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: PATCH: Add x32 abilist files


> While I like the aim of moving entirely to sysdeps files for this, the 
> move for x86_64 has affected the gnu/lib-names.h contents for x86 - I've 
> filed <http://sourceware.org/bugzilla/show_bug.cgi?id=14112> for this.  
> Any ideas for how the gnu/lib-names.h build for x86 can get information 
> from shlib-versions files in x86_64 sysdeps directories (and likewise for 
> all the other biarch/triarch cases as they move to sysdeps directories)?

I was vaguely concerned about this but forgot to verify that HJ had really
tested everything that's relevant.  HJ says he has a fix but has not shared
any hints.  Perhaps he has something cleaner than what I've thought of.

Options I see:

1. Put the x86_64.* patterns back into the main file, and use
   %ifdef __ILP32__
   to distinguish x32.  But if the i686 compiler defines __ILP32__ too,
   then this will get the wrong LD_SO.  However, that needs a solution
   anyway, since for i686/x86_64/x32 the one <gnu/lib-names.h> is supposed
   to be triarch.

2. Put them into a new common place like sysdeps/x86/shlib-versions.
   Then make sysdeps/{x86_64,i386}/Implies list x86.

I don't think we can really decide on what the right solution is until we
have a plan for generating triarch lib-names.h.

Unless we figure that out in the next few hours, I think the right thing
for now is to revert HJ's commit so that the i686 build does not regress
before we resolve this properly.


Thanks,
Roland


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