| Age | Commit message (Collapse) | Author | Files | Lines |
|---|
|
When running gdb.base/bg-exec-sigint-bp-cond.exp when SHELL is dash,
rather than bash, I get:
c&^M
Continuing.^M
(gdb) sh: 1: kill: Illegal option -S^M
^M
Breakpoint 2, foo () at /home/jenkins/smarchi/binutils-gdb/build/gdb/testsuite/../../../gdb/testsuite/gdb.base/bg-exec-sigint-bp-cond.c:23^M
23 return 0;^M
FAIL: gdb.base/bg-exec-sigint-bp-cond.exp: no force memory write: SIGINT does not interrupt background execution (timeout)
This is because it uses the kill command built-in the dash shell, and
using the SIG prefix with kill does not work with dash's kill. The
difference is listed in the documentation for bash's POSIX-correct mode
[1]:
The kill builtin does not accept signal names with a ‘SIG’ prefix.
Replace SIGINT with INT in that test.
By grepping, I found two other instances (gdb.base/sigwinch-notty.exp
and gdb.threads/detach-step-over.exp). Those were not problematic on my
system though. Since they are done through remote_exec, they don't go
through the shell and therefore invoke /bin/kill. On my Arch Linux,
it's:
$ /bin/kill --version
kill from util-linux 2.38.1 (with: sigqueue, pidfd)
and on my Ubuntu:
$ /bin/kill --version
kill from procps-ng 3.3.17
These two implementations accept "-SIGINT". But according to the POSIX
spec [2], the kill utility should recognize the signal name without the
SIG prefix (if it recognizes them with the SIG prefix, it's an
extension):
-s signal_name
Specify the signal to send, using one of the symbolic names defined
in the <signal.h> header. Values of signal_name shall be recognized
in a case-independent fashion, without the SIG prefix. In addition,
the symbolic name 0 shall be recognized, representing the signal
value zero. The corresponding signal shall be sent instead of SIGTERM.
-signal_name
[XSI] [Option Start]
Equivalent to -s signal_name. [Option End]
So, just in case some /bin/kill implementation happens to not recognize
the SIG prefixes, change these two other calls to remove the SIG
prefix.
[1] https://www.gnu.org/software/bash/manual/html_node/Bash-POSIX-Mode.html
[2] https://pubs.opengroup.org/onlinepubs/9699919799/utilities/kill.html
Change-Id: I81ccedd6c9428ab63b9261813f1905a18941f8da
Reviewed-By: Tom Tromey <tom@tromey.com>
|
|
The copyright years in the ROCm files (e.g. solib-rocm.c) are wrong,
they end in 2022 instead of 2023. I suppose because I posted (or at
least prepared) the patches in 2022 but merged them in 2023, and forgot
to update the year. I found a bunch of other files that are in the same
situation. Fix them all up.
Change-Id: Ia55f5b563606c2ba6a89046f22bc0bf1c0ff2e10
Reviewed-By: Tom Tromey <tom@tromey.com>
|
|
Background
----------
When a thread-specific breakpoint is deleted as a result of the
specific thread exiting the function remove_threaded_breakpoints is
called which sets the disposition of the breakpoint to
disp_del_at_next_stop and sets the breakpoint number to 0. Setting
the breakpoint number to zero has the effect of hiding the breakpoint
from the user. We also print a message indicating that the breakpoint
has been deleted.
It was brought to my attention during a review of another patch[1]
that setting a breakpoints number to zero will suppress the MI
breakpoint-deleted notification for that breakpoint, and indeed, this
can be seen to be true, in delete_breakpoint, if the breakpoint number
is zero, then GDB will not notify the breakpoint_deleted observer.
It seems wrong that a user created, thread-specific breakpoint, will
have a =breakpoint-created notification, but will not have a
=breakpoint-deleted notification. I suspect that this is a bug.
[1] https://sourceware.org/pipermail/gdb-patches/2023-February/196560.html
The First Problem
-----------------
During my initial testing I wanted to see how GDB handled the
breakpoint after it's number was set to zero. To do this I created
the testcase gdb.threads/thread-bp-deleted.exp. This test creates a
worker thread, which immediately exits. After the worker thread has
exited the main thread spins in a loop.
In GDB I break once the worker thread has been created and place a
thread-specific breakpoint, then use 'continue&' to resume the
inferior in non-stop mode. The worker thread then exits, but the main
thread never stops - instead it sits in the spin. I then tried to use
'maint info breakpoints' to see what GDB thought of the
thread-specific breakpoint.
Unfortunately, GDB crashed like this:
(gdb) continue&
Continuing.
(gdb) [Thread 0x7ffff7c5d700 (LWP 1202458) exited]
Thread-specific breakpoint 3 deleted - thread 2 no longer in the thread list.
maint info breakpoints
... snip some output ...
Fatal signal: Segmentation fault
----- Backtrace -----
0x5ffb62 gdb_internal_backtrace_1
../../src/gdb/bt-utils.c:122
0x5ffc05 _Z22gdb_internal_backtracev
../../src/gdb/bt-utils.c:168
0x89965e handle_fatal_signal
../../src/gdb/event-top.c:964
0x8997ca handle_sigsegv
../../src/gdb/event-top.c:1037
0x7f96f5971b1f ???
/usr/src/debug/glibc-2.30-2-gd74461fa34/nptl/../sysdeps/unix/sysv/linux/x86_64/sigaction.c:0
0xe602b0 _Z15print_thread_idP11thread_info
../../src/gdb/thread.c:1439
0x5b3d05 print_one_breakpoint_location
../../src/gdb/breakpoint.c:6542
0x5b462e print_one_breakpoint
../../src/gdb/breakpoint.c:6702
0x5b5354 breakpoint_1
../../src/gdb/breakpoint.c:6924
0x5b58b8 maintenance_info_breakpoints
../../src/gdb/breakpoint.c:7009
... etc ...
As the thread-specific breakpoint is set to disp_del_at_next_stop, and
GDB hasn't stopped yet, then the breakpoint still exists in the global
breakpoint list.
The breakpoint will not show in 'info breakpoints' as its number is
zero, but it will show in 'maint info breakpoints'.
