Things seem to converge in Linus' tree now.
I've rerun my ctags+grep based analysis for all architecures on
2.6.15-rc1-git6 and 2.6.5-rc2 with the following patches from -mm applied:
fix-missing-includes-for-2614-git11.patch
fix-missing-includes-for-2615-rc1.patch
dont-include-schedh-from-moduleh.patch
I also compile-tested these trees on alpha, arm, i386, ia64, mips, powerpc,
ppc64, and x86_64 and checked that make -i outputs the same diagnostic
messages (modulo line number and include order changes) for allnoconfig,
defconfig, and allmodconfig, whether or not
dont-include-schedh-from-moduleh.patch was applied to the tree.
None of this brought up any problems.
So from my point of view these patches are okay to go into mainline now.
Andrew, what are your plans for these patches?
Shall I send an updated dont-include-schedh-from-moduleh.patch whose
changelog reflects the current state of testing?
Thanks,
Tim
Tim Schmielau <[email protected]> wrote:
>
> Andrew, what are your plans for these patches?
Slam it in after 2.6.15, put fingers in ears?
> Shall I send an updated dont-include-schedh-from-moduleh.patch whose
> changelog reflects the current state of testing?
You can send a new changelog if you like, but I'd prefer not to do a
wholesale replacement of a tested patch.
On Wed, 23 Nov 2005, Andrew Morton wrote:
> Tim Schmielau <[email protected]> wrote:
> >
> > Andrew, what are your plans for these patches?
>
> Slam it in after 2.6.15, put fingers in ears?
I'll certainly listen when the patches go in.
I'll rerun my analysis to catch most of the problems. But I will allways
be late by two or three days because the analysis takes it's time.
The (well-known) problem is that at the time when patches are allowed to
go in, the rate of change is so high that I need to fix some dozen files
per day.
In a more quiet period (e.g., two or the days after most of the changes
are in), I'd probably be able to catch most (if not all) of the problems
by myself with
We'll see how it goes.
I'd really like to get the procedure right as my next patch to really cut
down sched.h (ab)use will affect about 10x as many files.
> > Shall I send an updated dont-include-schedh-from-moduleh.patch whose
> > changelog reflects the current state of testing?
>
> You can send a new changelog if you like, but I'd prefer not to do a
> wholesale replacement of a tested patch.
>
OK. I'll not update the changelog now as it will be out of date anyways
when 2.6.15 arrives.
Thanks,
Tim