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

[Fangrui Song <maskray@google.com>]

On 2022-04-29, Florian Weimer wrote:
>* Fangrui Song:
>
>> On 2022-04-29, Florian Weimer wrote:
>>>* Fangrui Song:
>>>
>>>>>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).
>>>
>>>How can we make sure that certain sections are covered by RELRO, and
>>>still get start/stop symbols for them?
>>
>> About __start_/__stop_:
>> __start_/__stop_ symbols are special. They don't need to be mentioned in a
>> linker script to take effects for lld and modern GNU ld. I haven't checked
>> whether there is an ancient GNU ld which requires symbol assignments when the
>> output section description is specified.
>
>According to the BFD ld documentation, __start_/__stop_ symbols are only
>generated if the resulting name is a C identifier and the name is the
>same.  That means that if we name the section .data.rel.ro.vtables or
>something like that (to benefit from .data.rel.ro.* placement), we won't
>get the start/stop symbols.  If we use a C identifier name for the
>section, it won't be covered by RELRO.

I know the rule.  I meant I don't find a way making lld build have
RELRO __libc_* without the help of ld.bfd --verbose ;-)

>> About RELRO:
>> It seems that there is no section name convention to make a section
>> RELRO.  Having a SECTIONS command with DATA_SEGMENT_ALIGN /
>> DATA_SEGMENT_RELRO_END can specify all RELRO sections. I do not know a
>> simpler approach than postprocessing ld.bfd --verbose output.
>>
>> To make GNU ld work, the following linker script is sufficient:
>>
>>   SECTIONS {
>>            __libc_subfreeres : { *(__libc_subfreeres) }
>>            __libc_atexit : { *(__libc_atexit) }
>>            __libc_IO_vtables : { *(__libc_IO_vtables) }
>>   } INSERT BEFORE .data.rel.ro;
>
>That might work.  But if we cut down the number of vtables, I think we
>can eliminate those special RELRO sections altogether.
>
>We have been slowly migrating off __libc_subfreeres and the like because
>having shutdown order depend on link is very non-obvious and
>problematic.  For example, for optimimum results, we want to shut down
>malloc last on a thread.

If there is a way to get rid of __libc_* output sections, it will
certainly be great!

>Thanks,
>Florian
>