x86: fix Autoconf caching of instruction support checks [BZ #27991]

  The Autoconf documentation for the AC_CACHE_CHECK macro states:

  The commands-to-set-it must have no side effects except for setting
  the variable cache-id, see below.

However, the tests for support of -msahf and -mmovbe were embedded in
the commands-to-set-it for lib_cv_include_x86_isa_level. This had the
consequence that libc_cv_have_x86_lahf_sahf and libc_cv_have_x86_movbe
were not defined whenever lib_cv_include_x86_isa_level was read from
cache. These variables' being undefined meant that their unquoted use
in later test expressions led to the 'test' built-in's misparsing its
arguments and emitting errors like "test: =: unexpected operator" or
"test: =: unary operator expected", depending on the particular shell.

This commit refactors the tests for LAHF/SAHF and MOVBE instruction
support into their own AC_CACHE_CHECK macro invocations to obey the
rule that the commands-to-set-it must have no side effects other than
setting the variable named by cache-id.
---
 sysdeps/x86/configure    | 62 ++++++++++++++++++++++++----------------
 sysdeps/x86/configure.ac | 34 +++++++++++-----------
 2 files changed, 56 insertions(+), 40 deletions(-)
  

On 17/06/2021 00:40, Matt Whitlock wrote:
> The Autoconf documentation for the AC_CACHE_CHECK macro states:
> 
>   The commands-to-set-it must have no side effects except for setting
>   the variable cache-id, see below.
> 
> However, the tests for support of -msahf and -mmovbe were embedded in
> the commands-to-set-it for lib_cv_include_x86_isa_level. This had the
> consequence that libc_cv_have_x86_lahf_sahf and libc_cv_have_x86_movbe
> were not defined whenever lib_cv_include_x86_isa_level was read from
> cache. These variables' being undefined meant that their unquoted use
> in later test expressions led to the 'test' built-in's misparsing its
> arguments and emitting errors like "test: =: unexpected operator" or
> "test: =: unary operator expected", depending on the particular shell.
> 
> This commit refactors the tests for LAHF/SAHF and MOVBE instruction
> support into their own AC_CACHE_CHECK macro invocations to obey the
> rule that the commands-to-set-it must have no side effects other than
> setting the variable named by cache-id.

LGTM, thanks.  Also, please confirm that you're the original author and 
are authorized to contribute this patch by adding either a DCO, i.e. add a 
Signed-off-by to indicate that; or by copyright assignment to the FSF
(check 2.1 and 2.2 on the Contribution Checklist [1])

