Re: [PATCH 3/6] ir-kbd-i2c: Switch to the new-style device binding model

[Jean Delvare <khali@linux-fr.org>]

On Mon, 06 Apr 2009 23:10:36 +0200, hermann pitton wrote:
> Am Montag, den 06.04.2009, 10:40 +0200 schrieb Jean Delvare:
> > Anyone out there with a MSI TV@nywhere Plus that could help with
> > testing?
> 
> Here is a link to one of the initial reports by Henry, others are close
> to it.
> 
> http://marc.info/?l=linux-video&m=113324147429459&w=2
> 
> There are two different variants of that MSI card, but that undocumented
> KS003 chip is the same on them.

Great, thanks for the pointer. If I understand correctly, the KS003
has a state machine flow which causes the chip to stop answering when
an invalid address is used on the bus and start answering again when a
valid address other than his own is used. As the old i2c model relied a
lot on probing, I am not surprised that this was a problem in the past.
But with the new model, probes should become infrequent, so I suspect
that the workaround may no longer be needed... except when i2c_scan=1
is used.

I'd rather keep the workaround in place for the time being, and only
once the ir-kbd-i2c changes have settled, try to remove it if someone
really cares.

-- 
Jean Delvare