[U-Boot] [linux-sunxi] Re: [PATCH v2 10/10] ARM: sun6i: Add Colombus board defconfig

[Ian Campbell]

On Sun, 2014-09-28 at 20:10 +0200, Hans de Goede wrote:
> On 09/28/2014 06:20 PM, Ian Campbell wrote:
> > On Sun, 2014-09-28 at 17:40 +0200, Hans de Goede wrote:

> >> Before you do that, note that I've just added 2 patches there, which I would
> >> like to get into v2014.10. Specifically I'm hoping that I can get some
> >> positive testing feedback on the bananapi gmac patch I've send (off-list),
> >> and I believe we really should try to get the bananapi fix into v2014.10,
> >> and if we're going todo a pull-req for v2014.10, we might as well include
> >> the 2 patches I've just added to next. Do you agree ?
> > 
> > You mean these two?
> >         sun7i: Add support for Olimex A20-OLinuXino-LIME2
> >         mmc: sunxi: add SDHC support for sun6i/sun7i/sun8i
> 
> Yes.
> 
> > The latter seems like a feature to me, or at least the changelog doesn't
> > give any rationale why it should go in now rather than waiting for the
> > next merge window (i.e. why it's a bugfix, what the upside is to justify
> > its inclusion now). How much testing has it had and what are the
> > potential downsides?
> 
> AFAIK the downside is that High Capacity cards will not work without it.
> 
> Looking at the code if this bit is set, then for some commands
> drivers/mmc/mmc.c or-s in OCR_HCS into the mmc cmdarg, so I guess you're
> right that this may cause some undesirable side effects, so lets delay
> this one.

OK.

> > WRT the new board (and new boards generally), I'm in two minds. On the
> > one hand they are pretty low risk (can't regress anything else, at least
> > not in this case), on the other we are 6 weeks past the close of the
> > merge window and 2 from the release date, so we are pretty far along.
> > Where do we draw the line?
> 
> Normally I would not include new boards at this moment in the cycle, but
> since we need to do a pull-req for the gmac anyways I thought it would
> be nice to have it included, esp. since many distros only spin things
> like sdcard boot images once, so if we do not include it now, many distros
> will not get it for a significant amount of time.

There's always Just One More Board(tm) ;-)

> Either way let me know how you want to proceed, if you think we should not
> include this, then I'll send a pull-req with only the gmac fix.

As I say I'm in two minds. I'm not really sure what the u-boot norm is
on this, I was hoping someone might chime in (although it's not been
very long and the thread topic doesn't exactly scream for attention).
Maybe run it by Albert/Tom and see how they feel about such things in
general?

Where run it by might be two alternate PRs? Or a PR structured so the
new board can trivially be dropped?

> > The gmac fix is a clear bug fix and once it is properly posted publicly
> > I will ack and then I agree it should go in.
> 
> I was hoping for Stephen to get around to testing it today, and then I wanted
> to send it out with his Tested-by. I'll just go and send it as is for now.

s/Stephen/Karsten/?

Ian.