Subversion Repositories shark

Rev

Rev 134 | Blame | Compare with Previous | Last modification | View Log | RSS feed

/* $Id: asm_common_x86.s,v 1.1 2003-04-24 13:36:02 giacomo Exp $ */

/*
 * Mesa 3-D graphics library
 * Version:  4.0.3
 *
 * Copyright (C) 1999-2002  Brian Paul   All Rights Reserved.
 *
 * Permission is hereby granted, free of charge, to any person obtaining a
 * copy of this software and associated documentation files (the "Software"),
 * to deal in the Software without restriction, including without limitation
 * the rights to use, copy, modify, merge, publish, distribute, sublicense,
 * and/or sell copies of the Software, and to permit persons to whom the
 * Software is furnished to do so, subject to the following conditions:
 *
 * The above copyright notice and this permission notice shall be included
 * in all copies or substantial portions of the Software.
 *
 * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
 * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
 * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
 * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN
 * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
 * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
 */

/*
 * Check extended CPU capabilities.  Now justs returns the raw CPUID
 * feature information, allowing the higher level code to interpret the
 * results.
 *
 * Written by Holger Waechtler <holger@akaflieg.extern.tu-berlin.de>
 *
 * Cleaned up and simplified by Gareth Hughes <gareth@valinux.com>
 */

/*
 * NOTE: Avoid using spaces in between '(' ')' and arguments, especially
 * with macros like CONST, LLBL that expand to CONCAT(...).  Putting spaces
 * in there will break the build on some platforms.
 */

#include "matypes.h"
#include "features_common_x86.h"


/* Intel vendor string
 */
#define GENU    0x756e6547      /* "Genu" */
#define INEI    0x49656e69      /* "ineI" */
#define NTEL    0x6c65746e      /* "ntel" */

/* AMD vendor string
 */
#define AUTH    0x68747541      /* "Auth" */
#define ENTI    0x69746e65      /* "enti" */
#define CAMD    0x444d4163      /* "cAMD" */


        SEG_DATA

/* We might want to print out some useful messages.
 */
GLNAME( found_intel ):  STRING( "Genuine Intel processor found\n\0" )
GLNAME( found_amd ):    STRING( "Authentic AMD processor found\n\0" )


        SEG_TEXT

ALIGNTEXT4
GLOBL GLNAME( _mesa_identify_x86_cpu_features )
GLNAME( _mesa_identify_x86_cpu_features ):

        PUSH_L  ( EBX )
        PUSH_L  ( ESI )

        /* Test for the CPUID command.  If the ID Flag bit in EFLAGS
         * (bit 21) is writable, the CPUID command is present.
         */
        PUSHF_L
        POP_L   ( EAX )
        MOV_L   ( EAX, ECX )
        XOR_L   ( CONST(0x00200000), EAX )
        PUSH_L  ( EAX )
        POPF_L
        PUSHF_L
        POP_L   ( EAX )

        /* Verify the ID Flag bit has been written.
         */
        CMP_L   ( ECX, EAX )
        JZ      ( LLBL (cpuid_done) )

        /* Get the CPU vendor info.
         */
        XOR_L   ( EAX, EAX )
        CPUID

        /* Test for Intel processors.  We must look for the
         * "GenuineIntel" string in EBX, ECX and EDX.
         */
        CMP_L   ( CONST(GENU), EBX )
        JNE     ( LLBL(cpuid_amd) )
        CMP_L   ( CONST(INEI), EDX )
        JNE     ( LLBL(cpuid_amd) )
        CMP_L   ( CONST(NTEL), ECX )
        JNE     ( LLBL(cpuid_amd) )

        /* We have an Intel processor, so we can get the feature
         * information with an CPUID input value of 1.
         */
        MOV_L   ( CONST(0x1), EAX )
        CPUID
        MOV_L   ( EDX, EAX )

        /* Mask out highest bit, which is used by AMD for 3dnow
         * Newer Intel have this bit set, but do not support 3dnow 
         */
        AND_L   ( CONST(0X7FFFFFFF), EAX)
        JMP     ( LLBL(cpuid_done) )

