Fix entry-values in C++ across CUs.
* dwarf2loc.c (call_site_to_target_addr) <FIELD_LOC_KIND_PHYSNAME>: Use
lookup_minimal_symbol. Add a comment.
* dwarf2read.c
(read_call_site_scope) <is_ref_attr> <die_is_declaration>: Prefer
DW_AT_linkage_name.
gdb/testsuite/
Fix entry-values in C++ across CUs.
* gdb.arch/amd64-tailcall-cxx.exp: New file.
* gdb.arch/amd64-tailcall-cxx1.S: New file.
* gdb.arch/amd64-tailcall-cxx1.cc: New file.
* gdb.arch/amd64-tailcall-cxx2.S: New file.
* gdb.arch/amd64-tailcall-cxx2.cc: New file.
From: Pedro Alves <palves@redhat.com>
The find command's patter/buffer that is passed to the target is a
binary blob, not a string.
$ make WERROR_CFLAGS="-Wpointer-sign -Werror" findcmd.o -k 2>&1 1>/dev/null
../../src/gdb/findcmd.c: In function ‘find_command’:
../../src/gdb/findcmd.c:278:6: error: pointer targets in passing argument 3 of ‘target_search_memory’ differ in signedness [-Werror=pointer-sign]
In file included from ../../src/gdb/findcmd.c:26:0:
../../src/gdb/target.h:1582:12: note: expected ‘const gdb_byte *’ but argument is of type ‘char *’
gdb/
2013-03-08 Pedro Alves <palves@redhat.com>
* findcmd.c (put_bits): Change type of parameter to 'gdb_byte *'.
(parse_find_args, find_command): Change type of pattern buffer
locals to 'gdb_byte *'.
Hafiz Abid Qadeer <abidh@codesourcery.com>
gdb/
* NEWS: Mention set and show trace-buffer-size commands.
Mention new packet.
* target.h (struct target_ops): New method
to_set_trace_buffer_size.
(target_set_trace_buffer_size): New macro.
* target.c (update_current_target): Set up new method.
* tracepoint.c (trace_buffer_size): New global.
(start_tracing): Send it to the target.
(set_trace_buffer_size): New function.
(_initialize_tracepoint): Add new setshow for trace-buffer-size.
* remote.c (remote_set_trace_buffer_size): New function.
(_initialize_remote): Use it.
(QTBuffer:size) New remote command.
(PACKET_QTBuffer_size): New enum.
(remote_protocol_features): Add an entry for
PACKET_QTBuffer_size.
gdb/gdbserver/
* tracepoint.c (trace_buffer_size): New global.
(DEFAULT_TRACE_BUFFER_SIZE): New define.
(init_trace_buffer): Change to one-argument function. Allocate
trace buffer memory.
(handle_tracepoint_general_set): Call cmd_bigqtbuffer_size to
handle QTBuffer:size packet.
(cmd_bigqtbuffer_size): New function.
(initialize_tracepoint): Call init_trace_buffer with
DEFAULT_TRACE_BUFFER_SIZE.
* server.c (handle_query): Add QTBuffer:size in the
supported packets.
gdb/doc/
* gdb.texinfo (Starting and Stopping Trace Experiments): Document
trace-buffer-size set and show commands.
(Tracepoint Packets): Document QTBuffer:size.
(General Query Packets): Document QTBuffer:size.
gdb/testsuite/
* gdb.trace/trace-buffer-size.exp: New file.
* gdb.trace/trace-buffer-size.c: New file.
$ make WERROR_CFLAGS="-Wpointer-sign -Werror" target.o -k 2>&1 1>/dev/null
../../src/gdb/target.c: In function ‘target_read_stralloc’:
../../src/gdb/target.c:2376:3: error: pointer targets in passing argument 1 of ‘strlen’ differ in signedness [-Werror=pointer-sign]
In file included from build-gnulib/import/string.h:27:0,
from ../../src/gdb/common/gdb_string.h:24,
from ../../src/gdb/target.c:24:
/usr/include/string.h:399:15: note: expected ‘const char *’ but argument is of type ‘gdb_byte *’
...
This is about the same as the previous patch.
Functions that take or return ascii-ish string arguments usually use
char* for parameters/return. That means that at points we call into
target methods that work with binary blobs, we need casts to/from
gdb_byte*/char*. To choose which type for the variables, I usually go
based on which requires the fewer casts, and what the contents of the
variable are supposed to hold, which often gives the same answer.
gdb/
2013-03-07 Pedro Alves <palves@redhat.com>
* target.c (target_read_stralloc, target_fileio_read_alloc):
*Cast pointer to 'gdb_byte *' in target call.
$ make WERROR_CFLAGS="-Wpointer-sign -Werror" corefile.o -k 2>&1 1>/dev/null
../../src/gdb/corefile.c: In function ‘read_memory_string’:
../../src/gdb/corefile.c:334:7: error: pointer targets in passing argument 2 of ‘read_memory’ differ in signedness [-Werror=pointer-sign]
../../src/gdb/corefile.c:217:1: note: expected ‘gdb_byte *’ but argument is of type ‘char *’
Functions that take or return ascii-ish string arguments usually use
char* for parameters/return. That means that at points we call into
target methods that work with binary blobs, we need casts to
gdb_byte*.
gdb/
2013-03-07 Pedro Alves <palves@redhat.com>
* corefile.c (read_memory_string): Cast pointer to gdb_byte* in
call.
(trace_pass_command): Likewise.
* cli/cli-cmds.c: Include cli/cli-utils.h.
