1 /* Optimized version of the standard strcpy() function.
2 This file is part of the GNU C Library.
3 Copyright (C) 2000, 2001 Free Software Foundation, Inc.
4 Contributed by Dan Pop <Dan.Pop@cern.ch>.
6 The GNU C Library is free software; you can redistribute it and/or
7 modify it under the terms of the GNU Library General Public License as
8 published by the Free Software Foundation; either version 2 of the
9 License, or (at your option) any later version.
11 The GNU C Library is distributed in the hope that it will be useful,
12 but WITHOUT ANY WARRANTY; without even the implied warranty of
13 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
14 Library General Public License for more details.
16 You should have received a copy of the GNU Library General Public
17 License along with the GNU C Library; see the file COPYING.LIB. If not,
18 write to the Free Software Foundation, Inc., 59 Temple Place - Suite 330,
19 Boston, MA 02111-1307, USA. */
27 In this form, it assumes little endian mode. For big endian mode, the
28 the two shifts in .l2 must be inverted:
30 shl value = r[1], sh1 // value = w0 << sh1
31 shr.u tmp = r[0], sh2 // tmp = w1 >> sh2
56 alloc r2 = ar.pfs, 2, 0, 30, 32
62 mov ret0 = in0 // return value = dest
64 mov saved_pr = pr // save the predicate registers
66 mov saved_lc = ar.lc // save the loop counter
68 sub tmp = r0, in0 ;; // tmp = -dest
69 mov dest = in0 // dest
71 and loopcnt = 7, tmp ;; // loopcnt = -dest % 8
72 cmp.eq p6, p0 = loopcnt, r0
73 adds loopcnt = -1, loopcnt // --loopcnt
74 (p6) br.cond.sptk .dest_aligned ;;
76 .l1: // copy -dest % 8 bytes
77 ld1 c = [src], 1 // c = *src++
79 st1 [dest] = c, 1 // *dest++ = c
81 (p6) br.cond.dpnt .restore_and_exit
84 and sh1 = 7, src // sh1 = src % 8
85 mov ar.lc = -1 // "infinite" loop
86 and asrc = -8, src ;; // asrc = src & -OPSIZ -- align src
88 mov pr.rot = 1 << 16 // set rotating predicates
89 cmp.ne p7, p0 = r0, r0 // clear p7
90 shl sh1 = sh1, 3 ;; // sh1 = 8 * (src % 8)
91 sub sh2 = 64, sh1 // sh2 = 64 - sh1
92 cmp.eq p6, p0 = sh1, r0 // is the src aligned?
93 (p6) br.cond.sptk .src_aligned ;;
94 ld8 r[1] = [asrc],8 ;;
98 ld8.s r[0] = [asrc], 8
99 shr.u value = r[1], sh1 ;; // value = w0 >> sh1
100 czx1.r pos = value ;; // do we have an "early" zero
101 cmp.lt p7, p0 = pos, thresh // in w0 >> sh1?
102 (p7) br.cond.dpnt .found0
103 chk.s r[0], .recovery2 // it is safe to do that only
104 .back2: // after the previous test
105 shl tmp = r[0], sh2 // tmp = w1 << sh2
107 or value = value, tmp ;; // value |= tmp
108 czx1.r pos = value ;;
109 cmp.ne p7, p0 = 8, pos
110 (p7) br.cond.dpnt .found0
111 st8 [dest] = value, 8 // store val to dest
115 (p[0]) ld8.s r[0] = [src], 8
116 (p[MEMLAT]) chk.s r[MEMLAT], .recovery3
118 (p[MEMLAT]) mov value = r[MEMLAT]
119 (p[MEMLAT]) czx1.r pos = r[MEMLAT] ;;
120 (p[MEMLAT]) cmp.ne p7, p0 = 8, pos
121 (p7) br.cond.dpnt .found0
122 (p[MEMLAT]) st8 [dest] = r[MEMLAT], 8
127 extr.u c = value, 0, 8 // c = value & 0xff
128 shr.u value = value, 8
133 mov ar.lc = saved_lc // restore the loop counter
134 mov pr = saved_pr, -1 // restore the predicate registers
137 add tmp = -8, asrc ;;
141 add tmp = -(MEMLAT + 1) * 8, src ;;
142 ld8 r[MEMLAT] = [tmp]