separate op switches, remove inline op code, strip whitespace

2 files changed, 368 insertions, 330 deletions

diff --git a/gen.rb b/gen.rb
index 4b02320..6d579ca 100644
--- a/gen.rb
+++ b/gen.rb
@@ -6,14 +6,15 @@ require 'erb'
 # load template data
 DATA = YAML.load_file(File.join(__dir__, 'ops.yaml'))
 
-switches = DATA['ops'].reduce([]) do |r, set|
-  prefix = (set.first == 'cb') ? 0xCB00 : 0x0000
-
-  set.last.reduce(r) do |r, op|
+switches = Hash.new { |h, k| h[k] = [] }
+DATA['ops'].each do |set_id, ops|
+  switches[set_id] = ops.select { |op|
+    op['op'] != 'PREFIX'
+  }.map { |op|
     # op hex string
-    hex = (prefix + op['hex']).to_s(16).upcase.rjust(4, '0')
+    hex = op['hex'].to_s(16).upcase.rjust(2, '0')
 
-    r << case op['op']
+    case op['op']
     when 'XX'
       [
         "case 0x%s: /* op: %s, cat: %s */",
@@ -50,8 +51,8 @@ switches = DATA['ops'].reduce([]) do |r, set|
         time_expr,
       ]
     end
-  end
-end.join("\n")
+  }.join("\n")
+end
 
