ref: 36f76f424f410f1d0427a9df55a34d8485e2ec8e
parent: 502e5b92ecb00d6993440ee82afb3c34657a8c7f
author: Ronald S. Bultje <rsbultje@gmail.com>
date: Mon Dec 17 04:06:20 EST 2018
Remove carriage return characters
--- a/include/compat/msvc/stdatomic.h
+++ b/include/compat/msvc/stdatomic.h
@@ -23,10 +23,10 @@
* (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
* SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*/
-
-#ifndef MSCVER_STDATOMIC_H_
-#define MSCVER_STDATOMIC_H_
-
+
+#ifndef MSCVER_STDATOMIC_H_
+#define MSCVER_STDATOMIC_H_
+
#if !defined(__cplusplus) && defined(_MSC_VER)
#pragma warning(push)
@@ -37,34 +37,34 @@
# include_next <stdatomic.h>
#else /* ! stdatomic.h */
-#include <windows.h>
-
-#include "common/attributes.h"
-
-typedef volatile LONG __declspec(align(32)) atomic_int;
-typedef volatile ULONG __declspec(align(32)) atomic_uint;
-
-typedef enum {
- memory_order_relaxed,
- memory_order_acquire
-} msvc_atomic_memory_order;
-
-#define atomic_init(p_a, v) do { *(p_a) = (v); } while(0)
-#define atomic_store(p_a, v) InterlockedExchange((LONG*)p_a, v)
-#define atomic_load(p_a) InterlockedCompareExchange((LONG*)p_a, 0, 0)
-#define atomic_load_explicit(p_a, mo) atomic_load(p_a)
-
-/*
- * TODO use a special call to increment/decrement
- * using InterlockedIncrement/InterlockedDecrement
- */
-#define atomic_fetch_add(p_a, inc) InterlockedExchangeAdd(p_a, inc)
-#define atomic_fetch_sub(p_a, dec) InterlockedExchangeAdd(p_a, -(dec))
-
-#endif /* ! stdatomic.h */
-
-#pragma warning(pop)
-
-#endif /* !defined(__cplusplus) && defined(_MSC_VER) */
-
-#endif /* MSCVER_STDATOMIC_H_ */
+#include <windows.h>
+
+#include "common/attributes.h"
+
+typedef volatile LONG __declspec(align(32)) atomic_int;
+typedef volatile ULONG __declspec(align(32)) atomic_uint;
+
+typedef enum {+ memory_order_relaxed,
+ memory_order_acquire
+} msvc_atomic_memory_order;
+
+#define atomic_init(p_a, v) do { *(p_a) = (v); } while(0)+#define atomic_store(p_a, v) InterlockedExchange((LONG*)p_a, v)
+#define atomic_load(p_a) InterlockedCompareExchange((LONG*)p_a, 0, 0)
+#define atomic_load_explicit(p_a, mo) atomic_load(p_a)
+
+/*
+ * TODO use a special call to increment/decrement
+ * using InterlockedIncrement/InterlockedDecrement
+ */
+#define atomic_fetch_add(p_a, inc) InterlockedExchangeAdd(p_a, inc)
+#define atomic_fetch_sub(p_a, dec) InterlockedExchangeAdd(p_a, -(dec))
+
+#endif /* ! stdatomic.h */
+
