Eric Botcazou [Wed, 4 Apr 2018 10:07:50 +0000 (12:07 +0200)]
Speed up direct linking with DLLs on Windows (1/2).
This patch deals with the auto-import feature. There are 2 versions
of this feature: the original one, which was piggybacked on the OS
loader with an optional help from the runtime (--enable-auto-import
--enable-runtime-pseudo-reloc-v1) and is still the one mostly
documented in the sources and manual; the enhanced one by Kai Tietz,
which is entirely piggybacked on the runtime (--enable-auto-import
--enable-runtime-pseudo-reloc-v2) and is the default for Mingw and
Cygwin nowadays.
The implementation is inefficient because of pe[p]_find_data_imports:
for every undefined symbol, the function walks the entire set of
relocations for all the input files and does a direct name comparison
for each of them.
This is easily fixable by using a hash-based map for v1 and a simple
hash table for v2. This patch leaves v1 alone and only changes v2.
It also factors out pe[p]_find_data_imports into a common function,
removes old cruft left and right, and attempts to better separate
the implementations of v1 and v2 in the code.
ld/
* emultempl/pe.em (U_SIZE): Delete.
(pe_data_import_dll): Likewise.
(make_import_fixup): Return void, take 4th parameter and pass it down
in call to pe_create_import_fixup.
(pe_find_data_imports): Move to...
(gld_${EMULATION_NAME}_after_open): Run the stdcall fixup pass after
the auto-import pass and add a guard before running the latter.
* emultempl/pep.em (U_SIZE): Delete.
(pep_data_import_dll): Likewise.
(make_import_fixup): Return void, take 4th parameter and pass it down
in call to pe_create_import_fixup.
(pep_find_data_imports): Move to...
(gld_${EMULATION_NAME}_after_open): Run the stdcall fixup pass after
the auto-import pass and add a guard before running the latter.
* pe-dll.c (runtime_pseudp_reloc_v2_init): Change type to bfd_boolean.
(pe_walk_relocs_of_symbol): Rename into...
(pe_walk_relocs): ...this. Add 2 more parameters,4th parameter to the
callback prototype and pass 4th parameter in calls to the callback.
If the import hash table is present, invoke the callback on the reloc
if the symbol name is in the table.
(pe_find_data_imports): ...here. Take 2 parameters. Build an import
hash table for the pseudo-relocation support version 2. When it is
built, walk the relocations only once at the end; when it is not, do
not build a fixup when the symbol isn't part of an import table.
Issue the associated warning only after a first fixup is built.
(tmp_seq2): Delete.
(make_singleton_name_imp): Likewise.
(make_import_fixup_mark): Return const char * and a stable string.
(make_import_fixup_entry): Do not deal with the pseudo-relocation
support version 2.
(make_runtime_pseudo_reloc): Factor out code and fix formatting.
(pe_create_import_fixup): Add 5th parameter. Clearly separate the
pseudo-relocation support version 2 from the rest. Fix formatting.
* pe-dll.h (pe_walk_relocs_of_symbol): Delete.
(pe_find_data_imports): Declare.
(pe_create_import_fixup): Add 5th parameter.
* pep-dll.c (pe_data_import_dll): Delete.
(pe_find_data_imports): Define.
(pe_walk_relocs_of_symbol): Delete.
* pep-dll.h (pep_walk_relocs_of_symbol): Delete.
(pep_find_data_imports): Declare.
(pep_create_import_fixup): Add 5th parameter.
* ld.texinfo (--enable-auto-import): Adjust to new implementation.
PR binutils/22875: HPPA/ELF: Also fail with relocation placeholders
Do not consider R_PARISC_UNIMPLEMENTED placeholder relocation entries of
the `elf_hppa_howto_table' table valid in `info_to_howto' HPPA handlers.
Instead issue an unsupported relocation type error and return a NULL
howto as with relocations whose number is R_PARISC_UNIMPLEMENTED or
beyond.
bfd/
* elf-hppa.h (elf_hppa_info_to_howto): Also return
unsuccessfully for unimplemented relocations.
(elf_hppa_info_to_howto_rel): Likewise.
Complement commit f3185997ac09 ("PR 22875: Stop strip corrupting unknown
relocs"), <https://sourceware.org/ml/binutils/2018-02/msg00445.html>,
and also set the `bfd_error_bad_value' error and report an unsupported
relocation type if a howto lookup fails with the i860 backend, fixing a
confusing `no error' error message and removing a binutils test failure:
failed with: <.../binutils/strip-new: tmpdir/bintest.o: no error>, expected: <.* bad value>
.../binutils/strip-new: tmpdir/bintest.o: no error
FAIL: binutils-all/strip-13
with the `i860-stardent-elf' target.
bfd/
* elf32-i860.c (lookup_howto): Add `abfd' parameter. Set the
`bfd_error_bad_value' error and call `_bfd_error_handler' on a
howto lookup failure.
(elf32_i860_reloc_type_lookup): Adjust `lookup_howto' call
accordingly.
(elf32_i860_info_to_howto_rela): Likewise.
(elf32_i860_relocate_splitn): Likewise.
(elf32_i860_relocate_pc16): Likewise.
(elf32_i860_relocate_pc26): Likewise.
(elf32_i860_relocate_section): Likewise.
PR binutils/22875: Visium/ELF: Prevent an out-of-bounds howto table access
Prevent an out-of-bounds `visium_elf_howto_table' table access in
`visium_info_to_howto_rela' by using the size of the table rather than
R_VISIUM_max to determine the number of entries in the contiguous
regular Visium relocation range defined and described in the table.
bfd/
* elf32-visium.c (visium_info_to_howto_rela): Correct the range
check for `visium_elf_howto_table' table access.
