X-Git-Url: https://notaz.gp2x.de/cgi-bin/gitweb.cgi?p=pcsx_rearmed.git;a=blobdiff_plain;f=plugins%2Fgpu_neon%2Fgpu.c;h=c31d9a1fe4ce4ca1832fd486dce2a2e965deba12;hp=339cdfdc0aae576c770d0152e0f3a7ea01e5d02e;hb=05740673341dbd61b23085c8cf1303f632e44f87;hpb=5440b88ee29ee5b8929fcf338dfc8f476468514d

diff --git a/plugins/gpu_neon/gpu.c b/plugins/gpu_neon/gpu.c
index 339cdfdc..c31d9a1f 100644
--- a/plugins/gpu_neon/gpu.c
+++ b/plugins/gpu_neon/gpu.c
@@ -1,5 +1,5 @@
 /*
- * (C) Gražvydas "notaz" Ignotas, 2011
+ * (C) Gražvydas "notaz" Ignotas, 2011-2012
  *
  * This work is licensed under the terms of any of these licenses
  * (at your option):
@@ -26,10 +26,29 @@
 
 struct psx_gpu gpu __attribute__((aligned(2048)));
 
+static noinline int do_cmd_buffer(uint32_t *data, int count);
+static void finish_vram_transfer(int is_read);
+
+static noinline void do_cmd_reset(void)
+{
+  if (unlikely(gpu.cmd_len > 0))
+    do_cmd_buffer(gpu.cmd_buffer, gpu.cmd_len);
+  gpu.cmd_len = 0;
+
+  if (unlikely(gpu.dma.h > 0))
+    finish_vram_transfer(gpu.dma_start.is_read);
+  gpu.dma.h = 0;
+}
+
 static noinline void do_reset(void)
 {
+  int i;
+
+  do_cmd_reset();
+
   memset(gpu.regs, 0, sizeof(gpu.regs));
-  memset(gpu.ex_regs, 0, sizeof(gpu.ex_regs));
+  for (i = 0; i < sizeof(gpu.ex_regs) / sizeof(gpu.ex_regs[0]); i++)
+    gpu.ex_regs[i] = (0xe0 + i) << 24;
   gpu.status.reg = 0x14802000;
   gpu.gp0 = 0;
   gpu.regs[3] = 1;
@@ -115,7 +134,10 @@ long GPUinit(void)
 
   gpu.state.frame_count = &gpu.zero;
   gpu.state.hcnt = &gpu.zero;
+  gpu.frameskip.active = 0;
+  gpu.cmd_len = 0;
   do_reset();
+
   return ret;
 }
 
@@ -131,7 +153,7 @@ void GPUwriteStatus(uint32_t data)
   uint32_t cmd = data >> 24;
 
   if (cmd < ARRAY_SIZE(gpu.regs)) {
-    if (cmd != 0 && cmd != 5 && gpu.regs[cmd] == data)
+    if (cmd > 1 && cmd != 5 && gpu.regs[cmd] == data)
       return;
     gpu.regs[cmd] = data;
   }
@@ -142,6 +164,9 @@ void GPUwriteStatus(uint32_t data)
     case 0x00:
       do_reset();
       break;
+    case 0x01:
+      do_cmd_reset();
+      break;
     case 0x03:
       gpu.status.blanking = data & 1;
       break;
@@ -189,9 +214,9 @@ const unsigned char cmd_lengths[256] =
 	0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
 	3, 3, 3, 3, 6, 6, 6, 6, 4, 4, 4, 4, 8, 8, 8, 8, // 20
 	5, 5, 5, 5, 8, 8, 8, 8, 7, 7, 7, 7, 11, 11, 11, 11,
-	2, 2, 2, 2, 0, 0, 0, 0, 3, 3, 3, 3, 3, 3, 3, 3, // 40
-	3, 3, 3, 3, 0, 0, 0, 0, 4, 4, 4, 4, 4, 4, 4, 4,
-	2, 2, 2, 2, 3, 3, 3, 3, 1, 1, 1, 1, 2, 2, 2, 2, // 60
+	2, 2, 2, 2, 2, 2, 2, 2, 3, 3, 3, 3, 3, 3, 3, 3, // 40
+	3, 3, 3, 3, 3, 3, 3, 3, 4, 4, 4, 4, 4, 4, 4, 4,
+	2, 2, 2, 2, 3, 3, 3, 3, 1, 1, 1, 1, 0, 0, 0, 0, // 60
 	1, 1, 1, 1, 2, 2, 2, 2, 1, 1, 1, 1, 2, 2, 2, 2,
 	3, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, // 80
 	0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
@@ -247,12 +272,16 @@ static int do_vram_io(uint32_t *data, int count, int is_read)
     do_vram_line(x, y, sdata, w, is_read);
   }
 