As GDB prints the breakpoint, the thread-id for the breakpoint is
printed as part of the 'stop only in thread ...' line. Printing the
thread-id involves calling find_thread_global_id to convert the global
thread-id into a thread_info*. Then calling print_thread_id to
convert the thread_info* into a string.
The problem is that find_thread_global_id returns nullptr as the
thread for the thread-specific breakpoint has exited. The
print_thread_id assumes it will be passed a non-nullptr. As a result
GDB crashes.
In this commit I've added an assert to print_thread_id (gdb/thread.c)
to check that the pointed passed in is not nullptr. This assert would
have triggered in the above case before GDB crashed.
MI Notifications: The Dangers Of Changing A Breakpoint's Number
---------------------------------------------------------------
Currently the delete_breakpoint function doesn't trigger the
breakpoint_deleted observer for any breakpoint with the number zero.
There is a comment explaining why this is the case in the code; it's
something about watchpoints. But I did consider just removing the 'is
the number zero' guard and always triggering the breakpoint_deleted
observer, figuring that I'd then fix the watchpoint issue some other
way.
But I realised this wasn't going to be good enough. When the MI
notification was delivered the number would be zero, so any frontend
parsing the notifications would not be able to match
=breakpoint-deleted notification to the earlier =breakpoint-created
notification.
What this means is that, at the point the breakpoint_deleted observer
is called, the breakpoint's number must be correct.
MI Notifications: The Dangers Of Delaying Deletion
--------------------------------------------------
The test I used to expose the above crash also brought another problem
to my attention. In the above test we used 'continue&' to resume,
after which a thread exited, but the inferior didn't stop. Recreating
the same test in the MI looks like this:
-break-insert -p 2 main
^done,bkpt={number="2",type="breakpoint",disp="keep",...<snip>...}
(gdb)
-exec-continue
^running
*running,thread-id="all"
(gdb)
~"[Thread 0x7ffff7c5d700 (LWP 987038) exited]\n"
=thread-exited,id="2",group-id="i1"
~"Thread-specific breakpoint 2 deleted - thread 2 no longer in the thread list.\n"
At this point the we have a single thread left, which is still
running:
-thread-info
^done,threads=[{id="1",target-id="Thread 0x7ffff7c5eb80 (LWP 987035)",name="thread-bp-delet",state="running",core="4"}],current-thread-id="1"
(gdb)
Notice that we got the =thread-exited notification from GDB as soon as
the thread exited. We also saw the CLI line from GDB, the line
explaining that breakpoint 2 was deleted. But, as expected, we didn't
see the =breakpoint-deleted notification.
I say "as expected" because the number was set to zero. But, even if
the number was not set to zero we still wouldn't see the
notification. The MI notification is driven by the breakpoint_deleted
observer, which is only called when we actually delete the breakpoint,
which is only done the next time GDB stops.
Now, maybe this is fine. The notification is delivered a little
late. But remember, by setting the number to zero the breakpoint will
be hidden from the user, for example, the breakpoint is removed from
the MI's -break-info command output.
This means that GDB is in a position where the breakpoint doesn't show
up in the breakpoint table, but a =breakpoint-deleted notification has
not yet been sent out. This doesn't seem right to me.
What this means is that, when the thread exits, we should immediately
be sending out the =breakpoint-deleted notification. We should not
wait for GDB to next stop before sending the notification.
The Solution
------------
My proposed solution is this; in remove_threaded_breakpoints, instead
of setting the disposition to disp_del_at_next_stop and setting the
number to zero, we now just call delete_breakpoint directly.
The notification will now be sent out immediately; as soon as the
thread exits.
As the number has not changed when delete_breakpoint is called, the
notification will have the correct number.
And as the breakpoint is immediately removed from the breakpoint list,
we no longer need to worry about 'maint info breakpoints' trying to
print the thread-id for an exited thread.
My only concern is that calling delete_breakpoint directly seems so
obvious that I wonder why the original patch (that added
remove_threaded_breakpoints) didn't take this approach. This code was
added in commit 49fa26b0411d, but the commit message offers no clues
to why this approach was taken, and the original email thread offers
no insights either[2]. There are no test regressions after making
this change, so I'm hopeful that this is going to be fine.
[2] https://sourceware.org/pipermail/gdb-patches/2013-September/106493.html
The Complication
----------------
Of course, it couldn't be that simple.
The script gdb.python/py-finish-breakpoint.exp had some regressions
during testing.
The problem was with the FinishBreakpoint.out_of_scope callback
implementation. This callback is supposed to trigger whenever the
FinishBreakpoint goes out of scope; and this includes when the thread
for the breakpoint exits.
The problem I ran into is the Python FinishBreakpoint implementation.
Specifically, after this change I was loosing some of the out_of_scope
calls.
The problem is that the out_of_scope call (of which I'm interested) is
triggered from the inferior_exit observer. Before my change the
observers were called in this order:
thread_exit
inferior_exit
breakpoint_deleted
The inferior_exit would trigger the out_of_scope call.
After my change the breakpoint_deleted notification (for
thread-specific breakpoints) occurs earlier, as soon as the
thread-exits, so now the order is:
thread_exit
breakpoint_deleted
inferior_exit
Currently, after the breakpoint_deleted call the Python object
associated with the breakpoint is released, so, when we get to the
inferior_exit observer, there's no longer a Python object to call the
out_of_scope method on.
My solution is to follow the model for how bpfinishpy_pre_stop_hook
and bpfinishpy_post_stop_hook are called, this is done from
gdbpy_breakpoint_cond_says_stop in py-breakpoint.c.
I've now added a new bpfinishpy_pre_delete_hook
gdbpy_breakpoint_deleted in py-breakpoint.c, and from this new hook
function I check and where needed call the out_of_scope method.
With this fix in place I now see the
gdb.python/py-finish-breakpoint.exp test fully passing again.
Testing
-------
Tested on x86-64/Linux with unix, native-gdbserver, and
native-extended-gdbserver boards.
New tests added to covers all the cases I've discussed above.
Approved-By: Pedro Alves <pedro@palves.net>
|
|
I noticed that several tests included copy & pasted code to run the
'maint show target-non-stop' command, and then switch based on the
result.
In this commit I factor this code out into a helper proc in
lib/gdb.exp, and update all the places I could find that used this
pattern to make use of the helper proc.
There should be no change in what is tested after this commit.
