MaxPacketSize from PR #417 might impact performance

[sessl3r]

PR #417 has fixed a issue with Rigol scopes and likely others by using the reported MaxPacketSize of the USB endpoints.
A similar PR was merged into usbtmc linux driver.

In the discussion over there it came up, that some Agilent scopes report a smaller MaxPacketSize then they can handle in reality and therefore this change decreased performance. I can not validate this and think it is bad practice to use those bigger packet sizes.

I completely dislike the way this was now solved in the usbtmc driver (default is 4k, user can overwrite to use the reported value). However I agree that if users have performance issues a overwrite option could be introduced for such quirks.

This is not a current issue - just a reminder to not loose this information.