Re: [RFA] The MIPS/Linux port

Re: [RFA] The MIPS/Linux port

[H . J . Lu]

On Wed, Jul 04, 2001 at 12:19:52PM -0700, Daniel Jacobowitz wrote:
>      addressed in either binutils or gcc, and HJ posted a potential patch
>      for it two days ago.

I have verified that this patch works for me. Since many gcc mips 
patches haven't been reviewed, I didn't send it to gcc.


H.J.
----
2001-07-02  H.J. Lu <hjl@gnu.org>

	* gcc/config/mips/linux.h (ASM_OUTPUT_SOURCE_LINE): Defined.

--- gcc/config/mips/linux.h.stabs	Mon Jul  2 21:53:09 2001
+++ gcc/config/mips/linux.h	Mon Jul  2 21:58:00 2001
@@ -182,6 +182,22 @@ Boston, MA 02111-1307, USA.  */
 #undef SET_ASM_OP
 #define SET_ASM_OP "\t.dummy\t"
 
+#undef  ASM_OUTPUT_SOURCE_LINE
+#define ASM_OUTPUT_SOURCE_LINE(FILE, LINE)				\
+do									\
+  {									\
+    static int sym_lineno = 1;						\
+    fprintf (FILE, "%sLM%d:\n\t%s 68,0,%d,%sLM%d",			\
+	     LOCAL_LABEL_PREFIX, sym_lineno, ASM_STABN_OP,		\
+	     LINE, LOCAL_LABEL_PREFIX, sym_lineno);			\
+    putc ('-', FILE);							\
+    assemble_name (FILE,						\
+		   XSTR (XEXP (DECL_RTL (current_function_decl), 0), 0));\
+    putc ('\n', FILE);							\
+    sym_lineno++;							\
+  }									\
+while (0)
+
 /* This is how we tell the assembler that two symbols have the
    same value.  */
 #undef ASM_OUTPUT_DEF

Re: [RFA] The MIPS/Linux port

[Mark Kettenis]

Looks like the MIPS port was started before a bit of reformatting of
the sources and conversion to ISO C took place.  Please

 * Format the copyright notice at the beginning of the file as all
   (most of) the other files, for example i386-linux-tdep.c.

 * Don't put a start on every line of multi-line comments.

 * Put two spaces after a full stop (.) in comments, even at the end
   of the comment.  Make the comments full senteces, including a full
   stop at the end of them.

 * Wrap long comments.

 * Use ISO C prototypes and function definitions, e.g.
   
      int
      get_longjmp_target (CORE_ADDR *pc)
      {

   and

      static void fetch_core_registers (char *, unsigned, int, CORE_ADDR);

   (get rid of PARAMS).

Good luck!

Mark

Re: [RFA] The MIPS/Linux port

[Daniel Jacobowitz]

On Wed, Jul 04, 2001 at 11:51:27PM +0200, Mark Kettenis wrote:
> Looks like the MIPS port was started before a bit of reformatting of
> the sources and conversion to ISO C took place.  Please

<snip>

Fixed.  Do you see any other formatting problems?

-- 
Daniel Jacobowitz                           Carnegie Mellon University
MontaVista Software                         Debian GNU/Linux Developer

2001-07-04  Daniel Jacobowitz  <drow@mvista.com>
	* mips-linux-tdep.c: New file.
	* mips-linux-nat.c: New file.
	* config/mips/linux.mh: New file.
	* config/mips/linux.mt: New file.
	* config/mips/xm-linux.h: New file.
	* config/mips/nm-linux.h: New file.
	* config/mips/tm-linux.h: New file.
	* configure.host: Recognize mips*-*-linux.
	* configure.tgt: Likewise.

Index: gdb/mips-linux-tdep.c
===================================================================
RCS file: N/A
diff -u /dev/null gdb/mips-linux-tdep.c
--- /dev/null	Wed Dec 31 16:00:00 1969
+++ gdb/mips-linux-tdep.c	Wed Jul  4 16:02:13 2001
@@ -0,0 +1,313 @@
+/* Target-dependent code for Linux/MIPS.
+   Copyright 1996, 2001 Free Software Foundation, Inc.
+
+   Originally contributed by David S. Miller (davem@caip.rutgers.edu) at
+   Rutgers University CAIP Research Center.
+
+   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 "gdbcore.h"
+#include "target.h"
+#include "solib-svr4.h"
+
+/* Copied from <asm/elf.h>.  */
+#define ELF_NGREG       45
+#define ELF_NFPREG      33
+
+typedef unsigned char elf_greg_t[4];
+typedef elf_greg_t elf_gregset_t[ELF_NGREG];
+
+typedef unsigned char elf_fpreg_t[8];
+typedef elf_fpreg_t elf_fpregset_t[ELF_NFPREG];
+
+/* 0 - 31 are integer registers, 32 - 63 are fp registers.  */
+#define FPR_BASE        32
+#define PC              64
+#define CAUSE           65
+#define BADVADDR        66
+#define MMHI            67
+#define MMLO            68
+#define FPC_CSR         69
+#define FPC_EIR         70
+
+#define EF_REG0			6
+#define EF_REG31		37
+#define EF_LO			38
+#define EF_HI			39
+#define EF_CP0_EPC		40
+#define EF_CP0_BADVADDR		41
+#define EF_CP0_STATUS		42
+#define EF_CP0_CAUSE		43
+
+#define EF_SIZE			180
+
+/* Figure out where the longjmp will land.
+   We expect the first arg to be a pointer to the jmp_buf structure from
+   which we extract the pc (JB_PC) that we will land at.  The pc is copied
+   into PC.  This routine returns 1 on success.  */
+
+int
+get_longjmp_target (CORE_ADDR *pc)
+{
+  CORE_ADDR jb_addr;
+  char buf[TARGET_PTR_BIT / TARGET_CHAR_BIT];
+
+  jb_addr = read_register (A0_REGNUM);
+
+  if (target_read_memory (jb_addr + JB_PC * JB_ELEMENT_SIZE, buf,
+			  TARGET_PTR_BIT / TARGET_CHAR_BIT))
+    return 0;
+
+  *pc = extract_address (buf, TARGET_PTR_BIT / TARGET_CHAR_BIT);
+
+  return 1;
+}
+
+/* Unpack an elf_gregset_t into GDB's register cache.  */
+
+void 
+supply_gregset (elf_gregset_t *gregsetp)
+{
+  int regi;
+  elf_greg_t *regp = *gregsetp;
+  static char zerobuf[MAX_REGISTER_RAW_SIZE] = {0};
+
+  for (regi = EF_REG0; regi <= EF_REG31; regi++)
+    supply_register ((regi - EF_REG0), (char *)(regp + regi));
+
+  supply_register (LO_REGNUM, (char *)(regp + EF_LO));
+  supply_register (HI_REGNUM, (char *)(regp + EF_HI));
+
+  supply_register (PC_REGNUM, (char *)(regp + EF_CP0_EPC));
+  supply_register (BADVADDR_REGNUM, (char *)(regp + EF_CP0_BADVADDR));
+  supply_register (PS_REGNUM, (char *)(regp + EF_CP0_STATUS));
+  supply_register (CAUSE_REGNUM, (char *)(regp + EF_CP0_CAUSE));
+
+  /* Fill inaccessible registers with zero.  */
+  supply_register (FP_REGNUM, zerobuf);
+  supply_register (UNUSED_REGNUM, zerobuf);
+  for (regi = FIRST_EMBED_REGNUM; regi < LAST_EMBED_REGNUM; regi++)
+    supply_register (regi, zerobuf);
+}
+
+/* Pack our registers (or one register) into an elf_gregset_t.  */
+
+void
+fill_gregset (elf_gregset_t *gregsetp, int regno)
+{
+  int regi;
+  elf_greg_t *regp = *gregsetp;
+
+  memset ((char *) regp, 0, sizeof (elf_gregset_t));
+
+#define fill(regp, regi, offset)			\
+	  memcpy((char *)((regp) + (offset)),		\
+		 &registers[REGISTER_BYTE (regi)],	\
+		 sizeof (elf_greg_t))
+
+  for (regi = EF_REG0; regi <= EF_REG31; regi++)
+    fill (regp, regi - EF_REG0, regi);
+
+  fill (regp, LO_REGNUM, EF_LO);
+  fill (regp, HI_REGNUM, EF_HI);
+
+  fill (regp, PC_REGNUM, EF_CP0_EPC);
+  fill (regp, BADVADDR_REGNUM, EF_CP0_BADVADDR);
+  fill (regp, PS_REGNUM, EF_CP0_STATUS);
+  fill (regp, CAUSE_REGNUM, EF_CP0_CAUSE);
+#undef fill
+}
+
+/* Likewise, unpack an elf_fpregset_t.  */
+
+void
+supply_fpregset (elf_fpregset_t *fpregsetp)
+{
+  register int regi;
+  static char zerobuf[MAX_REGISTER_RAW_SIZE] = {0};
+
+  for (regi = 0; regi < 32; regi++)
+    supply_register (FP0_REGNUM + regi,
+		     (char *)(*fpregsetp + regi));
+
+  supply_register (FCRCS_REGNUM, (char *)(*fpregsetp + 32));
+
+  /* FIXME: how can we supply FCRIR_REGNUM?  The ABI doesn't tell us. */
+  supply_register (FCRIR_REGNUM, zerobuf);
+}
+
+/* Likewise, pack one or all floating point registers into an
+   elf_fpregset_t.  */
+
+void
+fill_fpregset (elf_fpregset_t *fpregsetp, int regno)
+{
+  int regi;
+  char *from, *to;
+
+  for (regi = 0; regi < 32; regi++)
+    {
+      if ((regno == -1) || (regno == FP0_REGNUM + regi))
+	{
+	  from = (char *) &registers[REGISTER_BYTE (regi + FP0_REGNUM)];
+	  to = (char *) (*fpregsetp + regi);
+	  memcpy (to, from, REGISTER_RAW_SIZE (regi));
+	}
+    }
+
+  if ((regno == -1) || (regno == FCRCS_REGNUM))
+    {
+      from = (char *) &registers[REGISTER_BYTE (FCRCS_REGNUM)];
+      to = (char *) (*fpregsetp + 32);
+      memcpy (to, from, REGISTER_RAW_SIZE (FCRCS_REGNUM));
+    }
+}
+
+/* Map gdb internal register number to ptrace ``address''.
+   These ``addresses'' are defined in <asm/ptrace.h>.  */
+
+#define REGISTER_PTRACE_ADDR(regno)		\
+   (regno < 32 ? 		regno  		\
+  : (regno >= FP0_REGNUM && regno < FP0_REGNUM + 32) ?		\
+				FPR_BASE + (regno - FP0_REGNUM)	\
+  : regno == PC_REGNUM ?	PC		\
+  : regno == CAUSE_REGNUM ?	CAUSE		\
+  : regno == BADVADDR_REGNUM ?	BADVADDR	\
+  : regno == LO_REGNUM ?	MMLO		\
+  : regno == HI_REGNUM ?	MMHI		\
+  : regno == FCRCS_REGNUM ?	FPC_CSR		\
+  : regno == FCRIR_REGNUM ?	FPC_EIR		\
+  : 0)
+
+/* Return the ptrace ``address'' of register REGNO.  */
+
+CORE_ADDR
+register_addr (int regno, CORE_ADDR blockend)
+{
+  int regaddr;
+
+  if (regno < 0 || regno >= NUM_REGS)
+    error ("Bogon register number %d.", regno);
+
+  regaddr = REGISTER_PTRACE_ADDR (regno);
+
+  if (regno != 0 && regaddr == 0)
+    error ("Unknowable register number %d.", regno);
+
+  return regaddr;
+}
+
+static void fetch_core_registers (char *, unsigned, int, CORE_ADDR);
+
+void _initialize_core_regset (void);
+
+/*  Use a local version of this function to get the correct types for
+    regsets, until multi-arch core support is ready.  */
+
+static void
+fetch_core_registers (char *core_reg_sect, unsigned core_reg_size,
+		      int which, CORE_ADDR reg_addr)
+{
+  elf_gregset_t gregset;
+  elf_fpregset_t fpregset;
+
+  if (which == 0)
+    {
+      if (core_reg_size != sizeof (gregset))
+	{
+	  warning ("wrong size gregset struct in core file");
+	}
+      else
+	{
+	  memcpy ((char *) &gregset, core_reg_sect, sizeof (gregset));
+	  supply_gregset (&gregset);
+	}
+    }
+  else if (which == 2)
+    {
+      if (core_reg_size != sizeof (fpregset))
+	{
+	  warning ("wrong size fpregset struct in core file");
+	}
+      else
+	{
+	  memcpy ((char *) &fpregset, core_reg_sect, sizeof (fpregset));
+	  supply_fpregset (&fpregset);
+	}
+    }
+}
+
+/* Register that we are able to handle ELF file formats using standard
+   procfs "regset" structures.  */
+
+static struct core_fns regset_core_fns =
+{
+  bfd_target_elf_flavour,		/* core_flavour */
