Skip to content

HTTPS clone URL

Subversion checkout URL

You can clone with HTTPS or Subversion.

Download ZIP
branch: master
Fetching contributors…

Cannot retrieve contributors at this time

309 lines (270 sloc) 10.413 kb
# Copyright 2008, 2010 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/>.
set testfile "inline-cmds"
set srcfile "${testfile}.c"
set srcfile2 "inline-markers.c"
set fullsrcfile "${srcdir}/${subdir}/${srcfile}"
set fullsrcfile2 "${srcdir}/${subdir}/${srcfile2}"
set sources [list ${fullsrcfile} ${fullsrcfile2}]
set binfile ${objdir}/${subdir}/${testfile}
if { [gdb_compile $sources ${binfile} \
executable {debug optimize=-O2}] != "" } {
untested inline-cmds.exp
return -1
}
gdb_exit
gdb_start
gdb_reinitialize_dir $srcdir/$subdir
gdb_load ${binfile}
gdb_test_no_output "set listsize 1"
runto_main
get_compiler_info $binfile
get_debug_format
if { [skip_inline_frame_tests] } {
untested inline-cmds.exp
return
}
# First, check that the things we expected to be inlined really were,
# and those that shouldn't be weren't.
set line1 [gdb_get_line_number "set breakpoint 1 here" ${srcfile2}]
gdb_breakpoint $srcfile2:$line1
set line2 [gdb_get_line_number "set breakpoint 2 here" ${srcfile2}]
gdb_breakpoint $srcfile2:$line2
gdb_test "continue" ".*set breakpoint 1 here.*" "continue to bar (1)"
gdb_test "backtrace" "#0 bar.*#1 .*func1.*#2 .*main.*" \
"backtrace from bar (1)"
gdb_test "up" "#1 .*func1.*" "up from bar (1)"
gdb_test "info frame" ".*inlined into frame.*" "func1 inlined (1)"
gdb_test "continue" ".*set breakpoint 1 here.*" "continue to bar (2)"
gdb_test "backtrace" "#0 bar.*#1 .*func1.*#2 .*func2.*#3 .*main.*" \
"backtrace from bar (2)"
gdb_test "up" "#1 .*func1.*" "up from bar (2)"
gdb_test "info frame" ".*inlined into frame.*" "func1 inlined (2)"
gdb_test "up" "#2 .*func2.*" "up from func1 (2)"
gdb_test "info frame" ".*inlined into frame.*" "func2 inlined (2)"
gdb_test "continue" ".*set breakpoint 2 here.*" "continue to marker"
gdb_test "backtrace" "#0 marker.*#1 .*main.*" "backtrace from marker"
gdb_test "info frame" ".*called by frame.*" "marker not inlined"
# Next, check that we can next over inlined functions. We should not end up
# inside any of them.
delete_breakpoints
runto_main
# The lines before the first inlined call.
set first "x = 7|y = 8"
# Some extra lines that end up in our stepping due to code motion.
set opt "start of main|result = 0"
# We start this test with a "list" instead of a "next", in case the
# first non-prologue instruction in main comes from the inlined function.
set msg "next over inlined functions"
gdb_test_multiple "list" $msg {
-re "($first|result = func1|result = func2|$opt).*$gdb_prompt $" {
send_gdb "next\r"
exp_continue
}
-re "marker \\\(\\\);\r\n$gdb_prompt $" {
pass $msg
}
}
# Check that when next shows the call of func1, it has not happened yet.
runto_main
# Like the return value of gdb_test: -1 something is wrong, 0 passed, 1 failed.
set bt_test -1
set x_test -1
set func1_step -1
set last_was_func1_call 0
set msg "next past inlined func1"
gdb_test_multiple "list" $msg {
-re "($first|$opt).*$gdb_prompt $" {
set last_was_func1_call 0
send_gdb "next\r"
exp_continue
}
-re "result = func1 \\\(\\\);\r\n$gdb_prompt $" {
# Check whether x has been set. If 0, we may be doing something
# else associated with this line besides the inlined call - e.g.
# loading the address of result. If 7, we may be at the call site.
# If 15, though, we might be past the call and back at the store to
# result - that's OK, as long as we weren't just here (see
# func1_step above).
set x_val -1
gdb_test_multiple "print x" "" {
-re "\\\$$decimal = (\[0-9\]*)\r\n$gdb_prompt $" {
set x_val $expect_out(1,string)
}
-re "$gdb_prompt $" { }
}
if { $x_val == 0 || $x_val == 7 } {
if { $x_test != 1 } {
set x_test 0
}
} elseif { $x_val == 15 } {
if { $func1_step == -1 } {
# We passed func1 without stopping at the call site.
set x_test 1
}
} else {
set x_test 1
}
# func1 should not show up on backtraces if we are at its call
# site.
if { $bt_test != 1 } {
set bt_test [gdb_test "backtrace" "#0 \[^#]*main.*" ""]
}
# When we next over func1, we should not return to the same
# line. But we might go past the line, according to source
# code order, and then come back. A valid but odd layout is
# body of func1, load result's address into a register using
# the source location of "result = 0" several lines down, and
# then return to this line for the store. GCC 4.3 does that
# on ARM.
if { $last_was_func1_call } {
set func1_step 1
} elseif { $func1_step == -1 } {
set func1_step 0
}
set last_was_func1_call 1
send_gdb "next\r"
exp_continue
}
-re "result = func2 \\\(\\\);\r\n$gdb_prompt $" {
pass $msg
}
}
if { $x_test == 0 } {
pass "print x before func1"
} else {
fail "print x before func1"
}
if { $bt_test == 0 } {
pass "backtrace does not include func1"
} else {
fail "backtrace does not include func1"
}
if { $bt_test == 0 } {
pass "stepped over call to func1"
} else {
fail "stepped over call to func1"
}
# Next, check that we can single step into inlined functions. We should always
# "stop" at the call sites before entering them.
runto_main
set msg "step into func1"
set saw_call_site 0
gdb_test_multiple "list" $msg {
-re "($first|$opt).*$gdb_prompt $" {
send_gdb "step\r"
exp_continue
}
-re "result = func1.*$gdb_prompt $" {
set saw_call_site 1
send_gdb "step\r"
exp_continue
}
-re "func1 \\\(\\\) at .*\r\n$decimal.*bar \\\(\\\);\r\n$gdb_prompt $" {
if { $saw_call_site } {
pass $msg
} else {
fail $msg
}
}
}
# Check finish out of an inlined function.
set msg "finish from func1"
gdb_test_multiple "finish" $msg {
-re "result = func1 \\\(\\\);\r\n$gdb_prompt $" {
pass $msg
}
-re "($first|$opt).*$gdb_prompt $" {
# Whoops. We finished, but ended up back at an earlier line. Keep
# trying.
send_gdb "step\r"
exp_continue
}
-re "func1 \\\(\\\) at .*\r\n$decimal.*bar \\\(\\\);\r\n$gdb_prompt $" {
send_gdb "finish\r"
exp_continue
}
}
# Test some corner cases involving consecutive inlined functions.
set line3 [gdb_get_line_number "set breakpoint 3 here"]
gdb_breakpoint $line3
gdb_continue_to_breakpoint "consecutive func1"
gdb_test "next" ".*func1 .*first call.*" "next to first func1"
set msg "next to second func1"
gdb_test_multiple "next" $msg {
-re ".*func1 .*second call.*$gdb_prompt $" {
pass $msg
}
-re ".*marker .*$gdb_prompt $" {
# This assembles to two consecutive call instructions.
# Both appear to be at the same line, because they're
# in the body of the same inlined function. This is
# reasonable for the line table. GDB should take the
# containing block and/or function into account when
# deciding how far to step. The single line table entry
# is actually two consecutive instances of the same line.
kfail gdb/NNNN $msg
}
}
# It is easier when the two inlined functions are not on the same line.
set line4 [gdb_get_line_number "set breakpoint 4 here"]
gdb_breakpoint $line4
gdb_continue_to_breakpoint "func1 then func3"
gdb_test "next" ".*func1 \\\(\\\);" "next to func1 before func3"
gdb_test "next" ".*func3 \\\(\\\);" "next to func3"
# Test finishing out of one thing and into another.
set line5 [gdb_get_line_number "set breakpoint 5 here"]
gdb_breakpoint $line5
gdb_continue_to_breakpoint "finish into func1"
gdb_test "next" ".*marker \\\(\\\);" "next to finish marker"
gdb_test "step" ".*set breakpoint 2 here.*" "step into finish marker"
gdb_test "finish" "func1 \\\(\\\);" "finish from marker to func1"
gdb_test "step" "bar \\\(\\\);" "step into func1 for finish"
gdb_test "finish" "func3 \\\(\\\);" "finish from func1 to func3"
# Test a deeper call stack.
set line6 [gdb_get_line_number "set breakpoint 6 here"]
gdb_breakpoint $line6
gdb_continue_to_breakpoint "before the outer_inline call"
gdb_test "step" "marker \\\(\\\) at .*" "reach 1 the outer_inline call"
gdb_test "finish" "main \\\(\\\) at .*outer_inline2 \\\(\\\);" "reach outer_inline2"
gdb_test "bt" "#0 main.*" "backtrace at main of outer_inline"
gdb_test "step" "outer_inline2 \\\(\\\) at .*" "enter outer_inline2"
gdb_test "bt" "#0 outer_inline2.*#1 main.*" "backtrace at outer_inline2"
gdb_test "step" "outer_inline1 \\\(\\\) at .*" "enter outer_inline1 from outer_inline2"
set msg "backtrace at outer_inline1"
gdb_test_multiple "bt" $msg {
-re "#0 outer_inline1.*#1 outer_inline2.*#2 main.*$gdb_prompt $" {
pass $msg
}
-re "#0 $hex in outer_inline1.*#1 outer_inline2.*#2 main.*$gdb_prompt $" {
# Binutils PR gas/6717. Gas moves .loc past .p2align and the
# leading nop of the inlined call appears to be on the same line
# as main's call to marker.
xfail $msg
gdb_test "step" "noinline \\\(\\\);" "step to call of noinline"
}
}
gdb_test "step" "noinline \\\(\\\) at .*" "enter noinline from outer_inline1"
gdb_test "bt" "#0 noinline.*#1 .*outer_inline1.*#2 .*outer_inline2.*#3 main.*" "backtrace at noinline from outer_inline1"
gdb_test "step" "inlined_fn \\\(\\\) at .*" "enter inlined_fn from noinline"
gdb_test "bt" "#0 inlined_fn.*#1 noinline.*#2 .*outer_inline1.*#3 .*outer_inline2.*#4 main.*" "backtrace at inlined_fn from noinline"
gdb_test "info frame" ".*inlined into frame.*" "inlined_fn from noinline inlined"
gdb_test "up" "#1 noinline.*" "up to noinline"
gdb_test "info frame" ".*\n called by frame.*" "noinline from outer_inline1 not inlined"
gdb_test "up" "#2 .*outer_inline1.*" "up to outer_inline1"
gdb_test "info frame" ".*inlined into frame.*" "outer_inline1 inlined"
gdb_test "up" "#3 .*outer_inline2.*" "up to outer_inline2"
gdb_test "info frame" ".*inlined into frame.*" "outer_inline2 inlined"
gdb_test "up" "#4 main.*" "up from outer_inline2"
gdb_test "info frame" ".*\n caller of frame.*" "main not inlined"
Jump to Line
Something went wrong with that request. Please try again.