Re: [RFC PATCH v2 4/4] mmc: sdhci-esdhc-imx: using pinmux subsystem

[Shawn Guo]

On Thu, Dec 15, 2011 at 02:23:36PM +0100, Sascha Hauer wrote:
> On Thu, Dec 15, 2011 at 08:17:54PM +0800, Shawn Guo wrote:
> > On Thu, Dec 15, 2011 at 12:54:28PM +0100, Sascha Hauer wrote:
> > > On Thu, Dec 15, 2011 at 07:53:05PM +0800, Shawn Guo wrote:
> > > > On Thu, Dec 15, 2011 at 07:28:18PM +0800, Dong Aisheng-B29396 wrote:
> > > > > > -----Original Message-----
> > > > > > 
> > > > > My understanding is that pinmux_get will return an error if no proper pinmux
> > > > > Found without dummy pinmux. That's a real error.
> > > > > But with dummy pinmux, if no proper pinmux found, the pinctrl core may check
> > > > > If dummy pinmux is supported, if supported, it will fakely success with returning
> > > > > a dummy pinmux. Then real error is hiden.
> > > > > This is due to for supporting one single image, the dummy pinmux may also be enabled
> > > > > For platforms like mx6q with real pinmux.
> > > > > 
> > > > Got your point.  So your concern is about the fake success when dummy
> > > > pinmux comes to play rather than 'fake error'.  I'm not concerned about
> > > > that case much, since we will know the failure with a simple debug/info
> > > > message in the pinmux core, saying 'I do not find any available pinmux,
> > > > and I'm falling into the dummy pinmux'.  When you see this message
> > > > imx6q, you know something goes wrong.
> > > 
> > > So i.MX3/5 people must know that it's safe to ignore this message
> > > whereas i.MX6 people must know there's something wrong Adding messages
> > > saying "there might or might not be something wrong" is not good.
> > > People will frequently ask on the mailing list about these messages.
> > > 
> > Usually, people checking the serial output and asking question on list
> > are developers not users.  It's not too hard for developers to figure
> > out such message.  Or we can making it a debug message, and when some
> > driver is not working, turn on debug option, you know where it goes
> > wrong.
> 
> So we as developers have to figure out ourselves what can go wrong and
> the users, well who cares about users?
> 
It's really the type of problem that can be discovered and fixed in
the development phase, isn't it?

BTW, I did not mean to leave imx3 and imx5 there forever, but mean,
with the dummy pinmux support, we do not have to strictly migrate imx3,
imx5 and imx6 all at the same time.

-- 
Regards,
Shawn

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/