gdb/remote.c

gdb/remote.c - gdb

Source code

/* Remote target communications for serial-line targets in custom GDB protocol



   Copyright (C) 1988-2015 Free Software Foundation, Inc.



   This file is part of GDB.



   This program is free software; you can redistribute it and/or modify

   it under the terms of the GNU General Public License as published by

   the Free Software Foundation; either version 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/>.  */



/* See the GDB User Guide for details of the GDB remote protocol.  */



#include "defs.h"

#include <ctype.h>

#include <fcntl.h>

#include "inferior.h"

#include "infrun.h"

#include "bfd.h"

#include "symfile.h"

#include "target.h"

/*#include "terminal.h" */

#include "gdbcmd.h"

#include "objfiles.h"

#include "gdb-stabs.h"

#include "gdbthread.h"

#include "remote.h"

#include "remote-notif.h"

#include "regcache.h"

#include "value.h"

#include "observer.h"

#include "solib.h"

#include "cli/cli-decode.h"

#include "cli/cli-setshow.h"

#include "target-descriptions.h"

#include "gdb_bfd.h"

#include "filestuff.h"

#include "rsp-low.h"



#include <sys/time.h>



#include "event-loop.h"

#include "event-top.h"

#include "inf-loop.h"



#include <signal.h>

#include "serial.h"



#include "gdbcore.h" /* for exec_bfd */



#include "remote-fileio.h"

#include "gdb/fileio.h"

#include <sys/stat.h>

#include "xml-support.h"



#include "memory-map.h"



#include "tracepoint.h"

#include "ax.h"

#include "ax-gdb.h"

#include "agent.h"

#include "btrace.h"



/* Temp hacks for tracepoint encoding migration.  */

‌static char *target_buf;

‌static long target_buf_size;



/* The size to align memory write packets, when practical.  The protocol

   does not guarantee any alignment, and gdb will generate short

   writes and unaligned writes, but even as a best-effort attempt this

   can improve bulk transfers.  For instance, if a write is misaligned

   relative to the target's data bus, the stub may need to make an extra

   round trip fetching data from the target.  This doesn't make a

   huge difference, but it's easy to do, so we try to be helpful.



   The alignment chosen is arbitrary; usually data bus width is

   important here, not the possibly larger cache line size.  */

enum { REMOTE_ALIGN_WRITES = 16 };



/* Prototypes for local functions.  */

static void async_cleanup_sigint_signal_handler (void *dummy);

static int getpkt_sane (char **buf, long *sizeof_buf, int forever);

static int getpkt_or_notif_sane (char **buf, long *sizeof_buf,

                                 int forever, int *is_notif);



static void async_handle_remote_sigint (int);

static void async_handle_remote_sigint_twice (int);



static void remote_files_info (struct target_ops *ignore);



static void remote_prepare_to_store (struct target_ops *self,

                                     struct regcache *regcache);



static void remote_open_1 (const char *, int, struct target_ops *,

                           int extended_p);



static void remote_close (struct target_ops *self);



static void remote_mourn (struct target_ops *ops);



static void extended_remote_restart (void);



static void extended_remote_mourn (struct target_ops *);



static void remote_mourn_1 (struct target_ops *);



static void remote_send (char **buf, long *sizeof_buf_p);



static int readchar (int timeout);



static void remote_serial_write (const char *str, int len);



static void remote_kill (struct target_ops *ops);



static int remote_can_async_p (struct target_ops *);



static int remote_is_async_p (struct target_ops *);



static void remote_async (struct target_ops *ops,

                          void (*callback) (enum inferior_event_type event_type,

                                            void *context),

                          void *context);



static void sync_remote_interrupt_twice (int signo);



static void interrupt_query (void);



static void set_general_thread (struct ptid ptid);

static void set_continue_thread (struct ptid ptid);



static void get_offsets (void);



static void skip_frame (void);



static long read_frame (char **buf_p, long *sizeof_buf);



static int hexnumlen (ULONGEST num);



static void init_remote_ops (void);



static void init_extended_remote_ops (void);



static void remote_stop (struct target_ops *self, ptid_t);



static int stubhex (int ch);



static int hexnumstr (char *, ULONGEST);



static int hexnumnstr (char *, ULONGEST, int);



static CORE_ADDR remote_address_masked (CORE_ADDR);



static void print_packet (const char *);



static void compare_sections_command (char *, int);



static void packet_command (char *, int);



static int stub_unpack_int (char *buff, int fieldlength);



static ptid_t remote_current_thread (ptid_t oldptid);



static int putpkt_binary (const char *buf, int cnt);



static void check_binary_download (CORE_ADDR addr);



struct packet_config;



static void show_packet_config_cmd (struct packet_config *config);



static void show_remote_protocol_packet_cmd (struct ui_file *file,

                                             int from_tty,

                                             struct cmd_list_element *c,

                                             const char *value);



static char *write_ptid (char *buf, const char *endbuf, ptid_t ptid);

static ptid_t read_ptid (char *buf, char **obuf);



static void remote_set_permissions (struct target_ops *self);



struct remote_state;

static int remote_get_trace_status (struct target_ops *self,

                                    struct trace_status *ts);



static int remote_upload_tracepoints (struct target_ops *self,

                                      struct uploaded_tp **utpp);



static int remote_upload_trace_state_variables (struct target_ops *self,

                                                struct uploaded_tsv **utsvp);



static void remote_query_supported (void);



static void remote_check_symbols (void);



void _initialize_remote (void);



struct stop_reply;

static void stop_reply_xfree (struct stop_reply *);

static void remote_parse_stop_reply (char *, struct stop_reply *);

static void push_stop_reply (struct stop_reply *);

static void discard_pending_stop_replies_in_queue (struct remote_state *);

static int peek_stop_reply (ptid_t ptid);



static void remote_async_inferior_event_handler (gdb_client_data);



static void remote_terminal_ours (struct target_ops *self);



static int remote_read_description_p (struct target_ops *target);



static void remote_console_output (char *msg);



static int remote_supports_cond_breakpoints (struct target_ops *self);



static int remote_can_run_breakpoint_commands (struct target_ops *self);



/* For "remote".  */



‌static struct cmd_list_element *remote_cmdlist;



/* For "set remote" and "show remote".  */



‌static struct cmd_list_element *remote_set_cmdlist;

‌static struct cmd_list_element *remote_show_cmdlist;



/* Stub vCont actions support.



   Each field is a boolean flag indicating whether the stub reports

   support for the corresponding action.  */



‌struct vCont_action_support

{

  /* vCont;t */

  int t;



  /* vCont;r */

  int r;

};



/* Controls whether GDB is willing to use range stepping.  */



‌static int use_range_stepping = 1;



‌#define OPAQUETHREADBYTES 8



/* a 64 bit opaque identifier */

‌typedef unsigned char threadref[OPAQUETHREADBYTES];



/* About this many threadisds fit in a packet.  */



‌#define MAXTHREADLISTRESULTS 32



/* Description of the remote protocol state for the currently

   connected target.  This is per-target state, and independent of the

   selected architecture.  */



‌struct remote_state

{

  /* A buffer to use for incoming packets, and its current size.  The

     buffer is grown dynamically for larger incoming packets.

     Outgoing packets may also be constructed in this buffer.

     BUF_SIZE is always at least REMOTE_PACKET_SIZE;

     REMOTE_PACKET_SIZE should be used to limit the length of outgoing

     packets.  */

  char *buf;

  long buf_size;



  /* True if we're going through initial connection setup (finding out

     about the remote side's threads, relocating symbols, etc.).  */

  int starting_up;



  /* If we negotiated packet size explicitly (and thus can bypass

     heuristics for the largest packet size that will not overflow

     a buffer in the stub), this will be set to that packet size.

     Otherwise zero, meaning to use the guessed size.  */

  long explicit_packet_size;



  /* remote_wait is normally called when the target is running and

     waits for a stop reply packet.  But sometimes we need to call it

     when the target is already stopped.  We can send a "?" packet

     and have remote_wait read the response.  Or, if we already have

     the response, we can stash it in BUF and tell remote_wait to

     skip calling getpkt.  This flag is set when BUF contains a

     stop reply packet and the target is not waiting.  */

  int cached_wait_status;



  /* True, if in no ack mode.  That is, neither GDB nor the stub will

     expect acks from each other.  The connection is assumed to be

     reliable.  */

  int noack_mode;



  /* True if we're connected in extended remote mode.  */

  int extended;



  /* True if we resumed the target and we're waiting for the target to

     stop.  In the mean time, we can't start another command/query.

     The remote server wouldn't be ready to process it, so we'd

     timeout waiting for a reply that would never come and eventually

     we'd close the connection.  This can happen in asynchronous mode

     because we allow GDB commands while the target is running.  */

  int waiting_for_stop_reply;



  /* The status of the stub support for the various vCont actions.  */

  struct vCont_action_support supports_vCont;



  /* Nonzero if the user has pressed Ctrl-C, but the target hasn't

     responded to that.  */

  int ctrlc_pending_p;



  /* Descriptor for I/O to remote machine.  Initialize it to NULL so that

     remote_open knows that we don't have a file open when the program

     starts.  */

  struct serial *remote_desc;



  /* These are the threads which we last sent to the remote system.  The

     TID member will be -1 for all or -2 for not sent yet.  */

  ptid_t general_thread;

  ptid_t continue_thread;



  /* This is the traceframe which we last selected on the remote system.

     It will be -1 if no traceframe is selected.  */

  int remote_traceframe_number;



  char *last_pass_packet;



  /* The last QProgramSignals packet sent to the target.  We bypass

     sending a new program signals list down to the target if the new

     packet is exactly the same as the last we sent.  IOW, we only let

     the target know about program signals list changes.  */

  char *last_program_signals_packet;



  enum gdb_signal last_sent_signal;



  int last_sent_step;



  char *finished_object;

  char *finished_annex;

  ULONGEST finished_offset;



  /* Should we try the 'ThreadInfo' query packet?



     This variable (NOT available to the user: auto-detect only!)

     determines whether GDB will use the new, simpler "ThreadInfo"

     query or the older, more complex syntax for thread queries.

     This is an auto-detect variable (set to true at each connect,

     and set to false when the target fails to recognize it).  */

  int use_threadinfo_query;

  int use_threadextra_query;



  void (*async_client_callback) (enum inferior_event_type event_type,

                                 void *context);

  void *async_client_context;



  /* This is set to the data address of the access causing the target

     to stop for a watchpoint.  */

  CORE_ADDR remote_watch_data_address;



  /* This is non-zero if target stopped for a watchpoint.  */

  int remote_stopped_by_watchpoint_p;



  threadref echo_nextthread;

  threadref nextthread;

  threadref resultthreadlist[MAXTHREADLISTRESULTS];



  /* The state of remote notification.  */

  struct remote_notif_state *notif_state;

};



/* Private data that we'll store in (struct thread_info)->private.  */

‌struct private_thread_info

{

  char *extra;

  int core;

};



static void

‌free_private_thread_info (struct private_thread_info *info)

{

  xfree (info->extra);

  xfree (info);

}



/* This data could be associated with a target, but we do not always

   have access to the current target when we need it, so for now it is

   static.  This will be fine for as long as only one target is in use

   at a time.  */

‌static struct remote_state *remote_state;



static struct remote_state *

‌get_remote_state_raw (void)

{

  return remote_state;

}



/* Allocate a new struct remote_state with xmalloc, initialize it, and

   return it.  */



static struct remote_state *

‌new_remote_state (void)

{

  struct remote_state *result = XCNEW (struct remote_state);



  /* The default buffer size is unimportant; it will be expanded

     whenever a larger buffer is needed. */

  result->buf_size = 400;

  result->buf = xmalloc (result->buf_size);

  result->remote_traceframe_number = -1;

  result->last_sent_signal = GDB_SIGNAL_0;



  return result;

}



/* Description of the remote protocol for a given architecture.  */



‌struct packet_reg

{

  long offset; /* Offset into G packet.  */

  long regnum; /* GDB's internal register number.  */

  LONGEST pnum; /* Remote protocol register number.  */

  int in_g_packet; /* Always part of G packet.  */

  /* long size in bytes;  == register_size (target_gdbarch (), regnum);

     at present.  */

  /* char *name; == gdbarch_register_name (target_gdbarch (), regnum);

     at present.  */

};



‌struct remote_arch_state

{

  /* Description of the remote protocol registers.  */

  long sizeof_g_packet;



  /* Description of the remote protocol registers indexed by REGNUM

     (making an array gdbarch_num_regs in size).  */

  struct packet_reg *regs;



  /* This is the size (in chars) of the first response to the ``g''

     packet.  It is used as a heuristic when determining the maximum

     size of memory-read and memory-write packets.  A target will

     typically only reserve a buffer large enough to hold the ``g''

     packet.  The size does not include packet overhead (headers and

     trailers).  */

  long actual_register_packet_size;



  /* This is the maximum size (in chars) of a non read/write packet.

     It is also used as a cap on the size of read/write packets.  */

  long remote_packet_size;

};



/* Utility: generate error from an incoming stub packet.  */

static void

‌trace_error (char *buf)

{

  if (*buf++ != 'E')

    return;                        /* not an error msg */

  switch (*buf)

    {

    case '1':                        /* malformed packet error */

      if (*++buf == '0')        /*   general case: */

        error (_("remote.c: error in outgoing packet."));

      else

        error (_("remote.c: error in outgoing packet at field #%ld."),

               strtol (buf, NULL, 16));

    default:

      error (_("Target returns error code '%s'."), buf);

    }

}



/* Utility: wait for reply from stub, while accepting "O" packets.  */

static char *

‌remote_get_noisy_reply (char **buf_p,

                        long *sizeof_buf)

{

  do                                /* Loop on reply from remote stub.  */

    {

      char *buf;



      QUIT;                        /* Allow user to bail out with ^C.  */

      getpkt (buf_p, sizeof_buf, 0);

      buf = *buf_p;

      if (buf[0] == 'E')

        trace_error (buf);

      else if (strncmp (buf, "qRelocInsn:", strlen ("qRelocInsn:")) == 0)

        {

          ULONGEST ul;

          CORE_ADDR from, to, org_to;

          char *p, *pp;

          int adjusted_size = 0;

          volatile struct gdb_exception ex;



          p = buf + strlen ("qRelocInsn:");

          pp = unpack_varlen_hex (p, &ul);

          if (*pp != ';')

            error (_("invalid qRelocInsn packet: %s"), buf);

          from = ul;



          p = pp + 1;

          unpack_varlen_hex (p, &ul);

          to = ul;



          org_to = to;



          TRY_CATCH (ex, RETURN_MASK_ALL)

            {

              gdbarch_relocate_instruction (target_gdbarch (), &to, from);

            }

          if (ex.reason >= 0)

            {

              adjusted_size = to - org_to;



              xsnprintf (buf, *sizeof_buf, "qRelocInsn:%x", adjusted_size);

              putpkt (buf);

            }

          else if (ex.reason < 0 && ex.error == MEMORY_ERROR)

            {

              /* Propagate memory errors silently back to the target.

                 The stub may have limited the range of addresses we

                 can write to, for example.  */

              putpkt ("E01");

            }

          else

            {

              /* Something unexpectedly bad happened.  Be verbose so

                 we can tell what, and propagate the error back to the

                 stub, so it doesn't get stuck waiting for a

                 response.  */

              exception_fprintf (gdb_stderr, ex,

                                 _("warning: relocating instruction: "));

              putpkt ("E01");

            }

        }

      else if (buf[0] == 'O' && buf[1] != 'K')

        remote_console_output (buf + 1);        /* 'O' message from stub */

      else

        return buf;                /* Here's the actual reply.  */

    }

  while (1);

}



/* Handle for retreving the remote protocol data from gdbarch.  */

‌static struct gdbarch_data *remote_gdbarch_data_handle;



static struct remote_arch_state *

‌get_remote_arch_state (void)

{

  return gdbarch_data (target_gdbarch (), remote_gdbarch_data_handle);

}



/* Fetch the global remote target state.  */



static struct remote_state *

‌get_remote_state (void)

{

  /* Make sure that the remote architecture state has been

     initialized, because doing so might reallocate rs->buf.  Any

     function which calls getpkt also needs to be mindful of changes

     to rs->buf, but this call limits the number of places which run

     into trouble.  */

  get_remote_arch_state ();



  return get_remote_state_raw ();

}



static int

‌compare_pnums (const void *lhs_, const void *rhs_)

{

  const struct packet_reg * const *lhs = lhs_;

  const struct packet_reg * const *rhs = rhs_;



  if ((*lhs)->pnum < (*rhs)->pnum)

    return -1;

  else if ((*lhs)->pnum == (*rhs)->pnum)

    return 0;

  else

    return 1;

}



static int

‌map_regcache_remote_table (struct gdbarch *gdbarch, struct packet_reg *regs)

{

  int regnum, num_remote_regs, offset;

  struct packet_reg **remote_regs;



  for (regnum = 0; regnum < gdbarch_num_regs (gdbarch); regnum++)

    {

      struct packet_reg *r = &regs[regnum];



      if (register_size (gdbarch, regnum) == 0)

        /* Do not try to fetch zero-sized (placeholder) registers.  */

        r->pnum = -1;

      else

        r->pnum = gdbarch_remote_register_number (gdbarch, regnum);



      r->regnum = regnum;

    }



  /* Define the g/G packet format as the contents of each register

     with a remote protocol number, in order of ascending protocol

     number.  */



  remote_regs = alloca (gdbarch_num_regs (gdbarch)

                        * sizeof (struct packet_reg *));

  for (num_remote_regs = 0, regnum = 0;

       regnum < gdbarch_num_regs (gdbarch);

       regnum++)

    if (regs[regnum].pnum != -1)

      remote_regs[num_remote_regs++] = &regs[regnum];



  qsort (remote_regs, num_remote_regs, sizeof (struct packet_reg *),

         compare_pnums);



  for (regnum = 0, offset = 0; regnum < num_remote_regs; regnum++)

    {

      remote_regs[regnum]->in_g_packet = 1;

      remote_regs[regnum]->offset = offset;

      offset += register_size (gdbarch, remote_regs[regnum]->regnum);

    }



  return offset;

}



/* Given the architecture described by GDBARCH, return the remote

   protocol register's number and the register's offset in the g/G

   packets of GDB register REGNUM, in PNUM and POFFSET respectively.

   If the target does not have a mapping for REGNUM, return false,

   otherwise, return true.  */



int

‌remote_register_number_and_offset (struct gdbarch *gdbarch, int regnum,

                                   int *pnum, int *poffset)

{

  int sizeof_g_packet;

  struct packet_reg *regs;

  struct cleanup *old_chain;



  gdb_assert (regnum < gdbarch_num_regs (gdbarch));



  regs = xcalloc (gdbarch_num_regs (gdbarch), sizeof (struct packet_reg));

  old_chain = make_cleanup (xfree, regs);



  sizeof_g_packet = map_regcache_remote_table (gdbarch, regs);



  *pnum = regs[regnum].pnum;

  *poffset = regs[regnum].offset;



  do_cleanups (old_chain);



  return *pnum != -1;

}



static void *

‌init_remote_state (struct gdbarch *gdbarch)

{

  struct remote_state *rs = get_remote_state_raw ();

  struct remote_arch_state *rsa;



  rsa = GDBARCH_OBSTACK_ZALLOC (gdbarch, struct remote_arch_state);



  /* Use the architecture to build a regnum<->pnum table, which will be

     1:1 unless a feature set specifies otherwise.  */

  rsa->regs = GDBARCH_OBSTACK_CALLOC (gdbarch,

                                      gdbarch_num_regs (gdbarch),

                                      struct packet_reg);



  /* Record the maximum possible size of the g packet - it may turn out

     to be smaller.  */

  rsa->sizeof_g_packet = map_regcache_remote_table (gdbarch, rsa->regs);



  /* Default maximum number of characters in a packet body.  Many

     remote stubs have a hardwired buffer size of 400 bytes

     (c.f. BUFMAX in m68k-stub.c and i386-stub.c).  BUFMAX-1 is used

     as the maximum packet-size to ensure that the packet and an extra

     NUL character can always fit in the buffer.  This stops GDB

     trashing stubs that try to squeeze an extra NUL into what is

     already a full buffer (As of 1999-12-04 that was most stubs).  */

  rsa->remote_packet_size = 400 - 1;



  /* This one is filled in when a ``g'' packet is received.  */

  rsa->actual_register_packet_size = 0;



  /* Should rsa->sizeof_g_packet needs more space than the

     default, adjust the size accordingly.  Remember that each byte is

     encoded as two characters.  32 is the overhead for the packet

     header / footer.  NOTE: cagney/1999-10-26: I suspect that 8

     (``$NN:G...#NN'') is a better guess, the below has been padded a

     little.  */

  if (rsa->sizeof_g_packet > ((rsa->remote_packet_size - 32) / 2))

    rsa->remote_packet_size = (rsa->sizeof_g_packet * 2 + 32);



  /* Make sure that the packet buffer is plenty big enough for

     this architecture.  */

  if (rs->buf_size < rsa->remote_packet_size)

    {

      rs->buf_size = 2 * rsa->remote_packet_size;

      rs->buf = xrealloc (rs->buf, rs->buf_size);

    }



  return rsa;

}



/* Return the current allowed size of a remote packet.  This is

   inferred from the current architecture, and should be used to

   limit the length of outgoing packets.  */

static long

‌get_remote_packet_size (void)

{

  struct remote_state *rs = get_remote_state ();

  struct remote_arch_state *rsa = get_remote_arch_state ();



  if (rs->explicit_packet_size)

    return rs->explicit_packet_size;



  return rsa->remote_packet_size;

}



static struct packet_reg *

‌packet_reg_from_regnum (struct remote_arch_state *rsa, long regnum)

{

  if (regnum < 0 && regnum >= gdbarch_num_regs (target_gdbarch ()))

    return NULL;

  else

    {

      struct packet_reg *r = &rsa->regs[regnum];



      gdb_assert (r->regnum == regnum);

      return r;

    }

}



static struct packet_reg *

‌packet_reg_from_pnum (struct remote_arch_state *rsa, LONGEST pnum)

{

  int i;



  for (i = 0; i < gdbarch_num_regs (target_gdbarch ()); i++)

    {

      struct packet_reg *r = &rsa->regs[i];



      if (r->pnum == pnum)

        return r;

    }

  return NULL;

}



‌static struct target_ops remote_ops;



‌static struct target_ops extended_remote_ops;



/* FIXME: cagney/1999-09-23: Even though getpkt was called with

   ``forever'' still use the normal timeout mechanism.  This is

   currently used by the ASYNC code to guarentee that target reads

   during the initial connect always time-out.  Once getpkt has been

   modified to return a timeout indication and, in turn

   remote_wait()/wait_for_inferior() have gained a timeout parameter

   this can go away.  */

‌static int wait_forever_enabled_p = 1;



/* Allow the user to specify what sequence to send to the remote

   when he requests a program interruption: Although ^C is usually

   what remote systems expect (this is the default, here), it is

   sometimes preferable to send a break.  On other systems such

   as the Linux kernel, a break followed by g, which is Magic SysRq g

   is required in order to interrupt the execution.  */

‌const char interrupt_sequence_control_c[] = "Ctrl-C";

‌const char interrupt_sequence_break[] = "BREAK";

‌const char interrupt_sequence_break_g[] = "BREAK-g";

‌static const char *const interrupt_sequence_modes[] =

  {

    interrupt_sequence_control_c,

    interrupt_sequence_break,

    interrupt_sequence_break_g,

    NULL

  };

‌static const char *interrupt_sequence_mode = interrupt_sequence_control_c;



static void

‌show_interrupt_sequence (struct ui_file *file, int from_tty,

                         struct cmd_list_element *c,

                         const char *value)

{

  if (interrupt_sequence_mode == interrupt_sequence_control_c)

    fprintf_filtered (file,

                      _("Send the ASCII ETX character (Ctrl-c) "

                        "to the remote target to interrupt the "

                        "execution of the program.\n"));

  else if (interrupt_sequence_mode == interrupt_sequence_break)

    fprintf_filtered (file,

                      _("send a break signal to the remote target "

                        "to interrupt the execution of the program.\n"));

  else if (interrupt_sequence_mode == interrupt_sequence_break_g)

    fprintf_filtered (file,

                      _("Send a break signal and 'g' a.k.a. Magic SysRq g to "

                        "the remote target to interrupt the execution "

                        "of Linux kernel.\n"));

  else

    internal_error (__FILE__, __LINE__,

                    _("Invalid value for interrupt_sequence_mode: %s."),

                    interrupt_sequence_mode);

}



/* This boolean variable specifies whether interrupt_sequence is sent

   to the remote target when gdb connects to it.

   This is mostly needed when you debug the Linux kernel: The Linux kernel

   expects BREAK g which is Magic SysRq g for connecting gdb.  */

‌static int interrupt_on_connect = 0;



/* This variable is used to implement the "set/show remotebreak" commands.

   Since these commands are now deprecated in favor of "set/show remote

   interrupt-sequence", it no longer has any effect on the code.  */

‌static int remote_break;



static void

‌set_remotebreak (char *args, int from_tty, struct cmd_list_element *c)

{

  if (remote_break)

    interrupt_sequence_mode = interrupt_sequence_break;

  else

    interrupt_sequence_mode = interrupt_sequence_control_c;

}



static void

‌show_remotebreak (struct ui_file *file, int from_tty,

                  struct cmd_list_element *c,

                  const char *value)

{

}



/* This variable sets the number of bits in an address that are to be

   sent in a memory ("M" or "m") packet.  Normally, after stripping

   leading zeros, the entire address would be sent.  This variable

   restricts the address to REMOTE_ADDRESS_SIZE bits.  HISTORY: The

   initial implementation of remote.c restricted the address sent in

   memory packets to ``host::sizeof long'' bytes - (typically 32

   bits).  Consequently, for 64 bit targets, the upper 32 bits of an

   address was never sent.  Since fixing this bug may cause a break in

   some remote targets this variable is principly provided to

   facilitate backward compatibility.  */



‌static unsigned int remote_address_size;



/* Temporary to track who currently owns the terminal.  See

   remote_terminal_* for more details.  */



‌static int remote_async_terminal_ours_p;



/* The executable file to use for "run" on the remote side.  */



‌static char *remote_exec_file = "";





/* User configurable variables for the number of characters in a

   memory read/write packet.  MIN (rsa->remote_packet_size,

   rsa->sizeof_g_packet) is the default.  Some targets need smaller

   values (fifo overruns, et.al.) and some users need larger values

   (speed up transfers).  The variables ``preferred_*'' (the user

   request), ``current_*'' (what was actually set) and ``forced_*''

   (Positive - a soft limit, negative - a hard limit).  */



‌struct memory_packet_config

{

  char *name;

  long size;

  int fixed_p;

};



/* Compute the current size of a read/write packet.  Since this makes

   use of ``actual_register_packet_size'' the computation is dynamic.  */



static long

‌get_memory_packet_size (struct memory_packet_config *config)

{

  struct remote_state *rs = get_remote_state ();

  struct remote_arch_state *rsa = get_remote_arch_state ();



  /* NOTE: The somewhat arbitrary 16k comes from the knowledge (folk

     law?) that some hosts don't cope very well with large alloca()

     calls.  Eventually the alloca() code will be replaced by calls to

     xmalloc() and make_cleanups() allowing this restriction to either

     be lifted or removed.  */

#ifndef MAX_REMOTE_PACKET_SIZE

‌#define MAX_REMOTE_PACKET_SIZE 16384

#endif

  /* NOTE: 20 ensures we can write at least one byte.  */

#ifndef MIN_REMOTE_PACKET_SIZE

‌#define MIN_REMOTE_PACKET_SIZE 20

#endif

  long what_they_get;

  if (config->fixed_p)

    {

      if (config->size <= 0)

        what_they_get = MAX_REMOTE_PACKET_SIZE;

      else

        what_they_get = config->size;

    }

  else

    {

      what_they_get = get_remote_packet_size ();

      /* Limit the packet to the size specified by the user.  */

      if (config->size > 0

          && what_they_get > config->size)

        what_they_get = config->size;



      /* Limit it to the size of the targets ``g'' response unless we have

         permission from the stub to use a larger packet size.  */

      if (rs->explicit_packet_size == 0

          && rsa->actual_register_packet_size > 0

          && what_they_get > rsa->actual_register_packet_size)

        what_they_get = rsa->actual_register_packet_size;

    }

  if (what_they_get > MAX_REMOTE_PACKET_SIZE)

    what_they_get = MAX_REMOTE_PACKET_SIZE;

  if (what_they_get < MIN_REMOTE_PACKET_SIZE)

    what_they_get = MIN_REMOTE_PACKET_SIZE;



  /* Make sure there is room in the global buffer for this packet

     (including its trailing NUL byte).  */

  if (rs->buf_size < what_they_get + 1)

    {

      rs->buf_size = 2 * what_they_get;

      rs->buf = xrealloc (rs->buf, 2 * what_they_get);

    }



  return what_they_get;

}



/* Update the size of a read/write packet.  If they user wants

   something really big then do a sanity check.  */



static void

‌set_memory_packet_size (char *args, struct memory_packet_config *config)

{

  int fixed_p = config->fixed_p;

  long size = config->size;



  if (args == NULL)

    error (_("Argument required (integer, `fixed' or `limited')."));

  else if (strcmp (args, "hard") == 0

      || strcmp (args, "fixed") == 0)

    fixed_p = 1;

  else if (strcmp (args, "soft") == 0

           || strcmp (args, "limit") == 0)

    fixed_p = 0;

  else

    {

      char *end;



      size = strtoul (args, &end, 0);

      if (args == end)

        error (_("Invalid %s (bad syntax)."), config->name);

#if 0

      /* Instead of explicitly capping the size of a packet to

         MAX_REMOTE_PACKET_SIZE or dissallowing it, the user is

         instead allowed to set the size to something arbitrarily

         large.  */

      if (size > MAX_REMOTE_PACKET_SIZE)

        error (_("Invalid %s (too large)."), config->name);

#endif

    }

  /* Extra checks?  */

  if (fixed_p && !config->fixed_p)

    {

      if (! query (_("The target may not be able to correctly handle a %s\n"

                   "of %ld bytes. Change the packet size? "),

                   config->name, size))

        error (_("Packet size not changed."));

    }

  /* Update the config.  */

  config->fixed_p = fixed_p;

  config->size = size;

}



static void

‌show_memory_packet_size (struct memory_packet_config *config)

{

  printf_filtered (_("The %s is %ld. "), config->name, config->size);

  if (config->fixed_p)

    printf_filtered (_("Packets are fixed at %ld bytes.\n"),

                     get_memory_packet_size (config));

  else

    printf_filtered (_("Packets are limited to %ld bytes.\n"),

                     get_memory_packet_size (config));

}



‌static struct memory_packet_config memory_write_packet_config =

{

  "memory-write-packet-size",

};



static void

‌set_memory_write_packet_size (char *args, int from_tty)

{

  set_memory_packet_size (args, &memory_write_packet_config);

}



static void

‌show_memory_write_packet_size (char *args, int from_tty)

{

  show_memory_packet_size (&memory_write_packet_config);

}



static long

‌get_memory_write_packet_size (void)

{

  return get_memory_packet_size (&memory_write_packet_config);

}



‌static struct memory_packet_config memory_read_packet_config =

{

  "memory-read-packet-size",

};



static void

‌set_memory_read_packet_size (char *args, int from_tty)

{

  set_memory_packet_size (args, &memory_read_packet_config);

}



static void

‌show_memory_read_packet_size (char *args, int from_tty)

{

  show_memory_packet_size (&memory_read_packet_config);

}



static long

‌get_memory_read_packet_size (void)

{

  long size = get_memory_packet_size (&memory_read_packet_config);



  /* FIXME: cagney/1999-11-07: Functions like getpkt() need to get an

     extra buffer size argument before the memory read size can be

     increased beyond this.  */

  if (size > get_remote_packet_size ())

    size = get_remote_packet_size ();

  return size;

}





/* Generic configuration support for packets the stub optionally

   supports.  Allows the user to specify the use of the packet as well

   as allowing GDB to auto-detect support in the remote stub.  */



‌enum packet_support

  {

    PACKET_SUPPORT_UNKNOWN = 0,

    PACKET_ENABLE,

    PACKET_DISABLE

  };



‌struct packet_config

  {

    const char *name;

    const char *title;



    /* If auto, GDB auto-detects support for this packet or feature,

       either through qSupported, or by trying the packet and looking

       at the response.  If true, GDB assumes the target supports this

       packet.  If false, the packet is disabled.  Configs that don't

       have an associated command always have this set to auto.  */

    enum auto_boolean detect;



    /* Does the target support this packet?  */

    enum packet_support support;

  };



/* Analyze a packet's return value and update the packet config

   accordingly.  */



‌enum packet_result

{

  PACKET_ERROR,

  PACKET_OK,

  PACKET_UNKNOWN

};



static enum packet_support packet_config_support (struct packet_config *config);

static enum packet_support packet_support (int packet);



static void

‌show_packet_config_cmd (struct packet_config *config)

{

  char *support = "internal-error";



  switch (packet_config_support (config))

    {

    case PACKET_ENABLE:

      support = "enabled";

      break;

    case PACKET_DISABLE:

      support = "disabled";

      break;

    case PACKET_SUPPORT_UNKNOWN:

      support = "unknown";

      break;

    }

  switch (config->detect)

    {

    case AUTO_BOOLEAN_AUTO:

      printf_filtered (_("Support for the `%s' packet "

                         "is auto-detected, currently %s.\n"),

                       config->name, support);

      break;

    case AUTO_BOOLEAN_TRUE:

    case AUTO_BOOLEAN_FALSE:

      printf_filtered (_("Support for the `%s' packet is currently %s.\n"),

                       config->name, support);

      break;

    }

}



static void

‌add_packet_config_cmd (struct packet_config *config, const char *name,

                       const char *title, int legacy)

{

  char *set_doc;

  char *show_doc;

  char *cmd_name;



  config->name = name;

  config->title = title;

  set_doc = xstrprintf ("Set use of remote protocol `%s' (%s) packet",

                        name, title);

  show_doc = xstrprintf ("Show current use of remote "

                         "protocol `%s' (%s) packet",

                         name, title);

  /* set/show TITLE-packet {auto,on,off} */

  cmd_name = xstrprintf ("%s-packet", title);

  add_setshow_auto_boolean_cmd (cmd_name, class_obscure,

                                &config->detect, set_doc,

                                show_doc, NULL, /* help_doc */

                                NULL,

                                show_remote_protocol_packet_cmd,

                                &remote_set_cmdlist, &remote_show_cmdlist);

  /* The command code copies the documentation strings.  */

  xfree (set_doc);

  xfree (show_doc);

  /* set/show remote NAME-packet {auto,on,off} -- legacy.  */

  if (legacy)

    {

      char *legacy_name;



      legacy_name = xstrprintf ("%s-packet", name);

      add_alias_cmd (legacy_name, cmd_name, class_obscure, 0,

                     &remote_set_cmdlist);

      add_alias_cmd (legacy_name, cmd_name, class_obscure, 0,

                     &remote_show_cmdlist);

    }

}



static enum packet_result

‌packet_check_result (const char *buf)

{

  if (buf[0] != '\0')

    {

      /* The stub recognized the packet request.  Check that the

         operation succeeded.  */

      if (buf[0] == 'E'

          && isxdigit (buf[1]) && isxdigit (buf[2])

          && buf[3] == '\0')

        /* "Enn"  - definitly an error.  */

        return PACKET_ERROR;



      /* Always treat "E." as an error.  This will be used for

         more verbose error messages, such as E.memtypes.  */

      if (buf[0] == 'E' && buf[1] == '.')

        return PACKET_ERROR;



      /* The packet may or may not be OK.  Just assume it is.  */

      return PACKET_OK;

    }

  else

    /* The stub does not support the packet.  */

    return PACKET_UNKNOWN;

}



static enum packet_result

‌packet_ok (const char *buf, struct packet_config *config)

{

  enum packet_result result;



  if (config->detect != AUTO_BOOLEAN_TRUE

      && config->support == PACKET_DISABLE)

    internal_error (__FILE__, __LINE__,

                    _("packet_ok: attempt to use a disabled packet"));



  result = packet_check_result (buf);

  switch (result)

    {

    case PACKET_OK:

    case PACKET_ERROR:

      /* The stub recognized the packet request.  */

      if (config->support == PACKET_SUPPORT_UNKNOWN)

        {

          if (remote_debug)

            fprintf_unfiltered (gdb_stdlog,

                                "Packet %s (%s) is supported\n",

                                config->name, config->title);

          config->support = PACKET_ENABLE;

        }

      break;

    case PACKET_UNKNOWN:

      /* The stub does not support the packet.  */

      if (config->detect == AUTO_BOOLEAN_AUTO

          && config->support == PACKET_ENABLE)

        {

          /* If the stub previously indicated that the packet was

             supported then there is a protocol error.  */

          error (_("Protocol error: %s (%s) conflicting enabled responses."),

                 config->name, config->title);

        }

      else if (config->detect == AUTO_BOOLEAN_TRUE)

        {

          /* The user set it wrong.  */

          error (_("Enabled packet %s (%s) not recognized by stub"),

                 config->name, config->title);

        }



      if (remote_debug)

        fprintf_unfiltered (gdb_stdlog,

                            "Packet %s (%s) is NOT supported\n",

                            config->name, config->title);

      config->support = PACKET_DISABLE;

      break;

    }



  return result;

}



enum {

  PACKET_vCont = 0,

  PACKET_X,

  PACKET_qSymbol,

  PACKET_P,

  PACKET_p,

  PACKET_Z0,

  PACKET_Z1,

  PACKET_Z2,

  PACKET_Z3,

  PACKET_Z4,

  PACKET_vFile_open,

  PACKET_vFile_pread,

  PACKET_vFile_pwrite,

  PACKET_vFile_close,

  PACKET_vFile_unlink,

  PACKET_vFile_readlink,

  PACKET_qXfer_auxv,

  PACKET_qXfer_features,

  PACKET_qXfer_libraries,

  PACKET_qXfer_libraries_svr4,

  PACKET_qXfer_memory_map,

  PACKET_qXfer_spu_read,

  PACKET_qXfer_spu_write,

  PACKET_qXfer_osdata,

  PACKET_qXfer_threads,

  PACKET_qXfer_statictrace_read,

  PACKET_qXfer_traceframe_info,

  PACKET_qXfer_uib,

  PACKET_qGetTIBAddr,

  PACKET_qGetTLSAddr,

  PACKET_qSupported,

  PACKET_qTStatus,

  PACKET_QPassSignals,

  PACKET_QProgramSignals,

  PACKET_qCRC,

  PACKET_qSearch_memory,

  PACKET_vAttach,

  PACKET_vRun,

  PACKET_QStartNoAckMode,

  PACKET_vKill,

  PACKET_qXfer_siginfo_read,

  PACKET_qXfer_siginfo_write,

  PACKET_qAttached,



  /* Support for conditional tracepoints.  */

  PACKET_ConditionalTracepoints,



  /* Support for target-side breakpoint conditions.  */

  PACKET_ConditionalBreakpoints,



  /* Support for target-side breakpoint commands.  */

  PACKET_BreakpointCommands,



  /* Support for fast tracepoints.  */

  PACKET_FastTracepoints,



  /* Support for static tracepoints.  */

  PACKET_StaticTracepoints,



  /* Support for installing tracepoints while a trace experiment is

     running.  */

  PACKET_InstallInTrace,



  PACKET_bc,

  PACKET_bs,

  PACKET_TracepointSource,

  PACKET_QAllow,

  PACKET_qXfer_fdpic,

  PACKET_QDisableRandomization,

  PACKET_QAgent,

  PACKET_QTBuffer_size,

  PACKET_Qbtrace_off,

  PACKET_Qbtrace_bts,

  PACKET_qXfer_btrace,



  /* Support for the QNonStop packet.  */

  PACKET_QNonStop,



  /* Support for multi-process extensions.  */

  PACKET_multiprocess_feature,



  /* Support for enabling and disabling tracepoints while a trace

     experiment is running.  */

  PACKET_EnableDisableTracepoints_feature,



  /* Support for collecting strings using the tracenz bytecode.  */

  PACKET_tracenz_feature,



  /* Support for continuing to run a trace experiment while GDB is

     disconnected.  */

  PACKET_DisconnectedTracing_feature,



  /* Support for qXfer:libraries-svr4:read with a non-empty annex.  */

  PACKET_augmented_libraries_svr4_read_feature,



  PACKET_MAX

};



‌static struct packet_config remote_protocol_packets[PACKET_MAX];



/* Returns whether a given packet or feature is supported.  This takes

   into account the state of the corresponding "set remote foo-packet"

   command, which may be used to bypass auto-detection.  */



static enum packet_support

‌packet_config_support (struct packet_config *config)

{

  switch (config->detect)

    {

    case AUTO_BOOLEAN_TRUE:

      return PACKET_ENABLE;

    case AUTO_BOOLEAN_FALSE:

      return PACKET_DISABLE;

    case AUTO_BOOLEAN_AUTO:

      return config->support;

    default:

      gdb_assert_not_reached (_("bad switch"));

    }

}



/* Same as packet_config_support, but takes the packet's enum value as

   argument.  */



static enum packet_support

‌packet_support (int packet)

{

  struct packet_config *config = &remote_protocol_packets[packet];



  return packet_config_support (config);

}



static void

‌show_remote_protocol_packet_cmd (struct ui_file *file, int from_tty,

                                 struct cmd_list_element *c,

                                 const char *value)

{

  struct packet_config *packet;



  for (packet = remote_protocol_packets;

       packet < &remote_protocol_packets[PACKET_MAX];

       packet++)

    {

      if (&packet->detect == c->var)

        {

          show_packet_config_cmd (packet);

          return;

        }

    }

  internal_error (__FILE__, __LINE__, _("Could not find config for %s"),

                  c->name);

}



/* Should we try one of the 'Z' requests?  */



‌enum Z_packet_type

{

  Z_PACKET_SOFTWARE_BP,

  Z_PACKET_HARDWARE_BP,

  Z_PACKET_WRITE_WP,

  Z_PACKET_READ_WP,

  Z_PACKET_ACCESS_WP,

  NR_Z_PACKET_TYPES

};



/* For compatibility with older distributions.  Provide a ``set remote

   Z-packet ...'' command that updates all the Z packet types.  */



‌static enum auto_boolean remote_Z_packet_detect;



static void

‌set_remote_protocol_Z_packet_cmd (char *args, int from_tty,

                                  struct cmd_list_element *c)

{

  int i;



  for (i = 0; i < NR_Z_PACKET_TYPES; i++)

    remote_protocol_packets[PACKET_Z0 + i].detect = remote_Z_packet_detect;

}



static void

‌show_remote_protocol_Z_packet_cmd (struct ui_file *file, int from_tty,

                                   struct cmd_list_element *c,

                                   const char *value)

{

  int i;



  for (i = 0; i < NR_Z_PACKET_TYPES; i++)

    {

      show_packet_config_cmd (&remote_protocol_packets[PACKET_Z0 + i]);

    }

}



/* Returns true if the multi-process extensions are in effect.  */



static int

‌remote_multi_process_p (struct remote_state *rs)

{

  return packet_support (PACKET_multiprocess_feature) == PACKET_ENABLE;

}



/* Tokens for use by the asynchronous signal handlers for SIGINT.  */

‌static struct async_signal_handler *async_sigint_remote_twice_token;

‌static struct async_signal_handler *async_sigint_remote_token;





/* Asynchronous signal handle registered as event loop source for

   when we have pending events ready to be passed to the core.  */



‌static struct async_event_handler *remote_async_inferior_event_token;







‌static ptid_t magic_null_ptid;

‌static ptid_t not_sent_ptid;

‌static ptid_t any_thread_ptid;



/* Find out if the stub attached to PID (and hence GDB should offer to

   detach instead of killing it when bailing out).  */



static int

‌remote_query_attached (int pid)

{

  struct remote_state *rs = get_remote_state ();

  size_t size = get_remote_packet_size ();



  if (packet_support (PACKET_qAttached) == PACKET_DISABLE)

    return 0;



  if (remote_multi_process_p (rs))

    xsnprintf (rs->buf, size, "qAttached:%x", pid);

  else

    xsnprintf (rs->buf, size, "qAttached");



  putpkt (rs->buf);

  getpkt (&rs->buf, &rs->buf_size, 0);



  switch (packet_ok (rs->buf,

                     &remote_protocol_packets[PACKET_qAttached]))

    {

    case PACKET_OK:

      if (strcmp (rs->buf, "1") == 0)

        return 1;

      break;

    case PACKET_ERROR:

      warning (_("Remote failure reply: %s"), rs->buf);

      break;

    case PACKET_UNKNOWN:

      break;

    }



  return 0;

}



/* Add PID to GDB's inferior table.  If FAKE_PID_P is true, then PID

   has been invented by GDB, instead of reported by the target.  Since

   we can be connected to a remote system before before knowing about

   any inferior, mark the target with execution when we find the first

   inferior.  If ATTACHED is 1, then we had just attached to this

   inferior.  If it is 0, then we just created this inferior.  If it

   is -1, then try querying the remote stub to find out if it had

   attached to the inferior or not.  */



static struct inferior *

‌remote_add_inferior (int fake_pid_p, int pid, int attached)

{

  struct inferior *inf;



  /* Check whether this process we're learning about is to be

     considered attached, or if is to be considered to have been

     spawned by the stub.  */

  if (attached == -1)

    attached = remote_query_attached (pid);



  if (gdbarch_has_global_solist (target_gdbarch ()))

    {

      /* If the target shares code across all inferiors, then every

         attach adds a new inferior.  */

      inf = add_inferior (pid);



      /* ... and every inferior is bound to the same program space.

         However, each inferior may still have its own address

         space.  */

      inf->aspace = maybe_new_address_space ();

      inf->pspace = current_program_space;

    }

  else

    {

      /* In the traditional debugging scenario, there's a 1-1 match

         between program/address spaces.  We simply bind the inferior

         to the program space's address space.  */

      inf = current_inferior ();

      inferior_appeared (inf, pid);

    }



  inf->attach_flag = attached;

  inf->fake_pid_p = fake_pid_p;



  return inf;

}



/* Add thread PTID to GDB's thread list.  Tag it as executing/running

   according to RUNNING.  */



static void

‌remote_add_thread (ptid_t ptid, int running)

{

  struct remote_state *rs = get_remote_state ();



  /* GDB historically didn't pull threads in the initial connection

     setup.  If the remote target doesn't even have a concept of

     threads (e.g., a bare-metal target), even if internally we

     consider that a single-threaded target, mentioning a new thread

     might be confusing to the user.  Be silent then, preserving the

     age old behavior.  */

  if (rs->starting_up)

    add_thread_silent (ptid);

  else

    add_thread (ptid);



  set_executing (ptid, running);

  set_running (ptid, running);

}



/* Come here when we learn about a thread id from the remote target.

   It may be the first time we hear about such thread, so take the

   opportunity to add it to GDB's thread list.  In case this is the

   first time we're noticing its corresponding inferior, add it to

   GDB's inferior list as well.  */



static void

‌remote_notice_new_inferior (ptid_t currthread, int running)

{

  /* If this is a new thread, add it to GDB's thread list.

     If we leave it up to WFI to do this, bad things will happen.  */



  if (in_thread_list (currthread) && is_exited (currthread))

    {

      /* We're seeing an event on a thread id we knew had exited.

