/* Remote target communications for serial-line targets in custom GDB protocol Copyright (C) 1988-2023 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 . */ /* See the GDB User Guide for details of the GDB remote protocol. */ #include "defs.h" #include #include #include "inferior.h" #include "infrun.h" #include "bfd.h" #include "symfile.h" #include "target.h" #include "process-stratum-target.h" #include "gdbcmd.h" #include "objfiles.h" #include "gdbthread.h" #include "remote.h" #include "remote-notif.h" #include "regcache.h" #include "value.h" #include "observable.h" #include "solib.h" #include "cli/cli-decode.h" #include "cli/cli-setshow.h" #include "target-descriptions.h" #include "gdb_bfd.h" #include "gdbsupport/filestuff.h" #include "gdbsupport/rsp-low.h" #include "disasm.h" #include "location.h" #include "gdbsupport/gdb_sys_time.h" #include "gdbsupport/event-loop.h" #include "event-top.h" #include "inf-loop.h" #include #include "serial.h" #include "gdbcore.h" #include "remote-fileio.h" #include "gdbsupport/fileio.h" #include #include "xml-support.h" #include "memory-map.h" #include "tracepoint.h" #include "ax.h" #include "ax-gdb.h" #include "gdbsupport/agent.h" #include "btrace.h" #include "record-btrace.h" #include "gdbsupport/scoped_restore.h" #include "gdbsupport/environ.h" #include "gdbsupport/byte-vector.h" #include "gdbsupport/search.h" #include #include #include #include "async-event.h" #include "gdbsupport/selftest.h" /* The remote target. */ static const char remote_doc[] = N_("\ 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.)."); /* See remote.h */ bool remote_debug = false; #define OPAQUETHREADBYTES 8 /* a 64 bit opaque identifier */ typedef unsigned char threadref[OPAQUETHREADBYTES]; struct gdb_ext_thread_info; struct threads_listing_context; typedef int (*rmt_thread_action) (threadref *ref, void *context); struct protocol_feature; struct packet_reg; struct stop_reply; typedef std::unique_ptr stop_reply_up; /* 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 }; /* Convert the packet support auto_boolean to a name used for gdb printing. */ static const char * get_packet_support_name (auto_boolean support) { switch (support) { case AUTO_BOOLEAN_TRUE: return "on"; case AUTO_BOOLEAN_FALSE: return "off"; case AUTO_BOOLEAN_AUTO: return "auto"; default: gdb_assert_not_reached ("invalid var_auto_boolean"); } } /* Convert the target type (future remote target or currently connected target) to a name used for gdb printing. */ static const char * get_target_type_name (bool target_connected) { if (target_connected) return _("on the current remote target"); else return _("on future remote targets"); } /* Analyze a packet's return value and update the packet config accordingly. */ enum packet_result { PACKET_ERROR, PACKET_OK, PACKET_UNKNOWN }; /* Enumeration of packets for a remote target. */ enum { PACKET_vCont = 0, PACKET_X, PACKET_qSymbol, PACKET_P, PACKET_p, PACKET_Z0, PACKET_Z1, PACKET_Z2, PACKET_Z3, PACKET_Z4, PACKET_vFile_setfs, PACKET_vFile_open, PACKET_vFile_pread, PACKET_vFile_pwrite, PACKET_vFile_close, PACKET_vFile_unlink, PACKET_vFile_readlink, PACKET_vFile_fstat, PACKET_qXfer_auxv, PACKET_qXfer_features, PACKET_qXfer_exec_file, PACKET_qXfer_libraries, PACKET_qXfer_libraries_svr4, PACKET_qXfer_memory_map, 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_QCatchSyscalls, PACKET_QProgramSignals, PACKET_QSetWorkingDir, PACKET_QStartupWithShell, PACKET_QEnvironmentHexEncoded, PACKET_QEnvironmentReset, PACKET_QEnvironmentUnset, 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_Qbtrace_pt, PACKET_qXfer_btrace, /* Support for the QNonStop packet. */ PACKET_QNonStop, /* Support for the QThreadEvents packet. */ PACKET_QThreadEvents, /* 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, /* Support for the qXfer:btrace-conf:read packet. */ PACKET_qXfer_btrace_conf, /* Support for the Qbtrace-conf:bts:size packet. */ PACKET_Qbtrace_conf_bts_size, /* Support for swbreak+ feature. */ PACKET_swbreak_feature, /* Support for hwbreak+ feature. */ PACKET_hwbreak_feature, /* Support for fork events. */ PACKET_fork_event_feature, /* Support for vfork events. */ PACKET_vfork_event_feature, /* Support for the Qbtrace-conf:pt:size packet. */ PACKET_Qbtrace_conf_pt_size, /* Support for exec events. */ PACKET_exec_event_feature, /* Support for query supported vCont actions. */ PACKET_vContSupported, /* Support remote CTRL-C. */ PACKET_vCtrlC, /* Support TARGET_WAITKIND_NO_RESUMED. */ PACKET_no_resumed, /* Support for memory tagging, allocation tag fetch/store packets and the tag violation stop replies. */ PACKET_memory_tagging_feature, PACKET_MAX }; struct threads_listing_context; /* 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 */ bool t = false; /* vCont;r */ bool r = false; /* vCont;s */ bool s = false; /* vCont;S */ bool S = false; }; /* About this many threadids fit in a packet. */ #define MAXTHREADLISTRESULTS 32 /* Data for the vFile:pread readahead cache. */ struct readahead_cache { /* Invalidate the readahead cache. */ void invalidate (); /* Invalidate the readahead cache if it is holding data for FD. */ void invalidate_fd (int fd); /* Serve pread from the readahead cache. Returns number of bytes read, or 0 if the request can't be served from the cache. */ int pread (int fd, gdb_byte *read_buf, size_t len, ULONGEST offset); /* The file descriptor for the file that is being cached. -1 if the cache is invalid. */ int fd = -1; /* The offset into the file that the cache buffer corresponds to. */ ULONGEST offset = 0; /* The buffer holding the cache contents. */ gdb_byte *buf = nullptr; /* The buffer's size. We try to read as much as fits into a packet at a time. */ size_t bufsize = 0; /* Cache hit and miss counters. */ ULONGEST hit_count = 0; ULONGEST miss_count = 0; }; /* 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 { explicit remote_arch_state (struct gdbarch *gdbarch); /* 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). */ std::unique_ptr 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; }; /* Description of the remote protocol state for the currently connected target. This is per-target state, and independent of the selected architecture. */ class remote_state { public: remote_state (); ~remote_state (); /* Get the remote arch state for GDBARCH. */ struct remote_arch_state *get_remote_arch_state (struct gdbarch *gdbarch); public: /* data */ /* 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. The size of the buffer is always at least REMOTE_PACKET_SIZE; REMOTE_PACKET_SIZE should be used to limit the length of outgoing packets. */ gdb::char_vector buf; /* True if we're going through initial connection setup (finding out about the remote side's threads, relocating symbols, etc.). */ bool starting_up = false; /* 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 = 0; /* 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. */ bool noack_mode = false; /* True if we're connected in extended remote mode. */ bool extended = false; /* 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. */ bool waiting_for_stop_reply = false; /* The status of the stub support for the various vCont actions. */ vCont_action_support supports_vCont; /* True if the user has pressed Ctrl-C, but the target hasn't responded to that. */ bool ctrlc_pending_p = false; /* True if we saw a Ctrl-C while reading or writing from/to the remote descriptor. At that point it is not safe to send a remote interrupt packet, so we instead remember we saw the Ctrl-C and process it once we're done with sending/receiving the current packet, which should be shortly. If however that takes too long, and the user presses Ctrl-C again, we offer to disconnect. */ bool got_ctrlc_during_io = false; /* 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 = nullptr; /* 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 = null_ptid; ptid_t continue_thread = null_ptid; /* 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 = -1; char *last_pass_packet = nullptr; /* 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 = nullptr; gdb_signal last_sent_signal = GDB_SIGNAL_0; bool last_sent_step = false; /* The execution direction of the last resume we got. */ exec_direction_kind last_resume_exec_dir = EXEC_FORWARD; char *finished_object = nullptr; char *finished_annex = nullptr; ULONGEST finished_offset = 0; /* 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). */ bool use_threadinfo_query = false; bool use_threadextra_query = false; threadref echo_nextthread {}; threadref nextthread {}; threadref resultthreadlist[MAXTHREADLISTRESULTS] {}; /* The state of remote notification. */ struct remote_notif_state *notif_state = nullptr; /* The branch trace configuration. */ struct btrace_config btrace_config {}; /* The argument to the last "vFile:setfs:" packet we sent, used to avoid sending repeated unnecessary "vFile:setfs:" packets. Initialized to -1 to indicate that no "vFile:setfs:" packet has yet been sent. */ int fs_pid = -1; /* A readahead cache for vFile:pread. Often, reading a binary involves a sequence of small reads. E.g., when parsing an ELF file. A readahead cache helps mostly the case of remote debugging on a connection with higher latency, due to the request/reply nature of the RSP. We only cache data for a single file descriptor at a time. */ struct readahead_cache readahead_cache; /* 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. */ std::vector stop_reply_queue; /* Asynchronous signal handle registered as event loop source for when we have pending events ready to be passed to the core. */ struct async_event_handler *remote_async_inferior_event_token = nullptr; /* 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. */ int wait_forever_enabled_p = 1; private: /* Mapping of remote protocol data for each gdbarch. Usually there is only one entry here, though we may see more with stubs that support multi-process. */ std::unordered_map m_arch_states; }; static const target_info remote_target_info = { "remote", N_("Remote target using gdb-specific protocol"), remote_doc }; /* Description of a remote packet. */ struct packet_description { /* Name of the packet used for gdb output. */ const char *name; /* Title of the packet, used by the set/show remote name-packet commands to identify the individual packages and gdb output. */ const char *title; }; /* Configuration of a remote packet. */ struct packet_config { /* 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; }; /* 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 { const char *name; long size; int fixed_p; }; /* These global variables contain the default configuration for every new remote_feature object. */ static memory_packet_config memory_read_packet_config = { "memory-read-packet-size", }; static memory_packet_config memory_write_packet_config = { "memory-write-packet-size", }; /* This global array contains packet descriptions (name and title). */ static packet_description packets_descriptions[PACKET_MAX]; /* This global array contains the default configuration for every new per-remote target array. */ static packet_config remote_protocol_packets[PACKET_MAX]; /* Description of a remote target's features. It stores the configuration and provides functions to determine supported features of the target. */ struct remote_features { remote_features () { m_memory_read_packet_config = memory_read_packet_config; m_memory_write_packet_config = memory_write_packet_config; std::copy (std::begin (remote_protocol_packets), std::end (remote_protocol_packets), std::begin (m_protocol_packets)); } ~remote_features () = default; DISABLE_COPY_AND_ASSIGN (remote_features); /* Returns whether a given packet defined by its enum value is supported. */ enum packet_support packet_support (int) const; /* Returns the packet's corresponding "set remote foo-packet" command state. See struct packet_config for more details. */ enum auto_boolean packet_set_cmd_state (int packet) const { return m_protocol_packets[packet].detect; } /* Returns true if the multi-process extensions are in effect. */ int remote_multi_process_p () const { return packet_support (PACKET_multiprocess_feature) == PACKET_ENABLE; } /* Returns true if fork events are supported. */ int remote_fork_event_p () const { return packet_support (PACKET_fork_event_feature) == PACKET_ENABLE; } /* Returns true if vfork events are supported. */ int remote_vfork_event_p () const { return packet_support (PACKET_vfork_event_feature) == PACKET_ENABLE; } /* Returns true if exec events are supported. */ int remote_exec_event_p () const { return packet_support (PACKET_exec_event_feature) == PACKET_ENABLE; } /* Returns true if memory tagging is supported, false otherwise. */ bool remote_memory_tagging_p () const { return packet_support (PACKET_memory_tagging_feature) == PACKET_ENABLE; } /* Reset all packets back to "unknown support". Called when opening a new connection to a remote target. */ void reset_all_packet_configs_support (); /* Check result value in BUF for packet WHICH_PACKET and update the packet's support configuration accordingly. */ packet_result packet_ok (const char *buf, const int which_packet); packet_result packet_ok (const gdb::char_vector &buf, const int which_packet); /* Configuration of a remote target's memory read packet. */ memory_packet_config m_memory_read_packet_config; /* Configuration of a remote target's memory write packet. */ memory_packet_config m_memory_write_packet_config; /* The per-remote target array which stores a remote's packet configurations. */ packet_config m_protocol_packets[PACKET_MAX]; }; class remote_target : public process_stratum_target { public: remote_target () = default; ~remote_target () override; const target_info &info () const override { return remote_target_info; } const char *connection_string () override; thread_control_capabilities get_thread_control_capabilities () override { return tc_schedlock; } /* Open a remote connection. */ static void open (const char *, int); void close () override; void detach (inferior *, int) override; void disconnect (const char *, int) override; void commit_resumed () override; void resume (ptid_t, int, enum gdb_signal) override; ptid_t wait (ptid_t, struct target_waitstatus *, target_wait_flags) override; bool has_pending_events () override; void fetch_registers (struct regcache *, int) override; void store_registers (struct regcache *, int) override; void prepare_to_store (struct regcache *) override; int insert_breakpoint (struct gdbarch *, struct bp_target_info *) override; int remove_breakpoint (struct gdbarch *, struct bp_target_info *, enum remove_bp_reason) override; bool stopped_by_sw_breakpoint () override; bool supports_stopped_by_sw_breakpoint () override; bool stopped_by_hw_breakpoint () override; bool supports_stopped_by_hw_breakpoint () override; bool stopped_by_watchpoint () override; bool stopped_data_address (CORE_ADDR *) override; bool watchpoint_addr_within_range (CORE_ADDR, CORE_ADDR, int) override; int can_use_hw_breakpoint (enum bptype, int, int) override; int insert_hw_breakpoint (struct gdbarch *, struct bp_target_info *) override; int remove_hw_breakpoint (struct gdbarch *, struct bp_target_info *) override; int region_ok_for_hw_watchpoint (CORE_ADDR, int) override; int insert_watchpoint (CORE_ADDR, int, enum target_hw_bp_type, struct expression *) override; int remove_watchpoint (CORE_ADDR, int, enum target_hw_bp_type, struct expression *) override; void kill () override; void load (const char *, int) override; void mourn_inferior () override; void pass_signals (gdb::array_view) override; int set_syscall_catchpoint (int, bool, int, gdb::array_view) override; void program_signals (gdb::array_view) override; bool thread_alive (ptid_t ptid) override; const char *thread_name (struct thread_info *) override; void update_thread_list () override; std::string pid_to_str (ptid_t) override; const char *extra_thread_info (struct thread_info *) override; ptid_t get_ada_task_ptid (long lwp, ULONGEST thread) override; thread_info *thread_handle_to_thread_info (const gdb_byte *thread_handle, int handle_len, inferior *inf) override; gdb::byte_vector thread_info_to_thread_handle (struct thread_info *tp) override; void stop (ptid_t) override; void interrupt () override; void pass_ctrlc () override; enum target_xfer_status xfer_partial (enum target_object object, const char *annex, gdb_byte *readbuf, const gdb_byte *writebuf, ULONGEST offset, ULONGEST len, ULONGEST *xfered_len) override; ULONGEST get_memory_xfer_limit () override; void rcmd (const char *command, struct ui_file *output) override; const char *pid_to_exec_file (int pid) override; void log_command (const char *cmd) override { serial_log_command (this, cmd); } CORE_ADDR get_thread_local_address (ptid_t ptid, CORE_ADDR load_module_addr, CORE_ADDR offset) override; bool can_execute_reverse () override; std::vector memory_map () override; void flash_erase (ULONGEST address, LONGEST length) override; void flash_done () override; const struct target_desc *read_description () override; int search_memory (CORE_ADDR start_addr, ULONGEST search_space_len, const gdb_byte *pattern, ULONGEST pattern_len, CORE_ADDR *found_addrp) override; bool can_async_p () override; bool is_async_p () override; void async (bool) override; int async_wait_fd () override; void thread_events (int) override; int can_do_single_step () override; void terminal_inferior () override; void terminal_ours () override; bool supports_non_stop () override; bool supports_multi_process () override; bool supports_disable_randomization () override; bool filesystem_is_local () override; int fileio_open (struct inferior *inf, const char *filename, int flags, int mode, int warn_if_slow, fileio_error *target_errno) override; int fileio_pwrite (int fd, const gdb_byte *write_buf, int len, ULONGEST offset, fileio_error *target_errno) override; int fileio_pread (int fd, gdb_byte *read_buf, int len, ULONGEST offset, fileio_error *target_errno) override; int fileio_fstat (int fd, struct stat *sb, fileio_error *target_errno) override; int fileio_close (int fd, fileio_error *target_errno) override; int fileio_unlink (struct inferior *inf, const char *filename, fileio_error *target_errno) override; gdb::optional fileio_readlink (struct inferior *inf, const char *filename, fileio_error *target_errno) override; bool supports_enable_disable_tracepoint () override; bool supports_string_tracing () override; int remote_supports_cond_tracepoints (); bool supports_evaluation_of_breakpoint_conditions () override; int remote_supports_fast_tracepoints (); int remote_supports_static_tracepoints (); int remote_supports_install_in_trace (); bool can_run_breakpoint_commands () override; void trace_init () override; void download_tracepoint (struct bp_location *location) override; bool can_download_tracepoint () override; void download_trace_state_variable (const trace_state_variable &tsv) override; void enable_tracepoint (struct bp_location *location) override; void disable_tracepoint (struct bp_location *location) override; void trace_set_readonly_regions () override; void trace_start () override; int get_trace_status (struct trace_status *ts) override; void get_tracepoint_status (struct breakpoint *tp, struct uploaded_tp *utp) override; void trace_stop () override; int trace_find (enum trace_find_type type, int num, CORE_ADDR addr1, CORE_ADDR addr2, int *tpp) override; bool get_trace_state_variable_value (int tsv, LONGEST *val) override; int save_trace_data (const char *filename) override; int upload_tracepoints (struct uploaded_tp **utpp) override; int upload_trace_state_variables (struct uploaded_tsv **utsvp) override; LONGEST get_raw_trace_data (gdb_byte *buf, ULONGEST offset, LONGEST len) override; int get_min_fast_tracepoint_insn_len () override; void set_disconnected_tracing (int val) override; void set_circular_trace_buffer (int val) override; void set_trace_buffer_size (LONGEST val) override; bool set_trace_notes (const char *user, const char *notes, const char *stopnotes) override; int core_of_thread (ptid_t ptid) override; int verify_memory (const gdb_byte *data, CORE_ADDR memaddr, ULONGEST size) override; bool get_tib_address (ptid_t ptid, CORE_ADDR *addr) override; void set_permissions () override; bool static_tracepoint_marker_at (CORE_ADDR, struct static_tracepoint_marker *marker) override; std::vector static_tracepoint_markers_by_strid (const char *id) override; traceframe_info_up traceframe_info () override; bool use_agent (bool use) override; bool can_use_agent () override; struct btrace_target_info * enable_btrace (thread_info *tp, const struct btrace_config *conf) override; void disable_btrace (struct btrace_target_info *tinfo) override; void teardown_btrace (struct btrace_target_info *tinfo) override; enum btrace_error read_btrace (struct btrace_data *data, struct btrace_target_info *btinfo, enum btrace_read_type type) override; const struct btrace_config *btrace_conf (const struct btrace_target_info *) override; bool augmented_libraries_svr4_read () override; void follow_fork (inferior *, ptid_t, target_waitkind, bool, bool) override; void follow_exec (inferior *, ptid_t, const char *) override; int insert_fork_catchpoint (int) override; int remove_fork_catchpoint (int) override; int insert_vfork_catchpoint (int) override; int remove_vfork_catchpoint (int) override; int insert_exec_catchpoint (int) override; int remove_exec_catchpoint (int) override; enum exec_direction_kind execution_direction () override; bool supports_memory_tagging () override; bool fetch_memtags (CORE_ADDR address, size_t len, gdb::byte_vector &tags, int type) override; bool store_memtags (CORE_ADDR address, size_t len, const gdb::byte_vector &tags, int type) override; public: /* Remote specific methods. */ void remote_download_command_source (int num, ULONGEST addr, struct command_line *cmds); void remote_file_put (const char *local_file, const char *remote_file, int from_tty); void remote_file_get (const char *remote_file, const char *local_file, int from_tty); void remote_file_delete (const char *remote_file, int from_tty); int remote_hostio_pread (int fd, gdb_byte *read_buf, int len, ULONGEST offset, fileio_error *remote_errno); int remote_hostio_pwrite (int fd, const gdb_byte *write_buf, int len, ULONGEST offset, fileio_error *remote_errno); int remote_hostio_pread_vFile (int fd, gdb_byte *read_buf, int len, ULONGEST offset, fileio_error *remote_errno); int remote_hostio_send_command (int command_bytes, int which_packet, fileio_error *remote_errno, const char **attachment, int *attachment_len); int remote_hostio_set_filesystem (struct inferior *inf, fileio_error *remote_errno); /* We should get rid of this and use fileio_open directly. */ int remote_hostio_open (struct inferior *inf, const char *filename, int flags, int mode, int warn_if_slow, fileio_error *remote_errno); int remote_hostio_close (int fd, fileio_error *remote_errno); int remote_hostio_unlink (inferior *inf, const char *filename, fileio_error *remote_errno); struct remote_state *get_remote_state (); long get_remote_packet_size (void); long get_memory_packet_size (struct memory_packet_config *config); long get_memory_write_packet_size (); long get_memory_read_packet_size (); char *append_pending_thread_resumptions (char *p, char *endp, ptid_t ptid); static void open_1 (const char *name, int from_tty, int extended_p); void start_remote (int from_tty, int extended_p); void remote_detach_1 (struct inferior *inf, int from_tty); char *append_resumption (char *p, char *endp, ptid_t ptid, int step, gdb_signal siggnal); int remote_resume_with_vcont (ptid_t scope_ptid, int step, gdb_signal siggnal); thread_info *add_current_inferior_and_thread (const char *wait_status); ptid_t wait_ns (ptid_t ptid, struct target_waitstatus *status, target_wait_flags options); ptid_t wait_as (ptid_t ptid, target_waitstatus *status, target_wait_flags options); ptid_t process_stop_reply (struct stop_reply *stop_reply, target_waitstatus *status); ptid_t select_thread_for_ambiguous_stop_reply (const struct target_waitstatus &status); void remote_notice_new_inferior (ptid_t currthread, bool executing); void print_one_stopped_thread (thread_info *thread); void process_initial_stop_replies (int from_tty); thread_info *remote_add_thread (ptid_t ptid, bool running, bool executing, bool silent_p); void btrace_sync_conf (const btrace_config *conf); void remote_btrace_maybe_reopen (); void remove_new_fork_children (threads_listing_context *context); void kill_new_fork_children (inferior *inf); void discard_pending_stop_replies (struct inferior *inf); int stop_reply_queue_length (); void check_pending_events_prevent_wildcard_vcont (bool *may_global_wildcard_vcont); void discard_pending_stop_replies_in_queue (); struct stop_reply *remote_notif_remove_queued_reply (ptid_t ptid); struct stop_reply *queued_stop_reply (ptid_t ptid); int peek_stop_reply (ptid_t ptid); void remote_parse_stop_reply (const char *buf, stop_reply *event); void remote_stop_ns (ptid_t ptid); void remote_interrupt_as (); void remote_interrupt_ns (); char *remote_get_noisy_reply (); int remote_query_attached (int pid); inferior *remote_add_inferior (bool fake_pid_p, int pid, int attached, int try_open_exec); ptid_t remote_current_thread (ptid_t oldpid); ptid_t get_current_thread (const char *wait_status); void set_thread (ptid_t ptid, int gen); void set_general_thread (ptid_t ptid); void set_continue_thread (ptid_t ptid); void set_general_process (); char *write_ptid (char *buf, const char *endbuf, ptid_t ptid); int remote_unpack_thread_info_response (const char *pkt, threadref *expectedref, gdb_ext_thread_info *info); int remote_get_threadinfo (threadref *threadid, int fieldset, gdb_ext_thread_info *info); int parse_threadlist_response (const char *pkt, int result_limit, threadref *original_echo, threadref *resultlist, int *doneflag); int remote_get_threadlist (int startflag, threadref *nextthread, int result_limit, int *done, int *result_count, threadref *threadlist); int remote_threadlist_iterator (rmt_thread_action stepfunction, void *context, int looplimit); int remote_get_threads_with_ql (threads_listing_context *context); int remote_get_threads_with_qxfer (threads_listing_context *context); int remote_get_threads_with_qthreadinfo (threads_listing_context *context); void extended_remote_restart (); void get_offsets (); void remote_check_symbols (); void remote_supported_packet (const struct protocol_feature *feature, enum packet_support support, const char *argument); void remote_query_supported (); void remote_packet_size (const protocol_feature *feature, packet_support support, const char *value); void remote_serial_quit_handler (); void remote_detach_pid (int pid); void remote_vcont_probe (); void remote_resume_with_hc (ptid_t ptid, int step, gdb_signal siggnal); void send_interrupt_sequence (); void interrupt_query (); void remote_notif_get_pending_events (const notif_client *nc); int fetch_register_using_p (struct regcache *regcache, packet_reg *reg); int send_g_packet (); void process_g_packet (struct regcache *regcache); void fetch_registers_using_g (struct regcache *regcache); int store_register_using_P (const struct regcache *regcache, packet_reg *reg); void store_registers_using_G (const struct regcache *regcache); void set_remote_traceframe (); void check_binary_download (CORE_ADDR addr); target_xfer_status remote_write_bytes_aux (const char *header, CORE_ADDR memaddr, const gdb_byte *myaddr, ULONGEST len_units, int unit_size, ULONGEST *xfered_len_units, char packet_format, int use_length); target_xfer_status remote_write_bytes (CORE_ADDR memaddr, const gdb_byte *myaddr, ULONGEST len, int unit_size, ULONGEST *xfered_len); target_xfer_status remote_read_bytes_1 (CORE_ADDR memaddr, gdb_byte *myaddr, ULONGEST len_units, int unit_size, ULONGEST *xfered_len_units); target_xfer_status remote_xfer_live_readonly_partial (gdb_byte *readbuf, ULONGEST memaddr, ULONGEST len, int unit_size, ULONGEST *xfered_len); target_xfer_status remote_read_bytes (CORE_ADDR memaddr, gdb_byte *myaddr, ULONGEST len, int unit_size, ULONGEST *xfered_len); packet_result remote_send_printf (const char *format, ...) ATTRIBUTE_PRINTF (2, 3); target_xfer_status remote_flash_write (ULONGEST address, ULONGEST length, ULONGEST *xfered_len, const gdb_byte *data); int readchar (int timeout); void remote_serial_write (const char *str, int len); int putpkt (const char *buf); int putpkt_binary (const char *buf, int cnt); int putpkt (const gdb::char_vector &buf) { return putpkt (buf.data ()); } void skip_frame (); long read_frame (gdb::char_vector *buf_p); void getpkt (gdb::char_vector *buf, int forever); int getpkt_or_notif_sane_1 (gdb::char_vector *buf, int forever, int expecting_notif, int *is_notif); int getpkt_sane (gdb::char_vector *buf, int forever); int getpkt_or_notif_sane (gdb::char_vector *buf, int forever, int *is_notif); int remote_vkill (int pid); void remote_kill_k (); void extended_remote_disable_randomization (int val); int extended_remote_run (const std::string &args); void send_environment_packet (const char *action, const char *packet, const char *value); void extended_remote_environment_support (); void extended_remote_set_inferior_cwd (); target_xfer_status remote_write_qxfer (const char *object_name, const char *annex, const gdb_byte *writebuf, ULONGEST offset, LONGEST len, ULONGEST *xfered_len, const unsigned int which_packet); target_xfer_status remote_read_qxfer (const char *object_name, const char *annex, gdb_byte *readbuf, ULONGEST offset, LONGEST len, ULONGEST *xfered_len, const unsigned int which_packet); void push_stop_reply (struct stop_reply *new_event); bool vcont_r_supported (); remote_features m_features; private: bool start_remote_1 (int from_tty, int extended_p); /* The remote state. Don't reference this directly. Use the get_remote_state method instead. */ remote_state m_remote_state; }; static const target_info extended_remote_target_info = { "extended-remote", N_("Extended remote target using gdb-specific protocol"), remote_doc }; /* Set up the extended remote target by extending the standard remote target and adding to it. */ class extended_remote_target final : public remote_target { public: const target_info &info () const override { return extended_remote_target_info; } /* Open an extended-remote connection. */ static void open (const char *, int); bool can_create_inferior () override { return true; } void create_inferior (const char *, const std::string &, char **, int) override; void detach (inferior *, int) override; bool can_attach () override { return true; } void attach (const char *, int) override; void post_attach (int) override; bool supports_disable_randomization () override; }; struct stop_reply : public notif_event { ~stop_reply (); /* 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; /* The architecture associated with the expedited registers. */ gdbarch *arch; /* 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). */ std::vector regcache; enum target_stop_reason stop_reason; CORE_ADDR watch_data_address; int core; }; /* Return TARGET as a remote_target if it is one, else nullptr. */ static remote_target * as_remote_target (process_stratum_target *target) { return dynamic_cast (target); } /* See remote.h. */ bool is_remote_target (process_stratum_target *target) { return as_remote_target (target) != nullptr; } /* Per-program-space data key. */ static const registry::key> remote_pspace_data; /* The variable registered as the control variable used by the remote exec-file commands. While the remote exec-file setting is per-program-space, the set/show machinery uses this as the location of the remote exec-file value. */ static std::string remote_exec_file_var; /* 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 int hexnumlen (ULONGEST num); 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 int stub_unpack_int (const char *buff, int fieldlength); static void set_remote_protocol_packet_cmd (const char *args, int from_tty, cmd_list_element *c); static void show_packet_config_cmd (ui_file *file, const unsigned int which_packet, remote_target *remote); static void show_remote_protocol_packet_cmd (struct ui_file *file, int from_tty, struct cmd_list_element *c, const char *value); static ptid_t read_ptid (const char *buf, const char **obuf); static void remote_async_inferior_event_handler (gdb_client_data); static bool remote_read_description_p (struct target_ops *target); static void remote_console_output (const char *msg); static void remote_btrace_reset (remote_state *rs); static void remote_unpush_and_throw (remote_target *target); /* 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; /* Controls whether GDB is willing to use range stepping. */ static bool use_range_stepping = true; /* From the remote target's point of view, each thread is in one of these three states. */ enum class resume_state { /* Not resumed - we haven't been asked to resume this thread. */ NOT_RESUMED, /* We have been asked to resume this thread, but haven't sent a vCont action for it yet. We'll need to consider it next time commit_resume is called. */ RESUMED_PENDING_VCONT, /* We have been asked to resume this thread, and we have sent a vCont action for it. */ RESUMED, }; /* Information about a thread's pending vCont-resume. Used when a thread is in the remote_resume_state::RESUMED_PENDING_VCONT state. remote_target::resume stores this information which is then picked up by remote_target::commit_resume to know which is the proper action for this thread to include in the vCont packet. */ struct resumed_pending_vcont_info { /* True if the last resume call for this thread was a step request, false if a continue request. */ bool step; /* The signal specified in the last resume call for this thread. */ gdb_signal sig; }; /* Private data that we'll store in (struct thread_info)->priv. */ struct remote_thread_info : public private_thread_info { std::string extra; std::string name; int core = -1; /* Thread handle, perhaps a pthread_t or thread_t value, stored as a sequence of bytes. */ gdb::byte_vector thread_handle; /* Whether the target stopped for a breakpoint/watchpoint. */ enum target_stop_reason stop_reason = TARGET_STOPPED_BY_NO_REASON; /* This is set to the data address of the access causing the target to stop for a watchpoint. */ CORE_ADDR watch_data_address = 0; /* Get the thread's resume state. */ enum resume_state get_resume_state () const { return m_resume_state; } /* Put the thread in the NOT_RESUMED state. */ void set_not_resumed () { m_resume_state = resume_state::NOT_RESUMED; } /* Put the thread in the RESUMED_PENDING_VCONT state. */ void set_resumed_pending_vcont (bool step, gdb_signal sig) { m_resume_state = resume_state::RESUMED_PENDING_VCONT; m_resumed_pending_vcont_info.step = step; m_resumed_pending_vcont_info.sig = sig; } /* Get the information this thread's pending vCont-resumption. Must only be called if the thread is in the RESUMED_PENDING_VCONT resume state. */ const struct resumed_pending_vcont_info &resumed_pending_vcont_info () const { gdb_assert (m_resume_state == resume_state::RESUMED_PENDING_VCONT); return m_resumed_pending_vcont_info; } /* Put the thread in the VCONT_RESUMED state. */ void set_resumed () { m_resume_state = resume_state::RESUMED; } private: /* Resume state for this thread. This is used to implement vCont action coalescing (only when the target operates in non-stop mode). remote_target::resume moves the thread to the RESUMED_PENDING_VCONT state, which notes that this thread must be considered in the next commit_resume call. remote_target::commit_resume sends a vCont packet with actions for the threads in the RESUMED_PENDING_VCONT state and moves them to the VCONT_RESUMED state. When reporting a stop to the core for a thread, that thread is moved back to the NOT_RESUMED state. */ enum resume_state m_resume_state = resume_state::NOT_RESUMED; /* Extra info used if the thread is in the RESUMED_PENDING_VCONT state. */ struct resumed_pending_vcont_info m_resumed_pending_vcont_info; }; remote_state::remote_state () : buf (400) { } remote_state::~remote_state () { xfree (this->last_pass_packet); xfree (this->last_program_signals_packet); xfree (this->finished_object); xfree (this->finished_annex); } /* 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. */ char * remote_target::remote_get_noisy_reply () { struct remote_state *rs = get_remote_state (); do /* Loop on reply from remote stub. */ { char *buf; QUIT; /* Allow user to bail out with ^C. */ getpkt (&rs->buf, 0); buf = rs->buf.data (); if (buf[0] == 'E') trace_error (buf); else if (startswith (buf, "qRelocInsn:")) { ULONGEST ul; CORE_ADDR from, to, org_to; const char *p, *pp; int adjusted_size = 0; int relocated = 0; 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 { gdbarch_relocate_instruction (target_gdbarch (), &to, from); relocated = 1; } catch (const gdb_exception &ex) { if (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. */ } 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"); } if (relocated) { adjusted_size = to - org_to; xsnprintf (buf, rs->buf.size (), "qRelocInsn:%x", adjusted_size); putpkt (buf); } } 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); } struct remote_arch_state * remote_state::get_remote_arch_state (struct gdbarch *gdbarch) { remote_arch_state *rsa; auto it = this->m_arch_states.find (gdbarch); if (it == this->m_arch_states.end ()) { auto p = this->m_arch_states.emplace (std::piecewise_construct, std::forward_as_tuple (gdbarch), std::forward_as_tuple (gdbarch)); rsa = &p.first->second; /* Make sure that the packet buffer is plenty big enough for this architecture. */ if (this->buf.size () < rsa->remote_packet_size) this->buf.resize (2 * rsa->remote_packet_size); } else rsa = &it->second; return rsa; } /* Fetch the global remote target state. */ remote_state * remote_target::get_remote_state () { /* 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. */ m_remote_state.get_remote_arch_state (target_gdbarch ()); return &m_remote_state; } /* Fetch the remote exec-file from the current program space. */ static const char * get_remote_exec_file (void) { char *remote_exec_file; remote_exec_file = remote_pspace_data.get (current_program_space); if (remote_exec_file == NULL) return ""; return remote_exec_file; } /* Set the remote exec file for PSPACE. */ static void set_pspace_remote_exec_file (struct program_space *pspace, const char *remote_exec_file) { char *old_file = remote_pspace_data.get (pspace); xfree (old_file); remote_pspace_data.set (pspace, xstrdup (remote_exec_file)); } /* The "set/show remote exec-file" set command hook. */ static void set_remote_exec_file (const char *ignored, int from_tty, struct cmd_list_element *c) { set_pspace_remote_exec_file (current_program_space, remote_exec_file_var.c_str ()); } /* The "set/show remote exec-file" show command hook. */ static void show_remote_exec_file (struct ui_file *file, int from_tty, struct cmd_list_element *cmd, const char *value) { gdb_printf (file, "%s\n", get_remote_exec_file ()); } 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 = ®s[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 = XALLOCAVEC (struct packet_reg *, gdbarch_num_regs (gdbarch)); for (num_remote_regs = 0, regnum = 0; regnum < gdbarch_num_regs (gdbarch); regnum++) if (regs[regnum].pnum != -1) remote_regs[num_remote_regs++] = ®s[regnum]; std::sort (remote_regs, remote_regs + num_remote_regs, [] (const packet_reg *a, const packet_reg *b) { return a->pnum < b->pnum; }); 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) { gdb_assert (regnum < gdbarch_num_regs (gdbarch)); std::vector regs (gdbarch_num_regs (gdbarch)); map_regcache_remote_table (gdbarch, regs.data ()); *pnum = regs[regnum].pnum; *poffset = regs[regnum].offset; return *pnum != -1; } remote_arch_state::remote_arch_state (struct gdbarch *gdbarch) { /* Use the architecture to build a regnum<->pnum table, which will be 1:1 unless a feature set specifies otherwise. */ this->regs.reset (new packet_reg [gdbarch_num_regs (gdbarch)] ()); /* Record the maximum possible size of the g packet - it may turn out to be smaller. */ this->sizeof_g_packet = map_regcache_remote_table (gdbarch, this->regs.get ()); /* 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). */ this->remote_packet_size = 400 - 1; /* This one is filled in when a ``g'' packet is received. */ this->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 (this->sizeof_g_packet > ((this->remote_packet_size - 32) / 2)) this->remote_packet_size = (this->sizeof_g_packet * 2 + 32); } /* Get a pointer to the current remote target. If not connected to a remote target, return NULL. */ static remote_target * get_current_remote_target () { target_ops *proc_target = current_inferior ()->process_target (); return dynamic_cast (proc_target); } /* 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. */ long remote_target::get_remote_packet_size () { struct remote_state *rs = get_remote_state (); remote_arch_state *rsa = rs->get_remote_arch_state (target_gdbarch ()); if (rs->explicit_packet_size) return rs->explicit_packet_size; return rsa->remote_packet_size; } static struct packet_reg * packet_reg_from_regnum (struct gdbarch *gdbarch, struct remote_arch_state *rsa, long regnum) { if (regnum < 0 && regnum >= gdbarch_num_regs (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 gdbarch *gdbarch, struct remote_arch_state *rsa, LONGEST pnum) { int i; for (i = 0; i < gdbarch_num_regs (gdbarch); i++) { struct packet_reg *r = &rsa->regs[i]; if (r->pnum == pnum) return r; } return NULL; } /* 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) gdb_printf (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) gdb_printf (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) gdb_printf (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 (_("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 bool interrupt_on_connect = false; /* 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 bool remote_break; static void set_remotebreak (const 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 principally provided to facilitate backward compatibility. */ static unsigned int remote_address_size; /* The default max memory-write-packet-size, when the setting is "fixed". The 16k is historical. (It came from older GDB's using alloca for buffers and the knowledge (folklore?) that some hosts don't cope very well with large alloca calls.) */ #define DEFAULT_MAX_MEMORY_PACKET_SIZE_FIXED 16384 /* The minimum remote packet size for memory transfers. Ensures we can write at least one byte. */ #define MIN_MEMORY_PACKET_SIZE 20 /* Get the memory packet size, assuming it is fixed. */ static long get_fixed_memory_packet_size (struct memory_packet_config *config) { gdb_assert (config->fixed_p); if (config->size <= 0) return DEFAULT_MAX_MEMORY_PACKET_SIZE_FIXED; else return config->size; } /* Compute the current size of a read/write packet. Since this makes use of ``actual_register_packet_size'' the computation is dynamic. */ long remote_target::get_memory_packet_size (struct memory_packet_config *config) { struct remote_state *rs = get_remote_state (); remote_arch_state *rsa = rs->get_remote_arch_state (target_gdbarch ()); long what_they_get; if (config->fixed_p) what_they_get = get_fixed_memory_packet_size (config); 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 < MIN_MEMORY_PACKET_SIZE) what_they_get = MIN_MEMORY_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.resize (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 (const char *args, struct memory_packet_config *config, bool target_connected) { int fixed_p = config->fixed_p; long size = config->size; if (args == NULL) error (_("Argument required (integer, \"fixed\" or \"limit\").")); 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); /* Instead of explicitly capping the size of a packet to or disallowing it, the user is allowed to set the size to something arbitrarily large. */ } /* Extra checks? */ if (fixed_p && !config->fixed_p) { /* So that the query shows the correct value. */ long query_size = (size <= 0 ? DEFAULT_MAX_MEMORY_PACKET_SIZE_FIXED : size); if (target_connected && !query (_("The target may not be able to correctly handle a %s\n" "of %ld bytes. Change the packet size? "), config->name, query_size)) error (_("Packet size not changed.")); else if (!target_connected && !query (_("Future remote targets may not be able to " "correctly handle a %s\nof %ld bytes. Change the " "packet size for future remote targets? "), config->name, query_size)) error (_("Packet size not changed.")); } /* Update the config. */ config->fixed_p = fixed_p; config->size = size; const char *target_type = get_target_type_name (target_connected); gdb_printf (_("The %s %s is set to \"%s\".\n"), config->name, target_type, args); } /* Show the memory-read or write-packet size configuration CONFIG of the target REMOTE. If REMOTE is nullptr, the default configuration for future remote targets should be passed in CONFIG. */ static void show_memory_packet_size (memory_packet_config *config, remote_target *remote) { const char *target_type = get_target_type_name (remote != nullptr); if (config->size == 0) gdb_printf (_("The %s %s is 0 (default). "), config->name, target_type); else gdb_printf (_("The %s %s is %ld. "), config->name, target_type, config->size); if (config->fixed_p) gdb_printf (_("Packets are fixed at %ld bytes.\n"), get_fixed_memory_packet_size (config)); else { if (remote != nullptr) gdb_printf (_("Packets are limited to %ld bytes.\n"), remote->get_memory_packet_size (config)); else gdb_puts ("The actual limit will be further reduced " "dependent on the target.\n"); } } /* Configure the memory-write-packet size of the currently selected target. If no target is available, the default configuration for future remote targets is configured. */ static void set_memory_write_packet_size (const char *args, int from_tty) { remote_target *remote = get_current_remote_target (); if (remote != nullptr) { set_memory_packet_size (args, &remote->m_features.m_memory_write_packet_config, true); } else { memory_packet_config* config = &memory_write_packet_config; set_memory_packet_size (args, config, false); } } /* Display the memory-write-packet size of the currently selected target. If no target is available, the default configuration for future remote targets is shown. */ static void show_memory_write_packet_size (const char *args, int from_tty) { remote_target *remote = get_current_remote_target (); if (remote != nullptr) show_memory_packet_size (&remote->m_features.m_memory_write_packet_config, remote); else show_memory_packet_size (&memory_write_packet_config, nullptr); } /* Show the number of hardware watchpoints that can be used. */ static void show_hardware_watchpoint_limit (struct ui_file *file, int from_tty, struct cmd_list_element *c, const char *value) { gdb_printf (file, _("The maximum number of target hardware " "watchpoints is %s.\n"), value); } /* Show the length limit (in bytes) for hardware watchpoints. */ static void show_hardware_watchpoint_length_limit (struct ui_file *file, int from_tty, struct cmd_list_element *c, const char *value) { gdb_printf (file, _("The maximum length (in bytes) of a target " "hardware watchpoint is %s.\n"), value); } /* Show the number of hardware breakpoints that can be used. */ static void show_hardware_breakpoint_limit (struct ui_file *file, int from_tty, struct cmd_list_element *c, const char *value) { gdb_printf (file, _("The maximum number of target hardware " "breakpoints is %s.\n"), value); } /* Controls the maximum number of characters to display in the debug output for each remote packet. The remaining characters are omitted. */ static int remote_packet_max_chars = 512; /* Show the maximum number of characters to display for each remote packet when remote debugging is enabled. */ static void show_remote_packet_max_chars (struct ui_file *file, int from_tty, struct cmd_list_element *c, const char *value) { gdb_printf (file, _("Number of remote packet characters to " "display is %s.\n"), value); } long remote_target::get_memory_write_packet_size () { return get_memory_packet_size (&m_features.m_memory_write_packet_config); } /* Configure the memory-read-packet size of the currently selected target. If no target is available, the default configuration for future remote targets is adapted. */ static void set_memory_read_packet_size (const char *args, int from_tty) { remote_target *remote = get_current_remote_target (); if (remote != nullptr) set_memory_packet_size (args, &remote->m_features.m_memory_read_packet_config, true); else { memory_packet_config* config = &memory_read_packet_config; set_memory_packet_size (args, config, false); } } /* Display the memory-read-packet size of the currently selected target. If no target is available, the default configuration for future remote targets is shown. */ static void show_memory_read_packet_size (const char *args, int from_tty) { remote_target *remote = get_current_remote_target (); if (remote != nullptr) show_memory_packet_size (&remote->m_features.m_memory_read_packet_config, remote); else show_memory_packet_size (&memory_read_packet_config, nullptr); } long remote_target::get_memory_read_packet_size () { long size = get_memory_packet_size (&m_features.m_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; } static enum packet_support packet_config_support (const packet_config *config); static void set_remote_protocol_packet_cmd (const char *args, int from_tty, cmd_list_element *c) { remote_target *remote = get_current_remote_target (); gdb_assert (c->var.has_value ()); auto *default_config = static_cast (c->context ()); const int packet_idx = std::distance (remote_protocol_packets, default_config); if (packet_idx >= 0 && packet_idx < PACKET_MAX) { const char *name = packets_descriptions[packet_idx].name; const auto_boolean value = c->var->get (); const char *support = get_packet_support_name (value); const char *target_type = get_target_type_name (remote != nullptr); if (remote != nullptr) remote->m_features.m_protocol_packets[packet_idx].detect = value; else remote_protocol_packets[packet_idx].detect = value; gdb_printf (_("Support for the '%s' packet %s is set to \"%s\".\n"), name, target_type, support); return; } internal_error (_("Could not find config for %s"), c->name); } static void show_packet_config_cmd (ui_file *file, const unsigned int which_packet, remote_target *remote) { const char *support = "internal-error"; const char *target_type = get_target_type_name (remote != nullptr); packet_config *config; if (remote != nullptr) config = &remote->m_features.m_protocol_packets[which_packet]; else config = &remote_protocol_packets[which_packet]; 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: gdb_printf (file, _("Support for the '%s' packet %s is \"auto\", " "currently %s.\n"), packets_descriptions[which_packet].name, target_type, support); break; case AUTO_BOOLEAN_TRUE: case AUTO_BOOLEAN_FALSE: gdb_printf (file, _("Support for the '%s' packet %s is \"%s\".\n"), packets_descriptions[which_packet].name, target_type, get_packet_support_name (config->detect)); break; } } static void add_packet_config_cmd (const unsigned int which_packet, const char *name, const char *title, int legacy) { packets_descriptions[which_packet].name = name; packets_descriptions[which_packet].title = title; packet_config *config = &remote_protocol_packets[which_packet]; gdb::unique_xmalloc_ptr set_doc = xstrprintf ("Set use of remote protocol `%s' (%s) packet.", name, title); gdb::unique_xmalloc_ptr show_doc = xstrprintf ("Show current use of remote protocol `%s' (%s) packet.", name, title); /* set/show TITLE-packet {auto,on,off} */ gdb::unique_xmalloc_ptr cmd_name = xstrprintf ("%s-packet", title); set_show_commands cmds = add_setshow_auto_boolean_cmd (cmd_name.release (), class_obscure, &config->detect, set_doc.get (), show_doc.get (), NULL, /* help_doc */ set_remote_protocol_packet_cmd, show_remote_protocol_packet_cmd, &remote_set_cmdlist, &remote_show_cmdlist); cmds.show->set_context (config); cmds.set->set_context (config); /* set/show remote NAME-packet {auto,on,off} -- legacy. */ if (legacy) { /* It's not clear who should take ownership of the LEGACY_NAME string created below, so, for now, place the string into a static vector which ensures the strings is released when GDB exits. */ static std::vector> legacy_names; gdb::unique_xmalloc_ptr legacy_name = xstrprintf ("%s-packet", name); add_alias_cmd (legacy_name.get (), cmds.set, class_obscure, 0, &remote_set_cmdlist); add_alias_cmd (legacy_name.get (), cmds.show, class_obscure, 0, &remote_show_cmdlist); legacy_names.emplace_back (std::move (legacy_name)); } } 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" - definitely 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_check_result (const gdb::char_vector &buf) { return packet_check_result (buf.data ()); } packet_result remote_features::packet_ok (const char *buf, const int which_packet) { packet_config *config = &m_protocol_packets[which_packet]; packet_description *descr = &packets_descriptions[which_packet]; enum packet_result result; if (config->detect != AUTO_BOOLEAN_TRUE && config->support == PACKET_DISABLE) internal_error (_("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) { remote_debug_printf ("Packet %s (%s) is supported", descr->name, descr->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."), descr->name, descr->title); } else if (config->detect == AUTO_BOOLEAN_TRUE) { /* The user set it wrong. */ error (_("Enabled packet %s (%s) not recognized by stub"), descr->name, descr->title); } remote_debug_printf ("Packet %s (%s) is NOT supported", descr->name, descr->title); config->support = PACKET_DISABLE; break; } return result; } packet_result remote_features::packet_ok (const gdb::char_vector &buf, const int which_packet) { return packet_ok (buf.data (), which_packet); } /* 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 (const 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"); } } packet_support remote_features::packet_support (int packet) const { const packet_config *config = &m_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) { remote_target *remote = get_current_remote_target (); gdb_assert (c->var.has_value ()); auto *default_config = static_cast (c->context ()); const int packet_idx = std::distance (remote_protocol_packets, default_config); if (packet_idx >= 0 && packet_idx < PACKET_MAX) { show_packet_config_cmd (file, packet_idx, remote); return; } internal_error (_("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 (const char *args, int from_tty, struct cmd_list_element *c) { remote_target *remote = get_current_remote_target (); int i; for (i = 0; i < NR_Z_PACKET_TYPES; i++) { if (remote != nullptr) remote->m_features.m_protocol_packets[PACKET_Z0 + i].detect = remote_Z_packet_detect; else remote_protocol_packets[PACKET_Z0 + i].detect = remote_Z_packet_detect; } const char *support = get_packet_support_name (remote_Z_packet_detect); const char *target_type = get_target_type_name (remote != nullptr); gdb_printf (_("Use of Z packets %s is set to \"%s\".\n"), target_type, support); } static void show_remote_protocol_Z_packet_cmd (struct ui_file *file, int from_tty, struct cmd_list_element *c, const char *value) { remote_target *remote = get_current_remote_target (); int i; for (i = 0; i < NR_Z_PACKET_TYPES; i++) show_packet_config_cmd (file, PACKET_Z0 + i, remote); } /* Insert fork catchpoint target routine. If fork events are enabled then return success, nothing more to do. */ int remote_target::insert_fork_catchpoint (int pid) { return !m_features.remote_fork_event_p (); } /* Remove fork catchpoint target routine. Nothing to do, just return success. */ int remote_target::remove_fork_catchpoint (int pid) { return 0; } /* Insert vfork catchpoint target routine. If vfork events are enabled then return success, nothing more to do. */ int remote_target::insert_vfork_catchpoint (int pid) { return !m_features.remote_vfork_event_p (); } /* Remove vfork catchpoint target routine. Nothing to do, just return success. */ int remote_target::remove_vfork_catchpoint (int pid) { return 0; } /* Insert exec catchpoint target routine. If exec events are enabled, just return success. */ int remote_target::insert_exec_catchpoint (int pid) { return !m_features.remote_exec_event_p (); } /* Remove exec catchpoint target routine. Nothing to do, just return success. */ int remote_target::remove_exec_catchpoint (int pid) { return 0; } /* Take advantage of the fact that the TID field is not used, to tag special ptids with it set to != 0. */ static const ptid_t magic_null_ptid (42000, -1, 1); static const ptid_t not_sent_ptid (42000, -2, 1); static const ptid_t any_thread_ptid (42000, 0, 1); /* Find out if the stub attached to PID (and hence GDB should offer to detach instead of killing it when bailing out). */ int remote_target::remote_query_attached (int pid) { struct remote_state *rs = get_remote_state (); size_t size = get_remote_packet_size (); if (m_features.packet_support (PACKET_qAttached) == PACKET_DISABLE) return 0; if (m_features.remote_multi_process_p ()) xsnprintf (rs->buf.data (), size, "qAttached:%x", pid); else xsnprintf (rs->buf.data (), size, "qAttached"); putpkt (rs->buf); getpkt (&rs->buf, 0); switch (m_features.packet_ok (rs->buf, PACKET_qAttached)) { case PACKET_OK: if (strcmp (rs->buf.data (), "1") == 0) return 1; break; case PACKET_ERROR: warning (_("Remote failure reply: %s"), rs->buf.data ()); 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. If TRY_OPEN_EXEC is true then attempt to open this inferior's executable as the main executable if no main executable is open already. */ inferior * remote_target::remote_add_inferior (bool fake_pid_p, int pid, int attached, int try_open_exec) { 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 (); /* However, if the current inferior is already bound to a process, find some other empty inferior. */ if (inf->pid != 0) { inf = nullptr; for (inferior *it : all_inferiors ()) if (it->pid == 0) { inf = it; break; } } if (inf == nullptr) { /* Since all inferiors were already bound to a process, add a new inferior. */ inf = add_inferior_with_spaces (); } switch_to_inferior_no_thread (inf); inf->push_target (this); inferior_appeared (inf, pid); } inf->attach_flag = attached; inf->fake_pid_p = fake_pid_p; /* If no main executable is currently open then attempt to open the file that was executed to create this inferior. */ if (try_open_exec && get_exec_file (0) == NULL) exec_file_locate_attach (pid, 0, 1); /* Check for exec file mismatch, and let the user solve it. */ validate_exec_file (1); return inf; } static remote_thread_info *get_remote_thread_info (thread_info *thread); static remote_thread_info *get_remote_thread_info (remote_target *target, ptid_t ptid); /* Add thread PTID to GDB's thread list. Tag it as executing/running according to EXECUTING and RUNNING respectively. If SILENT_P (or the remote_state::starting_up flag) is true then the new thread is added silently, otherwise the new thread will be announced to the user. */ thread_info * remote_target::remote_add_thread (ptid_t ptid, bool running, bool executing, bool silent_p) { struct remote_state *rs = get_remote_state (); struct thread_info *thread; /* 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 || silent_p) thread = add_thread_silent (this, ptid); else thread = add_thread (this, ptid); /* We start by assuming threads are resumed. That state then gets updated when we process a matching stop reply. */ get_remote_thread_info (thread)->set_resumed (); set_executing (this, ptid, executing); set_running (this, ptid, running); return thread; } /* 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. EXECUTING indicates whether the thread is (internally) executing or stopped. */ void remote_target::remote_notice_new_inferior (ptid_t currthread, bool executing) { /* In non-stop mode, we assume new found threads are (externally) running until proven otherwise with a stop reply. In all-stop, we can only get here if all threads are stopped. */ bool running = target_is_non_stop_p (); /* 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. */ thread_info *tp = this->find_thread (currthread); if (tp != NULL && tp->state == THREAD_EXITED) { /* 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, executing, false); return; } if (!in_thread_list (this, currthread)) { struct inferior *inf = NULL; int pid = currthread.pid (); if (inferior_ptid.is_pid () && pid == inferior_ptid.pid ()) { /* 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 (this, ptid_t (pid))) thread_change_ptid (this, inferior_ptid, currthread); else { thread_info *thr = remote_add_thread (currthread, running, executing, false); switch_to_thread (thr); } return; } if (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 (this, 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 (find_inferior_pid (this, currthread.pid ()) == NULL) { bool fake_pid_p = !m_features.remote_multi_process_p (); inf = remote_add_inferior (fake_pid_p, currthread.pid (), -1, 1); } /* This is really a new thread. Add it. */ thread_info *new_thr = remote_add_thread (currthread, running, executing, false); /* 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 (!rs->starting_up) notice_new_inferior (new_thr, executing, 0); } } } /* Return THREAD's private thread data, creating it if necessary. */ static remote_thread_info * get_remote_thread_info (thread_info *thread) { gdb_assert (thread != NULL); if (thread->priv == NULL) thread->priv.reset (new remote_thread_info); return gdb::checked_static_cast (thread->priv.get ()); } /* Return PTID's private thread data, creating it if necessary. */ static remote_thread_info * get_remote_thread_info (remote_target *target, ptid_t ptid) { thread_info *thr = target->find_thread (ptid); return get_remote_thread_info (thr); } /* 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. */ void remote_target::pass_signals (gdb::array_view pass_signals) { if (m_features.packet_support (PACKET_QPassSignals) != PACKET_DISABLE) { char *pass_packet, *p; int count = 0; struct remote_state *rs = get_remote_state (); gdb_assert (pass_signals.size () < 256); for (size_t i = 0; i < pass_signals.size (); i++) { if (pass_signals[i]) count++; } pass_packet = (char *) xmalloc (count * 3 + strlen ("QPassSignals:") + 1); strcpy (pass_packet, "QPassSignals:"); p = pass_packet + strlen (pass_packet); for (size_t i = 0; i < pass_signals.size (); 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, 0); m_features.packet_ok (rs->buf, PACKET_QPassSignals); xfree (rs->last_pass_packet); rs->last_pass_packet = pass_packet; } else xfree (pass_packet); } } /* If 'QCatchSyscalls' is supported, tell the remote stub to report syscalls to GDB. */ int remote_target::set_syscall_catchpoint (int pid, bool needed, int any_count, gdb::array_view syscall_counts) { const char *catch_packet; enum packet_result result; int n_sysno = 0; if (m_features.packet_support (PACKET_QCatchSyscalls) == PACKET_DISABLE) { /* Not supported. */ return 1; } if (needed && any_count == 0) { /* Count how many syscalls are to be caught. */ for (size_t i = 0; i < syscall_counts.size (); i++) { if (syscall_counts[i] != 0) n_sysno++; } } remote_debug_printf ("pid %d needed %d any_count %d n_sysno %d", pid, needed, any_count, n_sysno); std::string built_packet; if (needed) { /* Prepare a packet with the sysno list, assuming max 8+1 characters for a sysno. If the resulting packet size is too big, fallback on the non-selective packet. */ const int maxpktsz = strlen ("QCatchSyscalls:1") + n_sysno * 9 + 1; built_packet.reserve (maxpktsz); built_packet = "QCatchSyscalls:1"; if (any_count == 0) { /* Add in each syscall to be caught. */ for (size_t i = 0; i < syscall_counts.size (); i++) { if (syscall_counts[i] != 0) string_appendf (built_packet, ";%zx", i); } } if (built_packet.size () > get_remote_packet_size ()) { /* catch_packet too big. Fallback to less efficient non selective mode, with GDB doing the filtering. */ catch_packet = "QCatchSyscalls:1"; } else catch_packet = built_packet.c_str (); } else catch_packet = "QCatchSyscalls:0"; struct remote_state *rs = get_remote_state (); putpkt (catch_packet); getpkt (&rs->buf, 0); result = m_features.packet_ok (rs->buf, PACKET_QCatchSyscalls); if (result == PACKET_OK) return 0; else return -1; } /* If 'QProgramSignals' is supported, tell the remote stub what signals it should pass through to the inferior when detaching. */ void remote_target::program_signals (gdb::array_view signals) { if (m_features.packet_support (PACKET_QProgramSignals) != PACKET_DISABLE) { char *packet, *p; int count = 0; struct remote_state *rs = get_remote_state (); gdb_assert (signals.size () < 256); for (size_t i = 0; i < signals.size (); i++) { if (signals[i]) count++; } packet = (char *) xmalloc (count * 3 + strlen ("QProgramSignals:") + 1); strcpy (packet, "QProgramSignals:"); p = packet + strlen (packet); for (size_t i = 0; i < signals.size (); 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, 0); m_features.packet_ok (rs->buf, 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. */ void remote_target::set_thread (ptid_t ptid, int gen) { struct remote_state *rs = get_remote_state (); ptid_t state = gen ? rs->general_thread : rs->continue_thread; char *buf = rs->buf.data (); char *endbuf = buf + get_remote_packet_size (); if (state == ptid) return; *buf++ = 'H'; *buf++ = gen ? 'g' : 'c'; if (ptid == magic_null_ptid) xsnprintf (buf, endbuf - buf, "0"); else if (ptid == any_thread_ptid) xsnprintf (buf, endbuf - buf, "0"); else if (ptid == minus_one_ptid) xsnprintf (buf, endbuf - buf, "-1"); else write_ptid (buf, endbuf, ptid); putpkt (rs->buf); getpkt (&rs->buf, 0); if (gen) rs->general_thread = ptid; else rs->continue_thread = ptid; } void remote_target::set_general_thread (ptid_t ptid) { set_thread (ptid, 1); } void remote_target::set_continue_thread (ptid_t 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. */ void remote_target::set_general_process () { /* If the remote can't handle multiple processes, don't bother. */ if (!m_features.remote_multi_process_p ()) return; remote_state *rs = get_remote_state (); /* We only need to change the remote current thread if it's pointing at some other process. */ if (rs->general_thread.pid () != inferior_ptid.pid ()) 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 (ptid_t ptid) { if (ptid == magic_null_ptid) /* The main thread is always alive. */ return 1; if (ptid.pid () != 0 && ptid.lwp () == 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. */ bool remote_target::thread_alive (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 (ptid)) return 1; p = rs->buf.data (); endp = p + get_remote_packet_size (); *p++ = 'T'; write_ptid (p, endp, ptid); putpkt (rs->buf); getpkt (&rs->buf, 0); return (rs->buf[0] == 'O' && rs->buf[1] == 'K'); } /* Return a pointer to a thread name if we know it and NULL otherwise. The thread_info object owns the memory for the name. */ const char * remote_target::thread_name (struct thread_info *info) { if (info->priv != NULL) { const std::string &name = get_remote_thread_info (info)->name; return !name.empty () ? name.c_str () : NULL; } return NULL; } /* About these extended threadlist and threadinfo packets. They are variable length packets but, the fields within them are often fixed length. They are redundant 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 const char *unpack_nibble (const char *buf, int *val); static const char *unpack_byte (const char *buf, int *value); static char *pack_int (char *buf, int value); static const char *unpack_int (const char *buf, int *value); static const char *unpack_string (const char *src, char *dest, int length); static char *pack_threadid (char *pkt, threadref *id); static const char *unpack_threadid (const 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 char *pack_threadlist_request (char *pkt, int startflag, int threadcount, threadref *nextthread); 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. */ char * remote_target::write_ptid (char *buf, const char *endbuf, ptid_t ptid) { int pid, tid; if (m_features.remote_multi_process_p ()) { pid = ptid.pid (); if (pid < 0) buf += xsnprintf (buf, endbuf - buf, "p-%x.", -pid); else buf += xsnprintf (buf, endbuf - buf, "p%x.", pid); } tid = ptid.lwp (); 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 one past the last parsed char. Returns null_ptid if no thread id is found, and throws an error if the thread id has an invalid format. */ static ptid_t read_ptid (const char *buf, const char **obuf) { const char *p = buf; const 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_t (pid, tid); } /* No multi-process. Just a tid. */ pp = unpack_varlen_hex (p, &tid); /* Return null_ptid when no thread id is found. */ if (p == pp) { if (obuf) *obuf = pp; return null_ptid; } /* Since the stub is not sending a process id, default to what's current_inferior, unless it doesn't have a PID yet. If so, then since there's no way to know the pid of the reported threads, use the magic number. */ inferior *inf = current_inferior (); if (inf->pid == 0) pid = magic_null_ptid.pid (); else pid = inf->pid; if (obuf) *obuf = pp; return ptid_t (pid, tid); } 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 (const 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 const char * unpack_nibble (const char *buf, int *val) { *val = fromhex (*buf++); return buf; } static const char * unpack_byte (const 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 const char * unpack_int (const 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 const char * unpack_string (const 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 const char * unpack_threadid (const char *inbuf, threadref *id) { char *altref; const 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. */ int remote_target::remote_unpack_thread_info_response (const char *pkt, threadref *expectedref, gdb_ext_thread_info *info) { struct remote_state *rs = get_remote_state (); int mask, length; int tag; threadref ref; const 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 something 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; } int remote_target::remote_get_threadinfo (threadref *threadid, int fieldset, gdb_ext_thread_info *info) { struct remote_state *rs = get_remote_state (); int result; pack_threadinfo_request (rs->buf.data (), fieldset, threadid); putpkt (rs->buf); getpkt (&rs->buf, 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)* */ int remote_target::parse_threadlist_response (const char *pkt, int result_limit, threadref *original_echo, threadref *resultlist, int *doneflag) { struct remote_state *rs = get_remote_state (); int count, resultcount, done; resultcount = 0; /* Assume the 'q' and 'M chars have been stripped. */ const char *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. */ int remote_target::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; /* Truncate 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.data (), startflag, result_limit, nextthread); putpkt (rs->buf); getpkt (&rs->buf, 0); if (rs->buf[0] == '\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. */ /* Possibly, there is a duplicate response. */ /* Possibilities : 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. */ int remote_target::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. */ struct thread_item { explicit thread_item (ptid_t ptid_) : ptid (ptid_) {} thread_item (thread_item &&other) = default; thread_item &operator= (thread_item &&other) = default; DISABLE_COPY_AND_ASSIGN (thread_item); /* The thread's PTID. */ ptid_t ptid; /* The thread's extra info. */ std::string extra; /* The thread's name. */ std::string name; /* The core the thread was running on. -1 if not known. */ int core = -1; /* The thread handle associated with the thread. */ gdb::byte_vector thread_handle; }; /* 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 { /* Return true if this object contains an entry for a thread with ptid PTID. */ bool contains_thread (ptid_t ptid) const { auto match_ptid = [&] (const thread_item &item) { return item.ptid == ptid; }; auto it = std::find_if (this->items.begin (), this->items.end (), match_ptid); return it != this->items.end (); } /* Remove the thread with ptid PTID. */ void remove_thread (ptid_t ptid) { auto match_ptid = [&] (const thread_item &item) { return item.ptid == ptid; }; auto it = std::remove_if (this->items.begin (), this->items.end (), match_ptid); if (it != this->items.end ()) this->items.erase (it); } /* The threads found on the remote target. */ std::vector items; }; static int remote_newthread_step (threadref *ref, void *data) { struct threads_listing_context *context = (struct threads_listing_context *) data; int pid = inferior_ptid.pid (); int lwp = threadref_to_int (ref); ptid_t ptid (pid, lwp); context->items.emplace_back (ptid); return 1; /* continue iterator */ } #define CRAZY_MAX_THREADS 1000 ptid_t remote_target::remote_current_thread (ptid_t oldpid) { struct remote_state *rs = get_remote_state (); putpkt ("qC"); getpkt (&rs->buf, 0); if (rs->buf[0] == 'Q' && rs->buf[1] == 'C') { const char *obuf; ptid_t result; result = read_ptid (&rs->buf[2], &obuf); if (*obuf != '\0') remote_debug_printf ("warning: garbage in qC reply"); return result; } else return oldpid; } /* List remote threads using the deprecated qL packet. */ int remote_target::remote_get_threads_with_ql (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, std::vector &attributes) { struct threads_listing_context *data = (struct threads_listing_context *) user_data; struct gdb_xml_value *attr; char *id = (char *) xml_find_attribute (attributes, "id")->value.get (); ptid_t ptid = read_ptid (id, NULL); data->items.emplace_back (ptid); thread_item &item = data->items.back (); attr = xml_find_attribute (attributes, "core"); if (attr != NULL) item.core = *(ULONGEST *) attr->value.get (); attr = xml_find_attribute (attributes, "name"); if (attr != NULL) item.name = (const char *) attr->value.get (); attr = xml_find_attribute (attributes, "handle"); if (attr != NULL) item.thread_handle = hex2bin ((const char *) attr->value.get ()); } 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 = (struct threads_listing_context *) user_data; if (body_text != NULL && *body_text != '\0') data->items.back ().extra = 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 }, { "name", GDB_XML_AF_OPTIONAL, NULL, NULL }, { "handle", GDB_XML_AF_OPTIONAL, NULL, 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. */ int remote_target::remote_get_threads_with_qxfer (threads_listing_context *context) { #if defined(HAVE_LIBEXPAT) if (m_features.packet_support (PACKET_qXfer_threads) == PACKET_ENABLE) { gdb::optional xml = target_read_stralloc (this, TARGET_OBJECT_THREADS, NULL); if (xml && (*xml)[0] != '\0') { gdb_xml_parse_quick (_("threads"), "threads.dtd", threads_elements, xml->data (), context); } return 1; } #endif return 0; } /* List remote threads using qfThreadInfo/qsThreadInfo. */ int remote_target::remote_get_threads_with_qthreadinfo (threads_listing_context *context) { struct remote_state *rs = get_remote_state (); if (rs->use_threadinfo_query) { const char *bufp; putpkt ("qfThreadInfo"); getpkt (&rs->buf, 0); bufp = rs->buf.data (); if (bufp[0] != '\0') /* q packet recognized */ { while (*bufp++ == 'm') /* reply contains one or more TID */ { do { ptid_t ptid = read_ptid (bufp, &bufp); context->items.emplace_back (ptid); } while (*bufp++ == ','); /* comma-separated list */ putpkt ("qsThreadInfo"); getpkt (&rs->buf, 0); bufp = rs->buf.data (); } return 1; } else { /* Packet not recognized. */ rs->use_threadinfo_query = 0; } } return 0; } /* Return true if INF only has one non-exited thread. */ static bool has_single_non_exited_thread (inferior *inf) { int count = 0; for (thread_info *tp ATTRIBUTE_UNUSED : inf->non_exited_threads ()) if (++count > 1) break; return count == 1; } /* Implement the to_update_thread_list function for the remote targets. */ void remote_target::update_thread_list () { struct threads_listing_context context; int got_list = 0; /* 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 (&context) || remote_get_threads_with_qthreadinfo (&context) || remote_get_threads_with_ql (&context)) { got_list = 1; if (context.items.empty () && remote_thread_always_alive (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. */ return; } /* CONTEXT now holds the current thread list on the remote target end. Delete GDB-side threads no longer found on the target. */ for (thread_info *tp : all_threads_safe ()) { if (tp->inf->process_target () != this) continue; if (!context.contains_thread (tp->ptid)) { /* Do not remove the thread if it is the last thread in the inferior. This situation happens when we have a pending exit process status to process. Otherwise we may end up with a seemingly live inferior (i.e. pid != 0) that has no threads. */ if (has_single_non_exited_thread (tp->inf)) continue; /* Not found. */ delete_thread (tp); } } /* Remove any unreported fork child threads from CONTEXT so that we don't interfere with follow fork, which is where creation of such threads is handled. */ remove_new_fork_children (&context); /* And now add threads we don't know about yet to our list. */ for (thread_item &item : context.items) { if (item.ptid != null_ptid) { /* In non-stop mode, we assume new found threads are executing until proven otherwise with a stop reply. In all-stop, we can only get here if all threads are stopped. */ bool executing = target_is_non_stop_p (); remote_notice_new_inferior (item.ptid, executing); thread_info *tp = this->find_thread (item.ptid); remote_thread_info *info = get_remote_thread_info (tp); info->core = item.core; info->extra = std::move (item.extra); info->name = std::move (item.name); info->thread_handle = std::move (item.thread_handle); } } } 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 (); } } /* * 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. */ const char * remote_target::extra_thread_info (thread_info *tp) { struct remote_state *rs = get_remote_state (); int set; threadref id; struct gdb_ext_thread_info threadinfo; if (rs->remote_desc == 0) /* paranoia */ internal_error (_("remote_threads_extra_info")); if (tp->ptid == magic_null_ptid || (tp->ptid.pid () != 0 && tp->ptid.lwp () == 0)) /* This is the main thread which was added by GDB. The remote server doesn't know about it. */ return NULL; std::string &extra = get_remote_thread_info (tp)->extra; /* If already have cached info, use it. */ if (!extra.empty ()) return extra.c_str (); if (m_features.packet_support (PACKET_qXfer_threads) == PACKET_ENABLE) { /* If we're using qXfer:threads:read, then the extra info is included in the XML. So if we didn't have anything cached, it's because there's really no extra info. */ return NULL; } if (rs->use_threadextra_query) { char *b = rs->buf.data (); char *endb = b + get_remote_packet_size (); xsnprintf (b, endb - b, "qThreadExtraInfo,"); b += strlen (b); write_ptid (b, endb, tp->ptid); putpkt (rs->buf); getpkt (&rs->buf, 0); if (rs->buf[0] != 0) { extra.resize (strlen (rs->buf.data ()) / 2); hex2bin (rs->buf.data (), (gdb_byte *) &extra[0], extra.size ()); return extra.c_str (); } } /* 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, tp->ptid.lwp ()); if (remote_get_threadinfo (&id, set, &threadinfo)) if (threadinfo.active) { if (*threadinfo.shortname) string_appendf (extra, " Name: %s", threadinfo.shortname); if (*threadinfo.display) { if (!extra.empty ()) extra += ','; string_appendf (extra, " State: %s", threadinfo.display); } if (*threadinfo.more_display) { if (!extra.empty ()) extra += ','; string_appendf (extra, " Priority: %s", threadinfo.more_display); } return extra.c_str (); } return NULL; } bool remote_target::static_tracepoint_marker_at (CORE_ADDR addr, struct static_tracepoint_marker *marker) { struct remote_state *rs = get_remote_state (); char *p = rs->buf.data (); xsnprintf (p, get_remote_packet_size (), "qTSTMat:"); p += strlen (p); p += hexnumstr (p, addr); putpkt (rs->buf); getpkt (&rs->buf, 0); p = rs->buf.data (); if (*p == 'E') error (_("Remote failure reply: %s"), p); if (*p++ == 'm') { parse_static_tracepoint_marker_definition (p, NULL, marker); return true; } return false; } std::vector remote_target::static_tracepoint_markers_by_strid (const char *strid) { struct remote_state *rs = get_remote_state (); std::vector markers; const char *p; static_tracepoint_marker marker; /* Ask for a first packet of static tracepoint marker definition. */ putpkt ("qTfSTM"); getpkt (&rs->buf, 0); p = rs->buf.data (); if (*p == 'E') error (_("Remote failure reply: %s"), p); while (*p++ == 'm') { do { parse_static_tracepoint_marker_definition (p, &p, &marker); if (strid == NULL || marker.str_id == strid) markers.push_back (std::move (marker)); } while (*p++ == ','); /* comma-separated list */ /* Ask for another packet of static tracepoint definition. */ putpkt ("qTsSTM"); getpkt (&rs->buf, 0); p = rs->buf.data (); } return markers; } /* Implement the to_get_ada_task_ptid function for the remote targets. */ ptid_t remote_target::get_ada_task_ptid (long lwp, ULONGEST thread) { return ptid_t (inferior_ptid.pid (), lwp); } /* Restart the remote side; this is an extended protocol operation. */ void remote_target::extended_remote_restart () { 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.data (), get_remote_packet_size (), "R%x", 0); putpkt (rs->buf); remote_fileio_reset (); } /* Clean up connection to a remote debugger. */ void remote_target::close () { /* Make sure we leave stdin registered in the event loop. */ terminal_ours (); trace_reset_local_state (); delete this; } remote_target::~remote_target () { struct remote_state *rs = get_remote_state (); /* Check for NULL because we may get here with a partially constructed target/connection. */ if (rs->remote_desc == nullptr) return; serial_close (rs->remote_desc); /* We are destroying the remote target, so we should discard everything of this target. */ discard_pending_stop_replies_in_queue (); if (rs->remote_async_inferior_event_token) delete_async_event_handler (&rs->remote_async_inferior_event_token); delete rs->notif_state; } /* Query the remote side for the text, data and bss offsets. */ void remote_target::get_offsets () { 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]; if (current_program_space->symfile_object_file == NULL) return; putpkt ("qOffsets"); getpkt (&rs->buf, 0); buf = rs->buf.data (); 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 (startswith (ptr, "Text=")) { ptr += 5; /* Don't use strtol, could lose on big values. */ while (*ptr && *ptr != ';') text_addr = (text_addr << 4) + fromhex (*ptr++); if (startswith (ptr, ";Data=")) { ptr += 6; while (*ptr && *ptr != ';') data_addr = (data_addr << 4) + fromhex (*ptr++); } else lose = 1; if (!lose && startswith (ptr, ";Bss=")) { 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 (startswith (ptr, "TextSeg=")) { 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 (startswith (ptr, ";DataSeg=")) { 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); objfile *objf = current_program_space->symfile_object_file; section_offsets offs = objf->section_offsets; symfile_segment_data_up data = get_symfile_segment_data (objf->obfd.get ()); 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 != nullptr && data->segments.size () == 2) { segments[0] = data->segments[0].base + text_addr; segments[1] = data->segments[1].base + 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->segments.size () == 1) { segments[0] = data->segments[0].base + 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 (objf->obfd.get (), data.get (), 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 (do_sections) { offs[SECT_OFF_TEXT (objf)] = 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[SECT_OFF_DATA (objf)] = data_addr; offs[SECT_OFF_BSS (objf)] = data_addr; } objfile_relocate (objf, offs); } /* Send interrupt_sequence to remote target. */ void remote_target::send_interrupt_sequence () { 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 (_("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 (const char *stop_reply) { if (stop_reply[0] == 'T' && strlen (stop_reply) > 3) { const char *p; /* Txx r:val ; r:val (...) */ p = &stop_reply[3]; /* Look for "register" named "thread". */ while (*p != '\0') { const 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. */ ptid_t remote_target::get_current_thread (const char *wait_status) { ptid_t ptid = null_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 == 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. The function returns pointer to the main thread of the inferior. */ thread_info * remote_target::add_current_inferior_and_thread (const char *wait_status) { bool fake_pid_p = false; switch_to_no_thread (); /* Now, if we have thread information, update the current thread's ptid. */ ptid_t curr_ptid = get_current_thread (wait_status); if (curr_ptid != null_ptid) { if (!m_features.remote_multi_process_p ()) fake_pid_p = true; } 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. */ curr_ptid = magic_null_ptid; fake_pid_p = true; } remote_add_inferior (fake_pid_p, curr_ptid.pid (), -1, 1); /* Add the main thread and switch to it. Don't try reading registers yet, since we haven't fetched the target description yet. */ thread_info *tp = add_thread_silent (this, curr_ptid); switch_to_thread_no_regs (tp); return tp; } /* Print info about a thread that was found already stopped on connection. */ void remote_target::print_one_stopped_thread (thread_info *thread) { target_waitstatus ws; /* If there is a pending waitstatus, use it. If there isn't it's because the thread's stop was reported with TARGET_WAITKIND_STOPPED / GDB_SIGNAL_0 and process_initial_stop_replies decided it wasn't interesting to save and report to the core. */ if (thread->has_pending_waitstatus ()) { ws = thread->pending_waitstatus (); thread->clear_pending_waitstatus (); } else { ws.set_stopped (GDB_SIGNAL_0); } switch_to_thread (thread); thread->set_stop_pc (get_frame_pc (get_current_frame ())); set_current_sal_from_frame (get_current_frame ()); /* For "info program". */ set_last_target_status (this, thread->ptid, ws); if (ws.kind () == TARGET_WAITKIND_STOPPED) { enum gdb_signal sig = ws.sig (); if (signal_print_state (sig)) gdb::observers::signal_received.notify (sig); } gdb::observers::normal_stop.notify (NULL, 1); } /* Process all initial stop replies the remote side sent in response to the ? packet. These indicate threads that were already stopped on initial connection. We mark these threads as stopped and print their current frame before giving the user the prompt. */ void remote_target::process_initial_stop_replies (int from_tty) { int pending_stop_replies = stop_reply_queue_length (); struct thread_info *selected = NULL; struct thread_info *lowest_stopped = NULL; struct thread_info *first = NULL; /* This is only used when the target is non-stop. */ gdb_assert (target_is_non_stop_p ()); /* Consume the initial pending events. */ while (pending_stop_replies-- > 0) { ptid_t waiton_ptid = minus_one_ptid; ptid_t event_ptid; struct target_waitstatus ws; int ignore_event = 0; event_ptid = target_wait (waiton_ptid, &ws, TARGET_WNOHANG); if (remote_debug) print_target_wait_results (waiton_ptid, event_ptid, ws); switch (ws.kind ()) { case TARGET_WAITKIND_IGNORE: case TARGET_WAITKIND_NO_RESUMED: case TARGET_WAITKIND_SIGNALLED: case TARGET_WAITKIND_EXITED: /* We shouldn't see these, but if we do, just ignore. */ remote_debug_printf ("event ignored"); ignore_event = 1; break; default: break; } if (ignore_event) continue; thread_info *evthread = this->find_thread (event_ptid); if (ws.kind () == TARGET_WAITKIND_STOPPED) { enum gdb_signal sig = ws.sig (); /* Stubs traditionally report SIGTRAP as initial signal, instead of signal 0. Suppress it. */ if (sig == GDB_SIGNAL_TRAP) sig = GDB_SIGNAL_0; evthread->set_stop_signal (sig); ws.set_stopped (sig); } if (ws.kind () != TARGET_WAITKIND_STOPPED || ws.sig () != GDB_SIGNAL_0) evthread->set_pending_waitstatus (ws); set_executing (this, event_ptid, false); set_running (this, event_ptid, false); get_remote_thread_info (evthread)->set_not_resumed (); } /* "Notice" the new inferiors before anything related to registers/memory. */ for (inferior *inf : all_non_exited_inferiors (this)) { inf->needs_setup = true; if (non_stop) { thread_info *thread = any_live_thread_of_inferior (inf); notice_new_inferior (thread, thread->state == THREAD_RUNNING, from_tty); } } /* If all-stop on top of non-stop, pause all threads. Note this records the threads' stop pc, so must be done after "noticing" the inferiors. */ if (!non_stop) { { /* At this point, the remote target is not async. It needs to be for the poll in stop_all_threads to consider events from it, so enable it temporarily. */ gdb_assert (!this->is_async_p ()); SCOPE_EXIT { target_async (false); }; target_async (true); stop_all_threads ("remote connect in all-stop"); } /* If all threads of an inferior were already stopped, we haven't setup the inferior yet. */ for (inferior *inf : all_non_exited_inferiors (this)) { if (inf->needs_setup) { thread_info *thread = any_live_thread_of_inferior (inf); switch_to_thread_no_regs (thread); setup_inferior (0); } } } /* Now go over all threads that are stopped, and print their current frame. If all-stop, then if there's a signalled thread, pick that as current. */ for (thread_info *thread : all_non_exited_threads (this)) { if (first == NULL) first = thread; if (!non_stop) thread->set_running (false); else if (thread->state != THREAD_STOPPED) continue; if (selected == nullptr && thread->has_pending_waitstatus ()) selected = thread; if (lowest_stopped == NULL || thread->inf->num < lowest_stopped->inf->num || thread->per_inf_num < lowest_stopped->per_inf_num) lowest_stopped = thread; if (non_stop) print_one_stopped_thread (thread); } /* In all-stop, we only print the status of one thread, and leave others with their status pending. */ if (!non_stop) { thread_info *thread = selected; if (thread == NULL) thread = lowest_stopped; if (thread == NULL) thread = first; print_one_stopped_thread (thread); } } /* Mark a remote_target as starting (by setting the starting_up flag within its remote_state) for the lifetime of this object. The reference count on the remote target is temporarily incremented, to prevent the target being deleted under our feet. */ struct scoped_mark_target_starting { /* Constructor, TARGET is the target to be marked as starting, its reference count will be incremented. */ scoped_mark_target_starting (remote_target *target) : m_remote_target (remote_target_ref::new_reference (target)), m_restore_starting_up (set_starting_up_flag (target)) { /* Nothing. */ } private: /* Helper function, set the starting_up flag on TARGET and return an object which, when it goes out of scope, will restore the previous value of the starting_up flag. */ static scoped_restore_tmpl set_starting_up_flag (remote_target *target) { remote_state *rs = target->get_remote_state (); gdb_assert (!rs->starting_up); return make_scoped_restore (&rs->starting_up, true); } /* A gdb::ref_ptr pointer to a remote_target. */ using remote_target_ref = gdb::ref_ptr; /* A reference to the target on which we are operating. */ remote_target_ref m_remote_target; /* An object which restores the previous value of the starting_up flag when it goes out of scope. */ scoped_restore_tmpl m_restore_starting_up; }; /* Helper for remote_target::start_remote, start the remote connection and sync state. Return true if everything goes OK, otherwise, return false. This function exists so that the scoped_restore created within it will expire before we return to remote_target::start_remote. */ bool remote_target::start_remote_1 (int from_tty, int extended_p) { REMOTE_SCOPED_DEBUG_ENTER_EXIT; struct remote_state *rs = get_remote_state (); /* Signal other parts that we're going through the initial setup, and so things may not be stable yet. E.g., we don't try to install tracepoints until we've relocated symbols. Also, a Ctrl-C before we're connected and synced up can't interrupt the target. Instead, it offers to drop the (potentially wedged) connection. */ scoped_mark_target_starting target_is_starting (this); QUIT; if (interrupt_on_connect) send_interrupt_sequence (); /* Ack any packet which the remote side has already sent. */ remote_serial_write ("+", 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 (); /* Check vCont support and set the remote state's vCont_action_support attribute. */ remote_vcont_probe (); /* If the stub wants to get a QAllow, compose one and send it. */ if (m_features.packet_support (PACKET_QAllow) != PACKET_DISABLE) set_permissions (); /* gdbserver < 7.7 (before its fix from 2013-12-11) did reply to any unknown 'v' packet with string "OK". "OK" gets interpreted by GDB as a reply to known packet. For packet "vFile:setfs:" it is an invalid reply and GDB would return error in remote_hostio_set_filesystem, making remote files access impossible. Disable "vFile:setfs:" in such case. Do not disable other 'v' packets as other "vFile" packets get correctly detected even on gdbserver < 7.7. */ { const char v_mustreplyempty[] = "vMustReplyEmpty"; putpkt (v_mustreplyempty); getpkt (&rs->buf, 0); if (strcmp (rs->buf.data (), "OK") == 0) { m_features.m_protocol_packets[PACKET_vFile_setfs].support = PACKET_DISABLE; } else if (strcmp (rs->buf.data (), "") != 0) error (_("Remote replied unexpectedly to '%s': %s"), v_mustreplyempty, rs->buf.data ()); } /* 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. */ if (m_features.packet_support (PACKET_QStartNoAckMode) != PACKET_DISABLE) { putpkt ("QStartNoAckMode"); getpkt (&rs->buf, 0); if (m_features.packet_ok (rs->buf, PACKET_QStartNoAckMode) == PACKET_OK) rs->noack_mode = 1; } if (extended_p) { /* Tell the remote that we are using the extended protocol. */ putpkt ("!"); getpkt (&rs->buf, 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, auto_solib_add); if (target_is_non_stop_p ()) { if (m_features.packet_support (PACKET_QNonStop) != PACKET_ENABLE) error (_("Non-stop mode requested, but remote " "does not support non-stop")); putpkt ("QNonStop:1"); getpkt (&rs->buf, 0); if (strcmp (rs->buf.data (), "OK") != 0) error (_("Remote refused setting non-stop mode with: %s"), rs->buf.data ()); /* 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. */ this->update_thread_list (); } else if (m_features.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, 0); if (strcmp (rs->buf.data (), "OK") != 0) error (_("Remote refused setting all-stop mode with: %s"), rs->buf.data ()); } /* 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 (get_trace_status (current_trace_status ()) != -1) { struct uploaded_tsv *uploaded_tsvs = NULL; upload_trace_state_variables (&uploaded_tsvs); merge_uploaded_trace_state_variables (&uploaded_tsvs); } /* Check whether the target is running now. */ putpkt ("?"); getpkt (&rs->buf, 0); if (!target_is_non_stop_p ()) { char *wait_status = NULL; if (rs->buf[0] == 'W' || rs->buf[0] == 'X') { if (!extended_p) error (_("The target is not running (try extended-remote?)")); return false; } else { /* Save the reply for later. */ wait_status = (char *) alloca (strlen (rs->buf.data ()) + 1); strcpy (wait_status, rs->buf.data ()); } /* 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 (this) == 0) { /* Target has no concept of threads at all. GDB treats non-threaded target as single-threaded; add a main thread. */ thread_info *tp = add_current_inferior_and_thread (wait_status); get_remote_thread_info (tp)->set_resumed (); } 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. */ ptid_t curr_thread = get_current_thread (wait_status); if (curr_thread == 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. */ remote_debug_printf ("warning: couldn't determine remote " "current thread; picking first in list."); for (thread_info *tp : all_non_exited_threads (this, minus_one_ptid)) { switch_to_thread (tp); break; } } else switch_to_thread (this->find_thread (curr_thread)); } /* 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 (this) && gdbarch_target_desc (target_gdbarch ()) == NULL) { target_clear_description (); target_find_description (); } /* Use the previously fetched status. */ gdb_assert (wait_status != NULL); struct notif_event *reply = remote_notif_parse (this, ¬if_client_stop, wait_status); push_stop_reply ((struct stop_reply *) reply); ::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.data (), "OK") != 0) { const notif_client *notif = ¬if_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 (this, notif, rs->buf.data ()); remote_notif_get_pending_events (notif); } if (thread_count (this) == 0) { if (!extended_p) error (_("The target is not running (try extended-remote?)")); return false; } /* Report all signals during attach/startup. */ pass_signals ({}); /* If there are already stopped threads, mark them stopped and report their stops before giving the prompt to the user. */ process_initial_stop_replies (from_tty); if (target_can_async_p ()) target_async (true); } /* Give the target a chance to look up symbols. */ for (inferior *inf : all_inferiors (this)) { /* The inferiors that exist at this point were created from what was found already running on the remote side, so we know they have execution. */ gdb_assert (this->has_execution (inf)); /* No use without a symbol-file. */ if (inf->pspace->symfile_object_file == nullptr) continue; /* Need to switch to a specific thread, because remote_check_symbols uses INFERIOR_PTID to set the general thread. */ scoped_restore_current_thread restore_thread; thread_info *thread = any_thread_of_inferior (inf); switch_to_thread (thread); this->remote_check_symbols (); } /* Possibly the target has been engaged in a trace run started previously; find out where things are at. */ if (get_trace_status (current_trace_status ()) != -1) { struct uploaded_tp *uploaded_tps = NULL; if (current_trace_status ()->running) gdb_printf (_("Trace is already running on the target.\n")); upload_tracepoints (&uploaded_tps); merge_uploaded_tracepoints (&uploaded_tps); } /* Possibly the target has been engaged in a btrace record started previously; find out where things are at. */ remote_btrace_maybe_reopen (); return true; } /* Start the remote connection and sync state. */ void remote_target::start_remote (int from_tty, int extended_p) { if (start_remote_1 (from_tty, extended_p) && breakpoints_should_be_inserted_now ()) insert_breakpoints (); } const char * remote_target::connection_string () { remote_state *rs = get_remote_state (); if (rs->remote_desc->name != NULL) return rs->remote_desc->name; else return NULL; } /* Open a connection to a remote debugger. NAME is the filename used for communication. */ void remote_target::open (const char *name, int from_tty) { open_1 (name, from_tty, 0); } /* Open a connection to a remote debugger using the extended remote gdb protocol. NAME is the filename used for communication. */ void extended_remote_target::open (const char *name, int from_tty) { open_1 (name, from_tty, 1 /*extended_p */); } void remote_features::reset_all_packet_configs_support () { int i; for (i = 0; i < PACKET_MAX; i++) m_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. */ void remote_target::remote_check_symbols () { char *tmp; int end; /* It doesn't make sense to send a qSymbol packet for an inferior that doesn't have execution, because the remote side doesn't know about inferiors without execution. */ gdb_assert (target_has_execution ()); if (m_features.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. */ gdb::char_vector msg (get_remote_packet_size ()); gdb::char_vector reply (get_remote_packet_size ()); /* Invite target to request symbol lookups. */ putpkt ("qSymbol::"); getpkt (&reply, 0); m_features.packet_ok (reply, PACKET_qSymbol); while (startswith (reply.data (), "qSymbol:")) { struct bound_minimal_symbol sym; tmp = &reply[8]; end = hex2bin (tmp, reinterpret_cast (msg.data ()), strlen (tmp) / 2); msg[end] = '\0'; sym = lookup_minimal_symbol (msg.data (), NULL, NULL); if (sym.minsym == NULL) xsnprintf (msg.data (), get_remote_packet_size (), "qSymbol::%s", &reply[8]); else { int addr_size = gdbarch_addr_bit (target_gdbarch ()) / 8; CORE_ADDR sym_addr = sym.value_address (); /* 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_inferior ()->top_target ()); xsnprintf (msg.data (), get_remote_packet_size (), "qSymbol:%s:%s", phex_nz (sym_addr, addr_size), &reply[8]); } putpkt (msg.data ()); getpkt (&reply, 0); } } 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 && startswith (name, "udp:")) { 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_target::set_permissions () { struct remote_state *rs = get_remote_state (); xsnprintf (rs->buf.data (), 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, 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.data (), "OK") != 0) warning (_("Remote refused setting permissions with: %s"), rs->buf.data ()); } /* 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) (remote_target *remote, 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 (remote_target *remote, 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->m_features.m_protocol_packets[feature->packet].support = support; } void remote_target::remote_packet_size (const 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; } /* Record the new maximum packet size. */ rs->explicit_packet_size = packet_size; } static void remote_packet_size (remote_target *remote, const protocol_feature *feature, enum packet_support support, const char *value) { remote->remote_packet_size (feature, support, value); } 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:exec-file:read", PACKET_DISABLE, remote_supported_packet, PACKET_qXfer_exec_file }, { "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: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 }, { "QCatchSyscalls", PACKET_DISABLE, remote_supported_packet, PACKET_QCatchSyscalls }, { "QProgramSignals", PACKET_DISABLE, remote_supported_packet, PACKET_QProgramSignals }, { "QSetWorkingDir", PACKET_DISABLE, remote_supported_packet, PACKET_QSetWorkingDir }, { "QStartupWithShell", PACKET_DISABLE, remote_supported_packet, PACKET_QStartupWithShell }, { "QEnvironmentHexEncoded", PACKET_DISABLE, remote_supported_packet, PACKET_QEnvironmentHexEncoded }, { "QEnvironmentReset", PACKET_DISABLE, remote_supported_packet, PACKET_QEnvironmentReset }, { "QEnvironmentUnset", PACKET_DISABLE, remote_supported_packet, PACKET_QEnvironmentUnset }, { "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 }, { "Qbtrace:pt", PACKET_DISABLE, remote_supported_packet, PACKET_Qbtrace_pt }, { "qXfer:btrace:read", PACKET_DISABLE, remote_supported_packet, PACKET_qXfer_btrace }, { "qXfer:btrace-conf:read", PACKET_DISABLE, remote_supported_packet, PACKET_qXfer_btrace_conf }, { "Qbtrace-conf:bts:size", PACKET_DISABLE, remote_supported_packet, PACKET_Qbtrace_conf_bts_size }, { "swbreak", PACKET_DISABLE, remote_supported_packet, PACKET_swbreak_feature }, { "hwbreak", PACKET_DISABLE, remote_supported_packet, PACKET_hwbreak_feature }, { "fork-events", PACKET_DISABLE, remote_supported_packet, PACKET_fork_event_feature }, { "vfork-events", PACKET_DISABLE, remote_supported_packet, PACKET_vfork_event_feature }, { "exec-events", PACKET_DISABLE, remote_supported_packet, PACKET_exec_event_feature }, { "Qbtrace-conf:pt:size", PACKET_DISABLE, remote_supported_packet, PACKET_Qbtrace_conf_pt_size }, { "vContSupported", PACKET_DISABLE, remote_supported_packet, PACKET_vContSupported }, { "QThreadEvents", PACKET_DISABLE, remote_supported_packet, PACKET_QThreadEvents }, { "no-resumed", PACKET_DISABLE, remote_supported_packet, PACKET_no_resumed }, { "memory-tagging", PACKET_DISABLE, remote_supported_packet, PACKET_memory_tagging_feature }, }; 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 *saveptr; char *p = strtok_r (copy, ",", &saveptr); do { if (strcmp (p, xml) == 0) { /* already there */ xfree (copy); return; } } while ((p = strtok_r (NULL, ",", &saveptr)) != NULL); xfree (copy); remote_support_xml = reconcat (remote_support_xml, remote_support_xml, ",", xml, (char *) NULL); } #endif } static void remote_query_supported_append (std::string *msg, const char *append) { if (!msg->empty ()) msg->append (";"); msg->append (append); } void remote_target::remote_query_supported () { 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 (m_features.packet_support (PACKET_qSupported) != PACKET_DISABLE) { std::string q; if (m_features.packet_set_cmd_state (PACKET_multiprocess_feature) != AUTO_BOOLEAN_FALSE) remote_query_supported_append (&q, "multiprocess+"); if (m_features.packet_set_cmd_state (PACKET_swbreak_feature) != AUTO_BOOLEAN_FALSE) remote_query_supported_append (&q, "swbreak+"); if (m_features.packet_set_cmd_state (PACKET_hwbreak_feature) != AUTO_BOOLEAN_FALSE) remote_query_supported_append (&q, "hwbreak+"); remote_query_supported_append (&q, "qRelocInsn+"); if (m_features.packet_set_cmd_state (PACKET_fork_event_feature) != AUTO_BOOLEAN_FALSE) remote_query_supported_append (&q, "fork-events+"); if (m_features.packet_set_cmd_state (PACKET_vfork_event_feature) != AUTO_BOOLEAN_FALSE) remote_query_supported_append (&q, "vfork-events+"); if (m_features.packet_set_cmd_state (PACKET_exec_event_feature) != AUTO_BOOLEAN_FALSE) remote_query_supported_append (&q, "exec-events+"); if (m_features.packet_set_cmd_state (PACKET_vContSupported) != AUTO_BOOLEAN_FALSE) remote_query_supported_append (&q, "vContSupported+"); if (m_features.packet_set_cmd_state (PACKET_QThreadEvents) != AUTO_BOOLEAN_FALSE) remote_query_supported_append (&q, "QThreadEvents+"); if (m_features.packet_set_cmd_state (PACKET_no_resumed) != AUTO_BOOLEAN_FALSE) remote_query_supported_append (&q, "no-resumed+"); if (m_features.packet_set_cmd_state (PACKET_memory_tagging_feature) != AUTO_BOOLEAN_FALSE) remote_query_supported_append (&q, "memory-tagging+"); /* Keep this one last to work around a gdbserver <= 7.10 bug in the qSupported:xmlRegisters=i386 handling. */ if (remote_support_xml != NULL && (m_features.packet_support (PACKET_qXfer_features) != PACKET_DISABLE)) remote_query_supported_append (&q, remote_support_xml); q = "qSupported:" + q; putpkt (q.c_str ()); getpkt (&rs->buf, 0); /* If an error occured, warn, but do not return - just reset the buffer to empty and go on to disable features. */ if (m_features.packet_ok (rs->buf, PACKET_qSupported) == PACKET_ERROR) { warning (_("Remote failure reply: %s"), rs->buf.data ()); rs->buf[0] = 0; } } memset (seen, 0, sizeof (seen)); next = rs->buf.data (); 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 (this, 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.resize (rs->explicit_packet_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 (this, feature, feature->default_support, NULL); } } /* Serial QUIT handler for the remote serial descriptor. Defers handling a Ctrl-C until we're done with the current command/response packet sequence, unless: - We're setting up the connection. Don't send a remote interrupt request, as we're not fully synced yet. Quit immediately instead. - The target has been resumed in the foreground (target_terminal::is_ours is false) with a synchronous resume packet, and we're blocked waiting for the stop reply, thus a Ctrl-C should be immediately sent to the target. - We get a second Ctrl-C while still within the same serial read or write. In that case the serial is seemingly wedged --- offer to quit/disconnect. - We see a second Ctrl-C without target response, after having previously interrupted the target. In that case the target/stub is probably wedged --- offer to quit/disconnect. */ void remote_target::remote_serial_quit_handler () { struct remote_state *rs = get_remote_state (); if (check_quit_flag ()) { /* If we're starting up, we're not fully synced yet. Quit immediately. */ if (rs->starting_up) quit (); else if (rs->got_ctrlc_during_io) { if (query (_("The target is not responding to GDB commands.