), KEY
+ # aesdec KEY, STATE
+ .byte 0x66, 0x0f, 0x38, 0xde, 0xc2
+ movaps 0x30(TKEYP), KEY
+ # aesdec KEY, STATE
+ .byte 0x66, 0x0f, 0x38, 0xde, 0xc2
+ movaps 0x40(TKEYP), KEY
+ # aesdec KEY, STATE
+ .byte 0x66, 0x0f, 0x38, 0xde, 0xc2
+ movaps 0x50(TKEYP), KEY
+ # aesdec KEY, STATE
+ .byte 0x66, 0x0f, 0x38, 0xde, 0xc2
+ movaps 0x60(TKEYP), KEY
+ # aesdec KEY, STATE
+ .byte 0x66, 0x0f, 0x38, 0xde, 0xc2
+ movaps 0x70(TKEYP), KEY
+ # aesdeclast KEY, STATE # last round
+ .byte 0x66, 0x0f, 0x38, 0xdf, 0xc2
+
+ pxor IV, STATE
+ movups STATE, (OUTP)
+ movaps IN, IV
+ sub $16, LEN
+ add $16, INP
+ add $16, OUTP
+ cmp $16, LEN
+ jge .Lcbc_dec_loop1
+.Lcbc_dec_ret:
+ emms
+ pop %ebp
+ pop %ebx
+ pop %esi
+ pop %edi
+ ret
+
+/*
+ * void intel_aes_cbc_enc_4x512(aes_context **, void *src, void *dst, void *iv)
+ *
+ * Stack after reg saves: 32(%esp) = void *iv
+ * 28(%esp) = void *dst
+ * 24(%esp) = void *src
+ * 20(%esp) = aes_context **
+ */
+ .align ALIGN64BYTES
+intel_aes_cbc_enc_4x512:
+ push %edi
+ push %esi
+ push %ebx
+ push %ebp
+ mov 20(%esp),%edi
+ mov 32(%esp),%esi
+ mov 24(%esp),%ebp
+ mov (%edi),%eax # pointer to context struct 1
+ mov 4(%edi),%ebx # pointer to context struct 2
+ mov 8(%edi),%ecx # pointer to context struct 3
+ mov 12(%edi),%edx # pointer to context struct 4
+ mov 4(%eax),%edi # number of rounds (10/12/14)
+ movups (%esi),%xmm0 # load IV as initial state
+ movups 0x10(%esi),%xmm1
+ movups 0x20(%esi),%xmm2
+ movups 0x30(%esi),%xmm3
+ sub $10,%edi
+ mov $0x200,%esi # 512 byte CBC chain
+ shl $4,%edi
+ add $0x38,%edi # 0x38 / 0x58 / 0x78
+.align 4
+.Lcbc_enc_loop4:
+ movups (%ebp),%xmm4 # load input
+ movups 0x200(%ebp),%xmm5
+ movups 0x400(%ebp),%xmm6
+ movups 0x600(%ebp),%xmm7
+ add $16,%ebp
+ mov %ebp,24(%esp)
+ mov 28(%esp),%ebp
+ pxor %xmm4,%xmm0 # CBC-mode XOR
+ pxor %xmm5,%xmm1
+ pxor %xmm6,%xmm2
+ pxor %xmm7,%xmm3
+
+ movaps 0x08(%eax),%xmm4 # round 0 key
+ movaps 0x08(%ebx),%xmm5
+ movaps 0x08(%ecx),%xmm6
+ movaps 0x08(%edx),%xmm7
+ pxor %xmm4,%xmm0 # round 0 XOR
+ pxor %xmm5,%xmm1
+ pxor %xmm6,%xmm2
+ pxor %xmm7,%xmm3
+
+ cmp $0x58,%edi
+ jb .L4enc128
+ je .L4enc192
+
+ movaps -0x60(%eax,%edi,1),%xmm4
+ movaps -0x60(%ebx,%edi,1),%xmm5
+ movaps -0x60(%ecx,%edi,1),%xmm6
+ movaps -0x60(%edx,%edi,1),%xmm7
+ # aesenc %xmm4,%xmm0
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xc4
+ # aesenc %xmm5,%xmm1
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xcd
+ # aesenc %xmm6,%xmm2
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xd6
+ # aesenc %xmm7,%xmm3
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xdf
+
+ movaps -0x50(%eax,%edi,1),%xmm4
+ movaps -0x50(%ebx,%edi,1),%xmm5
+ movaps -0x50(%ecx,%edi,1),%xmm6
+ movaps -0x50(%edx,%edi,1),%xmm7
+ # aesenc %xmm4,%xmm0
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xc4
+ # aesenc %xmm5,%xmm1
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xcd
+ # aesenc %xmm6,%xmm2
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xd6
+ # aesenc %xmm7,%xmm3
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xdf
+
+.align 4
+.L4enc192:
+ movaps -0x40(%eax,%edi,1),%xmm4
+ movaps -0x40(%ebx,%edi,1),%xmm5
+ movaps -0x40(%ecx,%edi,1),%xmm6
+ movaps -0x40(%edx,%edi,1),%xmm7
+ # aesenc %xmm4,%xmm0
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xc4
+ # aesenc %xmm5,%xmm1
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xcd
+ # aesenc %xmm6,%xmm2
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xd6
+ # aesenc %xmm7,%xmm3
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xdf
+
+ movaps -0x30(%eax,%edi,1),%xmm4
+ movaps -0x30(%ebx,%edi,1),%xmm5
+ movaps -0x30(%ecx,%edi,1),%xmm6
+ movaps -0x30(%edx,%edi,1),%xmm7
+ # aesenc %xmm4,%xmm0
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xc4
+ # aesenc %xmm5,%xmm1
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xcd
+ # aesenc %xmm6,%xmm2
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xd6
+ # aesenc %xmm7,%xmm3
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xdf
+
+.align 4
+.L4enc128:
+ movaps -0x20(%eax,%edi,1),%xmm4
+ movaps -0x20(%ebx,%edi,1),%xmm5
+ movaps -0x20(%ecx,%edi,1),%xmm6
+ movaps -0x20(%edx,%edi,1),%xmm7
+ # aesenc %xmm4,%xmm0
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xc4
+ # aesenc %xmm5,%xmm1
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xcd
+ # aesenc %xmm6,%xmm2
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xd6
+ # aesenc %xmm7,%xmm3
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xdf
+
+ movaps -0x10(%eax,%edi,1),%xmm4
+ movaps -0x10(%ebx,%edi,1),%xmm5
+ movaps -0x10(%ecx,%edi,1),%xmm6
+ movaps -0x10(%edx,%edi,1),%xmm7
+ # aesenc %xmm4,%xmm0
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xc4
+ # aesenc %xmm5,%xmm1
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xcd
+ # aesenc %xmm6,%xmm2
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xd6
+ # aesenc %xmm7,%xmm3
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xdf
+
+ movaps (%eax,%edi,1),%xmm4
+ movaps (%ebx,%edi,1),%xmm5
+ movaps (%ecx,%edi,1),%xmm6
+ movaps (%edx,%edi,1),%xmm7
+ # aesenc %xmm4,%xmm0
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xc4
+ # aesenc %xmm5,%xmm1
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xcd
+ # aesenc %xmm6,%xmm2
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xd6
+ # aesenc %xmm7,%xmm3
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xdf
+
+ movaps 0x10(%eax,%edi,1),%xmm4
+ movaps 0x10(%ebx,%edi,1),%xmm5
+ movaps 0x10(%ecx,%edi,1),%xmm6
+ movaps 0x10(%edx,%edi,1),%xmm7
+ # aesenc %xmm4,%xmm0
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xc4
+ # aesenc %xmm5,%xmm1
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xcd
+ # aesenc %xmm6,%xmm2
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xd6
+ # aesenc %xmm7,%xmm3
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xdf
+
+ movaps 0x20(%eax,%edi,1),%xmm4
+ movaps 0x20(%ebx,%edi,1),%xmm5
+ movaps 0x20(%ecx,%edi,1),%xmm6
+ movaps 0x20(%edx,%edi,1),%xmm7
+ # aesenc %xmm4,%xmm0
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xc4
+ # aesenc %xmm5,%xmm1
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xcd
+ # aesenc %xmm6,%xmm2
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xd6
+ # aesenc %xmm7,%xmm3
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xdf
+
+ movaps 0x30(%eax,%edi,1),%xmm4
+ movaps 0x30(%ebx,%edi,1),%xmm5
+ movaps 0x30(%ecx,%edi,1),%xmm6
+ movaps 0x30(%edx,%edi,1),%xmm7
+ # aesenc %xmm4,%xmm0
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xc4
+ # aesenc %xmm5,%xmm1
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xcd
+ # aesenc %xmm6,%xmm2
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xd6
+ # aesenc %xmm7,%xmm3
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xdf
+
+ movaps 0x40(%eax,%edi,1),%xmm4
+ movaps 0x40(%ebx,%edi,1),%xmm5
+ movaps 0x40(%ecx,%edi,1),%xmm6
+ movaps 0x40(%edx,%edi,1),%xmm7
+ # aesenc %xmm4,%xmm0
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xc4
+ # aesenc %xmm5,%xmm1
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xcd
+ # aesenc %xmm6,%xmm2
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xd6
+ # aesenc %xmm7,%xmm3
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xdf
+
+ movaps 0x50(%eax,%edi,1),%xmm4
+ movaps 0x50(%ebx,%edi,1),%xmm5
+ movaps 0x50(%ecx,%edi,1),%xmm6
+ movaps 0x50(%edx,%edi,1),%xmm7
+ # aesenc %xmm4,%xmm0
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xc4
+ # aesenc %xmm5,%xmm1
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xcd
+ # aesenc %xmm6,%xmm2
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xd6
+ # aesenc %xmm7,%xmm3
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xdf
+
+ movaps 0x60(%eax,%edi,1),%xmm4
+ movaps 0x60(%ebx,%edi,1),%xmm5
+ movaps 0x60(%ecx,%edi,1),%xmm6
+ movaps 0x60(%edx,%edi,1),%xmm7
+ # aesenc %xmm4,%xmm0
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xc4
+ # aesenc %xmm5,%xmm1
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xcd
+ # aesenc %xmm6,%xmm2
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xd6
+ # aesenc %xmm7,%xmm3
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xdf
+
+ movaps 0x70(%eax,%edi,1),%xmm4
+ movaps 0x70(%ebx,%edi,1),%xmm5
+ movaps 0x70(%ecx,%edi,1),%xmm6
+ movaps 0x70(%edx,%edi,1),%xmm7
+ # aesenclast %xmm4,%xmm0 # last round
+ .byte 0x66, 0x0f, 0x38, 0xdd, 0xc4
+ # aesenclast %xmm5,%xmm1
+ .byte 0x66, 0x0f, 0x38, 0xdd, 0xcd
+ # aesenclast %xmm6,%xmm2
+ .byte 0x66, 0x0f, 0x38, 0xdd, 0xd6
+ # aesenclast %xmm7,%xmm3
+ .byte 0x66, 0x0f, 0x38, 0xdd, 0xdf
+
+ sub $16,%esi
+ movups %xmm0,(%ebp) # store output
+ movups %xmm1,0x200(%ebp)
+ movups %xmm2,0x400(%ebp)
+ movups %xmm3,0x600(%ebp)
+ add $16,%ebp
+ mov %ebp,28(%esp)
+ mov 24(%esp),%ebp
+ cmp $16,%esi
+ jge .Lcbc_enc_loop4
+ emms
+ pop %ebp
+ pop %ebx
+ pop %esi
+ pop %edi
+ ret
+
+#if defined(__ELF__) && defined(SECTION_NOTE_GNU_STACK)
+ .section .note.GNU-stack,"",@progbits
+#endif
diff -urN linux-3.14-noloop/drivers/misc/aes-intel64.S linux-3.14-AES/drivers/misc/aes-intel64.S
--- linux-3.14-noloop/drivers/misc/aes-intel64.S 1970-01-01 02:00:00.000000000 +0200
+++ linux-3.14-AES/drivers/misc/aes-intel64.S 2014-03-31 22:20:36.000000000 +0300
@@ -0,0 +1,649 @@
+/*
+ * Implement AES algorithm in Intel AES-NI instructions.
+ *
+ * The white paper of AES-NI instructions can be downloaded from:
+ * http://softwarecommunity.intel.com/isn/downloads/intelavx/AES-Instructions-Set_WP.pdf
+ *
+ * Copyright (C) 2008, Intel Corp.
+ * Author: Huang Ying <ying.huang@intel.com>
+ * Vinodh Gopal <vinodh.gopal@intel.com>
+ * Kahraman Akdemir
+ *
+ * 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.
+ */
+
+/*
+ * Modified by Jari Ruusu, October 2009
+ * - Adapted for loop-AES
+ */
+
+/*
+ * Modified by Jari Ruusu, March 2010
+ * - Added parallelized 4x512 CBC encrypt
+ */
+
+#if !defined(ALIGN64BYTES)
+# define ALIGN64BYTES 64
+#endif
+
+ .file "aes-intel64.S"
+ .globl intel_aes_cbc_encrypt
+ .globl intel_aes_cbc_decrypt
+ .globl intel_aes_cbc_enc_4x512
+ .text
+
+#define STATE1 %xmm0
+#define STATE2 %xmm4
+#define STATE3 %xmm5
+#define STATE4 %xmm6
+#define STATE STATE1
+#define IN1 %xmm1
+#define IN2 %xmm7
+#define IN3 %xmm8
+#define IN4 %xmm9
+#define IN IN1
+#define KEY %xmm2
+#define IV %xmm3
+
+#define KEYP %rdi
+#define INP %rsi
+#define OUTP %rdx
+#define LEN %rcx
+#define IVP %r8
+#define NRND %r9d
+#define TKEYP %r10
+
+/*
+ * void intel_aes_cbc_encrypt(const aes_context *, void *src, void *dst, size_t len, void *iv)
+ *
+ * Parameters: %rdi = aes_context *
+ * %rsi = void *src
+ * %rdx = void *dst
+ * %rcx = size_t len
+ * %r8 = void *iv
+ */
+ .align ALIGN64BYTES
+intel_aes_cbc_encrypt:
+ mov 4(KEYP), NRND
+ add $8, KEYP
+ movups (IVP), STATE # load iv as initial state
+.align 4
+.Lcbc_enc_loop:
+ movups (INP), IN # load input
+ pxor IN, STATE
+
+ movaps (KEYP), KEY # key
+ mov KEYP, TKEYP
+ pxor KEY, STATE # round 0
+ add $0x30, TKEYP
+ cmp $12, NRND
+ jb .Lenc128
+ lea 0x20(TKEYP), TKEYP
+ je .Lenc192
+ add $0x20, TKEYP
+ movaps -0x60(TKEYP), KEY
+ # aesenc KEY, STATE
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xc2
+ movaps -0x50(TKEYP), KEY
+ # aesenc KEY, STATE
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xc2
+.align 4
+.Lenc192:
+ movaps -0x40(TKEYP), KEY
+ # aesenc KEY, STATE
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xc2
+ movaps -0x30(TKEYP), KEY
+ # aesenc KEY, STATE
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xc2
+.align 4
+.Lenc128:
+ movaps -0x20(TKEYP), KEY
+ # aesenc KEY, STATE
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xc2
+ movaps -0x10(TKEYP), KEY
+ # aesenc KEY, STATE
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xc2
+ movaps (TKEYP), KEY
+ # aesenc KEY, STATE
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xc2
+ movaps 0x10(TKEYP), KEY
+ # aesenc KEY, STATE
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xc2
+ movaps 0x20(TKEYP), KEY
+ # aesenc KEY, STATE
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xc2
+ movaps 0x30(TKEYP), KEY
+ # aesenc KEY, STATE
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xc2
+ movaps 0x40(TKEYP), KEY
+ # aesenc KEY, STATE
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xc2
+ movaps 0x50(TKEYP), KEY
+ # aesenc KEY, STATE
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xc2
+ movaps 0x60(TKEYP), KEY
+ # aesenc KEY, STATE
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xc2
+ movaps 0x70(TKEYP), KEY
+ # aesenclast KEY, STATE # last round
+ .byte 0x66, 0x0f, 0x38, 0xdd, 0xc2
+
+ movups STATE, (OUTP) # store output
+ sub $16, LEN
+ add $16, INP
+ add $16, OUTP
+ cmp $16, LEN
+ jge .Lcbc_enc_loop
+ emms
+ ret
+
+/*
+ * void intel_aes_cbc_decrypt(const aes_context *, void *src, void *dst, size_t len, void *iv)
+ *
+ * Parameters: %rdi = aes_context *
+ * %rsi = void *src
+ * %rdx = void *dst
+ * %rcx = size_t len
+ * %r8 = void *iv
+ */
+ .align ALIGN64BYTES
+intel_aes_cbc_decrypt:
+ mov 4(KEYP), NRND
+ add $264, KEYP
+ movups (IVP), IV
+ cmp $64, LEN
+ jb .Lcbc_dec_loop1
+.align 4
+.Lcbc_dec_loop4:
+ movups (INP), IN1
+ movaps IN1, STATE1
+ movups 0x10(INP), IN2
+ movaps IN2, STATE2
+ movups 0x20(INP), IN3
+ movaps IN3, STATE3
+ movups 0x30(INP), IN4
+ movaps IN4, STATE4
+
+ movaps (KEYP), KEY # key
+ mov KEYP, TKEYP
+ pxor KEY, STATE1 # round 0
+ pxor KEY, STATE2
+ pxor KEY, STATE3
+ pxor KEY, STATE4
+ add $0x30, TKEYP
+ cmp $12, NRND
+ jb .L4dec128
+ lea 0x20(TKEYP), TKEYP
+ je .L4dec192
+ add $0x20, TKEYP
+ movaps -0x60(TKEYP), KEY
+ # aesdec KEY, STATE1
+ .byte 0x66, 0x0f, 0x38, 0xde, 0xc2
+ # aesdec KEY, STATE2
+ .byte 0x66, 0x0f, 0x38, 0xde, 0xe2
+ # aesdec KEY, STATE3
+ .byte 0x66, 0x0f, 0x38, 0xde, 0xea
+ # aesdec KEY, STATE4
+ .byte 0x66, 0x0f, 0x38, 0xde, 0xf2
+ movaps -0x50(TKEYP), KEY
+ # aesdec KEY, STATE1
+ .byte 0x66, 0x0f, 0x38, 0xde, 0xc2
+ # aesdec KEY, STATE2
+ .byte 0x66, 0x0f, 0x38, 0xde, 0xe2
+ # aesdec KEY, STATE3
+ .byte 0x66, 0x0f, 0x38, 0xde, 0xea
+ # aesdec KEY, STATE4
+ .byte 0x66, 0x0f, 0x38, 0xde, 0xf2
+.align 4
+.L4dec192:
+ movaps -0x40(TKEYP), KEY
+ # aesdec KEY, STATE1
+ .byte 0x66, 0x0f, 0x38, 0xde, 0xc2
+ # aesdec KEY, STATE2
+ .byte 0x66, 0x0f, 0x38, 0xde, 0xe2
+ # aesdec KEY, STATE3
+ .byte 0x66, 0x0f, 0x38, 0xde, 0xea
+ # aesdec KEY, STATE4
+ .byte 0x66, 0x0f, 0x38, 0xde, 0xf2
+ movaps -0x30(TKEYP), KEY
+ # aesdec KEY, STATE1
+ .byte 0x66, 0x0f, 0x38, 0xde, 0xc2
+ # aesdec KEY, STATE2
+ .byte 0x66, 0x0f, 0x38, 0xde, 0xe2
+ # aesdec KEY, STATE3
+ .byte 0x66, 0x0f, 0x38, 0xde, 0xea
+ # aesdec KEY, STATE4
+ .byte 0x66, 0x0f, 0x38, 0xde, 0xf2
+.align 4
+.L4dec128:
+ movaps -0x20(TKEYP), KEY
+ # aesdec KEY, STATE1
+ .byte 0x66, 0x0f, 0x38, 0xde, 0xc2
+ # aesdec KEY, STATE2
+ .byte 0x66, 0x0f, 0x38, 0xde, 0xe2
+ # aesdec KEY, STATE3
+ .byte 0x66, 0x0f, 0x38, 0xde, 0xea
+ # aesdec KEY, STATE4
+ .byte 0x66, 0x0f, 0x38, 0xde, 0xf2
+ movaps -0x10(TKEYP), KEY
+ # aesdec KEY, STATE1
+ .byte 0x66, 0x0f, 0x38, 0xde, 0xc2
+ # aesdec KEY, STATE2
+ .byte 0x66, 0x0f, 0x38, 0xde, 0xe2
+ # aesdec KEY, STATE3
+ .byte 0x66, 0x0f, 0x38, 0xde, 0xea
+ # aesdec KEY, STATE4
+ .byte 0x66, 0x0f, 0x38, 0xde, 0xf2
+ movaps (TKEYP), KEY
+ # aesdec KEY, STATE1
+ .byte 0x66, 0x0f, 0x38, 0xde, 0xc2
+ # aesdec KEY, STATE2
+ .byte 0x66, 0x0f, 0x38, 0xde, 0xe2
+ # aesdec KEY, STATE3
+ .byte 0x66, 0x0f, 0x38, 0xde, 0xea
+ # aesdec KEY, STATE4
+ .byte 0x66, 0x0f, 0x38, 0xde, 0xf2
+ movaps 0x10(TKEYP), KEY
+ # aesdec KEY, STATE1
+ .byte 0x66, 0x0f, 0x38, 0xde, 0xc2
+ # aesdec KEY, STATE2
+ .byte 0x66, 0x0f, 0x38, 0xde, 0xe2
+ # aesdec KEY, STATE3
+ .byte 0x66, 0x0f, 0x38, 0xde, 0xea
+ # aesdec KEY, STATE4
+ .byte 0x66, 0x0f, 0x38, 0xde, 0xf2
+ movaps 0x20(TKEYP), KEY
+ # aesdec KEY, STATE1
+ .byte 0x66, 0x0f, 0x38, 0xde, 0xc2
+ # aesdec KEY, STATE2
+ .byte 0x66, 0x0f, 0x38, 0xde, 0xe2
+ # aesdec KEY, STATE3
+ .byte 0x66, 0x0f, 0x38, 0xde, 0xea
+ # aesdec KEY, STATE4
+ .byte 0x66, 0x0f, 0x38, 0xde, 0xf2
+ movaps 0x30(TKEYP), KEY
+ # aesdec KEY, STATE1
+ .byte 0x66, 0x0f, 0x38, 0xde, 0xc2
+ # aesdec KEY, STATE2
+ .byte 0x66, 0x0f, 0x38, 0xde, 0xe2
+ # aesdec KEY, STATE3
+ .byte 0x66, 0x0f, 0x38, 0xde, 0xea
+ # aesdec KEY, STATE4
+ .byte 0x66, 0x0f, 0x38, 0xde, 0xf2
+ movaps 0x40(TKEYP), KEY
+ # aesdec KEY, STATE1
+ .byte 0x66, 0x0f, 0x38, 0xde, 0xc2
+ # aesdec KEY, STATE2
+ .byte 0x66, 0x0f, 0x38, 0xde, 0xe2
+ # aesdec KEY, STATE3
+ .byte 0x66, 0x0f, 0x38, 0xde, 0xea
+ # aesdec KEY, STATE4
+ .byte 0x66, 0x0f, 0x38, 0xde, 0xf2
+ movaps 0x50(TKEYP), KEY
+ # aesdec KEY, STATE1
+ .byte 0x66, 0x0f, 0x38, 0xde, 0xc2
+ # aesdec KEY, STATE2
+ .byte 0x66, 0x0f, 0x38, 0xde, 0xe2
+ # aesdec KEY, STATE3
+ .byte 0x66, 0x0f, 0x38, 0xde, 0xea
+ # aesdec KEY, STATE4
+ .byte 0x66, 0x0f, 0x38, 0xde, 0xf2
+ movaps 0x60(TKEYP), KEY
+ # aesdec KEY, STATE1
+ .byte 0x66, 0x0f, 0x38, 0xde, 0xc2
+ # aesdec KEY, STATE2
+ .byte 0x66, 0x0f, 0x38, 0xde, 0xe2
+ # aesdec KEY, STATE3
+ .byte 0x66, 0x0f, 0x38, 0xde, 0xea
+ # aesdec KEY, STATE4
+ .byte 0x66, 0x0f, 0x38, 0xde, 0xf2
+ movaps 0x70(TKEYP), KEY
+ # aesdeclast KEY, STATE1 # last round
+ .byte 0x66, 0x0f, 0x38, 0xdf, 0xc2
+ # aesdeclast KEY, STATE2
+ .byte 0x66, 0x0f, 0x38, 0xdf, 0xe2
+ # aesdeclast KEY, STATE3
+ .byte 0x66, 0x0f, 0x38, 0xdf, 0xea
+ # aesdeclast KEY, STATE4
+ .byte 0x66, 0x0f, 0x38, 0xdf, 0xf2
+
+ pxor IV, STATE1
+ pxor IN1, STATE2
+ pxor IN2, STATE3
+ pxor IN3, STATE4
+ movaps IN4, IV
+ movups STATE1, (OUTP)
+ movups STATE2, 0x10(OUTP)
+ movups STATE3, 0x20(OUTP)
+ movups STATE4, 0x30(OUTP)
+ sub $64, LEN
+ add $64, INP
+ add $64, OUTP
+ cmp $64, LEN
+ jge .Lcbc_dec_loop4
+ cmp $16, LEN
+ jb .Lcbc_dec_ret
+.align 4
+.Lcbc_dec_loop1:
+ movups (INP), IN
+ movaps IN, STATE
+
+ movaps (KEYP), KEY # key
+ mov KEYP, TKEYP
+ pxor KEY, STATE # round 0
+ add $0x30, TKEYP
+ cmp $12, NRND
+ jb .Ldec128
+ lea 0x20(TKEYP), TKEYP
+ je .Ldec192
+ add $0x20, TKEYP
+ movaps -0x60(TKEYP), KEY
+ # aesdec KEY, STATE
+ .byte 0x66, 0x0f, 0x38, 0xde, 0xc2
+ movaps -0x50(TKEYP), KEY
+ # aesdec KEY, STATE
+ .byte 0x66, 0x0f, 0x38, 0xde, 0xc2
+.align 4
+.Ldec192:
+ movaps -0x40(TKEYP), KEY
+ # aesdec KEY, STATE
+ .byte 0x66, 0x0f, 0x38, 0xde, 0xc2
+ movaps -0x30(TKEYP), KEY
+ # aesdec KEY, STATE
+ .byte 0x66, 0x0f, 0x38, 0xde, 0xc2
+.align 4
+.Ldec128:
+ movaps -0x20(TKEYP), KEY
+ # aesdec KEY, STATE
+ .byte 0x66, 0x0f, 0x38, 0xde, 0xc2
+ movaps -0x10(TKEYP), KEY
+ # aesdec KEY, STATE
+ .byte 0x66, 0x0f, 0x38, 0xde, 0xc2
+ movaps (TKEYP), KEY
+ # aesdec KEY, STATE
+ .byte 0x66, 0x0f, 0x38, 0xde, 0xc2
+ movaps 0x10(TKEYP), KEY
+ # aesdec KEY, STATE
+ .byte 0x66, 0x0f, 0x38, 0xde, 0xc2
+ movaps 0x20(TKEYP), KEY
+ # aesdec KEY, STATE
+ .byte 0x66, 0x0f, 0x38, 0xde, 0xc2
+ movaps 0x30(TKEYP), KEY
+ # aesdec KEY, STATE
+ .byte 0x66, 0x0f, 0x38, 0xde, 0xc2
+ movaps 0x40(TKEYP), KEY
+ # aesdec KEY, STATE
