- Devel - libvirt List Archives

[libvirt] [PATCH] lxc: Ensure to kill init process by sending SIGKILL if needed
by Ryota Ozaki 22 Jul '10

22 Jul '10

Init process may remain after sending SIGTERM for some reason. For example, if original init program is used, it is definitely not killed by SIGTERM. --- src/lxc/lxc_controller.c | 4 ++++ 1 files changed, 4 insertions(+), 0 deletions(-) diff --git a/src/lxc/lxc_controller.c b/src/lxc/lxc_controller.c index 44bcc82..d8b7bc7 100644 --- a/src/lxc/lxc_controller.c +++ b/src/lxc/lxc_controller.c @@ -663,7 +663,11 @@ cleanup: close(containerPty); if (container > 1) { + int status; kill(container, SIGTERM); + if (!(waitpid(container, &status, WNOHANG) == 0 && + WIFEXITED(status))) + kill(container, SIGKILL); waitpid(container, NULL, 0); } return rc; -- 1.6.6.1

2 2

[libvirt] [PATCH] Fix incorrect use of private data in remote driver
by Daniel P. Berrange 22 Jul '10

22 Jul '10

For upstream bug https://bugzilla.redhat.com/show_bug.cgi?id=576796 The remote driver is using the wrong privateData field in a couple of functions. THis is harmless for stateful drivers like QEMU/UML/LXC, but will crash with Xen * src/remote/remote_driver.c: Fix use of privateData field --- src/remote/remote_driver.c | 10 +++++----- 1 files changed, 5 insertions(+), 5 deletions(-) diff --git a/src/remote/remote_driver.c b/src/remote/remote_driver.c index 7052bf1..c9129b1 100644 --- a/src/remote/remote_driver.c +++ b/src/remote/remote_driver.c @@ -4067,7 +4067,7 @@ remoteNetworkIsActive(virNetworkPtr network) int rv = -1; remote_network_is_active_args args; remote_network_is_active_ret ret; - struct private_data *priv = network->conn->privateData; + struct private_data *priv = network->conn->networkPrivateData; remoteDriverLock(priv); @@ -4091,7 +4091,7 @@ remoteNetworkIsPersistent(virNetworkPtr network) int rv = -1; remote_network_is_persistent_args args; remote_network_is_persistent_ret ret; - struct private_data *priv = network->conn->privateData; + struct private_data *priv = network->conn->networkPrivateData; remoteDriverLock(priv); @@ -4622,7 +4622,7 @@ remoteInterfaceIsActive(virInterfacePtr iface) int rv = -1; remote_interface_is_active_args args; remote_interface_is_active_ret ret; - struct private_data *priv = iface->conn->privateData; + struct private_data *priv = iface->conn->interfacePrivateData; remoteDriverLock(priv); @@ -5127,7 +5127,7 @@ remoteStoragePoolIsActive(virStoragePoolPtr pool) int rv = -1; remote_storage_pool_is_active_args args; remote_storage_pool_is_active_ret ret; - struct private_data *priv = pool->conn->privateData; + struct private_data *priv = pool->conn->storagePrivateData; remoteDriverLock(priv); @@ -5151,7 +5151,7 @@ remoteStoragePoolIsPersistent(virStoragePoolPtr pool) int rv = -1; remote_storage_pool_is_persistent_args args; remote_storage_pool_is_persistent_ret ret; - struct private_data *priv = pool->conn->privateData; + struct private_data *priv = pool->conn->storagePrivateData; remoteDriverLock(priv); -- 1.5.5.6 -- |: Red Hat, Engineering, London -o- http://people.redhat.com/berrange/ :| |: http://libvirt.org -o- http://virt-manager.org -o- http://deltacloud.org :| |: http://autobuild.org -o- http://search.cpan.org/~danberr/ :| |: GnuPG: 7D3B9505 -o- F3C9 553F A1DA 4AC2 5648 23C1 B3DF F742 7D3B 9505 :|

3 2

[libvirt] [PATCH (v2)] Fix PCI address assignment if no IDE controller is present
by Daniel P. Berrange 22 Jul '10

