From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 20087 invoked by alias); 5 Mar 2013 20:14:26 -0000 Received: (qmail 20042 invoked by uid 22791); 5 Mar 2013 20:14:25 -0000 X-SWARE-Spam-Status: No, hits=-6.5 required=5.0 tests=AWL,BAYES_00,KHOP_RCVD_UNTRUST,KHOP_SPAMHAUS_DROP,RCVD_IN_DNSWL_HI,RCVD_IN_HOSTKARMA_W,RP_MATCHES_RCVD,SPF_HELO_PASS X-Spam-Check-By: sourceware.org Received: from mx1.redhat.com (HELO mx1.redhat.com) (209.132.183.28) by sourceware.org (qpsmtpd/0.43rc1) with ESMTP; Tue, 05 Mar 2013 20:14:06 +0000 Received: from int-mx01.intmail.prod.int.phx2.redhat.com (int-mx01.intmail.prod.int.phx2.redhat.com [10.5.11.11]) by mx1.redhat.com (8.14.4/8.14.4) with ESMTP id r25KE4o8014986 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=OK); Tue, 5 Mar 2013 15:14:04 -0500 Received: from host2.jankratochvil.net (ovpn-116-50.ams2.redhat.com [10.36.116.50]) by int-mx01.intmail.prod.int.phx2.redhat.com (8.13.8/8.13.8) with ESMTP id r25KDqJS002149 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES128-SHA bits=128 verify=NO); Tue, 5 Mar 2013 15:13:54 -0500 Date: Tue, 05 Mar 2013 20:14:00 -0000 From: Jan Kratochvil To: Markus Metzger Cc: gdb-patches@sourceware.org, markus.t.metzger@gmail.com, Christian Himpel Subject: Re: [patch v9 22/23] testsuite, gdb.btrace: add btrace tests Message-ID: <20130305201351.GV2386@host2.jankratochvil.net> References: <1362416770-19750-1-git-send-email-markus.t.metzger@intel.com> <1362416770-19750-23-git-send-email-markus.t.metzger@intel.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <1362416770-19750-23-git-send-email-markus.t.metzger@intel.com> User-Agent: Mutt/1.5.21 (2010-09-15) X-IsSubscribed: yes Mailing-List: contact gdb-patches-help@sourceware.org; run by ezmlm Precedence: bulk List-Id: List-Subscribe: List-Archive: List-Post: List-Help: , Sender: gdb-patches-owner@sourceware.org X-SW-Source: 2013-03/txt/msg00186.txt.bz2 On Mon, 04 Mar 2013 18:06:09 +0100, Markus Metzger wrote: > From: Christian Himpel > > Recursive function calls are listed as only one function call. This is > marked as an XFAIL. > > 2013-03-04 Christian Himpel > > gdb/testsuite > * Makefile.in: Add btrace testsuite. > * configure: Regenerated. > * configure.ac: Add btrace testsuite. > * gdb.btrace/Makefile.in: New file. > * gdb.btrace/enable.c: New file. > * gdb.btrace/enable.exp: New file. > * gdb.btrace/function_call_history.c: New file. > * gdb.btrace/function_call_history.exp: New file. > * gdb.btrace/instruction_history.c: New file. > * gdb.btrace/instruction_history.exp: New file. > * gdb.btrace/x86-instruction_history.S: New file. > * lib/gdb.exp: Add btrace skip proc. [...] > --- /dev/null > +++ b/gdb/testsuite/gdb.btrace/enable.exp > @@ -0,0 +1,83 @@ > +# This testcase is part of GDB, the GNU debugger. > +# > +# Copyright 2013 Free Software Foundation, Inc. > +# > +# Contributed by Intel Corp. > +# > +# 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 . > + > +# check for btrace support > +if { [skip_btrace_tests] } { return -1 } > + > +# start fresh - without an executable > +gdb_exit > +gdb_start > + > +# record cannot be stopped, if it was never active > +gdb_test "record stop" "No record target is currently active\..*" "record stop without target" \ -> \\ First processing is by TCL, second processing is by regex. This is regex backslash. This way it acts the same as: gdb_test "record stop" "No record target is currently active..*" "record stop without target" One can also use {...} instead of "..." which avoids the Tcl processing, it would be then: gdb_test "record stop" {No record target is currently active\.} "record stop without target" And that trailing .