summaryrefslogtreecommitdiff
path: root/libc/string/sparc/sparc64
diff options
context:
space:
mode:
Diffstat (limited to 'libc/string/sparc/sparc64')
-rw-r--r--libc/string/sparc/sparc64/memchr.S261
-rw-r--r--libc/string/sparc/sparc64/memcpy.S918
-rw-r--r--libc/string/sparc/sparc64/memset.S313
-rw-r--r--libc/string/sparc/sparc64/sparcv9b/memcpy.S606
-rw-r--r--libc/string/sparc/sparc64/stpcpy.S272
-rw-r--r--libc/string/sparc/sparc64/strcat.S338
-rw-r--r--libc/string/sparc/sparc64/strchr.S480
-rw-r--r--libc/string/sparc/sparc64/strcmp.S278
-rw-r--r--libc/string/sparc/sparc64/strcpy.S243
-rw-r--r--libc/string/sparc/sparc64/strlen.S172
10 files changed, 3881 insertions, 0 deletions
diff --git a/libc/string/sparc/sparc64/memchr.S b/libc/string/sparc/sparc64/memchr.S
new file mode 100644
index 000000000..a10dfbc63
--- /dev/null
+++ b/libc/string/sparc/sparc64/memchr.S
@@ -0,0 +1,261 @@
+/* memchr (str, ch, n) -- Return pointer to first occurrence of CH in STR less
+ than N.
+ For SPARC v9.
+ Copyright (C) 1998, 1999, 2000, 2003 Free Software Foundation, Inc.
+ This file is part of the GNU C Library.
+ Contributed by Jan Vondrak <jvon4518@ss1000.ms.mff.cuni.cz> and
+ Jakub Jelinek <jj@ultra.linux.cz>.
+ This version is developed using the same algorithm as the fast C
+ version which carries the following introduction:
+ Based on strlen implementation by Torbjorn Granlund (tege@sics.se),
+ with help from Dan Sahlin (dan@sics.se) and
+ commentary by Jim Blandy (jimb@ai.mit.edu);
+ adaptation to memchr suggested by Dick Karpinski (dick@cca.ucsf.edu),
+ and implemented by Roland McGrath (roland@ai.mit.edu).
+
+ The GNU C Library is free software; you can redistribute it and/or
+ modify it under the terms of the GNU Lesser General Public
+ License as published by the Free Software Foundation; either
+ version 2.1 of the License, or (at your option) any later version.
+
+ The GNU C Library is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ Lesser General Public License for more details.
+
+ You should have received a copy of the GNU Lesser General Public
+ License along with the GNU C Library; if not, write to the Free
+ Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA
+ 02111-1307 USA. */
+
+#include <asm/asi.h>
+#ifndef XCC
+#define XCC xcc
+#define USE_BPR
+ .register %g2, #scratch
+ .register %g3, #scratch
+#endif
+
+ /* Normally, this uses
+ ((xword - 0x0101010101010101) & 0x8080808080808080) test
+ to find out if any byte in xword could be zero. This is fast, but
+ also gives false alarm for any byte in range 0x81-0xff. It does
+ not matter for correctness, as if this test tells us there could
+ be some zero byte, we check it byte by byte, but if bytes with
+ high bits set are common in the strings, then this will give poor
+ performance. You can #define EIGHTBIT_NOT_RARE and the algorithm
+ will use one tick slower, but more precise test
+ ((xword - 0x0101010101010101) & (~xword) & 0x8080808080808080),
+ which does not give any false alarms (but if some bits are set,
+ one cannot assume from it which bytes are zero and which are not).
+ It is yet to be measured, what is the correct default for glibc
+ in these days for an average user.
+ */
+
+ .text
+ .align 32
+ENTRY(__memchr)
+ and %o1, 0xff, %o1 /* IEU0 Group */
+#ifdef USE_BPR
+ brz,pn %o2, 12f /* CTI+IEU1 */
+#else
+ tst %o2 /* IEU1 */
+ be,pn %XCC, 12f /* CTI */
+#endif
+ sll %o1, 8, %g3 /* IEU0 Group */
+ add %o0, %o2, %o2 /* IEU1 */
+
+ sethi %hi(0x01010101), %g1 /* IEU0 Group */
+ or %g3, %o1, %g3 /* IEU1 */
+ ldub [%o0], %o3 /* Load */
+ sllx %g3, 16, %g5 /* IEU0 Group */
+
+ or %g1, %lo(0x01010101), %g1 /* IEU1 */
+ sllx %g1, 32, %g2 /* IEU0 Group */
+ or %g3, %g5, %g3 /* IEU1 */
+ sllx %g3, 32, %g5 /* IEU0 Group */
+
+ cmp %o3, %o1 /* IEU1 */
+ be,pn %xcc, 13f /* CTI */
+ or %g1, %g2, %g1 /* IEU0 Group */
+ andcc %o0, 7, %g0 /* IEU1 */
+
+ bne,a,pn %icc, 21f /* CTI */
+ add %o0, 1, %o0 /* IEU0 Group */
+ ldx [%o0], %o3 /* Load Group */
+ sllx %g1, 7, %g2 /* IEU0 */
+
+ or %g3, %g5, %g3 /* IEU1 */
+1: add %o0, 8, %o0 /* IEU0 Group */
+ xor %o3, %g3, %o4 /* IEU1 */
+ /* %g1 = 0101010101010101 *
+ * %g2 = 8080088080808080 *
+ * %g3 = c c c c c c c c *
+ * %o3 = value *
+ * %o4 = value XOR c */
+2: cmp %o0, %o2 /* IEU1 Group */
+
+ bg,pn %XCC, 11f /* CTI */
+ ldxa [%o0] ASI_PNF, %o3 /* Load */
+ sub %o4, %g1, %o5 /* IEU0 Group */
+ add %o0, 8, %o0 /* IEU1 */
+#ifdef EIGHTBIT_NOT_RARE
+ andn %o5, %o4, %o5 /* IEU0 Group */
+#endif
+
+ andcc %o5, %g2, %g0 /* IEU1 Group */
+ be,a,pt %xcc, 2b /* CTI */
+ xor %o3, %g3, %o4 /* IEU0 */
+ srlx %o4, 56, %g5 /* IEU0 */
+
+ andcc %g5, 0xff, %g0 /* IEU1 Group */
+ be,pn %icc, 3f /* CTI */
+ srlx %o4, 48, %g5 /* IEU0 */
+ andcc %g5, 0xff, %g0 /* IEU1 Group */
+
+ be,pn %icc, 4f /* CTI */
+ srlx %o4, 40, %g5 /* IEU0 */
+ andcc %g5, 0xff, %g0 /* IEU1 Group */
+ be,pn %icc, 5f /* CTI */
+
+ srlx %o4, 32, %g5 /* IEU0 */
+ andcc %g5, 0xff, %g0 /* IEU1 Group */
+ be,pn %icc, 6f /* CTI */
+ srlx %o4, 24, %g5 /* IEU0 */
+
+ andcc %g5, 0xff, %g0 /* IEU1 Group */
+ be,pn %icc, 7f /* CTI */
+ srlx %o4, 16, %g5 /* IEU0 */
+ andcc %g5, 0xff, %g0 /* IEU1 Group */
+
+ be,pn %icc, 8f /* CTI */
+ srlx %o4, 8, %g5 /* IEU0 */
+ andcc %g5, 0xff, %g0 /* IEU1 Group */
+ be,pn %icc, 9f /* CTI */
+
+ andcc %o4, 0xff, %g0 /* IEU1 Group */
+ bne,pt %icc, 2b /* CTI */
+ xor %o3, %g3, %o4 /* IEU0 */
+ retl /* CTI+IEU1 Group */
+
+ add %o0, -9, %o0 /* IEU0 */
+
+ .align 16
+3: retl /* CTI+IEU1 Group */
+ add %o0, -16, %o0 /* IEU0 */
+4: retl /* CTI+IEU1 Group */
+ add %o0, -15, %o0 /* IEU0 */
+
+5: retl /* CTI+IEU1 Group */
+ add %o0, -14, %o0 /* IEU0 */
+6: retl /* CTI+IEU1 Group */
+ add %o0, -13, %o0 /* IEU0 */
+
+7: retl /* CTI+IEU1 Group */
+ add %o0, -12, %o0 /* IEU0 */
+8: retl /* CTI+IEU1 Group */
+ add %o0, -11, %o0 /* IEU0 */
+
+9: retl /* CTI+IEU1 Group */
+ add %o0, -10, %o0 /* IEU0 */
+11: sub %o4, %g1, %o5 /* IEU0 Group */
+ sub %o0, 8, %o0 /* IEU1 */
+
+ andcc %o5, %g2, %g0 /* IEU1 Group */
+ be,pt %xcc, 12f /* CTI */
+ sub %o2, %o0, %o2 /* IEU0 */
+ tst %o2 /* IEU1 Group */
+
+ be,pn %XCC, 12f /* CTI */
+ srlx %o4, 56, %g5 /* IEU0 */
+ andcc %g5, 0xff, %g0 /* IEU1 Group */
+ be,pn %icc, 13f /* CTI */
+
+ cmp %o2, 1 /* IEU0 */
+ be,pn %XCC, 12f /* CTI Group */
+ srlx %o4, 48, %g5 /* IEU0 */
+ andcc %g5, 0xff, %g0 /* IEU1 Group */
+
+ be,pn %icc, 14f /* CTI */
+ cmp %o2, 2 /* IEU1 Group */
+ be,pn %XCC, 12f /* CTI */
+ srlx %o4, 40, %g5 /* IEU0 */
+
+ andcc %g5, 0xff, %g0 /* IEU1 Group */
+ be,pn %icc, 15f /* CTI */
+ cmp %o2, 3 /* IEU1 Group */
+ be,pn %XCC, 12f /* CTI */
+
+ srlx %o4, 32, %g5 /* IEU0 */
+ andcc %g5, 0xff, %g0 /* IEU1 Group */
+ be,pn %icc, 16f /* CTI */
+ cmp %o2, 4 /* IEU1 Group */
+
+ be,pn %XCC, 12f /* CTI */
+ srlx %o4, 24, %g5 /* IEU0 */
+ andcc %g5, 0xff, %g0 /* IEU1 Group */
+ be,pn %icc, 17f /* CTI */
+
+ cmp %o2, 5 /* IEU1 Group */
+ be,pn %XCC, 12f /* CTI */
+ srlx %o4, 16, %g5 /* IEU0 */
+ andcc %g5, 0xff, %g0 /* IEU1 Group */
+
+ be,pn %icc, 18f /* CTI */
+ cmp %o2, 6 /* IEU1 Group */
+ be,pn %XCC, 12f /* CTI */
+ srlx %o4, 8, %g5 /* IEU0 */
+
+ andcc %g5, 0xff, %g0 /* IEU1 Group */
+ be,pn %icc, 19f /* CTI */
+ nop /* IEU0 */
+12: retl /* CTI+IEU1 Group */
+
+ clr %o0 /* IEU0 */
+ nop /* Stub */
+13: retl /* CTI+IEU1 Group */
+ nop /* IEU0 */
+
+14: retl /* CTI+IEU1 Group */
+ add %o0, 1, %o0 /* IEU0 */
+15: retl /* CTI+IEU1 Group */
+ add %o0, 2, %o0 /* IEU0 */
+
+16: retl /* CTI+IEU1 Group */
+ add %o0, 3, %o0 /* IEU0 */
+17: retl /* CTI+IEU1 Group */
+ add %o0, 4, %o0 /* IEU0 */
+
+18: retl /* CTI+IEU1 Group */
+ add %o0, 5, %o0 /* IEU0 */
+19: retl /* CTI+IEU1 Group */
+ add %o0, 6, %o0 /* IEU0 */
+
+21: cmp %o0, %o2 /* IEU1 */
+ be,pn %XCC, 12b /* CTI */
+ sllx %g1, 7, %g2 /* IEU0 Group */
+ ldub [%o0], %o3 /* Load */
+
+ or %g3, %g5, %g3 /* IEU1 */
+22: andcc %o0, 7, %g0 /* IEU1 Group */
+ be,a,pn %icc, 1b /* CTI */
+ ldx [%o0], %o3 /* Load */
+
+ cmp %o3, %o1 /* IEU1 Group */
+ be,pn %xcc, 23f /* CTI */
+ add %o0, 1, %o0 /* IEU0 */
+ cmp %o0, %o2 /* IEU1 Group */
+
+ bne,a,pt %XCC, 22b /* CTI */
+ ldub [%o0], %o3 /* Load */
+ retl /* CTI+IEU1 Group */
+ clr %o0 /* IEU0 */
+
+23: retl /* CTI+IEU1 Group */
+ add %o0, -1, %o0 /* IEU0 */
+END(__memchr)
+
+weak_alias (__memchr, memchr)
+#if !__BOUNDED_POINTERS__
+weak_alias (__memchr, __ubp_memchr)
+#endif
diff --git a/libc/string/sparc/sparc64/memcpy.S b/libc/string/sparc/sparc64/memcpy.S
new file mode 100644
index 000000000..660fae6b2
--- /dev/null
+++ b/libc/string/sparc/sparc64/memcpy.S
@@ -0,0 +1,918 @@
+/* Copy SIZE bytes from SRC to DEST.
+ For UltraSPARC.
+ Copyright (C) 1996, 97, 98, 99, 2003 Free Software Foundation, Inc.
+ This file is part of the GNU C Library.
+ Contributed by David S. Miller (davem@caip.rutgers.edu) and
+ Jakub Jelinek (jakub@redhat.com).
+
+ The GNU C Library is free software; you can redistribute it and/or
+ modify it under the terms of the GNU Lesser General Public
+ License as published by the Free Software Foundation; either
+ version 2.1 of the License, or (at your option) any later version.
+
+ The GNU C Library is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ Lesser General Public License for more details.
+
+ You should have received a copy of the GNU Lesser General Public
+ License along with the GNU C Library; if not, write to the Free
+ Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA
+ 02111-1307 USA. */
+
+#include <asm/asi.h>
+#ifndef XCC
+#define USE_BPR
+ .register %g2, #scratch
+ .register %g3, #scratch
+ .register %g6, #scratch
+#define XCC xcc
+#endif
+#define FPRS_FEF 4
+
+#define FREG_FROB(f1, f2, f3, f4, f5, f6, f7, f8, f9) \
+ faligndata %f1, %f2, %f48; \
+ faligndata %f2, %f3, %f50; \
+ faligndata %f3, %f4, %f52; \
+ faligndata %f4, %f5, %f54; \
+ faligndata %f5, %f6, %f56; \
+ faligndata %f6, %f7, %f58; \
+ faligndata %f7, %f8, %f60; \
+ faligndata %f8, %f9, %f62;
+
+#define MAIN_LOOP_CHUNK(src, dest, fdest, fsrc, len, jmptgt) \
+ ldda [%src] %asi, %fdest; \
+ add %src, 0x40, %src; \
+ add %dest, 0x40, %dest; \
+ subcc %len, 0x40, %len; \
+ be,pn %xcc, jmptgt; \
+ stda %fsrc, [%dest - 0x40] %asi;
+
+#define LOOP_CHUNK1(src, dest, len, branch_dest) \
+ MAIN_LOOP_CHUNK(src, dest, f0, f48, len, branch_dest)
+#define LOOP_CHUNK2(src, dest, len, branch_dest) \
+ MAIN_LOOP_CHUNK(src, dest, f16, f48, len, branch_dest)
+#define LOOP_CHUNK3(src, dest, len, branch_dest) \
+ MAIN_LOOP_CHUNK(src, dest, f32, f48, len, branch_dest)
+
+#define STORE_SYNC(dest, fsrc) \
+ stda %fsrc, [%dest] %asi; \
+ add %dest, 0x40, %dest;
+
+#define STORE_JUMP(dest, fsrc, target) \
+ stda %fsrc, [%dest] %asi; \
+ add %dest, 0x40, %dest; \
+ ba,pt %xcc, target;
+
+#define VISLOOP_PAD nop; nop; nop; nop; \
+ nop; nop; nop; nop; \
+ nop; nop; nop; nop; \
+ nop; nop; nop;
+
+#define FINISH_VISCHUNK(dest, f0, f1, left) \
+ subcc %left, 8, %left; \
+ bl,pn %xcc, 205f; \
+ faligndata %f0, %f1, %f48; \
+ std %f48, [%dest]; \
+ add %dest, 8, %dest;
+
+#define UNEVEN_VISCHUNK(dest, f0, f1, left) \
+ subcc %left, 8, %left; \
+ bl,pn %xcc, 205f; \
+ fsrc1 %f0, %f1; \
+ ba,a,pt %xcc, 204f;
+
+ /* Macros for non-VIS memcpy code. */
+#define MOVE_BIGCHUNK(src, dst, offset, t0, t1, t2, t3) \
+ ldx [%src + offset + 0x00], %t0; \
+ ldx [%src + offset + 0x08], %t1; \
+ ldx [%src + offset + 0x10], %t2; \
+ ldx [%src + offset + 0x18], %t3; \
+ stw %t0, [%dst + offset + 0x04]; \
+ srlx %t0, 32, %t0; \
+ stw %t0, [%dst + offset + 0x00]; \
+ stw %t1, [%dst + offset + 0x0c]; \
+ srlx %t1, 32, %t1; \
+ stw %t1, [%dst + offset + 0x08]; \
+ stw %t2, [%dst + offset + 0x14]; \
+ srlx %t2, 32, %t2; \
+ stw %t2, [%dst + offset + 0x10]; \
+ stw %t3, [%dst + offset + 0x1c]; \
+ srlx %t3, 32, %t3; \
+ stw %t3, [%dst + offset + 0x18];
+
+#define MOVE_BIGALIGNCHUNK(src, dst, offset, t0, t1, t2, t3) \
+ ldx [%src + offset + 0x00], %t0; \
+ ldx [%src + offset + 0x08], %t1; \
+ ldx [%src + offset + 0x10], %t2; \
+ ldx [%src + offset + 0x18], %t3; \
+ stx %t0, [%dst + offset + 0x00]; \
+ stx %t1, [%dst + offset + 0x08]; \
+ stx %t2, [%dst + offset + 0x10]; \
+ stx %t3, [%dst + offset + 0x18]; \
+ ldx [%src + offset + 0x20], %t0; \
+ ldx [%src + offset + 0x28], %t1; \
+ ldx [%src + offset + 0x30], %t2; \
+ ldx [%src + offset + 0x38], %t3; \
+ stx %t0, [%dst + offset + 0x20]; \
+ stx %t1, [%dst + offset + 0x28]; \
+ stx %t2, [%dst + offset + 0x30]; \
+ stx %t3, [%dst + offset + 0x38];
+
+#define MOVE_LASTCHUNK(src, dst, offset, t0, t1, t2, t3) \
+ ldx [%src - offset - 0x10], %t0; \
+ ldx [%src - offset - 0x08], %t1; \
+ stw %t0, [%dst - offset - 0x0c]; \
+ srlx %t0, 32, %t2; \
+ stw %t2, [%dst - offset - 0x10]; \
+ stw %t1, [%dst - offset - 0x04]; \
+ srlx %t1, 32, %t3; \
+ stw %t3, [%dst - offset - 0x08];
+
+#define MOVE_LASTALIGNCHUNK(src, dst, offset, t0, t1) \
+ ldx [%src - offset - 0x10], %t0; \
+ ldx [%src - offset - 0x08], %t1; \
+ stx %t0, [%dst - offset - 0x10]; \
+ stx %t1, [%dst - offset - 0x08];
+
+ /* Macros for non-VIS memmove code. */
+#define RMOVE_BIGCHUNK(src, dst, offset, t0, t1, t2, t3) \
+ ldx [%src - offset - 0x20], %t0; \
+ ldx [%src - offset - 0x18], %t1; \
+ ldx [%src - offset - 0x10], %t2; \
+ ldx [%src - offset - 0x08], %t3; \
+ stw %t0, [%dst - offset - 0x1c]; \
+ srlx %t0, 32, %t0; \
+ stw %t0, [%dst - offset - 0x20]; \
+ stw %t1, [%dst - offset - 0x14]; \
+ srlx %t1, 32, %t1; \
+ stw %t1, [%dst - offset - 0x18]; \
+ stw %t2, [%dst - offset - 0x0c]; \
+ srlx %t2, 32, %t2; \
+ stw %t2, [%dst - offset - 0x10]; \
+ stw %t3, [%dst - offset - 0x04]; \
+ srlx %t3, 32, %t3; \
+ stw %t3, [%dst - offset - 0x08];
+
+#define RMOVE_BIGALIGNCHUNK(src, dst, offset, t0, t1, t2, t3) \
+ ldx [%src - offset - 0x20], %t0; \
+ ldx [%src - offset - 0x18], %t1; \
+ ldx [%src - offset - 0x10], %t2; \
+ ldx [%src - offset - 0x08], %t3; \
+ stx %t0, [%dst - offset - 0x20]; \
+ stx %t1, [%dst - offset - 0x18]; \
+ stx %t2, [%dst - offset - 0x10]; \
+ stx %t3, [%dst - offset - 0x08]; \
+ ldx [%src - offset - 0x40], %t0; \
+ ldx [%src - offset - 0x38], %t1; \
+ ldx [%src - offset - 0x30], %t2; \
+ ldx [%src - offset - 0x28], %t3; \
+ stx %t0, [%dst - offset - 0x40]; \
+ stx %t1, [%dst - offset - 0x38]; \
+ stx %t2, [%dst - offset - 0x30]; \
+ stx %t3, [%dst - offset - 0x28];
