| Age | Commit message (Collapse) | Author | Files | Lines |
|---|
|
When debugging testsuite failures, it can be awkward parsing gdb.log to
obtain all the commands run in order to manually re-run the test.
This patch adds the functionality to save all gdb commands to the file gdb.in
when the testsuite is run. The file is saved in the directory for the test and
if gdb is restarted then .1, .2, .3 etc is added to the filename.
Once a test has been run, the .in file can be used to re-run the test in the
following way:
gdb -x outputs/gdb.store/gdb.in outputs/gdb.store/store
The code works by intercepting send_gdb. I've added a TYPE to ensure that any
commands that would destroy the playback are kept from the log (for example the
Y from an answer to a y/n question).
Adds library function standard_output_file_with_gdb_instance to open a file
postfixed with count of the gdb instance. Ensure this count is reset when a new
.exp script is run.
I've re-run a random selection of .in files to check they do not error. Logs with
commands such as "attach <pid>" will not directly work when re-run.
gdb/testsuite/ChangeLog:
* lib/gdb.exp (gdb_unload): Mark Y as an answer.
(delete_breakpoints): Likewise.
(gdb_run_cmd): Likewise.
(gdb_start_cmd): Likewise.
(gdb_starti_cmd): Likewise.
(gdb_internal_error_resync): Likewise.
(gdb_test_multiple): Likewise.
(gdb_reinitialize_dir): Likewise.
(default_gdb_exit): Likewise.
(gdb_file_cmd): Mark kill as optional.
(default_gdb_start): Call gdb_stdin_log_init.
(send_gdb): Call gdb_stdin_log_write.
(rerun_to_main): Mark Y as an answer.
(gdb_stdin_log_init): New function.
(gdb_stdin_log_write): Likewise.
|
|
Fix up all failures encountered when running the testsuite with
GDB_DEBUG="infrun".
Some tests rely on enabling debugging for various components. With
debugging on, this will be lost to the debug file.
Disable separate tty for mi tests when debugging. This currently
does not work.
disasm.c should send errors to the stderr instead of the logfile.
Note that enabling debug for other components might still cause
additional errors above what has been fixed here.
gdb/ChangeLog:
* disasm.c (set_disassembler_options): Send errors to stderr.
gdb/testsuite/ChangeLog:
* gdb.base/breakpoint-in-ro-region.exp: Disable when debugging.
* gdb.base/debug-expr.exp: Likewise.
* gdb.base/foll-fork.exp: Likewise.
* gdb.base/foll-vfork.exp: Likewise.
* gdb.base/fork-print-inferior-events.exp: Likewise.
* gdb.base/gdb-sigterm.exp: Likewise.
* gdb.base/gdbinit-history.exp: Likewise.
* gdb.base/osabi.exp: Likewise.
* gdb.base/sss-bp-on-user-bp-2.exp: Likewise.
* gdb.base/ui-redirect.exp: Likewise.
* gdb.gdb/unittest.exp: Likewise.
* gdb.mi/mi-break.exp: Disable separate-mi-tty when debugging.
* gdb.mi/mi-watch.exp: Likewise.
* gdb.mi/new-ui-mi-sync.exp: Likewise.
* gdb.mi/user-selected-context-sync.exp: Likewise.
* gdb.python/python.exp: Disable debug test when debugging.
* gdb.threads/check-libthread-db.exp: Disable when debugging.
* gdb.threads/signal-while-stepping-over-bp-other-thread.exp:
Likewise.
* gdb.threads/stepi-random-signal.exp: Likewise.
|
|
Add both board option and environment variable which enables gdb
debug via a comma separated list and sends it to the file gdb.debug,
located in the output directory for the current test. Document this.
Add support for the environment variable in the Makefile.
The testsuite can be run with gdb debug enabled in the following way:
make check GDB_DEBUG="infrun,target,remote"
A Test with multiple invocations of GDB will all append debug to the
same log file.
gdb/testsuite/ChangeLog:
* Makefile.in: Pass through GDB_DEBUG.
* README (Testsuite Parameters): Add GDB_DEBUG.
(gdb,debug): Add board setting.
* lib/gdb.exp (default_gdb_start): Start debugging.
(gdb_debug_enabled): New procedure.
(gdb_debug_init): Likewise.
|
|
Currently, when logging is enabled, output will be sent to both a
logfile and standard terminal output. The redirect option sends output
only to the logfile. This includes all debug output.
Add the option to redirect debug output seperately to normal
output, using the cli command:
set logging debugredirect on
By setting this and enabling logging, all output and debug will
be sent to the logfile. The user will still see all output but
no debug output.
This causes a change in behaviour for anyone currently using
logging redirect, as now only output will be redirected. Users
will have to issue the additional command above to also redirect
debug.
Expand ui-redirect.exp cover the changes.
gdb/ChangeLog:
* cli/cli-interp.c (struct saved_output_files): Add saved entry.
(cli_interp_base::set_logging): Check debug_redirect.
* cli/cli-interp.h (set_logging): Add debug_redirect parameter.
* cli/cli-logging.c (debug_redirect): Add static variable.
(pop_output_files): Add default param.
(handle_redirections): Print debug setting.
(show_logging_command): Likewise.
(_initialize_cli_logging): Add debugredirect command.
* interps.c (current_interp_set_logging): Add debug_redirect
parameter.
* interps.h (set_logging): Add debug_redirect parameter.
(current_interp_set_logging): Likewise.
* mi/mi-common.h: Likewise.
* mi/mi-interp.c (mi_interp::set_logging): Likewise.
gdb/testsuite/ChangeLog:
* gdb.base/ui-redirect.exp: Add debug redirect tests.
|
|
Instead of using two bools to decide if the files should close when tee_file
is closed, make file one stay open and file two close. This simplifies the
use cases for it.
Inline the make_logging_output into the calling functions (the logic here
looks ugly in order to simplify a later change).
Expand ui-redirect.exp to cover the changes, similar to mi-logging.exp.
gdb/ChangeLog:
* cli/cli-interp.c (cli_interp_base::set_logging): Create tee_file
directly.
* cli/cli-interp.h (make_logging_output): Remove declaration.
* cli/cli-logging.c (make_logging_output): Remove function.
* mi/mi-interp.c (mi_interp::set_logging): Create tee_file
directly.
* ui-file.c (tee_file::tee_file): Remove bools.
(tee_file::~tee_file): Remove deletes.
* ui-file.h (tee_file): Remove bools.
gdb/testsuite/ChangeLog:
* gdb.base/ui-redirect.exp: Test redirection.
|
|
There is a CLI command 'complete' intended to use with emacs. Such a command
would also be useful for MI frontends, when separate CLI and MI channels cannot
be used. For example, on Windows (because of lack of PTYs) or when GDB is used
through SSH session.
This commit adds a new '-complete' MI command.
gdb/Changelog:
2019-01-28 Jan Vrany <jan.vrany@fit.cvut.cz>
* mi/mi-cmds.h (mi_cmd_complete): New function.
* mi/mi-main.c (mi_cmd_complete): Likewise.
* mi/mi-cmds.c: Define new MI command -complete.
* NEWS: Mention new -complete command.
gdb/doc/ChangeLog:
2019-01-28 Jan Vrany <jan.vrany@fit.cvut.cz>
* gdb.texinfo (Miscellaneous GDB/MI Commands): Document new
MI command -complete.
gdb/testsuite/ChangeLog:
2019-01-28 Jan Vrany <jan.vrany@fit.cvut.cz>
* gdb.mi/mi-complete.exp: New file.
* gdb.mi/mi-complete.cc: Likewise.
|
|
Add tests for calling sizeof on indexed and sliced arrays, and on
pointers to arrays. These are all things that currently work, but
were previously untested.
gdb/testsuite/ChangeLog:
* gdb.fortran/vla-sizeof.exp: Add tests of sizeof applied to
indexed and sliced arrays, and pointers to arrays.
|
|
This changes "info sharedlibrary" to add styling to the file name.
Tested on x86-64 Fedora 29.
gdb/ChangeLog
2019-05-14 Tom Tromey <tromey@adacore.com>
* solib.c (info_sharedlibrary_command): Style the file name.
gdb/testsuite/ChangeLog
2019-05-14 Tom Tromey <tromey@adacore.com>
* gdb.base/info-shared.exp (check_info_shared): Add "info shared"
styling test.
|
|
A base address selection entry in a location list consist of two (constant or
relocated) address offsets. The two offsets are the same size as an address
on the target machine.
The test-case gdb.dwarf2/dw2-skip-prologue.S encodes a base address selection
entry using .4byte, which is incorrect for 8-byte pointer size. [ Which
triggers an assert in dwz, see PR dwz/24172. ]
Fix this by using PTRBYTE instead.
Tested on x86_64-linux.
gdb/testsuite/ChangeLog:
2019-05-14 Tom de Vries <tdevries@suse.de>
* gdb.dwarf2/dw2-skip-prologue.S (.debug_loc): Fix base address
selection entry encoding.
|
|
This patch adds a completion function to the "catch exception"
and "catch handlers" commands.
Tested on x86-64 Fedora 29; reviewed off-list by Joel.
gdb/ChangeLog
2019-05-10 Tom Tromey <tromey@adacore.com>
* ada-lang.c (catch_ada_completer): New function.
(_initialize_ada_language): Use it.
gdb/testsuite/ChangeLog
2019-05-10 Tom Tromey <tromey@adacore.com>
* gdb.ada/info_exc.exp: Add "complete" test.
|
|
The test-case gdb.arch/amd64-tailcall-self.exp fails here:
...
if ![runto b] {
return -1
}
...
like:
...
(gdb) file build/gdb/testsuite/outputs/gdb.arch/amd64-tailcall-self/\
amd64-tailcall-self
Reading symbols from build/gdb/testsuite/outputs/gdb.arch/\
amd64-tailcall-self/amd64-tailcall-self...
Dwarf Error: Cannot find DIE at 0x1f5 referenced from DIE at 0x107 [in \
module build/gdb/testsuite/outputs/gdb.arch/amd64-tailcall-self/\
amd64-tailcall-self]
...
The problem is that in amd64-tailcall-self.S, CU-relative references are
assigned .debug_info section relative values. [ This is similar to the
problem fixed by "Fix gdb.arch/amd64-entry-value-paramref.S". ]
Fix this by assigning CU-relative references instead.
Tested on x86_64-linux.
gdb/testsuite/ChangeLog:
2019-05-09 Tom de Vries <tdevries@suse.de>
* gdb.arch/amd64-tailcall-self.S: Make DW_FORM_ref4 references
CU-relative.
|
|
The file gdb.arch/amd64-entry-value-paramref.S contains a DIE for function
bar:
...
DIE29: .uleb128 0x2 # (DIE (0x29) DW_TAG_subprogram)
.ascii "bar\0" # DW_AT_name
.byte 0x1 # DW_AT_decl_file (gdb.arch/amd64-entry-value-paramref.cc)
.byte 0x15 # DW_AT_decl_line
.long DIE45 # DW_AT_type
.byte 0x1 # DW_AT_inline
...
which refers to DIE45:
...
DIE45: .uleb128 0x4 # (DIE (0x45) DW_TAG_base_type)
.byte 0x4 # DW_AT_byte_size
.byte 0x5 # DW_AT_encoding
.ascii "int\0" # DW_AT_name
...
using a form DW_FORM_ref4:
...
.uleb128 0x2 # (abbrev code)
.uleb128 0x2e # (TAG: DW_TAG_subprogram)
.byte 0x1 # DW_children_yes
...
.uleb128 0x49 # (DW_AT_type)
.uleb128 0x13 # (DW_FORM_ref4)
...
However, the DW_FORM_ref4 is a CU-relative reference, while using a label for
the value will result in a section-relative value.
So, if linked in object files contain dwarf info and are placed in the
.debug_info section before the compilation units generated from
amd64-entry-value-paramref.S, then the referenced type is at 0x108:
...
<1><108>: Abbrev Number: 4 (DW_TAG_base_type)
<109> DW_AT_byte_size : 4
<10a> DW_AT_encoding : 5 (signed)
<10b> DW_AT_name : int
...
but the reference will point to a non-existing DIE at 0x1cf:
...
<1><f0>: Abbrev Number: 2 (DW_TAG_subprogram)
<f1> DW_AT_name : bar
<f5> DW_AT_decl_file : 1
<f6> DW_AT_decl_line : 21
<f7> DW_AT_type : <0x1cf>
<fb> DW_AT_inline : 1 (inlined)
...
which happens to cause a GDB internal error described in PR23270 - "GDB
internal error: dwarf2read.c:18656: internal-error: could not find partial
DIE 0x1b7 in cache".
Fix the invalid DWARF by making the reference value CU-relative:
...
- .long DIE45 # DW_AT_type
+ .long DIE45 - .Ldebug_info0 # DW_AT_type
...
Tested on x86_64-linux.
gdb/testsuite/ChangeLog:
2019-05-09 Tom de Vries <tdevries@suse.de>
* gdb.arch/amd64-entry-value-paramref.S: Make DW_FORM_ref4 references
CU-relative.
|
|
- set print frame-arguements all
- an Ada function named pck.call_me calls a C function named break_me
- you put a breakpoint in break_me and the program reaches this
breakpoint.
Now display the backtrace:
(gdb) bt
#0 break_me () at [...]
#1 0x000000000040243e in pck.call_me (
s={P_ARRAY = 0x7fffffffe21c, P_BOUNDS = 0x41e6e8}) at [...]
whereas we should expect:
(gdb) bt
#0 break_me () at [...]
#1 0x000000000040243e in pck.call_me (s="test") at [...]
The problem is that GDB prints the S parameter in the pck.call_me Ada
function using the current language, so the C one, because the program
is stopped in a C function, whereas it should use the pck.call_me frame
one. This behavior is ok when user manually changes the language but it's
not the right one when language is auto.
This patch fixes this problem so now when using auto language, all Ada
frame arguments are printed using Ada like syntax when the frame is part
of Ada code, even if the program is stopped in a frame using a different
language.
If the user explicitly sets a language (using "set language ...") then
no change here, all the Ada frame arguments are printed using this
language.
gdb/ChangeLog:
* ada-valprint.c (ada_val_print_gnat_array): Remove language
parameter and use Ada language definition instead.
(ada_val_print_ptr): Remove unused language parameter.
(ada_val_print_num): Remove language parameter and use Ada language
definition instead.
(ada_val_print_enum, ada_val_print_flt): Remove unused language
parameter.
(ada_val_print_struct_union, ada_val_print_ref): Remove language
parameter and use Ada language definition instead.
(ada_val_print_1): Update all ada_val_print_xxx calls.
Remove language parameter.
(ada_val_print): Update ada_val_print_1 call.
gdb/testsuite/ChangeLog:
* gdb.ada/frame_arg_lang.exp: New testcase.
* gdb.ada/frame_arg_lang/bla.adb: New file.
* gdb.ada/frame_arg_lang/pck.ads: New file.
* gdb.ada/frame_arg_lang/pck.adb: New file.
* gdb.ada/frame_arg_lang/foo.c: New file.
Tested on x86_64-linux, no regressions.
|
|
A user here noticed that the Python Value.string method did not work
for Ada arrays. I tracked this down to an oddity in value_as_address
-- namely, it calls coerce_array, but that function will not force
array coercion when the language has c_style_arrays=false, as Ada
does.
This patch fixes the problem by changing c_get_string so that arrays
take the "in GDB's memory" branch. The actual patch is somewhat more
complicated than you might think, because the caller can request more
array elements than the type allows. This is normal when the type is
using the C struct hack.
Tested on x86-64 Fedora 29.
gdb/ChangeLog
2019-05-08 Tom Tromey <tromey@adacore.com>
* c-lang.c (c_get_string): Handle non-C-style arrays.
gdb/testsuite/ChangeLog
2019-05-08 Tom Tromey <tromey@adacore.com>
* gdb.python/py-value.exp (test_value_in_inferior): Add Ada test.
|
|
Consider this short C example:
struct inner
{
unsigned x;
unsigned y : 3;
unsigned z : 3;
};
struct outer
{
unsigned char o : 3;
struct inner i __attribute__ ((packed));
};
When I use "ptype/o" on this, I get:
(gdb) ptype/o struct outer
/* offset | size */ type = struct outer {
/* 0: 5 | 1 */ unsigned char o : 3;
/* XXX 5-bit hole */
/* 1 | 8 */ struct inner {
/* 1 | 4 */ unsigned int x;
/* 5:29 | 4 */ unsigned int y : 3;
/* 5:26 | 4 */ unsigned int z : 3;
/* XXX 2-bit padding */
/* XXX 3-byte padding */
/* total size (bytes): 8 */
} i;
/* total size (bytes): 9 */
}
In the location of "o" ("0: 5"), the "5" means "there are 5 bits left
relative to the size of the underlying type.
I find this very difficult to follow. On irc, Sergio said that this
choice came because it is what pahole does. However, I think it's not
very useful, and maybe is just an artifact of the way that
DW_AT_bit_offset was defined in DWARF 3.
This patch changes ptype/o to print the offset of a bitfield in a more
natural way, that is, using the bit number according to the platform's
bit numbering.
With this patch, the output is now:
(gdb) ptype/o struct outer
/* offset | size */ type = struct outer {
/* 0: 0 | 1 */ unsigned char o : 3;
/* XXX 5-bit hole */
/* 1 | 8 */ struct inner {
/* 1 | 4 */ unsigned int x;
/* 5: 0 | 4 */ unsigned int y : 3;
/* 5: 3 | 4 */ unsigned int z : 3;
/* XXX 2-bit padding */
/* XXX 3-byte padding */
/* total size (bytes): 8 */
} i;
/* total size (bytes): 9 */
}
This is better, IMO, because now the "offset" of a bitfield is
consistent with the offset of an ordinary member, referring to its
offset from the start of the structure.
gdb/ChangeLog
2019-05-08 Tom Tromey <tromey@adacore.com>
* typeprint.c (print_offset_data::update): Print the bit offset,
not the number of bits remaining.
gdb/doc/ChangeLog
2019-05-08 Tom Tromey <tromey@adacore.com>
* gdb.texinfo (Symbols): Document change to ptype/o.
gdb/testsuite/ChangeLog
2019-05-08 Tom Tromey <tromey@adacore.com>
* gdb.base/ptype-offsets.exp: Update tests.
|
|
I noticed that ptype/o will print:
/* 3: 3 | 1 */ signed char a4 : 2;
/* XXX 3-bit hole */
That is, "*/" at the end of the "hole" message does not line up with
the other comment ends. I thought it would be a bit nicer if this did
line up, so I fixed it. Then, to my surprise, I found that I could
not make ptype-offsets.exp fail.
I still am not sure why it doesn't fail, but changing the tests to use
string_to_regexp and changing the quoting helped. This in turn showed
that some of the existing test cases were wrong, so I've also updated
them here.
gdb/ChangeLog
2019-05-08 Tom Tromey <tromey@adacore.com>
* typeprint.c (print_offset_data::maybe_print_hole): Add extra
padding at end of comment.
gdb/testsuite/ChangeLog
2019-05-08 Tom Tromey <tromey@adacore.com>
* gdb.base/ptype-offsets.exp: Use string_to_regexp. Fix test
cases.
* gdb.base/ptype-offsets.cc (struct abc) <my_int_type>: Now
"short".
|
|
While looking at a different Ada problem, I found that printing a
record containing a VLA did not work properly.
I tracked the problem down to dwarf2_evaluate_property trying, and
failing, to compare two types that differed only in qualifiers.
This patch changes dwarf2_evaluate_property to ignore qualifiers when
comparing types.
Tested on x86-64 Fedora 29.
gdb/ChangeLog
2019-05-08 Tom Tromey <tromey@adacore.com>
* dwarf2loc.c (dwarf2_evaluate_property) <PROP_ADDR_OFFSET>:
Compare main types.
gdb/testsuite/ChangeLog
2019-05-08 Tom Tromey <tromey@adacore.com>
* gdb.ada/vla.exp: New file.
* gdb.ada/vla/vla.adb: New file.
|
|
When adding a debug print here in index-cache.exp:
...
proc_with_prefix test_cache_disabled { cache_dir } {
lassign [ls_host $cache_dir] ret files_before
+ puts "before: '$files_before'"
+ exit
...
we have:
...
files_before: ''
...
When further adding:
...
proc_with_prefix test_cache_disabled { cache_dir } {
+ exec touch $cache_dir/foo.1 $cache_dir/foo.2 $cache_dir/foo.3
...
we have:
...
files_before: 'foo.1'
...
while we're expecting file_before to contain foo.[123].
Fix this by making the return statement in ls_host return a list rather than a
string (in accordance with the ls_host documentation), after which we have:
...
files_before: 'foo.1 foo.2 foo.3'
...
Tested on x86_64-linux.
gdb/testsuite/ChangeLog:
2019-05-07 Tom de Vries <tdevries@suse.de>
* gdb.base/index-cache.exp (ls_host): Fix return statement.
|
|
When running gdb.dlang/watch-loc.exp with target board cc-with-debug-names, we
run into:
...
FAIL: gdb.dlang/watch-loc.exp: disassemble _Dmain (GDB internal error)
...
in more detail:
...
(gdb) disassemble _Dmain
gdb/dwarf2read.c:5293: internal-error: \
compunit_symtab* dw2_find_pc_sect_compunit_symtab(objfile*, \
bound_minimal_symbol, CORE_ADDR, obj_section*, int): \
Assertion `result != NULL' failed.
...
The problem is that the .debug_aranges section in watch-loc.c contains a
debug_info_offset which is set to 0:
...
asm (
" .pushsection .debug_aranges,\"\",%progbits\n"
" .4byte .Laranges_end - .Laranges_start\n"
".Laranges_start:\n"
" .2byte 0x2\n"
" .4byte 0\n"
...
while the compilation unit at offset 0 in the .debug_section in the executable
is in fact not the compilation unit generated from watch-loc-dw.S.
[ Note: this is a non-trivial test-case. The file watch-loc-dw.S contains a
.debug_info section, but not an .debug_aranges section or any actual code.
The file watch-loc.c contains code and a .debug_aranges section, but no other
debug section. So, the intent for the .debug_aranges section in watch-loc.c
is to refer to a compilation unit in the .debug_info section in
watch-loc-dw.S. ]
This happens when linked in object files contain dwarf info and are placed in
the .debug_info section before the compilation units generated from
watch-loc.c and watch-loc-dw.S.
Fix this by defining the debug_info_offset field using a label .Lcu1_begin
that defines the start of an empty .debug_section compilation unit:
...
asm (
+ " .pushsection .debug_info,\"\",%progbits\n"
+ ".Lcu1_begin:"
+ " .popsection\n"
" .pushsection .debug_aranges,\"\",%progbits\n"
" .4byte .Laranges_end - .Laranges_start \n"
".Laranges_start:\n"
" .2byte 0x2\n"
- " .4byte 0\n"
+ " .4byte .Lcu1_begin\n"
...
which during linking merges with the start of the .debug_info section of
watch-loc-dw.S.
Tested on x86_64-linux with native, cc-with-gdb-index and cc-with-debug-names.
gdb/testsuite/ChangeLog:
2019-05-07 Tom de Vries <tdevries@suse.de>
PR testsuite/24522
* gdb.dlang/watch-loc.c: Fix debug_info_offset in .debug_aranges
section.
|
|
When running gdb.dwarf2/dw2-case-insensitive.exp with target board
cc-with-debug-names, we run into:
...
FAIL: gdb.dwarf2/dw2-case-insensitive.exp: regexp case-sensitive off \
(GDB internal error)
...
in more detail:
...
(gdb) info functions fUnC_lang
gdb/dwarf2read.c:5293: internal-error: \
compunit_symtab* dw2_find_pc_sect_compunit_symtab(objfile*, \
bound_minimal_symbol, CORE_ADDR, obj_section*, int): \
Assertion `result != NULL' failed.
...
The problem is that the .debug_aranges section in dw2-case-insensitive-debug.S
contains a debug_info_offset which is set to 0:
...
.section .debug_aranges,"",@progbits
.4byte .Laranges_end - .Laranges_start
.Laranges_start:
.2byte 0x2
.4byte 0
...
while the compilation unit at offset 0 in the .debug_section of the executable
is in fact not the compilation unit generated from
dw2-case-insensitive-debug.S.
This happens when linked in object files contain dwarf info and are placed in
the .debug_info section before the compilation unit generated from
dw2-case-insensitive-debug.S.
Fix this by defining the debug_info_offset field using the label .Lcu1_begin
that defines the start of the compilation unit:
...
- .4byte 0
+ .4byte .Lcu1_begin
...
Tested on x86_64-linux with native, cc-with-gdb-index and cc-with-debug-names.
gdb/testsuite/ChangeLog:
2019-05-07 Tom de Vries <tdevries@suse.de>
PR testsuite/24522
* gdb.dwarf2/dw2-case-insensitive-debug.S: Fix debug_info_offset in
.debug_aranges section.
|
|
When running gdb.dwarf2/multidictionary.exp with target board cc-with-dwz and
current dwz, we run into a dwz abort:
...
gdb compile failed, gdb/contrib/cc-with-tweaks.sh: line 188: 11484 Aborted \
(core dumped) $DWZ "$output_file" > /dev/null 2>&1
UNTESTED: gdb.dwarf2/multidictionary.exp: multidictionary.exp
...
The dwz abort (PR dwz/24169) is caused by an invalid DW_FORM_ref_addr in the
multidictionary binary.
The multidictionary binary is build from multidictionary.S which is generated
using the dwarf assembler, and multidictionary.S contains dwarf for 3
compilation units.
In multidictionary0.o (generated from multidictionary.S), we find a concrete
formal parameter DIE:
...
<2><dc>: Abbrev Number: 4 (DW_TAG_formal_parameter)
<dd> DW_AT_abstract_origin: <0xa6>
...
referring to an abstract formal parameter DIE at 0xa6:
...
<2><a6>: Abbrev Number: 8 (DW_TAG_formal_parameter)
<a7> DW_AT_name : msg
<ab> DW_AT_type : <0x92>
...
but in the multidictionary binary the concrete formal parameter DIE is still
referring to 0xa6:
...
<2><1a3>: Abbrev Number: 4 (DW_TAG_formal_parameter)
<1a4> DW_AT_abstract_origin: <0xa6>
...
while the abstract formal parameter DIE has moved to 0x16d:
...
<2><16d>: Abbrev Number: 8 (DW_TAG_formal_parameter)
<16e> DW_AT_name : msg
<172> DW_AT_type : <0x159>
...
The concrete formal parameter DIE is specified in multidictionary.S like this:
...
.Llabel21:
.uleb128 4
.4byte .Llabel17 - .Lcu1_begin
...
The problem is that the .Lcu1_begin label is assumed to mark the start of the
.debug_info section in the executable, but in fact it marks the start of the
first compilation unit from multidictionary.S in the executable. Usually
these two entities are the same, but they are not when linked in object files
contain dwarf info and are placed in the .debug_info section before the
compilation units generated from multidictionary.S.
Fix this in the dwarf assembler by generating instead the label itself:
...
.Llabel21:
.uleb128 4
.4byte .Llabel17
...
resulting in a relocation in the object file:
...
Offset Info Type Sym. Value Sym. Name + Addend
0000000000dd 00040000000a R_X86_64_32 0000000000000000 .debug_info + a6
...
and resulting in the correct offset in the executable:
...
<2><1a3>: Abbrev Number: 4 (DW_TAG_formal_parameter)
<1a4> DW_AT_abstract_origin: <0x16d>
...
Tested on x86_64-linux with native and cc-with-dwz.
gdb/testsuite/ChangeLog:
2019-05-07 Tom de Vries <tdevries@suse.de>
PR testsuite/24159
* lib/dwarf.exp: Fix handling of DW_FORM_ref_addr.
|
|
In gdb.base/index-cache.exp, handle the case that binfile contains either a
.gdb_index or .debug_names index section.
Tested on x86_64-linux with native, cc-with-gdb-index and cc-with-debug-names.
gdb/testsuite/ChangeLog:
2019-05-06 Tom de Vries <tdevries@suse.de>
* lib/gdb.exp (exec_has_index_section): New proc.
* gdb.base/index-cache.exp: Handle case that binfile contains an index
section.
|
|
Add a target board that makes it easy to run the test suite with a
.debug_names section added to executables.
gdb/ChangeLog:
2019-05-04 Tom de Vries <tdevries@suse.de>
* contrib/cc-with-tweaks.sh: Support -n arg.
gdb/testsuite/ChangeLog:
2019-05-04 Tom de Vries <tdevries@suse.de>
* boards/cc-with-debug-names.exp: New file.
|
|
An internal bug report points out that, when a global character enum
type is used, casting fails, like:
(gdb) print global_char_enum'('F')
$1 = 70
The bug here turns out to be that enumerators are qualified, so for
example the mangled name might be "pck__QU48", rather than "QU48".
This patch fixes the problem by only examining the suffix of the
enumerator. This is ok because the type is already known, and because
the mangling scheme ensures that there won't be clashes.
Tested on x86-64 Fedora 29.
gdb/ChangeLog
2019-05-03 Tom Tromey <tromey@adacore.com>
* ada-exp.y (convert_char_literal): Check suffix of each
enumerator.
gdb/testsuite/ChangeLog
2019-05-03 Tom Tromey <tromey@adacore.com>
* gdb.ada/char_enum/pck.ads (Global_Enum_Type): New type.
* gdb.ada/char_enum/foo.adb: Use Global_Enum_Type.
* gdb.ada/char_enum.exp: Add test.
|
|
Add a target board cc-with-gdb-index.exp, to make it easy to run cc-with-tweaks
with CC_WITH_TWEAKS_FLAGS='-i'.
Tested on x86_64-linux.
gdb/testsuite/ChangeLog:
2019-05-03 Tom de Vries <tdevries@suse.de>
* boards/cc-with-gdb-index.exp: New file.
|
|
When printing a rust structure that contains a string GDB can
currently fail to read the fields that define the string. This is
because GDB mistakenly treats a value that is the parent structure as
though it is the structure that defines the string, and then fails to
find the fields needed to extract a string.
The solution is to create a new value to represent the string field of
the parent value.
gdb/ChangeLog:
* rust-lang.c (val_print_struct): Handle printing structures
containing strings.
gdb/testsuite/ChangeLog:
* gdb.rust/simple.exp: Add new test case.
* gdb.rust/simple.rs (struct StringAtOffset): New struct.
(main): Initialise an instance of the new struct.
|
|
A user at AdaCore found a case where assignment to a nested packed
structure would fail. The bug is that ada_value_primitive_field
doesn't account for the situation where a field is not packed relative
to its containing structure, but where the structure itself is packed
in its parent.
gdb/ChangeLog
2019-05-01 Tom Tromey <tromey@adacore.com>
* ada-lang.c (ada_value_primitive_field): Treat more fields as
bitfields.
gdb/testsuite/ChangeLog
2019-05-01 Tom Tromey <tromey@adacore.com>
* gdb.ada/packed_array_assign/aggregates.ads (Nested_Packed): New
record.
(NPR): New variable.
* gdb.ada/packed_array_assign.exp: Add nested packed assignment
test.
|
|
A bug internal to AdaCore notes that assigning a non-scalar value to
an element of a packed array will sometimes fail.
The bug turns out to be that ada_value_assign incorrectly computes the
starting point for the assignment. This patch fixes the problem.
gdb/ChangeLog
2019-05-01 Tom Tromey <tromey@adacore.com>
* ada-lang.c (ada_value_assign): Correctly compute starting offset
for big-endian copies.
gdb/testsuite/ChangeLog
2019-05-01 Tom Tromey <tromey@adacore.com>
* gdb.ada/packed_array_assign.exp: Add packed assignment
regression test.
|
|
When running fullpath-expand.exp with target_board=dwarf4-gdb-index, we run
into:
...
$ make check-gdb RUNTESTFLAGS="--target_board=dwarf4-gdb-index fullpath-expand.exp"
Running src/gdb/testsuite/gdb.base/fullpath-expand.exp ...
gdb compile failed, cc-with-tweaks.sh: unable to find usable gdb
=== gdb Summary ===
nr of untested testcases 1
...
The same happens with fullname.exp.
The dwarf4-gdb-index.exp board file includes cc-with-tweaks.exp, which uses
cc-with-tweaks.sh, which calls gdb-add-index.sh.
The gdb-add-index.sh script uses a gdb executable, defaulting to gdb:
...
GDB=${GDB:=gdb}
...
The cc-with-tweaks.sh script tries to ensure that the build gdb executable is
used by gdb-add-index.sh:
...
if [ -z "$GDB" ]
then
if [ -f ./gdb ]
then
GDB="./gdb -data-directory data-directory"
elif [ -f ../gdb ]
then
GDB="../gdb -data-directory ../data-directory"
elif [ -f ../../gdb ]
then
GDB="../../gdb -data-directory ../../data-directory"
else
echo "$myname: unable to find usable gdb" >&2
exit 1
fi
fi
...
So, if the current directory is build/gdb/testsuite, then a gdb executable
build/gdb/testsuite/../gdb will be used.
However, in the case of fullpath-expand.exp the test cd's into the sources:
...
set saved_pwd [pwd]
cd $srcdir
set err [gdb_compile "${subdir}/${srcfile} ${subdir}/${srcfile2}" $binfile \
executable {debug}]
cd $saved_pwd
...
and cc-with-tweaks.sh generates the "unable to find usable gdb" error.
The same error occurs if we use --target_board=cc-with-dwz instead (only in
this case we actually don't need gdb, we just need the GDB variable to be set
in cc-with-tweaks.sh, which arguably is a bug in cc-with-tweaks.sh).
Fix both errors in cc-with-tweaks.exp by generating a gdb script gdb.sh using
$GDB, $GDBFLAGS and $INTERNAL_GDBFLAGS and passing this script to
cc-with-tweaks.sh by setting env(GDB).
Tested on x86_64-linux for gdb.base.
gdb/testsuite/ChangeLog:
2019-05-01 Tom de Vries <tdevries@suse.de>
* boards/cc-with-tweaks.exp: Generate gdb.sh, and pass it in env(GDB).
|
|
Board file dwarf4-gdb-index.exp contains all the commands from
cc-with-tweaks.exp (with CC_WITH_TWEAKS_FLAGS set to "-i").
Make dwarf4-gdb-index.exp smaller by including cc-with-tweaks.exp.
Tested on x86_64-linux for gdb.base.
gdb/testsuite/ChangeLog:
2019-05-01 Tom de Vries <tdevries@suse.de>
* boards/dwarf4-gdb-index.exp: Use cc-with-tweaks.exp.
|
|
Dwarf5 defines DW_FORM_strx1 and others, which are similar
to DW_FORM_strx but uses 1-4 bytes unsigned integers. This is
a small step towards supporting dwarf5 in gdb.
|
|
When an Ada program is dynamically linked against libgnat, and when
one of the standard exceptions is used, the exception object may be
referenced by the main executable using a copy relocation.
In this situation, a "catch exception" for those exceptions will not
manage to stop. This happens because, under the hood, "catch
exception" creates an expression object that examines the object
addresses -- but in this case, the address will be incorrect.
This patch fixes the problem by arranging for these filter expressions
to examine all the relevant minimal symbols. This way, the object
from libgnat will be found as well.
Tested on x86-64 Fedora 29.
gdb/ChangeLog
2019-04-30 Tom Tromey <tromey@adacore.com>
* ada-lang.c (ada_lookup_simple_minsyms): New function.
(create_excep_cond_exprs): Iterate over program spaces.
(ada_exception_catchpoint_cond_string): Examine all minimal
symbols for exception types.
gdb/testsuite/ChangeLog
2019-04-30 Tom Tromey <tromey@adacore.com>
* lib/ada.exp (find_ada_tool): New proc.
* lib/gdb.exp (gdb_compile_shlib): Allow .o files as inputs.
* gdb.ada/catch_ex_std.exp: New file.
* gdb.ada/catch_ex_std/foo.adb: New file.
* gdb.ada/catch_ex_std/some_package.adb: New file.
* gdb.ada/catch_ex_std/some_package.ads: New file.
|
|
PR c++/24470 concerns a crash in dwarf2read.c that occurs with a
particular test case.
The issue turns out to be that process_structure_scope will pass NULL
to symbol_symtab. This happens because new_symbol decided not to
create a symbol for the particular DIE.
This patch fixes the problem by finding another reasonably-appropriate
symtab to use instead; issuing a complaint if one cannot be found for
some reason.
As mentioned in the bug, I think there are other bugs here. For
example, when using "ptype" on the "l" object in the test case, I
think I would expect to see the template parameter. I didn't research
this too closely, since it seemed more important to fix the crash.
Tested on x86-64 Fedora 29.
I'd like to check this in to the 8.3 branch as well.
2019-04-30 Tom Tromey <tromey@adacore.com>
PR c++/24470:
* dwarf2read.c (process_structure_scope): Handle case where type
has template parameters but no symbol was created.
gdb/testsuite/ChangeLog
2019-04-30 Tom Tromey <tromey@adacore.com>
PR c++/24470:
* gdb.cp/temargs.cc: Add test code from PR.
|
|
Types in Fortran can have the 'allocatable' qualifier attached to
indicate that memory needs to be explicitly allocated by the user.
This patch extends GDB to show this qualifier when printing types.
Lots of tests results are then updated to include this new qualifier
in the expected results.
gdb/ChangeLog:
* f-typeprint.c (f_type_print_base): Print 'allocatable' type
qualifier.
* gdbtypes.h (TYPE_IS_ALLOCATABLE): Define.
gdb/testsuite/ChangeLog:
* gdb.fortran/vla-datatypes.exp: Update expected results.
* gdb.fortran/vla-ptype.exp: Likewise.
* gdb.fortran/vla-type.exp: Likewise.
* gdb.fortran/vla-value.exp: Likewise.
|
|
The whitespace produced as types are printed seems inconsistent. This
commit updates the rules in an attempt to make whitespace more
balanced and consistent. Expected results are updated.
gdb/ChangeLog:
* f-typeprint.c (f_print_type): Update rules for printing
whitespace.
(f_type_print_varspec_suffix): Likewise.
gdb/testsuite/ChangeLog:
* gdb.fortran/ptr-indentation.exp: Update expected results.
* gdb.fortran/ptype-on-functions.exp: Likewise.
* gdb.fortran/vla-ptr-info.exp: Likewise.
* gdb.fortran/vla-value.exp: Likewise.
|
|
Before this commit using ptype on a Fortran function will include
information about the functions return type, but not the expected
arguments as it would for C or C++. After this commit argument types
are included in the ptype output.
For example, before GDB prints:
(gdb) ptype fun1
type = integer(kind=4) ()
(gdb) ptype is_bigger
type = logical(kind=4) ()
and after GDB prints:
(gdb) ptype fun1
type = integer(kind=4) (integer(kind=4))
(gdb) ptype is_bigger
type = logical(kind=4) (integer(kind=4), integer(kind=4))
gdb/ChangeLog:
* f-typeprint.c (f_type_print_varspec_suffix): Handle printing
function arguments.
gdb/testsuite/ChangeLog:
* gdb.fortran/ptype-on-functions.exp: New file.
* gdb.fortran/ptype-on-functions.f90: New file.
|
|
For a program compiled with gfortran the base type names are written
as lower cases in the DWARF, and so GDB will display them as lower
case. Additionally, in most places where GDB supplies its own type
names (for example all of the types defined in f-lang.c in
`build_fortran_types`), the type names are all lower case.
An exception to this is where GDB prints the void type for Fortran.
In this case GDB uses upper case.
I'm not aware of any reason why this type should merit special
attention, and it looks our of place when printing types, so this
commit changes from 'VOID' to 'void' to match all the other types.
gdb/ChangeLog:
* f-lang.c (build_fortran_types): Change name of void type to
lower case.
* f-typeprint.c (f_type_print_base): Print the name of the void
type, rather than a fixed string.
* f-valprint.c (f_decorations): Use lower case void string.
gdb/testsuite/ChangeLog:
* gdb.fortran/exprs.exp (test_convenience_variables): Expect lower
case void string.
|
|
Currently when using $_creal and $_cimag to access the components of a
complex number the types of these components will have C type names
'float', 'double', etc. This is because the components of a complex
number are not given type names in DWARF, so GDB has to pick some
suitable names, and currently we always use the C names.
This commit changes the type names used based on the language, so for
Fortran we will now use the Fortran float types, and so will get the
Fortran float type names 'real', 'real*8', etc.
gdb/ChangeLog:
* dwarf2read.c (dwarf2_init_complex_target_type): Use different
types for Fortran.
gdb/testsuite/ChangeLog:
* gdb.fortran/complex.exp: Expand.
* gdb.fortran/complex.f: Renamed to...
* gdb.fortran/complex.f90: ...this, and extended to add more
complex values.
|
|
Add some additional builtin procedures for Fortran, these are MOD,
CEILING, FLOOR, MODULO, and CMPLX.
gdb/ChangeLog:
* f-exp.y (BINOP_INTRINSIC): New token.
(exp): New parser rule handling BINOP_INTRINSIC.
(f77_keywords): Add new builtin procedures.
* f-lang.c (evaluate_subexp_f): Handle BINOP_MOD, UNOP_FORTRAN_CEILING,
UNOP_FORTRAN_FLOOR, BINOP_FORTRAN_MODULO, BINOP_FORTRAN_CMPLX.
(operator_length_f): Handle UNOP_FORTRAN_CEILING,
UNOP_FORTRAN_FLOOR, BINOP_FORTRAN_MODULO, BINOP_FORTRAN_CMPLX.
(print_unop_subexp_f): New function.
(print_binop_subexp_f): New function.
(print_subexp_f): Handle UNOP_FORTRAN_CEILING, UNOP_FORTRAN_FLOOR,
BINOP_FORTRAN_MODULO, BINOP_FORTRAN_CMPLX.
(dump_subexp_body_f): Likewise.
(operator_check_f): Likewise.
* fortran-operator.def: Add UNOP_FORTRAN_CEILING, UNOP_FORTRAN_FLOOR,
BINOP_FORTRAN_MODULO, BINOP_FORTRAN_CMPLX
gdb/testsuite/ChangeLog:
* gdb.fortran/intrinsics.exp: Extend to cover MOD, CEILING, FLOOR,
MODULO, CMPLX.
|
|
Introduce a new print setting max-depth which can be set with 'set
print max-depth DEPTH'. The default value of DEPTH is 20, but this
can also be set to unlimited.
When GDB is printing a value containing nested structures GDB will
stop descending at depth DEPTH. Here is a small example:
typedef struct s1 { int a; } s1;
typedef struct s2 { s1 b; } s2;
typedef struct s3 { s2 c; } s3;
typedef struct s4 { s3 d; } s4;
s4 var = { { { { 3 } } } };
The following table shows how various depth settings affect printing
of 'var':
| Depth Setting | Result of 'p var' |
|---------------+--------------------------------|
| Unlimited | $1 = {d = {c = {b = {a = 3}}}} |
| 4 | $1 = {d = {c = {b = {a = 3}}}} |
| 3 | $1 = {d = {c = {b = {...}}}} |
| 2 | $1 = {d = {c = {...}}} |
| 1 | $1 = {d = {...}} |
| 0 | $1 = {...} |
Only structures, unions, and arrays are replaced in this way, scalars
and strings are not replaced.
The replacement is counted from the level at which you print, not from
the top level of the structure. So, consider the above example and
this GDB session:
(gdb) set print max-depth 2
(gdb) p var
$1 = {d = {c = {...}}}
(gdb) p var.d
$2 = {c = {b = {...}}}
(gdb) p var.d.c
$3 = {b = {a = 3}}
Setting the max-depth to 2 doesn't prevent the user from exploring
deeper into 'var' by asking for specific sub-fields to be printed.
The motivation behind this feature is to try and give the user more
control over how much is printed when examining large, complex data
structures.
The default max-depth of 20 means that there is a change in GDB's
default behaviour. Someone printing a data structure with 20 levels
of nesting will now see '{...}' instead of their data, they would need
to adjust the max depth, or call print again naming a specific field
in order to dig deeper into their data structure. If this is
considered a problem then we could increase the default, or even make
the default unlimited.
This commit relies on the previous commit, which added a new field to
the language structure, this new field was a string that contained the
pattern that should be used when a structure/union/array is replaced
in the output, this allows languages to use a syntax that is more
appropriate, mostly this will be selecting the correct types of
bracket '(...)' or '{...}', both of which are currently in use.
This commit should have no impact on MI output, expressions are
printed through the MI using -var-create and then -var-list-children.
As each use of -var-list-children only ever displays a single level of
an expression then the max-depth setting will have no impact.
This commit also adds the max-depth mechanism to the scripting
language pretty printers following basically the same rules as for the
built in value printing.
One quirk is that when printing a value using the display hint 'map',
if the keys of the map are structs then GDB will hide the keys one
depth level after it hides the values, this ensures that GDB produces
output like this:
$1 = map_object = {[{key1}] = {...}, [{key2}] = {...}}
Instead of this less helpful output:
$1 = map_object = {[{...}] = {...}, [{...}] = {...}}
This is covered by the new tests in gdb.python/py-nested-maps.exp.
gdb/ChangeLog:
* cp-valprint.c (cp_print_value_fields): Allow an additional level
of depth when printing anonymous structs or unions.
* guile/scm-pretty-print.c (gdbscm_apply_val_pretty_printer):
Don't print either the top-level value, or the children if the
max-depth is exceeded.
(ppscm_print_children): When printing the key of a map, allow one
extra level of depth.
* python/py-prettyprint.c (gdbpy_apply_val_pretty_printer): Don't
print either the top-level value, or the children if the max-depth
is exceeded.
(print_children): When printing the key of a map, allow one extra
level of depth.
* python/py-value.c (valpy_format_string): Add max_depth keyword.
* valprint.c: (PRINT_MAX_DEPTH_DEFAULT): Define.
(user_print_options): Initialise max_depth field.
(val_print_scalar_or_string_type_p): New function.
(val_print): Check to see if the max depth has been reached.
(val_print_check_max_depth): Define new function.
(show_print_max_depth): New function.
(_initialize_valprint): Add 'print max-depth' option.
* valprint.h (struct value_print_options) <max_depth>: New field.
(val_print_check_max_depth): Declare new function.
* NEWS: Document new feature.
gdb/doc/ChangeLog:
* gdb.texinfo (Print Settings): Document 'print max-depth'.
* guile.texi (Guile Pretty Printing API): Document that 'print
max-depth' can effect the display of a values children.
* python.texi (Pretty Printing API): Likewise.
(Values From Inferior): Document max_depth keyword.
gdb/testsuite/ChangeLog:
* gdb.base/max-depth.c: New file.
* gdb.base/max-depth.exp: New file.
* gdb.python/py-nested-maps.c: New file.
* gdb.python/py-nested-maps.exp: New file.
* gdb.python/py-nested-maps.py: New file.
* gdb.python/py-format-string.exp (test_max_depth): New proc.
(test_all_common): Call test_max_depth.
* gdb.fortran/max-depth.exp: New file.
* gdb.fortran/max-depth.f90: New file.
* gdb.go/max-depth.exp: New file.
* gdb.go/max-depth.go: New file.
* gdb.modula2/max-depth.exp: New file.
* gdb.modula2/max-depth.c: New file.
* lib/gdb.exp (get_print_expr_at_depths): New proc.
|
|
When running gdb-caching-proc.exp, if skip_opencl_tests fails like this:
...
(gdb) run
Starting program: \
build/gdb/testsuite/outputs/gdb.base/gdb-caching-proc/opencltest13530.x
CHK_ERR (clGetPlatformIDs (1, &platform, NULL), -1001)
src/gdb/testsuite/lib/opencl_hostapp.c:73 error: Unknown
[Inferior 1 (process 13600) exited with code 01]
(gdb)
skip_opencl_tests: OpenCL support not detected
...
then this regexp in skip_opencl_tests fails to match:
...
-re ".*$inferior_exited_re code.*${gdb_prompt} $" {
...
so instead we hit the default clause after a 30 seconds timeout. With the
iteration count set at 10, we end up taking 6 minutes to run this test-case.
Fix this by adding the missing "with" in the regexp, bring back the runtime to
half a minute.
Tested on x86_64-linux.
gdb/testsuite/ChangeLog:
2019-04-29 Tom de Vries <tdevries@suse.de>
* lib/opencl.exp (skip_opencl_tests): Add missing "with" in regexp.
|
|
Inferior function calls are powerful but might lead to undesired
results such as crashes when calling nested functions (frequently
used in particular in Ada).
This implements a GDB setting to disable calling inferior functions.
Note: the idea is that if/when the 'slash command' patch is pushed,
that this setting can be changed e.g. by using the shortcut /c.
This is version 2 of the patch. It handles all the received comments,
mostly replace 'can-call' by 'may-call', and avoid using
'inferior function call' in factor of 'calling function in the program'.
2019-04-26 Philippe Waroquiers <philippe.waroquiers@skynet.be>
gdb/ChangeLog
* NEWS: Mention the new set|show may-call-functions.
* infcall.c (may_call_functions_p): New variable.
(show_may_call_functions_p): New function.
(call_function_by_hand_dummy): Throws an error if not
may-call-functions.
(_initialize_infcall): Call add_setshow_boolean_cmd for
may-call-functions.
gdb/testsuite/ChangeLog
* gdb.base/callexit.exp: Test may-call-functions off.
gdb/doc/ChangeLog
* gdb.texinfo (Calling): Document the new
set|show may-call-functions.
|
|
This bug finds another usage where we end up segfaulting while
normalizing user input. inspect_type and replace_type recurse,
attempting to substitute the "real" symbol name for the typedef name.
However, since the both these names are the same, they keep calling
each other until the stack overflows.
A simple reproducer for it is given by
typedef struct foo foo;
int qux (foo *f) { return 0; }
(gdb) b qux(foo*)
Segmentation fault
inspect_type already contains some special handling to prevent a
similar situation from occurring with namespaces. I wonder, however,
whether we need be so pedantic about the exact nature of the substitution.
This patch implements this rather more aggressive assumption that these
substitutions should be avoided whenever the replacement symbol's name is
exactly the same as the one we're trying to substitute. [In the above
example, we're trying to substitute the tyepdef named "foo" with the symbol
named "foo" (a struct).]
gdb/ChangeLog:
PR c++/24367
* cp-support.c (inspect_type): Don't attempt substitutions
of symbol with the same name.
gdb/testsuite/ChangeLog:
PR c++/24367
* gdb.cp/meth-typedefs.cc (incomplete_struct)
(another_incomplete_struct, test_incomplete): New definitions.
(main): Use new definitions.
* gdb.cp/meth-typedefs.exp: Add new tests for `test_incomplete'
functions.
|
|
DW_FORM_strx is the new name of DW_FORM_GNU_str_index in the Dwarf 5 standard.
This is a small step towards supporting Dwarf 5 in gdb.
|
|
I forgot to include the ChangeLog entries in the commit
57e5e645010430b3d73f8c6a757d09f48dc8f8d5 ("Implement dump of mappings
with ELF headers by gcore").
|
|
This patch has a long story, but it all started back in 2015, with
commit df8411da087dc05481926f4c4a82deabc5bc3859 ("Implement support
for checking /proc/PID/coredump_filter"). The purpose of that commit
was to bring GDB's corefile generation closer to what the Linux kernel
does. However, back then, I did not implement the full support for
the dumping of memory mappings containing ELF headers (like mappings
of DSOs or executables). These mappings were being dumped most of
time, though, because the default value of /proc/PID/coredump_filter
is 0x33, which would cause anonymous private mappings (DSOs/executable
code mappings have this type) to be dumped. Well, until something
happened on binutils...
A while ago, I noticed something strange was happening with one of our
local testcases on Fedora GDB: it was failing due to some strange
build-id problem. On Fedora GDB, we (unfortunately) carry a bunch of
"local" patches, and some of these patches actually extend upstream's
build-id support in order to generate more useful information for the
user of a Fedora system (for example, when the user loads a corefile
into GDB, we detect whether the executable that generated that
corefile is present, and if it's not we issue a warning suggesting
that it should be installed, while also providing the build-id of the
executable). A while ago, Fedora GDB stopped printing those warnings.
I wanted to investigate this right away, and spent some time trying to
determine what was going on, but other things happened and I got
sidetracked. Meanwhile, the bug started to be noticed by some of our
users, and its priority started changing. Then, someone on IRC also
mentioned the problem, and when I tried helping him, I noticed he
wasn't running Fedora. Hm... So maybe the bug was *also* present
upstream.
After "some" time investigating, and with a lot of help from Keith and
others, I was finally able to determine that yes, the bug is also
present upstream, and that even though it started with a change in ld,
it is indeed a GDB issue.
So, as I said, the problem started with binutils, more specifically
after the following commit was pushed:
commit f6aec96dce1ddbd8961a3aa8a2925db2021719bb
Author: H.J. Lu <hjl.tools@gmail.com>
Date: Tue Feb 27 11:34:20 2018 -0800
ld: Add --enable-separate-code
This commit makes ld use "-z separate-code" by default on x86-64
machines. What this means is that code pages and data pages are now
separated in the binary, which is confusing GDB when it tries to decide
what to dump.
BTW, Fedora 28 binutils doesn't have this code, which means that
Fedora 28 GDB doesn't have the problem. From Fedora 29 on, binutils
was rebased and incorporated the commit above, which started causing
Fedora GDB to fail.
Anyway, the first thing I tried was to pass "-z max-page-size" and
specify a bigger page size (I saw a patch that did this and was
proposed to Linux, so I thought it might help). Obviously, this
didn't work, because the real "problem" is that ld will always use
separate pages for code and data. So I decided to look into how GDB
dumped the pages, and that's where I found the real issue.
What happens is that, because of "-z separate-code", the first two pages
of the ELF binary are (from /proc/PID/smaps):
00400000-00401000 r--p 00000000 fc:01 799548 /file
Size: 4 kB
KernelPageSize: 4 kB
MMUPageSize: 4 kB
Rss: 4 kB
Pss: 4 kB
Shared_Clean: 0 kB
Shared_Dirty: 0 kB
Private_Clean: 4 kB
Private_Dirty: 0 kB
Referenced: 4 kB
Anonymous: 0 kB
LazyFree: 0 kB
AnonHugePages: 0 kB
ShmemPmdMapped: 0 kB
Shared_Hugetlb: 0 kB
Private_Hugetlb: 0 kB
Swap: 0 kB
SwapPss: 0 kB
Locked: 0 kB
THPeligible: 0
VmFlags: rd mr mw me dw sd
00401000-00402000 r-xp 00001000 fc:01 799548 /file
Size: 4 kB
KernelPageSize: 4 kB
MMUPageSize: 4 kB
Rss: 4 kB
Pss: 4 kB
Shared_Clean: 0 kB
Shared_Dirty: 0 kB
Private_Clean: 0 kB
Private_Dirty: 4 kB
Referenced: 4 kB
Anonymous: 4 kB
LazyFree: 0 kB
AnonHugePages: 0 kB
ShmemPmdMapped: 0 kB
Shared_Hugetlb: 0 kB
Private_Hugetlb: 0 kB
Swap: 0 kB
SwapPss: 0 kB
Locked: 0 kB
THPeligible: 0
VmFlags: rd ex mr mw me dw sd
Whereas before, we had only one:
00400000-00401000 r-xp 00000000 fc:01 798593 /file
Size: 4 kB
KernelPageSize: 4 kB
MMUPageSize: 4 kB
Rss: 4 kB
Pss: 4 kB
Shared_Clean: 0 kB
Shared_Dirty: 0 kB
Private_Clean: 0 kB
Private_Dirty: 4 kB
Referenced: 4 kB
Anonymous: 4 kB
LazyFree: 0 kB
AnonHugePages: 0 kB
ShmemPmdMapped: 0 kB
Shared_Hugetlb: 0 kB
Private_Hugetlb: 0 kB
Swap: 0 kB
SwapPss: 0 kB
Locked: 0 kB
THPeligible: 0
VmFlags: rd ex mr mw me dw sd
Notice how we have "Anonymous" data mapped into the page. This will be
important.
So, the way GDB decides which pages it should dump has been revamped
by my patch in 2015, and now it takes the contents of
/proc/PID/coredump_filter into account. The default value for Linux
is 0x33, which means:
Dump anonymous private, anonymous shared, ELF headers and HugeTLB
private pages.
Or:
filter_flags filterflags = (COREFILTER_ANON_PRIVATE
| COREFILTER_ANON_SHARED
| COREFILTER_ELF_HEADERS
| COREFILTER_HUGETLB_PRIVATE);
Now, it is important to keep in mind that GDB doesn't always have *all*
of the necessary information to exactly determine the type of a page, so
the whole algorithm is based on heuristics (you can take a look at
linux-tdep.c:dump_mapping_p and
linux-tdep.c:linux_find_memory_regions_full for more info).
Before the patch to make ld use "-z separate-code", the (single) page
containing data and code was being flagged as an anonymous (due to the
non-zero "Anonymous:" field) private (due to the "r-xp" permission),
which means that it was being dumped into the corefile. That's why it
was working fine.
Now, as you can imagine, when "-z separate-code" is used, the *data*
page (which is where the ELF notes are, including the build-id one) now
doesn't have any "Anonymous:" mapping, so the heuristic is flagging it
as file-backed private, which is *not* dumped by default.
The next question I had to answer was: how come a corefile generated by
the Linux kernel was correct? Well, the answer is that GDB, unlike
Linux, doesn't actually implement the COREFILTER_ELF_HEADERS support.
On Linux, even though the data page is also treated as a file-backed
private mapping, it is also checked to see if there are any ELF headers
in the page, and then, because we *do* have ELF headers there, it is
dumped.
So, after more time trying to think of ways to fix this, I was able to
implement an algorithm that reads the first few bytes of the memory
mapping being processed, and checks to see if the ELF magic code is
present. This is basically what Linux does as well, except that, if
it finds the ELF magic code, it just dumps one page to the corefile,
whereas GDB will dump the whole mapping. But I don't think that's a
big issue, to be honest.
It's also important to explain that we *only* perform the ELF magic
code check if:
- The algorithm has decided *not* to dump the mapping so far, and;
- The mapping is private, and;
- The mapping's offset is zero, and;
- The user has requested us to dump mappings with ELF headers.
IOW, we're not going to blindly check every mapping.
As for the testcase, I struggled even more trying to write it. Since
our build-id support on upstream GDB is not very extensive, it's not
really possible to determine whether a corefile contains build-id
information or not just by using GDB. So, after thinking a lot about
the problem, I decided to rely on an external tool, eu-unstrip, in
order to verify whether the dump was successful. I verified the test
here on my machine, and everything seems to work as expected (i.e., it
fails without the patch, and works with the patch applied). We are
working hard to upstream our "local" Fedora GDB patches, and we intend
to submit our build-id extension patches "soon", so hopefully we'll be
able to use GDB itself to perform this verification.
I built and regtested this on the BuildBot, and no problems were
found.
gdb/ChangeLog:
2019-04-25 Sergio Durigan Junior <sergiodj@redhat.com>
PR corefiles/11608
PR corefiles/18187
* linux-tdep.c (dump_mapping_p): Add new parameters ADDR and
OFFSET. Verify if current mapping contains an ELF header.
(linux_find_memory_regions_full): Adjust call to
dump_mapping_p.
gdb/testsuite/ChangeLog:
2019-04-25 Sergio Durigan Junior <sergiodj@redhat.com>
PR corefiles/11608
PR corefiles/18187
* gdb.base/coredump-filter-build-id.exp: New file.
|
|
Add both board option and environment variable which enables gdbserver
debug and sends it to the file gdbserver.debug, located in the output
directory for the current test. Document this.
Add support for the environment variable in the Makefile.
The testsuite can be run with gdbserver debug enabled in the following way:
make check GDBSERVER_DEBUG=all
Disable tspeed.exp when debugging to prevent the log file filling
many gigabytes then timing out.
gdb/testsuite/ChangeLog:
* Makefile.in: Pass through GDBSERVER_DEBUG.
* README (Testsuite Parameters): Add GDBSERVER_DEBUG.
(gdbserver,debug): Add board setting.
* gdb.trace/tspeed.exp: Skip when debugging.
* lib/gdb.exp (gdbserver_debug_enabled): New procedure.
* lib/gdbserver-support.exp: Likewise
|
|
This changes the gdb test suite to omit -fno-stack-protector when
compiling Rust code. This makes Rust testing work again.
I think I saw this patch somewhere already, but I couldn't find it
again just now, so I'm checking this version in.
gdb/testsuite/ChangeLog
2019-04-24 Tom Tromey <tromey@adacore.com>
* lib/gdb.exp (gdb_compile): Don't add -fno-stack-protector for
Rust.
|
|
Commit 4aa866af ("Fix AMD64 return value ABI in expression
evaluation") introduced a regression when calling a function with a
structure that contains bitfields.
Because the caller of amd64_has_unaligned_fields handles bitfields
already, it seemed to me that the simplest fix was to ignore bitfields
here.
gdb/ChangeLog
2019-04-24 Tom Tromey <tromey@adacore.com>
* amd64-tdep.c (amd64_has_unaligned_fields): Ignore bitfields.
gdb/testsuite/ChangeLog
2019-04-24 Tom Tromey <tromey@adacore.com>
* gdb.arch/amd64-eval.exp: Test bitfield return.
* gdb.arch/amd64-eval.cc (struct Bitfields): New.
(class Foo) <return_bitfields>: New method.
(main): Call it.
|
|
Replaces use of aarch64_type_align with common type_align function.
Doing this fixes a bug in aarch64_type_align where static fields are
considered as part of the alignment calculation of a struct, which
results in arguments passed on the stack being misaligned. This bug
is exposed in the new test gdb.cp/many-args.exp.
Part of the old aarch64_type_align is retained and used as the gdbarch
type align callback in order to correctly align vectors.
gdb/ChangeLog:
* aarch64-tdep.c (aarch64_type_align): Only handle vector override
case.
(pass_on_stack): Use type_align.
(aarch64_gdbarch_init): Register aarch64_type_align gdbarch
function.
gdb/testsuite/ChangeLog:
* gdb.cp/many-args.cc: New file.
* gdb.cp/many-args.exp: New file.
|