+  default_check_format,			/* check_format */
+  default_core_sniffer,			/* core_sniffer */
+  fetch_core_registers,			/* core_read_registers */
+  NULL					/* next */
+};
+
+void
+_initialize_core_regset ()
+{
+  add_core_fns (&regset_core_fns);
+}
+
+
+/* Fetch (and possibly build) an appropriate link_map_offsets
+   structure for native Linux/MIPS targets using the struct offsets
+   defined in link.h (but without actual reference to that file).
+
+   This makes it possible to access Linux/MIPS shared libraries from a
+   GDB that was not built on an Linux/MIPS host (for cross debugging).  */
+
+struct link_map_offsets *
+mips_linux_svr4_fetch_link_map_offsets (void)
+{ 
+  static struct link_map_offsets lmo;
+  static struct link_map_offsets *lmp = NULL;
+
+  if (lmp == NULL)
+    { 
+      lmp = &lmo;
+
+      lmo.r_debug_size = 8;	/* The actual size is 20 bytes, but
+				   this is all we need.  */
+      lmo.r_map_offset = 4;
+      lmo.r_map_size   = 4;
+
+      lmo.link_map_size = 20;
+
+      lmo.l_addr_offset = 0;
+      lmo.l_addr_size   = 4;
+
+      lmo.l_name_offset = 4;
+      lmo.l_name_size   = 4;
+
+      lmo.l_next_offset = 12;
+      lmo.l_next_size   = 4;
+
+      lmo.l_prev_offset = 16;
+      lmo.l_prev_size   = 4;
+    }
+
+  return lmp;
+}
Index: gdb/mips-linux-nat.c
===================================================================
RCS file: N/A
diff -u /dev/null gdb/mips-linux-nat.c
--- /dev/null	Wed Dec 31 16:00:00 1969
+++ gdb/mips-linux-nat.c	Wed Jul  4 15:46:03 2001
@@ -0,0 +1,33 @@
+/* Native-dependent code for Linux/MIPS.
+   Copyright 1996, 2001 Free Software Foundation, Inc.
+
+   Originally contributed by David S. Miller (davem@caip.rutgers.edu) at
+   Rutgers University CAIP Research Center.
+
+   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 <sys/user.h>
+
+/* Return the size of the user struct.  */
+
+int
+kernel_u_size (void)
+{
+  return (sizeof (struct user));
+}
+  
Index: gdb/config/mips/xm-linux.h
===================================================================
RCS file: N/A
diff -u /dev/null gdb/config/mips/xm-linux.h
--- /dev/null	Wed Dec 31 16:00:00 1969
+++ gdb/config/mips/xm-linux.h	Wed Jul  4 16:27:53 2001
@@ -0,0 +1,43 @@
+/* Host definitions for Linux/MIPS.
+   Copyright 1996, 2001 Free Software Foundation, Inc.
+
+   Originally contributed by David S. Miller (davem@caip.rutgers.edu) at
+   Rutgers University CAIP Research Center.
+
+   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 XM_MIPSLINUX_H
+#define XM_MIPSLINUX_H
+
+#ifndef HOST_BYTE_ORDER
+#include <endian.h>
+#if __BYTE_ORDER == __BIG_ENDIAN
+#define HOST_BYTE_ORDER BIG_ENDIAN
+#else
+#define HOST_BYTE_ORDER LITTLE_ENDIAN
+#endif
+#endif
+
+#define HAVE_TERMIOS
+
+#define NEED_POSIX_SETPGID
+
+/* Need R_OK etc, but USG isn't defined.  */
+#include <unistd.h>
+
+#endif /* XM_MIPSLINUX_H */
Index: gdb/config/mips/tm-linux.h
===================================================================
RCS file: N/A
diff -u /dev/null gdb/config/mips/tm-linux.h
--- /dev/null	Wed Dec 31 16:00:00 1969
+++ gdb/config/mips/tm-linux.h	Wed Jul  4 16:26:23 2001
@@ -0,0 +1,94 @@
+/* Target-dependent definitions for Linux/MIPS.
+   Copyright 1996, 2001 Free Software Foundation, Inc.
+
+   Originally contributed by David S. Miller (davem@caip.rutgers.edu) at
+   Rutgers University CAIP Research Center.
+
+   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 TM_MIPSLINUX_H
+#define TM_MIPSLINUX_H
+
