[U-Boot] [PATCH v2 1/2] arm: socfpga: mmc: Enable calibration for drvsel and smpsel

Marek Vasut marex at denx.de
Fri Aug 21 07:21:26 CEST 2015


On Friday, August 21, 2015 at 07:00:46 AM, Chin Liang See wrote:
> Hi,

Hi,

> > > > > Any comment or ack for this patch?
> > > > 
> > > > Please don't expect that the reviewers/maintainers have nothing else
> > > > on their plate but to review your patch. Besides, this change is
> > > > really low priority one, since thus far the SD/MMC works fine on all
> > > > of the SoCFPGA boards to my knowledge.
> > > > 
> > > > Also, I thought we agreed to wait for Simon to comment on this patch.
> > > > You don't have to repost a patch before the discussion settles a bit.
> > > > This doesn't help, it only puts burden on the receiving side, chill
> > > > :) Let alone the fact that this code should be part of the DWMMC
> > > > core, as it is not socfpga specific.
> > > > 
> > > > What I would be really interested in is a proper description of the
> > > > calibration algorithm. What is the goal and how do you achieve it ?
> > > 
> > > Actually I am probing the efforts needed for this patch before moving
> > > to next one. While for the SDMMC calibration, it is not needed if the
> > > SD card is routed near to the chip. We have incident where SD access
> > > failed as they routed few inches away.
> > 
> > Does the calibration really help or is that a problem with the board ?
> > It'd be nice if you put effort into this patch to get it done correctly
> > :)
> 
> It's related to the board. When the trace route is not properly design,
> it might failed when communication between the controller and card at
> high speed. The calibration will ensure the correct data being driven
> and sampled at correct window. As this seems applicable for dwmmc, let
> me take a look into other vendor implementation. Thanks again for the
> advise :)

OK

Best regards,
Marek Vasut


More information about the U-Boot mailing list