\n" "Stop debugging it? "))) remote_unpush_and_throw (this); } /* If ^C has already been sent once, offer to disconnect. */ else if (!target_terminal::is_ours () && rs->ctrlc_pending_p) interrupt_query (); /* All-stop protocol, and blocked waiting for stop reply. Send an interrupt request. */ else if (!target_terminal::is_ours () && rs->waiting_for_stop_reply) target_interrupt (); else rs->got_ctrlc_during_io = 1; } } /* The remote_target that is current while the quit handler is overridden with remote_serial_quit_handler. */ static remote_target *curr_quit_handler_target; static void remote_serial_quit_handler () { curr_quit_handler_target->remote_serial_quit_handler (); } /* Remove the remote target from the target stack of each inferior that is using it. Upper targets depend on it so remove them first. */ static void remote_unpush_target (remote_target *target) { /* We have to unpush the target from all inferiors, even those that aren't running. */ scoped_restore_current_inferior restore_current_inferior; for (inferior *inf : all_inferiors (target)) { switch_to_inferior_no_thread (inf); inf->pop_all_targets_at_and_above (process_stratum); generic_mourn_inferior (); } /* Don't rely on target_close doing this when the target is popped from the last remote inferior above, because something may be holding a reference to the target higher up on the stack, meaning target_close won't be called yet. We lost the connection to the target, so clear these now, otherwise we may later throw TARGET_CLOSE_ERROR while trying to tell the remote target to close the file. */ fileio_handles_invalidate_target (target); } static void remote_unpush_and_throw (remote_target *target) { remote_unpush_target (target); throw_error (TARGET_CLOSE_ERROR, _("Disconnected from target.")); } void remote_target::open_1 (const char *name, int from_tty, int extended_p) { remote_target *curr_remote = get_current_remote_target (); 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.).")); /* 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 (curr_remote != NULL && !target_has_execution ()) { 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); remote_fileio_reset (); reopen_exec_file (); reread_symbols (from_tty); remote_target *remote = (extended_p ? new extended_remote_target () : new remote_target ()); target_ops_up target_holder (remote); remote_state *rs = remote->get_remote_state (); /* See FIXME above. */ if (!target_async_permitted) rs->wait_forever_enabled_p = 1; 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_setparity (rs->remote_desc, serial_parity); 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) { gdb_puts ("Remote debugging using "); gdb_puts (name); gdb_puts ("\n"); } /* Switch to using the remote target now. */ current_inferior ()->push_target (std::move (target_holder)); /* Register extra event sources in the event loop. */ rs->remote_async_inferior_event_token = create_async_event_handler (remote_async_inferior_event_handler, nullptr, "remote"); rs->notif_state = remote_notif_state_allocate (remote); /* Reset the target state; these things will be queried either by remote_query_supported or as they are needed. */ remote->m_features.reset_all_packet_configs_support (); 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->got_ctrlc_during_io = 0; rs->general_thread = not_sent_ptid; rs->continue_thread = not_sent_ptid; rs->remote_traceframe_number = -1; rs->last_resume_exec_dir = EXEC_FORWARD; /* Probe for ability to use "ThreadInfo" query, as required. */ rs->use_threadinfo_query = 1; rs->use_threadextra_query = 1; rs->readahead_cache.invalidate (); if (target_async_permitted) { /* 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. */ rs->wait_forever_enabled_p = 0; } /* First delete any symbols previously loaded from shared libraries. */ no_shared_libraries (NULL, 0); /* 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. */ { try { remote->start_remote (from_tty, extended_p); } catch (const gdb_exception &ex) { /* Pop the partially set up target - unless something else did already before throwing the exception. */ if (ex.error != TARGET_CLOSE_ERROR) remote_unpush_target (remote); throw; } } remote_btrace_reset (rs); if (target_async_permitted) rs->wait_forever_enabled_p = 1; } /* Determine if WS represents a fork status. */ static bool is_fork_status (target_waitkind kind) { return (kind == TARGET_WAITKIND_FORKED || kind == TARGET_WAITKIND_VFORKED); } /* Return THREAD's pending status if it is a pending fork parent, else return nullptr. */ static const target_waitstatus * thread_pending_fork_status (struct thread_info *thread) { const target_waitstatus &ws = (thread->has_pending_waitstatus () ? thread->pending_waitstatus () : thread->pending_follow); if (!is_fork_status (ws.kind ())) return nullptr; return &ws; } /* Detach the specified process. */ void remote_target::remote_detach_pid (int pid) { struct remote_state *rs = get_remote_state (); /* This should not be necessary, but the handling for D;PID in GDBserver versions prior to 8.2 incorrectly assumes that the selected process points to the same process we're detaching, leading to misbehavior (and possibly GDBserver crashing) when it does not. Since it's easy and cheap, work around it by forcing GDBserver to select GDB's current process. */ set_general_process (); if (m_features.remote_multi_process_p ()) xsnprintf (rs->buf.data (), get_remote_packet_size (), "D;%x", pid); else strcpy (rs->buf.data (), "D"); putpkt (rs->buf); getpkt (&rs->buf, 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.")); } /* This detaches a program to which we previously attached, using inferior_ptid to identify the process. 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. */ void remote_target::remote_detach_1 (inferior *inf, int from_tty) { int pid = inferior_ptid.pid (); struct remote_state *rs = get_remote_state (); int is_fork_parent; if (!target_has_execution ()) error (_("No process to detach from.")); target_announce_detach (from_tty); if (!gdbarch_has_global_breakpoints (target_gdbarch ())) { /* If we're in breakpoints-always-inserted mode, or the inferior is running, we have to remove breakpoints before detaching. We don't do this in common code instead because not all targets support removing breakpoints while the target is running. The remote target / gdbserver does, though. */ remove_breakpoints_inf (current_inferior ()); } /* Tell the remote target to detach. */ remote_detach_pid (pid); /* Exit only if this is the only active inferior. */ if (from_tty && !rs->extended && number_of_live_inferiors (this) == 1) gdb_puts (_("Ending remote debugging.\n")); /* See if any thread of the inferior we are detaching has a pending fork status. In that case, we must detach from the child resulting from that fork. */ for (thread_info *thread : inf->non_exited_threads ()) { const target_waitstatus *ws = thread_pending_fork_status (thread); if (ws == nullptr) continue; remote_detach_pid (ws->child_ptid ().pid ()); } /* Check also for any pending fork events in the stop reply queue. */ remote_notif_get_pending_events (¬if_client_stop); for (stop_reply_up &reply : rs->stop_reply_queue) { if (reply->ptid.pid () != pid) continue; if (!is_fork_status (reply->ws.kind ())) continue; remote_detach_pid (reply->ws.child_ptid ().pid ()); } thread_info *tp = this->find_thread (inferior_ptid); /* Check to see if we are detaching a fork parent. Note that if we are detaching a fork child, tp == NULL. */ is_fork_parent = (tp != NULL && tp->pending_follow.kind () == TARGET_WAITKIND_FORKED); /* If doing detach-on-fork, we don't mourn, because that will delete breakpoints that should be available for the followed inferior. */ if (!is_fork_parent) { /* Save the pid as a string before mourning, since that will unpush the remote target, and we need the string after. */ std::string infpid = target_pid_to_str (ptid_t (pid)); target_mourn_inferior (inferior_ptid); if (print_inferior_events) gdb_printf (_("[Inferior %d (%s) detached]\n"), inf->num, infpid.c_str ()); } else { switch_to_no_thread (); detach_inferior (current_inferior ()); } } void remote_target::detach (inferior *inf, int from_tty) { remote_detach_1 (inf, from_tty); } void extended_remote_target::detach (inferior *inf, int from_tty) { remote_detach_1 (inf, from_tty); } /* Target follow-fork function for remote targets. On entry, and at return, the current inferior is the fork parent. Note that although this is currently only used for extended-remote, it is named remote_follow_fork in anticipation of using it for the remote target as well. */ void remote_target::follow_fork (inferior *child_inf, ptid_t child_ptid, target_waitkind fork_kind, bool follow_child, bool detach_fork) { process_stratum_target::follow_fork (child_inf, child_ptid, fork_kind, follow_child, detach_fork); if ((fork_kind == TARGET_WAITKIND_FORKED && m_features.remote_fork_event_p ()) || (fork_kind == TARGET_WAITKIND_VFORKED && m_features.remote_vfork_event_p ())) { /* When following the parent and detaching the child, we detach the child here. For the case of following the child and detaching the parent, the detach is done in the target- independent follow fork code in infrun.c. We can't use target_detach when detaching an unfollowed child because the client side doesn't know anything about the child. */ if (detach_fork && !follow_child) { /* Detach the fork child. */ remote_detach_pid (child_ptid.pid ()); } } } /* Target follow-exec function for remote targets. Save EXECD_PATHNAME in the program space of the new inferior. */ void remote_target::follow_exec (inferior *follow_inf, ptid_t ptid, const char *execd_pathname) { process_stratum_target::follow_exec (follow_inf, ptid, execd_pathname); /* We know that this is a target file name, so if it has the "target:" prefix we strip it off before saving it in the program space. */ if (is_target_filename (execd_pathname)) execd_pathname += strlen (TARGET_SYSROOT_PREFIX); set_pspace_remote_exec_file (follow_inf->pspace, execd_pathname); } /* Same as remote_detach, but don't send the "D" packet; just disconnect. */ void remote_target::disconnect (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. Calling target_mourn_inferior won't unpush, and remote_target::mourn_inferior won't unpush if there is more than one inferior left. */ remote_unpush_target (this); if (from_tty) gdb_puts ("Ending remote debugging.\n"); } /* Attach to the process specified by ARGS. If FROM_TTY is non-zero, be chatty about it. */ void extended_remote_target::attach (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 (m_features.packet_support (PACKET_vAttach) == PACKET_DISABLE) error (_("This target does not support attaching to a process")); target_announce_attach (from_tty, pid); xsnprintf (rs->buf.data (), get_remote_packet_size (), "vAttach;%x", pid); putpkt (rs->buf); getpkt (&rs->buf, 0); switch (m_features.packet_ok (rs->buf, PACKET_vAttach)) { case PACKET_OK: if (!target_is_non_stop_p ()) { /* Save the reply for later. */ wait_status = (char *) alloca (strlen (rs->buf.data ()) + 1); strcpy (wait_status, rs->buf.data ()); } else if (strcmp (rs->buf.data (), "OK") != 0) error (_("Attaching to %s failed with: %s"), target_pid_to_str (ptid_t (pid)).c_str (), rs->buf.data ()); break; case PACKET_UNKNOWN: error (_("This target does not support attaching to a process")); default: error (_("Attaching to %s failed"), target_pid_to_str (ptid_t (pid)).c_str ()); } switch_to_inferior_no_thread (remote_add_inferior (false, pid, 1, 0)); inferior_ptid = ptid_t (pid); if (target_is_non_stop_p ()) { /* Get list of threads. */ update_thread_list (); thread_info *thread = first_thread_of_inferior (current_inferior ()); if (thread != nullptr) switch_to_thread (thread); /* Invalidate our notion of the remote current thread. */ record_currthread (rs, minus_one_ptid); } else { /* Now, if we have thread information, update the main thread's ptid. */ ptid_t curr_ptid = remote_current_thread (ptid_t (pid)); /* Add the main thread to the thread list. We add the thread silently in this case (the final true parameter). */ thread_info *thr = remote_add_thread (curr_ptid, true, true, true); switch_to_thread (thr); } /* Next, if the target can specify a description, read it. We do this before anything involving memory or registers. */ target_find_description (); if (!target_is_non_stop_p ()) { /* Use the previously fetched status. */ gdb_assert (wait_status != NULL); struct notif_event *reply = remote_notif_parse (this, ¬if_client_stop, wait_status); push_stop_reply ((struct stop_reply *) reply); } else { gdb_assert (wait_status == NULL); gdb_assert (target_can_async_p ()); } } /* Implementation of the to_post_attach method. */ void extended_remote_target::post_attach (int pid) { /* Get text, data & bss offsets. */ get_offsets (); /* 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 (current_program_space->symfile_object_file != NULL) remote_check_symbols(); } /* Check for the availability of vCont. This function should also check the response. */ void remote_target::remote_vcont_probe () { remote_state *rs = get_remote_state (); char *buf; strcpy (rs->buf.data (), "vCont?"); putpkt (rs->buf); getpkt (&rs->buf, 0); buf = rs->buf.data (); /* Make sure that the features we assume are supported. */ if (startswith (buf, "vCont")) { char *p = &buf[5]; int support_c, support_C; rs->supports_vCont.s = 0; rs->supports_vCont.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)) rs->supports_vCont.s = 1; else if (*p == 'S' && (*(p + 1) == ';' || *(p + 1) == 0)) rs->supports_vCont.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 c, and C are not all supported, we can't use vCont. Clearing BUF will make packet_ok disable the packet. */ if (!support_c || !support_C) buf[0] = 0; } m_features.packet_ok (rs->buf, 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. */ char * remote_target::append_resumption (char *p, char *endp, ptid_t ptid, int step, 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). */ && !(m_features.remote_multi_process_p () && ptid.is_pid ())) { struct thread_info *tp; if (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 = this->find_thread (magic_null_ptid); } else tp = this->find_thread (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 (m_features.remote_multi_process_p () && ptid.is_pid ()) { ptid_t nptid; /* All (-1) threads of process. */ nptid = ptid_t (ptid.pid (), -1); p += xsnprintf (p, endp - p, ":"); p = write_ptid (p, endp, nptid); } else if (ptid != minus_one_ptid) { p += xsnprintf (p, endp - p, ":"); p = write_ptid (p, endp, ptid); } return p; } /* Clear the thread's private info on resume. */ static void resume_clear_thread_private_info (struct thread_info *thread) { if (thread->priv != NULL) { remote_thread_info *priv = get_remote_thread_info (thread); priv->stop_reason = TARGET_STOPPED_BY_NO_REASON; priv->watch_data_address = 0; } } /* Append a vCont continue-with-signal action for threads that have a non-zero stop signal. */ char * remote_target::append_pending_thread_resumptions (char *p, char *endp, ptid_t ptid) { for (thread_info *thread : all_non_exited_threads (this, ptid)) if (inferior_ptid != thread->ptid && thread->stop_signal () != GDB_SIGNAL_0) { p = append_resumption (p, endp, thread->ptid, 0, thread->stop_signal ()); thread->set_stop_signal (GDB_SIGNAL_0); resume_clear_thread_private_info (thread); } return p; } /* Set the target running, using the packets that use Hc (c/s/C/S). */ void remote_target::remote_resume_with_hc (ptid_t ptid, int step, gdb_signal siggnal) { struct remote_state *rs = get_remote_state (); char *buf; rs->last_sent_signal = siggnal; rs->last_sent_step = step; /* The c/s/C/S resume packets use Hc, so set the continue thread. */ if (ptid == minus_one_ptid) set_continue_thread (any_thread_ptid); else set_continue_thread (ptid); for (thread_info *thread : all_non_exited_threads (this)) resume_clear_thread_private_info (thread); buf = rs->buf.data (); 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 && m_features.packet_support (PACKET_bs) == PACKET_DISABLE) error (_("Remote reverse-step not supported.")); if (!step && m_features.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); } /* Resume the remote inferior by using a "vCont" packet. SCOPE_PTID, STEP, and SIGGNAL have the same meaning as in target_resume. This function returns non-zero iff it resumes the inferior. This function issues a strict subset of all possible vCont commands at the moment. */ int remote_target::remote_resume_with_vcont (ptid_t scope_ptid, int step, enum gdb_signal siggnal) { struct remote_state *rs = get_remote_state (); char *p; char *endp; /* No reverse execution actions defined for vCont. */ if (::execution_direction == EXEC_REVERSE) return 0; if (m_features.packet_support (PACKET_vCont) == PACKET_DISABLE) return 0; p = rs->buf.data (); endp = p + 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 (scope_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 (scope_ptid == minus_one_ptid || scope_ptid.is_pid ()) { /* 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, scope_ptid); /* And continue others without a signal. */ append_resumption (p, endp, scope_ptid, /*step=*/ 0, GDB_SIGNAL_0); } else { /* Scheduler locking; resume only SCOPE_PTID. */ append_resumption (p, endp, scope_ptid, step, siggnal); } gdb_assert (strlen (rs->buf.data ()) < get_remote_packet_size ()); putpkt (rs->buf); if (target_is_non_stop_p ()) { /* 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, 0); if (strcmp (rs->buf.data (), "OK") != 0) error (_("Unexpected vCont reply in non-stop mode: %s"), rs->buf.data ()); } return 1; } /* Tell the remote machine to resume. */ void remote_target::resume (ptid_t scope_ptid, int step, enum gdb_signal siggnal) { struct remote_state *rs = get_remote_state (); /* When connected in non-stop mode, the core resumes threads individually. Resuming remote threads directly in target_resume would thus result in sending one packet per thread. Instead, to minimize roundtrip latency, here we just store the resume request (put the thread in RESUMED_PENDING_VCONT state); the actual remote resumption will be done in remote_target::commit_resume, where we'll be able to do vCont action coalescing. */ if (target_is_non_stop_p () && ::execution_direction != EXEC_REVERSE) { remote_thread_info *remote_thr = get_remote_thread_info (inferior_thread ()); /* We don't expect the core to ask to resume an already resumed (from its point of view) thread. */ gdb_assert (remote_thr->get_resume_state () == resume_state::NOT_RESUMED); remote_thr->set_resumed_pending_vcont (step, siggnal); /* There's actually nothing that says that the core can't request a wildcard resume in non-stop mode, though. It's just that we know it doesn't currently, so we don't bother with it. */ gdb_assert (scope_ptid == inferior_ptid); return; } /* 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 (!target_is_non_stop_p ()) remote_notif_process (rs->notif_state, ¬if_client_stop); rs->last_resume_exec_dir = ::execution_direction; /* Prefer vCont, and fallback to s/c/S/C, which use Hc. */ if (!remote_resume_with_vcont (scope_ptid, step, siggnal)) remote_resume_with_hc (scope_ptid, step, siggnal); /* Update resumed state tracked by the remote target. */ for (thread_info *tp : all_non_exited_threads (this, scope_ptid)) get_remote_thread_info (tp)->set_resumed (); /* 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 (!target_is_non_stop_p ()) rs->waiting_for_stop_reply = 1; } /* Private per-inferior info for target remote processes. */ struct remote_inferior : public private_inferior { /* Whether we can send a wildcard vCont for this process. */ bool may_wildcard_vcont = true; }; /* Get the remote private inferior data associated to INF. */ static remote_inferior * get_remote_inferior (inferior *inf) { if (inf->priv == NULL) inf->priv.reset (new remote_inferior); return gdb::checked_static_cast (inf->priv.get ()); } /* Class used to track the construction of a vCont packet in the outgoing packet buffer. This is used to send multiple vCont packets if we have more actions than would fit a single packet. */ class vcont_builder { public: explicit vcont_builder (remote_target *remote) : m_remote (remote) { restart (); } void flush (); void push_action (ptid_t ptid, bool step, gdb_signal siggnal); private: void restart (); /* The remote target. */ remote_target *m_remote; /* Pointer to the first action. P points here if no action has been appended yet. */ char *m_first_action; /* Where the next action will be appended. */ char *m_p; /* The end of the buffer. Must never write past this. */ char *m_endp; }; /* Prepare the outgoing buffer for a new vCont packet. */ void vcont_builder::restart () { struct remote_state *rs = m_remote->get_remote_state (); m_p = rs->buf.data (); m_endp = m_p + m_remote->get_remote_packet_size (); m_p += xsnprintf (m_p, m_endp - m_p, "vCont"); m_first_action = m_p; } /* If the vCont packet being built has any action, send it to the remote end. */ void vcont_builder::flush () { struct remote_state *rs; if (m_p == m_first_action) return; rs = m_remote->get_remote_state (); m_remote->putpkt (rs->buf); m_remote->getpkt (&rs->buf, 0); if (strcmp (rs->buf.data (), "OK") != 0) error (_("Unexpected vCont reply in non-stop mode: %s"), rs->buf.data ()); } /* The largest action is range-stepping, with its two addresses. This is more than sufficient. If a new, bigger action is created, it'll quickly trigger a failed assertion in append_resumption (and we'll just bump this). */ #define MAX_ACTION_SIZE 200 /* Append a new vCont action in the outgoing packet being built. If the action doesn't fit the packet along with previous actions, push what we've got so far to the remote end and start over a new vCont packet (with the new action). */ void vcont_builder::push_action (ptid_t ptid, bool step, gdb_signal siggnal) { char buf[MAX_ACTION_SIZE + 1]; char *endp = m_remote->append_resumption (buf, buf + sizeof (buf), ptid, step, siggnal); /* Check whether this new action would fit in the vCont packet along with previous actions. If not, send what we've got so far and start a new vCont packet. */ size_t rsize = endp - buf; if (rsize > m_endp - m_p) { flush (); restart (); /* Should now fit. */ gdb_assert (rsize <= m_endp - m_p); } memcpy (m_p, buf, rsize); m_p += rsize; *m_p = '\0'; } /* to_commit_resume implementation. */ void remote_target::commit_resumed () { /* If connected in all-stop mode, we'd send the remote resume request directly from remote_resume. Likewise if reverse-debugging, as there are no defined vCont actions for reverse execution. */ if (!target_is_non_stop_p () || ::execution_direction == EXEC_REVERSE) return; /* Try to send wildcard actions ("vCont;c" or "vCont;c:pPID.-1") instead of resuming all threads of each process individually. However, if any thread of a process must remain halted, we can't send wildcard resumes and must send one action per thread. Care must be taken to not resume threads/processes the server side already told us are stopped, but the core doesn't know about yet, because the events are still in the vStopped notification queue. For example: #1 => vCont s:p1.1;c #2 <= OK #3 <= %Stopped T05 p1.1 #4 => vStopped #5 <= T05 p1.2 #6 => vStopped #7 <= OK #8 (infrun handles the stop for p1.1 and continues stepping) #9 => vCont s:p1.1;c The last vCont above would resume thread p1.2 by mistake, because the server has no idea that the event for p1.2 had not been handled yet. The server side must similarly ignore resume actions for the thread that has a pending %Stopped notification (and any other threads with events pending), until GDB acks the notification with vStopped. Otherwise, e.g., the following case is mishandled: #1 => g (or any other packet) #2 <= [registers] #3 <= %Stopped T05 p1.2 #4 => vCont s:p1.1;c #5 <= OK Above, the server must not resume thread p1.2. GDB can't know that p1.2 stopped until it acks the %Stopped notification, and since from GDB's perspective all threads should be running, it sends a "c" action. Finally, special care must also be given to handling fork/vfork events. A (v)fork event actually tells us that two processes stopped -- the parent and the child. Until we follow the fork, we must not resume the child. Therefore, if we have a pending fork follow, we must not send a global wildcard resume action (vCont;c). We can still send process-wide wildcards though. */ /* Start by assuming a global wildcard (vCont;c) is possible. */ bool may_global_wildcard_vcont = true; /* And assume every process is individually wildcard-able too. */ for (inferior *inf : all_non_exited_inferiors (this)) { remote_inferior *priv = get_remote_inferior (inf); priv->may_wildcard_vcont = true; } /* Check for any pending events (not reported or processed yet) and disable process and global wildcard resumes appropriately. */ check_pending_events_prevent_wildcard_vcont (&may_global_wildcard_vcont); bool any_pending_vcont_resume = false; for (thread_info *tp : all_non_exited_threads (this)) { remote_thread_info *priv = get_remote_thread_info (tp); /* If a thread of a process is not meant to be resumed, then we can't wildcard that process. */ if (priv->get_resume_state () == resume_state::NOT_RESUMED) { get_remote_inferior (tp->inf)->may_wildcard_vcont = false; /* And if we can't wildcard a process, we can't wildcard everything either. */ may_global_wildcard_vcont = false; continue; } if (priv->get_resume_state () == resume_state::RESUMED_PENDING_VCONT) any_pending_vcont_resume = true; /* If a thread is the parent of an unfollowed fork, then we can't do a global wildcard, as that would resume the fork child. */ if (thread_pending_fork_status (tp) != nullptr) may_global_wildcard_vcont = false; } /* We didn't have any resumed thread pending a vCont resume, so nothing to do. */ if (!any_pending_vcont_resume) return; /* Now let's build the vCont packet(s). Actions must be appended from narrower to wider scopes (thread -> process -> global). If we end up with too many actions for a single packet vcont_builder flushes the current vCont packet to the remote side and starts a new one. */ struct vcont_builder vcont_builder (this); /* Threads first. */ for (thread_info *tp : all_non_exited_threads (this)) { remote_thread_info *remote_thr = get_remote_thread_info (tp); /* If the thread was previously vCont-resumed, no need to send a specific action for it. If we didn't receive a resume request for it, don't send an action for it either. */ if (remote_thr->get_resume_state () != resume_state::RESUMED_PENDING_VCONT) continue; gdb_assert (!thread_is_in_step_over_chain (tp)); /* We should never be commit-resuming a thread that has a stop reply. Otherwise, we would end up reporting a stop event for a thread while it is running on the remote target. */ remote_state *rs = get_remote_state (); for (const auto &stop_reply : rs->stop_reply_queue) gdb_assert (stop_reply->ptid != tp->ptid); const resumed_pending_vcont_info &info = remote_thr->resumed_pending_vcont_info (); /* Check if we need to send a specific action for this thread. If not, it will be included in a wildcard resume instead. */ if (info.step || info.sig != GDB_SIGNAL_0 || !get_remote_inferior (tp->inf)->may_wildcard_vcont) vcont_builder.push_action (tp->ptid, info.step, info.sig); remote_thr->set_resumed (); } /* Now check whether we can send any process-wide wildcard. This is to avoid sending a global wildcard in the case nothing is supposed to be resumed. */ bool any_process_wildcard = false; for (inferior *inf : all_non_exited_inferiors (this)) { if (get_remote_inferior (inf)->may_wildcard_vcont) { any_process_wildcard = true; break; } } if (any_process_wildcard) { /* If all processes are wildcard-able, then send a single "c" action, otherwise, send an "all (-1) threads of process" continue action for each running process, if any. */ if (may_global_wildcard_vcont) { vcont_builder.push_action (minus_one_ptid, false, GDB_SIGNAL_0); } else { for (inferior *inf : all_non_exited_inferiors (this)) { if (get_remote_inferior (inf)->may_wildcard_vcont) { vcont_builder.push_action (ptid_t (inf->pid), false, GDB_SIGNAL_0); } } } } vcont_builder.flush (); } /* Implementation of target_has_pending_events. */ bool remote_target::has_pending_events () { if (target_can_async_p ()) { remote_state *rs = get_remote_state (); if (async_event_handler_marked (rs->remote_async_inferior_event_token)) return true; /* Note that BUFCNT can be negative, indicating sticky error. */ if (rs->remote_desc->bufcnt != 0) return true; } return false; } /* 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. */ void remote_target::remote_stop_ns (ptid_t ptid) { struct remote_state *rs = get_remote_state (); char *p = rs->buf.data (); char *endp = p + get_remote_packet_size (); /* If any thread that needs to stop was resumed but pending a vCont resume, generate a phony stop_reply. However, first check whether the thread wasn't resumed with a signal. Generating a phony stop in that case would result in losing the signal. */ bool needs_commit = false; for (thread_info *tp : all_non_exited_threads (this, ptid)) { remote_thread_info *remote_thr = get_remote_thread_info (tp); if (remote_thr->get_resume_state () == resume_state::RESUMED_PENDING_VCONT) { const resumed_pending_vcont_info &info = remote_thr->resumed_pending_vcont_info (); if (info.sig != GDB_SIGNAL_0) { /* This signal must be forwarded to the inferior. We could commit-resume just this thread, but its simpler to just commit-resume everything. */ needs_commit = true; break; } } } if (needs_commit) commit_resumed (); else for (thread_info *tp : all_non_exited_threads (this, ptid)) { remote_thread_info *remote_thr = get_remote_thread_info (tp); if (remote_thr->get_resume_state () == resume_state::RESUMED_PENDING_VCONT) { remote_debug_printf ("Enqueueing phony stop reply for thread pending " "vCont-resume (%d, %ld, %s)", tp->ptid.pid(), tp->ptid.lwp (), pulongest (tp->ptid.tid ())); /* Check that the thread wasn't resumed with a signal. Generating a phony stop would result in losing the signal. */ const resumed_pending_vcont_info &info = remote_thr->resumed_pending_vcont_info (); gdb_assert (info.sig == GDB_SIGNAL_0); stop_reply *sr = new stop_reply (); sr->ptid = tp->ptid; sr->rs = rs; sr->ws.set_stopped (GDB_SIGNAL_0); sr->arch = tp->inf->gdbarch; sr->stop_reason = TARGET_STOPPED_BY_NO_REASON; sr->watch_data_address = 0; sr->core = 0; this->push_stop_reply (sr); /* Pretend that this thread was actually resumed on the remote target, then stopped. If we leave it in the RESUMED_PENDING_VCONT state and the commit_resumed method is called while the stop reply is still in the queue, we'll end up reporting a stop event to the core for that thread while it is running on the remote target... that would be bad. */ remote_thr->set_resumed (); } } if (!rs->supports_vCont.t) error (_("Remote server does not support stopping threads")); if (ptid == minus_one_ptid || (!m_features.remote_multi_process_p () && ptid.is_pid ())) p += xsnprintf (p, endp - p, "vCont;t"); else { ptid_t nptid; p += xsnprintf (p, endp - p, "vCont;t:"); if (ptid.is_pid ()) /* All (-1) threads of process. */ nptid = ptid_t (ptid.pid (), -1); 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, 0); if (strcmp (rs->buf.data (), "OK") != 0) error (_("Stopping %s failed: %s"), target_pid_to_str (ptid).c_str (), rs->buf.data ()); } /* All-stop version of target_interrupt. Sends a break or a ^C to interrupt the remote target. It is undefined which thread of which process reports the interrupt. */ void remote_target::remote_interrupt_as () { 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 pending stop events will be collected in remote_wait. */ if (stop_reply_queue_length () > 0) return; /* Send interrupt_sequence to remote target. */ send_interrupt_sequence (); } /* Non-stop version of target_interrupt. Uses `vCtrlC' to interrupt the remote target. It is undefined which thread of which process reports the interrupt. Throws an error if the packet is not supported by the server. */ void remote_target::remote_interrupt_ns () { struct remote_state *rs = get_remote_state (); char *p = rs->buf.data (); char *endp = p + get_remote_packet_size (); xsnprintf (p, endp - p, "vCtrlC"); /* In non-stop, we get an immediate OK reply. The stop reply will come in asynchronously by notification. */ putpkt (rs->buf); getpkt (&rs->buf, 0); switch (m_features.packet_ok (rs->buf, PACKET_vCtrlC)) { case PACKET_OK: break; case PACKET_UNKNOWN: error (_("No support for interrupting the remote target.")); case PACKET_ERROR: error (_("Interrupting target failed: %s"), rs->buf.data ()); } } /* Implement the to_stop function for the remote targets. */ void remote_target::stop (ptid_t ptid) { REMOTE_SCOPED_DEBUG_ENTER_EXIT; if (target_is_non_stop_p ()) remote_stop_ns (ptid); else { /* We don't currently have a way to transparently pause the remote target in all-stop mode. Interrupt it instead. */ remote_interrupt_as (); } } /* Implement the to_interrupt function for the remote targets. */ void remote_target::interrupt () { REMOTE_SCOPED_DEBUG_ENTER_EXIT; if (target_is_non_stop_p ()) remote_interrupt_ns (); else remote_interrupt_as (); } /* Implement the to_pass_ctrlc function for the remote targets. */ void remote_target::pass_ctrlc () { REMOTE_SCOPED_DEBUG_ENTER_EXIT; struct remote_state *rs = get_remote_state (); /* If we're starting up, we're not fully synced yet. Quit immediately. */ if (rs->starting_up) quit (); /* If ^C has already been sent once, offer to disconnect. */ else if (rs->ctrlc_pending_p) interrupt_query (); else target_interrupt (); } /* Ask the user what to do when an interrupt is received. */ void remote_target::interrupt_query () { struct remote_state *rs = get_remote_state (); if (rs->waiting_for_stop_reply && rs->ctrlc_pending_p) { if (query (_("The target is not responding to interrupt requests.\n" "Stop debugging it? "))) { remote_unpush_target (this); throw_error (TARGET_CLOSE_ERROR, _("Disconnected from target.")); } } else { if (query (_("Interrupted while waiting for the program.\n" "Give up waiting? "))) quit (); } } /* 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. */ void remote_target::terminal_inferior () { /* 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. */ } void remote_target::terminal_ours () { } static void remote_console_output (const char *msg) { const 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; gdb_stdtarg->puts (tb); } gdb_stdtarg->flush (); } /* Return the length of the stop reply queue. */ int remote_target::stop_reply_queue_length () { remote_state *rs = get_remote_state (); return rs->stop_reply_queue.size (); } static void remote_notif_stop_parse (remote_target *remote, const notif_client *self, const char *buf, struct notif_event *event) { remote->remote_parse_stop_reply (buf, (struct stop_reply *) event); } static void remote_notif_stop_ack (remote_target *remote, const notif_client *self, const char *buf, struct notif_event *event) { struct stop_reply *stop_reply = (struct stop_reply *) event; /* acknowledge */ putpkt (remote, self->ack_command); /* Kind can be TARGET_WAITKIND_IGNORE if we have meanwhile discarded the notification. It was left in the queue because we need to acknowledge it and pull the rest of the notifications out. */ if (stop_reply->ws.kind () != TARGET_WAITKIND_IGNORE) remote->push_stop_reply (stop_reply); } static int remote_notif_stop_can_get_pending_events (remote_target *remote, const 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. */ remote_state *rs = remote->get_remote_state (); mark_async_event_handler (rs->remote_async_inferior_event_token); return 0; } stop_reply::~stop_reply () { for (cached_reg_t ® : regcache) xfree (reg.data); } static notif_event_up remote_notif_stop_alloc_reply () { return notif_event_up (new struct stop_reply ()); } /* A client of notification Stop. */ const 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, }; /* If CONTEXT contains any fork child threads that have not been reported yet, remove them from the CONTEXT list. If such a thread exists it is because we are stopped at a fork catchpoint and have not yet called follow_fork, which will set up the host-side data structures for the new process. */ void remote_target::remove_new_fork_children (threads_listing_context *context) { const notif_client *notif = ¬if_client_stop; /* For any threads stopped at a fork event, remove the corresponding fork child threads from the CONTEXT list. */ for (thread_info *thread : all_non_exited_threads (this)) { const target_waitstatus *ws = thread_pending_fork_status (thread); if (ws == nullptr) continue; context->remove_thread (ws->child_ptid ()); } /* Check for any pending fork events (not reported or processed yet) in process PID and remove those fork child threads from the CONTEXT list as well. */ remote_notif_get_pending_events (notif); for (auto &event : get_remote_state ()->stop_reply_queue) if (event->ws.kind () == TARGET_WAITKIND_FORKED || event->ws.kind () == TARGET_WAITKIND_VFORKED) context->remove_thread (event->ws.child_ptid ()); else if (event->ws.kind () == TARGET_WAITKIND_THREAD_EXITED) context->remove_thread (event->ptid); } /* Check whether any event pending in the vStopped queue would prevent a global or process wildcard vCont action. Set *may_global_wildcard to false if we can't do a global wildcard (vCont;c), and clear the event inferior's may_wildcard_vcont flag if we can't do a process-wide wildcard resume (vCont;c:pPID.