[PATCH+doc] Fix PR threads/19422 - show which thread caused stop
Commit Message
On 01/04/2016 11:30 PM, Pedro Alves wrote:
>
> I'll work on documentation next if this looks agreeable.
>
I've added a short blurb to NEWS and manual now. Eli, how does that look?
From: Pedro Alves <palves@redhat.com>
Subject: [PATCH] Fix PR threads/19422 - show which thread caused stop
This commit changes GDB like this:
- Program received signal SIGINT, Interrupt.
+ Thread 1 "main" received signal SIGINT, Interrupt.
- Breakpoint 1 at 0x40087a: file threads.c, line 87.
+ Thread 3 "bar" hit Breakpoint 1 at 0x40087a: file threads.c, line 87.
... once the program goes multi-threaded. Until GDB sees a second
thread spawn, the output is still the same as before, per the
discussion back in 2012:
https://www.sourceware.org/ml/gdb/2012-11/msg00010.html
This helps non-stop mode, where you can't easily tell which thread hit
a breakpoint or received a signal:
(gdb) info threads
Id Target Id Frame
* 1 Thread 0x7ffff7fc1740 (LWP 19362) "main" (running)
2 Thread 0x7ffff7fc0700 (LWP 19366) "foo" (running)
3 Thread 0x7ffff77bf700 (LWP 19367) "bar" (running)
(gdb)
Program received signal SIGUSR1, User defined signal 1.
0x0000003616a09237 in pthread_join (threadid=140737353877248, thread_return=0x7fffffffd5b8) at pthread_join.c:92
92 lll_wait_tid (pd->tid);
(gdb) b threads.c:87
Breakpoint 1 at 0x40087a: file threads.c, line 87.
(gdb)
Breakpoint 1, thread_function1 (arg=0x1) at threads.c:87
87 usleep (1); /* Loop increment. */
The best the user can do is run "info threads" and try to figure
things out.
It actually also affects all-stop mode, in case of "handle SIG print
nostop":
...
Program received signal SIGUSR1, User defined signal 1.
Program received signal SIGUSR1, User defined signal 1.
Program received signal SIGUSR1, User defined signal 1.
Program received signal SIGUSR1, User defined signal 1.
...
The above doesn't give any clue that these were different threads
getting the SIGUSR1 signal.
I initially thought of lowercasing "breakpoint" in
"Thread 3 hit Breakpoint 1"
but then after trying it I realized that leaving "Breakpoint"
uppercase helps the eye quickly find the relevant information. It's
also easier to implement not showing anything about threads until the
program goes multi-threaded this way.
Here's a larger example session in non-stop mode:
(gdb) c -a&
Continuing.
(gdb) interrupt -a
(gdb)
Thread 1 "main" stopped.
0x0000003616a09237 in pthread_join (threadid=140737353877248, thread_return=0x7fffffffd5b8) at pthread_join.c:92
92 lll_wait_tid (pd->tid);
Thread 2 "foo" stopped.
0x0000003615ebc6ed in nanosleep () at ../sysdeps/unix/syscall-template.S:81
81 T_PSEUDO (SYSCALL_SYMBOL, SYSCALL_NAME, SYSCALL_NARGS)
Thread 3 "bar" stopped.
0x0000003615ebc6ed in nanosleep () at ../sysdeps/unix/syscall-template.S:81
81 T_PSEUDO (SYSCALL_SYMBOL, SYSCALL_NAME, SYSCALL_NARGS)
b threads.c:87
Breakpoint 4 at 0x40087a: file threads.c, line 87.
(gdb) b threads.c:67
Breakpoint 5 at 0x400811: file threads.c, line 67.
(gdb) c -a&
Continuing.
(gdb)
Thread 3 "bar" hit Breakpoint 4, thread_function1 (arg=0x1) at threads.c:87
87 usleep (1); /* Loop increment. */
Thread 2 "foo" hit Breakpoint 5, thread_function0 (arg=0x0) at threads.c:68
68 (*myp) ++;
info threads
Id Target Id Frame
* 1 Thread 0x7ffff7fc1740 (LWP 31957) "main" (running)
2 Thread 0x7ffff7fc0700 (LWP 31961) "foo" thread_function0 (arg=0x0) at threads.c:68
3 Thread 0x7ffff77bf700 (LWP 31962) "bar" thread_function1 (arg=0x1) at threads.c:87
(gdb) shell kill -SIGINT 31957
(gdb)
Thread 1 "main" received signal SIGINT, Interrupt.
0x0000003616a09237 in pthread_join (threadid=140737353877248, thread_return=0x7fffffffd5b8) at pthread_join.c:92
92 lll_wait_tid (pd->tid);
info threads
Id Target Id Frame
* 1 Thread 0x7ffff7fc1740 (LWP 31957) "main" 0x0000003616a09237 in pthread_join (threadid=140737353877248, thread_return=0x7fffffffd5b8) at pthread_join.c:92
2 Thread 0x7ffff7fc0700 (LWP 31961) "foo" thread_function0 (arg=0x0) at threads.c:68
3 Thread 0x7ffff77bf700 (LWP 31962) "bar" thread_function1 (arg=0x1) at threads.c:87
(gdb) t 2
[Switching to thread 2, Thread 0x7ffff7fc0700 (LWP 31961)]
#0 thread_function0 (arg=0x0) at threads.c:68
68 (*myp) ++;
(gdb) catch syscall
Catchpoint 6 (any syscall)
(gdb) c&
Continuing.
(gdb)
Thread 2 "foo" hit Catchpoint 6 (call to syscall nanosleep), 0x0000003615ebc6ed in nanosleep () at ../sysdeps/unix/syscall-template.S:81
81 T_PSEUDO (SYSCALL_SYMBOL, SYSCALL_NAME, SYSCALL_NARGS)
I'll work on documentation next if this looks agreeable.
This patch applies on top of the star wildcards thread IDs series:
https://sourceware.org/ml/gdb-patches/2016-01/msg00291.html
For convenience, I've pushed this to the
users/palves/show-which-thread-caused-stop branch.
gdb/doc/ChangeLog:
2016-01-14 Pedro Alves <palves@redhat.com>
* gdb.texinfo (Threads): Mention that GDB displays the ID and name
of the thread that hit a breakpoint or received a signal.
gdb/ChangeLog:
2016-01-14 Pedro Alves <palves@redhat.com>
* NEWS: Mention that GDB now displays the ID and name of the
thread that hit a breakpoint or received a signal.
