aboutsummaryrefslogtreecommitdiff
path: root/arch/i386/math-emu/poly.h
diff options
context:
space:
mode:
authorLinus Torvalds <torvalds@ppc970.osdl.org>2005-04-16 15:20:36 -0700
committerLinus Torvalds <torvalds@ppc970.osdl.org>2005-04-16 15:20:36 -0700
commit1da177e4c3f41524e886b7f1b8a0c1fc7321cac2 (patch)
tree0bba044c4ce775e45a88a51686b5d9f90697ea9d /arch/i386/math-emu/poly.h
Linux-2.6.12-rc2
Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
Diffstat (limited to 'arch/i386/math-emu/poly.h')
-rw-r--r--arch/i386/math-emu/poly.h121
1 files changed, 121 insertions, 0 deletions
diff --git a/arch/i386/math-emu/poly.h b/arch/i386/math-emu/poly.h
new file mode 100644
index 00000000000..4db79811492
--- /dev/null
+++ b/arch/i386/math-emu/poly.h
@@ -0,0 +1,121 @@
+/*---------------------------------------------------------------------------+
+ | poly.h |
+ | |
+ | Header file for the FPU-emu poly*.c source files. |
+ | |
+ | Copyright (C) 1994,1999 |
+ | W. Metzenthen, 22 Parker St, Ormond, Vic 3163, |
+ | Australia. E-mail billm@melbpc.org.au |
+ | |
+ | Declarations and definitions for functions operating on Xsig (12-byte |
+ | extended-significand) quantities. |
+ | |
+ +---------------------------------------------------------------------------*/
+
+#ifndef _POLY_H
+#define _POLY_H
+
+/* This 12-byte structure is used to improve the accuracy of computation
+ of transcendental functions.
+ Intended to be used to get results better than 8-byte computation
+ allows. 9-byte would probably be sufficient.
+ */
+typedef struct {
+ unsigned long lsw;
+ unsigned long midw;
+ unsigned long msw;
+} Xsig;
+
+asmlinkage void mul64(unsigned long long const *a, unsigned long long const *b,
+ unsigned long long *result);
+asmlinkage void polynomial_Xsig(Xsig *, const unsigned long long *x,
+ const unsigned long long terms[], const int n);
+
+asmlinkage void mul32_Xsig(Xsig *, const unsigned long mult);
+asmlinkage void mul64_Xsig(Xsig *, const unsigned long long *mult);
+asmlinkage void mul_Xsig_Xsig(Xsig *dest, const Xsig *mult);
+
+asmlinkage void shr_Xsig(Xsig *, const int n);
+asmlinkage int round_Xsig(Xsig *);
+asmlinkage int norm_Xsig(Xsig *);
+asmlinkage void div_Xsig(Xsig *x1, const Xsig *x2, const Xsig *dest);
+
+/* Macro to extract the most significant 32 bits from a long long */
+#define LL_MSW(x) (((unsigned long *)&x)[1])
+
+/* Macro to initialize an Xsig struct */
+#define MK_XSIG(a,b,c) { c, b, a }
+
+/* Macro to access the 8 ms bytes of an Xsig as a long long */
+#define XSIG_LL(x) (*(unsigned long long *)&x.midw)
+
+
+/*
+ Need to run gcc with optimizations on to get these to
+ actually be in-line.
+ */
+
+/* Multiply two fixed-point 32 bit numbers, producing a 32 bit result.
+ The answer is the ms word of the product. */
+/* Some versions of gcc make it difficult to stop eax from being clobbered.
+ Merely specifying that it is used doesn't work...
+ */
+static inline unsigned long mul_32_32(const unsigned long arg1,
+ const unsigned long arg2)
+{
+ int retval;
+ asm volatile ("mull %2; movl %%edx,%%eax" \
+ :"=a" (retval) \
+ :"0" (arg1), "g" (arg2) \
+ :"dx");
+ return retval;
+}
+
+
+/* Add the 12 byte Xsig x2 to Xsig dest, with no checks for overflow. */
+static inline void add_Xsig_Xsig(Xsig *dest, const Xsig *x2)
+{
+ asm volatile ("movl %1,%%edi; movl %2,%%esi;\n"
+ "movl (%%esi),%%eax; addl %%eax,(%%edi);\n"
+ "movl 4(%%esi),%%eax; adcl %%eax,4(%%edi);\n"
+ "movl 8(%%esi),%%eax; adcl %%eax,8(%%edi);\n"
+ :"=g" (*dest):"g" (dest), "g" (x2)
+ :"ax","si","di");
+}
+
+
+/* Add the 12 byte Xsig x2 to Xsig dest, adjust exp if overflow occurs. */
+/* Note: the constraints in the asm statement didn't always work properly
+ with gcc 2.5.8. Changing from using edi to using ecx got around the
+ problem, but keep fingers crossed! */
+static inline void add_two_Xsig(Xsig *dest, const Xsig *x2, long int *exp)
+{
+ asm volatile ("movl %2,%%ecx; movl %3,%%esi;\n"
+ "movl (%%esi),%%eax; addl %%eax,(%%ecx);\n"
+ "movl 4(%%esi),%%eax; adcl %%eax,4(%%ecx);\n"
+ "movl 8(%%esi),%%eax; adcl %%eax,8(%%ecx);\n"
+ "jnc 0f;\n"
+ "rcrl 8(%%ecx); rcrl 4(%%ecx); rcrl (%%ecx)\n"
+ "movl %4,%%ecx; incl (%%ecx)\n"
+ "movl $1,%%eax; jmp 1f;\n"
+ "0: xorl %%eax,%%eax;\n"
+ "1:\n"
+ :"=g" (*exp), "=g" (*dest)
+ :"g" (dest), "g" (x2), "g" (exp)
+ :"cx","si","ax");
+}
+
+
+/* Negate (subtract from 1.0) the 12 byte Xsig */
+/* This is faster in a loop on my 386 than using the "neg" instruction. */
+static inline void negate_Xsig(Xsig *x)
+{
+ asm volatile("movl %1,%%esi;\n"
+ "xorl %%ecx,%%ecx;\n"
+ "movl %%ecx,%%eax; subl (%%esi),%%eax; movl %%eax,(%%esi);\n"
+ "movl %%ecx,%%eax; sbbl 4(%%esi),%%eax; movl %%eax,4(%%esi);\n"
+ "movl %%ecx,%%eax; sbbl 8(%%esi),%%eax; movl %%eax,8(%%esi);\n"
+ :"=g" (*x):"g" (x):"si","ax","cx");
+}
+
+#endif /* _POLY_H */