File: glue.c

package info (click to toggle)
loop-aes 3.3a-1
  • links: PTS
  • area: main
  • in suites: squeeze
  • size: 1,856 kB
  • ctags: 833
  • sloc: ansic: 4,603; asm: 2,475; sh: 840; makefile: 467
file content (982 lines) | stat: -rw-r--r-- 32,227 bytes parent folder | download
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
/*
 *  glue.c
 *
 *  Written by Jari Ruusu, March 13 2010
 *
 *  Copyright 2001-2010 by Jari Ruusu.
 *  Redistribution of this file is permitted under the GNU Public License.
 */

#include <linux/version.h>
#include <linux/sched.h>
#include <linux/fs.h>
#include <linux/string.h>
#include <linux/types.h>
#include <linux/errno.h>
#if LINUX_VERSION_CODE >= 0x20600
# include <linux/bio.h>
# include <linux/blkdev.h>
#endif
#if LINUX_VERSION_CODE >= 0x20200
# include <linux/slab.h>
# include <linux/loop.h>
# include <asm/uaccess.h>
#else
# include <linux/malloc.h>
# include <asm/segment.h>
# include "patched-loop.h"
#endif
#if LINUX_VERSION_CODE >= 0x20400
# include <linux/spinlock.h>
#endif
#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
#include "aes.h"
#include "md5.h"

#if LINUX_VERSION_CODE >= 0x20600
typedef sector_t TransferSector_t;
# define LoopInfo_t struct loop_info64
#else
typedef int TransferSector_t;
# define LoopInfo_t struct loop_info
#endif

#if !defined(cpu_to_le32)
# if defined(__BIG_ENDIAN)
#  define cpu_to_le32(x) ({u_int32_t __x=(x);((u_int32_t)((((u_int32_t)(__x)&(u_int32_t)0x000000ffUL)<<24)|(((u_int32_t)(__x)&(u_int32_t)0x0000ff00UL)<<8)|(((u_int32_t)(__x)&(u_int32_t)0x00ff0000UL)>>8)|(((u_int32_t)(__x)&(u_int32_t)0xff000000UL)>>24)));})
# else
#  define cpu_to_le32(x) ((u_int32_t)(x))
# endif
#endif

#if LINUX_VERSION_CODE < 0x20200
# define copy_from_user(t,f,s) (verify_area(VERIFY_READ,f,s)?(s):(memcpy_fromfs(t,f,s),0))
#endif

#if !defined(LOOP_MULTI_KEY_SETUP)
# define LOOP_MULTI_KEY_SETUP 0x4C4D
#endif
#if !defined(LOOP_MULTI_KEY_SETUP_V3)
# define LOOP_MULTI_KEY_SETUP_V3 0x4C4E
#endif

#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 >= 0x20200
#if 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
#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;
}

void loop_compute_sector_iv(TransferSector_t devSect, u_int32_t *ivout)
{
    if(sizeof(TransferSector_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(TransferSector_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(TransferSector_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(TransferSector_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(TransferSector_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);
}

int transfer_aes(struct loop_device *lo, int cmd, char *raw_buf,
          char *loop_buf, int size, TransferSector_t devSect)
{
    aes_context     *a;
    AESmultiKey     *m;
    int             x;
    unsigned        y;
    u_int32_t       iv[8];

