file "__umodsi3.s"

// $NetBSD: umodsi3.S,v 1.2 2006/04/07 14:27:33 cherry Exp $

//-
// Copyright (c) 2000, Intel Corporation
// All rights reserved.
//
// Contributed 2/15/2000 by Marius Cornea, John Harrison, Cristina Iordache,
// Ted Kubaska, Bob Norin, and Shane Story of the Computational Software Lab,
// Intel Corporation.
//
// WARRANTY DISCLAIMER
//
// 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://developer.intel.com/opensource.
//

section .text

// 32-bit unsigned integer remainder

proc __umodsi3#
align 32
global __umodsi3#
align 32

__umodsi3:

{ .mii
 alloc r31=ar.pfs,2,0,0,0
 nop.i 0
 nop.i 0;;
} { .mii
 nop.m 0

 // 32-BIT UNSIGNED INTEGER REMAINDER BEGINS HERE

 // general register used:
 //    r32 - 32-bit unsigned integer dividend
 //    r33 - 32-bit unsigned integer divisor
 //    r8 - 32-bit unsigned integer result
 //    r2 - scratch register
 // floating-point registers used: f6, f7, f8, f9, f10, f11
 // predicate registers used: p6

 zxt4 r32=r32
 zxt4 r33=r33;;
} { .mmb
 setf.sig f11=r32
 setf.sig f7=r33
 nop.b 0;;
} { .mfi
 nop.m 0
 fcvt.xf f6=f11
 nop.i 0
} { .mfi
 // get 2's complement of b
 sub r33=r0,r33
 fcvt.xf f7=f7
 mov r2 = 0x0ffdd;;
} { .mfi
 setf.exp f9 = r2
 // (1) y0
 frcpa.s1 f8,p6=f6,f7
 nop.i 0;;
}  { .mfi
 nop.m 0
 // (2) q0 = a * y0
 (p6) fma.s1 f10=f6,f8,f0
 nop.i 0
} { .mfi
 nop.m 0
 // (3) e0 = 1 - b * y0
 (p6) fnma.s1 f8=f7,f8,f1
 nop.i 0;;
} { .mfi
 nop.m 0
 // (4) q1 = q0 + e0 * q0
 (p6) fma.s1 f10=f8,f10,f10
 nop.i 0
} { .mfi
 // get 2's complement of b
 setf.sig f7=r33
 // (5) e1 = e0 * e0 + 2^-34
 (p6) fma.s1 f8=f8,f8,f9
 nop.i 0;;
} { .mfi
 nop.m 0
 // (6) q2 = q1 + e1 * q1
 (p6) fma.s1 f8=f8,f10,f10
 nop.i 0;;
} { .mfi
 nop.m 0
 // (7) q = trunc(q2)
 fcvt.fxu.trunc.s1 f8=f8
 nop.i 0;;
}  { .mfi
 nop.m 0
 // (8) r = a + (-b) * q
 xma.l f8=f8,f7,f11
 nop.i 0;;
}  { .mmi
 // remainder will be in the least significant 32 bits of r8 (if b != 0)
 getf.sig r8=f8
 nop.m 0
 nop.i 0;;
}

 // 32-BIT UNSIGNED INTEGER REMAINDER ENDS HERE

{ .mmb
 nop.m 0
 nop.m 0
 br.ret.sptk b0;;
}

endp __umodsi3