* break-catch-sig.c (signal_catchpoint_print_it): Use
maybe_print_thread_hit_breakpoint.
* break-catch-syscall.c (print_it_catch_syscall): Likewise.
* break-catch-throw.c (print_it_exception_catchpoint): Likewise.
* breakpoint.c (maybe_print_thread_hit_breakpoint): New function.
(print_it_catch_fork, print_it_catch_vfork, print_it_catch_solib)
(print_it_catch_exec, print_it_ranged_breakpoint)
(print_it_watchpoint, print_it_masked_watchpoint, bkpt_print_it):
Use maybe_print_thread_hit_breakpoint.
* breakpoint.h (maybe_print_thread_hit_breakpoint): Declare.
* gdbthread.h (show_thread_that_caused_stop): Declare.
* infrun.c (print_signal_received_reason): Print which thread
received signal.
* thread.c (show_thread_that_caused_stop): New function.
gdb/testsuite/ChangeLog:
2016-01-14 Pedro Alves <palves@redhat.com>
* gdb.base/async-shell.exp: Adjust expected output.
* gdb.base/dprintf-non-stop.exp: Adjust expected output.
* gdb.base/siginfo-thread.exp: Adjust expected output.
* gdb.base/watchpoint-hw-hit-once.exp: Adjust expected output.
* gdb.java/jnpe.exp: Adjust expected output.
* gdb.threads/clone-new-thread-event.exp: Adjust expected output.
* gdb.threads/continue-pending-status.exp: Adjust expected output.
* gdb.threads/leader-exit.exp: Adjust expected output.
* gdb.threads/manythreads.exp: Adjust expected output.
* gdb.threads/pthreads.exp: Adjust expected output.
* gdb.threads/schedlock.exp: Adjust expected output.
* gdb.threads/siginfo-threads.exp: Adjust expected output.
* gdb.threads/signal-command-multiple-signals-pending.exp: Adjust
expected output.
* gdb.threads/signal-delivered-right-thread.exp: Adjust expected
output.
* gdb.threads/sigthread.exp: Adjust expected output.
* gdb.threads/watchpoint-fork.exp: Adjust expected output.
---
gdb/doc/gdb.texinfo | 14 +++++
gdb/NEWS | 7 +++
gdb/break-catch-sig.c | 4 +-
gdb/break-catch-syscall.c | 5 +-
gdb/break-catch-throw.c | 1 +
gdb/breakpoint.c | 73 +++++++++++++++++-----
gdb/breakpoint.h | 4 ++
gdb/gdbthread.h | 5 ++
gdb/infrun.c | 30 ++++++---
gdb/testsuite/gdb.base/async-shell.exp | 2 +-
gdb/testsuite/gdb.base/dprintf-non-stop.exp | 2 +-
gdb/testsuite/gdb.base/siginfo-thread.exp | 2 +-
gdb/testsuite/gdb.base/watchpoint-hw-hit-once.exp | 2 +-
gdb/testsuite/gdb.java/jnpe.exp | 2 +-
.../gdb.threads/clone-new-thread-event.exp | 2 +-
.../gdb.threads/continue-pending-status.exp | 2 +-
gdb/testsuite/gdb.threads/leader-exit.exp | 2 +-
gdb/testsuite/gdb.threads/manythreads.exp | 8 +--
gdb/testsuite/gdb.threads/pthreads.exp | 2 +-
gdb/testsuite/gdb.threads/schedlock.exp | 2 +-
gdb/testsuite/gdb.threads/siginfo-threads.exp | 2 +-
.../signal-command-multiple-signals-pending.exp | 4 +-
.../gdb.threads/signal-delivered-right-thread.exp | 2 +-
gdb/testsuite/gdb.threads/sigthread.exp | 2 +-
gdb/testsuite/gdb.threads/watchpoint-fork.exp | 4 +-
gdb/thread.c | 8 +++
26 files changed, 146 insertions(+), 47 deletions(-)
Comments
> Date: Thu, 14 Jan 2016 14:08:40 +0000
> From: Pedro Alves <palves@redhat.com>
>
> On 01/04/2016 11:30 PM, Pedro Alves wrote:
> >
> > I'll work on documentation next if this looks agreeable.
> >
>
> I've added a short blurb to NEWS and manual now. Eli, how does that look?
Looks good, thanks.
> From: Pedro Alves <palves@redhat.com>
> Subject: [PATCH] Fix PR threads/19422 - show which thread caused stop
>
> This commit changes GDB like this:
>
> - Program received signal SIGINT, Interrupt.
> + Thread 1 "main" received signal SIGINT, Interrupt.
>
> - Breakpoint 1 at 0x40087a: file threads.c, line 87.
> + Thread 3 "bar" hit Breakpoint 1 at 0x40087a: file threads.c, line 87.
Would it make sense to lose the "hit" part, and have this say
Thread 3 "bar": breakpoint 1 at 0x40087a: file threads.c, line 87.
(Note that I also avoided capitalizing "breakpoint".)
WDYT?
On 01/14/2016 04:36 PM, Eli Zaretskii wrote:
>
>> From: Pedro Alves <palves@redhat.com>
>> Subject: [PATCH] Fix PR threads/19422 - show which thread caused stop
>>
>> This commit changes GDB like this:
>>
>> - Program received signal SIGINT, Interrupt.
>> + Thread 1 "main" received signal SIGINT, Interrupt.
>>
>> - Breakpoint 1 at 0x40087a: file threads.c, line 87.
>> + Thread 3 "bar" hit Breakpoint 1 at 0x40087a: file threads.c, line 87.
>
> Would it make sense to lose the "hit" part, and have this say
>
> Thread 3 "bar": breakpoint 1 at 0x40087a: file threads.c, line 87.
>
Not sure. I kind of got used to how it was. Kind of the
counterpart of being explicit in saying "received", in the signal
case. If going that direction, I guess you'd also want:
Thread 1 "main": received signal SIGINT, Interrupt.
Thread 1 "main": signal SIGINT, Interrupt.
doesn't look right to me, for some reason. But it may just
be habituation.
> (Note that I also avoided capitalizing "breakpoint".)
>
Yeah, I initially thought of doing that as well, but then at least
with "hit", uppercase makes it easier to parse the message. At
least for me.
The second reason I didn't lower case is that it's easier to
implement not showing anything about threads until the program
goes multi-threaded this way, both in gdb and in the testsuite.
Otherwise we need extra logic.
The third reason I prefer the way it is, is that it's already
implemented this way. :-)
Thanks,
Pedro Alves
> Date: Thu, 14 Jan 2016 17:12:42 +0000
> From: Pedro Alves <palves@redhat.com>
> CC: gdb-patches@sourceware.org
>
> >> - Program received signal SIGINT, Interrupt.
> >> + Thread 1 "main" received signal SIGINT, Interrupt.
> >>
> >> - Breakpoint 1 at 0x40087a: file threads.c, line 87.
> >> + Thread 3 "bar" hit Breakpoint 1 at 0x40087a: file threads.c, line 87.
> >
> > Would it make sense to lose the "hit" part, and have this say
> >
> > Thread 3 "bar": breakpoint 1 at 0x40087a: file threads.c, line 87.
> >
>
> Not sure. I kind of got used to how it was. Kind of the
> counterpart of being explicit in saying "received", in the signal
> case. If going that direction, I guess you'd also want:
>
> Thread 1 "main": received signal SIGINT, Interrupt.
> Thread 1 "main": signal SIGINT, Interrupt.
No: we already announce signals with "Program received signal". But
with breakpoints, we just say "Breakpoint 1", not "Program hit
breakpoint 1".
Besides, "hit a breakpoint" is jargon, which is another reason I
wanted to get rid of it.
Thanks.
On 01/14/2016 06:25 PM, Eli Zaretskii wrote:
>> Date: Thu, 14 Jan 2016 17:12:42 +0000
>> From: Pedro Alves <palves@redhat.com>
>> CC: gdb-patches@sourceware.org
>>
>>>> - Program received signal SIGINT, Interrupt.
>>>> + Thread 1 "main" received signal SIGINT, Interrupt.
>>>>
>>>> - Breakpoint 1 at 0x40087a: file threads.c, line 87.
>>>> + Thread 3 "bar" hit Breakpoint 1 at 0x40087a: file threads.c, line 87.
>>>
>>> Would it make sense to lose the "hit" part, and have this say
>>>
>>> Thread 3 "bar": breakpoint 1 at 0x40087a: file threads.c, line 87.
>>>
>>
>> Not sure. I kind of got used to how it was. Kind of the
>> counterpart of being explicit in saying "received", in the signal
>> case. If going that direction, I guess you'd also want:
>>
>> Thread 1 "main": received signal SIGINT, Interrupt.
>> Thread 1 "main": signal SIGINT, Interrupt.
>
> No: we already announce signals with "Program received signal". But
> with breakpoints, we just say "Breakpoint 1", not "Program hit
> breakpoint 1".
Sure. Following your suggestion ends up with:
Thread 1 "main": breakpoint 1
Thread 1 "main" received signal SIGINT, Interrupt.
which seems inconsistent to me. If you disregard how the
current/single-threaded output looks, it seems better to me
to be consistent at least when debugging multiple threads.
> Besides, "hit a breakpoint" is jargon, which is another reason I
> wanted to get rid of it.
What do you mean, jargon? GDB already uses the term:
(gdb) info breakpoints
Num Type Disp Enb Address What
1 breakpoint keep y 0x000000000040073e in main at threads.c:40
breakpoint already hit 1 time
^^^^^^^^^^
Thanks,
Pedro Alves
> Date: Thu, 14 Jan 2016 18:59:57 +0000
> From: Pedro Alves <palves@redhat.com>
> CC: gdb-patches@sourceware.org
>
> > No: we already announce signals with "Program received signal". But
> > with breakpoints, we just say "Breakpoint 1", not "Program hit
> > breakpoint 1".
>
> Sure. Following your suggestion ends up with:
>
> Thread 1 "main": breakpoint 1
> Thread 1 "main" received signal SIGINT, Interrupt.
>
> which seems inconsistent to me.
Do we really have to be consistent here? We weren't before your
change.
> > Besides, "hit a breakpoint" is jargon, which is another reason I
> > wanted to get rid of it.
>
> What do you mean, jargon?
"Hit a breakpoint" is jargon. We don't really "hit" anything. A
breakpoint breaks, or triggers.
> GDB already uses the term:
>
> (gdb) info breakpoints
> Num Type Disp Enb Address What
> 1 breakpoint keep y 0x000000000040073e in main at threads.c:40
> breakpoint already hit 1 time
> ^^^^^^^^^^
Yes, but that one is pretty much confined to its corner. The message
that announces a breakpoint is much more visible.
Anyway, we can agree to disagree. No big deal.
On 01/14/2016 07:06 PM, Eli Zaretskii wrote:
>> > Date: Thu, 14 Jan 2016 18:59:57 +0000
>> > From: Pedro Alves <palves@redhat.com>
>> > CC: gdb-patches@sourceware.org
>> >
>>> > > No: we already announce signals with "Program received signal". But
>>> > > with breakpoints, we just say "Breakpoint 1", not "Program hit
>>> > > breakpoint 1".
>> >
>> > Sure. Following your suggestion ends up with:
>> >
>> > Thread 1 "main": breakpoint 1
>> > Thread 1 "main" received signal SIGINT, Interrupt.
>> >
>> > which seems inconsistent to me.
> Do we really have to be consistent here? We weren't before your
> change.
>
>>> > > Besides, "hit a breakpoint" is jargon, which is another reason I
>>> > > wanted to get rid of it.
>> >
>> > What do you mean, jargon?
> "Hit a breakpoint" is jargon. We don't really "hit" anything. A
> breakpoint breaks, or triggers.
Nothing actually "breaks" either. :-) I think it's as much jargon
as the word "breakpoint" itself. IOW, it may not be understood by
someone not familiar with a debugger, but anyone familiar with using
a debugger understands it.
