xref: /openbmc/linux/crypto/vmac.c (revision b6dcefde)
1 /*
2  * Modified to interface to the Linux kernel
3  * Copyright (c) 2009, Intel Corporation.
4  *
5  * This program is free software; you can redistribute it and/or modify it
6  * under the terms and conditions of the GNU General Public License,
7  * version 2, as published by the Free Software Foundation.
8  *
9  * This program is distributed in the hope it will be useful, but WITHOUT
10  * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
11  * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
12  * more details.
13  *
14  * You should have received a copy of the GNU General Public License along with
15  * this program; if not, write to the Free Software Foundation, Inc., 59 Temple
16  * Place - Suite 330, Boston, MA 02111-1307 USA.
17  */
18 
19 /* --------------------------------------------------------------------------
20  * VMAC and VHASH Implementation by Ted Krovetz (tdk@acm.org) and Wei Dai.
21  * This implementation is herby placed in the public domain.
22  * The authors offers no warranty. Use at your own risk.
23  * Please send bug reports to the authors.
24  * Last modified: 17 APR 08, 1700 PDT
25  * ----------------------------------------------------------------------- */
26 
27 #include <linux/init.h>
28 #include <linux/types.h>
29 #include <linux/crypto.h>
30 #include <linux/scatterlist.h>
31 #include <asm/byteorder.h>
32 #include <crypto/scatterwalk.h>
33 #include <crypto/vmac.h>
34 #include <crypto/internal/hash.h>
35 
36 /*
37  * Constants and masks
38  */
39 #define UINT64_C(x) x##ULL
40 const u64 p64   = UINT64_C(0xfffffffffffffeff);  /* 2^64 - 257 prime  */
41 const u64 m62   = UINT64_C(0x3fffffffffffffff);  /* 62-bit mask       */
42 const u64 m63   = UINT64_C(0x7fffffffffffffff);  /* 63-bit mask       */
43 const u64 m64   = UINT64_C(0xffffffffffffffff);  /* 64-bit mask       */
44 const u64 mpoly = UINT64_C(0x1fffffff1fffffff);  /* Poly key mask     */
45 
46 #ifdef __LITTLE_ENDIAN
47 #define INDEX_HIGH 1
48 #define INDEX_LOW 0
49 #else
50 #define INDEX_HIGH 0
51 #define INDEX_LOW 1
52 #endif
53 
54 /*
55  * The following routines are used in this implementation. They are
56  * written via macros to simulate zero-overhead call-by-reference.
57  *
58  * MUL64: 64x64->128-bit multiplication
59  * PMUL64: assumes top bits cleared on inputs
60  * ADD128: 128x128->128-bit addition
61  */
62 
63 #define ADD128(rh, rl, ih, il)						\
64 	do {								\
65 		u64 _il = (il);						\
66 		(rl) += (_il);						\
67 		if ((rl) < (_il))					\
68 			(rh)++;						\
69 		(rh) += (ih);						\
70 	} while (0)
71 
72 #define MUL32(i1, i2)	((u64)(u32)(i1)*(u32)(i2))
73 
74 #define PMUL64(rh, rl, i1, i2)	/* Assumes m doesn't overflow */	\
75 	do {								\
76 		u64 _i1 = (i1), _i2 = (i2);				\
77 		u64 m = MUL32(_i1, _i2>>32) + MUL32(_i1>>32, _i2);	\
78 		rh = MUL32(_i1>>32, _i2>>32);				\
79 		rl = MUL32(_i1, _i2);					\
80 		ADD128(rh, rl, (m >> 32), (m << 32));			\
81 	} while (0)
82 
83 #define MUL64(rh, rl, i1, i2)						\
84 	do {								\
85 		u64 _i1 = (i1), _i2 = (i2);				\
86 		u64 m1 = MUL32(_i1, _i2>>32);				\
87 		u64 m2 = MUL32(_i1>>32, _i2);				\
88 		rh = MUL32(_i1>>32, _i2>>32);				\
89 		rl = MUL32(_i1, _i2);					\
90 		ADD128(rh, rl, (m1 >> 32), (m1 << 32));			\
91 		ADD128(rh, rl, (m2 >> 32), (m2 << 32));			\
92 	} while (0)
93 
94 /*
95  * For highest performance the L1 NH and L2 polynomial hashes should be
96  * carefully implemented to take advantage of one's target architechture.
