Mirror of the gdb-patches mailing list
 help / color / mirror / Atom feed
* [PATCH/RFA] Move i387 register supply/fill to i387-tdep.c
@ 2002-04-11 16:05 Jason R Thorpe
  0 siblings, 0 replies; 8+ messages in thread
From: Jason R Thorpe @ 2002-04-11 16:05 UTC (permalink / raw)
  To: gdb-patches

[-- Attachment #1: Type: text/plain, Size: 1204 bytes --]

Per my query on gdb@ about cross-debugging of core files, I'd like to move
handling of i386 register supply/fill from i387-nat.c to i387-tdep.c.  This
renders i387-nat.c empty.

OK to commit?

	* Makefile.in (ALLDEPFILES): Remove i387-nat.c.
	(i387-nat.o): Delete rule.
	* i387-nat.c: Delete file, moving contents to...
	* i387-tdep.c: ...here.
	* i387-nat.h: Rename...
	* i387-tdep.h: ...to this.
	* go32-nat.c: Include i387-tdep.h instead of i387-nat.h.
	* i386-linux-nat.c: Likewise.
	* i386bsd-nat.c: Likewise.
	* i386gnu-nat.c: Likewise.
	* i386nbsd-nat.c: Likewise.
	* i386v4-nat.c: Likewise.
	* x86-64-linux-nat.c: Likewise.
	* config/i386/fbsd.mh (NATDEPFILES): Remove i387-nat.o.
	* config/i386/go32.mh (NATDEPFILES): Likewise.
	* config/i386/i386gnu.mh (NATDEPFILES): Likewise.
	* config/i386/i386sol2.mh (NATDEPFILES): Likewise.
	* config/i386/i386v42mp.mh (NATDEPFILES): Likewise.
	* config/i386/linux.mh (NATDEPFILES): Likewise.
	* config/i386/nbsd.mh (NATDEPFILES): Likewise.
	* config/i386/nbsdelf.mh (NATDEPFILES): Likewise.
	* config/i386/obsd.mh (NATDEPFILES): Likewise.
	* config/i386/x86-64linux.mh (NATDEPFILES): Likewise.

-- 
        -- Jason R. Thorpe <thorpej@wasabisystems.com>

[-- Attachment #2: i387-tdep.diff --]
[-- Type: text/plain, Size: 39107 bytes --]

Index: Makefile.in
===================================================================
RCS file: /cvs/src/src/gdb/Makefile.in,v
retrieving revision 1.169
diff -c -r1.169 Makefile.in
*** Makefile.in	9 Apr 2002 03:06:13 -0000	1.169
--- Makefile.in	11 Apr 2002 22:43:31 -0000
***************
*** 1184,1190 ****
  	i386-tdep.c i386b-nat.c i386mach-nat.c i386v-nat.c i386-linux-nat.c \
  	i386aix-nat.c i386m3-nat.c i386v4-nat.c i386ly-tdep.c \
  	i386bsd-nat.c i386bsd-tdep.c i386fbsd-nat.c \
! 	i387-tdep.c i387-nat.c \
  	i386-linux-tdep.c i386-nat.c \
  	i960-tdep.c \
  	ia64-linux-nat.c ia64-linux-tdep.c ia64-tdep.c \
--- 1184,1190 ----
  	i386-tdep.c i386b-nat.c i386mach-nat.c i386v-nat.c i386-linux-nat.c \
  	i386aix-nat.c i386m3-nat.c i386v4-nat.c i386ly-tdep.c \
  	i386bsd-nat.c i386bsd-tdep.c i386fbsd-nat.c \
! 	i387-tdep.c \
  	i386-linux-tdep.c i386-nat.c \
  	i960-tdep.c \
  	ia64-linux-nat.c ia64-linux-tdep.c ia64-tdep.c \
***************
*** 1611,1619 ****
  
  i387-tdep.o: i387-tdep.c $(floatformat_h) $(defs_h) $(gdbcore_h) \
  	$(inferior_h) $(language_h) $(regcache_h) $(doublest_h) i386-tdep.h
- 
- i387-nat.o: i387-nat.c $(defs_h) $(inferior_h) $(value_h) $(regcache_h) \
- 	i387-nat.h
  
  i960-tdep.o: i960-tdep.c $(floatformat_h) $(defs_h) $(expression_h) \
  	$(frame_h) $(gdbtypes_h) $(symtab_h) $(value_h) $(gdbcore_h) \
--- 1611,1616 ----
Index: go32-nat.c
===================================================================
RCS file: /cvs/src/src/gdb/go32-nat.c,v
retrieving revision 1.31
diff -c -r1.31 go32-nat.c
*** go32-nat.c	27 Mar 2002 21:35:35 -0000	1.31
--- go32-nat.c	11 Apr 2002 22:43:33 -0000
***************
*** 29,35 ****
  #include "gdbcmd.h"
  #include "floatformat.h"
  #include "buildsym.h"
! #include "i387-nat.h"
  #include "i386-tdep.h"
  #include "value.h"
  #include "regcache.h"
--- 29,35 ----
  #include "gdbcmd.h"
  #include "floatformat.h"
  #include "buildsym.h"
! #include "i387-tdep.h"
  #include "i386-tdep.h"
  #include "value.h"
  #include "regcache.h"
Index: i386-linux-nat.c
===================================================================
RCS file: /cvs/src/src/gdb/i386-linux-nat.c,v
retrieving revision 1.35
diff -c -r1.35 i386-linux-nat.c
*** i386-linux-nat.c	24 Feb 2002 22:14:33 -0000	1.35
--- i386-linux-nat.c	11 Apr 2002 22:43:34 -0000
***************
*** 57,63 ****
  #include "gregset.h"
  
  /* Prototypes for i387_supply_fsave etc.  */
! #include "i387-nat.h"
  
  /* Defines for XMM0_REGNUM etc. */
  #include "i386-tdep.h"
--- 57,63 ----
  #include "gregset.h"
  
  /* Prototypes for i387_supply_fsave etc.  */
! #include "i387-tdep.h"
  
  /* Defines for XMM0_REGNUM etc. */
  #include "i386-tdep.h"
Index: i386bsd-nat.c
===================================================================
RCS file: /cvs/src/src/gdb/i386bsd-nat.c,v
retrieving revision 1.11
diff -c -r1.11 i386bsd-nat.c
*** i386bsd-nat.c	5 Jan 2002 18:36:32 -0000	1.11
--- i386bsd-nat.c	11 Apr 2002 22:43:34 -0000
***************
*** 148,154 ****
        regcache_collect (i, REG_ADDR (gregsetp, i));
  }
  
! #include "i387-nat.h"
  
  /* Fill GDB's register array with the floating-point register values
     in *FPREGSETP.  */
--- 148,154 ----
        regcache_collect (i, REG_ADDR (gregsetp, i));
  }
  
! #include "i387-tdep.h"
  
  /* Fill GDB's register array with the floating-point register values
     in *FPREGSETP.  */
Index: i386gnu-nat.c
===================================================================
RCS file: /cvs/src/src/gdb/i386gnu-nat.c,v
retrieving revision 1.8
diff -c -r1.8 i386gnu-nat.c
*** i386gnu-nat.c	19 Jan 2002 12:51:04 -0000	1.8
--- i386gnu-nat.c	11 Apr 2002 22:43:34 -0000
***************
*** 34,42 ****
  #include <mach/exception.h>
  
  #include "i386-tdep.h"
  
  #include "gnu-nat.h"
- #include "i387-nat.h"
  
  
  /* Offset to the thread_state_t location where REG is stored.  */
--- 34,42 ----
  #include <mach/exception.h>
  
  #include "i386-tdep.h"
+ #include "i387-tdep.h"
  
  #include "gnu-nat.h"
  
  
  /* Offset to the thread_state_t location where REG is stored.  */
Index: i386nbsd-nat.c
===================================================================
RCS file: /cvs/src/src/gdb/i386nbsd-nat.c,v
retrieving revision 1.10
diff -c -r1.10 i386nbsd-nat.c
*** i386nbsd-nat.c	5 Jan 2002 18:36:32 -0000	1.10
--- i386nbsd-nat.c	11 Apr 2002 22:43:34 -0000
***************
*** 39,45 ****
  #include "gregset.h"
  
  /* Prototypes for i387_supply_fsave etc.  */
! #include "i387-nat.h"  
  
  struct md_core
  {
--- 39,45 ----
  #include "gregset.h"
  
  /* Prototypes for i387_supply_fsave etc.  */
! #include "i387-tdep.h"  
  
  struct md_core
  {
Index: i386v4-nat.c
===================================================================
RCS file: /cvs/src/src/gdb/i386v4-nat.c,v
retrieving revision 1.9
diff -c -r1.9 i386v4-nat.c
*** i386v4-nat.c	13 Feb 2002 08:10:30 -0000	1.9
--- i386v4-nat.c	11 Apr 2002 22:43:34 -0000
***************
*** 27,33 ****
  #ifdef HAVE_SYS_REG_H
  #include <sys/reg.h>
  #endif
! #include "i387-nat.h"
  
  
  #ifdef HAVE_SYS_PROCFS_H
--- 27,33 ----
  #ifdef HAVE_SYS_REG_H
  #include <sys/reg.h>
  #endif
! #include "i387-tdep.h"
  