 t = ERB.new(DATA['templates']['main'])
 puts t.run(binding)
diff --git a/ops.yaml b/ops.yaml
index 496e701..ed60519 100644
--- a/ops.yaml
+++ b/ops.yaml
@@ -366,13 +366,7 @@ ops:
         h:
         c:
       code: |
-        {
-          // get offset
-          const int8_t ofs = (int8_t) mmu_rb(ctx, old_pc + 1);
-
-          // jump
-          cpu_ww(ctx, RW_PC, cpu_rw(ctx, RW_PC) + ofs);
-        }
+        jr(ctx, old_pc + 1);
     - id: ADD HL, DE
       hex: 0x19
       cat: "16-bit math"
@@ -491,17 +485,9 @@ ops:
         h:
         c:
       code: |
-        {
-          if (!FLAG(ctx, Z)) {
-            // get offset
-            const int8_t ofs = (int8_t) mmu_rb(ctx, old_pc + 1);
-
-            // jump
-            cpu_ww(ctx, RW_PC, cpu_rw(ctx, RW_PC) + ofs);
-
-            // add true time
-            ctx->cpu.clock += 4; // FIXME: 4 should not be hard-coded
-          }
+        if (!FLAG(ctx, Z)) {
+          jr(ctx, old_pc + 1);
+          ctx->cpu.clock += 4; // FIXME
         }
     - id: LD HL, d16
       hex: 0x21
@@ -618,17 +604,9 @@ ops:
         h: "0"
         c: C
       code: |
-        {
-          if (FLAG(ctx, Z)) {
-            // get offset
-            const int8_t ofs = (int8_t) mmu_rb(ctx, old_pc + 1);
-
-            // jump
-            cpu_ww(ctx, RW_PC, cpu_rw(ctx, RW_PC) + ofs);
-
-            // add true time
-            ctx->cpu.clock += 4; // FIXME: 4 should not be hard-coded
-          }
+        if (FLAG(ctx, Z)) {
+          jr(ctx, old_pc + 1);
+          ctx->cpu.clock += 4; // FIXME: 4 should not be hard-coded
         }
     - id: ADD HL, HL
       hex: 0x29
@@ -749,17 +727,9 @@ ops:
         h:
         c:
       code: |
-        {
-          if (!(cpu_rb(ctx, RB_C) & F_Z)) {
-            // get offset
-            const int8_t ofs = (int8_t) mmu_rb(ctx, old_pc + 1);
-
-            // jump
-            cpu_ww(ctx, RW_PC, cpu_rw(ctx, RW_PC) + ofs);
-
-            // add true time
-            ctx->cpu.clock += 4; // FIXME: 4 should not be hard-coded
-          }
+        if (!FLAG(ctx, C)) {
+          jr(ctx, old_pc + 1);
+          ctx->cpu.clock += 4; // FIXME
         }
     - id: LD SP, d16
       hex: 0x31
@@ -790,10 +760,8 @@ ops:
         h:
         c:
       code: |
-        {
-          mmu_wb(ctx, cpu_rw(ctx, RW_HL), cpu_rb(ctx, RB_A));
-          cpu_wb(ctx, cpu_rw(ctx, RW_HL), cpu_rw(ctx, RW_HL) - 1);
-        }
+        mmu_wb(ctx, cpu_rw(ctx, RW_HL), cpu_rb(ctx, RB_A));
+        cpu_wb(ctx, cpu_rw(ctx, RW_HL), cpu_rw(ctx, RW_HL) - 1);
     - id: INC SP
       hex: 0x33
       cat: "16-bit math"
@@ -821,21 +789,7 @@ ops:
         h: H
         c:
       code: |
-        {
-          // write flags
-          const uint16_t addr = cpu_rw(ctx, RW_HL);
-          const uint8_t o = mmu_rb(ctx, addr),
-                        n = o + 1;
-
-          // write value
-          mmu_wb(ctx, addr, n);
-
-          // write flags
-          cpu_wf(ctx, F_Z | F_N | F_H, (
-            (n ? 0 : F_Z) | 
-            (((o & 0x0F) == 0x0F) ? F_H : 0)
-          ));
-        }
+        inc_hl_ptr(ctx);
     - id: DEC (HL)
       hex: 0x35
       cat: "8-bit math"
@@ -849,22 +803,7 @@ ops:
         h: H
         c:
       code: |
-        {
-          // write flags
-          const uint16_t addr = cpu_rw(ctx, RW_HL);
-          const uint8_t o = mmu_rb(ctx, addr),
-                        n = o - 1;
-
-          // write value
-          mmu_wb(ctx, addr, n);
-
-          // write flags
-          cpu_wf(ctx, F_Z | F_N | F_H, (
-            (n ? 0 : F_Z) | 
-            F_N |
-            (((o & 0x0F) == 0x00) ? F_H : 0)
-          ));
-        }
+        dec_hl_ptr(ctx);
     - id: LD (HL), d8
       hex: 0x36
       cat: "8-bit load/store/move"
@@ -911,17 +850,9 @@ ops:
         h:
         c:
       code: |
-        {
-          if (cpu_rb(ctx, RB_C) & F_Z) {
-            // get offset
-            const int8_t ofs = (int8_t) mmu_rb(ctx, old_pc + 1);
-
-            // jump
-            cpu_ww(ctx, RW_PC, cpu_rw(ctx, RW_PC) + ofs);
-
-            // add true time
-            ctx->cpu.clock += 4; // FIXME: 4 should not be hard-coded
-          }
+        if (FLAG(ctx, C)) {
+          jr(ctx, old_pc + 1);
+          ctx->cpu.clock += 4;
         }
     - id: ADD HL, SP
       hex: 0x39
@@ -1024,9 +955,7 @@ ops:
         h: "0"
         c: C
       code: |
-        cpu_wf(ctx, F_N | F_H | F_C, (
-          (FLAG(ctx, C) ? 0 : F_C)
-        ));
+        cpu_wf(ctx, F_N | F_H | F_C, FLAG(ctx, C) ? 0 : F_C);
     - id: LD B, B
       hex: 0x40
       cat: "8-bit load/store/move"
@@ -2090,22 +2019,7 @@ ops:
         h: H
         c: C
       code: |
-        {
-          // get old and new value
-          const uint8_t o = cpu_rb(ctx, RB_A),
-                        n = mmu_rw(ctx, cpu_rw(ctx, RW_HL));
-          const uint16_t v = o + n;
-
-          // write value
-          cpu_wb(ctx, RB_A, v);
-
-          // set flags
-          cpu_wf(ctx, F_Z | F_N | F_H | F_C, (
-            ((v & 0xFF) ? 0 : F_Z) |
-            ((((o & 0x0F) + (n & 0x0F)) & 0xF0) ? F_H : 0) |
-            ((v & 0x100) ? F_C : 0)
-          ));
-        }
+        add_hl_ptr(ctx);
     - id: ADD A, A
       hex: 0x87
       cat: "8-bit math"
@@ -2225,23 +2139,7 @@ ops:
         h: H
         c: C
       code: |
-        {
-          // get old and new value
-          const uint8_t o = cpu_rb(ctx, RB_A),
