On 9/13/2016, Ulf Hansson wrote: > Regarding the ocr_mask; How do these SoCs provides the power to the mmc/sd > card? > > Do note, I am *not* talking about the I/O voltage but the core power to > the card.
The SoC does not *provide* power to the card. It's not that fancy. How the 3.3v is supplied to the card slot is up to the board designer. But, the SoC spec says the I/O should be 3.3v, so we might as well fix the OCR as 3.3v because they don't have a choice. I did see all the DT regulator stuff for the other Renesas SoC, but all that is beyond how this chip is being used. Also note that this SoC (Renesas RZ/A1), unlike the others, is being used with XIP_KERNEL because it's possible to run a full kernel without any external memory (it comes in a 3MB, 5MB, or 10RM internal RAM versions). So, I'm trying to avoid adding extra DT "stuff" that just takes up more RAM resources and has no value for these simple embedded systems. I know in other DDR based system with Megs and Megs of RAM to spare this isn't a problem, but in an XIP_KERNEL environment, it all adds up. One of the reasons why I haven't been trying to push our BSP upstream and move everyone is because I'm afraid of what all the new DT bloat is going to do to the kernel RAM usage. Chris
