On Mon, Aug 7, 2017 at 9:21 PM, Michael Ellerman [off-list ref] wrote:

Rob Herring [off-list ref] writes:

quoted

On Tue, Jul 25, 2017 at 4:44 PM, Rob Herring [off-list ref] wrote:

quoted

This series is the last steps to remove storing the full path for every
DT node. Instead, we can create full path strings dynamically as needed
with printf %pOF specifiers (commit ce4fecf1fe15). There are a number of
remaining direct users of full_name after this series. I don't believe
there should be any functional impact for those users with the change to
only the node name (+unit-address). The majority are for struct
resource.name. This should only affect /proc/iomem display.

Patches 1 and 2 can be applied now for 4.14. For patches 3 and 4, my
target is 4.15 after all the dependencies have been merged.

PPC folks, Please test! The PPC parts are untested. A git branch with
all the dependencies is here[1].

PPC folks, any chance to test this?

I got stuck on your %pOF conversion, which broke the vio.c code, because of:

-               if (!strcmp(parent_node->full_name, "/ibm,platform-facilities"))
+               if (!strcmp(parent_node->full_name, "ibm,platform-facilities"))

But full_name hasn't been changed yet.

Ah, those lines need to be dropped from the %pOF conversion. I'll send
a new version.

What I was originally considering here was just:

strcmp(kbasename(parent_node->full_name), "ibm,platform-facilities")

That would work for "/foo/ibm,platform-facilities" too, but IMO
validation of the DT is not the kernel's job (if it is, we're doing a
horrible job). But in the end here, I decided to keep the existing
full path matching as patch 1 does.

But patch 1 here should fix that, so I'll pull it all together and try
and get it tested.

Thanks.

Rob