feat(avm): sha256_compression (#6452)

Please read [contributing guidelines](CONTRIBUTING.md) and remove this
line.

barretenberg/cpp/pil/avm/avm_main.pil

@@ -5,6 +5,7 @@ include "avm_binary.pil";
 include "constants.pil";
 include "avm_kernel.pil";
 include "gadgets/avm_conversion.pil";
+include "gadgets/avm_sha256.pil";
 
 namespace avm_main(256);
     //===== CONSTANT POLYNOMIALS ==================================================
@@ -48,6 +49,7 @@ namespace avm_main(256);
 
     //===== Gadget Selectors ======================================================
     pol commit sel_op_radix_le;
+    pol commit sel_op_sha256;
 
     //===== Fix Range Checks Selectors=============================================
     // We re-use the clk column for the lookup values of 8-bit resp. 16-bit range check.
@@ -202,6 +204,7 @@ namespace avm_main(256);
     sel_op_sstore * (1 - sel_op_sstore) = 0;
 
     sel_op_radix_le * (1 - sel_op_radix_le) = 0;
+    sel_op_sha256 * (1 - sel_op_sha256) = 0;
 
     sel_op_add * (1 - sel_op_add) = 0;
     sel_op_sub * (1 - sel_op_sub) = 0;
@@ -354,8 +357,10 @@ namespace avm_main(256);
 
     //===== CONTROL_FLOW_CONSISTENCY ============================================
     pol INTERNAL_CALL_STACK_SELECTORS = (first + sel_internal_call + sel_internal_return + sel_halt);
-    pol OPCODE_SELECTORS = (sel_op_add + sel_op_sub + sel_op_div + sel_op_fdiv + sel_op_mul + sel_op_not
-                          + sel_op_eq + sel_op_and + sel_op_or + sel_op_xor + sel_op_cast + KERNEL_INPUT_SELECTORS + KERNEL_OUTPUT_SELECTORS);
+    pol ALL_BINARY_SEL = sel_op_and + sel_op_or + sel_op_xor;
+    pol ALL_GADGET_SEL = sel_op_radix_le + sel_op_sha256;
+    pol ALL_MEMORY_SEL = sel_cmov + sel_mov;
+    pol OPCODE_SELECTORS = ALU_ALL_SEL + ALL_BINARY_SEL + ALL_MEMORY_SEL + ALL_GADGET_SEL + KERNEL_INPUT_SELECTORS + KERNEL_OUTPUT_SELECTORS;
 
     // Program counter must increment if not jumping or returning
     #[PC_INCREMENT]
@@ -549,6 +554,11 @@ namespace avm_main(256);
     is
     avm_conversion.to_radix_le_sel {avm_conversion.clk, avm_conversion.input, avm_conversion.radix, avm_conversion.num_limbs};
 
+    #[PERM_MAIN_SHA256]
+    sel_op_sha256 {clk, ia, ib, ic}
+    is
+    avm_sha256.sha256_compression_sel {avm_sha256.clk, avm_sha256.state, avm_sha256.input, avm_sha256.output};
+
     #[PERM_MAIN_MEM_A]
     mem_op_a {clk, space_id, mem_idx_a, ia, rwa
             , r_in_tag, w_in_tag, sel_mov_a, sel_cmov}

barretenberg/cpp/pil/avm/gadgets/avm_sha256.pil

@@ -0,0 +1,14 @@
+include "../avm_main.pil";
+
+namespace avm_sha256(256);
+
+    pol commit clk;
+
+    // Selector for Radix Operation
+    pol commit sha256_compression_sel;
+    sha256_compression_sel * (1 - sha256_compression_sel) = 0;
+
+    // These will all be arrays, but we just store the first element for permutation to the main trace for now
+    pol commit state;
+    pol commit input;
+    pol commit output;