many fixes

evm/src/cpu/kernel/asm/sha2/memory.asm

@@ -43,54 +43,54 @@
 %macro mload_kernel_general_u256
     // stack: offset
     DUP1
-    %mload_kernel_code_u32
+    %mload_kernel_general_u32
     // stack: c_7, offset
     %shl_const(32)
     // stack: c7 << 32, offset
     DUP2
-    %increment
+    %add_const(4)
     %mload_kernel_general_u32
     OR
     // stack: (c_7 << 32) | c_6, offset
     %shl_const(32)
     // stack: ((c_7 << 32) | c_6) << 32, offset
     DUP2
-    %add_const(2)
+    %add_const(8)
     %mload_kernel_general_u32
     OR
     // stack: (c_7 << 64) | (c_6 << 32) | c_5, offset
     %shl_const(32)
     // stack: ((c_7 << 64) | (c_6 << 32) | c_5) << 32, offset
     DUP2
-    %add_const(3)
+    %add_const(12)
     %mload_kernel_general_u32
     OR
     // stack: (c_7 << 96) | (c_6 << 64) | (c_5 << 32) | c_4, offset
     %shl_const(32)
     // stack: ((c_7 << 96) | (c_6 << 64) | (c_5 << 32) | c_4) << 32, offset
     DUP2
-    %add_const(4)
+    %add_const(16)
     %mload_kernel_general_u32
     OR
     // stack: (c_7 << 128) | (c_6 << 96) | (c_5 << 64) | (c_4 << 32) | c_3, offset
     %shl_const(32)
     // stack: ((c_7 << 128) | (c_6 << 96) | (c_5 << 64) | (c_4 << 32) | c_3) << 32, offset
     DUP2
-    %add_const(5)
+    %add_const(20)
     %mload_kernel_general_u32
     OR
     // stack: (c_7 << 160) | (c_6 << 128) | (c_5 << 96) | (c_4 << 64) | (c_3 << 32) | c_2, offset
     %shl_const(32)
     // stack: ((c_7 << 160) | (c_6 << 128) | (c_5 << 96) | (c_4 << 64) | (c_3 << 32) | c_2) << 32, offset
     DUP2
-    %add_const(6)
+    %add_const(24)
     %mload_kernel_general_u32
     OR
     // stack: (c_7 << 192) | (c_6 << 160) | (c_5 << 128) | (c_4 << 96) | (c_3 << 64) | (c_2 << 32) | c_1, offset
     %shl_const(32)
     // stack: ((c_7 << 192) | (c_6 << 160) | (c_5 << 128) | (c_4 << 96) | (c_3 << 64) | (c_2 << 32) | c_1) << 32, offset
     DUP2
-    %add_const(7)
+    %add_const(28)
     %mload_kernel_general_u32
     OR
     // stack: (c_7 << 224) | (c_6 << 192) | (c_5 << 160) | (c_4 << 128) | (c_3 << 96) | (c_2 << 64) | (c_1 << 32) | c_0, offset

evm/src/cpu/kernel/asm/sha2/sha2.asm

@@ -51,15 +51,16 @@ global sha2_pad:
     %mload_kernel_general
     // stack: num_bytes, retdest
     // STEP 1: append 1
-    // insert 128 (= 1 << 7) at x[num_bytes]
+    // insert 128 (= 1 << 7) at x[num_bytes+1]
     // stack: num_bytes, retdest
     push 1
     push 7
-    swap1 // TODO: remove once SHR implementation is fixed
     shl
     // stack: 128, num_bytes, retdest
     dup2
     // stack: num_bytes, 128, num_bytes, retdest
+    %increment
+    // stack: num_bytes+1, 128, num_bytes, retdest
     %mstore_kernel_general
     // stack: num_bytes, retdest
     // STEP 2: calculate num_blocks := (num_bytes+8)//64 + 1
@@ -70,20 +71,18 @@ global sha2_pad:
     