-                        n = cpu_rb(ctx, mmu_rb(ctx, cpu_rw(ctx, RW_HL))),
-                        c = FLAG(ctx, C) ? 1 : 0;
-          const uint16_t v = o + n + c;
-
-          // write value
-          cpu_wb(ctx, RB_A, v);
-
-          // set flags
-          cpu_wf(ctx, F_Z | F_N | F_H | F_C, (
-            ((v & 0xFF) ? 0 : F_Z) |
-            ((((o & 0x0F) + (n & 0x0F) + c) & 0xF0) ? F_H : 0) |
-            ((v & 0x100) ? F_C : 0)
-          ));
-        }
+        adc_hl_ptr(ctx);
     - id: ADC A, A
       hex: 0x8F
       cat: "8-bit math"
@@ -2361,23 +2259,7 @@ ops:
         h: H
         c: C
       code: |
-        {
-          // get old and new value
-          const uint8_t o = cpu_rb(ctx, RB_A),
-                        n = mmu_rb(ctx, cpu_rw(ctx, RW_HL)),
-                        v = o - n;
-
-          // write value
-          cpu_wb(ctx, RB_A, v);
-
-          // set flags
-          cpu_wf(ctx, F_Z | F_N | F_H | F_C, (
-            (v ? 0 : F_Z) |
-            F_N |
-            ((((o & 0x0F) - (n & 0x0F)) & 0xF0) ? F_H : 0) |
-            ((n > o) ? F_C : 0)
-          ));
-        }
+        sub_hl_ptr(ctx);
     - id: SUB A
       hex: 0x97
       cat: "8-bit math"
@@ -2495,24 +2377,7 @@ ops:
         h: H
         c: C
       code: |
-        {
-          // get old and new value
-          const uint8_t o = cpu_rb(ctx, RB_A),
-                        n = mmu_rb(ctx, cpu_rw(ctx, RW_HL)),
-                        c = FLAG(ctx, C) ? 1 : 0,
-                        v = o - (n + c);
-
-          // write value
-          cpu_wb(ctx, RB_A, v);
-
-          // set flags
-          cpu_wf(ctx, F_Z | F_N | F_H | F_C, (
-            (v ? 0 : F_Z) |
-            F_N |
-            ((((o & 0x0F) - (n & 0x0F)) & 0xF0) ? F_H : 0) |
-            ((n > o) ? F_C : 0) // FIXME: is this right for sub/sbc?
-          ));
-        }
+        sbc_hl_ptr(ctx);
     - id: SBC A, A
       hex: 0x9F
       cat: "8-bit math"
@@ -2625,19 +2490,7 @@ ops:
         h: "1"
         c: "0"
       code: |
-        {
-          // get value
-          const uint8_t v = cpu_rb(ctx, RB_A) & mmu_rb(ctx, cpu_rb(ctx, RW_HL));
-
-          // write value
-          cpu_wb(ctx, RB_A, v);
-
-          // set flags
-          cpu_wf(ctx, F_Z | F_N | F_H | F_C, (
-            (v ? 0 : F_Z) | 
-            F_H
-          ));
-        }
+        and_hl_ptr(ctx);
     - id: AND A
       hex: 0xA7
       cat: "8-bit math"
@@ -2749,18 +2602,7 @@ ops:
         h: "0"
         c: "0"
       code: |
-        {
-          // get value
-          const uint8_t v = cpu_rb(ctx, RB_A) ^ mmu_rb(ctx, cpu_rw(ctx, RW_HL));
-
-          // write value
-          cpu_wb(ctx, RB_A, v);
-
-          // set flags
-          cpu_wf(ctx, F_Z | F_N | F_H | F_C, (
-            (v ? 0 : F_Z)
-          ));
-        }
+        xor_hl_ptr(ctx);
     - id: XOR A
       hex: 0xAF
       cat: "8-bit math"
@@ -2872,18 +2714,7 @@ ops:
         h: "0"
         c: "0"
       code: |
-        {
-          // get value
-          const uint8_t v = cpu_rb(ctx, RB_A) | mmu_rb(ctx, cpu_rw(ctx, RW_HL));
-
-          // write value
-          cpu_wb(ctx, RB_A, v);
-
-          // set flags
-          cpu_wf(ctx, F_Z | F_N | F_H | F_C, (
-            (v ? 0 : F_Z)
-          ));
-        }
+        or_hl_ptr(ctx);
     - id: OR A
       hex: 0xB7
       cat: "8-bit math"
@@ -3002,20 +2833,7 @@ ops:
         h: H
         c: C
       code: |
-        {
-          // get old and new value
-          const uint8_t o = cpu_rb(ctx, RB_A),
-                        n = mmu_rb(ctx, cpu_rw(ctx, RW_HL)),
-                        v = o - n;
-
-          // set flags
-          cpu_wf(ctx, F_Z | F_N | F_H | F_C, (
-            (v ? 0 : F_Z) |
-            F_N |
-            ((((o & 0x0F) - (n & 0x0F)) & 0xF0) ? F_H : 0) |
-            ((n > o) ? F_C : 0)
-          ));
-        }
+        cp_hl_ptr(ctx);
     - id: CP A
       hex: 0xBF
       cat: "8-bit math"
@@ -3050,7 +2868,6 @@ ops:
       code: |
         if (!FLAG(ctx, Z)) {
           pop_rw(ctx, RW_PC);
-
           ctx->cpu.clock += 12;
         }
     - id: POP BC
@@ -3086,9 +2903,7 @@ ops:
         c:
       code: |
         if (!FLAG(ctx, Z)) {
-          // jump
           cpu_ww(ctx, RW_PC, mmu_rw(ctx, old_pc + 1));
-
           ctx->cpu.clock -= 16; // FIXME
         }
     - id: JP a16
@@ -3125,12 +2940,7 @@ ops:
         c:
       code: |
         if (!FLAG(ctx, Z)) {
-          // push pc
-          push_rw(ctx, RW_PC);
-
-          // jump
-          cpu_ww(ctx, RW_PC, mmu_rw(ctx, old_pc + 1));
-
+          call_a16(ctx, old_pc + 1);
           ctx->cpu.clock += 12; // FIXME
         }
     - id: PUSH BC
@@ -3162,23 +2972,7 @@ ops:
         h: H
         c: C
       code: |
-        {
-          // get old and new value
-          const uint8_t o = cpu_rb(ctx, RB_A),
-                        n = mmu_rb(ctx, old_pc + 1);
-          const uint16_t v = o + n;
-
-          // write value
-          cpu_wb(ctx, RB_A, v);
-
-          // set flags
-          cpu_wf(ctx, F_Z | F_N | F_H | F_C, (
-            ((v & 0xFF) ? 0 : F_Z) |
-            ((((o & 0x0F) + (n & 0x0F)) & 0xF0) ? F_H : 0) |
-            ((v & 0x100) ? F_C : 0)
-          ));
-        }
-        
+        add_d8(ctx, old_pc + 1);
     - id: RST 00H
       hex: 0xC7
       cat: "jumps/calls"
@@ -3192,6 +2986,8 @@ ops:
         n:
         h:
         c:
+      code: |
+        rst(ctx, 0x00);
     - id: RET Z
       hex: 0xC8
       cat: "jumps/calls"
@@ -3210,10 +3006,7 @@ ops:
         c:
       code: |
         if (FLAG(ctx, Z)) {
-          // pop pc
           pop_rw(ctx, RW_PC);
-
-          // update clock
           ctx->cpu.clock += 12;
         }
     - id: RET
@@ -3250,9 +3043,7 @@ ops:
         c:
       code: |
         if (FLAG(ctx, Z)) {
-          // jump
           cpu_ww(ctx, RW_PC, mmu_rw(ctx, old_pc + 1));
-
           ctx->cpu.clock -= 16; // FIXME
         }
     - id: PREFIX CB
@@ -3282,12 +3073,7 @@ ops:
         c:
       code: |
         if (FLAG(ctx, Z)) {
-          // push pc
-          push_rw(ctx, RW_PC);
-
-          // jump
-          cpu_ww(ctx, RW_PC, mmu_rw(ctx, old_pc + 1));
-
+          call_a16(ctx, old_pc + 1);
           ctx->cpu.clock += 12;
         }
     - id: CALL a16
@@ -3302,11 +3088,7 @@ ops:
         h:
         c:
       code: |
-        // push pc
-        push_rw(ctx, RW_PC);
-
-        // jump
-        cpu_ww(ctx, RW_PC, mmu_rw(ctx, old_pc + 1));
+        call_a16(ctx, old_pc + 1);
     - id: ADC A, d8
       hex: 0xCE
       cat: "8-bit math"
@@ -3332,6 +3114,8 @@ ops:
         n:
         h:
         c:
+      code: |
+        rst(ctx, 0x08);
     - id: RET NC
       hex: 0xD0
       cat: "jumps/calls"
@@ -3345,10 +3129,7 @@ ops:
         f: 8
       code: |
         if (!FLAG(ctx, C)) {
-          // pop pc
           pop_rw(ctx, RW_PC);
-
-          // update clock
           ctx->cpu.clock += 12;
         }
       flags:
@@ -3390,15 +3171,15 @@ ops:
         c:
       code: |
         if (!FLAG(ctx, C)) {
-          // jump
           cpu_ww(ctx, RW_PC, mmu_rw(ctx, old_pc + 1));
-
           ctx->cpu.clock -= 16; // FIXME
         }
     - id: XX
       hex: 0xD3
       cat: "invalid"
       op: XX
+      code: |
+        invalid(ctx, old_pc);
     - id: CALL NC, a16
       hex: 0xD4
       cat: "jumps/calls"
@@ -3416,7 +3197,10 @@ ops:
         h:
         c:
       code: |
-        invalid(ctx, old_pc);
+        if (!FLAG(ctx, C)) {
+          call_a16(ctx, old_pc + 1);
+          ctx->cpu.clock += 12; // FIXME
+        }
     - id: PUSH DE
       hex: 0xD5
       cat: "16-bit load/store/move"
@@ -3460,6 +3244,8 @@ ops:
         n: "1"
         h: H
         c: C
+      code: |
+        rst(ctx, 0x10);
     - id: RET C
       hex: 0xD8
       cat: "jumps/calls"
@@ -3477,10 +3263,7 @@ ops:
         c:
       code: |
         if (FLAG(ctx, C)) {
-          // pop pc
           pop_rw(ctx, RW_PC);
-
-          // update clock
           ctx->cpu.clock += 12;
         }
     - id: RETI
@@ -3497,11 +3280,9 @@ ops:
         h:
         c:
       code: |
-        // enable interrupts
-        ctx->cpu.ei = true;
-
-        // pop pc
+        // pop pc, enable interrupts
         pop_rw(ctx, RW_PC);
+        ctx->cpu.ei = true;
     - id: JP C, a16
       hex: 0xDA
       cat: "jumps/calls"
@@ -3520,10 +3301,7 @@ ops:
         c:
       code: |
         if (FLAG(ctx, C)) {
-          // jump
           cpu_ww(ctx, RW_PC, mmu_rw(ctx, old_pc + 1));
-
-          // update clock
           ctx->cpu.clock -= 16;
         }
     - id: XX
@@ -3550,13 +3328,7 @@ ops:
         c:
       code: |
         if (FLAG(ctx, C)) {
-          // push pc
-          push_rw(ctx, RW_PC);
-
-          // jump
-          cpu_ww(ctx, RW_PC, mmu_rw(ctx, old_pc + 1));
-
-          // update clock
+          call_a16(ctx, old_pc + 1);
           ctx->cpu.clock += 12;
         }
     - id: XX
@@ -3591,6 +3363,8 @@ ops:
         n:
         h:
         c:
+      code: |
+        rst(ctx, 0x16);
     - id: LDH (a8), A
       hex: 0xE0
       cat: "8-bit load/store/move"
@@ -3686,6 +3460,8 @@ ops:
         n:
         h:
         c:
+      code: |
+        rst(ctx, 0x20);
     - id: ADD SP, r8
       hex: 0xE8
       cat: "misc"
@@ -3769,6 +3545,8 @@ ops:
         n:
         h:
         c:
+      code: |
+        rst(ctx, 0x28);
     - id: LDH A, (a8)
       hex: 0xF0
       cat: "8-bit load/store/move"
@@ -3862,18 +3640,7 @@ ops:
         h: "0"
         c: "0"
       code: |
-        {
-          // get value
-          const uint8_t v = cpu_rb(ctx, RB_A) ^ mmu_rb(ctx, old_pc + 1);
-
-          // write value
-          cpu_wb(ctx, RB_A, v);
-
-          // set flags
-          cpu_wf(ctx, F_Z | F_N | F_H | F_C, (
-            (v ? 0 : F_Z)
-          ));
-        }
+        or_d8(ctx, old_pc + 1);
     - id: RST 30H
       hex: 0xF7
       cat: "jumps/calls"
@@ -3886,6 +3653,8 @@ ops:
         n:
         h:
         c:
+      code: |
+        rst(ctx, 0x30);
     - id: LD HL, SP+r8
       hex: 0xF8
       cat: "16-bit load/store/move"
@@ -3967,20 +3736,7 @@ ops:
         h: H
         c: C
       code: |
-        {
-          // get old and new value
-          const uint8_t o = cpu_rb(ctx, RB_A),
-                        n = mmu_rb(ctx, old_pc + 1),
-                        v = o - n;
-
-          // set flags
-          cpu_wf(ctx, F_Z | F_N | F_H | F_C, (
-            (v ? 0 : F_Z) |
-            F_N |
-            ((((o & 0x0F) - (n & 0x0F)) & 0xF0) ? F_H : 0) |
-            ((n > o) ? F_C : 0)
-          ));
-        }
+        cp_d8(ctx, old_pc + 1);
     - id: RST 38H
       hex: 0xFF
       cat: "jumps/calls"
@@ -3994,6 +3750,8 @@ ops:
         n:
         h:
         c:
+      code: |
+        rst(ctx, 0x38);
 
