Patchwork softfloat: fix for C99

login
register
mail settings
Submitter Avi Kivity
Date Dec. 27, 2011, 3:11 p.m.
Message ID <1324998680-16003-1-git-send-email-avi@redhat.com>
Download mbox | patch
Permalink /patch/133322/
State New
Headers show

Comments

Avi Kivity - Dec. 27, 2011, 3:11 p.m.
C99 appears to consider compound literals as non-constants, and complains
when they are used in static initializers.  Switch to ordinary initializer
syntax.

Signed-off-by: Avi Kivity <avi@redhat.com>
---
 fpu/softfloat-specialize.h |    8 ++++----
 fpu/softfloat.h            |    2 ++
 2 files changed, 6 insertions(+), 4 deletions(-)
Andreas Färber - Dec. 27, 2011, 4:15 p.m.
Am 27.12.2011 16:11, schrieb Avi Kivity:
> C99 appears to consider compound literals as non-constants, and complains
> when they are used in static initializers.  Switch to ordinary initializer
> syntax.
> 
> Signed-off-by: Avi Kivity <avi@redhat.com>

Acked-by: Andreas Färber <afaerber@suse.de>

For the record, tested with --extra-cflags=-std=gnu99.

> diff --git a/fpu/softfloat-specialize.h b/fpu/softfloat-specialize.h
> index c5e2dab..4902450 100644
> --- a/fpu/softfloat-specialize.h
> +++ b/fpu/softfloat-specialize.h
> @@ -89,8 +89,8 @@ const float64 float64_default_nan = const_float64(LIT64( 0xFFF8000000000000 ));
>  #define floatx80_default_nan_low  LIT64( 0xC000000000000000 )
>  #endif
>  
> -const floatx80 floatx80_default_nan = make_floatx80(floatx80_default_nan_high,
> -                                                    floatx80_default_nan_low);
> +const floatx80 floatx80_default_nan
> +    = make_floatx80_init(floatx80_default_nan_high, floatx80_default_nan_low);

Calling it init_floatx80 would avoid the line break, but I'm okay with
it either way.

Thanks,
Andreas


$ gcc --version
gcc (SUSE Linux) 4.6.2
Copyright (C) 2011 Free Software Foundation, Inc.
This is free software; see the source for copying conditions.  There is NO
warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
Andreas Färber - Feb. 9, 2012, 4:20 p.m.
Am 27.12.2011 17:15, schrieb Andreas Färber:
> Am 27.12.2011 16:11, schrieb Avi Kivity:
>> C99 appears to consider compound literals as non-constants, and complains
>> when they are used in static initializers.  Switch to ordinary initializer
>> syntax.
>>

Reported-by: Andreas Färber <andreas.faerber@web.de>

>> Signed-off-by: Avi Kivity <avi@redhat.com>
> 
> Acked-by: Andreas Färber <afaerber@suse.de>
> 
> For the record, tested with --extra-cflags=-std=gnu99.
> 
>> diff --git a/fpu/softfloat-specialize.h b/fpu/softfloat-specialize.h
>> index c5e2dab..4902450 100644
>> --- a/fpu/softfloat-specialize.h
>> +++ b/fpu/softfloat-specialize.h
>> @@ -89,8 +89,8 @@ const float64 float64_default_nan = const_float64(LIT64( 0xFFF8000000000000 ));
>>  #define floatx80_default_nan_low  LIT64( 0xC000000000000000 )
>>  #endif
>>  
>> -const floatx80 floatx80_default_nan = make_floatx80(floatx80_default_nan_high,
>> -                                                    floatx80_default_nan_low);
>> +const floatx80 floatx80_default_nan
>> +    = make_floatx80_init(floatx80_default_nan_high, floatx80_default_nan_low);
> 
> Calling it init_floatx80 would avoid the line break, but I'm okay with
> it either way.

Ping! Avi, you didn't indicate whether you were going to simplify this
patch or whether you're waiting for someone to apply it as is?