22 Jul '10

The previous patch was flawed because it forgot to take into account pre-reserved addresses The PCI slot 1 must be reserved at all times, since PIIX3 is always present, even if no IDE device is in use for guest disks * src/qemu/qemu_conf.c: Always reserve slot 1 for PIIX3 --- src/qemu/qemu_conf.c | 22 +++++++++++++++++----- 1 files changed, 17 insertions(+), 5 deletions(-) diff --git a/src/qemu/qemu_conf.c b/src/qemu/qemu_conf.c index 23e418b..0f1526d 100644 --- a/src/qemu/qemu_conf.c +++ b/src/qemu/qemu_conf.c @@ -2271,19 +2271,23 @@ int qemuDomainPCIAddressSetNextAddr(qemuDomainPCIAddressSetPtr addrs, * - VirtIO balloon * - Host device passthrough * - Watchdog + * + * Prior to this function being invoked, qemuCollectPCIAddress() will have + * added all existing PCI addresses from the 'def' to 'addrs'. Thus this + * function must only try to reserve addresses if info.type == NONE and + * skip over info.type == PCI */ int qemuAssignDevicePCISlots(virDomainDefPtr def, qemuDomainPCIAddressSetPtr addrs) { int i; + bool reservedIDE = false; /* Host bridge */ if (qemuDomainPCIAddressReserveSlot(addrs, 0) < 0) goto error; - /* PIIX3 (ISA bridge, IDE controller, something else unknown, USB controller) - * hardcoded slot=1, multifunction device - */ + /* Verify that first IDE controller (if any) is on the PIIX3, fn 1 */ for (i = 0; i < def->ncontrollers ; i++) { /* First IDE controller lives on the PIIX3 at slot=1, function=1 */ if (def->controllers[i]->type == VIR_DOMAIN_CONTROLLER_TYPE_IDE && @@ -2297,18 +2301,26 @@ qemuAssignDevicePCISlots(virDomainDefPtr def, qemuDomainPCIAddressSetPtr addrs) _("Primary IDE controller must have PCI address 0:0:1.1")); goto error; } + /* If TYPE==PCI, then then qemuCollectPCIAddress() function + * has already reserved the address, so we must skip */ + reservedIDE = true; } else { def->controllers[i]->info.type = VIR_DOMAIN_DEVICE_ADDRESS_TYPE_PCI; def->controllers[i]->info.addr.pci.domain = 0; def->controllers[i]->info.addr.pci.bus = 0; def->controllers[i]->info.addr.pci.slot = 1; def->controllers[i]->info.addr.pci.function = 1; - if (qemuDomainPCIAddressReserveSlot(addrs, 1) < 0) - goto error; } } } + /* PIIX3 (ISA bridge, IDE controller, something else unknown, USB controller) + * hardcoded slot=1, multifunction device + */ + if (!reservedIDE && + qemuDomainPCIAddressReserveSlot(addrs, 1) < 0) + goto error; + /* First VGA is hardcoded slot=2 */ if (def->nvideos > 0) { if (def->videos[0]->info.type == VIR_DOMAIN_DEVICE_ADDRESS_TYPE_PCI) { -- 1.7.1.1

2 1

[libvirt] [PATCH] Properly report failure to create raw storage volume files
by Laine Stump 21 Jul '10

21 Jul '10

We were previously checking for a return < 0 from virFileOperation(), but that function returns a standard errno, which is 0 on success, or some small positive number on failure. The result was that we wouldn't report failures to create storage volume files; instead they would appear to be created, but then would vanish as soon as a pool-refresh was done (or cause some later error as soon as someone tried to access the volume). The other uses of virFileOperation() were already properly checking for != 0. --- src/storage/storage_backend.c | 2 +- 1 files changed, 1 insertions(+), 1 deletions(-) diff --git a/src/storage/storage_backend.c b/src/storage/storage_backend.c index aba8937..9792eed 100644 --- a/src/storage/storage_backend.c +++ b/src/storage/storage_backend.c @@ -362,7 +362,7 @@ virStorageBackendCreateRaw(virConnectPtr conn ATTRIBUTE_UNUSED, createRawFileOpHook, &hdata, VIR_FILE_OP_FORCE_PERMS | (pool->def->type == VIR_STORAGE_POOL_NETFS - ? VIR_FILE_OP_AS_UID : 0))) < 0) { + ? VIR_FILE_OP_AS_UID : 0))) != 0) { virReportSystemError(createstat, _("cannot create path '%s'"), vol->target.path); -- 1.7.1.1

