| # Copyright 2021 Free Software Foundation, Inc. |
| |
| # 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/>. |
| |
| # Make an inferior call to a function which uses longjmp. However, |
| # the backtrace for the function that is called is broken at the point |
| # where the longjmp is handled. This test is checking to see if the |
| # inferior call still completes successfully. |
| # |
| # This test forces a broken backtrace using Python, but in real life a |
| # broken backtrace can easily occur when calling through code for |
| # which there is no debug information if the prologue unwinder fails, |
| # which can often happen if the code has been optimized. |
| # |
| # The problem was that, due to the broken backtrace, GDB failed to |
| # find the inferior call's dummy frame. GDB then concluded that the |
| # inferior had longjmp'd backward past the dummy frame and so garbage |
| # collected the dummy frame, this causes the breakpoint within the |
| # dummy frame to be deleted. |
| # |
| # When the inferior continued, and eventually returned to the dummy |
| # frame, it would try to execute instruction from the dummy frame |
| # (which for most, or even all, targets, is on the stack), and then |
| # experience undefined behaviuor, often a SIGSEGV. |
| |
| standard_testfile .c |
| |
| if { [prepare_for_testing "failed to prepare" $testfile $srcfile] } { |
| return -1 |
| } |
| |
| if ![runto_main] then { |
| return 0 |
| } |
| |
| # Skip this test if Python scripting is not enabled. |
| if { [skip_python_tests] } { continue } |
| |
| set pyfile [gdb_remote_download host ${srcdir}/${subdir}/${testfile}.py] |
| gdb_test_no_output "source ${pyfile}" "load python file" |
| |
| gdb_test "p some_func ()" " = 0" |
| |
| # When frame debugging is turned on, this test has (previously) |
| # revealed some crashes due to the Python frame unwinder trying to |
| # read registers. |
| # |
| # Enable frame debug and rerun the test. We don't bother checking the |
| # output of calling 'p some_func ()' as the output will be full of |
| # debug, to format of which isn't fixed. All we care about is that |
| # GDB is still running afterwards. |
| # |
| # All of the debug output makes this really slow when testing with the |
| # special read1 version of expect, hence the timeout factor. |
| with_read1_timeout_factor 10 { |
| gdb_test_no_output "set debug frame on" |
| gdb_test "p some_func ()" ".*" \ |
| "repeat p some_func () with frame debug on" |
| gdb_test_no_output "set debug frame off" |
| } |
| gdb_test "p 1 + 2 + 3" " = 6" |