As Pedro suggested on gdb-patches@ (see
https://sourceware.org/ml/gdb-patches/2015-05/msg00714.html), this
change makes symbol lookup functions return a structure that includes
both the symbol found and the block in which it was found. This makes
it possible to get rid of the block_found global variable and thus makes
block hunting explicit.
gdb/
* ada-exp.y (write_object_renaming): Replace struct
ada_symbol_info with struct block_symbol. Update field
references accordingly.
(block_lookup, select_possible_type_sym): Likewise.
(find_primitive_type): Likewise. Also update call to
ada_lookup_symbol to extract the symbol itself.
(write_var_or_type, write_name_assoc): Likewise.
* ada-lang.h (struct ada_symbol_info): Remove.
(ada_lookup_symbol_list): Replace struct ada_symbol_info with
struct block_symbol.
(ada_lookup_encoded_symbol, user_select_syms): Likewise.
(ada_lookup_symbol): Return struct block_symbol instead of a
mere symbol.
* ada-lang.c (defns_collected): Replace struct ada_symbol_info
with struct block_symbol.
(resolve_subexp, ada_resolve_function, sort_choices,
user_select_syms, is_nonfunction, add_defn_to_vec,
num_defns_collected, defns_collected,
symbols_are_identical_enums, remove_extra_symbols,
remove_irrelevant_renamings, add_lookup_symbol_list_worker,
ada_lookup_symbol_list, ada_iterate_over_symbols,
ada_lookup_encoded_symbol, get_var_value): Likewise.
(ada_lookup_symbol): Return a block_symbol instead of a mere
symbol. Replace struct ada_symbol_info with struct
block_symbol.
(ada_lookup_symbol_nonlocal): Likewise.
(standard_lookup): Make block passing explicit through
lookup_symbol_in_language.
* ada-tasks.c (get_tcb_types_info): Update the calls to
lookup_symbol_in_language to extract the mere symbol out of the
returned value.
(ada_tasks_inferior_data_sniffer): Likewise.
* ax-gdb.c (gen_static_field): Likewise for the call to
lookup_symbol.
(gen_maybe_namespace_elt): Deal with struct symbol_in_block from
lookup functions.
(gen_expr): Likewise.
* c-exp.y: Likewise. Remove uses of block_found.
(lex_one_token, classify_inner_name, c_print_token): Likewise.
(classify_name): Likewise. Rename the "sym" local variable to
"bsym".
* c-valprint.c (print_unpacked_pointer): Likewise.
* compile/compile-c-symbols.c (convert_symbol_sym): Promote the
"sym" parameter from struct symbol * to struct block_symbol.
Use it to remove uses of block_found. Deal with struct
symbol_in_block from lookup functions.
(gcc_convert_symbol): Likewise. Update the call to
convert_symbol_sym.
* compile/compile-object-load.c (compile_object_load): Deal with
struct symbol_in_block from lookup functions.
* cp-namespace.c (cp_lookup_nested_symbol_1,
cp_lookup_nested_symbol, cp_lookup_bare_symbol,
cp_search_static_and_baseclasses,
cp_lookup_symbol_in_namespace, cp_lookup_symbol_via_imports,
cp_lookup_symbol_imports_or_template,
cp_lookup_symbol_via_all_imports, cp_lookup_symbol_namespace,
lookup_namespace_scope, cp_lookup_nonlocal,
find_symbol_in_baseclass): Return struct symbol_in_block instead
of mere symbols and deal with struct symbol_in_block from lookup
functions.
* cp-support.c (inspect_type, replace_typedefs,
cp_lookup_rtti_type): Deal with struct symbol_in_block from
lookup functions.
* cp-support.h (cp_lookup_symbol_nonlocal,
cp_lookup_symbol_from_namespace,
cp_lookup_symbol_imports_or_template, cp_lookup_nested_symbol):
Return struct symbol_in_block instead of mere symbols.
* d-exp.y (d_type_from_name, d_module_from_name, push_variable,
push_module_name):
Deal with struct symbol_in_block from lookup functions. Remove
uses of block_found.
* eval.c (evaluate_subexp_standard): Update call to
cp_lookup_symbol_namespace.
* f-exp.y: Deal with struct symbol_in_block from lookup
functions. Remove uses of block_found.
(yylex): Likewise.
* gdbtypes.c (lookup_typename, lookup_struct, lookup_union,
lookup_enum, lookup_template_type, check_typedef): Deal with
struct symbol_in_block from lookup functions.
* guile/scm-frame.c (gdbscm_frame_read_var): Likewise.
* guile/scm-symbol.c (gdbscm_lookup_symbol): Likewise.
(gdbscm_lookup_global_symbol): Likewise.
* gnu-v3-abi.c (gnuv3_get_typeid_type): Likewise.
* go-exp.y: Likewise. Remove uses of block_found.
(package_name_p, classify_packaged_name, classify_name):
Likewise.
* infrun.c (insert_exception_resume_breakpoint): Likewise.
* jv-exp.y (push_variable): Likewise.
* jv-lang.c (java_lookup_class, get_java_object_type): Likewise.
* language.c (language_bool_type): Likewise.
* language.h (struct language_defn): Update
la_lookup_symbol_nonlocal to return a struct symbol_in_block
rather than a mere symbol.
* linespec.c (find_label_symbols): Deal with struct
symbol_in_block from lookup functions.
* m2-exp.y: Likewise. Remove uses of block_found.
(yylex): Likewise.
* mi/mi-cmd-stack.c (list_args_or_locals): Likewise.
* objc-lang.c (lookup_struct_typedef, find_imps): Likewise.
* p-exp.y: Likewise. Remove uses of block_found.
(yylex): Likewise.
