[U-Boot] [PATCH v2 0/6] Add support for Colibri Vybrid Modules

[Marek Vasut <marex@denx.de>]

On Monday, April 13, 2015 at 08:21:27 AM, maitysanchayan at gmail.com wrote:
> Hello,
> 
> On 15-04-13 01:28:07, Marek Vasut wrote:
> > On Sunday, April 12, 2015 at 07:33:41 AM, maitysanchayan at gmail.com wrote:
> > > Hi,
> > > 
> > > On 15-04-11 18:46:15, Marek Vasut wrote:
> > > > On Wednesday, April 08, 2015 at 03:20:04 PM, Sanchayan Maity wrote:
> > > > > Hello,
> > > > > 
> > > > > This is the second version of the patchset which adds support for
> > > > > the Toradex Colibri Vybrid VF50 and VF61 modules. Boot up has been
> > > > > tested using the serial loader over UART.
> > > > 
> > > > Hi!
> > > > 
> > > > Stefano, maybe you can pick 1,2,3,4,6 right away for next? This way,
> > > > we can solve 5/6 separately, since it might need some further
> > > > dicussion. I don't want that patch to block the rest of the series
> > > > though.
> > > 
> > > The sixth patch has some minor dependencies on the fifth patch like
> > > some #defines and configuration we enable for Colibri. In case the
> > > fifth patch will require some more discussion and time, then I guess I
> > > will need to rework the sixth patch and then everything related to USB
> > > can go in a later patch. The rework will be minor. So the v3 will be
> > > without USB support
> > 
> > Well, if you want to rework the 6th patch, that'd be very nice of you :)
> 
> Yes sure. Will put the USB support in the end or not, the first patches
> can go in atleast since there seem to be no complaints on them so far.
> 
> > > and USB will go in later once this gets applied. Perhaps we
> > > will send the USB along with some more other featureset addition like
> > > gpio support for Vybrid which my colleague did recently along with some
> > > driver model/dt stuff.
> > 
> > My impression is that the USB is just a matter of figuring out where the
> > PHY registers go -- whether in arch/arm/include/imx-common or into
> > include/usb/ . I'd be more inclined for the later, since the PHY regs
> > looks like chipidea IP specific thing, but I might be wrong.
> 
> I would say it looks like chipidea specific but I have not looked at IPs
> other than Freescale controllers to be sure.

I think it's CI-specific indeed, thus include/usb/ . +CC Peter, who might
chime in whether the PHY is CI-specific or IMX-specific.

> > > By the way, where the fifth patch is concerned, is it related to the
> > > discussion Fabio pointed out in the discussion on first version of the
> > > patch? uboot trying to get rid of the 'no register access via offset'
> > > rule? Anything required I can rework accordingly.
> > 
> > I was concerned about the placement of the PHY reg definition. I didn't
> > hear from Fabio on the result of this decision and I don't really mind
> > that either way.
> 
> Ok. So I assume the driver can resort to the way it was in the first
> version while being changed with regards to three points 1. Use of
> clrsetbits_ related functions 2. Register phy definitions going in
> either of imx-common/ or include/usb/ with chipidea name. 3. No need to
> use offset accesses with the struct{} in light of the discussion Fabio
> pointed out.

Yup, that'd be nice. Thanks!

Best regards,
Marek Vasut