summaryrefslogtreecommitdiff
path: root/graphics/ffmpeg/patches/patch-libavcodec_aarch64_vp9lpf_neon_S
diff options
context:
space:
mode:
Diffstat (limited to 'graphics/ffmpeg/patches/patch-libavcodec_aarch64_vp9lpf_neon_S')
-rw-r--r--graphics/ffmpeg/patches/patch-libavcodec_aarch64_vp9lpf_neon_S220
1 files changed, 220 insertions, 0 deletions
diff --git a/graphics/ffmpeg/patches/patch-libavcodec_aarch64_vp9lpf_neon_S b/graphics/ffmpeg/patches/patch-libavcodec_aarch64_vp9lpf_neon_S
new file mode 100644
index 0000000..5ca9c4a
--- /dev/null
+++ b/graphics/ffmpeg/patches/patch-libavcodec_aarch64_vp9lpf_neon_S
@@ -0,0 +1,220 @@
1Index: libavcodec/aarch64/vp9lpf_neon.S
2--- libavcodec/aarch64/vp9lpf_neon.S.orig
3+++ libavcodec/aarch64/vp9lpf_neon.S
4@@ -399,7 +399,7 @@
5 .endif
6 // If no pixels needed flat8in nor flat8out, jump to a
7 // writeout of the inner 4 pixels
8- br x14
9+ ret x14
10 1:
11
12 mov x5, v7.d[0]
13@@ -411,7 +411,7 @@
14 cbnz x5, 1f
15 .endif
16 // If no pixels need flat8out, jump to a writeout of the inner 6 pixels
17- br x15
18+ ret x15
19
20 1:
21 // flat8out
22@@ -532,32 +532,32 @@ function vp9_loop_filter_4
23 loop_filter 4, .8b, 0, v16, v17, v18, v19, v28, v29, v30, v31
24 ret
25 9:
26- br x10
27+ ret x10
28 endfunc
29
30 function vp9_loop_filter_4_16b_mix_44
31 loop_filter 4, .16b, 44, v16, v17, v18, v19, v28, v29, v30, v31
32 ret
33 9:
34- br x10
35+ ret x10
36 endfunc
37
38 function vp9_loop_filter_8
39 loop_filter 8, .8b, 0, v16, v17, v18, v19, v28, v29, v30, v31
40 ret
41 6:
42- br x13
43+ ret x13
44 9:
45- br x10
46+ ret x10
47 endfunc
48
49 function vp9_loop_filter_8_16b_mix
50 loop_filter 8, .16b, 88, v16, v17, v18, v19, v28, v29, v30, v31
51 ret
52 6:
53- br x13
54+ ret x13
55 9:
56- br x10
57+ ret x10
58 endfunc
59
60 function vp9_loop_filter_16
61@@ -568,7 +568,7 @@ function vp9_loop_filter_16
62 ldp d10, d11, [sp], 0x10
63 ldp d12, d13, [sp], 0x10
64 ldp d14, d15, [sp], 0x10
65- br x10
66+ ret x10
67 endfunc
68
69 function vp9_loop_filter_16_16b
70@@ -579,7 +579,7 @@ function vp9_loop_filter_16_16b
71 ldp d10, d11, [sp], 0x10
72 ldp d12, d13, [sp], 0x10
73 ldp d14, d15, [sp], 0x10
74- br x10
75+ ret x10
76 endfunc
77
78 .macro loop_filter_4
79@@ -648,7 +648,7 @@ function ff_vp9_loop_filter_v_4_8_neon, export=1
80 st1 {v23.8b}, [x9], x1
81 st1 {v25.8b}, [x0], x1
82
83- br x10
84+ ret x10
85 endfunc
86
87 function ff_vp9_loop_filter_v_44_16_neon, export=1
88@@ -672,7 +672,7 @@ function ff_vp9_loop_filter_v_44_16_neon, export=1
89 st1 {v23.16b}, [x9], x1
90 st1 {v25.16b}, [x0], x1
91
92- br x10
93+ ret x10
94 endfunc
95
96 function ff_vp9_loop_filter_h_4_8_neon, export=1
97@@ -714,7 +714,7 @@ function ff_vp9_loop_filter_h_4_8_neon, export=1
98 st1 {v25.s}[0], [x9], x1
99 st1 {v25.s}[1], [x0], x1
100
101- br x10
102+ ret x10
103 endfunc
104
105 function ff_vp9_loop_filter_h_44_16_neon, export=1
106@@ -766,7 +766,7 @@ function ff_vp9_loop_filter_h_44_16_neon, export=1
