Remove fno-unit-at-a-time make variable

  795985e4e751 in 2003 added -fno-unit-at-a-time to errlist.c and
siglist.c to "avoid reordering assembler output". -fno-toplevel-reorder
is a rough replacement for this legacy option
(https://sourceware.org/pipermail/gcc-patches/2006-January/186801.html).

The reordering requirement does not seem to be needed any longer.
---
 config.make.in        |  1 -
 configure             | 32 --------------------------------
 configure.ac          | 19 -------------------
 stdio-common/Makefile |  2 --
 4 files changed, 54 deletions(-)
  

On 30/03/2022 02:07, Fangrui Song wrote:
> 795985e4e751 in 2003 added -fno-unit-at-a-time to errlist.c and
> siglist.c to "avoid reordering assembler output". -fno-toplevel-reorder
> is a rough replacement for this legacy option
> (https://sourceware.org/pipermail/gcc-patches/2006-January/186801.html).
> 
> The reordering requirement does not seem to be needed any longer.

We still need them for otherwise DEFINE_COMPAT_ERRLIST used on errlist-compat.c
does not create _sys_errlist and _sys_siglist with expected sizes defined by
glibc ABI.

I am trying to fix without resorting to compiler options.

> ---
>  config.make.in        |  1 -
>  configure             | 32 --------------------------------
>  configure.ac          | 19 -------------------
>  stdio-common/Makefile |  2 --
>  4 files changed, 54 deletions(-)
> 
> diff --git a/config.make.in b/config.make.in
> index 6d43e691f7..ecde84a7da 100644
> --- a/config.make.in
> +++ b/config.make.in
> @@ -69,7 +69,6 @@ have-selinux = @have_selinux@
>  have-libaudit = @have_libaudit@
>  have-libcap = @have_libcap@
>  have-cc-with-libunwind = @libc_cv_cc_with_libunwind@
> -fno-unit-at-a-time = @fno_unit_at_a_time@
>  bind-now = @bindnow@
>  have-hash-style = @libc_cv_hashstyle@
>  use-default-link = @use_default_link@
> diff --git a/configure b/configure
> index 2a3cb49b0b..f836874a58 100755
> --- a/configure
> +++ b/configure
> @@ -620,7 +620,6 @@ libc_cv_cc_loop_to_function
>  libc_cv_cc_submachine
>  libc_cv_cc_nofma
>  libc_cv_mtls_dialect_gnu2
> -fno_unit_at_a_time
>  libc_cv_has_glob_dat
>  libc_cv_hashstyle
>  libc_cv_fpie
> @@ -6319,37 +6318,6 @@ fi
>  $as_echo "$libc_cv_has_glob_dat" >&6; }
>  
>  
> -{ $as_echo "$as_me:${as_lineno-$LINENO}: checking for -fno-toplevel-reorder -fno-section-anchors" >&5
> -$as_echo_n "checking for -fno-toplevel-reorder -fno-section-anchors... " >&6; }
> -if ${libc_cv_fno_toplevel_reorder+:} false; then :
> -  $as_echo_n "(cached) " >&6
> -else
> -  cat > conftest.c <<EOF
> -int foo;
> -EOF
> -if { ac_try='${CC-cc} $CFLAGS $CPPFLAGS -S -fno-toplevel-reorder -fno-section-anchors
> -			    conftest.c 1>&5'
> -  { { eval echo "\"\$as_me\":${as_lineno-$LINENO}: \"$ac_try\""; } >&5
> -  (eval $ac_try) 2>&5
> -  ac_status=$?
> -  $as_echo "$as_me:${as_lineno-$LINENO}: \$? = $ac_status" >&5
> -  test $ac_status = 0; }; }
> -then
> -  libc_cv_fno_toplevel_reorder=yes
> -else
> -  libc_cv_fno_toplevel_reorder=no
> -fi
> -rm -f conftest*
> -fi
> -{ $as_echo "$as_me:${as_lineno-$LINENO}: result: $libc_cv_fno_toplevel_reorder" >&5
> -$as_echo "$libc_cv_fno_toplevel_reorder" >&6; }
> -if test $libc_cv_fno_toplevel_reorder = yes; then
> -  fno_unit_at_a_time="-fno-toplevel-reorder -fno-section-anchors"
> -else
> -  fno_unit_at_a_time=-fno-unit-at-a-time
> -fi
> -
> -
>  { $as_echo "$as_me:${as_lineno-$LINENO}: checking for -mtls-dialect=gnu2" >&5
>  $as_echo_n "checking for -mtls-dialect=gnu2... " >&6; }
>  if ${libc_cv_mtls_dialect_gnu2+:} false; then :
> diff --git a/configure.ac b/configure.ac
> index fa7d3c025b..ec1db36d46 100644
> --- a/configure.ac
> +++ b/configure.ac
> @@ -1478,25 +1478,6 @@ fi
>  rm -f conftest*])
>  AC_SUBST(libc_cv_has_glob_dat)
>  
> -AC_CACHE_CHECK(for -fno-toplevel-reorder -fno-section-anchors, libc_cv_fno_toplevel_reorder, [dnl
> -cat > conftest.c <<EOF
> -int foo;
> -EOF
> -if AC_TRY_COMMAND([${CC-cc} $CFLAGS $CPPFLAGS -S -fno-toplevel-reorder -fno-section-anchors
> -			    conftest.c 1>&AS_MESSAGE_LOG_FD])
> -then
> -  libc_cv_fno_toplevel_reorder=yes
> -else
> -  libc_cv_fno_toplevel_reorder=no
> -fi
> -rm -f conftest*])
> -if test $libc_cv_fno_toplevel_reorder = yes; then
> -  fno_unit_at_a_time="-fno-toplevel-reorder -fno-section-anchors"
> -else
> -  fno_unit_at_a_time=-fno-unit-at-a-time
> -fi
> -AC_SUBST(fno_unit_at_a_time)
> -
>  AC_CACHE_CHECK([for -mtls-dialect=gnu2], libc_cv_mtls_dialect_gnu2,
>  [dnl
>  cat > conftest.c <<EOF
> diff --git a/stdio-common/Makefile b/stdio-common/Makefile
> index 435cd8904f..dbaecbd76e 100644
> --- a/stdio-common/Makefile
> +++ b/stdio-common/Makefile
> @@ -333,8 +333,6 @@ CFLAGS-isoc99_vfscanf.c += -fexceptions
>  CFLAGS-isoc99_vscanf.c += -fexceptions
>  CFLAGS-isoc99_fscanf.c += -fexceptions
>  CFLAGS-isoc99_scanf.c += -fexceptions
> -CFLAGS-errlist.c += $(fno-unit-at-a-time)
> -CFLAGS-siglist.c += $(fno-unit-at-a-time)
>  
>  # scanf14a.c and scanf16a.c test a deprecated extension which is no
>  # longer visible under most conformance levels; see the source files
  

On 2022-03-30, Adhemerval Zanella wrote:
>
>
>On 30/03/2022 02:07, Fangrui Song wrote:
>> 795985e4e751 in 2003 added -fno-unit-at-a-time to errlist.c and
>> siglist.c to "avoid reordering assembler output". -fno-toplevel-reorder
>> is a rough replacement for this legacy option
>> (https://sourceware.org/pipermail/gcc-patches/2006-January/186801.html).
>>
>> The reordering requirement does not seem to be needed any longer.
>
>We still need them for otherwise DEFINE_COMPAT_ERRLIST used on errlist-compat.c
>does not create _sys_errlist and _sys_siglist with expected sizes defined by
>glibc ABI.
>
>I am trying to fix without resorting to compiler options.

DEFINE_COMPAT_ERRLIST does not expand to code/data, just reordeable directives:


        .globl  __GLIBC_2_1_sys_errlist
.set __GLIBC_2_1_sys_errlist, _sys_errlist_internal
         .type   __GLIBC_2_1_sys_errlist,@object
         .size   __GLIBC_2_1_sys_errlist, 1000
         .globl  __GLIBC_2_1__sys_errlist
.set __GLIBC_2_1__sys_errlist, _sys_errlist_internal
         .type   __GLIBC_2_1__sys_errlist,@object
         .size   __GLIBC_2_1__sys_errlist, 1000
.symver __GLIBC_2_1_sys_nerr, sys_nerr@GLIBC_2.2.5
.symver __GLIBC_2_1__sys_nerr, _sys_nerr@GLIBC_2.2.5
.symver __GLIBC_2_1_sys_errlist, sys_errlist@GLIBC_2.2.5
.symver __GLIBC_2_1__sys_errlist, _sys_errlist@GLIBC_2.2.5
         .globl  __GLIBC_2_3_sys_errlist
.set __GLIBC_2_3_sys_errlist, _sys_errlist_internal
         .type   __GLIBC_2_3_sys_errlist,@object
         .size   __GLIBC_2_3_sys_errlist, 1008
         .globl  __GLIBC_2_3__sys_errlist
.set __GLIBC_2_3__sys_errlist, _sys_errlist_internal
         .type   __GLIBC_2_3__sys_errlist,@object
         .size   __GLIBC_2_3__sys_errlist, 1008
.symver __GLIBC_2_3_sys_nerr, sys_nerr@GLIBC_2.3
.symver __GLIBC_2_3__sys_nerr, _sys_nerr@GLIBC_2.3
.symver __GLIBC_2_3_sys_errlist, sys_errlist@GLIBC_2.3
.symver __GLIBC_2_3__sys_errlist, _sys_errlist@GLIBC_2.3
         .globl  __GLIBC_2_4_sys_errlist

I do not know whether GCC would reorder these macros. Even yes,
that'd just change the .symtab entries in the relocatable object file.
The linker behavior remains the same with reordering.