         This has to be a new thread reusing the old id.  Add it.  */

      remote_add_thread (currthread, running);

      return;

    }



  if (!in_thread_list (currthread))

    {

      struct inferior *inf = NULL;

      int pid = ptid_get_pid (currthread);



      if (ptid_is_pid (inferior_ptid)

          && pid == ptid_get_pid (inferior_ptid))

        {

          /* inferior_ptid has no thread member yet.  This can happen

             with the vAttach -> remote_wait,"TAAthread:" path if the

             stub doesn't support qC.  This is the first stop reported

             after an attach, so this is the main thread.  Update the

             ptid in the thread list.  */

          if (in_thread_list (pid_to_ptid (pid)))

            thread_change_ptid (inferior_ptid, currthread);

          else

            {

              remote_add_thread (currthread, running);

              inferior_ptid = currthread;

            }

          return;

        }



      if (ptid_equal (magic_null_ptid, inferior_ptid))

        {

          /* inferior_ptid is not set yet.  This can happen with the

             vRun -> remote_wait,"TAAthread:" path if the stub

             doesn't support qC.  This is the first stop reported

             after an attach, so this is the main thread.  Update the

             ptid in the thread list.  */

          thread_change_ptid (inferior_ptid, currthread);

          return;

        }



      /* When connecting to a target remote, or to a target

         extended-remote which already was debugging an inferior, we

         may not know about it yet.  Add it before adding its child

         thread, so notifications are emitted in a sensible order.  */

      if (!in_inferior_list (ptid_get_pid (currthread)))

        {

          struct remote_state *rs = get_remote_state ();

          int fake_pid_p = !remote_multi_process_p (rs);



          inf = remote_add_inferior (fake_pid_p,

                                     ptid_get_pid (currthread), -1);

        }



      /* This is really a new thread.  Add it.  */

      remote_add_thread (currthread, running);



      /* If we found a new inferior, let the common code do whatever

         it needs to with it (e.g., read shared libraries, insert

         breakpoints), unless we're just setting up an all-stop

         connection.  */

      if (inf != NULL)

        {

          struct remote_state *rs = get_remote_state ();



          if (non_stop || !rs->starting_up)

            notice_new_inferior (currthread, running, 0);

        }

    }

}



/* Return the private thread data, creating it if necessary.  */



static struct private_thread_info *

‌demand_private_info (ptid_t ptid)

{

  struct thread_info *info = find_thread_ptid (ptid);



  gdb_assert (info);



  if (!info->private)

    {

      info->private = xmalloc (sizeof (*(info->private)));

      info->private_dtor = free_private_thread_info;

      info->private->core = -1;

      info->private->extra = 0;

    }



  return info->private;

}



/* Call this function as a result of

   1) A halt indication (T packet) containing a thread id

   2) A direct query of currthread

   3) Successful execution of set thread */



static void

‌record_currthread (struct remote_state *rs, ptid_t currthread)

{

  rs->general_thread = currthread;

}



/* If 'QPassSignals' is supported, tell the remote stub what signals

   it can simply pass through to the inferior without reporting.  */



static void

‌remote_pass_signals (struct target_ops *self,

                     int numsigs, unsigned char *pass_signals)

{

  if (packet_support (PACKET_QPassSignals) != PACKET_DISABLE)

    {

      char *pass_packet, *p;

      int count = 0, i;

      struct remote_state *rs = get_remote_state ();



      gdb_assert (numsigs < 256);

      for (i = 0; i < numsigs; i++)

        {

          if (pass_signals[i])

            count++;

        }

      pass_packet = xmalloc (count * 3 + strlen ("QPassSignals:") + 1);

      strcpy (pass_packet, "QPassSignals:");

      p = pass_packet + strlen (pass_packet);

      for (i = 0; i < numsigs; i++)

        {

          if (pass_signals[i])

            {

              if (i >= 16)

                *p++ = tohex (i >> 4);

              *p++ = tohex (i & 15);

              if (count)

                *p++ = ';';

              else

                break;

              count--;

            }

        }

      *p = 0;

      if (!rs->last_pass_packet || strcmp (rs->last_pass_packet, pass_packet))

        {

          putpkt (pass_packet);

          getpkt (&rs->buf, &rs->buf_size, 0);

          packet_ok (rs->buf, &remote_protocol_packets[PACKET_QPassSignals]);

          if (rs->last_pass_packet)

            xfree (rs->last_pass_packet);

          rs->last_pass_packet = pass_packet;

        }

      else

        xfree (pass_packet);

    }

}



/* If 'QProgramSignals' is supported, tell the remote stub what

   signals it should pass through to the inferior when detaching.  */



static void

‌remote_program_signals (struct target_ops *self,

                        int numsigs, unsigned char *signals)

{

  if (packet_support (PACKET_QProgramSignals) != PACKET_DISABLE)

    {

      char *packet, *p;

      int count = 0, i;

      struct remote_state *rs = get_remote_state ();



      gdb_assert (numsigs < 256);

      for (i = 0; i < numsigs; i++)

        {

          if (signals[i])

            count++;

        }

      packet = xmalloc (count * 3 + strlen ("QProgramSignals:") + 1);

      strcpy (packet, "QProgramSignals:");

      p = packet + strlen (packet);

      for (i = 0; i < numsigs; i++)

        {

          if (signal_pass_state (i))

            {

              if (i >= 16)

                *p++ = tohex (i >> 4);

              *p++ = tohex (i & 15);

              if (count)

                *p++ = ';';

              else

                break;

              count--;

            }

        }

      *p = 0;

      if (!rs->last_program_signals_packet

          || strcmp (rs->last_program_signals_packet, packet) != 0)

        {

          putpkt (packet);

          getpkt (&rs->buf, &rs->buf_size, 0);

          packet_ok (rs->buf, &remote_protocol_packets[PACKET_QProgramSignals]);

          xfree (rs->last_program_signals_packet);

          rs->last_program_signals_packet = packet;

        }

      else

        xfree (packet);

    }

}



/* If PTID is MAGIC_NULL_PTID, don't set any thread.  If PTID is

   MINUS_ONE_PTID, set the thread to -1, so the stub returns the

   thread.  If GEN is set, set the general thread, if not, then set

   the step/continue thread.  */

static void

‌set_thread (struct ptid ptid, int gen)

{

  struct remote_state *rs = get_remote_state ();

  ptid_t state = gen ? rs->general_thread : rs->continue_thread;

  char *buf = rs->buf;

  char *endbuf = rs->buf + get_remote_packet_size ();



  if (ptid_equal (state, ptid))

    return;



  *buf++ = 'H';

  *buf++ = gen ? 'g' : 'c';

  if (ptid_equal (ptid, magic_null_ptid))

    xsnprintf (buf, endbuf - buf, "0");

  else if (ptid_equal (ptid, any_thread_ptid))

    xsnprintf (buf, endbuf - buf, "0");

  else if (ptid_equal (ptid, minus_one_ptid))

    xsnprintf (buf, endbuf - buf, "-1");

  else

    write_ptid (buf, endbuf, ptid);

  putpkt (rs->buf);

  getpkt (&rs->buf, &rs->buf_size, 0);

  if (gen)

    rs->general_thread = ptid;

  else

    rs->continue_thread = ptid;

}



static void

‌set_general_thread (struct ptid ptid)

{

  set_thread (ptid, 1);

}



static void

‌set_continue_thread (struct ptid ptid)

{

  set_thread (ptid, 0);

}



/* Change the remote current process.  Which thread within the process

   ends up selected isn't important, as long as it is the same process

   as what INFERIOR_PTID points to.



   This comes from that fact that there is no explicit notion of

   "selected process" in the protocol.  The selected process for

   general operations is the process the selected general thread

   belongs to.  */



static void

‌set_general_process (void)

{

  struct remote_state *rs = get_remote_state ();



  /* If the remote can't handle multiple processes, don't bother.  */

  if (!rs->extended || !remote_multi_process_p (rs))

    return;



  /* We only need to change the remote current thread if it's pointing

     at some other process.  */

  if (ptid_get_pid (rs->general_thread) != ptid_get_pid (inferior_ptid))

    set_general_thread (inferior_ptid);

}





/* Return nonzero if this is the main thread that we made up ourselves

   to model non-threaded targets as single-threaded.  */



static int

‌remote_thread_always_alive (struct target_ops *ops, ptid_t ptid)

{

  struct remote_state *rs = get_remote_state ();

  char *p, *endp;



  if (ptid_equal (ptid, magic_null_ptid))

    /* The main thread is always alive.  */

    return 1;



  if (ptid_get_pid (ptid) != 0 && ptid_get_lwp (ptid) == 0)

    /* The main thread is always alive.  This can happen after a

       vAttach, if the remote side doesn't support

       multi-threading.  */

    return 1;



  return 0;

}



/* Return nonzero if the thread PTID is still alive on the remote

   system.  */



static int

‌remote_thread_alive (struct target_ops *ops, ptid_t ptid)

{

  struct remote_state *rs = get_remote_state ();

  char *p, *endp;



  /* Check if this is a thread that we made up ourselves to model

     non-threaded targets as single-threaded.  */

  if (remote_thread_always_alive (ops, ptid))

    return 1;



  p = rs->buf;

  endp = rs->buf + get_remote_packet_size ();



  *p++ = 'T';

  write_ptid (p, endp, ptid);



  putpkt (rs->buf);

  getpkt (&rs->buf, &rs->buf_size, 0);

  return (rs->buf[0] == 'O' && rs->buf[1] == 'K');

}



/* About these extended threadlist and threadinfo packets.  They are

   variable length packets but, the fields within them are often fixed

   length.  They are redundent enough to send over UDP as is the

   remote protocol in general.  There is a matching unit test module

   in libstub.  */



/* WARNING: This threadref data structure comes from the remote O.S.,

   libstub protocol encoding, and remote.c.  It is not particularly

   changable.  */



/* Right now, the internal structure is int. We want it to be bigger.

   Plan to fix this.  */



‌typedef int gdb_threadref;        /* Internal GDB thread reference.  */



/* gdb_ext_thread_info is an internal GDB data structure which is

   equivalent to the reply of the remote threadinfo packet.  */



‌struct gdb_ext_thread_info

  {

    threadref threadid;                /* External form of thread reference.  */

    int active;                        /* Has state interesting to GDB?

                                   regs, stack.  */

    char display[256];                /* Brief state display, name,

                                   blocked/suspended.  */

    char shortname[32];                /* To be used to name threads.  */

    char more_display[256];        /* Long info, statistics, queue depth,

                                   whatever.  */

  };



/* The volume of remote transfers can be limited by submitting

   a mask containing bits specifying the desired information.

   Use a union of these values as the 'selection' parameter to

   get_thread_info.  FIXME: Make these TAG names more thread specific.  */



‌#define TAG_THREADID 1

‌#define TAG_EXISTS 2

‌#define TAG_DISPLAY 4

‌#define TAG_THREADNAME 8

‌#define TAG_MOREDISPLAY 16



‌#define BUF_THREAD_ID_SIZE (OPAQUETHREADBYTES * 2)



static char *unpack_nibble (char *buf, int *val);



static char *unpack_byte (char *buf, int *value);



static char *pack_int (char *buf, int value);



static char *unpack_int (char *buf, int *value);



static char *unpack_string (char *src, char *dest, int length);



static char *pack_threadid (char *pkt, threadref *id);



static char *unpack_threadid (char *inbuf, threadref *id);



void int_to_threadref (threadref *id, int value);



static int threadref_to_int (threadref *ref);



static void copy_threadref (threadref *dest, threadref *src);



static int threadmatch (threadref *dest, threadref *src);



static char *pack_threadinfo_request (char *pkt, int mode,

                                      threadref *id);



static int remote_unpack_thread_info_response (char *pkt,

                                               threadref *expectedref,

                                               struct gdb_ext_thread_info

                                               *info);





static int remote_get_threadinfo (threadref *threadid,

                                  int fieldset,        /*TAG mask */

                                  struct gdb_ext_thread_info *info);



static char *pack_threadlist_request (char *pkt, int startflag,

                                      int threadcount,

                                      threadref *nextthread);



static int parse_threadlist_response (char *pkt,

                                      int result_limit,

                                      threadref *original_echo,

                                      threadref *resultlist,

                                      int *doneflag);



static int remote_get_threadlist (int startflag,

                                  threadref *nextthread,

                                  int result_limit,

                                  int *done,

                                  int *result_count,

                                  threadref *threadlist);



‌typedef int (*rmt_thread_action) (threadref *ref, void *context);



static int remote_threadlist_iterator (rmt_thread_action stepfunction,

                                       void *context, int looplimit);



static int remote_newthread_step (threadref *ref, void *context);





/* Write a PTID to BUF.  ENDBUF points to one-passed-the-end of the

   buffer we're allowed to write to.  Returns

   BUF+CHARACTERS_WRITTEN.  */



static char *

‌write_ptid (char *buf, const char *endbuf, ptid_t ptid)

{

  int pid, tid;

  struct remote_state *rs = get_remote_state ();



  if (remote_multi_process_p (rs))

    {

      pid = ptid_get_pid (ptid);

      if (pid < 0)

        buf += xsnprintf (buf, endbuf - buf, "p-%x.", -pid);

      else

        buf += xsnprintf (buf, endbuf - buf, "p%x.", pid);

    }

  tid = ptid_get_lwp (ptid);

  if (tid < 0)

    buf += xsnprintf (buf, endbuf - buf, "-%x", -tid);

  else

    buf += xsnprintf (buf, endbuf - buf, "%x", tid);



  return buf;

}



/* Extract a PTID from BUF.  If non-null, OBUF is set to the to one

   passed the last parsed char.  Returns null_ptid on error.  */



static ptid_t

‌read_ptid (char *buf, char **obuf)

{

  char *p = buf;

  char *pp;

  ULONGEST pid = 0, tid = 0;



  if (*p == 'p')

    {

      /* Multi-process ptid.  */

      pp = unpack_varlen_hex (p + 1, &pid);

      if (*pp != '.')

        error (_("invalid remote ptid: %s"), p);



      p = pp;

      pp = unpack_varlen_hex (p + 1, &tid);

      if (obuf)

        *obuf = pp;

      return ptid_build (pid, tid, 0);

    }



  /* No multi-process.  Just a tid.  */

  pp = unpack_varlen_hex (p, &tid);



  /* Since the stub is not sending a process id, then default to

     what's in inferior_ptid, unless it's null at this point.  If so,

     then since there's no way to know the pid of the reported

     threads, use the magic number.  */

  if (ptid_equal (inferior_ptid, null_ptid))

    pid = ptid_get_pid (magic_null_ptid);

  else

    pid = ptid_get_pid (inferior_ptid);



  if (obuf)

    *obuf = pp;

  return ptid_build (pid, tid, 0);

}



static int

‌stubhex (int ch)

{

  if (ch >= 'a' && ch <= 'f')

    return ch - 'a' + 10;

  if (ch >= '0' && ch <= '9')

    return ch - '0';

  if (ch >= 'A' && ch <= 'F')

    return ch - 'A' + 10;

  return -1;

}



static int

‌stub_unpack_int (char *buff, int fieldlength)

{

  int nibble;

  int retval = 0;



  while (fieldlength)

    {

      nibble = stubhex (*buff++);

      retval |= nibble;

      fieldlength--;

      if (fieldlength)

        retval = retval << 4;

    }

  return retval;

}



static char *

‌unpack_nibble (char *buf, int *val)

{

  *val = fromhex (*buf++);

  return buf;

}



static char *

‌unpack_byte (char *buf, int *value)

{

  *value = stub_unpack_int (buf, 2);

  return buf + 2;

}



static char *

‌pack_int (char *buf, int value)

{

  buf = pack_hex_byte (buf, (value >> 24) & 0xff);

  buf = pack_hex_byte (buf, (value >> 16) & 0xff);

  buf = pack_hex_byte (buf, (value >> 8) & 0x0ff);

  buf = pack_hex_byte (buf, (value & 0xff));

  return buf;

}



static char *

‌unpack_int (char *buf, int *value)

{

  *value = stub_unpack_int (buf, 8);

  return buf + 8;

}



#if 0                        /* Currently unused, uncomment when needed.  */

static char *pack_string (char *pkt, char *string);



static char *

pack_string (char *pkt, char *string)

{

  char ch;

  int len;



  len = strlen (string);

  if (len > 200)

    len = 200;                /* Bigger than most GDB packets, junk???  */

  pkt = pack_hex_byte (pkt, len);

  while (len-- > 0)

    {

      ch = *string++;

      if ((ch == '\0') || (ch == '#'))

        ch = '*';                /* Protect encapsulation.  */

      *pkt++ = ch;

    }

  return pkt;

}

#endif /* 0 (unused) */



static char *

‌unpack_string (char *src, char *dest, int length)

{

  while (length--)

    *dest++ = *src++;

  *dest = '\0';

  return src;

}



static char *

‌pack_threadid (char *pkt, threadref *id)

{

  char *limit;

  unsigned char *altid;



  altid = (unsigned char *) id;

  limit = pkt + BUF_THREAD_ID_SIZE;

  while (pkt < limit)

    pkt = pack_hex_byte (pkt, *altid++);

  return pkt;

}





static char *

‌unpack_threadid (char *inbuf, threadref *id)

{

  char *altref;

  char *limit = inbuf + BUF_THREAD_ID_SIZE;

  int x, y;



  altref = (char *) id;



  while (inbuf < limit)

    {

      x = stubhex (*inbuf++);

      y = stubhex (*inbuf++);

      *altref++ = (x << 4) | y;

    }

  return inbuf;

}



/* Externally, threadrefs are 64 bits but internally, they are still

   ints.  This is due to a mismatch of specifications.  We would like

   to use 64bit thread references internally.  This is an adapter

   function.  */



void

‌int_to_threadref (threadref *id, int value)

{

  unsigned char *scan;



  scan = (unsigned char *) id;

  {

    int i = 4;

    while (i--)

      *scan++ = 0;

  }

  *scan++ = (value >> 24) & 0xff;

  *scan++ = (value >> 16) & 0xff;

  *scan++ = (value >> 8) & 0xff;

  *scan++ = (value & 0xff);

}



static int

‌threadref_to_int (threadref *ref)

{

  int i, value = 0;

  unsigned char *scan;



  scan = *ref;

  scan += 4;

  i = 4;

  while (i-- > 0)

    value = (value << 8) | ((*scan++) & 0xff);

  return value;

}



static void

‌copy_threadref (threadref *dest, threadref *src)

{

  int i;

  unsigned char *csrc, *cdest;



  csrc = (unsigned char *) src;

  cdest = (unsigned char *) dest;

  i = 8;

  while (i--)

    *cdest++ = *csrc++;

}



static int

‌threadmatch (threadref *dest, threadref *src)

{

  /* Things are broken right now, so just assume we got a match.  */

#if 0

  unsigned char *srcp, *destp;

  int i, result;

  srcp = (char *) src;

  destp = (char *) dest;



  result = 1;

  while (i-- > 0)

    result &= (*srcp++ == *destp++) ? 1 : 0;

  return result;

#endif

  return 1;

}



/*

   threadid:1,        # always request threadid

   context_exists:2,

   display:4,

   unique_name:8,

   more_display:16

 */



/* Encoding:  'Q':8,'P':8,mask:32,threadid:64 */



static char *

‌pack_threadinfo_request (char *pkt, int mode, threadref *id)

{

  *pkt++ = 'q';                                /* Info Query */

  *pkt++ = 'P';                                /* process or thread info */

  pkt = pack_int (pkt, mode);                /* mode */

  pkt = pack_threadid (pkt, id);        /* threadid */

  *pkt = '\0';                                /* terminate */

  return pkt;

}



/* These values tag the fields in a thread info response packet.  */

/* Tagging the fields allows us to request specific fields and to

   add more fields as time goes by.  */



‌#define TAG_THREADID 1                /* Echo the thread identifier.  */

‌#define TAG_EXISTS 2                /* Is this process defined enough to

                                   fetch registers and its stack?  */

‌#define TAG_DISPLAY 4                /* A short thing maybe to put on a window */

‌#define TAG_THREADNAME 8        /* string, maps 1-to-1 with a thread is.  */

‌#define TAG_MOREDISPLAY 16        /* Whatever the kernel wants to say about

                                   the process.  */



static int

‌remote_unpack_thread_info_response (char *pkt, threadref *expectedref,

                                    struct gdb_ext_thread_info *info)

{

  struct remote_state *rs = get_remote_state ();

  int mask, length;

  int tag;

  threadref ref;

  char *limit = pkt + rs->buf_size; /* Plausible parsing limit.  */

  int retval = 1;



  /* info->threadid = 0; FIXME: implement zero_threadref.  */

  info->active = 0;

  info->display[0] = '\0';

  info->shortname[0] = '\0';

  info->more_display[0] = '\0';



  /* Assume the characters indicating the packet type have been

     stripped.  */

  pkt = unpack_int (pkt, &mask);        /* arg mask */

  pkt = unpack_threadid (pkt, &ref);



  if (mask == 0)

    warning (_("Incomplete response to threadinfo request."));

  if (!threadmatch (&ref, expectedref))

    {                        /* This is an answer to a different request.  */

      warning (_("ERROR RMT Thread info mismatch."));

      return 0;

    }

  copy_threadref (&info->threadid, &ref);



  /* Loop on tagged fields , try to bail if somthing goes wrong.  */



  /* Packets are terminated with nulls.  */

  while ((pkt < limit) && mask && *pkt)

    {

      pkt = unpack_int (pkt, &tag);        /* tag */

      pkt = unpack_byte (pkt, &length);        /* length */

      if (!(tag & mask))                /* Tags out of synch with mask.  */

        {

          warning (_("ERROR RMT: threadinfo tag mismatch."));

          retval = 0;

          break;

        }

      if (tag == TAG_THREADID)

        {

          if (length != 16)

            {

              warning (_("ERROR RMT: length of threadid is not 16."));

              retval = 0;

              break;

            }

          pkt = unpack_threadid (pkt, &ref);

          mask = mask & ~TAG_THREADID;

          continue;

        }

      if (tag == TAG_EXISTS)

        {

          info->active = stub_unpack_int (pkt, length);

          pkt += length;

          mask = mask & ~(TAG_EXISTS);

          if (length > 8)

            {

              warning (_("ERROR RMT: 'exists' length too long."));

              retval = 0;

              break;

            }

          continue;

        }

      if (tag == TAG_THREADNAME)

        {

          pkt = unpack_string (pkt, &info->shortname[0], length);

          mask = mask & ~TAG_THREADNAME;

          continue;

        }

      if (tag == TAG_DISPLAY)

        {

          pkt = unpack_string (pkt, &info->display[0], length);

          mask = mask & ~TAG_DISPLAY;

          continue;

        }

      if (tag == TAG_MOREDISPLAY)

        {

          pkt = unpack_string (pkt, &info->more_display[0], length);

          mask = mask & ~TAG_MOREDISPLAY;

          continue;

        }

      warning (_("ERROR RMT: unknown thread info tag."));

      break;                        /* Not a tag we know about.  */

    }

  return retval;

}



static int

‌remote_get_threadinfo (threadref *threadid, int fieldset,        /* TAG mask */

                       struct gdb_ext_thread_info *info)

{

  struct remote_state *rs = get_remote_state ();

  int result;



  pack_threadinfo_request (rs->buf, fieldset, threadid);

  putpkt (rs->buf);

  getpkt (&rs->buf, &rs->buf_size, 0);



  if (rs->buf[0] == '\0')

    return 0;



  result = remote_unpack_thread_info_response (rs->buf + 2,

                                               threadid, info);

  return result;

}



/*    Format: i'Q':8,i"L":8,initflag:8,batchsize:16,lastthreadid:32   */



static char *

‌pack_threadlist_request (char *pkt, int startflag, int threadcount,

                         threadref *nextthread)

{

  *pkt++ = 'q';                        /* info query packet */

  *pkt++ = 'L';                        /* Process LIST or threadLIST request */

  pkt = pack_nibble (pkt, startflag);                /* initflag 1 bytes */

  pkt = pack_hex_byte (pkt, threadcount);        /* threadcount 2 bytes */

  pkt = pack_threadid (pkt, nextthread);        /* 64 bit thread identifier */

  *pkt = '\0';

  return pkt;

}



/* Encoding:   'q':8,'M':8,count:16,done:8,argthreadid:64,(threadid:64)* */



static int

‌parse_threadlist_response (char *pkt, int result_limit,

                           threadref *original_echo, threadref *resultlist,

                           int *doneflag)

{

  struct remote_state *rs = get_remote_state ();

  char *limit;

  int count, resultcount, done;



  resultcount = 0;

  /* Assume the 'q' and 'M chars have been stripped.  */

  limit = pkt + (rs->buf_size - BUF_THREAD_ID_SIZE);

  /* done parse past here */

  pkt = unpack_byte (pkt, &count);        /* count field */

  pkt = unpack_nibble (pkt, &done);

  /* The first threadid is the argument threadid.  */

  pkt = unpack_threadid (pkt, original_echo);        /* should match query packet */

  while ((count-- > 0) && (pkt < limit))

    {

      pkt = unpack_threadid (pkt, resultlist++);

      if (resultcount++ >= result_limit)

        break;

    }

  if (doneflag)

    *doneflag = done;

  return resultcount;

}



/* Fetch the next batch of threads from the remote.  Returns -1 if the

   qL packet is not supported, 0 on error and 1 on success.  */



static int

‌remote_get_threadlist (int startflag, threadref *nextthread, int result_limit,

                       int *done, int *result_count, threadref *threadlist)

{

  struct remote_state *rs = get_remote_state ();

  int result = 1;



  /* Trancate result limit to be smaller than the packet size.  */

  if ((((result_limit + 1) * BUF_THREAD_ID_SIZE) + 10)

      >= get_remote_packet_size ())

    result_limit = (get_remote_packet_size () / BUF_THREAD_ID_SIZE) - 2;



  pack_threadlist_request (rs->buf, startflag, result_limit, nextthread);

  putpkt (rs->buf);

  getpkt (&rs->buf, &rs->buf_size, 0);

  if (*rs->buf == '\0')

    {

      /* Packet not supported.  */

      return -1;

    }



  *result_count =

    parse_threadlist_response (rs->buf + 2, result_limit,

                               &rs->echo_nextthread, threadlist, done);



  if (!threadmatch (&rs->echo_nextthread, nextthread))

    {

      /* FIXME: This is a good reason to drop the packet.  */

      /* Possably, there is a duplicate response.  */

      /* Possabilities :

         retransmit immediatly - race conditions

         retransmit after timeout - yes

         exit

         wait for packet, then exit

       */

      warning (_("HMM: threadlist did not echo arg thread, dropping it."));

      return 0;                        /* I choose simply exiting.  */

    }

  if (*result_count <= 0)

    {

      if (*done != 1)

        {

          warning (_("RMT ERROR : failed to get remote thread list."));

          result = 0;

        }

      return result;                /* break; */

    }

  if (*result_count > result_limit)

    {

      *result_count = 0;

      warning (_("RMT ERROR: threadlist response longer than requested."));

      return 0;

    }

  return result;

}



/* Fetch the list of remote threads, with the qL packet, and call

   STEPFUNCTION for each thread found.  Stops iterating and returns 1

   if STEPFUNCTION returns true.  Stops iterating and returns 0 if the

   STEPFUNCTION returns false.  If the packet is not supported,

   returns -1.  */



static int

‌remote_threadlist_iterator (rmt_thread_action stepfunction, void *context,

                            int looplimit)

{

  struct remote_state *rs = get_remote_state ();

  int done, i, result_count;

  int startflag = 1;

  int result = 1;

  int loopcount = 0;



  done = 0;

  while (!done)

    {

      if (loopcount++ > looplimit)

        {

          result = 0;

          warning (_("Remote fetch threadlist -infinite loop-."));

          break;

        }

      result = remote_get_threadlist (startflag, &rs->nextthread,

                                      MAXTHREADLISTRESULTS,

                                      &done, &result_count,

                                      rs->resultthreadlist);

      if (result <= 0)

        break;

      /* Clear for later iterations.  */

      startflag = 0;

      /* Setup to resume next batch of thread references, set nextthread.  */

      if (result_count >= 1)

        copy_threadref (&rs->nextthread,

                        &rs->resultthreadlist[result_count - 1]);

      i = 0;

      while (result_count--)

        {

          if (!(*stepfunction) (&rs->resultthreadlist[i++], context))

            {

              result = 0;

              break;

            }

        }

    }

  return result;

}



/* A thread found on the remote target.  */



‌typedef struct thread_item

{

  /* The thread's PTID.  */

  ptid_t ptid;



  /* The thread's extra info.  May be NULL.  */

  char *extra;



  /* The core the thread was running on.  -1 if not known.  */

  int core;

‌} thread_item_t;

‌DEF_VEC_O(thread_item_t);



/* Context passed around to the various methods listing remote

   threads.  As new threads are found, they're added to the ITEMS

   vector.  */



‌struct threads_listing_context

{

  /* The threads found on the remote target.  */

  VEC (thread_item_t) *items;

};



/* Discard the contents of the constructed thread listing context.  */



static void

‌clear_threads_listing_context (void *p)

{

  struct threads_listing_context *context = p;

  int i;

  struct thread_item *item;



  for (i = 0; VEC_iterate (thread_item_t, context->items, i, item); ++i)

    xfree (item->extra);



  VEC_free (thread_item_t, context->items);

}



static int

‌remote_newthread_step (threadref *ref, void *data)

{

  struct threads_listing_context *context = data;

  struct thread_item item;

  int pid = ptid_get_pid (inferior_ptid);



  item.ptid = ptid_build (pid, threadref_to_int (ref), 0);

  item.core = -1;

  item.extra = NULL;



  VEC_safe_push (thread_item_t, context->items, &item);



  return 1;                        /* continue iterator */

}



‌#define CRAZY_MAX_THREADS 1000



static ptid_t

‌remote_current_thread (ptid_t oldpid)

{

  struct remote_state *rs = get_remote_state ();



  putpkt ("qC");

  getpkt (&rs->buf, &rs->buf_size, 0);

  if (rs->buf[0] == 'Q' && rs->buf[1] == 'C')

    return read_ptid (&rs->buf[2], NULL);

  else

    return oldpid;

}



/* List remote threads using the deprecated qL packet.  */



static int

‌remote_get_threads_with_ql (struct target_ops *ops,

                            struct threads_listing_context *context)

{

  if (remote_threadlist_iterator (remote_newthread_step, context,

                                  CRAZY_MAX_THREADS) >= 0)

    return 1;



  return 0;

}



#if defined(HAVE_LIBEXPAT)



static void

‌start_thread (struct gdb_xml_parser *parser,

              const struct gdb_xml_element *element,

              void *user_data, VEC(gdb_xml_value_s) *attributes)

{

  struct threads_listing_context *data = user_data;



  struct thread_item item;

  char *id;

  struct gdb_xml_value *attr;



  id = xml_find_attribute (attributes, "id")->value;

  item.ptid = read_ptid (id, NULL);



  attr = xml_find_attribute (attributes, "core");

  if (attr != NULL)

    item.core = *(ULONGEST *) attr->value;

  else

    item.core = -1;



  item.extra = 0;



  VEC_safe_push (thread_item_t, data->items, &item);

}



static void

‌end_thread (struct gdb_xml_parser *parser,

            const struct gdb_xml_element *element,

            void *user_data, const char *body_text)

{

  struct threads_listing_context *data = user_data;



  if (body_text && *body_text)

    VEC_last (thread_item_t, data->items)->extra = xstrdup (body_text);

}



‌const struct gdb_xml_attribute thread_attributes[] = {

  { "id", GDB_XML_AF_NONE, NULL, NULL },

  { "core", GDB_XML_AF_OPTIONAL, gdb_xml_parse_attr_ulongest, NULL },

  { NULL, GDB_XML_AF_NONE, NULL, NULL }

};



‌const struct gdb_xml_element thread_children[] = {

  { NULL, NULL, NULL, GDB_XML_EF_NONE, NULL, NULL }

};



‌const struct gdb_xml_element threads_children[] = {

  { "thread", thread_attributes, thread_children,

    GDB_XML_EF_REPEATABLE | GDB_XML_EF_OPTIONAL,

    start_thread, end_thread },

  { NULL, NULL, NULL, GDB_XML_EF_NONE, NULL, NULL }

};



‌const struct gdb_xml_element threads_elements[] = {

  { "threads", NULL, threads_children,

    GDB_XML_EF_NONE, NULL, NULL },

  { NULL, NULL, NULL, GDB_XML_EF_NONE, NULL, NULL }

};



#endif



/* List remote threads using qXfer:threads:read.  */



static int

‌remote_get_threads_with_qxfer (struct target_ops *ops,

                               struct threads_listing_context *context)

{

#if defined(HAVE_LIBEXPAT)

  if (packet_support (PACKET_qXfer_threads) == PACKET_ENABLE)

    {

      char *xml = target_read_stralloc (ops, TARGET_OBJECT_THREADS, NULL);

      struct cleanup *back_to = make_cleanup (xfree, xml);



      if (xml != NULL && *xml != '\0')

        {

          gdb_xml_parse_quick (_("threads"), "threads.dtd",

                               threads_elements, xml, context);

        }



      do_cleanups (back_to);

      return 1;

    }

#endif



  return 0;

}



/* List remote threads using qfThreadInfo/qsThreadInfo.  */



static int

‌remote_get_threads_with_qthreadinfo (struct target_ops *ops,

                                     struct threads_listing_context *context)

{

  struct remote_state *rs = get_remote_state ();



  if (rs->use_threadinfo_query)

    {

      char *bufp;



      putpkt ("qfThreadInfo");

      getpkt (&rs->buf, &rs->buf_size, 0);

      bufp = rs->buf;

      if (bufp[0] != '\0')                /* q packet recognized */

        {

          while (*bufp++ == 'm')        /* reply contains one or more TID */

            {

              do

                {

                  struct thread_item item;



                  item.ptid = read_ptid (bufp, &bufp);

                  item.core = -1;

                  item.extra = NULL;



                  VEC_safe_push (thread_item_t, context->items, &item);

                }

              while (*bufp++ == ',');        /* comma-separated list */

              putpkt ("qsThreadInfo");

              getpkt (&rs->buf, &rs->buf_size, 0);

              bufp = rs->buf;

            }

          return 1;

        }

      else

        {

          /* Packet not recognized.  */

          rs->use_threadinfo_query = 0;

        }

    }



  return 0;

}



/* Implement the to_update_thread_list function for the remote

   targets.  */



static void

‌remote_update_thread_list (struct target_ops *ops)

{

  struct remote_state *rs = get_remote_state ();

  struct threads_listing_context context;

  struct cleanup *old_chain;

  int got_list = 0;



  context.items = NULL;

  old_chain = make_cleanup (clear_threads_listing_context, &context);



  /* We have a few different mechanisms to fetch the thread list.  Try

     them all, starting with the most preferred one first, falling

     back to older methods.  */

  if (remote_get_threads_with_qxfer (ops, &context)

      || remote_get_threads_with_qthreadinfo (ops, &context)

      || remote_get_threads_with_ql (ops, &context))

    {

      int i;

      struct thread_item *item;

      struct thread_info *tp, *tmp;



      got_list = 1;



      if (VEC_empty (thread_item_t, context.items)

          && remote_thread_always_alive (ops, inferior_ptid))

        {

          /* Some targets don't really support threads, but still

             reply an (empty) thread list in response to the thread

             listing packets, instead of replying "packet not

             supported".  Exit early so we don't delete the main

             thread.  */

          do_cleanups (old_chain);

          return;

        }



      /* CONTEXT now holds the current thread list on the remote

         target end.  Delete GDB-side threads no longer found on the

         target.  */

      ALL_NON_EXITED_THREADS_SAFE (tp, tmp)

        {

          for (i = 0;

               VEC_iterate (thread_item_t, context.items, i, item);

               ++i)

            {

              if (ptid_equal (item->ptid, tp->ptid))

                break;

            }



          if (i == VEC_length (thread_item_t, context.items))

            {

              /* Not found.  */

              delete_thread (tp->ptid);

            }

        }



      /* And now add threads we don't know about yet to our list.  */

      for (i = 0;

           VEC_iterate (thread_item_t, context.items, i, item);

           ++i)

        {

          if (!ptid_equal (item->ptid, null_ptid))

            {

              struct private_thread_info *info;

              /* In non-stop mode, we assume new found threads are

                 running until proven otherwise with a stop reply.  In

                 all-stop, we can only get here if all threads are

                 stopped.  */

              int running = non_stop ? 1 : 0;



              remote_notice_new_inferior (item->ptid, running);



              info = demand_private_info (item->ptid);

              info->core = item->core;

              info->extra = item->extra;

              item->extra = NULL;

            }

        }

    }



  if (!got_list)

    {

      /* If no thread listing method is supported, then query whether

         each known thread is alive, one by one, with the T packet.

         If the target doesn't support threads at all, then this is a

         no-op.  See remote_thread_alive.  */

      prune_threads ();

    }



  do_cleanups (old_chain);

}



/*

 * Collect a descriptive string about the given thread.

 * The target may say anything it wants to about the thread

 * (typically info about its blocked / runnable state, name, etc.).

 * This string will appear in the info threads display.

 *

 * Optional: targets are not required to implement this function.

 */



static char *

‌remote_threads_extra_info (struct target_ops *self, struct thread_info *tp)

{

  struct remote_state *rs = get_remote_state ();

  int result;

  int set;

  threadref id;

  struct gdb_ext_thread_info threadinfo;

  static char display_buf[100];        /* arbitrary...  */

  int n = 0;                    /* position in display_buf */



  if (rs->remote_desc == 0)                /* paranoia */

    internal_error (__FILE__, __LINE__,

                    _("remote_threads_extra_info"));



  if (ptid_equal (tp->ptid, magic_null_ptid)

      || (ptid_get_pid (tp->ptid) != 0 && ptid_get_lwp (tp->ptid) == 0))

    /* This is the main thread which was added by GDB.  The remote

       server doesn't know about it.  */

    return NULL;



  if (packet_support (PACKET_qXfer_threads) == PACKET_ENABLE)

    {

      struct thread_info *info = find_thread_ptid (tp->ptid);



      if (info && info->private)

        return info->private->extra;

      else

        return NULL;

    }



  if (rs->use_threadextra_query)

    {

      char *b = rs->buf;

      char *endb = rs->buf + get_remote_packet_size ();



      xsnprintf (b, endb - b, "qThreadExtraInfo,");

      b += strlen (b);

      write_ptid (b, endb, tp->ptid);



      putpkt (rs->buf);

      getpkt (&rs->buf, &rs->buf_size, 0);

      if (rs->buf[0] != 0)

        {

          n = min (strlen (rs->buf) / 2, sizeof (display_buf));

          result = hex2bin (rs->buf, (gdb_byte *) display_buf, n);

          display_buf [result] = '\0';

          return display_buf;

        }

    }



  /* If the above query fails, fall back to the old method.  */

  rs->use_threadextra_query = 0;

  set = TAG_THREADID | TAG_EXISTS | TAG_THREADNAME

    | TAG_MOREDISPLAY | TAG_DISPLAY;

  int_to_threadref (&id, ptid_get_lwp (tp->ptid));

  if (remote_get_threadinfo (&id, set, &threadinfo))

    if (threadinfo.active)

      {

        if (*threadinfo.shortname)

          n += xsnprintf (&display_buf[0], sizeof (display_buf) - n,

                          " Name: %s,", threadinfo.shortname);

        if (*threadinfo.display)

          n += xsnprintf (&display_buf[n], sizeof (display_buf) - n,

                          " State: %s,", threadinfo.display);

        if (*threadinfo.more_display)

          n += xsnprintf (&display_buf[n], sizeof (display_buf) - n,

                          " Priority: %s", threadinfo.more_display);



        if (n > 0)

          {

            /* For purely cosmetic reasons, clear up trailing commas.  */

            if (',' == display_buf[n-1])

              display_buf[n-1] = ' ';

            return display_buf;

          }

      }

  return NULL;

}





static int

‌remote_static_tracepoint_marker_at (struct target_ops *self, CORE_ADDR addr,

                                    struct static_tracepoint_marker *marker)

{

  struct remote_state *rs = get_remote_state ();

  char *p = rs->buf;



  xsnprintf (p, get_remote_packet_size (), "qTSTMat:");

  p += strlen (p);

  p += hexnumstr (p, addr);

  putpkt (rs->buf);

  getpkt (&rs->buf, &rs->buf_size, 0);

  p = rs->buf;



  if (*p == 'E')

    error (_("Remote failure reply: %s"), p);



  if (*p++ == 'm')

    {

      parse_static_tracepoint_marker_definition (p, &p, marker);

      return 1;

    }



  return 0;

}



‌static VEC(static_tracepoint_marker_p) *

remote_static_tracepoint_markers_by_strid (struct target_ops *self,

                                           const char *strid)

{

  struct remote_state *rs = get_remote_state ();

  VEC(static_tracepoint_marker_p) *markers = NULL;

  struct static_tracepoint_marker *marker = NULL;

  struct cleanup *old_chain;

  char *p;



  /* Ask for a first packet of static tracepoint marker

     definition.  */

  putpkt ("qTfSTM");

  getpkt (&rs->buf, &rs->buf_size, 0);

  p = rs->buf;

  if (*p == 'E')

    error (_("Remote failure reply: %s"), p);



  old_chain = make_cleanup (free_current_marker, &marker);



  while (*p++ == 'm')

    {

      if (marker == NULL)

        marker = XCNEW (struct static_tracepoint_marker);



      do

        {

          parse_static_tracepoint_marker_definition (p, &p, marker);



          if (strid == NULL || strcmp (strid, marker->str_id) == 0)

            {

              VEC_safe_push (static_tracepoint_marker_p,

                             markers, marker);

              marker = NULL;

            }

          else

            {

              release_static_tracepoint_marker (marker);

              memset (marker, 0, sizeof (*marker));

            }

        }

      while (*p++ == ',');        /* comma-separated list */

      /* Ask for another packet of static tracepoint definition.  */

      putpkt ("qTsSTM");

      getpkt (&rs->buf, &rs->buf_size, 0);

      p = rs->buf;

    }



  do_cleanups (old_chain);

  return markers;

}





/* Implement the to_get_ada_task_ptid function for the remote targets.  */



static ptid_t

‌remote_get_ada_task_ptid (struct target_ops *self, long lwp, long thread)

{

  return ptid_build (ptid_get_pid (inferior_ptid), lwp, 0);

}





/* Restart the remote side; this is an extended protocol operation.  */



static void

‌extended_remote_restart (void)

{

  struct remote_state *rs = get_remote_state ();



  /* Send the restart command; for reasons I don't understand the

     remote side really expects a number after the "R".  */

  xsnprintf (rs->buf, get_remote_packet_size (), "R%x", 0);

  putpkt (rs->buf);



  remote_fileio_reset ();

}



/* Clean up connection to a remote debugger.  */



static void

‌remote_close (struct target_ops *self)

{

  struct remote_state *rs = get_remote_state ();



  if (rs->remote_desc == NULL)

    return; /* already closed */



  /* Make sure we leave stdin registered in the event loop, and we

     don't leave the async SIGINT signal handler installed.  */

  remote_terminal_ours (self);



  serial_close (rs->remote_desc);

  rs->remote_desc = NULL;



  /* We don't have a connection to the remote stub anymore.  Get rid

     of all the inferiors and their threads we were controlling.

     Reset inferior_ptid to null_ptid first, as otherwise has_stack_frame

     will be unable to find the thread corresponding to (pid, 0, 0).  */

  inferior_ptid = null_ptid;

  discard_all_inferiors ();



  /* We are closing the remote target, so we should discard

     everything of this target.  */

  discard_pending_stop_replies_in_queue (rs);



  if (remote_async_inferior_event_token)

    delete_async_event_handler (&remote_async_inferior_event_token);



  remote_notif_state_xfree (rs->notif_state);



  trace_reset_local_state ();

}



/* Query the remote side for the text, data and bss offsets.  */



static void

‌get_offsets (void)

{

  struct remote_state *rs = get_remote_state ();

  char *buf;

  char *ptr;

  int lose, num_segments = 0, do_sections, do_segments;

  CORE_ADDR text_addr, data_addr, bss_addr, segments[2];

  struct section_offsets *offs;

  struct symfile_segment_data *data;



  if (symfile_objfile == NULL)

    return;



  putpkt ("qOffsets");

  getpkt (&rs->buf, &rs->buf_size, 0);

  buf = rs->buf;



  if (buf[0] == '\000')

    return;                        /* Return silently.  Stub doesn't support

                                   this command.  */

  if (buf[0] == 'E')

    {

      warning (_("Remote failure reply: %s"), buf);

      return;

    }



  /* Pick up each field in turn.  This used to be done with scanf, but

     scanf will make trouble if CORE_ADDR size doesn't match

     conversion directives correctly.  The following code will work

     with any size of CORE_ADDR.  */

  text_addr = data_addr = bss_addr = 0;

  ptr = buf;

  lose = 0;



  if (strncmp (ptr, "Text=", 5) == 0)

    {

      ptr += 5;

      /* Don't use strtol, could lose on big values.  */

      while (*ptr && *ptr != ';')

        text_addr = (text_addr << 4) + fromhex (*ptr++);



      if (strncmp (ptr, ";Data=", 6) == 0)

        {

          ptr += 6;

          while (*ptr && *ptr != ';')

            data_addr = (data_addr << 4) + fromhex (*ptr++);

        }

      else

        lose = 1;



      if (!lose && strncmp (ptr, ";Bss=", 5) == 0)

        {

          ptr += 5;

          while (*ptr && *ptr != ';')

            bss_addr = (bss_addr << 4) + fromhex (*ptr++);



          if (bss_addr != data_addr)

            warning (_("Target reported unsupported offsets: %s"), buf);

        }

      else

        lose = 1;

    }

  else if (strncmp (ptr, "TextSeg=", 8) == 0)

    {

      ptr += 8;

      /* Don't use strtol, could lose on big values.  */

      while (*ptr && *ptr != ';')

        text_addr = (text_addr << 4) + fromhex (*ptr++);

      num_segments = 1;



      if (strncmp (ptr, ";DataSeg=", 9) == 0)

        {

          ptr += 9;

          while (*ptr && *ptr != ';')

            data_addr = (data_addr << 4) + fromhex (*ptr++);

          num_segments++;

        }

    }

  else

    lose = 1;



  if (lose)

    error (_("Malformed response to offset query, %s"), buf);

  else if (*ptr != '\0')

    warning (_("Target reported unsupported offsets: %s"), buf);



  offs = ((struct section_offsets *)

          alloca (SIZEOF_N_SECTION_OFFSETS (symfile_objfile->num_sections)));

  memcpy (offs, symfile_objfile->section_offsets,

          SIZEOF_N_SECTION_OFFSETS (symfile_objfile->num_sections));



  data = get_symfile_segment_data (symfile_objfile->obfd);

  do_segments = (data != NULL);

  do_sections = num_segments == 0;



  if (num_segments > 0)

    {

      segments[0] = text_addr;

      segments[1] = data_addr;

    }

  /* If we have two segments, we can still try to relocate everything

     by assuming that the .text and .data offsets apply to the whole

     text and data segments.  Convert the offsets given in the packet

     to base addresses for symfile_map_offsets_to_segments.  */

  else if (data && data->num_segments == 2)

    {

      segments[0] = data->segment_bases[0] + text_addr;

      segments[1] = data->segment_bases[1] + data_addr;

      num_segments = 2;

    }

  /* If the object file has only one segment, assume that it is text

     rather than data; main programs with no writable data are rare,

     but programs with no code are useless.  Of course the code might

     have ended up in the data segment... to detect that we would need

     the permissions here.  */

  else if (data && data->num_segments == 1)

    {

      segments[0] = data->segment_bases[0] + text_addr;

      num_segments = 1;

    }

  /* There's no way to relocate by segment.  */

  else

    do_segments = 0;



  if (do_segments)

    {

      int ret = symfile_map_offsets_to_segments (symfile_objfile->obfd, data,

                                                 offs, num_segments, segments);



      if (ret == 0 && !do_sections)

        error (_("Can not handle qOffsets TextSeg "

                 "response with this symbol file"));



      if (ret > 0)

        do_sections = 0;

    }



  if (data)

    free_symfile_segment_data (data);



  if (do_sections)

    {

      offs->offsets[SECT_OFF_TEXT (symfile_objfile)] = text_addr;



      /* This is a temporary kludge to force data and bss to use the

         same offsets because that's what nlmconv does now.  The real

         solution requires changes to the stub and remote.c that I

         don't have time to do right now.  */



      offs->offsets[SECT_OFF_DATA (symfile_objfile)] = data_addr;

      offs->offsets[SECT_OFF_BSS (symfile_objfile)] = data_addr;

    }



  objfile_relocate (symfile_objfile, offs);

}



/* Callback for iterate_over_threads.  Set the STOP_REQUESTED flags in

   threads we know are stopped already.  This is used during the

   initial remote connection in non-stop mode --- threads that are

   reported as already being stopped are left stopped.  */



static int

‌set_stop_requested_callback (struct thread_info *thread, void *data)

{

  /* If we have a stop reply for this thread, it must be stopped.  */

  if (peek_stop_reply (thread->ptid))

    set_stop_requested (thread->ptid, 1);



  return 0;

}



/* Send interrupt_sequence to remote target.  */

static void

‌send_interrupt_sequence (void)

{

  struct remote_state *rs = get_remote_state ();



  if (interrupt_sequence_mode == interrupt_sequence_control_c)

    remote_serial_write ("\x03", 1);

  else if (interrupt_sequence_mode == interrupt_sequence_break)

    serial_send_break (rs->remote_desc);

  else if (interrupt_sequence_mode == interrupt_sequence_break_g)

    {

      serial_send_break (rs->remote_desc);

      remote_serial_write ("g", 1);

    }

  else

    internal_error (__FILE__, __LINE__,

                    _("Invalid value for interrupt_sequence_mode: %s."),

                    interrupt_sequence_mode);

}





/* If STOP_REPLY is a T stop reply, look for the "thread" register,

   and extract the PTID.  Returns NULL_PTID if not found.  */



static ptid_t

‌stop_reply_extract_thread (char *stop_reply)

{

  if (stop_reply[0] == 'T' && strlen (stop_reply) > 3)

    {

      char *p;



      /* Txx r:val ; r:val (...)  */

      p = &stop_reply[3];



      /* Look for "register" named "thread".  */

      while (*p != '\0')

        {

          char *p1;



          p1 = strchr (p, ':');

          if (p1 == NULL)

            return null_ptid;



          if (strncmp (p, "thread", p1 - p) == 0)

            return read_ptid (++p1, &p);



          p1 = strchr (p, ';');

          if (p1 == NULL)

            return null_ptid;

          p1++;



          p = p1;

        }

    }



  return null_ptid;

}



/* Determine the remote side's current thread.  If we have a stop

   reply handy (in WAIT_STATUS), maybe it's a T stop reply with a

   "thread" register we can extract the current thread from.  If not,

   ask the remote which is the current thread with qC.  The former

   method avoids a roundtrip.  */



static ptid_t

‌get_current_thread (char *wait_status)

{

  ptid_t ptid;



  /* Note we don't use remote_parse_stop_reply as that makes use of

     the target architecture, which we haven't yet fully determined at

     this point.  */

  if (wait_status != NULL)

    ptid = stop_reply_extract_thread (wait_status);

  if (ptid_equal (ptid, null_ptid))

    ptid = remote_current_thread (inferior_ptid);



  return ptid;

}



/* Query the remote target for which is the current thread/process,

   add it to our tables, and update INFERIOR_PTID.  The caller is

   responsible for setting the state such that the remote end is ready

   to return the current thread.



