The introductory comment for partial_symtab has been mildly incorrect
since the partial symtab code was changed to allocate partial symtabs
with "new". This patch removes the incorrect text.
gdb/ChangeLog
2020-02-21 Tom Tromey <tom@tromey.com>
* psympriv.h (struct partial_symtab): Update comment.
With gccgo-6/7, we have:
...
FAIL: gdb.go/methods.exp: setting breakpoint at main.T.Foo
XFAIL: gdb.go/methods.exp: going to first breakpoint \
(the program exited)
FAIL: gdb.go/methods.exp: setting breakpoint at (*main.T).Bar
XFAIL: gdb.go/methods.exp: going to second breakpoint \
(the program is no longer running)
...
And with gccgo-8/9/10, we have:
...
PASS: gdb.go/methods.exp: setting breakpoint 1
XFAIL: gdb.go/methods.exp: going to first breakpoint
FAIL: gdb.go/methods.exp: setting breakpoint at (*main.T).Bar
XFAIL: gdb.go/methods.exp: going to second breakpoint \
(the program exited)
...
The first test passes and fails with different messages:
...
FAIL: gdb.go/methods.exp: setting breakpoint at main.T.Foo
...
or:
...
PASS: gdb.go/methods.exp: setting breakpoint 1
...
Fix this by removing the explicit pass call and using the message argument for
gdb_breakpoint, for both breakpoint locations.
The setup of the xfails is non-specific:
...
setup_xfail "*-*-*" ;# mangling issues IIRC
...
so let's start with removing these.
The first FAIL with gccgo-6:
...
FAIL: gdb.go/methods.exp: setting breakpoint at main.T.Foo
...
is due an incorrect DW_AT_name attribute:
...
# <554> DW_AT_name : main.Foo.N6_main.T
...
Fix this by recognizing the incorrect attribute, and xfailing the test.
Furthermore, if setting the breakpoint fails, there's not much point in trying
to continue to the breakpoint:
...
FAIL: gdb.go/methods.exp: setting breakpoint at main.T.Foo
FAIL: gdb.go/methods.exp: going to first breakpoint (the program exited)
...
Fix this by skipping the second test if the first one fails, also for the
second breakpoint.
With gccgo-10, we manage to set the first breakpoint, but continuing to
breakpoint test fails:
...
PASS: gdb.go/methods.exp: setting breakpoint 1
FAIL: gdb.go/methods.exp: going to first breakpoint
...
This is due to an incorrect regexp, requiring a colon in front of the
breakpoint location. Fix this for both breakpoints.
Setting the second breakpoint fails:
...
FAIL: gdb.go/methods.exp: setting breakpoint at (*main.T).Bar
...
presumably because the breakpoint location "(*main.T).Bar" does not follow the
naming convention explained at https://golang.org/doc/gdb#Naming. Fix this by
updating the breakpoint location to "main.(*T).Bar".
Still this test fails, for gccgo-6/7 because of an incorrect DW_AT_name
attribute:
...
# <529> DW_AT_name : main.Bar.pN6_main.T
...
and for gccgo-8/9/10 because of incorrect DW_AT_name/DW_AT_linkage_name
attributes (filed as gcc PR93866):
...
# <6e5> DW_AT_name : main.Bar..1main.T
# <6ec> DW_AT_linkage_name: main.T.Bar
..
Add xfails for both of these.
All in all, now we have with gccgo-6/7:
...
XFAIL: gdb.go/methods.exp: setting breakpoint at main.T.Foo
XFAIL: gdb.go/methods.exp: setting breakpoint at main.(*T).Bar
...
and with gccgo-8/9/10, we have:
...
PASS: gdb.go/methods.exp: setting breakpoint at main.T.Foo
PASS: gdb.go/methods.exp: going to first breakpoint
XFAIL: gdb.go/methods.exp: setting breakpoint at main.(*T).Bar
...
Tested on x86_64-linux with gccgo-6/7/8/9/10.
gdb/testsuite/ChangeLog:
2020-02-21 Tom de Vries <tdevries@suse.de>
PR go/18926
* lib/gdb.exp (bp_location2/bp_location2_regexp): Fix.
Remove blanket xfails. Use message argument for gdb_breakpoint.
Make continuing to breakpoint test conditional on setting breakpoint.
