source: trunk/libs/newlib/src/newlib/libc/machine/arc/memcmp.S @ 444

/*
   Copyright (c) 2015, Synopsys, Inc. All rights reserved.

   Redistribution and use in source and binary forms, with or without
   modification, are permitted provided that the following conditions are met:

   1) Redistributions of source code must retain the above copyright notice,
   this list of conditions and the following disclaimer.

   2) Redistributions in binary form must reproduce the above copyright notice,
   this list of conditions and the following disclaimer in the documentation
   and/or other materials provided with the distribution.

   3) Neither the name of the Synopsys, Inc., nor the names of its contributors
   may be used to endorse or promote products derived from this software
   without specific prior written permission.

   THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
   AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
   IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
   ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE
   LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
   CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
   SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
   INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
   CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
   ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
   POSSIBILITY OF SUCH DAMAGE.
*/

/* This implementation is optimized for performance.  For code size a generic
   implementation of this function from newlib/libc/string/memcmp.c will be
   used.  */
#if !defined (__OPTIMIZE_SIZE__) && !defined (PREFER_SIZE_OVER_SPEED)

#include "asm.h"

#if defined (__ARC601__) || !defined (__ARC_NORM__) \
    || !defined (__ARC_BARREL_SHIFTER__)

/* Addresses are unsigned, and at 0 is the vector table, so it's OK to assume
   that we can subtract 8 from a source end address without underflow.  */

ENTRY (memcmp)
        or      r12,r0,r1
        tst     r12,3
        breq    r2,0,.Lnil
        add_s   r3,r0,r2

/* This algorithm for big endian targets sometimes works incorrectly
   when sources are aligned. To be precise the last step is omitted.
   Just use a simple bytewise variant until the algorithm is reviewed
   and fixed.  */

#ifdef __LITTLE_ENDIAN__
        bne_s   .Lbytewise
#else /* BIG ENDIAN */
        b_s     .Lbytewise
#endif /* ENDIAN */
        sub     r6,r3,8
        ld      r4,[r0,0]
        ld      r5,[r1,0]
2:
        brhs    r0,r6,.Loop_end
        ld_s    r3,[r0,4]
        ld_s    r12,[r1,4]
        brne    r4,r5,.Leven
        ld.a    r4,[r0,8]
        breq.d  r3,r12,2b
        ld.a    r5,[r1,8]
#ifdef __LITTLE_ENDIAN__
        mov_s   r4,r3
        b.d     .Lodd
        mov_s   r5,r12
#else /* BIG ENDIAN */
        cmp_s   r3,r12
        j_s.d   [blink]
        rrc     r0,2
#endif /* ENDIAN */

        .balign 4
.Loop_end:
        sub     r3,r0,r6
        brhs    r3,4,.Last_cmp
        brne    r4,r5,.Leven
        ld      r4,[r0,4]
        ld      r5,[r1,4]
#ifdef __LITTLE_ENDIAN__
        .balign 4
.Last_cmp:
        mov_l   r0,24
        add3    r2,r0,r2
        xor     r0,r4,r5
        b.d     .Leven_cmp
        bset    r0,r0,r2
.Lodd:
.Leven:
        xor     r0,r4,r5
.Leven_cmp:
        mov_s   r1,0x80808080
        ; uses long immediate
        sub_s   r12,r0,1
        bic_s   r0,r0,r12
        sub     r0,r1,r0
        xor_s   r0,r0,r1
        and     r1,r5,r0
        and     r0,r4,r0
#else /* BIG ENDIAN */
.Last_cmp:
        mov_s   r3,0
        sub3    r2,r3,r2
        sub_s   r3,r3,1
        bclr    r3,r3,r2
        add_l   r3,r3,1
        and     r0,r4,r3
        and     r1,r5,r3
.Leven:
#endif /* ENDIAN */
        xor.f   0,r0,r1
        sub_s   r0,r0,r1
        j_s.d   [blink]
        mov.mi  r0,r1
        .balign 4
.Lbytewise:
        ldb     r4,[r0,0]
        ldb     r5,[r1,0]
        sub     r6,r3,2
3:
        brhs    r0,r6,.Lbyte_end
        ldb_s   r3,[r0,1]
        ldb_s   r12,[r1,1]
        brne    r4,r5,.Lbyte_even
        ldb.a   r4,[r0,2]
        breq.d  r3,r12,3b
        ldb.a   r5,[r1,2]
.Lbyte_odd:
        j_s.d   [blink]
        sub     r0,r3,r12
        .balign 4
.Lbyte_end:
        bbit1   r2,0,.Lbyte_even
        brne    r4,r5,.Lbyte_even
        ldb     r4,[r0,1]
        ldb     r5,[r1,1]
.Lbyte_even:
        j_s.d   [blink]
        sub     r0,r4,r5
.Lnil:
        j_s.d   [blink]
        mov_s   r0,0
ENDFUNC (memcmp)
#endif /* __ARC601__ || !__ARC_NORM__ || !__ARC_BARREL_SHIFTER__ */

#endif /* !__OPTIMIZE_SIZE__ && !PREFER_SIZE_OVER_SPEED */