   This function is called after handling the '?' or 'vRun' packets,

   whose response is a stop reply from which we can also try

   extracting the thread.  If the target doesn't support the explicit

   qC query, we infer the current thread from that stop reply, passed

   in in WAIT_STATUS, which may be NULL.  */



static void

‌add_current_inferior_and_thread (char *wait_status)

{

  struct remote_state *rs = get_remote_state ();

  int fake_pid_p = 0;

  ptid_t ptid = null_ptid;



  inferior_ptid = null_ptid;



  /* Now, if we have thread information, update inferior_ptid.  */

  ptid = get_current_thread (wait_status);



  if (!ptid_equal (ptid, null_ptid))

    {

      if (!remote_multi_process_p (rs))

        fake_pid_p = 1;



      inferior_ptid = ptid;

    }

  else

    {

      /* Without this, some commands which require an active target

         (such as kill) won't work.  This variable serves (at least)

         double duty as both the pid of the target process (if it has

         such), and as a flag indicating that a target is active.  */

      inferior_ptid = magic_null_ptid;

      fake_pid_p = 1;

    }



  remote_add_inferior (fake_pid_p, ptid_get_pid (inferior_ptid), -1);



  /* Add the main thread.  */

  add_thread_silent (inferior_ptid);

}



static void

‌remote_start_remote (int from_tty, struct target_ops *target, int extended_p)

{

  struct remote_state *rs = get_remote_state ();

  struct packet_config *noack_config;

  char *wait_status = NULL;



  immediate_quit++;                /* Allow user to interrupt it.  */

  QUIT;



  if (interrupt_on_connect)

    send_interrupt_sequence ();



  /* Ack any packet which the remote side has already sent.  */

  serial_write (rs->remote_desc, "+", 1);



  /* Signal other parts that we're going through the initial setup,

     and so things may not be stable yet.  */

  rs->starting_up = 1;



  /* The first packet we send to the target is the optional "supported

     packets" request.  If the target can answer this, it will tell us

     which later probes to skip.  */

  remote_query_supported ();



  /* If the stub wants to get a QAllow, compose one and send it.  */

  if (packet_support (PACKET_QAllow) != PACKET_DISABLE)

    remote_set_permissions (target);



  /* Next, we possibly activate noack mode.



     If the QStartNoAckMode packet configuration is set to AUTO,

     enable noack mode if the stub reported a wish for it with

     qSupported.



     If set to TRUE, then enable noack mode even if the stub didn't

     report it in qSupported.  If the stub doesn't reply OK, the

     session ends with an error.



     If FALSE, then don't activate noack mode, regardless of what the

     stub claimed should be the default with qSupported.  */



  noack_config = &remote_protocol_packets[PACKET_QStartNoAckMode];

  if (packet_config_support (noack_config) != PACKET_DISABLE)

    {

      putpkt ("QStartNoAckMode");

      getpkt (&rs->buf, &rs->buf_size, 0);

      if (packet_ok (rs->buf, noack_config) == PACKET_OK)

        rs->noack_mode = 1;

    }



  if (extended_p)

    {

      /* Tell the remote that we are using the extended protocol.  */

      putpkt ("!");

      getpkt (&rs->buf, &rs->buf_size, 0);

    }



  /* Let the target know which signals it is allowed to pass down to

     the program.  */

  update_signals_program_target ();



  /* Next, if the target can specify a description, read it.  We do

     this before anything involving memory or registers.  */

  target_find_description ();



  /* Next, now that we know something about the target, update the

     address spaces in the program spaces.  */

  update_address_spaces ();



  /* On OSs where the list of libraries is global to all

     processes, we fetch them early.  */

  if (gdbarch_has_global_solist (target_gdbarch ()))

    solib_add (NULL, from_tty, target, auto_solib_add);



  if (non_stop)

    {

      if (packet_support (PACKET_QNonStop) != PACKET_ENABLE)

        error (_("Non-stop mode requested, but remote "

                 "does not support non-stop"));



      putpkt ("QNonStop:1");

      getpkt (&rs->buf, &rs->buf_size, 0);



      if (strcmp (rs->buf, "OK") != 0)

        error (_("Remote refused setting non-stop mode with: %s"), rs->buf);



      /* Find about threads and processes the stub is already

         controlling.  We default to adding them in the running state.

         The '?' query below will then tell us about which threads are

         stopped.  */

      remote_update_thread_list (target);

    }

  else if (packet_support (PACKET_QNonStop) == PACKET_ENABLE)

    {

      /* Don't assume that the stub can operate in all-stop mode.

         Request it explicitly.  */

      putpkt ("QNonStop:0");

      getpkt (&rs->buf, &rs->buf_size, 0);



      if (strcmp (rs->buf, "OK") != 0)

        error (_("Remote refused setting all-stop mode with: %s"), rs->buf);

    }



  /* Upload TSVs regardless of whether the target is running or not.  The

     remote stub, such as GDBserver, may have some predefined or builtin

     TSVs, even if the target is not running.  */

  if (remote_get_trace_status (target, current_trace_status ()) != -1)

    {

      struct uploaded_tsv *uploaded_tsvs = NULL;



      remote_upload_trace_state_variables (target, &uploaded_tsvs);

      merge_uploaded_trace_state_variables (&uploaded_tsvs);

    }



  /* Check whether the target is running now.  */

  putpkt ("?");

  getpkt (&rs->buf, &rs->buf_size, 0);



  if (!non_stop)

    {

      ptid_t ptid;

      int fake_pid_p = 0;

      struct inferior *inf;



      if (rs->buf[0] == 'W' || rs->buf[0] == 'X')

        {

          if (!extended_p)

            error (_("The target is not running (try extended-remote?)"));



          /* We're connected, but not running.  Drop out before we

             call start_remote.  */

          rs->starting_up = 0;

          return;

        }

      else

        {

          /* Save the reply for later.  */

          wait_status = alloca (strlen (rs->buf) + 1);

          strcpy (wait_status, rs->buf);

        }



      /* Fetch thread list.  */

      target_update_thread_list ();



      /* Let the stub know that we want it to return the thread.  */

      set_continue_thread (minus_one_ptid);



      if (thread_count () == 0)

        {

          /* Target has no concept of threads at all.  GDB treats

             non-threaded target as single-threaded; add a main

             thread.  */

          add_current_inferior_and_thread (wait_status);

        }

      else

        {

          /* We have thread information; select the thread the target

             says should be current.  If we're reconnecting to a

             multi-threaded program, this will ideally be the thread

             that last reported an event before GDB disconnected.  */

          inferior_ptid = get_current_thread (wait_status);

          if (ptid_equal (inferior_ptid, null_ptid))

            {

              /* Odd... The target was able to list threads, but not

                 tell us which thread was current (no "thread"

                 register in T stop reply?).  Just pick the first

                 thread in the thread list then.  */

              inferior_ptid = thread_list->ptid;

            }

        }



      /* init_wait_for_inferior should be called before get_offsets in order

         to manage `inserted' flag in bp loc in a correct state.

         breakpoint_init_inferior, called from init_wait_for_inferior, set

         `inserted' flag to 0, while before breakpoint_re_set, called from

         start_remote, set `inserted' flag to 1.  In the initialization of

         inferior, breakpoint_init_inferior should be called first, and then

         breakpoint_re_set can be called.  If this order is broken, state of

         `inserted' flag is wrong, and cause some problems on breakpoint

         manipulation.  */

      init_wait_for_inferior ();



      get_offsets ();                /* Get text, data & bss offsets.  */



      /* If we could not find a description using qXfer, and we know

         how to do it some other way, try again.  This is not

         supported for non-stop; it could be, but it is tricky if

         there are no stopped threads when we connect.  */

      if (remote_read_description_p (target)

          && gdbarch_target_desc (target_gdbarch ()) == NULL)

        {

          target_clear_description ();

          target_find_description ();

        }



      /* Use the previously fetched status.  */

      gdb_assert (wait_status != NULL);

      strcpy (rs->buf, wait_status);

      rs->cached_wait_status = 1;



      immediate_quit--;

      start_remote (from_tty); /* Initialize gdb process mechanisms.  */

    }

  else

    {

      /* Clear WFI global state.  Do this before finding about new

         threads and inferiors, and setting the current inferior.

         Otherwise we would clear the proceed status of the current

         inferior when we want its stop_soon state to be preserved

         (see notice_new_inferior).  */

      init_wait_for_inferior ();



      /* In non-stop, we will either get an "OK", meaning that there

         are no stopped threads at this time; or, a regular stop

         reply.  In the latter case, there may be more than one thread

         stopped --- we pull them all out using the vStopped

         mechanism.  */

      if (strcmp (rs->buf, "OK") != 0)

        {

          struct notif_client *notif = &notif_client_stop;



          /* remote_notif_get_pending_replies acks this one, and gets

             the rest out.  */

          rs->notif_state->pending_event[notif_client_stop.id]

            = remote_notif_parse (notif, rs->buf);

          remote_notif_get_pending_events (notif);



          /* Make sure that threads that were stopped remain

             stopped.  */

          iterate_over_threads (set_stop_requested_callback, NULL);

        }



      if (target_can_async_p ())

        target_async (inferior_event_handler, 0);



      if (thread_count () == 0)

        {

          if (!extended_p)

            error (_("The target is not running (try extended-remote?)"));



          /* We're connected, but not running.  Drop out before we

             call start_remote.  */

          rs->starting_up = 0;

          return;

        }



      /* Let the stub know that we want it to return the thread.  */



      /* Force the stub to choose a thread.  */

      set_general_thread (null_ptid);



      /* Query it.  */

      inferior_ptid = remote_current_thread (minus_one_ptid);

      if (ptid_equal (inferior_ptid, minus_one_ptid))

        error (_("remote didn't report the current thread in non-stop mode"));



      get_offsets ();                /* Get text, data & bss offsets.  */



      /* In non-stop mode, any cached wait status will be stored in

         the stop reply queue.  */

      gdb_assert (wait_status == NULL);



      /* Report all signals during attach/startup.  */

      remote_pass_signals (target, 0, NULL);

    }



  /* If we connected to a live target, do some additional setup.  */

  if (target_has_execution)

    {

      if (symfile_objfile)         /* No use without a symbol-file.  */

        remote_check_symbols ();

    }



  /* Possibly the target has been engaged in a trace run started

     previously; find out where things are at.  */

  if (remote_get_trace_status (target, current_trace_status ()) != -1)

    {

      struct uploaded_tp *uploaded_tps = NULL;



      if (current_trace_status ()->running)

        printf_filtered (_("Trace is already running on the target.\n"));



      remote_upload_tracepoints (target, &uploaded_tps);



      merge_uploaded_tracepoints (&uploaded_tps);

    }



  /* The thread and inferior lists are now synchronized with the

     target, our symbols have been relocated, and we're merged the

     target's tracepoints with ours.  We're done with basic start

     up.  */

  rs->starting_up = 0;



  /* Maybe breakpoints are global and need to be inserted now.  */

  if (breakpoints_should_be_inserted_now ())

    insert_breakpoints ();

}



/* Open a connection to a remote debugger.

   NAME is the filename used for communication.  */



static void

‌remote_open (const char *name, int from_tty)

{

  remote_open_1 (name, from_tty, &remote_ops, 0);

}



/* Open a connection to a remote debugger using the extended

   remote gdb protocol.  NAME is the filename used for communication.  */



static void

‌extended_remote_open (const char *name, int from_tty)

{

  remote_open_1 (name, from_tty, &extended_remote_ops, 1 /*extended_p */);

}



/* Reset all packets back to "unknown support".  Called when opening a

   new connection to a remote target.  */



static void

‌reset_all_packet_configs_support (void)

{

  int i;



  for (i = 0; i < PACKET_MAX; i++)

    remote_protocol_packets[i].support = PACKET_SUPPORT_UNKNOWN;

}



/* Initialize all packet configs.  */



static void

‌init_all_packet_configs (void)

{

  int i;



  for (i = 0; i < PACKET_MAX; i++)

    {

      remote_protocol_packets[i].detect = AUTO_BOOLEAN_AUTO;

      remote_protocol_packets[i].support = PACKET_SUPPORT_UNKNOWN;

    }

}



/* Symbol look-up.  */



static void

‌remote_check_symbols (void)

{

  struct remote_state *rs = get_remote_state ();

  char *msg, *reply, *tmp;

  struct bound_minimal_symbol sym;

  int end;



  /* The remote side has no concept of inferiors that aren't running

     yet, it only knows about running processes.  If we're connected

     but our current inferior is not running, we should not invite the

     remote target to request symbol lookups related to its

     (unrelated) current process.  */

  if (!target_has_execution)

    return;



  if (packet_support (PACKET_qSymbol) == PACKET_DISABLE)

    return;



  /* Make sure the remote is pointing at the right process.  Note

     there's no way to select "no process".  */

  set_general_process ();



  /* Allocate a message buffer.  We can't reuse the input buffer in RS,

     because we need both at the same time.  */

  msg = alloca (get_remote_packet_size ());



  /* Invite target to request symbol lookups.  */



  putpkt ("qSymbol::");

  getpkt (&rs->buf, &rs->buf_size, 0);

  packet_ok (rs->buf, &remote_protocol_packets[PACKET_qSymbol]);

  reply = rs->buf;



  while (strncmp (reply, "qSymbol:", 8) == 0)

    {

      struct bound_minimal_symbol sym;



      tmp = &reply[8];

      end = hex2bin (tmp, (gdb_byte *) msg, strlen (tmp) / 2);

      msg[end] = '\0';

      sym = lookup_minimal_symbol (msg, NULL, NULL);

      if (sym.minsym == NULL)

        xsnprintf (msg, get_remote_packet_size (), "qSymbol::%s", &reply[8]);

      else

        {

          int addr_size = gdbarch_addr_bit (target_gdbarch ()) / 8;

          CORE_ADDR sym_addr = BMSYMBOL_VALUE_ADDRESS (sym);



          /* If this is a function address, return the start of code

             instead of any data function descriptor.  */

          sym_addr = gdbarch_convert_from_func_ptr_addr (target_gdbarch (),

                                                         sym_addr,

                                                         &current_target);



          xsnprintf (msg, get_remote_packet_size (), "qSymbol:%s:%s",

                     phex_nz (sym_addr, addr_size), &reply[8]);

        }



      putpkt (msg);

      getpkt (&rs->buf, &rs->buf_size, 0);

      reply = rs->buf;

    }

}



static struct serial *

‌remote_serial_open (const char *name)

{

  static int udp_warning = 0;



  /* FIXME: Parsing NAME here is a hack.  But we want to warn here instead

     of in ser-tcp.c, because it is the remote protocol assuming that the

     serial connection is reliable and not the serial connection promising

     to be.  */

  if (!udp_warning && strncmp (name, "udp:", 4) == 0)

    {

      warning (_("The remote protocol may be unreliable over UDP.\n"

                 "Some events may be lost, rendering further debugging "

                 "impossible."));

      udp_warning = 1;

    }



  return serial_open (name);

}



/* Inform the target of our permission settings.  The permission flags

   work without this, but if the target knows the settings, it can do

   a couple things.  First, it can add its own check, to catch cases

   that somehow manage to get by the permissions checks in target

   methods.  Second, if the target is wired to disallow particular

   settings (for instance, a system in the field that is not set up to

   be able to stop at a breakpoint), it can object to any unavailable

   permissions.  */



void

‌remote_set_permissions (struct target_ops *self)

{

  struct remote_state *rs = get_remote_state ();



  xsnprintf (rs->buf, get_remote_packet_size (), "QAllow:"

             "WriteReg:%x;WriteMem:%x;"

             "InsertBreak:%x;InsertTrace:%x;"

             "InsertFastTrace:%x;Stop:%x",

             may_write_registers, may_write_memory,

             may_insert_breakpoints, may_insert_tracepoints,

             may_insert_fast_tracepoints, may_stop);

  putpkt (rs->buf);

  getpkt (&rs->buf, &rs->buf_size, 0);



  /* If the target didn't like the packet, warn the user.  Do not try

     to undo the user's settings, that would just be maddening.  */

  if (strcmp (rs->buf, "OK") != 0)

    warning (_("Remote refused setting permissions with: %s"), rs->buf);

}



/* This type describes each known response to the qSupported

   packet.  */

‌struct protocol_feature

{

  /* The name of this protocol feature.  */

  const char *name;



  /* The default for this protocol feature.  */

  enum packet_support default_support;



  /* The function to call when this feature is reported, or after

     qSupported processing if the feature is not supported.

     The first argument points to this structure.  The second

     argument indicates whether the packet requested support be

     enabled, disabled, or probed (or the default, if this function

     is being called at the end of processing and this feature was

     not reported).  The third argument may be NULL; if not NULL, it

     is a NUL-terminated string taken from the packet following

     this feature's name and an equals sign.  */

  void (*func) (const struct protocol_feature *, enum packet_support,

                const char *);



  /* The corresponding packet for this feature.  Only used if

     FUNC is remote_supported_packet.  */

  int packet;

};



static void

‌remote_supported_packet (const struct protocol_feature *feature,

                         enum packet_support support,

                         const char *argument)

{

  if (argument)

    {

      warning (_("Remote qSupported response supplied an unexpected value for"

                 " \"%s\"."), feature->name);

      return;

    }



  remote_protocol_packets[feature->packet].support = support;

}



static void

‌remote_packet_size (const struct protocol_feature *feature,

                    enum packet_support support, const char *value)

{

  struct remote_state *rs = get_remote_state ();



  int packet_size;

  char *value_end;



  if (support != PACKET_ENABLE)

    return;



  if (value == NULL || *value == '\0')

    {

      warning (_("Remote target reported \"%s\" without a size."),

               feature->name);

      return;

    }



  errno = 0;

  packet_size = strtol (value, &value_end, 16);

  if (errno != 0 || *value_end != '\0' || packet_size < 0)

    {

      warning (_("Remote target reported \"%s\" with a bad size: \"%s\"."),

               feature->name, value);

      return;

    }



  if (packet_size > MAX_REMOTE_PACKET_SIZE)

    {

      warning (_("limiting remote suggested packet size (%d bytes) to %d"),

               packet_size, MAX_REMOTE_PACKET_SIZE);

      packet_size = MAX_REMOTE_PACKET_SIZE;

    }



  /* Record the new maximum packet size.  */

  rs->explicit_packet_size = packet_size;

}



‌static const struct protocol_feature remote_protocol_features[] = {

  { "PacketSize", PACKET_DISABLE, remote_packet_size, -1 },

  { "qXfer:auxv:read", PACKET_DISABLE, remote_supported_packet,

    PACKET_qXfer_auxv },

  { "qXfer:features:read", PACKET_DISABLE, remote_supported_packet,

    PACKET_qXfer_features },

  { "qXfer:libraries:read", PACKET_DISABLE, remote_supported_packet,

    PACKET_qXfer_libraries },

  { "qXfer:libraries-svr4:read", PACKET_DISABLE, remote_supported_packet,

    PACKET_qXfer_libraries_svr4 },

  { "augmented-libraries-svr4-read", PACKET_DISABLE,

    remote_supported_packet, PACKET_augmented_libraries_svr4_read_feature },

  { "qXfer:memory-map:read", PACKET_DISABLE, remote_supported_packet,

    PACKET_qXfer_memory_map },

  { "qXfer:spu:read", PACKET_DISABLE, remote_supported_packet,

    PACKET_qXfer_spu_read },

  { "qXfer:spu:write", PACKET_DISABLE, remote_supported_packet,

    PACKET_qXfer_spu_write },

  { "qXfer:osdata:read", PACKET_DISABLE, remote_supported_packet,

    PACKET_qXfer_osdata },

  { "qXfer:threads:read", PACKET_DISABLE, remote_supported_packet,

    PACKET_qXfer_threads },

  { "qXfer:traceframe-info:read", PACKET_DISABLE, remote_supported_packet,

    PACKET_qXfer_traceframe_info },

  { "QPassSignals", PACKET_DISABLE, remote_supported_packet,

    PACKET_QPassSignals },

  { "QProgramSignals", PACKET_DISABLE, remote_supported_packet,

    PACKET_QProgramSignals },

  { "QStartNoAckMode", PACKET_DISABLE, remote_supported_packet,

    PACKET_QStartNoAckMode },

  { "multiprocess", PACKET_DISABLE, remote_supported_packet,

    PACKET_multiprocess_feature },

  { "QNonStop", PACKET_DISABLE, remote_supported_packet, PACKET_QNonStop },

  { "qXfer:siginfo:read", PACKET_DISABLE, remote_supported_packet,

    PACKET_qXfer_siginfo_read },

  { "qXfer:siginfo:write", PACKET_DISABLE, remote_supported_packet,

    PACKET_qXfer_siginfo_write },

  { "ConditionalTracepoints", PACKET_DISABLE, remote_supported_packet,

    PACKET_ConditionalTracepoints },

  { "ConditionalBreakpoints", PACKET_DISABLE, remote_supported_packet,

    PACKET_ConditionalBreakpoints },

  { "BreakpointCommands", PACKET_DISABLE, remote_supported_packet,

    PACKET_BreakpointCommands },

  { "FastTracepoints", PACKET_DISABLE, remote_supported_packet,

    PACKET_FastTracepoints },

  { "StaticTracepoints", PACKET_DISABLE, remote_supported_packet,

    PACKET_StaticTracepoints },

  {"InstallInTrace", PACKET_DISABLE, remote_supported_packet,

   PACKET_InstallInTrace},

  { "DisconnectedTracing", PACKET_DISABLE, remote_supported_packet,

    PACKET_DisconnectedTracing_feature },

  { "ReverseContinue", PACKET_DISABLE, remote_supported_packet,

    PACKET_bc },

  { "ReverseStep", PACKET_DISABLE, remote_supported_packet,

    PACKET_bs },

  { "TracepointSource", PACKET_DISABLE, remote_supported_packet,

    PACKET_TracepointSource },

  { "QAllow", PACKET_DISABLE, remote_supported_packet,

    PACKET_QAllow },

  { "EnableDisableTracepoints", PACKET_DISABLE, remote_supported_packet,

    PACKET_EnableDisableTracepoints_feature },

  { "qXfer:fdpic:read", PACKET_DISABLE, remote_supported_packet,

    PACKET_qXfer_fdpic },

  { "qXfer:uib:read", PACKET_DISABLE, remote_supported_packet,

    PACKET_qXfer_uib },

  { "QDisableRandomization", PACKET_DISABLE, remote_supported_packet,

    PACKET_QDisableRandomization },

  { "QAgent", PACKET_DISABLE, remote_supported_packet, PACKET_QAgent},

  { "QTBuffer:size", PACKET_DISABLE,

    remote_supported_packet, PACKET_QTBuffer_size},

  { "tracenz", PACKET_DISABLE, remote_supported_packet, PACKET_tracenz_feature },

  { "Qbtrace:off", PACKET_DISABLE, remote_supported_packet, PACKET_Qbtrace_off },

  { "Qbtrace:bts", PACKET_DISABLE, remote_supported_packet, PACKET_Qbtrace_bts },

  { "qXfer:btrace:read", PACKET_DISABLE, remote_supported_packet,

    PACKET_qXfer_btrace }

};



‌static char *remote_support_xml;



/* Register string appended to "xmlRegisters=" in qSupported query.  */



void

‌register_remote_support_xml (const char *xml)

{

#if defined(HAVE_LIBEXPAT)

  if (remote_support_xml == NULL)

    remote_support_xml = concat ("xmlRegisters=", xml, (char *) NULL);

  else

    {

      char *copy = xstrdup (remote_support_xml + 13);

      char *p = strtok (copy, ",");



      do

        {

          if (strcmp (p, xml) == 0)

            {

              /* already there */

              xfree (copy);

              return;

            }

        }

      while ((p = strtok (NULL, ",")) != NULL);

      xfree (copy);



      remote_support_xml = reconcat (remote_support_xml,

                                     remote_support_xml, ",", xml,

                                     (char *) NULL);

    }

#endif

}



static char *

‌remote_query_supported_append (char *msg, const char *append)

{

  if (msg)

    return reconcat (msg, msg, ";", append, (char *) NULL);

  else

    return xstrdup (append);

}



static void

‌remote_query_supported (void)

{

  struct remote_state *rs = get_remote_state ();

  char *next;

  int i;

  unsigned char seen [ARRAY_SIZE (remote_protocol_features)];



  /* The packet support flags are handled differently for this packet

     than for most others.  We treat an error, a disabled packet, and

     an empty response identically: any features which must be reported

     to be used will be automatically disabled.  An empty buffer

     accomplishes this, since that is also the representation for a list

     containing no features.  */



  rs->buf[0] = 0;

  if (packet_support (PACKET_qSupported) != PACKET_DISABLE)

    {

      char *q = NULL;

      struct cleanup *old_chain = make_cleanup (free_current_contents, &q);



      q = remote_query_supported_append (q, "multiprocess+");



      if (remote_support_xml)

        q = remote_query_supported_append (q, remote_support_xml);



      q = remote_query_supported_append (q, "qRelocInsn+");



      q = reconcat (q, "qSupported:", q, (char *) NULL);

      putpkt (q);



      do_cleanups (old_chain);



      getpkt (&rs->buf, &rs->buf_size, 0);



      /* If an error occured, warn, but do not return - just reset the

         buffer to empty and go on to disable features.  */

      if (packet_ok (rs->buf, &remote_protocol_packets[PACKET_qSupported])

          == PACKET_ERROR)

        {

          warning (_("Remote failure reply: %s"), rs->buf);

          rs->buf[0] = 0;

        }

    }



  memset (seen, 0, sizeof (seen));



  next = rs->buf;

  while (*next)

    {

      enum packet_support is_supported;

      char *p, *end, *name_end, *value;



      /* First separate out this item from the rest of the packet.  If

         there's another item after this, we overwrite the separator

         (terminated strings are much easier to work with).  */

      p = next;

      end = strchr (p, ';');

      if (end == NULL)

        {

          end = p + strlen (p);

          next = end;

        }

      else

        {

          *end = '\0';

          next = end + 1;



          if (end == p)

            {

              warning (_("empty item in \"qSupported\" response"));

              continue;

            }

        }



      name_end = strchr (p, '=');

      if (name_end)

        {

          /* This is a name=value entry.  */

          is_supported = PACKET_ENABLE;

          value = name_end + 1;

          *name_end = '\0';

        }

      else

        {

          value = NULL;

          switch (end[-1])

            {

            case '+':

              is_supported = PACKET_ENABLE;

              break;



            case '-':

              is_supported = PACKET_DISABLE;

              break;



            case '?':

              is_supported = PACKET_SUPPORT_UNKNOWN;

              break;



            default:

              warning (_("unrecognized item \"%s\" "

                         "in \"qSupported\" response"), p);

              continue;

            }

          end[-1] = '\0';

        }



      for (i = 0; i < ARRAY_SIZE (remote_protocol_features); i++)

        if (strcmp (remote_protocol_features[i].name, p) == 0)

          {

            const struct protocol_feature *feature;



            seen[i] = 1;

            feature = &remote_protocol_features[i];

            feature->func (feature, is_supported, value);

            break;

          }

    }



  /* If we increased the packet size, make sure to increase the global

     buffer size also.  We delay this until after parsing the entire

     qSupported packet, because this is the same buffer we were

     parsing.  */

  if (rs->buf_size < rs->explicit_packet_size)

    {

      rs->buf_size = rs->explicit_packet_size;

      rs->buf = xrealloc (rs->buf, rs->buf_size);

    }



  /* Handle the defaults for unmentioned features.  */

  for (i = 0; i < ARRAY_SIZE (remote_protocol_features); i++)

    if (!seen[i])

      {

        const struct protocol_feature *feature;



        feature = &remote_protocol_features[i];

        feature->func (feature, feature->default_support, NULL);

      }

}



/* Remove any of the remote.c targets from target stack.  Upper targets depend

   on it so remove them first.  */



static void

‌remote_unpush_target (void)

{

  pop_all_targets_above (process_stratum - 1);

}



static void

‌remote_open_1 (const char *name, int from_tty,

               struct target_ops *target, int extended_p)

{

  struct remote_state *rs = get_remote_state ();



  if (name == 0)

    error (_("To open a remote debug connection, you need to specify what\n"

           "serial device is attached to the remote system\n"

           "(e.g. /dev/ttyS0, /dev/ttya, COM1, etc.)."));



  /* See FIXME above.  */

  if (!target_async_permitted)

    wait_forever_enabled_p = 1;



  /* If we're connected to a running target, target_preopen will kill it.

     Ask this question first, before target_preopen has a chance to kill

     anything.  */

  if (rs->remote_desc != NULL && !have_inferiors ())

    {

      if (from_tty

          && !query (_("Already connected to a remote target.  Disconnect? ")))

        error (_("Still connected."));

    }



  /* Here the possibly existing remote target gets unpushed.  */

  target_preopen (from_tty);



  /* Make sure we send the passed signals list the next time we resume.  */

  xfree (rs->last_pass_packet);

  rs->last_pass_packet = NULL;



  /* Make sure we send the program signals list the next time we

     resume.  */

  xfree (rs->last_program_signals_packet);

  rs->last_program_signals_packet = NULL;



  remote_fileio_reset ();

  reopen_exec_file ();

  reread_symbols ();



  rs->remote_desc = remote_serial_open (name);

  if (!rs->remote_desc)

    perror_with_name (name);



  if (baud_rate != -1)

    {

      if (serial_setbaudrate (rs->remote_desc, baud_rate))

        {

          /* The requested speed could not be set.  Error out to

             top level after closing remote_desc.  Take care to

             set remote_desc to NULL to avoid closing remote_desc

             more than once.  */

          serial_close (rs->remote_desc);

          rs->remote_desc = NULL;

          perror_with_name (name);

        }

    }



  serial_raw (rs->remote_desc);



  /* If there is something sitting in the buffer we might take it as a

     response to a command, which would be bad.  */

  serial_flush_input (rs->remote_desc);



  if (from_tty)

    {

      puts_filtered ("Remote debugging using ");

      puts_filtered (name);

      puts_filtered ("\n");

    }

  push_target (target);                /* Switch to using remote target now.  */



  /* Register extra event sources in the event loop.  */

  remote_async_inferior_event_token

    = create_async_event_handler (remote_async_inferior_event_handler,

                                  NULL);

  rs->notif_state = remote_notif_state_allocate ();



  /* Reset the target state; these things will be queried either by

     remote_query_supported or as they are needed.  */

  reset_all_packet_configs_support ();

  rs->cached_wait_status = 0;

  rs->explicit_packet_size = 0;

  rs->noack_mode = 0;

  rs->extended = extended_p;

  rs->waiting_for_stop_reply = 0;

  rs->ctrlc_pending_p = 0;



  rs->general_thread = not_sent_ptid;

  rs->continue_thread = not_sent_ptid;

  rs->remote_traceframe_number = -1;



  /* Probe for ability to use "ThreadInfo" query, as required.  */

  rs->use_threadinfo_query = 1;

  rs->use_threadextra_query = 1;



  if (target_async_permitted)

    {

      /* With this target we start out by owning the terminal.  */

      remote_async_terminal_ours_p = 1;



      /* FIXME: cagney/1999-09-23: During the initial connection it is

         assumed that the target is already ready and able to respond to

         requests.  Unfortunately remote_start_remote() eventually calls

         wait_for_inferior() with no timeout.  wait_forever_enabled_p gets

         around this.  Eventually a mechanism that allows

         wait_for_inferior() to expect/get timeouts will be

         implemented.  */

      wait_forever_enabled_p = 0;

    }



  /* First delete any symbols previously loaded from shared libraries.  */

  no_shared_libraries (NULL, 0);



  /* Start afresh.  */

  init_thread_list ();



  /* Start the remote connection.  If error() or QUIT, discard this

     target (we'd otherwise be in an inconsistent state) and then

     propogate the error on up the exception chain.  This ensures that

     the caller doesn't stumble along blindly assuming that the

     function succeeded.  The CLI doesn't have this problem but other

     UI's, such as MI do.



     FIXME: cagney/2002-05-19: Instead of re-throwing the exception,

     this function should return an error indication letting the

     caller restore the previous state.  Unfortunately the command

     ``target remote'' is directly wired to this function making that

     impossible.  On a positive note, the CLI side of this problem has

     been fixed - the function set_cmd_context() makes it possible for

     all the ``target ....'' commands to share a common callback

     function.  See cli-dump.c.  */

  {

    volatile struct gdb_exception ex;



    TRY_CATCH (ex, RETURN_MASK_ALL)

      {

        remote_start_remote (from_tty, target, extended_p);

      }

    if (ex.reason < 0)

      {

        /* Pop the partially set up target - unless something else did

           already before throwing the exception.  */

        if (rs->remote_desc != NULL)

          remote_unpush_target ();

        if (target_async_permitted)

          wait_forever_enabled_p = 1;

        throw_exception (ex);

      }

  }



  if (target_async_permitted)

    wait_forever_enabled_p = 1;

}



/* This takes a program previously attached to and detaches it.  After

   this is done, GDB can be used to debug some other program.  We

   better not have left any breakpoints in the target program or it'll

   die when it hits one.  */



static void

‌remote_detach_1 (const char *args, int from_tty, int extended)

{

  int pid = ptid_get_pid (inferior_ptid);

  struct remote_state *rs = get_remote_state ();



  if (args)

    error (_("Argument given to \"detach\" when remotely debugging."));



  if (!target_has_execution)

    error (_("No process to detach from."));



  if (from_tty)

    {

      char *exec_file = get_exec_file (0);

      if (exec_file == NULL)

        exec_file = "";

      printf_unfiltered (_("Detaching from program: %s, %s\n"), exec_file,

                         target_pid_to_str (pid_to_ptid (pid)));

      gdb_flush (gdb_stdout);

    }



  /* Tell the remote target to detach.  */

  if (remote_multi_process_p (rs))

    xsnprintf (rs->buf, get_remote_packet_size (), "D;%x", pid);

  else

    strcpy (rs->buf, "D");



  putpkt (rs->buf);

  getpkt (&rs->buf, &rs->buf_size, 0);



  if (rs->buf[0] == 'O' && rs->buf[1] == 'K')

    ;

  else if (rs->buf[0] == '\0')

    error (_("Remote doesn't know how to detach"));

  else

    error (_("Can't detach process."));



  if (from_tty && !extended)

    puts_filtered (_("Ending remote debugging.\n"));



  target_mourn_inferior ();

}



static void

‌remote_detach (struct target_ops *ops, const char *args, int from_tty)

{

  remote_detach_1 (args, from_tty, 0);

}



static void

‌extended_remote_detach (struct target_ops *ops, const char *args, int from_tty)

{

  remote_detach_1 (args, from_tty, 1);

}



/* Same as remote_detach, but don't send the "D" packet; just disconnect.  */



static void

‌remote_disconnect (struct target_ops *target, const char *args, int from_tty)

{

  if (args)

    error (_("Argument given to \"disconnect\" when remotely debugging."));



  /* Make sure we unpush even the extended remote targets; mourn

     won't do it.  So call remote_mourn_1 directly instead of

     target_mourn_inferior.  */

  remote_mourn_1 (target);



  if (from_tty)

    puts_filtered ("Ending remote debugging.\n");

}



/* Attach to the process specified by ARGS.  If FROM_TTY is non-zero,

   be chatty about it.  */



static void

‌extended_remote_attach_1 (struct target_ops *target, const char *args,

                          int from_tty)

{

  struct remote_state *rs = get_remote_state ();

  int pid;

  char *wait_status = NULL;



  pid = parse_pid_to_attach (args);



  /* Remote PID can be freely equal to getpid, do not check it here the same

     way as in other targets.  */



  if (packet_support (PACKET_vAttach) == PACKET_DISABLE)

    error (_("This target does not support attaching to a process"));



  if (from_tty)

    {

      char *exec_file = get_exec_file (0);



      if (exec_file)

        printf_unfiltered (_("Attaching to program: %s, %s\n"), exec_file,

                           target_pid_to_str (pid_to_ptid (pid)));

      else

        printf_unfiltered (_("Attaching to %s\n"),

                           target_pid_to_str (pid_to_ptid (pid)));



      gdb_flush (gdb_stdout);

    }



  xsnprintf (rs->buf, get_remote_packet_size (), "vAttach;%x", pid);

  putpkt (rs->buf);

  getpkt (&rs->buf, &rs->buf_size, 0);



  switch (packet_ok (rs->buf,

                     &remote_protocol_packets[PACKET_vAttach]))

    {

    case PACKET_OK:

      if (!non_stop)

        {

          /* Save the reply for later.  */

          wait_status = alloca (strlen (rs->buf) + 1);

          strcpy (wait_status, rs->buf);

        }

      else if (strcmp (rs->buf, "OK") != 0)

        error (_("Attaching to %s failed with: %s"),

               target_pid_to_str (pid_to_ptid (pid)),

               rs->buf);

      break;

    case PACKET_UNKNOWN:

      error (_("This target does not support attaching to a process"));

    default:

      error (_("Attaching to %s failed"),

             target_pid_to_str (pid_to_ptid (pid)));

    }



  set_current_inferior (remote_add_inferior (0, pid, 1));



  inferior_ptid = pid_to_ptid (pid);



  if (non_stop)

    {

      struct thread_info *thread;



      /* Get list of threads.  */

      remote_update_thread_list (target);



      thread = first_thread_of_process (pid);

      if (thread)

        inferior_ptid = thread->ptid;

      else

        inferior_ptid = pid_to_ptid (pid);



      /* Invalidate our notion of the remote current thread.  */

      record_currthread (rs, minus_one_ptid);

    }

  else

    {

      /* Now, if we have thread information, update inferior_ptid.  */

      inferior_ptid = remote_current_thread (inferior_ptid);



      /* Add the main thread to the thread list.  */

      add_thread_silent (inferior_ptid);

    }



  /* Next, if the target can specify a description, read it.  We do

     this before anything involving memory or registers.  */

  target_find_description ();



  if (!non_stop)

    {

      /* Use the previously fetched status.  */

      gdb_assert (wait_status != NULL);



      if (target_can_async_p ())

        {

          struct notif_event *reply

            =  remote_notif_parse (&notif_client_stop, wait_status);



          push_stop_reply ((struct stop_reply *) reply);



          target_async (inferior_event_handler, 0);

        }

      else

        {

          gdb_assert (wait_status != NULL);

          strcpy (rs->buf, wait_status);

          rs->cached_wait_status = 1;

        }

    }

  else

    gdb_assert (wait_status == NULL);

}



static void

‌extended_remote_attach (struct target_ops *ops, const char *args, int from_tty)

{

  extended_remote_attach_1 (ops, args, from_tty);

}



/* Implementation of the to_post_attach method.  */



static void

‌extended_remote_post_attach (struct target_ops *ops, int pid)

{

  /* In certain cases GDB might not have had the chance to start

     symbol lookup up until now.  This could happen if the debugged

     binary is not using shared libraries, the vsyscall page is not

     present (on Linux) and the binary itself hadn't changed since the

     debugging process was started.  */

  if (symfile_objfile != NULL)

    remote_check_symbols();

}





/* Check for the availability of vCont.  This function should also check

   the response.  */



static void

‌remote_vcont_probe (struct remote_state *rs)

{

  char *buf;



  strcpy (rs->buf, "vCont?");

  putpkt (rs->buf);

  getpkt (&rs->buf, &rs->buf_size, 0);

  buf = rs->buf;



  /* Make sure that the features we assume are supported.  */

  if (strncmp (buf, "vCont", 5) == 0)

    {

      char *p = &buf[5];

      int support_s, support_S, support_c, support_C;



      support_s = 0;

      support_S = 0;

      support_c = 0;

      support_C = 0;

      rs->supports_vCont.t = 0;

      rs->supports_vCont.r = 0;

      while (p && *p == ';')

        {

          p++;

          if (*p == 's' && (*(p + 1) == ';' || *(p + 1) == 0))

            support_s = 1;

          else if (*p == 'S' && (*(p + 1) == ';' || *(p + 1) == 0))

            support_S = 1;

          else if (*p == 'c' && (*(p + 1) == ';' || *(p + 1) == 0))

            support_c = 1;

          else if (*p == 'C' && (*(p + 1) == ';' || *(p + 1) == 0))

            support_C = 1;

          else if (*p == 't' && (*(p + 1) == ';' || *(p + 1) == 0))

            rs->supports_vCont.t = 1;

          else if (*p == 'r' && (*(p + 1) == ';' || *(p + 1) == 0))

            rs->supports_vCont.r = 1;



          p = strchr (p, ';');

        }



      /* If s, S, c, and C are not all supported, we can't use vCont.  Clearing

         BUF will make packet_ok disable the packet.  */

      if (!support_s || !support_S || !support_c || !support_C)

        buf[0] = 0;

    }



  packet_ok (buf, &remote_protocol_packets[PACKET_vCont]);

}



/* Helper function for building "vCont" resumptions.  Write a

   resumption to P.  ENDP points to one-passed-the-end of the buffer

   we're allowed to write to.  Returns BUF+CHARACTERS_WRITTEN.  The

   thread to be resumed is PTID; STEP and SIGGNAL indicate whether the

   resumed thread should be single-stepped and/or signalled.  If PTID

   equals minus_one_ptid, then all threads are resumed; if PTID

   represents a process, then all threads of the process are resumed;

   the thread to be stepped and/or signalled is given in the global

   INFERIOR_PTID.  */



static char *

‌append_resumption (char *p, char *endp,

                   ptid_t ptid, int step, enum gdb_signal siggnal)

{

  struct remote_state *rs = get_remote_state ();



  if (step && siggnal != GDB_SIGNAL_0)

    p += xsnprintf (p, endp - p, ";S%02x", siggnal);

  else if (step

           /* GDB is willing to range step.  */

           && use_range_stepping

           /* Target supports range stepping.  */

           && rs->supports_vCont.r

           /* We don't currently support range stepping multiple

              threads with a wildcard (though the protocol allows it,

              so stubs shouldn't make an active effort to forbid

              it).  */

           && !(remote_multi_process_p (rs) && ptid_is_pid (ptid)))

    {

      struct thread_info *tp;



      if (ptid_equal (ptid, minus_one_ptid))

        {

          /* If we don't know about the target thread's tid, then

             we're resuming magic_null_ptid (see caller).  */

          tp = find_thread_ptid (magic_null_ptid);

        }

      else

        tp = find_thread_ptid (ptid);

      gdb_assert (tp != NULL);



      if (tp->control.may_range_step)

        {

          int addr_size = gdbarch_addr_bit (target_gdbarch ()) / 8;



          p += xsnprintf (p, endp - p, ";r%s,%s",

                          phex_nz (tp->control.step_range_start,

                                   addr_size),

                          phex_nz (tp->control.step_range_end,

                                   addr_size));

        }

      else

        p += xsnprintf (p, endp - p, ";s");

    }

  else if (step)

    p += xsnprintf (p, endp - p, ";s");

  else if (siggnal != GDB_SIGNAL_0)

    p += xsnprintf (p, endp - p, ";C%02x", siggnal);

  else

    p += xsnprintf (p, endp - p, ";c");



  if (remote_multi_process_p (rs) && ptid_is_pid (ptid))

    {

      ptid_t nptid;



      /* All (-1) threads of process.  */

      nptid = ptid_build (ptid_get_pid (ptid), -1, 0);



      p += xsnprintf (p, endp - p, ":");

      p = write_ptid (p, endp, nptid);

    }

  else if (!ptid_equal (ptid, minus_one_ptid))

    {

      p += xsnprintf (p, endp - p, ":");

      p = write_ptid (p, endp, ptid);

    }



  return p;

}



/* Append a vCont continue-with-signal action for threads that have a

   non-zero stop signal.  */



static char *

‌append_pending_thread_resumptions (char *p, char *endp, ptid_t ptid)

{

  struct thread_info *thread;



  ALL_NON_EXITED_THREADS (thread)

    if (ptid_match (thread->ptid, ptid)

        && !ptid_equal (inferior_ptid, thread->ptid)

        && thread->suspend.stop_signal != GDB_SIGNAL_0)

      {

        p = append_resumption (p, endp, thread->ptid,

                               0, thread->suspend.stop_signal);

        thread->suspend.stop_signal = GDB_SIGNAL_0;

      }



  return p;

}



/* Resume the remote inferior by using a "vCont" packet.  The thread

   to be resumed is PTID; STEP and SIGGNAL indicate whether the

   resumed thread should be single-stepped and/or signalled.  If PTID

   equals minus_one_ptid, then all threads are resumed; the thread to

   be stepped and/or signalled is given in the global INFERIOR_PTID.

