Having been irritated by seeing bfd/elf{32,64}-aarch64.c to be re-
generated in x86-only builds, I came across 769a27ade5 ("Re: bfd
BLD-POTFILES.in dependencies"). I think this went slightly too far, as
outside of maintainer mode dependencies will cause the subset of files
to be (re-)generated which are actually needed for the build.
Generating them all is only needed when wanting to update certain files
under bfd/po/, i.e. in maintainer mode.
In the course of looking around in an attempt to try to understand how
things are meant to work, I further noticed that ld has got things
slightly wrong too: BLD-POTFILES.in depending on $(BLD_POTFILES) isn't
quite right (the output doesn't change when any of the enumerated files
changes; it's the mere presence which matters); like in bfd it looks
like we would better extend BUILT_SOURCES accordingly.
Furthermore it became apparent that ld fails to enumerate the .c files
generated from the .l and .y ones. While in their absence it was benign
whether translatable strings in the source files were actually marked as
such, this now becomes relevant. Mark respective strings at the same
time, but skipping ones which look to be of interest for debugging
purposes only (e.g. such used by printf() enclosed in #ifdef TRACE).
Trying to free malloc'd memory used by the stabs and coff debug info
parsers is complicated, and traversing the trees generated requires a
lot of code. It's better to bfd_alloc the memory which allows it all
to be freed without fuss when the bfd is closed. In the process of
doing this I reverted most of commit a633691333.
Some of the stabs handling code grows arrays of pointers with realloc,
to deal with arbitrary numbers of fields, function args, etc. The
code still does that but copies over to bfd_alloc memory when
finished. The alternative is to parse twice, once to size, then again
to populate the arrays. I think that complication is unwarranted.
Note that there is a greater than zero chance this patch breaks
something, eg. that I missed an attempt to free obj_alloc memory.
Also it seems there are no tests in the binutils testsuite aimed at
exercising objdump --debugging.
* budbg.h (finish_stab, parse_stab): Update prototypes
* debug.c: Include bucomm.h.
(struct debug_handle): Add "abfd" field.
(debug_init): Add "abfd" param. bfd_alloc handle.
(debug_xalloc, debug_xzalloc): New functions. Use throughout
in place of xmalloc and memset.
(debug_start_source): Remove "name_used" param.
* debug.h (debug_init, debug_start_source): Update prototypes.
(debug_xalloc, debug_xzalloc): Declare.
* objcopy.c (copy_object): Don't free dhandle.
* objdump.c (dump_bfd): Likewise.
* rdcoff.c (coff_get_slot): Add dhandle arg. debug_xzalloc
memory in place of xcalloc. Update callers.
(parse_coff_struct_type): Don't leak on error return. Copy
fields over to debug_xalloc memory.
(parse_coff_enum_type): Copy names and vals over the
debug_xalloc memory.
* rddbg.c (read_debugging_info): Adjust debug_init call.
Don't free dhandle.
(read_section_stabs_debugging_info): Don't free shandle.
Adjust parse_stab call. Call finish_stab on error return.
(read_symbol_stabs_debugging_info): Similarly.
* stabs.c (savestring): Delete unnecessary forward declaration.
Add dhandle param. debug_xalloc memory. Update callers.
(start_stab): Delete unnecessary casts.
(finish_stab): Add "emit" param. Free file_types, so_string,
and stabs handle.
(parse_stab): Delete string_used param. Revert code dealing
with string_used. Copy so_string passed to debug_set_filename
and stored as main_filename to debug_xalloc memory. Similarly
for string passed to debug_start_source and push_bincl. Copy
args to debug_xalloc memory. Don't leak args.
(parse_stab_enum_type): Copy names and values to debug_xalloc
memory. Don't free name.
(parse_stab_struct_type): Don't free fields.
(parse_stab_baseclasses): Delete unnecessary cast.
(parse_stab_struct_fields): Return debug_xalloc fields.
(parse_stab_cpp_abbrev): Use debug_xalloc for _vb$ type name.
(parse_stab_one_struct_field): Don't free name.
(parse_stab_members): Copy variants and methods to
debug_xalloc memory. Don't free name or argtypes.
(parse_stab_argtypes): Use debug_xalloc memory for physname
and args.
(push_bincl): Add dhandle param. Use debug_xalloc memory.
(stab_record_variable): Use debug_xalloc memory.
(stab_emit_pending_vars): Don't free var list.
(stab_find_slot): Add dhandle param. Use debug_xzalloc
memory. Update all callers.
(stab_find_tagged_type): Don't free name. Use debug_xzalloc.
(stab_demangle_qualified): Don't free name.
(stab_demangle_template): Don't free s1.
(stab_demangle_args): Tidy pvarargs refs. Copy *pargs on
success to debug_xalloc memory, free on failure.
(stab_demangle_fund_type): Don't free name.
(stab_demangle_v3_arglist): Copy args to debug_xalloc memory.
Don't free dt.
This adds the DAP readMemory and writeMemory requests. A small change
to the evaluation code is needed in order to test this -- this is one
of the few ways for a client to actually acquire a memory reference.