+
+#define RMOVE_LASTCHUNK(src, dst, offset, t0, t1, t2, t3) \
+ ldx [%src + offset + 0x00], %t0; \
+ ldx [%src + offset + 0x08], %t1; \
+ stw %t0, [%dst + offset + 0x04]; \
+ srlx %t0, 32, %t2; \
+ stw %t2, [%dst + offset + 0x00]; \
+ stw %t1, [%dst + offset + 0x0c]; \
+ srlx %t1, 32, %t3; \
+ stw %t3, [%dst + offset + 0x08];
+
+#define RMOVE_LASTALIGNCHUNK(src, dst, offset, t0, t1) \
+ ldx [%src + offset + 0x00], %t0; \
+ ldx [%src + offset + 0x08], %t1; \
+ stx %t0, [%dst + offset + 0x00]; \
+ stx %t1, [%dst + offset + 0x08];
+
+ .text
+ .align 32
+
+ENTRY(bcopy)
+ sub %o1, %o0, %o4 /* IEU0 Group */
+ mov %o0, %g3 /* IEU1 */
+ cmp %o4, %o2 /* IEU1 Group */
+ mov %o1, %o0 /* IEU0 */
+ bgeu,pt %XCC, 210f /* CTI */
+ mov %g3, %o1 /* IEU0 Group */
+#ifndef USE_BPR
+ srl %o2, 0, %o2 /* IEU1 */
+#endif
+ brnz,pn %o2, 220f /* CTI Group */
+ add %o0, %o2, %o0 /* IEU0 */
+ retl
+ nop
+END(bcopy)
+
+ .align 32
+200: be,pt %xcc, 201f /* CTI */
+ andcc %o0, 0x38, %g5 /* IEU1 Group */
+ mov 8, %g1 /* IEU0 */
+ sub %g1, %g2, %g2 /* IEU0 Group */
+ andcc %o0, 1, %g0 /* IEU1 */
+ be,pt %icc, 2f /* CTI */
+ sub %o2, %g2, %o2 /* IEU0 Group */
+1: ldub [%o1], %o5 /* Load Group */
+ add %o1, 1, %o1 /* IEU0 */
+ add %o0, 1, %o0 /* IEU1 */
+ subcc %g2, 1, %g2 /* IEU1 Group */
+ be,pn %xcc, 3f /* CTI */
+ stb %o5, [%o0 - 1] /* Store */
+2: ldub [%o1], %o5 /* Load Group */
+ add %o0, 2, %o0 /* IEU0 */
+ ldub [%o1 + 1], %g3 /* Load Group */
+ subcc %g2, 2, %g2 /* IEU1 Group */
+ stb %o5, [%o0 - 2] /* Store */
+ add %o1, 2, %o1 /* IEU0 */
+ bne,pt %xcc, 2b /* CTI Group */
+ stb %g3, [%o0 - 1] /* Store */
+3: andcc %o0, 0x38, %g5 /* IEU1 Group */
+201: be,pt %icc, 202f /* CTI */
+ mov 64, %g1 /* IEU0 */
+ fmovd %f0, %f2 /* FPU */
+ sub %g1, %g5, %g5 /* IEU0 Group */
+ alignaddr %o1, %g0, %g1 /* GRU Group */
+ ldd [%g1], %f4 /* Load Group */
+ sub %o2, %g5, %o2 /* IEU0 */
+1: ldd [%g1 + 0x8], %f6 /* Load Group */
+ add %g1, 0x8, %g1 /* IEU0 Group */
+ subcc %g5, 8, %g5 /* IEU1 */
+ faligndata %f4, %f6, %f0 /* GRU Group */
+ std %f0, [%o0] /* Store */
+ add %o1, 8, %o1 /* IEU0 Group */
+ be,pn %xcc, 202f /* CTI */
+ add %o0, 8, %o0 /* IEU1 */
+ ldd [%g1 + 0x8], %f4 /* Load Group */
+ add %g1, 8, %g1 /* IEU0 */
+ subcc %g5, 8, %g5 /* IEU1 */
+ faligndata %f6, %f4, %f0 /* GRU Group */
+ std %f0, [%o0] /* Store */
+ add %o1, 8, %o1 /* IEU0 */
+ bne,pt %xcc, 1b /* CTI Group */
+ add %o0, 8, %o0 /* IEU0 */
+202: membar #LoadStore | #StoreStore | #StoreLoad /* LSU Group */
+ wr %g0, ASI_BLK_P, %asi /* LSU Group */
+ subcc %o2, 0x40, %g6 /* IEU1 Group */
+ mov %o1, %g1 /* IEU0 */
+ andncc %g6, (0x40 - 1), %g6 /* IEU1 Group */
+ srl %g1, 3, %g2 /* IEU0 */
+ sub %o2, %g6, %g3 /* IEU0 Group */
+ andn %o1, (0x40 - 1), %o1 /* IEU1 */
+ and %g2, 7, %g2 /* IEU0 Group */
+ andncc %g3, 0x7, %g3 /* IEU1 */
+ fmovd %f0, %f2 /* FPU */
+ sub %g3, 0x10, %g3 /* IEU0 Group */
+ sub %o2, %g6, %o2 /* IEU1 */
+ alignaddr %g1, %g0, %g0 /* GRU Group */
+ add %g1, %g6, %g1 /* IEU0 Group */
+ subcc %o2, %g3, %o2 /* IEU1 */
+ ldda [%o1 + 0x00] %asi, %f0 /* LSU Group */
+ add %g1, %g3, %g1 /* IEU0 */
+ ldda [%o1 + 0x40] %asi, %f16 /* LSU Group */
+ sub %g6, 0x80, %g6 /* IEU0 */
+ ldda [%o1 + 0x80] %asi, %f32 /* LSU Group */
+ /* Clk1 Group 8-( */
+ /* Clk2 Group 8-( */
+ /* Clk3 Group 8-( */
+ /* Clk4 Group 8-( */
+203: rd %pc, %g5 /* PDU Group 8-( */
+ addcc %g5, %lo(300f - 203b), %g5 /* IEU1 Group */
+ sll %g2, 9, %g2 /* IEU0 */
+ jmpl %g5 + %g2, %g0 /* CTI Group brk forced*/
+ addcc %o1, 0xc0, %o1 /* IEU1 Group */
+
+ .align 512 /* OK, here comes the fun part... */
+300: FREG_FROB(f0, f2, f4, f6, f8, f10,f12,f14,f16) LOOP_CHUNK1(o1, o0, g6, 301f)
+ FREG_FROB(f16,f18,f20,f22,f24,f26,f28,f30,f32) LOOP_CHUNK2(o1, o0, g6, 302f)
+ FREG_FROB(f32,f34,f36,f38,f40,f42,f44,f46,f0) LOOP_CHUNK3(o1, o0, g6, 303f)
+ b,pt %xcc, 300b+4; faligndata %f0, %f2, %f48
+301: FREG_FROB(f16,f18,f20,f22,f24,f26,f28,f30,f32) STORE_SYNC(o0, f48) membar #Sync
+ FREG_FROB(f32,f34,f36,f38,f40,f42,f44,f46,f0) STORE_JUMP(o0, f48, 400f) membar #Sync
+302: FREG_FROB(f32,f34,f36,f38,f40,f42,f44,f46,f0) STORE_SYNC(o0, f48) membar #Sync
+ FREG_FROB(f0, f2, f4, f6, f8, f10,f12,f14,f16) STORE_JUMP(o0, f48, 416f) membar #Sync
+303: FREG_FROB(f0, f2, f4, f6, f8, f10,f12,f14,f16) STORE_SYNC(o0, f48) membar #Sync
+ FREG_FROB(f16,f18,f20,f22,f24,f26,f28,f30,f32) STORE_JUMP(o0, f48, 432f) membar #Sync
+ VISLOOP_PAD
+310: FREG_FROB(f2, f4, f6, f8, f10,f12,f14,f16,f18) LOOP_CHUNK1(o1, o0, g6, 311f)
+ FREG_FROB(f18,f20,f22,f24,f26,f28,f30,f32,f34) LOOP_CHUNK2(o1, o0, g6, 312f)
+ FREG_FROB(f34,f36,f38,f40,f42,f44,f46,f0, f2) LOOP_CHUNK3(o1, o0, g6, 313f)
+ b,pt %xcc, 310b+4; faligndata %f2, %f4, %f48
+311: FREG_FROB(f18,f20,f22,f24,f26,f28,f30,f32,f34) STORE_SYNC(o0, f48) membar #Sync
+ FREG_FROB(f34,f36,f38,f40,f42,f44,f46,f0, f2) STORE_JUMP(o0, f48, 402f) membar #Sync
+312: FREG_FROB(f34,f36,f38,f40,f42,f44,f46,f0, f2) STORE_SYNC(o0, f48) membar #Sync
+ FREG_FROB(f2, f4, f6, f8, f10,f12,f14,f16,f18) STORE_JUMP(o0, f48, 418f) membar #Sync
+313: FREG_FROB(f2, f4, f6, f8, f10,f12,f14,f16,f18) STORE_SYNC(o0, f48) membar #Sync
+ FREG_FROB(f18,f20,f22,f24,f26,f28,f30,f32,f34) STORE_JUMP(o0, f48, 434f) membar #Sync
+ VISLOOP_PAD
+320: FREG_FROB(f4, f6, f8, f10,f12,f14,f16,f18,f20) LOOP_CHUNK1(o1, o0, g6, 321f)
+ FREG_FROB(f20,f22,f24,f26,f28,f30,f32,f34,f36) LOOP_CHUNK2(o1, o0, g6, 322f)
+ FREG_FROB(f36,f38,f40,f42,f44,f46,f0, f2, f4) LOOP_CHUNK3(o1, o0, g6, 323f)
+ b,pt %xcc, 320b+4; faligndata %f4, %f6, %f48
+321: FREG_FROB(f20,f22,f24,f26,f28,f30,f32,f34,f36) STORE_SYNC(o0, f48) membar #Sync
+ FREG_FROB(f36,f38,f40,f42,f44,f46,f0, f2, f4) STORE_JUMP(o0, f48, 404f) membar #Sync
+322: FREG_FROB(f36,f38,f40,f42,f44,f46,f0, f2, f4) STORE_SYNC(o0, f48) membar #Sync
+ FREG_FROB(f4, f6, f8, f10,f12,f14,f16,f18,f20) STORE_JUMP(o0, f48, 420f) membar #Sync
+323: FREG_FROB(f4, f6, f8, f10,f12,f14,f16,f18,f20) STORE_SYNC(o0, f48) membar #Sync
+ FREG_FROB(f20,f22,f24,f26,f28,f30,f32,f34,f36) STORE_JUMP(o0, f48, 436f) membar #Sync
+ VISLOOP_PAD
+330: FREG_FROB(f6, f8, f10,f12,f14,f16,f18,f20,f22) LOOP_CHUNK1(o1, o0, g6, 331f)
+ FREG_FROB(f22,f24,f26,f28,f30,f32,f34,f36,f38) LOOP_CHUNK2(o1, o0, g6, 332f)
+ FREG_FROB(f38,f40,f42,f44,f46,f0, f2, f4, f6) LOOP_CHUNK3(o1, o0, g6, 333f)
+ b,pt %xcc, 330b+4; faligndata %f6, %f8, %f48
+331: FREG_FROB(f22,f24,f26,f28,f30,f32,f34,f36,f38) STORE_SYNC(o0, f48) membar #Sync
+ FREG_FROB(f38,f40,f42,f44,f46,f0, f2, f4, f6) STORE_JUMP(o0, f48, 406f) membar #Sync
+332: FREG_FROB(f38,f40,f42,f44,f46,f0, f2, f4, f6) STORE_SYNC(o0, f48) membar #Sync
+ FREG_FROB(f6, f8, f10,f12,f14,f16,f18,f20,f22) STORE_JUMP(o0, f48, 422f) membar #Sync
+333: FREG_FROB(f6, f8, f10,f12,f14,f16,f18,f20,f22) STORE_SYNC(o0, f48) membar #Sync
+ FREG_FROB(f22,f24,f26,f28,f30,f32,f34,f36,f38) STORE_JUMP(o0, f48, 438f) membar #Sync
+ VISLOOP_PAD
+340: FREG_FROB(f8, f10,f12,f14,f16,f18,f20,f22,f24) LOOP_CHUNK1(o1, o0, g6, 341f)
+ FREG_FROB(f24,f26,f28,f30,f32,f34,f36,f38,f40) LOOP_CHUNK2(o1, o0, g6, 342f)
+ FREG_FROB(f40,f42,f44,f46,f0, f2, f4, f6, f8) LOOP_CHUNK3(o1, o0, g6, 343f)
+ b,pt %xcc, 340b+4; faligndata %f8, %f10, %f48
+341: FREG_FROB(f24,f26,f28,f30,f32,f34,f36,f38,f40) STORE_SYNC(o0, f48) membar #Sync
+ FREG_FROB(f40,f42,f44,f46,f0, f2, f4, f6, f8) STORE_JUMP(o0, f48, 408f) membar #Sync
+342: FREG_FROB(f40,f42,f44,f46,f0, f2, f4, f6, f8) STORE_SYNC(o0, f48) membar #Sync
+ FREG_FROB(f8, f10,f12,f14,f16,f18,f20,f22,f24) STORE_JUMP(o0, f48, 424f) membar #Sync
+343: FREG_FROB(f8, f10,f12,f14,f16,f18,f20,f22,f24) STORE_SYNC(o0, f48) membar #Sync
+ FREG_FROB(f24,f26,f28,f30,f32,f34,f36,f38,f40) STORE_JUMP(o0, f48, 440f) membar #Sync
+ VISLOOP_PAD
+350: FREG_FROB(f10,f12,f14,f16,f18,f20,f22,f24,f26) LOOP_CHUNK1(o1, o0, g6, 351f)
+ FREG_FROB(f26,f28,f30,f32,f34,f36,f38,f40,f42) LOOP_CHUNK2(o1, o0, g6, 352f)
+ FREG_FROB(f42,f44,f46,f0, f2, f4, f6, f8, f10) LOOP_CHUNK3(o1, o0, g6, 353f)
+ b,pt %xcc, 350b+4; faligndata %f10, %f12, %f48
+351: FREG_FROB(f26,f28,f30,f32,f34,f36,f38,f40,f42) STORE_SYNC(o0, f48) membar #Sync
+ FREG_FROB(f42,f44,f46,f0, f2, f4, f6, f8, f10) STORE_JUMP(o0, f48, 410f) membar #Sync
+352: FREG_FROB(f42,f44,f46,f0, f2, f4, f6, f8, f10) STORE_SYNC(o0, f48) membar #Sync
+ FREG_FROB(f10,f12,f14,f16,f18,f20,f22,f24,f26) STORE_JUMP(o0, f48, 426f) membar #Sync
+353: FREG_FROB(f10,f12,f14,f16,f18,f20,f22,f24,f26) STORE_SYNC(o0, f48) membar #Sync
+ FREG_FROB(f26,f28,f30,f32,f34,f36,f38,f40,f42) STORE_JUMP(o0, f48, 442f) membar #Sync
+ VISLOOP_PAD
+360: FREG_FROB(f12,f14,f16,f18,f20,f22,f24,f26,f28) LOOP_CHUNK1(o1, o0, g6, 361f)
+ FREG_FROB(f28,f30,f32,f34,f36,f38,f40,f42,f44) LOOP_CHUNK2(o1, o0, g6, 362f)
+ FREG_FROB(f44,f46,f0, f2, f4, f6, f8, f10,f12) LOOP_CHUNK3(o1, o0, g6, 363f)
+ b,pt %xcc, 360b+4; faligndata %f12, %f14, %f48
+361: FREG_FROB(f28,f30,f32,f34,f36,f38,f40,f42,f44) STORE_SYNC(o0, f48) membar #Sync
+ FREG_FROB(f44,f46,f0, f2, f4, f6, f8, f10,f12) STORE_JUMP(o0, f48, 412f) membar #Sync
+362: FREG_FROB(f44,f46,f0, f2, f4, f6, f8, f10,f12) STORE_SYNC(o0, f48) membar #Sync
+ FREG_FROB(f12,f14,f16,f18,f20,f22,f24,f26,f28) STORE_JUMP(o0, f48, 428f) membar #Sync
+363: FREG_FROB(f12,f14,f16,f18,f20,f22,f24,f26,f28) STORE_SYNC(o0, f48) membar #Sync
+ FREG_FROB(f28,f30,f32,f34,f36,f38,f40,f42,f44) STORE_JUMP(o0, f48, 444f) membar #Sync
+ VISLOOP_PAD
+370: FREG_FROB(f14,f16,f18,f20,f22,f24,f26,f28,f30) LOOP_CHUNK1(o1, o0, g6, 371f)
+ FREG_FROB(f30,f32,f34,f36,f38,f40,f42,f44,f46) LOOP_CHUNK2(o1, o0, g6, 372f)
+ FREG_FROB(f46,f0, f2, f4, f6, f8, f10,f12,f14) LOOP_CHUNK3(o1, o0, g6, 373f)
+ b,pt %xcc, 370b+4; faligndata %f14, %f16, %f48
+371: FREG_FROB(f30,f32,f34,f36,f38,f40,f42,f44,f46) STORE_SYNC(o0, f48) membar #Sync
+ FREG_FROB(f46,f0, f2, f4, f6, f8, f10,f12,f14) STORE_JUMP(o0, f48, 414f) membar #Sync
+372: FREG_FROB(f46,f0, f2, f4, f6, f8, f10,f12,f14) STORE_SYNC(o0, f48) membar #Sync
+ FREG_FROB(f14,f16,f18,f20,f22,f24,f26,f28,f30) STORE_JUMP(o0, f48, 430f) membar #Sync
+373: FREG_FROB(f14,f16,f18,f20,f22,f24,f26,f28,f30) STORE_SYNC(o0, f48) membar #Sync
+ FREG_FROB(f30,f32,f34,f36,f38,f40,f42,f44,f46) STORE_JUMP(o0, f48, 446f) membar #Sync
+ VISLOOP_PAD
+400: FINISH_VISCHUNK(o0, f0, f2, g3)
+402: FINISH_VISCHUNK(o0, f2, f4, g3)
+404: FINISH_VISCHUNK(o0, f4, f6, g3)
+406: FINISH_VISCHUNK(o0, f6, f8, g3)
+408: FINISH_VISCHUNK(o0, f8, f10, g3)
+410: FINISH_VISCHUNK(o0, f10, f12, g3)
+412: FINISH_VISCHUNK(o0, f12, f14, g3)
+414: UNEVEN_VISCHUNK(o0, f14, f0, g3)
+416: FINISH_VISCHUNK(o0, f16, f18, g3)
+418: FINISH_VISCHUNK(o0, f18, f20, g3)
+420: FINISH_VISCHUNK(o0, f20, f22, g3)
+422: FINISH_VISCHUNK(o0, f22, f24, g3)
+424: FINISH_VISCHUNK(o0, f24, f26, g3)
+426: FINISH_VISCHUNK(o0, f26, f28, g3)
+428: FINISH_VISCHUNK(o0, f28, f30, g3)
+430: UNEVEN_VISCHUNK(o0, f30, f0, g3)
+432: FINISH_VISCHUNK(o0, f32, f34, g3)
+434: FINISH_VISCHUNK(o0, f34, f36, g3)
+436: FINISH_VISCHUNK(o0, f36, f38, g3)
+438: FINISH_VISCHUNK(o0, f38, f40, g3)
+440: FINISH_VISCHUNK(o0, f40, f42, g3)
+442: FINISH_VISCHUNK(o0, f42, f44, g3)
+444: FINISH_VISCHUNK(o0, f44, f46, g3)
+446: UNEVEN_VISCHUNK(o0, f46, f0, g3)
+204: ldd [%o1], %f2 /* Load Group */
+ add %o1, 8, %o1 /* IEU0 */
+ subcc %g3, 8, %g3 /* IEU1 */
+ faligndata %f0, %f2, %f8 /* GRU Group */
+ std %f8, [%o0] /* Store */
+ bl,pn %xcc, 205f /* CTI */
+ add %o0, 8, %o0 /* IEU0 Group */
+ ldd [%o1], %f0 /* Load Group */
+ add %o1, 8, %o1 /* IEU0 */
+ subcc %g3, 8, %g3 /* IEU1 */
+ faligndata %f2, %f0, %f8 /* GRU Group */
+ std %f8, [%o0] /* Store */
+ bge,pt %xcc, 204b /* CTI */
+ add %o0, 8, %o0 /* IEU0 Group */
+205: brz,pt %o2, 207f /* CTI Group */
+ mov %g1, %o1 /* IEU0 */
+206: ldub [%o1], %g5 /* LOAD */
+ add %o1, 1, %o1 /* IEU0 */
+ add %o0, 1, %o0 /* IEU1 */
+ subcc %o2, 1, %o2 /* IEU1 */
+ bne,pt %xcc, 206b /* CTI */
+ stb %g5, [%o0 - 1] /* Store Group */
+207: membar #StoreLoad | #StoreStore /* LSU Group */
+ wr %g0, FPRS_FEF, %fprs
+ retl
+ mov %g4, %o0
+
+208: andcc %o2, 1, %g0 /* IEU1 Group */
+ be,pt %icc, 2f+4 /* CTI */
+1: ldub [%o1], %g5 /* LOAD Group */
+ add %o1, 1, %o1 /* IEU0 */
+ add %o0, 1, %o0 /* IEU1 */
+ subcc %o2, 1, %o2 /* IEU1 Group */
+ be,pn %xcc, 209f /* CTI */
+ stb %g5, [%o0 - 1] /* Store */
+2: ldub [%o1], %g5 /* LOAD Group */
+ add %o0, 2, %o0 /* IEU0 */
+ ldub [%o1 + 1], %o5 /* LOAD Group */
+ add %o1, 2, %o1 /* IEU0 */
+ subcc %o2, 2, %o2 /* IEU1 Group */
+ stb %g5, [%o0 - 2] /* Store */
+ bne,pt %xcc, 2b /* CTI */
+ stb %o5, [%o0 - 1] /* Store */
+209: retl
+ mov %g4, %o0
+
+#ifdef USE_BPR
+
+ /* void *__align_cpy_4(void *dest, void *src, size_t n)
+ * SPARC v9 SYSV ABI
+ * Like memcpy, but results are undefined if (!n || ((dest | src | n) & 3))
+ */
+
+ .align 32
+ENTRY(__align_cpy_4)
+ mov %o0, %g4 /* IEU0 Group */
+ cmp %o2, 15 /* IEU1 */
+ bleu,pn %xcc, 208b /* CTI */
+ cmp %o2, (64 * 6) /* IEU1 Group */
+ bgeu,pn %xcc, 200b /* CTI */
+ andcc %o0, 7, %g2 /* IEU1 Group */
+ ba,pt %xcc, 216f /* CTI */
+ andcc %o1, 4, %g0 /* IEU1 Group */
+END(__align_cpy_4)
+
+ /* void *__align_cpy_8(void *dest, void *src, size_t n)
+ * SPARC v9 SYSV ABI
+ * Like memcpy, but results are undefined if (!n || ((dest | src | n) & 7))
+ */
+
+ .align 32
+ENTRY(__align_cpy_8)
+ mov %o0, %g4 /* IEU0 Group */
+ cmp %o2, 15 /* IEU1 */
+ bleu,pn %xcc, 208b /* CTI */
+ cmp %o2, (64 * 6) /* IEU1 Group */
+ bgeu,pn %xcc, 201b /* CTI */
+ andcc %o0, 0x38, %g5 /* IEU1 Group */
+ andcc %o2, -128, %g6 /* IEU1 Group */
+ bne,a,pt %xcc, 82f + 4 /* CTI */
+ ldx [%o1], %g1 /* Load */
+ ba,pt %xcc, 41f /* CTI Group */
+ andcc %o2, 0x70, %g6 /* IEU1 */
+END(__align_cpy_8)
+
+ /* void *__align_cpy_16(void *dest, void *src, size_t n)
+ * SPARC v9 SYSV ABI
+ * Like memcpy, but results are undefined if (!n || ((dest | src | n) & 15))
+ */
+
+ .align 32
+ENTRY(__align_cpy_16)
+ mov %o0, %g4 /* IEU0 Group */
+ cmp %o2, (64 * 6) /* IEU1 */
+ bgeu,pn %xcc, 201b /* CTI */
+ andcc %o0, 0x38, %g5 /* IEU1 Group */
+ andcc %o2, -128, %g6 /* IEU1 Group */
+ bne,a,pt %xcc, 82f + 4 /* CTI */
+ ldx [%o1], %g1 /* Load */
+ ba,pt %xcc, 41f /* CTI Group */
+ andcc %o2, 0x70, %g6 /* IEU1 */
+END(__align_cpy_16)
+
+#endif
+
+ .align 32
+ENTRY(memcpy)
+210:
+#ifndef USE_BPR