LLBL(cpuid_amd):

        /* Test for AMD processors.  We must look for the
         * "AuthenticAMD" string in EBX, ECX and EDX.
         */
        CMP_L   ( CONST(AUTH), EBX )
        JNE     ( LLBL(cpuid_other) )
        CMP_L   ( CONST(ENTI), EDX )
        JNE     ( LLBL(cpuid_other) )
        CMP_L   ( CONST(CAMD), ECX )
        JNE     ( LLBL(cpuid_other) )

        /* We have an AMD processor, so we can get the feature
         * information after we verify that the extended functions are
         * supported.
         */
        /* The features we need are almost all in the extended set.  The
         * exception is SSE enable, which is in the standard set (0x1).
         */
        MOV_L   ( CONST(0x1), EAX )
        CPUID
        TEST_L  ( EAX, EAX )
        JZ      ( LLBL (cpuid_failed) )
        MOV_L   ( EDX, ESI )

        MOV_L   ( CONST(0x80000000), EAX )
        CPUID
        TEST_L  ( EAX, EAX )
        JZ      ( LLBL (cpuid_failed) )

        MOV_L   ( CONST(0x80000001), EAX )
        CPUID
        MOV_L   ( EDX, EAX )
        
        AND_L   ( CONST(0x02000000), ESI )      /* OR in the SSE bit */
        OR_L    ( ESI, EAX )
        
        JMP     ( LLBL (cpuid_done) )

LLBL(cpuid_other):

        /* Test for other processors here when required.
         */

LLBL(cpuid_failed):

        /* If we can't determine the feature information, we must
         * return zero to indicate that no platform-specific
         * optimizations can be used.
         */
        MOV_L   ( CONST(0), EAX )

LLBL (cpuid_done):

        POP_L   ( ESI )
        POP_L   ( EBX )
        RET


#ifdef USE_SSE_ASM
/* Execute an SSE instruction to see if the operating system correctly
 * supports SSE.  A signal handler for SIGILL should have been set
 * before calling this function, otherwise this could kill the client
 * application.
 */
ALIGNTEXT4
GLOBL GLNAME( _mesa_test_os_sse_support )
GLNAME( _mesa_test_os_sse_support ):

        XORPS   ( XMM0, XMM0 )

        RET


/* Perform an SSE divide-by-zero to see if the operating system
 * correctly supports unmasked SIMD FPU exceptions.  Signal handlers for
 * SIGILL and SIGFPE should have been set before calling this function,
 * otherwise this could kill the client application.
 */
ALIGNTEXT4
GLOBL GLNAME( _mesa_test_os_sse_exception_support )
GLNAME( _mesa_test_os_sse_exception_support ):

        PUSH_L  ( EBP )
        MOV_L   ( ESP, EBP )
        SUB_L   ( CONST( 8 ), ESP )

        /* Save the original MXCSR register value.
         */
        STMXCSR ( REGOFF( -4, EBP ) )

        /* Unmask the divide-by-zero exception and perform one.
         */
        STMXCSR ( REGOFF( -8, EBP ) )
        AND_L   ( CONST( 0xfffffdff ), REGOFF( -8, EBP ) )
        LDMXCSR ( REGOFF( -8, EBP ) )

        XORPS   ( XMM0, XMM0 )

        PUSH_L  ( CONST( 0x3f800000 ) )
        PUSH_L  ( CONST( 0x3f800000 ) )
        PUSH_L  ( CONST( 0x3f800000 ) )
        PUSH_L  ( CONST( 0x3f800000 ) )

        MOVUPS  ( REGIND( ESP ), XMM1 )

        ADD_L   ( CONST( 32 ), ESP )

        DIVPS   ( XMM0, XMM1 )

        /* Restore the original MXCSR register value.
         */
        LDMXCSR ( REGOFF( -4, EBP ) )

        LEAVE
        RET

#endif