Nick Clifton [Tue, 27 Jan 2015 17:32:23 +0000 (17:32 +0000)]
Fixes for invalid memory accesses triggered by running windres on corrupt binaries.
PR binutils/17512
* rcparse.y: Add checks to avoid integer divide by zero.
* rescoff.c (read_coff_rsrc): Add check on the size of the
resource section.
(read_coff_res_dir): Add check on the nesting level.
Check for resource names overrunning the buffer.
* resrc.c (write_rc_messagetable): Update formatting.
Add check of 'elen' being zero.
Nick Clifton [Tue, 27 Jan 2015 15:49:12 +0000 (15:49 +0000)]
Fixes for memory access violations triggered by running addr2line on fuzzed binaries.
PR binutils/17512
* addr2line.c (slurp_symtab): If the symcount is zero, free the
symbol table pointer.
* dwarf2.c (concat_filename): Check for an empty directory table.
(scan_unit_for_symbols): Check for reading off the end of the
unit.
(parse_comp_unit): Check for a DW_AT_comp_dir attribute with a
non-string form.
* elf64-ppc.c (opd_entry_value): Fail if there are no relocs
available.
Joel Brobecker [Wed, 21 Jan 2015 09:29:09 +0000 (10:29 +0100)]
check gdb.lookup_type return value in gdb.python/py-lookup-type.exp
This further improves this testcase to check the output of
our calls to gdb.lookup_type.
gdb/ChangeLog:
* gdb.python/py-lookup-type.exp (test_lookup_type): Change
the second test to print the name attribute of value
returned by the call to gdb.lookup_type, and adjust
the expected output accordingly.
Mark Wielaard [Sun, 25 Jan 2015 15:42:05 +0000 (16:42 +0100)]
Remove testsuite compile errors with GCC5.
GCC5 defaults to the GNU11 standard for C and warns by default for
implicit function declarations and implicit return types.
https://gcc.gnu.org/gcc-5/porting_to.html
Fixing these issues in the testsuite turns 9 untested and 17 unsupported
testcases into 417 new passes when compiling with GCC5.
gdb/testsuite/ChangeLog:
* gdb.arch/i386-bp_permanent.c (standard): New declaration.
* gdb.base/disp-step-fork.c: Include unistd.h.
* gdb.base/siginfo-obj.c: Include stdio.h.
* gdb.base/siginfo-thread.c: Likewise.
* gdb.mi/non-stop.c: Include unistd.h.
* gdb.mi/nsthrexec.c: Include stdio.h.
* gdb.mi/pthreads.c: Include unistd.h.
* gdb.modula2/unbounded1.c (main): Declare returns int.
* gdb.reverse/consecutive-reverse.c: Likewise.
* gdb.threads/create-fail.c: Include unistd.h.
* gdb.threads/killed.c: Likewise.
* gdb.threads/linux-dp.c: Likewise.
* gdb.threads/non-ldr-exc-1.c: Include stdio.h and string.h.
* gdb.threads/non-ldr-exc-2.c: Likewise.
* gdb.threads/non-ldr-exc-3.c: Likewise.
* gdb.threads/non-ldr-exc-4.c: Likewise.
* gdb.threads/pthreads.c: Include unistd.h.
(main): Declare returns int.
* gdb.threads/tls-main.c (foo): New declaration.
* gdb.threads/watchpoint-fork-mt.c: Define _GNU_SOURCE.
Jan Kratochvil [Sat, 24 Jan 2015 14:44:52 +0000 (15:44 +0100)]
Fix 100x slowdown regression on DWZ files
Since Fedora started to use DWZ DWARF compressor:
http://fedoraproject.org/wiki/Features/DwarfCompressor
GDB has slowed down a lot. To make it clear - DWZ is DWARF structure
rearrangement, "compressor" does not mean any zlib style data compression.
This patch reduces LibreOffice backtrace from 5 minutes to 3 seconds (100x)
and it also reduces memory consumption 20x.
[ benchmark is at the bottom of this mail ]
One can see all DW_TAG_partial_unit have DW_AT_stmt_list 0x0 which causes
repeated decoding of that .debug_line unit on each DW_TAG_imported_unit.
This was OK before as each DW_TAG_compile_unit has its own .debug_line unit.
But since the introduction of DW_TAG_partial_unit by DWZ one should cache
read-in DW_AT_stmt_list .debug_line units.
Fortunately one does not need to cache whole
struct linetable *symtab->linetable
and other data from .debug_line mapping PC<->lines
------------------------------------------------------------------------------
Line Number Statements:
Extended opcode 2: set Address to 0x45c880
Advance Line by 25 to 26
Copy
------------------------------------------------------------------------------
as the only part of .debug_line which GDB needs for DW_TAG_partial_unit is:
------------------------------------------------------------------------------
The Directory Table:
../../gdb
/usr/include/bits
[...]
The File Name Table:
Entry Dir Time Size Name
1 1 0 0 gdb.c
2 2 0 0 string3.h
[...]
