On Mon, Feb 6, 2012 at 3:08 PM, Ian Campbell
<ijc@hellion.org.uk> wrote:
On Mon, 2012-02-06 at 11:42 -0600, Jonathan Nieder wrote:
> Ian Campbell wrote:
>
> > I suspect this is due to the lack of this u-boot patch:
> > http://lists.denx.de/pipermail/u-boot/2012-February/117020.html
> >
> > I found that without this my 3.2 dreamplug kernel would not boot (with
> > the 2011.12-2 package from debian). It's related to
> > CONFIG_ARM_PATCH_PHYS_VIRT.
>
> For the record, this proposed u-boot change has been filed as
> <http://bugs.debian.org/658904>.
>
> The inability to work with versions of u-boot without that change
> (workaround?) in the default configuration still would seem to be a
> regression in the kernel, so pointers toward solving it would be very
> welcome.
I hacked my kernel (the Kconfig files) to allow me to turn off
PATCH_PHYS_VIRT and it also worked.
However AIUI PATCH_VIRT_PHYS is part of the infrastructure which allows
a single kernel flavour to boot on multiple machines so turning it off
is not a viable workaround.
I think this is just a latent bug in u-boot which has been exposed by a
recent kernel changes and should be treated as a bug to be fixed in
u-boot. I personally wouldn't consider it a regression in the kernel (at
least based on my understanding of what's going on).
Ian.