I noticed a couple of places in infrun.c where we call
set_momentary_breakpoint_at_pc, and then set the newly created
breakpoint's thread field, these are in:
insert_exception_resume_breakpoint
insert_exception_resume_from_probe
Function set_momentary_breakpoint_at_pc calls
set_momentary_breakpoint, which always creates the breakpoint as
thread-specific for the current inferior_thread().
The two insert_* functions mentioned above take an arbitrary
thread_info* as an argument and set the breakpoint::thread to hold the
thread number of that arbitrary thread.
However, the insert_* functions store the breakpoint pointer within
the current inferior_thread(), so we know that the thread being passed
in must be the currently selected thread.
What this means is that we can:
1. Assert that the thread being passed in is the currently selected
thread, and
2. No longer adjust the breakpoint::thread field, this will already
have been set correctly be calling set_momentary_breakpoint_at_pc.
There should be no user visible changes after this commit.
Consider the following session:
(gdb) list some_func
1 int
2 some_func ()
3 {
4 int *p = 0;
5 return *p;
6 }
7
8 void
9 foo ()
10 {
(gdb) break foo if (some_func ())
Breakpoint 1 at 0x40111e: file bpcond.c, line 11.
(gdb) r
Starting program: /tmp/bpcond
Program received signal SIGSEGV, Segmentation fault.
0x0000000000401116 in some_func () at bpcond.c:5
5 return *p;
Error in testing condition for breakpoint 1:
The program being debugged stopped while in a function called from GDB.
Evaluation of the expression containing the function
(some_func) will be abandoned.
When the function is done executing, GDB will silently stop.
Breakpoint 1, 0x0000000000401116 in some_func () at bpcond.c:5
5 return *p;
(gdb)
What happens here is the breakpoint condition includes a call to an
inferior function, and the inferior function segfaults. We can see
that GDB reports the segfault, and then gives an error message that
indicates that an inferior function call was interrupted.
After this GDB appears to report that it is stopped at Breakpoint 1,
inside some_func.
I find this second stop report a little confusing. While it is true
that GDB stopped as a result of hitting breakpoint 1, I think the
message GDB currently prints might give the impression that GDB is
actually stopped at a location of breakpoint 1, which is not the case.
Also, I find the second stop message draws attention away from
the "Program received signal SIGSEGV, Segmentation fault" stop
message, and this second stop might be thought of as replacing in
someway the earlier message.
In short, I think things would be clearer if the second stop message
were not reported at all, so the output should, I think, look like
this:
(gdb) list some_func
1 int
2 some_func ()
3 {
4 int *p = 0;
5 return *p;
6 }
7
8 void
9 foo ()
10 {
(gdb) break foo if (some_func ())
Breakpoint 1 at 0x40111e: file bpcond.c, line 11.
(gdb) r
Starting program: /tmp/bpcond
Program received signal SIGSEGV, Segmentation fault.
0x0000000000401116 in some_func () at bpcond.c:5
5 return *p;
Error in testing condition for breakpoint 1:
The program being debugged stopped while in a function called from GDB.
Evaluation of the expression containing the function
(some_func) will be abandoned.
When the function is done executing, GDB will silently stop.
(gdb)
The user can still find the number of the breakpoint that triggered
the initial stop in this line:
Error in testing condition for breakpoint 1:
But there's now only one stop reason reported, the SIGSEGV, which I
think is much clearer.
To achieve this change I set the bpstat::print field when:
(a) a breakpoint condition evaluation failed, and
(b) the $pc of the thread changed during condition evaluation.
I've updated the existing tests that checked the error message printed
when a breakpoint condition evaluation failed.
Consider the following case:
(gdb) list some_func
1 int
2 some_func ()
3 {
4 int *p = 0;
5 return *p;
6 }
7
8 void
9 foo ()
10 {
(gdb) break foo if (some_func ())
Breakpoint 1 at 0x40111e: file bpcond.c, line 11.
(gdb) r
Starting program: /tmp/bpcond
Program received signal SIGSEGV, Segmentation fault.
0x0000000000401116 in some_func () at bpcond.c:5
5 return *p;
Error in testing breakpoint condition:
The program being debugged was signaled while in a function called from GDB.
GDB remains in the frame where the signal was received.
To change this behavior use "set unwindonsignal on".
Evaluation of the expression containing the function
(some_func) will be abandoned.
When the function is done executing, GDB will silently stop.
Program received signal SIGSEGV, Segmentation fault.
Breakpoint 1, 0x0000000000401116 in some_func () at bpcond.c:5
5 return *p;
(gdb)
Notice that this line:
Program received signal SIGSEGV, Segmentation fault.
Appears twice in the output. The first time is followed by the
current location. The second time is a little odd, why do we print
that?
Printing that line is controlled, in part, by a global variable,
stopped_by_random_signal. This variable is reset to zero in
handle_signal_stop, and is set if/when GDB figures out that the
inferior stopped due to some random signal.
The problem is, in our case, GDB first stops at the breakpoint for
foo, and enters handle_signal_stop and the stopped_by_random_signal
global is reset to 0.
Later within handle_signal_stop GDB calls bpstat_stop_status, it is
within this function (via bpstat_check_breakpoint_conditions) that the
breakpoint condition is checked, and, we end up calling the inferior
function (some_func in our example above).
