Tom Tromey [Wed, 14 Aug 2019 15:18:21 +0000 (09:18 -0600)]
Replace write_inferior_memory with target_write_memory
target_write_memory is just a simple wrapper for
write_inferior_memory. Because target_write_memory is needed for
gdbsupport, and because gdb uses the name "target_write_memory"
everywhere, this patch renames write_inferior_memory and removes the
wrapper. I think this brings gdb and gdbserver slightly more in sync.
gdb/gdbserver/ChangeLog
2019-08-15 Tom Tromey <tromey@adacore.com>
Tom Tromey [Wed, 14 Aug 2019 15:11:28 +0000 (09:11 -0600)]
Simplify write_inferior_memory
gdbserver's write_inferior_memory uses a static variable to avoid
memory leaks, and has a comment referring to the lack of cleanups.
This patch removes this comment and the code in favor of a
straightforward use of std::vector.
gdb/gdbserver/ChangeLog
2019-08-15 Tom Tromey <tromey@adacore.com>
* target.c (write_inferior_memory): Use std::vector.
Tom Tromey [Wed, 14 Aug 2019 20:01:57 +0000 (14:01 -0600)]
Fix CU overrun in scan_unit_for_symbols
A customer program had a DWARF CU that consisted of just a CU DIE,
without any children. In this situation, scan_unit_for_symbols will
try to read past the end of the current CU, and will take use the
first bytes of the next CU as an abbrev, printing an error message.
This patch fixes the bug by changing scan_unit_for_symbols to stop at
the end of the CU rather than the end of the .debug_info section.
bfd/ChangeLog
2019-08-15 Tom Tromey <tromey@adacore.com>
* dwarf2.c (scan_unit_for_symbols): Check for end of CU, not end
of section.
Tom de Vries [Thu, 15 Aug 2019 06:14:31 +0000 (08:14 +0200)]
[gdb] Fix gdb build on mips64-linux
When compiling for mips64-linux, we get:
...
src/gdb/linux-nat-trad.c:139:12: error: ‘gdbarch_num_regs’ was not declared \
in this scope
139 | regnum < gdbarch_num_regs (regcache->arch ());
...
Fix this by including gdbarch.h in linux-nat-trad.c, similar to commit b1c896b365 "Fix gdb build on macOS".
Alan Hayward [Wed, 14 Aug 2019 14:47:05 +0000 (15:47 +0100)]
AArch64: Allow additional sizes in prologue
When saving registers to the stack at the start of a function, not all state
needs to be saved. For example, only the first 64bits of float registers need
saving. However, a program may choose to store extra state if it wishes,
there is nothing preventing it doing so.
The aarch64_analyze_prologue will error if it detects extra state being
stored. Relex this restriction.
Tested via aarch64-prologue test.
gdb/ChangeLog:
* aarch64-tdep.c (aarch64_analyze_prologue): Allow any valid
register sizes.
gdb/testsuite/ChangeLog:
* gdb.arch/aarch64-prologue.c: New test.
* gdb.arch/aarch64-prologue.exp: New file.
Tom Tromey [Wed, 14 Aug 2019 14:40:18 +0000 (08:40 -0600)]
Fix gdb build on macOS
Internal testing showed that the macOS port did not build. The
breakage was caused by the patch to remove the gdbarch.h include from
defs.h. This patch fixes the problem.
gdb/ChangeLog
2019-08-14 Tom Tromey <tromey@adacore.com>
* darwin-nat.c: Include gdbarch.h.
* darwin-nat-info.c: Include gdbarch.h.
Alan Modra [Wed, 14 Aug 2019 01:21:22 +0000 (10:51 +0930)]
PR24623, DWARF errors
PR 24623
* dwarf2.c (stash_comp_unit): New function, extracted from..
(_bfd_dwarf2_find_nearest_line): ..here.
(find_abstract_instance): Parse comp units and decode line info
as needed.
This parameter might appear to be used to set up offset_size, but
since git commit 024b2372f5 offset_size is either set from the
debug_info data or is set to 4.
Tom Tromey [Sat, 6 Jul 2019 14:10:50 +0000 (08:10 -0600)]
Don't track the contents of the execution info window
The curses library keeps track of the contents of each window, and can
redraw the screen as needed. This means that in most cases is no need
for the TUI windows to also keep track of their contents. This patch
removes content tracking from the execution window, leaving that to
curses.
gdb/ChangeLog
2019-08-13 Tom Tromey <tom@tromey.com>
Tom Tromey [Sat, 6 Jul 2019 13:37:39 +0000 (07:37 -0600)]
Remove tui_clear_exec_info_content
After the previous patch, all calls to tui_clear_exec_info_content
come just after a call to tui_clear_source_content. Because these two
windows are linked, I think it makes sense to have
tui_clear_source_content simply do the work. So, this patch removes
tui_clear_exec_info_content.
gdb/ChangeLog
2019-08-13 Tom Tromey <tom@tromey.com>
Tom Tromey [Fri, 5 Jul 2019 23:36:16 +0000 (17:36 -0600)]
Remove tui_erase_exec_info_content
One call to tui_erase_exec_info_content can be removed. This call is
not needed because the function in question then immediately sets the
execution info window contents.
Once this is done, tui_clear_exec_info_content is just a wrapper for
the only call to tui_erase_exec_info_content, so
tui_erase_exec_info_content can be renamed and the wrapper function
removed.
gdb/ChangeLog
2019-08-13 Tom Tromey <tom@tromey.com>
Tom Tromey [Fri, 5 Jul 2019 23:35:13 +0000 (17:35 -0600)]
Turn tui_show_exec_info_content into a method
This changes tui_show_exec_info_content to be a method on
tui_source_window_base. As it is only called by other methods on this
class, it can be private.
gdb/ChangeLog
2019-08-13 Tom Tromey <tom@tromey.com>
Tom Tromey [Fri, 5 Jul 2019 23:32:19 +0000 (17:32 -0600)]
Move code to tui-winsource.h
This moves code related to the execution info window from tui-data.h
to tui-winsource.h. It fits better here because the execution info is
conceptually part of the source and disassembly windows, and
tui-winsource.h is where this common class lives.
gdb/ChangeLog
2019-08-13 Tom Tromey <tom@tromey.com>
Tom Tromey [Fri, 5 Jul 2019 23:14:46 +0000 (17:14 -0600)]
Change tui_check_and_display_highlight_if_needed to be a method
This changes tui_check_and_display_highlight_if_needed to be a method
on tui_win_info. This makes it clear that the NULL check in that
function is not needed, so it is removed here.
gdb/ChangeLog
2019-08-13 Tom Tromey <tom@tromey.com>
Tom Tromey [Fri, 5 Jul 2019 18:46:23 +0000 (12:46 -0600)]
Delete invisible TUI windows
This changes the TUI so that when the layout changes, any windows that
are invisible are now deleted. This makes it simpler to understand
window lifetimes.
gdb/ChangeLog
2019-08-13 Tom Tromey <tom@tromey.com>
Tom Tromey [Fri, 5 Jul 2019 18:34:31 +0000 (12:34 -0600)]
Two simplifications in tui-layout.c
This patch simplifies some code in tui-layout.c.
