diff options
| author | Jonathan Wright <jonathan.wright@arm.com> | 2020-09-28 13:35:51 +0100 |
|---|---|---|
| committer | Martin Storsjö <martin@martin.st> | 2021-11-16 13:43:56 +0200 |
| commit | 08b4716a9ed277afb4af69b07b751c9bd046aa70 (patch) | |
| tree | 541be310256c907c799c49a4aea070629e13a173 /libavcodec | |
| parent | 6f04cf54f532ad5695072303d74cddaa591f9aa1 (diff) | |
aarch64: Add Armv8.5-A BTI support
Add Branch Target Identifiers (BTIs) to all functions defined in
AArch64 assembly files. Most of the BTI landing pads are added
automatically by the 'function' macro.
BTI support is turned on or off at compile time based on the presence
of the __ARM_FEATURE_BTI_DEFAULT feature macro.
A binary compiled with BTI support can be executed on an Armv8-A
processor without BTI support because the instructions are defined in
NOP space.
Signed-off-by: Jonathan Wright <jonathan.wright@arm.com>
Signed-off-by: Elijah Ahmad <elijah.ahmad@arm.com>
Signed-off-by: Martin Storsjö <martin@martin.st>
Diffstat (limited to 'libavcodec')
| -rw-r--r-- | libavcodec/aarch64/fft_neon.S | 4 | ||||
| -rw-r--r-- | libavcodec/aarch64/h264idct_neon.S | 4 |
2 files changed, 8 insertions, 0 deletions
diff --git a/libavcodec/aarch64/fft_neon.S b/libavcodec/aarch64/fft_neon.S index 862039f97d..b4020fc8c7 100644 --- a/libavcodec/aarch64/fft_neon.S +++ b/libavcodec/aarch64/fft_neon.S @@ -36,6 +36,7 @@ function fft4_neon + AARCH64_VALID_JUMP_TARGET ld1 {v0.2s,v1.2s,v2.2s,v3.2s}, [x0] fadd v4.2s, v0.2s, v1.2s // r0+r1,i0+i1 @@ -58,6 +59,7 @@ function fft4_neon endfunc function fft8_neon + AARCH64_VALID_JUMP_TARGET mov x1, x0 ld1 {v0.2s, v1.2s, v2.2s, v3.2s}, [x0], #32 ld1 {v16.2s,v17.2s,v18.2s,v19.2s}, [x0] @@ -108,6 +110,7 @@ function fft8_neon endfunc function fft16_neon + AARCH64_VALID_JUMP_TARGET mov x1, x0 ld1 {v0.2s, v1.2s, v2.2s, v3.2s}, [x0], #32 ld1 {v16.2s,v17.2s,v18.2s,v19.2s}, [x0], #32 @@ -337,6 +340,7 @@ endfunc .macro def_fft n, n2, n4 function fft\n\()_neon, align=6 + AARCH64_VALID_JUMP_TARGET sub sp, sp, #16 stp x28, x30, [sp] add x28, x0, #\n4*2*8 diff --git a/libavcodec/aarch64/h264idct_neon.S b/libavcodec/aarch64/h264idct_neon.S index 7de44205d3..7d2879b0ce 100644 --- a/libavcodec/aarch64/h264idct_neon.S +++ b/libavcodec/aarch64/h264idct_neon.S @@ -24,6 +24,7 @@ function ff_h264_idct_add_neon, export=1 .L_ff_h264_idct_add_neon: + AARCH64_VALID_CALL_TARGET ld1 {v0.4H, v1.4H, v2.4H, v3.4H}, [x1] sxtw x2, w2 movi v30.8H, #0 @@ -79,6 +80,7 @@ endfunc function ff_h264_idct_dc_add_neon, export=1 .L_ff_h264_idct_dc_add_neon: + AARCH64_VALID_CALL_TARGET sxtw x2, w2 mov w3, #0 ld1r {v2.8H}, [x1] @@ -266,6 +268,7 @@ endfunc function ff_h264_idct8_add_neon, export=1 .L_ff_h264_idct8_add_neon: + AARCH64_VALID_CALL_TARGET movi v19.8H, #0 sxtw x2, w2 ld1 {v24.8H, v25.8H}, [x1] @@ -330,6 +333,7 @@ endfunc function ff_h264_idct8_dc_add_neon, export=1 .L_ff_h264_idct8_dc_add_neon: + AARCH64_VALID_CALL_TARGET mov w3, #0 sxtw x2, w2 ld1r {v31.8H}, [x1] |