This struct is used to augment the default full-frame AutoCirculate DMA transfer to accommodate multiple discontiguous "segments" of a video frame. The NTV2FieldBurn Demo demo app shows how this can be used. Other examples: More...

#include <ntv2publicinterface.h>

Public Member Functions
	NTV2SegmentedDMAInfo ()
	Constructs a disabled NTV2SegmentedDMAInfo struct. More...

	NTV2SegmentedDMAInfo (const ULWord inNumSegments, const ULWord inNumActiveBytesPerRow, const ULWord inHostBytesPerRow, const ULWord inDeviceBytesPerRow)
	Constructs an NTV2SegmentedDMAInfo struct from a segment count, a host pitch, and device pitch value. More...

void	Set (const ULWord inNumSegments, const ULWord inNumActiveBytesPerRow, const ULWord inHostBytesPerRow, const ULWord inDeviceBytesPerRow)
	Sets the NTV2SegmentedDMAInfo struct members. More...

void	Reset (void)
	Resets the NTV2SegmentedDMAInfo struct members to their default values (normal, non-segmented AutoCirculate DMA transfers). More...

ULWord	GetSegmentCount (void) const

bool	IsSegmented (void) const

bool	RPCEncode (NTV2_RPC_BLOB_TYPE &outBlob)

bool	RPCDecode (const NTV2_RPC_BLOB_TYPE &inBlob, size_t &inOutIndex)

Public Attributes
ULWord	acNumSegments
	Number of segments of size ::AUTOCIRCULATE_TRANSFER::acVideoBuffer.fByteCount to transfer. Zero or 1 means normal (unsegmented) transfer. More...

ULWord	acNumActiveBytesPerRow
	Number of active bytes in a row of video. More...

ULWord	acSegmentHostPitch
	Offset, in bytes, between the start of one host segment and the start of the next host segment (i.e. host memory bytes-per-row). More...

ULWord	acSegmentDevicePitch
	Offset, in bytes, between the start of one device segment and the start of the next device segment (i.e. device memory bytes-per-row). More...

Detailed Description

This struct is used to augment the default full-frame AutoCirculate DMA transfer to accommodate multiple discontiguous "segments" of a video frame. The NTV2FieldBurn Demo demo app shows how this can be used. Other examples:

An in-host-memory frame has extra "padding" bytes at the end of each row. In this case, set NTV2SegmentedDMAInfo::acNumActiveBytesPerRow to the number of active bytes per row, NTV2SegmentedDMAInfo::acNumSegments to the number of rows, NTV2SegmentedDMAInfo::acSegmentHostPitch to the number of bytes from the beginning of one row to the next. In this example, NTV2SegmentedDMAInfo::acSegmentDevicePitch would equal NTV2SegmentedDMAInfo::acNumActiveBytesPerRow (i.e. the frame is packed in device memory).
To DMA a sub-section of a frame, set NTV2SegmentedDMAInfo::acNumActiveBytesPerRow to the number of bytes that comprise one row of the subsection, then NTV2SegmentedDMAInfo::acNumSegments to the number of rows in the subsection, NTV2SegmentedDMAInfo::acSegmentHostPitch to the bytes-per-row of the entire frame in host memory, and NTV2SegmentedDMAInfo::acSegmentDevicePitch to the bytes-per-row of the entire frame in device memory.
Note
IMPORTANT: For segmented DMAs, the ::AUTOCIRCULATE_TRANSFER::acVideoBuffer.fByteCount field holds the segment byte count (i.e., the number of bytes to transfer per segment.

This struct uses a constructor to properly initialize itself. Do not use memset or bzero to initialize or "clear" it.

Todo:
Create a new field in the NTV2SegmentedDMAInfo structure to eliminate the necessity of setting ::AUTOCIRCULATE_TRANSFER::acVideoBuffer.fByteCount to store the segmented transfer's bytes-per-segment value.

Note: Setting NTV2SegmentedDMAInfo::acNumSegments to 0 or 1 defaults to normal non-segmented DMA behavior (i.e. DMA one complete, packed frame).

Definition at line 7079 of file ntv2publicinterface.h.

Constructor & Destructor Documentation

◆ NTV2SegmentedDMAInfo() [1/2]

NTV2SegmentedDMAInfo::NTV2SegmentedDMAInfo ( )

explicit

Constructs a disabled NTV2SegmentedDMAInfo struct.

Definition at line 2590 of file ntv2publicinterface.cpp.

