There is a tiny error left in dwarf.c:read_leb128 after Nick fixed the
signed overflow problem in code I wrote. It's to do with sleb128
values that have unnecessary excess bytes. For example, -1 is
represented as 0x7f, the most efficient encoding, but also as
0xff,0x7f or 0xff,0xff,0x7f and so on. None of these sequences
overflow any size signed value, but read_leb128 will report an
overflow given enough excess bytes. This patch fixes that problem,
and since the proper test for signed values with excess bytes can
easily be adapted to also test a sleb byte with just some bits that
overflow the result, I changed the code to not use signed right
shifts. (The C standard ISO/IEC 9899:1999 6.5.7 says signed right
shifts of negative values have an implementation defined value. A
long time ago I even used a C compiler for a certain microprocessor
that always did unsigned right shifts. Mind you, it is very unlikely
to be compiling binutils with such a compiler.)
bfd/
* wasm-module.c: Guard include of limits.h.
(CHAR_BIT): Provide backup define.
(wasm_read_leb128): Use CHAR_BIT to size "result" in bits.
Correct signed overflow checking.
opcodes/
* wasm32-dis.c: Include limits.h.
(CHAR_BIT): Provide backup define.
(wasm_read_leb128): Use CHAR_BIT to size "result" in bits.
Correct signed overflow checking.
binutils/
* dwarf.c: Include limits.h.
(CHAR_BIT): Provide backup define.
(read_leb128): Use CHAR_BIT to size "result" in bits. Correct
signed overflow checking.
* testsuite/binutils-all/pr26548.s,
* testsuite/binutils-all/pr26548.d,
* testsuite/binutils-all/pr26548e.d: New tests.
* testsuite/binutils-all/readelf.exp: Run them.
(readelf_test): Drop unused "xfails" parameter. Update all uses.
RISC-V only defines two float ABIs, soft-float and double-float, and the
value of soft-float is 0x0. But 0x0 usually means unknown/default setting
for many targets, and the non-ABI binary, which is generated by "ld/objcopy
-b binary", also has the 0x0 elf header flags, this may be confused.
We probably can define a new unknown/default ABI value to make them more
clear, but that will need more bits in the elf header flags, and also need
to discuss in the riscv psabi spec.
Training linker have a default ABI setting, and can be changed by ld
options or configure options is another solution, like what assemblr
usually do. So all objects, including the binary files, will have
explicit ABI setting. But the binary files will no longer be linked
with any object, users need to recompile them with the exactly ABI
they want. It may be inconvenience sometimes. Besides, I think linker
doesn't need to know the default arch/abi so far, just set them according
to the linked objects should be enough.
Therefore, without changing the riscv psabi, and keep the non-ABI binary
can be linked with any object, we don't check the ABI flags if no code
section in the PR24389. Just that we find the first input non-ABI binary
still cannot be linked with others in the PR27200. This patch fixs the
problem by delaying the elf_flags_init(obfd) check, since the flags of
non-ABI object with no code cannot be copyed to output BFD, we should
skip it, even if it is the first linked object.
However, there is a strange "break" at the end of loop in the PR24389.
The "break" cause the ld testcase "Link with zlib-gabi compressed debug
output 1" fails for rv64gc-linux toolchain, after applying the above
change. The root cause is that - the "break" make linker only checks
the "first" section of input BFD rather than the entire sections.
I have checked that AARCH64 and ARM both have the "break" at the end
of loop, but ARC doesn't. I suppose we should remove the "break" like
what ARC do, or use a pair of braces for the if statement.
I have passed the elf/linux toolchain regressions, so the change should
be fine.
bfd/
PR 27200
* elfnn-riscv.c (_bfd_riscv_elf_merge_private_bfd_data): Delay
copying the elf flags from input BFD to output BFD, until we have
checked if the input BFD has no code section or not. Also fix the
problem that we only check the first section rather than the entire
sections for input BFD.
Please consider output of objdump for the executable generated from pr13961.S
-------------
Contents of the .debug_info section:
...
