gr::digital::header_buffer Class Reference

Helper class for handling payload headers. More...

#include <gnuradio/digital/header_buffer.h>

Public Member Functions
	header_buffer (uint8_t *buffer=NULL)
	~header_buffer ()
void	clear ()
size_t	length () const
const uint8_t *	header () const
void	add_field8 (uint8_t data, int len=8, bool bs=false)
void	add_field16 (uint16_t data, int len=16, bool bs=false)
void	add_field32 (uint32_t data, int len=32, bool bs=false)
void	add_field64 (uint64_t data, int len=64, bool bs=false)
void	insert_bit (int bit)
uint8_t	extract_field8 (int pos, int len=8, bool bs=false)
uint16_t	extract_field16 (int pos, int len=16, bool bs=false)
uint32_t	extract_field32 (int pos, int len=32, bool bs=false)
uint64_t	extract_field64 (int pos, int len=64, bool bs=false)

Detailed Description

Helper class for handling payload headers.

This class is used by the header format blocks (e.g., digital::header_format_default) to make it easier to deal with payload headers. This class functions in two different ways depending on if it is used in a transmitter or receiver. When used in a transmitter, this class helps us build headers out of the fields of the protocol. When used in a receiver, this class helps us parse the received bits into the protocol's fields.

This page describes how to work with the different modes, transmit or receive. The class is instructed as to which mode it is in by how the constructor is called. If the constructor is passed a valid array (non NULL), then it is in transmit mode and will pack this buffer with the header fields. If that buffer is NULL, the object is in receive mode.

Transmit Mode

When passed a valid buffer in the constructor, this object is in transmit mode. We can then use the add_field[N] functions to add new fields to this header. The buffer MUST be large enough to hold the full header. As this class is meant to work mostly with the digital::header_format_default and child classes, the header length can be read from digital::header_format_default::header_nbytes().

Each field is a specific length of 8, 16, 32, or 64 bits that are to be transmitted in network byte order. We can adjust the direction of the bytes by setting the byte-swap flag, bs, to true or false.

The length argument (len) for all add_field[N] calls is the number of bytes actually accounted for in the data structure. Often, we would use the full size of the field, which is sizeof(dtype), and the add_field[N] call defaults to len=N. Occasionally, we may need to use fewer bytes than actually represented by the data type. An example would be the access code used in the header_format_default, which is a uint64_t type but may have fewer bytes used in the actual access code.

The function that calls this class is expected to handle the memory handling of the buffer – both allocating and deallocating.

As simple example of using this class in transmit mode:

uint8_t* buffer = (uint8_t*)volk_malloc(header_nbytes(),
                                        volk_get_alignment());

header_buffer hdr(buffer);
hdr.add_field64(sync_word, sync_word_len);
hdr.add_field16(payload_length);
hdr.add_field8(header_flags);
hdr.add_field8(header_options);

// Do something with the header

volk_free(buffer);

In this example, the header contains four fields:

  |0                           15|16          23|24          31|
  |                          sync word                         |
  |                                                            |
  |           length             |     flags    |   options    |

The sync word can be up to 64-bits, but the add_field64 is also passed the number of actual bytes in the sync word and so could be fewer.

Receive Mode

In receive mode, we build up the header as bits are received by inserting them with insert_bit. We can find out how long the current header is, in bits, using the call to length(). If the header is of the appropriate length, we can then start extracting the fields from it. When we are done with the current header, call clear() to reset the internal buffer to empty, which will mean that length() returns 0.

The header fields are extracted using the extract_field[N] functions. Like the add_field[N] functions, we specify the size (in bits) of the field we are extracting. We pass this function the bit-position of the expected field in the received header buffer. The extract_field[N] assumes that the number of bits for the field is N, but we can tell the function to use fewer bits if we want. Setting the length parameter of these functions greater than N is illegal, and it will throw an error.

For example, given a header of | length | seq. num. | where the length is 16 bits and the sequence number is 32 bits, we would use:

     uint16_t len = d_hdr_reg.extract_field16(0);
     uint32_t seq = d_hdr_reg.extract_field32(16);

The extract_field functions are specific to data types of the field and the number of bits for each field is inferred by the data type. So extract_field16 assumes we want all 16 bits in the field represented.

