Fix a bug with commit 4cc0665f24bb ("microMIPS support"),
<https://sourceware.org/ml/gdb-patches/2012-05/msg00724.html>, and add
missing microMIPS SYSCALL instruction decoding needed to determine the
location to put a breakpoint at when single-stepping though a syscall.
BFD/XCOFF: Fix storage class setting for weak defined symbols
Fix an issue with commit 8602d4fea60d ("Add AIX weak support"),
<https://sourceware.org/ml/binutils/2009-03/msg00189.html>, and use the
correct condition to set the storage class for weak defined symbols.
so clearly the inner condition can never be true. Correct the condition
then to check for the `bfd_link_hash_defweak' symbol type instead here,
and in a similar place a little further down in the same function.
bfd/
* xcofflink.c (xcoff_write_global_symbol): Fix symbol type
checks for defined weak symbols.
MIPS/GAS/testsuite: Correct whitespace issues with Loongson tests
Remove CR characters, trailing whitespace and space characters appearing
immediately before a tab character, and replace spaces with tabs, all
across Loongson GAS tests.
... qualified by their respective sizes, allowing to drop FirstXmm0 at
the same time.
folded RegXMM, RegYMM and RegZMM into RegSIMD, it's no longer impossible
to distinguish if Xmmword can represent a memory reference when operand
specification contains SIMD register. For example, template operands
specification like these
RegXMM|...|Xmmword|...
and
RegXMM|...
The Xmmword bitfield is always set by RegXMM which is represented by
"RegSIMD|Xmmword". This patch splits each of vcvtps2qq, vcvtps2uqq,
vcvttps2qq and vcvttps2uqq into 2 templates: one template only has
RegXMM source operand and the other only has mempry source operand.
gas/
PR gas/23418
* testsuite/gas/i386/xmmword.s: Add tests for vcvtps2qq,
vcvtps2uqq, vcvttps2qq and vcvttps2uqq.
* testsuite/gas/i386/xmmword.l: Updated.
Andreas Krebbel [Thu, 21 Dec 2017 12:12:03 +0000 (13:12 +0100)]
S/390: Improve partial relro support for 64 bit
Currently on S/390 the .got.plt always comes first which prevents the
GNU_RELRO segment from being extended across the non-plt GOT entries.
Just swapping both unfortunately is not that simple since our ABI
requires the _GLOBAL_OFFSET_TABLE_ symbol to point to the very
beginning of the entire GOT. Of the 3 magic GOT entries the first is
accessed via got pointer while second and third are being accessed via
DT_PLTGOT. In order to keep them together we make DT_PLTGOT to point
to the .got instead of .got.plt. However, this violates an assumption
in the dynamic linker prelink undo code about the GOTPLT entries
starting at DT_PLTGOT + 3. We got rid of this requirement with a
Glibc patch already in version 2.24:
https://sourceware.org/ml/libc-alpha/2016-06/msg01302.html
So the S/390 relro GOT layout will look like this with this patch:
The patch detects the current layout in size_dynamic_section in order
to deal also with linker scripts not generated by this ld version.
With partial relro enabled we pick a linker script where .got and
.got.plt are swapped which then triggers the rest of the logic.
ld/ChangeLog:
2018-07-18 Andreas Krebbel <krebbel@linux.ibm.com>
* emulparams/elf64_s390.sh: Define GENERATE_RELRO_SCRIPT and
SEPARATE_GOTPLT.
* testsuite/ld-s390/gotreloc_64-relro-1.dd: New test.
* testsuite/ld-s390/gotreloc_64-norelro-1.dd: Renamed from ...
* testsuite/ld-s390/gotreloc_64-1.dd: ... this.
* testsuite/ld-s390/s390.exp: Split the GOT testcase into two.
bfd/ChangeLog:
2018-07-18 Andreas Krebbel <krebbel@linux.ibm.com>
* elf-s390-common.c (s390_gotplt_after_got_p): New function.
(s390_got_pointer): New function.
(s390_got_offset): New function.
(s390_gotplt_offset): New function.
* elf64-s390.c (allocate_dynrelocs): Adjust comment.
(elf_s390_size_dynamic_sections): Move space for magic GOT entries
from .got.plt to .got if necessary and pick the right location for
_GLOBAL_OFFSET_TABLE_.
(elf_s390_relocate_section): Use the wrapper functions from
elf-s390-common.c to deal with both possible layouts (either .got
or .got.plt first).
(elf_s390_finish_dynamic_sections): Likewise.
(elf_s390_finish_dynamic_symbol): Make the location of the GOT
magic entries conditional.
Andreas Krebbel [Mon, 11 Jun 2018 11:23:00 +0000 (13:23 +0200)]
ld: Enable using separate linker script for -z relro
With this patch dedicated linker scripts can be generated for partial
relro triggered by defining GENERATE_RELRO_SCRIPT in the target
specific scripts.
This is necessary for e.g. S/390 where usually the .got.plt comes
first and prevents the relro segment from being extended across the
non-plt GOT entries.
The patch started with the work from Marcin taken from the mwk user
branches. However, the patch needed substantial changes due to the
'separate code' feature which got committed in the meantime.
ld/ChangeLog:
2018-07-18 Andreas Krebbel <krebbel@linux.ibm.com>
Marcin Kościelnicki <koriakin@0x04.net>
* emultempl/elf32.em: Add code to pick dedicated linker scripts
for partial relro.
* genscripts.sh: Generate dedicated linker scripts for partial relro.
Tom Tromey [Sun, 27 May 2018 05:26:39 +0000 (23:26 -0600)]
Return unique_xmalloc_ptr from gdbscm_safe_eval_string
This changes gdbscm_safe_eval_string to return a unique_xmalloc_ptr.
This allows for the removal of some cleanups. It also fixes a
potential latent memory leak in gdbscm_set_backtrace.
gdb/ChangeLog
2018-07-17 Tom Tromey <tom@tromey.com>
Tom Tromey [Sun, 27 May 2018 05:21:23 +0000 (23:21 -0600)]
Change gdbscm_exception_message_to_string to return a unique_xmalloc_ptr
This changes gdbscm_exception_message_to_string to return a
unique_xmalloc_ptr, allowing for the removal of some cleanups.
unique_xmalloc_ptr was chosen because at the root of the call chains
is a function from Guile that returns a malloc'd string.
gdb/ChangeLog
2018-07-17 Tom Tromey <tom@tromey.com>
BFD: Use `bfd_is_abs_symbol' to determine whether a symbol is absolute
Use `bfd_is_abs_symbol' to determine whether a symbol is absolute,
avoiding a problem with ordinary symbols defined in a linker script
outside an output section definition. Such symbols have its owning
section set to the absolute section up to the final link phase. A flag
has been added to the link hash to identify such symbols. Rather than
checking the flag by hand, use the macro that does it uniformly for all
users.
bfd/
* elf32-nds32.c (nds32_elf_relax_loadstore): Use
`bfd_is_abs_symbol' rather than `bfd_is_abs_section' in checking
whether the symbol is absolute.
