/* SPU target-dependent code for GDB, the GNU debugger.
   Copyright (C) 2006-2017 Free Software Foundation, Inc.

   Contributed by Ulrich Weigand <uweigand@de.ibm.com>.
   Based on a port by Sid Manning <sid@us.ibm.com>.

   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 3 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, see <http://www.gnu.org/licenses/>.  */

#include "defs.h"
#include "arch-utils.h"
#include "gdbtypes.h"
#include "gdbcmd.h"
#include "gdbcore.h"
#include "frame.h"
#include "frame-unwind.h"
#include "frame-base.h"
#include "trad-frame.h"
#include "symtab.h"
#include "symfile.h"
#include "value.h"
#include "inferior.h"
#include "dis-asm.h"
#include "disasm.h"
#include "objfiles.h"
#include "language.h"
#include "regcache.h"
#include "reggroups.h"
#include "floatformat.h"
#include "block.h"
#include "observer.h"
#include "infcall.h"
#include "dwarf2.h"
#include "dwarf2-frame.h"
#include "ax.h"
#include "spu-tdep.h"
#include "location.h"

/* The list of available "set spu " and "show spu " commands.  */
static struct cmd_list_element *setspucmdlist = NULL;
static struct cmd_list_element *showspucmdlist = NULL;

/* Whether to stop for new SPE contexts.  */
static int spu_stop_on_load_p = 0;
/* Whether to automatically flush the SW-managed cache.  */
static int spu_auto_flush_cache_p = 1;


/* The tdep structure.  */
struct gdbarch_tdep
{
  /* The spufs ID identifying our address space.  */
  int id;

  /* SPU-specific vector type.  */
  struct type *spu_builtin_type_vec128;
};


/* SPU-specific vector type.  */
static struct type *
spu_builtin_type_vec128 (struct gdbarch *gdbarch)
{
  struct gdbarch_tdep *tdep = gdbarch_tdep (gdbarch);

  if (!tdep->spu_builtin_type_vec128)
    {
      const struct builtin_type *bt = builtin_type (gdbarch);
      struct type *t;

      t = arch_composite_type (gdbarch,
			       "__spu_builtin_type_vec128", TYPE_CODE_UNION);
      append_composite_type_field (t, "uint128", bt->builtin_int128);
      append_composite_type_field (t, "v2_int64",
				   init_vector_type (bt->builtin_int64, 2));
      append_composite_type_field (t, "v4_int32",
				   init_vector_type (bt->builtin_int32, 4));
      append_composite_type_field (t, "v8_int16",
				   init_vector_type (bt->builtin_int16, 8));
      append_composite_type_field (t, "v16_int8",
				   init_vector_type (bt->builtin_int8, 16));
      append_composite_type_field (t, "v2_double",
				   init_vector_type (bt->builtin_double, 2));
      append_composite_type_field (t, "v4_float",
				   init_vector_type (bt->builtin_float, 4));

      TYPE_VECTOR (t) = 1;
      TYPE_NAME (t) = "spu_builtin_type_vec128";

      tdep->spu_builtin_type_vec128 = t;
    }

  return tdep->spu_builtin_type_vec128;
}


/* The list of available "info spu " commands.  */
static struct cmd_list_element *infospucmdlist = NULL;

/* Registers.  */

static const char *
spu_register_name (struct gdbarch *gdbarch, int reg_nr)
{
  static const char *register_names[] =
    {
      "r0", "r1", "r2", "r3", "r4", "r5", "r6", "r7",
      "r8", "r9", "r10", "r11", "r12", "r13", "r14", "r15",
      "r16", "r17", "r18", "r19", "r20", "r21", "r22", "r23",
      "r24", "r25", "r26", "r27", "r28", "r29", "r30", "r31",
      "r32", "r33", "r34", "r35", "r36", "r37", "r38", "r39",
      "r40", "r41", "r42", "r43", "r44", "r45", "r46", "r47",
      "r48", "r49", "r50", "r51", "r52", "r53", "r54", "r55",
      "r56", "r57", "r58", "r59", "r60", "r61", "r62", "r63",
      "r64", "r65", "r66", "r67", "r68", "r69", "r70", "r71",
      "r72", "r73", "r74", "r75", "r76", "r77", "r78", "r79",
      "r80", "r81", "r82", "r83", "r84", "r85", "r86", "r87",
      "r88", "r89", "r90", "r91", "r92", "r93", "r94", "r95",
      "r96", "r97", "r98", "r99", "r100", "r101", "r102", "r103",
      "r104", "r105", "r106", "r107", "r108", "r109", "r110", "r111",
      "r112", "r113", "r114", "r115", "r116", "r117", "r118", "r119",
      "r120", "r121", "r122", "r123", "r124", "r125", "r126", "r127",
      "id", "pc", "sp", "fpscr", "srr0", "lslr", "decr", "decr_status"
    };

  if (reg_nr < 0)
    return NULL;
  if (reg_nr >= sizeof register_names / sizeof *register_names)
    return NULL;

  return register_names[reg_nr];
}

static struct type *
spu_register_type (struct gdbarch *gdbarch, int reg_nr)
{
  if (reg_nr < SPU_NUM_GPRS)
    return spu_builtin_type_vec128 (gdbarch);

  switch (reg_nr)
    {
    case SPU_ID_REGNUM:
      return builtin_type (gdbarch)->builtin_uint32;

    case SPU_PC_REGNUM:
      return builtin_type (gdbarch)->builtin_func_ptr;

    case SPU_SP_REGNUM:
      return builtin_type (gdbarch)->builtin_data_ptr;

    case SPU_FPSCR_REGNUM:
      return builtin_type (gdbarch)->builtin_uint128;

    case SPU_SRR0_REGNUM:
      return builtin_type (gdbarch)->builtin_uint32;

    case SPU_LSLR_REGNUM:
      return builtin_type (gdbarch)->builtin_uint32;

    case SPU_DECR_REGNUM:
      return builtin_type (gdbarch)->builtin_uint32;

    case SPU_DECR_STATUS_REGNUM:
      return builtin_type (gdbarch)->builtin_uint32;

    default:
      internal_error (__FILE__, __LINE__, _("invalid regnum"));
    }
}

/* Pseudo registers for preferred slots - stack pointer.  */

static enum register_status
spu_pseudo_register_read_spu (struct regcache *regcache, const char *regname,
			      gdb_byte *buf)
{
  struct gdbarch *gdbarch = get_regcache_arch (regcache);
  enum bfd_endian byte_order = gdbarch_byte_order (gdbarch);
  enum register_status status;
  gdb_byte reg[32];
  char annex[32];
  ULONGEST id;
  ULONGEST ul;

  status = regcache_raw_read_unsigned (regcache, SPU_ID_REGNUM, &id);
  if (status != REG_VALID)
    return status;
  xsnprintf (annex, sizeof annex, "%d/%s", (int) id, regname);
  memset (reg, 0, sizeof reg);
  target_read (&current_target, TARGET_OBJECT_SPU, annex,
	       reg, 0, sizeof reg);

  ul = strtoulst ((char *) reg, NULL, 16);
  store_unsigned_integer (buf, 4, byte_order, ul);
  return REG_VALID;
}

static enum register_status
spu_pseudo_register_read (struct gdbarch *gdbarch, struct regcache *regcache,
                          int regnum, gdb_byte *buf)
{
  gdb_byte reg[16];
  char annex[32];
  ULONGEST id;
  enum register_status status;

  switch (regnum)
    {
    case SPU_SP_REGNUM:
      status = regcache_raw_read (regcache, SPU_RAW_SP_REGNUM, reg);
      if (status != REG_VALID)
	return status;
      memcpy (buf, reg, 4);
      return status;

    case SPU_FPSCR_REGNUM:
      status = regcache_raw_read_unsigned (regcache, SPU_ID_REGNUM, &id);
      if (status != REG_VALID)
	return status;
      xsnprintf (annex, sizeof annex, "%d/fpcr", (int) id);
      target_read (&current_target, TARGET_OBJECT_SPU, annex, buf, 0, 16);
      return status;

    case SPU_SRR0_REGNUM:
      return spu_pseudo_register_read_spu (regcache, "srr0", buf);

    case SPU_LSLR_REGNUM:
      return spu_pseudo_register_read_spu (regcache, "lslr", buf);

    case SPU_DECR_REGNUM:
      return spu_pseudo_register_read_spu (regcache, "decr", buf);

    case SPU_DECR_STATUS_REGNUM:
      return spu_pseudo_register_read_spu (regcache, "decr_status", buf);

    default:
      internal_error (__FILE__, __LINE__, _("invalid regnum"));
    }
}

static void
spu_pseudo_register_write_spu (struct regcache *regcache, const char *regname,
			       const gdb_byte *buf)
{
  struct gdbarch *gdbarch = get_regcache_arch (regcache);
  enum bfd_endian byte_order = gdbarch_byte_order (gdbarch);
  char reg[32];
  char annex[32];
  ULONGEST id;

  regcache_raw_read_unsigned (regcache, SPU_ID_REGNUM, &id);
  xsnprintf (annex, sizeof annex, "%d/%s", (int) id, regname);
  xsnprintf (reg, sizeof reg, "0x%s",
	     phex_nz (extract_unsigned_integer (buf, 4, byte_order), 4));
  target_write (&current_target, TARGET_OBJECT_SPU, annex,
		(gdb_byte *) reg, 0, strlen (reg));
}

static void
spu_pseudo_register_write (struct gdbarch *gdbarch, struct regcache *regcache,
                           int regnum, const gdb_byte *buf)
{
  gdb_byte reg[16];
  char annex[32];
  ULONGEST id;

  switch (regnum)
    {
    case SPU_SP_REGNUM:
      regcache_raw_read (regcache, SPU_RAW_SP_REGNUM, reg);
      memcpy (reg, buf, 4);
      regcache_raw_write (regcache, SPU_RAW_SP_REGNUM, reg);
      break;

    case SPU_FPSCR_REGNUM:
      regcache_raw_read_unsigned (regcache, SPU_ID_REGNUM, &id);
      xsnprintf (annex, sizeof annex, "%d/fpcr", (int) id);
      target_write (&current_target, TARGET_OBJECT_SPU, annex, buf, 0, 16);
      break;

    case SPU_SRR0_REGNUM:
      spu_pseudo_register_write_spu (regcache, "srr0", buf);
      break;

    case SPU_LSLR_REGNUM:
      spu_pseudo_register_write_spu (regcache, "lslr", buf);
      break;

    case SPU_DECR_REGNUM:
      spu_pseudo_register_write_spu (regcache, "decr", buf);
      break;

    case SPU_DECR_STATUS_REGNUM:
      spu_pseudo_register_write_spu (regcache, "decr_status", buf);
      break;

    default:
      internal_error (__FILE__, __LINE__, _("invalid regnum"));
    }
}

static int
spu_ax_pseudo_register_collect (struct gdbarch *gdbarch,
				struct agent_expr *ax, int regnum)
{
  switch (regnum)
    {
    case SPU_SP_REGNUM:
      ax_reg_mask (ax, SPU_RAW_SP_REGNUM);
      return 0;

    case SPU_FPSCR_REGNUM:
    case SPU_SRR0_REGNUM:
    case SPU_LSLR_REGNUM:
    case SPU_DECR_REGNUM:
    case SPU_DECR_STATUS_REGNUM:
      return -1;

    default:
      internal_error (__FILE__, __LINE__, _("invalid regnum"));
    }
}

static int
spu_ax_pseudo_register_push_stack (struct gdbarch *gdbarch,
				   struct agent_expr *ax, int regnum)
{
  switch (regnum)
    {
    case SPU_SP_REGNUM:
      ax_reg (ax, SPU_RAW_SP_REGNUM);
      return 0;

    case SPU_FPSCR_REGNUM:
    case SPU_SRR0_REGNUM:
    case SPU_LSLR_REGNUM:
    case SPU_DECR_REGNUM:
    case SPU_DECR_STATUS_REGNUM:
      return -1;

    default:
      internal_error (__FILE__, __LINE__, _("invalid regnum"));
    }
}


/* Value conversion -- access scalar values at the preferred slot.  */

static struct value *
spu_value_from_register (struct gdbarch *gdbarch, struct type *type,
			 int regnum, struct frame_id frame_id)
{
  struct value *value = default_value_from_register (gdbarch, type,
						     regnum, frame_id);
  LONGEST len = TYPE_LENGTH (type);

  if (regnum < SPU_NUM_GPRS && len < 16)
    {
      int preferred_slot = len < 4 ? 4 - len : 0;
      set_value_offset (value, preferred_slot);
    }

  return value;
}

/* Register groups.  */

static int
spu_register_reggroup_p (struct gdbarch *gdbarch, int regnum,
			 struct reggroup *group)
{
  /* Registers displayed via 'info regs'.  */
  if (group == general_reggroup)
    return 1;