+ .byte 0x66, 0x0f, 0x38, 0xde, 0xc2
+ movaps 0x50(TKEYP), KEY
+ # aesdec KEY, STATE
+ .byte 0x66, 0x0f, 0x38, 0xde, 0xc2
+ movaps 0x60(TKEYP), KEY
+ # aesdec KEY, STATE
+ .byte 0x66, 0x0f, 0x38, 0xde, 0xc2
+ movaps 0x70(TKEYP), KEY
+ # aesdeclast KEY, STATE # last round
+ .byte 0x66, 0x0f, 0x38, 0xdf, 0xc2
+
+ pxor IV, STATE
+ movups STATE, (OUTP)
+ movaps IN, IV
+ sub $16, LEN
+ add $16, INP
+ add $16, OUTP
+ cmp $16, LEN
+ jge .Lcbc_dec_loop1
+.Lcbc_dec_ret:
+ emms
+ ret
+
+/*
+ * void intel_aes_cbc_enc_4x512(aes_context **, void *src, void *dst, void *iv)
+ *
+ * Parameters: %rdi = aes_context **
+ * %rsi = void *src
+ * %rdx = void *dst
+ * %rcx = void *iv
+ */
+ .align ALIGN64BYTES
+intel_aes_cbc_enc_4x512:
+ mov (%rdi),%rax # pointer to context struct 1
+ mov 8(%rdi),%r8 # pointer to context struct 2
+ mov 16(%rdi),%r9 # pointer to context struct 3
+ mov 24(%rdi),%r10 # pointer to context struct 4
+ mov 4(%rax),%edi # number of rounds (10/12/14)
+ movups (%rcx),%xmm0 # load IV as initial state
+ movups 0x10(%rcx),%xmm1
+ movups 0x20(%rcx),%xmm2
+ movups 0x30(%rcx),%xmm3
+ sub $10,%edi
+ mov $0x200,%ecx # 512 byte CBC chain
+ shl $4,%edi
+ add $0x38,%edi # 0x38 / 0x58 / 0x78
+.align 4
+.Lcbc_enc_loop4:
+ movups (%rsi),%xmm4 # load input
+ movups 0x200(%rsi),%xmm5
+ movups 0x400(%rsi),%xmm6
+ movups 0x600(%rsi),%xmm7
+ add $16,%rsi
+ pxor %xmm4,%xmm0 # CBC-mode XOR
+ pxor %xmm5,%xmm1
+ pxor %xmm6,%xmm2
+ pxor %xmm7,%xmm3
+
+ movaps 0x08(%rax),%xmm4 # round 0 key
+ movaps 0x08(%r8),%xmm5
+ movaps 0x08(%r9),%xmm6
+ movaps 0x08(%r10),%xmm7
+ pxor %xmm4,%xmm0 # round 0 XOR
+ pxor %xmm5,%xmm1
+ pxor %xmm6,%xmm2
+ pxor %xmm7,%xmm3
+
+ cmp $0x58,%edi
+ jb .L4enc128
+ je .L4enc192
+
+ movaps -0x60(%rax,%rdi,1),%xmm4
+ movaps -0x60(%r8,%rdi,1),%xmm5
+ movaps -0x60(%r9,%rdi,1),%xmm6
+ movaps -0x60(%r10,%rdi,1),%xmm7
+ # aesenc %xmm4,%xmm0
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xc4
+ # aesenc %xmm5,%xmm1
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xcd
+ # aesenc %xmm6,%xmm2
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xd6
+ # aesenc %xmm7,%xmm3
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xdf
+
+ movaps -0x50(%rax,%rdi,1),%xmm4
+ movaps -0x50(%r8,%rdi,1),%xmm5
+ movaps -0x50(%r9,%rdi,1),%xmm6
+ movaps -0x50(%r10,%rdi,1),%xmm7
+ # aesenc %xmm4,%xmm0
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xc4
+ # aesenc %xmm5,%xmm1
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xcd
+ # aesenc %xmm6,%xmm2
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xd6
+ # aesenc %xmm7,%xmm3
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xdf
+
+.align 4
+.L4enc192:
+ movaps -0x40(%rax,%rdi,1),%xmm4
+ movaps -0x40(%r8,%rdi,1),%xmm5
+ movaps -0x40(%r9,%rdi,1),%xmm6
+ movaps -0x40(%r10,%rdi,1),%xmm7
+ # aesenc %xmm4,%xmm0
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xc4
+ # aesenc %xmm5,%xmm1
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xcd
+ # aesenc %xmm6,%xmm2
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xd6
+ # aesenc %xmm7,%xmm3
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xdf
+
+ movaps -0x30(%rax,%rdi,1),%xmm4
+ movaps -0x30(%r8,%rdi,1),%xmm5
+ movaps -0x30(%r9,%rdi,1),%xmm6
+ movaps -0x30(%r10,%rdi,1),%xmm7
+ # aesenc %xmm4,%xmm0
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xc4
+ # aesenc %xmm5,%xmm1
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xcd
+ # aesenc %xmm6,%xmm2
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xd6
+ # aesenc %xmm7,%xmm3
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xdf
+
+.align 4
+.L4enc128:
+ movaps -0x20(%rax,%rdi,1),%xmm4
+ movaps -0x20(%r8,%rdi,1),%xmm5
+ movaps -0x20(%r9,%rdi,1),%xmm6
+ movaps -0x20(%r10,%rdi,1),%xmm7
+ # aesenc %xmm4,%xmm0
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xc4
+ # aesenc %xmm5,%xmm1
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xcd
+ # aesenc %xmm6,%xmm2
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xd6
+ # aesenc %xmm7,%xmm3
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xdf
+
+ movaps -0x10(%rax,%rdi,1),%xmm4
+ movaps -0x10(%r8,%rdi,1),%xmm5
+ movaps -0x10(%r9,%rdi,1),%xmm6
+ movaps -0x10(%r10,%rdi,1),%xmm7
+ # aesenc %xmm4,%xmm0
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xc4
+ # aesenc %xmm5,%xmm1
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xcd
+ # aesenc %xmm6,%xmm2
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xd6
+ # aesenc %xmm7,%xmm3
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xdf
+
+ movaps (%rax,%rdi,1),%xmm4
+ movaps (%r8,%rdi,1),%xmm5
+ movaps (%r9,%rdi,1),%xmm6
+ movaps (%r10,%rdi,1),%xmm7
+ # aesenc %xmm4,%xmm0
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xc4
+ # aesenc %xmm5,%xmm1
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xcd
+ # aesenc %xmm6,%xmm2
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xd6
+ # aesenc %xmm7,%xmm3
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xdf
+
+ movaps 0x10(%rax,%rdi,1),%xmm4
+ movaps 0x10(%r8,%rdi,1),%xmm5
+ movaps 0x10(%r9,%rdi,1),%xmm6
+ movaps 0x10(%r10,%rdi,1),%xmm7
+ # aesenc %xmm4,%xmm0
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xc4
+ # aesenc %xmm5,%xmm1
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xcd
+ # aesenc %xmm6,%xmm2
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xd6
+ # aesenc %xmm7,%xmm3
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xdf
+
+ movaps 0x20(%rax,%rdi,1),%xmm4
+ movaps 0x20(%r8,%rdi,1),%xmm5
+ movaps 0x20(%r9,%rdi,1),%xmm6
+ movaps 0x20(%r10,%rdi,1),%xmm7
+ # aesenc %xmm4,%xmm0
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xc4
+ # aesenc %xmm5,%xmm1
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xcd
+ # aesenc %xmm6,%xmm2
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xd6
+ # aesenc %xmm7,%xmm3
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xdf
+
+ movaps 0x30(%rax,%rdi,1),%xmm4
+ movaps 0x30(%r8,%rdi,1),%xmm5
+ movaps 0x30(%r9,%rdi,1),%xmm6
+ movaps 0x30(%r10,%rdi,1),%xmm7
+ # aesenc %xmm4,%xmm0
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xc4
+ # aesenc %xmm5,%xmm1
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xcd
+ # aesenc %xmm6,%xmm2
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xd6
+ # aesenc %xmm7,%xmm3
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xdf
+
+ movaps 0x40(%rax,%rdi,1),%xmm4
+ movaps 0x40(%r8,%rdi,1),%xmm5
+ movaps 0x40(%r9,%rdi,1),%xmm6
+ movaps 0x40(%r10,%rdi,1),%xmm7
+ # aesenc %xmm4,%xmm0
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xc4
+ # aesenc %xmm5,%xmm1
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xcd
+ # aesenc %xmm6,%xmm2
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xd6
+ # aesenc %xmm7,%xmm3
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xdf
+
+ movaps 0x50(%rax,%rdi,1),%xmm4
+ movaps 0x50(%r8,%rdi,1),%xmm5
+ movaps 0x50(%r9,%rdi,1),%xmm6
+ movaps 0x50(%r10,%rdi,1),%xmm7
+ # aesenc %xmm4,%xmm0
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xc4
+ # aesenc %xmm5,%xmm1
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xcd
+ # aesenc %xmm6,%xmm2
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xd6
+ # aesenc %xmm7,%xmm3
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xdf
+
+ movaps 0x60(%rax,%rdi,1),%xmm4
+ movaps 0x60(%r8,%rdi,1),%xmm5
+ movaps 0x60(%r9,%rdi,1),%xmm6
+ movaps 0x60(%r10,%rdi,1),%xmm7
+ # aesenc %xmm4,%xmm0
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xc4
+ # aesenc %xmm5,%xmm1
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xcd
+ # aesenc %xmm6,%xmm2
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xd6
+ # aesenc %xmm7,%xmm3
+ .byte 0x66, 0x0f, 0x38, 0xdc, 0xdf
+
+ movaps 0x70(%rax,%rdi,1),%xmm4
+ movaps 0x70(%r8,%rdi,1),%xmm5
+ movaps 0x70(%r9,%rdi,1),%xmm6
+ movaps 0x70(%r10,%rdi,1),%xmm7
+ # aesenclast %xmm4,%xmm0 # last round
+ .byte 0x66, 0x0f, 0x38, 0xdd, 0xc4
+ # aesenclast %xmm5,%xmm1
+ .byte 0x66, 0x0f, 0x38, 0xdd, 0xcd
+ # aesenclast %xmm6,%xmm2
+ .byte 0x66, 0x0f, 0x38, 0xdd, 0xd6
+ # aesenclast %xmm7,%xmm3
+ .byte 0x66, 0x0f, 0x38, 0xdd, 0xdf
+
+ sub $16,%ecx
+ movups %xmm0,(%rdx) # store output
+ movups %xmm1,0x200(%rdx)
+ movups %xmm2,0x400(%rdx)
+ movups %xmm3,0x600(%rdx)
+ add $16,%rdx
+ cmp $16,%ecx
+ jge .Lcbc_enc_loop4
+ emms
+ ret
+
+#if defined(__ELF__) && defined(SECTION_NOTE_GNU_STACK)
+ .section .note.GNU-stack,"",@progbits
+#endif
diff -urN linux-3.14-noloop/drivers/misc/aes-x86.S linux-3.14-AES/drivers/misc/aes-x86.S
--- linux-3.14-noloop/drivers/misc/aes-x86.S 1970-01-01 02:00:00.000000000 +0200
+++ linux-3.14-AES/drivers/misc/aes-x86.S 2014-03-31 22:20:36.000000000 +0300
@@ -0,0 +1,927 @@
+//
+// Copyright (c) 2001, Dr Brian Gladman <brg@gladman.uk.net>, Worcester, UK.
+// All rights reserved.
+//
+// TERMS
+//
+// Redistribution and use in source and binary forms, with or without
+// modification, are permitted subject to the following conditions:
+//
+// 1. Redistributions of source code must retain the above copyright
+// notice, this list of conditions and the following disclaimer.
+//
+// 2. Redistributions in binary form must reproduce the above copyright
+// notice, this list of conditions and the following disclaimer in the
+// documentation and/or other materials provided with the distribution.
+//
+// 3. The copyright holder's name must not be used to endorse or promote
+// any products derived from this software without his specific prior
+// written permission.
+//
+// This software is provided 'as is' with no express or implied warranties
+// of correctness or fitness for purpose.
+
+// Modified by Jari Ruusu, December 24 2001
+// - Converted syntax to GNU CPP/assembler syntax
+// - C programming interface converted back to "old" API
+// - Minor portability cleanups and speed optimizations
+
+// Modified by Jari Ruusu, April 11 2002
+// - Added above copyright and terms to resulting object code so that
+// binary distributions can avoid legal trouble
+
+// An AES (Rijndael) implementation for x86 compatible processors. This
+// version uses i386 instruction set but instruction scheduling is optimized
+// for Pentium-2. This version only implements the standard AES block length
+// (128 bits, 16 bytes). This code does not preserve the eax, ecx or edx
+// registers or the artihmetic status flags. However, the ebx, esi, edi, and
+// ebp registers are preserved across calls.
+
+// void aes_set_key(aes_context *cx, const unsigned char key[], const int key_len, const int f)
+// void aes_encrypt(const aes_context *cx, const unsigned char in_blk[], unsigned char out_blk[])
+// void aes_decrypt(const aes_context *cx, const unsigned char in_blk[], unsigned char out_blk[])
+
+#if defined(USE_UNDERLINE)
+# define aes_set_key _aes_set_key
+# define aes_encrypt _aes_encrypt
+# define aes_decrypt _aes_decrypt
+#endif
+#if !defined(ALIGN32BYTES)
+# define ALIGN32BYTES 32
+#endif
+
+ .file "aes-x86.S"
+ .globl aes_set_key
+ .globl aes_encrypt
+ .globl aes_decrypt
+
+ .text
+copyright:
+ .ascii " \000"
+ .ascii "Copyright (c) 2001, Dr Brian Gladman <brg@gladman.uk.net>, Worcester, UK.\000"
+ .ascii "All rights reserved.\000"
+ .ascii " \000"
+ .ascii "TERMS\000"
+ .ascii " \000"
+ .ascii " Redistribution and use in source and binary forms, with or without\000"
+ .ascii " modification, are permitted subject to the following conditions:\000"
+ .ascii " \000"
+ .ascii " 1. Redistributions of source code must retain the above copyright\000"
+ .ascii " notice, this list of conditions and the following disclaimer.\000"
+ .ascii " \000"
+ .ascii " 2. Redistributions in binary form must reproduce the above copyright\000"
+ .ascii " notice, this list of conditions and the following disclaimer in the\000"
+ .ascii " documentation and/or other materials provided with the distribution.\000"
+ .ascii " \000"
+ .ascii " 3. The copyright holder's name must not be used to endorse or promote\000"
+ .ascii " any products derived from this software without his specific prior\000"
+ .ascii " written permission.\000"
+ .ascii " \000"
+ .ascii " This software is provided 'as is' with no express or implied warranties\000"
+ .ascii " of correctness or fitness for purpose.\000"
+ .ascii " \000"
+
+#define tlen 1024 // length of each of 4 'xor' arrays (256 32-bit words)
+
+// offsets to parameters with one register pushed onto stack
+
+#define ctx 8 // AES context structure
+#define in_blk 12 // input byte array address parameter
+#define out_blk 16 // output byte array address parameter
+
+// offsets in context structure
+
+#define nkey 0 // key length, size 4
+#define nrnd 4 // number of rounds, size 4
+#define ekey 8 // encryption key schedule base address, size 256
+#define dkey 264 // decryption key schedule base address, size 256
+
+// This macro performs a forward encryption cycle. It is entered with
+// the first previous round column values in %eax, %ebx, %esi and %edi and
+// exits with the final values in the same registers.
+
+#define fwd_rnd(p1,p2) \
+ mov %ebx,(%esp) ;\
+ movzbl %al,%edx ;\
+ mov %eax,%ecx ;\
+ mov p2(%ebp),%eax ;\
+ mov %edi,4(%esp) ;\
+ mov p2+12(%ebp),%edi ;\
+ xor p1(,%edx,4),%eax ;\
+ movzbl %ch,%edx ;\
+ shr $16,%ecx ;\
+ mov p2+4(%ebp),%ebx ;\
+ xor p1+tlen(,%edx,4),%edi ;\
+ movzbl %cl,%edx ;\
+ movzbl %ch,%ecx ;\
+ xor p1+3*tlen(,%ecx,4),%ebx ;\
+ mov %esi,%ecx ;\
+ mov p1+2*tlen(,%edx,4),%esi ;\
+ movzbl %cl,%edx ;\
+ xor p1(,%edx,4),%esi ;\
+ movzbl %ch,%edx ;\
+ shr $16,%ecx ;\
+ xor p1+tlen(,%edx,4),%ebx ;\
+ movzbl %cl,%edx ;\
+ movzbl %ch,%ecx ;\
+ xor p1+2*tlen(,%edx,4),%eax ;\
+ mov (%esp),%edx ;\
+ xor p1+3*tlen(,%ecx,4),%edi ;\
+ movzbl %dl,%ecx ;\
+ xor p2+8(%ebp),%esi ;\
+ xor p1(,%ecx,4),%ebx ;\
+ movzbl %dh,%ecx ;\
+ shr $16,%edx ;\
+ xor p1+tlen(,%ecx,4),%eax ;\
+ movzbl %dl,%ecx ;\
+ movzbl %dh,%edx ;\
+ xor p1+2*tlen(,%ecx,4),%edi ;\
+ mov 4(%esp),%ecx ;\
+ xor p1+3*tlen(,%edx,4),%esi ;\
+ movzbl %cl,%edx ;\
+ xor p1(,%edx,4),%edi ;\
+ movzbl %ch,%edx ;\
+ shr $16,%ecx ;\
+ xor p1+tlen(,%edx,4),%esi ;\
+ movzbl %cl,%edx ;\
+ movzbl %ch,%ecx ;\
+ xor p1+2*tlen(,%edx,4),%ebx ;\
+ xor p1+3*tlen(,%ecx,4),%eax
+
+// This macro performs an inverse encryption cycle. It is entered with
+// the first previous round column values in %eax, %ebx, %esi and %edi and
+// exits with the final values in the same registers.
+
+#define inv_rnd(p1,p2) \
+ movzbl %al,%edx ;\
+ mov %ebx,(%esp) ;\
+ mov %eax,%ecx ;\
+ mov p2(%ebp),%eax ;\
+ mov %edi,4(%esp) ;\
+ mov p2+4(%ebp),%ebx ;\
+ xor p1(,%edx,4),%eax ;\
+ movzbl %ch,%edx ;\
+ shr $16,%ecx ;\
+ mov p2+12(%ebp),%edi ;\
+ xor p1+tlen(,%edx,4),%ebx ;\
+ movzbl %cl,%edx ;\
+ movzbl %ch,%ecx ;\
+ xor p1+3*tlen(,%ecx,4),%edi ;\
+ mov %esi,%ecx ;\
+ mov p1+2*tlen(,%edx,4),%esi ;\
+ movzbl %cl,%edx ;\
+ xor p1(,%edx,4),%esi ;\
+ movzbl %ch,%edx ;\
+ shr $16,%ecx ;\
+ xor p1+tlen(,%edx,4),%edi ;\
+ movzbl %cl,%edx ;\
+ movzbl %ch,%ecx ;\
+ xor p1+2*tlen(,%edx,4),%eax ;\
+ mov (%esp),%edx ;\
+ xor p1+3*tlen(,%ecx,4),%ebx ;\
+ movzbl %dl,%ecx ;\
+ xor p2+8(%ebp),%esi ;\
+ xor p1(,%ecx,4),%ebx ;\
+ movzbl %dh,%ecx ;\
+ shr $16,%edx ;\
+ xor p1+tlen(,%ecx,4),%esi ;\
+ movzbl %dl,%ecx ;\
+ movzbl %dh,%edx ;\
+ xor p1+2*tlen(,%ecx,4),%edi ;\
+ mov 4(%esp),%ecx ;\
+ xor p1+3*tlen(,%edx,4),%eax ;\
+ movzbl %cl,%edx ;\
+ xor p1(,%edx,4),%edi ;\
+ movzbl %ch,%edx ;\
+ shr $16,%ecx ;\
+ xor p1+tlen(,%edx,4),%eax ;\
+ movzbl %cl,%edx ;\
+ movzbl %ch,%ecx ;\
+ xor p1+2*tlen(,%edx,4),%ebx ;\
+ xor p1+3*tlen(,%ecx,4),%esi
+
+// AES (Rijndael) Encryption Subroutine
+
+ .text
+ .align ALIGN32BYTES
+aes_encrypt:
+ push %ebp
+ mov ctx(%esp),%ebp // pointer to context
+ mov in_blk(%esp),%ecx
+ push %ebx
+ push %esi
+ push %edi
+ mov nrnd(%ebp),%edx // number of rounds
+ lea ekey+16(%ebp),%ebp // key pointer
+
+// input four columns and xor in first round key
+
+ mov (%ecx),%eax
+ mov 4(%ecx),%ebx
+ mov 8(%ecx),%esi
+ mov 12(%ecx),%edi
+ xor -16(%ebp),%eax
+ xor -12(%ebp),%ebx
+ xor -8(%ebp),%esi
+ xor -4(%ebp),%edi
+
+ sub $8,%esp // space for register saves on stack
+
+ sub $10,%edx
+ je aes_15
+ add $32,%ebp
+ sub $2,%edx
+ je aes_13
+ add $32,%ebp
+
+ fwd_rnd(aes_ft_tab,-64) // 14 rounds for 256-bit key
+ fwd_rnd(aes_ft_tab,-48)
+aes_13: fwd_rnd(aes_ft_tab,-32) // 12 rounds for 192-bit key
+ fwd_rnd(aes_ft_tab,-16)
+aes_15: fwd_rnd(aes_ft_tab,0) // 10 rounds for 128-bit key
+ fwd_rnd(aes_ft_tab,16)
+ fwd_rnd(aes_ft_tab,32)
+ fwd_rnd(aes_ft_tab,48)
+ fwd_rnd(aes_ft_tab,64)
+ fwd_rnd(aes_ft_tab,80)
+ fwd_rnd(aes_ft_tab,96)
+ fwd_rnd(aes_ft_tab,112)
+ fwd_rnd(aes_ft_tab,128)
+ fwd_rnd(aes_fl_tab,144) // last round uses a different table
+
+// move final values to the output array.
+
+ mov out_blk+20(%esp),%ebp
+ add $8,%esp
+ mov %eax,(%ebp)
+ mov %ebx,4(%ebp)
+ mov %esi,8(%ebp)
+ mov %edi,12(%ebp)
+ pop %edi
+ pop %esi
+ pop %ebx
+ pop %ebp
+ ret
+
+
+// AES (Rijndael) Decryption Subroutine
+
+ .align ALIGN32BYTES
+aes_decrypt:
+ push %ebp
+ mov ctx(%esp),%ebp // pointer to context
+ mov in_blk(%esp),%ecx
+ push %ebx
+ push %esi
+ push %edi
+ mov nrnd(%ebp),%edx // number of rounds
+ lea dkey+16(%ebp),%ebp // key pointer
+
+// input four columns and xor in first round key
+
+ mov (%ecx),%eax
+ mov 4(%ecx),%ebx
+ mov 8(%ecx),%esi
+ mov 12(%ecx),%edi
+ xor -16(%ebp),%eax
+ xor -12(%ebp),%ebx
+ xor -8(%ebp),%esi
+ xor -4(%ebp),%edi
+
+ sub $8,%esp // space for register saves on stack
+
+ sub $10,%edx
+ je aes_25
+ add $32,%ebp
+ sub $2,%edx
+ je aes_23
+ add $32,%ebp
+
+ inv_rnd(aes_it_tab,-64) // 14 rounds for 256-bit key
+ inv_rnd(aes_it_tab,-48)
+aes_23: inv_rnd(aes_it_tab,-32) // 12 rounds for 192-bit key
+ inv_rnd(aes_it_tab,-16)
+aes_25: inv_rnd(aes_it_tab,0) // 10 rounds for 128-bit key
+ inv_rnd(aes_it_tab,16)
+ inv_rnd(aes_it_tab,32)
+ inv_rnd(aes_it_tab,48)
+ inv_rnd(aes_it_tab,64)
+ inv_rnd(aes_it_tab,80)
+ inv_rnd(aes_it_tab,96)
+ inv_rnd(aes_it_tab,112)
+ inv_rnd(aes_it_tab,128)
+ inv_rnd(aes_il_tab,144) // last round uses a different table
+
+// move final values to the output array.
+
+ mov out_blk+20(%esp),%ebp
+ add $8,%esp
+ mov %eax,(%ebp)
+ mov %ebx,4(%ebp)
+ mov %esi,8(%ebp)
+ mov %edi,12(%ebp)
+ pop %edi
+ pop %esi
+ pop %ebx
+ pop %ebp
+ ret
+
+// AES (Rijndael) Key Schedule Subroutine
+
+// input/output parameters
+
+#define aes_cx 12 // AES context
+#define in_key 16 // key input array address
+#define key_ln 20 // key length, bytes (16,24,32) or bits (128,192,256)
+#define ed_flg 24 // 0=create both encr/decr keys, 1=create encr key only
+
+// offsets for locals
+
+#define cnt -4
+#define slen 8
+
+// This macro performs a column mixing operation on an input 32-bit
+// word to give a 32-bit result. It uses each of the 4 bytes in the
+// the input column to index 4 different tables of 256 32-bit words
+// that are xored together to form the output value.