+ srl %o2, 0, %o2 /* IEU1 Group */
+#endif
+ brz,pn %o2, 209b /* CTI Group */
+ mov %o0, %g4 /* IEU0 */
+218: cmp %o2, 15 /* IEU1 Group */
+ bleu,pn %xcc, 208b /* CTI */
+ cmp %o2, (64 * 6) /* IEU1 Group */
+ bgeu,pn %xcc, 200b /* CTI */
+ andcc %o0, 7, %g2 /* IEU1 Group */
+ sub %o0, %o1, %g5 /* IEU0 */
+ andcc %g5, 3, %o5 /* IEU1 Group */
+ bne,pn %xcc, 212f /* CTI */
+ andcc %o1, 3, %g0 /* IEU1 Group */
+ be,a,pt %xcc, 216f /* CTI */
+ andcc %o1, 4, %g0 /* IEU1 Group */
+ andcc %o1, 1, %g0 /* IEU1 Group */
+ be,pn %xcc, 4f /* CTI */
+ andcc %o1, 2, %g0 /* IEU1 Group */
+ ldub [%o1], %g2 /* Load Group */
+ add %o1, 1, %o1 /* IEU0 */
+ add %o0, 1, %o0 /* IEU1 */
+ sub %o2, 1, %o2 /* IEU0 Group */
+ bne,pn %xcc, 5f /* CTI Group */
+ stb %g2, [%o0 - 1] /* Store */
+4: lduh [%o1], %g2 /* Load Group */
+ add %o1, 2, %o1 /* IEU0 */
+ add %o0, 2, %o0 /* IEU1 */
+ sub %o2, 2, %o2 /* IEU0 */
+ sth %g2, [%o0 - 2] /* Store Group + bubble */
+5: andcc %o1, 4, %g0 /* IEU1 */
+216: be,a,pn %xcc, 2f /* CTI */
+ andcc %o2, -128, %g6 /* IEU1 Group */
+ lduw [%o1], %g5 /* Load Group */
+ add %o1, 4, %o1 /* IEU0 */
+ add %o0, 4, %o0 /* IEU1 */
+ sub %o2, 4, %o2 /* IEU0 Group */
+ stw %g5, [%o0 - 4] /* Store */
+ andcc %o2, -128, %g6 /* IEU1 Group */
+2: be,pn %xcc, 215f /* CTI */
+ andcc %o0, 4, %g0 /* IEU1 Group */
+ be,pn %xcc, 82f + 4 /* CTI Group */
+5: MOVE_BIGCHUNK(o1, o0, 0x00, g1, g3, g5, o5)
+ MOVE_BIGCHUNK(o1, o0, 0x20, g1, g3, g5, o5)
+ MOVE_BIGCHUNK(o1, o0, 0x40, g1, g3, g5, o5)
+ MOVE_BIGCHUNK(o1, o0, 0x60, g1, g3, g5, o5)
+35: subcc %g6, 128, %g6 /* IEU1 Group */
+ add %o1, 128, %o1 /* IEU0 */
+ bne,pt %xcc, 5b /* CTI */
+ add %o0, 128, %o0 /* IEU0 Group */
+215: andcc %o2, 0x70, %g6 /* IEU1 Group */
+41: be,pn %xcc, 80f /* CTI */
+ andcc %o2, 8, %g0 /* IEU1 Group */
+ /* Clk1 8-( */
+ /* Clk2 8-( */
+ /* Clk3 8-( */
+ /* Clk4 8-( */
+79: rd %pc, %o5 /* PDU Group */
+ sll %g6, 1, %g5 /* IEU0 Group */
+ add %o1, %g6, %o1 /* IEU1 */
+ sub %o5, %g5, %o5 /* IEU0 Group */
+ jmpl %o5 + %lo(80f - 79b), %g0 /* CTI Group brk forced*/
+ add %o0, %g6, %o0 /* IEU0 Group */
+36: MOVE_LASTCHUNK(o1, o0, 0x60, g2, g3, g5, o5)
+ MOVE_LASTCHUNK(o1, o0, 0x50, g2, g3, g5, o5)
+ MOVE_LASTCHUNK(o1, o0, 0x40, g2, g3, g5, o5)
+ MOVE_LASTCHUNK(o1, o0, 0x30, g2, g3, g5, o5)
+ MOVE_LASTCHUNK(o1, o0, 0x20, g2, g3, g5, o5)
+ MOVE_LASTCHUNK(o1, o0, 0x10, g2, g3, g5, o5)
+ MOVE_LASTCHUNK(o1, o0, 0x00, g2, g3, g5, o5)
+80: be,pt %xcc, 81f /* CTI */
+ andcc %o2, 4, %g0 /* IEU1 */
+ ldx [%o1], %g2 /* Load Group */
+ add %o0, 8, %o0 /* IEU0 */
+ stw %g2, [%o0 - 0x4] /* Store Group */
+ add %o1, 8, %o1 /* IEU1 */
+ srlx %g2, 32, %g2 /* IEU0 Group */
+ stw %g2, [%o0 - 0x8] /* Store */
+81: be,pt %xcc, 1f /* CTI */
+ andcc %o2, 2, %g0 /* IEU1 Group */
+ lduw [%o1], %g2 /* Load Group */
+ add %o1, 4, %o1 /* IEU0 */
+ stw %g2, [%o0] /* Store Group */
+ add %o0, 4, %o0 /* IEU0 */
+1: be,pt %xcc, 1f /* CTI */
+ andcc %o2, 1, %g0 /* IEU1 Group */
+ lduh [%o1], %g2 /* Load Group */
+ add %o1, 2, %o1 /* IEU0 */
+ sth %g2, [%o0] /* Store Group */
+ add %o0, 2, %o0 /* IEU0 */
+1: be,pt %xcc, 211f /* CTI */
+ nop /* IEU1 */
+ ldub [%o1], %g2 /* Load Group */
+ stb %g2, [%o0] /* Store Group + bubble */
+211: retl
+ mov %g4, %o0
+
+82: MOVE_BIGALIGNCHUNK(o1, o0, 0x00, g1, g3, g5, o5)
+ MOVE_BIGALIGNCHUNK(o1, o0, 0x40, g1, g3, g5, o5)
+37: subcc %g6, 128, %g6 /* IEU1 Group */
+ add %o1, 128, %o1 /* IEU0 */
+ bne,pt %xcc, 82b /* CTI */
+ add %o0, 128, %o0 /* IEU0 Group */
+ andcc %o2, 0x70, %g6 /* IEU1 */
+ be,pn %xcc, 84f /* CTI */
+ andcc %o2, 8, %g0 /* IEU1 Group */
+ /* Clk1 8-( */
+ /* Clk2 8-( */
+ /* Clk3 8-( */
+ /* Clk4 8-( */
+83: rd %pc, %o5 /* PDU Group */
+ add %o1, %g6, %o1 /* IEU0 Group */
+ sub %o5, %g6, %o5 /* IEU1 */
+ jmpl %o5 + %lo(84f - 83b), %g0 /* CTI Group brk forced*/
+ add %o0, %g6, %o0 /* IEU0 Group */
+38: MOVE_LASTALIGNCHUNK(o1, o0, 0x60, g2, g3)
+ MOVE_LASTALIGNCHUNK(o1, o0, 0x50, g2, g3)
+ MOVE_LASTALIGNCHUNK(o1, o0, 0x40, g2, g3)
+ MOVE_LASTALIGNCHUNK(o1, o0, 0x30, g2, g3)
+ MOVE_LASTALIGNCHUNK(o1, o0, 0x20, g2, g3)
+ MOVE_LASTALIGNCHUNK(o1, o0, 0x10, g2, g3)
+ MOVE_LASTALIGNCHUNK(o1, o0, 0x00, g2, g3)
+84: be,pt %xcc, 85f /* CTI Group */
+ andcc %o2, 4, %g0 /* IEU1 */
+ ldx [%o1], %g2 /* Load Group */
+ add %o0, 8, %o0 /* IEU0 */
+ add %o1, 8, %o1 /* IEU0 Group */
+ stx %g2, [%o0 - 0x8] /* Store */
+85: be,pt %xcc, 1f /* CTI */
+ andcc %o2, 2, %g0 /* IEU1 Group */
+ lduw [%o1], %g2 /* Load Group */
+ add %o0, 4, %o0 /* IEU0 */
+ add %o1, 4, %o1 /* IEU0 Group */
+ stw %g2, [%o0 - 0x4] /* Store */
+1: be,pt %xcc, 1f /* CTI */
+ andcc %o2, 1, %g0 /* IEU1 Group */
+ lduh [%o1], %g2 /* Load Group */
+ add %o0, 2, %o0 /* IEU0 */
+ add %o1, 2, %o1 /* IEU0 Group */
+ sth %g2, [%o0 - 0x2] /* Store */
+1: be,pt %xcc, 1f /* CTI */
+ nop /* IEU0 Group */
+ ldub [%o1], %g2 /* Load Group */
+ stb %g2, [%o0] /* Store Group + bubble */
+1: retl
+ mov %g4, %o0
+
+212: brz,pt %g2, 2f /* CTI Group */
+ mov 8, %g1 /* IEU0 */
+ sub %g1, %g2, %g2 /* IEU0 Group */
+ sub %o2, %g2, %o2 /* IEU0 Group */
+1: ldub [%o1], %g5 /* Load Group */
+ add %o1, 1, %o1 /* IEU0 */
+ add %o0, 1, %o0 /* IEU1 */
+ subcc %g2, 1, %g2 /* IEU1 Group */
+ bne,pt %xcc, 1b /* CTI */
+ stb %g5, [%o0 - 1] /* Store */
+2: andn %o2, 7, %g5 /* IEU0 Group */
+ and %o2, 7, %o2 /* IEU1 */
+ fmovd %f0, %f2 /* FPU */
+ alignaddr %o1, %g0, %g1 /* GRU Group */
+ ldd [%g1], %f4 /* Load Group */
+1: ldd [%g1 + 0x8], %f6 /* Load Group */
+ add %g1, 0x8, %g1 /* IEU0 Group */
+ subcc %g5, 8, %g5 /* IEU1 */
+ faligndata %f4, %f6, %f0 /* GRU Group */
+ std %f0, [%o0] /* Store */
+ add %o1, 8, %o1 /* IEU0 Group */
+ be,pn %xcc, 213f /* CTI */
+ add %o0, 8, %o0 /* IEU1 */
+ ldd [%g1 + 0x8], %f4 /* Load Group */
+ add %g1, 8, %g1 /* IEU0 */
+ subcc %g5, 8, %g5 /* IEU1 */
+ faligndata %f6, %f4, %f0 /* GRU Group */
+ std %f0, [%o0] /* Store */
+ add %o1, 8, %o1 /* IEU0 */
+ bne,pn %xcc, 1b /* CTI Group */
+ add %o0, 8, %o0 /* IEU0 */
+213: brz,pn %o2, 214f /* CTI Group */
+ nop /* IEU0 */
+ ldub [%o1], %g5 /* LOAD */
+ add %o1, 1, %o1 /* IEU0 */
+ add %o0, 1, %o0 /* IEU1 */
+ subcc %o2, 1, %o2 /* IEU1 */
+ bne,pt %xcc, 206b /* CTI */
+ stb %g5, [%o0 - 1] /* Store Group */
+214: wr %g0, FPRS_FEF, %fprs
+ retl
+ mov %g4, %o0
+END(memcpy)
+
+ .align 32
+228: andcc %o2, 1, %g0 /* IEU1 Group */
+ be,pt %icc, 2f+4 /* CTI */
+1: ldub [%o1 - 1], %o5 /* LOAD Group */
+ sub %o1, 1, %o1 /* IEU0 */
+ sub %o0, 1, %o0 /* IEU1 */
+ subcc %o2, 1, %o2 /* IEU1 Group */
+ be,pn %xcc, 229f /* CTI */
+ stb %o5, [%o0] /* Store */
+2: ldub [%o1 - 1], %o5 /* LOAD Group */
+ sub %o0, 2, %o0 /* IEU0 */
+ ldub [%o1 - 2], %g5 /* LOAD Group */
+ sub %o1, 2, %o1 /* IEU0 */
+ subcc %o2, 2, %o2 /* IEU1 Group */
+ stb %o5, [%o0 + 1] /* Store */
+ bne,pt %xcc, 2b /* CTI */
+ stb %g5, [%o0] /* Store */
+229: retl
+ mov %g4, %o0
+219: retl
+ nop
+
+ .align 32
+ENTRY(memmove)
+#ifndef USE_BPR
+ srl %o2, 0, %o2 /* IEU1 Group */
+#endif
+ brz,pn %o2, 219b /* CTI Group */
+ sub %o0, %o1, %o4 /* IEU0 */
+ cmp %o4, %o2 /* IEU1 Group */
+ bgeu,pt %XCC, 218b /* CTI */
+ mov %o0, %g4 /* IEU0 */
+ add %o0, %o2, %o0 /* IEU0 Group */
+220: add %o1, %o2, %o1 /* IEU1 */
+ cmp %o2, 15 /* IEU1 Group */
+ bleu,pn %xcc, 228b /* CTI */
+ andcc %o0, 7, %g2 /* IEU1 Group */
+ sub %o0, %o1, %g5 /* IEU0 */
+ andcc %g5, 3, %o5 /* IEU1 Group */
+ bne,pn %xcc, 232f /* CTI */
+ andcc %o1, 3, %g0 /* IEU1 Group */
+ be,a,pt %xcc, 236f /* CTI */
+ andcc %o1, 4, %g0 /* IEU1 Group */
+ andcc %o1, 1, %g0 /* IEU1 Group */
+ be,pn %xcc, 4f /* CTI */
+ andcc %o1, 2, %g0 /* IEU1 Group */
+ ldub [%o1 - 1], %g2 /* Load Group */
+ sub %o1, 1, %o1 /* IEU0 */
+ sub %o0, 1, %o0 /* IEU1 */
+ sub %o2, 1, %o2 /* IEU0 Group */
+ be,pn %xcc, 5f /* CTI Group */
+ stb %g2, [%o0] /* Store */
+4: lduh [%o1 - 2], %g2 /* Load Group */
+ sub %o1, 2, %o1 /* IEU0 */
+ sub %o0, 2, %o0 /* IEU1 */
+ sub %o2, 2, %o2 /* IEU0 */
+ sth %g2, [%o0] /* Store Group + bubble */
+5: andcc %o1, 4, %g0 /* IEU1 */
+236: be,a,pn %xcc, 2f /* CTI */
+ andcc %o2, -128, %g6 /* IEU1 Group */
+ lduw [%o1 - 4], %g5 /* Load Group */
+ sub %o1, 4, %o1 /* IEU0 */
+ sub %o0, 4, %o0 /* IEU1 */
+ sub %o2, 4, %o2 /* IEU0 Group */
+ stw %g5, [%o0] /* Store */
+ andcc %o2, -128, %g6 /* IEU1 Group */
+2: be,pn %xcc, 235f /* CTI */
+ andcc %o0, 4, %g0 /* IEU1 Group */
+ be,pn %xcc, 282f + 4 /* CTI Group */
+5: RMOVE_BIGCHUNK(o1, o0, 0x00, g1, g3, g5, o5)
+ RMOVE_BIGCHUNK(o1, o0, 0x20, g1, g3, g5, o5)
+ RMOVE_BIGCHUNK(o1, o0, 0x40, g1, g3, g5, o5)
+ RMOVE_BIGCHUNK(o1, o0, 0x60, g1, g3, g5, o5)
+ subcc %g6, 128, %g6 /* IEU1 Group */
+ sub %o1, 128, %o1 /* IEU0 */
+ bne,pt %xcc, 5b /* CTI */
+ sub %o0, 128, %o0 /* IEU0 Group */
+235: andcc %o2, 0x70, %g6 /* IEU1 Group */
+41: be,pn %xcc, 280f /* CTI */
+ andcc %o2, 8, %g0 /* IEU1 Group */
+ /* Clk1 8-( */
+ /* Clk2 8-( */
+ /* Clk3 8-( */
+ /* Clk4 8-( */
+279: rd %pc, %o5 /* PDU Group */
+ sll %g6, 1, %g5 /* IEU0 Group */
+ sub %o1, %g6, %o1 /* IEU1 */
+ sub %o5, %g5, %o5 /* IEU0 Group */
+ jmpl %o5 + %lo(280f - 279b), %g0 /* CTI Group brk forced*/
+ sub %o0, %g6, %o0 /* IEU0 Group */
+ RMOVE_LASTCHUNK(o1, o0, 0x60, g2, g3, g5, o5)
+ RMOVE_LASTCHUNK(o1, o0, 0x50, g2, g3, g5, o5)
+ RMOVE_LASTCHUNK(o1, o0, 0x40, g2, g3, g5, o5)
+ RMOVE_LASTCHUNK(o1, o0, 0x30, g2, g3, g5, o5)
+ RMOVE_LASTCHUNK(o1, o0, 0x20, g2, g3, g5, o5)
+ RMOVE_LASTCHUNK(o1, o0, 0x10, g2, g3, g5, o5)
+ RMOVE_LASTCHUNK(o1, o0, 0x00, g2, g3, g5, o5)
+280: be,pt %xcc, 281f /* CTI */
+ andcc %o2, 4, %g0 /* IEU1 */
+ ldx [%o1 - 8], %g2 /* Load Group */
+ sub %o0, 8, %o0 /* IEU0 */
+ stw %g2, [%o0 + 4] /* Store Group */
+ sub %o1, 8, %o1 /* IEU1 */
+ srlx %g2, 32, %g2 /* IEU0 Group */
+ stw %g2, [%o0] /* Store */
+281: be,pt %xcc, 1f /* CTI */
+ andcc %o2, 2, %g0 /* IEU1 Group */
+ lduw [%o1 - 4], %g2 /* Load Group */
+ sub %o1, 4, %o1 /* IEU0 */
+ stw %g2, [%o0 - 4] /* Store Group */
+ sub %o0, 4, %o0 /* IEU0 */
+1: be,pt %xcc, 1f /* CTI */
+ andcc %o2, 1, %g0 /* IEU1 Group */
+ lduh [%o1 - 2], %g2 /* Load Group */
+ sub %o1, 2, %o1 /* IEU0 */
+ sth %g2, [%o0 - 2] /* Store Group */
+ sub %o0, 2, %o0 /* IEU0 */
+1: be,pt %xcc, 211f /* CTI */
+ nop /* IEU1 */
+ ldub [%o1 - 1], %g2 /* Load Group */
+ stb %g2, [%o0 - 1] /* Store Group + bubble */
+211: retl
+ mov %g4, %o0
+
+282: RMOVE_BIGALIGNCHUNK(o1, o0, 0x00, g1, g3, g5, o5)
+ RMOVE_BIGALIGNCHUNK(o1, o0, 0x40, g1, g3, g5, o5)
+ subcc %g6, 128, %g6 /* IEU1 Group */
+ sub %o1, 128, %o1 /* IEU0 */
+ bne,pt %xcc, 282b /* CTI */
+ sub %o0, 128, %o0 /* IEU0 Group */
+ andcc %o2, 0x70, %g6 /* IEU1 */
+ be,pn %xcc, 284f /* CTI */
+ andcc %o2, 8, %g0 /* IEU1 Group */
+ /* Clk1 8-( */
+ /* Clk2 8-( */
+ /* Clk3 8-( */
+ /* Clk4 8-( */
+283: rd %pc, %o5 /* PDU Group */
+ sub %o1, %g6, %o1 /* IEU0 Group */
+ sub %o5, %g6, %o5 /* IEU1 */
+ jmpl %o5 + %lo(284f - 283b), %g0 /* CTI Group brk forced*/
+ sub %o0, %g6, %o0 /* IEU0 Group */
+ RMOVE_LASTALIGNCHUNK(o1, o0, 0x60, g2, g3)
+ RMOVE_LASTALIGNCHUNK(o1, o0, 0x50, g2, g3)
+ RMOVE_LASTALIGNCHUNK(o1, o0, 0x40, g2, g3)
+ RMOVE_LASTALIGNCHUNK(o1, o0, 0x30, g2, g3)
+ RMOVE_LASTALIGNCHUNK(o1, o0, 0x20, g2, g3)
+ RMOVE_LASTALIGNCHUNK(o1, o0, 0x10, g2, g3)
+ RMOVE_LASTALIGNCHUNK(o1, o0, 0x00, g2, g3)
+284: be,pt %xcc, 285f /* CTI Group */
+ andcc %o2, 4, %g0 /* IEU1 */
+ ldx [%o1 - 8], %g2 /* Load Group */
+ sub %o0, 8, %o0 /* IEU0 */
+ sub %o1, 8, %o1 /* IEU0 Group */
+ stx %g2, [%o0] /* Store */
+285: be,pt %xcc, 1f /* CTI */
+ andcc %o2, 2, %g0 /* IEU1 Group */
+ lduw [%o1 - 4], %g2 /* Load Group */
+ sub %o0, 4, %o0 /* IEU0 */
+ sub %o1, 4, %o1 /* IEU0 Group */
+ stw %g2, [%o0] /* Store */
+1: be,pt %xcc, 1f /* CTI */
+ andcc %o2, 1, %g0 /* IEU1 Group */
+ lduh [%o1 - 2], %g2 /* Load Group */
+ sub %o0, 2, %o0 /* IEU0 */
+ sub %o1, 2, %o1 /* IEU0 Group */
+ sth %g2, [%o0] /* Store */
+1: be,pt %xcc, 1f /* CTI */
+ nop /* IEU0 Group */
+ ldub [%o1 - 1], %g2 /* Load Group */
+ stb %g2, [%o0 - 1] /* Store Group + bubble */
+1: retl
+ mov %g4, %o0
+
+232: brz,pt %g2, 2f /* CTI Group */
+ sub %o2, %g2, %o2 /* IEU0 Group */
+1: ldub [%o1 - 1], %g5 /* Load Group */
+ sub %o1, 1, %o1 /* IEU0 */
+ sub %o0, 1, %o0 /* IEU1 */
+ subcc %g2, 1, %g2 /* IEU1 Group */
+ bne,pt %xcc, 1b /* CTI */
+ stb %g5, [%o0] /* Store */
+2: andn %o2, 7, %g5 /* IEU0 Group */
+ and %o2, 7, %o2 /* IEU1 */
+ fmovd %f0, %f2 /* FPU */
+ alignaddr %o1, %g0, %g1 /* GRU Group */
+ ldd [%g1], %f4 /* Load Group */
+1: ldd [%g1 - 8], %f6 /* Load Group */
+ sub %g1, 8, %g1 /* IEU0 Group */
+ subcc %g5, 8, %g5 /* IEU1 */
+ faligndata %f6, %f4, %f0 /* GRU Group */
+ std %f0, [%o0 - 8] /* Store */
+ sub %o1, 8, %o1 /* IEU0 Group */
+ be,pn %xcc, 233f /* CTI */
+ sub %o0, 8, %o0 /* IEU1 */
+ ldd [%g1 - 8], %f4 /* Load Group */
+ sub %g1, 8, %g1 /* IEU0 */
+ subcc %g5, 8, %g5 /* IEU1 */
+ faligndata %f4, %f6, %f0 /* GRU Group */
+ std %f0, [%o0 - 8] /* Store */
+ sub %o1, 8, %o1 /* IEU0 */
+ bne,pn %xcc, 1b /* CTI Group */
+ sub %o0, 8, %o0 /* IEU0 */
+233: brz,pn %o2, 234f /* CTI Group */
+ nop /* IEU0 */
+237: ldub [%o1 - 1], %g5 /* LOAD */
+ sub %o1, 1, %o1 /* IEU0 */
+ sub %o0, 1, %o0 /* IEU1 */
+ subcc %o2, 1, %o2 /* IEU1 */
+ bne,pt %xcc, 237b /* CTI */
+ stb %g5, [%o0] /* Store Group */
+234: wr %g0, FPRS_FEF, %fprs
+ retl
+ mov %g4, %o0
+END(memmove)
+
+#ifdef USE_BPR
+weak_alias (memcpy, __align_cpy_1)
+weak_alias (memcpy, __align_cpy_2)
+#endif
diff --git a/libc/string/sparc/sparc64/memset.S b/libc/string/sparc/sparc64/memset.S
new file mode 100644
index 000000000..2e23f92a8
--- /dev/null
+++ b/libc/string/sparc/sparc64/memset.S
@@ -0,0 +1,313 @@
+/* Set a block of memory to some byte value.
+ For UltraSPARC.
+ Copyright (C) 1996, 97, 98, 99, 2003 Free Software Foundation, Inc.
+ This file is part of the GNU C Library.
+ Contributed by David S. Miller (davem@caip.rutgers.edu) and
+ Jakub Jelinek (jj@ultra.linux.cz).
+
+ The GNU C Library is free software; you can redistribute it and/or
+ modify it under the terms of the GNU Lesser General Public
+ License as published by the Free Software Foundation; either
+ version 2.1 of the License, or (at your option) any later version.
+
+ The GNU C Library is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ Lesser General Public License for more details.
