While looking into the regressions reported by Luis Machado, I noticed
that null pathnames were being output in the warnings. E.g.
warning: Can't open file (null) during file-backed mapping note processing
I've changed the warning to output the pathname found in the note,
like this:
warning: Can't open file /var/lib/docker/aufs/diff/d07c...e21/lib/x86_64-linux-gnu/libc-2.27.so during file-backed mapping note processing
(I've shortened one of the path elements above.)
gdb/ChangeLog:
* corelow.c (core_target::build_file_mappings): Don't output
null pathname in warning.
I noticed some gdb.dwarf2 tests not running on my machine because of
this:
Running /home/simark/src/binutils-gdb/gdb/testsuite/gdb.dwarf2/dw2-single-line-discriminators.exp ...
gdb compile failed, /usr/bin/ld: /home/simark/build/binutils-gdb/gdb/testsuite/outputs/gdb.dwarf2/dw2-single-line-discriminators/dw2-single-line-discriminators0.o: relocation R_X86_64_32S against
symbol `x' can not be used when making a PIE object; recompile with -fPIE
collect2: error: ld returned 1 exit status
We get this when the target toolchain produces position-independent
executables by default. These tests are built from some assembly which
produces some relocations incompatible with position-independent
executables.
Add `nopie` to the compilation flags of these tests to force the
toolchain to produce non-position-independent executables. With this,
the changed tests run successfully on my machine.
gdb/ChangeLog:
* gdb.dwarf2/clztest.exp, gdb.dwarf2/dw2-common-block.exp,
gdb.dwarf2/dw2-dup-frame.exp, gdb.dwarf2/dw2-reg-undefined.exp,
gdb.dwarf2/dw2-single-line-discriminators.exp,
dw2-undefined-ret-addr.exp: Pass nopie to compilation options.
Change-Id: Ie06c946f8e56a6030be247d1c57f416fa8b67e4c
Older Rust compilers used special field names, rather than DWARF
features, to express the variant parts of Rust enums. This is handled
in gdb through a quirk recognizer that rewrites the types.
Tom de Vries pointed out in PR rust/26197 that the variant part
rewrite regressed this code. This patch fixes the problems:
* Univariant enums were not handled properly. Now we simply call
alloc_rust_variant for these as well.
* There was an off-by-one error in the handling of ordinary enums.
* Ordinary enums should have the size of their member types reset to
match the size of the enclosing enum. (It's not clear to me if this
is truly necessary, but it placates a test, and this is just legacy
handling in any case.)
Tested with Rust 1.12.0, 1.14.0, 1.19.0, 1.36.0, and 1.45.0 on x86-64
Fedora 32. There were some unrelated failures with 1.14.0 and 1.19,0;
but considering that these are fairly old releases, I don't plan to
look into them unless someone complains.
Note that this patch will not fix all the issues in the PR. In that
PR, Tom is using a somewhat unusual build of Rust -- in particular it
uses an older (pre-DWARF variant part) LLVM with a newer Rust. I
believe this compiler doesn't correctly implement the old-style name
fallback; the details are in the bug.
gdb/ChangeLog
2020-08-05 Tom Tromey <tromey@adacore.com>
PR rust/26197:
* dwarf2/read.c (alloc_rust_variant): Handle univariant case.
(quirk_rust_enum): Call alloc_rust_variant for univariant case.
Fix off-by-one and type size errors in ordinary case.
Operand sizes used for simulation of MSP430 hardware multiply
operations are not aligned with the sizes used on the target, resulting
in the simulator storing signed operands with too much precision.
Additionally, simulation of unsigned multiplication is missing explicit
casts to prevent any implicit sign extension.
gcc.c-torture/execute/pr91450-1.c uses unsigned widening multiplication
of 32-bit operands -4 and 2, to produce a 64-bit result:
0xffff fffc * 0x2 = 0x1 ffff fff8
If -4 is stored in 64-bit precision, then the multiplication is
essentially signed and the result is -8 in 64-bit precision
(0xffff ffff ffff fffc), which is not correct.
sim/msp430/ChangeLog:
* msp430-sim.c (put_op): For unsigned multiplication, explicitly cast
operands to the unsigned type before multiplying.