* is not needed there, gdb_test automatically matches trailing "\r\n$gdb_prompt $". > + > +# btrace cannot be enabled without a running inferior > +gdb_test "record btrace" "The program is not being run\." "record btrace without running program" Again \ -> \\ . > + > +# no function and no instruction history without btrace enabled > +gdb_test "record function-call-history" "No record target is currently active\..*" "record function-call-history without target" \..* -> \\. > +gdb_test "record instruction-history" "No record target is currently active\..*" "record instruction-history without target" \..* -> \\. > +gdb_test "info record" "No record target is currently active\..*" "info record without target" \..* -> \\. > + > +# start inferior > +standard_testfile standard_testfile should be at the top of file; it does not start/contol GDB any way. > +if [prepare_for_testing $testfile.exp $testfile {} {debug}] { > + return -1 > +} > +runto_main if ![runto_main] { return -1 } > + > +# enable btrace > +gdb_test_no_output "record btrace" "record btrace" > +gdb_test "record function-call-history" "No trace\." "record function-call-history without trace" \ -> \\ > +gdb_test "record instruction-history" "No trace\." "record instruction-history without trace" \ -> \\ > + > +# btrace cannot be enabled twice > +gdb_test "record btrace" "The process is already being recorded\." "record btrace the second time" \ -> \\ > + > +# full record cannot be activated as long as btrace is active > +gdb_test "record full" "Process record target already running\. Use \"record stop\" to stop record target first\." "record full cannot be enabled" \. -> \\. twice > + > +# no trace recorded yet > +gdb_test "info record" "Active record target: record-btrace\r\nRecorded 0 instructions in 0 functions for thread 1.*\." "info record without trace" \. -> \\. Many times below again and in the other *.exp files. > + > +# stop btrace record > +gdb_test "record stop" "Process record is stopped and all execution logs are deleted\." "record stop" > +gdb_test "record stop" "No record target is currently active\..*" "record stop the second time" > + > +# enable btrace again > +gdb_test_no_output "record btrace" "record btrace re-enable" > +gdb_test "record btrace" "The process is already being recorded\." "record btrace re-enable twice" > + > +# rerun > +send_gdb "run\n" > +gdb_expect { > + -re "The program .* has been started already.*y or n. $" { > + send_gdb "y\n" > + exp_continue > + } > + -re "Starting program.*$gdb_prompt $" { > + pass "rerun to main" > + } > + -re "$gdb_prompt $" { > + fail "rerun to main" > + } > + timeout { > + fail "(timeout) rerun to main" > + } > +} In general one should never use gdb_expect. And send_gdb is used only exceptionally. For complicated cases there is gdb_test_multiple. But in this case why there isn't just another runto_main? if ![runto_main] { return -1 } > diff --git a/gdb/testsuite/gdb.btrace/function_call_history.c b/gdb/testsuite/gdb.btrace/function_call_history.c > new file mode 100644 > index 0000000..a76b7c3 > --- /dev/null > +++ b/gdb/testsuite/gdb.btrace/function_call_history.c > @@ -0,0 +1,45 @@ > +/* This testcase is part of GDB, the GNU debugger. > + > + Copyright 2013 Free Software Foundation, Inc. > + > + Contributed by Intel Corp. > + > + 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 . */ > + > +int > +inc (int i) > +{ > + return i+1; > +} > + > +int > +fib (int n) > +{ > + if (n <= 1) > + return n; > + > + return fib(n-2) + fib(n-1); > +} > + > +int > +main (void) > +{ > + int i, j; > + > + for (i = 0; i < 10; i++) > + j += inc(i); > + > + j += fib(3); /* bp.1 */ > + return j; /* bp.2 */ > +} > diff --git a/gdb/testsuite/gdb.btrace/function_call_history.exp