Add remaining insturctions (rust-lang#1250)

* add vmmla vusmmla vsm4e vsm3 vrax1 vxar vsha512 vbcax veor3 neon instructions

* update runtime feature detect

* correct tests

* add `vrnd32x` `vrnd64x`

* add MISSING.md

crates/core_arch/MISSING.md

@@ -0,0 +1,194 @@
+## The following neon instructions are currently not implemented in stdarch
+
+### Can be implemented next:
+
+`vcls_u16`
+
+`vcls_u32`
+
+`vcls_u8`
+
+`vclsq_u16`
+
+`vclsq_u32`
+
+`vclsq_u8`
+
+`vcreate_s16`
+
+`vcreate_u16`
+
+`vpaddq_s64`
+
+`vpaddq_u64`
+
+`vreinterpretq_p128_f32`
+
+`vreinterpretq_p128_f64`
+
+`vreinterpretq_p128_p16`
+
+`vreinterpretq_p128_p8`
+
+`vreinterpretq_p128_s16`
+
+`vreinterpretq_p128_s32`
+
+`vreinterpretq_p128_s64`
+
+`vreinterpretq_p128_s8`
+
+`vreinterpretq_p128_u16`
+
+`vreinterpretq_p128_u32`
+
+`vreinterpretq_p128_u64`
+
+`vreinterpretq_p128_u8`
+
+`vslid_n_s64`
+
+`vslid_n_u64`
+
+`vsrid_n_s64`
+
+`vsrid_n_u64`
+
+### Not implemented on arm:
+
+`vcadd_rot270_f32`
+
+`vcadd_rot90_f32`
+
+`vcaddq_rot270_f32`
+
+`vcaddq_rot90_f32`
+
+`vdot_s32`
+
+`vdot_u32`
+
+`vdotq_s32`
+
+`vdotq_u32`
+
+`vdot_lane_s32`
+
+`vdot_lane_u32`
+
+`vdotq_lane_s32`
+
+`vdotq_lane_u32`
+
+`vcmla_f32`
+
+`vcmla_lane_f32`
+
+`vcmla_laneq_f32`
+
+`vcmla_rot180_f32`
+
+`vcmla_rot180_lane_f32`
+
+`vcmla_rot180_laneq_f32`
+
+`vcmla_rot270_f32`
+
+`vcmla_rot270_lane_f32`
+
+`vcmla_rot270_laneq_f32`
+
+`vcmla_rot90_f32`
+
+`vcmla_rot90_lane_f32`
+
+`vcmla_rot90_laneq_f32`
+
+`vcmlaq_f32`
+
+`vcmlaq_lane_f32`
+
+`vcmlaq_laneq_f32`
+
+`vcmlaq_rot180_f32`
+
+`vcmlaq_rot180_lane_f32`
+
+`vcmlaq_rot180_laneq_f32`
+
+`vcmlaq_rot270_f32`
+
+`vcmlaq_rot270_lane_f32`
+
+`vcmlaq_rot270_laneq_f32`
+
+`vcmlaq_rot90_f32`
+
+`vcmlaq_rot90_lane_f32`
+
+`vcmlaq_rot90_laneq_f32`
+
+### Not implemented in LLVM:
+
+`vrnd32x_f64`
+
+`vrnd32xq_f64`
+
+`vrnd32z_f64`
+
+`vrnd32zq_f64`
+
+`vrnd64x_f64`
+
+`vrnd64xq_f64`
+
+`vrnd64z_f64`
+
+`vrnd64zq_f64`
+
+### LLVM Select errors may occur:
+
+`vsudot_lane_s32`
+
+`vsudot_laneq_s32`
+
+`vsudotq_lane_s32`
+
+`vsudotq_laneq_s32`
+
+`vusdot_lane_s32`
+
+`vusdot_laneq_s32`
+
+`vusdot_s32`
+
+`vusdotq_lane_s32`
+
+`vusdotq_laneq_s32`
+
+`vusdotq_s32v`
+
+`vqshlu_n_s16`
+
+`vqshlu_n_s32`
+
+`vqshlu_n_s64`
+
+`vqshlu_n_s8`
+
+`vqshlub_n_s8`
+
+`vqshlud_n_s64`
+
+`vqshluh_n_s16`
+
+`vqshluq_n_s16`
+
+`vqshluq_n_s32`
+
+`vqshluq_n_s64`
+
+`vqshluq_n_s8`
+
+`vqshlus_n_s32`
+