* msp430-sim.h (struct msp430_cpu_state): Fix types used to store hwmult
operands.
sim/testsuite/sim/msp430/ChangeLog:
* mpyull_hwmult.s: New test.
After commit 66d6346b25 "gdb: remove TYPE_DYN_PROP_ADDR", I run into:
...
FAIL: gdb.fortran/class-allocatable-array.exp: print this%_data%b
...
(and 185 more FAILs, all for fortran test-cases).
The commit replaces "!x" by "x != 0".
Fix this by using "x == 0" instead.
Build and tested on x86_64-linux.
gdb/ChangeLog:
2020-08-05 Tom de Vries <tdevries@suse.de>
* gdbtypes.c (type_not_allocated, type_not_associated): Use
"prop->const_val () == 0" instead of "prop->const_val () != 0".
A malloc failure triggered by a fuzzed object file isn't a real
problem unless objdump doesn't exit cleanly after the failure, which
it does. However we have bfd_malloc_and_get_section to sanity check
size of uncompressed sections before allocating memory. Use it.
PR 26337
* objdump.c (load_specific_debug_section): Don't malloc space for
section contents, use bfd_malloc_and_get_section.
Problem found by Tadashi G. Takaoka.
2020-08-04 Christian Groessler <chris@groessler.org>
Tadashi G. Takaoka <tadashi.g.takaoka@gmail.com>
* z8kgen.c (opt): Fix "sout imm16,rs" and "soutb imm16,rbs"
opcodes (special "out" to absolute address).
* z8k-opc.h: Regenerate.
2020-08-04 Christian Groessler <chris@groessler.org>
* gas/testsuite/gas/z8k/inout.d: Adapt to correct encoding of
"sout/soutb #imm,reg"
Change instances of int variables and return values used as boolean
values to use the bool type.
Shorten the comments of a few functions, because I think they go a bit
too much in implementation details, which appear out of date anyway.
Make other misc changes to the functions that are already being changed,
such as using nullptr instead of NULL, dropping `struct` keywords and
declaring variables when first used.
gdb/ChangeLog:
* frame.h (frame_id_p): Return bool.
(frame_id_artificial_p): Return bool.
(frame_id_eq): Return bool.
(has_stack_frames): Return bool.
(get_selected_frame): Fix typo in comment.
(get_frame_pc_if_available): Return bool.
(get_frame_address_in_block_if_available): Return bool.
(get_frame_func_if_available): Return bool.
(read_frame_register_unsigned): Return bool.
(get_frame_register_bytes): Return bool.
(safe_frame_unwind_memory): Return bool.
(deprecated_frame_register_read): Return bool.
(frame_unwinder_is): Return bool.
* frame.c (struct frame_info) <prev_arch::p>: Change type to
bool.
<this_id::p>: Likewise.
<prev_p>: Likewise.
(frame_stash_add): Return bool.
(get_frame_id): Use bool.
(frame_id_build_special) Use bool.
(frame_id_build_unavailable_stack): Use bool.
(frame_id_build): Use bool.
(frame_id_p): Return bool, use true/false instead of 1/0.
(frame_id_artificial_p): Likewise.
(frame_id_eq): Likewise.
(frame_id_inner): Likewise.
(get_frame_func_if_available): Likewise.
(read_frame_register_unsigned): Likewise.
(deprecated_frame_register_read): Likewise.
(get_frame_register_bytes): Likewise.
(has_stack_frames): Likewise.
(inside_main_func): Likewise.
(inside_entry_func): Likewise.
(get_frame_pc_if_available): Likewise.
(get_frame_address_in_block_if_available): Likewise.
(frame_unwinder_is): Likewise.
(safe_frame_unwind_memory): Likewise.
(frame_unwind_arch): Likewise.
Change-Id: I6121fa56739b688be79d73d087d76b268ba5a46a
One might think that variable `frame_info::prev_func::p` is a simple
true/false value, but that's not the case, it can also have the value -1
to mean "unavaiable". Change it to use the `cached_copy_status` enum,
which seems designed exactly for this purpose.