+
+#define mix_col(p1) \
+ movzbl %bl,%ecx ;\
+ mov p1(,%ecx,4),%eax ;\
+ movzbl %bh,%ecx ;\
+ ror $16,%ebx ;\
+ xor p1+tlen(,%ecx,4),%eax ;\
+ movzbl %bl,%ecx ;\
+ xor p1+2*tlen(,%ecx,4),%eax ;\
+ movzbl %bh,%ecx ;\
+ xor p1+3*tlen(,%ecx,4),%eax
+
+// Key Schedule Macros
+
+#define ksc4(p1) \
+ rol $24,%ebx ;\
+ mix_col(aes_fl_tab) ;\
+ ror $8,%ebx ;\
+ xor 4*p1+aes_rcon_tab,%eax ;\
+ xor %eax,%esi ;\
+ xor %esi,%ebp ;\
+ mov %esi,16*p1(%edi) ;\
+ mov %ebp,16*p1+4(%edi) ;\
+ xor %ebp,%edx ;\
+ xor %edx,%ebx ;\
+ mov %edx,16*p1+8(%edi) ;\
+ mov %ebx,16*p1+12(%edi)
+
+#define ksc6(p1) \
+ rol $24,%ebx ;\
+ mix_col(aes_fl_tab) ;\
+ ror $8,%ebx ;\
+ xor 4*p1+aes_rcon_tab,%eax ;\
+ xor 24*p1-24(%edi),%eax ;\
+ mov %eax,24*p1(%edi) ;\
+ xor 24*p1-20(%edi),%eax ;\
+ mov %eax,24*p1+4(%edi) ;\
+ xor %eax,%esi ;\
+ xor %esi,%ebp ;\
+ mov %esi,24*p1+8(%edi) ;\
+ mov %ebp,24*p1+12(%edi) ;\
+ xor %ebp,%edx ;\
+ xor %edx,%ebx ;\
+ mov %edx,24*p1+16(%edi) ;\
+ mov %ebx,24*p1+20(%edi)
+
+#define ksc8(p1) \
+ rol $24,%ebx ;\
+ mix_col(aes_fl_tab) ;\
+ ror $8,%ebx ;\
+ xor 4*p1+aes_rcon_tab,%eax ;\
+ xor 32*p1-32(%edi),%eax ;\
+ mov %eax,32*p1(%edi) ;\
+ xor 32*p1-28(%edi),%eax ;\
+ mov %eax,32*p1+4(%edi) ;\
+ xor 32*p1-24(%edi),%eax ;\
+ mov %eax,32*p1+8(%edi) ;\
+ xor 32*p1-20(%edi),%eax ;\
+ mov %eax,32*p1+12(%edi) ;\
+ push %ebx ;\
+ mov %eax,%ebx ;\
+ mix_col(aes_fl_tab) ;\
+ pop %ebx ;\
+ xor %eax,%esi ;\
+ xor %esi,%ebp ;\
+ mov %esi,32*p1+16(%edi) ;\
+ mov %ebp,32*p1+20(%edi) ;\
+ xor %ebp,%edx ;\
+ xor %edx,%ebx ;\
+ mov %edx,32*p1+24(%edi) ;\
+ mov %ebx,32*p1+28(%edi)
+
+ .align ALIGN32BYTES
+aes_set_key:
+ pushfl
+ push %ebp
+ mov %esp,%ebp
+ sub $slen,%esp
+ push %ebx
+ push %esi
+ push %edi
+
+ mov aes_cx(%ebp),%edx // edx -> AES context
+
+ mov key_ln(%ebp),%ecx // key length
+ cmpl $128,%ecx
+ jb aes_30
+ shr $3,%ecx
+aes_30: cmpl $32,%ecx
+ je aes_32
+ cmpl $24,%ecx
+ je aes_32
+ mov $16,%ecx
+aes_32: shr $2,%ecx
+ mov %ecx,nkey(%edx)
+
+ lea 6(%ecx),%eax // 10/12/14 for 4/6/8 32-bit key length
+ mov %eax,nrnd(%edx)
+
+ mov in_key(%ebp),%esi // key input array
+ lea ekey(%edx),%edi // key position in AES context
+ cld
+ push %ebp
+ mov %ecx,%eax // save key length in eax
+ rep ; movsl // words in the key schedule
+ mov -4(%esi),%ebx // put some values in registers
+ mov -8(%esi),%edx // to allow faster code
+ mov -12(%esi),%ebp
+ mov -16(%esi),%esi
+
+ cmpl $4,%eax // jump on key size
+ je aes_36
+ cmpl $6,%eax
+ je aes_35
+
+ ksc8(0)
+ ksc8(1)
+ ksc8(2)
+ ksc8(3)
+ ksc8(4)
+ ksc8(5)
+ ksc8(6)
+ jmp aes_37
+aes_35: ksc6(0)
+ ksc6(1)
+ ksc6(2)
+ ksc6(3)
+ ksc6(4)
+ ksc6(5)
+ ksc6(6)
+ ksc6(7)
+ jmp aes_37
+aes_36: ksc4(0)
+ ksc4(1)
+ ksc4(2)
+ ksc4(3)
+ ksc4(4)
+ ksc4(5)
+ ksc4(6)
+ ksc4(7)
+ ksc4(8)
+ ksc4(9)
+aes_37: pop %ebp
+ mov aes_cx(%ebp),%edx // edx -> AES context
+ cmpl $0,ed_flg(%ebp)
+ jne aes_39
+
+// compile decryption key schedule from encryption schedule - reverse
+// order and do mix_column operation on round keys except first and last
+
+ mov nrnd(%edx),%eax // kt = cx->d_key + nc * cx->Nrnd
+ shl $2,%eax
+ lea dkey(%edx,%eax,4),%edi
+ lea ekey(%edx),%esi // kf = cx->e_key
+
+ movsl // copy first round key (unmodified)
+ movsl
+ movsl
+ movsl
+ sub $32,%edi
+ movl $1,cnt(%ebp)
+aes_38: // do mix column on each column of
+ lodsl // each round key
+ mov %eax,%ebx
+ mix_col(aes_im_tab)
+ stosl
+ lodsl
+ mov %eax,%ebx
+ mix_col(aes_im_tab)
+ stosl
+ lodsl
+ mov %eax,%ebx
+ mix_col(aes_im_tab)
+ stosl
+ lodsl
+ mov %eax,%ebx
+ mix_col(aes_im_tab)
+ stosl
+ sub $32,%edi
+
+ incl cnt(%ebp)
+ mov cnt(%ebp),%eax
+ cmp nrnd(%edx),%eax
+ jb aes_38
+
+ movsl // copy last round key (unmodified)
+ movsl
+ movsl
+ movsl
+aes_39: pop %edi
+ pop %esi
+ pop %ebx
+ mov %ebp,%esp
+ pop %ebp
+ popfl
+ ret
+
+
+// finite field multiplies by {02}, {04} and {08}
+
+#define f2(x) ((x<<1)^(((x>>7)&1)*0x11b))
+#define f4(x) ((x<<2)^(((x>>6)&1)*0x11b)^(((x>>6)&2)*0x11b))
+#define f8(x) ((x<<3)^(((x>>5)&1)*0x11b)^(((x>>5)&2)*0x11b)^(((x>>5)&4)*0x11b))
+
+// finite field multiplies required in table generation
+
+#define f3(x) (f2(x) ^ x)
+#define f9(x) (f8(x) ^ x)
+#define fb(x) (f8(x) ^ f2(x) ^ x)
+#define fd(x) (f8(x) ^ f4(x) ^ x)
+#define fe(x) (f8(x) ^ f4(x) ^ f2(x))
+
+// These defines generate the forward table entries
+
+#define u0(x) ((f3(x) << 24) | (x << 16) | (x << 8) | f2(x))
+#define u1(x) ((x << 24) | (x << 16) | (f2(x) << 8) | f3(x))
+#define u2(x) ((x << 24) | (f2(x) << 16) | (f3(x) << 8) | x)
+#define u3(x) ((f2(x) << 24) | (f3(x) << 16) | (x << 8) | x)
+
+// These defines generate the inverse table entries
+
+#define v0(x) ((fb(x) << 24) | (fd(x) << 16) | (f9(x) << 8) | fe(x))
+#define v1(x) ((fd(x) << 24) | (f9(x) << 16) | (fe(x) << 8) | fb(x))
+#define v2(x) ((f9(x) << 24) | (fe(x) << 16) | (fb(x) << 8) | fd(x))
+#define v3(x) ((fe(x) << 24) | (fb(x) << 16) | (fd(x) << 8) | f9(x))
+
+// These defines generate entries for the last round tables
+
+#define w0(x) (x)
+#define w1(x) (x << 8)
+#define w2(x) (x << 16)
+#define w3(x) (x << 24)
+
+// macro to generate inverse mix column tables (needed for the key schedule)
+
+#define im_data0(p1) \
+ .long p1(0x00),p1(0x01),p1(0x02),p1(0x03),p1(0x04),p1(0x05),p1(0x06),p1(0x07) ;\
+ .long p1(0x08),p1(0x09),p1(0x0a),p1(0x0b),p1(0x0c),p1(0x0d),p1(0x0e),p1(0x0f) ;\
+ .long p1(0x10),p1(0x11),p1(0x12),p1(0x13),p1(0x14),p1(0x15),p1(0x16),p1(0x17) ;\
+ .long p1(0x18),p1(0x19),p1(0x1a),p1(0x1b),p1(0x1c),p1(0x1d),p1(0x1e),p1(0x1f)
+#define im_data1(p1) \
+ .long p1(0x20),p1(0x21),p1(0x22),p1(0x23),p1(0x24),p1(0x25),p1(0x26),p1(0x27) ;\
+ .long p1(0x28),p1(0x29),p1(0x2a),p1(0x2b),p1(0x2c),p1(0x2d),p1(0x2e),p1(0x2f) ;\
+ .long p1(0x30),p1(0x31),p1(0x32),p1(0x33),p1(0x34),p1(0x35),p1(0x36),p1(0x37) ;\
+ .long p1(0x38),p1(0x39),p1(0x3a),p1(0x3b),p1(0x3c),p1(0x3d),p1(0x3e),p1(0x3f)
+#define im_data2(p1) \
+ .long p1(0x40),p1(0x41),p1(0x42),p1(0x43),p1(0x44),p1(0x45),p1(0x46),p1(0x47) ;\
+ .long p1(0x48),p1(0x49),p1(0x4a),p1(0x4b),p1(0x4c),p1(0x4d),p1(0x4e),p1(0x4f) ;\
+ .long p1(0x50),p1(0x51),p1(0x52),p1(0x53),p1(0x54),p1(0x55),p1(0x56),p1(0x57) ;\
+ .long p1(0x58),p1(0x59),p1(0x5a),p1(0x5b),p1(0x5c),p1(0x5d),p1(0x5e),p1(0x5f)
+#define im_data3(p1) \
+ .long p1(0x60),p1(0x61),p1(0x62),p1(0x63),p1(0x64),p1(0x65),p1(0x66),p1(0x67) ;\
+ .long p1(0x68),p1(0x69),p1(0x6a),p1(0x6b),p1(0x6c),p1(0x6d),p1(0x6e),p1(0x6f) ;\
+ .long p1(0x70),p1(0x71),p1(0x72),p1(0x73),p1(0x74),p1(0x75),p1(0x76),p1(0x77) ;\
+ .long p1(0x78),p1(0x79),p1(0x7a),p1(0x7b),p1(0x7c),p1(0x7d),p1(0x7e),p1(0x7f)
+#define im_data4(p1) \
+ .long p1(0x80),p1(0x81),p1(0x82),p1(0x83),p1(0x84),p1(0x85),p1(0x86),p1(0x87) ;\
+ .long p1(0x88),p1(0x89),p1(0x8a),p1(0x8b),p1(0x8c),p1(0x8d),p1(0x8e),p1(0x8f) ;\
+ .long p1(0x90),p1(0x91),p1(0x92),p1(0x93),p1(0x94),p1(0x95),p1(0x96),p1(0x97) ;\
+ .long p1(0x98),p1(0x99),p1(0x9a),p1(0x9b),p1(0x9c),p1(0x9d),p1(0x9e),p1(0x9f)
+#define im_data5(p1) \
+ .long p1(0xa0),p1(0xa1),p1(0xa2),p1(0xa3),p1(0xa4),p1(0xa5),p1(0xa6),p1(0xa7) ;\
+ .long p1(0xa8),p1(0xa9),p1(0xaa),p1(0xab),p1(0xac),p1(0xad),p1(0xae),p1(0xaf) ;\
+ .long p1(0xb0),p1(0xb1),p1(0xb2),p1(0xb3),p1(0xb4),p1(0xb5),p1(0xb6),p1(0xb7) ;\
+ .long p1(0xb8),p1(0xb9),p1(0xba),p1(0xbb),p1(0xbc),p1(0xbd),p1(0xbe),p1(0xbf)
+#define im_data6(p1) \
+ .long p1(0xc0),p1(0xc1),p1(0xc2),p1(0xc3),p1(0xc4),p1(0xc5),p1(0xc6),p1(0xc7) ;\
+ .long p1(0xc8),p1(0xc9),p1(0xca),p1(0xcb),p1(0xcc),p1(0xcd),p1(0xce),p1(0xcf) ;\
+ .long p1(0xd0),p1(0xd1),p1(0xd2),p1(0xd3),p1(0xd4),p1(0xd5),p1(0xd6),p1(0xd7) ;\
+ .long p1(0xd8),p1(0xd9),p1(0xda),p1(0xdb),p1(0xdc),p1(0xdd),p1(0xde),p1(0xdf)
+#define im_data7(p1) \
+ .long p1(0xe0),p1(0xe1),p1(0xe2),p1(0xe3),p1(0xe4),p1(0xe5),p1(0xe6),p1(0xe7) ;\
+ .long p1(0xe8),p1(0xe9),p1(0xea),p1(0xeb),p1(0xec),p1(0xed),p1(0xee),p1(0xef) ;\
+ .long p1(0xf0),p1(0xf1),p1(0xf2),p1(0xf3),p1(0xf4),p1(0xf5),p1(0xf6),p1(0xf7) ;\
+ .long p1(0xf8),p1(0xf9),p1(0xfa),p1(0xfb),p1(0xfc),p1(0xfd),p1(0xfe),p1(0xff)
+
+// S-box data - 256 entries
+
+#define sb_data0(p1) \
+ .long p1(0x63),p1(0x7c),p1(0x77),p1(0x7b),p1(0xf2),p1(0x6b),p1(0x6f),p1(0xc5) ;\
+ .long p1(0x30),p1(0x01),p1(0x67),p1(0x2b),p1(0xfe),p1(0xd7),p1(0xab),p1(0x76) ;\
+ .long p1(0xca),p1(0x82),p1(0xc9),p1(0x7d),p1(0xfa),p1(0x59),p1(0x47),p1(0xf0) ;\
+ .long p1(0xad),p1(0xd4),p1(0xa2),p1(0xaf),p1(0x9c),p1(0xa4),p1(0x72),p1(0xc0)
+#define sb_data1(p1) \
+ .long p1(0xb7),p1(0xfd),p1(0x93),p1(0x26),p1(0x36),p1(0x3f),p1(0xf7),p1(0xcc) ;\
+ .long p1(0x34),p1(0xa5),p1(0xe5),p1(0xf1),p1(0x71),p1(0xd8),p1(0x31),p1(0x15) ;\
+ .long p1(0x04),p1(0xc7),p1(0x23),p1(0xc3),p1(0x18),p1(0x96),p1(0x05),p1(0x9a) ;\
+ .long p1(0x07),p1(0x12),p1(0x80),p1(0xe2),p1(0xeb),p1(0x27),p1(0xb2),p1(0x75)
+#define sb_data2(p1) \
+ .long p1(0x09),p1(0x83),p1(0x2c),p1(0x1a),p1(0x1b),p1(0x6e),p1(0x5a),p1(0xa0) ;\
+ .long p1(0x52),p1(0x3b),p1(0xd6),p1(0xb3),p1(0x29),p1(0xe3),p1(0x2f),p1(0x84) ;\
+ .long p1(0x53),p1(0xd1),p1(0x00),p1(0xed),p1(0x20),p1(0xfc),p1(0xb1),p1(0x5b) ;\
+ .long p1(0x6a),p1(0xcb),p1(0xbe),p1(0x39),p1(0x4a),p1(0x4c),p1(0x58),p1(0xcf)
+#define sb_data3(p1) \
+ .long p1(0xd0),p1(0xef),p1(0xaa),p1(0xfb),p1(0x43),p1(0x4d),p1(0x33),p1(0x85) ;\
+ .long p1(0x45),p1(0xf9),p1(0x02),p1(0x7f),p1(0x50),p1(0x3c),p1(0x9f),p1(0xa8) ;\
+ .long p1(0x51),p1(0xa3),p1(0x40),p1(0x8f),p1(0x92),p1(0x9d),p1(0x38),p1(0xf5) ;\
+ .long p1(0xbc),p1(0xb6),p1(0xda),p1(0x21),p1(0x10),p1(0xff),p1(0xf3),p1(0xd2)
+#define sb_data4(p1) \
+ .long p1(0xcd),p1(0x0c),p1(0x13),p1(0xec),p1(0x5f),p1(0x97),p1(0x44),p1(0x17) ;\
+ .long p1(0xc4),p1(0xa7),p1(0x7e),p1(0x3d),p1(0x64),p1(0x5d),p1(0x19),p1(0x73) ;\
+ .long p1(0x60),p1(0x81),p1(0x4f),p1(0xdc),p1(0x22),p1(0x2a),p1(0x90),p1(0x88) ;\
+ .long p1(0x46),p1(0xee),p1(0xb8),p1(0x14),p1(0xde),p1(0x5e),p1(0x0b),p1(0xdb)
+#define sb_data5(p1) \
+ .long p1(0xe0),p1(0x32),p1(0x3a),p1(0x0a),p1(0x49),p1(0x06),p1(0x24),p1(0x5c) ;\
+ .long p1(0xc2),p1(0xd3),p1(0xac),p1(0x62),p1(0x91),p1(0x95),p1(0xe4),p1(0x79) ;\
+ .long p1(0xe7),p1(0xc8),p1(0x37),p1(0x6d),p1(0x8d),p1(0xd5),p1(0x4e),p1(0xa9) ;\
+ .long p1(0x6c),p1(0x56),p1(0xf4),p1(0xea),p1(0x65),p1(0x7a),p1(0xae),p1(0x08)
+#define sb_data6(p1) \
+ .long p1(0xba),p1(0x78),p1(0x25),p1(0x2e),p1(0x1c),p1(0xa6),p1(0xb4),p1(0xc6) ;\
+ .long p1(0xe8),p1(0xdd),p1(0x74),p1(0x1f),p1(0x4b),p1(0xbd),p1(0x8b),p1(0x8a) ;\
+ .long p1(0x70),p1(0x3e),p1(0xb5),p1(0x66),p1(0x48),p1(0x03),p1(0xf6),p1(0x0e) ;\
+ .long p1(0x61),p1(0x35),p1(0x57),p1(0xb9),p1(0x86),p1(0xc1),p1(0x1d),p1(0x9e)
+#define sb_data7(p1) \
+ .long p1(0xe1),p1(0xf8),p1(0x98),p1(0x11),p1(0x69),p1(0xd9),p1(0x8e),p1(0x94) ;\
+ .long p1(0x9b),p1(0x1e),p1(0x87),p1(0xe9),p1(0xce),p1(0x55),p1(0x28),p1(0xdf) ;\
+ .long p1(0x8c),p1(0xa1),p1(0x89),p1(0x0d),p1(0xbf),p1(0xe6),p1(0x42),p1(0x68) ;\
+ .long p1(0x41),p1(0x99),p1(0x2d),p1(0x0f),p1(0xb0),p1(0x54),p1(0xbb),p1(0x16)
+
+// Inverse S-box data - 256 entries
+
+#define ib_data0(p1) \
+ .long p1(0x52),p1(0x09),p1(0x6a),p1(0xd5),p1(0x30),p1(0x36),p1(0xa5),p1(0x38) ;\
+ .long p1(0xbf),p1(0x40),p1(0xa3),p1(0x9e),p1(0x81),p1(0xf3),p1(0xd7),p1(0xfb) ;\
+ .long p1(0x7c),p1(0xe3),p1(0x39),p1(0x82),p1(0x9b),p1(0x2f),p1(0xff),p1(0x87) ;\
+ .long p1(0x34),p1(0x8e),p1(0x43),p1(0x44),p1(0xc4),p1(0xde),p1(0xe9),p1(0xcb)
+#define ib_data1(p1) \
+ .long p1(0x54),p1(0x7b),p1(0x94),p1(0x32),p1(0xa6),p1(0xc2),p1(0x23),p1(0x3d) ;\
+ .long p1(0xee),p1(0x4c),p1(0x95),p1(0x0b),p1(0x42),p1(0xfa),p1(0xc3),p1(0x4e) ;\
+ .long p1(0x08),p1(0x2e),p1(0xa1),p1(0x66),p1(0x28),p1(0xd9),p1(0x24),p1(0xb2) ;\
+ .long p1(0x76),p1(0x5b),p1(0xa2),p1(0x49),p1(0x6d),p1(0x8b),p1(0xd1),p1(0x25)
+#define ib_data2(p1) \
+ .long p1(0x72),p1(0xf8),p1(0xf6),p1(0x64),p1(0x86),p1(0x68),p1(0x98),p1(0x16) ;\
+ .long p1(0xd4),p1(0xa4),p1(0x5c),p1(0xcc),p1(0x5d),p1(0x65),p1(0xb6),p1(0x92) ;\
+ .long p1(0x6c),p1(0x70),p1(0x48),p1(0x50),p1(0xfd),p1(0xed),p1(0xb9),p1(0xda) ;\
+ .long p1(0x5e),p1(0x15),p1(0x46),p1(0x57),p1(0xa7),p1(0x8d),p1(0x9d),p1(0x84)
+#define ib_data3(p1) \
+ .long p1(0x90),p1(0xd8),p1(0xab),p1(0x00),p1(0x8c),p1(0xbc),p1(0xd3),p1(0x0a) ;\
+ .long p1(0xf7),p1(0xe4),p1(0x58),p1(0x05),p1(0xb8),p1(0xb3),p1(0x45),p1(0x06) ;\
+ .long p1(0xd0),p1(0x2c),p1(0x1e),p1(0x8f),p1(0xca),p1(0x3f),p1(0x0f),p1(0x02) ;\
+ .long p1(0xc1),p1(0xaf),p1(0xbd),p1(0x03),p1(0x01),p1(0x13),p1(0x8a),p1(0x6b)
+#define ib_data4(p1) \
+ .long p1(0x3a),p1(0x91),p1(0x11),p1(0x41),p1(0x4f),p1(0x67),p1(0xdc),p1(0xea) ;\
+ .long p1(0x97),p1(0xf2),p1(0xcf),p1(0xce),p1(0xf0),p1(0xb4),p1(0xe6),p1(0x73) ;\
+ .long p1(0x96),p1(0xac),p1(0x74),p1(0x22),p1(0xe7),p1(0xad),p1(0x35),p1(0x85) ;\
+ .long p1(0xe2),p1(0xf9),p1(0x37),p1(0xe8),p1(0x1c),p1(0x75),p1(0xdf),p1(0x6e)
+#define ib_data5(p1) \
+ .long p1(0x47),p1(0xf1),p1(0x1a),p1(0x71),p1(0x1d),p1(0x29),p1(0xc5),p1(0x89) ;\
+ .long p1(0x6f),p1(0xb7),p1(0x62),p1(0x0e),p1(0xaa),p1(0x18),p1(0xbe),p1(0x1b) ;\
+ .long p1(0xfc),p1(0x56),p1(0x3e),p1(0x4b),p1(0xc6),p1(0xd2),p1(0x79),p1(0x20) ;\
+ .long p1(0x9a),p1(0xdb),p1(0xc0),p1(0xfe),p1(0x78),p1(0xcd),p1(0x5a),p1(0xf4)
+#define ib_data6(p1) \
+ .long p1(0x1f),p1(0xdd),p1(0xa8),p1(0x33),p1(0x88),p1(0x07),p1(0xc7),p1(0x31) ;\
+ .long p1(0xb1),p1(0x12),p1(0x10),p1(0x59),p1(0x27),p1(0x80),p1(0xec),p1(0x5f) ;\
+ .long p1(0x60),p1(0x51),p1(0x7f),p1(0xa9),p1(0x19),p1(0xb5),p1(0x4a),p1(0x0d) ;\
+ .long p1(0x2d),p1(0xe5),p1(0x7a),p1(0x9f),p1(0x93),p1(0xc9),p1(0x9c),p1(0xef)
+#define ib_data7(p1) \
+ .long p1(0xa0),p1(0xe0),p1(0x3b),p1(0x4d),p1(0xae),p1(0x2a),p1(0xf5),p1(0xb0) ;\
+ .long p1(0xc8),p1(0xeb),p1(0xbb),p1(0x3c),p1(0x83),p1(0x53),p1(0x99),p1(0x61) ;\
+ .long p1(0x17),p1(0x2b),p1(0x04),p1(0x7e),p1(0xba),p1(0x77),p1(0xd6),p1(0x26) ;\
+ .long p1(0xe1),p1(0x69),p1(0x14),p1(0x63),p1(0x55),p1(0x21),p1(0x0c),p1(0x7d)
+
+// The rcon_table (needed for the key schedule)
+//
+// Here is original Dr Brian Gladman's source code:
+// _rcon_tab:
+// %assign x 1
+// %rep 29
+// dd x
+// %assign x f2(x)
+// %endrep
+//
+// Here is precomputed output (it's more portable this way):
+
+ .section .rodata
+ .align ALIGN32BYTES
+aes_rcon_tab:
+ .long 0x01,0x02,0x04,0x08,0x10,0x20,0x40,0x80
+ .long 0x1b,0x36,0x6c,0xd8,0xab,0x4d,0x9a,0x2f
+ .long 0x5e,0xbc,0x63,0xc6,0x97,0x35,0x6a,0xd4
+ .long 0xb3,0x7d,0xfa,0xef,0xc5
+
+// The forward xor tables
+
+ .align ALIGN32BYTES
+aes_ft_tab:
+ sb_data0(u0)
+ sb_data1(u0)
+ sb_data2(u0)
+ sb_data3(u0)
+ sb_data4(u0)
+ sb_data5(u0)
+ sb_data6(u0)
+ sb_data7(u0)
+
+ sb_data0(u1)
+ sb_data1(u1)
+ sb_data2(u1)
+ sb_data3(u1)
+ sb_data4(u1)
+ sb_data5(u1)
+ sb_data6(u1)
+ sb_data7(u1)
+
+ sb_data0(u2)
+ sb_data1(u2)
+ sb_data2(u2)
+ sb_data3(u2)
+ sb_data4(u2)
+ sb_data5(u2)
+ sb_data6(u2)
+ sb_data7(u2)
+
+ sb_data0(u3)
+ sb_data1(u3)
+ sb_data2(u3)
+ sb_data3(u3)
+ sb_data4(u3)
+ sb_data5(u3)
+ sb_data6(u3)
+ sb_data7(u3)
+
+ .align ALIGN32BYTES
+aes_fl_tab:
+ sb_data0(w0)
+ sb_data1(w0)
+ sb_data2(w0)
+ sb_data3(w0)
+ sb_data4(w0)
+ sb_data5(w0)
+ sb_data6(w0)
+ sb_data7(w0)
+
+ sb_data0(w1)
+ sb_data1(w1)
+ sb_data2(w1)
+ sb_data3(w1)
+ sb_data4(w1)
+ sb_data5(w1)
+ sb_data6(w1)
+ sb_data7(w1)
+
+ sb_data0(w2)
+ sb_data1(w2)
+ sb_data2(w2)
+ sb_data3(w2)
+ sb_data4(w2)
+ sb_data5(w2)
+ sb_data6(w2)
+ sb_data7(w2)
+
+ sb_data0(w3)
+ sb_data1(w3)
+ sb_data2(w3)
+ sb_data3(w3)
+ sb_data4(w3)
+ sb_data5(w3)
+ sb_data6(w3)
+ sb_data7(w3)
+
+// The inverse xor tables
+
+ .align ALIGN32BYTES
+aes_it_tab:
+ ib_data0(v0)
+ ib_data1(v0)
+ ib_data2(v0)
+ ib_data3(v0)
+ ib_data4(v0)
+ ib_data5(v0)
+ ib_data6(v0)
+ ib_data7(v0)
+
+ ib_data0(v1)
+ ib_data1(v1)
+ ib_data2(v1)
+ ib_data3(v1)
+ ib_data4(v1)
+ ib_data5(v1)
+ ib_data6(v1)
+ ib_data7(v1)
+
+ ib_data0(v2)
+ ib_data1(v2)
+ ib_data2(v2)
+ ib_data3(v2)
+ ib_data4(v2)
+ ib_data5(v2)
+ ib_data6(v2)
+ ib_data7(v2)
+
+ ib_data0(v3)
+ ib_data1(v3)
+ ib_data2(v3)
+ ib_data3(v3)
+ ib_data4(v3)
+ ib_data5(v3)
+ ib_data6(v3)
+ ib_data7(v3)
+
+ .align ALIGN32BYTES
+aes_il_tab:
+ ib_data0(w0)
+ ib_data1(w0)
+ ib_data2(w0)
+ ib_data3(w0)
+ ib_data4(w0)
+ ib_data5(w0)
+ ib_data6(w0)
+ ib_data7(w0)
+
+ ib_data0(w1)
+ ib_data1(w1)
+ ib_data2(w1)
+ ib_data3(w1)
+ ib_data4(w1)
+ ib_data5(w1)
+ ib_data6(w1)
+ ib_data7(w1)
+
+ ib_data0(w2)
+ ib_data1(w2)
+ ib_data2(w2)
+ ib_data3(w2)
+ ib_data4(w2)
+ ib_data5(w2)
+ ib_data6(w2)
+ ib_data7(w2)
+
+ ib_data0(w3)
+ ib_data1(w3)
+ ib_data2(w3)
+ ib_data3(w3)
+ ib_data4(w3)
+ ib_data5(w3)
+ ib_data6(w3)
+ ib_data7(w3)
+
+// The inverse mix column tables
+
+ .align ALIGN32BYTES
+aes_im_tab:
+ im_data0(v0)
+ im_data1(v0)
+ im_data2(v0)
+ im_data3(v0)
+ im_data4(v0)
+ im_data5(v0)
+ im_data6(v0)
+ im_data7(v0)
+
+ im_data0(v1)
+ im_data1(v1)
+ im_data2(v1)
+ im_data3(v1)
+ im_data4(v1)
+ im_data5(v1)
+ im_data6(v1)
+ im_data7(v1)
+
+ im_data0(v2)
+ im_data1(v2)
+ im_data2(v2)
+ im_data3(v2)
+ im_data4(v2)
+ im_data5(v2)
+ im_data6(v2)
+ im_data7(v2)
+
+ im_data0(v3)
+ im_data1(v3)
+ im_data2(v3)
+ im_data3(v3)
+ im_data4(v3)
+ im_data5(v3)
+ im_data6(v3)
+ im_data7(v3)
+
+#if defined(__ELF__) && defined(SECTION_NOTE_GNU_STACK)
+ .section .note.GNU-stack,"",@progbits
+#endif
diff -urN linux-3.14-noloop/drivers/misc/aes.c linux-3.14-AES/drivers/misc/aes.c
--- linux-3.14-noloop/drivers/misc/aes.c 1970-01-01 02:00:00.000000000 +0200
+++ linux-3.14-AES/drivers/misc/aes.c 2014-03-31 22:20:36.000000000 +0300
@@ -0,0 +1,1479 @@
+// I retain copyright in this code but I encourage its free use provided
+// that I don't carry any responsibility for the results. I am especially
+// happy to see it used in free and open source software. If you do use
+// it I would appreciate an acknowledgement of its origin in the code or
+// the product that results and I would also appreciate knowing a little
+// about the use to which it is being put. I am grateful to Frank Yellin
+// for some ideas that are used in this implementation.
+//
+// Dr B. R. Gladman <brg@gladman.uk.net> 6th April 2001.
+//
+// This is an implementation of the AES encryption algorithm (Rijndael)
+// designed by Joan Daemen and Vincent Rijmen. This version is designed
+// to provide both fixed and dynamic block and key lengths and can also
+// run with either big or little endian internal byte order (see aes.h).
+// It inputs block and key lengths in bytes with the legal values being
+// 16, 24 and 32.
+
+/*
+ * Modified by Jari Ruusu, May 1 2001
+ * - Fixed some compile warnings, code was ok but gcc warned anyway.
+ * - Changed basic types: byte -> unsigned char, word -> u_int32_t
+ * - Major name space cleanup: Names visible to outside now begin
+ * with "aes_" or "AES_". A lot of stuff moved from aes.h to aes.c
+ * - Removed C++ and DLL support as part of name space cleanup.
+ * - Eliminated unnecessary recomputation of tables. (actual bug fix)
+ * - Merged precomputed constant tables to aes.c file.
+ * - Removed data alignment restrictions for portability reasons.
+ * - Made block and key lengths accept bit count (128/192/256)
+ * as well byte count (16/24/32).
+ * - Removed all error checks. This change also eliminated the need
+ * to preinitialize the context struct to zero.
+ * - Removed some totally unused constants.
+ */
+/*
+ * Modified by Jari Ruusu, April 21 2004
+ * - Added back code that avoids byte swaps on big endian boxes.
+ */
+
+#include "aes.h"
+
+// CONFIGURATION OPTIONS (see also aes.h)
+//
+// 1. Define UNROLL for full loop unrolling in encryption and decryption.
+// 2. Define PARTIAL_UNROLL to unroll two loops in encryption and decryption.
+// 3. Define FIXED_TABLES for compiled rather than dynamic tables.
+// 4. Define FF_TABLES to use tables for field multiplies and inverses.
+// Do not enable this without understanding stack space requirements.
+// 5. Define ARRAYS to use arrays to hold the local state block. If this
+// is not defined, individually declared 32-bit words are used.
+// 6. Define FAST_VARIABLE if a high speed variable block implementation
+// is needed (essentially three separate fixed block size code sequences)
+// 7. Define either ONE_TABLE or FOUR_TABLES for a fast table driven
+// version using 1 table (2 kbytes of table space) or 4 tables (8
+// kbytes of table space) for higher speed.
+// 8. Define either ONE_LR_TABLE or FOUR_LR_TABLES for a further speed
+// increase by using tables for the last rounds but with more table
+// space (2 or 8 kbytes extra).
+// 9. If neither ONE_TABLE nor FOUR_TABLES is defined, a compact but
+// slower version is provided.
+// 10. If fast decryption key scheduling is needed define ONE_IM_TABLE
+// or FOUR_IM_TABLES for higher speed (2 or 8 kbytes extra).