97  * Here these two hash functions are defined multiple time; once for
98  * 64-bit architectures, once for 32-bit SSE2 architectures, and once
99  * for the rest (32-bit) architectures.
100  * For each, nh_16 *must* be defined (works on multiples of 16 bytes).
101  * Optionally, nh_vmac_nhbytes can be defined (for multiples of
102  * VMAC_NHBYTES), and nh_16_2 and nh_vmac_nhbytes_2 (versions that do two
103  * NH computations at once).
104  */
105 
106 #ifdef CONFIG_64BIT
107 
108 #define nh_16(mp, kp, nw, rh, rl)					\
109 	do {								\
110 		int i; u64 th, tl;					\
111 		rh = rl = 0;						\
112 		for (i = 0; i < nw; i += 2) {				\
113 			MUL64(th, tl, le64_to_cpup((mp)+i)+(kp)[i],	\
114 				le64_to_cpup((mp)+i+1)+(kp)[i+1]);	\
115 			ADD128(rh, rl, th, tl);				\
116 		}							\
117 	} while (0)
118 
119 #define nh_16_2(mp, kp, nw, rh, rl, rh1, rl1)				\
120 	do {								\
121 		int i; u64 th, tl;					\
122 		rh1 = rl1 = rh = rl = 0;				\
123 		for (i = 0; i < nw; i += 2) {				\
124 			MUL64(th, tl, le64_to_cpup((mp)+i)+(kp)[i],	\
125 				le64_to_cpup((mp)+i+1)+(kp)[i+1]);	\
126 			ADD128(rh, rl, th, tl);				\
127 			MUL64(th, tl, le64_to_cpup((mp)+i)+(kp)[i+2],	\
128 				le64_to_cpup((mp)+i+1)+(kp)[i+3]);	\
129 			ADD128(rh1, rl1, th, tl);			\
130 		}							\
131 	} while (0)
132 
133 #if (VMAC_NHBYTES >= 64) /* These versions do 64-bytes of message at a time */
134 #define nh_vmac_nhbytes(mp, kp, nw, rh, rl)				\
135 	do {								\
136 		int i; u64 th, tl;					\
137 		rh = rl = 0;						\
138 		for (i = 0; i < nw; i += 8) {				\
139 			MUL64(th, tl, le64_to_cpup((mp)+i)+(kp)[i],	\
140 				le64_to_cpup((mp)+i+1)+(kp)[i+1]);	\
141 			ADD128(rh, rl, th, tl);				\
142 			MUL64(th, tl, le64_to_cpup((mp)+i+2)+(kp)[i+2],	\
143 				le64_to_cpup((mp)+i+3)+(kp)[i+3]);	\
144 			ADD128(rh, rl, th, tl);				\
145 			MUL64(th, tl, le64_to_cpup((mp)+i+4)+(kp)[i+4],	\
146 				le64_to_cpup((mp)+i+5)+(kp)[i+5]);	\
147 			ADD128(rh, rl, th, tl);				\
148 			MUL64(th, tl, le64_to_cpup((mp)+i+6)+(kp)[i+6],	\
149 				le64_to_cpup((mp)+i+7)+(kp)[i+7]);	\
150 			ADD128(rh, rl, th, tl);				\
151 		}							\
152 	} while (0)
153 
154 #define nh_vmac_nhbytes_2(mp, kp, nw, rh, rl, rh1, rl1)			\
155 	do {								\
156 		int i; u64 th, tl;					\
157 		rh1 = rl1 = rh = rl = 0;				\
158 		for (i = 0; i < nw; i += 8) {				\
159 			MUL64(th, tl, le64_to_cpup((mp)+i)+(kp)[i],	\
160 				le64_to_cpup((mp)+i+1)+(kp)[i+1]);	\
161 			ADD128(rh, rl, th, tl);				\