     %increment
     // stack: num_blocks = (num_bytes+8)//64 + 1, num_bytes, retdest
-    // STEP 3: calculate length := num_bytes*8+1
+    // STEP 3: calculate length := num_bytes*8
     swap1
     // stack: num_bytes, num_blocks, retdest
     push 8
     mul
-    %increment
-    // stack: length = num_bytes*8+1, num_blocks, retdest
-    // STEP 4: write length to x[num_blocks*64-8..num_blocks*64-1] 
+    // stack: length = num_bytes*8, num_blocks, retdest
+    // STEP 4: write length to x[num_blocks*64-7..num_blocks*64]
     dup2
     // stack: num_blocks, length, num_blocks, retdest
     push 64
     mul
-    %decrement
-    // stack: last_addr = num_blocks*64-1, length, num_blocks, retdest
+    // stack: last_addr = num_blocks*64, length, num_blocks, retdest
     %sha2_write_length
     // stack: num_blocks, retdest
     // STEP 5: write num_blocks to x[0]
@@ -114,6 +113,7 @@ global sha2_gen_message_schedule_from_block:
     %mload_kernel_general_u256
     // stack: block[1], block[0], output_addr, retdest
     swap2
+    STOP
     // stack: output_addr, block[0], block[1], retdest
     push 8
     // stack: counter=8, output_addr, block[0], block[1], retdest
@@ -123,10 +123,9 @@ sha2_gen_message_schedule_from_block_0_loop:
     // stack: counter, output_addr, block[0], block[1], retdest
     swap2
     // stack: block[0], output_addr, counter, block[1], retdest
-    push 15
     push 1
-    //shl
-    STOP
+    push 32
+    shl
     // stack: 1 << 32, block[0], output_addr, counter, block[1], retdest
     dup2
     dup2
@@ -138,21 +137,21 @@ sha2_gen_message_schedule_from_block_0_loop:
     swap2
     // stack: block[0], 1 << 32, block[0] % (1 << 32), output_addr, counter, block[1], retdest
     div
-    // stack: block[0] // (1 << 32), block[0] % (1 << 32), output_addr, counter, block[1], retdest
+    // stack: block[0] >> 32, block[0] % (1 << 32), output_addr, counter, block[1], retdest
     swap1
-    // stack: block[0] % (1 << 32), block[0] // (1 << 32), output_addr, counter, block[1], retdest
+    // stack: block[0] % (1 << 32), block[0] >> 32, output_addr, counter, block[1], retdest
     dup3
-    // stack: output_addr, block[0] % (1 << 32), block[0] // (1 << 32), output_addr, counter, block[1], retdest
+    // stack: output_addr, block[0] % (1 << 32), block[0] >> 32, output_addr, counter, block[1], retdest
     %mstore_kernel_general_u32
-    // stack: block[0] // (1 << 32), output_addr, counter, block[1], retdest
+    // stack: block[0] >> 32, output_addr, counter, block[1], retdest
     swap1
-    // stack: output_addr, block[0] // (1 << 32), counter, block[1], retdest
+    // stack: output_addr, block[0] >> 32, counter, block[1], retdest
     %add_const(4)
-    // stack: output_addr + 4, block[0] // (1 << 32), counter, block[1], retdest
+    // stack: output_addr + 4, block[0] >> 32, counter, block[1], retdest
     swap1
-    // stack: block[0] // (1 << 32), output_addr + 4, counter, block[1], retdest
+    // stack: block[0] >> 32, output_addr + 4, counter, block[1], retdest
     swap2
-    // stack: counter, output_addr + 4, block[0] // (1 << 32), block[1], retdest
+    // stack: counter, output_addr + 4, block[0] >> 32, block[1], retdest
     %decrement
     dup1
     iszero
@@ -177,7 +176,6 @@ sha2_gen_message_schedule_from_block_1_loop:
     // stack: block[1], output_addr, counter, block[0], retdest
     push 1
     push 32
-    swap1 // TODO: remove once SHL implementation is fixed
     shl
     // stack: 1 << 32, block[1], output_addr, counter, block[0], retdest
     dup2
@@ -190,21 +188,21 @@ sha2_gen_message_schedule_from_block_1_loop:
     swap2
     // stack: block[1], 1 << 32, block[1] % (1 << 32), output_addr, counter, block[0], retdest
     div
-    // stack: block[1] // (1 << 32), block[1] % (1 << 32), output_addr, counter, block[0], retdest
+    // stack: block[1] >> 32, block[1] % (1 << 32), output_addr, counter, block[0], retdest
     swap1
-    // stack: block[1] % (1 << 32), block[1] // (1 << 32), output_addr, counter, block[0], retdest
+    // stack: block[1] % (1 << 32), block[1] >> 32, output_addr, counter, block[0], retdest
     dup3
-    // stack: output_addr, block[1] % (1 << 32), block[1] // (1 << 32), output_addr, counter, block[0], retdest
+    // stack: output_addr, block[1] % (1 << 32), block[1] >> 32, output_addr, counter, block[0], retdest
     %mstore_kernel_general_u32
-    // stack: block[1] // (1 << 32), output_addr, counter, block[0], retdest
+    // stack: block[1] >> 32, output_addr, counter, block[0], retdest
     swap1
-    // stack: output_addr, block[1] // (1 << 32), counter, block[0], retdest
+    // stack: output_addr, block[1] >> 32, counter, block[0], retdest
     %add_const(4)
-    // stack: output_addr + 4, block[1] // (1 << 32), counter, block[0], retdest
+    // stack: output_addr + 4, block[1] >> 32, counter, block[0], retdest
     swap1
-    // stack: block[1] // (1 << 32), output_addr + 4, counter, block[0], retdest
+    // stack: block[1] >> 32, output_addr + 4, counter, block[0], retdest
     swap2
-    // stack: counter, output_addr + 4, block[1] // (1 << 32), block[0], retdest
+    // stack: counter, output_addr + 4, block[1] >> 32, block[0], retdest
     %decrement
     dup1
     iszero
@@ -307,7 +305,7 @@ sha2_gen_message_schedule_remaining_end:
     JUMP
 