Reviewed-By: Pedro Alves <pedro@palves.net>
|
|
With:
- catch a fork in thread 1
- select thread 2
- set follow-fork child
- next
... follow_fork notices that thread 1 had last stopped for a fork
which hasn't been followed yet, and because thread 1 is not the
current thread, GDB aborts the execution command, presenting the stop
in thread 1.
That makes sense, as only the forking thread (thread 1) survives in
the child, so better stop and let the user decide how to proceed.
However, with:
- catch a fork in thread 1
- select thread 2
- set follow-fork parent << note difference here
- next
... GDB does the same: follow_fork notices that thread 1 had last
stopped for a fork which hasn't been followed yet, and because thread
1 is not the current thread, GDB aborts the execution command,
presenting the stop in thread 1.
Aborting/stopping in this case doesn't make sense to me. As we're
following the parent, thread 2 will still continue to exist in the
parent. What the child does after we've followed the parent shouldn't
matter -- it can go on running free, be detached, etc., depending on
"set schedule-multiple", "set detach-on-fork", etc. That does not
influence the execution command that the user issued for the parent
thread.
So this patch changes GDB in that direction -- in follow_fork, if
following the parent, and we've switched threads meanwhile, switch
back to the unfollowed thread, follow it (stay with the parent), and
don't abort/stop. If we're following a fork (as opposed to vfork),
then switch back again to the thread that the user was trying to
resume. If following a vfork, however, stay with the vforking-thread
selected, as we will need to see a vfork_done event first, before we
can resume any other thread.
As I was working on this, I managed to end up calling target_resume
for a solo-thread resume (to collect the vfork_done event), with
scope_ptid pointing at the vfork parent thread, and inferior_ptid
pointing to the vfork child. For a solo-thread resume, the scope_ptid
argument to target_resume must the same as inferior_ptid. The mistake
was caught by the assertion in target_resume, like so:
...
[infrun] resume_1: step=0, signal=GDB_SIGNAL_0, trap_expected=0, current thread [1722839.1722839.0] at 0x5555555553c3
[infrun] do_target_resume: resume_ptid=1722839.1722939.0, step=0, sig=GDB_SIGNAL_0
../../src/gdb/target.c:2661: internal-error: target_resume: Assertion `inferior_ptid.matches (scope_ptid)' failed.
...
but I think it doesn't hurt to catch such a mistake earlier, hence the
change in internal_resume_ptid.
Change-Id: I896705506a16d2488b1bfb4736315dd966f4e412
|
|
Since commit 9af467b8240 ("[gdb/testsuite] Fix gdb.threads/schedlock.exp on
fast cpu"), the test-case fails for gcc 4.8.5.
The problem is that for gcc 4.8.5, the commit turned a two-line loop:
...
(gdb) next
78 while (*myp > 0)
(gdb) next
81 MAYBE_CALL_SOME_FUNCTION(); (*myp) ++;
(gdb) next
78 while (*myp > 0)
...
into a three-line loop:
...
(gdb) next
83 MAYBE_CALL_SOME_FUNCTION(); (*myp) ++;
(gdb) next
84 cnt++;
(gdb) next
85 }
(gdb) next
83 MAYBE_CALL_SOME_FUNCTION(); (*myp) ++;
(gdb)
...
and the test-case doesn't expect this.
Fix this by reverting back to the original loop shape as much as possible by:
- removing the cnt++ line
- replacing "while (1)" with "while (one)", where one is a volatile variable
set to 1.
Tested on x86_64-linux, using compilers:
- gcc 4.8.5, 7.5.0, 12.2.1
- clang 4.0.1, 13.0.1
|
|
Occasionally, I run into:
...
(gdb) PASS: gdb.threads/schedlock.exp: schedlock=on: cmd=continue: \
set scheduler-locking on
continue^M
Continuing.^M
PASS: gdb.threads/schedlock.exp: schedlock=on: cmd=continue: \
continue (with lock)
[Thread 0x7ffff746e700 (LWP 1339) exited]^M
No unwaited-for children left.^M
(gdb) Quit^M
(gdb) FAIL: gdb.threads/schedlock.exp: schedlock=on: cmd=continue: \
stop all threads (with lock) (timeout)
...
What happens is that this loop which is supposed to run "just short of forever":
...
/* Don't run forever. Run just short of it :) */
while (*myp > 0)
{
/* schedlock.exp: main loop. */
MAYBE_CALL_SOME_FUNCTION(); (*myp) ++;
}
...
finishes after 0x7fffffff iterations (when a signed wrap occurs), which on my
system takes only about 1.5 seconds.
Fix this by:
- changing the pointed-at type of myp from signed to unsigned, which makes the
wrap defined behaviour (and which also make the loop run twice as long,
which is already enough to make it impossible for me to reproduce the FAIL.
But let's try to solve this more structurally).
- changing the pointed-at type of myp from int to long long, making the wrap
unlikely.
- making sure the loop runs forever, by setting the loop condition to 1.
- making sure the loop still contains different lines (as far as debug info is
concerned) by incrementing a volatile counter in the loop.
- making sure the program doesn't run forever in case of trouble, by adding an
"alarm (30)".
Tested on x86_64-linux.
PR testsuite/30074
Bug: https://sourceware.org/bugzilla/show_bug.cgi?id=30074
|
|
When creating a breakpoint or watchpoint, the 'thread' and 'task'
keywords can be used to create a thread or task specific breakpoint or
watchpoint.
Currently, a thread or task specific breakpoint can only apply for a
single thread or task, if multiple threads or tasks are specified when
creating the breakpoint (or watchpoint), then the last specified id
will be used.
The exception to the above is that when the 'thread' keyword is used
during the creation of a watchpoint, GDB will give an error if
'thread' is given more than once.
In this commit I propose making this behaviour consistent, if the
'thread' or 'task' keywords are used more than once when creating
either a breakpoint or watchpoint, then GDB will give an error.
I haven't updated the manual, we don't explicitly say that these
keywords can be repeated, and (to me), given the keyword takes a
single id, I don't think it makes much sense to repeat the keyword.
As such, I see this more as adding a missing error to GDB, rather than
making some big change. However, I have added an entry to the NEWS
file as I guess it is possible that some people might hit this new
error with an existing (I claim, badly written) GDB script.