107 st1 {v25.s}[1], [x9], x1
108 st1 {v25.s}[3], [x0], x1
109
110- br x10
111+ ret x10
112 endfunc
113
114 function ff_vp9_loop_filter_v_8_8_neon, export=1
115@@ -793,14 +793,14 @@ function ff_vp9_loop_filter_v_8_8_neon, export=1
116 st1 {v23.8b}, [x9], x1
117 st1 {v26.8b}, [x0], x1
118
119- br x10
120+ ret x10
121 6:
122 sub x9, x0, x1, lsl #1
123 st1 {v22.8b}, [x9], x1
124 st1 {v24.8b}, [x0], x1
125 st1 {v23.8b}, [x9], x1
126 st1 {v25.8b}, [x0], x1
127- br x10
128+ ret x10
129 endfunc
130
131 .macro mix_v_16 mix
132@@ -828,14 +828,14 @@ function ff_vp9_loop_filter_v_\mix\()_16_neon, export=
133 st1 {v23.16b}, [x9], x1
134 st1 {v26.16b}, [x0], x1
135
136- br x10
137+ ret x10
138 6:
139 sub x9, x0, x1, lsl #1
140 st1 {v22.16b}, [x9], x1
141 st1 {v24.16b}, [x0], x1
142 st1 {v23.16b}, [x9], x1
143 st1 {v25.16b}, [x0], x1
144- br x10
145+ ret x10
146 endfunc
147 .endm
148
149@@ -876,7 +876,7 @@ function ff_vp9_loop_filter_h_8_8_neon, export=1
150 st1 {v23.8b}, [x9], x1
151 st1 {v27.8b}, [x0], x1
152
153- br x10
154+ ret x10
155 6:
156 // If we didn't need to do the flat8in part, we use the same writeback
157 // as in loop_filter_h_4_8.
158@@ -891,7 +891,7 @@ function ff_vp9_loop_filter_h_8_8_neon, export=1
159 st1 {v24.s}[1], [x0], x1
160 st1 {v25.s}[0], [x9], x1
161 st1 {v25.s}[1], [x0], x1
162- br x10
163+ ret x10
164 endfunc
165
166 .macro mix_h_16 mix
167@@ -942,7 +942,7 @@ function ff_vp9_loop_filter_h_\mix\()_16_neon, export=
168 st1 {v27.8b}, [x9], x1
169 st1 {v27.d}[1], [x0], x1
170
171- br x10
172+ ret x10
173 6:
174 add x9, x9, #2
175 add x0, x0, #2
176@@ -963,7 +963,7 @@ function ff_vp9_loop_filter_h_\mix\()_16_neon, export=
177 st1 {v24.s}[3], [x0], x1
178 st1 {v25.s}[1], [x9], x1
179 st1 {v25.s}[3], [x0], x1
180- br x10
181+ ret x10
182 endfunc
183 .endm
184
185@@ -1022,7 +1022,7 @@ function ff_vp9_loop_filter_v_16_8_neon, export=1
186 ldp d10, d11, [sp], 0x10
187 ldp d12, d13, [sp], 0x10
188 ldp d14, d15, [sp], 0x10
189- br x10
190+ ret x10
191 8:
192 add x9, x9, x1, lsl #2
193 // If we didn't do the flat8out part, the output is left in the
194@@ -1091,7 +1091,7 @@ function ff_vp9_loop_filter_v_16_16_neon, export=1
195 ldp d10, d11, [sp], 0x10
196 ldp d12, d13, [sp], 0x10
197 ldp d14, d15, [sp], 0x10
198- br x10
199+ ret x10
200 8:
201 add x9, x9, x1, lsl #2
202 st1 {v21.16b}, [x9], x1
203@@ -1168,7 +1168,7 @@ function ff_vp9_loop_filter_h_16_8_neon, export=1
204 ldp d10, d11, [sp], 0x10
205 ldp d12, d13, [sp], 0x10
206 ldp d14, d15, [sp], 0x10
207- br x10
208+ ret x10
209 8:
210 // The same writeback as in loop_filter_h_8_8
211 sub x9, x0, #4
212@@ -1287,7 +1287,7 @@ function ff_vp9_loop_filter_h_16_16_neon, export=1
213 ldp d10, d11, [sp], 0x10
214 ldp d12, d13, [sp], 0x10
215 ldp d14, d15, [sp], 0x10
216- br x10
217+ ret x10
218 8:
219 sub x9, x0, #4
220 add x0, x9, x1, lsl #3
generated by cgit v1.2.3 (git 2.25.1) at 2026-09-03 05:44:03 +0000