2015-01-01 09:32:14 +00:00
|
|
|
# Copyright (C) 2013-2015 Free Software Foundation, Inc.
|
GDB kills itself instead of interrupting inferior
When GDB is run with IO redirected to a pipe, the 'interrupt' command
causes it to kill its own process group instead of the inferior's.
The problem manifests itself in async mode, native debugging:
$ cat | gdb <file>
(gdb) set target-async on
(gdb) run &
(gdb) interrupt
A debugging session is active.
Inferior 1 [process 20584] will be killed.
Quit anyway? (y or n) [answered Y; input not from terminal]
In this case, GDB tells that its stdin isn't a tty and doesn't save
the inferior's process group in
inflow.c:terminal_init_inferior_with_pgrp. The 'interrupt' command
tries to 'kill' the inferior's process group in
`inf-ptrace.c:inf_ptrace_stop`, but since that wasn't saved in the
first place, GDB kills process group 0, meaning, its own process
group.
When GDB is used from a frontend, that means killing its own process
group including the frontend and possibly the X session. This was
originally seen with SublimeGDB:
https://github.com/quarnster/SublimeGDB/issues/29.
The patch makes GDB save the inferior pgid regardless of having a
terminal, as pgid is used not only to reset foreground process group,
but also to interrupt the inferior process. It also adds a regression
test. Luckily, we can emulate not having a terminal with "set
interactive-mode off", avoiding the need of special magic to spawn gdb
with a pipe.
Tested on x86_64 Fedora 17.
gdb/
2013-07-26 Cyril Nikolaev <cyril@nichtverstehen.de>
* inflow.c (terminal_init_inferior_with_pgrp): Save inferior
process group regardless of having tty on stdin.
gdb/testsuite/
2013-07-26 Pedro Alves <palves@redhat.com>
* gdb.base/interrupt-noterm.c, gdb.base/interrupt-noterm.exp: New
files.
2013-07-26 11:15:45 +00:00
|
|
|
|
|
|
|
# This program is free software; you can redistribute it and/or modify
|
|
|
|
# it under the terms of the GNU General Public License as published by
|
|
|
|
# the Free Software Foundation; either version 3 of the License, or
|
|
|
|
# (at your option) any later version.
|
|
|
|
#
|
|
|
|
# This program is distributed in the hope that it will be useful,
|
|
|
|
# but WITHOUT ANY WARRANTY; without even the implied warranty of
|
|
|
|
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
|
|
|
# GNU General Public License for more details.
|
|
|
|
#
|
|
|
|
# You should have received a copy of the GNU General Public License
|
|
|
|
# along with this program. If not, see <http://www.gnu.org/licenses/>.
|
|
|
|
|
|
|
|
standard_testfile
|
|
|
|
|
|
|
|
if [prepare_for_testing "failed to prepare for testing" \
|
|
|
|
${testfile} ${srcfile} {debug}] {
|
|
|
|
return -1
|
|
|
|
}
|
|
|
|
|
|
|
|
# Pretend there's no terminal.
|
|
|
|
gdb_test_no_output "set interactive-mode off"
|
|
|
|
|
|
|
|
if ![runto main] {
|
|
|
|
fail "Can't run to main"
|
|
|
|
return -1
|
|
|
|
}
|
|
|
|
|
|
|
|
# Delete breakpoints so that the next resume is a plain continue,
|
|
|
|
# instead of a step-over-breakpoint sequence just while GDB sends the
|
|
|
|
# interrupt request. If that's buggy on some targets (and it was on
|
|
|
|
# target remote for a while, where a ctrl-c at the wrong time will get
|
|
|
|
# lost), then it should get its own specific test. Disable
|
|
|
|
# confirmation, avoiding complications caused by the fact that we've
|
|
|
|
# disabled the terminal -- GDB would auto-answer "yes", confusing
|
|
|
|
# gdb_test_multiple.
|
|
|
|
gdb_test_no_output "set confirm off"
|
|
|
|
gdb_test_no_output "delete"
|
|
|
|
gdb_test_no_output "set confirm on"
|
|
|
|
|
|
|
|
set async_supported -1
|
|
|
|
set test "continue &"
|
|
|
|
gdb_test_multiple $test $test {
|
|
|
|
-re "Continuing\\.\r\n$gdb_prompt $" {
|
|
|
|
set async_supported 1
|
|
|
|
pass $test
|
|
|
|
}
|
|
|
|
-re ".*Asynchronous execution not supported on this target..*" {
|
|
|
|
unsupported $test
|
|
|
|
}
|
|
|
|
}
|
|
|
|
if { $async_supported < 0 } {
|
|
|
|
return 1
|
|
|
|
}
|
|
|
|
|
|
|
|
# With native debugging, and no terminal (emulated by interactive-mode
|
|
|
|
# off, above), GDB had a bug where "interrupt" would send SIGINT to
|
|
|
|
# its own process group, instead of the inferior's.
|
|
|
|
set test "interrupt"
|
|
|
|
gdb_test_multiple $test $test {
|
|
|
|
-re "interrupt\r\n$gdb_prompt " {
|
|
|
|
pass $test
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
set test "inferior received SIGINT"
|
|
|
|
gdb_test_multiple "" $test {
|
|
|
|
-re "\r\nProgram received signal SIGINT.*" {
|
|
|
|
# This appears after the prompt, which was already consumed
|
|
|
|
# above.
|
|
|
|
pass $test
|
|
|
|
}
|
|
|
|
}
|