lavu: add AESNI CPU flag

configure

@@ -368,6 +368,7 @@ Optimization options (experts only):
   --disable-fma3           disable FMA3 optimizations
   --disable-fma4           disable FMA4 optimizations
   --disable-avx2           disable AVX2 optimizations
+  --disable-aesni          disable AESNI optimizations
   --disable-armv5te        disable armv5te optimizations
   --disable-armv6          disable armv6 optimizations
   --disable-armv6t2        disable armv6t2 optimizations
@@ -1635,6 +1636,7 @@ ARCH_EXT_LIST_LOONGSON="
 "
 
 ARCH_EXT_LIST_X86_SIMD="
+    aesni
     amd3dnow
     amd3dnowext
     avx
@@ -2128,6 +2130,7 @@ sse3_deps="sse2"
 ssse3_deps="sse3"
 sse4_deps="ssse3"
 sse42_deps="sse4"
+aesni_deps="sse42"
 avx_deps="sse42"
 xop_deps="avx"
 fma3_deps="avx"
@@ -6011,6 +6014,7 @@ if enabled x86; then
     echo "3DNow! extended enabled   ${amd3dnowext-no}"
     echo "SSE enabled               ${sse-no}"
     echo "SSSE3 enabled             ${ssse3-no}"
+    echo "AESNI enabled             ${aesni-no}"
     echo "AVX enabled               ${avx-no}"
     echo "XOP enabled               ${xop-no}"
     echo "FMA3 enabled              ${fma3-no}"

doc/APIchanges

@@ -15,6 +15,9 @@ libavutil:     2015-08-28
 
 API changes, most recent first:
 
+2015-10-27 - xxxxxxx - lavu 55.5.100 - cpu.h
+  Add AV_CPU_FLAG_AESNI.
+
 2015-10-27 - xxxxxxx - lavc 57.12.100 / 57.8.0 - avcodec.h
   Deprecate av_free_packet(). Use av_packet_unref() as replacement,
   it resets the packet in a more consistent way.

libavutil/cpu.c

@@ -118,6 +118,7 @@ int av_parse_cpu_flags(const char *s)
 #define CPUFLAG_FMA4     (AV_CPU_FLAG_FMA4     | CPUFLAG_AVX)
 #define CPUFLAG_AVX2     (AV_CPU_FLAG_AVX2     | CPUFLAG_AVX)
 #define CPUFLAG_BMI2     (AV_CPU_FLAG_BMI2     | AV_CPU_FLAG_BMI1)
+#define CPUFLAG_AESNI    (AV_CPU_FLAG_AESNI    | CPUFLAG_SSE42)
     static const AVOption cpuflags_opts[] = {
         { "flags"   , NULL, 0, AV_OPT_TYPE_FLAGS, { .i64 = 0 }, INT64_MIN, INT64_MAX, .unit = "flags" },
 #if   ARCH_PPC
@@ -145,6 +146,7 @@ int av_parse_cpu_flags(const char *s)
         { "3dnow"   , NULL, 0, AV_OPT_TYPE_CONST, { .i64 = CPUFLAG_3DNOW        },    .unit = "flags" },
         { "3dnowext", NULL, 0, AV_OPT_TYPE_CONST, { .i64 = CPUFLAG_3DNOWEXT     },    .unit = "flags" },
         { "cmov",     NULL, 0, AV_OPT_TYPE_CONST, { .i64 = AV_CPU_FLAG_CMOV     },    .unit = "flags" },
+        { "aesni"   , NULL, 0, AV_OPT_TYPE_CONST, { .i64 = CPUFLAG_AESNI        },    .unit = "flags" },
 #elif ARCH_ARM
         { "armv5te",  NULL, 0, AV_OPT_TYPE_CONST, { .i64 = AV_CPU_FLAG_ARMV5TE  },    .unit = "flags" },
         { "armv6",    NULL, 0, AV_OPT_TYPE_CONST, { .i64 = AV_CPU_FLAG_ARMV6    },    .unit = "flags" },
@@ -205,6 +207,7 @@ int av_parse_cpu_caps(unsigned *flags, const char *s)
         { "3dnow"   , NULL, 0, AV_OPT_TYPE_CONST, { .i64 = AV_CPU_FLAG_3DNOW    },    .unit = "flags" },
         { "3dnowext", NULL, 0, AV_OPT_TYPE_CONST, { .i64 = AV_CPU_FLAG_3DNOWEXT },    .unit = "flags" },
         { "cmov",     NULL, 0, AV_OPT_TYPE_CONST, { .i64 = AV_CPU_FLAG_CMOV     },    .unit = "flags" },
+        { "aesni",    NULL, 0, AV_OPT_TYPE_CONST, { .i64 = AV_CPU_FLAG_AESNI    },    .unit = "flags" },
 
