Re: [PATCH v3 16/16] objtool,x86: Rewrite retpoline thunk calls

[Nathan Chancellor <nathan@kernel.org>]

On 6/7/2021 1:54 PM, 'Nick Desaulniers' via Clang Built Linux wrote:
> On Mon, Jun 7, 2021 at 2:46 AM Peter Zijlstra <peterz@infradead.org> wrote:
>>
> 
> Thanks, the below diff resolves the linker error reported in
> https://github.com/ClangBuiltLinux/linux/issues/1388
> 
> Both readelf implementations seem happy with the results, too.
> 
> Tested-by: Nick Desaulniers <ndesaulniers@google.com>
> 
> Nathan,
> Can you please test the below diff and see if that resolves your boot
> issue reported in:
> https://github.com/ClangBuiltLinux/linux/issues/1384

Unfortunately, it does not appear to resolve that issue.

$ git log -2 --decorate=no --oneline
eea6a9d6d277 Peter's fix
614124bea77e Linux 5.13-rc5

$ strings /mnt/c/Users/natec/Linux/kernel-investigation | grep microsoft
5.13.0-rc5-microsoft-standard-WSL2-00001-geea6a9d6d277 
(nathan@archlinux-ax161) #3 SMP Tue Jun 8 09:46:19 MST 2021

My VM still never makes it to userspace.

>> ---
>>   tools/objtool/elf.c | 25 ++++++++++++++++++++++++-
>>   1 file changed, 24 insertions(+), 1 deletion(-)
>>
>> diff --git a/tools/objtool/elf.c b/tools/objtool/elf.c
>> index 743c2e9d0f56..41bca1d13d8e 100644
>> --- a/tools/objtool/elf.c
>> +++ b/tools/objtool/elf.c
>> @@ -717,7 +717,7 @@ static int elf_add_string(struct elf *elf, struct section *strtab, char *str)
>>
>>   struct symbol *elf_create_undef_symbol(struct elf *elf, const char *name)
>>   {
>> -       struct section *symtab;
>> +       struct section *symtab, *symtab_shndx;
>>          struct symbol *sym;
>>          Elf_Data *data;
>>          Elf_Scn *s;
>> @@ -769,6 +769,29 @@ struct symbol *elf_create_undef_symbol(struct elf *elf, const char *name)
>>          symtab->len += data->d_size;
>>          symtab->changed = true;
>>
>> +       symtab_shndx = find_section_by_name(elf, ".symtab_shndx");
>> +       if (symtab_shndx) {
>> +               s = elf_getscn(elf->elf, symtab_shndx->idx);
>> +               if (!s) {
>> +                       WARN_ELF("elf_getscn");
>> +                       return NULL;
>> +               }
>> +
>> +               data = elf_newdata(s);
>> +               if (!data) {
>> +                       WARN_ELF("elf_newdata");
>> +                       return NULL;
>> +               }
>> +
>> +               data->d_buf = &sym->sym.st_size; /* conveniently 0 */
>> +               data->d_size = sizeof(Elf32_Word);
>> +               data->d_align = 4;
>> +               data->d_type = ELF_T_WORD;
>> +
>> +               symtab_shndx->len += 4;
>> +               symtab_shndx->changed = true;
>> +       }
>> +
>>          sym->sec = find_section_by_index(elf, 0);
>>
>>          elf_add_symbol(elf, sym);
> 
> 
> The only thing that's still different is that the `file` command still
> prints "no section header."
> 
> $ find . -name \*.lto.o | xargs file | rev | cut -d , -f 1 | rev |
> sort | uniq -c
>        1  no section header
>     8377  not stripped
>        1  too many section headers (33683)
>        1  too many section headers (50758)
> $ file --version
> file-5.39
> 
> That's drivers/gpu/drm/amd/amdgpu/amdgpu.lto.o, fs/xfs/xfs.lto.o,
> drivers/gpu/drm/i915/i915.lto.o, respectively.  I'm not sure that's a
> problem, yet, and whether 9bc0bb50727c8ac69fbb33fb937431cf3518ff37 is
> even related yet; those might just be huge drivers and figured it was
> reporting somewhere in case it ever comes up again.  CONFIG_LTO
> implies -ffunction-sections -fdata-sections, and
> CONFIG_LD_DEAD_CODE_DATA_ELIMINATION explicitly sets those, too.
>