4 17

Re: [libvirt] A problem in live migration
by Alex Jia 21 Jul '10

21 Jul '10

I am not sure if libvirt has resolved hostname and ip mapping problem, it had better that two hosts have same domain name, such as "dhcp-1-2-3.nay.redhat.com" and "dhcp-4-5-6.nay.redhat.com". In addition, I haven't seen your any storage pool information on two hosts, you need to create a sharing storage such as nfs sharing folder, and your guest is in the storage, of course, you may also manually mount your nfs folder to target host instead of creating storage pool. Best Regards, Alex ----- Original Message ----- From: "姚远" <sihuyy(a)gmail.com> To: libvir-list(a)redhat.com Sent: Wednesday, July 21, 2010 3:39:15 PM GMT +08:00 Beijing / Chongqing / Hong Kong / Urumqi Subject: [libvirt] A problem in live migration In virsh, I use command "migrate --live 1 qemu+ssh:// 10.1.10.60/system " to do live migration,but it shows error "internal error hostname on destination resolved to localhost, but migration requires an FQDN". I don't know why, and who do live migration successfully can help me. Thank you! My source and destination system are all fedora-13-x86_64, and the version of libvirt is 0.8.2 -- libvir-list mailing list libvir-list(a)redhat.com https://www.redhat.com/mailman/listinfo/libvir-list

3 2

[libvirt] [PATCH] Fix PCI address assignment if no IDE controller is present
by Daniel P. Berrange 21 Jul '10

21 Jul '10

The PCI slot 1 must be reserved at all times, since PIIX3 is always present, even if no IDE device is in use for guest disks * src/qemu/qemu_conf.c: Always reserve slot 1 for PIIX3 --- src/qemu/qemu_conf.c | 6 ++++-- 1 files changed, 4 insertions(+), 2 deletions(-) diff --git a/src/qemu/qemu_conf.c b/src/qemu/qemu_conf.c index 23e418b..6ce9c9e 100644 --- a/src/qemu/qemu_conf.c +++ b/src/qemu/qemu_conf.c @@ -2284,6 +2284,10 @@ qemuAssignDevicePCISlots(virDomainDefPtr def, qemuDomainPCIAddressSetPtr addrs) /* PIIX3 (ISA bridge, IDE controller, something else unknown, USB controller) * hardcoded slot=1, multifunction device */ + if (qemuDomainPCIAddressReserveSlot(addrs, 1) < 0) + goto error; + + /* Verify that first IDE controller (if any) is on the PIIX3, fn 1 */ for (i = 0; i < def->ncontrollers ; i++) { /* First IDE controller lives on the PIIX3 at slot=1, function=1 */ if (def->controllers[i]->type == VIR_DOMAIN_CONTROLLER_TYPE_IDE && @@ -2303,8 +2307,6 @@ qemuAssignDevicePCISlots(virDomainDefPtr def, qemuDomainPCIAddressSetPtr addrs) def->controllers[i]->info.addr.pci.bus = 0; def->controllers[i]->info.addr.pci.slot = 1; def->controllers[i]->info.addr.pci.function = 1; - if (qemuDomainPCIAddressReserveSlot(addrs, 1) < 0) - goto error; } } } -- 1.7.1.1

2 1

[libvirt] [PATCH] Document the memory balloon device
by Daniel P. Berrange 21 Jul '10

