nngxAddVramDmaCommand Function

NNGX_APICALL void NNGX_APIENTRY nngxAddVramDmaCommand(
     const GLvoid * srcaddr,
     GLvoid * dstaddr,
     GLsizei size
);

	Name	Description
in	srcaddr	Transfer source address (main memory).
in	dstaddr	Transfer destination address (VRAM).
in	size	Transfer data size.

There is no return value.

Issues a DMA transfer command request to the current command list.

DMA transfer command requests accumulate in the current command list. They are used when transferring data from main memory to VRAM without using the GL functions. This operation transfers size bytes of data from the address specified in srcaddr to the address specified in dstaddr.

There are no restrictions on the alignment of values that can be set in srcaddr, dstaddr, or size. However, both srcaddr and dstaddr can be transferred more efficiently by using pointers that have 8-byte alignment.

When using this function to perform a DMA transfer, cache coherecy is handled internally prior to the transfer and does not require support from the calling side. When a cache flush is unnecessary, it can be omitted by using the nngxAddVramDmaCommandNoCacheFlush function.

The following errors occur with this function.

GL_ERROR_8062_DMP	A valid command list is not currently bound.
GL_ERROR_8064_DMP	The size argument is negative.

nngxAddVramDmaCommandNoCacheFlush

2011/08/02

Added information about alignment and transfer efficiency.

2011/03/18

Added an explanation about the nngxAddVramDmaCommandNoCacheFlush function.

2011/02/18

Deleted indications specific to the 8-byte alignment restrictions for srcaddr and dstaddr.

2011/01/17

Explained cache coherency.

2010/06/09

Initial version.