------------------------------------------------------------------------------
specifically referenced in GDB for DW_AT_decl_file at a single place:
------------------------------------------------------------------------------
fe = &cu->line_header->file_names[file_index - 1];
SYMBOL_SYMTAB (sym) = fe->symtab;
------------------------------------------------------------------------------
This is because for some reason DW_TAG_partial_unit never contains PC-related
DWARF information. I do not know exactly why, the compression ratio is a bit
lower due to it but thanksfully currently it is that way:
dwz.c:
------------------------------------------------------------------------------
/* These attributes reference code, prevent moving
DIEs with them. */
case DW_AT_low_pc:
case DW_AT_high_pc:
case DW_AT_entry_pc:
case DW_AT_ranges:
die->die_ck_state = CK_BAD;
+
/* State of checksum computation. Not computed yet, computed and
suitable for moving into partial units, currently being computed
and finally determined unsuitable for moving into partial units. */
enum { CK_UNKNOWN, CK_KNOWN, CK_BEING_COMPUTED, CK_BAD } die_ck_state : 2;
------------------------------------------------------------------------------
I have also verified also real-world Fedora debuginfo files really comply with
that assumption with dwgrep
https://github.com/pmachata/dwgrep
using:
------------------------------------------------------------------------------
dwgrep -e 'entry ?DW_TAG_partial_unit child* ( ?DW_AT_low_pc , ?DW_AT_high_pc , ?DW_AT_ranges )' /usr/lib/debug/**
------------------------------------------------------------------------------
BTW I think GDB already does not support the whole DW_TAG_imported_unit and
DW_TAG_partial_unit usage possibilities as specified by the DWARF standard.
I think GDB would not work if DW_TAG_imported_unit was used in some inner
level and not at the CU level (readelf -wi level <1>) - this is how DWZ is
using DW_TAG_imported_unit. Therefore I do not think further assumptions
about DW_TAG_imported_unit and DW_TAG_partial_unit usage by DWZ are a problem
for GDB.
One could save the whole .debug_line decoded PC<->lines mapping (and not just
the DW_AT_decl_file table) but:
* there are some problematic corner cases so one could do it incorrectly
* there are no real world data to really test such patch extension
* such extension could be done perfectly incrementally on top of this patch
benchmark - on Fedora 20 x86_64 and FSF GDB HEAD:
echo -e 'thread apply all bt\nset confirm no\nq'|./gdb -p `pidof soffice.bin` -ex 'set pagination off' -ex 'maintenance set per-command
space' -ex 'maintenance set per-command symtab' -ex 'maintenance set per-command time'
FSF GDB HEAD ("thread apply all bt"):
Command execution time: 333.693000 (cpu), 335.587539 (wall)
---sec
Space used: 1736404992 (+1477189632 for this command)
----MB
vs.
THIS PATCH ("thread apply all bt"):
Command execution time: 2.595000 (cpu), 2.607573 (wall)
-sec
Space used: 340058112 (+85917696 for this command)
--MB
FSF GDB HEAD ("thread apply all bt full"):
Command execution time: 466.751000 (cpu), 468.345837 (wall)
---sec
Space used: 2330132480 (+2070974464 for this command)
----MB
vs.
THIS PATCH ("thread apply all bt full"):
Command execution time: 18.907000 (cpu), 18.964125 (wall)
--sec
Space used: 364462080 (+110325760 for this command)
---MB
Fix 100x slowdown regression on DWZ files.
* dwarf2read.c (struct dwarf2_per_objfile): Add line_header_hash.
(struct line_header): Add offset and offset_in_dwz.
(dwarf_decode_lines): Add parameter decode_mapping to the declaration.
(free_line_header_voidp): New declaration.
(line_header_hash, line_header_hash_voidp, line_header_eq_voidp): New
functions.
(dwarf2_build_include_psymtabs): Update dwarf_decode_lines caller.
(handle_DW_AT_stmt_list): Use line_header_hash.
(free_line_header_voidp): New function.
(dwarf_decode_line_header): Initialize offset and offset_in_dwz.
(dwarf_decode_lines): New parameter decode_mapping, use it.
(dwarf2_free_objfile): Free line_header_hash.
Simon Marchi [Fri, 23 Jan 2015 17:59:24 +0000 (12:59 -0500)]
Catch exception in value_rtti_indirect_type
In the situation described in bug 17416 [1]:
* "set print object" is on;
* The variable object is a pointer to a struct, and it contains an
invalid value (e.g. NULL, or random uninitialized value);
* The variable object (struct) has a child which is also a pointer to a
struct;
* We try to use "-var-list-children".
... an exception thrown in value_ind can propagate too far and leave an
half-built variable object, leading to a wrong state. This patch adds a
TRY_CATCH to catch it and makes value_rtti_indirect_type return NULL in
that case, meaning that the type of the pointed object could not be
found.
A test for the fix is also added.
New in v2:
* Added test.
* Restructured "catch" code.
* Added details about the bug in commit log.
gdb/Changelog:
* valops.c (value_rtti_indirect_type): Catch exception thrown by
value_ind.
gdb/testsuite/ChangeLog
* gdb.mi/mi-var-list-children-invalid-grandchild.c: New file.
* gdb.mi/mi-var-list-children-invalid-grandchild.exp: New file.