Fix continuing to breakpoint regexp. Add xfails for gccgo-6/7
DW_AT_name attribute. Add xfail for GCC PR93866.
My earlier patch to fix a declaration mismatch in mips-tdep.h missed a
couple of spots. Basically, I sent it too soon.
This patch fixes a couple more mismatches between a declaration (using
bfd_vma) and the definition (using CORE_ADDR).
gdb/ChangeLog
2020-02-21 Tom Tromey <tromey@adacore.com>
* mips-tdep.h (mips_pc_is_mips16, mips_pc_is_micromips): Parameter
type is CORE_ADDR.
I noticed a regression with board cc-with-dwz:
...
FAIL: gdb.cp/m-static.exp: static const int initialized elsewhere
FAIL: gdb.cp/m-static.exp: info variable everywhere
...
The problem started with commit 0494dbecdf "Consolidate partial symtab
dependency reading".
The commit replaces the dwarf2_psymtab::expand_psymtab specific reading of
dependencies, which contains a "dependencies[i]->user == NULL" test, with a
generic partial_symtab::read_dependencies call, which does not test the user
field.
This patch fixes the regression by adding back the test, in the generic
partial_symtab::read_dependencies.
Build and reg-tested on x86_64-linux.
Tested natively, as well as with boards cc-with-dwz and cc-with-dwz-m.
The patch fixes all 33 regressions with cc-with-dwz, and all 2929 regression
with cc-with-dwz-m.
gdb/ChangeLog:
2020-02-21 Tom de Vries <tdevries@suse.de>
PR gdb/25534
* psymtab.c (partial_symtab::read_dependencies): Don't read dependency
if dependencies[i]->user != NULL.
This adds a test for the commit:
commit 4f180d5396
Date: Fri Feb 21 08:19:21 2020 -0700
Check for null result from gdb_demangle
gdb/testsuite/ChangeLog:
* gdb.dwarf2/cpp-linkage-name.c: New file.
* gdb.dwarf2/cpp-linkage-name.exp: New file.
I am sending this patch on behalf of kmoy@google.com, who discovered the bug
and wrote the fix.
gdb_demangle can return null for strings that don't properly demangle. The null
check was mistakenly removed in commit 43816ebc33. Without this check, GDB
aborts when loading symbols from some binaries.
gdb/ChangeLog
2020-02-21 Ali Tamur <tamur@google.com>
* dwarf2/read.c (dwarf2_name): Add null check.
It's very easy to make bfd/mach-o.c allocate huge amounts of memory
with fuzzed binaries. This make it a little harder.
The patch also fixes a number of places where an attempt to detect
overflow of multiplication was done with code like
if (x * 4 < x)
/* overflow case */
That of course doesn't work. There are plenty of values of x that
overflow x * 4 but (x * 4) mod 2^n is greater than x. For example
with 16-bit types, 0x6000 * 4 = 0x18000 mod 2^16 = 0x8000.
* mach-o.c (bfd_mach_o_canonicalize_relocs): Fix ineffective
overflow check.
(bfd_mach_o_canonicalize_reloc): Likewise.
(bfd_mach_o_canonicalize_dynamic_reloc): Likewise. Sanity check
counts and offsets against file size.
(bfd_mach_o_build_dysymtab): Fix ineffective overflow check.
(bfd_mach_o_mangle_sections): Remove unnecessary overflow check.
(bfd_mach_o_read_symtab_symbols): Sanity check count and offset
against file size. Delete symbol table error message.
(bfd_mach_o_read_dysymtab): Sanity check counts and offsets
against file size.
(bfd_mach_o_read_symtab): Likewise.
(bfd_mach_o_read_command): Pass file size.
(bfd_mach_o_scan): Sanity check command count against file size.
For running the DejaGnu tests, some esoteric configurations
may require a testglue. This, for instance, is true about
testing ARC targets which uses its own DejaGnu board and
a simulator which does not support returning the program's exit
code. Therefore, for those tests that use "gdb_compile", a
"gdb_tg.o" file is compiled and linked into the final
executable.
There are tests that invoke "gdb_compile" from different
directories. Let's take a look at an example test:
gdb.base/fullname.exp. The purpose of this test is to build
the executable from different directories (absolute vs. relative
vs. other) and then check if gdb can handle setting breakpoints
accordingly.