<1><62>: Abbrev Number: 2 (DW_TAG_class_type)
<63> DW_AT_name : foo2
<68> DW_AT_byte_size : 4
<69> DW_AT_decl_file : 1
<6a> DW_AT_decl_line : 1
<6b> DW_AT_sibling : <0x3f> !!! There is no DIE <0x3f>
...
Contents of the .debug_types section:
...
<1><25>: Abbrev Number: 8 (DW_TAG_class_type) !! Hand-inserted of size=5
<26> DW_AT_specification: <0x2a>
<1><2a>: Abbrev Number: 2 (DW_TAG_class_type)
<2b> DW_AT_name : foo
<2f> DW_AT_byte_size : 4
<30> DW_AT_decl_file : 1
<31> DW_AT_decl_line : 1
<32> DW_AT_sibling : <0x3f> !!! There is no DIE <0x3f>, should be <44>
<2><36>: Abbrev Number: 3 (DW_TAG_member)
<37> DW_AT_name : bar
<3b> DW_AT_decl_file : 1
<3c> DW_AT_decl_line : 4
<3d> DW_AT_type : <0x3f> !!! There is no DIE <0x3f>
<41> DW_AT_data_member_location: 0
<42> DW_AT_accessibility: 1 (public)
<2><43>: Abbrev Number: 0
<1><44>: Abbrev Number: 4 (DW_TAG_base_type)
<45> DW_AT_byte_size : 4
<46> DW_AT_encoding : 5 (signed)
<47> DW_AT_name : int
...
---------------
The original assembly is generated from a source file and then
modified to insert DIE, with that the subsequent DIE references
should have been updated, which were not.
It is now getting updated to replace hardcoded DIE references with
label-calculated references.
gdb/testsuite/ChangeLog:
2021-02-16 Alok Kumar Sharma <AlokKumar.Sharma@amd.com>
* gdb.dwarf2/pr13961.S: Corrected invalide DIE references.
Since SHF_GNU_RETAIN is allowed on all sections, strip SHF_GNU_RETAIN
when checking incorrect section attributes.
PR gas/27412
* config/obj-elf.c (obj_elf_change_section): Strip SHF_GNU_RETAIN
when checking incorrect section attributes.
* testsuite/gas/elf/elf.exp: Run section28 and section29.
* testsuite/gas/elf/section28.d: New file.
* testsuite/gas/elf/section28.s: Likewise.
* testsuite/gas/elf/section29.d: Likewise.
* testsuite/gas/elf/section29.s: Likewise.
CVTPI2PD with a memory operand, unlike CVTPI2PS, doesn't engage MMX
logic. Therefore it
- has a proper AVX equivalent (CVTDQ2PD) and hence can be subject to
SSE2AVX translation and SSE checking,
- should not record MMX use in the respective ELF note.
This undoes a change to md_assemble() that 32930e4edb ("x86: Support
GNU_PROPERTY_X86_ISA_1_V[234] marker") did without any explanation. This
broke a CVTPI2PS property test that a subsequent test will add, and the
updates to existing tests also demonstrate what was wrong: For example,
AVX insns update the full YMM, even if a Vex128 variant is in use.
VZERO{ALL,UPPER} modify YMM registers despite having no operands.
While {,V}{LD,ST}MXCSR don't modify XMM registers, MXCSR and XMMn
collectively form underlying machine state.
opcodes/
* s390-mkopc.c (main): Accept arch14 as cpu string.
* s390-opc.txt: Add new arch14 instructions.
include/
* opcode/s390.h (enum s390_opcode_cpu_val): Add
S390_OPCODE_ARCH14.
gas/
* config/tc-s390.c (s390_parse_cpu): New entry for arch14.
* doc/c-s390.texi: Document arch14 march option.
* testsuite/gas/s390/s390.exp: Run the arch14 related tests.
* testsuite/gas/s390/zarch-arch14.d: New test.
* testsuite/gas/s390/zarch-arch14.s: New test.
"cp -p" has been observed to fail on Cygwin when the build tree is on a
local drive but the sources are on a Samba share. We don't really need
full copies of the files here - symlinks suffice.
* objdump.c (load_specific_debug_section): Don't call
bfd_cache_section_contents. Rearrange so that
bfd_get_full_section_contents is not called on path where
bfd_simple_get_relocated_section_contents is called.