-1). */ void remote_target::check_pending_events_prevent_wildcard_vcont (bool *may_global_wildcard) { const notif_client *notif = ¬if_client_stop; remote_notif_get_pending_events (notif); for (auto &event : get_remote_state ()->stop_reply_queue) { if (event->ws.kind () == TARGET_WAITKIND_NO_RESUMED || event->ws.kind () == TARGET_WAITKIND_NO_HISTORY) continue; if (event->ws.kind () == TARGET_WAITKIND_FORKED || event->ws.kind () == TARGET_WAITKIND_VFORKED) *may_global_wildcard = false; /* This may be the first time we heard about this process. Regardless, we must not do a global wildcard resume, otherwise we'd resume this process too. */ *may_global_wildcard = false; if (event->ptid != null_ptid) { inferior *inf = find_inferior_ptid (this, event->ptid); if (inf != NULL) get_remote_inferior (inf)->may_wildcard_vcont = false; } } } /* Discard all pending stop replies of inferior INF. */ void remote_target::discard_pending_stop_replies (struct inferior *inf) { 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 && reply->ptid.pid () == inf->pid) { /* Leave the notification pending, since the server expects that we acknowledge it with vStopped. But clear its contents, so that later on when we acknowledge it, we also discard it. */ remote_debug_printf ("discarding in-flight notification: ptid: %s, ws: %s\n", reply->ptid.to_string().c_str(), reply->ws.to_string ().c_str ()); reply->ws.set_ignore (); } /* Discard the stop replies we have already pulled with vStopped. */ auto iter = std::remove_if (rs->stop_reply_queue.begin (), rs->stop_reply_queue.end (), [=] (const stop_reply_up &event) { return event->ptid.pid () == inf->pid; }); for (auto it = iter; it != rs->stop_reply_queue.end (); ++it) remote_debug_printf ("discarding queued stop reply: ptid: %s, ws: %s\n", reply->ptid.to_string().c_str(), reply->ws.to_string ().c_str ()); rs->stop_reply_queue.erase (iter, rs->stop_reply_queue.end ()); } /* Discard the stop replies for RS in stop_reply_queue. */ void remote_target::discard_pending_stop_replies_in_queue () { remote_state *rs = get_remote_state (); /* Discard the stop replies we have already pulled with vStopped. */ auto iter = std::remove_if (rs->stop_reply_queue.begin (), rs->stop_reply_queue.end (), [=] (const stop_reply_up &event) { return event->rs == rs; }); rs->stop_reply_queue.erase (iter, rs->stop_reply_queue.end ()); } /* Remove the first reply in 'stop_reply_queue' which matches PTID. */ struct stop_reply * remote_target::remote_notif_remove_queued_reply (ptid_t ptid) { remote_state *rs = get_remote_state (); auto iter = std::find_if (rs->stop_reply_queue.begin (), rs->stop_reply_queue.end (), [=] (const stop_reply_up &event) { return event->ptid.matches (ptid); }); struct stop_reply *result; if (iter == rs->stop_reply_queue.end ()) result = nullptr; else { result = iter->release (); rs->stop_reply_queue.erase (iter); } if (notif_debug) gdb_printf (gdb_stdlog, "notif: discard queued event: 'Stop' in %s\n", ptid.to_string ().c_str ()); return result; } /* 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. */ struct stop_reply * remote_target::queued_stop_reply (ptid_t ptid) { remote_state *rs = get_remote_state (); struct stop_reply *r = remote_notif_remove_queued_reply (ptid); if (!rs->stop_reply_queue.empty () && target_can_async_p ()) { /* There's still at least an event left. */ mark_async_event_handler (rs->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. */ void remote_target::push_stop_reply (struct stop_reply *new_event) { remote_state *rs = get_remote_state (); rs->stop_reply_queue.push_back (stop_reply_up (new_event)); if (notif_debug) gdb_printf (gdb_stdlog, "notif: push 'Stop' %s to queue %d\n", new_event->ptid.to_string ().c_str (), int (rs->stop_reply_queue.size ())); /* Mark the pending event queue only if async mode is currently enabled. If async mode is not currently enabled, then, if it later becomes enabled, and there are events in this queue, we will mark the event token at that point, see remote_target::async. */ if (target_is_async_p ()) mark_async_event_handler (rs->remote_async_inferior_event_token); } /* Returns true if we have a stop reply for PTID. */ int remote_target::peek_stop_reply (ptid_t ptid) { remote_state *rs = get_remote_state (); for (auto &event : rs->stop_reply_queue) if (ptid == event->ptid && event->ws.kind () == TARGET_WAITKIND_STOPPED) return 1; return 0; } /* Helper for remote_parse_stop_reply. Return nonzero if the substring starting with P and ending with PEND matches PREFIX. */ static int strprefix (const char *p, const char *pend, const char *prefix) { for ( ; p < pend; p++, prefix++) if (*p != *prefix) return 0; return *prefix == '\0'; } /* Parse the stop reply in BUF. Either the function succeeds, and the result is stored in EVENT, or throws an error. */ void remote_target::remote_parse_stop_reply (const char *buf, stop_reply *event) { remote_arch_state *rsa = NULL; ULONGEST addr; const char *p; int skipregs = 0; event->ptid = null_ptid; event->rs = get_remote_state (); event->ws.set_ignore (); event->stop_reason = TARGET_STOPPED_BY_NO_REASON; event->regcache.clear (); 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) { const char *p1; int fieldsize; p1 = strchr (p, ':'); if (p1 == NULL) error (_("Malformed packet(a) (missing colon): %s\n\ Packet: '%s'\n"), p, buf); if (p == p1) error (_("Malformed packet(a) (missing register number): %s\n\ Packet: '%s'\n"), p, buf); /* Some "registers" are actually extended stop information. Note if you're adding a new entry here: GDB 7.9 and earlier assume that all register "numbers" that start with an hex digit are real register numbers. Make sure the server only sends such a packet if it knows the client understands it. */ if (strprefix (p, p1, "thread")) event->ptid = read_ptid (++p1, &p); else if (strprefix (p, p1, "syscall_entry")) { ULONGEST sysno; p = unpack_varlen_hex (++p1, &sysno); event->ws.set_syscall_entry ((int) sysno); } else if (strprefix (p, p1, "syscall_return")) { ULONGEST sysno; p = unpack_varlen_hex (++p1, &sysno); event->ws.set_syscall_return ((int) sysno); } else if (strprefix (p, p1, "watch") || strprefix (p, p1, "rwatch") || strprefix (p, p1, "awatch")) { event->stop_reason = TARGET_STOPPED_BY_WATCHPOINT; p = unpack_varlen_hex (++p1, &addr); event->watch_data_address = (CORE_ADDR) addr; } else if (strprefix (p, p1, "swbreak")) { event->stop_reason = TARGET_STOPPED_BY_SW_BREAKPOINT; /* Make sure the stub doesn't forget to indicate support with qSupported. */ if (m_features.packet_support (PACKET_swbreak_feature) != PACKET_ENABLE) error (_("Unexpected swbreak stop reason")); /* The value part is documented as "must be empty", though we ignore it, in case we ever decide to make use of it in a backward compatible way. */ p = strchrnul (p1 + 1, ';'); } else if (strprefix (p, p1, "hwbreak")) { event->stop_reason = TARGET_STOPPED_BY_HW_BREAKPOINT; /* Make sure the stub doesn't forget to indicate support with qSupported. */ if (m_features.packet_support (PACKET_hwbreak_feature) != PACKET_ENABLE) error (_("Unexpected hwbreak stop reason")); /* See above. */ p = strchrnul (p1 + 1, ';'); } else if (strprefix (p, p1, "library")) { event->ws.set_loaded (); p = strchrnul (p1 + 1, ';'); } else if (strprefix (p, p1, "replaylog")) { event->ws.set_no_history (); /* p1 will indicate "begin" or "end", but it makes no difference for now, so ignore it. */ p = strchrnul (p1 + 1, ';'); } else if (strprefix (p, p1, "core")) { ULONGEST c; p = unpack_varlen_hex (++p1, &c); event->core = c; } else if (strprefix (p, p1, "fork")) event->ws.set_forked (read_ptid (++p1, &p)); else if (strprefix (p, p1, "vfork")) event->ws.set_vforked (read_ptid (++p1, &p)); else if (strprefix (p, p1, "vforkdone")) { event->ws.set_vfork_done (); p = strchrnul (p1 + 1, ';'); } else if (strprefix (p, p1, "exec")) { ULONGEST ignored; int pathlen; /* Determine the length of the execd pathname. */ p = unpack_varlen_hex (++p1, &ignored); pathlen = (p - p1) / 2; /* Save the pathname for event reporting and for the next run command. */ gdb::unique_xmalloc_ptr pathname ((char *) xmalloc (pathlen + 1)); hex2bin (p1, (gdb_byte *) pathname.get (), pathlen); pathname.get ()[pathlen] = '\0'; /* This is freed during event handling. */ event->ws.set_execd (std::move (pathname)); /* Skip the registers included in this packet, since they may be for an architecture different from the one used by the original program. */ skipregs = 1; } else if (strprefix (p, p1, "create")) { event->ws.set_thread_created (); p = strchrnul (p1 + 1, ';'); } else { ULONGEST pnum; const char *p_temp; if (skipregs) { p = strchrnul (p1 + 1, ';'); p++; continue; } /* Maybe a real ``P'' register number. */ p_temp = unpack_varlen_hex (p, &pnum); /* If the first invalid character is the colon, we got a register number. Otherwise, it's an unknown stop reason. */ if (p_temp == p1) { /* If we haven't parsed the event's thread yet, find it now, in order to find the architecture of the reported expedited registers. */ if (event->ptid == null_ptid) { /* If there is no thread-id information then leave the event->ptid as null_ptid. Later in process_stop_reply we will pick a suitable thread. */ const char *thr = strstr (p1 + 1, ";thread:"); if (thr != NULL) event->ptid = read_ptid (thr + strlen (";thread:"), NULL); } if (rsa == NULL) { inferior *inf = (event->ptid == null_ptid ? NULL : find_inferior_ptid (this, event->ptid)); /* If this is the first time we learn anything about this process, skip the registers included in this packet, since we don't yet know which architecture to use to parse them. We'll determine the architecture later when we process the stop reply and retrieve the target description, via remote_notice_new_inferior -> post_create_inferior. */ if (inf == NULL) { p = strchrnul (p1 + 1, ';'); p++; continue; } event->arch = inf->gdbarch; rsa = event->rs->get_remote_arch_state (event->arch); } packet_reg *reg = packet_reg_from_pnum (event->arch, rsa, pnum); cached_reg_t cached_reg; 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; cached_reg.data = (gdb_byte *) xmalloc (register_size (event->arch, reg->regnum)); p = p1 + 1; fieldsize = hex2bin (p, cached_reg.data, register_size (event->arch, reg->regnum)); p += 2 * fieldsize; if (fieldsize < register_size (event->arch, reg->regnum)) warning (_("Remote reply is too short: %s"), buf); event->regcache.push_back (cached_reg); } else { /* Not a number. Silently skip unknown optional info. */ p = strchrnul (p1 + 1, ';'); } } 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; sig = (fromhex (buf[1]) << 4) + fromhex (buf[2]); if (GDB_SIGNAL_FIRST <= sig && sig < GDB_SIGNAL_LAST) event->ws.set_stopped ((enum gdb_signal) sig); else event->ws.set_stopped (GDB_SIGNAL_UNKNOWN); } break; case 'w': /* Thread exited. */ { ULONGEST value; p = unpack_varlen_hex (&buf[1], &value); event->ws.set_thread_exited (value); if (*p != ';') error (_("stop reply packet badly formatted: %s"), buf); event->ptid = read_ptid (++p, NULL); break; } case 'W': /* Target exited. */ case 'X': { 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.set_exited (value); } else { /* The remote process exited with a signal. */ if (GDB_SIGNAL_FIRST <= value && value < GDB_SIGNAL_LAST) event->ws.set_signalled ((enum gdb_signal) value); else event->ws.set_signalled (GDB_SIGNAL_UNKNOWN); } /* If no process is specified, return null_ptid, and let the caller figure out the right process to use. */ int pid = 0; if (*p == '\0') ; else if (*p == ';') { p++; if (*p == '\0') ; else if (startswith (p, "process:")) { 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 = ptid_t (pid); } break; case 'N': event->ws.set_no_resumed (); event->ptid = minus_one_ptid; break; } } /* 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 2.5) <-- (registers reply to step #2.3) Eventually 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_target::remote_notif_get_pending_events (const notif_client *nc) { struct remote_state *rs = get_remote_state (); if (rs->notif_state->pending_event[nc->id] != NULL) { if (notif_debug) gdb_printf (gdb_stdlog, "notif: process: '%s' ack pending event\n", nc->name); /* acknowledge */ nc->ack (this, nc, rs->buf.data (), rs->notif_state->pending_event[nc->id]); rs->notif_state->pending_event[nc->id] = NULL; while (1) { getpkt (&rs->buf, 0); if (strcmp (rs->buf.data (), "OK") == 0) break; else remote_notif_ack (this, nc, rs->buf.data ()); } } else { if (notif_debug) gdb_printf (gdb_stdlog, "notif: process: '%s' no pending reply\n", nc->name); } } /* Wrapper around remote_target::remote_notif_get_pending_events to avoid having to export the whole remote_target class. */ void remote_notif_get_pending_events (remote_target *remote, const notif_client *nc) { remote->remote_notif_get_pending_events (nc); } /* Called from process_stop_reply when the stop packet we are responding to didn't include a process-id or thread-id. STATUS is the stop event we are responding to. It is the task of this function to select a suitable thread (or process) and return its ptid, this is the thread (or process) we will assume the stop event came from. In some cases there isn't really any choice about which thread (or process) is selected, a basic remote with a single process containing a single thread might choose not to send any process-id or thread-id in its stop packets, this function will select and return the one and only thread. However, if a target supports multiple threads (or processes) and still doesn't include a thread-id (or process-id) in its stop packet then first, this is a badly behaving target, and second, we're going to have to select a thread (or process) at random and use that. This function will print a warning to the user if it detects that there is the possibility that GDB is guessing which thread (or process) to report. Note that this is called before GDB fetches the updated thread list from the target. So it's possible for the stop reply to be ambiguous and for GDB to not realize it. For example, if there's initially one thread, the target spawns a second thread, and then sends a stop reply without an id that concerns the first thread. GDB will assume the stop reply is about the first thread - the only thread it knows about - without printing a warning. Anyway, if the remote meant for the stop reply to be about the second thread, then it would be really broken, because GDB doesn't know about that thread yet. */ ptid_t remote_target::select_thread_for_ambiguous_stop_reply (const target_waitstatus &status) { REMOTE_SCOPED_DEBUG_ENTER_EXIT; /* Some stop events apply to all threads in an inferior, while others only apply to a single thread. */ bool process_wide_stop = (status.kind () == TARGET_WAITKIND_EXITED || status.kind () == TARGET_WAITKIND_SIGNALLED); remote_debug_printf ("process_wide_stop = %d", process_wide_stop); thread_info *first_resumed_thread = nullptr; bool ambiguous = false; /* Consider all non-exited threads of the target, find the first resumed one. */ for (thread_info *thr : all_non_exited_threads (this)) { remote_thread_info *remote_thr = get_remote_thread_info (thr); if (remote_thr->get_resume_state () != resume_state::RESUMED) continue; if (first_resumed_thread == nullptr) first_resumed_thread = thr; else if (!process_wide_stop || first_resumed_thread->ptid.pid () != thr->ptid.pid ()) ambiguous = true; } gdb_assert (first_resumed_thread != nullptr); remote_debug_printf ("first resumed thread is %s", pid_to_str (first_resumed_thread->ptid).c_str ()); remote_debug_printf ("is this guess ambiguous? = %d", ambiguous); /* Warn if the remote target is sending ambiguous stop replies. */ if (ambiguous) { static bool warned = false; if (!warned) { /* If you are seeing this warning then the remote target has stopped without specifying a thread-id, but the target does have multiple threads (or inferiors), and so GDB is having to guess which thread stopped. Examples of what might cause this are the target sending and 'S' stop packet, or a 'T' stop packet and not including a thread-id. Additionally, the target might send a 'W' or 'X packet without including a process-id, when the target has multiple running inferiors. */ if (process_wide_stop) warning (_("multi-inferior target stopped without " "sending a process-id, using first " "non-exited inferior")); else warning (_("multi-threaded target stopped without " "sending a thread-id, using first " "non-exited thread")); warned = true; } } /* If this is a stop for all threads then don't use a particular threads ptid, instead create a new ptid where only the pid field is set. */ if (process_wide_stop) return ptid_t (first_resumed_thread->ptid.pid ()); else return first_resumed_thread->ptid; } /* 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. */ ptid_t remote_target::process_stop_reply (struct stop_reply *stop_reply, struct target_waitstatus *status) { *status = stop_reply->ws; ptid_t ptid = stop_reply->ptid; /* If no thread/process was reported by the stub then select a suitable thread/process. */ if (ptid == null_ptid) ptid = select_thread_for_ambiguous_stop_reply (*status); gdb_assert (ptid != null_ptid); if (status->kind () != TARGET_WAITKIND_EXITED && status->kind () != TARGET_WAITKIND_SIGNALLED && status->kind () != TARGET_WAITKIND_NO_RESUMED) { /* Expedited registers. */ if (!stop_reply->regcache.empty ()) { struct regcache *regcache = get_thread_arch_regcache (this, ptid, stop_reply->arch); for (cached_reg_t ® : stop_reply->regcache) { regcache->raw_supply (reg.num, reg.data); xfree (reg.data); } stop_reply->regcache.clear (); } remote_notice_new_inferior (ptid, false); remote_thread_info *remote_thr = get_remote_thread_info (this, ptid); remote_thr->core = stop_reply->core; remote_thr->stop_reason = stop_reply->stop_reason; remote_thr->watch_data_address = stop_reply->watch_data_address; if (target_is_non_stop_p ()) { /* If the target works in non-stop mode, a stop-reply indicates that only this thread stopped. */ remote_thr->set_not_resumed (); } else { /* If the target works in all-stop mode, a stop-reply indicates that all the target's threads stopped. */ for (thread_info *tp : all_non_exited_threads (this)) get_remote_thread_info (tp)->set_not_resumed (); } } delete stop_reply; return ptid; } /* The non-stop mode version of target_wait. */ ptid_t remote_target::wait_ns (ptid_t ptid, struct target_waitstatus *status, target_wait_flags 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, 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.data ()); break; case 'O': /* Console output. */ remote_console_output (&rs->buf[1]); break; default: warning (_("Invalid remote reply: %s"), rs->buf.data ()); 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 (¬if_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->set_ignore (); return minus_one_ptid; } /* Otherwise do a blocking wait. */ ret = getpkt_or_notif_sane (&rs->buf, 1 /* forever */, &is_notif); } } /* Return the first resumed thread. */ static ptid_t first_remote_resumed_thread (remote_target *target) { for (thread_info *tp : all_non_exited_threads (target, minus_one_ptid)) if (tp->resumed ()) return tp->ptid; return null_ptid; } /* Wait until the remote machine stops, then return, storing status in STATUS just as `wait' would. */ ptid_t remote_target::wait_as (ptid_t ptid, target_waitstatus *status, target_wait_flags options) { struct remote_state *rs = get_remote_state (); ptid_t event_ptid = null_ptid; char *buf; struct stop_reply *stop_reply; again: status->set_ignore (); stop_reply = queued_stop_reply (ptid); if (stop_reply != NULL) { /* None of the paths that push a stop reply onto the queue should have set the waiting_for_stop_reply flag. */ gdb_assert (!rs->waiting_for_stop_reply); event_ptid = process_stop_reply (stop_reply, status); } else { int forever = ((options & TARGET_WNOHANG) == 0 && rs->wait_forever_enabled_p); if (!rs->waiting_for_stop_reply) { status->set_no_resumed (); return minus_one_ptid; } /* 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. */ int is_notif; int ret = getpkt_or_notif_sane (&rs->buf, forever, &is_notif); /* GDB gets a notification. Return to core as this event is not interesting. */ if (ret != -1 && is_notif) return minus_one_ptid; if (ret == -1 && (options & TARGET_WNOHANG) != 0) return minus_one_ptid; buf = rs->buf.data (); /* 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. */ rs->waiting_for_stop_reply = 0; warning (_("Remote failure reply: %s"), buf); status->set_stopped (GDB_SIGNAL_0); break; case 'F': /* File-I/O request. */ /* GDB may access the inferior memory while handling the File-I/O request, but we don't want GDB accessing memory while waiting for a stop reply. See the comments in putpkt_binary. Set waiting_for_stop_reply to 0 temporarily. */ rs->waiting_for_stop_reply = 0; remote_fileio_request (this, buf, rs->ctrlc_pending_p); rs->ctrlc_pending_p = 0; /* GDB handled the File-I/O request, and the target is running again. Keep waiting for events. */ rs->waiting_for_stop_reply = 1; break; case 'N': case 'T': case 'S': case 'X': case 'W': { /* There is a stop reply to handle. */ rs->waiting_for_stop_reply = 0; stop_reply = (struct stop_reply *) remote_notif_parse (this, ¬if_client_stop, rs->buf.data ()); event_ptid = process_stop_reply (stop_reply, status); break; } case 'O': /* Console output. */ remote_console_output (buf + 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 (); gdb_printf ("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 (buf, rs->last_sent_step ? "s" : "c"); putpkt (buf); break; } /* fallthrough */ default: warning (_("Invalid remote reply: %s"), buf); break; } } if (status->kind () == TARGET_WAITKIND_NO_RESUMED) return minus_one_ptid; else 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 (event_ptid != null_ptid) record_currthread (rs, event_ptid); else event_ptid = first_remote_resumed_thread (this); } else { /* A process exit. Invalidate our notion of current thread. */ record_currthread (rs, minus_one_ptid); /* It's possible that the packet did not include a pid. */ if (event_ptid == null_ptid) event_ptid = first_remote_resumed_thread (this); /* EVENT_PTID could still be NULL_PTID. Double-check. */ if (event_ptid == null_ptid) event_ptid = magic_null_ptid; } return event_ptid; } /* Wait until the remote machine stops, then return, storing status in STATUS just as `wait' would. */ ptid_t remote_target::wait (ptid_t ptid, struct target_waitstatus *status, target_wait_flags options) { REMOTE_SCOPED_DEBUG_ENTER_EXIT; remote_state *rs = get_remote_state (); /* Start by clearing the flag that asks for our wait method to be called, we'll mark it again at the end if needed. If the target is not in async mode then the async token should not be marked. */ if (target_is_async_p ()) clear_async_event_handler (rs->remote_async_inferior_event_token); else gdb_assert (!async_event_handler_marked (rs->remote_async_inferior_event_token)); ptid_t event_ptid; if (target_is_non_stop_p ()) event_ptid = wait_ns (ptid, status, options); else event_ptid = wait_as (ptid, status, options); if (target_is_async_p ()) { /* If there are events left in the queue, or unacknowledged notifications, then tell the event loop to call us again. */ if (!rs->stop_reply_queue.empty () || rs->notif_state->pending_event[notif_client_stop.id] != nullptr) mark_async_event_handler (rs->remote_async_inferior_event_token); } return event_ptid; } /* Fetch a single register using a 'p' packet. */ int remote_target::fetch_register_using_p (struct regcache *regcache, packet_reg *reg) { struct gdbarch *gdbarch = regcache->arch (); struct remote_state *rs = get_remote_state (); char *buf, *p; gdb_byte *regp = (gdb_byte *) alloca (register_size (gdbarch, reg->regnum)); int i; if (m_features.packet_support (PACKET_p) == PACKET_DISABLE) return 0; if (reg->pnum == -1) return 0; p = rs->buf.data (); *p++ = 'p'; p += hexnumstr (p, reg->pnum); *p++ = '\0'; putpkt (rs->buf); getpkt (&rs->buf, 0); buf = rs->buf.data (); switch (m_features.packet_ok (rs->buf, 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 (regcache->arch (), reg->regnum), buf); } /* If this register is unfetchable, tell the regcache. */ if (buf[0] == 'x') { regcache->raw_supply (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 (reg->regnum, regp); return 1; } /* Fetch the registers included in the target's 'g' packet. */ int remote_target::send_g_packet () { struct remote_state *rs = get_remote_state (); int buf_len; xsnprintf (rs->buf.data (), get_remote_packet_size (), "g"); putpkt (rs->buf); getpkt (&rs->buf, 0); if (packet_check_result (rs->buf) == PACKET_ERROR) error (_("Could not read registers; remote failure reply '%s'"), rs->buf.data ()); /* 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. */ { remote_debug_printf ("Bad register packet; fetching a new packet"); getpkt (&rs->buf, 0); } buf_len = strlen (rs->buf.data ()); /* Sanity check the received packet. */ if (buf_len % 2 != 0) error (_("Remote 'g' packet reply is of odd length: %s"), rs->buf.data ()); return buf_len / 2; } void remote_target::process_g_packet (struct regcache *regcache) { struct gdbarch *gdbarch = regcache->arch (); struct remote_state *rs = get_remote_state (); remote_arch_state *rsa = rs->get_remote_arch_state (gdbarch); int i, buf_len; char *p; char *regs; buf_len = strlen (rs->buf.data ()); /* Further sanity checks, with knowledge of the architecture. */ if (buf_len > 2 * rsa->sizeof_g_packet) error (_("Remote 'g' packet reply is too long (expected %ld bytes, got %d " "bytes): %s"), rsa->sizeof_g_packet, buf_len / 2, rs->buf.data ()); /* 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) { long sizeof_g_packet = buf_len / 2; for (i = 0; i < gdbarch_num_regs (gdbarch); i++) { long offset = rsa->regs[i].offset; long reg_size = register_size (gdbarch, i); if (rsa->regs[i].pnum == -1) continue; if (offset >= sizeof_g_packet) rsa->regs[i].in_g_packet = 0; else if (offset + reg_size > sizeof_g_packet) error (_("Truncated register %d in remote 'g' packet"), i); else rsa->regs[i].in_g_packet = 1; } /* Looks valid enough, we can assume this is the correct length for a 'g' packet. It's important not to adjust rsa->sizeof_g_packet if we have truncated registers otherwise this "if" won't be run the next time the method is called with a packet of the same size and one of the internal errors below will trigger instead. */ rsa->sizeof_g_packet = sizeof_g_packet; } regs = (char *) 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.data (); 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 (_("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]; long reg_size = register_size (gdbarch, i); if (r->in_g_packet) { if ((r->offset + reg_size) * 2 > strlen (rs->buf.data ())) /* This shouldn't happen - we adjusted in_g_packet above. */ internal_error (_("unexpected end of 'g' packet reply")); else if (rs->buf[r->offset * 2] == 'x') { gdb_assert (r->offset * 2 < strlen (rs->buf.data ())); /* The register isn't available, mark it as such (at the same time setting the value to zero). */ regcache->raw_supply (r->regnum, NULL); } else regcache->raw_supply (r->regnum, regs + r->offset); } } } void remote_target::fetch_registers_using_g (struct regcache *regcache) { send_g_packet (); process_g_packet (regcache); } /* Make the remote selected traceframe match GDB's selected traceframe. */ void remote_target::set_remote_traceframe () { 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")); } void remote_target::fetch_registers (struct regcache *regcache, int regnum) { struct gdbarch *gdbarch = regcache->arch (); struct remote_state *rs = get_remote_state (); remote_arch_state *rsa = rs->get_remote_arch_state (gdbarch); int i; set_remote_traceframe (); set_general_thread (regcache->ptid ()); if (regnum >= 0) { packet_reg *reg = packet_reg_from_regnum (gdbarch, 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 (reg->regnum, NULL); return; } fetch_registers_using_g (regcache); for (i = 0; i < gdbarch_num_regs (gdbarch); 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 (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. */ void remote_target::prepare_to_store (struct regcache *regcache) { struct remote_state *rs = get_remote_state (); remote_arch_state *rsa = rs->get_remote_arch_state (regcache->arch ()); int i; /* Make sure the entire registers array is valid. */ switch (m_features.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 (regcache->arch ()); i++) if (rsa->regs[i].in_g_packet) regcache->raw_update (rsa->regs[i].regnum); break; case PACKET_ENABLE: break; } } /* Helper: Attempt to store REGNUM using the P packet. Return fail IFF packet was not recognized. */ int remote_target::store_register_using_P (const struct regcache *regcache, packet_reg *reg) { struct gdbarch *gdbarch = regcache->arch (); struct remote_state *rs = get_remote_state (); /* Try storing a single register. */ char *buf = rs->buf.data (); gdb_byte *regp = (gdb_byte *) alloca (register_size (gdbarch, reg->regnum)); char *p; if (m_features.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 (reg->regnum, regp); bin2hex (regp, p, register_size (gdbarch, reg->regnum)); putpkt (rs->buf); getpkt (&rs->buf, 0); switch (m_features.packet_ok (rs->buf, 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.data ()); case PACKET_UNKNOWN: return 0; default: internal_error (_("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. */ void remote_target::store_registers_using_G (const struct regcache *regcache) { struct remote_state *rs = get_remote_state (); remote_arch_state *rsa = rs->get_remote_arch_state (regcache->arch ()); gdb_byte *regs; char *p; /* Extract all the registers in the regcache copying them into a local buffer. */ { int i; regs = (gdb_byte *) alloca (rsa->sizeof_g_packet); memset (regs, 0, rsa->sizeof_g_packet); for (i = 0; i < gdbarch_num_regs (regcache->arch ()); i++) { struct packet_reg *r = &rsa->regs[i]; if (r->in_g_packet) regcache->raw_collect (r->regnum, regs + r->offset); } } /* Command describes registers byte by byte, each byte encoded as two hex characters. */ p = rs->buf.data (); *p++ = 'G'; bin2hex (regs, p, rsa->sizeof_g_packet); putpkt (rs->buf); getpkt (&rs->buf, 0); if (packet_check_result (rs->buf) == PACKET_ERROR) error (_("Could not write registers; remote failure reply '%s'"), rs->buf.data ()); } /* Store register REGNUM, or all registers if REGNUM == -1, from the contents of the register cache buffer. FIXME: ignores errors. */ void remote_target::store_registers (struct regcache *regcache, int regnum) { struct gdbarch *gdbarch = regcache->arch (); struct remote_state *rs = get_remote_state (); remote_arch_state *rsa = rs->get_remote_arch_state (gdbarch); int i; set_remote_traceframe (); set_general_thread (regcache->ptid ()); if (regnum >= 0) { packet_reg *reg = packet_reg_from_regnum (gdbarch, 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 (gdbarch); 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 std::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". */ void remote_target::check_binary_download (CORE_ADDR addr) { struct remote_state *rs = get_remote_state (); switch (m_features.packet_support (PACKET_X)) { case PACKET_DISABLE: break; case PACKET_ENABLE: break; case PACKET_SUPPORT_UNKNOWN: { char *p; p = rs->buf.data (); *p++ = 'X'; p += hexnumstr (p, (ULONGEST) addr); *p++ = ','; p += hexnumstr (p, (ULONGEST) 0); *p++ = ':'; *p = '\0'; putpkt_binary (rs->buf.data (), (int) (p - rs->buf.data ())); getpkt (&rs->buf, 0); if (rs->buf[0] == '\0') { remote_debug_printf ("binary downloading NOT supported by target"); m_features.m_protocol_packets[PACKET_X].support = PACKET_DISABLE; } else { remote_debug_printf ("binary downloading supported by target"); m_features.m_protocol_packets[PACKET_X].support = PACKET_ENABLE; } break; } } } /* Helper function to resize the payload in order to try to get a good alignment. We try to write an amount of data such that the next write will start on an address aligned on REMOTE_ALIGN_WRITES. */ static int align_for_efficient_write (int todo, CORE_ADDR memaddr) { return ((memaddr + todo) & ~(REMOTE_ALIGN_WRITES - 1)) - memaddr; } /* 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_UNITS is the number of addressable units to write. UNIT_SIZE is the length in bytes of an addressable unit. 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
,: where encoding of is terminated by PACKET_FORMAT. If USE_LENGTH is 0, then the field and the preceding comma are omitted. Return the transferred status, error or OK (an 'enum target_xfer_status' value). Save the number of addressable units transferred in *XFERED_LEN_UNITS. Only transfer a single packet. On a platform with an addressable memory size of 2 bytes (UNIT_SIZE == 2), an exchange between gdb and the stub could look like (?? in place of the checksum): -> $m1000,4#?? <- aaaabbbbccccdddd -> $M1000,3:eeeeffffeeee#?? <- OK -> $m1000,4#?? <- eeeeffffeeeedddd */ target_xfer_status remote_target::remote_write_bytes_aux (const char *header, CORE_ADDR memaddr, const gdb_byte *myaddr, ULONGEST len_units, int unit_size, ULONGEST *xfered_len_units, char packet_format, int use_length) { struct remote_state *rs = get_remote_state (); char *p; char *plen = NULL; int plenlen = 0; int todo_units; int units_written; int payload_capacity_bytes; int payload_length_bytes; if (packet_format != 'X' && packet_format != 'M') internal_error (_("remote_write_bytes_aux: bad packet format")); if (len_units == 0) return TARGET_XFER_EOF; payload_capacity_bytes = get_memory_write_packet_size (); /* The packet buffer will be large enough for the payload; get_memory_packet_size ensures this. */ rs->buf[0] = '\0'; /* Compute the size of the actual payload by subtracting out the packet header and footer overhead: "$M,:...#nn". */ payload_capacity_bytes -= strlen ("$,:#NN"); if (!use_length) /* The comma won't be used. */ payload_capacity_bytes += 1; payload_capacity_bytes -= strlen (header); payload_capacity_bytes -= hexnumlen (memaddr); /* Construct the packet excluding the data: "