162 			MUL64(th, tl, le64_to_cpup((mp)+i)+(kp)[i+2],	\
163 				le64_to_cpup((mp)+i+1)+(kp)[i+3]);	\
164 			ADD128(rh1, rl1, th, tl);			\
165 			MUL64(th, tl, le64_to_cpup((mp)+i+2)+(kp)[i+2],	\
166 				le64_to_cpup((mp)+i+3)+(kp)[i+3]);	\
167 			ADD128(rh, rl, th, tl);				\
168 			MUL64(th, tl, le64_to_cpup((mp)+i+2)+(kp)[i+4],	\
169 				le64_to_cpup((mp)+i+3)+(kp)[i+5]);	\
170 			ADD128(rh1, rl1, th, tl);			\
171 			MUL64(th, tl, le64_to_cpup((mp)+i+4)+(kp)[i+4],	\
172 				le64_to_cpup((mp)+i+5)+(kp)[i+5]);	\
173 			ADD128(rh, rl, th, tl);				\
174 			MUL64(th, tl, le64_to_cpup((mp)+i+4)+(kp)[i+6],	\
175 				le64_to_cpup((mp)+i+5)+(kp)[i+7]);	\
176 			ADD128(rh1, rl1, th, tl);			\
177 			MUL64(th, tl, le64_to_cpup((mp)+i+6)+(kp)[i+6],	\
178 				le64_to_cpup((mp)+i+7)+(kp)[i+7]);	\
179 			ADD128(rh, rl, th, tl);				\
180 			MUL64(th, tl, le64_to_cpup((mp)+i+6)+(kp)[i+8],	\
181 				le64_to_cpup((mp)+i+7)+(kp)[i+9]);	\
182 			ADD128(rh1, rl1, th, tl);			\
183 		}							\
184 	} while (0)
185 #endif
186 
187 #define poly_step(ah, al, kh, kl, mh, ml)				\
188 	do {								\
189 		u64 t1h, t1l, t2h, t2l, t3h, t3l, z = 0;		\
190 		/* compute ab*cd, put bd into result registers */	\
191 		PMUL64(t3h, t3l, al, kh);				\
192 		PMUL64(t2h, t2l, ah, kl);				\
193 		PMUL64(t1h, t1l, ah, 2*kh);				\
194 		PMUL64(ah, al, al, kl);					\
195 		/* add 2 * ac to result */				\
196 		ADD128(ah, al, t1h, t1l);				\
197 		/* add together ad + bc */				\
198 		ADD128(t2h, t2l, t3h, t3l);				\
199 		/* now (ah,al), (t2l,2*t2h) need summing */		\
200 		/* first add the high registers, carrying into t2h */	\
201 		ADD128(t2h, ah, z, t2l);				\
202 		/* double t2h and add top bit of ah */			\
203 		t2h = 2 * t2h + (ah >> 63);				\
204 		ah &= m63;						\
205 		/* now add the low registers */				\
206 		ADD128(ah, al, mh, ml);					\
207 		ADD128(ah, al, z, t2h);					\
208 	} while (0)
209 
210 #else /* ! CONFIG_64BIT */
211 
212 #ifndef nh_16
213 #define nh_16(mp, kp, nw, rh, rl)					\
214 	do {								\
215 		u64 t1, t2, m1, m2, t;					\
216 		int i;							\
217 		rh = rl = t = 0;					\
218 		for (i = 0; i < nw; i += 2)  {				\
219 			t1 = le64_to_cpup(mp+i) + kp[i];		\
220 			t2 = le64_to_cpup(mp+i+1) + kp[i+1];		\
221 			m2 = MUL32(t1 >> 32, t2);			\
222 			m1 = MUL32(t1, t2 >> 32);			\
223 			ADD128(rh, rl, MUL32(t1 >> 32, t2 >> 32),	\
224 				MUL32(t1, t2));				\
225 			rh += (u64)(u32)(m1 >> 32)			\
226 				+ (u32)(m2 >> 32);			\
227 			t += (u64)(u32)m1 + (u32)m2;			\
228 		}							\