Reviewed-by: Adhemerval Zanella  <adhemerval.zanella@linaro.org>

[1] https://sourceware.org/glibc/wiki/Contribution%20checklist

> ---
>  sysdeps/x86/configure    | 62 ++++++++++++++++++++++++----------------
>  sysdeps/x86/configure.ac | 34 +++++++++++-----------
>  2 files changed, 56 insertions(+), 40 deletions(-)
> 
> diff --git a/sysdeps/x86/configure b/sysdeps/x86/configure
> index ead1295c38..62676bb686 100644
> --- a/sysdeps/x86/configure
> +++ b/sysdeps/x86/configure
> @@ -126,8 +126,6 @@ cat > conftest2.S <<EOF
>  4:
>  EOF
>  libc_cv_include_x86_isa_level=no
> -libc_cv_have_x86_lahf_sahf=no
> -libc_cv_have_x86_movbe=no
>  if { ac_try='${CC-cc} $CFLAGS $CPPFLAGS -nostartfiles -nostdlib -r -o conftest conftest1.S conftest2.S'
>    { { eval echo "\"\$as_me\":${as_lineno-$LINENO}: \"$ac_try\""; } >&5
>    (eval $ac_try) 2>&5
> @@ -137,24 +135,6 @@ if { ac_try='${CC-cc} $CFLAGS $CPPFLAGS -nostartfiles -nostdlib -r -o conftest c
>    count=`LC_ALL=C $READELF -n conftest | grep NT_GNU_PROPERTY_TYPE_0 | wc -l`
>    if test "$count" = 1; then
>      libc_cv_include_x86_isa_level=yes
> -    cat > conftest.c <<EOF
> -EOF
> -    if { ac_try='${CC-cc} $CFLAGS $CPPFLAGS -fverbose-asm -S -o - conftest.c'
> -  { { eval echo "\"\$as_me\":${as_lineno-$LINENO}: \"$ac_try\""; } >&5
> -  (eval $ac_try) 2>&5
> -  ac_status=$?
> -  $as_echo "$as_me:${as_lineno-$LINENO}: \$? = $ac_status" >&5
> -  test $ac_status = 0; }; } | grep -q "\-msahf"; then
> -      libc_cv_have_x86_lahf_sahf=yes
> -    fi
> -    if { ac_try='${CC-cc} $CFLAGS $CPPFLAGS -fverbose-asm -S -o - conftest.c'
> -  { { eval echo "\"\$as_me\":${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; }; } | grep -q "\-mmovbe"; then
> -      libc_cv_have_x86_movbe=yes
> -    fi
>    fi
>  fi
>  rm -f conftest*
> @@ -164,14 +144,48 @@ $as_echo "$libc_cv_include_x86_isa_level" >&6; }
>  if test $libc_cv_include_x86_isa_level = yes; then
>    $as_echo "#define INCLUDE_X86_ISA_LEVEL 1" >>confdefs.h
>  
> +  { $as_echo "$as_me:${as_lineno-$LINENO}: checking for LAHF/SAHF instruction support" >&5
> +$as_echo_n "checking for LAHF/SAHF instruction support... " >&6; }
> +if ${libc_cv_have_x86_lahf_sahf+:} false; then :
> +  $as_echo_n "(cached) " >&6
> +else
> +      libc_cv_have_x86_lahf_sahf=no
> +    if { ac_try='${CC-cc} $CFLAGS $CPPFLAGS -fverbose-asm -S -o - -x c /dev/null'
> +  { { eval echo "\"\$as_me\":${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; }; } | grep -q "\-msahf"; then
> +      libc_cv_have_x86_lahf_sahf=yes
> +    fi
>  fi
> -if test $libc_cv_have_x86_lahf_sahf = yes; then
> -  $as_echo "#define HAVE_X86_LAHF_SAHF 1" >>confdefs.h
> +{ $as_echo "$as_me:${as_lineno-$LINENO}: result: $libc_cv_have_x86_lahf_sahf" >&5
> +$as_echo "$libc_cv_have_x86_lahf_sahf" >&6; }
> +  if test $libc_cv_have_x86_lahf_sahf = yes; then
> +    $as_echo "#define HAVE_X86_LAHF_SAHF 1" >>confdefs.h
>  
> +  fi
> +  { $as_echo "$as_me:${as_lineno-$LINENO}: checking for MOVBE instruction support" >&5
> +$as_echo_n "checking for MOVBE instruction support... " >&6; }
> +if ${libc_cv_have_x86_movbe+:} false; then :
> +  $as_echo_n "(cached) " >&6
> +else
> +      libc_cv_have_x86_movbe=no
> +    if { ac_try='${CC-cc} $CFLAGS $CPPFLAGS -fverbose-asm -S -o - -x c /dev/null'
> +  { { eval echo "\"\$as_me\":${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; }; } | grep -q "\-mmovbe"; then
> +      libc_cv_have_x86_movbe=yes
> +    fi
>  fi
> -if test $libc_cv_have_x86_movbe = yes; then
> -  $as_echo "#define HAVE_X86_MOVBE 1" >>confdefs.h
> +{ $as_echo "$as_me:${as_lineno-$LINENO}: result: $libc_cv_have_x86_movbe" >&5
> +$as_echo "$libc_cv_have_x86_movbe" >&6; }
> +  if test $libc_cv_have_x86_movbe = yes; then
> +    $as_echo "#define HAVE_X86_MOVBE 1" >>confdefs.h
>  
> +  fi
>  fi
>  config_vars="$config_vars
>  enable-x86-isa-level = $libc_cv_include_x86_isa_level"
> diff --git a/sysdeps/x86/configure.ac b/sysdeps/x86/configure.ac
> index bca97fdc2f..04a12ab680 100644
> --- a/sysdeps/x86/configure.ac
> +++ b/sysdeps/x86/configure.ac
> @@ -98,30 +98,32 @@ cat > conftest2.S <<EOF
>  4:
>  EOF
>  libc_cv_include_x86_isa_level=no
> -libc_cv_have_x86_lahf_sahf=no
> -libc_cv_have_x86_movbe=no
>  if AC_TRY_COMMAND(${CC-cc} $CFLAGS $CPPFLAGS -nostartfiles -nostdlib -r -o conftest conftest1.S conftest2.S); then
>    count=`LC_ALL=C $READELF -n conftest | grep NT_GNU_PROPERTY_TYPE_0 | wc -l`
>    if test "$count" = 1; then
>      libc_cv_include_x86_isa_level=yes
> -    cat > conftest.c <<EOF
> -EOF
> -    if AC_TRY_COMMAND(${CC-cc} $CFLAGS $CPPFLAGS -fverbose-asm -S -o - conftest.c) | grep -q "\-msahf"; then
> -      libc_cv_have_x86_lahf_sahf=yes
> -    fi
> -    if AC_TRY_COMMAND(${CC-cc} $CFLAGS $CPPFLAGS -fverbose-asm -S -o - conftest.c) | grep -q "\-mmovbe"; then
> -      libc_cv_have_x86_movbe=yes
> -    fi
>    fi
>  fi
>  rm -f conftest*])
>  if test $libc_cv_include_x86_isa_level = yes; then
>    AC_DEFINE(INCLUDE_X86_ISA_LEVEL)
> -fi
> -if test $libc_cv_have_x86_lahf_sahf = yes; then
> -  AC_DEFINE(HAVE_X86_LAHF_SAHF)
> -fi
> -if test $libc_cv_have_x86_movbe = yes; then
> -  AC_DEFINE(HAVE_X86_MOVBE)
> +  AC_CACHE_CHECK([for LAHF/SAHF instruction support],
> +		 libc_cv_have_x86_lahf_sahf, [dnl
> +    libc_cv_have_x86_lahf_sahf=no
> +    if AC_TRY_COMMAND(${CC-cc} $CFLAGS $CPPFLAGS -fverbose-asm -S -o - -x c /dev/null) | grep -q "\-msahf"; then
> +      libc_cv_have_x86_lahf_sahf=yes
> +    fi])
> +  if test $libc_cv_have_x86_lahf_sahf = yes; then
> +    AC_DEFINE(HAVE_X86_LAHF_SAHF)
> +  fi
> +  AC_CACHE_CHECK([for MOVBE instruction support],
> +		 libc_cv_have_x86_movbe, [dnl
> +    libc_cv_have_x86_movbe=no
> +    if AC_TRY_COMMAND(${CC-cc} $CFLAGS $CPPFLAGS -fverbose-asm -S -o - -x c /dev/null) | grep -q "\-mmovbe"; then
> +      libc_cv_have_x86_movbe=yes
> +    fi])
> +  if test $libc_cv_have_x86_movbe = yes; then
> +    AC_DEFINE(HAVE_X86_MOVBE)
> +  fi
>  fi
>  LIBC_CONFIG_VAR([enable-x86-isa-level], [$libc_cv_include_x86_isa_level])
>
  

