[llvm-dev] RFC: LLD range extension thunks

[Peter Smith via llvm-dev]

I agree with Sean, most of the work is plumbing within the ELF linker.
I expect that the approach could be translated to the COFF linker but
the majority of the implementation details would be different.

On 19 January 2017 at 05:50, Sean Silva via llvm-dev
<llvm-dev at lists.llvm.org> wrote:
>
>
> On Wed, Jan 18, 2017 at 7:00 PM, Saleem Abdulrasool via llvm-dev
> <llvm-dev at lists.llvm.org> wrote:
>>
>> Out of curiosity, would this be done in a object format agnostic way?
>> Windows ARM requires the same branch island support, and the logic for the
>> detection and placement could be shared across ELF and COFF I suspect.
>
>
> I think that most of the code changes needed for this are going to be just
> refactoring to "put a loop around the Writer" and so is inherently specific
> to ELF and COFF (since they have different Writer's). Initial thunk
> placement heuristics/algorithms etc. are likely to be pretty simple (i.e.
> too small lines of code to really share meaningfully). If we start doing
> something really complicated for placing thunks it may make sense to factor
> out so it can be shared, but it may end up like ICF where the algorithms we
> use are the same, but the details of the data structures and semantics
> differ enough that it's not clear if sharing them would be a net savings or
> not.
>
> -- Sean Silva
>
>>
>>
>> On Wed, Jan 18, 2017 at 3:59 PM, Simon Atanasyan via llvm-dev
>> <llvm-dev at lists.llvm.org> wrote:
>>>
>>>
>>>
>>> On Jan 19, 2017 2:48 AM, "Ed Maste" <emaste at freebsd.org> wrote:
>>>
>>> On 4 January 2017 at 13:34, Peter Smith via llvm-dev
>>> <llvm-dev at lists.llvm.org> wrote:
>>> > I'm about to start working on range extension thunks in lld. This is
>>> > an attempt to summarize the approach I'd like to take and what the
>>> > impact will be on lld outside of thunks.
>>>
>>> Now that LLD works well for FreeBSD/amd64 (and arm64 is very close)
>>> I'm looking at other architectures, starting with mips64. The
>>> statically-linked toolchain components currently fail to link with an
>>> out of range jump, so I'm very interested in seeing this work
>>> progress. Are you looking at only arm and AArch64? Once the
>>> infrastructure is in I'll try to take a look at mips if nobody else
>>> does first.
>>>
>>>
>>> I'm waiting for this changes too. Now mips thunks places at the end of
>>> the corresponding section. Not sure about FreeBSD but on Linux that leads to
>>> incorrect code in case of static linking -- a thunk goes between crt*.o
>>> files which needs to be "joined" together. Gnu linker puts thunks to the
>>> separate section. We need to do the same thing.
>>>
>>> _______________________________________________
>>> LLVM Developers mailing list
>>> llvm-dev at lists.llvm.org
>>> http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-dev
>>>
>>
>>
>>
>> --
>> Saleem Abdulrasool
>> compnerd (at) compnerd (dot) org
>>
>> _______________________________________________
>> LLVM Developers mailing list
>> llvm-dev at lists.llvm.org
>> http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-dev
>>
>
>
> _______________________________________________
> LLVM Developers mailing list
> llvm-dev at lists.llvm.org
> http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-dev
>