Version 1.1 - See readme.md for details.

Makefile

@@ -9,7 +9,7 @@ include $(DEVKITARM)/ds_rules
 
 export TARGET		:=	GimliDS
 export TOPDIR		:=	$(CURDIR)
-export VERSION		:=  1.0a
+export VERSION		:=  1.1
 
 ICON 		:= -b $(CURDIR)/C64_icon.bmp "GimliDS $(VERSION);wavemotion-dave;https://github.com/wavemotion-dave/GimliDS" 
 

README.md

@@ -147,6 +147,12 @@ Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
 ```
 ## Change Log
 
+Version 1.1 release 11-May-2025 by wavemotion-dave
+  * Fix to CIA vars initialization so games like Chuckie Egg start properly.
+  * Fix to TOD clock counters (previous versions would not clock at all... fixes some utilities mostly).
+  * Improved REU handling for a bit more speed (gain of 1-2 frames on REU games)
+  * Improved VIC handling for a bit more speed (gain of 1-2 frames across the board)
+
 Version 1.0 release 09-May-2025 by wavemotion-dave
   * Added TOGGLE ZOOM button map to scale screen 1:1 temporarily (to make text games readable)
   * Another proper fix for CPU jump commands not handling the extra cycle it takes.

arm9/source/C64.cpp

@@ -1351,7 +1351,7 @@ void C64::main_loop(void)
         // -----------------------------------------------------------------
         if (ThePrefs.TrueDrive)
         {
-            int cycles_1541 = FLOPPY_CYCLES_PER_LINE + CycleDeltas[myConfig.flopCycles];
+            int cycles_1541 = FLOPPY_CYCLES_PER_LINE;
             TheCPU1541->CountVIATimers(cycles_1541);
 
             if (!TheCPU1541->Idle)

arm9/source/CIA.cpp

@@ -84,7 +84,7 @@ void MOS6526::Reset(void)
     pra = prb = ddra = ddrb = 0;
 
     ta = tb = 0xffff;
-    latcha = latchb = 1;
+    latcha = latchb = 0xffff;
 
     tod_10ths = tod_sec = tod_min = 0; tod_hr = 1;
     alm_10ths = alm_sec = alm_min = alm_hr = 0;
@@ -559,7 +559,6 @@ void MOS6526::CountTOD(void)
         tod_divider--;
     else
     {
-
         // Reload divider according to 50/60 Hz flag
         if (cra & 0x80)
             tod_divider = 4;

arm9/source/CPU_emulline.h

@@ -947,7 +947,7 @@
         case 0x7c:
         case 0xdc:
         case 0xfc:
-#if PRECISE_CPU_CYCLES
+#if PRECISE_CPU_CYCLES_NO // Save the effort and we need the ITCM_CODE space
             read_byte_abs_x();
 #else
             pc+=2;

arm9/source/REU.cpp

@@ -51,8 +51,11 @@
 
 extern u8 myRAM[];
 
+// We are emulating the REU-1764 (256K) only
 u8 REU_RAM[256 * 1024];
 
+#define RAM_MASK    0x3FFFF
+
 /*
  *  REU constructor
  */
@@ -61,7 +64,6 @@ REU::REU(MOS6510 * cpu) : the_cpu(cpu)
 {
     // Allocate expansion RAM
     ram_size = 0x40000; // 256K only
-    ram_mask = ram_size - 1;
     
     ex_ram = REU_RAM;
 
@@ -208,6 +210,11 @@ void REU::WriteIO2(uint16_t adr, uint8_t byte)
             {
                 execute_dma();
             } 
+            if ((byte & 0x90) == 0x80) 
+            {
+                // This would normally be the 'delayed' write... not fully handled by GimliDS as we can't afford the trigger check on FF00
+                execute_dma();
+            }
             break;
         default:
             regs[adr & 0xf] = byte;
@@ -236,45 +243,44 @@ void REU::execute_dma()
     // Get transfer length
     uint16_t length = regs[7] | (regs[8] << 8);
 
-    // Calculate address increments
+    // We are skipping the memory increment of regs[10] as there
-    unsigned c64_inc = (regs[10] & 0x80) ? 0 : 1;
+    // are no known practical uses of it... and we need the speed.