229 		ADD128(rh, rl, (t >> 32), (t << 32));			\
230 	} while (0)
231 #endif
232 
233 static void poly_step_func(u64 *ahi, u64 *alo,
234 			const u64 *kh, const u64 *kl,
235 			const u64 *mh, const u64 *ml)
236 {
237 #define a0 (*(((u32 *)alo)+INDEX_LOW))
238 #define a1 (*(((u32 *)alo)+INDEX_HIGH))
239 #define a2 (*(((u32 *)ahi)+INDEX_LOW))
240 #define a3 (*(((u32 *)ahi)+INDEX_HIGH))
241 #define k0 (*(((u32 *)kl)+INDEX_LOW))
242 #define k1 (*(((u32 *)kl)+INDEX_HIGH))
243 #define k2 (*(((u32 *)kh)+INDEX_LOW))
244 #define k3 (*(((u32 *)kh)+INDEX_HIGH))
245 
246 	u64 p, q, t;
247 	u32 t2;
248 
249 	p = MUL32(a3, k3);
250 	p += p;
251 	p += *(u64 *)mh;
252 	p += MUL32(a0, k2);
253 	p += MUL32(a1, k1);
254 	p += MUL32(a2, k0);
255 	t = (u32)(p);
256 	p >>= 32;
257 	p += MUL32(a0, k3);
258 	p += MUL32(a1, k2);
259 	p += MUL32(a2, k1);
260 	p += MUL32(a3, k0);
261 	t |= ((u64)((u32)p & 0x7fffffff)) << 32;
262 	p >>= 31;
263 	p += (u64)(((u32 *)ml)[INDEX_LOW]);
264 	p += MUL32(a0, k0);
265 	q =  MUL32(a1, k3);
266 	q += MUL32(a2, k2);
267 	q += MUL32(a3, k1);
268 	q += q;
269 	p += q;
270 	t2 = (u32)(p);
271 	p >>= 32;
272 	p += (u64)(((u32 *)ml)[INDEX_HIGH]);
273 	p += MUL32(a0, k1);
274 	p += MUL32(a1, k0);
275 	q =  MUL32(a2, k3);
276 	q += MUL32(a3, k2);
277 	q += q;
278 	p += q;
279 	*(u64 *)(alo) = (p << 32) | t2;
280 	p >>= 32;
281 	*(u64 *)(ahi) = p + t;
282 
283 #undef a0
284 #undef a1
285 #undef a2
286 #undef a3
287 #undef k0
288 #undef k1
289 #undef k2
290 #undef k3
291 }
292 
293 #define poly_step(ah, al, kh, kl, mh, ml)				\
294 	poly_step_func(&(ah), &(al), &(kh), &(kl), &(mh), &(ml))
295 
296 #endif  /* end of specialized NH and poly definitions */
297 
298 /* At least nh_16 is defined. Defined others as needed here */
299 #ifndef nh_16_2
300 #define nh_16_2(mp, kp, nw, rh, rl, rh2, rl2)				\
301 	do { 								\
302 		nh_16(mp, kp, nw, rh, rl);				\
303 		nh_16(mp, ((kp)+2), nw, rh2, rl2);			\
304 	} while (0)
305 #endif
306 #ifndef nh_vmac_nhbytes
307 #define nh_vmac_nhbytes(mp, kp, nw, rh, rl)				\
308 	nh_16(mp, kp, nw, rh, rl)
309 #endif
310 #ifndef nh_vmac_nhbytes_2
311 #define nh_vmac_nhbytes_2(mp, kp, nw, rh, rl, rh2, rl2)			\
312 	do {								\
313 		nh_vmac_nhbytes(mp, kp, nw, rh, rl);			\
314 		nh_vmac_nhbytes(mp, ((kp)+2), nw, rh2, rl2);		\
315 	} while (0)
316 #endif
317 
318 static void vhash_abort(struct vmac_ctx *ctx)
319 {
320 	ctx->polytmp[0] = ctx->polykey[0] ;
321 	ctx->polytmp[1] = ctx->polykey[1] ;
322 	ctx->first_block_processed = 0;