PR binutils/22875: IQ2000/ELF: Prevent an out-of-bounds howto table access
Prevent an out-of-bounds `iq2000_elf_howto_table' table access in
`iq2000_info_to_howto_rela' by using the size of the table rather than
R_IQ2000_max to determine the number of entries in the contiguous
regular IQ2000 relocation range defined and described in the table.
bfd/
* elf32-iq2000.c (iq2000_info_to_howto_rela): Correct the range
check for `iq2000_elf_howto_table' table access.
PR binutils/22875: FRV/ELF: Prevent an out-of-bounds howto table access
Prevent an out-of-bounds `elf32_frv_howto_table' table access in
`frv_info_to_howto_rela' by using the size of the table rather than
R_FRV_max to determine the number of entries in the contiguous regular
FRV relocation range defined and described in the table.
bfd/
* elf32-frv.c (frv_info_to_howto_rela): Correct the range check
for `elf32_frv_howto_table' table access.
PR binutils/22875: MIPS/ELF: Also fail with relocation placeholders
Do not consider placeholder EMPTY_HOWTO relocation entries valid in
`rtype_to_howto' MIPS handlers. Instead issue an unsupported relocation
type error and return a NULL howto as with relocations outside the three
ISA-specific min-max ranges.
bfd/
* elf32-mips.c (mips_elf32_rtype_to_howto): Also return
unsuccessfully for placeholder howtos.
* elf64-mips.c (mips_elf64_rtype_to_howto): Likewise.
* elfn32-mips.c (mips_elf_n32_rtype_to_howto): Likewise.
MIPS/LD/testsuite: Correct LD emulations for `mips*-*-kfreebsd*-gnu'
Complement commit 86b24e15c45b ("MIPS/LD/testsuite: Correct
comm-data.exp test ABI/emul/endian arrangement") and set LD emulations
correctly for `mips*-*-kfreebsd*-gnu' targets in comm-data.exp, removing
test suite failures:
FAIL: MIPS o32/copyreloc common symbol override test (auxiliary shared object build)
FAIL: MIPS o32/copyreloc common symbol override test
FAIL: MIPS o32/nocopyreloc common symbol override test (auxiliary shared object build)
FAIL: MIPS o32/nocopyreloc common symbol override test
ld/
* testsuite/ld-mips-elf/comm-data.exp: Correct support for
`mips*-*-kfreebsd*-gnu' targets.
Fix problem where mixed section types can cause internal error during a -r link.
During a -r (or --emit-relocs) link, if two sections had the same name but
different section types, gold would put relocations for both sections into
the same relocation section even though the data sections remained separate.
For .eh_frame sections, when one section is PROGBITS and another is
X86_64_UNWIND, we really should be using the UNWIND section type and
combining the sections anyway. For other sections, we should be
creating one relocation section for each output data section.
gold/
PR gold/23016
* incremental.cc (can_incremental_update): Check for unwind section
type.
* layout.h (Layout::layout): Add sh_type parameter.
* layout.cc (Layout::layout): Likewise.
(Layout::layout_reloc): Create new output reloc section if data
section does not already have one.
(Layout::layout_eh_frame): Check for unwind section type.
(Layout::make_eh_frame_section): Use unwind section type for .eh_frame
and .eh_frame_hdr.
* object.h (Sized_relobj_file::Shdr_write): New typedef.
(Sized_relobj_file::layout_section): Add sh_type parameter.
(Sized_relobj_file::Deferred_layout::Deferred_layout): Add sh_type
parameter.
* object.cc (Sized_relobj_file::check_eh_frame_flags): Check for
unwind section type.
(Sized_relobj_file::layout_section): Add sh_type parameter; pass it
to Layout::layout.
(Sized_relobj_file::do_layout): Make local copy of sh_type.
Force .eh_frame sections to unwind section type.
Pass sh_type to layout_section.
(Sized_relobj_file<size, big_endian>::do_layout_deferred_sections):
Pass sh_type to layout_section.
* output.cc (Output_section::Output_section): Initialize reloc_section_.
* output.h (Output_section::reloc_section): New method.
(Output_section::set_reloc_section): New method.
(Output_section::reloc_section_): New data member.
* target.h (Target::unwind_section_type): New method.
(Target::Target_info::unwind_section_type): New data member.
* testsuite/Makefile.am (pr23016_1, pr23016_2): New test cases.
* testsuite/Makefile.in: Regenerate.
* testsuite/testfile.cc: Add unwind_section_type.
* testsuite/pr23016_1.sh: New test script.
* testsuite/pr23016_1a.s: New source file.
* testsuite/pr23016_1b.s: New source file.
* testsuite/pr23016_2.sh: New test script.
* testsuite/pr23016_2a.s: New source file.
* testsuite/pr23016_2b.s: New source file.
Weimin Pan [Wed, 28 Mar 2018 19:23:48 +0000 (13:23 -0600)]
Fix infinite recursion when printing static member with typedef
The original problem was fixed (see related PR 22242). But using a typedef
as the declared type for a static member variable, as commented in this PR,
is still causing gdb to get into infinite loop when printing the static
member's value. This problem can be reproduced as follows:
% cat t.cc
class A {
typedef A type;
public:
bool operator==(const type& other) { return true; }
static const type INSTANCE;
};
const A A::INSTANCE;
int main() {
A a;
if (a == A::INSTANCE) {
return -1;
}
return 0;
}
% g++ -g t.cc
% gdb -ex "start" -ex "p a" a.out
The fix is rather trivial - in cp_print_static_field(), should call
check_typedef() to get the static member's real type and use it to
check whether it's a struct or an array.
As Simon suggested, I've added a new test case to the testsuite
and am passing the original type, not the real type, as argument
to both cp_print_value_fields() and val_print().
Re-tested on both aarch64-linux-gnu and amd64-linux-gnu. No regressions.
Simon Marchi [Fri, 30 Mar 2018 21:18:54 +0000 (17:18 -0400)]
Use std::vector in uploaded_tp
This patch changes the VEC(char_ptr) fields in uploaded_tp to use
std::vector<char *>. At first, I wanted to creep in more changes, like
using std::string, but it was making the patch too big and less focused,
so I decided to keep it to just that.
It also looks like the strings in those vectors are never free'd. If
so, we can fix that in another patch.
gdb/ChangeLog:
* tracepoint.h (struct uploaded_tp): Initialize fields.
<actions, step_actions, cmd_strings>: Change type to
std::vector<char *>.
* tracepoint.c (get_uploaded_tp): Allocate with new.
(free_uploaded_tps): Free with delete.
(parse_tracepoint_definition): Adjust to std::vector change.
* breakpoint.c (read_uploaded_action): Likewise.
(create_tracepoint_from_upload): Likewise.
* ctf.c (ctf_write_uploaded_tp): Likewise.
(SET_ARRAY_FIELD): Likewise.
* tracefile-tfile.c (tfile_write_uploaded_tp): Likewise.
Tom Tromey [Wed, 28 Mar 2018 21:35:46 +0000 (15:35 -0600)]
Remove free_dwo_file_cleanup
This removes free_dwo_file_cleanup, the last cleanup in dwarf2read.c.
This is replaced with a unique_ptr; which, despite the fact that a
dwo_file is obstack-allocated, seemed like the best fit.
* dwarf2read.c (class free_cached_comp_units): New class.
(dw2_instantiate_symtab, dwarf2_build_psymtabs_hard): Use it.
(free_cached_comp_units): Remove function.
Tom Tromey [Wed, 28 Mar 2018 21:49:24 +0000 (15:49 -0600)]
Remove make_cleanup_unpush_target
This removes make_cleanup_unpush_target, replacing it with a
unique_ptr. This may seem odd, because the object in question is not
actually freed, but unique_ptr provided the necessary functionality.
Remove a duplicate `unsupported relocation type' message and the setting
of the `bfd_error_bad_value' error from `mips_elf32_rtype_to_howto',
added with commit f3185997ac09 ("PR 22875: Stop strip corrupting unknown
relocs"), <https://sourceware.org/ml/binutils/2018-02/msg00445.html>.
This message is already produced and the `bfd_error_bad_value' error set
by `mips_elf32_rtype_to_howto' before a NULL howto is returned, so there
is no need to repeat these actions here.
bfd/
* elf32-mips.c (mips_info_to_howto_rel): Remove the calls to
`_bfd_error_handler' and to set the `bfd_error_bad_value' error.
BFD/PA: Remove ATTRIBUTE_UNUSED from `elf_hppa_info_to_howto_rel'
Remove ATTRIBUTE_UNUSED annotation from the `abfd' parameter in
`elf_hppa_info_to_howto' now that commit f3185997ac09 ("PR 22875: Stop
strip corrupting unknown relocs"),
<https://sourceware.org/ml/binutils/2018-02/msg00445.html>, made it
used.
bfd/
* elf-hppa.h (elf_hppa_info_to_howto_rel): Remove
ATTRIBUTE_UNUSED from `abfd'.
MIPS/BFD: Call `mips_elf32_rtype_to_howto' directly with o32
Call `mips_elf32_rtype_to_howto' directly rather than via the
`->elf_backend_mips_rtype_to_howto' method in the o32 backend,
complementing commit 861fb55ab50a ("Defer allocation of R_MIPS_REL32 GOT
slots"), <https://sourceware.org/ml/binutils/2008-08/msg00096.html>, and
reverting the change to `mips_info_to_howto_rel' originally made with
commit 0a44bf6950b3 ("mips-vxworks support"),
<https://sourceware.org/ml/binutils/2006-03/msg00179.html>.
With `mips_vxworks_rtype_to_howto' gone there is a single backend method
used across all o32 targets, so there in no need for the indirection and
the associated extra cost. This also makes the o32 backend consistent
with the n32 and n64 backends.
bfd/
* elf32-mips.c (mips_info_to_howto_rel): Call
`mips_elf32_rtype_to_howto' directly rather than via
`->elf_backend_mips_rtype_to_howto'.
Renlin Li [Wed, 28 Mar 2018 17:06:05 +0000 (18:06 +0100)]
[2/2][LD][AARCH64]Add BFD_RELOC_AARCH64_TLSLE_LDST8/16/32/64_TPREL_LO12 support in LD.
This patch adds the following relocation support into binutils bfd linker.
BFD_RELOC_AARCH64_TLSLE_LDST16_TPREL_LO12,
BFD_RELOC_AARCH64_TLSLE_LDST16_TPREL_LO12_NC,
BFD_RELOC_AARCH64_TLSLE_LDST32_TPREL_LO12,
BFD_RELOC_AARCH64_TLSLE_LDST32_TPREL_LO12_NC,
BFD_RELOC_AARCH64_TLSLE_LDST64_TPREL_LO12,
BFD_RELOC_AARCH64_TLSLE_LDST64_TPREL_LO12_NC,
BFD_RELOC_AARCH64_TLSLE_LDST8_TPREL_LO12,
BFD_RELOC_AARCH64_TLSLE_LDST8_TPREL_LO12_NC.
Those relocations includes both ip64 and ilp32 variant.
Renlin Li [Wed, 28 Mar 2018 17:03:55 +0000 (18:03 +0100)]
[1/2][GAS][AARCH64]Add BFD_RELOC_AARCH64_TLSLE_LDST8/16/32/64_TPREL_LO12 support in GAS.
This patch adds the following relocation support into binutils gas.
BFD_RELOC_AARCH64_TLSLE_LDST16_TPREL_LO12,
BFD_RELOC_AARCH64_TLSLE_LDST16_TPREL_LO12_NC,
BFD_RELOC_AARCH64_TLSLE_LDST32_TPREL_LO12,
BFD_RELOC_AARCH64_TLSLE_LDST32_TPREL_LO12_NC,
BFD_RELOC_AARCH64_TLSLE_LDST64_TPREL_LO12,
BFD_RELOC_AARCH64_TLSLE_LDST64_TPREL_LO12_NC,
BFD_RELOC_AARCH64_TLSLE_LDST8_TPREL_LO12,
BFD_RELOC_AARCH64_TLSLE_LDST8_TPREL_LO12_NC.
Those relocations includes both ip64 and ilp32 variant.
Jan Beulich [Wed, 28 Mar 2018 12:25:07 +0000 (14:25 +0200)]
x86: drop VecESize
It again can be inferred from other information.
The vpopcntd templates all need to have Dword added to their memory
operands; the lack thereof was actually a bug preventing certain Intel
syntax code to assemble, so test cases get extended.
Jan Beulich [Wed, 28 Mar 2018 12:22:00 +0000 (14:22 +0200)]
x86: don't show suffixes for to-scalar-int conversion insns
In the course of folding their patterns (possible now that the pointless
and partly even bogus VecESize are no longer in the way) I've noticed
that vcvt*2usi, other than their vcvt*2si counterparts, don't allow for
any suffixes. As that is supposedly intentional, make the disassembler
consistently omit suffixes for all to-scalar-int conversion insns.
Eric Botcazou [Wed, 28 Mar 2018 10:17:15 +0000 (12:17 +0200)]
PR ld/22972 on SPARC.
This is a regression for the corner case of a hidden symbol in a PIC/PIE
binary which is subject to both a new-style GOTDATA relocation and an
old-style GOT relocation. In this case, depending on the link order,
the R_SPARC_RELATIVE dynamic relocation for the GOT slot needed because
of the old-style relocation can be replaced with R_SPARC_NONE coming
from the GOTDATA relocation.
The fix simply records whether an old-style GOT relocation is seen for a
symbol and prevents the R_SPARC_NONE from being generated in this case.
bfd/
* elfxx-sparc.c (struct _bfd_sparc_elf_link_hash_entry): Add new flag
has_old_style_got_reloc.
(_bfd_sparc_elf_check_relocs) <GOT relocations>: Set it for old-style
relocations. Fix a couple of long lines.
(_bfd_sparc_elf_relocate_section) <R_SPARC_GOTDATA_OP>: Do not generate
a R_SPARC_NONE for the GOT slot if the symbol is also subject to
old-style GOT relocations.
ld/
* testsuite/ld-sparc/sparc.exp: Add test for mixed GOTDATA/GOT relocs.
* testsuite/ld-sparc/gotop-hidden.c: New file.
* testsuite/ld-sparc/got-hidden32.s: Likewise.
* testsuite/ld-sparc/got-hidden64.s: Likewise.
* testsuite/ld-sparc/pass.out: Likewise.
Tom Tromey [Thu, 22 Mar 2018 16:31:41 +0000 (10:31 -0600)]
Remove cleanups from gdb_readline_wrapper
This removes some cleanups from gdb_readline_wrapper by changing the
existing gdb_readline_wrapper_cleanup struct to have a constructor and
destructor, and then changing gdb_readline_wrapper to simply
instantiate it on the stack.
Tom Tromey [Wed, 14 Mar 2018 22:38:02 +0000 (16:38 -0600)]
C++-ify typedef hash
This changes the typedef_hash_table structure to be a C++ class. It
adds constructors and destructors and changes some functions to be
methods of the class. Then it changes the various users of this class
to adapt. This allows for the removal of some cleanups.
* typeprint.h (struct type_print_options) <local_typedefs,
global_typedefs>: Remove "struct" keyword.
(class typedef_hash_table): New class.
(recursively_update_typedef_hash, add_template_parameters)
(create_typedef_hash, free_typedef_hash, copy_typedef_hash)
(find_typedef_in_hash): Don't declare.
* typeprint.c (struct typedef_hash_table): Move to typeprint.h.
(typedef_hash_table::recursively_update): Rename from
recursively_update_typedef_hash. Now a member.
(typedef_hash_table::add_template_parameters): Rename from
add_template_parameters. Now a member.
(typedef_hash_table::typedef_hash_table): Now a constructor;
rename from create_typedef_hash.
(typedef_hash_table::~typedef_hash_table): Now a destructor;
rename from free_typedef_hash.
(do_free_typedef_hash, make_cleanup_free_typedef_hash)
(do_free_global_table): Remove.
(typedef_hash_table::typedef_hash_table): New constructor; renamed
from copy_type_recursive.
(create_global_typedef_table): Remove.
(typedef_hash_table::find_global_typedef): Now a member of
typedef_hash_table.
(typedef_hash_table::find_typedef): Rename from
find_typedef_in_hash; now a member.
(whatis_exp): Update.
* extension.h (struct ext_lang_type_printers): Add constructor and
destructor.
(start_ext_lang_type_printers, free_ext_lang_type_printers): Don't
declare.
* extension.c (ext_lang_type_printers::ext_lang_type_printers):
Now a constructor; rename from start_ext_lang_type_printers.
(ext_lang_type_printers): Now a destructor; rename from
free_ext_lang_type_printers.
* c-typeprint.c (find_typedef_for_canonicalize, c_print_type_1):
Update.