+#include "mips/tm-mips.h"
+
+/* We don't want to inherit tm-mips.h's shared library trampoline code.  */
+
+#undef IN_SOLIB_CALL_TRAMPOLINE
+#undef IN_SOLIB_RETURN_TRAMPOLINE
+#undef SKIP_TRAMPOLINE_CODE
+#undef IGNORE_HELPER_CALL
+
+/* FIXME: Do not include "tm-linux.h", because we do not want the host's
+   <signal.h>.  Instead, copy the non-signal bits for now.  */
+
+/* We need this file for the SOLIB_TRAMPOLINE stuff.  */
+
+#include "tm-sysv4.h"
+
+/* We define SVR4_SHARED_LIBS unconditionally, on the assumption that
+   link.h is available on all linux platforms.  For I386 and SH3/4,
+   we hard-code the information rather than use link.h anyway (for
+   the benefit of cross-debugging).  We may move to doing that for
+   other architectures as well.  */
+
+#define SVR4_SHARED_LIBS
+#include "solib.h"              /* Support for shared libraries.  */
+
+/* End from "tm-linux.h".  */
+
+/* There's an E_MIPS_ABI_O32 flag in e_flags, but we don't use it - in
+   fact, using it may violate the o32 ABI.  */
+
+#define MIPS_DEFAULT_ABI MIPS_ABI_O32
+
+/* Linux/MIPS has __SIGRTMAX == 127.  */
+
+#define REALTIME_LO 32
+#define REALTIME_HI 128
+
+/* Use target_specific function to define link map offsets.  */
+
+extern struct link_map_offsets *mips_linux_svr4_fetch_link_map_offsets (void);
+#define SVR4_FETCH_LINK_MAP_OFFSETS() \
+  mips_linux_svr4_fetch_link_map_offsets ()
+
+/* Details about jmp_buf.  */
+
+#define JB_ELEMENT_SIZE 4
+#define JB_PC 0
+
+/* Figure out where the longjmp will land.  Slurp the arguments out of the
+   stack.  We expect the first arg to be a pointer to the jmp_buf structure
+   from which we extract the pc (JB_PC) that we will land at.  The pc is
+   copied into ADDR.  This routine returns 1 on success.  */
+
+#define GET_LONGJMP_TARGET(ADDR) get_longjmp_target(ADDR)
+extern int get_longjmp_target (CORE_ADDR *);
+
+/* We do single stepping in software.  */
+
+#define SOFTWARE_SINGLE_STEP_P() 1
+extern void mips_software_single_step (unsigned int, int);
+#define SOFTWARE_SINGLE_STEP(sig,bp_p) mips_software_single_step (sig, bp_p)
+
+/* FIXME: This still needs to be implemented.  */
+
+#undef  IN_SIGTRAMP
+#define IN_SIGTRAMP(pc, name)	(0)
+
+#endif /* TM_MIPSLINUX_H */
Index: gdb/config/mips/nm-linux.h
===================================================================
RCS file: N/A
diff -u /dev/null gdb/config/mips/nm-linux.h
--- /dev/null	Wed Dec 31 16:00:00 1969
+++ gdb/config/mips/nm-linux.h	Wed Jul  4 16:20:05 2001
@@ -0,0 +1,76 @@
+/* Native-dependent definitions for Linux/MIPS.
+   Copyright 1996, 2001 Free Software Foundation, Inc.
+
+   Originally contributed by David S. Miller (davem@caip.rutgers.edu) at
+   Rutgers University CAIP Research Center.
+
+   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 NM_MIPSLINUX_H
+#define NM_MIPSLINUX_H
+
+#define MIPS_GNULINUX_TARGET
+
+#include "nm-linux.h"
+
+/* Return sizeof user struct to callers in less machine dependent
+   routines.  */
+
+extern int kernel_u_size (void);
+#define KERNEL_U_SIZE kernel_u_size()
+
+/* ptrace register ``addresses'' are absolute.  */
+
+#define U_REGS_OFFSET 0
+
+/* ptrace transfers longs, and expects addresses as longs.  */
+
+#define PTRACE_ARG3_TYPE long
+#define PTRACE_XFER_TYPE long
+
+#define REGISTER_U_ADDR(addr, blockend, regno) (addr) = mips_register_addr ((regno),(blockend))
+
+/* Pseudo registers can not be read.  ptrace does not provide a way to
+   read (or set) PS_REGNUM, and there's no point in reading or setting