    if(!size || (size & 511)) {
        return -EINVAL;
    }
    m = (AESmultiKey *)lo->key_data;
    y = m->keyMask;
    if(cmd == READ) {
        while(size) {
#ifdef CONFIG_BLK_DEV_LOOP_KEYSCRUB
            read_lock(&m->rwlock);
#endif
            a = m->keyPtr[((unsigned)devSect) & y];
            if(y) {
                memcpy(&iv[0], raw_buf, 16);
                raw_buf += 16;
                loop_buf += 16;
            } else {
                loop_compute_sector_iv(devSect, &iv[0]);
            }
            x = 15;
            do {
                memcpy(&iv[4], raw_buf, 16);
                aes_decrypt(a, raw_buf, loop_buf);
                *((u_int32_t *)(&loop_buf[ 0])) ^= iv[0];
                *((u_int32_t *)(&loop_buf[ 4])) ^= iv[1];
                *((u_int32_t *)(&loop_buf[ 8])) ^= iv[2];
                *((u_int32_t *)(&loop_buf[12])) ^= iv[3];
                if(y && !x) {
                    raw_buf -= 496;
                    loop_buf -= 496;
                    memcpy(&iv[4], &m->partialMD5[0], 16);
                    loop_compute_md5_iv_v3(devSect, &iv[4], (u_int32_t *)(&loop_buf[16]));
                } else {
                    raw_buf += 16;
                    loop_buf += 16;
                    memcpy(&iv[0], raw_buf, 16);
                }
                aes_decrypt(a, raw_buf, loop_buf);
                *((u_int32_t *)(&loop_buf[ 0])) ^= iv[4];
                *((u_int32_t *)(&loop_buf[ 4])) ^= iv[5];
                *((u_int32_t *)(&loop_buf[ 8])) ^= iv[6];
                *((u_int32_t *)(&loop_buf[12])) ^= iv[7];
                if(y && !x) {
                    raw_buf += 512;
                    loop_buf += 512;
                } else {
                    raw_buf += 16;
                    loop_buf += 16;
                }
            } while(--x >= 0);
#ifdef CONFIG_BLK_DEV_LOOP_KEYSCRUB
            read_unlock(&m->rwlock);
#endif
#if LINUX_VERSION_CODE >= 0x20600
            cond_resched();
#elif LINUX_VERSION_CODE >= 0x20400
            if(current->need_resched) {set_current_state(TASK_RUNNING);schedule();}
#elif LINUX_VERSION_CODE >= 0x20200
            if(current->need_resched) {current->state=TASK_RUNNING;schedule();}
#else
            if(need_resched) schedule();
#endif
            size -= 512;
            devSect++;
        }
    } else {
        while(size) {
#ifdef CONFIG_BLK_DEV_LOOP_KEYSCRUB
            read_lock(&m->rwlock);
#endif
            a = m->keyPtr[((unsigned)devSect) & y];
            if(y) {
#if LINUX_VERSION_CODE < 0x20400
                /* on 2.2 and older kernels, real raw_buf may be doing */
                /* writes at any time, so this needs to be stack buffer */
                u_int32_t tmp_raw_buf[128];
                char *TMP_RAW_BUF = (char *)(&tmp_raw_buf[0]);
#else
                /* on 2.4 and later kernels, real raw_buf is not doing */
                /* any writes now so it can be used as temp buffer */
