Re: [PPC64/Power7 - 2.6.35-rc5] Bad relocation warnings whileBuilding a CONFIG_RELOCATABLE kernel with CONFIG_ISERIES enabled

[Alexander Graf <agraf@suse.de>]

Milton Miller wrote:
> I wrote:
>   
>> On Mon Jul 19 2010 at about 03:36:51 EST, Alexander Graf wrote:
>>     
>>> On 19.07.2010, at 03:11, Benjamin Herrenschmidt wrote:
>>>
>>>       
>>>> On Thu, 2010-07-15 at 17:05 +0530, Subrata Modak wrote:
>>>>         
>>>>> commit e62cee42e66dcca83aae02748535f62e0f564a0c solved the problem for
>>>>> 2.6.34-rc6. However some other bad relocation warnings generated against
>>>>> 2.6.35-rc5 on Power7/ppc64 below:
>>>>>
>>>>> MODPOST 2004 modules^M
>>>>> WARNING: 2 bad relocations^M
>>>>> c000000000008590 R_PPC64_ADDR32 .text+0x4000000000008460^M
>>>>> c000000000008594 R_PPC64_ADDR32 .text+0x4000000000008598^M
>>>>>           
>>>> I think this is KVM + CONFIG_RELOCATABLE. Caused by:
>>>>
>>>> .global kvmppc_trampoline_lowmem
>>>> kvmppc_trampoline_lowmem:
>>>> .long kvmppc_handler_lowmem_trampoline - CONFIG_KERNEL_START
>>>>
>>>> .global kvmppc_trampoline_enter
>>>> kvmppc_trampoline_enter:
>>>> .long kvmppc_handler_trampoline_enter - CONFIG_KERNEL_START
>>>>
>>>> Alex, can you turn these into 64-bit on ppc64 so the relocator
>>>> can grok them ?
>>>>         
>>> If I turn them into 64-bit, will the values be > RMA? In that case
>>> things would break anyways. How does relocation work on PPC? Are the
>>> first few megs copied over to low memory? Would I have to mask anything
>>> in the above code to make sure I use the real values? 
>>>
>>> Alex
>>>
>>>       
>> You can still do the subtraction, but you have to allocate 64 bits for
>> storage.  Relocatable ppc64 kernels work by adjusting PPC64_RELOC_RELATIVE
>> entries during early boot (reloc in reloc_64.S called from head_64.S).
>>
>> The code purposely only supports 64 bit relative addressing.
>>     
>
> Oh yea, and for book-3s, the code copies from 0x100 to __end_interrupts
> in arch/powerpc/kernel/exceptions-64s.h down to the real 0, but the rest
> of the kernel is at some disjointed address.  The interrupt will go to
> the copy at the real zero.  Any references to code outside that region
> must be done via a full indrect branch (not a relative one), simiar to
> the secondary startup (via following the function pointer in a descriptor
> set in very low memory), or syscall entry and exception vectors via paca.
>   

That would still break on normal PPC boxes, as any address accessed in
real mode has to be inside the RMA. And the #include for
kvm/book3s_rmhandlers.S happens after __end_interrupts. So I'd end up
with code that gets executed outside of the RMA after a relocation, right?


Alex