Skip to content

L
linux
Commit 97ccc379 authored by Linus Torvalds's avatar Linus Torvalds
Browse files
Options

Import 2.1.73

parent 62e3b9df
Expand all Hide whitespace changes
Inline Side-by-side
Showing with 23211 additions and 32172 deletions
Documentation/Configure.help
View file @ 97ccc379
......@@ -248,11 +248,13 @@ CONFIG_BLK_DEV_IDETAPE
Include IDE/ATAPI FLOPPY support
CONFIG_BLK_DEV_IDEFLOPPY
If you have an IDE floppy drive which uses the ATAPI protocol, say
Y. Chances are that you don't, because these animals are rare.
If you have an IDE floppy drive which uses the ATAPI protocol, say Y.
ATAPI is a new protocol used by IDE CDROM/tape/floppy drives,
similar to the SCSI protocol. At boot time, the FLOPPY drive will
be identified along with other IDE devices, as "hdb" or "hdc", or
similar to the SCSI protocol. IDE floppy drives include the
LS-120 and the ATAPI ZIP (ATAPI PD-CD/CDR drives are not supported
by this driver; support for PD-CD/CDR drives is available through
the SCSI emulation). At boot time, the FLOPPY drive will be
identified along with other IDE devices, as "hdb" or "hdc", or
something similar. If you want to compile the driver as a module ( =
code which can be inserted in and removed from the running kernel
whenever you want), say M here and read Documentation/modules.txt.
......@@ -263,10 +265,11 @@ CONFIG_BLK_DEV_IDESCSI
This will provide SCSI host adapter emulation for IDE ATAPI devices,
and will allow you to use a SCSI device driver instead of a native
ATAPI driver. This is useful if you have an ATAPI device for which
no native driver has been written; you can then use this emulation
together with an appropriate SCSI device driver. If both this SCSI
emulation and native ATAPI support are compiled into the kernel, the
native support will be used. Normally, say N.
no native driver has been written (for example, an ATAPI PD-CD or
CDR drive); you can then use this emulation together with an
appropriate SCSI device driver. If both this SCSI emulation and
native ATAPI support are compiled into the kernel, the native
support will be used. Normally, say N.
CMD640 chipset bugfix/support
CONFIG_BLK_DEV_CMD640
......@@ -303,17 +306,24 @@ CONFIG_BLK_DEV_RZ1000
Linux. This may slow disk throughput by a few percent, but at least
things will operate 100% reliably. If unsure, say Y.
Intel 82371 PIIX (Triton I/II), VIA VP-1 DMA support
Generic PCI IDE chipset support
CONFIG_BLK_DEV_IDEPCI
Enable this for PCI systems which use IDE drive(s).
This option helps the IDE driver to automatically detect and
configure all PCI-based IDE interfaces in your system.
It is safe to say Y to this question.
Generic PCI bus-master DMA support
CONFIG_BLK_DEV_IDEDMA
If your PCI system uses IDE drive(s) (as opposed to SCSI, say)
and is capable of bus-master DMA operation (most Pentium PCI
systems), you will want to enable this option to allow use of
bus-mastering DMA data transfers. Read the comments at the
beginning of drivers/block/idedma.c and Documentation/ide.txt.
You can get the latest version of the hdparm utility via
ftp (user: anonymous) from
sunsite.unc.edu/pub/Linux/kernel/patches/diskdrives/; it is
used to tune your harddisk.
If your PCI IDE controller is capable of bus-master DMA
(Direct Memory Access) transfers (most newer systems),
then you will want to enable this option to reduce CPU overhead.
With this option, Linux will automatically enable DMA transfers
in most cases, noting this with "DMA" appended to the drive
identification info. You can also use the "hdparm" utility to
enable DMA for drives which were not enabled automatically.
You can get the latest version of the hdparm utility via anonymous
FTP from sunsite.unc.edu/pub/Linux/system/hardware/
It is safe to say Y to this question.
Other IDE chipset support
......@@ -371,6 +381,12 @@ CONFIG_BLK_DEV_OPTI621
for drives attached to an OPTi MIDE controller.
Please read the comments at the top of drivers/block/opti621.c.
NS87415 support (EXPERIMENTAL)
CONFIG_BLK_DEV_NS87415
This driver adds detection and support for the NS87415 chip
(used in SPARC64, among others).
Please read the comments at the top of drivers/block/ns87415.c.
QDI QD6580 support
CONFIG_BLK_DEV_QD6580
This driver is enabled at runtime using the "ide0=qd6580" kernel
......@@ -875,31 +891,8 @@ CONFIG_BINFMT_AOUT
Kernel support for JAVA binaries
CONFIG_BINFMT_JAVA
JAVA(tm) is an object oriented programming language developed by
SUN; JAVA programs are compiled into "JAVA bytecode" binaries which
can then be interpreted by run time systems on many different
operating systems. These JAVA binaries are becoming a universal
executable format. If you want to execute JAVA binaries, read the
Java on Linux HOWTO, available via ftp (user: anonymous) at
sunsite.unc.edu:/pub/Linux/docs/HOWTO. You will then need to install
the run time system contained in the Java Developers Kit (JDK) as
described in the HOWTO. This is completely independent of the Linux
kernel and you do NOT need to say Y here for this to work.
Saying Y here allows you to execute a JAVA bytecode binary just like
any other Linux program: by simply typing in its name. (You also
need to have the JDK installed for this to work). As more and more
Java programs become available, the use for this will gradually
increase. You can even execute HTML files containing JAVA applets (=
JAVA binaries) if those files start with the string
"<!--applet-->". If you want to use this, say Y here and read
Documentation/java.txt. If you disable this option it will reduce
your kernel by about 4kB. This is not much and by itself does not
warrant removing support. However its removal is a good idea if you
do not have the JDK installed. You may answer M for module support
and later load the module when you install the JDK or find an
interesting Java program that you can't live without. The module
will be called binfmt_java.o. If you don't know what to answer at
this point then answer Y.
This option is obsolete. Use binfmt_misc instead. It is more
flexible.
Kernel support for Linux/Intel ELF binaries
CONFIG_BINFMT_EM86
......@@ -1332,6 +1325,37 @@ CONFIG_IP_MASQUERADE
inserted in and removed from the running kernel whenever you want;
read Documentation/modules.txt for details.
IP: ICMP masquerading
CONFIG_IP_MASQUERADE_ICMP
The basic masquerade code described for CONFIG_IP_MASQUERADE only
handles TCP or UDP packets (and ICMP errors for existing
connections). This option adds additional support for masquerading
ICMP packets, such as ping or the probes used by the Windows 95
tracert program.
If you want this, say Y.
IP: ipautofw masquerade support
CONFIG_IP_MASQUERADE_IPAUTOFW (Experimental)
ipautofw is a program by Richard Lynch allowing additional
support for masquerading protocols which do not (as yet)
have additional protocol helpers.
Information and source for ipautofw is available from
ftp://ftp.netis.com/pub/members/rlynch/
The ipautofw code is still under development and so is currently
marked EXPERIMENTAL.
If you want this, say Y.
IP: ipportfw masquerade support
CONFIG_IP_MASQUERADE_IPPORTFW
ipportfw is an addition to IP Masquerading written by Steven Clarke
to allow some forwarding of packets from outside to inside a
firewall on given ports. Information and source for ipportfw is
available from
http://www.monmouth.demon.co.uk/ipsubs/portforwarding.html
The portfw code is still under development and so is currently
marked EXPERIMENTAL.
If you want this, say Y.
IP: always defragment
CONFIG_IP_ALWAYS_DEFRAG
This option means that all incoming fragments (= parts of IP packets
......
Documentation/filesystems/sysv-fs.txt
View file @ 97ccc379
......@@ -28,9 +28,9 @@ Bugs in the present implementation:
Please report any bugs and suggestions to
Bruno Haible <haible@ma2s2.mathematik.uni-karlsruhe.de> or
Pascal Haible <haible@izfm.uni-stuttgart.de> .
Bruno Haible <haible@ma2s2.mathematik.uni-karlsruhe.de>
Pascal Haible <haible@izfm.uni-stuttgart.de>
Krzysztof G. Baranowski <kgb@manjak.knm.org.pl>
Bruno Haible
<haible@ma2s2.mathematik.uni-karlsruhe.de>
......
MAINTAINERS
View file @ 97ccc379
......@@ -279,6 +279,12 @@ M: andersee@debian.org
L: linux-kernel@vger.rutgers.edu
S: Maintained
IDE/ATAPI TAPE/FLOPPY DRIVERS
P: Gadi Oxman
M: Gadi Oxman <gadio@netvision.net.il>
L: linux-kernel@vger.rutgers.edu
S: Maintained
ISDN SUBSYSTEM
P: Fritz Elfert
M: fritz@wuemaus.franken.de
......@@ -551,6 +557,11 @@ P: Pavel Machek
M: pavel@atrey.karlin.mff.cuni.cz
S: Maintained
SYSV FILESYSTEM
P: Krzysztof G. Baranowski
M: kgb@manjak.knm.org.pl
S: Maintained
REST:
P: Linus Torvalds
S: Buried alive in diapers
Makefile
View file @ 97ccc379
VERSION = 2
PATCHLEVEL = 1
SUBLEVEL = 72
SUBLEVEL = 73
ARCH := $(shell uname -m | sed -e s/i.86/i386/ -e s/sun4u/sparc64/)
......
arch/i386/defconfig
View file @ 97ccc379
......@@ -58,6 +58,7 @@ CONFIG_BLK_DEV_IDECD=y
CONFIG_BLK_DEV_CMD640=y
# CONFIG_BLK_DEV_CMD640_ENHANCED is not set
CONFIG_BLK_DEV_RZ1000=y
CONFIG_BLK_DEV_IDEPCI=y
CONFIG_BLK_DEV_IDEDMA=y
# CONFIG_IDE_CHIPSETS is not set
......@@ -221,7 +222,6 @@ CONFIG_ISO9660_FS=y
# CONFIG_VFAT_FS is not set
CONFIG_PROC_FS=y
CONFIG_NFS_FS=y
# CONFIG_ROOT_NFS is not set
CONFIG_NFSD=y
CONFIG_SUNRPC=y
CONFIG_LOCKD=y
......
arch/i386/kernel/process.c
View file @ 97ccc379
......@@ -40,6 +40,10 @@
#include <asm/system.h>
#include <asm/io.h>
#include <asm/ldt.h>
#include <asm/processor.h>
#ifdef CONFIG_MATH_EMULATION
#include <asm/math_emu.h>
#endif
#ifdef __SMP__
asmlinkage void ret_from_smpfork(void) __asm__("ret_from_smpfork");
......@@ -525,19 +529,16 @@ int dump_fpu (struct pt_regs * regs, struct user_i387_struct* fpu)
{
int fpvalid;
/* Flag indicating the math stuff is valid. We don't support this for the
soft-float routines yet */
if (hard_math) {
if ((fpvalid = current->used_math) != 0) {
if (last_task_used_math == current)
__asm__("clts ; fnsave %0": :"m" (*fpu));
if ((fpvalid = current->used_math) != 0) {
if (hard_math) {
if (last_task_used_math == current) {
__asm__("clts ; fsave %0; fwait": :"m" (*fpu));
}
else
memcpy(fpu,&current->tss.i387.hard,sizeof(*fpu));
} else {
memcpy(fpu,&current->tss.i387.hard,sizeof(*fpu));
}
} else {
/* we should dump the emulator state here, but we need to
convert it into standard 387 format first.. */
fpvalid = 0;
}
return fpvalid;
......
arch/i386/kernel/ptrace.c
View file @ 97ccc379
......@@ -16,6 +16,7 @@
#include <asm/uaccess.h>
#include <asm/pgtable.h>
#include <asm/system.h>
#include <asm/processor.h>
/*
* does not yet catch signals sent when the child dies.
......@@ -561,6 +562,102 @@ asmlinkage int sys_ptrace(long request, long pid, long addr, long data)
goto out;
}
case PTRACE_GETREGS: { /* Get all gp regs from the child. */
if (!access_ok(VERIFY_WRITE, (unsigned *)data,
17*sizeof(long)))
{
ret = -EIO;
goto out;
}
for ( i = 0; i < 17*sizeof(long); i += sizeof(long) )
{
__put_user(getreg(child, i),(unsigned long *) data);
data += sizeof(long);
}
ret = 0;
goto out;
};
case PTRACE_SETREGS: { /* Set all gp regs in the child. */
unsigned long tmp;
if (!access_ok(VERIFY_READ, (unsigned *)data,
17*sizeof(long)))
{
ret = -EIO;
goto out;
}
for ( i = 0; i < 17*sizeof(long); i += sizeof(long) )
{
__get_user(tmp, (unsigned long *) data);
putreg(child, i, tmp);
data += sizeof(long);
}
ret = 0;
goto out;
};
case PTRACE_GETFPREGS: { /* Get the child FPU state. */
if (!access_ok(VERIFY_WRITE, (unsigned *)data,
sizeof(struct user_i387_struct)))
{
ret = -EIO;
goto out;
}
ret = 0;
if ( !child->used_math ) {
/* Simulate an empty FPU. */
child->tss.i387.hard.cwd = 0xffff037f;
child->tss.i387.hard.swd = 0xffff0000;
child->tss.i387.hard.twd = 0xffffffff;
}
#ifdef CONFIG_MATH_EMULATION
if ( hard_math ) {
#endif
if (last_task_used_math == child) {
clts();
__asm__("fnsave %0; fwait":"=m" (child->tss.i387.hard));
last_task_used_math = NULL;
stts();
}
__copy_to_user((void *)data, &child->tss.i387.hard,
sizeof(struct user_i387_struct));
#ifdef CONFIG_MATH_EMULATION
} else {
save_i387_soft(&child->tss.i387.soft,
(struct _fpstate *)data);
}
#endif
goto out;
};
case PTRACE_SETFPREGS: { /* Set the child FPU state. */
if (!access_ok(VERIFY_READ, (unsigned *)data,
sizeof(struct user_i387_struct)))
{
ret = -EIO;
goto out;
}
child->used_math = 1;
#ifdef CONFIG_MATH_EMULATION
if ( hard_math ) {
#endif
if (last_task_used_math == child) {
/* Discard the state of the FPU */
last_task_used_math = NULL;
}
__copy_from_user(&child->tss.i387.hard, (void *)data,
sizeof(struct user_i387_struct));
child->flags &= ~PF_USEDFPU;
#ifdef CONFIG_MATH_EMULATION
} else {
restore_i387_soft(&child->tss.i387.soft,
(struct _fpstate *)data);
}
#endif
ret = 0;
goto out;
};
default:
ret = -EIO;
goto out;
......
arch/i386/kernel/signal.c
View file @ 97ccc379
......@@ -156,7 +156,6 @@ static inline void restore_i387_hard(struct _fpstate *buf)
stts();
}
#endif
current->used_math = 1;
current->flags &= ~PF_USEDFPU;
__copy_from_user(&current->tss.i387.hard, buf, sizeof(*buf));
}
......@@ -169,8 +168,9 @@ static inline void restore_i387(struct _fpstate *buf)
if (hard_math)
restore_i387_hard(buf);
else
restore_i387_soft(buf);
restore_i387_soft(&current->tss.i387.soft, buf);
#endif
current->used_math = 1;
}
static int
......@@ -309,7 +309,6 @@ static inline struct _fpstate * save_i387_hard(struct _fpstate * buf)
#endif
current->tss.i387.hard.status = current->tss.i387.hard.swd;
copy_to_user(buf, &current->tss.i387.hard, sizeof(*buf));
current->used_math = 0;
return buf;
}
......@@ -318,10 +317,16 @@ static struct _fpstate * save_i387(struct _fpstate *buf)
if (!current->used_math)
return NULL;
/* This will cause a "finit" to be triggered by the next
attempted FPU operation by the 'current' process.
*/
current->used_math = 0;
#ifndef CONFIG_MATH_EMULATION
return save_i387_hard(buf);
#else
return hard_math ? save_i387_hard(buf) : save_i387_soft(buf);
return hard_math ? save_i387_hard(buf)
: save_i387_soft(&current->tss.i387.soft, buf);
#endif
}
......
arch/i386/math-emu/Makefile
View file @ 97ccc379
......@@ -12,18 +12,23 @@ CFLAGS := $(CFLAGS) $(PARANOID) $(DEBUG) -fno-builtin $(MATH_EMULATION)
.S.o:
$(CC) -D__ASSEMBLY__ $(PARANOID) -c $<
L_OBJS =fpu_entry.o div_small.o errors.o \
fpu_arith.o fpu_aux.o fpu_etc.o fpu_trig.o \
# From 'C' language sources:
C_OBJS =fpu_entry.o errors.o \
fpu_arith.o fpu_aux.o fpu_etc.o fpu_tags.o fpu_trig.o \
load_store.o get_address.o \
poly_atan.o poly_l2.o poly_2xm1.o poly_sin.o poly_tan.o \
reg_add_sub.o reg_compare.o reg_constant.o reg_ld_str.o \
reg_div.o reg_mul.o reg_norm.o \
reg_u_add.o reg_u_div.o reg_u_mul.o reg_u_sub.o \
reg_round.o \
reg_add_sub.o reg_compare.o reg_constant.o reg_convert.o \
reg_ld_str.o reg_divide.o reg_mul.o
# From 80x86 assembler sources:
A_OBJS =reg_u_add.o reg_u_div.o reg_u_mul.o reg_u_sub.o \
div_small.o reg_norm.o reg_round.o \
wm_shrx.o wm_sqrt.o \
div_Xsig.o polynom_Xsig.o round_Xsig.o \
shr_Xsig.o mul_Xsig.o
L_OBJS =$(C_OBJS) $(A_OBJS)
include $(TOPDIR)/Rules.make
proto:
......
arch/i386/math-emu/README
View file @ 97ccc379
+---------------------------------------------------------------------------+
| wm-FPU-emu an FPU emulator for 80386 and 80486SX microprocessors. |
| |
| Copyright (C) 1992,1993,1994,1995,1996 |
| Copyright (C) 1992,1993,1994,1995,1996,1997 |
| W. Metzenthen, 22 Parker St, Ormond, Vic 3163, |
| Australia. E-mail billm@suburbia.net |
| |
......@@ -44,9 +44,12 @@ some differences.
Please report bugs, etc to me at:
billm@suburbia.net
For more information on the emulator and on floating point topics, see
my web pages, currently at http://www.suburbia.net/~billm/
--Bill Metzenthen
October 1996
December 1997
----------------------- Internals of wm-FPU-emu -----------------------
......@@ -95,8 +98,9 @@ form of re-entrancy which is required by the Linux kernel.
----------------------- Limitations of wm-FPU-emu -----------------------
There are a number of differences between the current wm-FPU-emu
(version 1.20) and the 80486 FPU (apart from bugs). Some of the more
important differences are listed below:
(version 2.00) and the 80486 FPU (apart from bugs). The differences
are fewer than those which applied to the 1.xx series of the emulator.
Some of the more important differences are listed below:
The Roundup flag does not have much meaning for the transcendental
functions and its 80486 value with these functions is likely to differ
......@@ -122,18 +126,6 @@ and Unnormals. None of these will be generated by an 80486 or by the
emulator. Do not use them. The emulator treats them differently in
detail from the way an 80486 does.
The emulator treats PseudoDenormals differently from an 80486. These
numbers are in fact properly normalised numbers with the exponent
offset by 1, and the emulator treats them as such. Unlike the 80486,
the emulator does not generate a Denormal Operand exception for these
numbers. The arithmetical results produced when using such a number as
an operand are the same for the emulator and a real 80486 (apart from
any slight precision difference for the transcendental functions).
Neither the emulator nor an 80486 produces one of these numbers as the
result of any arithmetic operation. An 80486 can keep one of these
numbers in an FPU register with its identity as a PseudoDenormal, but
the emulator will not; they are always converted to a valid number.
