diff options
| author | das <das@FreeBSD.org> | 2005-03-05 20:34:45 +0000 |
|---|---|---|
| committer | das <das@FreeBSD.org> | 2005-03-05 20:34:45 +0000 |
| commit | 062f662fe278c9446b719d13d73d55acefeb1ff3 (patch) | |
| tree | 8c2afc0654fec0a65cb52419b675e5d0aac281d3 /lib/msun | |
| parent | 9f6d0603853138731f15e3874369b807e66f2f16 (diff) | |
| download | FreeBSD-src-062f662fe278c9446b719d13d73d55acefeb1ff3.zip FreeBSD-src-062f662fe278c9446b719d13d73d55acefeb1ff3.tar.gz | |
Work around a gcc bug. This fixes feholdexcept() et al. at -O1.
Symptoms of the problem included assembler warnings and
nondeterministic runtime behavior when a fe*() call that affects the
fpsr is closely followed by a float point op.
The bug (at least, I think it's a bug) is that gcc does not insert a
break between a volatile asm and a dependent instruction if the
volatile asm came from an inlined function. Volatile asms seem to be
fine in other circumstances, even without -mvolatile-asm-stop, so
perhaps the compiler adds the stop bits before inlining takes place.
The problem does not occur at -O0 because inlining is disabled, and it
doesn't happen at -O2 because -fschedule-insns2 knows better.
Diffstat (limited to 'lib/msun')
| -rw-r--r-- | lib/msun/ia64/fenv.h | 2 |
1 files changed, 1 insertions, 1 deletions
diff --git a/lib/msun/ia64/fenv.h b/lib/msun/ia64/fenv.h index f1415f1..98992b1 100644 --- a/lib/msun/ia64/fenv.h +++ b/lib/msun/ia64/fenv.h @@ -61,7 +61,7 @@ extern const fenv_t __fe_dfl_env; #define _FPUSW_SHIFT 13 #define __stfpsr(__r) __asm __volatile("mov %0=ar.fpsr" : "=r" (*(__r))) -#define __ldfpsr(__r) __asm __volatile("mov ar.fpsr=%0" : : "r" (__r)) +#define __ldfpsr(__r) __asm __volatile("mov ar.fpsr=%0;;" : : "r" (__r)) static __inline int feclearexcept(int __excepts) |