I've added some new tests to check for the new error.
Just one test needed updating, gdb.linespec/keywords.exp, this test
did use the 'thread' keyword twice, and expected the breakpoint to be
created. Looking at what this test was for though, it was checking
the use of '-force-condition', and I don't think that being able to
repeat 'thread' was actually a critical part of this test.
As such, I've updated this test to expect the error when 'thread' is
repeated.
|
|
The test gdb.threads/thread-specific-bp.exp tries to set non-stop mode
on a running target, something which the manual makes clear is not
allowed.
This commit restructures the test a little, we now set the non-stop
mode as part of the GDBFLAGS, so the mode will be set before GDB
connects to the target. As a consequence I'm able to move the
with_test_prefix out of the check_thread_specific_breakpoint proc.
The check_thread_specific_breakpoint proc is now called within a loop.
After this commit the gdb.threads/thread-specific-bp.exp test still
has some failures, this is because of an issue GDB currently has
printing "Thread ... exited" messages. This problem should be
addressed by this patch:
https://sourceware.org/pipermail/gdb-patches/2022-December/194694.html
when it is merged.
|
|
This patch applies the appropriate FIXME notes described in commit 5b6d1e4
"Multi-target support".
"You'll notice that remote.c includes some FIXME notes. These refer to
the fact that the global arrays that hold data for the remote packets
supported are still globals. For example, if we connect to two
different servers/stubs, then each might support different remote
protocol features. They might even be different architectures, like
e.g., one ARM baremetal stub, and a x86 gdbserver, to debug a
host/controller scenario as a single program. That isn't going to
work correctly today, because of said globals. I'm leaving fixing
that for another pass, since it does not appear to be trivial, and I'd
rather land the base work first. It's already useful to be able to
debug multiple instances of the same server (e.g., a distributed
cluster, where you have full control over the servers installed), so I
think as is it's already reasonable incremental progress."
Using this patch it is possible to configure per-remote targets'
feature packets.
Given the following setup for two gdbservers:
~~~~
gdbserver --multi :1234
gdbserver --disable-packet=vCont --multi :2345
~~~~
Before this patch configuring of range-stepping was not possible for one
of two connected remote targets with different support for the vCont
packet. As one of the targets supports vCont, it should be possible to
configure "set range-stepping". However, the output of GDB looks like:
(gdb) target extended-remote :1234
Remote debugging using :1234
(gdb) add-inferior -no-connection
[New inferior 2]
Added inferior 2
(gdb) inferior 2
[Switching to inferior 2 [<null>] (<noexec>)]
(gdb) target extended-remote :2345
Remote debugging using :2345
(gdb) set range-stepping on
warning: Range stepping is not supported by the current target
(gdb) inferior 1
[Switching to inferior 1 [<null>] (<noexec>)]
(gdb) set range-stepping on
warning: Range stepping is not supported by the current target
~~~~
Two warnings are shown. The warning for inferior 1 should not appear
as it is connected to a target supporting the vCont package.
~~~~
(gdb) target extended-remote :1234
Remote debugging using :1234
(gdb) add-inferior -no-connection
[New inferior 2]
Added inferior 2
(gdb) inferior 2
[Switching to inferior 2 [<null>] (<noexec>)]
(gdb) target extended-remote :2345
Remote debugging using :2345
(gdb) set range-stepping on
warning: Range stepping is not supported by the current target
(gdb) inferior 1
[Switching to inferior 1 [<null>] (<noexec>)]
(gdb) set range-stepping on
(gdb)
~~~~
Now only one warning is shown for inferior 2, which is connected to
a target not supporting vCont.
The per-remote target feature array is realized by a new class
remote_features, which stores the per-remote target array and
provides functions to determine supported features of the target.
A remote_target object now has a new member of that class.
Each time a new remote_target object is initialized, a new per-remote
target array is constructed based on the global remote_protocol_packets
array. The global array is initialized in the function _initialize_remote
and can be configured using the command line. Before this patch the
command line configuration affected current targets and future remote
targets (due to the global feature array used by all remote
targets). This behavior is different and the configuration applies as
follows:
- If a target is connected, the command line configuration affects the
current connection. All other existing remote targets are not
affected.
- If not connected, the command line configuration affects future
connections.
The show command displays the current remote target's configuration. If no
remote target is selected the default configuration for future
connections is shown.
If we have for instance the following setup with inferior 2 being
selected:
~~~~
(gdb) info inferiors
Num Description Connection Executable
1 <null> 1 (extended-remote :1234)
* 2 <null> 2 (extended-remote :2345)
~~~~
Before this patch, if we run 'set remote multiprocess-feature-packet', the
following configuration was set:
The feature array of all remote targets (in this setup the two connected
targets) and all future remote connections are affected.
After this patch, it will be configured as follows:
The feature array of target with port :2345 which is currently selected
will be configured. All other existing remote targets are not affected.
The show command 'show remote multiprocess-feature-packet' will display
the configuration of target with port :2345.
Due to this configuration change, it is required to adapt the test
"gdb/testsuite/gdb.multi/multi-target-info-inferiors.exp" to configure the
multiprocess-feature-packet before the connections are created.
To inform the gdb user about the new behaviour of the 'show remote
PACKET-NAME' commands and the new configuration impact for remote
targets using the 'set remote PACKET-NAME' commands the commands'
outputs are adapted. Due to this change it is required to adapt each
test using the set/show remote 'PACKET-NAME' commands.
|
|
Change gdb.threads to use clean_restart more consistently.
|
|
A number of tests end with "return". However, this is unnecessary.
This patch removes all of these.
|
|
This changes some tests to use require with 'is_remote', rather than
an explicit test. This adds uniformity helps clean up more spots
where a test might exit early without any notification.
|
|
A number of tests will exit early without saying why. This patch adds
"unsupported" at spots like this that I could readily find.