Self modifying code can cause the emulator to fail. An example of such
code is:
movl %esp,[%ebx]
......
arch/i386/math-emu/div_small.S
View file @ 97ccc379
......@@ -12,13 +12,13 @@
+---------------------------------------------------------------------------*/
/*---------------------------------------------------------------------------+
| unsigned long div_small(unsigned long long *x, unsigned long y) |
| unsigned long FPU_div_small(unsigned long long *x, unsigned long y) |
+---------------------------------------------------------------------------*/
#include "fpu_emu.h"
.text
ENTRY(div_small)
ENTRY(FPU_div_small)
pushl %ebp
movl %esp,%ebp
......
arch/i386/math-emu/errors.c
View file @ 97ccc379
This diff is collapsed.
arch/i386/math-emu/fpu_arith.c
View file @ 97ccc379
......@@ -3,9 +3,9 @@
| |
| Code to implement the FPU register/register arithmetic instructions |
| |
| Copyright (C) 1992,1993 |
| W. Metzenthen, 22 Parker St, Ormond, Vic 3163, |
| Australia. E-mail billm@vaxc.cc.monash.edu.au |
| Copyright (C) 1992,1993,1997 |
| W. Metzenthen, 22 Parker St, Ormond, Vic 3163, Australia |
| E-mail billm@suburbia.net |
| |
| |
+---------------------------------------------------------------------------*/
......@@ -19,16 +19,18 @@
void fadd__()
{
/* fadd st,st(i) */
int i = FPU_rm;
clear_C1();
reg_add(&st(0), &st(FPU_rm), &st(0), control_word);
FPU_add(&st(i), FPU_gettagi(i), 0, control_word);
}
void fmul__()
{
/* fmul st,st(i) */
int i = FPU_rm;
clear_C1();
reg_mul(&st(0), &st(FPU_rm), &st(0), control_word);
FPU_mul(&st(i), FPU_gettagi(i), 0, control_word);
}
......@@ -37,7 +39,7 @@ void fsub__()
{
/* fsub st,st(i) */
clear_C1();
reg_sub(&st(0), &st(FPU_rm), &st(0), control_word);
FPU_sub(0, FPU_rm, control_word);
}
......@@ -45,7 +47,7 @@ void fsubr_()
{
/* fsubr st,st(i) */
clear_C1();
reg_sub(&st(FPU_rm), &st(0), &st(0), control_word);
FPU_sub(REV, FPU_rm, control_word);
}
......@@ -53,7 +55,7 @@ void fdiv__()
{
/* fdiv st,st(i) */
clear_C1();
reg_div(&st(0), &st(FPU_rm), &st(0), control_word);
FPU_div(0, FPU_rm, control_word);
}
......@@ -61,7 +63,7 @@ void fdivr_()
{
/* fdivr st,st(i) */
clear_C1();
reg_div(&st(FPU_rm), &st(0), &st(0), control_word);
FPU_div(REV, FPU_rm, control_word);
}
......@@ -69,8 +71,9 @@ void fdivr_()
void fadd_i()
{
/* fadd st(i),st */
int i = FPU_rm;
clear_C1();
reg_add(&st(0), &st(FPU_rm), &st(FPU_rm), control_word);
FPU_add(&st(i), FPU_gettagi(i), i, control_word);
}
......@@ -78,27 +81,23 @@ void fmul_i()
{
/* fmul st(i),st */
clear_C1();
reg_mul(&st(0), &st(FPU_rm), &st(FPU_rm), control_word);
FPU_mul(&st(0), FPU_gettag0(), FPU_rm, control_word);
}
void fsubri()
{
/* fsubr st(i),st */
/* This is the sense of the 80486 manual
reg_sub(&st(FPU_rm), &st(0), &st(FPU_rm), control_word); */
clear_C1();
reg_sub(&st(0), &st(FPU_rm), &st(FPU_rm), control_word);
FPU_sub(DEST_RM, FPU_rm, control_word);
}
void fsub_i()
{
/* fsub st(i),st */
/* This is the sense of the 80486 manual
reg_sub(&st(0), &st(FPU_rm), &st(FPU_rm), control_word); */
clear_C1();
reg_sub(&st(FPU_rm), &st(0), &st(FPU_rm), control_word);
FPU_sub(REV|DEST_RM, FPU_rm, control_word);
}
......@@ -106,7 +105,7 @@ void fdivri()
{
/* fdivr st(i),st */
clear_C1();
reg_div(&st(0), &st(FPU_rm), &st(FPU_rm), control_word);
FPU_div(DEST_RM, FPU_rm, control_word);
}
......@@ -114,7 +113,7 @@ void fdiv_i()
{
/* fdiv st(i),st */
clear_C1();
reg_div(&st(FPU_rm), &st(0), &st(FPU_rm), control_word);
FPU_div(REV|DEST_RM, FPU_rm, control_word);
}
......@@ -122,9 +121,10 @@ void fdiv_i()
void faddp_()
{
/* faddp st(i),st */
int i = FPU_rm;
clear_C1();
if ( !reg_add(&st(0), &st(FPU_rm), &st(FPU_rm), control_word) )
pop();
if ( FPU_add(&st(i), FPU_gettagi(i), i, control_word) >= 0 )
FPU_pop();
}
......@@ -132,8 +132,8 @@ void fmulp_()
{
/* fmulp st(i),st */
clear_C1();
if ( !reg_mul(&st(0), &st(FPU_rm), &st(FPU_rm), control_word) )
pop();
if ( FPU_mul(&st(0), FPU_gettag0(), FPU_rm, control_word) >= 0 )
FPU_pop();
}
......@@ -141,22 +141,18 @@ void fmulp_()
void fsubrp()
{
/* fsubrp st(i),st */
/* This is the sense of the 80486 manual
reg_sub(&st(FPU_rm), &st(0), &st(FPU_rm), control_word); */
clear_C1();
if ( !reg_sub(&st(0), &st(FPU_rm), &st(FPU_rm), control_word) )
pop();
if ( FPU_sub(DEST_RM, FPU_rm, control_word) >= 0 )
FPU_pop();
}
void fsubp_()
{
/* fsubp st(i),st */
/* This is the sense of the 80486 manual
reg_sub(&st(0), &st(FPU_rm), &st(FPU_rm), control_word); */
clear_C1();
if ( !reg_sub(&st(FPU_rm), &st(0), &st(FPU_rm), control_word) )
pop();
if ( FPU_sub(REV|DEST_RM, FPU_rm, control_word) >= 0 )
FPU_pop();
}
......@@ -164,8 +160,8 @@ void fdivrp()
{
/* fdivrp st(i),st */
clear_C1();
if ( !reg_div(&st(0), &st(FPU_rm), &st(FPU_rm), control_word) )
pop();
if ( FPU_div(DEST_RM, FPU_rm, control_word) >= 0 )
FPU_pop();
}
......@@ -173,7 +169,6 @@ void fdivp_()
{
/* fdivp st(i),st */
clear_C1();
if ( !reg_div(&st(FPU_rm), &st(0), &st(FPU_rm), control_word) )
pop();
if ( FPU_div(REV|DEST_RM, FPU_rm, control_word) >= 0 )
FPU_pop();
}
arch/i386/math-emu/fpu_asm.h
View file @ 97ccc379
/*---------------------------------------------------------------------------+
| fpu_asm.h |
| |
| Copyright (C) 1992,1995 |
| Copyright (C) 1992,1995,1997 |
| W. Metzenthen, 22 Parker St, Ormond, Vic 3163, |
| Australia. E-mail billm@jacobi.maths.monash.edu.au |
| Australia. E-mail billm@suburbia.net |
| |
+---------------------------------------------------------------------------*/
......@@ -19,13 +19,14 @@
#define PARAM2 12(%ebp)
#define PARAM3 16(%ebp)
#define PARAM4 20(%ebp)
#define PARAM5 24(%ebp)
#define PARAM6 28(%ebp)
#define PARAM7 32(%ebp)
#define SIGL_OFFSET 8
#define SIGN(x) (x)
#define TAG(x) 1(x)
#define EXP(x) 4(x)
#define SIGL_OFFSET 0
#define EXP(x) 8(x)
#define SIG(x) SIGL_OFFSET##(x)
#define SIGL(x) SIGL_OFFSET##(x)
#define SIGH(x) 12(x)
#define SIGH(x) 4(x)
#endif _FPU_ASM_H_
arch/i386/math-emu/fpu_aux.c
View file @ 97ccc379
......@@ -3,9 +3,9 @@
| |
| Code to implement some of the FPU auxiliary instructions. |
| |
| Copyright (C) 1992,1993,1994 |
| W. Metzenthen, 22 Parker St, Ormond, Vic 3163, |
| Australia. E-mail billm@vaxc.cc.monash.edu.au |
| Copyright (C) 1992,1993,1994,1997 |
| W. Metzenthen, 22 Parker St, Ormond, Vic 3163, Australia |
| E-mail billm@suburbia.net |
| |
| |
+---------------------------------------------------------------------------*/
......@@ -32,15 +32,11 @@ void fclex(void)
/* Needs to be externally visible */
void finit()
{
int r;
control_word = 0x037f;
partial_status = 0;
top = 0; /* We don't keep top in the status word internally. */
for (r = 0; r < 8; r++)
{
regs[r].tag = TW_Empty;
}
/* The behaviour is different to that detailed in
fpu_tag_word = 0xffff;
/* The behaviour is different from that detailed in
Section 15.1.6 of the Intel manual */
operand_address.offset = 0;
operand_address.selector = 0;
......@@ -99,19 +95,27 @@ void fp_nop()
void fld_i_()
{
FPU_REG *st_new_ptr;
int i;
u_char tag;
if ( STACK_OVERFLOW )
{ stack_overflow(); return; }
{ FPU_stack_overflow(); return; }
/* fld st(i) */
if ( NOT_EMPTY(FPU_rm) )
{ reg_move(&st(FPU_rm), st_new_ptr); push(); }
i = FPU_rm;
if ( NOT_EMPTY(i) )
{
reg_copy(&st(i), st_new_ptr);
tag = FPU_gettagi(i);
push();
FPU_settag0(tag);
}
else
{
if ( control_word & CW_Invalid )
{
/* The masked response */
stack_underflow();
FPU_stack_underflow();
}
else
EXCEPTION(EX_StackUnder);
......@@ -124,61 +128,77 @@ void fxch_i()
{
/* fxch st(i) */
FPU_REG t;
register FPU_REG *sti_ptr = &st(FPU_rm), *st0_ptr = &st(0);
if ( st0_ptr->tag == TW_Empty )
int i = FPU_rm;
FPU_REG *st0_ptr = &st(0), *sti_ptr = &st(i);
long tag_word = fpu_tag_word;
int regnr = top & 7, regnri = ((regnr + i) & 7);
u_char st0_tag = (tag_word >> (regnr*2)) & 3;
u_char sti_tag = (tag_word >> (regnri*2)) & 3;
if ( st0_tag == TAG_Empty )
{
if ( sti_ptr->tag == TW_Empty )
if ( sti_tag == TAG_Empty )
{
stack_underflow();
stack_underflow_i(FPU_rm);
FPU_stack_underflow();
FPU_stack_underflow_i(i);
return;
}
if ( control_word & CW_Invalid )
reg_move(sti_ptr, st0_ptr); /* Masked response */
stack_underflow_i(FPU_rm);
{
/* Masked response */
FPU_copy_to_reg0(sti_ptr, sti_tag);
}
FPU_stack_underflow_i(i);
return;
}
if ( sti_ptr->tag == TW_Empty )
if ( sti_tag == TAG_Empty )
{
if ( control_word & CW_Invalid )
reg_move(st0_ptr, sti_ptr); /* Masked response */
stack_underflow();
{
/* Masked response */
FPU_copy_to_regi(st0_ptr, st0_tag, i);
}
FPU_stack_underflow();
return;
}
clear_C1();
reg_move(st0_ptr, &t);
reg_move(sti_ptr, st0_ptr);
reg_move(&t, sti_ptr);
reg_copy(st0_ptr, &t);
reg_copy(sti_ptr, st0_ptr);
reg_copy(&t, sti_ptr);
tag_word &= ~(3 << (regnr*2)) & ~(3 << (regnri*2));
tag_word |= (sti_tag << (regnr*2)) | (st0_tag << (regnri*2));
fpu_tag_word = tag_word;
}
void ffree_()
{
/* ffree st(i) */
st(FPU_rm).tag = TW_Empty;
FPU_settagi(FPU_rm, TAG_Empty);
}
void ffreep()
{
/* ffree st(i) + pop - unofficial code */
st(FPU_rm).tag = TW_Empty;
pop();
FPU_settagi(FPU_rm, TAG_Empty);
FPU_pop();
}
void fst_i_()
{
/* fst st(i) */
reg_move(&st(0), &st(FPU_rm));
FPU_copy_to_regi(&st(0), FPU_gettag0(), FPU_rm);
}
void fstp_i()
{
/* fstp st(i) */
reg_move(&st(0), &st(FPU_rm));
pop();
FPU_copy_to_regi(&st(0), FPU_gettag0(), FPU_rm);
FPU_pop();
}
arch/i386/math-emu/fpu_emu.h
View file @ 97ccc379
/*---------------------------------------------------------------------------+
| fpu_emu.h |
| |
| Copyright (C) 1992,1993,1994 |
| Copyright (C) 1992,1993,1994,1997 |
| W. Metzenthen, 22 Parker St, Ormond, Vic 3163, |
| Australia. E-mail billm@vaxc.cc.monash.edu.au |
| Australia. E-mail billm@suburbia.net |
| |
+---------------------------------------------------------------------------*/
......@@ -11,14 +11,6 @@
#ifndef _FPU_EMU_H_
#define _FPU_EMU_H_
/*
* Define DENORM_OPERAND to make the emulator detect denormals
* and use the denormal flag of the status word. Note: this only
* affects the flag and corresponding interrupt, the emulator
* will always generate denormals and operate upon them as required.
*/
#define DENORM_OPERAND
/*
* Define PECULIAR_486 to get a closer approximation to 80486 behaviour,
* rather than behaviour which appears to be cleaner.
......@@ -38,28 +30,51 @@
#define EXP_BIAS Const(0)
#define EXP_OVER Const(0x4000) /* smallest invalid large exponent */
#define EXP_UNDER Const(-0x3fff) /* largest invalid small exponent */
#define EXP_WAY_UNDER Const(-0x6000) /* Below the smallest denormal, but
still a 16 bit nr. */
#define EXP_Infinity EXP_OVER
#define EXP_NaN EXP_OVER
#define EXTENDED_Ebias Const(0x3fff)
#define EXTENDED_Emin (-0x3ffe) /* smallest valid exponent */
#define SIGN_POS Const(0)
#define SIGN_NEG Const(1)
#define SIGN_NEG Const(0x80)
#define SIGN_Positive Const(0)
#define SIGN_Negative Const(0x8000)
/* Keep the order TW_Valid, TW_Zero, TW_Denormal */
#define TW_Valid Const(0) /* valid */
#define TW_Zero Const(1) /* zero */
/* Keep the order TAG_Valid, TAG_Zero, TW_Denormal */
/* The following fold to 2 (Special) in the Tag Word */
/* #define TW_Denormal Const(4) */ /* De-normal */
#define TW_Denormal Const(4) /* De-normal */
#define TW_Infinity Const(5) /* + or - infinity */
#define TW_NaN Const(6) /* Not a Number */
#define TW_Unsupported Const(7) /* Not supported by an 80486 */
#define TAG_Valid Const(0) /* valid */
#define TAG_Zero Const(1) /* zero */
#define TAG_Special Const(2) /* De-normal, + or - infinity,
or Not a Number */
#define TAG_Empty Const(3) /* empty */
#define LOADED_DATA Const(10101) /* Special st() number to identify
loaded data (not on stack). */
/* A few flags (must be >= 0x10). */
#define REV 0x10
#define DEST_RM 0x20
#define LOADED 0x40
#define TW_Empty Const(7) /* empty */
#define FPU_Exception Const(0x80000000) /* Added to tag returns. */
#ifndef __ASSEMBLY__
#include <asm/sigcontext.h> /* for struct _fpstate */
#include <asm/math_emu.h>
#include "fpu_system.h"
#include <asm/sigcontext.h> /* for struct _fpstate */
#include <asm/math_emu.h>
#include <linux/linkage.h>
/*
......@@ -67,7 +82,7 @@
*/
#ifdef RE_ENTRANT_CHECKING
extern char emulating;
extern u_char emulating;
# define RE_ENTRANT_CHECK_OFF emulating = 0
# define RE_ENTRANT_CHECK_ON emulating = 1
#else
......@@ -97,18 +112,24 @@ extern char emulating;
struct address {
unsigned int offset;
unsigned short selector;
unsigned short opcode:11,
empty:5;
unsigned int selector:16;
unsigned int opcode:11;
unsigned int empty:5;
};
struct fpu__reg {
unsigned sigl;
unsigned sigh;
short exp;
};
typedef void (*FUNC)(void);
typedef struct fpu_reg FPU_REG;
typedef void (*FUNC_ST0)(FPU_REG *st0_ptr);
typedef struct { unsigned char address_size, operand_size, segment; }
typedef struct fpu__reg FPU_REG;
typedef void (*FUNC_ST0)(FPU_REG *st0_ptr, u_char st0_tag);
typedef struct { u_char address_size, operand_size, segment; }
overrides;
/* This structure is 32 bits: */
typedef struct { overrides override;
unsigned char default_mode; } fpu_addr_modes;
u_char default_mode; } fpu_addr_modes;
/* PROTECTED has a restricted meaning in the emulator; it is used
to signal that the emulator needs to do special things to ensure
that protection is respected in a segmented model. */
......@@ -117,27 +138,50 @@ typedef struct { overrides override;
#define VM86 SIXTEEN
#define PM16 (SIXTEEN | PROTECTED)
#define SEG32 PROTECTED
extern unsigned char const data_sizes_16[32];
extern u_char const data_sizes_16[32];
#define register_base ((u_char *) registers )
#define fpu_register(x) ( * ((FPU_REG *)( register_base + 10 * (x & 7) )) )
#define st(x) ( * ((FPU_REG *)( register_base + 10 * ((top+x) & 7) )) )
#define st(x) ( regs[((top+x) &7 )] )
#define STACK_OVERFLOW (FPU_stackoverflow(&st_new_ptr))
#define NOT_EMPTY(i) (!FPU_empty_i(i))
#define STACK_OVERFLOW (st_new_ptr = &st(-1), st_new_ptr->tag != TW_Empty)
#define NOT_EMPTY(i) (st(i).tag != TW_Empty)
#define NOT_EMPTY_ST0 (st0_tag ^ TW_Empty)
#define NOT_EMPTY_ST0 (st0_tag ^ TAG_Empty)
#define pop() { regs[(top++ & 7 )].tag = TW_Empty; }
#define poppop() { regs[((top + 1) & 7 )].tag \
= regs[(top & 7 )].tag = TW_Empty; \
top += 2; }
#define poppop() { FPU_pop(); FPU_pop(); }
/* push() does not affect the tags */
#define push() { top--; }
#define signbyte(a) (((u_char *)(a))[9])
#define getsign(a) (signbyte(a) & 0x80)
#define setsign(a,b) { if (b) signbyte(a) |= 0x80; else signbyte(a) &= 0x7f; }
#define copysign(a,b) { if (getsign(a)) signbyte(b) |= 0x80; \
else signbyte(b) &= 0x7f; }
#define changesign(a) { signbyte(a) ^= 0x80; }
#define setpositive(a) { signbyte(a) &= 0x7f; }
#define setnegative(a) { signbyte(a) |= 0x80; }
#define signpositive(a) ( (signbyte(a) & 0x80) == 0 )
#define signnegative(a) (signbyte(a) & 0x80)
#include "fpu_proto.h"
static inline void reg_copy(FPU_REG const *x, FPU_REG *y)
{
*(short *)&(y->exp) = *(const short *)&(x->exp);
*(long long *)&(y->sigl) = *(const long long *)&(x->sigl);
}
#define exponent(x) (((*(short *)&((x)->exp)) & 0x7fff) - EXTENDED_Ebias)
#define setexponentpos(x,y) { (*(short *)&((x)->exp)) = \
((y) + EXTENDED_Ebias) & 0x7fff; }
#define exponent16(x) (*(short *)&((x)->exp))
#define setexponent16(x,y) { (*(short *)&((x)->exp)) = (y); }
#define addexponent(x,y) { (*(short *)&((x)->exp)) += (y); }
#define stdexp(x) { (*(short *)&((x)->exp)) += EXTENDED_Ebias; }
#define reg_move(x, y) { \
*(short *)&((y)->sign) = *(const short *)&((x)->sign); \
*(long *)&((y)->exp) = *(const long *)&((x)->exp); \
*(long long *)&((y)->sigl) = *(const long long *)&((x)->sigl); }
#define isdenormal(ptr) (exponent(ptr) == EXP_BIAS+EXP_UNDER)
#define significand(x) ( ((unsigned long long *)&((x)->sigl))[0] )
......@@ -145,24 +189,26 @@ extern unsigned char const data_sizes_16[32];
/*----- Prototypes for functions written in assembler -----*/
/* extern void reg_move(FPU_REG *a, FPU_REG *b); */
asmlinkage void normalize(FPU_REG *x);
asmlinkage void normalize_nuo(FPU_REG *x);
asmlinkage int reg_div(FPU_REG const *arg1, FPU_REG const *arg2,
FPU_REG *answ, unsigned int control_w);
asmlinkage int reg_u_sub(FPU_REG const *arg1, FPU_REG const *arg2,
FPU_REG *answ, unsigned int control_w);
asmlinkage int reg_u_mul(FPU_REG const *arg1, FPU_REG const *arg2,
FPU_REG *answ, unsigned int control_w);
asmlinkage int reg_u_div(FPU_REG const *arg1, FPU_REG const *arg2,
FPU_REG *answ, unsigned int control_w);
asmlinkage int reg_u_add(FPU_REG const *arg1, FPU_REG const *arg2,
FPU_REG *answ, unsigned int control_w);
asmlinkage int wm_sqrt(FPU_REG *n, unsigned int control_w);
asmlinkage unsigned shrx(void *l, unsigned x);
asmlinkage unsigned shrxs(void *v, unsigned x);
asmlinkage unsigned long div_small(unsigned long long *x, unsigned long y);
asmlinkage void round_reg(FPU_REG *arg, unsigned int extent,
unsigned int control_w);
asmlinkage int FPU_normalize(FPU_REG *x);
asmlinkage int FPU_normalize_nuo(FPU_REG *x);
asmlinkage int FPU_u_sub(FPU_REG const *arg1, FPU_REG const *arg2,
FPU_REG *answ, unsigned int control_w, u_char sign,
int expa, int expb);
asmlinkage int FPU_u_mul(FPU_REG const *arg1, FPU_REG const *arg2,
FPU_REG *answ, unsigned int control_w, u_char sign,
int expon);
asmlinkage int FPU_u_div(FPU_REG const *arg1, FPU_REG const *arg2,
FPU_REG *answ, unsigned int control_w, u_char sign);
asmlinkage int FPU_u_add(FPU_REG const *arg1, FPU_REG const *arg2,
FPU_REG *answ, unsigned int control_w, u_char sign,
int expa, int expb);
asmlinkage int wm_sqrt(FPU_REG *n, int dummy1, int dummy2,
unsigned int control_w, u_char sign);
asmlinkage unsigned FPU_shrx(void *l, unsigned x);
asmlinkage unsigned FPU_shrxs(void *v, unsigned x);
asmlinkage unsigned long FPU_div_small(unsigned long long *x, unsigned long y);
asmlinkage int FPU_round(FPU_REG *arg, unsigned int extent, int dummy,
unsigned int control_w, u_char sign);
#ifndef MAKING_PROTO
#include "fpu_proto.h"
......
