Merge branch 'ikjordan:audio' into audio

software/apps/moon/CMakeLists.txt

@@ -11,6 +11,7 @@ target_compile_definitions(moon PRIVATE
 	DVI_VERTICAL_REPEAT=1
 	DVI_N_TMDS_BUFFERS=3
 	DVI_MONOCHROME_TMDS
+	DVI_1BPP_BUFFER
 )
 
 target_link_libraries(moon
@@ -34,6 +35,7 @@ target_compile_definitions(moon_pio_encode PRIVATE
 	DVI_VERTICAL_REPEAT=1
 	DVI_N_TMDS_BUFFERS=3
 	DVI_MONOCHROME_TMDS
+	DVI_1BPP_BUFFER
 	USE_PIO_TMDS_ENCODE
 )
 

software/apps/moon_double_audio/CMakeLists.txt

@@ -3,7 +3,6 @@
 add_executable(moon_double_audio
 	main.c
 	tmds_double.S
-	tmds_double.h
 )
 
 target_compile_options(moon_double_audio PRIVATE -Wall)

software/apps/moon_double_audio/main.c

@@ -41,9 +41,9 @@ static void core1_main();
 // but playback set to 32kHz
 
 // Pick one:
-#define MODE_640x480_60Hz
+//#define MODE_640x480_60Hz
 //#define MODE_720x540_50Hz
-//#define MODE_720x576_50Hz
+#define MODE_720x576_50Hz
 
 #if defined(MODE_640x480_60Hz)
 #define FRAME_WIDTH 640
@@ -129,7 +129,7 @@ int main()
 
     dvi_set_audio_freq(&dvi0, AUDIO_RATE, dvi0.timing->bit_clk_khz*HDMI_N/(AUDIO_RATE/100)/128, HDMI_N);
 
-	printf("FReq %i Set CTS %i\n", dvi0.timing->bit_clk_khz, dvi0.timing->bit_clk_khz*HDMI_N/(AUDIO_RATE/100)/128);
+	printf("Freq %i Set CTS %i\n", dvi0.timing->bit_clk_khz, dvi0.timing->bit_clk_khz*HDMI_N/(AUDIO_RATE/100)/128);
 
 
     sem_init(&fifty_hz, 0, 1);
@@ -154,16 +154,9 @@ static void __not_in_flash_func(render_loop)()
     {
 		for (uint y = 0; y < FRAME_HEIGHT/2; ++y)
         {
-			const uint32_t *colourbuf = &((const uint32_t*)moon_img)[(y + (FRAME_HEIGHT >> 3))* IMAGE_WIDTH / 32];
+			const uint8_t *colourbuf = (const uint8_t*)&moon_img[(y + (FRAME_HEIGHT >> 3))* IMAGE_WIDTH / 8];
 			queue_remove_blocking_u32(&dvi0.q_tmds_free, &tmdsbuf);
 
-            // 32 pixels create 8 * 2 32 bit words
-            // tmds_double_1bpp requires multiple of 32 pixels
-            // For input of 360 to generate 720 pixels:
-            // 360 pixels -> 11.25 32 bit words
-            // Will therefore generate 12 * 32 * 2 pixels = 768
-            // tmdsbuf size must be rounded up to multiple of 64
-            // in dvi.c
 			tmds_double_1bpp(colourbuf, tmdsbuf, FRAME_WIDTH);
 			queue_add_blocking_u32(&dvi0.q_tmds_valid, &tmdsbuf);
 		}

software/apps/moon_double_audio/tmds_double.S

@@ -28,13 +28,13 @@
 // Fast Half res 1bpp black/white encoder
 
 // Taking the encoder from DVI spec, with initial balance 0:
-// 
+//
 // - Encoding either 0x00 or 0xff will produce a running balance of -8, with
 //   output symbol of 0x100 or 0x200
-// 
+//
 // - Subsequently encoding either 0x01 or 0xfe will return the balance to 0, with
 //  output symbol of 0x1ff or 0x2ff
-// 
+//
 // So we can do 1bpp encode with a lookup of x coordinate LSB, and input
 // colour bit. If we process pixels in even-sized blocks, only the colour
 // lookup is needed.
@@ -56,14 +56,14 @@
 
 .endm
 
-// r0: input buffer (word-aligned)
+// r0: input buffer (byte-aligned)
 // r1: output buffer (word-aligned)
 // r2: output pixel count
 decl_func tmds_double_1bpp
 	push {r4-r7, lr}        // Store scratch registers
 	mov r7, r8
 	push {r7}
-	lsls r2, #1             // Double the output pixel count - initially set to 640
+	lsls r2, #1             // Double the output pixel count
 	add r2, r1              // Now points to end of output buffer
 	mov ip, r2              // ip is r12 - store end of output buffer
 	adr r4, tmds_d1bpp_table // r4 contains pointer to look up table
@@ -71,30 +71,21 @@ decl_func tmds_double_1bpp
 	// Mask: 2 bit index, 8 bytes per entry
 	movs r3, #0x18          // bits 4 to 5 checked
 	b 2f                    // Jump over for the first pass
+
 1:
-	ldmia r0!, {r2}         // Loads r2 with contents pointed by r0 then inc r0
+	ldrb r2, [r0]
+	adds r0, #1
 #if !DVI_1BPP_BIT_REVERSE
 	tmds_double_1bpp_body lsls 3  lsls 1
 	tmds_double_1bpp_body lsrs 1  lsrs 3
-	tmds_double_1bpp_body lsrs 5  lsrs 7
-	tmds_double_1bpp_body lsrs 9  lsrs 11
-	tmds_double_1bpp_body lsrs 13 lsrs 15
-	tmds_double_1bpp_body lsrs 17 lsrs 19
-	tmds_double_1bpp_body lsrs 21 lsrs 23
-	tmds_double_1bpp_body lsrs 25 lsrs 27
 #else
 	tmds_double_1bpp_body lsrs 3   lsrs 1
 	tmds_double_1bpp_body lsls 1   lsls 3
-	tmds_double_1bpp_body lsrs 11  lsrs 9
-	tmds_double_1bpp_body lsrs 7   lsrs 5
-	tmds_double_1bpp_body lsrs 19  lsrs 17
-	tmds_double_1bpp_body lsrs 15  lsrs 13
-	tmds_double_1bpp_body lsrs 27  lsrs 25
-	tmds_double_1bpp_body lsrs 23  lsrs 21
 #endif
+
 2:
 	cmp r1, ip              // Has all of the output data been generated
-	blo 1b                  // If not, back to generate more 
+	blo 1b                  // If not, back to generate more
 
