101     b0.val[0] = vcombine_s16(  in vtrnq_s64_to_s16()
103     b0.val[1] = vcombine_s16(  in vtrnq_s64_to_s16()
131         vtrn_s32(vreinterpret_s32_s16(b0.val[0]), vreinterpret_s32_s16(b1.val[0]));  in transpose_s16_4x4d()
133         vtrn_s32(vreinterpret_s32_s16(b0.val[1]), vreinterpret_s32_s16(b1.val[1]));  in transpose_s16_4x4d()
135     *a0 = vreinterpret_s16_s32(c0.val[0]);  in transpose_s16_4x4d()
136     *a1 = vreinterpret_s16_s32(c1.val[0]);  in transpose_s16_4x4d()
137     *a2 = vreinterpret_s16_s32(c0.val[1]);  in transpose_s16_4x4d()
138     *a3 = vreinterpret_s16_s32(c1.val[1]);  in transpose_s16_4x4d()
164         vtrnq_s32(vreinterpretq_s32_s16(b0.val[0]), vreinterpretq_s32_s16(b1.val[0]));  in transpose_s16_4x4q()
166         vtrnq_s32(vreinterpretq_s32_s16(b0.val[1]), vreinterpretq_s32_s16(b1.val[1]));  in transpose_s16_4x4q()
168     *a0 = vreinterpretq_s16_s32(c0.val[0]);  in transpose_s16_4x4q()
169     *a1 = vreinterpretq_s16_s32(c1.val[0]);  in transpose_s16_4x4q()
170     *a2 = vreinterpretq_s16_s32(c0.val[1]);  in transpose_s16_4x4q()
171     *a3 = vreinterpretq_s16_s32(c1.val[1]);  in transpose_s16_4x4q()
220     c0 = vtrnq_s32(vreinterpretq_s32_s16(b0.val[0]), vreinterpretq_s32_s16(b1.val[0]));  in transpose_s16_8x8()
221     c1 = vtrnq_s32(vreinterpretq_s32_s16(b0.val[1]), vreinterpretq_s32_s16(b1.val[1]));  in transpose_s16_8x8()
222     c2 = vtrnq_s32(vreinterpretq_s32_s16(b2.val[0]), vreinterpretq_s32_s16(b3.val[0]));  in transpose_s16_8x8()
223     c3 = vtrnq_s32(vreinterpretq_s32_s16(b2.val[1]), vreinterpretq_s32_s16(b3.val[1]));  in transpose_s16_8x8()
235     d0 = vtrnq_s64_to_s16(c0.val[0], c2.val[0]);  in transpose_s16_8x8()
236     d1 = vtrnq_s64_to_s16(c1.val[0], c3.val[0]);  in transpose_s16_8x8()
237     d2 = vtrnq_s64_to_s16(c0.val[1], c2.val[1]);  in transpose_s16_8x8()
238     d3 = vtrnq_s64_to_s16(c1.val[1], c3.val[1]);  in transpose_s16_8x8()
240     *a0 = d0.val[0];  in transpose_s16_8x8()
241     *a1 = d1.val[0];  in transpose_s16_8x8()
242     *a2 = d2.val[0];  in transpose_s16_8x8()
243     *a3 = d3.val[0];  in transpose_s16_8x8()
244     *a4 = d0.val[1];  in transpose_s16_8x8()
245     *a5 = d1.val[1];  in transpose_s16_8x8()
246     *a6 = d2.val[1];  in transpose_s16_8x8()
247     *a7 = d3.val[1];  in transpose_s16_8x8()
253     b0.val[0] = vcombine_s32(vget_low_s32(a0), vget_low_s32(a1));  in vtrnq_s64_to_s32()
254     b0.val[1] = vcombine_s32(vget_high_s32(a0), vget_high_s32(a1));  in vtrnq_s64_to_s32()
280     const int32x4x2_t c0 = vtrnq_s64_to_s32(b0.val[0], b1.val[0]);  in transpose_s32_4x4()