Rename to `status` to be consistent with `prev_pc::status` (and be cause
`p` means `predicate`, which implies boolean, which this is not).
gdb/ChangeLog:
* frame.c (frame_info) <prev_func> <p>: Rename to status, change
type to cached_copy_status.
(fprintf_frame): Adjust.
(get_frame_func_if_available): Adjust.
(frame_cleanup_after_sniffer): Adjust.
Change-Id: I50c6ebef6c0acb076e25c741f7f417bfd101d953
This patch adds basic support for the eBPF target: tdep and build
machinery. The accompanying simulator is introduced in subsequent
patches.
gdb/ChangeLog:
2020-08-04 Weimin Pan <weimin.pan@oracle.com>
Jose E. Marchesi <jose.marchesi@oracle.com>
* configure.tgt: Add entry for bpf-*-*.
* Makefile.in (ALL_TARGET_OBS): Add bpf-tdep.o
(ALLDEPFILES): Add bpf-tdep.c.
* bpf-tdep.c: New file.
* MAINTAINERS: Add bpf target and maintainer.
gdb/doc/ChangeLog:
2020-08-04 Jose E. Marchesi <jose.marchesi@oracle.com>
* gdb.texinfo (Contributors): Add information for the eBPF
support.
(BPF): New section.
In the check-test-names.exp library 'unset' was being used to unset an
array variable. Though this seems to work fine on tcl 8.6, it was
discovered on a CentOS 7.8.2003 machine, running tcl 8.5, that this
doesn't work and 'array unset' should be used instead.
Using 'array unset' should work fine for newer and older versions of
tcl (since 8.3, releases ~2000).
gdb/testsuite/ChangeLog:
* lib/check-test-names.exp (do_reset_vars): Use 'array unset' to
unset the array variable.
For DWARF4 DW_AT_high_pc can be expressed as constant offset from
DW_AT_low_pc which saves a relocation. Use DW_FORM_udate (uleb128)
to keep the constant value as small as possible.
gas/ChangeLog:
* dwarf2dbg.c (out_debug_abbrev): When DWARF2_VERSION >= 4, use
DW_FORM_udata for DW_AT_high_pc.
(out_debug_info): Use emit_leb128_expr for DW_AT_high_pc, when
DWARF2_VERSION >= 4.
* read.c (emit_leb128_exp): No longer static.
* read.h (emit_leb128_exp): Define.
For DWARF5 the zero file list entry in the .debug_line table represents
the compile unit main file. It can be set with .file 0 when -gdwarf-5
is given. But since this directive is illegal for older versions, this
is almost never set. To make sure it is always set (so DW_AT_name of
the compile unit can be set) use file (and dir) 1 if that is defined
(otherwise fall back to pwd, to match DW_AT_comp_dir).
gas/ChangeLog:
* gas/dwarf2dbg.c (out_dir_and_file_list): For DWARF5 emit at
least one directory if there is at least one file. Use dirs[1]
if dirs[0] is not set, or if there is no dirs[1] the current
working directory. Use files[1] filename, when files[0] filename
isn't set.
DWARF5 CU headers have a new unit type field and move the abbrev offset
to the end of the header.
gas/ChangeLog:
* dwarf2dbg.c (out_debug_info): Emit unit type and abbrev offset
for DWARF5.
* gas/testsuite/gas/elf/dwarf-4-cu.d: New file.
* gas/testsuite/gas/elf/dwarf-4-cu.s: Likewise.
* gas/testsuite/gas/elf/dwarf-5-cu.d: Likewise.
* gas/testsuite/gas/elf/dwarf-5-cu.s: Likewise.
* testsuite/gas/elf/elf.exp: Run dwarf-4-cu and dwarf-5-cu.
When reverting commit 9cfd2b89bd "[gdb/testsuite] Fix
gdb.arch/amd64-entry-value-paramref.S", we run into an internal-error:
...
(gdb) file amd64-entry-value-paramref^M
Reading symbols from amd64-entry-value-paramref...^M
src/gdb/dwarf2/read.c:18903: internal-error: could not find partial DIE
0x1b7 in cache [from module amd64-entry-value-paramref]^M
A problem internal to GDB has been detected,^M
further debugging may prove unreliable.^M
...
because of invalid dwarf.
