Re: [PATCH 1/5] ARM: phyCORE-AM335x: Updated SPI NOR dts entries

From: Alexander Aring <alex.aring@gmail.com>
To: Sascha Hauer <s.hauer@pengutronix.de>
Cc: barebox@lists.infradead.org
Subject: Re: [PATCH 1/5] ARM: phyCORE-AM335x: Updated SPI NOR dts entries
Date: Fri, 5 Sep 2014 09:18:13 +0200	[thread overview]
Message-ID: <20140905071811.GB23258@omega> (raw)
In-Reply-To: <20140905065122.GB5352@pengutronix.de>

On Fri, Sep 05, 2014 at 08:51:22AM +0200, Sascha Hauer wrote:
> On Fri, Sep 05, 2014 at 08:32:00AM +0200, Alexander Aring wrote:
> > Hi Sascha,
> > 
> > On Fri, Sep 05, 2014 at 07:37:33AM +0200, Sascha Hauer wrote:
> > > Hi Teresa,
> > > 
> > > On Thu, Sep 04, 2014 at 08:50:36AM +0200, Teresa Gámez wrote:
> > > > Updated the SPI NOR flash entries. NOR flash got
> > > > detected but did not work.
> > > > 
> > > > - Updated muxing
> > > > - Fixed frequency
> > > > - Fixed CS
> > > > - Removed first compatible entry (the flashes used is changing frequently)
> > > > 
> > > > Signed-off-by: Teresa Gámez <t.gamez@phytec.de>
> > > 
> > > Applied up to 4/5. The Root partition start seems indeed wrong.
> > > 
> > 
> > yea. But sometimes you also do a "fixed while applying". Why not now?
> > Okay, maybe on the "good" days.
> 
> Maybe because you didn't write the correct start offset and I didn't
> feel like doing maths before the first coffee ;)
> 

hey, I wrote a correct start offset with a calculation example. In
my last mail to Teresa. I can also give you the advice, to use a calculator
when it's too early in the morning.

My problem is to write some english, but I can't use google translator
or something like that to check it. (Sometimes the author of lilo "Werner
Almesberger" review my bad english and helps me there). :-)

> > 
> > Nevertheless, isn't there any mechanism to check on conflicts while
> > creating partitions like request memory mapped regions?
> 
> No. Right now creating multiple conflicting partitions is a feature, see
> for example /dev/env0 and /dev/nor0.bareboxenv. Those are two
> partitions which both conflict but must both exist.
> 

Sounds for me like doing symlink magic. But I don't feel like to doing
magic in the morning. I don't drink coffee since I stopped working at
phytec. Does devfs support symlinks? -> First look: "no".

> I must admit it's a misfeature though and I wish we could do without it
> as it really can hide problems.
> 

If symlink magic doesn't work, maybe add a flag/type which allow to have a
conflict partition? This sounds more like a feature which doesn't help
to check on misconfigured partitions, you can simple switch it off.

I would prefer the symlink magic idea.

- Alex

_______________________________________________
barebox mailing list
barebox@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/barebox