-next fails to boot as of today on S3C6410

[Jamie Iles]

On Wed, Nov 23, 2011 at 01:50:34PM +0000, Mark Brown wrote:
> On Wed, Nov 23, 2011 at 01:32:37PM +0000, Jamie Iles wrote:
> > On Wed, Nov 23, 2011 at 01:01:35PM +0000, Mark Brown wrote:
> 
> > > I tried adding some yesterday around the VIC registration in the CPU
> > > code but it didn't actually appear on the console so I'm none the wiser.
> 
> > OK.  You mentioned you had trouble reverting the patch yesterday - I've 
> > pushed a branch (git://github.com/jamieiles/linux-2.6-ji.git 
> > vic-dt-revert-next) which is today's next with all of the VIC patches 
> > reverted if you want to try that to be sure.  I've built s3c6400_defconfig 
> > successfully, but can't test it unless there's a qemu model somewhere?
> 
> It wasn't the mechanics of reverting that caused problems, it was the
> fact that when I reverted things still didn't work.  Actually, I've gone
> back and retested with the for-next branch from today and it appears
> that for some reason the issue has gone away there which is a bit
> confusing though as I write this it occurs to me that I did turn device
> tree back on (my board doesn't use it but I typically build it in) which
> I had off when I tested yesterday and would explain the issue.

OK, I've just tried hacking something on picoxcell that uses IRQ numbers 
starting at 32 rather than 0 and registering the vic manually rather 
than using the device tree binding and it all runs nicely.  So USE_OF is 
enabled, but it's as close as I can get to your system and it isn't 
using any of the DT stuff in the VIC driver.

> It looks like there's a second bug breaking the boot in there - testing
> your commit the board boots but if I test the tip of Russell's for-next
> then that breaks again.  I'm just starting another bisect and
> considering sending a patch to add a git finger-point alias.

Thanks for your patience Mark!  Please let me know if there's anything I 
can do to help.

Jamie