blob: 4ff982b0f89eb8dc0a80759799cef4095383c8b2 [file] [log] [blame]
/*
* Copyright (C) 2008 The Android Open Source Project
* All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* * Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* * 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.
*
* 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 OWNER 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.
*/
/*
* Copyright (c) 2013 ARM Ltd
* 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. The name of the company may not be used to endorse or promote
* products derived from this software without specific prior written
* permission.
*
* THIS SOFTWARE IS PROVIDED BY ARM LTD ``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 ARM LTD 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.
*/
.L_memcpy_base:
// Assumes that n >= 0, and dst, src are valid pointers.
// For any sizes less than 832 use the neon code that doesn't
// care about the src alignment. This avoids any checks
// for src alignment, and offers the best improvement since
// smaller sized copies are dominated by the overhead of
// the pre and post main loop.
// For larger copies, if src and dst cannot both be aligned to
// word boundaries, use the neon code.
// For all other copies, align dst to a double word boundary
// and copy using LDRD/STRD instructions.
cmp r2, #16
blo .L_copy_less_than_16_unknown_align
.L_copy_unknown_alignment:
// Unknown alignment of src and dst.
// Assumes that the first few bytes have already been prefetched.
// Align destination to 128 bits. The mainloop store instructions
// require this alignment or they will throw an exception.
rsb r3, r0, #0
ands r3, r3, #0xF
beq 2f
// Copy up to 15 bytes (count in r3).
sub r2, r2, r3
movs ip, r3, lsl #31
itt mi
ldrbmi lr, [r1], #1
strbmi lr, [r0], #1
itttt cs
ldrbcs ip, [r1], #1
ldrbcs lr, [r1], #1
strbcs ip, [r0], #1
strbcs lr, [r0], #1
movs ip, r3, lsl #29
bge 1f
// Copies 4 bytes, dst 32 bits aligned before, at least 64 bits after.
vld4.8 {d0[0], d1[0], d2[0], d3[0]}, [r1]!
vst4.8 {d0[0], d1[0], d2[0], d3[0]}, [r0, :32]!
1: bcc 2f
// Copies 8 bytes, dst 64 bits aligned before, at least 128 bits after.
vld1.8 {d0}, [r1]!
vst1.8 {d0}, [r0, :64]!
2: cmp r2, #256
ble .L_copy_loop
// Make sure DST is 64 BYTE aligned.
rsb r3, r0, #0
ands r3, r3, #0x30
beq .L_copy_loop
sub r2, r2, r3
cmp r3, #0x10
beq .L_copy_16
vld1.8 {d0 - d3}, [r1]!
vst1.8 {d0 - d3}, [r0, :128]!
ands r3, r3, #0x10
beq .L_copy_loop
.L_copy_16:
vld1.8 {d0, d1}, [r1]!
vst1.8 {d0, d1}, [r0, :128]!
.L_copy_loop:
// Make sure we have at least 64 bytes to copy.
subs r2, r2, #64
blo 2f
1: // The main loop copies 64 bytes at a time.
vld1.8 {d0 - d3}, [r1]!
vst1.8 {d0 - d3}, [r0, :128]!
pld [r1, #(64*4)]
subs r2, r2, #64
vld1.8 {d4 - d7}, [r1]!
vst1.8 {d4 - d7}, [r0, :128]!
bhs 1b
2: // Fix-up the remaining count and make sure we have >= 32 bytes left.
adds r2, r2, #32
blo 3f
// 32 bytes. These cache lines were already preloaded.
vld1.8 {d0 - d3}, [r1]!
sub r2, r2, #32
vst1.8 {d0 - d3}, [r0, :128]!
3: // Less than 32 left.
add r2, r2, #32
tst r2, #0x10
beq .L_copy_less_than_16_unknown_align
// Copies 16 bytes, destination 128 bits aligned.
vld1.8 {d0, d1}, [r1]!
vst1.8 {d0, d1}, [r0, :128]!
.L_copy_less_than_16_unknown_align:
// Copy up to 15 bytes (count in r2).
movs ip, r2, lsl #29
bcc 1f
vld1.8 {d0}, [r1]!
vst1.8 {d0}, [r0]!
1: bge 2f
vld4.8 {d0[0], d1[0], d2[0], d3[0]}, [r1]!
vst4.8 {d0[0], d1[0], d2[0], d3[0]}, [r0]!
2: // Copy 0 to 4 bytes.
lsls r2, r2, #31
itt ne
ldrbne lr, [r1], #1
strbne lr, [r0], #1
itttt cs
ldrbcs ip, [r1], #1
ldrbcs lr, [r1]
strbcs ip, [r0], #1
strbcs lr, [r0]
pop {r0, pc}