+
+#define UNROLL
+//#define PARTIAL_UNROLL
+
+#define FIXED_TABLES
+//#define FF_TABLES
+//#define ARRAYS
+#define FAST_VARIABLE
+
+//#define ONE_TABLE
+#define FOUR_TABLES
+
+//#define ONE_LR_TABLE
+#define FOUR_LR_TABLES
+
+//#define ONE_IM_TABLE
+#define FOUR_IM_TABLES
+
+#if defined(UNROLL) && defined (PARTIAL_UNROLL)
+#error both UNROLL and PARTIAL_UNROLL are defined
+#endif
+
+#if defined(ONE_TABLE) && defined (FOUR_TABLES)
+#error both ONE_TABLE and FOUR_TABLES are defined
+#endif
+
+#if defined(ONE_LR_TABLE) && defined (FOUR_LR_TABLES)
+#error both ONE_LR_TABLE and FOUR_LR_TABLES are defined
+#endif
+
+#if defined(ONE_IM_TABLE) && defined (FOUR_IM_TABLES)
+#error both ONE_IM_TABLE and FOUR_IM_TABLES are defined
+#endif
+
+#if defined(AES_BLOCK_SIZE) && AES_BLOCK_SIZE != 16 && AES_BLOCK_SIZE != 24 && AES_BLOCK_SIZE != 32
+#error an illegal block size has been specified
+#endif
+
+/* INTERNAL_BYTE_ORDER: 0=unknown, 1=little endian, 2=big endian */
+#if defined(INTERNAL_BYTE_ORDER)
+#elif defined(__i386__)||defined(__i386)||defined(__x86_64__)||defined(__x86_64)||defined(__amd64__)||defined(__amd64)||defined(__AMD64__)||defined(__AMD64)
+# define INTERNAL_BYTE_ORDER 1
+# undef DATA_ALWAYS_ALIGNED
+# define DATA_ALWAYS_ALIGNED 1 /* unaligned access is always ok */
+#elif defined(__ppc__)||defined(__ppc)||defined(__PPC__)||defined(__PPC)||defined(__powerpc__)||defined(__powerpc)||defined(__POWERPC__)||defined(__POWERPC)||defined(__PowerPC__)||defined(__PowerPC)||defined(__ppc64__)||defined(__ppc64)||defined(__PPC64__)||defined(__PPC64)||defined(__powerpc64__)||defined(__powerpc64)||defined(__s390__)||defined(__s390)
+# define INTERNAL_BYTE_ORDER 2
+# undef DATA_ALWAYS_ALIGNED
+# define DATA_ALWAYS_ALIGNED 1 /* unaligned access is always ok */
+#elif defined(__alpha__)||defined(__alpha)||defined(__ia64__)||defined(__ia64)
+# define INTERNAL_BYTE_ORDER 1
+#elif defined(__hppa__)||defined(__hppa)||defined(__HPPA__)||defined(__HPPA)||defined(__parisc__)||defined(__parisc)||defined(__sparc__)||defined(__sparc)||defined(__sparc_v9__)||defined(__sparc_v9)||defined(__sparc64__)||defined(__sparc64)||defined(__mc68000__)||defined(__mc68000)
+# define INTERNAL_BYTE_ORDER 2
+#elif defined(CONFIGURE_DETECTS_BYTE_ORDER)
+# if WORDS_BIGENDIAN
+# define INTERNAL_BYTE_ORDER 2
+# else
+# define INTERNAL_BYTE_ORDER 1
+# endif
+#elif defined(__linux__) && defined(__KERNEL__)
+# include <asm/byteorder.h>
+# if defined(__BIG_ENDIAN)
+# define INTERNAL_BYTE_ORDER 2
+# else
+# define INTERNAL_BYTE_ORDER 1
+# endif
+#else
+# include <sys/param.h>
+# if (defined(BYTE_ORDER) && defined(LITTLE_ENDIAN) && (BYTE_ORDER == LITTLE_ENDIAN)) || (defined(__BYTE_ORDER) && defined(__LITTLE_ENDIAN) && (__BYTE_ORDER == __LITTLE_ENDIAN))
+# define INTERNAL_BYTE_ORDER 1
+# elif WORDS_BIGENDIAN || defined(__BIG_ENDIAN__) || (defined(BYTE_ORDER) && defined(BIG_ENDIAN) && (BYTE_ORDER == BIG_ENDIAN)) || (defined(__BYTE_ORDER) && defined(__BIG_ENDIAN) && (__BYTE_ORDER == __BIG_ENDIAN))
+# define INTERNAL_BYTE_ORDER 2
+# else
+# define INTERNAL_BYTE_ORDER 0
+# endif
+#endif
+
+#if defined(DATA_ALWAYS_ALIGNED) && (INTERNAL_BYTE_ORDER > 0)
+# define word_in(x) *(u_int32_t*)(x)
+# define word_out(x,v) *(u_int32_t*)(x) = (v)
+#elif defined(__linux__) && defined(__KERNEL__)
+# include <asm/unaligned.h>
+# define word_in(x) get_unaligned((u_int32_t*)(x))
+# define word_out(x,v) put_unaligned((v),(u_int32_t*)(x))
+#else
+/* unknown endianness and/or unable to handle unaligned data */
+# undef INTERNAL_BYTE_ORDER
+# define INTERNAL_BYTE_ORDER 1
+# define word_in(x) ((u_int32_t)(((unsigned char *)(x))[0])|((u_int32_t)(((unsigned char *)(x))[1])<<8)|((u_int32_t)(((unsigned char *)(x))[2])<<16)|((u_int32_t)(((unsigned char *)(x))[3])<<24))
+# define word_out(x,v) ((unsigned char *)(x))[0]=(v),((unsigned char *)(x))[1]=((v)>>8),((unsigned char *)(x))[2]=((v)>>16),((unsigned char *)(x))[3]=((v)>>24)
+#endif
+
+// upr(x,n): rotates bytes within words by n positions, moving bytes
+// to higher index positions with wrap around into low positions
+// ups(x,n): moves bytes by n positions to higher index positions in
+// words but without wrap around
+// bval(x,n): extracts a byte from a word
+
+#if (INTERNAL_BYTE_ORDER < 2)
+/* little endian */
+#define upr(x,n) (((x) << 8 * (n)) | ((x) >> (32 - 8 * (n))))
+#define ups(x,n) ((x) << 8 * (n))
+#define bval(x,n) ((unsigned char)((x) >> 8 * (n)))
+#define bytes2word(b0, b1, b2, b3) \
+ ((u_int32_t)(b3) << 24 | (u_int32_t)(b2) << 16 | (u_int32_t)(b1) << 8 | (b0))
+#else
+/* big endian */
+#define upr(x,n) (((x) >> 8 * (n)) | ((x) << (32 - 8 * (n))))
+#define ups(x,n) ((x) >> 8 * (n)))
+#define bval(x,n) ((unsigned char)((x) >> (24 - 8 * (n))))
+#define bytes2word(b0, b1, b2, b3) \
+ ((u_int32_t)(b0) << 24 | (u_int32_t)(b1) << 16 | (u_int32_t)(b2) << 8 | (b3))
+#endif
+
+// Disable at least some poor combinations of options
+
+#if !defined(ONE_TABLE) && !defined(FOUR_TABLES)
+#define FIXED_TABLES
+#undef UNROLL
+#undef ONE_LR_TABLE
+#undef FOUR_LR_TABLES
+#undef ONE_IM_TABLE
+#undef FOUR_IM_TABLES
+#elif !defined(FOUR_TABLES)
+#ifdef FOUR_LR_TABLES
+#undef FOUR_LR_TABLES
+#define ONE_LR_TABLE
+#endif
+#ifdef FOUR_IM_TABLES
+#undef FOUR_IM_TABLES
+#define ONE_IM_TABLE
+#endif
+#elif !defined(AES_BLOCK_SIZE)
+#if defined(UNROLL)
+#define PARTIAL_UNROLL
+#undef UNROLL
+#endif
+#endif
+
+// the finite field modular polynomial and elements
+
+#define ff_poly 0x011b
+#define ff_hi 0x80
+
+// multiply four bytes in GF(2^8) by 'x' {02} in parallel
+
+#define m1 0x80808080
+#define m2 0x7f7f7f7f
+#define m3 0x0000001b
+#define FFmulX(x) ((((x) & m2) << 1) ^ ((((x) & m1) >> 7) * m3))
+
+// The following defines provide alternative definitions of FFmulX that might
+// give improved performance if a fast 32-bit multiply is not available. Note
+// that a temporary variable u needs to be defined where FFmulX is used.
+
+// #define FFmulX(x) (u = (x) & m1, u |= (u >> 1), ((x) & m2) << 1) ^ ((u >> 3) | (u >> 6))
+// #define m4 0x1b1b1b1b
+// #define FFmulX(x) (u = (x) & m1, ((x) & m2) << 1) ^ ((u - (u >> 7)) & m4)
+
+// perform column mix operation on four bytes in parallel
+
+#define fwd_mcol(x) (f2 = FFmulX(x), f2 ^ upr(x ^ f2,3) ^ upr(x,2) ^ upr(x,1))
+
+#if defined(FIXED_TABLES)
+
+// the S-Box table
+
+static const unsigned char s_box[256] =
+{
+ 0x63, 0x7c, 0x77, 0x7b, 0xf2, 0x6b, 0x6f, 0xc5,
+ 0x30, 0x01, 0x67, 0x2b, 0xfe, 0xd7, 0xab, 0x76,
+ 0xca, 0x82, 0xc9, 0x7d, 0xfa, 0x59, 0x47, 0xf0,
+ 0xad, 0xd4, 0xa2, 0xaf, 0x9c, 0xa4, 0x72, 0xc0,
+ 0xb7, 0xfd, 0x93, 0x26, 0x36, 0x3f, 0xf7, 0xcc,
+ 0x34, 0xa5, 0xe5, 0xf1, 0x71, 0xd8, 0x31, 0x15,
+ 0x04, 0xc7, 0x23, 0xc3, 0x18, 0x96, 0x05, 0x9a,
+ 0x07, 0x12, 0x80, 0xe2, 0xeb, 0x27, 0xb2, 0x75,
+ 0x09, 0x83, 0x2c, 0x1a, 0x1b, 0x6e, 0x5a, 0xa0,
+ 0x52, 0x3b, 0xd6, 0xb3, 0x29, 0xe3, 0x2f, 0x84,
+ 0x53, 0xd1, 0x00, 0xed, 0x20, 0xfc, 0xb1, 0x5b,
+ 0x6a, 0xcb, 0xbe, 0x39, 0x4a, 0x4c, 0x58, 0xcf,
+ 0xd0, 0xef, 0xaa, 0xfb, 0x43, 0x4d, 0x33, 0x85,
+ 0x45, 0xf9, 0x02, 0x7f, 0x50, 0x3c, 0x9f, 0xa8,
+ 0x51, 0xa3, 0x40, 0x8f, 0x92, 0x9d, 0x38, 0xf5,
+ 0xbc, 0xb6, 0xda, 0x21, 0x10, 0xff, 0xf3, 0xd2,
+ 0xcd, 0x0c, 0x13, 0xec, 0x5f, 0x97, 0x44, 0x17,
+ 0xc4, 0xa7, 0x7e, 0x3d, 0x64, 0x5d, 0x19, 0x73,
+ 0x60, 0x81, 0x4f, 0xdc, 0x22, 0x2a, 0x90, 0x88,
+ 0x46, 0xee, 0xb8, 0x14, 0xde, 0x5e, 0x0b, 0xdb,
+ 0xe0, 0x32, 0x3a, 0x0a, 0x49, 0x06, 0x24, 0x5c,
+ 0xc2, 0xd3, 0xac, 0x62, 0x91, 0x95, 0xe4, 0x79,
+ 0xe7, 0xc8, 0x37, 0x6d, 0x8d, 0xd5, 0x4e, 0xa9,
+ 0x6c, 0x56, 0xf4, 0xea, 0x65, 0x7a, 0xae, 0x08,
+ 0xba, 0x78, 0x25, 0x2e, 0x1c, 0xa6, 0xb4, 0xc6,
+ 0xe8, 0xdd, 0x74, 0x1f, 0x4b, 0xbd, 0x8b, 0x8a,
+ 0x70, 0x3e, 0xb5, 0x66, 0x48, 0x03, 0xf6, 0x0e,
+ 0x61, 0x35, 0x57, 0xb9, 0x86, 0xc1, 0x1d, 0x9e,
+ 0xe1, 0xf8, 0x98, 0x11, 0x69, 0xd9, 0x8e, 0x94,
+ 0x9b, 0x1e, 0x87, 0xe9, 0xce, 0x55, 0x28, 0xdf,
+ 0x8c, 0xa1, 0x89, 0x0d, 0xbf, 0xe6, 0x42, 0x68,
+ 0x41, 0x99, 0x2d, 0x0f, 0xb0, 0x54, 0xbb, 0x16
+};
+
+// the inverse S-Box table
+
+static const unsigned char inv_s_box[256] =
+{
+ 0x52, 0x09, 0x6a, 0xd5, 0x30, 0x36, 0xa5, 0x38,
+ 0xbf, 0x40, 0xa3, 0x9e, 0x81, 0xf3, 0xd7, 0xfb,
+ 0x7c, 0xe3, 0x39, 0x82, 0x9b, 0x2f, 0xff, 0x87,
+ 0x34, 0x8e, 0x43, 0x44, 0xc4, 0xde, 0xe9, 0xcb,
+ 0x54, 0x7b, 0x94, 0x32, 0xa6, 0xc2, 0x23, 0x3d,
+ 0xee, 0x4c, 0x95, 0x0b, 0x42, 0xfa, 0xc3, 0x4e,
+ 0x08, 0x2e, 0xa1, 0x66, 0x28, 0xd9, 0x24, 0xb2,
+ 0x76, 0x5b, 0xa2, 0x49, 0x6d, 0x8b, 0xd1, 0x25,
+ 0x72, 0xf8, 0xf6, 0x64, 0x86, 0x68, 0x98, 0x16,
+ 0xd4, 0xa4, 0x5c, 0xcc, 0x5d, 0x65, 0xb6, 0x92,
+ 0x6c, 0x70, 0x48, 0x50, 0xfd, 0xed, 0xb9, 0xda,
+ 0x5e, 0x15, 0x46, 0x57, 0xa7, 0x8d, 0x9d, 0x84,
+ 0x90, 0xd8, 0xab, 0x00, 0x8c, 0xbc, 0xd3, 0x0a,
+ 0xf7, 0xe4, 0x58, 0x05, 0xb8, 0xb3, 0x45, 0x06,
+ 0xd0, 0x2c, 0x1e, 0x8f, 0xca, 0x3f, 0x0f, 0x02,
+ 0xc1, 0xaf, 0xbd, 0x03, 0x01, 0x13, 0x8a, 0x6b,
+ 0x3a, 0x91, 0x11, 0x41, 0x4f, 0x67, 0xdc, 0xea,
+ 0x97, 0xf2, 0xcf, 0xce, 0xf0, 0xb4, 0xe6, 0x73,
+ 0x96, 0xac, 0x74, 0x22, 0xe7, 0xad, 0x35, 0x85,
+ 0xe2, 0xf9, 0x37, 0xe8, 0x1c, 0x75, 0xdf, 0x6e,
+ 0x47, 0xf1, 0x1a, 0x71, 0x1d, 0x29, 0xc5, 0x89,
+ 0x6f, 0xb7, 0x62, 0x0e, 0xaa, 0x18, 0xbe, 0x1b,
+ 0xfc, 0x56, 0x3e, 0x4b, 0xc6, 0xd2, 0x79, 0x20,
+ 0x9a, 0xdb, 0xc0, 0xfe, 0x78, 0xcd, 0x5a, 0xf4,
+ 0x1f, 0xdd, 0xa8, 0x33, 0x88, 0x07, 0xc7, 0x31,
+ 0xb1, 0x12, 0x10, 0x59, 0x27, 0x80, 0xec, 0x5f,
+ 0x60, 0x51, 0x7f, 0xa9, 0x19, 0xb5, 0x4a, 0x0d,
+ 0x2d, 0xe5, 0x7a, 0x9f, 0x93, 0xc9, 0x9c, 0xef,
+ 0xa0, 0xe0, 0x3b, 0x4d, 0xae, 0x2a, 0xf5, 0xb0,
+ 0xc8, 0xeb, 0xbb, 0x3c, 0x83, 0x53, 0x99, 0x61,
+ 0x17, 0x2b, 0x04, 0x7e, 0xba, 0x77, 0xd6, 0x26,
+ 0xe1, 0x69, 0x14, 0x63, 0x55, 0x21, 0x0c, 0x7d
+};
+
+// used to ensure table is generated in the right format
+// depending on the internal byte order required
+
+#if (INTERNAL_BYTE_ORDER < 2)
+/* little endian */
+#define w0(p) 0x000000##p
+#else
+/* big endian */
+#define w0(p) 0x##p##000000
+#endif
+
+// Number of elements required in this table for different
+// block and key lengths is:
+//
+// Nk = 4 6 8
+// ----------
+// Nb = 4 | 10 8 7
+// 6 | 19 12 11
+// 8 | 29 19 14
+//
+// this table can be a table of bytes if the key schedule
+// code is adjusted accordingly
+
+static const u_int32_t rcon_tab[29] =
+{
+ w0(01), w0(02), w0(04), w0(08),
+ w0(10), w0(20), w0(40), w0(80),
+ w0(1b), w0(36), w0(6c), w0(d8),
+ w0(ab), w0(4d), w0(9a), w0(2f),
+ w0(5e), w0(bc), w0(63), w0(c6),
+ w0(97), w0(35), w0(6a), w0(d4),
+ w0(b3), w0(7d), w0(fa), w0(ef),
+ w0(c5)
+};
+
+#undef w0
+
+// used to ensure table is generated in the right format
+// depending on the internal byte order required
+
+#if (INTERNAL_BYTE_ORDER < 2)
+/* little endian */
+#define r0(p,q,r,s) 0x##p##q##r##s
+#define r1(p,q,r,s) 0x##q##r##s##p
+#define r2(p,q,r,s) 0x##r##s##p##q
+#define r3(p,q,r,s) 0x##s##p##q##r
+#define w0(p) 0x000000##p
+#define w1(p) 0x0000##p##00
+#define w2(p) 0x00##p##0000
+#define w3(p) 0x##p##000000
+#else
+/* big endian */
+#define r0(p,q,r,s) 0x##s##r##q##p
+#define r1(p,q,r,s) 0x##p##s##r##q
+#define r2(p,q,r,s) 0x##q##p##s##r
+#define r3(p,q,r,s) 0x##r##q##p##s
+#define w0(p) 0x##p##000000
+#define w1(p) 0x00##p##0000
+#define w2(p) 0x0000##p##00
+#define w3(p) 0x000000##p
+#endif
+
+#if defined(FIXED_TABLES) && (defined(ONE_TABLE) || defined(FOUR_TABLES))
+
+// data for forward tables (other than last round)
+
+#define f_table \
+ r(a5,63,63,c6), r(84,7c,7c,f8), r(99,77,77,ee), r(8d,7b,7b,f6),\
+ r(0d,f2,f2,ff), r(bd,6b,6b,d6), r(b1,6f,6f,de), r(54,c5,c5,91),\
+ r(50,30,30,60), r(03,01,01,02), r(a9,67,67,ce), r(7d,2b,2b,56),\
+ r(19,fe,fe,e7), r(62,d7,d7,b5), r(e6,ab,ab,4d), r(9a,76,76,ec),\
+ r(45,ca,ca,8f), r(9d,82,82,1f), r(40,c9,c9,89), r(87,7d,7d,fa),\
+ r(15,fa,fa,ef), r(eb,59,59,b2), r(c9,47,47,8e), r(0b,f0,f0,fb),\
+ r(ec,ad,ad,41), r(67,d4,d4,b3), r(fd,a2,a2,5f), r(ea,af,af,45),\
+ r(bf,9c,9c,23), r(f7,a4,a4,53), r(96,72,72,e4), r(5b,c0,c0,9b),\
+ r(c2,b7,b7,75), r(1c,fd,fd,e1), r(ae,93,93,3d), r(6a,26,26,4c),\
+ r(5a,36,36,6c), r(41,3f,3f,7e), r(02,f7,f7,f5), r(4f,cc,cc,83),\
+ r(5c,34,34,68), r(f4,a5,a5,51), r(34,e5,e5,d1), r(08,f1,f1,f9),\
+ r(93,71,71,e2), r(73,d8,d8,ab), r(53,31,31,62), r(3f,15,15,2a),\
+ r(0c,04,04,08), r(52,c7,c7,95), r(65,23,23,46), r(5e,c3,c3,9d),\
+ r(28,18,18,30), r(a1,96,96,37), r(0f,05,05,0a), r(b5,9a,9a,2f),\
+ r(09,07,07,0e), r(36,12,12,24), r(9b,80,80,1b), r(3d,e2,e2,df),\
+ r(26,eb,eb,cd), r(69,27,27,4e), r(cd,b2,b2,7f), r(9f,75,75,ea),\
+ r(1b,09,09,12), r(9e,83,83,1d), r(74,2c,2c,58), r(2e,1a,1a,34),\
+ r(2d,1b,1b,36), r(b2,6e,6e,dc), r(ee,5a,5a,b4), r(fb,a0,a0,5b),\
+ r(f6,52,52,a4), r(4d,3b,3b,76), r(61,d6,d6,b7), r(ce,b3,b3,7d),\
+ r(7b,29,29,52), r(3e,e3,e3,dd), r(71,2f,2f,5e), r(97,84,84,13),\
+ r(f5,53,53,a6), r(68,d1,d1,b9), r(00,00,00,00), r(2c,ed,ed,c1),\
+ r(60,20,20,40), r(1f,fc,fc,e3), r(c8,b1,b1,79), r(ed,5b,5b,b6),\
+ r(be,6a,6a,d4), r(46,cb,cb,8d), r(d9,be,be,67), r(4b,39,39,72),\
+ r(de,4a,4a,94), r(d4,4c,4c,98), r(e8,58,58,b0), r(4a,cf,cf,85),\
+ r(6b,d0,d0,bb), r(2a,ef,ef,c5), r(e5,aa,aa,4f), r(16,fb,fb,ed),\
+ r(c5,43,43,86), r(d7,4d,4d,9a), r(55,33,33,66), r(94,85,85,11),\
+ r(cf,45,45,8a), r(10,f9,f9,e9), r(06,02,02,04), r(81,7f,7f,fe),\
+ r(f0,50,50,a0), r(44,3c,3c,78), r(ba,9f,9f,25), r(e3,a8,a8,4b),\
+ r(f3,51,51,a2), r(fe,a3,a3,5d), r(c0,40,40,80), r(8a,8f,8f,05),\
+ r(ad,92,92,3f), r(bc,9d,9d,21), r(48,38,38,70), r(04,f5,f5,f1),\
+ r(df,bc,bc,63), r(c1,b6,b6,77), r(75,da,da,af), r(63,21,21,42),\
+ r(30,10,10,20), r(1a,ff,ff,e5), r(0e,f3,f3,fd), r(6d,d2,d2,bf),\
+ r(4c,cd,cd,81), r(14,0c,0c,18), r(35,13,13,26), r(2f,ec,ec,c3),\
+ r(e1,5f,5f,be), r(a2,97,97,35), r(cc,44,44,88), r(39,17,17,2e),\
+ r(57,c4,c4,93), r(f2,a7,a7,55), r(82,7e,7e,fc), r(47,3d,3d,7a),\
+ r(ac,64,64,c8), r(e7,5d,5d,ba), r(2b,19,19,32), r(95,73,73,e6),\
+ r(a0,60,60,c0), r(98,81,81,19), r(d1,4f,4f,9e), r(7f,dc,dc,a3),\
+ r(66,22,22,44), r(7e,2a,2a,54), r(ab,90,90,3b), r(83,88,88,0b),\
+ r(ca,46,46,8c), r(29,ee,ee,c7), r(d3,b8,b8,6b), r(3c,14,14,28),\
+ r(79,de,de,a7), r(e2,5e,5e,bc), r(1d,0b,0b,16), r(76,db,db,ad),\
+ r(3b,e0,e0,db), r(56,32,32,64), r(4e,3a,3a,74), r(1e,0a,0a,14),\
+ r(db,49,49,92), r(0a,06,06,0c), r(6c,24,24,48), r(e4,5c,5c,b8),\
+ r(5d,c2,c2,9f), r(6e,d3,d3,bd), r(ef,ac,ac,43), r(a6,62,62,c4),\
+ r(a8,91,91,39), r(a4,95,95,31), r(37,e4,e4,d3), r(8b,79,79,f2),\
+ r(32,e7,e7,d5), r(43,c8,c8,8b), r(59,37,37,6e), r(b7,6d,6d,da),\
+ r(8c,8d,8d,01), r(64,d5,d5,b1), r(d2,4e,4e,9c), r(e0,a9,a9,49),\
+ r(b4,6c,6c,d8), r(fa,56,56,ac), r(07,f4,f4,f3), r(25,ea,ea,cf),\
+ r(af,65,65,ca), r(8e,7a,7a,f4), r(e9,ae,ae,47), r(18,08,08,10),\
+ r(d5,ba,ba,6f), r(88,78,78,f0), r(6f,25,25,4a), r(72,2e,2e,5c),\
+ r(24,1c,1c,38), r(f1,a6,a6,57), r(c7,b4,b4,73), r(51,c6,c6,97),\
+ r(23,e8,e8,cb), r(7c,dd,dd,a1), r(9c,74,74,e8), r(21,1f,1f,3e),\
+ r(dd,4b,4b,96), r(dc,bd,bd,61), r(86,8b,8b,0d), r(85,8a,8a,0f),\
+ r(90,70,70,e0), r(42,3e,3e,7c), r(c4,b5,b5,71), r(aa,66,66,cc),\
+ r(d8,48,48,90), r(05,03,03,06), r(01,f6,f6,f7), r(12,0e,0e,1c),\
+ r(a3,61,61,c2), r(5f,35,35,6a), r(f9,57,57,ae), r(d0,b9,b9,69),\
+ r(91,86,86,17), r(58,c1,c1,99), r(27,1d,1d,3a), r(b9,9e,9e,27),\
+ r(38,e1,e1,d9), r(13,f8,f8,eb), r(b3,98,98,2b), r(33,11,11,22),\
+ r(bb,69,69,d2), r(70,d9,d9,a9), r(89,8e,8e,07), r(a7,94,94,33),\
+ r(b6,9b,9b,2d), r(22,1e,1e,3c), r(92,87,87,15), r(20,e9,e9,c9),\
+ r(49,ce,ce,87), r(ff,55,55,aa), r(78,28,28,50), r(7a,df,df,a5),\
+ r(8f,8c,8c,03), r(f8,a1,a1,59), r(80,89,89,09), r(17,0d,0d,1a),\
+ r(da,bf,bf,65), r(31,e6,e6,d7), r(c6,42,42,84), r(b8,68,68,d0),\
+ r(c3,41,41,82), r(b0,99,99,29), r(77,2d,2d,5a), r(11,0f,0f,1e),\
+ r(cb,b0,b0,7b), r(fc,54,54,a8), r(d6,bb,bb,6d), r(3a,16,16,2c)
+
+// data for inverse tables (other than last round)
+
+#define i_table \
+ r(50,a7,f4,51), r(53,65,41,7e), r(c3,a4,17,1a), r(96,5e,27,3a),\
+ r(cb,6b,ab,3b), r(f1,45,9d,1f), r(ab,58,fa,ac), r(93,03,e3,4b),\
+ r(55,fa,30,20), r(f6,6d,76,ad), r(91,76,cc,88), r(25,4c,02,f5),\
+ r(fc,d7,e5,4f), r(d7,cb,2a,c5), r(80,44,35,26), r(8f,a3,62,b5),\
+ r(49,5a,b1,de), r(67,1b,ba,25), r(98,0e,ea,45), r(e1,c0,fe,5d),\
+ r(02,75,2f,c3), r(12,f0,4c,81), r(a3,97,46,8d), r(c6,f9,d3,6b),\
+ r(e7,5f,8f,03), r(95,9c,92,15), r(eb,7a,6d,bf), r(da,59,52,95),\
+ r(2d,83,be,d4), r(d3,21,74,58), r(29,69,e0,49), r(44,c8,c9,8e),\
+ r(6a,89,c2,75), r(78,79,8e,f4), r(6b,3e,58,99), r(dd,71,b9,27),\
+ r(b6,4f,e1,be), r(17,ad,88,f0), r(66,ac,20,c9), r(b4,3a,ce,7d),\
+ r(18,4a,df,63), r(82,31,1a,e5), r(60,33,51,97), r(45,7f,53,62),\
+ r(e0,77,64,b1), r(84,ae,6b,bb), r(1c,a0,81,fe), r(94,2b,08,f9),\
+ r(58,68,48,70), r(19,fd,45,8f), r(87,6c,de,94), r(b7,f8,7b,52),\
+ r(23,d3,73,ab), r(e2,02,4b,72), r(57,8f,1f,e3), r(2a,ab,55,66),\
+ r(07,28,eb,b2), r(03,c2,b5,2f), r(9a,7b,c5,86), r(a5,08,37,d3),\
+ r(f2,87,28,30), r(b2,a5,bf,23), r(ba,6a,03,02), r(5c,82,16,ed),\
+ r(2b,1c,cf,8a), r(92,b4,79,a7), r(f0,f2,07,f3), r(a1,e2,69,4e),\
+ r(cd,f4,da,65), r(d5,be,05,06), r(1f,62,34,d1), r(8a,fe,a6,c4),\
+ r(9d,53,2e,34), r(a0,55,f3,a2), r(32,e1,8a,05), r(75,eb,f6,a4),\
+ r(39,ec,83,0b), r(aa,ef,60,40), r(06,9f,71,5e), r(51,10,6e,bd),\
+ r(f9,8a,21,3e), r(3d,06,dd,96), r(ae,05,3e,dd), r(46,bd,e6,4d),\
+ r(b5,8d,54,91), r(05,5d,c4,71), r(6f,d4,06,04), r(ff,15,50,60),\
+ r(24,fb,98,19), r(97,e9,bd,d6), r(cc,43,40,89), r(77,9e,d9,67),\
+ r(bd,42,e8,b0), r(88,8b,89,07), r(38,5b,19,e7), r(db,ee,c8,79),\
+ r(47,0a,7c,a1), r(e9,0f,42,7c), r(c9,1e,84,f8), r(00,00,00,00),\
+ r(83,86,80,09), r(48,ed,2b,32), r(ac,70,11,1e), r(4e,72,5a,6c),\
+ r(fb,ff,0e,fd), r(56,38,85,0f), r(1e,d5,ae,3d), r(27,39,2d,36),\