Don't set section->user_data.
(free_debug_section): Always free section->start. Don't twiddle
section flags.
* readelf.c (load_specific_debug_section): Don't set user_data.
* dwarf.h (struct dwarf_section): Remove use_data field.
* dwarf.c (NO_ABBREVS, ABBREV): Adjust to suit.
nds32_elf_get_relocated_section_contents uses nds32_get_section_contents
to read sections contents, but nds32_get_section_contents has the wrong
behaviour as it calls bfd_malloc_and_get_section. That function always
mallocs its output buffer, whereas get_relocated_section_contents must
support an already allocated buffer.
bfd/
* elf32-nds32.c (nds32_get_section_contents): Replace
bfd_malloc_and_get_section with bfd_get_full_section_contents.
(nds32_elf_relax_delete_blanks): Init contents.
(nds32_elf_relax_section, nds32_relax_fp_as_gp): Likewise.
binutils/
* testsuite/binutils-all/compress.exp: Remove nds32 xfails.
* testsuite/binutils-all/objdump.exp: Likewise.
Use bfd_get_full_section_contents and tidy the start of this function
to match current generic get_relocated_section_contents.
* coff-alpha.c (alpha_ecoff_get_relocated_section_contents): Use
bfd_get_full_section_contents.
In particular, bfd_get_full_section_contents rather than
bfd_get_section_contents so that compressed sections are handled
properly.
Necessary for mips if objdump is to not cache debug sections.
* elfxx-mips.c (_bfd_elf_mips_get_relocated_section_contents): Apply
all fixes to bfd_generic_get_relocated_section_contents since this
function was split out.
Space for a NULL is there in every backend bfd_get_symtab_upper_bound
or bfd_get_dynamic_symtab_upper_bound when the symbol count is non-zero,
and placed as a terminator by bfd_canonicalize_symtab. Many backends
even return a single NULL entry array for zero symbol count, and while
there are a few that return a NULL array for no symbols, that case is
handled fine in objdump. So don't have objdump add yet another NULL
entry.
* objdump.c (slurp_symtab): Don't add an extra entry for NULL
to the symbol array.
(slurp_dynamic_symtab): Likewise.
(dump_bfd): Formatting. Copy terminating NULL from extra_syms.
"unresolved" as a test result means runtest returns an error, which
can be confusing when there is no apparent error unless you look in
.log files. In particular many tests are skipped without reporting an
error if no target C compiler is found, but if a target C compiler is
found but won't compile a testcase for some reason we used to mark the
test as unresolved. Which is no more worthy of an error than when
lacking a C compiler entirely.
* testsuite/ld-cdtest/cdtest.exp,
* testsuite/ld-checks/checks.exp,
* testsuite/ld-elf/binutils.exp,
* testsuite/ld-elf/compress.exp,
* testsuite/ld-elf/dwarf.exp,
* testsuite/ld-elf/exclude.exp,
* testsuite/ld-elf/frame.exp,
* testsuite/ld-elf/indirect.exp,
* testsuite/ld-elf/linux-x86.exp,
* testsuite/ld-elf/sec-to-seg.exp,
* testsuite/ld-elf/tls_common.exp,
* testsuite/ld-elfcomm/elfcomm.exp,
* testsuite/ld-elfvers/vers.exp,
* testsuite/ld-elfvsb/elfvsb.exp,
* testsuite/ld-elfweak/elfweak.exp,
* testsuite/ld-ifunc/binutils.exp,
* testsuite/ld-mips-elf/mips-elf-flags.exp,
* testsuite/ld-misc/defsym.exp,
* testsuite/ld-mn10300/mn10300.exp,
* testsuite/ld-plugin/lto.exp,
* testsuite/ld-plugin/plugin.exp,
* testsuite/ld-scripts/align.exp,
* testsuite/ld-scripts/assert.exp,
* testsuite/ld-scripts/crossref.exp,
* testsuite/ld-scripts/defined.exp,
* testsuite/ld-scripts/extern.exp,
* testsuite/ld-scripts/log2.exp,
* testsuite/ld-scripts/map-address.exp,
* testsuite/ld-scripts/phdrs.exp,
* testsuite/ld-scripts/phdrs2.exp,
* testsuite/ld-scripts/script.exp,
* testsuite/ld-scripts/section-flags.exp,
* testsuite/ld-scripts/sizeof.exp,
* testsuite/ld-scripts/weak.exp,
* testsuite/ld-selective/selective.exp,
* testsuite/ld-sh/sh.exp,
* testsuite/ld-shared/shared.exp,
* testsuite/ld-srec/srec.exp,
* testsuite/ld-tic6x/tic6x.exp,
* testsuite/ld-undefined/undefined.exp,
* testsuite/ld-undefined/weak-undef.exp,
* testsuite/lib/ld-lib.exp: Don't use unresolved except after
perror. Instead report "unsupported" or "fail".