Mark Wielaard [Tue, 9 Dec 2014 10:45:41 +0000 (11:45 +0100)]
Use GCC5/DWARF5 DW_AT_noreturn to mark functions that don't return normally.
Add a flag field is_noreturn to struct func_type. Make calling_convention
a small bit field to not increase the size of the struct. Set is_noreturn
if the new GCC5/DWARF5 DW_AT_noreturn is set on a DW_TAG_subprogram.
Use this information to warn the user before doing a finish or return from
a function that does not return normally to its caller.
(gdb) finish
warning: Function endless does not return normally.
Try to finish anyway? (y or n)
(gdb) return
warning: Function does not return normally to caller.
Make endless return now? (y or n)
gdb/ChangeLog
* dwarf2read.c (read_subroutine_type): Set TYPE_NO_RETURN from
DW_AT_noreturn.
* gdbtypes.h (struct func_type): Add is_noreturn field flag. Make
calling_convention an 8 bit bit field.
(TYPE_NO_RETURN): New macro.
* infcmd.c (finish_command): Query if function does not return
normally.
* stack.c (return_command): Likewise.
gdb/testsuite/ChangeLog
* gdb.base/noreturn-return.c: New file.
* gdb.base/noreturn-return.exp: New file.
* gdb.base/noreturn-finish.c: New file.
* gdb.base/noreturn-finish.exp: New file.
include/ChangeLog
* dwarf2.def (DW_AT_noreturn): New DWARF5 attribute.
The dwarf2.h addition and the code to emit the new attribute is already in
the gcc tree.
Pedro Alves [Fri, 23 Jan 2015 11:12:39 +0000 (11:12 +0000)]
Linux: make target_is_async_p return false when async is off
linux_nat_is_async_p currently always returns true, even when the
target is _not_ async. That confuses
gdb_readline_wrapper/gdb_readline_wrapper_cleanup, which
force-disables target-async while the secondary prompt is active. As
a result, when gdb_readline_wrapper returns, the target is left async,
even through it was sync to begin with.
That can result in weird bugs, like the one the test added by this
commit exposes.
* linux-nat.c (linux_is_async_p): New macro.
(linux_nat_is_async_p):
(linux_nat_terminal_inferior): Check whether the target can async
instead of whether it is already async.
(linux_nat_terminal_ours): Don't check whether the target is
async.
(linux_async_pipe): Use linux_is_async_p.
Anders Granlund [Thu, 22 Jan 2015 19:33:04 +0000 (14:33 -0500)]
Introduce gdb_interact in testsuite
gdb_interact is a small utility that we have found quite useful to debug
test cases.
Putting gdb_interact in a test suspends it and allows to interact with
gdb to inspect whatever you want. You can then type ">>>" to resume the
test execution. Of course, this is only for gdb devs. It wouldn't make
sense to leave a gdb_interact permanently in a test case.
When starting the interaction with the user, the script prints this
banner:
+------------------------------------------+
| Script interrupted, you can now interact |
| with by gdb. Type >>> to continue. |
+------------------------------------------+
Notes:
* When gdb is launched, the gdb_spawn_id variable (lib/gdb.exp) is
assigned -1. Given the name, I would expect it to contain the gdb
expect spawn id, which is needed for interact. I changed all places
that set gdb_spawn_id to -1 to set it to the actual gdb spawn id
instead.
* When entering the "interact" mode, the last (gdb) prompt is already
eaten by expect, so it doesn't show up on the terminal. Subsequent
prompts do appear though. We tried to print "(gdb)" just before the
interact to replace it. However, it could be misleading if you are
debugging an MI test case, it makes you think that you are typing in a
CLI prompt, when in reality it's MI. In the end I decided that since
the feature is for developers who know what they're doing and that one
is normally consciously using gdb_interact, the script doesn't need
to babysit the user.
* There are probably some quirks depending on where in the script
gdb_interact appears (e.g. it could interfere with following
commands and make them fail), but it works for most cases. Quirks can
always be fixed later.
The idea and original implementation was contributed by Anders
Granlund, a colleague of mine. Thanks to him.
Jan Kratochvil [Thu, 22 Jan 2015 20:04:53 +0000 (21:04 +0100)]
Sort threads for thread apply all
downstream Fedora request:
Please make it easier to find the backtrace of the crashing thread
https://bugzilla.redhat.com/show_bug.cgi?id=1024504
Currently after loading a core file GDB prints:
Core was generated by `./threadcrash1'.
Program terminated with signal SIGSEGV, Segmentation fault.
8 *(volatile int *)0=0;
(gdb) _
there is nowhere seen which of the threads had crashed. In reality GDB always
numbers that thread as #1 and it is the current thread that time. But after
dumping all the info into a file for later analysis it is no longer obvious.
'thread apply all bt' even puts the thread #1 to the _end_ of the output!!!
I find maybe as good enough and with no risk of UI change flamewar to just
sort the threads by their number. Currently they are printed as they happen
in the internal GDB list which has no advantage. Printing thread #1 as the
first one with assumed 'thread apply all bt' (after the core file is loaded)
should make the complaint resolved I guess.
