On Fri, Jan 6, 2023 at 1:11 PM Rob Kossler <rkoss...@nd.edu> wrote: > On Thu, Jan 5, 2023 at 7:46 PM Brian Padalino <bpadal...@gmail.com> wrote: > > Before RFNoC 4.0, there was a generic NoC shell that was used instead of > one being generated for each of the blocks. > > > > I see there is a noc_shell_generic_ctrlport_pyld_chdr, but it isn't used > anywhere. > > > > Is there a reason that a generic one isn't utilized and that ones are > being generated individually for each block? > > With RFNoC 4.0, you have various options for which interface you would > like for payload data and packet context data. These options, which > are in the block XML file, determine which signals get passed in and > out of your block. So, the reason it is custom is to handle the > variety of interface options. >
There are no more XML files, right? It's all YAML? Right? In the specification: https://files.ettus.com/app_notes/RFNoC_Specification.pdf ... section 2.5.5.3 shows axis_chdr, axis_pyld_ctxt, and axis_data. But beyond that, I don't see any descriptions of the differences. Moreover, I am unsure if I change the YAML and use rfnoc_image_builder to build things, does that regenerate the underlying noc_shell_<block>.v file in the OOT description? Or is it just an input to some other tool? I am also confused by Figure 18 at the end of the document. I'm not sure I see any XML files or the rfnocmodtool GUI? Lastly, I see rfnoc_blocktool in uhd under host/utils/rfnoc_blocktool which has rfnoc_create_verilog.py which looks are fpga_iface but I didn't see anything get installed from UHD. Is rfrnocmodtool from gr-ettus not the right thing to use anymore? I might just be getting confused by out of date tutorials and specifications, but figured I'd try to get my bearings straight. Thanks, Brian
_______________________________________________ USRP-users mailing list -- usrp-users@lists.ettus.com To unsubscribe send an email to usrp-users-le...@lists.ettus.com