>
>> > GDB already uses the term:
>> >
>> > (gdb) info breakpoints
>> > Num Type Disp Enb Address What
>> > 1 breakpoint keep y 0x000000000040073e in main at threads.c:40
>> > breakpoint already hit 1 time
>> > ^^^^^^^^^^
> Yes, but that one is pretty much confined to its corner. The message
> that announces a breakpoint is much more visible.
>
> Anyway, we can agree to disagree. No big deal.
Yes, I think so.
Looking up online for documentation of several debuggers, tutorials, etc., I
see "breakpoint hit" mentioned all over the place. Some examples:
http://bashdb.sourceforge.net/bashdb.html
"There is also a notion of a “one-time” breakpoint which gets deleted as soon as
it is hit, so that that breakpoint is executed once only. "
http://lldb.llvm.org/lldb-gdb.html
"Run until we hit line 12 or control leaves the current function."
http://docs.roguewave.com/totalview/8.15.10/html/User_Guides/UsingthegSpecifier.html
"Thread 1.1 hit breakpoint 1 at line 35 in ".breakpoint_here""
https://msdn.microsoft.com/en-us/library/5557y8b4.aspx
"When you run this code in the debugger, execution stops whenever
the breakpoint is hit, before the code on that line is executed."
etc.
I really honestly believe that nobody will be confused.
Thanks,
Pedro Alves
On Thu, 14 Jan 2016 15:08:40 +0100, Pedro Alves wrote:
[...]
> * gdb.base/watchpoint-hw-hit-once.exp: Adjust expected output.
[...]
> diff --git a/gdb/testsuite/gdb.base/watchpoint-hw-hit-once.exp b/gdb/testsuite/gdb.base/watchpoint-hw-hit-once.exp
> index 5c1be00..46d5c62 100644
> --- a/gdb/testsuite/gdb.base/watchpoint-hw-hit-once.exp
> +++ b/gdb/testsuite/gdb.base/watchpoint-hw-hit-once.exp
> @@ -31,7 +31,7 @@ gdb_test "rwatch watchee"
>
> gdb_breakpoint [gdb_get_line_number "dummy = 2"]
>
> -gdb_test "continue" "Continuing.\r\nHardware read watchpoint \[0-9\]+: watchee\r\n\r\nValue = 0\r\n.*"
> +gdb_test "continue" "Continuing.\r\n\r\nHardware read watchpoint \[0-9\]+: watchee\r\n\r\nValue = 0\r\n.*"
I haven't seen a comment why such change is necessary. It probably happens
from some of the changes involved but not sure if it was really required.
Thanks,
Jan
On 01/22/2016 04:44 PM, Jan Kratochvil wrote:
> On Thu, 14 Jan 2016 15:08:40 +0100, Pedro Alves wrote:
> [...]
>> * gdb.base/watchpoint-hw-hit-once.exp: Adjust expected output.
> [...]
>> diff --git a/gdb/testsuite/gdb.base/watchpoint-hw-hit-once.exp b/gdb/testsuite/gdb.base/watchpoint-hw-hit-once.exp
>> index 5c1be00..46d5c62 100644
>> --- a/gdb/testsuite/gdb.base/watchpoint-hw-hit-once.exp
>> +++ b/gdb/testsuite/gdb.base/watchpoint-hw-hit-once.exp
>> @@ -31,7 +31,7 @@ gdb_test "rwatch watchee"
>>
>> gdb_breakpoint [gdb_get_line_number "dummy = 2"]
>>
>> -gdb_test "continue" "Continuing.\r\nHardware read watchpoint \[0-9\]+: watchee\r\n\r\nValue = 0\r\n.*"
>> +gdb_test "continue" "Continuing.\r\n\r\nHardware read watchpoint \[0-9\]+: watchee\r\n\r\nValue = 0\r\n.*"
>
> I haven't seen a comment why such change is necessary. It probably happens
> from some of the changes involved but not sure if it was really required.
Previously, all breakpoint hits had that extra line, except watchpoints.
The patch made every breakpoint kind call the same function to
print the breakpoint hit, which had the side effect of making that inconsistency
disappear. That is, before we had:
(gdb) c
Continuing.
Breakpoint 2, thread_function0 (arg=0x0) at /home/pedro/gdb/tests/threads.c:87
87 int_return ();
(gdb) c
Continuing.
Hardware watchpoint 4: args[i]
Old value = 0
New value = 1
main () at /home/pedro/gdb/tests/threads.c:57
57 res = pthread_create(&threads[i],
Now watchpoint hits are consistent with all other breakpoints in that aspect:
(gdb) c
Continuing.
Hardware watchpoint 2: args[i]
Old value = 0
New value = 1
main () at /home/pedro/gdb/tests/threads.c:57
57 res = pthread_create(&threads[i],
Thanks,
Pedro Alves
On Fri, 22 Jan 2016 17:55:07 +0100, Pedro Alves wrote:
> Now watchpoint hits are consistent with all other breakpoints in that aspect:
OK, that makes sense - so that it does not mess with the possible inferior
output.
Thanks for the explanation,
Jan
On Thu, 14 Jan 2016 15:08:40 +0100, Pedro Alves wrote:
> gdb/ChangeLog:
> 2016-01-14 Pedro Alves <palves@redhat.com>
>
> * NEWS: Mention that GDB now displays the ID and name of the
> thread that hit a breakpoint or received a signal.
> * break-catch-sig.c (signal_catchpoint_print_it): Use
> maybe_print_thread_hit_breakpoint.
> * break-catch-syscall.c (print_it_catch_syscall): Likewise.
> * break-catch-throw.c (print_it_exception_catchpoint): Likewise.
> * breakpoint.c (maybe_print_thread_hit_breakpoint): New function.
> (print_it_catch_fork, print_it_catch_vfork, print_it_catch_solib)
> (print_it_catch_exec, print_it_ranged_breakpoint)
> (print_it_watchpoint, print_it_masked_watchpoint, bkpt_print_it):
> Use maybe_print_thread_hit_breakpoint.
> * breakpoint.h (maybe_print_thread_hit_breakpoint): Declare.
> * gdbthread.h (show_thread_that_caused_stop): Declare.
> * infrun.c (print_signal_received_reason): Print which thread
> received signal.
> * thread.c (show_thread_that_caused_stop): New function.
There was already before a regression if --with-guile (which is default if
Guile is found) was used:
backtrace^M
#0 0x00007ffff6078da0 in __sigprocmask (how=2, set=0x7fffffffcc40, oset=0x0) at ../sysdeps/unix/sysv/linux/x86_64/sigprocmask.c:39^M
#1 0x0000000000966ce9 in _rl_handle_signal (sig=2) at signals.c:228^M
#2 0x0000000000966c05 in rl_signal_handler (sig=2) at signals.c:149^M
#3 <signal handler called>^M
#4 0x00007ffff613afc0 in __poll_nocancel () at ../sysdeps/unix/syscall-template.S:84^M
#5 0x00000000007d9a88 in gdb_wait_for_event (block=1) at event-loop.c:746^M
#6 0x00000000007d8e87 in gdb_do_one_event () at event-loop.c:323^M
#7 0x00000000007d8ed8 in start_event_loop () at event-loop.c:347^M
#8 0x00000000007da9dc in cli_command_loop (data=0x0) at event-top.c:186^M
#9 0x00000000007d0b4c in current_interp_command_loop () at interps.c:317^M
#10 0x00000000007d1f56 in captured_command_loop (data=0x0) at main.c:318^M
#11 0x00000000007cd6c9 in catch_errors (func=0x7d1f3b <captured_command_loop>, func_args=0x0, errstring=0x1167f75 "", mask=RETURN_MASK_ALL) at exceptions.c:240^M
#12 0x00000000007d3514 in captured_main (data=0x7fffffffd600) at main.c:1157^M
#13 0x00000000007cd6c9 in catch_errors (func=0x7d23d5 <captured_main>, func_args=0x7fffffffd600, errstring=0x1167f75 "", mask=RETURN_MASK_ALL) at exceptions.c:240^M
#14 0x00000000007d353d in gdb_main (args=0x7fffffffd600) at main.c:1165^M
#15 0x000000000049ae8c in main (argc=5, argv=0x7fffffffd708) at gdb.c:32^M
(gdb) PASS: gdb.gdb/selftest.exp: backtrace through signal handler
->
backtrace^M
#0 pthread_cond_wait@@GLIBC_2.3.2 () at ../sysdeps/unix/sysv/linux/x86_64/pthread_cond_wait.S:185^M
#1 0x00007ffff6db32b7 in GC_wait_marker () at pthread_support.c:2036^M
#2 0x00007ffff6da92ba in GC_help_marker (my_mark_no=my_mark_no@entry=4) at mark.c:1168^M
#3 0x00007ffff6db15ef in GC_mark_thread (id=<optimized out>) at pthread_support.c:389^M
#4 0x00007ffff6b8360a in start_thread (arg=0x7ffff337e700) at pthread_create.c:334^M
#5 0x00007ffff5847a4d in clone () at ../sysdeps/unix/sysv/linux/x86_64/clone.S:109^M
(gdb) FAIL: gdb.gdb/selftest.exp: backtrace through signal handler
Additionally this patchset added a new regression:
Program received signal SIGINT, Interrupt.^M
0x00007ffff613afc0 in __poll_nocancel () at ../sysdeps/unix/syscall-template.S:84^M
84 T_PSEUDO (SYSCALL_SYMBOL, SYSCALL_NAME, SYSCALL_NARGS)^M
(gdb) PASS: gdb.gdb/selftest.exp: send ^C to child process
->
Thread 1 "xgdb" received signal SIGINT, Interrupt.^M
0x00007ffff583bfdd in poll () at ../sysdeps/unix/syscall-template.S:84^M
84 T_PSEUDO (SYSCALL_SYMBOL, SYSCALL_NAME, SYSCALL_NARGS)^M
(gdb) FAIL: gdb.gdb/selftest.exp: send ^C to child process
OK to check-in the fix for both of these problems?
Tested on x86_64-fedora23-linux-gnu.
Thanks,
Jan
@@ -2969,6 +2969,20 @@ useful in writing breakpoint conditional expressions, command scripts,
and so forth. @xref{Convenience Vars,, Convenience Variables}, for
general information on convenience variables.
+If @value{GDBN} detects the program is multi-threaded, it augments the
+usual message about stopping at a breakpoint with the ID and name of
+the thread that hit the breakpoint.
+
+@smallexample
+Thread 2 "client" hit Breakpoint 1, send_message () at client.c:68
+@end smallexample
+
+Likewise when the program receives a signal:
+
+@smallexample
+Thread 1 "main" received signal SIGINT, Interrupt.
+@end smallexample
+
@table @code
@kindex info threads
@item info threads @r{[}@var{thread-id-list}@r{]}
@@ -49,6 +49,13 @@
* The new convenience variable $_inferior holds the number of the
current inferior.
+* GDB now displays the ID and name of the thread that hit a breakpoint
+ or received a signal, if your program is multi-threaded. For
+ example:
+
+ Thread 3 "bar" hit Breakpoint 1 at 0x40087a: file program.c, line 20.
+ Thread 1 "main" received signal SIGINT, Interrupt.
+
* Record btrace now supports non-stop mode.
* Support for tracepoints on aarch64-linux was added in GDBserver.
