This is an automated email from the git hooks/post-receive script. Git pushed a commit to branch master in repository ffmpeg.
commit 0e49820959bf914413ccb8dbf1b7b1a36272bb65 Author: Niklas Haas <[email protected]> AuthorDate: Thu Oct 22 20:15:54 2026 +0200 Commit: Niklas Haas <[email protected]> CommitDate: Sun Oct 25 02:23:01 2026 +0000 swscale: generate disjoint uop macros for the C and x86 backends Previously, the uop macros were generated from the union of uops over all backends, so each backend would also instantiate uops that only the other one can ever use. Split off a separate "x86" backend, and generate a separate C header for it, together with its NASM macros in a single pass. The x86 backend only enables SWS_UOP_FLAG_FMA on CPUs with FMA3, so generate its uops both with and without FMA. This does not reduce the checkasm coverage: the uops removed from uops_macros.h are all LINEAR_FMA, READ_PLANAR_FV_FMA or RW_SHUFFLE, which checkasm does not test, since the C reference backend cannot run them. The only tests that disappear are the SSE4 and AVX2 variants of two U8 PERMUTE kernels, which the x86 backend no longer builds because it never uses them. Co-Authored-By: Claude Opus 5.5 <[email protected]> Signed-off-by: Niklas Haas <[email protected]> --- libswscale/uops_macros.h | 144 +++---------------------------------- libswscale/uops_macros_gen.c | 11 +-- libswscale/x86/ops.c | 2 +- libswscale/x86/uops_macros.asm | 2 - libswscale/{ => x86}/uops_macros.h | 10 +-- tests/fate/libswscale.mak | 15 ++-- tests/ref/fate/source | 1 + 7 files changed, 32 insertions(+), 153 deletions(-) diff --git a/libswscale/uops_macros.h b/libswscale/uops_macros.h index 9e300fbac1..5082b45237 100644 --- a/libswscale/uops_macros.h +++ b/libswscale/uops_macros.h @@ -49,16 +49,8 @@ MACRO(__VA_ARGS__, u8_read_planar_fv_xy_f32 , .type = SWS_PIXEL_U8 , .uop = SWS_UOP_READ_PLANAR_FV , .mask = 0x3, .par.filter.type = SWS_PIXEL_F32) \ MACRO(__VA_ARGS__, u8_read_planar_fv_xyz_f32 , .type = SWS_PIXEL_U8 , .uop = SWS_UOP_READ_PLANAR_FV , .mask = 0x7, .par.filter.type = SWS_PIXEL_F32) \ MACRO(__VA_ARGS__, u8_read_planar_fv_xyzw_f32 , .type = SWS_PIXEL_U8 , .uop = SWS_UOP_READ_PLANAR_FV , .mask = 0xf, .par.filter.type = SWS_PIXEL_F32) -#define SWS_FOR_U8_READ_PLANAR_FV_FMA(MACRO, ...) \ - MACRO(__VA_ARGS__, u8_read_planar_fv_fma_x_f32 , SWS_PIXEL_U8 , SWS_UOP_READ_PLANAR_FV_FMA, 0x1, SWS_PIXEL_F32) \ - MACRO(__VA_ARGS__, u8_read_planar_fv_fma_xy_f32 , SWS_PIXEL_U8 , SWS_UOP_READ_PLANAR_FV_FMA, 0x3, SWS_PIXEL_F32) \ - MACRO(__VA_ARGS__, u8_read_planar_fv_fma_xyz_f32 , SWS_PIXEL_U8 , SWS_UOP_READ_PLANAR_FV_FMA, 0x7, SWS_PIXEL_F32) \ - MACRO(__VA_ARGS__, u8_read_planar_fv_fma_xyzw_f32 , SWS_PIXEL_U8 , SWS_UOP_READ_PLANAR_FV_FMA, 0xf, SWS_PIXEL_F32) -#define SWS_FOR_STRUCT_U8_READ_PLANAR_FV_FMA(MACRO, ...) \ - MACRO(__VA_ARGS__, u8_read_planar_fv_fma_x_f32 , .type = SWS_PIXEL_U8 , .uop = SWS_UOP_READ_PLANAR_FV_FMA, .mask = 0x1, .par.filter.type = SWS_PIXEL_F32) \ - MACRO(__VA_ARGS__, u8_read_planar_fv_fma_xy_f32 , .type = SWS_PIXEL_U8 , .uop = SWS_UOP_READ_PLANAR_FV_FMA, .mask = 0x3, .par.filter.type = SWS_PIXEL_F32) \ - MACRO(__VA_ARGS__, u8_read_planar_fv_fma_xyz_f32 , .type = SWS_PIXEL_U8 , .uop = SWS_UOP_READ_PLANAR_FV_FMA, .mask = 0x7, .par.filter.type = SWS_PIXEL_F32) \ - MACRO(__VA_ARGS__, u8_read_planar_fv_fma_xyzw_f32 , .type = SWS_PIXEL_U8 , .uop = SWS_UOP_READ_PLANAR_FV_FMA, .mask = 0xf, .par.filter.type = SWS_PIXEL_F32) +#define SWS_FOR_U8_READ_PLANAR_FV_FMA(MACRO, ...) +#define SWS_FOR_STRUCT_U8_READ_PLANAR_FV_FMA(MACRO, ...) #define SWS_FOR_U8_READ_PACKED(MACRO, ...) \ MACRO(__VA_ARGS__, u8_read_packed_xy , SWS_PIXEL_U8 , SWS_UOP_READ_PACKED , 0x3) \ MACRO(__VA_ARGS__, u8_read_packed_xyz , SWS_PIXEL_U8 , SWS_UOP_READ_PACKED , 0x7) \ @@ -105,52 +97,8 @@ MACRO(__VA_ARGS__, u8_write_bit_x , SWS_PIXEL_U8 , SWS_UOP_WRITE_BIT , 0x1) #define SWS_FOR_STRUCT_U8_WRITE_BIT(MACRO, ...) \ MACRO(__VA_ARGS__, u8_write_bit_x , .type = SWS_PIXEL_U8 , .uop = SWS_UOP_WRITE_BIT , .mask = 0x1) -#define SWS_FOR_U8_RW_SHUFFLE(MACRO, ...) \ - MACRO(__VA_ARGS__, u8_rw_shuffle_x_0_2_12 , SWS_PIXEL_U8 , SWS_UOP_RW_SHUFFLE , 0x1, 0x0, 2, 12) \ - MACRO(__VA_ARGS__, u8_rw_shuffle_x_0_4_12 , SWS_PIXEL_U8 , SWS_UOP_RW_SHUFFLE , 0x1, 0x0, 4, 12) \ - MACRO(__VA_ARGS__, u8_rw_shuffle_x_0_4_16 , SWS_PIXEL_U8 , SWS_UOP_RW_SHUFFLE , 0x1, 0x0, 4, 16) \ - MACRO(__VA_ARGS__, u8_rw_shuffle_x_0_5_15 , SWS_PIXEL_U8 , SWS_UOP_RW_SHUFFLE , 0x1, 0x0, 5, 15) \ - MACRO(__VA_ARGS__, u8_rw_shuffle_x_0_6_12 , SWS_PIXEL_U8 , SWS_UOP_RW_SHUFFLE , 0x1, 0x0, 6, 12) \ - MACRO(__VA_ARGS__, u8_rw_shuffle_x_0_8_12 , SWS_PIXEL_U8 , SWS_UOP_RW_SHUFFLE , 0x1, 0x0, 8, 12) \ - MACRO(__VA_ARGS__, u8_rw_shuffle_x_0_8_16 , SWS_PIXEL_U8 , SWS_UOP_RW_SHUFFLE , 0x1, 0x0, 8, 16) \ - MACRO(__VA_ARGS__, u8_rw_shuffle_x_0_10_15 , SWS_PIXEL_U8 , SWS_UOP_RW_SHUFFLE , 0x1, 0x0, 10, 15) \ - MACRO(__VA_ARGS__, u8_rw_shuffle_x_0_12_12 , SWS_PIXEL_U8 , SWS_UOP_RW_SHUFFLE , 0x1, 0x0, 12, 12) \ - MACRO(__VA_ARGS__, u8_rw_shuffle_x_0_12_16 , SWS_PIXEL_U8 , SWS_UOP_RW_SHUFFLE , 0x1, 0x0, 12, 16) \ - MACRO(__VA_ARGS__, u8_rw_shuffle_x_0_15_5 , SWS_PIXEL_U8 , SWS_UOP_RW_SHUFFLE , 0x1, 0x0, 15, 5) \ - MACRO(__VA_ARGS__, u8_rw_shuffle_x_0_15_15 , SWS_PIXEL_U8 , SWS_UOP_RW_SHUFFLE , 0x1, 0x0, 15, 15) \ - MACRO(__VA_ARGS__, u8_rw_shuffle_x_0_16_4 , SWS_PIXEL_U8 , SWS_UOP_RW_SHUFFLE , 0x1, 0x0, 16, 4) \ - MACRO(__VA_ARGS__, u8_rw_shuffle_x_0_16_8 , SWS_PIXEL_U8 , SWS_UOP_RW_SHUFFLE , 0x1, 0x0, 16, 8) \ - MACRO(__VA_ARGS__, u8_rw_shuffle_x_0_16_12 , SWS_PIXEL_U8 , SWS_UOP_RW_SHUFFLE , 0x1, 0x0, 16, 12) \ - MACRO(__VA_ARGS__, u8_rw_shuffle_x_0_16_16 , SWS_PIXEL_U8 , SWS_UOP_RW_SHUFFLE , 0x1, 0x0, 16, 16) \ - MACRO(__VA_ARGS__, u8_rw_shuffle_x_ff_2_16 , SWS_PIXEL_U8 , SWS_UOP_RW_SHUFFLE , 0x1, 0xff, 2, 16) \ - MACRO(__VA_ARGS__, u8_rw_shuffle_x_ff_4_16 , SWS_PIXEL_U8 , SWS_UOP_RW_SHUFFLE , 0x1, 0xff, 4, 16) \ - MACRO(__VA_ARGS__, u8_rw_shuffle_x_ff_6_16 , SWS_PIXEL_U8 , SWS_UOP_RW_SHUFFLE , 0x1, 0xff, 6, 16) \ - MACRO(__VA_ARGS__, u8_rw_shuffle_x_ff_8_16 , SWS_PIXEL_U8 , SWS_UOP_RW_SHUFFLE , 0x1, 0xff, 8, 16) \ - MACRO(__VA_ARGS__, u8_rw_shuffle_x_ff_12_16 , SWS_PIXEL_U8 , SWS_UOP_RW_SHUFFLE , 0x1, 0xff, 12, 16) \ - MACRO(__VA_ARGS__, u8_rw_shuffle_x_ff_16_16 , SWS_PIXEL_U8 , SWS_UOP_RW_SHUFFLE , 0x1, 0xff, 16, 16) -#define SWS_FOR_STRUCT_U8_RW_SHUFFLE(MACRO, ...) \ - MACRO(__VA_ARGS__, u8_rw_shuffle_x_0_2_12 , .type = SWS_PIXEL_U8 , .uop = SWS_UOP_RW_SHUFFLE , .mask = 0x1, .par.shuffle.clear_value = 0x0, .par.shuffle.read_size = 2, .par.shuffle.write_size = 12) \ - MACRO(__VA_ARGS__, u8_rw_shuffle_x_0_4_12 , .type = SWS_PIXEL_U8 , .uop = SWS_UOP_RW_SHUFFLE , .mask = 0x1, .par.shuffle.clear_value = 0x0, .par.shuffle.read_size = 4, .par.shuffle.write_size = 12) \ - MACRO(__VA_ARGS__, u8_rw_shuffle_x_0_4_16 , .type = SWS_PIXEL_U8 , .uop = SWS_UOP_RW_SHUFFLE , .mask = 0x1, .par.shuffle.clear_value = 0x0, .par.shuffle.read_size = 4, .par.shuffle.write_size = 16) \ - MACRO(__VA_ARGS__, u8_rw_shuffle_x_0_5_15 , .type = SWS_PIXEL_U8 , .uop = SWS_UOP_RW_SHUFFLE , .mask = 0x1, .par.shuffle.clear_value = 0x0, .par.shuffle.read_size = 5, .par.shuffle.write_size = 15) \ - MACRO(__VA_ARGS__, u8_rw_shuffle_x_0_6_12 , .type = SWS_PIXEL_U8 , .uop = SWS_UOP_RW_SHUFFLE , .mask = 0x1, .par.shuffle.clear_value = 0x0, .par.shuffle.read_size = 6, .par.shuffle.write_size = 12) \ - MACRO(__VA_ARGS__, u8_rw_shuffle_x_0_8_12 , .type = SWS_PIXEL_U8 , .uop = SWS_UOP_RW_SHUFFLE , .mask = 0x1, .par.shuffle.clear_value = 0x0, .par.shuffle.read_size = 8, .par.shuffle.write_size = 12) \ - MACRO(__VA_ARGS__, u8_rw_shuffle_x_0_8_16 , .type = SWS_PIXEL_U8 , .uop = SWS_UOP_RW_SHUFFLE , .mask = 0x1, .par.shuffle.clear_value = 0x0, .par.shuffle.read_size = 8, .par.shuffle.write_size = 16) \ - MACRO(__VA_ARGS__, u8_rw_shuffle_x_0_10_15 , .type = SWS_PIXEL_U8 , .uop = SWS_UOP_RW_SHUFFLE , .mask = 0x1, .par.shuffle.clear_value = 0x0, .par.shuffle.read_size = 10, .par.shuffle.write_size = 15) \ - MACRO(__VA_ARGS__, u8_rw_shuffle_x_0_12_12 , .type = SWS_PIXEL_U8 , .uop = SWS_UOP_RW_SHUFFLE , .mask = 0x1, .par.shuffle.clear_value = 0x0, .par.shuffle.read_size = 12, .par.shuffle.write_size = 12) \ - MACRO(__VA_ARGS__, u8_rw_shuffle_x_0_12_16 , .type = SWS_PIXEL_U8 , .uop = SWS_UOP_RW_SHUFFLE , .mask = 0x1, .par.shuffle.clear_value = 0x0, .par.shuffle.read_size = 12, .par.shuffle.write_size = 