2 Copyright (C) 2012 Paul Davis
4 This program is free software; you can redistribute it and/or modify
5 it under the terms of the GNU General Public License as published by
6 the Free Software Foundation; either version 2 of the License, or
7 (at your option) any later version.
9 This program is distributed in the hope that it will be useful,
10 but WITHOUT ANY WARRANTY; without even the implied warranty of
11 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
12 GNU General Public License for more details.
14 You should have received a copy of the GNU General Public License
15 along with this program; if not, write to the Free Software
16 Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
20 #include "libpbd-config.h"
22 #define _XOPEN_SOURCE 600
23 #include <cstring> // for memset
29 #include "pbd/error.h"
38 unsigned long cpuflags = 0;
42 #if !( (defined __x86_64__) || (defined __i386__) ) // !ARCH_X86
47 #ifndef _LP64 //USE_X86_64_ASM
56 : "%eax", "%ecx", "%edx"
61 /* asm notes: although we explicitly save&restore rbx, we must tell
62 gcc that ebx,rbx is clobbered so that it doesn't try to use it as an intermediate
63 register when storing rbx. gcc 4.3 didn't make this "mistake", but gcc 4.4
64 does, at least on x86_64.
75 : "%rax", "%rbx", "%rcx", "%rdx"
78 #endif /* USE_X86_64_ASM */
80 if (cpuflags & (1<<25)) {
81 _flags = Flags (_flags | (HasSSE|HasFlushToZero));
84 if (cpuflags & (1<<26)) {
85 _flags = Flags (_flags | HasSSE2);
88 if (cpuflags & (1 << 24)) {
92 /* DAZ wasn't available in the first version of SSE. Since
93 setting a reserved bit in MXCSR causes a general protection
94 fault, we need to be able to check the availability of this
95 feature without causing problems. To do this, one needs to
96 set up a 512-byte area of memory to save the SSE state to,
97 using fxsave, and then one needs to inspect bytes 28 through
98 31 for the MXCSR_MASK value. If bit 6 is set, DAZ is
99 supported, otherwise, it isn't.
102 #ifndef HAVE_POSIX_MEMALIGN
103 fxbuf = (char **) malloc (sizeof (char *));
105 *fxbuf = (char *) malloc (512);
108 posix_memalign ((void **) &fxbuf, 16, sizeof (char *));
110 posix_memalign ((void **) fxbuf, 16, 512);
114 memset (*fxbuf, 0, 512);
123 uint32_t mxcsr_mask = *((uint32_t*) &((*fxbuf)[28]));
125 /* if the mask is zero, set its default value (from intel specs) */
127 if (mxcsr_mask == 0) {
131 if (mxcsr_mask & (1<<6)) {
132 _flags = Flags (_flags | HasDenormalsAreZero);
145 #else // COMPILER_MSVC
146 const char* pbd_fpu = "pbd/msvc/fpu.cc takes precedence over this file";
147 #endif // COMPILER_MSVC