arch/i386/math-emu/fpu_entry.c
View file @ 97ccc379
This diff is collapsed.
arch/i386/math-emu/fpu_etc.c
View file @ 97ccc379
......@@ -3,9 +3,9 @@
| |
| Implement a few FPU instructions. |
| |
| Copyright (C) 1992,1993,1994 |
| Copyright (C) 1992,1993,1994,1997 |
| W. Metzenthen, 22 Parker St, Ormond, Vic 3163, |
| Australia. E-mail billm@vaxc.cc.monash.edu.au |
| Australia. E-mail billm@suburbia.net |
| |
| |
+---------------------------------------------------------------------------*/
......@@ -17,102 +17,116 @@
#include "reg_constant.h"
static void fchs(FPU_REG *st0_ptr)
static void fchs(FPU_REG *st0_ptr, u_char st0tag)
{
if ( st0_ptr->tag ^ TW_Empty )
if ( st0tag ^ TAG_Empty )
{
st0_ptr->sign ^= SIGN_POS^SIGN_NEG;
signbyte(st0_ptr) ^= SIGN_NEG;
clear_C1();
}
else
stack_underflow();
FPU_stack_underflow();
}
static void fabs(FPU_REG *st0_ptr)
static void fabs(FPU_REG *st0_ptr, u_char st0tag)
{
if ( st0_ptr->tag ^ TW_Empty )
if ( st0tag ^ TAG_Empty )
{
st0_ptr->sign = SIGN_POS;
setpositive(st0_ptr);
clear_C1();
}
else
stack_underflow();
FPU_stack_underflow();
}
static void ftst_(FPU_REG *st0_ptr)
static void ftst_(FPU_REG *st0_ptr, u_char st0tag)
{
switch (st0_ptr->tag)
switch (st0tag)
{
case TW_Zero:
case TAG_Zero:
setcc(SW_C3);
break;
case TW_Valid:
if (st0_ptr->sign == SIGN_POS)
case TAG_Valid:
if (getsign(st0_ptr) == SIGN_POS)
setcc(0);
else
setcc(SW_C0);
#ifdef DENORM_OPERAND
if ( (st0_ptr->exp <= EXP_UNDER) && (denormal_operand()) )
break;
case TAG_Special:
switch ( FPU_Special(st0_ptr) )
{
case TW_Denormal:
if (getsign(st0_ptr) == SIGN_POS)
setcc(0);
else
setcc(SW_C0);
if ( denormal_operand() < 0 )
{
#ifdef PECULIAR_486
/* This is weird! */
if (st0_ptr->sign == SIGN_POS)
setcc(SW_C3);
/* This is weird! */
if (getsign(st0_ptr) == SIGN_POS)
setcc(SW_C3);
#endif PECULIAR_486
return;
return;
}
break;
case TW_NaN:
setcc(SW_C0|SW_C2|SW_C3); /* Operand is not comparable */
EXCEPTION(EX_Invalid);
break;
case TW_Infinity:
if (getsign(st0_ptr) == SIGN_POS)
setcc(0);
else
setcc(SW_C0);
break;
default:
setcc(SW_C0|SW_C2|SW_C3); /* Operand is not comparable */
EXCEPTION(EX_INTERNAL|0x14);
break;
}
#endif DENORM_OPERAND
break;
case TW_NaN:
setcc(SW_C0|SW_C2|SW_C3); /* Operand is not comparable */
EXCEPTION(EX_Invalid);
break;
case TW_Infinity:
if (st0_ptr->sign == SIGN_POS)
setcc(0);
else
setcc(SW_C0);
break;
case TW_Empty:
case TAG_Empty:
setcc(SW_C0|SW_C2|SW_C3);
EXCEPTION(EX_StackUnder);
break;
default:
setcc(SW_C0|SW_C2|SW_C3); /* Operand is not comparable */
EXCEPTION(EX_INTERNAL|0x14);
break;
}
}
static void fxam(FPU_REG *st0_ptr)
static void fxam(FPU_REG *st0_ptr, u_char st0tag)
{
int c=0;
switch (st0_ptr->tag)
int c = 0;
switch (st0tag)
{
case TW_Empty:
case TAG_Empty:
c = SW_C3|SW_C0;
break;
case TW_Zero:
case TAG_Zero:
c = SW_C3;
break;
case TW_Valid:
/* This will need to be changed if TW_Denormal is ever used. */
if ( st0_ptr->exp <= EXP_UNDER )
c = SW_C2|SW_C3; /* Denormal */
else
c = SW_C2;
break;
case TW_NaN:
c = SW_C0;
break;
case TW_Infinity:
c = SW_C2|SW_C0;
case TAG_Valid:
c = SW_C2;
break;
case TAG_Special:
switch ( FPU_Special(st0_ptr) )
{
case TW_Denormal:
c = SW_C2|SW_C3; /* Denormal */
break;
case TW_NaN:
/* We also use NaN for unsupported types. */
if ( (st0_ptr->sigh & 0x80000000) && (exponent(st0_ptr) == EXP_OVER) )
c = SW_C0;
break;
case TW_Infinity:
c = SW_C2|SW_C0;
break;
}
}
if (st0_ptr->sign == SIGN_NEG)
if ( getsign(st0_ptr) == SIGN_NEG )
c |= SW_C1;
setcc(c);
}
......@@ -123,7 +137,7 @@ static FUNC_ST0 const fp_etc_table[] = {
ftst_, fxam, (FUNC_ST0)FPU_illegal, (FUNC_ST0)FPU_illegal
};
void fp_etc()
void FPU_etc()
{
(fp_etc_table[FPU_rm])(&st(0));
(fp_etc_table[FPU_rm])(&st(0), FPU_gettag0());
}
arch/i386/math-emu/fpu_proto.h
View file @ 97ccc379
#ifndef _FPU_PROTO_H
#define _FPU_PROTO_H
/* errors.c */
extern void Un_impl(void);
extern void FPU_illegal(void);
extern void emu_printall(void);
extern void stack_overflow(void);
extern void stack_underflow(void);
extern void stack_underflow_i(int i);
extern void stack_underflow_pop(int i);
extern int set_precision_flag(int flags);
extern void FPU_printall(void);
asmlinkage void FPU_exception(int n);
asmlinkage int real_2op_NaN(FPU_REG const *a, FPU_REG const *b, FPU_REG *dest);
asmlinkage int arith_invalid(FPU_REG *dest);
asmlinkage int divide_by_zero(int sign, FPU_REG *dest);
asmlinkage void set_precision_flag_up(void);
asmlinkage void set_precision_flag_down(void);
asmlinkage int denormal_operand(void);
asmlinkage int arith_overflow(FPU_REG *dest);
asmlinkage int arith_underflow(FPU_REG *dest);
extern int real_1op_NaN(FPU_REG *a);
extern int real_2op_NaN(FPU_REG const *b, u_char tagb, int deststnr,
FPU_REG const *defaultNaN);
extern int arith_invalid(int deststnr);
extern int FPU_divide_by_zero(int deststnr, u_char sign);
extern int set_precision_flag(int flags);
extern void set_precision_flag_up(void);
extern void set_precision_flag_down(void);
extern int denormal_operand(void);
extern int arith_overflow(FPU_REG *dest);
extern int arith_underflow(FPU_REG *dest);
extern void FPU_stack_overflow(void);
extern void FPU_stack_underflow(void);
extern void FPU_stack_underflow_i(int i);
extern void FPU_stack_underflow_pop(int i);
/* fpu_arith.c */
extern void fadd__(void);
extern void fmul__(void);
......@@ -36,7 +40,6 @@ extern void fsubrp(void);
extern void fsubp_(void);
extern void fdivrp(void);
extern void fdivp_(void);
/* fpu_aux.c */
extern void fclex(void);
extern void finit(void);
......@@ -49,89 +52,92 @@ extern void ffree_(void);
extern void ffreep(void);
extern void fst_i_(void);
extern void fstp_i(void);
/* fpu_entry.c */
asmlinkage void math_emulate(long arg);
extern void math_emulate(long arg);
extern void math_abort(struct info *info, unsigned int signal);
/* fpu_etc.c */
extern void fp_etc(void);
extern void FPU_etc(void);
/* fpu_tags.c */
extern int FPU_gettag0(void);
extern int FPU_gettagi(int stnr);
extern int FPU_gettag(int regnr);
extern void FPU_settag0(int tag);
extern void FPU_settagi(int stnr, int tag);
extern void FPU_settag(int regnr, int tag);
extern int FPU_Special(FPU_REG const *ptr);
extern int isNaN(FPU_REG const *ptr);
extern void FPU_pop(void);
extern int FPU_empty_i(int stnr);
extern int FPU_stackoverflow(FPU_REG **st_new_ptr);
extern void FPU_sync_tags(void);
extern void FPU_copy_to_regi(FPU_REG const *r, u_char tag, int stnr);
extern void FPU_copy_to_reg1(FPU_REG const *r, u_char tag);
extern void FPU_copy_to_reg0(FPU_REG const *r, u_char tag);
/* fpu_trig.c */
extern void convert_l2reg(long const *arg, FPU_REG *dest);
extern void trig_a(void);
extern void trig_b(void);
extern void FPU_triga(void);
extern void FPU_trigb(void);
/* get_address.c */
extern void *get_address(unsigned char FPU_modrm, unsigned long *fpu_eip,
struct address *addr,
fpu_addr_modes);
extern void *get_address_16(unsigned char FPU_modrm, unsigned long *fpu_eip,
struct address *addr,
fpu_addr_modes);
extern void *FPU_get_address(u_char FPU_modrm, unsigned long *fpu_eip,
struct address *addr, fpu_addr_modes addr_modes);
extern void *FPU_get_address_16(u_char FPU_modrm, unsigned long *fpu_eip,
struct address *addr, fpu_addr_modes addr_modes);
/* load_store.c */
extern int load_store_instr(unsigned char type, fpu_addr_modes addr_modes,
void *address);
extern int FPU_load_store(u_char type, fpu_addr_modes addr_modes,
void *data_address);
/* poly_2xm1.c */
extern int poly_2xm1(FPU_REG const *arg, FPU_REG *result);
extern int poly_2xm1(u_char sign, FPU_REG *arg, FPU_REG *result);
/* poly_atan.c */
extern void poly_atan(FPU_REG *arg1, FPU_REG *arg2, FPU_REG *result);
extern void poly_atan(FPU_REG *st0_ptr, u_char st0_tag, FPU_REG *st1_ptr,
u_char st1_tag);
/* poly_l2.c */
extern void poly_l2(FPU_REG const *arg, FPU_REG const *y, FPU_REG *result);
extern int poly_l2p1(FPU_REG const *arg, FPU_REG const *y, FPU_REG *result);
extern void poly_l2(FPU_REG *st0_ptr, FPU_REG *st1_ptr, u_char st1_sign);
extern int poly_l2p1(u_char s0, u_char s1, FPU_REG *r0, FPU_REG *r1,
FPU_REG *d);
/* poly_sin.c */
extern void poly_sine(FPU_REG const *arg, FPU_REG *result);
extern void poly_cos(FPU_REG const *arg, FPU_REG *result);
extern void poly_sine(FPU_REG *st0_ptr);
extern void poly_cos(FPU_REG *st0_ptr);
/* poly_tan.c */
extern void poly_tan(FPU_REG const *arg, FPU_REG *result);
extern void poly_tan(FPU_REG *st0_ptr);
/* reg_add_sub.c */
extern int reg_add(FPU_REG const *a, FPU_REG const *b,
FPU_REG *dest, int control_w);
extern int reg_sub(FPU_REG const *a, FPU_REG const *b,
FPU_REG *dest, int control_w);
extern int FPU_add(FPU_REG const *b, u_char tagb, int destrnr, int control_w);
extern int FPU_sub(int flags, int rm, int control_w);
/* reg_compare.c */
extern int compare(FPU_REG const *b);
extern int compare_st_data(FPU_REG const *b);
extern int FPU_compare_st_data(FPU_REG const *loaded_data, u_char loaded_tag);
extern void fcom_st(void);
extern void fcompst(void);
extern void fcompp(void);
extern void fucom_(void);
extern void fucomp(void);
extern void fucompp(void);
/* reg_constant.c */
extern void fconst(void);
/* reg_ld_str.c */
extern int reg_load_extended(long double *addr, FPU_REG *loaded_data);
extern int reg_load_double(double *dfloat, FPU_REG *loaded_data);
extern int reg_load_single(float *single, FPU_REG *loaded_data);
extern void reg_load_int64(long long *_s, FPU_REG *loaded_data);
extern void reg_load_int32(long *_s, FPU_REG *loaded_data);
extern void reg_load_int16(short *_s, FPU_REG *loaded_data);
extern void reg_load_bcd(char *s, FPU_REG *loaded_data);
extern int reg_store_extended(long double *d, FPU_REG *st0_ptr);
extern int reg_store_double(double *dfloat, FPU_REG *st0_ptr);
extern int reg_store_single(float *single, FPU_REG *st0_ptr);
extern int reg_store_int64(long long *d, FPU_REG *st0_ptr);
extern int reg_store_int32(long *d, FPU_REG *st0_ptr);
extern int reg_store_int16(short *d, FPU_REG *st0_ptr);
extern int reg_store_bcd(char *d, FPU_REG *st0_ptr);
extern int round_to_int(FPU_REG *r);
extern char *fldenv(fpu_addr_modes addr_modes, char *address);
extern void frstor(fpu_addr_modes addr_modes, char *address);
extern unsigned short tag_word(void);
extern char *fstenv(fpu_addr_modes addr_modes, char *address);
extern void fsave(fpu_addr_modes addr_modes, char *address);
extern int FPU_load_extended(long double *s, int stnr);
extern int FPU_load_double(double *dfloat, FPU_REG *loaded_data);
extern int FPU_load_single(float *single, FPU_REG *loaded_data);
extern int FPU_load_int64(long long *_s);
extern int FPU_load_int32(long *_s, FPU_REG *loaded_data);
extern int FPU_load_int16(short *_s, FPU_REG *loaded_data);
extern int FPU_load_bcd(u_char *s);
extern int FPU_store_extended(FPU_REG *st0_ptr, u_char st0_tag,
long double *d);
extern int FPU_store_double(FPU_REG *st0_ptr, u_char st0_tag, double *dfloat);
extern int FPU_store_single(FPU_REG *st0_ptr, u_char st0_tag, float *single);
extern int FPU_store_int64(FPU_REG *st0_ptr, u_char st0_tag, long long *d);
extern int FPU_store_int32(FPU_REG *st0_ptr, u_char st0_tag, long *d);
extern int FPU_store_int16(FPU_REG *st0_ptr, u_char st0_tag, short *d);
extern int FPU_store_bcd(FPU_REG *st0_ptr, u_char st0_tag, u_char *d);
extern int FPU_round_to_int(FPU_REG *r, u_char tag);
extern u_char *fldenv(fpu_addr_modes addr_modes, u_char *s);
extern void frstor(fpu_addr_modes addr_modes, u_char *data_address);
extern u_char *fstenv(fpu_addr_modes addr_modes, u_char *d);
extern void fsave(fpu_addr_modes addr_modes, u_char *data_address);
extern int FPU_tagof(FPU_REG *ptr);
/* reg_mul.c */
extern int reg_mul(FPU_REG const *a, FPU_REG const *b,
FPU_REG *dest, unsigned int control_w);
extern int FPU_mul(FPU_REG const *b, u_char tagb, int deststnr, int control_w);
extern int FPU_div(int flags, int regrm, int control_w);
/* reg_convert.c */
extern int FPU_to_exp16(FPU_REG const *a, FPU_REG *x);
#endif /* _FPU_PROTO_H */
arch/i386/math-emu/fpu_system.h
View file @ 97ccc379
/*---------------------------------------------------------------------------+
| fpu_system.h |
| |
| Copyright (C) 1992,1994 |
| Copyright (C) 1992,1994,1997 |
| W. Metzenthen, 22 Parker St, Ormond, Vic 3163, |
| Australia. E-mail billm@vaxc.cc.monash.edu.au |
| Australia. E-mail billm@suburbia.net |
| |
+---------------------------------------------------------------------------*/
......@@ -18,19 +18,19 @@
/* This sets the pointer FPU_info to point to the argument part
of the stack frame of math_emulate() */
#define SETUP_DATA_AREA(arg) FPU_info = (struct info *) &arg
#define LDT_DESCRIPTOR(s) (current->ldt[(s) >> 3])
#define SEG_D_SIZE(x) ((x).b & (3 << 21))
#define SEG_G_BIT(x) ((x).b & (1 << 23))
#define SEG_GRANULARITY(x) (((x).b & (1 << 23)) ? 4096 : 1)
#define SEG_286_MODE(x) ((x).b & ( 0xff000000 | 0xf0000 | (1 << 23)))
#define SEG_BASE_ADDR(s) (((s).b & 0xff000000) \
#define SETUP_DATA_AREA(arg) FPU_info = (struct info *) &arg
#define LDT_DESCRIPTOR(s) (current->ldt[(s) >> 3])
#define SEG_D_SIZE(x) ((x).b & (3 << 21))
#define SEG_G_BIT(x) ((x).b & (1 << 23))
#define SEG_GRANULARITY(x) (((x).b & (1 << 23)) ? 4096 : 1)
#define SEG_286_MODE(x) ((x).b & ( 0xff000000 | 0xf0000 | (1 << 23)))
#define SEG_BASE_ADDR(s) (((s).b & 0xff000000) \
| (((s).b & 0xff) << 16) | ((s).a >> 16))
#define SEG_LIMIT(s) (((s).b & 0xff0000) | ((s).a & 0xffff))
#define SEG_EXECUTE_ONLY(s) (((s).b & ((1 << 11) | (1 << 9))) == (1 << 11))
#define SEG_WRITE_PERM(s) (((s).b & ((1 << 11) | (1 << 9))) == (1 << 9))
#define SEG_EXPAND_DOWN(s) (((s).b & ((1 << 11) | (1 << 10))) \
#define SEG_LIMIT(s) (((s).b & 0xff0000) | ((s).a & 0xffff))
#define SEG_EXECUTE_ONLY(s) (((s).b & ((1 << 11) | (1 << 9))) == (1 << 11))
#define SEG_WRITE_PERM(s) (((s).b & ((1 << 11) | (1 << 9))) == (1 << 9))
#define SEG_EXPAND_DOWN(s) (((s).b & ((1 << 11) | (1 << 10))) \
== (1 << 10))
#define I387 (current->tss.i387)
......@@ -48,23 +48,24 @@
/* nz if ip_offset and cs_selector are not to be set for the current
instruction. */
#define no_ip_update (((char *)&(I387.soft.twd))[0])
#define FPU_rm (((unsigned char *)&(I387.soft.twd))[1])
#define no_ip_update (*(u_char *)&(I387.soft.no_update))
#define FPU_rm (*(u_char *)&(I387.soft.rm))
/* Number of bytes of data which can be legally accessed by the current
instruction. This only needs to hold a number <= 108, so a byte will do. */
#define access_limit (((unsigned char *)&(I387.soft.twd))[2])
#define access_limit (*(u_char *)&(I387.soft.alimit))
#define partial_status (I387.soft.swd)
#define partial_status (I387.soft.swd)
#define control_word (I387.soft.cwd)
#define regs (I387.soft.regs)
#define top (I387.soft.top)
#define fpu_tag_word (I387.soft.twd)
#define registers (I387.soft.st_space)
#define top (I387.soft.ftop)
#define instruction_address (*(struct address *)&I387.soft.fip)
#define operand_address (*(struct address *)&I387.soft.foo)
#define instruction_address (*(struct address *)&I387.soft.fip)
#define operand_address (*(struct address *)&I387.soft.foo)
#define FPU_verify_area(x,y,z) if ( verify_area(x,y,z) ) \
math_abort(FPU_info,SIGSEGV)
#define FPU_verify_area(x,y,z) if ( verify_area(x,y,z) ) \
math_abort(FPU_info,SIGSEGV)
#undef FPU_IGNORE_CODE_SEGV
#ifdef FPU_IGNORE_CODE_SEGV
......@@ -80,4 +81,7 @@
#define FPU_code_verify_area(z) FPU_verify_area(VERIFY_READ,(void *)FPU_EIP,z)
#endif
#define FPU_get_user(x,y) get_user((x),(y))
#define FPU_put_user(x,y) put_user((x),(y))
#endif
arch/i386/math-emu/fpu_tags.c 0 → 100644
View file @ 97ccc379
/*---------------------------------------------------------------------------+
| fpu_tags.c |
| |
| Set FPU register tags. |
| |
| Copyright (C) 1997 |
| W. Metzenthen, 22 Parker St, Ormond, Vic 3163, Australia |
| E-mail billm@jacobi.maths.monash.edu.au |
| |
| |
+---------------------------------------------------------------------------*/
#include "fpu_emu.h"
#include "fpu_system.h"
#include "exception.h"
void FPU_pop(void)
{
fpu_tag_word |= 3 << ((top & 7)*2);
top++;
}
int FPU_gettag0(void)
{
return (fpu_tag_word >> ((top & 7)*2)) & 3;
}
int FPU_gettagi(int stnr)
{
return (fpu_tag_word >> (((top+stnr) & 7)*2)) & 3;
}
int FPU_gettag(int regnr)
{
return (fpu_tag_word >> ((regnr & 7)*2)) & 3;
}
void FPU_settag0(int tag)
{
int regnr = top;
regnr &= 7;
fpu_tag_word &= ~(3 << (regnr*2));
fpu_tag_word |= (tag & 3) << (regnr*2);
}
void FPU_settagi(int stnr, int tag)
{
int regnr = stnr+top;
regnr &= 7;
fpu_tag_word &= ~(3 << (regnr*2));
fpu_tag_word |= (tag & 3) << (regnr*2);
}
void FPU_settag(int regnr, int tag)
{
regnr &= 7;
fpu_tag_word &= ~(3 << (regnr*2));
fpu_tag_word |= (tag & 3) << (regnr*2);
}
int FPU_Special(FPU_REG const *ptr)
{
int exp = exponent(ptr);
if ( exp == EXP_BIAS+EXP_UNDER )
return TW_Denormal;
else if ( exp != EXP_BIAS+EXP_OVER )
return TW_NaN;
else if ( (ptr->sigh == 0x80000000) && (ptr->sigl == 0) )
return TW_Infinity;
return TW_NaN;
}
int isNaN(FPU_REG const *ptr)
{
return ( (exponent(ptr) == EXP_BIAS+EXP_OVER)
&& !((ptr->sigh == 0x80000000) && (ptr->sigl == 0)) );
}
int FPU_empty_i(int stnr)
{
int regnr = (top+stnr) & 7;
return ((fpu_tag_word >> (regnr*2)) & 3) == TAG_Empty;
}
int FPU_stackoverflow(FPU_REG **st_new_ptr)
{
*st_new_ptr = &st(-1);
return ((fpu_tag_word >> (((top - 1) & 7)*2)) & 3) != TAG_Empty;
}
void FPU_copy_to_regi(FPU_REG const *r, u_char tag, int stnr)
{
reg_copy(r, &st(stnr));
FPU_settagi(stnr, tag);
}
void FPU_copy_to_reg1(FPU_REG const *r, u_char tag)
{
reg_copy(r, &st(1));
FPU_settagi(1, tag);
}
void FPU_copy_to_reg0(FPU_REG const *r, u_char tag)
{
int regnr = top;
regnr &= 7;
reg_copy(r, &st(0));
fpu_tag_word &= ~(3 << (regnr*2));
fpu_tag_word |= (tag & 3) << (regnr*2);
}
arch/i386/math-emu/fpu_trig.c
View file @ 97ccc379
This diff is collapsed.