323 }
324 
325 static u64 l3hash(u64 p1, u64 p2,
326 			u64 k1, u64 k2, u64 len)
327 {
328 	u64 rh, rl, t, z = 0;
329 
330 	/* fully reduce (p1,p2)+(len,0) mod p127 */
331 	t = p1 >> 63;
332 	p1 &= m63;
333 	ADD128(p1, p2, len, t);
334 	/* At this point, (p1,p2) is at most 2^127+(len<<64) */
335 	t = (p1 > m63) + ((p1 == m63) && (p2 == m64));
336 	ADD128(p1, p2, z, t);
337 	p1 &= m63;
338 
339 	/* compute (p1,p2)/(2^64-2^32) and (p1,p2)%(2^64-2^32) */
340 	t = p1 + (p2 >> 32);
341 	t += (t >> 32);
342 	t += (u32)t > 0xfffffffeu;
343 	p1 += (t >> 32);
344 	p2 += (p1 << 32);
345 
346 	/* compute (p1+k1)%p64 and (p2+k2)%p64 */
347 	p1 += k1;
348 	p1 += (0 - (p1 < k1)) & 257;
349 	p2 += k2;
350 	p2 += (0 - (p2 < k2)) & 257;
351 
352 	/* compute (p1+k1)*(p2+k2)%p64 */
353 	MUL64(rh, rl, p1, p2);
354 	t = rh >> 56;
355 	ADD128(t, rl, z, rh);
356 	rh <<= 8;
357 	ADD128(t, rl, z, rh);
358 	t += t << 8;
359 	rl += t;
360 	rl += (0 - (rl < t)) & 257;
361 	rl += (0 - (rl > p64-1)) & 257;
362 	return rl;
363 }
364 
365 static void vhash_update(const unsigned char *m,
366 			unsigned int mbytes, /* Pos multiple of VMAC_NHBYTES */
367 			struct vmac_ctx *ctx)
368 {
369 	u64 rh, rl, *mptr;
370 	const u64 *kptr = (u64 *)ctx->nhkey;
371 	int i;
372 	u64 ch, cl;
373 	u64 pkh = ctx->polykey[0];
374 	u64 pkl = ctx->polykey[1];
375 
376 	mptr = (u64 *)m;
377 	i = mbytes / VMAC_NHBYTES;  /* Must be non-zero */
378 
379 	ch = ctx->polytmp[0];
380 	cl = ctx->polytmp[1];
381 
382 	if (!ctx->first_block_processed) {
383 		ctx->first_block_processed = 1;
384 		nh_vmac_nhbytes(mptr, kptr, VMAC_NHBYTES/8, rh, rl);
385 		rh &= m62;
386 		ADD128(ch, cl, rh, rl);
387 		mptr += (VMAC_NHBYTES/sizeof(u64));
388 		i--;
389 	}
390 
391 	while (i--) {
392 		nh_vmac_nhbytes(mptr, kptr, VMAC_NHBYTES/8, rh, rl);
393 		rh &= m62;
394 		poly_step(ch, cl, pkh, pkl, rh, rl);
395 		mptr += (VMAC_NHBYTES/sizeof(u64));
396 	}
397 
398 	ctx->polytmp[0] = ch;
399 	ctx->polytmp[1] = cl;
400 }
401 
402 static u64 vhash(unsigned char m[], unsigned int mbytes,
403 			u64 *tagl, struct vmac_ctx *ctx)
404 {
405 	u64 rh, rl, *mptr;
406 	const u64 *kptr = (u64 *)ctx->nhkey;
407 	int i, remaining;
408 	u64 ch, cl;
409 	u64 pkh = ctx->polykey[0];
410 	u64 pkl = ctx->polykey[1];
411 
412 	mptr = (u64 *)m;
413 	i = mbytes / VMAC_NHBYTES;
414 	remaining = mbytes % VMAC_NHBYTES;
415 
416 	if (ctx->first_block_processed) {
417 		ch = ctx->polytmp[0];
418 		cl = ctx->polytmp[1];
419 	} else if (i) {