In show_layout, all the layout settings can be handled by a single
switch statement. In show_source_disasm_command and
show_source_or_disasm_and_command, there is no need to check the
current layout, as the caller has already done so.
gdb/ChangeLog
2019-08-13 Tom Tromey <tom@tromey.com>
* tui/tui-layout.c (show_layout): Unify all layout cases into a
single switch.
(show_source_disasm_command, show_source_or_disasm_and_command):
Don't check current layout.
Tom Tromey [Fri, 5 Jul 2019 18:29:18 +0000 (12:29 -0600)]
Simplify tui_make_all_invisible
This simplifies the implementation of tui_make_all_invisible. Also,
because show_data is only called by show_layout, this hoists the call
to tui_make_all_invisible and removes the call from show_data.
gdb/ChangeLog
2019-08-13 Tom Tromey <tom@tromey.com>
Tom Tromey [Fri, 5 Jul 2019 18:13:40 +0000 (12:13 -0600)]
Move current_layout to tui-layout.c
This moves the current_layout global to tui-layout.c. This allows for
the removal of an accessor function; but also it just seems clearer to
have it here.
gdb/ChangeLog
2019-08-13 Tom Tromey <tom@tromey.com>
* tui/tui-layout.c (current_layout, tui_current_layout): Move from
tui-data.c.
(show_source_disasm_command, show_data)
(show_source_or_disasm_and_command): Don't use
tui_set_current_layout_to.
* tui/tui-data.h (tui_set_current_layout_to): Don't declare.
* tui/tui-data.c (current_layout, tui_current_layout): Move to
tui-layout.c.
(tui_set_current_layout_to): Remove.
Tom Tromey [Fri, 5 Jul 2019 18:00:06 +0000 (12:00 -0600)]
clear_detail can only be called on TUI source windows
The clear_detail method can only be called on source windows, so
remove definitions from the base of the class hierarchy, leaving only
a single non-virtual method.
gdb/ChangeLog
2019-08-13 Tom Tromey <tom@tromey.com>
Tom Tromey [Mon, 12 Aug 2019 18:06:20 +0000 (12:06 -0600)]
Don't include readline headers from tracepoint.c
I was curious why updating readline caused so much to be to rebuilt,
so I look at all the uses of the readline headers. Most are included
for valid reasons (either readline directory or for tilde_expand); but
the includes in tracepoint.c didn't seem to be used.
This patch removes them. Tested by rebuilding.
gdb/ChangeLog
2019-08-13 Tom Tromey <tromey@adacore.com>
* tracepoint.c: Don't include readline.h or history.h.
Tom de Vries [Tue, 13 Aug 2019 16:57:50 +0000 (18:57 +0200)]
[gdb/testsuite] Fix gdb.gdb/selftest.exp regexp
With gdb.gdb/selftest.exp, we get:
...
(xgdb) PASS: gdb.gdb/selftest.exp: send SIGINT signal to child process
^M
Thread 1 "xgdb" received signal SIGINT, Interrupt.^M
0x00007ffff5bf01db in poll () from /lib64/libc.so.6^M
(gdb) FAIL: gdb.gdb/selftest.exp: send ^C to child process again
...
The failure is due to gdb printing 'Thread 1 "xgdb" received signal SIGINT',
but the regexp in the test-case expecting 'Program received signal SIGINT'.
Fix this by updating the regexp, similar to how that is done earlier in the
test-case.
gdb/testsuite/ChangeLog:
2019-08-13 Tom de Vries <tdevries@suse.de>
* gdb.gdb/selftest.exp (send ^C to child process again): Accept also
Thread.
Sandra Loosemore [Tue, 13 Aug 2019 16:09:30 +0000 (09:09 -0700)]
Fixes for gdb.python tests on remote Windows host.
This patch fixes several test ERRORs and FAILs seen from running
gdb.python tests on a remote Windows host. The problems fixed
generally fall into these categories:
- Failure to copy the .py script to the host.
- Confusion between build and host pathnames.
- Assuming pathnames printed on the host include "/" as a directory
separator.
- Tests that need to be conditionally disabled due to missing features
on the host, etc.
gdb/testsuite/
* gdb.python/py-completion.exp: Download the .py file to the host
and use its host pathname. Conditionalize tests that use
tab completion and manipulate files on the build machine.
* gdb.python/py-events.exp: Download the .py file to the host
and use its host pathname.
* gdb.python/py-evsignal.exp: Likewise.
* gdb.python/py-evthreads.exp: Likewise.
* gdb.python/py-framefilter-invalidarg.exp: Match Windows
pathname syntax.
* gdb.python/py-framefilter.exp: Download the .py file to the right
place on the host. Match Windows pathname syntax.
* gdb.python/py-mi-var-info-path-expression.exp: Download the
.py file to the host and use its host pathname.
* gdb.python/py-objfile-script.exp: Match Windows pathname syntax.
* gdb.python/py-objfile.exp: Expect a host pathname, not a
build pathname. Skip symlink test on Windows host. Add missing
newline at end of file.
* gdb.python/py-pp-maint.exp: Download the .py file to the host
and use its host pathname.
* gdb.python/py-pp-registration.exp: Match Windows pathname syntax.
* gdb.python/py-section-script.exp: Use host location of binfile
on safe-path. Use correct path separator on Windows host.
Reorder alternatives in gdb_test_multiple to prevent matching
the wrong alternative on success.
* gdb.python/py-symtab.exp: Match Windows pathname syntax.