 #define CPU_FLAG_P2 AV_CPU_FLAG_CMOV | AV_CPU_FLAG_MMX
 #define CPU_FLAG_P3 CPU_FLAG_P2 | AV_CPU_FLAG_MMX2 | AV_CPU_FLAG_SSE
@@ -340,6 +343,7 @@ static const struct {
     { AV_CPU_FLAG_AVX2,      "avx2"       },
     { AV_CPU_FLAG_BMI1,      "bmi1"       },
     { AV_CPU_FLAG_BMI2,      "bmi2"       },
+    { AV_CPU_FLAG_AESNI,     "aesni"      },
 #endif
     { 0 }
 };

libavutil/cpu.h

@@ -42,6 +42,7 @@
 #define AV_CPU_FLAG_ATOM     0x10000000 ///< Atom processor, some SSSE3 instructions are slower
 #define AV_CPU_FLAG_SSE4         0x0100 ///< Penryn SSE4.1 functions
 #define AV_CPU_FLAG_SSE42        0x0200 ///< Nehalem SSE4.2 functions
+#define AV_CPU_FLAG_AESNI       0x80000 ///< Advanced Encryption Standard functions
 #define AV_CPU_FLAG_AVX          0x4000 ///< AVX functions: requires OS support even if YMM registers aren't used
 #define AV_CPU_FLAG_AVXSLOW   0x8000000 ///< AVX supported, but slow when using YMM registers (e.g. Bulldozer)
 #define AV_CPU_FLAG_XOP          0x0400 ///< Bulldozer XOP functions

libavutil/version.h

@@ -56,7 +56,7 @@
  */
 
 #define LIBAVUTIL_VERSION_MAJOR  55
-#define LIBAVUTIL_VERSION_MINOR   4
+#define LIBAVUTIL_VERSION_MINOR   5
 #define LIBAVUTIL_VERSION_MICRO 100
 
 #define LIBAVUTIL_VERSION_INT   AV_VERSION_INT(LIBAVUTIL_VERSION_MAJOR, \

libavutil/x86/cpu.c

@@ -126,6 +126,8 @@ int ff_get_cpu_flags_x86(void)
             rval |= AV_CPU_FLAG_SSE4;
         if (ecx & 0x00100000 )
             rval |= AV_CPU_FLAG_SSE42;
+        if (ecx & 0x01000000 )
+            rval |= AV_CPU_FLAG_AESNI;
 #if HAVE_AVX
         /* Check OXSAVE and AVX bits */
         if ((ecx & 0x18000000) == 0x18000000) {

libavutil/x86/cpu.h

@@ -47,6 +47,7 @@
 #define X86_FMA3(flags)             CPUEXT(flags, FMA3)
 #define X86_FMA4(flags)             CPUEXT(flags, FMA4)
 #define X86_AVX2(flags)             CPUEXT(flags, AVX2)
+#define X86_AESNI(flags)            CPUEXT(flags, AESNI)
 
 #define EXTERNAL_AMD3DNOW(flags)    CPUEXT_SUFFIX(flags, _EXTERNAL, AMD3DNOW)
 #define EXTERNAL_AMD3DNOWEXT(flags) CPUEXT_SUFFIX(flags, _EXTERNAL, AMD3DNOWEXT)
@@ -69,6 +70,7 @@
 #define EXTERNAL_FMA3(flags)        CPUEXT_SUFFIX(flags, _EXTERNAL, FMA3)
 #define EXTERNAL_FMA4(flags)        CPUEXT_SUFFIX(flags, _EXTERNAL, FMA4)
 #define EXTERNAL_AVX2(flags)        CPUEXT_SUFFIX(flags, _EXTERNAL, AVX2)
+#define EXTERNAL_AESNI(flags)       CPUEXT_SUFFIX(flags, _EXTERNAL, AESNI)
 
