1// SPDX-License-Identifier: GPL-2.0+
2/*
3 * FIPS-180-2 compliant SHA-512 and SHA-384 implementation
4 *
5 * SHA-512 code by Jean-Luc Cooke <jlcooke@certainkey.com>
6 *
7 * Copyright (c) Jean-Luc Cooke <jlcooke@certainkey.com>
8 * Copyright (c) Andrew McDonald <andrew@mcdonald.org.uk>
9 * Copyright (c) 2003 Kyle McMartin <kyle@debian.org>
10 * Copyright (c) 2020 Reuben Dowle <reuben.dowle@4rf.com>
11 */
12
13#ifndef USE_HOSTCC
14#include <cyclic.h>
15#endif /* USE_HOSTCC */
16#include <compiler.h>
17#include <u-boot/sha512.h>
18
19const uint8_t sha384_der_prefix[SHA384_DER_LEN] = {
20	0x30, 0x41, 0x30, 0x0d, 0x06, 0x09, 0x60, 0x86,
21	0x48, 0x01, 0x65, 0x03, 0x04, 0x02, 0x02, 0x05,
22	0x00, 0x04, 0x30
23};
24
25const uint8_t sha512_der_prefix[SHA512_DER_LEN] = {
26	0x30, 0x51, 0x30, 0x0d, 0x06, 0x09, 0x60, 0x86,
27	0x48, 0x01, 0x65, 0x03, 0x04, 0x02, 0x03, 0x05,
28	0x00, 0x04, 0x40
29};
30
31#define SHA384_H0	0xcbbb9d5dc1059ed8ULL
32#define SHA384_H1	0x629a292a367cd507ULL
33#define SHA384_H2	0x9159015a3070dd17ULL
34#define SHA384_H3	0x152fecd8f70e5939ULL
35#define SHA384_H4	0x67332667ffc00b31ULL
36#define SHA384_H5	0x8eb44a8768581511ULL
37#define SHA384_H6	0xdb0c2e0d64f98fa7ULL
38#define SHA384_H7	0x47b5481dbefa4fa4ULL
39
40#define SHA512_H0	0x6a09e667f3bcc908ULL
41#define SHA512_H1	0xbb67ae8584caa73bULL
42#define SHA512_H2	0x3c6ef372fe94f82bULL
43#define SHA512_H3	0xa54ff53a5f1d36f1ULL
44#define SHA512_H4	0x510e527fade682d1ULL
45#define SHA512_H5	0x9b05688c2b3e6c1fULL
46#define SHA512_H6	0x1f83d9abfb41bd6bULL
47#define SHA512_H7	0x5be0cd19137e2179ULL
48
49static inline uint64_t Ch(uint64_t x, uint64_t y, uint64_t z)
50{
51        return z ^ (x & (y ^ z));
52}
53
54static inline uint64_t Maj(uint64_t x, uint64_t y, uint64_t z)
55{
56        return (x & y) | (z & (x | y));
57}
58
59static const uint64_t sha512_K[80] = {
60        0x428a2f98d728ae22ULL, 0x7137449123ef65cdULL, 0xb5c0fbcfec4d3b2fULL,
61        0xe9b5dba58189dbbcULL, 0x3956c25bf348b538ULL, 0x59f111f1b605d019ULL,
62        0x923f82a4af194f9bULL, 0xab1c5ed5da6d8118ULL, 0xd807aa98a3030242ULL,
63        0x12835b0145706fbeULL, 0x243185be4ee4b28cULL, 0x550c7dc3d5ffb4e2ULL,
64        0x72be5d74f27b896fULL, 0x80deb1fe3b1696b1ULL, 0x9bdc06a725c71235ULL,
65        0xc19bf174cf692694ULL, 0xe49b69c19ef14ad2ULL, 0xefbe4786384f25e3ULL,
66        0x0fc19dc68b8cd5b5ULL, 0x240ca1cc77ac9c65ULL, 0x2de92c6f592b0275ULL,
67        0x4a7484aa6ea6e483ULL, 0x5cb0a9dcbd41fbd4ULL, 0x76f988da831153b5ULL,
68        0x983e5152ee66dfabULL, 0xa831c66d2db43210ULL, 0xb00327c898fb213fULL,
69        0xbf597fc7beef0ee4ULL, 0xc6e00bf33da88fc2ULL, 0xd5a79147930aa725ULL,
70        0x06ca6351e003826fULL, 0x142929670a0e6e70ULL, 0x27b70a8546d22ffcULL,
71        0x2e1b21385c26c926ULL, 0x4d2c6dfc5ac42aedULL, 0x53380d139d95b3dfULL,
72        0x650a73548baf63deULL, 0x766a0abb3c77b2a8ULL, 0x81c2c92e47edaee6ULL,
73        0x92722c851482353bULL, 0xa2bfe8a14cf10364ULL, 0xa81a664bbc423001ULL,
74        0xc24b8b70d0f89791ULL, 0xc76c51a30654be30ULL, 0xd192e819d6ef5218ULL,
75        0xd69906245565a910ULL, 0xf40e35855771202aULL, 0x106aa07032bbd1b8ULL,
