xref: /third_party/ffmpeg/libavutil/x86/cpu.c (revision cabdff1a)
1/*
2 * CPU detection code, extracted from mmx.h
3 * (c)1997-99 by H. Dietz and R. Fisher
4 * Converted to C and improved by Fabrice Bellard.
5 *
6 * This file is part of FFmpeg.
7 *
8 * FFmpeg is free software; you can redistribute it and/or
9 * modify it under the terms of the GNU Lesser General Public
10 * License as published by the Free Software Foundation; either
11 * version 2.1 of the License, or (at your option) any later version.
12 *
13 * FFmpeg is distributed in the hope that it will be useful,
14 * but WITHOUT ANY WARRANTY; without even the implied warranty of
15 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
16 * Lesser General Public License for more details.
17 *
18 * You should have received a copy of the GNU Lesser General Public
19 * License along with FFmpeg; if not, write to the Free Software
20 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
21 */
22
23#include <stdlib.h>
24#include <string.h>
25
26#include "libavutil/x86/asm.h"
27#include "libavutil/x86/cpu.h"
28#include "libavutil/cpu.h"
29#include "libavutil/cpu_internal.h"
30
31#if HAVE_X86ASM
32
33#define cpuid(index, eax, ebx, ecx, edx)        \
34    ff_cpu_cpuid(index, &eax, &ebx, &ecx, &edx)
35
36#define xgetbv(index, eax, edx)                 \
37    ff_cpu_xgetbv(index, &eax, &edx)
38
39#elif HAVE_INLINE_ASM
40
41/* ebx saving is necessary for PIC. gcc seems unable to see it alone */
42#define cpuid(index, eax, ebx, ecx, edx)                        \
43    __asm__ volatile (                                          \
44        "mov    %%"FF_REG_b", %%"FF_REG_S" \n\t"                \
45        "cpuid                       \n\t"                      \
46        "xchg   %%"FF_REG_b", %%"FF_REG_S                       \
47        : "=a" (eax), "=S" (ebx), "=c" (ecx), "=d" (edx)        \
48        : "0" (index), "2"(0))
49
50#define xgetbv(index, eax, edx)                                 \
51    __asm__ (".byte 0x0f, 0x01, 0xd0" : "=a"(eax), "=d"(edx) : "c" (index))
52
53#define get_eflags(x)                           \
54    __asm__ volatile ("pushfl     \n"           \
55                      "pop    %0  \n"           \
56                      : "=r"(x))
57
58#define set_eflags(x)                           \
59    __asm__ volatile ("push    %0 \n"           \
60                      "popfl      \n"           \
61                      :: "r"(x))
62
63#endif /* HAVE_INLINE_ASM */
64
65#if ARCH_X86_64
66
67#define cpuid_test() 1
68
69#elif HAVE_X86ASM
70
71#define cpuid_test ff_cpu_cpuid_test
72
73#elif HAVE_INLINE_ASM
74
75static int cpuid_test(void)
76{
77    x86_reg a, c;
78
79    /* Check if CPUID is supported by attempting to toggle the ID bit in
80     * the EFLAGS register. */
81    get_eflags(a);
82    set_eflags(a ^ 0x200000);
83    get_eflags(c);
84
85    return a != c;
86}
87#endif
88
89/* Function to test if multimedia instructions are supported...  */
90int ff_get_cpu_flags_x86(void)
91{
92    int rval = 0;
93
94#ifdef cpuid
95
96    int eax, ebx, ecx, edx;
97    int max_std_level, max_ext_level, std_caps = 0, ext_caps = 0;
98    int family = 0, model = 0;
99    union { int i[3]; char c[12]; } vendor;
100    int xcr0_lo = 0, xcr0_hi = 0;
101
102    if (!cpuid_test())
103        return 0; /* CPUID not supported */
104
105    cpuid(0, max_std_level, vendor.i[0], vendor.i[2], vendor.i[1]);
106
107    if (max_std_level >= 1) {
108        cpuid(1, eax, ebx, ecx, std_caps);
109        family = ((eax >> 8) & 0xf) + ((eax >> 20) & 0xff);
110        model  = ((eax >> 4) & 0xf) + ((eax >> 12) & 0xf0);