   cb:
     - id: RLC B
@@ -7941,9 +7699,9 @@ templates:
     ) {
       const uint8_t of = cpu_rb(ctx, RB_F);
       cpu_wb(ctx, RB_F, (
-        (((mask & F_Z) ? flags : of) & F_Z) | 
-        (((mask & F_N) ? flags : of) & F_N) | 
-        (((mask & F_H) ? flags : of) & F_H) | 
+        (((mask & F_Z) ? flags : of) & F_Z) |
+        (((mask & F_N) ? flags : of) & F_N) |
+        (((mask & F_H) ? flags : of) & F_H) |
         (((mask & F_C) ? flags : of) & F_C)
       ));
     }
@@ -7959,8 +7717,8 @@ templates:
 
     static void
     not_implemented(
-      gb_t * const ctx, 
-      const uint16_t pc, 
+      gb_t * const ctx,
+      const uint16_t pc,
       const uint16_t op
     ) {
       UNUSED(op);
@@ -7968,15 +7726,6 @@ templates:
       cpu_ww(ctx, RW_PC, pc); // FIXME?
     }
 
-    static uint16_t
-    gb_next_op(
-      const gb_t * const ctx
-    ) {
-      const uint16_t addr = cpu_rw(ctx, RW_PC);
-      const uint8_t op = mmu_rb(ctx, addr);
-      return (op == 0xCB) ? 0xCB00 | mmu_rb(ctx, addr + 1) : op;
-    }
-
     static void
     inc_rb(
       gb_t * const ctx,
@@ -7991,7 +7740,26 @@ templates:
 