   This function returns non-zero iff it resumes the inferior.



   This function issues a strict subset of all possible vCont commands at the

   moment.  */



static int

‌remote_vcont_resume (ptid_t ptid, int step, enum gdb_signal siggnal)

{

  struct remote_state *rs = get_remote_state ();

  char *p;

  char *endp;



  if (packet_support (PACKET_vCont) == PACKET_SUPPORT_UNKNOWN)

    remote_vcont_probe (rs);



  if (packet_support (PACKET_vCont) == PACKET_DISABLE)

    return 0;



  p = rs->buf;

  endp = rs->buf + get_remote_packet_size ();



  /* If we could generate a wider range of packets, we'd have to worry

     about overflowing BUF.  Should there be a generic

     "multi-part-packet" packet?  */



  p += xsnprintf (p, endp - p, "vCont");



  if (ptid_equal (ptid, magic_null_ptid))

    {

      /* MAGIC_NULL_PTID means that we don't have any active threads,

         so we don't have any TID numbers the inferior will

         understand.  Make sure to only send forms that do not specify

         a TID.  */

      append_resumption (p, endp, minus_one_ptid, step, siggnal);

    }

  else if (ptid_equal (ptid, minus_one_ptid) || ptid_is_pid (ptid))

    {

      /* Resume all threads (of all processes, or of a single

         process), with preference for INFERIOR_PTID.  This assumes

         inferior_ptid belongs to the set of all threads we are about

         to resume.  */

      if (step || siggnal != GDB_SIGNAL_0)

        {

          /* Step inferior_ptid, with or without signal.  */

          p = append_resumption (p, endp, inferior_ptid, step, siggnal);

        }



      /* Also pass down any pending signaled resumption for other

         threads not the current.  */

      p = append_pending_thread_resumptions (p, endp, ptid);



      /* And continue others without a signal.  */

      append_resumption (p, endp, ptid, /*step=*/ 0, GDB_SIGNAL_0);

    }

  else

    {

      /* Scheduler locking; resume only PTID.  */

      append_resumption (p, endp, ptid, step, siggnal);

    }



  gdb_assert (strlen (rs->buf) < get_remote_packet_size ());

  putpkt (rs->buf);



  if (non_stop)

    {

      /* In non-stop, the stub replies to vCont with "OK".  The stop

         reply will be reported asynchronously by means of a `%Stop'

         notification.  */

      getpkt (&rs->buf, &rs->buf_size, 0);

      if (strcmp (rs->buf, "OK") != 0)

        error (_("Unexpected vCont reply in non-stop mode: %s"), rs->buf);

    }



  return 1;

}



/* Tell the remote machine to resume.  */



static void

‌remote_resume (struct target_ops *ops,

               ptid_t ptid, int step, enum gdb_signal siggnal)

{

  struct remote_state *rs = get_remote_state ();

  char *buf;



  /* In all-stop, we can't mark REMOTE_ASYNC_GET_PENDING_EVENTS_TOKEN

     (explained in remote-notif.c:handle_notification) so

     remote_notif_process is not called.  We need find a place where

     it is safe to start a 'vNotif' sequence.  It is good to do it

     before resuming inferior, because inferior was stopped and no RSP

     traffic at that moment.  */

  if (!non_stop)

    remote_notif_process (rs->notif_state, &notif_client_stop);



  rs->last_sent_signal = siggnal;

  rs->last_sent_step = step;



  /* The vCont packet doesn't need to specify threads via Hc.  */

  /* No reverse support (yet) for vCont.  */

  if (execution_direction != EXEC_REVERSE)

    if (remote_vcont_resume (ptid, step, siggnal))

      goto done;



  /* All other supported resume packets do use Hc, so set the continue

     thread.  */

  if (ptid_equal (ptid, minus_one_ptid))

    set_continue_thread (any_thread_ptid);

  else

    set_continue_thread (ptid);



  buf = rs->buf;

  if (execution_direction == EXEC_REVERSE)

    {

      /* We don't pass signals to the target in reverse exec mode.  */

      if (info_verbose && siggnal != GDB_SIGNAL_0)

        warning (_(" - Can't pass signal %d to target in reverse: ignored."),

                 siggnal);



      if (step && packet_support (PACKET_bs) == PACKET_DISABLE)

        error (_("Remote reverse-step not supported."));

      if (!step && packet_support (PACKET_bc) == PACKET_DISABLE)

        error (_("Remote reverse-continue not supported."));



      strcpy (buf, step ? "bs" : "bc");

    }

  else if (siggnal != GDB_SIGNAL_0)

    {

      buf[0] = step ? 'S' : 'C';

      buf[1] = tohex (((int) siggnal >> 4) & 0xf);

      buf[2] = tohex (((int) siggnal) & 0xf);

      buf[3] = '\0';

    }

  else

    strcpy (buf, step ? "s" : "c");



  putpkt (buf);



 done:

  /* We are about to start executing the inferior, let's register it

     with the event loop.  NOTE: this is the one place where all the

     execution commands end up.  We could alternatively do this in each

     of the execution commands in infcmd.c.  */

  /* FIXME: ezannoni 1999-09-28: We may need to move this out of here

     into infcmd.c in order to allow inferior function calls to work

     NOT asynchronously.  */

  if (target_can_async_p ())

    target_async (inferior_event_handler, 0);



  /* We've just told the target to resume.  The remote server will

     wait for the inferior to stop, and then send a stop reply.  In

     the mean time, we can't start another command/query ourselves

     because the stub wouldn't be ready to process it.  This applies

     only to the base all-stop protocol, however.  In non-stop (which

     only supports vCont), the stub replies with an "OK", and is

     immediate able to process further serial input.  */

  if (!non_stop)

    rs->waiting_for_stop_reply = 1;

}





/* Set up the signal handler for SIGINT, while the target is

   executing, ovewriting the 'regular' SIGINT signal handler.  */

static void

‌async_initialize_sigint_signal_handler (void)

{

  signal (SIGINT, async_handle_remote_sigint);

}



/* Signal handler for SIGINT, while the target is executing.  */

static void

‌async_handle_remote_sigint (int sig)

{

  signal (sig, async_handle_remote_sigint_twice);

  /* Note we need to go through gdb_call_async_signal_handler in order

     to wake up the event loop on Windows.  */

  gdb_call_async_signal_handler (async_sigint_remote_token, 0);

}



/* Signal handler for SIGINT, installed after SIGINT has already been

   sent once.  It will take effect the second time that the user sends

   a ^C.  */

static void

‌async_handle_remote_sigint_twice (int sig)

{

  signal (sig, async_handle_remote_sigint);

  /* See note in async_handle_remote_sigint.  */

  gdb_call_async_signal_handler (async_sigint_remote_twice_token, 0);

}



/* Perform the real interruption of the target execution, in response

   to a ^C.  */

static void

‌async_remote_interrupt (gdb_client_data arg)

{

  if (remote_debug)

    fprintf_unfiltered (gdb_stdlog, "async_remote_interrupt called\n");



  target_stop (inferior_ptid);

}



/* Perform interrupt, if the first attempt did not succeed.  Just give

   up on the target alltogether.  */

static void

‌async_remote_interrupt_twice (gdb_client_data arg)

{

  if (remote_debug)

    fprintf_unfiltered (gdb_stdlog, "async_remote_interrupt_twice called\n");



  interrupt_query ();

}



/* Reinstall the usual SIGINT handlers, after the target has

   stopped.  */

static void

‌async_cleanup_sigint_signal_handler (void *dummy)

{

  signal (SIGINT, handle_sigint);

}



/* Send ^C to target to halt it.  Target will respond, and send us a

   packet.  */

‌static void (*ofunc) (int);



/* The command line interface's stop routine.  This function is installed

   as a signal handler for SIGINT.  The first time a user requests a

   stop, we call remote_stop to send a break or ^C.  If there is no

   response from the target (it didn't stop when the user requested it),

   we ask the user if he'd like to detach from the target.  */

static void

‌sync_remote_interrupt (int signo)

{

  /* If this doesn't work, try more severe steps.  */

  signal (signo, sync_remote_interrupt_twice);



  gdb_call_async_signal_handler (async_sigint_remote_token, 1);

}



/* The user typed ^C twice.  */



static void

‌sync_remote_interrupt_twice (int signo)

{

  signal (signo, ofunc);

  gdb_call_async_signal_handler (async_sigint_remote_twice_token, 1);

  signal (signo, sync_remote_interrupt);

}



/* Non-stop version of target_stop.  Uses `vCont;t' to stop a remote

   thread, all threads of a remote process, or all threads of all

   processes.  */



static void

‌remote_stop_ns (ptid_t ptid)

{

  struct remote_state *rs = get_remote_state ();

  char *p = rs->buf;

  char *endp = rs->buf + get_remote_packet_size ();



  if (packet_support (PACKET_vCont) == PACKET_SUPPORT_UNKNOWN)

    remote_vcont_probe (rs);



  if (!rs->supports_vCont.t)

    error (_("Remote server does not support stopping threads"));



  if (ptid_equal (ptid, minus_one_ptid)

      || (!remote_multi_process_p (rs) && ptid_is_pid (ptid)))

    p += xsnprintf (p, endp - p, "vCont;t");

  else

    {

      ptid_t nptid;



      p += xsnprintf (p, endp - p, "vCont;t:");



      if (ptid_is_pid (ptid))

          /* All (-1) threads of process.  */

        nptid = ptid_build (ptid_get_pid (ptid), -1, 0);

      else

        {

          /* Small optimization: if we already have a stop reply for

             this thread, no use in telling the stub we want this

             stopped.  */

          if (peek_stop_reply (ptid))

            return;



          nptid = ptid;

        }



      write_ptid (p, endp, nptid);

    }



  /* In non-stop, we get an immediate OK reply.  The stop reply will

     come in asynchronously by notification.  */

  putpkt (rs->buf);

  getpkt (&rs->buf, &rs->buf_size, 0);

  if (strcmp (rs->buf, "OK") != 0)

    error (_("Stopping %s failed: %s"), target_pid_to_str (ptid), rs->buf);

}



/* All-stop version of target_stop.  Sends a break or a ^C to stop the

   remote target.  It is undefined which thread of which process

   reports the stop.  */



static void

‌remote_stop_as (ptid_t ptid)

{

  struct remote_state *rs = get_remote_state ();



  rs->ctrlc_pending_p = 1;



  /* If the inferior is stopped already, but the core didn't know

     about it yet, just ignore the request.  The cached wait status

     will be collected in remote_wait.  */

  if (rs->cached_wait_status)

    return;



  /* Send interrupt_sequence to remote target.  */

  send_interrupt_sequence ();

}



/* This is the generic stop called via the target vector.  When a target

   interrupt is requested, either by the command line or the GUI, we

   will eventually end up here.  */



static void

‌remote_stop (struct target_ops *self, ptid_t ptid)

{

  if (remote_debug)

    fprintf_unfiltered (gdb_stdlog, "remote_stop called\n");



  if (non_stop)

    remote_stop_ns (ptid);

  else

    remote_stop_as (ptid);

}



/* Ask the user what to do when an interrupt is received.  */



static void

‌interrupt_query (void)

{

  target_terminal_ours ();



  if (target_can_async_p ())

    {

      signal (SIGINT, handle_sigint);

      quit ();

    }

  else

    {

      if (query (_("Interrupted while waiting for the program.\n\

Give up (and stop debugging it)? ")))

        {

          remote_unpush_target ();

          quit ();

        }

    }



  target_terminal_inferior ();

}



/* Enable/disable target terminal ownership.  Most targets can use

   terminal groups to control terminal ownership.  Remote targets are

   different in that explicit transfer of ownership to/from GDB/target

   is required.  */



static void

‌remote_terminal_inferior (struct target_ops *self)

{

  if (!target_async_permitted)

    /* Nothing to do.  */

    return;



  /* FIXME: cagney/1999-09-27: Make calls to target_terminal_*()

     idempotent.  The event-loop GDB talking to an asynchronous target

     with a synchronous command calls this function from both

     event-top.c and infrun.c/infcmd.c.  Once GDB stops trying to

     transfer the terminal to the target when it shouldn't this guard

     can go away.  */

  if (!remote_async_terminal_ours_p)

    return;

  delete_file_handler (input_fd);

  remote_async_terminal_ours_p = 0;

  async_initialize_sigint_signal_handler ();

  /* NOTE: At this point we could also register our selves as the

     recipient of all input.  Any characters typed could then be

     passed on down to the target.  */

}



static void

‌remote_terminal_ours (struct target_ops *self)

{

  if (!target_async_permitted)

    /* Nothing to do.  */

    return;



  /* See FIXME in remote_terminal_inferior.  */

  if (remote_async_terminal_ours_p)

    return;

  async_cleanup_sigint_signal_handler (NULL);

  add_file_handler (input_fd, stdin_event_handler, 0);

  remote_async_terminal_ours_p = 1;

}



static void

‌remote_console_output (char *msg)

{

  char *p;



  for (p = msg; p[0] && p[1]; p += 2)

    {

      char tb[2];

      char c = fromhex (p[0]) * 16 + fromhex (p[1]);



      tb[0] = c;

      tb[1] = 0;

      fputs_unfiltered (tb, gdb_stdtarg);

    }

  gdb_flush (gdb_stdtarg);

}



‌typedef struct cached_reg

{

  int num;

  gdb_byte data[MAX_REGISTER_SIZE];

‌} cached_reg_t;



‌DEF_VEC_O(cached_reg_t);



‌typedef struct stop_reply

{

  struct notif_event base;



  /* The identifier of the thread about this event  */

  ptid_t ptid;



  /* The remote state this event is associated with.  When the remote

     connection, represented by a remote_state object, is closed,

     all the associated stop_reply events should be released.  */

  struct remote_state *rs;



  struct target_waitstatus ws;



  /* Expedited registers.  This makes remote debugging a bit more

     efficient for those targets that provide critical registers as

     part of their normal status mechanism (as another roundtrip to

     fetch them is avoided).  */

  VEC(cached_reg_t) *regcache;



  int stopped_by_watchpoint_p;

  CORE_ADDR watch_data_address;



  int core;

‌} *stop_reply_p;



‌DECLARE_QUEUE_P (stop_reply_p);

‌DEFINE_QUEUE_P (stop_reply_p);

/* The list of already fetched and acknowledged stop events.  This

   queue is used for notification Stop, and other notifications

   don't need queue for their events, because the notification events

   of Stop can't be consumed immediately, so that events should be

   queued first, and be consumed by remote_wait_{ns,as} one per

   time.  Other notifications can consume their events immediately,

   so queue is not needed for them.  */

static QUEUE (stop_reply_p) *stop_reply_queue;



static void

‌stop_reply_xfree (struct stop_reply *r)

{

  notif_event_xfree ((struct notif_event *) r);

}



static void

‌remote_notif_stop_parse (struct notif_client *self, char *buf,

                         struct notif_event *event)

{

  remote_parse_stop_reply (buf, (struct stop_reply *) event);

}



static void

‌remote_notif_stop_ack (struct notif_client *self, char *buf,

                       struct notif_event *event)

{

  struct stop_reply *stop_reply = (struct stop_reply *) event;



  /* acknowledge */

  putpkt ((char *) self->ack_command);



  if (stop_reply->ws.kind == TARGET_WAITKIND_IGNORE)

      /* We got an unknown stop reply.  */

      error (_("Unknown stop reply"));



  push_stop_reply (stop_reply);

}



static int

‌remote_notif_stop_can_get_pending_events (struct notif_client *self)

{

  /* We can't get pending events in remote_notif_process for

     notification stop, and we have to do this in remote_wait_ns

     instead.  If we fetch all queued events from stub, remote stub

     may exit and we have no chance to process them back in

     remote_wait_ns.  */

  mark_async_event_handler (remote_async_inferior_event_token);

  return 0;

}



static void

‌stop_reply_dtr (struct notif_event *event)

{

  struct stop_reply *r = (struct stop_reply *) event;



  VEC_free (cached_reg_t, r->regcache);

}



static struct notif_event *

‌remote_notif_stop_alloc_reply (void)

{

  struct notif_event *r

    = (struct notif_event *) XNEW (struct stop_reply);



  r->dtr = stop_reply_dtr;



  return r;

}



/* A client of notification Stop.  */



‌struct notif_client notif_client_stop =

{

  "Stop",

  "vStopped",

  remote_notif_stop_parse,

  remote_notif_stop_ack,

  remote_notif_stop_can_get_pending_events,

  remote_notif_stop_alloc_reply,

  REMOTE_NOTIF_STOP,

};



/* A parameter to pass data in and out.  */



‌struct queue_iter_param

{

  void *input;

  struct stop_reply *output;

};



/* Remove stop replies in the queue if its pid is equal to the given

   inferior's pid.  */



static int

‌remove_stop_reply_for_inferior (QUEUE (stop_reply_p) *q,

                                QUEUE_ITER (stop_reply_p) *iter,

                                stop_reply_p event,

                                void *data)

{

  struct queue_iter_param *param = data;

  struct inferior *inf = param->input;



  if (ptid_get_pid (event->ptid) == inf->pid)

    {

      stop_reply_xfree (event);

      QUEUE_remove_elem (stop_reply_p, q, iter);

    }



  return 1;

}



/* Discard all pending stop replies of inferior INF.  */



static void

‌discard_pending_stop_replies (struct inferior *inf)

{

  int i;

  struct queue_iter_param param;

  struct stop_reply *reply;

  struct remote_state *rs = get_remote_state ();

  struct remote_notif_state *rns = rs->notif_state;



  /* This function can be notified when an inferior exists.  When the

     target is not remote, the notification state is NULL.  */

  if (rs->remote_desc == NULL)

    return;



  reply = (struct stop_reply *) rns->pending_event[notif_client_stop.id];



  /* Discard the in-flight notification.  */

  if (reply != NULL && ptid_get_pid (reply->ptid) == inf->pid)

    {

      stop_reply_xfree (reply);

      rns->pending_event[notif_client_stop.id] = NULL;

    }



  param.input = inf;

  param.output = NULL;

  /* Discard the stop replies we have already pulled with

     vStopped.  */

  QUEUE_iterate (stop_reply_p, stop_reply_queue,

                 remove_stop_reply_for_inferior, &param);

}



/* If its remote state is equal to the given remote state,

   remove EVENT from the stop reply queue.  */



static int

‌remove_stop_reply_of_remote_state (QUEUE (stop_reply_p) *q,

                                   QUEUE_ITER (stop_reply_p) *iter,

                                   stop_reply_p event,

                                   void *data)

{

  struct queue_iter_param *param = data;

  struct remote_state *rs = param->input;



  if (event->rs == rs)

    {

      stop_reply_xfree (event);

      QUEUE_remove_elem (stop_reply_p, q, iter);

    }



  return 1;

}



/* Discard the stop replies for RS in stop_reply_queue.  */



static void

‌discard_pending_stop_replies_in_queue (struct remote_state *rs)

{

  struct queue_iter_param param;



  param.input = rs;

  param.output = NULL;

  /* Discard the stop replies we have already pulled with

     vStopped.  */

  QUEUE_iterate (stop_reply_p, stop_reply_queue,

                 remove_stop_reply_of_remote_state, &param);

}



/* A parameter to pass data in and out.  */



static int

‌remote_notif_remove_once_on_match (QUEUE (stop_reply_p) *q,

                                   QUEUE_ITER (stop_reply_p) *iter,

                                   stop_reply_p event,

                                   void *data)

{

  struct queue_iter_param *param = data;

  ptid_t *ptid = param->input;



  if (ptid_match (event->ptid, *ptid))

    {

      param->output = event;

      QUEUE_remove_elem (stop_reply_p, q, iter);

      return 0;

    }



  return 1;

}



/* Remove the first reply in 'stop_reply_queue' which matches

   PTID.  */



static struct stop_reply *

‌remote_notif_remove_queued_reply (ptid_t ptid)

{

  struct queue_iter_param param;



  param.input = &ptid;

  param.output = NULL;



  QUEUE_iterate (stop_reply_p, stop_reply_queue,

                 remote_notif_remove_once_on_match, &param);

  if (notif_debug)

    fprintf_unfiltered (gdb_stdlog,

                        "notif: discard queued event: 'Stop' in %s\n",

                        target_pid_to_str (ptid));



  return param.output;

}



/* Look for a queued stop reply belonging to PTID.  If one is found,

   remove it from the queue, and return it.  Returns NULL if none is

   found.  If there are still queued events left to process, tell the

   event loop to get back to target_wait soon.  */



static struct stop_reply *

‌queued_stop_reply (ptid_t ptid)

{

  struct stop_reply *r = remote_notif_remove_queued_reply (ptid);



  if (!QUEUE_is_empty (stop_reply_p, stop_reply_queue))

    /* There's still at least an event left.  */

    mark_async_event_handler (remote_async_inferior_event_token);



  return r;

}



/* Push a fully parsed stop reply in the stop reply queue.  Since we

   know that we now have at least one queued event left to pass to the

   core side, tell the event loop to get back to target_wait soon.  */



static void

‌push_stop_reply (struct stop_reply *new_event)

{

  QUEUE_enque (stop_reply_p, stop_reply_queue, new_event);



  if (notif_debug)

    fprintf_unfiltered (gdb_stdlog,

                        "notif: push 'Stop' %s to queue %d\n",

                        target_pid_to_str (new_event->ptid),

                        QUEUE_length (stop_reply_p,

                                      stop_reply_queue));



  mark_async_event_handler (remote_async_inferior_event_token);

}



static int

‌stop_reply_match_ptid_and_ws (QUEUE (stop_reply_p) *q,

                              QUEUE_ITER (stop_reply_p) *iter,

                              struct stop_reply *event,

                              void *data)

{

  ptid_t *ptid = data;



  return !(ptid_equal (*ptid, event->ptid)

           && event->ws.kind == TARGET_WAITKIND_STOPPED);

}



/* Returns true if we have a stop reply for PTID.  */



static int

‌peek_stop_reply (ptid_t ptid)

{

  return !QUEUE_iterate (stop_reply_p, stop_reply_queue,

                         stop_reply_match_ptid_and_ws, &ptid);

}



/* Parse the stop reply in BUF.  Either the function succeeds, and the

   result is stored in EVENT, or throws an error.  */



static void

‌remote_parse_stop_reply (char *buf, struct stop_reply *event)

{

  struct remote_arch_state *rsa = get_remote_arch_state ();

  ULONGEST addr;

  char *p;



  event->ptid = null_ptid;

  event->rs = get_remote_state ();

  event->ws.kind = TARGET_WAITKIND_IGNORE;

  event->ws.value.integer = 0;

  event->stopped_by_watchpoint_p = 0;

  event->regcache = NULL;

  event->core = -1;



  switch (buf[0])

    {

    case 'T':                /* Status with PC, SP, FP, ...        */

      /* Expedited reply, containing Signal, {regno, reg} repeat.  */

      /*  format is:  'Tssn...:r...;n...:r...;n...:r...;#cc', where

            ss = signal number

            n... = register number

            r... = register contents

      */



      p = &buf[3];        /* after Txx */

      while (*p)

        {

          char *p1;

          char *p_temp;

          int fieldsize;

          LONGEST pnum = 0;



          /* If the packet contains a register number, save it in

             pnum and set p1 to point to the character following it.

             Otherwise p1 points to p.  */



          /* If this packet is an awatch packet, don't parse the 'a'

             as a register number.  */



          if (strncmp (p, "awatch", strlen("awatch")) != 0

              && strncmp (p, "core", strlen ("core") != 0))

            {

              /* Read the ``P'' register number.  */

              pnum = strtol (p, &p_temp, 16);

              p1 = p_temp;

            }

          else

            p1 = p;



          if (p1 == p)        /* No register number present here.  */

            {

              p1 = strchr (p, ':');

              if (p1 == NULL)

                error (_("Malformed packet(a) (missing colon): %s\n\

Packet: '%s'\n"),

                       p, buf);

              if (strncmp (p, "thread", p1 - p) == 0)

                event->ptid = read_ptid (++p1, &p);

              else if ((strncmp (p, "watch", p1 - p) == 0)

                       || (strncmp (p, "rwatch", p1 - p) == 0)

                       || (strncmp (p, "awatch", p1 - p) == 0))

                {

                  event->stopped_by_watchpoint_p = 1;

                  p = unpack_varlen_hex (++p1, &addr);

                  event->watch_data_address = (CORE_ADDR) addr;

                }

              else if (strncmp (p, "library", p1 - p) == 0)

                {

                  p1++;

                  p_temp = p1;

                  while (*p_temp && *p_temp != ';')

                    p_temp++;



                  event->ws.kind = TARGET_WAITKIND_LOADED;

                  p = p_temp;

                }

              else if (strncmp (p, "replaylog", p1 - p) == 0)

                {

                  event->ws.kind = TARGET_WAITKIND_NO_HISTORY;

                  /* p1 will indicate "begin" or "end", but it makes

                     no difference for now, so ignore it.  */

                  p_temp = strchr (p1 + 1, ';');

                  if (p_temp)

                    p = p_temp;

                }

              else if (strncmp (p, "core", p1 - p) == 0)

                {

                  ULONGEST c;



                  p = unpack_varlen_hex (++p1, &c);

                  event->core = c;

                }

              else

                {

                  /* Silently skip unknown optional info.  */

                  p_temp = strchr (p1 + 1, ';');

                  if (p_temp)

                    p = p_temp;

                }

            }

          else

            {

              struct packet_reg *reg = packet_reg_from_pnum (rsa, pnum);

              cached_reg_t cached_reg;



              p = p1;



              if (*p != ':')

                error (_("Malformed packet(b) (missing colon): %s\n\

Packet: '%s'\n"),

                       p, buf);

              ++p;



              if (reg == NULL)

                error (_("Remote sent bad register number %s: %s\n\

Packet: '%s'\n"),

                       hex_string (pnum), p, buf);



              cached_reg.num = reg->regnum;



              fieldsize = hex2bin (p, cached_reg.data,

                                   register_size (target_gdbarch (),

                                                  reg->regnum));

              p += 2 * fieldsize;

              if (fieldsize < register_size (target_gdbarch (),

                                             reg->regnum))

                warning (_("Remote reply is too short: %s"), buf);



              VEC_safe_push (cached_reg_t, event->regcache, &cached_reg);

            }



          if (*p != ';')

            error (_("Remote register badly formatted: %s\nhere: %s"),

                   buf, p);

          ++p;

        }



      if (event->ws.kind != TARGET_WAITKIND_IGNORE)

        break;



      /* fall through */

    case 'S':                /* Old style status, just signal only.  */

      {

        int sig;



        event->ws.kind = TARGET_WAITKIND_STOPPED;

        sig = (fromhex (buf[1]) << 4) + fromhex (buf[2]);

        if (GDB_SIGNAL_FIRST <= sig && sig < GDB_SIGNAL_LAST)

          event->ws.value.sig = (enum gdb_signal) sig;

        else

          event->ws.value.sig = GDB_SIGNAL_UNKNOWN;

      }

      break;

    case 'W':                /* Target exited.  */

    case 'X':

      {

        char *p;

        int pid;

        ULONGEST value;



        /* GDB used to accept only 2 hex chars here.  Stubs should

           only send more if they detect GDB supports multi-process

           support.  */

        p = unpack_varlen_hex (&buf[1], &value);



        if (buf[0] == 'W')

          {

            /* The remote process exited.  */

            event->ws.kind = TARGET_WAITKIND_EXITED;

            event->ws.value.integer = value;

          }

        else

          {

            /* The remote process exited with a signal.  */

            event->ws.kind = TARGET_WAITKIND_SIGNALLED;

            if (GDB_SIGNAL_FIRST <= value && value < GDB_SIGNAL_LAST)

              event->ws.value.sig = (enum gdb_signal) value;

            else

              event->ws.value.sig = GDB_SIGNAL_UNKNOWN;

          }



        /* If no process is specified, assume inferior_ptid.  */

        pid = ptid_get_pid (inferior_ptid);

        if (*p == '\0')

          ;

        else if (*p == ';')

          {

            p++;



            if (p == '\0')

              ;

            else if (strncmp (p,

                              "process:", sizeof ("process:") - 1) == 0)

              {

                ULONGEST upid;



                p += sizeof ("process:") - 1;

                unpack_varlen_hex (p, &upid);

                pid = upid;

              }

            else

              error (_("unknown stop reply packet: %s"), buf);

          }

        else

          error (_("unknown stop reply packet: %s"), buf);

        event->ptid = pid_to_ptid (pid);

      }

      break;

    }



  if (non_stop && ptid_equal (event->ptid, null_ptid))

    error (_("No process or thread specified in stop reply: %s"), buf);

}



/* When the stub wants to tell GDB about a new notification reply, it

   sends a notification (%Stop, for example).  Those can come it at

   any time, hence, we have to make sure that any pending

   putpkt/getpkt sequence we're making is finished, before querying

   the stub for more events with the corresponding ack command

   (vStopped, for example).  E.g., if we started a vStopped sequence

   immediately upon receiving the notification, something like this

   could happen:



    1.1) --> Hg 1

    1.2) <-- OK

    1.3) --> g

    1.4) <-- %Stop

    1.5) --> vStopped

    1.6) <-- (registers reply to step #1.3)



   Obviously, the reply in step #1.6 would be unexpected to a vStopped

   query.



   To solve this, whenever we parse a %Stop notification successfully,

   we mark the REMOTE_ASYNC_GET_PENDING_EVENTS_TOKEN, and carry on

   doing whatever we were doing:



    2.1) --> Hg 1

    2.2) <-- OK

    2.3) --> g

    2.4) <-- %Stop

      <GDB marks the REMOTE_ASYNC_GET_PENDING_EVENTS_TOKEN>

    2.5) <-- (registers reply to step #2.3)



   Eventualy after step #2.5, we return to the event loop, which

   notices there's an event on the

   REMOTE_ASYNC_GET_PENDING_EVENTS_TOKEN event and calls the

   associated callback --- the function below.  At this point, we're

   always safe to start a vStopped sequence. :



    2.6) --> vStopped

    2.7) <-- T05 thread:2

    2.8) --> vStopped

    2.9) --> OK

*/



void

‌remote_notif_get_pending_events (struct notif_client *nc)

{

  struct remote_state *rs = get_remote_state ();



  if (rs->notif_state->pending_event[nc->id] != NULL)

    {

      if (notif_debug)

        fprintf_unfiltered (gdb_stdlog,

                            "notif: process: '%s' ack pending event\n",

                            nc->name);



      /* acknowledge */

      nc->ack (nc, rs->buf, rs->notif_state->pending_event[nc->id]);

      rs->notif_state->pending_event[nc->id] = NULL;



      while (1)

        {

          getpkt (&rs->buf, &rs->buf_size, 0);

          if (strcmp (rs->buf, "OK") == 0)

            break;

          else

            remote_notif_ack (nc, rs->buf);

        }

    }

  else

    {

      if (notif_debug)

        fprintf_unfiltered (gdb_stdlog,

                            "notif: process: '%s' no pending reply\n",

                            nc->name);

    }

}



/* Called when it is decided that STOP_REPLY holds the info of the

   event that is to be returned to the core.  This function always

   destroys STOP_REPLY.  */



static ptid_t

‌process_stop_reply (struct stop_reply *stop_reply,

                    struct target_waitstatus *status)

{

  ptid_t ptid;



  *status = stop_reply->ws;

  ptid = stop_reply->ptid;



  /* If no thread/process was reported by the stub, assume the current

     inferior.  */

  if (ptid_equal (ptid, null_ptid))

    ptid = inferior_ptid;



  if (status->kind != TARGET_WAITKIND_EXITED

      && status->kind != TARGET_WAITKIND_SIGNALLED)

    {

      struct remote_state *rs = get_remote_state ();



      /* Expedited registers.  */

      if (stop_reply->regcache)

        {

          struct regcache *regcache

            = get_thread_arch_regcache (ptid, target_gdbarch ());

          cached_reg_t *reg;

          int ix;



          for (ix = 0;

               VEC_iterate(cached_reg_t, stop_reply->regcache, ix, reg);

               ix++)

            regcache_raw_supply (regcache, reg->num, reg->data);

          VEC_free (cached_reg_t, stop_reply->regcache);

        }



      rs->remote_stopped_by_watchpoint_p = stop_reply->stopped_by_watchpoint_p;

      rs->remote_watch_data_address = stop_reply->watch_data_address;



      remote_notice_new_inferior (ptid, 0);

      demand_private_info (ptid)->core = stop_reply->core;

    }



  stop_reply_xfree (stop_reply);

  return ptid;

}



/* The non-stop mode version of target_wait.  */



static ptid_t

‌remote_wait_ns (ptid_t ptid, struct target_waitstatus *status, int options)

{

  struct remote_state *rs = get_remote_state ();

  struct stop_reply *stop_reply;

  int ret;

  int is_notif = 0;



  /* If in non-stop mode, get out of getpkt even if a

     notification is received.        */



  ret = getpkt_or_notif_sane (&rs->buf, &rs->buf_size,

                              0 /* forever */, &is_notif);

  while (1)

    {

      if (ret != -1 && !is_notif)

        switch (rs->buf[0])

          {

          case 'E':                /* Error of some sort.        */

            /* We're out of sync with the target now.  Did it continue

               or not?  We can't tell which thread it was in non-stop,

               so just ignore this.  */

            warning (_("Remote failure reply: %s"), rs->buf);

            break;

          case 'O':                /* Console output.  */

            remote_console_output (rs->buf + 1);

            break;

          default:

            warning (_("Invalid remote reply: %s"), rs->buf);

            break;

          }



      /* Acknowledge a pending stop reply that may have arrived in the

         mean time.  */

      if (rs->notif_state->pending_event[notif_client_stop.id] != NULL)

        remote_notif_get_pending_events (&notif_client_stop);



      /* If indeed we noticed a stop reply, we're done.  */

      stop_reply = queued_stop_reply (ptid);

      if (stop_reply != NULL)

        return process_stop_reply (stop_reply, status);



      /* Still no event.  If we're just polling for an event, then

         return to the event loop.  */

      if (options & TARGET_WNOHANG)

        {

          status->kind = TARGET_WAITKIND_IGNORE;

          return minus_one_ptid;

        }



      /* Otherwise do a blocking wait.  */

      ret = getpkt_or_notif_sane (&rs->buf, &rs->buf_size,

                                  1 /* forever */, &is_notif);

    }

}



/* Wait until the remote machine stops, then return, storing status in

   STATUS just as `wait' would.  */



static ptid_t

‌remote_wait_as (ptid_t ptid, struct target_waitstatus *status, int options)

{

  struct remote_state *rs = get_remote_state ();

  ptid_t event_ptid = null_ptid;

  char *buf;

  struct stop_reply *stop_reply;



 again:



  status->kind = TARGET_WAITKIND_IGNORE;

  status->value.integer = 0;



  stop_reply = queued_stop_reply (ptid);

  if (stop_reply != NULL)

    return process_stop_reply (stop_reply, status);



  if (rs->cached_wait_status)

    /* Use the cached wait status, but only once.  */

    rs->cached_wait_status = 0;

  else

    {

      int ret;

      int is_notif;



      if (!target_is_async_p ())

        {

          ofunc = signal (SIGINT, sync_remote_interrupt);

          /* If the user hit C-c before this packet, or between packets,

             pretend that it was hit right here.  */

          if (check_quit_flag ())

            {

              clear_quit_flag ();

              sync_remote_interrupt (SIGINT);

            }

        }



      /* FIXME: cagney/1999-09-27: If we're in async mode we should

         _never_ wait for ever -> test on target_is_async_p().

         However, before we do that we need to ensure that the caller

         knows how to take the target into/out of async mode.  */

      ret = getpkt_or_notif_sane (&rs->buf, &rs->buf_size,

                                  wait_forever_enabled_p, &is_notif);



      if (!target_is_async_p ())

        signal (SIGINT, ofunc);



      /* GDB gets a notification.  Return to core as this event is

         not interesting.  */

      if (ret != -1 && is_notif)

        return minus_one_ptid;

    }



  buf = rs->buf;



  rs->remote_stopped_by_watchpoint_p = 0;



  /* We got something.  */

  rs->waiting_for_stop_reply = 0;



  /* Assume that the target has acknowledged Ctrl-C unless we receive

     an 'F' or 'O' packet.  */

  if (buf[0] != 'F' && buf[0] != 'O')

    rs->ctrlc_pending_p = 0;



  switch (buf[0])

    {

    case 'E':                /* Error of some sort.        */

      /* We're out of sync with the target now.  Did it continue or

         not?  Not is more likely, so report a stop.  */

      warning (_("Remote failure reply: %s"), buf);

      status->kind = TARGET_WAITKIND_STOPPED;

      status->value.sig = GDB_SIGNAL_0;

      break;

    case 'F':                /* File-I/O request.  */

      remote_fileio_request (buf, rs->ctrlc_pending_p);

      rs->ctrlc_pending_p = 0;

      break;

    case 'T': case 'S': case 'X': case 'W':

      {

        struct stop_reply *stop_reply

          = (struct stop_reply *) remote_notif_parse (&notif_client_stop,

                                                      rs->buf);



        event_ptid = process_stop_reply (stop_reply, status);

        break;

      }

    case 'O':                /* Console output.  */

      remote_console_output (buf + 1);



      /* The target didn't really stop; keep waiting.  */

      rs->waiting_for_stop_reply = 1;



      break;

    case '\0':

      if (rs->last_sent_signal != GDB_SIGNAL_0)

        {

          /* Zero length reply means that we tried 'S' or 'C' and the

             remote system doesn't support it.  */

          target_terminal_ours_for_output ();

          printf_filtered

            ("Can't send signals to this remote system.  %s not sent.\n",

             gdb_signal_to_name (rs->last_sent_signal));

          rs->last_sent_signal = GDB_SIGNAL_0;

          target_terminal_inferior ();



          strcpy ((char *) buf, rs->last_sent_step ? "s" : "c");

          putpkt ((char *) buf);



          /* We just told the target to resume, so a stop reply is in

             order.  */

          rs->waiting_for_stop_reply = 1;

          break;

        }

      /* else fallthrough */

    default:

      warning (_("Invalid remote reply: %s"), buf);

      /* Keep waiting.  */

      rs->waiting_for_stop_reply = 1;

      break;

    }



  if (status->kind == TARGET_WAITKIND_IGNORE)

    {

      /* Nothing interesting happened.  If we're doing a non-blocking

         poll, we're done.  Otherwise, go back to waiting.  */

      if (options & TARGET_WNOHANG)

        return minus_one_ptid;

      else

        goto again;

    }

  else if (status->kind != TARGET_WAITKIND_EXITED

           && status->kind != TARGET_WAITKIND_SIGNALLED)

    {

      if (!ptid_equal (event_ptid, null_ptid))

        record_currthread (rs, event_ptid);

      else

        event_ptid = inferior_ptid;

    }

  else

    /* A process exit.  Invalidate our notion of current thread.  */

    record_currthread (rs, minus_one_ptid);



  return event_ptid;

}



/* Wait until the remote machine stops, then return, storing status in

   STATUS just as `wait' would.  */



static ptid_t

‌remote_wait (struct target_ops *ops,

             ptid_t ptid, struct target_waitstatus *status, int options)

{

  ptid_t event_ptid;



  if (non_stop)

    event_ptid = remote_wait_ns (ptid, status, options);

  else

    event_ptid = remote_wait_as (ptid, status, options);



  if (target_can_async_p ())

    {

      /* If there are are events left in the queue tell the event loop

         to return here.  */

      if (!QUEUE_is_empty (stop_reply_p, stop_reply_queue))

        mark_async_event_handler (remote_async_inferior_event_token);

    }



  return event_ptid;

}



/* Fetch a single register using a 'p' packet.  */



static int

‌fetch_register_using_p (struct regcache *regcache, struct packet_reg *reg)

{

  struct remote_state *rs = get_remote_state ();

  char *buf, *p;

  char regp[MAX_REGISTER_SIZE];

  int i;



  if (packet_support (PACKET_p) == PACKET_DISABLE)

    return 0;



  if (reg->pnum == -1)

    return 0;



  p = rs->buf;

  *p++ = 'p';

  p += hexnumstr (p, reg->pnum);

  *p++ = '\0';

  putpkt (rs->buf);

  getpkt (&rs->buf, &rs->buf_size, 0);



  buf = rs->buf;



  switch (packet_ok (buf, &remote_protocol_packets[PACKET_p]))

    {

    case PACKET_OK:

      break;

    case PACKET_UNKNOWN:

      return 0;

    case PACKET_ERROR:

      error (_("Could not fetch register \"%s\"; remote failure reply '%s'"),

             gdbarch_register_name (get_regcache_arch (regcache),

                                    reg->regnum),

             buf);

    }



  /* If this register is unfetchable, tell the regcache.  */

  if (buf[0] == 'x')

    {

      regcache_raw_supply (regcache, reg->regnum, NULL);

      return 1;

    }



  /* Otherwise, parse and supply the value.  */

  p = buf;

  i = 0;

  while (p[0] != 0)

    {

      if (p[1] == 0)

        error (_("fetch_register_using_p: early buf termination"));



      regp[i++] = fromhex (p[0]) * 16 + fromhex (p[1]);

      p += 2;

    }

  regcache_raw_supply (regcache, reg->regnum, regp);

  return 1;

}



/* Fetch the registers included in the target's 'g' packet.  */



static int

‌send_g_packet (void)

{

  struct remote_state *rs = get_remote_state ();

  int buf_len;



  xsnprintf (rs->buf, get_remote_packet_size (), "g");

  remote_send (&rs->buf, &rs->buf_size);



  /* We can get out of synch in various cases.  If the first character

     in the buffer is not a hex character, assume that has happened

     and try to fetch another packet to read.  */

  while ((rs->buf[0] < '0' || rs->buf[0] > '9')

         && (rs->buf[0] < 'A' || rs->buf[0] > 'F')

         && (rs->buf[0] < 'a' || rs->buf[0] > 'f')

         && rs->buf[0] != 'x')        /* New: unavailable register value.  */

    {

      if (remote_debug)

        fprintf_unfiltered (gdb_stdlog,

                            "Bad register packet; fetching a new packet\n");

      getpkt (&rs->buf, &rs->buf_size, 0);

    }



  buf_len = strlen (rs->buf);



  /* Sanity check the received packet.  */

  if (buf_len % 2 != 0)

    error (_("Remote 'g' packet reply is of odd length: %s"), rs->buf);



  return buf_len / 2;

}



static void

‌process_g_packet (struct regcache *regcache)

{

  struct gdbarch *gdbarch = get_regcache_arch (regcache);

  struct remote_state *rs = get_remote_state ();

  struct remote_arch_state *rsa = get_remote_arch_state ();

  int i, buf_len;

  char *p;

  char *regs;



  buf_len = strlen (rs->buf);



  /* Further sanity checks, with knowledge of the architecture.  */

  if (buf_len > 2 * rsa->sizeof_g_packet)

    error (_("Remote 'g' packet reply is too long: %s"), rs->buf);



  /* Save the size of the packet sent to us by the target.  It is used

     as a heuristic when determining the max size of packets that the

     target can safely receive.  */

  if (rsa->actual_register_packet_size == 0)

    rsa->actual_register_packet_size = buf_len;



  /* If this is smaller than we guessed the 'g' packet would be,

     update our records.  A 'g' reply that doesn't include a register's

     value implies either that the register is not available, or that

     the 'p' packet must be used.  */

  if (buf_len < 2 * rsa->sizeof_g_packet)

    {

      rsa->sizeof_g_packet = buf_len / 2;



      for (i = 0; i < gdbarch_num_regs (gdbarch); i++)

        {

          if (rsa->regs[i].pnum == -1)

            continue;



          if (rsa->regs[i].offset >= rsa->sizeof_g_packet)

            rsa->regs[i].in_g_packet = 0;

          else

            rsa->regs[i].in_g_packet = 1;

        }

    }



  regs = alloca (rsa->sizeof_g_packet);



  /* Unimplemented registers read as all bits zero.  */

  memset (regs, 0, rsa->sizeof_g_packet);



  /* Reply describes registers byte by byte, each byte encoded as two

     hex characters.  Suck them all up, then supply them to the

     register cacheing/storage mechanism.  */



  p = rs->buf;

  for (i = 0; i < rsa->sizeof_g_packet; i++)

    {

      if (p[0] == 0 || p[1] == 0)

        /* This shouldn't happen - we adjusted sizeof_g_packet above.  */

        internal_error (__FILE__, __LINE__,

                        _("unexpected end of 'g' packet reply"));



      if (p[0] == 'x' && p[1] == 'x')

        regs[i] = 0;                /* 'x' */

      else

        regs[i] = fromhex (p[0]) * 16 + fromhex (p[1]);

      p += 2;

    }



  for (i = 0; i < gdbarch_num_regs (gdbarch); i++)

    {

      struct packet_reg *r = &rsa->regs[i];



      if (r->in_g_packet)

        {

          if (r->offset * 2 >= strlen (rs->buf))

            /* This shouldn't happen - we adjusted in_g_packet above.  */

            internal_error (__FILE__, __LINE__,

                            _("unexpected end of 'g' packet reply"));

          else if (rs->buf[r->offset * 2] == 'x')

            {

              gdb_assert (r->offset * 2 < strlen (rs->buf));

              /* The register isn't available, mark it as such (at

                 the same time setting the value to zero).  */

              regcache_raw_supply (regcache, r->regnum, NULL);

            }

          else

            regcache_raw_supply (regcache, r->regnum,

                                 regs + r->offset);

        }

    }

}



static void

‌fetch_registers_using_g (struct regcache *regcache)

{

  send_g_packet ();

  process_g_packet (regcache);

}



/* Make the remote selected traceframe match GDB's selected

   traceframe.  */



static void

‌set_remote_traceframe (void)

{

  int newnum;

  struct remote_state *rs = get_remote_state ();



  if (rs->remote_traceframe_number == get_traceframe_number ())

    return;



  /* Avoid recursion, remote_trace_find calls us again.  */

  rs->remote_traceframe_number = get_traceframe_number ();



  newnum = target_trace_find (tfind_number,

                              get_traceframe_number (), 0, 0, NULL);



  /* Should not happen.  If it does, all bets are off.  */

  if (newnum != get_traceframe_number ())

    warning (_("could not set remote traceframe"));

}



static void

‌remote_fetch_registers (struct target_ops *ops,

                        struct regcache *regcache, int regnum)

{

  struct remote_arch_state *rsa = get_remote_arch_state ();

  int i;



  set_remote_traceframe ();

  set_general_thread (inferior_ptid);



  if (regnum >= 0)

    {

      struct packet_reg *reg = packet_reg_from_regnum (rsa, regnum);



      gdb_assert (reg != NULL);



      /* If this register might be in the 'g' packet, try that first -

         we are likely to read more than one register.  If this is the

         first 'g' packet, we might be overly optimistic about its

         contents, so fall back to 'p'.  */

      if (reg->in_g_packet)

        {

          fetch_registers_using_g (regcache);

          if (reg->in_g_packet)

            return;

        }



      if (fetch_register_using_p (regcache, reg))

        return;



      /* This register is not available.  */

      regcache_raw_supply (regcache, reg->regnum, NULL);



      return;

    }



  fetch_registers_using_g (regcache);



  for (i = 0; i < gdbarch_num_regs (get_regcache_arch (regcache)); i++)

    if (!rsa->regs[i].in_g_packet)

      if (!fetch_register_using_p (regcache, &rsa->regs[i]))

        {

          /* This register is not available.  */

          regcache_raw_supply (regcache, i, NULL);

        }

}



/* Prepare to store registers.  Since we may send them all (using a

   'G' request), we have to read out the ones we don't want to change

   first.  */



static void

‌remote_prepare_to_store (struct target_ops *self, struct regcache *regcache)

{

  struct remote_arch_state *rsa = get_remote_arch_state ();

  int i;

  gdb_byte buf[MAX_REGISTER_SIZE];



  /* Make sure the entire registers array is valid.  */

  switch (packet_support (PACKET_P))

    {

    case PACKET_DISABLE:

    case PACKET_SUPPORT_UNKNOWN:

      /* Make sure all the necessary registers are cached.  */

      for (i = 0; i < gdbarch_num_regs (get_regcache_arch (regcache)); i++)

        if (rsa->regs[i].in_g_packet)

          regcache_raw_read (regcache, rsa->regs[i].regnum, buf);

      break;

    case PACKET_ENABLE:

      break;

    }

}



/* Helper: Attempt to store REGNUM using the P packet.  Return fail IFF

   packet was not recognized.  */



static int

‌store_register_using_P (const struct regcache *regcache,

                        struct packet_reg *reg)

{

  struct gdbarch *gdbarch = get_regcache_arch (regcache);

  struct remote_state *rs = get_remote_state ();

  /* Try storing a single register.  */

  char *buf = rs->buf;

  gdb_byte regp[MAX_REGISTER_SIZE];

  char *p;



  if (packet_support (PACKET_P) == PACKET_DISABLE)

    return 0;



  if (reg->pnum == -1)

    return 0;



  xsnprintf (buf, get_remote_packet_size (), "P%s=", phex_nz (reg->pnum, 0));

  p = buf + strlen (buf);

  regcache_raw_collect (regcache, reg->regnum, regp);

  bin2hex (regp, p, register_size (gdbarch, reg->regnum));

  putpkt (rs->buf);

  getpkt (&rs->buf, &rs->buf_size, 0);



  switch (packet_ok (rs->buf, &remote_protocol_packets[PACKET_P]))

    {

    case PACKET_OK:

      return 1;

    case PACKET_ERROR:

      error (_("Could not write register \"%s\"; remote failure reply '%s'"),

             gdbarch_register_name (gdbarch, reg->regnum), rs->buf);

    case PACKET_UNKNOWN:

      return 0;

    default:

      internal_error (__FILE__, __LINE__, _("Bad result from packet_ok"));

    }

}



/* Store register REGNUM, or all registers if REGNUM == -1, from the

   contents of the register cache buffer.  FIXME: ignores errors.  */



static void

‌store_registers_using_G (const struct regcache *regcache)

{

  struct remote_state *rs = get_remote_state ();

  struct remote_arch_state *rsa = get_remote_arch_state ();

  gdb_byte *regs;

  char *p;



  /* Extract all the registers in the regcache copying them into a

     local buffer.  */

  {

    int i;



    regs = alloca (rsa->sizeof_g_packet);

    memset (regs, 0, rsa->sizeof_g_packet);

    for (i = 0; i < gdbarch_num_regs (get_regcache_arch (regcache)); i++)

      {

        struct packet_reg *r = &rsa->regs[i];



        if (r->in_g_packet)

          regcache_raw_collect (regcache, r->regnum, regs + r->offset);

      }

  }



  /* Command describes registers byte by byte,

     each byte encoded as two hex characters.  */

  p = rs->buf;

  *p++ = 'G';

  /* remote_prepare_to_store insures that rsa->sizeof_g_packet gets

     updated.  */

  bin2hex (regs, p, rsa->sizeof_g_packet);

  putpkt (rs->buf);

  getpkt (&rs->buf, &rs->buf_size, 0);

  if (packet_check_result (rs->buf) == PACKET_ERROR)

    error (_("Could not write registers; remote failure reply '%s'"),

           rs->buf);

}



/* Store register REGNUM, or all registers if REGNUM == -1, from the contents

   of the register cache buffer.  FIXME: ignores errors.  */



static void

‌remote_store_registers (struct target_ops *ops,

                        struct regcache *regcache, int regnum)

{

  struct remote_arch_state *rsa = get_remote_arch_state ();

  int i;



  set_remote_traceframe ();

  set_general_thread (inferior_ptid);



  if (regnum >= 0)

    {

      struct packet_reg *reg = packet_reg_from_regnum (rsa, regnum);



      gdb_assert (reg != NULL);



      /* Always prefer to store registers using the 'P' packet if

         possible; we often change only a small number of registers.

