1 /* gf128mul.c - GF(2^128) multiplication functions
17  ---------------------------------------------------------------------------
44  ---------------------------------------------------------------------------
95  * 16-bit value that must be XOR-ed into the low-degree end of the
99  * the "be" convention where the highest-order bit is the coefficient of
100  * the highest-degree polynomial term, and one for the "le" convention
101  * where the highest-order bit is the coefficient of the lowest-degree
130 static const u16 gf128mul_table_le[256] = gf128mul_dat(xda_le);
131 static const u16 gf128mul_table_be[256] = gf128mul_dat(xda_be);
135  * the polynomial field representation.  They use 64-bit word operations
142 	u64 a = be64_to_cpu(x->a);  in gf128mul_x8_lle()
143 	u64 b = be64_to_cpu(x->b);  in gf128mul_x8_lle()
146 	x->b = cpu_to_be64((b >> 8) | (a << 56));  in gf128mul_x8_lle()
147 	x->a = cpu_to_be64((a >> 8) ^ (_tt << 48));  in gf128mul_x8_lle()
153 	u64 a = be64_to_cpu(x->a);  in gf128mul_x8_lle_ti()
154 	u64 b = be64_to_cpu(x->b);  in gf128mul_x8_lle_ti()
157 	x->b = cpu_to_be64((b >> 8) | (a << 56));  in gf128mul_x8_lle_ti()
158 	x->a = cpu_to_be64((a >> 8) ^ (_tt << 48));  in gf128mul_x8_lle_ti()
163 	u64 a = be64_to_cpu(x->a);  in gf128mul_x8_bbe()
164 	u64 b = be64_to_cpu(x->b);  in gf128mul_x8_bbe()
167 	x->a = cpu_to_be64((a << 8) | (b >> 56));  in gf128mul_x8_bbe()
168 	x->b = cpu_to_be64((b << 8) ^ _tt);  in gf128mul_x8_bbe()
173 	u64 a = le64_to_cpu(x->a);  in gf128mul_x8_ble()
174 	u64 b = le64_to_cpu(x->b);  in gf128mul_x8_ble()
177 	r->a = cpu_to_le64((a << 8) | (b >> 56));  in gf128mul_x8_ble()
178 	r->b = cpu_to_le64((b << 8) ^ _tt);  in gf128mul_x8_ble()
185 	 * The p array should be aligned to twice the size of its element type,  in gf128mul_lle()
201 	be128 *p = PTR_ALIGN(&array[0], 2 * sizeof(be128));  in gf128mul_lle()  local
204 	p[0] = *r;  in gf128mul_lle()
206 		gf128mul_x_lle(&p[2 * i + 2], &p[2 * i]);  in gf128mul_lle()
210 		u8 ch = ((u8 *)b)[15 - i];  in gf128mul_lle()
212 		be128_xor(r, r, &p[ 0 + !(ch & 0x80)]);  in gf128mul_lle()
213 		be128_xor(r, r, &p[ 2 + !(ch & 0x40)]);  in gf128mul_lle()
214 		be128_xor(r, r, &p[ 4 + !(ch & 0x20)]);  in gf128mul_lle()
215 		be128_xor(r, r, &p[ 6 + !(ch & 0x10)]);  in gf128mul_lle()
216 		be128_xor(r, r, &p[ 8 + !(ch & 0x08)]);  in gf128mul_lle()
217 		be128_xor(r, r, &p[10 + !(ch & 0x04)]);  in gf128mul_lle()
218 		be128_xor(r, r, &p[12 + !(ch & 0x02)]);  in gf128mul_lle()
219 		be128_xor(r, r, &p[14 + !(ch & 0x01)]);  in gf128mul_lle()
233     the 256 16 byte values that result from the 256 values
253 		t->t[i] = kzalloc(sizeof(*t->t[i]), GFP_KERNEL);  in gf128mul_init_64k_bbe()
254 		if (!t->t[i]) {  in gf128mul_init_64k_bbe()
261 	t->t[0]->t[1] = *g;  in gf128mul_init_64k_bbe()
263 		gf128mul_x_bbe(&t->t[0]->t[j + j], &t->t[0]->t[j]);  in gf128mul_init_64k_bbe()
266 		for (j = 2; j < 256; j += j)  in gf128mul_init_64k_bbe()
268 				be128_xor(&t->t[i]->t[j + k],  in gf128mul_init_64k_bbe()
269 					  &t->t[i]->t[j], &t->t[i]->t[k]);  in gf128mul_init_64k_bbe()
275 			t->t[i]->t[j] = t->t[i - 1]->t[j];  in gf128mul_init_64k_bbe()
276 			gf128mul_x8_bbe(&t->t[i]->t[j]);  in gf128mul_init_64k_bbe()
290 		kfree_sensitive(t->t[i]);  in gf128mul_free_64k()
301 	*r = t->t[0]->t[ap[15]];  in gf128mul_64k_bbe()
303 		be128_xor(r, r, &t->t[i]->t[ap[15 - i]]);  in gf128mul_64k_bbe()
311     single byte, we can construct a table of the 256 16 byte
312     values that result from the 256 values of this byte.
333 	t->t[128] = *g;  in gf128mul_init_4k_lle()
335 		gf128mul_x_lle(&t->t[j], &t->t[j+j]);  in gf128mul_init_4k_lle()
337 	for (j = 2; j < 256; j += j)  in gf128mul_init_4k_lle()
339 			be128_xor(&t->t[j + k], &t->t[j], &t->t[k]);  in gf128mul_init_4k_lle()
352 	*r = t->t[ap[15]];  in gf128mul_4k_lle()
353 	while (i--) {  in gf128mul_4k_lle()
355 		be128_xor(r, r, &t->t[ap[i]]);  in gf128mul_4k_lle()