+ r(64,d9,0f,0a), r(21,a6,5c,68), r(d1,54,5b,9b), r(3a,2e,36,24),\
+ r(b1,67,0a,0c), r(0f,e7,57,93), r(d2,96,ee,b4), r(9e,91,9b,1b),\
+ r(4f,c5,c0,80), r(a2,20,dc,61), r(69,4b,77,5a), r(16,1a,12,1c),\
+ r(0a,ba,93,e2), r(e5,2a,a0,c0), r(43,e0,22,3c), r(1d,17,1b,12),\
+ r(0b,0d,09,0e), r(ad,c7,8b,f2), r(b9,a8,b6,2d), r(c8,a9,1e,14),\
+ r(85,19,f1,57), r(4c,07,75,af), r(bb,dd,99,ee), r(fd,60,7f,a3),\
+ r(9f,26,01,f7), r(bc,f5,72,5c), r(c5,3b,66,44), r(34,7e,fb,5b),\
+ r(76,29,43,8b), r(dc,c6,23,cb), r(68,fc,ed,b6), r(63,f1,e4,b8),\
+ r(ca,dc,31,d7), r(10,85,63,42), r(40,22,97,13), r(20,11,c6,84),\
+ r(7d,24,4a,85), r(f8,3d,bb,d2), r(11,32,f9,ae), r(6d,a1,29,c7),\
+ r(4b,2f,9e,1d), r(f3,30,b2,dc), r(ec,52,86,0d), r(d0,e3,c1,77),\
+ r(6c,16,b3,2b), r(99,b9,70,a9), r(fa,48,94,11), r(22,64,e9,47),\
+ r(c4,8c,fc,a8), r(1a,3f,f0,a0), r(d8,2c,7d,56), r(ef,90,33,22),\
+ r(c7,4e,49,87), r(c1,d1,38,d9), r(fe,a2,ca,8c), r(36,0b,d4,98),\
+ r(cf,81,f5,a6), r(28,de,7a,a5), r(26,8e,b7,da), r(a4,bf,ad,3f),\
+ r(e4,9d,3a,2c), r(0d,92,78,50), r(9b,cc,5f,6a), r(62,46,7e,54),\
+ r(c2,13,8d,f6), r(e8,b8,d8,90), r(5e,f7,39,2e), r(f5,af,c3,82),\
+ r(be,80,5d,9f), r(7c,93,d0,69), r(a9,2d,d5,6f), r(b3,12,25,cf),\
+ r(3b,99,ac,c8), r(a7,7d,18,10), r(6e,63,9c,e8), r(7b,bb,3b,db),\
+ r(09,78,26,cd), r(f4,18,59,6e), r(01,b7,9a,ec), r(a8,9a,4f,83),\
+ r(65,6e,95,e6), r(7e,e6,ff,aa), r(08,cf,bc,21), r(e6,e8,15,ef),\
+ r(d9,9b,e7,ba), r(ce,36,6f,4a), r(d4,09,9f,ea), r(d6,7c,b0,29),\
+ r(af,b2,a4,31), r(31,23,3f,2a), r(30,94,a5,c6), r(c0,66,a2,35),\
+ r(37,bc,4e,74), r(a6,ca,82,fc), r(b0,d0,90,e0), r(15,d8,a7,33),\
+ r(4a,98,04,f1), r(f7,da,ec,41), r(0e,50,cd,7f), r(2f,f6,91,17),\
+ r(8d,d6,4d,76), r(4d,b0,ef,43), r(54,4d,aa,cc), r(df,04,96,e4),\
+ r(e3,b5,d1,9e), r(1b,88,6a,4c), r(b8,1f,2c,c1), r(7f,51,65,46),\
+ r(04,ea,5e,9d), r(5d,35,8c,01), r(73,74,87,fa), r(2e,41,0b,fb),\
+ r(5a,1d,67,b3), r(52,d2,db,92), r(33,56,10,e9), r(13,47,d6,6d),\
+ r(8c,61,d7,9a), r(7a,0c,a1,37), r(8e,14,f8,59), r(89,3c,13,eb),\
+ r(ee,27,a9,ce), r(35,c9,61,b7), r(ed,e5,1c,e1), r(3c,b1,47,7a),\
+ r(59,df,d2,9c), r(3f,73,f2,55), r(79,ce,14,18), r(bf,37,c7,73),\
+ r(ea,cd,f7,53), r(5b,aa,fd,5f), r(14,6f,3d,df), r(86,db,44,78),\
+ r(81,f3,af,ca), r(3e,c4,68,b9), r(2c,34,24,38), r(5f,40,a3,c2),\
+ r(72,c3,1d,16), r(0c,25,e2,bc), r(8b,49,3c,28), r(41,95,0d,ff),\
+ r(71,01,a8,39), r(de,b3,0c,08), r(9c,e4,b4,d8), r(90,c1,56,64),\
+ r(61,84,cb,7b), r(70,b6,32,d5), r(74,5c,6c,48), r(42,57,b8,d0)
+
+// generate the required tables in the desired endian format
+
+#undef r
+#define r r0
+
+#if defined(ONE_TABLE)
+static const u_int32_t ft_tab[256] =
+ { f_table };
+#elif defined(FOUR_TABLES)
+static const u_int32_t ft_tab[4][256] =
+{ { f_table },
+#undef r
+#define r r1
+ { f_table },
+#undef r
+#define r r2
+ { f_table },
+#undef r
+#define r r3
+ { f_table }
+};
+#endif
+
+#undef r
+#define r r0
+#if defined(ONE_TABLE)
+static const u_int32_t it_tab[256] =
+ { i_table };
+#elif defined(FOUR_TABLES)
+static const u_int32_t it_tab[4][256] =
+{ { i_table },
+#undef r
+#define r r1
+ { i_table },
+#undef r
+#define r r2
+ { i_table },
+#undef r
+#define r r3
+ { i_table }
+};
+#endif
+
+#endif
+
+#if defined(FIXED_TABLES) && (defined(ONE_LR_TABLE) || defined(FOUR_LR_TABLES))
+
+// data for inverse tables (last round)
+
+#define li_table \
+ w(52), w(09), w(6a), w(d5), w(30), w(36), w(a5), w(38),\
+ w(bf), w(40), w(a3), w(9e), w(81), w(f3), w(d7), w(fb),\
+ w(7c), w(e3), w(39), w(82), w(9b), w(2f), w(ff), w(87),\
+ w(34), w(8e), w(43), w(44), w(c4), w(de), w(e9), w(cb),\
+ w(54), w(7b), w(94), w(32), w(a6), w(c2), w(23), w(3d),\
+ w(ee), w(4c), w(95), w(0b), w(42), w(fa), w(c3), w(4e),\
+ w(08), w(2e), w(a1), w(66), w(28), w(d9), w(24), w(b2),\
+ w(76), w(5b), w(a2), w(49), w(6d), w(8b), w(d1), w(25),\
+ w(72), w(f8), w(f6), w(64), w(86), w(68), w(98), w(16),\
+ w(d4), w(a4), w(5c), w(cc), w(5d), w(65), w(b6), w(92),\
+ w(6c), w(70), w(48), w(50), w(fd), w(ed), w(b9), w(da),\
+ w(5e), w(15), w(46), w(57), w(a7), w(8d), w(9d), w(84),\
+ w(90), w(d8), w(ab), w(00), w(8c), w(bc), w(d3), w(0a),\
+ w(f7), w(e4), w(58), w(05), w(b8), w(b3), w(45), w(06),\
+ w(d0), w(2c), w(1e), w(8f), w(ca), w(3f), w(0f), w(02),\
+ w(c1), w(af), w(bd), w(03), w(01), w(13), w(8a), w(6b),\
+ w(3a), w(91), w(11), w(41), w(4f), w(67), w(dc), w(ea),\
+ w(97), w(f2), w(cf), w(ce), w(f0), w(b4), w(e6), w(73),\
+ w(96), w(ac), w(74), w(22), w(e7), w(ad), w(35), w(85),\
+ w(e2), w(f9), w(37), w(e8), w(1c), w(75), w(df), w(6e),\
+ w(47), w(f1), w(1a), w(71), w(1d), w(29), w(c5), w(89),\
+ w(6f), w(b7), w(62), w(0e), w(aa), w(18), w(be), w(1b),\
+ w(fc), w(56), w(3e), w(4b), w(c6), w(d2), w(79), w(20),\
+ w(9a), w(db), w(c0), w(fe), w(78), w(cd), w(5a), w(f4),\
+ w(1f), w(dd), w(a8), w(33), w(88), w(07), w(c7), w(31),\
+ w(b1), w(12), w(10), w(59), w(27), w(80), w(ec), w(5f),\
+ w(60), w(51), w(7f), w(a9), w(19), w(b5), w(4a), w(0d),\
+ w(2d), w(e5), w(7a), w(9f), w(93), w(c9), w(9c), w(ef),\
+ w(a0), w(e0), w(3b), w(4d), w(ae), w(2a), w(f5), w(b0),\
+ w(c8), w(eb), w(bb), w(3c), w(83), w(53), w(99), w(61),\
+ w(17), w(2b), w(04), w(7e), w(ba), w(77), w(d6), w(26),\
+ w(e1), w(69), w(14), w(63), w(55), w(21), w(0c), w(7d),
+
+// generate the required tables in the desired endian format
+
+#undef r
+#define r(p,q,r,s) w0(q)
+#if defined(ONE_LR_TABLE)
+static const u_int32_t fl_tab[256] =
+ { f_table };
+#elif defined(FOUR_LR_TABLES)
+static const u_int32_t fl_tab[4][256] =
+{ { f_table },
+#undef r
+#define r(p,q,r,s) w1(q)
+ { f_table },
+#undef r
+#define r(p,q,r,s) w2(q)
+ { f_table },
+#undef r
+#define r(p,q,r,s) w3(q)
+ { f_table }
+};
+#endif
+
+#undef w
+#define w w0
+#if defined(ONE_LR_TABLE)
+static const u_int32_t il_tab[256] =
+ { li_table };
+#elif defined(FOUR_LR_TABLES)
+static const u_int32_t il_tab[4][256] =
+{ { li_table },
+#undef w
+#define w w1
+ { li_table },
+#undef w
+#define w w2
+ { li_table },
+#undef w
+#define w w3
+ { li_table }
+};
+#endif
+
+#endif
+
+#if defined(FIXED_TABLES) && (defined(ONE_IM_TABLE) || defined(FOUR_IM_TABLES))
+
+#define m_table \
+ r(00,00,00,00), r(0b,0d,09,0e), r(16,1a,12,1c), r(1d,17,1b,12),\
+ r(2c,34,24,38), r(27,39,2d,36), r(3a,2e,36,24), r(31,23,3f,2a),\
+ r(58,68,48,70), r(53,65,41,7e), r(4e,72,5a,6c), r(45,7f,53,62),\
+ r(74,5c,6c,48), r(7f,51,65,46), r(62,46,7e,54), r(69,4b,77,5a),\
+ r(b0,d0,90,e0), r(bb,dd,99,ee), r(a6,ca,82,fc), r(ad,c7,8b,f2),\
+ r(9c,e4,b4,d8), r(97,e9,bd,d6), r(8a,fe,a6,c4), r(81,f3,af,ca),\
+ r(e8,b8,d8,90), r(e3,b5,d1,9e), r(fe,a2,ca,8c), r(f5,af,c3,82),\
+ r(c4,8c,fc,a8), r(cf,81,f5,a6), r(d2,96,ee,b4), r(d9,9b,e7,ba),\
+ r(7b,bb,3b,db), r(70,b6,32,d5), r(6d,a1,29,c7), r(66,ac,20,c9),\
+ r(57,8f,1f,e3), r(5c,82,16,ed), r(41,95,0d,ff), r(4a,98,04,f1),\
+ r(23,d3,73,ab), r(28,de,7a,a5), r(35,c9,61,b7), r(3e,c4,68,b9),\
+ r(0f,e7,57,93), r(04,ea,5e,9d), r(19,fd,45,8f), r(12,f0,4c,81),\
+ r(cb,6b,ab,3b), r(c0,66,a2,35), r(dd,71,b9,27), r(d6,7c,b0,29),\
+ r(e7,5f,8f,03), r(ec,52,86,0d), r(f1,45,9d,1f), r(fa,48,94,11),\
+ r(93,03,e3,4b), r(98,0e,ea,45), r(85,19,f1,57), r(8e,14,f8,59),\
+ r(bf,37,c7,73), r(b4,3a,ce,7d), r(a9,2d,d5,6f), r(a2,20,dc,61),\
+ r(f6,6d,76,ad), r(fd,60,7f,a3), r(e0,77,64,b1), r(eb,7a,6d,bf),\
+ r(da,59,52,95), r(d1,54,5b,9b), r(cc,43,40,89), r(c7,4e,49,87),\
+ r(ae,05,3e,dd), r(a5,08,37,d3), r(b8,1f,2c,c1), r(b3,12,25,cf),\
+ r(82,31,1a,e5), r(89,3c,13,eb), r(94,2b,08,f9), r(9f,26,01,f7),\
+ r(46,bd,e6,4d), r(4d,b0,ef,43), r(50,a7,f4,51), r(5b,aa,fd,5f),\
+ r(6a,89,c2,75), r(61,84,cb,7b), r(7c,93,d0,69), r(77,9e,d9,67),\
+ r(1e,d5,ae,3d), r(15,d8,a7,33), r(08,cf,bc,21), r(03,c2,b5,2f),\
+ r(32,e1,8a,05), r(39,ec,83,0b), r(24,fb,98,19), r(2f,f6,91,17),\
+ r(8d,d6,4d,76), r(86,db,44,78), r(9b,cc,5f,6a), r(90,c1,56,64),\
+ r(a1,e2,69,4e), r(aa,ef,60,40), r(b7,f8,7b,52), r(bc,f5,72,5c),\
+ r(d5,be,05,06), r(de,b3,0c,08), r(c3,a4,17,1a), r(c8,a9,1e,14),\
+ r(f9,8a,21,3e), r(f2,87,28,30), r(ef,90,33,22), r(e4,9d,3a,2c),\
+ r(3d,06,dd,96), r(36,0b,d4,98), r(2b,1c,cf,8a), r(20,11,c6,84),\
+ r(11,32,f9,ae), r(1a,3f,f0,a0), r(07,28,eb,b2), r(0c,25,e2,bc),\
+ r(65,6e,95,e6), r(6e,63,9c,e8), r(73,74,87,fa), r(78,79,8e,f4),\
+ r(49,5a,b1,de), r(42,57,b8,d0), r(5f,40,a3,c2), r(54,4d,aa,cc),\
+ r(f7,da,ec,41), r(fc,d7,e5,4f), r(e1,c0,fe,5d), r(ea,cd,f7,53),\
+ r(db,ee,c8,79), r(d0,e3,c1,77), r(cd,f4,da,65), r(c6,f9,d3,6b),\
+ r(af,b2,a4,31), r(a4,bf,ad,3f), r(b9,a8,b6,2d), r(b2,a5,bf,23),\
+ r(83,86,80,09), r(88,8b,89,07), r(95,9c,92,15), r(9e,91,9b,1b),\
+ r(47,0a,7c,a1), r(4c,07,75,af), r(51,10,6e,bd), r(5a,1d,67,b3),\
+ r(6b,3e,58,99), r(60,33,51,97), r(7d,24,4a,85), r(76,29,43,8b),\
+ r(1f,62,34,d1), r(14,6f,3d,df), r(09,78,26,cd), r(02,75,2f,c3),\
+ r(33,56,10,e9), r(38,5b,19,e7), r(25,4c,02,f5), r(2e,41,0b,fb),\
+ r(8c,61,d7,9a), r(87,6c,de,94), r(9a,7b,c5,86), r(91,76,cc,88),\
+ r(a0,55,f3,a2), r(ab,58,fa,ac), r(b6,4f,e1,be), r(bd,42,e8,b0),\
+ r(d4,09,9f,ea), r(df,04,96,e4), r(c2,13,8d,f6), r(c9,1e,84,f8),\
+ r(f8,3d,bb,d2), r(f3,30,b2,dc), r(ee,27,a9,ce), r(e5,2a,a0,c0),\
+ r(3c,b1,47,7a), r(37,bc,4e,74), r(2a,ab,55,66), r(21,a6,5c,68),\
+ r(10,85,63,42), r(1b,88,6a,4c), r(06,9f,71,5e), r(0d,92,78,50),\
+ r(64,d9,0f,0a), r(6f,d4,06,04), r(72,c3,1d,16), r(79,ce,14,18),\
+ r(48,ed,2b,32), r(43,e0,22,3c), r(5e,f7,39,2e), r(55,fa,30,20),\
+ r(01,b7,9a,ec), r(0a,ba,93,e2), r(17,ad,88,f0), r(1c,a0,81,fe),\
+ r(2d,83,be,d4), r(26,8e,b7,da), r(3b,99,ac,c8), r(30,94,a5,c6),\
+ r(59,df,d2,9c), r(52,d2,db,92), r(4f,c5,c0,80), r(44,c8,c9,8e),\
+ r(75,eb,f6,a4), r(7e,e6,ff,aa), r(63,f1,e4,b8), r(68,fc,ed,b6),\
+ r(b1,67,0a,0c), r(ba,6a,03,02), r(a7,7d,18,10), r(ac,70,11,1e),\
+ r(9d,53,2e,34), r(96,5e,27,3a), r(8b,49,3c,28), r(80,44,35,26),\
+ r(e9,0f,42,7c), r(e2,02,4b,72), r(ff,15,50,60), r(f4,18,59,6e),\
+ r(c5,3b,66,44), r(ce,36,6f,4a), r(d3,21,74,58), r(d8,2c,7d,56),\
+ r(7a,0c,a1,37), r(71,01,a8,39), r(6c,16,b3,2b), r(67,1b,ba,25),\
+ r(56,38,85,0f), r(5d,35,8c,01), r(40,22,97,13), r(4b,2f,9e,1d),\
+ r(22,64,e9,47), r(29,69,e0,49), r(34,7e,fb,5b), r(3f,73,f2,55),\
+ r(0e,50,cd,7f), r(05,5d,c4,71), r(18,4a,df,63), r(13,47,d6,6d),\
+ r(ca,dc,31,d7), r(c1,d1,38,d9), r(dc,c6,23,cb), r(d7,cb,2a,c5),\
+ r(e6,e8,15,ef), r(ed,e5,1c,e1), r(f0,f2,07,f3), r(fb,ff,0e,fd),\
+ r(92,b4,79,a7), r(99,b9,70,a9), r(84,ae,6b,bb), r(8f,a3,62,b5),\
+ r(be,80,5d,9f), r(b5,8d,54,91), r(a8,9a,4f,83), r(a3,97,46,8d)
+
+#undef r
+#define r r0
+
+#if defined(ONE_IM_TABLE)
+static const u_int32_t im_tab[256] =
+ { m_table };
+#elif defined(FOUR_IM_TABLES)
+static const u_int32_t im_tab[4][256] =
+{ { m_table },
+#undef r
+#define r r1
+ { m_table },
+#undef r
+#define r r2
+ { m_table },
+#undef r
+#define r r3
+ { m_table }
+};
+#endif
+
+#endif
+
+#else
+
+static int tab_gen = 0;
+
+static unsigned char s_box[256]; // the S box
+static unsigned char inv_s_box[256]; // the inverse S box
+static u_int32_t rcon_tab[AES_RC_LENGTH]; // table of round constants
+
+#if defined(ONE_TABLE)
+static u_int32_t ft_tab[256];
+static u_int32_t it_tab[256];
+#elif defined(FOUR_TABLES)
+static u_int32_t ft_tab[4][256];
+static u_int32_t it_tab[4][256];
+#endif
+
+#if defined(ONE_LR_TABLE)
+static u_int32_t fl_tab[256];
+static u_int32_t il_tab[256];
+#elif defined(FOUR_LR_TABLES)
+static u_int32_t fl_tab[4][256];
+static u_int32_t il_tab[4][256];
+#endif
+
+#if defined(ONE_IM_TABLE)
+static u_int32_t im_tab[256];
+#elif defined(FOUR_IM_TABLES)
+static u_int32_t im_tab[4][256];
+#endif
+
+// Generate the tables for the dynamic table option
+
+#if !defined(FF_TABLES)
+
+// It will generally be sensible to use tables to compute finite
+// field multiplies and inverses but where memory is scarse this
+// code might sometimes be better.
+
+// return 2 ^ (n - 1) where n is the bit number of the highest bit
+// set in x with x in the range 1 < x < 0x00000200. This form is
+// used so that locals within FFinv can be bytes rather than words
+
+static unsigned char hibit(const u_int32_t x)
+{ unsigned char r = (unsigned char)((x >> 1) | (x >> 2));
+
+ r |= (r >> 2);
+ r |= (r >> 4);
+ return (r + 1) >> 1;
+}
+
+// return the inverse of the finite field element x
+
+static unsigned char FFinv(const unsigned char x)
+{ unsigned char p1 = x, p2 = 0x1b, n1 = hibit(x), n2 = 0x80, v1 = 1, v2 = 0;
+
+ if(x < 2) return x;
+
+ for(;;)
+ {
+ if(!n1) return v1;
+
+ while(n2 >= n1)
+ {
+ n2 /= n1; p2 ^= p1 * n2; v2 ^= v1 * n2; n2 = hibit(p2);
+ }
+
+ if(!n2) return v2;
+
+ while(n1 >= n2)
+ {
+ n1 /= n2; p1 ^= p2 * n1; v1 ^= v2 * n1; n1 = hibit(p1);
+ }
+ }
+}
+
+// define the finite field multiplies required for Rijndael
+
+#define FFmul02(x) ((((x) & 0x7f) << 1) ^ ((x) & 0x80 ? 0x1b : 0))
+#define FFmul03(x) ((x) ^ FFmul02(x))
+#define FFmul09(x) ((x) ^ FFmul02(FFmul02(FFmul02(x))))
+#define FFmul0b(x) ((x) ^ FFmul02((x) ^ FFmul02(FFmul02(x))))
+#define FFmul0d(x) ((x) ^ FFmul02(FFmul02((x) ^ FFmul02(x))))
+#define FFmul0e(x) FFmul02((x) ^ FFmul02((x) ^ FFmul02(x)))
+
+#else
+
+#define FFinv(x) ((x) ? pow[255 - log[x]]: 0)
+
+#define FFmul02(x) (x ? pow[log[x] + 0x19] : 0)
+#define FFmul03(x) (x ? pow[log[x] + 0x01] : 0)
+#define FFmul09(x) (x ? pow[log[x] + 0xc7] : 0)
+#define FFmul0b(x) (x ? pow[log[x] + 0x68] : 0)
+#define FFmul0d(x) (x ? pow[log[x] + 0xee] : 0)
+#define FFmul0e(x) (x ? pow[log[x] + 0xdf] : 0)
+
+#endif
+
+// The forward and inverse affine transformations used in the S-box
+
+#define fwd_affine(x) \
+ (w = (u_int32_t)x, w ^= (w<<1)^(w<<2)^(w<<3)^(w<<4), 0x63^(unsigned char)(w^(w>>8)))
+
+#define inv_affine(x) \
+ (w = (u_int32_t)x, w = (w<<1)^(w<<3)^(w<<6), 0x05^(unsigned char)(w^(w>>8)))
+
+static void gen_tabs(void)
+{ u_int32_t i, w;
+
+#if defined(FF_TABLES)
+
+ unsigned char pow[512], log[256];
+
+ // log and power tables for GF(2^8) finite field with
+ // 0x011b as modular polynomial - the simplest primitive
+ // root is 0x03, used here to generate the tables
+
+ i = 0; w = 1;
+ do
+ {
+ pow[i] = (unsigned char)w;
+ pow[i + 255] = (unsigned char)w;
+ log[w] = (unsigned char)i++;
+ w ^= (w << 1) ^ (w & ff_hi ? ff_poly : 0);
+ }
+ while (w != 1);
+
+#endif
+
+ for(i = 0, w = 1; i < AES_RC_LENGTH; ++i)
+ {
+ rcon_tab[i] = bytes2word(w, 0, 0, 0);
+ w = (w << 1) ^ (w & ff_hi ? ff_poly : 0);
+ }
+
+ for(i = 0; i < 256; ++i)
+ { unsigned char b;
+
+ s_box[i] = b = fwd_affine(FFinv((unsigned char)i));
+
+ w = bytes2word(b, 0, 0, 0);
+#if defined(ONE_LR_TABLE)
+ fl_tab[i] = w;
+#elif defined(FOUR_LR_TABLES)
+ fl_tab[0][i] = w;
+ fl_tab[1][i] = upr(w,1);
+ fl_tab[2][i] = upr(w,2);
+ fl_tab[3][i] = upr(w,3);
+#endif
+ w = bytes2word(FFmul02(b), b, b, FFmul03(b));
+#if defined(ONE_TABLE)
+ ft_tab[i] = w;
+#elif defined(FOUR_TABLES)
+ ft_tab[0][i] = w;
+ ft_tab[1][i] = upr(w,1);
+ ft_tab[2][i] = upr(w,2);
+ ft_tab[3][i] = upr(w,3);
+#endif
+ inv_s_box[i] = b = FFinv(inv_affine((unsigned char)i));
+
+ w = bytes2word(b, 0, 0, 0);
+#if defined(ONE_LR_TABLE)
+ il_tab[i] = w;
+#elif defined(FOUR_LR_TABLES)
+ il_tab[0][i] = w;
+ il_tab[1][i] = upr(w,1);
+ il_tab[2][i] = upr(w,2);
+ il_tab[3][i] = upr(w,3);
+#endif
+ w = bytes2word(FFmul0e(b), FFmul09(b), FFmul0d(b), FFmul0b(b));
+#if defined(ONE_TABLE)
+ it_tab[i] = w;
+#elif defined(FOUR_TABLES)
+ it_tab[0][i] = w;
+ it_tab[1][i] = upr(w,1);
+ it_tab[2][i] = upr(w,2);
+ it_tab[3][i] = upr(w,3);
+#endif
+#if defined(ONE_IM_TABLE)
+ im_tab[b] = w;
+#elif defined(FOUR_IM_TABLES)
+ im_tab[0][b] = w;
+ im_tab[1][b] = upr(w,1);
+ im_tab[2][b] = upr(w,2);
+ im_tab[3][b] = upr(w,3);
+#endif
+
+ }
+}
+
+#endif
+
+#define no_table(x,box,vf,rf,c) bytes2word( \
+ box[bval(vf(x,0,c),rf(0,c))], \
+ box[bval(vf(x,1,c),rf(1,c))], \
+ box[bval(vf(x,2,c),rf(2,c))], \
+ box[bval(vf(x,3,c),rf(3,c))])
+
+#define one_table(x,op,tab,vf,rf,c) \
+ ( tab[bval(vf(x,0,c),rf(0,c))] \
+ ^ op(tab[bval(vf(x,1,c),rf(1,c))],1) \
+ ^ op(tab[bval(vf(x,2,c),rf(2,c))],2) \
+ ^ op(tab[bval(vf(x,3,c),rf(3,c))],3))
+
+#define four_tables(x,tab,vf,rf,c) \
+ ( tab[0][bval(vf(x,0,c),rf(0,c))] \
+ ^ tab[1][bval(vf(x,1,c),rf(1,c))] \
+ ^ tab[2][bval(vf(x,2,c),rf(2,c))] \
+ ^ tab[3][bval(vf(x,3,c),rf(3,c))])
+
+#define vf1(x,r,c) (x)
+#define rf1(r,c) (r)
+#define rf2(r,c) ((r-c)&3)
+
+#if defined(FOUR_LR_TABLES)
+#define ls_box(x,c) four_tables(x,fl_tab,vf1,rf2,c)
+#elif defined(ONE_LR_TABLE)
+#define ls_box(x,c) one_table(x,upr,fl_tab,vf1,rf2,c)
+#else
+#define ls_box(x,c) no_table(x,s_box,vf1,rf2,c)
+#endif
+
+#if defined(FOUR_IM_TABLES)
+#define inv_mcol(x) four_tables(x,im_tab,vf1,rf1,0)
+#elif defined(ONE_IM_TABLE)
+#define inv_mcol(x) one_table(x,upr,im_tab,vf1,rf1,0)
+#else
+#define inv_mcol(x) \
+ (f9 = (x),f2 = FFmulX(f9), f4 = FFmulX(f2), f8 = FFmulX(f4), f9 ^= f8, \
+ f2 ^= f4 ^ f8 ^ upr(f2 ^ f9,3) ^ upr(f4 ^ f9,2) ^ upr(f9,1))
+#endif
+
+// Subroutine to set the block size (if variable) in bytes, legal
+// values being 16, 24 and 32.
+
+#if defined(AES_BLOCK_SIZE)
+#define nc (AES_BLOCK_SIZE / 4)
+#else
+#define nc (cx->aes_Ncol)
+
+void aes_set_blk(aes_context *cx, int n_bytes)
+{
+#if !defined(FIXED_TABLES)
+ if(!tab_gen) { gen_tabs(); tab_gen = 1; }
+#endif
+
+ switch(n_bytes) {
+ case 32: /* bytes */
+ case 256: /* bits */
+ nc = 8;
+ break;
+ case 24: /* bytes */
+ case 192: /* bits */
+ nc = 6;
+ break;
+ case 16: /* bytes */
+ case 128: /* bits */
+ default:
+ nc = 4;
+ break;
+ }
+}
+
+#endif
+
+// Initialise the key schedule from the user supplied key. The key
+// length is now specified in bytes - 16, 24 or 32 as appropriate.
+// This corresponds to bit lengths of 128, 192 and 256 bits, and
+// to Nk values of 4, 6 and 8 respectively.
