> > Here is my RFC to support DMA with the i2c-sh_mobile core. DMA works nicely > > with my tests so far and we save 1 interrupt per transferred byte, yay! > > Do we have an idea of how much power (or CPU time ?) DMA support could save ?
I have some numbers from the function tracer, but I don't trust them. According to them, DMA is always better (50us per interrupt vs. 23us for DMA setup). This is why I say the threshold is still under test.
signature.asc
Description: Digital signature
