aboutsummaryrefslogtreecommitdiffstats
path: root/contrib/libs/isa-l/erasure_code/gf_vect_mad_avx.asm
blob: 1a252c474fe6949bfc354bfd51c76f732a319841 (plain) (blame)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
;  Copyright(c) 2011-2015 Intel Corporation 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.
;    * Neither the name of Intel Corporation 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
;  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.
;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;

;;;
;;; gf_vect_mad_avx(len, vec, vec_i, mul_array, src, dest);
;;;

%include "reg_sizes.asm"

%ifidn __OUTPUT_FORMAT__, win64
 %define arg0  rcx
 %define arg0.w ecx
 %define arg1  rdx
 %define arg2  r8
 %define arg3  r9
 %define arg4  r12
 %define arg5  r15
 %define tmp   r11
 %define return rax
 %define return.w eax
 %define PS 8
 %define stack_size 16*3 + 3*8
 %define arg(x)      [rsp + stack_size + PS + PS*x]
 %define func(x) proc_frame x

%macro FUNC_SAVE 0
	sub	rsp, stack_size
	vmovdqa	[rsp+16*0],xmm6
	vmovdqa	[rsp+16*1],xmm7
	vmovdqa	[rsp+16*2],xmm8
	save_reg	r12,  3*16 + 0*8
	save_reg	r15,  3*16 + 1*8
	end_prolog
	mov	arg4, arg(4)
	mov	arg5, arg(5)
%endmacro

%macro FUNC_RESTORE 0
	vmovdqa	xmm6, [rsp+16*0]
	vmovdqa	xmm7, [rsp+16*1]
	vmovdqa	xmm8, [rsp+16*2]
	mov	r12,  [rsp + 3*16 + 0*8]
	mov	r15,  [rsp + 3*16 + 1*8]
	add	rsp, stack_size
%endmacro

%elifidn __OUTPUT_FORMAT__, elf64
 %define arg0  rdi
 %define arg0.w edi
 %define arg1  rsi
 %define arg2  rdx
 %define arg3  rcx
 %define arg4  r8
 %define arg5  r9
 %define tmp   r11
 %define return rax
 %define return.w eax

 %define func(x) x:
 %define FUNC_SAVE
 %define FUNC_RESTORE
%endif

;;; gf_vect_mad_avx(len, vec, vec_i, mul_array, src, dest)
%define len   arg0
%define len.w arg0.w
%define vec    arg1
%define vec_i    arg2
%define mul_array arg3
%define	src   arg4
%define dest  arg5
%define pos   return
%define pos.w return.w

%ifndef EC_ALIGNED_ADDR
;;; Use Un-aligned load/store
 %define XLDR vmovdqu
 %define XSTR vmovdqu
%else
;;; Use Non-temporal load/stor
 %ifdef NO_NT_LDST
  %define XLDR vmovdqa
  %define XSTR vmovdqa
 %else
  %define XLDR vmovntdqa
  %define XSTR vmovntdq
 %endif
%endif


default rel

[bits 64]
section .text

%define xmask0f  xmm8
%define xgft_lo  xmm7
%define xgft_hi  xmm6

%define x0     xmm0
%define xtmpa  xmm1
%define xtmph  xmm2
%define xtmpl  xmm3
%define xd     xmm4
%define xtmpd  xmm5

align 16
global gf_vect_mad_avx:ISAL_SYM_TYPE_FUNCTION
func(gf_vect_mad_avx)
%ifidn __OUTPUT_FORMAT__, macho64
global _gf_vect_mad_avx:ISAL_SYM_TYPE_FUNCTION
func(_gf_vect_mad_avx)
%endif

	FUNC_SAVE
	sub	len, 16
	jl	.return_fail

	xor	pos, pos
	vmovdqa	xmask0f, [mask0f]	;Load mask of lower nibble in each byte

	sal	vec_i, 5		;Multiply by 32
	vmovdqu	xgft_lo, [vec_i+mul_array]	;Load array Cx{00}, Cx{01}, Cx{02}, ...
	vmovdqu	xgft_hi, [vec_i+mul_array+16]	; " Cx{00}, Cx{10}, Cx{20}, ... , Cx{f0}

	XLDR	xtmpd, [dest+len]	;backup the last 16 bytes in dest

.loop16:
	XLDR	xd, [dest+pos]		;Get next dest vector
.loop16_overlap:
	XLDR	x0, [src+pos]		;Get next source vector

	vpand	xtmpa, x0, xmask0f	;Mask low src nibble in bits 4-0
	vpsraw	x0, x0, 4		;Shift to put high nibble into bits 4-0
	vpand	x0, x0, xmask0f		;Mask high src nibble in bits 4-0

	vpshufb	xtmph, xgft_hi, x0	;Lookup mul table of high nibble
	vpshufb	xtmpl, xgft_lo, xtmpa	;Lookup mul table of low nibble
	vpxor	xtmph, xtmph, xtmpl ;GF add high and low partials
	vpxor	xd, xd, xtmph		;xd += partial

	XSTR	[dest+pos], xd
	add	pos, 16			;Loop on 16 bytes at a time
	cmp	pos, len
	jle	.loop16

	lea	tmp, [len + 16]
	cmp	pos, tmp
	je	.return_pass

	;; Tail len
	mov	pos, len	;Overlapped offset length-16
	vmovdqa	xd, xtmpd	;Restore xd
	jmp	.loop16_overlap	;Do one more overlap pass

.return_pass:
	mov	return, 0
	FUNC_RESTORE
	ret

.return_fail:
	mov	return, 1
	FUNC_RESTORE
	ret

endproc_frame

section .data

align 16

mask0f: dq 0x0f0f0f0f0f0f0f0f, 0x0f0f0f0f0f0f0f0f

;;;       func            core, ver, snum
slversion gf_vect_mad_avx, 02,  01,  0201