+
+#define mx(t,f) (*t++ = inv_mcol(*f),f++)
+#define cp(t,f) *t++ = *f++
+
+#if AES_BLOCK_SIZE == 16
+#define cpy(d,s) cp(d,s); cp(d,s); cp(d,s); cp(d,s)
+#define mix(d,s) mx(d,s); mx(d,s); mx(d,s); mx(d,s)
+#elif AES_BLOCK_SIZE == 24
+#define cpy(d,s) cp(d,s); cp(d,s); cp(d,s); cp(d,s); \
+ cp(d,s); cp(d,s)
+#define mix(d,s) mx(d,s); mx(d,s); mx(d,s); mx(d,s); \
+ mx(d,s); mx(d,s)
+#elif AES_BLOCK_SIZE == 32
+#define cpy(d,s) cp(d,s); cp(d,s); cp(d,s); cp(d,s); \
+ cp(d,s); cp(d,s); cp(d,s); cp(d,s)
+#define mix(d,s) mx(d,s); mx(d,s); mx(d,s); mx(d,s); \
+ mx(d,s); mx(d,s); mx(d,s); mx(d,s)
+#else
+
+#define cpy(d,s) \
+switch(nc) \
+{ case 8: cp(d,s); cp(d,s); \
+ case 6: cp(d,s); cp(d,s); \
+ case 4: cp(d,s); cp(d,s); \
+ cp(d,s); cp(d,s); \
+}
+
+#define mix(d,s) \
+switch(nc) \
+{ case 8: mx(d,s); mx(d,s); \
+ case 6: mx(d,s); mx(d,s); \
+ case 4: mx(d,s); mx(d,s); \
+ mx(d,s); mx(d,s); \
+}
+
+#endif
+
+void aes_set_key(aes_context *cx, const unsigned char in_key[], int n_bytes, const int f)
+{ u_int32_t *kf, *kt, rci;
+
+#if !defined(FIXED_TABLES)
+ if(!tab_gen) { gen_tabs(); tab_gen = 1; }
+#endif
+
+ switch(n_bytes) {
+ case 32: /* bytes */
+ case 256: /* bits */
+ cx->aes_Nkey = 8;
+ break;
+ case 24: /* bytes */
+ case 192: /* bits */
+ cx->aes_Nkey = 6;
+ break;
+ case 16: /* bytes */
+ case 128: /* bits */
+ default:
+ cx->aes_Nkey = 4;
+ break;
+ }
+
+ cx->aes_Nrnd = (cx->aes_Nkey > nc ? cx->aes_Nkey : nc) + 6;
+
+ cx->aes_e_key[0] = word_in(in_key );
+ cx->aes_e_key[1] = word_in(in_key + 4);
+ cx->aes_e_key[2] = word_in(in_key + 8);
+ cx->aes_e_key[3] = word_in(in_key + 12);
+
+ kf = cx->aes_e_key;
+ kt = kf + nc * (cx->aes_Nrnd + 1) - cx->aes_Nkey;
+ rci = 0;
+
+ switch(cx->aes_Nkey)
+ {
+ case 4: do
+ { kf[4] = kf[0] ^ ls_box(kf[3],3) ^ rcon_tab[rci++];
+ kf[5] = kf[1] ^ kf[4];
+ kf[6] = kf[2] ^ kf[5];
+ kf[7] = kf[3] ^ kf[6];
+ kf += 4;
+ }
+ while(kf < kt);
+ break;
+
+ case 6: cx->aes_e_key[4] = word_in(in_key + 16);
+ cx->aes_e_key[5] = word_in(in_key + 20);
+ do
+ { kf[ 6] = kf[0] ^ ls_box(kf[5],3) ^ rcon_tab[rci++];
+ kf[ 7] = kf[1] ^ kf[ 6];
+ kf[ 8] = kf[2] ^ kf[ 7];
+ kf[ 9] = kf[3] ^ kf[ 8];
+ kf[10] = kf[4] ^ kf[ 9];
+ kf[11] = kf[5] ^ kf[10];
+ kf += 6;
+ }
+ while(kf < kt);
+ break;
+
+ case 8: cx->aes_e_key[4] = word_in(in_key + 16);
+ cx->aes_e_key[5] = word_in(in_key + 20);
+ cx->aes_e_key[6] = word_in(in_key + 24);
+ cx->aes_e_key[7] = word_in(in_key + 28);
+ do
+ { kf[ 8] = kf[0] ^ ls_box(kf[7],3) ^ rcon_tab[rci++];
+ kf[ 9] = kf[1] ^ kf[ 8];
+ kf[10] = kf[2] ^ kf[ 9];
+ kf[11] = kf[3] ^ kf[10];
+ kf[12] = kf[4] ^ ls_box(kf[11],0);
+ kf[13] = kf[5] ^ kf[12];
+ kf[14] = kf[6] ^ kf[13];
+ kf[15] = kf[7] ^ kf[14];
+ kf += 8;
+ }
+ while (kf < kt);
+ break;
+ }
+
+ if(!f)
+ { u_int32_t i;
+
+ kt = cx->aes_d_key + nc * cx->aes_Nrnd;
+ kf = cx->aes_e_key;
+
+ cpy(kt, kf); kt -= 2 * nc;
+
+ for(i = 1; i < cx->aes_Nrnd; ++i)
+ {
+#if defined(ONE_TABLE) || defined(FOUR_TABLES)
+#if !defined(ONE_IM_TABLE) && !defined(FOUR_IM_TABLES)
+ u_int32_t f2, f4, f8, f9;
+#endif
+ mix(kt, kf);
+#else
+ cpy(kt, kf);
+#endif
+ kt -= 2 * nc;
+ }
+
+ cpy(kt, kf);
+ }
+}
+
+// y = output word, x = input word, r = row, c = column
+// for r = 0, 1, 2 and 3 = column accessed for row r
+
+#if defined(ARRAYS)
+#define s(x,c) x[c]
+#else
+#define s(x,c) x##c
+#endif
+
+// I am grateful to Frank Yellin for the following constructions
+// which, given the column (c) of the output state variable that
+// is being computed, return the input state variables which are
+// needed for each row (r) of the state
+
+// For the fixed block size options, compilers reduce these two
+// expressions to fixed variable references. For variable block
+// size code conditional clauses will sometimes be returned
+
+#define unused 77 // Sunset Strip
+
+#define fwd_var(x,r,c) \
+ ( r==0 ? \
+ ( c==0 ? s(x,0) \
+ : c==1 ? s(x,1) \
+ : c==2 ? s(x,2) \
+ : c==3 ? s(x,3) \
+ : c==4 ? s(x,4) \
+ : c==5 ? s(x,5) \
+ : c==6 ? s(x,6) \
+ : s(x,7)) \
+ : r==1 ? \
+ ( c==0 ? s(x,1) \
+ : c==1 ? s(x,2) \
+ : c==2 ? s(x,3) \
+ : c==3 ? nc==4 ? s(x,0) : s(x,4) \
+ : c==4 ? s(x,5) \
+ : c==5 ? nc==8 ? s(x,6) : s(x,0) \
+ : c==6 ? s(x,7) \
+ : s(x,0)) \
+ : r==2 ? \
+ ( c==0 ? nc==8 ? s(x,3) : s(x,2) \
+ : c==1 ? nc==8 ? s(x,4) : s(x,3) \
+ : c==2 ? nc==4 ? s(x,0) : nc==8 ? s(x,5) : s(x,4) \
+ : c==3 ? nc==4 ? s(x,1) : nc==8 ? s(x,6) : s(x,5) \
+ : c==4 ? nc==8 ? s(x,7) : s(x,0) \
+ : c==5 ? nc==8 ? s(x,0) : s(x,1) \
+ : c==6 ? s(x,1) \
+ : s(x,2)) \
+ : \
+ ( c==0 ? nc==8 ? s(x,4) : s(x,3) \
+ : c==1 ? nc==4 ? s(x,0) : nc==8 ? s(x,5) : s(x,4) \
+ : c==2 ? nc==4 ? s(x,1) : nc==8 ? s(x,6) : s(x,5) \
+ : c==3 ? nc==4 ? s(x,2) : nc==8 ? s(x,7) : s(x,0) \
+ : c==4 ? nc==8 ? s(x,0) : s(x,1) \
+ : c==5 ? nc==8 ? s(x,1) : s(x,2) \
+ : c==6 ? s(x,2) \
+ : s(x,3)))
+
+#define inv_var(x,r,c) \
+ ( r==0 ? \
+ ( c==0 ? s(x,0) \
+ : c==1 ? s(x,1) \
+ : c==2 ? s(x,2) \
+ : c==3 ? s(x,3) \
+ : c==4 ? s(x,4) \
+ : c==5 ? s(x,5) \
+ : c==6 ? s(x,6) \
+ : s(x,7)) \
+ : r==1 ? \
+ ( c==0 ? nc==4 ? s(x,3) : nc==8 ? s(x,7) : s(x,5) \
+ : c==1 ? s(x,0) \
+ : c==2 ? s(x,1) \
+ : c==3 ? s(x,2) \
+ : c==4 ? s(x,3) \
+ : c==5 ? s(x,4) \
+ : c==6 ? s(x,5) \
+ : s(x,6)) \
+ : r==2 ? \
+ ( c==0 ? nc==4 ? s(x,2) : nc==8 ? s(x,5) : s(x,4) \
+ : c==1 ? nc==4 ? s(x,3) : nc==8 ? s(x,6) : s(x,5) \
+ : c==2 ? nc==8 ? s(x,7) : s(x,0) \
+ : c==3 ? nc==8 ? s(x,0) : s(x,1) \
+ : c==4 ? nc==8 ? s(x,1) : s(x,2) \
+ : c==5 ? nc==8 ? s(x,2) : s(x,3) \
+ : c==6 ? s(x,3) \
+ : s(x,4)) \
+ : \
+ ( c==0 ? nc==4 ? s(x,1) : nc==8 ? s(x,4) : s(x,3) \
+ : c==1 ? nc==4 ? s(x,2) : nc==8 ? s(x,5) : s(x,4) \
+ : c==2 ? nc==4 ? s(x,3) : nc==8 ? s(x,6) : s(x,5) \
+ : c==3 ? nc==8 ? s(x,7) : s(x,0) \
+ : c==4 ? nc==8 ? s(x,0) : s(x,1) \
+ : c==5 ? nc==8 ? s(x,1) : s(x,2) \
+ : c==6 ? s(x,2) \
+ : s(x,3)))
+
+#define si(y,x,k,c) s(y,c) = word_in(x + 4 * c) ^ k[c]
+#define so(y,x,c) word_out(y + 4 * c, s(x,c))
+
+#if defined(FOUR_TABLES)
+#define fwd_rnd(y,x,k,c) s(y,c)= (k)[c] ^ four_tables(x,ft_tab,fwd_var,rf1,c)
+#define inv_rnd(y,x,k,c) s(y,c)= (k)[c] ^ four_tables(x,it_tab,inv_var,rf1,c)
+#elif defined(ONE_TABLE)
+#define fwd_rnd(y,x,k,c) s(y,c)= (k)[c] ^ one_table(x,upr,ft_tab,fwd_var,rf1,c)
+#define inv_rnd(y,x,k,c) s(y,c)= (k)[c] ^ one_table(x,upr,it_tab,inv_var,rf1,c)
+#else
+#define fwd_rnd(y,x,k,c) s(y,c) = fwd_mcol(no_table(x,s_box,fwd_var,rf1,c)) ^ (k)[c]
+#define inv_rnd(y,x,k,c) s(y,c) = inv_mcol(no_table(x,inv_s_box,inv_var,rf1,c) ^ (k)[c])
+#endif
+
+#if defined(FOUR_LR_TABLES)
+#define fwd_lrnd(y,x,k,c) s(y,c)= (k)[c] ^ four_tables(x,fl_tab,fwd_var,rf1,c)
+#define inv_lrnd(y,x,k,c) s(y,c)= (k)[c] ^ four_tables(x,il_tab,inv_var,rf1,c)
+#elif defined(ONE_LR_TABLE)
+#define fwd_lrnd(y,x,k,c) s(y,c)= (k)[c] ^ one_table(x,ups,fl_tab,fwd_var,rf1,c)
+#define inv_lrnd(y,x,k,c) s(y,c)= (k)[c] ^ one_table(x,ups,il_tab,inv_var,rf1,c)
+#else
+#define fwd_lrnd(y,x,k,c) s(y,c) = no_table(x,s_box,fwd_var,rf1,c) ^ (k)[c]
+#define inv_lrnd(y,x,k,c) s(y,c) = no_table(x,inv_s_box,inv_var,rf1,c) ^ (k)[c]
+#endif
+
+#if AES_BLOCK_SIZE == 16
+
+#if defined(ARRAYS)
+#define locals(y,x) x[4],y[4]
+#else
+#define locals(y,x) x##0,x##1,x##2,x##3,y##0,y##1,y##2,y##3
+// the following defines prevent the compiler requiring the declaration
+// of generated but unused variables in the fwd_var and inv_var macros
+#define b04 unused
+#define b05 unused
+#define b06 unused
+#define b07 unused
+#define b14 unused
+#define b15 unused
+#define b16 unused
+#define b17 unused
+#endif
+#define l_copy(y, x) s(y,0) = s(x,0); s(y,1) = s(x,1); \
+ s(y,2) = s(x,2); s(y,3) = s(x,3);
+#define state_in(y,x,k) si(y,x,k,0); si(y,x,k,1); si(y,x,k,2); si(y,x,k,3)
+#define state_out(y,x) so(y,x,0); so(y,x,1); so(y,x,2); so(y,x,3)
+#define round(rm,y,x,k) rm(y,x,k,0); rm(y,x,k,1); rm(y,x,k,2); rm(y,x,k,3)
+
+#elif AES_BLOCK_SIZE == 24
+
+#if defined(ARRAYS)
+#define locals(y,x) x[6],y[6]
+#else
+#define locals(y,x) x##0,x##1,x##2,x##3,x##4,x##5, \
+ y##0,y##1,y##2,y##3,y##4,y##5
+#define b06 unused
+#define b07 unused
+#define b16 unused
+#define b17 unused
+#endif
+#define l_copy(y, x) s(y,0) = s(x,0); s(y,1) = s(x,1); \
+ s(y,2) = s(x,2); s(y,3) = s(x,3); \
+ s(y,4) = s(x,4); s(y,5) = s(x,5);
+#define state_in(y,x,k) si(y,x,k,0); si(y,x,k,1); si(y,x,k,2); \
+ si(y,x,k,3); si(y,x,k,4); si(y,x,k,5)
+#define state_out(y,x) so(y,x,0); so(y,x,1); so(y,x,2); \
+ so(y,x,3); so(y,x,4); so(y,x,5)
+#define round(rm,y,x,k) rm(y,x,k,0); rm(y,x,k,1); rm(y,x,k,2); \
+ rm(y,x,k,3); rm(y,x,k,4); rm(y,x,k,5)
+#else
+
+#if defined(ARRAYS)
+#define locals(y,x) x[8],y[8]
+#else
+#define locals(y,x) x##0,x##1,x##2,x##3,x##4,x##5,x##6,x##7, \
+ y##0,y##1,y##2,y##3,y##4,y##5,y##6,y##7
+#endif
+#define l_copy(y, x) s(y,0) = s(x,0); s(y,1) = s(x,1); \
+ s(y,2) = s(x,2); s(y,3) = s(x,3); \
+ s(y,4) = s(x,4); s(y,5) = s(x,5); \
+ s(y,6) = s(x,6); s(y,7) = s(x,7);
+
+#if AES_BLOCK_SIZE == 32
+
+#define state_in(y,x,k) si(y,x,k,0); si(y,x,k,1); si(y,x,k,2); si(y,x,k,3); \
+ si(y,x,k,4); si(y,x,k,5); si(y,x,k,6); si(y,x,k,7)
+#define state_out(y,x) so(y,x,0); so(y,x,1); so(y,x,2); so(y,x,3); \
+ so(y,x,4); so(y,x,5); so(y,x,6); so(y,x,7)
+#define round(rm,y,x,k) rm(y,x,k,0); rm(y,x,k,1); rm(y,x,k,2); rm(y,x,k,3); \
+ rm(y,x,k,4); rm(y,x,k,5); rm(y,x,k,6); rm(y,x,k,7)
+#else
+
+#define state_in(y,x,k) \
+switch(nc) \
+{ case 8: si(y,x,k,7); si(y,x,k,6); \
+ case 6: si(y,x,k,5); si(y,x,k,4); \
+ case 4: si(y,x,k,3); si(y,x,k,2); \
+ si(y,x,k,1); si(y,x,k,0); \
+}
+
+#define state_out(y,x) \
+switch(nc) \
+{ case 8: so(y,x,7); so(y,x,6); \
+ case 6: so(y,x,5); so(y,x,4); \
+ case 4: so(y,x,3); so(y,x,2); \
+ so(y,x,1); so(y,x,0); \
+}
+
+#if defined(FAST_VARIABLE)
+
+#define round(rm,y,x,k) \
+switch(nc) \
+{ case 8: rm(y,x,k,7); rm(y,x,k,6); \
+ rm(y,x,k,5); rm(y,x,k,4); \
+ rm(y,x,k,3); rm(y,x,k,2); \
+ rm(y,x,k,1); rm(y,x,k,0); \
+ break; \
+ case 6: rm(y,x,k,5); rm(y,x,k,4); \
+ rm(y,x,k,3); rm(y,x,k,2); \
+ rm(y,x,k,1); rm(y,x,k,0); \
+ break; \
+ case 4: rm(y,x,k,3); rm(y,x,k,2); \
+ rm(y,x,k,1); rm(y,x,k,0); \
+ break; \
+}
+#else
+
+#define round(rm,y,x,k) \
+switch(nc) \
+{ case 8: rm(y,x,k,7); rm(y,x,k,6); \
+ case 6: rm(y,x,k,5); rm(y,x,k,4); \
+ case 4: rm(y,x,k,3); rm(y,x,k,2); \
+ rm(y,x,k,1); rm(y,x,k,0); \
+}
+
+#endif
+
+#endif
+#endif
+
+void aes_encrypt(const aes_context *cx, const unsigned char in_blk[], unsigned char out_blk[])
+{ u_int32_t locals(b0, b1);
+ const u_int32_t *kp = cx->aes_e_key;
+
+#if !defined(ONE_TABLE) && !defined(FOUR_TABLES)
+ u_int32_t f2;
+#endif
+
+ state_in(b0, in_blk, kp); kp += nc;
+
+#if defined(UNROLL)
+
+ switch(cx->aes_Nrnd)
+ {
+ case 14: round(fwd_rnd, b1, b0, kp );
+ round(fwd_rnd, b0, b1, kp + nc ); kp += 2 * nc;
+ case 12: round(fwd_rnd, b1, b0, kp );
+ round(fwd_rnd, b0, b1, kp + nc ); kp += 2 * nc;
+ case 10: round(fwd_rnd, b1, b0, kp );
+ round(fwd_rnd, b0, b1, kp + nc);
+ round(fwd_rnd, b1, b0, kp + 2 * nc);
+ round(fwd_rnd, b0, b1, kp + 3 * nc);
+ round(fwd_rnd, b1, b0, kp + 4 * nc);
+ round(fwd_rnd, b0, b1, kp + 5 * nc);
+ round(fwd_rnd, b1, b0, kp + 6 * nc);
+ round(fwd_rnd, b0, b1, kp + 7 * nc);
+ round(fwd_rnd, b1, b0, kp + 8 * nc);
+ round(fwd_lrnd, b0, b1, kp + 9 * nc);
+ }
+
+#elif defined(PARTIAL_UNROLL)
+ { u_int32_t rnd;
+
+ for(rnd = 0; rnd < (cx->aes_Nrnd >> 1) - 1; ++rnd)
+ {
+ round(fwd_rnd, b1, b0, kp);
+ round(fwd_rnd, b0, b1, kp + nc); kp += 2 * nc;
+ }
+
+ round(fwd_rnd, b1, b0, kp);
+ round(fwd_lrnd, b0, b1, kp + nc);
+ }
+#else
+ { u_int32_t rnd;
+
+ for(rnd = 0; rnd < cx->aes_Nrnd - 1; ++rnd)
+ {
+ round(fwd_rnd, b1, b0, kp);
+ l_copy(b0, b1); kp += nc;
+ }
+
+ round(fwd_lrnd, b0, b1, kp);
+ }
+#endif
+
+ state_out(out_blk, b0);
+}
+
+void aes_decrypt(const aes_context *cx, const unsigned char in_blk[], unsigned char out_blk[])
+{ u_int32_t locals(b0, b1);
+ const u_int32_t *kp = cx->aes_d_key;
+
+#if !defined(ONE_TABLE) && !defined(FOUR_TABLES)
+ u_int32_t f2, f4, f8, f9;
+#endif
+
+ state_in(b0, in_blk, kp); kp += nc;
+
+#if defined(UNROLL)
+
+ switch(cx->aes_Nrnd)
+ {
+ case 14: round(inv_rnd, b1, b0, kp );
+ round(inv_rnd, b0, b1, kp + nc ); kp += 2 * nc;
+ case 12: round(inv_rnd, b1, b0, kp );
+ round(inv_rnd, b0, b1, kp + nc ); kp += 2 * nc;
+ case 10: round(inv_rnd, b1, b0, kp );
+ round(inv_rnd, b0, b1, kp + nc);
+ round(inv_rnd, b1, b0, kp + 2 * nc);
+ round(inv_rnd, b0, b1, kp + 3 * nc);
+ round(inv_rnd, b1, b0, kp + 4 * nc);
+ round(inv_rnd, b0, b1, kp + 5 * nc);
+ round(inv_rnd, b1, b0, kp + 6 * nc);
+ round(inv_rnd, b0, b1, kp + 7 * nc);
+ round(inv_rnd, b1, b0, kp + 8 * nc);
+ round(inv_lrnd, b0, b1, kp + 9 * nc);
+ }
+
+#elif defined(PARTIAL_UNROLL)
+ { u_int32_t rnd;
+
+ for(rnd = 0; rnd < (cx->aes_Nrnd >> 1) - 1; ++rnd)
+ {
+ round(inv_rnd, b1, b0, kp);
+ round(inv_rnd, b0, b1, kp + nc); kp += 2 * nc;
+ }
+
+ round(inv_rnd, b1, b0, kp);
+ round(inv_lrnd, b0, b1, kp + nc);
+ }
+#else
+ { u_int32_t rnd;
+
+ for(rnd = 0; rnd < cx->aes_Nrnd - 1; ++rnd)
+ {
+ round(inv_rnd, b1, b0, kp);
+ l_copy(b0, b1); kp += nc;
+ }
+
+ round(inv_lrnd, b0, b1, kp);
+ }
+#endif
+
+ state_out(out_blk, b0);
+}
diff -urN linux-3.14-noloop/drivers/misc/aes.h linux-3.14-AES/drivers/misc/aes.h
--- linux-3.14-noloop/drivers/misc/aes.h 1970-01-01 02:00:00.000000000 +0200
+++ linux-3.14-AES/drivers/misc/aes.h 2014-03-31 22:20:36.000000000 +0300
@@ -0,0 +1,112 @@
+// I retain copyright in this code but I encourage its free use provided
+// that I don't carry any responsibility for the results. I am especially
+// happy to see it used in free and open source software. If you do use
+// it I would appreciate an acknowledgement of its origin in the code or
+// the product that results and I would also appreciate knowing a little
+// about the use to which it is being put. I am grateful to Frank Yellin
+// for some ideas that are used in this implementation.
+//
+// Dr B. R. Gladman <brg@gladman.uk.net> 6th April 2001.
+//
+// This is an implementation of the AES encryption algorithm (Rijndael)
+// designed by Joan Daemen and Vincent Rijmen. This version is designed
+// to provide both fixed and dynamic block and key lengths and can also
+// run with either big or little endian internal byte order (see aes.h).
+// It inputs block and key lengths in bytes with the legal values being
+// 16, 24 and 32.
+
+/*
+ * Modified by Jari Ruusu, May 1 2001
+ * - Fixed some compile warnings, code was ok but gcc warned anyway.
+ * - Changed basic types: byte -> unsigned char, word -> u_int32_t
+ * - Major name space cleanup: Names visible to outside now begin
+ * with "aes_" or "AES_". A lot of stuff moved from aes.h to aes.c
+ * - Removed C++ and DLL support as part of name space cleanup.
+ * - Eliminated unnecessary recomputation of tables. (actual bug fix)
+ * - Merged precomputed constant tables to aes.c file.
+ * - Removed data alignment restrictions for portability reasons.
+ * - Made block and key lengths accept bit count (128/192/256)
+ * as well byte count (16/24/32).
+ * - Removed all error checks. This change also eliminated the need
+ * to preinitialize the context struct to zero.
+ * - Removed some totally unused constants.
+ */
+
+#ifndef _AES_H
+#define _AES_H
+
+#include <linux/types.h>
+#include <linux/linkage.h>
+#include <linux/module.h>
+
+// CONFIGURATION OPTIONS (see also aes.c)
+//
+// Define AES_BLOCK_SIZE to set the cipher block size (16, 24 or 32) or
+// leave this undefined for dynamically variable block size (this will
+// result in much slower code).
+// IMPORTANT NOTE: AES_BLOCK_SIZE is in BYTES (16, 24, 32 or undefined). If
+// left undefined a slower version providing variable block length is compiled
+
+#define AES_BLOCK_SIZE 16
+
+// The number of key schedule words for different block and key lengths
+// allowing for method of computation which requires the length to be a
+// multiple of the key length
+//
+// Nk = 4 6 8
+// -------------
+// Nb = 4 | 60 60 64
+// 6 | 96 90 96
+// 8 | 120 120 120
+
+#if !defined(AES_BLOCK_SIZE) || (AES_BLOCK_SIZE == 32)
+#define AES_KS_LENGTH 120
+#define AES_RC_LENGTH 29
+#else
+#define AES_KS_LENGTH 4 * AES_BLOCK_SIZE
+#define AES_RC_LENGTH (9 * AES_BLOCK_SIZE) / 8 - 8
+#endif
+
+typedef struct
+{
+ u_int32_t aes_Nkey; // the number of words in the key input block
+ u_int32_t aes_Nrnd; // the number of cipher rounds
+ u_int32_t aes_e_key[AES_KS_LENGTH]; // the encryption key schedule
+ u_int32_t aes_d_key[AES_KS_LENGTH]; // the decryption key schedule
+#if !defined(AES_BLOCK_SIZE)
+ u_int32_t aes_Ncol; // the number of columns in the cipher state
+#endif
+} aes_context;
+
+// avoid global name conflict with mainline kernel
+#define aes_set_key _aes_set_key
+#define aes_encrypt _aes_encrypt
+#define aes_decrypt _aes_decrypt
+
+// THE CIPHER INTERFACE
+
+#if !defined(AES_BLOCK_SIZE)
+extern void aes_set_blk(aes_context *, const int);
+#endif
+
+#if defined(CONFIG_X86) || defined(CONFIG_X86_64)
+ asmlinkage
+#endif
+extern void aes_set_key(aes_context *, const unsigned char [], const int, const int);
+
+#if defined(CONFIG_X86) || defined(CONFIG_X86_64)
+ asmlinkage
+#endif
+extern void aes_encrypt(const aes_context *, const unsigned char [], unsigned char []);
+
+#if defined(CONFIG_X86) || defined(CONFIG_X86_64)
+ asmlinkage
+#endif
+extern void aes_decrypt(const aes_context *, const unsigned char [], unsigned char []);
+
+// The block length inputs to aes_set_block and aes_set_key are in numbers
+// of bytes or bits. The calls to subroutines must be made in the above
+// order but multiple calls can be made without repeating earlier calls
+// if their parameters have not changed.
+
+#endif // _AES_H
diff -urN linux-3.14-noloop/drivers/misc/crypto-ksym.c linux-3.14-AES/drivers/misc/crypto-ksym.c
--- linux-3.14-noloop/drivers/misc/crypto-ksym.c 1970-01-01 02:00:00.000000000 +0200
+++ linux-3.14-AES/drivers/misc/crypto-ksym.c 2014-03-31 22:20:36.000000000 +0300
@@ -0,0 +1,18 @@
+#include <linux/module.h>
+#include "aes.h"
+#include "md5.h"
+EXPORT_SYMBOL(aes_set_key);
+EXPORT_SYMBOL(aes_encrypt);
+EXPORT_SYMBOL(aes_decrypt);
+EXPORT_SYMBOL(md5_transform_CPUbyteorder);
+#if defined(CONFIG_X86_64)
+EXPORT_SYMBOL(md5_transform_CPUbyteorder_2x);
+#endif
+#if defined(CONFIG_BLK_DEV_LOOP_INTELAES) && (defined(CONFIG_X86) || defined(CONFIG_X86_64))
+asmlinkage extern void intel_aes_cbc_encrypt(const aes_context *, void *src, void *dst, size_t len, void *iv);
+asmlinkage extern void intel_aes_cbc_decrypt(const aes_context *, void *src, void *dst, size_t len, void *iv);
+asmlinkage extern void intel_aes_cbc_enc_4x512(aes_context **, void *src, void *dst, void *iv);
+EXPORT_SYMBOL(intel_aes_cbc_encrypt);
+EXPORT_SYMBOL(intel_aes_cbc_decrypt);
+EXPORT_SYMBOL(intel_aes_cbc_enc_4x512);
+#endif
diff -urN linux-3.14-noloop/drivers/misc/md5-2x-amd64.S linux-3.14-AES/drivers/misc/md5-2x-amd64.S
--- linux-3.14-noloop/drivers/misc/md5-2x-amd64.S 1970-01-01 02:00:00.000000000 +0200
+++ linux-3.14-AES/drivers/misc/md5-2x-amd64.S 2014-03-31 22:20:36.000000000 +0300
@@ -0,0 +1,223 @@
+//
+// md5-2x-amd64.S
+//
+// Written by Jari Ruusu, October 1 2003
+//
+// Copyright 2003 by Jari Ruusu.