In our case above the thread performing the inferior function call
segfaults in some_func. GDB catches the SIGSEGV and handles the stop,
this causes us to reenter handle_signal_stop. The global variable
stopped_by_random_signal is updated, this time it is set to true
because the thread stopped due to SIGSEGV. As a result of this we
print the first instance of the line (as seen above in the example).
Finally we unwind GDB's call stack, the inferior function call is
complete, and we return to the original handle_signal_stop. However,
the stopped_by_random_signal global is still carrying the value as
computed for the inferior function call's stop, which is why we now
print a second instance of the line, as seen in the example.
To prevent this, I propose adding a scoped_restore before we start an
inferior function call. This will save and restore the global
stopped_by_random_signal value.
With this done, the output from our example is now this:
(gdb) list some_func
1 int
2 some_func ()
3 {
4 int *p = 0;
5 return *p;
6 }
7
8 void
9 foo ()
10 {
(gdb) break foo if (some_func ())
Breakpoint 1 at 0x40111e: file bpcond.c, line 11.
(gdb) r
Starting program: /tmp/bpcond
Program received signal SIGSEGV, Segmentation fault.
0x0000000000401116 in some_func () at bpcond.c:5
5 return *p;
Error in testing condition for breakpoint 1:
The program being debugged stopped while in a function called from GDB.
Evaluation of the expression containing the function
(some_func) will be abandoned.
When the function is done executing, GDB will silently stop.
Breakpoint 1, 0x0000000000401116 in some_func () at bpcond.c:5
5 return *p;
(gdb)
We now only see the 'Program received signal SIGSEGV, ...' line once,
which I think makes more sense.
Finally, I'm aware that the last few lines, that report the stop as
being at 'Breakpoint 1', when this is not where the thread is actually
located anymore, is not great. I'll address that in the next commit.
This commit extends gdbserver to take account of a failed memory
access from agent_mem_read, and to return a new eval_result_type
expr_eval_invalid_memory_access.
I have only updated the agent_mem_read calls related directly to
reading memory, I have not updated any of the calls related to
tracepoint data collection. This is just because I'm not familiar
with that area of gdb/gdbserver, and I don't want to break anything,
so leaving the existing behaviour untouched seems like the safest
approach.
I've then updated gdb.base/bp-cond-failure.exp to test evaluating the
breakpoints on the target, and have also extended the test so that it
checks for different sizes of memory access.
Currently the gdbserver function agent_mem_read ignores any errors
from calling read_inferior_memory. This means that if there is an
attempt to access invalid memory then this will appear to succeed.
In this patch I update agent_mem_read so that if read_inferior_memory
fails, agent_mem_read will return an error code.
However, none of the callers of agent_mem_read actually check the
return value, so this commit will have no effect on anything. In the
next commit I will update the users of agent_mem_read to check for the
error code.
I've also updated the header comments on agent_mem_read to better
reflect what the function does, and its possible return values.
When GDB fails to test the condition of a conditional breakpoint, for
whatever reason, the error message looks like this:
(gdb) break foo if (*(int *) 0) == 1
Breakpoint 1 at 0x40111e: file bpcond.c, line 11.
(gdb) r
Starting program: /tmp/bpcond
Error in testing breakpoint condition:
Cannot access memory at address 0x0
Breakpoint 1, foo () at bpcond.c:11
11 int a = 32;
(gdb)
The line I'm interested in for this commit is this one:
Error in testing breakpoint condition:
In the case above we can figure out that the problematic breakpoint
was #1 because in the final line of the message GDB reports the stop
at breakpoint #1.
However, in the next few patches I plan to change this. In some cases
I don't think it makes sense for GDB to report the stop as being at
breakpoint #1, consider this case:
(gdb) list some_func
1 int
2 some_func ()
3 {
4 int *p = 0;
5 return *p;
6 }
7
8 void
9 foo ()
10 {
(gdb) break foo if (some_func ())
Breakpoint 1 at 0x40111e: file bpcond.c, line 11.
(gdb) r
Starting program: /tmp/bpcond
Program received signal SIGSEGV, Segmentation fault.
0x0000000000401116 in some_func () at bpcond.c:5
5 return *p;
Error in testing breakpoint condition:
The program being debugged was signaled while in a function called from GDB.
GDB remains in the frame where the signal was received.
To change this behavior use "set unwindonsignal on".
Evaluation of the expression containing the function
(some_func) will be abandoned.
When the function is done executing, GDB will silently stop.
Program received signal SIGSEGV, Segmentation fault.
Breakpoint 1, 0x0000000000401116 in some_func () at bpcond.c:5
5 return *p;
(gdb)
Notice that, the final lines of output reports the stop as being at
breakpoint #1, even though the inferior in not located within
some_func, and it's certainly not located at the breakpoint location.
I find this behaviour confusing, and propose that this should be
changed. However, if I make that change then every reference to
breakpoint #1 will be lost from the error message.
So, in this commit, in preparation for the later commits, I propose to
change the 'Error in testing breakpoint condition:' line to this:
Error in testing condition for breakpoint NUMBER:
where NUMBER will be filled in as appropriate. Here's the first
example with the updated error:
(gdb) break foo if (*(int *) 0) == 0
Breakpoint 1 at 0x40111e: file bpcond.c, line 11.