         Sometimes we change a larger number; we'd need help from a

         higher layer to know to use 'G'.  */

      if (store_register_using_P (regcache, reg))

        return;



      /* For now, don't complain if we have no way to write the

         register.  GDB loses track of unavailable registers too

         easily.  Some day, this may be an error.  We don't have

         any way to read the register, either...  */

      if (!reg->in_g_packet)

        return;



      store_registers_using_G (regcache);

      return;

    }



  store_registers_using_G (regcache);



  for (i = 0; i < gdbarch_num_regs (get_regcache_arch (regcache)); i++)

    if (!rsa->regs[i].in_g_packet)

      if (!store_register_using_P (regcache, &rsa->regs[i]))

        /* See above for why we do not issue an error here.  */

        continue;

}





/* Return the number of hex digits in num.  */



static int

‌hexnumlen (ULONGEST num)

{

  int i;



  for (i = 0; num != 0; i++)

    num >>= 4;



  return max (i, 1);

}



/* Set BUF to the minimum number of hex digits representing NUM.  */



static int

‌hexnumstr (char *buf, ULONGEST num)

{

  int len = hexnumlen (num);



  return hexnumnstr (buf, num, len);

}





/* Set BUF to the hex digits representing NUM, padded to WIDTH characters.  */



static int

‌hexnumnstr (char *buf, ULONGEST num, int width)

{

  int i;



  buf[width] = '\0';



  for (i = width - 1; i >= 0; i--)

    {

      buf[i] = "0123456789abcdef"[(num & 0xf)];

      num >>= 4;

    }



  return width;

}



/* Mask all but the least significant REMOTE_ADDRESS_SIZE bits.  */



static CORE_ADDR

‌remote_address_masked (CORE_ADDR addr)

{

  unsigned int address_size = remote_address_size;



  /* If "remoteaddresssize" was not set, default to target address size.  */

  if (!address_size)

    address_size = gdbarch_addr_bit (target_gdbarch ());



  if (address_size > 0

      && address_size < (sizeof (ULONGEST) * 8))

    {

      /* Only create a mask when that mask can safely be constructed

         in a ULONGEST variable.  */

      ULONGEST mask = 1;



      mask = (mask << address_size) - 1;

      addr &= mask;

    }

  return addr;

}



/* Determine whether the remote target supports binary downloading.

   This is accomplished by sending a no-op memory write of zero length

   to the target at the specified address. It does not suffice to send

   the whole packet, since many stubs strip the eighth bit and

   subsequently compute a wrong checksum, which causes real havoc with

   remote_write_bytes.



   NOTE: This can still lose if the serial line is not eight-bit

   clean.  In cases like this, the user should clear "remote

   X-packet".  */



static void

‌check_binary_download (CORE_ADDR addr)

{

  struct remote_state *rs = get_remote_state ();



  switch (packet_support (PACKET_X))

    {

    case PACKET_DISABLE:

      break;

    case PACKET_ENABLE:

      break;

    case PACKET_SUPPORT_UNKNOWN:

      {

        char *p;



        p = rs->buf;

        *p++ = 'X';

        p += hexnumstr (p, (ULONGEST) addr);

        *p++ = ',';

        p += hexnumstr (p, (ULONGEST) 0);

        *p++ = ':';

        *p = '\0';



        putpkt_binary (rs->buf, (int) (p - rs->buf));

        getpkt (&rs->buf, &rs->buf_size, 0);



        if (rs->buf[0] == '\0')

          {

            if (remote_debug)

              fprintf_unfiltered (gdb_stdlog,

                                  "binary downloading NOT "

                                  "supported by target\n");

            remote_protocol_packets[PACKET_X].support = PACKET_DISABLE;

          }

        else

          {

            if (remote_debug)

              fprintf_unfiltered (gdb_stdlog,

                                  "binary downloading supported by target\n");

            remote_protocol_packets[PACKET_X].support = PACKET_ENABLE;

          }

        break;

      }

    }

}



/* Write memory data directly to the remote machine.

   This does not inform the data cache; the data cache uses this.

   HEADER is the starting part of the packet.

   MEMADDR is the address in the remote memory space.

   MYADDR is the address of the buffer in our space.

   LEN is the number of bytes.

   PACKET_FORMAT should be either 'X' or 'M', and indicates if we

   should send data as binary ('X'), or hex-encoded ('M').



   The function creates packet of the form

       <HEADER><ADDRESS>,<LENGTH>:<DATA>



   where encoding of <DATA> is termined by PACKET_FORMAT.



   If USE_LENGTH is 0, then the <LENGTH> field and the preceding comma

   are omitted.



   Return the transferred status, error or OK (an

   'enum target_xfer_status' value).  Save the number of bytes

   transferred in *XFERED_LEN.  Only transfer a single packet.  */



static enum target_xfer_status

‌remote_write_bytes_aux (const char *header, CORE_ADDR memaddr,

                        const gdb_byte *myaddr, ULONGEST len,

                        ULONGEST *xfered_len, char packet_format,

                        int use_length)

{

  struct remote_state *rs = get_remote_state ();

  char *p;

  char *plen = NULL;

  int plenlen = 0;

  int todo;

  int nr_bytes;

  int payload_size;

  int payload_length;

  int header_length;



  if (packet_format != 'X' && packet_format != 'M')

    internal_error (__FILE__, __LINE__,

                    _("remote_write_bytes_aux: bad packet format"));



  if (len == 0)

    return TARGET_XFER_EOF;



  payload_size = get_memory_write_packet_size ();



  /* The packet buffer will be large enough for the payload;

     get_memory_packet_size ensures this.  */

  rs->buf[0] = '\0';



  /* Compute the size of the actual payload by subtracting out the

     packet header and footer overhead: "$M<memaddr>,<len>:...#nn".  */



  payload_size -= strlen ("$,:#NN");

  if (!use_length)

    /* The comma won't be used.  */

    payload_size += 1;

  header_length = strlen (header);

  payload_size -= header_length;

  payload_size -= hexnumlen (memaddr);



  /* Construct the packet excluding the data: "<header><memaddr>,<len>:".  */



  strcat (rs->buf, header);

  p = rs->buf + strlen (header);



  /* Compute a best guess of the number of bytes actually transfered.  */

  if (packet_format == 'X')

    {

      /* Best guess at number of bytes that will fit.  */

      todo = min (len, payload_size);

      if (use_length)

        payload_size -= hexnumlen (todo);

      todo = min (todo, payload_size);

    }

  else

    {

      /* Num bytes that will fit.  */

      todo = min (len, payload_size / 2);

      if (use_length)

        payload_size -= hexnumlen (todo);

      todo = min (todo, payload_size / 2);

    }



  if (todo <= 0)

    internal_error (__FILE__, __LINE__,

                    _("minimum packet size too small to write data"));



  /* If we already need another packet, then try to align the end

     of this packet to a useful boundary.  */

  if (todo > 2 * REMOTE_ALIGN_WRITES && todo < len)

    todo = ((memaddr + todo) & ~(REMOTE_ALIGN_WRITES - 1)) - memaddr;



  /* Append "<memaddr>".  */

  memaddr = remote_address_masked (memaddr);

  p += hexnumstr (p, (ULONGEST) memaddr);



  if (use_length)

    {

      /* Append ",".  */

      *p++ = ',';



      /* Append <len>.  Retain the location/size of <len>.  It may need to

         be adjusted once the packet body has been created.  */

      plen = p;

      plenlen = hexnumstr (p, (ULONGEST) todo);

      p += plenlen;

    }



  /* Append ":".  */

  *p++ = ':';

  *p = '\0';



  /* Append the packet body.  */

  if (packet_format == 'X')

    {

      /* Binary mode.  Send target system values byte by byte, in

         increasing byte addresses.  Only escape certain critical

         characters.  */

      payload_length = remote_escape_output (myaddr, todo, (gdb_byte *) p,

                                             &nr_bytes, payload_size);



      /* If not all TODO bytes fit, then we'll need another packet.  Make

         a second try to keep the end of the packet aligned.  Don't do

         this if the packet is tiny.  */

      if (nr_bytes < todo && nr_bytes > 2 * REMOTE_ALIGN_WRITES)

        {

          int new_nr_bytes;



          new_nr_bytes = (((memaddr + nr_bytes) & ~(REMOTE_ALIGN_WRITES - 1))

                          - memaddr);

          if (new_nr_bytes != nr_bytes)

            payload_length = remote_escape_output (myaddr, new_nr_bytes,

                                                   (gdb_byte *) p, &nr_bytes,

                                                   payload_size);

        }



      p += payload_length;

      if (use_length && nr_bytes < todo)

        {

          /* Escape chars have filled up the buffer prematurely,

             and we have actually sent fewer bytes than planned.

             Fix-up the length field of the packet.  Use the same

             number of characters as before.  */

          plen += hexnumnstr (plen, (ULONGEST) nr_bytes, plenlen);

          *plen = ':';  /* overwrite \0 from hexnumnstr() */

        }

    }

  else

    {

      /* Normal mode: Send target system values byte by byte, in

         increasing byte addresses.  Each byte is encoded as a two hex

         value.  */

      nr_bytes = bin2hex (myaddr, p, todo);

      p += 2 * nr_bytes;

    }



  putpkt_binary (rs->buf, (int) (p - rs->buf));

  getpkt (&rs->buf, &rs->buf_size, 0);



  if (rs->buf[0] == 'E')

    return TARGET_XFER_E_IO;



  /* Return NR_BYTES, not TODO, in case escape chars caused us to send

     fewer bytes than we'd planned.  */

  *xfered_len = (ULONGEST) nr_bytes;

  return TARGET_XFER_OK;

}



/* Write memory data directly to the remote machine.

   This does not inform the data cache; the data cache uses this.

   MEMADDR is the address in the remote memory space.

   MYADDR is the address of the buffer in our space.

   LEN is the number of bytes.



   Return the transferred status, error or OK (an

   'enum target_xfer_status' value).  Save the number of bytes

   transferred in *XFERED_LEN.  Only transfer a single packet.  */



static enum target_xfer_status

‌remote_write_bytes (CORE_ADDR memaddr, const gdb_byte *myaddr, ULONGEST len,

                    ULONGEST *xfered_len)

{

  char *packet_format = 0;



  /* Check whether the target supports binary download.  */

  check_binary_download (memaddr);



  switch (packet_support (PACKET_X))

    {

    case PACKET_ENABLE:

      packet_format = "X";

      break;

    case PACKET_DISABLE:

      packet_format = "M";

      break;

    case PACKET_SUPPORT_UNKNOWN:

      internal_error (__FILE__, __LINE__,

                      _("remote_write_bytes: bad internal state"));

    default:

      internal_error (__FILE__, __LINE__, _("bad switch"));

    }



  return remote_write_bytes_aux (packet_format,

                                 memaddr, myaddr, len, xfered_len,

                                 packet_format[0], 1);

}



/* Read memory data directly from the remote machine.

   This does not use the data cache; the data cache uses this.

   MEMADDR is the address in the remote memory space.

   MYADDR is the address of the buffer in our space.

   LEN is the number of bytes.



   Return the transferred status, error or OK (an

   'enum target_xfer_status' value).  Save the number of bytes

   transferred in *XFERED_LEN.  */



static enum target_xfer_status

‌remote_read_bytes_1 (CORE_ADDR memaddr, gdb_byte *myaddr, ULONGEST len,

                     ULONGEST *xfered_len)

{

  struct remote_state *rs = get_remote_state ();

  int max_buf_size;                /* Max size of packet output buffer.  */

  char *p;

  int todo;

  int i;



  max_buf_size = get_memory_read_packet_size ();

  /* The packet buffer will be large enough for the payload;

     get_memory_packet_size ensures this.  */



  /* Number if bytes that will fit.  */

  todo = min (len, max_buf_size / 2);



  /* Construct "m"<memaddr>","<len>".  */

  memaddr = remote_address_masked (memaddr);

  p = rs->buf;

  *p++ = 'm';

  p += hexnumstr (p, (ULONGEST) memaddr);

  *p++ = ',';

  p += hexnumstr (p, (ULONGEST) todo);

  *p = '\0';

  putpkt (rs->buf);

  getpkt (&rs->buf, &rs->buf_size, 0);

  if (rs->buf[0] == 'E'

      && isxdigit (rs->buf[1]) && isxdigit (rs->buf[2])

      && rs->buf[3] == '\0')

    return TARGET_XFER_E_IO;

  /* Reply describes memory byte by byte, each byte encoded as two hex

     characters.  */

  p = rs->buf;

  i = hex2bin (p, myaddr, todo);

  /* Return what we have.  Let higher layers handle partial reads.  */

  *xfered_len = (ULONGEST) i;

  return TARGET_XFER_OK;

}



/* Using the set of read-only target sections of remote, read live

   read-only memory.



   For interface/parameters/return description see target.h,

   to_xfer_partial.  */



static enum target_xfer_status

‌remote_xfer_live_readonly_partial (struct target_ops *ops, gdb_byte *readbuf,

                                   ULONGEST memaddr, ULONGEST len,

                                   ULONGEST *xfered_len)

{

  struct target_section *secp;

  struct target_section_table *table;



  secp = target_section_by_addr (ops, memaddr);

  if (secp != NULL

      && (bfd_get_section_flags (secp->the_bfd_section->owner,

                                 secp->the_bfd_section)

          & SEC_READONLY))

    {

      struct target_section *p;

      ULONGEST memend = memaddr + len;



      table = target_get_section_table (ops);



      for (p = table->sections; p < table->sections_end; p++)

        {

          if (memaddr >= p->addr)

            {

              if (memend <= p->endaddr)

                {

                  /* Entire transfer is within this section.  */

                  return remote_read_bytes_1 (memaddr, readbuf, len,

                                              xfered_len);

                }

              else if (memaddr >= p->endaddr)

                {

                  /* This section ends before the transfer starts.  */

                  continue;

                }

              else

                {

                  /* This section overlaps the transfer.  Just do half.  */

                  len = p->endaddr - memaddr;

                  return remote_read_bytes_1 (memaddr, readbuf, len,

                                              xfered_len);

                }

            }

        }

    }



  return TARGET_XFER_EOF;

}



/* Similar to remote_read_bytes_1, but it reads from the remote stub

   first if the requested memory is unavailable in traceframe.

   Otherwise, fall back to remote_read_bytes_1.  */



static enum target_xfer_status

‌remote_read_bytes (struct target_ops *ops, CORE_ADDR memaddr,

                   gdb_byte *myaddr, ULONGEST len, ULONGEST *xfered_len)

{

  if (len == 0)

    return TARGET_XFER_EOF;



  if (get_traceframe_number () != -1)

    {

      VEC(mem_range_s) *available;



      /* If we fail to get the set of available memory, then the

         target does not support querying traceframe info, and so we

         attempt reading from the traceframe anyway (assuming the

         target implements the old QTro packet then).  */

      if (traceframe_available_memory (&available, memaddr, len))

        {

          struct cleanup *old_chain;



          old_chain = make_cleanup (VEC_cleanup(mem_range_s), &available);



          if (VEC_empty (mem_range_s, available)

              || VEC_index (mem_range_s, available, 0)->start != memaddr)

            {

              enum target_xfer_status res;



              /* Don't read into the traceframe's available

                 memory.  */

              if (!VEC_empty (mem_range_s, available))

                {

                  LONGEST oldlen = len;



                  len = VEC_index (mem_range_s, available, 0)->start - memaddr;

                  gdb_assert (len <= oldlen);

                }



              do_cleanups (old_chain);



              /* This goes through the topmost target again.  */

              res = remote_xfer_live_readonly_partial (ops, myaddr, memaddr,

                                                       len, xfered_len);

              if (res == TARGET_XFER_OK)

                return TARGET_XFER_OK;

              else

                {

                  /* No use trying further, we know some memory starting

                     at MEMADDR isn't available.  */

                  *xfered_len = len;

                  return TARGET_XFER_UNAVAILABLE;

                }

            }



          /* Don't try to read more than how much is available, in

             case the target implements the deprecated QTro packet to

             cater for older GDBs (the target's knowledge of read-only

             sections may be outdated by now).  */

          len = VEC_index (mem_range_s, available, 0)->length;



          do_cleanups (old_chain);

        }

    }



  return remote_read_bytes_1 (memaddr, myaddr, len, xfered_len);

}







/* Sends a packet with content determined by the printf format string

   FORMAT and the remaining arguments, then gets the reply.  Returns

   whether the packet was a success, a failure, or unknown.  */



static enum packet_result

‌remote_send_printf (const char *format, ...)

{

  struct remote_state *rs = get_remote_state ();

  int max_size = get_remote_packet_size ();

  va_list ap;



  va_start (ap, format);



  rs->buf[0] = '\0';

  if (vsnprintf (rs->buf, max_size, format, ap) >= max_size)

    internal_error (__FILE__, __LINE__, _("Too long remote packet."));



  if (putpkt (rs->buf) < 0)

    error (_("Communication problem with target."));



  rs->buf[0] = '\0';

  getpkt (&rs->buf, &rs->buf_size, 0);



  return packet_check_result (rs->buf);

}



static void

‌restore_remote_timeout (void *p)

{

  int value = *(int *)p;



  remote_timeout = value;

}



/* Flash writing can take quite some time.  We'll set

   effectively infinite timeout for flash operations.

   In future, we'll need to decide on a better approach.  */

‌static const int remote_flash_timeout = 1000;



static void

‌remote_flash_erase (struct target_ops *ops,

                    ULONGEST address, LONGEST length)

{

  int addr_size = gdbarch_addr_bit (target_gdbarch ()) / 8;

  int saved_remote_timeout = remote_timeout;

  enum packet_result ret;

  struct cleanup *back_to = make_cleanup (restore_remote_timeout,

                                          &saved_remote_timeout);



  remote_timeout = remote_flash_timeout;



  ret = remote_send_printf ("vFlashErase:%s,%s",

                            phex (address, addr_size),

                            phex (length, 4));

  switch (ret)

    {

    case PACKET_UNKNOWN:

      error (_("Remote target does not support flash erase"));

    case PACKET_ERROR:

      error (_("Error erasing flash with vFlashErase packet"));

    default:

      break;

    }



  do_cleanups (back_to);

}



static enum target_xfer_status

‌remote_flash_write (struct target_ops *ops, ULONGEST address,

                    ULONGEST length, ULONGEST *xfered_len,

                    const gdb_byte *data)

{

  int saved_remote_timeout = remote_timeout;

  enum target_xfer_status ret;

  struct cleanup *back_to = make_cleanup (restore_remote_timeout,

                                          &saved_remote_timeout);



  remote_timeout = remote_flash_timeout;

  ret = remote_write_bytes_aux ("vFlashWrite:", address, data, length,

                                xfered_len,'X', 0);

  do_cleanups (back_to);



  return ret;

}



static void

‌remote_flash_done (struct target_ops *ops)

{

  int saved_remote_timeout = remote_timeout;

  int ret;

  struct cleanup *back_to = make_cleanup (restore_remote_timeout,

                                          &saved_remote_timeout);



  remote_timeout = remote_flash_timeout;

  ret = remote_send_printf ("vFlashDone");

  do_cleanups (back_to);



  switch (ret)

    {

    case PACKET_UNKNOWN:

      error (_("Remote target does not support vFlashDone"));

    case PACKET_ERROR:

      error (_("Error finishing flash operation"));

    default:

      break;

    }

}



static void

‌remote_files_info (struct target_ops *ignore)

{

  puts_filtered ("Debugging a target over a serial line.\n");

}



/* Stuff for dealing with the packets which are part of this protocol.

   See comment at top of file for details.  */



/* Close/unpush the remote target, and throw a TARGET_CLOSE_ERROR

   error to higher layers.  Called when a serial error is detected.

   The exception message is STRING, followed by a colon and a blank,

   the system error message for errno at function entry and final dot

   for output compatibility with throw_perror_with_name.  */



static void

‌unpush_and_perror (const char *string)

{

  int saved_errno = errno;



  remote_unpush_target ();

  throw_error (TARGET_CLOSE_ERROR, "%s: %s.", string,

               safe_strerror (saved_errno));

}



/* Read a single character from the remote end.  */



static int

‌readchar (int timeout)

{

  int ch;

  struct remote_state *rs = get_remote_state ();



  ch = serial_readchar (rs->remote_desc, timeout);



  if (ch >= 0)

    return ch;



  switch ((enum serial_rc) ch)

    {

    case SERIAL_EOF:

      remote_unpush_target ();

      throw_error (TARGET_CLOSE_ERROR, _("Remote connection closed"));

      /* no return */

    case SERIAL_ERROR:

      unpush_and_perror (_("Remote communication error.  "

                           "Target disconnected."));

      /* no return */

    case SERIAL_TIMEOUT:

      break;

    }

  return ch;

}



/* Wrapper for serial_write that closes the target and throws if

   writing fails.  */



static void

‌remote_serial_write (const char *str, int len)

{

  struct remote_state *rs = get_remote_state ();



  if (serial_write (rs->remote_desc, str, len))

    {

      unpush_and_perror (_("Remote communication error.  "

                           "Target disconnected."));

    }

}



/* Send the command in *BUF to the remote machine, and read the reply

   into *BUF.  Report an error if we get an error reply.  Resize

   *BUF using xrealloc if necessary to hold the result, and update

   *SIZEOF_BUF.  */



static void

‌remote_send (char **buf,

             long *sizeof_buf)

{

  putpkt (*buf);

  getpkt (buf, sizeof_buf, 0);



  if ((*buf)[0] == 'E')

    error (_("Remote failure reply: %s"), *buf);

}



/* Return a pointer to an xmalloc'ed string representing an escaped

   version of BUF, of len N.  E.g. \n is converted to \\n, \t to \\t,

   etc.  The caller is responsible for releasing the returned

   memory.  */



static char *

‌escape_buffer (const char *buf, int n)

{

  struct cleanup *old_chain;

  struct ui_file *stb;

  char *str;



  stb = mem_fileopen ();

  old_chain = make_cleanup_ui_file_delete (stb);



  fputstrn_unfiltered (buf, n, '\\', stb);

  str = ui_file_xstrdup (stb, NULL);

  do_cleanups (old_chain);

  return str;

}



/* Display a null-terminated packet on stdout, for debugging, using C

   string notation.  */



static void

‌print_packet (const char *buf)

{

  puts_filtered ("\"");

  fputstr_filtered (buf, '"', gdb_stdout);

  puts_filtered ("\"");

}



int

‌putpkt (const char *buf)

{

  return putpkt_binary (buf, strlen (buf));

}



/* Send a packet to the remote machine, with error checking.  The data

   of the packet is in BUF.  The string in BUF can be at most

   get_remote_packet_size () - 5 to account for the $, # and checksum,

   and for a possible /0 if we are debugging (remote_debug) and want

   to print the sent packet as a string.  */



static int

‌putpkt_binary (const char *buf, int cnt)

{

  struct remote_state *rs = get_remote_state ();

  int i;

  unsigned char csum = 0;

  char *buf2 = alloca (cnt + 6);



  int ch;

  int tcount = 0;

  char *p;

  char *message;



  /* Catch cases like trying to read memory or listing threads while

     we're waiting for a stop reply.  The remote server wouldn't be

     ready to handle this request, so we'd hang and timeout.  We don't

     have to worry about this in synchronous mode, because in that

     case it's not possible to issue a command while the target is

     running.  This is not a problem in non-stop mode, because in that

     case, the stub is always ready to process serial input.  */

  if (!non_stop && target_can_async_p () && rs->waiting_for_stop_reply)

    {

      error (_("Cannot execute this command while the target is running.\n"

               "Use the \"interrupt\" command to stop the target\n"

               "and then try again."));

    }



  /* We're sending out a new packet.  Make sure we don't look at a

     stale cached response.  */

  rs->cached_wait_status = 0;



  /* Copy the packet into buffer BUF2, encapsulating it

     and giving it a checksum.  */



  p = buf2;

  *p++ = '$';



  for (i = 0; i < cnt; i++)

    {

      csum += buf[i];

      *p++ = buf[i];

    }

  *p++ = '#';

  *p++ = tohex ((csum >> 4) & 0xf);

  *p++ = tohex (csum & 0xf);



  /* Send it over and over until we get a positive ack.  */



  while (1)

    {

      int started_error_output = 0;



      if (remote_debug)

        {

          struct cleanup *old_chain;

          char *str;



          *p = '\0';

          str = escape_buffer (buf2, p - buf2);

          old_chain = make_cleanup (xfree, str);

          fprintf_unfiltered (gdb_stdlog, "Sending packet: %s...", str);

          gdb_flush (gdb_stdlog);

          do_cleanups (old_chain);

        }

      remote_serial_write (buf2, p - buf2);



      /* If this is a no acks version of the remote protocol, send the

         packet and move on.  */

      if (rs->noack_mode)

        break;



      /* Read until either a timeout occurs (-2) or '+' is read.

         Handle any notification that arrives in the mean time.  */

      while (1)

        {

          ch = readchar (remote_timeout);



          if (remote_debug)

            {

              switch (ch)

                {

                case '+':

                case '-':

                case SERIAL_TIMEOUT:

                case '$':

                case '%':

                  if (started_error_output)

                    {

                      putchar_unfiltered ('\n');

                      started_error_output = 0;

                    }

                }

            }



          switch (ch)

            {

            case '+':

              if (remote_debug)

                fprintf_unfiltered (gdb_stdlog, "Ack\n");

              return 1;

            case '-':

              if (remote_debug)

                fprintf_unfiltered (gdb_stdlog, "Nak\n");

              /* FALLTHROUGH */

            case SERIAL_TIMEOUT:

              tcount++;

              if (tcount > 3)

                return 0;

              break;                /* Retransmit buffer.  */

            case '$':

              {

                if (remote_debug)

                  fprintf_unfiltered (gdb_stdlog,

                                      "Packet instead of Ack, ignoring it\n");

                /* It's probably an old response sent because an ACK

                   was lost.  Gobble up the packet and ack it so it

                   doesn't get retransmitted when we resend this

                   packet.  */

                skip_frame ();

                remote_serial_write ("+", 1);

                continue;        /* Now, go look for +.  */

              }



            case '%':

              {

                int val;



                /* If we got a notification, handle it, and go back to looking

                   for an ack.  */

                /* We've found the start of a notification.  Now

                   collect the data.  */

                val = read_frame (&rs->buf, &rs->buf_size);

                if (val >= 0)

                  {

                    if (remote_debug)

                      {

                        struct cleanup *old_chain;

                        char *str;



                        str = escape_buffer (rs->buf, val);

                        old_chain = make_cleanup (xfree, str);

                        fprintf_unfiltered (gdb_stdlog,

                                            "  Notification received: %s\n",

                                            str);

                        do_cleanups (old_chain);

                      }

                    handle_notification (rs->notif_state, rs->buf);

                    /* We're in sync now, rewait for the ack.  */

                    tcount = 0;

                  }

                else

                  {

                    if (remote_debug)

                      {

                        if (!started_error_output)

                          {

                            started_error_output = 1;

                            fprintf_unfiltered (gdb_stdlog, "putpkt: Junk: ");

                          }

                        fputc_unfiltered (ch & 0177, gdb_stdlog);

                        fprintf_unfiltered (gdb_stdlog, "%s", rs->buf);

                      }

                  }

                continue;

              }

              /* fall-through */

            default:

              if (remote_debug)

                {

                  if (!started_error_output)

                    {

                      started_error_output = 1;

                      fprintf_unfiltered (gdb_stdlog, "putpkt: Junk: ");

                    }

                  fputc_unfiltered (ch & 0177, gdb_stdlog);

                }

              continue;

            }

          break;                /* Here to retransmit.  */

        }



#if 0

      /* This is wrong.  If doing a long backtrace, the user should be

         able to get out next time we call QUIT, without anything as

         violent as interrupt_query.  If we want to provide a way out of

         here without getting to the next QUIT, it should be based on

         hitting ^C twice as in remote_wait.  */

      if (quit_flag)

        {

          quit_flag = 0;

          interrupt_query ();

        }

#endif

    }

  return 0;

}



/* Come here after finding the start of a frame when we expected an

   ack.  Do our best to discard the rest of this packet.  */



static void

‌skip_frame (void)

{

  int c;



  while (1)

    {

      c = readchar (remote_timeout);

      switch (c)

        {

        case SERIAL_TIMEOUT:

          /* Nothing we can do.  */

          return;

        case '#':

          /* Discard the two bytes of checksum and stop.  */

          c = readchar (remote_timeout);

          if (c >= 0)

            c = readchar (remote_timeout);



          return;

        case '*':                /* Run length encoding.  */

          /* Discard the repeat count.  */

          c = readchar (remote_timeout);

          if (c < 0)

            return;

          break;

        default:

          /* A regular character.  */

          break;

        }

    }

}



/* Come here after finding the start of the frame.  Collect the rest

   into *BUF, verifying the checksum, length, and handling run-length

   compression.  NUL terminate the buffer.  If there is not enough room,

   expand *BUF using xrealloc.



   Returns -1 on error, number of characters in buffer (ignoring the

   trailing NULL) on success. (could be extended to return one of the

   SERIAL status indications).  */



static long

‌read_frame (char **buf_p,

            long *sizeof_buf)

{

  unsigned char csum;

  long bc;

  int c;

  char *buf = *buf_p;

  struct remote_state *rs = get_remote_state ();



  csum = 0;

  bc = 0;



  while (1)

    {

      c = readchar (remote_timeout);

      switch (c)

        {

        case SERIAL_TIMEOUT:

          if (remote_debug)

            fputs_filtered ("Timeout in mid-packet, retrying\n", gdb_stdlog);

          return -1;

        case '$':

          if (remote_debug)

            fputs_filtered ("Saw new packet start in middle of old one\n",

                            gdb_stdlog);

          return -1;                /* Start a new packet, count retries.  */

        case '#':

          {

            unsigned char pktcsum;

            int check_0 = 0;

            int check_1 = 0;



            buf[bc] = '\0';



            check_0 = readchar (remote_timeout);

            if (check_0 >= 0)

              check_1 = readchar (remote_timeout);



            if (check_0 == SERIAL_TIMEOUT || check_1 == SERIAL_TIMEOUT)

              {

                if (remote_debug)

                  fputs_filtered ("Timeout in checksum, retrying\n",

                                  gdb_stdlog);

                return -1;

              }

            else if (check_0 < 0 || check_1 < 0)

              {

                if (remote_debug)

                  fputs_filtered ("Communication error in checksum\n",

                                  gdb_stdlog);

                return -1;

              }



            /* Don't recompute the checksum; with no ack packets we

               don't have any way to indicate a packet retransmission

               is necessary.  */

            if (rs->noack_mode)

              return bc;



            pktcsum = (fromhex (check_0) << 4) | fromhex (check_1);

            if (csum == pktcsum)

              return bc;



            if (remote_debug)

              {

                struct cleanup *old_chain;

                char *str;



                str = escape_buffer (buf, bc);

                old_chain = make_cleanup (xfree, str);

                fprintf_unfiltered (gdb_stdlog,

                                    "Bad checksum, sentsum=0x%x, "

                                    "csum=0x%x, buf=%s\n",

                                    pktcsum, csum, str);

                do_cleanups (old_chain);

              }

            /* Number of characters in buffer ignoring trailing

               NULL.  */

            return -1;

          }

        case '*':                /* Run length encoding.  */

          {

            int repeat;



             csum += c;

            c = readchar (remote_timeout);

            csum += c;

            repeat = c - ' ' + 3;        /* Compute repeat count.  */



            /* The character before ``*'' is repeated.  */



            if (repeat > 0 && repeat <= 255 && bc > 0)

              {

                if (bc + repeat - 1 >= *sizeof_buf - 1)

                  {

                    /* Make some more room in the buffer.  */

                    *sizeof_buf += repeat;

                    *buf_p = xrealloc (*buf_p, *sizeof_buf);

                    buf = *buf_p;

                  }



                memset (&buf[bc], buf[bc - 1], repeat);

                bc += repeat;

                continue;

              }



            buf[bc] = '\0';

            printf_filtered (_("Invalid run length encoding: %s\n"), buf);

            return -1;

          }

        default:

          if (bc >= *sizeof_buf - 1)

            {

              /* Make some more room in the buffer.  */

              *sizeof_buf *= 2;

              *buf_p = xrealloc (*buf_p, *sizeof_buf);

              buf = *buf_p;

            }



          buf[bc++] = c;

          csum += c;

          continue;

        }

    }

}



/* Read a packet from the remote machine, with error checking, and

   store it in *BUF.  Resize *BUF using xrealloc if necessary to hold

   the result, and update *SIZEOF_BUF.  If FOREVER, wait forever

   rather than timing out; this is used (in synchronous mode) to wait

   for a target that is is executing user code to stop.  */

/* FIXME: ezannoni 2000-02-01 this wrapper is necessary so that we

   don't have to change all the calls to getpkt to deal with the

   return value, because at the moment I don't know what the right

   thing to do it for those.  */

void

‌getpkt (char **buf,

        long *sizeof_buf,

        int forever)

{

  int timed_out;



  timed_out = getpkt_sane (buf, sizeof_buf, forever);

}





/* Read a packet from the remote machine, with error checking, and

   store it in *BUF.  Resize *BUF using xrealloc if necessary to hold

   the result, and update *SIZEOF_BUF.  If FOREVER, wait forever

   rather than timing out; this is used (in synchronous mode) to wait

   for a target that is is executing user code to stop.  If FOREVER ==

   0, this function is allowed to time out gracefully and return an

   indication of this to the caller.  Otherwise return the number of

   bytes read.  If EXPECTING_NOTIF, consider receiving a notification

   enough reason to return to the caller.  *IS_NOTIF is an output

   boolean that indicates whether *BUF holds a notification or not

   (a regular packet).  */



static int

‌getpkt_or_notif_sane_1 (char **buf, long *sizeof_buf, int forever,

                        int expecting_notif, int *is_notif)

{

  struct remote_state *rs = get_remote_state ();

  int c;

  int tries;

  int timeout;

  int val = -1;



  /* We're reading a new response.  Make sure we don't look at a

     previously cached response.  */

  rs->cached_wait_status = 0;



  strcpy (*buf, "timeout");



  if (forever)

    timeout = watchdog > 0 ? watchdog : -1;

  else if (expecting_notif)

    timeout = 0; /* There should already be a char in the buffer.  If

                    not, bail out.  */

  else

    timeout = remote_timeout;



‌#define MAX_TRIES 3



  /* Process any number of notifications, and then return when

     we get a packet.  */

  for (;;)

    {

      /* If we get a timeout or bad checksum, retry up to MAX_TRIES

         times.  */

      for (tries = 1; tries <= MAX_TRIES; tries++)

        {

          /* This can loop forever if the remote side sends us

             characters continuously, but if it pauses, we'll get

             SERIAL_TIMEOUT from readchar because of timeout.  Then

             we'll count that as a retry.



             Note that even when forever is set, we will only wait

             forever prior to the start of a packet.  After that, we

             expect characters to arrive at a brisk pace.  They should

             show up within remote_timeout intervals.  */

          do

            c = readchar (timeout);

          while (c != SERIAL_TIMEOUT && c != '$' && c != '%');



          if (c == SERIAL_TIMEOUT)

            {

              if (expecting_notif)

                return -1; /* Don't complain, it's normal to not get

                              anything in this case.  */



              if (forever)        /* Watchdog went off?  Kill the target.  */

                {

                  QUIT;

                  remote_unpush_target ();

                  throw_error (TARGET_CLOSE_ERROR,

                               _("Watchdog timeout has expired.  "

                                 "Target detached."));

                }

              if (remote_debug)

                fputs_filtered ("Timed out.\n", gdb_stdlog);

            }

          else

            {

              /* We've found the start of a packet or notification.

                 Now collect the data.  */

              val = read_frame (buf, sizeof_buf);

              if (val >= 0)

                break;

            }



          remote_serial_write ("-", 1);

        }



      if (tries > MAX_TRIES)

        {

          /* We have tried hard enough, and just can't receive the

             packet/notification.  Give up.  */

          printf_unfiltered (_("Ignoring packet error, continuing...\n"));



          /* Skip the ack char if we're in no-ack mode.  */

          if (!rs->noack_mode)

            remote_serial_write ("+", 1);

          return -1;

        }



      /* If we got an ordinary packet, return that to our caller.  */

      if (c == '$')

        {

          if (remote_debug)

            {

             struct cleanup *old_chain;

             char *str;



             str = escape_buffer (*buf, val);

             old_chain = make_cleanup (xfree, str);

             fprintf_unfiltered (gdb_stdlog, "Packet received: %s\n", str);

             do_cleanups (old_chain);

            }



          /* Skip the ack char if we're in no-ack mode.  */

          if (!rs->noack_mode)

            remote_serial_write ("+", 1);

          if (is_notif != NULL)

            *is_notif = 0;

          return val;

        }



       /* If we got a notification, handle it, and go back to looking

         for a packet.  */

      else

        {

          gdb_assert (c == '%');



          if (remote_debug)

            {

              struct cleanup *old_chain;

              char *str;



              str = escape_buffer (*buf, val);

              old_chain = make_cleanup (xfree, str);

              fprintf_unfiltered (gdb_stdlog,

                                  "  Notification received: %s\n",

                                  str);

              do_cleanups (old_chain);

            }

          if (is_notif != NULL)

            *is_notif = 1;



          handle_notification (rs->notif_state, *buf);



          /* Notifications require no acknowledgement.  */



          if (expecting_notif)

            return val;

        }

    }

}



static int

‌getpkt_sane (char **buf, long *sizeof_buf, int forever)

{

  return getpkt_or_notif_sane_1 (buf, sizeof_buf, forever, 0, NULL);

}



static int

‌getpkt_or_notif_sane (char **buf, long *sizeof_buf, int forever,

                      int *is_notif)

{

  return getpkt_or_notif_sane_1 (buf, sizeof_buf, forever, 1,

                                 is_notif);

}





static void

‌remote_kill (struct target_ops *ops)

{

  volatile struct gdb_exception ex;



  /* Catch errors so the user can quit from gdb even when we

     aren't on speaking terms with the remote system.  */

  TRY_CATCH (ex, RETURN_MASK_ERROR)

    {

      putpkt ("k");

    }

  if (ex.reason < 0)

    {

      if (ex.error == TARGET_CLOSE_ERROR)

        {

          /* If we got an (EOF) error that caused the target

             to go away, then we're done, that's what we wanted.

             "k" is susceptible to cause a premature EOF, given

             that the remote server isn't actually required to

             reply to "k", and it can happen that it doesn't

             even get to reply ACK to the "k".  */

          return;

        }



        /* Otherwise, something went wrong.  We didn't actually kill

           the target.  Just propagate the exception, and let the

           user or higher layers decide what to do.  */

        throw_exception (ex);

    }



  /* We've killed the remote end, we get to mourn it.  Since this is

     target remote, single-process, mourning the inferior also

     unpushes remote_ops.  */

  target_mourn_inferior ();

}



static int

‌remote_vkill (int pid, struct remote_state *rs)

{

  if (packet_support (PACKET_vKill) == PACKET_DISABLE)

    return -1;



  /* Tell the remote target to detach.  */

  xsnprintf (rs->buf, get_remote_packet_size (), "vKill;%x", pid);

  putpkt (rs->buf);

  getpkt (&rs->buf, &rs->buf_size, 0);



  switch (packet_ok (rs->buf,

                     &remote_protocol_packets[PACKET_vKill]))

    {

    case PACKET_OK:

      return 0;

    case PACKET_ERROR:

      return 1;

    case PACKET_UNKNOWN:

      return -1;

    default:

      internal_error (__FILE__, __LINE__, _("Bad result from packet_ok"));

    }

}



static void

‌extended_remote_kill (struct target_ops *ops)

{

  int res;

  int pid = ptid_get_pid (inferior_ptid);

  struct remote_state *rs = get_remote_state ();



  res = remote_vkill (pid, rs);

  if (res == -1 && !(rs->extended && remote_multi_process_p (rs)))

    {

      /* Don't try 'k' on a multi-process aware stub -- it has no way

         to specify the pid.  */



      putpkt ("k");

#if 0

      getpkt (&rs->buf, &rs->buf_size, 0);

      if (rs->buf[0] != 'O' || rs->buf[0] != 'K')

        res = 1;

#else

      /* Don't wait for it to die.  I'm not really sure it matters whether

         we do or not.  For the existing stubs, kill is a noop.  */

      res = 0;

#endif

    }



  if (res != 0)

    error (_("Can't kill process"));



  target_mourn_inferior ();

}



static void

‌remote_mourn (struct target_ops *ops)

{

  remote_mourn_1 (ops);

}



/* Worker function for remote_mourn.  */

static void

‌remote_mourn_1 (struct target_ops *target)

{

  unpush_target (target);



  /* remote_close takes care of doing most of the clean up.  */

  generic_mourn_inferior ();

}



static void

‌extended_remote_mourn_1 (struct target_ops *target)

{

  struct remote_state *rs = get_remote_state ();



  /* In case we got here due to an error, but we're going to stay

     connected.  */

  rs->waiting_for_stop_reply = 0;



  /* If the current general thread belonged to the process we just

     detached from or has exited, the remote side current general

     thread becomes undefined.  Considering a case like this:



     - We just got here due to a detach.

     - The process that we're detaching from happens to immediately

       report a global breakpoint being hit in non-stop mode, in the

       same thread we had selected before.

     - GDB attaches to this process again.

     - This event happens to be the next event we handle.



     GDB would consider that the current general thread didn't need to

     be set on the stub side (with Hg), since for all it knew,

     GENERAL_THREAD hadn't changed.



     Notice that although in all-stop mode, the remote server always

     sets the current thread to the thread reporting the stop event,

     that doesn't happen in non-stop mode; in non-stop, the stub *must

     not* change the current thread when reporting a breakpoint hit,

     due to the decoupling of event reporting and event handling.



