[PATCH 1/2] xenon_sdhci: support for HS200 mode

[Pali Rohár]

On Monday 31 August 2020 10:10:50 Andre Heider wrote:
> On 31/08/2020 09:50, Pali Rohár wrote:
> > On Monday 31 August 2020 05:25:37 Andre Heider wrote:
> > > From: Wojciech Macek <wma at semihalf.com>
> > > 
> > > Add support for Marvell Xenon SDHCI HS200 mode.
> > > 
> > > Changes focue mostly on correct PHY initialization.
> > > All procedure is similar to the one done by Linux
> > > driver, but simplified.
> > > 
> > > Change-Id: I5e2396eeb23784f495abc18ea5a2eb7a92390730
> > > Signed-off-by: Wojciech Macek <wma at semihalf.com>
> > > Reviewed-on: http://vgitil04.il.marvell.com:8080/59230
> > > Tested-by: iSoC Platform CI <ykjenk at marvell.com>
> > > Reviewed-by: Grzegorz Jaszczyk <jaz at semihalf.com>
> > > Reviewed-by: Kostya Porotchkin <kostap at marvell.com>
> > > Reviewed-by: Igal Liberman <igall at marvell.com>
> > > [a.heider: adapt to mainline]
> > > Signed-off-by: Andre Heider <a.heider at gmail.com>
> > > ---
> > > Missing downstream patch, noticed while diffing branches:
> > > https://github.com/MarvellEmbeddedProcessors/u-boot-marvell/commit/387232507a0d9dda3990284221eaf87d7541dd02
> > 
> > Hello Andre! Why is this patch needed? Or what it is fixing?
> > I tested upstream U-Boot with all previous patches on Turris MOX and
> > Espressobin and SD cards worked fine.
> 
> My primary aim was to reduces the diff against the downstream fork, to make
> upstream work just as well.

I think the point is not to copy+paste code from Marvell's fork to
upstream U-Boot as is, but rather to provide missing functionality or
fixing bugs in upstream U-Boot.

I guess that U-Boot maintainers do not want new code which nobody knows
what is doing or do not know why is needed or do not know if it even
works for specified HW (like 3720 DB or uDPU board).

> But the HS200/400 mode is not used by those two boards, but by:
> arch/arm/dts/armada-3720-uDPU.dts
> arch/arm/dts/armada-3720-db.dts
> 
> And my other patch for espressobin-emmc is using it too. They're
> independent, so it works without it, just not as fast.
> 
> Having said that, I don't have either of those boards, so I haven't tested
> it myself. But the patches have been downstream for over two years, so I'd
> say they're well tested since everybody runs a downstream firmware (which
> will hopefully change now ;)

Yea, they are probably tested, but with historic U-Boot version -- which
is not enough.

The best thing would be to find somebody who wants to test changes on
real HW.