Re: [PATCH v3 2/2] Default to --with-default-link=no (bug 25812)

[Fangrui Song <maskray@google.com>]

On 2022-04-28, Florian Weimer wrote:
>* Fangrui Song:
>
>> I use chromium's prebuilt clang/lld and do something like:
>>
>>   curl -s https://raw.githubusercontent.com/chromium/chromium/main/tools/clang/scripts/update.py | python3 - --output-dir=~/Stable
>>   sudo ln -sf ~/Stable/bin/lld /usr/local/bin/ld
>>
>> so that gcc will use ld.lld.
>>
>> ld.lld doesn't have a built-in linker script (ld.bfd --verbose), so $@T will be empty and `test -s $@T` will fail.
>> I wonder how to better support ld.lld with the default --with-default-link=no ...
>
>How does lld place sections without a linker script?  Purely based on
>names?

It entirely uses code to describe the built-in rules like gold, while
GNU ld uses an internal linker script plus built-in code.
(some rules cannot be described by the linker script language and needs
code anyway).

There is no support dumping lld/gold's built-in rules into a linker
script.

I will add DATA_SEGMENT_RELRO_END support to ld.lld:
https://reviews.llvm.org/D124656 . With the patch, ld.lld can read shlib.lds
transformed from ld.bfd --verbose output.

I can think of 3 ways for lld:

* ask lld users to use --with-default-link=yes
* remove `test -s $@T` (if lld)
* add -fuse-ld=bfd

>> Besides, the position of .hash in --with-default-link=no seems weird.
>> 71213dc2589554dc8f8061e9b04e80c55d098b6a (2006-09) moved it to the end
>> of the RO segment.
>
>I think it's odd that we have an old-style hash table at all.  Do we
>really need it?  All tools should have migrated by now.

We can remove .hash for non-mips architectures.

(I am a bit sad .MIPS.xhash support went in, but thankfully it doesn't take
too much code.  The translation table could be solved in a better way. I
suspect the whole thing is unnecessary with RELR.)