Merge "citrix" branch into "master".

This merge is long overdue, simply because I forgot that there were
outstanding changes on "citrix" that had not yet been merged.

The important fix here is the addition of mlockall.  This fixes some
bugs seen under stressful conditions in XenServer.

diff --git a/ChangeLog b/ChangeLog
index 1acda09..e9ba255 100644 (file)
--- a/ChangeLog
+++ b/ChangeLog
@@ -12,6 +12,11 @@ v0.99.0 - 14 Jan 2010
 ---------------------
     - User-space forwarding engine
     - Bug fixes
+       
+v0.90.7 - 29 Nov 2009
+---------------------
+    - Add support for NetFlow active timeouts
+    - Bug fixes
 
 v0.90.6 - 6 Oct 2009
 --------------------

diff --git a/configure.ac b/configure.ac
index 6d49484..e8497e1 100644 (file)
--- a/configure.ac
+++ b/configure.ac
@@ -53,6 +53,7 @@ OVS_CHECK_IF_PACKET
 OVS_CHECK_STRTOK_R
 AC_CHECK_MEMBERS([struct stat.st_mtim.tv_nsec, struct stat.st_mtimensec],
   [], [], [[#include <sys/stat.h>]])
+AC_CHECK_FUNCS([mlockall])
 
 OVS_CHECK_PKIDIR
 OVS_CHECK_RUNDIR

diff --git a/vswitchd/ovs-vswitchd.8.in b/vswitchd/ovs-vswitchd.8.in
index cbfea13..0ec1ecd 100644 (file)
--- a/vswitchd/ovs-vswitchd.8.in
+++ b/vswitchd/ovs-vswitchd.8.in
@@ -74,6 +74,17 @@ Open vSwitch distribution for instructions on how to build and load
 the Open vSwitch kernel module.
 .PP
 .SH OPTIONS
+.IP "\fB--mlockall\fR"
+Causes \fBovs\-vswitchd\fR to call the \fBmlockall()\fR function, to
+attempt to lock all of its process memory into physical RAM,
+preventing the kernel from paging any of its memory to disk.  This
+helps to avoid networking interruptions due to system memory pressure.
+.IP
+Some systems do not support \fBmlockall()\fR at all, and other systems
+only allow privileged users, such as the superuser, to use it.
+\fBovs\-vswitchd\fR emits a log message if \fBmlockall()\fR is
+unavailable or unsuccessful.
+.
 .IP "\fB--fake-proc-net\fR"
 Causes \fBovs\-vswitchd\fR to simulate some files in \fB/proc/net/vlan\fR
 and \fB/proc/net/bonding\fR that some legacy software expects to

diff --git a/vswitchd/ovs-vswitchd.c b/vswitchd/ovs-vswitchd.c
index 7b0661e..c1acfc4 100644 (file)
--- a/vswitchd/ovs-vswitchd.c
+++ b/vswitchd/ovs-vswitchd.c
@@ -22,6 +22,9 @@
 #include <signal.h>
 #include <stdlib.h>
 #include <string.h>
+#ifdef HAVE_MLOCKALL
+#include <sys/mman.h>
+#endif
 
 #include "bridge.h"
 #include "command-line.h"
@@ -136,6 +139,7 @@ parse_options(int argc, char *argv[])
 {
     enum {
         OPT_PEER_CA_CERT = UCHAR_MAX + 1,
+        OPT_MLOCKALL,
         OPT_FAKE_PROC_NET,
         VLOG_OPTION_ENUMS,
         LEAK_CHECKER_OPTION_ENUMS,
@@ -144,6 +148,7 @@ parse_options(int argc, char *argv[])
     static struct option long_options[] = {
         {"help",        no_argument, 0, 'h'},
         {"version",     no_argument, 0, 'V'},
+        {"mlockall",    no_argument, 0, OPT_MLOCKALL},
         {"fake-proc-net", no_argument, 0, OPT_FAKE_PROC_NET},
         DAEMON_LONG_OPTIONS,
         VLOG_LONG_OPTIONS,
@@ -175,6 +180,16 @@ parse_options(int argc, char *argv[])
             OVS_PRINT_VERSION(OFP_VERSION, OFP_VERSION);
             exit(EXIT_SUCCESS);
 
+        case OPT_MLOCKALL:
+#ifdef HAVE_MLOCKALL
+            if (mlockall(MCL_CURRENT | MCL_FUTURE)) {
+                VLOG_ERR("mlockall failed: %s", strerror(errno));
+            }
+#else
+            VLOG_ERR("mlockall not supported on this system");
+#endif
+            break;
+
         case OPT_FAKE_PROC_NET:
             error = proc_net_compat_init();
             if (error) {

diff --git a/xenserver/etc_init.d_vswitch b/xenserver/etc_init.d_vswitch
index 7a8b83e..9b05879 100755 (executable)
--- a/xenserver/etc_init.d_vswitch
+++ b/xenserver/etc_init.d_vswitch
@@ -50,6 +50,7 @@ test -e /etc/sysconfig/vswitch && . /etc/sysconfig/vswitch
 : ${VSWITCHD_PIDFILE:=/var/run/ovs-vswitchd.pid}
 : ${VSWITCHD_RUN_DIR:=/var/xen/vswitch}
 : ${VSWITCHD_PRIORITY:=-10}
+: ${VSWITCHD_MLOCKALL:=yes}
 : ${VSWITCHD_LOGFILE:=/var/log/ovs-vswitchd.log}
 : ${VSWITCHD_FILE_LOGLEVEL:=INFO}
 : ${VSWITCHD_SYSLOG_LOGLEVEL:=ERR}
@@ -218,12 +219,15 @@ function start_vswitchd {
     if [ "$ENABLE_FAKE_PROC_NET" = "y" ]; then
         fake_proc_net_opt="--fake-proc-net"
     fi
+    if [ "$VSWITCHD_MLOCKALL" != "no" ]; then
+        mlockall_opt="--mlockall"
+    fi
     if [ "$daemonize" != "y" ]; then
         # Start in background and force a "success" message
         action "Starting ovs-vswitchd ($strace_opt$valgrind_opt)" true
-        (nice -n "$VSWITCHD_PRIORITY" $strace_opt $valgrind_opt "$vswitchd" --pidfile="$VSWITCHD_PIDFILE" --detach $monitor_opt --no-chdir $fake_proc_net_opt -vANY:CONSOLE:EMER $syslog_opt $logfile_level_opt $logfile_file_opt $leak_opt "$VSWITCHD_OVSDB_SERVER") &
+        (nice -n "$VSWITCHD_PRIORITY" $strace_opt $valgrind_opt "$vswitchd" --pidfile="$VSWITCHD_PIDFILE" --detach $monitor_opt --no-chdir $fake_proc_net_opt -vANY:CONSOLE:EMER $syslog_opt $logfile_level_opt $logfile_file_opt $leak_opt $mlockall_opt "$VSWITCHD_OVSDB_SERVER") &
     else
-        action "Starting ovs-vswitchd" nice -n "$VSWITCHD_PRIORITY" "$vswitchd" --pidfile="$VSWITCHD_PIDFILE" --detach $monitor_opt --no-chdir $fake_proc_net_opt -vANY:CONSOLE:EMER $syslog_opt $logfile_level_opt $logfile_file_opt $leak_opt "$VSWITCHD_OVSDB_SERVER"
+        action "Starting ovs-vswitchd" nice -n "$VSWITCHD_PRIORITY" "$vswitchd" --pidfile="$VSWITCHD_PIDFILE" --detach $monitor_opt --no-chdir $fake_proc_net_opt -vANY:CONSOLE:EMER $syslog_opt $logfile_level_opt $logfile_file_opt $leak_opt $mlockall_opt "$VSWITCHD_OVSDB_SERVER"
     fi
 }
 

diff --git a/xenserver/usr_share_vswitch_scripts_sysconfig.template b/xenserver/usr_share_vswitch_scripts_sysconfig.template
index 8b740c2..4d5efca 100644 (file)
--- a/xenserver/usr_share_vswitch_scripts_sysconfig.template
+++ b/xenserver/usr_share_vswitch_scripts_sysconfig.template
@@ -117,6 +117,13 @@
 #     processes.
 # VSWITCHD_PRIORITY=-10
 
+# VSWITCHD_MLOCKALL: Whether to pass ovs-vswitchd the --mlockall option.
+#     This option should be set to "yes" or "no".  The default is "yes".
+#     Enabling this option can avoid networking interruptions due to
+#     system memory pressure in extraordinary situations, such as multiple
+#     concurrent VM import operations.
+# VSWITCHD_MLOCKALL=yes
+
 # VSWITCHD_LOGFILE: File to send the FILE_LOGLEVEL log messages to.
 # VSWITCHD_LOGFILE=/var/log/ovs-vswitchd.log