@@ -220,14 +220,16 @@ signal_catchpoint_print_it (bpstat bs)
ptid_t ptid;
struct target_waitstatus last;
const char *signal_name;
+ struct ui_out *uiout = current_uiout;
get_last_target_status (&ptid, &last);
signal_name = signal_to_name_or_int (last.value.sig);
annotate_catchpoint (b->number);
+ maybe_print_thread_hit_breakpoint (uiout);
- printf_filtered (_("\nCatchpoint %d (signal %s), "), b->number, signal_name);
+ printf_filtered (_("Catchpoint %d (signal %s), "), b->number, signal_name);
return PRINT_SRC_AND_LOC;
}
@@ -254,11 +254,12 @@ print_it_catch_syscall (bpstat bs)
get_syscall_by_number (gdbarch, last.value.syscall_number, &s);
annotate_catchpoint (b->number);
+ maybe_print_thread_hit_breakpoint (uiout);
if (b->disposition == disp_del)
- ui_out_text (uiout, "\nTemporary catchpoint ");
+ ui_out_text (uiout, "Temporary catchpoint ");
else
- ui_out_text (uiout, "\nCatchpoint ");
+ ui_out_text (uiout, "Catchpoint ");
if (ui_out_is_mi_like_p (uiout))
{
ui_out_field_string (uiout, "reason",
@@ -264,6 +264,7 @@ print_it_exception_catchpoint (bpstat bs)
enum exception_event_kind kind = classify_exception_breakpoint (b);
annotate_catchpoint (b->number);
+ maybe_print_thread_hit_breakpoint (uiout);
bp_temp = b->disposition == disp_del;
ui_out_text (uiout,
@@ -4790,6 +4790,37 @@ watchpoint_value_print (struct value *val, struct ui_file *stream)
}
}
+/* Print the "Thread ID hit" part of "Thread ID hit Breakpoint N" if
+ debugging multiple threads. */
+
+void
+maybe_print_thread_hit_breakpoint (struct ui_out *uiout)
+{
+ if (ui_out_is_mi_like_p (uiout))
+ return;
+
+ ui_out_text (uiout, "\n");
+
+ if (show_thread_that_caused_stop ())
+ {
+ const char *name;
+ struct thread_info *thr = inferior_thread ();
+
+ ui_out_text (uiout, "Thread ");
+ ui_out_field_fmt (uiout, "thread-id", "%s", print_thread_id (thr));
+
+ name = thr->name != NULL ? thr->name : target_thread_name (thr);
+ if (name != NULL)
+ {
+ ui_out_text (uiout, " \"");
+ ui_out_field_fmt (uiout, "name", "%s", name);
+ ui_out_text (uiout, "\"");
+ }
+
+ ui_out_text (uiout, " hit ");
+ }
+}
+
/* Generic routine for printing messages indicating why we
stopped. The behavior of this function depends on the value
'print_it' in the bpstat structure. Under some circumstances we
@@ -8093,10 +8124,11 @@ print_it_catch_fork (bpstat bs)
struct fork_catchpoint *c = (struct fork_catchpoint *) bs->breakpoint_at;
annotate_catchpoint (b->number);
+ maybe_print_thread_hit_breakpoint (uiout);
if (b->disposition == disp_del)
- ui_out_text (uiout, "\nTemporary catchpoint ");
+ ui_out_text (uiout, "Temporary catchpoint ");
else
- ui_out_text (uiout, "\nCatchpoint ");
+ ui_out_text (uiout, "Catchpoint ");
if (ui_out_is_mi_like_p (uiout))
{
ui_out_field_string (uiout, "reason",
@@ -8210,10 +8242,11 @@ print_it_catch_vfork (bpstat bs)
struct fork_catchpoint *c = (struct fork_catchpoint *) b;
annotate_catchpoint (b->number);
+ maybe_print_thread_hit_breakpoint (uiout);
if (b->disposition == disp_del)
- ui_out_text (uiout, "\nTemporary catchpoint ");
+ ui_out_text (uiout, "Temporary catchpoint ");
else
- ui_out_text (uiout, "\nCatchpoint ");
+ ui_out_text (uiout, "Catchpoint ");
if (ui_out_is_mi_like_p (uiout))
{
ui_out_field_string (uiout, "reason",
@@ -8406,10 +8439,11 @@ print_it_catch_solib (bpstat bs)
struct ui_out *uiout = current_uiout;
annotate_catchpoint (b->number);
+ maybe_print_thread_hit_breakpoint (uiout);
if (b->disposition == disp_del)
- ui_out_text (uiout, "\nTemporary catchpoint ");
+ ui_out_text (uiout, "Temporary catchpoint ");
else
- ui_out_text (uiout, "\nCatchpoint ");
+ ui_out_text (uiout, "Catchpoint ");
ui_out_field_int (uiout, "bkptno", b->number);
ui_out_text (uiout, "\n");
if (ui_out_is_mi_like_p (uiout))
@@ -8675,10 +8709,11 @@ print_it_catch_exec (bpstat bs)
struct exec_catchpoint *c = (struct exec_catchpoint *) b;
annotate_catchpoint (b->number);
+ maybe_print_thread_hit_breakpoint (uiout);
if (b->disposition == disp_del)
- ui_out_text (uiout, "\nTemporary catchpoint ");
+ ui_out_text (uiout, "Temporary catchpoint ");
else
- ui_out_text (uiout, "\nCatchpoint ");
+ ui_out_text (uiout, "Catchpoint ");
if (ui_out_is_mi_like_p (uiout))
{
ui_out_field_string (uiout, "reason",
@@ -10207,10 +10242,13 @@ print_it_ranged_breakpoint (bpstat bs)
gdb_assert (bl && bl->next == NULL);
annotate_breakpoint (b->number);
+
+ maybe_print_thread_hit_breakpoint (uiout);
+
if (b->disposition == disp_del)
- ui_out_text (uiout, "\nTemporary ranged breakpoint ");
+ ui_out_text (uiout, "Temporary ranged breakpoint ");
else
- ui_out_text (uiout, "\nRanged breakpoint ");
+ ui_out_text (uiout, "Ranged breakpoint ");