  /* Registers displayed via 'info float'.  */
  if (group == float_reggroup)
    return 0;

  /* Registers that need to be saved/restored in order to
     push or pop frames.  */
  if (group == save_reggroup || group == restore_reggroup)
    return 1;

  return default_register_reggroup_p (gdbarch, regnum, group);
}

/* DWARF-2 register numbers.  */

static int
spu_dwarf_reg_to_regnum (struct gdbarch *gdbarch, int reg)
{
  /* Use cooked instead of raw SP.  */
  return (reg == SPU_RAW_SP_REGNUM)? SPU_SP_REGNUM : reg;
}


/* Address handling.  */

static int
spu_gdbarch_id (struct gdbarch *gdbarch)
{
  struct gdbarch_tdep *tdep = gdbarch_tdep (gdbarch);
  int id = tdep->id;

  /* The objfile architecture of a standalone SPU executable does not
     provide an SPU ID.  Retrieve it from the objfile's relocated
     address range in this special case.  */
  if (id == -1
      && symfile_objfile && symfile_objfile->obfd
      && bfd_get_arch (symfile_objfile->obfd) == bfd_arch_spu
      && symfile_objfile->sections != symfile_objfile->sections_end)
    id = SPUADDR_SPU (obj_section_addr (symfile_objfile->sections));

  return id;
}

static int
spu_address_class_type_flags (int byte_size, int dwarf2_addr_class)
{
  if (dwarf2_addr_class == 1)
    return TYPE_INSTANCE_FLAG_ADDRESS_CLASS_1;
  else
    return 0;
}

static const char *
spu_address_class_type_flags_to_name (struct gdbarch *gdbarch, int type_flags)
{
  if (type_flags & TYPE_INSTANCE_FLAG_ADDRESS_CLASS_1)
    return "__ea";
  else
    return NULL;
}

static int
spu_address_class_name_to_type_flags (struct gdbarch *gdbarch,
				      const char *name, int *type_flags_ptr)
{
  if (strcmp (name, "__ea") == 0)
    {
      *type_flags_ptr = TYPE_INSTANCE_FLAG_ADDRESS_CLASS_1;
      return 1;
    }
  else
   return 0;
}

static void
spu_address_to_pointer (struct gdbarch *gdbarch,
			struct type *type, gdb_byte *buf, CORE_ADDR addr)
{
  enum bfd_endian byte_order = gdbarch_byte_order (gdbarch);
  store_unsigned_integer (buf, TYPE_LENGTH (type), byte_order,
			  SPUADDR_ADDR (addr));
}

static CORE_ADDR
spu_pointer_to_address (struct gdbarch *gdbarch,
			struct type *type, const gdb_byte *buf)
{
  int id = spu_gdbarch_id (gdbarch);
  enum bfd_endian byte_order = gdbarch_byte_order (gdbarch);
  ULONGEST addr
    = extract_unsigned_integer (buf, TYPE_LENGTH (type), byte_order);

  /* Do not convert __ea pointers.  */
  if (TYPE_ADDRESS_CLASS_1 (type))
    return addr;

  return addr? SPUADDR (id, addr) : 0;
}

static CORE_ADDR
spu_integer_to_address (struct gdbarch *gdbarch,
			struct type *type, const gdb_byte *buf)
{
  int id = spu_gdbarch_id (gdbarch);
  ULONGEST addr = unpack_long (type, buf);

  return SPUADDR (id, addr);
}


/* Decoding SPU instructions.  */

enum
  {
    op_lqd   = 0x34,
    op_lqx   = 0x3c4,
    op_lqa   = 0x61,
    op_lqr   = 0x67,
    op_stqd  = 0x24,
    op_stqx  = 0x144,
    op_stqa  = 0x41,
    op_stqr  = 0x47,

    op_il    = 0x081,
    op_ila   = 0x21,
    op_a     = 0x0c0,
    op_ai    = 0x1c,

    op_selb  = 0x8,

    op_br    = 0x64,
    op_bra   = 0x60,
    op_brsl  = 0x66,
    op_brasl = 0x62,
    op_brnz  = 0x42,
    op_brz   = 0x40,
    op_brhnz = 0x46,
    op_brhz  = 0x44,
    op_bi    = 0x1a8,
    op_bisl  = 0x1a9,
    op_biz   = 0x128,
    op_binz  = 0x129,
    op_bihz  = 0x12a,
    op_bihnz = 0x12b,
  };

static int
is_rr (unsigned int insn, int op, int *rt, int *ra, int *rb)
{
  if ((insn >> 21) == op)
    {
      *rt = insn & 127;
      *ra = (insn >> 7) & 127;
      *rb = (insn >> 14) & 127;
      return 1;
    }

  return 0;
}

static int
is_rrr (unsigned int insn, int op, int *rt, int *ra, int *rb, int *rc)
{
  if ((insn >> 28) == op)
    {
      *rt = (insn >> 21) & 127;
      *ra = (insn >> 7) & 127;
      *rb = (insn >> 14) & 127;
      *rc = insn & 127;
      return 1;
    }

  return 0;
}

static int
is_ri7 (unsigned int insn, int op, int *rt, int *ra, int *i7)
{
  if ((insn >> 21) == op)
    {
      *rt = insn & 127;
      *ra = (insn >> 7) & 127;
      *i7 = (((insn >> 14) & 127) ^ 0x40) - 0x40;
      return 1;
    }

  return 0;
}

static int
is_ri10 (unsigned int insn, int op, int *rt, int *ra, int *i10)
{
  if ((insn >> 24) == op)
    {
      *rt = insn & 127;
      *ra = (insn >> 7) & 127;
      *i10 = (((insn >> 14) & 0x3ff) ^ 0x200) - 0x200;
      return 1;
    }

  return 0;
}

static int
is_ri16 (unsigned int insn, int op, int *rt, int *i16)
{
  if ((insn >> 23) == op)
    {
      *rt = insn & 127;
      *i16 = (((insn >> 7) & 0xffff) ^ 0x8000) - 0x8000;
      return 1;
    }

  return 0;
}

static int
is_ri18 (unsigned int insn, int op, int *rt, int *i18)
{
  if ((insn >> 25) == op)
    {
      *rt = insn & 127;
      *i18 = (((insn >> 7) & 0x3ffff) ^ 0x20000) - 0x20000;
      return 1;
    }

  return 0;
}

static int
is_branch (unsigned int insn, int *offset, int *reg)
{
  int rt, i7, i16;

  if (is_ri16 (insn, op_br, &rt, &i16)
      || is_ri16 (insn, op_brsl, &rt, &i16)
      || is_ri16 (insn, op_brnz, &rt, &i16)
      || is_ri16 (insn, op_brz, &rt, &i16)
      || is_ri16 (insn, op_brhnz, &rt, &i16)
      || is_ri16 (insn, op_brhz, &rt, &i16))
    {
      *reg = SPU_PC_REGNUM;
      *offset = i16 << 2;
      return 1;
    }

  if (is_ri16 (insn, op_bra, &rt, &i16)
      || is_ri16 (insn, op_brasl, &rt, &i16))
    {
      *reg = -1;
      *offset = i16 << 2;
      return 1;
    }

  if (is_ri7 (insn, op_bi, &rt, reg, &i7)
      || is_ri7 (insn, op_bisl, &rt, reg, &i7)
      || is_ri7 (insn, op_biz, &rt, reg, &i7)
      || is_ri7 (insn, op_binz, &rt, reg, &i7)
      || is_ri7 (insn, op_bihz, &rt, reg, &i7)
      || is_ri7 (insn, op_bihnz, &rt, reg, &i7))
    {
      *offset = 0;
      return 1;
    }

  return 0;
}


/* Prolog parsing.  */

struct spu_prologue_data
  {
    /* Stack frame size.  -1 if analysis was unsuccessful.  */
    int size;

    /* How to find the CFA.  The CFA is equal to SP at function entry.  */
    int cfa_reg;
    int cfa_offset;

    /* Offset relative to CFA where a register is saved.  -1 if invalid.  */
    int reg_offset[SPU_NUM_GPRS];
  };

static CORE_ADDR
spu_analyze_prologue (struct gdbarch *gdbarch,
		      CORE_ADDR start_pc, CORE_ADDR end_pc,
                      struct spu_prologue_data *data)
{
  enum bfd_endian byte_order = gdbarch_byte_order (gdbarch);
  int found_sp = 0;
  int found_fp = 0;
  int found_lr = 0;
  int found_bc = 0;
  int reg_immed[SPU_NUM_GPRS];
  gdb_byte buf[16];
  CORE_ADDR prolog_pc = start_pc;
  CORE_ADDR pc;
  int i;


  /* Initialize DATA to default values.  */
  data->size = -1;

  data->cfa_reg = SPU_RAW_SP_REGNUM;
  data->cfa_offset = 0;

  for (i = 0; i < SPU_NUM_GPRS; i++)
    data->reg_offset[i] = -1;

  /* Set up REG_IMMED array.  This is non-zero for a register if we know its
     preferred slot currently holds this immediate value.  */
  for (i = 0; i < SPU_NUM_GPRS; i++)
      reg_immed[i] = 0;

  /* Scan instructions until the first branch.

     The following instructions are important prolog components:

	- The first instruction to set up the stack pointer.
	- The first instruction to set up the frame pointer.
	- The first instruction to save the link register.
	- The first instruction to save the backchain.

     We return the instruction after the latest of these four,
     or the incoming PC if none is found.  The first instruction
     to set up the stack pointer also defines the frame size.

     Note that instructions saving incoming arguments to their stack
     slots are not counted as important, because they are hard to
     identify with certainty.  This should not matter much, because
     arguments are relevant only in code compiled with debug data,
     and in such code the GDB core will advance until the first source
     line anyway, using SAL data.

     For purposes of stack unwinding, we analyze the following types
     of instructions in addition:

      - Any instruction adding to the current frame pointer.
      - Any instruction loading an immediate constant into a register.
      - Any instruction storing a register onto the stack.

     These are used to compute the CFA and REG_OFFSET output.  */

  for (pc = start_pc; pc < end_pc; pc += 4)
    {
      unsigned int insn;
      int rt, ra, rb, rc, immed;

      if (target_read_memory (pc, buf, 4))
	break;
      insn = extract_unsigned_integer (buf, 4, byte_order);

      /* AI is the typical instruction to set up a stack frame.
         It is also used to initialize the frame pointer.  */
      if (is_ri10 (insn, op_ai, &rt, &ra, &immed))
	{
	  if (rt == data->cfa_reg && ra == data->cfa_reg)
	    data->cfa_offset -= immed;

	  if (rt == SPU_RAW_SP_REGNUM && ra == SPU_RAW_SP_REGNUM
	      && !found_sp)
	    {
	      found_sp = 1;
	      prolog_pc = pc + 4;

	      data->size = -immed;
	    }
	  else if (rt == SPU_FP_REGNUM && ra == SPU_RAW_SP_REGNUM
		   && !found_fp)
	    {
	      found_fp = 1;
	      prolog_pc = pc + 4;

	      data->cfa_reg = SPU_FP_REGNUM;
	      data->cfa_offset -= immed;
	    }
	}

      /* A is used to set up stack frames of size >= 512 bytes.
         If we have tracked the contents of the addend register,
         we can handle this as well.  */
      else if (is_rr (insn, op_a, &rt, &ra, &rb))
	{
	  if (rt == data->cfa_reg && ra == data->cfa_reg)
	    {
	      if (reg_immed[rb] != 0)
		data->cfa_offset -= reg_immed[rb];
	      else
		data->cfa_reg = -1;  /* We don't know the CFA any more.  */
	    }

	  if (rt == SPU_RAW_SP_REGNUM && ra == SPU_RAW_SP_REGNUM
	      && !found_sp)
	    {
	      found_sp = 1;
	      prolog_pc = pc + 4;

	      if (reg_immed[rb] != 0)
		data->size = -reg_immed[rb];
	    }
	}

      /* We need to track IL and ILA used to load immediate constants
         in case they are later used as input to an A instruction.  */
      else if (is_ri16 (insn, op_il, &rt, &immed))
	{
	  reg_immed[rt] = immed;

	  if (rt == SPU_RAW_SP_REGNUM && !found_sp)
	    found_sp = 1;
	}

      else if (is_ri18 (insn, op_ila, &rt, &immed))
	{
	  reg_immed[rt] = immed & 0x3ffff;

	  if (rt == SPU_RAW_SP_REGNUM && !found_sp)
	    found_sp = 1;
	}

      /* STQD is used to save registers to the stack.  */
      else if (is_ri10 (insn, op_stqd, &rt, &ra, &immed))
	{
	  if (ra == data->cfa_reg)
	    data->reg_offset[rt] = data->cfa_offset - (immed << 4);

	  if (ra == data->cfa_reg && rt == SPU_LR_REGNUM
              && !found_lr)
	    {
	      found_lr = 1;
	      prolog_pc = pc + 4;
	    }

	  if (ra == SPU_RAW_SP_REGNUM
	      && (found_sp? immed == 0 : rt == SPU_RAW_SP_REGNUM)
	      && !found_bc)
	    {
	      found_bc = 1;
	      prolog_pc = pc + 4;
	    }
	}