+
+ You should have received a copy of the GNU Lesser General Public
+ License along with the GNU C Library; if not, write to the Free
+ Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA
+ 02111-1307 USA. */
+
+#include <asm/asi.h>
+#ifndef XCC
+#define XCC xcc
+#define USE_BPR
+#endif
+#define FPRS_FEF 4
+
+#define SET_BLOCKS(base, offset, source) \
+ stx source, [base - offset - 0x18]; \
+ stx source, [base - offset - 0x10]; \
+ stx source, [base - offset - 0x08]; \
+ stx source, [base - offset - 0x00];
+
+ /* Well, memset is a lot easier to get right than bcopy... */
+ .text
+ .align 32
+ENTRY(memset)
+ andcc %o1, 0xff, %o1
+ mov %o0, %o5
+ be,a,pt %icc, 50f
+#ifndef USE_BPR
+ srl %o2, 0, %o1
+#else
+ mov %o2, %o1
+#endif
+ cmp %o2, 7
+#ifndef USE_BPR
+ srl %o2, 0, %o2
+#endif
+ bleu,pn %XCC, 17f
+ andcc %o0, 3, %g5
+ be,pt %xcc, 4f
+ and %o1, 0xff, %o1
+ cmp %g5, 3
+ be,pn %xcc, 2f
+ stb %o1, [%o0 + 0x00]
+ cmp %g5, 2
+ be,pt %xcc, 2f
+ stb %o1, [%o0 + 0x01]
+ stb %o1, [%o0 + 0x02]
+2: sub %g5, 4, %g5
+ sub %o0, %g5, %o0
+ add %o2, %g5, %o2
+4: sllx %o1, 8, %g1
+ andcc %o0, 4, %g0
+ or %o1, %g1, %o1
+ sllx %o1, 16, %g1
+ or %o1, %g1, %o1
+ be,pt %xcc, 2f
+ sllx %o1, 32, %g1
+ stw %o1, [%o0]
+ sub %o2, 4, %o2
+ add %o0, 4, %o0
+2: cmp %o2, 128
+ or %o1, %g1, %o1
+ blu,pn %xcc, 9f
+ andcc %o0, 0x38, %g5
+ be,pn %icc, 6f
+ mov 64, %o4
+ andcc %o0, 8, %g0
+ be,pn %icc, 1f
+ sub %o4, %g5, %o4
+ stx %o1, [%o0]
+ add %o0, 8, %o0
+1: andcc %o4, 16, %g0
+ be,pn %icc, 1f
+ sub %o2, %o4, %o2
+ stx %o1, [%o0]
+ stx %o1, [%o0 + 8]
+ add %o0, 16, %o0
+1: andcc %o4, 32, %g0
+ be,pn %icc, 7f
+ andncc %o2, 0x3f, %o3
+ stw %o1, [%o0]
+ stw %o1, [%o0 + 4]
+ stw %o1, [%o0 + 8]
+ stw %o1, [%o0 + 12]
+ stw %o1, [%o0 + 16]
+ stw %o1, [%o0 + 20]
+ stw %o1, [%o0 + 24]
+ stw %o1, [%o0 + 28]
+ add %o0, 32, %o0
+7: be,pn %xcc, 9f
+ nop
+ ldd [%o0 - 8], %f0
+18: wr %g0, ASI_BLK_P, %asi
+ membar #StoreStore | #LoadStore
+ andcc %o3, 0xc0, %g5
+ and %o2, 0x3f, %o2
+ fmovd %f0, %f2
+ fmovd %f0, %f4
+ andn %o3, 0xff, %o3
+ fmovd %f0, %f6
+ cmp %g5, 64
+ fmovd %f0, %f8
+ fmovd %f0, %f10
+ fmovd %f0, %f12
+ brz,pn %g5, 10f
+ fmovd %f0, %f14
+ be,pn %icc, 2f
+ stda %f0, [%o0 + 0x00] %asi
+ cmp %g5, 128
+ be,pn %icc, 2f
+ stda %f0, [%o0 + 0x40] %asi
+ stda %f0, [%o0 + 0x80] %asi
+2: brz,pn %o3, 12f
+ add %o0, %g5, %o0
+10: stda %f0, [%o0 + 0x00] %asi
+ stda %f0, [%o0 + 0x40] %asi
+ stda %f0, [%o0 + 0x80] %asi
+ stda %f0, [%o0 + 0xc0] %asi
+11: subcc %o3, 256, %o3
+ bne,pt %xcc, 10b
+ add %o0, 256, %o0
+12: wr %g0, FPRS_FEF, %fprs
+ membar #StoreLoad | #StoreStore
+9: andcc %o2, 0x78, %g5
+ be,pn %xcc, 13f
+ andcc %o2, 7, %o2
+14: rd %pc, %o4
+ srl %g5, 1, %o3
+ sub %o4, %o3, %o4
+ jmpl %o4 + (13f - 14b), %g0
+ add %o0, %g5, %o0
+12: SET_BLOCKS (%o0, 0x68, %o1)
+ SET_BLOCKS (%o0, 0x48, %o1)
+ SET_BLOCKS (%o0, 0x28, %o1)
+ SET_BLOCKS (%o0, 0x08, %o1)
+13: be,pn %xcc, 8f
+ andcc %o2, 4, %g0
+ be,pn %xcc, 1f
+ andcc %o2, 2, %g0
+ stw %o1, [%o0]
+ add %o0, 4, %o0
+1: be,pn %xcc, 1f
+ andcc %o2, 1, %g0
+ sth %o1, [%o0]
+ add %o0, 2, %o0
+1: bne,a,pn %xcc, 8f
+ stb %o1, [%o0]
+8: retl
+ mov %o5, %o0
+17: brz,pn %o2, 0f
+8: add %o0, 1, %o0
+ subcc %o2, 1, %o2
+ bne,pt %xcc, 8b
+ stb %o1, [%o0 - 1]
+0: retl
+ mov %o5, %o0
+
+6: stx %o1, [%o0]
+ andncc %o2, 0x3f, %o3
+ be,pn %xcc, 9b
+ nop
+ ba,pt %xcc, 18b
+ ldd [%o0], %f0
+END(memset)
+
+#define ZERO_BLOCKS(base, offset, source) \
+ stx source, [base - offset - 0x38]; \
+ stx source, [base - offset - 0x30]; \
+ stx source, [base - offset - 0x28]; \
+ stx source, [base - offset - 0x20]; \
+ stx source, [base - offset - 0x18]; \
+ stx source, [base - offset - 0x10]; \
+ stx source, [base - offset - 0x08]; \
+ stx source, [base - offset - 0x00];
+
+ .text
+ .align 32
+ENTRY(__bzero)
+#ifndef USE_BPR
+ srl %o1, 0, %o1
+#endif
+ mov %o0, %o5
+50: cmp %o1, 7
+ bleu,pn %xcc, 17f
+ andcc %o0, 3, %o2
+ be,a,pt %xcc, 4f
+ andcc %o0, 4, %g0
+ cmp %o2, 3
+ be,pn %xcc, 2f
+ stb %g0, [%o0 + 0x00]
+ cmp %o2, 2
+ be,pt %xcc, 2f
+ stb %g0, [%o0 + 0x01]
+ stb %g0, [%o0 + 0x02]
+2: sub %o2, 4, %o2
+ sub %o0, %o2, %o0
+ add %o1, %o2, %o1
+ andcc %o0, 4, %g0
+4: be,pt %xcc, 2f
+ cmp %o1, 128
+ stw %g0, [%o0]
+ sub %o1, 4, %o1
+ add %o0, 4, %o0
+2: blu,pn %xcc, 9f
+ andcc %o0, 0x38, %o2
+ be,pn %icc, 6f
+ mov 64, %o4
+ andcc %o0, 8, %g0
+ be,pn %icc, 1f
+ sub %o4, %o2, %o4
+ stx %g0, [%o0]
+ add %o0, 8, %o0
+1: andcc %o4, 16, %g0
+ be,pn %icc, 1f
+ sub %o1, %o4, %o1
+ stx %g0, [%o0]
+ stx %g0, [%o0 + 8]
+ add %o0, 16, %o0
+1: andcc %o4, 32, %g0
+ be,pn %icc, 7f
+ andncc %o1, 0x3f, %o3
+ stx %g0, [%o0]
+ stx %g0, [%o0 + 8]
+ stx %g0, [%o0 + 16]
+ stx %g0, [%o0 + 24]
+ add %o0, 32, %o0
+6: andncc %o1, 0x3f, %o3
+7: be,pn %xcc, 9f
+ wr %g0, ASI_BLK_P, %asi
+ membar #StoreLoad | #StoreStore | #LoadStore
+ fzero %f0
+ andcc %o3, 0xc0, %o2
+ and %o1, 0x3f, %o1
+ fzero %f2
+ andn %o3, 0xff, %o3
+ faddd %f0, %f2, %f4
+ fmuld %f0, %f2, %f6
+ cmp %o2, 64
+ faddd %f0, %f2, %f8
+ fmuld %f0, %f2, %f10
+ faddd %f0, %f2, %f12
+ brz,pn %o2, 10f
+ fmuld %f0, %f2, %f14
+ be,pn %icc, 2f
+ stda %f0, [%o0 + 0x00] %asi
+ cmp %o2, 128
+ be,pn %icc, 2f
+ stda %f0, [%o0 + 0x40] %asi
+ stda %f0, [%o0 + 0x80] %asi
+2: brz,pn %o3, 12f
+ add %o0, %o2, %o0
+10: stda %f0, [%o0 + 0x00] %asi
+ stda %f0, [%o0 + 0x40] %asi
+ stda %f0, [%o0 + 0x80] %asi
+ stda %f0, [%o0 + 0xc0] %asi
+11: subcc %o3, 256, %o3
+ bne,pt %xcc, 10b
+ add %o0, 256, %o0
+12: wr %g0, FPRS_FEF, %fprs
+ membar #StoreLoad | #StoreStore
+9: andcc %o1, 0xf8, %o2
+ be,pn %xcc, 13f
+ andcc %o1, 7, %o1
+14: rd %pc, %o4
+ srl %o2, 1, %o3
+ sub %o4, %o3, %o4
+ jmpl %o4 + (13f - 14b), %g0
+ add %o0, %o2, %o0
+12: ZERO_BLOCKS (%o0, 0xc8, %g0)
+ ZERO_BLOCKS (%o0, 0x88, %g0)
+ ZERO_BLOCKS (%o0, 0x48, %g0)
+ ZERO_BLOCKS (%o0, 0x08, %g0)
+13: be,pn %xcc, 8f
+ andcc %o1, 4, %g0
+ be,pn %xcc, 1f
+ andcc %o1, 2, %g0
+ stw %g0, [%o0]
+ add %o0, 4, %o0
+1: be,pn %xcc, 1f
+ andcc %o1, 1, %g0
+ sth %g0, [%o0]
+ add %o0, 2, %o0
+1: bne,a,pn %xcc, 8f
+ stb %g0, [%o0]
+8: retl
+ mov %o5, %o0
+17: be,pn %xcc, 13b
+ orcc %o1, 0, %g0
+ be,pn %xcc, 0f
+8: add %o0, 1, %o0
+ subcc %o1, 1, %o1
+ bne,pt %xcc, 8b
+ stb %g0, [%o0 - 1]
+0: retl
+ mov %o5, %o0
+END(__bzero)
+
+weak_alias (__bzero, bzero)
diff --git a/libc/string/sparc/sparc64/sparcv9b/memcpy.S b/libc/string/sparc/sparc64/sparcv9b/memcpy.S
new file mode 100644
index 000000000..7f36b7c2c
--- /dev/null
+++ b/libc/string/sparc/sparc64/sparcv9b/memcpy.S
@@ -0,0 +1,606 @@
+/* Copy SIZE bytes from SRC to DEST.
+ For UltraSPARC-III.
+ Copyright (C) 2001, 2003 Free Software Foundation, Inc.
+ This file is part of the GNU C Library.
+ Contributed by David S. Miller (davem@redhat.com)
+
+ The GNU C Library is free software; you can redistribute it and/or
+ modify it under the terms of the GNU Lesser General Public
+ License as published by the Free Software Foundation; either
+ version 2.1 of the License, or (at your option) any later version.
+
+ The GNU C Library is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ Lesser General Public License for more details.
+
+ You should have received a copy of the GNU Lesser General Public
+ License along with the GNU C Library; if not, write to the Free
+ Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA
+ 02111-1307 USA. */
+
+#define ASI_BLK_P 0xf0
+#define FPRS_FEF 0x04
+#define VISEntryHalf rd %fprs, %o5; wr %g0, FPRS_FEF, %fprs
+#define VISExitHalf and %o5, FPRS_FEF, %o5; wr %o5, 0x0, %fprs
+
+#ifndef XCC
+#define USE_BPR
+#define XCC xcc
+#endif
+
+ .register %g2,#scratch
+ .register %g3,#scratch
+ .register %g6,#scratch
+
+ .text
+ .align 32
+
+ENTRY(bcopy)
+ sub %o1, %o0, %o4
+ mov %o0, %g4
+ cmp %o4, %o2
+ mov %o1, %o0
+ bgeu,pt %XCC, 100f
+ mov %g4, %o1
+#ifndef USE_BPR
+ srl %o2, 0, %o2
+#endif
+ brnz,pn %o2, 220f
+ add %o0, %o2, %o0
+ retl
+ nop
+END(bcopy)
+
+ /* Special/non-trivial issues of this code:
+ *
+ * 1) %o5 is preserved from VISEntryHalf to VISExitHalf
+ * 2) Only low 32 FPU registers are used so that only the
+ * lower half of the FPU register set is dirtied by this
+ * code. This is especially important in the kernel.
+ * 3) This code never prefetches cachelines past the end
+ * of the source buffer.
+ *
+ * The cheetah's flexible spine, oversized liver, enlarged heart,
+ * slender muscular body, and claws make it the swiftest hunter
+ * in Africa and the fastest animal on land. Can reach speeds
+ * of up to 2.4GB per second.
+ */
+ .align 32
+ENTRY(memcpy)
+
+100: /* %o0=dst, %o1=src, %o2=len */
+ mov %o0, %g5
+ cmp %o2, 0
+ be,pn %XCC, out
+218: or %o0, %o1, %o3
+ cmp %o2, 16
+ bleu,a,pn %XCC, small_copy
+ or %o3, %o2, %o3
+
+ cmp %o2, 256
+ blu,pt %XCC, medium_copy
+ andcc %o3, 0x7, %g0
+
+ ba,pt %xcc, enter
+ andcc %o0, 0x3f, %g2
+
+ /* Here len >= 256 and condition codes reflect execution
+ * of "andcc %o0, 0x7, %g2", done by caller.
+ */
+ .align 64
+enter:
+ /* Is 'dst' already aligned on an 64-byte boundary? */
+ be,pt %XCC, 2f
+
+ /* Compute abs((dst & 0x3f) - 0x40) into %g2. This is the number
+ * of bytes to copy to make 'dst' 64-byte aligned. We pre-
+ * subtract this from 'len'.
+ */
+ sub %g2, 0x40, %g2
+ sub %g0, %g2, %g2
+ sub %o2, %g2, %o2
+
+ /* Copy %g2 bytes from src to dst, one byte at a time. */
+1: ldub [%o1 + 0x00], %o3
+ add %o1, 0x1, %o1
+ add %o0, 0x1, %o0
+ subcc %g2, 0x1, %g2
+
+ bg,pt %XCC, 1b
+ stb %o3, [%o0 + -1]
+
+2: VISEntryHalf
+ and %o1, 0x7, %g1
+ ba,pt %xcc, begin
+ alignaddr %o1, %g0, %o1
+
+ .align 64
+begin:
+ prefetch [%o1 + 0x000], #one_read
+ prefetch [%o1 + 0x040], #one_read
+ andn %o2, (0x40 - 1), %o4
+ prefetch [%o1 + 0x080], #one_read
+ prefetch [%o1 + 0x0c0], #one_read
+ ldd [%o1 + 0x000], %f0
+ prefetch [%o1 + 0x100], #one_read
+ ldd [%o1 + 0x008], %f2
+ prefetch [%o1 + 0x140], #one_read
+ ldd [%o1 + 0x010], %f4
+ prefetch [%o1 + 0x180], #one_read
+ faligndata %f0, %f2, %f16
+ ldd [%o1 + 0x018], %f6
+ faligndata %f2, %f4, %f18
+ ldd [%o1 + 0x020], %f8
+ faligndata %f4, %f6, %f20
+ ldd [%o1 + 0x028], %f10
+ faligndata %f6, %f8, %f22
+
+ ldd [%o1 + 0x030], %f12
+ faligndata %f8, %f10, %f24
+ ldd [%o1 + 0x038], %f14
+ faligndata %f10, %f12, %f26
+ ldd [%o1 + 0x040], %f0
+
+ sub %o4, 0x80, %o4
+ add %o1, 0x40, %o1
+ ba,pt %xcc, loop
+ srl %o4, 6, %o3
+
+ .align 64
+loop:
+ ldd [%o1 + 0x008], %f2
+ faligndata %f12, %f14, %f28
+ ldd [%o1 + 0x010], %f4
+ faligndata %f14, %f0, %f30
+ stda %f16, [%o0] ASI_BLK_P
+ ldd [%o1 + 0x018], %f6
+ faligndata %f0, %f2, %f16
+
+ ldd [%o1 + 0x020], %f8
+ faligndata %f2, %f4, %f18
+ ldd [%o1 + 0x028], %f10
+ faligndata %f4, %f6, %f20
+ ldd [%o1 + 0x030], %f12
+ faligndata %f6, %f8, %f22
+ ldd [%o1 + 0x038], %f14
+ faligndata %f8, %f10, %f24
+
+ ldd [%o1 + 0x040], %f0
+ prefetch [%o1 + 0x180], #one_read
+ faligndata %f10, %f12, %f26
+ subcc %o3, 0x01, %o3
+ add %o1, 0x40, %o1
+ bg,pt %XCC, loop
+ add %o0, 0x40, %o0
+
+ /* Finally we copy the last full 64-byte block. */
+loopfini:
+ ldd [%o1 + 0x008], %f2
+ faligndata %f12, %f14, %f28
+ ldd [%o1 + 0x010], %f4
+ faligndata %f14, %f0, %f30
+ stda %f16, [%o0] ASI_BLK_P
+ ldd [%o1 + 0x018], %f6
+ faligndata %f0, %f2, %f16
+ ldd [%o1 + 0x020], %f8
+ faligndata %f2, %f4, %f18
+ ldd [%o1 + 0x028], %f10
+ faligndata %f4, %f6, %f20
+ ldd [%o1 + 0x030], %f12
+ faligndata %f6, %f8, %f22
+ ldd [%o1 + 0x038], %f14
+ faligndata %f8, %f10, %f24
+ cmp %g1, 0
+ be,pt %XCC, 1f
+ add %o0, 0x40, %o0
+ ldd [%o1 + 0x040], %f0
+1: faligndata %f10, %f12, %f26
+ faligndata %f12, %f14, %f28
+ faligndata %f14, %f0, %f30
+ stda %f16, [%o0] ASI_BLK_P
+ add %o0, 0x40, %o0
+ add %o1, 0x40, %o1
+ membar #Sync
+
+ /* Now we copy the (len modulo 64) bytes at the end.
+ * Note how we borrow the %f0 loaded above.
+ *
+ * Also notice how this code is careful not to perform a
+ * load past the end of the src buffer.
+ */
+loopend:
+ and %o2, 0x3f, %o2
+ andcc %o2, 0x38, %g2
+ be,pn %XCC, endcruft
+ subcc %g2, 0x8, %g2
+ be,pn %XCC, endcruft
+ cmp %g1, 0
+
+ be,a,pt %XCC, 1f
+ ldd [%o1 + 0x00], %f0
+
+1: ldd [%o1 + 0x08], %f2
+ add %o1, 0x8, %o1
+ sub %o2, 0x8, %o2
+ subcc %g2, 0x8, %g2
+ faligndata %f0, %f2, %f8
+ std %f8, [%o0 + 0x00]
+ be,pn %XCC, endcruft
+ add %o0, 0x8, %o0
+ ldd [%o1 + 0x08], %f0
+ add %o1, 0x8, %o1
+ sub %o2, 0x8, %o2
+ subcc %g2, 0x8, %g2
+ faligndata %f2, %f0, %f8
+ std %f8, [%o0 + 0x00]
+ bne,pn %XCC, 1b
+ add %o0, 0x8, %o0
+
+ /* If anything is left, we copy it one byte at a time.
+ * Note that %g1 is (src & 0x3) saved above before the
+ * alignaddr was performed.