In contrast, when using -readnow, we have:
...
(gdb) file -readnow amd64-entry-value-paramref
Reading symbols from amd64-entry-value-paramref...
Expanding full symbols from amd64-entry-value-paramref...
Dwarf Error: Cannot find DIE at 0x1b7 referenced from DIE at 0x11a \
[in module amd64-entry-value-paramref]
(gdb)
...
Change the internal error into a Dwarf Error, such that we have:
...
(gdb) file amd64-entry-value-paramref^M
Reading symbols from amd64-entry-value-paramref...^M
Dwarf Error: Cannot not find DIE at 0x1b7 \
[from module amd64-entry-value-paramref]^M
^M
(No debugging symbols found in amd64-entry-value-paramref)^M
(gdb)
...
Build and tested on x86_64-linux.
gdb/ChangeLog:
2020-08-04 Tom de Vries <tdevries@suse.de>
PR symtab/23270
* dwarf2/read.c (find_partial_die): Change internal error into Dwarf
Error.
This matches the current set of system calls in the FreeBSD head
development branch as of r363367. Some of these system calls
were also included in 12.1 release.
gdb/ChangeLog:
* syscalls/freebsd.xml: Regenerate.
The MSP430 GAS option "-md" is supposed to indicate that the CRT startup
code should copy data from ROM to RAM at startup. However, this option
has no effect; GAS handles the related behaviour automatically by
looking for the presence of certain symbols in the input file.
gas/ChangeLog:
* config/tc-msp430.c (OPTION_MOVE_DATA): Remove.
(md_parse_option): Remove case for OPTION_MOVE_DATA.
(md_longopts): Remove "md" entry.
(md_show_usage): Likewise.
When reading an exec with a .debug_line section containing a vendor-specific
extended opcode, we get:
...
$ gdb -batch -iex "set complaints 10" dw2-vendor-extended-opcode
During symbol reading: mangled .debug_line section
...
and reading of the .debug_line section is abandoned.
The vendor-specific extended opcode should be ignored, as specified in the
DWARF standard (7.1 Vendor Extensibility). [ FWIW, vendor-specific
standard opcodes are already ignored. ]
Fix this by ignoring all vendor-specific extended opcodes.
Build and tested on x86_64-linux.
gdb/ChangeLog:
2020-08-03 Tom de Vries <tdevries@suse.de>
PR symtab/26333
* dwarf2/read.c (dwarf_decode_lines_1): Ignore
DW_LNE_lo_user/DW_LNE_hi_user range.
gdb/testsuite/ChangeLog:
2020-08-03 Tom de Vries <tdevries@suse.de>
PR symtab/26333
* lib/dwarf.exp (DW_LNE_user): New proc.
* gdb.dwarf2/dw2-vendor-extended-opcode.c: New test.
* gdb.dwarf2/dw2-vendor-extended-opcode.exp: New file.
As far as I can tell, the following comment is false nowadays.
/* Calls to m-alloc get turned by sed into xm-alloc. */
Remove it, and call xmalloc.
* ldlex.l (yy_create_string_buffer): Use xmalloc rather than malloc.
* lexsup.c (parse_args): Likewise.