>> ---
>>  config.make.in        |  1 -
>>  configure             | 32 --------------------------------
>>  configure.ac          | 19 -------------------
>>  stdio-common/Makefile |  2 --
>>  4 files changed, 54 deletions(-)
>>
>> diff --git a/config.make.in b/config.make.in
>> index 6d43e691f7..ecde84a7da 100644
>> --- a/config.make.in
>> +++ b/config.make.in
>> @@ -69,7 +69,6 @@ have-selinux = @have_selinux@
>>  have-libaudit = @have_libaudit@
>>  have-libcap = @have_libcap@
>>  have-cc-with-libunwind = @libc_cv_cc_with_libunwind@
>> -fno-unit-at-a-time = @fno_unit_at_a_time@
>>  bind-now = @bindnow@
>>  have-hash-style = @libc_cv_hashstyle@
>>  use-default-link = @use_default_link@
>> diff --git a/configure b/configure
>> index 2a3cb49b0b..f836874a58 100755
>> --- a/configure
>> +++ b/configure
>> @@ -620,7 +620,6 @@ libc_cv_cc_loop_to_function
>>  libc_cv_cc_submachine
>>  libc_cv_cc_nofma
>>  libc_cv_mtls_dialect_gnu2
>> -fno_unit_at_a_time
>>  libc_cv_has_glob_dat
>>  libc_cv_hashstyle
>>  libc_cv_fpie
>> @@ -6319,37 +6318,6 @@ fi
>>  $as_echo "$libc_cv_has_glob_dat" >&6; }
>>
>>
>> -{ $as_echo "$as_me:${as_lineno-$LINENO}: checking for -fno-toplevel-reorder -fno-section-anchors" >&5
>> -$as_echo_n "checking for -fno-toplevel-reorder -fno-section-anchors... " >&6; }
>> -if ${libc_cv_fno_toplevel_reorder+:} false; then :
>> -  $as_echo_n "(cached) " >&6
>> -else
>> -  cat > conftest.c <<EOF
>> -int foo;
>> -EOF
>> -if { ac_try='${CC-cc} $CFLAGS $CPPFLAGS -S -fno-toplevel-reorder -fno-section-anchors
>> -			    conftest.c 1>&5'
>> -  { { eval echo "\"\$as_me\":${as_lineno-$LINENO}: \"$ac_try\""; } >&5
>> -  (eval $ac_try) 2>&5
>> -  ac_status=$?
>> -  $as_echo "$as_me:${as_lineno-$LINENO}: \$? = $ac_status" >&5
>> -  test $ac_status = 0; }; }
>> -then
>> -  libc_cv_fno_toplevel_reorder=yes
>> -else
>> -  libc_cv_fno_toplevel_reorder=no
>> -fi
>> -rm -f conftest*
>> -fi
>> -{ $as_echo "$as_me:${as_lineno-$LINENO}: result: $libc_cv_fno_toplevel_reorder" >&5
>> -$as_echo "$libc_cv_fno_toplevel_reorder" >&6; }
>> -if test $libc_cv_fno_toplevel_reorder = yes; then
>> -  fno_unit_at_a_time="-fno-toplevel-reorder -fno-section-anchors"
>> -else
>> -  fno_unit_at_a_time=-fno-unit-at-a-time
>> -fi
>> -
>> -
>>  { $as_echo "$as_me:${as_lineno-$LINENO}: checking for -mtls-dialect=gnu2" >&5
>>  $as_echo_n "checking for -mtls-dialect=gnu2... " >&6; }
>>  if ${libc_cv_mtls_dialect_gnu2+:} false; then :
>> diff --git a/configure.ac b/configure.ac
>> index fa7d3c025b..ec1db36d46 100644
>> --- a/configure.ac
>> +++ b/configure.ac
>> @@ -1478,25 +1478,6 @@ fi
>>  rm -f conftest*])
>>  AC_SUBST(libc_cv_has_glob_dat)
>>
>> -AC_CACHE_CHECK(for -fno-toplevel-reorder -fno-section-anchors, libc_cv_fno_toplevel_reorder, [dnl
>> -cat > conftest.c <<EOF
>> -int foo;
>> -EOF
>> -if AC_TRY_COMMAND([${CC-cc} $CFLAGS $CPPFLAGS -S -fno-toplevel-reorder -fno-section-anchors
>> -			    conftest.c 1>&AS_MESSAGE_LOG_FD])
>> -then
>> -  libc_cv_fno_toplevel_reorder=yes
>> -else
>> -  libc_cv_fno_toplevel_reorder=no
>> -fi
>> -rm -f conftest*])
>> -if test $libc_cv_fno_toplevel_reorder = yes; then
>> -  fno_unit_at_a_time="-fno-toplevel-reorder -fno-section-anchors"
>> -else
>> -  fno_unit_at_a_time=-fno-unit-at-a-time
>> -fi
>> -AC_SUBST(fno_unit_at_a_time)
>> -
>>  AC_CACHE_CHECK([for -mtls-dialect=gnu2], libc_cv_mtls_dialect_gnu2,
>>  [dnl
>>  cat > conftest.c <<EOF
>> diff --git a/stdio-common/Makefile b/stdio-common/Makefile
>> index 435cd8904f..dbaecbd76e 100644
>> --- a/stdio-common/Makefile
>> +++ b/stdio-common/Makefile
>> @@ -333,8 +333,6 @@ CFLAGS-isoc99_vfscanf.c += -fexceptions
>>  CFLAGS-isoc99_vscanf.c += -fexceptions
>>  CFLAGS-isoc99_fscanf.c += -fexceptions
>>  CFLAGS-isoc99_scanf.c += -fexceptions
>> -CFLAGS-errlist.c += $(fno-unit-at-a-time)
>> -CFLAGS-siglist.c += $(fno-unit-at-a-time)
>>
>>  # scanf14a.c and scanf16a.c test a deprecated extension which is no
>>  # longer visible under most conformance levels; see the source files
  

On 30/03/2022 13:23, Fangrui Song wrote:
> On 2022-03-30, Adhemerval Zanella wrote:
>>
>>
>> On 30/03/2022 02:07, Fangrui Song wrote:
>>> 795985e4e751 in 2003 added -fno-unit-at-a-time to errlist.c and
>>> siglist.c to "avoid reordering assembler output". -fno-toplevel-reorder
>>> is a rough replacement for this legacy option
>>> (https://sourceware.org/pipermail/gcc-patches/2006-January/186801.html).
>>>
>>> The reordering requirement does not seem to be needed any longer.
>>
>> We still need them for otherwise DEFINE_COMPAT_ERRLIST used on errlist-compat.c
>> does not create _sys_errlist and _sys_siglist with expected sizes defined by
>> glibc ABI.
>>
>> I am trying to fix without resorting to compiler options.
> 
> DEFINE_COMPAT_ERRLIST does not expand to code/data, just reordeable directives:
> 
> 
>        .globl  __GLIBC_2_1_sys_errlist
> .set __GLIBC_2_1_sys_errlist, _sys_errlist_internal
>         .type   __GLIBC_2_1_sys_errlist,@object
>         .size   __GLIBC_2_1_sys_errlist, 1000
>         .globl  __GLIBC_2_1__sys_errlist
> .set __GLIBC_2_1__sys_errlist, _sys_errlist_internal
>         .type   __GLIBC_2_1__sys_errlist,@object
>         .size   __GLIBC_2_1__sys_errlist, 1000
> .symver __GLIBC_2_1_sys_nerr, sys_nerr@GLIBC_2.2.5
> .symver __GLIBC_2_1__sys_nerr, _sys_nerr@GLIBC_2.2.5
> .symver __GLIBC_2_1_sys_errlist, sys_errlist@GLIBC_2.2.5
> .symver __GLIBC_2_1__sys_errlist, _sys_errlist@GLIBC_2.2.5
>         .globl  __GLIBC_2_3_sys_errlist
> .set __GLIBC_2_3_sys_errlist, _sys_errlist_internal
>         .type   __GLIBC_2_3_sys_errlist,@object
>         .size   __GLIBC_2_3_sys_errlist, 1008
>         .globl  __GLIBC_2_3__sys_errlist
> .set __GLIBC_2_3__sys_errlist, _sys_errlist_internal
>         .type   __GLIBC_2_3__sys_errlist,@object
>         .size   __GLIBC_2_3__sys_errlist, 1008
> .symver __GLIBC_2_3_sys_nerr, sys_nerr@GLIBC_2.3
> .symver __GLIBC_2_3__sys_nerr, _sys_nerr@GLIBC_2.3
> .symver __GLIBC_2_3_sys_errlist, sys_errlist@GLIBC_2.3
> .symver __GLIBC_2_3__sys_errlist, _sys_errlist@GLIBC_2.3
>         .globl  __GLIBC_2_4_sys_errlist
> 
> I do not know whether GCC would reorder these macros. Even yes,
> that'd just change the .symtab entries in the relocatable object file.
> The linker behavior remains the same with reordering.

It does not seem to, just remove the -fno-unit-at-a-time and issue make
check-abi and you will see that object size for the compat symbols 
reference to _sys_err_internal instead of the define compat ones.
  