,:". */ strcat (rs->buf.data (), header); p = rs->buf.data () + 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_units = std::min (len_units, (ULONGEST) payload_capacity_bytes / unit_size); if (use_length) payload_capacity_bytes -= hexnumlen (todo_units); todo_units = std::min (todo_units, payload_capacity_bytes / unit_size); } else { /* Number of bytes that will fit. */ todo_units = std::min (len_units, (ULONGEST) (payload_capacity_bytes / unit_size) / 2); if (use_length) payload_capacity_bytes -= hexnumlen (todo_units); todo_units = std::min (todo_units, (payload_capacity_bytes / unit_size) / 2); } if (todo_units <= 0) internal_error (_("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_units > 2 * REMOTE_ALIGN_WRITES && todo_units < len_units) todo_units = align_for_efficient_write (todo_units, memaddr); /* Append "". */ memaddr = remote_address_masked (memaddr); p += hexnumstr (p, (ULONGEST) memaddr); if (use_length) { /* Append ",". */ *p++ = ','; /* Append the length and retain its location and size. It may need to be adjusted once the packet body has been created. */ plen = p; plenlen = hexnumstr (p, (ULONGEST) todo_units); 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_bytes = remote_escape_output (myaddr, todo_units, unit_size, (gdb_byte *) p, &units_written, payload_capacity_bytes); /* If not all TODO units 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 (units_written < todo_units && units_written > 2 * REMOTE_ALIGN_WRITES) { int new_todo_units; new_todo_units = align_for_efficient_write (units_written, memaddr); if (new_todo_units != units_written) payload_length_bytes = remote_escape_output (myaddr, new_todo_units, unit_size, (gdb_byte *) p, &units_written, payload_capacity_bytes); } p += payload_length_bytes; if (use_length && units_written < todo_units) { /* Escape chars have filled up the buffer prematurely, and we have actually sent fewer units than planned. Fix-up the length field of the packet. Use the same number of characters as before. */ plen += hexnumnstr (plen, (ULONGEST) units_written, 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. */ p += 2 * bin2hex (myaddr, p, todo_units * unit_size); units_written = todo_units; } putpkt_binary (rs->buf.data (), (int) (p - rs->buf.data ())); getpkt (&rs->buf, 0); if (rs->buf[0] == 'E') return TARGET_XFER_E_IO; /* Return UNITS_WRITTEN, not TODO_UNITS, in case escape chars caused us to send fewer units than we'd planned. */ *xfered_len_units = (ULONGEST) units_written; return (*xfered_len_units != 0) ? TARGET_XFER_OK : TARGET_XFER_EOF; } /* 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. */ target_xfer_status remote_target::remote_write_bytes (CORE_ADDR memaddr, const gdb_byte *myaddr, ULONGEST len, int unit_size, ULONGEST *xfered_len) { const char *packet_format = NULL; /* Check whether the target supports binary download. */ check_binary_download (memaddr); switch (m_features.packet_support (PACKET_X)) { case PACKET_ENABLE: packet_format = "X"; break; case PACKET_DISABLE: packet_format = "M"; break; case PACKET_SUPPORT_UNKNOWN: internal_error (_("remote_write_bytes: bad internal state")); default: internal_error (_("bad switch")); } return remote_write_bytes_aux (packet_format, memaddr, myaddr, len, unit_size, 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_UNITS is the number of addressable memory units to read.. UNIT_SIZE is the length in bytes of an addressable unit. Return the transferred status, error or OK (an 'enum target_xfer_status' value). Save the number of bytes transferred in *XFERED_LEN_UNITS. See the comment of remote_write_bytes_aux for an example of memory read/write exchange between gdb and the stub. */ target_xfer_status remote_target::remote_read_bytes_1 (CORE_ADDR memaddr, gdb_byte *myaddr, ULONGEST len_units, int unit_size, ULONGEST *xfered_len_units) { struct remote_state *rs = get_remote_state (); int buf_size_bytes; /* Max size of packet output buffer. */ char *p; int todo_units; int decoded_bytes; buf_size_bytes = get_memory_read_packet_size (); /* The packet buffer will be large enough for the payload; get_memory_packet_size ensures this. */ /* Number of units that will fit. */ todo_units = std::min (len_units, (ULONGEST) (buf_size_bytes / unit_size) / 2); /* Construct "m"","". */ memaddr = remote_address_masked (memaddr); p = rs->buf.data (); *p++ = 'm'; p += hexnumstr (p, (ULONGEST) memaddr); *p++ = ','; p += hexnumstr (p, (ULONGEST) todo_units); *p = '\0'; putpkt (rs->buf); getpkt (&rs->buf, 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.data (); decoded_bytes = hex2bin (p, myaddr, todo_units * unit_size); /* Return what we have. Let higher layers handle partial reads. */ *xfered_len_units = (ULONGEST) (decoded_bytes / unit_size); return (*xfered_len_units != 0) ? TARGET_XFER_OK : TARGET_XFER_EOF; } /* 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. */ target_xfer_status remote_target::remote_xfer_live_readonly_partial (gdb_byte *readbuf, ULONGEST memaddr, ULONGEST len, int unit_size, ULONGEST *xfered_len) { const struct target_section *secp; secp = target_section_by_addr (this, memaddr); if (secp != NULL && (bfd_section_flags (secp->the_bfd_section) & SEC_READONLY)) { ULONGEST memend = memaddr + len; const target_section_table *table = target_get_section_table (this); for (const target_section &p : *table) { if (memaddr >= p.addr) { if (memend <= p.endaddr) { /* Entire transfer is within this section. */ return remote_read_bytes_1 (memaddr, readbuf, len, unit_size, 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, unit_size, 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. */ target_xfer_status remote_target::remote_read_bytes (CORE_ADDR memaddr, gdb_byte *myaddr, ULONGEST len, int unit_size, ULONGEST *xfered_len) { if (len == 0) return TARGET_XFER_EOF; if (get_traceframe_number () != -1) { std::vector 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)) { if (available.empty () || available[0].start != memaddr) { enum target_xfer_status res; /* Don't read into the traceframe's available memory. */ if (!available.empty ()) { LONGEST oldlen = len; len = available[0].start - memaddr; gdb_assert (len <= oldlen); } /* This goes through the topmost target again. */ res = remote_xfer_live_readonly_partial (myaddr, memaddr, len, unit_size, 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 (*xfered_len != 0) ? TARGET_XFER_UNAVAILABLE : TARGET_XFER_EOF; } } /* 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 = available[0].length; } } return remote_read_bytes_1 (memaddr, myaddr, len, unit_size, 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. */ packet_result remote_target::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'; int size = vsnprintf (rs->buf.data (), max_size, format, ap); va_end (ap); if (size >= max_size) internal_error (_("Too long remote packet.")); if (putpkt (rs->buf) < 0) error (_("Communication problem with target.")); rs->buf[0] = '\0'; getpkt (&rs->buf, 0); return packet_check_result (rs->buf); } /* 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; void remote_target::flash_erase (ULONGEST address, LONGEST length) { int addr_size = gdbarch_addr_bit (target_gdbarch ()) / 8; enum packet_result ret; scoped_restore restore_timeout = make_scoped_restore (&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; } } target_xfer_status remote_target::remote_flash_write (ULONGEST address, ULONGEST length, ULONGEST *xfered_len, const gdb_byte *data) { scoped_restore restore_timeout = make_scoped_restore (&remote_timeout, remote_flash_timeout); return remote_write_bytes_aux ("vFlashWrite:", address, data, length, 1, xfered_len,'X', 0); } void remote_target::flash_done () { int ret; scoped_restore restore_timeout = make_scoped_restore (&remote_timeout, remote_flash_timeout); ret = remote_send_printf ("vFlashDone"); switch (ret) { case PACKET_UNKNOWN: error (_("Remote target does not support vFlashDone")); case PACKET_ERROR: error (_("Error finishing flash operation")); default: break; } } /* 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 (remote_target *target, const char *string) { int saved_errno = errno; remote_unpush_target (target); throw_error (TARGET_CLOSE_ERROR, "%s: %s.", string, safe_strerror (saved_errno)); } /* Read a single character from the remote end. The current quit handler is overridden to avoid quitting in the middle of packet sequence, as that would break communication with the remote server. See remote_serial_quit_handler for more detail. */ int remote_target::readchar (int timeout) { int ch; struct remote_state *rs = get_remote_state (); { scoped_restore restore_quit_target = make_scoped_restore (&curr_quit_handler_target, this); scoped_restore restore_quit = make_scoped_restore (&quit_handler, ::remote_serial_quit_handler); rs->got_ctrlc_during_io = 0; ch = serial_readchar (rs->remote_desc, timeout); if (rs->got_ctrlc_during_io) set_quit_flag (); } if (ch >= 0) return ch; switch ((enum serial_rc) ch) { case SERIAL_EOF: remote_unpush_target (this); throw_error (TARGET_CLOSE_ERROR, _("Remote connection closed")); /* no return */ case SERIAL_ERROR: unpush_and_perror (this, _("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. The current quit handler is overridden to avoid quitting in the middle of packet sequence, as that would break communication with the remote server. See remote_serial_quit_handler for more detail. */ void remote_target::remote_serial_write (const char *str, int len) { struct remote_state *rs = get_remote_state (); scoped_restore restore_quit_target = make_scoped_restore (&curr_quit_handler_target, this); scoped_restore restore_quit = make_scoped_restore (&quit_handler, ::remote_serial_quit_handler); rs->got_ctrlc_during_io = 0; if (serial_write (rs->remote_desc, str, len)) { unpush_and_perror (this, _("Remote communication error. " "Target disconnected.")); } if (rs->got_ctrlc_during_io) set_quit_flag (); } /* Return a string representing an escaped version of BUF, of len N. E.g. \n is converted to \\n, \t to \\t, etc. */ static std::string escape_buffer (const char *buf, int n) { string_file stb; stb.putstrn (buf, n, '\\'); return stb.release (); } int remote_target::putpkt (const char *buf) { return putpkt_binary (buf, strlen (buf)); } /* Wrapper around remote_target::putpkt to avoid exporting remote_target. */ int putpkt (remote_target *remote, const char *buf) { return remote->putpkt (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 $, # and checksum, and for a possible /0 if we are debugging (remote_debug) and want to print the sent packet as a string. */ int remote_target::putpkt_binary (const char *buf, int cnt) { struct remote_state *rs = get_remote_state (); int i; unsigned char csum = 0; gdb::def_vector data (cnt + 6); char *buf2 = data.data (); int ch; int tcount = 0; char *p; /* Catch cases like trying to read memory or listing threads while we're waiting for a stop reply. The remote server wouldn't be ready to handle this request, so we'd hang and timeout. We don't have to worry about this in synchronous mode, because in that case it's not possible to issue a command while the target is running. This is not a problem in non-stop mode, because in that case, the stub is always ready to process serial input. */ if (!target_is_non_stop_p () && target_is_async_p () && rs->waiting_for_stop_reply) { error (_("Cannot execute this command while the target is running.\n" "Use the \"interrupt\" command to stop the target\n" "and then try again.")); } /* Copy the packet into buffer BUF2, encapsulating it and giving it a checksum. */ p = buf2; *p++ = '$'; 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) { if (remote_debug) { *p = '\0'; int len = (int) (p - buf2); int max_chars; if (remote_packet_max_chars < 0) max_chars = len; else max_chars = remote_packet_max_chars; std::string str = escape_buffer (buf2, std::min (len, max_chars)); if (len > max_chars) remote_debug_printf_nofunc ("Sending packet: %s [%d bytes omitted]", str.c_str (), len - max_chars); else remote_debug_printf_nofunc ("Sending packet: %s", str.c_str ()); } 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); switch (ch) { case '+': remote_debug_printf_nofunc ("Received Ack"); return 1; case '-': remote_debug_printf_nofunc ("Received Nak"); /* FALLTHROUGH */ case SERIAL_TIMEOUT: tcount++; if (tcount > 3) return 0; break; /* Retransmit buffer. */ case '$': { remote_debug_printf ("Packet instead of Ack, ignoring it"); /* 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); if (val >= 0) { remote_debug_printf_nofunc (" Notification received: %s", escape_buffer (rs->buf.data (), val).c_str ()); handle_notification (rs->notif_state, rs->buf.data ()); /* We're in sync now, rewait for the ack. */ tcount = 0; } else remote_debug_printf_nofunc ("Junk: %c%s", ch & 0177, rs->buf.data ()); continue; } /* fall-through */ default: remote_debug_printf_nofunc ("Junk: %c%s", ch & 0177, rs->buf.data ()); 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. */ void remote_target::skip_frame () { 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. 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). */ long remote_target::read_frame (gdb::char_vector *buf_p) { unsigned char csum; long bc; int c; char *buf = buf_p->data (); struct remote_state *rs = get_remote_state (); csum = 0; bc = 0; while (1) { c = readchar (remote_timeout); switch (c) { case SERIAL_TIMEOUT: remote_debug_printf ("Timeout in mid-packet, retrying"); return -1; case '$': remote_debug_printf ("Saw new packet start in middle of old one"); 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) { remote_debug_printf ("Timeout in checksum, retrying"); return -1; } else if (check_0 < 0 || check_1 < 0) { remote_debug_printf ("Communication error in checksum"); 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; remote_debug_printf ("Bad checksum, sentsum=0x%x, csum=0x%x, buf=%s", pktcsum, csum, escape_buffer (buf, bc).c_str ()); /* 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 >= buf_p->size () - 1) { /* Make some more room in the buffer. */ buf_p->resize (buf_p->size () + repeat); buf = buf_p->data (); } memset (&buf[bc], buf[bc - 1], repeat); bc += repeat; continue; } buf[bc] = '\0'; gdb_printf (_("Invalid run length encoding: %s\n"), buf); return -1; } default: if (bc >= buf_p->size () - 1) { /* Make some more room in the buffer. */ buf_p->resize (buf_p->size () * 2); buf = buf_p->data (); } buf[bc++] = c; csum += c; continue; } } } /* Set this to the maximum number of seconds to wait instead of waiting forever in target_wait(). If this timer times out, then it generates an error and the command is aborted. This replaces most of the need for timeouts in the GDB test suite, and makes it possible to distinguish between a hung target and one with slow communications. */ static int watchdog = 0; static void show_watchdog (struct ui_file *file, int from_tty, struct cmd_list_element *c, const char *value) { gdb_printf (file, _("Watchdog timer is %s.\n"), value); } /* Read a packet from the remote machine, with error checking, and store it in *BUF. Resize *BUF if necessary to hold the result. 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 remote_target::getpkt (gdb::char_vector *buf, int forever) { getpkt_sane (buf, forever); } /* Read a packet from the remote machine, with error checking, and store it in *BUF. Resize *BUF if necessary to hold the result. 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). */ int remote_target::getpkt_or_notif_sane_1 (gdb::char_vector *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; strcpy (buf->data (), "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 != '$' && c != '%'); if (c == SERIAL_TIMEOUT) { if (expecting_notif) return -1; /* Don't complain, it's normal to not get anything in this case. */ if (forever) /* Watchdog went off? Kill the target. */ { remote_unpush_target (this); throw_error (TARGET_CLOSE_ERROR, _("Watchdog timeout has expired. " "Target detached.")); } remote_debug_printf ("Timed out."); } else { /* We've found the start of a packet or notification. Now collect the data. */ val = read_frame (buf); if (val >= 0) break; } remote_serial_write ("-", 1); } if (tries > MAX_TRIES) { /* We have tried hard enough, and just can't receive the packet/notification. Give up. */ gdb_printf (_("Ignoring packet error, continuing...\n")); /* Skip the ack char if we're in no-ack mode. */ if (!rs->noack_mode) remote_serial_write ("+", 1); return -1; } /* If we got an ordinary packet, return that to our caller. */ if (c == '$') { if (remote_debug) { int max_chars; if (remote_packet_max_chars < 0) max_chars = val; else max_chars = remote_packet_max_chars; std::string str = escape_buffer (buf->data (), std::min (val, max_chars)); if (val > max_chars) remote_debug_printf_nofunc ("Packet received: %s [%d bytes omitted]", str.c_str (), val - max_chars); else remote_debug_printf_nofunc ("Packet received: %s", str.c_str ()); } /* 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 == '%'); remote_debug_printf_nofunc (" Notification received: %s", escape_buffer (buf->data (), val).c_str ()); if (is_notif != NULL) *is_notif = 1; handle_notification (rs->notif_state, buf->data ()); /* Notifications require no acknowledgement. */ if (expecting_notif) return val; } } } int remote_target::getpkt_sane (gdb::char_vector *buf, int forever) { return getpkt_or_notif_sane_1 (buf, forever, 0, NULL); } int remote_target::getpkt_or_notif_sane (gdb::char_vector *buf, int forever, int *is_notif) { return getpkt_or_notif_sane_1 (buf, forever, 1, is_notif); } /* Kill any new fork children of inferior INF that haven't been processed by follow_fork. */ void remote_target::kill_new_fork_children (inferior *inf) { remote_state *rs = get_remote_state (); const notif_client *notif = ¬if_client_stop; /* Kill the fork child threads of any threads in inferior INF that are stopped at a fork event. */ for (thread_info *thread : inf->non_exited_threads ()) { const target_waitstatus *ws = thread_pending_fork_status (thread); if (ws == nullptr) continue; int child_pid = ws->child_ptid ().pid (); int res = remote_vkill (child_pid); if (res != 0) error (_("Can't kill fork child process %d"), child_pid); } /* Check for any pending fork events (not reported or processed yet) in inferior INF and kill those fork child threads as well. */ remote_notif_get_pending_events (notif); for (auto &event : rs->stop_reply_queue) { if (event->ptid.pid () != inf->pid) continue; if (!is_fork_status (event->ws.kind ())) continue; int child_pid = event->ws.child_ptid ().pid (); int res = remote_vkill (child_pid); if (res != 0) error (_("Can't kill fork child process %d"), child_pid); } } /* Target hook to kill the current inferior. */ void remote_target::kill () { int res = -1; inferior *inf = find_inferior_pid (this, inferior_ptid.pid ()); gdb_assert (inf != nullptr); if (m_features.packet_support (PACKET_vKill) != PACKET_DISABLE) { /* If we're stopped while forking and we haven't followed yet, kill the child task. We need to do this before killing the parent task because if this is a vfork then the parent will be sleeping. */ kill_new_fork_children (inf); res = remote_vkill (inf->pid); if (res == 0) { target_mourn_inferior (inferior_ptid); return; } } /* If we are in 'target remote' mode and we are killing the only inferior, then we will tell gdbserver to exit and unpush the target. */ if (res == -1 && !m_features.remote_multi_process_p () && number_of_live_inferiors (this) == 1) { remote_kill_k (); /* We've killed the remote end, we get to mourn it. If we are not in extended mode, mourning the inferior also unpushes remote_ops from the target stack, which closes the remote connection. */ target_mourn_inferior (inferior_ptid); return; } error (_("Can't kill process")); } /* Send a kill request to the target using the 'vKill' packet. */ int remote_target::remote_vkill (int pid) { if (m_features.packet_support (PACKET_vKill) == PACKET_DISABLE) return -1; remote_state *rs = get_remote_state (); /* Tell the remote target to detach. */ xsnprintf (rs->buf.data (), get_remote_packet_size (), "vKill;%x", pid); putpkt (rs->buf); getpkt (&rs->buf, 0); switch (m_features.packet_ok (rs->buf, PACKET_vKill)) { case PACKET_OK: return 0; case PACKET_ERROR: return 1; case PACKET_UNKNOWN: return -1; default: internal_error (_("Bad result from packet_ok")); } } /* Send a kill request to the target using the 'k' packet. */ void remote_target::remote_kill_k () { /* Catch errors so the user can quit from gdb even when we aren't on speaking terms with the remote system. */ try { putpkt ("k"); } catch (const gdb_exception_error &ex) { 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; } } void remote_target::mourn_inferior () { struct remote_state *rs = get_remote_state (); /* We're no longer interested in notification events of an inferior that exited or was killed/detached. */ discard_pending_stop_replies (current_inferior ()); /* In 'target remote' mode with one inferior, we close the connection. */ if (!rs->extended && number_of_live_inferiors (this) <= 1) { remote_unpush_target (this); return; } /* 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); /* Call common code to mark the inferior as not running. */ generic_mourn_inferior (); } bool extended_remote_target::supports_disable_randomization () { return (m_features.packet_support (PACKET_QDisableRandomization) == PACKET_ENABLE); } void remote_target::extended_remote_disable_randomization (int val) { struct remote_state *rs = get_remote_state (); char *reply; xsnprintf (rs->buf.data (), get_remote_packet_size (), "QDisableRandomization:%x", val); putpkt (rs->buf); reply = remote_get_noisy_reply (); if (*reply == '\0') error (_("Target does not support QDisableRandomization.")); if (strcmp (reply, "OK") != 0) error (_("Bogus QDisableRandomization reply from target: %s"), reply); } int remote_target::extended_remote_run (const std::string &args) { struct remote_state *rs = get_remote_state (); int len; const char *remote_exec_file = get_remote_exec_file (); /* If the user has disabled vRun support, or we have detected that support is not available, do not try it. */ if (m_features.packet_support (PACKET_vRun) == PACKET_DISABLE) return -1; strcpy (rs->buf.data (), "vRun;"); len = strlen (rs->buf.data ()); 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.data () + len, strlen (remote_exec_file)); if (!args.empty ()) { int i; gdb_argv argv (args.c_str ()); 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.data () + len, strlen (argv[i])); } } rs->buf[len++] = '\0'; putpkt (rs->buf); getpkt (&rs->buf, 0); switch (m_features.packet_ok (rs->buf, 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"); } } /* Helper function to send set/unset environment packets. ACTION is either "set" or "unset". PACKET is either "QEnvironmentHexEncoded" or "QEnvironmentUnsetVariable". VALUE is the variable to be sent. */ void remote_target::send_environment_packet (const char *action, const char *packet, const char *value) { remote_state *rs = get_remote_state (); /* Convert the environment variable to an hex string, which is the best format to be transmitted over the wire. */ std::string encoded_value = bin2hex ((const gdb_byte *) value, strlen (value)); xsnprintf (rs->buf.data (), get_remote_packet_size (), "%s:%s", packet, encoded_value.c_str ()); putpkt (rs->buf); getpkt (&rs->buf, 0); if (strcmp (rs->buf.data (), "OK") != 0) warning (_("Unable to %s environment variable '%s' on remote."), action, value); } /* Helper function to handle the QEnvironment* packets. */ void remote_target::extended_remote_environment_support () { remote_state *rs = get_remote_state (); if (m_features.packet_support (PACKET_QEnvironmentReset) != PACKET_DISABLE) { putpkt ("QEnvironmentReset"); getpkt (&rs->buf, 0); if (strcmp (rs->buf.data (), "OK") != 0) warning (_("Unable to reset environment on remote.")); } gdb_environ *e = ¤t_inferior ()->environment; if (m_features.packet_support (PACKET_QEnvironmentHexEncoded) != PACKET_DISABLE) { for (const std::string &el : e->user_set_env ()) send_environment_packet ("set", "QEnvironmentHexEncoded", el.c_str ()); } if (m_features.packet_support (PACKET_QEnvironmentUnset) != PACKET_DISABLE) for (const std::string &el : e->user_unset_env ()) send_environment_packet ("unset", "QEnvironmentUnset", el.c_str ()); } /* Helper function to set the current working directory for the inferior in the remote target. */ void remote_target::extended_remote_set_inferior_cwd () { if (m_features.packet_support (PACKET_QSetWorkingDir) != PACKET_DISABLE) { const std::string &inferior_cwd = current_inferior ()->cwd (); remote_state *rs = get_remote_state (); if (!inferior_cwd.empty ()) { std::string hexpath = bin2hex ((const gdb_byte *) inferior_cwd.data (), inferior_cwd.size ()); xsnprintf (rs->buf.data (), get_remote_packet_size (), "QSetWorkingDir:%s", hexpath.c_str ()); } else { /* An empty inferior_cwd means that the user wants us to reset the remote server's inferior's cwd. */ xsnprintf (rs->buf.data (), get_remote_packet_size (), "QSetWorkingDir:"); } putpkt (rs->buf); getpkt (&rs->buf, 0); if (m_features.packet_ok (rs->buf, PACKET_QSetWorkingDir) != PACKET_OK) error (_("\ Remote replied unexpectedly while setting the inferior's working\n\ directory: %s"), rs->buf.data ()); } } /* 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. */ void extended_remote_target::create_inferior (const char *exec_file, const std::string &args, char **env, int from_tty) { int run_worked; char *stop_reply; struct remote_state *rs = get_remote_state (); const char *remote_exec_file = get_remote_exec_file (); /* If running asynchronously, register the target file descriptor with the event loop. */ if (target_can_async_p ()) target_async (true); /* Disable address space randomization if requested (and supported). */ if (supports_disable_randomization ()) extended_remote_disable_randomization (disable_randomization); /* If startup-with-shell is on, we inform gdbserver to start the remote inferior using a shell. */ if (m_features.packet_support (PACKET_QStartupWithShell) != PACKET_DISABLE) { xsnprintf (rs->buf.data (), get_remote_packet_size (), "QStartupWithShell:%d", startup_with_shell ? 1 : 0); putpkt (rs->buf); getpkt (&rs->buf, 0); if (strcmp (rs->buf.data (), "OK") != 0) error (_("\ Remote replied unexpectedly while setting startup-with-shell: %s"), rs->buf.data ()); } extended_remote_environment_support (); extended_remote_set_inferior_cwd (); /* 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.empty ()) error (_("Remote target does not support \"set args\" or run ARGS")); /* Fall back to "R". */ extended_remote_restart (); } /* vRun's success return is a stop reply. */ stop_reply = run_worked ? rs->buf.data () : 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) { if (bp_tgt->conditions.empty ()) return 0; buf += strlen (buf); xsnprintf (buf, buf_end - buf, "%s", ";"); buf++; /* Send conditions to the target. */ for (agent_expr *aexpr : bp_tgt->conditions) { xsnprintf (buf, buf_end - buf, "X%x,", aexpr->len); buf += strlen (buf); for (int 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) { if (bp_tgt->tcommands.empty ()) 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 (agent_expr *aexpr : bp_tgt->tcommands) { sprintf (buf, "X%x,", aexpr->len); buf += strlen (buf); for (int 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. */ int remote_target::insert_breakpoint (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 (m_features.packet_support (PACKET_Z0) != PACKET_DISABLE) { CORE_ADDR addr = bp_tgt->reqstd_address; struct remote_state *rs; char *p, *endbuf; /* 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.data (); endbuf = p + get_remote_packet_size (); *(p++) = 'Z'; *(p++) = '0'; *(p++) = ','; addr = (ULONGEST) remote_address_masked (addr); p += hexnumstr (p, addr); xsnprintf (p, endbuf - p, ",%d", bp_tgt->kind); if (supports_evaluation_of_breakpoint_conditions ()) remote_add_target_side_condition (gdbarch, bp_tgt, p, endbuf); if (can_run_breakpoint_commands ()) remote_add_target_side_commands (gdbarch, bp_tgt, p); putpkt (rs->buf); getpkt (&rs->buf, 0); switch (m_features.packet_ok (rs->buf, PACKET_Z0)) { case PACKET_ERROR: return -1; case PACKET_OK: return 0; case PACKET_UNKNOWN: break; } } /* If this breakpoint has target-side commands but this stub doesn't support Z0 packets, throw error. */ if (!bp_tgt->tcommands.empty ()) throw_error (NOT_SUPPORTED_ERROR, _("\ Target doesn't support breakpoints that have target side commands.")); return memory_insert_breakpoint (this, gdbarch, bp_tgt); } int remote_target::remove_breakpoint (struct gdbarch *gdbarch, struct bp_target_info *bp_tgt, enum remove_bp_reason reason) { CORE_ADDR addr = bp_tgt->placed_address; struct remote_state *rs = get_remote_state (); if (m_features.packet_support (PACKET_Z0) != PACKET_DISABLE) { char *p = rs->buf.data (); char *endbuf = p + 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->kind); putpkt (rs->buf); getpkt (&rs->buf, 0); return (rs->buf[0] == 'E'); } return memory_remove_breakpoint (this, gdbarch, bp_tgt, reason); } static enum Z_packet_type 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 (_("hw_bp_to_z: bad watchpoint type %d"), type); } } int remote_target::insert_watchpoint (CORE_ADDR addr, int len, enum target_hw_bp_type type, struct expression *cond) { struct remote_state *rs = get_remote_state (); char *endbuf = rs->buf.data () + get_remote_packet_size (); char *p; enum Z_packet_type packet = watchpoint_to_Z_packet (type); if (m_features.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.data (), endbuf - rs->buf.data (), "Z%x,", packet); p = strchr (rs->buf.data (), '\0'); addr = remote_address_masked (addr); p += hexnumstr (p, (ULONGEST) addr); xsnprintf (p, endbuf - p, ",%x", len); putpkt (rs->buf); getpkt (&rs->buf, 0); switch (m_features.packet_ok (rs->buf, PACKET_Z0 + packet)) { case PACKET_ERROR: return -1; case PACKET_UNKNOWN: return 1; case PACKET_OK: return 0; } internal_error (_("remote_insert_watchpoint: reached end of function")); } bool remote_target::watchpoint_addr_within_range (CORE_ADDR addr, CORE_ADDR start, int length) { CORE_ADDR diff = remote_address_masked (addr - start); return diff < length; } int remote_target::remove_watchpoint (CORE_ADDR addr, int len, enum target_hw_bp_type type, struct expression *cond) { struct remote_state *rs = get_remote_state (); char *endbuf = rs->buf.data () + get_remote_packet_size (); char *p; enum Z_packet_type packet = watchpoint_to_Z_packet (type); if (m_features.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.data (), endbuf - rs->buf.data (), "z%x,", packet); p = strchr (rs->buf.data (), '\0'); addr = remote_address_masked (addr); p += hexnumstr (p, (ULONGEST) addr); xsnprintf (p, endbuf - p, ",%x", len); putpkt (rs->buf); getpkt (&rs->buf, 0); switch (m_features.packet_ok (rs->buf, PACKET_Z0 + packet)) { case PACKET_ERROR: case PACKET_UNKNOWN: return -1; case PACKET_OK: return 0; } internal_error (_("remote_remove_watchpoint: reached end of function")); } static int remote_hw_watchpoint_limit = -1; static int remote_hw_watchpoint_length_limit = -1; static int remote_hw_breakpoint_limit = -1; int remote_target::region_ok_for_hw_watchpoint (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; } int remote_target::can_use_hw_breakpoint (enum bptype 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; } /* The to_stopped_by_sw_breakpoint method of target remote. */ bool remote_target::stopped_by_sw_breakpoint () { struct thread_info *thread = inferior_thread (); return (thread->priv != NULL && (get_remote_thread_info (thread)->stop_reason == TARGET_STOPPED_BY_SW_BREAKPOINT)); } /* The to_supports_stopped_by_sw_breakpoint method of target remote. */ bool remote_target::supports_stopped_by_sw_breakpoint () { return (m_features.packet_support (PACKET_swbreak_feature) == PACKET_ENABLE); } /* The to_stopped_by_hw_breakpoint method of target remote. */ bool remote_target::stopped_by_hw_breakpoint () { struct thread_info *thread = inferior_thread (); return (thread->priv != NULL && (get_remote_thread_info (thread)->stop_reason == TARGET_STOPPED_BY_HW_BREAKPOINT)); } /* The to_supports_stopped_by_hw_breakpoint method of target remote. */ bool remote_target::supports_stopped_by_hw_breakpoint () { return (m_features.packet_support (PACKET_hwbreak_feature) == PACKET_ENABLE); } bool remote_target::stopped_by_watchpoint () { struct thread_info *thread = inferior_thread (); return (thread->priv != NULL && (get_remote_thread_info (thread)->stop_reason == TARGET_STOPPED_BY_WATCHPOINT)); } bool remote_target::stopped_data_address (CORE_ADDR *addr_p) { struct thread_info *thread = inferior_thread (); if (thread->priv != NULL && (get_remote_thread_info (thread)->stop_reason == TARGET_STOPPED_BY_WATCHPOINT)) { *addr_p = get_remote_thread_info (thread)->watch_data_address; return true; } return false; } int remote_target::insert_hw_breakpoint (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; if (m_features.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.data (); endbuf = p + get_remote_packet_size (); *(p++) = 'Z'; *(p++) = '1'; *(p++) = ','; addr = remote_address_masked (addr); p += hexnumstr (p, (ULONGEST) addr); xsnprintf (p, endbuf - p, ",%x", bp_tgt->kind); if (supports_evaluation_of_breakpoint_conditions ()) remote_add_target_side_condition (gdbarch, bp_tgt, p, endbuf); if (can_run_breakpoint_commands ()) remote_add_target_side_commands (gdbarch, bp_tgt, p); putpkt (rs->buf); getpkt (&rs->buf, 0); switch (m_features.packet_ok (rs->buf, 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: return 0; } internal_error (_("remote_insert_hw_breakpoint: reached end of function")); } int remote_target::remove_hw_breakpoint (struct gdbarch *gdbarch, struct bp_target_info *bp_tgt) { CORE_ADDR addr; struct remote_state *rs = get_remote_state (); char *p = rs->buf.data (); char *endbuf = p + get_remote_packet_size (); if (m_features.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->kind); putpkt (rs->buf); getpkt (&rs->buf, 0); switch (m_features.packet_ok (rs->buf, PACKET_Z1)) { case PACKET_ERROR: case PACKET_UNKNOWN: return -1; case PACKET_OK: return 0; } internal_error (_("remote_remove_hw_breakpoint: reached end of function")); } /* Verify memory using the "qCRC:" request. */ int remote_target::verify_memory (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 () && m_features.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.data (), 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, 0); result = m_features.packet_ok (rs->buf, 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 (this, 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 (const char *args, int from_tty) { asection *s; const char *sectname; bfd_size_type size; bfd_vma lma; int matched = 0; int mismatched = 0; int res; int read_only = 0; if (!current_program_space->exec_bfd ()) error (_("command cannot be used without an exec file")); if (args != NULL && strcmp (args, "-r") == 0) { read_only = 1; args = NULL; } for (s = current_program_space->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_section_size (s); if (size == 0) continue; /* Skip zero-length section. */ sectname = bfd_section_name (s); if (args && strcmp (args, sectname) != 0) continue; /* Not the section selected by user. */ matched = 1; /* Do this section. */ lma = s->lma; gdb::byte_vector sectdata (size); bfd_get_section_contents (current_program_space->exec_bfd (), s, sectdata.data (), 0, size); res = target_verify_memory (sectdata.data (), lma, size); if (res == -1) error (_("target memory fault, section %s, range %s -- %s"), sectname, paddress (target_gdbarch (), lma), paddress (target_gdbarch (), lma + size)); gdb_printf ("Section %s, range %s -- %s: ", sectname, paddress (target_gdbarch (), lma), paddress (target_gdbarch (), lma + size)); if (res) gdb_printf ("matched.\n"); else { gdb_printf ("MIS-MATCHED!\n"); mismatched++; } } if (mismatched > 0) warning (_("One or more sections of the target image does " "not match the loaded file")); if (args && !matched) gdb_printf (_("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. */ target_xfer_status remote_target::remote_write_qxfer (const char *object_name, const char *annex, const gdb_byte *writebuf, ULONGEST offset, LONGEST len, ULONGEST *xfered_len, const unsigned int which_packet) { int i, buf_len; ULONGEST n; struct remote_state *rs = get_remote_state (); int max_size = get_memory_write_packet_size (); if (m_features.packet_support (which_packet) == PACKET_DISABLE) return TARGET_XFER_E_IO; /* Insert header. */ i = snprintf (rs->buf.data (), 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, 1, (gdb_byte *) rs->buf.data () + i, &max_size, max_size); if (putpkt_binary (rs->buf.data (), i + buf_len) < 0 || getpkt_sane (&rs->buf, 0) < 0 || m_features.packet_ok (rs->buf, which_packet) != PACKET_OK) return TARGET_XFER_E_IO; unpack_varlen_hex (rs->buf.data (), &n); *xfered_len = n; return (*xfered_len != 0) ? TARGET_XFER_OK : TARGET_XFER_EOF; } /* 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. */ target_xfer_status remote_target::remote_read_qxfer (const char *object_name, const char *annex, gdb_byte *readbuf, ULONGEST offset, LONGEST len, ULONGEST *xfered_len, const unsigned int which_packet) { struct remote_state *rs = get_remote_state (); LONGEST i, n, packet_len; if (m_features.packet_support (which_packet) == 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 = std::min (get_remote_packet_size () - 5, len); snprintf (rs->buf.data (), 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, 0); if (packet_len < 0 || m_features.packet_ok (rs->buf, which_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.data ()); /* '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.data () + 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; } } enum target_xfer_status remote_target::xfer_partial (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; int unit_size = gdbarch_addressable_memory_unit_size (target_gdbarch ()); 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, unit_size, xfered_len); else return remote_read_bytes (offset, readbuf, len, unit_size, xfered_len); } /* Handle extra signal info using qxfer packets. */ if (object == TARGET_OBJECT_SIGNAL_INFO) { if (readbuf) return remote_read_qxfer ("siginfo", annex, readbuf, offset, len, xfered_len, PACKET_qXfer_siginfo_read); else return remote_write_qxfer ("siginfo", annex, writebuf, offset, len, xfered_len, PACKET_qXfer_siginfo_write); } if (object == TARGET_OBJECT_STATIC_TRACE_DATA) { if (readbuf) return remote_read_qxfer ("statictrace", annex, readbuf, offset, len, xfered_len, PACKET_qXfer_statictrace_read); else return TARGET_XFER_E_IO; } /* Only handle flash writes. */ if (writebuf != NULL) { switch (object) { case TARGET_OBJECT_FLASH: return remote_flash_write (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 ("auxv", annex, readbuf, offset, len, xfered_len, PACKET_qXfer_auxv); case TARGET_OBJECT_AVAILABLE_FEATURES: return remote_read_qxfer ("features", annex, readbuf, offset, len, xfered_len, PACKET_qXfer_features); case TARGET_OBJECT_LIBRARIES: return remote_read_qxfer ("libraries", annex, readbuf, offset, len, xfered_len, PACKET_qXfer_libraries); case TARGET_OBJECT_LIBRARIES_SVR4: return remote_read_qxfer ("libraries-svr4", annex, readbuf, offset, len, xfered_len, PACKET_qXfer_libraries_svr4); case TARGET_OBJECT_MEMORY_MAP: gdb_assert (annex == NULL); return remote_read_qxfer ("memory-map", annex, readbuf, offset, len, xfered_len, 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 ("osdata", annex, readbuf, offset, len, xfered_len, PACKET_qXfer_osdata); case TARGET_OBJECT_THREADS: gdb_assert (annex == NULL); return remote_read_qxfer ("threads", annex, readbuf, offset, len, xfered_len, PACKET_qXfer_threads); case TARGET_OBJECT_TRACEFRAME_INFO: gdb_assert (annex == NULL); return remote_read_qxfer ("traceframe-info", annex, readbuf, offset, len, xfered_len, PACKET_qXfer_traceframe_info); case TARGET_OBJECT_FDPIC: return remote_read_qxfer ("fdpic", annex, readbuf, offset, len, xfered_len, PACKET_qXfer_fdpic); case TARGET_OBJECT_OPENVMS_UIB: return remote_read_qxfer ("uib", annex, readbuf, offset, len, xfered_len, PACKET_qXfer_uib); case TARGET_OBJECT_BTRACE: return remote_read_qxfer ("btrace", annex, readbuf, offset, len, xfered_len, PACKET_qXfer_btrace); case TARGET_OBJECT_BTRACE_CONF: return remote_read_qxfer ("btrace-conf", annex, readbuf, offset, len, xfered_len, PACKET_qXfer_btrace_conf); case TARGET_OBJECT_EXEC_FILE: return remote_read_qxfer ("exec-file", annex, readbuf, offset, len, xfered_len, PACKET_qXfer_exec_file); 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.data (); *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, 0); strcpy ((char *) readbuf, rs->buf.data ()); *xfered_len = strlen ((char *) readbuf); return (*xfered_len != 0) ? TARGET_XFER_OK : TARGET_XFER_EOF; } /* Implementation of to_get_memory_xfer_limit. */ ULONGEST remote_target::get_memory_xfer_limit () { return get_memory_write_packet_size (); } int remote_target::search_memory (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 (); /* 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; auto read_memory = [=] (CORE_ADDR addr, gdb_byte *result, size_t len) { return (target_read (this, TARGET_OBJECT_MEMORY, NULL, result, addr, len) == len); }; /* 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 (m_features.packet_support (PACKET_qSearch_memory) == 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 (read_memory, 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.data (), 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, 1, (gdb_byte *) rs->buf.data () + 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.data (), i + escaped_pattern_len) < 0 || getpkt_sane (&rs->buf, 0) < 0 || m_features.packet_ok (rs->buf, PACKET_qSearch_memory) != PACKET_OK) { /* The request may not have worked because the command is not supported. If so, fall back to the simple way. */ if (m_features.packet_support (PACKET_qSearch_memory) == PACKET_DISABLE) { return simple_search_memory (read_memory, 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.data ()); unpack_varlen_hex (&rs->buf[2], &found_addr); *found_addrp = found_addr; } else error (_("Unknown qSearch:memory reply: %s"), rs->buf.data ()); return found; } void remote_target::rcmd (const char *command, struct ui_file *outbuf) { struct remote_state *rs = get_remote_state (); char *p = rs->buf.data (); 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.data (), "qRcmd,"); p = strchr (rs->buf.data (), '\0'); if ((strlen (rs->buf.data ()) + 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, 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.data (); 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' && isxdigit (buf[1]) && isxdigit (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]); gdb_putc (c, outbuf); } break; } } std::vector remote_target::memory_map () { std::vector result; gdb::optional text = target_read_stralloc (current_inferior ()->top_target (), TARGET_OBJECT_MEMORY_MAP, NULL); if (text) result = parse_memory_map (text->data ()); return result; } /* Set of callbacks used to implement the 'maint packet' command. */ struct cli_packet_command_callbacks : public send_remote_packet_callbacks { /* Called before the packet is sent. BUF is the packet content before the protocol specific prefix, suffix, and escaping is added. */ void sending (gdb::array_view &buf) override { gdb_puts ("sending: "); print_packet (buf); gdb_puts ("\n"); } /* Called with BUF, the reply from the remote target. */ void received (gdb::array_view &buf) override { gdb_puts ("received: \""); print_packet (buf); gdb_puts ("\"\n"); } private: /* Print BUF o gdb_stdout. Any non-printable bytes in BUF are printed as '\x??' with '??' replaced by the hexadecimal value of the byte. */ static void print_packet (gdb::array_view &buf) { string_file stb; for (int i = 0; i < buf.size (); ++i) { gdb_byte c = buf[i]; if (isprint (c)) gdb_putc (c, &stb); else gdb_printf (&stb, "\\x%02x", (unsigned char) c); } gdb_puts (stb.string ().c_str ()); } }; /* See remote.h. */ void send_remote_packet (gdb::array_view &buf, send_remote_packet_callbacks *callbacks) { if (buf.size () == 0 || buf.data ()[0] == '\0') error (_("a remote packet must not be empty")); remote_target *remote = get_current_remote_target (); if (remote == nullptr) error (_("packets can only be sent to a remote target")); callbacks->sending (buf); remote->putpkt_binary (buf.data (), buf.size ()); remote_state *rs = remote->get_remote_state (); int bytes = remote->getpkt_sane (&rs->buf, 0); if (bytes < 0) error (_("error while fetching packet from remote target")); gdb::array_view view (&rs->buf[0], bytes); callbacks->received (view); } /* Entry point for the 'maint packet' command. */ static void cli_packet_command (const char *args, int from_tty) { cli_packet_command_callbacks cb; gdb::array_view view = gdb::make_array_view (args, args == nullptr ? 0 : strlen (args)); send_remote_packet (view, &cb); } #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 (const char *cmd, int tty) { int sample_thread = SAMPLE_THREAD; gdb_printf (_("Remote threadset test\n")); set_general_thread (sample_thread); } static void threadalive_test (const char *cmd, int tty) { int sample_thread = SAMPLE_THREAD; int pid = inferior_ptid.pid (); ptid_t ptid = ptid_t (pid, sample_thread, 0); if (remote_thread_alive (ptid)) gdb_printf ("PASS: Thread alive test\n"); else gdb_printf ("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 thread id into hex. */ hexid[16] = 0; gdb_printf ("%s %s\n", title, (&hexid[0])); } static void threadlist_test_cmd (const char *cmd, int tty) { int startflag = 1; threadref nextthread; int done, result_count; threadref threadlist[3]; gdb_printf ("Remote Threadlist test\n"); if (!remote_get_threadlist (startflag, &nextthread, 3, &done, &result_count, &threadlist[0])) gdb_printf ("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); gdb_printf ("Name: %s\n ", info->shortname); gdb_printf ("State: %s\n", info->display); gdb_printf ("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 (const char *cmd, int tty) { int athread = SAMPLE_THREAD; threadref thread; int set; int_to_threadref (&thread, athread); gdb_printf ("Remote Threadinfo test\n"); if (!get_and_display_threadinfo (&thread)) gdb_printf ("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 (const char *cmd, int tty) { gdb_printf ("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. */ std::string remote_target::pid_to_str (ptid_t ptid) { if (ptid == null_ptid) return normal_pid_to_str (ptid); else if (ptid.is_pid ()) { /* 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 (!m_features.remote_multi_process_p ()) return "Remote target"; return normal_pid_to_str (ptid); } else { if (magic_null_ptid == ptid) return "Thread
