[i2c] Where do all the patches go?

[Jean Delvare]

Hi Nishanth,

On Wed, 25 Apr 2007 19:38:59 -0500, Nishanth Menon wrote:
> Hi Jean,
> 
> Am not sure if this question has been asked before.. but checking the
> archives and the FAQs did not help me :(.
> 
> What is the development procedure for i2c developers?
> a) Patches such as David's patch which provide intrusive changes.
> b) Minor updates/patches to the rest of drivers (adapters, chips etc).
> 
> I do see pending patches being kept here:
> http://khali.linux-fr.org/devel/linux-2.6/jdelvare-i2c/ . This I believe
> applies for (a) and (b). I suppose if a patch in the list ends up here,
> we can assume it is "accepted" for mainline?

Almost, yes. The patches staging at the location above are the ones
that I want Andrew Morton to include in his -mm kernel. As long as the
patch isn't in Linus' tree, there still is a chance that it is removed
from my patch queue before I send the patches to Linus, if anyone
objects to that particular patch, for example because testing in -mm
revealed problems. Ideally, I send them all to Linus once every two or
three months, during the merge window for the next upstream kernel.

> However, am not clear on which git tree is the first recipient for the
> patches so that some dim blokes like me can start pulling the latest
> code base (SVN repo I believe is 2.4 kernel only)? I suppose all trees
> finally end up in Linus's tree.. but then, which is the first kernel
> which includes the patches(I cant find something like a i2c git tree - I
> dont think it exists..)?

I am building a new temporary git tree each time I need to send patches
to Linus, I delete it when Linus is done. The rest of the time, the i2c
patches are only available as a quilt series, at the location above. No
git tree.

And you are right, the SVN i2c repository is for Linux 2.4 and has
nothing to do with Linux 2.6.

> Further how about the lm_sensor tools..i2cdump etc.. do the
> "corresponding" version remain in svn for the latest (kernel+patches in
> khali website)?

The user-space i2c tools are essentially independent from the kernel
version these days, so this shouldn't be a problem.

-- 
Jean Delvare