# define TMP_RAW_BUF raw_buf
#endif
                memcpy(TMP_RAW_BUF, loop_buf, 512);
                memcpy(&iv[0], &m->partialMD5[0], 16);
                loop_compute_md5_iv_v3(devSect, &iv[0], (u_int32_t *)(&TMP_RAW_BUF[16]));
                x = 15;
                do {
                    iv[0] ^= *((u_int32_t *)(&TMP_RAW_BUF[ 0]));
                    iv[1] ^= *((u_int32_t *)(&TMP_RAW_BUF[ 4]));
                    iv[2] ^= *((u_int32_t *)(&TMP_RAW_BUF[ 8]));
                    iv[3] ^= *((u_int32_t *)(&TMP_RAW_BUF[12]));
                    aes_encrypt(a, (unsigned char *)(&iv[0]), raw_buf);
                    memcpy(&iv[0], raw_buf, 16);
                    raw_buf += 16;
#if LINUX_VERSION_CODE < 0x20400
                    TMP_RAW_BUF += 16;
#endif
                    iv[0] ^= *((u_int32_t *)(&TMP_RAW_BUF[ 0]));
                    iv[1] ^= *((u_int32_t *)(&TMP_RAW_BUF[ 4]));
                    iv[2] ^= *((u_int32_t *)(&TMP_RAW_BUF[ 8]));
                    iv[3] ^= *((u_int32_t *)(&TMP_RAW_BUF[12]));
                    aes_encrypt(a, (unsigned char *)(&iv[0]), raw_buf);
                    memcpy(&iv[0], raw_buf, 16);
                    raw_buf += 16;
#if LINUX_VERSION_CODE < 0x20400
                    TMP_RAW_BUF += 16;
#endif
                } while(--x >= 0);
                loop_buf += 512;
            } else {
                loop_compute_sector_iv(devSect, &iv[0]);
                x = 15;
                do {
                    iv[0] ^= *((u_int32_t *)(&loop_buf[ 0]));
                    iv[1] ^= *((u_int32_t *)(&loop_buf[ 4]));
                    iv[2] ^= *((u_int32_t *)(&loop_buf[ 8]));
                    iv[3] ^= *((u_int32_t *)(&loop_buf[12]));
                    aes_encrypt(a, (unsigned char *)(&iv[0]), raw_buf);
                    memcpy(&iv[0], raw_buf, 16);
                    loop_buf += 16;
                    raw_buf += 16;
                    iv[0] ^= *((u_int32_t *)(&loop_buf[ 0]));
                    iv[1] ^= *((u_int32_t *)(&loop_buf[ 4]));
                    iv[2] ^= *((u_int32_t *)(&loop_buf[ 8]));
                    iv[3] ^= *((u_int32_t *)(&loop_buf[12]));
                    aes_encrypt(a, (unsigned char *)(&iv[0]), raw_buf);
                    memcpy(&iv[0], raw_buf, 16);
                    loop_buf += 16;
                    raw_buf += 16;
                } while(--x >= 0);
            }
#ifdef CONFIG_BLK_DEV_LOOP_KEYSCRUB
            read_unlock(&m->rwlock);
#endif
#if LINUX_VERSION_CODE >= 0x20600
            cond_resched();
#elif LINUX_VERSION_CODE >= 0x20400
            if(current->need_resched) {set_current_state(TASK_RUNNING);schedule();}
#elif LINUX_VERSION_CODE >= 0x20200
            if(current->need_resched) {current->state=TASK_RUNNING;schedule();}
#else
            if(need_resched) schedule();
#endif
            size -= 512;
            devSect++;
        }
    }
    return(0);
}