+// Redistribution of this file is permitted under the GNU Public License.
+//
+
+// Modified by Jari Ruusu, June 12 2004
+// - Converted 32 bit x86 code to 64 bit AMD64 code
+
+// Modified by Jari Ruusu, April 11 2010
+// - Added another parallel MD5 transform computation
+
+// A MD5 transform implementation for AMD64 compatible processors.
+// This code does not preserve the rax, rcx, rdx, rsi, rdi or r8-r11
+// registers or the artihmetic status flags. However, the rbx, rbp and
+// r12-r15 registers are preserved across calls.
+
+// void md5_transform_CPUbyteorder_2x(u_int32_t *hashAB, u_int32_t *inA, u_int32_t *inB)
+
+#if defined(USE_UNDERLINE)
+# define md5_transform_CPUbyteorder_2x _md5_transform_CPUbyteorder_2x
+#endif
+#if !defined(ALIGN64BYTES)
+# define ALIGN64BYTES 64
+#endif
+
+ .file "md5-2x-amd64.S"
+ .globl md5_transform_CPUbyteorder_2x
+
+// rdi = pointer to u_int32_t hash[4 + 4] array which is read and written
+// hash[0...3] are for first MD5, hash[4...7] are for second MD5
+// rsi = pointer to u_int32_t in[16] array, first MD5, read only
+// rdx = pointer to u_int32_t in[16] array, second MD5, read only
+
+ .text
+ .align ALIGN64BYTES
+md5_transform_CPUbyteorder_2x:
+ push %rbx
+ push %rbp
+ push %r12
+ push %r13
+ push %r14
+ push %r15
+
+ movl 12(%rdi),%eax ; movl 12+16(%rdi),%ebx
+ movl 8(%rdi),%ecx ; movl 8+16(%rdi),%r13d
+ movl (%rdi),%r8d ; movl 16(%rdi),%r11d
+ movl 4(%rdi),%r9d ; movl 4+16(%rdi),%r12d
+ movl (%rsi),%r10d ; movl (%rdx),%ebp
+ prefetcht0 60(%rsi) ; prefetcht0 60(%rdx)
+ movl %eax,%r15d ; movl %ebx,%r14d
+ xorl %ecx,%eax ; xorl %r13d,%ebx
+
+#define REPEAT1(p1Aw,p1Bw,p2Ax,p2Bx,p3Az,p3Bz,p4c,p5s,p6Nin,p7ANz,p7BNz,p8ANy,p8BNy) \
+ addl $p4c,p1Aw ; addl $p4c,p1Bw ;\
+ andl p2Ax,%eax ; andl p2Bx,%ebx ;\
+ addl %r10d,p1Aw ; addl %ebp,p1Bw ;\
+ xorl p3Az,%eax ; xorl p3Bz,%ebx ;\
+ movl p6Nin*4(%rsi),%r10d ; movl p6Nin*4(%rdx),%ebp ;\
+ addl %eax,p1Aw ; addl %ebx,p1Bw ;\
+ movl p7ANz,%eax ; movl p7BNz,%ebx ;\
+ roll $p5s,p1Aw ; roll $p5s,p1Bw ;\
+ xorl p8ANy,%eax ; xorl p8BNy,%ebx ;\
+ addl p2Ax,p1Aw ; addl p2Bx,p1Bw
+
+ REPEAT1(%r8d,%r11d,%r9d,%r12d,%r15d,%r14d,0xd76aa478,7,1,%ecx,%r13d,%r9d,%r12d)
+ REPEAT1(%r15d,%r14d,%r8d,%r11d,%ecx,%r13d,0xe8c7b756,12,2,%r9d,%r12d,%r8d,%r11d)
+ REPEAT1(%ecx,%r13d,%r15d,%r14d,%r9d,%r12d,0x242070db,17,3,%r8d,%r11d,%r15d,%r14d)
+ REPEAT1(%r9d,%r12d,%ecx,%r13d,%r8d,%r11d,0xc1bdceee,22,4,%r15d,%r14d,%ecx,%r13d)
+ REPEAT1(%r8d,%r11d,%r9d,%r12d,%r15d,%r14d,0xf57c0faf,7,5,%ecx,%r13d,%r9d,%r12d)
+ REPEAT1(%r15d,%r14d,%r8d,%r11d,%ecx,%r13d,0x4787c62a,12,6,%r9d,%r12d,%r8d,%r11d)
+ REPEAT1(%ecx,%r13d,%r15d,%r14d,%r9d,%r12d,0xa8304613,17,7,%r8d,%r11d,%r15d,%r14d)
+ REPEAT1(%r9d,%r12d,%ecx,%r13d,%r8d,%r11d,0xfd469501,22,8,%r15d,%r14d,%ecx,%r13d)
+ REPEAT1(%r8d,%r11d,%r9d,%r12d,%r15d,%r14d,0x698098d8,7,9,%ecx,%r13d,%r9d,%r12d)
+ REPEAT1(%r15d,%r14d,%r8d,%r11d,%ecx,%r13d,0x8b44f7af,12,10,%r9d,%r12d,%r8d,%r11d)
+ REPEAT1(%ecx,%r13d,%r15d,%r14d,%r9d,%r12d,0xffff5bb1,17,11,%r8d,%r11d,%r15d,%r14d)
+ REPEAT1(%r9d,%r12d,%ecx,%r13d,%r8d,%r11d,0x895cd7be,22,12,%r15d,%r14d,%ecx,%r13d)
+ REPEAT1(%r8d,%r11d,%r9d,%r12d,%r15d,%r14d,0x6b901122,7,13,%ecx,%r13d,%r9d,%r12d)
+ REPEAT1(%r15d,%r14d,%r8d,%r11d,%ecx,%r13d,0xfd987193,12,14,%r9d,%r12d,%r8d,%r11d)
+ REPEAT1(%ecx,%r13d,%r15d,%r14d,%r9d,%r12d,0xa679438e,17,15,%r8d,%r11d,%r15d,%r14d)
+
+ addl $0x49b40821,%r9d ; addl $0x49b40821,%r12d
+ andl %ecx,%eax ; andl %r13d,%ebx
+ addl %r10d,%r9d ; addl %ebp,%r12d
+ xorl %r8d,%eax ; xorl %r11d,%ebx
+ movl 1*4(%rsi),%r10d ; movl 1*4(%rdx),%ebp
+ addl %eax,%r9d ; addl %ebx,%r12d
+ movl %ecx,%eax ; movl %r13d,%ebx
+ roll $22,%r9d ; roll $22,%r12d
+ addl %ecx,%r9d ; addl %r13d,%r12d
+
+#define REPEAT2(p1Aw,p1Bw,p2Ax,p2Bx,p3Ay,p3By,p4Az,p4Bz,p5c,p6s,p7Nin,p8ANy,p8BNy) \
+ xorl p2Ax,%eax ; xorl p2Bx,%ebx ;\
+ addl $p5c,p1Aw ; addl $p5c,p1Bw ;\
+ andl p4Az,%eax ; andl p4Bz,%ebx ;\
+ addl %r10d,p1Aw ; addl %ebp,p1Bw ;\
+ xorl p3Ay,%eax ; xorl p3By,%ebx ;\
+ movl p7Nin*4(%rsi),%r10d ; movl p7Nin*4(%rdx),%ebp ;\
+ addl %eax,p1Aw ; addl %ebx,p1Bw ;\
+ movl p8ANy,%eax ; movl p8BNy,%ebx ;\
+ roll $p6s,p1Aw ; roll $p6s,p1Bw ;\
+ addl p2Ax,p1Aw ; addl p2Bx,p1Bw
+
+ REPEAT2(%r8d,%r11d,%r9d,%r12d,%ecx,%r13d,%r15d,%r14d,0xf61e2562,5,6,%r9d,%r12d)
+ REPEAT2(%r15d,%r14d,%r8d,%r11d,%r9d,%r12d,%ecx,%r13d,0xc040b340,9,11,%r8d,%r11d)
+ REPEAT2(%ecx,%r13d,%r15d,%r14d,%r8d,%r11d,%r9d,%r12d,0x265e5a51,14,0,%r15d,%r14d)
+ REPEAT2(%r9d,%r12d,%ecx,%r13d,%r15d,%r14d,%r8d,%r11d,0xe9b6c7aa,20,5,%ecx,%r13d)
+ REPEAT2(%r8d,%r11d,%r9d,%r12d,%ecx,%r13d,%r15d,%r14d,0xd62f105d,5,10,%r9d,%r12d)
+ REPEAT2(%r15d,%r14d,%r8d,%r11d,%r9d,%r12d,%ecx,%r13d,0x02441453,9,15,%r8d,%r11d)
+ REPEAT2(%ecx,%r13d,%r15d,%r14d,%r8d,%r11d,%r9d,%r12d,0xd8a1e681,14,4,%r15d,%r14d)
+ REPEAT2(%r9d,%r12d,%ecx,%r13d,%r15d,%r14d,%r8d,%r11d,0xe7d3fbc8,20,9,%ecx,%r13d)
+ REPEAT2(%r8d,%r11d,%r9d,%r12d,%ecx,%r13d,%r15d,%r14d,0x21e1cde6,5,14,%r9d,%r12d)
+ REPEAT2(%r15d,%r14d,%r8d,%r11d,%r9d,%r12d,%ecx,%r13d,0xc33707d6,9,3,%r8d,%r11d)
+ REPEAT2(%ecx,%r13d,%r15d,%r14d,%r8d,%r11d,%r9d,%r12d,0xf4d50d87,14,8,%r15d,%r14d)
+ REPEAT2(%r9d,%r12d,%ecx,%r13d,%r15d,%r14d,%r8d,%r11d,0x455a14ed,20,13,%ecx,%r13d)
+ REPEAT2(%r8d,%r11d,%r9d,%r12d,%ecx,%r13d,%r15d,%r14d,0xa9e3e905,5,2,%r9d,%r12d)
+ REPEAT2(%r15d,%r14d,%r8d,%r11d,%r9d,%r12d,%ecx,%r13d,0xfcefa3f8,9,7,%r8d,%r11d)
+ REPEAT2(%ecx,%r13d,%r15d,%r14d,%r8d,%r11d,%r9d,%r12d,0x676f02d9,14,12,%r15d,%r14d)
+
+ xorl %ecx,%eax ; xorl %r13d,%ebx
+ addl $0x8d2a4c8a,%r9d ; addl $0x8d2a4c8a,%r12d
+ andl %r8d,%eax ; andl %r11d,%ebx
+ addl %r10d,%r9d ; addl %ebp,%r12d
+ xorl %r15d,%eax ; xorl %r14d,%ebx
+ movl 5*4(%rsi),%r10d ; movl 5*4(%rdx),%ebp
+ addl %eax,%r9d ; addl %ebx,%r12d
+ movl %ecx,%eax ; movl %r13d,%ebx
+ roll $20,%r9d ; roll $20,%r12d
+ xorl %r15d,%eax ; xorl %r14d,%ebx
+ addl %ecx,%r9d ; addl %r13d,%r12d
+
+#define REPEAT3(p1Aw,p1Bw,p2Ax,p2Bx,p3c,p4s,p5Nin,p6ANy,p6BNy,p7ANz,p7BNz) \
+ addl $p3c,p1Aw ; addl $p3c,p1Bw ;\
+ xorl p2Ax,%eax ; xorl p2Bx,%ebx ;\
+ addl %r10d,p1Aw ; addl %ebp,p1Bw ;\
+ movl p5Nin*4(%rsi),%r10d ; movl p5Nin*4(%rdx),%ebp ;\
+ addl %eax,p1Aw ; addl %ebx,p1Bw ;\
+ movl p6ANy,%eax ; movl p6BNy,%ebx ;\
+ roll $p4s,p1Aw ; roll $p4s,p1Bw ;\
+ xorl p7ANz,%eax ; xorl p7BNz,%ebx ;\
+ addl p2Ax,p1Aw ; addl p2Bx,p1Bw
+
+ REPEAT3(%r8d,%r11d,%r9d,%r12d,0xfffa3942,4,8,%r9d,%r12d,%ecx,%r13d)
+ REPEAT3(%r15d,%r14d,%r8d,%r11d,0x8771f681,11,11,%r8d,%r11d,%r9d,%r12d)
+ REPEAT3(%ecx,%r13d,%r15d,%r14d,0x6d9d6122,16,14,%r15d,%r14d,%r8d,%r11d)
+ REPEAT3(%r9d,%r12d,%ecx,%r13d,0xfde5380c,23,1,%ecx,%r13d,%r15d,%r14d)
+ REPEAT3(%r8d,%r11d,%r9d,%r12d,0xa4beea44,4,4,%r9d,%r12d,%ecx,%r13d)
+ REPEAT3(%r15d,%r14d,%r8d,%r11d,0x4bdecfa9,11,7,%r8d,%r11d,%r9d,%r12d)
+ REPEAT3(%ecx,%r13d,%r15d,%r14d,0xf6bb4b60,16,10,%r15d,%r14d,%r8d,%r11d)
+ REPEAT3(%r9d,%r12d,%ecx,%r13d,0xbebfbc70,23,13,%ecx,%r13d,%r15d,%r14d)
+ REPEAT3(%r8d,%r11d,%r9d,%r12d,0x289b7ec6,4,0,%r9d,%r12d,%ecx,%r13d)
+ REPEAT3(%r15d,%r14d,%r8d,%r11d,0xeaa127fa,11,3,%r8d,%r11d,%r9d,%r12d)
+ REPEAT3(%ecx,%r13d,%r15d,%r14d,0xd4ef3085,16,6,%r15d,%r14d,%r8d,%r11d)
+ REPEAT3(%r9d,%r12d,%ecx,%r13d,0x04881d05,23,9,%ecx,%r13d,%r15d,%r14d)
+ REPEAT3(%r8d,%r11d,%r9d,%r12d,0xd9d4d039,4,12,%r9d,%r12d,%ecx,%r13d)
+ REPEAT3(%r15d,%r14d,%r8d,%r11d,0xe6db99e5,11,15,%r8d,%r11d,%r9d,%r12d)
+ REPEAT3(%ecx,%r13d,%r15d,%r14d,0x1fa27cf8,16,2,%r15d,%r14d,%r8d,%r11d)
+
+ addl $0xc4ac5665,%r9d ; addl $0xc4ac5665,%r12d
+ xorl %ecx,%eax ; xorl %r13d,%ebx
+ addl %r10d,%r9d ; addl %ebp,%r12d
+ movl (%rsi),%r10d ; movl (%rdx),%ebp
+ addl %eax,%r9d ; addl %ebx,%r12d
+ movl %r15d,%eax ; movl %r14d,%ebx
+ roll $23,%r9d ; roll $23,%r12d
+ notl %eax ; notl %ebx
+ addl %ecx,%r9d ; addl %r13d,%r12d
+
+#define REPEAT4(p1Aw,p1Bw,p2Ax,p2Bx,p3Ay,p3By,p4c,p5s,p6Nin,p7ANz,p7BNz) \
+ addl $p4c,p1Aw ; addl $p4c,p1Bw ;\
+ orl p2Ax,%eax ; orl p2Bx,%ebx ;\
+ addl %r10d,p1Aw ; addl %ebp,p1Bw ;\
+ xorl p3Ay,%eax ; xorl p3By,%ebx ;\
+ movl p6Nin*4(%rsi),%r10d ; movl p6Nin*4(%rdx),%ebp ;\
+ addl %eax,p1Aw ; addl %ebx,p1Bw ;\
+ movl p7ANz,%eax ; movl p7BNz,%ebx ;\
+ roll $p5s,p1Aw ; roll $p5s,p1Bw ;\
+ notl %eax ; notl %ebx ;\
+ addl p2Ax,p1Aw ; addl p2Bx,p1Bw
+
+ REPEAT4(%r8d,%r11d,%r9d,%r12d,%ecx,%r13d,0xf4292244,6,7,%ecx,%r13d)
+ REPEAT4(%r15d,%r14d,%r8d,%r11d,%r9d,%r12d,0x432aff97,10,14,%r9d,%r12d)
+ REPEAT4(%ecx,%r13d,%r15d,%r14d,%r8d,%r11d,0xab9423a7,15,5,%r8d,%r11d)
+ REPEAT4(%r9d,%r12d,%ecx,%r13d,%r15d,%r14d,0xfc93a039,21,12,%r15d,%r14d)
+ REPEAT4(%r8d,%r11d,%r9d,%r12d,%ecx,%r13d,0x655b59c3,6,3,%ecx,%r13d)
+ REPEAT4(%r15d,%r14d,%r8d,%r11d,%r9d,%r12d,0x8f0ccc92,10,10,%r9d,%r12d)
+ REPEAT4(%ecx,%r13d,%r15d,%r14d,%r8d,%r11d,0xffeff47d,15,1,%r8d,%r11d)
+ REPEAT4(%r9d,%r12d,%ecx,%r13d,%r15d,%r14d,0x85845dd1,21,8,%r15d,%r14d)
+ REPEAT4(%r8d,%r11d,%r9d,%r12d,%ecx,%r13d,0x6fa87e4f,6,15,%ecx,%r13d)
+ REPEAT4(%r15d,%r14d,%r8d,%r11d,%r9d,%r12d,0xfe2ce6e0,10,6,%r9d,%r12d)
+ REPEAT4(%ecx,%r13d,%r15d,%r14d,%r8d,%r11d,0xa3014314,15,13,%r8d,%r11d)
+ REPEAT4(%r9d,%r12d,%ecx,%r13d,%r15d,%r14d,0x4e0811a1,21,4,%r15d,%r14d)
+ REPEAT4(%r8d,%r11d,%r9d,%r12d,%ecx,%r13d,0xf7537e82,6,11,%ecx,%r13d)
+ REPEAT4(%r15d,%r14d,%r8d,%r11d,%r9d,%r12d,0xbd3af235,10,2,%r9d,%r12d)
+ REPEAT4(%ecx,%r13d,%r15d,%r14d,%r8d,%r11d,0x2ad7d2bb,15,9,%r8d,%r11d)
+
+ addl $0xeb86d391,%r9d ; addl $0xeb86d391,%r12d
+ orl %ecx,%eax ; orl %r13d,%ebx
+ addl %r10d,%r9d ; addl %ebp,%r12d
+ xorl %r15d,%eax ; xorl %r14d,%ebx
+ addl %eax,%r9d ; addl %ebx,%r12d
+ roll $21,%r9d ; roll $21,%r12d
+ addl %ecx,%r9d ; addl %r13d,%r12d
+
+ addl %r8d,(%rdi) ; addl %r11d,16(%rdi)
+ addl %r9d,4(%rdi) ; addl %r12d,4+16(%rdi)
+ addl %ecx,8(%rdi) ; addl %r13d,8+16(%rdi)
+ addl %r15d,12(%rdi) ; addl %r14d,12+16(%rdi)
+
+ pop %r15
+ pop %r14
+ pop %r13
+ pop %r12
+ pop %rbp
+ pop %rbx
+ ret
+
+#if defined(__ELF__) && defined(SECTION_NOTE_GNU_STACK)
+ .section .note.GNU-stack,"",@progbits
+#endif
diff -urN linux-3.14-noloop/drivers/misc/md5-amd64.S linux-3.14-AES/drivers/misc/md5-amd64.S
--- linux-3.14-noloop/drivers/misc/md5-amd64.S 1970-01-01 02:00:00.000000000 +0200
+++ linux-3.14-AES/drivers/misc/md5-amd64.S 2014-03-31 22:20:36.000000000 +0300
@@ -0,0 +1,204 @@
+//
+// md5-amd64.S
+//
+// Written by Jari Ruusu, October 1 2003
+//
+// Copyright 2003 by Jari Ruusu.
+// Redistribution of this file is permitted under the GNU Public License.
+//
+
+// Modified by Jari Ruusu, June 12 2004
+// - Converted 32 bit x86 code to 64 bit AMD64 code
+
+// A MD5 transform implementation for AMD64 compatible processors.
+// This code does not preserve the rax, rcx, rdx, rsi, rdi or r8-r11
+// registers or the artihmetic status flags. However, the rbx, rbp and
+// r12-r15 registers are preserved across calls.
+
+// void md5_transform_CPUbyteorder(u_int32_t *hash, u_int32_t *in)
+
+#if defined(USE_UNDERLINE)
+# define md5_transform_CPUbyteorder _md5_transform_CPUbyteorder
+#endif
+#if !defined(ALIGN64BYTES)
+# define ALIGN64BYTES 64
+#endif
+
+ .file "md5-amd64.S"
+ .globl md5_transform_CPUbyteorder
+
+// rdi = pointer to hash[4] array which is read and written
+// rsi = pointer to in[16] array which is read only
+
+ .text
+ .align ALIGN64BYTES
+md5_transform_CPUbyteorder:
+ movl 12(%rdi),%eax
+ movl 8(%rdi),%ecx
+ movl (%rdi),%r8d
+ movl 4(%rdi),%r9d
+ movl (%rsi),%r10d
+ prefetcht0 60(%rsi)
+ movl %eax,%edx
+ xorl %ecx,%eax
+
+#define REPEAT1(p1w,p2x,p3z,p4c,p5s,p6Nin,p7Nz,p8Ny) \
+ addl $p4c,p1w ;\
+ andl p2x,%eax ;\
+ addl %r10d,p1w ;\
+ xorl p3z,%eax ;\
+ movl p6Nin*4(%rsi),%r10d ;\
+ addl %eax,p1w ;\
+ movl p7Nz,%eax ;\
+ roll $p5s,p1w ;\
+ xorl p8Ny,%eax ;\
+ addl p2x,p1w
+
+ REPEAT1(%r8d,%r9d,%edx,0xd76aa478, 7, 1,%ecx,%r9d)
+ REPEAT1(%edx,%r8d,%ecx,0xe8c7b756,12, 2,%r9d,%r8d)
+ REPEAT1(%ecx,%edx,%r9d,0x242070db,17, 3,%r8d,%edx)
+ REPEAT1(%r9d,%ecx,%r8d,0xc1bdceee,22, 4,%edx,%ecx)
+ REPEAT1(%r8d,%r9d,%edx,0xf57c0faf, 7, 5,%ecx,%r9d)
+ REPEAT1(%edx,%r8d,%ecx,0x4787c62a,12, 6,%r9d,%r8d)
+ REPEAT1(%ecx,%edx,%r9d,0xa8304613,17, 7,%r8d,%edx)
+ REPEAT1(%r9d,%ecx,%r8d,0xfd469501,22, 8,%edx,%ecx)
+ REPEAT1(%r8d,%r9d,%edx,0x698098d8, 7, 9,%ecx,%r9d)
+ REPEAT1(%edx,%r8d,%ecx,0x8b44f7af,12,10,%r9d,%r8d)
+ REPEAT1(%ecx,%edx,%r9d,0xffff5bb1,17,11,%r8d,%edx)
+ REPEAT1(%r9d,%ecx,%r8d,0x895cd7be,22,12,%edx,%ecx)
+ REPEAT1(%r8d,%r9d,%edx,0x6b901122, 7,13,%ecx,%r9d)
+ REPEAT1(%edx,%r8d,%ecx,0xfd987193,12,14,%r9d,%r8d)
+ REPEAT1(%ecx,%edx,%r9d,0xa679438e,17,15,%r8d,%edx)
+
+ addl $0x49b40821,%r9d
+ andl %ecx,%eax
+ addl %r10d,%r9d
+ xorl %r8d,%eax
+ movl 1*4(%rsi),%r10d
+ addl %eax,%r9d
+ movl %ecx,%eax
+ roll $22,%r9d
+ addl %ecx,%r9d
+
+#define REPEAT2(p1w,p2x,p3y,p4z,p5c,p6s,p7Nin,p8Ny) \
+ xorl p2x,%eax ;\
+ addl $p5c,p1w ;\
+ andl p4z,%eax ;\
+ addl %r10d,p1w ;\
+ xorl p3y,%eax ;\
+ movl p7Nin*4(%rsi),%r10d ;\
+ addl %eax,p1w ;\
+ movl p8Ny,%eax ;\
+ roll $p6s,p1w ;\
+ addl p2x,p1w
+
+ REPEAT2(%r8d,%r9d,%ecx,%edx,0xf61e2562, 5, 6,%r9d)
+ REPEAT2(%edx,%r8d,%r9d,%ecx,0xc040b340, 9,11,%r8d)
+ REPEAT2(%ecx,%edx,%r8d,%r9d,0x265e5a51,14, 0,%edx)
+ REPEAT2(%r9d,%ecx,%edx,%r8d,0xe9b6c7aa,20, 5,%ecx)
+ REPEAT2(%r8d,%r9d,%ecx,%edx,0xd62f105d, 5,10,%r9d)
+ REPEAT2(%edx,%r8d,%r9d,%ecx,0x02441453, 9,15,%r8d)
+ REPEAT2(%ecx,%edx,%r8d,%r9d,0xd8a1e681,14, 4,%edx)
+ REPEAT2(%r9d,%ecx,%edx,%r8d,0xe7d3fbc8,20, 9,%ecx)
+ REPEAT2(%r8d,%r9d,%ecx,%edx,0x21e1cde6, 5,14,%r9d)
+ REPEAT2(%edx,%r8d,%r9d,%ecx,0xc33707d6, 9, 3,%r8d)
+ REPEAT2(%ecx,%edx,%r8d,%r9d,0xf4d50d87,14, 8,%edx)
+ REPEAT2(%r9d,%ecx,%edx,%r8d,0x455a14ed,20,13,%ecx)
+ REPEAT2(%r8d,%r9d,%ecx,%edx,0xa9e3e905, 5, 2,%r9d)
+ REPEAT2(%edx,%r8d,%r9d,%ecx,0xfcefa3f8, 9, 7,%r8d)
+ REPEAT2(%ecx,%edx,%r8d,%r9d,0x676f02d9,14,12,%edx)
+
+ xorl %ecx,%eax
+ addl $0x8d2a4c8a,%r9d
+ andl %r8d,%eax
+ addl %r10d,%r9d
+ xorl %edx,%eax
+ movl 5*4(%rsi),%r10d
+ addl %eax,%r9d
+ movl %ecx,%eax
+ roll $20,%r9d
+ xorl %edx,%eax
+ addl %ecx,%r9d
+
+#define REPEAT3(p1w,p2x,p3c,p4s,p5Nin,p6Ny,p7Nz) \
+ addl $p3c,p1w ;\
+ xorl p2x,%eax ;\
+ addl %r10d,p1w ;\
+ movl p5Nin*4(%rsi),%r10d ;\
+ addl %eax,p1w ;\
+ movl p6Ny,%eax ;\
+ roll $p4s,p1w ;\
+ xorl p7Nz,%eax ;\
+ addl p2x,p1w
+
+ REPEAT3(%r8d,%r9d,0xfffa3942, 4, 8,%r9d,%ecx)
+ REPEAT3(%edx,%r8d,0x8771f681,11,11,%r8d,%r9d)
+ REPEAT3(%ecx,%edx,0x6d9d6122,16,14,%edx,%r8d)
+ REPEAT3(%r9d,%ecx,0xfde5380c,23, 1,%ecx,%edx)
+ REPEAT3(%r8d,%r9d,0xa4beea44, 4, 4,%r9d,%ecx)
+ REPEAT3(%edx,%r8d,0x4bdecfa9,11, 7,%r8d,%r9d)
+ REPEAT3(%ecx,%edx,0xf6bb4b60,16,10,%edx,%r8d)
+ REPEAT3(%r9d,%ecx,0xbebfbc70,23,13,%ecx,%edx)
+ REPEAT3(%r8d,%r9d,0x289b7ec6, 4, 0,%r9d,%ecx)
+ REPEAT3(%edx,%r8d,0xeaa127fa,11, 3,%r8d,%r9d)
+ REPEAT3(%ecx,%edx,0xd4ef3085,16, 6,%edx,%r8d)
+ REPEAT3(%r9d,%ecx,0x04881d05,23, 9,%ecx,%edx)
+ REPEAT3(%r8d,%r9d,0xd9d4d039, 4,12,%r9d,%ecx)
+ REPEAT3(%edx,%r8d,0xe6db99e5,11,15,%r8d,%r9d)
+ REPEAT3(%ecx,%edx,0x1fa27cf8,16, 2,%edx,%r8d)
+
+ addl $0xc4ac5665,%r9d
+ xorl %ecx,%eax
+ addl %r10d,%r9d
+ movl (%rsi),%r10d
+ addl %eax,%r9d
+ movl %edx,%eax
+ roll $23,%r9d
+ notl %eax
+ addl %ecx,%r9d
+
+#define REPEAT4(p1w,p2x,p3y,p4c,p5s,p6Nin,p7Nz) \
+ addl $p4c,p1w ;\
+ orl p2x,%eax ;\
+ addl %r10d,p1w ;\
+ xorl p3y,%eax ;\
+ movl p6Nin*4(%rsi),%r10d ;\
+ addl %eax,p1w ;\
+ movl p7Nz,%eax ;\
+ roll $p5s,p1w ;\
+ notl %eax ;\
+ addl p2x,p1w
+
+ REPEAT4(%r8d,%r9d,%ecx,0xf4292244, 6, 7,%ecx)
+ REPEAT4(%edx,%r8d,%r9d,0x432aff97,10,14,%r9d)
+ REPEAT4(%ecx,%edx,%r8d,0xab9423a7,15, 5,%r8d)
+ REPEAT4(%r9d,%ecx,%edx,0xfc93a039,21,12,%edx)
+ REPEAT4(%r8d,%r9d,%ecx,0x655b59c3, 6, 3,%ecx)
+ REPEAT4(%edx,%r8d,%r9d,0x8f0ccc92,10,10,%r9d)
+ REPEAT4(%ecx,%edx,%r8d,0xffeff47d,15, 1,%r8d)
+ REPEAT4(%r9d,%ecx,%edx,0x85845dd1,21, 8,%edx)
+ REPEAT4(%r8d,%r9d,%ecx,0x6fa87e4f, 6,15,%ecx)
+ REPEAT4(%edx,%r8d,%r9d,0xfe2ce6e0,10, 6,%r9d)
+ REPEAT4(%ecx,%edx,%r8d,0xa3014314,15,13,%r8d)
+ REPEAT4(%r9d,%ecx,%edx,0x4e0811a1,21, 4,%edx)
+ REPEAT4(%r8d,%r9d,%ecx,0xf7537e82, 6,11,%ecx)
+ REPEAT4(%edx,%r8d,%r9d,0xbd3af235,10, 2,%r9d)
+ REPEAT4(%ecx,%edx,%r8d,0x2ad7d2bb,15, 9,%r8d)
+
+ addl $0xeb86d391,%r9d
+ orl %ecx,%eax
+ addl %r10d,%r9d
+ xorl %edx,%eax
+ addl %eax,%r9d
+ roll $21,%r9d
+ addl %ecx,%r9d
+
+ addl %r8d,(%rdi)
+ addl %r9d,4(%rdi)
+ addl %ecx,8(%rdi)
+ addl %edx,12(%rdi)
+ ret
+
+#if defined(__ELF__) && defined(SECTION_NOTE_GNU_STACK)
+ .section .note.GNU-stack,"",@progbits
+#endif
diff -urN linux-3.14-noloop/drivers/misc/md5-x86.S linux-3.14-AES/drivers/misc/md5-x86.S
--- linux-3.14-noloop/drivers/misc/md5-x86.S 1970-01-01 02:00:00.000000000 +0200
+++ linux-3.14-AES/drivers/misc/md5-x86.S 2014-03-31 22:20:36.000000000 +0300
@@ -0,0 +1,211 @@
+//
+// md5-x86.S
+//
+// Written by Jari Ruusu, October 1 2003
+//
+// Copyright 2003 by Jari Ruusu.
