Re: Do defconfigs work for 2.6.32-rc5 mainline kernels?

[Tony Lindgren <tony@atomide.com>]

* Aguirre Rodriguez, Sergio Alberto <saaguirre@ti.com> [091102 12:32]:
>  
> 
> > -----Original Message-----
> > From: linux-omap-owner@vger.kernel.org 
> > [mailto:linux-omap-owner@vger.kernel.org] On Behalf Of Tim Bird
> > Sent: Monday, November 02, 2009 12:21 PM
> > To: Bill Gatliff
> > Cc: linux-omap@vger.kernel.org
> > Subject: Re: Do defconfigs work for 2.6.32-rc5 mainline kernels?
> > 
> > Bill Gatliff wrote:
> > > I'm trying to build a kernel for beagleboard, using the 
> > defconfig found
> > > in 2.6.32-rc5 and the 2.6.32-rc5 source code.
> > > 
> > > Said kernel hangs after the uncompressor, I'm just trying 
> > to figure out
> > > if this is a known issue or not.  The defconfig in the 
> > linux-omap tree
> > > does seem to work with the linux-omap source code, FWIW...
> > 
> > Short story:
> > I found that the omap OSK defconfig is broken in mainline also.
> > 
> > Long story:
> > I found that the defconfig in the mainline 2.6.32-rc1 kernel
> > didn't work on an OMAP OSK board.  Thinking there was some
> > problem in the code itself, I bisected from a known working
> > config on 2.6.29 all the way back up to 2.6.32-rc1.  When
> > I was done I was at the current code base, but with a .config
> > that had about 650 differences.
> > 
> > I haven't had time to go back and find the individual config
> > option that changed to break things.  I would need some kind
> > of "config bisect" to find this efficiently.
> > 
> > I didn't think to try the OSK defconfig in the omap tree.
> > (Silly me, might have saved me some grief. :-)
> > 
> > Is there a maintainer for individual defconfig files?
> > In other words, if I repair the OSK defconfig, who do
> > I send it to?

Just send the fixes to this list as usual.
 
> Why don't you guys enable Kernel Low level prints and send the logs?

Right, maybe checkout omap-debug branch, and enable DEBUG_LL to see
what happens? Also, some compilers have bugs with CONFIG_ARM_UNWIND=y
AFAIK.
 
> Otherwise its pretty hard to see what's wrong.

Agreed. I've been booting various boards with the mainline kernel,
including OSK here and there. There should not be anything seriously
wrong. And we still have time to send in fixes for this -rc cycle.

Tony