Comment 3 for bug 1753314

As I said before:

"I'm not submitting this as a patch to the development list as I'm not fully certain it complies with the hardware spec and doesn't break any other functionality."

What I'm trying to say here is that while I may have been able to cobble together a hack to make the UART nominally compatible with VxWorks, I do not understand the hardware or QEMU well enough to really fix this the right way. Even with my hack, every once in a while when printing lots of data on the console, the output from the UART will stall unless I press a key on the serial port, and I don't know why it does that. I did try to investigate it a little but wasn't able to make much progress. (My suspicion is that it has something to do with the fact that the imx_serial module doesn't implement FIFO support, but even if that's true I don't know how to fix it.)

Even so, I figured it was still worth it to attach my changes to the bug report so that somebody who is better at this than me could use it as a starting point, and so that anybody else who might want to experiment with VxWorks using the QEMU sabrelite machine model wouldn't be totally stuck like I was.

In short, the changes I made are good enough for my own needs (the output stalls don't happen often enough to really bother me), but they're not a fully debugged fix. That's why I filed a bug report instead of just submitting a patch in the first place: I wanted somebody sexier than me to create a fully debugged fix.