16) \ - MACRO(__VA_ARGS__, u8_rw_shuffle_x_0_15_5 , .type = SWS_PIXEL_U8 , .uop = SWS_UOP_RW_SHUFFLE , .mask = 0x1, .par.shuffle.clear_value = 0x0, .par.shuffle.read_size = 15, .par.shuffle.write_size = 5) \ - MACRO(__VA_ARGS__, u8_rw_shuffle_x_0_15_15 , .type = SWS_PIXEL_U8 , .uop = SWS_UOP_RW_SHUFFLE , .mask = 0x1, .par.shuffle.clear_value = 0x0, .par.shuffle.read_size = 15, .par.shuffle.write_size = 15) \ - MACRO(__VA_ARGS__, u8_rw_shuffle_x_0_16_4 , .type = SWS_PIXEL_U8 , .uop = SWS_UOP_RW_SHUFFLE , .mask = 0x1, .par.shuffle.clear_value = 0x0, .par.shuffle.read_size = 16, .par.shuffle.write_size = 4) \ - MACRO(__VA_ARGS__, u8_rw_shuffle_x_0_16_8 , .type = SWS_PIXEL_U8 , .uop = SWS_UOP_RW_SHUFFLE , .mask = 0x1, .par.shuffle.clear_value = 0x0, .par.shuffle.read_size = 16, .par.shuffle.write_size = 8) \ - MACRO(__VA_ARGS__, u8_rw_shuffle_x_0_16_12 , .type = SWS_PIXEL_U8 , .uop = SWS_UOP_RW_SHUFFLE , .mask = 0x1, .par.shuffle.clear_value = 0x0, .par.shuffle.read_size = 16, .par.shuffle.write_size = 12) \ - MACRO(__VA_ARGS__, u8_rw_shuffle_x_0_16_16 , .type = SWS_PIXEL_U8 , .uop = SWS_UOP_RW_SHUFFLE , .mask = 0x1, .par.shuffle.clear_value = 0x0, .par.shuffle.read_size = 16, .par.shuffle.write_size = 16) \ - MACRO(__VA_ARGS__, u8_rw_shuffle_x_ff_2_16 , .type = SWS_PIXEL_U8 , .uop = SWS_UOP_RW_SHUFFLE , .mask = 0x1, .par.shuffle.clear_value = 0xff, .par.shuffle.read_size = 2, .par.shuffle.write_size = 16) \ - MACRO(__VA_ARGS__, u8_rw_shuffle_x_ff_4_16 , .type = SWS_PIXEL_U8 , .uop = SWS_UOP_RW_SHUFFLE , .mask = 0x1, .par.shuffle.clear_value = 0xff, .par.shuffle.read_size = 4, .par.shuffle.write_size = 16) \ - MACRO(__VA_ARGS__, u8_rw_shuffle_x_ff_6_16 , .type = SWS_PIXEL_U8 , .uop = SWS_UOP_RW_SHUFFLE , .mask = 0x1, .par.shuffle.clear_value = 0xff, .par.shuffle.read_size = 6, .par.shuffle.write_size = 16) \ - MACRO(__VA_ARGS__, u8_rw_shuffle_x_ff_8_16 , .type = SWS_PIXEL_U8 , .uop = SWS_UOP_RW_SHUFFLE , .mask = 0x1, .par.shuffle.clear_value = 0xff, .par.shuffle.read_size = 8, .par.shuffle.write_size = 16) \ - MACRO(__VA_ARGS__, u8_rw_shuffle_x_ff_12_16 , .type = SWS_PIXEL_U8 , .uop = SWS_UOP_RW_SHUFFLE , .mask = 0x1, .par.shuffle.clear_value = 0xff, .par.shuffle.read_size = 12, .par.shuffle.write_size = 16) \ - MACRO(__VA_ARGS__, u8_rw_shuffle_x_ff_16_16 , .type = SWS_PIXEL_U8 , .uop = SWS_UOP_RW_SHUFFLE , .mask = 0x1, .par.shuffle.clear_value = 0xff, .par.shuffle.read_size = 16, .par.shuffle.write_size = 16) +#define SWS_FOR_U8_RW_SHUFFLE(MACRO, ...) +#define SWS_FOR_STRUCT_U8_RW_SHUFFLE(MACRO, ...) #define SWS_FOR_U8_PERMUTE(MACRO, ...) \ MACRO(__VA_ARGS__, u8_permute_xy_xy_zw , SWS_PIXEL_U8 , SWS_UOP_PERMUTE , 0x3, 2, 0, 1, 0, 0, 0, 0, 2, 3, 0, 0, 0, 0) \ MACRO(__VA_ARGS__, u8_permute_xyz_x_w , SWS_PIXEL_U8 , SWS_UOP_PERMUTE , 0x7, 1, 0, 0, 0, 0, 0, 0, 3, 0, 0, 0, 0, 0) \ @@ -385,16 +333,8 @@ MACRO(__VA_ARGS__, u16_read_planar_fv_xy_f32 , .type = SWS_PIXEL_U16, .uop = SWS_UOP_READ_PLANAR_FV , .mask = 0x3, .par.filter.type = SWS_PIXEL_F32) \ MACRO(__VA_ARGS__, u16_read_planar_fv_xyz_f32 , .type = SWS_PIXEL_U16, .uop = SWS_UOP_READ_PLANAR_FV , .mask = 0x7, .par.filter.type = SWS_PIXEL_F32) \ MACRO(__VA_ARGS__, u16_read_planar_fv_xyzw_f32 , .type = SWS_PIXEL_U16, .uop = SWS_UOP_READ_PLANAR_FV , .mask = 0xf, .par.filter.type = SWS_PIXEL_F32) -#define SWS_FOR_U16_READ_PLANAR_FV_FMA(MACRO, ...) \ - MACRO(__VA_ARGS__, u16_read_planar_fv_fma_x_f32 , SWS_PIXEL_U16, SWS_UOP_READ_PLANAR_FV_FMA, 0x1, SWS_PIXEL_F32) \ - MACRO(__VA_ARGS__, u16_read_planar_fv_fma_xy_f32 , SWS_PIXEL_U16, SWS_UOP_READ_PLANAR_FV_FMA, 0x3, SWS_PIXEL_F32) \ - MACRO(__VA_ARGS__, u16_read_planar_fv_fma_xyz_f32 , SWS_PIXEL_U16, SWS_UOP_READ_PLANAR_FV_FMA, 0x7, SWS_PIXEL_F32) \ - MACRO(__VA_ARGS__, u16_read_planar_fv_fma_xyzw_f32 , SWS_PIXEL_U16, SWS_UOP_READ_PLANAR_FV_FMA, 0xf, SWS_PIXEL_F32) -#define SWS_FOR_STRUCT_U16_READ_PLANAR_FV_FMA(MACRO, ...) \ - MACRO(__VA_ARGS__, u16_read_planar_fv_fma_x_f32 , .type = SWS_PIXEL_U16, .uop = SWS_UOP_READ_PLANAR_FV_FMA, .mask = 0x1, .par.filter.type = SWS_PIXEL_F32) \ - MACRO(__VA_ARGS__, u16_read_planar_fv_fma_xy_f32 , .type = SWS_PIXEL_U16, .uop = SWS_UOP_READ_PLANAR_FV_FMA, .mask = 0x3, .par.filter.type = SWS_PIXEL_F32) \ - MACRO(__VA_ARGS__, u16_read_planar_fv_fma_xyz_f32 , .type = SWS_PIXEL_U16, .uop = SWS_UOP_READ_PLANAR_FV_FMA, .mask = 0x7, .par.filter.type = SWS_PIXEL_F32) \ - MACRO(__VA_ARGS__, u16_read_planar_fv_fma_xyzw_f32 , .type = SWS_PIXEL_U16, .uop = SWS_UOP_READ_PLANAR_FV_FMA, .mask = 0xf, .par.filter.type = SWS_PIXEL_F32) +#define SWS_FOR_U16_READ_PLANAR_FV_FMA(MACRO, ...) +#define SWS_FOR_STRUCT_U16_READ_PLANAR_FV_FMA(MACRO, ...) #define SWS_FOR_U16_READ_PACKED(MACRO, ...) \ MACRO(__VA_ARGS__, u16_read_packed_xy , SWS_PIXEL_U16, SWS_UOP_READ_PACKED , 0x3) \ MACRO(__VA_ARGS__, u16_read_packed_xyz , SWS_PIXEL_U16, SWS_UOP_READ_PACKED , 0x7) \ @@ -897,16 +837,8 @@ MACRO(__VA_ARGS__, f32_read_planar_fv_xy_f32 , .type = SWS_PIXEL_F32, .uop = SWS_UOP_READ_PLANAR_FV , .mask = 0x3, .par.filter.type = SWS_PIXEL_F32) \ MACRO(__VA_ARGS__, f32_read_planar_fv_xyz_f32 , .type = SWS_PIXEL_F32, .uop = SWS_UOP_READ_PLANAR_FV , .mask = 0x7, .par.filter.type = SWS_PIXEL_F32) \ MACRO(__VA_ARGS__, f32_read_planar_fv_xyzw_f32 , .type = SWS_PIXEL_F32, .uop = SWS_UOP_READ_PLANAR_FV , .mask = 0xf, .par.filter.type = SWS_PIXEL_F32) -#define SWS_FOR_F32_READ_PLANAR_FV_FMA(MACRO, ...) \ - MACRO(__VA_ARGS__, f32_read_planar_fv_fma_x_f32 , SWS_PIXEL_F32, SWS_UOP_READ_PLANAR_FV_FMA, 0x1, SWS_PIXEL_F32) \ - MACRO(__VA_ARGS__, f32_read_planar_fv_fma_xy_f32 , SWS_PIXEL_F32, SWS_UOP_READ_PLANAR_FV_FMA, 0x3, SWS_PIXEL_F32) \ - MACRO(__VA_ARGS__, f32_read_planar_fv_fma_xyz_f32 , SWS_PIXEL_F32, SWS_UOP_READ_PLANAR_FV_FMA, 0x7, SWS_PIXEL_F32) \ - MACRO(__VA_ARGS__, f32_read_planar_fv_fma_xyzw_f32 , SWS_PIXEL_F32, SWS_UOP_READ_PLANAR_FV_FMA, 0xf, SWS_PIXEL_F32) -#define SWS_FOR_STRUCT_F32_READ_PLANAR_FV_FMA(MACRO, ...) \ - MACRO(__VA_ARGS__, f32_read_planar_fv_fma_x_f32 , .type = SWS_PIXEL_F32, .uop = SWS_UOP_READ_PLANAR_FV_FMA, .mask = 0x1, .par.filter.type = SWS_PIXEL_F32) \ - MACRO(__VA_ARGS__, f32_read_planar_fv_fma_xy_f32 , .type = SWS_PIXEL_F32, .uop = SWS_UOP_READ_PLANAR_FV_FMA, .mask = 0x3, .par.filter.type = SWS_PIXEL_F32) \ - MACRO(__VA_ARGS__, f32_read_planar_fv_fma_xyz_f32 , .type = SWS_PIXEL_F32, .uop = SWS_UOP_READ_PLANAR_FV_FMA, .mask = 0x7, .par.filter.type = SWS_PIXEL_F32) \ - MACRO(__VA_ARGS__, f32_read_planar_fv_fma_xyzw_f32 , .type = SWS_PIXEL_F32, .uop = SWS_UOP_READ_PLANAR_FV_FMA, .mask = 0xf, .par.filter.type = SWS_PIXEL_F32) +#define SWS_FOR_F32_READ_PLANAR_FV_FMA(MACRO, ...) +#define SWS_FOR_STRUCT_F32_READ_PLANAR_FV_FMA(MACRO, ...) #define SWS_FOR_F32_READ_PACKED(MACRO, ...) #define SWS_FOR_STRUCT_F32_READ_PACKED(MACRO, ...) #define SWS_FOR_F32_READ_NIBBLE(MACRO, ...) @@ -1089,64 +1021,8 @@ MACRO(__VA_ARGS__, f32_linear_xyzw_xxx0x_xxx0x_xxx0x_000x0 , .type = SWS_PIXEL_F32, .uop = SWS_UOP_LINEAR , .mask = 0xf, .par.lin.one = 0x00000, .par.lin.zero = 0xba108) \ MACRO(__VA_ARGS__, f32_linear_xyzw_x0x0x_xxx0x_xx00x_000x0 , .type = SWS_PIXEL_F32, .uop = SWS_UOP_LINEAR , .mask = 0xf, .par.lin.one = 0x00000, .par.lin.zero = 0xbb10a) \ MACRO(__VA_ARGS__, f32_linear_xyzw_x0000_0x000_00x00_000x0 , .type = SWS_PIXEL_F32, .uop = SWS_UOP_LINEAR , .mask = 0xf, .par.lin.one = 0x00000, .par.lin.zero = 0xbefbe) -#define SWS_FOR_F32_LINEAR_FMA(MACRO, ...) \ - MACRO(__VA_ARGS__, f32_linear_fma_x_xxx0x , SWS_PIXEL_F32, SWS_UOP_LINEAR_FMA , 0x1, 0x00000, 0xfffe8, 0xfffe8) \ - MACRO(__VA_ARGS__, f32_linear_fma_x_XXX0x , SWS_PIXEL_F32, SWS_UOP_LINEAR_FMA , 0x1, 0x00000, 0xfffe8, 0xfffef) \ - MACRO(__VA_ARGS__, f32_linear_fma_x_x000x , SWS_PIXEL_F32, SWS_UOP_LINEAR_FMA , 0x1, 0x00000, 0xfffee, 0xfffee) \ - MACRO(__VA_ARGS__, f32_linear_fma_x_X000x , SWS_PIXEL_F32, SWS_UOP_LINEAR_FMA , 0x1, 0x00000, 0xfffee, 0xfffef) \ - MACRO(__VA_ARGS__, f32_linear_fma_x_xxx00 , SWS_PIXEL_F32, SWS_UOP_LINEAR_FMA , 0x1, 0x00000, 0xffff8, 0xffff8) \ - MACRO(__VA_ARGS__, f32_linear_fma_x_xXx00 , SWS_PIXEL_F32, SWS_UOP_LINEAR_FMA , 0x1, 0x00000, 0xffff8, 0xffffa) \ - MACRO(__VA_ARGS__, f32_linear_fma_x_xxX00 , SWS_PIXEL_F32, SWS_UOP_LINEAR_FMA , 0x1, 0x00000, 0xffff8, 0xffffc) \ - MACRO(__VA_ARGS__, f32_linear_fma_x_xXX00 , SWS_PIXEL_F32, SWS_UOP_LINEAR_FMA , 0x1, 0x00000, 0xffff8, 0xffffe) \ - MACRO(__VA_ARGS__, f32_linear_fma_y_0x000 , SWS_PIXEL_F32, SWS_UOP_LINEAR_FMA , 0x2, 0x00000, 0xfffbf, 0xfffbf) \ - MACRO(__VA_ARGS__, f32_linear_fma_xyz_xxx0x_xxx0x_xxx0x , SWS_PIXEL_F32, SWS_UOP_LINEAR_FMA , 0x7, 0x00000, 0xfa108, 0xfa108) \ - MACRO(__VA_ARGS__, f32_linear_fma_xyz_XXX0x_XxX0x_XXX0x , SWS_PIXEL_F32, SWS_UOP_LINEAR_FMA , 0x7, 0x00000, 0xfa108, 0xfbdaf) \ - MACRO(__VA_ARGS__, f32_linear_fma_xyz_XXX0x_XXX0x_XXX0x , SWS_PIXEL_F32, SWS_UOP_LINEAR_FMA , 0x7, 0x00000, 0xfa108, 0xfbdef) \ - MACRO(__VA_ARGS__, f32_linear_fma_xyz_X0X0x_XXX0x_XX00x , SWS_PIXEL_F32, SWS_UOP_LINEAR_FMA , 0x7, 0x00000, 0xfb10a, 0xfbdef) \ - MACRO(__VA_ARGS__, f32_linear_fma_xyz_xxx00_xxx0x_xxx0x , SWS_PIXEL_F32, SWS_UOP_LINEAR_FMA , 0x7, 0x00000, 0xfa118, 0xfa118) \ - MACRO(__VA_ARGS__, f32_linear_fma_xyz_xXX00_XxX0x_XXX0x , SWS_PIXEL_F32, SWS_UOP_LINEAR_FMA , 0x7, 0x00000, 0xfa118, 0xfbdbe) \ - MACRO(__VA_ARGS__, f32_linear_fma_xyz_xXX00_XXX0x_XXX0x , SWS_PIXEL_F32, SWS_UOP_LINEAR_FMA , 0x7, 0x00000, 0xfa118, 0xfbdfe) \ - MACRO(__VA_ARGS__, f32_linear_fma_xyz_X000x_0X00x_00X0x , SWS_PIXEL_F32, SWS_UOP_LINEAR_FMA , 0x7, 0x00000, 0xfadae, 0xfbdef) \ - MACRO(__VA_ARGS__, f32_linear_fma_xyz_x0000_0x000_00x00 , SWS_PIXEL_F32, SWS_UOP_LINEAR_FMA , 0x7, 0x00000, 0xfefbe, 0xfefbe) \ - MACRO(__VA_ARGS__, f32_linear_fma_xyz_10X0x_1XX0x_1X00x , SWS_PIXEL_F32, SWS_UOP_LINEAR_FMA , 0x7, 0x00421, 0xfb10a, 0xfbdef) \ - MACRO(__VA_ARGS__, f32_linear_fma_w_000x0 , SWS_PIXEL_F32, SWS_UOP_LINEAR_FMA , 0x8, 0x00000, 0xbffff, 0xbffff) \ - MACRO(__VA_ARGS__, f32_linear_fma_xw_x000x_000x0 , SWS_PIXEL_F32, SWS_UOP_LINEAR_FMA , 0x9, 0x00000, 0xbffee, 0xbffee) \ - MACRO(__VA_ARGS__, f32_linear_fma_xw_X000x_000x0 , SWS_PIXEL_F32, SWS_UOP_LINEAR_FMA , 0x9, 0x00000, 0xbffee, 0xbffef) \ - MACRO(__VA_ARGS__, f32_linear_fma_xw_xxx00_000x0 , SWS_PIXEL_F32, SWS_UOP_LINEAR_FMA , 0x9, 0x00000, 0xbfff8, 0xbfff8) \ - MACRO(__VA_ARGS__, f32_linear_fma_xw_xXX00_000x0 , SWS_PIXEL_F32, SWS_UOP_LINEAR_FMA , 0x9, 0x00000, 0xbfff8, 0xbfffe) \ - MACRO(__VA_ARGS__, f32_linear_fma_xyzw_xxx0x_xxx0x_xxx0x_000x0, SWS_PIXEL_F32, SWS_UOP_LINEAR_FMA , 0xf, 0x00000, 0xba108, 0xba108) \ - MACRO(__VA_ARGS__, f32_linear_fma_xyzw_XXX0x_XXX0x_XXX0x_000x0, SWS_PIXEL_F32, SWS_UOP_LINEAR_FMA , 0xf, 0x00000, 0xba108, 0xbbdef) \ - MACRO(__VA_ARGS__, f32_linear_fma_xyzw_X0X0x_XXX0x_XX00x_000x0, SWS_PIXEL_F32, SWS_UOP_LINEAR_FMA , 0xf, 0x00000, 0xbb10a, 0xbbdef) \ - MACRO(__VA_ARGS__, f32_linear_fma_xyzw_x0000_0x000_00x00_000x0, SWS_PIXEL_F32, SWS_UOP_LINEAR_FMA , 0xf, 0x00000, 0xbefbe, 0xbefbe) -#define SWS_FOR_STRUCT_F32_LINEAR_FMA(MACRO, ...) \ - MACRO(__VA_ARGS__, f32_linear_fma_x_xxx0x , .type = SWS_PIXEL_F32, .uop = SWS_UOP_LINEAR_FMA , .mask = 0x1, .par.lin.one = 0x00000, .par.lin.zero = 0xfffe8, .par.lin.exact = 0xfffe8) \ - MACRO(__VA_ARGS__, f32_linear_fma_x_XXX0x , .type = SWS_PIXEL_F32, .uop = SWS_UOP_LINEAR_FMA , .mask = 0x1, .par.lin.one = 0x00000, .par.lin.zero = 0xfffe8, .par.lin.exact = 0xfffef) \ - MACRO(__VA_ARGS__, f32_linear_fma_x_x000x , .type = SWS_PIXEL_F32, .uop = SWS_UOP_LINEAR_FMA , .mask = 0x1, .par.lin.one = 0x00000, .par.lin.zero = 0xfffee, .par.lin.exact = 0xfffee) \ - MACRO(__VA_ARGS__, f32_linear_fma_x_X000x , .type = SWS_PIXEL_F32, .uop = SWS_UOP_LINEAR_FMA , .mask = 0x1, .par.lin.one = 0x00000, .par.lin.zero = 0xfffee, .par.lin.exact = 0xfffef) \ - MACRO(__VA_ARGS__, f32_linear_fma_x_xxx00 , .type = SWS_PIXEL_F32, .uop = SWS_UOP_LINEAR_FMA , .mask = 0x1, .par.lin.one = 0x00000, .par.lin.zero = 0xffff8, .par.lin.exact = 0xffff8) \ - MACRO(__VA_ARGS__, f32_linear_fma_x_xXx00 , .type = SWS_PIXEL_F32, .uop = SWS_UOP_LINEAR_FMA , .mask = 0x1, .par.lin.one = 0x00000, .par.lin.zero = 0xffff8, .par.lin.exact = 0xffffa) \ - MACRO(__VA_ARGS__, f32_linear_fma_x_xxX00 , .type = SWS_PIXEL_F32, .uop = SWS_UOP_LINEAR_FMA , .mask = 0x1, .par.lin.one = 0x00000, .par.lin.zero = 0xffff8, .par.lin.exact = 0xffffc) \ - MACRO(__VA_ARGS__, f32_linear_fma_x_xXX00 , .type = SWS_PIXEL_F32, .uop = SWS_UOP_LINEAR_FMA , .mask = 0x1, .par.lin.one = 0x00000, .par.lin.zero = 0xffff8, .par.lin.exact = 0xffffe) \ - MACRO(__VA_ARGS__, f32_linear_fma_y_0x000 , .type = SWS_PIXEL_F32, .uop = SWS_UOP_LINEAR_FMA , .mask = 0x2, .par.lin.one = 0x00000, .par.lin.zero = 0xfffbf, .par.lin.exact = 0xfffbf) \ - MACRO(__VA_ARGS__, f32_linear_fma_xyz_xxx0x_xxx0x_xxx0x , .type = SWS_PIXEL_F32, .uop = SWS_UOP_LINEAR_FMA , .mask = 0x7, .par.lin.one = 0x00000, .par.lin.zero = 0xfa108, .par.lin.exact = 0xfa108) \ - MACRO(__VA_ARGS__, f32_linear_fma_xyz_XXX0x_XxX0x_XXX0x , .type = SWS_PIXEL_F32, .uop = SWS_UOP_LINEAR_FMA , .mask = 0x7, .par.lin.one = 0x00000, .par.lin.zero = 0xfa108, .par.lin.exact = 0xfbdaf) \ - MACRO(__VA_ARGS__, f32_linear_fma_xyz_XXX0x_XXX0x_XXX0x , .type = SWS_PIXEL_F32, .uop = SWS_UOP_LINEAR_FMA , .mask = 0x7, .par.lin.one = 0x00000, .par.lin.zero = 0xfa108, .par.lin.exact = 0xfbdef) \ - MACRO(__VA_ARGS__, f32_linear_fma_xyz_X0X0x_XXX0x_XX00x , .type = SWS_PIXEL_F32, .uop = SWS_UOP_LINEAR_FMA , .mask = 0x7, .par.lin.one = 0x00000, .par.lin.zero = 0xfb10a, .par.lin.exact = 0xfbdef) \ - MACRO(__VA_ARGS__, f32_linear_fma_xyz_xxx00_xxx0x_xxx0x , .type = SWS_PIXEL_F32, .uop = SWS_UOP_LINEAR_FMA , .mask = 0x7, .par.lin.one = 0x00000, .par.lin.zero = 0xfa118, .par.lin.exact = 0xfa118) \ - MACRO(__VA_ARGS__, f32_linear_fma_xyz_xXX00_XxX0x_XXX0x , .type = SWS_PIXEL_F32, .uop = SWS_UOP_LINEAR_FMA , .mask = 0x7, .par.lin.one = 0x00000, .par.lin.zero = 0xfa118, .par.lin.exact = 0xfbdbe) \ - MACRO(__VA_ARGS__, f32_linear_fma_xyz_xXX00_XXX0x_XXX0x , .type = SWS_PIXEL_F32, .uop = SWS_UOP_LINEAR_FMA , .mask = 0x7, .par.lin.one = 0x00000, .par.lin.zero = 0xfa118, .par.lin.exact = 0xfbdfe) \ - MACRO(__VA_ARGS__, f32_linear_fma_xyz_X000x_0X00x_00X0x , .type = SWS_PIXEL_F32, .uop = SWS_UOP_LINEAR_FMA , .mask = 0x7, .par.lin.one = 0x00000, .par.lin.zero = 0xfadae, .par.lin.exact = 0xfbdef) \ - MACRO(__VA_ARGS__, f32_linear_fma_xyz_x0000_0x000_00x00 , .type = SWS_PIXEL_F32, .uop = SWS_UOP_LINEAR_FMA , .mask = 0x7, .par.lin.one = 0x00000, .par.lin.zero = 0xfefbe, .par.lin.exact = 0xfefbe) \ - MACRO(__VA_ARGS__, f32_linear_fma_xyz_10X0x_1XX0x_1X00x , .type = SWS_PIXEL_F32, .uop = SWS_UOP_LINEAR_FMA , .mask = 0x7, .par.lin.one = 0x00421, .par.lin.zero = 0xfb10a, .par.lin.exact = 0xfbdef) \ - MACRO(__VA_ARGS__, f32_linear_fma_w_000x0 , .type = SWS_PIXEL_F32, .uop = SWS_UOP_LINEAR_FMA , .mask = 0x8, .par.lin.one = 0x00000, .par.lin.zero = 0xbffff, .par.lin.exact = 0xbffff) \ - MACRO(__VA_ARGS__, f32_linear_fma_xw_x000x_000x0 , .type = SWS_PIXEL_F32, .uop = SWS_UOP_LINEAR_FMA , .mask = 0x9, .par.lin.one = 0x00000, .par.lin.zero = 0xbffee, .par.lin.exact = 0xbffee) \ - MACRO(__VA_ARGS__, f32_linear_fma_xw_X000x_000x0 , .type = SWS_PIXEL_F32, .uop = SWS_UOP_LINEAR_FMA , .mask = 0x9, .par.lin.one = 0x00000, .par.lin.zero = 0xbffee, .par.lin.exact = 0xbffef) \ - MACRO(__VA_ARGS__, f32_linear_fma_xw_xxx00_000x0 , .type = SWS_PIXEL_F32, .uop = SWS_UOP_LINEAR_FMA , .mask = 0x9, .par.lin.one = 0x00000, .par.lin.zero = 0xbfff8, .par.lin.exact = 0xbfff8) \ - MACRO(__VA_ARGS__, f32_linear_fma_xw_xXX00_000x0 , .type = SWS_PIXEL_F32, .uop = SWS_UOP_LINEAR_FMA , .mask = 0x9, .par.lin.one = 0x00000, .par.lin.zero = 0xbfff8, .par.lin.exact = 0xbfffe) \ - MACRO(__VA_ARGS__, f32_linear_fma_xyzw_xxx0x_xxx0x_xxx0x_000x0, .type = SWS_PIXEL_F32, .uop = SWS_UOP_LINEAR_FMA , .mask = 0xf, .par.lin.one = 0x00000, .par.lin.zero = 0xba108, .par.lin.exact = 0xba108) \ - MACRO(__VA_ARGS__, f32_linear_fma_xyzw_XXX0x_XXX0x_XXX0x_000x0, .type = SWS_PIXEL_F32, .uop = SWS_UOP_LINEAR_FMA , .mask = 0xf, .par.lin.one = 0x00000, .par.lin.zero = 0xba108, .par.lin.exact = 0xbbdef) \ - MACRO(__VA_ARGS__, f32_linear_fma_xyzw_X0X0x_XXX0x_XX00x_000x0, .type = SWS_PIXEL_F32, .uop = SWS_UOP_LINEAR_FMA , .mask = 0xf, .par.lin.one = 0x00000, .par.lin.zero = 0xbb10a, .par.lin.exact = 0xbbdef) \ - MACRO(__VA_ARGS__, f32_linear_fma_xyzw_x0000_0x000_00x00_000x0, .type = SWS_PIXEL_F32, .uop = SWS_UOP_LINEAR_FMA , .mask = 0xf, .par.lin.one = 0x00000, .par.lin.zero = 0xbefbe, .par.lin.exact = 0xbefbe) +#define SWS_FOR_F32_LINEAR_FMA(MACRO, ...) +#define SWS_FOR_STRUCT_F32_LINEAR_FMA(MACRO, ...) #define SWS_FOR_F32_DITHER(MACRO, ...) \ MACRO(__VA_ARGS__, f32_dither_x_0_16x16 , SWS_PIXEL_F32, SWS_UOP_DITHER , 0x1, 0, 0, 0, 0, 4) \ MACRO(__VA_ARGS__, f32_dither_y_3_16x16 , SWS_PIXEL_F32, SWS_UOP_DITHER , 0x2, 0, 3, 0, 0, 4) \ diff --git a/libswscale/uops_macros_gen.c