On 2022-03-30, Adhemerval Zanella wrote:
>
>
>On 30/03/2022 13:23, Fangrui Song wrote:
>> On 2022-03-30, Adhemerval Zanella wrote:
>>>
>>>
>>> On 30/03/2022 02:07, Fangrui Song wrote:
>>>> 795985e4e751 in 2003 added -fno-unit-at-a-time to errlist.c and
>>>> siglist.c to "avoid reordering assembler output". -fno-toplevel-reorder
>>>> is a rough replacement for this legacy option
>>>> (https://sourceware.org/pipermail/gcc-patches/2006-January/186801.html).
>>>>
>>>> The reordering requirement does not seem to be needed any longer.
>>>
>>> We still need them for otherwise DEFINE_COMPAT_ERRLIST used on errlist-compat.c
>>> does not create _sys_errlist and _sys_siglist with expected sizes defined by
>>> glibc ABI.
>>>
>>> I am trying to fix without resorting to compiler options.
>>
>> DEFINE_COMPAT_ERRLIST does not expand to code/data, just reordeable directives:
>>
>>
>>        .globl  __GLIBC_2_1_sys_errlist
>> .set __GLIBC_2_1_sys_errlist, _sys_errlist_internal
>>         .type   __GLIBC_2_1_sys_errlist,@object
>>         .size   __GLIBC_2_1_sys_errlist, 1000
>>         .globl  __GLIBC_2_1__sys_errlist
>> .set __GLIBC_2_1__sys_errlist, _sys_errlist_internal
>>         .type   __GLIBC_2_1__sys_errlist,@object
>>         .size   __GLIBC_2_1__sys_errlist, 1000
>> .symver __GLIBC_2_1_sys_nerr, sys_nerr@GLIBC_2.2.5
>> .symver __GLIBC_2_1__sys_nerr, _sys_nerr@GLIBC_2.2.5
>> .symver __GLIBC_2_1_sys_errlist, sys_errlist@GLIBC_2.2.5
>> .symver __GLIBC_2_1__sys_errlist, _sys_errlist@GLIBC_2.2.5
>>         .globl  __GLIBC_2_3_sys_errlist
>> .set __GLIBC_2_3_sys_errlist, _sys_errlist_internal
>>         .type   __GLIBC_2_3_sys_errlist,@object
>>         .size   __GLIBC_2_3_sys_errlist, 1008
>>         .globl  __GLIBC_2_3__sys_errlist
>> .set __GLIBC_2_3__sys_errlist, _sys_errlist_internal
>>         .type   __GLIBC_2_3__sys_errlist,@object
>>         .size   __GLIBC_2_3__sys_errlist, 1008
>> .symver __GLIBC_2_3_sys_nerr, sys_nerr@GLIBC_2.3
>> .symver __GLIBC_2_3__sys_nerr, _sys_nerr@GLIBC_2.3
>> .symver __GLIBC_2_3_sys_errlist, sys_errlist@GLIBC_2.3
>> .symver __GLIBC_2_3__sys_errlist, _sys_errlist@GLIBC_2.3
>>         .globl  __GLIBC_2_4_sys_errlist
>>
>> I do not know whether GCC would reorder these macros. Even yes,
>> that'd just change the .symtab entries in the relocatable object file.
>> The linker behavior remains the same with reordering.
>
>It does not seem to, just remove the -fno-unit-at-a-time and issue make
>check-abi and you will see that object size for the compat symbols
>reference to _sys_err_internal instead of the define compat ones.

I see. I think this is a brittle behavior in GNU assembler. Filed
https://sourceware.org/bugzilla/show_bug.cgi?id=29012 with detailed
information. I have created a patch but I know that will not solve
glibc's problem :(
  

On 31/03/2022 00:43, Fangrui Song wrote:
> On 2022-03-30, Adhemerval Zanella wrote:
>>
>>
>> On 30/03/2022 13:23, Fangrui Song wrote:
>>> On 2022-03-30, Adhemerval Zanella wrote:
>>>>
>>>>
>>>> On 30/03/2022 02:07, Fangrui Song wrote:
>>>>> 795985e4e751 in 2003 added -fno-unit-at-a-time to errlist.c and
>>>>> siglist.c to "avoid reordering assembler output". -fno-toplevel-reorder
>>>>> is a rough replacement for this legacy option
>>>>> (https://sourceware.org/pipermail/gcc-patches/2006-January/186801.html).
>>>>>
>>>>> The reordering requirement does not seem to be needed any longer.
>>>>
>>>> We still need them for otherwise DEFINE_COMPAT_ERRLIST used on errlist-compat.c
>>>> does not create _sys_errlist and _sys_siglist with expected sizes defined by
>>>> glibc ABI.
>>>>
>>>> I am trying to fix without resorting to compiler options.
>>>
>>> DEFINE_COMPAT_ERRLIST does not expand to code/data, just reordeable directives:
>>>
>>>
>>>        .globl  __GLIBC_2_1_sys_errlist
>>> .set __GLIBC_2_1_sys_errlist, _sys_errlist_internal
>>>         .type   __GLIBC_2_1_sys_errlist,@object
>>>         .size   __GLIBC_2_1_sys_errlist, 1000
>>>         .globl  __GLIBC_2_1__sys_errlist
>>> .set __GLIBC_2_1__sys_errlist, _sys_errlist_internal
>>>         .type   __GLIBC_2_1__sys_errlist,@object
>>>         .size   __GLIBC_2_1__sys_errlist, 1000
>>> .symver __GLIBC_2_1_sys_nerr, sys_nerr@GLIBC_2.2.5
>>> .symver __GLIBC_2_1__sys_nerr, _sys_nerr@GLIBC_2.2.5
>>> .symver __GLIBC_2_1_sys_errlist, sys_errlist@GLIBC_2.2.5
>>> .symver __GLIBC_2_1__sys_errlist, _sys_errlist@GLIBC_2.2.5
>>>         .globl  __GLIBC_2_3_sys_errlist
>>> .set __GLIBC_2_3_sys_errlist, _sys_errlist_internal
>>>         .type   __GLIBC_2_3_sys_errlist,@object
>>>         .size   __GLIBC_2_3_sys_errlist, 1008
>>>         .globl  __GLIBC_2_3__sys_errlist
>>> .set __GLIBC_2_3__sys_errlist, _sys_errlist_internal
>>>         .type   __GLIBC_2_3__sys_errlist,@object
>>>         .size   __GLIBC_2_3__sys_errlist, 1008
>>> .symver __GLIBC_2_3_sys_nerr, sys_nerr@GLIBC_2.3
>>> .symver __GLIBC_2_3__sys_nerr, _sys_nerr@GLIBC_2.3
>>> .symver __GLIBC_2_3_sys_errlist, sys_errlist@GLIBC_2.3
>>> .symver __GLIBC_2_3__sys_errlist, _sys_errlist@GLIBC_2.3
>>>         .globl  __GLIBC_2_4_sys_errlist
>>>
>>> I do not know whether GCC would reorder these macros. Even yes,
>>> that'd just change the .symtab entries in the relocatable object file.
>>> The linker behavior remains the same with reordering.
>>
>> It does not seem to, just remove the -fno-unit-at-a-time and issue make
>> check-abi and you will see that object size for the compat symbols
>> reference to _sys_err_internal instead of the define compat ones.
> 
> I see. I think this is a brittle behavior in GNU assembler. Filed
> https://sourceware.org/bugzilla/show_bug.cgi?id=29012 with detailed
> information. I have created a patch but I know that will not solve
> glibc's problem :(

It would be good to have this fixes, but unfortunately we need a way
to handle this on older binutils.  I am kind worried that the only
possible way to actually fix this without resorting to any compiler 
flags is coding the array definitions in assembly direct...
  

On 2022-03-31, Adhemerval Zanella wrote:
>
>
>On 31/03/2022 00:43, Fangrui Song wrote:
>> On 2022-03-30, Adhemerval Zanella wrote:
>>>
>>>
>>> On 30/03/2022 13:23, Fangrui Song wrote:
>>>> On 2022-03-30, Adhemerval Zanella wrote:
>>>>>
>>>>>
>>>>> On 30/03/2022 02:07, Fangrui Song wrote:
>>>>>> 795985e4e751 in 2003 added -fno-unit-at-a-time to errlist.c and
>>>>>> siglist.c to "avoid reordering assembler output". -fno-toplevel-reorder
>>>>>> is a rough replacement for this legacy option
>>>>>> (https://sourceware.org/pipermail/gcc-patches/2006-January/186801.html).
>>>>>>
>>>>>> The reordering requirement does not seem to be needed any longer.
>>>>>
>>>>> We still need them for otherwise DEFINE_COMPAT_ERRLIST used on errlist-compat.c
>>>>> does not create _sys_errlist and _sys_siglist with expected sizes defined by
>>>>> glibc ABI.
>>>>>
>>>>> I am trying to fix without resorting to compiler options.
>>>>
>>>> DEFINE_COMPAT_ERRLIST does not expand to code/data, just reordeable directives:
>>>>
>>>>
>>>>        .globl  __GLIBC_2_1_sys_errlist
>>>> .set __GLIBC_2_1_sys_errlist, _sys_errlist_internal
>>>>         .type   __GLIBC_2_1_sys_errlist,@object
>>>>         .size   __GLIBC_2_1_sys_errlist, 1000
>>>>         .globl  __GLIBC_2_1__sys_errlist
>>>> .set __GLIBC_2_1__sys_errlist, _sys_errlist_internal
>>>>         .type   __GLIBC_2_1__sys_errlist,@object
>>>>         .size   __GLIBC_2_1__sys_errlist, 1000
>>>> .symver __GLIBC_2_1_sys_nerr, sys_nerr@GLIBC_2.2.5
>>>> .symver __GLIBC_2_1__sys_nerr, _sys_nerr@GLIBC_2.2.5
>>>> .symver __GLIBC_2_1_sys_errlist, sys_errlist@GLIBC_2.2.5
>>>> .symver __GLIBC_2_1__sys_errlist, _sys_errlist@GLIBC_2.2.5
>>>>         .globl  __GLIBC_2_3_sys_errlist
>>>> .set __GLIBC_2_3_sys_errlist, _sys_errlist_internal
>>>>         .type   __GLIBC_2_3_sys_errlist,@object
>>>>         .size   __GLIBC_2_3_sys_errlist, 1008
>>>>         .globl  __GLIBC_2_3__sys_errlist
>>>> .set __GLIBC_2_3__sys_errlist, _sys_errlist_internal
>>>>         .type   __GLIBC_2_3__sys_errlist,@object
>>>>         .size   __GLIBC_2_3__sys_errlist, 1008
>>>> .symver __GLIBC_2_3_sys_nerr, sys_nerr@GLIBC_2.3
>>>> .symver __GLIBC_2_3__sys_nerr, _sys_nerr@GLIBC_2.3
>>>> .symver __GLIBC_2_3_sys_errlist, sys_errlist@GLIBC_2.3
>>>> .symver __GLIBC_2_3__sys_errlist, _sys_errlist@GLIBC_2.3
>>>>         .globl  __GLIBC_2_4_sys_errlist
>>>>
>>>> I do not know whether GCC would reorder these macros. Even yes,
>>>> that'd just change the .symtab entries in the relocatable object file.
>>>> The linker behavior remains the same with reordering.
>>>
>>> It does not seem to, just remove the -fno-unit-at-a-time and issue make
>>> check-abi and you will see that object size for the compat symbols
>>> reference to _sys_err_internal instead of the define compat ones.
>>
>> I see. I think this is a brittle behavior in GNU assembler. Filed
>> https://sourceware.org/bugzilla/show_bug.cgi?id=29012 with detailed
>> information. I have created a patch but I know that will not solve
>> glibc's problem :(
>
>It would be good to have this fixes, but unfortunately we need a way
>to handle this on older binutils.  I am kind worried that the only
>possible way to actually fix this without resorting to any compiler
>flags is coding the array definitions in assembly direct...