(source_command): Use skip-spaces.
(disassemble_command): Likewise.
* findcmd.c: Include cli/cli-utils.h.
(parse_find_args): Use skip_spaces.
* go32-nat.c: Include cli/cli-utils.h.
(go32_sldt): Use skip_spaces.
(go32_sgdt): Likewise.
(go32_sidt): Likewise.
(go32_pde): Likewise.
(go32_pte): Likewise.
(go32_pte_for_address): Likewise.
* infcmd.c: Include cli/cli-utils.h.
(registers_info): Use skip_spaces.
* linux-tdep.c (read_mapping): Use skip_spaces_const.
(linux_info_proc): Likewise.
* linux-thread-db.c: Include cli/cli-utils.h.
(info_auto_load_libthread_db): Use skip_spaces_const.
* m32r-rom.c: Include cli/cli-utils.h.
(m32r_upload_command): Use skip_spaces.
* maint.c: Include cli/cli-utils.h.
(maintenance_translate_address): Use skip_spaces.
* mi/mi-parse.c: Include cli/cli-utils.h.
(mi_parse_argv): Use skip_spaces.
(mi_parse): Likewise.
* minsyms.c: Include cli/cli-utils.h.
(msymbol_hash_iw): Use skip_spaces_const.
* objc-lang.c: Include cli/cli-utils.h.
(parse_selector): Use skip_spaces.
(parse_method): Likewise.
* python/python.c: Include cli/cli-utils.h.
(python_interactive_command)[HAVE_PYTHON]: Use skip_spaces.
(python_command)[HAVE_PYTHON]: Likewise.
(python_interactive_command)[!HAVE_PYTHON]: Likewise.
* remote-m32r-sdi.c: Include cli/cli-utils.h.
(m32r_load): Use skip_spaces.
* serial.c: Include cli/cli-utils.h.
(serial_open): Use skip_spaces_const.
* stack.c: Include cli/cli-utils.h.
(parse_frame_specification_1): Use skip_spaces_const.
* symfile.c: Include cli/cli-utils.h.
(set_ext_lang_command): Use skip_spaces.
* symtab.c: Include cli/cli-utils.h.
(rbreak_command): Use skip_spaces.
* thread.c (thread_name_command): Use skip_spaces.
* tracepoint.c (validate_actionline): Use skip_spaces.
(encode_actions_1): Likewise.
(trace_find_range_command): Likewise.
(trace_find_outside_command): Likewise.
(trace_dump_actions): Likewise.
Trimmed for brevity:
$ make WERROR_CFLAGS="-Wpointer-sign -Werror" c-lang.o expprint.o utils.o valprint.o varobj.o -k 2>&1 1>/dev/null
../../src/gdb/c-lang.c: In function ‘parse_one_string’:
../../src/gdb/c-lang.c:540:8: error: pointer targets in passing argument 3 of ‘convert_between_encodings’ differ in signedness [-Werror=pointer-sign]
In file included from ../../src/gdb/c-lang.c:30:0:
../../src/gdb/charset.h:64:6: note: expected ‘const gdb_byte *’ but argument is of type ‘char *’
../../src/gdb/expprint.c: In function ‘print_subexp_standard’:
../../src/gdb/expprint.c:205:2: error: pointer targets in passing argument 3 of ‘current_language->la_printstr’ differ in signedness [-Werror=pointer-sign]
../../src/gdb/expprint.c:205:2: note: expected ‘const gdb_byte *’ but argument is of type ‘char *’
cc1: all warnings being treated as errors
make: *** [expprint.o] Error 1
../../src/gdb/utils.c: In function ‘host_char_to_target’:
../../src/gdb/utils.c:1474:9: error: pointer targets in passing argument 3 of ‘convert_between_encodings’ differ in signedness [-Werror=pointer-sign]
../../src/gdb/varobj.c: In function ‘value_get_print_value’:
../../src/gdb/varobj.c:2934:8: error: pointer targets in return differ in signedness [-Werror=pointer-sign]
../../src/gdb/varobj.c:2968:12: error: pointer targets in assignment differ in signedness [-Werror=pointer-sign]
../../src/gdb/varobj.c:2971:3: error: pointer targets in return differ in signedness [-Werror=pointer-sign]
cc1: all warnings being treated as errors
make: *** [varobj.o] Error 1
As with the previous patch, the encoding conversion code works with
gdb_byte arrays as the generic buffers that hold strings of any
encoding/width. This patch adds casts where appropriate.
gdb/
2013-03-07 Pedro Alves <palves@redhat.com>
* c-lang.c (parse_one_string): Cast argument to gdb_byte *.
* expprint.c (print_subexp_standard): Likewise.
* utils.c (host_char_to_target): Likewise.
* valprint.c (generic_emit_char, generic_printstr): Likewise.
* varobj.c (value_get_print_value): Change type of local to char*.
Cast it gdb_byte * in call to language printer.
$ make WERROR_CFLAGS="-Wpointer-sign -Werror" charset.o 2>&1 1>/dev/null
../../src/gdb/charset.c: In function ‘wchar_iterate’:
../../src/gdb/charset.c:665:13: error: pointer targets in assignment differ in signedness [-Werror=pointer-sign]
../../src/gdb/charset.c:691:13: error: pointer targets in assignment differ in signedness [-Werror=pointer-sign]
../../src/gdb/charset.c:706:12: error: pointer targets in assignment differ in signedness [-Werror=pointer-sign]
The encoding conversion code works with gdb_byte arrays as the generic
buffers that hold strings of any encoding/width. Changing the type of
this field to gdb_byte* removes the need for one cast, and makes
everything work with the same types. That's good -- WRT to strings,
"char *" is (almost) consistently throughout GDB only used for
ascii-ish strings.
gdb/
2013-03-07 Pedro Alves <palves@redhat.com>
* charset.c (struct wchar_iterator) <input>: Change type to 'const
gdb_byte *'.
