binutils-gdb/gdb/testsuite/gdb.threads/schedlock.exp
Pedro Alves 856e7dd698 Make "set scheduler-locking step" depend on user intention, only
Currently, "set scheduler-locking step" is a bit odd.  The manual
documents it as being optimized for stepping, so that focus of
debugging does not change unexpectedly, but then it says that
sometimes other threads may run, and thus focus may indeed change
unexpectedly...  A user can then be excused to get confused and wonder
why does GDB behave like this.

I don't think a user should have to know about details of how "next"
or whatever other run control command is implemented internally to
understand when does the "scheduler-locking step" setting take effect.

This patch completes a transition that the code has been moving
towards for a while.  It makes "set scheduler-locking step" hold
threads depending on whether the _command_ the user entered was a
stepping command [step/stepi/next/nexti], or not.

Before, GDB could end up locking threads even on "continue" if for
some reason run control decides a thread needs to be single stepped
(e.g., for a software watchpoint).

After, if a "continue" happens to need to single-step for some reason,
we won't lock threads (unless when stepping over a breakpoint,
naturally).  And if a stepping command wants to continue a thread for
bit, like when skipping a function to a step-resume breakpoint, we'll
still lock threads, so focus of debugging doesn't change.

In order to make this work, we need to record in the thread structure
whether what set it running was a stepping command.

(A follow up patch will remove the "step" parameters of 'proceed' and 'resume')

FWIW, Fedora GDB, which defaults to "scheduler-locking step" (mainline
defaults to "off") carries a different patch that goes in this
direction as well.

Tested on x86_64 Fedora 20, native and gdbserver.

gdb/ChangeLog:
2015-03-24  Pedro Alves  <palves@redhat.com>

	* gdbthread.h (struct thread_control_state) <stepping_command>:
	New field.
	* infcmd.c (step_once): Pass step=1 to clear_proceed_status.  Set
	the thread's stepping_command field.
	* infrun.c (resume): Check the thread's stepping_command flag to
	determine which threads should be resumed.  Rename 'entry_step'
	local to user_step.
	(clear_proceed_status_thread): Clear 'stepping_command'.
	(schedlock_applies): Change parameter type to struct thread_info
	pointer.  Adjust.
	(find_thread_needs_step_over): Remove 'step' parameter.  Adjust.
	(switch_back_to_stepped_thread): Adjust calls to
	'schedlock_applies'.
	(_initialize_infrun): Adjust "set scheduler-locking step" help.

gdb/testsuite/ChangeLog:
2015-03-24  Pedro Alves  <palves@redhat.com>

	* gdb.threads/schedlock.exp (test_step): No longer expect that
	"set scheduler-locking step" with "next" over a function call runs
	threads unlocked.

gdb/doc/ChangeLog:
2015-03-24  Pedro Alves  <palves@redhat.com>

	* gdb.texinfo (test_step) <set scheduler-locking step>: No longer
	mention that threads may sometimes run unlocked.
2015-03-24 17:50:31 +00:00

