ARROW-10914: [Rust] Refactor simd arithmetic kernels to use chunked iteration

This should avoid out of bounds reads in those kernels and also removes some unsafe blocks. The performance actually seems to increase slightly, I will post the numbers later.

The division kernel looks a bit more complex now, but follows the same pattern as used in the simd aggregation functions, with two separate codepaths for the nullable/non-null cases.

The comparison kernels might also do some out of bounds reads, I will look at those in a separate ticket/PR.

Closes apache#8929 from jhorstmann/ARROW-10914-simd-arithmetic-read-out-of-bounds

Authored-by: Jörn Horstmann <[email protected]>
Signed-off-by: Jorge C. Leitao <[email protected]>

Author: jhorstmann

rust/arrow/src/buffer.rs

@@ -824,7 +824,7 @@ impl MutableBuffer {
     }
 
     /// View buffer as typed slice.
-    pub fn typed_data_mut<T: ArrowNativeType + num::Num>(&mut self) -> &mut [T] {
+    pub fn typed_data_mut<T: ArrowNativeType>(&mut self) -> &mut [T] {
         assert_eq!(self.len() % mem::size_of::<T>(), 0);
         assert!(memory::is_ptr_aligned::<T>(self.raw_data() as *const T));
         unsafe {