      /* _start uses SELB to set up the stack pointer.  */
      else if (is_rrr (insn, op_selb, &rt, &ra, &rb, &rc))
	{
	  if (rt == SPU_RAW_SP_REGNUM && !found_sp)
	    found_sp = 1;
	}

      /* We terminate if we find a branch.  */
      else if (is_branch (insn, &immed, &ra))
	break;
    }


  /* If we successfully parsed until here, and didn't find any instruction
     modifying SP, we assume we have a frameless function.  */
  if (!found_sp)
    data->size = 0;

  /* Return cooked instead of raw SP.  */
  if (data->cfa_reg == SPU_RAW_SP_REGNUM)
    data->cfa_reg = SPU_SP_REGNUM;

  return prolog_pc;
}

/* Return the first instruction after the prologue starting at PC.  */
static CORE_ADDR
spu_skip_prologue (struct gdbarch *gdbarch, CORE_ADDR pc)
{
  struct spu_prologue_data data;
  return spu_analyze_prologue (gdbarch, pc, (CORE_ADDR)-1, &data);
}

/* Return the frame pointer in use at address PC.  */
static void
spu_virtual_frame_pointer (struct gdbarch *gdbarch, CORE_ADDR pc,
			   int *reg, LONGEST *offset)
{
  struct spu_prologue_data data;
  spu_analyze_prologue (gdbarch, pc, (CORE_ADDR)-1, &data);

  if (data.size != -1 && data.cfa_reg != -1)
    {
      /* The 'frame pointer' address is CFA minus frame size.  */
      *reg = data.cfa_reg;
      *offset = data.cfa_offset - data.size;
    }
  else
    {
      /* ??? We don't really know ...  */
      *reg = SPU_SP_REGNUM;
      *offset = 0;
    }
}

/* Implement the stack_frame_destroyed_p gdbarch method.

   1) scan forward from the point of execution:
       a) If you find an instruction that modifies the stack pointer
          or transfers control (except a return), execution is not in
          an epilogue, return.
       b) Stop scanning if you find a return instruction or reach the
          end of the function or reach the hard limit for the size of
          an epilogue.
   2) scan backward from the point of execution:
        a) If you find an instruction that modifies the stack pointer,
            execution *is* in an epilogue, return.
        b) Stop scanning if you reach an instruction that transfers
           control or the beginning of the function or reach the hard
           limit for the size of an epilogue.  */

static int
spu_stack_frame_destroyed_p (struct gdbarch *gdbarch, CORE_ADDR pc)
{
  enum bfd_endian byte_order = gdbarch_byte_order (gdbarch);
  CORE_ADDR scan_pc, func_start, func_end, epilogue_start, epilogue_end;
  bfd_byte buf[4];
  unsigned int insn;
  int rt, ra, rb, immed;

  /* Find the search limits based on function boundaries and hard limit.
     We assume the epilogue can be up to 64 instructions long.  */

  const int spu_max_epilogue_size = 64 * 4;

  if (!find_pc_partial_function (pc, NULL, &func_start, &func_end))
    return 0;

  if (pc - func_start < spu_max_epilogue_size)
    epilogue_start = func_start;
  else
    epilogue_start = pc - spu_max_epilogue_size;

  if (func_end - pc < spu_max_epilogue_size)
    epilogue_end = func_end;
  else
    epilogue_end = pc + spu_max_epilogue_size;

  /* Scan forward until next 'bi $0'.  */

  for (scan_pc = pc; scan_pc < epilogue_end; scan_pc += 4)
    {
      if (target_read_memory (scan_pc, buf, 4))
	return 0;
      insn = extract_unsigned_integer (buf, 4, byte_order);

      if (is_branch (insn, &immed, &ra))
	{
	  if (immed == 0 && ra == SPU_LR_REGNUM)
	    break;

	  return 0;
	}

      if (is_ri10 (insn, op_ai, &rt, &ra, &immed)
	  || is_rr (insn, op_a, &rt, &ra, &rb)
	  || is_ri10 (insn, op_lqd, &rt, &ra, &immed))
	{
	  if (rt == SPU_RAW_SP_REGNUM)
	    return 0;
	}
    }

  if (scan_pc >= epilogue_end)
    return 0;

  /* Scan backward until adjustment to stack pointer (R1).  */

  for (scan_pc = pc - 4; scan_pc >= epilogue_start; scan_pc -= 4)
    {
      if (target_read_memory (scan_pc, buf, 4))
	return 0;
      insn = extract_unsigned_integer (buf, 4, byte_order);

      if (is_branch (insn, &immed, &ra))
	return 0;

      if (is_ri10 (insn, op_ai, &rt, &ra, &immed)
	  || is_rr (insn, op_a, &rt, &ra, &rb)
	  || is_ri10 (insn, op_lqd, &rt, &ra, &immed))
	{
	  if (rt == SPU_RAW_SP_REGNUM)
	    return 1;
	}
    }

  return 0;
}


/* Normal stack frames.  */

struct spu_unwind_cache
{
  CORE_ADDR func;
  CORE_ADDR frame_base;
  CORE_ADDR local_base;

  struct trad_frame_saved_reg *saved_regs;
};

static struct spu_unwind_cache *
spu_frame_unwind_cache (struct frame_info *this_frame,
			void **this_prologue_cache)
{
  struct gdbarch *gdbarch = get_frame_arch (this_frame);
  struct gdbarch_tdep *tdep = gdbarch_tdep (gdbarch);
  enum bfd_endian byte_order = gdbarch_byte_order (gdbarch);
  struct spu_unwind_cache *info;
  struct spu_prologue_data data;
  CORE_ADDR id = tdep->id;
  gdb_byte buf[16];

  if (*this_prologue_cache)
    return (struct spu_unwind_cache *) *this_prologue_cache;

  info = FRAME_OBSTACK_ZALLOC (struct spu_unwind_cache);
  *this_prologue_cache = info;
  info->saved_regs = trad_frame_alloc_saved_regs (this_frame);
  info->frame_base = 0;
  info->local_base = 0;

  /* Find the start of the current function, and analyze its prologue.  */
  info->func = get_frame_func (this_frame);
  if (info->func == 0)
    {
      /* Fall back to using the current PC as frame ID.  */
      info->func = get_frame_pc (this_frame);
      data.size = -1;
    }
  else
    spu_analyze_prologue (gdbarch, info->func, get_frame_pc (this_frame),
			  &data);

  /* If successful, use prologue analysis data.  */
  if (data.size != -1 && data.cfa_reg != -1)
    {
      CORE_ADDR cfa;
      int i;

      /* Determine CFA via unwound CFA_REG plus CFA_OFFSET.  */
      get_frame_register (this_frame, data.cfa_reg, buf);
      cfa = extract_unsigned_integer (buf, 4, byte_order) + data.cfa_offset;
      cfa = SPUADDR (id, cfa);

      /* Call-saved register slots.  */
      for (i = 0; i < SPU_NUM_GPRS; i++)
	if (i == SPU_LR_REGNUM
	    || (i >= SPU_SAVED1_REGNUM && i <= SPU_SAVEDN_REGNUM))
	  if (data.reg_offset[i] != -1)
	    info->saved_regs[i].addr = cfa - data.reg_offset[i];

      /* Frame bases.  */
      info->frame_base = cfa;
      info->local_base = cfa - data.size;
    }

  /* Otherwise, fall back to reading the backchain link.  */
  else
    {
      CORE_ADDR reg;
      LONGEST backchain;
      ULONGEST lslr;
      int status;

      /* Get local store limit.  */
      lslr = get_frame_register_unsigned (this_frame, SPU_LSLR_REGNUM);
      if (!lslr)
	lslr = (ULONGEST) -1;

      /* Get the backchain.  */
      reg = get_frame_register_unsigned (this_frame, SPU_SP_REGNUM);
      status = safe_read_memory_integer (SPUADDR (id, reg), 4, byte_order,
					 &backchain);

      /* A zero backchain terminates the frame chain.  Also, sanity
         check against the local store size limit.  */
      if (status && backchain > 0 && backchain <= lslr)
	{
	  /* Assume the link register is saved into its slot.  */
	  if (backchain + 16 <= lslr)
	    info->saved_regs[SPU_LR_REGNUM].addr = SPUADDR (id,
							    backchain + 16);

          /* Frame bases.  */
	  info->frame_base = SPUADDR (id, backchain);
	  info->local_base = SPUADDR (id, reg);
	}
    }

  /* If we didn't find a frame, we cannot determine SP / return address.  */
  if (info->frame_base == 0)
    return info;

  /* The previous SP is equal to the CFA.  */
  trad_frame_set_value (info->saved_regs, SPU_SP_REGNUM,
			SPUADDR_ADDR (info->frame_base));

  /* Read full contents of the unwound link register in order to
     be able to determine the return address.  */
  if (trad_frame_addr_p (info->saved_regs, SPU_LR_REGNUM))
    target_read_memory (info->saved_regs[SPU_LR_REGNUM].addr, buf, 16);
  else
    get_frame_register (this_frame, SPU_LR_REGNUM, buf);

  /* Normally, the return address is contained in the slot 0 of the
     link register, and slots 1-3 are zero.  For an overlay return,
     slot 0 contains the address of the overlay manager return stub,
     slot 1 contains the partition number of the overlay section to
     be returned to, and slot 2 contains the return address within
     that section.  Return the latter address in that case.  */
  if (extract_unsigned_integer (buf + 8, 4, byte_order) != 0)
    trad_frame_set_value (info->saved_regs, SPU_PC_REGNUM,
			  extract_unsigned_integer (buf + 8, 4, byte_order));
  else
    trad_frame_set_value (info->saved_regs, SPU_PC_REGNUM,
			  extract_unsigned_integer (buf, 4, byte_order));
 
  return info;
}

static void
spu_frame_this_id (struct frame_info *this_frame,
		   void **this_prologue_cache, struct frame_id *this_id)
{
  struct spu_unwind_cache *info =
    spu_frame_unwind_cache (this_frame, this_prologue_cache);

  if (info->frame_base == 0)
    return;

  *this_id = frame_id_build (info->frame_base, info->func);
}

static struct value *
spu_frame_prev_register (struct frame_info *this_frame,
			 void **this_prologue_cache, int regnum)
{
  struct spu_unwind_cache *info
    = spu_frame_unwind_cache (this_frame, this_prologue_cache);

  /* Special-case the stack pointer.  */
  if (regnum == SPU_RAW_SP_REGNUM)
    regnum = SPU_SP_REGNUM;

  return trad_frame_get_prev_register (this_frame, info->saved_regs, regnum);
}

static const struct frame_unwind spu_frame_unwind = {
  NORMAL_FRAME,
  default_frame_unwind_stop_reason,
  spu_frame_this_id,
  spu_frame_prev_register,
  NULL,
  default_frame_sniffer
};

static CORE_ADDR
spu_frame_base_address (struct frame_info *this_frame, void **this_cache)
{
  struct spu_unwind_cache *info
    = spu_frame_unwind_cache (this_frame, this_cache);
  return info->local_base;
}

static const struct frame_base spu_frame_base = {
  &spu_frame_unwind,
  spu_frame_base_address,
  spu_frame_base_address,
  spu_frame_base_address
};

static CORE_ADDR
spu_unwind_pc (struct gdbarch *gdbarch, struct frame_info *next_frame)
{
  struct gdbarch_tdep *tdep = gdbarch_tdep (gdbarch);
  CORE_ADDR pc = frame_unwind_register_unsigned (next_frame, SPU_PC_REGNUM);
  /* Mask off interrupt enable bit.  */
  return SPUADDR (tdep->id, pc & -4);
}

static CORE_ADDR
spu_unwind_sp (struct gdbarch *gdbarch, struct frame_info *next_frame)
{
  struct gdbarch_tdep *tdep = gdbarch_tdep (gdbarch);
  CORE_ADDR sp = frame_unwind_register_unsigned (next_frame, SPU_SP_REGNUM);
  return SPUADDR (tdep->id, sp);
}

static CORE_ADDR
spu_read_pc (struct regcache *regcache)
{
  struct gdbarch_tdep *tdep = gdbarch_tdep (get_regcache_arch (regcache));
  ULONGEST pc;
  regcache_cooked_read_unsigned (regcache, SPU_PC_REGNUM, &pc);
  /* Mask off interrupt enable bit.  */
  return SPUADDR (tdep->id, pc & -4);
}

static void
spu_write_pc (struct regcache *regcache, CORE_ADDR pc)
{
  /* Keep interrupt enabled state unchanged.  */
  ULONGEST old_pc;

  regcache_cooked_read_unsigned (regcache, SPU_PC_REGNUM, &old_pc);
  regcache_cooked_write_unsigned (regcache, SPU_PC_REGNUM,
				  (SPUADDR_ADDR (pc) & -4) | (old_pc & 3));
}


/* Cell/B.E. cross-architecture unwinder support.  */

struct spu2ppu_cache
{
  struct frame_id frame_id;
  struct regcache *regcache;
};

static struct gdbarch *
spu2ppu_prev_arch (struct frame_info *this_frame, void **this_cache)
{
  struct spu2ppu_cache *cache = (struct spu2ppu_cache *) *this_cache;
  return get_regcache_arch (cache->regcache);
}

static void
spu2ppu_this_id (struct frame_info *this_frame,
		 void **this_cache, struct frame_id *this_id)
{
  struct spu2ppu_cache *cache = (struct spu2ppu_cache *) *this_cache;
  *this_id = cache->frame_id;
}

static struct value *
spu2ppu_prev_register (struct frame_info *this_frame,
		       void **this_cache, int regnum)
{
  struct spu2ppu_cache *cache = (struct spu2ppu_cache *) *this_cache;
  struct gdbarch *gdbarch = get_regcache_arch (cache->regcache);
  gdb_byte *buf;

  buf = (gdb_byte *) alloca (register_size (gdbarch, regnum));
  regcache_cooked_read (cache->regcache, regnum, buf);
  return frame_unwind_got_bytes (this_frame, regnum, buf);
}

static int
spu2ppu_sniffer (const struct frame_unwind *self,
		 struct frame_info *this_frame, void **this_prologue_cache)
{
  struct gdbarch *gdbarch = get_frame_arch (this_frame);
  enum bfd_endian byte_order = gdbarch_byte_order (gdbarch);
  CORE_ADDR base, func, backchain;
  gdb_byte buf[4];

  if (gdbarch_bfd_arch_info (target_gdbarch ())->arch == bfd_arch_spu)
    return 0;

  base = get_frame_sp (this_frame);
  func = get_frame_pc (this_frame);
  if (target_read_memory (base, buf, 4))
    return 0;
  backchain = extract_unsigned_integer (buf, 4, byte_order);

  if (!backchain)
    {
      struct frame_info *fi;

      struct spu2ppu_cache *cache
	= FRAME_OBSTACK_CALLOC (1, struct spu2ppu_cache);

      cache->frame_id = frame_id_build (base + 16, func);

      for (fi = get_next_frame (this_frame); fi; fi = get_next_frame (fi))
	if (gdbarch_bfd_arch_info (get_frame_arch (fi))->arch != bfd_arch_spu)
	  break;

      if (fi)
	{
	  cache->regcache = frame_save_as_regcache (fi);
	  *this_prologue_cache = cache;
	  return 1;
	}
      else
	{
	  struct regcache *regcache;
	  regcache = get_thread_arch_regcache (inferior_ptid, target_gdbarch ());
	  cache->regcache = regcache_dup (regcache);
	  *this_prologue_cache = cache;
	  return 1;
	}
    }

  return 0;
}

static void
spu2ppu_dealloc_cache (struct frame_info *self, void *this_cache)
{
  struct spu2ppu_cache *cache = (struct spu2ppu_cache *) this_cache;
  regcache_xfree (cache->regcache);
}

static const struct frame_unwind spu2ppu_unwind = {
  ARCH_FRAME,
  default_frame_unwind_stop_reason,
  spu2ppu_this_id,
  spu2ppu_prev_register,
  NULL,
  spu2ppu_sniffer,
  spu2ppu_dealloc_cache,
  spu2ppu_prev_arch,
};


/* Function calling convention.  */

static CORE_ADDR
spu_frame_align (struct gdbarch *gdbarch, CORE_ADDR sp)
{
  return sp & ~15;
}

static CORE_ADDR
spu_push_dummy_code (struct gdbarch *gdbarch, CORE_ADDR sp, CORE_ADDR funaddr,
		     struct value **args, int nargs, struct type *value_type,
		     CORE_ADDR *real_pc, CORE_ADDR *bp_addr,
		     struct regcache *regcache)
{
  /* Allocate space sufficient for a breakpoint, keeping the stack aligned.  */
  sp = (sp - 4) & ~15;
  /* Store the address of that breakpoint */
  *bp_addr = sp;
  /* The call starts at the callee's entry point.  */
  *real_pc = funaddr;

  return sp;
}

static int
spu_scalar_value_p (struct type *type)
{
  switch (TYPE_CODE (type))
    {
    case TYPE_CODE_INT:
    case TYPE_CODE_ENUM:
    case TYPE_CODE_RANGE:
    case TYPE_CODE_CHAR:
    case TYPE_CODE_BOOL:
    case TYPE_CODE_PTR:
    case TYPE_CODE_REF:
    case TYPE_CODE_RVALUE_REF:
      return TYPE_LENGTH (type) <= 16;

    default:
      return 0;
    }
}

static void
spu_value_to_regcache (struct regcache *regcache, int regnum,
		       struct type *type, const gdb_byte *in)
{
  int len = TYPE_LENGTH (type);

  if (spu_scalar_value_p (type))
    {
      int preferred_slot = len < 4 ? 4 - len : 0;
      regcache_cooked_write_part (regcache, regnum, preferred_slot, len, in);
    }
  else
    {
      while (len >= 16)
	{
	  regcache_cooked_write (regcache, regnum++, in);
	  in += 16;
	  len -= 16;
	}

      if (len > 0)
	regcache_cooked_write_part (regcache, regnum, 0, len, in);
    }
}

static void
spu_regcache_to_value (struct regcache *regcache, int regnum,
		       struct type *type, gdb_byte *out)
{
  int len = TYPE_LENGTH (type);

  if (spu_scalar_value_p (type))
    {
      int preferred_slot = len < 4 ? 4 - len : 0;
      regcache_cooked_read_part (regcache, regnum, preferred_slot, len, out);
    }
  else
    {
      while (len >= 16)
	{
	  regcache_cooked_read (regcache, regnum++, out);
	  out += 16;
	  len -= 16;
	}

      if (len > 0)
	regcache_cooked_read_part (regcache, regnum, 0, len, out);
    }
}

static CORE_ADDR
spu_push_dummy_call (struct gdbarch *gdbarch, struct value *function,
		     struct regcache *regcache, CORE_ADDR bp_addr,
		     int nargs, struct value **args, CORE_ADDR sp,
		     int struct_return, CORE_ADDR struct_addr)
{
  enum bfd_endian byte_order = gdbarch_byte_order (gdbarch);
  CORE_ADDR sp_delta;
  int i;
  int regnum = SPU_ARG1_REGNUM;
  int stack_arg = -1;
  gdb_byte buf[16];

  /* Set the return address.  */
  memset (buf, 0, sizeof buf);
  store_unsigned_integer (buf, 4, byte_order, SPUADDR_ADDR (bp_addr));
  regcache_cooked_write (regcache, SPU_LR_REGNUM, buf);

  /* If STRUCT_RETURN is true, then the struct return address (in
     STRUCT_ADDR) will consume the first argument-passing register.
     Both adjust the register count and store that value.  */
  if (struct_return)
    {
      memset (buf, 0, sizeof buf);
      store_unsigned_integer (buf, 4, byte_order, SPUADDR_ADDR (struct_addr));
      regcache_cooked_write (regcache, regnum++, buf);
    }

  /* Fill in argument registers.  */
  for (i = 0; i < nargs; i++)
    {
      struct value *arg = args[i];
      struct type *type = check_typedef (value_type (arg));
      const gdb_byte *contents = value_contents (arg);
      int n_regs = align_up (TYPE_LENGTH (type), 16) / 16;

      /* If the argument doesn't wholly fit into registers, it and
	 all subsequent arguments go to the stack.  */
      if (regnum + n_regs - 1 > SPU_ARGN_REGNUM)
	{
	  stack_arg = i;
	  break;
	}

      spu_value_to_regcache (regcache, regnum, type, contents);
      regnum += n_regs;
    }

  /* Overflow arguments go to the stack.  */
  if (stack_arg != -1)
    {
      CORE_ADDR ap;

      /* Allocate all required stack size.  */
      for (i = stack_arg; i < nargs; i++)
	{
	  struct type *type = check_typedef (value_type (args[i]));
	  sp -= align_up (TYPE_LENGTH (type), 16);
	}

      /* Fill in stack arguments.  */
      ap = sp;
      for (i = stack_arg; i < nargs; i++)
	{
	  struct value *arg = args[i];
	  struct type *type = check_typedef (value_type (arg));
	  int len = TYPE_LENGTH (type);
	  int preferred_slot;
	  
	  if (spu_scalar_value_p (type))
	    preferred_slot = len < 4 ? 4 - len : 0;
	  else
	    preferred_slot = 0;

	  target_write_memory (ap + preferred_slot, value_contents (arg), len);
	  ap += align_up (TYPE_LENGTH (type), 16);
	}
    }

  /* Allocate stack frame header.  */
  sp -= 32;

  /* Store stack back chain.  */
  regcache_cooked_read (regcache, SPU_RAW_SP_REGNUM, buf);
  target_write_memory (sp, buf, 16);

  /* Finally, update all slots of the SP register.  */
  sp_delta = sp - extract_unsigned_integer (buf, 4, byte_order);
  for (i = 0; i < 4; i++)
    {
      CORE_ADDR sp_slot = extract_unsigned_integer (buf + 4*i, 4, byte_order);
      store_unsigned_integer (buf + 4*i, 4, byte_order, sp_slot + sp_delta);
    }
  regcache_cooked_write (regcache, SPU_RAW_SP_REGNUM, buf);

  return sp;
}

static struct frame_id
spu_dummy_id (struct gdbarch *gdbarch, struct frame_info *this_frame)
{
  struct gdbarch_tdep *tdep = gdbarch_tdep (gdbarch);
  CORE_ADDR pc = get_frame_register_unsigned (this_frame, SPU_PC_REGNUM);
  CORE_ADDR sp = get_frame_register_unsigned (this_frame, SPU_SP_REGNUM);
  return frame_id_build (SPUADDR (tdep->id, sp), SPUADDR (tdep->id, pc & -4));
}

/* Function return value access.  */

static enum return_value_convention
spu_return_value (struct gdbarch *gdbarch, struct value *function,
		  struct type *type, struct regcache *regcache,
		  gdb_byte *out, const gdb_byte *in)
{
  struct type *func_type = function ? value_type (function) : NULL;
  enum return_value_convention rvc;
  int opencl_vector = 0;

  if (func_type)
    {
      func_type = check_typedef (func_type);

      if (TYPE_CODE (func_type) == TYPE_CODE_PTR)
	func_type = check_typedef (TYPE_TARGET_TYPE (func_type));

      if (TYPE_CODE (func_type) == TYPE_CODE_FUNC
	  && TYPE_CALLING_CONVENTION (func_type) == DW_CC_GDB_IBM_OpenCL
	  && TYPE_CODE (type) == TYPE_CODE_ARRAY
	  && TYPE_VECTOR (type))
	opencl_vector = 1;
    }

  if (TYPE_LENGTH (type) <= (SPU_ARGN_REGNUM - SPU_ARG1_REGNUM + 1) * 16)
    rvc = RETURN_VALUE_REGISTER_CONVENTION;
  else
    rvc = RETURN_VALUE_STRUCT_CONVENTION;

  if (in)
    {
      switch (rvc)
	{
	case RETURN_VALUE_REGISTER_CONVENTION:
	  if (opencl_vector && TYPE_LENGTH (type) == 2)
	    regcache_cooked_write_part (regcache, SPU_ARG1_REGNUM, 2, 2, in);
	  else
	    spu_value_to_regcache (regcache, SPU_ARG1_REGNUM, type, in);
	  break;

	case RETURN_VALUE_STRUCT_CONVENTION:
	  error (_("Cannot set function return value."));
	  break;
	}
    }
  else if (out)
    {
      switch (rvc)
	{
	case RETURN_VALUE_REGISTER_CONVENTION:
	  if (opencl_vector && TYPE_LENGTH (type) == 2)
	    regcache_cooked_read_part (regcache, SPU_ARG1_REGNUM, 2, 2, out);
	  else
	    spu_regcache_to_value (regcache, SPU_ARG1_REGNUM, type, out);
	  break;

	case RETURN_VALUE_STRUCT_CONVENTION:
	  error (_("Function return value unknown."));
	  break;
	}
    }

  return rvc;
}


/* Breakpoints.  */
constexpr gdb_byte spu_break_insn[] = { 0x00, 0x00, 0x3f, 0xff };

typedef BP_MANIPULATION (spu_break_insn) spu_breakpoint;

static int
spu_memory_remove_breakpoint (struct gdbarch *gdbarch,
			      struct bp_target_info *bp_tgt)
{
  /* We work around a problem in combined Cell/B.E. debugging here.  Consider
     that in a combined application, we have some breakpoints inserted in SPU
     code, and now the application forks (on the PPU side).  GDB common code
     will assume that the fork system call copied all breakpoints into the new
     process' address space, and that all those copies now need to be removed
     (see breakpoint.c:detach_breakpoints).

