Update architecture code to use safe accessors

- [x] 2. Phase 2: Update architecture code to use safe accessors
  - [x] 2.1 Update SIMD folding code in src/arch/ to use params.get_key()
    - Replace direct keys[index] access with params.get_key(index) in algorithm.rs
    - Update VPCLMULQDQ code to use safe key access methods
    - Update aarch64 and x86 architecture-specific code
    - _Requirements: 3.1, 5.2_

  - [x] 2.2 Update CRC32 algorithm code to use safe accessors
    - Modify src/crc32/algorithm.rs to use params.get_key() instead of keys[index]
    - Update fusion code in src/crc32/fusion/ if it accesses keys directly
    - _Requirements: 3.1, 5.2_

  - [x] 2.3 Update CRC64 algorithm code to use safe accessors
    - Modify src/crc64/algorithm.rs to use params.get_key() instead of keys[index]
    - Update any other CRC64-specific code that accesses keys directly
    - _Requirements: 3.1, 5.2_

  - [x] 2.4 Run performance benchmarks to verify zero overhead
    - Benchmark key access performance before and after changes
    - Verify compiler optimizations eliminate any performance regression
    - Document that performance remains identical to direct array access
    - _Requirements: 2.2, 4.4_

Author: onethumb

.kiro/specs/future-proof-crc-params/tasks.md

@@ -8,24 +8,24 @@
   - Write comprehensive unit tests for CrcKeysStorage functionality
   - _Requirements: 4.1, 4.2, 4.4, 5.1_
 
-- [ ] 2. Phase 2: Update architecture code to use safe accessors
-  - [ ] 2.1 Update SIMD folding code in src/arch/ to use params.get_key()
+- [x] 2. Phase 2: Update architecture code to use safe accessors
+  - [x] 2.1 Update SIMD folding code in src/arch/ to use params.get_key()
     - Replace direct keys[index] access with params.get_key(index) in algorithm.rs
     - Update VPCLMULQDQ code to use safe key access methods
     - Update aarch64 and x86 architecture-specific code
     - _Requirements: 3.1, 5.2_
 
-  - [ ] 2.2 Update CRC32 algorithm code to use safe accessors
+  - [x] 2.2 Update CRC32 algorithm code to use safe accessors
     - Modify src/crc32/algorithm.rs to use params.get_key() instead of keys[index]
     - Update fusion code in src/crc32/fusion/ if it accesses keys directly
     - _Requirements: 3.1, 5.2_
 
-  - [ ] 2.3 Update CRC64 algorithm code to use safe accessors
+  - [x] 2.3 Update CRC64 algorithm code to use safe accessors
     - Modify src/crc64/algorithm.rs to use params.get_key() instead of keys[index]
     - Update any other CRC64-specific code that accesses keys directly
     - _Requirements: 3.1, 5.2_
 
-  - [ ] 2.4 Run performance benchmarks to verify zero overhead
+  - [x] 2.4 Run performance benchmarks to verify zero overhead
     - Benchmark key access performance before and after changes
     - Verify compiler optimizations eliminate any performance regression
     - Document that performance remains identical to direct array access

src/algorithm.rs

@@ -21,6 +21,37 @@ use crate::structs::CrcState;
 use crate::traits::{ArchOps, EnhancedCrcWidth};
 use crate::{crc32, crc64, CrcParams};
 
+/// Extract keys from CrcParams using safe accessor methods
+/// This ensures bounds checking and future compatibility
+#[inline(always)]
+fn extract_keys_array(params: CrcParams) -> [u64; 23] {
+    [
+        params.get_key(0),
+        params.get_key(1),
+        params.get_key(2),
+        params.get_key(3),
+        params.get_key(4),
+        params.get_key(5),
+        params.get_key(6),
+        params.get_key(7),
+        params.get_key(8),
+        params.get_key(9),
+        params.get_key(10),
+        params.get_key(11),
+        params.get_key(12),
+        params.get_key(13),
+        params.get_key(14),
+        params.get_key(15),
+        params.get_key(16),
+        params.get_key(17),
+        params.get_key(18),
+        params.get_key(19),
+        params.get_key(20),
+        params.get_key(21),
+        params.get_key(22),
+    ]
+}
+
 /// Main entry point that works for both CRC-32 and CRC-64
 #[inline]
 #[cfg_attr(
@@ -65,13 +96,19 @@ where
             bytes,
             &mut crc_state,
             reflector,
-            params.keys,
+            extract_keys_array(params),
             ops,
         );
     }
 
     // Process large inputs with SIMD-optimized approach
-    process_large_aligned::<T, W>(bytes, &mut crc_state, reflector, params.keys, ops)
+    process_large_aligned::<T, W>(
+        bytes,
+        &mut crc_state,
+        reflector,
+        extract_keys_array(params),
+        ops,
+    )
 }
 
 /// Process data with the selected strategy