On Thu, 15 Jan 2015 20:29:07 +0100, Doug Evans wrote:
No objection to sorting the list, but if thread #1 is the important one,
then a concern could be it'll have scrolled off the screen (such a
concern has been voiced in another thread in another context),
and if not lost (say it's in an emacs buffer) one would still have
to scroll back to see it.
So one *could* still want #1 to be last.
Do we want an option to choose the sort direction?
Jan Kratochvil [Thu, 22 Jan 2015 20:02:24 +0000 (21:02 +0100)]
Print current thread after loading a core file
downstream Fedora request:
Please make it easier to find the backtrace of the crashing thread
https://bugzilla.redhat.com/show_bug.cgi?id=1024504
Currently after loading a core file GDB prints:
Core was generated by `./threadcrash1'.
Program terminated with signal SIGSEGV, Segmentation fault.
8 *(volatile int *)0=0;
(gdb) _
there is nowhere seen which of the threads had crashed. In reality GDB always
numbers that thread as #1 and it is the current thread that time. But after
dumping all the info into a file for later analysis it is no longer obvious.
'thread apply all bt' even puts the thread #1 to the _end_ of the output!!!
Should GDB always print after loading a core file what "thread" command would
print?
[Current thread is 1 (Thread 0x7fcbe28fe700 (LWP 15453))]
BTW I think it will print the thread even when loading single/non-threaded
core file when other inferior(s) exist. But that currently crashes
[Bug threads/12074] multi-inferior internal error
https://sourceware.org/bugzilla/show_bug.cgi?id=12074
plus I think that would be a correct behavior anyway.
* corelow.c (core_open): Call also thread_command.
* gdbthread.h (thread_command): New prototype moved from ...
* thread.c (thread_command): ... here.
(thread_command): Make it global.
Pedro Alves [Thu, 22 Jan 2015 18:30:01 +0000 (18:30 +0000)]
mingw32: fix windows-termcap/curses check
When GDB is configured with "--without-tui --with-curses" or "--with-tui",
$prefer_curses is set to yes. But, that still doesn't mean that curses
will be used. configure will still search for the curses library, and
continue building without it. That's done here:
curses_found=no
if test x"$prefer_curses" = xyes; then
...
AC_SEARCH_LIBS(waddstr, [ncurses cursesX curses])
if test "$ac_cv_search_waddstr" != no; then
curses_found=yes
fi
fi
So if waddstr is not found, meaning curses is not really
available, even though it'd be preferred, $prefer_curses is
'yes', but $curses_found is 'no'.
So the right fix to tell whether we're linking with curses is
$curses_found=yes.
* gdb/tui/tui.c (tui_enable) [__MINGW32__]: If the call to 'newterm'
fails with the 1st arg NULL, try again with "unknown". Don't test
the "cup" capability: it isn't supported by the Windows port of
ncurses, but the Windows console driver is still capable of
supporting TUI.
Jan Kratochvil [Thu, 22 Jan 2015 18:18:16 +0000 (19:18 +0100)]
compile: Fix function pointers
TBH while I always comment reasons for each of the compilation options in
reality I tried them all and chose that combination that needs the most simple
compile/compile-object-load.c (ld.so emulation) implementation.
Nick Clifton [Thu, 22 Jan 2015 12:06:04 +0000 (12:06 +0000)]
Fixes memory access violations triggered by running dlltool on fuzzed binaries.
PR binutils/17512
* coffcode.h (handle_COMDAT): When searching for the section
symbol, make sure that there is space left in the symbol table.
* vms-alpha.c (_bfd_vms_slurp_ehdr): Add range checks.
Nick Clifton [Wed, 21 Jan 2015 17:37:23 +0000 (17:37 +0000)]
Fix memory access violations triggered by running strip on fuzzed binaries.
PR binutils/17512
* coffcode.h (coff_set_arch_mach_hook): Check return value from
bfd_malloc.
(coff_slurp_line_table): Return FALSE if the line number
information was corrupt.
(coff_slurp_symbol_table): Return FALSE if the symbol information
was corrupt.
* mach-o.c (bfd_mach_o_bfd_copy_private_header_data): Always
initialise the fields of the dyld_info structure.
(bfd_mach_o_build_exec_seg_command): Replace assertion with an
error message and a return value.
(bfd_mach_o_layout_commands): Change the function to boolean.
Return FALSE if the function fails.
(bfd_mach_o_build_commands): Fail if bfd_mach_o_layout_commands
fails.
(bfd_mach_o_read_command): Fail if an unrecognised command is
encountered.
* peXXigen.c (_bfd_XXi_swap_aouthdr_in): Set bfd_error if the
read fails.
(slurp_symtab): Check the return from bfd_malloc.
(_bfd_XX_bfd_copy_private_bfd_data_common): Fail if the copy
encountered an error.
(_bfd_XXi_final_link_postscript): Fail if a section could not be
copied.
* peicode.h (pe_bfd_object_p): Fail if the header could not be
swapped in.
* tekhex.c (first_phase): Fail if the section is too big.
* versados.c (struct esdid): Add content_size field.
(process_otr): Use and check the new field.