21 Jul '10

* formatdomain.html.in: Document <memballoon> element --- docs/formatdomain.html.in | 47 +++++++++++++++++++++++++++++++++++++++++++++ 1 files changed, 47 insertions(+), 0 deletions(-) diff --git a/docs/formatdomain.html.in b/docs/formatdomain.html.in index 4cca67f..3e80312 100644 --- a/docs/formatdomain.html.in +++ b/docs/formatdomain.html.in @@ -1506,6 +1506,53 @@ qemu-kvm -net nic,model=? /dev/null </dd> </dl> + <h4><a name="elementsMemBalloon">Memory balloon device</a></h4> + + <p> + A virtual memory balloon device is added to all Xen and KVM/QEMU + guests. It will be seen as <code>memballoon</code> element. + It will be automatically added when appropriate, so there is no + need to explicitly add this element in the guest XML unless a + specific PCI slot needs to be assigned. + <span class="since">Since 0.8.3, Xen, QEMU and KVM only</span> + </p> + + <p> + Example automatically added device with KVM + </p> +<pre> + ... + <devices> + <memballoon model='virtio'/> + </devices> + ...</pre> + + <p> + Example manually added device with static PCI slot 2 requested + </p> +<pre> + ... + <devices> + <watchdog model='virtio'/> + <address type='pci' domain='0x0000' bus='0x00' slot='0x02' function='0x0'/> + </devices> +</domain></pre> + + <dl> + <dt><code>model</code></dt> + <dd> + <p> + The required <code>model</code> attribute specifies what type + of balloon device is provided. Valid values are specific to + the virtualization platform + </p> + <ul> + <li>'virtio' — default with QEMU/KVM</li> + <li>'xen' — default with Xen</li> + </ul> + </dd> + </dl> + <h2><a name="examples">Example configs</a></h2> <p> -- 1.7.1.1

2 1

[libvirt] [PATCH] Implement support for virtio plan9fs filesystem passthrough in QEMU
by Daniel P. Berrange 21 Jul '10