b/libswscale/uops_macros_gen.c index 6985451168..40d3e888b9 100644 --- a/libswscale/uops_macros_gen.c +++ b/libswscale/uops_macros_gen.c @@ -228,12 +228,15 @@ typedef struct Backend { static const Backend backends[] = { { - /* Currently also covers all uops used by the x86 backend */ .name = "c", .guard = "SWSCALE_UOPS_MACROS_H", + .flags = { COMMON_FLAGS }, + }, { + .name = "x86", + .guard = "SWSCALE_X86_UOPS_MACROS_H", .flags = { - COMMON_FLAGS, - COMMON_FLAGS | SWS_UOP_FLAG_PSHUFB | SWS_UOP_FLAG_FMA, + COMMON_FLAGS | SWS_UOP_FLAG_PSHUFB, + COMMON_FLAGS | SWS_UOP_FLAG_PSHUFB | SWS_UOP_FLAG_FMA, /* FMA3 */ }, }, }; @@ -490,7 +493,7 @@ static int usage(void) { fprintf(stderr, "Usage: uops_macros_gen [-backend BACKEND] [-nasm FILE]\n" "Writes C macros to stdout, and optionally NASM macros to FILE\n" - "BACKEND is one of: c (default)\n"); + "BACKEND is one of: c (default), x86\n"); return 1; } diff --git a/libswscale/x86/ops.c b/libswscale/x86/ops.c index 23785c58ca..7985b7a026 100644 --- a/libswscale/x86/ops.c +++ b/libswscale/x86/ops.c @@ -26,7 +26,7 @@ #include "../ops_chain.h" #include "../uops.h" -#include "../uops_macros.h" +#include "uops_macros.h" static int setup_rw_packed(const SwsImplParams *params, SwsImplResult *out) { diff --git a/libswscale/x86/uops_macros.asm b/libswscale/x86/uops_macros.asm index 6df56d0e8f..d7d539f286 100644 --- a/libswscale/x86/uops_macros.asm +++ b/libswscale/x86/uops_macros.asm @@ -108,8 +108,6 @@ DECL_OP %1, u8_permute_xyz_yx_xw , SWS_PIXEL_U8 , SWS_UOP_PERMUTE , 0x7, 2, 1, 0, 0, 0, 0, 0, 0, 3, 0, 0, 0, 0 DECL_OP %1, u8_permute_xyz_xyz_yzw , SWS_PIXEL_U8 , SWS_UOP_PERMUTE , 0x7, 3, 0, 1, 2, 0, 0, 0, 1, 2, 3, 0, 0, 0 DECL_OP %1, u8_permute_xyz_xtyz_wyzt , SWS_PIXEL_U8 , SWS_UOP_PERMUTE , 0x7, 4, 0, -1, 1, 2, 0, 0, 3, 1, 2, -1, 0, 0 - DECL_OP %1, u8_permute_xyw_x_z , SWS_PIXEL_U8 , SWS_UOP_PERMUTE , 0xb, 1, 0, 0, 0, 0, 0, 0, 2, 0, 0, 0, 0, 0 - DECL_OP %1, u8_permute_xzw_x_y , SWS_PIXEL_U8 , SWS_UOP_PERMUTE , 0xd, 1, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0 DECL_OP %1, u8_permute_yzw_y_x , SWS_PIXEL_U8 , SWS_UOP_PERMUTE , 0xe, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 DECL_OP %1, u8_permute_yzw_z_x , SWS_PIXEL_U8 , SWS_UOP_PERMUTE , 0xe, 1, 2, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 DECL_OP %1, u8_permute_yzw_w_x , SWS_PIXEL_U8 , SWS_UOP_PERMUTE , 0xe, 1, 3, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 diff --git a/libswscale/uops_macros.h b/libswscale/x86/uops_macros.h similarity index 99% copy from libswscale/uops_macros.h copy to libswscale/x86/uops_macros.h index 9e300fbac1..91499a1d86 100644 --- a/libswscale/uops_macros.h +++ b/libswscale/x86/uops_macros.h @@ -3,8 +3,8 @@ * To regenerate, run: make fate-sws-uops-macros GEN=1 */ -#ifndef SWSCALE_UOPS_MACROS_H -#define SWSCALE_UOPS_MACROS_H +#ifndef SWSCALE_X86_UOPS_MACROS_H +#define SWSCALE_X86_UOPS_MACROS_H /** * Boilerplate helper macros, for template-based backends. These will be @@ -159,8 +159,6 @@ MACRO(__VA_ARGS__, u8_permute_xyz_yx_xw , SWS_PIXEL_U8 , SWS_UOP_PERMUTE , 0x7, 2, 1, 0, 0, 0, 0, 0, 0, 3, 0, 0, 0, 0) \ MACRO(__VA_ARGS__, u8_permute_xyz_xyz_yzw , SWS_PIXEL_U8 , SWS_UOP_PERMUTE , 0x7, 3, 0, 1, 2, 0, 0, 0, 1, 2, 3, 0, 0, 0) \ MACRO(__VA_ARGS__, u8_permute_xyz_xtyz_wyzt , SWS_PIXEL_U8 , SWS_UOP_PERMUTE , 0x7, 4, 0, -1, 1, 2, 0, 0, 3, 1, 2, -1, 0, 0) \ - MACRO(__VA_ARGS__, u8_permute_xyw_x_z , SWS_PIXEL_U8 , SWS_UOP_PERMUTE , 0xb, 1, 0, 0, 0, 0, 0, 0, 2, 0, 0, 0, 0, 0) \ - MACRO(__VA_ARGS__, u8_permute_xzw_x_y , SWS_PIXEL_U8 , SWS_UOP_PERMUTE , 0xd, 1, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0) \ MACRO(__VA_ARGS__, u8_permute_yzw_y_x , SWS_PIXEL_U8 , SWS_UOP_PERMUTE , 0xe, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0) \ MACRO(__VA_ARGS__, u8_permute_yzw_z_x , SWS_PIXEL_U8 , SWS_UOP_PERMUTE , 0xe, 1, 2, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0) \ MACRO(__VA_ARGS__, u8_permute_yzw_w_x , SWS_PIXEL_U8 , SWS_UOP_PERMUTE , 0xe, 1, 3, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0) \ @@ -189,8 +187,6 @@ MACRO(__VA_ARGS__, u8_permute_xyz_yx_xw , .type = SWS_PIXEL_U8 , .uop = SWS_UOP_PERMUTE , .mask = 0x7, .par.move.num_moves = 2, .par.move.dst = {1, 0, 0, 0, 0, 0}, .par.move.src = {0, 3, 0, 0, 0, 0}) \ MACRO(__VA_ARGS__, u8_permute_xyz_xyz_yzw , .type = SWS_PIXEL_U8 , .uop = SWS_UOP_PERMUTE , .mask = 0x7, .par.move.num_moves = 3, .par.move.dst = {0, 1, 2, 0, 0, 0}, .par.move.src = {1, 2, 3, 0, 0, 0}) \ MACRO(__VA_ARGS__, u8_permute_xyz_xtyz_wyzt , .type = SWS_PIXEL_U8 , .uop = SWS_UOP_PERMUTE , .mask = 0x7, .par.move.num_moves = 4, .par.move.dst = {0, -1, 1, 2, 0, 0}, .par.move.src = {3, 1, 2, -1, 0, 0}) \ - MACRO(__VA_ARGS__, u8_permute_xyw_x_z , .type = SWS_PIXEL_U8 , .uop = SWS_UOP_PERMUTE , .mask = 0xb, .par.move.num_moves = 1, .par.move.dst = {0, 0, 0, 0, 0, 0}, .par.move.src = {2, 0, 0, 0, 0, 0}) \ - MACRO(__VA_ARGS__, u8_permute_xzw_x_y , .type = SWS_PIXEL_U8 , .uop = SWS_UOP_PERMUTE , .mask = 0xd, .par.move.num_moves = 1, .par.move.dst = {0, 0, 0, 0, 0, 0}, .par.move.src = {1, 0, 0, 0, 0, 0}) \ MACRO(__VA_ARGS__, u8_permute_yzw_y_x , .type = SWS_PIXEL_U8 , .uop = SWS_UOP_PERMUTE , .mask = 0xe, .par.move.num_moves = 1, .par.move.dst = {1, 0, 0, 0, 0, 0}, .par.move.src = {0, 0, 0, 0, 0, 0}) \ MACRO(__VA_ARGS__, u8_permute_yzw_z_x , .type = SWS_PIXEL_U8 , .uop = SWS_UOP_PERMUTE , .mask = 0xe, .par.move.num_moves = 1, .par.move.dst = {2, 0, 0, 0, 0, 0}, .par.move.src = {0, 0, 0, 0, 0, 0}) \ MACRO(__VA_ARGS__, u8_permute_yzw_w_x , .type = SWS_PIXEL_U8 , .uop = SWS_UOP_PERMUTE , .mask = 0xe, .par.move.num_moves = 1, .par.move.dst = {3, 0, 0, 0, 0, 0}, .par.move.src = {0, 0, 0, 0, 0, 0}) \ @@ -1208,4 +1204,4 @@ MACRO(__VA_ARGS__, f32_lut_3d_xyzw_static , .type = SWS_PIXEL_F32, .uop = SWS_UOP_LUT_3D , .mask = 0xf, .par.lut3d.dynamic = 0) \ MACRO(__VA_ARGS__, f32_lut_3d_xyzw_dynamic , .type = SWS_PIXEL_F32, .uop = SWS_UOP_LUT_3D , .mask = 0xf, .par.lut3d.dynamic = 1) -#endif /* SWSCALE_UOPS_MACROS_H */ +#endif /* SWSCALE_X86_UOPS_MACROS_H */ diff --git a/tests/fate/libswscale.mak b/tests/fate/libswscale.mak index 5ccab1ad4f..802ac72a03 100644 --- a/tests/fate/libswscale.mak +++ b/tests/fate/libswscale.mak @@ -103,6 +103,7 @@ fate-sws-ops-list: libswscale/tests/sws_ops$(EXESUF) fate-sws-ops-list: CMD = run libswscale/tests/sws_ops$(EXESUF) | do_md5sum | cut -d" " -f1 FATE_SWS_UOPS_MACROS = fate-sws-uops-macros-c \ + fate-sws-uops-macros-x86 \ fate-sws-uops-macros-x86-nasm \ ifeq ($(HAVE_INT128),yes) @@ -114,13 +115,17 @@ endif # (and thus output) due to sorting by memcmp() on the parameters struct. fate-sws-uops-macros-c: libswscale/uops_macros_gen$(EXESUF) fate-sws-uops-macros-c: REF = $(SRC_PATH)/libswscale/uops_macros.h -fate-sws-uops-macros-c: CMD = run libswscale/uops_macros_gen$(EXESUF) \ - -nasm $(TARGET_PATH)/tests/data/fate/sws-uops-macros.asm +fate-sws-uops-macros-c: CMD = run libswscale/uops_macros_gen$(EXESUF) -# Generated together with fate-sws-uops-macros-c, since enumeration is slow -fate-sws-uops-macros-x86-nasm: fate-sws-uops-macros-c +fate-sws-uops-macros-x86: libswscale/uops_macros_gen$(EXESUF) +fate-sws-uops-macros-x86: REF = $(SRC_PATH)/libswscale/x86/uops_macros.h +fate-sws-uops-macros-x86: CMD = run libswscale/uops_macros_gen$(EXESUF) -backend x86 \ + -nasm $(TARGET_PATH)/tests/data/fate/sws-uops-macros-x86.asm + +# Generated together with fate-sws-uops-macros-x86, since enumeration is slow +fate-sws-uops-macros-x86-nasm: fate-sws-uops-macros-x86 fate-sws-uops-macros-x86-nasm: REF = $(SRC_PATH)/libswscale/x86/uops_macros.asm -fate-sws-uops-macros-x86-nasm: CMD = cat tests/data/fate/sws-uops-macros.asm +fate-sws-uops-macros-x86-nasm: CMD = cat tests/data/fate/sws-uops-macros-x86.asm fate-sws-uops-macros: $(FATE_SWS_UOPS_MACROS) diff --git a/tests/ref/fate/source b/tests/ref/fate/source index 9fdd18cd70..614f6f7183 100644 --- a/tests/ref/fate/source +++ b/tests/ref/fate/source @@ -22,6 +22,7 @@ libswscale/log2_tab.c libswscale/riscv/cpu_common.c libswscale/uops_macros.h libswscale/x86/uops_macros.asm +libswscale/x86/uops_macros.h tools/uncoded_frame.c tools/yuvcmp.c Headers without standard inclusion guards: -- To stop receiving notification emails like this one, please contact [email protected]. _______________________________________________ ffmpeg-cvslog mailing list -- [email protected] To unsubscribe send an email to [email protected]