(gdb) r
Starting program: /tmp/bpcond
Error in testing condition for breakpoint 1:
Cannot access memory at address 0x0
Breakpoint 1, foo () at bpcond.c:11
11 int a = 32;
(gdb)
The breakpoint number does now appear twice in the output, but I don't
see that as a negative.
This commit just changes the one line of the error, and updates the
few tests that either included the old error in comments, or actually
checked for the error in the expected output.
As the only test that checked the line I modified is a Python test,
I've added a new test that doesn't rely on Python that checks the
error message in detail.
While working on the new test, I spotted that it would fail when run
with native-gdbserver and native-extended-gdbserver target boards.
This turns out to be due to a gdbserver bug. To avoid cluttering this
commit I've added a work around to the new test script so that the
test passes for the remote boards, in the next few commits I will fix
gdbserver, and update the test script to remove the work around.
This commit builds on the previous one to fix all the remaining
failures in gdb.base/unwind-on-each-insn.exp for RISC-V.
The problem we have in gdb.base/unwind-on-each-insn.exp is that, when
we are in the function epilogue, the previous frame and stack pointer
values are being restored, and so, the values that we calculated
during the function prologue are no longer suitable.
Here's an example from the function 'bar' in the mentioned test. This
was compiled for 64-bit RISC-V with compressed instruction support:
Dump of assembler code for function bar:
0x000000000001018a <+0>: add sp,sp,-32
0x000000000001018c <+2>: sd ra,24(sp)
0x000000000001018e <+4>: sd fp,16(sp)
0x0000000000010190 <+6>: add fp,sp,32
0x0000000000010192 <+8>: sd a0,-24(fp)
0x0000000000010196 <+12>: ld a0,-24(fp)
0x000000000001019a <+16>: jal 0x10178 <foo>
0x000000000001019e <+20>: nop
0x00000000000101a0 <+22>: ld ra,24(sp)
0x00000000000101a2 <+24>: ld fp,16(sp)
0x00000000000101a4 <+26>: add sp,sp,32
0x00000000000101a6 <+28>: ret
End of assembler dump.
When we are at address 0x101a4 the previous instruction has restored
the frame-pointer, as such GDB's (current) preference for using the
frame-pointer as the frame base address is clearly not going to work.
We need to switch to using the stack-pointer instead.
At address 0x101a6 the previous instruction has restored the
stack-pointer value. Currently GDB will not understand this and so
will still assume the stack has been decreased by 32 bytes in this
function.
My proposed solution is to extend GDB such that GDB will scan the
instructions at the current $pc looking for this pattern:
ld fp,16(sp)
add sp,sp,32
ret
Obviously the immediates can change, but the basic pattern indicates
that the function is in the process of restoring state before
returning. If GDB sees this pattern then GDB can use the inferior's
position within this instruction sequence to help calculate the
correct frame-id.
With this implemented then gdb.base/unwind-on-each-insn.exp now fully
passes.
Obviously what I've implemented is just a heuristic. It's not going
to work for every function. If the compiler reorders the
instructions, or merges the epilogue back into the function body then
GDB is once again going to get the frame-id wrong.
I'm OK with that, we're no worse off that we are right now in that
situation (plus we can always improve the heuristic later).
Remember, this is for debugging code without debug information,
and (in our imagined situation) with more aggressive levels of
optimisation being used. Obviously GDB is going to struggle in these
situations.
My thinking is, lets get something in place now. Then, later, if
possible, we might be able to improve the logic to cover more
situations -- if there's an interest in doing so. But I figure we
need something in place as a starting point.
After this commit gdb.base/unwind-on-each-insn.exp passes with no
failures on RV64.
Add support to the RISC-V prologue scanner for c.ldsp and c.lwsp
instructions.
This fixes some of the failures in gdb.base/unwind-on-each-insn.exp,
though there are further failures that are not fixed by this commit.
This change started as a wider fix that would address all the failures
in gdb.base/unwind-on-each-insn.exp, however, that wider fix needed
support for the two additional compressed instructions.
When I added support for those two compressed instructions I noticed
that some of the failures in gdb.base/unwind-on-each-insn.exp resolved
themselves!
Here's what's going on:
The reason for the failures is that GDB is trying to build the
frame-id during the last few instructions of the function. These are
the instructions that restore the frame and stack pointers just prior
to the return instruction itself.
By the time we reach the function epilogue the stack offset that we
calculated during the prologue scan is no longer valid, and so we
calculate the wrong frame-id.
However, in the particular case of interest here, the test function
'foo', the function is so simple and short (the empty function) that
GDB's prologue scan could, in theory, scan every instruction of the
function.
I say "could, in theory," because currently GDB stops the prologue
scan early when it hits an unknown instruction. The unknown
instruction happens to be one of the compressed instructions that I'm
adding support for in this commit.
Now that GDB understands the compressed instructions the prologue scan
really does go from the start of the function right up to the current
program counter. As such, GDB sees that the stack frame has been
allocated, and then deallocated, and so builds the correct frame-id.
Of course, most real functions are not as simple as the test function
'foo'. As such, we can't usually rely on scanning right up to the end
of the function -- there are some instructions we always need to stop
at because GDB can't reason about how they change the inferior
state (e.g. a function call). The test function 'bar' is just such an
example.
After this commit, we can now build the frame-id correctly for every
instruction in 'foo', but there are some tests still failing in 'bar'.
