The argument to gdb_init is not used, remove it.
There should be no user visible changes after this commit.
gdb/ChangeLog:
* main.c (captured_main_1): Don't pass argument to gdb_init.
* top.c (gdb_init): Remove unused argument, and add header
comment.
* top.h (gdb_init): Remove argument.
PR gas/27763
While the comment in output_jump() was basically correct prior to the
introduction of 64-bit mode, both that and the not-JMP-like behavior of
XBEGIN require adjustments: Branches with 32-bit displacement do not
wrap at 4G in 64-bit mode, and XBEGIN with 16-bit operand size doesn't
wrap at 64k. Similarly %rip-relative addressing doesn't wrap at 4G.
The new testcase points out that for PE/COFF object_64bit didn't get
set so far, preventing in particular the check at the end of
md_convert_frag() to take effect.
For Mach-O the new testcase fails (bogusly), in that only the first two
of the expected errors get raised. Since for Mach-O many testcases
already fail, and since an x86_64-darwin target can't even be configured
for, I didn't think I need to bother.
Note that there are further issues in this area, in particular for
branches with operand size overrides. Such branches, which truncate
%rip / %eip, can't be correctly expressed with ordinary PC-relative
relocations. It's not really clear what to do with them - perhaps the
best we can do is to carry through all associated relocations, leaving
it to the linker (or even loader) to decide (once the final address
layout is known). Same perhaps goes for relocations associated with
32-bit addressing in 64-bit mode.
I'm seeing timeouts from gdb.rust/traits.exp when we attempt to print
things with "maint print objfiles".
This happens for two reasons:
1 - GDB does not explicitly split each entry into its own line, but rather
relies on the terminal's width to insert line breaks.
2 - When running the GDB testsuite, such width may be unlimited, which will
prevent GDB from inserting any line breaks.
As a result, the output may be too lengthy and will come in big lines. Tweak
the support library to match the patterns line-by-line, which gives us more
time to match things. Also fix GDB's output to print one entry per line,
regardless of the terminal width.
A similar approach was used in another testcase using the same command (commit
eaeaf44cfdc9a4096a0dd52fa0606f29d4bfd48e). With the new line breaks, we don't
need a particular pattern, so clean up that test as well.
gdb/ChangeLog:
2021-04-27 Luis Machado <luis.machado@linaro.org>
* psymtab.c (psymbol_functions::dump): Output newline.
* symmisc.c (dump_objfile): Likewise.
gdb/testsuite/ChangeLog:
2021-04-27 Luis Machado <luis.machado@linaro.org>
* gdb.base/maint.exp: Drop a pattern that is not needed.
* lib/gdb.exp (readnow): Match line-by-line.
For aarch64, the "info all-registers" output is very verbose. If you run
this test using read1, it will timeout before the command output is done being
read.
Read line-by-line so we don't run into timeouts.
gdb/testsuite/ChangeLog:
2021-04-27 Luis Machado <luis.machado@linaro.org>
* gdb.xml/tdesc-reload.exp: Pass -lbl.
Without any explicit dependencies specified, the observers attached
to the 'gdb::observers::new_objfile' observable are always notified
in the order in which they have been attached.
The new_objfile observer callback to auto-load scripts is attached in
'_initialize_auto_load'.
The new_objfile observer callback that propagates the new_objfile event
to the Python side is attached in 'gdbpy_initialize_inferior', which is
called via '_initialize_python'.
With '_initialize_python' happening before '_initialize_auto_load',
the consequence was that the new_objfile event was emitted on the Python
side before autoloaded scripts had been executed when a new objfile was
loaded.
As a result, trying to access the objfile's pretty printers (defined in
the autoloaded script) from a handler for the Python-side
'new_objfile' event would fail. Those would only be initialized later on
(when the 'auto_load_new_objfile' callback was called).
To make sure that the objfile passed to the Python event handler
is properly initialized (including its 'pretty_printers' member),
make sure that the 'auto_load_new_objfile' observer is notified
before the 'python_new_objfile' one that propagates the event
to the Python side.
To do this, make use of the mechanism to explicitly specify
dependencies between observers (introduced in a preparatory commit).
Add a corresponding testcase that involves a test library with an autoloaded
Python script and a handler for the Python 'new_objfile' event.
(The real world use case where I came across this issue was in an attempt
to extend handling for GDB pretty printers for dynamically loaded
objfiles in the Qt Creator IDE, s. [1] and [2] for more background.)
[1] https://bugreports.qt.io/browse/QTCREATORBUG-25339
[2] https://codereview.qt-project.org/c/qt-creator/qt-creator/+/333857/1
Tested on x86_64-linux (Debian testing).
gdb/ChangeLog:
* gdb/auto-load.c (_initialize_auto_load): 'Specify token
when attaching the 'auto_load_new_objfile' observer, so
other observers can specify it as a dependency.
* gdb/auto-load.h (struct token): Declare
'auto_load_new_objfile_observer_token' as token to be used
for the 'auto_load_new_objfile' observer.
* gdb/python/py-inferior.c (gdbpy_initialize_inferior): Make
'python_new_objfile' observer depend on 'auto_load_new_objfile'
observer, so it gets notified after the latter.
gdb/testsuite/ChangeLog:
* gdb.python/libpy-autoloaded-pretty-printers-in-newobjfile-event.so-gdb.py: New test.
* gdb.python/py-autoloaded-pretty-printers-in-newobjfile-event-lib.cc: New test.
* gdb.python/py-autoloaded-pretty-printers-in-newobjfile-event-lib.h: New test.
* gdb.python/py-autoloaded-pretty-printers-in-newobjfile-event-main.cc: New test.
* gdb.python/py-autoloaded-pretty-printers-in-newobjfile-event.exp: New test.
* gdb.python/py-autoloaded-pretty-printers-in-newobjfile-event.py: New test.
Change-Id: I8275b3f4c3bec32e56dd7892f9a59d89544edf89
Previously, the observers attached to an observable were always notified
in the order in which they had been attached. That order is not easily
controlled, because observers are typically attached in _initialize_*
functions, which are called in an undefined order.
However, an observer may require that another observer attached only
later is called before itself is.
Therefore, extend the 'observable' class to allow explicitly specifying
dependencies when attaching observers, by adding the possibility to
specify tokens for observers that it depends on.
To make sure dependencies are notified before observers depending on
them, the vector holding the observers is sorted in a way that
dependencies come before observers depending on them. The current
implementation for sorting uses the depth-first search algorithm for
topological sorting as described at [1].
Extend the observable unit tests to cover this case as well. Check that
this works for a few different orders in which the observers are
attached.
This newly introduced mechanism to explicitly specify dependencies will
be used in a follow-up commit.
[1] https://en.wikipedia.org/wiki/Topological_sorting#Depth-first_search
Tested on x86_64-linux (Debian testing).
gdb/ChangeLog:
* unittests/observable-selftests.c (dependency_test_counters):
New.
(observer_token0, observer_token1, observer_token2,
observer_token3, observer_token4, observer_token5): New.
(struct dependency_observer_data): New struct.
(observer_dependency_test_callback): New function.
(test_observers): New.
(run_dependency_test): New function.
(test_dependency): New.
(_initialize_observer_selftest): Register dependency test.
gdbsupport/ChangeLog:
* observable.h (class observable): Extend to allow specifying
dependencies between observers, keep vector holding observers
sorted so that dependencies are notified before observers
depending on them.
Change-Id: I5399def1eeb69ca99e28c9f1fdf321d78b530bdb
Add () to !i.prefix[ADDR_PREFIX] to silence GCC 5:
gas/config/tc-i386.c:4152:31: error: logical not is only applied to the left hand side of comparison [-Werror=logical-not-parentheses]
&& !i.prefix[ADDR_PREFIX] != (flag_code == CODE_32BIT))
^
* config/tc-i386.c (optimize_encoding): Add () to silence GCC 5.
Since libgloss provides a default syscall table for arches, use that
to provide the default syscall table for ports. Only the exceptions
need to be enumerated now with the common logic as the default.
Force this on for all ports. We have a few common models that can
be used, so make them generally available. If the port doesn't use
any hardware (the default), then behavior is unchanged.
PR gdb/27743 points out a gdb crash when expanding partial symtabs,
where one of the compilation units uses DW_TAG_imported_unit.
This crash happens for gdb 10, but not git trunk. This patch pulls
over the new test case only.
gdb/testsuite/ChangeLog
2021-04-26 Tom Tromey <tromey@adacore.com>
PR gdb/27743:
* gdb.dwarf2/imported-unit-bp.exp: New file.
* gdb.dwarf2/imported-unit-bp-main.c: New file.
* gdb.dwarf2/imported-unit-bp-alt.c: New file.
Bug 27773 shows that passing a filename in a non-existent directory to
the "dump binary" command leads to a gdb crash. This is because the
gdb_fopen_cloexec in dump_binary_file fails (returns nullptr) and the
return value is not checked. Fix that by erroring out if
gdb_fopen_cloexec fails.
gdb/ChangeLog:
PR gdb/27773
* cli/cli-dump.c (dump_binary_file): Check result of
gdb_fopen_cloexec.
gdb/testsuite/ChangeLog:
PR gdb/27773
* gdb.base/dump.exp: Test dump to non-existent dir.
Change-Id: Iea89a3bf9e6b9dcc31142faa5ae17bc855759328
The test checks for a particular ARCH level, but it needs to check for
ARCH levels from a minimum and upwards.
gdb/testsuite/ChangeLog:
2021-04-26 Luis Machado <luis.machado@linaro.org>
* gdb.arch/aarch64-dbreg-contents.c (set_watchpoint): Fix arch level
comparison.
The only insn requiring a truly 16-bit PC-relative relocation outside of
16-bit mode is XBEGIN (with an operand size override). For it, the
relocation generated should behave similar to 8- and (for 64-bit) 32-bit
PC-relatives ones, i.e. be checked for a signed value to fit the field.
This same mode is also correct for 16-bit code. Outside of 16-bit code,
branches with operand size overrides act in a truly PC-relative way only
when living in the low 32k of address space, as they truncate rIP to 16
bits. This can't be expressed by a PC-relative relocation.
Putting in place a new testcase, I'd like to note that the two existing
ones (pcrel16 and pcrel16abs) appear to be pretty pointless: They don't
expect any error despite supposedly checking for overflow, and in fact
there can't possibly be any error for the
- former since gas doesn't emit any relocation in the first place there,
- latter because the way the relocation gets expressed by gas doesn't
allow the linker to notice the overflow; it should be detected by gas
if at all, but see above (an error would be reported here for x86-64
afaict, but this test doesn't get re-used there).
Over the years I've seen a number of instances where people used
lea (%reg1), %reg2
or
lea symbol, %reg
despite the same thing being expressable via MOV. Since additionally
LEA often has restrictions towards the ports it can be issued to, while
MOV typically gets dealt with simply by register renaming, transform to
MOV when possible (without growing opcode size and without altering
involved relocation types).
Note that for Mach-O the new 64-bit testcases would fail (for
BFD_RELOC_X86_64_32S not having a representation), and hence get skipped
there.
Constants not known at the time an individual insn gets assembled and
going into a sign-extended field still shouldn't be silently truncated
at the time the respective fixup gets resolved.
On sparc build failed as:
```
gdb/sparc-linux-nat.c: In member function
'virtual void sparc_linux_nat_target::fetch_registers(regcache*, int)':
gdb/sparc-linux-nat.c:36:37:
error: cannot convert 'regcache*' to 'process_stratum_target*'
36 | { sparc_fetch_inferior_registers (regcache, regnum); }
| ^~~~~~~~
| |
| regcache*
```
The fix adopts gdb/sparc-nat.h API change in d1e93af64a6
("gdb: set current thread in sparc_{fetch,collect}_inferior_registers").
gdb/ChangeLog:
* sparc-linux-nat.c (sparc_linux_nat_target): fix sparc build
by passing `process_stratum_target*` parameter.
This commit adds a flag to the ptype command in order to print the
offsets and sizes of struct members using the hexadecimal notation. The
'x' flag ensures use of the hexadecimal notation while the 'd' flag
ensures use of the decimal notation. The default is to use decimal
notation.
Before this patch, gdb only uses decimal notation, as pointed out in PR
gdb/22640.
Here is an example of this new behavior with hex output turned on:
(gdb) ptype /ox struct type_print_options
/* offset | size */ type = struct type_print_options {
/* 0x0000: 0x0 | 0x0004 */ unsigned int raw : 1;
/* 0x0000: 0x1 | 0x0004 */ unsigned int print_methods : 1;
/* 0x0000: 0x2 | 0x0004 */ unsigned int print_typedefs : 1;
/* 0x0000: 0x3 | 0x0004 */ unsigned int print_offsets : 1;
/* 0x0000: 0x4 | 0x0004 */ unsigned int print_in_hex : 1;
/* XXX 3-bit hole */
/* XXX 3-byte hole */
/* 0x0004 | 0x0004 */ int print_nested_type_limit;
/* 0x0008 | 0x0008 */ typedef_hash_table *local_typedefs;
/* 0x0010 | 0x0008 */ typedef_hash_table *global_typedefs;
/* 0x0018 | 0x0008 */ ext_lang_type_printers *global_printers;
/* total size (bytes): 32 */
}
This patch also adds the 'set print type hex' and 'show print type hex'
commands in order to set and inspect the default behavior regarding the
use of decimal or hexadecimal notation when printing struct sizes and
offsets.
Tested using on x86_64.
gdb/ChangeLog:
PR gdb/22640
* typeprint.h (struct type_print_options): Add print_in_hex
flag.
(struct print_offset_data): Add print_in_hex flag, add a
constructor accepting a type_print_options* argument.
* typeprint.c (type_print_raw_options, default_ptype_flags): Set
default value for print_in_hex.
(print_offset_data::indentation): Allow more horizontal space.
(print_offset_data::print_offset_data): Add ctor.
(print_offset_data::maybe_print_hole, print_offset_data::update):
Handle the print_in_hex flag.
(whatis_exp): Handle 'x' and 'd' flags.
(print_offsets_and_sizes_in_hex): Declare.
(set_print_offsets_and_sizes_in_hex): Create.
(show_print_offsets_and_sizes_in_hex): Create.
(_initialize_typeprint): Update help message for the ptype
command, register the 'set print type hex' and 'show print type
hex' commands.
* c-typeprint.c (c_print_type, c_type_print_base_struct_union)
(c_type_print_base): Construct the print_offset_data
object using the type_print_optons parameter.
* rust-lang.c (rust_language::print_type): Construct the
print_offset_data object using the type_print_optons parameter.
* NEWS: Mention the new flags of the ptype command.
gdb/doc/ChangeLog:
PR gdb/22640
* gdb.texinfo (Symbols): Describe the 'x' and 'd' flags of the
ptype command, describe 'set print type hex' and 'show print
type hex' commands. Update 'ptype/o' examples.
gdb/testsuite/ChangeLog:
PR gdb/22640
* gdb.base/ptype-offsets.exp: Add tests to verify the behavior
of 'ptype/ox' and 'ptype/od'. Check that 'set print type hex'
changes the default behavior of 'ptype/o'. Update to take into
account new horizontal layout.
* gdb.rust/simple.exp: Update ptype test to check new horizontal
layout.
* gdb.rust/union.exp: Same.
Move the declaration of struct type_print_raw_options before struct
print_offset_data to ease upcoming changes. This is a helper commit
intended to make it easier to build a print_offset_data object from
configurations given by a type_print_raw_options.
gdb/ChangeLog:
* typeprint.h (struct type_print_options): Move before
print_offset_data.
A rebuild showed that an earlier change of mine missed a built header
file -- cris/engv32.h. This patch fixes the problem.
sim/cris/ChangeLog
2021-04-25 Tom Tromey <tom@tromey.com>
* Makefile.in (SIM_EXTRA_DEPS): Add engv32.h.
This changes sim-options.c to use the libiberty hash table, rather
than its own custom hash table.
sim/common/ChangeLog
2021-04-25 Tom Tromey <tom@tromey.com>
* sim-options.c (compare_strings): New function.
(ARG_HASH_SIZE, ARG_HASH): Remove.
(dup_arg_p): Use htab_t.
(sim_parse_args): Remove assert.
Switch observer to use the "new" debug printf mechanism and sprinkle a
few debug prints. Here's a small example of the output with "infrun"
and "observer" debug output enabled:
[infrun] proceed: enter
[observer] notify: start: observable target_resumed notify() called
[observer] notify: start: calling observer mi-interp of observable target_resumed
[observer] notify: end: calling observer mi-interp of observable target_resumed
[observer] notify: start: calling observer py-inferior of observable target_resumed
[observer] notify: end: calling observer py-inferior of observable target_resumed
[observer] notify: end: observable target_resumed notify() called
...
gdbsupport/ChangeLog:
* observable.h (observer_debug_printf,
OBSERVER_SCOPED_DEBUG_START_END): New.
(class observable) <notify, attach>: Use them.
Change-Id: If3ae4b6b65450ca3b7cae56698a87fc526688b86
A little thing that bothers me with scoped_debug_start_end is that it's
not possible to pass a format string to add context to the messages: the
start and end messages are fixed.
It was done like this at the time because there's the risk that debug
output is not enabled on entry (when the constructor runs) but is
enabled on exit (when the destructor runs). For example, a user
debugging from a top-gdb may manually enable a debug_foo variable. If
debug output is disabled while the constructor runs, we won't render the
format string (to minimize overhead) so it won't be available in the
destructor.
I think it would be nice to be able to use a format string along with
scoped_debug_start_end, and I think it's unfortunate that such a narrow
use case prevents it. So with this patch, I propose that we allow
passing a format string to scoped_debug_start_end, and if the rare
situation described above happens, then we just show a "sorry, message
not available" kind of message.
The following patch makes use of this.
gdbsupport/ChangeLog:
* common-debug.h (struct scoped_debug_start_end)
<scoped_debug_start_end>: Change start_msg/end_msg for
start_prefix/end_prefix. Add format string parameter and make
variadic.
<~scoped_debug_start_end>: Adjust.
<m_end_msg>: Rename to...
<m_end_prefix>: ... this.
<m_with_format>: New.
<m_msg>: New.
(scoped_debug_start_end): Make variadic.
(scoped_debug_enter_exit): Adjust.
Change-Id: I9427ce8877a246a46694b3a1fec3837dc6954d6e
Give a name to each observer, this will help produce more meaningful
debug message.
gdbsupport/ChangeLog:
* observable.h (class observable) <struct observer> <observer>:
Add name parameter.
<name>: New field.
<attach>: Add name parameter, update all callers.
Change-Id: Ie0cc4664925215b8d2b09e026011b7803549fba0
Instead of using a pair. This allows keeping more data per observer in
a structured way, and using field names is clearer than first/second.
gdbsupport/ChangeLog:
* observable.h (class observable) <struct observer>: New.
<detach, notify>: Update.
<m_observers>: Change type to vector of observers.
Change-Id: Iadf7d1fa25049cfb089e6b1b429ddebc548825ab
As we turn on more modules by default for all ports, the number of
options has been increasing. The sim-options module has a limit on
the number of options it can support, and if it's exceeded, it likes
to go into an infinite loop. Increase the ceiling and add an assert
so we abort right away instead of hanging.
This will be needed to turn on hw support for v850 as it will then
exceed the current limit.
Every port using this sets the 1st arg to yes and the 2nd arg to "".
These are the defaults we probably want anyways in order to unify the
codebase, so move them to the macro and only allow ports to declare
extra hardware models.
Don't create a symbolic link to tmpdir/ldscripts if it exists.
PR ld/27771
* testsuite/ld-bootstrap/bootstrap.exp: Create a symbolic link
to tmpdir/ldscripts only if it doesn't exist.
While doing some changes, some code failed to compile because it used
the scoped_debug_start_end macro, but couldn't find the CONCAT macro.
Fix that by making common-debug.h include preprocessor.h, the header
file that provides CONCAT.
gdbsupport/ChangeLog:
* common-debug.h: Include preprocessor.h.
Change-Id: Ibf863a932a18cba9a57b4bd72df538ef52d39127
The change implementing the .persistent family of sections broke the
existing support for the .persistent.bss section in the compiler:
int a __attribute__ ((section (".persistent.bss")));
t.s: Assembler messages:
t.s:4: Warning: setting incorrect section type for .persistent.bss
The compiler encodes it as @nobits but the assembler expects @progbits.
The assembler is incorrect and should treat the section like the compiler.
bfd/
* elf.c (special_sections_p): Add .persistent.bss.
gas/
* testsuite/gas/elf/section25.d: Run it everywhere.
* testsuite/gas/elf/section26.d: Likewise.
* testsuite/gas/elf/section26.s: Add test for .persistent.bss.
Display literal value loaded with l32r opcode as a part of disassembly.
This significantly simplifies reading of disassembly output.
2020-04-23 Max Filippov <jcmvbkbc@gmail.com>
opcodes/
* xtensa-dis.c (print_xtensa_operand): For PC-relative operand
of l32r fetch and display referenced literal value.
Output literals as 4-byte words, not as separate bytes.
2021-04-23 Max Filippov <jcmvbkbc@gmail.com>
opcodes/
* xtensa-dis.c (print_insn_xtensa): Set info->bytes_per_chunk
to 4 for literal disassembly.
While working on some changes to 'info sources' I ran into a situation
where I was seeing the same source files reported twice in the output
of the 'info sources' command when using either .gdb_index or the
.debug_name index.
I traced the problem back to some caching in
dwarf2_base_index_functions::map_symbol_filenames; when called GDB
caches the set of filenames, but, filesnames are not removed as the
index entries are expanded into full symtabs. As a result we can end
up seeing filenames reported both from a full symtab _and_ from
a (stale) previously cached index entry.
Now, obviously, when seeing a problem like this the "correct" fix is
to remove the stale entries from the cache, however, I ran a few
experiments to see why this wasn't really hitting us anywhere, and, as
far as I can tell, ::map_symbol_filenames is only called from three
places:
1. The mi command -file-list-exec-source-files,
2. The 'info sources' command, and
3. Filename completion
However, the result of this "bug" is that we will see duplicate
filenames, and readline's completion mechanism already removes
duplicates, so for case #3 we will never see any problems.
Cases #1 and #2 are basically the same, and in each case, to see a
problem we need to ensure we craft the test in a particular way, start
up ensuring we have some unexpected symtabs, then run one of the
commands to populate the cache, then expand one of the symtabs, and
list the sources again. At this point you'll see duplicate entries in
the results. Hardly surprising we haven't randomly hit this situation
in testing.
So, considering that use cases #1 and #2 are certainly not "high
performance" code (i.e. I don't think these justify the need for
caching) this leaves use case #3. Does this use justify the need for
caching? Well the psymbol_functions::map_symbol_filenames function
doesn't seem to do any extra caching, and within
dwarf2_base_index_functions::map_symbol_filenames, the only expensive
bit appears to be the call to dw2_get_file_names, and this already
does its own caching via this_cu->v.quick->file_names.
The upshot of all this analysis was that I'm not convinced the need
for the additional caching is justified, and so, I propose that to fix
the bug in GDB, I just remove the extra caching (for now).
If we later find that the caching _was_ useful, then we can
reintroduce it, but add it back such that it doesn't reintroduce this
bug.
As I was changing dwarf2_base_index_functions::map_symbol_filenames I
replaced the use of htab_up with std::unordered_set.
Tested using target_boards cc-with-debug-names and dwarf4-gdb-index.
gdb/ChangeLog:
* dwarf2/read.c: Add 'unordered_set' include.
(dwarf2_base_index_functions::map_symbol_filenames): Replace
'visited' hash table with 'qfn_cache' unordered_set. Remove use
of per_Bfd->filenames_cache cache, and use function local
filenames_cache instead. Reindent.
* dwarf2/read.h (struct dwarf2_per_bfd) <filenames_cache>: Delete.
gdb/testsuite/ChangeLog:
* gdb.base/info_sources.exp: Add new tests.
These are marker relocations, so together with their bit size being zero
their byte size should be zero as well. This is expressed by a value of
3, not 0.
LEA behavior without a 64-bit destination is independent of address size
- in particular LEA with 32-bit addressing and 64-bit destination is the
same as LEA with 64-bit addressing and 32-bit destination. IOW checking
merely i.prefix[ADDR_PREFIX] is insufficient. This also means wrong
relocation types (R_X86_64_32S when R_X86_64_32 is needed) were used so
far in such cases.
Note that in one case in build_modrm_byte() the 64-bit check came too
early altogether, and hence gets dropped in favor of the one included in
the new helper. This is benign to non-64-bit code from all I can tell,
but the failure to clear disp16 could have been a latent problem.
In preparation for extending the conditions here defer this check until
operands have been parsed, as certain further attributes will need to
be known for determinig applicability of this check to be correct to
LEA.
