This is the mail archive of the libc-hacker@sources.redhat.com mailing list for the glibc project.

Note that libc-hacker is a closed list. You may look at the archives of this list, but subscription and posting are not open.


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

Re: [PATCH] GLIBC prelinking part


Jakub Jelinek <jakub@redhat.com> writes:

> a) whether something like this could be acceptable for glibc 2.3
>    (with l_machine_specific replaced with what David Mosberger suggested,
>     ie. architecture specific sub-structure of link_map)
>    Note that non-REL architectures will probably need far less sysdep
>    changes than IA-32 (well, aside from mips)

I think this is fine.  Don't know about the machine specific
definitions, yet, since I haven't seen it.

> b) if my choice of new SHT_ constant and new DT_ constants is right, or
>    whether that should be coordinated with Sun somehow, or whether I should
>    choose constants far from the numbers used by Sun

When it's time to add the patch I'll ask the people at Sun what they
have allocated and reserve some numbers close by.  I don't want to use
a new range since this means additional tests when accessing the data.
This is why I'm not suggesting to reuse the DT_MIPS_ and SHT_MIPS_
numbers.

> c) whether you agree on:
>    the format of .gnu.conflict (ie. ElfW(Rela) with ELFW(R_SYM) zero)
>    the format of .gnu.liblist (ie. ElfW(Lib))

In general, yes.  But I want to check what SGI did.



Two problems:

- all uses of _dl_debug_mask should be protected with __builtin_expect

- the patch definitely should be split.  The caching stuff should be
  independent from the rest.

-- 
---------------.                          ,-.   1325 Chesapeake Terrace
Ulrich Drepper  \    ,-------------------'   \  Sunnyvale, CA 94089 USA
Red Hat          `--' drepper at redhat.com   `------------------------


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