Re: [PATCH] hda: add sigmatel dig mic support and refactor 9227

[Tobin Davis <tdavis@dsl-only.net>]

On Wed, 2006-10-25 at 17:41 -0400, Matt Porter wrote:
> On Wed, Oct 25, 2006 at 01:53:56PM -0700, Tobin Davis wrote:
> > Couple of things, which version is this against?  I tried patching
> > against the tip and all areas were rejected.
>  
> Hrm, interesting. I pulled and update my hg tree. I'll take a look.
> 
> > Also, the original 9227 and 927x code was written specifically for Intel
> > D965 board support.  The information I had was from the Stac 9227
> > developers reference and several systems I was working with at the time
> > (I was a contractor at Intel).  From what I can tell, your patch removes
> 
> Actually, the core code wasn't written specifically for the D965
> support. I wrote and submitted support for the 927x code against
> the sigmatel reference codec board before these patches for D965
> got in. The 927x support was committed 23 Jan 2006, 9227/8/9 codec
> support via the 927x patch path was committed 06 July 2006. Those
> are the portions I authored. It looks like the D965/9227 specific
> support showed up on 22 Aug 2006. All according to the hg logs.
> 
Yes, this was my patch specifically for the new Intel motherboards.

> > the 3stack configuration and replaces it with the ref config, which
> > could break the 3-stack systems.  Has this patch been tested on these
> > systems?
> 
> This patch does no such thing. The reason i say "refactor" is that
> it properly separates 927x and 9227 codecs. They are different. The
> Intel patch just hacked in D965 support for the 9227 using the
> 927x path. If you look closely, this patch moves all of those 9227-based
> D965 systems over to the 9227 table leaving the 927x table with just
> the reference board.  This is needed because the pin widgets on the
> 9227 and 927x differ in some ways including the digital mic pin
> widget nids.
> 
Unfortunately, the D965 5stack 7.1 surround sound boards have the 9271
chips, not the 9227.  I broke the 9227 board away from the 922x
configuration because the codec changed functionally between the
9221/2/3 (found on Intel D945 boards) and 9227 (found on D965 3 stack
boards).  But the functionality between the 9227 and the 927x was
similar enough to keep together (I only had 3 days with a 5 stack board
before my contract ended).  So if the 927x is being split from the 9227
code, the D965-5stack configuration needs to follow, as that is what
will be detected by the preset function.

I wish I had access to the new codec specs.  It would have made my last
week at Intel much easier.

-- 
Tobin Davis <tdavis@dsl-only.net>


-------------------------------------------------------------------------
Using Tomcat but need to do more? Need to support web services, security?
Get stuff done quickly with pre-integrated technology to make your job easier
Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo
http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642