Convert the RISC-V specific debug settings to use the new debug
printing scheme. This updates the following settings:
set/show debug riscv breakpoints
set/show debug riscv gdbarch
set/show debug riscv infcall
set/show debug riscv unwinder
All of these settings now take a boolean rather than an integer, and
all print their output using the new debug scheme.
There should be no visible change for anyone not turning on debug.
While I was working on the disassembler styling for ARM I noticed that
the whitespace in the cpsie instruction was inconsistent with most of
the other ARM disassembly output, the disassembly for cpsie looks like
this:
cpsie if,#10
notice there's no space before the '#10' immediate, most other ARM
instructions have a space before each operand.
This commit updates the disassembler to add the missing space, and
updates the tests I found that tested this instruction.
This commit follows on from the following two commits:
commit 80dc83fd0e
Date: Fri Jun 11 11:30:47 2021 +0100
gdb/remote: handle target dying just before a stepi
And:
commit 079f190d4c
Date: Thu Mar 9 10:45:03 2023 +0100
[gdb/testsuite] Fix gdb.server/server-kill.exp for remote target
The first of these commits fixed an issue in GDB and tried to extend
the gdb.server/server-kill.exp test to cover the GDB fix.
Unfortunately, the changes to gdb.server/server-kill.exp were not
correct, and were causing problems when trying to run with the
remote-gdbserver-on-localhost board file.
The second commit reverts some of the gdb.server/server-kill.exp
changes introduced in the first commit so that the test will now work
correctly with the remote-gdbserver-on-localhost board file.
The second commit is just about GDB's testing infrastructure -- it's
not about the original fix to GDB from the first commit, the actual
GDB change was fine.
While reviewing the second commit I wanted to check that the problem
fixed in the first commit is still being tested by the
gdb.server/server-kill.exp script, so I reverted the change to
breakpoint.c that is the core of the first commit and ran the test
script ..... and saw no failures.
The first commit is about GDB discovering that gdbserver has died
while trying to insert a breakpoint. As soon as GDB spots that
gdbserver is gone we mourn the remote inferior, which ends up deleting
all the breakpoints associated with the remote inferiors. We then
throw an exception which is caught in the insert breakpoints code, and
we try to display an error that includes the breakpoint number
.... but the breakpoint has already been deleted ... and so GDB
crashes.
After digging a little, what I found is that today, when the test does
'stepi' the first thing we end up doing is calculating the frame-id as
part of the stepi logic, it is during this frame-id calculation that
we mourn the remote inferior, delete the breakpoints, and throw an
exception. The exception is caught by the top level interpreter loop,
and so we never try to print the breakpoint number which is what
caused the original crash.
If I add an 'info frame' command to the test script, prior to killing
gdbserver, then now when we 'stepi' GDB already has the frame-id
calculated, and the first thing we do is try to insert the
breakpoints, this will trigger the original bug.
In order to reproduce this experiment you'll need to change a function
in breakpoint.c, like this:
static void
rethrow_on_target_close_error (const gdb_exception &e)
{
return;
}
Then run gdb.server/server-kill.exp with and without this patch. You
should find that without this patch there are zero test failures,
while with this patch there will be one failure like this:
(gdb) PASS: gdb.server/server-kill.exp: test_stepi: info frame
Executing on target: kill -9 4513 (timeout = 300)
builtin_spawn -ignore SIGHUP kill -9 4513
stepi
../../src/gdb/breakpoint.c:2863: internal-error: insert_bp_location: Assertion `bl->owner != nullptr' failed.
A problem internal to GDB has been detected,
further debugging may prove unreliable.
----- Backtrace -----
...
A potential test failure was introduced with commit:
commit 6bf5f25bb1
Date: Wed Mar 8 16:11:30 2023 +0000
gdb/python: make the gdb.unwinder.Unwinder class more robust
In this commit a new test was added, however the expected output
pattern varies depending on which Python version GDB is linked
against.
Older versions of Python result in output like this:
(gdb) python global_test_unwinder.name = "foo"
Traceback (most recent call last):
File "<string>", line 1, in <module>
AttributeError: can't set attribute
Error while executing Python code.
(gdb)
While more recent versions of Python give a similar, but slightly more
verbose error message, like this:
(gdb) python global_test_unwinder.name = "foo"
Traceback (most recent call last):
File "<string>", line 1, in <module>
AttributeError: can't set attribute 'name'
Error while executing Python code.
(gdb)
The test was only accepting the first version of the output. This
commit extends the test pattern so that either version will be
accepted.
The detection logic for TPIDR2 was implemented incorrectly. Originally
the detection was supposed to be through a ptrace error code, but in reality,
for backwards compatibility, the detection should be based on the size of
the returned iovec.
For instance, if a target supports both TPIDR and TPIDR2, ptrace will return a
iovec size of 16. If a target only supports TPIDR and not TPIDR2, it will
return a iovec size of 8, even if we asked for 16 bytes.
This patch fixes this issue in code that is shared between gdb and gdbserver,
therefore both gdb and gdbserver are fixed.
Tested on AArch64/Linux Ubuntu 20.04.
A case of a string section ending with an unterminated string. Fix it
by allocating one more byte and making it zero. Also make functions
reading the data return void* so that casts are not needed.