The GNU assembler issue has been fixed https://sourceware.org/bugzilla/show_bug.cgi?id=29012
(milestone 2.39, way larger than the current required version: 2.25)

Switching to assembly output doesn't seem bad :-)

If you keep the compiler driver option but need to refactor the nearby
code, you may drop -fno-unit-at-a-time. It was added in 2006 (somewhere
between GCC 4.1 and 4.4), while glibc requires GCC>=6.2.
  

On 04/04/2022 12:57, Fangrui Song wrote:
> 
> On 2022-03-31, Adhemerval Zanella wrote:
>>
>>
>> On 31/03/2022 00:43, Fangrui Song wrote:
>>> On 2022-03-30, Adhemerval Zanella wrote:
>>>>
>>>>
>>>> On 30/03/2022 13:23, Fangrui Song wrote:
>>>>> On 2022-03-30, Adhemerval Zanella wrote:
>>>>>>
>>>>>>
>>>>>> On 30/03/2022 02:07, Fangrui Song wrote:
>>>>>>> 795985e4e751 in 2003 added -fno-unit-at-a-time to errlist.c and
>>>>>>> siglist.c to "avoid reordering assembler output". -fno-toplevel-reorder
>>>>>>> is a rough replacement for this legacy option
>>>>>>> (https://sourceware.org/pipermail/gcc-patches/2006-January/186801.html).
>>>>>>>
>>>>>>> The reordering requirement does not seem to be needed any longer.
>>>>>>
>>>>>> We still need them for otherwise DEFINE_COMPAT_ERRLIST used on errlist-compat.c
>>>>>> does not create _sys_errlist and _sys_siglist with expected sizes defined by
>>>>>> glibc ABI.
>>>>>>
>>>>>> I am trying to fix without resorting to compiler options.
>>>>>
>>>>> DEFINE_COMPAT_ERRLIST does not expand to code/data, just reordeable directives:
>>>>>
>>>>>
>>>>>        .globl  __GLIBC_2_1_sys_errlist
>>>>> .set __GLIBC_2_1_sys_errlist, _sys_errlist_internal
>>>>>         .type   __GLIBC_2_1_sys_errlist,@object
>>>>>         .size   __GLIBC_2_1_sys_errlist, 1000
>>>>>         .globl  __GLIBC_2_1__sys_errlist
>>>>> .set __GLIBC_2_1__sys_errlist, _sys_errlist_internal
>>>>>         .type   __GLIBC_2_1__sys_errlist,@object
>>>>>         .size   __GLIBC_2_1__sys_errlist, 1000
>>>>> .symver __GLIBC_2_1_sys_nerr, sys_nerr@GLIBC_2.2.5
>>>>> .symver __GLIBC_2_1__sys_nerr, _sys_nerr@GLIBC_2.2.5
>>>>> .symver __GLIBC_2_1_sys_errlist, sys_errlist@GLIBC_2.2.5
>>>>> .symver __GLIBC_2_1__sys_errlist, _sys_errlist@GLIBC_2.2.5
>>>>>         .globl  __GLIBC_2_3_sys_errlist
>>>>> .set __GLIBC_2_3_sys_errlist, _sys_errlist_internal
>>>>>         .type   __GLIBC_2_3_sys_errlist,@object
>>>>>         .size   __GLIBC_2_3_sys_errlist, 1008
>>>>>         .globl  __GLIBC_2_3__sys_errlist
>>>>> .set __GLIBC_2_3__sys_errlist, _sys_errlist_internal
>>>>>         .type   __GLIBC_2_3__sys_errlist,@object
>>>>>         .size   __GLIBC_2_3__sys_errlist, 1008
>>>>> .symver __GLIBC_2_3_sys_nerr, sys_nerr@GLIBC_2.3
>>>>> .symver __GLIBC_2_3__sys_nerr, _sys_nerr@GLIBC_2.3
>>>>> .symver __GLIBC_2_3_sys_errlist, sys_errlist@GLIBC_2.3
>>>>> .symver __GLIBC_2_3__sys_errlist, _sys_errlist@GLIBC_2.3
>>>>>         .globl  __GLIBC_2_4_sys_errlist
>>>>>
>>>>> I do not know whether GCC would reorder these macros. Even yes,
>>>>> that'd just change the .symtab entries in the relocatable object file.
>>>>> The linker behavior remains the same with reordering.
>>>>
>>>> It does not seem to, just remove the -fno-unit-at-a-time and issue make
>>>> check-abi and you will see that object size for the compat symbols
>>>> reference to _sys_err_internal instead of the define compat ones.
>>>
>>> I see. I think this is a brittle behavior in GNU assembler. Filed
>>> https://sourceware.org/bugzilla/show_bug.cgi?id=29012 with detailed
>>> information. I have created a patch but I know that will not solve
>>> glibc's problem :(
>>
>> It would be good to have this fixes, but unfortunately we need a way
>> to handle this on older binutils.  I am kind worried that the only
>> possible way to actually fix this without resorting to any compiler
>> flags is coding the array definitions in assembly direct...
> 
> The GNU assembler issue has been fixed https://sourceware.org/bugzilla/show_bug.cgi?id=29012
> (milestone 2.39, way larger than the current required version: 2.25)
> 
> Switching to assembly output doesn't seem bad :-)
> 
> If you keep the compiler driver option but need to refactor the nearby
> code, you may drop -fno-unit-at-a-time. It was added in 2006 (somewhere
> between GCC 4.1 and 4.4), while glibc requires GCC>=6.2.

Good to know we won't need to rely on compiler flags to get the expected
correct asm directives. I am still struggling to get a fix without resorting
to compiler flags, but without much success.  Trying to move it to assembly
might be tricky, I am not sure if the data directives would be architecture
agnostic.
  

On Tue, Apr 5, 2022 at 8:35 AM Adhemerval Zanella
<adhemerval.zanella@linaro.org> wrote:
>
>
>
> On 04/04/2022 12:57, Fangrui Song wrote:
> >
> > On 2022-03-31, Adhemerval Zanella wrote:
> >>
> >>
> >> On 31/03/2022 00:43, Fangrui Song wrote:
> >>> On 2022-03-30, Adhemerval Zanella wrote:
> >>>>
> >>>>
> >>>> On 30/03/2022 13:23, Fangrui Song wrote:
> >>>>> On 2022-03-30, Adhemerval Zanella wrote:
> >>>>>>
> >>>>>>
> >>>>>> On 30/03/2022 02:07, Fangrui Song wrote:
> >>>>>>> 795985e4e751 in 2003 added -fno-unit-at-a-time to errlist.c and
> >>>>>>> siglist.c to "avoid reordering assembler output". -fno-toplevel-reorder
> >>>>>>> is a rough replacement for this legacy option
> >>>>>>> (https://sourceware.org/pipermail/gcc-patches/2006-January/186801.html).
> >>>>>>>
> >>>>>>> The reordering requirement does not seem to be needed any longer.
> >>>>>>
> >>>>>> We still need them for otherwise DEFINE_COMPAT_ERRLIST used on errlist-compat.c
> >>>>>> does not create _sys_errlist and _sys_siglist with expected sizes defined by
> >>>>>> glibc ABI.
> >>>>>>
> >>>>>> I am trying to fix without resorting to compiler options.
> >>>>>
> >>>>> DEFINE_COMPAT_ERRLIST does not expand to code/data, just reordeable directives:
> >>>>>
> >>>>>
> >>>>>        .globl  __GLIBC_2_1_sys_errlist
> >>>>> .set __GLIBC_2_1_sys_errlist, _sys_errlist_internal
> >>>>>         .type   __GLIBC_2_1_sys_errlist,@object
> >>>>>         .size   __GLIBC_2_1_sys_errlist, 1000
> >>>>>         .globl  __GLIBC_2_1__sys_errlist
> >>>>> .set __GLIBC_2_1__sys_errlist, _sys_errlist_internal
> >>>>>         .type   __GLIBC_2_1__sys_errlist,@object
> >>>>>         .size   __GLIBC_2_1__sys_errlist, 1000
> >>>>> .symver __GLIBC_2_1_sys_nerr, sys_nerr@GLIBC_2.2.5
> >>>>> .symver __GLIBC_2_1__sys_nerr, _sys_nerr@GLIBC_2.2.5
> >>>>> .symver __GLIBC_2_1_sys_errlist, sys_errlist@GLIBC_2.2.5
> >>>>> .symver __GLIBC_2_1__sys_errlist, _sys_errlist@GLIBC_2.2.5
> >>>>>         .globl  __GLIBC_2_3_sys_errlist
> >>>>> .set __GLIBC_2_3_sys_errlist, _sys_errlist_internal
> >>>>>         .type   __GLIBC_2_3_sys_errlist,@object
> >>>>>         .size   __GLIBC_2_3_sys_errlist, 1008
> >>>>>         .globl  __GLIBC_2_3__sys_errlist
> >>>>> .set __GLIBC_2_3__sys_errlist, _sys_errlist_internal
> >>>>>         .type   __GLIBC_2_3__sys_errlist,@object
> >>>>>         .size   __GLIBC_2_3__sys_errlist, 1008
> >>>>> .symver __GLIBC_2_3_sys_nerr, sys_nerr@GLIBC_2.3
> >>>>> .symver __GLIBC_2_3__sys_nerr, _sys_nerr@GLIBC_2.3
> >>>>> .symver __GLIBC_2_3_sys_errlist, sys_errlist@GLIBC_2.3
> >>>>> .symver __GLIBC_2_3__sys_errlist, _sys_errlist@GLIBC_2.3
> >>>>>         .globl  __GLIBC_2_4_sys_errlist
> >>>>>
> >>>>> I do not know whether GCC would reorder these macros. Even yes,
> >>>>> that'd just change the .symtab entries in the relocatable object file.
> >>>>> The linker behavior remains the same with reordering.
> >>>>
> >>>> It does not seem to, just remove the -fno-unit-at-a-time and issue make
> >>>> check-abi and you will see that object size for the compat symbols
> >>>> reference to _sys_err_internal instead of the define compat ones.
> >>>
> >>> I see. I think this is a brittle behavior in GNU assembler. Filed
> >>> https://sourceware.org/bugzilla/show_bug.cgi?id=29012 with detailed
> >>> information. I have created a patch but I know that will not solve
> >>> glibc's problem :(
> >>
> >> It would be good to have this fixes, but unfortunately we need a way
> >> to handle this on older binutils.  I am kind worried that the only
> >> possible way to actually fix this without resorting to any compiler
> >> flags is coding the array definitions in assembly direct...
> >
> > The GNU assembler issue has been fixed https://sourceware.org/bugzilla/show_bug.cgi?id=29012
> > (milestone 2.39, way larger than the current required version: 2.25)
> >
> > Switching to assembly output doesn't seem bad :-)
> >
> > If you keep the compiler driver option but need to refactor the nearby
> > code, you may drop -fno-unit-at-a-time. It was added in 2006 (somewhere
> > between GCC 4.1 and 4.4), while glibc requires GCC>=6.2.
>
> Good to know we won't need to rely on compiler flags to get the expected
> correct asm directives. I am still struggling to get a fix without resorting
> to compiler flags, but without much success.  Trying to move it to assembly
> might be tricky, I am not sure if the data directives would be architecture
> agnostic.

