Firmware RTL question about ip_send.v

[mr.ri...@gmail.com]

Hi Steve and Group,

I'm trying to address an issue against the OpenHPSDR TAPR protocol 2 repo:

https://github.com/TAPR/OpenHPSDR-Firmware/issues/3

Much of that ethernet code was borrowed from the HL2 project. The person

who posted that issue is suggests I re-introduce ip_id:

"fix the hardcoded IP Identification field (0x0000 on every
packet) regardless of which of the above is chosen — per RFC 791 this should be
unique per packet (or at least incrementing) so that debugging tools and any
future fragmentation/reassembly logic can tell packets apart."

The ip_id was removed in the commit below with the comment:

"Removed ip_id to fix timing"

I'm wondering what timing problems it caused before I re-introduce it in the P2

code. Maybe it was something specific to P1 (protocol 1)?

https://github.com/softerhardware/Hermes-Lite2/commit/9772b1faca8b68ecf06de9ac1fdd831343d6cf00

-Rick / N1GP

[Steve Haynal]

Hi Rick,

I'm sorry but I don't remember the details of that patch from almost 6 years ago. :)

The patch looks fairly small. You can apply it in reverse, build the gateware with Quartus, and see if/what timing violations show up.

73,

Steve

kf7o

[mr.ri...@gmail.com]

Tnx for the reply Steve. I'll try that out. I wasn't sure if it was a quartus build timing issue or

maybe a runtime timing issue. But I'll find out...