Regards,
Andreas

> 
> Thanks,
> Andreas
> 
> 
> $ gcc --version
> gcc (SUSE Linux) 4.6.2
> Copyright (C) 2011 Free Software Foundation, Inc.
> This is free software; see the source for copying conditions.  There is NO
> warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
Avi Kivity - Feb. 9, 2012, 4:38 p.m.
On 02/09/2012 06:20 PM, Andreas Färber wrote:
> Am 27.12.2011 17:15, schrieb Andreas Färber:
> > Am 27.12.2011 16:11, schrieb Avi Kivity:
> >> C99 appears to consider compound literals as non-constants, and complains
> >> when they are used in static initializers.  Switch to ordinary initializer
> >> syntax.
> >>
>
> Reported-by: Andreas Färber <andreas.faerber@web.de>
>
> >> Signed-off-by: Avi Kivity <avi@redhat.com>
> > 
> > Acked-by: Andreas Färber <afaerber@suse.de>
> > 
> > For the record, tested with --extra-cflags=-std=gnu99.
> > 
> >> diff --git a/fpu/softfloat-specialize.h b/fpu/softfloat-specialize.h
> >> index c5e2dab..4902450 100644
> >> --- a/fpu/softfloat-specialize.h
> >> +++ b/fpu/softfloat-specialize.h
> >> @@ -89,8 +89,8 @@ const float64 float64_default_nan = const_float64(LIT64( 0xFFF8000000000000 ));
> >>  #define floatx80_default_nan_low  LIT64( 0xC000000000000000 )
> >>  #endif
> >>  
> >> -const floatx80 floatx80_default_nan = make_floatx80(floatx80_default_nan_high,
> >> -                                                    floatx80_default_nan_low);
> >> +const floatx80 floatx80_default_nan
> >> +    = make_floatx80_init(floatx80_default_nan_high, floatx80_default_nan_low);
> > 
> > Calling it init_floatx80 would avoid the line break, but I'm okay with
> > it either way.
>
> Ping! Avi, you didn't indicate whether you were going to simplify this
> patch or whether you're waiting for someone to apply it as is?
>

Actually I forgot all about it.  If everyone's okay with it as is I'd
like it to be applied, otherwise I'll respin.  Copying some maintainers
completely at random.
Andreas Färber - March 16, 2012, 10:01 a.m.
Am 09.02.2012 17:38, schrieb Avi Kivity:
> On 02/09/2012 06:20 PM, Andreas Färber wrote:
>> Am 27.12.2011 17:15, schrieb Andreas Färber:
>>> Am 27.12.2011 16:11, schrieb Avi Kivity:
>>>> C99 appears to consider compound literals as non-constants, and complains
>>>> when they are used in static initializers.  Switch to ordinary initializer
>>>> syntax.
>>>>
>>
>> Reported-by: Andreas Färber <andreas.faerber@web.de>
>>
>>>> Signed-off-by: Avi Kivity <avi@redhat.com>
>>>
>>> Acked-by: Andreas Färber <afaerber@suse.de>
>>>
>>> For the record, tested with --extra-cflags=-std=gnu99.
>>>
>>>> diff --git a/fpu/softfloat-specialize.h b/fpu/softfloat-specialize.h
>>>> index c5e2dab..4902450 100644
>>>> --- a/fpu/softfloat-specialize.h
>>>> +++ b/fpu/softfloat-specialize.h
>>>> @@ -89,8 +89,8 @@ const float64 float64_default_nan = const_float64(LIT64( 0xFFF8000000000000 ));
>>>>  #define floatx80_default_nan_low  LIT64( 0xC000000000000000 )
>>>>  #endif
>>>>  
>>>> -const floatx80 floatx80_default_nan = make_floatx80(floatx80_default_nan_high,
>>>> -                                                    floatx80_default_nan_low);
>>>> +const floatx80 floatx80_default_nan
>>>> +    = make_floatx80_init(floatx80_default_nan_high, floatx80_default_nan_low);
>>>
>>> Calling it init_floatx80 would avoid the line break, but I'm okay with
>>> it either way.
>>
>> Ping! Avi, you didn't indicate whether you were going to simplify this
>> patch or whether you're waiting for someone to apply it as is?
>>
> 
> Actually I forgot all about it.  If everyone's okay with it as is I'd
> like it to be applied, otherwise I'll respin.  Copying some maintainers
> completely at random.

Ping?

Andreas
Blue Swirl - March 17, 2012, 4:24 p.m.
Thanks, applied.