 #define INLINE_AMD3DNOW(flags)      CPUEXT_SUFFIX(flags, _INLINE, AMD3DNOW)
 #define INLINE_AMD3DNOWEXT(flags)   CPUEXT_SUFFIX(flags, _INLINE, AMD3DNOWEXT)
@@ -91,6 +93,7 @@
 #define INLINE_FMA3(flags)          CPUEXT_SUFFIX(flags, _INLINE, FMA3)
 #define INLINE_FMA4(flags)          CPUEXT_SUFFIX(flags, _INLINE, FMA4)
 #define INLINE_AVX2(flags)          CPUEXT_SUFFIX(flags, _INLINE, AVX2)
+#define INLINE_AESNI(flags)         CPUEXT_SUFFIX(flags, _INLINE, AESNI)
 
 void ff_cpu_cpuid(int index, int *eax, int *ebx, int *ecx, int *edx);
 void ff_cpu_xgetbv(int op, int *eax, int *edx);

libavutil/x86/x86inc.asm

@@ -772,6 +772,7 @@ BRANCH_INSTR jz, je, jnz, jne, jl, jle, jnl, jnle, jg, jge, jng, jnge, ja, jae,
 %assign cpuflags_atom     (1<<21)
 %assign cpuflags_bmi1     (1<<22)|cpuflags_lzcnt
 %assign cpuflags_bmi2     (1<<23)|cpuflags_bmi1
+%assign cpuflags_aesni    (1<<24)|cpuflags_sse42
 
 ; Returns a boolean value expressing whether or not the specified cpuflag is enabled.
 %define    cpuflag(x) (((((cpuflags & (cpuflags_ %+ x)) ^ (cpuflags_ %+ x)) - 1) >> 31) & 1)
@@ -1171,12 +1172,12 @@ AVX_INSTR addsd, sse2, 1, 0, 1
 AVX_INSTR addss, sse, 1, 0, 1
 AVX_INSTR addsubpd, sse3, 1, 0, 0
 AVX_INSTR addsubps, sse3, 1, 0, 0
-AVX_INSTR aesdec, fnord, 0, 0, 0
-AVX_INSTR aesdeclast, fnord, 0, 0, 0
-AVX_INSTR aesenc, fnord, 0, 0, 0
-AVX_INSTR aesenclast, fnord, 0, 0, 0
-AVX_INSTR aesimc
-AVX_INSTR aeskeygenassist
+AVX_INSTR aesdec, aesni, 0, 0, 0
+AVX_INSTR aesdeclast, aesni, 0, 0, 0
+AVX_INSTR aesenc, aesni, 0, 0, 0
+AVX_INSTR aesenclast, aesni, 0, 0, 0
+AVX_INSTR aesimc, aesni
+AVX_INSTR aeskeygenassist, aesni
 AVX_INSTR andnpd, sse2, 1, 0, 0
 AVX_INSTR andnps, sse, 1, 0, 0
 AVX_INSTR andpd, sse2, 1, 0, 1

tests/checkasm/checkasm.c

@@ -117,6 +117,7 @@ static const struct {
     { "SSSE3",    "ssse3",    AV_CPU_FLAG_SSSE3|AV_CPU_FLAG_ATOM },
     { "SSE4.1",   "sse4",     AV_CPU_FLAG_SSE4 },
     { "SSE4.2",   "sse42",    AV_CPU_FLAG_SSE42 },
+    { "AES-NI",   "aesni",    AV_CPU_FLAG_AESNI },
     { "AVX",      "avx",      AV_CPU_FLAG_AVX },
     { "XOP",      "xop",      AV_CPU_FLAG_XOP },
     { "FMA3",     "fma3",     AV_CPU_FLAG_FMA3 },