(c_type_print_base_struct_union): Update. Remove cleanups.
Tom Tromey [Tue, 27 Mar 2018 14:55:13 +0000 (08:55 -0600)]
Include <cmath> in dwarf-index-write.c
On x86-64 Fedora 26, when building with the system gcc, I get:
../../binutils-gdb/gdb/dwarf-index-write.c: In member function ‘void debug_names::build()’:
../../binutils-gdb/gdb/dwarf-index-write.c:705:13: error: ‘pow’ is not a member of ‘std’
There are actually more messages, but this is sufficient to show the
problem.
The fix is to include <cmath>.
I'm checking this in as obvious. Tested by building.
Joel Brobecker [Tue, 27 Mar 2018 14:17:45 +0000 (09:17 -0500)]
set varsize-limit: New GDB setting for maximum dynamic object size
This is a command we somehow forgot to contribute at the time the Ada
language was first contributed to the FSF. This command allows
the user to change the maximum size we allow when reading memory
from dynamic objects (the default is 65536 bytes).
At the moment, this limit is only used by Ada, and so the implementation
is kept inside ada-lang.c. However, it is conceivable that other language
might want to use it also to handle the same kind of issues; for instance,
this might be useful when handling dynamic types in C. So the name
of the setting was made language-neutral, to allow for this.
Note that an alias for "set var" needs to be introduced as well.
We are not adding a test for that, since this is a feature that is
already exercized by numerous existing tests.
gdb/ChangeLog
* NEWS: Add entry describing new "set|show varsize-limit" command.
* ada-lang.c (_initialize_ada_language): Add "set/show varsize-limit"
command.
* printcmd.c (_initialize_printcmd): Add "set var" alias of
"set variable".
Simon Marchi [Tue, 27 Mar 2018 14:07:44 +0000 (10:07 -0400)]
Move DWARF index-related things to a separate file
I want to add a DWARF index-related feature (automatically produce index
files when loading objfiles in GDB), but I don't want to add many
hundred lines to the already too big dwarf2read.c. I thought it would
be a logical split to move everything related to the DWARF index to its
own file.
I first tried to move everything that reads and writes DWARF indices to
a separate file, but found that the "read" part is a little bit
entangled with the rest of dwarf2read.c, so the line is hard to draw
about where to split. The write part is quite isolated though, so I
moved this part to a new file, dwarf-index-write.c. Some things are
necessary to both reading and writing indices, so I placed them in
dwarf-index-common.{c,h}. The idea would be to have a
dwarf-index-read.c eventually that would use it too (for now that code
is still in dwarf2read.c).
This required moving some things to a new dwarf2read.h header, so they
can be read by the code that writes the index.
The patch is big in number of lines, but it's all existing code being
moved around. The only changes are that some functions are not static
anymore, a declaration is added in a .h file, and therefore the comment
is moved there.
I built-tested it with a little and big endian target.
This patch is also available on the users/simark/split-dwarf2read
branch.
Joel Brobecker [Tue, 27 Mar 2018 13:57:16 +0000 (08:57 -0500)]
problem looking up some symbols when they have a linkage name
This patch fixes a known failure in gdb.ada/maint_with_ada.exp
(maintenance check-psymtabs). Another way to witness the same
issue is by considering the following Ada declarations...
type Wrapper is record
A : Integer;
end record;
u00045 : constant Wrapper := (A => 16#060287af#);
pragma Export (C, u00045, "symada__cS");
... which declares a variable name "u00045" but with a linkage
name which is "symada__cS". This variable is a record with one
component, the Ada equivalent of a struct with one field in C.
Trying to print that variable's value currently yields:
(gdb) p /x <symada__cS>
'symada(char, signed)' has unknown type; cast it to its declared type
This indicates that GDB was only able to find the minimal symbol,
but not the full symbol. The expected output is:
(gdb) print /x <symada__cS>
$1 = (a => 0x60287af)
The error message gives a hint about what's happening: We processed
the symbol through gdb_demangle, which in the case of this particular
symbol name, ends up matching the C++ naming scheme. As a result,
the demangler transforms our symbol name into 'symada(char, signed)',
thus breaking Ada lookups.
This patch fixes the issue by first introducing a new language_defn
attribute called la_store_sym_names_in_linkage_form_p, which is a boolean
to be set to true for the few languages that do not want their symbols
to have their names stored in demangled form, and false otherwise.
We then use this language attribute to skip the call to gdb_demangle
for all languages whose la_store_sym_names_in_linkage_form_p is true.
In terms of the selection of languages for which the new attribute
is set to true, the selection errs on the side of preserving the
existing behavior, and only changes the behavior for the languages
where we are certain storing symbol names in demangling form is not
needed. It is conceivable that other languages might be in the same
situation, but I not knowing in detail the symbol name enconding
strategy, I decided to play it safe and let other language maintainers
potentially adjust their language if it makes sense to do so.
gdb/ChangeLog:
PR gdb/22670
* dwarf2read.c (dwarf2_physname): Do not return the demangled
symbol name if the CU's language stores symbol names in linkage
format.
* language.h (struct language_defn)
<la_store_sym_names_in_linkage_form_p>: New field. Adjust
all instances of this struct.
Tom Tromey [Fri, 23 Mar 2018 16:40:00 +0000 (10:40 -0600)]
Remove verbose code from backtrace command
In https://sourceware.org/ml/gdb-patches/2017-06/msg00741.html,
Pedro asks:
> Doesn't the "info verbose on" bit affect frame filters too?
The answer is that yes, it could. However, it's not completely
effective, because the C code can't guess how many frames might need
to be unwound to satisfy the request -- a frame filter will request as
many frames as it needs.
Also, I tried removing this code from backtrace, and I think the
result is better without it. In particular, now the expansion line
occurs just before the frame that caused the expansion, like:
(gdb) bt no-filters
#0 0x00007ffff576cecd in poll () from /lib64/libc.so.6
Reading in symbols for ../../binutils-gdb/gdb/event-loop.c...done.
#1 0x00000000007ecc33 in gdb_wait_for_event (block=1)
at ../../binutils-gdb/gdb/event-loop.c:772
#2 0x00000000007ec006 in gdb_do_one_event ()
at ../../binutils-gdb/gdb/event-loop.c:347
#3 0x00000000007ec03e in start_event_loop ()
at ../../binutils-gdb/gdb/event-loop.c:371
Reading in symbols for ../../binutils-gdb/gdb/main.c...done.
#4 0x000000000086693d in captured_command_loop (
Reading in symbols for ../../binutils-gdb/gdb/exceptions.c...done.
data=0x0) at ../../binutils-gdb/gdb/main.c:325
So, I am proposing this patch to simply remove this code.
Tom Tromey [Fri, 23 Mar 2018 16:29:28 +0000 (10:29 -0600)]
Improve "backtrace" help text
This improves help text in stack.c in two ways. First, it removes
trailing newlines from various help strings. I think these are never
needed. Second, it adds a "Usage" line to the "backtrace" text, as
suggested by Pedro.
Tom Tromey [Tue, 25 Apr 2017 03:59:43 +0000 (21:59 -0600)]
Call wrap_hint in one more spot in py-framefilter.c
PR python/16486 notes that "bt" output is still wrapped differently
when a frame filter is in use. This patch brings it a bit closer by
adding one more wrap_hint call, in a place where stack.c does this as
well.
Tom Tromey [Tue, 25 Apr 2017 03:41:56 +0000 (21:41 -0600)]
Return EXT_LANG_BT_ERROR in one more spot in py-framefilter.c
While reading py-framefilter.c, I found one spot where an exception
could be caught but then not be turned into EXT_LANG_BT_ERROR. This
patch fixes this spot.
Tom Tromey [Tue, 25 Apr 2017 03:03:22 +0000 (21:03 -0600)]
Move some code later in backtrace_command_1
PR backtrace/15584 notes that some code in backtrace_command_1 is not
useful when frame filters are in use. This patch moves this code into
the no-frame-filters "if". This also removes the unused local
"trailing_level", which I noticed while moving the code around.
Tom Tromey [Tue, 25 Apr 2017 03:47:59 +0000 (21:47 -0600)]
Throw a "quit" on a KeyboardException in py-framefilter.c
If a C-c comes while the Python code for a frame filter is running, it
will be turned into a Python KeyboardException. It seems good for
this to be treated like a GDB quit, so this patch changes
py-framefilter.c to notice this situation and call throw_quit in this
case.
* gdb.python/py-framefilter.exp: Add test for KeyboardInterrupt.
* gdb.python/py-framefilter.py (name_error): New global.
(ErrorInName.function): Use name_error.
Tom Tromey [Tue, 25 Apr 2017 02:59:37 +0000 (20:59 -0600)]
Allow C-c to work in backtrace in more cases
PR cli/17716 notes that it is difficult to C-c (or "q" at a pagination
prompt) while backtracing using a frame filter. One reason for this
is that many places in py-framefilter.c use RETURN_MASK_ALL in a
try/catch.
This patch changes these spots to use RETURN_MASK_ERROR instead. This
is safe to do because this entire file is exception safe now.
Tom Tromey [Sun, 23 Apr 2017 17:07:26 +0000 (11:07 -0600)]
Remove EXT_LANG_BT_COMPLETED
While looking at the frame filter code, I noticed that
EXT_LANG_BT_COMPLETED is not really needed. Semantically there is no
difference between the "completed" and "ok" results. So, this patch
removes this constant.
Tom Tromey [Sun, 23 Apr 2017 17:03:57 +0000 (11:03 -0600)]
Allow hiding of some filtered frames
When a frame filter elides some frames, they are still printed by
"bt", indented a few spaces. PR backtrace/15582 notes that it would
be nice for users if elided frames could simply be dropped. This
patch adds this capability.
Tom Tromey [Sun, 23 Apr 2017 17:02:52 +0000 (11:02 -0600)]
Change backtrace_command_1 calling to use flags
The next patch will add more flags to backtrace_command_1; and rather
than add another boolean argument, this patch changes it to accept a
flags value.
Tom Tromey [Sun, 23 Apr 2017 16:54:33 +0000 (10:54 -0600)]
Rationalize "backtrace" command line parsing
The backtrace command has peculiar command-line parsing. In
particular, it splits the command line, then loops over the arguments.
If it sees a word it recognizes, like "full", it effectively drops
this word from the argument vector. Then, it pastes together the
remaining arguments, passing them on to backtrace_command_1, which in
turn passes the resulting string to parse_and_eval_long.
The documentation doesn't mention the parse_and_eval_long at all, so
it is a bit of a hidden feature that you can "bt 3*2". The strange
algorithm above also means you can "bt 3 * no-filters 2" and get 6
frames...
This patch changes backtrace's command line parsing to be a bit more
rational. Now, special words like "full" are only recognized at the
start of the command.
This also updates the documentation to describe the various bt options
individually.
Cary Coutant [Fri, 23 Mar 2018 16:03:34 +0000 (09:03 -0700)]
Fix case where IR file provides symbol visibility but replacement file does not.
In PR 22868, two IR files provide conflicting visibility for a symbol.
When a def with PROTECTED visibility is seen after a def with DEFAULT
visibility, gold does not override the visibility. Later, if the
replacement object define the symbol with DEFAULT visibility, the symbol
remains DEFAULT. This was caused by a recent change to allow multiply-defined
absolute symbols, combined with the fact that the plugin framework was using
SHN_ABS as the section index for placeholder symbols. The solution is to
use a real (but arbitrary) section index.
gold/
PR gold/22868
* plugin.cc (Sized_pluginobj::do_add_symbols): Use a real section
index instead of SHN_ABS for defined symbols.
* testsuite/Makefile.am (plugin_pr22868): New test case.
* testsuite/Makefile.in: Regenerate
* testsuite/plugin_pr22868.sh: New test script.
* testsuite/plugin_pr22868_a.c: New source file.
* testsuite/plugin_pr22868_b.c: New source file.
Keith Seitz [Mon, 26 Mar 2018 17:34:28 +0000 (10:34 -0700)]
Remove struct keyword from section_addr_info
Buildbot pointed out a failiure in windows-nat.c:
../../binutils-gdb/gdb/windows-nat.c:582:10: error: using typedef-name 'section_addr_info' after 'struct'
struct section_addr_info *addrs;
^~~~~~~~~~~~~~~~~
In file included from ../../binutils-gdb/gdb/windows-nat.c:49:0:
../../binutils-gdb/gdb/symfile.h:75:37: note: 'section_addr_info' has a previous declaration here
typedef std::vector<other_sections> section_addr_info;
^~~~~~~~~~~~~~~~~
A recursive grep of the sources for "struct section_addr_info" reveals one
additional reference in a comment. In both cases, this patch simply removes
the struct keyword.
gdb/ChangeLog:
* symfile.c (place_section): Remove "struct" from section_addr_info
in comment.
* windows-nat.c (struct safe_symbol_file_add_args) <addrs>: Remove
"struct" keyword from section_addr_info.
Pedro Alves [Sun, 25 Mar 2018 17:02:43 +0000 (18:02 +0100)]
eval.c: reverse minsym and sym
I noticed that in evaluate_funcall, where we handle
OP_VAR_MSYM_VALUE/OP_VAR_VALUE to figure out the symbol's name gets
the minimal_symbol/symbol backwards. Happens to be harmless in
practice because the symbol name is recorded in the common initial
sequence (in the general_symbol_info field).
H.J. Lu [Sun, 25 Mar 2018 12:02:17 +0000 (05:02 -0700)]
x86-64: Add ENDBR64 to the TLSDESC PLT entry
The TLSDESC entry in a lazy procedure linkage table is called indirectly
with "callq *(%rax)". This patch adds an ENDBR64 to support indirect
branch tracking in Intel CET. The TLSDESC PLT entry now looks like:
Weimin Pan [Fri, 3 Nov 2017 00:38:36 +0000 (18:38 -0600)]
aarch64: Make "info address" resolve TLS variables
TLS variables can't be resolved on aarch64-linux-gnu
Running the test case with upstream gdb shows two failures:
(1) Receiving different error messages when printing TLS variable before
program runs - because the ARM compiler does not emit dwarf attribute
DW_AT_location for TLS, the result is expected and the baseline may
need to be changed for aarch64.
(2) Using "info address" command on C++ static TLS object resulted in
"symbol unresolved" error - below is a snippet from the test case:
class K {
public:
static __thread int another_thread_local;
};
__thread int K::another_thread_local;
(gdb) info address K::another_thread_local
Symbol "K::another_thread_local" is unresolved.
This patch contains fix for (2).
Function info_address_command() handles the "info address" command and
calls lookup_minimal_symbol_and_objfile() to find sym's symbol entry in
mininal symbol table if SYMBOL_COMPUTED_OPS (sym) is false. Problem is
that function lookup_minimal_symbol_and_objfile() only looked up an
objfile's minsym ordinary hash table, not its demangled hash table, which
was the reason why the C++ name was not found.
The fix is to call lookup_minimal_symbol(), which already looks up entries
in both minsym's hash tables, to find names when traversing the object file
list in lookup_minimal_symbol_and_objfile().
Tested in both aarch64-linux-gnu and amd64-linux-gnu. No regressions.
Keith Seitz [Thu, 22 Mar 2018 23:49:48 +0000 (16:49 -0700)]
Add psymbols for nested types
c++/22968 involves the inability of ptype to find a type definition for
a type defined inside another type. I recently added some additional
support for nested type definitions, but I apparently overlooked psymbols.
The user reports that using -readnow fixes the problem:
$ gdb 22968 -ex "ptype Outer::Inner"
There is no field named Inner
$ gdb -readnow 22968 -ex "ptype Outer::Inner"
type = struct Outer::Inner {
<no data field>
}
We clearly did not find a psymbol for Outer::Inner because it was located
in another CU. This patch addresses this problem by scanning structs
for additional psymbols. Rust is already doing this.
With this patch, the identical result to "-readnow" is given (without
using `-readnow', of course).
gdb/ChangeLog:
PR c++/22968
* dwarf2read.c (scan_partial_symbols): Scan structs/classes for
nested type definitions for C++, too.
gdb/testsuite/ChangeLog:
PR c++/22968
* gdb.cp/subtypes.exp: New file.
* gdb.cp/subtypes.h: New file.
* gdb.cp/subtypes.cc: New file.
* gdb.cp/subtypes-2.cc: New file.
Cary Coutant [Fri, 23 Mar 2018 17:05:38 +0000 (10:05 -0700)]
Add --debug=plugin option to record plugin actions.
With --debug=plugin, gold will log all plugin actions to a log file,
and make copies of plugin-supplied replacement files in a temporary
directory created under the current working directory.
gold/
* debug.h (DEBUG_PLUGIN): New constant.
(DEBUG_ALL): Add DEBUG_PLUGIN.
(debug_string_to_enum): Likewise.
* plugin.cc (make_sized_plugin_object): Add filename parameter.
(Plugin_recorder): New class.
(Plugin_manager::~Plugin_manager): Delete recorder_.
(Plugin_manager::load_plugins): Create and initialize recorder_.
(Plugin_manager::claim_file): Record claimed and unclaimed files.
(Plugin_manager::make_plugin_object): Use object name as name for
plugin object, if available.
(Plugin_manager::add_input_file): Record replacement files.
(Sized_pluginobj::do_add_symbols): Record plugin symbols.
(Plugin_finish::run): Call Plugin_recorder::finish().
(make_sized_plugin_object): Add filename parameter and pass to
Sized_pluginobj constructor.
* plugin.h (Plugin::filename): New method.
(Plugin::recorder): New method.
(Plugin::recorder_): New data member.
* machoread.c (struct oso_el): Add a constructor. Don't define as
a typedef.
(macho_register_oso): Remove.
(macho_symtab_read): Take a std::vector.
(oso_el_compare_name): Now a std::sort comparator.
(macho_symfile_read_all_oso): Take a std::vector.
(macho_symfile_read): Use std::vector. Remove cleanups.
Nick Clifton [Fri, 23 Mar 2018 12:20:03 +0000 (12:20 +0000)]
Improve readelf's selection of a file start symbol when displaying a gnu build attribute.
* readelf.c (get_symbol_for_build_attribute): Skip ARM mapping
symbols.
(print_gnu_build_attribute_description): If no file start symbol
could be found, look for one two bytes into the file.
* testsuite/binutils-all/note-4-64.s: Set the address of the file
start symbol to two bytes into the file.
* testsuite/binutils-all/note-4-32.s: Likewise.
Andrew Burgess [Mon, 19 Mar 2018 00:33:50 +0000 (00:33 +0000)]
gdb: Fix testsuite issue in gdb.arch/amd64-disp-step-avx.exp
This test starts up and confirms that $xmm0 has the value 0, it then
modifies $xmm0 (in the inferior) and confirms that the new value can
be read (in GDB).
On some machines I was noticing that this test would occasionally
fail, and on investigation I believe that the reason for this is that
the test is linked as a dynamically linked executable and makes use of
the system libraries during startup. The reason that this causes
problems is that both the runtime linker and the startup code run
before main can, and do (on at least some platforms) make use of the
XMM registers.
In this commit I modify the test program slightly to allow it to be
linked statically, without using the startup libraries. Now by the
time GDB reaches the symbol main we have only executed one 'nop'
instruction, and the XMM registers should all have the value 0. I've
extended the test script to confirm that $xmm0 to $xmm15 are all
initially 0, and I also check that at the point after $xmm0 has been
modified, all the other XMM registers ($xmm1 to $xmm15) are still 0.
The test program is still linked against libc in order that we can
call the exit function, however, we now call _exit rather than exit in
order to avoid all of the usual cleanup that exit does. This clean up
tries to tear down things that are usually setup during the startup
code, but now this isn't called calling exit will just result in a
crash.
gdb/testsuite/ChangeLog:
* gdb.arch/amd64-disp-step-avx.S: Add '_start' label.
(done): Call '_exit' not 'exit' to avoid atexit handlers.
* gdb.arch/amd64-disp-step-avx.exp: Pass -static, and
-nostartfiles when compiling the test. Confirm that all registers
xmm0 to xmm15 are initially 0, and that xmm1 to xmm15 are 0 after.
Andrew Burgess [Mon, 19 Mar 2018 00:37:31 +0000 (00:37 +0000)]
gdb: Minor cleanup in some gdb.arch/* tests
A small number of tests incorrectly tried to pass -Wa,-g through to
GCC as an extra compile time flag, either to gdb_compile or
prepare_for_testing.
The problem is that the syntax used for passing the flags was
incorrect, and as a result these extra flags were being ignored.
Luckily, the 'debug' flag was being passed in each case anyway, which
means that the '-g' flag would already be added.
Given that all these tests pass 'debug', and the invalid flag has been
ignored for some time, I'm just removing the flags in this commit.
I've also changed the tests from using gdb_compile to
prepare_for_testing, which allows some extra code to be removed from a
couple of tests scripts.
There should be no change in the test results after this commit.
gdb/testsuite/ChangeLog:
* gdb.arch/amd64-disp-step-avx.exp: Remove unneeded assembler flag
option, syntax was wrong anyway.
* gdb.arch/arm-disp-step.exp: Likewise.
* gdb.arch/sparc64-regs.exp: Likewise.
* gdb.arch/amd64-disp-step.exp: Remove unneeded assembler flag
option, syntax was wrong anyway, switch to use
prepare_for_testing.
* gdb.arch/i386-disp-step.exp: Likewise.
Andreas Arnez [Fri, 23 Mar 2018 09:38:23 +0000 (10:38 +0100)]
Testsuite: fully migrate to use_gdb_stub convenience func
In the GDB test suite, there are still multiple invocations of
"target_info exists use_gdb_stub". However, the recommended way of
checking for use_gdb_stub is to call the convenience function of the same
name.
Replace these occurrences and just call "use_gdb_stub" instead.
This patch merges the masks for matching the stwux and stdux
instructions in rs6000-tdep.c:skip_prologue into a single mask that
only matches these two instructions.
Commit 72dd273062 fixed the warning described in PR tdep/18295, this
patch addresses the comment in the same PR indicating that the mask
was too permissive.
gdb/Changelog:
PR tdep/18295
* rs6000-tdep.c (skip_prologue): Match both stwux and stdux
a single mask.
projects / binutils.git / log