On Tue, Dec 27, 2011 at 15:11, Avi Kivity <avi@redhat.com> wrote:
> C99 appears to consider compound literals as non-constants, and complains
> when they are used in static initializers.  Switch to ordinary initializer
> syntax.
>
> Signed-off-by: Avi Kivity <avi@redhat.com>
> ---
>  fpu/softfloat-specialize.h |    8 ++++----
>  fpu/softfloat.h            |    2 ++
>  2 files changed, 6 insertions(+), 4 deletions(-)
>
> diff --git a/fpu/softfloat-specialize.h b/fpu/softfloat-specialize.h
> index c5e2dab..4902450 100644
> --- a/fpu/softfloat-specialize.h
> +++ b/fpu/softfloat-specialize.h
> @@ -89,8 +89,8 @@ const float64 float64_default_nan = const_float64(LIT64( 0xFFF8000000000000 ));
>  #define floatx80_default_nan_low  LIT64( 0xC000000000000000 )
>  #endif
>
> -const floatx80 floatx80_default_nan = make_floatx80(floatx80_default_nan_high,
> -                                                    floatx80_default_nan_low);
> +const floatx80 floatx80_default_nan
> +    = make_floatx80_init(floatx80_default_nan_high, floatx80_default_nan_low);
>
>  /*----------------------------------------------------------------------------
>  | The pattern for a default generated quadruple-precision NaN.  The `high' and
> @@ -104,8 +104,8 @@ const floatx80 floatx80_default_nan = make_floatx80(floatx80_default_nan_high,
>  #define float128_default_nan_low  LIT64( 0x0000000000000000 )
>  #endif
>
> -const float128 float128_default_nan = make_float128(float128_default_nan_high,
> -                                                    float128_default_nan_low);
> +const float128 float128_default_nan
> +    = make_float128_init(float128_default_nan_high, float128_default_nan_low);
>
>  /*----------------------------------------------------------------------------
>  | Raises the exceptions specified by `flags'.  Floating-point traps can be
> diff --git a/fpu/softfloat.h b/fpu/softfloat.h
> index 07c2929..2ce4110 100644
> --- a/fpu/softfloat.h
> +++ b/fpu/softfloat.h
> @@ -129,6 +129,7 @@ typedef struct {
>     uint16_t high;
>  } floatx80;
>  #define make_floatx80(exp, mant) ((floatx80) { mant, exp })
> +#define make_floatx80_init(exp, mant) { .low = mant, .high = exp }
>  typedef struct {
>  #ifdef HOST_WORDS_BIGENDIAN
>     uint64_t high, low;
> @@ -137,6 +138,7 @@ typedef struct {
>  #endif
>  } float128;
>  #define make_float128(high_, low_) ((float128) { .high = high_, .low = low_ })
> +#define make_float128_init(high_, low_) { .high = high_, .low = low_ }
>
>  /*----------------------------------------------------------------------------
>  | Software IEC/IEEE floating-point underflow tininess-detection mode.
> --
> 1.7.7.1
>
>

Patch

diff --git a/fpu/softfloat-specialize.h b/fpu/softfloat-specialize.h
index c5e2dab..4902450 100644
--- a/fpu/softfloat-specialize.h
+++ b/fpu/softfloat-specialize.h
@@ -89,8 +89,8 @@  const float64 float64_default_nan = const_float64(LIT64( 0xFFF8000000000000 ));
 #define floatx80_default_nan_low  LIT64( 0xC000000000000000 )
 #endif
 
-const floatx80 floatx80_default_nan = make_floatx80(floatx80_default_nan_high,
-                                                    floatx80_default_nan_low);
+const floatx80 floatx80_default_nan
+    = make_floatx80_init(floatx80_default_nan_high, floatx80_default_nan_low);
 
 /*----------------------------------------------------------------------------
 | The pattern for a default generated quadruple-precision NaN.  The `high' and
@@ -104,8 +104,8 @@  const floatx80 floatx80_default_nan = make_floatx80(floatx80_default_nan_high,
 #define float128_default_nan_low  LIT64( 0x0000000000000000 )
 #endif
 
-const float128 float128_default_nan = make_float128(float128_default_nan_high,
-                                                    float128_default_nan_low);
+const float128 float128_default_nan
+    = make_float128_init(float128_default_nan_high, float128_default_nan_low);
 
 /*----------------------------------------------------------------------------
 | Raises the exceptions specified by `flags'.  Floating-point traps can be
diff --git a/fpu/softfloat.h b/fpu/softfloat.h
index 07c2929..2ce4110 100644
--- a/fpu/softfloat.h
+++ b/fpu/softfloat.h
@@ -129,6 +129,7 @@  typedef struct {
     uint16_t high;
 } floatx80;
 #define make_floatx80(exp, mant) ((floatx80) { mant, exp })
+#define make_floatx80_init(exp, mant) { .low = mant, .high = exp }
 typedef struct {
 #ifdef HOST_WORDS_BIGENDIAN
     uint64_t high, low;
@@ -137,6 +138,7 @@  typedef struct {
 #endif
 } float128;
 #define make_float128(high_, low_) ((float128) { .high = high_, .low = low_ })
+#define make_float128_init(high_, low_) { .high = high_, .low = low_ }
 
 /*----------------------------------------------------------------------------
 | Software IEC/IEEE floating-point underflow tininess-detection mode.