Many directives are architecture-independent:
https://sourceware.org/binutils/docs/as/Pseudo-Ops.html#Pseudo-Ops
binutils-gdb/gas/read.c:346 The `portable[]` array.

To support Clang, no refactoring is probably needed: just change
fno_unit_at_a_time to only specify -fno-toplevel-reorder (and rename
it), not th legacy -fno-unit-at-a-time.
  

On 05/04/2022 13:22, Fāng-ruì Sòng wrote:
> On Tue, Apr 5, 2022 at 8:35 AM Adhemerval Zanella
> <adhemerval.zanella@linaro.org> wrote:
>>
>>
>>
>> On 04/04/2022 12:57, Fangrui Song wrote:
>>>
>>> On 2022-03-31, Adhemerval Zanella wrote:
>>>>
>>>>
>>>> On 31/03/2022 00:43, Fangrui Song wrote:
>>>>> On 2022-03-30, Adhemerval Zanella wrote:
>>>>>>
>>>>>>
>>>>>> On 30/03/2022 13:23, Fangrui Song wrote:
>>>>>>> On 2022-03-30, Adhemerval Zanella wrote:
>>>>>>>>
>>>>>>>>
>>>>>>>> On 30/03/2022 02:07, Fangrui Song wrote:
>>>>>>>>> 795985e4e751 in 2003 added -fno-unit-at-a-time to errlist.c and
>>>>>>>>> siglist.c to "avoid reordering assembler output". -fno-toplevel-reorder
>>>>>>>>> is a rough replacement for this legacy option
>>>>>>>>> (https://sourceware.org/pipermail/gcc-patches/2006-January/186801.html).
>>>>>>>>>
>>>>>>>>> The reordering requirement does not seem to be needed any longer.
>>>>>>>>
>>>>>>>> We still need them for otherwise DEFINE_COMPAT_ERRLIST used on errlist-compat.c
>>>>>>>> does not create _sys_errlist and _sys_siglist with expected sizes defined by
>>>>>>>> glibc ABI.
>>>>>>>>
>>>>>>>> I am trying to fix without resorting to compiler options.
>>>>>>>
>>>>>>> DEFINE_COMPAT_ERRLIST does not expand to code/data, just reordeable directives:
>>>>>>>
>>>>>>>
>>>>>>>        .globl  __GLIBC_2_1_sys_errlist
>>>>>>> .set __GLIBC_2_1_sys_errlist, _sys_errlist_internal
>>>>>>>         .type   __GLIBC_2_1_sys_errlist,@object
>>>>>>>         .size   __GLIBC_2_1_sys_errlist, 1000
>>>>>>>         .globl  __GLIBC_2_1__sys_errlist
>>>>>>> .set __GLIBC_2_1__sys_errlist, _sys_errlist_internal
>>>>>>>         .type   __GLIBC_2_1__sys_errlist,@object
>>>>>>>         .size   __GLIBC_2_1__sys_errlist, 1000
>>>>>>> .symver __GLIBC_2_1_sys_nerr, sys_nerr@GLIBC_2.2.5
>>>>>>> .symver __GLIBC_2_1__sys_nerr, _sys_nerr@GLIBC_2.2.5
>>>>>>> .symver __GLIBC_2_1_sys_errlist, sys_errlist@GLIBC_2.2.5
>>>>>>> .symver __GLIBC_2_1__sys_errlist, _sys_errlist@GLIBC_2.2.5
>>>>>>>         .globl  __GLIBC_2_3_sys_errlist
>>>>>>> .set __GLIBC_2_3_sys_errlist, _sys_errlist_internal
>>>>>>>         .type   __GLIBC_2_3_sys_errlist,@object
>>>>>>>         .size   __GLIBC_2_3_sys_errlist, 1008
>>>>>>>         .globl  __GLIBC_2_3__sys_errlist
>>>>>>> .set __GLIBC_2_3__sys_errlist, _sys_errlist_internal
>>>>>>>         .type   __GLIBC_2_3__sys_errlist,@object
>>>>>>>         .size   __GLIBC_2_3__sys_errlist, 1008
>>>>>>> .symver __GLIBC_2_3_sys_nerr, sys_nerr@GLIBC_2.3
>>>>>>> .symver __GLIBC_2_3__sys_nerr, _sys_nerr@GLIBC_2.3
>>>>>>> .symver __GLIBC_2_3_sys_errlist, sys_errlist@GLIBC_2.3
>>>>>>> .symver __GLIBC_2_3__sys_errlist, _sys_errlist@GLIBC_2.3
>>>>>>>         .globl  __GLIBC_2_4_sys_errlist
>>>>>>>
>>>>>>> I do not know whether GCC would reorder these macros. Even yes,
>>>>>>> that'd just change the .symtab entries in the relocatable object file.
>>>>>>> The linker behavior remains the same with reordering.
>>>>>>
>>>>>> It does not seem to, just remove the -fno-unit-at-a-time and issue make
>>>>>> check-abi and you will see that object size for the compat symbols
>>>>>> reference to _sys_err_internal instead of the define compat ones.
>>>>>
>>>>> I see. I think this is a brittle behavior in GNU assembler. Filed
>>>>> https://sourceware.org/bugzilla/show_bug.cgi?id=29012 with detailed
>>>>> information. I have created a patch but I know that will not solve
>>>>> glibc's problem :(
>>>>
>>>> It would be good to have this fixes, but unfortunately we need a way
>>>> to handle this on older binutils.  I am kind worried that the only
>>>> possible way to actually fix this without resorting to any compiler
>>>> flags is coding the array definitions in assembly direct...
>>>
>>> The GNU assembler issue has been fixed https://sourceware.org/bugzilla/show_bug.cgi?id=29012
>>> (milestone 2.39, way larger than the current required version: 2.25)
>>>
>>> Switching to assembly output doesn't seem bad :-)
>>>
>>> If you keep the compiler driver option but need to refactor the nearby
>>> code, you may drop -fno-unit-at-a-time. It was added in 2006 (somewhere
>>> between GCC 4.1 and 4.4), while glibc requires GCC>=6.2.
>>
>> Good to know we won't need to rely on compiler flags to get the expected
>> correct asm directives. I am still struggling to get a fix without resorting
>> to compiler flags, but without much success.  Trying to move it to assembly
>> might be tricky, I am not sure if the data directives would be architecture
>> agnostic.
> 
> Many directives are architecture-independent:
> https://sourceware.org/binutils/docs/as/Pseudo-Ops.html#Pseudo-Ops
> binutils-gdb/gas/read.c:346 The `portable[]` array.

I think we can make it work with asciz directive.

> 
> To support Clang, no refactoring is probably needed: just change
> fno_unit_at_a_time to only specify -fno-toplevel-reorder (and rename
> it), not th legacy -fno-unit-at-a-time.