Tom Tromey [Sun, 4 Aug 2019 23:39:27 +0000 (17:39 -0600)]
Remove readline hack from gdb_select
As discussed on gdb-patches, this removes the readline hack from the
mingw-hdep.c version of gdb_select. It's believed that this is not
needed any more. See:
Patrick Palka [Sun, 17 Mar 2019 14:32:16 +0000 (08:32 -0600)]
Fix gdb's selftest.exp after readline import
After the sync there is one testsuite regression, the test
"signal SIGINT" in gdb.gdb/selftest.exp which now FAILs. Previously,
the readline 6.2 SIGINT handler would temporarily reinstall the
underlying application's SIGINT handler and immediately re-raise SIGINT
so that the orginal handler gets invoked. But now (since readline 6.3)
its SIGINT handler does not re-raise SIGINT or directly invoke the
original handler; it now sets a flag marking that SIGINT was raised, and
waits until readline explicitly has control to call the application's
SIGINT handler. Anyway, because SIGINT is no longer re-raised from
within readline's SIGINT handler, doing "signal SIGINT" with a stopped
inferior gdb process will no longer resume and then immediately stop the
process (since there is no 2nd SIGINT to immediately catch). Instead,
the inferior gdb process will now just print "Quit" and continue to run.
So with this commit, this particular test case is adjusted to reflect
this change in behavior (we now have to send a 2nd SIGINT manually to
stop it).
gdb/testsuite/ChangeLog
2019-08-12 Patrick Palka <patrick@parcs.ath.cx>
* gdb.gdb/selftest.exp (test_with_self): Update test to now
expect the GDB inferior to no longer immediately stop after
being resumed with "signal SIGINT".
Tom Tromey [Sun, 7 Oct 2018 21:24:12 +0000 (15:24 -0600)]
Remove gdb workaround from readline/complete.c
This removes a gdb-local patch from readline's get_y_or_n. The code
references a gdb test that continues to work when I remove this patch.
So, I think it is not needed any more.
readline/ChangeLog.gdb
2018-10-07 Tom Tromey <tom@tromey.com>
Tom Tromey [Sun, 7 Oct 2018 18:52:13 +0000 (12:52 -0600)]
Import readline 7.0 (patch 5)
This imports readline 7.0 (up to patch 5) while preserving all
gdb-local changes.
This was done by checking out the readline git repository, making a
branch based on the gdb baseline revision, applying the gdb changes to
that branch, and then merging from readline 7.
readline/ChangeLog.gdb
2019-08-12 Tom Tromey <tom@tromey.com>
Modify the ARM encoding and decoding of SQRSHRL and UQRSHLL MVE instructions.
This is a change to the first published specifications [1][a] but since there is no hardware
out there that uses the old instructions we do not want to support the old variant.
This changes are done based on the latest published specifications [1][b].
[1] https://developer.arm.com/architectures/cpu-architecture/m-profile/docs/ddi0553/latest/armv81-m-architecture-reference-manual
[a] version bf
[b] version bh
gas * config/tc-arm.c (enum operand_parse_code): Add the entry OP_I48_I64.
(po_imm1_or_imm2_or_fail): Marco to check the immediate is either of
48 or 64.
(parse_operands): Add case OP_I48_I64.
(do_mve_scalar_shift1): Add function to encode the MVE shift
instructions with 4 arguments.
* testsuite/gas/arm/mve-shift-bad.l: Modify.
* testsuite/gas/arm/mve-shift-bad.s: Likewise.
* testsuite/gas/arm/mve-shift.d: Likewise.
* testsuite/gas/arm/mve-shift.s: Likewise.
opcodes * arm-dis.c (struct mopcode32 mve_opcodes): Modify the mask for
cases MVE_SQRSHRL and MVE_UQRSHLL.
(print_insn_mve): Add case for specifier 'k' to check
specific bit of the instruction.
Barnaby Wilks [Mon, 12 Aug 2019 10:08:36 +0000 (11:08 +0100)]
Add generic and ARM specific support for half-precision IEEE 754 floating point numbers to the assembler.
Half precision floating point numbers will be encoded using the IEEE 754
half precision floating point format - 16 bits in total, 1 for sign, 5
for exponent and 10 bits of mantissa.
This patch implements the float16 directive for both the IEEE 754 format
and the Arm alternative format for the Arm backend.
The Arm alternative format is almost identical to the IEEE 754 format,
except that it doesn't encode for NaNs or Infinity (instead an exponent
of 0x1F represents a normalized number in the range 65536 to 131008).
The alternative format is documented in the reference manual:
Which format is used is controlled by the .float16_format <format>
directive, where if <format> = ieee, then use the IEEE 754
half-precision format else if <format> = alternative, then use the
Arm alternative format
Or the format can be set on the command line via the -mfp16-format
option that has a similar syntax. -mfp16-format=<ieee|alternative>.
This also fixes the format and it cannot be changed by any directives.
Once the format has been set (either by the command line option or a directive) it cannot be changed,
and any attempts to change it (i.e. with the float16_format directive) will result in a warning and the
line being ignored.
For ELF targets the appropriate EABI attribute will be written out at the end of assembling
if the format has been explicitly specified. If no format has been explicitly specified then no
EABI attributes will be written.
If the format is not explicitly specified then any float16 directives are encoding using the IEEE 754-2008
format by default until the format is fixed or changed with the float16_format directive.
gas * config/tc-arm.c (enum fp_16bit_format): Add enum to represent the 2 float16 encodings.
(md_atof): Set precision for float16 type.
(arm_is_largest_exponent_ok): Check for whether to encode with the IEEE or alternative
format.
(set_fp16_format): Parse a float16_format directive.
(arm_parse_fp16_opt): Parse the fp16-format command line option.
(aeabi_set_public_attributes): For ELF encode the FP16 format EABI attribute.
* config/tc-arm.h (TC_LARGEST_EXPONENT_IS_NORMAL): Macro that expands to
arm_is_largest_exponent_ok.
(arm_is_largest_exponent_ok): Add prototype for arm_is_largest_exponent_ok function.
* doc/c-arm.texi: Add documentation for .float16, .float16_format and -mfp16-format=
* testsuite/gas/arm/float16-bad.d: New test.
* testsuite/gas/arm/float16-bad.l: New test.
* testsuite/gas/arm/float16-bad.s: New test.
* testsuite/gas/arm/float16-be.d: New test.
* testsuite/gas/arm/float16-format-bad.d: New test.
* testsuite/gas/arm/float16-format-bad.l: New test.
* testsuite/gas/arm/float16-format-bad.s: New test.
* testsuite/gas/arm/float16-format-opt-bad.d: New test.
* testsuite/gas/arm/float16-format-opt-bad.l: New test.
* testsuite/gas/arm/float16-le.d: New test.
* testsuite/gas/arm/float16.s: New test.