int keySetup_aes(struct loop_device *lo, LoopInfo_t *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);
}

int keyClean_aes(struct loop_device *lo)
{
    if(lo->key_data) {
        clearAndFreeMultiKey((AESmultiKey *)lo->key_data);
        lo->key_data = 0;
    }
    return(0);
}

#if defined(CONFIG_BLK_DEV_LOOP_PADLOCK) && (defined(CONFIG_X86) || defined(CONFIG_X86_64))
#if LINUX_VERSION_CODE < 0x20400
#error "this code does not support padlock crypto instructions on 2.2 or older kernels"
#endif

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 {
    u_int32_t   iv[4];
    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, TransferSector_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));

    if(!size || (size & 511) || (((unsigned long)raw_buf | (unsigned long)loop_buf) & 15)) {
        return -EINVAL;
    }
    m = (AESmultiKey *)lo->key_data;
    y = m->keyMask;
    if(cmd == READ) {
        while(size) {
            padlock_flush_key_context();
            ivcw->cw[0] = m->padlock_cw_d;
            ivcw->cw[3] = ivcw->cw[2] = ivcw->cw[1] = 0;
#ifdef CONFIG_BLK_DEV_LOOP_KEYSCRUB
            read_lock(&m->rwlock);
#endif
            a = m->keyPtr[((unsigned)devSect) & y];
            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, &ivcw->iv[0], (u_int32_t *)(&loop_buf[16]));
                /* XOR with correct IV now */
                *((u_int32_t *)(&loop_buf[ 0])) ^= ivcw->iv[0];
                *((u_int32_t *)(&loop_buf[ 4])) ^= ivcw->iv[1];
                *((u_int32_t *)(&loop_buf[ 8])) ^= ivcw->iv[2];
                *((u_int32_t *)(&loop_buf[12])) ^= ivcw->iv[3];
            } else {
                loop_compute_sector_iv(devSect, &ivcw->iv[0]);
                padlock_rep_xcryptcbc(&ivcw->cw[0], &a->aes_d_key[0], raw_buf, loop_buf, &ivcw->iv[0], 32);
            }
#ifdef CONFIG_BLK_DEV_LOOP_KEYSCRUB
            read_unlock(&m->rwlock);
#endif
#if LINUX_VERSION_CODE >= 0x20600
            cond_resched();
#else
            if(current->need_resched) {set_current_state(TASK_RUNNING);schedule();}
#endif
            size -= 512;
            raw_buf += 512;
            loop_buf += 512;
            devSect++;
        }
    } else {
        while(size) {
            padlock_flush_key_context();
            ivcw->cw[0] = m->padlock_cw_e;
            ivcw->cw[3] = ivcw->cw[2] = ivcw->cw[1] = 0;
#ifdef CONFIG_BLK_DEV_LOOP_KEYSCRUB
            read_lock(&m->rwlock);
#endif
            a = m->keyPtr[((unsigned)devSect) & y];
            if(y) {
                memcpy(raw_buf, loop_buf, 512);
                memcpy(&ivcw->iv[0], &m->partialMD5[0], 16);
                loop_compute_md5_iv_v3(devSect, &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, &ivcw->iv[0]);
                padlock_rep_xcryptcbc(&ivcw->cw[0], &a->aes_e_key[0], loop_buf, raw_buf, &ivcw->iv[0], 32);
            }
#ifdef CONFIG_BLK_DEV_LOOP_KEYSCRUB
            read_unlock(&m->rwlock);
#endif
#if LINUX_VERSION_CODE >= 0x20600
            cond_resched();
#else
            if(current->need_resched) {set_current_state(TASK_RUNNING);schedule();}
#endif
            size -= 512;
            raw_buf += 512;
            loop_buf += 512;
            devSect++;
        }
    }
    return(0);
}
#endif

#if defined(CONFIG_BLK_DEV_LOOP_INTELAES) && (defined(CONFIG_X86) || defined(CONFIG_X86_64))
#if LINUX_VERSION_CODE < 0x20400
#error "this code does not support Intel AES crypto instructions on 2.2 or older kernels"
#endif

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);

static int transfer_intel_aes(struct loop_device *lo, int cmd, char *raw_buf,
          char *loop_buf, int size, TransferSector_t devSect)
{
    aes_context     *a;
    AESmultiKey     *m;
    unsigned        y;
    u_int32_t       ivua[4+4];
    u_int32_t       *iv;

