17 typedef __uint128_t u128;  typedef
53 static __always_inline void fproduct_copy_from_wide_(u64 *output, u128 *input)  in fproduct_copy_from_wide_()
56 		u128 xi = input[0];  in fproduct_copy_from_wide_()
60 		u128 xi = input[1];  in fproduct_copy_from_wide_()
64 		u128 xi = input[2];  in fproduct_copy_from_wide_()
68 		u128 xi = input[3];  in fproduct_copy_from_wide_()
72 		u128 xi = input[4];  in fproduct_copy_from_wide_()
78 fproduct_sum_scalar_multiplication_(u128 *output, u64 *input, u64 s)  in fproduct_sum_scalar_multiplication_()
80 	output[0] += (u128)input[0] * s;  in fproduct_sum_scalar_multiplication_()
81 	output[1] += (u128)input[1] * s;  in fproduct_sum_scalar_multiplication_()
82 	output[2] += (u128)input[2] * s;  in fproduct_sum_scalar_multiplication_()
83 	output[3] += (u128)input[3] * s;  in fproduct_sum_scalar_multiplication_()
84 	output[4] += (u128)input[4] * s;  in fproduct_sum_scalar_multiplication_()
87 static __always_inline void fproduct_carry_wide_(u128 *tmp)  in fproduct_carry_wide_()
91 		u128 tctr = tmp[ctr];  in fproduct_carry_wide_()
92 		u128 tctrp1 = tmp[ctr + 1];  in fproduct_carry_wide_()
94 		u128 c = ((tctr) >> (51));  in fproduct_carry_wide_()
95 		tmp[ctr] = ((u128)(r0));  in fproduct_carry_wide_()
100 		u128 tctr = tmp[ctr];  in fproduct_carry_wide_()
101 		u128 tctrp1 = tmp[ctr + 1];  in fproduct_carry_wide_()
103 		u128 c = ((tctr) >> (51));  in fproduct_carry_wide_()
104 		tmp[ctr] = ((u128)(r0));  in fproduct_carry_wide_()
110 		u128 tctr = tmp[ctr];  in fproduct_carry_wide_()
111 		u128 tctrp1 = tmp[ctr + 1];  in fproduct_carry_wide_()
113 		u128 c = ((tctr) >> (51));  in fproduct_carry_wide_()
114 		tmp[ctr] = ((u128)(r0));  in fproduct_carry_wide_()
119 		u128 tctr = tmp[ctr];  in fproduct_carry_wide_()
120 		u128 tctrp1 = tmp[ctr + 1];  in fproduct_carry_wide_()
122 		u128 c = ((tctr) >> (51));  in fproduct_carry_wide_()
123 		tmp[ctr] = ((u128)(r0));  in fproduct_carry_wide_()
157 static __always_inline void fmul_mul_shift_reduce_(u128 *output, u64 *input,  in fmul_mul_shift_reduce_()
191 		u128 b4;  in fmul_fmul()
192 		u128 b0;  in fmul_fmul()
193 		u128 b4_;  in fmul_fmul()
194 		u128 b0_;  in fmul_fmul()
199 		u128 t[5] = { 0 };  in fmul_fmul()
204 		b4_ = ((b4) & (((u128)(0x7ffffffffffffLLU))));  in fmul_fmul()
205 		b0_ = ((b0) + (((u128)(19) * (((u64)(((b4) >> (51))))))));  in fmul_fmul()
218 static __always_inline void fsquare_fsquare__(u128 *tmp, u64 *output)  in fsquare_fsquare__()
230 	u128 s0 = ((((((u128)(r0) * (r0))) + (((u128)(d4) * (r1))))) +  in fsquare_fsquare__()
231 		   (((u128)(d2) * (r3))));  in fsquare_fsquare__()
232 	u128 s1 = ((((((u128)(d0) * (r1))) + (((u128)(d4) * (r2))))) +  in fsquare_fsquare__()
233 		   (((u128)(r3 * 19) * (r3))));  in fsquare_fsquare__()
234 	u128 s2 = ((((((u128)(d0) * (r2))) + (((u128)(r1) * (r1))))) +  in fsquare_fsquare__()
235 		   (((u128)(d4) * (r3))));  in fsquare_fsquare__()
236 	u128 s3 = ((((((u128)(d0) * (r3))) + (((u128)(d1) * (r2))))) +  in fsquare_fsquare__()
237 		   (((u128)(r4) * (d419))));  in fsquare_fsquare__()
238 	u128 s4 = ((((((u128)(d0) * (r4))) + (((u128)(d1) * (r3))))) +  in fsquare_fsquare__()
239 		   (((u128)(r2) * (r2))));  in fsquare_fsquare__()
247 static __always_inline void fsquare_fsquare_(u128 *tmp, u64 *output)  in fsquare_fsquare_()
249 	u128 b4;  in fsquare_fsquare_()
250 	u128 b0;  in fsquare_fsquare_()
251 	u128 b4_;  in fsquare_fsquare_()
252 	u128 b0_;  in fsquare_fsquare_()
261 	b4_ = ((b4) & (((u128)(0x7ffffffffffffLLU))));  in fsquare_fsquare_()
262 	b0_ = ((b0) + (((u128)(19) * (((u64)(((b4) >> (51))))))));  in fsquare_fsquare_()
274 static __always_inline void fsquare_fsquare_times_(u64 *output, u128 *tmp,  in fsquare_fsquare_times_()
286 	u128 t[5];  in fsquare_fsquare_times()
294 	u128 t[5];  in fsquare_fsquare_times_inplace()
399 	u128 tmp[5];  in fscalar()
400 	u128 b4;  in fscalar()
401 	u128 b0;  in fscalar()
402 	u128 b4_;  in fscalar()
403 	u128 b0_;  in fscalar()
406 		tmp[0] = ((u128)(xi) * (s));  in fscalar()
410 		tmp[1] = ((u128)(xi) * (s));  in fscalar()
414 		tmp[2] = ((u128)(xi) * (s));  in fscalar()
418 		tmp[3] = ((u128)(xi) * (s));  in fscalar()
422 		tmp[4] = ((u128)(xi) * (s));  in fscalar()
427 	b4_ = ((b4) & (((u128)(0x7ffffffffffffLLU))));  in fscalar()
428 	b0_ = ((b0) + (((u128)(19) * (((u64)(((b4) >> (51))))))));  in fscalar()