 // Precodition: memory, starting at 0, contains num_blocks, block0[0], ..., block0[63], block1[0], ..., blocklast[63]
-                stack contains output_addr
+//              stack contains output_addr
 // Postcondition: 
 global sha2_gen_all_message_schedules: 
     JUMPDEST

evm/src/cpu/kernel/asm/sha2/write_length.asm

@@ -4,7 +4,6 @@
     // stack: length, last_addr
     push 1
     push 8
-    swap1 // TODO: remove once SHL implementation is fixed
     shl
 
     // stack: 1 << 8, length, last_addr
@@ -21,10 +20,9 @@
     // stack: 1 << 8, length, last_addr
     dup1
     // stack: 1 << 8, 1 << 8, length, last_addr
-    dup2
+    dup3
     // stack: length, 1 << 8, 1 << 8, length, last_addr
     push 8
-    swap1 // TODO: remove once SHR implementation is fixed
     shr
     // stack: length >> 8, 1 << 8, 1 << 8, length, last_addr
     mod
@@ -40,10 +38,9 @@
     // stack: 1 << 8, length, last_addr
     dup1
     // stack: 1 << 8, 1 << 8, length, last_addr
-    dup2
+    dup3
     // stack: length, 1 << 8, 1 << 8, length, last_addr
     push 16
-    swap1 // TODO: remove once SHR implementation is fixed
     shr
     // stack: length >> 16, 1 << 8, 1 << 8, length, last_addr
     mod
@@ -59,10 +56,9 @@
     // stack: 1 << 8, length, last_addr
     dup1
     // stack: 1 << 8, 1 << 8, length, last_addr
-    dup2
+    dup3
     // stack: length, 1 << 8, 1 << 8, length, last_addr
     push 24
-    swap1 // TODO: remove once SHR implementation is fixed
     shr
     // stack: length >> 24, 1 << 8, 1 << 8, length, last_addr
     mod
@@ -72,16 +68,15 @@
     push 3
     swap1
     sub
-    // stack: last_addr - 1, (length >> 24) % (1 << 8), 1 << 8, length, last_addr
+    // stack: last_addr - 3, (length >> 24) % (1 << 8), 1 << 8, length, last_addr
     %mstore_kernel_general
 