     To keep things simple, we always invalidate our notion of the

     current thread.  */

  record_currthread (rs, minus_one_ptid);



  /* Unlike "target remote", we do not want to unpush the target; then

     the next time the user says "run", we won't be connected.  */



  /* Call common code to mark the inferior as not running.        */

  generic_mourn_inferior ();



  if (!have_inferiors ())

    {

      if (!remote_multi_process_p (rs))

        {

          /* Check whether the target is running now - some remote stubs

             automatically restart after kill.        */

          putpkt ("?");

          getpkt (&rs->buf, &rs->buf_size, 0);



          if (rs->buf[0] == 'S' || rs->buf[0] == 'T')

            {

              /* Assume that the target has been restarted.  Set

                 inferior_ptid so that bits of core GDB realizes

                 there's something here, e.g., so that the user can

                 say "kill" again.  */

              inferior_ptid = magic_null_ptid;

            }

        }

    }

}



static void

‌extended_remote_mourn (struct target_ops *ops)

{

  extended_remote_mourn_1 (ops);

}



static int

‌extended_remote_supports_disable_randomization (struct target_ops *self)

{

  return packet_support (PACKET_QDisableRandomization) == PACKET_ENABLE;

}



static void

‌extended_remote_disable_randomization (int val)

{

  struct remote_state *rs = get_remote_state ();

  char *reply;



  xsnprintf (rs->buf, get_remote_packet_size (), "QDisableRandomization:%x",

             val);

  putpkt (rs->buf);

  reply = remote_get_noisy_reply (&target_buf, &target_buf_size);

  if (*reply == '\0')

    error (_("Target does not support QDisableRandomization."));

  if (strcmp (reply, "OK") != 0)

    error (_("Bogus QDisableRandomization reply from target: %s"), reply);

}



static int

‌extended_remote_run (char *args)

{

  struct remote_state *rs = get_remote_state ();

  int len;



  /* If the user has disabled vRun support, or we have detected that

     support is not available, do not try it.  */

  if (packet_support (PACKET_vRun) == PACKET_DISABLE)

    return -1;



  strcpy (rs->buf, "vRun;");

  len = strlen (rs->buf);



  if (strlen (remote_exec_file) * 2 + len >= get_remote_packet_size ())

    error (_("Remote file name too long for run packet"));

  len += 2 * bin2hex ((gdb_byte *) remote_exec_file, rs->buf + len,

                      strlen (remote_exec_file));



  gdb_assert (args != NULL);

  if (*args)

    {

      struct cleanup *back_to;

      int i;

      char **argv;



      argv = gdb_buildargv (args);

      back_to = make_cleanup_freeargv (argv);

      for (i = 0; argv[i] != NULL; i++)

        {

          if (strlen (argv[i]) * 2 + 1 + len >= get_remote_packet_size ())

            error (_("Argument list too long for run packet"));

          rs->buf[len++] = ';';

          len += 2 * bin2hex ((gdb_byte *) argv[i], rs->buf + len,

                              strlen (argv[i]));

        }

      do_cleanups (back_to);

    }



  rs->buf[len++] = '\0';



  putpkt (rs->buf);

  getpkt (&rs->buf, &rs->buf_size, 0);



  switch (packet_ok (rs->buf, &remote_protocol_packets[PACKET_vRun]))

    {

    case PACKET_OK:

      /* We have a wait response.  All is well.  */

      return 0;

    case PACKET_UNKNOWN:

      return -1;

    case PACKET_ERROR:

      if (remote_exec_file[0] == '\0')

        error (_("Running the default executable on the remote target failed; "

                 "try \"set remote exec-file\"?"));

      else

        error (_("Running \"%s\" on the remote target failed"),

               remote_exec_file);

    default:

      gdb_assert_not_reached (_("bad switch"));

    }

}



/* In the extended protocol we want to be able to do things like

   "run" and have them basically work as expected.  So we need

   a special create_inferior function.  We support changing the

   executable file and the command line arguments, but not the

   environment.  */



static void

‌extended_remote_create_inferior (struct target_ops *ops,

                                 char *exec_file, char *args,

                                 char **env, int from_tty)

{

  int run_worked;

  char *stop_reply;

  struct remote_state *rs = get_remote_state ();



  /* If running asynchronously, register the target file descriptor

     with the event loop.  */

  if (target_can_async_p ())

    target_async (inferior_event_handler, 0);



  /* Disable address space randomization if requested (and supported).  */

  if (extended_remote_supports_disable_randomization (ops))

    extended_remote_disable_randomization (disable_randomization);



  /* Now restart the remote server.  */

  run_worked = extended_remote_run (args) != -1;

  if (!run_worked)

    {

      /* vRun was not supported.  Fail if we need it to do what the

         user requested.  */

      if (remote_exec_file[0])

        error (_("Remote target does not support \"set remote exec-file\""));

      if (args[0])

        error (_("Remote target does not support \"set args\" or run <ARGS>"));



      /* Fall back to "R".  */

      extended_remote_restart ();

    }



  if (!have_inferiors ())

    {

      /* Clean up from the last time we ran, before we mark the target

         running again.  This will mark breakpoints uninserted, and

         get_offsets may insert breakpoints.  */

      init_thread_list ();

      init_wait_for_inferior ();

    }



  /* vRun's success return is a stop reply.  */

  stop_reply = run_worked ? rs->buf : NULL;

  add_current_inferior_and_thread (stop_reply);



  /* Get updated offsets, if the stub uses qOffsets.  */

  get_offsets ();

}





/* Given a location's target info BP_TGT and the packet buffer BUF,  output

   the list of conditions (in agent expression bytecode format), if any, the

   target needs to evaluate.  The output is placed into the packet buffer

   started from BUF and ended at BUF_END.  */



static int

‌remote_add_target_side_condition (struct gdbarch *gdbarch,

                                  struct bp_target_info *bp_tgt, char *buf,

                                  char *buf_end)

{

  struct agent_expr *aexpr = NULL;

  int i, ix;

  char *pkt;

  char *buf_start = buf;



  if (VEC_empty (agent_expr_p, bp_tgt->conditions))

    return 0;



  buf += strlen (buf);

  xsnprintf (buf, buf_end - buf, "%s", ";");

  buf++;



  /* Send conditions to the target and free the vector.  */

  for (ix = 0;

       VEC_iterate (agent_expr_p, bp_tgt->conditions, ix, aexpr);

       ix++)

    {

      xsnprintf (buf, buf_end - buf, "X%x,", aexpr->len);

      buf += strlen (buf);

      for (i = 0; i < aexpr->len; ++i)

        buf = pack_hex_byte (buf, aexpr->buf[i]);

      *buf = '\0';

    }

  return 0;

}



static void

‌remote_add_target_side_commands (struct gdbarch *gdbarch,

                                 struct bp_target_info *bp_tgt, char *buf)

{

  struct agent_expr *aexpr = NULL;

  int i, ix;



  if (VEC_empty (agent_expr_p, bp_tgt->tcommands))

    return;



  buf += strlen (buf);



  sprintf (buf, ";cmds:%x,", bp_tgt->persist);

  buf += strlen (buf);



  /* Concatenate all the agent expressions that are commands into the

     cmds parameter.  */

  for (ix = 0;

       VEC_iterate (agent_expr_p, bp_tgt->tcommands, ix, aexpr);

       ix++)

    {

      sprintf (buf, "X%x,", aexpr->len);

      buf += strlen (buf);

      for (i = 0; i < aexpr->len; ++i)

        buf = pack_hex_byte (buf, aexpr->buf[i]);

      *buf = '\0';

    }

}



/* Insert a breakpoint.  On targets that have software breakpoint

   support, we ask the remote target to do the work; on targets

   which don't, we insert a traditional memory breakpoint.  */



static int

‌remote_insert_breakpoint (struct target_ops *ops,

                          struct gdbarch *gdbarch,

                          struct bp_target_info *bp_tgt)

{

  /* Try the "Z" s/w breakpoint packet if it is not already disabled.

     If it succeeds, then set the support to PACKET_ENABLE.  If it

     fails, and the user has explicitly requested the Z support then

     report an error, otherwise, mark it disabled and go on.  */



  if (packet_support (PACKET_Z0) != PACKET_DISABLE)

    {

      CORE_ADDR addr = bp_tgt->reqstd_address;

      struct remote_state *rs;

      char *p, *endbuf;

      int bpsize;

      struct condition_list *cond = NULL;



      /* Make sure the remote is pointing at the right process, if

         necessary.  */

      if (!gdbarch_has_global_breakpoints (target_gdbarch ()))

        set_general_process ();



      gdbarch_remote_breakpoint_from_pc (gdbarch, &addr, &bpsize);



      rs = get_remote_state ();

      p = rs->buf;

      endbuf = rs->buf + get_remote_packet_size ();



      *(p++) = 'Z';

      *(p++) = '0';

      *(p++) = ',';

      addr = (ULONGEST) remote_address_masked (addr);

      p += hexnumstr (p, addr);

      xsnprintf (p, endbuf - p, ",%d", bpsize);



      if (remote_supports_cond_breakpoints (ops))

        remote_add_target_side_condition (gdbarch, bp_tgt, p, endbuf);



      if (remote_can_run_breakpoint_commands (ops))

        remote_add_target_side_commands (gdbarch, bp_tgt, p);



      putpkt (rs->buf);

      getpkt (&rs->buf, &rs->buf_size, 0);



      switch (packet_ok (rs->buf, &remote_protocol_packets[PACKET_Z0]))

        {

        case PACKET_ERROR:

          return -1;

        case PACKET_OK:

          bp_tgt->placed_address = addr;

          bp_tgt->placed_size = bpsize;

          return 0;

        case PACKET_UNKNOWN:

          break;

        }

    }



  /* If this breakpoint has target-side commands but this stub doesn't

     support Z0 packets, throw error.  */

  if (!VEC_empty (agent_expr_p, bp_tgt->tcommands))

    throw_error (NOT_SUPPORTED_ERROR, _("\

Target doesn't support breakpoints that have target side commands."));



  return memory_insert_breakpoint (ops, gdbarch, bp_tgt);

}



static int

‌remote_remove_breakpoint (struct target_ops *ops,

                          struct gdbarch *gdbarch,

                          struct bp_target_info *bp_tgt)

{

  CORE_ADDR addr = bp_tgt->placed_address;

  struct remote_state *rs = get_remote_state ();



  if (packet_support (PACKET_Z0) != PACKET_DISABLE)

    {

      char *p = rs->buf;

      char *endbuf = rs->buf + get_remote_packet_size ();



      /* Make sure the remote is pointing at the right process, if

         necessary.  */

      if (!gdbarch_has_global_breakpoints (target_gdbarch ()))

        set_general_process ();



      *(p++) = 'z';

      *(p++) = '0';

      *(p++) = ',';



      addr = (ULONGEST) remote_address_masked (bp_tgt->placed_address);

      p += hexnumstr (p, addr);

      xsnprintf (p, endbuf - p, ",%d", bp_tgt->placed_size);



      putpkt (rs->buf);

      getpkt (&rs->buf, &rs->buf_size, 0);



      return (rs->buf[0] == 'E');

    }



  return memory_remove_breakpoint (ops, gdbarch, bp_tgt);

}



static int

‌watchpoint_to_Z_packet (int type)

{

  switch (type)

    {

    case hw_write:

      return Z_PACKET_WRITE_WP;

      break;

    case hw_read:

      return Z_PACKET_READ_WP;

      break;

    case hw_access:

      return Z_PACKET_ACCESS_WP;

      break;

    default:

      internal_error (__FILE__, __LINE__,

                      _("hw_bp_to_z: bad watchpoint type %d"), type);

    }

}



static int

‌remote_insert_watchpoint (struct target_ops *self,

                          CORE_ADDR addr, int len, int type,

                          struct expression *cond)

{

  struct remote_state *rs = get_remote_state ();

  char *endbuf = rs->buf + get_remote_packet_size ();

  char *p;

  enum Z_packet_type packet = watchpoint_to_Z_packet (type);



  if (packet_support (PACKET_Z0 + packet) == PACKET_DISABLE)

    return 1;



  /* Make sure the remote is pointing at the right process, if

     necessary.  */

  if (!gdbarch_has_global_breakpoints (target_gdbarch ()))

    set_general_process ();



  xsnprintf (rs->buf, endbuf - rs->buf, "Z%x,", packet);

  p = strchr (rs->buf, '\0');

  addr = remote_address_masked (addr);

  p += hexnumstr (p, (ULONGEST) addr);

  xsnprintf (p, endbuf - p, ",%x", len);



  putpkt (rs->buf);

  getpkt (&rs->buf, &rs->buf_size, 0);



  switch (packet_ok (rs->buf, &remote_protocol_packets[PACKET_Z0 + packet]))

    {

    case PACKET_ERROR:

      return -1;

    case PACKET_UNKNOWN:

      return 1;

    case PACKET_OK:

      return 0;

    }

  internal_error (__FILE__, __LINE__,

                  _("remote_insert_watchpoint: reached end of function"));

}



static int

‌remote_watchpoint_addr_within_range (struct target_ops *target, CORE_ADDR addr,

                                     CORE_ADDR start, int length)

{

  CORE_ADDR diff = remote_address_masked (addr - start);



  return diff < length;

}





static int

‌remote_remove_watchpoint (struct target_ops *self,

                          CORE_ADDR addr, int len, int type,

                          struct expression *cond)

{

  struct remote_state *rs = get_remote_state ();

  char *endbuf = rs->buf + get_remote_packet_size ();

  char *p;

  enum Z_packet_type packet = watchpoint_to_Z_packet (type);



  if (packet_support (PACKET_Z0 + packet) == PACKET_DISABLE)

    return -1;



  /* Make sure the remote is pointing at the right process, if

     necessary.  */

  if (!gdbarch_has_global_breakpoints (target_gdbarch ()))

    set_general_process ();



  xsnprintf (rs->buf, endbuf - rs->buf, "z%x,", packet);

  p = strchr (rs->buf, '\0');

  addr = remote_address_masked (addr);

  p += hexnumstr (p, (ULONGEST) addr);

  xsnprintf (p, endbuf - p, ",%x", len);

  putpkt (rs->buf);

  getpkt (&rs->buf, &rs->buf_size, 0);



  switch (packet_ok (rs->buf, &remote_protocol_packets[PACKET_Z0 + packet]))

    {

    case PACKET_ERROR:

    case PACKET_UNKNOWN:

      return -1;

    case PACKET_OK:

      return 0;

    }

  internal_error (__FILE__, __LINE__,

                  _("remote_remove_watchpoint: reached end of function"));

}





‌int remote_hw_watchpoint_limit = -1;

‌int remote_hw_watchpoint_length_limit = -1;

‌int remote_hw_breakpoint_limit = -1;



static int

‌remote_region_ok_for_hw_watchpoint (struct target_ops *self,

                                    CORE_ADDR addr, int len)

{

  if (remote_hw_watchpoint_length_limit == 0)

    return 0;

  else if (remote_hw_watchpoint_length_limit < 0)

    return 1;

  else if (len <= remote_hw_watchpoint_length_limit)

    return 1;

  else

    return 0;

}



static int

‌remote_check_watch_resources (struct target_ops *self,

                              int type, int cnt, int ot)

{

  if (type == bp_hardware_breakpoint)

    {

      if (remote_hw_breakpoint_limit == 0)

        return 0;

      else if (remote_hw_breakpoint_limit < 0)

        return 1;

      else if (cnt <= remote_hw_breakpoint_limit)

        return 1;

    }

  else

    {

      if (remote_hw_watchpoint_limit == 0)

        return 0;

      else if (remote_hw_watchpoint_limit < 0)

        return 1;

      else if (ot)

        return -1;

      else if (cnt <= remote_hw_watchpoint_limit)

        return 1;

    }

  return -1;

}



static int

‌remote_stopped_by_watchpoint (struct target_ops *ops)

{

  struct remote_state *rs = get_remote_state ();



  return rs->remote_stopped_by_watchpoint_p;

}



static int

‌remote_stopped_data_address (struct target_ops *target, CORE_ADDR *addr_p)

{

  struct remote_state *rs = get_remote_state ();

  int rc = 0;



  if (remote_stopped_by_watchpoint (target))

    {

      *addr_p = rs->remote_watch_data_address;

      rc = 1;

    }



  return rc;

}





static int

‌remote_insert_hw_breakpoint (struct target_ops *self, struct gdbarch *gdbarch,

                             struct bp_target_info *bp_tgt)

{

  CORE_ADDR addr = bp_tgt->reqstd_address;

  struct remote_state *rs;

  char *p, *endbuf;

  char *message;

  int bpsize;



  /* The length field should be set to the size of a breakpoint

     instruction, even though we aren't inserting one ourselves.  */



  gdbarch_remote_breakpoint_from_pc (gdbarch, &addr, &bpsize);



  if (packet_support (PACKET_Z1) == PACKET_DISABLE)

    return -1;



  /* Make sure the remote is pointing at the right process, if

     necessary.  */

  if (!gdbarch_has_global_breakpoints (target_gdbarch ()))

    set_general_process ();



  rs = get_remote_state ();

  p = rs->buf;

  endbuf = rs->buf + get_remote_packet_size ();



  *(p++) = 'Z';

  *(p++) = '1';

  *(p++) = ',';



  addr = remote_address_masked (addr);

  p += hexnumstr (p, (ULONGEST) addr);

  xsnprintf (p, endbuf - p, ",%x", bpsize);



  if (remote_supports_cond_breakpoints (self))

    remote_add_target_side_condition (gdbarch, bp_tgt, p, endbuf);



  if (remote_can_run_breakpoint_commands (self))

    remote_add_target_side_commands (gdbarch, bp_tgt, p);



  putpkt (rs->buf);

  getpkt (&rs->buf, &rs->buf_size, 0);



  switch (packet_ok (rs->buf, &remote_protocol_packets[PACKET_Z1]))

    {

    case PACKET_ERROR:

      if (rs->buf[1] == '.')

        {

          message = strchr (rs->buf + 2, '.');

          if (message)

            error (_("Remote failure reply: %s"), message + 1);

        }

      return -1;

    case PACKET_UNKNOWN:

      return -1;

    case PACKET_OK:

      bp_tgt->placed_address = addr;

      bp_tgt->placed_size = bpsize;

      return 0;

    }

  internal_error (__FILE__, __LINE__,

                  _("remote_insert_hw_breakpoint: reached end of function"));

}





static int

‌remote_remove_hw_breakpoint (struct target_ops *self, struct gdbarch *gdbarch,

                             struct bp_target_info *bp_tgt)

{

  CORE_ADDR addr;

  struct remote_state *rs = get_remote_state ();

  char *p = rs->buf;

  char *endbuf = rs->buf + get_remote_packet_size ();



  if (packet_support (PACKET_Z1) == PACKET_DISABLE)

    return -1;



  /* Make sure the remote is pointing at the right process, if

     necessary.  */

  if (!gdbarch_has_global_breakpoints (target_gdbarch ()))

    set_general_process ();



  *(p++) = 'z';

  *(p++) = '1';

  *(p++) = ',';



  addr = remote_address_masked (bp_tgt->placed_address);

  p += hexnumstr (p, (ULONGEST) addr);

  xsnprintf (p, endbuf  - p, ",%x", bp_tgt->placed_size);



  putpkt (rs->buf);

  getpkt (&rs->buf, &rs->buf_size, 0);



  switch (packet_ok (rs->buf, &remote_protocol_packets[PACKET_Z1]))

    {

    case PACKET_ERROR:

    case PACKET_UNKNOWN:

      return -1;

    case PACKET_OK:

      return 0;

    }

  internal_error (__FILE__, __LINE__,

                  _("remote_remove_hw_breakpoint: reached end of function"));

}



/* Verify memory using the "qCRC:" request.  */



static int

‌remote_verify_memory (struct target_ops *ops,

                      const gdb_byte *data, CORE_ADDR lma, ULONGEST size)

{

  struct remote_state *rs = get_remote_state ();

  unsigned long host_crc, target_crc;

  char *tmp;



  /* It doesn't make sense to use qCRC if the remote target is

     connected but not running.  */

  if (target_has_execution && packet_support (PACKET_qCRC) != PACKET_DISABLE)

    {

      enum packet_result result;



      /* Make sure the remote is pointing at the right process.  */

      set_general_process ();



      /* FIXME: assumes lma can fit into long.  */

      xsnprintf (rs->buf, get_remote_packet_size (), "qCRC:%lx,%lx",

                 (long) lma, (long) size);

      putpkt (rs->buf);



      /* Be clever; compute the host_crc before waiting for target

         reply.  */

      host_crc = xcrc32 (data, size, 0xffffffff);



      getpkt (&rs->buf, &rs->buf_size, 0);



      result = packet_ok (rs->buf,

                          &remote_protocol_packets[PACKET_qCRC]);

      if (result == PACKET_ERROR)

        return -1;

      else if (result == PACKET_OK)

        {

          for (target_crc = 0, tmp = &rs->buf[1]; *tmp; tmp++)

            target_crc = target_crc * 16 + fromhex (*tmp);



          return (host_crc == target_crc);

        }

    }



  return simple_verify_memory (ops, data, lma, size);

}



/* compare-sections command



   With no arguments, compares each loadable section in the exec bfd

   with the same memory range on the target, and reports mismatches.

   Useful for verifying the image on the target against the exec file.  */



static void

‌compare_sections_command (char *args, int from_tty)

{

  asection *s;

  struct cleanup *old_chain;

  gdb_byte *sectdata;

  const char *sectname;

  bfd_size_type size;

  bfd_vma lma;

  int matched = 0;

  int mismatched = 0;

  int res;

  int read_only = 0;



  if (!exec_bfd)

    error (_("command cannot be used without an exec file"));



  /* Make sure the remote is pointing at the right process.  */

  set_general_process ();



  if (args != NULL && strcmp (args, "-r") == 0)

    {

      read_only = 1;

      args = NULL;

    }



  for (s = exec_bfd->sections; s; s = s->next)

    {

      if (!(s->flags & SEC_LOAD))

        continue;                /* Skip non-loadable section.  */



      if (read_only && (s->flags & SEC_READONLY) == 0)

        continue;                /* Skip writeable sections */



      size = bfd_get_section_size (s);

      if (size == 0)

        continue;                /* Skip zero-length section.  */



      sectname = bfd_get_section_name (exec_bfd, s);

      if (args && strcmp (args, sectname) != 0)

        continue;                /* Not the section selected by user.  */



      matched = 1;                /* Do this section.  */

      lma = s->lma;



      sectdata = xmalloc (size);

      old_chain = make_cleanup (xfree, sectdata);

      bfd_get_section_contents (exec_bfd, s, sectdata, 0, size);



      res = target_verify_memory (sectdata, lma, size);



      if (res == -1)

        error (_("target memory fault, section %s, range %s -- %s"), sectname,

               paddress (target_gdbarch (), lma),

               paddress (target_gdbarch (), lma + size));



      printf_filtered ("Section %s, range %s -- %s: ", sectname,

                       paddress (target_gdbarch (), lma),

                       paddress (target_gdbarch (), lma + size));

      if (res)

        printf_filtered ("matched.\n");

      else

        {

          printf_filtered ("MIS-MATCHED!\n");

          mismatched++;

        }



      do_cleanups (old_chain);

    }

  if (mismatched > 0)

    warning (_("One or more sections of the target image does not match\n\

the loaded file\n"));

  if (args && !matched)

    printf_filtered (_("No loaded section named '%s'.\n"), args);

}



/* Write LEN bytes from WRITEBUF into OBJECT_NAME/ANNEX at OFFSET

   into remote target.  The number of bytes written to the remote

   target is returned, or -1 for error.  */



static enum target_xfer_status

‌remote_write_qxfer (struct target_ops *ops, const char *object_name,

                    const char *annex, const gdb_byte *writebuf,

                    ULONGEST offset, LONGEST len, ULONGEST *xfered_len,

                    struct packet_config *packet)

{

  int i, buf_len;

  ULONGEST n;

  struct remote_state *rs = get_remote_state ();

  int max_size = get_memory_write_packet_size ();



  if (packet->support == PACKET_DISABLE)

    return TARGET_XFER_E_IO;



  /* Insert header.  */

  i = snprintf (rs->buf, max_size,

                "qXfer:%s:write:%s:%s:",

                object_name, annex ? annex : "",

                phex_nz (offset, sizeof offset));

  max_size -= (i + 1);



  /* Escape as much data as fits into rs->buf.  */

  buf_len = remote_escape_output

    (writebuf, len, (gdb_byte *) rs->buf + i, &max_size, max_size);



  if (putpkt_binary (rs->buf, i + buf_len) < 0

      || getpkt_sane (&rs->buf, &rs->buf_size, 0) < 0

      || packet_ok (rs->buf, packet) != PACKET_OK)

    return TARGET_XFER_E_IO;



  unpack_varlen_hex (rs->buf, &n);



  *xfered_len = n;

  return TARGET_XFER_OK;

}



/* Read OBJECT_NAME/ANNEX from the remote target using a qXfer packet.

   Data at OFFSET, of up to LEN bytes, is read into READBUF; the

   number of bytes read is returned, or 0 for EOF, or -1 for error.

   The number of bytes read may be less than LEN without indicating an

   EOF.  PACKET is checked and updated to indicate whether the remote

   target supports this object.  */



static enum target_xfer_status

‌remote_read_qxfer (struct target_ops *ops, const char *object_name,

                   const char *annex,

                   gdb_byte *readbuf, ULONGEST offset, LONGEST len,

                   ULONGEST *xfered_len,

                   struct packet_config *packet)

{

  struct remote_state *rs = get_remote_state ();

  LONGEST i, n, packet_len;



  if (packet->support == PACKET_DISABLE)

    return TARGET_XFER_E_IO;



  /* Check whether we've cached an end-of-object packet that matches

     this request.  */

  if (rs->finished_object)

    {

      if (strcmp (object_name, rs->finished_object) == 0

          && strcmp (annex ? annex : "", rs->finished_annex) == 0

          && offset == rs->finished_offset)

        return TARGET_XFER_EOF;





      /* Otherwise, we're now reading something different.  Discard

         the cache.  */

      xfree (rs->finished_object);

      xfree (rs->finished_annex);

      rs->finished_object = NULL;

      rs->finished_annex = NULL;

    }



  /* Request only enough to fit in a single packet.  The actual data

     may not, since we don't know how much of it will need to be escaped;

     the target is free to respond with slightly less data.  We subtract

     five to account for the response type and the protocol frame.  */

  n = min (get_remote_packet_size () - 5, len);

  snprintf (rs->buf, get_remote_packet_size () - 4, "qXfer:%s:read:%s:%s,%s",

            object_name, annex ? annex : "",

            phex_nz (offset, sizeof offset),

            phex_nz (n, sizeof n));

  i = putpkt (rs->buf);

  if (i < 0)

    return TARGET_XFER_E_IO;



  rs->buf[0] = '\0';

  packet_len = getpkt_sane (&rs->buf, &rs->buf_size, 0);

  if (packet_len < 0 || packet_ok (rs->buf, packet) != PACKET_OK)

    return TARGET_XFER_E_IO;



  if (rs->buf[0] != 'l' && rs->buf[0] != 'm')

    error (_("Unknown remote qXfer reply: %s"), rs->buf);



  /* 'm' means there is (or at least might be) more data after this

     batch.  That does not make sense unless there's at least one byte

     of data in this reply.  */

  if (rs->buf[0] == 'm' && packet_len == 1)

    error (_("Remote qXfer reply contained no data."));



  /* Got some data.  */

  i = remote_unescape_input ((gdb_byte *) rs->buf + 1,

                             packet_len - 1, readbuf, n);



  /* 'l' is an EOF marker, possibly including a final block of data,

     or possibly empty.  If we have the final block of a non-empty

     object, record this fact to bypass a subsequent partial read.  */

  if (rs->buf[0] == 'l' && offset + i > 0)

    {

      rs->finished_object = xstrdup (object_name);

      rs->finished_annex = xstrdup (annex ? annex : "");

      rs->finished_offset = offset + i;

    }



  if (i == 0)

    return TARGET_XFER_EOF;

  else

    {

      *xfered_len = i;

      return TARGET_XFER_OK;

    }

}



static enum target_xfer_status

‌remote_xfer_partial (struct target_ops *ops, enum target_object object,

                     const char *annex, gdb_byte *readbuf,

                     const gdb_byte *writebuf, ULONGEST offset, ULONGEST len,

                     ULONGEST *xfered_len)

{

  struct remote_state *rs;

  int i;

  char *p2;

  char query_type;



  set_remote_traceframe ();

  set_general_thread (inferior_ptid);



  rs = get_remote_state ();



  /* Handle memory using the standard memory routines.  */

  if (object == TARGET_OBJECT_MEMORY)

    {

      /* If the remote target is connected but not running, we should

         pass this request down to a lower stratum (e.g. the executable

         file).  */

      if (!target_has_execution)

        return TARGET_XFER_EOF;



      if (writebuf != NULL)

        return remote_write_bytes (offset, writebuf, len, xfered_len);

      else

        return remote_read_bytes (ops, offset, readbuf, len, xfered_len);

    }



  /* Handle SPU memory using qxfer packets.  */

  if (object == TARGET_OBJECT_SPU)

    {

      if (readbuf)

        return remote_read_qxfer (ops, "spu", annex, readbuf, offset, len,

                                  xfered_len, &remote_protocol_packets

                                  [PACKET_qXfer_spu_read]);

      else

        return remote_write_qxfer (ops, "spu", annex, writebuf, offset, len,

                                   xfered_len, &remote_protocol_packets

                                   [PACKET_qXfer_spu_write]);

    }



  /* Handle extra signal info using qxfer packets.  */

  if (object == TARGET_OBJECT_SIGNAL_INFO)

    {

      if (readbuf)

        return remote_read_qxfer (ops, "siginfo", annex, readbuf, offset, len,

                                  xfered_len, &remote_protocol_packets

                                  [PACKET_qXfer_siginfo_read]);

      else

        return remote_write_qxfer (ops, "siginfo", annex,

                                   writebuf, offset, len, xfered_len,

                                   &remote_protocol_packets

                                   [PACKET_qXfer_siginfo_write]);

    }



  if (object == TARGET_OBJECT_STATIC_TRACE_DATA)

    {

      if (readbuf)

        return remote_read_qxfer (ops, "statictrace", annex,

                                  readbuf, offset, len, xfered_len,

                                  &remote_protocol_packets

                                  [PACKET_qXfer_statictrace_read]);

      else

        return TARGET_XFER_E_IO;

    }



  /* Only handle flash writes.  */

  if (writebuf != NULL)

    {

      LONGEST xfered;



      switch (object)

        {

        case TARGET_OBJECT_FLASH:

          return remote_flash_write (ops, offset, len, xfered_len,

                                     writebuf);



        default:

          return TARGET_XFER_E_IO;

        }

    }



  /* Map pre-existing objects onto letters.  DO NOT do this for new

     objects!!!  Instead specify new query packets.  */

  switch (object)

    {

    case TARGET_OBJECT_AVR:

      query_type = 'R';

      break;



    case TARGET_OBJECT_AUXV:

      gdb_assert (annex == NULL);

      return remote_read_qxfer (ops, "auxv", annex, readbuf, offset, len,

                                xfered_len,

                                &remote_protocol_packets[PACKET_qXfer_auxv]);



    case TARGET_OBJECT_AVAILABLE_FEATURES:

      return remote_read_qxfer

        (ops, "features", annex, readbuf, offset, len, xfered_len,

         &remote_protocol_packets[PACKET_qXfer_features]);



    case TARGET_OBJECT_LIBRARIES:

      return remote_read_qxfer

        (ops, "libraries", annex, readbuf, offset, len, xfered_len,

         &remote_protocol_packets[PACKET_qXfer_libraries]);



    case TARGET_OBJECT_LIBRARIES_SVR4:

      return remote_read_qxfer

        (ops, "libraries-svr4", annex, readbuf, offset, len, xfered_len,

         &remote_protocol_packets[PACKET_qXfer_libraries_svr4]);



    case TARGET_OBJECT_MEMORY_MAP:

      gdb_assert (annex == NULL);

      return remote_read_qxfer (ops, "memory-map", annex, readbuf, offset, len,

                                 xfered_len,

                                &remote_protocol_packets[PACKET_qXfer_memory_map]);



    case TARGET_OBJECT_OSDATA:

      /* Should only get here if we're connected.  */

      gdb_assert (rs->remote_desc);

      return remote_read_qxfer

        (ops, "osdata", annex, readbuf, offset, len, xfered_len,

        &remote_protocol_packets[PACKET_qXfer_osdata]);



    case TARGET_OBJECT_THREADS:

      gdb_assert (annex == NULL);

      return remote_read_qxfer (ops, "threads", annex, readbuf, offset, len,

                                xfered_len,

                                &remote_protocol_packets[PACKET_qXfer_threads]);



    case TARGET_OBJECT_TRACEFRAME_INFO:

      gdb_assert (annex == NULL);

      return remote_read_qxfer

        (ops, "traceframe-info", annex, readbuf, offset, len, xfered_len,

         &remote_protocol_packets[PACKET_qXfer_traceframe_info]);



    case TARGET_OBJECT_FDPIC:

      return remote_read_qxfer (ops, "fdpic", annex, readbuf, offset, len,

                                xfered_len,

                                &remote_protocol_packets[PACKET_qXfer_fdpic]);



    case TARGET_OBJECT_OPENVMS_UIB:

      return remote_read_qxfer (ops, "uib", annex, readbuf, offset, len,

                                xfered_len,

                                &remote_protocol_packets[PACKET_qXfer_uib]);



    case TARGET_OBJECT_BTRACE:

      return remote_read_qxfer (ops, "btrace", annex, readbuf, offset, len,

                                xfered_len,

        &remote_protocol_packets[PACKET_qXfer_btrace]);



    default:

      return TARGET_XFER_E_IO;

    }



  /* Minimum outbuf size is get_remote_packet_size ().  If LEN is not

     large enough let the caller deal with it.  */

  if (len < get_remote_packet_size ())

    return TARGET_XFER_E_IO;

  len = get_remote_packet_size ();



  /* Except for querying the minimum buffer size, target must be open.  */

  if (!rs->remote_desc)

    error (_("remote query is only available after target open"));



  gdb_assert (annex != NULL);

  gdb_assert (readbuf != NULL);



  p2 = rs->buf;

  *p2++ = 'q';

  *p2++ = query_type;



  /* We used one buffer char for the remote protocol q command and

     another for the query type.  As the remote protocol encapsulation

     uses 4 chars plus one extra in case we are debugging

     (remote_debug), we have PBUFZIZ - 7 left to pack the query

     string.  */

  i = 0;

  while (annex[i] && (i < (get_remote_packet_size () - 8)))

    {

      /* Bad caller may have sent forbidden characters.  */

      gdb_assert (isprint (annex[i]) && annex[i] != '$' && annex[i] != '#');

      *p2++ = annex[i];

      i++;

    }

  *p2 = '\0';

  gdb_assert (annex[i] == '\0');



  i = putpkt (rs->buf);

  if (i < 0)

    return TARGET_XFER_E_IO;



  getpkt (&rs->buf, &rs->buf_size, 0);

  strcpy ((char *) readbuf, rs->buf);



  *xfered_len = strlen ((char *) readbuf);

  return TARGET_XFER_OK;

}



static int

‌remote_search_memory (struct target_ops* ops,

                      CORE_ADDR start_addr, ULONGEST search_space_len,

                      const gdb_byte *pattern, ULONGEST pattern_len,

                      CORE_ADDR *found_addrp)

{

  int addr_size = gdbarch_addr_bit (target_gdbarch ()) / 8;

  struct remote_state *rs = get_remote_state ();

  int max_size = get_memory_write_packet_size ();

  struct packet_config *packet =

    &remote_protocol_packets[PACKET_qSearch_memory];

  /* Number of packet bytes used to encode the pattern;

     this could be more than PATTERN_LEN due to escape characters.  */

  int escaped_pattern_len;

  /* Amount of pattern that was encodable in the packet.  */

  int used_pattern_len;

  int i;

  int found;

  ULONGEST found_addr;



  /* Don't go to the target if we don't have to.

     This is done before checking packet->support to avoid the possibility that

     a success for this edge case means the facility works in general.  */

  if (pattern_len > search_space_len)

    return 0;

  if (pattern_len == 0)

    {

      *found_addrp = start_addr;

      return 1;

    }



  /* If we already know the packet isn't supported, fall back to the simple

     way of searching memory.  */



  if (packet_config_support (packet) == PACKET_DISABLE)

    {

      /* Target doesn't provided special support, fall back and use the

         standard support (copy memory and do the search here).  */

      return simple_search_memory (ops, start_addr, search_space_len,

                                   pattern, pattern_len, found_addrp);

    }



  /* Make sure the remote is pointing at the right process.  */

  set_general_process ();



  /* Insert header.  */

  i = snprintf (rs->buf, max_size,

                "qSearch:memory:%s;%s;",

                phex_nz (start_addr, addr_size),

                phex_nz (search_space_len, sizeof (search_space_len)));

  max_size -= (i + 1);



  /* Escape as much data as fits into rs->buf.  */

  escaped_pattern_len =

    remote_escape_output (pattern, pattern_len, (gdb_byte *) rs->buf + i,

                          &used_pattern_len, max_size);



  /* Bail if the pattern is too large.  */

  if (used_pattern_len != pattern_len)

    error (_("Pattern is too large to transmit to remote target."));



  if (putpkt_binary (rs->buf, i + escaped_pattern_len) < 0

      || getpkt_sane (&rs->buf, &rs->buf_size, 0) < 0

      || packet_ok (rs->buf, packet) != PACKET_OK)

    {

      /* The request may not have worked because the command is not

         supported.  If so, fall back to the simple way.  */

      if (packet->support == PACKET_DISABLE)

        {

          return simple_search_memory (ops, start_addr, search_space_len,

                                       pattern, pattern_len, found_addrp);

        }

      return -1;

    }



  if (rs->buf[0] == '0')

    found = 0;

  else if (rs->buf[0] == '1')

    {

      found = 1;

      if (rs->buf[1] != ',')

        error (_("Unknown qSearch:memory reply: %s"), rs->buf);

      unpack_varlen_hex (rs->buf + 2, &found_addr);

      *found_addrp = found_addr;

    }

  else

    error (_("Unknown qSearch:memory reply: %s"), rs->buf);



  return found;

}



static void

‌remote_rcmd (struct target_ops *self, const char *command,

             struct ui_file *outbuf)

{

  struct remote_state *rs = get_remote_state ();

  char *p = rs->buf;



  if (!rs->remote_desc)

    error (_("remote rcmd is only available after target open"));



  /* Send a NULL command across as an empty command.  */

  if (command == NULL)

    command = "";



  /* The query prefix.  */

  strcpy (rs->buf, "qRcmd,");

  p = strchr (rs->buf, '\0');



  if ((strlen (rs->buf) + strlen (command) * 2 + 8/*misc*/)

      > get_remote_packet_size ())

    error (_("\"monitor\" command ``%s'' is too long."), command);



  /* Encode the actual command.  */

  bin2hex ((const gdb_byte *) command, p, strlen (command));



  if (putpkt (rs->buf) < 0)

    error (_("Communication problem with target."));



  /* get/display the response */

  while (1)

    {

      char *buf;



      /* XXX - see also remote_get_noisy_reply().  */

      QUIT;                        /* Allow user to bail out with ^C.  */

      rs->buf[0] = '\0';

      if (getpkt_sane (&rs->buf, &rs->buf_size, 0) == -1)

        {

          /* Timeout.  Continue to (try to) read responses.

             This is better than stopping with an error, assuming the stub

             is still executing the (long) monitor command.

             If needed, the user can interrupt gdb using C-c, obtaining

             an effect similar to stop on timeout.  */

          continue;

        }

      buf = rs->buf;

      if (buf[0] == '\0')

        error (_("Target does not support this command."));

      if (buf[0] == 'O' && buf[1] != 'K')

        {

          remote_console_output (buf + 1); /* 'O' message from stub.  */

          continue;

        }

      if (strcmp (buf, "OK") == 0)

        break;

      if (strlen (buf) == 3 && buf[0] == 'E'

          && isdigit (buf[1]) && isdigit (buf[2]))

        {

          error (_("Protocol error with Rcmd"));

        }

      for (p = buf; p[0] != '\0' && p[1] != '\0'; p += 2)

        {

          char c = (fromhex (p[0]) << 4) + fromhex (p[1]);



          fputc_unfiltered (c, outbuf);

        }

      break;

    }

}



‌static VEC(mem_region_s) *

remote_memory_map (struct target_ops *ops)

{

  VEC(mem_region_s) *result = NULL;

  char *text = target_read_stralloc (&current_target,

                                     TARGET_OBJECT_MEMORY_MAP, NULL);



  if (text)

    {

      struct cleanup *back_to = make_cleanup (xfree, text);



      result = parse_memory_map (text);

      do_cleanups (back_to);

    }



  return result;

}



static void

‌packet_command (char *args, int from_tty)

{

  struct remote_state *rs = get_remote_state ();



  if (!rs->remote_desc)

    error (_("command can only be used with remote target"));



  if (!args)

    error (_("remote-packet command requires packet text as argument"));



  puts_filtered ("sending: ");

  print_packet (args);

  puts_filtered ("\n");

  putpkt (args);



  getpkt (&rs->buf, &rs->buf_size, 0);

  puts_filtered ("received: ");

  print_packet (rs->buf);

  puts_filtered ("\n");

}



#if 0

/* --------- UNIT_TEST for THREAD oriented PACKETS ------------------- */



static void display_thread_info (struct gdb_ext_thread_info *info);



static void threadset_test_cmd (char *cmd, int tty);



static void threadalive_test (char *cmd, int tty);



static void threadlist_test_cmd (char *cmd, int tty);



int get_and_display_threadinfo (threadref *ref);



static void threadinfo_test_cmd (char *cmd, int tty);



static int thread_display_step (threadref *ref, void *context);



static void threadlist_update_test_cmd (char *cmd, int tty);



static void init_remote_threadtests (void);



#define SAMPLE_THREAD  0x05060708        /* Truncated 64 bit threadid.  */



static void

threadset_test_cmd (char *cmd, int tty)

{

  int sample_thread = SAMPLE_THREAD;



  printf_filtered (_("Remote threadset test\n"));

  set_general_thread (sample_thread);

}





static void

threadalive_test (char *cmd, int tty)

{

  int sample_thread = SAMPLE_THREAD;

  int pid = ptid_get_pid (inferior_ptid);

  ptid_t ptid = ptid_build (pid, sample_thread, 0);



  if (remote_thread_alive (ptid))

    printf_filtered ("PASS: Thread alive test\n");

  else

    printf_filtered ("FAIL: Thread alive test\n");

}



void output_threadid (char *title, threadref *ref);



void

output_threadid (char *title, threadref *ref)

{

  char hexid[20];



  pack_threadid (&hexid[0], ref);        /* Convert threead id into hex.  */

  hexid[16] = 0;

  printf_filtered ("%s  %s\n", title, (&hexid[0]));

}



static void

threadlist_test_cmd (char *cmd, int tty)

{

  int startflag = 1;

  threadref nextthread;

  int done, result_count;

  threadref threadlist[3];



  printf_filtered ("Remote Threadlist test\n");

  if (!remote_get_threadlist (startflag, &nextthread, 3, &done,

                              &result_count, &threadlist[0]))

    printf_filtered ("FAIL: threadlist test\n");

  else

    {

      threadref *scan = threadlist;

      threadref *limit = scan + result_count;



      while (scan < limit)

        output_threadid (" thread ", scan++);

    }

}



void

display_thread_info (struct gdb_ext_thread_info *info)

{

  output_threadid ("Threadid: ", &info->threadid);

  printf_filtered ("Name: %s\n ", info->shortname);

  printf_filtered ("State: %s\n", info->display);

  printf_filtered ("other: %s\n\n", info->more_display);

}



int

get_and_display_threadinfo (threadref *ref)

{

  int result;

  int set;

  struct gdb_ext_thread_info threadinfo;



  set = TAG_THREADID | TAG_EXISTS | TAG_THREADNAME

    | TAG_MOREDISPLAY | TAG_DISPLAY;

  if (0 != (result = remote_get_threadinfo (ref, set, &threadinfo)))

    display_thread_info (&threadinfo);

  return result;

}



static void

threadinfo_test_cmd (char *cmd, int tty)

{

  int athread = SAMPLE_THREAD;

  threadref thread;

  int set;



  int_to_threadref (&thread, athread);

  printf_filtered ("Remote Threadinfo test\n");

  if (!get_and_display_threadinfo (&thread))

    printf_filtered ("FAIL cannot get thread info\n");

}



static int

thread_display_step (threadref *ref, void *context)

{

  /* output_threadid(" threadstep ",ref); *//* simple test */

  return get_and_display_threadinfo (ref);

}



static void

threadlist_update_test_cmd (char *cmd, int tty)

{

  printf_filtered ("Remote Threadlist update test\n");

  remote_threadlist_iterator (thread_display_step, 0, CRAZY_MAX_THREADS);

}



static void

init_remote_threadtests (void)

{

  add_com ("tlist", class_obscure, threadlist_test_cmd,

           _("Fetch and print the remote list of "

             "thread identifiers, one pkt only"));

  add_com ("tinfo", class_obscure, threadinfo_test_cmd,

           _("Fetch and display info about one thread"));

  add_com ("tset", class_obscure, threadset_test_cmd,

           _("Test setting to a different thread"));

  add_com ("tupd", class_obscure, threadlist_update_test_cmd,

           _("Iterate through updating all remote thread info"));

  add_com ("talive", class_obscure, threadalive_test,

           _(" Remote thread alive test "));

}



#endif /* 0 */



/* Convert a thread ID to a string.  Returns the string in a static

   buffer.  */



static char *

‌remote_pid_to_str (struct target_ops *ops, ptid_t ptid)

{

  static char buf[64];

  struct remote_state *rs = get_remote_state ();



  if (ptid_equal (ptid, null_ptid))

    return normal_pid_to_str (ptid);

  else if (ptid_is_pid (ptid))

    {

      /* Printing an inferior target id.  */



      /* When multi-process extensions are off, there's no way in the

         remote protocol to know the remote process id, if there's any

         at all.  There's one exception --- when we're connected with

         target extended-remote, and we manually attached to a process

         with "attach PID".  We don't record anywhere a flag that

         allows us to distinguish that case from the case of

         connecting with extended-remote and the stub already being

         attached to a process, and reporting yes to qAttached, hence

         no smart special casing here.  */

      if (!remote_multi_process_p (rs))

        {

          xsnprintf (buf, sizeof buf, "Remote target");

          return buf;

        }



      return normal_pid_to_str (ptid);

    }

  else

    {

      if (ptid_equal (magic_null_ptid, ptid))

        xsnprintf (buf, sizeof buf, "Thread <main>");

      else if (rs->extended && remote_multi_process_p (rs))

        xsnprintf (buf, sizeof buf, "Thread %d.%ld",

                   ptid_get_pid (ptid), ptid_get_lwp (ptid));

      else

        xsnprintf (buf, sizeof buf, "Thread %ld",

                   ptid_get_lwp (ptid));

      return buf;

    }

}



/* Get the address of the thread local variable in OBJFILE which is

   stored at OFFSET within the thread local storage for thread PTID.  */



static CORE_ADDR

‌remote_get_thread_local_address (struct target_ops *ops,

                                 ptid_t ptid, CORE_ADDR lm, CORE_ADDR offset)

{

  if (packet_support (PACKET_qGetTLSAddr) != PACKET_DISABLE)

    {

      struct remote_state *rs = get_remote_state ();

      char *p = rs->buf;

      char *endp = rs->buf + get_remote_packet_size ();

      enum packet_result result;



      strcpy (p, "qGetTLSAddr:");

      p += strlen (p);

      p = write_ptid (p, endp, ptid);

      *p++ = ',';

      p += hexnumstr (p, offset);

      *p++ = ',';

      p += hexnumstr (p, lm);

      *p++ = '\0';



      putpkt (rs->buf);

      getpkt (&rs->buf, &rs->buf_size, 0);

      result = packet_ok (rs->buf,

                          &remote_protocol_packets[PACKET_qGetTLSAddr]);

      if (result == PACKET_OK)

        {

          ULONGEST result;



          unpack_varlen_hex (rs->buf, &result);

          return result;

        }

      else if (result == PACKET_UNKNOWN)

        throw_error (TLS_GENERIC_ERROR,

                     _("Remote target doesn't support qGetTLSAddr packet"));

      else

        throw_error (TLS_GENERIC_ERROR,

                     _("Remote target failed to process qGetTLSAddr request"));

    }

  else

    throw_error (TLS_GENERIC_ERROR,

                 _("TLS not supported or disabled on this target"));

  /* Not reached.  */

  return 0;

}



/* Provide thread local base, i.e. Thread Information Block address.

