Bob Copeland <me@...> writes:
Hi Bob, Thank you for contributing and supporting so actively the mesh development. I understood the issues with TSF, so I stopped using it. My aim is to approximate how much time the wireless channel to which the card is tuned in, is busy. I use ieee80211_frame_duration for each in/outcoming packet. I use the beacon triggers for periodically averaging and storing samples. I worked around that error assuming a fixed beacon time interval instead of using TSF, which is pretty solid in a controlled environment. However, the approximation would improve a lot counting the used time instead of computing it afterwards. Pursuing a similar purpose, somebody has already implemented that for Atheros, using kernel time and reading registers. Here is the presentation mail https://lists.ath9k.org/pipermail/ath9k-devel/2015-August/014141.html and the project link https://github.com/thuehn/RegMon . The author is Dr.-Ing. Thomas Hühn. I plan to use it in the mid future. Thank you, Ferran