 	pop {r7}                // Restore registers and return
 	mov r8, r7
@@ -143,7 +134,7 @@ tmds_d1bpp_table:
 	ldr \rd, [r0, \rd]
 .endm
 
-// r0: input buffer (word-aligned)
+// r0: input buffer (byte-aligned)
 // r1: output buffer (word-aligned)
 // r2: output pixel count
 decl_func tmds_double_2bpp
@@ -161,28 +152,14 @@ decl_func tmds_double_2bpp
 	b 2f
 1:
 	mov r4, r8
-	ldmia r4!, {r2}
+	ldrb r2, [r4]
+	adds r4, #1
 	mov r8, r4
 	tmds_double_2bpp_body lsls 2  r7
 	tmds_double_2bpp_body lsrs 0  r6
 	tmds_double_2bpp_body lsrs 2  r5
 	tmds_double_2bpp_body lsrs 4  r4
 	stmia r1!, {r4-r7}
-	tmds_double_2bpp_body lsrs 6  r7
-	tmds_double_2bpp_body lsrs 8  r6
-	tmds_double_2bpp_body lsrs 10 r5
-	tmds_double_2bpp_body lsrs 12 r4
-	stmia r1!, {r4-r7}
-	tmds_double_2bpp_body lsrs 14 r7
-	tmds_double_2bpp_body lsrs 16 r6
-	tmds_double_2bpp_body lsrs 18 r5
-	tmds_double_2bpp_body lsrs 20 r4
-	stmia r1!, {r4-r7}
-	tmds_double_2bpp_body lsrs 22 r7
-	tmds_double_2bpp_body lsrs 24 r6
-	tmds_double_2bpp_body lsrs 26 r5
-	tmds_double_2bpp_body lsrs 28 r4
-	stmia r1!, {r4-r7}
 2:
 	cmp r1, ip
 	blo 1b
@@ -196,3 +173,23 @@ tmds_d2bpp_table:
 	.word 0xb3e30 // 10, 10
 	.word 0x73d30 // 01, 01
 	.word 0x7f103 // 00, 00
+
+// r0: start of tdms buffer for plane 1
+// r1: pixel count to copy
+decl_func tmds_clone
+    push {lr}
+    lsls r1, #2     // convert to 32 bit words
+    adds r2, r0, r1 // Start of plane 2
+    mov ip, r2      // End when get to plane 2
+    adds r3, r2, r1 // Start of plane 3
+    b 2f
+
+1:
+    ldm r0!, {r1}
+    stm r2!, {r1}
+    stm r3!, {r1}
+
+2:
+    cmp r0, ip
+    blo 1b
+    pop {pc}

software/apps/moon_double_audio/tmds_double.h

@@ -4,7 +4,9 @@
 #include "dvi_config_defs.h"
 
 // Functions from tmds_double.S
-void tmds_double_1bpp(const uint32_t *pixbuf, uint32_t *symbuf, size_t n_pix);
-void tmds_double_2bpp(const uint32_t *pixbuf, uint32_t *symbuf, size_t n_pix);
+void tmds_double_1bpp(const uint8_t *pixbuf, uint32_t *symbuf, size_t n_pix);
+void tmds_double_2bpp(const uint8_t *pixbuf, uint32_t *symbuf, size_t n_pix);
+
+void tmds_clone(const uint32_t *symbuf, size_t n_pix);
 
 #endif

software/libdvi/dvi.c

@@ -55,7 +55,11 @@ void dvi_init(struct dvi_inst *inst, uint spinlock_tmds_queue, uint spinlock_col
     dvi_setup_scanline_for_active(inst->timing, inst->dma_cfg, NULL, &inst->dma_list_error, false);
     dvi_setup_scanline_for_active(inst->timing, inst->dma_cfg, NULL, &inst->dma_list_active_blank, true);
 
-    uint16_t mask = 0x3f;  // To account for worst case of monochrome horizontal pixel doubling
+    uint16_t mask = 0;
+
+#ifdef DVI_1BPP_BUFFER
+    mask = 0x1f;  // To account for worst case of 1bpp horizontal pixels generated 32 bits at a time (e.g. 720x568)
+#endif
 
     for (int i = 0; i < DVI_N_TMDS_BUFFERS; ++i) {
 #if DVI_MONOCHROME_TMDS