Commit MD parts of a loosely functional AMD64 port. This is based on

a heavily stripped down FreeBSD/i386 (brutally stripped down actually) to
attempt to get a stable base to start from.  There is a lot missing still.
Worth noting:
- The kernel runs at 1GB in order to cheat with the pmap code.  pmap uses
  a variation of the PAE code in order to avoid having to worry about 4
  levels of page tables yet.
- It boots in 64 bit "long mode" with a tiny trampoline embedded in the
  i386 loader.  This simplifies locore.s greatly.
- There are still quite a few fragments of i386-specific code that have
  not been translated yet, and some that I cheated and wrote dumb C
  versions of (bcopy etc).
- It has both int 0x80 for syscalls (but using registers for argument
  passing, as is native on the amd64 ABI), and the 'syscall' instruction
  for syscalls.  int 0x80 preserves all registers, 'syscall' does not.
- I have tried to minimize looking at the NetBSD code, except in a couple
  of places (eg: to find which register they use to replace the trashed
  %rcx register in the syscall instruction).  As a result, there is not a
  lot of similarity.  I did look at NetBSD a few times while debugging to
  get some ideas about what I might have done wrong in my first attempt.

1 files changed, 15 insertions, 58 deletions

diff --git a/sys/amd64/include/fpu.h b/sys/amd64/include/fpu.h
index 8e0f9bc..db1fe40 100644
--- a/sys/amd64/include/fpu.h
+++ b/sys/amd64/include/fpu.h
@@ -45,81 +45,38 @@
 #ifndef _MACHINE_NPX_H_
 #define	_MACHINE_NPX_H_
 
-/* Environment information of floating point unit */
-struct env87 {
-	long	en_cw;		/* control word (16bits) */
-	long	en_sw;		/* status word (16bits) */
-	long	en_tw;		/* tag word (16bits) */
-	long	en_fip;		/* floating point instruction pointer */
-	u_short	en_fcs;		/* floating code segment selector */
-	u_short	en_opcode;	/* opcode last executed (11 bits ) */
-	long	en_foo;		/* floating operand offset */
-	long	en_fos;		/* floating operand segment selector */
-};
-
-/* Contents of each floating point accumulator */
+/* Contents of each x87 floating point accumulator */
 struct fpacc87 {
-#ifdef dontdef /* too unportable */
-	u_long	fp_mantlo;	/* mantissa low (31:0) */
-	u_long	fp_manthi;	/* mantissa high (63:32) */
-	int	fp_exp:15;	/* exponent */
-	int	fp_sgn:1;	/* mantissa sign */
-#else
 	u_char	fp_bytes[10];
-#endif
 };
 
-/* Floating point context */
-struct save87 {
-	struct	env87 sv_env;	/* floating point control/status */
-	struct	fpacc87	sv_ac[8];	/* accumulator contents, 0-7 */
-	u_char	sv_pad0[4];	/* padding for (now unused) saved status word */
-	/*
-	 * Bogus padding for emulators.  Emulators should use their own
-	 * struct and arrange to store into this struct (ending here)
-	 * before it is inspected for ptracing or for core dumps.  Some
-	 * emulators overwrite the whole struct.  We have no good way of
-	 * knowing how much padding to leave.  Leave just enough for the
-	 * GPL emulator's i387_union (176 bytes total).
-	 */
-	u_char	sv_pad[64];	/* padding; used by emulators */
+/* Contents of each SSE extended accumulator */
+struct  xmmacc {
+	u_char	xmm_bytes[16];
 };
 
 struct  envxmm {
 	u_int16_t	en_cw;		/* control word (16bits) */
 	u_int16_t	en_sw;		/* status word (16bits) */
-	u_int16_t	en_tw;		/* tag word (16bits) */
+	u_int8_t	en_tw;		/* tag word (8bits) */
+	u_int8_t	en_zero;
 	u_int16_t	en_opcode;	/* opcode last executed (11 bits ) */
-	u_int32_t	en_fip;		/* floating point instruction pointer */
-	u_int16_t	en_fcs;		/* floating code segment selector */
-	u_int16_t	en_pad0;	/* padding */
-	u_int32_t	en_foo;		/* floating operand offset */
-	u_int16_t	en_fos;		/* floating operand segment selector */
-	u_int16_t	en_pad1;	/* padding */
+	u_int64_t	en_rip;		/* floating point instruction pointer */
+	u_int64_t	en_rdp;		/* floating operand pointer */
 	u_int32_t	en_mxcsr;	/* SSE sontorol/status register */
-	u_int32_t	en_pad2;	/* padding */
+	u_int32_t	en_mxcsr_mask;	/* valid bits in mxcsr */
 };
 
-/* Contents of each SSE extended accumulator */
-struct  xmmacc {
-	u_char	xmm_bytes[16];
-};
-
-struct  savexmm {
+struct  savefpu {
 	struct	envxmm	sv_env;
 	struct {
 		struct fpacc87	fp_acc;
 		u_char		fp_pad[6];      /* padding */
 	} sv_fp[8];
-	struct xmmacc	sv_xmm[8];
-	u_char sv_pad[224];
+	struct xmmacc	sv_xmm[16];
+	u_char sv_pad[96];
 } __aligned(16);
 
-union	savefpu {
-	struct	save87	sv_87;
-	struct	savexmm	sv_xmm;
-};
-
 /*
  * The hardware default control word for i387's and later coprocessors is
  * 0x37F, giving:
@@ -144,10 +101,10 @@ int	npxdna(void);
 void	npxdrop(void);
 void	npxexit(struct thread *td);
 int	npxformat(void);
-int	npxgetregs(struct thread *td, union savefpu *addr);
+int	npxgetregs(struct thread *td, struct savefpu *addr);
 void	npxinit(u_short control);
-void	npxsave(union savefpu *addr);
-void	npxsetregs(struct thread *td, union savefpu *addr);
+void	npxsave(struct savefpu *addr);
+void	npxsetregs(struct thread *td, struct savefpu *addr);
 int	npxtrap(void);
 #endif