(make_wchar_iterator): Remove cast to char*.
(wchar_iterate): Change type of local.
$ make WERROR_CFLAGS="-Wpointer-sign -Werror" regcache.o 2>&1 1>/dev/null
../../src/gdb/regcache.c: In function ‘regcache_xmalloc_1’:
../../src/gdb/regcache.c:228:2: error: pointer targets in assignment differ in signedness [-Werror=pointer-sign]
../../src/gdb/regcache.c:235:2: error: pointer targets in assignment differ in signedness [-Werror=pointer-sign]
cc1: all warnings being treated as errors
regcache->register_status is "signed char".
gdb/
2013-03-07 Pedro Alves <palves@redhat.com>
* regcache.c (regcache_xmalloc_1): Call XCALLOC with signed char
for 'regcache->register_status'.
$ make WERROR_CFLAGS="-Wpointer-sign -Werror" breakpoint.o 2>&1 1>/dev/null
../../src/gdb/breakpoint.c: In function ‘breakpoint_xfer_memory’:
../../src/gdb/breakpoint.c:1578:2: error: pointer targets in passing argument 3 of ‘gdbarch_breakpoint_from_pc’ differ in signedness [-Werror=pointer-sign]
In file included from ../../src/gdb/defs.h:644:0,
from ../../src/gdb/breakpoint.c:20:
../../src/gdb/gdbarch.h:495:25: note: expected ‘int *’ but argument is of type ‘unsigned int *’
target_info.placed_size is an 'int', and gdbarch_breakpoint_from_pc
takes an int too.
gdb/
2013-03-07 Pedro Alves <palves@redhat.com>
* breakpoint.c.c (breakpoint_xfer_memory): Change type of local to
int.
$ make WERROR_CFLAGS="-Wpointer-sign -Werror" stap-probe.o 2>&1 1>/dev/null
../../src/gdb/stap-probe.c: In function ‘handle_stap_probe’:
../../src/gdb/stap-probe.c:1306:19: error: pointer targets in assignment differ in signedness [-Werror=pointer-sign]
cc1: all warnings being treated as errors
make: *** [stap-probe.o] Error 1
provider is a string, so it's rightfully a char*. 'data' holds raw
bytes (bfd_byte), so a cast is the right thing to do.
gdb/
2013-03-07 Pedro Alves <palves@redhat.com>
* stap-probe.c (handle_stap_probe): Add cast to char*.
$ make WERROR_CFLAGS="-Wpointer-sign -Werror" linux-record.o 2>&1 1>/dev/null
...
../../src/gdb/linux-record.c: In function ‘record_linux_system_call’:
../../src/gdb/linux-record.c:1152:9: error: pointer targets in passing argument 3 of ‘regcache_raw_read_signed’ differ in signedness [-Werror=pointer-sign]
In file included from ../../src/gdb/linux-record.c:23:0:
../../src/gdb/regcache.h:76:3: note: expected ‘long int *’ but argument is of type ‘long unsigned int *’
../../src/gdb/linux-record.c:1186:13: error: pointer targets in passing argument 3 of ‘regcache_raw_read_signed’ differ in signedness [-Werror=pointer-sign]
In file included from ../../src/gdb/linux-record.c:23:0:
../../src/gdb/regcache.h:76:3: note: expected ‘long int *’ but argument is of type ‘long unsigned int *’
I believe the read_signed calls are correct, and the variables are in
the wrong.
Tested on x86_64 Fedora 17.
gdb/
2013-03-07 Pedro Alves <palves@redhat.com>
* linux-record.c (record_linux_system_call) <gdb_sys_msgrcv,
RECORD_MSGRCV>: Pass a signed variable to
regcache_raw_read_signed, instead of an unsigned one.
tcp_retry_limit is installed as variable of an unsigned command:
add_setshow_uinteger_cmd ("connect-timeout", class_obscure,
&tcp_retry_limit, _("\
and I found no uses of the variable treating it as signed (like < 0
checks or some such).
2013-03-07 Pedro Alves <palves@redhat.com>
* ser-tcp.c (tcp_retry_limit): Change type to unsigned int.
Move the declarations to a header, rather than declaring them in
(multiple) .c files.
gdb/
2013-03-07 Pedro Alves <palves@redhat.com>
* remote.c (hex2bin, bin2hex): Move extern declarations to ...
* remote.h (hex2bin, bin2hex): ... here.
* tracepoint.c (hex2bin, bin2hex): Remove extern declarations.
The strlen in this function looked suspicious, for the function
documents that STR is in the target's charset (what if it is utf-32,
etc.).
On closer inspection, nothing is calling the function, and it's been
that way ever since the function was added.
gdb/
2013-03-06 Pedro Alves <palves@redhat.com>
* python/py-utils.c (target_string_to_unicode): Delete function.
* python/python-internal.h (target_string_to_unicode): Delete
declaration.
* cli/cli-decode.c (add_setshow_zuinteger_unlimited_cmd): Change
parameter VAR's type from "unsigned int" to "int".
* command.h (var_zuinteger_unlimited): Update its comments.
(add_setshow_zuinteger_unlimited_cmd): Update the declaration.
