Kai Krakow kai@kaishome.de writes:
Device reports may come in faster than our consumers could possibly read them, this is especially true for multi-axis events: When you move a stick across its range, it will always generate at least two events, one for the x axis, and one for the y axis. This is not really an error situation, so let's report this situation only once at most. If we already know the multi-event situation, let's skip logging this completely: We won't loose any information anyway because the report contains a complete device state and only axis positions were updated.
Thus, this commit adds a parameter to process_hid_report() to let it know if we are currently processing reports that are known to be sent multiple times in sequence (with updated reports).
Also, if our consumers are slow to respond, then report the issue only once and not per each occurrence during the duration of one consumer read cycle.
Finally, this is not really an error, it's a warning at most, thus degrade the error message to a warning to not pollute peoples consoles and logs with unimportant stuff.
Is it really necessary to add all that complexity then? Why not simply remove the message?