[PATCH v3 1/1] Add bigobj support to AArch64 COFF

[Jan Beulich]

On 07.11.2025 11:58, Richard Earnshaw (lists) wrote:
> On 07/11/2025 10:53, Jan Beulich wrote:
>> On 27.10.2025 21:13, Evgeny Karpov wrote:
>>> During Boost library testing on aarch64-w64-mingw32, it appeared that 2^16
>>> sections are not enough. It can be handled by using the bigobj format to extend
>>> the total amount of sections to 2^32. This patch adds bigobj support to
>>> AArch64 COFF in a similar way to how it is done for x86_64.
>>>
>>> Co-authored-by: Martin Vejbora <mvejbora@microsoft.com>
>>
>> Unless you have a copyright assignment on record with the FSF, this needs your
>> S-o-b.
>>
>> Further, your mailer has line-wrapped the patch, which means it won't apply if
>> someone (me?) would want to commit this on your behalf. (Besides the patch
>> being more difficult to review this way, of course.)
>>
>>> --- a/bfd/config.bfd
>>> +++ b/bfd/config.bfd
>>> @@ -252,7 +252,7 @@ case "${targ}" in
>>>      ;;
>>>    aarch64-*-pe* | aarch64-*-mingw*)
>>>      targ_defvec=aarch64_pe_le_vec
>>> -    targ_selvecs="aarch64_pe_le_vec aarch64_pei_le_vec
>>> aarch64_elf64_le_vec aarch64_elf64_be_vec aarch64_elf32_le_vec
>>> aarch64_elf32_be_vec arm_elf32_le_vec arm_elf32_be_vec pdb_vec"
>>> +    targ_selvecs="aarch64_pe_le_vec aarch64_pei_le_vec
>>> aarch64_pe_big_le_vec aarch64_elf64_le_vec aarch64_elf64_be_vec
>>> aarch64_elf32_le_vec aarch64_elf32_be_vec arm_elf32_le_vec
>>> arm_elf32_be_vec pdb_vec"
>>
>> I continue to be in trouble with the chosen naming. The side by side appearance
>> of "big" and "le" is disturbing. No matter that x86 uses merely "big", I think
>> "bigobj" would be better for architectures where endianness also needs recording
>> in the names.
>>
>> Otherwise looks okay to me, but will of course need an Arm64 maintainer's okay
>> as well.
> 
> If bigobj is making symbols too long, then how about 'bo' instead?

I wouldn't mind that very much (it's a bit cryptic), but I don't think name length
really matters here?

Jan