-    unsigned reu_inc = (regs[10] & 0x40) ? 0 : 1;
     
-    // Do transfer
+    if ((regs[1] & 3) == 0)
-    bool verify_error = false;
-    while (!verify_error) 
     {
-        switch (regs[1] & 3) 
+        while (length--)
         {
-            case 0:     // C64 -> REU
+            REU_RAM[reu_adr++ & RAM_MASK] = the_cpu->REUReadByte(c64_adr++);
-                ex_ram[reu_adr & ram_mask] = the_cpu->REUReadByte(c64_adr);
-                break;
-            case 1:     // C64 <- REU
-                the_cpu->REUWriteByte(c64_adr, ex_ram[reu_adr & ram_mask]);
-                break;
-            case 2: {   // C64 <-> REU
-                uint8_t tmp = the_cpu->REUReadByte(c64_adr);
-                the_cpu->REUWriteByte(c64_adr, ex_ram[reu_adr & ram_mask]);
-                ex_ram[reu_adr & ram_mask] = tmp;
-                break;
         }
-            case 3:     // Compare
+    }
-                if (ex_ram[reu_adr & ram_mask] != the_cpu->REUReadByte(c64_adr)) 
+    else if ((regs[1] & 3) == 1)
+    {
+        while (length--)
+        {
+            the_cpu->REUWriteByte(c64_adr++, REU_RAM[reu_adr++ & RAM_MASK]);
+        }
+    }
+    else if ((regs[1] & 3) == 2)
+    {
+        while (length--)
+        {
+            uint8_t tmp = the_cpu->REUReadByte(c64_adr);
+            the_cpu->REUWriteByte(c64_adr++, REU_RAM[reu_adr & RAM_MASK]);
+            REU_RAM[reu_adr++ & RAM_MASK] = tmp;
+        }
+    }
+    else // Compare
+    {
+        while (length--) 
+        {
+            if (REU_RAM[reu_adr++ & RAM_MASK] != the_cpu->REUReadByte(c64_adr++)) 
             {
                 regs[0] |= 0x20;    // Verify error
-                    verify_error = true;
-                }
                 break;
             }
-
+        }
-        c64_adr += c64_inc;
+    }
-        reu_adr += reu_inc;
-        if (--length == 0)
-        {
     regs[0] |= 0x40;    // Transfer finished
-            break;
-        }
-    }
 
     // Update address and length registers
     if (regs[1] & 0x20)
@@ -289,7 +295,7 @@ void REU::execute_dma()
     }
     else 
     {
-        reu_adr &= ram_mask;
+        reu_adr &= RAM_MASK;
         regs[2] = c64_adr & 0xff;
         regs[3] = c64_adr >> 8;
         regs[4] = reu_adr & 0xff;
@@ -308,7 +314,7 @@ void REU::execute_dma()
 void REU::GetState(REUState *rs)
 {
     rs->ram_size              = ram_size;
-    rs->ram_mask              = ram_mask;
+    rs->ram_mask              = RAM_MASK;
     rs->autoload_c64_adr_lo   = autoload_c64_adr_lo;
     rs->autoload_c64_adr_hi   = autoload_c64_adr_hi;
     rs->autoload_reu_adr_lo   = autoload_reu_adr_lo;
@@ -327,7 +333,7 @@ void REU::GetState(REUState *rs)
 void REU::SetState(REUState *rs)
 {
     ram_size              = rs->ram_size;
-    ram_size              = rs->ram_mask;
+    ram_mask              = RAM_MASK;
     autoload_c64_adr_lo   = rs->autoload_c64_adr_lo;
     autoload_c64_adr_hi   = rs->autoload_c64_adr_hi;
     autoload_reu_adr_lo   = rs->autoload_reu_adr_lo;

arm9/source/VIC.cpp

@@ -1018,7 +1018,7 @@ __attribute__ ((noinline))  ITCM_CODE void MOS6569::el_ecm_text(uint8 *p, uint8
 }
 
 
-void MOS6569::el_std_idle(uint8 *p, uint8 *r)
+__attribute__ ((noinline))  ITCM_CODE void MOS6569::el_std_idle(uint8 *p, uint8 *r)
 {
     uint8 data = *get_physical(ctrl1 & 0x40 ? 0x39ff : 0x3fff);
     uint32 *lp = (uint32 *)p;
@@ -1431,7 +1431,7 @@ int MOS6569::EmulateLine(void)
