bqt%softjar.se@localhost (Johnny Billquist) writes:
This whole thread and implementation is broken, I think.
The driver should keep the DTR low for as long as approriate. But there
is no reason why the close should hang on that.
If you open the port again, while the DTR is low, then the open should
block until it is ready. I honestly do not see the point in why this
should be tied together with the close call.
I fully agree. However, the issue was that rapidly opening and closing
the device is slowed down. And that's true with either algorithm.