420 		nh_vmac_nhbytes(mptr, kptr, VMAC_NHBYTES/8, ch, cl);
421 		ch &= m62;
422 		ADD128(ch, cl, pkh, pkl);
423 		mptr += (VMAC_NHBYTES/sizeof(u64));
424 		i--;
425 	} else if (remaining) {
426 		nh_16(mptr, kptr, 2*((remaining+15)/16), ch, cl);
427 		ch &= m62;
428 		ADD128(ch, cl, pkh, pkl);
429 		mptr += (VMAC_NHBYTES/sizeof(u64));
430 		goto do_l3;
431 	} else {/* Empty String */
432 		ch = pkh; cl = pkl;
433 		goto do_l3;
434 	}
435 
436 	while (i--) {
437 		nh_vmac_nhbytes(mptr, kptr, VMAC_NHBYTES/8, rh, rl);
438 		rh &= m62;
439 		poly_step(ch, cl, pkh, pkl, rh, rl);
440 		mptr += (VMAC_NHBYTES/sizeof(u64));
441 	}
442 	if (remaining) {
443 		nh_16(mptr, kptr, 2*((remaining+15)/16), rh, rl);
444 		rh &= m62;
445 		poly_step(ch, cl, pkh, pkl, rh, rl);
446 	}
447 
448 do_l3:
449 	vhash_abort(ctx);
450 	remaining *= 8;
451 	return l3hash(ch, cl, ctx->l3key[0], ctx->l3key[1], remaining);
452 }
453 
454 static u64 vmac(unsigned char m[], unsigned int mbytes,
455 			unsigned char n[16], u64 *tagl,
456 			struct vmac_ctx_t *ctx)
457 {
458 	u64 *in_n, *out_p;
459 	u64 p, h;
460 	int i;
461 
462 	in_n = ctx->__vmac_ctx.cached_nonce;
463 	out_p = ctx->__vmac_ctx.cached_aes;
464 
465 	i = n[15] & 1;
466 	if ((*(u64 *)(n+8) != in_n[1]) || (*(u64 *)(n) != in_n[0])) {
467 		in_n[0] = *(u64 *)(n);
468 		in_n[1] = *(u64 *)(n+8);
469 		((unsigned char *)in_n)[15] &= 0xFE;
470 		crypto_cipher_encrypt_one(ctx->child,
471 			(unsigned char *)out_p, (unsigned char *)in_n);
472 
473 		((unsigned char *)in_n)[15] |= (unsigned char)(1-i);
474 	}
475 	p = be64_to_cpup(out_p + i);
476 	h = vhash(m, mbytes, (u64 *)0, &ctx->__vmac_ctx);
477 	return p + h;
478 }
479 
480 static int vmac_set_key(unsigned char user_key[], struct vmac_ctx_t *ctx)
481 {
482 	u64 in[2] = {0}, out[2];
483 	unsigned i;
484 	int err = 0;
485 
486 	err = crypto_cipher_setkey(ctx->child, user_key, VMAC_KEY_LEN);
487 	if (err)
488 		return err;
489 
490 	/* Fill nh key */
491 	((unsigned char *)in)[0] = 0x80;
492 	for (i = 0; i < sizeof(ctx->__vmac_ctx.nhkey)/8; i += 2) {
493 		crypto_cipher_encrypt_one(ctx->child,
494 			(unsigned char *)out, (unsigned char *)in);
495 		ctx->__vmac_ctx.nhkey[i] = be64_to_cpup(out);
496 		ctx->__vmac_ctx.nhkey[i+1] = be64_to_cpup(out+1);
497 		((unsigned char *)in)[15] += 1;
498 	}
499 
500 	/* Fill poly key */
501 	((unsigned char *)in)[0] = 0xC0;
502 	in[1] = 0;
503 	for (i = 0; i < sizeof(ctx->__vmac_ctx.polykey)/8; i += 2) {
504 		crypto_cipher_encrypt_one(ctx->child,
