Re: [PATCH v3 5/9] OMAP4 : DSS2 : HDMI: HDMI driver header file addition

[Tomi Valkeinen <tomi.valkeinen@ti.com>]

On Tue, 2011-03-08 at 02:17 -0600, K, Mythri P wrote:
> Hi Tomi,
> 
> On Mon, Mar 7, 2011 at 3:16 PM, Tomi Valkeinen <tomi.valkeinen@ti.com> wrote:
> > On Fri, 2011-03-04 at 01:48 -0600, K, Mythri P wrote:
> >> Adding the hdmi interface driver header file (hdmi.h) to the dss driver.
> >> Register and timing declaration to be used by the corresponding c file
> >> is added in this file.
> >
> > The subject and description are wrong. Always before sending patches do
> > a quick sanity check, by check the subjects, looking at the stats in the
> > intro letter, etc.

<snip>

> >> +        * SW HACK : DDC needs time to stablize else it sometimes reads 0 values
> >> +        * or right shifted values.
> >> +        */
> >> +       usleep_range(800, 1000);
> >
> > This is still unclear. Is it an OMAP HW problem? OMAP HW feature? A
> > feature in HDMI TVs? A HDMI spec delay? Or unknown reason?
> >
> there is an internal DDC (i2c bus) , without this delay I have
> sometimes seen that while reading EDID
> we sometimes (Not always and only with some TV's , even with a
> particular TV it is not consistent)
> read a shifted value or EDID value is 0. I can clarify with the h/w
> team , but  can we add this as a s/w hack for time being?

Sure, I have nothing against this work-around as such. I just want it to
be clearly described in the comment. The comment should explain as much
as possible about the hack/workaround, because if the problem is not
described in the TRM/errata, the comment is the only place where the
reader can get information about it.

 Tomi