1 
2 /*---------------------------------------------------------------*/
3 /*--- begin                             host_generic_simd64.h ---*/
4 /*---------------------------------------------------------------*/
5 
6 /*
7    This file is part of Valgrind, a dynamic binary instrumentation
8    framework.
9 
10    Copyright (C) 2004-2013 OpenWorks LLP
11       info@open-works.net
12 
13    This program is free software; you can redistribute it and/or
14    modify it under the terms of the GNU General Public License as
15    published by the Free Software Foundation; either version 2 of the
16    License, or (at your option) any later version.
17 
18    This program is distributed in the hope that it will be useful, but
19    WITHOUT ANY WARRANTY; without even the implied warranty of
20    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
21    General Public License for more details.
22 
23    You should have received a copy of the GNU General Public License
24    along with this program; if not, write to the Free Software
25    Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA
26    02110-1301, USA.
27 
28    The GNU General Public License is contained in the file COPYING.
29 
30    Neither the names of the U.S. Department of Energy nor the
31    University of California nor the names of its contributors may be
32    used to endorse or promote products derived from this software
33    without prior written permission.
34 */
35 
36 /* Generic helper functions for doing 64-bit SIMD arithmetic in cases
37    where the instruction selectors cannot generate code in-line.
38    These are purely back-end entities and cannot be seen/referenced
39    as clean helper functions from IR.
40 
41    These will get called from generated code and therefore should be
42    well behaved -- no floating point or mmx insns, just straight
43    integer code.
44 
45    Each function implements the correspondingly-named IR primop.
46 */
47 
48 #ifndef __VEX_HOST_GENERIC_SIMD64_H
49 #define __VEX_HOST_GENERIC_SIMD64_H
50 
51 #include "libvex_basictypes.h"
52 
53 /* DO NOT MAKE THESE INTO REGPARM FNS!  THIS WILL BREAK CALLING
54    SEQUENCES GENERATED BY host-x86/isel.c. */
55 
56 extern ULong h_generic_calc_Add32x2 ( ULong, ULong );
57 extern ULong h_generic_calc_Add16x4 ( ULong, ULong );
58 extern ULong h_generic_calc_Add8x8  ( ULong, ULong );
59 
60 extern ULong h_generic_calc_QAdd16Sx4 ( ULong, ULong );
61 extern ULong h_generic_calc_QAdd8Sx8  ( ULong, ULong );
62 extern ULong h_generic_calc_QAdd16Ux4 ( ULong, ULong );
63 extern ULong h_generic_calc_QAdd8Ux8  ( ULong, ULong );
64 
65 extern ULong h_generic_calc_Sub32x2 ( ULong, ULong );
66 extern ULong h_generic_calc_Sub16x4 ( ULong, ULong );
67 extern ULong h_generic_calc_Sub8x8  ( ULong, ULong );
68 
69 extern ULong h_generic_calc_QSub16Sx4 ( ULong, ULong );
70 extern ULong h_generic_calc_QSub8Sx8  ( ULong, ULong );
71 extern ULong h_generic_calc_QSub16Ux4 ( ULong, ULong );
72 extern ULong h_generic_calc_QSub8Ux8  ( ULong, ULong );
73 
74 extern ULong h_generic_calc_Mul16x4    ( ULong, ULong );
75 extern ULong h_generic_calc_Mul32x2    ( ULong, ULong );
76 extern ULong h_generic_calc_MulHi16Sx4 ( ULong, ULong );
77 extern ULong h_generic_calc_MulHi16Ux4 ( ULong, ULong );
78 
79 extern ULong h_generic_calc_CmpEQ32x2  ( ULong, ULong );
80 extern ULong h_generic_calc_CmpEQ16x4  ( ULong, ULong );
81 extern ULong h_generic_calc_CmpEQ8x8   ( ULong, ULong );
82 extern ULong h_generic_calc_CmpGT32Sx2 ( ULong, ULong );
83 extern ULong h_generic_calc_CmpGT16Sx4 ( ULong, ULong );
84 extern ULong h_generic_calc_CmpGT8Sx8  ( ULong, ULong );
85 
86 extern ULong h_generic_calc_CmpNEZ32x2 ( ULong );
87 extern ULong h_generic_calc_CmpNEZ16x4 ( ULong );
88 extern ULong h_generic_calc_CmpNEZ8x8  ( ULong );
89 
90 extern ULong h_generic_calc_QNarrowBin32Sto16Sx4 ( ULong, ULong );
91 extern ULong h_generic_calc_QNarrowBin16Sto8Sx8  ( ULong, ULong );
92 extern ULong h_generic_calc_QNarrowBin16Sto8Ux8  ( ULong, ULong );
93 extern ULong h_generic_calc_NarrowBin32to16x4    ( ULong, ULong );
94 extern ULong h_generic_calc_NarrowBin16to8x8     ( ULong, ULong );
95 
96 extern ULong h_generic_calc_InterleaveHI8x8 ( ULong, ULong );
97 extern ULong h_generic_calc_InterleaveLO8x8 ( ULong, ULong );
98 extern ULong h_generic_calc_InterleaveHI16x4 ( ULong, ULong );
99 extern ULong h_generic_calc_InterleaveLO16x4 ( ULong, ULong );
100 extern ULong h_generic_calc_InterleaveHI32x2 ( ULong, ULong );
101 extern ULong h_generic_calc_InterleaveLO32x2 ( ULong, ULong );
102 
103 extern ULong h_generic_calc_CatOddLanes16x4 ( ULong, ULong );
104 extern ULong h_generic_calc_CatEvenLanes16x4 ( ULong, ULong );
105 extern ULong h_generic_calc_Perm8x8 ( ULong, ULong );
106 
107 extern ULong h_generic_calc_ShlN8x8  ( ULong, UInt );
108 extern ULong h_generic_calc_ShlN16x4 ( ULong, UInt );
109 extern ULong h_generic_calc_ShlN32x2 ( ULong, UInt );
110 
111 extern ULong h_generic_calc_ShrN16x4 ( ULong, UInt );
112 extern ULong h_generic_calc_ShrN32x2 ( ULong, UInt );
113 
114 extern ULong h_generic_calc_SarN8x8  ( ULong, UInt );
115 extern ULong h_generic_calc_SarN16x4 ( ULong, UInt );
116 extern ULong h_generic_calc_SarN32x2 ( ULong, UInt );
117 
118 extern ULong h_generic_calc_Avg8Ux8  ( ULong, ULong );
119 extern ULong h_generic_calc_Avg16Ux4 ( ULong, ULong );
120 
121 extern ULong h_generic_calc_Max16Sx4 ( ULong, ULong );
122 extern ULong h_generic_calc_Max8Ux8  ( ULong, ULong );
123 extern ULong h_generic_calc_Min16Sx4 ( ULong, ULong );
124 extern ULong h_generic_calc_Min8Ux8  ( ULong, ULong );
125 
126 extern UInt  h_generic_calc_GetMSBs8x8 ( ULong );
127 
128 /* 32-bit SIMD HELPERS */
129 
130 extern UInt h_generic_calc_Add16x2   ( UInt, UInt );
131 extern UInt h_generic_calc_Sub16x2   ( UInt, UInt );
132 
133 extern UInt h_generic_calc_HAdd16Ux2 ( UInt, UInt );
134 extern UInt h_generic_calc_HAdd16Sx2 ( UInt, UInt );
135 extern UInt h_generic_calc_HSub16Ux2 ( UInt, UInt );
136 extern UInt h_generic_calc_HSub16Sx2 ( UInt, UInt );
137 
138 extern UInt h_generic_calc_QAdd16Ux2 ( UInt, UInt );
139 extern UInt h_generic_calc_QAdd16Sx2 ( UInt, UInt );
140 extern UInt h_generic_calc_QSub16Ux2 ( UInt, UInt );
141 extern UInt h_generic_calc_QSub16Sx2 ( UInt, UInt );
142 
143 extern UInt h_generic_calc_Add8x4   ( UInt, UInt );
144 extern UInt h_generic_calc_Sub8x4   ( UInt, UInt );
145 
146 extern UInt h_generic_calc_HAdd8Ux4 ( UInt, UInt );
147 extern UInt h_generic_calc_HAdd8Sx4 ( UInt, UInt );
148 extern UInt h_generic_calc_HSub8Ux4 ( UInt, UInt );
149 extern UInt h_generic_calc_HSub8Sx4 ( UInt, UInt );
150 
151 extern UInt h_generic_calc_QAdd8Ux4 ( UInt, UInt );
152 extern UInt h_generic_calc_QAdd8Sx4 ( UInt, UInt );
153 extern UInt h_generic_calc_QSub8Ux4 ( UInt, UInt );
154 extern UInt h_generic_calc_QSub8Sx4 ( UInt, UInt );
155 
156 extern UInt h_generic_calc_Sad8Ux4  ( UInt, UInt );
157 
158 extern UInt h_generic_calc_QAdd32S  ( UInt, UInt );
159 extern UInt h_generic_calc_QSub32S  ( UInt, UInt );
160 
161 extern UInt h_generic_calc_CmpNEZ16x2 ( UInt );
162 extern UInt h_generic_calc_CmpNEZ8x4  ( UInt );
163 
164 extern ULong h_calc_DPBtoBCD ( ULong dpb );
165 extern ULong h_calc_BCDtoDPB ( ULong bcd );
166 
167 // Signed and unsigned integer division, that behave like
168 // the ARMv7 UDIV and SDIV instructions.
169 extern UInt  h_calc_udiv32_w_arm_semantics ( UInt,  UInt  );
170 extern ULong h_calc_udiv64_w_arm_semantics ( ULong, ULong );
171 extern Int   h_calc_sdiv32_w_arm_semantics ( Int,   Int   );
172 extern Long  h_calc_sdiv64_w_arm_semantics ( Long,  Long  );
173 
174 
175 #endif /* ndef __VEX_HOST_GENERIC_SIMD64_H */
176 
177 /*---------------------------------------------------------------*/
178 /*--- end                               host_generic_simd64.h ---*/
179 /*---------------------------------------------------------------*/
180