11:02 a.m.
From: "Richard W.M. Jones" <rjones(a)redhat.com>
This commit allows you to run the daemon under valgrind. You have to
enable it at configure time:
./configure --enable-valgrind-daemon
This should *not* be done for production builds.
When this feature is enabled, valgrind is added to the appliance and
the daemon is run under valgrind. Log messages from valgrind are
passed back over a virtio-serial channel into a file called
'valgrind.log.$PID' in the top build directory.
Running 'make check', 'make extra-tests' etc causes many
valgrind.log.* files to be created which must be examined by hand.
---
.gitignore | 1 +
appliance/Makefile.am | 9 +++++++--
appliance/init | 10 +++++++++-
appliance/packagelist.in | 4 ++++
configure.ac | 15 +++++++++++++++
src/guestfs.c | 8 ++++++++
src/launch.c | 10 ++++++++++
7 files changed, 54 insertions(+), 3 deletions(-)
diff --git a/.gitignore b/.gitignore
index b75cafb..fc732bc 100644
--- a/.gitignore
+++ b/.gitignore
@@ -387,3 +387,4 @@ test-tool/libguestfs-test-tool-helper
tools/test.img
tools/virt-*.1
tools/virt-*.pl
+/valgrind.log.*
diff --git a/appliance/Makefile.am b/appliance/Makefile.am
index e2e8b96..99896be 100644
--- a/appliance/Makefile.am
+++ b/appliance/Makefile.am
@@ -42,13 +42,18 @@ make.sh: make.sh.in
chmod +x $@-t
mv $@-t $@
+PACKAGELIST_CPP_FLAGS = -D$(DISTRO)=1
+if VALGRIND_DAEMON
+PACKAGELIST_CPP_FLAGS += -DVALGRIND_DAEMON=1
+endif
+
packagelist: packagelist.in
- cpp -undef -D$(DISTRO)=1 < $< | \
+ cpp -undef $(PACKAGELIST_CPP_FLAGS) < $< | \
grep -v '^[[:space:]]*$$' | grep -v '^#' > $@-t
mv $@-t $@
excludelist: excludelist.in
- cpp -undef -D$(DISTRO)=1 < $< | \
+ cpp -undef $(PACKAGELIST_CPP_FLAGS) < $< | \
grep -v '^[[:space:]]*$$' | grep -v '^#' > $@-t
mv $@-t $@
diff --git a/appliance/init b/appliance/init
index 0f32a55..1d69339 100755
--- a/appliance/init
+++ b/appliance/init
@@ -106,8 +106,16 @@ if grep -sq guestfs_verbose=1 /proc/cmdline; then
fi
if ! grep -sq guestfs_rescue=1 /proc/cmdline; then
+ # Run the daemon under valgrind if ./configure --enable-valgrind-daemon
+ vg_channel=/dev/virtio-ports/org.libguestfs.valgrind
+ if [ -w $vg_channel ]; then
+ exec 3>$vg_channel
+ vg="valgrind --leak-check=full --log-fd=3 --error-exitcode=119
--max-stackframe=8388608 --child-silent-after-fork=yes"
+ echo "enabling valgrind: $vg"
+ fi
+
# The host will kill qemu abruptly if guestfsd shuts down normally
- guestfsd
+ $vg guestfsd
# Otherwise we try to clean up gracefully. For example, this ensures that a
# core dump generated by the guest daemon will be written to disk.
diff --git a/appliance/packagelist.in b/appliance/packagelist.in
index 5735a5a..b42c510 100644
--- a/appliance/packagelist.in
+++ b/appliance/packagelist.in
@@ -130,3 +130,7 @@ tar
xfsprogs
#endif
zerofree
+
+#ifdef VALGRIND_DAEMON
+valgrind
+#endif
diff --git a/configure.ac b/configure.ac
index 3e364b1..ebc20aa 100644
--- a/configure.ac
+++ b/configure.ac
@@ -264,8 +264,23 @@ if test "x$enable_daemon" = "xyes"; then
[],
[enable_install_daemon=no])
AC_MSG_RESULT([$enable_install_daemon])
+
+ dnl Enable valgrind in the daemon.
+ AC_MSG_CHECKING([if we should run the daemon under valgrind])
+ AC_ARG_ENABLE([valgrind-daemon],
+ [AS_HELP_STRING([--enable-valgrind-daemon],
+ [run the daemon under valgrind (developers only)
@<:@default=no@:>@])],
+ [],
+ [enable_valgrind_daemon=no])
+ AC_MSG_RESULT([$enable_valgrind_daemon])
+
+ if test "x$enable_valgrind_daemon" = "xyes"; then
+ AC_DEFINE([VALGRIND_DAEMON],[1],[Define to 1 to run the daemon under
valgrind])
+ AC_DEFINE_UNQUOTED([VALGRIND_LOG_PATH],["$(pwd)"],[Path to save
valgrind log files])
+ fi
fi
AM_CONDITIONAL([INSTALL_DAEMON],[test "x$enable_install_daemon" =
"xyes"])
+AM_CONDITIONAL([VALGRIND_DAEMON],[test "x$enable_valgrind_daemon" =
"xyes"])
dnl Build the appliance?
AC_MSG_CHECKING([if we should build the appliance])
diff --git a/src/guestfs.c b/src/guestfs.c
index 31968e4..3607eaa 100644
--- a/src/guestfs.c
+++ b/src/guestfs.c
@@ -190,9 +190,17 @@ guestfs_close (guestfs_h *g)
if (g->autosync && g->state == READY)
guestfs_internal_autosync (g);
+ /* If we are valgrinding the daemon, then we *don't* want to kill
+ * the subprocess because we want the final valgrind messages sent
+ * when we close sockets below. However for normal production use,
+ * killing the subprocess is the right thing to do (in case the
+ * daemon or qemu is not responding).
+ */
+#ifndef VALGRIND_DAEMON
/* Kill the qemu subprocess. */
if (g->state != CONFIG)
guestfs_kill_subprocess (g);
+#endif
/* Run user close callbacks. */
guestfs___call_callbacks_void (g, GUESTFS_EVENT_CLOSE);
diff --git a/src/launch.c b/src/launch.c
index 1af74b9..4e2fba9 100644
--- a/src/launch.c
+++ b/src/launch.c
@@ -703,6 +703,16 @@ launch_appliance (guestfs_h *g)
add_cmdline (g, "-device");
add_cmdline (g,
"virtserialport,chardev=channel0,name=org.libguestfs.channel.0");
+#ifdef VALGRIND_DAEMON
+ /* Set up virtio-serial channel for valgrind messages. */
+ add_cmdline (g, "-chardev");
+ snprintf (buf, sizeof buf, "file,path=%s/valgrind.log.%d,id=valgrind",
+ VALGRIND_LOG_PATH, getpid ());
+ add_cmdline (g, buf);
+ add_cmdline (g, "-device");
+ add_cmdline (g,
"virtserialport,chardev=valgrind,name=org.libguestfs.valgrind");
+#endif
+
/* Enable user networking. */
if (g->enable_network) {
add_cmdline (g, "-netdev");
--
1.7.6