xref: /linux/arch/arm64/include/asm/fpsimdmacros.h (revision e0bf6c5ca2d3281f231c5f0c9bf145e9513644de)
1 /*
2  * FP/SIMD state saving and restoring macros
3  *
4  * Copyright (C) 2012 ARM Ltd.
5  * Author: Catalin Marinas <catalin.marinas@arm.com>
6  *
7  * This program is free software; you can redistribute it and/or modify
8  * it under the terms of the GNU General Public License version 2 as
9  * published by the Free Software Foundation.
10  *
11  * This program is distributed in the hope that it will be useful,
12  * but WITHOUT ANY WARRANTY; without even the implied warranty of
13  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
14  * GNU General Public License for more details.
15  *
16  * You should have received a copy of the GNU General Public License
17  * along with this program.  If not, see <http://www.gnu.org/licenses/>.
18  */
19 
20 .macro fpsimd_save state, tmpnr
21 	stp	q0, q1, [\state, #16 * 0]
22 	stp	q2, q3, [\state, #16 * 2]
23 	stp	q4, q5, [\state, #16 * 4]
24 	stp	q6, q7, [\state, #16 * 6]
25 	stp	q8, q9, [\state, #16 * 8]
26 	stp	q10, q11, [\state, #16 * 10]
27 	stp	q12, q13, [\state, #16 * 12]
28 	stp	q14, q15, [\state, #16 * 14]
29 	stp	q16, q17, [\state, #16 * 16]
30 	stp	q18, q19, [\state, #16 * 18]
31 	stp	q20, q21, [\state, #16 * 20]
32 	stp	q22, q23, [\state, #16 * 22]
33 	stp	q24, q25, [\state, #16 * 24]
34 	stp	q26, q27, [\state, #16 * 26]
35 	stp	q28, q29, [\state, #16 * 28]
36 	stp	q30, q31, [\state, #16 * 30]!
37 	mrs	x\tmpnr, fpsr
38 	str	w\tmpnr, [\state, #16 * 2]
39 	mrs	x\tmpnr, fpcr
40 	str	w\tmpnr, [\state, #16 * 2 + 4]
41 .endm
42 
43 .macro fpsimd_restore_fpcr state, tmp
44 	/*
45 	 * Writes to fpcr may be self-synchronising, so avoid restoring
46 	 * the register if it hasn't changed.
47 	 */
48 	mrs	\tmp, fpcr
49 	cmp	\tmp, \state
50 	b.eq	9999f
51 	msr	fpcr, \state
52 9999:
53 .endm
54 
55 /* Clobbers \state */
56 .macro fpsimd_restore state, tmpnr
57 	ldp	q0, q1, [\state, #16 * 0]
58 	ldp	q2, q3, [\state, #16 * 2]
59 	ldp	q4, q5, [\state, #16 * 4]
60 	ldp	q6, q7, [\state, #16 * 6]
61 	ldp	q8, q9, [\state, #16 * 8]
62 	ldp	q10, q11, [\state, #16 * 10]
63 	ldp	q12, q13, [\state, #16 * 12]
64 	ldp	q14, q15, [\state, #16 * 14]
65 	ldp	q16, q17, [\state, #16 * 16]
66 	ldp	q18, q19, [\state, #16 * 18]
67 	ldp	q20, q21, [\state, #16 * 20]
68 	ldp	q22, q23, [\state, #16 * 22]
69 	ldp	q24, q25, [\state, #16 * 24]
70 	ldp	q26, q27, [\state, #16 * 26]
71 	ldp	q28, q29, [\state, #16 * 28]
72 	ldp	q30, q31, [\state, #16 * 30]!
73 	ldr	w\tmpnr, [\state, #16 * 2]
74 	msr	fpsr, x\tmpnr
75 	ldr	w\tmpnr, [\state, #16 * 2 + 4]
76 	fpsimd_restore_fpcr x\tmpnr, \state
77 .endm
78 
79 .macro fpsimd_save_partial state, numnr, tmpnr1, tmpnr2
80 	mrs	x\tmpnr1, fpsr
81 	str	w\numnr, [\state, #8]
82 	mrs	x\tmpnr2, fpcr
83 	stp	w\tmpnr1, w\tmpnr2, [\state]
84 	adr	x\tmpnr1, 0f
85 	add	\state, \state, x\numnr, lsl #4
86 	sub	x\tmpnr1, x\tmpnr1, x\numnr, lsl #1
87 	br	x\tmpnr1
88 	stp	q30, q31, [\state, #-16 * 30 - 16]
89 	stp	q28, q29, [\state, #-16 * 28 - 16]
90 	stp	q26, q27, [\state, #-16 * 26 - 16]
91 	stp	q24, q25, [\state, #-16 * 24 - 16]
92 	stp	q22, q23, [\state, #-16 * 22 - 16]
93 	stp	q20, q21, [\state, #-16 * 20 - 16]
94 	stp	q18, q19, [\state, #-16 * 18 - 16]
95 	stp	q16, q17, [\state, #-16 * 16 - 16]
96 	stp	q14, q15, [\state, #-16 * 14 - 16]
97 	stp	q12, q13, [\state, #-16 * 12 - 16]
98 	stp	q10, q11, [\state, #-16 * 10 - 16]
99 	stp	q8, q9, [\state, #-16 * 8 - 16]
100 	stp	q6, q7, [\state, #-16 * 6 - 16]
101 	stp	q4, q5, [\state, #-16 * 4 - 16]
102 	stp	q2, q3, [\state, #-16 * 2 - 16]
103 	stp	q0, q1, [\state, #-16 * 0 - 16]
104 0:
105 .endm
106 
107 .macro fpsimd_restore_partial state, tmpnr1, tmpnr2
108 	ldp	w\tmpnr1, w\tmpnr2, [\state]
109 	msr	fpsr, x\tmpnr1
110 	fpsimd_restore_fpcr x\tmpnr2, x\tmpnr1
111 	adr	x\tmpnr1, 0f
112 	ldr	w\tmpnr2, [\state, #8]
113 	add	\state, \state, x\tmpnr2, lsl #4
114 	sub	x\tmpnr1, x\tmpnr1, x\tmpnr2, lsl #1
115 	br	x\tmpnr1
116 	ldp	q30, q31, [\state, #-16 * 30 - 16]
117 	ldp	q28, q29, [\state, #-16 * 28 - 16]
118 	ldp	q26, q27, [\state, #-16 * 26 - 16]
119 	ldp	q24, q25, [\state, #-16 * 24 - 16]
120 	ldp	q22, q23, [\state, #-16 * 22 - 16]
121 	ldp	q20, q21, [\state, #-16 * 20 - 16]
122 	ldp	q18, q19, [\state, #-16 * 18 - 16]
123 	ldp	q16, q17, [\state, #-16 * 16 - 16]
124 	ldp	q14, q15, [\state, #-16 * 14 - 16]
125 	ldp	q12, q13, [\state, #-16 * 12 - 16]
126 	ldp	q10, q11, [\state, #-16 * 10 - 16]
127 	ldp	q8, q9, [\state, #-16 * 8 - 16]
128 	ldp	q6, q7, [\state, #-16 * 6 - 16]
129 	ldp	q4, q5, [\state, #-16 * 4 - 16]
130 	ldp	q2, q3, [\state, #-16 * 2 - 16]
131 	ldp	q0, q1, [\state, #-16 * 0 - 16]
132 0:
133 .endm
134