Hi Rob, Mark, Jon,


Setting aside the future plan, anyway we should avoid
relying on an undocumented compatible string.
(and we should avoid its spread as Rob pointed out.)

I rephrase the git-log  (no update in the code-diff).

Could you check v2, please?
https://patchwork.kernel.org/patch/8529701/

Better?





2016-03-04 5:23 GMT+09:00 Rob Herring [off-list ref]:

On Thu, Mar 3, 2016 at 4:12 AM, Jon Medhurst (Tixy) [off-list ref] wrote:

quoted

On Thu, 2016-03-03 at 12:07 +0900, Masahiro Yamada wrote:
[...]

quoted

This patch is derived from Rob Herring' comment
"BTW, we should also kill off "amba-bus" which is an ambiguous term"
in the following thread:
http://lkml.iu.edu/hypermail/linux/kernel/1601.0/01822.html


So, the plan would be like this:

[1] Make device trees not depend on "arm,amba-bus"   (this commit)
[2] New device trees should no longer use "arm,amba-bus" alone.
[3] Go though some releases until we do not care about the backward
compatibility

Why would we stop caring about backwards compatibility? If I was a user
of any of the platforms in question and updated my kernel, I wouldn't
expect to have to debug why it was broken, then install a new dtb to fix
it - which may be a tricky thing to do, depending on the firmware used
to boot Linux.

If no one notices, then there is no ABI. Things break all the time
when nobody cares. It will be more than a few releases for sure. We
may never get to step 3, but we should do step 1 so people stop
copy-n-pasting this in step 2.

Rob
--
To unsubscribe from this list: send the line "unsubscribe devicetree" in
the body of a message to majordomo at vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

-- 
Best Regards
Masahiro Yamada