From: Yao Qi <qiyaoltc@gmail.com>
To: gdb-patches@sourceware.org
Subject: [PATCH 7/8] Add a test case for fast tracepoints' locking mechanism
Date: Fri, 18 Sep 2015 12:43:00 -0000 [thread overview]
Message-ID: <1442580184-22562-8-git-send-email-yao.qi@linaro.org> (raw)
In-Reply-To: <1442580184-22562-1-git-send-email-yao.qi@linaro.org>
From: Pierre Langlois <pierre.langlois@arm.com>
When installing a fast tracepoint, we create a jump pad with a
spin-lock. This way, only one thread can collect a given tracepoint at
any time. This test case checks that this lock actually works as
expected.
This test works by creating a function which overrides the in-process
agent library's gdb_collect function. On start up, GDBserver will ask
GDB with the 'qSymbol' packet about symbols present in the inferior.
GDB will reply with the gdb_agent_gdb_collect function from the test
case instead of the one from the agent.
gdb/testsuite/ChangeLog:
* gdb.trace/ftrace-lock.c: New file.
* gdb.trace/ftrace-lock.exp: New file.
---
gdb/testsuite/gdb.trace/ftrace-lock.c | 106 ++++++++++++++++++++++++++++++++
gdb/testsuite/gdb.trace/ftrace-lock.exp | 88 ++++++++++++++++++++++++++
2 files changed, 194 insertions(+)
create mode 100644 gdb/testsuite/gdb.trace/ftrace-lock.c
create mode 100644 gdb/testsuite/gdb.trace/ftrace-lock.exp
diff --git a/gdb/testsuite/gdb.trace/ftrace-lock.c b/gdb/testsuite/gdb.trace/ftrace-lock.c
new file mode 100644
index 0000000..8ed45f4
--- /dev/null
+++ b/gdb/testsuite/gdb.trace/ftrace-lock.c
@@ -0,0 +1,106 @@
+/* This testcase is part of GDB, the GNU debugger.
+
+ Copyright 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/>. */
+
+#include <pthread.h>
+
+#ifdef SYMBOL_PREFIX
+#define SYMBOL(str) SYMBOL_PREFIX #str
+#else
+#define SYMBOL(str) #str
+#endif
+
+/* Called if the testcase failed. */
+static void
+fail (void)
+{
+}
+
+static pthread_mutex_t mutex = PTHREAD_MUTEX_INITIALIZER;
+
+/* This function overrides gdb_collect in the in-process agent library.
+ See gdbserver/tracepoint.c (gdb_collect). We want this function to
+ be ran instead of the one from the library to easily check that only
+ one thread is tracing at a time.
+
+ This works as expected because GDBserver will ask GDB about symbols
+ present in the inferior with the 'qSymbol' packet. And GDB will
+ reply with the address of this function instead of the one from the
+ in-process agent library. */
+
+void
+gdb_agent_gdb_collect (void *tpoint, unsigned char *regs)
+{
+ /* If we cannot acquire a lock, then this means another thread is
+ tracing and the lock implemented by the jump pad is not working! */
+ if (pthread_mutex_trylock (&mutex) != 0)
+ {
+ fail ();
+ return;
+ }
+
+ sleep (1);
+
+ if (pthread_mutex_unlock (&mutex) != 0)
+ {
+ fail ();
+ return;
+ }
+}
+
+/* Called from asm. */
+static void __attribute__((used))
+func (void)
+{
+}
+
+static void *
+thread_function (void *arg)
+{
+ /* `set_point' is the label at which to set a fast tracepoint. The
+ insn at the label must be large enough to fit a fast tracepoint
+ jump. */
+ asm (" .global " SYMBOL (set_point) "\n"
+ SYMBOL (set_point) ":\n"
+#if (defined __x86_64__ || defined __i386__)
+ " call " SYMBOL (func) "\n"
+#elif (defined __aarch64__)
+ " nop\n"
+#endif
+ );
+}
+
+static void
+end (void)
+{
+}
+
+int
+main (int argc, char *argv[], char *envp[])
+{
+ pthread_t threads[NUM_THREADS];
+ int i;
+
+ for (i = 0; i < NUM_THREADS; i++)
+ pthread_create (&threads[i], NULL, thread_function, NULL);
+
+ for (i = 0; i < NUM_THREADS; i++)
+ pthread_join (threads[i], NULL);
+
+ end ();
+
+ return 0;
+}
diff --git a/gdb/testsuite/gdb.trace/ftrace-lock.exp b/gdb/testsuite/gdb.trace/ftrace-lock.exp
new file mode 100644
index 0000000..c67739d
--- /dev/null
+++ b/gdb/testsuite/gdb.trace/ftrace-lock.exp
@@ -0,0 +1,88 @@
+# Copyright 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/>.
+
+load_lib "trace-support.exp"
+
+standard_testfile
+set executable $testfile
+set expfile $testfile.exp
+
+# make check RUNTESTFLAGS='gdb.trace/ftrace-lock.exp NUM_THREADS=2'
+if ![info exists NUM_THREADS] {
+ set NUM_THREADS 2
+}
+
+# Some targets have leading underscores on assembly symbols.
+set additional_flags [gdb_target_symbol_prefix_flags]
+set options [list debug [gdb_target_symbol_prefix_flags] \
+ additional_flags=-DNUM_THREADS=$NUM_THREADS]
+
+# Check that the target supports trace.