if (ui_out_is_mi_like_p (uiout))
{
ui_out_field_string (uiout, "reason",
@@ -10719,11 +10757,13 @@ print_it_watchpoint (bpstat bs)
stb = mem_fileopen ();
old_chain = make_cleanup_ui_file_delete (stb);
+ annotate_watchpoint (b->number);
+ maybe_print_thread_hit_breakpoint (uiout);
+
switch (b->type)
{
case bp_watchpoint:
case bp_hardware_watchpoint:
- annotate_watchpoint (b->number);
if (ui_out_is_mi_like_p (uiout))
ui_out_field_string
(uiout, "reason",
@@ -10758,7 +10798,6 @@ print_it_watchpoint (bpstat bs)
case bp_access_watchpoint:
if (bs->old_val != NULL)
{
- annotate_watchpoint (b->number);
if (ui_out_is_mi_like_p (uiout))
ui_out_field_string
(uiout, "reason",
@@ -10935,10 +10974,12 @@ print_it_masked_watchpoint (bpstat bs)
/* Masked watchpoints have only one location. */
gdb_assert (b->loc && b->loc->next == NULL);
+ annotate_watchpoint (b->number);
+ maybe_print_thread_hit_breakpoint (uiout);
+
switch (b->type)
{
case bp_hardware_watchpoint:
- annotate_watchpoint (b->number);
if (ui_out_is_mi_like_p (uiout))
ui_out_field_string
(uiout, "reason",
@@ -13120,10 +13161,12 @@ bkpt_print_it (bpstat bs)
bl->address,
b->number, 1);
annotate_breakpoint (b->number);
+ maybe_print_thread_hit_breakpoint (uiout);
+
if (bp_temp)
- ui_out_text (uiout, "\nTemporary breakpoint ");
+ ui_out_text (uiout, "Temporary breakpoint ");
else
- ui_out_text (uiout, "\nBreakpoint ");
+ ui_out_text (uiout, "Breakpoint ");
if (ui_out_is_mi_like_p (uiout))
{
ui_out_field_string (uiout, "reason",
@@ -1622,4 +1622,8 @@ extern void breakpoint_free_objfile (struct objfile *objfile);
extern char *ep_parse_optional_if_clause (char **arg);
+/* Print the "Thread ID hit" part of "Thread ID hit Breakpoint N" to
+ UIOUT iff debugging multiple threads. */
+extern void maybe_print_thread_hit_breakpoint (struct ui_out *uiout);
+
#endif /* !defined (BREAKPOINT_H) */
@@ -630,6 +630,11 @@ extern void thread_cancel_execution_command (struct thread_info *thr);
executing). */
extern void validate_registers_access (void);
+/* Returns whether to show which thread hit the breakpoint, received a
+ signal, etc. and ended up causing a user-visible stop. This is
+ true iff we ever detected multiple threads. */
+extern int show_thread_that_caused_stop (void);
+
extern struct thread_info *thread_list;
#endif /* GDBTHREAD_H */
@@ -7896,21 +7896,35 @@ print_exited_reason (struct ui_out *uiout, int exitstatus)
void
print_signal_received_reason (struct ui_out *uiout, enum gdb_signal siggnal)
{
+ struct thread_info *thr = inferior_thread ();
+
annotate_signal ();
- if (siggnal == GDB_SIGNAL_0 && !ui_out_is_mi_like_p (uiout))
+ if (ui_out_is_mi_like_p (uiout))
+ ;
+ else if (show_thread_that_caused_stop ())
{
- struct thread_info *t = inferior_thread ();
+ const char *name;
- ui_out_text (uiout, "\n[");
- ui_out_field_string (uiout, "thread-name",
- target_pid_to_str (t->ptid));
- ui_out_field_fmt (uiout, "thread-id", "] #%s", print_thread_id (t));
- ui_out_text (uiout, " stopped");
+ ui_out_text (uiout, "\nThread ");
+ ui_out_field_fmt (uiout, "thread-id", "%s", print_thread_id (thr));
+
+ name = thr->name != NULL ? thr->name : target_thread_name (thr);
+ if (name != NULL)
+ {
+ ui_out_text (uiout, " \"");
+ ui_out_field_fmt (uiout, "name", "%s", name);
+ ui_out_text (uiout, "\"");
+ }
}
else
+ ui_out_text (uiout, "\nProgram");
+
+ if (siggnal == GDB_SIGNAL_0 && !ui_out_is_mi_like_p (uiout))
+ ui_out_text (uiout, " stopped");
+ else
{
- ui_out_text (uiout, "\nProgram received signal ");
+ ui_out_text (uiout, " received signal ");
annotate_signal_name ();
if (ui_out_is_mi_like_p (uiout))
ui_out_field_string
@@ -48,7 +48,7 @@ gdb_test_multiple $test $test {
set test "process stopped"
gdb_test_multiple "" $test {
- -re "\r\n\\\[process \[0-9\]+\\\] #1 stopped\\\.\r\n" {
+ -re "Program stopped\\\.\r\n" {
pass $test
}
}
@@ -64,7 +64,7 @@ gdb_test_multiple $test $test {
set test "inferior stopped"
gdb_test_multiple "" $test {
- -re "\r\n\\\[.*\\\] #1 stopped\\\.\r\n" {
+ -re "Program stopped\\\.\r\n" {
pass $test
}
}
@@ -41,7 +41,7 @@ if { ![runto_main] } then {
}
# Run to the signal.
-gdb_test "continue" ".*Program received signal SIGSEGV.*" "continue to signal"
+gdb_test "continue" "Thread .* received signal SIGSEGV.*" "continue to signal"
# Try to generate a core file, for a later test.
set gcorefile [standard_output_file $testfile.gcore]
@@ -31,7 +31,7 @@ gdb_test "rwatch watchee"
gdb_breakpoint [gdb_get_line_number "dummy = 2"]
-gdb_test "continue" "Continuing.\r\nHardware read watchpoint \[0-9\]+: watchee\r\n\r\nValue = 0\r\n.*"
+gdb_test "continue" "Continuing.\r\n\r\nHardware read watchpoint \[0-9\]+: watchee\r\n\r\nValue = 0\r\n.*"