         if (raster >= FIRST_DMA_LINE && raster <= LAST_DMA_LINE && ((raster & 7) == y_scroll) && bad_lines_enabled) 
         {
             is_bad_line = true;
-            cycles_left = BAD_CYCLES_PER_LINE + CycleDeltas[myConfig.cpuCycles];
+            cycles_left = BAD_CYCLES_PER_LINE + CycleDeltas[myConfig.cpuCycles] + CycleDeltas[myConfig.badCycles];
         }
         goto VIC_nop;
     }
@@ -1452,7 +1452,7 @@ int MOS6569::EmulateLine(void)
         {
             // Turn on display
             display_state = is_bad_line = true;
-            cycles_left = BAD_CYCLES_PER_LINE + CycleDeltas[myConfig.cpuCycles];
+            cycles_left = BAD_CYCLES_PER_LINE + CycleDeltas[myConfig.cpuCycles] + CycleDeltas[myConfig.badCycles];
             rc = 0;
 
             // Read and latch 40 bytes from video matrix and color RAM

arm9/source/mainmenu.cpp

@@ -386,7 +386,7 @@ void SetDefaultGameConfig(void)
     myConfig.reserved7   = 0;
     myConfig.reserved8   = 0xA5;             // So it's easy to spot on an "upgrade" and we can re-default it
     myConfig.cpuCycles   = 0;                // Normal 63 - this is the delta adjustment to that
-    myConfig.flopCycles  = 0;                // Normal 64 - this is the delta adjustment to that
+    myConfig.badCycles   = 0;                // Normal 23 - this is the delta adjustment to that
     
     myConfig.offsetX     = 32;              // Push the side border off the main display
     myConfig.offsetY     = 19;              // Push the top border off the main display
@@ -394,7 +394,7 @@ void SetDefaultGameConfig(void)
     myConfig.scaleY      = 200;             // Scale the 200 pixels of C64 display to the DS 200 (yes, there is only 192 so this will cut... use PAN UP/DN)
 }
 
-s16 CycleDeltas[] = {0,1,2,3,4,5,6,7,8,9,-9,-8,-7,-6,-5,-4,-3,-2,-1};  // Used with myConfig.cpuCycles and myConfig.flopCycles
+s16 CycleDeltas[] = {0,1,2,3,4,5,6,7,8,9,-9,-8,-7,-6,-5,-4,-3,-2,-1};  // Used with myConfig.cpuCycles and myConfig.badCycles
 
 // ----------------------------------------------------------------------
 // Read file twice and ensure we get the same CRC... if not, do it again
@@ -592,7 +592,7 @@ const struct options_t Option_Table[1][20] =
         {"LCD JITTER",     {"NONE", "LIGHT", "HEAVY"},                                                  &myConfig.jitter,      3},
         {"DISK SOUND",     {"SFX OFF", "SFX ON"},                                                       &myConfig.diskSFX,     2},
         {"CPU CYCLES",     {CYCLE_DELTA_STR},                                                           &myConfig.cpuCycles,   19},
-        {"1541 CYCLES",    {CYCLE_DELTA_STR},                                                           &myConfig.flopCycles,  19},
+        {"BAD CYCLES" ,    {CYCLE_DELTA_STR},                                                           &myConfig.badCycles,   19},
         {"POUND KEY",      {"POUND", "LEFT ARROW", "UP ARROW", "C= COMMODORE"},                         &myConfig.poundKey,    4},        
 
         {"D-PAD UP",       {KEY_MAP_OPTIONS},                                                           &myConfig.key_map[0],  65},

arm9/source/mainmenu.h

@@ -21,7 +21,7 @@ struct __attribute__((__packed__)) Config_t
     u8  reserved7;
     u8  reserved8;
     u8  cpuCycles;
-    u8  flopCycles;
+    u8  badCycles;
     s8  offsetX;
     s8  offsetY;
     s16 scaleX;