On Wednesday, August 11, 2021 12:44:09 PM EDT, Adhemerval Zanella wrote:
>
> On 17/06/2021 00:40, Matt Whitlock wrote:
>> The Autoconf documentation for the AC_CACHE_CHECK macro states:
>> 
>>   The commands-to-set-it must have no side effects except for setting
>>   the variable cache-id, see below.
>> 
>> However, the tests for support of -msahf and -mmovbe were embedded in
>> the commands-to-set-it for lib_cv_include_x86_isa_level. This had the ...
>
> LGTM, thanks.  Also, please confirm that you're the original author and 
> are authorized to contribute this patch by adding either a DCO, i.e. add a 
> Signed-off-by to indicate that; or by copyright assignment to the FSF
> (check 2.1 and 2.2 on the Contribution Checklist [1])

The lawyers are ruining free software development. Not being one myself, I 
make no legal claims about my contribution, but I do assert that it is the 
product of my own work that I performed without interactive assistance or 
collaboration and, as I do not believe in the concept of intellectual 
property, that it is my intention to refuse in perpetuity any and all 
claims of ownership that anyone may ever have believed that I had over this 
material. If you need a few extra bytes at the end of this email message to 
help you feel good, then here they are:

Signed-off-by: Matt Whitlock <sourceware@mattwhitlock.name>