76        0x19a4c116b8d2d0c8ULL, 0x1e376c085141ab53ULL, 0x2748774cdf8eeb99ULL,
77        0x34b0bcb5e19b48a8ULL, 0x391c0cb3c5c95a63ULL, 0x4ed8aa4ae3418acbULL,
78        0x5b9cca4f7763e373ULL, 0x682e6ff3d6b2b8a3ULL, 0x748f82ee5defb2fcULL,
79        0x78a5636f43172f60ULL, 0x84c87814a1f0ab72ULL, 0x8cc702081a6439ecULL,
80        0x90befffa23631e28ULL, 0xa4506cebde82bde9ULL, 0xbef9a3f7b2c67915ULL,
81        0xc67178f2e372532bULL, 0xca273eceea26619cULL, 0xd186b8c721c0c207ULL,
82        0xeada7dd6cde0eb1eULL, 0xf57d4f7fee6ed178ULL, 0x06f067aa72176fbaULL,
83        0x0a637dc5a2c898a6ULL, 0x113f9804bef90daeULL, 0x1b710b35131c471bULL,
84        0x28db77f523047d84ULL, 0x32caab7b40c72493ULL, 0x3c9ebe0a15c9bebcULL,
85        0x431d67c49c100d4cULL, 0x4cc5d4becb3e42b6ULL, 0x597f299cfc657e2aULL,
86        0x5fcb6fab3ad6faecULL, 0x6c44198c4a475817ULL,
87};
88
89static inline uint64_t ror64(uint64_t word, unsigned int shift)
90{
91	return (word >> (shift & 63)) | (word << ((-shift) & 63));
92}
93
94#define e0(x)       (ror64(x,28) ^ ror64(x,34) ^ ror64(x,39))
95#define e1(x)       (ror64(x,14) ^ ror64(x,18) ^ ror64(x,41))
96#define s0(x)       (ror64(x, 1) ^ ror64(x, 8) ^ (x >> 7))
97#define s1(x)       (ror64(x,19) ^ ror64(x,61) ^ (x >> 6))
98
99/*
100 * 64-bit integer manipulation macros (big endian)
101 */
102#ifndef GET_UINT64_BE
103#define GET_UINT64_BE(n,b,i) {				\
104	(n) = ( (unsigned long long) (b)[(i)    ] << 56 )	\
105	    | ( (unsigned long long) (b)[(i) + 1] << 48 )	\
106	    | ( (unsigned long long) (b)[(i) + 2] << 40 )	\
107	    | ( (unsigned long long) (b)[(i) + 3] << 32 )	\
108	    | ( (unsigned long long) (b)[(i) + 4] << 24 )	\
109	    | ( (unsigned long long) (b)[(i) + 5] << 16 )	\
110	    | ( (unsigned long long) (b)[(i) + 6] <<  8 )	\
111	    | ( (unsigned long long) (b)[(i) + 7]       );	\
112}
113#endif
114#ifndef PUT_UINT64_BE
115#define PUT_UINT64_BE(n,b,i) {				\
116	(b)[(i)    ] = (unsigned char) ( (n) >> 56 );	\
117	(b)[(i) + 1] = (unsigned char) ( (n) >> 48 );	\
118	(b)[(i) + 2] = (unsigned char) ( (n) >> 40 );	\
119	(b)[(i) + 3] = (unsigned char) ( (n) >> 32 );	\
120	(b)[(i) + 4] = (unsigned char) ( (n) >> 24 );	\
121	(b)[(i) + 5] = (unsigned char) ( (n) >> 16 );	\
122	(b)[(i) + 6] = (unsigned char) ( (n) >>  8 );	\
123	(b)[(i) + 7] = (unsigned char) ( (n)       );	\
124}
125#endif
126
127static inline void LOAD_OP(int I, uint64_t *W, const uint8_t *input)
128{
129	GET_UINT64_BE(W[I], input, I*8);
130}
131
132static inline void BLEND_OP(int I, uint64_t *W)
133{
134	W[I & 15] += s1(W[(I-2) & 15]) + W[(I-7) & 15] + s0(W[(I-15) & 15]);
135}
136
137static void
138sha512_transform(uint64_t *state, const uint8_t *input)
139{
140	uint64_t a, b, c, d, e, f, g, h, t1, t2;
141
142	int i;
143	uint64_t W[16];
144
145	/* load the state into our registers */
146	a=state[0];   b=state[1];   c=state[2];   d=state[3];
147	e=state[4];   f=state[5];   g=state[6];   h=state[7];
148
149	/* now iterate */
150	for (i=0; i<80; i+=8) {
151		if (!(i & 8)) {
152			int j;
153
154			if (i < 16) {
155				/* load the input */
156				for (j = 0; j < 16; j++)
157					LOAD_OP(i + j, W, input);
