X-Git-Url: https://notaz.gp2x.de/cgi-bin/gitweb.cgi?p=pcsx_rearmed.git;a=blobdiff_plain;f=plugins%2Fgpu_neon%2Fpsx_gpu%2Fpsx_gpu_arm_neon.S;h=ffbea043c2cb64ddde846414f0ea62390ebf7d19;hp=da47756efcaea45cd3704fd5338e88c3895daa75;hb=refs%2Fheads%2Fmaster;hpb=26e3e2aa7525fd4e63e64192dfbb68950e0e4c5a

diff --git a/plugins/gpu_neon/psx_gpu/psx_gpu_arm_neon.S b/plugins/gpu_neon/psx_gpu/psx_gpu_arm_neon.S
index da47756e..ffbea043 100644
--- a/plugins/gpu_neon/psx_gpu/psx_gpu_arm_neon.S
+++ b/plugins/gpu_neon/psx_gpu/psx_gpu_arm_neon.S
@@ -13,15 +13,9 @@
  * General Public License for more details.
  */
 
-#define MAX_SPANS                                         512
-#define MAX_BLOCKS                                        64
-#define MAX_BLOCKS_PER_ROW                                128
-
-#define RENDER_STATE_MASK_EVALUATE                        0x20
-#define RENDER_FLAGS_MODULATE_TEXELS                      0x1
-#define RENDER_FLAGS_BLEND                                0x2
 #define RENDER_INTERLACE_ENABLED                          0x1
 
+#include "psx_gpu.h"
 #include "psx_gpu_offsets.h"
 
 #define psx_gpu_b_dx_offset (psx_gpu_b_block_span_offset + 4)
@@ -228,7 +222,6 @@
 
 #ifdef __MACH__
 #define flush_render_block_buffer _flush_render_block_buffer
-#define setup_sprite_untextured_simple _setup_sprite_untextured_simple
 #define update_texture_8bpp_cache _update_texture_8bpp_cache
 #endif
 
@@ -369,8 +362,8 @@ function(compute_all_gradients)
   sub r14, r14, #(62 - 12)           @ r14 = shift - (62 - FIXED_BITS)
 
   vshll.u16 uvrg_base, uvrg0, #16    @ uvrg_base = uvrg0 << 16
-  vdup.u32 r_shift, r14              @ r_shift = { shift, shift, shift, shift }
-
+  vdup.u32 r_shift, r14              @ r_shift = { shift, shift*, shift, shift* }
+                                     @ * - vshl.u64: ignored by hw
   vadd.u32 uvrg_base, uvrgb_phase
   vabs.s32 ga_uvrg_x, ga_uvrg_x      @ ga_uvrg_x = abs(ga_uvrg_x)
 
@@ -565,6 +558,8 @@ function(compute_all_gradients)
 #define left_x_32_low                            d22
 #define left_x_32_high                           d23
 
+#define tmp_max_blocks                           d20
+
 #define edges_xy                                 q0
 #define edges_dx_dy                              d2
 #define edge_shifts                              d3
@@ -819,8 +814,10 @@ function(compute_all_gradients)
   str b, [span_b_offset], #4;                                                  \
   setup_spans_adjust_interpolants_##direction();                               \
                                                                                \
+  vmov.u16 tmp_max_blocks, #MAX_BLOCKS_PER_ROW;                                \
   vshr.u16 left_right_x_16_high, left_right_x_16_high, #3;                     \
   vshl.u16 span_shifts, c_0xFFFE, span_shifts;                                 \
+  vmin.u16 left_right_x_16_high, left_right_x_16_high, tmp_max_blocks;         \
                                                                                \
   vst4.u16 { left_right_x_16, span_shifts_y }, [span_edge_data]!;              \
                                                                                \
@@ -867,8 +864,10 @@ function(compute_all_gradients)
   str b, [span_b_offset], #4;                                                  \
   setup_spans_adjust_interpolants_##direction();                               \
                                                                                \
-  vshl.u16 span_shifts, c_0xFFFE, span_shifts;                                 \
+  vmov.u16 tmp_max_blocks, #MAX_BLOCKS_PER_ROW;                                \
   vshr.u16 left_right_x_16_high, left_right_x_16_high, #3;                     \
+  vshl.u16 span_shifts, c_0xFFFE, span_shifts;                                 \
+  vmin.u16 left_right_x_16_high, left_right_x_16_high, tmp_max_blocks;         \
                                                                                \
   vst4.u16 { left_right_x_16, span_shifts_y }, [span_edge_data]!;              \
                                                                                \
@@ -908,7 +907,9 @@ function(compute_all_gradients)
   ble 1f;                                                                      \
                                                                                \
   orr temp, y_a, y_a, lsl #16;                                                 \
+  cmp height, #512;                                                            \
   add temp, temp, #(1 << 16);                                                  \
+  movgt height, #512;                                                          \
   add y_a, temp, #2;                                                           \
   add y_a, y_a, #(2 << 16);                                                    \
   vmov y_x4, temp, y_a;                                                        \
@@ -963,7 +964,9 @@ function(compute_all_gradients)
   ble 1f;                                                                      \
                                                                                \
   orr temp, y_a, y_a, lsl #16;                                                 \
+  cmp height, #512;                                                            \
   sub temp, temp, #(1 << 16);                                                  \
+  movgt height, #512;                                                          \
   sub y_a, temp, #2;                                                           \
   sub y_a, y_a, #(2 << 16);                                                    \
   vmov y_x4, temp, y_a;                                                        \
@@ -5826,14 +5829,7 @@ function(setup_sprite_16bpp_4x)
 
 .align 3
 
-function(setup_sprite_untextured)
-  ldrh r12, [psx_gpu, #psx_gpu_render_state_offset]
-  tst r12, #(RENDER_STATE_MASK_EVALUATE | RENDER_FLAGS_MODULATE_TEXELS         \
-    | RENDER_FLAGS_BLEND)
-  ldrbeq r12, [psx_gpu, #psx_gpu_render_mode_offset]
-  tsteq r12, #RENDER_INTERLACE_ENABLED
-  beq setup_sprite_untextured_simple
-
+function(setup_sprite_untextured_512)
   stmdb sp!, { r4 - r11, r14 }
 
   ldr width, [sp, #40]
@@ -6133,6 +6129,7 @@ function(scale2x_tiles8)
   mov r14, r2
 
 0:
+  pld [r1, #1024*2]
   vld1.u16 { q0 }, [r1, :128]!
   vld1.u16 { q2 }, [r1, :128]!
   vmov q1, q0