This is the mail archive of the mailing list for the glibc project.

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: glibc 2.3 on linux-mips

On Sat, Oct 19, 2002 at 06:39:47PM +0200, Guido Guenther wrote:

> On Mon, Oct 07, 2002 at 01:15:02AM +0200, Guido Guenther wrote:
> > make[4]: *** [/glibc/glibc-package-2.3/mips-linux/obj/math/test-fpucw.out] Error 1
> This test fails since the read back fpu control word is 0x80000 instead
> of 0x0. I wonder if this patch is correct:
> diff -u -r1.6 fpu_control.h
> --- sysdeps/mips//fpu_control.h	6 Jul 2001 04:56:00 -0000	1.6
> +++ sysdeps/mips//fpu_control.h	19 Oct 2002 16:03:35 -0000
> @@ -74,7 +74,7 @@
>  #define _FPU_RC_UP      0x2
>  #define _FPU_RC_DOWN    0x3
> -#define _FPU_RESERVED 0xfe3c0000  /* Reserved bits in cw */
> +#define _FPU_RESERVED   0xfebc0000  /* Reserved bits in cw */
>  /* The fdlibm code requires strict IEEE double precision arithmetic,
> which additionally masks out the condition bit 23 - note that the other
> condition bits (25-31) are masked out too?

The value of _FPU_RESERVED reflects some particular MIPS CPU; I think I
choose the R4600 back in the dark ages when the universe was formed ;-)

As different CPU declare reserve different bits in fcr31 I'm not sure which
bits we should set in _FPU_RESERVED.  Two possibilities, the bits that are
reserved on all FPUs or those that are reserved on the lowest end supported

Imho test-fpucw.c as well; it assumes that reserved bits have to be
zero will fail on anything that doesn't fulfill that assumption.  That hold
for MIPS FPUs but probably not on all FPUs of all architectures out there.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]