158			} else {
159				for (j = 0; j < 16; j++) {
160					BLEND_OP(i + j, W);
161				}
162			}
163		}
164
165		t1 = h + e1(e) + Ch(e,f,g) + sha512_K[i  ] + W[(i & 15)];
166		t2 = e0(a) + Maj(a,b,c);    d+=t1;    h=t1+t2;
167		t1 = g + e1(d) + Ch(d,e,f) + sha512_K[i+1] + W[(i & 15) + 1];
168		t2 = e0(h) + Maj(h,a,b);    c+=t1;    g=t1+t2;
169		t1 = f + e1(c) + Ch(c,d,e) + sha512_K[i+2] + W[(i & 15) + 2];
170		t2 = e0(g) + Maj(g,h,a);    b+=t1;    f=t1+t2;
171		t1 = e + e1(b) + Ch(b,c,d) + sha512_K[i+3] + W[(i & 15) + 3];
172		t2 = e0(f) + Maj(f,g,h);    a+=t1;    e=t1+t2;
173		t1 = d + e1(a) + Ch(a,b,c) + sha512_K[i+4] + W[(i & 15) + 4];
174		t2 = e0(e) + Maj(e,f,g);    h+=t1;    d=t1+t2;
175		t1 = c + e1(h) + Ch(h,a,b) + sha512_K[i+5] + W[(i & 15) + 5];
176		t2 = e0(d) + Maj(d,e,f);    g+=t1;    c=t1+t2;
177		t1 = b + e1(g) + Ch(g,h,a) + sha512_K[i+6] + W[(i & 15) + 6];
178		t2 = e0(c) + Maj(c,d,e);    f+=t1;    b=t1+t2;
179		t1 = a + e1(f) + Ch(f,g,h) + sha512_K[i+7] + W[(i & 15) + 7];
180		t2 = e0(b) + Maj(b,c,d);    e+=t1;    a=t1+t2;
181	}
182
183	state[0] += a; state[1] += b; state[2] += c; state[3] += d;
184	state[4] += e; state[5] += f; state[6] += g; state[7] += h;
185
186	/* erase our data */
187	a = b = c = d = e = f = g = h = t1 = t2 = 0;
188}
189
190static void sha512_block_fn(sha512_context *sst, const uint8_t *src,
191				    int blocks)
192{
193	while (blocks--) {
194		sha512_transform(sst->state, src);
195		src += SHA512_BLOCK_SIZE;
196	}
197}
198
199static void sha512_base_do_update(sha512_context *sctx,
200					const uint8_t *data,
201					unsigned int len)
202{
203	unsigned int partial = sctx->count[0] % SHA512_BLOCK_SIZE;
204
205	sctx->count[0] += len;
206	if (sctx->count[0] < len)
207		sctx->count[1]++;
208
209	if (unlikely((partial + len) >= SHA512_BLOCK_SIZE)) {
210		int blocks;
211
212		if (partial) {
213			int p = SHA512_BLOCK_SIZE - partial;
214
215			memcpy(sctx->buf + partial, data, p);
216			data += p;
217			len -= p;
218
219			sha512_block_fn(sctx, sctx->buf, 1);
220		}
221
222		blocks = len / SHA512_BLOCK_SIZE;
223		len %= SHA512_BLOCK_SIZE;
224
225		if (blocks) {
226			sha512_block_fn(sctx, data, blocks);
227			data += blocks * SHA512_BLOCK_SIZE;
228		}
229		partial = 0;
230	}
231	if (len)
232		memcpy(sctx->buf + partial, data, len);
233}
234
235static void sha512_base_do_finalize(sha512_context *sctx)
236{
237	const int bit_offset = SHA512_BLOCK_SIZE - sizeof(uint64_t[2]);
238	uint64_t *bits = (uint64_t *)(sctx->buf + bit_offset);
239	unsigned int partial = sctx->count[0] % SHA512_BLOCK_SIZE;
240
241	sctx->buf[partial++] = 0x80;
242	if (partial > bit_offset) {
243		memset(sctx->buf + partial, 0x0, SHA512_BLOCK_SIZE - partial);
244		partial = 0;
245
246		sha512_block_fn(sctx, sctx->buf, 1);
247	}
248
249	memset(sctx->buf + partial, 0x0, bit_offset - partial);
250	bits[0] = cpu_to_be64(sctx->count[1] << 3 | sctx->count[0] >> 61);
251	bits[1] = cpu_to_be64(sctx->count[0] << 3);
252	sha512_block_fn(sctx, sctx->buf, 1);
253}
254
255#if defined(CONFIG_SHA384)
256void sha384_starts(sha512_context * ctx)
257{
258	ctx->state[0] = SHA384_H0;