* linespec.c (struct linespec_canonical_name): New.
(struct linespec_state): Change canonical_names type to it.
(add_sal_to_sals): Change variable canonical_name to canonical. Change
xrealloc element size. Initialize the different CANONICAL fields.
(canonical_to_fullform): New.
(filter_results): Use it. Add variables canonical, fullform and
cleanup.
(struct decode_line_2_item, decode_line_2_compare_items): New.
(decode_line_2): Remove variables iter and item_names, add variables
items and items_count. Modify the code for these new variables.
gdb/testsuite/
* gdb.linespec/base/one/thefile.cc (twodup): New.
(m): Call it.
* gdb.linespec/base/two/thefile.cc (dupname): New.
(n): Call it.
* gdb.linespec/break-ask.exp: New file.
* gdb.linespec/lspec.cc (body_elsewhere): New comment marker.
gdb_byte should be used for bytes from the program being debugged. We
have many places using char or unsigned char instead all over the
existing ports, and more ends up added over time due to copy/paste as
new code is based on old code.
I've greped the tree for "char buf[", and fixed all I found.
Tested by building with --enable-targets=all.
2013-03-01 Pedro Alves <palves@redhat.com>
Use gdb_byte for bytes from the program being debugged.
* arm-tdep.c (arm_store_return_value, arm_get_longjmp_target):
Change type of local 'buf' to gdb_byte.
* avr-tdep.c (avr_frame_prev_register, avr_push_dummy_call): Likewise.
* bfin-tdep.c (bfin_push_dummy_call): Likewise.
* cris-tdep.c (cris_sigcontext_addr)
(cris_sigtramp_frame_unwind_cache): Likewise.
* frv-linux-tdep.c (frv_linux_pc_in_sigtramp)
(frv_linux_sigcontext_reg_addr, frv_linux_sigtramp_frame_cache):
Likewise.
* frv-tdep.c (frv_pseudo_register_write, frv_analyze_prologue): Likewise.
* hppa-hpux-tdep.c (hppa32_hpux_find_global_pointer)
(hppa32_hpux_search_dummy_call_sequence)
(hppa_hpux_supply_save_state): Likewise.
* hppa-linux-tdep.c (insns_match_pattern)
(hppa_linux_find_global_pointer): Likewise.
* hppa-tdep.c (hppa_in_function_epilogue_p)
(skip_prologue_hard_way, hppa_frame_cache): Likewise.
* i386-nto-tdep.c (i386nto_sigcontext_addr): Likewise.
* i386fbsd-tdep.c (i386fbsd_supply_uthread)
(i386fbsd_collect_uthread): Likewise.
* ia64-hpux-tdep.c (ia64_hpux_push_dummy_code): Likewise.
* ia64-linux-tdep.c (ia64_linux_sigcontext_register_address): Likewise.
* ia64-tdep.c (examine_prologue, ia64_frame_cache)
(ia64_frame_prev_register, ia64_sigtramp_frame_cache)
(ia64_sigtramp_frame_prev_register, ia64_access_reg)
(ia64_access_rse_reg, ia64_libunwind_frame_this_id)
(ia64_libunwind_frame_prev_register)
(ia64_libunwind_sigtramp_frame_this_id)
(ia64_find_global_pointer_from_dynamic_section)
(find_extant_func_descr, find_func_descr, ia64_dummy_id)
(ia64_unwind_pc): Likewise.
* iq2000-tdep.c (iq2000_store_return_value): Likewise.
* m68hc11-tdep.c (m68hc11_push_dummy_call)
(m68hc11_extract_return_value): Likewise.
* m68klinux-nat.c (fetch_register, store_register): Likewise.
* mep-tdep.c (mep_pseudo_cr32_read, mep_pseudo_cr32_write)
(mep_get_insn, mep_push_dummy_call): Likewise.
* mips-linux-tdep.c (mips_linux_get_longjmp_target)
(mips_linux_in_dynsym_stub): Likewise.
* mn10300-tdep.c (mep_pseudo_cr32_write): Likewise.
* ppc-linux-nat.c (fetch_register, store_register): Likewise.
* regcache.c (dump_endian_bytes): Change type of parameter 'buf'
to gdb_byte.
* remote-mips.c (mips_set_register): Likewise.
* remote-sim.c (gdbsim_fetch_register): Likewise.
* score-tdep.c (score7_fetch_inst): Change type of parameter
'memblock' and local 'buf' to gdb_byte.
(score7_malloc_and_get_memblock): Change return type to gdb_byte.
Change type of local 'buf' to gdb_byte. Adjust.
(score7_adjust_memblock_ptr): Change type of parameter 'memblock'
to gdb_byte**.
(score7_analyze_prologue): Change type of 'memblock' and
'memblock_ptr' locals to gdb_byte*.
* sh64-tdep.c (sh64_extract_return_value)
(sh64_store_return_value): Change type of local 'buf' to gdb_byte.
* solib-darwin.c (darwin_current_sos, darwin_read_exec_load_addr):
* solib-pa64.c (pa64_solib_create_inferior_hook)
(pa64_open_symbol_file_object): Remove local 'buf'.
* solib-som.c (som_solib_create_inferior_hook, link_map_start)
(som_open_symbol_file_object): Likewise.
* solib-spu.c (spu_current_sos): Likewise.
* spu-linux-nat.c (spu_fetch_inferior_registers): Likewise.
* spu-multiarch.c (parse_spufs_run, spu_fetch_registers)
(spu_store_registers): Likewise.