arch/i386/math-emu/get_address.c
View file @ 97ccc379
......@@ -3,9 +3,9 @@
| |
| Get the effective address from an FPU instruction. |
| |
| Copyright (C) 1992,1993,1994 |
| Copyright (C) 1992,1993,1994,1997 |
| W. Metzenthen, 22 Parker St, Ormond, Vic 3163, |
| Australia. E-mail billm@vaxc.cc.monash.edu.au |
| Australia. E-mail billm@suburbia.net |
| |
| |
+---------------------------------------------------------------------------*/
......@@ -41,7 +41,7 @@ static int reg_offset[] = {
offsetof(struct info,___edi)
};
#define REG_(x) (*(long *)(reg_offset[(x)]+(char *) FPU_info))
#define REG_(x) (*(long *)(reg_offset[(x)]+(u_char *) FPU_info))
static int reg_offset_vm86[] = {
offsetof(struct info,___cs),
......@@ -54,7 +54,7 @@ static int reg_offset_vm86[] = {
};
#define VM86_REG_(x) (*(unsigned short *) \
(reg_offset_vm86[((unsigned)x)]+(char *) FPU_info))
(reg_offset_vm86[((unsigned)x)]+(u_char *) FPU_info))
/* These are dummy, fs and gs are not saved on the stack. */
#define ___FS ___ds
......@@ -71,18 +71,18 @@ static int reg_offset_pm[] = {
};
#define PM_REG_(x) (*(unsigned short *) \
(reg_offset_pm[((unsigned)x)]+(char *) FPU_info))
(reg_offset_pm[((unsigned)x)]+(u_char *) FPU_info))
/* Decode the SIB byte. This function assumes mod != 0 */
static int sib(int mod, unsigned long *fpu_eip)
{
unsigned char ss,index,base;
u_char ss,index,base;
long offset;
RE_ENTRANT_CHECK_OFF;
FPU_code_verify_area(1);
get_user(base, (unsigned char *) (*fpu_eip)); /* The SIB byte */
FPU_get_user(base, (u_char *) (*fpu_eip)); /* The SIB byte */
RE_ENTRANT_CHECK_ON;
(*fpu_eip)++;
ss = base >> 6;
......@@ -112,7 +112,7 @@ static int sib(int mod, unsigned long *fpu_eip)
long displacement;
RE_ENTRANT_CHECK_OFF;
FPU_code_verify_area(1);
get_user(displacement, (signed char *) (*fpu_eip));
FPU_get_user(displacement, (signed char *) (*fpu_eip));
offset += displacement;
RE_ENTRANT_CHECK_ON;
(*fpu_eip)++;
......@@ -123,7 +123,7 @@ static int sib(int mod, unsigned long *fpu_eip)
long displacement;
RE_ENTRANT_CHECK_OFF;
FPU_code_verify_area(4);
get_user(displacement, (signed long *) (*fpu_eip));
FPU_get_user(displacement, (long *) (*fpu_eip));
offset += displacement;
RE_ENTRANT_CHECK_ON;
(*fpu_eip) += 4;
......@@ -133,7 +133,7 @@ static int sib(int mod, unsigned long *fpu_eip)
}
static unsigned long vm86_segment(unsigned char segment,
static unsigned long vm86_segment(u_char segment,
unsigned short *selector)
{
segment--;
......@@ -150,7 +150,7 @@ static unsigned long vm86_segment(unsigned char segment,
/* This should work for 16 and 32 bit protected mode. */
static long pm_address(unsigned char FPU_modrm, unsigned char segment,
static long pm_address(u_char FPU_modrm, u_char segment,
unsigned short *selector, long offset)
{
struct desc_struct descriptor;
......@@ -233,12 +233,11 @@ static long pm_address(unsigned char FPU_modrm, unsigned char segment,
*/
void *get_address(unsigned char FPU_modrm, unsigned long *fpu_eip,
void *FPU_get_address(u_char FPU_modrm, unsigned long *fpu_eip,
struct address *addr,
/* unsigned short *selector, unsigned long *offset, */
fpu_addr_modes addr_modes)
{
unsigned char mod;
u_char mod;
unsigned rm = FPU_modrm & 7;
long *cpu_reg_ptr;
int address = 0; /* Initialized just to stop compiler warnings. */
......@@ -270,7 +269,7 @@ void *get_address(unsigned char FPU_modrm, unsigned long *fpu_eip,
/* Special case: disp32 */
RE_ENTRANT_CHECK_OFF;
FPU_code_verify_area(4);
get_user(address, (unsigned long *) (*fpu_eip));
FPU_get_user(address, (unsigned long *) (*fpu_eip));
(*fpu_eip) += 4;
RE_ENTRANT_CHECK_ON;
addr->offset = address;
......@@ -287,7 +286,7 @@ void *get_address(unsigned char FPU_modrm, unsigned long *fpu_eip,
/* 8 bit signed displacement */
RE_ENTRANT_CHECK_OFF;
FPU_code_verify_area(1);
get_user(address, (signed char *) (*fpu_eip));
FPU_get_user(address, (signed char *) (*fpu_eip));
RE_ENTRANT_CHECK_ON;
(*fpu_eip)++;
break;
......@@ -295,7 +294,7 @@ void *get_address(unsigned char FPU_modrm, unsigned long *fpu_eip,
/* 32 bit displacement */
RE_ENTRANT_CHECK_OFF;
FPU_code_verify_area(4);
get_user(address, (long *) (*fpu_eip));
FPU_get_user(address, (long *) (*fpu_eip));
(*fpu_eip) += 4;
RE_ENTRANT_CHECK_ON;
break;
......@@ -329,12 +328,11 @@ void *get_address(unsigned char FPU_modrm, unsigned long *fpu_eip,
}
void *get_address_16(unsigned char FPU_modrm, unsigned long *fpu_eip,
void *FPU_get_address_16(u_char FPU_modrm, unsigned long *fpu_eip,
struct address *addr,
/* unsigned short *selector, unsigned long *offset, */
fpu_addr_modes addr_modes)
{
unsigned char mod;
u_char mod;
unsigned rm = FPU_modrm & 7;
int address = 0; /* Default used for mod == 0 */
......@@ -358,7 +356,7 @@ void *get_address_16(unsigned char FPU_modrm, unsigned long *fpu_eip,
/* Special case: disp16 */
RE_ENTRANT_CHECK_OFF;
FPU_code_verify_area(2);
get_user(address, (unsigned short *) (*fpu_eip));
FPU_get_user(address, (unsigned short *) (*fpu_eip));
(*fpu_eip) += 2;
RE_ENTRANT_CHECK_ON;
goto add_segment;
......@@ -368,7 +366,7 @@ void *get_address_16(unsigned char FPU_modrm, unsigned long *fpu_eip,
/* 8 bit signed displacement */
RE_ENTRANT_CHECK_OFF;
FPU_code_verify_area(1);
get_user(address, (signed char *) (*fpu_eip));
FPU_get_user(address, (signed char *) (*fpu_eip));
RE_ENTRANT_CHECK_ON;
(*fpu_eip)++;
break;
......@@ -376,7 +374,7 @@ void *get_address_16(unsigned char FPU_modrm, unsigned long *fpu_eip,
/* 16 bit displacement */
RE_ENTRANT_CHECK_OFF;
FPU_code_verify_area(2);
get_user(address, (unsigned short *) (*fpu_eip));
FPU_get_user(address, (unsigned short *) (*fpu_eip));
(*fpu_eip) += 2;
RE_ENTRANT_CHECK_ON;
break;
......
arch/i386/math-emu/load_store.c
View file @ 97ccc379
......@@ -4,9 +4,9 @@
| This file contains most of the code to interpret the FPU instructions |
| which load and store from user memory. |
| |
| Copyright (C) 1992,1993,1994 |
| Copyright (C) 1992,1993,1994,1997 |
| W. Metzenthen, 22 Parker St, Ormond, Vic 3163, |
| Australia. E-mail billm@vaxc.cc.monash.edu.au |
| Australia. E-mail billm@suburbia.net |
| |
| |
+---------------------------------------------------------------------------*/
......@@ -32,10 +32,10 @@
#define _PUSH_ 3 /* Need to check for space to push onto stack */
#define _null_ 4 /* Function illegal or not implemented */
#define pop_0() { st0_ptr->tag = TW_Empty; top++; }
#define pop_0() { FPU_settag0(TAG_Empty); top++; }
static unsigned char const type_table[32] = {
static u_char const type_table[32] = {
_PUSH_, _PUSH_, _PUSH_, _PUSH_,
_null_, _null_, _null_, _null_,
_REG0_, _REG0_, _REG0_, _REG0_,
......@@ -46,25 +46,27 @@ static unsigned char const type_table[32] = {
_NONE_, _REG0_, _NONE_, _REG0_
};
unsigned char const data_sizes_16[32] = {
u_char const data_sizes_16[32] = {
4, 4, 8, 2, 0, 0, 0, 0,
4, 4, 8, 2, 4, 4, 8, 2,
14, 0, 94, 10, 2, 10, 0, 8,
14, 0, 94, 10, 2, 10, 2, 8
};
unsigned char const data_sizes_32[32] = {
u_char const data_sizes_32[32] = {
4, 4, 8, 2, 0, 0, 0, 0,
4, 4, 8, 2, 4, 4, 8, 2,
28, 0,108, 10, 2, 10, 0, 8,
28, 0,108, 10, 2, 10, 2, 8
};
int load_store_instr(unsigned char type, fpu_addr_modes addr_modes,
int FPU_load_store(u_char type, fpu_addr_modes addr_modes,
void *data_address)
{
FPU_REG loaded_data;
FPU_REG *st0_ptr;
u_char st0_tag = TAG_Empty; /* This is just to stop a gcc warning. */
u_char loaded_tag;
st0_ptr = NULL; /* Initialized just to stop compiler warnings. */
......@@ -93,13 +95,14 @@ int load_store_instr(unsigned char type, fpu_addr_modes addr_modes,
case _REG0_:
st0_ptr = &st(0); /* Some of these instructions pop after
storing */
st0_tag = FPU_gettag0();
break;
case _PUSH_:
{
st0_ptr = &st(-1);
if ( st0_ptr->tag != TW_Empty )
{ stack_overflow(); return 0; }
if ( FPU_gettagi(-1) != TAG_Empty )
{ FPU_stack_overflow(); return 0; }
top--;
st0_ptr = &st(0);
}
break;
case _null_:
......@@ -116,92 +119,97 @@ int load_store_instr(unsigned char type, fpu_addr_modes addr_modes,
{
case 000: /* fld m32real */
clear_C1();
reg_load_single((float *)data_address, &loaded_data);
if ( (loaded_data.tag == TW_NaN) &&
real_2op_NaN(&loaded_data, &loaded_data, &loaded_data) )
loaded_tag = FPU_load_single((float *)data_address, &loaded_data);
if ( (loaded_tag == TAG_Special)
&& isNaN(&loaded_data)
&& (real_1op_NaN(&loaded_data) < 0) )
{
top++;
break;
}
reg_move(&loaded_data, st0_ptr);
FPU_copy_to_reg0(&loaded_data, loaded_tag);
break;
case 001: /* fild m32int */
clear_C1();
reg_load_int32((long *)data_address, st0_ptr);
loaded_tag = FPU_load_int32((long *)data_address, &loaded_data);
FPU_copy_to_reg0(&loaded_data, loaded_tag);
break;
case 002: /* fld m64real */
clear_C1();
reg_load_double((double *)data_address, &loaded_data);
if ( (loaded_data.tag == TW_NaN) &&
real_2op_NaN(&loaded_data, &loaded_data, &loaded_data) )
loaded_tag = FPU_load_double((double *)data_address, &loaded_data);
if ( (loaded_tag == TAG_Special)
&& isNaN(&loaded_data)
&& (real_1op_NaN(&loaded_data) < 0) )
{
top++;
break;
}
reg_move(&loaded_data, st0_ptr);
FPU_copy_to_reg0(&loaded_data, loaded_tag);
break;
case 003: /* fild m16int */
clear_C1();
reg_load_int16((short *)data_address, st0_ptr);
loaded_tag = FPU_load_int16((short *)data_address, &loaded_data);
FPU_copy_to_reg0(&loaded_data, loaded_tag);
break;
case 010: /* fst m32real */
clear_C1();
reg_store_single((float *)data_address, st0_ptr);
FPU_store_single(st0_ptr, st0_tag, (float *)data_address);
break;
case 011: /* fist m32int */
clear_C1();
reg_store_int32((long *)data_address, st0_ptr);
FPU_store_int32(st0_ptr, st0_tag, (long *)data_address);
break;
case 012: /* fst m64real */
clear_C1();
reg_store_double((double *)data_address, st0_ptr);
FPU_store_double(st0_ptr, st0_tag, (double *)data_address);
break;
case 013: /* fist m16int */
clear_C1();
reg_store_int16((short *)data_address, st0_ptr);
FPU_store_int16(st0_ptr, st0_tag, (short *)data_address);
break;
case 014: /* fstp m32real */
clear_C1();
if ( reg_store_single((float *)data_address, st0_ptr) )
if ( FPU_store_single(st0_ptr, st0_tag, (float *)data_address) )
pop_0(); /* pop only if the number was actually stored
(see the 80486 manual p16-28) */
break;
case 015: /* fistp m32int */
clear_C1();
if ( reg_store_int32((long *)data_address, st0_ptr) )
if ( FPU_store_int32(st0_ptr, st0_tag, (long *)data_address) )
pop_0(); /* pop only if the number was actually stored
(see the 80486 manual p16-28) */
break;
case 016: /* fstp m64real */
clear_C1();
if ( reg_store_double((double *)data_address, st0_ptr) )
if ( FPU_store_double(st0_ptr, st0_tag, (double *)data_address) )
pop_0(); /* pop only if the number was actually stored
(see the 80486 manual p16-28) */
break;
case 017: /* fistp m16int */
clear_C1();
if ( reg_store_int16((short *)data_address, st0_ptr) )
if ( FPU_store_int16(st0_ptr, st0_tag, (short *)data_address) )
pop_0(); /* pop only if the number was actually stored
(see the 80486 manual p16-28) */
break;
case 020: /* fldenv m14/28byte */
fldenv(addr_modes, (char *)data_address);
fldenv(addr_modes, (u_char *)data_address);
/* Ensure that the values just loaded are not changed by
fix-up operations. */
return 1;
case 022: /* frstor m94/108byte */
frstor(addr_modes, (char *)data_address);
frstor(addr_modes, (u_char *)data_address);
/* Ensure that the values just loaded are not changed by
fix-up operations. */
return 1;
case 023: /* fbld m80dec */
clear_C1();
reg_load_bcd((char *)data_address, st0_ptr);
loaded_tag = FPU_load_bcd((u_char *)data_address);
FPU_settag0(loaded_tag);
break;
case 024: /* fldcw */
RE_ENTRANT_CHECK_OFF;
FPU_verify_area(VERIFY_READ, data_address, 2);
get_user(control_word, (unsigned short *) data_address);
FPU_get_user(control_word, (unsigned short *) data_address);
RE_ENTRANT_CHECK_ON;
if ( partial_status & ~control_word & CW_Exceptions )
partial_status |= (SW_Summary | SW_Backward);
......@@ -213,45 +221,47 @@ int load_store_instr(unsigned char type, fpu_addr_modes addr_modes,
return 1;
case 025: /* fld m80real */
clear_C1();
reg_load_extended((long double *)data_address, st0_ptr);
loaded_tag = FPU_load_extended((long double *)data_address, 0);
FPU_settag0(loaded_tag);
break;
case 027: /* fild m64int */
clear_C1();
reg_load_int64((long long *)data_address, st0_ptr);
loaded_tag = FPU_load_int64((long long *)data_address);
FPU_settag0(loaded_tag);
break;
case 030: /* fstenv m14/28byte */
fstenv(addr_modes, (char *)data_address);
fstenv(addr_modes, (u_char *)data_address);
return 1;
case 032: /* fsave */
fsave(addr_modes, (char *)data_address);
fsave(addr_modes, (u_char *)data_address);
return 1;
case 033: /* fbstp m80dec */
clear_C1();
if ( reg_store_bcd((char *)data_address, st0_ptr) )
if ( FPU_store_bcd(st0_ptr, st0_tag, (u_char *)data_address) )
pop_0(); /* pop only if the number was actually stored
(see the 80486 manual p16-28) */
break;
case 034: /* fstcw m16int */
RE_ENTRANT_CHECK_OFF;
FPU_verify_area(VERIFY_WRITE,data_address,2);
put_user(control_word, (unsigned short *) data_address);
FPU_put_user(control_word, (unsigned short *) data_address);
RE_ENTRANT_CHECK_ON;
return 1;
case 035: /* fstp m80real */
clear_C1();
if ( reg_store_extended((long double *)data_address, st0_ptr) )
if ( FPU_store_extended(st0_ptr, st0_tag, (long double *)data_address) )
pop_0(); /* pop only if the number was actually stored
(see the 80486 manual p16-28) */
break;
case 036: /* fstsw m2byte */
RE_ENTRANT_CHECK_OFF;
FPU_verify_area(VERIFY_WRITE,data_address,2);
put_user(status_word(),(unsigned short *) data_address);
FPU_put_user(status_word(),(unsigned short *) data_address);
RE_ENTRANT_CHECK_ON;
return 1;
case 037: /* fistp m64int */
clear_C1();
if ( reg_store_int64((long long *)data_address, st0_ptr) )
if ( FPU_store_int64(st0_ptr, st0_tag, (long long *)data_address) )
pop_0(); /* pop only if the number was actually stored
(see the 80486 manual p16-28) */
break;
......
