jwagner Posted November 23, 2012 Report Posted November 23, 2012 Hello, I tried using the `uvm_unpack arrayN macro, but I ran into a problem. The DUT uses a SPI for accessing a SRAM. The length of the access is dynamic and ends by disabling the chip select. The access command and start address are held in the first three bytes, so I know exactly the start of the payload. Because of the defined protocol, no metadata can be encoded in the bytestream. So i tried to use `uvm_unpack_arrayN(data,8) to read back my array. The thing is, when i run this macro without having the packer set to use meta data, the array data will always have zero entries, because sz__ will be always zero. The macro expansion is: `uvm_unpack_arrayN(data, 8); begin int sz__; if (packer.use_metadata) begin if (packer.big_endian) begin int cnt__ = packer.count + 32; uvm_bitstream_t tmp__ = sz__; for (int i=0; i<32; i++) tmp__ = packer.m_bits[cnt__ - i - 1]; sz__ = tmp__; end else begin sz__ = packer.m_bits[packer.count +: 32]; end packer.count += 32; end data = new[sz__]; begin foreach (data ) begin ...... It is easy to work around my issue. I just computed the length and did a foreach: data = new[packer.get_packed_size() / 8 - 3]; foreach(data) `uvm_unpack_intN(data, 8); I just wanted to ask if this kind of behaviour is intended by the macro? Maybe it would make sense to init the array only, if metainformation is encoded or if no meta information is encoded and the array is not initialized? Quote
Recommended Posts
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.