Some headers have fields that are not standard sizes of integers, like a 1 bit, 4 bit, or even 12 bit fields. We can ask for fewer bits for each field. say:

  |0          15|16     19|20          31|
  |     len     |  flags  |   options    |

We would use the following extraction functions:

     uint16_t len   = d_hdr_reg.extract_field16(0);
     uint8_t  flags = d_hdr_reg.extract_field8(16, 4);
     uint16_t opts  = d_hdr_reg.extract_field16(20, 12);

See also

header_format_default

header_format_counter

header_format_crc

Constructor & Destructor Documentation

header_buffer()

gr::digital::header_buffer::header_buffer

(

uint8_t *

buffer = NULL

)

Create a header buffer object with a pre-allocated buffer, buffer, to hold the formatted header data.

If buffer is set to NULL, then this object is in receive mode meant to receive bits from an incoming data stream and provide the ability to extract fields. In this mode, calls to add_field are invalid and will be nops.

~header_buffer()

gr::digital::header_buffer::~header_buffer

(

)

Class destructor.

Member Function Documentation

add_field16()

void gr::digital::header_buffer::add_field16	(	uint16_t	data,
		int	len = 16,
		bool	bs = false
	)

Add an 16-bit field to the header.

Parameters

data	The 16-bit data item.
len	Length (in bits) of data.
bs	Set to 'true' to byte swap the data.

add_field32()

void gr::digital::header_buffer::add_field32	(	uint32_t	data,
		int	len = 32,
		bool	bs = false
	)

Add an 32-bit field to the header.

Parameters

data	The 32-bit data item.
len	Length (in bits) of data.
bs	Set to 'true' to byte swap the data.

add_field64()

void gr::digital::header_buffer::add_field64	(	uint64_t	data,
		int	len = 64,
		bool	bs = false
	)

Add an 64-bit field to the header.

Parameters

data	The 64-bit data item.
len	Length (in bits) of data.
bs	Set to 'true' to byte swap the data.

add_field8()

void gr::digital::header_buffer::add_field8	(	uint8_t	data,
		int	len = 8,
		bool	bs = false
	)

Add an 8-bit field to the header.

Parameters

data	The 8-bit data item.
len	Length (in bits) of data.
bs	Set to 'true' to byte swap the data.

clear()

void gr::digital::header_buffer::clear

(

)

Clears the header.

In transmit mode, this resets the current offset so new add_field functions start adding data to the start of the buffer.

In receive mode, this clears the buffer that we have inserted bits in to.

extract_field16()

uint16_t gr::digital::header_buffer::extract_field16	(	int	pos,
		int	len = 16,
		bool	bs = false
	)

Returns up to a 16-bit field in the packet header.

Parameters

pos	Bit position of the start of the field.
len	The number of bits in the field.
bs	Set to 'true' to byte swap the data.

extract_field32()

uint32_t gr::digital::header_buffer::extract_field32	(	int	pos,
		int	len = 32,
		bool	bs = false
	)

Returns up to a 32-bit field in the packet header.

Parameters

pos	Bit position of the start of the field.
len	The number of bits in the field.
bs	Set to 'true' to byte swap the data.

extract_field64()

uint64_t gr::digital::header_buffer::extract_field64	(	int	pos,
		int	len = 64,
		bool	bs = false
	)

Returns up to a 64-bit field in the packet header.

Parameters

pos	Bit position of the start of the field.
len	The number of bits in the field.
bs	Set to 'true' to byte swap the data.

extract_field8()

uint8_t gr::digital::header_buffer::extract_field8	(	int	pos,
		int	len = 8,
		bool	bs = false
	)

Returns up to an 8-bit field in the packet header.

Parameters

pos	Bit position of the start of the field.
len	The number of bits in the field.
bs	Set to 'true' to byte swap the data.

header()

const uint8_t* gr::digital::header_buffer::header

(

)

const

Returns a constant pointer to the buffer.

insert_bit()

void gr::digital::header_buffer::insert_bit

(

int

bit

)

Insert a new bit on the back of the input buffer. This function is used in receive mode to add new bits as they are received for later use of the extract_field functions.

Parameters

bit	New bit to add.

length()

size_t gr::digital::header_buffer::length

(

)

const

In transmit mode, this returns the length of the data in the buffer (not the allocated buffer length).

In receiving mode, this returns the current length in bits of the received header.

---

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

• header_buffer.h