[4/6] nptl: Move lowlevellock into libc

And export __lll_lock_wait and __lll_lock_wait_private as
GLIBC_PRIVATE.  This should eliminate the need for the previous
riscv kludge.
---
 nptl/Makefile                          |  3 +--
 nptl/Versions                          |  2 ++
 nptl/libc-lowlevellock.c               | 20 --------------------
 nptl/lowlevellock.c                    |  6 ++----
 sysdeps/nptl/lowlevellock.h            |  6 ++++--
 sysdeps/riscv/nptl/libc-lowlevellock.c |  8 --------
 6 files changed, 9 insertions(+), 36 deletions(-)
 delete mode 100644 nptl/libc-lowlevellock.c
 delete mode 100644 sysdeps/riscv/nptl/libc-lowlevellock.c

On 17/02/2021 14:21, Florian Weimer via Libc-alpha wrote:
> And export __lll_lock_wait and __lll_lock_wait_private as
> GLIBC_PRIVATE.  This should eliminate the need for the previous
> riscv kludge.

Should it mark BZ#15648 as fixed (Carlos has opened BZ#15658 to track
the LTO issue)?

LGTM, thanks

> ---
>  nptl/Makefile                          |  3 +--
>  nptl/Versions                          |  2 ++
>  nptl/libc-lowlevellock.c               | 20 --------------------
>  nptl/lowlevellock.c                    |  6 ++----
>  sysdeps/nptl/lowlevellock.h            |  6 ++++--
>  sysdeps/riscv/nptl/libc-lowlevellock.c |  8 --------
>  6 files changed, 9 insertions(+), 36 deletions(-)
>  delete mode 100644 nptl/libc-lowlevellock.c
>  delete mode 100644 sysdeps/riscv/nptl/libc-lowlevellock.c
> 
> diff --git a/nptl/Makefile b/nptl/Makefile
> index 5a7558d943..62f368af1e 100644
> --- a/nptl/Makefile
> +++ b/nptl/Makefile
> @@ -33,9 +33,9 @@ routines = \
>    forward \
>    libc-cancellation \
>    libc-cleanup \
> -  libc-lowlevellock \
>    libc_multiple_threads \
>    libc_pthread_init \
> +  lowlevellock \
>    old_pthread_cond_destroy \
>    old_pthread_cond_init \
>    pthread_atfork \

Ok.

> @@ -100,7 +100,6 @@ libpthread-routines = \
>    futex-internal \
>    herrno \
>    libpthread-compat \
> -  lowlevellock \
>    nptl-init \
>    nptlfreeres \
>    old_pthread_atfork \

Ok.

> diff --git a/nptl/Versions b/nptl/Versions
> index 7254ca4750..694747bb44 100644
> --- a/nptl/Versions
> +++ b/nptl/Versions
> @@ -77,6 +77,8 @@ libc {
>      __libc_current_sigrtmin_private;
>      __libc_dl_error_tsd;
>      __libc_pthread_init;
> +    __lll_lock_wait;
> +    __lll_lock_wait_private;
>      __pthread_attr_copy;
>      __pthread_attr_destroy;
>      __pthread_attr_init;

Ok.

> diff --git a/nptl/libc-lowlevellock.c b/nptl/libc-lowlevellock.c
> deleted file mode 100644
> index e840399c41..0000000000
> --- a/nptl/libc-lowlevellock.c
> +++ /dev/null
> @@ -1,20 +0,0 @@
> -/* Copyright (C) 2003-2021 Free Software Foundation, Inc.
> -   This file is part of the GNU C Library.
> -   Contributed by Paul Mackerras <paulus@au.ibm.com>, 2003.
> -
> -   The GNU C Library is free software; you can redistribute it and/or
> -   modify it under the terms of the GNU Lesser General Public
> -   License as published by the Free Software Foundation; either
> -   version 2.1 of the License, or (at your option) any later version.
> -
> -   The GNU C Library is distributed in the hope that it will be useful,
> -   but WITHOUT ANY WARRANTY; without even the implied warranty of
> -   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.	 See the GNU
> -   Lesser General Public License for more details.
> -
> -   You should have received a copy of the GNU Lesser General Public
> -   License along with the GNU C Library; if not, see
> -   <https://www.gnu.org/licenses/>.  */
> -
> -/* No difference to lowlevellock.c, except we lose a couple of functions.  */
> -#include <lowlevellock.c>

Ok.

> diff --git a/nptl/lowlevellock.c b/nptl/lowlevellock.c
> index 1ea397d5cb..7976c6d8f5 100644
> --- a/nptl/lowlevellock.c
> +++ b/nptl/lowlevellock.c
> @@ -35,10 +35,8 @@ __lll_lock_wait_private (int *futex)
>        futex_wait ((unsigned int *) futex, 2, LLL_PRIVATE); /* Wait if *futex == 2.  */
>      }
>  }
> +libc_hidden_def (__lll_lock_wait_private)
>  
> -
> -/* This function doesn't get included in libc.  */
> -#if IS_IN (libpthread)
>  void
>  __lll_lock_wait (int *futex, int private)
>  {
> @@ -52,4 +50,4 @@ __lll_lock_wait (int *futex, int private)
>        futex_wait ((unsigned int *) futex, 2, private); /* Wait if *futex == 2.  */
>      }
>  }
> -#endif
> +libc_hidden_def (__lll_lock_wait)

Ok.

> diff --git a/sysdeps/nptl/lowlevellock.h b/sysdeps/nptl/lowlevellock.h
> index 0a2b413f86..176ba96251 100644
> --- a/sysdeps/nptl/lowlevellock.h
> +++ b/sysdeps/nptl/lowlevellock.h
> @@ -75,8 +75,10 @@
>  #define lll_cond_trylock(lock)	\
>    __glibc_unlikely (atomic_compare_and_exchange_bool_acq (&(lock), 2, 0))
>  
> -extern void __lll_lock_wait_private (int *futex) attribute_hidden;
> -extern void __lll_lock_wait (int *futex, int private) attribute_hidden;
> +extern void __lll_lock_wait_private (int *futex);
> +libc_hidden_proto (__lll_lock_wait_private)
> +extern void __lll_lock_wait (int *futex, int private);
> +libc_hidden_proto (__lll_lock_wait)
>  
>  /* This is an expression rather than a statement even though its value is
>     void, so that it can be used in a comma expression or as an expression

Ok.

> diff --git a/sysdeps/riscv/nptl/libc-lowlevellock.c b/sysdeps/riscv/nptl/libc-lowlevellock.c
> deleted file mode 100644
> index 9523fb46a5..0000000000
> --- a/sysdeps/riscv/nptl/libc-lowlevellock.c
> +++ /dev/null
> @@ -1,8 +0,0 @@
> -/* This kludge works around a libpthread static linking problem:
> -   https://sourceware.org/bugzilla/show_bug.cgi?id=15648.  */
> -
> -#ifndef SHARED
> -# define __lll_lock_wait_private weak_function __lll_lock_wait_private
> -#endif
> -
> -#include <lowlevellock.c>
> 

Ok.

* Adhemerval Zanella:

> On 17/02/2021 14:21, Florian Weimer via Libc-alpha wrote:
>> And export __lll_lock_wait and __lll_lock_wait_private as
>> GLIBC_PRIVATE.  This should eliminate the need for the previous
>> riscv kludge.
>
> Should it mark BZ#15648 as fixed (Carlos has opened BZ#15658 to track
> the LTO issue)?

Makes sense, I've added the [BZ #15648] reference.  The multiple
definition issue will definitely be gone after this change.

Thanks,
Florian