     // stack: 1 << 8, length, last_addr
     dup1
     // stack: 1 << 8, 1 << 8, length, last_addr
-    dup2
+    dup3
     // stack: length, 1 << 8, 1 << 8, length, last_addr
     push 32
-    swap1 // TODO: remove once SHR implementation is fixed
     shr
     // stack: length >> 32, 1 << 8, 1 << 8, length, last_addr
     mod
@@ -91,13 +86,13 @@
     push 4
     swap1
     sub
-    // stack: last_addr - 1, (length >> 32) % (1 << 8), 1 << 8, length, last_addr
+    // stack: last_addr - 4, (length >> 32) % (1 << 8), 1 << 8, length, last_addr
     %mstore_kernel_general
 
     // stack: 1 << 8, length, last_addr
     dup1
     // stack: 1 << 8, 1 << 8, length, last_addr
-    dup2
+    dup3
     // stack: length, 1 << 8, 1 << 8, length, last_addr
     push 40
     shr
@@ -109,16 +104,15 @@
     push 5
     swap1
     sub
-    // stack: last_addr - 1, (length >> 40) % (1 << 8), 1 << 8, length, last_addr
+    // stack: last_addr - 5, (length >> 40) % (1 << 8), 1 << 8, length, last_addr
     %mstore_kernel_general
 
     // stack: 1 << 8, length, last_addr
     dup1
     // stack: 1 << 8, 1 << 8, length, last_addr
-    dup2
+    dup3
     // stack: length, 1 << 8, 1 << 8, length, last_addr
     push 48
-    swap1 // TODO: remove once SHR implementation is fixed
     shr
     // stack: length >> 48, 1 << 8, 1 << 8, length, last_addr
     mod
@@ -128,16 +122,15 @@
     push 6
     swap1
     sub
-    // stack: last_addr - 1, (length >> 48) % (1 << 8), 1 << 8, length, last_addr
+    // stack: last_addr - 6, (length >> 48) % (1 << 8), 1 << 8, length, last_addr
     %mstore_kernel_general
 
     // stack: 1 << 8, length, last_addr
     dup1
     // stack: 1 << 8, 1 << 8, length, last_addr
-    dup2
+    dup3
     // stack: length, 1 << 8, 1 << 8, length, last_addr
     push 56
-    swap1 // TODO: remove once SHR implementation is fixed
     shr
     // stack: length >> 56, 1 << 8, 1 << 8, length, last_addr
     mod
@@ -147,7 +140,7 @@
     push 7
     swap1
     sub
-    // stack: last_addr - 1, (length >> 56) % (1 << 8), 1 << 8, length, last_addr
+    // stack: last_addr - 7, (length >> 56) % (1 << 8), 1 << 8, length, last_addr
     %mstore_kernel_general
     %pop3
     // stack: (empty)

evm/src/cpu/kernel/tests/sha2.rs

@@ -28,6 +28,9 @@ fn test_sha2_store() -> Result<()> {
     dbg!(num_bytes);
     dbg!(bytes.clone());
 
+    let num_bytes = 3;
+    let bytes: Vec<U256> = vec![97.into(), 98.into(), 99.into()];
+
     let mut store_initial_stack = vec![U256::from(num_bytes)];
     store_initial_stack.extend(bytes);
     store_initial_stack.push(U256::from_str("0xdeadbeef").unwrap());
@@ -48,7 +51,7 @@ fn test_sha2_store() -> Result<()> {
     let mem = memory_after_storing.context_memory[0].segments[Segment::KernelGeneral as usize]
         .content
         .clone();
-    dbg!(&mem[0..66]);
+    dbg!(&mem[0..65]);
 
     dbg!(&mem[100..353]);