+#pragma warning(pop)
+
+#endif /* !defined(__cplusplus) && defined(_MSC_VER) */
+
+#endif /* MSCVER_STDATOMIC_H_ */
--- a/src/x86/itx_ssse3.asm
+++ b/src/x86/itx_ssse3.asm
@@ -31,76 +31,76 @@
deint_shuf: db 0, 1, 4, 5, 8, 9, 12, 13, 2, 3, 6, 7, 10, 11, 14, 15
-qw_2896x8: times 8 dw 2896*8
-qw_1567_m3784: times 4 dw 1567, -3784
-qw_3784_1567: times 4 dw 3784, 1567
-
-qw_1321_3803: times 4 dw 1321, 3803
-qw_2482_m1321: times 4 dw 2482, -1321
-qw_3344_2482: times 4 dw 3344, 2482
-qw_3344_m3803: times 4 dw 3344, -3803
-qw_m6688_m3803: times 4 dw -6688, -3803
-qw_3344x8: times 8 dw 3344*8
-qw_5793x4: times 8 dw 5793*4
-
-pd_2048: times 4 dd 2048
-qw_2048: times 8 dw 2048
-
+qw_2896x8: times 8 dw 2896*8
+qw_1567_m3784: times 4 dw 1567, -3784
+qw_3784_1567: times 4 dw 3784, 1567
+
+qw_1321_3803: times 4 dw 1321, 3803
+qw_2482_m1321: times 4 dw 2482, -1321
+qw_3344_2482: times 4 dw 3344, 2482
+qw_3344_m3803: times 4 dw 3344, -3803
+qw_m6688_m3803: times 4 dw -6688, -3803
+qw_3344x8: times 8 dw 3344*8
+qw_5793x4: times 8 dw 5793*4
+
+pd_2048: times 4 dd 2048
+qw_2048: times 8 dw 2048
+
iadst4_dconly1a: times 2 dw 10568, 19856, 26752, 30424
-iadst4_dconly1b: times 2 dw 30424, 26752, 19856, 10568
+iadst4_dconly1b: times 2 dw 30424, 26752, 19856, 10568
iadst4_dconly2a: dw 10568, 10568, 10568, 10568, 19856, 19856, 19856, 19856
-iadst4_dconly2b: dw 26752, 26752, 26752, 26752, 30424, 30424, 30424, 30424
-
-SECTION .text
-
-%define m(x) mangle(private_prefix %+ _ %+ x %+ SUFFIX)
-
-%macro ITX4_END 4-5 2048 ; row[1-4], rnd
-%if %5
- mova m2, [qw_%5]
- pmulhrsw m0, m2
- pmulhrsw m1, m2
-%endif
- lea r2, [dstq+strideq*2]
-%assign %%i 1
-%rep 4
- %if %1 & 2
- CAT_XDEFINE %%row_adr, %%i, r2 + strideq*(%1&1)
- %else
- CAT_XDEFINE %%row_adr, %%i, dstq + strideq*(%1&1)
- %endif
- %assign %%i %%i + 1
- %rotate 1
-%endrep
-
- movd m2, [%%row_adr1] ;dst0
- movd m4, [%%row_adr2] ;dst1
- punpckldq m2, m4 ;high: dst1 :low: dst0
- movd m3, [%%row_adr3] ;dst2
- movd m4, [%%row_adr4] ;dst3
- punpckldq m3, m4 ;high: dst3 :low: dst2
-
- pxor m4, m4
- punpcklbw m2, m4 ;extend byte to word
- punpcklbw m3, m4 ;extend byte to word
-
- paddw m0, m2 ;high: dst1 + out1 ;low: dst0 + out0
- paddw m1, m3 ;high: dst3 + out3 ;low: dst2 + out2
-
- packuswb m0, m1 ;high->low: dst3 + out3, dst2 + out2, dst1 + out1, dst0 + out0
-
- movd [%%row_adr1], m0 ;store dst0 + out0
- pshuflw m1, m0, q1032
- movd [%%row_adr2], m1 ;store dst1 + out1
