Closed edersondisouza closed 4 years ago
I was wondering about this myself. This should be worded very carefully. Of course you can build a single listener that uses FIFO levels to control playback speed and you can do a pretty good job with that. However, if you want precise playback times, or you have two or more listeners that require synchronization you will have a bit of a tough time accomplishing that. You could do some things with 1588 that will give you pretty good sync, but this project is about AVB/TSN using AS, at least that's what I thought. Additionally, with the introduction of 802.1CB you could have a particular packet of samples be delivered out of order. If you are using FIFO levels to control playback speed you will need to have clever enough logic that realizes a packet is missing and doesn't wrongly assume it needs to speed up playback because the FIFO level has dropped. Then at a later time the packet will arrive and the FIFO level would be back to where it belongs. On Tuesday, December 17, 2019, 05:56:11 AM MST, andrew-elder notifications@github.com wrote:
@andrew-elder commented on this pull request.
In source/avb.rst:
@@ -37,7 +37,10 @@ System Requirements
This tutorial has been validated on two desktop machines with Intel(R) Ethernet -Controller I210 connected back-to-back and Linux kernel version 4.19. +Controller I210 connected back-to-back and Linux kernel version 4.19. However, +only the machine acting as "Talker" needs a TSN capabable Ethernet controller
The listener needs to support gPTP h/w timestamping, right? I guess I think changing this text has the potential to create confusion. It is fine to state the validation setup is two i210s. Actual requirements for a listener should probably go somewhere else completely. Or provide a reference to external documentation.
— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub, or unsubscribe.
Fair @andrew-elder and @craiggunther. I'll drop this PR and bring something different later. Thanks!
Thank you for working to improve the documentation!
Only the Talker machnine on a TSN system needs capable HW and kernel, so, let's make it clear on documentation.