[PATCH v2] scripts: Do not strip .rela.dyn section

[Alexandre Ghiti]

On 07/04/2025 10:15, Alexandre Ghiti wrote:
> Hi Masahiro,
>
> On Fri, Apr 4, 2025 at 5:25 PM Masahiro Yamada <masahiroy at kernel.org> wrote:
>> On Fri, Apr 4, 2025 at 12:45 AM Alexandre Ghiti <alex at ghiti.fr> wrote:
>>> Hi Ard,
>>>
>>> On 03/04/2025 17:11, Ard Biesheuvel wrote:
>>>> On Thu, 3 Apr 2025 at 16:42, Alexandre Ghiti <alexghiti at rivosinc.com> wrote:
>>>>> riscv uses the .rela.dyn section to relocate the kernel at runtime but
>>>>> that section is stripped from vmlinux. That prevents kexec to
>>>>> successfully load vmlinux since it does not contain the relocations info
>>>>> needed.
>>>>>
>>>> Maybe explain that .rela.dyn contains runtime relocations, which are
>>>> only emitted if they are actually needed - as opposed to the static
>>>> relocations that are not emitted as SHF_ALLOC sections, and are not
>>>> considered to be part of the runtime image in the first place.
>>>
>>> Ok I'll do.
>>>
>>>
>>>> It
>>>> would be nice if we could use --remove-relocations= here, which only
>>>> removes static relocations, but unfortunately, llvm-objcopy does not
>>>> support this.
>>>>
>>>> Also, I wonder if this should apply to all of .rel.dyn, .rela.dyn and
>>>> .relr.dyn, as they all carry runtime relocations.
>>>
>>> Ok, I'll add them to the next version.
>>>
>>>
>>>> Finally, I'd be curious to know why RISC-V relies on --emit-relocs in
>>>> the first place? Is the relocs check really needed? If not, it would
>>>> be a nice opportunity to get rid of Makefile.postlink entirely.
>>>
>>> So I had to check and it happens that this was an issue with the
>>> toolchain, I should check if that still happens with newer ones.
>>>
>>> commit 559d1e45a16dcf1542e430ea3dce9ab625be98d0
>>> Author: Alexandre Ghiti <alexghiti at rivosinc.com>
>>> Date:   Wed Mar 29 06:53:29 2023 +0200
>>>
>>>       riscv: Use --emit-relocs in order to move .rela.dyn in init
>>
>>
>>
>> So,
>>
>> Fixes: 559d1e45a16d ("riscv: Use --emit-relocs in order to move
>> .rela.dyn in init")
>>
>> Is this the correct tag?
> This is the initial culprit yes, but if we use this tag, the fix won't
> apply. So I decided to pick Ard's patch so that this fix can be easily
> backported to 6.14, and I'll come up with a new version for previous
> releases. Is that ok with you?


And I have just looked at 6.15-rc1 and noticed that the relocation 
stripping was moved to Makefile.vmlinux, so this fix won't apply to 6.14 
neither.

I'll then use the initial culprit, which is Fixes: 559d1e45a16d ("riscv: 
Use --emit-relocs in order to move.rela.dyn in init").

Sorry for the noise,

Alex


>
> Thanks,
>
> Alex
>
>>
>>
>>
>>
>> --
>> Best Regards
>> Masahiro Yamada
> _______________________________________________
> linux-riscv mailing list
> linux-riscv at lists.infradead.org
> http://lists.infradead.org/mailman/listinfo/linux-riscv