505 			(unsigned char *)out, (unsigned char *)in);
506 		ctx->__vmac_ctx.polytmp[i] =
507 			ctx->__vmac_ctx.polykey[i] =
508 				be64_to_cpup(out) & mpoly;
509 		ctx->__vmac_ctx.polytmp[i+1] =
510 			ctx->__vmac_ctx.polykey[i+1] =
511 				be64_to_cpup(out+1) & mpoly;
512 		((unsigned char *)in)[15] += 1;
513 	}
514 
515 	/* Fill ip key */
516 	((unsigned char *)in)[0] = 0xE0;
517 	in[1] = 0;
518 	for (i = 0; i < sizeof(ctx->__vmac_ctx.l3key)/8; i += 2) {
519 		do {
520 			crypto_cipher_encrypt_one(ctx->child,
521 				(unsigned char *)out, (unsigned char *)in);
522 			ctx->__vmac_ctx.l3key[i] = be64_to_cpup(out);
523 			ctx->__vmac_ctx.l3key[i+1] = be64_to_cpup(out+1);
524 			((unsigned char *)in)[15] += 1;
525 		} while (ctx->__vmac_ctx.l3key[i] >= p64
526 			|| ctx->__vmac_ctx.l3key[i+1] >= p64);
527 	}
528 
529 	/* Invalidate nonce/aes cache and reset other elements */
530 	ctx->__vmac_ctx.cached_nonce[0] = (u64)-1; /* Ensure illegal nonce */
531 	ctx->__vmac_ctx.cached_nonce[1] = (u64)0;  /* Ensure illegal nonce */
532 	ctx->__vmac_ctx.first_block_processed = 0;
533 
534 	return err;
535 }
536 
537 static int vmac_setkey(struct crypto_shash *parent,
538 		const u8 *key, unsigned int keylen)
539 {
540 	struct vmac_ctx_t *ctx = crypto_shash_ctx(parent);
541 
542 	if (keylen != VMAC_KEY_LEN) {
543 		crypto_shash_set_flags(parent, CRYPTO_TFM_RES_BAD_KEY_LEN);
544 		return -EINVAL;
545 	}
546 
547 	return vmac_set_key((u8 *)key, ctx);
548 }
549 
550 static int vmac_init(struct shash_desc *pdesc)
551 {
552 	struct crypto_shash *parent = pdesc->tfm;
553 	struct vmac_ctx_t *ctx = crypto_shash_ctx(parent);
554 
555 	memset(&ctx->__vmac_ctx, 0, sizeof(struct vmac_ctx));
556 	return 0;
557 }
558 
559 static int vmac_update(struct shash_desc *pdesc, const u8 *p,
560 		unsigned int len)
561 {
562 	struct crypto_shash *parent = pdesc->tfm;
563 	struct vmac_ctx_t *ctx = crypto_shash_ctx(parent);
564 
565 	vhash_update(p, len, &ctx->__vmac_ctx);
566 
567 	return 0;
568 }
569 
570 static int vmac_final(struct shash_desc *pdesc, u8 *out)
571 {
572 	struct crypto_shash *parent = pdesc->tfm;
573 	struct vmac_ctx_t *ctx = crypto_shash_ctx(parent);
574 	vmac_t mac;
575 	u8 nonce[16] = {};
576 
577 	mac = vmac(NULL, 0, nonce, NULL, ctx);
578 	memcpy(out, &mac, sizeof(vmac_t));
579 	memset(&mac, 0, sizeof(vmac_t));
580 	memset(&ctx->__vmac_ctx, 0, sizeof(struct vmac_ctx));
581 	return 0;
582 }
583 
584 static int vmac_init_tfm(struct crypto_tfm *tfm)
585 {
586 	struct crypto_cipher *cipher;