     While this is certainly true for PPU side breakpoints, it is not true
     for SPU side breakpoints.  fork will clone the SPU context file
     descriptors, so that all the existing SPU contexts are in accessible
     in the new process.  However, the contents of the SPU contexts themselves
     are *not* cloned.  Therefore the effect of detach_breakpoints is to
     remove SPU breakpoints from the *original* SPU context's local store
     -- this is not the correct behaviour.

     The workaround is to check whether the PID we are asked to remove this
     breakpoint from (i.e. ptid_get_pid (inferior_ptid)) is different from the
     PID of the current inferior (i.e. current_inferior ()->pid).  This is only
     true in the context of detach_breakpoints.  If so, we simply do nothing.
     [ Note that for the fork child process, it does not matter if breakpoints
     remain inserted, because those SPU contexts are not runnable anyway --
     the Linux kernel allows only the original process to invoke spu_run.  */

  if (ptid_get_pid (inferior_ptid) != current_inferior ()->pid) 
    return 0;

  return default_memory_remove_breakpoint (gdbarch, bp_tgt);
}


/* Software single-stepping support.  */

static std::vector<CORE_ADDR>
spu_software_single_step (struct regcache *regcache)
{
  struct gdbarch *gdbarch = get_regcache_arch (regcache);
  enum bfd_endian byte_order = gdbarch_byte_order (gdbarch);
  CORE_ADDR pc, next_pc;
  unsigned int insn;
  int offset, reg;
  gdb_byte buf[4];
  ULONGEST lslr;
  std::vector<CORE_ADDR> next_pcs;

  pc = regcache_read_pc (regcache);

  if (target_read_memory (pc, buf, 4))
    throw_error (MEMORY_ERROR, _("Could not read instruction at %s."),
		 paddress (gdbarch, pc));

  insn = extract_unsigned_integer (buf, 4, byte_order);

  /* Get local store limit.  */
  lslr = regcache_raw_get_unsigned (regcache, SPU_LSLR_REGNUM);
  if (!lslr)
    lslr = (ULONGEST) -1;

  /* Next sequential instruction is at PC + 4, except if the current
     instruction is a PPE-assisted call, in which case it is at PC + 8.
     Wrap around LS limit to be on the safe side.  */
  if ((insn & 0xffffff00) == 0x00002100)
    next_pc = (SPUADDR_ADDR (pc) + 8) & lslr;
  else
    next_pc = (SPUADDR_ADDR (pc) + 4) & lslr;

  next_pcs.push_back (SPUADDR (SPUADDR_SPU (pc), next_pc));

  if (is_branch (insn, &offset, &reg))
    {
      CORE_ADDR target = offset;

      if (reg == SPU_PC_REGNUM)
	target += SPUADDR_ADDR (pc);
      else if (reg != -1)
	target += regcache_raw_get_unsigned (regcache, reg) & -4;

      target = target & lslr;
      if (target != next_pc)
	next_pcs.push_back (SPUADDR (SPUADDR_SPU (pc), target));
    }

  return next_pcs;
}


/* Longjmp support.  */

static int
spu_get_longjmp_target (struct frame_info *frame, CORE_ADDR *pc)
{
  struct gdbarch *gdbarch = get_frame_arch (frame);
  struct gdbarch_tdep *tdep = gdbarch_tdep (gdbarch);
  enum bfd_endian byte_order = gdbarch_byte_order (gdbarch);
  gdb_byte buf[4];
  CORE_ADDR jb_addr;
  int optim, unavail;

  /* Jump buffer is pointed to by the argument register $r3.  */
  if (!get_frame_register_bytes (frame, SPU_ARG1_REGNUM, 0, 4, buf,
				 &optim, &unavail))
    return 0;

  jb_addr = extract_unsigned_integer (buf, 4, byte_order);
  if (target_read_memory (SPUADDR (tdep->id, jb_addr), buf, 4))
    return 0;

  *pc = extract_unsigned_integer (buf, 4, byte_order);
  *pc = SPUADDR (tdep->id, *pc);
  return 1;
}


/* Disassembler.  */

struct spu_dis_asm_info : disassemble_info
{
  int id;
};

static void
spu_dis_asm_print_address (bfd_vma addr, struct disassemble_info *info)
{
  struct spu_dis_asm_info *data = (struct spu_dis_asm_info *) info;
  gdb_disassembler *di
    = static_cast<gdb_disassembler *>(info->application_data);

  print_address (di->arch (), SPUADDR (data->id, addr),
		 (struct ui_file *) info->stream);
}

static int
gdb_print_insn_spu (bfd_vma memaddr, struct disassemble_info *info)
{
  /* The opcodes disassembler does 18-bit address arithmetic.  Make
     sure the SPU ID encoded in the high bits is added back when we
     call print_address.  */
  struct spu_dis_asm_info spu_info;

  memcpy (&spu_info, info, sizeof (*info));
  spu_info.id = SPUADDR_SPU (memaddr);
  spu_info.print_address_func = spu_dis_asm_print_address;
  return default_print_insn (memaddr, &spu_info);
}


/* Target overlays for the SPU overlay manager.

   See the documentation of simple_overlay_update for how the
   interface is supposed to work.

   Data structures used by the overlay manager:

   struct ovly_table
     {
        u32 vma;
        u32 size;
        u32 pos;
        u32 buf;
     } _ovly_table[];   -- one entry per overlay section

   struct ovly_buf_table
     {
        u32 mapped;
     } _ovly_buf_table[];  -- one entry per overlay buffer

   _ovly_table should never change.

   Both tables are aligned to a 16-byte boundary, the symbols
   _ovly_table and _ovly_buf_table are of type STT_OBJECT and their
   size set to the size of the respective array. buf in _ovly_table is
   an index into _ovly_buf_table.

   mapped is an index into _ovly_table.  Both the mapped and buf indices start
   from one to reference the first entry in their respective tables.  */

/* Using the per-objfile private data mechanism, we store for each
   objfile an array of "struct spu_overlay_table" structures, one
   for each obj_section of the objfile.  This structure holds two
   fields, MAPPED_PTR and MAPPED_VAL.  If MAPPED_PTR is zero, this
   is *not* an overlay section.  If it is non-zero, it represents
   a target address.  The overlay section is mapped iff the target
   integer at this location equals MAPPED_VAL.  */

static const struct objfile_data *spu_overlay_data;

struct spu_overlay_table
  {
    CORE_ADDR mapped_ptr;
    CORE_ADDR mapped_val;
  };

/* Retrieve the overlay table for OBJFILE.  If not already cached, read
   the _ovly_table data structure from the target and initialize the
   spu_overlay_table data structure from it.  */
static struct spu_overlay_table *
spu_get_overlay_table (struct objfile *objfile)
{
  enum bfd_endian byte_order = bfd_big_endian (objfile->obfd)?
		   BFD_ENDIAN_BIG : BFD_ENDIAN_LITTLE;
  struct bound_minimal_symbol ovly_table_msym, ovly_buf_table_msym;
  CORE_ADDR ovly_table_base, ovly_buf_table_base;
  unsigned ovly_table_size, ovly_buf_table_size;
  struct spu_overlay_table *tbl;
  struct obj_section *osect;
  gdb_byte *ovly_table;
  int i;

  tbl = (struct spu_overlay_table *) objfile_data (objfile, spu_overlay_data);
  if (tbl)
    return tbl;

  ovly_table_msym = lookup_minimal_symbol ("_ovly_table", NULL, objfile);
  if (!ovly_table_msym.minsym)
    return NULL;

  ovly_buf_table_msym = lookup_minimal_symbol ("_ovly_buf_table",
					       NULL, objfile);
  if (!ovly_buf_table_msym.minsym)
    return NULL;

  ovly_table_base = BMSYMBOL_VALUE_ADDRESS (ovly_table_msym);
  ovly_table_size = MSYMBOL_SIZE (ovly_table_msym.minsym);

  ovly_buf_table_base = BMSYMBOL_VALUE_ADDRESS (ovly_buf_table_msym);
  ovly_buf_table_size = MSYMBOL_SIZE (ovly_buf_table_msym.minsym);

  ovly_table = (gdb_byte *) xmalloc (ovly_table_size);
  read_memory (ovly_table_base, ovly_table, ovly_table_size);

  tbl = OBSTACK_CALLOC (&objfile->objfile_obstack,
			objfile->sections_end - objfile->sections,
			struct spu_overlay_table);

  for (i = 0; i < ovly_table_size / 16; i++)
    {
      CORE_ADDR vma  = extract_unsigned_integer (ovly_table + 16*i + 0,
						 4, byte_order);
      CORE_ADDR size = extract_unsigned_integer (ovly_table + 16*i + 4,
						 4, byte_order);
      CORE_ADDR pos  = extract_unsigned_integer (ovly_table + 16*i + 8,
						 4, byte_order);
      CORE_ADDR buf  = extract_unsigned_integer (ovly_table + 16*i + 12,
						 4, byte_order);

      if (buf == 0 || (buf - 1) * 4 >= ovly_buf_table_size)
	continue;

      ALL_OBJFILE_OSECTIONS (objfile, osect)
	if (vma == bfd_section_vma (objfile->obfd, osect->the_bfd_section)
	    && pos == osect->the_bfd_section->filepos)
	  {
	    int ndx = osect - objfile->sections;
	    tbl[ndx].mapped_ptr = ovly_buf_table_base + (buf - 1) * 4;
	    tbl[ndx].mapped_val = i + 1;
	    break;
	  }
    }

  xfree (ovly_table);
  set_objfile_data (objfile, spu_overlay_data, tbl);
  return tbl;
}

/* Read _ovly_buf_table entry from the target to dermine whether
   OSECT is currently mapped, and update the mapped state.  */
static void
spu_overlay_update_osect (struct obj_section *osect)
{
  enum bfd_endian byte_order = bfd_big_endian (osect->objfile->obfd)?
		   BFD_ENDIAN_BIG : BFD_ENDIAN_LITTLE;
  struct spu_overlay_table *ovly_table;
  CORE_ADDR id, val;

  ovly_table = spu_get_overlay_table (osect->objfile);
  if (!ovly_table)
    return;

  ovly_table += osect - osect->objfile->sections;
  if (ovly_table->mapped_ptr == 0)
    return;

  id = SPUADDR_SPU (obj_section_addr (osect));
  val = read_memory_unsigned_integer (SPUADDR (id, ovly_table->mapped_ptr),
				      4, byte_order);
  osect->ovly_mapped = (val == ovly_table->mapped_val);
}

/* If OSECT is NULL, then update all sections' mapped state.
   If OSECT is non-NULL, then update only OSECT's mapped state.  */
static void
spu_overlay_update (struct obj_section *osect)
{
  /* Just one section.  */
  if (osect)
    spu_overlay_update_osect (osect);

  /* All sections.  */
  else
    {
      struct objfile *objfile;

      ALL_OBJSECTIONS (objfile, osect)
	if (section_is_overlay (osect))
	  spu_overlay_update_osect (osect);
    }
}

/* Whenever a new objfile is loaded, read the target's _ovly_table.
   If there is one, go through all sections and make sure for non-
   overlay sections LMA equals VMA, while for overlay sections LMA
   is larger than SPU_OVERLAY_LMA.  */
static void
spu_overlay_new_objfile (struct objfile *objfile)
{
  struct spu_overlay_table *ovly_table;
  struct obj_section *osect;

  /* If we've already touched this file, do nothing.  */
  if (!objfile || objfile_data (objfile, spu_overlay_data) != NULL)
    return;

  /* Consider only SPU objfiles.  */
  if (bfd_get_arch (objfile->obfd) != bfd_arch_spu)
    return;

  /* Check if this objfile has overlays.  */
  ovly_table = spu_get_overlay_table (objfile);
  if (!ovly_table)
    return;

