This is the mail archive of the libc-alpha@sourceware.org mailing list for the glibc project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: [COMMITTED] alpha: Fix static gettimeofday symbol


* Adhemerval Zanella:

> On 13/02/2020 09:04, Florian Weimer wrote:
>> * Adhemerval Zanella:
>> 
>>> By undef strong_alias on alpha implementation, the
>>> default_symbol_version macro becomes an empty macro on static build.
>>> It fixes the issue introduced at c953219420.
>> 
>> Aren't we using a different mechanism for this in other cases?
>> Like defining macros for the alias targets?
>> 
>> In any case, SET_VERSION for this macro isn't very descriptive.
>
> I don't have a strong opinion here, on some ifunc variant definitions
> some implementation undef/redefine the alias macros.  The problems with 
> this approach are:
>
>   1. once you have composed macros (such as default_symbol_version).
>     
>   2. if you want to use the same macro you need to undef after the
>      common case inclusion. 
>
> On both cases you will need to redefined it using libc-symbol.h 
> internals which defeats the usage of such macros imho (for instance
> the usage of _weak_alias or any other internal macros).

I thought we had cases where we redefined the *target* identifiers of
these aliases, but maybe I'm wrong about that.

But I'm sure we have cases somewhere where we have a -common.c file
which contains the code, without the aliases.

Thanks,
Florian


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]