  
  #ifdef HAVE_SYS_PROCFS_H
Index: i387-nat.c
===================================================================
RCS file: i387-nat.c
diff -N i387-nat.c
*** i387-nat.c	27 Dec 2001 15:22:27 -0000	1.9
--- /dev/null	1 Jan 1970 00:00:00 -0000
***************
*** 1,346 ****
- /* Native-dependent code for the i387.
-    Copyright 2000, 2001 Free Software Foundation, Inc.
- 
-    This file is part of GDB.
- 
-    This program is free software; you can redistribute it and/or modify
-    it under the terms of the GNU General Public License as published by
-    the Free Software Foundation; either version 2 of the License, or
-    (at your option) any later version.
- 
-    This program is distributed in the hope that it will be useful,
-    but WITHOUT ANY WARRANTY; without even the implied warranty of
-    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
-    GNU General Public License for more details.
- 
-    You should have received a copy of the GNU General Public License
-    along with this program; if not, write to the Free Software
-    Foundation, Inc., 59 Temple Place - Suite 330,
-    Boston, MA 02111-1307, USA.  */
- 
- #include "defs.h"
- #include "inferior.h"
- #include "value.h"
- #include "regcache.h"
- 
- #include "i387-nat.h"
- #include "i386-tdep.h"
- 
- /* FIXME: kettenis/2000-05-21: Right now more than a few i386 targets
-    define their own routines to manage the floating-point registers in
-    GDB's register array.  Most (if not all) of these targets use the
-    format used by the "fsave" instruction in their communication with
-    the OS.  They should all be converted to use the routines below.  */
- 
- /* At fsave_offset[REGNUM] you'll find the offset to the location in
-    the data structure used by the "fsave" instruction where GDB
-    register REGNUM is stored.  */
- 
- static int fsave_offset[] =
- {
-   28 + 0 * FPU_REG_RAW_SIZE,	/* FP0_REGNUM through ...  */
-   28 + 1 * FPU_REG_RAW_SIZE,  
-   28 + 2 * FPU_REG_RAW_SIZE,  
-   28 + 3 * FPU_REG_RAW_SIZE,  
-   28 + 4 * FPU_REG_RAW_SIZE,  
-   28 + 5 * FPU_REG_RAW_SIZE,  
-   28 + 6 * FPU_REG_RAW_SIZE,  
-   28 + 7 * FPU_REG_RAW_SIZE,	/* ... FP7_REGNUM.  */
-   0,				/* FCTRL_REGNUM (16 bits).  */
-   4,				/* FSTAT_REGNUM (16 bits).  */
-   8,				/* FTAG_REGNUM (16 bits).  */
-   16,				/* FISEG_REGNUM (16 bits).  */
-   12,				/* FIOFF_REGNUM.  */
-   24,				/* FOSEG_REGNUM.  */
-   20,				/* FOOFF_REGNUM.  */
-   18				/* FOP_REGNUM (bottom 11 bits).  */
- };
- 
- #define FSAVE_ADDR(fsave, regnum) (fsave + fsave_offset[regnum - FP0_REGNUM])
- \f
- 
- /* Fill register REGNUM in GDB's register array with the appropriate
-    value from *FSAVE.  This function masks off any of the reserved
-    bits in *FSAVE.  */
- 
- void
- i387_supply_register (int regnum, char *fsave)
- {
-   /* Most of the FPU control registers occupy only 16 bits in
-      the fsave area.  Give those a special treatment.  */
-   if (regnum >= FPC_REGNUM
-       && regnum != FIOFF_REGNUM && regnum != FOOFF_REGNUM)
-     {
-       unsigned int val = *(unsigned short *) (FSAVE_ADDR (fsave, regnum));
- 
-       if (regnum == FOP_REGNUM)
- 	{
- 	  val &= ((1 << 11) - 1);
- 	  supply_register (regnum, (char *) &val);
- 	}
-       else
- 	supply_register (regnum, (char *) &val);
-     }
-   else
-     supply_register (regnum, FSAVE_ADDR (fsave, regnum));
- }
- 
- /* Fill GDB's register array with the floating-point register values
-    in *FSAVE.  This function masks off any of the reserved
-    bits in *FSAVE.  */
- 
- void
- i387_supply_fsave (char *fsave)
- {
-   int i;
- 
-   for (i = FP0_REGNUM; i < XMM0_REGNUM; i++)
-     i387_supply_register (i, fsave);
- }
- 
- /* Fill register REGNUM (if it is a floating-point register) in *FSAVE
-    with the value in GDB's register array.  If REGNUM is -1, do this
-    for all registers.  This function doesn't touch any of the reserved
-    bits in *FSAVE.  */
- 
- void
- i387_fill_fsave (char *fsave, int regnum)
- {
-   int i;
- 
-   for (i = FP0_REGNUM; i < XMM0_REGNUM; i++)
-     if (regnum == -1 || regnum == i)
-       {
- 	/* Most of the FPU control registers occupy only 16 bits in
-            the fsave area.  Give those a special treatment.  */
- 	if (i >= FPC_REGNUM
- 	    && i != FIOFF_REGNUM && i != FOOFF_REGNUM)
- 	  {
- 	    if (i == FOP_REGNUM)
- 	      {
- 		unsigned short oldval, newval;
- 
- 		/* The opcode occupies only 11 bits.  */
- 		oldval = (*(unsigned short *) (FSAVE_ADDR (fsave, i)));
- 		newval = *(unsigned short *) &registers[REGISTER_BYTE (i)];
- 		newval &= ((1 << 11) - 1);
- 		newval |= oldval & ~((1 << 11) - 1);
- 		memcpy (FSAVE_ADDR (fsave, i), &newval, 2);
- 	      }
- 	    else
- 	      memcpy (FSAVE_ADDR (fsave, i), &registers[REGISTER_BYTE (i)], 2);
- 	  }
- 	else
- 	  memcpy (FSAVE_ADDR (fsave, i), &registers[REGISTER_BYTE (i)],
- 		  REGISTER_RAW_SIZE (i));
-       }
- }
- \f
- 
- /* At fxsave_offset[REGNUM] you'll find the offset to the location in
-    the data structure used by the "fxsave" instruction where GDB
-    register REGNUM is stored.  */
- 
- static int fxsave_offset[] =
- {
-   32,				/* FP0_REGNUM through ...  */
-   48,
-   64,
-   80,
-   96,
-   112,
-   128,
-   144,				/* ... FP7_REGNUM (80 bits each).  */
-   0,				/* FCTRL_REGNUM (16 bits).  */
-   2,				/* FSTAT_REGNUM (16 bits).  */
-   4,				/* FTAG_REGNUM (16 bits).  */
-   12,				/* FISEG_REGNUM (16 bits).  */
-   8,				/* FIOFF_REGNUM.  */
-   20,				/* FOSEG_REGNUM (16 bits).  */
-   16,				/* FOOFF_REGNUM.  */
-   6,				/* FOP_REGNUM (bottom 11 bits).  */
-   160,				/* XMM0_REGNUM through ...  */
-   176,
-   192,
-   208,
-   224,
-   240,
-   256,
-   272,				/* ... XMM7_REGNUM (128 bits each).  */
-   24,				/* MXCSR_REGNUM.  */
- };
- 
- #define FXSAVE_ADDR(fxsave, regnum) \
-   (fxsave + fxsave_offset[regnum - FP0_REGNUM])
- 
- static int i387_tag (unsigned char *raw);
- \f
- 
- /* Fill GDB's register array with the floating-point and SSE register
-    values in *FXSAVE.  This function masks off any of the reserved
-    bits in *FXSAVE.  */
- 
- void
- i387_supply_fxsave (char *fxsave)
- {
-   int i;
- 
-   for (i = FP0_REGNUM; i <= MXCSR_REGNUM; i++)
-     {
-       /* Most of the FPU control registers occupy only 16 bits in
- 	 the fxsave area.  Give those a special treatment.  */
-       if (i >= FPC_REGNUM && i < XMM0_REGNUM
- 	  && i != FIOFF_REGNUM && i != FOOFF_REGNUM)
- 	{
- 	  unsigned long val = *(unsigned short *) (FXSAVE_ADDR (fxsave, i));
- 
- 	  if (i == FOP_REGNUM)
- 	    {
- 	      val &= ((1 << 11) - 1);
- 	      supply_register (i, (char *) &val);
- 	    }
- 	  else if (i== FTAG_REGNUM)
- 	    {
- 	      /* The fxsave area contains a simplified version of the
-                  tag word.  We have to look at the actual 80-bit FP
-                  data to recreate the traditional i387 tag word.  */
- 
- 	      unsigned long ftag = 0;
- 	      unsigned long fstat;
- 	      int fpreg;
- 	      int top;
- 
- 	      fstat = *(unsigned short *) (FXSAVE_ADDR (fxsave, FSTAT_REGNUM));
- 	      top = ((fstat >> 11) & 0x7);
- 
- 	      for (fpreg = 7; fpreg >= 0; fpreg--)
- 		{
- 		  int tag;
- 
- 		  if (val & (1 << fpreg))
- 		    {
- 		      int regnum = (fpreg + 8 - top) % 8 + FP0_REGNUM;
- 		      tag = i387_tag (FXSAVE_ADDR (fxsave, regnum));
- 		    }
- 		  else
- 		    tag = 3;		/* Empty */
- 
- 		  ftag |= tag << (2 * fpreg);
- 		}
- 	      supply_register (i, (char *) &ftag);
- 	    }
- 	  else
- 	    supply_register (i, (char *) &val);
- 	}
-       else
- 	supply_register (i, FXSAVE_ADDR (fxsave, i));
-     }
- }
- 
- /* Fill register REGNUM (if it is a floating-point or SSE register) in
-    *FXSAVE with the value in GDB's register array.  If REGNUM is -1, do
-    this for all registers.  This function doesn't touch any of the
-    reserved bits in *FXSAVE.  */
- 
- void
- i387_fill_fxsave (char *fxsave, int regnum)
- {
-   int i;
- 
-   for (i = FP0_REGNUM; i <= MXCSR_REGNUM; i++)
-     if (regnum == -1 || regnum == i)
-       {
- 	/* Most of the FPU control registers occupy only 16 bits in
-            the fxsave area.  Give those a special treatment.  */
- 	if (i >= FPC_REGNUM && i < XMM0_REGNUM
- 	    && i != FIOFF_REGNUM && i != FDOFF_REGNUM)
- 	  {
- 	    if (i == FOP_REGNUM)
- 	      {
- 		unsigned short oldval, newval;
- 
- 		/* The opcode occupies only 11 bits.  */
- 		oldval = (*(unsigned short *) (FXSAVE_ADDR (fxsave, i)));
- 		newval = *(unsigned short *) &registers[REGISTER_BYTE (i)];
- 		newval &= ((1 << 11) - 1);
- 		newval |= oldval & ~((1 << 11) - 1);
- 		memcpy (FXSAVE_ADDR (fxsave, i), &newval, 2);
- 	      }
- 	    else if (i == FTAG_REGNUM)
- 	      {
- 		/* Converting back is much easier.  */
- 
- 		unsigned short val = 0;
- 		unsigned short ftag;
- 		int fpreg;
- 
- 		ftag = *(unsigned short *) &registers[REGISTER_BYTE (i)];
- 
- 		for (fpreg = 7; fpreg >= 0; fpreg--)
- 		  {
- 		    int tag = (ftag >> (fpreg * 2)) & 3;
- 
- 		    if (tag != 3)
- 		      val |= (1 << fpreg);
- 		  }
- 
- 		memcpy (FXSAVE_ADDR (fxsave, i), &val, 2);
- 	      }
- 	    else
- 	      memcpy (FXSAVE_ADDR (fxsave, i),
- 		      &registers[REGISTER_BYTE (i)], 2);
- 	  }
- 	else
- 	  memcpy (FXSAVE_ADDR (fxsave, i), &registers[REGISTER_BYTE (i)],
- 		  REGISTER_RAW_SIZE (i));
-       }
- }
- 
- /* Recreate the FTW (tag word) valid bits from the 80-bit FP data in
-    *RAW.  */
- 
- static int
- i387_tag (unsigned char *raw)
- {
-   int integer;
-   unsigned int exponent;
-   unsigned long fraction[2];
- 
-   integer = raw[7] & 0x80;
-   exponent = (((raw[9] & 0x7f) << 8) | raw[8]);
-   fraction[0] = ((raw[3] << 24) | (raw[2] << 16) | (raw[1] << 8) | raw[0]);
-   fraction[1] = (((raw[7] & 0x7f) << 24) | (raw[6] << 16)
- 		 | (raw[5] << 8) | raw[4]);
- 
-   if (exponent == 0x7fff)
-     {
-       /* Special.  */
-       return (2);
-     }
-   else if (exponent == 0x0000)
-     {
-       if (fraction[0] == 0x0000 && fraction[1] == 0x0000 && !integer)
- 	{
- 	  /* Zero.  */
- 	  return (1);
- 	}
-       else
- 	{
- 	  /* Special.  */
- 	  return (2);
- 	}
-     }
-   else
-     {
-       if (integer)
- 	{
- 	  /* Valid.  */
- 	  return (0);
- 	}
-       else
- 	{
- 	  /* Special.  */
- 	  return (2);
- 	}
-     }
- }
--- 0 ----
Index: i387-nat.h
===================================================================
RCS file: i387-nat.h
diff -N i387-nat.h
*** i387-nat.h	17 Mar 2001 23:02:10 -0000	1.4
--- /dev/null	1 Jan 1970 00:00:00 -0000
***************
*** 1,56 ****
- /* Native-dependent code for the i387.
-    Copyright 2000, 2001 Free Software Foundation, Inc.
- 
-    This file is part of GDB.
- 
-    This program is free software; you can redistribute it and/or modify
-    it under the terms of the GNU General Public License as published by
-    the Free Software Foundation; either version 2 of the License, or
-    (at your option) any later version.
- 
-    This program is distributed in the hope that it will be useful,
-    but WITHOUT ANY WARRANTY; without even the implied warranty of
-    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
-    GNU General Public License for more details.
- 
-    You should have received a copy of the GNU General Public License
-    along with this program; if not, write to the Free Software
-    Foundation, Inc., 59 Temple Place - Suite 330,
-    Boston, MA 02111-1307, USA.  */
- 
- #ifndef I387_NAT_H
- #define I387_NAT_H
- 
- /* Fill register REGNUM in GDB's register array with the appropriate
-    value from *FSAVE.  This function masks off any of the reserved
-    bits in *FSAVE.  */
- 
- extern void i387_supply_register (int regnum, char *fsave);
- 
- /* Fill GDB's register array with the floating-point register values
-    in *FSAVE.  This function masks off any of the reserved
-    bits in *FSAVE.  */
- 
- extern void i387_supply_fsave (char *fsave);
- 
- /* Fill register REGNUM (if it is a floating-point register) in *FSAVE
-    with the value in GDB's register array.  If REGNUM is -1, do this
-    for all registers.  This function doesn't touch any of the reserved
-    bits in *FSAVE.  */
- 
- extern void i387_fill_fsave (char *fsave, int regnum);
- 
- /* Fill GDB's register array with the floating-point and SSE register
-    values in *FXSAVE.  This function masks off any of the reserved
-    bits in *FXSAVE.  */
- 
- extern void i387_supply_fxsave (char *fxsave);
- 
- /* Fill register REGNUM (if it is a floating-point or SSE register) in
-    *FXSAVE with the value in GDB's register array.  If REGNUM is -1, do