21 Jul '10

Make use of the existing <filesystem> element to support plan9fs filesystem passthrough in the QEMU driver <filesystem type='mount'> <source dir='/export/to/guest'/> <target dir='/import/from/host'/> </filesystem> NB, the target is not actually a directory, it is merely a arbitrary string tag that is exported to the guest as a hint for where to mount it. --- src/qemu/qemu_conf.c | 96 +++++++++++++++++++++++++ src/qemu/qemu_conf.h | 5 ++ tests/qemuxml2argvdata/qemuxml2argv-fs9p.args | 1 + tests/qemuxml2argvdata/qemuxml2argv-fs9p.xml | 28 +++++++ tests/qemuxml2argvtest.c | 2 + 5 files changed, 132 insertions(+), 0 deletions(-) create mode 100644 tests/qemuxml2argvdata/qemuxml2argv-fs9p.args create mode 100644 tests/qemuxml2argvdata/qemuxml2argv-fs9p.xml diff --git a/src/qemu/qemu_conf.c b/src/qemu/qemu_conf.c index 3e35c57..23e418b 100644 --- a/src/qemu/qemu_conf.c +++ b/src/qemu/qemu_conf.c @@ -1205,6 +1205,8 @@ static unsigned long long qemudComputeCmdFlags(const char *help, flags |= QEMUD_CMD_FLAG_NO_KVM_PIT; if (strstr(help, "-tdf")) flags |= QEMUD_CMD_FLAG_TDF; + if (strstr(help, "-fsdev")) + flags |= QEMUD_CMD_FLAG_FSDEV; /* Keep disabled till we're actually ready to turn on netdev mode * The plan is todo it in 0.13.0 QEMU, but lets wait & see... */ @@ -1987,6 +1989,10 @@ qemuAssignDeviceAliases(virDomainDefPtr def, unsigned long long qemuCmdFlags) if (!(qemuCmdFlags & QEMUD_CMD_FLAG_DEVICE)) return 0; + for (i = 0; i < def->nfss ; i++) { + if (virAsprintf(&def->fss[i]->info.alias, "fs%d", i) < 0) + goto no_memory; + } for (i = 0; i < def->nsounds ; i++) { if (virAsprintf(&def->sounds[i]->info.alias, "sound%d", i) < 0) goto no_memory; @@ -2324,6 +2330,15 @@ qemuAssignDevicePCISlots(virDomainDefPtr def, qemuDomainPCIAddressSetPtr addrs) goto error; } } + for (i = 0; i < def->nfss ; i++) { + if (def->fss[i]->info.type != VIR_DOMAIN_DEVICE_ADDRESS_TYPE_NONE) + continue; + + /* Only support VirtIO-9p-pci so far. If that changes, + * we might need to skip devices here */ + if (qemuDomainPCIAddressSetNextAddr(addrs, &def->fss[i]->info) < 0) + goto error; + } /* Network interfaces */ for (i = 0; i < def->nnets ; i++) { @@ -2713,6 +2728,64 @@ error: } +char *qemuBuildFSStr(virDomainFSDefPtr fs, + unsigned long long qemuCmdFlags ATTRIBUTE_UNUSED) +{ + virBuffer opt = VIR_BUFFER_INITIALIZER; + + if (fs->type != VIR_DOMAIN_FS_TYPE_MOUNT) { + qemuReportError(VIR_ERR_CONFIG_UNSUPPORTED, "%s", + _("can only passthrough directories")); + goto error; + } + + virBufferAddLit(&opt, "local,security_model=mapped"); + virBufferVSprintf(&opt, ",id=%s%s", QEMU_FSDEV_HOST_PREFIX, fs->info.alias); + virBufferVSprintf(&opt, ",path=%s", fs->src); + + if (virBufferError(&opt)) { + virReportOOMError(); + goto error; + } + + return virBufferContentAndReset(&opt); + +error: + virBufferFreeAndReset(&opt); + return NULL; +} + + +char * +qemuBuildFSDevStr(virDomainFSDefPtr fs) +{ + virBuffer opt = VIR_BUFFER_INITIALIZER; + + if (fs->type != VIR_DOMAIN_FS_TYPE_MOUNT) { + qemuReportError(VIR_ERR_CONFIG_UNSUPPORTED, "%s", + _("can only passthrough directories")); + goto error; + } + + virBufferAddLit(&opt, "virtio-9p-pci"); + virBufferVSprintf(&opt, ",id=%s", fs->info.alias); + virBufferVSprintf(&opt, ",fsdev=%s%s", QEMU_FSDEV_HOST_PREFIX, fs->info.alias); + virBufferVSprintf(&opt, ",mount_tag=%s", fs->dst); + qemuBuildDeviceAddressStr(&opt, &fs->info); + + if (virBufferError(&opt)) { + virReportOOMError(); + goto error; + } + + return virBufferContentAndReset(&opt); + +error: + virBufferFreeAndReset(&opt); + return NULL; +} + + char * qemuBuildControllerDevStr(virDomainControllerDefPtr def) { @@ -4274,6 +4347,29 @@ int qemudBuildCommandLine(virConnectPtr conn, } } + if (qemuCmdFlags & QEMUD_CMD_FLAG_FSDEV) { + for (i = 0 ; i < def->nfss ; i++) { + char *optstr; + virDomainFSDefPtr fs = def->fss[i]; + + ADD_ARG_LIT("-fsdev"); + if (!(optstr = qemuBuildFSStr(fs, qemuCmdFlags))) + goto error; + ADD_ARG(optstr); + + ADD_ARG_LIT("-device"); + if (!