Hi Ben,

On Thu, 16 Jul 2015, I wrote:

On Wed, 15 Jul 2015, I wrote:

quoted

On Tue, 14 Jul 2015, Benjamin Herrenschmidt wrote:

quoted

Maybe we should have a dedicated accessor for "mac_xpram" ...

...

The arch_nvram_ops methods don't deal with structures like partitions ...

Instead of the accessor you suggested, perhaps it would be better to add a 
method like arch_nvram_ops.get_partition, to replace the 
pmac_get_partition() exported function?

The call sites for pmac_get_partition() are in the implementation of the 
IOC_NVRAM_GET_OFFSET ioctl that's used with /dev/nvram, and in 
pmac_xpram_read(). pmac_xpram_write() has no caller and could be removed.

But this doesn't have much to do with linux-fbdev. I think the old 
NV_CMODE/NV_VMODE issues*, which this patch avoids, are irrelevant to the 
problem of nvram module re-use, which is the aim of this patch series.

But if those issues really are relevant then we should move the discussion 
to the revised patch, that is, [RFC v6 16/25] powerpc, fbdev: Use NV_CMODE 
and NV_VMODE only when CONFIG_PPC32 and CONFIG_PPC_PMAC and CONFIG_NVRAM.

(There was no response to any patch in RFC v6 from any PowerPC 
maintainers, which is why I've revived this thread.)

* https://lists.ozlabs.org/pipermail/linuxppc-dev/2001-November/012662.html

--