       // set flags
       cpu_wf(ctx, F_Z | F_N | F_H, (
-        (n ? F_Z : 0) | 
+        (n ? F_Z : 0) |
+        (((o & 0x0F) == 0x0F) ? F_H : 0)
+      ));
+    }
+
+    static void
+    inc_hl_ptr(
+      gb_t * const ctx
+    ) {
+      // write flags
+      const uint16_t addr = cpu_rw(ctx, RW_HL);
+      const uint8_t o = mmu_rb(ctx, addr),
+                    n = o + 1;
+
+      // write value
+      mmu_wb(ctx, addr, n);
+
+      // write flags
+      cpu_wf(ctx, F_Z | F_N | F_H, (
+        (n ? 0 : F_Z) |
         (((o & 0x0F) == 0x0F) ? F_H : 0)
       ));
     }
@@ -8010,13 +7778,33 @@ templates:
 
       // set flags
       cpu_wf(ctx, F_Z | F_N | F_H, (
-        (n ? F_Z : 0) | 
-        F_N | 
+        (n ? F_Z : 0) |
+        F_N |
         (((o ^ n) & 0x10) ? F_H : 0)
       ));
     }
 
     static void
+    dec_hl_ptr(
+      gb_t * const ctx
+    ) {
+      // write flags
+      const uint16_t addr = cpu_rw(ctx, RW_HL);
+      const uint8_t o = mmu_rb(ctx, addr),
+                    n = o - 1;
+
+      // write value
+      mmu_wb(ctx, addr, n);
+
+      // write flags
+      cpu_wf(ctx, F_Z | F_N | F_H, (
+        (n ? 0 : F_Z) |
+        F_N |
+        (((o & 0x0F) == 0x00) ? F_H : 0)
+      ));
+    }
+
+    static void
     add_rw_rw(
       gb_t * const ctx,
       const rw_t dst_reg,
@@ -8026,7 +7814,7 @@ templates:
       const uint16_t dst_val = cpu_rw(ctx, dst_reg),
                      src_val = cpu_rw(ctx, src_reg);
       const uint32_t n = ((uint32_t) dst_val) + ((uint32_t) src_val);
-      
+
       // set new value
       cpu_ww(ctx, dst_reg, (uint16_t) (n & 0xFFFF));
 
@@ -8059,6 +7847,47 @@ templates:
     }
 
     static void
+    add_d8(
+      gb_t * const ctx,
+      const uint16_t addr
+    ) {
+      // get old and new value
+      const uint8_t o = cpu_rb(ctx, RB_A),
+                    n = mmu_rb(ctx, addr);
+      const uint16_t v = o + n;
+
+      // write value
+      cpu_wb(ctx, RB_A, v);
+
+      // set flags
+      cpu_wf(ctx, F_Z | F_N | F_H | F_C, (
+        ((v & 0xFF) ? 0 : F_Z) |
+        ((((o & 0x0F) + (n & 0x0F)) & 0xF0) ? F_H : 0) |
+        ((v & 0x100) ? F_C : 0)
+      ));
+    }
+
+    static void
+    add_hl_ptr(
+      gb_t * const ctx
+    ) {
+      // get old and new value
+      const uint8_t o = cpu_rb(ctx, RB_A),
+                    n = mmu_rw(ctx, cpu_rw(ctx, RW_HL));
+      const uint16_t v = o + n;
+
+      // write value
+      cpu_wb(ctx, RB_A, v);
+
+      // set flags
+      cpu_wf(ctx, F_Z | F_N | F_H | F_C, (
+        ((v & 0xFF) ? 0 : F_Z) |
+        ((((o & 0x0F) + (n & 0x0F)) & 0xF0) ? F_H : 0) |
+        ((v & 0x100) ? F_C : 0)
+      ));
+    }
+
+    static void
     adc_rb(
       gb_t * const ctx,
       const rb_t reg
@@ -8081,6 +7910,27 @@ templates:
     }
 
