Re: [patch 2/3] CONFIG_HIGHPTE vs. sub-page page tables.

[Russell King <rmk+lkml@arm.linux.org.uk>]

On Mon, Feb 04, 2008 at 02:51:33AM -0800, Andrew Morton wrote:
> If this situation (conflicting changes and poor code quality) persists into
> the 2.6.25 cycle I will toss all the subsystem trees out of -mm, shall
> rebase -mm on mainline and shall merge first.  I had decided today to
> actually just do this, but on reflection I'll give it just one more shot.

Can I too whinge about that?

Shortly after the 2.6 merge window opened, various changes went in which
completely broke a number of the merged changes in the ARM tree.  That
resulted in the stuff which I thought was safe to merge becoming unsafe,
and with that I dropped all the changes which conflicted.

In some cases, these merge conflicts came about due to a bug fix I had
to put in to make the kernel bootable on ARM again.

I'm still in the middle of rebuilding the resulting mess from that - and
we're not yet back to where we were prior to the 2.6.24 release.  So, the
current version of the ARM tree which you most likely pulled for -mm1 is
incomplete with respect to what was planned to go in.  Therefore, you can
expect to see quite a number of apparantly "new" changes appearing in it
as these problems are resolved.

They're not really new, they're just the old stuff with the merge conflicts
fixed.

I don't see any end to these bun fights at the start of the merge window.
I believe it's inevitable given the work flow that we're now using.

-- 
Russell King
 Linux kernel    2.6 ARM Linux   - http://www.arm.linux.org.uk/
 maintainer of: