On Tue, 2015-09-15 at 18:30 -0700, Mitchel Humpherys wrote:
> Any overlap in the reserved memory regions (those specified in the
> reserved-memory DT node) is a bug.
Can you expand a bit on why you think it's a bug? I assume it was discussed at
some point on the list but I didn't see it sorry.
There's nothing I can see in the binding document[1] about whether regions can
overlap, or what it would mean if they did.
If we want to declare that overlapping regions are always a bug then there
should be some text in the binding explaining that. There's also the
possibility that we have existing device trees in the wild that contain
overlapping regions, and whether we think it's OK to retrospectively declare
that they're incorrect.
cheers
[1]: https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/tree/Documentation/devicetree/bindings/reserved-memory/reserved-memory.txt
On Thu, Nov 12 2015 at 01:19:59 PM, Michael Ellerman <[email protected]> wrote:
> On Tue, 2015-09-15 at 18:30 -0700, Mitchel Humpherys wrote:
>
>> Any overlap in the reserved memory regions (those specified in the
>> reserved-memory DT node) is a bug.
>
> Can you expand a bit on why you think it's a bug? I assume it was discussed at
> some point on the list but I didn't see it sorry.
The reason I think it's a bug is because the overlapping memory could be
handed out to multiple firmwares, which generally ends in "random"
firmware crashes. We've found by sad experience that root-causing such
a crash can be quite difficult.
Is there a valid use case for overlapping regions? I can't think of
one...
> There's nothing I can see in the binding document[1] about whether regions can
> overlap, or what it would mean if they did.
You're right, the bindings document doesn't say anything about
overlapping memory regions. I can submit something unless someone comes
up with a reason why we should allow overlapping memory regions.
> If we want to declare that overlapping regions are always a bug then there
> should be some text in the binding explaining that. There's also the
> possibility that we have existing device trees in the wild that contain
> overlapping regions, and whether we think it's OK to retrospectively declare
> that they're incorrect.
I did a quick survey of in-tree users of reserved-memory and couldn't
find any overlapping regions.
-Mitch
--
Qualcomm Innovation Center, Inc.
The Qualcomm Innovation Center, Inc. is a member of the Code Aurora Forum,
a Linux Foundation Collaborative Project