(versados_get_section_contents): Check that the section exists and
that the requested data is available.
Alan Modra [Tue, 20 Jan 2015 10:41:37 +0000 (21:11 +1030)]
Reorder more powerpc64 sections again for -z relro
.toc1 is the second level TOC section used by gcc's -mminimal-toc. It
too should be read-only after relocation. Also, the last patch
description mentioned .sbss moving but didn't actually do that, so fix
that problem. .tocbss (whatever that is) was before .sbss previously,
so move that one too.
* emulparams/elf64ppc.sh (OTHER_SDATA_SECTIONS): Use in place of..
(OTHER_BSS_SYMBOLS): ..this.
(OTHER_PLT_RELOC_SECTIONS): Don't define.
(OTHER_GOT_RELOC_SECTIONS): Add rela.toc1 and rela.tocbss.
(OTHER_READWRITE_SECTIONS): Don't define. Move .toc1 to..
(OTHER_RELRO_SECTIONS_2): ..here.
* scripttempl/elf.sc: Move SBSS too when DATA_SDATA.
Andrew Burgess [Wed, 7 Jan 2015 10:51:35 +0000 (10:51 +0000)]
ld: Don't evaluate unneeded PROVIDE expressions.
When creating a linker mapfile (using -Map=MAPFILE), we previously would
always try to evaluate the expression from a PROVIDE statement.
However, this is not always safe, consider:
PROVIDE (foo = 0x10);
PROVIDE (bar = foo);
In this example, if neither 'foo' or 'bar' is needed, then while
generating the linker mapfile evaluating the expression for 'foo' is
harmless (just the value 0x10). However, evaluating the expression for
'bar' requires the symbol 'foo', which is undefined. This used to cause
a fatal error.
This patch changes the behaviour, so that when the destination of the
PROVIDE is not defined (that is the PROVIDE is not going to provide
anything) the expression is not evaluated, and instead a special string
is displayed to indicate that the linker is discarding the PROVIDE
statement.
This change not only fixes the spurious undefined symbol error, but also
means that a user can now tell if a PROVIDE statement has provided
anything by inspecting the linker mapfile, something that could not be
done before.
ld/ChangeLog:
* ldlang.c (print_assignment): Only evaluate the expression for a
PROVIDE'd assignment when the destination is being defined.
Display a special message for PROVIDE'd symbols that are not being
provided.
ld/testsuite/ChangeLog:
* ld-scripts/provide-4.d: New file.
* ld-scripts/provide-4-map.d: New file.
* ld-scripts/provide-4.t: New file.
* ld-scripts/provide-5.d: New file.
* ld-scripts/provide-5.s: New file.
* ld-scripts/provide-5-map.d: New file.
* ld-scripts/provide-5.t: New file.
* ld-scripts/provide.exp: Run the provide-4.d and provide-5.d
tests.
Andrew Burgess [Tue, 6 Jan 2015 23:40:48 +0000 (23:40 +0000)]
ld/testing: run_dump_test can now check linker mapfiles.
Add a new option 'map' to the ld run_dump_test mechanism. When the
'map' option is given run_dump_test will ensure that there is a
-Map=MAPFILE present in the linker command line, adding one if needed.
The MAPFILE is then compared with the file passed to the new 'map'
option using the regexp_diff function. This should make it slightly
easier to write tests that check the linker mapfile output.
The only test I found that already compares mapfile content is updated
to use the new mechanism.
ld/testsuite/ChangeLog:
* ld-scripts/overlay-size.d: Add 'map' option.
* ld-scripts/overlay-size.exp: Remove manual check of mapfile.
* lib/ld-lib.exp (run_dump_test): Add support for new 'map'
option, checking linker mapfile output.
Alan Modra [Tue, 20 Jan 2015 06:49:15 +0000 (17:19 +1030)]
Reorder more powerpc64 sections for -z relro
This moves .got too, which requires .sdata and .sbss to move with it,
because these sections share addressing via the toc pointer and with
small-model code must be within a 16-bit signed offset. .plt, .iplt
and .branch_lt must also be moved since they are addressed via a
32-bit offset from the toc pointer, and we might have a very large
.data section.
This change means we may have some bss style sections before the data
segment, necessitating another PT_LOAD header. Also, since _edata is
defined at the end of the data segment it's possible with an empty
.data to have _edata at the end of .plt which looks a little unusual
since .plt is a bss style section. That should only happen rarely in
real world binaries, but does occur in the ld testsuite.
Alan Modra [Sat, 17 Jan 2015 11:03:43 +0000 (21:33 +1030)]
Fix garbage collection of common symbols
Running lang_common before garbage collection means slightly less work
in garbage collection code, since common symbols should no longer
appear there. It does have the side effect of keeping linker script
symbols (at least those defined outside of sections) global too,
hence some testsuite churn.
bfd/
PR 17165
* elf-bfd.h (ELF_COMMON_DEF): Note that this might be true for
linker script assignments too.
* elflink.c (elf_gc_sweep_symbol): Don't drop ELF_COMMON_DEF syms.