arch/i386/math-emu/poly_2xm1.c
View file @ 97ccc379
......@@ -3,9 +3,9 @@
| |
| Function to compute 2^x-1 by a polynomial approximation. |
| |
| Copyright (C) 1992,1993,1994 |
| W. Metzenthen, 22 Parker St, Ormond, Vic 3163, |
| Australia. E-mail billm@vaxc.cc.monash.edu.au |
| Copyright (C) 1992,1993,1994,1997 |
| W. Metzenthen, 22 Parker St, Ormond, Vic 3163, Australia |
| E-mail billm@suburbia.net |
| |
| |
+---------------------------------------------------------------------------*/
......@@ -13,6 +13,7 @@
#include "exception.h"
#include "reg_constant.h"
#include "fpu_emu.h"
#include "fpu_system.h"
#include "control_w.h"
#include "poly.h"
......@@ -48,20 +49,19 @@ static const Xsig *shiftterm[] = { &shiftterm0, &shiftterm1,
/*--- poly_2xm1() -----------------------------------------------------------+
| Requires an argument which is TW_Valid and < 1. |
| Requires st(0) which is TAG_Valid and < 1. |
+---------------------------------------------------------------------------*/
int poly_2xm1(FPU_REG const *arg, FPU_REG *result)
int poly_2xm1(u_char sign, FPU_REG *arg, FPU_REG *result)
{
long int exponent, shift;
unsigned long long Xll;
Xsig accumulator, Denom, argSignif;
long int exponent, shift;
unsigned long long Xll;
Xsig accumulator, Denom, argSignif;
u_char tag;
exponent = arg->exp - EXP_BIAS;
exponent = exponent16(arg);
#ifdef PARANOID
if ( (exponent >= 0) /* Don't want a |number| >= 1.0 */
|| (arg->tag != TW_Valid) )
if ( exponent >= 0 ) /* Don't want a |number| >= 1.0 */
{
/* Number negative, too large, or not Valid. */
EXCEPTION(EX_INTERNAL|0x127);
......@@ -94,7 +94,7 @@ int poly_2xm1(FPU_REG const *arg, FPU_REG *result)
if ( exponent < -2 )
{
/* Shift the argument right by the required places. */
if ( shrx(&Xll, -2-exponent) >= 0x80000000U )
if ( FPU_shrx(&Xll, -2-exponent) >= 0x80000000U )
Xll++; /* round up */
}
......@@ -118,7 +118,7 @@ int poly_2xm1(FPU_REG const *arg, FPU_REG *result)
exponent = 1;
}
if ( arg->sign != SIGN_POS )
if ( sign != SIGN_POS )
{
/* The argument is negative, use the identity:
f(-x) = -f(x) / (1 + f(x))
......@@ -142,10 +142,14 @@ int poly_2xm1(FPU_REG const *arg, FPU_REG *result)
/* Convert to 64 bit signed-compatible */
exponent += round_Xsig(&accumulator);
result = &st(0);
significand(result) = XSIG_LL(accumulator);
result->tag = TW_Valid;
result->exp = exponent + EXP_BIAS;
result->sign = arg->sign;
setexponent16(result, exponent);
tag = FPU_round(result, 1, 0, FULL_PRECISION, sign);
setsign(result, sign);
FPU_settag0(tag);
return 0;
......
arch/i386/math-emu/poly_atan.c
View file @ 97ccc379
......@@ -3,9 +3,9 @@
| |
| Compute the arctan of a FPU_REG, using a polynomial approximation. |
| |
| Copyright (C) 1992,1993,1994 |
| W. Metzenthen, 22 Parker St, Ormond, Vic 3163, |
| Australia. E-mail billm@vaxc.cc.monash.edu.au |
| Copyright (C) 1992,1993,1994,1997 |
| W. Metzenthen, 22 Parker St, Ormond, Vic 3163, Australia |
| E-mail billm@suburbia.net |
| |
| |
+---------------------------------------------------------------------------*/
......@@ -13,6 +13,7 @@
#include "exception.h"
#include "reg_constant.h"
#include "fpu_emu.h"
#include "fpu_system.h"
#include "status_w.h"
#include "control_w.h"
#include "poly.h"
......@@ -51,31 +52,57 @@ static const Xsig pi_signif = MK_XSIG(0xc90fdaa2, 0x2168c234, 0xc4c6628b);
/*--- poly_atan() -----------------------------------------------------------+
| |
+---------------------------------------------------------------------------*/
void poly_atan(FPU_REG *arg1, FPU_REG *arg2, FPU_REG *result)
void poly_atan(FPU_REG *st0_ptr, u_char st0_tag,
FPU_REG *st1_ptr, u_char st1_tag)
{
char transformed, inverted,
sign1 = arg1->sign, sign2 = arg2->sign;
long int exponent, dummy_exp;
Xsig accumulator, Numer, Denom, accumulatore, argSignif,
argSq, argSqSq;
u_char transformed, inverted,
sign1, sign2;
int exponent;
long int dummy_exp;
Xsig accumulator, Numer, Denom, accumulatore, argSignif,
argSq, argSqSq;
u_char tag;
sign1 = getsign(st0_ptr);
sign2 = getsign(st1_ptr);
if ( st0_tag == TAG_Valid )
{
exponent = exponent(st0_ptr);
}
else
{
/* This gives non-compatible stack contents... */
FPU_to_exp16(st0_ptr, st0_ptr);
exponent = exponent16(st0_ptr);
}
if ( st1_tag == TAG_Valid )
{
exponent -= exponent(st1_ptr);
}
else
{
/* This gives non-compatible stack contents... */
FPU_to_exp16(st1_ptr, st1_ptr);
exponent -= exponent16(st1_ptr);
}
arg1->sign = arg2->sign = SIGN_POS;
if ( (compare(arg2) & ~COMP_Denormal) == COMP_A_lt_B )
if ( (exponent < 0) || ((exponent == 0) &&
((st0_ptr->sigh < st1_ptr->sigh) ||
((st0_ptr->sigh == st1_ptr->sigh) &&
(st0_ptr->sigl < st1_ptr->sigl))) ) )
{
inverted = 1;
exponent = arg1->exp - arg2->exp;
Numer.lsw = Denom.lsw = 0;
XSIG_LL(Numer) = significand(arg1);
XSIG_LL(Denom) = significand(arg2);
XSIG_LL(Numer) = significand(st0_ptr);
XSIG_LL(Denom) = significand(st1_ptr);
}
else
{
inverted = 0;
exponent = arg2->exp - arg1->exp;
exponent = -exponent;
Numer.lsw = Denom.lsw = 0;
XSIG_LL(Numer) = significand(arg2);
XSIG_LL(Denom) = significand(arg1);
XSIG_LL(Numer) = significand(st1_ptr);
XSIG_LL(Denom) = significand(st0_ptr);
}
div_Xsig(&Numer, &Denom, &argSignif);
exponent += norm_Xsig(&argSignif);
......@@ -189,9 +216,14 @@ void poly_atan(FPU_REG *arg1, FPU_REG *arg2, FPU_REG *result)
}
exponent += round_Xsig(&accumulator);
significand(result) = XSIG_LL(accumulator);
result->exp = exponent + EXP_BIAS;
result->tag = TW_Valid;
result->sign = sign2;
significand(st1_ptr) = XSIG_LL(accumulator);
setexponent16(st1_ptr, exponent);
tag = FPU_round(st1_ptr, 1, 0, FULL_PRECISION, sign2);
FPU_settagi(1, tag);
set_precision_flag_up(); /* We do not really know if up or down,
use this as the default. */
}
arch/i386/math-emu/poly_l2.c
View file @ 97ccc379
......@@ -3,9 +3,9 @@
| |
| Compute the base 2 log of a FPU_REG, using a polynomial approximation. |
| |
| Copyright (C) 1992,1993,1994 |
| W. Metzenthen, 22 Parker St, Ormond, Vic 3163, |
| Australia. E-mail billm@vaxc.cc.monash.edu.au |
| Copyright (C) 1992,1993,1994,1997 |
| W. Metzenthen, 22 Parker St, Ormond, Vic 3163, Australia |
| E-mail billm@suburbia.net |
| |
| |
+---------------------------------------------------------------------------*/
......@@ -14,96 +14,101 @@
#include "exception.h"
#include "reg_constant.h"
#include "fpu_emu.h"
#include "fpu_system.h"
#include "control_w.h"
#include "poly.h"
static void log2_kernel(FPU_REG const *arg,
static void log2_kernel(FPU_REG const *arg, u_char argsign,
Xsig *accum_result, long int *expon);
/*--- poly_l2() -------------------------------------------------------------+
| Base 2 logarithm by a polynomial approximation. |
+---------------------------------------------------------------------------*/
void poly_l2(FPU_REG const *arg, FPU_REG const *y, FPU_REG *result)
void poly_l2(FPU_REG *st0_ptr, FPU_REG *st1_ptr, u_char st1_sign)
{
long int exponent, expon, expon_expon;
Xsig accumulator, expon_accum, yaccum;
char sign;
u_char sign, argsign;
FPU_REG x;
int tag;
exponent = exponent16(st0_ptr);
exponent = arg->exp - EXP_BIAS;
/* From arg, make a number > sqrt(2)/2 and < sqrt(2) */
if ( arg->sigh > (unsigned)0xb504f334 )
/* From st0_ptr, make a number > sqrt(2)/2 and < sqrt(2) */
if ( st0_ptr->sigh > (unsigned)0xb504f334 )
{
/* Treat as sqrt(2)/2 < arg < 1 */
significand(&x) = - significand(arg);
x.sign = SIGN_NEG;
x.tag = TW_Valid;
x.exp = EXP_BIAS-1;
/* Treat as sqrt(2)/2 < st0_ptr < 1 */
significand(&x) = - significand(st0_ptr);
setexponent16(&x, -1);
exponent++;
normalize(&x);
argsign = SIGN_NEG;
}
else
{
/* Treat as 1 <= arg < sqrt(2) */
x.sigh = arg->sigh - 0x80000000;
x.sigl = arg->sigl;
x.sign = SIGN_POS;
x.tag = TW_Valid;
x.exp = EXP_BIAS;
normalize(&x);
/* Treat as 1 <= st0_ptr < sqrt(2) */
x.sigh = st0_ptr->sigh - 0x80000000;
x.sigl = st0_ptr->sigl;
setexponent16(&x, 0);
argsign = SIGN_POS;
}
tag = FPU_normalize_nuo(&x);
if ( x.tag == TW_Zero )
if ( tag == TAG_Zero )
{
expon = 0;
accumulator.msw = accumulator.midw = accumulator.lsw = 0;
}
else
{
log2_kernel(&x, &accumulator, &expon);
log2_kernel(&x, argsign, &accumulator, &expon);
}
sign = exponent < 0;
if ( sign ) exponent = -exponent;
if ( exponent < 0 )
{
sign = SIGN_NEG;
exponent = -exponent;
}
else
sign = SIGN_POS;
expon_accum.msw = exponent; expon_accum.midw = expon_accum.lsw = 0;
if ( exponent )
{
expon_expon = 31 + norm_Xsig(&expon_accum);
shr_Xsig(&accumulator, expon_expon - expon);
if ( sign ^ (x.sign == SIGN_NEG) )
if ( sign ^ argsign )
negate_Xsig(&accumulator);
add_Xsig_Xsig(&accumulator, &expon_accum);
}
else
{
expon_expon = expon;
sign = x.sign;
sign = argsign;
}
yaccum.lsw = 0; XSIG_LL(yaccum) = significand(y);
yaccum.lsw = 0; XSIG_LL(yaccum) = significand(st1_ptr);
mul_Xsig_Xsig(&accumulator, &yaccum);
expon_expon += round_Xsig(&accumulator);
if ( accumulator.msw == 0 )
{
reg_move(&CONST_Z, y);
}
else
{
result->exp = expon_expon + y->exp + 1;
significand(result) = XSIG_LL(accumulator);
result->tag = TW_Valid; /* set the tags to Valid */
result->sign = sign ^ y->sign;
FPU_copy_to_reg1(&CONST_Z, TAG_Zero);
return;
}
significand(st1_ptr) = XSIG_LL(accumulator);
setexponent16(st1_ptr, expon_expon + exponent16(st1_ptr) + 1);
tag = FPU_round(st1_ptr, 1, 0, FULL_PRECISION, sign ^ st1_sign);
FPU_settagi(1, tag);
set_precision_flag_up(); /* 80486 appears to always do this */
return;
}
......@@ -111,47 +116,62 @@ void poly_l2(FPU_REG const *arg, FPU_REG const *y, FPU_REG *result)
| Base 2 logarithm by a polynomial approximation. |
| log2(x+1) |
+---------------------------------------------------------------------------*/
int poly_l2p1(FPU_REG const *arg, FPU_REG const *y, FPU_REG *result)
int poly_l2p1(u_char sign0, u_char sign1,
FPU_REG *st0_ptr, FPU_REG *st1_ptr, FPU_REG *dest)
{
char sign;
long int exponent;
Xsig accumulator, yaccum;
u_char tag;
long int exponent;
Xsig accumulator, yaccum;
sign = arg->sign;
if ( arg->exp < EXP_BIAS )
if ( exponent16(st0_ptr) < 0 )
{
log2_kernel(arg, &accumulator, &exponent);
log2_kernel(st0_ptr, sign0, &accumulator, &exponent);
yaccum.lsw = 0;
XSIG_LL(yaccum) = significand(y);
XSIG_LL(yaccum) = significand(st1_ptr);
mul_Xsig_Xsig(&accumulator, &yaccum);
exponent += round_Xsig(&accumulator);
result->exp = exponent + y->exp + 1;
significand(result) = XSIG_LL(accumulator);
result->tag = TW_Valid; /* set the tags to Valid */
result->sign = sign ^ y->sign;
exponent += exponent16(st1_ptr) + 1;
if ( exponent < EXP_WAY_UNDER ) exponent = EXP_WAY_UNDER;
significand(dest) = XSIG_LL(accumulator);
setexponent16(dest, exponent);
return 0;
tag = FPU_round(dest, 1, 0, FULL_PRECISION, sign0 ^ sign1);
FPU_settagi(1, tag);
if ( tag == TAG_Valid )
set_precision_flag_up(); /* 80486 appears to always do this */
}
else
{
/* The magnitude of arg is far too large. */
reg_move(y, result);
if ( sign != SIGN_POS )
/* The magnitude of st0_ptr is far too large. */
if ( sign0 != SIGN_POS )
{
/* Trying to get the log of a negative number. */
return 1;
#ifdef PECULIAR_486 /* Stupid 80486 doesn't worry about log(negative). */
changesign(st1_ptr);
#else
if ( arith_invalid(1) < 0 )
return 1;
#endif PECULIAR_486
}
/* 80486 appears to do this */
if ( sign0 == SIGN_NEG )
set_precision_flag_down();
else
{
return 0;
}
set_precision_flag_up();
}
if ( exponent(dest) <= EXP_UNDER )
EXCEPTION(EX_Underflow);
return 0;
}
......@@ -180,20 +200,17 @@ static const unsigned long leadterm = 0xb8000000;
| Base 2 logarithm by a polynomial approximation. |
| log2(x+1) |
+---------------------------------------------------------------------------*/
static void log2_kernel(FPU_REG const *arg, Xsig *accum_result,
static void log2_kernel(FPU_REG const *arg, u_char argsign, Xsig *accum_result,
long int *expon)
{
char sign;
long int exponent, adj;
unsigned long long Xsq;
Xsig accumulator, Numer, Denom, argSignif, arg_signif;
sign = arg->sign;
exponent = arg->exp - EXP_BIAS;
exponent = exponent16(arg);
Numer.lsw = Denom.lsw = 0;
XSIG_LL(Numer) = XSIG_LL(Denom) = significand(arg);
if ( sign == SIGN_POS )
if ( argsign == SIGN_POS )
{
shr_Xsig(&Denom, 2 - (1 + exponent));
Denom.msw |= 0x80000000;
......
