diff options
Diffstat (limited to 'sysdeps/ia64/fpu/s_fdiml.S')
| -rw-r--r-- | sysdeps/ia64/fpu/s_fdiml.S | 227 |
1 files changed, 227 insertions, 0 deletions
diff --git a/sysdeps/ia64/fpu/s_fdiml.S b/sysdeps/ia64/fpu/s_fdiml.S new file mode 100644 index 0000000000..00876c3904 --- /dev/null +++ b/sysdeps/ia64/fpu/s_fdiml.S @@ -0,0 +1,227 @@ +.file "fdiml.s" + + +// Copyright (c) 2001 - 2003, Intel Corporation +// All rights reserved. +// +// Contributed 2001 by the Intel Numerics Group, Intel Corporation +// +// Redistribution and use in source and binary forms, with or without +// modification, are permitted provided that the following conditions are +// met: +// +// * Redistributions of source code must retain the above copyright +// notice, this list of conditions and the following disclaimer. +// +// * Redistributions in binary form must reproduce the above copyright +// notice, this list of conditions and the following disclaimer in the +// documentation and/or other materials provided with the distribution. +// +// * The name of Intel Corporation may not be used to endorse or promote +// products derived from this software without specific prior written +// permission. + +// THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS +// "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT +// LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR +// A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL INTEL OR ITS +// CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, +// EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, +// PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR +// PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY +// OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY OR TORT (INCLUDING +// NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS +// SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. +// +// Intel Corporation is the author of this code, and requests that all +// problem reports or change requests be submitted to it directly at +// http://www.intel.com/software/products/opensource/libraries/num.htm. +// +// History +//============================================================== +// 06/08/01 Initial version +// 08/23/01 Corrected error tag number +// 05/20/02 Cleaned up namespace and sf0 syntax +// 01/28/03 Improved performance; fixed parameters for call to error routine +// +// API +//============================================================== +// long double fdiml( long double x, long double y ); +// input floating point f8, f9 +// output floating point f8 +// +// +// Overview of operation +//============================================================== +// fdiml determines the positive difference between the arguments +// Result = x - y if x > y +// = +0 if x <= y +// +// Error support is called if x-y overflows for x > y +// + +// Registers used +//============================================================== +// General purpose registers: r14, r32 - r39 + +rExpBig = r14 + +// r36-39 parameters for libm_error_support +GR_SAVE_B0 = r33 +GR_SAVE_GP = r34 +GR_SAVE_PFS = r35 + +GR_Parameter_X = r36 +GR_Parameter_Y = r37 +GR_Parameter_RESULT = r38 +GR_Parameter_TAG = r39 + +// Floating-point registers: f8 - f12 + +f_tmp_result = f10 +fBig = f11 +fNormX = f12 + +// Predicate registers: p6 - p10 + + +.section .text +GLOBAL_LIBM_ENTRY(fdiml) + +{ .mfi + mov rExpBig = 0x13fff // Exponent to indicate overflow + fcmp.le.s1 p6,p7 = f8, f9 // Is x <= y? + nop.i 0 +} +{ .mfi + nop.m 0 + fnorm.s1 fNormX = f8 // Save x + nop.i 0 +} +;; + +{ .mfi + setf.exp fBig = rExpBig // Constant to test for overflow + fcmp.eq.s0 p8,p0 = f8, f9 // Dummy op to set Denormal or Invalid + nop.i 0 +} +;; + +{ .mfi + nop.m 0 + fclass.m p9,p10 = f8, 0x1e3 // Test for x natval, nan, inf + nop.i 0 +} +;; + +{ .mfi + nop.m 0 +(p6) fmerge.s f8 = f0, f0 // Result is +0 if x <= y + nop.i 0 +} +{ .mfi + nop.m 0 +(p7) fms.s0 f8 = f8, f1, f9 // Result is x - y if x > y + nop.i 0 +} +;; + +{ .mfi + nop.m 0 +(p10) fclass.m p9,p10 = f9, 0x1e3 // Test for y natval, nan, inf + nop.i 0 +} +;; + +{ .mfi + nop.m 0 +(p10) fcmp.ge.s1 p8,p0 = f8, fBig // Test result for overflow + nop.i 0 +} +;; + +{ .mbb +(p9) cmp.ne p8,p0 = r0,r0 // Clear p8 if x or y natval,nan,inf +(p8) br.cond.spnt FDIM_OVERFLOW // Branch if result overflows + br.ret.sptk b0 // Normal return +} +;; + + +// Here if result will overflow +FDIM_OVERFLOW: +{ .mfi + alloc r32=ar.pfs,2,2,4,0 + fms.s0 f_tmp_result = f8,f1,f9 // Normalize result force overflow + nop.i 0 +} +{ .mfb + mov GR_Parameter_TAG = 195 // Error code + nop.f 0 + br.cond.sptk __libm_error_region // Branch to error code +} +;; + +GLOBAL_LIBM_END(fdiml) + +LOCAL_LIBM_ENTRY(__libm_error_region) +// Call error support to report possible range error +.prologue + +{ .mfi + add GR_Parameter_Y=-32,sp // Parameter 2 value + nop.f 0 +.save ar.pfs,GR_SAVE_PFS + mov GR_SAVE_PFS=ar.pfs // Save ar.pfs +} +{ .mfi +.fframe 64 + add sp=-64,sp // Create new stack + nop.f 0 + mov GR_SAVE_GP=gp // Save gp +};; + +{ .mmi + stfe [GR_Parameter_Y] = f9,16 // STORE Parameter 2 on stack + add GR_Parameter_X = 16,sp // Parameter 1 address +.save b0, GR_SAVE_B0 + mov GR_SAVE_B0=b0 // Save b0 +};; + +.body +{ .mib + stfe [GR_Parameter_X] = fNormX // STORE Parameter 1 on stack + add GR_Parameter_RESULT = 0,GR_Parameter_Y // Parameter 3 address + nop.b 0 +} +{ .mib + stfe [GR_Parameter_Y] = f_tmp_result // STORE Parameter 3 on stack + add GR_Parameter_Y = -16,GR_Parameter_Y + br.call.sptk b0=__libm_error_support# // Call error handling function +};; + +{ .mmi + add GR_Parameter_RESULT = 48,sp + nop.m 0 + nop.i 0 +};; + +{ .mmi + ldfe f8 = [GR_Parameter_RESULT] // Get return result off stack +.restore sp + add sp = 64,sp // Restore stack pointer + mov b0 = GR_SAVE_B0 // Restore return address +};; + +{ .mib + mov gp = GR_SAVE_GP // Restore gp + mov ar.pfs = GR_SAVE_PFS // Restore ar.pfs + br.ret.sptk b0 // Return +};; + +LOCAL_LIBM_END(__libm_error_region) + + +.type __libm_error_support#,@function +.global __libm_error_support# + |