(nds32_elf_relax_lo12): Likewise.
* elfnn-aarch64.c (elfNN_aarch64_final_link_relocate): Likewise.
(elfNN_aarch64_check_relocs): Likewise.
* xcofflink.c (xcoff_need_ldrel_p): Likewise.
(bfd_xcoff_import_symbol): Likewise.
(xcoff_write_global_symbol): Likewise.
LD: Export relative-from-absolute symbol marking to BFD
It is usually possible to tell absolute and ordinary symbols apart in
BFD throughout the link, by checking whether the section that owns the
symbol is absolute or not.
That however does not work for ordinary symbols defined in a linker
script outside an output section statement. Initially such symbols are
entered into to the link hash as absolute symbols, owned by the absolute
section. A flag is set in the internal linker expression defining such
symbols to tell the linker to convert them to section-relative ones in
the final phase of the link. That flag is however not accessible to BFD
linker code, including BFD target code in particular.
Add a flag to the link hash then to copy the information held in the
linker expression. Define a macro, `bfd_is_abs_symbol', for BFD code to
use where determining whether a symbol is absolute or ordinary is
required before the final link phase.
This macro will correctly identify the special `__ehdr_start' symbol as
ordinary throughout link, for example, even though early on it will be
assigned to the absolute section. Of course this does not let BFD code
identify what the symbol's ultimate section will be before the final
link phase has converted this symbol (in `update_definedness').
Jim Wilson [Tue, 17 Jul 2018 16:44:42 +0000 (09:44 -0700)]
RISC-V: Don't decrement pc after break.
gdb/
* riscv-tdep.c (riscv_has_feature): Delete comment that refers to
set_gdbarch_decr_pc_after_break. Call riscv_read_misa_reg always.
(riscv_gdbarch_init): Delete local has_compressed_isa. Delete now
unecessary braces after EF_RISCV_RVC test. Delete call to
set_gdbarch_decr_pc_after_break.
Tom Tromey [Thu, 28 Jun 2018 17:19:33 +0000 (11:19 -0600)]
Remove --disable-gdbcli
I think it doesn't really make sense to allow building gdb without the
CLI. Perhaps at one time this was a goal, but libgdb is long gone and
the CLI is intrinsic to gdb.
So, this patch removes the implementation of this configure option.
It is still recognized (this is autoconf's default), but does nothing.
This simplifies configure.ac and Makefile.in a bit.
Tested by rebuilding.
gdb/ChangeLog
2018-07-17 Tom Tromey <tom@tromey.com>
* configure.ac: Remove --disable-gdbcli.
* configure: Rebuild.
* Makefile.in (SUBDIR_CLI_DEPS, SUBDIR_CLI_LDFLAGS)
(SUBDIR_CLI_CFLAGS): Remove.
(SFILES): Use SUBDIR_CLI_SRCS.
(COMMON_OBS): Use SUBDIR_CLI_OBS.
Tom Tromey [Fri, 29 Jun 2018 21:10:04 +0000 (15:10 -0600)]
Use scoped_free_pendings in coff_symtab_read
PR gdb/18624 concerns an assertion failure that occurs when setting a
breakpoint in a Go program on Windows.
What happens here is that coff_symtab_read uses buildsym but does not
instantiate scoped_free_pendings. So, the struct pending objects are
never released. Later, dwarf2read.c calls buildsym_init, which
asserts.
This patch fixes the problem by instantiating scoped_free_pendings in
coff_symtab_read.
Tested using the test executable from the PR. I don't know how to
test this more fully.
gdb/ChangeLog
2018-07-17 Tom Tromey <tom@tromey.com>
PR gdb/18624:
* coffread.c (coff_symtab_read): Use scoped_free_pendings.
Simon Marchi [Mon, 16 Jul 2018 18:15:09 +0000 (14:15 -0400)]
Fix compilation error in ia64-vms-tdep.c with libunwind-ia64
I found this when doing a --enable-targets=all build with libunwind-ia64
properly configured.
CXX ia64-vms-tdep.o
/home/emaisin/src/binutils-gdb/gdb/ia64-vms-tdep.c: In function ‘int ia64_vms_find_proc_info_x(unw_addr_space_t, unw_word_t, unw_proc_info_t*, int, void*)’:
/home/emaisin/src/binutils-gdb/gdb/ia64-vms-tdep.c:79:33: error: invalid conversion from ‘void*’ to ‘gdb_byte* {aka unsigned char*}’ [-fpermissive]
pi->unwind_info, pi->unwind_info_size);
^
In file included from /home/emaisin/src/binutils-gdb/gdb/target.h:70:0,
from /home/emaisin/src/binutils-gdb/gdb/exec.h:23,
from /home/emaisin/src/binutils-gdb/gdb/gdbcore.h:29,
from /home/emaisin/src/binutils-gdb/gdb/ia64-vms-tdep.c:25:
/home/emaisin/src/binutils-gdb/gdb/target/target.h:35:12: note: initializing argument 2 of ‘int target_read_memory(CORE_ADDR, gdb_byte*, ssize_t)’
extern int target_read_memory (CORE_ADDR memaddr, gdb_byte *myaddr,
^
Tom Tromey [Mon, 21 May 2018 04:45:44 +0000 (22:45 -0600)]
Make free_pending_blocks static
free_pending_blocks can be static because scoped_free_pendings (et al)
arrange for it to be NULL in the "steady state". This removes a
couple of unnecessary calls to free_pending_blocks and changes it to
be static.
gdb/ChangeLog
2018-07-16 Tom Tromey <tom@tromey.com>
Tom Tromey [Sun, 20 May 2018 17:25:20 +0000 (11:25 -0600)]
Remove buildsym_new_init
buildsym_new_init is just an alias for buildsym_init. This removes
it. In the long run buildsym_init will also go away; this patch just
helps make things a bit clearer in the meantime.
gdb/ChangeLog
2018-07-16 Tom Tromey <tom@tromey.com>
Tom Tromey [Sun, 20 May 2018 17:19:14 +0000 (11:19 -0600)]
Move within_function to stabsread
The global within_function is only used by a few symbol readers. This
patch moves the global out of buildsym and into stabsread, which
seemed like a better fit. It also arranges for the existing readers
to clear the global at the appropriate time.
gdb/ChangeLog
2018-07-16 Tom Tromey <tom@tromey.com>
Tom Tromey [Sun, 20 May 2018 17:16:57 +0000 (11:16 -0600)]
Move processing_gcc to stabsread
processing_gcc is also only used by stabsread -- it is set by the
DWARF reader, but this turns out not to be needed. So, this patch
moves processing_gcc and removes the assignment from the DWARF reader.
gdb/ChangeLog
2018-07-16 Tom Tromey <tom@tromey.com>
* stabsread.h (processing_gcc_compilation): Move from buildsym.h.
* dwarf2read.c (dwarf2_start_symtab): Don't set
processing_gcc_compilation.
* buildsym.h (processing_gcc_compilation): Move to stabsread.h.
