newlib-cygwin/winsup/cygwin/math/ceill.S

65 lines
1.4 KiB
ArmAsm

/**
* This file has no copyright assigned and is placed in the Public Domain.
* This file is part of the mingw-w64 runtime package.
* No warranty is given; refer to the file DISCLAIMER.PD within this package.
*/
#include <_mingw_mac.h>
.file "ceill.S"
.text
#ifdef __x86_64__
.align 8
#else
.align 4
#endif
.globl __MINGW_USYMBOL(ceill)
.def __MINGW_USYMBOL(ceill); .scl 2; .type 32; .endef
__MINGW_USYMBOL(ceill):
#if defined(_AMD64_) || defined(__x86_64__)
fldt (%rdx)
subq $24,%rsp
fstcw 8(%rsp) /* store fpu control word */
/* We use here %edx although only the low 1 bits are defined.
But none of the operations should care and they are faster
than the 16 bit operations. */
movl $0x0800,%edx /* round towards +oo */
orl 8(%rsp),%edx
andl $0xfbff,%edx
movl %edx,(%rsp)
fldcw (%rsp) /* load modified control word */
frndint /* round */
fldcw 8(%rsp) /* restore original control word */
addq $24,%rsp
movq %rcx,%rax
movq $0,8(%rcx)
fstpt (%rcx)
ret
#elif defined(_ARM_) || defined(__arm__)
vmrs r1, fpscr
bic r0, r1, #0x00c00000
orr r0, r0, #0x00400000 /* Round towards Plus Infinity */
vmsr fpscr, r0
vcvtr.s32.f64 s0, d0
vcvt.f64.s32 d0, s0
vmsr fpscr, r1
bx lr
#elif defined(_X86_) || defined(__i386__)
fldt 4(%esp)
subl $8,%esp
fstcw 4(%esp)
movl $0x0800,%edx
orl 4(%esp),%edx
andl $0xfbff,%edx
movl %edx,(%esp)
fldcw (%esp)
frndint
fldcw 4(%esp)
addl $8,%esp
ret
#endif