-    this for all registers.  This function doesn't touch any of the
-    reserved bits in *FXSAVE.  */
- 
- extern void i387_fill_fxsave (char *fxsave, int regnum);
- 
- #endif /* i387-nat.h */
--- 0 ----
Index: i387-tdep.c
===================================================================
RCS file: /cvs/src/src/gdb/i387-tdep.c,v
retrieving revision 1.16
diff -c -r1.16 i387-tdep.c
*** i387-tdep.c	12 Nov 2001 22:27:35 -0000	1.16
--- i387-tdep.c	11 Apr 2002 22:43:34 -0000
***************
*** 388,390 ****
--- 388,709 ----
    printf_filtered ("Opcode:              %s\n",
  		   local_hex_string_custom (fop ? (fop | 0xd800) : 0, "04"));
  }
+ 
+ /* FIXME: kettenis/2000-05-21: Right now more than a few i386 targets
+    define their own routines to manage the floating-point registers in
+    GDB's register array.  Most (if not all) of these targets use the
+    format used by the "fsave" instruction in their communication with
+    the OS.  They should all be converted to use the routines below.  */
+ 
+ /* At fsave_offset[REGNUM] you'll find the offset to the location in
+    the data structure used by the "fsave" instruction where GDB
+    register REGNUM is stored.  */
+ 
+ static int fsave_offset[] =
+ {
+   28 + 0 * FPU_REG_RAW_SIZE,	/* FP0_REGNUM through ...  */
+   28 + 1 * FPU_REG_RAW_SIZE,  
+   28 + 2 * FPU_REG_RAW_SIZE,  
+   28 + 3 * FPU_REG_RAW_SIZE,  
+   28 + 4 * FPU_REG_RAW_SIZE,  
+   28 + 5 * FPU_REG_RAW_SIZE,  
+   28 + 6 * FPU_REG_RAW_SIZE,  
+   28 + 7 * FPU_REG_RAW_SIZE,	/* ... FP7_REGNUM.  */
+   0,				/* FCTRL_REGNUM (16 bits).  */
+   4,				/* FSTAT_REGNUM (16 bits).  */
+   8,				/* FTAG_REGNUM (16 bits).  */
+   16,				/* FISEG_REGNUM (16 bits).  */
+   12,				/* FIOFF_REGNUM.  */
+   24,				/* FOSEG_REGNUM.  */
+   20,				/* FOOFF_REGNUM.  */
+   18				/* FOP_REGNUM (bottom 11 bits).  */
+ };
+ 
+ #define FSAVE_ADDR(fsave, regnum) (fsave + fsave_offset[regnum - FP0_REGNUM])
+ \f
+ 
+ /* Fill register REGNUM in GDB's register array with the appropriate
+    value from *FSAVE.  This function masks off any of the reserved
+    bits in *FSAVE.  */
+ 
+ void
+ i387_supply_register (int regnum, char *fsave)
+ {
+   /* Most of the FPU control registers occupy only 16 bits in
+      the fsave area.  Give those a special treatment.  */
+   if (regnum >= FPC_REGNUM
+       && regnum != FIOFF_REGNUM && regnum != FOOFF_REGNUM)
+     {
+       unsigned int val = *(unsigned short *) (FSAVE_ADDR (fsave, regnum));
+ 
+       if (regnum == FOP_REGNUM)
+ 	{
+ 	  val &= ((1 << 11) - 1);
+ 	  supply_register (regnum, (char *) &val);
+ 	}
+       else
+ 	supply_register (regnum, (char *) &val);
+     }
+   else
+     supply_register (regnum, FSAVE_ADDR (fsave, regnum));
+ }
+ 
+ /* Fill GDB's register array with the floating-point register values
+    in *FSAVE.  This function masks off any of the reserved
+    bits in *FSAVE.  */
+ 
+ void
+ i387_supply_fsave (char *fsave)
+ {
+   int i;
+ 
+   for (i = FP0_REGNUM; i < XMM0_REGNUM; i++)
+     i387_supply_register (i, fsave);
+ }
+ 
+ /* Fill register REGNUM (if it is a floating-point register) in *FSAVE
+    with the value in GDB's register array.  If REGNUM is -1, do this
+    for all registers.  This function doesn't touch any of the reserved
+    bits in *FSAVE.  */
+ 
+ void
+ i387_fill_fsave (char *fsave, int regnum)
+ {
+   int i;
+ 
+   for (i = FP0_REGNUM; i < XMM0_REGNUM; i++)
+     if (regnum == -1 || regnum == i)
+       {
+ 	/* Most of the FPU control registers occupy only 16 bits in
+            the fsave area.  Give those a special treatment.  */
+ 	if (i >= FPC_REGNUM
+ 	    && i != FIOFF_REGNUM && i != FOOFF_REGNUM)
+ 	  {
+ 	    if (i == FOP_REGNUM)
+ 	      {
+ 		unsigned short oldval, newval;
+ 
+ 		/* The opcode occupies only 11 bits.  */
+ 		oldval = (*(unsigned short *) (FSAVE_ADDR (fsave, i)));
+ 		newval = *(unsigned short *) &registers[REGISTER_BYTE (i)];
+ 		newval &= ((1 << 11) - 1);
+ 		newval |= oldval & ~((1 << 11) - 1);
+ 		memcpy (FSAVE_ADDR (fsave, i), &newval, 2);
+ 	      }
+ 	    else
+ 	      memcpy (FSAVE_ADDR (fsave, i), &registers[REGISTER_BYTE (i)], 2);
+ 	  }
+ 	else
+ 	  memcpy (FSAVE_ADDR (fsave, i), &registers[REGISTER_BYTE (i)],
+ 		  REGISTER_RAW_SIZE (i));
+       }
+ }
+ \f
+ 
+ /* At fxsave_offset[REGNUM] you'll find the offset to the location in
+    the data structure used by the "fxsave" instruction where GDB
+    register REGNUM is stored.  */
+ 
+ static int fxsave_offset[] =
+ {
+   32,				/* FP0_REGNUM through ...  */
+   48,
+   64,
+   80,
+   96,
+   112,
+   128,
+   144,				/* ... FP7_REGNUM (80 bits each).  */
+   0,				/* FCTRL_REGNUM (16 bits).  */
+   2,				/* FSTAT_REGNUM (16 bits).  */
+   4,				/* FTAG_REGNUM (16 bits).  */
+   12,				/* FISEG_REGNUM (16 bits).  */
+   8,				/* FIOFF_REGNUM.  */
+   20,				/* FOSEG_REGNUM (16 bits).  */
+   16,				/* FOOFF_REGNUM.  */
+   6,				/* FOP_REGNUM (bottom 11 bits).  */
+   160,				/* XMM0_REGNUM through ...  */
+   176,
+   192,
+   208,
+   224,
+   240,
+   256,
+   272,				/* ... XMM7_REGNUM (128 bits each).  */
+   24,				/* MXCSR_REGNUM.  */
+ };
+ 
+ #define FXSAVE_ADDR(fxsave, regnum) \
+   (fxsave + fxsave_offset[regnum - FP0_REGNUM])
+ 
+ static int i387_tag (unsigned char *raw);
+ \f
+ 
+ /* Fill GDB's register array with the floating-point and SSE register
+    values in *FXSAVE.  This function masks off any of the reserved
+    bits in *FXSAVE.  */
+ 
+ void
+ i387_supply_fxsave (char *fxsave)
+ {
+   int i;
+ 
+   for (i = FP0_REGNUM; i <= MXCSR_REGNUM; i++)
+     {
+       /* Most of the FPU control registers occupy only 16 bits in
+ 	 the fxsave area.  Give those a special treatment.  */
+       if (i >= FPC_REGNUM && i < XMM0_REGNUM
+ 	  && i != FIOFF_REGNUM && i != FOOFF_REGNUM)
+ 	{
+ 	  unsigned long val = *(unsigned short *) (FXSAVE_ADDR (fxsave, i));
+ 
+ 	  if (i == FOP_REGNUM)
+ 	    {
+ 	      val &= ((1 << 11) - 1);
+ 	      supply_register (i, (char *) &val);
+ 	    }
+ 	  else if (i== FTAG_REGNUM)
+ 	    {
+ 	      /* The fxsave area contains a simplified version of the
+                  tag word.  We have to look at the actual 80-bit FP
+                  data to recreate the traditional i387 tag word.  */
+ 
+ 	      unsigned long ftag = 0;
+ 	      unsigned long fstat;
+ 	      int fpreg;
+ 	      int top;
+ 
+ 	      fstat = *(unsigned short *) (FXSAVE_ADDR (fxsave, FSTAT_REGNUM));
+ 	      top = ((fstat >> 11) & 0x7);
+ 
+ 	      for (fpreg = 7; fpreg >= 0; fpreg--)
+ 		{
+ 		  int tag;
+ 
+ 		  if (val & (1 << fpreg))
+ 		    {
+ 		      int regnum = (fpreg + 8 - top) % 8 + FP0_REGNUM;
+ 		      tag = i387_tag (FXSAVE_ADDR (fxsave, regnum));
+ 		    }
+ 		  else
+ 		    tag = 3;		/* Empty */
+ 
+ 		  ftag |= tag << (2 * fpreg);
+ 		}
+ 	      supply_register (i, (char *) &ftag);
+ 	    }
+ 	  else
+ 	    supply_register (i, (char *) &val);
+ 	}
+       else
+ 	supply_register (i, FXSAVE_ADDR (fxsave, i));
+     }
+ }
+ 
+ /* Fill register REGNUM (if it is a floating-point or SSE register) in
+    *FXSAVE with the value in GDB's register array.  If REGNUM is -1, do
+    this for all registers.  This function doesn't touch any of the
+    reserved bits in *FXSAVE.  */
+ 
+ void
+ i387_fill_fxsave (char *fxsave, int regnum)
+ {
+   int i;
+ 
+   for (i = FP0_REGNUM; i <= MXCSR_REGNUM; i++)
+     if (regnum == -1 || regnum == i)
+       {
+ 	/* Most of the FPU control registers occupy only 16 bits in
+            the fxsave area.  Give those a special treatment.  */
+ 	if (i >= FPC_REGNUM && i < XMM0_REGNUM
+ 	    && i != FIOFF_REGNUM && i != FDOFF_REGNUM)
+ 	  {
+ 	    if (i == FOP_REGNUM)
+ 	      {
+ 		unsigned short oldval, newval;
+ 
+ 		/* The opcode occupies only 11 bits.  */
+ 		oldval = (*(unsigned short *) (FXSAVE_ADDR (fxsave, i)));
+ 		newval = *(unsigned short *) &registers[REGISTER_BYTE (i)];
+ 		newval &= ((1 << 11) - 1);
+ 		newval |= oldval & ~((1 << 11) - 1);
+ 		memcpy (FXSAVE_ADDR (fxsave, i), &newval, 2);
+ 	      }
+ 	    else if (i == FTAG_REGNUM)
+ 	      {
+ 		/* Converting back is much easier.  */
+ 
+ 		unsigned short val = 0;
+ 		unsigned short ftag;
+ 		int fpreg;
+ 
+ 		ftag = *(unsigned short *) &registers[REGISTER_BYTE (i)];
+ 
+ 		for (fpreg = 7; fpreg >= 0; fpreg--)
+ 		  {
+ 		    int tag = (ftag >> (fpreg * 2)) & 3;
+ 
+ 		    if (tag != 3)
+ 		      val |= (1 << fpreg);
+ 		  }
+ 
+ 		memcpy (FXSAVE_ADDR (fxsave, i), &val, 2);
+ 	      }
+ 	    else
+ 	      memcpy (FXSAVE_ADDR (fxsave, i),
+ 		      &registers[REGISTER_BYTE (i)], 2);
+ 	  }
+ 	else
+ 	  memcpy (FXSAVE_ADDR (fxsave, i), &registers[REGISTER_BYTE (i)],
+ 		  REGISTER_RAW_SIZE (i));
+       }
+ }
+ 
+ /* Recreate the FTW (tag word) valid bits from the 80-bit FP data in
+    *RAW.  */
+ 
+ static int
+ i387_tag (unsigned char *raw)
+ {
+   int integer;
+   unsigned int exponent;
+   unsigned long fraction[2];
+ 
+   integer = raw[7] & 0x80;
+   exponent = (((raw[9] & 0x7f) << 8) | raw[8]);
+   fraction[0] = ((raw[3] << 24) | (raw[2] << 16) | (raw[1] << 8) | raw[0]);
+   fraction[1] = (((raw[7] & 0x7f) << 24) | (raw[6] << 16)
+ 		 | (raw[5] << 8) | raw[4]);
+ 
+   if (exponent == 0x7fff)
+     {
+       /* Special.  */
+       return (2);
+     }
+   else if (exponent == 0x0000)
+     {
+       if (fraction[0] == 0x0000 && fraction[1] == 0x0000 && !integer)
+ 	{
+ 	  /* Zero.  */
+ 	  return (1);
+ 	}
+       else
+ 	{
+ 	  /* Special.  */
+ 	  return (2);
+ 	}
+     }
+   else
+     {
+       if (integer)
+ 	{
+ 	  /* Valid.  */
+ 	  return (0);
+ 	}
+       else
+ 	{
+ 	  /* Special.  */
+ 	  return (2);
+ 	}
+     }
+ }
Index: i387-tdep.h
===================================================================
RCS file: i387-tdep.h
diff -N i387-tdep.h
*** /dev/null	1 Jan 1970 00:00:00 -0000
--- i387-tdep.h	11 Apr 2002 22:43:34 -0000
***************
*** 0 ****
--- 1,56 ----
+ /* Target-dependent code for the i387.
+    Copyright 2000, 2001 Free Software Foundation, Inc.
+ 
+    This file is part of GDB.
+ 
+    This program is free software; you can redistribute it and/or modify
+    it under the terms of the GNU General Public License as published by
+    the Free Software Foundation; either version 2 of the License, or
+    (at your option) any later version.
+ 
+    This program is distributed in the hope that it will be useful,
+    but WITHOUT ANY WARRANTY; without even the implied warranty of
+    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+    GNU General Public License for more details.
+ 
+    You should have received a copy of the GNU General Public License
+    along with this program; if not, write to the Free Software
+    Foundation, Inc., 59 Temple Place - Suite 330,
+    Boston, MA 02111-1307, USA.  */
+ 
+ #ifndef I387_TDEP_H
+ #define I387_TDEP_H
+ 
+ /* Fill register REGNUM in GDB's register array with the appropriate
+    value from *FSAVE.  This function masks off any of the reserved
+    bits in *FSAVE.  */
+ 
+ extern void i387_supply_register (int regnum, char *fsave);
+ 
+ /* Fill GDB's register array with the floating-point register values
+    in *FSAVE.  This function masks off any of the reserved
+    bits in *FSAVE.  */
+ 
+ extern void i387_supply_fsave (char *fsave);
+ 
+ /* Fill register REGNUM (if it is a floating-point register) in *FSAVE
+    with the value in GDB's register array.  If REGNUM is -1, do this
+    for all registers.  This function doesn't touch any of the reserved
+    bits in *FSAVE.  */
+ 
+ extern void i387_fill_fsave (char *fsave, int regnum);
+ 
+ /* Fill GDB's register array with the floating-point and SSE register
+    values in *FXSAVE.  This function masks off any of the reserved
+    bits in *FXSAVE.  */
+ 
+ extern void i387_supply_fxsave (char *fxsave);
+ 
+ /* Fill register REGNUM (if it is a floating-point or SSE register) in
+    *FXSAVE with the value in GDB's register array.  If REGNUM is -1, do
+    this for all registers.  This function doesn't touch any of the
+    reserved bits in *FXSAVE.  */
+ 
+ extern void i387_fill_fxsave (char *fxsave, int regnum);
+ 
+ #endif /* i387-tdep.h */
Index: x86-64-linux-nat.c
===================================================================
RCS file: /cvs/src/src/gdb/x86-64-linux-nat.c,v
retrieving revision 1.9
diff -c -r1.9 x86-64-linux-nat.c
*** x86-64-linux-nat.c	10 Apr 2002 12:12:33 -0000	1.9
--- x86-64-linux-nat.c	11 Apr 2002 22:43:35 -0000
***************
*** 25,31 ****
  #include "inferior.h"
  #include "gdbcore.h"
  #include "regcache.h"
! #include "i387-nat.h"
  #include "gdb_assert.h"
  #include "x86-64-tdep.h"
  
--- 25,31 ----
  #include "inferior.h"
  #include "gdbcore.h"
  #include "regcache.h"
! #include "i387-tdep.h"
  #include "gdb_assert.h"
  #include "x86-64-tdep.h"
  
Index: config/i386/fbsd.mh
===================================================================
RCS file: /cvs/src/src/gdb/config/i386/fbsd.mh,v
retrieving revision 1.11
diff -c -r1.11 fbsd.mh
*** config/i386/fbsd.mh	31 Mar 2002 17:47:17 -0000	1.11
--- config/i386/fbsd.mh	11 Apr 2002 22:43:36 -0000
***************
*** 4,7 ****
  
  NAT_FILE= nm-fbsd.h
  # NOTE: Do not spread NATDEPFILES over several lines - it hurts BSD make.
! NATDEPFILES= fork-child.o infptrace.o inftarg.o solib.o solib-svr4.o solib-legacy.o corelow.o core-aout.o core-regset.o i386-nat.o i387-nat.o i386bsd-nat.o i386fbsd-nat.o gcore.o fbsd-proc.o
--- 4,7 ----
  
  NAT_FILE= nm-fbsd.h
  # NOTE: Do not spread NATDEPFILES over several lines - it hurts BSD make.
! NATDEPFILES= fork-child.o infptrace.o inftarg.o solib.o solib-svr4.o solib-legacy.o corelow.o core-aout.o core-regset.o i386-nat.o i386bsd-nat.o i386fbsd-nat.o gcore.o fbsd-proc.o
Index: config/i386/go32.mh
===================================================================
RCS file: /cvs/src/src/gdb/config/i386/go32.mh,v
retrieving revision 1.5
diff -c -r1.5 go32.mh
*** config/i386/go32.mh	18 Jan 2002 04:50:59 -0000	1.5
--- config/i386/go32.mh	11 Apr 2002 22:43:36 -0000
***************
*** 7,13 ****
  XM_FILE= xm-go32.h
  
  NAT_FILE= nm-go32.h
! NATDEPFILES= go32-nat.o i386-nat.o i387-nat.o
  
  TERMCAP=
  HOST_IPC=
--- 7,13 ----
  XM_FILE= xm-go32.h
  
  NAT_FILE= nm-go32.h
! NATDEPFILES= go32-nat.o i386-nat.o
  
  TERMCAP=
  HOST_IPC=
Index: config/i386/i386gnu.mh
===================================================================
RCS file: /cvs/src/src/gdb/config/i386/i386gnu.mh,v
retrieving revision 1.6
diff -c -r1.6 i386gnu.mh
*** config/i386/i386gnu.mh	18 Jan 2002 04:50:59 -0000	1.6
--- config/i386/i386gnu.mh	11 Apr 2002 22:43:36 -0000
***************
*** 1,5 ****
  # Host: Intel 386 running the GNU Hurd
! NATDEPFILES= i386gnu-nat.o i387-nat.o gnu-nat.o fork-child.o solib.o solib-svr4.o solib-legacy.o corelow.o notify_S.o process_reply_S.o msg_reply_S.o msg_U.o exc_request_U.o exc_request_S.o
  XM_FILE= xm-i386gnu.h
  NAT_FILE= nm-gnu.h
  MH_CFLAGS = -D_GNU_SOURCE
--- 1,5 ----
  # Host: Intel 386 running the GNU Hurd
! NATDEPFILES= i386gnu-nat.o gnu-nat.o fork-child.o solib.o solib-svr4.o solib-legacy.o corelow.o notify_S.o process_reply_S.o msg_reply_S.o msg_U.o exc_request_U.o exc_request_S.o
  XM_FILE= xm-i386gnu.h
  NAT_FILE= nm-gnu.h
  MH_CFLAGS = -D_GNU_SOURCE
Index: config/i386/i386sol2.mh
===================================================================
RCS file: /cvs/src/src/gdb/config/i386/i386sol2.mh,v
retrieving revision 1.4
diff -c -r1.4 i386sol2.mh
*** config/i386/i386sol2.mh	14 Feb 2002 08:13:33 -0000	1.4
--- config/i386/i386sol2.mh	11 Apr 2002 22:43:36 -0000
***************
*** 4,8 ****
  XM_CLIBS= -lsocket -lnsl
  
  NAT_FILE= nm-i386sol2.h
! NATDEPFILES= core-regset.o fork-child.o i386v4-nat.o i387-nat.o corelow.o \
  	procfs.o proc-api.o proc-events.o proc-flags.o proc-why.o gcore.o
--- 4,8 ----
  XM_CLIBS= -lsocket -lnsl
  
  NAT_FILE= nm-i386sol2.h
! NATDEPFILES= core-regset.o fork-child.o i386v4-nat.o corelow.o \
  	procfs.o proc-api.o proc-events.o proc-flags.o proc-why.o gcore.o
Index: config/i386/i386v42mp.mh
===================================================================
RCS file: /cvs/src/src/gdb/config/i386/i386v42mp.mh,v
retrieving revision 1.7
diff -c -r1.7 i386v42mp.mh
*** config/i386/i386v42mp.mh	11 Apr 2002 18:13:21 -0000	1.7
--- config/i386/i386v42mp.mh	11 Apr 2002 22:43:36 -0000
***************
*** 14,20 ****
  # continuation character (backslash) to extend a commented line.  As a
  # consequence, make considers subsequent tab-indented lines to be
  # some sort of error.
! NATDEPFILES= corelow.o core-regset.o fork-child.o i386v4-nat.o i387-nat.o \
  	gcore.o solib.o solib-svr4.o solib-legacy.o procfs.o proc-api.o \
  	proc-events.o proc-flags.o proc-why.o uw-thread.o
  
--- 14,20 ----
  # continuation character (backslash) to extend a commented line.  As a
  # consequence, make considers subsequent tab-indented lines to be
  # some sort of error.
! NATDEPFILES= corelow.o core-regset.o fork-child.o i386v4-nat.o \
  	gcore.o solib.o solib-svr4.o solib-legacy.o procfs.o proc-api.o \
  	proc-events.o proc-flags.o proc-why.o uw-thread.o
  
Index: config/i386/linux.mh
===================================================================
RCS file: /cvs/src/src/gdb/config/i386/linux.mh,v
retrieving revision 1.11
diff -c -r1.11 linux.mh
*** config/i386/linux.mh	18 Jan 2002 04:50:59 -0000	1.11
--- config/i386/linux.mh	11 Apr 2002 22:43:36 -0000
***************
*** 4,10 ****
  
  NAT_FILE= nm-linux.h
  NATDEPFILES= infptrace.o inftarg.o fork-child.o corelow.o linux-proc.o \
! 	core-aout.o i386-nat.o i386-linux-nat.o i387-nat.o \
  	proc-service.o thread-db.o lin-lwp.o linux-proc.o gcore.o
  
  # The dynamically loaded libthread_db needs access to symbols in the
--- 4,10 ----
  
  NAT_FILE= nm-linux.h
  NATDEPFILES= infptrace.o inftarg.o fork-child.o corelow.o linux-proc.o \
! 	core-aout.o i386-nat.o i386-linux-nat.o \
  	proc-service.o thread-db.o lin-lwp.o linux-proc.o gcore.o
  
  # The dynamically loaded libthread_db needs access to symbols in the
Index: config/i386/nbsd.mh
===================================================================
RCS file: /cvs/src/src/gdb/config/i386/nbsd.mh,v
retrieving revision 1.12
diff -c -r1.12 nbsd.mh
*** config/i386/nbsd.mh	18 Jan 2002 04:50:59 -0000	1.12
--- config/i386/nbsd.mh	11 Apr 2002 22:43:36 -0000
***************
*** 6,12 ****
  # continuation character (backslash) to extend a commented line.  As a
  # consequence, BSD make considers subsequent tab-indented lines to be
  # "unassociated shell commands".
! NATDEPFILES= fork-child.o infptrace.o inftarg.o corelow.o i387-nat.o i386bsd-nat.o i386nbsd-nat.o solib.o solib-sunos.o
  
  XM_FILE= xm-nbsd.h
  NAT_FILE= nm-nbsd.h
--- 6,12 ----
  # continuation character (backslash) to extend a commented line.  As a
  # consequence, BSD make considers subsequent tab-indented lines to be
  # "unassociated shell commands".
! NATDEPFILES= fork-child.o infptrace.o inftarg.o corelow.o i386bsd-nat.o i386nbsd-nat.o solib.o solib-sunos.o
  
  XM_FILE= xm-nbsd.h
  NAT_FILE= nm-nbsd.h
Index: config/i386/nbsdelf.mh
===================================================================
RCS file: /cvs/src/src/gdb/config/i386/nbsdelf.mh,v
retrieving revision 1.8
diff -c -r1.8 nbsdelf.mh
*** config/i386/nbsdelf.mh	18 Jan 2002 04:50:59 -0000	1.8
--- config/i386/nbsdelf.mh	11 Apr 2002 22:43:36 -0000
***************
*** 1,4 ****
  # Host: Intel 386 running NetBSD
! NATDEPFILES= fork-child.o infptrace.o inftarg.o corelow.o i387-nat.o i386bsd-nat.o i386nbsd-nat.o solib.o solib-svr4.o solib-legacy.o
  XM_FILE= xm-nbsd.h
  NAT_FILE= nm-nbsdelf.h
--- 1,4 ----
  # Host: Intel 386 running NetBSD
! NATDEPFILES= fork-child.o infptrace.o inftarg.o corelow.o i386bsd-nat.o i386nbsd-nat.o solib.o solib-svr4.o solib-legacy.o
  XM_FILE= xm-nbsd.h
  NAT_FILE= nm-nbsdelf.h
Index: config/i386/obsd.mh
===================================================================
RCS file: /cvs/src/src/gdb/config/i386/obsd.mh,v
retrieving revision 1.2
diff -c -r1.2 obsd.mh
*** config/i386/obsd.mh	18 Jan 2002 04:50:59 -0000	1.2
--- config/i386/obsd.mh	11 Apr 2002 22:43:36 -0000
***************
*** 4,7 ****
  
  NAT_FILE= nm-obsd.h
  # NOTE: Do not spread NATDEPFILES over several lines - it hurts BSD make.
! NATDEPFILES= fork-child.o infptrace.o inftarg.o solib.o solib-sunos.o corelow.o core-aout.o i386-nat.o i387-nat.o i386bsd-nat.o
--- 4,7 ----
  
  NAT_FILE= nm-obsd.h
  # NOTE: Do not spread NATDEPFILES over several lines - it hurts BSD make.
! NATDEPFILES= fork-child.o infptrace.o inftarg.o solib.o solib-sunos.o corelow.o core-aout.o i386-nat.o i386bsd-nat.o
Index: config/i386/x86-64linux.mh
===================================================================
RCS file: /cvs/src/src/gdb/config/i386/x86-64linux.mh,v
retrieving revision 1.5
diff -c -r1.5 x86-64linux.mh
*** config/i386/x86-64linux.mh	21 Feb 2002 12:48:36 -0000	1.5
--- config/i386/x86-64linux.mh	11 Apr 2002 22:43:36 -0000
***************
*** 5,11 ****
  NAT_FILE= nm-x86-64.h
  NATDEPFILES= infptrace.o inftarg.o fork-child.o corelow.o \
  	core-aout.o i386-nat.o x86-64-linux-nat.o \
! 	i387-nat.o proc-service.o thread-db.o lin-lwp.o \
  	linux-proc.o gcore.o 
  
  LOADLIBES = -ldl -rdynamic
--- 5,11 ----
  NAT_FILE= nm-x86-64.h
  NATDEPFILES= infptrace.o inftarg.o fork-child.o corelow.o \
  	core-aout.o i386-nat.o x86-64-linux-nat.o \
! 	proc-service.o thread-db.o lin-lwp.o \
  	linux-proc.o gcore.o 
  
  LOADLIBES = -ldl -rdynamic

^ permalink raw reply	[flat|nested] 8+ messages in thread
* [PATCH/RFA] Move i387 register supply/fill to i387-tdep.c
@ 2002-05-09 16:45 Jason R Thorpe
  2002-05-09 17:12 ` Daniel Jacobowitz
  2002-05-10  0:26 ` Eli Zaretskii
  0 siblings, 2 replies; 8+ messages in thread
From: Jason R Thorpe @ 2002-05-09 16:45 UTC (permalink / raw)
  To: gdb-patches

[-- Attachment #1: Type: text/plain, Size: 1507 bytes --]

This is a re-submit of an earlier patch, which takes into account
Mark's recent changes to i387-nat.c.

Basically, this moves all of the register supply/fill stuff for i387
from i387-nat.c to i387-tdep.c, to facilitate cross-debugging of core
files.  This renders i387-nat.c empty, and so it is removed in this
patch.

OK to commit?

        * Makefile.in (ALLDEPFILES): Remove i387-nat.c.
        (i387-nat.o): Delete rule.
        * i387-nat.c: Delete file, moving contents to...
        * i387-tdep.c: ...here.
        * i387-nat.h: Rename...
        * i387-tdep.h: ...to this.
        * go32-nat.c: Include i387-tdep.h instead of i387-nat.h.
        * i386-linux-nat.c: Likewise.
        * i386bsd-nat.c: Likewise.
        * i386gnu-nat.c: Likewise. 
        * i386nbsd-nat.c: Likewise.
        * i386v4-nat.c: Likewise.
        * x86-64-linux-nat.c: Likewise.
        * config/i386/fbsd.mh (NATDEPFILES): Remove i387-nat.o.
        * config/i386/go32.mh (NATDEPFILES): Likewise.
        * config/i386/i386gnu.mh (NATDEPFILES): Likewise.
        * config/i386/i386sol2.mh (NATDEPFILES): Likewise.
        * config/i386/i386v42mp.mh (NATDEPFILES): Likewise.
        * config/i386/linux.mh (NATDEPFILES): Likewise.
        * config/i386/nbsd.mh (NATDEPFILES): Likewise.
        * config/i386/nbsdelf.mh (NATDEPFILES): Likewise.
        * config/i386/obsd.mh (NATDEPFILES): Likewise.
        * config/i386/x86-64linux.mh (NATDEPFILES): Likewise.

-- 
        -- Jason R. Thorpe <thorpej@wasabisystems.com>

[-- Attachment #2: i387-patch --]
[-- Type: text/plain, Size: 33399 bytes --]

Index: go32-nat.c
===================================================================
RCS file: /cvs/src/src/gdb/go32-nat.c,v
retrieving revision 1.32
diff -u -r1.32 go32-nat.c
--- go32-nat.c	19 Apr 2002 19:28:48 -0000	1.32
+++ go32-nat.c	9 May 2002 23:33:58 -0000
@@ -29,7 +29,7 @@
 #include "gdbcmd.h"
 #include "floatformat.h"
 #include "buildsym.h"
-#include "i387-nat.h"
+#include "i387-tdep.h"
 #include "i386-tdep.h"
 #include "value.h"
 #include "regcache.h"
Index: i386-linux-nat.c
===================================================================
RCS file: /cvs/src/src/gdb/i386-linux-nat.c,v
retrieving revision 1.35
diff -u -r1.35 i386-linux-nat.c
--- i386-linux-nat.c	24 Feb 2002 22:14:33 -0000	1.35
+++ i386-linux-nat.c	9 May 2002 23:33:58 -0000
@@ -57,7 +57,7 @@
 #include "gregset.h"
 
 /* Prototypes for i387_supply_fsave etc.  */
-#include "i387-nat.h"
+#include "i387-tdep.h"
 
 /* Defines for XMM0_REGNUM etc. */
 #include "i386-tdep.h"
Index: i386bsd-nat.c
===================================================================
RCS file: /cvs/src/src/gdb/i386bsd-nat.c,v
retrieving revision 1.11
diff -u -r1.11 i386bsd-nat.c
--- i386bsd-nat.c	5 Jan 2002 18:36:32 -0000	1.11
+++ i386bsd-nat.c	9 May 2002 23:33:58 -0000
@@ -148,7 +148,7 @@
       regcache_collect (i, REG_ADDR (gregsetp, i));
 }
 
-#include "i387-nat.h"
+#include "i387-tdep.h"
 
 /* Fill GDB's register array with the floating-point register values
    in *FPREGSETP.  */
Index: i386gnu-nat.c
===================================================================
RCS file: /cvs/src/src/gdb/i386gnu-nat.c,v
retrieving revision 1.9
diff -u -r1.9 i386gnu-nat.c
--- i386gnu-nat.c	24 Apr 2002 21:44:46 -0000	1.9
+++ i386gnu-nat.c	9 May 2002 23:33:59 -0000
@@ -34,9 +34,9 @@
 #include <mach/exception.h>
 
 #include "i386-tdep.h"
+#include "i387-tdep.h"
 
 #include "gnu-nat.h"
-#include "i387-nat.h"
 
 #ifdef HAVE_SYS_PROCFS_H
 # include <sys/procfs.h>
Index: i386nbsd-nat.c
===================================================================
RCS file: /cvs/src/src/gdb/i386nbsd-nat.c,v
retrieving revision 1.10
diff -u -r1.10 i386nbsd-nat.c
--- i386nbsd-nat.c	5 Jan 2002 18:36:32 -0000	1.10
+++ i386nbsd-nat.c	9 May 2002 23:33:59 -0000
@@ -39,7 +39,7 @@
 #include "gregset.h"
 
 /* Prototypes for i387_supply_fsave etc.  */
-#include "i387-nat.h"  
+#include "i387-tdep.h"  
 
 struct md_core
 {
Index: i386v4-nat.c
===================================================================
RCS file: /cvs/src/src/gdb/i386v4-nat.c,v
retrieving revision 1.9
diff -u -r1.9 i386v4-nat.c
--- i386v4-nat.c	13 Feb 2002 08:10:30 -0000	1.9
+++ i386v4-nat.c	9 May 2002 23:33:59 -0000
@@ -27,7 +27,7 @@
 #ifdef HAVE_SYS_REG_H
 #include <sys/reg.h>
 #endif
-#include "i387-nat.h"
+#include "i387-tdep.h"
 