Tom Tromey [Sun, 20 May 2018 17:13:12 +0000 (11:13 -0600)]
Move some code from buildsym to stabsread
A few things that currently reside in buildsym.c turn out to be
specific to the stabs reader. This patch moves these from
buildsym.[ch] to stabsread.[ch].
gdb/ChangeLog
2018-07-16 Tom Tromey <tom@tromey.com>
* stabsread.h (HASHSIZE, hashname, symnum, next_symbol_text)
(next_symbol_text_func): Move from buildsym.h.
* stabsread.c (hashname): Move from buildsym.c.
* buildsym.h (HASHSIZE, symnum, next_symbol_text)
(next_symbol_text_func, hashname): Move to stabsread.h.
* buildsym.c: Don't include bcache.h
(hashname): Move to stasbread.c.
Jan Beulich [Mon, 16 Jul 2018 06:19:21 +0000 (08:19 +0200)]
x86: fix operand size checking
Currently mov to/from control, debug, and test register insns accept any
size GPR operand (general pattern: templates with D set and both
operands being registers in distinct register files). This is due to
improper checking of the reverse case, including not informing the
caller whether a straight and/or reverse match was successful.
The helper functions need to be told two indexes: One to index the given
operand types array, and the other to index the template one. The caller
must attempt a further straight match only if the function reported a
straight match (and respectively for reverse matches).
Fix internal error when using --emit-relocs with plugins.
When an .eh_frame section has deferred layout because of plugins, gold
was neglecting to mark the section as deferred. When we later processed
the corresponding relocation section, we were then ignoring it, causing
the internal error later on in the link.
gold/
PR gold/23397
* object.cc (Sized_relobj_file::do_layout): Mark section as deferred.
* testsuite/Makefile.am (plugin_test_1): Add --emit-relocs option to
existing test case.
* testsuite/Makefile.in: Regenerate.
Fix problem causing duplicated linker-generated symbols with versions.
When generating _end, _edata, etc. symbols, and a version script provides
a version name, and we are linking against another shared library that
provides those symbols with a different version, gold ends up trying to
resolve the other shared library's symbols to the new definitions, resulting
in two copies of each symbol, one as default, and one as non-default.
This patch tests for that condition, and ignores the symbols provided
by the other shared library.
gold/
PR gold/23409
* symtab.cc (Symbol_table::define_special_symbol): Add check for
version name on existing symbol.
* testsuite/Makefile.am (ver_test_pr23409): New test case.
* testsuite/Makefile.in: Regenerate.
* testsuite/ver_test_pr23409.sh: New test script.
* testsuite/ver_test_pr23409_1.script: New version script.
* testsuite/ver_test_pr23409_2.script: New version script.
Simon Marchi [Sat, 14 Jul 2018 02:06:34 +0000 (22:06 -0400)]
Remove struct keyword in range-based for loop
Fix this with gcc 6.3.0:
/home/simark/src/binutils-gdb/gdb/symfile.c: In function 'void set_objfile_default_section_offset(objfile*, const section_addr_info&, CORE_ADDR)':
/home/simark/src/binutils-gdb/gdb/symfile.c:2114:14: error: types may not be defined in a for-range-declaration [-Werror]
for (const struct other_sections *objf_sect : objf_addrs_sorted)
^~~~~~
gdb/ChangeLog:
Expect for another variant of error message when gdbserver cannot resolve hostname
I've noticed that on a few hosts, when given an invalid hostname,
gdbserver fails with:
spawn /../../gdb/gdbserver/gdbserver --once tcp8:123:2353 /gdb/build/fedora-s390x/build/gdb/testsuite/outputs/gdb.server/server-connect/server-connect
tcp8:123:2353: cannot resolve name: No address associated with hostname
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
Exiting
Unfortunately, this causes a fail on the new
gdb.server/server-connect.exp test (introduced by the IPv6 patch):
FAIL: gdb.server/server-connect.exp: tcp8: start gdbserver: gdbserver should fail but did not
This happens because we're expecting for another variant of this error
message:
cannot resolve name: Name or service not known
Therefore, this patch extends the helper function 'gdbserver_start' to
also recognize the "No address associated with hostname" message.
This "fixes" the testcase on the hosts that use this variant.
* lib/gdbserver-support.exp (gdbserver_start): Expect for the
message "No address associated with hostname" when gdbserver
cannot resolve the hostname.
Nick Clifton [Fri, 13 Jul 2018 14:48:49 +0000 (15:48 +0100)]
Add a test that relocs are correctly generated for missing build notes.
* testsuite/gas/elf/missing-build-notes.s: New test. Checks that
relocs are correctly generated for missing build notes.
* testsuite/gas/elf/missing-build-notes.d: New file. Expected
output from objdump.
* testsuite/gas/elf/elf.exp: Run the new test.
Nick Clifton [Fri, 13 Jul 2018 10:46:44 +0000 (11:46 +0100)]
Allow bit-patterns in the immediate field of ARM neon mov instructions.
* config/tc-arm.c (do_neon_mov): When converting an integer
immediate into a floating point value, check that the conversion
is valid. Also warn if the immediate is valid as both a floating
point value and a bit pattern.
* testsuite/gas/arm/vfp-mov-enc.s: Add instructions that use
floating point bit patterns.
* testsuite/gas/arm/vfp-mov-enc.d: Add regexps for the disassembly
of the new insns.
Andreas Arnez [Fri, 13 Jul 2018 10:46:14 +0000 (12:46 +0200)]
S390: Fix displaced stepping of "basr r,0"
The BASR instruction behaves differently depending on whether the second
operand is a number from 1 to 15, or zero. In the former case BASR jumps
to the address contained in the general register of that number, but in
the latter case no jump is performed. GDB's displaced-stepping logic does
not distinguish these cases, although it should.
This is fixed. In the case where no jump is performed the PC is adjusted
to point back after the original instruction. Otherwise the PC is left
alone.
gdb/ChangeLog:
* s390-tdep.c (s390_displaced_step_fixup): Adjust PC for a
non-branching basr.
Pedro Alves [Fri, 13 Jul 2018 09:28:47 +0000 (10:28 +0100)]
GDBserver: Pass process_info pointer to target_kill
We start from a process_info pointer, pass down process->pid, and
then the target_kill implementations need to find the process from the
pid again. Pass the process_info pointer down directly instead.
gdb/gdbserver/ChangeLog:
2018-07-13 Pedro Alves <palves@redhat.com>
* linux-low.c (linux_kill): Change parameter to process_info
pointer instead of pid. Adjust.
* lynx-low.c (lynx_kill): Likewise.
* nto-low.c (nto_kill): Likewise.
* spu-low.c (spu_kill): Likewise.
* win32-low.c (win32_kill): Likewise.
* server.c (handle_v_kill, kill_inferior_callback)
(detach_or_kill_for_exit): Adjust.
* target.c (kill_inferior): Change parameter to process_info
pointer instead of pid. Adjust.
* target.h (struct target_ops) <kill>: Change parameter to
process_info pointer instead of pid. Adjust all implementations
and callers.
(kill_inferior): Likewise.