259	ctx->state[1] = SHA384_H1;
260	ctx->state[2] = SHA384_H2;
261	ctx->state[3] = SHA384_H3;
262	ctx->state[4] = SHA384_H4;
263	ctx->state[5] = SHA384_H5;
264	ctx->state[6] = SHA384_H6;
265	ctx->state[7] = SHA384_H7;
266	ctx->count[0] = ctx->count[1] = 0;
267}
268
269void sha384_update(sha512_context *ctx, const uint8_t *input, uint32_t length)
270{
271	sha512_base_do_update(ctx, input, length);
272}
273
274void sha384_finish(sha512_context * ctx, uint8_t digest[SHA384_SUM_LEN])
275{
276	int i;
277
278	sha512_base_do_finalize(ctx);
279	for(i=0; i<SHA384_SUM_LEN / sizeof(uint64_t); i++)
280		PUT_UINT64_BE(ctx->state[i], digest, i * 8);
281}
282
283/*
284 * Output = SHA-512( input buffer ). Trigger the watchdog every 'chunk_sz'
285 * bytes of input processed.
286 */
287void sha384_csum_wd(const unsigned char *input, unsigned int ilen,
288		unsigned char *output, unsigned int chunk_sz)
289{
290	sha512_context ctx;
291#if defined(CONFIG_HW_WATCHDOG) || defined(CONFIG_WATCHDOG)
292	const unsigned char *end;
293	unsigned char *curr;
294	int chunk;
295#endif
296
297	sha384_starts(&ctx);
298
299#if defined(CONFIG_HW_WATCHDOG) || defined(CONFIG_WATCHDOG)
300	curr = (unsigned char *)input;
301	end = input + ilen;
302	while (curr < end) {
303		chunk = end - curr;
304		if (chunk > chunk_sz)
305			chunk = chunk_sz;
306		sha384_update(&ctx, curr, chunk);
307		curr += chunk;
308		schedule();
309	}
310#else
311	sha384_update(&ctx, input, ilen);
312#endif
313
314	sha384_finish(&ctx, output);
315}
316
317#endif
318
319void sha512_starts(sha512_context * ctx)
320{
321	ctx->state[0] = SHA512_H0;
322	ctx->state[1] = SHA512_H1;
323	ctx->state[2] = SHA512_H2;
324	ctx->state[3] = SHA512_H3;
325	ctx->state[4] = SHA512_H4;
326	ctx->state[5] = SHA512_H5;
327	ctx->state[6] = SHA512_H6;
328	ctx->state[7] = SHA512_H7;
329	ctx->count[0] = ctx->count[1] = 0;
330}
331
332void sha512_update(sha512_context *ctx, const uint8_t *input, uint32_t length)
333{
334	sha512_base_do_update(ctx, input, length);
335}
336
337void sha512_finish(sha512_context * ctx, uint8_t digest[SHA512_SUM_LEN])
338{
339	int i;
340
341	sha512_base_do_finalize(ctx);
342	for(i=0; i<SHA512_SUM_LEN / sizeof(uint64_t); i++)
343		PUT_UINT64_BE(ctx->state[i], digest, i * 8);
344}
345
346/*
347 * Output = SHA-512( input buffer ). Trigger the watchdog every 'chunk_sz'
348 * bytes of input processed.
349 */
350void sha512_csum_wd(const unsigned char *input, unsigned int ilen,
351		unsigned char *output, unsigned int chunk_sz)
352{
353	sha512_context ctx;
354#if defined(CONFIG_HW_WATCHDOG) || defined(CONFIG_WATCHDOG)
355	const unsigned char *end;
356	unsigned char *curr;
357	int chunk;
358#endif
359
360	sha512_starts(&ctx);
361
362#if defined(CONFIG_HW_WATCHDOG) || defined(CONFIG_WATCHDOG)
363	curr = (unsigned char *)input;
364	end = input + ilen;
365	while (curr < end) {
366		chunk = end - curr;
367		if (chunk > chunk_sz)
368			chunk = chunk_sz;
369		sha512_update(&ctx, curr, chunk);
370		curr += chunk;
371		schedule();
372	}
373#else
374	sha512_update(&ctx, input, ilen);
375#endif
376
377	sha512_finish(&ctx, output);
378}
379