* testsuite/gas/arm/float16-eabi-alternative-format.d: New test.
* testsuite/gas/arm/float16-eabi-ieee-format.d: New test.
* testsuite/gas/arm/float16-eabi-no-format.d: New test.
* testsuite/gas/arm/float16-eabi.s: New test.
* config/atof-ieee.c (H_PRECISION): Macro for precision of float16
type.
(atof_ieee): Set precision and exponent bits for encoding float16
types.
(gen_to_words): NaN and Infinity encoding for float16.
(ieee_md_atof): Set precision for encoding float16 type.
Alan Modra [Sat, 10 Aug 2019 01:00:19 +0000 (10:30 +0930)]
lang_input_statement_type next pointers
"next" and "next_real_file" in lang_input_statement_type always point
to another lang_input_statement_type, so it makes sense for these to
not be the generic lang_statement_union_type. This patch also updates
a number of variables in ldlang.c for the same reason, and modifies
lang_statement_append to reduce the need for casts.
* ldlang.h (lang_input_statement_type): Make next
and next_real_file a lang_input_statement_type pointer.
(lang_statement_append): Delete prototype.
(LANG_FOR_EACH_INPUT_STATEMENT): Update for lang_input_statement_type
change.
* ldmain.c (add_archive_element): Likewise.
* ldlang.c: Likewise throughout.
(lang_statement_append): Make static. Make element and field
void pointers. Remove casts in calls.
(lang_check): Use a lang_input_statement_type pointer for "file".
(find_rescan_insertion): Similarly for "iter" and return value.
(lang_process): Similarly for "insert", "iter" and "temp".
* emultempl/spuelf.em (embedded_spu_file): Likewise.
* emultempl/aix.em (gld${EMULATION_NAME}_before_allocation): Expand
lang_statment_append call.
Fix access to uninitialized variable in fill_in_stop_func
This patch changes find_pc_partial_function so that *block is set to
nullptr when it fails, so that fill_in_stop_func won't access an
uninitialized variable.
gdb/ChangeLog:
2019-08-09 Pedro Franco de Carvalho <pedromfc@linux.ibm.com>
* blockframe.c (find_pc_partial_function): Set *block to nullptr
when the function fails.
Andreas Arnez [Fri, 9 Aug 2019 18:27:03 +0000 (20:27 +0200)]
s390: Implement 'type_align' gdbarch method
The align.exp test case yields many FAILs on s390x, since GDB's _Alignoff
doesn't always agree with the compiler's. On s390x, the maximum alignment
is 8, but GDB returns an alignment of 16 for 16-byte data types such as
"long double".
This is fixed by implementing the type_align gdbarch method. The new
method returns an alignment of 8 for all integer, floating-point, and
vector types larger than 8 bytes. With this change, all align.exp tests
pass.
gdb/ChangeLog:
* s390-tdep.c (s390_type_align): New function.
(s390_gdbarch_init): Set it as type_align gdbarch method.
* cp-demangle (d_print_comp_inner): Guard against a NULL 'typed_name'.
* testsuite/demangle-expected: Add testcase.
* cp-demangle.c (d_encoding): Guard against NULL return values from
d_right (dc).
* testsuite/demangle-expected: Add testcase.
2019-04-29 Ben L <bobsayshilol@live.co.uk>
* cp-demangle.c (d_expression_1): Don't peek ahead unless the current
char is valid.
* testsuite/demangle-expected: Add testcase.
2019-04-10 Nick Clifton <nickc@redhat.com>
PR 89394
* cp-demangle.c (cplus_demangle_fill_name): Reject negative
lengths.
(d_count_templates_scopes): Replace num_templates and num_scopes
parameters with a struct d_print_info pointer parameter. Adjust
body of the function accordingly. Add recursion counter and check
that the recursion limit is not reached.
(d_print_init): Pass dpi parameter to d_count_templates_scopes.
Reset recursion counter afterwards, unless the recursion limit was
reached.
Add support for a MIPS specific .MIPS.xhash section.
This patch is a reimplementation of [1] which was submitted in 2015 by
Neil Schellenberger. Copyright issue was sorted out [2] last year.
It proposed a new section (.gnu.xhash) and related dynamic tag
(DT_GNU_XHASH). The new section would be virtually identical to the
existing .gnu.hash except for the translation table (xlat) which would
contain correct MIPS .dynsym indexes corresponding to the hashvals in
chains. This is because MIPS ABI imposes a different ordering on the
dynsyms than the one expected by the .gnu.hash section. Another addition
would be a leading word (ngnusyms) which would contain the number of
entries in the translation table.
In this patch, the new section name and dynamic tag are changed to
reflect the fact that the section should be treated as MIPS-specific
(.MIPS.xhash and DT_MIPS_XHASH).
This patch addresses the alignment issue as reported in [3], which is
caused by the leading word added to the .MIPS.xhash section. Leading word
is removed in this patch, and the number of entries in the translation
table is now calculated using DT_MIPS_SYMTABNO dynamic tag (this is
addressed by the corresponding glibc patch).
Suggestions on coding style in [4] were taken into account. Existing
GNU hash testcase was covered, and another one was added in the MIPS
part of the testsuite.
The other major change is reserving MIPS ABI version 5 for .MIPS.xhash,
marking the need of support for .MIPS.xhash in the dynamic linker (again,
addressed in the corresponding glibc patch). This is something which I
am not sure of, especially after reading [5]. I am confused on whether
this ABI version is reserved for IFUNC, or it can be used for this
purpose.
Already mentioned glibc patch is submitted at:
https://sourceware.org/ml/libc-alpha/2019-06/msg00456.html
bfd * elf-bfd.h (struct elf_backend_data): New members.
* elflink.c (_bfd_elf_link_create_dynamic_sections): Create
.gnu.hash section if necessary.
(struct collect_gnu_hash_codes): New member.
(elf_gnu_hash_process_symidx): New function name.
(elf_renumber_gnu_hash_syms): Ignore local and undefined
symbols. Record xlat location for every symbol which should have
a .MIPS.xhash entry.
(bfd_elf_size_dynamic_sections): Add DT_GNU_HASH dynamic tag to
dynamic section if necessary.
(GNU_HASH_SECTION_NAME): New define.
(bfd_elf_size_dynsym_hash_dynstr): Get .MIPS.xhash section.
Update the section size info.
* elfxx-mips.c (struct mips_elf_hash_sort_data): New members.
(struct mips_elf_link_hash_entry): New member.
(mips_elf_link_hash_newfunc): Initialize .MIPS.xhash translation
table location.
