From e502abe001306ad8c95ea5b32a437c53d90eccfa Mon Sep 17 00:00:00 2001
From: Michael Forney <mforney@mforney.org>
Date: Mon, 8 Feb 2021 04:58:41 +0100
Subject: [PATCH] games/gb: various HDMA fixes H-blank DMA should only transfer
 16 bytes per h-blank, rather than waiting for the first h-blank and then
 transferring the whole size.

HDMAC should read 0xff when the transfer is finished, and 0 in the
high bit when the transfer is ongoing. Also, if 0 is written in the
high bit, the current transfer should be aborted.

Introduce two flags, DMAREADY and DMAHBLANK rather than special
constants 1 and -1. If dma is non-zero, there is an ongoing DMA. If
DMAREADY is set, the next chunk is ready to transfer.

Reference: https://gbdev.io/pandocs/#ff55-hdma5-cgb-mode-only-new-dma-length-mode-start

Tested with pokemon crystal.

What was happening is that when the game was loading N background tiles
into vram (each 16 bytes, so one per h-blank), it did something like
this:
- start an hdma transfer for N+1 tiles
- after the Nth tile is transferred, it would read HDMA5, clear the
high bit, then write it back to abort the transfer.

games/gb would instead transfer all N+1 tiles at once, overwriting one
extra tile with whatever was 1 past the end of the source array, and
then would interpret the cancel request as the start of a new transfer
of 16 bytes, which would copy an additional tile past the end. The end
result is that every transfer would end up copying N+2 tiles instead
of just N, overwriting certain tiles with whatever was after the end
of the source data.
---
 sys/src/games/gb/dat.h |  5 ++++-
 sys/src/games/gb/gb.c  |  2 +-
 sys/src/games/gb/mem.c | 23 ++++++++++++++---------
 sys/src/games/gb/ppu.c |  4 ++--
 4 files changed, 21 insertions(+), 13 deletions(-)

diff --git a/sys/src/games/gb/dat.h b/sys/src/games/gb/dat.h
index 68c03372c..890856663 100644
--- a/sys/src/games/gb/dat.h
+++ b/sys/src/games/gb/dat.h
@@ -8,7 +8,7 @@ extern MBC3Timer timer;
 extern uchar vram[16384];
 extern int nrom, nback, nbackbank;
 extern u32int pal[64];
-extern s8int dma;
+extern u8int dma;
 extern u32int divclock;
 
 extern Event *elist;
@@ -113,6 +113,9 @@ enum {
 	FEATRAM = 1,
 	FEATBAT = 2,
 	FEATTIM = 4,
+
+	DMAREADY = 1,
+	DMAHBLANK = 2,
 	
 	INIT = -1,
 	SAVE = -2,
diff --git a/sys/src/games/gb/gb.c b/sys/src/games/gb/gb.c
index 5e038ddf4..0c1119bb1 100644
--- a/sys/src/games/gb/gb.c
+++ b/sys/src/games/gb/gb.c
@@ -308,7 +308,7 @@ threadmain(int argc, char **argv)
 			qlock(&pauselock);
 			qunlock(&pauselock);
 		}
-		if(dma > 0)
+		if(dma & DMAREADY)
 			t = dmastep();
 		else
 			t = step();
diff --git a/sys/src/games/gb/mem.c b/sys/src/games/gb/mem.c
index 9098dbe18..05e4754ea 100644
--- a/sys/src/games/gb/mem.c
+++ b/sys/src/games/gb/mem.c
@@ -15,7 +15,7 @@ int nrom, nback, nbackbank;
 u32int divclock;
 int prish;
 MBC3Timer timer, timerl;
-s8int dma;
+u8int dma;
 u32int white;
 u32int moncols[4];
 
@@ -175,7 +175,14 @@ regwrite(u8int a, u8int v)
 	case HDMAC:
 		if((mode & COL) == 0)
 			goto ff;
-		dma = (v & 0x80) != 0 ? -1 : 1;
+		if(v & 0x80){
+			v &= 0x7f;
+			dma = DMAHBLANK;
+		}else if(dma){
+			v |= 0x80;
+			dma = 0;
+		}else
+			dma = DMAREADY;
 		break;
 	case NR10: v |= 0x80; goto snd;
 	case NR14: case NR24: v |= 0x38; goto snd;
@@ -534,6 +541,7 @@ meminit(void)
 	reg[VBK] = 0xfe;
 	reg[SVBK] = 0xf8;
 	reg[IF] = 0xe0;
+	reg[HDMAC] = 0xff;
 }
 
 void
@@ -572,7 +580,7 @@ dmastep(void)
 {
 	int i;
 	u16int sa, da;
-	
+
 	sa = (reg[HDMASL] | reg[HDMASH] << 8) & 0xfff0;
 	da = (reg[HDMADL] | reg[HDMADH] << 8) & 0x1ff0 | 0x8000;
 	for(i = 0; i < 16; i++)
@@ -583,12 +591,9 @@ dmastep(void)
 	reg[HDMADL] += 16;
 	if((reg[HDMADL] & 0xf0) == 0)
 		reg[HDMADH]++;
-	if((reg[HDMAC] & 0x7f) == 0)
+	if(--reg[HDMAC] == 0xff)
 		dma = 0;
-	else{
-		reg[HDMAC]--;
-		if((reg[HDMAC] & 0x80) != 0)
-			dma = 1;
-	}
+	else if(dma & DMAHBLANK)
+		dma &= ~DMAREADY;
 	return 64;
 }
diff --git a/sys/src/games/gb/ppu.c b/sys/src/games/gb/ppu.c
index f24e9db81..9043420eb 100644
--- a/sys/src/games/gb/ppu.c
+++ b/sys/src/games/gb/ppu.c
@@ -338,8 +338,8 @@ hblanktick(void *)
 			reg[IF] |= IRQLCDS;
 		t = hblclock + 456 * 2 - clock;
 		addevent(&evhblank, t < 0 ? 456 * 2 : t);
-		if(dma < 0)
-			dma = 1;
+		if(dma & DMAHBLANK)
+			dma |= DMAREADY;
 		break;
 	}
 }