* p-valprint.c (pascal_val_print): Likewise.
* parse.c (write_dollar_variable): Likewise. Remove uses of
block_found.
* parser-defs.h (struct symtoken): Turn the SYM field into a
struct symbol_in_block.
* printcmd.c (address_info): Deal with struct symbol_in_block
from lookup functions.
* python/py-frame.c (frapy_read_var): Likewise.
* python/py-symbol.c (gdbpy_lookup_symbol,
gdbpy_lookup_global_symbol): Likewise.
* skip.c (skip_function_command): Likewise.
* solib-darwin.c (darwin_lookup_lib_symbol): Return a struct
symbol_in_block instead of a mere symbol.
* solib-spu.c (spu_lookup_lib_symbol): Likewise.
* solib-svr4.c (elf_lookup_lib_symbol): Likewise.
* solib.c (solib_global_lookup): Likewise.
* solist.h (solib_global_lookup): Likewise.
(struct target_so_ops): Update lookup_lib_global_symbol to
return a struct symbol_in_block rather than a mere symbol.
* source.c (select_source_symtab): Deal with struct
symbol_in_block from lookup functions.
* stack.c (print_frame_args, iterate_over_block_arg_vars):
Likewise.
* symfile.c (set_initial_language): Likewise.
* symtab.c (SYMBOL_LOOKUP_FAILED): Turn into a struct
symbol_in_block.
(SYMBOL_LOOKUP_FAILED_P): New predicate as a macro.
(struct symbol_cache_slot): Turn the FOUND field into a struct
symbol_in_block.
(block_found): Remove.
(eq_symbol_entry): Update to deal with struct symbol_in_block in
cache slots.
(symbol_cache_lookup): Return a struct symbol_in_block rather
than a mere symbol.
(symbol_cache_mark_found): Add a BLOCK parameter to fill
appropriately the cache slots. Update callers.
(symbol_cache_dump): Update cache slots handling to the type
change.
(lookup_symbol_in_language, lookup_symbol, lookup_language_this,
lookup_symbol_aux, lookup_local_symbol,
lookup_symbol_in_objfile, lookup_global_symbol_from_objfile,
lookup_symbol_in_objfile_symtabs,
lookup_symbol_in_objfile_from_linkage_name,
lookup_symbol_via_quick_fns, basic_lookup_symbol_nonlocal,
lookup_symbol_in_static_block, lookup_static_symbol,
lookup_global_symbol):
Return a struct symbol_in_block rather than a mere symbol. Deal
with struct symbol_in_block from other lookup functions. Remove
uses of block_found.
(lookup_symbol_in_block): Remove uses of block_found.
(struct global_sym_lookup_data): Turn the RESULT field into a
struct symbol_in_block.
(lookup_symbol_global_iterator_cb): Update references to the
RESULT field.
(search_symbols): Deal with struct symbol_in_block from lookup
functions.
* symtab.h (struct symbol_in_block): New structure.
(block_found): Remove.
(lookup_symbol_in_language, lookup_symbol,
basic_lookup_symbol_nonlocal, lookup_symbol_in_static_block,
looku_static_symbol, lookup_global_symbol,
lookup_symbol_in_block, lookup_language_this,
lookup_global_symbol_from_objfile): Return a struct
symbol_in_block rather than just a mere symbol. Update comments
to remove mentions of block_found.
* valops.c (find_function_in_inferior,
value_struct_elt_for_reference, value_maybe_namespace_elt,
value_of_this): Deal with struct symbol_in_block from lookup
functions.
* value.c (value_static_field, value_fn_field): Likewise.
This change should have been in the previous patch (Mostly trivial enum
fixes).
gdb/ChangeLog:
* remote-m32r-sdi.c (m32r_remove_watchpoint): Use enum type
instead of integer.
2015-07-30 Sandra Loosemore <sandra@codesourcery.com>
gdb/
* nios2-tdep.c (nios2_analyze_prologue): Do what the comment
already says and disallow non-stack memory writes in the prologue.
2015-07-30 Sandra Loosemore <sandra@codesourcery.com>
gdb/
* nios2-tdep.c (nios2_analyze_prologue): Update comments to
reflect how current GCC emits stack overflow checks. Match
both trap and break instructions for backward compatibility.
Disallow other trap and break instructions in the prologue.
The new gdb.threads/fork-plus-threads.exp test exposes one more
problem. When one types "info inferiors" after running the program,
one see's a couple inferior left still, while there should only be
inferior #1 left. E.g.:
(gdb) info inferiors
Num Description Executable
4 process 8393 /home/pedro/bugs/src/test
2 process 8388 /home/pedro/bugs/src/test
* 1 <null> /home/pedro/bugs/src/test
(gdb) info threads
Calling prune_inferiors() manually at this point (from a top gdb) does
not remove them, because they still have inf->pid != 0 (while they
shouldn't). This suggests that we never mourned those inferiors.
Enabling logs (master + previous patch) we see:
...
WL: waitpid Thread 0x7ffff7fc2740 (LWP 9513) received Trace/breakpoint trap (stopped)
WL: Handling extended status 0x03057f
LHEW: Got clone event from LWP 9513, new child is LWP 9579
[New Thread 0x7ffff37b8700 (LWP 9579)]
WL: waitpid Thread 0x7ffff7fc2740 (LWP 9508) received 0 (exited)
WL: Thread 0x7ffff7fc2740 (LWP 9508) exited.
^^^^^^^^
[Thread 0x7ffff7fc2740 (LWP 9508) exited]
WL: waitpid Thread 0x7ffff7fc2740 (LWP 9499) received 0 (exited)
WL: Thread 0x7ffff7fc2740 (LWP 9499) exited.
