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
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
|
/*
* Copyright (c) 2013-2022, Arm Limited and Contributors. All rights reserved.
*
* SPDX-License-Identifier: BSD-3-Clause
*/
#include <arch.h>
#include <asm_macros.S>
.globl flush_dcache_range
.globl flush_dcache_to_popa_range
.globl clean_dcache_range
.globl inv_dcache_range
.globl dcsw_op_louis
.globl dcsw_op_all
.globl dcsw_op_level1
.globl dcsw_op_level2
.globl dcsw_op_level3
/*
* This macro can be used for implementing various data cache operations `op`
*/
.macro do_dcache_maintenance_by_mva op
/* Exit early if size is zero */
cbz x1, exit_loop_\op
dcache_line_size x2, x3
add x1, x0, x1
sub x3, x2, #1
bic x0, x0, x3
loop_\op:
dc \op, x0
add x0, x0, x2
cmp x0, x1
b.lo loop_\op
dsb sy
exit_loop_\op:
ret
.endm
.macro check_plat_can_cmo
#if CONDITIONAL_CMO
mov x3, x30
mov x2, x0
bl plat_can_cmo
mov x30, x3
cbnz x0, 1f
ret
1:
mov x0, x2
#endif
.endm
/* ------------------------------------------
* Clean+Invalidate from base address till
* size. 'x0' = addr, 'x1' = size
* ------------------------------------------
*/
func flush_dcache_range
check_plat_can_cmo
do_dcache_maintenance_by_mva civac
endfunc flush_dcache_range
/* ------------------------------------------
* Clean from base address till size.
* 'x0' = addr, 'x1' = size
* ------------------------------------------
*/
func clean_dcache_range
check_plat_can_cmo
do_dcache_maintenance_by_mva cvac
endfunc clean_dcache_range
/* ------------------------------------------
* Invalidate from base address till
* size. 'x0' = addr, 'x1' = size
* ------------------------------------------
*/
func inv_dcache_range
check_plat_can_cmo
do_dcache_maintenance_by_mva ivac
endfunc inv_dcache_range
/*
* On implementations with FEAT_MTE2,
* Root firmware must issue DC_CIGDPAPA instead of DC_CIPAPA ,
* in order to additionally clean and invalidate Allocation Tags
* associated with the affected locations.
*
* ------------------------------------------
* Clean+Invalidate by PA to POPA
* from base address till size.
* 'x0' = addr, 'x1' = size
* ------------------------------------------
*/
func flush_dcache_to_popa_range
/* Exit early if size is zero */
cbz x1, exit_loop_dc_cipapa
check_plat_can_cmo
dcache_line_size x2, x3
sub x3, x2, #1
bic x0, x0, x3
add x1, x1, x0
loop_dc_cipapa:
sys #6, c7, c14, #1, x0 /* DC CIPAPA,<Xt> */
add x0, x0, x2
cmp x0, x1
b.lo loop_dc_cipapa
dsb osh
exit_loop_dc_cipapa:
ret
endfunc flush_dcache_to_popa_range
/* ---------------------------------------------------------------
* Data cache operations by set/way to the level specified
*
* The main function, do_dcsw_op requires:
* x0: The operation type (0-2), as defined in arch.h
* x3: The last cache level to operate on
* x9: clidr_el1
* x10: The cache level to begin operation from
* and will carry out the operation on each data cache from level 0
* to the level in x3 in sequence
*
* The dcsw_op macro sets up the x3 and x9 parameters based on
* clidr_el1 cache information before invoking the main function
* ---------------------------------------------------------------
*/
.macro dcsw_op shift, fw, ls
mrs x9, clidr_el1
ubfx x3, x9, \shift, \fw
lsl x3, x3, \ls
mov x10, xzr
b do_dcsw_op
.endm
func do_dcsw_op
cbz x3, exit
mrs x12, ID_AA64MMFR2_EL1 // stash FEAT_CCIDX identifier in x12
ubfx x12, x12, #ID_AA64MMFR2_EL1_CCIDX_SHIFT, #ID_AA64MMFR2_EL1_CCIDX_LENGTH
adr x14, dcsw_loop_table // compute inner loop address
add x14, x14, x0, lsl #5 // inner loop is 8x32-bit instructions
#if ENABLE_BTI