arch/i386/math-emu/poly_sin.c
View file @ 97ccc379
......@@ -4,9 +4,9 @@
| Computation of an approximation of the sin function and the cosine |
| function by a polynomial. |
| |
| Copyright (C) 1992,1993,1994 |
| W. Metzenthen, 22 Parker St, Ormond, Vic 3163, |
| Australia. E-mail billm@vaxc.cc.monash.edu.au |
| Copyright (C) 1992,1993,1994,1997 |
| W. Metzenthen, 22 Parker St, Ormond, Vic 3163, Australia |
| E-mail billm@suburbia.net |
| |
| |
+---------------------------------------------------------------------------*/
......@@ -15,6 +15,7 @@
#include "exception.h"
#include "reg_constant.h"
#include "fpu_emu.h"
#include "fpu_system.h"
#include "control_w.h"
#include "poly.h"
......@@ -62,35 +63,26 @@ static const unsigned long long neg_terms_h[N_COEFF_NH] =
/*--- poly_sine() -----------------------------------------------------------+
| |
+---------------------------------------------------------------------------*/
void poly_sine(FPU_REG const *arg, FPU_REG *result)
void poly_sine(FPU_REG *st0_ptr)
{
int exponent, echange;
Xsig accumulator, argSqrd, argTo4;
unsigned long fix_up, adj;
unsigned long long fixed_arg;
FPU_REG result;
#ifdef PARANOID
if ( arg->tag == TW_Zero )
{
/* Return 0.0 */
reg_move(&CONST_Z, result);
return;
}
#endif PARANOID
exponent = arg->exp - EXP_BIAS;
exponent = exponent(st0_ptr);
accumulator.lsw = accumulator.midw = accumulator.msw = 0;
/* Split into two ranges, for arguments below and above 1.0 */
/* The boundary between upper and lower is approx 0.88309101259 */
if ( (exponent < -1) || ((exponent == -1) && (arg->sigh <= 0xe21240aa)) )
if ( (exponent < -1) || ((exponent == -1) && (st0_ptr->sigh <= 0xe21240aa)) )
{
/* The argument is <= 0.88309101259 */
argSqrd.msw = arg->sigh; argSqrd.midw = arg->sigl; argSqrd.lsw = 0;
mul64_Xsig(&argSqrd, &significand(arg));
argSqrd.msw = st0_ptr->sigh; argSqrd.midw = st0_ptr->sigl; argSqrd.lsw = 0;
mul64_Xsig(&argSqrd, &significand(st0_ptr));
shr_Xsig(&argSqrd, 2*(-1-exponent));
argTo4.msw = argSqrd.msw; argTo4.midw = argSqrd.midw;
argTo4.lsw = argSqrd.lsw;
......@@ -107,29 +99,29 @@ void poly_sine(FPU_REG const *arg, FPU_REG *result)
shr_Xsig(&accumulator, 2); /* Divide by four */
accumulator.msw |= 0x80000000; /* Add 1.0 */
mul64_Xsig(&accumulator, &significand(arg));
mul64_Xsig(&accumulator, &significand(arg));
mul64_Xsig(&accumulator, &significand(arg));
mul64_Xsig(&accumulator, &significand(st0_ptr));
mul64_Xsig(&accumulator, &significand(st0_ptr));
mul64_Xsig(&accumulator, &significand(st0_ptr));
/* Divide by four, FPU_REG compatible, etc */
exponent = 3*exponent + EXP_BIAS;
exponent = 3*exponent;
/* The minimum exponent difference is 3 */
shr_Xsig(&accumulator, arg->exp - exponent);
shr_Xsig(&accumulator, exponent(st0_ptr) - exponent);
negate_Xsig(&accumulator);
XSIG_LL(accumulator) += significand(arg);
XSIG_LL(accumulator) += significand(st0_ptr);
echange = round_Xsig(&accumulator);
result->exp = arg->exp + echange;
setexponentpos(&result, exponent(st0_ptr) + echange);
}
else
{
/* The argument is > 0.88309101259 */
/* We use sin(arg) = cos(pi/2-arg) */
/* We use sin(st(0)) = cos(pi/2-st(0)) */
fixed_arg = significand(arg);
fixed_arg = significand(st0_ptr);
if ( exponent == 0 )
{
......@@ -192,16 +184,16 @@ void poly_sine(FPU_REG const *arg, FPU_REG *result)
echange = round_Xsig(&accumulator);
result->exp = EXP_BIAS - 1 + echange;
setexponentpos(&result, echange - 1);
}
significand(result) = XSIG_LL(accumulator);
result->tag = TW_Valid;
result->sign = arg->sign;
significand(&result) = XSIG_LL(accumulator);
setsign(&result, getsign(st0_ptr));
FPU_copy_to_reg0(&result, TAG_Valid);
#ifdef PARANOID
if ( (result->exp >= EXP_BIAS)
&& (significand(result) > 0x8000000000000000LL) )
if ( (exponent(&result) >= 0)
&& (significand(&result) > 0x8000000000000000LL) )
{
EXCEPTION(EX_INTERNAL|0x150);
}
......@@ -214,42 +206,36 @@ void poly_sine(FPU_REG const *arg, FPU_REG *result)
/*--- poly_cos() ------------------------------------------------------------+
| |
+---------------------------------------------------------------------------*/
void poly_cos(FPU_REG const *arg, FPU_REG *result)
void poly_cos(FPU_REG *st0_ptr)
{
FPU_REG result;
long int exponent, exp2, echange;
Xsig accumulator, argSqrd, fix_up, argTo4;
unsigned long adj;
unsigned long long fixed_arg;
#ifdef PARANOID
if ( arg->tag == TW_Zero )
{
/* Return 1.0 */
reg_move(&CONST_1, result);
return;
}
if ( (arg->exp > EXP_BIAS)
|| ((arg->exp == EXP_BIAS)
&& (significand(arg) > 0xc90fdaa22168c234LL)) )
if ( (exponent(st0_ptr) > 0)
|| ((exponent(st0_ptr) == 0)
&& (significand(st0_ptr) > 0xc90fdaa22168c234LL)) )
{
EXCEPTION(EX_Invalid);
reg_move(&CONST_QNaN, result);
FPU_copy_to_reg0(&CONST_QNaN, TAG_Special);
return;
}
#endif PARANOID
exponent = arg->exp - EXP_BIAS;
exponent = exponent(st0_ptr);
accumulator.lsw = accumulator.midw = accumulator.msw = 0;
if ( (exponent < -1) || ((exponent == -1) && (arg->sigh <= 0xb00d6f54)) )
if ( (exponent < -1) || ((exponent == -1) && (st0_ptr->sigh <= 0xb00d6f54)) )
{
/* arg is < 0.687705 */
argSqrd.msw = arg->sigh; argSqrd.midw = arg->sigl; argSqrd.lsw = 0;
mul64_Xsig(&argSqrd, &significand(arg));
argSqrd.msw = st0_ptr->sigh; argSqrd.midw = st0_ptr->sigl;
argSqrd.lsw = 0;
mul64_Xsig(&argSqrd, &significand(st0_ptr));
if ( exponent < -1 )
{
......@@ -270,8 +256,8 @@ void poly_cos(FPU_REG const *arg, FPU_REG *result)
N_COEFF_PH-1);
negate_Xsig(&accumulator);
mul64_Xsig(&accumulator, &significand(arg));
mul64_Xsig(&accumulator, &significand(arg));
mul64_Xsig(&accumulator, &significand(st0_ptr));
mul64_Xsig(&accumulator, &significand(st0_ptr));
shr_Xsig(&accumulator, -2*(1+exponent));
shr_Xsig(&accumulator, 3);
......@@ -290,20 +276,20 @@ void poly_cos(FPU_REG const *arg, FPU_REG *result)
if ( accumulator.msw == 0 )
{
/* The result is 1.0 */
reg_move(&CONST_1, result);
FPU_copy_to_reg0(&CONST_1, TAG_Valid);
return;
}
else
{
significand(result) = XSIG_LL(accumulator);
significand(&result) = XSIG_LL(accumulator);
/* will be a valid positive nr with expon = -1 */
*(short *)&(result->sign) = 0;
result->exp = EXP_BIAS - 1;
setexponentpos(&result, -1);
}
}
else
{
fixed_arg = significand(arg);
fixed_arg = significand(st0_ptr);
if ( exponent == 0 )
{
......@@ -392,14 +378,15 @@ void poly_cos(FPU_REG const *arg, FPU_REG *result)
echange = round_Xsig(&accumulator);
result->exp = exp2 + EXP_BIAS + echange;
*(short *)&(result->sign) = 0; /* Is a valid positive nr */
significand(result) = XSIG_LL(accumulator);
setexponentpos(&result, exp2 + echange);
significand(&result) = XSIG_LL(accumulator);
}
FPU_copy_to_reg0(&result, TAG_Valid);
#ifdef PARANOID
if ( (result->exp >= EXP_BIAS)
&& (significand(result) > 0x8000000000000000LL) )
if ( (exponent(&result) >= 0)
&& (significand(&result) > 0x8000000000000000LL) )
{
EXCEPTION(EX_INTERNAL|0x151);
}
......
arch/i386/math-emu/poly_tan.c
View file @ 97ccc379
......@@ -3,9 +3,9 @@
| |
| Compute the tan of a FPU_REG, using a polynomial approximation. |
| |
| Copyright (C) 1992,1993,1994 |
| Copyright (C) 1992,1993,1994,1997 |
| W. Metzenthen, 22 Parker St, Ormond, Vic 3163, |
| Australia. E-mail billm@vaxc.cc.monash.edu.au |
| Australia. E-mail billm@suburbia.net |
| |
| |
+---------------------------------------------------------------------------*/
......@@ -13,6 +13,7 @@
#include "exception.h"
#include "reg_constant.h"
#include "fpu_emu.h"
#include "fpu_system.h"
#include "control_w.h"
#include "poly.h"
......@@ -52,7 +53,7 @@ static const unsigned long long twothirds = 0xaaaaaaaaaaaaaaabLL;
/*--- poly_tan() ------------------------------------------------------------+
| |
+---------------------------------------------------------------------------*/
void poly_tan(FPU_REG const *arg, FPU_REG *result)
void poly_tan(FPU_REG *st0_ptr)
{
long int exponent;
int invert;
......@@ -60,20 +61,20 @@ void poly_tan(FPU_REG const *arg, FPU_REG *result)
argSignif, fix_up;
unsigned long adj;
exponent = arg->exp - EXP_BIAS;
exponent = exponent(st0_ptr);
#ifdef PARANOID
if ( arg->sign != 0 ) /* Can't hack a number < 0.0 */
{ arith_invalid(result); return; } /* Need a positive number */
if ( signnegative(st0_ptr) ) /* Can't hack a number < 0.0 */
{ arith_invalid(0); return; } /* Need a positive number */
#endif PARANOID
/* Split the problem into two domains, smaller and larger than pi/4 */
if ( (exponent == 0) || ((exponent == -1) && (arg->sigh > 0xc90fdaa2)) )
if ( (exponent == 0) || ((exponent == -1) && (st0_ptr->sigh > 0xc90fdaa2)) )
{
/* The argument is greater than (approx) pi/4 */
invert = 1;
accum.lsw = 0;
XSIG_LL(accum) = significand(arg);
XSIG_LL(accum) = significand(st0_ptr);
if ( exponent == 0 )
{
......@@ -92,12 +93,12 @@ void poly_tan(FPU_REG const *arg, FPU_REG *result)
{
invert = 0;
argSignif.lsw = 0;
XSIG_LL(accum) = XSIG_LL(argSignif) = significand(arg);
XSIG_LL(accum) = XSIG_LL(argSignif) = significand(st0_ptr);
if ( exponent < -1 )
{
/* shift the argument right by the required places */
if ( shrx(&XSIG_LL(accum), -1-exponent) >= 0x80000000U )
if ( FPU_shrx(&XSIG_LL(accum), -1-exponent) >= 0x80000000U )
XSIG_LL(accum) ++; /* round up */
}
}
......@@ -206,8 +207,8 @@ void poly_tan(FPU_REG const *arg, FPU_REG *result)
/* Transfer the result */
round_Xsig(&accum);
*(short *)&(result->sign) = 0;
significand(result) = XSIG_LL(accum);
result->exp = EXP_BIAS + exponent;
FPU_settag0(TAG_Valid);
significand(st0_ptr) = XSIG_LL(accum);
setexponent16(st0_ptr, exponent + EXTENDED_Ebias); /* Result is positive. */
}
arch/i386/math-emu/reg_add_sub.c
View file @ 97ccc379
This diff is collapsed.
arch/i386/math-emu/reg_compare.c
View file @ 97ccc379
......@@ -3,9 +3,9 @@
| |
| Compare two floating point registers |
| |
| Copyright (C) 1992,1993,1994 |
| W. Metzenthen, 22 Parker St, Ormond, Vic 3163, |
| Australia. E-mail billm@vaxc.cc.monash.edu.au |
| Copyright (C) 1992,1993,1994,1997 |
| W. Metzenthen, 22 Parker St, Ormond, Vic 3163, Australia |
| E-mail billm@suburbia.net |
| |
| |
+---------------------------------------------------------------------------*/
......@@ -21,86 +21,87 @@
#include "status_w.h"
int compare(FPU_REG const *b)
static int compare(FPU_REG const *b, int tagb)
{
int diff;
char st0_tag;
FPU_REG *st0_ptr;
int diff, exp0, expb;
u_char st0_tag;
FPU_REG *st0_ptr;
FPU_REG x, y;
u_char st0_sign, signb = getsign(b);
st0_ptr = &st(0);
st0_tag = st0_ptr->tag;
st0_tag = FPU_gettag0();
st0_sign = getsign(st0_ptr);
if ( st0_tag | b->tag )
if ( tagb == TAG_Special )
tagb = FPU_Special(b);
if ( st0_tag == TAG_Special )
st0_tag = FPU_Special(st0_ptr);
if ( ((st0_tag != TAG_Valid) && (st0_tag != TW_Denormal))
|| ((tagb != TAG_Valid) && (tagb != TW_Denormal)) )
{
if ( st0_tag == TW_Zero )
if ( st0_tag == TAG_Zero )
{
if ( b->tag == TW_Zero ) return COMP_A_eq_B;
if ( b->tag == TW_Valid )
{
return ((b->sign == SIGN_POS) ? COMP_A_lt_B : COMP_A_gt_B)
#ifdef DENORM_OPERAND
| ((b->exp <= EXP_UNDER) ?
COMP_Denormal : 0)
#endif DENORM_OPERAND
;
}
if ( tagb == TAG_Zero ) return COMP_A_eq_B;
if ( tagb == TAG_Valid )
return ((signb == SIGN_POS) ? COMP_A_lt_B : COMP_A_gt_B);
if ( tagb == TW_Denormal )
return ((signb == SIGN_POS) ? COMP_A_lt_B : COMP_A_gt_B)
| COMP_Denormal;
}
else if ( b->tag == TW_Zero )
else if ( tagb == TAG_Zero )
{
if ( st0_tag == TW_Valid )
{
return ((st0_ptr->sign == SIGN_POS) ? COMP_A_gt_B
: COMP_A_lt_B)
#ifdef DENORM_OPERAND
| ((st0_ptr->exp <= EXP_UNDER )
? COMP_Denormal : 0 )
#endif DENORM_OPERAND
;
}
if ( st0_tag == TAG_Valid )
return ((st0_sign == SIGN_POS) ? COMP_A_gt_B : COMP_A_lt_B);
if ( st0_tag == TW_Denormal )
return ((st0_sign == SIGN_POS) ? COMP_A_gt_B : COMP_A_lt_B)
| COMP_Denormal;
}
if ( st0_tag == TW_Infinity )
{
if ( (b->tag == TW_Valid) || (b->tag == TW_Zero) )
{
return ((st0_ptr->sign == SIGN_POS) ? COMP_A_gt_B
: COMP_A_lt_B)
#ifdef DENORM_OPERAND
| (((b->tag == TW_Valid) && (b->exp <= EXP_UNDER)) ?
COMP_Denormal : 0 )
#endif DENORM_OPERAND
;
}
else if ( b->tag == TW_Infinity )
if ( (tagb == TAG_Valid) || (tagb == TAG_Zero) )
return ((st0_sign == SIGN_POS) ? COMP_A_gt_B : COMP_A_lt_B);
else if ( tagb == TW_Denormal )
return ((st0_sign == SIGN_POS) ? COMP_A_gt_B : COMP_A_lt_B)
| COMP_Denormal;
else if ( tagb == TW_Infinity )
{
/* The 80486 book says that infinities can be equal! */
return (st0_ptr->sign == b->sign) ? COMP_A_eq_B :
((st0_ptr->sign == SIGN_POS) ? COMP_A_gt_B : COMP_A_lt_B);
return (st0_sign == signb) ? COMP_A_eq_B :
((st0_sign == SIGN_POS) ? COMP_A_gt_B : COMP_A_lt_B);
}
/* Fall through to the NaN code */
}
else if ( b->tag == TW_Infinity )
else if ( tagb == TW_Infinity )
{
if ( (st0_tag == TW_Valid) || (st0_tag == TW_Zero) )
{
return ((b->sign == SIGN_POS) ? COMP_A_lt_B : COMP_A_gt_B)
#ifdef DENORM_OPERAND
| (((st0_tag == TW_Valid)
&& (st0_ptr->exp <= EXP_UNDER)) ?
COMP_Denormal : 0)
#endif DENORM_OPERAND
;
}
if ( (st0_tag == TAG_Valid) || (st0_tag == TAG_Zero) )
return ((signb == SIGN_POS) ? COMP_A_lt_B : COMP_A_gt_B);
if ( st0_tag == TW_Denormal )
return ((signb == SIGN_POS) ? COMP_A_lt_B : COMP_A_gt_B)
| COMP_Denormal;
/* Fall through to the NaN code */
}
/* The only possibility now should be that one of the arguments
is a NaN */
if ( (st0_tag == TW_NaN) || (b->tag == TW_NaN) )
if ( (st0_tag == TW_NaN) || (tagb == TW_NaN) )
{
if ( ((st0_tag == TW_NaN) && !(st0_ptr->sigh & 0x40000000))
|| ((b->tag == TW_NaN) && !(b->sigh & 0x40000000)) )
/* At least one arg is a signaling NaN */
int signalling = 0, unsupported = 0;
if ( st0_tag == TW_NaN )
{
signalling = (st0_ptr->sigh & 0xc0000000) == 0x80000000;
unsupported = !((exponent(st0_ptr) == EXP_OVER)
&& (st0_ptr->sigh & 0x80000000));
}
if ( tagb == TW_NaN )
{
signalling |= (b->sigh & 0xc0000000) == 0x80000000;
unsupported |= !((exponent(b) == EXP_OVER)
&& (b->sigh & 0x80000000));
}
if ( signalling || unsupported )
return COMP_No_Comp | COMP_SNaN | COMP_NaN;
else
/* Neither is a signaling NaN */
......@@ -110,24 +111,34 @@ int compare(FPU_REG const *b)
EXCEPTION(EX_Invalid);
}
if (st0_sign != signb)
{
return ((st0_sign == SIGN_POS) ? COMP_A_gt_B : COMP_A_lt_B)
| ( ((st0_tag == TW_Denormal) || (tagb == TW_Denormal)) ?
COMP_Denormal : 0);
}
if ( (st0_tag == TW_Denormal) || (tagb == TW_Denormal) )
{
FPU_to_exp16(st0_ptr, &x);
FPU_to_exp16(b, &y);
st0_ptr = &x;
b = &y;
exp0 = exponent16(st0_ptr);
expb = exponent16(b);
}
else
{
exp0 = exponent(st0_ptr);
expb = exponent(b);
}
#ifdef PARANOID
if (!(st0_ptr->sigh & 0x80000000)) EXCEPTION(EX_Invalid);
if (!(b->sigh & 0x80000000)) EXCEPTION(EX_Invalid);
#endif PARANOID
if (st0_ptr->sign != b->sign)
{
return ((st0_ptr->sign == SIGN_POS) ? COMP_A_gt_B : COMP_A_lt_B)
#ifdef DENORM_OPERAND
|
( ((st0_ptr->exp <= EXP_UNDER) || (b->exp <= EXP_UNDER)) ?
COMP_Denormal : 0)
#endif DENORM_OPERAND
;
}
diff = st0_ptr->exp - b->exp;
diff = exp0 - expb;
if ( diff == 0 )
{
diff = st0_ptr->sigh - b->sigh; /* Works only if ms bits are
......@@ -142,42 +153,30 @@ int compare(FPU_REG const *b)
if ( diff > 0 )
{
return ((st0_ptr->sign == SIGN_POS) ? COMP_A_gt_B : COMP_A_lt_B)
#ifdef DENORM_OPERAND
|
( ((st0_ptr->exp <= EXP_UNDER) || (b->exp <= EXP_UNDER)) ?
COMP_Denormal : 0)
#endif DENORM_OPERAND
;
return ((st0_sign == SIGN_POS) ? COMP_A_gt_B : COMP_A_lt_B)
| ( ((st0_tag == TW_Denormal) || (tagb == TW_Denormal)) ?
COMP_Denormal : 0);
}
if ( diff < 0 )
{
return ((st0_ptr->sign == SIGN_POS) ? COMP_A_lt_B : COMP_A_gt_B)
#ifdef DENORM_OPERAND
|
( ((st0_ptr->exp <= EXP_UNDER) || (b->exp <= EXP_UNDER)) ?
COMP_Denormal : 0)
#endif DENORM_OPERAND
;
return ((st0_sign == SIGN_POS) ? COMP_A_lt_B : COMP_A_gt_B)
| ( ((st0_tag == TW_Denormal) || (tagb == TW_Denormal)) ?
COMP_Denormal : 0);
}
return COMP_A_eq_B
#ifdef DENORM_OPERAND
|
( ((st0_ptr->exp <= EXP_UNDER) || (b->exp <= EXP_UNDER)) ?
