The purpose of this concept is to turn the load of debugging
information off, either globally (via the '--readnever' option), or
objfile-specific. The implementation proposed here is an extension of
the patch distributed with Fedora GDB; looking at the Fedora patch
itself and the history, one can see some reasons why it was never
resubmitted:
- The patch appears to have been introduced as a workaround, at
least initially;
- The patch is far from perfect, as it simply shunts the load of
DWARF debugging information, without really worrying about the
other debug format.
- Who really does non-symbolic debugging anyways?
One use of this feature is when a user simply wants to do the
following sequence: attach, dump core, detach. Loading the debugging
information in this case is an unnecessary cause of delay.
This patch expands the version shipped with Fedora GDB in order to
make the feature available for all the debuginfo backends, not only
for DWARF. It also implements a per-objfile flag which can be
activated by using the "-readnever" command when using the
'add-symbol-file' or 'symbol-file' commands.
It's also worth mentioning that this patch tests whether GDB correctly
fails to initialize if both '--readnow' and '--readnever' options are
passed.
Tested on the BuildBot.
gdb/ChangeLog:
2017-12-01 Andrew Cagney <cagney@redhat.com>
Joel Brobecker <brobecker@adacore.com>
Sergio Durigan Junior <sergiodj@redhat.com>
* NEWS (Changes since GDB 8.0: Mention new '--readnever'
feature.
* coffread.c (coff_symfile_read): Do not map over sections with
'coff_locate_sections' if readnever is on.
* dwarf2read.c (dwarf2_has_info): Return 0 if
readnever is on.
* elfread.c (elf_symfile_read): Do not map over sections with
'elf_locate_sections' if readnever is on.
* main.c (validate_readnow_readnever): New function.
(captured_main_1): Add support for --readnever.
(print_gdb_help): Document --readnever.
* objfile-flags.h (enum objfile_flag) <OBJF_READNEVER>: New
flag.
* symfile.c (readnever_symbol_files): New global.
(symbol_file_add_with_addrs): Set 'OBJF_READNEVER' when
'READNEVER_SYMBOL_FILES' is set.
(validate_readnow_readnever): New function.
(symbol_file_command): Handle '-readnever' option.
Call 'validate_readnow_readnever'.
(add_symbol_file_command): Handle '-readnever' option.
Call 'validate_readnow_readnever'.
(_initialize_symfile): Document new '-readnever' option for
both 'symbol-file' and 'add-symbol-file' commands.
* top.h (readnever_symbol_files): New extern global.
* xcoffread.c (xcoff_initial_scan): Do not read debug
information if readnever is on.
gdb/doc/ChangeLog:
2017-12-01 Andrew Cagney <cagney@redhat.com>
Joel Brobecker <brobecker@adacore.com>
Sergio Durigan Junior <sergiodj@redhat.com>
* gdb.texinfo (File Options): Document --readnever.
(Commands to Specify Files): Likewise, for 'symbol-file' and
'add-symbol-file'.
gdb/testsuite/ChangeLog:
2017-12-01 Joel Brobecker <brobecker@adacore.com>
Sergio Durigan Junior <sergiodj@redhat.com>
Pedro Alves <palves@redhat.com>
* gdb.base/readnever.c, gdb.base/readnever.exp: New files.
This is a bug that's been detected while doing the readnever work.
If you use 'symbol-file' or 'add-symbol-file', the position of each
argument passed to the command matters. This means that if you do:
(gdb) symbol-file -readnow /foo/bar
The symbol file specified will (correctly) have all of its symbols
read by GDB (because of the -readnow flag). However, if you do:
(gdb) symbol-file /foo/bar -readnow
GDB will silently ignore the -readnow flag, because it was specified
after the filename. This is not a good thing to do and may confuse
the user.
To address that, I've modified the argument parsing mechanisms of
symbol_file_command and add_symbol_file_command to be
"position-independent". I have also added one error call at the end
of add_symbol_file_command's argument parsing logic, which now clearly
complains if no filename has been specified. Both commands now
support the "--" option to stop argument processing.
This patch provides a testcase for both commands, in order to make
sure that the argument order does not matter. It has been
regression-tested on BuildBot.
gdb/ChangeLog:
2017-12-01 Sergio Durigan Junior <sergiodj@redhat.com>
* symfile.c (symbol_file_command): Call
'symbol_file_add_main_1' only after processing all command
line options.
(add_symbol_file_command): Modify logic to make arguments
position-independent.
gdb/testsuite/ChangeLog:
2017-12-01 Sergio Durigan Junior <sergiodj@redhat.com>
* gdb.base/relocate.exp: Add tests to guarantee that arguments
to 'symbol-file' and 'add-symbol-file' can be
position-independent.
One of our users reported that trying to print the following expression,
caused GDB to SEGV:
(gdb) print some_package.some_type (val)
In this particular instance, the crash occurred inside ada_args_match
because it is given a NULL "func", leading to the SEGV because of:
struct type *func_type = SYMBOL_TYPE (func);
This NULL symbol comes from a list of symbols which was given to
ada_resolve_function (parameter called "syms") which then iterates
over each of them to discard the ones that don't match the actuals:
for (k = 0; k < nsyms; k += 1)
{
struct type *type = ada_check_typedef (SYMBOL_TYPE (syms[k].symbol));
if (ada_args_match (syms[k].symbol, args, nargs)
&& (fallback || return_match (type, context_type)))
[...]
}
What's really interesting is that, when entering the block above for
the first time, all entries in SYMS have a valid (non-NULL) symbol.
However, once we return from the call to ada_check_typedef, the first
entry of our SYMS table gets set to all zeros:
(gdb) p syms[0]
$2 = {symbol = 0x0, block = 0x0}
Hence the call to ada_args_match with a NULL symbol, and the ensuing
SEGV.
To find out why this happen, we need to step back a little and look
at how syms was allocated. This list of symbols comes from a symbol
lookup, which means ada_lookup_symbol_list_worker. We have our first
hint when we look at the function's documentation and see:
This vector is transient---good only to the next call of
ada_lookup_symbol_list.
Implementation-wise, this is done by using a static global obstack,
which we just re-initialize each time ada_lookup_symbol_list_worker
gets called:
obstack_free (&symbol_list_obstack, NULL);
obstack_init (&symbol_list_obstack);
This property was probably established in order to facilitate the use
of the returned vector, since the users of that function would not have
to worry about releasing that memory when no longer needed. However,
I found during this investigation that it is all to easy to indirectly
trigger another symbol lookup while still using the results of a previous
lookup.
In our particular case, there is the call to ada_check_typedef, which
leads to check_typedef. As it happens, my first symbol had a type which
was a typedef to a stub type, so check_typedef calls lookup_symbol to
find the non-stub version. This in turn eventually leads us back to
ada_lookup_symbol_list_worker, where the first thing it does is free
the memory area when our list of symbols have been residing and then
recreates a new one. in other words, SYMS then becomes a dangling
pointer!
This patch fixes the issue by having ada_lookup_symbol_list_worker
return a copy of the list of symbols, with the responsibility of
deallocating that list now transfered to the users of that list.
More generally speaking, it is absolutely amazing that we haven't seen
consequences of this issue before. This can happen fairly frequently.
For instance, I found that ada-exp.y::write_var_or_type calls
ada_lookup_symbol_list, and then, while processing that list, calls
select_possible_type_sym, which leads to ada_prefer_type, eventually
leading to ada_check_typedef again (via eg. ada_is_array_descriptor_type).
Even more amazing is the fact that, while I was able to produce multiple
scenarios where the corruption occurs, none of them leads to incorrect
behavior at the user level. In other words, it requires a very precise
set of conditions for the corruption to become user-visible, and
despite having a megalarge program where the crash occured, using that
as a template for creating a reproducer did not work (pb goes away).
This is why this patch does not come with a reproducer. On the other hand,
this should not be a problem in terms of testing coverage, as the changes
are made in common areas which, at least for the most part, are routinely
exercised during testing.
gdb/ChangeLog:
* ada-lang.c (symbol_list_obstack): Delete.
(resolve_subexp): Make sure "candidates" gets xfree'ed.
(ada_lookup_symbol_list_worker): Remove the limitation that
the result is only good until the next call, now making it
the responsibility of the caller to free the result when no
longer needed. Adjust the function's intro comment accordingly.
(ada_lookup_symbol_list): Adjust the function's intro comment.
(ada_iterate_over_symbols): Make sure "results" gets xfree'ed.
(ada_lookup_encoded_symbol, get_var_value): Likewise.
(_initialize_ada_language): Remove symbol_list_obstack
initialization.
* ada-exp.y (block_lookup): Make sure "syms" gets xfree'ed.
(write_var_or_type, write_name_assoc): Likewise.
Tested on x86_64-linux.
The purpose of this concept is to turn the load of debugging
information off, either globally (via the '--readnever' option), or
objfile-specific. The implementation proposed here is an extension of
the patch distributed with Fedora GDB; looking at the Fedora patch
itself and the history, one can see some reasons why it was never
resubmitted:
- The patch appears to have been introduced as a workaround, at
least initially;
- The patch is far from perfect, as it simply shunts the load of
DWARF debugging information, without really worrying about the
other debug format.
- Who really does non-symbolic debugging anyways?
One use of this feature is when a user simply wants to do the
following sequence: attach, dump core, detach. Loading the debugging
information in this case is an unnecessary cause of delay.
This patch expands the version shipped with Fedora GDB in order to
make the feature available for all the debuginfo backends, not only
for DWARF. It also implements a per-objfile flag which can be
activated by using the "-readnever" command when using the
'add-symbol-file' or 'symbol-file' commands.
It's also worth mentioning that this patch tests whether GDB correctly
fails to initialize if both '--readnow' and '--readnever' options are
passed.
Tested on the BuildBot.
gdb/ChangeLog:
2017-12-01 Andrew Cagney <cagney@redhat.com>
Joel Brobecker <brobecker@adacore.com>
Sergio Durigan Junior <sergiodj@redhat.com>
* NEWS (Changes since GDB 8.0: Mention new '--readnever'
feature.
* coffread.c (coff_symfile_read): Do not map over sections with
'coff_locate_sections' if readnever is on.
* dwarf2read.c (dwarf2_has_info): Return 0 if
readnever is on.
* elfread.c (elf_symfile_read): Do not map over sections with
'elf_locate_sections' if readnever is on.
* main.c (validate_readnow_readnever): New function.
(captured_main_1): Add support for --readnever.
(print_gdb_help): Document --readnever.
* objfile-flags.h (enum objfile_flag) <OBJF_READNEVER>: New
flag.
* symfile.c (readnever_symbol_files): New global.
(symbol_file_add_with_addrs): Set 'OBJF_READNEVER' when
'READNEVER_SYMBOL_FILES' is set.
(validate_readnow_readnever): New function.
(symbol_file_command): Handle '-readnever' option.
Call 'validate_readnow_readnever'.
(add_symbol_file_command): Handle '-readnever' option.
Call 'validate_readnow_readnever'.
(_initialize_symfile): Document new '-readnever' option for
both 'symbol-file' and 'add-symbol-file' commands.
* top.h (readnever_symbol_files): New extern global.
* xcoffread.c (xcoff_initial_scan): Do not read debug
information if readnever is on.
gdb/doc/ChangeLog:
2017-12-01 Andrew Cagney <cagney@redhat.com>
Joel Brobecker <brobecker@adacore.com>
Sergio Durigan Junior <sergiodj@redhat.com>
* gdb.texinfo (File Options): Document --readnever.
(Commands to Specify Files): Likewise, for 'symbol-file' and
'add-symbol-file'.
gdb/testsuite/ChangeLog:
2017-12-01 Joel Brobecker <brobecker@adacore.com>
Sergio Durigan Junior <sergiodj@redhat.com>
Pedro Alves <palves@redhat.com>
* gdb.base/readnever.c, gdb.base/readnever.exp: New files.
On irc, Pedro pointed out that dependencies for objects in
subdirectories didn't seem to be working.
The bug was that the "-include" for .deps files was using the wrong file
name for subdirectory objects; e.g., for cli/cli-decode.o it was trying
to open .deps/cli/cli-decode.o, whereas the correct file is
cli/.deps/cli-decode.o.
This patch changes how the dep files are found. Tested by touching a
source file and rebuilding cli/cli-decode.o.
2017-12-01 Tom Tromey <tom@tromey.com>
* Makefile.in (all_deps_files): New variable.
Include .Po files using all_deps_files.
This is the previously mentioned patch to get rid of
unstructured/ioctl-based procfs support in procfs.c. Given that support
for structured procfs was introduced in Solaris 2.6 back in 1997 and
we're just removing support for Solaris < 10, there's no point in
carrying that baggage (and tons of support for IRIX and OSF/1 as well)
around any longer.
Most of the patch should be straightforward (removing support for
!NEW_PROC_API, non-Solaris OSes and pre-Solaris 10 quirks).
Only a few points need explanations:
* <sys/syscall.h> was already included unconditionally in most places,
so there's no need to have guards in a few remaining ones.
* configure.host already obsoletes i?86-*-sysv4.2, i?86-*-sysv5, so
NEW_PROC_API detection for those in configure.ac can go.
* I'm still including <sys/procfs.h> with #define _STRUCTURED_PROC 1.
Theoretically, it would be better to include <procfs.h> on Solaris
(which includes that define), but that breaks the build over
<procfs.h> vs. gdb's "procfs.h", and doesn't exist on Linux.
* I've regenerated syscall_table[] in proc-events.c with a small script
from Solaris 10, 11.3, 11.4 <sys/syscall.h>, so there should be no
traces of older Solaris versions and other OSes left.
* prsysent_t and DYNAMIC_SYSCALLS was only used for AIX 5, but AIX
doesn't use procfs.c any longer, so all related code can go.
The patch was generated with diff -w so one can easier see changes
without being distracted by simple reindentations.
So far, it has only been compiled and smoke-tested on
amd64-pc-solaris2.1[01], sparcv9-sun-solaris2.1[01], and
x86_64-pc-linux-gnu. Certainly needs more testing (Solaris 11.3
vs. 11.4, 32-bit gdb, testsuite once I've figured out what's wrong on
Solaris 10 etc.), but it's enough to get a first impression how much
cleanup is possible here.
* configure.ac Don't check for sys/fault.h, sys/syscall.h,
sys/proc.h.
(NEW_PROC_API): Remove.
(prsysent_t, pr_sigset_t, pr_sigaction64_t, pr_siginfo64_t):
Likewise.
* common/common.m4 (GDB_AC_COMMON): Don't check for sys/syscall.h.
* configure: Regenerate.
* config.in: Regenerate.
* gdbserver/configure: Regenerate.
* gdbserver/config.in: Regenerate.
* i386-sol2-nat.c (_initialize_amd64_sol2_nat): Remove
NEW_PROC_API test.
* sparc-sol2-nat.c (_initialize_sparc_sol2_nat): Likewise.
* linux-btrace.c: Remove HAVE_SYS_SYSCALL_H test.
* proc-api.c: Remove !NEW_PROC_API support.
Remove HAVE_SYS_PROC_H and HAVE_SYS_USER_H tests.
Remove tests for macros always defined on Solaris.
* proc-events.c: Remove !NEW_PROC_API support.
Remove Remove HAVE_SYS_SYSCALL_H, HAVE_SYS_PROC_H and
HAVE_SYS_USER_H tests.
(init_syscall_table): Remove non-Solaris syscalls.
Remove tests for syscalls present on all Solaris versions.
Add missing Solaris 10+ syscalls.
(signal_table): Remove non-Solaris signals.
Remove tests for signals present on all Solaris versions.
(fault_table): Remove non-Solaris faults.
Remove tests for faults present on all Solaris versions.
* proc-flags.c: Remove !NEW_PROC_API support.
(pr_flag_table): Remove non-Solaris and pre-Solaris 7 comments.
Remove non-Solaris flags.
* proc-why.c: Remove !NEW_PROC_API support.
(pr_why_table): Remove meaningless comments.
Remove tests for reasons present on all Solaris versions.
Remove OSF/1 cases.
(proc_prettyfprint_why): Likewise.
* procfs.c: Remove !NEW_PROC_API and DYNAMIC_SYSCALLS support.
Remove HAVE_SYS_FAULT_H and HAVE_SYS_SYSCALL_H tests.
Remove WA_READ test, IRIX watchpoint support.
(gdb_sigset_t, gdb_sigaction_t, gdb_siginfo_t): Replace by base
types. Change users.
(gdb_praddset, gdb_prdelset, gdb_premptysysset, gdb_praddsysset)
(gdb_prdelset, gdb_pr_issyssetmember): Replace by base macros.
Change callers.
Remove CTL_PROC_NAME_FMT tests.
(gdb_prstatus_t, gdb_lwpstatus_t): Replace by base types. Change
users.
(sysset_t_size): Remove. Use sizeof (sysset_t) in callers.
Remove PROCFS_DONT_PIOCSSIG_CURSIG support.
(proc_modify_flag): Replace GDBRESET by PCUNSET.
Remove PR_ASYNC, PR_KLC tests.
(proc_unset_inherit_on_fork): Remove PR_ASYNC test.
(proc_parent_pid): Remove PCWATCH etc. tests.
(proc_set_watchpoint): Remove !PCWATCH && !PIOCSWATCH support.
Remove PCAGENT test.
(proc_get_nthreads) [PIOCNTHR && PIOCTLIST]: Remove.
Remove SYS_lwpcreate || SYS_lwp_create test.
(proc_get_current_thread): Likewise.
[PIOCNTHR && PIOCTLIST]: Remove.
[PIOCLSTATUS]: Remove.
(procfs_debug_inferior): Remove non-Solaris cases, conditionals.
[PRFS_STOPEXEC]: Remove.
(syscall_is_lwp_exit): Remove non-Solaris cases, conditionals.
(syscall_is_exit): Likewise.
(syscall_is_exec): Likewise.
(syscall_is_lwp_create): Likewise.
Remove SYS_syssgi support.
(procfs_wait): Remove PR_ASYNC, !PIOCSSPCACT tests.
[SYS_syssgi]: Remove.
Remove non-Solaris cases, conditionals.
(unconditionally_kill_inferior) [PROCFS_NEED_PIOCSSIG_FOR_KILL]:
Remove.
(procfs_init_inferior) [SYS_syssgi]: Remove.
(procfs_set_exec_trap) [PRFS_STOPEXEC]: Remove.
(procfs_inferior_created) [SYS_syssgi]: Remove.
(procfs_set_watchpoint): Remove !AIX5 test.
(procfs_stopped_by_watchpoint): Remove FLTWATCH test, FLTKWATCH
case.
(mappingflags) [MA_PHYS]: Remove.
(info_mappings_callback): Remove PCAGENT test.
Remove PIOCOPENLWP || PCAGENT test.
Building current gdb mainline with gcc 7.1 on Solaris 11.4 fails:
/vol/src/gnu/gdb/gdb/dist/gdb/sol-thread.c: In function `void _initialize_sol_thread()':
/vol/src/gnu/gdb/gdb/dist/gdb/sol-thread.c:1229:66: error: invalid conversion from `void (*)(char*, int)' to `void (*)(const char*, int)' [-fpermissive]
_("Show info on Solaris user threads."), &maintenanceinfolist);
^
In file included from /vol/src/gnu/gdb/gdb/dist/gdb/completer.h:21:0,
from /vol/src/gnu/gdb/gdb/dist/gdb/symtab.h:31,
from /vol/src/gnu/gdb/gdb/dist/gdb/language.h:26,
from /vol/src/gnu/gdb/gdb/dist/gdb/frame.h:72,
from /vol/src/gnu/gdb/gdb/dist/gdb/gdbarch.h:39,
from /vol/src/gnu/gdb/gdb/dist/gdb/defs.h:557,
from /vol/src/gnu/gdb/gdb/dist/gdb/sol-thread.c:51:
/vol/src/gnu/gdb/gdb/dist/gdb/command.h:140:33: note: initializing argument 3 of `cmd_list_element* add_cmd(const char*, command_class, void (*)(const char*, int), const char*, cmd_list_element**)'
extern struct cmd_list_element *add_cmd (const char *, enum command_class,
^~~~~~~
The following patch allows compilation to succeed on i386-pc-solaris2.11
and sparc-sun-solaris2.11.
* sol-thread.c (info_solthreads): Constify args.
Cast args to void *.
mpfr.h uses a non-portable test to guess if intmax_t is available and
if API functions using intmax_t should be exposed. Define
MPFR_USE_INTMAX_T to override the non-portable test and always expose
these functions. This fixes the build on platforms where the test
guesses incorrectly.
gdb/ChangeLog:
* target-float.c [HAVE_LIBMPFR]: Define MPFR_USE_INTMAX_T.
This patch fixes a potential issue which was noticed by code inspection:
ada-lang.c::to_fixed_range_type uses gdbtypes.c::create_static_range_type
to create most of the range type, which relies on create_range_type to
do most of the work. The latter has the following piece of code which
sets the length of the range type to match the length of the index_type:
if (TYPE_STUB (index_type))
TYPE_TARGET_STUB (result_type) = 1;
else
TYPE_LENGTH (result_type) = TYPE_LENGTH (check_typedef (index_type));
In Ada, it is actually possible to have a range type whose size
is smaller than its base type. For instance, with:
type Unsigned2_T is range 0 .. 2 ** 16 - 1;
for Unsigned2_T'SIZE use 16;
The compiler generates the following DWARF:
.uleb128 0x3 # (DIE (0x4e) DW_TAG_subrange_type)
.byte 0x2 # DW_AT_byte_size
.byte 0 # DW_AT_lower_bound
.value 0xffff # DW_AT_upper_bound
.long .LASF64 # DW_AT_name: "try__unsigned2_t___XDLU_0__65535"
.long 0x616 # DW_AT_type
... which points to the following base type...
.uleb128 0x1d # (DIE (0x616) DW_TAG_base_type)
.byte 0x4 # DW_AT_byte_size
.byte 0x5 # DW_AT_encoding
.long .LASF57 # DW_AT_name: "try__Tunsigned2_tB"
# DW_AT_artificial
... which has a size of 4 bytes.
With a type like this one, create_range_type returns a type whose
size is 4 bytes, instead of 2, which is not what we we would normally
expect.
Currently, this function is only used to handle array index types,
so the length of the type actually does not matter and there should
not be any user-visible consequences of the current behavior. But
it seems best to plug this latent bug now, rather than wait for it
to surface....
gdb/ChangeLog:
* ada-lang.c (to_fixed_range_type): Make sure that the size
of the range type being returned is the same as the size
of the range type being fixed.
Tested on x86_64-linux, no regression.
Trying to set a breakpoint in a function with an ABI tag does not work
currently. E.g., debugging gdb itself, we see this with the
"string_printf" function:
(top-gdb) b string_print [TAB]
(top-gdb) b string_printf[abi:cxx11](char const*, ...) [RET]
No source file named string_printf[abi.
Make breakpoint pending on future shared library load? (y or [n])
Quoting doesn't help:
(top-gdb) b 'string_printf[abi:cxx11]'(char const*, ...)
malformed linespec error: unexpected string, "(char const*, ...)"
(top-gdb) b 'string_printf[abi:cxx11](char const*, ...)'
No source file named string_printf[abi.
Make breakpoint pending on future shared library load? (y or [n]) n
This patch fixes this, and takes it a bit further.
The actual symbol name as demangled by libiberty's demangler is really
string_printf[abi:cxx11](char const*, ...)
however, this patch makes it possible to set the breakpoint with
string_printf(char const*, ...)
too. I.e., ignoring the ABI tag.
And to match, it teaches the completer to complete the symbol name
without the ABI tag, i.e.,
"string_pri<TAB>" -> "string_printf(char const*, ...)"
If however, you really want to break on a symbol with the tag, then
you simply start writing the tag, and GDB will preserve it, like:
"string_printf[a<TAB>" -> "string_printf[abi:cxx11](char const*, ...)"
Grows the gdb.linespec/ tests like this:
-# of expected passes 8977
+# of expected passes 9176
gdb/ChangeLog:
2017-11-29 Pedro Alves <palves@redhat.com>
PR c++/19436
* NEWS: Mention setting breakpoints on functions with C++ ABI
tags.
* completer.h (completion_match_for_lcd) <match,
mark_ignored_range>: New methods.
<finish>: Consider ignored ranges.
<clear>: Clear ignored ranges.
<m_ignored_ranges, m_finished_storage>: New fields.
* cp-support.c (cp_search_name_hash): Ignore ABI tags.
(cp_symbol_name_matches_1, cp_fq_symbol_name_matches): Pass the
completion_match_for_lcd pointer to strncmp_iw_with_mode.
(test_cp_symbol_name_cmp): Add [abi:...] tags unit tests.
* language.c (default_symbol_name_matcher): Pass the
completion_match_for_lcd pointer to strncmp_iw_with_mode.
* linespec.c (linespec_lexer_lex_string): Don't tokenize ABI tags.
* utils.c (skip_abi_tag): New function.
(strncmp_iw_with_mode): Add completion_match_for_lcd parameter.
Handle ABI tags.
* utils.h (strncmp_iw_with_mode): Add completion_match_for_lcd
parameter.
gdb/testsuite/ChangeLog:
2017-11-29 Pedro Alves <palves@redhat.com>
PR c++/19436
* gdb.linespec/cpls-abi-tag.cc: New file.
* gdb.linespec/cpls-abi-tag.exp: New file.
gdb/doc/ChangeLog:
2017-11-29 Pedro Alves <palves@redhat.com>
PR c++/19436
* gdb.texinfo (Debugging C Plus Plus): Document setting
breakpoints in functions with ABI tags.
This patch teaches GDB about setting breakpoints in all scopes
(namespaces and classes) by default.
Here's a contrived example:
(gdb) b func<tab>
(anonymous namespace)::A::function() Bn::(anonymous namespace)::B::function() function(int, int)
(anonymous namespace)::B::function() Bn::(anonymous namespace)::function() gdb::(anonymous namespace)::A::function()
(anonymous namespace)::B::function() const Bn::(anonymous namespace)::function(int, int) gdb::(anonymous namespace)::function()
(anonymous namespace)::function() Bn::B::func() gdb::(anonymous namespace)::function(int, int)
(anonymous namespace)::function(int, int) Bn::B::function() gdb::A::func()
A::func() Bn::func() gdb::A::function()
A::function() Bn::function() gdb::func()
B::func() Bn::function(int, int) gdb::function()
B::function() Bn::function(long) gdb::function(int, int)
B::function() const func() gdb::function(long)
B::function_const() const function()
(gdb) b function
Breakpoint 1 at 0x4005ce: function. (26 locations)
(gdb) b B::function<tab>
(anonymous namespace)::B::function() B::function() const Bn::B::function()
(anonymous namespace)::B::function() const B::function_const() const
B::function() Bn::(anonymous namespace)::B::function()
(gdb) b B::function
Breakpoint 1 at 0x40072c: B::function. (6 locations)
To get back the original behavior of interpreting the function name as
a fully-qualified name, you can use the new "-qualified" (or "-q")
option/flag (added by this commit). For example:
(gdb) b B::function
(anonymous namespace)::B::function() B::function() const Bn::B::function()
(anonymous namespace)::B::function() const B::function_const() const
B::function() Bn::(anonymous namespace)::B::function()
vs:
(gdb) b -qualified B::function
B::function() B::function() const B::function_const() const
I've chosen "-qualified" / "-q" because "-f" (for "full" or
"fully-qualified") is already taken for "-function".
Note: the "-qualified" option works with both linespecs and explicit
locations. I.e., these are equivalent:
(gdb) b -q func
(gdb) b -q -f func
and so are these:
(gdb) b -q filename.cc:func
(gdb) b -q -s filename.cc -f func
(gdb) b -s filename.cc -q -f func
(gdb) b -s filename.cc -f func -q
To better understand why I consider wild matching the better default,
consider what happens when we get to the point when _all_ of GDB is
wrapped under "namespace gdb {}". I have a patch series that does
that, and when I started debugging that GDB, I immediately became
frustrated. You'd have to write "b gdb::internal_error", "b
gdb::foo", "b gdb::bar", etc. etc., which gets annoying pretty
quickly. OTOH, consider how this makes it very easy to set
breakpoints in classes wrapped in anonymous namespaces. You just
don't think of them, GDB finds the symbols for you automatically.
(At the Cauldron a couple months ago, several people told me that they
run into a similar issue when debugging other C++ projects. One
example was when debugging LLVM, which puts all its code under the
"llvm" namespace.)
Implementation-wise, what the patch does is:
- makes C++ symbol name hashing only consider the last component of
a symbol name. (so that we can look up symbol names by
last-component name only).
- adds a C++ symbol name matcher for symbol_name_match_type::WILD,
which ignores missing leading specifiers / components.
- adjusts a few preexisting testsuite tests to use "-qualified" when
they mean it.
- adds new testsuite tests.
- adds unit tests.
Grows the gdb.linespec/ tests like this:
-# of expected passes 7823
+# of expected passes 8977
gdb/ChangeLog:
2017-11-29 Pedro Alves <palves@redhat.com>
* NEWS: Mention that breakpoints on C++ functions are now set on
on all namespaces/classes by default, and mention "break
-qualified".
* ax-gdb.c (agent_command_1): Adjust to pass a
symbol_name_match_type to new_linespec_location.
* breakpoint.c (parse_breakpoint_sals): Adjust to
get_linespec_location's return type change.
(strace_marker_create_sals_from_location): Adjust to pass a
symbol_name_match_type to new_linespec_location.
(strace_marker_decode_location): Adjust to get_linespec_location's
return type change.
(strace_command): Adjust to pass a symbol_name_match_type to
new_linespec_location.
(LOCATION_HELP_STRING): Add paragraph about wildmatching, and
mention "-qualified".
* c-lang.c (cplus_language_defn): Install cp_search_name_hash.
* completer.c (explicit_location_match_type::MATCH_QUALIFIED): New
enumerator.
(complete_address_and_linespec_locations): New parameter
'match_type'. Pass it down.
(explicit_options): Add "-qualified".
(collect_explicit_location_matches): Pass the requested match type
to the linespec completers. Handle MATCH_QUALIFIED.
(location_completer): Handle "-qualified" combined with linespecs.
* cp-support.c (cp_search_name_hash): New.
(cp_symbol_name_matches_1): Implement wild matching for C++.
(cp_fq_symbol_name_matches): Reimplement.
(cp_get_symbol_name_matcher): Return different matchers depending
on the lookup name's match type.
(selftests::test_cp_symbol_name_matches): Add wild matching tests.
* cp-support.h (cp_search_name_hash): New declaration.
* dwarf2read.c
(selftests::dw2_expand_symtabs_matching::test_symbols): Add
symbols.
(test_dw2_expand_symtabs_matching_symbol): Add wild matching
tests.
* guile/scm-breakpoint.c (gdbscm_register_breakpoint_x): Adjust to
pass a symbol_name_match_type to new_linespec_location.
* linespec.c (linespec_parse_basic): Lookup function symbols using
the parser's symbol name match type.
(convert_explicit_location_to_linespec): New
symbol_name_match_type parameter. Pass it down to
find_linespec_symbols.
(convert_explicit_location_to_sals): Pass the location's name
match type to convert_explicit_location_to_linespec.
(parse_linespec): New match_type parameter. Save it in the
parser.
(linespec_parser_new): Default to symbol_name_match_type::WILD.
(linespec_complete_function): New symbol_name_match_type
parameter. Use it.
(complete_linespec_component): Pass down the parser's recorded
name match type.
(linespec_complete_label): New symbol_name_match_type parameter.
Use it.
(linespec_complete): New symbol_name_match_type parameter. Save
it in the parser and pass it down. Adjust to
get_linespec_location's prototype change.
(find_function_symbols, find_linespec_symbols): New
symbol_name_match_type parameter. Pass it down instead of
assuming symbol_name_match_type::WILD.
* linespec.h (linespec_complete, linespec_complete_function)
(linespec_complete_label): New symbol_name_match_type parameter.
* location.c (event_location::linespec_location): Now a struct
linespec_location.
(EL_LINESPEC): Adjust.
(initialize_explicit_location): Default to
symbol_name_match_type::WILD.
(new_linespec_location): New symbol_name_match_type parameter.
Record it in the location.
(get_linespec_location): Now returns a struct linespec_location.
(new_explicit_location): Also copy func_name_match_type.
(explicit_to_string_internal)
(string_to_explicit_location): Handle "-qualified".
(copy_event_location): Adjust to LINESPEC_LOCATION type change.
Copy symbol_name_match_type fields.
(event_location_deleter::operator()): Adjust to LINESPEC_LOCATION
type change.
(event_location_to_string): Adjust to LINESPEC_LOCATION type
change. Handle "-qualfied".
(string_to_explicit_location): Handle "-qualified".
(string_to_event_location_basic): New symbol_name_match_type
parameter. Pass it down.
(string_to_event_location): Handle "-qualified".
* location.h (struct linespec_location): New.
(explicit_location::func_name_match_type): New field.
(new_linespec_location): Now returns a const linespec_location *.
(string_to_event_location_basic): New symbol_name_match_type
parameter.
(explicit_completion_info::saw_explicit_location_option): New
field.
* mi/mi-cmd-break.c (mi_cmd_break_insert_1): Adjust to pass a
symbol_name_match_type to new_linespec_location.
* python/py-breakpoint.c (bppy_init): Likewise.
* python/python.c (gdbpy_decode_line): Likewise.
gdb/testsuite/ChangeLog:
2017-11-29 Pedro Alves <palves@redhat.com>
* gdb.base/langs.exp: Use -qualified.
* gdb.cp/meth-typedefs.exp: Use -qualified, and add tests without
it.
* gdb.cp/namespace.exp: Use -qualified.
* gdb.linespec/cpcompletion.exp (overload-2, fqn, fqn-2)
(overload-3, template-overload, template-ret-type, const-overload)
(const-overload-quoted, anon-ns, ambiguous-prefix): New
procedures.
(test_driver): Call them.
* gdb.cp/save-bp-qualified.cc: New.
* gdb.cp/save-bp-qualified.exp: New.
* gdb.linespec/explicit.exp: Test -qualified.
* lib/completion-support.exp (completion::explicit_opts_list): Add
"-qualified".
* lib/gdb.exp (gdb_breakpoint): Handle "qualified".
gdb/doc/ChangeLog:
2017-11-29 Pedro Alves <palves@redhat.com>
* gdb.texinfo (Linespec Locations): Document how "function" is
interpreted in C++ and Ada. Document "-qualified".
(Explicit Locations): Document how "-function" is interpreted in
C++ and Ada. Document "-qualified".
A following patch will add support for wild matching for C++ symbols,
making completing on "b push_ba" on a C++ program complete to
std::vector<...>::push_back, std::string::push_back etc., like:
(gdb) b push_ba[TAB]
std::vector<...>::push_back(....)
std::string<...>::push_back(....)
Currently, we compute the "lowest common denominator" between all
completion candidates (what the input line is adjusted to) as the
common prefix of all matches. That's problematic with wild matching
as above, as then we'd end up with TAB changing the input line to
"b std::", losing the original input, like:
(gdb) b push_ba[TAB]
std::vector<...>::push_back(....)
std::string<...>::push_back(....)
(gdb) b std::
while obviously we'd want it to adjust itself to "b push_back(" instead:
(gdb) b push_ba[TAB]
std::vector<...>::push_back(....)
std::string<...>::push_back(....)
(gdb) b push_back(
This patch adds the core code necessary to support this, though
nothing really makes use of it yet in this patch.
gdb/ChangeLog:
2017-11-29 Pedro Alves <palves@redhat.com>
* ada-lang.c (ada_lookup_name_info::matches): Change type of
parameter from completion_match to completion_match_result.
Adjust.
(do_wild_match, do_full_match, ada_symbol_name_matches): Likewise.
* completer.c (completion_tracker::maybe_add_completion): Add
match_for_lcd parameter and use it.
(completion_tracker::add_completion): Likewise.
* completer.h (class completion_match_for_lcd): New class.
(completion_match_result::match_for_lcd): New field.
(completion_match_result::set_match): New method.
(completion_tracker): Add comments.
(completion_tracker::add_completion): Add match_for_lcd parameter.
(completion_tracker::reset_completion_match_result): Reset
match_for_lcd too.
(completion_tracker::maybe_add_completion): Add match_for_lcd
parameter.
(completion_tracker::m_lowest_common_denominator_unique): Extend
comments.
* cp-support.c (cp_symbol_name_matches_1)
(cp_fq_symbol_name_matches): Change type of parameter from
completion_match to completion_match_result. Adjust.
* language.c (default_symbol_name_matcher): Change type of
parameter from completion_match to completion_match_result.
Adjust.
* language.h (completion_match_for_lcd): Forward declare.
(default_symbol_name_matcher): Change type of parameter from
completion_match to completion_match_result.
* symtab.c (compare_symbol_name): Adjust.
(completion_list_add_name): Pass the match_for_lcd to the tracker.
* symtab.h (ada_lookup_name_info::matches): Change type of
parameter from completion_match to completion_match_result.
(symbol_name_matcher_ftype): Likewise, and update comments.
The recent-ish commit e5f25bc5d6 ('Fix "list ambiguous_variable"')
caused a serious regression on PPC64. See
<https://sourceware.org/ml/gdb-patches/2017-11/msg00666.html>.
Basically, after that patch, GDB sets breakpoints in function
descriptors instead of where the descriptors point to, which is
incorrect.
The problem is that GDB now only runs a minsym's address through
gdbarch_convert_from_func_ptr_addr if msymbol_is_text returns true.
However, if the symbol points to a function descriptor,
msymbol_is_text is false since function descriptors are in fact
outside the text section.
The fix is to also run a non-text address through
gdbarch_convert_from_func_ptr_addr, and if that detects that it was
indeed a function descriptor, treat the resulting address as a
function.
While implementing that directly in linespec.c:minsym_found (where the
bad msymbol_is_text check is) fixes the issue, I noticed that
linespec.c:add_minsym has some code that also basically needs to do
the same checks, however it's implemented differently. Also,
add_minsym is calling find_pc_sect_line on non-function symbols, which
also doesn't look right.
So I introduced msymbol_is_function, so that we have a simple place to
consider minsyms and function descriptors.
And then, the only other use of msymbol_is_text is in
find_function_alias_target, which turns out to also be incorrect.
Changing that one to use msymbol_is_function, i.e., to consider
function descriptors too fixes (on PPC64):
-FAIL: gdb.base/symbol-alias.exp: p func_alias
-FAIL: gdb.base/symbol-alias.exp: p *func_alias()
+PASS: gdb.base/symbol-alias.exp: p func_alias
+PASS: gdb.base/symbol-alias.exp: p *func_alias()
And then after that, msymbol_is_text is no longer used anywhere, so it
can be removed.
Tested on x86_64 GNU/Linux, no regressions. Tested on PPC64 GNU/Linux
and results compared to a testrun of e5f25bc5d6db^ (before the
offending commit), also no regressions. (there's a couple new FAILs
and some new symbol name matching unit tests are crashing, but that
looks unrelated).
gdb/ChangeLog:
2017-11-29 Pedro Alves <palves@redhat.com>
* linespec.c (minsym_found, add_minsym): Use msymbol_is_function.
* minsyms.c (msymbol_is_text): Delete.
(msymbol_is_function): New function.
* minsyms.h (msymbol_is_text): Delete.
(msymbol_is_function): New declaration.
* symtab.c (find_function_alias_target): Use msymbol_is_function.
Joel pointed out that gdb snapshots were broken by my Makefile patch
series. The bug is that rmdir in distclean was failing, because the
directory did not exist. This fixes the bug by only invoking rmdir when
the directory exists.
Tested using "src-release.sh gdb".
2017-11-29 Tom Tromey <tom@tromey.com>
* Makefile.in (distclean): Handle the case where rmdir fails.
This updates the usage text for the add-symbol-file, symbol-file, and
load commands.
gdb/ChangeLog
2017-11-29 Tom Tromey <tom@tromey.com>
* symfile.c (_initialize_symfile): Update usage text for
add-symbol-file, symbol-file, load.
This patch updates add-symbol-file help and error text.
It changes add-symbol-file to throw an exception if "-s" is seen but
not all of the arguments are given. Previously this was silently
ignored.
It changes the unrecognized argument message to more clearly state
what went wrong.
Finally, it updates the usage line in the help text to follow GNU
style regarding "metasyntactic variables"; a change I believe should
be made to all gdb help messages.
gdb/ChangeLog
2017-11-29 Tom Tromey <tom@tromey.com>
* symfile.c (add_symbol_file_command): Error if some arguments to
-s are missing. Change unrecognized-argument error message.
(_initialize_symfile): Fix usage text for add-symbol-file.
gdb/testsuite/ChangeLog
2017-11-29 Tom Tromey <tom@tromey.com>
* gdb.base/relocate.exp: Update invalid argument test.
Add new tests for invalid arguments.
This removes REMOTE_OBS from the Makefile. It is no longer needed, as
remote support is always built into gdb. The relevant sources are now
added to COMMON_SFILES, where they are treated like other ordinary
sources.
ChangeLog
2017-11-27 Tom Tromey <tom@tromey.com>
* Makefile.in (REMOTE_OBS): Remove.
(SFILES): Remove remote sources.
(COMMON_SFILES): Add remote sources.
(ALLDEPFILES): Remove dcache.c.
Move the object files corresponding to target/*.c to the target
subdirectory in the build tree.
ChangeLog
2017-11-27 Tom Tromey <tom@tromey.com>
* Makefile.in (SUBDIR_TARGET_SRCS, SUBDIR_TARGET_OBS): New
variables.
(SFILES): Use SUBDIR_TARGET_SRCS.
(COMMON_OBS): Use SUBDIR_TARGET_OBS. Remove waitstatus.o.
(CONFIG_SRC_SUBDIR): Add target.
(%.o): Remove target rule.
While working on the previous patch, I found a few .o files whose
corresponding .c file was not mentioned in Makefile.in. This patch
fixes the problem. I pulled this out separately to make it simpler to
review.
ChangeLog
2017-11-27 Tom Tromey <tom@tromey.com>
* Makefile.in (COMMON_OBS): Remove filename-seen-cache.o,
registry.o, thread-fsm.o, debug.o.
(COMMON_SFILES): Add filename-seen-cache.c, registry.c,
thread-fsm.c, debug.c.
This introduces a new COMMON_SFILES variable, and then defines some of
COMMON_OBS in terms of this new variable. This simpifies adding a new
ordinary source file.
ChangeLog
2017-11-27 Tom Tromey <tom@tromey.com>
* Makefile.in (COMMON_SFILES): New.
(SFILES): Move some entries to COMMON_SFILES.
(COMMON_OBS): Use COMMON_SFILES.
Change YYOBJ to be defined in terms of YYFILES.
ChangeLog
2017-11-27 Tom Tromey <tom@tromey.com>
* Makefile.in (YYFILES): Update comment.
(YYOBJ): Redefine.
Move the object files corresponding to python/*.c to the python
subdirectory in the build tree.
Because special CFLAGS are passed just to Python compilations, this
patch also required the addition of a pattern rule to update
INTERNAL_CFLAGS for here.
ChangeLog
2017-11-27 Tom Tromey <tom@tromey.com>
* Makefile.in (SUBDIR_PYTHON_OBS): Redefine.
(CONFIG_SRC_SUBDIR): Add python.
(%.o): Remove python rule.
(python/%.o): New rule.
* configure: Rebuild.
* configure.ac (CONFIG_OBS): Refer to python/python.o
Move the object files corresponding to guile/*.c to the guile
subdirectory in the build tree.
ChangeLog
2017-11-27 Tom Tromey <tom@tromey.com>
* configure: Rebuild.
* configure.ac (CONFIG_OBS): Refer to guile/guile.o.
* Makefile.in (SUBDIR_GUILE_OBS): Redefine.
(CONFIG_SRC_SUBDIR): Add guile.
(%.o): Remove guile rule.
Move the object files corresponding to unittests/*.c to the unittests
subdirectory in the build tree.
ChangeLog
2017-11-27 Tom Tromey <tom@tromey.com>
* Makefile.in (SUBDIR_UNITTESTS_OBS): Redefine.
(%.o): Remove unittests rule.
(CONFIG_SRC_SUBDIR): Add unittests.
Move the object files corresponding to tui/*.c to the tui subdirectory
in the build tree.
ChangeLog
2017-11-27 Tom Tromey <tom@tromey.com>
* Makefile.in (SUBDIR_TUI_OBS): Redefine.
(CONFIG_SRC_SUBDIR): Add tui.
(%.o): Remove tui rule.
Move the object files corresponding to compile/*.c to the compile
subdirectory in the build tree.
ChangeLog
2017-11-27 Tom Tromey <tom@tromey.com>
* Makefile.in (SUBDIR_GCC_COMPILE_OBS): Redefine.
(%.o): Remove compile rule.
(CONFIG_SRC_SUBDIR): Add compile.
Move object files corresponding to mi/*.c to a subdirectory in the
build tree.
ChangeLog
2017-11-27 Tom Tromey <tom@tromey.com>
* Makefile.in (SUBDIR_MI_OBS): Redefine.
(%.o): Remove mi rule.
(CONFIG_SRC_SUBDIR): Add mi.
(COMMON_OBS): Use mi/mi-common.o
Following the "arch" move, this moves the object files corresponding
to the cli/*.c source files to the "cli" build directory.
ChangeLog
2017-11-27 Tom Tromey <tom@tromey.com>
* Makefile.in (SUBDIR_CLI_OBS): Redefine.
(%.o): Remove cli rule.
(CONFIG_SRC_SUBDIR): Add cli.
This implements a simpler way to make the "arch" build directory --
namely, now it is done as an order-only dependency in the Makefile,
rather than being created when config.status is run. This simpler
because it means that the build directories can be changed without
re-running autoconf.
ChangeLog
2017-11-27 Tom Tromey <tom@tromey.com>
* configure.ac (CONFIG_SRC_SUBDIR): Don't subst.
* configure: Rebuild.
* Makefile.in (CONFIG_SRC_SUBDIR): Redefine.
(CONFIG_DEP_SUBDIR): New variable.
(%.o): Add order-only dependency.
($(CONFIG_DEP_SUBDIR)): New target.
This patch updates the `find` command help and docs description to show
how to search for not null terminated strings when current language's
strings includes it.
gdb/ChangeLog:
PR gdb/21945
* findcmd.c (_initialize_mem_search): Update find command help
text.
gdb/doc/ChangeLog:
PR gdb/21945
* gdb.texinfo (Search Memory): Update description and example
about how to search a string without NULL terminator.
While playing with valgrind, I noticed that with Python 3, the progname
variable in do_start_initialization is not being freed (concat returns a
malloc'ed string). This patch uses unique_xmalloc_ptr to manage it.
With Python 2, we pass progname it directly to Py_SetProgramName, so it
should not be freed. We therefore release it before passing it.
gdb/ChangeLog:
* python/python.c (do_start_initialization): Change progname
type to gdb::unique_xmalloc_ptr. Release the pointer when using
Python 2.
This changes maybe_disable_address_space_randomization to be an RAII
class, rather than having it return a cleanup.
Regression tested by the buildbot.
ChangeLog
2017-11-26 Tom Tromey <tom@tromey.com>
* nat/linux-personality.h (class
maybe_disable_address_space_randomization): New class.
(maybe_disable_address_space_randomization): Don't declare
function.
* nat/linux-personality.c (restore_personality)
(make_disable_asr_cleanup): Remove.
(maybe_disable_address_space_randomization): Now a constructor.
(~maybe_disable_address_space_randomization): New destructor.
* linux-nat.c (linux_nat_create_inferior): Update.
gdbserver/ChangeLog
2017-11-26 Tom Tromey <tom@tromey.com>
* linux-low.c (linux_create_inferior): Update.
This removes a cleanup from gcore.c, replacing it with
unique_xmalloc_ptr.
Regression tested by the buildbot.
ChangeLog
2017-11-26 Tom Tromey <tom@tromey.com>
* gcore.c (write_gcore_file_1): Use gdb::unique_xmalloc_ptr.
Switching spu_software_single_step to use a regcache instead of a frame:
https://sourceware.org/ml/gdb-patches/2016-11/msg00355.html
cause a serious regression to SPU single-stepping.
There were two separate problems:
- SPU_LSLR_REGNUM is a pseudo register, so we must use the "cooked"
regcache methods instead of the "raw" ones to access it.
- When accessing a branch target register, we must only use the first
four bytes of the 16-byte vector register. This was done automatically
by the frame routines, but not by the regcache routines.
gdb/ChangeLog:
2017-11-26 Ulrich Weigand <uweigand@de.ibm.com>
* spu-tdep.c (spu_software_single_step): Access SPU_LSLR_REGNUM as
"cooked" register. Access only first four bytes of branch target
registers.
The support for setting breakpoint in functions with ABI tags patch
will add a use of SYMBOL_HASH_NEXT in cp-support.c, which fails to
compile with:
src/gdb/cp-support.c:38:0:
src/gdb/cp-support.c: In function ‘unsigned int cp_search_name_hash(const char*)’:
src/gdb/../include/safe-ctype.h:148:20: error: ‘do_not_use_tolower_with_safe_ctype’ was not declared in this scope
#define tolower(c) do_not_use_tolower_with_safe_ctype
^
src/gdb/minsyms.h:174:18: note: in expansion of macro ‘tolower’
((hash) * 67 + tolower ((unsigned char) (c)) - 113)
^
src/gdb/cp-support.c:1677:14: note: in expansion of macro ‘SYMBOL_HASH_NEXT’
hash = SYMBOL_HASH_NEXT (hash, *string);
^
This fixes the problem before it happens.
I was somewhat worried about whether this might have an impact with
languages that are case insensitive, but I convinced myself that it
doesn't. As bonus, this improves GDB's minsym interning performance a
bit (3%-10%). See
<https://sourceware.org/ml/gdb/2017-11/msg00021.html>.
gdb/ChangeLog:
2017-11-25 Pedro Alves <palves@redhat.com>
* dictionary.c: Include "safe-ctype.h".
* minsyms.c: Include "safe-ctype.h".
* minsyms.c (SYMBOL_HASH_NEXT): Use TOLOWER instead of tolower.
Earlier while working on the big completer rework series, I managed to
break
(gdb) [TAB]
locally, and make GDB crash, but only notice a few weeks down the
road, because we have no test for that...
I also noticed that:
(gdb) [TAB]
didn't work (didn't show all commands as matches), even though
entering a command with leading whitespace works:
(gdb) help
This commit fixes the latter and adds a testcase that covers both
issues.
The gdb.base/completion.exp change is necessary because the new test
has a file name that also starts with "gdb.base/complet", making that
particular test ambiguous. Adding another letter disambiguates.
gdb/ChangeLog:
2017-11-25 Pedro Alves <palves@redhat.com>
* completer.c (complete_line_internal_1): Skip spaces until the
start of the command.
gdb/testsuite/ChangeLog:
2017-11-25 Pedro Alves <palves@redhat.com>
* gdb.base/complete-empty.exp: New file.
* gdb.base/completion.exp: Adjust.
currently "b func tion" manages to set a breakpoint at "function" !
All these years I had never noticed this, but now that the linespec
completer actually works, this easily happens by accident, with:
"b func t<tab>"
expecting to get "thread", but getting instead:
"b func tion"
...
Also, this:
"b rettypefunc<int>"
manages to set a breakpoint on "rettype func<int>()".
These things happen due to strcmp_iw "magic".
Fix it by teaching strcmp_iw about when can it skip whitespace. This
required handling user-defined operators, and scope operators,
complicating the code a bit, unfortunately. I added unit tests for
all the corner cases I stumbled on, as I was developing this, and then
in the end wrote a testsuite testcase covering many of the same things
and more (to be added later).
gdb/ChangeLog:
2017-11-24 Pedro Alves <palves@redhat.com>
* cp-support.c (cp_symbol_name_matches_1): New, factored out from
cp_fq_symbol_name_matches. Pass language_cplus to
strncmp_with_mode.
(cp_fq_symbol_name_matches): Call cp_symbol_name_matches_1.
(selftests::test_cp_symbol_name_cmp): New.
(_initialize_cp_support): Register "cp_symbol_name_matches"
selftests.
* language.c (default_symbol_name_matcher): Pass language_minimal
to strncmp_iw_with_mode.
* utils.c: Include "cp-support.h" and <algorithm>.
(valid_identifier_name_char, cp_skip_operator_token, skip_ws)
(cp_is_operator): New functions.
(strncmp_iw_with_mode): Use them. Add language parameter. Don't
skip whitespace in the symbol name when the lookup name doesn't
have spaces, and vice versa.
(strncmp_iw, strcmp_iw): Pass language to strncmp_iw_with_mode.
* utils.h (strncmp_iw_with_mode): Add language parameter.
This patch enhances the debugger to print the exception message, when
available, as part of an exception catchpoint hit notification (both
GDB/CLI and GDB/MI). For instance, with the following code...
procedure A is
begin
raise Constraint_Error with "hello world";
end A;
... instead of printing...
Catchpoint 1, CONSTRAINT_ERROR at 0x000000000040245c in a () at a.adb:3
... it now prints:
Catchpoint 1, CONSTRAINT_ERROR (hello world) at 0x000000000040245c in a ()
^^^^^^^^^^^^^
This enhancement requires runtime support. If not present, the debugger
just behaves as before.
In GDB/MI mode, if the exception message is available, it is provided
as an extra field named "exception-message" in the catchpoint notification:
*stopped,bkptno="1",[...],exception-name="CONSTRAINT_ERROR",
exception-message="hello world",[...]
gdb/ChangeLog:
* ada-lang.c (ada_exception_message_1, ada_exception_message):
New functions.
(print_it_exception): If available, display the exception
message as well.
* NEWS: Document new feature.
gdb/doc/ChangeLog:
* gdb.texinfo (GDB/MI Ada Exception Information): Document
new "exception-message" field.
gdb/testsuite/ChangeLog:
* gdb.ada/catch_ex.exp, gdb.ada/mi_catch_ex.exp,
gdb.ada/mi_ex_cond.exp: Accept optional exception message in
when hitting an exception catchpoint.
Building GDB with GCC 6.2.1 gives multiple errors like
gdb/dtrace-probe.c: In member function ‘void dtrace_probe::build_arg_exprs(gdbarch*)’:
gdb/dtrace-probe.c:627:8: error: types may not be defined in a for-range-declaration [-Werror]
for (struct dtrace_probe_arg &arg : m_args
Fix it by removing the 'struct' keyword.
A similar Bug was already fixed for GCC 6.3.1
https://sourceware.org/ml/gdb-patches/2017-10/msg00442.html
gdb/ChangeLog:
* dtrace-probe.c (dtrace_probe::build_arg_exprs)
(dtrace_probe::is_enabled, dtrace_probe::enable)
(dtrace_probe::disable): Remove keyword 'struct' at for-range
variable
* probe.c (gen_ui_out_table_header_info)
(print_ui_out_not_applicables): Remove keyword 'struct' at
for-range variable
This patch (finally!) makes it so that trying to use XNEW with a type
that requires "new" will cause a compilation error. The criterion I
initially used to allow a type to use XNEW (which calls malloc in the
end) was std::is_trivially_constructible, but then realized that gcc 4.8
did not have it. Instead, I went with:
using IsMallocatable = std::is_pod<T>;
which is just a bit more strict, which doesn't hurt. A similar thing is
done for macros that free instead of allocated, the criterion is:
using IsFreeable = gdb::Or<std::is_trivially_destructible<T>, std::is_void<T>>;
Trying to use XNEW on a type that requires new will result in an error
like this:
In file included from /home/simark/src/binutils-gdb/gdb/common/common-utils.h:26:0,
from /home/simark/src/binutils-gdb/gdb/common/common-defs.h:78,
from /home/simark/src/binutils-gdb/gdb/defs.h:28,
from /home/simark/src/binutils-gdb/gdb/lala.c:1:
/home/simark/src/binutils-gdb/gdb/common/poison.h: In instantiation of ‘T* xnew() [with T = bar]’:
/home/simark/src/binutils-gdb/gdb/lala.c:13:3: required from here
/home/simark/src/binutils-gdb/gdb/common/poison.h:103:3: error: static assertion failed: Trying to use XNEW with a non-POD data type. Use operator new instead.
static_assert (IsMallocatable<T>::value, "Trying to use XNEW with a non-POD\
^~~~~~~~~~~~~
Generated-code-wise, it adds one more function call (xnew<T>) when using
XNEW and building with -O0, but it all goes away with optimizations
enabled.
gdb/ChangeLog:
* common/common-utils.h: Include poison.h.
(xfree): Remove declaration, add definition with static_assert.
* common/common-utils.c (xfree): Remove.
* common/poison.h (IsMallocatable): Define.
(IsFreeable): Define.
(free): Delete for non-freeable types.
(xnew): New.
(XNEW): Undef and redefine.
(xcnew): New.
(XCNEW): Undef and redefine.
(xdelete): New.
(XDELETE): Undef and redefine.
(xnewvec): New.
(XNEWVEC): Undef and redefine.
(xcnewvec): New.
(XCNEWVEC): Undef and redefine.
(xresizevec): New.
(XRESIZEVEC): Undef and redefine.
(xdeletevec): New.
(XDELETEVEC): Undef and redefine.
(xnewvar): New.
(XNEWVAR): Undef and redefine.
(xcnewvar): New.
(XCNEWVAR): Undef and redefine.
(xresizevar): New.
(XRESIZEVAR): Undef and redefine.
There are multiple definitions of the private_thread_info structure
compiled in the same GDB build. Because of the one definition rule, we
need to change this if we want to be able to make them non-POD (e.g. use
std::vector fields). This patch creates a class hierarchy, with
private_thread_info being an abstract base class, and all the specific
implementations inheriting from it.
In order to poison XNEW/xfree for non-POD types, it is also needed to
get rid of the xfree in thread_info::~thread_info, which operates on an
opaque type. This is replaced by thread_info::priv now being a
unique_ptr, which calls the destructor of the private_thread_info
subclass when the thread is being destroyed.
Including gdbthread.h from darwin-nat.h gave these errors:
/Users/simark/src/binutils-gdb/gdb/gdbthread.h:609:3: error: must use 'class' tag to refer to type 'thread_info' in this scope
thread_info *m_thread;
^
class
/usr/include/mach/thread_act.h:240:15: note: class 'thread_info' is hidden by a non-type declaration of 'thread_info' here
kern_return_t thread_info
^
It turns out that there is a thread_info function in the Darwin/XNU/mach API:
http://web.mit.edu/darwin/src/modules/xnu/osfmk/man/thread_info.html
Therefore, I had to add the class keyword at a couple of places in gdbthread.h,
I don't really see a way around it.
gdb/ChangeLog:
* gdbthread.h (private_thread_info): Define structure type, add
virtual pure destructor.
(thread_info) <priv>: Change type to unique_ptr.
<private_dtor>: Remove.
* thread.c (add_thread_with_info): Adjust to use of unique_ptr.
(private_thread_info::~private_thread_info): Provide default
implementation.
(thread_info::~thread_info): Don't call private_dtor nor
manually free priv.
* aix-thread.c (private_thread_info): Rename to ...
(aix_thread_info): ... this.
(get_aix_thread_info): New.
(sync_threadlists): Adjust.
(iter_tid): Adjust.
(aix_thread_resume): Adjust.
(aix_thread_fetch_registers): Adjust.
(aix_thread_store_registers): Adjust.
(aix_thread_extra_thread_info): Adjust.
* darwin-nat.h (private_thread_info): Rename to ...
(darwin_thread_info): ... this.
(get_darwin_thread_info): New.
* darwin-nat.c (darwin_init_thread_list): Adjust.
(darwin_check_new_threads): Adjust.
(thread_info_from_private_thread_info): Adjust.
* linux-thread-db.c (private_thread_info): Rename to ...
(thread_db_thread_info): ... this, initialize fields.
(get_thread_db_thread_info): New.
<dying>: Change type to bool.
(update_thread_state): Adjust to type rename.
(record_thread): Adjust to type rename an use of unique_ptr.
(thread_db_pid_to_str): Likewise.
(thread_db_extra_thread_info): Likewise.
(thread_db_thread_handle_to_thread_info): Likewise.
(thread_db_get_thread_local_address): Likewise.
* nto-tdep.h (private_thread_info): Rename to ...
(nto_thread_info): ... this, initialize fields.
(get_nto_thread_info): New.
<name>: Change type to std::string.
* nto-tdep.c (nto_extra_thread_info): Adjust to type rename and
use of unique_ptr.
* nto-procfs.c (update_thread_private_data_name): Adjust to
std::string change, allocate nto_private_thread_info with new.
(update_thread_private_data): Adjust to unique_ptr.
* remote.c (private_thread_info): Rename to ...
(remote_thread_info): ... this, initialize data members with
default values.
<extra, name>: Change type to std::string.
<thread_handle>: Change type to non-pointer.
(free_private_thread_info): Remove.
(get_private_info_thread): Rename to...
(get_remote_thread_info): ... this, change return type, adjust to
use of unique_ptr, use remote_thread_info constructor.
(remote_add_thread): Adjust.
(get_private_info_ptid): Rename to...
(get_remote_thread_info): ...this, change return type.
(remote_thread_name): Use get_remote_thread_info, adjust to
change to std::string.
(struct thread_item) <~thread_item>: Remove.
<thread_handle>: Make non pointer.
(start_thread): Adjust to thread_item::thread_handle type
change.
(remote_update_thread_list): Adjust to type name change, move
strings from temporary to long-lived object instead of
duplicating.
(remote_threads_extra_info): Use get_remote_thread_info.
(process_initial_stop_replies): Likewise.
(resume_clear_thread_private_info): Likewise.
(remote_resume): Adjust to type name change.
(remote_commit_resume): Use get_remote_thread_info.
(process_stop_reply): Adjust to type name change.
(remote_stopped_by_sw_breakpoint): Use get_remote_thread_info.
(remote_stopped_by_hw_breakpoint): Likewise.
(remote_stopped_by_watchpoint): Likewise.
(remote_stopped_data_address): Likewise.
(remote_core_of_thread): Likewise.
(remote_thread_handle_to_thread_info): Use
get_private_info_thread, adjust to thread_handle field type
change.
This patch C++ifies the thread_item and threads_listing_context
structures in remote.c. thread_item::{extra,name} are changed to
std::string. As a result, there's a bit of awkwardness in
remote_update_thread_list, where we have to xstrdup those strings when
filling the private_thread_info structure. This is removed in the
following patch, where private_thread_info is also C++ified and its
corresponding fields made std::string too. The xstrdup then becomes an
std::move.
Other than that there's nothing really special, it's a usual day-to-day
VEC -> vector and char* -> std::string change. It allows removing a
cleanup in remote_update_thread_list.
Note that an overload of hex2bin that returns a gdb::byte_vector is
added, with corresponding selftests.
gdb/ChangeLog:
* remote.c (struct thread_item): Add constructor, disable copy
construction and copy assignment, define default move
construction and move assignment.
<extra, name>: Change type to std::string.
<core>: Initialize.
<thread_handle>: Make non-pointer.
(thread_item_t): Remove typedef.
(DEF_VEC_O(thread_item_t)): Remove.
(threads_listing_context) <contains_thread>: New method.
<remove_thread>: New method.
<items>: Change type to std::vector.
(clear_threads_listing_context): Remove.
(threads_listing_context_remove): Remove.
(remote_newthread_step): Use thread_item constructor, adjust to
change to std::vector.
(start_thread): Use thread_item constructor, adjust to change to
std::vector.
(end_thread): Adjust to change to std::vector and std::string.
(remote_get_threads_with_qthreadinfo): Use thread_item
constructor, adjust to std::vector.
(remote_update_thread_list): Adjust to change to std::vector and
std::string, use threads_listing_context methods.
(remove_child_of_pending_fork): Adjust.
(remove_new_fork_children): Adjust.
* Makefile.in (SUBDIR_UNITTESTS_SRCS): Add rsp-low-selftests.c.
(SUBDIR_UNITTESTS_OBS): Add rsp-low-selftests.o.
* unittests/rsp-low-selftests.c: New file.
* common/rsp-low.h: Include common/byte-vector.h.
(hex2bin): New overload.
* common/rsp-low.c (hex2bin): New overload.
There are currently multiple definitions of private_inferior, defined in
remote.c and darwin-nat.h. The patch that poisons XNEW and friends for
non-POD types trips on that, because private_inferior is freed in
~inferior(), where it is an opaque type. Since the compiler can't tell
whether the type is POD, it gives an error. Also, we can't start using
C++ features in these structures (make them non-POD) as long as there
are multiple definitions with the same name. For these reasons, this
patch makes a class hierarchy, with private_inferior being the abstract
base class, and darwin_inferior & remote_inferior inheriting from it.
Destruction is done through the virtual destructor.
I stumbled on some suspicious code in the darwin implementation though.
darwin_check_new_threads does an XCNEW(darwin_thread_t) when it finds a
new thread, allocating a new structure for it (darwin_thread_t is a
typedef for private_thread_info). It then VEC_safe_pushes it in a
vector defined as DEF_VEC_O (a vector of objects). This means that the
structure content gets copied in the vector. The thread_info object is
created with the XCNEW'ed structure as the private thread info, while
the rest of the code works with the instance in the vector. We have
therefore two distinct instances of darwin_thread_t/private_thread_info
for each thread. This is not really a problem in practice, because
thread_info::priv is not used in the darwin code. I still find it weird
and far from ideal, so I tried to fix it by changing the vector to be a
vector of pointers. There should now be a single instance of the
structure for each thread. The deallocation of the
darwin_thread_t/private_thread_info structure is done by the thread_info
destructor.
I am able to build on macOS, but not really test, since the port seems a
bit broken. I am not able to debug reliably on the machine I have
access to, which runs macOS 10.12.6.
gdb/ChangeLog:
* inferior.h (private_inferior): Define structure type, add
virtual pure destructor.
(inferior) <priv>: Change type to unique_ptr.
* inferior.c (private_inferior::~private_inferior): Provide
default implementation.
(inferior::~inferior): Don't free priv field.
(exit_inferior_1): Likewise.
* darwin-nat.h (struct darwin_exception_info): Initialize fields.
(darwin_exception_info): Remove typedef.
(DEF_VEC_O (darwin_thread_t)); Remove.
(private_inferior): Rename to ...
(darwin_private_inferior): ... this, extend private_inferior.
(get_darwin_inferior): New.
<threads>: Change type to std::vector of darwin_thread_t pointers.
* darwin-nat.c (darwin_check_new_threads): Adjust.
(find_inferior_task_it): Adjust.
(darwin_find_thread); Adjust.
(darwin_suspend_inferior): Adjust.
(darwin_resume_inferior): Adjust.
(darwin_find_new_inferior): Adjust.
(darwin_decode_notify_message): Adjust.
(darwin_send_reply): Adjust.
(darwin_resume_inferior_threads): Adjust.
(darwin_suspend_inferior_threads): Adjust.
(darwin_decode_message): Adjust.
(darwin_wait): Adjust.
(darwin_interrupt): Adjust.
(darwin_deallocate_threads): Adjust.
(darwin_mourn_inferior): Adjust, don't free private data.
(darwin_reply_to_all_pending_messages): Adjust.
(darwin_stop_inferior): Adjust.
(darwin_setup_exceptions): Adjust.
(darwin_kill_inferior): Adjust.
(darwin_setup_request_notification): Adjust.
(darwin_attach_pid): Adjust.
(darwin_init_thread_list): Adjust.
(darwin_setup_fake_stop_event): Adjust.
(darwin_attach): Adjust.
(darwin_detach): Adjust.
(darwin_xfer_partial): Adjust.
(set_enable_mach_exceptions): Adjust.
(darwin_pid_to_exec_file): Adjust.
(darwin_get_ada_task_ptid): Adjust.
* darwin-nat-info.c (get_task_from_args): Adjust.
(info_mach_ports_command): Adjust.
(info_mach_region_command): Adjust.
(info_mach_exceptions_command): Adjust.
* remote.c (private_inferior): Rename to ...
(remote_private_inferior): ... this, initialize fields.
(get_remote_inferior); New.
(remote_commit_resume): Use get_remote_inferior.
(check_pending_event_prevents_wildcard_vcont_callback): Likewise.