     static void
+    adc_hl_ptr(
+      gb_t * const ctx
+    ) {
+      // get old and new value
+      const uint8_t o = cpu_rb(ctx, RB_A),
+                    n = cpu_rb(ctx, mmu_rb(ctx, cpu_rw(ctx, RW_HL))),
+                    c = FLAG(ctx, C) ? 1 : 0;
+      const uint16_t v = o + n + c;
+
+      // write value
+      cpu_wb(ctx, RB_A, v);
+
+      // set flags
+      cpu_wf(ctx, F_Z | F_N | F_H | F_C, (
+        ((v & 0xFF) ? 0 : F_Z) |
+        ((((o & 0x0F) + (n & 0x0F) + c) & 0xF0) ? F_H : 0) |
+        ((v & 0x100) ? F_C : 0)
+      ));
+    }
+
+    static void
     sub_rb(
       gb_t * const ctx,
       const rb_t reg
@@ -8103,6 +7953,27 @@ templates:
     }
 
     static void
+    sub_hl_ptr(
+      gb_t * const ctx
+    ) {
+      // get old and new value
+      const uint8_t o = cpu_rb(ctx, RB_A),
+                    n = mmu_rb(ctx, cpu_rw(ctx, RW_HL)),
+                    v = o - n;
+
+      // write value
+      cpu_wb(ctx, RB_A, v);
+
+      // set flags
+      cpu_wf(ctx, F_Z | F_N | F_H | F_C, (
+        (v ? 0 : F_Z) |
+        F_N |
+        ((((o & 0x0F) - (n & 0x0F)) & 0xF0) ? F_H : 0) |
+        ((n > o) ? F_C : 0)
+      ));
+    }
+
+    static void
     sub_d8(
       gb_t * const ctx,
       const uint16_t addr
@@ -8148,6 +8019,28 @@ templates:
     }
 
     static void
+    sbc_hl_ptr(
+      gb_t * const ctx
+    ) {
+      // get old and new value
+      const uint8_t o = cpu_rb(ctx, RB_A),
+                    n = mmu_rb(ctx, cpu_rw(ctx, RW_HL)),
+                    c = FLAG(ctx, C) ? 1 : 0,
+                    v = o - (n + c);
+
+      // write value
+      cpu_wb(ctx, RB_A, v);
+
+      // set flags
+      cpu_wf(ctx, F_Z | F_N | F_H | F_C, (
+        (v ? 0 : F_Z) |
+        F_N |
+        ((((o & 0x0F) - (n & 0x0F)) & 0xF0) ? F_H : 0) |
+        ((n > o) ? F_C : 0) // FIXME: is this right for sub/sbc?
+      ));
+    }
+
+    static void
     and_rb(
       gb_t * const ctx,
       const rb_t reg
@@ -8160,7 +8053,24 @@ templates:
 
       // set flags
       cpu_wf(ctx, F_Z | F_N | F_H | F_C, (
-        (v ? 0 : F_Z) | 
+        (v ? 0 : F_Z) |
+        F_H
+      ));
+    }
+
+    static void
+    and_hl_ptr(
+      gb_t * const ctx
+    ) {
+      // get value
+      const uint8_t v = cpu_rb(ctx, RB_A) & mmu_rb(ctx, cpu_rb(ctx, RW_HL));
+
+      // write value
+      cpu_wb(ctx, RB_A, v);
+
+      // set flags
+      cpu_wf(ctx, F_Z | F_N | F_H | F_C, (
+        (v ? 0 : F_Z) |
         F_H
       ));
     }
@@ -8183,6 +8093,22 @@ templates:
     }
 
     static void
+    xor_hl_ptr(
+      gb_t * const ctx
+    ) {
+      // get value
+      const uint8_t v = cpu_rb(ctx, RB_A) ^ mmu_rb(ctx, cpu_rw(ctx, RW_HL));
+
+      // write value
+      cpu_wb(ctx, RB_A, v);
+
+      // set flags
+      cpu_wf(ctx, F_Z | F_N | F_H | F_C, (
+        (v ? 0 : F_Z)
+      ));
+    }
+
+    static void
     or_rb(
       gb_t * const ctx,
       const rb_t reg
@@ -8198,6 +8124,39 @@ templates:
     }
 