This is needed to move to automake & its dejagnu-provided logic,
and eventually by the unified sim logic. The $arch is used only
to figure out which `run` program to use when running tests, and
as we move to a single top-level build, we can delete this and
use sim/run directly.
Current toolchains warn about unused result from fread, so mitigate
the edge case if fread returns short data. It's not great, but it
gets things building again.
Rather than hand maintain m4 includes in various autotool files,
use AC_CONFIG_MACRO_DIRS to declare the relevant search paths.
This simplifies the code, makes it more robust, and cleans out
unused logic from configure.
Any code using AC_DEBUGINFOD from this dir is using -I../config when
running aclocal, so an explicit include on pkg.m4 is unnecessary:
aclocal will find the pkg.m4 in this dir just as easily. This is
seen in the only two dirs that use AC_DEBUGINFOD (binutils & gdb)
as their aclocal.m4 already has m4_include on config m4 files.
The include as written only works if aclocal is run on a dir that is
at the same level of config/. Any other depth will fail.
./
|-- config/
|-- binutils/ # works
|-- gdb/ # works
`-- sim/ # works
`-- <port>/ # fails
It fails even if AC_DEBUGINFOD itself isn't used:
sim/bfin/ $ aclocal -I../../config
aclocal-1.15: error: ../../config/debuginfod.m4:8: file '../config/pkg.m4' does not exist
I've been using gdbreplay to help debug an intermittent failure, and I
wanted it to be a little simpler to use. This patch adds support for
"-" as the "address" argument. With this patch you can do:
(gdb) target remote | gdbreplay logfile -
... and not have to start gdbreplay in a separate shell.
2021-02-12 Tom Tromey <tromey@adacore.com>
* gdbreplay.cc (remote_desc): Remove.
(remote_desc_in, remote_desc_out): New globals.
(remote_close): Update.
(remote_open): Handle "-".
(remote_open): Update.
(logchar): Log to stderr.
(expect, play): Update.
When running test-case gdb.threads/create-fail.exp on openSUSE Factory
(with glibc version 2.32) I run into:
...
(gdb) continue
Continuing.
[New Thread 0x7ffff7c83700 (LWP 626354)]
[New Thread 0x7ffff7482700 (LWP 626355)]
[Thread 0x7ffff7c83700 (LWP 626354) exited]
[New Thread 0x7ffff6c81700 (LWP 626356)]
[Thread 0x7ffff7482700 (LWP 626355) exited]
[New Thread 0x7ffff6480700 (LWP 626357)]
[Thread 0x7ffff6c81700 (LWP 626356) exited]
[New Thread 0x7ffff5c7f700 (LWP 626358)]
[Thread 0x7ffff6480700 (LWP 626357) exited]
pthread_create: 22: Invalid argument
Thread 6 "create-fail" received signal SIG32, Real-time event 32.
[Switching to Thread 0x7ffff5c7f700 (LWP 626358)]
0x00007ffff7d87695 in clone () from /lib64/libc.so.6
(gdb) FAIL: gdb.threads/create-fail.exp: iteration 1: run till end
...
The problem is that glibc-internal signal SIGCANCEL is not recognized by gdb.
There's code in check_thread_signals that is supposed to take care of that,
but it's not working because this code in lin_thread_get_thread_signals has
stopped working:
...
/* NPTL reserves the first two RT signals, but does not provide any
way for the debugger to query the signal numbers - fortunately
they don't change. */
sigaddset (set, __SIGRTMIN);
sigaddset (set, __SIGRTMIN + 1);
...