- punpckhqdq m0, m0
- movd [%%row_adr3], m0 ;store dst2 + out2
- psrlq m0, 32
- movd [%%row_adr4], m0 ;store dst3 + out3
-
- ret
-%endmacro
-
-
+iadst4_dconly2b: dw 26752, 26752, 26752, 26752, 30424, 30424, 30424, 30424
+
+SECTION .text
+
+%define m(x) mangle(private_prefix %+ _ %+ x %+ SUFFIX)
+
+%macro ITX4_END 4-5 2048 ; row[1-4], rnd
+%if %5
+ mova m2, [qw_%5]
+ pmulhrsw m0, m2
+ pmulhrsw m1, m2
+%endif
+ lea r2, [dstq+strideq*2]
+%assign %%i 1
+%rep 4
+ %if %1 & 2
+ CAT_XDEFINE %%row_adr, %%i, r2 + strideq*(%1&1)
+ %else
+ CAT_XDEFINE %%row_adr, %%i, dstq + strideq*(%1&1)
+ %endif
+ %assign %%i %%i + 1
+ %rotate 1
+%endrep
+
+ movd m2, [%%row_adr1] ;dst0
+ movd m4, [%%row_adr2] ;dst1
+ punpckldq m2, m4 ;high: dst1 :low: dst0
+ movd m3, [%%row_adr3] ;dst2
+ movd m4, [%%row_adr4] ;dst3
+ punpckldq m3, m4 ;high: dst3 :low: dst2
+
+ pxor m4, m4
+ punpcklbw m2, m4 ;extend byte to word
+ punpcklbw m3, m4 ;extend byte to word
+
+ paddw m0, m2 ;high: dst1 + out1 ;low: dst0 + out0
+ paddw m1, m3 ;high: dst3 + out3 ;low: dst2 + out2
+
+ packuswb m0, m1 ;high->low: dst3 + out3, dst2 + out2, dst1 + out1, dst0 + out0
+
+ movd [%%row_adr1], m0 ;store dst0 + out0
+ pshuflw m1, m0, q1032
+ movd [%%row_adr2], m1 ;store dst1 + out1
+ punpckhqdq m0, m0
+ movd [%%row_adr3], m0 ;store dst2 + out2
+ psrlq m0, 32
+ movd [%%row_adr4], m0 ;store dst3 + out3
+
+ ret
+%endmacro
+
+
; flags: 1 = swap, 2: coef_regs
%macro ITX_MUL2X_PACK 5-6 0 ; dst/src, tmp[1], rnd, coef[1-2], flags
%if %6 & 2
@@ -118,27 +118,27 @@
psrad m%2, 12
psrad m%1, 12
packssdw m%1, m%2
-%endmacro
-
-%macro IDCT4_1D_PACKED 0-1 ;qw_2896x8
- punpckhwd m2, m0, m1 ;unpacked in1 in3
- psubw m3, m0, m1
- paddw m0, m1
- punpcklqdq m0, m3 ;high: in0-in2 ;low: in0+in2
-
- mova m3, [pd_2048]
- ITX_MUL2X_PACK 2, 1, 3, 1567, 3784
-
-%if %0 == 1
- pmulhrsw m0, m%1
-%else
- pmulhrsw m0, [qw_2896x8] ;high: t1 ;low: t0
-%endif
-
+%endmacro
+
+%macro IDCT4_1D_PACKED 0-1 ;qw_2896x8
+ punpckhwd m2, m0, m1 ;unpacked in1 in3
+ psubw m3, m0, m1
+ paddw m0, m1
+ punpcklqdq m0, m3 ;high: in0-in2 ;low: in0+in2
+
+ mova m3, [pd_2048]
+ ITX_MUL2X_PACK 2, 1, 3, 1567, 3784
+
+%if %0 == 1
+ pmulhrsw m0, m%1
+%else
+ pmulhrsw m0, [qw_2896x8] ;high: t1 ;low: t0
+%endif
+
psubsw m1, m0, m2 ;high: out2 ;low: out3
paddsw m0, m2 ;high: out1 ;low: out0
-%endmacro
-
+%endmacro
+
%macro IADST4_1D_PACKED 0
punpcklwd m2, m0, m1 ;unpacked in0 in2
punpckhwd m3, m0, m1 ;unpacked in1 in3