b/gdb/testsuite/gdb.btrace/function_call_history.exp > new file mode 100644 > index 0000000..711dd20 > --- /dev/null > +++ b/gdb/testsuite/gdb.btrace/function_call_history.exp > @@ -0,0 +1,211 @@ > +# This testcase is part of GDB, the GNU debugger. > +# > +# Copyright 2013 Free Software Foundation, Inc. > +# > +# Contributed by Intel Corp. > +# > +# 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 . > + > +# check for btrace support > +if { [skip_btrace_tests] } { return -1 } > + > +# start inferior > +standard_testfile > +if [prepare_for_testing function_call_history.exp $testfile {} {debug}] { > + return -1 > +} > +runto_main if ![runto_main] { return -1 } > + > +# start btrace > +gdb_test_no_output "record btrace" > + > +# set bp after increment loop and continue > +set bp_location [gdb_get_line_number "bp.1" $testfile.c] > +gdb_breakpoint $bp_location > +gdb_continue_to_breakpoint "cont to $bp_location" ".*$testfile.c:$bp_location.*" > + > +# show function call history with unlimited size, we expect to see all 21 entries > +gdb_test_no_output "set record function-call-history-size 0" > +gdb_test "record function-call-history" "0 main\r > +1 inc\r > +2 main\r > +3 inc\r > +4 main\r > +5 inc\r > +6 main\r > +7 inc\r > +8 main\r > +9 inc\r > +10 main\r > +11 inc\r > +12 main\r > +13 inc\r > +14 main\r > +15 inc\r > +16 main\r > +17 inc\r > +18 main\r > +19 inc\r > +20 main\r" "record function-call-history - with size unlimited" > + > +# show function call history with size of 21, we expect to see all 21 entries > +gdb_test_no_output "set record function-call-history-size 21" > +# show function call history > +gdb_test "record function-call-history 0" "0 main\r > +1 inc\r > +2 main\r > +3 inc\r > +4 main\r > +5 inc\r > +6 main\r > +7 inc\r > +8 main\r > +9 inc\r > +10 main\r > +11 inc\r > +12 main\r > +13 inc\r > +14 main\r > +15 inc\r > +16 main\r > +17 inc\r > +18 main\r > +19 inc\r > +20 main\r" "record function-call-history - show all 21 entries" > + > +# show first 15 entries > +gdb_test_no_output "set record function-call-history-size 15" > +gdb_test "record function-call-history 0" "0 main\r > +1 inc\r > +2 main\r > +3 inc\r > +4 main\r > +5 inc\r > +6 main\r > +7 inc\r > +8 main\r > +9 inc\r > +10 main\r > +11 inc\r > +12 main\r > +13 inc\r > +14 main\r" "record function-call-history - show first 15 entries" > + > +# show last 6 entries > +gdb_test "record function-call-history +" "15 inc\r > +16 main\r > +17 inc\r > +18 main\r > +19 inc\r > +20 main\r" "record function-call-history - show last 6 entries" > + > +# moving further should not work > +gdb_test "record function-call-history +" "At the end of the branch trace record\." "record function-call-history - at the end (1)" > + > +# make sure we cannot move any further a second time > +gdb_test "record function-call-history +" "At the end of the branch trace record\." "record function-call-history - at the end (2)" > + > +# moving back showing the latest 15 function calls > +gdb_test "record function-call-history -" "6 main\r > +7 inc\r > +8 main\r > +9 inc\r > +10 main\r > +11 inc\r > +12 main\r > +13 inc\r > +14 main\r > +15 inc\r > +16 main\r > +17 inc\r > +18 main\r > +19 inc\r > +20 main\r" "record function-call-history - show last 15 entries" > + > +# moving further back shows the 6 first function calls > +gdb_test "record function-call-history -" "0 main\r > +1 inc\r > +2 main\r > +3 inc\r > +4 main\r > +5 inc\r" "record function-call-history - show first 6 entries" > + > +# moving further back shouldn't work > +gdb_test "record function-call-history -" "At the start of the branch trace record\." "record function-call-history - at the start (1)" > + > +# make sure we cannot move any further back > +gdb_test "record function-call-history -" "At the start of the branch trace record\." "record function-call-history - at the start (2)" > + > +# moving forward again, but this time with file and line number, expected to see the first 15 entries > +gdb_test "record function-call-history /l +" " > +.