There are definitely more of these; for example dw2-ranges.exp fails
silently because a compilation fails. I didn't attempt to address
these as that is a much larger task.
|
|
This changes skip_hw_breakpoint_tests to invert the sense, and renames
it to allow_hw_breakpoint_tests. This also converts some tests to use
"require" -- I missed this particular check in the first series.
|
|
This changes skip_shlib_tests to invert the sense, and renames it to
allow_shlib_tests.
|
|
This changes skip_python_tests to invert the sense, and renames it to
allow_python_tests.
|
|
This changes skip_hw_watchpoint_tests to invert the sense, and renames
it to allow_hw_watchpoint_tests.
|
|
This changes skip_hw_watchpoint_multi_tests to invert the sense, and
renames it to allow_hw_watchpoint_multi_tests.
|
|
This changes skip_hw_watchpoint_access_tests to invert the sense, and
renames it to allow_hw_watchpoint_access_tests.
|
|
This changes various tests to use "require" for the Python feature.
|
|
This changes some tests to use "require !gdb_debug_enabled".
|
|
This changes some tests to use "require isnative".
|
|
This changes some tests to use "require can_spawn_for_attach".
|
|
This changes some tests to use "require !use_gdb_stub".
|
|
This changes some tests to use "require !skip_hw_watchpoint_tests".
|
|
This changes some tests to use "require !skip_shlib_tests".
|
|
Fix regexp in gdb.threads/dlopen-libpthread.exp:
'libpthread\\.so' -> '/libpthread\\.so'.
Tested on x86_64-linux.
|
|
On an x86_64 laptop running ubuntu 22.04.1 with unity desktop:
...
$ echo $XDG_CURRENT_DESKTOP
Unity:Unity7:ubuntu
...
I have:
...
$ echo $LD_PRELOAD
libgtk3-nocsd.so.0
...
due to package gtk3-nocsd, a package recommended by unity-session.
Consequently, for each exec these dependencies are pulled in, including
libpthread.so.0:
...
$ lddtree /lib/x86_64-linux-gnu/libgtk3-nocsd.so.0
libgtk3-nocsd.so.0 => /lib/x86_64-linux-gnu/libgtk3-nocsd.so.0 (interpreter => none)
libdl.so.2 => /lib/x86_64-linux-gnu/libdl.so.2
libpthread.so.0 => /lib/x86_64-linux-gnu/libpthread.so.0
libc.so.6 => /lib/x86_64-linux-gnu/libc.so.6
ld-linux-x86-64.so.2 => /lib/x86_64-linux-gnu/ld-linux-x86-64.so.2
...
So, while test-case gdb.threads/dlopen-libpthread.exp appears to run ok:
...
# of expected passes 12
# of unsupported tests 1
...
with LD_PRELOAD="" we have instead:
...
(gdb) PASS: gdb.threads/dlopen-libpthread.exp: continue to breakpoint: notify
info sharedlibrary^M
From To Syms Read Shared Object Library^M
$hex $hex Yes /lib64/ld-linux-x86-64.so.2^M
$hex $hex Yes /lib/x86_64-linux-gnu/libc.so.6^M
$hex $hex Yes dlopen-libpthread.so^M
(gdb) FAIL: gdb.threads/dlopen-libpthread.exp: libpthread.so found
...
The problem is that libpthread is expected as dependency of
dlopen-libpthread.so, but it's missing:
...
$ lddtree dlopen-libpthread.so
dlopen-libpthread.so => ./dlopen-libpthread.so (interpreter => none)
libc.so.6 => $outputs/gdb.threads/dlopen-libpthread/dlopen-libpthread.so.d/libc.so.6
ld-linux-x86-64.so.2 => /lib/x86_64-linux-gnu/ld-linux-x86-64.so.2
...
due to having glibc 2.35, which has libpthread integrated into libc.
Fix this by:
- adding a proc has_dependency
- using [has_dependency $exec libpthread.so] as hint that libpthread
may be preloaded
- using ![has_dependency $shlib libpthread.so] to detect that
the libpthread.so dependency is missing.
Also add a missing return after untested "no matching probes".
Tested on x86_64-linux, with and without LD_PRELOAD="".
|
|
This commit is the result of running the gdb/copyright.py script,
which automated the update of the copyright year range for all
source files managed by the GDB project to be updated to include
year 2023.
|
|
gdb.threads/watchpoint-fork.exp
On Ubuntu 22.04.1 x86_64, with glibc 2.35 I run into:
...
watchpoint-fork-mt.c: In function 'start':^M
watchpoint-fork-mt.c:67:7: warning: 'pthread_yield' is deprecated: \
pthread_yield is deprecated, use sched_yield instead \
[-Wdeprecated-declarations]^M
67 | i = pthread_yield ();^M
| ^^M
...
Fix this as suggested, by using sched_yield instead.
Tested on x86_64-linux.
|
|
On ubuntu 22.04.1 x86_64, I run into:
...
(gdb) info probes all rtld rtld_map_complete^M
No probes matched.^M
(gdb) XFAIL: gdb.threads/dlopen-libpthread.exp: info probes all rtld rtld_map_complete
UNTESTED: gdb.threads/dlopen-libpthread.exp: no matching probes
...
This has been filed as PR testsuite/17016.
The problem is that the name rtld_map_complete is used, which was only
available in Fedora 17, and upstream the name map_complete was used.
In the email thread discussing a proposed patch (
https://sourceware.org/legacy-ml/gdb-patches/2014-09/msg00712.html ) it was
suggested to make the test-case handle both names.
So, handle both names: map_complete and rtld_map_complete.
This exposes the following FAIL:
...
(gdb) info sharedlibrary^M
From To Syms Read Shared Object Library^M
$hex $hex Yes /lib64/ld-linux-x86-64.so.2^M
$hex $hex Yes (*) /lib/x86_64-linux-gnu/libgtk3-nocsd.so.0^M
$hex $hex Yes /lib/x86_64-linux-gnu/libc.so.6^M
$hex $hex Yes /lib/x86_64-linux-gnu/libdl.so.2^M
$hex $hex Yes /lib/x86_64-linux-gnu/libpthread.so.0^M
(*): Shared library is missing debugging information.^M
(gdb) FAIL: gdb.threads/dlopen-libpthread.exp: libpthread.so not found
...
due to using a glibc (v2.35) that has libpthread integrated into libc.
Fix this by changing the FAIL into UNSUPPORTED.
Tested on x86_64-linux.
Bug: https://sourceware.org/bugzilla/show_bug.cgi?id=17016
|
|
Once in a while I run into:
...
FAIL: gdb.threads/detach-step-over.exp: \
breakpoint-condition-evaluation=host: target-non-stop=off: non-stop=off: \
displaced=off: iter 1: all threads running
...