◆ NTV2SegmentedDMAInfo() [2/2]

NTV2SegmentedDMAInfo::NTV2SegmentedDMAInfo	(	const ULWord	inNumSegments,
		const ULWord	inNumActiveBytesPerRow,
		const ULWord	inHostBytesPerRow,
		const ULWord	inDeviceBytesPerRow
	)

explicit

Constructs an NTV2SegmentedDMAInfo struct from a segment count, a host pitch, and device pitch value.

Parameters

[in]	inNumSegments	Specifies the number of segments (lines) to DMA.
[in]	inNumActiveBytesPerRow	Specifies the number of active bytes in a row of video.
[in]	inHostBytesPerRow	Specifies the offset, in bytes, between two adjacent segments on the host.
[in]	inDeviceBytesPerRow	Specifies the offset, in bytes, between two adjacent segments on the device.

Definition at line 2596 of file ntv2publicinterface.cpp.

Member Function Documentation

◆ GetSegmentCount()

ULWord NTV2SegmentedDMAInfo::GetSegmentCount ( void ) const

inline

Returns: My segment count.

Definition at line 7121 of file ntv2publicinterface.h.

◆ IsSegmented()

bool NTV2SegmentedDMAInfo::IsSegmented ( void ) const

inline

Returns: True if I'm currently active (i.e., I have more than one segment; otherwise false.

Definition at line 7126 of file ntv2publicinterface.h.

◆ Reset()

void NTV2SegmentedDMAInfo::Reset ( void )

Resets the NTV2SegmentedDMAInfo struct members to their default values (normal, non-segmented AutoCirculate DMA transfers).

Definition at line 2616 of file ntv2publicinterface.cpp.

◆ RPCDecode()

bool NTV2SegmentedDMAInfo::RPCDecode	(	const NTV2_RPC_BLOB_TYPE &	inBlob,
		size_t &	inOutIndex
	)

Definition at line 3908 of file ntv2publicinterface.cpp.

◆ RPCEncode()

bool NTV2SegmentedDMAInfo::RPCEncode ( NTV2_RPC_BLOB_TYPE & outBlob )

Definition at line 3899 of file ntv2publicinterface.cpp.

◆ Set()

void NTV2SegmentedDMAInfo::Set	(	const ULWord	inNumSegments,
		const ULWord	inNumActiveBytesPerRow,
		const ULWord	inHostBytesPerRow,
		const ULWord	inDeviceBytesPerRow
	)

Sets the NTV2SegmentedDMAInfo struct members.

Parameters

[in]	inNumSegments	Specifies the number of segments (lines) to DMA. If 1 or zero, performs a Reset, and the remaining parameters are ignored.
[in]	inNumActiveBytesPerRow	Specifies the number of active bytes in a row of video.
[in]	inHostBytesPerRow	Specifies the offset, in bytes, between two adjacent segments on the host.
[in]	inDeviceBytesPerRow	Specifies the offset, in bytes, between two adjacent segments on the device.

Definition at line 2602 of file ntv2publicinterface.cpp.

Member Data Documentation

◆ acNumActiveBytesPerRow

ULWord NTV2SegmentedDMAInfo::acNumActiveBytesPerRow

Number of active bytes in a row of video.

Definition at line 7082 of file ntv2publicinterface.h.

◆ acNumSegments

ULWord NTV2SegmentedDMAInfo::acNumSegments

Number of segments of size ::AUTOCIRCULATE_TRANSFER::acVideoBuffer.fByteCount to transfer. Zero or 1 means normal (unsegmented) transfer.

Definition at line 7080 of file ntv2publicinterface.h.

◆ acSegmentDevicePitch

ULWord NTV2SegmentedDMAInfo::acSegmentDevicePitch

Offset, in bytes, between the start of one device segment and the start of the next device segment (i.e. device memory bytes-per-row).

Definition at line 7085 of file ntv2publicinterface.h.

◆ acSegmentHostPitch

ULWord NTV2SegmentedDMAInfo::acSegmentHostPitch

Offset, in bytes, between the start of one host segment and the start of the next host segment (i.e. host memory bytes-per-row).

Definition at line 7083 of file ntv2publicinterface.h.

The documentation for this class was generated from the following files:

/home/tcbuilder/buildAgent/work/df8e8cce5bebfd58/libajantv2/ajantv2/includes/ntv2publicinterface.h
/home/tcbuilder/buildAgent/work/df8e8cce5bebfd58/libajantv2/ajantv2/src/ntv2publicinterface.cpp

Public Member Functions

Public Attributes