281     const int32x4x2_t c1 = vtrnq_s64_to_s32(b0.val[1], b1.val[1]);  in transpose_s32_4x4()
283     *a0 = c0.val[0];  in transpose_s32_4x4()
284     *a1 = c1.val[0];  in transpose_s32_4x4()
285     *a2 = c0.val[1];  in transpose_s32_4x4()
286     *a3 = c1.val[1];  in transpose_s32_4x4()
318     const int32x4x2_t b0 = vtrnq_s32(a0->val[0], a1->val[0]);  in transpose_s32_8x8()
319     const int32x4x2_t b1 = vtrnq_s32(a2->val[0], a3->val[0]);  in transpose_s32_8x8()
320     const int32x4x2_t b2 = vtrnq_s32(a4->val[0], a5->val[0]);  in transpose_s32_8x8()
321     const int32x4x2_t b3 = vtrnq_s32(a6->val[0], a7->val[0]);  in transpose_s32_8x8()
322     const int32x4x2_t b4 = vtrnq_s32(a0->val[1], a1->val[1]);  in transpose_s32_8x8()
323     const int32x4x2_t b5 = vtrnq_s32(a2->val[1], a3->val[1]);  in transpose_s32_8x8()
324     const int32x4x2_t b6 = vtrnq_s32(a4->val[1], a5->val[1]);  in transpose_s32_8x8()
325     const int32x4x2_t b7 = vtrnq_s32(a6->val[1], a7->val[1]);  in transpose_s32_8x8()
336     const int32x4x2_t c0 = vtrnq_s64_to_s32(b0.val[0], b1.val[0]);  in transpose_s32_8x8()
337     const int32x4x2_t c1 = vtrnq_s64_to_s32(b0.val[1], b1.val[1]);  in transpose_s32_8x8()
338     const int32x4x2_t c2 = vtrnq_s64_to_s32(b2.val[0], b3.val[0]);  in transpose_s32_8x8()
339     const int32x4x2_t c3 = vtrnq_s64_to_s32(b2.val[1], b3.val[1]);  in transpose_s32_8x8()
340     const int32x4x2_t c4 = vtrnq_s64_to_s32(b4.val[0], b5.val[0]);  in transpose_s32_8x8()
341     const int32x4x2_t c5 = vtrnq_s64_to_s32(b4.val[1], b5.val[1]);  in transpose_s32_8x8()
342     const int32x4x2_t c6 = vtrnq_s64_to_s32(b6.val[0], b7.val[0]);  in transpose_s32_8x8()
343     const int32x4x2_t c7 = vtrnq_s64_to_s32(b6.val[1], b7.val[1]);  in transpose_s32_8x8()
354     a0->val[0] = c0.val[0];  in transpose_s32_8x8()
355     a0->val[1] = c2.val[0];  in transpose_s32_8x8()
356     a1->val[0] = c1.val[0];  in transpose_s32_8x8()
357     a1->val[1] = c3.val[0];  in transpose_s32_8x8()
358     a2->val[0] = c0.val[1];  in transpose_s32_8x8()
359     a2->val[1] = c2.val[1];  in transpose_s32_8x8()
360     a3->val[0] = c1.val[1];  in transpose_s32_8x8()
361     a3->val[1] = c3.val[1];  in transpose_s32_8x8()
362     a4->val[0] = c4.val[0];  in transpose_s32_8x8()
363     a4->val[1] = c6.val[0];  in transpose_s32_8x8()
364     a5->val[0] = c5.val[0];  in transpose_s32_8x8()
365     a5->val[1] = c7.val[0];  in transpose_s32_8x8()
366     a6->val[0] = c4.val[1];  in transpose_s32_8x8()
367     a6->val[1] = c6.val[1];  in transpose_s32_8x8()
368     a7->val[0] = c5.val[1];  in transpose_s32_8x8()
369     a7->val[1] = c7.val[1];  in transpose_s32_8x8()