Re: [hatari-devel] FPU update

[Nicolas Pomarède <npomarede@xxxxxxxxxxxx>]

Le 21/04/2017 à 15:33, Nicolas Pomarède a écrit :
>                                                          ^
> In file included from
> /home/npomarede/src/hatari-fpu/src/cpu/fpp_softfloat.c:30:0:
> /home/npomarede/src/hatari-fpu/src/cpu/softfloat/softfloat-macros.h: In
> function 'add192':
> /home/npomarede/src/hatari-fpu/src/cpu/softfloat/softfloat-macros.h:439:16:
> warning: comparison between signed and unsigned integer expressions
> [-Wsign-compare]
>      z0 += ( z1 < carry1 );

Hi

for this one in add192(), carry1 should be declared as uint8_t not 
int8_t, because we have :
    carry1 = ( z2 < a2 );

which means carry1 can only be 0 or 1, not negative. In similar 
sub192(), uint8_t is correctly used.

diff -r d84dc2131312 src/cpu/softfloat/softfloat-macros.h
--- a/src/cpu/softfloat/softfloat-macros.h      Sat Apr 22 15:58:33 2017 
+0200
+++ b/src/cpu/softfloat/softfloat-macros.h      Sat Apr 22 19:58:10 2017 
+0200
@@ -428,7 +428,7 @@
  )
 {
     uint64_t z0, z1, z2;
-    int8_t carry0, carry1;
+    uint8_t carry0, carry1;


Nicolas