 
 #ifdef HAVE_SYS_PROCFS_H
Index: i387-nat.c
===================================================================
RCS file: i387-nat.c
diff -N i387-nat.c
--- i387-nat.c	6 May 2002 11:39:51 -0000	1.10
+++ /dev/null	1 Jan 1970 00:00:00 -0000
@@ -1,351 +0,0 @@
-/* Native-dependent code for the i387.
-   Copyright 2000, 2001 Free Software Foundation, Inc.
-
-   This file is part of GDB.
-
-   This program is free software; you can redistribute it and/or modify
-   it under the terms of the GNU General Public License as published by
-   the Free Software Foundation; either version 2 of the License, or
-   (at your option) any later version.
-
-   This program is distributed in the hope that it will be useful,
-   but WITHOUT ANY WARRANTY; without even the implied warranty of
-   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
-   GNU General Public License for more details.
-
-   You should have received a copy of the GNU General Public License
-   along with this program; if not, write to the Free Software
-   Foundation, Inc., 59 Temple Place - Suite 330,
-   Boston, MA 02111-1307, USA.  */
-
-#include "defs.h"
-#include "inferior.h"
-#include "value.h"
-#include "regcache.h"
-
-#include "i387-nat.h"
-#include "i386-tdep.h"
-
-/* FIXME: kettenis/2000-05-21: Right now more than a few i386 targets
-   define their own routines to manage the floating-point registers in
-   GDB's register array.  Most (if not all) of these targets use the
-   format used by the "fsave" instruction in their communication with
-   the OS.  They should all be converted to use the routines below.  */
-
-/* At fsave_offset[REGNUM] you'll find the offset to the location in
-   the data structure used by the "fsave" instruction where GDB
-   register REGNUM is stored.  */
-
-static int fsave_offset[] =
-{
-  28 + 0 * FPU_REG_RAW_SIZE,	/* FP0_REGNUM through ...  */
-  28 + 1 * FPU_REG_RAW_SIZE,  
-  28 + 2 * FPU_REG_RAW_SIZE,  
-  28 + 3 * FPU_REG_RAW_SIZE,  
-  28 + 4 * FPU_REG_RAW_SIZE,  
-  28 + 5 * FPU_REG_RAW_SIZE,  
-  28 + 6 * FPU_REG_RAW_SIZE,  
-  28 + 7 * FPU_REG_RAW_SIZE,	/* ... FP7_REGNUM.  */
-  0,				/* FCTRL_REGNUM (16 bits).  */
-  4,				/* FSTAT_REGNUM (16 bits).  */
-  8,				/* FTAG_REGNUM (16 bits).  */
-  16,				/* FISEG_REGNUM (16 bits).  */
-  12,				/* FIOFF_REGNUM.  */
-  24,				/* FOSEG_REGNUM.  */
-  20,				/* FOOFF_REGNUM.  */
-  18				/* FOP_REGNUM (bottom 11 bits).  */
-};
-
-#define FSAVE_ADDR(fsave, regnum) (fsave + fsave_offset[regnum - FP0_REGNUM])
-\f
-
-/* Fill register REGNUM in GDB's register array with the appropriate
-   value from *FSAVE.  This function masks off any of the reserved
-   bits in *FSAVE.  */
-
-void
-i387_supply_register (int regnum, char *fsave)
-{
-  /* Most of the FPU control registers occupy only 16 bits in
-     the fsave area.  Give those a special treatment.  */
-  if (regnum >= FPC_REGNUM
-      && regnum != FIOFF_REGNUM && regnum != FOOFF_REGNUM)
-    {
-      unsigned int val = *(unsigned short *) (FSAVE_ADDR (fsave, regnum));
-
-      if (regnum == FOP_REGNUM)
-	{
-	  val &= ((1 << 11) - 1);
-	  supply_register (regnum, (char *) &val);
-	}
-      else
-	supply_register (regnum, (char *) &val);
-    }
-  else
-    supply_register (regnum, FSAVE_ADDR (fsave, regnum));
-}
-
-/* Fill GDB's register array with the floating-point register values
-   in *FSAVE.  This function masks off any of the reserved
-   bits in *FSAVE.  */
-
-void
-i387_supply_fsave (char *fsave)
-{
-  int i;
-
-  for (i = FP0_REGNUM; i < XMM0_REGNUM; i++)
-    i387_supply_register (i, fsave);
-}
-
-/* Fill register REGNUM (if it is a floating-point register) in *FSAVE
-   with the value in GDB's register array.  If REGNUM is -1, do this
-   for all registers.  This function doesn't touch any of the reserved
-   bits in *FSAVE.  */
-
-void
-i387_fill_fsave (char *fsave, int regnum)
-{
-  int i;
-
-  for (i = FP0_REGNUM; i < XMM0_REGNUM; i++)
-    if (regnum == -1 || regnum == i)
-      {
-	/* Most of the FPU control registers occupy only 16 bits in
-           the fsave area.  Give those a special treatment.  */
-	if (i >= FPC_REGNUM
-	    && i != FIOFF_REGNUM && i != FOOFF_REGNUM)
-	  {
-	    char buf[4];
-
-	    regcache_collect (i, buf);
-
-	    if (i == FOP_REGNUM)
-	      {
-		unsigned short oldval, newval;
-
-		/* The opcode occupies only 11 bits.  */
-		oldval = (*(unsigned short *) (FSAVE_ADDR (fsave, i)));
-		newval = *(unsigned short *) buf;
-		newval &= ((1 << 11) - 1);
-		newval |= oldval & ~((1 << 11) - 1);
-		memcpy (FSAVE_ADDR (fsave, i), &newval, 2);
-	      }
-	    else
-	      memcpy (FSAVE_ADDR (fsave, i), buf, 2);
-	  }
-	else
-	  regcache_collect (i, FSAVE_ADDR (fsave, i));
-      }
-}
-\f
-
-/* At fxsave_offset[REGNUM] you'll find the offset to the location in
-   the data structure used by the "fxsave" instruction where GDB
-   register REGNUM is stored.  */
-
-static int fxsave_offset[] =
-{
-  32,				/* FP0_REGNUM through ...  */
-  48,
-  64,
-  80,
-  96,
-  112,
-  128,
-  144,				/* ... FP7_REGNUM (80 bits each).  */
-  0,				/* FCTRL_REGNUM (16 bits).  */
-  2,				/* FSTAT_REGNUM (16 bits).  */
-  4,				/* FTAG_REGNUM (16 bits).  */
-  12,				/* FISEG_REGNUM (16 bits).  */
-  8,				/* FIOFF_REGNUM.  */
-  20,				/* FOSEG_REGNUM (16 bits).  */
-  16,				/* FOOFF_REGNUM.  */
-  6,				/* FOP_REGNUM (bottom 11 bits).  */
-  160,				/* XMM0_REGNUM through ...  */
-  176,
-  192,
-  208,
-  224,
-  240,
-  256,
-  272,				/* ... XMM7_REGNUM (128 bits each).  */
-  24,				/* MXCSR_REGNUM.  */
-};
-
-#define FXSAVE_ADDR(fxsave, regnum) \
-  (fxsave + fxsave_offset[regnum - FP0_REGNUM])
-
-static int i387_tag (unsigned char *raw);
-\f
-
-/* Fill GDB's register array with the floating-point and SSE register
-   values in *FXSAVE.  This function masks off any of the reserved
-   bits in *FXSAVE.  */
-
-void
-i387_supply_fxsave (char *fxsave)
-{
-  int i;
-
-  for (i = FP0_REGNUM; i <= MXCSR_REGNUM; i++)
-    {
-      /* Most of the FPU control registers occupy only 16 bits in
-	 the fxsave area.  Give those a special treatment.  */
-      if (i >= FPC_REGNUM && i < XMM0_REGNUM
-	  && i != FIOFF_REGNUM && i != FOOFF_REGNUM)
-	{
-	  unsigned long val = *(unsigned short *) (FXSAVE_ADDR (fxsave, i));
-
-	  if (i == FOP_REGNUM)
-	    {
-	      val &= ((1 << 11) - 1);
-	      supply_register (i, (char *) &val);
-	    }
-	  else if (i== FTAG_REGNUM)
-	    {
-	      /* The fxsave area contains a simplified version of the
-                 tag word.  We have to look at the actual 80-bit FP
-                 data to recreate the traditional i387 tag word.  */
-
-	      unsigned long ftag = 0;
-	      unsigned long fstat;
-	      int fpreg;
-	      int top;
-
-	      fstat = *(unsigned short *) (FXSAVE_ADDR (fxsave, FSTAT_REGNUM));
-	      top = ((fstat >> 11) & 0x7);
-
-	      for (fpreg = 7; fpreg >= 0; fpreg--)
-		{
-		  int tag;
-
-		  if (val & (1 << fpreg))
-		    {
-		      int regnum = (fpreg + 8 - top) % 8 + FP0_REGNUM;
-		      tag = i387_tag (FXSAVE_ADDR (fxsave, regnum));
-		    }
-		  else
-		    tag = 3;		/* Empty */
-
-		  ftag |= tag << (2 * fpreg);
-		}
-	      supply_register (i, (char *) &ftag);
-	    }
-	  else
-	    supply_register (i, (char *) &val);
-	}
-      else
-	supply_register (i, FXSAVE_ADDR (fxsave, i));
-    }
-}
-
-/* Fill register REGNUM (if it is a floating-point or SSE register) in
-   *FXSAVE with the value in GDB's register array.  If REGNUM is -1, do
-   this for all registers.  This function doesn't touch any of the
-   reserved bits in *FXSAVE.  */
-
-void
-i387_fill_fxsave (char *fxsave, int regnum)
-{
-  int i;
-
-  for (i = FP0_REGNUM; i <= MXCSR_REGNUM; i++)
-    if (regnum == -1 || regnum == i)
-      {
-	/* Most of the FPU control registers occupy only 16 bits in
-           the fxsave area.  Give those a special treatment.  */
-	if (i >= FPC_REGNUM && i < XMM0_REGNUM
-	    && i != FIOFF_REGNUM && i != FDOFF_REGNUM)
-	  {
-	    char buf[4];
-
-	    regcache_collect (i, buf);
-
-	    if (i == FOP_REGNUM)
-	      {
-		unsigned short oldval, newval;
-
-		/* The opcode occupies only 11 bits.  */
-		oldval = (*(unsigned short *) (FXSAVE_ADDR (fxsave, i)));
-		newval = *(unsigned short *) buf;
-		newval &= ((1 << 11) - 1);
-		newval |= oldval & ~((1 << 11) - 1);
-		memcpy (FXSAVE_ADDR (fxsave, i), &newval, 2);
-	      }
-	    else if (i == FTAG_REGNUM)
-	      {
-		/* Converting back is much easier.  */
-
-		unsigned short val = 0;
-		unsigned short ftag;
-		int fpreg;
-
-		ftag = *(unsigned short *) buf;
-
-		for (fpreg = 7; fpreg >= 0; fpreg--)
-		  {
-		    int tag = (ftag >> (fpreg * 2)) & 3;
-
-		    if (tag != 3)
-		      val |= (1 << fpreg);
-		  }
-
-		memcpy (FXSAVE_ADDR (fxsave, i), &val, 2);
-	      }
-	    else
-	      memcpy (FXSAVE_ADDR (fxsave, i), buf, 2);
-	  }
-	else
-	  regcache_collect (i, FXSAVE_ADDR (fxsave, i));
-      }
-}
-
-/* Recreate the FTW (tag word) valid bits from the 80-bit FP data in
-   *RAW.  */
-
-static int
-i387_tag (unsigned char *raw)
-{
-  int integer;
-  unsigned int exponent;
-  unsigned long fraction[2];
-
-  integer = raw[7] & 0x80;
-  exponent = (((raw[9] & 0x7f) << 8) | raw[8]);
-  fraction[0] = ((raw[3] << 24) | (raw[2] << 16) | (raw[1] << 8) | raw[0]);
-  fraction[1] = (((raw[7] & 0x7f) << 24) | (raw[6] << 16)
-		 | (raw[5] << 8) | raw[4]);
-
-  if (exponent == 0x7fff)
-    {
-      /* Special.  */
-      return (2);
-    }
-  else if (exponent == 0x0000)
-    {
-      if (fraction[0] == 0x0000 && fraction[1] == 0x0000 && !integer)
-	{
-	  /* Zero.  */
-	  return (1);
-	}
-      else
-	{
-	  /* Special.  */
-	  return (2);
-	}
-    }
-  else
-    {
-      if (integer)
-	{
-	  /* Valid.  */
-	  return (0);
-	}
-      else
-	{
-	  /* Special.  */
-	  return (2);
-	}
-    }
-}
Index: i387-nat.h
===================================================================
RCS file: i387-nat.h
diff -N i387-nat.h
--- i387-nat.h	17 Mar 2001 23:02:10 -0000	1.4
+++ /dev/null	1 Jan 1970 00:00:00 -0000
@@ -1,56 +0,0 @@
-/* Native-dependent code for the i387.
-   Copyright 2000, 2001 Free Software Foundation, Inc.
-
-   This file is part of GDB.
-
-   This program is free software; you can redistribute it and/or modify
-   it under the terms of the GNU General Public License as published by
-   the Free Software Foundation; either version 2 of the License, or
-   (at your option) any later version.
-
-   This program is distributed in the hope that it will be useful,
-   but WITHOUT ANY WARRANTY; without even the implied warranty of
-   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
-   GNU General Public License for more details.
-
-   You should have received a copy of the GNU General Public License
-   along with this program; if not, write to the Free Software
-   Foundation, Inc., 59 Temple Place - Suite 330,
-   Boston, MA 02111-1307, USA.  */
-
-#ifndef I387_NAT_H
-#define I387_NAT_H
-
-/* Fill register REGNUM in GDB's register array with the appropriate
-   value from *FSAVE.  This function masks off any of the reserved
-   bits in *FSAVE.  */
-
-extern void i387_supply_register (int regnum, char *fsave);
-
-/* Fill GDB's register array with the floating-point register values
-   in *FSAVE.  This function masks off any of the reserved
-   bits in *FSAVE.  */
-
-extern void i387_supply_fsave (char *fsave);
-
-/* Fill register REGNUM (if it is a floating-point register) in *FSAVE
-   with the value in GDB's register array.  If REGNUM is -1, do this
-   for all registers.  This function doesn't touch any of the reserved
-   bits in *FSAVE.  */
-
-extern void i387_fill_fsave (char *fsave, int regnum);
-
-/* Fill GDB's register array with the floating-point and SSE register
-   values in *FXSAVE.  This function masks off any of the reserved
-   bits in *FXSAVE.  */
-
-extern void i387_supply_fxsave (char *fxsave);
-
-/* Fill register REGNUM (if it is a floating-point or SSE register) in
-   *FXSAVE with the value in GDB's register array.  If REGNUM is -1, do
