Section mismatches in arch/powerpc/kernel/head_64.S

[Michael Ellerman]

Hi Laura,

On Thu, 2015-11-12 at 13:39 -0800, Laura Abbott wrote:
> Hi,
 
> There seem to be section mismatches coming from head_64.S
> 
> WARNING: vmlinux.o(.text+0x8994): Section mismatch in reference from the
> variable __boot_from_prom to the function .init.text:prom_init()
> The function __boot_from_prom() references
> the function __init prom_init().
> This is often because __boot_from_prom lacks a __init
> annotation or the annotation of prom_init is wrong.
> 
> WARNING: vmlinux.o(.text+0x8c30): Section mismatch in reference from the
> variable start_here_multiplatform to the function .init.text:early_setup()
> The function start_here_multiplatform() references
> the function __init early_setup().
> This is often because start_here_multiplatform lacks a __init
> annotation or the annotation of early_setup is wrong.
> 
> WARNING: vmlinux.o(.text+0x8c54): Section mismatch in reference from the
> variable start_here_common to the function .init.text:setup_system()
> The function start_here_common() references
> the function __init setup_system().
> This is often because start_here_common lacks a __init
> annotation or the annotation of setup_system is wrong.
> 
> WARNING: vmlinux.o(.text+0x8c68): Section mismatch in reference from the
> variable start_here_common to the function .init.text:start_kernel()
> The function start_here_common() references
> the function __init start_kernel().
> This is often because start_here_common lacks a __init
> annotation or the annotation of start_kernel is wrong.

These all look correct, ie. they are references from non-init code to init
code. I don't think any of them are real bugs, because that is all early boot
code anyway which should never run after boot, or if it did that would be a big
anyway.

I'm not sure how we fix them. Some of that code might be able to be moved out
of head_64.S and made __init, but then you just have the same problem again.

It looks like x86 calls through a pointer and marks that as __REFDATA, but
that's kind of gross just to avoid a section mismatch warning.

So I'm not sure what the solution is here.

cheers