+ */
+endcruft:
+ cmp %o2, 0
+ add %o1, %g1, %o1
+ VISExitHalf
+ be,pn %XCC, out
+ sub %o0, %o1, %o3
+
+ andcc %g1, 0x7, %g0
+ bne,pn %icc, small_copy_unaligned
+ andcc %o2, 0x8, %g0
+ be,pt %icc, 1f
+ nop
+ ldx [%o1], %o5
+ stx %o5, [%o1 + %o3]
+ add %o1, 0x8, %o1
+
+1: andcc %o2, 0x4, %g0
+ be,pt %icc, 1f
+ nop
+ lduw [%o1], %o5
+ stw %o5, [%o1 + %o3]
+ add %o1, 0x4, %o1
+
+1: andcc %o2, 0x2, %g0
+ be,pt %icc, 1f
+ nop
+ lduh [%o1], %o5
+ sth %o5, [%o1 + %o3]
+ add %o1, 0x2, %o1
+
+1: andcc %o2, 0x1, %g0
+ be,pt %icc, out
+ nop
+ ldub [%o1], %o5
+ ba,pt %xcc, out
+ stb %o5, [%o1 + %o3]
+
+medium_copy: /* 16 < len <= 64 */
+ bne,pn %XCC, small_copy_unaligned
+ sub %o0, %o1, %o3
+
+medium_copy_aligned:
+ andn %o2, 0x7, %o4
+ and %o2, 0x7, %o2
+1: subcc %o4, 0x8, %o4
+ ldx [%o1], %o5
+ stx %o5, [%o1 + %o3]
+ bgu,pt %XCC, 1b
+ add %o1, 0x8, %o1
+ andcc %o2, 0x4, %g0
+ be,pt %XCC, 1f
+ nop
+ sub %o2, 0x4, %o2
+ lduw [%o1], %o5
+ stw %o5, [%o1 + %o3]
+ add %o1, 0x4, %o1
+1: cmp %o2, 0
+ be,pt %XCC, out
+ nop
+ ba,pt %xcc, small_copy_unaligned
+ nop
+
+small_copy: /* 0 < len <= 16 */
+ andcc %o3, 0x3, %g0
+ bne,pn %XCC, small_copy_unaligned
+ sub %o0, %o1, %o3
+
+small_copy_aligned:
+ subcc %o2, 4, %o2
+ lduw [%o1], %g1
+ stw %g1, [%o1 + %o3]
+ bgu,pt %XCC, small_copy_aligned
+ add %o1, 4, %o1
+
+out: retl
+ mov %g5, %o0
+
+ .align 32
+small_copy_unaligned:
+ subcc %o2, 1, %o2
+ ldub [%o1], %g1
+ stb %g1, [%o1 + %o3]
+ bgu,pt %XCC, small_copy_unaligned
+ add %o1, 1, %o1
+ retl
+ mov %g5, %o0
+
+END(memcpy)
+
+#define RMOVE_BIGCHUNK(src, dst, offset, t0, t1, t2, t3) \
+ ldx [%src - offset - 0x20], %t0; \
+ ldx [%src - offset - 0x18], %t1; \
+ ldx [%src - offset - 0x10], %t2; \
+ ldx [%src - offset - 0x08], %t3; \
+ stw %t0, [%dst - offset - 0x1c]; \
+ srlx %t0, 32, %t0; \
+ stw %t0, [%dst - offset - 0x20]; \
+ stw %t1, [%dst - offset - 0x14]; \
+ srlx %t1, 32, %t1; \
+ stw %t1, [%dst - offset - 0x18]; \
+ stw %t2, [%dst - offset - 0x0c]; \
+ srlx %t2, 32, %t2; \
+ stw %t2, [%dst - offset - 0x10]; \
+ stw %t3, [%dst - offset - 0x04]; \
+ srlx %t3, 32, %t3; \
+ stw %t3, [%dst - offset - 0x08];
+
+#define RMOVE_BIGALIGNCHUNK(src, dst, offset, t0, t1, t2, t3) \
+ ldx [%src - offset - 0x20], %t0; \
+ ldx [%src - offset - 0x18], %t1; \
+ ldx [%src - offset - 0x10], %t2; \
+ ldx [%src - offset - 0x08], %t3; \
+ stx %t0, [%dst - offset - 0x20]; \
+ stx %t1, [%dst - offset - 0x18]; \
+ stx %t2, [%dst - offset - 0x10]; \
+ stx %t3, [%dst - offset - 0x08]; \
+ ldx [%src - offset - 0x40], %t0; \
+ ldx [%src - offset - 0x38], %t1; \
+ ldx [%src - offset - 0x30], %t2; \
+ ldx [%src - offset - 0x28], %t3; \
+ stx %t0, [%dst - offset - 0x40]; \
+ stx %t1, [%dst - offset - 0x38]; \
+ stx %t2, [%dst - offset - 0x30]; \
+ stx %t3, [%dst - offset - 0x28];
+
+#define RMOVE_LASTCHUNK(src, dst, offset, t0, t1, t2, t3) \
+ ldx [%src + offset + 0x00], %t0; \
+ ldx [%src + offset + 0x08], %t1; \
+ stw %t0, [%dst + offset + 0x04]; \
+ srlx %t0, 32, %t2; \
+ stw %t2, [%dst + offset + 0x00]; \
+ stw %t1, [%dst + offset + 0x0c]; \
+ srlx %t1, 32, %t3; \
+ stw %t3, [%dst + offset + 0x08];
+
+#define RMOVE_LASTALIGNCHUNK(src, dst, offset, t0, t1) \
+ ldx [%src + offset + 0x00], %t0; \
+ ldx [%src + offset + 0x08], %t1; \
+ stx %t0, [%dst + offset + 0x00]; \
+ stx %t1, [%dst + offset + 0x08];
+
+ .align 32
+228: andcc %o2, 1, %g0 /* IEU1 Group */
+ be,pt %icc, 2f+4 /* CTI */
+1: ldub [%o1 - 1], %o5 /* LOAD Group */
+ sub %o1, 1, %o1 /* IEU0 */
+ sub %o0, 1, %o0 /* IEU1 */
+ subcc %o2, 1, %o2 /* IEU1 Group */
+ be,pn %xcc, 229f /* CTI */
+ stb %o5, [%o0] /* Store */
+2: ldub [%o1 - 1], %o5 /* LOAD Group */
+ sub %o0, 2, %o0 /* IEU0 */
+ ldub [%o1 - 2], %g5 /* LOAD Group */
+ sub %o1, 2, %o1 /* IEU0 */
+ subcc %o2, 2, %o2 /* IEU1 Group */
+ stb %o5, [%o0 + 1] /* Store */
+ bne,pt %xcc, 2b /* CTI */
+ stb %g5, [%o0] /* Store */
+229: retl
+ mov %g4, %o0
+
+ .align 32
+ENTRY(memmove)
+ mov %o0, %g5
+#ifndef USE_BPR
+ srl %o2, 0, %o2 /* IEU1 Group */
+#endif
+ brz,pn %o2, out /* CTI Group */
+ sub %o0, %o1, %o4 /* IEU0 */
+ cmp %o4, %o2 /* IEU1 Group */
+ bgeu,pt %XCC, 218b /* CTI */
+ mov %o0, %g4 /* IEU0 */
+ add %o0, %o2, %o0 /* IEU0 Group */
+220: add %o1, %o2, %o1 /* IEU1 */
+ cmp %o2, 15 /* IEU1 Group */
+ bleu,pn %xcc, 228b /* CTI */
+ andcc %o0, 7, %g2 /* IEU1 Group */
+ sub %o0, %o1, %g5 /* IEU0 */
+ andcc %g5, 3, %o5 /* IEU1 Group */
+ bne,pn %xcc, 232f /* CTI */
+ andcc %o1, 3, %g0 /* IEU1 Group */
+ be,a,pt %xcc, 236f /* CTI */
+ andcc %o1, 4, %g0 /* IEU1 Group */
+ andcc %o1, 1, %g0 /* IEU1 Group */
+ be,pn %xcc, 4f /* CTI */
+ andcc %o1, 2, %g0 /* IEU1 Group */
+ ldub [%o1 - 1], %g2 /* Load Group */
+ sub %o1, 1, %o1 /* IEU0 */
+ sub %o0, 1, %o0 /* IEU1 */
+ sub %o2, 1, %o2 /* IEU0 Group */
+ be,pn %xcc, 5f /* CTI Group */
+ stb %g2, [%o0] /* Store */
+4: lduh [%o1 - 2], %g2 /* Load Group */
+ sub %o1, 2, %o1 /* IEU0 */
+ sub %o0, 2, %o0 /* IEU1 */
+ sub %o2, 2, %o2 /* IEU0 */
+ sth %g2, [%o0] /* Store Group + bubble */
+5: andcc %o1, 4, %g0 /* IEU1 */
+236: be,a,pn %xcc, 2f /* CTI */
+ andcc %o2, -128, %g6 /* IEU1 Group */
+ lduw [%o1 - 4], %g5 /* Load Group */
+ sub %o1, 4, %o1 /* IEU0 */
+ sub %o0, 4, %o0 /* IEU1 */
+ sub %o2, 4, %o2 /* IEU0 Group */
+ stw %g5, [%o0] /* Store */
+ andcc %o2, -128, %g6 /* IEU1 Group */
+2: be,pn %xcc, 235f /* CTI */
+ andcc %o0, 4, %g0 /* IEU1 Group */
+ be,pn %xcc, 282f + 4 /* CTI Group */
+5: RMOVE_BIGCHUNK(o1, o0, 0x00, g1, g3, g5, o5)
+ RMOVE_BIGCHUNK(o1, o0, 0x20, g1, g3, g5, o5)
+ RMOVE_BIGCHUNK(o1, o0, 0x40, g1, g3, g5, o5)
+ RMOVE_BIGCHUNK(o1, o0, 0x60, g1, g3, g5, o5)
+ subcc %g6, 128, %g6 /* IEU1 Group */
+ sub %o1, 128, %o1 /* IEU0 */
+ bne,pt %xcc, 5b /* CTI */
+ sub %o0, 128, %o0 /* IEU0 Group */
+235: andcc %o2, 0x70, %g6 /* IEU1 Group */
+41: be,pn %xcc, 280f /* CTI */
+ andcc %o2, 8, %g0 /* IEU1 Group */
+ /* Clk1 8-( */
+ /* Clk2 8-( */
+ /* Clk3 8-( */
+ /* Clk4 8-( */
+279: rd %pc, %o5 /* PDU Group */
+ sll %g6, 1, %g5 /* IEU0 Group */
+ sub %o1, %g6, %o1 /* IEU1 */
+ sub %o5, %g5, %o5 /* IEU0 Group */
+ jmpl %o5 + %lo(280f - 279b), %g0 /* CTI Group brk forced*/
+ sub %o0, %g6, %o0 /* IEU0 Group */
+ RMOVE_LASTCHUNK(o1, o0, 0x60, g2, g3, g5, o5)
+ RMOVE_LASTCHUNK(o1, o0, 0x50, g2, g3, g5, o5)
+ RMOVE_LASTCHUNK(o1, o0, 0x40, g2, g3, g5, o5)
+ RMOVE_LASTCHUNK(o1, o0, 0x30, g2, g3, g5, o5)
+ RMOVE_LASTCHUNK(o1, o0, 0x20, g2, g3, g5, o5)
+ RMOVE_LASTCHUNK(o1, o0, 0x10, g2, g3, g5, o5)
+ RMOVE_LASTCHUNK(o1, o0, 0x00, g2, g3, g5, o5)
+280: be,pt %xcc, 281f /* CTI */
+ andcc %o2, 4, %g0 /* IEU1 */
+ ldx [%o1 - 8], %g2 /* Load Group */
+ sub %o0, 8, %o0 /* IEU0 */
+ stw %g2, [%o0 + 4] /* Store Group */
+ sub %o1, 8, %o1 /* IEU1 */
+ srlx %g2, 32, %g2 /* IEU0 Group */
+ stw %g2, [%o0] /* Store */
+281: be,pt %xcc, 1f /* CTI */
+ andcc %o2, 2, %g0 /* IEU1 Group */
+ lduw [%o1 - 4], %g2 /* Load Group */
+ sub %o1, 4, %o1 /* IEU0 */
+ stw %g2, [%o0 - 4] /* Store Group */
+ sub %o0, 4, %o0 /* IEU0 */
+1: be,pt %xcc, 1f /* CTI */
+ andcc %o2, 1, %g0 /* IEU1 Group */
+ lduh [%o1 - 2], %g2 /* Load Group */
+ sub %o1, 2, %o1 /* IEU0 */
+ sth %g2, [%o0 - 2] /* Store Group */
+ sub %o0, 2, %o0 /* IEU0 */
+1: be,pt %xcc, 211f /* CTI */
+ nop /* IEU1 */
+ ldub [%o1 - 1], %g2 /* Load Group */
+ stb %g2, [%o0 - 1] /* Store Group + bubble */
+211: retl
+ mov %g4, %o0
+
+282: RMOVE_BIGALIGNCHUNK(o1, o0, 0x00, g1, g3, g5, o5)
+ RMOVE_BIGALIGNCHUNK(o1, o0, 0x40, g1, g3, g5, o5)
+ subcc %g6, 128, %g6 /* IEU1 Group */
+ sub %o1, 128, %o1 /* IEU0 */
+ bne,pt %xcc, 282b /* CTI */
+ sub %o0, 128, %o0 /* IEU0 Group */
+ andcc %o2, 0x70, %g6 /* IEU1 */
+ be,pn %xcc, 284f /* CTI */
+ andcc %o2, 8, %g0 /* IEU1 Group */
+ /* Clk1 8-( */
+ /* Clk2 8-( */
+ /* Clk3 8-( */
+ /* Clk4 8-( */
+283: rd %pc, %o5 /* PDU Group */
+ sub %o1, %g6, %o1 /* IEU0 Group */
+ sub %o5, %g6, %o5 /* IEU1 */
+ jmpl %o5 + %lo(284f - 283b), %g0 /* CTI Group brk forced*/
+ sub %o0, %g6, %o0 /* IEU0 Group */
+ RMOVE_LASTALIGNCHUNK(o1, o0, 0x60, g2, g3)
+ RMOVE_LASTALIGNCHUNK(o1, o0, 0x50, g2, g3)
+ RMOVE_LASTALIGNCHUNK(o1, o0, 0x40, g2, g3)
+ RMOVE_LASTALIGNCHUNK(o1, o0, 0x30, g2, g3)
+ RMOVE_LASTALIGNCHUNK(o1, o0, 0x20, g2, g3)
+ RMOVE_LASTALIGNCHUNK(o1, o0, 0x10, g2, g3)
+ RMOVE_LASTALIGNCHUNK(o1, o0, 0x00, g2, g3)
+284: be,pt %xcc, 285f /* CTI Group */
+ andcc %o2, 4, %g0 /* IEU1 */
+ ldx [%o1 - 8], %g2 /* Load Group */
+ sub %o0, 8, %o0 /* IEU0 */
+ sub %o1, 8, %o1 /* IEU0 Group */
+ stx %g2, [%o0] /* Store */
+285: be,pt %xcc, 1f /* CTI */
+ andcc %o2, 2, %g0 /* IEU1 Group */
+ lduw [%o1 - 4], %g2 /* Load Group */
+ sub %o0, 4, %o0 /* IEU0 */
+ sub %o1, 4, %o1 /* IEU0 Group */
+ stw %g2, [%o0] /* Store */
+1: be,pt %xcc, 1f /* CTI */
+ andcc %o2, 1, %g0 /* IEU1 Group */
+ lduh [%o1 - 2], %g2 /* Load Group */
+ sub %o0, 2, %o0 /* IEU0 */
+ sub %o1, 2, %o1 /* IEU0 Group */
+ sth %g2, [%o0] /* Store */
+1: be,pt %xcc, 1f /* CTI */
+ nop /* IEU0 Group */
+ ldub [%o1 - 1], %g2 /* Load Group */
+ stb %g2, [%o0 - 1] /* Store Group + bubble */
+1: retl
+ mov %g4, %o0
+
+232: brz,pt %g2, 2f /* CTI Group */
+ sub %o2, %g2, %o2 /* IEU0 Group */
+1: ldub [%o1 - 1], %g5 /* Load Group */
+ sub %o1, 1, %o1 /* IEU0 */
+ sub %o0, 1, %o0 /* IEU1 */
+ subcc %g2, 1, %g2 /* IEU1 Group */
+ bne,pt %xcc, 1b /* CTI */
+ stb %g5, [%o0] /* Store */
+2: andn %o2, 7, %g5 /* IEU0 Group */
+ and %o2, 7, %o2 /* IEU1 */
+ fmovd %f0, %f2 /* FPU */
+ alignaddr %o1, %g0, %g1 /* GRU Group */
+ ldd [%g1], %f4 /* Load Group */
+1: ldd [%g1 - 8], %f6 /* Load Group */
+ sub %g1, 8, %g1 /* IEU0 Group */
+ subcc %g5, 8, %g5 /* IEU1 */
+ faligndata %f6, %f4, %f0 /* GRU Group */
+ std %f0, [%o0 - 8] /* Store */
+ sub %o1, 8, %o1 /* IEU0 Group */
+ be,pn %xcc, 233f /* CTI */
+ sub %o0, 8, %o0 /* IEU1 */
+ ldd [%g1 - 8], %f4 /* Load Group */
+ sub %g1, 8, %g1 /* IEU0 */
+ subcc %g5, 8, %g5 /* IEU1 */
+ faligndata %f4, %f6, %f0 /* GRU Group */
+ std %f0, [%o0 - 8] /* Store */
+ sub %o1, 8, %o1 /* IEU0 */
+ bne,pn %xcc, 1b /* CTI Group */
+ sub %o0, 8, %o0 /* IEU0 */
+233: brz,pn %o2, 234f /* CTI Group */
+ nop /* IEU0 */
+237: ldub [%o1 - 1], %g5 /* LOAD */
+ sub %o1, 1, %o1 /* IEU0 */
+ sub %o0, 1, %o0 /* IEU1 */
+ subcc %o2, 1, %o2 /* IEU1 */
+ bne,pt %xcc, 237b /* CTI */
+ stb %g5, [%o0] /* Store Group */
+234: wr %g0, FPRS_FEF, %fprs
+ retl
+ mov %g4, %o0
+END(memmove)
+
+#ifdef USE_BPR
+weak_alias (memcpy, __align_cpy_1)
+weak_alias (memcpy, __align_cpy_2)
+weak_alias (memcpy, __align_cpy_4)
+weak_alias (memcpy, __align_cpy_8)
+weak_alias (memcpy, __align_cpy_16)
+#endif
diff --git a/libc/string/sparc/sparc64/stpcpy.S b/libc/string/sparc/sparc64/stpcpy.S
new file mode 100644
index 000000000..a66f9d152
--- /dev/null
+++ b/libc/string/sparc/sparc64/stpcpy.S
@@ -0,0 +1,272 @@
+/* Copy SRC to DEST returning the address of the terminating '\0' in DEST.
+ For SPARC v9.
+ Copyright (C) 1998, 1999, 2002, 2003, 2004 Free Software Foundation, Inc.
+ This file is part of the GNU C Library.
+ Contributed by Jan Vondrak <jvon4518@ss1000.ms.mff.cuni.cz> and
+ Jakub Jelinek <jj@ultra.linux.cz>.
+
+ The GNU C Library is free software; you can redistribute it and/or
+ modify it under the terms of the GNU Lesser General Public
+ License as published by the Free Software Foundation; either
+ version 2.1 of the License, or (at your option) any later version.
+
+ The GNU C Library is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ Lesser General Public License for more details.
+
+ You should have received a copy of the GNU Lesser General Public
+ License along with the GNU C Library; if not, write to the Free
+ Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA
+ 02111-1307 USA. */
+
+#include <asm/asi.h>
+#ifndef XCC
+ .register %g2, #scratch
+ .register %g3, #scratch
+ .register %g6, #scratch
+#endif
+
+ /* Normally, this uses
+ ((xword - 0x0101010101010101) & 0x8080808080808080) test
+ to find out if any byte in xword could be zero. This is fast, but
+ also gives false alarm for any byte in range 0x81-0xff. It does
+ not matter for correctness, as if this test tells us there could
+ be some zero byte, we check it byte by byte, but if bytes with
+ high bits set are common in the strings, then this will give poor
+ performance. You can #define EIGHTBIT_NOT_RARE and the algorithm
+ will use one tick slower, but more precise test
+ ((xword - 0x0101010101010101) & (~xword) & 0x8080808080808080),
+ which does not give any false alarms (but if some bits are set,
+ one cannot assume from it which bytes are zero and which are not).
+ It is yet to be measured, what is the correct default for glibc
+ in these days for an average user.
+ */
+
+ .text
+ .align 32
+ENTRY(__stpcpy)
+ sethi %hi(0x01010101), %g1 /* IEU0 Group */
+ or %g1, %lo(0x01010101), %g1 /* IEU0 Group */
+ andcc %o0, 7, %g0 /* IEU1 */
+ sllx %g1, 32, %g2 /* IEU0 Group */
+
+ bne,pn %icc, 12f /* CTI */
+ andcc %o1, 7, %g3 /* IEU1 */
+ or %g1, %g2, %g1 /* IEU0 Group */
+ bne,pn %icc, 14f /* CTI */
+
+ sllx %g1, 7, %g2 /* IEU0 Group */
+1: ldx [%o1], %o3 /* Load */
+ add %o1, 8, %o1 /* IEU1 */
+2: mov %o3, %g3 /* IEU0 Group */
+
+ sub %o3, %g1, %o2 /* IEU1 */
+3: ldxa [%o1] ASI_PNF, %o3 /* Load */
+#ifdef EIGHTBIT_NOT_RARE
+ andn %o2, %g3, %o2 /* IEU0 Group */
+#endif
+ add %o0, 8, %o0 /* IEU0 Group */
+ andcc %o2, %g2, %g0 /* IEU1 */
+
+ add %o1, 8, %o1 /* IEU0 Group */
+ be,a,pt %xcc, 2b /* CTI */
+ stx %g3, [%o0 - 8] /* Store */
+ srlx %g3, 56, %g5 /* IEU0 Group */
+
+ andcc %g5, 0xff, %g0 /* IEU1 Group */
+ be,pn %icc, 11f /* CTI */
+ srlx %g3, 48, %g4 /* IEU0 */
+ andcc %g4, 0xff, %g0 /* IEU1 Group */
+
+ be,pn %icc, 10f /* CTI */
+ srlx %g3, 40, %g5 /* IEU0 */
+ andcc %g5, 0xff, %g0 /* IEU1 Group */
+ be,pn %icc, 9f /* CTI */
+
+ srlx %g3, 32, %g4 /* IEU0 */
+ andcc %g4, 0xff, %g0 /* IEU1 Group */
+ be,pn %icc, 8f /* CTI */
+ srlx %g3, 24, %g5 /* IEU0 */
+
+ andcc %g5, 0xff, %g0 /* IEU1 Group */
+ be,pn %icc, 7f /* CTI */
+ srlx %g3, 16, %g4 /* IEU0 */
+ andcc %g4, 0xff, %g0 /* IEU1 Group */
+
+ be,pn %icc, 6f /* CTI */
+ srlx %g3, 8, %g5 /* IEU0 */
+ andcc %g5, 0xff, %g0 /* IEU1 Group */
+ be,pn %icc, 5f /* CTI */
+
+ sub %o3, %g1, %o2 /* IEU0 */
+ stx %g3, [%o0 - 8] /* Store Group */
+ andcc %g3, 0xff, %g0 /* IEU1 */
+ bne,pt %icc, 3b /* CTI */
+
+ mov %o3, %g3 /* IEU0 Group */
+4: retl /* CTI+IEU1 Group */
+ sub %o0, 1, %o0 /* IEU0 */
+
+ .align 16
+6: ba,pt %xcc, 23f /* CTI Group */
+ sub %o0, 3, %g6 /* IEU0 */
+5: sub %o0, 2, %g6 /* IEU0 Group */
+ stb %g5, [%o0 - 2] /* Store */
+
+ srlx %g3, 16, %g4 /* IEU0 Group */
+23: sth %g4, [%o0 - 4] /* Store */
+ srlx %g3, 32, %g4 /* IEU0 Group */
+ stw %g4, [%o0 - 8] /* Store */
+
+ retl /* CTI+IEU1 Group */
+ mov %g6, %o0 /* IEU0 */
+8: ba,pt %xcc, 24f /* CTI Group */
+ sub %o0, 5, %g6 /* IEU0 */
+
+7: sub %o0, 4, %g6 /* IEU0 Group */
+ stb %g5, [%o0 - 4] /* Store */
+ srlx %g3, 32, %g4 /* IEU0 Group */
+24: stw %g4, [%o0 - 8] /* Store */
+
+ retl /* CTI+IEU1 Group */
+ mov %g6, %o0 /* IEU0 */
+10: ba,pt %xcc, 25f /* CTI Group */
+ sub %o0, 7, %g6 /* IEU0 */
+
+9: sub %o0, 6, %g6 /* IEU0 Group */
+ stb %g5, [%o0 - 6] /* Store */
+ srlx %g3, 48, %g4 /* IEU0 */
+25: sth %g4, [%o0 - 8] /* Store Group */
+
+ retl /* CTI+IEU1 Group */
+ mov %g6, %o0 /* IEU0 */
+11: stb %g5, [%o0 - 8] /* Store Group */
+ retl /* CTI+IEU1 Group */
+
+ sub %o0, 8, %o0 /* IEU0 */
+
+ .align 16
+12: or %g1, %g2, %g1 /* IEU0 Group */
+ ldub [%o1], %o3 /* Load */
+ sllx %g1, 7, %g2 /* IEU0 Group */
+ stb %o3, [%o0] /* Store Group */
+
+13: add %o0, 1, %o0 /* IEU0 */
+ add %o1, 1, %o1 /* IEU1 */
+ andcc %o3, 0xff, %g0 /* IEU1 Group */
+ be,pn %icc, 4b /* CTI */
+
+ lduba [%o1] ASI_PNF, %o3 /* Load */
+ andcc %o0, 7, %g0 /* IEU1 Group */
+ bne,a,pt %icc, 13b /* CTI */
+ stb %o3, [%o0] /* Store */
+
+ andcc %o1, 7, %g3 /* IEU1 Group */
+ be,a,pt %icc, 1b /* CTI */
+ ldx [%o1], %o3 /* Load */
+14: orcc %g0, 64, %g4 /* IEU1 Group */
+
+ sllx %g3, 3, %g5 /* IEU0 */
+ sub %o1, %g3, %o1 /* IEU0 Group */
+ sub %g4, %g5, %g4 /* IEU1 */
+ /* %g1 = 0101010101010101 *
+ * %g2 = 8080808080808080 *
+ * %g3 = source alignment *
+ * %g5 = number of bits to shift left *
+ * %g4 = number of bits to shift right */
+ ldxa [%o1] ASI_PNF, %o5 /* Load Group */
+
+ addcc %o1, 8, %o1 /* IEU1 */
+15: sllx %o5, %g5, %o3 /* IEU0 Group */
+ ldxa [%o1] ASI_PNF, %o5 /* Load */
+ srlx %o5, %g4, %o4 /* IEU0 Group */
+
+ add %o0, 8, %o0 /* IEU1 */
+ or %o3, %o4, %o3 /* IEU0 Group */
+ add %o1, 8, %o1 /* IEU1 */
+ sub %o3, %g1, %o4 /* IEU0 Group */
+
+#ifdef EIGHTBIT_NOT_RARE
+ andn %o4, %o3, %o4 /* IEU0 Group */
+#endif
+ andcc %o4, %g2, %g0 /* IEU1 Group */
+ be,a,pt %xcc, 15b /* CTI */
+ stx %o3, [%o0 - 8] /* Store */
+ srlx %o3, 56, %o4 /* IEU0 Group */
+
+ andcc %o4, 0xff, %g0 /* IEU1 Group */
+ be,pn %icc, 22f /* CTI */
+ srlx %o3, 48, %o4 /* IEU0 */
+ andcc %o4, 0xff, %g0 /* IEU1 Group */
+
+ be,pn %icc, 21f /* CTI */
+ srlx %o3, 40, %o4 /* IEU0 */
+ andcc %o4, 0xff, %g0 /* IEU1 Group */
+ be,pn %icc, 20f /* CTI */
+
+ srlx %o3, 32, %o4 /* IEU0 */
+ andcc %o4, 0xff, %g0 /* IEU1 Group */
+ be,pn %icc, 19f /* CTI */
+ srlx %o3, 24, %o4 /* IEU0 */
+
+ andcc %o4, 0xff, %g0 /* IEU1 Group */
+ be,pn %icc, 18f /* CTI */
+ srlx %o3, 16, %o4 /* IEU0 */
+ andcc %o4, 0xff, %g0 /* IEU1 Group */
+
+ be,pn %icc, 17f /* CTI */
+ srlx %o3, 8, %o4 /* IEU0 */
+ andcc %o4, 0xff, %g0 /* IEU1 Group */
+ be,pn %icc, 16f /* CTI */
+
+ andcc %o3, 0xff, %g0 /* IEU1 Group */
+ bne,pn %icc, 15b /* CTI */
+ stx %o3, [%o0 - 8] /* Store */
+ retl /* CTI+IEU1 Group */
+
+ sub %o0, 1, %o0 /* IEU0 */
+
+ .align 16
+17: ba,pt %xcc, 26f /* CTI Group */
+ subcc %o0, 3, %g6 /* IEU1 */
+18: ba,pt %xcc, 27f /* CTI Group */
+ subcc %o0, 4, %g6 /* IEU1 */
+
+19: ba,pt %xcc, 28f /* CTI Group */
+ subcc %o0, 5, %g6 /* IEU1 */
+16: subcc %o0, 2, %g6 /* IEU1 Group */
+ srlx %o3, 8, %o4 /* IEU0 */
+
+ stb %o4, [%o0 - 2] /* Store */
+26: srlx %o3, 16, %o4 /* IEU0 Group */
+ stb %o4, [%o0 - 3] /* Store */
+27: srlx %o3, 24, %o4 /* IEU0 Group */
+
+ stb %o4, [%o0 - 4] /* Store */
+28: srlx %o3, 32, %o4 /* IEU0 Group */
+ stw %o4, [%o0 - 8] /* Store */
+ retl /* CTI+IEU1 Group */
+
+ mov %g6, %o0 /* IEU0 */
+
+ .align 16
+21: ba,pt %xcc, 29f /* CTI Group */
+ subcc %o0, 7, %g6 /* IEU1 */
+22: ba,pt %xcc, 30f /* CTI Group */
+ subcc %o0, 8, %g6 /* IEU1 */
+
+20: subcc %o0, 6, %g6 /* IEU1 Group */
+ srlx %o3, 40, %o4 /* IEU0 */
+ stb %o4, [%o0 - 6] /* Store */
+29: srlx %o3, 48, %o4 /* IEU0 Group */
+
+ stb %o4, [%o0 - 7] /* Store */
+30: srlx %o3, 56, %o4 /* IEU0 Group */
+ stb %o4, [%o0 - 8] /* Store */
+ retl /* CTI+IEU1 Group */
+
+ mov %g6, %o0 /* IEU0 */
+END(__stpcpy)
+
+weak_alias (__stpcpy, stpcpy)
diff --git a/libc/string/sparc/sparc64/strcat.S b/libc/string/sparc/sparc64/strcat.S
new file mode 100644
index 000000000..21ff119c4
--- /dev/null
+++ b/libc/string/sparc/sparc64/strcat.S
@@ -0,0 +1,338 @@
+/* strcat (dest, src) -- Append SRC on the end of DEST.