-   this for all registers.  This function doesn't touch any of the
-   reserved bits in *FXSAVE.  */
-
-extern void i387_fill_fxsave (char *fxsave, int regnum);
-
-#endif /* i387-nat.h */
Index: i387-tdep.c
===================================================================
RCS file: /cvs/src/src/gdb/i387-tdep.c,v
retrieving revision 1.16
diff -u -r1.16 i387-tdep.c
--- i387-tdep.c	12 Nov 2001 22:27:35 -0000	1.16
+++ i387-tdep.c	9 May 2002 23:33:59 -0000
@@ -388,3 +388,327 @@
   printf_filtered ("Opcode:              %s\n",
 		   local_hex_string_custom (fop ? (fop | 0xd800) : 0, "04"));
 }
+
+/* FIXME: kettenis/2000-05-21: Right now more than a few i386 targets
+   define their own routines to manage the floating-point registers in
+   GDB's register array.  Most (if not all) of these targets use the
+   format used by the "fsave" instruction in their communication with
+   the OS.  They should all be converted to use the routines below.  */
+
+/* At fsave_offset[REGNUM] you'll find the offset to the location in
+   the data structure used by the "fsave" instruction where GDB
+   register REGNUM is stored.  */
+
+static int fsave_offset[] =
+{
+  28 + 0 * FPU_REG_RAW_SIZE,	/* FP0_REGNUM through ...  */
+  28 + 1 * FPU_REG_RAW_SIZE,  
+  28 + 2 * FPU_REG_RAW_SIZE,  
+  28 + 3 * FPU_REG_RAW_SIZE,  
+  28 + 4 * FPU_REG_RAW_SIZE,  
+  28 + 5 * FPU_REG_RAW_SIZE,  
+  28 + 6 * FPU_REG_RAW_SIZE,  
+  28 + 7 * FPU_REG_RAW_SIZE,	/* ... FP7_REGNUM.  */
+  0,				/* FCTRL_REGNUM (16 bits).  */
+  4,				/* FSTAT_REGNUM (16 bits).  */
+  8,				/* FTAG_REGNUM (16 bits).  */
+  16,				/* FISEG_REGNUM (16 bits).  */
+  12,				/* FIOFF_REGNUM.  */
+  24,				/* FOSEG_REGNUM.  */
+  20,				/* FOOFF_REGNUM.  */
+  18				/* FOP_REGNUM (bottom 11 bits).  */
+};
+
+#define FSAVE_ADDR(fsave, regnum) (fsave + fsave_offset[regnum - FP0_REGNUM])
+\f
+
+/* Fill register REGNUM in GDB's register array with the appropriate
+   value from *FSAVE.  This function masks off any of the reserved
+   bits in *FSAVE.  */
+
+void
+i387_supply_register (int regnum, char *fsave)
+{
+  /* Most of the FPU control registers occupy only 16 bits in
+     the fsave area.  Give those a special treatment.  */
+  if (regnum >= FPC_REGNUM
+      && regnum != FIOFF_REGNUM && regnum != FOOFF_REGNUM)
+    {
+      unsigned int val = *(unsigned short *) (FSAVE_ADDR (fsave, regnum));
+
+      if (regnum == FOP_REGNUM)
+	{
+	  val &= ((1 << 11) - 1);
+	  supply_register (regnum, (char *) &val);
+	}
+      else
+	supply_register (regnum, (char *) &val);
+    }
+  else
+    supply_register (regnum, FSAVE_ADDR (fsave, regnum));
+}
+
+/* Fill GDB's register array with the floating-point register values
+   in *FSAVE.  This function masks off any of the reserved
+   bits in *FSAVE.  */
+
+void
+i387_supply_fsave (char *fsave)
+{
+  int i;
+
+  for (i = FP0_REGNUM; i < XMM0_REGNUM; i++)
+    i387_supply_register (i, fsave);
+}
+
+/* Fill register REGNUM (if it is a floating-point register) in *FSAVE
+   with the value in GDB's register array.  If REGNUM is -1, do this
+   for all registers.  This function doesn't touch any of the reserved
+   bits in *FSAVE.  */
+
+void
+i387_fill_fsave (char *fsave, int regnum)
+{
+  int i;
+
+  for (i = FP0_REGNUM; i < XMM0_REGNUM; i++)
+    if (regnum == -1 || regnum == i)
+      {
+	/* Most of the FPU control registers occupy only 16 bits in
+           the fsave area.  Give those a special treatment.  */
+	if (i >= FPC_REGNUM
+	    && i != FIOFF_REGNUM && i != FOOFF_REGNUM)
+	  {
+	    char buf[4];
+
+	    regcache_collect (i, buf);
+
+	    if (i == FOP_REGNUM)
+	      {
+		unsigned short oldval, newval;
+
+		/* The opcode occupies only 11 bits.  */
+		oldval = (*(unsigned short *) (FSAVE_ADDR (fsave, i)));
+		newval = *(unsigned short *) buf;
+		newval &= ((1 << 11) - 1);
+		newval |= oldval & ~((1 << 11) - 1);
+		memcpy (FSAVE_ADDR (fsave, i), &newval, 2);
+	      }
+	    else
+	      memcpy (FSAVE_ADDR (fsave, i), buf, 2);
+	  }
+	else
+	  regcache_collect (i, FSAVE_ADDR (fsave, i));
+      }
+}
+\f
+
+/* At fxsave_offset[REGNUM] you'll find the offset to the location in
+   the data structure used by the "fxsave" instruction where GDB
+   register REGNUM is stored.  */
+
+static int fxsave_offset[] =
+{
+  32,				/* FP0_REGNUM through ...  */
+  48,
+  64,
+  80,
+  96,
+  112,
+  128,
+  144,				/* ... FP7_REGNUM (80 bits each).  */
+  0,				/* FCTRL_REGNUM (16 bits).  */
+  2,				/* FSTAT_REGNUM (16 bits).  */
+  4,				/* FTAG_REGNUM (16 bits).  */
+  12,				/* FISEG_REGNUM (16 bits).  */
+  8,				/* FIOFF_REGNUM.  */
+  20,				/* FOSEG_REGNUM (16 bits).  */
+  16,				/* FOOFF_REGNUM.  */
+  6,				/* FOP_REGNUM (bottom 11 bits).  */
+  160,				/* XMM0_REGNUM through ...  */
+  176,
+  192,
+  208,
+  224,
+  240,
+  256,
+  272,				/* ... XMM7_REGNUM (128 bits each).  */
+  24,				/* MXCSR_REGNUM.  */
+};
+
+#define FXSAVE_ADDR(fxsave, regnum) \
+  (fxsave + fxsave_offset[regnum - FP0_REGNUM])
+
+static int i387_tag (unsigned char *raw);
+\f
+
+/* Fill GDB's register array with the floating-point and SSE register
+   values in *FXSAVE.  This function masks off any of the reserved
+   bits in *FXSAVE.  */
+
+void
+i387_supply_fxsave (char *fxsave)
+{
+  int i;
+
+  for (i = FP0_REGNUM; i <= MXCSR_REGNUM; i++)
+    {
+      /* Most of the FPU control registers occupy only 16 bits in
+	 the fxsave area.  Give those a special treatment.  */
+      if (i >= FPC_REGNUM && i < XMM0_REGNUM
+	  && i != FIOFF_REGNUM && i != FOOFF_REGNUM)
+	{
+	  unsigned long val = *(unsigned short *) (FXSAVE_ADDR (fxsave, i));
+
+	  if (i == FOP_REGNUM)
+	    {
+	      val &= ((1 << 11) - 1);
+	      supply_register (i, (char *) &val);
+	    }
+	  else if (i== FTAG_REGNUM)
+	    {
+	      /* The fxsave area contains a simplified version of the
+                 tag word.  We have to look at the actual 80-bit FP
+                 data to recreate the traditional i387 tag word.  */
+
+	      unsigned long ftag = 0;
+	      unsigned long fstat;
+	      int fpreg;
+	      int top;
+
+	      fstat = *(unsigned short *) (FXSAVE_ADDR (fxsave, FSTAT_REGNUM));
+	      top = ((fstat >> 11) & 0x7);
+
+	      for (fpreg = 7; fpreg >= 0; fpreg--)
+		{
+		  int tag;
+
+		  if (val & (1 << fpreg))
+		    {
+		      int regnum = (fpreg + 8 - top) % 8 + FP0_REGNUM;
+		      tag = i387_tag (FXSAVE_ADDR (fxsave, regnum));
+		    }
+		  else
+		    tag = 3;		/* Empty */
+
+		  ftag |= tag << (2 * fpreg);
+		}
+	      supply_register (i, (char *) &ftag);
+	    }
+	  else
+	    supply_register (i, (char *) &val);
+	}
+      else
+	supply_register (i, FXSAVE_ADDR (fxsave, i));
+    }
+}
+
+/* Fill register REGNUM (if it is a floating-point or SSE register) in
+   *FXSAVE with the value in GDB's register array.  If REGNUM is -1, do
+   this for all registers.  This function doesn't touch any of the
+   reserved bits in *FXSAVE.  */
+
+void
+i387_fill_fxsave (char *fxsave, int regnum)
+{
+  int i;
+
+  for (i = FP0_REGNUM; i <= MXCSR_REGNUM; i++)
+    if (regnum == -1 || regnum == i)
+      {
+	/* Most of the FPU control registers occupy only 16 bits in
+           the fxsave area.  Give those a special treatment.  */
+	if (i >= FPC_REGNUM && i < XMM0_REGNUM
+	    && i != FIOFF_REGNUM && i != FDOFF_REGNUM)
+	  {
+	    char buf[4];
+
+	    regcache_collect (i, buf);
+
+	    if (i == FOP_REGNUM)
+	      {
+		unsigned short oldval, newval;
+
+		/* The opcode occupies only 11 bits.  */
+		oldval = (*(unsigned short *) (FXSAVE_ADDR (fxsave, i)));
+		newval = *(unsigned short *) buf;
+		newval &= ((1 << 11) - 1);
+		newval |= oldval & ~((1 << 11) - 1);
+		memcpy (FXSAVE_ADDR (fxsave, i), &newval, 2);
+	      }
+	    else if (i == FTAG_REGNUM)
+	      {
+		/* Converting back is much easier.  */
+
+		unsigned short val = 0;
+		unsigned short ftag;
+		int fpreg;
+
+		ftag = *(unsigned short *) buf;
+
+		for (fpreg = 7; fpreg >= 0; fpreg--)
+		  {
+		    int tag = (ftag >> (fpreg * 2)) & 3;
+
+		    if (tag != 3)
+		      val |= (1 << fpreg);
+		  }
+
+		memcpy (FXSAVE_ADDR (fxsave, i), &val, 2);
+	      }
+	    else
+	      memcpy (FXSAVE_ADDR (fxsave, i), buf, 2);
+	  }
+	else
+	  regcache_collect (i, FXSAVE_ADDR (fxsave, i));
+      }
+}
+
+/* Recreate the FTW (tag word) valid bits from the 80-bit FP data in
+   *RAW.  */
+
+static int
+i387_tag (unsigned char *raw)
+{
+  int integer;
+  unsigned int exponent;
+  unsigned long fraction[2];
+
+  integer = raw[7] & 0x80;
+  exponent = (((raw[9] & 0x7f) << 8) | raw[8]);
+  fraction[0] = ((raw[3] << 24) | (raw[2] << 16) | (raw[1] << 8) | raw[0]);
+  fraction[1] = (((raw[7] & 0x7f) << 24) | (raw[6] << 16)
+		 | (raw[5] << 8) | raw[4]);
+
+  if (exponent == 0x7fff)
+    {
+      /* Special.  */
+      return (2);
+    }
+  else if (exponent == 0x0000)
+    {
+      if (fraction[0] == 0x0000 && fraction[1] == 0x0000 && !integer)
+	{
+	  /* Zero.  */
+	  return (1);
+	}
+      else
+	{
+	  /* Special.  */
+	  return (2);
+	}
+    }
+  else
+    {
+      if (integer)
+	{
+	  /* Valid.  */
+	  return (0);
+	}
+      else
+	{
+	  /* Special.  */
+	  return (2);
+	}
+    }
+}
Index: i387-tdep.h
===================================================================
RCS file: i387-tdep.h
diff -N i387-tdep.h
--- /dev/null	1 Jan 1970 00:00:00 -0000
+++ i387-tdep.h	9 May 2002 23:33:59 -0000
@@ -0,0 +1,56 @@
+/* Target-dependent code for the i387.
+   Copyright 2000, 2001 Free Software Foundation, Inc.
+
+   This file is part of GDB.
+
+   This program is free software; you can redistribute it and/or modify
+   it under the terms of the GNU General Public License as published by
+   the Free Software Foundation; either version 2 of the License, or
+   (at your option) any later version.
+
+   This program is distributed in the hope that it will be useful,
+   but WITHOUT ANY WARRANTY; without even the implied warranty of
+   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+   GNU General Public License for more details.
+
+   You should have received a copy of the GNU General Public License
+   along with this program; if not, write to the Free Software
+   Foundation, Inc., 59 Temple Place - Suite 330,
+   Boston, MA 02111-1307, USA.  */
+
+#ifndef I387_TDEP_H
+#define I387_TDEP_H
+
+/* Fill register REGNUM in GDB's register array with the appropriate
+   value from *FSAVE.  This function masks off any of the reserved
+   bits in *FSAVE.  */
+
+extern void i387_supply_register (int regnum, char *fsave);
+
+/* Fill GDB's register array with the floating-point register values
+   in *FSAVE.  This function masks off any of the reserved
+   bits in *FSAVE.  */
+
+extern void i387_supply_fsave (char *fsave);
+
+/* Fill register REGNUM (if it is a floating-point register) in *FSAVE
+   with the value in GDB's register array.  If REGNUM is -1, do this
+   for all registers.  This function doesn't touch any of the reserved
+   bits in *FSAVE.  */
+
+extern void i387_fill_fsave (char *fsave, int regnum);
+
+/* Fill GDB's register array with the floating-point and SSE register
+   values in *FXSAVE.  This function masks off any of the reserved
+   bits in *FXSAVE.  */
+
+extern void i387_supply_fxsave (char *fxsave);
+
+/* Fill register REGNUM (if it is a floating-point or SSE register) in
+   *FXSAVE with the value in GDB's register array.  If REGNUM is -1, do
+   this for all registers.  This function doesn't touch any of the
+   reserved bits in *FXSAVE.  */
+
+extern void i387_fill_fxsave (char *fxsave, int regnum);
+
+#endif /* i387-tdep.h */
Index: x86-64-linux-nat.c
===================================================================
RCS file: /cvs/src/src/gdb/x86-64-linux-nat.c,v
retrieving revision 1.11
diff -u -r1.11 x86-64-linux-nat.c
--- x86-64-linux-nat.c	7 May 2002 11:22:54 -0000	1.11
+++ x86-64-linux-nat.c	9 May 2002 23:34:00 -0000
@@ -25,7 +25,7 @@
 #include "inferior.h"
 #include "gdbcore.h"
 #include "regcache.h"
-#include "i387-nat.h"
+#include "i387-tdep.h"
 #include "gdb_assert.h"
 #include "x86-64-tdep.h"
 