* ecoff.c (READ): Delete type param. Allocate one extra byte
to terminate string sections with a NUL. Adjust invocation.
* elfxx-mips.c (READ): Likewise.
* libbfd-in.h (_bfd_alloc_and_read): Return a void*.
(_bfd_malloc_and_read): Likewise.
* libbfd.h: Regenerate.
tc-aarch64.c:1473:27: runtime error: left shift of 7 by 30 places
cannot be represented in type 'int'.
* config/tc-aarch64.c (parse_vector_reg_list): Avoid UB left
shift.
This should sort out some very old FIXMEs in code handling stabs
debug info. Necessary if we are to fuss over freeing up memory before
objdump and objcopy exit. It is of course better from a user
viewpoint to *not* free memory, which takes some time, and leave that
to process exit. The only reason to do so is that having many memory
leaks in binutils/ code tends to hide leaks in bfd/ or opcodes/, which
we should care about.
* budbg.h (parse_stab): Update prototype.
* debug.h (debug_start_source): Update prototype.
* debug.c (debug_start_source): Add name_used. Set if stashed.
* rddbg.c (read_symbol_stabs_debugging_info): Always malloc
stab string passed to parse_stab. Free stab string when
unreferenced.
(read_section_stabs_debugging_info): Likewise, and strings
section contents.
* stabs.c (parse_stab): Add string_used param. Set if string
stashed. Pass to debug_start_source. Realloc file_types
array rather that using malloc. Clarify comment about
debug_make_indirect_type.
We may have added some abbrevs to the list before hitting an error.
Free the list elements too. free_abbrev_list returns list->next so we
need to init it earlier to avoid an uninitialised memory access.
* dwarf.c (process_abbrev_set): Call free_abbrev_list on errors.
Set list->next earlier.
I noticed the prefix parameter was unused in print_doc_of_command. And
when removing it, it becomes unused in apropos_cmd.
Change-Id: Id72980b03fe091b22931e6b85945f412b274ed5e
GDB expected PC should point right after the SVC instruction when the
syscall is active. But some active syscalls keep PC pointing to the SVC
instruction itself.
This leads to a broken backtrace like:
Backtrace stopped: previous frame identical to this frame (corrupt stack?)
#0 0xb6f8681c in pthread_cond_timedwait@@GLIBC_2.4 () from /lib/arm-linux-gnueabihf/libpthread.so.0
#1 0xb6e21f80 in ?? ()
The reason is that .ARM.exidx unwinder gives up if PC does not point
right after the SVC (syscall) instruction. I did not investigate why but
some syscalls will point PC to the SVC instruction itself. This happens
for the "futex" syscall used by pthread_cond_timedwait.
That normally does not matter as ARM prologue unwinder gets called
instead of the .ARM.exidx one. Unfortunately some glibc calls have more
complicated prologue where the GDB unwinder fails to properly determine
the return address (that is in fact an orthogonal GDB bug). I expect it
is due to the "vpush" there in this case but I did not investigate it more:
Dump of assembler code for function pthread_cond_timedwait@@GLIBC_2.4:
0xb6f8757c <+0>: push {r4, r5, r6, r7, r8, r9, r10, r11, lr}
0xb6f87580 <+4>: mov r10, r2
0xb6f87584 <+8>: vpush {d8}
Regression tested on armv7l kernel 5.15.32-v7l+ (Raspbian 11).
Approved-By: Luis Machado <luis.machado@arm.com>
Linker adds indirect symbols for versioned symbol aliases, which are
created by ".symver foo, foo@FOO", by checking symbol type, value and
section so that references to foo will be replaced by references to
foo@FOO if foo and foo@FOO have the same symbol type, value and section.
But in IR, since all symbols of the same type have the same value and
section, we can't tell if a symbol is an alias of another symbol by
their types, values and sections. We shouldn't add indirect symbols
for versioned symbol aliases in IR.
bfd/
PR ld/30281
* elflink.c (elf_link_add_object_symbols): Don't add indirect
symbols for ".symver foo, foo@FOO" aliases in IR.
ld/
PR ld/30281
* testsuite/ld-plugin/lto.exp: Add PR ld/30281 test.
* testsuite/ld-plugin/pr30281.t: New file.
* testsuite/ld-plugin/pr30281.c: Likewise.
A recent patch caused my system gcc (Fedora 36, so gcc 12.2.1) to warn
about sym_addr being possibly uninitialized in frame.c. It isn't, but
the compiler can't tell. So, this patch initializes the variable. I
also fixed a formatting buglet that I missed in review.
With test-case gdb.base/trace-commands.exp and editing off, I run into fails
because multi-line commands are issued using gdb_test_sequence, which
doesn't handle them correctly.
Fix this by using gdb_test instead.
Tested on x86_64-linux.
PR testsuite/30288
Bug: https://sourceware.org/bugzilla/show_bug.cgi?id=30288
With test-case gdb.threads/threadapply.exp and editing set to on, we have:
...
(gdb) define remove^M
Type commands for definition of "remove".^M
End with a line saying just "end".^M
>remove-inferiors 3^M
>end^M
(gdb)
...
but with editing set to off, we run into:
...
