Pipe CPU consumption


I have a question related to pipe transfers :

During a “okUsbFrontPanel_ReadFromPipeOut” call , the caller thread is blocked until completion. OK, clear.

But what about driver code, i mean is it only a DMA transfer programming ?
‘Who’ polls the DMA tx completion flag ?
Is it the driver JOB ?



Yes, on all three operating systems (Mac, Linux, Windows), the underlying call to perform a bulk transfer is a call to the operating system. It is the job of this layer to handle the completion determination.