   Returns 1 if ptid is found and thread_local_base is non zero.  */



static int

‌remote_get_tib_address (struct target_ops *self, ptid_t ptid, CORE_ADDR *addr)

{

  if (packet_support (PACKET_qGetTIBAddr) != PACKET_DISABLE)

    {

      struct remote_state *rs = get_remote_state ();

      char *p = rs->buf;

      char *endp = rs->buf + get_remote_packet_size ();

      enum packet_result result;



      strcpy (p, "qGetTIBAddr:");

      p += strlen (p);

      p = write_ptid (p, endp, ptid);

      *p++ = '\0';



      putpkt (rs->buf);

      getpkt (&rs->buf, &rs->buf_size, 0);

      result = packet_ok (rs->buf,

                          &remote_protocol_packets[PACKET_qGetTIBAddr]);

      if (result == PACKET_OK)

        {

          ULONGEST result;



          unpack_varlen_hex (rs->buf, &result);

          if (addr)

            *addr = (CORE_ADDR) result;

          return 1;

        }

      else if (result == PACKET_UNKNOWN)

        error (_("Remote target doesn't support qGetTIBAddr packet"));

      else

        error (_("Remote target failed to process qGetTIBAddr request"));

    }

  else

    error (_("qGetTIBAddr not supported or disabled on this target"));

  /* Not reached.  */

  return 0;

}



/* Support for inferring a target description based on the current

   architecture and the size of a 'g' packet.  While the 'g' packet

   can have any size (since optional registers can be left off the

   end), some sizes are easily recognizable given knowledge of the

   approximate architecture.  */



‌struct remote_g_packet_guess

{

  int bytes;

  const struct target_desc *tdesc;

};

‌typedef struct remote_g_packet_guess remote_g_packet_guess_s;

‌DEF_VEC_O(remote_g_packet_guess_s);



‌struct remote_g_packet_data

{

  VEC(remote_g_packet_guess_s) *guesses;

};



‌static struct gdbarch_data *remote_g_packet_data_handle;



static void *

‌remote_g_packet_data_init (struct obstack *obstack)

{

  return OBSTACK_ZALLOC (obstack, struct remote_g_packet_data);

}



void

‌register_remote_g_packet_guess (struct gdbarch *gdbarch, int bytes,

                                const struct target_desc *tdesc)

{

  struct remote_g_packet_data *data

    = gdbarch_data (gdbarch, remote_g_packet_data_handle);

  struct remote_g_packet_guess new_guess, *guess;

  int ix;



  gdb_assert (tdesc != NULL);



  for (ix = 0;

       VEC_iterate (remote_g_packet_guess_s, data->guesses, ix, guess);

       ix++)

    if (guess->bytes == bytes)

      internal_error (__FILE__, __LINE__,

                      _("Duplicate g packet description added for size %d"),

                      bytes);



  new_guess.bytes = bytes;

  new_guess.tdesc = tdesc;

  VEC_safe_push (remote_g_packet_guess_s, data->guesses, &new_guess);

}



/* Return 1 if remote_read_description would do anything on this target

   and architecture, 0 otherwise.  */



static int

‌remote_read_description_p (struct target_ops *target)

{

  struct remote_g_packet_data *data

    = gdbarch_data (target_gdbarch (), remote_g_packet_data_handle);



  if (!VEC_empty (remote_g_packet_guess_s, data->guesses))

    return 1;



  return 0;

}



static const struct target_desc *

‌remote_read_description (struct target_ops *target)

{

  struct remote_g_packet_data *data

    = gdbarch_data (target_gdbarch (), remote_g_packet_data_handle);



  /* Do not try this during initial connection, when we do not know

     whether there is a running but stopped thread.  */

  if (!target_has_execution || ptid_equal (inferior_ptid, null_ptid))

    return target->beneath->to_read_description (target->beneath);



  if (!VEC_empty (remote_g_packet_guess_s, data->guesses))

    {

      struct remote_g_packet_guess *guess;

      int ix;

      int bytes = send_g_packet ();



      for (ix = 0;

           VEC_iterate (remote_g_packet_guess_s, data->guesses, ix, guess);

           ix++)

        if (guess->bytes == bytes)

          return guess->tdesc;



      /* We discard the g packet.  A minor optimization would be to

         hold on to it, and fill the register cache once we have selected

         an architecture, but it's too tricky to do safely.  */

    }



  return target->beneath->to_read_description (target->beneath);

}



/* Remote file transfer support.  This is host-initiated I/O, not

   target-initiated; for target-initiated, see remote-fileio.c.  */



/* If *LEFT is at least the length of STRING, copy STRING to

   *BUFFER, update *BUFFER to point to the new end of the buffer, and

   decrease *LEFT.  Otherwise raise an error.  */



static void

‌remote_buffer_add_string (char **buffer, int *left, char *string)

{

  int len = strlen (string);



  if (len > *left)

    error (_("Packet too long for target."));



  memcpy (*buffer, string, len);

  *buffer += len;

  *left -= len;



  /* NUL-terminate the buffer as a convenience, if there is

     room.  */

  if (*left)

    **buffer = '\0';

}



/* If *LEFT is large enough, hex encode LEN bytes from BYTES into

   *BUFFER, update *BUFFER to point to the new end of the buffer, and

   decrease *LEFT.  Otherwise raise an error.  */



static void

‌remote_buffer_add_bytes (char **buffer, int *left, const gdb_byte *bytes,

                         int len)

{

  if (2 * len > *left)

    error (_("Packet too long for target."));



  bin2hex (bytes, *buffer, len);

  *buffer += 2 * len;

  *left -= 2 * len;



  /* NUL-terminate the buffer as a convenience, if there is

     room.  */

  if (*left)

    **buffer = '\0';

}



/* If *LEFT is large enough, convert VALUE to hex and add it to

   *BUFFER, update *BUFFER to point to the new end of the buffer, and

   decrease *LEFT.  Otherwise raise an error.  */



static void

‌remote_buffer_add_int (char **buffer, int *left, ULONGEST value)

{

  int len = hexnumlen (value);



  if (len > *left)

    error (_("Packet too long for target."));



  hexnumstr (*buffer, value);

  *buffer += len;

  *left -= len;



  /* NUL-terminate the buffer as a convenience, if there is

     room.  */

  if (*left)

    **buffer = '\0';

}



/* Parse an I/O result packet from BUFFER.  Set RETCODE to the return

   value, *REMOTE_ERRNO to the remote error number or zero if none

   was included, and *ATTACHMENT to point to the start of the annex

   if any.  The length of the packet isn't needed here; there may

   be NUL bytes in BUFFER, but they will be after *ATTACHMENT.



   Return 0 if the packet could be parsed, -1 if it could not.  If

   -1 is returned, the other variables may not be initialized.  */



static int

‌remote_hostio_parse_result (char *buffer, int *retcode,

                            int *remote_errno, char **attachment)

{

  char *p, *p2;



  *remote_errno = 0;

  *attachment = NULL;



  if (buffer[0] != 'F')

    return -1;



  errno = 0;

  *retcode = strtol (&buffer[1], &p, 16);

  if (errno != 0 || p == &buffer[1])

    return -1;



  /* Check for ",errno".  */

  if (*p == ',')

    {

      errno = 0;

      *remote_errno = strtol (p + 1, &p2, 16);

      if (errno != 0 || p + 1 == p2)

        return -1;

      p = p2;

    }



  /* Check for ";attachment".  If there is no attachment, the

     packet should end here.  */

  if (*p == ';')

    {

      *attachment = p + 1;

      return 0;

    }

  else if (*p == '\0')

    return 0;

  else

    return -1;

}



/* Send a prepared I/O packet to the target and read its response.

   The prepared packet is in the global RS->BUF before this function

   is called, and the answer is there when we return.



   COMMAND_BYTES is the length of the request to send, which may include

   binary data.  WHICH_PACKET is the packet configuration to check

   before attempting a packet.  If an error occurs, *REMOTE_ERRNO

   is set to the error number and -1 is returned.  Otherwise the value

   returned by the function is returned.



   ATTACHMENT and ATTACHMENT_LEN should be non-NULL if and only if an

   attachment is expected; an error will be reported if there's a

   mismatch.  If one is found, *ATTACHMENT will be set to point into

   the packet buffer and *ATTACHMENT_LEN will be set to the

   attachment's length.  */



static int

‌remote_hostio_send_command (int command_bytes, int which_packet,

                            int *remote_errno, char **attachment,

                            int *attachment_len)

{

  struct remote_state *rs = get_remote_state ();

  int ret, bytes_read;

  char *attachment_tmp;



  if (!rs->remote_desc

      || packet_support (which_packet) == PACKET_DISABLE)

    {

      *remote_errno = FILEIO_ENOSYS;

      return -1;

    }



  putpkt_binary (rs->buf, command_bytes);

  bytes_read = getpkt_sane (&rs->buf, &rs->buf_size, 0);



  /* If it timed out, something is wrong.  Don't try to parse the

     buffer.  */

  if (bytes_read < 0)

    {

      *remote_errno = FILEIO_EINVAL;

      return -1;

    }



  switch (packet_ok (rs->buf, &remote_protocol_packets[which_packet]))

    {

    case PACKET_ERROR:

      *remote_errno = FILEIO_EINVAL;

      return -1;

    case PACKET_UNKNOWN:

      *remote_errno = FILEIO_ENOSYS;

      return -1;

    case PACKET_OK:

      break;

    }



  if (remote_hostio_parse_result (rs->buf, &ret, remote_errno,

                                  &attachment_tmp))

    {

      *remote_errno = FILEIO_EINVAL;

      return -1;

    }



  /* Make sure we saw an attachment if and only if we expected one.  */

  if ((attachment_tmp == NULL && attachment != NULL)

      || (attachment_tmp != NULL && attachment == NULL))

    {

      *remote_errno = FILEIO_EINVAL;

      return -1;

    }



  /* If an attachment was found, it must point into the packet buffer;

     work out how many bytes there were.  */

  if (attachment_tmp != NULL)

    {

      *attachment = attachment_tmp;

      *attachment_len = bytes_read - (*attachment - rs->buf);

    }



  return ret;

}



/* Open FILENAME on the remote target, using FLAGS and MODE.  Return a

   remote file descriptor, or -1 if an error occurs (and set

   *REMOTE_ERRNO).  */



static int

‌remote_hostio_open (struct target_ops *self,

                    const char *filename, int flags, int mode,

                    int *remote_errno)

{

  struct remote_state *rs = get_remote_state ();

  char *p = rs->buf;

  int left = get_remote_packet_size () - 1;



  remote_buffer_add_string (&p, &left, "vFile:open:");



  remote_buffer_add_bytes (&p, &left, (const gdb_byte *) filename,

                           strlen (filename));

  remote_buffer_add_string (&p, &left, ",");



  remote_buffer_add_int (&p, &left, flags);

  remote_buffer_add_string (&p, &left, ",");



  remote_buffer_add_int (&p, &left, mode);



  return remote_hostio_send_command (p - rs->buf, PACKET_vFile_open,

                                     remote_errno, NULL, NULL);

}



/* Write up to LEN bytes from WRITE_BUF to FD on the remote target.

   Return the number of bytes written, or -1 if an error occurs (and

   set *REMOTE_ERRNO).  */



static int

‌remote_hostio_pwrite (struct target_ops *self,

                      int fd, const gdb_byte *write_buf, int len,

                      ULONGEST offset, int *remote_errno)

{

  struct remote_state *rs = get_remote_state ();

  char *p = rs->buf;

  int left = get_remote_packet_size ();

  int out_len;



  remote_buffer_add_string (&p, &left, "vFile:pwrite:");



  remote_buffer_add_int (&p, &left, fd);

  remote_buffer_add_string (&p, &left, ",");



  remote_buffer_add_int (&p, &left, offset);

  remote_buffer_add_string (&p, &left, ",");



  p += remote_escape_output (write_buf, len, (gdb_byte *) p, &out_len,

                             get_remote_packet_size () - (p - rs->buf));



  return remote_hostio_send_command (p - rs->buf, PACKET_vFile_pwrite,

                                     remote_errno, NULL, NULL);

}



/* Read up to LEN bytes FD on the remote target into READ_BUF

   Return the number of bytes read, or -1 if an error occurs (and

   set *REMOTE_ERRNO).  */



static int

‌remote_hostio_pread (struct target_ops *self,

                     int fd, gdb_byte *read_buf, int len,

                     ULONGEST offset, int *remote_errno)

{

  struct remote_state *rs = get_remote_state ();

  char *p = rs->buf;

  char *attachment;

  int left = get_remote_packet_size ();

  int ret, attachment_len;

  int read_len;



  remote_buffer_add_string (&p, &left, "vFile:pread:");



  remote_buffer_add_int (&p, &left, fd);

  remote_buffer_add_string (&p, &left, ",");



  remote_buffer_add_int (&p, &left, len);

  remote_buffer_add_string (&p, &left, ",");



  remote_buffer_add_int (&p, &left, offset);



  ret = remote_hostio_send_command (p - rs->buf, PACKET_vFile_pread,

                                    remote_errno, &attachment,

                                    &attachment_len);



  if (ret < 0)

    return ret;



  read_len = remote_unescape_input ((gdb_byte *) attachment, attachment_len,

                                    read_buf, len);

  if (read_len != ret)

    error (_("Read returned %d, but %d bytes."), ret, (int) read_len);



  return ret;

}



/* Close FD on the remote target.  Return 0, or -1 if an error occurs

   (and set *REMOTE_ERRNO).  */



static int

‌remote_hostio_close (struct target_ops *self, int fd, int *remote_errno)

{

  struct remote_state *rs = get_remote_state ();

  char *p = rs->buf;

  int left = get_remote_packet_size () - 1;



  remote_buffer_add_string (&p, &left, "vFile:close:");



  remote_buffer_add_int (&p, &left, fd);



  return remote_hostio_send_command (p - rs->buf, PACKET_vFile_close,

                                     remote_errno, NULL, NULL);

}



/* Unlink FILENAME on the remote target.  Return 0, or -1 if an error

   occurs (and set *REMOTE_ERRNO).  */



static int

‌remote_hostio_unlink (struct target_ops *self,

                      const char *filename, int *remote_errno)

{

  struct remote_state *rs = get_remote_state ();

  char *p = rs->buf;

  int left = get_remote_packet_size () - 1;



  remote_buffer_add_string (&p, &left, "vFile:unlink:");



  remote_buffer_add_bytes (&p, &left, (const gdb_byte *) filename,

                           strlen (filename));



  return remote_hostio_send_command (p - rs->buf, PACKET_vFile_unlink,

                                     remote_errno, NULL, NULL);

}



/* Read value of symbolic link FILENAME on the remote target.  Return

   a null-terminated string allocated via xmalloc, or NULL if an error

   occurs (and set *REMOTE_ERRNO).  */



static char *

‌remote_hostio_readlink (struct target_ops *self,

                        const char *filename, int *remote_errno)

{

  struct remote_state *rs = get_remote_state ();

  char *p = rs->buf;

  char *attachment;

  int left = get_remote_packet_size ();

  int len, attachment_len;

  int read_len;

  char *ret;



  remote_buffer_add_string (&p, &left, "vFile:readlink:");



  remote_buffer_add_bytes (&p, &left, (const gdb_byte *) filename,

                           strlen (filename));



  len = remote_hostio_send_command (p - rs->buf, PACKET_vFile_readlink,

                                    remote_errno, &attachment,

                                    &attachment_len);



  if (len < 0)

    return NULL;



  ret = xmalloc (len + 1);



  read_len = remote_unescape_input ((gdb_byte *) attachment, attachment_len,

                                    (gdb_byte *) ret, len);

  if (read_len != len)

    error (_("Readlink returned %d, but %d bytes."), len, read_len);



  ret[len] = '\0';

  return ret;

}



static int

‌remote_fileio_errno_to_host (int errnum)

{

  switch (errnum)

    {

      case FILEIO_EPERM:

        return EPERM;

      case FILEIO_ENOENT:

        return ENOENT;

      case FILEIO_EINTR:

        return EINTR;

      case FILEIO_EIO:

        return EIO;

      case FILEIO_EBADF:

        return EBADF;

      case FILEIO_EACCES:

        return EACCES;

      case FILEIO_EFAULT:

        return EFAULT;

      case FILEIO_EBUSY:

        return EBUSY;

      case FILEIO_EEXIST:

        return EEXIST;

      case FILEIO_ENODEV:

        return ENODEV;

      case FILEIO_ENOTDIR:

        return ENOTDIR;

      case FILEIO_EISDIR:

        return EISDIR;

      case FILEIO_EINVAL:

        return EINVAL;

      case FILEIO_ENFILE:

        return ENFILE;

      case FILEIO_EMFILE:

        return EMFILE;

      case FILEIO_EFBIG:

        return EFBIG;

      case FILEIO_ENOSPC:

        return ENOSPC;

      case FILEIO_ESPIPE:

        return ESPIPE;

      case FILEIO_EROFS:

        return EROFS;

      case FILEIO_ENOSYS:

        return ENOSYS;

      case FILEIO_ENAMETOOLONG:

        return ENAMETOOLONG;

    }

  return -1;

}



static char *

‌remote_hostio_error (int errnum)

{

  int host_error = remote_fileio_errno_to_host (errnum);



  if (host_error == -1)

    error (_("Unknown remote I/O error %d"), errnum);

  else

    error (_("Remote I/O error: %s"), safe_strerror (host_error));

}



static void

‌remote_hostio_close_cleanup (void *opaque)

{

  int fd = *(int *) opaque;

  int remote_errno;



  remote_hostio_close (find_target_at (process_stratum), fd, &remote_errno);

}





static void *

‌remote_bfd_iovec_open (struct bfd *abfd, void *open_closure)

{

  const char *filename = bfd_get_filename (abfd);

  int fd, remote_errno;

  int *stream;



  gdb_assert (remote_filename_p (filename));



  fd = remote_hostio_open (find_target_at (process_stratum),

                           filename + 7, FILEIO_O_RDONLY, 0, &remote_errno);

  if (fd == -1)

    {

      errno = remote_fileio_errno_to_host (remote_errno);

      bfd_set_error (bfd_error_system_call);

      return NULL;

    }



  stream = xmalloc (sizeof (int));

  *stream = fd;

  return stream;

}



static int

‌remote_bfd_iovec_close (struct bfd *abfd, void *stream)

{

  int fd = *(int *)stream;

  int remote_errno;



  xfree (stream);



  /* Ignore errors on close; these may happen if the remote

     connection was already torn down.  */

  remote_hostio_close (find_target_at (process_stratum), fd, &remote_errno);



  /* Zero means success.  */

  return 0;

}



static file_ptr

‌remote_bfd_iovec_pread (struct bfd *abfd, void *stream, void *buf,

                        file_ptr nbytes, file_ptr offset)

{

  int fd = *(int *)stream;

  int remote_errno;

  file_ptr pos, bytes;



  pos = 0;

  while (nbytes > pos)

    {

      bytes = remote_hostio_pread (find_target_at (process_stratum),

                                   fd, (gdb_byte *) buf + pos, nbytes - pos,

                                   offset + pos, &remote_errno);

      if (bytes == 0)

        /* Success, but no bytes, means end-of-file.  */

        break;

      if (bytes == -1)

        {

          errno = remote_fileio_errno_to_host (remote_errno);

          bfd_set_error (bfd_error_system_call);

          return -1;

        }



      pos += bytes;

    }



  return pos;

}



static int

‌remote_bfd_iovec_stat (struct bfd *abfd, void *stream, struct stat *sb)

{

  /* FIXME: We should probably implement remote_hostio_stat.  */

  sb->st_size = INT_MAX;

  return 0;

}



int

‌remote_filename_p (const char *filename)

{

  return strncmp (filename,

                  REMOTE_SYSROOT_PREFIX,

                  sizeof (REMOTE_SYSROOT_PREFIX) - 1) == 0;

}



bfd *

‌remote_bfd_open (const char *remote_file, const char *target)

{

  bfd *abfd = gdb_bfd_openr_iovec (remote_file, target,

                                   remote_bfd_iovec_open, NULL,

                                   remote_bfd_iovec_pread,

                                   remote_bfd_iovec_close,

                                   remote_bfd_iovec_stat);



  return abfd;

}



void

‌remote_file_put (const char *local_file, const char *remote_file, int from_tty)

{

  struct cleanup *back_to, *close_cleanup;

  int retcode, fd, remote_errno, bytes, io_size;

  FILE *file;

  gdb_byte *buffer;

  int bytes_in_buffer;

  int saw_eof;

  ULONGEST offset;

  struct remote_state *rs = get_remote_state ();



  if (!rs->remote_desc)

    error (_("command can only be used with remote target"));



  file = gdb_fopen_cloexec (local_file, "rb");

  if (file == NULL)

    perror_with_name (local_file);

  back_to = make_cleanup_fclose (file);



  fd = remote_hostio_open (find_target_at (process_stratum),

                           remote_file, (FILEIO_O_WRONLY | FILEIO_O_CREAT

                                         | FILEIO_O_TRUNC),

                           0700, &remote_errno);

  if (fd == -1)

    remote_hostio_error (remote_errno);



  /* Send up to this many bytes at once.  They won't all fit in the

     remote packet limit, so we'll transfer slightly fewer.  */

  io_size = get_remote_packet_size ();

  buffer = xmalloc (io_size);

  make_cleanup (xfree, buffer);



  close_cleanup = make_cleanup (remote_hostio_close_cleanup, &fd);



  bytes_in_buffer = 0;

  saw_eof = 0;

  offset = 0;

  while (bytes_in_buffer || !saw_eof)

    {

      if (!saw_eof)

        {

          bytes = fread (buffer + bytes_in_buffer, 1,

                         io_size - bytes_in_buffer,

                         file);

          if (bytes == 0)

            {

              if (ferror (file))

                error (_("Error reading %s."), local_file);

              else

                {

                  /* EOF.  Unless there is something still in the

                     buffer from the last iteration, we are done.  */

                  saw_eof = 1;

                  if (bytes_in_buffer == 0)

                    break;

                }

            }

        }

      else

        bytes = 0;



      bytes += bytes_in_buffer;

      bytes_in_buffer = 0;



      retcode = remote_hostio_pwrite (find_target_at (process_stratum),

                                      fd, buffer, bytes,

                                      offset, &remote_errno);



      if (retcode < 0)

        remote_hostio_error (remote_errno);

      else if (retcode == 0)

        error (_("Remote write of %d bytes returned 0!"), bytes);

      else if (retcode < bytes)

        {

          /* Short write.  Save the rest of the read data for the next

             write.  */

          bytes_in_buffer = bytes - retcode;

          memmove (buffer, buffer + retcode, bytes_in_buffer);

        }



      offset += retcode;

    }



  discard_cleanups (close_cleanup);

  if (remote_hostio_close (find_target_at (process_stratum), fd, &remote_errno))

    remote_hostio_error (remote_errno);



  if (from_tty)

    printf_filtered (_("Successfully sent file \"%s\".\n"), local_file);

  do_cleanups (back_to);

}



void

‌remote_file_get (const char *remote_file, const char *local_file, int from_tty)

{

  struct cleanup *back_to, *close_cleanup;

  int fd, remote_errno, bytes, io_size;

  FILE *file;

  gdb_byte *buffer;

  ULONGEST offset;

  struct remote_state *rs = get_remote_state ();



  if (!rs->remote_desc)

    error (_("command can only be used with remote target"));



  fd = remote_hostio_open (find_target_at (process_stratum),

                           remote_file, FILEIO_O_RDONLY, 0, &remote_errno);

  if (fd == -1)

    remote_hostio_error (remote_errno);



  file = gdb_fopen_cloexec (local_file, "wb");

  if (file == NULL)

    perror_with_name (local_file);

  back_to = make_cleanup_fclose (file);



  /* Send up to this many bytes at once.  They won't all fit in the

     remote packet limit, so we'll transfer slightly fewer.  */

  io_size = get_remote_packet_size ();

  buffer = xmalloc (io_size);

  make_cleanup (xfree, buffer);



  close_cleanup = make_cleanup (remote_hostio_close_cleanup, &fd);



  offset = 0;

  while (1)

    {

      bytes = remote_hostio_pread (find_target_at (process_stratum),

                                   fd, buffer, io_size, offset, &remote_errno);

      if (bytes == 0)

        /* Success, but no bytes, means end-of-file.  */

        break;

      if (bytes == -1)

        remote_hostio_error (remote_errno);



      offset += bytes;



      bytes = fwrite (buffer, 1, bytes, file);

      if (bytes == 0)

        perror_with_name (local_file);

    }



  discard_cleanups (close_cleanup);

  if (remote_hostio_close (find_target_at (process_stratum), fd, &remote_errno))

    remote_hostio_error (remote_errno);



  if (from_tty)

    printf_filtered (_("Successfully fetched file \"%s\".\n"), remote_file);

  do_cleanups (back_to);

}



void

‌remote_file_delete (const char *remote_file, int from_tty)

{

  int retcode, remote_errno;

  struct remote_state *rs = get_remote_state ();



  if (!rs->remote_desc)

    error (_("command can only be used with remote target"));



  retcode = remote_hostio_unlink (find_target_at (process_stratum),

                                  remote_file, &remote_errno);

  if (retcode == -1)

    remote_hostio_error (remote_errno);



  if (from_tty)

    printf_filtered (_("Successfully deleted file \"%s\".\n"), remote_file);

}



static void

‌remote_put_command (char *args, int from_tty)

{

  struct cleanup *back_to;

  char **argv;



  if (args == NULL)

    error_no_arg (_("file to put"));



  argv = gdb_buildargv (args);

  back_to = make_cleanup_freeargv (argv);

  if (argv[0] == NULL || argv[1] == NULL || argv[2] != NULL)

    error (_("Invalid parameters to remote put"));



  remote_file_put (argv[0], argv[1], from_tty);



  do_cleanups (back_to);

}



static void

‌remote_get_command (char *args, int from_tty)

{

  struct cleanup *back_to;

  char **argv;



  if (args == NULL)

    error_no_arg (_("file to get"));



  argv = gdb_buildargv (args);

  back_to = make_cleanup_freeargv (argv);

  if (argv[0] == NULL || argv[1] == NULL || argv[2] != NULL)

    error (_("Invalid parameters to remote get"));



  remote_file_get (argv[0], argv[1], from_tty);



  do_cleanups (back_to);

}



static void

‌remote_delete_command (char *args, int from_tty)

{

  struct cleanup *back_to;

  char **argv;



  if (args == NULL)

    error_no_arg (_("file to delete"));



  argv = gdb_buildargv (args);

  back_to = make_cleanup_freeargv (argv);

  if (argv[0] == NULL || argv[1] != NULL)

    error (_("Invalid parameters to remote delete"));



  remote_file_delete (argv[0], from_tty);



  do_cleanups (back_to);

}



static void

‌remote_command (char *args, int from_tty)

{

  help_list (remote_cmdlist, "remote ", all_commands, gdb_stdout);

}



static int

‌remote_can_execute_reverse (struct target_ops *self)

{

  if (packet_support (PACKET_bs) == PACKET_ENABLE

      || packet_support (PACKET_bc) == PACKET_ENABLE)

    return 1;

  else

    return 0;

}



static int

‌remote_supports_non_stop (struct target_ops *self)

{

  return 1;

}



static int

‌remote_supports_disable_randomization (struct target_ops *self)

{

  /* Only supported in extended mode.  */

  return 0;

}



static int

‌remote_supports_multi_process (struct target_ops *self)

{

  struct remote_state *rs = get_remote_state ();



  /* Only extended-remote handles being attached to multiple

     processes, even though plain remote can use the multi-process

     thread id extensions, so that GDB knows the target process's

     PID.  */

  return rs->extended && remote_multi_process_p (rs);

}



static int

‌remote_supports_cond_tracepoints (void)

{

  return packet_support (PACKET_ConditionalTracepoints) == PACKET_ENABLE;

}



static int

‌remote_supports_cond_breakpoints (struct target_ops *self)

{

  return packet_support (PACKET_ConditionalBreakpoints) == PACKET_ENABLE;

}



static int

‌remote_supports_fast_tracepoints (void)

{

  return packet_support (PACKET_FastTracepoints) == PACKET_ENABLE;

}



static int

‌remote_supports_static_tracepoints (void)

{

  return packet_support (PACKET_StaticTracepoints) == PACKET_ENABLE;

}



static int

‌remote_supports_install_in_trace (void)

{

  return packet_support (PACKET_InstallInTrace) == PACKET_ENABLE;

}



static int

‌remote_supports_enable_disable_tracepoint (struct target_ops *self)

{

  return (packet_support (PACKET_EnableDisableTracepoints_feature)

          == PACKET_ENABLE);

}



static int

‌remote_supports_string_tracing (struct target_ops *self)

{

  return packet_support (PACKET_tracenz_feature) == PACKET_ENABLE;

}



static int

‌remote_can_run_breakpoint_commands (struct target_ops *self)

{

  return packet_support (PACKET_BreakpointCommands) == PACKET_ENABLE;

}



static void

‌remote_trace_init (struct target_ops *self)

{

  putpkt ("QTinit");

  remote_get_noisy_reply (&target_buf, &target_buf_size);

  if (strcmp (target_buf, "OK") != 0)

    error (_("Target does not support this command."));

}



static void free_actions_list (char **actions_list);

static void free_actions_list_cleanup_wrapper (void *);

static void

‌free_actions_list_cleanup_wrapper (void *al)

{

  free_actions_list (al);

}



static void

‌free_actions_list (char **actions_list)

{

  int ndx;



  if (actions_list == 0)

    return;



  for (ndx = 0; actions_list[ndx]; ndx++)

    xfree (actions_list[ndx]);



  xfree (actions_list);

}



/* Recursive routine to walk through command list including loops, and

   download packets for each command.  */



static void

‌remote_download_command_source (int num, ULONGEST addr,

                                struct command_line *cmds)

{

  struct remote_state *rs = get_remote_state ();

  struct command_line *cmd;



  for (cmd = cmds; cmd; cmd = cmd->next)

    {

      QUIT;        /* Allow user to bail out with ^C.  */

      strcpy (rs->buf, "QTDPsrc:");

      encode_source_string (num, addr, "cmd", cmd->line,

                            rs->buf + strlen (rs->buf),

                            rs->buf_size - strlen (rs->buf));

      putpkt (rs->buf);

      remote_get_noisy_reply (&target_buf, &target_buf_size);

      if (strcmp (target_buf, "OK"))

        warning (_("Target does not support source download."));



      if (cmd->control_type == while_control

          || cmd->control_type == while_stepping_control)

        {

          remote_download_command_source (num, addr, *cmd->body_list);



          QUIT;        /* Allow user to bail out with ^C.  */

          strcpy (rs->buf, "QTDPsrc:");

          encode_source_string (num, addr, "cmd", "end",

                                rs->buf + strlen (rs->buf),

                                rs->buf_size - strlen (rs->buf));

          putpkt (rs->buf);

          remote_get_noisy_reply (&target_buf, &target_buf_size);

          if (strcmp (target_buf, "OK"))

            warning (_("Target does not support source download."));

        }

    }

}



static void

‌remote_download_tracepoint (struct target_ops *self, struct bp_location *loc)

{

‌#define BUF_SIZE 2048



  CORE_ADDR tpaddr;

  char addrbuf[40];

  char buf[BUF_SIZE];

  char **tdp_actions;

  char **stepping_actions;

  int ndx;

  struct cleanup *old_chain = NULL;

  struct agent_expr *aexpr;

  struct cleanup *aexpr_chain = NULL;

  char *pkt;

  struct breakpoint *b = loc->owner;

  struct tracepoint *t = (struct tracepoint *) b;



  encode_actions_rsp (loc, &tdp_actions, &stepping_actions);

  old_chain = make_cleanup (free_actions_list_cleanup_wrapper,

                            tdp_actions);

  (void) make_cleanup (free_actions_list_cleanup_wrapper,

                       stepping_actions);



  tpaddr = loc->address;

  sprintf_vma (addrbuf, tpaddr);

  xsnprintf (buf, BUF_SIZE, "QTDP:%x:%s:%c:%lx:%x", b->number,

             addrbuf, /* address */

             (b->enable_state == bp_enabled ? 'E' : 'D'),

             t->step_count, t->pass_count);

  /* Fast tracepoints are mostly handled by the target, but we can

     tell the target how big of an instruction block should be moved

     around.  */

  if (b->type == bp_fast_tracepoint)

    {

      /* Only test for support at download time; we may not know

         target capabilities at definition time.  */

      if (remote_supports_fast_tracepoints ())

        {

          int isize;



          if (gdbarch_fast_tracepoint_valid_at (target_gdbarch (),

                                                tpaddr, &isize, NULL))

            xsnprintf (buf + strlen (buf), BUF_SIZE - strlen (buf), ":F%x",

                       isize);

          else

            /* If it passed validation at definition but fails now,

               something is very wrong.  */

            internal_error (__FILE__, __LINE__,

                            _("Fast tracepoint not "

                              "valid during download"));

        }

      else

        /* Fast tracepoints are functionally identical to regular

           tracepoints, so don't take lack of support as a reason to

           give up on the trace run.  */

        warning (_("Target does not support fast tracepoints, "

                   "downloading %d as regular tracepoint"), b->number);

    }

  else if (b->type == bp_static_tracepoint)

    {

      /* Only test for support at download time; we may not know

         target capabilities at definition time.  */

      if (remote_supports_static_tracepoints ())

        {

          struct static_tracepoint_marker marker;



          if (target_static_tracepoint_marker_at (tpaddr, &marker))

            strcat (buf, ":S");

          else

            error (_("Static tracepoint not valid during download"));

        }

      else

        /* Fast tracepoints are functionally identical to regular

           tracepoints, so don't take lack of support as a reason

           to give up on the trace run.  */

        error (_("Target does not support static tracepoints"));

    }

  /* If the tracepoint has a conditional, make it into an agent

     expression and append to the definition.  */

  if (loc->cond)

    {

      /* Only test support at download time, we may not know target

         capabilities at definition time.  */

      if (remote_supports_cond_tracepoints ())

        {

          aexpr = gen_eval_for_expr (tpaddr, loc->cond);

          aexpr_chain = make_cleanup_free_agent_expr (aexpr);

          xsnprintf (buf + strlen (buf), BUF_SIZE - strlen (buf), ":X%x,",

                     aexpr->len);

          pkt = buf + strlen (buf);

          for (ndx = 0; ndx < aexpr->len; ++ndx)

            pkt = pack_hex_byte (pkt, aexpr->buf[ndx]);

          *pkt = '\0';

          do_cleanups (aexpr_chain);

        }

      else

        warning (_("Target does not support conditional tracepoints, "

                   "ignoring tp %d cond"), b->number);

    }



  if (b->commands || *default_collect)

    strcat (buf, "-");

  putpkt (buf);

  remote_get_noisy_reply (&target_buf, &target_buf_size);

  if (strcmp (target_buf, "OK"))

    error (_("Target does not support tracepoints."));



  /* do_single_steps (t); */

  if (tdp_actions)

    {

      for (ndx = 0; tdp_actions[ndx]; ndx++)

        {

          QUIT;        /* Allow user to bail out with ^C.  */

          xsnprintf (buf, BUF_SIZE, "QTDP:-%x:%s:%s%c",

                     b->number, addrbuf, /* address */

                     tdp_actions[ndx],

                     ((tdp_actions[ndx + 1] || stepping_actions)

                      ? '-' : 0));

          putpkt (buf);

          remote_get_noisy_reply (&target_buf,

                                  &target_buf_size);

          if (strcmp (target_buf, "OK"))

            error (_("Error on target while setting tracepoints."));

        }

    }

  if (stepping_actions)

    {

      for (ndx = 0; stepping_actions[ndx]; ndx++)

        {

          QUIT;        /* Allow user to bail out with ^C.  */

          xsnprintf (buf, BUF_SIZE, "QTDP:-%x:%s:%s%s%s",

                     b->number, addrbuf, /* address */

                     ((ndx == 0) ? "S" : ""),

                     stepping_actions[ndx],

                     (stepping_actions[ndx + 1] ? "-" : ""));

          putpkt (buf);

          remote_get_noisy_reply (&target_buf,

                                  &target_buf_size);

          if (strcmp (target_buf, "OK"))

            error (_("Error on target while setting tracepoints."));

        }

    }



  if (packet_support (PACKET_TracepointSource) == PACKET_ENABLE)

    {

      if (b->addr_string)

        {

          strcpy (buf, "QTDPsrc:");

          encode_source_string (b->number, loc->address,

                                "at", b->addr_string, buf + strlen (buf),

                                2048 - strlen (buf));



          putpkt (buf);

          remote_get_noisy_reply (&target_buf, &target_buf_size);

          if (strcmp (target_buf, "OK"))

            warning (_("Target does not support source download."));

        }

      if (b->cond_string)

        {

          strcpy (buf, "QTDPsrc:");

          encode_source_string (b->number, loc->address,

                                "cond", b->cond_string, buf + strlen (buf),

                                2048 - strlen (buf));

          putpkt (buf);

          remote_get_noisy_reply (&target_buf, &target_buf_size);

          if (strcmp (target_buf, "OK"))

            warning (_("Target does not support source download."));

        }

      remote_download_command_source (b->number, loc->address,

                                      breakpoint_commands (b));

    }



  do_cleanups (old_chain);

}



static int

‌remote_can_download_tracepoint (struct target_ops *self)

{

  struct remote_state *rs = get_remote_state ();

  struct trace_status *ts;

  int status;



  /* Don't try to install tracepoints until we've relocated our

     symbols, and fetched and merged the target's tracepoint list with

     ours.  */

  if (rs->starting_up)

    return 0;



  ts = current_trace_status ();

  status = remote_get_trace_status (self, ts);



  if (status == -1 || !ts->running_known || !ts->running)

    return 0;



  /* If we are in a tracing experiment, but remote stub doesn't support

     installing tracepoint in trace, we have to return.  */

  if (!remote_supports_install_in_trace ())

    return 0;



  return 1;

}





static void

‌remote_download_trace_state_variable (struct target_ops *self,

                                      struct trace_state_variable *tsv)

{

  struct remote_state *rs = get_remote_state ();

  char *p;



  xsnprintf (rs->buf, get_remote_packet_size (), "QTDV:%x:%s:%x:",

             tsv->number, phex ((ULONGEST) tsv->initial_value, 8),

             tsv->builtin);

  p = rs->buf + strlen (rs->buf);

  if ((p - rs->buf) + strlen (tsv->name) * 2 >= get_remote_packet_size ())

    error (_("Trace state variable name too long for tsv definition packet"));

  p += 2 * bin2hex ((gdb_byte *) (tsv->name), p, strlen (tsv->name));

  *p++ = '\0';

  putpkt (rs->buf);

  remote_get_noisy_reply (&target_buf, &target_buf_size);

  if (*target_buf == '\0')

    error (_("Target does not support this command."));

  if (strcmp (target_buf, "OK") != 0)

    error (_("Error on target while downloading trace state variable."));

}



static void

‌remote_enable_tracepoint (struct target_ops *self,

                          struct bp_location *location)

{

  struct remote_state *rs = get_remote_state ();

  char addr_buf[40];



  sprintf_vma (addr_buf, location->address);

  xsnprintf (rs->buf, get_remote_packet_size (), "QTEnable:%x:%s",

             location->owner->number, addr_buf);

  putpkt (rs->buf);

  remote_get_noisy_reply (&rs->buf, &rs->buf_size);

  if (*rs->buf == '\0')

    error (_("Target does not support enabling tracepoints while a trace run is ongoing."));

  if (strcmp (rs->buf, "OK") != 0)

    error (_("Error on target while enabling tracepoint."));

}



static void

‌remote_disable_tracepoint (struct target_ops *self,

                           struct bp_location *location)

{

  struct remote_state *rs = get_remote_state ();

  char addr_buf[40];



  sprintf_vma (addr_buf, location->address);

  xsnprintf (rs->buf, get_remote_packet_size (), "QTDisable:%x:%s",

             location->owner->number, addr_buf);

  putpkt (rs->buf);

  remote_get_noisy_reply (&rs->buf, &rs->buf_size);

  if (*rs->buf == '\0')

    error (_("Target does not support disabling tracepoints while a trace run is ongoing."));

  if (strcmp (rs->buf, "OK") != 0)

    error (_("Error on target while disabling tracepoint."));

}



static void

‌remote_trace_set_readonly_regions (struct target_ops *self)

{

  asection *s;

  bfd *abfd = NULL;

  bfd_size_type size;

  bfd_vma vma;

  int anysecs = 0;

  int offset = 0;



  if (!exec_bfd)

    return;                        /* No information to give.  */



  strcpy (target_buf, "QTro");

  offset = strlen (target_buf);

  for (s = exec_bfd->sections; s; s = s->next)

    {

      char tmp1[40], tmp2[40];

      int sec_length;



      if ((s->flags & SEC_LOAD) == 0 ||

      /*  (s->flags & SEC_CODE) == 0 || */

          (s->flags & SEC_READONLY) == 0)

        continue;



      anysecs = 1;

      vma = bfd_get_section_vma (abfd, s);

      size = bfd_get_section_size (s);

      sprintf_vma (tmp1, vma);

      sprintf_vma (tmp2, vma + size);

      sec_length = 1 + strlen (tmp1) + 1 + strlen (tmp2);

      if (offset + sec_length + 1 > target_buf_size)

        {

          if (packet_support (PACKET_qXfer_traceframe_info) != PACKET_ENABLE)

            warning (_("\

Too many sections for read-only sections definition packet."));

          break;

        }

      xsnprintf (target_buf + offset, target_buf_size - offset, ":%s,%s",

                 tmp1, tmp2);

      offset += sec_length;

    }

  if (anysecs)

    {

      putpkt (target_buf);

      getpkt (&target_buf, &target_buf_size, 0);

    }

}



static void

‌remote_trace_start (struct target_ops *self)

{

  putpkt ("QTStart");

  remote_get_noisy_reply (&target_buf, &target_buf_size);

  if (*target_buf == '\0')

    error (_("Target does not support this command."));

  if (strcmp (target_buf, "OK") != 0)

    error (_("Bogus reply from target: %s"), target_buf);

}



static int

‌remote_get_trace_status (struct target_ops *self, struct trace_status *ts)

{

  /* Initialize it just to avoid a GCC false warning.  */

  char *p = NULL;

  /* FIXME we need to get register block size some other way.  */

  extern int trace_regblock_size;

  volatile struct gdb_exception ex;

  enum packet_result result;



  if (packet_support (PACKET_qTStatus) == PACKET_DISABLE)

    return -1;



  trace_regblock_size = get_remote_arch_state ()->sizeof_g_packet;



  putpkt ("qTStatus");



  TRY_CATCH (ex, RETURN_MASK_ERROR)

    {

      p = remote_get_noisy_reply (&target_buf, &target_buf_size);

    }

  if (ex.reason < 0)

    {

      if (ex.error != TARGET_CLOSE_ERROR)

        {

          exception_fprintf (gdb_stderr, ex, "qTStatus: ");

          return -1;

        }

      throw_exception (ex);

    }



  result = packet_ok (p, &remote_protocol_packets[PACKET_qTStatus]);



  /* If the remote target doesn't do tracing, flag it.  */

  if (result == PACKET_UNKNOWN)

    return -1;



  /* We're working with a live target.  */

  ts->filename = NULL;



  if (*p++ != 'T')

    error (_("Bogus trace status reply from target: %s"), target_buf);



  /* Function 'parse_trace_status' sets default value of each field of

     'ts' at first, so we don't have to do it here.  */

  parse_trace_status (p, ts);



  return ts->running;

}



static void

‌remote_get_tracepoint_status (struct target_ops *self, struct breakpoint *bp,

                              struct uploaded_tp *utp)

{

  struct remote_state *rs = get_remote_state ();

  char *reply;

  struct bp_location *loc;

  struct tracepoint *tp = (struct tracepoint *) bp;

  size_t size = get_remote_packet_size ();



  if (tp)

    {

      tp->base.hit_count = 0;

      tp->traceframe_usage = 0;

      for (loc = tp->base.loc; loc; loc = loc->next)

        {

          /* If the tracepoint was never downloaded, don't go asking for

             any status.  */

          if (tp->number_on_target == 0)

            continue;

          xsnprintf (rs->buf, size, "qTP:%x:%s", tp->number_on_target,

                     phex_nz (loc->address, 0));

          putpkt (rs->buf);

          reply = remote_get_noisy_reply (&target_buf, &target_buf_size);

          if (reply && *reply)

            {

              if (*reply == 'V')

                parse_tracepoint_status (reply + 1, bp, utp);

            }

        }

    }

  else if (utp)

    {

      utp->hit_count = 0;

      utp->traceframe_usage = 0;

      xsnprintf (rs->buf, size, "qTP:%x:%s", utp->number,

                 phex_nz (utp->addr, 0));

      putpkt (rs->buf);

      reply = remote_get_noisy_reply (&target_buf, &target_buf_size);

      if (reply && *reply)

        {

          if (*reply == 'V')

            parse_tracepoint_status (reply + 1, bp, utp);

        }

    }

}



static void

‌remote_trace_stop (struct target_ops *self)

{

  putpkt ("QTStop");

  remote_get_noisy_reply (&target_buf, &target_buf_size);

  if (*target_buf == '\0')

    error (_("Target does not support this command."));

  if (strcmp (target_buf, "OK") != 0)

    error (_("Bogus reply from target: %s"), target_buf);

}



static int

‌remote_trace_find (struct target_ops *self,

                   enum trace_find_type type, int num,

                   CORE_ADDR addr1, CORE_ADDR addr2,

                   int *tpp)

{

  struct remote_state *rs = get_remote_state ();

  char *endbuf = rs->buf + get_remote_packet_size ();

  char *p, *reply;

  int target_frameno = -1, target_tracept = -1;



  /* Lookups other than by absolute frame number depend on the current

     trace selected, so make sure it is correct on the remote end

     first.  */

  if (type != tfind_number)

    set_remote_traceframe ();



  p = rs->buf;

  strcpy (p, "QTFrame:");

  p = strchr (p, '\0');

  switch (type)

    {

    case tfind_number:

      xsnprintf (p, endbuf - p, "%x", num);

      break;

    case tfind_pc:

      xsnprintf (p, endbuf - p, "pc:%s", phex_nz (addr1, 0));

      break;

    case tfind_tp:

      xsnprintf (p, endbuf - p, "tdp:%x", num);

      break;

    case tfind_range:

      xsnprintf (p, endbuf - p, "range:%s:%s", phex_nz (addr1, 0),

                 phex_nz (addr2, 0));

      break;

    case tfind_outside:

      xsnprintf (p, endbuf - p, "outside:%s:%s", phex_nz (addr1, 0),

                 phex_nz (addr2, 0));

      break;

    default:

      error (_("Unknown trace find type %d"), type);

    }



  putpkt (rs->buf);

  reply = remote_get_noisy_reply (&(rs->buf), &rs->buf_size);

  if (*reply == '\0')

    error (_("Target does not support this command."));



  while (reply && *reply)

    switch (*reply)

      {

      case 'F':

        p = ++reply;

        target_frameno = (int) strtol (p, &reply, 16);

        if (reply == p)

          error (_("Unable to parse trace frame number"));

        /* Don't update our remote traceframe number cache on failure

           to select a remote traceframe.  */

        if (target_frameno == -1)

          return -1;

        break;

      case 'T':

        p = ++reply;

        target_tracept = (int) strtol (p, &reply, 16);

        if (reply == p)

          error (_("Unable to parse tracepoint number"));

        break;

      case 'O':                /* "OK"? */

        if (reply[1] == 'K' && reply[2] == '\0')

          reply += 2;

        else

          error (_("Bogus reply from target: %s"), reply);

        break;

      default:

        error (_("Bogus reply from target: %s"), reply);

      }

  if (tpp)

    *tpp = target_tracept;



  rs->remote_traceframe_number = target_frameno;

  return target_frameno;

}



static int

‌remote_get_trace_state_variable_value (struct target_ops *self,

                                       int tsvnum, LONGEST *val)

{

  struct remote_state *rs = get_remote_state ();

  char *reply;

  ULONGEST uval;



  set_remote_traceframe ();



  xsnprintf (rs->buf, get_remote_packet_size (), "qTV:%x", tsvnum);

  putpkt (rs->buf);

  reply = remote_get_noisy_reply (&target_buf, &target_buf_size);

  if (reply && *reply)

