* arm.cc (Arm_input_section::Arm_input_section): For a
SHT_ARM_EXIDX section, always keeps the input sections.
(Arm_input_section::set_exidx_section_link): New method.
(Arm_exidx_input_section::Arm_exidx_input_section): Initialize
has_errors_ to false.
(Arm_exidx_input_section::has_errors,
Arm_exidx_input_section::set_has_errors): New methods.
(Arm_exidx_input_section::has_errors_): New data member.
(Arm_relobj::get_exidx_shndx_list): New method.
(Arm_output_section::append_text_sections_to_list): Do not skip
section without SHF_EXECINSTR.
(Arm_output_section::fix_exidx_coverage): Skip input sections with
errors.
(Arm_relobj::make_exidx_input_section): Add new parameter for text
section header. Make error messages more verbose. Check for
a non-executable section linked to an EXIDX section.
(Arm_relobj::do_read_symbols): Remove error checking, which has been
moved to Arm_relobj::make_exidx_input_section. Add an assertion to
check that there is no deferred EXIDX section if we exit early.
Instead of not making an EXIDX section in case of an error, make one
and set the has_errors flag of it.
(Target_arm::do_finalize_sections): Fix up links of EXIDX sections
in a relocatable link.
(Target_arm::do_relax): Look for the EXIDX output section instead of
assuming that it is called .ARM.exidx.
(Target_arm::fix_exidx_coverage): Add a new parameter for input
section list. Do not check for SHF_EXECINSTR section flags but
skip any input section with errors.
* output.cc (Output_section::Output_section): Initialize
always_keeps_input_sections_ to false.
(Output_section::add_input_section): Check for
always_keeps_input_sections_.
* output.h (Output_section::always_keeps_input_sections,
Output_section::set_always_keeps_input_sections): New methods.
(Output_section::always_keeps_input_sections): New data member.
weak when resolving to a dynamic def.
(Symbol_table::should_override): Add adjust_dyndef flag; set it
for weak undef/dynamic def cases. Adjust callers.
* symtab.cc (Symbol::init_fields): Initialize undef_binding_set_ and
undef_binding_weak_.
(Symbol_table::sized_write_globals): Adjust symbol binding.
(Symbol_table::sized_write_symbol): Add binding parameter.
* symtab.h (Symbol::set_undef_binding): New method.
(Symbol::is_undef_binding_weak): New method.
(Symbol::undef_binding_set_, Symbol::undef_binding_weak_): New members.
(Symbol_table::should_override): Add new parameter.
(Symbol_table::sized_write_symbol): Add new parameter.
* testsuite/weak_undef_file1.cc: Add new test case.
* testsuite/weak_undef_file2.cc: Fix header comment.
* testsuite/weak_undef_test.cc: Add new test case.
This is just a copy of the Python 2.7 version of python-config.py.
Tested with versions 2.5 and 2.6. It's nearly identical to the
previous version, except it's written in a more pythonic way,
so it should be fine for 2.4 as well.
gdb/ChangeLog:
* python/python-config.py: Resync with Python 2.7 version of this
script.
Thiago Jung Bauermann <bauerman@br.ibm.com>
Support for hw accelerated condition watchpoints in booke powerpc.
* breakpoint.c (fetch_watchpoint_value): Rename to fetch_subexp_value
and move to eval.c. Change callers.
(insert_bp_location): Pass watchpoint condition in
target_insert_watchpoint.
(remove_breakpoint_1) Pass watchpoint condition in
target_remove_watchpoint.
(watchpoint_locations_match): Call
target_can_accel_watchpoint_condition.
* eval.c: Include wrapper.h.
(fetch_subexp_value): Moved from breakpoint.c.
* ppc-linux-nat.c (ppc_linux_region_ok_for_hw_watchpoint):
Formatting fix.
(can_use_watchpoint_cond_accel): New function.
(calculate_dvc): Likewise.
(num_memory_accesses): Likewise.
(check_condition): Likewise.
(ppc_linux_can_accel_watchpoint_condition): Likewise
(ppc_linux_insert_watchpoint): Call can_use_watchpoint_cond_accel,
check_condition and calculate_dvc.
(ppc_linux_remove_watchpoint): Likewise.
(_initialize_ppc_linux_nat): Set to_can_accel_watchpoint_condition to
ppc_linux_can_accel_watchpoint_condition
* target.c (debug_to_insert_watchpoint): Add argument for watchpoint
condition.
(debug_to_remove_watchpoint): Likewise.
(debug_to_can_accel_watchpoint_condition): New function.
(update_current_target): Set to_can_accel_watchpoint_condition.
(setup_target_debug): Set to_can_accel_watchpoint_condition.
* target.h: Add opaque declaration for struct expression.
(struct target_ops) <to_insert_watchpoint>,
<to_remove_watchpoint>: Add new arguments to pass the watchpoint
<to_can_accel_watchpoint_condition>: New member.
condition. Update all callers and implementations.
(target_can_accel_watchpoint_condition): New macro.
* value.c (free_value_chain): New function.
* value.h (fetch_subexp_value): New prototype.
(free_value_chain): Likewise.
This is on GNU/Linux.
GDBserver does not exit properly when the inferior exits, as demonstrated
with any program using the procedure below:
% gdbserver-head :4444 simple_main
Process simple_main created; pid = 25681
Listening on port 4444
Then, in another terminal, start GDB, connect to GDBserver, and run
the program to completion:
% gdb-head simple_main
(gdb) tar rem :4444
(gdb) cont
Continuing.
Program exited normally.
Going back to the terminal where GDBserver is running, we see the following
output:
Child exited with status 0
readchar: Got EOF
Remote side has terminated connection. GDBserver will reopen the connection.
Listening on port 4444
The problem is that we're missing a call to mourn_inferior. As a result,
after we've handled the vCont packet, we fail to notice that there are
no process left to debug (target_running() returns true), and thus try
to continue reading from the remote socket. However, since GDB just
disconnected after having received the "exit with status 0" reply to the
vCont request, the read triggers the EOF exception.
gdb/gdbserver/ChangeLog:
* server.c (handle_v_cont): Call mourn_inferior if process
just exited.
(myresume): Likewise.