Since glibc commit d2dc5467c6 "Filter out NPTL internal signals (BZ #22391)"
(first released as part of glibc 2.28), a sigaddset with a glibc-internal
signal has no other effect than setting errno to EINVALID.
Fix this by eliminating the usage of sigset_t in check_thread_signals and
lin_thread_get_thread_signals.
The same problem was observed on Ubuntu 20.04.
Tested on x86_64-linux, openSUSE Factory.
Tested on aarch64-linux, Ubuntu 20.04 and Ubuntu 18.04.
gdb/ChangeLog:
2021-02-12 Tom de Vries <tdevries@suse.de>
PR threads/26228
* linux-nat.c (lin_thread_get_thread_signals): Remove.
(lin_thread_signals): New static var.
(lin_thread_get_thread_signal_num, lin_thread_get_thread_signal):
New function.
* linux-nat.h (lin_thread_get_thread_signals): Remove.
(lin_thread_get_thread_signal_num, lin_thread_get_thread_signal):
Declare.
* linux-thread-db.c (check_thread_signals): Use
lin_thread_get_thread_signal_num and lin_thread_get_thread_signal.
* configure.ac (follow-debug-links): Add option to enable or
disable the following of debug links by default. Set the
default for the option to be 'follow'.
* dwarf.c (do_follow_links): Initialise with DEFAULT_FOR_FOLLOW_LINKS.
(dwarf_select_sections_by_names): Add no-follow-links option.
(dwarf_select_sections_by_letter): Add 'N' option.
* objdump.c (usage): Add conditional text describing the
follow links option.
(slurp_symtab): Ensure that there is a NULL entry at the end
of the symbol table.
(slurp_dynamic_symtab): Likewise.
(dump_bfd): When extending the symbol table, ensure that there
is still a NULL entry at the end.
* readelf.c (usage): Add conditional text describing the
follow links option.
* doc/binutils.texi: Update documentation for objcopy and
readelf.
* doc/debug.options.texi: Update documentation of the
follow-links option.
* config.in: Regenerate.
* configure: Regenerate.
* testsuite/binutils-all/compress.exp: Add the -WN option to
objdump command lines that are not expecting to follow links.
* testsuite/binutils-all/readelf.exp: Add the
--debug-dump=no-follow-links option to tests that are not
expecting to follow debug links.
gas * testsuite/gas/mach-o/sections-1.d: Stop automatic debug link
following.
* testsuite/gas/xgate/insns-dwarf2.d: Likewise.
ld * testsuite/ld-elf/sec64k.exp: Stop readelf from automatically
following debug links.
You'd think "unresolved" would be correct for an objcopy test when the
assembler refuses to assemble one of our source files. After all, the
test of objcopy hasn't been run. However, "unresolved" results in
runtest returning with an error status. If instead we report
"unsupported", runtest returns success. Which is a little less
confusing to a user who doesn't see any errors reported unless they
look in log files.
* testsuite/binutils-all/objcopy.exp: Report "unsupported" when
gas or ld fails to build a testcase rather than "unresolved".
Report "fail" when readelf returns an error status rather than
"unresolved".
* testsuite/binutils-all/ar.exp: Likewise.
* testsuite/binutils-all/compress.exp: Likewise.
* testsuite/binutils-all/readelf.exp: Likewise.
This fixes the test for z80, which predefine register symbols "a" and
"c" among others.
* testsuite/binutils-all/pr25662.s: Replace "a" with "aaa" and
"c" with "ccc" labels.
Some hppa gas targets treat anything starting in the first column as a
label, so directives can't start there. Also, binutils_assemble and
run_dump_test cleverly edit test source to suit the hppa .comm
directive which has a different syntax to most targets. The editing
means we can't match source file names in dumps. Finally, hppa gas
complains if instructions are emitted without a ".text" or similar
directive.
* testsuite/gas/all/pr27381.err: Don't match source file name.
* testsuite/gas/all/pr27381.s: Don't start directive in first column.
* testsuite/gas/all/pr27384.err: Don't match source file name.
Adjust line number.
* testsuite/gas/all/pr27384.s: Add ".text" directive.