* target.c (debug_print_register): Likewise.
* tic6x-tdep.c (tic6x_get_longjmp_target): Likewise.
* xstormy16-tdep.c (xstormy16_store_return_value)
(xstormy16_push_dummy_call, xstormy16_resolve_jmp_table_entry)
(xstormy16_find_jmp_table_entry): Likewise.
printf-like functions to avoid type related warnings on all
platforms.
(handle_output_debug_string): Fetch context information address
from debug string using string_to_core_addr.
* regformats/reg-tilegx.dat (name): Change abi name to "tilegx".
* regformats/reg-tilegx32.dat: New.
gdbserver/
* Makefile.in (clean): Remove reg-tilegx.c, reg-tilegx32.c.
(reg-tilegx32.c): New rule.
* configure.srv (tilegx-*-linux*): Add reg-tilegx32.o to srv_regobj.
* linux-tile-low.c (tile_arch_setup): New function. Invoke
different register info initializer according to elf class.
(init_registers_tilgx32): New function. The tilegx32 register info
initializer.
(tile_fill_gregset): Use "uint_reg_t" to represent register size.
(tile_store_gregset): Likewise.
2013-02-27 Yao Qi <yao@codesourcery.com>
Pedro Alves <palves@redhat.com>
* tracepoint.c (tfile_trace_find): For tfind
pc/tp/range/outside, look for the next trace frame instead of
always starting from frame 0.
The "you have broken" bit of this text reads to me as if I had broken
it myself somehow. This patch eliminates that ambiguity.
2013-02-25 Pedro Alves <palves@redhat.com>
* common/linux-ptrace.c (linux_ptrace_test_ret_to_nx): Change
warning text.
symbol size.
* ppc64-tdep.c (ppc64_elf_make_msymbol_special): New function.
* ppc64-tdep.h (ppc64_elf_make_msymbol_special): Declare.
* ppc-linux-tdep.c (ppc_linux_init_abi): Set up to use the above.
* ppcfbsd-tdep.c (ppcfbsd_init_abi): Likewise.
Code cleanup.
* elfread.c (build_id_bfd_get): Make the return type const.
(build_id_verify): Make the check parameter const.
(build_id_to_debug_filename): Make the build_id parameter and variable
data const.
(find_separate_debug_file_by_buildid): Make the variable build_id const.
* python/py-arch.c (archpy_disassmble): Implementation of the
new method gdb.Architecture.disassemble.
(arch_object_methods): Add entry for the new method.
* doc/gdb.texinfo (Architectures In Python): Add description
about the new method gdb.Architecture.disassemble.
* testsuite/gdb.python/py-arch.c: New test case
* testsuite/gdb.python/py-arch.exp: New tests to test
gdb.Architecture.disassemble
* testsuite/gdb.python/Makefile.in: Add py-arch to the list of
EXECUTABLES.
While touching monitor_load in a previous patch, I noticed this method
is no longer set to anything other than NULL in the tree, so we can
remove it.
Tested by building with --enable-targets=all.
2013-02-19 Pedro Alves <palves@redhat.com>
Garbage collect 'struct monitor_ops'::load_routine.
* monitor.h (struct monitor_ops) <load_routine>: Remove field.
* monitor.c (monitor_load): No longer call
current_monitor->load_routine.
* dbug-rom.c (init_dbug_cmds): Don't set 'load_routine'.
* m32r-rom.c (init_m32r_cmds): Don't set 'load_routine'.
* ppcbug-rom.c (init_ppc_cmds): Don't set 'load_routine'.
Harmonize this old-looking code with generic_load, which fixes several
issues.
2013-02-19 Pedro Alves <palves@redhat.com>
PR gdb/15161
Harmonize with generic_load.
* monitor.c: Include "readline/readline.h".
(monitor_load): Rename parameter 'file' to 'args'. Use build_argv
instead of sscanf. Use CORE_ADDR/strtoulst instead of unsigned
long/strtol for the 'load_offset' local. Error out if no argument
is given or if too many arguments are given. Tilde expand the
passed in file name.
* symfile.c (load_section_data): Change type of load_offset
to CORE_ADDR.
(generic_load): User strtoulst instead of strtoul for conversion
of load_offset.
(tilegx_write_pc): New function.
(tilegx_cannot_reference_register): Return zero if REGNO
is TILEGX_FAULTNUM_REGNUM.
(tilegx_gdbarch_init): Add call to set_gdbarch_write_pc.
(tilegx_register_name): Add handling of "faultnum" register.
* tilegx-tdep.h (enum tilegx_regnum): Add TILEGX_FAULTNUM_REGNUM.
* tilegx-linux-tdep.c (tilegx_linux_supply_regset): Add
handling of TILEGX_FAULTNUM_REGNUM.
* tilegx-linux-nat.c (regmap): Add entry for TILEGX_FAULTNUM_REGNUM.
Hafiz Abid Qadeer <abidh@codesourcery.com>
gdb/
* NEWS: Mention new field "trace-file".
* tracepoint.c (trace_status_mi): Output "trace-file" field.
(tfile_open): Record the trace file's filename in the trace
status.
(tfile_files_info): Mention the name of the trace file.
Check the "filename" field explicitely.
(trace_status_command): Explicitely check "filename" field.
(trace_find_command): Ditto.
(trace_find_pc_command): Ditto.
(trace_find_tracepoint_command): Ditto.
(trace_find_line_command): Ditto.
(trace_find_range_command): Ditto.
(trace_find_outside_command): Ditto.
