Re: [PATCH] hppa: use hidden_def with internal pthread funcs

["Carlos O'Donell" <carlos at systemhalted dot org>]

On Mon, Nov 19, 2012 at 9:32 AM, Joseph S. Myers
<joseph@codesourcery.com> wrote:
> On Mon, 19 Nov 2012, Carlos O'Donell wrote:
>
>> Hopefully things are a little better now for hppa, all I have left are
>> the _STACK_GROWS_UP patches for nptl/.
>
> I think points (b), (e), (g), (h), (k), (l), (m), (n), (o), (q), (r), (s),
> (t), (u), (v), (x) and (y) from my previous list
> <http://sourceware.org/ml/libc-ports/2012-06/msg00048.html> are still
> current, along with points not on that list:
>
>   (z) There is no c++-types.data for hppa.
>
>   (aa) The TLS configure checks should be removed from
>   sysdeps/hppa/configure.in, now the rule is that we assume a
>   recent-enough binutils and such checks are no longer needed for each
>   architecture and have been removed for other architectures.
>
>   (ab) __ASSUME_LWS_CAS (definition and uses) should be removed now
>   2.6.16 is the global minimum kernel version.
>
>   (ac) lll_futex_timed_wait_bitset should be defined for hppa.
>
> (Some of the bits/*.h files where my previous list notes the hppa versions
> are out of date should perhaps be reworked along the lines used for
> bits/fcntl.h to reduce duplication between architecture versions - but
> such a reworking wouldn't be appropriate during a release freeze, whereas
> straightforward updates of the hppa versions would be appropriate then.)

Thanks, that is correct and matches my list of remaining things that
need fixing.

My comment above is simply that with the patches for nptl
_STACK_GROWS_UP cases hppa will be able to compile and work
out-of-the-box without patches.

Cheers,
Carlos.