These registers are responsible for controlling how framebuffer data can be DMA'd from the DS GPU, and also for configuring the upscaling matrix.

Registers

The physical address can be calculated by subtracting 0xEB00000 from the virtual address.

Control

There are two matrix units, one at +0x200 for vertical (Y) scaling, and the other one at +0x300 for horizontal (X) scaling.

VAddress	Name	Width	Description
0x1EC1xn00	KRN_WIDTH	4	Kernel width - 1 is written here, 1 <= width <= 8 This decides how many pixels are written each batch.
0x1EC1xn04	KRN_PATTERN_BITS	4	If the corresponding bit for the current batch iteration index is set then a new pixel is read. The amount of set bits determine how many pixels are read each batch. Any bit indexes past KRN_WIDTH are ignored. This value is 8 bits, but it has to be written with a 32bit write.
0x1EC1xn40	KRN_MTX	0xC0	int kerneldata[6][8]; - matrix data is written here, height is always 6

Bit(s)	Description
0	Enable bit
1	Enable vertical matrix
2	Enable horizontal matrix
4	???
5	???
8-9	Input pixel mode? 0 = 4byte color, 1 = 3byte color, 2 = 2byte color, 3 = 2byte color
10-11	Output framebuffer rotation: 0 = normal, 1 = 90° CW (right), 2 = 180° CW (upside down, not mirrored), 3 = 270° CW (left)
12	Output tiling for use with the GPU. When set, the output width and height must be a multiple of 8.
15	Start bit (setting this will eventually raise MTX interrupt 0)
16	Data still available flag (?)

Bit(s)	Description
0-8	Output framebuffer width - 1 is written here, 1 <= width <= 512
16-25	Output framebuffer height - 1 is written here, 1 <= height <= 512

Reading this register will return pending interrupts. Writing this register will acknowledge pending interrupts where the bits are set.

Bit(s)	Description
0	FIFO ready (signal to start DMA)
1	FIFO overrun(?) (occurs if DMA is too slow)
2	FIFO underrun(?) (occurs on VBlank)

Interrupt Enable for the above interrupts.