Re: Thread exit error : gdb7.2 in FreeBSD (built from ports)

[Jusctsch <schumjs@gmail.com>]

Pedro Alves-10 wrote:
> 
> Please don't top post.
> 
> I understand that the whole program exited, and you're getting
> that error from withing get_current_frame, right?  Please give
> me the most detail as possible so that I don't have to try
> to guess things.
> 
> Why is GDB calling get_current_frame at all?  Please show the
> same but with "set debug infrun 1".  Sounds like a FreeBSD
> backend bug.
> 
> -- 
> Pedro Alves
> 
> 

The virgin code for gdb7.2, after FreeBSD patches, is currently running.

This issue occurs after a thread has exited, but the logic in
delete_thread_1 dictates that this if statement is fallen through:
thread.c:259-260
  if (tp->refcount > 0 || ptid_equal (tp->ptid, inferior_ptid))
In which we do not delete the internal bookkeeping of the thread, just flag
it as exited. It is my understanding that we then soon hit the
get_current_frame() for this exited thread (as the exited thread's ptid is
the inferior_ptid). This flag is hit in this code, which prompts the user to
switch threads.
This series of events is what I consider to be the bug.

This is the backtrace at the point of the error in get current_frame():
1777  	error (_("Invalid selected thread."));
(gdb) bt
#0  get_current_frame () at frame.c:1177
#1  0x000000000053d5f4 in handle_inferior_event (ecs=0x7fffffffe300) at
infrun.c:3697
#2  0x000000000053ae53 in wait_for_inferior (treat_exec_as_sigtrap=0) at
infrun.c:2551
#3  0x000000000053a12c in proceed (addr=34386749296,
siggnal=TARGET_SIGNAL_0, step=0) at infrun.c:2064
#4  0x0000000000533760 in run_command_1 (args=0x801314084 "-d", from_tty=1,
tbreak_at_main=0) at infcmd.c:586
#5  0x00000000005337a0 in run_command (args=0x801314084 "-d", from_tty=1) at
infcmd.c:596
#6  0x00000000004aba04 in do_cfunc (c=0x8013a5700, args=0x801314084 "-d",
from_tty=1) at ./cli/cli-decode.c:67
#7  0x00000000004ae9c5 in cmd_func (cmd=0x8013a5700, args=0x801314084 "-d",
from_tty=1) at ./cli/cli-decode.c:1771
#8  0x000000000044d24c in execute_command (p=0x801314085 "d", from_tty=1) at
top.c:422
#9  0x000000000055533e in command_handler (command=0x801314080 "") at
event-top.c:498
#10 0x0000000000555922 in command_line_handler (rl=0x801313c68 "run -d") at
event-top.c:702
#11 0x0000000800ab7317 in rl_callback_read_char () from
/lib/libreadline.so.8
#12 0x00000000005549a1 in rl_callback_read_char_wrapper (client_data=0x0) at
event-top.c:178
#13 0x0000000000555209 in stdin_event_handler (error=0, client_data=0x0) at
event-top.c:433
#14 0x00000000005539ab in handle_file_event (data=...) at event-loop.c:817
#15 0x0000000000553022 in process_event () at event-loop.c:399
#16 0x000000000055310b in gdb_do_one_event (data=0x0) at event-loop.c:464
#17 0x000000000054d91f in catch_errors (func=0x553040 <gdb_do_one_event>,
func_args=0x0, errstring=0x741f23 "",
    mask=6) at exceptions.c:518
#18 0x00000000004c2144 in tui_command_loop (data=0x0) at
./tui/tui-interp.c:171
#19 0x000000000054e097 in current_interp_command_loop () at interps.c:291
#20 0x0000000000442d61 in captured_command_loop (data=0x0) at ./main.c:227
#21 0x000000000054d91f in catch_errors (func=0x442d50
<captured_command_loop>, func_args=0x0,
    errstring=0x7246a3 "", mask=6) at exceptions.c:518
#22 0x0000000000443c8a in captured_main (data=0x7fffffffea40) at
./main.c:925
#23 0x000000000054d91f in catch_errors (func=0x442da0 <captured_main>,
func_args=0x7fffffffea40,
    errstring=0x7246a3 "", mask=6) at exceptions.c:518
#24 0x0000000000443d11 in gdb_main (args=0x7fffffffea40) at ./main.c:934
#25 0x0000000000442a8e in main (argc=2, argv=0x7fffffffeab0) at gdb.c:43
(gdb)

Any thoughts? Is this the regular operation of gdb7.2? Maybe that
tp->refcount field is never decremented to 0 (for the exiting thread for
some reason). I will take a look at that and append my results, as well as
running gdb under "set debug infrun 1".

-- 
View this message in context: http://old.nabble.com/Thread-exit-error-%3A-gdb7.2-in-FreeBSD-%28built-from-ports%29-tp32463912p32503829.html
Sent from the Sourceware - gdb list mailing list archive at Nabble.com.