Upsampling = converting sampling rate/size from one to another, for example 16bit/44.1KHz to 24bit/96KHz.
Oversampling = this is a bit more complicated than upsampling. It basically is a way to use DAC that does not have as much bit "width" to function as if it does.
One simple example can be seen with one of the earliest DAC, TDA-1540 (recalling off the top of my head, so pardon for any errors in my memory). The DAC is a 14bit DAC, but the data is 16bit data... How do you make a 14bit DAC to function as a 16bit DAC? Simple, just make it run 4 times as fast and use the digital filter to generate the proper data for the DAC to produce the effect of 16bit DAC.
For example, a DAC is able to output voltage level in increments of 1 volt... So how does it output say 1.25V? By quickly switching the output between 1 and 2 volt in the ratio of a 2 volt output with three 1 volt output. With a smoothing circuit at the end (which is what LPF does), and you get something that's really really looking like 1.25V to you.
That's my understand of the principles, however this is all recalled off the top of my head, so feel free to let me know if I have any mistakes.

Thanks
David