(bfd_elf_gc_mark_dynamic_ref_symbol): Similarly.
ld/
PR 17165
* ldlang.c (lang_process): Run lang_common before lang_gc_sections.
ld/testsuite/
* ld-gc/pr14265.d,
* ld-cris/tls-gc-68.d,
* ld-cris/tls-gc-69.d,
* ld-cris/tls-gc-70.d,
* ld-cris/tls-gc-71.d,
* ld-cris/tls-gc-75.d,
* ld-cris/tls-gc-76.d,
* ld-cris/tls-gc-79.d,
* ld-mmix/bpo-10.d,
* ld-mmix/bpo-11.d: Update.
Alan Modra [Mon, 19 Jan 2015 08:27:04 +0000 (18:57 +1030)]
Extend .reloc to accept some BFD_RELOCs
Tests that bfd_perform_reloc doesn't freak over a NONE reloc at end
of section.
gas/
* read.c (s_reloc): Match BFD_RELOC_NONE, BFD_RELOC{8,16,32,64}.
* write.c (get_frag_for_reloc): Allow match just past end of frag.
gas/testsuite/
* gas/all/none.s,
* gas/all/none.d: New test.
* gas/all/gas.exp: Run it.
Alan Modra [Mon, 19 Jan 2015 00:06:26 +0000 (10:36 +1030)]
Fallout from recent bfd_reloc_outofrange changes
Commit ec93045b and cd21f5da introduced a large number of tic4x and
tic54x regressions, due to the new checks being wrong for targets
with octets_per_byte != 1. To fix that I introduced a new
bfd_get_section_limit_octets and performed the check on octets rather
than byte adresses, reducing the number of bfd_octets_per_byte calls.
bfd_octets_per_byte is rather expensive..
I then wondered why the same bfd_reloc_outofrange check added to
bfd_perform_relocation wasn't also added to bfd_install_relocation.
The two functions are virtually identical and ought to remain that
way. However, adding the same check to bfd_install_relocation
resulted in ld-elf "FAIL Link eh-group.o to eh-group" on many ELF
targets, including x64_64-linux. The reason being that eh-group.o
has NONE relocs at the end of a section, and most targets give NONE
relocs a non-zero size. So if we are to keep the new outofrange
check it appears that NONE relocs must have a zero size.
This commit fixes the regression on RHEL-5 systems introduced by
nat/linux-personality.c's check of HAVE_DECL_ADDR_NO_RANDOMIZE.
RHEL-5 systems define HAVE_DECL_ADDR_NO_RANDOMIZE as zero, so we
cannot use #ifndef; instead this patch uses the "#if !" construction.
The regression was reported by Ulrich Weigand here:
* tui/tui-win.c (tui_rehighlight_all, tui_set_var_cmd): New
functions.
(_initialize_tui_win) <border-kind, border-mode>:
<active-border-mode>: Use tui_set_var_cmd as the "set" function.
* tui/tui-win.h: Add prototype for tui_rehighlight_all.
* tui/tui-win.c (tui_set_tab_width_command): Delete and
recreate the source and the disassembly windows, to show the
effect of the changed tab size immediately.
Eli Zaretskii [Fri, 16 Jan 2015 11:31:36 +0000 (13:31 +0200)]
Fix TUI-related documentation.
tui/tui-win.c (tui_scroll_left_command, tui_scroll_right_command):
Doc fix.
doc/gdb.texinfo (TUI Commands): Document the possible
values of NAME argument to 'winheight' command. Explain the
effect of 'tabset' setting better.
Andreas Krebbel [Fri, 16 Jan 2015 11:19:21 +0000 (12:19 +0100)]
S/390: Add support for IBM z13.
- 32 128 bit vector registers (overlapping with the existing 16 64 bit
floating point registers)
- vector double instructions
- vector integer instructions
- scalar vector instructions (allowing to have more floating point
registers for scalar operations)
- vector string instructions
gas/ChangeLog:
* config/tc-s390.c (struct pd_reg): Remove.
(pre_defined_registers): Remove.
(REG_NAME_CNT): Remove.
(reg_name_search): Calculate the register number instead of doing
a lookup.
(register_name, tc_s390_regname_to_dw2regnum): Adopt to the new
reg_name_search signature.
(s390_parse_cpu): Support the new arch string z13.
(s390_insert_operand): Support for vector registers with the extra
field for the fifth bit of each vector register operand.
(md_gather_operand): Adjust to the new handling of optional
parameters.
* doc/as.texinfo: Document the z13 cpu string.
gas/testsuite/ChangeLog:
* gas/s390/esa-g5.d: Add a variant without the optional operand.
* gas/s390/esa-g5.s: Likewise.
* gas/s390/esa-z9-109.d: Likewise.
* gas/s390/esa-z9-109.s: Likewise.
* gas/s390/zarch-z9-109.d: Likewise.
* gas/s390/zarch-z9-109.s: Likewise.
* gas/s390/zarch-z10.d: For variants with a zero optional argument
it is not dumped by objdump anymore.
* gas/s390/zarch-zEC12.d: Likewise.
* gas/s390/zarch-z13.d: New file.
* gas/s390/zarch-z13.s: New file.
* gas/s390/s390.exp: Run the test for the z13 files.