In can easily reproduce this by doing:
...
# Wait a bit, to give time for the threads to hit the
# breakpoint.
- sleep 1
return true
...
Fix this by counting the running threads in a loop, effectively allowing 10
seconds (instead of 1) for the threads to start running, but only sleeping if
needed.
Reduces total execution time from 1m27s to 56s.
Tested on x86_64-linux.
|
|
The canonical form of 'if' in modern TCL is 'if {} {}'. But there's
still a bunch of places in the testsuite where we make use of the
'then' keyword, and sometimes these get copies into new tests, which
just spreads poor practice.
This commit removes all use of the 'then' keyword from the gdb.threads/
test script directory.
There should be no changes in what is tested after this commit.
|
|
This commit addresses one of the issues identified in PR gdb/28275.
Bug gdb/28275 identifies a number of situations in which this assert:
Assertion `!proc_target->commit_resumed_state' failed.
could be triggered. There's actually a number of similar places where
this assert is found in GDB, the two of interest in gdb/28275 are in
target_wait and target_stop.
In one of the comments:
https://sourceware.org/bugzilla/show_bug.cgi?id=28275#c1
steps to trigger the assertion within target_stop were identified when
using a modified version of the gdb.threads/detach-step-over.exp test
script.
In the gdb.threads/detach-step-over.exp test, we attach to a
multi-threaded inferior, and continue the inferior in asynchronous
(background) mode. Each thread is continuously hitting a conditional
breakpoint where the condition is always false. While the inferior is
running we detach. The goal is that we detach while GDB is performing a
step-over for the conditional breakpoint in at least one thread.
While detaching, if a step-over is in progress, then GDB has to
complete the step over before we can detach. This involves calling
target_stop and target_wait (see prepare_for_detach).
As far as gdb/28275 is concerned, the interesting part here, is the
the process_stratum_target::commit_resumed_state variable must be
false when target_stop and target_wait are called.
This is currently ensured because, in detach_command (infrun.c), we
create an instance of scoped_disable_commit_resumed, this ensures that
when target_detach is called, ::commit_resumed_state will be false.
The modification to the test that I propose here, and which exposed
the bug, is that, instead of using "detach" to detach from the
inferior, we instead use "quit". Quitting GDB after attaching to an
inferior will cause GDB to first detach, and then exit.
When we quit GDB we end up calling target_detach via a different code
path, the stack looks like:
#0 target_detach
#1 kill_or_detach
#2 quit_force
#3 quit_command
Along this path there is no scoped_disable_commit_resumed created.
::commit_resumed_state can be true when we reach prepare_for_detach,
which calls target_wait and target_stop, so the assertion will trigger.
In this commit, I propose fixing this by adding the creation of a
scoped_disable_commit_resumed into target_detach. This will ensure
that ::commit_resumed_state is false when calling prepare_for_detach
from within target_detach.
I did consider placing the scoped_disable_commit_resumed in
prepare_for_detach, however, placing it in target_detach ensures that
the target's commit_resumed_state flag is left to false if the detached
inferior was the last one for that target. It's the same rationale as
for patch "gdb: disable commit resumed in target_kill" that comes later
in this series, but for detach instead of kill.
detach_command still includes a scoped_disable_commit_resumed too, but I
think it is still relevant to cover the resumption at the end of the
function.
Co-Authored-By: Simon Marchi <simon.marchi@efficios.com>
Bug: https://sourceware.org/bugzilla/show_bug.cgi?id=28275
Change-Id: Ie128f7aba6ef0e018859275eca372e6ea738e96f
|
|
Factor out some bits of gdb.threads/detach-step-over.exp to procs in
preparation to adding some new variations of the test. Rename the
existing "test" proc and make it use proc_with_prefix.
Co-Authored-By: Simon Marchi <simon.marchi@efficios.com>
Change-Id: Ib4412545c81c8556029e0f7bfa9dd48d7a9f3189
|
|
Before doing further changes to this file, change to use the :: notation
instead of declaring global variables with the `global` keyword.
Change-Id: I72301fd8f4693fea61aac054ba17245a1f4442fb
Approved-By: Andrew Burgess <aburgess@redhat.com>
|
|
I noticed a plain gcore command in test-case gdb.threads/gcore-thread.exp:
...
gdb_test "gcore $core0file" "Saved corefile .*" \
"save a zeroed-threads corefile"
...
Use gdb_gcore_cmd instead.
Tested on x86_64-linux.
|
|
The hardware watchpoint check has been updated in a couple of recent
patches. This patch updates the hardware watchpoint test in the remaining
gdb tests.
The issue is the PowerPC processors support hardware watchpoints with the
exception of Power 9. The hardware watchpoint support is disabled on
Power 9. The test skip_hw_watchpoint_tests must be used to correctly
determine if the PowerPC processor supports hardware watchpoints.
This patch fixes 6 test failures in test gdb.threads/watchpoint-fork.exp.
Test gdb.base/watch-vfork.exp runs with can-use-hw-watchpoints set to
true and false. When the test is run with can-use-hw-watchpoints set to
true, gdb just falls back to using software watchpoints. The
patch reduces the number of expected passes by 2 since because it now
only runs once with can-use-hw-watchpoints set to false.
Test gdb.mi/mi-watch.exp runs the test with argument hw and sw. If the
argument is hw and hardware watchpoints are not supported the test exits.
The number of expected passes is cut in half with the patch as it now only
runs the test using software breakpoints. Previously the pass to use
hardware watchpoints was not skipped and the test actually ran using
software watchpoints.