     static void
+    or_hl_ptr(
+      gb_t * const ctx
+    ) {
+      // get value
+      const uint8_t v = cpu_rb(ctx, RB_A) | mmu_rb(ctx, cpu_rw(ctx, RW_HL));
+
+      // write value
+      cpu_wb(ctx, RB_A, v);
+
+      // set flags
+      cpu_wf(ctx, F_Z | F_N | F_H | F_C, (
+        (v ? 0 : F_Z)
+      ));
+    }
+
+    static void
+    or_d8(
+      gb_t * const ctx,
+      const uint16_t addr
+    ) {
+      // get value
+      const uint8_t v = cpu_rb(ctx, RB_A) ^ mmu_rb(ctx, addr);
+
+      // write value
+      cpu_wb(ctx, RB_A, v);
+
+      // set flags
+      cpu_wf(ctx, F_Z | F_N | F_H | F_C, (
+        (v ? 0 : F_Z)
+      ));
+    }
+
+    static void
     cp_rb(
       gb_t * const ctx,
       const rb_t reg
@@ -8217,6 +8176,43 @@ templates:
     }
 
     static void
+    cp_hl_ptr(
+      gb_t * const ctx
+    ) {
+      // get old and new value
+      const uint8_t o = cpu_rb(ctx, RB_A),
+                    n = mmu_rb(ctx, cpu_rw(ctx, RW_HL)),
+                    v = o - n;
+
+      // set flags
+      cpu_wf(ctx, F_Z | F_N | F_H | F_C, (
+        (v ? 0 : F_Z) |
+        F_N |
+        ((((o & 0x0F) - (n & 0x0F)) & 0xF0) ? F_H : 0) |
+        ((n > o) ? F_C : 0)
+      ));
+    }
+
+    static void
+    cp_d8(
+      gb_t * const ctx,
+      const uint16_t addr
+    ) {
+      // get old and new value
+      const uint8_t o = cpu_rb(ctx, RB_A),
+                    n = mmu_rb(ctx, addr),
+                    v = o - n;
+
+      // set flags
+      cpu_wf(ctx, F_Z | F_N | F_H | F_C, (
+        (v ? 0 : F_Z) |
+        F_N |
+        ((((o & 0x0F) - (n & 0x0F)) & 0xF0) ? F_H : 0) |
+        ((n > o) ? F_C : 0)
+      ));
+    }
+
+    static void
     push_rw(
       gb_t * const ctx,
       const rw_t reg
@@ -8405,7 +8401,7 @@ templates:
       // set flags
       cpu_wf(ctx, F_Z | F_N | F_H, (
         (v ? 0 : F_Z) |
-        0 | 
+        0 |
         F_H
       ));
     }
@@ -8422,7 +8418,7 @@ templates:
       // set flags
       cpu_wf(ctx, F_Z | F_N | F_H, (
         (v ? 0 : F_Z) |
-        0 | 
+        0 |
         F_H
       ));
     }
@@ -8485,17 +8481,58 @@ templates:
       ));
     }
 
+    static void
+    jr(
+      gb_t * const ctx,
+      const uint16_t ofs_addr
+    ) {
+      // get offset
+      const int8_t ofs = (int8_t) mmu_rb(ctx, ofs_addr);
+
+      // jump
+      cpu_ww(ctx, RW_PC, cpu_rw(ctx, RW_PC) + ofs);
+    }
+
+    static void
+    rst(
+      gb_t * const ctx,
+      const uint8_t val
+    ) {
+      UNUSED(ctx);
+      UNUSED(val);
+    }
+
+    static void
+    call_a16(
+      gb_t * const ctx,
+      const uint16_t addr
+    ) {
+      // push pc
+      push_rw(ctx, RW_PC);
+
+      // jump
+      cpu_ww(ctx, RW_PC, mmu_rw(ctx, addr));
+    }
+
     void
     gb_step(
       gb_t * const ctx
     ) {
       const uint16_t old_pc = cpu_rw(ctx, RW_PC),
-                     op = gb_next_op(ctx);
+                     op = mmu_rb(ctx, old_pc);
 
       // TODO: handle interrupts
 
       switch (op) {
-      <%= switches %>
+      <%= switches['main'] %>
+      case 0xCB:
+        switch (mmu_rb(ctx, old_pc + 1)) {
+        <%= switches['cb'] %>
+        default:
+          not_implemented(ctx, old_pc, op);
+        }
+
+        break;
       default:
         not_implemented(ctx, old_pc, op);
       }