* ld-s390/tlsbin.dd: The nopr register operand is optional and not
printed if 0 anymore.
opcodes/ChangeLog:
* s390-dis.c (s390_extract_operand): Support vector register
operands.
(s390_print_insn_with_opcode): Support new operands types and add
new handling of optional operands.
* s390-mkopc.c (s390_opcode_mode_val, s390_opcode_cpu_val): Remove
and include opcode/s390.h instead.
(struct op_struct): New field `flags'.
(insertOpcode, insertExpandedMnemonic): New parameter `flags'.
(dumpTable): Dump flags.
(main): Parse flags from the s390-opc.txt file. Add z13 as cpu
string.
* s390-opc.c: Add new operands types, instruction formats, and
instruction masks.
(s390_opformats): Add new formats for .insn.
* s390-opc.txt: Add new instructions.
Eli Zaretskii [Fri, 16 Jan 2015 11:24:20 +0000 (13:24 +0200)]
Leave more space in TUI mode for thread ID.
gdb/tui/tui-data.h (LINE_PREFIX): Make shorter
(MAX_PID_WIDTH): Enlarge from 14 to 19, to leave enough space for
"Thread NNNNN.XXXX" thread ID notation on Windows.
Jan Kratochvil [Fri, 16 Jan 2015 05:39:47 +0000 (06:39 +0100)]
Fix gcc-5 compilation
With gcc-5.0 pre-release one gets:
hppa-tdep.c: In function ‘inst_saves_gr’:
hppa-tdep.c:1406:30: error: comparison of constant ‘9’ with boolean expression is always false [-Werror=bool-compare]
This patch moves the shared code present on
gdb/linux-nat.c:linux_nat_create_inferior and
gdb/gdbserver/linux-low.c:linux_create_inferior to
nat/linux-personality.c. This code is responsible for disabling
address space randomization based on user setting, and using
<sys/personality.h> to do that. I decided to put the prototype of the
maybe_disable_address_space_randomization on nat/linux-osdata.h
because it seemed the best place to put it.
I regression-tested this patch on Fedora 20 x86_64, and found no
regressions.
* Makefile.in (SFILES): Add linux-personality.c.
(linux-personality.o): New rule.
* configure.srv (srv_linux_obj): Add linux-personality.o to the
list of objects to be built.
* linux-low.c: Include nat/linux-personality.h.
(linux_create_inferior): Remove code to disable address space
randomization (moved to ../nat/linux-personality.c). Create
cleanup to disable address space randomization.
This patch moves safe_strerror from the gdb/{posix,mingw}-hdep.c files
to the respective common/{posix,mingw}-strerror.c files. This is a
preparation for the next patch, which shares a common code (to disable
address space randomization when creating a new inferior).
The patch has been regtested on Fedora 20 x86_64, and no regressions
were found.
* Makefile.in (ALLDEPFILES): Including common/mingw-strerror.c and
common/posix-strerror.c.
(posix-strerror.o): New rule.
(mingw-strerror.o): Likewise.
* common/common-utils.h (safe_strerror): Move prototype to here,
from utils.h.
* common/common.host: New file.
* common/mingw-strerror.c: Likewise.
* common/posix-strerror.c: Likewise.
* configure: Regenerated.
* configure.ac: Source common/common.host. Add variable
common_host_obs to gdb_host_obs.
* contrib/ari/gdb_ari.sh: Mention gdb/common/mingw-strerror.c and
gdb/common/posix-strerror.c when warning about the use of
strerror.
* mingw-hdep.c (safe_strerror): Remove definition; move it to
common/mingw-strerror.c.
* posix-hdep.c (safe_strerror): Remove definition; move it to
common/posix-hdep.c.
* utils.h (safe_strerror): Remove prototype; move to
common/common-utils.h.
Don Breazeal [Thu, 15 Jan 2015 18:47:31 +0000 (10:47 -0800)]
Skip two more attach tests when testing against stub-like targets
This patch updates two attach tests to use utility procs for checking if
the attach test should run and for launching the program to be attached, as
follows:
1) Use can_spawn_for_attach instead of is_remote target
2) Use spawn_wait_for_attach instead of exec/sleep
Tested (1) with i686-mingw32 host and i686-pc-linux-gnu build/target and
both with x86_64 Ubuntu.
gdb/testsuite/ChangeLog:
* gdb.base/attach-pie-noexec.exp: Use can_spawn_for_attach
instead of checking whether the target board is remote and
use spawn_wait_for_attach instead of exec/sleep.
* gdb.base/attach-twice.exp: Likewise.
Alan Modra [Thu, 15 Jan 2015 08:37:33 +0000 (19:07 +1030)]
Omit section dynsyms for any linker created section
This is a simplification, and fixes a testcase I had where an empty
powerpc64 .branch_lt section was chosen for the data_index_section
dynamic symbol and thus wasn't removed.
* elflink.c (_bfd_elf_link_omit_section_dynsym): Return true for
any output section matching a linker created dynobj section.