There are compilation warnings / errors when compiling coremaker2.c
for the gdb.base/corefile2.exp tests. Here's the command to use
on x86_64 linux:
make check RUNTESTFLAGS="--target_board unix/-m32" \
TESTS="gdb.base/corefile2.exp"
These are the warnings / errors - I've shortened the paths somewhat:
gdb compile failed, gdb/testsuite/gdb.base/coremaker2.c: In function 'main':
gdb/testsuite/gdb.base/coremaker2.c:106:11: warning: cast from pointer to integer of different size [-Wpointer-to-int-cast]
106 | addr = ((unsigned long long) buf_ro + pagesize) & ~(pagesize - 1);
| ^
gdb/testsuite/gdb.base/coremaker2.c:108:15: warning: cast from pointer to integer of different size [-Wpointer-to-int-cast]
108 | if (addr <= (unsigned long long) buf_ro
| ^
gdb/testsuite/gdb.base/coremaker2.c:109:18: warning: cast from pointer to integer of different size [-Wpointer-to-int-cast]
109 | || addr >= (unsigned long long) buf_ro + sizeof (buf_ro))
| ^
gdb/testsuite/gdb.base/coremaker2.c:115:19: warning: cast to pointer from integer of different size [-Wint-to-pointer-cast]
115 | mbuf_ro = mmap ((void *) addr, pagesize, PROT_READ,
| ^
gdb/testsuite/gdb.base/coremaker2.c:130:11: warning: cast from pointer to integer of different size [-Wpointer-to-int-cast]
130 | addr = ((unsigned long long) buf_rw + pagesize) & ~(pagesize - 1);
| ^
gdb/testsuite/gdb.base/coremaker2.c:132:15: warning: cast from pointer to integer of different size [-Wpointer-to-int-cast]
132 | if (addr <= (unsigned long long) buf_rw
| ^
gdb/testsuite/gdb.base/coremaker2.c:133:18: warning: cast from pointer to integer of different size [-Wpointer-to-int-cast]
133 | || addr >= (unsigned long long) buf_rw + sizeof (buf_rw))
| ^
gdb/testsuite/gdb.base/coremaker2.c:139:19: warning: cast to pointer from integer of different size [-Wint-to-pointer-cast]
139 | mbuf_rw = mmap ((void *) addr, pagesize, PROT_READ,
| ^
These were fixed by changing unsigned long long to uintptr_t.
Tested on either rawhide or Fedora 32 with architectures: x86_64,
x86_64/-m32, aarch64, s390x, and ppc64le.
gdb/testsuite/ChangeLog:
* gdb.base/coremaker2.c: Change all uses of 'unsigned long long'
to 'uintptr_t'
(inttypes.h): Include.
It turns out that the recently added gdb.base/corefile2.exp test won't
run on ppc64le linux. The test case fails the internal checks which
ensure that a mmap'd region can be placed within the statically
allocated regions buf_rw[] and buf_ro[].
ppc64le linux apparently has 64k pages, which is much larger than
the 24k regions originally allocated for buf_rw[] and buf_ro[].
This patch increases the size of each region to 256 KiB.
Tested on either rawhide or Fedora 32 for these architectures: x86_64,
x86_64/-m32, ppc64le, aarch64, and s390x.
gdb/testsuite/ChangeLog:
* gdb.base/coremaker2.c (buf_rw): Increase size to 256 KiB.
(C5_24k): Delete.
(C5_8k, C5_64k, C5_256k): New macros.
(buf_ro): Allocate 256 KiB of initialized data.
LTO can be used to build binutils with
$ CC="gcc -flto -ffat-lto-objects -Wl,--as-needed" CXX="g++ -flto -ffat-lto-objects -Wl,--as-needed" .../configure
But not all linker tests are compatible with LTO. Pass -fno-lto to CC
to disable LTO on linker tests by default. -flto is passed explicitly
to CC in linker LTO tests.
* testsuite/ld-elf/indirect.exp: Append -fno-lto to CC.
* testsuite/ld-elfvers/vers.exp: Likewise.
* testsuite/ld-elfweak/elfweak.exp: Likewise.
* testsuite/ld-ifunc/ifunc.exp: Likewise.
* testsuite/ld-plugin/lto.exp (no_lto): New.
Add $no_lto to build pr15146c.so.
* testsuite/lib/ld-lib.exp (at_least_gcc_version): Filter out
-Wl,xxx options.
(check_gcc_plugin_enabled): Likewise.
(run_ld_link_exec_tests): Prepend -fno-lto to $cflags.
(run_cc_link_tests): Likewise.
gcc -O2 -g -o ar -Wl,--as-needed arparse.o arlex.o ar.o not-ranlib.o arsup.o rename.o binemul.o emul_vanilla.o bucomm.o version.o filemode.o libbfd-2.35-3.fc33.so libiberty.a -Wl,-R,.