COMP_Denormal : 0)
#endif DENORM_OPERAND
;
| ( ((st0_tag == TW_Denormal) || (tagb == TW_Denormal)) ?
COMP_Denormal : 0);
}
/* This function requires that st(0) is not empty */
int compare_st_data(FPU_REG const *loaded_data)
int FPU_compare_st_data(FPU_REG const *loaded_data, u_char loaded_tag)
{
int f, c;
c = compare(loaded_data);
c = compare(loaded_data, loaded_tag);
if (c & COMP_NaN)
{
......@@ -209,7 +208,7 @@ int compare_st_data(FPU_REG const *loaded_data)
setcc(f);
if (c & COMP_Denormal)
{
return denormal_operand();
return denormal_operand() < 0;
}
return 0;
}
......@@ -218,6 +217,7 @@ int compare_st_data(FPU_REG const *loaded_data)
static int compare_st_st(int nr)
{
int f, c;
FPU_REG *st_ptr;
if ( !NOT_EMPTY(0) || !NOT_EMPTY(nr) )
{
......@@ -227,7 +227,8 @@ static int compare_st_st(int nr)
return !(control_word & CW_Invalid);
}
c = compare(&st(nr));
st_ptr = &st(nr);
c = compare(st_ptr, FPU_gettagi(nr));
if (c & COMP_NaN)
{
setcc(SW_C3 | SW_C2 | SW_C0);
......@@ -259,7 +260,7 @@ static int compare_st_st(int nr)
setcc(f);
if (c & COMP_Denormal)
{
return denormal_operand();
return denormal_operand() < 0;
}
return 0;
}
......@@ -268,6 +269,7 @@ static int compare_st_st(int nr)
static int compare_u_st_st(int nr)
{
int f, c;
FPU_REG *st_ptr;
if ( !NOT_EMPTY(0) || !NOT_EMPTY(nr) )
{
......@@ -277,7 +279,8 @@ static int compare_u_st_st(int nr)
return !(control_word & CW_Invalid);
}
c = compare(&st(nr));
st_ptr = &st(nr);
c = compare(st_ptr, FPU_gettagi(nr));
if (c & COMP_NaN)
{
setcc(SW_C3 | SW_C2 | SW_C0);
......@@ -314,7 +317,7 @@ static int compare_u_st_st(int nr)
setcc(f);
if (c & COMP_Denormal)
{
return denormal_operand();
return denormal_operand() < 0;
}
return 0;
}
......@@ -332,7 +335,7 @@ void fcompst()
{
/* fcomp st(i) */
if ( !compare_st_st(FPU_rm) )
pop();
FPU_pop();
}
......@@ -361,7 +364,7 @@ void fucomp()
{
/* fucomp st(i) */
if ( !compare_u_st_st(FPU_rm) )
pop();
FPU_pop();
}
......
arch/i386/math-emu/reg_constant.c
View file @ 97ccc379
......@@ -3,9 +3,9 @@
| |
| All of the constant FPU_REGs |
| |
| Copyright (C) 1992,1993,1994,1996 |
| Copyright (C) 1992,1993,1994,1997 |
| W. Metzenthen, 22 Parker St, Ormond, Vic 3163, |
| Australia. E-mail billm@jacobi.maths.monash.edu.au |
| Australia. E-mail billm@suburbia.net |
| |
| |
+---------------------------------------------------------------------------*/
......@@ -17,59 +17,52 @@
#include "control_w.h"
FPU_REG const CONST_1 = { SIGN_POS, TW_Valid, EXP_BIAS,
0x00000000, 0x80000000 };
FPU_REG const CONST_2 = { SIGN_POS, TW_Valid, EXP_BIAS+1,
0x00000000, 0x80000000 };
FPU_REG const CONST_HALF = { SIGN_POS, TW_Valid, EXP_BIAS-1,
0x00000000, 0x80000000 };
FPU_REG const CONST_L2T = { SIGN_POS, TW_Valid, EXP_BIAS+1,
0xcd1b8afe, 0xd49a784b };
FPU_REG const CONST_L2E = { SIGN_POS, TW_Valid, EXP_BIAS,
0x5c17f0bc, 0xb8aa3b29 };
FPU_REG const CONST_PI = { SIGN_POS, TW_Valid, EXP_BIAS+1,
0x2168c235, 0xc90fdaa2 };
FPU_REG const CONST_PI2 = { SIGN_POS, TW_Valid, EXP_BIAS,
0x2168c235, 0xc90fdaa2 };
FPU_REG const CONST_PI4 = { SIGN_POS, TW_Valid, EXP_BIAS-1,
0x2168c235, 0xc90fdaa2 };
FPU_REG const CONST_LG2 = { SIGN_POS, TW_Valid, EXP_BIAS-2,
0xfbcff799, 0x9a209a84 };
FPU_REG const CONST_LN2 = { SIGN_POS, TW_Valid, EXP_BIAS-1,
0xd1cf79ac, 0xb17217f7 };
#define MAKE_REG(s,e,l,h) { l, h, \
((EXTENDED_Ebias+(e)) | ((SIGN_##s != 0)*0x8000)) }
FPU_REG const CONST_1 = MAKE_REG(POS, 0, 0x00000000, 0x80000000);
FPU_REG const CONST_2 = MAKE_REG(POS, 1, 0x00000000, 0x80000000);
FPU_REG const CONST_HALF = MAKE_REG(POS, -1, 0x00000000, 0x80000000);
FPU_REG const CONST_L2T = MAKE_REG(POS, 1, 0xcd1b8afe, 0xd49a784b);
FPU_REG const CONST_L2E = MAKE_REG(POS, 0, 0x5c17f0bc, 0xb8aa3b29);
FPU_REG const CONST_PI = MAKE_REG(POS, 1, 0x2168c235, 0xc90fdaa2);
FPU_REG const CONST_PI2 = MAKE_REG(POS, 0, 0x2168c235, 0xc90fdaa2);
FPU_REG const CONST_PI4 = MAKE_REG(POS, -1, 0x2168c235, 0xc90fdaa2);
FPU_REG const CONST_LG2 = MAKE_REG(POS, -2, 0xfbcff799, 0x9a209a84);
FPU_REG const CONST_LN2 = MAKE_REG(POS, -1, 0xd1cf79ac, 0xb17217f7);
/* Extra bits to take pi/2 to more than 128 bits precision. */
FPU_REG const CONST_PI2extra = { SIGN_NEG, TW_Valid, EXP_BIAS-66,
0xfc8f8cbb, 0xece675d1 };
FPU_REG const CONST_PI2extra = MAKE_REG(NEG, -66,
0xfc8f8cbb, 0xece675d1);
/* Only the sign (and tag) is used in internal zeroes */
FPU_REG const CONST_Z = { SIGN_POS, TW_Zero, EXP_UNDER, 0x0, 0x0 };
FPU_REG const CONST_Z = MAKE_REG(POS, EXP_UNDER, 0x0, 0x0);
/* Only the sign and significand (and tag) are used in internal NaNs */
/* The 80486 never generates one of these
FPU_REG const CONST_SNAN = { SIGN_POS, TW_NaN, EXP_OVER, 0x00000001, 0x80000000 };
FPU_REG const CONST_SNAN = MAKE_REG(POS, EXP_OVER, 0x00000001, 0x80000000);
*/
/* This is the real indefinite QNaN */
FPU_REG const CONST_QNaN = { SIGN_NEG, TW_NaN, EXP_OVER, 0x00000000, 0xC0000000 };
FPU_REG const CONST_QNaN = MAKE_REG(NEG, EXP_OVER, 0x00000000, 0xC0000000);
/* Only the sign (and tag) is used in internal infinities */
FPU_REG const CONST_INF = { SIGN_POS, TW_Infinity, EXP_OVER, 0x00000000, 0x80000000 };
FPU_REG const CONST_INF = MAKE_REG(POS, EXP_OVER, 0x00000000, 0x80000000);
static void fld_const(FPU_REG const *c, int adj)
static void fld_const(FPU_REG const *c, int adj, u_char tag)
{
FPU_REG *st_new_ptr;
if ( STACK_OVERFLOW )
{
stack_overflow();
FPU_stack_overflow();
return;
}
push();
reg_move(c, st_new_ptr);
reg_copy(c, st_new_ptr);
st_new_ptr->sigl += adj; /* For all our fldxxx constants, we don't need to
borrow or carry. */
FPU_settag0(tag);
clear_C1();
}
......@@ -80,37 +73,37 @@ static void fld_const(FPU_REG const *c, int adj)
static void fld1(int rc)
{
fld_const(&CONST_1, 0);
fld_const(&CONST_1, 0, TAG_Valid);
}
static void fldl2t(int rc)
{
fld_const(&CONST_L2T, (rc == RC_UP) ? 1 : 0);
fld_const(&CONST_L2T, (rc == RC_UP) ? 1 : 0, TAG_Valid);
}
static void fldl2e(int rc)
{
fld_const(&CONST_L2E, DOWN_OR_CHOP(rc) ? -1 : 0);
fld_const(&CONST_L2E, DOWN_OR_CHOP(rc) ? -1 : 0, TAG_Valid);
}
static void fldpi(int rc)
{
fld_const(&CONST_PI, DOWN_OR_CHOP(rc) ? -1 : 0);
fld_const(&CONST_PI, DOWN_OR_CHOP(rc) ? -1 : 0, TAG_Valid);
}
static void fldlg2(int rc)
{
fld_const(&CONST_LG2, DOWN_OR_CHOP(rc) ? -1 : 0);
fld_const(&CONST_LG2, DOWN_OR_CHOP(rc) ? -1 : 0, TAG_Valid);
}
static void fldln2(int rc)
{
fld_const(&CONST_LN2, DOWN_OR_CHOP(rc) ? -1 : 0);
fld_const(&CONST_LN2, DOWN_OR_CHOP(rc) ? -1 : 0, TAG_Valid);
}
static void fldz(int rc)
{
fld_const(&CONST_Z, 0);
fld_const(&CONST_Z, 0, TAG_Zero);
}
typedef void (*FUNC_RC)(int);
......
arch/i386/math-emu/reg_convert.c 0 → 100644
View file @ 97ccc379
/*---------------------------------------------------------------------------+
| reg_convert.c |
| |
| Convert register representation. |
| |
| Copyright (C) 1992,1993,1994,1996,1997 |
| W. Metzenthen, 22 Parker St, Ormond, Vic 3163, Australia |
| E-mail billm@suburbia.net |
| |
| |
+---------------------------------------------------------------------------*/
#include "exception.h"
#include "fpu_emu.h"
int FPU_to_exp16(FPU_REG const *a, FPU_REG *x)
{
int sign = getsign(a);
*(long long *)&(x->sigl) = *(const long long *)&(a->sigl);
/* Set up the exponent as a 16 bit quantity. */
setexponent16(x, exponent(a));
if ( exponent16(x) == EXP_UNDER )
{
/* The number is a de-normal or pseudodenormal. */
/* We only deal with the significand and exponent. */
if (x->sigh & 0x80000000)
{
/* Is a pseudodenormal. */
/* This is non-80486 behaviour because the number
loses its 'denormal' identity. */
addexponent(x, 1);
}
else
{
/* Is a denormal. */
addexponent(x, 1);
FPU_normalize_nuo(x);
}
}
if ( !(x->sigh & 0x80000000) )
{
EXCEPTION(EX_INTERNAL | 0x180);
}
return sign;
}
arch/i386/math-emu/reg_div.S deleted 100644 → 0
View file @ 62e3b9df
This diff is collapsed.
arch/i386/math-emu/reg_divide.c 0 → 100644
View file @ 97ccc379
This diff is collapsed.
arch/i386/math-emu/reg_ld_str.c
View file @ 97ccc379
This diff is collapsed.
arch/i386/math-emu/reg_mul.c
View file @ 97ccc379
This diff is collapsed.
arch/i386/math-emu/reg_norm.S
View file @ 97ccc379
This diff is collapsed.
arch/i386/math-emu/reg_round.S
View file @ 97ccc379
This diff is collapsed.
arch/i386/math-emu/reg_u_add.S
View file @ 97ccc379
This diff is collapsed.
arch/i386/math-emu/reg_u_div.S
View file @ 97ccc379
This diff is collapsed.
arch/i386/math-emu/reg_u_mul.S
View file @ 97ccc379
This diff is collapsed.
arch/i386/math-emu/reg_u_sub.S
View file @ 97ccc379
This diff is collapsed.
arch/i386/math-emu/version.h
View file @ 97ccc379
......@@ -2,11 +2,11 @@
| version.h |
| |
| |
| Copyright (C) 1992,1993,1994,1996 |
| Copyright (C) 1992,1993,1994,1996,1997 |
| W. Metzenthen, 22 Parker St, Ormond, Vic 3163, Australia |
| E-mail billm@jacobi.maths.monash.edu.au |
| E-mail billm@suburbia.net |
| |
| |
+---------------------------------------------------------------------------*/
#define FPU_VERSION "wm-FPU-emu version 1.22"
#define FPU_VERSION "wm-FPU-emu version 2.00"
arch/i386/math-emu/wm_shrx.S
View file @ 97ccc379
......@@ -9,9 +9,9 @@
| Australia. E-mail billm@jacobi.maths.monash.edu.au |
| |
| Call from C as: |
| unsigned shrx(void *arg1, unsigned arg2) |
| unsigned FPU_shrx(void *arg1, unsigned arg2) |
| and |
| unsigned shrxs(void *arg1, unsigned arg2) |
| unsigned FPU_shrxs(void *arg1, unsigned arg2) |
| |
+---------------------------------------------------------------------------*/
......@@ -19,7 +19,7 @@
.text
/*---------------------------------------------------------------------------+
| unsigned shrx(void *arg1, unsigned arg2) |
| unsigned FPU_shrx(void *arg1, unsigned arg2) |
| |
| Extended shift right function. |
| Fastest for small shifts. |
......@@ -32,7 +32,7 @@
| Results returned in the 64 bit arg and eax. |
+---------------------------------------------------------------------------*/
ENTRY(shrx)
ENTRY(FPU_shrx)
push %ebp
movl %esp,%ebp
pushl %esi
......@@ -95,7 +95,7 @@ L_more_than_95:
/*---------------------------------------------------------------------------+
| unsigned shrxs(void *arg1, unsigned arg2) |
| unsigned FPU_shrxs(void *arg1, unsigned arg2) |
| |
| Extended shift right function (optimized for small floating point |
| integers). |
......@@ -110,7 +110,7 @@ L_more_than_95:
| part which has been shifted out of the arg. |
| Results returned in the 64 bit arg and eax. |
+---------------------------------------------------------------------------*/
ENTRY(shrxs)
ENTRY(FPU_shrxs)
push %ebp
movl %esp,%ebp
pushl %esi
......
arch/i386/math-emu/wm_sqrt.S
View file @ 97ccc379
This diff is collapsed.
arch/m68k/amiga/amifb.c deleted 100644 → 0
View file @ 62e3b9df
This diff is collapsed.
arch/m68k/amiga/amikeyb.c deleted 100644 → 0
View file @ 62e3b9df
This diff is collapsed.
arch/m68k/amiga/cyberfb.c deleted 100644 → 0
View file @ 62e3b9df
This diff is collapsed.
arch/m68k/amiga/retz3fb.c deleted 100644 → 0
View file @ 62e3b9df
This diff is collapsed.
arch/m68k/amiga/retz3fb.h deleted 100644 → 0
View file @ 62e3b9df
This diff is collapsed.
arch/m68k/atari/atafb.c deleted 100644 → 0
View file @ 62e3b9df
This diff is collapsed.
arch/m68k/atari/atafb.h deleted 100644 → 0
View file @ 62e3b9df
This diff is collapsed.
arch/m68k/console/fbcon.c deleted 100644 → 0
View file @ 62e3b9df
This diff is collapsed.
arch/m68k/console/font_8x16.c deleted 100644 → 0
View file @ 62e3b9df
This diff is collapsed.
arch/m68k/console/font_8x8.c deleted 100644 → 0
View file @ 62e3b9df
This diff is collapsed.
arch/m68k/console/fonts.c deleted 100644 → 0
View file @ 62e3b9df
This diff is collapsed.
arch/m68k/console/pearl_8x8.c deleted 100644 → 0
View file @ 62e3b9df
This diff is collapsed.
arch/m68k/console/txtcon.c deleted 100644 → 0
View file @ 62e3b9df
This diff is collapsed.
arch/mips/.gdbinit 0 → 100644
View file @ 97ccc379
echo Setting up the environment for debugging vmlinux...\n
echo set remotedebug 0 \n
set remotedebug 0
echo cd arch/mips/kernel \n
cd arch/mips/kernel
echo target remote /dev/ttyS0 \n
target remote /dev/ttyS0
arch/mips/Makefile
View file @ 97ccc379
This diff is collapsed.
arch/mips/boot/Makefile
View file @ 97ccc379
This diff is collapsed.
arch/mips/config.in
View file @ 97ccc379
This diff is collapsed.
arch/mips/defconfig
View file @ 97ccc379
This diff is collapsed.
arch/mips/deskstation/hw-access.c
View file @ 97ccc379
This diff is collapsed.
arch/mips/deskstation/int-handler.S
View file @ 97ccc379
This diff is collapsed.
arch/mips/deskstation/setup.c
View file @ 97ccc379
This diff is collapsed.
arch/mips/jazz/g364.c
View file @ 97ccc379
This diff is collapsed.
arch/mips/jazz/hw-access.c
View file @ 97ccc379
This diff is collapsed.
arch/mips/jazz/int-handler.S
View file @ 97ccc379
This diff is collapsed.
arch/mips/jazz/jazzdma.c
View file @ 97ccc379
This diff is collapsed.
arch/mips/jazz/setup.c
View file @ 97ccc379
This diff is collapsed.
arch/mips/kernel/Makefile
View file @ 97ccc379
This diff is collapsed.
arch/mips/kernel/branch.c
View file @ 97ccc379
This diff is collapsed.
arch/mips/kernel/entry.S
View file @ 97ccc379
This diff is collapsed.
arch/mips/kernel/fpe.c 0 → 100644
View file @ 97ccc379
This diff is collapsed.
arch/mips/kernel/gdb-low.S
View file @ 97ccc379
This diff is collapsed.
arch/mips/kernel/gdb-stub.c
View file @ 97ccc379
This diff is collapsed.
arch/mips/kernel/head.S
View file @ 97ccc379
This diff is collapsed.
arch/mips/kernel/irix5sys.h
View file @ 97ccc379
This diff is collapsed.
arch/mips/kernel/irixelf.c
View file @ 97ccc379
This diff is collapsed.
arch/mips/kernel/irixinv.c 0 → 100644
View file @ 97ccc379
This diff is collapsed.
arch/mips/kernel/irixioctl.c
View file @ 97ccc379
This diff is collapsed.
arch/mips/kernel/irixsig.c
View file @ 97ccc379
This diff is collapsed.
arch/mips/kernel/irq.c
View file @ 97ccc379
This diff is collapsed.
arch/mips/kernel/mips_ksyms.c
View file @ 97ccc379
This diff is collapsed.
arch/mips/kernel/pci.c
View file @ 97ccc379
This diff is collapsed.
arch/mips/kernel/proc.c
View file @ 97ccc379
This diff is collapsed.
arch/mips/kernel/process.c
View file @ 97ccc379
This diff is collapsed.
arch/mips/kernel/ptrace.c
View file @ 97ccc379
This diff is collapsed.
arch/mips/kernel/r2300_fpu.S
View file @ 97ccc379
This diff is collapsed.
arch/mips/kernel/r2300_misc.S
View file @ 97ccc379
This diff is collapsed.
arch/mips/kernel/r2300_scall.S
View file @ 97ccc379
This diff is collapsed.
arch/mips/kernel/r2300_switch.S
View file @ 97ccc379
This diff is collapsed.
arch/mips/kernel/r4k_fpu.S
View file @ 97ccc379
This diff is collapsed.
arch/mips/kernel/r4k_misc.S
View file @ 97ccc379
This diff is collapsed.
arch/mips/kernel/r4k_scall.S
View file @ 97ccc379
This diff is collapsed.
arch/mips/kernel/r6000_fpu.S
View file @ 97ccc379
This diff is collapsed.
arch/mips/kernel/setup.c
View file @ 97ccc379
This diff is collapsed.
arch/mips/kernel/signal.c
View file @ 97ccc379
This diff is collapsed.
arch/mips/kernel/syscall.c
View file @ 97ccc379
This diff is collapsed.
