DMA: Increase precision for large transfers

Also gets rid of the delay on the GPU side for writing to VRAM (doesn't
make sense), and it's not needed since we slice the block transfers now.

Fixes palette corruption in Vigilante 8, and missing rider in
Championship Motocross 2001 featuring Ricky Carmichael.

data/database/gamesettings.ini

@@ -597,3 +597,45 @@ DMAMaxSliceTicks = 100
 # SLUS-00912 (Destruction Derby Raw (USA))
 [SLUS-00912]
 ForceInterlacing = true
+
+
+# SLUS-00510 (Vigilante 8 (USA) (Rev 1))
+[SLUS-00510]
+DMAMaxSliceTicks = 300
+DMAHaltTicks = 100
+
+
+# SLES-01212 (Vigilante 8 (Europe))
+[SLES-01212]
+DMAMaxSliceTicks = 300
+DMAHaltTicks = 100
+
+
+# SLES-01214 (Vigilante 8 (Germany))
+[SLES-01214]
+DMAMaxSliceTicks = 300
+DMAHaltTicks = 100
+
+
+# SLES-01213 (Vigilante 8 (France))
+[SLES-01213]
+DMAMaxSliceTicks = 300
+DMAHaltTicks = 100
+
+
+# SLES-01215 (Vigilante 8 (Italy))
+[SLES-01215]
+DMAMaxSliceTicks = 300
+DMAHaltTicks = 100
+
+
+# SLPS-01703 (Vigilante 8 (Japan))
+[SLPS-01703]
+DMAMaxSliceTicks = 300
+DMAHaltTicks = 100
+
+
+# SLPS-91182 (Vigilante 8 (Japan))
+[SLPS-91182]
+DMAMaxSliceTicks = 300
+DMAHaltTicks = 100

src/core/dma.cpp

@@ -284,7 +284,6 @@ bool DMA::TransferChannel(Channel channel)
 
     case SyncMode::LinkedList:
     {
-      TickCount used_ticks = 0;
       if (!copy_to_device)
       {
         Panic("Linked list not implemented for DMA reads");
@@ -295,12 +294,13 @@ bool DMA::TransferChannel(Channel channel)
                       current_address & ADDRESS_MASK);
 
       u8* ram_pointer = Bus::g_ram;
-      bool halt_transfer = false;
-      while (cs.request)
+      TickCount remaining_ticks = m_max_slice_ticks;
+      while (cs.request && remaining_ticks > 0)
       {
         u32 header;
         std::memcpy(&header, &ram_pointer[current_address & ADDRESS_MASK], sizeof(header));
-        used_ticks += 10;
+        CPU::AddPendingTicks(10);
+        remaining_ticks -= 10;
 
         const u32 word_count = header >> 24;
         const u32 next_address = header & UINT32_C(0x00FFFFFF);
@@ -308,35 +308,26 @@ bool DMA::TransferChannel(Channel channel)
                         word_count * UINT32_C(4), word_count, next_address);
         if (word_count > 0)
         {
-          used_ticks += 5;
-          used_ticks +=
+          CPU::AddPendingTicks(5);
+          remaining_ticks -= 5;
+
+          const TickCount block_ticks =
             TransferMemoryToDevice(channel, (current_address + sizeof(header)) & ADDRESS_MASK, 4, word_count);
-        }
-        else if ((current_address & ADDRESS_MASK) == (next_address & ADDRESS_MASK))
-        {
-          current_address = next_address;
-          halt_transfer = true;
-          break;
+          CPU::AddPendingTicks(block_ticks);
+          remaining_ticks -= block_ticks;
         }
 
         current_address = next_address;
         if (current_address & UINT32_C(0x800000))
           break;
-
-        if (used_ticks >= m_max_slice_ticks)
-        {
-          halt_transfer = true;
-          break;
-        }
       }
 
       cs.base_address = current_address;
-      CPU::AddPendingTicks(used_ticks);
 
       if (current_address & UINT32_C(0x800000))
         break;
 
-      if (halt_transfer)
+      if (cs.request)
       {
         // stall the transfer for a bit if we ran for too long
         HaltTransfer(m_halt_ticks);
@@ -359,35 +350,55 @@ bool DMA::TransferChannel(Channel channel)
 
       const u32 block_size = cs.block_control.request.GetBlockSize();
       u32 blocks_remaining = cs.block_control.request.GetBlockCount();
-      TickCount used_ticks = 0;
+      TickCount ticks_remaining = m_max_slice_ticks;
 
       if (copy_to_device)
       {
         do
         {
           blocks_remaining--;
-          used_ticks += TransferMemoryToDevice(channel, current_address & ADDRESS_MASK, increment, block_size);
+
+          const TickCount ticks =
+            TransferMemoryToDevice(channel, current_address & ADDRESS_MASK, increment, block_size);
+          CPU::AddPendingTicks(ticks);
+          ticks_remaining -= ticks;
+
           current_address = (current_address + (increment * block_size));
-        } while (cs.request && blocks_remaining > 0);
+        } while (cs.request && blocks_remaining > 0 && ticks_remaining > 0);
       }
       else
       {
         do
         {
           blocks_remaining--;
-          used_ticks += TransferDeviceToMemory(channel, current_address & ADDRESS_MASK, increment, block_size);
+
+          const TickCount ticks =
+            TransferDeviceToMemory(channel, current_address & ADDRESS_MASK, increment, block_size);
+          CPU::AddPendingTicks(ticks);
+          ticks_remaining -= ticks;
+
           current_address = (current_address + (increment * block_size));
-        } while (cs.request && blocks_remaining > 0);
+        } while (cs.request && blocks_remaining > 0 && ticks_remaining > 0);
       }
 
       cs.base_address = current_address & BASE_ADDRESS_MASK;
       cs.block_control.request.block_count = blocks_remaining;
-      CPU::AddPendingTicks(used_ticks);
 
       // finish transfer later if the request was cleared
       if (blocks_remaining > 0)
+      {
+        if (cs.request)
+        {
+          // we got halted
+          if (!m_unhalt_event->IsActive())
+            HaltTransfer(m_halt_ticks);
+
+          return false;
+        }
+
         return true;
       }
+    }
     break;
 
     default:

src/core/gpu.h

@@ -167,7 +167,7 @@ protected:
   // The GPU internally appears to run at 2x the system clock.
   ALWAYS_INLINE static constexpr TickCount GPUTicksToSystemTicks(TickCount gpu_ticks)
   {
-    return std::max<TickCount>(gpu_ticks >> 1, 1);
+    return std::max<TickCount>((gpu_ticks + 1) >> 1, 1);
   }
   ALWAYS_INLINE static constexpr TickCount SystemTicksToGPUTicks(TickCount sysclk_ticks) { return sysclk_ticks << 1; }
 

src/core/gpu_commands.cpp

@@ -49,7 +49,6 @@ void GPU::ExecuteCommands()
           for (u32 i = 0; i < words_to_copy; i++)
             m_blit_buffer.push_back(FifoPop());
           m_blit_remaining_words -= words_to_copy;
-          AddCommandTicks(words_to_copy);
 
           Log_DebugPrintf("VRAM write burst of %u words, %u words remaining", words_to_copy, m_blit_remaining_words);
           if (m_blit_remaining_words == 0)