Bring in the following patch from FSF binutils mainline:

2003-06-13  Richard Earnshaw  <rearnsha@arm.com>

	* tc-arm.c (FPU_DEFAULT, case TE_LINUX): Default to FPU_ARCH_FPA.
	(FPU_DEFAULT, case TE_NetBSD): Default to FPU_ARCH_VFP for ELF,
	FPU_ARCH_FPA for AOUT.
	(md_begin): Don't try to guess the floating point architecture from
	the CPU if the OS ABI (Linux, NetBSD) mandates a particular form.

...per Richard's suggestion.
This commit is contained in:
thorpej 2003-06-14 16:02:57 +00:00
parent 17ae477780
commit 1d99a54efe
1 changed files with 18 additions and 0 deletions

View File

@ -132,6 +132,18 @@
#endif
#endif
#ifdef TE_LINUX
#define FPU_DEFAULT FPU_ARCH_FPA
#endif
#ifdef TE_NetBSD
#ifdef OBJ_ELF
#define FPU_DEFAULT FPU_ARCH_VFP /* Soft-float, but VFP order. */
#else
#define FPU_DEFAULT FPU_ARCH_FPA /* Soft-float, but FPA order. */
#endif
#endif
/* For backwards compatibility we default to the FPA. */
#ifndef FPU_DEFAULT
#define FPU_DEFAULT FPU_ARCH_FPA
@ -9253,10 +9265,16 @@ md_begin ()
}
else if (mfpu_opt == -1)
{
#if !(defined (TE_LINUX) || defined(TE_NetBSD))
/* Some environments specify a default FPU. If they don't, infer
it from the processor. */
if (mcpu_fpu_opt != -1)
mfpu_opt = mcpu_fpu_opt;
else
mfpu_opt = march_fpu_opt;
#else
mfpu_opt = FPU_DEFAULT;
#endif
}
if (mfpu_opt == -1)