This is the mail archive of the
gdb-patches@sourceware.org
mailing list for the GDB project.
Re: [PATCH] testsuite: Add (extensive) hardware breakpoint testing
On Fri, 11 Nov 2011, Doug Evans wrote:
> {i386,amd64}-linux are important enough targets that I think this
> should be fixed for 7.4.
OK, so we've got this part sorted out -- and I verified the two failures
observed are also present with break.exp, so they're a generic
x86_64-linux-gnu problem and nothing wrong with this test.
> One high level comment:
> There's a lot of basic linespec testing that would be nice to make table driven.
> E.g., {hbreak,thbreak} x {function, file:lineno, ...}, followed by
> delete_breakpoints.
> [I might not do this when testing running to the breakpoints, dunno if
> that is easily made data-driven, though that would be nice too.]
> If we're going to do (semi-)exhaustive linespec testing, I'd rather
> have a table than a lot of repetitive code.
>
> The same applies to break,tbreak.
> They too could be table driven.
>
> OTOH, maybe it's better to do simple linespec testing separately.
> There I would move {break,tbreak,hbreak,thbreak} x {function,
> file:lineno, ...} into its own file,
> have a table to drive it, and do a lot of testing with less code.
> It'd be more easily extensible too.
>
> I wouldn't make this a requirement, just mentioning it.
I agree lots of older test cases ask for cleaning up -- I refrained from
diverging from break.exp too much, although I did some obvious tidy-ups,
like removing the remaining send_gdb/gdb_expect pieces in favour to
gdb_test/gdb_test_multiple. Interestingly enough someone did that to
break.exp too at one point, but didn't catch all the places. I think
it'll make sense to complete it using hbreak2.exp as a reference for a
change. I can't commit to doing it now, but I'll try as my time permits.
> > @@ -0,0 +1,629 @@
> > +# ? Copyright 1988, 1990, 1991, 1992, 1994, 1995, 1996, 1997, 1998, 1999,
> > +# ? 2000, 2002, 2003, 2007, 2008, 2009, 2010, 2011
> > +# ? Free Software Foundation, Inc.
>
> This is a new file. I'm not sure what the rule is, but I think this
> should just say 2011.
I wasn't sure -- the contents of the file certainly are not new, as this
is break.exp taken as a whole and then edited. Not even quite heavily --
you can run a diff or compare the files side by side and they are still
pretty close to each other. I have removed the older years now; they can
be still added back if needed.
> > +# FIXME: The rest of this test doesn't work with anything that can't
> > +# handle arguments.
> > +# Huh? ?There doesn't *appear* to be anything that passes arguments
> > +# below.
>
> I think I'd just remove the FIXME and Huh? comments.
> And then ideally remove the mips-idt-* test.
Again, this was taken from break.exp verbatim. I have removed it now as
it seems to make little sense indeed, but in this case I think the former
should be updated accordingly.
> > + ? ?if [istarget "mips*tx39-*"] {
> > + ? ? ? set timeout 60
>
> I would do:
>
> set oldtimeout $timeout
> if [istarget ...] {
> set timeout 60
> }
>
> and then restore timeout at the end of the function.
Hmm, the TX39 is an R3000 clone and probably does not have hardware
breakpoints in the first place. ;) So this is another break.exp
compatibility place -- I'm leaving it in place for this reason.
As to your observation -- does it really matter? The timeout variable is
local to this function (accessed with upvar from gdb_test), so it's wiped
away at the point the procedure returns. I'm therefore leaving it as it
is unless you have further notes.
> > +# Reset the default arguments for VxWorks
> > +if [istarget "*-*-vxworks*"] {
> > + ? ?set timeout 10
> > + ? ?verbose "Timeout is now $timeout seconds" 2
>
> Timeout restoration is no longer needed at the end of a file.
> I would delete these lines.
>
> > + ? ?gdb_test_no_output "set args main"
> > +}
Hmm, this actually looks like a blind copy&paste from gdb.base/a2-run.exp
with the timeout bit added later on. I have therefore removed the whole
conditional -- it's not like we change the arguments anywhere here.
Here's an updated version -- any further comments?
2011-11-11 Maciej W. Rozycki <macro@mips.com>
Maciej W. Rozycki <macro@codesourcery.com>
Daniel Jacobowitz <dan@codesourcery.com>
* gdb.base/hbreak2.exp: New test for "hbreak" and "thbreak".
Maciej
gdb-test-hbreak.diff
Index: gdb-fsf-trunk-quilt/gdb/testsuite/gdb.base/hbreak2.exp
===================================================================
--- /dev/null 1970-01-01 00:00:00.000000000 +0000
+++ gdb-fsf-trunk-quilt/gdb/testsuite/gdb.base/hbreak2.exp 2011-11-11 19:47:24.495606633 +0000
@@ -0,0 +1,612 @@
+# Copyright 2011 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/>.
+
+# Based on break.exp by Rob Savoye. (rob@cygnus.com)
+# Hardware breakpoint support by Maciej W. Rozycki and Daniel Jacobowitz.
+# Only one hardware breakpoint is set at a time as targets may limit
+# the number available.
+
+
+if { [prepare_for_testing hbreak2.exp "break" {break.c break1.c} {debug nowarnings}] } {
+ return -1
+}
+set srcfile break.c
+set srcfile1 break1.c
+
+if ![runto_main] then { fail "break tests suppressed" }
+delete_breakpoints
+
+#
+# Test whether the target supports hardware breakpoints at all.
+#
+gdb_test_multiple "hbreak main" "hardware breakpoint support" {
+ -re "No hardware breakpoint support in the target.*$gdb_prompt $" {
+ unsupported "hardware breakpoints"
+ return
+ }
+ -re "Hardware breakpoints used exceeds limit.*$gdb_prompt $" {
+ unsupported "hardware breakpoints"
+ return
+ }
+ -re "Hardware assisted breakpoint.*at.* file .*$srcfile, line.*$gdb_prompt $" {
+ pass "hardware breakpoint support"
+ }
+}
+gdb_run_cmd
+gdb_test_multiple "" "hardware breakpoint insertion" {
+ -re "Warning:\[\r\n\]+Cannot insert hardware breakpoint \[0-9\]+\.\[\r\n\]+Could not insert hardware breakpoints:\[\r\n\]+You may have requested too many hardware breakpoints/watchpoints\.\[\r\n\]+.*$gdb_prompt $" {
+ unsupported "hardware breakpoint insertion"
+ return
+ }
+ -re "Breakpoint \[0-9\]+,.*main .*argc.*argv.* at .*$srcfile:.*\[\r\n\]+.*\[\t \]+if .argc.* \{.*$gdb_prompt $" {
+ pass "hardware breakpoint insertion"
+ }
+}
+delete_breakpoints
+
+#
+# Test simple hardware breakpoint setting commands.
+#
+
+#
+# Test break at function.
+#
+gdb_test "hbreak main" \
+ "Hardware assisted breakpoint.*at.* file .*$srcfile, line.*" \
+ "hardware breakpoint function"
+delete_breakpoints
+
+#
+# Test break at quoted function.
+#
+gdb_test "hbreak \"marker2\"" \
+ "Hardware assisted breakpoint.*at.* file .*$srcfile1, line.*" \
+ "hardware breakpoint quoted function"
+delete_breakpoints
+
+#
+# Test break at function in file.
+#
+gdb_test "hbreak $srcfile:factorial" \
+ "Hardware assisted breakpoint.*at.* file .*$srcfile, line.*" \
+ "hardware breakpoint function in file"
+delete_breakpoints
+
+set bp_location1 [gdb_get_line_number "set breakpoint 1 here"]
+
+#
+# Test break at line number.
+#
+# Note that the default source file is the last one whose source text
+# was printed. For native debugging, before we've executed the
+# program, this is the file containing main, but for remote debugging,
+# it's wherever the processor was stopped when we connected to the
+# board. So, to be sure, we do a list command.
+#
+gdb_test "list main" \
+ ".*main \\(argc, argv, envp\\).*" \
+ "use `list' to establish default source file"
+gdb_test "hbreak $bp_location1" \
+ "Hardware assisted breakpoint.*at.* file .*$srcfile, line $bp_location1\\." \
+ "hardware breakpoint line number"
+delete_breakpoints
+
+set bp_location2 [gdb_get_line_number "set breakpoint 2 here"]
+
+#
+# Test break at line number in file.
+#
+gdb_test "hbreak $srcfile:$bp_location2" \
+ "Hardware assisted breakpoint.*at.* file .*$srcfile, line $bp_location2\\." \
+ "hardware breakpoint line number in file"
+delete_breakpoints
+
+set bp_location3 [gdb_get_line_number "set breakpoint 3 here"]
+set bp_location4 [gdb_get_line_number "set breakpoint 4 here"]
+
+#
+# Test putting a break at the start of a multi-line if conditional.
+# Verify the breakpoint was put at the start of the conditional.
+#
+gdb_test "hbreak multi_line_if_conditional" \
+ "Hardware assisted breakpoint.*at.* file .*$srcfile, line $bp_location3\\." \
+ "hardware breakpoint at start of multi line if conditional"
+delete_breakpoints
+
+gdb_test "hbreak multi_line_while_conditional" \
+ "Hardware assisted breakpoint.*at.* file .*$srcfile, line $bp_location4\\." \
+ "hardware breakpoint at start of multi line while conditional"
+
+set bp_location5 [gdb_get_line_number "set breakpoint 5 here"]
+set bp_location6 [gdb_get_line_number "set breakpoint 6 here"]
+
+#
+# Check to see what breakpoints are set.
+#
+if [target_info exists gdb_stub] {
+ set main_line $bp_location5
+} else {
+ set main_line $bp_location6
+}
+
+if {$hp_aCC_compiler} {
+ set proto "\\(int\\)"
+} else {
+ set proto ""
+}
+
+set bp_location7 [gdb_get_line_number "set breakpoint 7 here"]
+set bp_location8 [gdb_get_line_number "set breakpoint 8 here" $srcfile1]
+set bp_location9 [gdb_get_line_number "set breakpoint 9 here" $srcfile1]
+
+gdb_test "info break" \
+ "Num Type\[ \]+Disp Enb Address\[ \]+What.*
+\[0-9\]+\[\t \]+hw breakpoint keep y.* in multi_line_while_conditional at .*$srcfile:$bp_location4" \
+ "hardware breakpoint info"
+delete_breakpoints
+
+#
+# Run until the breakpoint at main is hit. For non-stubs-using targets.
+#
+gdb_test "hbreak main" \
+ "Hardware assisted breakpoint.*at.* file .*$srcfile, line.*" \
+ "hardware breakpoint function (2)"
+gdb_run_cmd
+gdb_test "" \
+ "Breakpoint \[0-9\]+,.*main .*argc.*argv.* at .*$srcfile:$bp_location6.*$bp_location6\[\t \]+if .argc.* \{.*" \
+ "run until function breakpoint"
+delete_breakpoints
+
+#
+# Run until the breakpoint at a line number.
+#
+gdb_test "hbreak $bp_location1" \
+ "Hardware assisted breakpoint.*at.* file .*$srcfile, line $bp_location1\\." \
+ "hardware breakpoint line number (2)"
+gdb_test "continue" \
+ "Continuing\\..*Breakpoint \[0-9\]+, main \\(argc=.*, argv=.*, envp=.*\\) at .*$srcfile:$bp_location1.*$bp_location1\[\t \]+printf.*factorial.*" \
+ "run until breakpoint set at a line number"
+delete_breakpoints
+
+#
+# Run until the breakpoint set in a function in a file.
+#
+gdb_test "hbreak $srcfile:factorial" \
+ "Hardware assisted breakpoint.*at.* file .*$srcfile, line.*" \
+ "hardware breakpoint function in file (2)"
+for {set i 6} {$i >= 1} {incr i -1} {
+ gdb_test "continue" \
+ "Continuing\\..*Breakpoint \[0-9\]+, factorial \\(value=$i\\) at .*$srcfile:$bp_location7.*$bp_location7\[\t \]+.*if .value > 1. \{.*" \
+ "run until file:function($i) breakpoint"
+}
+delete_breakpoints
+
+#
+# Run until the breakpoint set at a quoted function.
+#
+gdb_test "hbreak \"marker2\"" \
+ "Hardware assisted breakpoint.*at.* file .*$srcfile1, line.*" \
+ "hardware breakpoint quoted function (2)"
+gdb_test "continue" \
+ "Continuing\\..*Breakpoint \[0-9\]+, (0x\[0-9a-f\]+ in )?marker2 \\(a=43\\) at .*$srcfile1:($bp_location8|$bp_location9).*" \
+ "run until quoted breakpoint"
+delete_breakpoints
+#
+# Run until the file:function breakpoint at a line number in a file.
+#
+gdb_test "hbreak $srcfile:$bp_location2" \
+ "Hardware assisted breakpoint.*at.* file .*$srcfile, line $bp_location2\\." \
+ "hardware breakpoint line number in file (2)"
+gdb_test "continue" \
+ "Continuing\\..*Breakpoint \[0-9\]+, main \\(argc=.*, argv=.*, envp=.*\\) at .*$srcfile:$bp_location2.*$bp_location2\[\t \]+argc = \\(argc == 12345\\);.*" \
+ "run until file:linenum breakpoint"
+delete_breakpoints
+
+# Test break at offset +1.
+set bp_location10 [gdb_get_line_number "set breakpoint 10 here"]
+
+gdb_test "hbreak +1" \
+ "Hardware assisted breakpoint.*at.* file .*$srcfile, line $bp_location10\\." \
+ "hardware breakpoint offset +1"
+
+# Check to see if breakpoint is hit when stepped onto.
+
+gdb_test "step" \
+ ".*Breakpoint \[0-9\]+, main \\(argc=.*, argv=.*, envp=.*\\) at .*$srcfile:$bp_location10.*$bp_location10\[\t \]+return argc;.*breakpoint 10 here.*" \
+ "step onto hardware breakpoint"
+delete_breakpoints
+
+# Check to see if breakpoint can be set on ending brace of function.
+set bp_location10a [gdb_get_line_number "set breakpoint 10a here"]
+
+gdb_test "hbreak $bp_location10a" \
+ "Hardware assisted breakpoint.*at.* file .*$srcfile, line $bp_location10a\\." \
+ "setting hardware breakpoint at }"
+
+gdb_test "continue" \
+ ".*Breakpoint \[0-9\]+, main \\(argc=.*, argv=.*, envp=.*\\) at .*$srcfile:$bp_location10a.*$bp_location10a\[\t \]+}.*breakpoint 10a here.*" \
+ "continue to hardware breakpoint at }"
+
+#
+# Delete all breakpoints so we can start over, course this can be a test too.
+#
+delete_breakpoints
+
+#
+# Test temporary breakpoint at function.
+#
+
+gdb_test "thbreak main" \
+ "Hardware assisted breakpoint.*at.* file .*$srcfile, line.*" \
+ "temporary hardware breakpoint function"
+delete_breakpoints
+
+#
+# Test break at function in file.
+#
+
+gdb_test "thbreak $srcfile:factorial" \
+ "Hardware assisted breakpoint.*at.* file .*$srcfile, line.*" \
+ "temporary hardware breakpoint function in file"
+delete_breakpoints
+
+#
+# Test break at line number.
+#
+gdb_test "thbreak $bp_location1" \
+ "Hardware assisted breakpoint.*at.* file .*$srcfile, line $bp_location1.*" \
+ "temporary hardware breakpoint line number #1"
+delete_breakpoints
+
+gdb_test "thbreak $bp_location6" \
+ "Hardware assisted breakpoint.*at.* file .*$srcfile, line $bp_location6.*" \
+ "temporary hardware breakpoint line number #2"
+delete_breakpoints
+
+#
+# Test break at line number in file.
+#
+gdb_test "thbreak $srcfile:$bp_location2" \
+ "Hardware assisted breakpoint.*at.* file .*$srcfile, line $bp_location2.*" \
+ "temporary hardware breakpoint line number in file #1"
+delete_breakpoints
+
+set bp_location11 [gdb_get_line_number "set breakpoint 11 here"]
+gdb_test "thbreak $srcfile:$bp_location11" \
+ "Hardware assisted breakpoint.*at.* file .*$srcfile, line $bp_location11.*" \
+ "temporary hardware breakpoint line number in file #2"
+
+#
+# Check to see what breakpoints are set (temporary this time).
+#
+gdb_test "info break" \
+ "Num Type.*Disp Enb Address.*What.*\[\r\n\]
+\[0-9\]+\[\t \]+hw breakpoint del.*y.*in main at .*$srcfile:$bp_location11.*" \
+ "temporary hardware breakpoint info"
+
+
+#***********
+
+if ![runto_main] then { fail "break tests suppressed" }
+
+# Verify that GDB responds gracefully when asked to set a breakpoint
+# on a nonexistent source line.
+#
+gdb_test "hbreak 999" \
+ "No line 999 in file .*" \
+ "hardware break on non-existent source line"
+
+# Run to the desired default location. If not positioned here, the
+# tests below don't work.
+#
+gdb_test "until $bp_location1" "main .* at .*:$bp_location1.*" \
+ "until bp_location1"
+
+
+# Verify that GDB allows one to just say "hbreak", which is treated
+# as the "default" breakpoint.
+#
+gdb_test "hbreak" "Hardware assisted breakpoint \[0-9\]*.*" \
+ "hardware break on default location"
+
+# Verify that a "silent" breakpoint can be set, and that GDB is indeed
+# "silent" about its triggering.
+#
+if ![runto_main] then { fail "break tests suppressed" }
+
+gdb_test_multiple "hbreak $bp_location1" \
+ "set to-be-silent hardware break bp_location1" {
+ -re "Hardware assisted breakpoint (\[0-9\]*) at .*, line $bp_location1.*$gdb_prompt $" {
+ pass "set to-be-silent hardware break bp_location1"
+ }
+ }
+
+send_gdb "commands $expect_out(1,string)\n"
+send_gdb "silent\n"
+send_gdb "end\n"
+gdb_test "" "set silent hardware break bp_location1"
+
+gdb_test "info break $expect_out(1,string)" \
+ "\[0-9\]*\[ \t\]*hw breakpoint.*:$bp_location1\r\n\[ \t\]*silent.*" \
+ "info silent hardware break bp_location1"
+
+gdb_test "continue" "Continuing." \
+ "hit silent hardware break bp_location1"
+
+gdb_test "bt" "#0 main .* at .*:$bp_location1.*" \
+ "stopped for silent hardware break bp_location1"
+
+# Verify that GDB can at least parse a breakpoint with the
+# "thread" keyword. (We won't attempt to test here that a
+# thread-specific breakpoint really triggers appropriately.
+# The gdb.threads subdirectory contains tests for that.)
+#
+set bp_location12 [gdb_get_line_number "set breakpoint 12 here"]
+gdb_test "hbreak $bp_location12 thread 999" "Unknown thread 999.*" \
+ "thread-specific hardware breakpoint on non-existent thread disallowed"
+
+gdb_test "hbreak $bp_location12 thread foo" \
+ "Junk after thread keyword.*" \
+ "thread-specific hardware breakpoint on bogus thread ID disallowed"
+
+# Verify that GDB responds gracefully to a breakpoint command with
+# trailing garbage.
+#
+gdb_test "hbreak $bp_location12 foo" \
+ "Junk at end of arguments.*" \
+ "hardware breakpoint with trailing garbage disallowed"
+
+# Verify that GDB responds gracefully to a "clear" command that has
+# no matching breakpoint. (First, get us off the current source line,
+# which we know has a breakpoint.)
+#
+gdb_test "next" "marker1.*" "step over hardware breakpoint"
+
+gdb_test "clear 81" "No breakpoint at 81.*" \
+ "clear line has no breakpoint disallowed"
+
+gdb_test "clear" "No breakpoint at this line.*" \
+ "clear current line has no breakpoint disallowed"
+delete_breakpoints
+
+# Verify that a breakpoint can be set via a convenience variable.
+#
+gdb_test_no_output "set \$foo=$bp_location11" \
+ "set convenience variable \$foo to bp_location11"
+
+gdb_test "hbreak \$foo" \
+ "Hardware assisted breakpoint (\[0-9\]*) at .*, line $bp_location11.*" \
+ "set hardware breakpoint via convenience variable"
+delete_breakpoints
+
+# Verify that GDB responds gracefully to an attempt to set a
+# breakpoint via a convenience variable whose type is not integer.
+#
+gdb_test_no_output "set \$foo=81.5" \
+ "set convenience variable \$foo to 81.5"
+
+gdb_test "hbreak \$foo" \
+ "Convenience variables used in line specs must have integer values.*" \
+ "set hardware breakpoint via non-integer convenience variable disallowed"
+
+# Verify that we can set and trigger a breakpoint in a user-called function.
+#
+gdb_test "hbreak marker2" \
+ "Hardware assisted breakpoint (\[0-9\]*) at .*, line ($bp_location8|$bp_location9).*" \
+ "set hardware breakpoint on to-be-called function"
+
+gdb_test "print marker2(99)" \
+ "The program being debugged stopped while in a function called from GDB.\r\nEvaluation of the expression containing the function\r\n.marker2$proto. will be abandoned.\r\nWhen the function is done executing, GDB will silently stop.*" \
+ "hit hardware breakpoint on called function"
+
+# As long as we're stopped (breakpointed) in a called function,
+# verify that we can successfully backtrace & such from here.
+#
+# In this and the following test, the _sr4export check apparently is needed
+# for hppa*-*-hpux.
+#
+gdb_test_multiple "bt" "backtrace while in called function" {
+ -re "#0\[ \t\]*($hex in )?marker2.*:($bp_location8|$bp_location9)\r\n#1.*_sr4export.*$gdb_prompt $" {
+ pass "backtrace while in called function"
+ }
+ -re "#0\[ \t\]*($hex in )?marker2.*:($bp_location8|$bp_location9)\r\n#1.*function called from gdb.*$gdb_prompt $" {
+ pass "backtrace while in called function"
+ }
+}
+
+# Return from the called function. For remote targets, it's important to do
+# this before runto_main, which otherwise may silently stop on the dummy
+# breakpoint inserted by GDB at the program's entry point.
+#
+gdb_test_multiple "finish" "finish from called function" {
+ -re "Run till exit from .*marker2.* at .*($bp_location8|$bp_location9)\r\n.* in _sr4export.*$gdb_prompt $" {
+ pass "finish from called function"
+ }
+ -re "Run till exit from .*marker2.* at .*($bp_location8|$bp_location9)\r\n.*function called from gdb.*$gdb_prompt $" {
+ pass "finish from called function"
+ }
+ -re "Run till exit from .*marker2.* at .*($bp_location8|$bp_location9)\r\n.*Value returned.*$gdb_prompt $" {
+ pass "finish from called function"
+ }
+}
+
+#********
+
+
+#
+# Test "next" over recursive function call.
+#
+
+proc test_next_with_recursion {} {
+ global gdb_prompt
+ global decimal
+ global binfile
+
+ delete_breakpoints
+
+ # Can't set a hardware breakpoint without a live target, so do it now
+ # before it's killed below.
+ gdb_test "hbreak factorial" \
+ "Hardware assisted breakpoint $decimal at .*" \
+ "hardware break at factorial"
+
+ gdb_test "kill" "" "kill program" \
+ "Kill the program being debugged.*y or n. $" "y"
+
+ # Run until we call factorial with 6
+
+ gdb_run_cmd
+ if [gdb_test "" \
+ "Break.* factorial .value=6. .*" \
+ "run to factorial(6)"] then { gdb_suppress_tests }
+
+ # Continue until we call factorial recursively with 5.
+
+ if [gdb_test "continue" \
+ "Continuing.*Break.* factorial .value=5. .*" \
+ "continue to factorial(5)"] then { gdb_suppress_tests }
+
+ # Do a backtrace just to confirm how many levels deep we are.
+
+ if [gdb_test "backtrace" \
+ "#0\[ \t\]+ factorial .value=5..*" \
+ "backtrace from factorial(5)"] then { gdb_suppress_tests }
+
+ # Now a "next" should position us at the recursive call, which
+ # we will be performing with 4.
+
+ if [gdb_test "next" \
+ ".* factorial .value - 1.;.*" \
+ "next to recursive call"] then { gdb_suppress_tests }
+
+ # Disable the breakpoint at the entry to factorial by deleting them all.
+ # The "next" should run until we return to the next line from this
+ # recursive call to factorial with 4.
+ # Buggy versions of gdb will stop instead at the innermost frame on
+ # the line where we are trying to "next" to.
+
+ delete_breakpoints
+
+ if [istarget "mips*tx39-*"] {
+ set timeout 60
+ }
+ # We used to set timeout here for all other targets as well. This
+ # is almost certainly wrong. The proper timeout depends on the
+ # target system in use, and how we communicate with it, so there
+ # is no single value appropriate for all targets. The timeout
+ # should be established by the Dejagnu config file(s) for the
+ # board, and respected by the test suite.
+ #
+ # For example, if I'm running GDB over an SSH tunnel talking to a
+ # portmaster in California talking to an ancient 68k board running
+ # a crummy ROM monitor (a situation I can only wish were
+ # hypothetical), then I need a large timeout. But that's not the
+ # kind of knowledge that belongs in this file.
+
+ gdb_test next "\[0-9\]*\[\t \]+return \\(value\\);.*" \
+ "next over recursive call"
+
+ # OK, we should be back in the same stack frame we started from.
+ # Do a backtrace just to confirm.
+
+ set result [gdb_test "backtrace" \
+ "#0\[ \t\]+ factorial .value=120.*\r\n#1\[ \t\]+ \[0-9a-fx\]+ in factorial .value=6..*" \
+ "backtrace from factorial(5.1)"]
+ if { $result != 0 } { gdb_suppress_tests }
+
+ if [target_info exists gdb,noresults] { gdb_suppress_tests }
+ gdb_continue_to_end "recursive next test"
+ gdb_stop_suppressing_tests;
+}
+
+test_next_with_recursion
+
+
+#********
+
+# Build a new file with optimization enabled so that we can try breakpoints
+# on targets with optimized prologues.
+
+if { [prepare_for_testing hbreak2.exp "breako2" {break.c break1.c} {debug nowarnings optimize=-O2}] } {
+ return -1
+}
+
+if ![runto_main] then { fail "break tests suppressed" }
+delete_breakpoints
+
+#
+# Test break at function.
+#
+gdb_test "hbreak main" \
+ "Hardware assisted breakpoint.*at.* file .*, line.*" \
+ "hardware breakpoint function, optimized file"
+
+#
+# Run until the breakpoint at main is hit. For non-stubs-using targets.
+#
+gdb_run_cmd
+gdb_test_multiple "" "run until hardware function breakpoint, optimized file" {
+ -re "Breakpoint \[0-9\]+,.*main .*argc.*argv.* at .*$srcfile:$bp_location6.*$bp_location6\[\t \]+if .argc.* \{.*$gdb_prompt $" {
+ pass "run until hardware function breakpoint, optimized file"
+ }
+ -re "Breakpoint \[0-9\]+,.*main .*argc.*argv.* at .*$gdb_prompt $" {
+ pass "run until hardware function breakpoint, optimized file (code motion)"
+ }
+}
+delete_breakpoints
+
+#
+# Test break at function.
+#
+gdb_test "hbreak marker4" \
+ "Hardware assisted breakpoint.*at.* file .*$srcfile1, line.*" \
+ "hardware breakpoint small function, optimized file"
+
+#
+# Run until the breakpoint at a small function.
+#
+
+#
+# Add a second pass pattern. The behavior differs here between stabs
+# and dwarf for one-line functions. Stabs preserves two line symbols
+# (one before the prologue and one after) with the same line number,
+# but dwarf regards these as duplicates and discards one of them.
+# Therefore the address after the prologue (where the breakpoint is)
+# has no exactly matching line symbol, and GDB reports the breakpoint
+# as if it were in the middle of a line rather than at the beginning.
+
+set bp_location13 [gdb_get_line_number "set breakpoint 13 here" $srcfile1]
+set bp_location14 [gdb_get_line_number "set breakpoint 14 here" $srcfile1]
+
+gdb_test_multiple "continue" \
+ "run until hardware breakpoint set at small function, optimized file" {
+ -re "Breakpoint $decimal, marker4 \\(d=(d@entry=)?177601976\\) at .*$srcfile1:$bp_location13\[\r\n\]+$bp_location13\[\t \]+void marker4.*" {
+ pass "run until hardware breakpoint set at small function, optimized file"
+ }
+ -re "Breakpoint $decimal, $hex in marker4 \\(d=(d@entry=)?177601976\\) at .*$srcfile1:$bp_location13\[\r\n\]+$bp_location13\[\t \]+void marker4.*" {
+ pass "run until hardware breakpoint set at small function, optimized file"
+ }
+ -re "Breakpoint $decimal, marker4 \\(d=(d@entry=)?177601976\\) at .*$srcfile1:$bp_location14\[\r\n\]+$bp_location14\[\t \]+void marker4.*" {
+ # marker4() is defined at line 46 when compiled with -DPROTOTYPES
+ pass "run until hardware breakpoint set at small function, optimized file (line bp_location14)"
+ }
+ -re "Breakpoint $decimal, factorial \\(.*\\) .*\{\r\n$gdb_prompt" {
+ # GCC 4.3 emits bad line number information - see gcc/36748.
+ if { [test_compiler_info "gcc-4-3-*"] } {
+ setup_xfail *-*-*
+ }
+ fail "run until hardware breakpoint set at small function, optimized file"
+ }
+}