+// Redistribution of this file is permitted under the GNU Public License.
+//
+
+// A MD5 transform implementation for x86 compatible processors. This
+// version uses i386 instruction set but instruction scheduling is optimized
+// for Pentium-2. This code does not preserve the eax, ecx or edx registers
+// or the artihmetic status flags. However, the ebx, esi, edi, and ebp
+// registers are preserved across calls.
+
+// void md5_transform_CPUbyteorder(u_int32_t *hash, u_int32_t *in)
+
+#if defined(USE_UNDERLINE)
+# define md5_transform_CPUbyteorder _md5_transform_CPUbyteorder
+#endif
+#if !defined(ALIGN32BYTES)
+# define ALIGN32BYTES 32
+#endif
+
+ .file "md5-x86.S"
+ .globl md5_transform_CPUbyteorder
+ .text
+ .align ALIGN32BYTES
+
+md5_transform_CPUbyteorder:
+ push %ebp
+ mov 4+4(%esp),%eax // pointer to 'hash' input
+ mov 8+4(%esp),%ebp // pointer to 'in' array
+ push %ebx
+ push %esi
+ push %edi
+
+ mov (%eax),%esi
+ mov 4(%eax),%edi
+ mov 8(%eax),%ecx
+ mov 12(%eax),%eax
+ mov (%ebp),%ebx
+ mov %eax,%edx
+ xor %ecx,%eax
+
+#define REPEAT1(p1w,p2x,p3z,p4c,p5s,p6Nin,p7Nz,p8Ny) \
+ add $p4c,p1w ;\
+ and p2x,%eax ;\
+ add %ebx,p1w ;\
+ xor p3z,%eax ;\
+ mov p6Nin*4(%ebp),%ebx ;\
+ add %eax,p1w ;\
+ mov p7Nz,%eax ;\
+ rol $p5s,p1w ;\
+ xor p8Ny,%eax ;\
+ add p2x,p1w
+
+ REPEAT1(%esi,%edi,%edx,0xd76aa478, 7, 1,%ecx,%edi)
+ REPEAT1(%edx,%esi,%ecx,0xe8c7b756,12, 2,%edi,%esi)
+ REPEAT1(%ecx,%edx,%edi,0x242070db,17, 3,%esi,%edx)
+ REPEAT1(%edi,%ecx,%esi,0xc1bdceee,22, 4,%edx,%ecx)
+ REPEAT1(%esi,%edi,%edx,0xf57c0faf, 7, 5,%ecx,%edi)
+ REPEAT1(%edx,%esi,%ecx,0x4787c62a,12, 6,%edi,%esi)
+ REPEAT1(%ecx,%edx,%edi,0xa8304613,17, 7,%esi,%edx)
+ REPEAT1(%edi,%ecx,%esi,0xfd469501,22, 8,%edx,%ecx)
+ REPEAT1(%esi,%edi,%edx,0x698098d8, 7, 9,%ecx,%edi)
+ REPEAT1(%edx,%esi,%ecx,0x8b44f7af,12,10,%edi,%esi)
+ REPEAT1(%ecx,%edx,%edi,0xffff5bb1,17,11,%esi,%edx)
+ REPEAT1(%edi,%ecx,%esi,0x895cd7be,22,12,%edx,%ecx)
+ REPEAT1(%esi,%edi,%edx,0x6b901122, 7,13,%ecx,%edi)
+ REPEAT1(%edx,%esi,%ecx,0xfd987193,12,14,%edi,%esi)
+ REPEAT1(%ecx,%edx,%edi,0xa679438e,17,15,%esi,%edx)
+
+ add $0x49b40821,%edi
+ and %ecx,%eax
+ add %ebx,%edi
+ xor %esi,%eax
+ mov 1*4(%ebp),%ebx
+ add %eax,%edi
+ mov %ecx,%eax
+ rol $22,%edi
+ add %ecx,%edi
+
+#define REPEAT2(p1w,p2x,p3y,p4z,p5c,p6s,p7Nin,p8Ny) \
+ xor p2x,%eax ;\
+ add $p5c,p1w ;\
+ and p4z,%eax ;\
+ add %ebx,p1w ;\
+ xor p3y,%eax ;\
+ mov p7Nin*4(%ebp),%ebx ;\
+ add %eax,p1w ;\
+ mov p8Ny,%eax ;\
+ rol $p6s,p1w ;\
+ add p2x,p1w
+
+ REPEAT2(%esi,%edi,%ecx,%edx,0xf61e2562, 5, 6,%edi)
+ REPEAT2(%edx,%esi,%edi,%ecx,0xc040b340, 9,11,%esi)
+ REPEAT2(%ecx,%edx,%esi,%edi,0x265e5a51,14, 0,%edx)
+ REPEAT2(%edi,%ecx,%edx,%esi,0xe9b6c7aa,20, 5,%ecx)
+ REPEAT2(%esi,%edi,%ecx,%edx,0xd62f105d, 5,10,%edi)
+ REPEAT2(%edx,%esi,%edi,%ecx,0x02441453, 9,15,%esi)
+ REPEAT2(%ecx,%edx,%esi,%edi,0xd8a1e681,14, 4,%edx)
+ REPEAT2(%edi,%ecx,%edx,%esi,0xe7d3fbc8,20, 9,%ecx)
+ REPEAT2(%esi,%edi,%ecx,%edx,0x21e1cde6, 5,14,%edi)
+ REPEAT2(%edx,%esi,%edi,%ecx,0xc33707d6, 9, 3,%esi)
+ REPEAT2(%ecx,%edx,%esi,%edi,0xf4d50d87,14, 8,%edx)
+ REPEAT2(%edi,%ecx,%edx,%esi,0x455a14ed,20,13,%ecx)
+ REPEAT2(%esi,%edi,%ecx,%edx,0xa9e3e905, 5, 2,%edi)
+ REPEAT2(%edx,%esi,%edi,%ecx,0xfcefa3f8, 9, 7,%esi)
+ REPEAT2(%ecx,%edx,%esi,%edi,0x676f02d9,14,12,%edx)
+
+ xor %ecx,%eax
+ add $0x8d2a4c8a,%edi
+ and %esi,%eax
+ add %ebx,%edi
+ xor %edx,%eax
+ mov 5*4(%ebp),%ebx
+ add %eax,%edi
+ mov %ecx,%eax
+ rol $20,%edi
+ xor %edx,%eax
+ add %ecx,%edi
+
+#define REPEAT3(p1w,p2x,p3c,p4s,p5Nin,p6Ny,p7Nz) \
+ add $p3c,p1w ;\
+ xor p2x,%eax ;\
+ add %ebx,p1w ;\
+ mov p5Nin*4(%ebp),%ebx ;\
+ add %eax,p1w ;\
+ mov p6Ny,%eax ;\
+ rol $p4s,p1w ;\
+ xor p7Nz,%eax ;\
+ add p2x,p1w
+
+ REPEAT3(%esi,%edi,0xfffa3942, 4, 8,%edi,%ecx)
+ REPEAT3(%edx,%esi,0x8771f681,11,11,%esi,%edi)
+ REPEAT3(%ecx,%edx,0x6d9d6122,16,14,%edx,%esi)
+ REPEAT3(%edi,%ecx,0xfde5380c,23, 1,%ecx,%edx)
+ REPEAT3(%esi,%edi,0xa4beea44, 4, 4,%edi,%ecx)
+ REPEAT3(%edx,%esi,0x4bdecfa9,11, 7,%esi,%edi)
+ REPEAT3(%ecx,%edx,0xf6bb4b60,16,10,%edx,%esi)
+ REPEAT3(%edi,%ecx,0xbebfbc70,23,13,%ecx,%edx)
+ REPEAT3(%esi,%edi,0x289b7ec6, 4, 0,%edi,%ecx)
+ REPEAT3(%edx,%esi,0xeaa127fa,11, 3,%esi,%edi)
+ REPEAT3(%ecx,%edx,0xd4ef3085,16, 6,%edx,%esi)
+ REPEAT3(%edi,%ecx,0x04881d05,23, 9,%ecx,%edx)
+ REPEAT3(%esi,%edi,0xd9d4d039, 4,12,%edi,%ecx)
+ REPEAT3(%edx,%esi,0xe6db99e5,11,15,%esi,%edi)
+ REPEAT3(%ecx,%edx,0x1fa27cf8,16, 2,%edx,%esi)
+
+ add $0xc4ac5665,%edi
+ xor %ecx,%eax
+ add %ebx,%edi
+ mov (%ebp),%ebx
+ add %eax,%edi
+ mov %edx,%eax
+ rol $23,%edi
+ not %eax
+ add %ecx,%edi
+
+#define REPEAT4(p1w,p2x,p3y,p4c,p5s,p6Nin,p7Nz) \
+ add $p4c,p1w ;\
+ or p2x,%eax ;\
+ add %ebx,p1w ;\
+ xor p3y,%eax ;\
+ mov p6Nin*4(%ebp),%ebx ;\
+ add %eax,p1w ;\
+ mov p7Nz,%eax ;\
+ rol $p5s,p1w ;\
+ not %eax ;\
+ add p2x,p1w
+
+ REPEAT4(%esi,%edi,%ecx,0xf4292244, 6, 7,%ecx)
+ REPEAT4(%edx,%esi,%edi,0x432aff97,10,14,%edi)
+ REPEAT4(%ecx,%edx,%esi,0xab9423a7,15, 5,%esi)
+ REPEAT4(%edi,%ecx,%edx,0xfc93a039,21,12,%edx)
+ REPEAT4(%esi,%edi,%ecx,0x655b59c3, 6, 3,%ecx)
+ REPEAT4(%edx,%esi,%edi,0x8f0ccc92,10,10,%edi)
+ REPEAT4(%ecx,%edx,%esi,0xffeff47d,15, 1,%esi)
+ REPEAT4(%edi,%ecx,%edx,0x85845dd1,21, 8,%edx)
+ REPEAT4(%esi,%edi,%ecx,0x6fa87e4f, 6,15,%ecx)
+ REPEAT4(%edx,%esi,%edi,0xfe2ce6e0,10, 6,%edi)
+ REPEAT4(%ecx,%edx,%esi,0xa3014314,15,13,%esi)
+ REPEAT4(%edi,%ecx,%edx,0x4e0811a1,21, 4,%edx)
+ REPEAT4(%esi,%edi,%ecx,0xf7537e82, 6,11,%ecx)
+ REPEAT4(%edx,%esi,%edi,0xbd3af235,10, 2,%edi)
+ REPEAT4(%ecx,%edx,%esi,0x2ad7d2bb,15, 9,%esi)
+
+ add $0xeb86d391,%edi
+ or %ecx,%eax
+ add %ebx,%edi
+ xor %edx,%eax
+ mov 4+16(%esp),%ebp // pointer to 'hash' output
+ add %eax,%edi
+ rol $21,%edi
+ add %ecx,%edi
+
+ add %esi,(%ebp)
+ add %edi,4(%ebp)
+ add %ecx,8(%ebp)
+ add %edx,12(%ebp)
+
+ pop %edi
+ pop %esi
+ pop %ebx
+ pop %ebp
+ ret
+
+#if defined(__ELF__) && defined(SECTION_NOTE_GNU_STACK)
+ .section .note.GNU-stack,"",@progbits
+#endif
diff -urN linux-3.14-noloop/drivers/misc/md5.c linux-3.14-AES/drivers/misc/md5.c
--- linux-3.14-noloop/drivers/misc/md5.c 1970-01-01 02:00:00.000000000 +0200
+++ linux-3.14-AES/drivers/misc/md5.c 2014-03-31 22:20:36.000000000 +0300
@@ -0,0 +1,106 @@
+/*
+ * MD5 Message Digest Algorithm (RFC1321).
+ *
+ * Derived from cryptoapi implementation, originally based on the
+ * public domain implementation written by Colin Plumb in 1993.
+ *
+ * Copyright (c) Cryptoapi developers.
+ * Copyright (c) 2002 James Morris <jmorris@intercode.com.au>
+ *
+ * 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.
+ */
+
+#include "md5.h"
+
+#define MD5_F1(x, y, z) (z ^ (x & (y ^ z)))
+#define MD5_F2(x, y, z) MD5_F1(z, x, y)
+#define MD5_F3(x, y, z) (x ^ y ^ z)
+#define MD5_F4(x, y, z) (y ^ (x | ~z))
+#define MD5_STEP(f, w, x, y, z, in, s) \
+ (w += f(x, y, z) + in, w = (w<<s | w>>(32-s)) + x)
+
+void md5_transform_CPUbyteorder(u_int32_t *hash, u_int32_t const *in)
+{
+ u_int32_t a, b, c, d;
+
+ a = hash[0];
+ b = hash[1];
+ c = hash[2];
+ d = hash[3];
+
+ MD5_STEP(MD5_F1, a, b, c, d, in[0] + 0xd76aa478, 7);
+ MD5_STEP(MD5_F1, d, a, b, c, in[1] + 0xe8c7b756, 12);
+ MD5_STEP(MD5_F1, c, d, a, b, in[2] + 0x242070db, 17);
+ MD5_STEP(MD5_F1, b, c, d, a, in[3] + 0xc1bdceee, 22);
+ MD5_STEP(MD5_F1, a, b, c, d, in[4] + 0xf57c0faf, 7);
+ MD5_STEP(MD5_F1, d, a, b, c, in[5] + 0x4787c62a, 12);
+ MD5_STEP(MD5_F1, c, d, a, b, in[6] + 0xa8304613, 17);
+ MD5_STEP(MD5_F1, b, c, d, a, in[7] + 0xfd469501, 22);
+ MD5_STEP(MD5_F1, a, b, c, d, in[8] + 0x698098d8, 7);
+ MD5_STEP(MD5_F1, d, a, b, c, in[9] + 0x8b44f7af, 12);
+ MD5_STEP(MD5_F1, c, d, a, b, in[10] + 0xffff5bb1, 17);
+ MD5_STEP(MD5_F1, b, c, d, a, in[11] + 0x895cd7be, 22);
+ MD5_STEP(MD5_F1, a, b, c, d, in[12] + 0x6b901122, 7);
+ MD5_STEP(MD5_F1, d, a, b, c, in[13] + 0xfd987193, 12);
+ MD5_STEP(MD5_F1, c, d, a, b, in[14] + 0xa679438e, 17);
+ MD5_STEP(MD5_F1, b, c, d, a, in[15] + 0x49b40821, 22);
+
+ MD5_STEP(MD5_F2, a, b, c, d, in[1] + 0xf61e2562, 5);
+ MD5_STEP(MD5_F2, d, a, b, c, in[6] + 0xc040b340, 9);
+ MD5_STEP(MD5_F2, c, d, a, b, in[11] + 0x265e5a51, 14);
+ MD5_STEP(MD5_F2, b, c, d, a, in[0] + 0xe9b6c7aa, 20);
+ MD5_STEP(MD5_F2, a, b, c, d, in[5] + 0xd62f105d, 5);
+ MD5_STEP(MD5_F2, d, a, b, c, in[10] + 0x02441453, 9);
+ MD5_STEP(MD5_F2, c, d, a, b, in[15] + 0xd8a1e681, 14);
+ MD5_STEP(MD5_F2, b, c, d, a, in[4] + 0xe7d3fbc8, 20);
+ MD5_STEP(MD5_F2, a, b, c, d, in[9] + 0x21e1cde6, 5);
+ MD5_STEP(MD5_F2, d, a, b, c, in[14] + 0xc33707d6, 9);
+ MD5_STEP(MD5_F2, c, d, a, b, in[3] + 0xf4d50d87, 14);
+ MD5_STEP(MD5_F2, b, c, d, a, in[8] + 0x455a14ed, 20);
+ MD5_STEP(MD5_F2, a, b, c, d, in[13] + 0xa9e3e905, 5);
+ MD5_STEP(MD5_F2, d, a, b, c, in[2] + 0xfcefa3f8, 9);
+ MD5_STEP(MD5_F2, c, d, a, b, in[7] + 0x676f02d9, 14);
+ MD5_STEP(MD5_F2, b, c, d, a, in[12] + 0x8d2a4c8a, 20);
+
+ MD5_STEP(MD5_F3, a, b, c, d, in[5] + 0xfffa3942, 4);
+ MD5_STEP(MD5_F3, d, a, b, c, in[8] + 0x8771f681, 11);
+ MD5_STEP(MD5_F3, c, d, a, b, in[11] + 0x6d9d6122, 16);
+ MD5_STEP(MD5_F3, b, c, d, a, in[14] + 0xfde5380c, 23);
+ MD5_STEP(MD5_F3, a, b, c, d, in[1] + 0xa4beea44, 4);
+ MD5_STEP(MD5_F3, d, a, b, c, in[4] + 0x4bdecfa9, 11);
+ MD5_STEP(MD5_F3, c, d, a, b, in[7] + 0xf6bb4b60, 16);
+ MD5_STEP(MD5_F3, b, c, d, a, in[10] + 0xbebfbc70, 23);
+ MD5_STEP(MD5_F3, a, b, c, d, in[13] + 0x289b7ec6, 4);
+ MD5_STEP(MD5_F3, d, a, b, c, in[0] + 0xeaa127fa, 11);
+ MD5_STEP(MD5_F3, c, d, a, b, in[3] + 0xd4ef3085, 16);
+ MD5_STEP(MD5_F3, b, c, d, a, in[6] + 0x04881d05, 23);
+ MD5_STEP(MD5_F3, a, b, c, d, in[9] + 0xd9d4d039, 4);
+ MD5_STEP(MD5_F3, d, a, b, c, in[12] + 0xe6db99e5, 11);
+ MD5_STEP(MD5_F3, c, d, a, b, in[15] + 0x1fa27cf8, 16);
+ MD5_STEP(MD5_F3, b, c, d, a, in[2] + 0xc4ac5665, 23);
+
+ MD5_STEP(MD5_F4, a, b, c, d, in[0] + 0xf4292244, 6);
+ MD5_STEP(MD5_F4, d, a, b, c, in[7] + 0x432aff97, 10);
+ MD5_STEP(MD5_F4, c, d, a, b, in[14] + 0xab9423a7, 15);
+ MD5_STEP(MD5_F4, b, c, d, a, in[5] + 0xfc93a039, 21);
+ MD5_STEP(MD5_F4, a, b, c, d, in[12] + 0x655b59c3, 6);
+ MD5_STEP(MD5_F4, d, a, b, c, in[3] + 0x8f0ccc92, 10);
+ MD5_STEP(MD5_F4, c, d, a, b, in[10] + 0xffeff47d, 15);
+ MD5_STEP(MD5_F4, b, c, d, a, in[1] + 0x85845dd1, 21);
+ MD5_STEP(MD5_F4, a, b, c, d, in[8] + 0x6fa87e4f, 6);
+ MD5_STEP(MD5_F4, d, a, b, c, in[15] + 0xfe2ce6e0, 10);
+ MD5_STEP(MD5_F4, c, d, a, b, in[6] + 0xa3014314, 15);
+ MD5_STEP(MD5_F4, b, c, d, a, in[13] + 0x4e0811a1, 21);
+ MD5_STEP(MD5_F4, a, b, c, d, in[4] + 0xf7537e82, 6);
+ MD5_STEP(MD5_F4, d, a, b, c, in[11] + 0xbd3af235, 10);
+ MD5_STEP(MD5_F4, c, d, a, b, in[2] + 0x2ad7d2bb, 15);
+ MD5_STEP(MD5_F4, b, c, d, a, in[9] + 0xeb86d391, 21);
+
+ hash[0] += a;
+ hash[1] += b;
+ hash[2] += c;
+ hash[3] += d;
+}
diff -urN linux-3.14-noloop/drivers/misc/md5.h linux-3.14-AES/drivers/misc/md5.h
--- linux-3.14-noloop/drivers/misc/md5.h 1970-01-01 02:00:00.000000000 +0200
+++ linux-3.14-AES/drivers/misc/md5.h 2014-03-31 22:20:36.000000000 +0300
@@ -0,0 +1,15 @@
+/* md5.h */
+
+#include <linux/types.h>
+#include <linux/linkage.h>
+#include <linux/module.h>
+
+#if defined(CONFIG_X86) || defined(CONFIG_X86_64)
+ asmlinkage
+#endif
+extern void md5_transform_CPUbyteorder(u_int32_t *, u_int32_t const *);
+
+#if defined(CONFIG_X86) || defined(CONFIG_X86_64)
+ asmlinkage
+#endif
+extern void md5_transform_CPUbyteorder_2x(u_int32_t *, u_int32_t const *, u_int32_t const *);
diff -urN linux-3.14-noloop/include/linux/loop.h linux-3.14-AES/include/linux/loop.h
--- linux-3.14-noloop/include/linux/loop.h 1970-01-01 02:00:00.000000000 +0200
+++ linux-3.14-AES/include/linux/loop.h 2014-03-31 22:20:36.000000000 +0300
@@ -0,0 +1,171 @@
+#ifndef _LINUX_LOOP_H
+#define _LINUX_LOOP_H
+
+/*
+ * include/linux/loop.h
+ *
+ * Written by Theodore Ts'o, 3/29/93.
+ *
+ * Copyright 1993 by Theodore Ts'o. Redistribution of this file is
+ * permitted under the GNU General Public License.
+ */
+
+#define LO_NAME_SIZE 64
+#define LO_KEY_SIZE 32
+
+#ifdef __KERNEL__
+#include <linux/version.h>
+#include <linux/bio.h>
+#include <linux/blkdev.h>
+#include <linux/module.h>
+#include <linux/spinlock.h>
+
+struct loop_func_table;
+
+struct loop_device {
+ int lo_number;
+ int lo_refcnt;
+ loff_t lo_offset;
+ loff_t lo_sizelimit;
+ int lo_flags;
+ int (*transfer)(struct loop_device *, int cmd,
+ char *raw_buf, char *loop_buf, int size,
+ sector_t real_block);
+ struct loop_func_table *lo_encryption;
+ char lo_file_name[LO_NAME_SIZE];
+ char lo_crypt_name[LO_NAME_SIZE];
+ char lo_encrypt_key[LO_KEY_SIZE];
+ int lo_encrypt_key_size;
+#if LINUX_VERSION_CODE >= 0x30600
+ kuid_t lo_key_owner; /* Who set the key */
+#else
+ uid_t lo_key_owner; /* Who set the key */
+#endif
+ __u32 lo_init[2];
+ int (*ioctl)(struct loop_device *, int cmd,
+ unsigned long arg);
+
+ struct file * lo_backing_file;
+ struct block_device *lo_device;
+ void *key_data;
+
+ int old_gfp_mask;
+
+ spinlock_t lo_lock;
+ struct completion lo_done;
+ atomic_t lo_pending;
+
+ struct request_queue *lo_queue;
+
+ struct bio *lo_bio_que0;
+ struct bio *lo_bio_free0;
+ struct bio *lo_bio_free1;
+ int lo_bio_flshMax;
+ int lo_bio_flshCnt;
+ wait_queue_head_t lo_bio_wait;
+ wait_queue_head_t lo_buf_wait;
+ sector_t lo_offs_sec;
+ sector_t lo_iv_remove;
+ spinlock_t lo_ioctl_spin;
+ int lo_ioctl_busy;
+ wait_queue_head_t lo_ioctl_wait;
+ struct request_queue *lo_backingQueue;
+#ifdef CONFIG_BLK_DEV_LOOP_KEYSCRUB
+ void (*lo_keyscrub_fn)(void *);
+ void *lo_keyscrub_ptr;
+#endif
+};
+
+#endif /* __KERNEL__ */
+
+/*
+ * Loop flags
+ */
+#define LO_FLAGS_DO_BMAP 1
+#define LO_FLAGS_READ_ONLY 2
+
+#include <asm/posix_types.h> /* for __kernel_old_dev_t */
+#include <asm/types.h> /* for __u64 */
+
+/* Backwards compatibility version */
+struct loop_info {
+ int lo_number; /* ioctl r/o */
+ __kernel_old_dev_t lo_device; /* ioctl r/o */
+ unsigned long lo_inode; /* ioctl r/o */
+ __kernel_old_dev_t lo_rdevice; /* ioctl r/o */
+ int lo_offset;
+ int lo_encrypt_type;
+ int lo_encrypt_key_size; /* ioctl w/o */
+ int lo_flags; /* ioctl r/o */
+ char lo_name[LO_NAME_SIZE];
+ unsigned char lo_encrypt_key[LO_KEY_SIZE]; /* ioctl w/o */
+ unsigned long lo_init[2];
+ char reserved[4];
+};
+
+struct loop_info64 {
+ __u64 lo_device; /* ioctl r/o */
+ __u64 lo_inode; /* ioctl r/o */
+ __u64 lo_rdevice; /* ioctl r/o */
+ __u64 lo_offset;
+ __u64 lo_sizelimit;/* bytes, 0 == max available */
+ __u32 lo_number; /* ioctl r/o */
+ __u32 lo_encrypt_type;
+ __u32 lo_encrypt_key_size; /* ioctl w/o */
+ __u32 lo_flags; /* ioctl r/o */
+ __u8 lo_file_name[LO_NAME_SIZE];
+ __u8 lo_crypt_name[LO_NAME_SIZE];
+ __u8 lo_encrypt_key[LO_KEY_SIZE]; /* ioctl w/o */
+ __u64 lo_init[2];
+};
+
+/*
+ * Loop filter types
+ */
+
+#define LO_CRYPT_NONE 0
+#define LO_CRYPT_XOR 1
+#define LO_CRYPT_DES 2
+#define LO_CRYPT_FISH2 3 /* Twofish encryption */
+#define LO_CRYPT_BLOW 4
+#define LO_CRYPT_CAST128 5
+#define LO_CRYPT_IDEA 6
+#define LO_CRYPT_DUMMY 9
+#define LO_CRYPT_SKIPJACK 10
+#define LO_CRYPT_AES 16
+#define LO_CRYPT_CRYPTOAPI 18
+#define MAX_LO_CRYPT 20
+
+#ifdef __KERNEL__
+/* Support for loadable transfer modules */
+struct loop_func_table {
+ int number; /* filter type */
+ int (*transfer)(struct loop_device *lo, int cmd, char *raw_buf,
+ char *loop_buf, int size, sector_t real_block);
+ int (*init)(struct loop_device *, struct loop_info64 *);
+ /* release is called from loop_unregister_transfer or clr_fd */
+ int (*release)(struct loop_device *);
+ int (*ioctl)(struct loop_device *, int cmd, unsigned long arg);
+ struct module *owner;
+};
+
+int loop_register_transfer(struct loop_func_table *funcs);
+int loop_unregister_transfer(int number);
+
+#endif
+/*
+ * IOCTL commands --- we will commandeer 0x4C ('L')
+ */
+
+#define LOOP_SET_FD 0x4C00
+#define LOOP_CLR_FD 0x4C01
+#define LOOP_SET_STATUS 0x4C02
+#define LOOP_GET_STATUS 0x4C03
+#define LOOP_SET_STATUS64 0x4C04
+#define LOOP_GET_STATUS64 0x4C05
+#define LOOP_CHANGE_FD 0x4C06
+
+#define LOOP_MULTI_KEY_SETUP 0x4C4D
+#define LOOP_MULTI_KEY_SETUP_V3 0x4C4E
+#define LOOP_RECOMPUTE_DEV_SIZE 0x4C52
+#endif
Reference in a new issue Copy permalink