Alan Modra [Thu, 15 Jan 2015 09:12:59 +0000 (19:42 +1030)]
Fix ARM fail of gap test
ld-elf/gap test was failing due to the ARM backend attempting to output
arch symbols when ld -s (strip all symbols) is in force. This patch
stops that happening and tidies the code a little.
PR 17842
* elflink.c (elf_link_output_sym): Assert elf_onesymtab set.
(bfd_elf_final_link): Always create a symbol table when emit_relocs.
Don't assign symtab file position unless symbols will be output.
Merge blocks with condition in common. Don't call
elf_backend_output_arch_local_syms or elf_backend_output_arch_syms
unless other symbols are output. Move assignment of symtab_shndx
file position. Localize variable.
Joel Brobecker [Thu, 15 Jan 2015 06:09:32 +0000 (10:09 +0400)]
[Ada] 'first/'last/'length of array whose bound is a discriminant
Consider the following code:
type Table is array (Positive range <>) of Integer;
type Object (N : Integer) is record
Data : Table (1 .. N);
end record;
My_Object : Object := (N => 3, Data => (3, 5, 8));
Trying to print the range and length of the My_Object.Data array yields:
The first one is correct, and that is thanks to the fact that
the lower bound is statically known. However, for the upper
bound, and consequently the array's length, the values are incorrect.
It should be:
What happens here is that ada_array_bound_from_type sees that
our array has a parallel "___XA" type, and therefore tries to
use it. In particular, it described our array's index type as:
[...]___XDLU_1__n, which means lower bound = 1, and upper bound
is value of "n". Unfortunately, ada_array_bound_from_type does
not have access to the discriminant, and is therefore unable to
compute the bound correctly.
Fortunately, at this stage, the bound has already been computed
a while ago, and therefore doesn't need to be re-computed here.
This patch fixes the issue by ignoring that ___XA type if the array
is marked as already fixed.
This also fixes the same issue with packed arrays.
gdb/ChangeLog:
* ada-lang.c (ada_array_bound_from_type): Ignore array's parallel
___XA type if the array has already been fixed.
Joel Brobecker [Thu, 15 Jan 2015 03:14:45 +0000 (07:14 +0400)]
Build failure in sim/rx/gdb-if.c on windows
This should fix a build failure reported on x86_64-mingw32 by Daniel
Calcoen due to conflicting declarations of "open". This patch just
renames the static global in sim/rx/gdb-if.c into "rx_sim_is_open".
sim/rx/ChangeLog:
* gdb-if.c (open): Rename to...
(rx_sim_is_open): This. Replace all uses of "open" by uses of
"rx_sim_is_open" throughout.
This pulls in some missing prototypes and also adds corresponding entries
to the ChangeLog file. Please note that for one function, strverscmp(),
the ChangeLog entry was already there, but the actual prototype wasn't.
Yao Qi [Fri, 28 Nov 2014 01:18:26 +0000 (09:18 +0800)]
Detect 64-bit-ness in PowerPC Book III-E
This patch is to teach both GDB and GDBServer to detect 64-bit inferior
correctly. We find a problem that GDBServer is unable to detect on a
e5500 core processor. Current GDBServer assumes that MSR is a 64-bit
register, but MSR is a 32-bit register in Book III-E. This patch is
to fix this problem by checking the right bit in MSR, in order to handle
both Book III-S and Book III-E. In order to detect Book III-S and
Book III-E, we check the PPC_FEATURE_BOOKE from the host's HWCAP (by
getauxval on glibc >= 2.16. If getauxval doesn't exist, we implement
the fallback by parsing /proc/self/auxv), because it should an invariant
on the same machine cross different processes.
In order to share code, I add nat/ppc-linux.c for both GDB and
GDBserver side.
Yao Qi [Thu, 8 Jan 2015 09:23:55 +0000 (17:23 +0800)]
Move some ppc macros to nat/ppc-linux.h
When I use PPC_FEATURE_BOOKE in GDBserver, I find it is defined in GDB
but not in GDBserver. After taking a further look, I find some macros
are duplicated between ppc-linux-nat.c and linux-ppc-low.c, so this
patch is to move them into nat/ppc-linux.h.
Pedro Alves [Mon, 12 Jan 2015 19:30:08 +0000 (19:30 +0000)]
PR17525 - breakpoint commands not executed when program run from -x script
Executing a gdb script that runs the inferior (from the command line
with -x), and has it hit breakpoints with breakpoint commands that
themselves run the target, is currently broken on async targets
(Linux, remote).
While we're executing a command list or a script, we force the
interpreter to be sync, which results in some functions nesting an
event loop and waiting for the target to stop, instead of returning
immediately and having the top level event loop handle the stop.
The issue with this bug is simply that bpstat_do_actions misses
checking whether the interpreter is sync. When we get here, in the
case of executing a script (or, when the interpreter is sync), the
program has already advanced to the next breakpoint, through
maybe_wait_sync_command_done. We need to process its breakpoints
immediately, just like with a sync target.
PR gdb/17525
* gdb.base/bp-cmds-execution-x-script.c: New file.
* gdb.base/bp-cmds-execution-x-script.exp: New file.
* gdb.base/bp-cmds-execution-x-script.gdb: New file.
projects / binutils.git / log