Introduce Value::as_i32 (scroll-tech#732)

To help make scroll-tech#596 easier to read
and reason about.

Also introduce a few more conversion helpers.

ceno_zkvm/src/instructions/riscv/mul.rs

@@ -432,9 +432,7 @@ impl<E: ExtensionField> Signed<E> {
             lkm,
             *val.as_u16_limbs().last().unwrap() as u64,
         )?;
-        let signed_val = val.as_u32() as i32;
-
-        Ok(signed_val)
+        Ok(i32::from(val))
     }
 
     pub fn expr(&self) -> Expression<E> {

ceno_zkvm/src/uint.rs

@@ -606,6 +606,30 @@ pub struct Value<'a, T: Into<u64> + From<u32> + Copy + Default> {
     pub limbs: Cow<'a, [u16]>,
 }
 
+impl<'a, T: Into<u64> + From<u32> + Copy + Default> From<&'a Value<'a, T>> for &'a [u16] {
+    fn from(v: &'a Value<'a, T>) -> Self {
+        v.as_u16_limbs()
+    }
+}
+
+impl<'a, T: Into<u64> + From<u32> + Copy + Default> From<&Value<'a, T>> for u64 {
+    fn from(v: &Value<'a, T>) -> Self {
+        v.as_u64()
+    }
+}
+
+impl<'a, T: Into<u64> + From<u32> + Copy + Default> From<&Value<'a, T>> for u32 {
+    fn from(v: &Value<'a, T>) -> Self {
+        v.as_u32()
+    }
+}
+
+impl<'a, T: Into<u64> + From<u32> + Copy + Default> From<&Value<'a, T>> for i32 {
+    fn from(v: &Value<'a, T>) -> Self {
+        v.as_i32()
+    }
+}
+
 // TODO generalize to support non 16 bit limbs
 // TODO optimize api with fixed size array
 impl<'a, T: Into<u64> + From<u32> + Copy + Default> Value<'a, T> {
@@ -616,10 +640,7 @@ impl<'a, T: Into<u64> + From<u32> + Copy + Default> Value<'a, T> {
     const LIMBS: usize = (Self::M + 15) / 16;
 
     pub fn new(val: T, lkm: &mut LkMultiplicity) -> Self {
-        let uint = Value::<T> {
-            val,
-            limbs: Cow::Owned(Self::split_to_u16(val)),
-        };
+        let uint = Self::new_unchecked(val);
         Self::assert_u16(&uint.limbs, lkm);
         uint
     }
@@ -684,6 +705,11 @@ impl<'a, T: Into<u64> + From<u32> + Copy + Default> Value<'a, T> {
         self.as_u64() as u32
     }
 
+    /// Convert the limbs to an i32 value
+    pub fn as_i32(&self) -> i32 {
+        self.as_u32() as i32
+    }
+
     pub fn u16_fields<F: SmallField>(&self) -> Vec<F> {
         self.limbs.iter().map(|v| F::from(*v as u64)).collect_vec()
     }