[PATCH v7 3/4] drivers: of: add initialization code for dma reserved memory

On 9/18/2013 5:48 AM, Grant Likely wrote:

On Mon, 16 Sep 2013 06:43:08 -0700, Grant Likely [off-list ref] wrote:

quoted

On Mon, Sep 16, 2013 at 12:25 AM, Benjamin Herrenschmidt
[off-list ref] wrote:

quoted

On Mon, 2013-09-16 at 09:12 +0200, Marek Szyprowski wrote:

quoted

Fixing those issues requires further discussion. Frankly, right now I
really have no idea which way should I go. The /reserved-ranges node seems
to be easy to match particular reserved memory region with a given device.
I'm also not really convinced if it makes sense to add a code for finding
and matching a reserved memory region to every device driver, which might
need it. I would really like to get some more feedback on the Ben's
comment.

In any case, the code will also change significantly, so I assume that the
best, what can be done now is to revert the current version and start from
the scratch with a new, complete proposal.

Yes, let's do that. I will try to help. It shouldn't be that a big deal.

I think we can get the concerns sorted quickly. I'm at meetings all
day today in SJC, but I'll take a look at it this evening. I'll be in
Edinburgh. We're holding an ARM summit at the same time so a bunch of
affected ARM folks will be there.

After reading through the entirety of the other thread, I think this can
be fixed. My preference would be to fix in now rather than waiting
another cycle. If you haven't sent the revert yet then please hold off
for a bit (I can't tell at the moment since I'm currently on an airplane
and I didn't update my tree before leaving the ground).

I've didn't send revert yet, I can wait a few days more if you see that the
issues can be fixed instead of redoing all from scratch.

Best regards
-- 
Marek Szyprowski
Samsung R&D Institute Poland