* tracepoint.h (struct trace_status) <from_file>: Rename it
to "filename" and make it hold the trace file's filename
instead of a boolean.
* remote.c (remote_get_trace_status): Initialize "filename"
field with NULL instead of 0.
gdb/doc/
* gdb.texinfo (GDB/MI Tracepoint Commands) <-trace-status>:
Document the "trace-file" field.
gdb/testsuite/
* gdb.trace/tfile.exp: Add test for -trace-status command.
This makes gdbserver share gdb's savestring, instead of baking its own.
Tested on x86_64 Fedora 17.
gdb/
2013-02-14 Pedro Alves <palves@redhat.com>
* utils.c (savestring): Don't #undef it. Move function to
common/common-utils.c.
* common/common-utils.c: Include gdb_string.h.
(savestring): Move here from utils.c.
* common/common-utils.h (savestring): Declare.
gdb/gdbserver/
2013-02-14 Pedro Alves <palves@redhat.com>
* tracepoint.c (save_string): Delete.
(add_tracepoint_action): Use savestring instead of save_string.
It's better to avoid needless confusion, and call string length,
length, instead of size, which is usually used to refer to sizeof of
the string (len+1):
size_t len = strlen (str);
size_t size = sizeof (str);
Tested on x86_64 Fedora 17.
2013-02-14 Pedro Alves <palves@redhat.com>
* utils.c (savestring): Rename parameter 'size' to 'len'.
Yufeng Zhang <yufeng.zhang@arm.com>
* aarch64-linux-nat.c (aarch64_init_debug_reg_state): Delete.
(aarch64_inferior_data, struct aarch64_inferior_data):
Delete.
(struct aarch64_process_info): New.
(aarch64_process_list): New global.
(aarch64_find_process_pid, aarch64_add_process)
(aarch64_process_info_get): New functions.
(aarch64_inferior_data_get): Delete.
(aarch64_process_info_get): New function.
(aarch64_forget_process): New function.
(aarch64_get_debug_reg_state): New parameter 'pid'. Reimplement.
(aarch64_linux_prepare_to_resume): Pass the lwp's pid to
aarch64_get_debug_reg_state.
(aarch64_notify_debug_reg_change): Use iterate_over_lwps
instead of linux_nat_iterate_watchpoint_lwps.
(aarch64_linux_new_fork): New function.
(aarch64_linux_child_post_startup_inferior): Use
aarch64_forget_process instead of aarch64_init_debug_reg_state.
(aarch64_handle_breakpoint, aarch64_linux_insert_hw_breakpoint)
(aarch64_linux_remove_hw_breakpoint)
(aarch64_handle_aligned_watchpoint)
(aarch64_handle_unaligned_watchpoint)
(aarch64_linux_insert_watchpoint)
(aarch64_linux_remove_watchpoint)
(aarch64_linux_stopped_data_address): Adjust to pass the current
process id to aarch64_debug_reg_state.
(_initialize_aarch64_linux_nat): Install aarch64_linux_new_fork as
linux_nat_new_fork hook, and aarch64_forget_process as
linux_nat_forget_process hook; remove the call to
register_inferior_data_with_cleanup.
I happened to notice a bug with ptype &Ref, and found out userdef.exp
actually exercises the bug. With:
class Container
{
public:
Member m;
Member& operator* ();
};
Member& Container::operator* ()
{
return this->m;
}
And 'c' is of type Container:
(gdb) p c
$1 = {m = {z = -9192}}
(gdb) p *c
$2 = (Member &) @0x7fffffffda20: {z = -9192}
(gdb) ptype *c
type = class Member {
public:
int z;
} &
(gdb) p &*c
$3 = (Member *) 0x7fffffffda20
(gdb) ptype &*c
type = class Member {
public:
int z;
} &*
(gdb)
Notice that last print (&*c) on says the type is a pointer - that's
how you get the address behind a reference. But notice the last ptype
instead says the type of the same expression is a pointer _reference_.
This looks like a bug to me.
This patch fixes it. The issue is that we're entering the VALUE_LVAL
(x) == lval_memory branch by mistake for references. The fix is just
to swap the tests so references are checked first, like value_addr
also handles references first.
Tested on x86_64 Fedora 17.
2013-02-14 Pedro Alves <palves@redhat.com>
* eval.c (evaluate_subexp_for_address) <default_case_after_eval,
EVAL_AVOID_SIDE_EFFECTS>: Swap and handle TYPE_CODE_REF before
lval_memory.
2013-02-14 Pedro Alves <palves@redhat.com>
* gdb.cp/userdef.exp (ptype &*c): Don't expect an &.
Hafiz Abid Qadeer <abidh@codesourcery.com>
gdb/
* tracepoint.h (validate_trace_state_variable_name): Declare.
* tracepoint.c (validate_trace_state_variable_name): New.
(trace_variable_command): Parse the trace state variable's name
without using parse_expression. Do several validations.
* mi/mi-main.c (mi_cmd_trace_define_variable): Don't parse the
trace state variable's name with parse_expression. Validate it.
gdb/testsuite/
* gdb.trace/tsv.exp: Adjust tests, and add a few more.