(gdb) define remove^M
Type commands for definition of "remove".^M
End with a line saying just "end".^M
>remove-inferiors 3^M
end^M
>(gdb) FAIL: gdb.threads/threadapply.exp: thread_set=all: try remove: \
define remove (timeout)
...
The commands are issued by this test:
...
gdb_define_cmd "remove" {
"remove-inferiors 3"
}
...
which does:
- gdb_test_multiple "define remove", followed by
- gdb_test_multiple "remove-inferiors 3\nend".
Proc gdb_test_multiple has special handling for multi-line commands, which
splits it up into subcommands, and for each subcommand issues it and then
waits for the resulting prompt (the secondary prompt ">" for all but the last
subcommand).
However, that doesn't work as expected in this case because the initial
gdb_test_multiple "define remove" fails to match all resulting output, and
consequently the secondary prompt resulting from "define remove" is counted as
if it was the one resulting from "remove-inferiors 3".
Fix this by matching the entire output of "define remove", including the
secondary prompt.
Tested on x86_64-linux.
PR testsuite/30288
Bug: https://sourceware.org/bugzilla/show_bug.cgi?id=30288
I noticed that language_demangle shadows the global
"current_language". When I went to fix this, though, I then saw that
language_demangle is only called in two places, and has a comment
saying it should be removed. This patch removes it. Note that the
NULL check in language_demangle is not needed by either of the
existing callers.
Regression tested on x86-64 Fedora 36.
Approved-By: Simon Marchi <simon.marchi@efficios.com>
Tom de Vries pointed out a bug in the index-cache background writer --
sometimes it will fail. He also noted that it fails when the number
of worker threads is set to zero. These turn out to be the same
problem -- the cache can't be written to until the per-BFD's
"index_table" member is set.
This patch avoids the race by rearranging the code slightly, to ensure
the cache cannot possibly be written before the member is set.
Bug: https://sourceware.org/bugzilla/show_bug.cgi?id=30261
Allow consumers of GDB to extract the name of the main method. This is
most useful for Fortran programs which have a variable main method.
Used by both MAP and DDT e.g. it is used to detect the presence of debug
information.
Co-Authored-By: Maciej W. Rozycki <macro@embecosm.com>
Fix a functional regression and restore the handling of DW_CC_program
code of DW_AT_calling_convention attribute for determining the name of
the starting function of the program where the DW_AT_main_subprogram
attribute has not been provided, such as with Fortran code compiled with
GCC versions 4.5.4 and below, or where DWARF version 3 or below has been
requested. Without it "main" is considered the starting function. Cf.
GCC PR fortran/43414.
Original code was removed with commit 6209cde4dd ("Delete DWARF
psymtab code"), and then an update to complement commit 81873cc81e
("[gdb/symtab] Support DW_AT_main_subprogram with -readnow.") has also
been included here.
In the case where a Fortran program has a program name of "main" and
there is also a minimal symbol called main, such as with programs built
with GCC version 4.4.7 or below, the backtrace will erroneously stop at
the minimal symbol rather than the user specified main, e.g.:
(gdb) bt
#0 bar () at .../gdb/testsuite/gdb.fortran/backtrace.f90:17
#1 0x0000000000402556 in foo () at .../gdb/testsuite/gdb.fortran/backtrace.f90:21
#2 0x0000000000402575 in main () at .../gdb/testsuite/gdb.fortran/backtrace.f90:31
#3 0x00000000004025aa in main ()
(gdb)
This patch fixes this issue by increasing the precedence of the full
symbol when the language of the current frame is Fortran.
Newer versions of GCC transform the program name to "MAIN__" in this
case, avoiding the problem.
Co-Authored-By: Maciej W. Rozycki <macro@embecosm.com>
This commit intends to move operands that require very special handling or
operand types that are so minor (e.g. only useful on a few instructions)
under "W". I also intend this "W" to be "temporary" operand storage until
we can find good two character (or less) operand type.
In this commit, prefetch offset operand "f" for 'Zicbop' extension is moved
to "Wif" because of its special handling (and allocating single character
"f" for this operand type seemed too much).
Current expected allocation guideline is as follows:
1. 'W'
2. The most closely related single-letter extension in lowercase
(strongly recommended but not mandatory)
3. Identify operand type
The author currently plans to allocate following three-character operand
types (for operands including instructions from unratified extensions).
1. "Wif" ('Zicbop': fetch offset)
2. "Wfv" (unratified 'Zfa': value operand from FLI.[HSDQ] instructions)
3. "Wfm" / "WfM"
'Zfh', 'F', 'D', 'Q': rounding modes "m" with special handling
solely for widening conversion instructions.
gas/ChangeLog:
* config/tc-riscv.c (validate_riscv_insn, riscv_ip): Move from
"f" to "Wif".
opcodes/ChangeLog:
* riscv-dis.c (print_insn_args): Move from "f" to "Wif".
* riscv-opc.c (riscv_opcodes): Reflect new operand type.
This can't be done for all insns currently encoded with .byte. For one
outside of 64-bit mode unused (typically ignored) register encoding bits
in VEX/XOP/EVEX prefixes can't be set to their non-default values, since
the necessary registers cannot be specified (and some of these bits
can't even be used outside of 64-bit mode). And then there are odd tests
like the first one in bad-bcast.s: Its purpose is to illegaly set EVEX.b
together with EVEX.W (which could be expressed; note though EVEX.W set
is invalid on its own), but then it also clears EVEX.B and EVEX.R' plus
it sets EVEX.vvvv to other than 0xf (rendering the test ambiguous,
because that's another #UD reason).