  /* Now go and fiddle with all the LMAs.  */
  ALL_OBJFILE_OSECTIONS (objfile, osect)
    {
      bfd *obfd = objfile->obfd;
      asection *bsect = osect->the_bfd_section;
      int ndx = osect - objfile->sections;

      if (ovly_table[ndx].mapped_ptr == 0)
	bfd_section_lma (obfd, bsect) = bfd_section_vma (obfd, bsect);
      else
	bfd_section_lma (obfd, bsect) = SPU_OVERLAY_LMA + bsect->filepos;
    }
}


/* Insert temporary breakpoint on "main" function of newly loaded
   SPE context OBJFILE.  */
static void
spu_catch_start (struct objfile *objfile)
{
  struct bound_minimal_symbol minsym;
  struct compunit_symtab *cust;
  CORE_ADDR pc;

  /* Do this only if requested by "set spu stop-on-load on".  */
  if (!spu_stop_on_load_p)
    return;

  /* Consider only SPU objfiles.  */
  if (!objfile || bfd_get_arch (objfile->obfd) != bfd_arch_spu)
    return;

  /* The main objfile is handled differently.  */
  if (objfile == symfile_objfile)
    return;

  /* There can be multiple symbols named "main".  Search for the
     "main" in *this* objfile.  */
  minsym = lookup_minimal_symbol ("main", NULL, objfile);
  if (!minsym.minsym)
    return;

  /* If we have debugging information, try to use it -- this
     will allow us to properly skip the prologue.  */
  pc = BMSYMBOL_VALUE_ADDRESS (minsym);
  cust
    = find_pc_sect_compunit_symtab (pc, MSYMBOL_OBJ_SECTION (minsym.objfile,
							     minsym.minsym));
  if (cust != NULL)
    {
      const struct blockvector *bv = COMPUNIT_BLOCKVECTOR (cust);
      struct block *block = BLOCKVECTOR_BLOCK (bv, GLOBAL_BLOCK);
      struct symbol *sym;
      struct symtab_and_line sal;

      sym = block_lookup_symbol (block, "main", VAR_DOMAIN);
      if (sym)
	{
	  fixup_symbol_section (sym, objfile);
	  sal = find_function_start_sal (sym, 1);
	  pc = sal.pc;
	}
    }

  /* Use a numerical address for the set_breakpoint command to avoid having
     the breakpoint re-set incorrectly.  */
  event_location_up location = new_address_location (pc, NULL, 0);
  create_breakpoint (get_objfile_arch (objfile), location.get (),
		     NULL /* cond_string */, -1 /* thread */,
		     NULL /* extra_string */,
		     0 /* parse_condition_and_thread */, 1 /* tempflag */,
		     bp_breakpoint /* type_wanted */,
		     0 /* ignore_count */,
		     AUTO_BOOLEAN_FALSE /* pending_break_support */,
		     &bkpt_breakpoint_ops /* ops */, 0 /* from_tty */,
		     1 /* enabled */, 0 /* internal  */, 0);
}


/* Look up OBJFILE loaded into FRAME's SPU context.  */
static struct objfile *
spu_objfile_from_frame (struct frame_info *frame)
{
  struct gdbarch *gdbarch = get_frame_arch (frame);
  struct gdbarch_tdep *tdep = gdbarch_tdep (gdbarch);
  struct objfile *obj;

  if (gdbarch_bfd_arch_info (gdbarch)->arch != bfd_arch_spu)
    return NULL;

  ALL_OBJFILES (obj)
    {
      if (obj->sections != obj->sections_end
	  && SPUADDR_SPU (obj_section_addr (obj->sections)) == tdep->id)
	return obj;
    }

  return NULL;
}

/* Flush cache for ea pointer access if available.  */
static void
flush_ea_cache (void)
{
  struct bound_minimal_symbol msymbol;
  struct objfile *obj;

  if (!has_stack_frames ())
    return;

  obj = spu_objfile_from_frame (get_current_frame ());
  if (obj == NULL)
    return;

  /* Lookup inferior function __cache_flush.  */
  msymbol = lookup_minimal_symbol ("__cache_flush", NULL, obj);
  if (msymbol.minsym != NULL)
    {
      struct type *type;
      CORE_ADDR addr;

      type = objfile_type (obj)->builtin_void;
      type = lookup_function_type (type);
      type = lookup_pointer_type (type);
      addr = BMSYMBOL_VALUE_ADDRESS (msymbol);

      call_function_by_hand (value_from_pointer (type, addr), 0, NULL);
    }
}

/* This handler is called when the inferior has stopped.  If it is stopped in
   SPU architecture then flush the ea cache if used.  */
static void
spu_attach_normal_stop (struct bpstats *bs, int print_frame)
{
  if (!spu_auto_flush_cache_p)
    return;

  /* Temporarily reset spu_auto_flush_cache_p to avoid recursively
     re-entering this function when __cache_flush stops.  */
  spu_auto_flush_cache_p = 0;
  flush_ea_cache ();
  spu_auto_flush_cache_p = 1;
}


/* "info spu" commands.  */

static void
info_spu_event_command (char *args, int from_tty)
{
  struct frame_info *frame = get_selected_frame (NULL);
  ULONGEST event_status = 0;
  ULONGEST event_mask = 0;
  struct cleanup *chain;
  gdb_byte buf[100];
  char annex[32];
  LONGEST len;
  int id;

  if (gdbarch_bfd_arch_info (get_frame_arch (frame))->arch != bfd_arch_spu)
    error (_("\"info spu\" is only supported on the SPU architecture."));

  id = get_frame_register_unsigned (frame, SPU_ID_REGNUM);

  xsnprintf (annex, sizeof annex, "%d/event_status", id);
  len = target_read (&current_target, TARGET_OBJECT_SPU, annex,
		     buf, 0, (sizeof (buf) - 1));
  if (len <= 0)
    error (_("Could not read event_status."));
  buf[len] = '\0';
  event_status = strtoulst ((char *) buf, NULL, 16);
 
  xsnprintf (annex, sizeof annex, "%d/event_mask", id);
  len = target_read (&current_target, TARGET_OBJECT_SPU, annex,
		     buf, 0, (sizeof (buf) - 1));
  if (len <= 0)
    error (_("Could not read event_mask."));
  buf[len] = '\0';
  event_mask = strtoulst ((char *) buf, NULL, 16);
 
  chain = make_cleanup_ui_out_tuple_begin_end (current_uiout, "SPUInfoEvent");

  if (current_uiout->is_mi_like_p ())
    {
      current_uiout->field_fmt ("event_status",
				"0x%s", phex_nz (event_status, 4));
      current_uiout->field_fmt ("event_mask",
				"0x%s", phex_nz (event_mask, 4));
    }
  else
    {
      printf_filtered (_("Event Status 0x%s\n"), phex (event_status, 4));
      printf_filtered (_("Event Mask   0x%s\n"), phex (event_mask, 4));
    }

  do_cleanups (chain);
}

static void
info_spu_signal_command (char *args, int from_tty)
{
  struct frame_info *frame = get_selected_frame (NULL);
  struct gdbarch *gdbarch = get_frame_arch (frame);
  enum bfd_endian byte_order = gdbarch_byte_order (gdbarch);
  ULONGEST signal1 = 0;
  ULONGEST signal1_type = 0;
  int signal1_pending = 0;
  ULONGEST signal2 = 0;
  ULONGEST signal2_type = 0;
  int signal2_pending = 0;
  struct cleanup *chain;
  char annex[32];
  gdb_byte buf[100];
  LONGEST len;
  int id;

  if (gdbarch_bfd_arch_info (gdbarch)->arch != bfd_arch_spu)
    error (_("\"info spu\" is only supported on the SPU architecture."));

  id = get_frame_register_unsigned (frame, SPU_ID_REGNUM);

  xsnprintf (annex, sizeof annex, "%d/signal1", id);
  len = target_read (&current_target, TARGET_OBJECT_SPU, annex, buf, 0, 4);
  if (len < 0)
    error (_("Could not read signal1."));
  else if (len == 4)
    {
      signal1 = extract_unsigned_integer (buf, 4, byte_order);
      signal1_pending = 1;
    }
    
  xsnprintf (annex, sizeof annex, "%d/signal1_type", id);
  len = target_read (&current_target, TARGET_OBJECT_SPU, annex,
		     buf, 0, (sizeof (buf) - 1));
  if (len <= 0)
    error (_("Could not read signal1_type."));
  buf[len] = '\0';
  signal1_type = strtoulst ((char *) buf, NULL, 16);

  xsnprintf (annex, sizeof annex, "%d/signal2", id);
  len = target_read (&current_target, TARGET_OBJECT_SPU, annex, buf, 0, 4);
  if (len < 0)
    error (_("Could not read signal2."));
  else if (len == 4)
    {
      signal2 = extract_unsigned_integer (buf, 4, byte_order);
      signal2_pending = 1;
    }
    
  xsnprintf (annex, sizeof annex, "%d/signal2_type", id);
  len = target_read (&current_target, TARGET_OBJECT_SPU, annex,
		     buf, 0, (sizeof (buf) - 1));
  if (len <= 0)
    error (_("Could not read signal2_type."));
  buf[len] = '\0';
  signal2_type = strtoulst ((char *) buf, NULL, 16);

  chain = make_cleanup_ui_out_tuple_begin_end (current_uiout, "SPUInfoSignal");

  if (current_uiout->is_mi_like_p ())
    {
      current_uiout->field_int ("signal1_pending", signal1_pending);
      current_uiout->field_fmt ("signal1", "0x%s", phex_nz (signal1, 4));
      current_uiout->field_int ("signal1_type", signal1_type);
      current_uiout->field_int ("signal2_pending", signal2_pending);
      current_uiout->field_fmt ("signal2", "0x%s", phex_nz (signal2, 4));
      current_uiout->field_int ("signal2_type", signal2_type);
    }
  else
    {
      if (signal1_pending)
	printf_filtered (_("Signal 1 control word 0x%s "), phex (signal1, 4));
      else
	printf_filtered (_("Signal 1 not pending "));

      if (signal1_type)
	printf_filtered (_("(Type Or)\n"));
      else
	printf_filtered (_("(Type Overwrite)\n"));

      if (signal2_pending)
	printf_filtered (_("Signal 2 control word 0x%s "), phex (signal2, 4));
      else
	printf_filtered (_("Signal 2 not pending "));

      if (signal2_type)
	printf_filtered (_("(Type Or)\n"));
      else
	printf_filtered (_("(Type Overwrite)\n"));
    }

  do_cleanups (chain);
}

static void
info_spu_mailbox_list (gdb_byte *buf, int nr, enum bfd_endian byte_order,
		       const char *field, const char *msg)
{
  struct cleanup *chain;
  int i;

  if (nr <= 0)
    return;

  chain = make_cleanup_ui_out_table_begin_end (current_uiout, 1, nr, "mbox");

  current_uiout->table_header (32, ui_left, field, msg);
  current_uiout->table_body ();

  for (i = 0; i < nr; i++)
    {
      struct cleanup *val_chain;
      ULONGEST val;
      val_chain = make_cleanup_ui_out_tuple_begin_end (current_uiout, "mbox");
      val = extract_unsigned_integer (buf + 4*i, 4, byte_order);
      current_uiout->field_fmt (field, "0x%s", phex (val, 4));
      do_cleanups (val_chain);

      if (!current_uiout->is_mi_like_p ())
	printf_filtered ("\n");
    }

  do_cleanups (chain);
}

static void
info_spu_mailbox_command (char *args, int from_tty)
{
  struct frame_info *frame = get_selected_frame (NULL);
  struct gdbarch *gdbarch = get_frame_arch (frame);
  enum bfd_endian byte_order = gdbarch_byte_order (gdbarch);
  struct cleanup *chain;
  char annex[32];
  gdb_byte buf[1024];
  LONGEST len;
  int id;

  if (gdbarch_bfd_arch_info (gdbarch)->arch != bfd_arch_spu)
    error (_("\"info spu\" is only supported on the SPU architecture."));

  id = get_frame_register_unsigned (frame, SPU_ID_REGNUM);

  chain = make_cleanup_ui_out_tuple_begin_end (current_uiout, "SPUInfoMailbox");

  xsnprintf (annex, sizeof annex, "%d/mbox_info", id);
  len = target_read (&current_target, TARGET_OBJECT_SPU, annex,
		     buf, 0, sizeof buf);
  if (len < 0)
    error (_("Could not read mbox_info."));

  info_spu_mailbox_list (buf, len / 4, byte_order,
			 "mbox", "SPU Outbound Mailbox");

  xsnprintf (annex, sizeof annex, "%d/ibox_info", id);
  len = target_read (&current_target, TARGET_OBJECT_SPU, annex,
		     buf, 0, sizeof buf);
  if (len < 0)
    error (_("Could not read ibox_info."));