"; else if (m_features.remote_multi_process_p ()) if (ptid.lwp () == 0) return normal_pid_to_str (ptid); else return string_printf ("Thread %d.%ld", ptid.pid (), ptid.lwp ()); else return string_printf ("Thread %ld", ptid.lwp ()); } } /* Get the address of the thread local variable in OBJFILE which is stored at OFFSET within the thread local storage for thread PTID. */ CORE_ADDR remote_target::get_thread_local_address (ptid_t ptid, CORE_ADDR lm, CORE_ADDR offset) { if (m_features.packet_support (PACKET_qGetTLSAddr) != PACKET_DISABLE) { struct remote_state *rs = get_remote_state (); char *p = rs->buf.data (); char *endp = p + 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, 0); result = m_features.packet_ok (rs->buf, PACKET_qGetTLSAddr); if (result == PACKET_OK) { ULONGEST addr; unpack_varlen_hex (rs->buf.data (), &addr); return addr; } 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. */ bool remote_target::get_tib_address (ptid_t ptid, CORE_ADDR *addr) { if (m_features.packet_support (PACKET_qGetTIBAddr) != PACKET_DISABLE) { struct remote_state *rs = get_remote_state (); char *p = rs->buf.data (); char *endp = p + 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, 0); result = m_features.packet_ok (rs->buf, PACKET_qGetTIBAddr); if (result == PACKET_OK) { ULONGEST val; unpack_varlen_hex (rs->buf.data (), &val); if (addr) *addr = (CORE_ADDR) val; return true; } 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 false; } /* 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 { remote_g_packet_guess (int bytes_, const struct target_desc *tdesc_) : bytes (bytes_), tdesc (tdesc_) { } int bytes; const struct target_desc *tdesc; }; struct remote_g_packet_data { std::vector guesses; }; static const registry::key remote_g_packet_data_handle; static struct remote_g_packet_data * get_g_packet_data (struct gdbarch *gdbarch) { struct remote_g_packet_data *data = remote_g_packet_data_handle.get (gdbarch); if (data == nullptr) data = remote_g_packet_data_handle.emplace (gdbarch); return data; } void register_remote_g_packet_guess (struct gdbarch *gdbarch, int bytes, const struct target_desc *tdesc) { struct remote_g_packet_data *data = get_g_packet_data (gdbarch); gdb_assert (tdesc != NULL); for (const remote_g_packet_guess &guess : data->guesses) if (guess.bytes == bytes) internal_error (_("Duplicate g packet description added for size %d"), bytes); data->guesses.emplace_back (bytes, tdesc); } /* Return true if remote_read_description would do anything on this target and architecture, false otherwise. */ static bool remote_read_description_p (struct target_ops *target) { struct remote_g_packet_data *data = get_g_packet_data (target_gdbarch ()); return !data->guesses.empty (); } const struct target_desc * remote_target::read_description () { struct remote_g_packet_data *data = get_g_packet_data (target_gdbarch ()); /* Do not try this during initial connection, when we do not know whether there is a running but stopped thread. */ if (!target_has_execution () || inferior_ptid == null_ptid) return beneath ()->read_description (); if (!data->guesses.empty ()) { int bytes = send_g_packet (); for (const remote_g_packet_guess &guess : data->guesses) 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 beneath ()->read_description (); } /* 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, const 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 FILEIO_SUCCESS 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 (const char *buffer, int *retcode, fileio_error *remote_errno, const char **attachment) { char *p, *p2; *remote_errno = FILEIO_SUCCESS; *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 = (fileio_error) 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. */ int remote_target::remote_hostio_send_command (int command_bytes, int which_packet, fileio_error *remote_errno, const char **attachment, int *attachment_len) { struct remote_state *rs = get_remote_state (); int ret, bytes_read; const char *attachment_tmp; if (m_features.packet_support (which_packet) == PACKET_DISABLE) { *remote_errno = FILEIO_ENOSYS; return -1; } putpkt_binary (rs->buf.data (), command_bytes); bytes_read = getpkt_sane (&rs->buf, 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 (m_features.packet_ok (rs->buf, 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.data (), &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.data ()); } return ret; } /* See declaration.h. */ void readahead_cache::invalidate () { this->fd = -1; } /* See declaration.h. */ void readahead_cache::invalidate_fd (int fd) { if (this->fd == fd) this->fd = -1; } /* Set the filesystem remote_hostio functions that take FILENAME arguments will use. Return 0 on success, or -1 if an error occurs (and set *REMOTE_ERRNO). */ int remote_target::remote_hostio_set_filesystem (struct inferior *inf, fileio_error *remote_errno) { struct remote_state *rs = get_remote_state (); int required_pid = (inf == NULL || inf->fake_pid_p) ? 0 : inf->pid; char *p = rs->buf.data (); int left = get_remote_packet_size () - 1; char arg[9]; int ret; if (m_features.packet_support (PACKET_vFile_setfs) == PACKET_DISABLE) return 0; if (rs->fs_pid != -1 && required_pid == rs->fs_pid) return 0; remote_buffer_add_string (&p, &left, "vFile:setfs:"); xsnprintf (arg, sizeof (arg), "%x", required_pid); remote_buffer_add_string (&p, &left, arg); ret = remote_hostio_send_command (p - rs->buf.data (), PACKET_vFile_setfs, remote_errno, NULL, NULL); if (m_features.packet_support (PACKET_vFile_setfs) == PACKET_DISABLE) return 0; if (ret == 0) rs->fs_pid = required_pid; return ret; } /* Implementation of to_fileio_open. */ int remote_target::remote_hostio_open (inferior *inf, const char *filename, int flags, int mode, int warn_if_slow, fileio_error *remote_errno) { struct remote_state *rs = get_remote_state (); char *p = rs->buf.data (); int left = get_remote_packet_size () - 1; if (warn_if_slow) { static int warning_issued = 0; gdb_printf (_("Reading %s from remote target...\n"), filename); if (!warning_issued) { warning (_("File transfers from remote targets can be slow." " Use \"set sysroot\" to access files locally" " instead.")); warning_issued = 1; } } if (remote_hostio_set_filesystem (inf, remote_errno) != 0) return -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.data (), PACKET_vFile_open, remote_errno, NULL, NULL); } int remote_target::fileio_open (struct inferior *inf, const char *filename, int flags, int mode, int warn_if_slow, fileio_error *remote_errno) { return remote_hostio_open (inf, filename, flags, mode, warn_if_slow, remote_errno); } /* Implementation of to_fileio_pwrite. */ int remote_target::remote_hostio_pwrite (int fd, const gdb_byte *write_buf, int len, ULONGEST offset, fileio_error *remote_errno) { struct remote_state *rs = get_remote_state (); char *p = rs->buf.data (); int left = get_remote_packet_size (); int out_len; rs->readahead_cache.invalidate_fd (fd); 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, 1, (gdb_byte *) p, &out_len, (get_remote_packet_size () - (p - rs->buf.data ()))); return remote_hostio_send_command (p - rs->buf.data (), PACKET_vFile_pwrite, remote_errno, NULL, NULL); } int remote_target::fileio_pwrite (int fd, const gdb_byte *write_buf, int len, ULONGEST offset, fileio_error *remote_errno) { return remote_hostio_pwrite (fd, write_buf, len, offset, remote_errno); } /* Helper for the implementation of to_fileio_pread. Read the file from the remote side with vFile:pread. */ int remote_target::remote_hostio_pread_vFile (int fd, gdb_byte *read_buf, int len, ULONGEST offset, fileio_error *remote_errno) { struct remote_state *rs = get_remote_state (); char *p = rs->buf.data (); const 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.data (), 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; } /* See declaration.h. */ int readahead_cache::pread (int fd, gdb_byte *read_buf, size_t len, ULONGEST offset) { if (this->fd == fd && this->offset <= offset && offset < this->offset + this->bufsize) { ULONGEST max = this->offset + this->bufsize; if (offset + len > max) len = max - offset; memcpy (read_buf, this->buf + offset - this->offset, len); return len; } return 0; } /* Implementation of to_fileio_pread. */ int remote_target::remote_hostio_pread (int fd, gdb_byte *read_buf, int len, ULONGEST offset, fileio_error *remote_errno) { int ret; struct remote_state *rs = get_remote_state (); readahead_cache *cache = &rs->readahead_cache; ret = cache->pread (fd, read_buf, len, offset); if (ret > 0) { cache->hit_count++; remote_debug_printf ("readahead cache hit %s", pulongest (cache->hit_count)); return ret; } cache->miss_count++; remote_debug_printf ("readahead cache miss %s", pulongest (cache->miss_count)); cache->fd = fd; cache->offset = offset; cache->bufsize = get_remote_packet_size (); cache->buf = (gdb_byte *) xrealloc (cache->buf, cache->bufsize); ret = remote_hostio_pread_vFile (cache->fd, cache->buf, cache->bufsize, cache->offset, remote_errno); if (ret <= 0) { cache->invalidate_fd (fd); return ret; } cache->bufsize = ret; return cache->pread (fd, read_buf, len, offset); } int remote_target::fileio_pread (int fd, gdb_byte *read_buf, int len, ULONGEST offset, fileio_error *remote_errno) { return remote_hostio_pread (fd, read_buf, len, offset, remote_errno); } /* Implementation of to_fileio_close. */ int remote_target::remote_hostio_close (int fd, fileio_error *remote_errno) { struct remote_state *rs = get_remote_state (); char *p = rs->buf.data (); int left = get_remote_packet_size () - 1; rs->readahead_cache.invalidate_fd (fd); remote_buffer_add_string (&p, &left, "vFile:close:"); remote_buffer_add_int (&p, &left, fd); return remote_hostio_send_command (p - rs->buf.data (), PACKET_vFile_close, remote_errno, NULL, NULL); } int remote_target::fileio_close (int fd, fileio_error *remote_errno) { return remote_hostio_close (fd, remote_errno); } /* Implementation of to_fileio_unlink. */ int remote_target::remote_hostio_unlink (inferior *inf, const char *filename, fileio_error *remote_errno) { struct remote_state *rs = get_remote_state (); char *p = rs->buf.data (); int left = get_remote_packet_size () - 1; if (remote_hostio_set_filesystem (inf, remote_errno) != 0) return -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.data (), PACKET_vFile_unlink, remote_errno, NULL, NULL); } int remote_target::fileio_unlink (struct inferior *inf, const char *filename, fileio_error *remote_errno) { return remote_hostio_unlink (inf, filename, remote_errno); } /* Implementation of to_fileio_readlink. */ gdb::optional remote_target::fileio_readlink (struct inferior *inf, const char *filename, fileio_error *remote_errno) { struct remote_state *rs = get_remote_state (); char *p = rs->buf.data (); const char *attachment; int left = get_remote_packet_size (); int len, attachment_len; int read_len; if (remote_hostio_set_filesystem (inf, remote_errno) != 0) return {}; 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.data (), PACKET_vFile_readlink, remote_errno, &attachment, &attachment_len); if (len < 0) return {}; std::string ret (len, '\0'); read_len = remote_unescape_input ((gdb_byte *) attachment, attachment_len, (gdb_byte *) &ret[0], len); if (read_len != len) error (_("Readlink returned %d, but %d bytes."), len, read_len); return ret; } /* Implementation of to_fileio_fstat. */ int remote_target::fileio_fstat (int fd, struct stat *st, fileio_error *remote_errno) { struct remote_state *rs = get_remote_state (); char *p = rs->buf.data (); int left = get_remote_packet_size (); int attachment_len, ret; const char *attachment; struct fio_stat fst; int read_len; remote_buffer_add_string (&p, &left, "vFile:fstat:"); remote_buffer_add_int (&p, &left, fd); ret = remote_hostio_send_command (p - rs->buf.data (), PACKET_vFile_fstat, remote_errno, &attachment, &attachment_len); if (ret < 0) { if (*remote_errno != FILEIO_ENOSYS) return ret; /* Strictly we should return -1, ENOSYS here, but when "set sysroot remote:" was implemented in August 2008 BFD's need for a stat function was sidestepped with this hack. This was not remedied until March 2015 so we retain the previous behavior to avoid breaking compatibility. Note that the memset is a March 2015 addition; older GDBs set st_size *and nothing else* so the structure would have garbage in all other fields. This might break something but retaining the previous behavior here would be just too wrong. */ memset (st, 0, sizeof (struct stat)); st->st_size = INT_MAX; return 0; } read_len = remote_unescape_input ((gdb_byte *) attachment, attachment_len, (gdb_byte *) &fst, sizeof (fst)); if (read_len != ret) error (_("vFile:fstat returned %d, but %d bytes."), ret, read_len); if (read_len != sizeof (fst)) error (_("vFile:fstat returned %d bytes, but expecting %d."), read_len, (int) sizeof (fst)); remote_fileio_to_host_stat (&fst, st); return 0; } /* Implementation of to_filesystem_is_local. */ bool remote_target::filesystem_is_local () { /* Valgrind GDB presents itself as a remote target but works on the local filesystem: it does not implement remote get and users are not expected to set a sysroot. To handle this case we treat the remote filesystem as local if the sysroot is exactly TARGET_SYSROOT_PREFIX and if the stub does not support vFile:open. */ if (gdb_sysroot == TARGET_SYSROOT_PREFIX) { packet_support ps = m_features.packet_support (PACKET_vFile_open); if (ps == PACKET_SUPPORT_UNKNOWN) { int fd; fileio_error remote_errno; /* Try opening a file to probe support. The supplied filename is irrelevant, we only care about whether the stub recognizes the packet or not. */ fd = remote_hostio_open (NULL, "just probing", FILEIO_O_RDONLY, 0700, 0, &remote_errno); if (fd >= 0) remote_hostio_close (fd, &remote_errno); ps = m_features.packet_support (PACKET_vFile_open); } if (ps == PACKET_DISABLE) { static int warning_issued = 0; if (!warning_issued) { warning (_("remote target does not support file" " transfer, attempting to access files" " from local filesystem.")); warning_issued = 1; } return true; } } return false; } static char * remote_hostio_error (fileio_error errnum) { int host_error = fileio_error_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)); } /* A RAII wrapper around a remote file descriptor. */ class scoped_remote_fd { public: scoped_remote_fd (remote_target *remote, int fd) : m_remote (remote), m_fd (fd) { } ~scoped_remote_fd () { if (m_fd != -1) { try { fileio_error remote_errno; m_remote->remote_hostio_close (m_fd, &remote_errno); } catch (...) { /* Swallow exception before it escapes the dtor. If something goes wrong, likely the connection is gone, and there's nothing else that can be done. */ } } } DISABLE_COPY_AND_ASSIGN (scoped_remote_fd); /* Release ownership of the file descriptor, and return it. */ ATTRIBUTE_UNUSED_RESULT int release () noexcept { int fd = m_fd; m_fd = -1; return fd; } /* Return the owned file descriptor. */ int get () const noexcept { return m_fd; } private: /* The remote target. */ remote_target *m_remote; /* The owned remote I/O file descriptor. */ int m_fd; }; void remote_file_put (const char *local_file, const char *remote_file, int from_tty) { remote_target *remote = get_current_remote_target (); if (remote == nullptr) error (_("command can only be used with remote target")); remote->remote_file_put (local_file, remote_file, from_tty); } void remote_target::remote_file_put (const char *local_file, const char *remote_file, int from_tty) { int retcode, bytes, io_size; fileio_error remote_errno; int bytes_in_buffer; int saw_eof; ULONGEST offset; gdb_file_up file = gdb_fopen_cloexec (local_file, "rb"); if (file == NULL) perror_with_name (local_file); scoped_remote_fd fd (this, remote_hostio_open (NULL, remote_file, (FILEIO_O_WRONLY | FILEIO_O_CREAT | FILEIO_O_TRUNC), 0700, 0, &remote_errno)); if (fd.get () == -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 (); gdb::byte_vector buffer (io_size); bytes_in_buffer = 0; saw_eof = 0; offset = 0; while (bytes_in_buffer || !saw_eof) { if (!saw_eof) { bytes = fread (buffer.data () + bytes_in_buffer, 1, io_size - bytes_in_buffer, file.get ()); if (bytes == 0) { if (ferror (file.get ())) 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 (fd.get (), buffer.data (), 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.data (), buffer.data () + retcode, bytes_in_buffer); } offset += retcode; } if (remote_hostio_close (fd.release (), &remote_errno)) remote_hostio_error (remote_errno); if (from_tty) gdb_printf (_("Successfully sent file \"%s\".\n"), local_file); } void remote_file_get (const char *remote_file, const char *local_file, int from_tty) { remote_target *remote = get_current_remote_target (); if (remote == nullptr) error (_("command can only be used with remote target")); remote->remote_file_get (remote_file, local_file, from_tty); } void remote_target::remote_file_get (const char *remote_file, const char *local_file, int from_tty) { fileio_error remote_errno; int bytes, io_size; ULONGEST offset; scoped_remote_fd fd (this, remote_hostio_open (NULL, remote_file, FILEIO_O_RDONLY, 0, 0, &remote_errno)); if (fd.get () == -1) remote_hostio_error (remote_errno); gdb_file_up file = gdb_fopen_cloexec (local_file, "wb"); if (file == NULL) perror_with_name (local_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 (); gdb::byte_vector buffer (io_size); offset = 0; while (1) { bytes = remote_hostio_pread (fd.get (), buffer.data (), 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.data (), 1, bytes, file.get ()); if (bytes == 0) perror_with_name (local_file); } if (remote_hostio_close (fd.release (), &remote_errno)) remote_hostio_error (remote_errno); if (from_tty) gdb_printf (_("Successfully fetched file \"%s\".\n"), remote_file); } void remote_file_delete (const char *remote_file, int from_tty) { remote_target *remote = get_current_remote_target (); if (remote == nullptr) error (_("command can only be used with remote target")); remote->remote_file_delete (remote_file, from_tty); } void remote_target::remote_file_delete (const char *remote_file, int from_tty) { int retcode; fileio_error remote_errno; retcode = remote_hostio_unlink (NULL, remote_file, &remote_errno); if (retcode == -1) remote_hostio_error (remote_errno); if (from_tty) gdb_printf (_("Successfully deleted file \"%s\".\n"), remote_file); } static void remote_put_command (const char *args, int from_tty) { if (args == NULL) error_no_arg (_("file to put")); gdb_argv argv (args); 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); } static void remote_get_command (const char *args, int from_tty) { if (args == NULL) error_no_arg (_("file to get")); gdb_argv argv (args); 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); } static void remote_delete_command (const char *args, int from_tty) { if (args == NULL) error_no_arg (_("file to delete")); gdb_argv argv (args); if (argv[0] == NULL || argv[1] != NULL) error (_("Invalid parameters to remote delete")); remote_file_delete (argv[0], from_tty); } bool remote_target::can_execute_reverse () { if (m_features.packet_support (PACKET_bs) == PACKET_ENABLE || m_features.packet_support (PACKET_bc) == PACKET_ENABLE) return true; else return false; } bool remote_target::supports_non_stop () { return true; } bool remote_target::supports_disable_randomization () { /* Only supported in extended mode. */ return false; } bool remote_target::supports_multi_process () { return m_features.remote_multi_process_p (); } int remote_target::remote_supports_cond_tracepoints () { return (m_features.packet_support (PACKET_ConditionalTracepoints) == PACKET_ENABLE); } bool remote_target::supports_evaluation_of_breakpoint_conditions () { return (m_features.packet_support (PACKET_ConditionalBreakpoints) == PACKET_ENABLE); } int remote_target::remote_supports_fast_tracepoints () { return m_features.packet_support (PACKET_FastTracepoints) == PACKET_ENABLE; } int remote_target::remote_supports_static_tracepoints () { return m_features.packet_support (PACKET_StaticTracepoints) == PACKET_ENABLE; } int remote_target::remote_supports_install_in_trace () { return m_features.packet_support (PACKET_InstallInTrace) == PACKET_ENABLE; } bool remote_target::supports_enable_disable_tracepoint () { return (m_features.packet_support (PACKET_EnableDisableTracepoints_feature) == PACKET_ENABLE); } bool remote_target::supports_string_tracing () { return m_features.packet_support (PACKET_tracenz_feature) == PACKET_ENABLE; } bool remote_target::can_run_breakpoint_commands () { return m_features.packet_support (PACKET_BreakpointCommands) == PACKET_ENABLE; } void remote_target::trace_init () { struct remote_state *rs = get_remote_state (); putpkt ("QTinit"); remote_get_noisy_reply (); if (strcmp (rs->buf.data (), "OK") != 0) error (_("Target does not support this command.")); } /* Recursive routine to walk through command list including loops, and download packets for each command. */ void remote_target::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.data (), "QTDPsrc:"); encode_source_string (num, addr, "cmd", cmd->line, rs->buf.data () + strlen (rs->buf.data ()), rs->buf.size () - strlen (rs->buf.data ())); putpkt (rs->buf); remote_get_noisy_reply (); if (strcmp (rs->buf.data (), "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_0.get ()); QUIT; /* Allow user to bail out with ^C. */ strcpy (rs->buf.data (), "QTDPsrc:"); encode_source_string (num, addr, "cmd", "end", rs->buf.data () + strlen (rs->buf.data ()), rs->buf.size () - strlen (rs->buf.data ())); putpkt (rs->buf); remote_get_noisy_reply (); if (strcmp (rs->buf.data (), "OK")) warning (_("Target does not support source download.")); } } } void remote_target::download_tracepoint (struct bp_location *loc) { CORE_ADDR tpaddr; char addrbuf[40]; std::vector tdp_actions; std::vector stepping_actions; char *pkt; struct breakpoint *b = loc->owner; struct tracepoint *t = (struct tracepoint *) b; struct remote_state *rs = get_remote_state (); int ret; const char *err_msg = _("Tracepoint packet too large for target."); size_t size_left; /* We use a buffer other than rs->buf because we'll build strings across multiple statements, and other statements in between could modify rs->buf. */ gdb::char_vector buf (get_remote_packet_size ()); encode_actions_rsp (loc, &tdp_actions, &stepping_actions); tpaddr = loc->address; strcpy (addrbuf, phex (tpaddr, sizeof (CORE_ADDR))); ret = snprintf (buf.data (), 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); if (ret < 0 || ret >= buf.size ()) error ("%s", err_msg); /* 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 ()) { if (gdbarch_fast_tracepoint_valid_at (loc->gdbarch, tpaddr, NULL)) { size_left = buf.size () - strlen (buf.data ()); ret = snprintf (buf.data () + strlen (buf.data ()), size_left, ":F%x", gdb_insn_length (loc->gdbarch, tpaddr)); if (ret < 0 || ret >= size_left) error ("%s", err_msg); } else /* If it passed validation at definition but fails now, something is very wrong. */ internal_error (_("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 || b->type == bp_static_marker_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)) { size_left = buf.size () - strlen (buf.data ()); ret = snprintf (buf.data () + strlen (buf.data ()), size_left, ":S"); if (ret < 0 || ret >= size_left) error ("%s", err_msg); } 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 ()) { agent_expr_up aexpr = gen_eval_for_expr (tpaddr, loc->cond.get ()); size_left = buf.size () - strlen (buf.data ()); ret = snprintf (buf.data () + strlen (buf.data ()), size_left, ":X%x,", aexpr->len); if (ret < 0 || ret >= size_left) error ("%s", err_msg); size_left = buf.size () - strlen (buf.data ()); /* Two bytes to encode each aexpr byte, plus the terminating null byte. */ if (aexpr->len * 2 + 1 > size_left) error ("%s", err_msg); pkt = buf.data () + strlen (buf.data ()); for (int ndx = 0; ndx < aexpr->len; ++ndx) pkt = pack_hex_byte (pkt, aexpr->buf[ndx]); *pkt = '\0'; } else warning (_("Target does not support conditional tracepoints, " "ignoring tp %d cond"), b->number); } if (b->commands || !default_collect.empty ()) { size_left = buf.size () - strlen (buf.data ()); ret = snprintf (buf.data () + strlen (buf.data ()), size_left, "-"); if (ret < 0 || ret >= size_left) error ("%s", err_msg); } putpkt (buf.data ()); remote_get_noisy_reply (); if (strcmp (rs->buf.data (), "OK")) error (_("Target does not support tracepoints.")); /* do_single_steps (t); */ for (auto action_it = tdp_actions.begin (); action_it != tdp_actions.end (); action_it++) { QUIT; /* Allow user to bail out with ^C. */ bool has_more = ((action_it + 1) != tdp_actions.end () || !stepping_actions.empty ()); ret = snprintf (buf.data (), buf.size (), "QTDP:-%x:%s:%s%c", b->number, addrbuf, /* address */ action_it->c_str (), has_more ? '-' : 0); if (ret < 0 || ret >= buf.size ()) error ("%s", err_msg); putpkt (buf.data ()); remote_get_noisy_reply (); if (strcmp (rs->buf.data (), "OK")) error (_("Error on target while setting tracepoints.")); } for (auto action_it = stepping_actions.begin (); action_it != stepping_actions.end (); action_it++) { QUIT; /* Allow user to bail out with ^C. */ bool is_first = action_it == stepping_actions.begin (); bool has_more = (action_it + 1) != stepping_actions.end (); ret = snprintf (buf.data (), buf.size (), "QTDP:-%x:%s:%s%s%s", b->number, addrbuf, /* address */ is_first ? "S" : "", action_it->c_str (), has_more ? "-" : ""); if (ret < 0 || ret >= buf.size ()) error ("%s", err_msg); putpkt (buf.data ()); remote_get_noisy_reply (); if (strcmp (rs->buf.data (), "OK")) error (_("Error on target while setting tracepoints.")); } if (m_features.packet_support (PACKET_TracepointSource) == PACKET_ENABLE) { if (b->locspec != nullptr) { ret = snprintf (buf.data (), buf.size (), "QTDPsrc:"); if (ret < 0 || ret >= buf.size ()) error ("%s", err_msg); const char *str = b->locspec->to_string (); encode_source_string (b->number, loc->address, "at", str, buf.data () + strlen (buf.data ()), buf.size () - strlen (buf.data ())); putpkt (buf.data ()); remote_get_noisy_reply (); if (strcmp (rs->buf.data (), "OK")) warning (_("Target does not support source download.")); } if (b->cond_string) { ret = snprintf (buf.data (), buf.size (), "QTDPsrc:"); if (ret < 0 || ret >= buf.size ()) error ("%s", err_msg); encode_source_string (b->number, loc->address, "cond", b->cond_string.get (), buf.data () + strlen (buf.data ()), buf.size () - strlen (buf.data ())); putpkt (buf.data ()); remote_get_noisy_reply (); if (strcmp (rs->buf.data (), "OK")) warning (_("Target does not support source download.")); } remote_download_command_source (b->number, loc->address, breakpoint_commands (b)); } } bool remote_target::can_download_tracepoint () { 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 false; ts = current_trace_status (); status = get_trace_status (ts); if (status == -1 || !ts->running_known || !ts->running) return false; /* 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 false; return true; } void remote_target::download_trace_state_variable (const trace_state_variable &tsv) { struct remote_state *rs = get_remote_state (); char *p; xsnprintf (rs->buf.data (), get_remote_packet_size (), "QTDV:%x:%s:%x:", tsv.number, phex ((ULONGEST) tsv.initial_value, 8), tsv.builtin); p = rs->buf.data () + strlen (rs->buf.data ()); if ((p - rs->buf.data ()) + tsv.name.length () * 2 >= get_remote_packet_size ()) error (_("Trace state variable name too long for tsv definition packet")); p += 2 * bin2hex ((gdb_byte *) (tsv.name.data ()), p, tsv.name.length ()); *p++ = '\0'; putpkt (rs->buf); remote_get_noisy_reply (); if (rs->buf[0] == '\0') error (_("Target does not support this command.")); if (strcmp (rs->buf.data (), "OK") != 0) error (_("Error on target while downloading trace state variable.")); } void remote_target::enable_tracepoint (struct bp_location *location) { struct remote_state *rs = get_remote_state (); xsnprintf (rs->buf.data (), get_remote_packet_size (), "QTEnable:%x:%s", location->owner->number, phex (location->address, sizeof (CORE_ADDR))); putpkt (rs->buf); remote_get_noisy_reply (); if (rs->buf[0] == '\0') error (_("Target does not support enabling tracepoints while a trace run is ongoing.")); if (strcmp (rs->buf.data (), "OK") != 0) error (_("Error on target while enabling tracepoint.")); } void remote_target::disable_tracepoint (struct bp_location *location) { struct remote_state *rs = get_remote_state (); xsnprintf (rs->buf.data (), get_remote_packet_size (), "QTDisable:%x:%s", location->owner->number, phex (location->address, sizeof (CORE_ADDR))); putpkt (rs->buf); remote_get_noisy_reply (); if (rs->buf[0] == '\0') error (_("Target does not support disabling tracepoints while a trace run is ongoing.")); if (strcmp (rs->buf.data (), "OK") != 0) error (_("Error on target while disabling tracepoint.")); } void remote_target::trace_set_readonly_regions () { asection *s; bfd_size_type size; bfd_vma vma; int anysecs = 0; int offset = 0; bfd *abfd = current_program_space->exec_bfd (); if (!abfd) return; /* No information to give. */ struct remote_state *rs = get_remote_state (); strcpy (rs->buf.data (), "QTro"); offset = strlen (rs->buf.data ()); for (s = abfd->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_section_vma (s); size = bfd_section_size (s); bfd_sprintf_vma (abfd, tmp1, vma); bfd_sprintf_vma (abfd, tmp2, vma + size); sec_length = 1 + strlen (tmp1) + 1 + strlen (tmp2); if (offset + sec_length + 1 > rs->buf.size ()) { if (m_features.packet_support (PACKET_qXfer_traceframe_info) != PACKET_ENABLE) warning (_("\ Too many sections for read-only sections definition packet.")); break; } xsnprintf (rs->buf.data () + offset, rs->buf.size () - offset, ":%s,%s", tmp1, tmp2); offset += sec_length; } if (anysecs) { putpkt (rs->buf); getpkt (&rs->buf, 0); } } void remote_target::trace_start () { struct remote_state *rs = get_remote_state (); putpkt ("QTStart"); remote_get_noisy_reply (); if (rs->buf[0] == '\0') error (_("Target does not support this command.")); if (strcmp (rs->buf.data (), "OK") != 0) error (_("Bogus reply from target: %s"), rs->buf.data ()); } int remote_target::get_trace_status (struct trace_status *ts) { /* Initialize it just to avoid a GCC false warning. */ char *p = NULL; enum packet_result result; struct remote_state *rs = get_remote_state (); if (m_features.packet_support (PACKET_qTStatus) == PACKET_DISABLE) return -1; /* FIXME we need to get register block size some other way. */ trace_regblock_size = rs->get_remote_arch_state (target_gdbarch ())->sizeof_g_packet; putpkt ("qTStatus"); try { p = remote_get_noisy_reply (); } catch (const gdb_exception_error &ex) { if (ex.error != TARGET_CLOSE_ERROR) { exception_fprintf (gdb_stderr, ex, "qTStatus: "); return -1; } throw; } result = m_features.packet_ok (p, 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"), rs->buf.data ()); /* 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; } void remote_target::get_tracepoint_status (struct breakpoint *bp, struct uploaded_tp *utp) { struct remote_state *rs = get_remote_state (); char *reply; struct tracepoint *tp = (struct tracepoint *) bp; size_t size = get_remote_packet_size (); if (tp) { tp->hit_count = 0; tp->traceframe_usage = 0; for (bp_location *loc : tp->locations ()) { /* If the tracepoint was never downloaded, don't go asking for any status. */ if (tp->number_on_target == 0) continue; xsnprintf (rs->buf.data (), size, "qTP:%x:%s", tp->number_on_target, phex_nz (loc->address, 0)); putpkt (rs->buf); reply = remote_get_noisy_reply (); 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.data (), size, "qTP:%x:%s", utp->number, phex_nz (utp->addr, 0)); putpkt (rs->buf); reply = remote_get_noisy_reply (); if (reply && *reply) { if (*reply == 'V') parse_tracepoint_status (reply + 1, bp, utp); } } } void remote_target::trace_stop () { struct remote_state *rs = get_remote_state (); putpkt ("QTStop"); remote_get_noisy_reply (); if (rs->buf[0] == '\0') error (_("Target does not support this command.")); if (strcmp (rs->buf.data (), "OK") != 0) error (_("Bogus reply from target: %s"), rs->buf.data ()); } int remote_target::trace_find (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.data () + 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.data (); 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 (); 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; } bool remote_target::get_trace_state_variable_value (int tsvnum, LONGEST *val) { struct remote_state *rs = get_remote_state (); char *reply; ULONGEST uval; set_remote_traceframe (); xsnprintf (rs->buf.data (), get_remote_packet_size (), "qTV:%x", tsvnum); putpkt (rs->buf); reply = remote_get_noisy_reply (); if (reply && *reply) { if (*reply == 'V') { unpack_varlen_hex (reply + 1, &uval); *val = (LONGEST) uval; return true; } } return false; } int remote_target::save_trace_data (const char *filename) { struct remote_state *rs = get_remote_state (); char *p, *reply; p = rs->buf.data (); strcpy (p, "QTSave:"); p += strlen (p); if ((p - rs->buf.data ()) + 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 (); 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. */ LONGEST remote_target::get_raw_trace_data (gdb_byte *buf, ULONGEST offset, LONGEST len) { struct remote_state *rs = get_remote_state (); char *reply; char *p; int rslt; p = rs->buf.data (); 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 (); 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 (reply, buf, len); return rslt; } /* Something went wrong, flag as an error. */ return -1; } void remote_target::set_disconnected_tracing (int val) { struct remote_state *rs = get_remote_state (); if (m_features.packet_support (PACKET_DisconnectedTracing_feature) == PACKET_ENABLE) { char *reply; xsnprintf (rs->buf.data (), get_remote_packet_size (), "QTDisconnected:%x", val); putpkt (rs->buf); reply = remote_get_noisy_reply (); 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.")); } int remote_target::core_of_thread (ptid_t ptid) { thread_info *info = this->find_thread (ptid); if (info != NULL && info->priv != NULL) return get_remote_thread_info (info)->core; return -1; } void remote_target::set_circular_trace_buffer (int val) { struct remote_state *rs = get_remote_state (); char *reply; xsnprintf (rs->buf.data (), get_remote_packet_size (), "QTBuffer:circular:%x", val); putpkt (rs->buf); reply = remote_get_noisy_reply (); if (*reply == '\0') error (_("Target does not support this command.")); if (strcmp (reply, "OK") != 0) error (_("Bogus reply from target: %s"), reply); } traceframe_info_up remote_target::traceframe_info () { gdb::optional text = target_read_stralloc (current_inferior ()->top_target (), TARGET_OBJECT_TRACEFRAME_INFO, NULL); if (text) return parse_traceframe_info (text->data ()); 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. */ int remote_target::get_min_fast_tracepoint_insn_len () { 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.data (), get_remote_packet_size (), "qTMinFTPILen"); putpkt (rs->buf); reply = remote_get_noisy_reply (); if (*reply == '\0') return -1; else { ULONGEST min_insn_len; unpack_varlen_hex (reply, &min_insn_len); return (int) min_insn_len; } } void remote_target::set_trace_buffer_size (LONGEST val) { if (m_features.packet_support (PACKET_QTBuffer_size) != PACKET_DISABLE) { struct remote_state *rs = get_remote_state (); char *buf = rs->buf.data (); char *endbuf = 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 (); result = m_features.packet_ok (rs->buf, PACKET_QTBuffer_size); if (result != PACKET_OK) warning (_("Bogus reply from target: %s"), rs->buf.data ()); } } bool remote_target::set_trace_notes (const char *user, const char *notes, const char *stop_notes) { struct remote_state *rs = get_remote_state (); char *reply; char *buf = rs->buf.data (); char *endbuf = 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 (); if (*reply == '\0') return false; if (strcmp (reply, "OK") != 0) error (_("Bogus reply from target: %s"), reply); return true; } bool remote_target::use_agent (bool use) { if (m_features.packet_support (PACKET_QAgent) != PACKET_DISABLE) { struct remote_state *rs = get_remote_state (); /* If the stub supports QAgent. */ xsnprintf (rs->buf.data (), get_remote_packet_size (), "QAgent:%d", use); putpkt (rs->buf); getpkt (&rs->buf, 0); if (strcmp (rs->buf.data (), "OK") == 0) { ::use_agent = use; return true; } } return false; } bool remote_target::can_use_agent () { return (m_features.packet_support (PACKET_QAgent) != PACKET_DISABLE); } struct btrace_target_info { /* The ptid of the traced thread. */ ptid_t ptid; /* The obtained branch trace configuration. */ struct btrace_config conf; }; /* Reset our idea of our target's btrace configuration. */ static void remote_btrace_reset (remote_state *rs) { memset (&rs->btrace_config, 0, sizeof (rs->btrace_config)); } /* Synchronize the configuration with the target. */ void remote_target::btrace_sync_conf (const btrace_config *conf) { struct remote_state *rs; char *buf, *pos, *endbuf; rs = get_remote_state (); buf = rs->buf.data (); endbuf = buf + get_remote_packet_size (); if (m_features.packet_support (PACKET_Qbtrace_conf_bts_size) == PACKET_ENABLE && conf->bts.size != rs->btrace_config.bts.size) { pos = buf; pos += xsnprintf (pos, endbuf - pos, "%s=0x%x", packets_descriptions[PACKET_Qbtrace_conf_bts_size].name, conf->bts.size); putpkt (buf); getpkt (&rs->buf, 0); if (m_features.packet_ok (buf, PACKET_Qbtrace_conf_bts_size) == PACKET_ERROR) { if (buf[0] == 'E' && buf[1] == '.') error (_("Failed to configure the BTS buffer size: %s"), buf + 2); else error (_("Failed to configure the BTS buffer size.")); } rs->btrace_config.bts.size = conf->bts.size; } if (m_features.packet_support (PACKET_Qbtrace_conf_pt_size) == PACKET_ENABLE && conf->pt.size != rs->btrace_config.pt.size) { pos = buf; pos += xsnprintf (pos, endbuf - pos, "%s=0x%x", packets_descriptions[PACKET_Qbtrace_conf_pt_size].name, conf->pt.size); putpkt (buf); getpkt (&rs->buf, 0); if (m_features.packet_ok (buf, PACKET_Qbtrace_conf_pt_size) == PACKET_ERROR) { if (buf[0] == 'E' && buf[1] == '.') error (_("Failed to configure the trace buffer size: %s"), buf + 2); else error (_("Failed to configure the trace buffer size.")); } rs->btrace_config.pt.size = conf->pt.size; } } /* Read TP's btrace configuration from the target and store it into CONF. */ static void btrace_read_config (thread_info *tp, struct btrace_config *conf) { /* target_read_stralloc relies on INFERIOR_PTID. */ scoped_restore_current_thread restore_thread; switch_to_thread (tp); gdb::optional xml = target_read_stralloc (current_inferior ()->top_target (), TARGET_OBJECT_BTRACE_CONF, ""); if (xml) parse_xml_btrace_conf (conf, xml->data ()); } /* Maybe reopen target btrace. */ void remote_target::remote_btrace_maybe_reopen () { struct remote_state *rs = get_remote_state (); int btrace_target_pushed = 0; #if !defined (HAVE_LIBIPT) int warned = 0; #endif /* Don't bother walking the entirety of the remote thread list when we know the feature isn't supported by the remote. */ if (m_features.packet_support (PACKET_qXfer_btrace_conf) != PACKET_ENABLE) return; for (thread_info *tp : all_non_exited_threads (this)) { memset (&rs->btrace_config, 0x00, sizeof (struct btrace_config)); btrace_read_config (tp, &rs->btrace_config); if (rs->btrace_config.format == BTRACE_FORMAT_NONE) continue; #if !defined (HAVE_LIBIPT) if (rs->btrace_config.format == BTRACE_FORMAT_PT) { if (!warned) { warned = 1; warning (_("Target is recording using Intel Processor Trace " "but support was disabled at compile time.")); } continue; } #endif /* !defined (HAVE_LIBIPT) */ /* Push target, once, but before anything else happens. This way our changes to the threads will be cleaned up by unpushing the target in case btrace_read_config () throws. */ if (!btrace_target_pushed) { btrace_target_pushed = 1; record_btrace_push_target (); gdb_printf (_("Target is recording using %s.\n"), btrace_format_string (rs->btrace_config.format)); } tp->btrace.target = XCNEW (struct btrace_target_info); tp->btrace.target->ptid = tp->ptid; tp->btrace.target->conf = rs->btrace_config; } } /* Enable branch tracing. */ struct btrace_target_info * remote_target::enable_btrace (thread_info *tp, const struct btrace_config *conf) { struct btrace_target_info *tinfo = NULL; struct packet_config *packet = NULL; struct remote_state *rs = get_remote_state (); char *buf = rs->buf.data (); char *endbuf = buf + get_remote_packet_size (); unsigned int which_packet; switch (conf->format) { case BTRACE_FORMAT_BTS: which_packet = PACKET_Qbtrace_bts; break; case BTRACE_FORMAT_PT: which_packet = PACKET_Qbtrace_pt; break; default: internal_error (_("Bad branch btrace format: %u."), (unsigned int) conf->format); } packet = &m_features.m_protocol_packets[which_packet]; if (packet == NULL || packet_config_support (packet) != PACKET_ENABLE) error (_("Target does not support branch tracing.")); btrace_sync_conf (conf); ptid_t ptid = tp->ptid; set_general_thread (ptid); buf += xsnprintf (buf, endbuf - buf, "%s", packets_descriptions[which_packet].name); putpkt (rs->buf); getpkt (&rs->buf, 0); if (m_features.packet_ok (rs->buf, which_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).c_str (), &rs->buf[2]); else error (_("Could not enable branch tracing for %s."), target_pid_to_str (ptid).c_str ()); } tinfo = XCNEW (struct btrace_target_info); tinfo->ptid = ptid; /* If we fail to read the configuration, we lose some information, but the tracing itself is not impacted. */ try { btrace_read_config (tp, &tinfo->conf); } catch (const gdb_exception_error &err) { if (err.message != NULL) warning ("%s", err.what ()); } return tinfo; } /* Disable branch tracing. */ void remote_target::disable_btrace (struct btrace_target_info *tinfo) { struct remote_state *rs = get_remote_state (); char *buf = rs->buf.data (); char *endbuf = buf + get_remote_packet_size (); if (m_features.packet_support (PACKET_Qbtrace_off) != PACKET_ENABLE) error (_("Target does not support branch tracing.")); set_general_thread (tinfo->ptid); buf += xsnprintf (buf, endbuf - buf, "%s", packets_descriptions[PACKET_Qbtrace_off].name); putpkt (rs->buf); getpkt (&rs->buf, 0); if (m_features.packet_ok (rs->buf, PACKET_Qbtrace_off) == 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).c_str (), &rs->buf[2]); else error (_("Could not disable branch tracing for %s."), target_pid_to_str (tinfo->ptid).c_str ()); } xfree (tinfo); } /* Teardown branch tracing. */ void remote_target::teardown_btrace (struct btrace_target_info *tinfo) { /* We must not talk to the target during teardown. */ xfree (tinfo); } /* Read the branch trace. */ enum btrace_error remote_target::read_btrace (struct btrace_data *btrace, struct btrace_target_info *tinfo, enum btrace_read_type type) { const char *annex; if (m_features.packet_support (PACKET_qXfer_btrace) != 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 (_("Bad branch tracing read type: %u."), (unsigned int) type); } gdb::optional xml = target_read_stralloc (current_inferior ()->top_target (), TARGET_OBJECT_BTRACE, annex); if (!xml) return BTRACE_ERR_UNKNOWN; parse_xml_btrace (btrace, xml->data ()); return BTRACE_ERR_NONE; } const struct btrace_config * remote_target::btrace_conf (const struct btrace_target_info *tinfo) { return &tinfo->conf; } bool remote_target::augmented_libraries_svr4_read () { return (m_features.packet_support (PACKET_augmented_libraries_svr4_read_feature) == PACKET_ENABLE); } /* Implementation of to_load. */ void remote_target::load (const char *name, int from_tty) { generic_load (name, from_tty); } /* Accepts an integer PID; returns a string representing a file that can be opened on the remote side to get the symbols for the child process. Returns NULL if the operation is not supported. */ const char * remote_target::pid_to_exec_file (int pid) { static gdb::optional filename; char *annex = NULL; if (m_features.packet_support (PACKET_qXfer_exec_file) != PACKET_ENABLE) return NULL; inferior *inf = find_inferior_pid (this, pid); if (inf == NULL) internal_error (_("not currently attached to process %d"), pid); if (!inf->fake_pid_p) { const int annex_size = 9; annex = (char *) alloca (annex_size); xsnprintf (annex, annex_size, "%x", pid); } filename = target_read_stralloc (current_inferior ()->top_target (), TARGET_OBJECT_EXEC_FILE, annex); return filename ? filename->data () : nullptr; } /* Implement the to_can_do_single_step target_ops method. */ int remote_target::can_do_single_step () { /* We can only tell whether target supports single step or not by supported s and S vCont actions if the stub supports vContSupported feature. If the stub doesn't support vContSupported feature, we have conservatively to think target doesn't supports single step. */ if (m_features.packet_support (PACKET_vContSupported) == PACKET_ENABLE) { struct remote_state *rs = get_remote_state (); return rs->supports_vCont.s && rs->supports_vCont.S; } else return 0; } /* Implementation of the to_execution_direction method for the remote target. */ enum exec_direction_kind remote_target::execution_direction () { struct remote_state *rs = get_remote_state (); return rs->last_resume_exec_dir; } /* Return pointer to the thread_info struct which corresponds to THREAD_HANDLE (having length HANDLE_LEN). */ thread_info * remote_target::thread_handle_to_thread_info (const gdb_byte *thread_handle, int handle_len, inferior *inf) { for (thread_info *tp : all_non_exited_threads (this)) { remote_thread_info *priv = get_remote_thread_info (tp); if (tp->inf == inf && priv != NULL) { if (handle_len != priv->thread_handle.size ()) error (_("Thread handle size mismatch: %d vs %zu (from remote)"), handle_len, priv->thread_handle.size ()); if (memcmp (thread_handle, priv->thread_handle.data (), handle_len) == 0) return tp; } } return NULL; } gdb::byte_vector remote_target::thread_info_to_thread_handle (struct thread_info *tp) { remote_thread_info *priv = get_remote_thread_info (tp); return priv->thread_handle; } bool remote_target::can_async_p () { /* This flag should be checked in the common target.c code. */ gdb_assert (target_async_permitted); /* We're async whenever the serial device can. */ struct remote_state *rs = get_remote_state (); return serial_can_async_p (rs->remote_desc); } bool remote_target::is_async_p () { /* We're async whenever the serial device is. */ struct remote_state *rs = get_remote_state (); 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) { /* Don't propogate error information up to the client. Instead let the client find out about the error by querying the target. */ inferior_event_handler (INF_REG_EVENT); } static void remote_async_inferior_event_handler (gdb_client_data data) { inferior_event_handler (INF_REG_EVENT); } int remote_target::async_wait_fd () { struct remote_state *rs = get_remote_state (); return rs->remote_desc->fd; } void remote_target::async (bool enable) { struct remote_state *rs = get_remote_state (); if (enable) { serial_async (rs->remote_desc, remote_async_serial_handler, rs); /* If there are pending events in the stop reply queue tell the event loop to process them. */ if (!rs->stop_reply_queue.empty ()) mark_async_event_handler (rs->remote_async_inferior_event_token); /* For simplicity, below we clear the pending events token without remembering whether it is marked, so here we always mark it. If there's actually no pending notification to process, this ends up being a no-op (other than a spurious event-loop wakeup). */ if (target_is_non_stop_p ()) mark_async_event_handler (rs->notif_state->get_pending_events_token); } else { serial_async (rs->remote_desc, NULL, NULL); /* If the core is disabling async, it doesn't want to be disturbed with target events. Clear all async event sources too. */ clear_async_event_handler (rs->remote_async_inferior_event_token); if (target_is_non_stop_p ()) clear_async_event_handler (rs->notif_state->get_pending_events_token); } } /* Implementation of the to_thread_events method. */ void remote_target::thread_events (int enable) { struct remote_state *rs = get_remote_state (); size_t size = get_remote_packet_size (); if (m_features.packet_support (PACKET_QThreadEvents) == PACKET_DISABLE) return; xsnprintf (rs->buf.data (), size, "QThreadEvents:%x", enable ? 1 : 0); putpkt (rs->buf); getpkt (&rs->buf, 0); switch (m_features.packet_ok (rs->buf, PACKET_QThreadEvents)) { case PACKET_OK: if (strcmp (rs->buf.data (), "OK") != 0) error (_("Remote refused setting thread events: %s"), rs->buf.data ()); break; case PACKET_ERROR: warning (_("Remote failure reply: %s"), rs->buf.data ()); break; case PACKET_UNKNOWN: break; } } static void show_remote_cmd (const 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 cmd_list_element *list = remote_show_cmdlist; struct ui_out *uiout = current_uiout; ui_out_emit_tuple tuple_emitter (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 { ui_out_emit_tuple option_emitter (uiout, "option"); uiout->field_string ("name", list->name); uiout->text (": "); if (list->type == show_cmd) do_show_command (NULL, from_tty, list); else cmd_func (list, NULL, from_tty); } } /* Function to be called whenever a new objfile (shlib) is detected. */ static void remote_new_objfile (struct objfile *objfile) { /* The objfile change happened in that program space. */ program_space *pspace = current_program_space; /* The affected program space is possibly shared by multiple inferiors. Consider sending a qSymbol packet for each of the inferiors using that program space. */ for (inferior *inf : all_inferiors ()) { if (inf->pspace != pspace) continue; /* Check whether the inferior's process target is a remote target. */ remote_target *remote = as_remote_target (inf->process_target ()); if (remote == nullptr) continue; /* When we are attaching or handling a fork child and the shared library subsystem reads the list of loaded libraries, we receive new objfile events in between each found library. The libraries are read in an undefined order, so if we gave the remote side a chance to look up symbols between each objfile, we might give it an inconsistent picture of the inferior. It could appear that a library A appears loaded but a library B does not, even though library A requires library B. That would present a state that couldn't normally exist in the inferior. So, skip these events, we'll give the remote a chance to look up symbols once all the loaded libraries and their symbols are known to GDB. */ if (inf->in_initial_library_scan) continue; if (!remote->has_execution (inf)) continue; /* Need to switch to a specific thread, because remote_check_symbols will set the general thread using INFERIOR_PTID. It's possible to have inferiors with no thread here, because we are called very early in the connection process, while the inferior is being set up, before threads are added. Just skip it, start_remote_1 also calls remote_check_symbols when it's done setting things up. */ thread_info *thread = any_thread_of_inferior (inf); if (thread != nullptr) { scoped_restore_current_thread restore_thread; switch_to_thread (thread); remote->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. */ int remote_target::upload_tracepoints (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, 0); p = rs->buf.data (); while (*p && *p != 'l') { parse_tracepoint_definition (p, utpp); /* Ask for another packet of tracepoint definition. */ putpkt ("qTsP"); getpkt (&rs->buf, 0); p = rs->buf.data (); } return 0; } int remote_target::upload_trace_state_variables (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, 0); p = rs->buf.data (); while (*p && *p != 'l') { parse_tsv_definition (p, utsvp); /* Ask for another packet of variable definition. */ putpkt ("qTsV"); getpkt (&rs->buf, 0); p = rs->buf.data (); } 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) { gdb_printf (file, _("Debugger's willingness to use range stepping " "is %s.\n"), value); } /* Return true if the vCont;r action is supported by the remote stub. */ bool remote_target::vcont_r_supported () { return (m_features.packet_support (PACKET_vCont) == PACKET_ENABLE && get_remote_state ()->supports_vCont.r); } /* The "set/show range-stepping" set hook. */ static void set_range_stepping (const char *ignore_args, int from_tty, struct cmd_list_element *c) { /* When enabling, check whether range stepping is actually supported by the target, and warn if not. */ if (use_range_stepping) { remote_target *remote = get_current_remote_target (); if (remote == NULL || !remote->vcont_r_supported ()) warning (_("Range stepping is not supported by the current target")); } } static void show_remote_debug (struct ui_file *file, int from_tty, struct cmd_list_element *c, const char *value) { gdb_printf (file, _("Debugging of remote protocol is %s.\n"), value); } static void show_remote_timeout (struct ui_file *file, int from_tty, struct cmd_list_element *c, const char *value) { gdb_printf (file, _("Timeout limit to wait for target to respond is %s.\n"), value); } /* Implement the "supports_memory_tagging" target_ops method. */ bool remote_target::supports_memory_tagging () { return m_features.remote_memory_tagging_p (); } /* Create the qMemTags packet given ADDRESS, LEN and TYPE. */ static void create_fetch_memtags_request (gdb::char_vector &packet, CORE_ADDR address, size_t len, int type) { int addr_size = gdbarch_addr_bit (target_gdbarch ()) / 8; std::string request = string_printf ("qMemTags:%s,%s:%s", phex_nz (address, addr_size), phex_nz (len, sizeof (len)), phex_nz (type, sizeof (type))); strcpy (packet.data (), request.c_str ()); } /* Parse the qMemTags packet reply into TAGS. Return true if successful, false otherwise. */ static bool parse_fetch_memtags_reply (const gdb::char_vector &reply, gdb::byte_vector &tags) { if (reply.empty () || reply[0] == 'E' || reply[0] != 'm') return false; /* Copy the tag data. */ tags = hex2bin (reply.data () + 1); return true; } /* Create the QMemTags packet given ADDRESS, LEN, TYPE and TAGS. */ static void create_store_memtags_request (gdb::char_vector &packet, CORE_ADDR address, size_t len, int type, const gdb::byte_vector &tags) { int addr_size = gdbarch_addr_bit (target_gdbarch ()) / 8; /* Put together the main packet, address and length. */ std::string request = string_printf ("QMemTags:%s,%s:%s:", phex_nz (address, addr_size), phex_nz (len, sizeof (len)), phex_nz (type, sizeof (type))); request += bin2hex (tags.data (), tags.size ()); /* Check if we have exceeded the maximum packet size. */ if (packet.size () < request.length ()) error (_("Contents too big for packet QMemTags.")); strcpy (packet.data (), request.c_str ()); } /* Implement the "fetch_memtags" target_ops method. */ bool remote_target::fetch_memtags (CORE_ADDR address, size_t len, gdb::byte_vector &tags, int type) { /* Make sure the qMemTags packet is supported. */ if (!m_features.remote_memory_tagging_p ()) gdb_assert_not_reached ("remote fetch_memtags called with packet disabled"); struct remote_state *rs = get_remote_state (); create_fetch_memtags_request (rs->buf, address, len, type); putpkt (rs->buf); getpkt (&rs->buf, 0); return parse_fetch_memtags_reply (rs->buf, tags); } /* Implement the "store_memtags" target_ops method. */ bool remote_target::store_memtags (CORE_ADDR address, size_t len, const gdb::byte_vector &tags, int type) { /* Make sure the QMemTags packet is supported. */ if (!m_features.remote_memory_tagging_p ()) gdb_assert_not_reached ("remote store_memtags called with packet disabled"); struct remote_state *rs = get_remote_state (); create_store_memtags_request (rs->buf, address, len, type, tags); putpkt (rs->buf); getpkt (&rs->buf, 0); /* Verify if the request was successful. */ return packet_check_result (rs->buf.data ()) == PACKET_OK; } /* Return true if remote target T is non-stop. */ bool remote_target_is_non_stop_p (remote_target *t) { scoped_restore_current_thread restore_thread; switch_to_target_no_thread (t); return target_is_non_stop_p (); } #if GDB_SELF_TEST namespace selftests { static void test_memory_tagging_functions () { remote_target remote; struct packet_config *config = &remote.m_features.m_protocol_packets[PACKET_memory_tagging_feature]; scoped_restore restore_memtag_support_ = make_scoped_restore (&config->support); /* Test memory tagging packet support. */ config->support = PACKET_SUPPORT_UNKNOWN; SELF_CHECK (remote.supports_memory_tagging () == false); config->support = PACKET_DISABLE; SELF_CHECK (remote.supports_memory_tagging () == false); config->support = PACKET_ENABLE; SELF_CHECK (remote.supports_memory_tagging () == true); /* Setup testing. */ gdb::char_vector packet; gdb::byte_vector tags, bv; std::string expected, reply; packet.resize (32000); /* Test creating a qMemTags request. */ expected = "qMemTags:0,0:0"; create_fetch_memtags_request (packet, 0x0, 0x0, 0); SELF_CHECK (strcmp (packet.data (), expected.c_str ()) == 0); expected = "qMemTags:deadbeef,10:1"; create_fetch_memtags_request (packet, 0xdeadbeef, 16, 1); SELF_CHECK (strcmp (packet.data (), expected.c_str ()) == 0); /* Test parsing a qMemTags reply. */ /* Error reply, tags vector unmodified. */ reply = "E00"; strcpy (packet.data (), reply.c_str ()); tags.resize (0); SELF_CHECK (parse_fetch_memtags_reply (packet, tags) == false); SELF_CHECK (tags.size () == 0); /* Valid reply, tags vector updated. */ tags.resize (0); bv.resize (0); for (int i = 0; i < 5; i++) bv.push_back (i); reply = "m" + bin2hex (bv.data (), bv.size ()); strcpy (packet.data (), reply.c_str ()); SELF_CHECK (parse_fetch_memtags_reply (packet, tags) == true); SELF_CHECK (tags.size () == 5); for (int i = 0; i < 5; i++) SELF_CHECK (tags[i] == i); /* Test creating a QMemTags request. */ /* Empty tag data. */ tags.resize (0); expected = "QMemTags:0,0:0:"; create_store_memtags_request (packet, 0x0, 0x0, 0, tags); SELF_CHECK (memcmp (packet.data (), expected.c_str (), expected.length ()) == 0); /* Non-empty tag data. */ tags.resize (0); for (int i = 0; i < 5; i++) tags.push_back (i); expected = "QMemTags:deadbeef,ff:1:0001020304"; create_store_memtags_request (packet, 0xdeadbeef, 255, 1, tags); SELF_CHECK (memcmp (packet.data (), expected.c_str (), expected.length ()) == 0); } } // namespace selftests #endif /* GDB_SELF_TEST */ void _initialize_remote (); void _initialize_remote () { add_target (remote_target_info, remote_target::open); add_target (extended_remote_target_info, extended_remote_target::open); /* Hook into new objfile notification. */ gdb::observers::new_objfile.attach (remote_new_objfile, "remote"); #if 0 init_remote_threadtests (); #endif /* set/show remote ... */ add_basic_prefix_cmd ("remote", class_maintenance, _("\ Remote protocol specific variables.\n\ Configure various remote-protocol specific variables such as\n\ the packets being used."), &remote_set_cmdlist, 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, 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, cli_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); set_show_commands remotebreak_cmds = 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); deprecate_cmd (remotebreak_cmds.set, "set remote interrupt-sequence"); deprecate_cmd (remotebreak_cmds.show, "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_zuinteger_unlimited_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 \"unlimited\" for unlimited hardware watchpoints."), NULL, show_hardware_watchpoint_limit, &remote_set_cmdlist, &remote_show_cmdlist); add_setshow_zuinteger_unlimited_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 \"unlimited\" to allow watchpoints of unlimited size."), NULL, show_hardware_watchpoint_length_limit, &remote_set_cmdlist, &remote_show_cmdlist); add_setshow_zuinteger_unlimited_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 \"unlimited\" for unlimited hardware breakpoints."), NULL, show_hardware_breakpoint_limit, &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 (PACKET_X, "X", "binary-download", 1); add_packet_config_cmd (PACKET_vCont, "vCont", "verbose-resume", 0); add_packet_config_cmd (PACKET_QPassSignals, "QPassSignals", "pass-signals", 0); add_packet_config_cmd (PACKET_QCatchSyscalls, "QCatchSyscalls", "catch-syscalls", 0); add_packet_config_cmd (PACKET_QProgramSignals, "QProgramSignals", "program-signals", 0); add_packet_config_cmd (PACKET_QSetWorkingDir, "QSetWorkingDir", "set-working-dir", 0); add_packet_config_cmd (PACKET_QStartupWithShell, "QStartupWithShell", "startup-with-shell", 0); add_packet_config_cmd (PACKET_QEnvironmentHexEncoded,"QEnvironmentHexEncoded", "environment-hex-encoded", 0); add_packet_config_cmd (PACKET_QEnvironmentReset, "QEnvironmentReset", "environment-reset", 0); add_packet_config_cmd (PACKET_QEnvironmentUnset, "QEnvironmentUnset", "environment-unset", 0); add_packet_config_cmd (PACKET_qSymbol, "qSymbol", "symbol-lookup", 0); add_packet_config_cmd (PACKET_P, "P", "set-register", 1); add_packet_config_cmd (PACKET_p, "p", "fetch-register", 1); add_packet_config_cmd (PACKET_Z0, "Z0", "software-breakpoint", 0); add_packet_config_cmd (PACKET_Z1, "Z1", "hardware-breakpoint", 0); add_packet_config_cmd (PACKET_Z2, "Z2", "write-watchpoint", 0); add_packet_config_cmd (PACKET_Z3, "Z3", "read-watchpoint", 0); add_packet_config_cmd (PACKET_Z4, "Z4", "access-watchpoint", 0); add_packet_config_cmd (PACKET_qXfer_auxv, "qXfer:auxv:read", "read-aux-vector", 0); add_packet_config_cmd (PACKET_qXfer_exec_file, "qXfer:exec-file:read", "pid-to-exec-file", 0); add_packet_config_cmd (PACKET_qXfer_features, "qXfer:features:read", "target-features", 0); add_packet_config_cmd (PACKET_qXfer_libraries, "qXfer:libraries:read", "library-info", 0); add_packet_config_cmd (PACKET_qXfer_libraries_svr4, "qXfer:libraries-svr4:read", "library-info-svr4", 0); add_packet_config_cmd (PACKET_qXfer_memory_map, "qXfer:memory-map:read", "memory-map", 0); add_packet_config_cmd (PACKET_qXfer_osdata, "qXfer:osdata:read", "osdata", 0); add_packet_config_cmd (PACKET_qXfer_threads, "qXfer:threads:read", "threads", 0); add_packet_config_cmd (PACKET_qXfer_siginfo_read, "qXfer:siginfo:read", "read-siginfo-object", 0); add_packet_config_cmd (PACKET_qXfer_siginfo_write, "qXfer:siginfo:write", "write-siginfo-object", 0); add_packet_config_cmd (PACKET_qXfer_traceframe_info, "qXfer:traceframe-info:read", "traceframe-info", 0); add_packet_config_cmd (PACKET_qXfer_uib, "qXfer:uib:read", "unwind-info-block", 0); add_packet_config_cmd (PACKET_qGetTLSAddr, "qGetTLSAddr", "get-thread-local-storage-address", 0); add_packet_config_cmd (PACKET_qGetTIBAddr, "qGetTIBAddr", "get-thread-information-block-address", 0); add_packet_config_cmd (PACKET_bc, "bc", "reverse-continue", 0); add_packet_config_cmd (PACKET_bs, "bs", "reverse-step", 0); add_packet_config_cmd (PACKET_qSupported, "qSupported", "supported-packets", 0); add_packet_config_cmd (PACKET_qSearch_memory, "qSearch:memory", "search-memory", 0); add_packet_config_cmd (PACKET_qTStatus, "qTStatus", "trace-status", 0); add_packet_config_cmd (PACKET_vFile_setfs, "vFile:setfs", "hostio-setfs", 0); add_packet_config_cmd (PACKET_vFile_open, "vFile:open", "hostio-open", 0); add_packet_config_cmd (PACKET_vFile_pread, "vFile:pread", "hostio-pread", 0); add_packet_config_cmd (PACKET_vFile_pwrite, "vFile:pwrite", "hostio-pwrite", 0); add_packet_config_cmd (PACKET_vFile_close, "vFile:close", "hostio-close", 0); add_packet_config_cmd (PACKET_vFile_unlink, "vFile:unlink", "hostio-unlink", 0); add_packet_config_cmd (PACKET_vFile_readlink, "vFile:readlink", "hostio-readlink", 0); add_packet_config_cmd (PACKET_vFile_fstat, "vFile:fstat", "hostio-fstat", 0); add_packet_config_cmd (PACKET_vAttach, "vAttach", "attach", 0); add_packet_config_cmd (PACKET_vRun, "vRun", "run", 0); add_packet_config_cmd (PACKET_QStartNoAckMode, "QStartNoAckMode", "noack", 0); add_packet_config_cmd (PACKET_vKill, "vKill", "kill", 0); add_packet_config_cmd (PACKET_qAttached, "qAttached", "query-attached", 0); add_packet_config_cmd (PACKET_ConditionalTracepoints, "ConditionalTracepoints", "conditional-tracepoints", 0); add_packet_config_cmd (PACKET_ConditionalBreakpoints, "ConditionalBreakpoints", "conditional-breakpoints", 0); add_packet_config_cmd (PACKET_BreakpointCommands, "BreakpointCommands", "breakpoint-commands", 0); add_packet_config_cmd (PACKET_FastTracepoints, "FastTracepoints", "fast-tracepoints", 0); add_packet_config_cmd (PACKET_TracepointSource, "TracepointSource", "TracepointSource", 0); add_packet_config_cmd (PACKET_QAllow, "QAllow", "allow", 0); add_packet_config_cmd (PACKET_StaticTracepoints, "StaticTracepoints", "static-tracepoints", 0); add_packet_config_cmd (PACKET_InstallInTrace, "InstallInTrace", "install-in-trace", 0); add_packet_config_cmd (PACKET_qXfer_statictrace_read, "qXfer:statictrace:read", "read-sdata-object", 0); add_packet_config_cmd (PACKET_qXfer_fdpic, "qXfer:fdpic:read", "read-fdpic-loadmap", 0); add_packet_config_cmd (PACKET_QDisableRandomization, "QDisableRandomization", "disable-randomization", 0); add_packet_config_cmd (PACKET_QAgent, "QAgent", "agent", 0); add_packet_config_cmd (PACKET_QTBuffer_size, "QTBuffer:size", "trace-buffer-size", 0); add_packet_config_cmd (PACKET_Qbtrace_off, "Qbtrace:off", "disable-btrace", 0); add_packet_config_cmd (PACKET_Qbtrace_bts, "Qbtrace:bts", "enable-btrace-bts", 0); add_packet_config_cmd (PACKET_Qbtrace_pt, "Qbtrace:pt", "enable-btrace-pt", 0); add_packet_config_cmd (PACKET_qXfer_btrace, "qXfer:btrace", "read-btrace", 0); add_packet_config_cmd (PACKET_qXfer_btrace_conf, "qXfer:btrace-conf", "read-btrace-conf", 0); add_packet_config_cmd (PACKET_Qbtrace_conf_bts_size, "Qbtrace-conf:bts:size", "btrace-conf-bts-size", 0); add_packet_config_cmd (PACKET_multiprocess_feature, "multiprocess-feature", "multiprocess-feature", 0); add_packet_config_cmd (PACKET_swbreak_feature, "swbreak-feature", "swbreak-feature", 0); add_packet_config_cmd (PACKET_hwbreak_feature, "hwbreak-feature", "hwbreak-feature", 0); add_packet_config_cmd (PACKET_fork_event_feature, "fork-event-feature", "fork-event-feature", 0); add_packet_config_cmd (PACKET_vfork_event_feature, "vfork-event-feature", "vfork-event-feature", 0); add_packet_config_cmd (PACKET_Qbtrace_conf_pt_size, "Qbtrace-conf:pt:size", "btrace-conf-pt-size", 0); add_packet_config_cmd (PACKET_vContSupported, "vContSupported", "verbose-resume-supported", 0); add_packet_config_cmd (PACKET_exec_event_feature, "exec-event-feature", "exec-event-feature", 0); add_packet_config_cmd (PACKET_vCtrlC, "vCtrlC", "ctrl-c", 0); add_packet_config_cmd (PACKET_QThreadEvents, "QThreadEvents", "thread-events", 0); add_packet_config_cmd (PACKET_no_resumed, "N stop reply", "no-resumed-stop-reply", 0); add_packet_config_cmd (PACKET_memory_tagging_feature, "memory-tagging-feature", "memory-tagging-feature", 0); /* Assert that we've registered "set remote foo-packet" 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_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 == (packets_descriptions[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_basic_prefix_cmd ("remote", class_files, _("\ Manipulate files on the remote system.\n\ Transfer files to and from the remote target system."), &remote_cmdlist, 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); add_setshow_string_noescape_cmd ("exec-file", class_files, &remote_exec_file_var, _("\ Set the remote pathname for \"run\"."), _("\ Show the remote pathname for \"run\"."), NULL, set_remote_exec_file, show_remote_exec_file, &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); add_setshow_zinteger_cmd ("watchdog", class_maintenance, &watchdog, _("\ Set watchdog timer."), _("\ Show watchdog timer."), _("\ When non-zero, this timeout is used instead of waiting forever for a target\n\ to finish a low-level step or continue operation. If the specified amount\n\ of time passes without a response from the target, an error occurs."), NULL, show_watchdog, &setlist, &showlist); add_setshow_zuinteger_unlimited_cmd ("remote-packet-max-chars", no_class, &remote_packet_max_chars, _("\ Set the maximum number of characters to display for each remote packet."), _("\ Show the maximum number of characters to display for each remote packet."), _("\ Specify \"unlimited\" to display all the characters."), NULL, show_remote_packet_max_chars, &setdebuglist, &showdebuglist); add_setshow_boolean_cmd ("remote", no_class, &remote_debug, _("Set debugging of remote protocol."), _("Show debugging of remote protocol."), _("\ When enabled, each packet sent or received with the remote target\n\ is displayed."), NULL, show_remote_debug, &setdebuglist, &showdebuglist); add_setshow_zuinteger_unlimited_cmd ("remotetimeout", no_class, &remote_timeout, _("\ Set timeout limit to wait for target to respond."), _("\ Show timeout limit to wait for target to respond."), _("\ This value is used to set the time limit for gdb to wait for a response\n\ from the target."), NULL, show_remote_timeout, &setlist, &showlist); /* Eventually initialize fileio. See fileio.c */ initialize_remote_fileio (&remote_set_cmdlist, &remote_show_cmdlist); #if GDB_SELF_TEST selftests::register_test ("remote_memory_tagging", selftests::test_memory_tagging_functions); #endif }