    /* make iv 16 byte aligned */
    iv = (u_int32_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;
    if(cmd == READ) {
        while(size) {
#ifdef CONFIG_BLK_DEV_LOOP_KEYSCRUB
            read_lock(&m->rwlock);
#endif
            a = m->keyPtr[((unsigned)devSect) & y];
            if(y) {
                /* decrypt using fake all-zero IV */
                memset(&iv[0], 0, 16);
                kernel_fpu_begin();
                intel_aes_cbc_decrypt(a, raw_buf, loop_buf, 512, iv);
                kernel_fpu_end();
                /* compute correct IV */
                memcpy(&iv[0], &m->partialMD5[0], 16);
                loop_compute_md5_iv_v3(devSect, &iv[0], (u_int32_t *)(&loop_buf[16]));
                /* XOR with correct IV now */
                *((u_int32_t *)(&loop_buf[ 0])) ^= iv[0];
                *((u_int32_t *)(&loop_buf[ 4])) ^= iv[1];
                *((u_int32_t *)(&loop_buf[ 8])) ^= iv[2];
                *((u_int32_t *)(&loop_buf[12])) ^= iv[3];
            } else {
                loop_compute_sector_iv(devSect, &iv[0]);
                kernel_fpu_begin();
                intel_aes_cbc_decrypt(a, raw_buf, loop_buf, 512, iv);
                kernel_fpu_end();
            }
#ifdef CONFIG_BLK_DEV_LOOP_KEYSCRUB
            read_unlock(&m->rwlock);
#endif
#if LINUX_VERSION_CODE >= 0x20600
            cond_resched();
#else
            if(current->need_resched) {set_current_state(TASK_RUNNING);schedule();}
#endif
            size -= 512;
            raw_buf += 512;
            loop_buf += 512;
            devSect++;
        }
    } else {
        while(size) {
#ifdef CONFIG_BLK_DEV_LOOP_KEYSCRUB
            read_lock(&m->rwlock);
#endif
            a = m->keyPtr[((unsigned)devSect) & y];
            if(y) {
                memcpy(raw_buf, loop_buf, 512);
                memcpy(&iv[0], &m->partialMD5[0], 16);
                loop_compute_md5_iv_v3(devSect, &iv[0], (u_int32_t *)(&raw_buf[16]));
                kernel_fpu_begin();
                intel_aes_cbc_encrypt(a, raw_buf, raw_buf, 512, iv);
                kernel_fpu_end();
            } else {
                loop_compute_sector_iv(devSect, &iv[0]);
                kernel_fpu_begin();
                intel_aes_cbc_encrypt(a, loop_buf, raw_buf, 512, iv);
                kernel_fpu_end();
            }
#ifdef CONFIG_BLK_DEV_LOOP_KEYSCRUB
            read_unlock(&m->rwlock);
#endif
#if LINUX_VERSION_CODE >= 0x20600
            cond_resched();
#else
            if(current->need_resched) {set_current_state(TASK_RUNNING);schedule();}
#endif
            size -= 512;
            raw_buf += 512;
            loop_buf += 512;
            devSect++;
        }
    }
    return(0);
}
#endif

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;
}

#if LINUX_VERSION_CODE >= 0x20200

static struct loop_func_table funcs_aes = {
    number:     16,     /* 16 == AES */
    transfer:   (void *) transfer_aes,
    init:       (void *) keySetup_aes,
    release:    keyClean_aes,
    ioctl:      (void *) 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:   (void *) transfer_padlock_aes,
    init:       (void *) keySetup_aes,
    release:    keyClean_aes,
    ioctl:      (void *) 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:   (void *) transfer_intel_aes,
    init:       (void *) keySetup_aes,
    release:    keyClean_aes,
    ioctl:      (void *) handleIoctl_aes
};
#endif

int init_module_aes(void)
{
#if defined(CONFIG_BLK_DEV_LOOP_PADLOCK) && (defined(CONFIG_X86) || defined(CONFIG_X86_64))
    if((boot_cpu_data.x86 >= 6) && (boot_cpu_data.x86_vendor == X86_VENDOR_CENTAUR)
      && (cpuid_eax(0xC0000000) >= 0xC0000001) && ((cpuid_edx(0xC0000001) & 0xC0) == 0xC0)) {
        if(loop_register_transfer(&funcs_padlock_aes)) {
            printk("loop: unable to register padlock AES transfer\n");
            return -EIO;
        }
        printk("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)) {
        if(loop_register_transfer(&funcs_intel_aes)) {
            printk("loop: unable to register Intel AES transfer\n");
            return -EIO;
        }
        printk("loop: Intel hardware AES enabled\n");
    } else
#endif
    if(loop_register_transfer(&funcs_aes)) {
        printk("loop: unable to register AES transfer\n");
        return -EIO;
    }
#ifdef CONFIG_BLK_DEV_LOOP_KEYSCRUB
    printk("loop: AES key scrubbing enabled\n");
#endif
    return 0;
}

void cleanup_module_aes(void)
{
    if(loop_unregister_transfer(funcs_aes.number)) {
        printk("loop: unable to unregister AES transfer\n");
    }
}

#endif