The following tests run the same with and without the patch. The tests
are supposed to execute the gdb command "set can-use-hw-watchpoints 0" if
the processor does not support hardware bwatchpoints. However the command
was not being executed and gdb was falling back to using software
watchpoints since the Power 9 watchpoint resource check fails. With the
patch, the tests now execute the command and the test runs using software
watchpoints as it did previously. The tests are:
gdb.base/commands.exp
gdb.base/cond-eval-mode.exp
gdb.base/display.exp
gdb.base/gdb11531.exp
gdb.base/recurse.exp
gdb.base/value-double-free.exp
gdb.base/watch-bitfields.exp
gdb.base/watch-cond-infcall.exp
gdb.base/watch-cond.exp
gdb.base/watchpoint-solib.exp
gdb.base/watchpoints.exp
The following two tests are not supported on the Power 9 system used to
test the changes. The patch does not change the tests results for these
tests:
gdb.python/py-breakpoint.exp
gdb.mi/mi-watch-nonstop.exp
|
|
With test-case gdb.base/infoline-reloc-main-from-zero.exp and clang I run into:
...
gdb compile failed, clang-13.0: warning: -e main: 'linker' input unused \
[-Wunused-command-line-argument]
clang-13.0: warning: -Wl,-Ttext=0x00: 'linker' input unused \
[-Wunused-command-line-argument]
clang-13.0: warning: -Wl,-N: 'linker' input unused \
[-Wunused-command-line-argument]
UNTESTED: gdb.base/infoline-reloc-main-from-zero.exp: \
infoline-reloc-main-from-zero.exp
UNTESTED: gdb.base/infoline-reloc-main-from-zero.exp: failed to compile
...
Fix this by using ldflags instead of additional_flags.
Likewise, fix all occurrences of:
...
$ find gdb/testsuite -name *.exp | xargs grep additional_flags.*Wl
...
Tested on x86_64-linux.
|
|
On ubuntu 22.04 with the libc6-dbg package installed, I have the
following failure:
where
#0 print_philosopher (n=3, left=33 '!', right=33 '!') at .../gdb/testsuite/gdb.threads/linux-dp.c:105
#1 0x000055555555576a in philosopher (data=0x55555555937c) at .../gdb/testsuite/gdb.threads/linux-dp.c:148
#2 0x00007ffff7e11b43 in start_thread (arg=<optimized out>) at ./nptl/pthread_create.c:442
#3 0x00007ffff7ea3a00 in clone3 () at ../sysdeps/unix/sysv/linux/x86_64/clone3.S:81
(gdb) FAIL: gdb.threads/linux-dp.exp: first thread-specific breakpoint hit
The regex for this test accounts for different situations (with /
without debug symbol) but assumes that if debug info is present the
backtrace shows execution under pthread_create. However, for the
implementation under test, we are under start_thread.
Update the regex to accept start_thread.
Tested on Ubuntu-22.04 x86_64 with and without libc6-dbg debug symbols
available.
Change-Id: I1e1536279890bca2cd07f038e026b41e46af44e0
|
|
In next-fork-other-thread.c, there's this loop:
...
do
{
ret = waitpid (pid, &stat, 0);
} while (ret == EINTR);
...
The loop condition tests for "ret == EINTR" but waitpid signals EINTR by
returning -1 and setting errno to EINTR.
Fix this by changing the loop condition to "ret == -1 && errno == EINTR".
Tested on x86_64-linux.
|
|
On openSUSE Tumbleweed (using glibc 2.36), I run into:
...
(gdb) print /d (int) munmap (4198400, 4096)^M
Invalid cast.^M
(gdb) FAIL: gdb.base/break-main-file-remove-fail.exp: cmdline: \
get integer valueof "(int) munmap (4198400, 4096)"
...
The problem is that after starting the executable, the symbol has type
"void (*) (void)":
...
(gdb) p munmap
$1 = {<text variable, no debug info>} 0x401030 <munmap@plt>
(gdb) start
...
(gdb) p munmap
$2 = {void (void)} 0x7ffff7feb9a0 <__GI_munmap>
...
which causes the "Invalid cast" error.
Looking at the debug info for glibc for symbol __GI_munmap:
...
<0><189683>: Abbrev Number: 1 (DW_TAG_compile_unit)
<189691> DW_AT_name : ../sysdeps/unix/syscall-template.S
<189699> DW_AT_producer : GNU AS 2.39.0
<1><1896ae>: Abbrev Number: 2 (DW_TAG_subprogram)
<1896af> DW_AT_name : __GI___munmap
<1896b3> DW_AT_external : 1
<1896b4> DW_AT_low_pc : 0x10cad0
<1896bc> DW_AT_high_pc : 37
...
that's probably caused by this bit (or similar bits for other munmap aliases).
This is fixed in gas on trunk by commit 5578fbf672e ("GAS: Add a return type
tag to DWARF DIEs generated for function symbols").
Work around this (for say gas 2.39) by explicitly specifying the prototype for
munmap.
Likewise for getpid in a couple of other test-cases.
Tested on x86_64-linux.
|
|
On aarch64 (and likewise on arm), I run into:
...
(gdb) PASS: gdb.threads/killed-outside.exp: get pid of inferior
Executing on target: kill -9 11516 (timeout = 300)
builtin_spawn -ignore SIGHUP kill -9 11516^M
continue^M
Continuing.^M
Unable to fetch general registers: No such process.^M
(gdb) [Thread 0xfffff7d511e0 (LWP 11518) exited]^M
^M
Program terminated with signal SIGKILL, Killed.^M
The program no longer exists.^M
FAIL: gdb.threads/killed-outside.exp: prompt after first continue (timeout)
...
due to a mismatch between the actual "No such process" line and the expected
one:
...
set no_such_process_msg "Couldn't get registers: No such process\."
...
Fix this by updating the regexp.
Tested on aarch64-linux, and x86_64-linux.
|
|
In all-stop mode, when the target is itself in non-stop mode (like
GNU/Linux), if you use the "step N" (or "stepi/next/nexti N") to step
a thread a number of times:
(gdb) help step
step, s
Step program until it reaches a different source line.
Usage: step [N]
Argument N means step N times (or till program stops for another reason).
... GDB prematurely stops all threads after the first step, and
doesn't re-resume them for the subsequent N-1 steps. It's as if for
the 2nd and subsequent steps, the command was running with
scheduler-locking enabled.
This can be observed with the testcase added by this commit, which
looks like this:
static pthread_barrier_t barrier;
static void *
thread_func (void *arg)
{
pthread_barrier_wait (&barrier);
return NULL;
}
int
main ()
{
pthread_t thread;
int ret;
pthread_barrier_init (&barrier, NULL, 2);
/* We run to this line below, and then issue "next 3". That should
step over the 3 lines below and land on the return statement. If
GDB prematurely stops the thread_func thread after the first of
the 3 nexts (and never resumes it again), then the join won't
ever return. */
pthread_create (&thread, NULL, thread_func, NULL); /* set break here */
pthread_barrier_wait (&barrier);
pthread_join (thread, NULL);
return 0;
}
The test hangs and times out without the GDB fix:
(gdb) next 3
[New Thread 0x7ffff7d89700 (LWP 525772)]
FAIL: gdb.threads/step-N-all-progress.exp: non-stop=off: target-non-stop=on: next 3 (timeout)
The problem is a core gdb bug.
