Re: [GIT PULL] DT/core: cpu_ofnode updates for v3.12

On 13/08/13 20:45, Rafael J. Wysocki wrote:

On Tuesday, August 13, 2013 01:44:23 PM Rob Herring wrote:

quoted

On Tue, Aug 13, 2013 at 10:40 AM, Sudeep KarkadaNagesha
[off-list ref] wrote:

quoted

Adding PowerPC list

On 13/08/13 14:00, Rafael J. Wysocki wrote:

quoted

On Monday, August 12, 2013 02:27:47 PM Sudeep KarkadaNagesha wrote:

quoted

The following changes since commit
d4e4ab86bcba5a72779c43dc1459f71fea3d89c8:

Linux 3.11-rc5 (2013-08-11 18:04:20 -0700)

are available in the git repository at:

git://linux-arm.org/linux-skn.git cpu_of_node

[snip]

quoted

quoted

All error/warnings:

warning: (MPC836x_RDK && MTD_NAND_FSL_ELBC && MTD_NAND_FSL_UPM)
selects FSL_LBC which has unmet direct dependencies (FSL_SOC)
warning: (MPC836x_RDK && MTD_NAND_FSL_ELBC && MTD_NAND_FSL_UPM)
selects FSL_LBC which has unmet direct dependencies (FSL_SOC)
In file included from arch/powerpc/include/asm/kvm_para.h:26:0, from
include/uapi/linux/kvm_para.h:26, from include/linux/kvm_para.h:4,
from include/linux/kvm_host.h:30, from
arch/powerpc/kernel/asm-offsets.c:53:
include/linux/of.h:269:28: error: conflicting types for
'of_get_cpu_node'
extern struct device_node *of_get_cpu_node(int cpu); ^ In file
included from include/linux/of.h:139:0, from
arch/powerpc/include/asm/kvm_para.h:26, from
include/uapi/linux/kvm_para.h:26, from include/linux/kvm_para.h:4,
from include/linux/kvm_host.h:30, from
arch/powerpc/kernel/asm-offsets.c:53:
arch/powerpc/include/asm/prom.h:47:21: note: previous declaration
of 'of_get_cpu_node' was here
struct device_node *of_get_cpu_node(int cpu, unsigned int *thread);
^ make[2]: *** [arch/powerpc/kernel/asm-offsets.s] Error 1 make[2]:
Target `__build' not remade because of errors. make[1]: ***
[prepare0] Error 2 make[1]: Target `prepare' not remade because of
errors. make: *** [sub-make] Error 2

There seems to be conflict in the new function "of_get_cpu_node" added.
PowerPC also defines the same function name. Further microblaze and
openrisc declares it(can be removed) but doesn't define it.
To fix this:
1. I can rename the newly added function to something different like
   `of_get_cpunode` or
2. If of_* namespace should be used by only OF/FDT and not by any
   architecture specific code, then the arch specific version can be
   renamed to some thing like arch_of_get_cpu_node.
   Also most of the calls to arch specific function can be moved to
   generic code.

Let me know your thoughts.

It is up to Rafael if he is willing/able to rebase his tree, but I
would drop this series until this is sorted out.

=20
Yeah, I've just done that.
=20

Thanks Rafael, sorry for the trouble. I didn't expect of_* name-space to
be used in ARCH specific code.

quoted

I think the new common function should be and can be generalized to work=

 for

quoted

powerpc.
It would need to make reg property optional and pass in the device
node to the arch specific function.

A short term solution would be just to make the function "#ifndef CONFIG=

_PPC".

=20
I wouldn't do that, it's almost guaranteed to be messy going forward.
=20
I'd go for 1 above personally.

Even though it's easier approach, I would go for fixing PPC and converge
at generic of_get_cpu_node implementation if possible.

Regards,
Sudeep