+ For SPARC v9.
+ Copyright (C) 1998, 1999, 2003 Free Software Foundation, Inc.
+ This file is part of the GNU C Library.
+ Contributed by Jakub Jelinek <jj@ultra.linux.cz> and
+ Jan Vondrak <jvon4518@ss1000.ms.mff.cuni.cz>.
+
+ The GNU C Library is free software; you can redistribute it and/or
+ modify it under the terms of the GNU Lesser General Public
+ License as published by the Free Software Foundation; either
+ version 2.1 of the License, or (at your option) any later version.
+
+ The GNU C Library is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ Lesser General Public License for more details.
+
+ You should have received a copy of the GNU Lesser General Public
+ License along with the GNU C Library; if not, write to the Free
+ Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA
+ 02111-1307 USA. */
+
+#include <asm/asi.h>
+#ifndef XCC
+#define XCC xcc
+#define USE_BPR
+ .register %g2, #scratch
+ .register %g3, #scratch
+ .register %g6, #scratch
+#endif
+
+ /* Normally, this uses
+ ((xword - 0x0101010101010101) & 0x8080808080808080) test
+ to find out if any byte in xword could be zero. This is fast, but
+ also gives false alarm for any byte in range 0x81-0xff. It does
+ not matter for correctness, as if this test tells us there could
+ be some zero byte, we check it byte by byte, but if bytes with
+ high bits set are common in the strings, then this will give poor
+ performance. You can #define EIGHTBIT_NOT_RARE and the algorithm
+ will use one tick slower, but more precise test
+ ((xword - 0x0101010101010101) & (~xword) & 0x8080808080808080),
+ which does not give any false alarms (but if some bits are set,
+ one cannot assume from it which bytes are zero and which are not).
+ It is yet to be measured, what is the correct default for glibc
+ in these days for an average user.
+ */
+
+ .text
+ .align 32
+ENTRY(strcat)
+ sethi %hi(0x01010101), %g1 /* IEU0 Group */
+ ldub [%o0], %o3 /* Load */
+ or %g1, %lo(0x01010101), %g1 /* IEU0 Group */
+ mov %o0, %g6 /* IEU1 */
+
+ sllx %g1, 32, %g2 /* IEU0 Group */
+ andcc %o0, 7, %g0 /* IEU1 */
+ or %g1, %g2, %g1 /* IEU0 Group */
+ bne,pn %icc, 32f /* CTI */
+
+ sllx %g1, 7, %g2 /* IEU0 Group */
+ brz,pn %o3, 30f /* CTI+IEU1 */
+ ldx [%o0], %o3 /* Load */
+48: add %o0, 8, %o0 /* IEU0 Group */
+
+49: sub %o3, %g1, %o2 /* IEU0 Group */
+#ifdef EIGHTBIT_NOT_RARE
+ andn %o2, %o3, %g5 /* IEU0 Group */
+ ldxa [%o0] ASI_PNF, %o3 /* Load */
+ andcc %g5, %g2, %g0 /* IEU1 Group */
+#else
+ ldxa [%o0] ASI_PNF, %o3 /* Load */
+ andcc %o2, %g2, %g0 /* IEU1 Group */
+#endif
+ be,pt %xcc, 49b /* CTI */
+
+ add %o0, 8, %o0 /* IEU0 */
+ addcc %o2, %g1, %g3 /* IEU1 Group */
+ srlx %o2, 32, %o2 /* IEU0 */
+50: andcc %o2, %g2, %g0 /* IEU1 Group */
+
+ be,pn %xcc, 51f /* CTI */
+ srlx %g3, 56, %o2 /* IEU0 */
+ andcc %o2, 0xff, %g0 /* IEU1 Group */
+ be,pn %icc, 29f /* CTI */
+
+ srlx %g3, 48, %o2 /* IEU0 */
+ andcc %o2, 0xff, %g0 /* IEU1 Group */
+ be,pn %icc, 28f /* CTI */
+ srlx %g3, 40, %o2 /* IEU0 */
+
+ andcc %o2, 0xff, %g0 /* IEU1 Group */
+ be,pn %icc, 27f /* CTI */
+ srlx %g3, 32, %o2 /* IEU0 */
+ andcc %o2, 0xff, %g0 /* IEU1 Group */
+
+ be,pn %icc, 26f /* CTI */
+51: srlx %g3, 24, %o2 /* IEU0 */
+ andcc %o2, 0xff, %g0 /* IEU1 Group */
+ be,pn %icc, 25f /* CTI */
+
+ srlx %g3, 16, %o2 /* IEU0 */
+ andcc %o2, 0xff, %g0 /* IEU1 Group */
+ be,pn %icc, 24f /* CTI */
+ srlx %g3, 8, %o2 /* IEU0 */
+
+ andcc %o2, 0xff, %g0 /* IEU1 Group */
+ be,pn %icc, 23f /* CTI */
+ sub %o3, %g1, %o2 /* IEU0 */
+ andcc %g3, 0xff, %g0 /* IEU1 Group */
+
+ be,pn %icc, 52f /* CTI */
+ ldxa [%o0] ASI_PNF, %o3 /* Load */
+ andcc %o2, %g2, %g0 /* IEU1 Group */
+ be,pt %xcc, 49b /* CTI */
+
+ add %o0, 8, %o0 /* IEU0 */
+ addcc %o2, %g1, %g3 /* IEU1 Group */
+ ba,pt %xcc, 50b /* CTI */
+ srlx %o2, 32, %o2 /* IEU0 */
+
+ .align 16
+52: ba,pt %xcc, 12f /* CTI Group */
+ add %o0, -9, %o0 /* IEU0 */
+23: ba,pt %xcc, 12f /* CTI Group */
+ add %o0, -10, %o0 /* IEU0 */
+
+24: ba,pt %xcc, 12f /* CTI Group */
+ add %o0, -11, %o0 /* IEU0 */
+25: ba,pt %xcc, 12f /* CTI Group */
+ add %o0, -12, %o0 /* IEU0 */
+
+26: ba,pt %xcc, 12f /* CTI Group */
+ add %o0, -13, %o0 /* IEU0 */
+27: ba,pt %xcc, 12f /* CTI Group */
+ add %o0, -14, %o0 /* IEU0 */
+
+28: ba,pt %xcc, 12f /* CTI Group */
+ add %o0, -15, %o0 /* IEU0 */
+29: add %o0, -16, %o0 /* IEU0 Group */
+30: andcc %o1, 7, %g3 /* IEU1 */
+
+31: bne,pn %icc, 14f /* CTI */
+ orcc %g0, 64, %g4 /* IEU1 Group */
+1: ldx [%o1], %o3 /* Load */
+ add %o1, 8, %o1 /* IEU1 */
+
+2: mov %o3, %g3 /* IEU0 Group */
+3: sub %o3, %g1, %o2 /* IEU1 */
+ ldxa [%o1] ASI_PNF, %o3 /* Load */
+#ifdef EIGHTBIT_NOT_RARE
+ andn %o2, %g3, %o2 /* IEU0 Group */
+#endif
+ add %o0, 8, %o0 /* IEU0 Group */
+
+ andcc %o2, %g2, %g0 /* IEU1 */
+ add %o1, 8, %o1 /* IEU0 Group */
+ be,a,pt %xcc, 2b /* CTI */
+ stx %g3, [%o0 - 8] /* Store */
+
+ srlx %g3, 56, %g5 /* IEU0 Group */
+ andcc %g5, 0xff, %g0 /* IEU1 Group */
+ be,pn %icc, 11f /* CTI */
+ srlx %g3, 48, %g4 /* IEU0 */
+
+ andcc %g4, 0xff, %g0 /* IEU1 Group */
+ be,pn %icc, 10f /* CTI */
+ srlx %g3, 40, %g5 /* IEU0 */
+ andcc %g5, 0xff, %g0 /* IEU1 Group */
+
+ be,pn %icc, 9f /* CTI */
+ srlx %g3, 32, %g4 /* IEU0 */
+ andcc %g4, 0xff, %g0 /* IEU1 Group */
+ be,pn %icc, 8f /* CTI */
+
+ srlx %g3, 24, %g5 /* IEU0 */
+ andcc %g5, 0xff, %g0 /* IEU1 Group */
+ be,pn %icc, 7f /* CTI */
+ srlx %g3, 16, %g4 /* IEU0 */
+
+ andcc %g4, 0xff, %g0 /* IEU1 Group */
+ be,pn %icc, 6f /* CTI */
+ srlx %g3, 8, %g5 /* IEU0 */
+ andcc %g5, 0xff, %g0 /* IEU1 Group */
+
+ be,pn %icc, 5f /* CTI */
+ sub %o3, %g1, %o2 /* IEU0 */
+ stx %g3, [%o0 - 8] /* Store Group */
+ andcc %g3, 0xff, %g0 /* IEU1 */
+
+ bne,pt %icc, 3b /* CTI */
+ mov %o3, %g3 /* IEU0 Group */
+4: retl /* CTI+IEU1 Group */
+ mov %g6, %o0 /* IEU0 */
+
+ .align 16
+5: stb %g5, [%o0 - 2] /* Store Group */
+ srlx %g3, 16, %g4 /* IEU0 */
+6: sth %g4, [%o0 - 4] /* Store Group */
+ srlx %g3, 32, %g4 /* IEU0 */
+
+ stw %g4, [%o0 - 8] /* Store Group */
+ retl /* CTI+IEU1 Group */
+ mov %g6, %o0 /* IEU0 */
+7: stb %g5, [%o0 - 4] /* Store Group */
+
+ srlx %g3, 32, %g4 /* IEU0 */
+8: stw %g4, [%o0 - 8] /* Store Group */
+ retl /* CTI+IEU1 Group */
+ mov %g6, %o0 /* IEU0 */
+
+9: stb %g5, [%o0 - 6] /* Store Group */
+ srlx %g3, 48, %g4 /* IEU0 */
+10: sth %g4, [%o0 - 8] /* Store Group */
+ retl /* CTI+IEU1 Group */
+
+ mov %g6, %o0 /* IEU0 */
+11: stb %g5, [%o0 - 8] /* Store Group */
+ retl /* CTI+IEU1 Group */
+ mov %g6, %o0 /* IEU0 */
+
+ .align 16
+32: andcc %o0, 7, %g0 /* IEU1 Group */
+ be,a,pn %icc, 48b /* CTI */
+ ldx [%o0], %o3 /* Load */
+ add %o0, 1, %o0 /* IEU0 Group */
+
+ brnz,a,pt %o3, 32b /* CTI+IEU1 */
+ lduba [%o0] ASI_PNF, %o3 /* Load */
+ add %o0, -1, %o0 /* IEU0 Group */
+ andcc %o0, 7, %g0 /* IEU1 Group */
+
+ be,a,pn %icc, 31b /* CTI */
+ andcc %o1, 7, %g3 /* IEU1 Group */
+12: ldub [%o1], %o3 /* Load */
+ stb %o3, [%o0] /* Store Group */
+
+13: add %o0, 1, %o0 /* IEU0 */
+ add %o1, 1, %o1 /* IEU1 */
+ andcc %o3, 0xff, %g0 /* IEU1 Group */
+ be,pn %icc, 4b /* CTI */
+
+ lduba [%o1] ASI_PNF, %o3 /* Load */
+ andcc %o0, 7, %g0 /* IEU1 Group */
+ bne,a,pt %icc, 13b /* CTI */
+ stb %o3, [%o0] /* Store */
+
+ andcc %o1, 7, %g3 /* IEU1 Group */
+ be,a,pt %icc, 1b /* CTI */
+ ldx [%o1], %o3 /* Load */
+ orcc %g0, 64, %g4 /* IEU1 Group */
+
+14: sllx %g3, 3, %g5 /* IEU0 */
+ sub %o1, %g3, %o1 /* IEU0 Group */
+ sub %g4, %g5, %g4 /* IEU1 */
+ /* %g1 = 0101010101010101 *
+ * %g2 = 8080808080808080 *
+ * %g3 = source alignment *
+ * %g5 = number of bits to shift left *
+ * %g4 = number of bits to shift right */
+ ldxa [%o1] ASI_PNF, %o5 /* Load Group */
+
+ addcc %o1, 8, %o1 /* IEU1 */
+15: sllx %o5, %g5, %o3 /* IEU0 Group */
+ ldxa [%o1] ASI_PNF, %o5 /* Load */
+ srlx %o5, %g4, %o4 /* IEU0 Group */
+
+ add %o0, 8, %o0 /* IEU1 */
+ or %o3, %o4, %o3 /* IEU0 Group */
+ add %o1, 8, %o1 /* IEU1 */
+ sub %o3, %g1, %o4 /* IEU0 Group */
+
+#ifdef EIGHTBIT_NOT_RARE
+ andn %o4, %o3, %o4 /* IEU0 Group */
+#endif
+ andcc %o4, %g2, %g0 /* IEU1 Group */
+ be,a,pt %xcc, 15b /* CTI */
+ stx %o3, [%o0 - 8] /* Store */
+ srlx %o3, 56, %o4 /* IEU0 Group */
+
+ andcc %o4, 0xff, %g0 /* IEU1 Group */
+ be,pn %icc, 22f /* CTI */
+ srlx %o3, 48, %o4 /* IEU0 */
+ andcc %o4, 0xff, %g0 /* IEU1 Group */
+
+ be,pn %icc, 21f /* CTI */
+ srlx %o3, 40, %o4 /* IEU0 */
+ andcc %o4, 0xff, %g0 /* IEU1 Group */
+ be,pn %icc, 20f /* CTI */
+
+ srlx %o3, 32, %o4 /* IEU0 */
+ andcc %o4, 0xff, %g0 /* IEU1 Group */
+ be,pn %icc, 19f /* CTI */
+ srlx %o3, 24, %o4 /* IEU0 */
+
+ andcc %o4, 0xff, %g0 /* IEU1 Group */
+ be,pn %icc, 18f /* CTI */
+ srlx %o3, 16, %o4 /* IEU0 */
+ andcc %o4, 0xff, %g0 /* IEU1 Group */
+
+ be,pn %icc, 17f /* CTI */
+ srlx %o3, 8, %o4 /* IEU0 */
+ andcc %o4, 0xff, %g0 /* IEU1 Group */
+ be,pn %icc, 16f /* CTI */
+
+ andcc %o3, 0xff, %g0 /* IEU1 Group */
+ bne,pn %icc, 15b /* CTI */
+ stx %o3, [%o0 - 8] /* Store */
+ retl /* CTI+IEU1 Group */
+
+ mov %g6, %o0 /* IEU0 */
+
+ .align 16
+16: srlx %o3, 8, %o4 /* IEU0 Group */
+ stb %o4, [%o0 - 2] /* Store */
+17: srlx %o3, 16, %o4 /* IEU0 Group */
+ stb %o4, [%o0 - 3] /* Store */
+
+18: srlx %o3, 24, %o4 /* IEU0 Group */
+ stb %o4, [%o0 - 4] /* Store */
+19: srlx %o3, 32, %o4 /* IEU0 Group */
+ stw %o4, [%o0 - 8] /* Store */
+
+ retl /* CTI+IEU1 Group */
+ mov %g6, %o0 /* IEU0 */
+ nop
+ nop
+
+20: srlx %o3, 40, %o4 /* IEU0 Group */
+ stb %o4, [%o0 - 6] /* Store */
+21: srlx %o3, 48, %o4 /* IEU0 Group */
+ stb %o4, [%o0 - 7] /* Store */
+
+22: srlx %o3, 56, %o4 /* IEU0 Group */
+ stb %o4, [%o0 - 8] /* Store */
+ retl /* CTI+IEU1 Group */
+ mov %g6, %o0 /* IEU0 */
+END(strcat)
diff --git a/libc/string/sparc/sparc64/strchr.S b/libc/string/sparc/sparc64/strchr.S
new file mode 100644
index 000000000..784b7fe4d
--- /dev/null
+++ b/libc/string/sparc/sparc64/strchr.S
@@ -0,0 +1,480 @@
+/* strchr (str, ch) -- Return pointer to first occurrence of CH in STR.
+ For SPARC v9.
+ Copyright (C) 1998, 1999, 2003 Free Software Foundation, Inc.
+ This file is part of the GNU C Library.
+ Contributed by Jan Vondrak <jvon4518@ss1000.ms.mff.cuni.cz> and
+ Jakub Jelinek <jj@ultra.linux.cz>.
+
+ The GNU C Library is free software; you can redistribute it and/or
+ modify it under the terms of the GNU Lesser General Public
+ License as published by the Free Software Foundation; either
+ version 2.1 of the License, or (at your option) any later version.
+
+ The GNU C Library is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ Lesser General Public License for more details.
+
+ You should have received a copy of the GNU Lesser General Public
+ License along with the GNU C Library; if not, write to the Free
+ Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA
+ 02111-1307 USA. */
+
+#include <asm/asi.h>
+#ifndef XCC
+#define XCC xcc
+#define USE_BPR
+ .register %g2, #scratch
+ .register %g3, #scratch
+ .register %g6, #scratch
+#endif
+
+ /* Normally, this uses
+ ((xword - 0x0101010101010101) & 0x8080808080808080) test
+ to find out if any byte in xword could be zero. This is fast, but
+ also gives false alarm for any byte in range 0x81-0xff. It does
+ not matter for correctness, as if this test tells us there could
+ be some zero byte, we check it byte by byte, but if bytes with
+ high bits set are common in the strings, then this will give poor
+ performance. You can #define EIGHTBIT_NOT_RARE and the algorithm
+ will use one tick slower, but more precise test
+ ((xword - 0x0101010101010101) & (~xword) & 0x8080808080808080),
+ which does not give any false alarms (but if some bits are set,
+ one cannot assume from it which bytes are zero and which are not).
+ It is yet to be measured, what is the correct default for glibc
+ in these days for an average user.