(mips_elf_sort_hash_table): Initialize the pointer to the
.MIPS.xhash section.
(mips_elf_sort_hash_table_f): Populate the .MIPS.xhash
translation table entry with the symbol dynindx.
(_bfd_mips_elf_section_from_shdr): Add SHT_MIPS_XHASH.
(_bfd_mips_elf_fake_sections): Initialize .MIPS.xhash section
info.
(_bfd_mips_elf_create_dynamic_sections): Create .MIPS.xhash
section.
(_bfd_mips_elf_size_dynamic_sections): Add DT_MIPS_XHASH tag to
dynamic section.
(_bfd_mips_elf_finish_synamic_sections): Add DT_MIPS_XHASH.
(_bfd_mips_elf_final_write_processing): Set .MIPS.xhash section
sh_link info.
(_bfd_mips_elf_get_target_dtag): Get DT_MIPS_XHASH tag.
(MIPS_LIBC_ABI_XHASH): New ABI version enum value.
(_bfd_mips_post_process_headers): Mark the ABI version as
MIPS_LIBC_ABI_XHASH if there exists a .MIPS.xhash section,
but not a .hash section.
(_bfd_mips_elf_record_xhash_symbol): New function. Record a
position in the translation table, associated with the hash
entry.
* elfxx-mips.h (literal_reloc_p): Define
elf_backend_record_xhash_symbol backend hook.
* elfxx-target.h: Initialize elf_backend_record_xhash_symbol
backend hook.
include * elf/mips.h (SHT_GNU_XHASH): New define.
(DT_GNU_XHASH): New define.
binutils * readelf.c (get_mips_dynamic_type): Return MIPS_XHASH dynamic type.
(get_mips_section_type_name): Return MI{S_XHASH name string.
(dynamic_section_mips_val): Initialize the .MIPS.xhash dynamic
info.
(process_symbol_table): Initialize the .MIPS.xhash section
pointer. Adjust the readelf output to support the new section.
(process_object): Set the .MIPS.xhash dynamic info to zero.
Jan Beulich [Fri, 9 Aug 2019 07:57:57 +0000 (09:57 +0200)]
x86-64: generalize SIMD test expectations
In order to be able to add/remove insns to/from the middle of these
tests, generalize the patterns for the symbol reference comments of RIP-
relative operands.
Tom de Vries [Fri, 9 Aug 2019 04:49:04 +0000 (06:49 +0200)]
[gdb] Fix gdb.dwarf2/varval.exp with -fPIE/-pie
With target board unix/-fPIE/-pie, we get:
...
FAIL: gdb.dwarf2/varval.exp: print varval2
...
This is due comparing a get_frame_pc result (which includes the for PIE
non-zero relocation offset) with pc_high and pc_low obtained using
get_scope_pc_bounds (which do not include the relocation offset).
Fix this by adjusting pc_high and pc_low with the relocation offset.
Tested on x86_64-linux with target board unix/-fPIE/-pie.
gdb/ChangeLog:
2019-08-09 Tom de Vries <tdevries@suse.de>
PR gdb/24591
* dwarf2read.c (dwarf2_fetch_die_loc_sect_off): Adjust pc_high and
pc_low with relocation offset.
Tom de Vries [Thu, 8 Aug 2019 20:26:28 +0000 (22:26 +0200)]
[gdb/testsuite] Fix gdb.tui/basic.exp with check-read1
With gdb.tui/basic.exp and check-read1, we run into (using -v for
verbose log):
...
^[[0+++ _csi_0 <<<>>>
ERROR: (DejaGnu) proc "_csi_0" does not exist.
...
In contrast, without check-read1, we have:
...
^[[0;10m<SNIP>+++ _csi_m <<<0;10>>>
...
The problem is that this regexp in _accept:
...
-re "^\x1b\\\[(\[0-9;\]*)(\[0-9a-zA-Z@\])" {
...
while matching the longer sequence '^[' '[' '0' ';' '1' '0' 'm', also matches
the shorter sequence '^[' '[' '0'.
The regexp attempts to match a CSI (Control Sequence Introducer) sequence, and
the final byte of such a sequence cannot be a digit.
gas * testsuite/gas/i386/property-1.d: Adjust for new output format
from readelf.
* testsuite/gas/i386/property-2.d: Likewise.
* testsuite/gas/i386/x86-64-property-1.d: Likewise.
* testsuite/gas/i386/x86-64-property-2.d: Likewise.
When printing unknown note types, readelf prints the raw description
section byte-by-byte. However, it does not mask appropriately, e.g. it
prints the byte 'ba' as 'ffffffba'.
* readelf.c (process_note): Mask unknown description data bytes.
Prevent objdump from aborting when asked to disassemble an unknown type of ARC binary file.
PR 24854
* arc-dis.c (arc_insn_length): Return 0 rather than aborting when
encountering an unknown machine type.
(print_insn_arc): Handle arc_insn_length returning 0. In error
cases return -1 rather than calling abort.
Alan Modra [Wed, 7 Aug 2019 14:07:49 +0000 (23:37 +0930)]
Integer overflows in readelf get_data
I noticed the test for overflow of amt = size * nmemb in get_data
wasn't effective. An obvious example of nmemb = 3 and size = half max
value overflows but doesn't result in amt < nmemb. This patch fixes
this problem and reports a size truncation or overflow rather than out
of memory in more cases.
Nick Clifton [Wed, 7 Aug 2019 14:15:59 +0000 (15:15 +0100)]
Improve the formatting of the title strings of the binutils manual pages.
PR 24777
* doc/binutils.texi: Ensure consistent formating of title strings
for man pages. Extend the title of the size man page to be more
informative.
Alan Hayward [Wed, 7 Aug 2019 08:47:57 +0000 (09:47 +0100)]
AArch64 pauth: Indicate unmasked addresses in backtrace
Armv8.3-a Pointer Authentication causes the function return address to be
obfuscated on entry to some functions. GDB must unmask the link register in
order to produce a backtrace.
The following patch adds markers of [PAC] to the bracktrace, to indicate
which addresses needed unmasking. This includes the backtrace when using MI.
For example, consider the following backtrace:
(gdb) bt
0 0x0000000000400490 in puts@plt ()
1 0x00000000004005dc in foo ("hello") at cbreak-lib.c:6
2 0x0000000000400604 [PAC] in bar () at cbreak-lib.c:12
3 0x0000000000400620 [PAC] in main2 () at cbreak.c:17
4 0x00000000004005b4 in main () at cbreak-3.c:10
The functions in cbreak-lib use pointer auth, which masks the return address
to the previous function, causing the addresses of bar (in the library) and main2
(in the main binary) to require unmasking in order to unwind the backtrace.
An extra bool is added alongside the prev_pc in the frame structure. At the
point at which the link register is unmasked, the AArch64 port calls into frame
to sets the bool. This is the most efficient way of doing it.
The marker is also added to the python frame printer, which is always printed if
set. The marker is not explicitly exposed to the python code.
I expect this will potentially cause issues with some tests in the testsuite
when Armv8.3 pointer authentication is used. This should be fixed up in the
the future once real hardware is available for full testsuite testing.
gdb/ChangeLog:
* NEWS: Expand the Pointer Authentication entry.
* aarch64-tdep.c (aarch64_frame_unmask_address): Rename from this.
(aarch64_frame_unmask_lr): ... to this.
(aarch64_prologue_prev_register, aarch64_dwarf2_prev_register):
Call aarch64_frame_unmask_lr.
* frame.c (struct frame_info): Add "masked" variable.
(frame_set_previous_pc_masked) (frame_get_pc_masked): New functions.
(fprint_frame): Check for masked pc.
* frame.h (frame_set_previous_pc_masked) (frame_get_pc_masked): New
declarations.
* python/py-framefilter.c (py_print_frame): Check for masked pc.
* stack.c (print_frame): Check for masked pc.
gdb/doc/ChangeLog:
* gdb.texinfo (AArch64 Pointer Authentication): New subsection.
This patch makes the elf64bpf emulation to use it's own linker script,
based on elf.sc. At the moment, the only change is that the BPF
executable doesn't define an entry symbol (BPF programs feature
several entry points scattered in several sections.)
This is a step towards the goal of generating proper ELF executables
that would be loaded by the kernel's libbpf. We are not there yet:
BPF "programs" should still be linked with -r.
This change removes a warning while linking executables, decreases the
number of unsupported tests in the target from 47 to 29, and increases
the number of expected passes from 104 to 145.
Regtested in x86_64 for all targets.
No regressions.
ld/ChangeLog:
2019-08-07 Jose E. Marchesi <jose.marchesi@oracle.com>
* scripttempl/elf64bpf.sc: Adapted from elf.sc.
* emulparams/elf64bpf.sh (SCRIPT_NAME): Use elf64bpf.
(EMBEDDED): Define.
* testsuite/ld-bpf/call-1.d: Do not expect a warning regarding an
undefined entry symbol.
* testsuite/ld-bpf/jump-1.d: Likewise.
* testsuite/ld-undefined/undefined.exp: Do not pass '-e entry' to
ld in BPF targets, and do not expect line number information.
* testsuite/ld-srec/srec.exp (run_srec_test): xfail s-record tests
in BPF targets.
Omair Majid [Wed, 7 Aug 2019 09:20:56 +0000 (10:20 +0100)]
Fix the binutils test for .NET assembly support so that it distinguishing between targets which do not support the x86 PE format at all, and those that do support it, but which do not recognise the .NET assembly variants.
* testsuite/binutils-all/objdump.exp
(test_objdump_dotnet_assemblies): Fix test to distinguish errors
in parsing simple pei-i386 and pei-x86-64 vs parsing the newly
introduced machine types.
* testsuite/gentestdlls.c (write_simple_dll): New function.
(main): Generate simple and Linux-specific variants of pei-i386
and pei-x86-64 files so both can be used by tests.
Jan Beulich [Wed, 7 Aug 2019 08:46:52 +0000 (10:46 +0200)]
x86: drop stray FloatMF
The flag is supposed to be used in templates which allow for both a
"short" and a "long" format memory operand. Drop it from templates not
matching this pattern. In the control/status word cases it was (ab)used
in place of the intended IgnoreSize.
Tom Tromey [Sat, 13 Jul 2019 18:13:15 +0000 (12:13 -0600)]
Introduce obstack_strndup
This introduces obstack_strndup and changes gdb to use it.
Note that obstack_strndup works like savestring, and not exactly like
xstrndup. The difference is that obstack_strndup uses the passed-in
length, while xstrndup uses strnlen to choose the length.
gdb/ChangeLog
2019-08-06 Tom Tromey <tom@tromey.com>
* stabsread.c (patch_block_stabs, read_one_struct_field)
(read_enum_type): Use obstack_strndup.
* rust-exp.y (rust_parser::copy_name): Use obstack_strndup.
* gdb_obstack.h (obstack_strndup): Use obstack_strndup.
* dwarf2read.c (guess_full_die_structure_name)
(anonymous_struct_prefix): Use obstack_strndup.
* dbxread.c (cp_set_block_scope): Use obstack_strndup.
* c-exp.y (yylex): Use obstack_strndup.
* ada-exp.y (write_object_renaming, write_ambiguous_var)
(write_var_or_type): Use obstack_strndup.
Tom Tromey [Sat, 13 Jul 2019 18:03:07 +0000 (12:03 -0600)]
Make obstack_strdup inline
This changes obstack_strdup to be an inline function. This seems
better to me, considering how small it is; but also it follows what
the code did before the previous patch.
gdb/ChangeLog
2019-08-06 Tom Tromey <tom@tromey.com>
Add a selftest that checks documentation invariants.
Several approaches were discussed (mail or irc) to verify the invariants of
the GDB help documentation : checking with apropos ., modifying add_cmd
to do the check and output a warning, implement maintenance check-doc.
A selftest was finally chosen as:
* this can be run on demand, including by users if they want
to check user defined commands.
* it does not interact with the normal behaviour of apropos, define,
python, ...
(such as output warnings when a user defines a command help that
does not respect the doc).
* when the selftest runs, it checks the user defined and python
defined commands currently defined.
gdb/ChangeLog
* unittests/help-doc-selftests.c: New file.
* Makefile.in: Add the new file.
Make first and last lines of 'command help documentation' consistent.
With this patch, the help docs now respect 2 invariants:
* The first line of a command help is terminated by a '.' character.
* The last character of a command help is not a newline character.
Note that the changes for the last invariant were done by Tom, as part of :
[PATCH] Remove trailing newlines from help text
https://sourceware.org/ml/gdb-patches/2019-06/msg00050.html
but some occurrences have been re-introduced since then.
Some help docs had to be rephrased/restructured to respect the above
invariants.
Before this patch, print_doc_line was printing the first line
of a command help documentation, but stopping at the first '.'
or ',' character.
This was giving inconsistent results :
* The first line of command helps was sometimes '.' terminated,
sometimes not.
* The first line of command helps was not always designed to be
readable/understandable/unambiguous when stopping at the first
'.' or ',' character.
This e.g. created the following inconsistencies/problems:
< catch exception -- Catch Ada exceptions
< catch handlers -- Catch Ada exceptions
< catch syscall -- Catch system calls by their names
< down-silently -- Same as the `down' command
while the new help is:
> catch exception -- Catch Ada exceptions, when raised.
> catch handlers -- Catch Ada exceptions, when handled.
> catch syscall -- Catch system calls by their names, groups and/or numbers.
> down-silently -- Same as the `down' command, but does not print anything.
Also, the command help doc should not be terminated by a newline
character, but this was not respected by all commands.
The cli-option -OPT framework re-introduced some occurences.
So, the -OPT build help framework was changed to not output newlines at the
end of %OPTIONS% replacement.
This patch changes the help documentations to ensure the 2 invariants
given above.
It implied to slightly rephrase or restructure some help docs.
Based on the above invariants, print_doc_line (called by
'apropos' and 'help' commands to print the first line of a command
help) now outputs the full first line of a command help.
This all results in a lot of small changes in the produced help docs.
There are less code changes than changes in the help docs, as a lot
of docs are produced by some code (e.g. the remote packet usage settings).
gdb/ChangeLog
2019-08-07 Philippe Waroquiers <philippe.waroquiers@skynet.be>
* cli/cli-decode.h (print_doc_line): Add for_value_prefix argument.
* cli/cli-decode.c (print_doc_line): Likewise. It now prints
the full first line, except when FOR_VALUE_PREFIX. In this case,
the trailing '.' is not output, and the first character is uppercased.
(print_help_for_command): Update call to print_doc_line.
(print_doc_of_command): Likewise.
* cli/cli-setshow.c (deprecated_show_value_hack): Likewise.
* cli/cli-option.c (append_indented_doc): Do not append newline.
(build_help_option): Append newline after first appended_indented_doc
only if a second call is done.
(build_help): Append 2 new lines before each option, except the first
one.
* compile/compile.c (_initialize_compile): Add new lines after
%OPTIONS%, when not at the end of the help.
Change help doc or code
producing the help doc to respect the invariants.
* maint-test-options.c (_initialize_maint_test_options): Likewise.
Also removed the new line after 'Options:', as all other commands
do not put an empty line between 'Options:' and the first option.
* printcmd.c (_initialize_printcmd): Likewise.
* stack.c (_initialize_stack): Likewise.
* interps.c (interpreter_exec_cmd): Fix "Usage:" line that was
incorrectly telling COMMAND is optional.
* ada-lang.c (_initialize_ada_language): Change help doc or code
producing the help doc to respect the invariants.
* ada-tasks.c (_initialize_ada_tasks): Likewise.
* breakpoint.c (_initialize_breakpoint): Likewise.
* cli/cli-cmds.c (_initialize_cli_cmds): Likewise.
* cli/cli-logging.c (_initialize_cli_logging): Likewise.
* cli/cli-setshow.c (_initialize_cli_setshow): Likewise.
* cli/cli-style.c (cli_style_option::add_setshow_commands,
_initialize_cli_style): Likewise.
* corelow.c (core_target_info): Likewise.
* dwarf-index-cache.c (_initialize_index_cache): Likewise.
* dwarf2read.c (_initialize_dwarf2_read): Likewise.
* filesystem.c (_initialize_filesystem): Likewise.
* frame.c (_initialize_frame): Likewise.
* gnu-nat.c (add_task_commands): Likewise.
* infcall.c (_initialize_infcall): Likewise.
* infcmd.c (_initialize_infcmd): Likewise.
* interps.c (_initialize_interpreter): Likewise.
* language.c (_initialize_language): Likewise.
* linux-fork.c (_initialize_linux_fork): Likewise.
* maint-test-settings.c (_initialize_maint_test_settings): Likewise.
* maint.c (_initialize_maint_cmds): Likewise.
* memattr.c (_initialize_mem): Likewise.
* printcmd.c (_initialize_printcmd): Likewise.
* python/lib/gdb/function/strfns.py (_MemEq, _StrLen, _StrEq,
_RegEx): Likewise.
* ravenscar-thread.c (_initialize_ravenscar): Likewise.
* record-btrace.c (_initialize_record_btrace): Likewise.
* record-full.c (_initialize_record_full): Likewise.
* record.c (_initialize_record): Likewise.
* regcache-dump.c (_initialize_regcache_dump): Likewise.
* regcache.c (_initialize_regcache): Likewise.
* remote.c (add_packet_config_cmd, init_remote_threadtests,
_initialize_remote): Likewise.
* ser-tcp.c (_initialize_ser_tcp): Likewise.
* serial.c (_initialize_serial): Likewise.
* skip.c (_initialize_step_skip): Likewise.
* source.c (_initialize_source): Likewise.
* stack.c (_initialize_stack): Likewise.
* symfile.c (_initialize_symfile): Likewise.
* symtab.c (_initialize_symtab): Likewise.
* target-descriptions.c (_initialize_target_descriptions): Likewise.
* top.c (init_main): Likewise.
* tracefile-tfile.c (tfile_target_info): Likewise.
* tracepoint.c (_initialize_tracepoint): Likewise.
* tui/tui-win.c (_initialize_tui_win): Likewise.
* utils.c (add_internal_problem_command): Likewise.
* valprint.c (value_print_option_defs): Likewise.
gdb/testsuite/ChangeLog
2019-08-07 Philippe Waroquiers <philippe.waroquiers@skynet.be>
* gdb.base/style.exp: Update tests for help doc new invariants.
* gdb.base/help.exp: Likewise.
This patch drops gdb's configury support for glibc's mcheck function.
It has been observed to cause false abort()s, because it is
thread-unsafe yet interposes every malloc/free operation. So if any
library transitively used by gdb also uses threads, then these
functions can easily corrupt their own checking data. These days, gcc
ASAN and valgrind provide high quality checking, and mcheck is
apparently itself being slowly deprecated.
So, let's stop linking to it. Attached patch drops the
autoconf/Makefile machinery for both gdb and gdbserver. No
testsuite-visible impact. IMHO not worth mentioning in NEWS.
See also: https://sourceware.org/bugzilla/show_bug.cgi?id=9939
gdb/ChangeLog
PR build/24886
* configure.ac: Drop enable-libmcheck support.
* configure, config.in: Rebuild.
* libmcheck.m4: Remove.
* acinclude.m4: Don't include it.
* Makefile.in: Don't distribute it.
* top.c (print_gdb_configuration): Don't mention it.
gdb/gdbserver/ChangeLog
PR build/24886
* configure.ac: Drop enable-libmcheck support.
* configure, config.in: Rebuild.
* acinclude.m4: Don't include it.
Tom Tromey [Fri, 19 Jul 2019 04:07:10 +0000 (22:07 -0600)]
Add more styling to "disassemble"
This adds more styling to the disassemble command. In particular,
addresses and function names in the disassembly are now styled.
This required fixing a small latent bug in set_output_style. This
function always passed NULL to emit_style_escape; but when writing to
a file other than gdb_stdout, it should emit the style escape
directly. (FWIW this is another argument for better integrating the
pager with ui_file and getting rid of this entire layer.)
gdb/ChangeLog
2019-08-06 Tom Tromey <tom@tromey.com>
Tom Tromey [Mon, 22 Jul 2019 21:20:24 +0000 (15:20 -0600)]
Clean up source file error reporting
print_source_lines_base reopens the source file every time that a
source line is to be printed. However, there's no need to do this so
frequently -- it's enough to do it when switching source files, and
otherwise rely on the cache.
The code seems to try to avoid these multiple opens; at a guess I'd
say something just got confused along the way.
This patch fixes the problem by reorganizing the code both to make it
more clear, and to ensure that reopens only occur when the "last
source visited" changes.
gdb/ChangeLog
2019-08-06 Tom Tromey <tromey@adacore.com>
* source.c (last_source_error): Now bool.
(print_source_lines_base): Make "noprint" bool. Only open
source file when last_source_visited changes.
Tom Tromey [Mon, 22 Jul 2019 20:31:43 +0000 (14:31 -0600)]
Add file offsets to the source cache
Currently, gdb stores the number of lines and an array of file offsets
for the start of each line in struct symtab. This patch moves this
information to the source cache. This has two benefits.
First, it allows gdb to read a source file less frequently.
Currently, a source file may be read multiple times: once when
computing the file offsets, once when highlighting, and then pieces
may be read again while printing source lines. With this change, the
file is read once for its source text and file offsets; and then
perhaps read again if it is evicted from the cache.
Second, if multiple symtabs cover the same source file, then this will
share the file offsets between them. I'm not sure whether this
happens in practice.
gdb/ChangeLog
2019-08-06 Tom Tromey <tromey@adacore.com>
* annotate.c (annotate_source_line): Use g_source_cache.
* source-cache.c (source_cache::get_plain_source_lines): Change
parameters. Populate m_offset_cache.
(source_cache::ensure): New method.
(source_cache::get_line_charpos): New method.
(extract_lines): Move lower. Change parameters.
(source_cache::get_source_lines): Move lower.
* source-cache.h (class source_cache): Update comment.
<get_line_charpos>: New method.
<get_source_lines>: Update comment.
<clear>: Clear m_offset_cache.
<get_plain_source_lines>: Change parameters.
<ensure>: New method
<m_offset_cache>: New member.
* source.c (forget_cached_source_info_for_objfile): Update.
(info_source_command): Use g_source_cache.
(find_source_lines, open_source_file_with_line_charpos): Remove.
(print_source_lines_base, search_command_helper): Use g_source_cache.
* source.h (open_source_file_with_line_charpos): Don't declare.
* symtab.h (struct symtab) <nlines, line_charpos>: Remove.
* tui/tui-source.c (tui_source_window::do_scroll_vertical):
Use g_source_cache.
Tom Tromey [Mon, 22 Jul 2019 18:41:23 +0000 (12:41 -0600)]
Save plain text in the source cache
Currently the source cache will only store highlighted text. However,
there's no reason it could not also store plain text, when styling is
turned off.
This patch makes this change. This also simplifies the source cache
code somewhat.
gdb/ChangeLog
2019-08-06 Tom Tromey <tromey@adacore.com>
Tom Tromey [Tue, 23 Jul 2019 14:22:50 +0000 (08:22 -0600)]
Fix latent bug in source cache
The source cache was not returning the final \n of the requested range
of lines. This caused regressions with later patches in this series,
so this patch pre-emptively fixes the bug.
This adds a self-test of "extract_lines" to the source cache code. To
make it simpler to test, I changed extract_lines to be a static
function, and changed it's API a bit.
gdb/ChangeLog
2019-08-06 Tom Tromey <tromey@adacore.com>
* source-cache.c (extract_lines): No longer a method.
Changed type of parameter. Include final newline.
(selftests::extract_lines_test): New function.
(_initialize_source_cache): Likewise.
* source-cache.h (class source_cache)
<extract_lines>: Don't declare.
Alan Modra [Tue, 6 Aug 2019 06:10:08 +0000 (15:40 +0930)]
LTO object insertion point
This is a mostly cosmetic fix for cases like PR24873 where LTO
recompiled objects were supposed to be inserted inside a group. The
specific case handled by this patch is when the first file inside a
group is an archive, the first file claimed by the plugin. Prior to
this patch we would have inserted the recompiled objects before the
group, which doesn't matter really since the entire group will be
reloaded, but it looks a little wrong in map files.
PR 24873
* ldlang.c (find_replacements_insert_point): Return "before" flag.
(find_next_input_statement): New function.
(lang_process): When placing recompiled LTO objects before a
claimed archive, place them immediately before in the statement
list.
Alan Modra [Tue, 6 Aug 2019 05:53:28 +0000 (15:23 +0930)]
PR24873, gcc -flto objects result in --start-group .. --end-group failure to include --as-needed libraries
Reloading of archives (and checking --as-needed DSOs again) is
disabled until we hit the plugin insert point. It's necessary to do
that because in a case like lib1.a lto.o lib2.a where lib1.a and
lib2.a contain duplicate symbols, we want the lto.o recompiled object
to pull in objects from lib2.a as necessary, but not from lib1.a.
Unfortunately this heuristic fails when the insert point is inside a
group, because ld actually loads the symbols from the recompiled
object before running over the contours of the script, thus missing
the fact that new undefs appeared in the group.
PR 24873
* ldlang.c (plugin_undefs): New static var.
(open_input_bfds <lang_group_statement_enum>): Loop on
plugin_undefs and hitting plugin_insert point.
(lang_process <lto_plugin_active>): Set plugin_undefs.
projects / thirdparty / binutils-gdb.git / log
