From c60251de732a63ce3fa4c8723fbb429fb4a35fac Mon Sep 17 00:00:00 2001
From: Jeffrey Pfau <jeffrey@endrift.com>
Date: Fri, 5 Feb 2016 20:44:49 -0800
Subject: [PATCH] GB Video: Fix oam-bug 1
---
src/gb/video.c | 12 +++++++-----
1 file changed, 7 insertions(+), 5 deletions(-)
diff --git a/src/gb/video.c b/src/gb/video.c
index 7d3413750..28a50bcff 100644
--- a/src/gb/video.c
+++ b/src/gb/video.c
@@ -201,14 +201,16 @@ static void _cleanOAM(struct GBVideo* video, int y) {
void GBVideoWriteLCDC(struct GBVideo* video, GBRegisterLCDC value) {
if (!GBRegisterLCDCIsEnable(video->p->memory.io[REG_LCDC]) && GBRegisterLCDCIsEnable(value)) {
- // TODO: Does enabling the LCD start in vblank?
video->mode = 2;
- video->nextMode = GB_VIDEO_MODE_2_LENGTH;
+ video->nextMode = GB_VIDEO_MODE_2_LENGTH - 5; // TODO: Why is this fudge factor needed? Might be related to T-cycles for load/store differing
video->nextEvent = video->nextMode;
- video->eventDiff = 0;
- video->stat = GBRegisterSTATSetMode(video->stat, video->mode);
+ video->eventDiff = -video->p->cpu->cycles;
+ // TODO: Does this read as 0 for 4 T-cycles?
+ video->stat = GBRegisterSTATSetMode(video->stat, 2);
video->p->memory.io[REG_STAT] = video->stat;
- video->eventDiff = 0;
+ video->ly = 0;
+ video->p->memory.io[REG_LY] = 0;
+
if (video->p->cpu->cycles + video->nextEvent < video->p->cpu->nextEvent) {
video->p->cpu->nextEvent = video->p->cpu->cycles + video->nextEvent;
}