*$srcfile:40-41 main\r > +.*$srcfile:22-24 inc\r > +.*$srcfile:40-41 main\r > +.*$srcfile:22-24 inc\r > +.*$srcfile:40-41 main\r > +.*$srcfile:22-24 inc\r > +.*$srcfile:40-41 main\r > +.*$srcfile:22-24 inc\r > +.*$srcfile:40-41 main\r > +.*$srcfile:22-24 inc\r > +.*$srcfile:40-41 main\r > +.*$srcfile:22-24 inc\r > +.*$srcfile:40-41 main\r > +.*$srcfile:22-24 inc\r > +.*$srcfile:40-41 main\r" "record function-call-history /l - show first 15 entries" > + > +# moving forward and expect to see the latest 6 entries > +gdb_test "record function-call-history /l +" " > +.*$srcfile:22-24 inc\r > +.*$srcfile:40-41 main\r > +.*$srcfile:22-24 inc\r > +.*$srcfile:40-41 main\r > +.*$srcfile:22-24 inc\r > +.*$srcfile:40-43 main\r" "record function-call-history /l - show last 6 entries" > + > +# moving further forward shouldn't work > +gdb_test "record function-call-history /l +" "At the end of the branch trace record\." "record function-call-history /l - at the end (1)" > +gdb_test "record function-call-history /l" "At the end of the branch trace record\." "record function-call-history /l - at the end (2)" > + > +set expected_range "3 inc\r > +4 main\r > +5 inc\r > +6 main\r > +7 inc\r > +8 main\r > +9 inc\r" > + > +# show functions in instruction range > +gdb_test "record function-call-history 3,10" $expected_range "absolute instruction range" > +gdb_test "record function-call-history 3,+7" $expected_range "relative positive instruction range" > +gdb_test "record function-call-history 10,-7" $expected_range "relative negative instruction range" > + > +# set bp after fib recursion and continue > +set bp_location [gdb_get_line_number "bp.2" $testfile.c] > +gdb_breakpoint $bp_location > +gdb_continue_to_breakpoint "cont to $bp_location" ".*$testfile.c:$bp_location.*" > + > +# at this point we expect to have main, fib, ..., fib, main, where fib occurs 8 times, > +# so we limit the output to only show the latest 10 function calls > +gdb_test_no_output "set record function-call-history-size 10" > +set message "show recursive function call history" > +gdb_test_multiple "record function-call-history" $message { > + -re "13 main\r\n14 fib\r\n15 fib\r\n16 fib\r\n17 fib\r\n18 fib\r\n19 fib\r\n20 fib\r\n21 fib\r\n22 main\r\n$gdb_prompt " { > + pass $message > + } > + -re "13 inc\r\n14 main\r\n15 inc\r\n16 main\r\n17 inc\r\n18 main\r\n19 inc\r\n20 main\r\n21 fib\r\n22 main\r\n$gdb_prompt " { > + # recursive function calls appear only as 1 call > + xfail $message kfail, as already discussed before. A nitpick but there can (and therefore should) be trailing $ match so that no more accidental input is accepted there, therefore in both cases: main\r\n$gdb_prompt " { -> main\r\n$gdb_prompt $" { > + } > +} > diff --git a/gdb/testsuite/gdb.btrace/instruction_history.c b/gdb/testsuite/gdb.btrace/instruction_history.c > new file mode 100644 > index 0000000..2dc3c6b > --- /dev/null > +++ b/gdb/testsuite/gdb.btrace/instruction_history.c > @@ -0,0 +1,26 @@ > +/* This testcase is part of GDB, the GNU debugger. > + > + Copyright 2013 Free Software Foundation, Inc. > + > + Contributed by Intel Corp. > + > + 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 . */ > + > +int > +main (void) > +{ > + // loop is declared in x86-instruction_history.S Rather /* comments */ ; although one will probably never use