In {,x86-64-}disassem.s many bogus encodings exist - some with ModR/M
byte but insufficient displacement bytes, some using SIB encoding with
the SIB byte actually being the supposed immediate. Some of these could
be expressed by .insn, but I don't want to introduce bogus examples.
These will all need adjustment anyway once the disassembler is improved
in the way it deals with unrecognized encodings.
Generally generated code is meant to remain the same. {,x86-64-}nops.d
are exceptions because insn prefixes are emitted in a different order.
opcode{,-intel,-suffix}.d are also adjusted (along with an according
correction to opcode.s) to cover an apparent typo in the original tests
(xor when or was meant).
Where necessary --divide is added as gas option, to allow for the use
of the extension opcode functionality.
Comments are being adjusted where obviously wrong/misleading.
Since we have no insn suffix and it's also not realistic to infer
immediate size from the size of other (typically register) operands
(like optimize_imm() does), and since we also don't have a template
telling us permitted size(s), a new syntax construct is introduced to
allow size (and signedness) specification. In the absence of such, the
size is inferred from significant bits (which obviously may yield
inconsistent results at least for effectively negative values, depending
on whether BFD64 is enabled), and only if supplied expressions can be
evaluated at parsing time. Being explicit is generally recommended to
users.
Size specification is permitted at bit granularity, but of course the
eventually emitted immediate values will be padded up to 8-, 16-, 32-,
or 64-bit fields.
In particular the scaling factor cannot always be determined from pre-
existing operand attributes. Introduce a new {:d<N>} vector operand
syntax extension, restricted to .insn only, to allow specifying this in
(at least) otherwise ambiguous cases.
Deal with register and memory operands; immediate operands will follow
later, as will the handling of EVEX embedded broadcast and EVEX Disp8
scaling.
Note that because we can't really know how to encode their use, %cr8 and
up cannot be used with .insn outside of 64-bit mode. Users would need to
specify an explicit LOCK prefix in combination with %cr0 etc.
So called "short form" encoding is specified by a trailing "+r", whereas
a possible extension opcode is specified by the usual "/<digit>". Take
these off the expression before handing it to get_absolute_expression().
Note that on targets where / starts a comment, --divide needs passing to
gas in order to make use of the extension opcode functionality.
All encoding spaces can be used this way; there's a certain risk that
the bits presently reserved could be used for other purposes down the
road, but people using .insn are expected to know what they're doing
anyway. Plus this way there's at least _some_ way to have those bits
set.
For now this will only allow operand-less insns to be encoded this way.
While only a secondary issue there, the testcase of PR gas/27212 exposes
an oversight in relocation handling: Just like e.g. Arm32, which has a
similar comment and a similar check, relocations against STN_UNDEF have
to be permitted to satisfy the ELF spec.
PR 30219 shows an internal error due to a "Bad switch" in
print_exception() in gdb/exceptions.c. The switch in question
contains cases for RETURN_QUIT and RETURN_ERROR, but is missing a case
for the recently added RETURN_FORCED_QUIT. This commit adds that case.
Making the above change allows the errant test case to pass, but does
not fix the underlying problem, which I'll describe shortly. Even
though the addition of a case for RETURN_FORCED_QUIT isn't the actual
fix, I still think it's important to add this case so that other
situations which lead to print_exeption() being called won't generate
that "Bad switch" internal error.
In order to understand the underlying problem, please examine
this portion of the backtrace from the bug report:
0x5576e4ff5780 print_exception
/home/smarchi/src/binutils-gdb/gdb/exceptions.c:100
0x5576e4ff5930 exception_print(ui_file*, gdb_exception const&)
/home/smarchi/src/binutils-gdb/gdb/exceptions.c:110
0x5576e6a896dd quit_force(int*, int)
/home/smarchi/src/binutils-gdb/gdb/top.c:1849
The real problem is in quit_force; here's the try/catch which
eventually leads to the internal error:
/* Get out of tfind mode, and kill or detach all inferiors. */
try
{
disconnect_tracing ();
for (inferior *inf : all_inferiors ())
kill_or_detach (inf, from_tty);
}
catch (const gdb_exception &ex)
{
exception_print (gdb_stderr, ex);
}
While running the calls in the try-block, a QUIT check is being
performed. This check finds that sync_quit_force_run is (still) set,
causing a gdb_exception_forced_quit to be thrown. The exception
gdb_exception_forced_quit is derived from gdb_exception, causing
exception_print to be called. As shown by the backtrace,
print_exception is then called, leading to the internal error.
The actual fix, also implemented by this commit, is to clear
sync_quit_force_run along with the quit flag. This will allow the
various cleanup code, called by quit_force, to run without triggering
a gdb_exception_forced_quit. (Though, if another SIGTERM is sent to
the gdb process, these flags will be set again and a QUIT check in the
cleanup code will detect it and throw the exception.)
Bug: https://sourceware.org/bugzilla/show_bug.cgi?id=30219
Approved-By: Simon Marchi <simon.marchi@efficios.com>
