openharmony/third_party_ffmpeg
1
0
Fork 0
mirror of https://gitee.com/openharmony/third_party_ffmpeg synced 2024-11-27 05:00:37 +00:00
Code Issues Actions Packages Projects Releases Wiki Activity

sws/aarch64/yuv2rgb: honor iOS calling convention

Browse Source 
y_offset and y_coeff being successive 32-bit integers, they are packed
into 8 bytes instead of 2x8 bytes.

See https://developer.apple.com/library/ios/documentation/Xcode/Conceptual/iPhoneOSABIReference/Articles/ARM64FunctionCallingConventions.html

> iOS diverges from Procedure Call Standard for the ARM 64-bit
> Architecture in several ways
[...]
> In the generic procedure call standard, all function arguments passed
> on the stack consume slots in multiples of 8 bytes. In iOS, this
> requirement is dropped, and values consume only the space required.
[...]
> Padding is still inserted on the stack to satisfy arguments’ alignment
> requirements.
This commit is contained in:
Clément Bœsch 2016-04-08 13:39:22 +02:00
parent 259879d32d
commit cab9661dba
1 changed files with 12 additions and 6 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

18
libswscale/aarch64/yuv2rgb_neon.S
View File

@ -21,10 +21,18 @@
#include "libavutil/aarch64/asm.S"
.macro load_yoff_ycoeff yoff ycoeff
#if defined(__APPLE__)
ldp w9, w10, [sp, #\yoff]
#else
ldr w9, [sp, #\yoff]
ldr w10, [sp, #\ycoeff]
#endif
.endm
.macro load_args_nv12
ldr x8, [sp] // table
ldr w9, [sp, #8] // y_offset
ldr w10, [sp, #16] // y_coeff
load_yoff_ycoeff 8, 16 // y_offset, y_coeff
ld1 {v1.1D}, [x8]
dup v0.8H, w10
dup v3.8H, w9
@ -42,8 +50,7 @@
ldr x13, [sp] // srcV
ldr w14, [sp, #8] // linesizeV
ldr x8, [sp, #16] // table
ldr w9, [sp, #24] // y_offset
ldr w10, [sp, #32] // y_coeff
load_yoff_ycoeff 24, 32 // y_offset, y_coeff
ld1 {v1.1D}, [x8]
dup v0.8H, w10
dup v3.8H, w9
@ -59,8 +66,7 @@
ldr x13, [sp] // srcV
ldr w14, [sp, #8] // linesizeV
ldr x8, [sp, #16] // table
ldr w9, [sp, #24] // y_offset
ldr w10, [sp, #32] // y_coeff
load_yoff_ycoeff 24, 32 // y_offset, y_coeff
ld1 {v1.1D}, [x8]
dup v0.8H, w10
dup v3.8H, w9