All of the above .o files are lto, leading to libbfd-2.35-3.fc33.so
not being found needed when loading the IR objects. That's problem
number one: We exclude IR references when deciding a shared library
is needed. See PR15146. Thus none of the libbfd.so symbols are
loaded before libiberty.a is scanned, and libbfd.so contains copies of
libiberty.a functions. We ought to be using the libbfd.so copies
rather than extracting them from the archive (an object is extracted
even to satisfy IR symbols). After lto recompilation, libbfd.so is of
course found to be needed and loaded. But that causes more problems.
The lto recompilation didn't see symbol references from libbfd.so and
variables like _xexit_cleanup are made local in the recompiled
objects. Oops, two copies of them. Finally, those silly undefined
symbols in the lto output debug files, combined with definitions in
both libbfd.so and IR objects result in IR symbols being made
dynamic.
The main fix here is to revert the PR15146 change to
elf_link_add_object_symbols.
PR 26314
* elflink.c (bfd_elf_link_record_dynamic_symbol): Don't allow
IR symbols to become dynamic.
(elf_link_add_object_symbols): Don't exclude IR symbols when
deciding whether an as-needed shared library is needed.
With this patch, ld/pr24511 test passes for ARC.
At first glance, the test was failing because the order of
"__init_array_start" and "__fini_array_start" weak symbols were
reversed:
$ nm -n dump.out
expected output | real output
00002104 D __init_array_start | 00002104 D __fini_array_start
0000210c D __fini_array_start | 00002104 D __init_array_start
The order of the symbols are different as a side effect of both
symbols being mapped to the _same_ address (0x2104). Looking
further into the mapping logs [1] revealed that the linker
script must consider all instances of ".init_array" (in other
words ".init_array.*") inside its relevant section. Same logic
holds for ".fini_array".
Therefore, adding "KEEP (*(SORT(.init_array.*)))" to the linker
script, along with the one for ".finit_array.*", resolved the
problem. While at it, I took the liberty of refactoring the
script a little bit and made those pieces of script macros.
[1] Linker's mapping for the relevant part of the test
---------------------------------------------------------------
.init_array 0x2104 0x0
0x2104 PROVIDE (__init_array_start = .)
*(.init_array)
[!provide] PROVIDE (__init_array_end = .)
.fini_array 0x2104 0x0
0x2104 PROVIDE (__fini_array_start = .)
*(.fini_array)
[!provide] PROVIDE (__fini_array_end = .)
.data 0x2104 0x0
*(.data .data.* .gnu.linkonce.d.*)
.data 0x2104 0x0 pr24511.o
.init_array.01000
0x2104 0x8
.init_array.01000
0x2104 0x8 pr24511.o
.fini_array.01000
0x210c 0x8
.fini_array.01000
0x210c 0x8 pr24511.o
---------------------------------------------------------------
ld:
* scripttempl/elfarc.sc (.init_array): Keep ".init_array.*".
(.fini_array): Keep ".fini_array.*".
Signed-off-by: Claudiu Zissulescu <claziss@gmail.com>
PR 26318 shows that running `maint print symbols` on an Ada binary,
compiled with an Ada distribution that includes debug info for the
standard library, triggers this assertion:
/home/simark/src/binutils-gdb/gdb/gdbtypes.h:526: internal-error: LONGEST dynamic_prop::const_val() const: Assertion `m_kind == PROP_CONST' failed.
The problem is in particular when printing type
`system__object_reader__decoded_ada_name__TTdecodedSP1___XDL_0`, which
has a dynamic high bound (PROP_LOCLIST kind). When printing a concrete
value of this type, this type gets resolved to a type with a constant
high bound, so ada_modulus can return this constant value.
However, when printing the dynamic range type on its own, such as with
`maint print symbols`, the high bound is still of kind PROP_LOCLIST.
When ada_modulus tries to access the property as a const value, the
assert triggers.
There's no sensible numerical value to return in this case. Ideally,
ada_modulus would return something to the caller indicating that the
value is dynamic and therefore can't be returned as an integer. The
callers would handle it, for example `maint print symbols` would say
that the high bound of the type is dynamic.
However, this patch implements the simpler fix of returning 0 in that
case. It kind of restores the previous behavior of before we validated
the dynamic property kind in the getters, where we would just return
whatever random integer value was in `const_val`. Except now it's
consistently 0.