* testsuite/gas/elf/pr27355.err: Don't match source file name.
With exec:
...
$ clang -gdwarf-5 ./src/gdb/testsuite/gdb.dwarf2/fission-mix*.c
...
we have:
...
$ readelf -w a.out
...
Contents of the .debug_str_offsets section:
Length: 0x24
Version: 0x5
Index Offset [String]
0 1d0 clang version 10.0.1
1 1e6 src/gdb/testsuite/gdb.dwarf2/fission-mix-2.c
2 213 /home/vries/gdb_versions/devel
3 232 bar
4 236 x
5 61 int
6 238 s
7 23a func2
8 2c ild/BUILD/glibc-2.26/csu
9 5 sdeps/x86_64/start.S
10 1d0 clang version 10.0.1
11 240 src/gdb/testsuite/gdb.dwarf2/fission-mix.c
12 213 /home/vries/gdb_versions/devel
13 26b foo
14 236 x
15 61 int
16 238 s
17 26f func
18 274 main
19 279 arg
...
The section consists of two parts, one for each CU, each with a header, but
the printing only reads the first header as a header, and prints the second
header as:
...
8 2c ild/BUILD/glibc-2.26/csu
9 5 sdeps/x86_64/start.S
...
Fix this in display_debug_str_offsets such that we have:
...
6 238 s
7 23a func2
Length: 0x2c
Version: 0x5
Index Offset [String]
0 1d0 clang version 10.0.1
1 240 src/gdb/testsuite/gdb.dwarf2/fission-mix.c
...
binutils/ChangeLog:
2021-02-12 Tom de Vries <tdevries@suse.de>
* dwarf.c (display_debug_str_offsets): Handle multiple sets of
entries.
With exec:
...
$ gcc -gsplit-dwarf ~/hello.c -gdwarf-5
...
a dwarf-5 DW_UT_skeleton CU is generated, but the corresponding DWO ID is not
printed by readelf -wi.
Add this, such that we have:
....
Compilation Unit @ offset 0xc7:
Length: 0x31 (32-bit)
Version: 5
Unit Type: DW_UT_skeleton (4)
Abbrev Offset: 0x64
Pointer Size: 8
+ DWO ID: 0x4756ae3ac4348f21
<0><db>: Abbrev Number: 1 (DW_TAG_skeleton_unit)
...
binutils/ChangeLog:
2021-02-12 Tom de Vries <tdevries@suse.de>
* dwarf.c (process_debug_info): Print DWO ID.
When running test-case gdb.dwarf2/fission-mix.exp using gcc-11, I run into:
...
(gdb) file fission-mix^M
Reading symbols from fission-mix...^M
src/gdb/dwarf2/attribute.h:258: internal-error: \
void attribute::set_unsigned_reprocess(ULONGEST): \
Assertion `form_requires_reprocessing ()' failed.^M
...
This happens when calling set_unsigned_reprocess on an attribute with form
DW_FORM_strx. The assert triggers because DW_FORM_strx is not listed in
form_requires_reprocessing.
Fix this by adding DW_FORM_strx in form_requires_reprocessing.
Tested on x86_64-linux.
gdb/ChangeLog:
2021-02-11 Tom de Vries <tdevries@suse.de>
PR symtab/27353
* dwarf2/attribute.c (attribute::form_requires_reprocessing):
Return true for DW_FORM_strx.
Luis pointed out that an earlier patch of mine caused two regressions
in gdb.fortran. This patch fixes the problem.
Regression tested on x86-64 Fedora 32.
gdb/ChangeLog
2021-02-11 Tom Tromey <tromey@adacore.com>
PR gdb/27383:
* parse.c (write_exp_symbol_reference): Write sym.block.
In commit:
commit e92c8eb86d
Date: Tue Feb 9 15:46:13 2021 +0000
gdb/fortran: add parser support for lbound and ubound
When I created the test gdb/testsuite/gdb.fortran/lbound-ubound.exp, I
copied the script from a different file and failed to delete the test
description comment at the top (even though I added a new
description). Fixed in this commit.
gdb/testsuite/ChangeLog:
* gdb.fortran/lbound-ubound.exp: Remove old comment.
