I want to know exactly what are the advantages of using the do_pack and do_unpack functions on my uvm_sequence_item, instead of creating custom made pack and unpack functions inside the item class (ignoring the uvm_packer class tool).
The items I’m working with today contains several dynamic fields (like extensible bit vectors and queues for example) and interdependencies between fields.
I have been reading about how to implement those methods correctly, but it seems to me that its more complicated than to simply create new userdefined functions that pack/unpack all the fields to/from a bit string.
Should I make an effort to study and use the predefined methods? What are the advantages over an independent user defined implementation?
Well if you implement it well than you can skip the default do_pack/unpack functions.
As far as I know do_pack/unpack methods provides some policy classes, which adds some checks in pack/unpack.
Also using do_pack/unpack would allow you to use standard way of UVM coding, which is always recommended whereas using custom way.
I do not think it is much complicated to implement do_pack, unpack using predefined methods. May be your could share your questions here.
How you suggest I could implement the unpacking of eight-bit Extensible bit vectors(EBV-8) using the uvm_packer functions?
An extensible bit vector (EBV) is a data structure with an extensible data range.
An EBV is an array of blocks. Each block contains a single extension bit followed by a specific number of data bits. If B represents the total number of bits in one block, then a block contains B – 1 data bits. This protocol I’m working with shall use blocks of length 8 bits (EBV-8). The data value represented by an EBV is simply the bit string formed by the data bits as read from left-to-right, ignoring the extension bits. See image:
The field in the item is an integer, but it has to pack/unpack to/from an array/queue with this specifications. I already have functions to get an EBV-8 array from an integer and vice-versa (get the integer from an array passed by reference, so I can have the resulting array also, after poping the EBV part). Any clues on how to solve this using the do_pack/do_unpack methods?