@@ -166,27 +166,27 @@
psrad m2, 12 ;out3
packssdw m0, m5 ;high: out1 ;low: out0
packssdw m2, m2 ;high: out3 ;low: out3
-%endmacro
-
-%macro INV_TXFM_FN 4 ; type1, type2, fast_thresh, size
-cglobal inv_txfm_add_%1_%2_%4, 4, 5, 0, dst, stride, coeff, eob, tx2
- %undef cmp
- lea tx2q, [m(i%2_%4_internal).pass2]
-%if %3 > 0
- cmp eobd, %3
- jle %%end
-%elif %3 == 0
- test eobd, eobd
- jz %%end
-%endif
+%endmacro
+
+%macro INV_TXFM_FN 4 ; type1, type2, fast_thresh, size
+cglobal inv_txfm_add_%1_%2_%4, 4, 5, 0, dst, stride, coeff, eob, tx2
+ %undef cmp
+ lea tx2q, [m(i%2_%4_internal).pass2]
+%if %3 > 0
+ cmp eobd, %3
+ jle %%end
+%elif %3 == 0
+ test eobd, eobd
+ jz %%end
+%endif
call i%1_%4_internal
- RET
-ALIGN function_align
-%%end:
-%endmacro
-
-%macro INV_TXFM_4X4_FN 2-3 -1 ; type1, type2, fast_thresh
- INV_TXFM_FN %1, %2, %3, 4x4
+ RET
+ALIGN function_align
+%%end:
+%endmacro
+
+%macro INV_TXFM_4X4_FN 2-3 -1 ; type1, type2, fast_thresh
+ INV_TXFM_FN %1, %2, %3, 4x4
%ifidn %1_%2, dct_identity
mova m0, [qw_2896x8]
pmulhrsw m0, [coeffq]
@@ -238,35 +238,35 @@
RET
%endif
%endif
-%endmacro
-
-
-INIT_XMM ssse3
-
-cglobal idct_4x4_internal, 0, 0, 4, dst, stride, coeff, eob, tx2
- mova m0, [coeffq+16*0] ;high: in1 ;low: in0
- mova m1, [coeffq+16*1] ;high: in3 ;low in2
-
- IDCT4_1D_PACKED
-
- mova m2, [deint_shuf]
- shufps m3, m0, m1, q1331
- shufps m0, m1, q0220
- pshufb m0, m2 ;high: in1 ;low: in0
- pshufb m1, m3, m2 ;high: in3 ;low :in2
+%endmacro
+
+
+INIT_XMM ssse3
+
+cglobal idct_4x4_internal, 0, 0, 4, dst, stride, coeff, eob, tx2
+ mova m0, [coeffq+16*0] ;high: in1 ;low: in0
+ mova m1, [coeffq+16*1] ;high: in3 ;low in2
+
+ IDCT4_1D_PACKED
+
+ mova m2, [deint_shuf]
+ shufps m3, m0, m1, q1331
+ shufps m0, m1, q0220
+ pshufb m0, m2 ;high: in1 ;low: in0
+ pshufb m1, m3, m2 ;high: in3 ;low :in2
jmp tx2q
-.pass2:
- IDCT4_1D_PACKED
-
- pxor m2, m2
- mova [coeffq+16*0], m2
- mova [coeffq+16*1], m2 ;memset(coeff, 0, sizeof(*coeff) * sh * sw);
-
- ITX4_END 0, 1, 3, 2
-
-INV_TXFM_4X4_FN dct, dct, 0
-
+.pass2:
+ IDCT4_1D_PACKED
+
+ pxor m2, m2
+ mova [coeffq+16*0], m2
+ mova [coeffq+16*1], m2 ;memset(coeff, 0, sizeof(*coeff) * sh * sw);
+
+ ITX4_END 0, 1, 3, 2
+
+INV_TXFM_4X4_FN dct, dct, 0
+
cglobal iadst_4x4_internal, 0, 0, 6, dst, stride, coeff, eob, tx2
mova m0, [coeffq+16*0]
mova m1, [coeffq+16*1]
@@ -391,4 +391,4 @@
IWHT4_1D_PACKED
shufpd m0, m2, 0x01
- ITX4_END 0, 3, 2, 1, 0
+ ITX4_END 0, 3, 2, 1, 0