+   ZERO_REGNUM.  We also can not set BADVADDR, CAUSE, or FCRIR via
+   ptrace().  These macros only apply to the ptrace() code, as kgdb
+   has different restrictions.  */
+
+#define CANNOT_FETCH_REGISTER(regno) \
+  ((regno) == PS_REGNUM || (regno) == ZERO_REGNUM || (regno) >= FP_REGNUM)
+#define CANNOT_STORE_REGISTER(regno) \
+  ((regno) == PS_REGNUM || (regno) == ZERO_REGNUM || (regno) >= FP_REGNUM \
+   || (regno) == BADVADDR_REGNUM || (regno) == CAUSE_REGNUM || (regno) == FCRIR_REGNUM)
+
+/* FIXME: This should be moved to ../nm-linux.h once we have converted all
+   Linux targets to use the new threads stuff (without the #undef of
+   course).  */
+
+extern int lin_lwp_prepare_to_proceed (void);
+#undef PREPARE_TO_PROCEED
+#define PREPARE_TO_PROCEED(select_it) lin_lwp_prepare_to_proceed ()
+
+extern void lin_lwp_attach_lwp (ptid_t ptid, int verbose);
+#define ATTACH_LWP(ptid, verbose) lin_lwp_attach_lwp ((ptid), (verbose))
+
+#include <signal.h>
+
+extern void lin_thread_get_thread_signals (sigset_t *mask);
+#define GET_THREAD_SIGNALS(mask) lin_thread_get_thread_signals (mask)
+
+#endif /* NM_MIPSLINUX_H */
Index: gdb/config/mips/linux.mh
===================================================================
RCS file: N/A
diff -u /dev/null gdb/config/mips/linux.mh
--- /dev/null	Wed Dec 31 16:00:00 1969
+++ gdb/config/mips/linux.mh	Wed Jul  4 16:05:51 2001
@@ -0,0 +1,8 @@
+# Host: Linux/MIPS
+XDEPFILES=
+XM_FILE= xm-linux.h
+NAT_FILE= nm-linux.h
+NATDEPFILES= infptrace.o inftarg.o fork-child.o mips-linux-nat.o \
+	thread-db.o lin-lwp.o proc-service.o
+
+LOADLIBES = -ldl -rdynamic
Index: gdb/config/mips/linux.mt
===================================================================
RCS file: N/A
diff -u /dev/null gdb/config/mips/linux.mt
--- /dev/null	Wed Dec 31 16:00:00 1969
+++ gdb/config/mips/linux.mt	Wed Jul  4 16:06:00 2001
@@ -0,0 +1,9 @@
+# Target: Linux/MIPS
+TDEPFILES= mips-tdep.o mips-linux-tdep.o corelow.o \
+	solib.o solib-svr4.o
+TM_FILE= tm-linux.h
+
+GDBSERVER_DEPFILES= low-linux.o
+
+SIM_OBS = remote-sim.o
+SIM = ../sim/mips/libsim.a
Index: gdb/configure.host
===================================================================
RCS file: /cvs/src/src/gdb/configure.host,v
retrieving revision 1.22
diff -u -r1.22 configure.host
--- gdb/configure.host	2001/06/15 19:23:45	1.22
+++ gdb/configure.host	2001/07/04 23:28:18
@@ -16,6 +16,7 @@
 i[3456]86*)		gdb_host_cpu=i386 ;;
 m68*)			gdb_host_cpu=m68k ;;
 m88*)			gdb_host_cpu=m88k ;;
+mips*)			gdb_host_cpu=mips ;;
 powerpc*)		gdb_host_cpu=powerpc ;;
 sparc64)		gdb_host_cpu=sparc ;;
 s390*)			gdb_host_cpu=s390 ;;
@@ -118,6 +119,7 @@
 mips-sgi-irix5*)	gdb_host=irix5 ;;
 mips-sgi-irix6*)	gdb_host=irix6 ;;
 mips-sony-*)		gdb_host=news-mips ;;
+mips*-*-linux*)		gdb_host=linux ;;
 mips-*-mach3*)		gdb_host=mipsm3 ;;
 mips-*-sysv4*)		gdb_host=mipsv4 ;;
 mips-*-sysv*)		gdb_host=riscos ;;
Index: gdb/configure.tgt
===================================================================
RCS file: /cvs/src/src/gdb/configure.tgt,v
retrieving revision 1.30
diff -u -r1.30 configure.tgt
--- gdb/configure.tgt	2001/06/15 19:23:45	1.30
+++ gdb/configure.tgt	2001/07/04 23:28:19
@@ -211,6 +211,7 @@
 mips*-sgi-irix6*)	gdb_target=irix6 ;;
 mips*-sgi-*)		gdb_target=irix3 ;;
 mips*-sony-*)		gdb_target=bigmips ;;
+mips*-*-linux*)		gdb_target=linux ;;
 mips*-*-mach3*)		gdb_target=mipsm3 ;;
 mips*-*-sysv4*)		gdb_target=mipsv4 ;;
 mips*-*-sysv*)		gdb_target=bigmips ;;