316 lines
8.0 KiB
Plaintext

# Copyright (C) 1996-2015 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/>.
# This file was written by Daniel Jacobowitz <drow@mvista.com>
# (parts based on pthreads.exp by Fred Fish (fnf@cygnus.com).
#
# This test covers the various forms of "set scheduler-locking".
standard_testfile
# The number of threads, including the main thread.
set NUM 2
if {[gdb_compile_pthreads "${srcdir}/${subdir}/${srcfile}" "${binfile}" executable debug] != "" } {
return -1
}
# Now we can proceed with the real testing.
# Get the current contents of the `args` array in the test program.
# Description is appended to the test message.
proc get_args { description } {
global gdb_prompt
global NUM
set pattern "(\[0-9\]+)"
for {set i 1} {[expr $i < $NUM]} {incr i} {
append pattern ", (\[0-9\]+)"
}
set test "listed args ($description)"
gdb_test_multiple "print args" $test {
-re "\\\$\[0-9\]+ = {$pattern}.*$gdb_prompt $" {
pass $test
set result ""
for {set i 1} {[expr $i <= $NUM]} {incr i} {
lappend result $expect_out($i,string)
}
return $result
}
}
}
proc stop_process { description } {
global gdb_prompt
# For this to work we must be sure to consume the "Continuing."
# message first, or GDB's signal handler may not be in place.
after 1000 {send_gdb "\003"}
gdb_expect {
-re "Program received signal SIGINT.*$gdb_prompt $"
{
pass $description
}
timeout
{
fail "$description (timeout)"
}
}
}
proc get_current_thread { description } {
global gdb_prompt
set test "find current thread ($description)"
gdb_test_multiple "bt" $test {
-re "thread_function \\(arg=0x(\[0-9\])\\).*$gdb_prompt $" {
pass $test
return $expect_out(1,string)
}
}
return ""
}
# Make sure we're stopped in the loop, in one of the non-main threads.
proc goto_loop { msg } {
gdb_breakpoint [concat [gdb_get_line_number "schedlock.exp: main loop"] " if arg != 0"]
set test "return to loop"
if {$msg != ""} {
set test "$test ($msg)"
}
gdb_continue_to_breakpoint $test
delete_breakpoints
}
proc my_continue { msg } {
set test "continue ($msg)"
gdb_test_multiple "continue" $test {
-re "Continuing" {
pass $test
}
}
stop_process "stop all threads ($msg)"
goto_loop $msg
}
# Use CMD to step the loop 10 times. CMD may be "step" or "next".
proc step_ten_loops { cmd } {
global gdb_prompt
for {set i 0} {[expr $i < 10]} {set i [expr $i + 1]} {
set other_step 0
set test "$cmd to increment ($i)"
gdb_test_multiple $cmd $test {
-re ".*myp\\) \\+\\+;\[\r\n\]+$gdb_prompt $" {
pass $test
}
-re "$gdb_prompt $" {
if {$other_step == 0} {
set other_step 1
send_gdb "$cmd\n"
exp_continue
} else {
fail $test
# FIXME cascade?
}
}
}
}
}
# Start with a fresh gdb.
gdb_exit
gdb_start
gdb_reinitialize_dir $srcdir/$subdir
# We'll need this when we send_gdb a ^C to GDB. Need to do it before we
# run the program and gdb starts saving and restoring tty states.
gdb_test "shell stty intr '^C'" ".*"
gdb_load ${binfile}
gdb_test_no_output "set print sevenbit-strings"
gdb_test_no_output "set width 0"
runto_main
# See if scheduler locking is available on this target.
global gdb_prompt
gdb_test_multiple "set scheduler-locking off" "scheduler locking set to none" {
-re "Target .* cannot support this command" {
unsupported "target does not support scheduler locking"
return
}
-re "$gdb_prompt $" {
pass "scheduler locking set to none"
}
timeout {
unsupported "target does not support scheduler locking (timeout)"
return
}
}
gdb_breakpoint [gdb_get_line_number "schedlock.exp: last thread start"]
gdb_continue_to_breakpoint "all threads started"
set start_args [get_args "before initial"]
# First make sure that all threads are alive.
my_continue "initial"
set cont_args [get_args "after initial"]
set bad 0
for {set i 0} {[expr $i < $NUM]} {set i [expr $i + 1]} {
if {[lindex $start_args $i] == [lindex $cont_args $i]} {
incr bad
}
}
if { $bad == 0 } {
pass "all threads alive"
} else {
fail "all threads alive ($bad/$NUM did not run)"
}
# Compare the previous thread and args with the current thread and
# args. Check that we didn't switch threads, and that the threads
# incremented their args counter the amounts expected. CMD is the
# command being tested. BEFORE_THREAD is the thread that was selected
# before the command was run. BEFORE_ARGS is the value of the
# thread's args before the command was run. LOCKED indicates whether
# we expect threads other than the selected thread remained locked.
proc check_result { cmd before_thread before_args locked } {
global NUM
# Make sure we're still in the same thread.
set newthread [get_current_thread "after"]
set test "$cmd does not change thread"
if {$before_thread == $newthread} {
pass "$test"
} else {
fail "$test (switched to thread $newthread)"
}
set after_args [get_args "after"]
set test "current thread advanced"
if { $locked } {
set test "$test - locked"
} else {
set test "$test - unlocked"
}
set num_other_threads 0
for {set i 0} {$i < $NUM} {incr i} {
if {[lindex $before_args $i] == [lindex $after_args $i]} {
if {$i == $before_thread} {
fail "$test (didn't run)"
}
} else {
if {$i == $before_thread} {
if {$cmd == "continue"
|| [lindex $before_args $i] == [expr [lindex $after_args $i] - 10]} {
pass "$test"
} else {
fail "$test (wrong amount)"
}
} else {
incr num_other_threads
}
}
}
if { $locked } {
gdb_assert {$num_other_threads == 0} "other threads didn't run - locked"
} else {
gdb_assert {$num_other_threads > 0} "other threads ran - unlocked"
}
}
with_test_prefix "schedlock=on: cmd=continue" {
# Use whichever we stopped in.
set curthread [get_current_thread "before"]
# Test continue with scheduler locking.
gdb_test "set scheduler-locking on" ""
my_continue "with lock"
check_result "continue" $curthread $cont_args 1
}
# Test stepping/nexting with different modes of scheduler locking.
proc test_step { schedlock cmd call_function } {
global NUM
gdb_test_no_output "set scheduler-locking off"
goto_loop ""
set curthread [get_current_thread "before"]
# No need to set to off again. This avoids a duplicate message.
if {$schedlock != "off"} {
gdb_test_no_output "set scheduler-locking $schedlock"
}
gdb_test "print call_function = $call_function" \
" = $call_function"
set before_args [get_args "before"]
step_ten_loops $cmd
if { $schedlock == "on" || $schedlock == "step" } {
set locked 1
} else {
set locked 0
}
check_result $cmd $curthread $before_args $locked
}
# Test stepping/nexting with different modes of scheduler locking.
foreach schedlock {"off" "step" "on"} {
with_test_prefix "schedlock=$schedlock" {
with_test_prefix "cmd=step" {
test_step $schedlock "step" 0
}
with_test_prefix "cmd=next" {
# In GDB <= 7.9, with schedlock "step", "next" would
# unlock threads when stepping over a function call. This
# exercises "next" with and without a function call. WRT
# "schedlock step", "next" should behave just like "step".
foreach call_function {0 1} {
with_test_prefix "call_function=$call_function" {
test_step $schedlock "next" $call_function
}
}
}
}
}