# Here should be no repeated notification of the read watchpoint.
gdb_test "continue" \
@@ -51,5 +51,5 @@ gdb_test "next" \
set line [gdb_get_line_number "stop point"]
gdb_breakpoint $line
-gdb_test "continue" "Continuing.\[\r\n\]*success\[\r\n\]*Breakpoint .*:$line\[\r\n\]*.*// stop point\[\r\n\]*" \
+gdb_test "continue" "Continuing.\[\r\n\]*success\[\r\n\]*Thread .* hit Breakpoint .*:$line\[\r\n\]*.*// stop point\[\r\n\]*" \
"continue to success for next-over-NPE"
@@ -30,5 +30,5 @@ if { ![runto_main] } {
}
gdb_test "continue" \
- "Program received signal SIGUSR1, User defined signal 1.*" \
+ "Thread 2 received signal SIGUSR1, User defined signal 1.*" \
"catch SIGUSR1"
@@ -116,7 +116,7 @@ for {set i 0} {$i < $attempts} {incr i} {
set msg "caught interrupt"
gdb_test_multiple "" $msg {
- -re "Program received signal SIGINT.*$gdb_prompt $" {
+ -re "Thread .* received signal SIGINT.*$gdb_prompt $" {
pass $msg
}
}
@@ -55,7 +55,7 @@ send_gdb "\003"
set test "caught interrupt"
gdb_test_multiple "" $test {
- -re "Program received signal SIGINT.*$gdb_prompt $" {
+ -re "Thread .* received signal SIGINT.*$gdb_prompt $" {
pass $test
}
}
@@ -70,14 +70,14 @@ proc interrupt_and_wait { message } {
-re "\\\[\[^\]\]* exited\\\]\r\n" {
exp_continue
}
- -re "Program received signal SIGINT.*$gdb_prompt $" {
+ -re " received signal SIGINT.*$gdb_prompt $" {
pass "$message"
}
-re "$gdb_prompt $" {
# Note that with this regex order, if GDB emits [New
# Thread ...] output between "Program received signal" and
- # the prompt, the "Program received signal" regex won't
- # match. That's good, as if we see that happening, it's a
+ # the prompt, the "received signal" regex won't match.
+ # That's good, as if we see that happening, it's a
# regression.
#
# GDB makes sure to notify about signal stops, end of
@@ -92,7 +92,7 @@ proc interrupt_and_wait { message } {
# ... more new threads output ...
# [New Thread NNN]
# [New Thread NNN]
- # Program received signal SIGINT, Interrupt.
+ # Thread NNN received signal SIGINT, Interrupt.
# [New Thread NNN]
# [New Thread NNN]
# ... pages of new threads output ...
@@ -202,7 +202,7 @@ proc check_control_c {} {
send_gdb "\003"
set description "Stopped with a ^C"
gdb_expect {
- -re "Program received signal SIGINT.*$gdb_prompt $" {
+ -re "Thread .* received signal SIGINT.*$gdb_prompt $" {
pass $description
}
-re "Quit.*$gdb_prompt $" {
@@ -69,7 +69,7 @@ proc stop_process { description } {
# message first, or GDB's signal handler may not be in place.
after 1000 {send_gdb "\003"}
gdb_expect {
- -re "Program received signal SIGINT.*$gdb_prompt $"
+ -re "Thread .* received signal SIGINT.*$gdb_prompt $"
{
pass $description
}
@@ -51,7 +51,7 @@ for {set sigcount 0} {$sigcount < 4} {incr sigcount} {
set test "catch signal $sigcount"
set sigusr ""
gdb_test_multiple "continue" $test {
- -re "Program received signal SIGUSR(\[12\]), User defined signal \[12\]\\.\r\n.*\r\n$gdb_prompt $" {
+ -re "Thread .* received signal SIGUSR(\[12\]), User defined signal \[12\]\\.\r\n.*\r\n$gdb_prompt $" {
set sigusr $expect_out(1,string)
pass $test
}
@@ -68,9 +68,9 @@ proc test { schedlock } {
# signal.
gdb_test "thread 3" "Switching to thread 3.*"
- gdb_test "continue" "Program received signal SIGUSR2.*" "stop with SIGUSR2"
+ gdb_test "continue" "Thread 3 .*received signal SIGUSR2.*" "stop with SIGUSR2"
gdb_test "thread 2" "Switching to thread 2.*"
- gdb_test "continue" "Program received signal SIGUSR1.*" "stop with SIGUSR1"
+ gdb_test "continue" "Thread 2 .*received signal SIGUSR1.*" "stop with SIGUSR1"
gdb_test "break handler_sigusr1" "Breakpoint .* at .*$srcfile.*"
gdb_test "break handler_sigusr2" "Breakpoint .* at .*$srcfile.*"
@@ -41,7 +41,7 @@ proc test { command } {
gdb_test "handle SIGUSR1 stop print pass"
- gdb_test "continue" "Program received signal SIGUSR1.*" "stop with SIGUSR1"
+ gdb_test "continue" "Thread 2 .*received signal SIGUSR1.*" "stop with SIGUSR1"
set pattern "\\\* 2\[ \t\]+Thread.*"
@@ -49,4 +49,4 @@ after 500 {send_gdb "\003"}
# Make sure we do not get an internal error from hitting Control-C
# while many signals are flying back and forth.
-gdb_test "" "Program received signal SIGINT.*" "stop with control-c"
+gdb_test "" "Thread .* received signal SIGINT.*" "stop with control-c"
@@ -75,7 +75,7 @@ proc test {type symbol} {
"reakpoint \[0-9\]+, marker.*" "breakpoint after the second fork"
gdb_test "continue" \
"atchpoint \[0-9\]+: var.*Old value = 2.*New value = 3.*mark_exit \\(\\);" "watchpoint after the second fork"
- gdb_test "continue" "Continuing\\..*\r\nBreakpoint \[0-9\]+, mark_exit .*" "finish"
+ gdb_test "continue" "Continuing\\..*\r\n(Thread .* hit )?Breakpoint \[0-9\]+, mark_exit .*" "finish"
}
# threads
@@ -129,7 +129,7 @@ proc test {type symbol} {
"atchpoint \[0-9\]+: var.*Old value = 5.*New value = 6.*after-fork2-A.*" "watchpoint A after the second fork"
gdb_test "continue" \
"atchpoint \[0-9\]+: var.*Old value = 6.*New value = 7.*after-fork2-B.*" "watchpoint B after the second fork"
- gdb_test "continue" "Continuing\\..*\r\nBreakpoint \[0-9\]+, mark_exit .*" "finish"
+ gdb_test "continue" "Continuing\\..*\r\nThread .* hit Breakpoint \[0-9\]+, mark_exit .*" "finish"
}
}
}
@@ -1685,6 +1685,14 @@ make_cleanup_restore_current_thread (void)
/* See gdbthread.h. */
int
+show_thread_that_caused_stop (void)
+{
+ return highest_thread_num > 1;
+}
+
+/* See gdbthread.h. */
+
+int
show_inferior_qualified_tids (void)
{
return (inferior_list->next != NULL || inferior_list->num != 1);