When you do "step/stepi/next/nexti N", GDB internally creates a
thread_fsm object and associates it with the stepping thread. For the
stepping commands, the FSM's class is step_command_fsm. That object
is what keeps track of how many steps are left to make. When one step
finishes, handle_inferior_event calls stop_waiting and returns, and
then fetch_inferior_event calls the "should_stop" method of the event
thread's FSM. The implementation of that method decrements the
steps-left counter. If the counter is 0, it returns true and we
proceed to presenting the stop to the user. If it isn't 0 yet, then
the method returns false, indicating to fetch_inferior_event to "keep
going".
Focusing now on when the first step finishes -- we're in "all-stop"
mode, with the target in non-stop mode. When a step finishes,
handle_inferior_event calls stop_waiting, which itself calls
stop_all_threads to stop everything. I.e., after the first step
completes, all threads are stopped, before handle_inferior_event
returns. And after that, now in fetch_inferior_event, we consult the
thread's thread_fsm::should_stop, which as we've seen, for the first
step returns false -- i.e., we need to keep_going for another step.
However, since the target is in non-stop mode, keep_going resumes
_only_ the current thread. All the other threads remain stopped,
inadvertently.
If the target is in non-stop mode, we don't actually need to stop all
threads right after each step finishes, and then re-resume them again.
We can instead defer stopping all threads until all the steps are
completed.
So fix this by delaying the stopping of all threads until after we
called the FSM's "should_stop" method. I.e., move it from
stop_waiting, to handle_inferior_events's callers,
fetch_inferior_event and wait_for_inferior.
New test included. Tested on x86-64 GNU/Linux native and gdbserver.
Change-Id: Iaad50dcfea4464c84bdbac853a89df92ade6ae01
|
|
A WIP version of a patch
(https://sourceware.org/pipermail/gdb-patches/2022-June/190202.html)
resulted in a bug that went unnoticed by the testuite, like so:
(gdb) PASS: gdb.threads/no-unwaited-for-left.exp: enable scheduler-locking, for main thread
continue
Continuing.
[New Thread 1251861.1251861]
No unwaited-for children left.
(gdb) PASS: gdb.threads/no-unwaited-for-left.exp: continue stops when the main thread exits
info threads
Id Target Id Frame
3 Thread 1251861.1251863 "no-unwaited-for" __pthread_clockjoin_ex (threadid=140737351558976, thread_return=0x0, clockid=<optimized out>, abstime=<optimized out>, block=<optimized out>) at pthread_join_common.c:145
4 Thread 1251861.1251861 "no-unwaited-for" <unavailable> in ?? ()
The current thread <Thread ID 1> has terminated. See `help thread'.
(gdb) PASS: gdb.threads/no-unwaited-for-left.exp: only thread 3 left, main thread terminated
Somehow, above, GDB re-added the zombie leader back before printing
"No unwaited-for children left.". The "only thread 3 left, main
thread terminated" test should have caught this, but didn't. That is
because the test's regexp has a ".*" after the part that matches
thread 3. This commit tightens that regexp to catch such a bug. It
also tightens the "only main thread left, thread 2 terminated" test's
regexp in the same way.
Change-Id: I8744f327a0aa0e2669d1ddda88247e99b91cefff
|
|
When building gdb with -fsanitize=address we run into:
...
builtin_spawn gdb -nw -nx -iex set height 0 -iex set width 0 -data-directory \
build/gdb/data-directory^M
==10637==ASan runtime does not come first in initial library list; you \
should either link runtime to your application or manually preload it with \
LD_PRELOAD.^M
ERROR: GDB process no longer exists
...
Prevent the ASan runtime error by using
ASAN_OPTIONS=verify_asan_link_order=0. This makes both test-cases pass.
Tested on x86_64-linux.
Bug: https://sourceware.org/bugzilla/show_bug.cgi?id=29358
|
|
It is not necessary to call get_compiler_info before calling
test_compiler_info, and, after recent commits that removed setting up
the gcc_compiled, true, and false globals from get_compiler_info,
there is now no longer any need for any test script to call
get_compiler_info directly.
As a result every call to get_compiler_info outside of lib/gdb.exp is
redundant, and this commit removes them all.
There should be no change in what is tested after this commit.
|
|
After this commit the gcc_compiled global is no longer exported from
lib/gdb.exp. In theory we could switch over all uses of gcc_compiled
to instead call test_compiler_info directly, however, I have instead
added a new proc to gdb.exp: 'is_c_compiler_gcc'. I've then updated
the testsuite to call this proc instead of using the global.
Having a new proc specifically for this task means that we have a
single consistent pattern for detecting gcc. By wrapping this logic
within a proc that calls test_compiler_info, rather than using the
global, means that test scripts don't need to call get_compiler_info
before they read the global, simply calling the new proc does
everything in one go.
As a result I've been able to remove the get_compiler_info calls from
all the test scripts that I've touched in this commit.
In some of the tests e.g. gdb.dwarf2/*.exp, the $gcc_compiled flag was
being checked at the top of the script to decide if the whole script
should be skipped or not. In these cases I've called the new proc
directly and removed all uses of gcc_compiled.
In other cases, e.g. most of the gdb.base scripts, there were many
uses of gcc_compiled. In these cases I set a new global gcc_compiled
near the top of the script, and leave the rest of the script
unchanged.
There should be no changes in what is tested after this commit.
|
|
While running the gdb.threads/tls.exp test with a GDB configured
without Python, I noticed some duplicate test names.
This is caused by a call to skip_python_tests that is within a proc
that is called multiple times by the test script. Each call to
skip_python_tests results in a call to 'unsupported', and this causes
the duplicate test names.
After this commit we now call skip_python_tests just once and place
the result into a variable. Now, instead of calling skip_python_tests
multiple times, we just check the variable.
There should be no change in what is tested after this commit.
|