111        if (std_caps & (1 << 15))
112            rval |= AV_CPU_FLAG_CMOV;
113        if (std_caps & (1 << 23))
114            rval |= AV_CPU_FLAG_MMX;
115        if (std_caps & (1 << 25))
116            rval |= AV_CPU_FLAG_MMXEXT;
117#if HAVE_SSE
118        if (std_caps & (1 << 25))
119            rval |= AV_CPU_FLAG_SSE;
120        if (std_caps & (1 << 26))
121            rval |= AV_CPU_FLAG_SSE2;
122        if (ecx & 1)
123            rval |= AV_CPU_FLAG_SSE3;
124        if (ecx & 0x00000200 )
125            rval |= AV_CPU_FLAG_SSSE3;
126        if (ecx & 0x00080000 )
127            rval |= AV_CPU_FLAG_SSE4;
128        if (ecx & 0x00100000 )
129            rval |= AV_CPU_FLAG_SSE42;
130        if (ecx & 0x02000000 )
131            rval |= AV_CPU_FLAG_AESNI;
132#if HAVE_AVX
133        /* Check OXSAVE and AVX bits */
134        if ((ecx & 0x18000000) == 0x18000000) {
135            /* Check for OS support */
136            xgetbv(0, xcr0_lo, xcr0_hi);
137            if ((xcr0_lo & 0x6) == 0x6) {
138                rval |= AV_CPU_FLAG_AVX;
139                if (ecx & 0x00001000)
140                    rval |= AV_CPU_FLAG_FMA3;
141            }
142        }
143#endif /* HAVE_AVX */
144#endif /* HAVE_SSE */
145    }
146    if (max_std_level >= 7) {
147        cpuid(7, eax, ebx, ecx, edx);
148#if HAVE_AVX2
149        if ((rval & AV_CPU_FLAG_AVX) && (ebx & 0x00000020))
150            rval |= AV_CPU_FLAG_AVX2;
151#if HAVE_AVX512 /* F, CD, BW, DQ, VL */
152        if ((xcr0_lo & 0xe0) == 0xe0) { /* OPMASK/ZMM state */
153            if ((rval & AV_CPU_FLAG_AVX2) && (ebx & 0xd0030000) == 0xd0030000) {
154                rval |= AV_CPU_FLAG_AVX512;
155#if HAVE_AVX512ICL
156                if ((ebx & 0xd0200000) == 0xd0200000 && (ecx & 0x5f42) == 0x5f42)
157                    rval |= AV_CPU_FLAG_AVX512ICL;
158#endif /* HAVE_AVX512ICL */
159            }
160        }
161#endif /* HAVE_AVX512 */
162#endif /* HAVE_AVX2 */
163        /* BMI1/2 don't need OS support */
164        if (ebx & 0x00000008) {
165            rval |= AV_CPU_FLAG_BMI1;
166            if (ebx & 0x00000100)
167                rval |= AV_CPU_FLAG_BMI2;
168        }
169    }
170
171    cpuid(0x80000000, max_ext_level, ebx, ecx, edx);
172
173    if (max_ext_level >= 0x80000001) {
174        cpuid(0x80000001, eax, ebx, ecx, ext_caps);
175        if (ext_caps & (1U << 31))
176            rval |= AV_CPU_FLAG_3DNOW;
177        if (ext_caps & (1 << 30))
178            rval |= AV_CPU_FLAG_3DNOWEXT;
179        if (ext_caps & (1 << 23))
180            rval |= AV_CPU_FLAG_MMX;
181        if (ext_caps & (1 << 22))
182            rval |= AV_CPU_FLAG_MMXEXT;
183
184        if (!strncmp(vendor.c, "AuthenticAMD", 12)) {
185        /* Allow for selectively disabling SSE2 functions on AMD processors
186           with SSE2 support but not SSE4a. This includes Athlon64, some
187           Opteron, and some Sempron processors. MMX, SSE, or 3DNow! are faster
188           than SSE2 often enough to utilize this special-case flag.
189           AV_CPU_FLAG_SSE2 and AV_CPU_FLAG_SSE2SLOW are both set in this case
190           so that SSE2 is used unless explicitly disabled by checking
191           AV_CPU_FLAG_SSE2SLOW. */
192            if (rval & AV_CPU_FLAG_SSE2 && !(ecx & 0x00000040))
193                rval |= AV_CPU_FLAG_SSE2SLOW;
194
195        /* Similar to the above but for AVX functions on AMD processors.
196           This is necessary only for functions using YMM registers on Bulldozer
197           and Jaguar based CPUs as they lack 256-bit execution units. SSE/AVX
198           functions using XMM registers are always faster on them.