-  if (h > 0 && count > 0) {
-    y &= 511;
-    do_vram_line(x, y, sdata, count, is_read);
-    o = count;
-    count = 0;
+  if (h > 0) {
+    if (count > 0) {
+      y &= 511;
+      do_vram_line(x, y, sdata, count, is_read);
+      o = count;
+      count = 0;
+    }
   }
+  else
+    finish_vram_transfer(is_read);
   gpu.dma.y = y;
   gpu.dma.h = h;
   gpu.dma.offset = o;
@@ -267,9 +296,11 @@ static void start_vram_transfer(uint32_t pos_word, uint32_t size_word, int is_re
 
   gpu.dma.x = pos_word & 0x3ff;
   gpu.dma.y = (pos_word >> 16) & 0x1ff;
-  gpu.dma.w = size_word & 0x3ff;
-  gpu.dma.h = (size_word >> 16) & 0x1ff;
+  gpu.dma.w = ((size_word - 1) & 0x3ff) + 1;
+  gpu.dma.h = (((size_word >> 16) - 1) & 0x1ff) + 1;
   gpu.dma.offset = 0;
+  gpu.dma.is_read = is_read;
+  gpu.dma_start = gpu.dma;
 
   renderer_flush_queues();
   if (is_read) {
@@ -278,15 +309,21 @@ static void start_vram_transfer(uint32_t pos_word, uint32_t size_word, int is_re
     memcpy(&gpu.gp0, VRAM_MEM_XY(gpu.dma.x, gpu.dma.y), 4);
     gpu.state.last_vram_read_frame = *gpu.state.frame_count;
   }
-  else {
-    renderer_invalidate_caches(gpu.dma.x, gpu.dma.y, gpu.dma.w, gpu.dma.h);
-  }
 
   log_io("start_vram_transfer %c (%d, %d) %dx%d\n", is_read ? 'r' : 'w',
     gpu.dma.x, gpu.dma.y, gpu.dma.w, gpu.dma.h);
 }
 
-static int check_cmd(uint32_t *data, int count)
+static void finish_vram_transfer(int is_read)
+{
+  if (is_read)
+    gpu.status.img = 0;
+  else
+    renderer_update_caches(gpu.dma_start.x, gpu.dma_start.y,
+                           gpu.dma_start.w, gpu.dma_start.h);
+}
+
+static noinline int do_cmd_buffer(uint32_t *data, int count)
 {
   int len, cmd, start, pos;
   int vram_dirty = 0;
@@ -366,7 +403,7 @@ static int check_cmd(uint32_t *data, int count)
 
 static void flush_cmd_buffer(void)
 {
-  int left = check_cmd(gpu.cmd_buffer, gpu.cmd_len);
+  int left = do_cmd_buffer(gpu.cmd_buffer, gpu.cmd_len);
   if (left > 0)
     memmove(gpu.cmd_buffer, gpu.cmd_buffer + gpu.cmd_len - left, left * 4);
   gpu.cmd_len = left;
@@ -381,7 +418,7 @@ void GPUwriteDataMem(uint32_t *mem, int count)
   if (unlikely(gpu.cmd_len > 0))
     flush_cmd_buffer();
 
-  left = check_cmd(mem, count);
+  left = do_cmd_buffer(mem, count);
   if (left)
     log_anomaly("GPUwriteDataMem: discarded %d/%d words\n", left, count);
 }
@@ -432,7 +469,7 @@ long GPUdmaChain(uint32_t *rambase, uint32_t start_addr)
     list[0] |= 0x800000;
 
     if (len) {
-      left = check_cmd(list + 1, len);
+      left = do_cmd_buffer(list + 1, len);
       if (left)
         log_anomaly("GPUdmaChain: discarded %d/%d words\n", left, len);
     }
@@ -519,7 +556,6 @@ long GPUfreeze(uint32_t type, struct GPUFreeze *freeze)
       freeze->ulStatus = gpu.status.reg;
       break;
     case 0: // load
-      renderer_invalidate_caches(0, 0, 1024, 512);
       memcpy(gpu.vram, freeze->psxVRam, sizeof(gpu.vram));
       memcpy(gpu.regs, freeze->ulControl, sizeof(gpu.regs));
       memcpy(gpu.ex_regs, freeze->ulControl + 0xe0, sizeof(gpu.ex_regs));
@@ -529,6 +565,7 @@ long GPUfreeze(uint32_t type, struct GPUFreeze *freeze)
         GPUwriteStatus((i << 24) | (gpu.regs[i] ^ 1));
       }
       renderer_sync_ecmds(gpu.ex_regs);
+      renderer_update_caches(0, 0, 1024, 512);
       break;
   }