add x14, x14, x0, lsl #2 // inner loop is + "bti j" instruction
#endif
mov x0, x9
mov w8, #1
loop1:
add x2, x10, x10, lsr #1 // work out 3x current cache level
lsr x1, x0, x2 // extract cache type bits from clidr
and x1, x1, #7 // mask the bits for current cache only
cmp x1, #2 // see what cache we have at this level
b.lo level_done // nothing to do if no cache or icache
msr csselr_el1, x10 // select current cache level in csselr
isb // isb to sych the new cssr&csidr
mrs x1, ccsidr_el1 // read the new ccsidr
and x2, x1, #7 // extract the length of the cache lines
add x2, x2, #4 // add 4 (line length offset)
cbz x12, 1f // check for FEAT_CCIDX for Associativity
ubfx x4, x1, #3, #21 // x4 = associativity CCSIDR_EL1[23:3]
b 2f
1:
ubfx x4, x1, #3, #10 // x4 = associativity CCSIDR_EL1[12:3]
2:
clz w5, w4 // bit position of way size increment
lsl w9, w4, w5 // w9 = aligned max way number
lsl w16, w8, w5 // w16 = way number loop decrement
orr w9, w10, w9 // w9 = combine way and cache number
cbz x12, 3f // check for FEAT_CCIDX for NumSets
ubfx x6, x1, #32, #24 // x6 (w6) = numsets CCSIDR_EL1[55:32]
// ISA will not allow x->w ubfx
b 4f
3:
ubfx w6, w1, #13, #15 // w6 = numsets CCSIDR_EL1[27:13]
4:
lsl w17, w8, w2 // w17 = set number loop decrement
dsb sy // barrier before we start this level
br x14 // jump to DC operation specific loop
.macro dcsw_loop _op
#if ENABLE_BTI
bti j
#endif
loop2_\_op:
lsl w7, w6, w2 // w7 = aligned max set number
loop3_\_op:
orr w11, w9, w7 // combine cache, way and set number
dc \_op, x11
subs w7, w7, w17 // decrement set number
b.hs loop3_\_op
subs x9, x9, x16 // decrement way number
b.hs loop2_\_op
b level_done
.endm
level_done:
add x10, x10, #2 // increment cache number
cmp x3, x10
b.hi loop1
msr csselr_el1, xzr // select cache level 0 in csselr
dsb sy // barrier to complete final cache operation
isb
exit:
ret
endfunc do_dcsw_op
dcsw_loop_table:
dcsw_loop isw
dcsw_loop cisw
dcsw_loop csw
func dcsw_op_louis
check_plat_can_cmo
dcsw_op #LOUIS_SHIFT, #CLIDR_FIELD_WIDTH, #LEVEL_SHIFT
endfunc dcsw_op_louis
func dcsw_op_all
check_plat_can_cmo
dcsw_op #LOC_SHIFT, #CLIDR_FIELD_WIDTH, #LEVEL_SHIFT
endfunc dcsw_op_all
/* ---------------------------------------------------------------
* Helper macro for data cache operations by set/way for the
* level specified
* ---------------------------------------------------------------
*/
.macro dcsw_op_level level
mrs x9, clidr_el1
mov x3, \level
sub x10, x3, #2
b do_dcsw_op
.endm
/* ---------------------------------------------------------------
* Data cache operations by set/way for level 1 cache
*
* The main function, do_dcsw_op requires:
* x0: The operation type (0-2), as defined in arch.h
* ---------------------------------------------------------------
*/
func dcsw_op_level1
check_plat_can_cmo
dcsw_op_level #(1 << LEVEL_SHIFT)
endfunc dcsw_op_level1
/* ---------------------------------------------------------------
* Data cache operations by set/way for level 2 cache
*
* The main function, do_dcsw_op requires:
* x0: The operation type (0-2), as defined in arch.h
* ---------------------------------------------------------------
*/
func dcsw_op_level2
check_plat_can_cmo
dcsw_op_level #(2 << LEVEL_SHIFT)
endfunc dcsw_op_level2
/* ---------------------------------------------------------------
* Data cache operations by set/way for level 3 cache
*
* The main function, do_dcsw_op requires:
* x0: The operation type (0-2), as defined in arch.h
* ---------------------------------------------------------------
*/
func dcsw_op_level3
check_plat_can_cmo
dcsw_op_level #(3 << LEVEL_SHIFT)
endfunc dcsw_op_level3
|