While reviewing the native AArch64 patch, I noticed a problem:
On 02/06/2013 08:46 PM, Pedro Alves wrote:
>
>> > +static void
>> > +aarch64_linux_prepare_to_resume (struct lwp_info *lwp)
>> > +{
>> > + struct arch_lwp_info *info = lwp->arch_private;
>> > +
>> > + /* NULL means this is the main thread still going through the shell,
>> > + or, no watchpoint has been set yet. In that case, there's
>> > + nothing to do. */
>> > + if (info == NULL)
>> > + return;
>> > +
>> > + if (DR_HAS_CHANGED (info->dr_changed_bp)
>> > + || DR_HAS_CHANGED (info->dr_changed_wp))
>> > + {
>> > + int tid = GET_LWP (lwp->ptid);
>> > + struct aarch64_debug_reg_state *state = aarch64_get_debug_reg_state ();
> Hmm. This is always fetching the debug_reg_state of
> the current inferior, but may not be the inferior of lwp.
> I see the same bug on x86. Sorry about that. I'll fix it.
A natural fix would be to make xxx_get_debug_reg_state take an
inferior argument, but that doesn't work because of the case where we
detach breakpoints/watchpoints from the child fork, at a time there's
no inferior for the child fork at all. We do a nasty hack in
i386_inferior_data_get, but that relies on all callers pointing the
current inferior to the correct inferior, which isn't actually being
done by all callers, and I don't think we want to enforce that -- deep
in the bowls of linux-nat.c, there are many cases we resume lwps
behind the scenes, and it's be better to not have that code rely on
global state (as it doesn't today).
The fix is to decouple the watchpoints code from inferiors, making it
track target processes instead. This way, we can freely keep track of
the watchpoint mirrors for these processes behind the core's back.
Checkpoints also play dirty tricks with swapping the process behind
the inferior, so they get special treatment too in the patch (which
just amounts to calling a new hook). Instead of the old hack in
i386_inferior_data_get, where we returned a copy of the current
inferior's debug registers mirror, as soon as we detect a fork in the
target, we copy the debug register mirror from the parent to the child
process.
I don't have an old kernel handy to test, but I stepped through gdb doing
the watchpoint removal in the fork child in the watchpoint-fork test
seeing that the debug registers end up cleared in the child.
I didn't find the need for linux_nat_iterate_watchpoint_lwps. If
we use plain iterate_over_lwps instead, what happens is that
when removing watchpoints, that iterate_over_lwps doesn't actually
iterate over anything, since the fork child is not added to the
lwp list until later, at detach time, in linux_child_follow_fork.
And if we don't iterate over that lwp, we don't mark its debug
registers as needing update. But linux_child_follow_fork takes
care of doing that explicitly:
child_lp = add_lwp (inferior_ptid);
child_lp->stopped = 1;
child_lp->last_resume_kind = resume_stop;
make_cleanup (delete_lwp_cleanup, child_lp);
/* CHILD_LP has new PID, therefore linux_nat_new_thread is not called for it.
See i386_inferior_data_get for the Linux kernel specifics.
Ensure linux_nat_prepare_to_resume will reset the hardware debug
registers. It is done by the linux_nat_new_thread call, which is
being skipped in add_lwp above for the first lwp of a pid. */
gdb_assert (num_lwps (GET_PID (child_lp->ptid)) == 1);
if (linux_nat_new_thread != NULL)
linux_nat_new_thread (child_lp);
if (linux_nat_prepare_to_resume != NULL)
linux_nat_prepare_to_resume (child_lp);
ptrace (PTRACE_DETACH, child_pid, 0, 0);
so unless I'm missing something (quite possible) it ends up all
the same. But, the !detach-on-fork, and the "follow-fork child" paths
should also call linux_nat_new_thread, and they don't presently. It
seems to me in those cases we're not clearing debug regs correctly
when that's needed. Instead of copying that bit that works around
add_lwp bypassing the linux_nat_new_thread call, I thought it'd
be better to add an add_initial_lwp call to be used in the case we
really need to bypass linux_nat_new_thread, and make
add_lwp always call linux_nat_new_thread.
i386_cleanup_dregs is rewritten to forget about the current process
debug mirrors, which takes cares of other i386 ports. Only a couple
of extra tweaks here and there were needed, as some targets wheren't
actually calling i386_cleanup_dregs.
Tested on Fedora 17 x86_64 -m64/-m32.
GDBserver already fetches the i386_debug_reg_state from the right
process, and, it doesn't handle forks at all, so no fix is needed over
there.
gdb/
2013-02-13 Pedro Alves <palves@redhat.com>
* amd64-linux-nat.c (update_debug_registers_callback):
Update comment.
(amd64_linux_dr_set_control, amd64_linux_dr_set_addr): Use
iterate_over_lwps.
(amd64_linux_prepare_to_resume): Pass the lwp's pid to
i386_debug_reg_state.
(amd64_linux_new_fork): New function.
(_initialize_amd64_linux_nat): Install amd64_linux_new_fork as
linux_nat_new_fork hook, and i386_forget_process as
linux_nat_forget_process hook.
* i386-linux-nat.c (update_debug_registers_callback):
Update comment.
(amd64_linux_dr_set_control, amd64_linux_dr_set_addr): Use
iterate_over_lwps.
(i386_linux_prepare_to_resume): Pass the lwp's pid to
i386_debug_reg_state.
(i386_linux_new_fork): New function.
(_initialize_i386_linux_nat): Install i386_linux_new_fork as
linux_nat_new_fork hook, and i386_forget_process as
linux_nat_forget_process hook.
* i386-nat.c (i386_init_dregs): Delete.
(i386_inferior_data, struct i386_inferior_data):
Delete.
(struct i386_process_info): New.
(i386_process_list): New global.
(i386_find_process_pid, i386_add_process, i386_process_info_get):
New functions.
(i386_inferior_data_get): Delete.