  info_spu_mailbox_list (buf, len / 4, byte_order,
			 "ibox", "SPU Outbound Interrupt Mailbox");

  xsnprintf (annex, sizeof annex, "%d/wbox_info", id);
  len = target_read (&current_target, TARGET_OBJECT_SPU, annex,
		     buf, 0, sizeof buf);
  if (len < 0)
    error (_("Could not read wbox_info."));

  info_spu_mailbox_list (buf, len / 4, byte_order,
			 "wbox", "SPU Inbound Mailbox");

  do_cleanups (chain);
}

static ULONGEST
spu_mfc_get_bitfield (ULONGEST word, int first, int last)
{
  ULONGEST mask = ~(~(ULONGEST)0 << (last - first + 1));
  return (word >> (63 - last)) & mask;
}

static void
info_spu_dma_cmdlist (gdb_byte *buf, int nr, enum bfd_endian byte_order)
{
  static const char *spu_mfc_opcode[256] =
    {
    /* 00 */ NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL,
             NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL,
    /* 10 */ NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL,
             NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL,
    /* 20 */ "put", "putb", "putf", NULL, "putl", "putlb", "putlf", NULL,
             "puts", "putbs", "putfs", NULL, NULL, NULL, NULL, NULL,
    /* 30 */ "putr", "putrb", "putrf", NULL, "putrl", "putrlb", "putrlf", NULL,
             NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL,
    /* 40 */ "get", "getb", "getf", NULL, "getl", "getlb", "getlf", NULL,
             "gets", "getbs", "getfs", NULL, NULL, NULL, NULL, NULL,
    /* 50 */ NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL,
             NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL,
    /* 60 */ NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL,
             NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL,
    /* 70 */ NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL,
             NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL,
    /* 80 */ "sdcrt", "sdcrtst", NULL, NULL, NULL, NULL, NULL, NULL,
             NULL, "sdcrz", NULL, NULL, NULL, "sdcrst", NULL, "sdcrf",
    /* 90 */ NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL,
             NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL,
    /* a0 */ "sndsig", "sndsigb", "sndsigf", NULL, NULL, NULL, NULL, NULL,
             NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL,
    /* b0 */ "putlluc", NULL, NULL, NULL, "putllc", NULL, NULL, NULL,
             "putqlluc", NULL, NULL, NULL, NULL, NULL, NULL, NULL,
    /* c0 */ "barrier", NULL, NULL, NULL, NULL, NULL, NULL, NULL,
             "mfceieio", NULL, NULL, NULL, "mfcsync", NULL, NULL, NULL,
    /* d0 */ "getllar", NULL, NULL, NULL, NULL, NULL, NULL, NULL,
             NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL,
    /* e0 */ NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL,
             NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL,
    /* f0 */ NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL,
             NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL,
    };

  int *seq = XALLOCAVEC (int, nr);
  int done = 0;
  struct cleanup *chain;
  int i, j;


  /* Determine sequence in which to display (valid) entries.  */
  for (i = 0; i < nr; i++)
    {
      /* Search for the first valid entry all of whose
	 dependencies are met.  */
      for (j = 0; j < nr; j++)
	{
          ULONGEST mfc_cq_dw3;
	  ULONGEST dependencies;

	  if (done & (1 << (nr - 1 - j)))
	    continue;

	  mfc_cq_dw3
	    = extract_unsigned_integer (buf + 32*j + 24,8, byte_order);
	  if (!spu_mfc_get_bitfield (mfc_cq_dw3, 16, 16))
	    continue;

	  dependencies = spu_mfc_get_bitfield (mfc_cq_dw3, 0, nr - 1);
	  if ((dependencies & done) != dependencies)
	    continue;

	  seq[i] = j;
	  done |= 1 << (nr - 1 - j);
	  break;
	}

      if (j == nr)
	break;
    }

  nr = i;


  chain = make_cleanup_ui_out_table_begin_end (current_uiout, 10, nr,
					       "dma_cmd");

  current_uiout->table_header (7, ui_left, "opcode", "Opcode");
  current_uiout->table_header (3, ui_left, "tag", "Tag");
  current_uiout->table_header (3, ui_left, "tid", "TId");
  current_uiout->table_header (3, ui_left, "rid", "RId");
  current_uiout->table_header (18, ui_left, "ea", "EA");
  current_uiout->table_header (7, ui_left, "lsa", "LSA");
  current_uiout->table_header (7, ui_left, "size", "Size");
  current_uiout->table_header (7, ui_left, "lstaddr", "LstAddr");
  current_uiout->table_header (7, ui_left, "lstsize", "LstSize");
  current_uiout->table_header (1, ui_left, "error_p", "E");

  current_uiout->table_body ();

  for (i = 0; i < nr; i++)
    {
      struct cleanup *cmd_chain;
      ULONGEST mfc_cq_dw0;
      ULONGEST mfc_cq_dw1;
      ULONGEST mfc_cq_dw2;
      int mfc_cmd_opcode, mfc_cmd_tag, rclass_id, tclass_id;
      int list_lsa, list_size, mfc_lsa, mfc_size;
      ULONGEST mfc_ea;
      int list_valid_p, qw_valid_p, ea_valid_p, cmd_error_p;

      /* Decode contents of MFC Command Queue Context Save/Restore Registers.
	 See "Cell Broadband Engine Registers V1.3", section 3.3.2.1.  */

      mfc_cq_dw0
	= extract_unsigned_integer (buf + 32*seq[i], 8, byte_order);
      mfc_cq_dw1
	= extract_unsigned_integer (buf + 32*seq[i] + 8, 8, byte_order);
      mfc_cq_dw2
	= extract_unsigned_integer (buf + 32*seq[i] + 16, 8, byte_order);

      list_lsa = spu_mfc_get_bitfield (mfc_cq_dw0, 0, 14);
      list_size = spu_mfc_get_bitfield (mfc_cq_dw0, 15, 26);
      mfc_cmd_opcode = spu_mfc_get_bitfield (mfc_cq_dw0, 27, 34);
      mfc_cmd_tag = spu_mfc_get_bitfield (mfc_cq_dw0, 35, 39);
      list_valid_p = spu_mfc_get_bitfield (mfc_cq_dw0, 40, 40);
      rclass_id = spu_mfc_get_bitfield (mfc_cq_dw0, 41, 43);
      tclass_id = spu_mfc_get_bitfield (mfc_cq_dw0, 44, 46);

      mfc_ea = spu_mfc_get_bitfield (mfc_cq_dw1, 0, 51) << 12
		| spu_mfc_get_bitfield (mfc_cq_dw2, 25, 36);

      mfc_lsa = spu_mfc_get_bitfield (mfc_cq_dw2, 0, 13);
      mfc_size = spu_mfc_get_bitfield (mfc_cq_dw2, 14, 24);
      qw_valid_p = spu_mfc_get_bitfield (mfc_cq_dw2, 38, 38);
      ea_valid_p = spu_mfc_get_bitfield (mfc_cq_dw2, 39, 39);
      cmd_error_p = spu_mfc_get_bitfield (mfc_cq_dw2, 40, 40);

      cmd_chain = make_cleanup_ui_out_tuple_begin_end (current_uiout, "cmd");

      if (spu_mfc_opcode[mfc_cmd_opcode])
	current_uiout->field_string ("opcode", spu_mfc_opcode[mfc_cmd_opcode]);
      else
	current_uiout->field_int ("opcode", mfc_cmd_opcode);

      current_uiout->field_int ("tag", mfc_cmd_tag);
      current_uiout->field_int ("tid", tclass_id);
      current_uiout->field_int ("rid", rclass_id);

      if (ea_valid_p)
	current_uiout->field_fmt ("ea", "0x%s", phex (mfc_ea, 8));
      else
	current_uiout->field_skip ("ea");

      current_uiout->field_fmt ("lsa", "0x%05x", mfc_lsa << 4);
      if (qw_valid_p)
	current_uiout->field_fmt ("size", "0x%05x", mfc_size << 4);
      else
	current_uiout->field_fmt ("size", "0x%05x", mfc_size);

      if (list_valid_p)
	{
	  current_uiout->field_fmt ("lstaddr", "0x%05x", list_lsa << 3);
	  current_uiout->field_fmt ("lstsize", "0x%05x", list_size << 3);
	}
      else
	{
	  current_uiout->field_skip ("lstaddr");
	  current_uiout->field_skip ("lstsize");
	}

      if (cmd_error_p)
	current_uiout->field_string ("error_p", "*");
      else
	current_uiout->field_skip ("error_p");

      do_cleanups (cmd_chain);

      if (!current_uiout->is_mi_like_p ())
	printf_filtered ("\n");
    }

  do_cleanups (chain);
}

static void
info_spu_dma_command (char *args, int from_tty)
{
  struct frame_info *frame = get_selected_frame (NULL);
  struct gdbarch *gdbarch = get_frame_arch (frame);
  enum bfd_endian byte_order = gdbarch_byte_order (gdbarch);
  ULONGEST dma_info_type;
  ULONGEST dma_info_mask;
  ULONGEST dma_info_status;
  ULONGEST dma_info_stall_and_notify;
  ULONGEST dma_info_atomic_command_status;
  struct cleanup *chain;
  char annex[32];
  gdb_byte buf[1024];
  LONGEST len;
  int id;

  if (gdbarch_bfd_arch_info (get_frame_arch (frame))->arch != bfd_arch_spu)
    error (_("\"info spu\" is only supported on the SPU architecture."));

  id = get_frame_register_unsigned (frame, SPU_ID_REGNUM);

  xsnprintf (annex, sizeof annex, "%d/dma_info", id);
  len = target_read (&current_target, TARGET_OBJECT_SPU, annex,
		     buf, 0, 40 + 16 * 32);
  if (len <= 0)
    error (_("Could not read dma_info."));

  dma_info_type
    = extract_unsigned_integer (buf, 8, byte_order);
  dma_info_mask
    = extract_unsigned_integer (buf + 8, 8, byte_order);
  dma_info_status
    = extract_unsigned_integer (buf + 16, 8, byte_order);
  dma_info_stall_and_notify
    = extract_unsigned_integer (buf + 24, 8, byte_order);
  dma_info_atomic_command_status
    = extract_unsigned_integer (buf + 32, 8, byte_order);
  
  chain = make_cleanup_ui_out_tuple_begin_end (current_uiout, "SPUInfoDMA");

  if (current_uiout->is_mi_like_p ())
    {
      current_uiout->field_fmt ("dma_info_type", "0x%s",
				phex_nz (dma_info_type, 4));
      current_uiout->field_fmt ("dma_info_mask", "0x%s",
				phex_nz (dma_info_mask, 4));
      current_uiout->field_fmt ("dma_info_status", "0x%s",
				phex_nz (dma_info_status, 4));
      current_uiout->field_fmt ("dma_info_stall_and_notify", "0x%s",
				phex_nz (dma_info_stall_and_notify, 4));
      current_uiout->field_fmt ("dma_info_atomic_command_status", "0x%s",
				phex_nz (dma_info_atomic_command_status, 4));
    }
  else
    {
      const char *query_msg = _("no query pending");

      if (dma_info_type & 4)
	switch (dma_info_type & 3)
	  {
	    case 1: query_msg = _("'any' query pending"); break;
	    case 2: query_msg = _("'all' query pending"); break;
	    default: query_msg = _("undefined query type"); break;
	  }

      printf_filtered (_("Tag-Group Status  0x%s\n"),
		       phex (dma_info_status, 4));
      printf_filtered (_("Tag-Group Mask    0x%s (%s)\n"),
		       phex (dma_info_mask, 4), query_msg);
      printf_filtered (_("Stall-and-Notify  0x%s\n"),
		       phex (dma_info_stall_and_notify, 4));
      printf_filtered (_("Atomic Cmd Status 0x%s\n"),
		       phex (dma_info_atomic_command_status, 4));
      printf_filtered ("\n");
    }

  info_spu_dma_cmdlist (buf + 40, 16, byte_order);
  do_cleanups (chain);
}

static void
info_spu_proxydma_command (char *args, int from_tty)
{
  struct frame_info *frame = get_selected_frame (NULL);
  struct gdbarch *gdbarch = get_frame_arch (frame);
  enum bfd_endian byte_order = gdbarch_byte_order (gdbarch);
  ULONGEST dma_info_type;
  ULONGEST dma_info_mask;
  ULONGEST dma_info_status;
  struct cleanup *chain;
  char annex[32];
  gdb_byte buf[1024];
  LONGEST len;
  int id;

  if (gdbarch_bfd_arch_info (gdbarch)->arch != bfd_arch_spu)
    error (_("\"info spu\" is only supported on the SPU architecture."));