+ */
+
+ .text
+ .align 32
+ENTRY(strchr)
+ andcc %o1, 0xff, %o1 /* IEU1 Group */
+ be,pn %icc, 17f /* CTI */
+ sllx %o1, 8, %g3 /* IEU0 Group */
+ sethi %hi(0x01010101), %g1 /* IEU1 */
+
+ or %g3, %o1, %g3 /* IEU0 Group */
+ ldub [%o0], %o3 /* Load */
+ sllx %g3, 16, %g5 /* IEU0 Group */
+ or %g1, %lo(0x01010101), %g1 /* IEU1 */
+
+ sllx %g1, 32, %g2 /* IEU0 Group */
+ brz,pn %o3, 5f /* CTI+IEU1 */
+ orcc %g3, %g5, %g3 /* IEU1 Group */
+ sllx %g3, 32, %g5 /* IEU0 */
+
+ cmp %o3, %o1 /* IEU1 Group */
+ be,pn %xcc, 14f /* CTI */
+ or %g1, %g2, %g1 /* IEU0 */
+ andcc %o0, 7, %g0 /* IEU1 Group */
+
+ bne,a,pn %icc, 15f /* CTI */
+ add %o0, 1, %o0 /* IEU0 */
+ ldx [%o0], %o3 /* Load Group */
+1: sllx %g1, 7, %g2 /* IEU0 */
+
+ or %g3, %g5, %g3 /* IEU1 */
+ add %o0, 8, %o0 /* IEU0 Group */
+ xor %o3, %g3, %o4 /* IEU1 */
+ /* %g1 = 0101010101010101 *
+ * %g2 = 8080088080808080 *
+ * %g3 = c c c c c c c c *
+ * %o3 = value *
+ * %o4 = value XOR c */
+2: sub %o3, %g1, %o2 /* IEU0 Group */
+
+ sub %o4, %g1, %o5 /* IEU1 */
+#ifdef EIGHTBIT_NOT_RARE
+ andn %o2, %o3, %g6 /* IEU0 Group */
+ andn %o5, %o4, %o5 /* IEU1 */
+ ldxa [%o0] ASI_PNF, %o3 /* Load */
+ or %o5, %g6, %o5 /* IEU0 Group */
+#else
+ ldxa [%o0] ASI_PNF, %o3 /* Load */
+ or %o5, %o2, %o5 /* IEU0 Group */
+#endif
+ add %o0, 8, %o0 /* IEU1 */
+
+ andcc %o5, %g2, %g0 /* IEU1 Group */
+ be,a,pt %xcc, 2b /* CTI */
+ xor %o3, %g3, %o4 /* IEU0 */
+ srlx %o5, 32, %g5 /* IEU0 Group */
+
+ add %o2, %g1, %o2 /* IEU1 */
+3: andcc %g5, %g2, %g0 /* IEU1 Group */
+ be,pn %xcc, 4f /* CTI */
+ srlx %o2, 56, %g5 /* IEU0 */
+
+ andcc %g5, 0xff, %g0 /* IEU1 Group */
+ be,pn %icc, 5f /* CTI */
+ srlx %o4, 56, %g5 /* IEU0 */
+ andcc %g5, 0xff, %g0 /* IEU1 Group */
+
+ be,pn %icc, 6f /* CTI */
+ srlx %o2, 48, %g5 /* IEU0 */
+ andcc %g5, 0xff, %g0 /* IEU1 Group */
+ be,pn %icc, 5f /* CTI */
+
+ srlx %o4, 48, %g5 /* IEU0 */
+ andcc %g5, 0xff, %g0 /* IEU1 Group */
+ be,pn %icc, 7f /* CTI */
+ srlx %o2, 40, %g5 /* IEU0 */
+
+ andcc %g5, 0xff, %g0 /* IEU1 Group */
+ be,pn %icc, 5f /* CTI */
+ srlx %o4, 40, %g5 /* IEU0 */
+ andcc %g5, 0xff, %g0 /* IEU1 Group */
+
+ be,pn %icc, 8f /* CTI */
+ srlx %o2, 32, %g5 /* IEU0 */
+ andcc %g5, 0xff, %g0 /* IEU1 Group */
+ be,pn %icc, 5f /* CTI */
+
+ srlx %o4, 32, %g5 /* IEU0 */
+ andcc %g5, 0xff, %g0 /* IEU1 Group */
+ be,pn %icc, 9f /* CTI */
+4: srlx %o2, 24, %g5 /* IEU0 */
+
+ andcc %g5, 0xff, %g0 /* IEU1 Group */
+ be,pn %icc, 5f /* CTI */
+ srlx %o4, 24, %g5 /* IEU0 */
+ andcc %g5, 0xff, %g0 /* IEU1 Group */
+
+ be,pn %icc, 10f /* CTI */
+ srlx %o2, 16, %g5 /* IEU0 */
+ andcc %g5, 0xff, %g0 /* IEU1 Group */
+ be,pn %icc, 5f /* CTI */
+
+ srlx %o4, 16, %g5 /* IEU0 */
+ andcc %g5, 0xff, %g0 /* IEU1 Group */
+ be,pn %icc, 11f /* CTI */
+ srlx %o2, 8, %g5 /* IEU0 */
+
+ andcc %g5, 0xff, %g0 /* IEU1 Group */
+ be,pn %icc, 5f /* CTI */
+ srlx %o4, 8, %g5 /* IEU0 */
+ andcc %g5, 0xff, %g0 /* IEU1 Group */
+
+ be,pn %icc, 12f /* CTI */
+ andcc %o2, 0xff, %g0 /* IEU1 Group */
+ be,pn %icc, 5f /* CTI */
+ sub %o3, %g1, %o2 /* IEU0 */
+
+ andcc %o4, 0xff, %g0 /* IEU1 Group */
+ be,pn %icc, 13f /* CTI */
+ xor %o3, %g3, %o4 /* IEU0 */
+ ldxa [%o0] ASI_PNF, %o3 /* Load Group */
+
+ sub %o4, %g1, %o5 /* IEU0 */
+ or %o5, %o2, %o5 /* IEU1 */
+ add %o0, 8, %o0 /* IEU0 Group */
+ andcc %o5, %g2, %g0 /* IEU1 */
+
+ be,a,pt %xcc, 2b /* CTI */
+ xor %o3, %g3, %o4 /* IEU0 Group */
+ srlx %o5, 32, %g5 /* IEU0 Group */
+ ba,pt %xcc, 3b /* CTI */
+
+ add %o2, %g1, %o2 /* IEU1 */
+
+ .align 16
+5: retl /* CTI+IEU1 Group */
+ clr %o0 /* IEU0 */
+6: retl /* CTI+IEU1 Group */
+ add %o0, -16, %o0 /* IEU0 */
+
+7: retl /* CTI+IEU1 Group */
+ add %o0, -15, %o0 /* IEU0 */
+8: retl /* CTI+IEU1 Group */
+ add %o0, -14, %o0 /* IEU0 */
+
+9: retl /* CTI+IEU1 Group */
+ add %o0, -13, %o0 /* IEU0 */
+10: retl /* CTI+IEU1 Group */
+ add %o0, -12, %o0 /* IEU0 */
+
+11: retl /* CTI+IEU1 Group */
+ add %o0, -11, %o0 /* IEU0 */
+12: retl /* CTI+IEU1 Group */
+ add %o0, -10, %o0 /* IEU0 */
+
+13: retl /* CTI+IEU1 Group */
+ add %o0, -9, %o0 /* IEU0 */
+14: retl /* CTI+IEU1 Group */
+ nop /* IEU0 */
+
+ .align 16
+15: ldub [%o0], %o3 /* Load Group */
+16: andcc %o0, 7, %g0 /* IEU1 */
+ be,a,pn %icc, 1b /* CTI */
+ ldx [%o0], %o3 /* Load Group */
+
+ andcc %o3, 0xff, %g0 /* IEU1 Group */
+ be,pn %icc, 5b /* CTI */
+ add %o0, 1, %o0 /* IEU0 */
+ cmp %o3, %o1 /* IEU1 Group */
+
+ bne,a,pn %icc, 16b /* CTI */
+ ldub [%o0], %o3 /* Load */
+ retl /* CTI+IEU1 Group */
+ add %o0, -1, %o0 /* IEU0 */
+
+ /* strchr (str, 0) */
+ .align 32
+ nop
+ .align 16
+17: sethi %hi(0x01010101), %g1 /* IEU0 Group */
+ ldub [%o0], %o3 /* Load */
+ or %g1, %lo(0x01010101), %g1 /* IEU0 Group */
+ sllx %g1, 32, %g2 /* IEU0 Group */
+
+ andcc %o0, 7, %g0 /* IEU1 */
+ or %g1, %g2, %g1 /* IEU0 Group */
+ bne,pn %icc, 32f /* CTI */
+ sllx %g1, 7, %g2 /* IEU0 Group */
+
+ brz,pn %o3, 30f /* CTI+IEU1 */
+ ldx [%o0], %o3 /* Load */
+18: add %o0, 8, %o0 /* IEU0 Group */
+19: sub %o3, %g1, %o2 /* IEU0 Group */
+
+#ifdef EIGHTBIT_NOT_RARE
+ andn %o2, %o3, %g6 /* IEU0 Group */
+ ldxa [%o0] ASI_PNF, %o3 /* Load */
+ andcc %g6, %g2, %g0 /* IEU1 Group */
+#else
+ ldxa [%o0] ASI_PNF, %o3 /* Load */
+ andcc %o2, %g2, %g0 /* IEU1 Group */
+#endif
+ be,pt %xcc, 19b /* CTI */
+ add %o0, 8, %o0 /* IEU0 */
+
+ addcc %o2, %g1, %g3 /* IEU1 Group */
+ srlx %o2, 32, %o2 /* IEU0 */
+20: andcc %o2, %g2, %g0 /* IEU1 Group */
+ be,pn %xcc, 21f /* CTI */
+
+ srlx %g3, 56, %o2 /* IEU0 */
+ andcc %o2, 0xff, %g0 /* IEU1 Group */
+ be,pn %icc, 29f /* CTI */
+ srlx %g3, 48, %o2 /* IEU0 */
+
+ andcc %o2, 0xff, %g0 /* IEU1 Group */
+ be,pn %icc, 28f /* CTI */
+ srlx %g3, 40, %o2 /* IEU0 */
+ andcc %o2, 0xff, %g0 /* IEU1 Group */
+
+ be,pn %icc, 27f /* CTI */
+ srlx %g3, 32, %o2 /* IEU0 */
+ andcc %o2, 0xff, %g0 /* IEU1 Group */
+ be,pn %icc, 26f /* CTI */
+
+21: srlx %g3, 24, %o2 /* IEU0 */
+ andcc %o2, 0xff, %g0 /* IEU1 Group */
+ be,pn %icc, 25f /* CTI */
+ srlx %g3, 16, %o2 /* IEU0 */
+
+ andcc %o2, 0xff, %g0 /* IEU1 Group */
+ be,pn %icc, 24f /* CTI */
+ srlx %g3, 8, %o2 /* IEU0 */
+ andcc %o2, 0xff, %g0 /* IEU1 Group */
+
+ be,pn %icc, 23f /* CTI */
+ sub %o3, %g1, %o2 /* IEU0 */
+ andcc %g3, 0xff, %g0 /* IEU1 Group */
+ be,pn %icc, 22f /* CTI */
+
+ ldxa [%o0] ASI_PNF, %o3 /* Load */
+ andcc %o2, %g2, %g0 /* IEU1 Group */
+ be,pt %xcc, 19b /* CTI */
+ add %o0, 8, %o0 /* IEU0 */
+
+ addcc %o2, %g1, %g3 /* IEU1 Group */
+ ba,pt %xcc, 20b /* CTI */
+ srlx %o2, 32, %o2 /* IEU0 */
+
+ .align 16
+22: retl /* CTI+IEU1 Group */
+ add %o0, -9, %o0 /* IEU0 */
+23: retl /* CTI+IEU1 Group */
+ add %o0, -10, %o0 /* IEU0 */
+
+24: retl /* CTI+IEU1 Group */
+ add %o0, -11, %o0 /* IEU0 */
+25: retl /* CTI+IEU1 Group */
+ add %o0, -12, %o0 /* IEU0 */
+
+26: retl /* CTI+IEU1 Group */
+ add %o0, -13, %o0 /* IEU0 */
+27: retl /* CTI+IEU1 Group */
+ add %o0, -14, %o0 /* IEU0 */
+
+28: retl /* CTI+IEU1 Group */
+ add %o0, -15, %o0 /* IEU0 */
+29: retl /* CTI+IEU1 Group */
+ add %o0, -16, %o0 /* IEU0 */
+
+30: retl /* CTI+IEU1 Group */
+ nop /* IEU0 */
+
+ .align 16
+32: andcc %o0, 7, %g0 /* IEU1 Group */
+ be,a,pn %icc, 18b /* CTI */
+ ldx [%o0], %o3 /* Load */
+ add %o0, 1, %o0 /* IEU0 Group */
+
+ brnz,a,pt %o3, 32b /* CTI+IEU1 */
+ lduba [%o0] ASI_PNF, %o3 /* Load */
+ retl /* CTI+IEU1 Group */
+ add %o0, -1, %o0 /* IEU0 */
+END(strchr)
+
+ .align 32
+ENTRY(strrchr)
+ andcc %o1, 0xff, %o1 /* IEU1 Group */
+ be,pn %icc, 17b /* CTI */
+ clr %g4 /* IEU0 */
+ andcc %o0, 7, %g0 /* IEU1 Group */
+
+ bne,pn %icc, 13f /* CTI */
+ sllx %o1, 8, %g3 /* IEU0 */
+ ldx [%o0], %o3 /* Load Group */
+1: sethi %hi(0x01010101), %g1 /* IEU0 */
+
+ or %g3, %o1, %g3 /* IEU1 */
+ sllx %g3, 16, %g5 /* IEU0 Group */
+ or %g1, %lo(0x01010101), %g1 /* IEU1 */
+ sllx %g1, 32, %g2 /* IEU0 Group */
+
+ or %g3, %g5, %g3 /* IEU1 */
+ sllx %g3, 32, %g5 /* IEU0 Group */
+ or %g1, %g2, %g1 /* IEU1 */
+ sllx %g1, 7, %g2 /* IEU0 Group */
+
+ or %g3, %g5, %g3 /* IEU1 */
+ add %o0, 8, %o0 /* IEU0 Group */
+ xor %o3, %g3, %o4 /* IEU1 */
+ /* %g1 = 0101010101010101 *
+ * %g2 = 8080088080808080 *
+ * %g3 = c c c c c c c c *
+ * %o3 = value *
+ * %o4 = value XOR c */
+2: sub %o3, %g1, %o2 /* IEU0 Group */
+
+3: sub %o4, %g1, %o5 /* IEU1 */
+#ifdef EIGHTBIT_NOT_RARE
+ andn %o2, %o3, %g6 /* IEU0 Group */
+ andn %o5, %o4, %o5 /* IEU1 */
+ ldxa [%o0] ASI_PNF, %o3 /* Load */
+
+ or %o5, %g6, %o5 /* IEU0 Group */
+#else
+ ldxa [%o0] ASI_PNF, %o3 /* Load */
+
+ or %o5, %o2, %o5 /* IEU0 Group */
+#endif
+ add %o0, 8, %o0 /* IEU1 */
+ andcc %o5, %g2, %g0 /* IEU1 Group */
+ be,a,pt %xcc, 2b /* CTI */
+
+ xor %o3, %g3, %o4 /* IEU0 */
+ srlx %o5, 32, %g5 /* IEU0 Group */
+ add %o2, %g1, %o2 /* IEU1 */
+ andcc %g5, %g2, %g0 /* IEU1 Group */
+
+ be,pn %xcc, 7f /* CTI */
+ srlx %o2, 56, %g5 /* IEU0 */
+ andcc %g5, 0xff, %g0 /* IEU1 Group */
+ be,pn %icc, 12f /* CTI */
+
+ srlx %o4, 56, %g5 /* IEU0 */
+ andcc %g5, 0xff, %g0 /* IEU1 Group */
+ srlx %o2, 48, %g5 /* IEU0 */
+ be,a,pn %icc, 4f /* CTI */
+
+ add %o0, -16, %g4 /* IEU0 Group */
+4: andcc %g5, 0xff, %g0 /* IEU1 Group */
+ be,pn %icc, 12f /* CTI */
+ srlx %o4, 48, %g5 /* IEU0 */
+
+ andcc %g5, 0xff, %g0 /* IEU1 Group */
+ srlx %o2, 40, %g5 /* IEU0 */
+ be,a,pn %icc, 5f /* CTI */
+ add %o0, -15, %g4 /* IEU0 Group */
+
+5: andcc %g5, 0xff, %g0 /* IEU1 Group */
+ be,pn %icc, 12f /* CTI */
+ srlx %o4, 40, %g5 /* IEU0 */
+ andcc %g5, 0xff, %g0 /* IEU1 Group */
+
+ srlx %o2, 32, %g5 /* IEU0 */
+ be,a,pn %icc, 6f /* CTI */
+ add %o0, -14, %g4 /* IEU0 Group */
+6: andcc %g5, 0xff, %g0 /* IEU1 Group */
+
+ be,pn %icc, 12f /* CTI */
+ srlx %o4, 32, %g5 /* IEU0 */
+ andcc %g5, 0xff, %g0 /* IEU1 Group */
+ be,a,pn %icc, 7f /* CTI */
+
+ add %o0, -13, %g4 /* IEU0 */
+7: srlx %o2, 24, %g5 /* IEU0 */
+ andcc %g5, 0xff, %g0 /* IEU1 Group */
+ be,pn %icc, 12f /* CTI */
+
+ srlx %o4, 24, %g5 /* IEU0 */
+ andcc %g5, 0xff, %g0 /* IEU1 Group */
+ srlx %o2, 16, %g5 /* IEU0 */
+ be,a,pn %icc, 8f /* CTI */
+
+ add %o0, -12, %g4 /* IEU0 Group */
+8: andcc %g5, 0xff, %g0 /* IEU1 Group */
+ be,pn %icc, 12f /* CTI */
+ srlx %o4, 16, %g5 /* IEU0 */
+
+ andcc %g5, 0xff, %g0 /* IEU1 Group */
+ srlx %o2, 8, %g5 /* IEU0 */
+ be,a,pn %icc, 9f /* CTI */
+ add %o0, -11, %g4 /* IEU0 Group */
+
+9: andcc %g5, 0xff, %g0 /* IEU1 Group */
+ be,pn %icc, 12f /* CTI */
+ srlx %o4, 8, %g5 /* IEU0 */
+ andcc %g5, 0xff, %g0 /* IEU1 Group */
+
+ be,a,pn %icc, 10f /* CTI */
+ add %o0, -10, %g4 /* IEU0 */
+10: andcc %o2, 0xff, %g0 /* IEU1 Group */
+ be,pn %icc, 12f /* CTI */
+
+ sub %o3, %g1, %o2 /* IEU0 */
+ andcc %o4, 0xff, %g0 /* IEU1 Group */
+ be,a,pn %icc, 11f /* CTI */
+ add %o0, -9, %g4 /* IEU0 */
+
+11: ba,pt %xcc, 3b /* CTI Group */
+ xor %o3, %g3, %o4 /* IEU0 Group */
+12: retl /* CTI+IEU1 Group */
+ mov %g4, %o0 /* IEU0 */
+
+ .align 16
+13: ldub [%o0], %o3 /* Load Group */
+ add %o0, 1, %o0 /* IEU0 */
+14: andcc %o3, 0xff, %g0 /* IEU1 Group */
+ be,pn %icc, 12b /* CTI */
+
+ cmp %o3, %o1 /* IEU1 Group */
+ ldub [%o0], %o3 /* Load */
+ be,a,pn %icc, 15f /* CTI */
+ add %o0, -1, %g4 /* IEU0 Group */
+
+15: andcc %o0, 7, %g0 /* IEU1 Group */
+ bne,a,pt %icc, 14b /* CTI */
+ add %o0, 1, %o0 /* IEU0 */
+ ba,pt %xcc, 1b /* CTI Group */
+
+ ldx [%o0], %o3 /* Load */
+END(strrchr)
+
+weak_alias (strchr, index)
+weak_alias (strrchr, rindex)
diff --git a/libc/string/sparc/sparc64/strcmp.S b/libc/string/sparc/sparc64/strcmp.S
new file mode 100644
index 000000000..5866778d3
--- /dev/null
+++ b/libc/string/sparc/sparc64/strcmp.S
@@ -0,0 +1,278 @@
+/* Compare two strings for differences.
+ For SPARC v9.
+ Copyright (C) 1997, 1999, 2003 Free Software Foundation, Inc.
+ This file is part of the GNU C Library.
+ Contributed by Jan Vondrak <jvon4518@ss1000.ms.mff.cuni.cz> and
+ Jakub Jelinek <jj@ultra.linux.cz>.
+
+ The GNU C Library is free software; you can redistribute it and/or
+ modify it under the terms of the GNU Lesser General Public
+ License as published by the Free Software Foundation; either
+ version 2.1 of the License, or (at your option) any later version.
+
+ The GNU C Library is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ Lesser General Public License for more details.
+
+ You should have received a copy of the GNU Lesser General Public
+ License along with the GNU C Library; if not, write to the Free
+ Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA
+ 02111-1307 USA. */
+
+#include <asm/asi.h>
+#ifndef XCC
+ .register %g2, #scratch
+ .register %g3, #scratch
+ .register %g6, #scratch
+#endif
+
+ /* Normally, this uses
+ ((xword - 0x0101010101010101) & 0x8080808080808080) test
+ to find out if any byte in xword could be zero. This is fast, but
+ also gives false alarm for any byte in range 0x81-0xff. It does
+ not matter for correctness, as if this test tells us there could
+ be some zero byte, we check it byte by byte, but if bytes with
+ high bits set are common in the strings, then this will give poor
+ performance. You can #define EIGHTBIT_NOT_RARE and the algorithm
+ will use one tick slower, but more precise test
+ ((xword - 0x0101010101010101) & (~xword) & 0x8080808080808080),
+ which does not give any false alarms (but if some bits are set,
+ one cannot assume from it which bytes are zero and which are not).
+ It is yet to be measured, what is the correct default for glibc
+ in these days for an average user.
+ */
+
+ .text
+ .align 32
+ENTRY(strcmp)
+ sethi %hi(0x01010101), %g1 /* IEU0 Group */
+ andcc %o0, 7, %g0 /* IEU1 */
+ bne,pn %icc, 7f /* CTI */
+ or %g1, %lo(0x01010101), %g1 /* IEU0 Group */
+
+ andcc %o1, 7, %g3 /* IEU1 */
+ bne,pn %icc, 9f /* CTI */
+ sllx %g1, 32, %g2 /* IEU0 Group */
+ ldx [%o0], %o2 /* Load */
+
+ or %g1, %g2, %g1 /* IEU0 Group */
+1: ldx [%o1], %o3 /* Load */
+ sub %o1, %o0, %o1 /* IEU1 */
+ sllx %g1, 7, %g2 /* IEU0 Group */
+
+2: add %o0, 8, %o0 /* IEU1 */
+ sub %o2, %g1, %g3 /* IEU0 Group */
+ subcc %o2, %o3, %g0 /* IEU1 */
+ bne,pn %xcc, 13f /* CTI */
+
+#ifdef EIGHTBIT_NOT_RARE
+ andn %g3, %o2, %g4 /* IEU0 Group */
+ ldxa [%o0] ASI_PNF, %o2 /* Load */
+ andcc %g4, %g2, %g0 /* IEU1 Group */
+#else
+ ldxa [%o0] ASI_PNF, %o2 /* Load Group */
+ andcc %g3, %g2, %g0 /* IEU1 */
+#endif
+ be,a,pt %xcc, 2b /* CTI */
+ ldxa [%o1 + %o0] ASI_PNF, %o3 /* Load Group */
+
+ addcc %g3, %g1, %o4 /* IEU1 */
+ srlx %g3, 32, %g3 /* IEU0 */
+ andcc %g3, %g2, %g0 /* IEU1 Group */
+ be,pt %xcc, 3f /* CTI */
+
+ srlx %o4, 56, %o5 /* IEU0 */
+ andcc %o5, 0xff, %g0 /* IEU1 Group */
+ be,pn %icc, 4f /* CTI */
+ srlx %o4, 48, %o5 /* IEU0 */
+
+ andcc %o5, 0xff, %g0 /* IEU1 Group */
+ be,pn %icc, 4f /* CTI */
+ srlx %o4, 40, %o5 /* IEU0 */
+ andcc %o5, 0xff, %g0 /* IEU1 Group */
+
+ be,pn %icc, 4f /* CTI */
+ srlx %o4, 32, %o5 /* IEU0 */
+ andcc %o5, 0xff, %g0 /* IEU1 Group */
+ be,pn %icc, 4f /* CTI */
+
+3: srlx %o4, 24, %o5 /* IEU0 */
+ andcc %o5, 0xff, %g0 /* IEU1 Group */
+ be,pn %icc, 4f /* CTI */
+ srlx %o4, 16, %o5 /* IEU0 */
+
+ andcc %o5, 0xff, %g0 /* IEU1 Group */
+ be,pn %icc, 4f /* CTI */
+ srlx %o4, 8, %o5 /* IEU0 */
+ andcc %o5, 0xff, %g0 /* IEU1 Group */
+
+ be,pn %icc, 4f /* CTI */
+ andcc %o4, 0xff, %g0 /* IEU1 Group */
+ bne,a,pn %icc, 2b /* CTI */
+ ldxa [%o1 + %o0] ASI_PNF, %o3 /* Load */
+
+4: retl /* CTI+IEU1 Group */
+ clr %o0 /* IEU0 */
+
+ .align 32
+13: mov 0xff, %g6 /* IEU0 Group */
+#ifdef EIGHTBIT_NOT_RARE
+ andcc %g4, %g2, %g0 /* IEU1 */
+#else
+ andcc %g3, %g2, %g0 /* IEU1 */
+#endif
+ be,pt %xcc, 25f /* CTI */
+ addcc %g3, %g1, %o4 /* IEU1 Group */
+
+ srlx %g3, 32, %g3 /* IEU0 */
+ andcc %g3, %g2, %g0 /* IEU1 Group */
+ be,pt %xcc, 23f /* CTI */
+ sllx %g6, 56, %o5 /* IEU0 */
+
+ andcc %o4, %o5, %g0 /* IEU1 Group */
+ be,pn %xcc, 24f /* CTI */
+ sllx %g6, 48, %o5 /* IEU0 */
+ andcc %o4, %o5, %g0 /* IEU1 Group */
+
+ be,pn %xcc, 24f /* CTI */
+ sllx %g6, 40, %o5 /* IEU0 */
+ andcc %o4, %o5, %g0 /* IEU1 Group */
+ be,pn %xcc, 24f /* CTI */
+
+ sllx %g6, 32, %o5 /* IEU0 */
+ andcc %o4, %o5, %g0 /* IEU1 Group */
+ be,pn %xcc, 24f /* CTI */
+23: sllx %g6, 24, %o5 /* IEU0 */
+
+ andcc %o4, %o5, %g0 /* IEU1 Group */
+ be,pn %icc, 24f /* CTI */
+ sllx %g6, 16, %o5 /* IEU0 */
+ andcc %o4, %o5, %g0 /* IEU1 Group */
+
+ be,pn %icc, 24f /* CTI */
+ sllx %g6, 8, %o5 /* IEU0 */
+ andcc %o4, %o5, %g0 /* IEU1 Group */
+ be,pn %icc, 24f /* CTI */
+
+ mov %g6, %o5 /* IEU0 */
+25: cmp %o4, %o3 /* IEU1 Group */
+5: mov -1, %o0 /* IEU0 */
+ retl /* CTI+IEU1 Group */
+
+ movgu %xcc, 1, %o0 /* Single Group */
+
+ .align 16
+24: sub %o5, 1, %g6 /* IEU0 Group */
+ clr %o0 /* IEU1 */
+ or %o5, %g6, %o5 /* IEU0 Group */
+ andn %o4, %o5, %o4 /* IEU0 Group */
+
+ andn %o3, %o5, %o3 /* IEU1 */
+ cmp %o4, %o3 /* IEU1 Group */
+ movgu %xcc, 1, %o0 /* Single Group */
+ retl /* CTI+IEU1 Group */
+
+ movlu %xcc, -1, %o0 /* Single Group */
+6: retl /* CTI+IEU1 Group */
+ mov %o4, %o0 /* IEU0 */
+
+ .align 16
+7: ldub [%o0], %o2 /* Load */
+ add %o0, 1, %o0 /* IEU1 */
+ ldub [%o1], %o3 /* Load Group */
+ sllx %g1, 32, %g2 /* IEU0 */
+
+8: add %o1, 1, %o1 /* IEU1 */
+ subcc %o2, %o3, %o4 /* IEU1 Group */
+ bne,pn %xcc, 6b /* CTI */
+ lduba [%o0] ASI_PNF, %o2 /* Load */
+
+ brz,pn %o3, 4b /* CTI+IEU1 Group */
+ lduba [%o1] ASI_PNF, %o3 /* Load */
+ andcc %o0, 7, %g0 /* IEU1 Group */
+ bne,a,pn %icc, 8b /* CTI */
+
+ add %o0, 1, %o0 /* IEU0 */
+ or %g1, %g2, %g1 /* IEU0 Group */
+ andcc %o1, 7, %g3 /* IEU1 */
+ be,a,pn %icc, 1b /* CTI */
+
+ ldxa [%o0] ASI_PNF, %o2 /* Load Group */
+9: sllx %g3, 3, %g5 /* IEU0 */
+ mov 64, %o5 /* IEU1 */
+ sub %o1, %g3, %o1 /* IEU0 Group */
+
+ sub %o5, %g5, %o5 /* IEU1 */
+ ldxa [%o1] ASI_PNF, %g6 /* Load Group */
+ or %g1, %g2, %g1 /* IEU0 */
+ sub %o1, %o0, %o1 /* IEU1 */
+
+ sllx %g1, 7, %g2 /* IEU0 Group */
+ add %o1, 8, %o1 /* IEU1 */
+ /* %g1 = 0101010101010101
+ * %g2 = 8080808080800880
+ * %g5 = number of bits to shift left
+ * %o5 = number of bits to shift right */
+10: sllx %g6, %g5, %o3 /* IEU0 Group */
+ ldxa [%o1 + %o0] ASI_PNF, %g6 /* Load */
+
+11: srlx %g6, %o5, %o4 /* IEU0 Group */
+ ldxa [%o0] ASI_PNF, %o2 /* Load */
+ or %o3, %o4, %o3 /* IEU1 */
+ add %o0, 8, %o0 /* IEU0 Group */
+
+ subcc %o2, %o3, %g0 /* IEU1 */
+#ifdef EIGHTBIT_NOT_RARE
+ sub %o2, %g1, %g3 /* IEU0 Group */
+ bne,pn %xcc, 13b /* CTI */
+ andn %g3, %o2, %g4 /* IEU0 Group */
+
+ andcc %g4, %g2, %g0 /* IEU1 Group */
+ be,pt %xcc, 10b /* CTI */
+ srlx %g4, 32, %g4 /* IEU0 */
+ andcc %g4, %g2, %g0 /* IEU1 Group */
+#else
+ bne,pn %xcc, 13b /* CTI */
+ sub %o2, %g1, %g3 /* IEU0 Group */
+ andcc %g3, %g2, %g0 /* IEU1 Group */
+
+ be,pt %xcc, 10b /* CTI */
+ srlx %g3, 32, %g3 /* IEU0 */
+ andcc %g3, %g2, %g0 /* IEU1 Group */
+#endif
+ be,pt %xcc, 12f /* CTI */
+
+ srlx %o2, 56, %g3 /* IEU0 */
+ andcc %g3, 0xff, %g0 /* IEU1 Group */
+ be,pn %icc, 4b /* CTI */
+ srlx %o2, 48, %g3 /* IEU0 */
+
+ andcc %g3, 0xff, %g0 /* IEU1 Group */
+ be,pn %icc, 4b /* CTI */
+ srlx %o2, 40, %g3 /* IEU0 */
+ andcc %g3, 0xff, %g0 /* IEU1 Group */
+
+ be,pn %icc, 4b /* CTI */
+ srlx %o2, 32, %g3 /* IEU0 */
+ andcc %g3, 0xff, %g0 /* IEU1 Group */
+ be,pn %icc, 4b /* CTI */
+
+12: srlx %o2, 24, %g3 /* IEU0 */
+ andcc %g3, 0xff, %g0 /* IEU1 Group */
+ be,pn %icc, 4b /* CTI */
+ srlx %o2, 16, %g3 /* IEU0 */
+
+ andcc %g3, 0xff, %g0 /* IEU1 Group */
+ be,pn %icc, 4b /* CTI */
+ srlx %o2, 8, %g3 /* IEU0 */
+ andcc %g3, 0xff, %g0 /* IEU1 Group */
+
+ be,pn %icc, 4b /* CTI */
+ andcc %o2, 0xff, %g0 /* IEU1 Group */
+ be,pn %icc, 4b /* CTI */
+ sllx %g6, %g5, %o3 /* IEU0 */
+
+ ba,pt %xcc, 11b /* CTI Group */
+ ldxa [%o1 + %o0] ASI_PNF, %g6 /* Load */
+END(strcmp)
diff --git a/libc/string/sparc/sparc64/strcpy.S b/libc/string/sparc/sparc64/strcpy.S
new file mode 100644
index 000000000..aa5d2aee3
--- /dev/null
+++ b/libc/string/sparc/sparc64/strcpy.S
@@ -0,0 +1,243 @@
+/* Copy SRC to DEST returning DEST.