Re: [RFA] The MIPS/Linux port

[Mark Kettenis]

   Date: Wed, 4 Jul 2001 16:34:23 -0700
   From: Daniel Jacobowitz <dmj+@andrew.cmu.edu>

   Fixed.  Do you see any other formatting problems?

Yep :-).  It would be great if you could wrap the definitions of
REGISTER_U_ADDR and CANNOT_STORE_REGISTER such that the lines are not
too long for 80-character wide terminal.

Otherwise, things look really fine.

Mark

Re: [RFA] The MIPS/Linux port

[John Court]

Daniel Jacobowitz wrote:

>On Wed, Jul 04, 2001 at 11:51:27PM +0200, Mark Kettenis wrote:
>
>>Looks like the MIPS port was started before a bit of reformatting of
>>the sources and conversion to ISO C took place.  Please
>>
>
><snip>
>
>Fixed.  Do you see any other formatting problems?
>
Hi Daniel,

You have a bug in your  supply_fpregset() and fill_fpregset()
routines.  You are being passed an elf_fpregset_t * and incrementing
the counter by 264 od bytes each time through the loops.  You
need to offload the arguement something like the following before
using it with regi.

  elf_fpreg_t  *fpregp = &(*fpregsetp)[0];


This wont show up very well until you try and load a core dump
files which will actually cause gdb to die as it spins off into invalid
memory.    I would be surprised if your floating point regs in
gdb look any good the way the code is at the moment.
I had to look at the assembler to work this out a few days ago.