Afaik llvm does not support -fno-toplevel-reorder
  

On Tue, Apr 5, 2022 at 10:40 AM Adhemerval Zanella
<adhemerval.zanella@linaro.org> wrote:
>
>
>
> On 05/04/2022 13:22, Fāng-ruì Sòng wrote:
> > On Tue, Apr 5, 2022 at 8:35 AM Adhemerval Zanella
> > <adhemerval.zanella@linaro.org> wrote:
> >>
> >>
> >>
> >> On 04/04/2022 12:57, Fangrui Song wrote:
> >>>
> >>> On 2022-03-31, Adhemerval Zanella wrote:
> >>>>
> >>>>
> >>>> On 31/03/2022 00:43, Fangrui Song wrote:
> >>>>> On 2022-03-30, Adhemerval Zanella wrote:
> >>>>>>
> >>>>>>
> >>>>>> On 30/03/2022 13:23, Fangrui Song wrote:
> >>>>>>> On 2022-03-30, Adhemerval Zanella wrote:
> >>>>>>>>
> >>>>>>>>
> >>>>>>>> On 30/03/2022 02:07, Fangrui Song wrote:
> >>>>>>>>> 795985e4e751 in 2003 added -fno-unit-at-a-time to errlist.c and
> >>>>>>>>> siglist.c to "avoid reordering assembler output". -fno-toplevel-reorder
> >>>>>>>>> is a rough replacement for this legacy option
> >>>>>>>>> (https://sourceware.org/pipermail/gcc-patches/2006-January/186801.html).
> >>>>>>>>>
> >>>>>>>>> The reordering requirement does not seem to be needed any longer.
> >>>>>>>>
> >>>>>>>> We still need them for otherwise DEFINE_COMPAT_ERRLIST used on errlist-compat.c
> >>>>>>>> does not create _sys_errlist and _sys_siglist with expected sizes defined by
> >>>>>>>> glibc ABI.
> >>>>>>>>
> >>>>>>>> I am trying to fix without resorting to compiler options.
> >>>>>>>
> >>>>>>> DEFINE_COMPAT_ERRLIST does not expand to code/data, just reordeable directives:
> >>>>>>>
> >>>>>>>
> >>>>>>>        .globl  __GLIBC_2_1_sys_errlist
> >>>>>>> .set __GLIBC_2_1_sys_errlist, _sys_errlist_internal
> >>>>>>>         .type   __GLIBC_2_1_sys_errlist,@object
> >>>>>>>         .size   __GLIBC_2_1_sys_errlist, 1000
> >>>>>>>         .globl  __GLIBC_2_1__sys_errlist
> >>>>>>> .set __GLIBC_2_1__sys_errlist, _sys_errlist_internal
> >>>>>>>         .type   __GLIBC_2_1__sys_errlist,@object
> >>>>>>>         .size   __GLIBC_2_1__sys_errlist, 1000
> >>>>>>> .symver __GLIBC_2_1_sys_nerr, sys_nerr@GLIBC_2.2.5
> >>>>>>> .symver __GLIBC_2_1__sys_nerr, _sys_nerr@GLIBC_2.2.5
> >>>>>>> .symver __GLIBC_2_1_sys_errlist, sys_errlist@GLIBC_2.2.5
> >>>>>>> .symver __GLIBC_2_1__sys_errlist, _sys_errlist@GLIBC_2.2.5
> >>>>>>>         .globl  __GLIBC_2_3_sys_errlist
> >>>>>>> .set __GLIBC_2_3_sys_errlist, _sys_errlist_internal
> >>>>>>>         .type   __GLIBC_2_3_sys_errlist,@object
> >>>>>>>         .size   __GLIBC_2_3_sys_errlist, 1008
> >>>>>>>         .globl  __GLIBC_2_3__sys_errlist
> >>>>>>> .set __GLIBC_2_3__sys_errlist, _sys_errlist_internal
> >>>>>>>         .type   __GLIBC_2_3__sys_errlist,@object
> >>>>>>>         .size   __GLIBC_2_3__sys_errlist, 1008
> >>>>>>> .symver __GLIBC_2_3_sys_nerr, sys_nerr@GLIBC_2.3
> >>>>>>> .symver __GLIBC_2_3__sys_nerr, _sys_nerr@GLIBC_2.3
> >>>>>>> .symver __GLIBC_2_3_sys_errlist, sys_errlist@GLIBC_2.3
> >>>>>>> .symver __GLIBC_2_3__sys_errlist, _sys_errlist@GLIBC_2.3
> >>>>>>>         .globl  __GLIBC_2_4_sys_errlist
> >>>>>>>
> >>>>>>> I do not know whether GCC would reorder these macros. Even yes,
> >>>>>>> that'd just change the .symtab entries in the relocatable object file.
> >>>>>>> The linker behavior remains the same with reordering.
> >>>>>>
> >>>>>> It does not seem to, just remove the -fno-unit-at-a-time and issue make
> >>>>>> check-abi and you will see that object size for the compat symbols
> >>>>>> reference to _sys_err_internal instead of the define compat ones.
> >>>>>
> >>>>> I see. I think this is a brittle behavior in GNU assembler. Filed
> >>>>> https://sourceware.org/bugzilla/show_bug.cgi?id=29012 with detailed
> >>>>> information. I have created a patch but I know that will not solve
> >>>>> glibc's problem :(
> >>>>
> >>>> It would be good to have this fixes, but unfortunately we need a way
> >>>> to handle this on older binutils.  I am kind worried that the only
> >>>> possible way to actually fix this without resorting to any compiler
> >>>> flags is coding the array definitions in assembly direct...
> >>>
> >>> The GNU assembler issue has been fixed https://sourceware.org/bugzilla/show_bug.cgi?id=29012
> >>> (milestone 2.39, way larger than the current required version: 2.25)
> >>>
> >>> Switching to assembly output doesn't seem bad :-)
> >>>
> >>> If you keep the compiler driver option but need to refactor the nearby
> >>> code, you may drop -fno-unit-at-a-time. It was added in 2006 (somewhere
> >>> between GCC 4.1 and 4.4), while glibc requires GCC>=6.2.
> >>
> >> Good to know we won't need to rely on compiler flags to get the expected
> >> correct asm directives. I am still struggling to get a fix without resorting
> >> to compiler flags, but without much success.  Trying to move it to assembly
> >> might be tricky, I am not sure if the data directives would be architecture
> >> agnostic.
> >
> > Many directives are architecture-independent:
> > https://sourceware.org/binutils/docs/as/Pseudo-Ops.html#Pseudo-Ops
> > binutils-gdb/gas/read.c:346 The `portable[]` array.
>
> I think we can make it work with asciz directive.
>
> >
> > To support Clang, no refactoring is probably needed: just change
> > fno_unit_at_a_time to only specify -fno-toplevel-reorder (and rename
> > it), not th legacy -fno-unit-at-a-time.
>
> Afaik llvm does not support -fno-toplevel-reorder

It doesn't, but its integrated assembler does not need the option to
have the desired semantics. I have a note that the new GNU as behavior
is quite similar to LLVM's integrated assembler since 2014-03:
https://sourceware.org/bugzilla/show_bug.cgi?id=29012#c1
  

On 05/04/2022 13:44, Fāng-ruì Sòng wrote:
> On Tue, Apr 5, 2022 at 10:40 AM Adhemerval Zanella
> <adhemerval.zanella@linaro.org> wrote:
>>
>>
>>
>> On 05/04/2022 13:22, Fāng-ruì Sòng wrote:
>>> On Tue, Apr 5, 2022 at 8:35 AM Adhemerval Zanella
>>> <adhemerval.zanella@linaro.org> wrote:
>>>>
>>>>
>>>>
>>>> On 04/04/2022 12:57, Fangrui Song wrote:
>>>>>
>>>>> On 2022-03-31, Adhemerval Zanella wrote:
>>>>>>
>>>>>>
>>>>>> On 31/03/2022 00:43, Fangrui Song wrote:
>>>>>>> On 2022-03-30, Adhemerval Zanella wrote:
>>>>>>>>
>>>>>>>>
>>>>>>>> On 30/03/2022 13:23, Fangrui Song wrote:
>>>>>>>>> On 2022-03-30, Adhemerval Zanella wrote:
>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>> On 30/03/2022 02:07, Fangrui Song wrote:
>>>>>>>>>>> 795985e4e751 in 2003 added -fno-unit-at-a-time to errlist.c and
>>>>>>>>>>> siglist.c to "avoid reordering assembler output". -fno-toplevel-reorder
>>>>>>>>>>> is a rough replacement for this legacy option
>>>>>>>>>>> (https://sourceware.org/pipermail/gcc-patches/2006-January/186801.html).
>>>>>>>>>>>
>>>>>>>>>>> The reordering requirement does not seem to be needed any longer.
>>>>>>>>>>
>>>>>>>>>> We still need them for otherwise DEFINE_COMPAT_ERRLIST used on errlist-compat.c
>>>>>>>>>> does not create _sys_errlist and _sys_siglist with expected sizes defined by
>>>>>>>>>> glibc ABI.
>>>>>>>>>>
>>>>>>>>>> I am trying to fix without resorting to compiler options.
>>>>>>>>>
>>>>>>>>> DEFINE_COMPAT_ERRLIST does not expand to code/data, just reordeable directives:
>>>>>>>>>
>>>>>>>>>
>>>>>>>>>        .globl  __GLIBC_2_1_sys_errlist
>>>>>>>>> .set __GLIBC_2_1_sys_errlist, _sys_errlist_internal
>>>>>>>>>         .type   __GLIBC_2_1_sys_errlist,@object
>>>>>>>>>         .size   __GLIBC_2_1_sys_errlist, 1000
>>>>>>>>>         .globl  __GLIBC_2_1__sys_errlist
>>>>>>>>> .set __GLIBC_2_1__sys_errlist, _sys_errlist_internal
>>>>>>>>>         .type   __GLIBC_2_1__sys_errlist,@object
>>>>>>>>>         .size   __GLIBC_2_1__sys_errlist, 1000
>>>>>>>>> .symver __GLIBC_2_1_sys_nerr, sys_nerr@GLIBC_2.2.5
>>>>>>>>> .symver __GLIBC_2_1__sys_nerr, _sys_nerr@GLIBC_2.2.5
>>>>>>>>> .symver __GLIBC_2_1_sys_errlist, sys_errlist@GLIBC_2.2.5
>>>>>>>>> .symver __GLIBC_2_1__sys_errlist, _sys_errlist@GLIBC_2.2.5
>>>>>>>>>         .globl  __GLIBC_2_3_sys_errlist
>>>>>>>>> .set __GLIBC_2_3_sys_errlist, _sys_errlist_internal
>>>>>>>>>         .type   __GLIBC_2_3_sys_errlist,@object
>>>>>>>>>         .size   __GLIBC_2_3_sys_errlist, 1008
>>>>>>>>>         .globl  __GLIBC_2_3__sys_errlist
>>>>>>>>> .set __GLIBC_2_3__sys_errlist, _sys_errlist_internal
>>>>>>>>>         .type   __GLIBC_2_3__sys_errlist,@object
>>>>>>>>>         .size   __GLIBC_2_3__sys_errlist, 1008
>>>>>>>>> .symver __GLIBC_2_3_sys_nerr, sys_nerr@GLIBC_2.3
>>>>>>>>> .symver __GLIBC_2_3__sys_nerr, _sys_nerr@GLIBC_2.3
>>>>>>>>> .symver __GLIBC_2_3_sys_errlist, sys_errlist@GLIBC_2.3
>>>>>>>>> .symver __GLIBC_2_3__sys_errlist, _sys_errlist@GLIBC_2.3
>>>>>>>>>         .globl  __GLIBC_2_4_sys_errlist
>>>>>>>>>
>>>>>>>>> I do not know whether GCC would reorder these macros. Even yes,
>>>>>>>>> that'd just change the .symtab entries in the relocatable object file.
>>>>>>>>> The linker behavior remains the same with reordering.
>>>>>>>>
>>>>>>>> It does not seem to, just remove the -fno-unit-at-a-time and issue make
>>>>>>>> check-abi and you will see that object size for the compat symbols
>>>>>>>> reference to _sys_err_internal instead of the define compat ones.
>>>>>>>
>>>>>>> I see. I think this is a brittle behavior in GNU assembler. Filed
>>>>>>> https://sourceware.org/bugzilla/show_bug.cgi?id=29012 with detailed
>>>>>>> information. I have created a patch but I know that will not solve
>>>>>>> glibc's problem :(
>>>>>>
>>>>>> It would be good to have this fixes, but unfortunately we need a way
>>>>>> to handle this on older binutils.  I am kind worried that the only
>>>>>> possible way to actually fix this without resorting to any compiler
>>>>>> flags is coding the array definitions in assembly direct...
>>>>>
>>>>> The GNU assembler issue has been fixed https://sourceware.org/bugzilla/show_bug.cgi?id=29012
>>>>> (milestone 2.39, way larger than the current required version: 2.25)
>>>>>
>>>>> Switching to assembly output doesn't seem bad :-)
>>>>>
>>>>> If you keep the compiler driver option but need to refactor the nearby
>>>>> code, you may drop -fno-unit-at-a-time. It was added in 2006 (somewhere
>>>>> between GCC 4.1 and 4.4), while glibc requires GCC>=6.2.
>>>>
>>>> Good to know we won't need to rely on compiler flags to get the expected
>>>> correct asm directives. I am still struggling to get a fix without resorting
>>>> to compiler flags, but without much success.  Trying to move it to assembly
>>>> might be tricky, I am not sure if the data directives would be architecture
>>>> agnostic.
>>>
>>> Many directives are architecture-independent:
>>> https://sourceware.org/binutils/docs/as/Pseudo-Ops.html#Pseudo-Ops
>>> binutils-gdb/gas/read.c:346 The `portable[]` array.
>>
>> I think we can make it work with asciz directive.
>>
>>>
>>> To support Clang, no refactoring is probably needed: just change
>>> fno_unit_at_a_time to only specify -fno-toplevel-reorder (and rename
>>> it), not th legacy -fno-unit-at-a-time.
>>
>> Afaik llvm does not support -fno-toplevel-reorder
> 
> It doesn't, but its integrated assembler does not need the option to
> have the desired semantics. I have a note that the new GNU as behavior
> is quite similar to LLVM's integrated assembler since 2014-03:
> https://sourceware.org/bugzilla/show_bug.cgi?id=29012#c1

This is not what I am seeing on my clang branch, where clang with integrated
moves the global asm directives to the top of the file.  And it makes 
check-abi fail due the _sys_errlist/sys_errlist being with wrong value.
  

* Adhemerval Zanella via Libc-alpha:

> This is not what I am seeing on my clang branch, where clang with integrated
> moves the global asm directives to the top of the file.  And it makes 
> check-abi fail due the _sys_errlist/sys_errlist being with wrong value.

You can move the directives to the bottom of the file by creating a
temporary .s file and #include that from a wrapper .S file that
#includes the .s file.

Or perhaps generate the assembler file directly, as in
elf/dl-debug-symbols.S, which solves the same problem.  Linux <errno.h>
should be includable directly from assembler, hopefully that's true for
the Hurd version as well.

Thanks,
Florian
  

On 05/04/2022 14:53, Florian Weimer wrote:
> * Adhemerval Zanella via Libc-alpha:
> 
>> This is not what I am seeing on my clang branch, where clang with integrated
>> moves the global asm directives to the top of the file.  And it makes 
>> check-abi fail due the _sys_errlist/sys_errlist being with wrong value.
> 
> You can move the directives to the bottom of the file by creating a
> temporary .s file and #include that from a wrapper .S file that
> #includes the .s file.

Not sure I am following your suggestion here, do you mean get the -S
output file from compiler?

> 
> Or perhaps generate the assembler file directly, as in
> elf/dl-debug-symbols.S, which solves the same problem.  Linux <errno.h>
> should be includable directly from assembler, hopefully that's true for
> the Hurd version as well.

Making an assembly file is really tricky here because _sys_errlist_internal
contains internal gap (the errno values are not strictly sequential). The
compiler handles by adding '.zero' commands, which is not easily.  I think
we might generate it though a script.
  

* Adhemerval Zanella:

> On 05/04/2022 14:53, Florian Weimer wrote:
>> * Adhemerval Zanella via Libc-alpha:
>> 
>>> This is not what I am seeing on my clang branch, where clang with integrated
>>> moves the global asm directives to the top of the file.  And it makes 
>>> check-abi fail due the _sys_errlist/sys_errlist being with wrong value.
>> 
>> You can move the directives to the bottom of the file by creating a
>> temporary .s file and #include that from a wrapper .S file that
>> #includes the .s file.
>
> Not sure I am following your suggestion here, do you mean get the -S
> output file from compiler?

Yes, H.J. did that in earlier patches.  But I didn't like it back then. 8-p

> Making an assembly file is really tricky here because _sys_errlist_internal
> contains internal gap (the errno values are not strictly sequential). The
> compiler handles by adding '.zero' commands, which is not easily.  I think
> we might generate it though a script.

I think the input file is already generated by a script.

Thanks,
Florian
  

On 05/04/2022 17:26, Florian Weimer wrote:
> * Adhemerval Zanella:
> 
>> On 05/04/2022 14:53, Florian Weimer wrote:
>>> * Adhemerval Zanella via Libc-alpha:
>>>
>>>> This is not what I am seeing on my clang branch, where clang with integrated
>>>> moves the global asm directives to the top of the file.  And it makes 
>>>> check-abi fail due the _sys_errlist/sys_errlist being with wrong value.
>>>
>>> You can move the directives to the bottom of the file by creating a
>>> temporary .s file and #include that from a wrapper .S file that
>>> #includes the .s file.
>>
>> Not sure I am following your suggestion here, do you mean get the -S
>> output file from compiler?
> 
> Yes, H.J. did that in earlier patches.  But I didn't like it back then. 8-p

For dl-debug-symbols.S? I might take it as an example.

> 
>> Making an assembly file is really tricky here because _sys_errlist_internal
>> contains internal gap (the errno values are not strictly sequential). The
>> compiler handles by adding '.zero' commands, which is not easily.  I think
>> we might generate it though a script.
> 
> I think the input file is already generated by a script.

No, what is scripted is the tst-errno-manual (which checks the errlist.h
against manual/errno.texi).
  

* Adhemerval Zanella:

> On 05/04/2022 17:26, Florian Weimer wrote:
>> * Adhemerval Zanella:
>> 
>>> On 05/04/2022 14:53, Florian Weimer wrote:
>>>> * Adhemerval Zanella via Libc-alpha:
>>>>
>>>>> This is not what I am seeing on my clang branch, where clang with integrated
>>>>> moves the global asm directives to the top of the file.  And it makes 
>>>>> check-abi fail due the _sys_errlist/sys_errlist being with wrong value.
>>>>
>>>> You can move the directives to the bottom of the file by creating a
>>>> temporary .s file and #include that from a wrapper .S file that
>>>> #includes the .s file.
>>>
>>> Not sure I am following your suggestion here, do you mean get the -S
>>> output file from compiler?
>> 
>> Yes, H.J. did that in earlier patches.  But I didn't like it back then. 8-p
>
> For dl-debug-symbols.S? I might take it as an example.

Yes, here's a link to one of the earlier patches:

  [PATCH] Extend struct r_debug to support multiple namespaces
  <https://sourceware.org/pipermail/libc-alpha/2021-August/130149.html>

In the end, I think it was easier to use a manually written assembler
file in that case, with the constants filled in via the .sym expander.

Thanks,
Florian
  

On 05/04/2022 17:40, Florian Weimer wrote:
> * Adhemerval Zanella:
> 
>> On 05/04/2022 17:26, Florian Weimer wrote:
>>> * Adhemerval Zanella:
>>>
>>>> On 05/04/2022 14:53, Florian Weimer wrote:
>>>>> * Adhemerval Zanella via Libc-alpha:
>>>>>
>>>>>> This is not what I am seeing on my clang branch, where clang with integrated
>>>>>> moves the global asm directives to the top of the file.  And it makes 
>>>>>> check-abi fail due the _sys_errlist/sys_errlist being with wrong value.
>>>>>
>>>>> You can move the directives to the bottom of the file by creating a
>>>>> temporary .s file and #include that from a wrapper .S file that
>>>>> #includes the .s file.
>>>>
>>>> Not sure I am following your suggestion here, do you mean get the -S
>>>> output file from compiler?
>>>
>>> Yes, H.J. did that in earlier patches.  But I didn't like it back then. 8-p
>>
>> For dl-debug-symbols.S? I might take it as an example.
> 
> Yes, here's a link to one of the earlier patches:
> 
>   [PATCH] Extend struct r_debug to support multiple namespaces
>   <https://sourceware.org/pipermail/libc-alpha/2021-August/130149.html>

Thanks, I found it on the my local inbox history.

> 
> In the end, I think it was easier to use a manually written assembler
> file in that case, with the constants filled in via the .sym expander.

I think for this case, it would be simpler to use H.J strategy, I really
don't want to re-implement what compiler might do to lay out a C array
to assembler.
  

