More Dumb Uart Questions

I have a design that I have never been able to use the UART on. Well, yesterday I finally figured out the problem. I am using MicroC/OS-II and the JTAG Uart and normal Uart take more OS resources than I originally thought. OK problem solved. It would be nice if this info was available somewhere. By the way, here are the resources used:

1x Semaphore for fd_list_lock

2x Semaphore for each UART (jtag or normal)

1x flag for each UART (jtag or normal)

OK, now this is for the main app that runs on my hardware. I also have a custom bootloader than uses the single-thread sys_lib. When I try to use the normal UART on the bootloader, I get garbage out. I have read all of the posts here about the UART and I can't find any answers. I started with the code copied out of the multi-thread code but that didn't work so I started playing with the library properties. I've set a sys_clock and not set it. I've used reduced device drivers and the full drivers. And every combination of those two. Am I missing something? I'm running the UART at 9600, 8-bit, no parity, 1 stop bit, no flow control. Any help would be greatly appreciated.