Cheers

John Court

Re: [RFA] The MIPS/Linux port

[Daniel Jacobowitz]

On Fri, Jul 06, 2001 at 09:57:26AM +1000, John Court wrote:
> Daniel Jacobowitz wrote:
> 
> >On Wed, Jul 04, 2001 at 11:51:27PM +0200, Mark Kettenis wrote:
> >
> >>Looks like the MIPS port was started before a bit of reformatting of
> >>the sources and conversion to ISO C took place.  Please
> >>
> >
> ><snip>
> >
> >Fixed.  Do you see any other formatting problems?
> >
> Hi Daniel,
> 
> You have a bug in your  supply_fpregset() and fill_fpregset()
> routines.  You are being passed an elf_fpregset_t * and incrementing
> the counter by 264 od bytes each time through the loops.  You
> need to offload the arguement something like the following before
> using it with regi.
> 
>   elf_fpreg_t  *fpregp = &(*fpregsetp)[0];
> 
> 
> This wont show up very well until you try and load a core dump
> files which will actually cause gdb to die as it spins off into invalid
> memory.    I would be surprised if your floating point regs in
> gdb look any good the way the code is at the moment.
> I had to look at the assembler to work this out a few days ago.

No, you're wrong.  Note that I say:
  supply_register (FCRCS_REGNUM, (char *)(*fpregsetp + 32));

&(*fpregsetp)[0] is just fancy notation for *fpregsetp, in this case at
least.

I specifcally fixed that bug at least three times :)


-- 
Daniel Jacobowitz                           Carnegie Mellon University
MontaVista Software                         Debian GNU/Linux Developer

Re: [RFA] The MIPS/Linux port

[John Court]

Daniel Jacobowitz wrote:

>On Fri, Jul 06, 2001 at 09:57:26AM +1000, John Court wrote:
>
>>Daniel Jacobowitz wrote:
>>
>>>On Wed, Jul 04, 2001 at 11:51:27PM +0200, Mark Kettenis wrote:
>>>
>>>>Looks like the MIPS port was started before a bit of reformatting of
>>>>the sources and conversion to ISO C took place.  Please
>>>>
>>><snip>
>>>
>>>Fixed.  Do you see any other formatting problems?
>>>
>>Hi Daniel,
>>
>>You have a bug in your  supply_fpregset() and fill_fpregset()
>>routines.  You are being passed an elf_fpregset_t * and incrementing
>>the counter by 264 od bytes each time through the loops.  You
>>need to offload the arguement something like the following before
>>using it with regi.
>>
>>  elf_fpreg_t  *fpregp = &(*fpregsetp)[0];
>>
>>
>>This wont show up very well until you try and load a core dump
>>files which will actually cause gdb to die as it spins off into invalid
>>memory.    I would be surprised if your floating point regs in
>>gdb look any good the way the code is at the moment.
>>I had to look at the assembler to work this out a few days ago.
>>
>
>No, you're wrong.  Note that I say:
>  supply_register (FCRCS_REGNUM, (char *)(*fpregsetp + 32));
>
>&(*fpregsetp)[0] is just fancy notation for *fpregsetp, in this case at
>least.
>
>I specifcally fixed that bug at least three times :)
>
>
In this case I am happy to be wrong :-)    I was working with someone 
elses patches
on 5.0 and missed the extra "*" in your core.    I should have checked more
closely.....I have now and you are right  :-(.

Cheers

John

Re: [RFA] The MIPS/Linux port

[Daniel Jacobowitz]

On Fri, Jul 06, 2001 at 10:44:38AM +1000, John Court wrote:
> In this case I am happy to be wrong :-)    I was working with someone 
> elses patches
> on 5.0 and missed the extra "*" in your core.    I should have checked more
> closely.....I have now and you are right  :-(.

Indeed.  You've presumably got the patches that "everyone is using". 
That's the least of the things wrong with them, and one of the reasons
I'm merging at last :)

-- 
Daniel Jacobowitz                           Carnegie Mellon University
MontaVista Software                         Debian GNU/Linux Developer

Re: [RFA] The MIPS/Linux port

[Eli Zaretskii]

On Wed, 4 Jul 2001, Daniel Jacobowitz wrote:

> Comments welcome!  (approval even more so)

An entry for NEWS (where new configurations are listed) would be nice.