such new feature with old compilers not supporting // for .c . > + loop (); > + return 0; > +} > diff --git a/gdb/testsuite/gdb.btrace/instruction_history.exp b/gdb/testsuite/gdb.btrace/instruction_history.exp > new file mode 100644 > index 0000000..92288eb > --- /dev/null > +++ b/gdb/testsuite/gdb.btrace/instruction_history.exp > @@ -0,0 +1,190 @@ > +# This testcase is part of GDB, the GNU debugger. > +# > +# Copyright 2013 Free Software Foundation, Inc. > +# > +# Contributed by Intel Corp. > +# > +# 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 . > + > +# check for btrace support > +if { [skip_btrace_tests] } { return -1 } > + > +# compile and run to main > +standard_testfile > +if [prepare_for_testing instruction_history.exp $testfile "$testfile.c x86-$testfile.S" {debug}] { standard_testfile has parameters for this purpose: standard_testfile .c x86-instruction_history.S if [prepare_for_testing $testfile.exp $testfile "$srcfile $srcfile2" {debug}] { And please rename x86-instruction_history.S to instruction_history-x86.S so that completion works for all the files belonging to a testcase. > + return -1 > +} > + > +runto_main if ![runto_main] { return -1 } > + > +# set bp before loop and continue > +set bp_location [gdb_get_line_number "bp.1" x86-$testfile.S] > +gdb_breakpoint x86-$testfile.S:$bp_location > +gdb_continue_to_breakpoint "cont to $bp_location" ".*x86-$testfile.S:$bp_location.*" > + > +# start btrace > +gdb_test_no_output "record btrace" > + > +# set bp after loop and continue > +set bp_location [gdb_get_line_number "bp.2" x86-$testfile.S] > +gdb_breakpoint x86-$testfile.S:$bp_location > +gdb_continue_to_breakpoint "cont to $bp_location" ".*x86-$testfile.S:$bp_location.*" > + > +# The following test cases test if "browsing" through the > +# instruction history works as expected. So for the tests > +# it is necessary to count the number of lines that are > +# shown by the "record instruction-history" command. > + > +set testname "determine number of recorded instructions" > +gdb_test_multiple "info record" $testname { > + -re "Active record target: record-btrace\r\nRecorded \(\[0-9\]*\) instructions in \(\[0-9\]*\) functions for thread 1 .*\.\r\n$gdb_prompt " { Again trailing $gdb_prompt $" like in the other file. > + set traced $expect_out(1,string) > + set traced_functions $expect_out(2,string) > + pass $testname > + } > +} > + > +# we have exactly 7 instructions here > +set message "exactly 7 instructions" > +if { $traced != 7 } { > + fail $message > +} else { > + pass $message > +} > + > +# test that we see the expected instructions > +gdb_test "record instruction-history 1,6" "1\t 0x\[0-9a-f\]+ :\tje 0x\[0-9a-f\]+ \r It is OK as is but FYI one can do the escaping automatically at runtime for any string using 'string_to_regexp' function. > +2\t 0x\[0-9a-f\]+ :\tdec %eax\r > +3\t 0x\[0-9a-f\]+ :\tjmp 0x\[0-9a-f\]+ \r > +4\t 0x\[0-9a-f\]+ :\tcmp \\\$0x0,%eax\r > +5\t 0x\[0-9a-f\]+ :\tje 0x\[0-9a-f\]+ \r" > + > +gdb_test "record instruction-history /f 1,+5" "1\t 0x\[0-9a-f\]+ <\\+\[0-9\]+>:\tje 0x\[0-9a-f\]+ \r > +2\t 0x\[0-9a-f\]+ <\\+\[0-9\]+>:\tdec %eax\r > +3\t 0x\[0-9a-f\]+ <\\+\[0-9\]+>:\tjmp 0x\[0-9a-f\]+ \r > +4\t 0x\[0-9a-f\]+ <\\+\[0-9\]+>:\tcmp \\\$0x0,%eax\r > +5\t 0x\[0-9a-f\]+ <\\+\[0-9\]+>:\tje 0x\[0-9a-f\]+ \r" > + > +gdb_test "record instruction-history /p 6,-5" "1\t 0x\[0-9a-f\]+ :\tje 0x\[0-9a-f\]+ \r > +2\t 0x\[0-9a-f\]+ :\tdec %eax\r > +3\t 0x\[0-9a-f\]+ :\tjmp 0x\[0-9a-f\]+ \r > +4\t 0x\[0-9a-f\]+ :\tcmp \\\$0x0,%eax\r > +5\t 0x\[0-9a-f\]+ :\tje 0x\[0-9a-f\]+ \r" > + > +gdb_test "record instruction-history /pf 1,6" "1\t 0x\[0-9a-f\]+ <\\+\[0-9\]+>:\tje 0x\[0-9a-f\]+ \r > +2\t 0x\[0-9a-f\]+ <\\+\[0-9\]+>:\tdec %eax\r > +3\t 0x\[0-9a-f\]+ <\\+\[0-9\]+>:\tjmp 0x\[0-9a-f\]+ \r > +4\t 0x\[0-9a-f\]+ <\\+\[0-9\]+>:\tcmp \\\$0x0,%eax\r > +5\t 0x\[0-9a-f\]+ <\\+\[0-9\]+>:\tje 0x\[0-9a-f\]+ \r" > + > +# the following tests are checking the iterators > +# to avoid lots of regexps, we just check the number of lines that > +# were printed during command execution. > + > +# test_lines_output returns the output lines from command as a list. > +proc test_lines_output { command message } { > + global gdb_prompt > + set message "test_lines_output: $message" > + gdb_test_multiple $command $message { > + -re "\n\(.*\)\r\n$gdb_prompt " { trailing $ again. > + return [split [string trim $expect_out(1,string)] "\n"] > + } > + } > +} > + > +# test_lines_length returns the number of lines from command. > +proc test_lines_length { command message } { > + return [llength [test_lines_output $command $message]] > +} > + > +# show instruction history with unlimited size, we expect to see > +# all $traced instructions > +gdb_test_no_output "set record instruction-history-size 0" > +set message "record instruction-history - unlimited" > +set lines [test_lines_length "record instruction-history 0" $message] > +if { $traced != $lines } { > + fail $message > +} else { > + pass $message > +} > + > +gdb_test_no_output "set record instruction-history-size $traced" > +set message "record instruction-history - traced" > +set lines [test_lines_length "record instruction-history 0" $message] > +if { $traced != $lines } { > + fail $message > +} else { > + pass $message > +} > + > +# test that the iterator works > +set history_size 3 > +gdb_test_no_output "set record instruction-history-size $history_size" > +set message "browse history forward start" > +set lines [test_lines_length "record instruction-history 0" $message] > +if { $lines != $history_size } { > + fail $message > +} else { > + pass $message > +} > + > +set message "browse history forward middle" > +set lines [test_lines_length "record instruction-history +" $message] > +if { $lines != $history_size } { > + fail $message > +} else { > + pass $message > +} > + > +set message "browse history forward last" > +set lines [test_lines_length "record instruction-history +" $message] > +if { $lines != 1 } { > + fail $message > +} else { > + pass $message > +} > + > +gdb_test "record instruction-history" "At the end of the branch trace record\." "browse history forward beyond 1" > + > +# make sure we cannot move further > +gdb_test "record instruction-history" "At the end of the branch trace record\." "browse history forward beyond 2" > + > +set message "browse history backward last" > +set lines [test_lines_length "record instruction-history -" $message] > +if { $lines != $history_size } { > + fail $message > +} else { > + pass $message > +} > + > +set message "browse history backward middle" > +set lines [test_lines_length "record instruction-history -" $message] > +if { $lines != $history_size } { > + fail $message > +} else { > + pass $message > +} > + > +set message "browse history backward first" > +set lines [test_lines_length "record instruction-history -" $message] > +if { $lines != 1 } { > + fail $message > +} else { > + pass $message > +} > + > +gdb_test "record instruction-history -" "At the start of the branch trace record\." "browse history backward beyond 1" > + > +# make sure we cannot move further back > +gdb_test "record instruction-history -" "At the start of the branch trace record\." "browse history backward beyond 2" > diff --git a/gdb/testsuite/gdb.btrace/x86-instruction_history.S b/gdb/testsuite/gdb.btrace/x86-instruction_history.S > new file mode 100644 > index 0000000..5c74b46 > --- /dev/null > +++ b/gdb/testsuite/gdb.btrace/x86-instruction_history.S > @@ -0,0 +1,32 @@ > +/* This testcase is part of GDB, the GNU debugger. > + > + Copyright 2013 Free Software Foundation, Inc. > + > + Contributed by Intel Corp. > + > + 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 . */ > + > + .text > + .globl loop > + .type loop, @function > + > +loop: > + movl $0x2, %eax /* bp.1 */ > +.L1: > + cmpl $0, %eax > + je .L2 > + decl %eax > + jmp .L1 > +.L2: > + ret /* bp.2 */ > diff --git a/gdb/testsuite/lib/gdb.exp b/gdb/testsuite/lib/gdb.exp > index 8b16b38..1393455 100644 > --- a/gdb/testsuite/lib/gdb.exp > +++ b/gdb/testsuite/lib/gdb.exp > @@ -2097,6 +2097,75 @@ proc skip_vsx_tests {} { > return $skip_vsx_tests_saved > } > > +# Run a test on the target to see if it supports btrace hardware. Return 0 if so, > +# 1 if it does not. Based on 'check_vmx_hw_available' from the GCC testsuite. > + > +proc skip_btrace_tests {} { > + global skip_btrace_tests_saved > + global srcdir subdir gdb_prompt inferior_exited_re > + > + # Use the cached value, if it exists. > + set me "skip_btrace_tests" > + if [info exists skip_btrace_tests_saved] { > + verbose "$me: returning saved $skip_btrace_tests_saved" 2 > + return $skip_btrace_tests_saved > + } > + > + if { ![istarget "i?86-*-*"] && ![istarget "x86_64-*-*"] } { > + verbose "$me: target does not support btrace, returning 1" 2 > + return [set skip_btrace_tests_saved 1] > + } > + > + # Set up, compile, and execute a test program. > + # Include the current process ID in the file names to prevent conflicts > + # with invocations for multiple testsuites. > + set src btrace[pid].c > + set exe btrace[pid].x Use: set src [standard_output_file btrace[pid].c] set exe [standard_output_file btrace[pid].x] > + > + set f [open $src "w"] > + puts $f "int main() { return 0; }" For C there should be: puts $f "int main(void) { return 0; }" > + close $f > + > + verbose "$me: compiling testfile $src" 2 > + set compile_flags {debug nowarnings quiet} > + set lines [gdb_compile $src $exe executable $compile_flags] > + file delete $src > + > + if ![string match "" $lines] then { > + verbose "$me: testfile compilation failed, returning 1" 2 > + return [set skip_btrace_tests_saved 1] > + } > + > + # No error message, compilation succeeded so now run it via gdb. > + > + gdb_exit > + gdb_start > + gdb_reinitialize_dir $srcdir/$subdir > + gdb_load "$exe" This can be replaced by: clean_restart btrace[pid].x Unfortunately one cannot use $exe as that one has standard_output_file already applied which must not be done for clean_restart. > + runto_main if ![runto_main] { return [...] } > + # In case of an unexpected output, we return 2 as a fail value. > + set skip_btrace_tests_saved 2 > + gdb_test_multiple "record btrace" "check btrace support" { > + -re "You can't do that when your target is.*\r\n$gdb_prompt $" { > + set skip_btrace_tests_saved 1 > + } > + -re "Target does not support branch tracing.*\r\n$gdb_prompt $" { > + set skip_btrace_tests_saved 1 > + } > + -re "Could not enable branch tracing.*\r\n$gdb_prompt $" { > + set skip_btrace_tests_saved 1 > + } > + -re "$gdb_prompt $" { To test no output has been produced (instead of any output has been produced): -re "^record btrace\r\n$gdb_prompt $" { > + set skip_btrace_tests_saved 0 > + } > + } > + gdb_exit > + remote_file build delete $exe > + > + verbose "$me: returning $skip_btrace_tests_saved" 2 > + return $skip_btrace_tests_saved > +} > + > # Skip all the tests in the file if you are not on an hppa running > # hpux target. > > -- > 1.7.1 If you follow the advices there shold not be an issue but it would be probably worth another review. Thanks, Jan