    {

      if (*reply == 'V')

        {

          unpack_varlen_hex (reply + 1, &uval);

          *val = (LONGEST) uval;

          return 1;

        }

    }

  return 0;

}



static int

‌remote_save_trace_data (struct target_ops *self, const char *filename)

{

  struct remote_state *rs = get_remote_state ();

  char *p, *reply;



  p = rs->buf;

  strcpy (p, "QTSave:");

  p += strlen (p);

  if ((p - rs->buf) + strlen (filename) * 2 >= get_remote_packet_size ())

    error (_("Remote file name too long for trace save packet"));

  p += 2 * bin2hex ((gdb_byte *) filename, p, strlen (filename));

  *p++ = '\0';

  putpkt (rs->buf);

  reply = remote_get_noisy_reply (&target_buf, &target_buf_size);

  if (*reply == '\0')

    error (_("Target does not support this command."));

  if (strcmp (reply, "OK") != 0)

    error (_("Bogus reply from target: %s"), reply);

  return 0;

}



/* This is basically a memory transfer, but needs to be its own packet

   because we don't know how the target actually organizes its trace

   memory, plus we want to be able to ask for as much as possible, but

   not be unhappy if we don't get as much as we ask for.  */



static LONGEST

‌remote_get_raw_trace_data (struct target_ops *self,

                           gdb_byte *buf, ULONGEST offset, LONGEST len)

{

  struct remote_state *rs = get_remote_state ();

  char *reply;

  char *p;

  int rslt;



  p = rs->buf;

  strcpy (p, "qTBuffer:");

  p += strlen (p);

  p += hexnumstr (p, offset);

  *p++ = ',';

  p += hexnumstr (p, len);

  *p++ = '\0';



  putpkt (rs->buf);

  reply = remote_get_noisy_reply (&target_buf, &target_buf_size);

  if (reply && *reply)

    {

      /* 'l' by itself means we're at the end of the buffer and

         there is nothing more to get.  */

      if (*reply == 'l')

        return 0;



      /* Convert the reply into binary.  Limit the number of bytes to

         convert according to our passed-in buffer size, rather than

         what was returned in the packet; if the target is

         unexpectedly generous and gives us a bigger reply than we

         asked for, we don't want to crash.  */

      rslt = hex2bin (target_buf, buf, len);

      return rslt;

    }



  /* Something went wrong, flag as an error.  */

  return -1;

}



static void

‌remote_set_disconnected_tracing (struct target_ops *self, int val)

{

  struct remote_state *rs = get_remote_state ();



  if (packet_support (PACKET_DisconnectedTracing_feature) == PACKET_ENABLE)

    {

      char *reply;



      xsnprintf (rs->buf, get_remote_packet_size (), "QTDisconnected:%x", val);

      putpkt (rs->buf);

      reply = remote_get_noisy_reply (&target_buf, &target_buf_size);

      if (*reply == '\0')

        error (_("Target does not support this command."));

      if (strcmp (reply, "OK") != 0)

        error (_("Bogus reply from target: %s"), reply);

    }

  else if (val)

    warning (_("Target does not support disconnected tracing."));

}



static int

‌remote_core_of_thread (struct target_ops *ops, ptid_t ptid)

{

  struct thread_info *info = find_thread_ptid (ptid);



  if (info && info->private)

    return info->private->core;

  return -1;

}



static void

‌remote_set_circular_trace_buffer (struct target_ops *self, int val)

{

  struct remote_state *rs = get_remote_state ();

  char *reply;



  xsnprintf (rs->buf, get_remote_packet_size (), "QTBuffer:circular:%x", val);

  putpkt (rs->buf);

  reply = remote_get_noisy_reply (&target_buf, &target_buf_size);

  if (*reply == '\0')

    error (_("Target does not support this command."));

  if (strcmp (reply, "OK") != 0)

    error (_("Bogus reply from target: %s"), reply);

}



static struct traceframe_info *

‌remote_traceframe_info (struct target_ops *self)

{

  char *text;



  text = target_read_stralloc (&current_target,

                               TARGET_OBJECT_TRACEFRAME_INFO, NULL);

  if (text != NULL)

    {

      struct traceframe_info *info;

      struct cleanup *back_to = make_cleanup (xfree, text);



      info = parse_traceframe_info (text);

      do_cleanups (back_to);

      return info;

    }



  return NULL;

}



/* Handle the qTMinFTPILen packet.  Returns the minimum length of

   instruction on which a fast tracepoint may be placed.  Returns -1

   if the packet is not supported, and 0 if the minimum instruction

   length is unknown.  */



static int

‌remote_get_min_fast_tracepoint_insn_len (struct target_ops *self)

{

  struct remote_state *rs = get_remote_state ();

  char *reply;



  /* If we're not debugging a process yet, the IPA can't be

     loaded.  */

  if (!target_has_execution)

    return 0;



  /* Make sure the remote is pointing at the right process.  */

  set_general_process ();



  xsnprintf (rs->buf, get_remote_packet_size (), "qTMinFTPILen");

  putpkt (rs->buf);

  reply = remote_get_noisy_reply (&target_buf, &target_buf_size);

  if (*reply == '\0')

    return -1;

  else

    {

      ULONGEST min_insn_len;



      unpack_varlen_hex (reply, &min_insn_len);



      return (int) min_insn_len;

    }

}



static void

‌remote_set_trace_buffer_size (struct target_ops *self, LONGEST val)

{

  if (packet_support (PACKET_QTBuffer_size) != PACKET_DISABLE)

    {

      struct remote_state *rs = get_remote_state ();

      char *buf = rs->buf;

      char *endbuf = rs->buf + get_remote_packet_size ();

      enum packet_result result;



      gdb_assert (val >= 0 || val == -1);

      buf += xsnprintf (buf, endbuf - buf, "QTBuffer:size:");

      /* Send -1 as literal "-1" to avoid host size dependency.  */

      if (val < 0)

        {

          *buf++ = '-';

          buf += hexnumstr (buf, (ULONGEST) -val);

        }

      else

        buf += hexnumstr (buf, (ULONGEST) val);



      putpkt (rs->buf);

      remote_get_noisy_reply (&rs->buf, &rs->buf_size);

      result = packet_ok (rs->buf,

                  &remote_protocol_packets[PACKET_QTBuffer_size]);



      if (result != PACKET_OK)

        warning (_("Bogus reply from target: %s"), rs->buf);

    }

}



static int

‌remote_set_trace_notes (struct target_ops *self,

                        const char *user, const char *notes,

                        const char *stop_notes)

{

  struct remote_state *rs = get_remote_state ();

  char *reply;

  char *buf = rs->buf;

  char *endbuf = rs->buf + get_remote_packet_size ();

  int nbytes;



  buf += xsnprintf (buf, endbuf - buf, "QTNotes:");

  if (user)

    {

      buf += xsnprintf (buf, endbuf - buf, "user:");

      nbytes = bin2hex ((gdb_byte *) user, buf, strlen (user));

      buf += 2 * nbytes;

      *buf++ = ';';

    }

  if (notes)

    {

      buf += xsnprintf (buf, endbuf - buf, "notes:");

      nbytes = bin2hex ((gdb_byte *) notes, buf, strlen (notes));

      buf += 2 * nbytes;

      *buf++ = ';';

    }

  if (stop_notes)

    {

      buf += xsnprintf (buf, endbuf - buf, "tstop:");

      nbytes = bin2hex ((gdb_byte *) stop_notes, buf, strlen (stop_notes));

      buf += 2 * nbytes;

      *buf++ = ';';

    }

  /* Ensure the buffer is terminated.  */

  *buf = '\0';



  putpkt (rs->buf);

  reply = remote_get_noisy_reply (&target_buf, &target_buf_size);

  if (*reply == '\0')

    return 0;



  if (strcmp (reply, "OK") != 0)

    error (_("Bogus reply from target: %s"), reply);



  return 1;

}



static int

‌remote_use_agent (struct target_ops *self, int use)

{

  if (packet_support (PACKET_QAgent) != PACKET_DISABLE)

    {

      struct remote_state *rs = get_remote_state ();



      /* If the stub supports QAgent.  */

      xsnprintf (rs->buf, get_remote_packet_size (), "QAgent:%d", use);

      putpkt (rs->buf);

      getpkt (&rs->buf, &rs->buf_size, 0);



      if (strcmp (rs->buf, "OK") == 0)

        {

          use_agent = use;

          return 1;

        }

    }



  return 0;

}



static int

‌remote_can_use_agent (struct target_ops *self)

{

  return (packet_support (PACKET_QAgent) != PACKET_DISABLE);

}



‌struct btrace_target_info

{

  /* The ptid of the traced thread.  */

  ptid_t ptid;

};



/* Check whether the target supports branch tracing.  */



static int

‌remote_supports_btrace (struct target_ops *self)

{

  if (packet_support (PACKET_Qbtrace_off) != PACKET_ENABLE)

    return 0;

  if (packet_support (PACKET_Qbtrace_bts) != PACKET_ENABLE)

    return 0;

  if (packet_support (PACKET_qXfer_btrace) != PACKET_ENABLE)

    return 0;



  return 1;

}



/* Enable branch tracing.  */



static struct btrace_target_info *

‌remote_enable_btrace (struct target_ops *self, ptid_t ptid)

{

  struct btrace_target_info *tinfo = NULL;

  struct packet_config *packet = &remote_protocol_packets[PACKET_Qbtrace_bts];

  struct remote_state *rs = get_remote_state ();

  char *buf = rs->buf;

  char *endbuf = rs->buf + get_remote_packet_size ();



  if (packet_config_support (packet) != PACKET_ENABLE)

    error (_("Target does not support branch tracing."));



  set_general_thread (ptid);



  buf += xsnprintf (buf, endbuf - buf, "%s", packet->name);

  putpkt (rs->buf);

  getpkt (&rs->buf, &rs->buf_size, 0);



  if (packet_ok (rs->buf, packet) == PACKET_ERROR)

    {

      if (rs->buf[0] == 'E' && rs->buf[1] == '.')

        error (_("Could not enable branch tracing for %s: %s"),

               target_pid_to_str (ptid), rs->buf + 2);

      else

        error (_("Could not enable branch tracing for %s."),

               target_pid_to_str (ptid));

    }



  tinfo = xzalloc (sizeof (*tinfo));

  tinfo->ptid = ptid;



  return tinfo;

}



/* Disable branch tracing.  */



static void

‌remote_disable_btrace (struct target_ops *self,

                       struct btrace_target_info *tinfo)

{

  struct packet_config *packet = &remote_protocol_packets[PACKET_Qbtrace_off];

  struct remote_state *rs = get_remote_state ();

  char *buf = rs->buf;

  char *endbuf = rs->buf + get_remote_packet_size ();



  if (packet_config_support (packet) != PACKET_ENABLE)

    error (_("Target does not support branch tracing."));



  set_general_thread (tinfo->ptid);



  buf += xsnprintf (buf, endbuf - buf, "%s", packet->name);

  putpkt (rs->buf);

  getpkt (&rs->buf, &rs->buf_size, 0);



  if (packet_ok (rs->buf, packet) == PACKET_ERROR)

    {

      if (rs->buf[0] == 'E' && rs->buf[1] == '.')

        error (_("Could not disable branch tracing for %s: %s"),

               target_pid_to_str (tinfo->ptid), rs->buf + 2);

      else

        error (_("Could not disable branch tracing for %s."),

               target_pid_to_str (tinfo->ptid));

    }



  xfree (tinfo);

}



/* Teardown branch tracing.  */



static void

‌remote_teardown_btrace (struct target_ops *self,

                        struct btrace_target_info *tinfo)

{

  /* We must not talk to the target during teardown.  */

  xfree (tinfo);

}



/* Read the branch trace.  */



static enum btrace_error

‌remote_read_btrace (struct target_ops *self,

                    VEC (btrace_block_s) **btrace,

                    struct btrace_target_info *tinfo,

                    enum btrace_read_type type)

{

  struct packet_config *packet = &remote_protocol_packets[PACKET_qXfer_btrace];

  struct remote_state *rs = get_remote_state ();

  struct cleanup *cleanup;

  const char *annex;

  char *xml;



  if (packet_config_support (packet) != PACKET_ENABLE)

    error (_("Target does not support branch tracing."));



#if !defined(HAVE_LIBEXPAT)

  error (_("Cannot process branch tracing result. XML parsing not supported."));

#endif



  switch (type)

    {

    case BTRACE_READ_ALL:

      annex = "all";

      break;

    case BTRACE_READ_NEW:

      annex = "new";

      break;

    case BTRACE_READ_DELTA:

      annex = "delta";

      break;

    default:

      internal_error (__FILE__, __LINE__,

                      _("Bad branch tracing read type: %u."),

                      (unsigned int) type);

    }



  xml = target_read_stralloc (&current_target,

                              TARGET_OBJECT_BTRACE, annex);

  if (xml == NULL)

    return BTRACE_ERR_UNKNOWN;



  cleanup = make_cleanup (xfree, xml);

  *btrace = parse_xml_btrace (xml);

  do_cleanups (cleanup);



  return BTRACE_ERR_NONE;

}



static int

‌remote_augmented_libraries_svr4_read (struct target_ops *self)

{

  return (packet_support (PACKET_augmented_libraries_svr4_read_feature)

          == PACKET_ENABLE);

}



/* Implementation of to_load.  */



static void

‌remote_load (struct target_ops *self, const char *name, int from_tty)

{

  generic_load (name, from_tty);

}



static void

‌init_remote_ops (void)

{

  remote_ops.to_shortname = "remote";

  remote_ops.to_longname = "Remote serial target in gdb-specific protocol";

  remote_ops.to_doc =

    "Use a remote computer via a serial line, using a gdb-specific protocol.\n\

Specify the serial device it is connected to\n\

(e.g. /dev/ttyS0, /dev/ttya, COM1, etc.).";

  remote_ops.to_open = remote_open;

  remote_ops.to_close = remote_close;

  remote_ops.to_detach = remote_detach;

  remote_ops.to_disconnect = remote_disconnect;

  remote_ops.to_resume = remote_resume;

  remote_ops.to_wait = remote_wait;

  remote_ops.to_fetch_registers = remote_fetch_registers;

  remote_ops.to_store_registers = remote_store_registers;

  remote_ops.to_prepare_to_store = remote_prepare_to_store;

  remote_ops.to_files_info = remote_files_info;

  remote_ops.to_insert_breakpoint = remote_insert_breakpoint;

  remote_ops.to_remove_breakpoint = remote_remove_breakpoint;

  remote_ops.to_stopped_by_watchpoint = remote_stopped_by_watchpoint;

  remote_ops.to_stopped_data_address = remote_stopped_data_address;

  remote_ops.to_watchpoint_addr_within_range =

    remote_watchpoint_addr_within_range;

  remote_ops.to_can_use_hw_breakpoint = remote_check_watch_resources;

  remote_ops.to_insert_hw_breakpoint = remote_insert_hw_breakpoint;

  remote_ops.to_remove_hw_breakpoint = remote_remove_hw_breakpoint;

  remote_ops.to_region_ok_for_hw_watchpoint

     = remote_region_ok_for_hw_watchpoint;

  remote_ops.to_insert_watchpoint = remote_insert_watchpoint;

  remote_ops.to_remove_watchpoint = remote_remove_watchpoint;

  remote_ops.to_kill = remote_kill;

  remote_ops.to_load = remote_load;

  remote_ops.to_mourn_inferior = remote_mourn;

  remote_ops.to_pass_signals = remote_pass_signals;

  remote_ops.to_program_signals = remote_program_signals;

  remote_ops.to_thread_alive = remote_thread_alive;

  remote_ops.to_update_thread_list = remote_update_thread_list;

  remote_ops.to_pid_to_str = remote_pid_to_str;

  remote_ops.to_extra_thread_info = remote_threads_extra_info;

  remote_ops.to_get_ada_task_ptid = remote_get_ada_task_ptid;

  remote_ops.to_stop = remote_stop;

  remote_ops.to_xfer_partial = remote_xfer_partial;

  remote_ops.to_rcmd = remote_rcmd;

  remote_ops.to_log_command = serial_log_command;

  remote_ops.to_get_thread_local_address = remote_get_thread_local_address;

  remote_ops.to_stratum = process_stratum;

  remote_ops.to_has_all_memory = default_child_has_all_memory;

  remote_ops.to_has_memory = default_child_has_memory;

  remote_ops.to_has_stack = default_child_has_stack;

  remote_ops.to_has_registers = default_child_has_registers;

  remote_ops.to_has_execution = default_child_has_execution;

  remote_ops.to_has_thread_control = tc_schedlock;    /* can lock scheduler */

  remote_ops.to_can_execute_reverse = remote_can_execute_reverse;

  remote_ops.to_magic = OPS_MAGIC;

  remote_ops.to_memory_map = remote_memory_map;

  remote_ops.to_flash_erase = remote_flash_erase;

  remote_ops.to_flash_done = remote_flash_done;

  remote_ops.to_read_description = remote_read_description;

  remote_ops.to_search_memory = remote_search_memory;

  remote_ops.to_can_async_p = remote_can_async_p;

  remote_ops.to_is_async_p = remote_is_async_p;

  remote_ops.to_async = remote_async;

  remote_ops.to_terminal_inferior = remote_terminal_inferior;

  remote_ops.to_terminal_ours = remote_terminal_ours;

  remote_ops.to_supports_non_stop = remote_supports_non_stop;

  remote_ops.to_supports_multi_process = remote_supports_multi_process;

  remote_ops.to_supports_disable_randomization

    = remote_supports_disable_randomization;

  remote_ops.to_fileio_open = remote_hostio_open;

  remote_ops.to_fileio_pwrite = remote_hostio_pwrite;

  remote_ops.to_fileio_pread = remote_hostio_pread;

  remote_ops.to_fileio_close = remote_hostio_close;

  remote_ops.to_fileio_unlink = remote_hostio_unlink;

  remote_ops.to_fileio_readlink = remote_hostio_readlink;

  remote_ops.to_supports_enable_disable_tracepoint = remote_supports_enable_disable_tracepoint;

  remote_ops.to_supports_string_tracing = remote_supports_string_tracing;

  remote_ops.to_supports_evaluation_of_breakpoint_conditions = remote_supports_cond_breakpoints;

  remote_ops.to_can_run_breakpoint_commands = remote_can_run_breakpoint_commands;

  remote_ops.to_trace_init = remote_trace_init;

  remote_ops.to_download_tracepoint = remote_download_tracepoint;

  remote_ops.to_can_download_tracepoint = remote_can_download_tracepoint;

  remote_ops.to_download_trace_state_variable

    = remote_download_trace_state_variable;

  remote_ops.to_enable_tracepoint = remote_enable_tracepoint;

  remote_ops.to_disable_tracepoint = remote_disable_tracepoint;

  remote_ops.to_trace_set_readonly_regions = remote_trace_set_readonly_regions;

  remote_ops.to_trace_start = remote_trace_start;

  remote_ops.to_get_trace_status = remote_get_trace_status;

  remote_ops.to_get_tracepoint_status = remote_get_tracepoint_status;

  remote_ops.to_trace_stop = remote_trace_stop;

  remote_ops.to_trace_find = remote_trace_find;

  remote_ops.to_get_trace_state_variable_value

    = remote_get_trace_state_variable_value;

  remote_ops.to_save_trace_data = remote_save_trace_data;

  remote_ops.to_upload_tracepoints = remote_upload_tracepoints;

  remote_ops.to_upload_trace_state_variables

    = remote_upload_trace_state_variables;

  remote_ops.to_get_raw_trace_data = remote_get_raw_trace_data;

  remote_ops.to_get_min_fast_tracepoint_insn_len = remote_get_min_fast_tracepoint_insn_len;

  remote_ops.to_set_disconnected_tracing = remote_set_disconnected_tracing;

  remote_ops.to_set_circular_trace_buffer = remote_set_circular_trace_buffer;

  remote_ops.to_set_trace_buffer_size = remote_set_trace_buffer_size;

  remote_ops.to_set_trace_notes = remote_set_trace_notes;

  remote_ops.to_core_of_thread = remote_core_of_thread;

  remote_ops.to_verify_memory = remote_verify_memory;

  remote_ops.to_get_tib_address = remote_get_tib_address;

  remote_ops.to_set_permissions = remote_set_permissions;

  remote_ops.to_static_tracepoint_marker_at

    = remote_static_tracepoint_marker_at;

  remote_ops.to_static_tracepoint_markers_by_strid

    = remote_static_tracepoint_markers_by_strid;

  remote_ops.to_traceframe_info = remote_traceframe_info;

  remote_ops.to_use_agent = remote_use_agent;

  remote_ops.to_can_use_agent = remote_can_use_agent;

  remote_ops.to_supports_btrace = remote_supports_btrace;

  remote_ops.to_enable_btrace = remote_enable_btrace;

  remote_ops.to_disable_btrace = remote_disable_btrace;

  remote_ops.to_teardown_btrace = remote_teardown_btrace;

  remote_ops.to_read_btrace = remote_read_btrace;

  remote_ops.to_augmented_libraries_svr4_read =

    remote_augmented_libraries_svr4_read;

}



/* Set up the extended remote vector by making a copy of the standard

   remote vector and adding to it.  */



static void

‌init_extended_remote_ops (void)

{

  extended_remote_ops = remote_ops;



  extended_remote_ops.to_shortname = "extended-remote";

  extended_remote_ops.to_longname =

    "Extended remote serial target in gdb-specific protocol";

  extended_remote_ops.to_doc =

    "Use a remote computer via a serial line, using a gdb-specific protocol.\n\

Specify the serial device it is connected to (e.g. /dev/ttya).";

  extended_remote_ops.to_open = extended_remote_open;

  extended_remote_ops.to_create_inferior = extended_remote_create_inferior;

  extended_remote_ops.to_mourn_inferior = extended_remote_mourn;

  extended_remote_ops.to_detach = extended_remote_detach;

  extended_remote_ops.to_attach = extended_remote_attach;

  extended_remote_ops.to_post_attach = extended_remote_post_attach;

  extended_remote_ops.to_kill = extended_remote_kill;

  extended_remote_ops.to_supports_disable_randomization

    = extended_remote_supports_disable_randomization;

}



static int

‌remote_can_async_p (struct target_ops *ops)

{

  struct remote_state *rs = get_remote_state ();



  if (!target_async_permitted)

    /* We only enable async when the user specifically asks for it.  */

    return 0;



  /* We're async whenever the serial device is.  */

  return serial_can_async_p (rs->remote_desc);

}



static int

‌remote_is_async_p (struct target_ops *ops)

{

  struct remote_state *rs = get_remote_state ();



  if (!target_async_permitted)

    /* We only enable async when the user specifically asks for it.  */

    return 0;



  /* We're async whenever the serial device is.  */

  return serial_is_async_p (rs->remote_desc);

}



/* Pass the SERIAL event on and up to the client.  One day this code

   will be able to delay notifying the client of an event until the

   point where an entire packet has been received.  */



‌static serial_event_ftype remote_async_serial_handler;



static void

‌remote_async_serial_handler (struct serial *scb, void *context)

{

  struct remote_state *rs = context;



  /* Don't propogate error information up to the client.  Instead let

     the client find out about the error by querying the target.  */

  rs->async_client_callback (INF_REG_EVENT, rs->async_client_context);

}



static void

‌remote_async_inferior_event_handler (gdb_client_data data)

{

  inferior_event_handler (INF_REG_EVENT, NULL);

}



static void

‌remote_async (struct target_ops *ops,

              void (*callback) (enum inferior_event_type event_type,

                                void *context),

              void *context)

{

  struct remote_state *rs = get_remote_state ();



  if (callback != NULL)

    {

      serial_async (rs->remote_desc, remote_async_serial_handler, rs);

      rs->async_client_callback = callback;

      rs->async_client_context = context;

    }

  else

    serial_async (rs->remote_desc, NULL, NULL);

}



static void

‌set_remote_cmd (char *args, int from_tty)

{

  help_list (remote_set_cmdlist, "set remote ", all_commands, gdb_stdout);

}



static void

‌show_remote_cmd (char *args, int from_tty)

{

  /* We can't just use cmd_show_list here, because we want to skip

     the redundant "show remote Z-packet" and the legacy aliases.  */

  struct cleanup *showlist_chain;

  struct cmd_list_element *list = remote_show_cmdlist;

  struct ui_out *uiout = current_uiout;



  showlist_chain = make_cleanup_ui_out_tuple_begin_end (uiout, "showlist");

  for (; list != NULL; list = list->next)

    if (strcmp (list->name, "Z-packet") == 0)

      continue;

    else if (list->type == not_set_cmd)

      /* Alias commands are exactly like the original, except they

         don't have the normal type.  */

      continue;

    else

      {

        struct cleanup *option_chain

          = make_cleanup_ui_out_tuple_begin_end (uiout, "option");



        ui_out_field_string (uiout, "name", list->name);

        ui_out_text (uiout, ":  ");

        if (list->type == show_cmd)

          do_show_command ((char *) NULL, from_tty, list);

        else

          cmd_func (list, NULL, from_tty);

        /* Close the tuple.  */

        do_cleanups (option_chain);

      }



  /* Close the tuple.  */

  do_cleanups (showlist_chain);

}





/* Function to be called whenever a new objfile (shlib) is detected.  */

static void

‌remote_new_objfile (struct objfile *objfile)

{

  struct remote_state *rs = get_remote_state ();



  if (rs->remote_desc != 0)                /* Have a remote connection.  */

    remote_check_symbols ();

}



/* Pull all the tracepoints defined on the target and create local

   data structures representing them.  We don't want to create real

   tracepoints yet, we don't want to mess up the user's existing

   collection.  */



static int

‌remote_upload_tracepoints (struct target_ops *self, struct uploaded_tp **utpp)

{

  struct remote_state *rs = get_remote_state ();

  char *p;



  /* Ask for a first packet of tracepoint definition.  */

  putpkt ("qTfP");

  getpkt (&rs->buf, &rs->buf_size, 0);

  p = rs->buf;

  while (*p && *p != 'l')

    {

      parse_tracepoint_definition (p, utpp);

      /* Ask for another packet of tracepoint definition.  */

      putpkt ("qTsP");

      getpkt (&rs->buf, &rs->buf_size, 0);

      p = rs->buf;

    }

  return 0;

}



static int

‌remote_upload_trace_state_variables (struct target_ops *self,

                                     struct uploaded_tsv **utsvp)

{

  struct remote_state *rs = get_remote_state ();

  char *p;



  /* Ask for a first packet of variable definition.  */

  putpkt ("qTfV");

  getpkt (&rs->buf, &rs->buf_size, 0);

  p = rs->buf;

  while (*p && *p != 'l')

    {

      parse_tsv_definition (p, utsvp);

      /* Ask for another packet of variable definition.  */

      putpkt ("qTsV");

      getpkt (&rs->buf, &rs->buf_size, 0);

      p = rs->buf;

    }

  return 0;

}



/* The "set/show range-stepping" show hook.  */



static void

‌show_range_stepping (struct ui_file *file, int from_tty,

                     struct cmd_list_element *c,

                     const char *value)

{

  fprintf_filtered (file,

                    _("Debugger's willingness to use range stepping "

                      "is %s.\n"), value);

}



/* The "set/show range-stepping" set hook.  */



static void

‌set_range_stepping (char *ignore_args, int from_tty,

                    struct cmd_list_element *c)

{

  struct remote_state *rs = get_remote_state ();



  /* Whene enabling, check whether range stepping is actually

     supported by the target, and warn if not.  */

  if (use_range_stepping)

    {

      if (rs->remote_desc != NULL)

        {

          if (packet_support (PACKET_vCont) == PACKET_SUPPORT_UNKNOWN)

            remote_vcont_probe (rs);



          if (packet_support (PACKET_vCont) == PACKET_ENABLE

              && rs->supports_vCont.r)

            return;

        }



      warning (_("Range stepping is not supported by the current target"));

    }

}



void

‌_initialize_remote (void)

{

  struct remote_state *rs;

  struct cmd_list_element *cmd;

  const char *cmd_name;



  /* architecture specific data */

  remote_gdbarch_data_handle =

    gdbarch_data_register_post_init (init_remote_state);

  remote_g_packet_data_handle =

    gdbarch_data_register_pre_init (remote_g_packet_data_init);



  /* Initialize the per-target state.  At the moment there is only one

     of these, not one per target.  Only one target is active at a

     time.  */

  remote_state = new_remote_state ();



  init_remote_ops ();

  add_target (&remote_ops);



  init_extended_remote_ops ();

  add_target (&extended_remote_ops);



  /* Hook into new objfile notification.  */

  observer_attach_new_objfile (remote_new_objfile);

  /* We're no longer interested in notification events of an inferior

     when it exits.  */

  observer_attach_inferior_exit (discard_pending_stop_replies);



  /* Set up signal handlers.  */

  async_sigint_remote_token =

    create_async_signal_handler (async_remote_interrupt, NULL);

  async_sigint_remote_twice_token =

    create_async_signal_handler (async_remote_interrupt_twice, NULL);



#if 0

  init_remote_threadtests ();

#endif



  stop_reply_queue = QUEUE_alloc (stop_reply_p, stop_reply_xfree);

  /* set/show remote ...  */



  add_prefix_cmd ("remote", class_maintenance, set_remote_cmd, _("\

Remote protocol specific variables\n\

Configure various remote-protocol specific variables such as\n\

the packets being used"),

                  &remote_set_cmdlist, "set remote ",

                  0 /* allow-unknown */, &setlist);

  add_prefix_cmd ("remote", class_maintenance, show_remote_cmd, _("\

Remote protocol specific variables\n\

Configure various remote-protocol specific variables such as\n\

the packets being used"),

                  &remote_show_cmdlist, "show remote ",

                  0 /* allow-unknown */, &showlist);



  add_cmd ("compare-sections", class_obscure, compare_sections_command, _("\

Compare section data on target to the exec file.\n\

Argument is a single section name (default: all loaded sections).\n\

To compare only read-only loaded sections, specify the -r option."),

           &cmdlist);



  add_cmd ("packet", class_maintenance, packet_command, _("\

Send an arbitrary packet to a remote target.\n\

   maintenance packet TEXT\n\

If GDB is talking to an inferior via the GDB serial protocol, then\n\

this command sends the string TEXT to the inferior, and displays the\n\

response packet.  GDB supplies the initial `$' character, and the\n\

terminating `#' character and checksum."),

           &maintenancelist);



  add_setshow_boolean_cmd ("remotebreak", no_class, &remote_break, _("\

Set whether to send break if interrupted."), _("\

Show whether to send break if interrupted."), _("\

If set, a break, instead of a cntrl-c, is sent to the remote target."),

                           set_remotebreak, show_remotebreak,

                           &setlist, &showlist);

  cmd_name = "remotebreak";

  cmd = lookup_cmd (&cmd_name, setlist, "", -1, 1);

  deprecate_cmd (cmd, "set remote interrupt-sequence");

  cmd_name = "remotebreak"; /* needed because lookup_cmd updates the pointer */

  cmd = lookup_cmd (&cmd_name, showlist, "", -1, 1);

  deprecate_cmd (cmd, "show remote interrupt-sequence");



  add_setshow_enum_cmd ("interrupt-sequence", class_support,

                        interrupt_sequence_modes, &interrupt_sequence_mode,

                        _("\

Set interrupt sequence to remote target."), _("\

Show interrupt sequence to remote target."), _("\

Valid value is \"Ctrl-C\", \"BREAK\" or \"BREAK-g\". The default is \"Ctrl-C\"."),

                        NULL, show_interrupt_sequence,

                        &remote_set_cmdlist,

                        &remote_show_cmdlist);



  add_setshow_boolean_cmd ("interrupt-on-connect", class_support,

                           &interrupt_on_connect, _("\

Set whether interrupt-sequence is sent to remote target when gdb connects to."), _("                \

Show whether interrupt-sequence is sent to remote target when gdb connects to."), _("                \

If set, interrupt sequence is sent to remote target."),

                           NULL, NULL,

                           &remote_set_cmdlist, &remote_show_cmdlist);



  /* Install commands for configuring memory read/write packets.  */



  add_cmd ("remotewritesize", no_class, set_memory_write_packet_size, _("\

Set the maximum number of bytes per memory write packet (deprecated)."),

           &setlist);

  add_cmd ("remotewritesize", no_class, show_memory_write_packet_size, _("\

Show the maximum number of bytes per memory write packet (deprecated)."),

           &showlist);

  add_cmd ("memory-write-packet-size", no_class,

           set_memory_write_packet_size, _("\

Set the maximum number of bytes per memory-write packet.\n\

Specify the number of bytes in a packet or 0 (zero) for the\n\

default packet size.  The actual limit is further reduced\n\

dependent on the target.  Specify ``fixed'' to disable the\n\

further restriction and ``limit'' to enable that restriction."),

           &remote_set_cmdlist);

  add_cmd ("memory-read-packet-size", no_class,

           set_memory_read_packet_size, _("\

Set the maximum number of bytes per memory-read packet.\n\

Specify the number of bytes in a packet or 0 (zero) for the\n\

default packet size.  The actual limit is further reduced\n\

dependent on the target.  Specify ``fixed'' to disable the\n\

further restriction and ``limit'' to enable that restriction."),

           &remote_set_cmdlist);

  add_cmd ("memory-write-packet-size", no_class,

           show_memory_write_packet_size,

           _("Show the maximum number of bytes per memory-write packet."),

           &remote_show_cmdlist);

  add_cmd ("memory-read-packet-size", no_class,

           show_memory_read_packet_size,

           _("Show the maximum number of bytes per memory-read packet."),

           &remote_show_cmdlist);



  add_setshow_zinteger_cmd ("hardware-watchpoint-limit", no_class,

                            &remote_hw_watchpoint_limit, _("\

Set the maximum number of target hardware watchpoints."), _("\

Show the maximum number of target hardware watchpoints."), _("\

Specify a negative limit for unlimited."),

                            NULL, NULL, /* FIXME: i18n: The maximum

                                           number of target hardware

                                           watchpoints is %s.  */

                            &remote_set_cmdlist, &remote_show_cmdlist);

  add_setshow_zinteger_cmd ("hardware-watchpoint-length-limit", no_class,

                            &remote_hw_watchpoint_length_limit, _("\

Set the maximum length (in bytes) of a target hardware watchpoint."), _("\

Show the maximum length (in bytes) of a target hardware watchpoint."), _("\

Specify a negative limit for unlimited."),

                            NULL, NULL, /* FIXME: i18n: The maximum

                                           length (in bytes) of a target

                                           hardware watchpoint is %s.  */

                            &remote_set_cmdlist, &remote_show_cmdlist);

  add_setshow_zinteger_cmd ("hardware-breakpoint-limit", no_class,

                            &remote_hw_breakpoint_limit, _("\

Set the maximum number of target hardware breakpoints."), _("\

Show the maximum number of target hardware breakpoints."), _("\

Specify a negative limit for unlimited."),

                            NULL, NULL, /* FIXME: i18n: The maximum

                                           number of target hardware

                                           breakpoints is %s.  */

                            &remote_set_cmdlist, &remote_show_cmdlist);



  add_setshow_zuinteger_cmd ("remoteaddresssize", class_obscure,

                             &remote_address_size, _("\

Set the maximum size of the address (in bits) in a memory packet."), _("\

Show the maximum size of the address (in bits) in a memory packet."), NULL,

                             NULL,

                             NULL, /* FIXME: i18n: */

                             &setlist, &showlist);



  init_all_packet_configs ();



  add_packet_config_cmd (&remote_protocol_packets[PACKET_X],

                         "X", "binary-download", 1);



  add_packet_config_cmd (&remote_protocol_packets[PACKET_vCont],

                         "vCont", "verbose-resume", 0);



  add_packet_config_cmd (&remote_protocol_packets[PACKET_QPassSignals],

                         "QPassSignals", "pass-signals", 0);



  add_packet_config_cmd (&remote_protocol_packets[PACKET_QProgramSignals],

                         "QProgramSignals", "program-signals", 0);



  add_packet_config_cmd (&remote_protocol_packets[PACKET_qSymbol],

                         "qSymbol", "symbol-lookup", 0);



  add_packet_config_cmd (&remote_protocol_packets[PACKET_P],

                         "P", "set-register", 1);



  add_packet_config_cmd (&remote_protocol_packets[PACKET_p],

                         "p", "fetch-register", 1);



  add_packet_config_cmd (&remote_protocol_packets[PACKET_Z0],

                         "Z0", "software-breakpoint", 0);



  add_packet_config_cmd (&remote_protocol_packets[PACKET_Z1],

                         "Z1", "hardware-breakpoint", 0);



  add_packet_config_cmd (&remote_protocol_packets[PACKET_Z2],

                         "Z2", "write-watchpoint", 0);



  add_packet_config_cmd (&remote_protocol_packets[PACKET_Z3],

                         "Z3", "read-watchpoint", 0);



  add_packet_config_cmd (&remote_protocol_packets[PACKET_Z4],

                         "Z4", "access-watchpoint", 0);



  add_packet_config_cmd (&remote_protocol_packets[PACKET_qXfer_auxv],

                         "qXfer:auxv:read", "read-aux-vector", 0);



  add_packet_config_cmd (&remote_protocol_packets[PACKET_qXfer_features],

                         "qXfer:features:read", "target-features", 0);



  add_packet_config_cmd (&remote_protocol_packets[PACKET_qXfer_libraries],

                         "qXfer:libraries:read", "library-info", 0);



  add_packet_config_cmd (&remote_protocol_packets[PACKET_qXfer_libraries_svr4],

                         "qXfer:libraries-svr4:read", "library-info-svr4", 0);



  add_packet_config_cmd (&remote_protocol_packets[PACKET_qXfer_memory_map],

                         "qXfer:memory-map:read", "memory-map", 0);



  add_packet_config_cmd (&remote_protocol_packets[PACKET_qXfer_spu_read],

                         "qXfer:spu:read", "read-spu-object", 0);



  add_packet_config_cmd (&remote_protocol_packets[PACKET_qXfer_spu_write],

                         "qXfer:spu:write", "write-spu-object", 0);



  add_packet_config_cmd (&remote_protocol_packets[PACKET_qXfer_osdata],

                        "qXfer:osdata:read", "osdata", 0);



  add_packet_config_cmd (&remote_protocol_packets[PACKET_qXfer_threads],

                         "qXfer:threads:read", "threads", 0);



  add_packet_config_cmd (&remote_protocol_packets[PACKET_qXfer_siginfo_read],

                         "qXfer:siginfo:read", "read-siginfo-object", 0);



  add_packet_config_cmd (&remote_protocol_packets[PACKET_qXfer_siginfo_write],

                         "qXfer:siginfo:write", "write-siginfo-object", 0);



  add_packet_config_cmd

    (&remote_protocol_packets[PACKET_qXfer_traceframe_info],

     "qXfer:traceframe-info:read", "traceframe-info", 0);



  add_packet_config_cmd (&remote_protocol_packets[PACKET_qXfer_uib],

                         "qXfer:uib:read", "unwind-info-block", 0);



  add_packet_config_cmd (&remote_protocol_packets[PACKET_qGetTLSAddr],

                         "qGetTLSAddr", "get-thread-local-storage-address",

                         0);



  add_packet_config_cmd (&remote_protocol_packets[PACKET_qGetTIBAddr],

                         "qGetTIBAddr", "get-thread-information-block-address",

                         0);



  add_packet_config_cmd (&remote_protocol_packets[PACKET_bc],

                         "bc", "reverse-continue", 0);



  add_packet_config_cmd (&remote_protocol_packets[PACKET_bs],

                         "bs", "reverse-step", 0);



  add_packet_config_cmd (&remote_protocol_packets[PACKET_qSupported],

                         "qSupported", "supported-packets", 0);



  add_packet_config_cmd (&remote_protocol_packets[PACKET_qSearch_memory],

                         "qSearch:memory", "search-memory", 0);



  add_packet_config_cmd (&remote_protocol_packets[PACKET_qTStatus],

                         "qTStatus", "trace-status", 0);



  add_packet_config_cmd (&remote_protocol_packets[PACKET_vFile_open],

                         "vFile:open", "hostio-open", 0);



  add_packet_config_cmd (&remote_protocol_packets[PACKET_vFile_pread],

                         "vFile:pread", "hostio-pread", 0);



  add_packet_config_cmd (&remote_protocol_packets[PACKET_vFile_pwrite],

                         "vFile:pwrite", "hostio-pwrite", 0);



  add_packet_config_cmd (&remote_protocol_packets[PACKET_vFile_close],

                         "vFile:close", "hostio-close", 0);



  add_packet_config_cmd (&remote_protocol_packets[PACKET_vFile_unlink],

                         "vFile:unlink", "hostio-unlink", 0);



  add_packet_config_cmd (&remote_protocol_packets[PACKET_vFile_readlink],

                         "vFile:readlink", "hostio-readlink", 0);



  add_packet_config_cmd (&remote_protocol_packets[PACKET_vAttach],

                         "vAttach", "attach", 0);



  add_packet_config_cmd (&remote_protocol_packets[PACKET_vRun],

                         "vRun", "run", 0);



  add_packet_config_cmd (&remote_protocol_packets[PACKET_QStartNoAckMode],

                         "QStartNoAckMode", "noack", 0);



  add_packet_config_cmd (&remote_protocol_packets[PACKET_vKill],

                         "vKill", "kill", 0);



  add_packet_config_cmd (&remote_protocol_packets[PACKET_qAttached],

                         "qAttached", "query-attached", 0);



  add_packet_config_cmd (&remote_protocol_packets[PACKET_ConditionalTracepoints],

                         "ConditionalTracepoints",

                         "conditional-tracepoints", 0);



  add_packet_config_cmd (&remote_protocol_packets[PACKET_ConditionalBreakpoints],

                         "ConditionalBreakpoints",

                         "conditional-breakpoints", 0);



  add_packet_config_cmd (&remote_protocol_packets[PACKET_BreakpointCommands],

                         "BreakpointCommands",

                         "breakpoint-commands", 0);



  add_packet_config_cmd (&remote_protocol_packets[PACKET_FastTracepoints],

                         "FastTracepoints", "fast-tracepoints", 0);



  add_packet_config_cmd (&remote_protocol_packets[PACKET_TracepointSource],

                         "TracepointSource", "TracepointSource", 0);



  add_packet_config_cmd (&remote_protocol_packets[PACKET_QAllow],

                         "QAllow", "allow", 0);



  add_packet_config_cmd (&remote_protocol_packets[PACKET_StaticTracepoints],

                         "StaticTracepoints", "static-tracepoints", 0);



  add_packet_config_cmd (&remote_protocol_packets[PACKET_InstallInTrace],

                         "InstallInTrace", "install-in-trace", 0);



  add_packet_config_cmd (&remote_protocol_packets[PACKET_qXfer_statictrace_read],

                         "qXfer:statictrace:read", "read-sdata-object", 0);



  add_packet_config_cmd (&remote_protocol_packets[PACKET_qXfer_fdpic],

                         "qXfer:fdpic:read", "read-fdpic-loadmap", 0);



  add_packet_config_cmd (&remote_protocol_packets[PACKET_QDisableRandomization],

                         "QDisableRandomization", "disable-randomization", 0);



  add_packet_config_cmd (&remote_protocol_packets[PACKET_QAgent],

                         "QAgent", "agent", 0);



  add_packet_config_cmd (&remote_protocol_packets[PACKET_QTBuffer_size],

                         "QTBuffer:size", "trace-buffer-size", 0);



  add_packet_config_cmd (&remote_protocol_packets[PACKET_Qbtrace_off],

       "Qbtrace:off", "disable-btrace", 0);



  add_packet_config_cmd (&remote_protocol_packets[PACKET_Qbtrace_bts],

       "Qbtrace:bts", "enable-btrace", 0);



  add_packet_config_cmd (&remote_protocol_packets[PACKET_qXfer_btrace],

       "qXfer:btrace", "read-btrace", 0);



  /* Assert that we've registered commands for all packet configs.  */

  {

    int i;



    for (i = 0; i < PACKET_MAX; i++)

      {

        /* Ideally all configs would have a command associated.  Some

           still don't though.  */

        int excepted;



        switch (i)

          {

          case PACKET_QNonStop:

          case PACKET_multiprocess_feature:

          case PACKET_EnableDisableTracepoints_feature:

          case PACKET_tracenz_feature:

          case PACKET_DisconnectedTracing_feature:

          case PACKET_augmented_libraries_svr4_read_feature:

          case PACKET_qCRC:

            /* Additions to this list need to be well justified:

               pre-existing packets are OK; new packets are not.  */

            excepted = 1;

            break;

          default:

            excepted = 0;

            break;

          }



        /* This catches both forgetting to add a config command, and

           forgetting to remove a packet from the exception list.  */

        gdb_assert (excepted == (remote_protocol_packets[i].name == NULL));

      }

  }



  /* Keep the old ``set remote Z-packet ...'' working.  Each individual

     Z sub-packet has its own set and show commands, but users may

     have sets to this variable in their .gdbinit files (or in their

     documentation).  */

  add_setshow_auto_boolean_cmd ("Z-packet", class_obscure,

                                &remote_Z_packet_detect, _("\

Set use of remote protocol `Z' packets"), _("\

Show use of remote protocol `Z' packets "), _("\

When set, GDB will attempt to use the remote breakpoint and watchpoint\n\

packets."),

                                set_remote_protocol_Z_packet_cmd,

                                show_remote_protocol_Z_packet_cmd,

                                /* FIXME: i18n: Use of remote protocol

                                   `Z' packets is %s.  */

                                &remote_set_cmdlist, &remote_show_cmdlist);



  add_prefix_cmd ("remote", class_files, remote_command, _("\

Manipulate files on the remote system\n\

Transfer files to and from the remote target system."),

                  &remote_cmdlist, "remote ",

                  0 /* allow-unknown */, &cmdlist);



  add_cmd ("put", class_files, remote_put_command,

           _("Copy a local file to the remote system."),

           &remote_cmdlist);



  add_cmd ("get", class_files, remote_get_command,

           _("Copy a remote file to the local system."),

           &remote_cmdlist);



  add_cmd ("delete", class_files, remote_delete_command,

           _("Delete a remote file."),

           &remote_cmdlist);



  remote_exec_file = xstrdup ("");

  add_setshow_string_noescape_cmd ("exec-file", class_files,

                                   &remote_exec_file, _("\

Set the remote pathname for \"run\""), _("\

Show the remote pathname for \"run\""), NULL, NULL, NULL,

                                   &remote_set_cmdlist, &remote_show_cmdlist);



  add_setshow_boolean_cmd ("range-stepping", class_run,

                           &use_range_stepping, _("\

Enable or disable range stepping."), _("\

Show whether target-assisted range stepping is enabled."), _("\

If on, and the target supports it, when stepping a source line, GDB\n\

tells the target to step the corresponding range of addresses itself instead\n\

of issuing multiple single-steps.  This speeds up source level\n\

stepping.  If off, GDB always issues single-steps, even if range\n\

stepping is supported by the target.  The default is on."),

                           set_range_stepping,

                           show_range_stepping,

                           &setlist,

                           &showlist);



  /* Eventually initialize fileio.  See fileio.c */

  initialize_remote_fileio (remote_set_cmdlist, remote_show_cmdlist);



  /* Take advantage of the fact that the TID field is not used, to tag

     special ptids with it set to != 0.  */

  magic_null_ptid = ptid_build (42000, -1, 1);

  not_sent_ptid = ptid_build (42000, -2, 1);

  any_thread_ptid = ptid_build (42000, 0, 1);



  target_buf_size = 2048;

  target_buf = xmalloc (target_buf_size);

}
gdb/remote.c - gdb

Global variables defined

Data types defined

Functions defined

Macros defined

Source code
