ChimeraTK::numeric_address Namespace Reference

Functions
RegisterPath	BAR ()
	The numeric_address::BAR() function can be used to directly access registers by numeric addresses, instead of using register names (e.g.

Function Documentation

BAR()

RegisterPath ChimeraTK::numeric_address::BAR

(

)

The numeric_address::BAR() function can be used to directly access registers by numeric addresses, instead of using register names (e.g.

when no map file exists). The address will form a special RegisterPath which can be used in any place where a register path name is expected, if the backend supports it.

The syntax is as follows

• BAR()/<barNumber>/<addressInBytes> -> access 4-byte register in bar <barNumber> at byte offset <addressInBytes>
• BAR()/<barNumber>/<addressInBytes>*<lengthInBytes> -> access register with arbitrary length of <lengthInBytes>
• BAR()/<barNumber>/<addressInBytes>*<nBytes>u<bitWidth> -> access an unsigned register of <nBytes> bytes whose element width is <bitWidth> (8, 16, 32 or 64)
• BAR()/<barNumber>/<addressInBytes>*<nBytes>u<bitWidth>p<pitchBits> -> access a strided register of <nBytes> bytes: the element pitch is <pitchBits> bits (a multiple of 8) and the number of elements is <nBytes>*8/<pitchBits>

Every part after <addressInBytes> is optional: *<nBytes> is the total span in bytes (default: the element size), u<bitWidth> selects an unsigned and s<bitWidth> a signed element of width 8, 16, 32 or 64 (default: signed 32-bit), and p<pitchBits> sets the element pitch in bits (a multiple of 8). With p<pitchBits> the register is strided: each element has a distance of pitchBits bits in memory, and nElements = nBytes*8/pitchBits.

Definition at line 7 of file NumericAddress.cc.