freebsd-skq/sys/libkern/ia64/__moddi3.S
schweikh 804e8df48c Improve a comment and undo a bogus s/a/an/ in a comment. An asm guru
should add more comments explaining which registers hold which variables.

Spotted by:	bde
2003-02-11 20:15:11 +00:00

159 lines
3.4 KiB
ArmAsm

.file "__moddi3.s"
// $FreeBSD$
//
// 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
// 64-bit signed integer remainder
.proc __moddi3#
.align 32
.global __moddi3#
.align 32
__moddi3:
{ .mii
alloc r31=ar.pfs,3,0,0,0
nop.i 0
nop.i 0
} { .mmb
// 64-BIT SIGNED INTEGER REMAINDER BEGINS HERE
// general register used:
// r32 - 64-bit signed integer dividend, called a below
// r33 - 64-bit signed integer divisor, called b below
// r8 - 64-bit signed integer result
// r2 - scratch register
// floating-point registers used: f6, f7, f8, f9, f10, f11, f12
// predicate registers used: p6
setf.sig f12=r32 // holds a in integer form
setf.sig f7=r33
nop.b 0
} { .mlx
nop.m 0
//movl r2=0x8000000000000000;;
movl r2=0xffffffffffffffff;;
} { .mfi
// get the 2's complement of b
sub r33=r0,r33
fcvt.xf f6=f12
nop.i 0
} { .mfi
nop.m 0
fcvt.xf f7=f7
nop.i 0;;
} { .mfi
nop.m 0
// Step (1)
// y0 = 1 / b in f8
frcpa.s1 f8,p6=f6,f7
nop.i 0;;
} { .mfi
nop.m 0
// Step (2)
// q0 = a * y0 in f10
(p6) fma.s1 f10=f6,f8,f0
nop.i 0
} { .mfi
nop.m 0
// Step (3)
// e0 = 1 - b * y0 in f9
(p6) fnma.s1 f9=f7,f8,f1
nop.i 0;;
} { .mfi
nop.m 0
// Step (4)
// q1 = q0 + e0 * q0 in f10
(p6) fma.s1 f10=f9,f10,f10
nop.i 0
} { .mfi
nop.m 0
// Step (5)
// e1 = e0 * e0 in f11
(p6) fma.s1 f11=f9,f9,f0
nop.i 0;;
} { .mfi
nop.m 0
// Step (6)
// y1 = y0 + e0 * y0 in f8
(p6) fma.s1 f8=f9,f8,f8
nop.i 0;;
} { .mfi
nop.m 0
// Step (7)
// q2 = q1 + e1 * q1 in f9
(p6) fma.s1 f9=f11,f10,f10
nop.i 0;;
} { .mfi
nop.m 0
// Step (8)
// y2 = y1 + e1 * y1 in f8
(p6) fma.s1 f8=f11,f8,f8
nop.i 0;;
} { .mfi
nop.m 0
// Step (9)
// r2 = a - b * q2 in f10
(p6) fnma.s1 f10=f7,f9,f6
nop.i 0;;
} { .mfi
setf.sig f7=r33
// Step (10)
// q3 = q2 + r2 * y2 in f8
(p6) fma.s1 f8=f10,f8,f9
nop.i 0;;
} { .mfi
nop.m 0
// (11) q = trunc(q3)
fcvt.fx.trunc.s1 f8=f8
nop.i 0;;
} { .mfi
nop.m 0
// (12) r = a + (-b) * q
xma.l f8=f8,f7,f12
nop.i 0;;
} { .mib
getf.sig r8=f8
nop.i 0
nop.b 0
}
// 64-BIT SIGNED INTEGER REMAINDER ENDS HERE
{ .mib
nop.m 0
nop.i 0
br.ret.sptk b0;;
}
.endp __moddi3