When "gdb_compile" generates the "gdb_tg.o", it does not
do it again for the same test. Although this might seem
efficient, it can lead to problems when changing directories
before the next compile:
gdb compile failed, arc-elf32-gcc: error: gdb_tg.o:
No such file or directory
This patch checks if the wrapper file ("gdb_tg.o") is still in
reach and if it is not, it will stimulate the regeneration of
the wrapper.
It is worth mentioning that GCC's DejaGnu tests handle these
scenarios as well and they seem to be more efficient in doing so
by saving the library paths and manipulating them if necessary
[1]. However, for GDB tests, that require less compilations,
I think the proposed solution should be fine compared to a more
full fledged solution from GCC. The glue file in our case is
only 2 KiB.
Last but not least, I ran the x86_64 tests on an x86_64 host and
found no regression.
[1]
Avid coders may look for "set_ld_library_path_env_vars" in
gcc/testsuite/lib/target-libpath.exp.
gdb/testsuite/ChangeLog:
* lib/gdb.exp (gdb_wrapper_init): Reset
"gdb_wrapper_initialized" to 0 if "wrapper_file" does
not exist.
dwarf2_find_containing_comp_unit has this in its binary search:
if (mid_cu->is_dwz > offset_in_dwz
|| (mid_cu->is_dwz == offset_in_dwz
&& mid_cu->sect_off + mid_cu->length >= sect_off))
high = mid;
The intent here is to determine whether SECT_OFF appears in or before
MID_CU.
I believe this has an off-by-one error, and that the check should use
">" rather than ">=". If the two side are equal, then SECT_OFF
actually appears at the start of the next CU.
I've had this patch kicking around for ages but I forget how I found
the problem.
gdb/ChangeLog
2020-02-20 Tom Tromey <tom@tromey.com>
* dwarf2/read.c (dwarf2_find_containing_comp_unit): Use ">", not
">=", in binary search.
(dwarf2_find_containing_comp_unit): New overload.
(run_test): New self-test.
(_initialize_dwarf2_read): Register new test.
CSRRW and CSRRWI always write CSR. CSRRS, CSRRC, CSRRSI and CSRRCI write CSR
when RS1 isn't zero. The CSR is read only if the [11:10] bits of CSR address
is 0x3. The read-only CSR can not be written by the CSR instructions.
gas/
* config/tc-riscv.c (riscv_ip): New boolean insn_with_csr to indicate
we are assembling instruction with CSR. Call riscv_csr_read_only_check
after parsing all arguments.
(enum csr_insn_type): New enum is used to classify the CSR instruction.
(riscv_csr_insn_type, riscv_csr_read_only_check): New functions. These
are used to check if we write a read-only CSR by the CSR instruction.
* testsuite/gas/riscv/priv-reg-fail-read-only-01.s: New testcase. Test
all CSR for the read-only CSR checking.
* testsuite/gas/riscv/priv-reg-fail-read-only-01.d: Likewise.
* testsuite/gas/riscv/priv-reg-fail-read-only-01.l: Likewise.
* testsuite/gas/riscv/priv-reg-fail-read-only-02.s: New testcase. Test
all CSR instructions for the read-only CSR checking.
* testsuite/gas/riscv/priv-reg-fail-read-only-02.d: Likewise.
* testsuite/gas/riscv/priv-reg-fail-read-only-02.l: Likewise.
Add new .option `csr-check/no-csr-check` and GAS option `-mcsr-check
/-mno-csr-check` to enbale/disable the CSR checking. Disable the CSR
checking by default.
gas/
* config/tc-riscv.c: Add new .option and GAS options to enbale/disable
the CSR checking. We disable the CSR checking by default.
(reg_lookup_internal): Check the `riscv_opts.csr_check`
before we doing the CSR checking.
* doc/c-riscv.texi: Add description for the new .option and assembler
options.
* testsuite/gas/riscv/priv-reg-fail-fext.d: Add `-mcsr-check` to enable
the CSR checking.
* testsuite/gas/riscv/priv-reg-fail-rv32-only.d: Likewise.
According to the riscv privilege spec, some CSR are only valid when rv32 or
the specific extension is set. We extend the DECLARE_CSR and DECLARE_CSR_ALIAS
to record more informaton we need, and then check whether the CSR is valid
according to these information. We report warning message when the CSR is
invalid, so we have a choice between error and warning by --fatal-warnings
option. Also, a --no-warn/-W option is used to turn the warnings off, if
people don't want the warnings.
gas/
* config/tc-riscv.c (enum riscv_csr_class): New enum. Used to decide
whether or not this CSR is legal in the current ISA string.
(struct riscv_csr_extra): New structure to hold all extra information
of CSR.
(riscv_init_csr_hash): New function. According to the DECLARE_CSR and
DECLARE_CSR_ALIAS, insert CSR extra information into csr_extra_hash.
Call hash_reg_name to insert CSR address into reg_names_hash.
(md_begin): Call riscv_init_csr_hashes for each DECLARE_CSR.
(reg_csr_lookup_internal, riscv_csr_class_check): New functions.
Decide whether the CSR is valid according to the csr_extra_hash.
(init_opcode_hash): Update 'if (hash_error != NULL)' as hash_error is
not a boolean. This is same as riscv_init_csr_hash, so keep the
consistent usage.
* testsuite/gas/riscv/csr-dw-regnums.d: Add -march=rv32if option.
* testsuite/gas/riscv/priv-reg.d: Add f-ext by -march option.
* testsuite/gas/riscv/priv-reg-fail-fext.d: New testcase. The source
file is `priv-reg.s`, and the ISA is rv32i without f-ext, so the
f-ext CSR are not allowed.
* testsuite/gas/riscv/priv-reg-fail-fext.l: Likewise.
* testsuite/gas/riscv/priv-reg-fail-rv32-only.d: New testcase. The
source file is `priv-reg.s`, and the ISA is rv64if, so the
rv32-only CSR are not allowed.
* testsuite/gas/riscv/priv-reg-fail-rv32-only.l: Likewise.
include/
* opcode/riscv-opc.h: Extend DECLARE_CSR and DECLARE_CSR_ALIAS to
record riscv_csr_class.
opcodes/
* riscv-dis.c (print_insn_args): Updated since the DECLARE_CSR is changed.
gdb/
* riscv-tdep.c: Updated since the DECLARE_CSR is changed.
* riscv-tdep.h: Likewise.
* features/riscv/rebuild-csr-xml.sh: Generate the 64bit-csr.xml without
rv32-only CSR.
* features/riscv/64bit-csr.xml: Regernated.
binutils/
* dwarf.c: Updated since the DECLARE_CSR is changed.
This allows pdp11 to handle 32-bit fixups that can be applied by gas,
and improves the error message emitted when the required reloc isn't
available.
* config/tc-pdp11.c (md_apply_fix): Handle BFD_RELOC_32.
(tc_gen_reloc): Only give a BAD_CASE assertion on pcrel relocs.
This patch fixes an ancient wart in aout support, in that text and
data section sizes are rounded up for alignment rather that just the
corresponding header sizes. Changing section sizes could conceivably
result in buffer overflows if section contents were held in memory.
Also, keeping the original section sizes allows this PR to be fixed
nicely.
bfd/
PR 25569
* aoutx.h (adjust_o_magic, adjust_z_magic, adjust_n_magic): Use
"text", "data" and "bss" section pointer vars. Don't update
section size, just exec header sizes.
(adjust_sizes_and_vmas): Don't update text section size. Set
initial exec header a_text. Print exec headers sizes.
* pdp11.c (adjust_o_magic, adjust_z_magic, adjust_n_magic),
(adjust_sizes_and_vmas): Similarly. Formatting.
(final_link): Correct final file extension.
gas/
PR 25569
* config/obj-aout.c (obj_aout_frob_file_before_fix): Don't loop
on section size adjustment, instead perform another write if
exec header size is larger than section size.
There is currently a regression when using
'{putchar,fputc}_unfiltered' with 'puts_unfiltered' which was
introduced by one of the commits that reworked the unfiltered print
code.
The regression makes it impossible to use '{putchar,fputc}_unfiltered'
with 'puts_unfiltered', because the former writes directly to the
ui_file stream using 'stream->write', while the latter uses a buffered
mechanism (see 'wrap_buffer') and delays the printing.