  id = get_frame_register_unsigned (frame, SPU_ID_REGNUM);

  xsnprintf (annex, sizeof annex, "%d/proxydma_info", id);
  len = target_read (&current_target, TARGET_OBJECT_SPU, annex,
		     buf, 0, 24 + 8 * 32);
  if (len <= 0)
    error (_("Could not read proxydma_info."));

  dma_info_type = extract_unsigned_integer (buf, 8, byte_order);
  dma_info_mask = extract_unsigned_integer (buf + 8, 8, byte_order);
  dma_info_status = extract_unsigned_integer (buf + 16, 8, byte_order);
  
  chain = make_cleanup_ui_out_tuple_begin_end (current_uiout,
					       "SPUInfoProxyDMA");

  if (current_uiout->is_mi_like_p ())
    {
      current_uiout->field_fmt ("proxydma_info_type", "0x%s",
				phex_nz (dma_info_type, 4));
      current_uiout->field_fmt ("proxydma_info_mask", "0x%s",
				phex_nz (dma_info_mask, 4));
      current_uiout->field_fmt ("proxydma_info_status", "0x%s",
				phex_nz (dma_info_status, 4));
    }
  else
    {
      const char *query_msg;

      switch (dma_info_type & 3)
	{
	case 0: query_msg = _("no query pending"); break;
	case 1: query_msg = _("'any' query pending"); break;
	case 2: query_msg = _("'all' query pending"); break;
	default: query_msg = _("undefined query type"); break;
	}

      printf_filtered (_("Tag-Group Status  0x%s\n"),
		       phex (dma_info_status, 4));
      printf_filtered (_("Tag-Group Mask    0x%s (%s)\n"),
		       phex (dma_info_mask, 4), query_msg);
      printf_filtered ("\n");
    }

  info_spu_dma_cmdlist (buf + 24, 8, byte_order);
  do_cleanups (chain);
}

static void
info_spu_command (char *args, int from_tty)
{
  printf_unfiltered (_("\"info spu\" must be followed by "
		       "the name of an SPU facility.\n"));
  help_list (infospucmdlist, "info spu ", all_commands, gdb_stdout);
}


/* Root of all "set spu "/"show spu " commands.  */

static void
show_spu_command (char *args, int from_tty)
{
  help_list (showspucmdlist, "show spu ", all_commands, gdb_stdout);
}

static void
set_spu_command (char *args, int from_tty)
{
  help_list (setspucmdlist, "set spu ", all_commands, gdb_stdout);
}

static void
show_spu_stop_on_load (struct ui_file *file, int from_tty,
                       struct cmd_list_element *c, const char *value)
{
  fprintf_filtered (file, _("Stopping for new SPE threads is %s.\n"),
                    value);
}

static void
show_spu_auto_flush_cache (struct ui_file *file, int from_tty,
			   struct cmd_list_element *c, const char *value)
{
  fprintf_filtered (file, _("Automatic software-cache flush is %s.\n"),
                    value);
}


/* Set up gdbarch struct.  */

static struct gdbarch *
spu_gdbarch_init (struct gdbarch_info info, struct gdbarch_list *arches)
{
  struct gdbarch *gdbarch;
  struct gdbarch_tdep *tdep;
  int id = -1;

  /* Which spufs ID was requested as address space?  */
  if (info.tdep_info)
    id = *(int *)info.tdep_info;
  /* For objfile architectures of SPU solibs, decode the ID from the name.
     This assumes the filename convention employed by solib-spu.c.  */
  else if (info.abfd)
    {
      const char *name = strrchr (info.abfd->filename, '@');
      if (name)
	sscanf (name, "@0x%*x <%d>", &id);
    }

  /* Find a candidate among extant architectures.  */
  for (arches = gdbarch_list_lookup_by_info (arches, &info);
       arches != NULL;
       arches = gdbarch_list_lookup_by_info (arches->next, &info))
    {
      tdep = gdbarch_tdep (arches->gdbarch);
      if (tdep && tdep->id == id)
	return arches->gdbarch;
    }

  /* None found, so create a new architecture.  */
  tdep = XCNEW (struct gdbarch_tdep);
  tdep->id = id;
  gdbarch = gdbarch_alloc (&info, tdep);

  /* Disassembler.  */
  set_gdbarch_print_insn (gdbarch, gdb_print_insn_spu);

  /* Registers.  */
  set_gdbarch_num_regs (gdbarch, SPU_NUM_REGS);
  set_gdbarch_num_pseudo_regs (gdbarch, SPU_NUM_PSEUDO_REGS);
  set_gdbarch_sp_regnum (gdbarch, SPU_SP_REGNUM);
  set_gdbarch_pc_regnum (gdbarch, SPU_PC_REGNUM);
  set_gdbarch_read_pc (gdbarch, spu_read_pc);
  set_gdbarch_write_pc (gdbarch, spu_write_pc);
  set_gdbarch_register_name (gdbarch, spu_register_name);
  set_gdbarch_register_type (gdbarch, spu_register_type);
  set_gdbarch_pseudo_register_read (gdbarch, spu_pseudo_register_read);
  set_gdbarch_pseudo_register_write (gdbarch, spu_pseudo_register_write);
  set_gdbarch_value_from_register (gdbarch, spu_value_from_register);
  set_gdbarch_register_reggroup_p (gdbarch, spu_register_reggroup_p);
  set_gdbarch_dwarf2_reg_to_regnum (gdbarch, spu_dwarf_reg_to_regnum);
  set_gdbarch_ax_pseudo_register_collect
    (gdbarch, spu_ax_pseudo_register_collect);
  set_gdbarch_ax_pseudo_register_push_stack
    (gdbarch, spu_ax_pseudo_register_push_stack);

  /* Data types.  */
  set_gdbarch_char_signed (gdbarch, 0);
  set_gdbarch_ptr_bit (gdbarch, 32);
  set_gdbarch_addr_bit (gdbarch, 32);
  set_gdbarch_short_bit (gdbarch, 16);
  set_gdbarch_int_bit (gdbarch, 32);
  set_gdbarch_long_bit (gdbarch, 32);
  set_gdbarch_long_long_bit (gdbarch, 64);
  set_gdbarch_float_bit (gdbarch, 32);
  set_gdbarch_double_bit (gdbarch, 64);
  set_gdbarch_long_double_bit (gdbarch, 64);
  set_gdbarch_float_format (gdbarch, floatformats_ieee_single);
  set_gdbarch_double_format (gdbarch, floatformats_ieee_double);
  set_gdbarch_long_double_format (gdbarch, floatformats_ieee_double);

  /* Address handling.  */
  set_gdbarch_address_to_pointer (gdbarch, spu_address_to_pointer);
  set_gdbarch_pointer_to_address (gdbarch, spu_pointer_to_address);
  set_gdbarch_integer_to_address (gdbarch, spu_integer_to_address);
  set_gdbarch_address_class_type_flags (gdbarch, spu_address_class_type_flags);
  set_gdbarch_address_class_type_flags_to_name
    (gdbarch, spu_address_class_type_flags_to_name);
  set_gdbarch_address_class_name_to_type_flags
    (gdbarch, spu_address_class_name_to_type_flags);


  /* Inferior function calls.  */
  set_gdbarch_call_dummy_location (gdbarch, ON_STACK);
  set_gdbarch_frame_align (gdbarch, spu_frame_align);
  set_gdbarch_frame_red_zone_size (gdbarch, 2000);
  set_gdbarch_push_dummy_code (gdbarch, spu_push_dummy_code);
  set_gdbarch_push_dummy_call (gdbarch, spu_push_dummy_call);
  set_gdbarch_dummy_id (gdbarch, spu_dummy_id);
  set_gdbarch_return_value (gdbarch, spu_return_value);

  /* Frame handling.  */
  set_gdbarch_inner_than (gdbarch, core_addr_lessthan);
  dwarf2_append_unwinders (gdbarch);
  frame_unwind_append_unwinder (gdbarch, &spu_frame_unwind);
  frame_base_set_default (gdbarch, &spu_frame_base);
  set_gdbarch_unwind_pc (gdbarch, spu_unwind_pc);
  set_gdbarch_unwind_sp (gdbarch, spu_unwind_sp);
  set_gdbarch_virtual_frame_pointer (gdbarch, spu_virtual_frame_pointer);
  set_gdbarch_frame_args_skip (gdbarch, 0);
  set_gdbarch_skip_prologue (gdbarch, spu_skip_prologue);
  set_gdbarch_stack_frame_destroyed_p (gdbarch, spu_stack_frame_destroyed_p);

  /* Cell/B.E. cross-architecture unwinder support.  */
  frame_unwind_prepend_unwinder (gdbarch, &spu2ppu_unwind);

  /* Breakpoints.  */
  set_gdbarch_decr_pc_after_break (gdbarch, 4);
  set_gdbarch_breakpoint_kind_from_pc (gdbarch, spu_breakpoint::kind_from_pc);
  set_gdbarch_sw_breakpoint_from_kind (gdbarch, spu_breakpoint::bp_from_kind);
  set_gdbarch_memory_remove_breakpoint (gdbarch, spu_memory_remove_breakpoint);
  set_gdbarch_software_single_step (gdbarch, spu_software_single_step);
  set_gdbarch_get_longjmp_target (gdbarch, spu_get_longjmp_target);

  /* Overlays.  */
  set_gdbarch_overlay_update (gdbarch, spu_overlay_update);

  return gdbarch;
}

/* Provide a prototype to silence -Wmissing-prototypes.  */
extern initialize_file_ftype _initialize_spu_tdep;

void
_initialize_spu_tdep (void)
{
  register_gdbarch_init (bfd_arch_spu, spu_gdbarch_init);

  /* Add ourselves to objfile event chain.  */
  observer_attach_new_objfile (spu_overlay_new_objfile);
  spu_overlay_data = register_objfile_data ();

  /* Install spu stop-on-load handler.  */
  observer_attach_new_objfile (spu_catch_start);

  /* Add ourselves to normal_stop event chain.  */
  observer_attach_normal_stop (spu_attach_normal_stop);

  /* Add root prefix command for all "set spu"/"show spu" commands.  */
  add_prefix_cmd ("spu", no_class, set_spu_command,
		  _("Various SPU specific commands."),
		  &setspucmdlist, "set spu ", 0, &setlist);
  add_prefix_cmd ("spu", no_class, show_spu_command,
		  _("Various SPU specific commands."),
		  &showspucmdlist, "show spu ", 0, &showlist);

  /* Toggle whether or not to add a temporary breakpoint at the "main"
     function of new SPE contexts.  */
  add_setshow_boolean_cmd ("stop-on-load", class_support,
                          &spu_stop_on_load_p, _("\
Set whether to stop for new SPE threads."),
                           _("\
Show whether to stop for new SPE threads."),
                           _("\
Use \"on\" to give control to the user when a new SPE thread\n\
enters its \"main\" function.\n\
Use \"off\" to disable stopping for new SPE threads."),
                          NULL,
                          show_spu_stop_on_load,
                          &setspucmdlist, &showspucmdlist);

  /* Toggle whether or not to automatically flush the software-managed
     cache whenever SPE execution stops.  */
  add_setshow_boolean_cmd ("auto-flush-cache", class_support,
                          &spu_auto_flush_cache_p, _("\
Set whether to automatically flush the software-managed cache."),
                           _("\
Show whether to automatically flush the software-managed cache."),
                           _("\
Use \"on\" to automatically flush the software-managed cache\n\
whenever SPE execution stops.\n\
Use \"off\" to never automatically flush the software-managed cache."),
                          NULL,
                          show_spu_auto_flush_cache,
                          &setspucmdlist, &showspucmdlist);

  /* Add root prefix command for all "info spu" commands.  */
  add_prefix_cmd ("spu", class_info, info_spu_command,
		  _("Various SPU specific commands."),
		  &infospucmdlist, "info spu ", 0, &infolist);

  /* Add various "info spu" commands.  */
  add_cmd ("event", class_info, info_spu_event_command,
	   _("Display SPU event facility status.\n"),
	   &infospucmdlist);
  add_cmd ("signal", class_info, info_spu_signal_command,
	   _("Display SPU signal notification facility status.\n"),
	   &infospucmdlist);
  add_cmd ("mailbox", class_info, info_spu_mailbox_command,
	   _("Display SPU mailbox facility status.\n"),
	   &infospucmdlist);
  add_cmd ("dma", class_info, info_spu_dma_command,
	   _("Display MFC DMA status.\n"),
	   &infospucmdlist);
  add_cmd ("proxydma", class_info, info_spu_proxydma_command,
	   _("Display MFC Proxy-DMA status.\n"),
	   &infospucmdlist);
}