You next step should be to start a simple serial IO project - just serial IO, no PWM, nothing else.
(probably also start a new thread - with a reference back to this one).

Start with something really simple that just configures the UART (baud rate, etc) and sends a simple message such as "Hello, World!" - don't try interrupts yet, just send 1 character, wait for TI, send the next character, etc...
I wouldn't even try to do it in a loop at this stage.

Run this on a simulator, and you should see the message come up in the simulated serial output.

Connect this to a PC and hypoterminal (or whatever) - if you see the message correctly, you know that you have the UART configured correctly, and all your hardware & cables are working.

Once this is working, be sure to save this project so that you can go back to it if everything goes terribly wrong later...

Now, take a look at that interrupt-driven serial code I found earlier (or some other example, if you prefer):
http://www.8052.com/forum/read.phtml?id=100452

Adapt your initial project to use this interrupt-driven interface to send the message; Simulate; Test in-target.

Again, save this project once it's working.

Now, try making it wait for a receied character (via the interrupt stuff) before sending the message.
Simulate; Test in-target; Save the project.

Now you could try a crude performance test:
Have your main loop wait for received characters(via the interrupt stuff); each time it receives a character, have it transmit that character straight back again (also via the interrupt stuff).
Now, anything you type at hypoterminal you should see echoed straight back.
Take you receive & transmit wires, and connect them to separate COM: ports on your PC (or separate PCs, if you don't have multiple COM: ports):

COM1:  Tx ---------------> 8051 Rx
                              |
                              |
                              V
                             your
                             code
                              |
                              |
                              V
COM2: Rx <---------------- 8051 Tx

Now, anything you type at the hypoterminal on COM1: (say) should appear on the hypoterminal on COM2:.

Type suffifient characters to check that your buffer wrap-around is working.

Now you can test the performance by sending a file from COM1: to COM2:, then use the PC to compare the two files - you don't want to see any missing or corrupted bytes!