If you do a quick & dirty hack on e.g. top.c:show_gdb_datadir:
@@ -2088,6 +2088,13 @@ static void
show_gdb_datadir (struct ui_file *file, int from_tty,
struct cmd_list_element *c, const char *value)
{
+ putchar_unfiltered ('\n');
+ puts_unfiltered ("TEST");
+ putchar_unfiltered ('>');
+ puts_unfiltered ("PUTS");
+ putchar_unfiltered ('\n');
rebuild GDB and invoke the "show data-directory" command, you will
see:
(gdb) show data-directory
>
TESTPUTSGDB's data directory is "/usr/local/share/gdb".
Note how the '>' was printed before the output, and "TEST" and "PUTS"
were printed together.
My first attempt to fix this was to always call 'flush_wrap_buffer' at
the end of 'fputs_maybe_filtered', since it seemed to me that the
function should always print what was requested. But I wasn't sure
this was the right thing to do, so I talked to Tom on IRC and he gave
me another, simpler idea: make '{putchar,fputc}_unfiltered' call into
the already existing 'fputs_unfiltered' function.
This patch implements the idea. I regtested it on the Buildbot, and
no regressions were detected.
gdb/ChangeLog:
2020-02-20 Sergio Durigan Junior <sergiodj@redhat.com>
Tom Tromey <tom@tromey.com>
* utils.c (fputs_maybe_filtered): Call 'stream->puts' instead
of 'fputc_unfiltered'.
(putchar_unfiltered): Call 'fputc_unfiltered'.
(fputc_unfiltered): Call 'fputs_unfiltered'.
With gdb.go/hello.go, we run into an xpass:
...
Thread 1 "hello" hit Breakpoint 1, main.main () at hello.go:7^M
7 func main () {^M
(gdb) print st^M
$1 = 0x0 ""^M
(gdb) XPASS: gdb.go/hello.exp: starting string check
...
The xfail is setup as follows:
...
\# This used to print "", i.e., the local "st" initialized as "".
setup_xfail "*-*-*"
gdb_test "print st" \
".* = $hex \"\"" \
"starting string check"
...
It's not clear what gccgo/gc PR this xfail refers to.
It's also not clear why the empty string is both:
- listed as reason for xfail, and
- used in the pass pattern.
Furthermore, there's a comment in the hello.go testcase:
...
st := "Hello, world!" // this intentionally shadows the global "st"
...
while there's no global st variable present, only a variable myst:
...
var myst = "Shall we?"
...
Fix this by splitting up the test-case in two test-cases, hello.{go,exp} and
global-local-var-shadow.{go,exp}.
In hello.exp we no longer attempt to print st before its declaration. In
hello.go we remove the myst variable as well the comment related to shadowing.
In global-local-var-shadow.go, we rename myst to st, such that the comment
related to shadowing is correct. In global-local-var-shadow.exp we attempt to
print the value of st before the local definition, which should print the
value of the global definition, and xfail this with reference to GCC PR93844.
Tested on x86_64-linux, with gccgo 10.
gdb/testsuite/ChangeLog:
2020-02-20 Tom de Vries <tdevries@suse.de>
PR go/17018
* gdb.go/hello.exp: Copy ...
* gdb.go/global-local-var-shadow.exp: ... here. New file. Expect
print of st to print value of global definition. Add xfail for GCC
PR93844.
* gdb.go/hello.exp: Remove printing of st before definition.
* gdb.go/hello.go: Copy ...
* gdb.go/global-local-var-shadow.go: ... here. New test. Rename myst
to st.
* gdb.go/hello.go: Remove myst. Remove comment about shadowing.
The 'set_target_ops' function takes a target op vector and creates a
clone of it via XNEW and memcpy. This is not necessary. 'the_target'
is a singleton, and the argument that is passed to 'set_target_ops' is
always the address of a global, static object. Therefore, update the
implementation to simply copy the pointer.
gdbserver/ChangeLog:
2020-02-20 Pedro Alves <palves@redhat.com>
* target.cc (set_target_ops): Simply copy the given target pointer
instead of creating a copy of the pointed object.
Tom Tromey
Tom de Vries
Tom Tromey
Andrew Burgess
Ali Tamur via gdb-patches
Alan Modra
Shahab Vahedi
Nelson Chu
GDB Administrator
Sergio Durigan Junior
Tankut Baris Aktemur
Pedro Alves