On 05/04/2022 18:15, Adhemerval Zanella wrote:
> 
> 
> On 05/04/2022 17:40, Florian Weimer wrote:
>> * Adhemerval Zanella:
>>
>>> On 05/04/2022 17:26, Florian Weimer wrote:
>>>> * Adhemerval Zanella:
>>>>
>>>>> On 05/04/2022 14:53, Florian Weimer wrote:
>>>>>> * Adhemerval Zanella via Libc-alpha:
>>>>>>
>>>>>>> This is not what I am seeing on my clang branch, where clang with integrated
>>>>>>> moves the global asm directives to the top of the file.  And it makes 
>>>>>>> check-abi fail due the _sys_errlist/sys_errlist being with wrong value.
>>>>>>
>>>>>> You can move the directives to the bottom of the file by creating a
>>>>>> temporary .s file and #include that from a wrapper .S file that
>>>>>> #includes the .s file.
>>>>>
>>>>> Not sure I am following your suggestion here, do you mean get the -S
>>>>> output file from compiler?
>>>>
>>>> Yes, H.J. did that in earlier patches.  But I didn't like it back then. 8-p
>>>
>>> For dl-debug-symbols.S? I might take it as an example.
>>
>> Yes, here's a link to one of the earlier patches:
>>
>>   [PATCH] Extend struct r_debug to support multiple namespaces
>>   <https://sourceware.org/pipermail/libc-alpha/2021-August/130149.html>
> 
> Thanks, I found it on the my local inbox history.
> 
>>
>> In the end, I think it was easier to use a manually written assembler
>> file in that case, with the constants filled in via the .sym expander.
> 
> I think for this case, it would be simpler to use H.J strategy, I really
> don't want to re-implement what compiler might do to lay out a C array
> to assembler.

And so it seems it does not really help clang, the integrated assembler
still reorganizes the global asm directives which breaks the required
ABI size directives.
  

On 2022-04-05, Adhemerval Zanella wrote:
>
>
>On 05/04/2022 13:44, Fāng-ruì Sòng wrote:
>> On Tue, Apr 5, 2022 at 10:40 AM Adhemerval Zanella
>> <adhemerval.zanella@linaro.org> wrote:
>>>
>>>
>>>
>>> On 05/04/2022 13:22, Fāng-ruì Sòng wrote:
>>>> On Tue, Apr 5, 2022 at 8:35 AM Adhemerval Zanella
>>>> <adhemerval.zanella@linaro.org> wrote:
>>>>>
>>>>>
>>>>>
>>>>> On 04/04/2022 12:57, Fangrui Song wrote:
>>>>>>
>>>>>> On 2022-03-31, Adhemerval Zanella wrote:
>>>>>>>
>>>>>>>
>>>>>>> On 31/03/2022 00:43, Fangrui Song wrote:
>>>>>>>> On 2022-03-30, Adhemerval Zanella wrote:
>>>>>>>>>
>>>>>>>>>
>>>>>>>>> On 30/03/2022 13:23, Fangrui Song wrote:
>>>>>>>>>> On 2022-03-30, Adhemerval Zanella wrote:
>>>>>>>>>>>
>>>>>>>>>>>
>>>>>>>>>>> On 30/03/2022 02:07, Fangrui Song wrote:
>>>>>>>>>>>> 795985e4e751 in 2003 added -fno-unit-at-a-time to errlist.c and
>>>>>>>>>>>> siglist.c to "avoid reordering assembler output". -fno-toplevel-reorder
>>>>>>>>>>>> is a rough replacement for this legacy option
>>>>>>>>>>>> (https://sourceware.org/pipermail/gcc-patches/2006-January/186801.html).
>>>>>>>>>>>>
>>>>>>>>>>>> The reordering requirement does not seem to be needed any longer.
>>>>>>>>>>>
>>>>>>>>>>> We still need them for otherwise DEFINE_COMPAT_ERRLIST used on errlist-compat.c
>>>>>>>>>>> does not create _sys_errlist and _sys_siglist with expected sizes defined by
>>>>>>>>>>> glibc ABI.
>>>>>>>>>>>
>>>>>>>>>>> I am trying to fix without resorting to compiler options.
>>>>>>>>>>
>>>>>>>>>> DEFINE_COMPAT_ERRLIST does not expand to code/data, just reordeable directives:
>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>>        .globl  __GLIBC_2_1_sys_errlist
>>>>>>>>>> .set __GLIBC_2_1_sys_errlist, _sys_errlist_internal
>>>>>>>>>>         .type   __GLIBC_2_1_sys_errlist,@object
>>>>>>>>>>         .size   __GLIBC_2_1_sys_errlist, 1000
>>>>>>>>>>         .globl  __GLIBC_2_1__sys_errlist
>>>>>>>>>> .set __GLIBC_2_1__sys_errlist, _sys_errlist_internal
>>>>>>>>>>         .type   __GLIBC_2_1__sys_errlist,@object
>>>>>>>>>>         .size   __GLIBC_2_1__sys_errlist, 1000
>>>>>>>>>> .symver __GLIBC_2_1_sys_nerr, sys_nerr@GLIBC_2.2.5
>>>>>>>>>> .symver __GLIBC_2_1__sys_nerr, _sys_nerr@GLIBC_2.2.5
>>>>>>>>>> .symver __GLIBC_2_1_sys_errlist, sys_errlist@GLIBC_2.2.5
>>>>>>>>>> .symver __GLIBC_2_1__sys_errlist, _sys_errlist@GLIBC_2.2.5
>>>>>>>>>>         .globl  __GLIBC_2_3_sys_errlist
>>>>>>>>>> .set __GLIBC_2_3_sys_errlist, _sys_errlist_internal
>>>>>>>>>>         .type   __GLIBC_2_3_sys_errlist,@object
>>>>>>>>>>         .size   __GLIBC_2_3_sys_errlist, 1008
>>>>>>>>>>         .globl  __GLIBC_2_3__sys_errlist
>>>>>>>>>> .set __GLIBC_2_3__sys_errlist, _sys_errlist_internal
>>>>>>>>>>         .type   __GLIBC_2_3__sys_errlist,@object
>>>>>>>>>>         .size   __GLIBC_2_3__sys_errlist, 1008
>>>>>>>>>> .symver __GLIBC_2_3_sys_nerr, sys_nerr@GLIBC_2.3
>>>>>>>>>> .symver __GLIBC_2_3__sys_nerr, _sys_nerr@GLIBC_2.3
>>>>>>>>>> .symver __GLIBC_2_3_sys_errlist, sys_errlist@GLIBC_2.3
>>>>>>>>>> .symver __GLIBC_2_3__sys_errlist, _sys_errlist@GLIBC_2.3
>>>>>>>>>>         .globl  __GLIBC_2_4_sys_errlist
>>>>>>>>>>
>>>>>>>>>> I do not know whether GCC would reorder these macros. Even yes,
>>>>>>>>>> that'd just change the .symtab entries in the relocatable object file.
>>>>>>>>>> The linker behavior remains the same with reordering.
>>>>>>>>>
>>>>>>>>> It does not seem to, just remove the -fno-unit-at-a-time and issue make
>>>>>>>>> check-abi and you will see that object size for the compat symbols
>>>>>>>>> reference to _sys_err_internal instead of the define compat ones.
>>>>>>>>
>>>>>>>> I see. I think this is a brittle behavior in GNU assembler. Filed
>>>>>>>> https://sourceware.org/bugzilla/show_bug.cgi?id=29012 with detailed
>>>>>>>> information. I have created a patch but I know that will not solve
>>>>>>>> glibc's problem :(
>>>>>>>
>>>>>>> It would be good to have this fixes, but unfortunately we need a way
>>>>>>> to handle this on older binutils.  I am kind worried that the only
>>>>>>> possible way to actually fix this without resorting to any compiler
>>>>>>> flags is coding the array definitions in assembly direct...
>>>>>>
>>>>>> The GNU assembler issue has been fixed https://sourceware.org/bugzilla/show_bug.cgi?id=29012
>>>>>> (milestone 2.39, way larger than the current required version: 2.25)
>>>>>>
>>>>>> Switching to assembly output doesn't seem bad :-)
>>>>>>
>>>>>> If you keep the compiler driver option but need to refactor the nearby
>>>>>> code, you may drop -fno-unit-at-a-time. It was added in 2006 (somewhere
>>>>>> between GCC 4.1 and 4.4), while glibc requires GCC>=6.2.
>>>>>
>>>>> Good to know we won't need to rely on compiler flags to get the expected
>>>>> correct asm directives. I am still struggling to get a fix without resorting
>>>>> to compiler flags, but without much success.  Trying to move it to assembly
>>>>> might be tricky, I am not sure if the data directives would be architecture
>>>>> agnostic.
>>>>
>>>> Many directives are architecture-independent:
>>>> https://sourceware.org/binutils/docs/as/Pseudo-Ops.html#Pseudo-Ops
>>>> binutils-gdb/gas/read.c:346 The `portable[]` array.
>>>
>>> I think we can make it work with asciz directive.
>>>
>>>>
>>>> To support Clang, no refactoring is probably needed: just change
>>>> fno_unit_at_a_time to only specify -fno-toplevel-reorder (and rename
>>>> it), not th legacy -fno-unit-at-a-time.
>>>
>>> Afaik llvm does not support -fno-toplevel-reorder
>>
>> It doesn't, but its integrated assembler does not need the option to
>> have the desired semantics. I have a note that the new GNU as behavior
>> is quite similar to LLVM's integrated assembler since 2014-03:
>> https://sourceware.org/bugzilla/show_bug.cgi?id=29012#c1
>
>This is not what I am seeing on my clang branch, where clang with integrated
>moves the global asm directives to the top of the file.  And it makes
>check-abi fail due the _sys_errlist/sys_errlist being with wrong value.

I see. For gas/testsuite/gas/elf/size.s, LLVM integrated assembler's
behavior matches GNU as. The glibc errlist.os example is more complex
due to a symbol assignment implied by .symver .

https://reviews.llvm.org/D123283 should fix it. It isn't a perfect fix but
works with most cases. I can request a backport to LLVM 14.0.1 if it lands.