Pedro Alves [Fri, 13 Jul 2018 09:28:46 +0000 (10:28 +0100)]
GDBserver: Pass process_info pointer to target_detach and target_join
We start from a process_info pointer, pass down process->pid, and then
the target_detach and target_join implementations need to find the
process from the pid again. Pass the process_info pointer down
directly instead.
gdb/gdbserver/ChangeLog:
2018-07-13 Pedro Alves <palves@redhat.com>
* linux-low.c (linux_detach, linux_join): Change parameter to
process_info pointer instead of pid. Adjust.
* lynx-low.c (lynx_detach, lynx_join): Likewise.
* nto-low.c (nto_detach): Likewise.
* spu-low.c (spu_detach, spu_join): Likewise.
* win32-low.c (win32_detach, win32_join): Likewise.
* server.c (handle_detach, detach_or_kill_for_exit): Adjust.
* target.h (struct target_ops) <detach, join>: Change parameter to
process_info pointer instead of pid. Adjust all implementations
and callers.
(detach_inferior, join_inferior): Rename 'pid' parameter to
'proc'.
Enhance 'thread apply' command to also accept [FLAG]... arguments.
An example usage for this new argument:
thread apply all -s frame apply all -s p some_local_var_somewhere
Prints the thread id, frame location and some_local_var_somewhere
value in frames of threads that have such local var.
To make the life of the user easier, the most typical use cases
have shortcuts :
taas : shortcut for 'thread apply all -s'
tfaas : shortcut for 'thread apply all -s frame apply all -s"
An example usage :
tfaas p some_local_var_somewhere
same as the longer:
'thread apply all -s frame apply all -s p some_local_var_somewhere'
gdb/ChangeLog
2018-07-12 Philippe Waroquiers <philippe.waroquiers@skynet.be>
* thread.c (thr_try_catch_cmd): New function.
(thread_apply_all_command): Handle qcs flags.
(thread_apply_command): Handle qcs flags.
(taas_command): New function.
(tfaas_command): New function.
(_initialize_thread): Update to setup the new commands 'taas
and 'tfaas'. Change doc string for 'thread apply'.
Implement frame apply [all | COUNT | -COUNT | level LEVEL... ] [FLAG]... COMMAND.
Also implement the command 'faas COMMAND', a shortcut for
'frame apply all -s COMMAND'.
The syntax of 'frame apply' to specify some innermost or outermost
frames is similar to 'backtrace' command, using COUNT or -COUNT.
To apply a COMMAND to a more specific set of frames, the following
new command and syntax can be used:
frame apply level LEVEL... [FLAG]... COMMAND
where LEVEL is one or more frame levels or range of frame levels.
The new command 'frame apply' allows to apply a COMMAND to a number of frames,
or to all frames, or to a set of frames.
The optional [FLAG]... arguments allow to control what output to produce
and how to handle errors raised when applying COMMAND to a frame.
Some example usages for this new command:
frame apply all info frame
Produce info frame for all frames.
frame apply all p $sp
For each frame, print the location, followed by the frame sp.
frame apply all -q p $sp
Same as before, but -q flag (q = quiet) indicates to only print
the frames sp.
frame apply all p some_local_var_somewhere
Print some_local_var_somewhere in all frames. 'frame apply'
will abort as soon as the print command fails.
frame apply all -c p some_local_var_somewhere
Same as before, but -c flag (c = continue) means to
print the error and continue applying command in case the
print command fails.
frame apply all -s p some_local_var_somewhere
Same as before, but -s flag (s = silent) means to
be silent for frames where the print command fails.
In other words, this allows to 'search' the frame in which
some_local_var_somewhere can be printed.
frame apply all -s -q p some_local_var_somewhere
Same as before, but does not print the frame info in which
the variable is found.
frame apply level 2-4 5 8-10 -s p i = i + 1
Increments i in the identified frames.
gdb/ChangeLog
2018-07-12 Philippe Waroquiers <philippe.waroquiers@skynet.be>
* stack.c: (trailing_outermost_frame): New function, mostly
extracted from backtrace_command_1.
(leading_innermost_frame): New function.
(backtrace_command_1): Update to call trailing_outermost_frame.
(frame_apply_command_count): New function.
(frame_apply_level_command): New function.
(frame_apply_all_command): New function.
(frame_apply_command): New function.
(faas_command): New function.
(frame_cmd_list): New variable.
(_initialize_stack): Update to setup the new commands 'frame apply'
and 'faas'.
Add helper functions parse_flags and parse_flags_qcs
Add helper functions parse_flags and parse_flags_qcs.
parse_flags helper function allows to look for a set of flags at
the start of a string.
A flag must be given individually.
parse_flags_qcs is a specialised helper function to handle
the flags -q, -c and -s, that are used in the new command 'frame apply'
and in the command 'thread apply.
Modify number_or_range_parser::get_number to differentiate a
- followed by digits from a - followed by an alpha (i.e. a flag or an option).
That is needed for the addition of the [FLAG]... arguments to
thread apply ID... [FLAG]... COMMAND
Remove bool number_or_range_parser::m_finished, rather
implement the 'finished' logic inside number_or_range_parser::finished.
The new logic properly detects the end of parsing even if not at
end of the string. This ensures that number_or_range_parser::cur_tok
really points past the last parsed token when parsing is finished.
Before, it was always pointing at the end of the string.
As parsing now is finished directly when not positioned on a number,
number_is_in_list must do an error check before the loop getting all
numbers.
The error message for 'thread apply -$unknownconvvar p 1'
is now the more clear:
Convenience variable must have integer value.
Invalid thread ID: -$unknownconvvar p 1
instead of previously:
negative value
gdb/ChangeLog
2018-07-12 Philippe Waroquiers <philippe.waroquiers@skynet.be>
* cli-utils.c (number_or_range_parser::get_number): Only handle
numbers or convenience var as numbers.
(parse_flags): New function.
(parse_flags_qcs): New function.
(number_or_range_parser::finished): Ensure parsing end is detected
before end of string.
* cli-utils.h (parse_flags): New function.
(parse_flags_qcs): New function.
(number_or_range_parser): Remove m_finished bool.
(number_or_range_parser::skip_range): Set m_in_range to false.
gdb/testsuite/ChangeLog
2018-07-12 Philippe Waroquiers <philippe.waroquiers@skynet.be>
MIPS/LD/testsuite: Correct mips-elf.exp test ABI/emul/endian arrangement
Similarly to commit 86b24e15c45b ("MIPS/LD/testsuite: Correct
comm-data.exp test ABI/emul/endian arrangement") update the mips-elf.exp
test script to:
- correctly select emulations for targets using non-traditional MIPS
emulations,
- correctly select ABIs for targets that do not support all of them,
- use the default endianness selection where possible to benefit targets
that support only one,
- simplify test invocation by providing ABI-specific `run_dump_test'
wrappers, specifically `run_dump_test_o32', `run_dump_test_n32',
`run_dump_test_n64' and `run_dump_test_eabi', which remove the need to
use conditionals across the Expect script or to repeat ABI-specific
GAS and LD flags with each invocation,
removing numerous test failures for `mips-sgi-irix6', `mips64-openbsd',
`mips64el-openbsd' and `mips64el-ps2-elf' targets and broadening
coverage for several MIPS targets.
There are some new failures for the `mips64el-ps2-elf' target with tests
that were not previously run for that target:
FAIL: MIPS16 link branch to absolute expression (n32)
FAIL: MIPS16 link branch to absolute expression 1 (n32)
FAIL: MIPS16 link branch to absolute expression 2 (n32)
FAIL: microMIPS link branch to absolute expression (n32)
FAIL: MIPS ELF got reloc n32
FAIL: MIPS ELF xgot reloc n32
FAIL: undefined weak symbol overflow (n32)
FAIL: R_MIPS16_HI16 and R_MIPS16_LO16 relocs n32
FAIL: ld-mips-elf/attr-gnu-4-0-n32-ph
FAIL: ld-mips-elf/attr-gnu-4-1-n32-ph
FAIL: ld-mips-elf/attr-gnu-4-2-n32-ph
FAIL: ld-mips-elf/attr-gnu-4-3-n32-ph
FAIL: MIPSr6 JALR reloc unaligned/cross-mode link test (n32)
which are mostly due to dump discrepancies caused by mapping differences
coming from the default linker scripts used by these test cases, or
sometimes because of the specific MIPS processor architecture recorded
in the ELF file taking precedence over the general MIPS ISA level also
recorded. Finally, the R_MIPS16_HI16/R_MIPS16_LO16 relocation test
failure is a preexisting issue with the IRIX style emulation.
These failures will have to be addressed separately.
MIPS/LD/testsuite: Limit reloc 6 tests to DSO targets
Complement commit 500c05ddbe98 ("MIPS/ld/testsuite: Limit export class
call reloc tests to DSO targets") and also check for target's shared
library support in determining whether to run reloc 6 tests, which
require it. No changes in results as all targets currently enabled for
NewABI testing support shared libraries.
Declare 'ioarg' (from ser-tcp.c:try_connect) as 'u_long' when on Windows (and unbreak build on mingw32)
When building GDB on mingw32, it fails with:
../../binutils-gdb/gdb/ser-tcp.c: In function 'int try_connect(const addrinfo*, unsigned int*)':
../../binutils-gdb/gdb/ser-tcp.c:176:25: error: invalid conversion from 'int*' to 'u_long* {aka long unsigned int*}' [-fpermissive]
ioctl (sock, FIONBIO, &ioarg);
^~~~~~
In file included from ../../binutils-gdb/gdb/serial.h:23:0,
from ../../binutils-gdb/gdb/ser-tcp.c:21:
/usr/x86_64-w64-mingw32/sys-root/mingw/include/winsock2.h:977:34: note: initializing argument 3 of 'int ioctlsocket(SOCKET, long int, u_long*)'
WINSOCK_API_LINKAGE int WSAAPI ioctlsocket(SOCKET s,__LONG32 cmd,u_long *argp);
^~~~~~~~~~~
make[2]: *** [Makefile:1610: ser-tcp.o] Error 1
The problem happens because the IPv6
commit (c7ab0aef11d91b637bf091aa9176b8dc4aadee46) wrongly removed the
code responsible for declaring 'ioarg' with a different type if
building for Windows. This patch restores that.
Tamar Christina [Thu, 12 Jul 2018 09:28:46 +0000 (10:28 +0100)]
Add remainder of Em16 restrictions for AArch64 gas.
This adds the missing Em16 constraints the rest of the instructions requiring them
and also adds a testcase to test all the instructions so these are checked from
now on.
The Em16 operand constrains the valid registers to the lower 16 registers when used
with a half precision qualifier.
The list has been cross checked (by hand) through the Arm ARM version Ca.
This patch implements IPv6 support for both GDB and gdbserver. Based
on my research, it is the fourth attempt to do that since 2006. Since
I used ideas from all of the previous patches, I also added their
authors's names on the ChangeLogs as a way to recognize their
efforts. For reference sake, you can find the previous attempts at:
The basic idea behind the patch is to start using the new
'getaddrinfo'/'getnameinfo' calls, which are responsible for
translating names and addresses in a protocol-independent way. This
means that if we ever have a new version of the IP protocol, we won't
need to change the code again (or, at least, won't have to change the
majority of the code).
The function 'getaddrinfo' returns a linked list of possible addresses
to connect to. Dealing with multiple addresses proved to be a hard
task with the current TCP auto-retry mechanism implemented on
ser-tcp:net_open. For example, when gdbserver listened only on an
IPv4 socket:
$ ./gdbserver --once 127.0.0.1:1234 ./a.out
and GDB was instructed to try to connect to both IPv6 and IPv4
sockets:
the user would notice a somewhat big delay before GDB was able to
connect to the IPv4 socket. This happened because GDB was trying to
connect to the IPv6 socket first, and had to wait until the connection
timed out before it tried to connect to the IPv4 socket.
For that reason, I had to rewrite the main loop and implement a new
method for handling multiple connections. After some discussion,
Pedro and I agreed on the following algorithm:
1) For each entry returned by 'getaddrinfo', we try to open a socket
and connect to it.
2.a) If we have a successful 'connect', we just use that connection.
2.b) If we don't have a successfull 'connect', but if we've got a
ECONNREFUSED (meaning the the connection was refused), we keep track
of this fact by using a flag.
2.c) If we don't have a successfull 'connect', but if we've got a
EINPROGRESS (meaning that the connection is in progress), we perform
a 'select' call on the socket until we have a result (either a
successful connection, or an error on the socket).
3) If tcp_auto_retry is true, and we haven't gotten a successful
connection, and at least one of our attempts failed with
ECONNREFUSED, then we wait a little bit (i.e., call
'wait_for_connect'), check to see if there was a
timeout/interruption (in which case we bail out), and then go back
to (1).
After multiple tests, I was able to connect without delay on the
scenario described above, and was also able to connect in all other
types of scenarios.
I also implemented some hostname parsing functions (along with their
corresponding unit tests) which are used to help GDB and gdbserver to
parse hostname strings provided by the user. These new functions are
living inside common/netstuff.[ch]. I've had to do that since IPv6
introduces a new URL scheme, which defines that square brackets can be
used to enclose the host part and differentiate it from the
port (e.g., "[::1]:1234" means "host ::1, port 1234"). I spent some
time thinking about a reasonable way to interpret what the user wants,
and I came up with the following:
- If the user has provided a prefix that doesn't specify the protocol
version (i.e., "tcp:" or "udp:"), or if the user has not provided
any prefix, don't make any assumptions (i.e., assume AF_UNSPEC when
dealing with 'getaddrinfo') *unless* the host starts with "[" (in
which case, assume it's an IPv6 host).
- If the user has provided a prefix that does specify the protocol
version (i.e., "tcp4:", "tcp6:", "udp4:" or "udp6:"), then respect
that.
This method doesn't follow strictly what RFC 2732 proposes (that
literal IPv6 addresses should be provided enclosed in "[" and "]")
because IPv6 addresses still can be provided without square brackets
in our case, but since we have prefixes to specify protocol versions I
think this is not an issue.
Another thing worth mentioning is the new 'GDB_TEST_SOCKETHOST'
testcase parameter, which makes it possible to specify the
hostname (without the port) to be used when testing GDB and
gdbserver. For example, to run IPv6 tests:
$ make check-gdb RUNTESTFLAGS='GDB_TEST_SOCKETHOST=tcp6:[::1]'
Or, to run IPv4 tests:
$ make check-gdb RUNTESTFLAGS='GDB_TEST_SOCKETHOST=tcp4:127.0.0.1'
This required a few changes on the gdbserver-base.exp, and also a
minimal adjustment on gdb.server/run-without-local-binary.exp.
Finally, I've implemented a new testcase,
gdb.server/server-connect.exp, which is supposed to run on the native
host and perform various "smoke tests" using different connection
methods.
This patch has been regression-tested on BuildBot and locally, and
also built using a x86_64-w64-mingw32 GCC, and no problems were found.
gdb/ChangeLog:
2018-07-11 Sergio Durigan Junior <sergiodj@redhat.com>
Jan Kratochvil <jan.kratochvil@redhat.com>
Paul Fertser <fercerpav@gmail.com>
Tsutomu Seki <sekiriki@gmail.com>
Pedro Alves <palves@redhat.com>
* Makefile.in (SUBDIR_UNITTESTS_SRCS): Add
'unittests/parse-connection-spec-selftests.c'.
(COMMON_SFILES): Add 'common/netstuff.c'.
(HFILES_NO_SRCDIR): Add 'common/netstuff.h'.
* NEWS (Changes since GDB 8.2): Mention IPv6 support.
* common/netstuff.c: New file.
* common/netstuff.h: New file.
* ser-tcp.c: Include 'netstuff.h' and 'wspiapi.h'.
(wait_for_connect): Update comment. New parameter
'gdb::optional<int> sock' instead of 'struct serial *scb'.
Use 'sock' directly instead of 'scb->fd'.
(try_connect): New function, with code from 'net_open'.
(net_open): Rewrite main loop to deal with multiple
sockets/addresses. Handle IPv6-style hostnames; implement
support for IPv6 connections.
* unittests/parse-connection-spec-selftests.c: New file.
gdb/gdbserver/ChangeLog:
2018-07-11 Sergio Durigan Junior <sergiodj@redhat.com>
Jan Kratochvil <jan.kratochvil@redhat.com>
Paul Fertser <fercerpav@gmail.com>
Tsutomu Seki <sekiriki@gmail.com>
* Makefile.in (SFILES): Add '$(srcdir)/common/netstuff.c'.
(OBS): Add 'common/netstuff.o'.
(GDBREPLAY_OBS): Likewise.
* gdbreplay.c: Include 'wspiapi.h' and 'netstuff.h'.
(remote_open): Implement support for IPv6
connections.
* remote-utils.c: Include 'netstuff.h', 'filestuff.h'
and 'wspiapi.h'.
(handle_accept_event): Accept connections from IPv6 sources.
(remote_prepare): Handle IPv6-style hostnames; implement
support for IPv6 connections.
(remote_open): Implement support for printing connections from
IPv6 sources.
gdb/testsuite/ChangeLog:
2018-07-11 Sergio Durigan Junior <sergiodj@redhat.com>
Jan Kratochvil <jan.kratochvil@redhat.com>
Paul Fertser <fercerpav@gmail.com>
Tsutomu Seki <sekiriki@gmail.com>
* README (Testsuite Parameters): Mention new 'GDB_TEST_SOCKETHOST'
parameter.
* boards/native-extended-gdbserver.exp: Do not set 'sockethost'
by default.
* boards/native-gdbserver.exp: Likewise.
* gdb.server/run-without-local-binary.exp: Improve regexp used
for detecting when a remote debugging connection succeeds.
* gdb.server/server-connect.exp: New file.
* lib/gdbserver-support.exp (gdbserver_default_get_comm_port):
Do not prefix the port number with ":".
(gdbserver_start): New global GDB_TEST_SOCKETHOST. Implement
support for detecting and using it. Add '$debughost_gdbserver'
to the list of arguments used to start gdbserver. Handle case
when gdbserver cannot resolve a network name.
gdb/doc/ChangeLog:
2018-07-11 Sergio Durigan Junior <sergiodj@redhat.com>
Jan Kratochvil <jan.kratochvil@redhat.com>
Paul Fertser <fercerpav@gmail.com>
Tsutomu Seki <sekiriki@gmail.com>
* gdb.texinfo (Remote Connection Commands): Add explanation
about new IPv6 support. Add new connection prefixes.
Pedro Alves [Wed, 11 Jul 2018 22:31:44 +0000 (23:31 +0100)]
GDB: Work around D;PID handling bug in older GDBservers (PR gdb/23377)
This commit adds a GDB workaround for the GDBserver bug exposed by
commit f2ffa92bbce9 ("gdb: Eliminate the 'stop_pc' global"), so that
newer GDBs can continue working with older GDBservers.
gdb/ChangeLog:
2018-07-11 Pedro Alves <palves@redhat.com>
Pedro Alves [Wed, 11 Jul 2018 22:31:44 +0000 (23:31 +0100)]
GDBserver: Don't assume a current process in D;PID implementation (PR gdb/23377)
This fixes a gdb.base/multi-forks.exp regression with GDBserver.
Git commit f2ffa92bbce9 ("gdb: Eliminate the 'stop_pc' global") caused
the regression by exposing a latent bug in gdbserver.
The bug is that GDBserver's implementation of the D;PID packet
incorrectly assumes that the selected thread points to the process
being detached. This happens via the any_persistent_commands call,
which calls current_process:
(gdb) bt
#0 0x000000000040a57e in internal_error(char const*, int, char const*, ...)
(file=0x4a53c0 "src/gdb/gdbserver/inferiors.c", line=212, fmt=0x4a539e "%s:
Assertion `%s' failed.") at src/gdb/gdbserver/../common/errors.c:54
#1 0x0000000000420acf in current_process() () at
src/gdb/gdbserver/inferiors.c:212
#2 0x00000000004226a0 in any_persistent_commands() () at
gdb/gdbserver/mem-break.c:308
#3 0x000000000042cb43 in handle_detach(char*) (own_buf=0x6f0280 "D;62ea") at
src/gdb/gdbserver/server.c:1210
#4 0x0000000000433af3 in process_serial_event() () at
src/gdb/gdbserver/server.c:4055
#5 0x0000000000434878 in handle_serial_event(int, void*) (err=0,
client_data=0x0)
The "eliminate stop_pc" commit exposes the problem because before that
commit, GDB's switch_to_thread always read the newly-selected thread's
PC, and that would end up forcing GDBserver's selected thread to
change accordingly as side effect. After that commit, GDB no longer
reads the thread's PC, and GDBserver does not switch the thread.
Fix this by removing the assumption from GDBserver.
gdb/gdbserver/ChangeLog:
2018-07-11 Pedro Alves <palves@redhat.com>
PR gdb/23377
* mem-break.c (any_persistent_commands): Add process_info
parameter and use it instead of relying on the current process.
Change return type to bool.
* mem-break.h (any_persistent_commands): Add process_info
parameter and change return type to bool.
* server.c (handle_detach): Remove require_running_or_return call.
Look up the process_info for the process we're about to detach.
If not found, return back error to GDB. Adjust
any_persistent_commands call to pass down a process pointer.
Pedro Alves [Wed, 11 Jul 2018 19:39:12 +0000 (20:39 +0100)]
gdb: Remove traces of h8300 ecoff support
There's this little bit of ecoff support in the h8300 port that I ran
into that I think we could remove. Grepping around BFD, I don't think
ECOFF is supported for h8300.
And if not, I'm failing to see why we'd support ECOFF debug info in
ELF for h8300 (.mdebug). It kind of seems like someone just installed
the set_gdbarch_ecoff_reg_to_regnum hook at the same time the dwarf2
and stabs hooks were installed.
I'm removing this so that greps for ecoff stuff only lead to MIPS, as
ISTR that Linux/MIPS support still relies on some subset of ECOFF for
something.
gdb/ChangeLog:
2018-07-11 Pedro Alves <palves@redhat.com>
Fix PR c++/23373: GDB hangs when printing a struct with a static member of itself
This patch fixes a failure that happens when a structure has a static
member whose type is the same as itself. From the bug report:
Example code:
struct A
{
static A Empty;
int a;
};
int main(void) {
A a;
return 0;
}
Output:
(gdb) ptype/o A
/* offset | size */ type = struct A {
static struct A {
static struct A {
static struct A {
static struct A {
static struct A {
static struct A {
... # infinite loop
The problem here is that GDB is not taking into account the fact that
static members inside a class/struct are not stored in the
class/struct, and therefore they should not be accounted for during
the display of the offsets/sizes. The fix is simple: we just check if
the field we're dealing with (on
c-typeprint.c:c_type_print_base_struct_union) is static, and if it is
then we don't iterate over it.
This patch also adds a new test for this case, and doesn't introduce
any regressions. I believe it is important enough to be included in
the 8.2 branch.
PR c++/23373
* gdb.base/ptype-offsets.cc (struct static_member): New
struct.
(main) <stmember>: New variable.
* gdb.base/ptype-offsets.exp: Add test for printing a struct
with a static member in it.
Pedro Alves [Wed, 11 Jul 2018 18:49:19 +0000 (19:49 +0100)]
GDBserver: Fix "Cond. jump or move depends on uninit value" in x87 code
Running gdbserver under Valgrind I get:
==26925== Conditional jump or move depends on uninitialised value(s)
==26925== at 0x473E7F: i387_cache_to_xsave(regcache*, void*) (i387-fp.c:579)
==26925== by 0x46E3ED: x86_fill_xstateregset(regcache*, void*) (linux-x86-low.c:418)
==26925== by 0x45E747: regsets_store_inferior_registers(regsets_info*, regcache*) (linux-low.c:5456)
==26925== by 0x45EEF8: linux_store_registers(regcache*, int) (linux-low.c:5731)
==26925== by 0x426441: regcache_invalidate_thread(thread_info*) (regcache.c:89)
==26925== by 0x45CCAF: linux_resume_one_lwp_throw(lwp_info*, int, int, siginfo_t*) (linux-low.c:4447)
==26925== by 0x45CE2A: linux_resume_one_lwp(lwp_info*, int, int, siginfo_t*) (linux-low.c:4519)
==26925== by 0x45E17C: proceed_one_lwp(thread_info*, lwp_info*) (linux-low.c:5216)
==26925== by 0x45DC81: linux_resume_one_thread(thread_info*, bool) (linux-low.c:5031)
==26925== by 0x45DD34: linux_resume(thread_resume*, unsigned long)::{lambda(thread_info*)#2}::operator()(thread_info*) const (linux-low.c:5095)
==26925== by 0x462907: void for_each_thread<linux_resume(thread_resume*, unsigned long)::{lambda(thread_info*)#2}>(linux_resume(thread_resume*, unsigned long)::{lambda(thread_info*)#2}) (gdbthread.h:150)
==26925== by 0x45DE62: linux_resume(thread_resume*, unsigned long) (linux-low.c:5093)
==26925==
==26925== Conditional jump or move depends on uninitialised value(s)
==26925== at 0x473EBD: i387_cache_to_xsave(regcache*, void*) (i387-fp.c:586)
==26925== by 0x46E3ED: x86_fill_xstateregset(regcache*, void*) (linux-x86-low.c:418)
==26925== by 0x45E747: regsets_store_inferior_registers(regsets_info*, regcache*) (linux-low.c:5456)
==26925== by 0x45EEF8: linux_store_registers(regcache*, int) (linux-low.c:5731)
==26925== by 0x426441: regcache_invalidate_thread(thread_info*) (regcache.c:89)
==26925== by 0x45CCAF: linux_resume_one_lwp_throw(lwp_info*, int, int, siginfo_t*) (linux-low.c:4447)
==26925== by 0x45CE2A: linux_resume_one_lwp(lwp_info*, int, int, siginfo_t*) (linux-low.c:4519)
==26925== by 0x45E17C: proceed_one_lwp(thread_info*, lwp_info*) (linux-low.c:5216)
==26925== by 0x45DC81: linux_resume_one_thread(thread_info*, bool) (linux-low.c:5031)
==26925== by 0x45DD34: linux_resume(thread_resume*, unsigned long)::{lambda(thread_info*)#2}::operator()(thread_info*) const (linux-low.c:5095)
==26925== by 0x462907: void for_each_thread<linux_resume(thread_resume*, unsigned long)::{lambda(thread_info*)#2}>(linux_resume(thread_resume*, unsigned long)::{lambda(thread_info*)#2}) (gdbthread.h:150)
==26925== by 0x45DE62: linux_resume(thread_resume*, unsigned long) (linux-low.c:5093)
The problem is a type/width mismatch in code like this, in
gdbserver/i387-fp.c:
/* Some registers are 16-bit. */
collect_register_by_name (regcache, "fctrl", &val);
fp->fctrl = val;
In the above code:
#1 - 'val' is a 64-bit unsigned long.
#2 - "fctrl" is 32-bit in the register cache, thus half of 'val' is
left uninitialized by collect_register_by_name, which works with
an untyped raw buffer output (i.e., void*).
#3 - fp->fctrl is an unsigned short (16-bit). For some such
registers we're masking off the uninitialized bits with 0xffff,
but not in all cases.
We end up in such a fragile situation because
collect_registers_by_name works with an untyped output buffer pointer,
making it easy to pass a pointer to a variable of the wrong size.
Fix this by using regcache_raw_get_unsigned instead (actually a new
regcache_raw_get_unsigned_by_name wrapper), which always returns a
zero-extended ULONGEST register value. It ends up simplifying the
i387-tdep.c code a bit, even.
gdb/gdbserver/ChangeLog:
2018-07-11 Pedro Alves <palves@redhat.com>
* i387-fp.c (i387_cache_to_fsave, cache_to_fxsave)
(i387_cache_to_xsave): Use regcache_raw_get_unsigned_by_name
instead of collect_register_by_name.
* regcache.c (regcache_raw_get_unsigned_by_name): New.
* regcache.h (regcache_raw_get_unsigned_by_name): New.
MIPS/BFD: Fix TLS relocation resolution for regular executables
Correct an issue with commit 0f20cc3522e7 ("TLS support for MIPS"),
<https://sourceware.org/ml/binutils/2005-02/msg00607.html>, where a
condition used to determine whether to use a dynamic symbol for GD, LD
and IE TLS dynamic relocations against a symbol that has been defined
locally has been incorrectly reversed.
It's executables rather than dynamic shared objects where no symbol is
required, because such symbols cannot be preempted and therefore their
values (thread pointer offsets) are fixed at the static link time as is
the associated module ID of the main executable, so the original
condition should have been `shared' instead of `!shared'. This wrong
condition was then later converted from `!shared' to `!bfd_link_pic',
with commit 0e1862bb401f ("Add output_type to bfd_link_info").
Use the correct `bfd_link_dll' condition then, and adjust code for the
dynamic symbol index possibly being -1 as with symbols that have been
forced local, removing unnecessary dynamic relocations from dynamic
regular executables. PIE executables are unaffected as the existing
condition excluded them by chance due to the conversion mentioned above.
Adjust test cases accordingly.
bfd/
* elfxx-mips.c (mips_tls_got_relocs): Use `bfd_link_dll' rather
than `!bfd_link_pic' in determining the dynamic symbol index.
Avoid the index of -1.
(mips_elf_initialize_tls_slots): Likewise. Flatten code by
moving `dyn' to the beginning of the function block.
PR ld/22570: MIPS/BFD: Fix TLS relocation resolution for PIE executables
Correct a commit 0e1862bb401f ("Add output_type to bfd_link_info") issue
and use `bfd_link_dll' rather than `bfd_link_pic' in determining whether
to fully resolve GD, LD and IE TLS relocations referring to symbols
locally defined rather than deferring them to the load time by means of
dynamic relocations.
Such symbols cannot be preempted in PIE executables, which are
necessarily PIC, and therefore their values (thread pointer offsets) are
fixed at the static link time as is the associated module ID of the main
executable.
Given the `tlsbin-o32.s' and `tlsdyn-o32.s' sources from our test suite
this removes the absolute TLS relocations from the static:
DYNAMIC RELOCATION RECORDS
OFFSET TYPE VALUE 00000000 R_MIPS_NONE *ABS* 1000002c R_MIPS_TLS_TPREL32 *ABS* 10000030 R_MIPS_TLS_DTPMOD32 *ABS* 10000038 R_MIPS_TLS_DTPMOD32 *ABS*
and the dynamic:
DYNAMIC RELOCATION RECORDS
OFFSET TYPE VALUE 00000000 R_MIPS_NONE *ABS* 1000002c R_MIPS_TLS_TPREL32 *ABS* 10000038 R_MIPS_TLS_DTPMOD32 *ABS* 10000044 R_MIPS_TLS_DTPMOD32 *ABS* 10000030 R_MIPS_TLS_DTPMOD32 tlsvar_gd 10000034 R_MIPS_TLS_DTPREL32 tlsvar_gd 10000040 R_MIPS_TLS_TPREL32 tlsvar_ie
PIE executable respectively, as reported by `objdump -R', and fills the
corresponding GOT slots with the values expected, as recorded with the
test cases added. The new output from `objdump -R' is:
DYNAMIC RELOCATION RECORDS (none)
and:
DYNAMIC RELOCATION RECORDS
OFFSET TYPE VALUE 00000000 R_MIPS_NONE *ABS* 10000030 R_MIPS_TLS_DTPMOD32 tlsvar_gd 10000034 R_MIPS_TLS_DTPREL32 tlsvar_gd 10000040 R_MIPS_TLS_TPREL32 tlsvar_ie
for the static and the dynamic executable respectively.
2018-07-11 Maciej W. Rozycki <macro@mips.com>
Rich Felker <bugdal@aerifal.cx>
bfd/
PR ld/22570
* elfxx-mips.c (mips_tls_got_relocs): Use `bfd_link_dll' rather
than `bfd_link_pic' to determine whether dynamic relocations are
to be produced.
(mips_elf_initialize_tls_slots): Likewise.
ld/
PR ld/22570
* testsuite/ld-mips-elf/tlsbin-pie-o32.d: New test.
* testsuite/ld-mips-elf/tlsbin-pie-o32.got: New test.
* testsuite/ld-mips-elf/tlsdyn-pie-o32.d: New test.
* testsuite/ld-mips-elf/tlsdyn-pie-o32.got: New test.
* testsuite/ld-mips-elf/mips-elf.exp: Run the new tests.
Franz Sirl [Wed, 24 Jan 2018 13:39:11 +0000 (13:39 +0000)]
Fix printing the size of GOLD's memory areana on Cygwin based systems.
I just stumbled over this with 2.29.1 while building a cross-toolchain, on Cygwin64, but it's still the same for 2.30. m.arena has size_t on Cygwin64 and thus errors out due to -Werror=format.
Jan Beulich [Wed, 11 Jul 2018 08:30:00 +0000 (10:30 +0200)]
x86: adjust monitor/mwait templates
Architecturally, MONITOR's and MONITORX'es memory operand is a 16- or
32-bit register outside of 64-bit mode, and a 64- or 32-bit register
inside 64-bit mode. The other register operands, including all of them
for MWAIT and MWAITX, are uniformly 32-bit, irrespective of mode. Retain
the original 64-bit MONITOR{,X} templates for compatibility only, and
fold the MWAIT{,X} ones.
Jan Beulich [Wed, 11 Jul 2018 08:28:56 +0000 (10:28 +0200)]
x86: drop {,reg16_}inoutportreg variables
The checking against reg16_inoutportreg can be had with a simple test of
a bit, and the value setting from inoutportreg can be replaced by using
the actual register's reg_type field.
Note that the so far redundant 2nd instance of OPERAND_TYPE_INOUTPORTREG
is left in place, for its use in type_names[].
Jan Beulich [Wed, 11 Jul 2018 08:28:16 +0000 (10:28 +0200)]
x86: simplify legacy prefix emission
The check_prefix label was bogus from the beginning: The special
checking is supposed to happen for PadLock insns only; no
3-opcode-byte insn should go this path.
Jan Beulich [Wed, 11 Jul 2018 08:24:44 +0000 (10:24 +0200)]
x86: replace off-by-one OTMax
With its name it should equal to the largest prior enumerator, which is
not very helpful as that will change every once in a while. Rename it to
OTNum instead to make name and value match, and correct use sites to no
longer use one too large a value.
projects / thirdparty / binutils-gdb.git / log