199           AV_CPU_FLAG_AVX and AV_CPU_FLAG_AVXSLOW are both set so that AVX is
200           used unless explicitly disabled by checking AV_CPU_FLAG_AVXSLOW. */
201            if ((family == 0x15 || family == 0x16) && (rval & AV_CPU_FLAG_AVX))
202                rval |= AV_CPU_FLAG_AVXSLOW;
203
204        /* Zen 3 and earlier have slow gather */
205            if ((family <= 0x19) && (rval & AV_CPU_FLAG_AVX2))
206                rval |= AV_CPU_FLAG_SLOW_GATHER;
207        }
208
209        /* XOP and FMA4 use the AVX instruction coding scheme, so they can't be
210         * used unless the OS has AVX support. */
211        if (rval & AV_CPU_FLAG_AVX) {
212            if (ecx & 0x00000800)
213                rval |= AV_CPU_FLAG_XOP;
214            if (ecx & 0x00010000)
215                rval |= AV_CPU_FLAG_FMA4;
216        }
217    }
218
219    if (!strncmp(vendor.c, "GenuineIntel", 12)) {
220        if (family == 6 && (model == 9 || model == 13 || model == 14)) {
221            /* 6/9 (pentium-m "banias"), 6/13 (pentium-m "dothan"), and
222             * 6/14 (core1 "yonah") theoretically support sse2, but it's
223             * usually slower than mmx, so let's just pretend they don't.
224             * AV_CPU_FLAG_SSE2 is disabled and AV_CPU_FLAG_SSE2SLOW is
225             * enabled so that SSE2 is not used unless explicitly enabled
226             * by checking AV_CPU_FLAG_SSE2SLOW. The same situation
227             * applies for AV_CPU_FLAG_SSE3 and AV_CPU_FLAG_SSE3SLOW. */
228            if (rval & AV_CPU_FLAG_SSE2)
229                rval ^= AV_CPU_FLAG_SSE2SLOW | AV_CPU_FLAG_SSE2;
230            if (rval & AV_CPU_FLAG_SSE3)
231                rval ^= AV_CPU_FLAG_SSE3SLOW | AV_CPU_FLAG_SSE3;
232        }
233        /* The Atom processor has SSSE3 support, which is useful in many cases,
234         * but sometimes the SSSE3 version is slower than the SSE2 equivalent
235         * on the Atom, but is generally faster on other processors supporting
236         * SSSE3. This flag allows for selectively disabling certain SSSE3
237         * functions on the Atom. */
238        if (family == 6 && model == 28)
239            rval |= AV_CPU_FLAG_ATOM;
240
241        /* Conroe has a slow shuffle unit. Check the model number to ensure not
242         * to include crippled low-end Penryns and Nehalems that lack SSE4. */
243        if ((rval & AV_CPU_FLAG_SSSE3) && !(rval & AV_CPU_FLAG_SSE4) &&
244            family == 6 && model < 23)
245            rval |= AV_CPU_FLAG_SSSE3SLOW;
246
247        /* Haswell has slow gather */
248        if ((rval & AV_CPU_FLAG_AVX2) && family == 6 && model < 70)
249            rval |= AV_CPU_FLAG_SLOW_GATHER;
250    }
251
252#endif /* cpuid */
253
254    return rval;
255}
256
257size_t ff_get_cpu_max_align_x86(void)
258{
259    int flags = av_get_cpu_flags();
260
261    if (flags & AV_CPU_FLAG_AVX512)
262        return 64;
263    if (flags & (AV_CPU_FLAG_AVX2      |
264                 AV_CPU_FLAG_AVX       |
265                 AV_CPU_FLAG_XOP       |
266                 AV_CPU_FLAG_FMA4      |
267                 AV_CPU_FLAG_FMA3      |
268                 AV_CPU_FLAG_AVXSLOW))
269        return 32;
270    if (flags & (AV_CPU_FLAG_AESNI     |
271                 AV_CPU_FLAG_SSE42     |
272                 AV_CPU_FLAG_SSE4      |
273                 AV_CPU_FLAG_SSSE3     |
274                 AV_CPU_FLAG_SSE3      |
275                 AV_CPU_FLAG_SSE2      |
276                 AV_CPU_FLAG_SSE       |
277                 AV_CPU_FLAG_ATOM      |
278                 AV_CPU_FLAG_SSSE3SLOW |
279                 AV_CPU_FLAG_SSE3SLOW  |
280                 AV_CPU_FLAG_SSE2SLOW))
281        return 16;
282
283    return 8;
284}
285