+if { [gdb_compile_pthreads "$srcdir/$subdir/$srcfile" $binfile executable $options] != "" } {
+ untested "Couldn't compile test program"
+ return -1
+}
+
+clean_restart ${testfile}
+
+if ![runto_main] {
+ fail "Can't run to main to check for trace support"
+ return -1
+}
+
+if ![gdb_target_supports_trace] {
+ unsupported "target does not support trace"
+ return -1
+}
+
+# Compile the test case with the in-process agent library.
+set libipa [get_in_proc_agent]
+gdb_load_shlibs $libipa
+
+lappend options shlib=$libipa
+
+if { [gdb_compile_pthreads "$srcdir/$subdir/$srcfile" $binfile executable $options] != "" } {
+ untested "Couldn't compile test program with in-process agent library"
+ return -1
+}
+
+clean_restart ${executable}
+
+if ![runto_main] {
+ fail "Can't run to main to check for trace support"
+ return -1
+}
+
+if { [gdb_test "info sharedlibrary" ".*${libipa}.*" "IPA loaded"] != 0 } {
+ untested "Could not find IPA lib loaded"
+ return 1
+}
+
+gdb_test "break end" ""
+gdb_test "break fail" ""
+
+gdb_test "ftrace set_point" "Fast tracepoint .*" \
+ "fast tracepoint at a long insn"
+
+gdb_test "tstart" ""
+
+# If NUM_THREADS is high then this test case may timeout. Increase the
+# timeout temporarily.
+with_timeout_factor $NUM_THREADS {
+ # If the fail function is hit, then the testcase will fail.
+ gdb_test "continue" ".*Breakpoint \[0-9\]+, end \(\).*" \
+ "do not hit the fail function"
+}
+
+gdb_test "tstop" ""
--
1.9.1
next prev parent reply other threads:[~2015-09-18 12:43 UTC|newest]
Thread overview: 37+ messages / expand[flat|nested] mbox.gz Atom feed top
2015-09-14 11:31 [PATCH 0/8] [AArch64] Support fast tracepoints Pierre Langlois
2015-09-14 11:31 ` [PATCH 1/8] [AArch64] Use debug_printf instead of fprintf_unfiltered Pierre Langlois
2015-09-15 10:06 ` Yao Qi
2015-09-15 11:07 ` Pierre Langlois
2015-09-15 11:19 ` Yao Qi
2015-09-15 11:43 ` [PATCH 2/8 v2] " Pierre Langlois
2015-09-15 17:40 ` Pierre Langlois
2015-09-14 11:31 ` [PATCH 3/8] [AArch64] Make aarch64_decode_adrp handle both ADR and ADRP instructions Pierre Langlois
2015-09-15 10:12 ` Yao Qi
2015-09-15 12:05 ` [PATCH 3/8 v2] " Pierre Langlois
2015-09-14 11:32 ` [PATCH 7/8] [testsuite] Add a gdb.trace test for instruction relocation Pierre Langlois
2015-09-15 11:27 ` Yao Qi
2015-09-14 11:32 ` [PATCH 5/8] [GDBserver][AArch64] Implement target_emit_ops Pierre Langlois
2015-09-14 11:32 ` [PATCH 6/8] [testsuite][AArch64] Enable fast tracepoint tests Pierre Langlois
2015-09-15 10:18 ` Yao Qi
2015-09-14 11:33 ` [PATCH 4/8] [GDBserver][AArch64] Add support for fast tracepoints Pierre Langlois
2015-09-15 13:01 ` Yao Qi
2015-09-14 11:38 ` [PATCH 2/8] [AArch64] Move instruction decoding into new arch/ directory Pierre Langlois
2015-09-15 10:10 ` Yao Qi
2015-09-15 12:02 ` [PATCH 2/8 v2] " Pierre Langlois
2015-09-14 11:38 ` [PATCH 8/8] [testsuite] Add a test case for fast tracepoints' locking mechanism Pierre Langlois
2015-09-14 16:45 ` [PATCH] Add NEWS entry for fast tracepoint support on aarch64-linux Pierre Langlois
2015-09-18 12:43 ` [PATCH 0/8 V2] [AArch64] Support fast tracepoints Yao Qi
2015-09-18 12:43 ` [PATCH 4/8] Implement target_emit_ops Yao Qi
2015-09-18 12:57 ` Pierre Langlois
2016-02-05 20:09 ` Antoine Tremblay
2016-02-08 17:30 ` Yao Qi
2016-02-08 17:59 ` Antoine Tremblay
2015-09-18 12:43 ` [PATCH 8/8] Add NEWS entry for fast tracepoint support on aarch64-linux Yao Qi
2015-09-18 14:07 ` Eli Zaretskii
2015-09-18 12:43 ` [PATCH 1/8] Move instruction decoding into new arch/ directory Yao Qi
2015-09-18 12:43 ` Yao Qi [this message]
2015-09-18 12:43 ` [PATCH 5/8] Enable fast tracepoint tests Yao Qi
2015-09-18 12:43 ` [PATCH 2/8] Make aarch64_decode_adrp handle both ADR and ADRP instructions Yao Qi
2015-09-18 12:44 ` [PATCH 6/8] Add a gdb.trace test for instruction relocation Yao Qi
2015-09-18 12:44 ` [PATCH 3/8] Add support for fast tracepoints Yao Qi
2015-09-21 14:06 ` [PATCH 0/8 V2] [AArch64] Support " Yao Qi
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=1442580184-22562-8-git-send-email-yao.qi@linaro.org \
--to=qiyaoltc@gmail.com \
--cc=gdb-patches@sourceware.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox