Commit Graph

99194 Commits

Author SHA1 Message Date
Andrew Eikum
2f1575ea6f Fix the size of the dos_message field in the internal_extra_pe_filehdr structure on hosts where sizeof(long) == 8.
* coff/internal.h (struct internal_extra_pe_filehdr): Use ints
	instead of longs to hold dos_message.
2019-10-29 08:02:34 +00:00
Alan Modra
b9d8f5601b Re: Optimise away eh_frame advance_loc 0
If we happen to get the fixed and variable parts of the advance_loc
in different frags, bad things happen when subtracting one from a
fr_fix of zero.

	PR 25125
	* dw2gencfi.c (output_cfi_insn): Don't allow DW_CFA_advance_loc4
	to be placed in a different frag to the rs_cfa.
2019-10-29 12:20:54 +10:30
GDB Administrator
d7b510baa8 Automatic date update in version.in 2019-10-29 00:00:34 +00:00
Christian Biesinger
87f34879e5 Add a string_view version of startswith
Makes sure that the string is longer than prefix, so that strncmp will
do the right thing even if the string is not null-terminated.

For use in my string_view conversion patch:
https://sourceware.org/ml/gdb-patches/2019-10/msg00030.html
https://gnutoolchain-gerrit.osci.io/r/c/binutils-gdb/+/125

gdb/ChangeLog:

2019-10-28  Christian Biesinger  <cbiesinger@google.com>

	* gdbsupport/common-utils.h (startswith): Add an overloaded version
	that takes gdb::string_view arguments.

Change-Id: I5389855de2fd70e7065a789a79374b0693651b71
2019-10-28 12:20:13 -05:00
Nick Clifton
1ee3542c0a Fix potentially illegal shift and assign operation in CSKY disassembler.
* csky-dis.c (csky_chars_to_number): Check for a negative
	count. Use an unsigned integer to construct the return value.
2019-10-28 16:45:55 +00:00
Nick Clifton
bbf9a0b5ee Fix buffer overrun in TIC30 disassembler.
* tic30-dis.c (OPERAND_BUFFER_LEN): Define.  Use as length of
	operand buffer.  Set value to 15 not 13.
	(get_register_operand): Use OPERAND_BUFFER_LEN.
	(get_indirect_operand): Likewise.
	(print_two_operand): Likewise.
	(print_three_operand): Likewise.
	(print_oar_insn): Likewise.
2019-10-28 16:15:34 +00:00
Nick Clifton
d1e304bc27 Stop potential illegal memory access in the NS32K disassembler.
* ns32k-dis.c (bit_extract): Add sanitiy check of parameters.
	(bit_extract_simple): Likewise.
	(bit_copy): Likewise.
	(pirnt_insn_ns32k): Ensure that uninitialised elements in the
	index_offset array are not accessed.
2019-10-28 15:44:23 +00:00
Nick Clifton
dee334510f Prevent an illegal memory access in the xgate disassembler.
* xgate-dis.c (print_insn): Fix decoding of the XGATE_OP_DYA
	operand.
2019-10-28 15:06:32 +00:00
GDB Administrator
12c3e9173e Automatic date update in version.in 2019-10-28 00:00:43 +00:00
GDB Administrator
e9c999a28f Automatic date update in version.in 2019-10-27 00:02:01 +00:00
John David Anglin
f612a98f9f Add some missing casts to suppress implicit cast warnings
PR gas/25121
	* config/tc-hppa.c (tc_gen_reloc): Cast some enums to int.
	(md_assemble): Likewise.
2019-10-26 16:48:59 -04:00
Alan Modra
6f69abb049 Optimise away eh_frame advance_loc 0
These can be generated when multiple cfi directives are emitted for an
instruction and the insn frag is closed off between directives, as
happens when listings are enabled.  No doubt the advance_loc of zero
could be avoided by backtracking over frags in dw2gencfi.c before
calling cfi_add_advance_loc, but that seems like more work than
cleaning up afterwards as this patch does.

Noticed when looking at the testcase in PR25125.

	PR 25125
	* dw2gencfi.c (output_cfi_insn): Don't output DW_CFA_advance_loc+0.
	* ehopt.c (eh_frame_estimate_size_before_relax): Return -1 for
	an advance_loc of zero.
	(eh_frame_relax_frag): Translate fr_subtype of 7 to size -1.
	(eh_frame_convert_frag): Handle fr_subtype of 7.  Abort on
	unexpected fr_subtype.
2019-10-26 20:19:28 +10:30
Tom de Vries
30baf67b65 [gdb] Fix more typos in comments (2)
Fix typos in comments.  NFC.

Tested on x86_64-linux.

gdb/ChangeLog:

2019-10-26  Tom de Vries  <tdevries@suse.de>

	* aarch64-linux-tdep.c: Fix typos in comments.
	* aarch64-tdep.c: Same.
	* ada-lang.c: Same.
	* amd64-nat.c: Same.
	* arc-tdep.c: Same.
	* arch/aarch64-insn.c: Same.
	* block.c: Same.
	* breakpoint.h: Same.
	* btrace.h: Same.
	* c-varobj.c: Same.
	* cli/cli-decode.c: Same.
	* cli/cli-script.c: Same.
	* cli/cli-utils.h: Same.
	* coff-pe-read.c: Same.
	* coffread.c: Same.
	* compile/compile-cplus-symbols.c: Same.
	* compile/compile-object-run.c: Same.
	* completer.c: Same.
	* corelow.c: Same.
	* cp-support.c: Same.
	* demangle.c: Same.
	* dwarf-index-write.c: Same.
	* dwarf2-frame.c: Same.
	* dwarf2-frame.h: Same.
	* eval.c: Same.
	* frame-base.h: Same.
	* frame.h: Same.
	* gdbcmd.h: Same.
	* gdbtypes.h: Same.
	* gnu-nat.c: Same.
	* guile/scm-objfile.c: Same.
	* i386-tdep.c: Same.
	* i386-tdep.h: Same.
	* infcall.c: Same.
	* infcall.h: Same.
	* linux-nat.c: Same.
	* m68k-tdep.c: Same.
	* macroexp.c: Same.
	* memattr.c: Same.
	* mi/mi-cmd-disas.c: Same.
	* mi/mi-getopt.h: Same.
	* mi/mi-main.c: Same.
	* minsyms.c: Same.
	* nat/aarch64-sve-linux-sigcontext.h: Same.
	* objfiles.h: Same.
	* ppc-linux-nat.c: Same.
	* ppc-linux-tdep.c: Same.
	* ppc-tdep.h: Same.
	* progspace.h: Same.
	* prologue-value.h: Same.
	* python/py-evtregistry.c: Same.
	* python/py-instruction.h: Same.
	* record-btrace.c: Same.
	* record-full.c: Same.
	* remote.c: Same.
	* rs6000-tdep.c: Same.
	* ser-tcp.c: Same.
	* sol-thread.c: Same.
	* sparc-sol2-tdep.c: Same.
	* sparc64-tdep.c: Same.
	* stabsread.c: Same.
	* symfile.c: Same.
	* symtab.h: Same.
	* target.c: Same.
	* tracepoint.c: Same.
	* tui/tui-data.h: Same.
	* tui/tui-io.c: Same.
	* tui/tui-win.c: Same.
	* tui/tui.c: Same.
	* unittests/rsp-low-selftests.c: Same.
	* user-regs.h: Same.
	* utils.c: Same.
	* utils.h: Same.
	* valarith.c: Same.
	* valops.c: Same.
	* valprint.c: Same.
	* valprint.h: Same.
	* value.c: Same.
	* value.h: Same.
	* varobj.c: Same.
	* x86-nat.h: Same.
	* xtensa-tdep.c: Same.

gdb/gdbserver/ChangeLog:

2019-10-26  Tom de Vries  <tdevries@suse.de>

	* linux-aarch64-low.c: Fix typos in comments.
	* linux-arm-low.c: Same.
	* linux-low.c: Same.
	* linux-ppc-low.c: Same.
	* proc-service.c: Same.
	* regcache.h: Same.
	* server.c: Same.
	* tracepoint.c: Same.
	* win32-low.c: Same.

gdb/stubs/ChangeLog:

2019-10-26  Tom de Vries  <tdevries@suse.de>

	* ia64vms-stub.c: Fix typos in comments.
	* m32r-stub.c: Same.
	* m68k-stub.c: Same.
	* sh-stub.c: Same.

gdb/testsuite/ChangeLog:

2019-10-26  Tom de Vries  <tdevries@suse.de>

	* gdb.base/bigcore.c: Fix typos in comments.
	* gdb.base/ctf-ptype.c: Same.
	* gdb.base/long_long.c: Same.
	* gdb.dwarf2/dw2-op-out-param.S: Same.
	* gdb.python/py-evthreads.c: Same.
	* gdb.reverse/i387-stack-reverse.c: Same.
	* gdb.trace/tfile.c: Same.
	* lib/compiler.c: Same.
	* lib/compiler.cc: Same.

Change-Id: I8573d84a577894270179ae30f46c48d806fc1beb
2019-10-26 09:55:32 +02:00
Alan Modra
1cc3da765d sort_gnu_build_notes fix
* objcopy.c (sort_gnu_build_notes): Correct sort of deleted
	note2.
2019-10-26 13:11:02 +10:30
GDB Administrator
25294a18fd Automatic date update in version.in 2019-10-26 00:00:44 +00:00
H.J. Lu
760ddef46d elfedit: Report unknown x86 feature
Report unknown x86 feature:

$ ./elfedit --disable-x86-feature foo a.out
elfedit: Error: Unknown x86 feature: foo
$

	* elfedit.c (elf_x86_feature): Report unknown x86 feature.
2019-10-25 15:13:21 -07:00
H.J. Lu
256878f318 elfedit.c: Replace BYTE_PUT with byte_put
Since BYTE_PUT is defined as

 #define BYTE_PUT(field, val) byte_put (field, val, sizeof (field))

use byte_put, instead of BYTE_PUT, to put 4-byte bitmask at ptr with
"byte_put (ptr, bitmask, 4)", instead of "BYTE_PUT (ptr, bitmask)", to
work with "unsigned char *ptr".

	* elfedit.c (update_gnu_property): Replace BYTE_PUT with byte_put.
2019-10-25 15:06:39 -07:00
Ali Tamur
1834d45f0f Fix find_charset_names.
The patch f2aec7f6d1 changed the return type of relocate_gdb_directory to
std::string, but the change is not reflected in find_charset_names function.
(Probably missed because the broken code is behind an #ifdef).

gdb/ChangeLog

	* charset.c (find_charset_names): Reflect API change.
2019-10-25 13:57:05 -07:00
Christian Biesinger
92174eeae8 Revert unintentional change in symtab.c
In the previous commit, I accidentally changed the wrong line;
this reverts it to what it should be.

gdb/ChangeLog:

2019-10-25  Christian Biesinger  <cbiesinger@google.com>

	* symtab.c (symbol_set_names): Revert unintentional change in the
	Ada case.

Change-Id: I9abf174927687e74c7435bd4607aab7f248c6e79
2019-10-25 14:47:19 -05:00
Christian Biesinger
5396ae1717 Don't make an extra copy + allocation of the demangled name
We can just keep around the malloc()-ed name we got from bfd and free
it later.

gdb/ChangeLog:

2019-10-25  Christian Biesinger  <cbiesinger@google.com>

	* symtab.c (struct demangled_name_entry): Change demangled name
	to a unique_xmalloc_ptr<char>, now that we don't allocate it as
	part of the struct anymore.
	(symbol_set_names): No longer obstack allocate + copy the demangled
	name, just store the allocated name from bfd.

Change-Id: Ie6ad50e1e1e73509f55d756f0a437897bb93e3b0
2019-10-25 14:09:02 -05:00
Nick Clifton
5c49f2cd78 Improve objcopy's note mergeing capabilities.
* objcopy.c (struct merged_note_section): New structure.  Used to
	chain together details of mergeable note sections.
	(is_merged_note_section): Rename to is_megreable_note_section and
	return true for note sections that use GNU_BUILD_ATTRS_SECTION_NAME
	as a prefix.
	(num_bytes): Delete
	(objcoopy_internal_note): Add padded_namesz field.
	(DEBUG_MERGE): New macro.  Set to non-zero to enable debugging of
	the note merging code.
	(gap_exists): Rename to overlaps_or_adjoins and return TRUE for
	overlapping notes or adjoining notes.
	(contained_by, is_deleted_note, is_version_note)
	(compare_gnu_build_notes, sort_gnu_build_notes): New functions.
	(merge_gnu_build_notes):  Rework.  Sort notes into a mergeable
	order first.  Merge them.  Then sort them into an ascending
	address order before writing them out.
	(copy_object): Handle more than one mergeable note section.
	* testsuite/binutils-all/note-2-32.d: Update for new merging
	behaviour.
	* testsuite/binutils-all/note-2-32.s: Likewise.
	* testsuite/binutils-all/note-2-64.d: Likewise.
	* testsuite/binutils-all/note-2-64.s: Likewise.
	* testsuite/binutils-all/note-3-32.d: Likewise.
	* testsuite/binutils-all/note-3-32.s: Likewise.
	* testsuite/binutils-all/note-3-64.d: Likewise.
	* testsuite/binutils-all/note-3-64.s: Likewise.
	* testsuite/binutils-all/note-4-32.d: Likewise.
	* testsuite/binutils-all/note-4-32.s: Likewise.
	* testsuite/binutils-all/note-4-64.d: Likewise.
	* testsuite/binutils-all/note-4-64.s: Likewise.
	* testsuite/binutils-all/note-6-32.s: New test source file.
	* testsuite/binutils-all/note-6-64.s: New test source file.
	* testsuite/binutils-all/note-6-32.d: New test driver file.
	* testsuite/binutils-all/note-6-64.d: New test driver file.
	* testsuite/binutils-all/objcopy.exp: Run the new test.
2019-10-25 16:46:07 +01:00
Nick Clifton
27cee81d06 Fix potential undefined behaviour in the RX disassembler.
* rx-dis.c (print_insn_rx): Use parenthesis to ensure correct
	access to opcodes.op array element.
2019-10-25 16:10:04 +01:00
Tom Tromey
93878f4717 Allow out-of-order reads of CIEs
Currently gdb has an assertion that requires CIEs to be read in the
order in which they appear in the debug info:

   gdb_assert (n < 1
               || cie_table->entries[n - 1]->cie_pointer < cie->cie_pointer);

This assertion ensures that the table will be sorted, which is
important because it is later searched using bsearch.

However, a customer provided an executable that causes this assertion
to trigger.  This executable causes decode_frame_entry_1 to call
decode_frame_entry to find the CIE, resulting in an out-of-order read.

I don't know a good way to construct a reproducer, but this can happen
if the FDE appears before its CIE.  See
https://sourceware.org/bugzilla/show_bug.cgi?id=16563

This patch fixes the problem by storing CIEs in an unordered map.  The
CIE table is discarded after the frame section is parsed, so this
seemed both simple and straightforward.

gdb/ChangeLog
2019-10-25  Tom Tromey  <tromey@adacore.com>

	* dwarf2-frame.c (dwarf2_cie_table): Now a typedef.
	(bsearch_cie_cmp, add_cie): Remove.
	(find_cie): Reimplement.
	(decode_frame_entry_1, decode_frame_entry): Change type.  Update.
	(dwarf2_build_frame_info): Update.

Change-Id: I4a99597fa4b1398a9d105b683a36d992d506485c
2019-10-25 08:29:44 -06:00
Tom Tromey
52c64cf72d gdbserver does not need xstrdup
gdbserver has its own implementation of xstrdup.  However, because
gdbserver links against libiberty now, I think this is not needed.
This patch removes it.

gdb/gdbserver/ChangeLog
2019-10-25  Tom Tromey  <tromey@adacore.com>

	* utils.c (xstrdup): Remove.

Change-Id: I2aa56d18d0f9af8e70a00dff431d2fda5705a5d5
2019-10-25 08:22:44 -06:00
Alan Modra
37a5888387 PR25125, relaxation chooses wrong branch size
The patch I made for PR12049 didn't test for a "negative" branch
properly.  "if (target < address)" ought to have been
"if (target < address + fragP->fr_fix)".  Rather than making that
change, this patch adds fragP->fr_fix into address earlier.  The patch
also avoids running into a bad interaction with the m68k
md_prepare_relax_scan by returning zero growth immediately, since the
adjusted target expression would result in a zero "aim".

	PR gas/25125
	PR gas/12049
	* write.c (relax_frag): Correct calculation of delta for
	positive branches where "stretch" would make the branch
	negative.  Return zero immediately in that case.  Correct
	TC_PCREL_ADJUST comment.
2019-10-25 23:19:24 +10:30
Tom de Vries
158da0d12a [gdb/testsuite] Use -wrap and $gdb_test_name in gdb_test_multiple calls (2)
Make gdb_test_multiple calls shorter by using new gdb_test_multiple variable
$gdb_test_name and new gdb_test_multiple pattern flag -wrap.

Tested on x86_64-linux.

gdb/testsuite/ChangeLog:

2019-10-25  Tom de Vries  <tdevries@suse.de>

	* gdb.reverse/sigall-precsave.exp: Use -wrap and $gdb_test_name in
	gdb_test_multiple calls.
	* gdb.reverse/sigall-reverse.exp: Same.
	* gdb.reverse/solib-precsave.exp: Same.
	* gdb.reverse/solib-reverse.exp: Same.
	* gdb.reverse/until-precsave.exp: Same.
	* gdb.reverse/until-reverse.exp: Same.

Change-Id: I67bb327d069dbc439410996bcfe6c7f905b2ca52
2019-10-25 10:48:12 +02:00
Alan Modra
30fe183248 PR4499, assign file positions assumes segment offsets increasing
This rewrites much of assign_file_positions_for_non_load_sections to
allow objcopy and strip to handle cases like that in PR4499 where
program headers were not in their usual position immediately after the
ELF file header, and PT_LOAD headers were not sorted by paddr.

	PR 4499
include/
	* elf/internal.h (struct elf_segment_map): Delete header_size.
	Add no_sort_lma and idx.
bfd/
	* elf-nacl.c (nacl_modify_segment_map): Set no_sort_lma for all
	PT_LOAD segments.
	* elf32-spu.c (spu_elf_modify_segment_map): Likewise on overlay
	PT_LOAD segments.
	* elf.c (elf_sort_segments): New function.
	(assign_file_positions_except_relocs): Use shortcuts to elfheader
	and elf_tdata.  Seek to e_phoff not sizeof_ehdr to write program
	headers.  Move PT_PHDR check..
	(assign_file_positions_for_non_load_sections): ..and code setting
	PT_PHDR p_vaddr and p_paddr, and code setting __ehdr_start value..
	(assign_file_positions_for_load_sections): ..to here.  Sort
	PT_LOAD headers.  Delete header_pad code.  Use actual number of
	headers rather than allocated in calculating size for program
	headers.  Don't assume program headers follow ELF file header.
	Simplify pt_load_count code.  Only set "off" for PT_LOAD or
	PT_NOTE in cores.
	(rewrite_elf_program_header): Set p_vaddr_offset for segments
	that include file and program headers.
	(copy_elf_program_header): Likewise, replacing header_size code.
2019-10-25 13:30:05 +10:30
Alan Modra
c0c121b01c readelf PT_PHDR check
When PT_PHDR isn't covered by a PT_LOAD header, p_vaddr in PT_PHDR
isn't valid but the value might just pass a vaddr test.  So test
p_offset as well.

	* readelf.c (process_program_headers): Check PT_PHDR p_offset
	as well as p_vaddr.  Use p_filesz, not p_memsz, in vaddr test.
2019-10-25 13:29:44 +10:30
GDB Administrator
e2ffdcea6c Automatic date update in version.in 2019-10-25 00:01:28 +00:00
H.J. Lu
7b71fc971b Call forget_cached_source_info to clear the stale source cache
Clear the stale source cache when re-reading symbols.

	PR gdb/25126
	* symfile.c (reread_symbols): Call forget_cached_source_info to
	clear the stale source cache.
2019-10-24 15:43:21 -07:00
Christian Biesinger
cbb5a2ea49 Remove python_has_threads check in configure.ac
The only use of python_has_threads has been removed in
commit 404f29021a

gdb/ChangeLog:

2019-10-24  Christian Biesinger  <cbiesinger@google.com>

	* configure: Rebuild.
	* configure.ac: Remove code that sets python_has_threads.

Change-Id: I75f1b873562bc2abc6f2db17699a3e82fcfd2de3
2019-10-24 14:14:07 -05:00
Christian Biesinger
71737c435d Simplify Python checks in configure.ac
The version checking code is not necessary. It is only used to define
HAVE_LIBPYTHON2_6 or HAVE_LIBPYTHON2_7, which is not used anywhere.

If a version check is desired, the PY_{MAJOR,MINOR}_VERSION macro from
the Python headers can be (and is) used, which does not require updating
configure.ac whenever a new Python version is released.

gdb/ChangeLog:

2019-10-24  Christian Biesinger  <cbiesinger@google.com>

	* config.in: Regenerate.
	* configure: Regenerate.
	* configure.ac: Remove the code that uses sed to get the python
	version and defines HAVE_LIBPYTHON2_6 / HAVE_LIBPYTHON2_7.

Change-Id: I07073870d9040c2bc8519882c8b3c1368edd4513
2019-10-24 13:47:37 -05:00
Tom de Vries
4ccdfbec50 [gdb/testsuite] Add -wrap pattern flag to gdb_test_multiple
Currently, in order to rewrite:
...
gdb_test <command> <pattern> <message>
...
using gdb_test_multiple, we get:
...
gdb_test_multiple <command> <message> {
    -re "\[\r\n\]*(?:<pattern>)\[\r\n\]+$gdb_prompt $" {
    	pass $gdb_test_name
    }
}
...

Add a '-wrap pattern flag to gdb_test_multiple, that wraps the regexp
pattern as gdb_test wraps its message argument.

This allows us to rewrite into the more compact:
...
gdb_test_multiple <command> <message> {
    -re -wrap <pattern> {
        pass $gdb_test_name
    }
}
...

Tested on x86_64-linux.

gdb/testsuite/ChangeLog:

2019-10-24  Tom de Vries  <tdevries@suse.de>

	* lib/gdb.exp (gdb_test_multiple): Add -wrap pattern flag.
	* gdb.reverse/step-precsave.exp: Rewrite gdb_test_multiple containing
	kfail using -wrap pattern flag and convenience variable
	gdb_test_name.

Change-Id: Ie42c97d5ab7acf6db351299ccd23a83540fe6e1a
2019-10-24 18:43:46 +02:00
Andrew Burgess
33d569b709 gdb/python: Return None from Progspace.block_for_pc on error
The documentation for Progspace.block_for_pc says:

  Return the innermost gdb.Block containing the given pc value. If the
  block cannot be found for the pc value specified, the function will
  return None.

However, the implementation actually throws an error for invalid
addresses, like this:

    (gdb) python print gdb.current_progspace ().block_for_pc (1)
    Traceback (most recent call last):
      File "<string>", line 1, in <module>
    RuntimeError: Cannot locate object file for block.
    Error while executing Python code.
    (gdb)

This has been the behaviour since the command was first added (when
the documentation was still as above) in this commit:

    commit f3e9a8177c
    Date:   Wed Feb 24 21:18:28 2010 +0000

Since that commit the code in question has moved around, but the
important parts are largely unchanged.  The function in question is
now in py-progspace.c:pspy_block_for_pc.

Examining the code shows that the real state is more complex than just
the function throws an error instead of returning None, instead the
real situation is:

  1. If we can't find a compilation unit for the $pc value then we
  throw an error, but

  2. If we can find a compilation unit, but can't find a block within
  the compilation unit for the $pc then return None.

I suspect for most users of the Python API this distinction is
irrelevant, and I propose that we standardise on one single failure
mechanism.

Given the function can currently return None in some cases, and is
documented to return None on error, I propose we make that the case
for all error paths, which is what this patch does.

As the Progspace.block_for_pc method is currently untested, I've added
some basic tests including for a call with an invalid $pc.

This is potentially an API breaking change, though an undocumented
part of the API.  Also, users should have been checking and handling a
None return value anyway, so my hope is that this shouldn't be too
disruptive.

gdb/ChangeLog:

	* python/py-progspace.c (pspy_block_for_pc): Return None for all
	error paths.

gdb/testsuite/ChangeLog:

	* gdb.python/py-progspace.exp: Add tests for the
	Progspace.block_for_pc method.

Change-Id: I9cea8d2132902bcad0013d1fd39080dd5423cc57
2019-10-24 15:27:02 +01:00
GDB Administrator
867c09822c Automatic date update in version.in 2019-10-24 00:00:24 +00:00
Tom Tromey
f16f7b7c74 Fix opcodes includes
Now that gdb can unconditionally use a -I pointing at the top of the
source tree, we can remove the ugly "../opcodes/" formulation that was
needed earlier.  This patch adds the -I and cleans up these includes.

gdb/ChangeLog
2019-10-23  Tom Tromey  <tom@tromey.com>

	* arc-tdep.c: Remove ".." from include.
	* frv-tdep.c: Remove ".." from include.
	* lm32-tdep.c: Remove ".." from include.
	* microblaze-tdep.c: Remove ".." from include.
	* or1k-tdep.h: Remove ".." from include.
	* s12z-tdep.c: Remove ".." from include.
	* Makefile.in (OPCODES_CFLAGS): Add comment.
	(TOP_CFLAGS): New variable.
	(INTERNAL_CFLAGS_BASE): Add TOP_CFLAGS.

Change-Id: I21428726d55f9fab0c9da90b56f6664f258cf91a
2019-10-23 15:16:49 -06:00
Tom Tromey
6999161a2a Move readline to the readline/readline subdirectory
readline turns out to be a bit of a stumbling block for the project to
move gdbsupport (and then gdbserver) to the top-level.

The issue is that readline headers are intended to be included with
names like "readline/readline.h".  To support this, gdb effectively
adds a -I option pointing to the top-level source directory -- but,
importantly, this option is not used when the system readline is used.

For gdbsupport, a -I option like this would always be needed, but that
in turn would break the system readline case.  This was PR build/17077,
fixed in commit a8a5dbcab8.

Previously, we had discussed this on the gdb-patches list in terms of
removing readline from the tree

    https://sourceware.org/ml/gdb-patches/2019-09/msg00317.html

However, Eli expressed some concerns, and Joel did as well (off-list).

Given those concerns, and the fact that a patch-free local readline is
relatively new in gdb (it was locally patched for years), I changed my
mind and decided to handle this situation by moving the readline
sources down a level.

That is, upstream readline is now in readline/readline, and the
top-level readline directory just contains the minimal configury
needed to build that.

This fixes the problem because, when gdb unconditionally adds a
-I$(top_srcdir), this will not find readline headers.  A separate -I
will be needed instead, which is exactly what's needed for
--with-system-readline.

gdb/ChangeLog
2019-10-23  Tom Tromey  <tom@tromey.com>

	* Makefile.in (READLINE_DIR): Update.

gdb/doc/ChangeLog
2019-10-23  Tom Tromey  <tom@tromey.com>

	* Makefile.in (READLINE_DIR): Update.

readline/ChangeLog
2019-10-23  Tom Tromey  <tom@tromey.com>

	Move old contents to readline/ subdirectory.
	* aclocal.m4, configure, configure.ac, .gitignore, Makefile.am,
	Makefile.in, README: New files.

Change-Id: Ice156a2ee09ea68722b48f64d97146d7428ea9e4
2019-10-23 15:16:48 -06:00
Tankut Baris Aktemur
12e7c35ec3 infcall: refactor 'call_function_by_hand_dummy'
Extract out the code region that reserves stack space to a separate
function.

Fix the comment of 'call_function_by_hand_dummy' to remove reference
to the NARGS argument that was removed in commit (e71585ffe2 "Use
gdb:array_view in call_function_by_hand & friends").

gdb/ChangeLog:
2019-10-23  Tankut Baris Aktemur  <tankut.baris.aktemur@intel.com>

	* infcall.c (call_function_by_hand_dummy): Fix the function
	comment.  And extract out a code section into...
	(reserve_stack_space): ...this new function.

Change-Id: I8938ef4134aff68a0a21724aaa2406bfe453438a
2019-10-23 20:58:42 +02:00
Tankut Baris Aktemur
37055cada8 infcall: remove unused parameter in 'value_arg_coerce'
Remove the unused SP parameter from the auxiliary function
'value_arg_coerce'.

gdb/ChangeLog:
2019-10-23  Tankut Baris Aktemur  <tankut.baris.aktemur@intel.com>

	* infcall.c (value_arg_coerce): Remove an unused parameter.
	(call_function_by_hand_dummy): Update the call to
	'value_arg_coerce'.

Change-Id: If324a1dda3fa5d4c145790b92bd3f656c00296f4
2019-10-23 20:40:02 +02:00
Tankut Baris Aktemur
39bcc47c7e infcall: move assertions in 'call_function_by_hand_dummy' to an earlier spot
This is a refactoring that performs type assertions on the callee
function at the beginning of 'call_function_by_hand_dummy' rather than
at a later point so that

- the checks are grouped together at the beginning of the function for
improved readability, and

- we don't have to align and push things on the stack only to find out
later that the function call is illegal.

gdb/ChangeLog:
2019-10-23  Tankut Baris Aktemur  <tankut.baris.aktemur@intel.com>

	* infcall.c (call_function_by_hand_dummy): Refactor.

Change-Id: I411ac083ac6a9ee6eb93c4b82393a81a4fc927be
2019-10-23 20:40:02 +02:00
Tankut Baris Aktemur
bd888c0fe2 Add myself to the gdb/MAINTAINERS write-after-approval list
gdb/ChangeLog:
2019-10-23  Tankut Baris Aktemur  <tankut.baris.aktemur@intel.com>

	* MAINTAINERS (Write After Approval): Add Tankut Baris Aktemur.
2019-10-23 20:40:01 +02:00
Christian Biesinger
14796d2978 Add a note for how to get the list of threads
It's not immediately obvious how to get the list of threads,
so add a note about that in the "Threads in Python" section.

gdb/doc/ChangeLog:

2019-10-23  Christian Biesinger  <cbiesinger@google.com>

	* python.texi (Threads In Python): Add a note for how to get the
	list of threads.

Change-Id: I0fef8a7aff161fc347c09052319048c907a6e8c3
2019-10-23 11:27:28 -05:00
Tom Tromey
c12d372d26 Check for sigprocmask in common.m4
I noticed that gdbsupport uses HAVE_SIGPROCMASK, but common.m4 does
not check for it.  This means that gdbserver may not compile some
gdbsupport code properly.  This patch fixes this error.

gdb/ChangeLog
2019-10-23  Tom Tromey  <tom@tromey.com>

	* configure: Rebuild.
	* configure.ac: Don't check for sigprocmask.
	* gdbsupport/common.m4 (GDB_AC_COMMON): Check for sigprocmask.

gdb/gdbserver/ChangeLog
2019-10-23  Tom Tromey  <tom@tromey.com>

	* configure, config.in: Rebuild.

Change-Id: I2c0a4dd2c376507b9483b38707a242382faa8163
2019-10-23 09:06:21 -06:00
Tom Tromey
4d0b984b97 Use m4_include, not sinclude in .m4 files
Pedro pointed out that sinclude does not error if a file is missing.
This patch changes gdb to only use m4_include, which seems more
correct.

gdb/ChangeLog
2019-10-23  Tom Tromey  <tom@tromey.com>

	* configure: Rebuild.
	* acinclude.m4: Use m4_include, not sinclude.

gdb/gdbserver/ChangeLog
2019-10-23  Tom Tromey  <tom@tromey.com>

	* configure: Rebuild.
	* acinclude.m4: Use m4_include, not sinclude.

gdb/testsuite/ChangeLog
2019-10-23  Tom Tromey  <tom@tromey.com>

	* configure: Rebuild.
	* aclocal.m4: Use m4_include, not sinclude.

Change-Id: I970362e0af7875f9f72796401126acf0ff6dba11
2019-10-23 09:06:20 -06:00
Nick Clifton
de6d8dc25c Fix typo in RX disassembler error messages.
* rx-dis.c (get_register_name): Fix spelling typo in error
	message.
	(get_condition_name, get_flag_name, get_double_register_name)
	(get_double_register_high_name, get_double_register_low_name)
	(get_double_control_register_name, get_double_condition_name)
	(get_opsize_name, get_size_name): Likewise.
2019-10-23 10:17:21 +01:00
Tom de Vries
7e785608ce [gdb/breakpoints] Fix fullname.exp when run from symlink dir
I run into this error with gdb.base/fullname.exp:
...
(gdb) file /data/gdb_versions/devel/build/gdb/testsuite/outputs/\
  gdb.base/fullname/fullname
Reading symbols from /data/gdb_versions/devel/build/gdb/testsuite/outputs/\
  gdb.base/fullname/fullname...
(gdb) break /data/gdb_versions/devel/build/gdb/testsuite/\
  outputs/gdb.base/fullname/tmp-fullname.c:21
No source file named /data/gdb_versions/devel/build/gdb/testsuite/outputs/\
  gdb.base/fullname/tmp-fullname.c.
Make breakpoint pending on future shared library load? (y or [n]) n
(gdb) FAIL: gdb.base/fullname.exp: set breakpoint by full path before loading symbols - built relative
...

The FAIL is due to this comparison in iterate_over_some_symtabs failing:
...
481                   if (FILENAME_CMP (real_path, fullname) == 0)
(gdb) p real_path
$2 = 0x1a201f0 "/data/gdb_versions/devel/build/gdb/testsuite/outputs/\
  gdb.base/fullname/tmp-fullname.c"
(gdb) p fullname
$3 = 0x1a1de80 "/home/vries/gdb_versions/devel/build/gdb/testsuite/outputs/\
  gdb.base/fullname/tmp-fullname.c"
...