arch/mips/kernel/syscalls.h
View file @ 97ccc379
This diff is collapsed.
arch/mips/kernel/sysirix.c
View file @ 97ccc379
This diff is collapsed.
arch/mips/kernel/sysmips.c
View file @ 97ccc379
This diff is collapsed.
arch/mips/kernel/time.c
View file @ 97ccc379
This diff is collapsed.
arch/mips/kernel/traps.c
View file @ 97ccc379
This diff is collapsed.
arch/mips/kernel/unaligned.c
View file @ 97ccc379
This diff is collapsed.
arch/mips/lib/Makefile
View file @ 97ccc379
This diff is collapsed.
arch/mips/lib/checksum.c
View file @ 97ccc379
This diff is collapsed.
arch/mips/lib/copy_user.S
View file @ 97ccc379
This diff is collapsed.
arch/mips/lib/io.c deleted 100644 → 0
View file @ 62e3b9df
This diff is collapsed.
arch/mips/mm/andes.c
View file @ 97ccc379
This diff is collapsed.
arch/mips/mm/loadmmu.c
View file @ 97ccc379
This diff is collapsed.
arch/mips/mm/r2300.c
View file @ 97ccc379
This diff is collapsed.
arch/mips/mm/r4xx0.c
View file @ 97ccc379
This diff is collapsed.
arch/mips/mm/r6000.c
View file @ 97ccc379
This diff is collapsed.
arch/mips/mm/tfp.c
View file @ 97ccc379
This diff is collapsed.
arch/mips/sgi/kernel/Makefile
View file @ 97ccc379
This diff is collapsed.
arch/mips/sgi/kernel/indyIRQ.S
View file @ 97ccc379
This diff is collapsed.
arch/mips/sgi/kernel/indy_hpc.c
View file @ 97ccc379
This diff is collapsed.
arch/mips/sgi/kernel/indy_int.c
View file @ 97ccc379
This diff is collapsed.
arch/mips/sgi/kernel/indy_mc.c
View file @ 97ccc379
This diff is collapsed.
arch/mips/sgi/kernel/indy_timer.c
View file @ 97ccc379
This diff is collapsed.
arch/mips/sgi/kernel/setup.c
View file @ 97ccc379
This diff is collapsed.
arch/mips/sgi/kernel/system.c
View file @ 97ccc379
This diff is collapsed.
arch/mips/sgi/kernel/time.c
View file @ 97ccc379
This diff is collapsed.
arch/mips/sgi/prom/Makefile
View file @ 97ccc379
This diff is collapsed.
arch/mips/sgi/prom/cmdline.c
View file @ 97ccc379
This diff is collapsed.
arch/mips/sgi/prom/console.c
View file @ 97ccc379
This diff is collapsed.
arch/mips/sgi/prom/env.c
View file @ 97ccc379
This diff is collapsed.
arch/mips/sgi/prom/file.c
View file @ 97ccc379
This diff is collapsed.
arch/mips/sgi/prom/init.c
View file @ 97ccc379
This diff is collapsed.
arch/mips/sgi/prom/memory.c
View file @ 97ccc379
This diff is collapsed.
arch/mips/sgi/prom/printf.c
View file @ 97ccc379
This diff is collapsed.
arch/mips/sgi/prom/salone.c
View file @ 97ccc379
This diff is collapsed.
arch/mips/sgi/prom/tags.c
View file @ 97ccc379
This diff is collapsed.
arch/mips/sgi/prom/time.c
View file @ 97ccc379
This diff is collapsed.
arch/mips/sgi/prom/tree.c
View file @ 97ccc379
This diff is collapsed.
arch/mips/sni/hw-access.c
View file @ 97ccc379
This diff is collapsed.
arch/mips/sni/int-handler.S
View file @ 97ccc379
This diff is collapsed.
arch/mips/sni/pci.c
View file @ 97ccc379
This diff is collapsed.
arch/mips/sni/setup.c
View file @ 97ccc379
This diff is collapsed.
arch/mips/tools/Makefile
View file @ 97ccc379
This diff is collapsed.
arch/mips/tools/offset.c
View file @ 97ccc379
This diff is collapsed.
drivers/block/Config.in
View file @ 97ccc379
This diff is collapsed.
drivers/block/Makefile
View file @ 97ccc379
This diff is collapsed.
drivers/block/ali14xx.c
View file @ 97ccc379
This diff is collapsed.
drivers/block/cmd640.c
View file @ 97ccc379
This diff is collapsed.
drivers/block/dtc2278.c
View file @ 97ccc379
This diff is collapsed.
drivers/block/ht6560b.c
View file @ 97ccc379
This diff is collapsed.
drivers/block/ide-cd.c
View file @ 97ccc379
This diff is collapsed.
drivers/block/ide-cd.h
View file @ 97ccc379
This diff is collapsed.
drivers/block/ide-disk.c
View file @ 97ccc379
This diff is collapsed.
drivers/block/ide-dma.c
View file @ 97ccc379
This diff is collapsed.
drivers/block/ide-floppy.c
View file @ 97ccc379
This diff is collapsed.
drivers/block/ide-pci.c 0 → 100644
View file @ 97ccc379
This diff is collapsed.
drivers/block/ide-probe.c
View file @ 97ccc379
This diff is collapsed.
drivers/block/ide-proc.c 0 → 100644
View file @ 97ccc379
This diff is collapsed.
drivers/block/ide-tape.c
View file @ 97ccc379
This diff is collapsed.
drivers/block/ide.c
View file @ 97ccc379
This diff is collapsed.
drivers/block/ide.h
View file @ 97ccc379
This diff is collapsed.
drivers/block/ll_rw_blk.c
View file @ 97ccc379
This diff is collapsed.
drivers/block/ns87415.c 0 → 100644
View file @ 97ccc379
This diff is collapsed.
drivers/block/opti621.c
View file @ 97ccc379
This diff is collapsed.
drivers/block/pdc4030.c
View file @ 97ccc379
This diff is collapsed.
drivers/block/qd6580.c
View file @ 97ccc379
This diff is collapsed.
drivers/block/rz1000.c
View file @ 97ccc379
This diff is collapsed.
drivers/block/trm290.c
View file @ 97ccc379
This diff is collapsed.
drivers/block/umc8672.c
View file @ 97ccc379
This diff is collapsed.
drivers/char/pc_keyb.c
View file @ 97ccc379
This diff is collapsed.
drivers/char/pc_keyb.h
View file @ 97ccc379
This diff is collapsed.
drivers/char/pcwd.c
View file @ 97ccc379
This diff is collapsed.
drivers/net/3c505.c
View file @ 97ccc379
This diff is collapsed.
drivers/net/3c59x.c
View file @ 97ccc379
This diff is collapsed.
drivers/net/Config.in
View file @ 97ccc379
This diff is collapsed.
drivers/net/eth16i.c
View file @ 97ccc379
This diff is collapsed.
drivers/net/hamradio/Config.in
View file @ 97ccc379
This diff is collapsed.
drivers/net/hamradio/dmascc.c 0 → 100644
View file @ 97ccc379
This diff is collapsed.
drivers/net/ipddp.c
View file @ 97ccc379
This diff is collapsed.
drivers/pci/pci.c
View file @ 97ccc379
This diff is collapsed.
drivers/scsi/ide-scsi.c
View file @ 97ccc379
This diff is collapsed.
drivers/scsi/ide-scsi.h
View file @ 97ccc379
This diff is collapsed.
drivers/scsi/st.c
View file @ 97ccc379
This diff is collapsed.
drivers/sgi/Makefile 0 → 100644
View file @ 97ccc379
This diff is collapsed.
drivers/sgi/char/cons_newport.c 0 → 100644
View file @ 97ccc379
This diff is collapsed.
drivers/sgi/char/gconsole.h 0 → 100644
View file @ 97ccc379
This diff is collapsed.
drivers/sgi/char/graphics.c 0 → 100644
View file @ 97ccc379
This diff is collapsed.
drivers/sgi/char/graphics.h 0 → 100644
View file @ 97ccc379
This diff is collapsed.
drivers/sgi/char/linux_logo.h 0 → 100644
View file @ 97ccc379
This diff is collapsed.
drivers/sgi/char/newport.c 0 → 100644
View file @ 97ccc379
This diff is collapsed.
drivers/sgi/char/newport.h 0 → 100644
View file @ 97ccc379
This diff is collapsed.
drivers/sgi/char/rrm.c 0 → 100644
View file @ 97ccc379
This diff is collapsed.
drivers/sgi/char/sgicons.c 0 → 100644
View file @ 97ccc379
This diff is collapsed.
drivers/sgi/char/sgiserial.c 0 → 100644
View file @ 97ccc379
This diff is collapsed.
drivers/sgi/char/sgiserial.h 0 → 100644
View file @ 97ccc379
This diff is collapsed.
drivers/sgi/char/shmiq.c 0 → 100644
View file @ 97ccc379
This diff is collapsed.
drivers/sgi/char/streamable.c 0 → 100644
View file @ 97ccc379
This diff is collapsed.
drivers/sgi/char/usema.c 0 → 100644
View file @ 97ccc379
This diff is collapsed.
drivers/sgi/char/vga_font.c 0 → 100644
View file @ 97ccc379
This diff is collapsed.
drivers/sound/lowlevel/awe_wave.c
View file @ 97ccc379
This diff is collapsed.
fs/dcache.c
View file @ 97ccc379
This diff is collapsed.
fs/ext2/fsync.c
View file @ 97ccc379
This diff is collapsed.
fs/fat/inode.c
View file @ 97ccc379
This diff is collapsed.
fs/isofs/dir.c
View file @ 97ccc379
This diff is collapsed.
fs/isofs/inode.c
View file @ 97ccc379
This diff is collapsed.
fs/isofs/rock.c
View file @ 97ccc379
This diff is collapsed.
fs/msdos/namei.c
View file @ 97ccc379
This diff is collapsed.
fs/namei.c
View file @ 97ccc379
This diff is collapsed.
fs/nfs/dir.c
View file @ 97ccc379
This diff is collapsed.
fs/nfs/inode.c
View file @ 97ccc379
This diff is collapsed.
fs/nfs/write.c
View file @ 97ccc379
This diff is collapsed.
fs/nfsd/vfs.c
View file @ 97ccc379
This diff is collapsed.
fs/proc/generic.c
View file @ 97ccc379
This diff is collapsed.
fs/smbfs/dir.c
View file @ 97ccc379
This diff is collapsed.
fs/smbfs/file.c
View file @ 97ccc379
This diff is collapsed.
fs/smbfs/inode.c
View file @ 97ccc379
This diff is collapsed.
fs/smbfs/ioctl.c
View file @ 97ccc379
This diff is collapsed.
fs/smbfs/proc.c
View file @ 97ccc379
This diff is collapsed.
fs/smbfs/sock.c
View file @ 97ccc379
This diff is collapsed.
fs/sysv/namei.c
View file @ 97ccc379
This diff is collapsed.
include/asm-i386/math_emu.h
View file @ 97ccc379
This diff is collapsed.
include/asm-i386/processor.h
View file @ 97ccc379
This diff is collapsed.
include/asm-i386/ptrace.h
View file @ 97ccc379
This diff is collapsed.
include/asm-i386/system.h
View file @ 97ccc379
This diff is collapsed.
include/asm-mips/asmmacro.h
View file @ 97ccc379
This diff is collapsed.
include/asm-mips/atomic.h
View file @ 97ccc379
This diff is collapsed.
include/asm-mips/bitops.h
View file @ 97ccc379
This diff is collapsed.
include/asm-mips/bootinfo.h
View file @ 97ccc379
This diff is collapsed.
include/asm-mips/bugs.h
View file @ 97ccc379
This diff is collapsed.
include/asm-mips/byteorder.h
View file @ 97ccc379
This diff is collapsed.
include/asm-mips/checksum.h
View file @ 97ccc379
This diff is collapsed.
include/asm-mips/cpu.h
View file @ 97ccc379
This diff is collapsed.
include/asm-mips/decstation.h deleted 100644 → 0
View file @ 62e3b9df
This diff is collapsed.
include/asm-mips/deskstation.h deleted 100644 → 0
View file @ 62e3b9df
This diff is collapsed.
include/asm-mips/dma.h
View file @ 97ccc379
This diff is collapsed.
include/asm-mips/elf.h
View file @ 97ccc379
This diff is collapsed.
include/asm-mips/floppy.h
View file @ 97ccc379
This diff is collapsed.
include/asm-mips/gdb-stub.h
View file @ 97ccc379
This diff is collapsed.
include/asm-mips/gfx.h
View file @ 97ccc379
This diff is collapsed.
include/asm-mips/inventory.h 0 → 100644
View file @ 97ccc379
This diff is collapsed.
include/asm-mips/io.h
View file @ 97ccc379
This diff is collapsed.
include/asm-mips/ioctls.h
View file @ 97ccc379
This diff is collapsed.
include/asm-mips/irq.h
View file @ 97ccc379
This diff is collapsed.
include/asm-mips/jazz.h
View file @ 97ccc379
This diff is collapsed.
include/asm-mips/jazzdma.h
View file @ 97ccc379
This diff is collapsed.
include/asm-mips/keyboard.h
View file @ 97ccc379
This diff is collapsed.
include/asm-mips/mipsconfig.h
View file @ 97ccc379
This diff is collapsed.
include/asm-mips/mipsregs.h
View file @ 97ccc379
This diff is collapsed.
include/asm-mips/namei.h
View file @ 97ccc379
This diff is collapsed.
include/asm-mips/ng1.h
View file @ 97ccc379
This diff is collapsed.
include/asm-mips/offset.h
View file @ 97ccc379
This diff is collapsed.
include/asm-mips/pci.h
View file @ 97ccc379
This diff is collapsed.
include/asm-mips/pgtable.h
View file @ 97ccc379
This diff is collapsed.
include/asm-mips/posix_types.h
View file @ 97ccc379
This diff is collapsed.
include/asm-mips/prctl.h 0 → 100644
View file @ 97ccc379
This diff is collapsed.
include/asm-mips/processor.h
View file @ 97ccc379
This diff is collapsed.
include/asm-mips/r4kcache.h
View file @ 97ccc379
This diff is collapsed.
include/asm-mips/scatterlist.h
View file @ 97ccc379
This diff is collapsed.
include/asm-mips/sgi.h
View file @ 97ccc379
This diff is collapsed.
include/asm-mips/sgialib.h
View file @ 97ccc379
This diff is collapsed.
include/asm-mips/sgiarcs.h
View file @ 97ccc379
This diff is collapsed.
include/asm-mips/sgihpc.h
View file @ 97ccc379
This diff is collapsed.
include/asm-mips/sgimc.h
View file @ 97ccc379
This diff is collapsed.
include/asm-mips/sgint23.h
View file @ 97ccc379
This diff is collapsed.
include/asm-mips/shmiq.h
View file @ 97ccc379
This diff is collapsed.
include/asm-mips/sigcontext.h
View file @ 97ccc379
This diff is collapsed.
include/asm-mips/siginfo.h 0 → 100644
View file @ 97ccc379
This diff is collapsed.
include/asm-mips/signal.h
View file @ 97ccc379
This diff is collapsed.
include/asm-mips/socket.h
View file @ 97ccc379
This diff is collapsed.
include/asm-mips/stackframe.h
View file @ 97ccc379
This diff is collapsed.
include/asm-mips/string.h
View file @ 97ccc379
This diff is collapsed.
include/asm-mips/system.h
View file @ 97ccc379
This diff is collapsed.
include/asm-mips/termbits.h
View file @ 97ccc379
This diff is collapsed.
include/asm-mips/uaccess.h
View file @ 97ccc379
This diff is collapsed.
include/asm-mips/unaligned.h
View file @ 97ccc379
This diff is collapsed.
include/asm-mips/usioctl.h 0 → 100644
View file @ 97ccc379
This diff is collapsed.
include/asm-mips/vector.h
View file @ 97ccc379
This diff is collapsed.
include/asm-mips/watch.h
View file @ 97ccc379
This diff is collapsed.
include/linux/blkdev.h
View file @ 97ccc379
This diff is collapsed.
include/linux/fs.h
View file @ 97ccc379
This diff is collapsed.
include/linux/hdreg.h
View file @ 97ccc379
This diff is collapsed.
include/linux/ip_fw.h
View file @ 97ccc379
This diff is collapsed.
include/linux/mm.h
View file @ 97ccc379
This diff is collapsed.
include/linux/nfs_fs.h
View file @ 97ccc379
This diff is collapsed.
include/linux/pci.h
View file @ 97ccc379
This diff is collapsed.
include/linux/proc_fs.h
View file @ 97ccc379
This diff is collapsed.
include/linux/smb_fs.h
View file @ 97ccc379
This diff is collapsed.
include/linux/socket.h
View file @ 97ccc379
This diff is collapsed.
include/linux/sysctl.h
View file @ 97ccc379
This diff is collapsed.
include/net/ip_autofw.h 0 → 100644
View file @ 97ccc379
This diff is collapsed.
include/net/ip_masq.h
View file @ 97ccc379
This diff is collapsed.
include/net/ip_masq_mod.h 0 → 100644
View file @ 97ccc379
This diff is collapsed.
include/net/ip_portfw.h 0 → 100644
View file @ 97ccc379
This diff is collapsed.
include/net/llc.h
View file @ 97ccc379
This diff is collapsed.
include/net/tcp.h
View file @ 97ccc379
This diff is collapsed.
init/main.c
View file @ 97ccc379
This diff is collapsed.
kernel/ksyms.c
View file @ 97ccc379
This diff is collapsed.
kernel/printk.c
View file @ 97ccc379
This diff is collapsed.
kernel/signal.c
View file @ 97ccc379
This diff is collapsed.
net/Config.in
View file @ 97ccc379
This diff is collapsed.
net/appletalk/ddp.c
View file @ 97ccc379
This diff is collapsed.
net/core/iovec.c
View file @ 97ccc379
This diff is collapsed.
net/ipv4/Config.in
View file @ 97ccc379
This diff is collapsed.
net/ipv4/Makefile
View file @ 97ccc379
This diff is collapsed.
net/ipv4/arp.c
View file @ 97ccc379
This diff is collapsed.
net/ipv4/fib_frontend.c
View file @ 97ccc379
This diff is collapsed.
net/ipv4/icmp.c
View file @ 97ccc379
This diff is collapsed.
net/ipv4/ip_forward.c
View file @ 97ccc379
This diff is collapsed.
net/ipv4/ip_masq.c
View file @ 97ccc379
This diff is collapsed.
net/ipv4/ip_masq_app.c
View file @ 97ccc379
This diff is collapsed.
net/ipv4/ip_masq_autofw.c 0 → 100644
View file @ 97ccc379
This diff is collapsed.
net/ipv4/ip_masq_cuseeme.c 0 → 100644
View file @ 97ccc379
This diff is collapsed.
net/ipv4/ip_masq_ftp.c
View file @ 97ccc379
This diff is collapsed.
net/ipv4/ip_masq_irc.c
View file @ 97ccc379
This diff is collapsed.
net/ipv4/ip_masq_mod.c 0 → 100644
View file @ 97ccc379
This diff is collapsed.
net/ipv4/ip_masq_portfw.c 0 → 100644
View file @ 97ccc379
This diff is collapsed.
net/ipv4/ip_masq_quake.c
View file @ 97ccc379
This diff is collapsed.
net/ipv4/ip_masq_raudio.c
View file @ 97ccc379
This diff is collapsed.
net/ipv4/ip_masq_vdolive.c 0 → 100644
View file @ 97ccc379
This diff is collapsed.
net/ipv4/ip_output.c
View file @ 97ccc379
This diff is collapsed.
net/ipv4/ip_sockglue.c
View file @ 97ccc379
This diff is collapsed.
net/ipv4/route.c
View file @ 97ccc379
This diff is collapsed.
net/ipv4/sysctl_net_ipv4.c
View file @ 97ccc379
This diff is collapsed.
net/ipv4/tcp_ipv4.c
View file @ 97ccc379
This diff is collapsed.
net/ipv4/tcp_timer.c
View file @ 97ccc379
This diff is collapsed.
net/ipv4/udp.c
View file @ 97ccc379
This diff is collapsed.
net/ipv6/addrconf.c
View file @ 97ccc379
This diff is collapsed.
net/sunrpc/svc.c
View file @ 97ccc379
This diff is collapsed.
Markdown is supported
0% or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment
GitLab Nexedi Edition | About GitLab | About Nexedi | 沪ICP备2021021310号-2 | 沪ICP备2021021310号-7