(i386_process_info_get): New function.
(i386_debug_reg_state): New parameter 'pid'. Reimplement.
(i386_forget_process): New function.
(i386_cleanup_dregs): Rewrite.
(i386_update_inferior_debug_regs, i386_insert_watchpoint)
(i386_remove_watchpoint, i386_region_ok_for_watchpoint)
(i386_stopped_data_address, i386_insert_hw_breakpoint)
(i386_remove_hw_breakpoint): Adjust to pass the current process id
to i386_debug_reg_state.
(i386_use_watchpoints): Don't register inferior data.
* i386-nat.h (i386_debug_reg_state): Add new 'pid' parameter, and
adjust comment.
(i386_forget_process): Declare.
* linux-fork.c (delete_fork): Call linux_nat_forget_process.
* linux-nat.c (linux_nat_new_fork, linux_nat_forget_process_hook):
New static globals.
(linux_child_follow_fork): Don't call linux_nat_new_thread here.
(add_initial_lwp): New, factored out from ...
(add_lwp): ... this. Don't check the number of lwps before
calling linux_nat_new_thread.
(linux_nat_iterate_watchpoint_lwps): Delete.
(linux_nat_attach): Use add_initial_lwp instead of add_lwp.
(linux_handle_extended_wait): Call the linux_nat_new_fork hook on
forks and vforks.
(linux_nat_wait_1): Use add_initial_lwp instead of add_lwp for the
initial lwp.
(linux_nat_kill, linux_nat_mourn_inferior): Call
linux_nat_forget_process.
(linux_nat_set_new_fork, linux_nat_set_forget_process)
(linux_nat_forget_process): New functions.
* linux-nat.h (linux_nat_iterate_watchpoint_lwps_ftype): Delete
type.
(linux_nat_iterate_watchpoint_lwps): Delete declaration.
(linux_nat_new_fork_ftype, linux_nat_forget_process_ftype): New
types.
(linux_nat_set_new_fork, linux_nat_set_forget_process)
(linux_nat_forget_process): New declarations.
* amd64fbsd-nat.c (super_mourn_inferior): New global.
(amd64fbsd_mourn_inferior): New function.
(_initialize_amd64fbsd_nat): Override to_mourn_inferior.
* windows-nat.c (windows_detach): Call i386_cleanup_dregs.
2013-02-13 Marcus Shawcroft <marcus.shawcroft@arm.com>
* aarch64-linux-nat.c (debug_reg_change_callback)
(aarch64_linux_get_debug_reg_capacity): ARI fix: Replace %llx with
%s and phex().
* c-exp.y (lex_one_token): Initialize other fields of yylval on
NAME return.
(classify_inner_name): Remove 'first_name' argument, add
'context'. Remove unused variable.
(yylex): Explicitly maintain the context type. Exit loop earlier
if NAME result is seen.
gdb/testsuite
* gdb.cp/m-static.cc (gnu_obj_1::~gnu_obj_1): New destructor.
* gdb.cp/m-static.exp: Add tests to print quoted destructor.
Some files managed to get in the tree with outdated copyright years.
This fixes it. Applied.
gdb/
2013-02-12 Pedro Alves <palves@redhat.com>
* break-catch-sig.c: Update copyright years.
gdb/testsuite/
2013-02-12 Pedro Alves <palves@redhat.com>
* gdb.base/catch-signal.c: Update copyright years.
* gdb.base/catch-signal.exp: Update copyright years.
* gdb.dwarf2/dw2-dir-file-name.c: Update copyright years.
* gdb.dwarf2/dw2-dir-file-name.exp: Update copyright years.
* gdb.dwarf2/dw2-empty-pc-range.S: Update copyright years.
* gdb.dwarf2/dw2-error.S: Update copyright years.
* gdb.dwarf2/dw2-error.c: Update copyright years.
* gdb.dwarf2/dw2-restrict.S: Update copyright years.
* gdb.dwarf2/dw2-restrict.c: Update copyright years.
* gdb.dwarf2/dw2-restrict.exp: Update copyright years.
provide a ui_out destructor.
* ui-out.h: Declare the new ui_out destructor.
(ui_out_impl): Add a field for data destructor in ui_out_impl.
* ui-out.c (default_data_destroy): Add a default data destructor
which does nothing.
(default_ui_out_impl): Set the new data_destroy field to
default_data_destroy
(uo_data_destroy): Local function which invokes the data
destructor if present.
(clear_table): Local function which clears the table data of a
ui_out object.
(ui_out_destroy): Public function which frees a ui_out object.
(ui_out_table_end): Use the new clear_table function.
* cli-out.c (cli_ui_out_impl): Set the new data_destroy field to
NULL.
* mi/mi-out.c (mi_ui_out_impl): Set the new data_destroy field
to NULL.
(printf_decfloat): New function. Broken out from ui_printf.
Remove unnecessary code to shift the entire format string down.
(printf_pointer): New function.
(ui_printf): Code to print C strings, wide C strings, decfloats,
and pointers moved to separate functions.
Jan Kratochvil
Hafiz Abid Qadeer
Yao Qi
Pedro Alves
Tom Tromey
Keith Seitz
Eli Zaretskii
Doug Evans
Pierre Muller
Corinna Vinschen
Hui Zhu
Kevin Buettner
Jiong Wang
Anthony Green
Maciej W. Rozycki
Alan Modra
Siva Chandra Reddy
Kai Tietz
Walter Lee
Lei Liu
Yufeng Zhang
Marcus Shawcroft
Sergio Durigan Junior