The difference in pathnames is due to having a symlink dir:
...
$ ls -la /home/vries/gdb_versions
lrwxrwxrwx 1 vries users 18 26 jun  2018 /home/vries/gdb_versions -> /data/gdb_versions
...
and the test passses when eliminating it:
...
$ ( cd $(pwd -P); make check RUNTESTFLAGS=gdb.base/fullname.exp )
...

The FAIL is a regression from commit a0c1ffedcf "Only compute realpath when
basenames_may_differ is set".  Before, find_and_open_source was returning a
real-path, resulting in variable 'fullname' being the same as varible
'real_path' in the comparison listed above.  But after, that's no longer the
case.

Fix the FAIL by applying gdb_realpath on the fullname variable before the
comparison.

Tested on x86_64-linux.

I wasn't able to write a test-case.  The FAIL starts at:
...
$ cd build/gdb
$ mv testsuite testsuite.bla
$ ln -s testsuite.bla testsuite
...
but already this doesn't trigger it anymore:
...
$ cd build/gdb/outputs
$ mv outputs outputs.bla
$ ln -s outputs.bla outputs
...

gdb/ChangeLog:

2019-10-23  Tom de Vries  <tdevries@suse.de>

	PR breakpoints/24687
	* symtab.c (iterate_over_some_symtabs): Apply gdb_realpath on fullname.

Change-Id: I1ace62a234458781e958980f3b425edf1490df27
2019-10-23 09:33:46 +02:00
GDB Administrator
6a093f341b Automatic date update in version.in 2019-10-23 00:00:20 +00:00
Christian Biesinger
403772ef61 Make demangled_name_entry::language not a bitfield
Having it as a bitfield causes extra work, and this is not memory-sensitive.
Furthermore, once https://sourceware.org/ml/gdb-patches/2019-10/msg00812.html
lands, the bitfield won't even save any memory at all.

gdb/ChangeLog:

2019-10-22  Christian Biesinger  <cbiesinger@google.com>

	* symtab.c (struct demangled_name_entry) <language>: Change from
	bitfield to regular variable.

Change-Id: I4ea31d1cfcbe0f09a09bd058cd304862308dc388
2019-10-22 15:36:37 -05:00
Christian Biesinger
3a49427939 Fix compile error & incorrect push
I accidentally pushed the wrong version of the patch for commit
7bb4305982 (where the review
comments were not fixed), and I did a bad conflict resolution
for ccb1ba6229 leading to a
compile error when libxxhash is available. This fixes both
issues.

gdb/ChangeLog:

2019-10-22  Christian Biesinger  <cbiesinger@google.com>

	* symtab.c (struct demangled_name_entry): Add a constructor.
	(free_demangled_name_entry): New function to call the destructor
	for demangled_name_entry.
	(create_demangled_names_hash): Pass free_demangled_name_entry to
	htab_create_alloc.
	(symbol_set_names): Call placement new for demangled_name_entry.
	* utils.c: No longer include xxhash.h here, now that fast_hash
	is inlined in the header.
	* utils.h: Instead, include it here.

Change-Id: If776099d39a65a12733d42efcb859feca1b07a39
2019-10-22 13:41:12 -05:00
Christian Biesinger
ccb1ba6229 Use libxxhash for hashing, if present
XXHash is faster than htab_hash_string:
------------------------------------------------------------
Benchmark                     Time           CPU Iterations
------------------------------------------------------------
BM_xxh3                      11 ns         11 ns   65887249
BM_xxh32                     19 ns         19 ns   36511877
BM_xxh64                     16 ns         16 ns   42964585
BM_hash_string              182 ns        182 ns    3853125
BM_iterative_hash            77 ns         77 ns    9087638

Unfortunately, XXH3 is still experimental (see
https://github.com/Cyan4973/xxHash#user-content-new-experimental-hash-algorithm)

However, regular XXH64 is still a lot faster than
htab_hash_string per my benchmark above. I used the
following string for the benchmark:
static constexpr char str[] = "_ZZZL13make_gdb_typeP7gdbarchP10tdesc_typeEN16gdb_type_creator19make_gdb_type_flagsEPK22tdesc_type_with_fieldsE19__PRETTY_FUNCTION__";

htab_hash_string is currently 4.35% + 7.98% (rehashing) of gdb
startup when attaching to Chrome's content_shell.

An additional 5.21% is spent in msymbol_hash, which does not use
this hash function. Unfortunately, since it has to lowercase the
string, it can't use this hash function.
BM_msymbol_hash                52 ns         52 ns   13281495

It may be worth investigating if strlen+XXHash is still faster than
htab_hash_string, which would make it easier to use in more places.

Debian ships xxhash as libxxhash{0,-dev}. Fedora ships it as xxhash-devel.

gdb/ChangeLog:

2019-10-22  Christian Biesinger  <cbiesinger@google.com>

	* Makefile.in: Link with libxxhash.
	* config.in: Regenerate.
	* configure: Regenerate.
	* configure.ac: Search for libxxhash.
	* utils.c (fast_hash): Use xxhash if present.

Change-Id: Icab218388b9f829522ed3977f04301ae6d4fc4ca
2019-10-22 11:47:24 -05:00