+ For SPARC v9.
+ Copyright (C) 1998, 1999, 2003 Free Software Foundation, Inc.
+ This file is part of the GNU C Library.
+ Contributed by Jan Vondrak <jvon4518@ss1000.ms.mff.cuni.cz> and
+ Jakub Jelinek <jj@ultra.linux.cz>.
+
+ The GNU C Library is free software; you can redistribute it and/or
+ modify it under the terms of the GNU Lesser General Public
+ License as published by the Free Software Foundation; either
+ version 2.1 of the License, or (at your option) any later version.
+
+ The GNU C Library is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ Lesser General Public License for more details.
+
+ You should have received a copy of the GNU Lesser General Public
+ License along with the GNU C Library; if not, write to the Free
+ Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA
+ 02111-1307 USA. */
+
+#include <asm/asi.h>
+#ifndef XCC
+ .register %g2, #scratch
+ .register %g3, #scratch
+ .register %g6, #scratch
+#endif
+
+ /* Normally, this uses
+ ((xword - 0x0101010101010101) & 0x8080808080808080) test
+ to find out if any byte in xword could be zero. This is fast, but
+ also gives false alarm for any byte in range 0x81-0xff. It does
+ not matter for correctness, as if this test tells us there could
+ be some zero byte, we check it byte by byte, but if bytes with
+ high bits set are common in the strings, then this will give poor
+ performance. You can #define EIGHTBIT_NOT_RARE and the algorithm
+ will use one tick slower, but more precise test
+ ((xword - 0x0101010101010101) & (~xword) & 0x8080808080808080),
+ which does not give any false alarms (but if some bits are set,
+ one cannot assume from it which bytes are zero and which are not).
+ It is yet to be measured, what is the correct default for glibc
+ in these days for an average user.
+ */
+
+ .text
+ .align 32
+ENTRY(strcpy)
+ sethi %hi(0x01010101), %g1 /* IEU0 Group */
+ mov %o0, %g6 /* IEU1 */
+ or %g1, %lo(0x01010101), %g1 /* IEU0 Group */
+ andcc %o0, 7, %g0 /* IEU1 */
+
+ sllx %g1, 32, %g2 /* IEU0 Group */
+ bne,pn %icc, 12f /* CTI */
+ andcc %o1, 7, %g3 /* IEU1 */
+ or %g1, %g2, %g1 /* IEU0 Group */
+
+ bne,pn %icc, 14f /* CTI */
+ sllx %g1, 7, %g2 /* IEU0 Group */
+1: ldx [%o1], %o3 /* Load */
+ add %o1, 8, %o1 /* IEU1 */
+
+2: mov %o3, %g3 /* IEU0 Group */
+3: sub %o3, %g1, %o2 /* IEU1 */
+ ldxa [%o1] ASI_PNF, %o3 /* Load */
+#ifdef EIGHTBIT_NOT_RARE
+ andn %o2, %g3, %o2 /* IEU0 Group */
+#endif
+ add %o0, 8, %o0 /* IEU0 Group */
+
+ andcc %o2, %g2, %g0 /* IEU1 */
+ add %o1, 8, %o1 /* IEU0 Group */
+ be,a,pt %xcc, 2b /* CTI */
+ stx %g3, [%o0 - 8] /* Store */
+
+ srlx %g3, 56, %g5 /* IEU0 Group */
+ andcc %g5, 0xff, %g0 /* IEU1 Group */
+ be,pn %icc, 11f /* CTI */
+ srlx %g3, 48, %g4 /* IEU0 */
+
+ andcc %g4, 0xff, %g0 /* IEU1 Group */
+ be,pn %icc, 10f /* CTI */
+ srlx %g3, 40, %g5 /* IEU0 */
+ andcc %g5, 0xff, %g0 /* IEU1 Group */
+
+ be,pn %icc, 9f /* CTI */
+ srlx %g3, 32, %g4 /* IEU0 */
+ andcc %g4, 0xff, %g0 /* IEU1 Group */
+ be,pn %icc, 8f /* CTI */
+
+ srlx %g3, 24, %g5 /* IEU0 */
+ andcc %g5, 0xff, %g0 /* IEU1 Group */
+ be,pn %icc, 7f /* CTI */
+ srlx %g3, 16, %g4 /* IEU0 */
+
+ andcc %g4, 0xff, %g0 /* IEU1 Group */
+ be,pn %icc, 6f /* CTI */
+ srlx %g3, 8, %g5 /* IEU0 */
+ andcc %g5, 0xff, %g0 /* IEU1 Group */
+
+ be,pn %icc, 5f /* CTI */
+ sub %o3, %g1, %o2 /* IEU0 */
+ stx %g3, [%o0 - 8] /* Store Group */
+ andcc %g3, 0xff, %g0 /* IEU1 */
+
+ bne,pt %icc, 3b /* CTI */
+ mov %o3, %g3 /* IEU0 Group */
+4: retl /* CTI+IEU1 Group */
+ mov %g6, %o0 /* IEU0 */
+
+ .align 16
+5: stb %g5, [%o0 - 2] /* Store Group */
+ srlx %g3, 16, %g4 /* IEU0 */
+6: sth %g4, [%o0 - 4] /* Store Group */
+ srlx %g3, 32, %g4 /* IEU0 */
+
+ stw %g4, [%o0 - 8] /* Store Group */
+ retl /* CTI+IEU1 Group */
+ mov %g6, %o0 /* IEU0 */
+7: stb %g5, [%o0 - 4] /* Store Group */
+
+ srlx %g3, 32, %g4 /* IEU0 */
+8: stw %g4, [%o0 - 8] /* Store Group */
+ retl /* CTI+IEU1 Group */
+ mov %g6, %o0 /* IEU0 */
+
+9: stb %g5, [%o0 - 6] /* Store Group */
+ srlx %g3, 48, %g4 /* IEU0 */
+10: sth %g4, [%o0 - 8] /* Store Group */
+ retl /* CTI+IEU1 Group */
+
+ mov %g6, %o0 /* IEU0 */
+11: stb %g5, [%o0 - 8] /* Store Group */
+ retl /* CTI+IEU1 Group */
+ mov %g6, %o0 /* IEU0 */
+
+12: or %g1, %g2, %g1 /* IEU0 Group */
+ ldub [%o1], %o3 /* Load */
+ sllx %g1, 7, %g2 /* IEU0 Group */
+ stb %o3, [%o0] /* Store Group */
+
+13: add %o0, 1, %o0 /* IEU0 */
+ add %o1, 1, %o1 /* IEU1 */
+ andcc %o3, 0xff, %g0 /* IEU1 Group */
+ be,pn %icc, 4b /* CTI */
+
+ lduba [%o1] ASI_PNF, %o3 /* Load */
+ andcc %o0, 7, %g0 /* IEU1 Group */
+ bne,a,pt %icc, 13b /* CTI */
+ stb %o3, [%o0] /* Store */
+
+ andcc %o1, 7, %g3 /* IEU1 Group */
+ be,a,pt %icc, 1b /* CTI */
+ ldx [%o1], %o3 /* Load */
+14: orcc %g0, 64, %g4 /* IEU1 Group */
+
+ sllx %g3, 3, %g5 /* IEU0 */
+ sub %o1, %g3, %o1 /* IEU0 Group */
+ sub %g4, %g5, %g4 /* IEU1 */
+ /* %g1 = 0101010101010101 *
+ * %g2 = 8080808080808080 *
+ * %g3 = source alignment *
+ * %g5 = number of bits to shift left *
+ * %g4 = number of bits to shift right */
+ ldxa [%o1] ASI_PNF, %o5 /* Load Group */
+
+ addcc %o1, 8, %o1 /* IEU1 */
+15: sllx %o5, %g5, %o3 /* IEU0 Group */
+ ldxa [%o1] ASI_PNF, %o5 /* Load */
+ srlx %o5, %g4, %o4 /* IEU0 Group */
+
+ add %o0, 8, %o0 /* IEU1 */
+ or %o3, %o4, %o3 /* IEU0 Group */
+ add %o1, 8, %o1 /* IEU1 */
+ sub %o3, %g1, %o4 /* IEU0 Group */
+
+#ifdef EIGHTBIT_NOT_RARE
+ andn %o4, %o3, %o4 /* IEU0 Group */
+#endif
+ andcc %o4, %g2, %g0 /* IEU1 Group */
+ be,a,pt %xcc, 15b /* CTI */
+ stx %o3, [%o0 - 8] /* Store */
+ srlx %o3, 56, %o4 /* IEU0 Group */
+
+ andcc %o4, 0xff, %g0 /* IEU1 Group */
+ be,pn %icc, 22f /* CTI */
+ srlx %o3, 48, %o4 /* IEU0 */
+ andcc %o4, 0xff, %g0 /* IEU1 Group */
+
+ be,pn %icc, 21f /* CTI */
+ srlx %o3, 40, %o4 /* IEU0 */
+ andcc %o4, 0xff, %g0 /* IEU1 Group */
+ be,pn %icc, 20f /* CTI */
+
+ srlx %o3, 32, %o4 /* IEU0 */
+ andcc %o4, 0xff, %g0 /* IEU1 Group */
+ be,pn %icc, 19f /* CTI */
+ srlx %o3, 24, %o4 /* IEU0 */
+
+ andcc %o4, 0xff, %g0 /* IEU1 Group */
+ be,pn %icc, 18f /* CTI */
+ srlx %o3, 16, %o4 /* IEU0 */
+ andcc %o4, 0xff, %g0 /* IEU1 Group */
+
+ be,pn %icc, 17f /* CTI */
+ srlx %o3, 8, %o4 /* IEU0 */
+ andcc %o4, 0xff, %g0 /* IEU1 Group */
+ be,pn %icc, 16f /* CTI */
+
+ andcc %o3, 0xff, %g0 /* IEU1 Group */
+ bne,pn %icc, 15b /* CTI */
+ stx %o3, [%o0 - 8] /* Store */
+ retl /* CTI+IEU1 Group */
+
+ mov %g6, %o0 /* IEU0 */
+
+ .align 16
+16: srlx %o3, 8, %o4 /* IEU0 Group */
+ stb %o4, [%o0 - 2] /* Store */
+17: srlx %o3, 16, %o4 /* IEU0 Group */
+ stb %o4, [%o0 - 3] /* Store */
+
+18: srlx %o3, 24, %o4 /* IEU0 Group */
+ stb %o4, [%o0 - 4] /* Store */
+19: srlx %o3, 32, %o4 /* IEU0 Group */
+ stw %o4, [%o0 - 8] /* Store */
+
+ retl /* CTI+IEU1 Group */
+ mov %g6, %o0 /* IEU0 */
+ nop
+ nop
+
+20: srlx %o3, 40, %o4 /* IEU0 Group */
+ stb %o4, [%o0 - 6] /* Store */
+21: srlx %o3, 48, %o4 /* IEU0 Group */
+ stb %o4, [%o0 - 7] /* Store */
+
+22: srlx %o3, 56, %o4 /* IEU0 Group */
+ stb %o4, [%o0 - 8] /* Store */
+ retl /* CTI+IEU1 Group */
+ mov %g6, %o0 /* IEU0 */
+END(strcpy)
diff --git a/libc/string/sparc/sparc64/strlen.S b/libc/string/sparc/sparc64/strlen.S
new file mode 100644
index 000000000..5167ad5ba
--- /dev/null
+++ b/libc/string/sparc/sparc64/strlen.S
@@ -0,0 +1,172 @@
+/* Determine the length of a string. For SPARC v9.
+ Copyright (C) 1998, 1999, 2003 Free Software Foundation, Inc.
+ This file is part of the GNU C Library.
+ Contributed by Jan Vondrak <jvon4518@ss1000.ms.mff.cuni.cz> and
+ Jakub Jelinek <jj@ultra.linux.cz>.
+
+ The GNU C Library is free software; you can redistribute it and/or
+ modify it under the terms of the GNU Lesser General Public
+ License as published by the Free Software Foundation; either
+ version 2.1 of the License, or (at your option) any later version.
+
+ The GNU C Library is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ Lesser General Public License for more details.
+
+ You should have received a copy of the GNU Lesser General Public
+ License along with the GNU C Library; if not, write to the Free
+ Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA
+ 02111-1307 USA. */
+
+#include <asm/asi.h>
+
+ /* Normally, this uses
+ ((xword - 0x0101010101010101) & 0x8080808080808080) test
+ to find out if any byte in xword could be zero. This is fast, but
+ also gives false alarm for any byte in range 0x81-0xff. It does
+ not matter for correctness, as if this test tells us there could
+ be some zero byte, we check it byte by byte, but if bytes with
+ high bits set are common in the strings, then this will give poor
+ performance. You can #define EIGHTBIT_NOT_RARE and the algorithm
+ will use one tick slower, but more precise test
+ ((xword - 0x0101010101010101) & (~xword) & 0x8080808080808080),
+ which does not give any false alarms (but if some bits are set,
+ one cannot assume from it which bytes are zero and which are not).
+ It is yet to be measured, what is the correct default for glibc
+ in these days for an average user.
+ */
+
+ .text
+ .align 32
+ENTRY(strlen)
+ sethi %hi(0x01010101), %g1 /* IEU0 Group */
+ ldub [%o0], %o3 /* Load */
+ or %g1, %lo(0x01010101), %g1 /* IEU0 Group */
+ mov %o0, %o1 /* IEU1 */
+
+ sllx %g1, 32, %g4 /* IEU0 Group */
+ andcc %o0, 7, %g0 /* IEU1 */
+ or %g1, %g4, %g1 /* IEU0 Group */
+ brz,pn %o3, 13f /* CTI+IEU1 */
+
+ sllx %g1, 7, %g4 /* IEU0 Group */
+ bne,a,pn %icc, 15f /* CTI */
+ add %o0, 1, %o0 /* IEU1 */
+ /* %g1 = 0x0101010101010101 *
+ * %g4 = 0x8080808080808080 *
+ * %o0 = string pointer *
+ * %o1 = start of string */
+1: ldx [%o0], %o3 /* Load Group */
+
+ add %o0, 8, %o0 /* IEU1 */
+2: sub %o3, %g1, %o2 /* IEU0 Group */
+#ifdef EIGHTBIT_NOT_RARE
+ andn %o2, %o3, %o5 /* IEU0 Group */
+ ldxa [%o0] ASI_PNF, %o3 /* Load */
+ andcc %o5, %g4, %g0 /* IEU1 Group */
+#else
+ ldxa [%o0] ASI_PNF, %o3 /* Load */
+ andcc %o2, %g4, %g0 /* IEU1 Group */
+#endif
+
+ be,pt %xcc, 2b /* CTI */
+ add %o0, 8, %o0 /* IEU0 */
+ addcc %o2, %g1, %g5 /* IEU1 Group */
+#ifdef EIGHTBIT_NOT_RARE
+ srlx %o5, 32, %o5 /* IEU0 */
+
+3: andcc %o5, %g4, %g0 /* IEU1 Group */
+#else
+ srlx %o2, 32, %o2 /* IEU0 */
+
+3: andcc %o2, %g4, %g0 /* IEU1 Group */
+#endif
+ be,pn %xcc, 4f /* CTI */
+ srlx %g5, 56, %o2 /* IEU0 */
+ andcc %o2, 0xff, %g0 /* IEU1 Group */
+
+ be,pn %icc, 12f /* CTI */
+ srlx %g5, 48, %o2 /* IEU0 */
+ andcc %o2, 0xff, %g0 /* IEU1 Group */
+ be,pn %icc, 11f /* CTI */
+
+ srlx %g5, 40, %o2 /* IEU0 */
+ andcc %o2, 0xff, %g0 /* IEU1 Group */
+ be,pn %icc, 10f /* CTI */
+ srlx %g5, 32, %o2 /* IEU0 */
+
+ andcc %o2, 0xff, %g0 /* IEU1 Group */
+ be,pn %icc, 9f /* CTI */
+4: srlx %g5, 24, %o2 /* IEU0 */
+ andcc %o2, 0xff, %g0 /* IEU1 Group */
+
+ be,pn %icc, 8f /* CTI */
+ srlx %g5, 16, %o2 /* IEU0 */
+ andcc %o2, 0xff, %g0 /* IEU1 Group */
+ be,pn %icc, 7f /* CTI */
+
+ srlx %g5, 8, %o2 /* IEU0 */
+ andcc %o2, 0xff, %g0 /* IEU1 Group */
+ be,pn %icc, 6f /* CTI */
+ sub %o3, %g1, %o2 /* IEU0 */
+
+ andcc %g5, 0xff, %g0 /* IEU1 Group */
+ be,pn %icc, 5f /* CTI */
+ ldxa [%o0] ASI_PNF, %o3 /* Load */
+ andcc %o2, %g4, %g0 /* IEU1 Group */
+
+ be,pt %xcc, 2b /* CTI */
+ add %o0, 8, %o0 /* IEU0 */
+ addcc %o2, %g1, %g5 /* IEU1 Group */
+ ba,pt %xcc, 3b /* CTI */
+
+ srlx %o2, 32, %o2 /* IEU0 */
+5: add %o0, -9, %o0 /* IEU0 Group */
+ retl /* CTI+IEU1 Group */
+ sub %o0, %o1, %o0 /* IEU0 */
+
+6: add %o0, -10, %o0 /* IEU0 Group */
+ retl /* CTI+IEU1 Group */
+ sub %o0, %o1, %o0 /* IEU0 */
+7: add %o0, -11, %o0 /* IEU0 Group */
+
+ retl /* CTI+IEU1 Group */
+ sub %o0, %o1, %o0 /* IEU0 */
+8: add %o0, -12, %o0 /* IEU0 Group */
+ retl /* CTI+IEU1 Group */
+
+ sub %o0, %o1, %o0 /* IEU0 */
+9: add %o0, -13, %o0 /* IEU0 Group */
+ retl /* CTI+IEU1 Group */
+ sub %o0, %o1, %o0 /* IEU0 */
+
+10: add %o0, -14, %o0 /* IEU0 Group */
+ retl /* CTI+IEU1 Group */
+ sub %o0, %o1, %o0 /* IEU0 */
+11: add %o0, -15, %o0 /* IEU0 Group */
+
+ retl /* CTI+IEU1 Group */
+ sub %o0, %o1, %o0 /* IEU0 */
+12: add %o0, -16, %o0 /* IEU0 Group */
+ retl /* CTI+IEU1 Group */
+
+ sub %o0, %o1, %o0 /* IEU0 */
+13: retl /* CTI+IEU1 Group */
+ mov 0, %o0 /* IEU0 */
+ nop
+
+15: ldub [%o0], %o3 /* Load Group */
+16: andcc %o0, 7, %g0 /* IEU1 */
+ be,pn %icc, 1b /* CTI */
+ nop /* IEU0 Group */
+
+ add %o0, 1, %o0 /* IEU1 */
+ andcc %o3, 0xff, %g0 /* IEU1 Group */
+ bne,a,pt %icc, 16b /* CTI */
+ lduba [%o0] ASI_PNF, %o3 /* Load */
+
+ add %o0, -1, %o0 /* IEU0 Group */
+ retl /* CTI+IEU1 Group */
+ sub %o0, %o1, %o0 /* IEU0 */
+END(strlen)