i.MX EHCI initialization

i.MX EHCI initialization

[Sascha Hauer]

Hi Daniel et all,

I see an increasing number of patches trying to change the
initialization sequence of the i.MX ehci controller. It's getting harder
and harder to keep track what the changes are, what other boards
and SoCs a single patch affects and whether it just reverts another
patch.

I thought about making the initialization completely board specific and
creating a set of helper functions which the board can call if it wants
to to prevent too much code duplication.

What do you think?

Sascha

-- 
Pengutronix e.K.                           |                             |
Industrial Linux Solutions                 | http://www.pengutronix.de/  |
Peiner Str. 6-8, 31137 Hildesheim, Germany | Phone: +49-5121-206917-0    |
Amtsgericht Hildesheim, HRA 2686           | Fax:   +49-5121-206917-5555 |

i.MX EHCI initialization

[Daniel Mack]

Hi Sascha,

On Mon, May 17, 2010 at 11:36:52AM +0200, Sascha Hauer wrote:
> I see an increasing number of patches trying to change the
> initialization sequence of the i.MX ehci controller. It's getting harder
> and harder to keep track what the changes are, what other boards
> and SoCs a single patch affects and whether it just reverts another
> patch.
> 
> I thought about making the initialization completely board specific and
> creating a set of helper functions which the board can call if it wants
> to to prevent too much code duplication.

If we really need this, I'm pro adding that kind of interface, sure.
However, at least for the MX51 stuff, altering the init order wasn't
necessary at all eventually, and I'm not sure about other SoCs.

I also didn't follow all patches in this area I fear - could you
summarize which SoC wants it differently than the current code does it?

Thanks,
Daniel

i.MX EHCI initialization

[Sascha Hauer]

On Mon, May 17, 2010 at 11:59:55AM +0200, Daniel Mack wrote:
> Hi Sascha,
> 
> On Mon, May 17, 2010 at 11:36:52AM +0200, Sascha Hauer wrote:
> > I see an increasing number of patches trying to change the
> > initialization sequence of the i.MX ehci controller. It's getting harder
> > and harder to keep track what the changes are, what other boards
> > and SoCs a single patch affects and whether it just reverts another
> > patch.
> > 
> > I thought about making the initialization completely board specific and
> > creating a set of helper functions which the board can call if it wants
> > to to prevent too much code duplication.
> 
> If we really need this, I'm pro adding that kind of interface, sure.
> However, at least for the MX51 stuff, altering the init order wasn't
> necessary at all eventually, and I'm not sure about other SoCs.
> 
> I also didn't follow all patches in this area I fear - could you
> summarize which SoC wants it differently than the current code does it?

Valentin changed it with 2f0e40aba1cafe3a834bfcbac8f1e704d496dab9 and
Dinh changed it in his original patch. It may be a feeling at the
moment, but I think we should have an eye on the different SoC and board
specific things we add to the probe function.

Sascha


-- 
Pengutronix e.K.                           |                             |
Industrial Linux Solutions                 | http://www.pengutronix.de/  |
Peiner Str. 6-8, 31137 Hildesheim, Germany | Phone: +49-5121-206917-0    |
Amtsgericht Hildesheim, HRA 2686           | Fax:   +49-5121-206917-5555 |