| *** Skip has joined #litex | 00:00 | |
| *** tpb has joined #litex | 00:00 | |
| *** captain_morgan has quit IRC | 00:43 | |
| *** captain_morgan has joined #litex | 00:45 | |
| *** scanakci has quit IRC | 01:18 | |
| *** Skip has quit IRC | 02:10 | |
| *** Degi has quit IRC | 03:33 | |
| *** Degi has joined #litex | 03:34 | |
| *** HoloIRCUser2 has joined #litex | 03:52 | |
| *** HoloIRCUser has quit IRC | 03:52 | |
| *** scanakci has joined #litex | 04:20 | |
| *** benh_ has joined #litex | 04:25 | |
| *** benh has quit IRC | 04:26 | |
| *** CarlFK has quit IRC | 06:25 | |
| *** CarlFK has joined #litex | 07:32 | |
| *** CarlFK1 has joined #litex | 08:37 | |
| *** CarlFK has quit IRC | 08:37 | |
| *** benh_ is now known as benh | 09:43 | |
| *** FFY00 has quit IRC | 10:43 | |
| *** FFY00 has joined #litex | 10:44 | |
| *** scanakci has quit IRC | 11:24 | |
| zyp | hmm, I just realized that the first/last mechanism of the streams doesn't support signalling a zero length packet | 12:07 |
|---|---|---|
| _florent_ | zyp: can you describe what you want to achieve? | 12:10 |
| zyp | I'm currently trying to reuse LiteEthMACCRCInserter for usb (with a suitable CRC16 class), and in the case of a zero length packet, the output should consist of only the two crc bytes | 12:12 |
| _florent_ | ok, LiteEthMACCRCInserter is indeed probably expecting at least a data, you could use a modified version that would only insert the CRC is there is no data (you could add a be or length field to the payload to indicate data validity). | 12:20 |
| zyp | yeah, I just realized that this is gonna be a problem all over the place, not just for CRC | 12:21 |
| _florent_ | this signaling indeed needs to be added by the user as a payload parameter | 12:23 |
| zyp | length field sounds like a good idea, that'll also be useful later to increase the datapath width | 12:25 |
| zyp | so a zlp will be signalled by first=1, last=1, length=0 | 12:25 |
| zyp | by the way, is there any documentation on the stream api anywhere? I'm still a bit uncertain how the timings are supposed to work | 12:37 |
| *** Skip has joined #litex | 13:05 | |
| shuffle2 | agree | 13:39 |
| shuffle2 | i suspect everything is just yolo'd, so you have to read how any pre-existing code expects to use the stream | 13:44 |
| zyp | I figured it out, I just misread some code that led me to doubt that my understanding was correct | 13:45 |
| *** CarlFK1 is now known as CarlK | 14:01 | |
| *** Skip has quit IRC | 14:46 | |
| _florent_ | it could be better, but there is this: https://github.com/enjoy-digital/litex/wiki/Streams | 15:55 |
| tpb | Title: Streams · enjoy-digital/litex Wiki · GitHub (at github.com) | 15:55 |
| _florent_ | the streams are basically a simplified AXI (valid/ready/first/last/payload/param) that you can customize depending the need/project | 15:57 |
| *** Skip has joined #litex | 18:00 | |
| *** Skip has quit IRC | 19:21 | |
| *** rohitksingh has joined #litex | 19:48 | |
| *** HoloIRCUser has joined #litex | 20:37 | |
| *** HoloIRCUser2 has quit IRC | 20:40 | |
| *** HoloIRCUser1 has joined #litex | 22:45 | |
| *** HoloIRCUser2 has joined #litex | 22:46 | |
| *** HoloIRCUser has quit IRC | 22:48 | |
| *** HoloIRCUser1 has quit IRC | 22:49 | |
| *** CarlK has quit IRC | 23:06 | |
| *** FFY00 has quit IRC | 23:12 | |
| *** FFY00 has joined #litex | 23:12 | |
| *** CarlFK has joined #litex | 23:25 | |
| *** HoloIRCUser has joined #litex | 23:56 | |
| *** HoloIRCUser2 has quit IRC | 23:56 | |
Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!