587 	struct crypto_instance *inst = (void *)tfm->__crt_alg;
588 	struct crypto_spawn *spawn = crypto_instance_ctx(inst);
589 	struct vmac_ctx_t *ctx = crypto_tfm_ctx(tfm);
590 
591 	cipher = crypto_spawn_cipher(spawn);
592 	if (IS_ERR(cipher))
593 		return PTR_ERR(cipher);
594 
595 	ctx->child = cipher;
596 	return 0;
597 }
598 
599 static void vmac_exit_tfm(struct crypto_tfm *tfm)
600 {
601 	struct vmac_ctx_t *ctx = crypto_tfm_ctx(tfm);
602 	crypto_free_cipher(ctx->child);
603 }
604 
605 static int vmac_create(struct crypto_template *tmpl, struct rtattr **tb)
606 {
607 	struct shash_instance *inst;
608 	struct crypto_alg *alg;
609 	int err;
610 
611 	err = crypto_check_attr_type(tb, CRYPTO_ALG_TYPE_SHASH);
612 	if (err)
613 		return err;
614 
615 	alg = crypto_get_attr_alg(tb, CRYPTO_ALG_TYPE_CIPHER,
616 			CRYPTO_ALG_TYPE_MASK);
617 	if (IS_ERR(alg))
618 		return PTR_ERR(alg);
619 
620 	inst = shash_alloc_instance("vmac", alg);
621 	err = PTR_ERR(inst);
622 	if (IS_ERR(inst))
623 		goto out_put_alg;
624 
625 	err = crypto_init_spawn(shash_instance_ctx(inst), alg,
626 			shash_crypto_instance(inst),
627 			CRYPTO_ALG_TYPE_MASK);
628 	if (err)
629 		goto out_free_inst;
630 
631 	inst->alg.base.cra_priority = alg->cra_priority;
632 	inst->alg.base.cra_blocksize = alg->cra_blocksize;
633 	inst->alg.base.cra_alignmask = alg->cra_alignmask;
634 
635 	inst->alg.digestsize = sizeof(vmac_t);
636 	inst->alg.base.cra_ctxsize = sizeof(struct vmac_ctx_t);
637 	inst->alg.base.cra_init = vmac_init_tfm;
638 	inst->alg.base.cra_exit = vmac_exit_tfm;
639 
640 	inst->alg.init = vmac_init;
641 	inst->alg.update = vmac_update;
642 	inst->alg.final = vmac_final;
643 	inst->alg.setkey = vmac_setkey;
644 
645 	err = shash_register_instance(tmpl, inst);
646 	if (err) {
647 out_free_inst:
648 		shash_free_instance(shash_crypto_instance(inst));
649 	}
650 
651 out_put_alg:
652 	crypto_mod_put(alg);
653 	return err;
654 }
655 
656 static struct crypto_template vmac_tmpl = {
657 	.name = "vmac",
658 	.create = vmac_create,
659 	.free = shash_free_instance,
660 	.module = THIS_MODULE,
661 };
662 
663 static int __init vmac_module_init(void)
664 {
665 	return crypto_register_template(&vmac_tmpl);
666 }
667 
668 static void __exit vmac_module_exit(void)
669 {
670 	crypto_unregister_template(&vmac_tmpl);
671 }
672 
673 module_init(vmac_module_init);
674 module_exit(vmac_module_exit);
675 
676 MODULE_LICENSE("GPL");
677 MODULE_DESCRIPTION("VMAC hash algorithm");
678 
679