Index: config/i386/fbsd.mh
===================================================================
RCS file: /cvs/src/src/gdb/config/i386/fbsd.mh,v
retrieving revision 1.11
diff -u -r1.11 fbsd.mh
--- config/i386/fbsd.mh	31 Mar 2002 17:47:17 -0000	1.11
+++ config/i386/fbsd.mh	9 May 2002 23:34:00 -0000
@@ -4,4 +4,4 @@
 
 NAT_FILE= nm-fbsd.h
 # NOTE: Do not spread NATDEPFILES over several lines - it hurts BSD make.
-NATDEPFILES= fork-child.o infptrace.o inftarg.o solib.o solib-svr4.o solib-legacy.o corelow.o core-aout.o core-regset.o i386-nat.o i387-nat.o i386bsd-nat.o i386fbsd-nat.o gcore.o fbsd-proc.o
+NATDEPFILES= fork-child.o infptrace.o inftarg.o solib.o solib-svr4.o solib-legacy.o corelow.o core-aout.o core-regset.o i386-nat.o i386bsd-nat.o i386fbsd-nat.o gcore.o fbsd-proc.o
Index: config/i386/go32.mh
===================================================================
RCS file: /cvs/src/src/gdb/config/i386/go32.mh,v
retrieving revision 1.5
diff -u -r1.5 go32.mh
--- config/i386/go32.mh	18 Jan 2002 04:50:59 -0000	1.5
+++ config/i386/go32.mh	9 May 2002 23:34:01 -0000
@@ -7,7 +7,7 @@
 XM_FILE= xm-go32.h
 
 NAT_FILE= nm-go32.h
-NATDEPFILES= go32-nat.o i386-nat.o i387-nat.o
+NATDEPFILES= go32-nat.o i386-nat.o
 
 TERMCAP=
 HOST_IPC=
Index: config/i386/i386gnu.mh
===================================================================
RCS file: /cvs/src/src/gdb/config/i386/i386gnu.mh,v
retrieving revision 1.7
diff -u -r1.7 i386gnu.mh
--- config/i386/i386gnu.mh	24 Apr 2002 21:44:47 -0000	1.7
+++ config/i386/i386gnu.mh	9 May 2002 23:34:01 -0000
@@ -1,5 +1,5 @@
 # Host: Intel 386 running the GNU Hurd
-NATDEPFILES= i386gnu-nat.o i387-nat.o gnu-nat.o corelow.o core-regset.o \
+NATDEPFILES= i386gnu-nat.o gnu-nat.o corelow.o core-regset.o \
 	     fork-child.o solib.o solib-svr4.o solib-legacy.o \
 	     notify_S.o process_reply_S.o msg_reply_S.o \
 	     msg_U.o exc_request_U.o exc_request_S.o
Index: config/i386/i386sol2.mh
===================================================================
RCS file: /cvs/src/src/gdb/config/i386/i386sol2.mh,v
retrieving revision 1.4
diff -u -r1.4 i386sol2.mh
--- config/i386/i386sol2.mh	14 Feb 2002 08:13:33 -0000	1.4
+++ config/i386/i386sol2.mh	9 May 2002 23:34:01 -0000
@@ -4,5 +4,5 @@
 XM_CLIBS= -lsocket -lnsl
 
 NAT_FILE= nm-i386sol2.h
-NATDEPFILES= core-regset.o fork-child.o i386v4-nat.o i387-nat.o corelow.o \
+NATDEPFILES= core-regset.o fork-child.o i386v4-nat.o corelow.o \
 	procfs.o proc-api.o proc-events.o proc-flags.o proc-why.o gcore.o
Index: config/i386/i386v42mp.mh
===================================================================
RCS file: /cvs/src/src/gdb/config/i386/i386v42mp.mh,v
retrieving revision 1.7
diff -u -r1.7 i386v42mp.mh
--- config/i386/i386v42mp.mh	11 Apr 2002 18:13:21 -0000	1.7
+++ config/i386/i386v42mp.mh	9 May 2002 23:34:01 -0000
@@ -14,7 +14,7 @@
 # continuation character (backslash) to extend a commented line.  As a
 # consequence, make considers subsequent tab-indented lines to be
 # some sort of error.
-NATDEPFILES= corelow.o core-regset.o fork-child.o i386v4-nat.o i387-nat.o \
+NATDEPFILES= corelow.o core-regset.o fork-child.o i386v4-nat.o \
 	gcore.o solib.o solib-svr4.o solib-legacy.o procfs.o proc-api.o \
 	proc-events.o proc-flags.o proc-why.o uw-thread.o
 
Index: config/i386/linux.mh
===================================================================
RCS file: /cvs/src/src/gdb/config/i386/linux.mh,v
retrieving revision 1.11
diff -u -r1.11 linux.mh
--- config/i386/linux.mh	18 Jan 2002 04:50:59 -0000	1.11
+++ config/i386/linux.mh	9 May 2002 23:34:01 -0000
@@ -4,7 +4,7 @@
 
 NAT_FILE= nm-linux.h
 NATDEPFILES= infptrace.o inftarg.o fork-child.o corelow.o linux-proc.o \
-	core-aout.o i386-nat.o i386-linux-nat.o i387-nat.o \
+	core-aout.o i386-nat.o i386-linux-nat.o \
 	proc-service.o thread-db.o lin-lwp.o linux-proc.o gcore.o
 
 # The dynamically loaded libthread_db needs access to symbols in the
Index: config/i386/nbsd.mh
===================================================================
RCS file: /cvs/src/src/gdb/config/i386/nbsd.mh,v
retrieving revision 1.12
diff -u -r1.12 nbsd.mh
--- config/i386/nbsd.mh	18 Jan 2002 04:50:59 -0000	1.12
+++ config/i386/nbsd.mh	9 May 2002 23:34:01 -0000
@@ -6,7 +6,7 @@
 # continuation character (backslash) to extend a commented line.  As a
 # consequence, BSD make considers subsequent tab-indented lines to be
 # "unassociated shell commands".
-NATDEPFILES= fork-child.o infptrace.o inftarg.o corelow.o i387-nat.o i386bsd-nat.o i386nbsd-nat.o solib.o solib-sunos.o
+NATDEPFILES= fork-child.o infptrace.o inftarg.o corelow.o i386bsd-nat.o i386nbsd-nat.o solib.o solib-sunos.o
 
 XM_FILE= xm-nbsd.h
 NAT_FILE= nm-nbsd.h
Index: config/i386/nbsdelf.mh
===================================================================
RCS file: /cvs/src/src/gdb/config/i386/nbsdelf.mh,v
retrieving revision 1.8
diff -u -r1.8 nbsdelf.mh
--- config/i386/nbsdelf.mh	18 Jan 2002 04:50:59 -0000	1.8
+++ config/i386/nbsdelf.mh	9 May 2002 23:34:01 -0000
@@ -1,4 +1,4 @@
 # Host: Intel 386 running NetBSD
-NATDEPFILES= fork-child.o infptrace.o inftarg.o corelow.o i387-nat.o i386bsd-nat.o i386nbsd-nat.o solib.o solib-svr4.o solib-legacy.o
+NATDEPFILES= fork-child.o infptrace.o inftarg.o corelow.o i386bsd-nat.o i386nbsd-nat.o solib.o solib-svr4.o solib-legacy.o
 XM_FILE= xm-nbsd.h
 NAT_FILE= nm-nbsdelf.h
Index: config/i386/obsd.mh
===================================================================
RCS file: /cvs/src/src/gdb/config/i386/obsd.mh,v
retrieving revision 1.2
diff -u -r1.2 obsd.mh
--- config/i386/obsd.mh	18 Jan 2002 04:50:59 -0000	1.2
+++ config/i386/obsd.mh	9 May 2002 23:34:01 -0000
@@ -4,4 +4,4 @@
 
 NAT_FILE= nm-obsd.h
 # NOTE: Do not spread NATDEPFILES over several lines - it hurts BSD make.
-NATDEPFILES= fork-child.o infptrace.o inftarg.o solib.o solib-sunos.o corelow.o core-aout.o i386-nat.o i387-nat.o i386bsd-nat.o
+NATDEPFILES= fork-child.o infptrace.o inftarg.o solib.o solib-sunos.o corelow.o core-aout.o i386-nat.o i386bsd-nat.o
Index: config/i386/x86-64linux.mh
===================================================================
RCS file: /cvs/src/src/gdb/config/i386/x86-64linux.mh,v
retrieving revision 1.5
diff -u -r1.5 x86-64linux.mh
--- config/i386/x86-64linux.mh	21 Feb 2002 12:48:36 -0000	1.5
+++ config/i386/x86-64linux.mh	9 May 2002 23:34:01 -0000
@@ -5,7 +5,7 @@
 NAT_FILE= nm-x86-64.h
 NATDEPFILES= infptrace.o inftarg.o fork-child.o corelow.o \
 	core-aout.o i386-nat.o x86-64-linux-nat.o \
-	i387-nat.o proc-service.o thread-db.o lin-lwp.o \
+	proc-service.o thread-db.o lin-lwp.o \
 	linux-proc.o gcore.o 
 
 LOADLIBES = -ldl -rdynamic

^ permalink raw reply	[flat|nested] 8+ messages in thread

end of thread, other threads:[~2002-05-10 15:10 UTC | newest]

Thread overview: 8+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2002-04-11 16:05 [PATCH/RFA] Move i387 register supply/fill to i387-tdep.c Jason R Thorpe
2002-05-09 16:45 Jason R Thorpe
2002-05-09 17:12 ` Daniel Jacobowitz
2002-05-09 17:23   ` Jason R Thorpe
2002-05-09 18:28     ` Daniel Jacobowitz
2002-05-10  3:54     ` Mark Kettenis
2002-05-10  8:10       ` Jason R Thorpe
2002-05-10  0:26 ` Eli Zaretskii

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox