gdb: Don't allow annotations to influence what else GDB prints
A change was accidentally made that moved a call to do_gdb_disassembly
out of an if block guarded by 'if (source_print && sal.symtab)'. The
result was that if a user has 'set disassemble-next-line on' then the
backtrace would now include some disassembly of a few instructions in
each frame.
This change was not intentional, but was not spotted by any tests.
This commit restores the old behaviour and adds a test to ensure this
doesn't break again in the future.
Shahab Vahedi [Mon, 6 Jan 2020 14:27:32 +0000 (15:27 +0100)]
GDB: Fix the overflow in addr/line_is_displayed()
In tui_disasm_window::addr_is_displayed(), there can be situations
where "content" is empty. For instance, it can happen when the
"content" was not filled in tui_disasm_window::set_contents(),
because tui_disassemble() threw an exception. Usually this exception
is the result of fetching invalid PC addresses like the ones beyond
the end of the program.
Having "content.size ()" zero leads to an overflow in this condition
check inside tui_disasm_window::addr_is_displayed():
int i = 0;
while (i < content.size () - threshold ...) {
... content[i] ...
}
"threshold" is 2 and there are times that "content.size ()" is 0.
This results into an overflow and the loop is entered whereas it
should have been skipped. Finally, "content[i]" access leads to
a segmentation fault.
Same problem applies to tui_source_window::line_is_displayed().
The issue has been discussed at length in bug 25345:
https://sourceware.org/bugzilla/show_bug.cgi?id=25345
This commit avoids the segmentation faults with an early check:
if (content.size () < SCROLL_THRESHOLD)
return false;
Moreover, those functions have been overhauled to a leaner code.
* tui/tui-disasm.c (tui_disasm_window::addr_is_displayed): Avoid
overflow by an early check of content vs threshold.
* tui/tui-source.c (tui_source_window::line_is_displayed):
Likewise.
Jon Turney [Thu, 2 Jan 2020 00:25:56 +0000 (00:25 +0000)]
Fix a crash with a malformed PE header
Don't try to read the PE export table when no section contains the RVA
for it.
(I have a PE executable [1] packed with UPX, where the export table data
directory entry contains a RVA which doesn't correspond to any section.
Mistakenly trying to debug this with gdb makes it crash.)
Nick Clifton [Mon, 6 Jan 2020 16:24:23 +0000 (16:24 +0000)]
Add -fcommon compiler command line option to linker tests that need common symbols.
PR 25327
* testsuite/ld-elf/shared.exp: Add -fcommon option to compiler
command line when building libcomm1.o and pr13250 tests.
* testsuite/ld-plugin/lto.exp: Likewise for pr20267 tests.
Hannes Domani [Sun, 22 Dec 2019 13:58:22 +0000 (14:58 +0100)]
Fix search in TUI
The variable last_line_listed is never set when print_source_lines_base is
called in TUI mode, so the search always started from the last line printed
outside of TUI mode.
gdb/ChangeLog:
2020-01-06 Hannes Domani <ssbssa@yahoo.de>
* source.c (print_source_lines_base): Set last_line_listed.
Eli Zaretskii [Mon, 6 Jan 2020 11:51:54 +0000 (11:51 +0000)]
Improve process exit status macros on MinGW
When a Windows program is terminated by a fatal exception, its exit
code is the value of that exception, as defined by the various
EXCEPTION_* symbols in the Windows API headers. This commit emulates
WTERMSIG etc. by translating the fatal exception codes to more-or-less
equivalent Posix signals.
gdb/ChangeLog:
2020-01-06 Eli Zaretskii <eliz@gnu.org>
Pedro Alves <palves@redhat.com>
* Makefile.in (COMMON_SFILES): Add gdbsupport/gdb_wait.c.
* windows-tdep.c: New enumeration of WINDOWS_SIG* signals.
(windows_gdb_signal_to_target): New function, uses the above
enumeration to convert GDB internal signal codes to equivalent
Windows codes.
(windows_init_abi): Call set_gdbarch_gdb_signal_to_target.
* windows-nat.c: Include "gdb_wait.h".
(get_windows_debug_event): Extract the fatal exception from the
exit status and convert to the equivalent Posix signal number.
* cli/cli-cmds.c (exit_status_set_internal_vars): Account for the
possibility that WTERMSIG returns GDB_SIGNAL_UNKNOWN.
* gdbsupport/gdb_wait.c: New file, implements
windows_status_to_termsig.
* gdbsupport/gdb_wait.h (WIFEXITED, WIFSIGNALED, WEXITSTATUS)
(WTERMSIG) [__MINGW32__]: Separate definitions for MinGW.
gdb/gdbserver/ChangeLog:
2020-01-06 Eli Zaretskii <eliz@gnu.org>
Pedro Alves <palves@redhat.com>
* win32-low.c (get_child_debug_event): Extract the fatal exception
from the exit status and convert to the equivalent Posix signal
number.
(win32_wait): Allow TARGET_WAITKIND_SIGNALLED status as well.
* Makefile.in (OBS, SFILES): Add gdb_wait.[co].
Alan Modra [Mon, 6 Jan 2020 09:21:22 +0000 (19:51 +1030)]
bfd_check_format: ignore errors from coff_real_object_p
Since 1993-11-05 git commit c188b0bec3b, bfd_check_format has failed
if any of the target object_p functions returns false with any error
but bfd_error_wrong_format. That's just weird. There is really no
reason why coff_real_object_p should be fixed to only return that
error instead of numerous other possible errors. Even an out of
memory condition for one target doesn't necessarily mean other targets
can't match, assuming the failing target nicely returns all memory it
might have used.
* format.c (bfd_check_format_matches): Ignore bfd_error on target
match failures. Don't init to bfd_error_wrong_format before
calling _bfd_check_format.
Alan Modra [Mon, 6 Jan 2020 07:59:14 +0000 (18:29 +1030)]
alpha-vms: don't exit on stack underflow/overflow
BFD is not supposed to exit or abort on anything the user can do.
* vms-alpha.c (_bfd_vms_push, _bfd_vms_pop): Return pass/fail
status rather than exiting on stack overflow or underflow.
(_bfd_vms_slurp_etir): Adjust to suit.
Alan Modra [Sat, 4 Jan 2020 09:23:19 +0000 (19:53 +1030)]
ubsan: m32c: left shift of negative value
There are probably a lot more of these still here.
cpu/
* m32c.cpu (f-dsp-8-u16, f-dsp-8-s16): Rearrange to mask any sign
bits before shifting rather than masking after shifting.
(f-dsp-16-u16, f-dsp-16-s16, f-dsp-32-u16, f-dsp-32-s16): Likewise.
(f-dsp-40-u16, f-dsp-40-s16, f-dsp-48-u16, f-dsp-48-s16): Likewise.
(f-dsp-64-u16, f-dsp-8-s24): Likewise.
(f-bitbase32-16-s19-unprefixed): Avoid signed left shift.
opcodes/
* m32c-ibld.c: Regenerate.
Alan Modra [Sun, 5 Jan 2020 22:22:39 +0000 (08:52 +1030)]
PR25344, z80 disassembler recursion
PR 25344
* z80-dis.c (suffix): Don't use a local struct buffer copy.
Peek at next byte to prevent recursion on repeated prefix bytes.
Ensure uninitialised "mybuf" is not accessed.
(print_insn_z80): Don't zero n_fetch and n_used here,..
(print_insn_z80_buf): ..do it here instead.
Andrew Burgess [Sun, 22 Dec 2019 23:45:31 +0000 (23:45 +0000)]
gdb: use tui_set_layout not show_layout to fix window focus
When calling tui_add_win_to_layout, use tui_set_layout not show_layout
so that window focus is correctly updated. If the focus is not
correctly maintained then GDB can be crashed like this:
start
tui enable
layout asm
list SOME_FUNCTION
At this point GDB will have "popped up" the source window to
display SOME_FUNCTION. Previously no window would have focus at this
point, and so if the user now does 'focus next' or 'focus prev', then
GDB would crash.
Calling tui_set_layout ensures that focus is correctly calculated as
the source window is "popped up", and this fixes the issue.
gdb/ChangeLog:
* tui/tui-layout.c (tui_add_win_to_layout): Use tui_set_layout not
show_layout.
gdb/testsuite/ChangeLog:
* gdb.tui/list.exp: Test 'focus next' after 'list main'.
Luis Machado [Fri, 3 Jan 2020 19:08:16 +0000 (16:08 -0300)]
[AArch64] Fix erroneous use of spu architecture bfd
While investigating some SVE code, i noticed the use of two spu bfd variables.
This looks like an oversight, as the "id" field is available for non-spu
architectures as well, even though its primary use was the Cell BE
architecture.
gdb/ChangeLog:
2020-01-05 Luis Machado <luis.machado@linaro.org>
* aarch64-linux-nat.c
(aarch64_linux_nat_target::thread_architecture): Use bfd_arch_aarch64
and bfd_mach_aarch64.
Eli Zaretskii [Sun, 5 Jan 2020 05:50:27 +0000 (09:50 +0400)]
libctf: Add configure check for asprintf (for MinGW)
This commit fixes a compilation warning when compiling libctf
on MinGW:
libctf/ctf-dump.c:118:8: warning: implicit declaration of function
'asprintf'; did you mean 'vasprintf'? [-Wimplicit-function-declaration]
if (asprintf (&bit, " %lx: [slice 0x%x:0x%x]",
^~~~~~~~
vasprintf
MinGW doesn't provide that function, so we depend on the one provided
by libiberty. However, the declaration is guarded by HAVE_DECL_ASPRINTF,
which we do not have in libctf's config.h.
Alan Modra [Thu, 2 Jan 2020 07:09:32 +0000 (17:39 +1030)]
Release bfd_alloc memory in bfd_check_format_matches
It's a little tricky. We can release any memory back when we have a
match failure, but after a match success which we might want to
preserve for later use the high water mark must change to that of the
matched bfd.
* format.c (bfd_check_format_matches): Add preserve_match.
Save initial bfd state in "preserve", matched bfd state in
"preserve_match". Save just the first match. Release
bfd_alloc memory. Restore and finish preserved state as
appropriate on all function exit paths.
Alan Modra [Tue, 31 Dec 2019 23:58:42 +0000 (10:28 +1030)]
coff: free malloc'd memory on successful target match too
object_p functions cannot allocate memory by malloc and not free it
before returning. Even a successful target match may not be the best
match. If a match isn't used then those malloc'd blocks won't be
freed.
* coffgen.c (coff_real_object_p): Free malloc'd memory on target
match too.
While handling the comments of Tom related to
[RFC] Have an option to tell GDB to detect and possibly handle mismatched exec-files.
https://sourceware.org/ml/gdb-patches/2019-12/msg00621.html
I saw that GDB warnings are produced ignoring the given styles.
This patch:
* ensures that style markups are properly handled by "warning".
* changes 'set/show data-directory' so that file style is used
in warnings and in 'show message'
* changes all other messages in top.c to use file style when appropriate.
* Uses the above data-directory changes in gdb.base/style.exp
2020-01-03 Philippe Waroquiers <philippe.waroquiers@skynet.be>
* ui-file.c (stdio_file::can_emit_style_escape)
(tee_file::can_emit_style_escape): Ensure style is used also on
gdb_stderr when gdb_stderr is a tty supporting styling, similarly
to gdb_stdout.
* main.c (set_gdb_data_directory): Use file style to output the
warning that the given pathname is not a directory.
* top.c (show_history_filename, gdb_safe_append_history)
(show_gdb_datadir): Use file style.
2020-01-03 Philippe Waroquiers <philippe.waroquiers@skynet.be>
* gdb.base/style.exp: Test that warnings are styled.
Nick Clifton [Fri, 3 Jan 2020 16:17:53 +0000 (16:17 +0000)]
Fix potential illegal memory access when parsing a corrupt PEF format file.
PR 25307
(bfd_pef_parse_function_stubs): Correct the test that ensures that
there is enough data remaining in the code buffer before
attempting to read a function stub.
Hannes Domani [Fri, 3 Jan 2020 12:55:12 +0000 (12:55 +0000)]
For PE format files, the base relocation table is necessary if the image is loaded at a different image base than specified in the PE header. This patch provides a new option --enable-reloc-section to force the generation of this section.
* emultempl/pe.em: Add new option --enable-reloc-section.
* emultempl/pep.em: Likewise.
* ld.texi: Document --enable-reloc-section.
* pe-dll.c (pe_dll_build_sections): Use pe_dll_enable_reloc_section.
(pe_dll_fill_sections): Simplify by calling pe_exe_fill_sections.
* pe-dll.h: Add extern declaration of option flag.
* pep-dll.c (pe_dll_enable_reloc_section):
Add alias define for pep_dll_enable_reloc_section.
* pep-dll.h: Add extern declaration of option flag.
Fix ld/PR25316 for the ia64 target by refusing to support binary merging.
ld/PR25316
* elfnn-ia64.c (elfNN_ia64_merge_private_bfd_data): don't fail
on binary inputs ld/PR25316.
(is_ia64_elf): new helper to filter on ia64 objects.
Jan Beulich [Fri, 3 Jan 2020 09:14:16 +0000 (10:14 +0100)]
Arm64: correct {su,us}dot SIMD encodings
According to the specification these permit the Q bit to control the
vector length operated on, and hence this bit should not already be set
in the opcode table entries (it rather needs setting dynamically). Note
how the test case output did also not match its input. Besides
correcting the test case also extend it to cover both forms.
Add support for the GBZ80, Z180, and eZ80 variants of the Z80 architecure. Add an ELF based target for these as well.
PR 25224
bfd * Makefile.am: Add z80-elf target support.
* configure.ac: Likewise.
* targets.c: Likewise.
* config.bfd: Add z80-elf target support and new arches: ez80 and z180.
* elf32-z80.c: New file.
* archures.c: Add new z80 architectures: eZ80 and Z180.
* coffcode.h: Likewise.
* cpu-z80.c: Likewise.
* bfd-in2.h: Likewise plus additional Z80 relocations.
* coff-z80.c: Add new relocations for Z80 target and local label check.
gas * config/tc-z80.c: Add new architectures: Z180 and eZ80. Add support
for assembler code generated by SDCC. Add new relocation types. Add
z80-elf target support.
* config/tc-z80.h: Add z80-elf target support. Enable dollar local
labels. Local labels starts from ".L".
* testsuite/gas/all/fwdexp.d: Fix failure due to symbol conflict.
* testsuite/gas/all/fwdexp.s: Likewise.
* testsuite/gas/z80/suffix.d: Fix failure on ELF target.
* testsuite/gas/z80/z80.exp: Add new tests
* testsuite/gas/z80/dollar.d: New file.
* testsuite/gas/z80/dollar.s: New file.
* testsuite/gas/z80/ez80_adl_all.d: New file.
* testsuite/gas/z80/ez80_adl_all.s: New file.
* testsuite/gas/z80/ez80_adl_suf.d: New file.
* testsuite/gas/z80/ez80_isuf.s: New file.
* testsuite/gas/z80/ez80_z80_all.d: New file.
* testsuite/gas/z80/ez80_z80_all.s: New file.
* testsuite/gas/z80/ez80_z80_suf.d: New file.
* testsuite/gas/z80/r800_extra.d: New file.
* testsuite/gas/z80/r800_extra.s: New file.
* testsuite/gas/z80/r800_ii8.d: New file.
* testsuite/gas/z80/r800_z80_doc.d: New file.
* testsuite/gas/z80/z180.d: New file.
* testsuite/gas/z80/z180.s: New file.
* testsuite/gas/z80/z180_z80_doc.d: New file.
* testsuite/gas/z80/z80_doc.d: New file.
* testsuite/gas/z80/z80_doc.s: New file.
* testsuite/gas/z80/z80_ii8.d: New file.
* testsuite/gas/z80/z80_ii8.s: New file.
* testsuite/gas/z80/z80_in_f_c.d: New file.
* testsuite/gas/z80/z80_in_f_c.s: New file.
* testsuite/gas/z80/z80_op_ii_ld.d: New file.
* testsuite/gas/z80/z80_op_ii_ld.s: New file.
* testsuite/gas/z80/z80_out_c_0.d: New file.
* testsuite/gas/z80/z80_out_c_0.s: New file.
* testsuite/gas/z80/z80_reloc.d: New file.
* testsuite/gas/z80/z80_reloc.s: New file.
* testsuite/gas/z80/z80_sli.d: New file.
* testsuite/gas/z80/z80_sli.s: New file.
Tamar Christina [Thu, 2 Jan 2020 14:06:01 +0000 (14:06 +0000)]
AArch64: Set the correct ELF class for AArch64 stubs (PR/25210)
This fixes PR 25210 by specifying the the correct ELF class for AArch64 stubs.
After doing this the stub section starts behaving like a normal object file
loaded from disk. That is SEC_LINKER_CREATED causes us to have to write the
section manually.
This flag was added as a fix for PR 24753. I believe that
fix to still be correct as linker created sections don't have a size on disk
and it fixes the Arm bootstrap regression. But in this case specifying the
correct section class also makes the stub section not be considered by
compress.c.
So I'm partially revert this change so that we don't have to manage the section
manually as implied by SEC_LINKER_CREATED.
bfd/ChangeLog:
PR 25210
PR 24753
* elfnn-aarch64.c (_bfd_aarch64_create_stub_section): Set ELF class.
Alan Modra [Tue, 31 Dec 2019 11:54:31 +0000 (22:24 +1030)]
asan: alpha-vms: Heap-buffer-overflow
This fixes yet more errors in the alpha-vms buffer size checks.
* vms-alpha.c (_bfd_vms_slurp_eisd): Don't overflow when checking
offset. Don't overflow when checking rec_size, and do allow
rec_size to the end of the buffer. Ensure eisd->type can be
accessed, not just the first 32 bytes. Don't call
_bfd_vms_save_counted_string with zero length remaining. Fail
on empty string section name.
(_bfd_vms_slurp_egsd): Formatting. Catch more reads past end
of record size. Correct remaining length calculation. Fail
on empty string section name. Consolidate error paths.
Alan Modra [Mon, 30 Dec 2019 01:18:20 +0000 (11:48 +1030)]
archive.c bfd_zalloc
Quite a few bfd_zalloc calls are wasting time clearing memory, and
should be bfd_alloc instead.
* archive.c (do_slurp_bsd_armap): Use bfd_alloc rather than
bfd_zalloc when memory is all written after the call.
(do_slurp_coff_armap): Likewise. Set bfd_error on ridiculously
large allocations that overflow bfd_size_type. Use just one
bfd_release on error exit.
(_bfd_slurp_extended_name_table): Use bfd_alloc for extended_names,
clear last byte rather than the entire array. Use bfd_alloc for
string table. Rearrange and simplify code copying file names.
Bernd Edlinger [Sat, 23 Nov 2019 06:37:26 +0000 (07:37 +0100)]
Fix setting breakpoints or stepping on line 65535
This removes code that was present from the very first git revisison 7b4ac7e1ed2c4616bce56d1760807798be87ac9e from 1988. It was in the
gdb/dbxread.c at the time (and makes more sense for dbx line info format
since line numbers are 16-bit entities in that debug format and debugging
files with more than 65535 lines would not work anyway) but moved from
there to gdb/buildsym.c which is used for dwarf line info as well, and
excluding an arbitrary line number does certainly not make sense nowadays.
Alan Modra [Sun, 29 Dec 2019 02:26:04 +0000 (12:56 +1030)]
ubsan: alpha-vms: shift exponent is too large
* vms-alpha.c (_bfd_vms_slurp_egsd): Make base_addr a bfd_vma.
Limit alignment power. Correct and simplify alignment expression.
(evax_bfd_print_relocation_records): Avoid signed shift left.
Simon Marchi [Sat, 28 Dec 2019 01:58:42 +0000 (20:58 -0500)]
[PATCH] Adjust test gdb.ada/ptype_tagged_param.exp for when GNAT runtime does not have debug info
This test verifies that GDB correctly identifies the run-time type of
"s" as being the type "Circle". However, that can only be done
correctly if the GNAT runtime has been compiled and shipped with debug
information, so that GDB can poke in its internal data structures.
Currently the test fails when when running against a GNAT runtime
without debug info. This is the case, for example, on Arch Linux using
the distribution package.
This patch adds a helper in lib/ada.exp to check whether the GNAT
runtime has debug info or not. It then uses it in
gdb.ada/ptype_tagged_param.exp to expect a different result, depending
on whether we have debug info or not in the runtime.
At first, I made it so we would XFAIL the test, in the absence of debug
info, but then I thought that we might as well test for the output we
expect in the absence of debug info instead.
gdb/testsuite/ChangeLog:
* lib/ada.exp (gnat_runtime_has_debug_info): New proc.
* lib/gnat_debug_info_test.adb: New file.
* gdb.ada/ptype_tagged_param.exp: Use
gnat_runtime_has_debug_info, expect a different output if
runtime does not have debug info.
Jan Beulich [Fri, 27 Dec 2019 08:39:58 +0000 (09:39 +0100)]
x86: adjust ignored prefix warning for branches
There's no reason to not also issue them in Intel syntax mode, and it
can be quite helpful to mention the actual insn (after all there can be
multiple on a single line).
Jan Beulich [Fri, 27 Dec 2019 08:39:17 +0000 (09:39 +0100)]
x86-64: correct / adjust prefix emission
First and foremost REX must come last. Next JumpInterSegment branches
can't possibly have a REX prefix, as they're consistently CpuNo64. And
finally make BND prefix handling in output_branch() consistent with that
of other prefixes in the same function, and make its placement among
prefixes consistent with output_jump() (which, oddly enough, still isn't
the supposedly canonical order specified by the *_PREFIX definitions).
Jan Beulich [Fri, 27 Dec 2019 08:38:34 +0000 (09:38 +0100)]
x86-64: fix Intel64 handling of branch with data16 prefix
The expectation of x86-64-branch-3 for "call" / "jmp" with an obvious
direct destination to translate to an indirect _far_ branch is plain
wrong. The operand size prefix should have no effect at all on the
interpretation of the operand. The main underlying issue here is that
the Intel64 templates of the direct branches don't include Disp16, yet
various assumptions exist that it would always be there when there's
also Disp32/Disp32S, toggled by the operand size prefix (which is
being ignored by direct branches in Intel64 mode).
Along these lines it was also wrong to base the displacement width
decision solely on the operand size prefix: REX.W cancels this effect
and hence needs taking into consideration, too.
A disassembler change is needed here as well: XBEGIN was wrongly treated
the same as direct CALL/JMP, which isn't the case - the operand size
prefix does affect displacement size there, it's merely ignored when it
comes to updating [ER]IP.
Jan Beulich [Fri, 27 Dec 2019 08:22:03 +0000 (09:22 +0100)]
x86: consolidate Disp<NN> handling a little
In memory operand addressing, which forms of displacement are permitted
besides Disp8 is pretty clearly limited
- outside of 64-bit mode, Disp16 or Disp32 only, depending on address
size (MPX being special in not allowing Disp16),
- in 64-bit mode, Disp32s or Disp64 without address size override, and
solely Disp32 with one.
Adjust assembler and i386-gen to match this, observing that templates
already get adjusted before trying to match them against input depending
on the presence of an address size prefix.
This adjustment logic gets extended to all cases, as certain DispNN
values should also be dropped when there's no such prefix. In fact
behavior of the assembler, perhaps besides the exact diagnostics wording,
should not differ between there being templates applicable to 64-bit and
non-64-bit at the same time, or there being fully separate sets of
templates, with their DispNN settings already reduced accordingly.
This adjustment logic further gets guarded such that there wouldn't be
and Disp<N> conversion based on address size prefix when this prefix
doesn't control the width of the displacement (on branches other than
absolute ones).
These adjustments then also allow folding two MOV templates, which had
been split between 64-bit and non-64-bits variants so far.
Once in this area also
- drop the bogus DispNN from JumpByte templates, leaving just the
correct Disp8 there (compensated by i386_finalize_displacement()
now setting Disp8 on their operands),
- add the missing Disp32S to XBEGIN.
Note that the changes make it necessary to temporarily mark a test as
XFAIL; this will get taken care of by a subsequent patch. The failing
parts are entirely bogus and will get replaced.
This also renames it to make it clearer that this is not a cheap
function (to compute_and_set_names). Also renames name to m_name
to make the implementation of the renamed function more readable.
Most of the places that access sym->m_name directly were also changed
to call linkage_name () instead, to make it clearer which name they
are accessing.
gdb/ChangeLog:
2019-12-26 Christian Biesinger <cbiesinger@google.com>