Hi Mark, On Wed, Oct 04, 2017 at 04:44:29PM +0100, Mark Rutland wrote: > On Wed, Oct 04, 2017 at 05:16:16PM +0200, Marcus Folkesson wrote: > > The uniq ID is usually exported as ATAG_SERIAL but in case of > > devicetrees, this information is not passed. > > > > The uniq ID is stored in OCOTP memory (bank 0, Word 1,2) on imx CPUs. > > Read the ID and set as system serial number. > > > > This function is activated by setting `read-system-serial` > > dt property for the imx-ocotp node. > > > > Signed-off-by: Marcus Folkesson <[email protected]> > > --- > > > > Comment: > > I'm not sure if this functionality really should be in a device driver. > > Maybe it should be in arch/arm/mach-imx/. However, it's not > > completely wrong so I give it a shot. > > > > .../devicetree/bindings/nvmem/imx-ocotp.txt | 1 + > > drivers/nvmem/imx-ocotp.c | 38 > > ++++++++++++++++++++++ > > 2 files changed, 39 insertions(+) > > > > diff --git a/Documentation/devicetree/bindings/nvmem/imx-ocotp.txt > > b/Documentation/devicetree/bindings/nvmem/imx-ocotp.txt > > index 70d791b03ea1..bf0cc9e5e50f 100644 > > --- a/Documentation/devicetree/bindings/nvmem/imx-ocotp.txt > > +++ b/Documentation/devicetree/bindings/nvmem/imx-ocotp.txt > > @@ -16,6 +16,7 @@ Required properties: > > > > Optional properties: > > - read-only: disable write access > > +- read-system-serial: read uniq ID and export as system serial number. > > Available in /proc/cpuinfo > > As described, this is a pure SW option ( > > /proc/cpuinfo is a not a HW detail, and whether or not to export this > information there is a policy. So as described, this shouldn't be in the > DT.
Ok, I agree with that. > > Are you just trying to describe whether hte information is available in > the OCOTP memory? No, the information is available for all concerned iMX CPUs what I can tell. My intention was to flag if the driver should make use of it, but the DT seems to be the wrong way to go. > > e.g. maybe this should be something like: > > - has-unique-id: OCOTP bank0, Words 1 & 2 contain a unique ID > identifying the system. > > Thanks, > Mark. I will come up with an v2 where this is not configurable. Thanks, Marcus Folkesson