[Thread 0x7ffff7fc2740 (LWP 9499) exited]
RSRL: resuming stopped-resumed LWP Thread 0x7ffff37b8700 (LWP 9579) at 0x3615ef4ce1: step=0
...
(gdb) info inferiors
Num Description Executable
5 process 9508 /home/pedro/bugs/src/test
^^^^
4 process 9503 /home/pedro/bugs/src/test
3 process 9500 /home/pedro/bugs/src/test
2 process 9499 /home/pedro/bugs/src/test
* 1 <null> /home/pedro/bugs/src/test
(gdb)
...
Note the "Thread 0x7ffff7fc2740 (LWP 9508) exited." line.
That's this in wait_lwp:
/* Check if the thread has exited. */
if (WIFEXITED (status) || WIFSIGNALED (status))
{
thread_dead = 1;
if (debug_linux_nat)
fprintf_unfiltered (gdb_stdlog, "WL: %s exited.\n",
target_pid_to_str (lp->ptid));
}
}
That was the leader thread reporting an exit, meaning the whole
process is gone. So the problem is that this code doesn't understand
that an WIFEXITED status of the leader LWP should be reported to
infrun as process exit.
gdb/ChangeLog:
2015-07-30 Pedro Alves <palves@redhat.com>
PR threads/18600
* linux-nat.c (wait_lwp): Report to the core when thread group
leader exits.
gdb/testsuite/ChangeLog:
2015-07-30 Pedro Alves <palves@redhat.com>
PR threads/18600
* gdb.threads/fork-plus-threads.exp: Test that "info inferiors"
only shows inferior 1.
When a program forks and another process start threads while gdb is
handling the fork event, newly created threads are left stuck stopped
by gdb, even though gdb presents them as "running", to the user.
This can be seen with the test added by this patch. The test has the
inferior fork a certain number of times and waits for all children to
exit. Each fork child spawns a number of threads that do nothing and
joins them immediately. Normally, the program should run unimpeded
(from the point of view of the user) and exit very quickly. Without
this fix, it doesn't because of some threads left stopped by gdb, so
inferior 1 never exits.
The program triggers when a new clone thread is found while inside the
linux_stop_and_wait_all_lwps call in linux-thread-db.c:
linux_stop_and_wait_all_lwps ();
ALL_LWPS (lp)
if (ptid_get_pid (lp->ptid) == pid)
thread_from_lwp (lp->ptid);
linux_unstop_all_lwps ();
Within linux_stop_and_wait_all_lwps, we reach
linux_handle_extended_wait with the "stopping" parameter set to 1, and
because of that we don't mark the new lwp as resumed. As consequence,
the subsequent resume_stopped_resumed_lwps, called from
linux_unstop_all_lwps, never resumes the new LWP.
There's lots of cruft in linux_handle_extended_wait that no longer
makes sense. On systems with CLONE events support, we don't rely on
libthread_db for thread listing anymore, so the code that preserves
stop_requested and the handling of last_resume_kind is all dead.
So the fix is to remove all that, and simply always mark the new LWP
as resumed, so that resume_stopped_resumed_lwps re-resumes it.
gdb/ChangeLog:
2015-07-30 Pedro Alves <palves@redhat.com>
Simon Marchi <simon.marchi@ericsson.com>
PR threads/18600
* linux-nat.c (linux_handle_extended_wait): On CLONE event, always
mark the new thread as resumed. Remove STOPPING parameter.
(wait_lwp): Adjust call to linux_handle_extended_wait.
(linux_nat_filter_event): Adjust call to
linux_handle_extended_wait.
(resume_stopped_resumed_lwps): Add debug output.
gdb/testsuite/ChangeLog:
2015-07-30 Simon Marchi <simon.marchi@ericsson.com>
Pedro Alves <palves@redhat.com>
PR threads/18600
* gdb.threads/fork-plus-threads.c: New file.
* gdb.threads/fork-plus-threads.exp: New file.
This patch removes the isize output argument from the
fast_tracepoint_valid_at gdbarch hook. It was used to return the size
of the instruction that needs to be replaced when installing a fast
tracepoint. Instead of getting this value from the
fast_tracepoint_valid_at hook, we can call the gdb_insn_length function.
If we do not do this, then architectures which do not have a restriction
on where to install the fast tracepoint will send uninitialized memory
off to GDBserver. See remote_download_tracepoint:
~~~
int isize;
if (gdbarch_fast_tracepoint_valid_at (target_gdbarch (),
tpaddr, &isize, NULL))
xsnprintf (buf + strlen (buf), BUF_SIZE - strlen (buf), ":F%x",
isize);
~~~
The default implementation of fast_tracepoint_valid_at will not set
isize resulting in uninitialized memory being sent. Later on, GDBserver
could use this information to compute a jump offset.
gdb/ChangeLog:
* arch-utils.c (default_fast_tracepoint_valid_at): Remove unused
isize argument.
* arch-utils.h (default_fast_tracepoint_valid_at): Likewise.
* breakpoint.c (check_fast_tracepoint_sals): Adjust call to
gdbarch_fast_tracepoint_valid_at.
* gdbarch.sh (fast_tracepoint_valid_at): Remove isize argument.
* gdbarch.h: Regenerate.
* gdbarch.c: Regenerate.
* i386-tdep.c (i386_fast_tracepoint_valid_at): Remove isize
argument. Do not set it.
* remote.c (remote_download_tracepoint): Adjust call to
gdbarch_fast_tracepoint_valid_at. Call gdb_insn_length to get
the instruction length.
This patch moves ARM register numbers enum to arch/arm.h, so that it
can used by GDBserver too.
This patch also creates a new directory gdb/arch in which arch-specific
or target-specific files are placed.
gdb:
2015-07-30 Yao Qi <yao.qi@linaro.org>
* arm-tdep.h (enum gdb_regnum): Move it to ...
* arch/arm.h: ... here. New file.
* Makefile.in (HFILES_NO_SRCDIR): Add arch/arm.h.
This patch cleans up the decoding functions using booleans when they can
decode two instructions. The boolean argument is used to know which of
the two instructions was decoded.
The instructions affected are BR/BLR, B/BL, CBZ/CBNZ and TBZ/TBNZ.
These arguments would be named after a named bit in the instruction
encoding, this patch renames them to 'is_XXX'. Furthermore, the
'unsigned' type would be used to describe a boolean while
aarch64_decode_cb would use 'int' (see the 'is64' argument). This patch
makes all booleans be 'int' and decoded bitfields be 'unsigned'.
gdb/ChangeLog:
* aarch64-tdep.c (decode_b): Rename link argument to is_bl.
Change its type to int *.
(decode_br): Rename link argument to is_blr. Change its type to
int *.
(decode_cb): Rename op argument to is_cbnz. Change its type to
int *.
(decode_tb): Rename op argument to is_tbnz. Change its type to
int *. Set is_tbnz to either 1 or 0.
(aarch64_analyze_prologue): Change type of is_link to int. Add
new variables is_cbnz and is_tbnz. Adjust call to
aarch64_decode_cb and aarch64_decode_tb.
Since Pedro's ptrace cleanups, the MIPS buildbot compilation fails.
Code in MIPS native uses ptrace with 3 arguments, where ptrace requires
4. When looking at the definition of ptrace in
/usr/include/sys/ptrace.h, it shows that it takes a variable number of
arguments. The wrapper macro in nat/gdb_ptrace.h takes a fixed number
of arguments (4). That would explain why it used to work and stopped.
I am pushing this as obvious, tell me if there is any problem.
I built-tested this with a MIPS toolchain (ct-ng), but I don't have any
setup to test it. At least it should put back the buildbot builder in a
better shape.
gdb/ChangeLog:
* mips-linux-nat.c (write_watchpoint_regs): Add NULL as ptrace's 4th
parameter.
(mips_linux_new_thread): Likewise.
* nat/mips-linux-watch.c (mips_linux_read_watch_registers): Likewise.
gdb/gdbserver/ChangeLog:
* linux-mips-low.c (mips_linux_prepare_to_resume): Add NULL as
ptrace's 4th parameter.
When exiting GDB -- whether it's via the "quit" command, via a SIGTERM,
or otherwise -- we should leave the terminal in the state we acquired
it. To that end, we have to undo any modifications that may have been
made by the TUI (ncurses) or by the CLI (readline).
Tested on x86_64 Debian Stretch.
gdb/ChangeLog:
* top.c: Include "tui/tui.h".
(undo_terminal_modifications_before_exit): New static function.
(quit_force): Use it.
gdb/testsuite/ChangeLog:
* gdb.base/batch-preserve-term-settings.exp
(test_terminal_settings_preserved_after_cli_exit): New test.
Right now this variable is initialized to 0 i.e. terminal_is_inferior
and does not get set to terminal_is_ours until target_terminal_init() is
called. This function however only gets called when an inferior is
first created. In the meantime, terminal_state would wrongly remain set
to terminal_is_inferior.
Tested on x86_64 Debian Stretch -- native, gdbserver and
extended-gdbserver.
gdb/ChangeLog:
* target.c (terminal_state): Initialize to terminal_is_ours.
We see the following regressions in testing on x86_64-linux,
reverse-step^M
Cannot access memory at address 0x2aaaaaed26c0^M
(gdb) FAIL: gdb.reverse/solib-precsave.exp: reverse-step into solib function one
when GDB reverse step into a function, GDB wants to skip prologue so
it requests TARGET_OBJECT_CODE_MEMORY to read some code memory in
memory_xfer_partial_1. However in dcache_read_memory_partial, the object
becomes TARGET_OBJECT_MEMORY
return ops->to_xfer_partial (ops, TARGET_OBJECT_MEMORY, NULL,
myaddr, NULL, memaddr, len,
xfered_len);
in reverse debugging, ops->to_xfer_partial is record_full_core_xfer_partial
and it will return TARGET_XFER_E_IO because it can't find any records.
The test fails.
At this moment, the delegate relationship is like
dcache -> record-core -> core -> exec
and we want to GDB read memory across targets, which means if the
requested memory isn't found in record-core, GDB can read memory from
core, and exec even further if needed. I find raw_memory_xfer_partial
is exactly what I want.
gdb:
2015-07-29 Yao Qi <yao.qi@linaro.org>
PR record/18691
* dcache.c (dcache_read_memory_partial): Call
raw_memory_xfer_partial.
* target.c (raw_memory_xfer_partial): Make it non-static.
* target.h (raw_memory_xfer_partial): Declare.
This patch updates various value handling functions to make them
consider the addressable memory unit size of the current architecture.
This allows to correctly extract and print values on architectures whose
addressable memory unit is not 8 bits.
The patch doesn't cover all the code that would ideally need to be
adjusted, only the code paths that we happen to use, plus a few obvious
ones. Specifically, those areas are not covered by this patch:
- Management of unavailable bits
- Bitfields
- C++ stuff
Regression-tested on x86-64 Ubuntu 14.04. I saw no related test result
change.
gdb/ChangeLog:
* c-valprint.c (c_val_print_array): Consider addressable memory
unit size.
(c_val_print_ptr): Likewise.
(c_val_print_int): Likewise.
* findvar.c (read_frame_register_value): Likewise.
* valarith.c (find_size_for_pointer_math): Likewise.
(value_ptrdiff): Likewise.
(value_subscripted_rvalue): Likewise.
* valops.c (read_value_memory): Likewise (and rename variables).
(value_assign): Likewise.
(value_repeat): Likewise.
(value_array): Likewise.
(value_slice): Likewise.
* valprint.c (generic_val_print_ptr): Likewise.
(generic_val_print_enum): Likewise.
(generic_val_print_bool): Likewise.
(generic_val_print_int): Likewise.
(generic_val_print_char): Likewise.
(generic_val_print_float): Likewise.
(generic_val_print_decfloat): Likewise.
(generic_val_print_complex): Likewise.
(val_print_scalar_formatted): Likewise.
(val_print_array_elements): Likewise.
* value.c (set_value_parent): Likewise.
(value_contents_copy_raw): Likewise.
(set_internalvar_component): Likewise.
(value_primitive_field): Likewise.
(value_fetch_lazy): Likewise.
* value.h (read_value_memory): Update comment.
Similar to get_type_arch, used to get the gdbarch associated to a
struct value.
gdb/ChangeLog:
* value.c (get_value_arch): New function.
* value.h (get_value_arch): New declaration.
This patch tries to clean up a bit the blur around the length field in
struct type, regarding its use with architectures with non-8-bits
addressable memory. It clarifies that the field is expressed in host
bytes, which is what is the closest to the current reality.
It also introduces a new function to get the length of the type in
target addressable memory units.
gdb/ChangeLog:
* gdbtypes.c (type_length_units): New function.
* gdbtypes.h (type_length_units): New declaration.
(struct type) <length>: Update comment.
GDB currently does not promptly quit after receiving a SIGTERM while no
proper target is active. This is because in handle_sigterm we currently
look at target_can_async_p to determine whether to asynchronously quit
GDB using an async signal handler or to asynchronously quit using the
quit flag. However, target_can_async_p is always false under the dummy
target, so under this target we always use the quit flag and not the
async signal handler to signal that GDB should quit. So GDB won't quit
until a code path that checks the quit flag is executed.
To fix this issue, this patch makes the SIGTERM handler no longer
inspect target_can_async_p, and instead makes the handler
unconditionally set the quit flag _and_ mark the corresponding async
signal handler, so that if the target is async (or if it's the dummy
target) then we will likely quit through the async signal handler, and
if it's not async then we will likely quit through the quit flag. This
redundant approach is similar to how we handle SIGINT.
gdb/ChangeLog:
* event-top.c (handle_sigterm): Don't inspect
target_can_async_p. Always set the quit flag and always mark
the async signal handler.
gdb/testsuite/ChangeLog:
* gdb.base/gdb-sigterm-2.exp: New test.
When using GDB to debug an RX target using the GDB remote protocol,
using a Renesas supplied debug agent, I encountered the following
assertion error:
thread.c:85: internal-error: inferior_thread: Assertion `tp' failed.
A problem internal to GDB has been detected,
further debugging may prove unreliable.
Create a core file of GDB? (y or n) n
Command aborted.
This assertion error occurs due to the fact that the value associated
with inferior_ptid is not on the thread list.
The remote debug output (obtained with "set debug remote 1") is fairly
short, so I will include it up to the point where things go wrong -
which is somewhat before the assertion failure:
(gdb) target remote coyote.lan:61234
Remote debugging using coyote.lan:61234
Sending packet: $qSupported:multiprocess+;swbreak+;hwbreak+;qRelocInsn+#c9...Ack
Packet received: PacketSize=c00;qXfer:memory-map:read-;qXfer:features:read-;QStartNoAckMode+;multiprocess+;QNonStop+
Packet qSupported (supported-packets) is supported
Sending packet: $QStartNoAckMode#b0...Ack
Packet received: OK
Sending packet: $Hgp0.0#ad...Packet received: OK
Sending packet: $QNonStop:0#8c...Packet received: OK
Sending packet: $qTStatus#49...Packet received:
Packet qTStatus (trace-status) is NOT supported
Sending packet: $?#3f...Packet received: S02
Sending packet: $qfThreadInfo#bb...Packet received: m1
Sending packet: $qsThreadInfo#c8...Packet received: l
Sending packet: $qAttached:a410#bf...Packet received: 0
Packet qAttached (query-attached) is supported
Sending packet: $Hc-1#09...Packet received: OK
Sending packet: $qC#b4...Packet received: QC not supported
Above is the trace starting from the invocation of "target remote"
through the call of get_current_thread() in remote_start_remote().
Below, I've pasted this line of code along with additional lines of
context. The test following the call is especially important to
understanding both the problem and my patch.
/* We have thread information; select the thread the target
says should be current. If we're reconnecting to a
multi-threaded program, this will ideally be the thread
that last reported an event before GDB disconnected. */
inferior_ptid = get_current_thread (wait_status);
if (ptid_equal (inferior_ptid, null_ptid))
{
/* Odd... The target was able to list threads, but not
tell us which thread was current (no "thread"
register in T stop reply?). Just pick the first
thread in the thread list then. */
inferior_ptid = thread_list->ptid;
}
}
Prior to getting to the code pasted above, remote_start_remote()
made a call to target_update_thread_list(). This corresponds to the
following lines from the above trace:
Sending packet: $qfThreadInfo#bb...Packet received: m1
Sending packet: $qsThreadInfo#c8...Packet received: l
Sending packet: $qAttached:a410#bf...Packet received: 0
Packet qAttached (query-attached) is supported
Once target_update_thread_list has completed, the thread list
contains a single entry: {pid = 42000, lwp = 1, tid = 0}.
remote_start_remote() then makes a call to set_continue_thread(),
accounting for this line of the trace:
Sending packet: $Hc-1#09...Packet received: OK
Finally, the call to get_current_thread() is responsible for the last
line of the trace that I provided above:
Sending packet: $qC#b4...Packet received: QC not supported
get_current_thread() calls stop_reply_extract_thread() with the wait
status. This returns null_ptid.
get_current_thread() then calls remote_current_thread with a null
inferior_ptid. After the calls to putpkt() and getpkt(), rs->buf[0]
is 'Q', so read_ptid() is called and its result is returned.
The buffer passed to read_ptid() is " not supported". read_ptid ultimately
returns a ptid of {pid = 4200, lwp = 0, tid = 0}.
However, this thread is not on the thread list. As noted earlier, the
call to target_update_thread_list() had placed {pid = 42000, lwp = 1,
tid = 0} on the list. This is the only thread in the list.
When these calls ultimately return to remote_start_remote(),
inferior_ptid gets set to {pid = 4200, lwp = 0, tid = 0}, which
(again) is not on the thread list.
It appears to me that the string " not supported" is coming from the
debug agent. If so, it should be fixed, but I don't see a reason to
not consult the thread list in order to place a valid thread id in
inferior_ptid.
This (consultation of the thread list) is what is done when
inferior_ptid is null_ptid:
if (ptid_equal (inferior_ptid, null_ptid))
{
/* Odd... The target was able to list threads, but not
tell us which thread was current (no "thread"
register in T stop reply?). Just pick the first
thread in the thread list then. */
inferior_ptid = thread_list->ptid;
}
My patch causes a null inferior_ptid to be returned by read_ptid when
no thread id is found in the response from the debug agent. This
return value ends up being returned by remote_current_thread() and
then by get_current_thread. The assignment then places this null
value into inferior_ptid. That, in turn, allows the ptid_equal test
(noted above) to fetch a valid thread from the thread list. I no
longer see the assertion failure due a good value (which is on the
thread list) being placed in inferior_ptid.
This patch also adds two log warnings that may be output when "set
debug remote 1" is used. When running against the Renesas debug agent
mentioned earlier, this is the relevant portion of the log output:
Sending packet: $qC#b4...Packet received: QC not supported
warning: garbage in qC reply
warning: couldn't determine remote current thread; picking first in list.
gdb/ChangeLog:
* remote.c (read_ptid): Return null_ptid when no thread id
is found.
(remote_current_thread): Add log warning for malformed
qC reply.
(remote_start_remote): Add log warning when current thread
not found.
This reverts commit b558ff043d.
This reverts commit 4a11f20659.
The initial import commit failed to retain local changes made to
readline's configure.in (and the commit message erroneously stated that
there were no local changes that needed to be reapplied). Also the
import caused a couple of build errors and a scattering of testsuite
regressions throughout many arches. It's probably better to start over
with this import, hopefully more carefully next time.
This patch syncs our upstream copy of readline from version 6.2 to the
latest version, 7.0 alpha (released July 10 2015).
I essentially copied what was done the last time readline was synced,
when Jan updated to readline 6.2 in 2011:
http://sourceware.org/ml/gdb-patches/2011-05/msg00003.html
Procedure:
1. I extracted the readline-7.0-alpha tarball on top of readline/.
2. I deleted all the new files under doc/ that were deliberately omitted
before.
3. I regenerated readline/configure and readline/examples/rlfe/configure
using autoconf 2.64. No other configure files need regenerating.
4. I updated the function gdb_printable_part in completer.c with a
trivial change made to the readline function it is based off of,
printable_part in readline/complete.c. There is more work to be done in
completer.c to sync it with readline/complete.c, but it is non-trivial
and should probably be done separately anyway.
Local patches that had to be reapplied:
None. readline 7.0 alpha contains all of our local readline
patches.
New files in readline/:
colors.{c,h}
examples/{hist_erasedups,hist_purgecmd,rl-callbacktest,rlbasic}.c
parse-colors.{c,h}
readline.pc.in
configure.ac
Deleted files in readline/:
configure.in
Regressions:
After the sync there is one testsuite regression, the test
"signal SIGINT" in gdb.gdb/selftest.exp which now FAILs. Previously,
the readline 6.2 SIGINT handler would temporarily reinstall the
underlying application's SIGINT handler and immediately re-raise SIGINT
so that the orginal handler gets invoked. But now (since readline 6.3)
its SIGINT handler does not re-raise SIGINT or directly invoke the
original handler; it now sets a flag marking that SIGINT was raised, and
waits until readline explicitly has control to call the application's
SIGINT handler. Anyway, because SIGINT is no longer re-raised from
within readline's SIGINT handler, doing "signal SIGINT" with a stopped
inferior gdb process will no longer resume and then immediately stop the
process (since there is no 2nd SIGINT to immediately catch). Instead,
the inferior gdb process will now just print "Quit" and continue to run.
So with this commit, this particular test case is adjusted to reflect
this change in behavior (we now have to send a 2nd SIGINT manually to
stop it).
Aside from this one testsuite regression, I personally noticed no
regression in user-visible behavior. Though I only tested on x86_64
and on i686 Debian Stretch.
Getting this kind of change in at the start of the GDB 7.11 development
cycle will allow us to get a lot of passive testing from developers and
from bleeding-edge users.
readline/ChangeLog.gdb:
Import readline 7.0 alpha
* configure: Regenerate.
* examples/rlfe/configure: Regenerate.
gdb/ChangeLog:
* completer.c (gdb_printable_part): Sync with readline function
it is based off of.
gdb/testsuite/ChangeLog:
* gdb.gdb/selftest.exp (test_with_self): Update test to now
expect the GDB inferior to no longer immediately stop after
being resumed with "signal SIGINT".
The buildbot noticed that the enum __ptrace_request series broke the
s390 GNU/Linux build:
../../binutils-gdb/gdb/s390-linux-nat.c: In function 'fetch_regs':
../../binutils-gdb/gdb/s390-linux-nat.c:226:54: error: macro "ptrace" requires 4 arguments, but only 3 given
if (ptrace (PTRACE_PEEKUSR_AREA, tid, (long) &parea) < 0)
^
../../binutils-gdb/gdb/s390-linux-nat.c: In function 'store_regs':
../../binutils-gdb/gdb/s390-linux-nat.c:243:54: error: macro "ptrace" requires 4 arguments, but only 3 given
if (ptrace (PTRACE_PEEKUSR_AREA, tid, (long) &parea) < 0)
^
Fix this the same way it's handled everywhere else -- just pass 0 as
forth argument, which also handles non-varargs ptrace prototypes in
non-glibc libcs, e.g., Bionic (if it ever gets a s390 port...).
gdb/ChangeLog:
2015-07-24 Pedro Alves <palves@redhat.com>
* s390-linux-nat.c (fetch_regs, store_regs, fetch_fpregs)
(s390_stopped_by_watchpoint, s390_prepare_to_resume): Pass 0 as
forth argument to ptrace PTRACE_PEEKUSR_AREA/PTRACE_POKEUSR_AREA.
If a non-leader thread exits the process while all other threads are
ptrace-stopped, native gdb fails an assertion. The test added by this
commit catches it:
/home/pedro/gdb/mygit/build/../src/gdb/linux-nat.c:3198: internal-error: linux_nat_filter_event: Assertion `lp->resumed' failed.
A problem internal to GDB has been detected,
further debugging may prove unreliable.
Quit this debugging session? (y or n)
FAIL: gdb.threads/non-leader-exit-process.exp: program exits normally (GDB internal error)
The fix is just to remove the assertion.
With that out of the way, neither GDB not GDBserver handle this
perfectly though, so I'm adding a KFAIL:
(gdb) continue
Continuing.
[Thread 0x7ffff7fc0700 (LWP 15350) exited]
No unwaited-for children left.
Couldn't get registers: No such process.
(gdb) KFAIL: gdb.threads/non-ldr-exit.exp: program exits normally (PRMS: gdb/18717)
gdb/ChangeLog:
2015-07-24 Pedro Alves <palves@redhat.com>
PR gdb/18717
* linux-nat.c (linux_nat_filter_event): Don't assert that the lwp
is resumed, and extend the debug log.
gdb/testsuite/ChangeLog:
2015-07-24 Pedro Alves <palves@redhat.com>
PR gdb/18717
* gdb.threads/non-ldr-exit.c: New file.
* gdb.threads/non-ldr-exit.exp: New file.
Ref: https://sourceware.org/ml/gdb-patches/2015-07/msg00629.html
This fixes the bogus command line in the error message shown when the
SHELL environment variable points somewhere that's not something that
resembles a shell:
$ SHELL=/nonexisting gdb /home/pedro/a.out
(gdb) r
Starting program: /home/pedro/a.out
- Cannot exec /home/pedro/a.out -c exec /home/pedro/a.out .
+ Cannot exec /nonexisting -c exec /home/pedro/a.out .
Error: No such file or directory
During startup program exited with code 127.
(gdb)
gdb/ChangeLog:
2015-07-24 Pedro Alves <palves@redhat.com>
* fork-child.c (fork_inferior): Print argv[0] instead of exec_file.
Building in C++ mode issues ~40 warnings like this:
../../src/gdb/linux-nat.c: In function ‘int linux_handle_extended_wait(lwp_info*, int, int)’:
../../src/gdb/linux-nat.c:2016:51: warning: invalid conversion from ‘int’ to ‘__ptrace_request’ [-fpermissive]
ptrace (PTRACE_GETEVENTMSG, pid, 0, &new_pid);
The issue is that in glibc, ptrace's first parameter is an enum.
That's not a problem if we pick the PTRACE_XXX requests from
sys/ptrace.h, as those will be values of the corresponding enum.
However, we have fallback definitions for PTRACE_XXX symbols when the
system headers miss them (such as PTRACE_GETEVENTMSG above), and those
are plain integer constants. E.g., nat/linux-ptrace.h:
#define PTRACE_GETEVENTMSG 0x4201
One idea would be to fix this by defining those fallbacks like:
-#define PTRACE_GETEVENTMSG 0x4201
+#define PTRACE_GETEVENTMSG ((enum __ptrace_request) 0x4201)
However, while glibc's ptrace uses enum __ptrace_request for first
parameter:
extern long int ptrace (enum __ptrace_request __request, ...) __THROW;
other libc's, like e.g., Android's bionic do not -- in that case, the
first parameter is int:
long ptrace(int request, pid_t pid, void * addr, void * data);
So the fix I came up is to make configure/ptrace.m4 also detect the
type of the ptrace's first parameter and defin PTRACE_TYPE_ARG1, as
already does the for parameters 3-4, and then simply wrap ptrace with
a macro that casts the first argument to the detected type. (I'm
leaving adding a nicer wrapper for when we drop building in C).
While this adds the wrapper, GNU/Linux files won't use it until the
next patch, which makes all native GNU/Linux files include
gdb_ptrace.h.
gdb/ChangeLog:
2015-07-24 Pedro Alves <palves@redhat.com>
* ptrace.m4 (ptrace tests): Test in C++ mode. Try with 'enum
__ptrace_request as first parameter type instead of int.
(PTRACE_TYPE_ARG1): Define.
* nat/gdb_ptrace.h [!PTRACE_TYPE_ARG5] (ptrace): Define as wrapper
that casts first argument to PTRACE_TYPE_ARG1.
* config.in: Regenerate.
* configure: Regenerate.
gdb/gdbserver/ChangeLog:
2015-07-24 Pedro Alves <palves@redhat.com>
* config.in: Regenerate.
* configure: Regenerate.
Now that gdbserver's configure defines PTRACE_TYPE_ARGx etc., we'll be
able to make gdbserver use gdb_ptrace.h too. Move it to the native
target files directory.
gdb/ChangeLog:
2015-07-24 Pedro Alves <palves@redhat.com>
* gdb_ptrace.h: Move ...
* nat/gdb_ptrace.h: ... here.
* inf-ptrace.c: Adjust.
This factors the ptrace checks out of gdb's configure.ac to a new
ptrace.m4 file, and then makes gdbserver's configure.ac source it too.
gdb/ChangeLog:
2015-07-24 Pedro Alves <palves@redhat.com>
* acinclude.m4: Include ptrace.m4.
* configure.ac: Call GDB_AC_PTRACE and move ptrace checks ...
* ptrace.m4: ... to this new file.
gdb/gdbserver/ChangeLog:
2015-07-24 Pedro Alves <palves@redhat.com>
* acinclude.m4: Include ../ptrace.m4.
* configure.ac: Call GDB_AC_PTRACE.
* config.in, configure: Regenerate.
Dummy CUs are used by the incremental linker to pre-allocate space
in the output file. They have a DWARF header but no contents.
gdb/ChangeLog:
* dwarf2read.c (dwarf2_per_cu_data): Add comment.
(load_cu): Handle dummy CUs.
(dw2_do_instantiate_symtab, process_queuef): Ditto.
(dwarf2_fetch_die_loc_sect_off, dwarf2_fetch_constant_bytes): Ditto.
gdb/testsuite/ChangeLog:
* gdb.dwarf2/dw2-dummy-cu.S: New file.
* gdb.dwarf2/dw2-dummy-cu.exp: New file.
The ltpy_get_all_source_lines function, use to implement
the gdb.LineTable.source_lines method, returns a list:
source_list = PyDict_Keys (source_dict);
return source_list;
This patch fixes the function's documentation as well as its docstring
to say that it returns a list rather than a FrozenSet.
gdb/ChangeLog:
* py-linetable.c (ltpy_get_all_source_lines): Adjust function
documentation to say that it returns a list rather than
a FrozenSet.
(linetable_object_methods): Update the docstring of the
"source_line" entry.
Tested on x86_64-linux.
When a dynamic array type contains a typedef-wrapped array, an assertion
failure occurs during type resolution. This is what happens in the
following Ada case:
type Rec_Type is record
I : Integer;
B : Boolean;
end record;
type Vec_Type is array (1 .. 4) of Rec_Type;
type Array_Type is array (Positive range <>) of Vec_Type;
If users try to print or even pass to an inferior call a variable A of
type Array_Type, GDB will raise an error:
(gdb) print a
../../src/gdb/gdbtypes.c:1807: internal-error:
resolve_dynamic_array: Assertion `TYPE_CODE (type) ==
TYPE_CODE_ARRAY' failed.
A problem internal to GDB has been detected,
further debugging may prove unreliable.
Quit this debugging session? (y or n)
What happens is that during dynamic array type resolution, we first peel
TYPE_CODE_TYPEDEF layers wrapping the array element type and check if
its type is itself TYPE_CODE_ARRAY. If it is, we pass the
typedef-wrapped type to a recursive call to resolve_dynamic_array
whereas this function expects only TYPE_CODE_ARRAY types.
This patch makes it pass the peeled type to the recursive call so that
type resolution can continue smoothly.
gdb/ChangeLog:
* gdbtypes.c (resolve_dynamic_array): Pass the peeled element
type to the recursive call instead of the original (maybe
TYPE_CODE_TYPEDEF) type.
gdb/testsuite/ChangeLog:
* gdb.ada/var_arr_typedef.exp: New testcase.
* gdb.ada/var_arr_typedef/pack.adb: New file.
* gdb.ada/var_arr_typedef/pack.ads: New file.
* gdb.ada/var_arr_typedef/var_arr_typedef.adb: New file.
Nowadays aarch64_linux_can_use_hw_breakpoint always return one, but it
can be smarter, say, if GDB knows target doesn't support HW watchpoint
or breakpoint because HW watchpoint/breakpoint is disabled in linux
kernel, for example, it can safely return zero.
gdb:
2015-07-23 Yao Qi <yao.qi@linaro.org>
* aarch64-linux-nat.c (aarch64_linux_can_use_hw_breakpoint): If
TYPE is watchpoint, return zero if aarch64_num_wp_regs is zero.
If TYPE is breakpoint, return zero if arch64_num_bp_regs is zero.
There are also some duplication on getting HW watchpoint/breakpoint
registers info between GDB and GDBserver. This patch moves them
to nat/aarch64-linux-hw-point.c.
Note that ENABLE_NLS is not defined in GDBserver, so it should be OK
to use _( markup.
gdb:
2015-07-21 Yao Qi <yao.qi@linaro.org>
* aarch64-linux-nat.c (aarch64_linux_get_debug_reg_capacity):
Move it to nat/aarch64-linux-hw-point.c.
(aarch64_linux_child_post_startup_inferior): Update.
* nat/aarch64-linux-hw-point.c (aarch64_linux_get_debug_reg_capacity):
New function.
* nat/aarch64-linux-hw-point.h (aarch64_linux_get_debug_reg_capacity):
Declare it.
gdb/gdbserver:
2015-07-21 Yao Qi <yao.qi@linaro.org>
* linux-aarch64-low.c (aarch64_arch_setup): Remove code and call
aarch64_linux_get_debug_reg_capacity.