This is what we had before we added dynamic property getters:
$ ./gdb -q ~/foo -ex "maint expand-symtabs" -ex "maint print symbols" -batch | grep 'typedef <system__object_reader__decoded_ada_name__TTdecodedSP1'
typedef <system__object_reader__decoded_ada_name__TTdecodedSP1: mod 107820865988257;
and this is what we have now:
$ ./gdb -q ~/foo -ex "maint expand-symtabs" -ex "maint print symbols" -batch | grep 'typedef <system__object_reader__decoded_ada_name__TTdecodedSP1'
typedef <system__object_reader__decoded_ada_name__TTdecodedSP1: mod 0;
The value 107820865988257 is the `baton` field of the property's union
interpreted as an integer, so a bogus value.
gdb/ChangeLog:
PR ada/26318
* ada-lang.c (ada_modulus): Return 0 if property is not of const
kind.
Change-Id: I3f6d343a9c3cd7cd62a4fc591943a43541223d50
Apply minor refactoring to 'set_breakpoint_condition'.
gdb/ChangeLog:
2020-07-30 Tankut Baris Aktemur <tankut.baris.aktemur@intel.com>
* breakpoint.c (set_breakpoint_condition): Do minor refactoring.
In 'set_breakpoint_condition', GDB resets the condition expressions
before parsing the condition input by the user. This leads to the
problem of losing the condition expressions if the new condition
does not parse successfully and is thus rejected.
For instance:
$ gdb ./test
Reading symbols from ./test...
(gdb) start
Temporary breakpoint 1 at 0x114e: file test.c, line 4.
Starting program: test
Temporary breakpoint 1, main () at test.c:4
4 int a = 10;
(gdb) break 5
Breakpoint 2 at 0x555555555155: file test.c, line 5.
Now define a condition that would evaluate to false. Next, attempt
to overwrite that with an invalid condition:
(gdb) cond 2 a == 999
(gdb) cond 2 gibberish
No symbol "gibberish" in current context.
(gdb) info breakpoints
Num Type Disp Enb Address What
2 breakpoint keep y 0x0000555555555155 in main at test.c:5
stop only if a == 999
It appears as if the bad condition is successfully rejected. But if we
resume the program, we see that we hit the breakpoint although the condition
would evaluate to false.
(gdb) continue
Continuing.
Breakpoint 2, main () at test.c:5
5 a = a + 1; /* break-here */
Fix the problem by not resetting the condition expressions before
parsing the condition input.
Suppose the fix is applied. A similar problem could occur if the
condition is valid, but has "junk" at the end. In this case, parsing
succeeds, but an error is raised immediately after. It is too late,
though; the condition expression is already updated.
For instance:
$ gdb ./test
Reading symbols from ./test...
(gdb) start
Temporary breakpoint 1 at 0x114e: file test.c, line 4.
Starting program: test
Temporary breakpoint 1, main () at test.c:4
4 int a = 10;
(gdb) break 5
Breakpoint 2 at 0x555555555155: file test.c, line 5.
(gdb) cond 2 a == 999
(gdb) cond 2 a == 10 if
Junk at end of expression
(gdb) info breakpoints
Num Type Disp Enb Address What
2 breakpoint keep y 0x0000555555555155 in main at test.c:5
stop only if a == 999
(gdb) c
Continuing.
Breakpoint 2, main () at test.c:5
5 a = a + 1; /* break-here */
(gdb)
We should not have hit the breakpoint because the condition would
evaluate to false.
Fix this problem by updating the condition expression of the breakpoint
after parsing the input successfully and checking that there is no
remaining junk.
gdb/ChangeLog:
2020-07-30 Tankut Baris Aktemur <tankut.baris.aktemur@intel.com>
* breakpoint.c (set_breakpoint_condition): Update the condition
expressions after checking that the input condition string parses
successfully and does not contain junk at the end.
gdb/testsuite/ChangeLog:
2020-07-30 Tankut Baris Aktemur <tankut.baris.aktemur@intel.com>
* gdb.base/condbreak-bad.exp: Extend the test with scenarios
that attempt to overwrite an existing condition with a condition
that fails parsing and also with a condition that parses fine
but contains junk at the end.