(optstr = qemuBuildFSDevStr(fs))) + goto error; + ADD_ARG(optstr); + } + } else { + if (def->nfss) { + qemuReportError(VIR_ERR_CONFIG_UNSUPPORTED, "%s", + _("filesystem passthrough not supported by this QEMU")); + goto error; + } + } + if (!def->nnets) { /* If we have -device, then we set -nodefault already */ if (!(qemuCmdFlags & QEMUD_CMD_FLAG_DEVICE)) { diff --git a/src/qemu/qemu_conf.h b/src/qemu/qemu_conf.h index ca4559f..68f10eb 100644 --- a/src/qemu/qemu_conf.h +++ b/src/qemu/qemu_conf.h @@ -91,6 +91,7 @@ enum qemud_cmd_flags { QEMUD_CMD_FLAG_TDF = (1LL << 35), /* -tdf flag (user-mode pit catchup) */ QEMUD_CMD_FLAG_PCI_CONFIGFD = (1LL << 36), /* pci-assign.configfd */ QEMUD_CMD_FLAG_NODEFCONFIG = (1LL << 37), /* -nodefconfig */ + QEMUD_CMD_FLAG_FSDEV = (1LL << 38), /* -fstype filesystem passthrough */ }; /* Main driver state */ @@ -175,6 +176,7 @@ typedef qemuDomainPCIAddressSet *qemuDomainPCIAddressSetPtr; # define QEMU_DRIVE_HOST_PREFIX "drive-" # define QEMU_VIRTIO_SERIAL_PREFIX "virtio-serial" +# define QEMU_FSDEV_HOST_PREFIX "fsdev-" # define qemuReportError(code, ...) \ virReportErrorHelper(NULL, VIR_FROM_QEMU, code, __FILE__, \ @@ -235,9 +237,12 @@ char *qemuDeviceDriveHostAlias(virDomainDiskDefPtr disk, char *qemuBuildDriveStr(virDomainDiskDefPtr disk, int bootable, unsigned long long qemuCmdFlags); +char *qemuBuildFSStr(virDomainFSDefPtr fs, + unsigned long long qemuCmdFlags); /* Current, best practice */ char * qemuBuildDriveDevStr(virDomainDiskDefPtr disk); +char * qemuBuildFSDevStr(virDomainFSDefPtr fs); /* Current, best practice */ char * qemuBuildControllerDevStr(virDomainControllerDefPtr def); diff --git a/tests/qemuxml2argvdata/qemuxml2argv-fs9p.args b/tests/qemuxml2argvdata/qemuxml2argv-fs9p.args new file mode 100644 index 0000000..42c26a2 --- /dev/null +++ b/tests/qemuxml2argvdata/qemuxml2argv-fs9p.args @@ -0,0 +1 @@ +LC_ALL=C PATH=/bin HOME=/home/test USER=test LOGNAME=test /usr/bin/qemu -S -M pc -m 214 -smp 1 -nographic -nodefaults -monitor unix:/tmp/test-monitor,server,nowait -no-acpi -boot c -hda /dev/HostVG/QEMUGuest1 -fsdev local,id=fsdev-fs0,path=/export/to/guest -device virtio-9p-pci,id=fs0,fsdev=fsdev-fs0,mount_tag=/import/from/host,bus=pci.0,addr=0x4 -usb -device virtio-balloon-pci,id=balloon0,bus=pci.0,addr=0x3 diff --git a/tests/qemuxml2argvdata/qemuxml2argv-fs9p.xml b/tests/qemuxml2argvdata/qemuxml2argv-fs9p.xml new file mode 100644 index 0000000..9072ead --- /dev/null +++ b/tests/qemuxml2argvdata/qemuxml2argv-fs9p.xml @@ -0,0 +1,28 @@ +<domain type='qemu'> + <name>QEMUGuest1</name> + <uuid>c7a5fdbd-edaf-9455-926a-d65c16db1809</uuid> + <memory>219200</memory> + <currentMemory>219200</currentMemory> + <vcpu>1</vcpu> + <os> + <type arch='i686' machine='pc'>hvm</type> + <boot dev='hd'/> + </os> + <clock offset='utc'/> + <on_poweroff>destroy</on_poweroff> + <on_reboot>restart</on_reboot> + <on_crash>destroy</on_crash> + <devices> + <emulator>/usr/bin/qemu</emulator> + <disk type='block' device='disk'> + <source dev='/dev/HostVG/QEMUGuest1'/> + <target dev='hda' bus='ide'/> + <address type='drive' controller='0' bus='0' unit='0'/> + </disk> + <controller type='ide' index='0'/> + <filesystem type='mount'> + <source dir='/export/to/guest'/> + <target dir='/import/from/host'/> + </filesystem> + </devices> +</domain> diff --git a/tests/qemuxml2argvtest.c b/tests/qemuxml2argvtest.c index 24002b8..10d1342 100644 --- a/tests/qemuxml2argvtest.c +++ b/tests/qemuxml2argvtest.c @@ -357,6 +357,8 @@ mymain(int argc, char **argv) DO_TEST("sound", 0); DO_TEST("sound-device", QEMUD_CMD_FLAG_DEVICE | QEMUD_CMD_FLAG_NODEFCONFIG); + DO_TEST("fs9p", QEMUD_CMD_FLAG_DEVICE | + QEMUD_CMD_FLAG_NODEFCONFIG | QEMUD_CMD_FLAG_FSDEV); DO_TEST("hostdev-usb-address", 0); DO_TEST("hostdev-usb-address-device", QEMUD_CMD_FLAG_DEVICE | -- 1.7.1.1

2 1

[libvirt] [PATCH 0/4] Re-arrange PCI device ordering
by Daniel P. Berrange 21 Jul '10

21 Jul '10

In an attempt to provide better compatability with older QEMU where static PCI device ordering wasn't available, we re-arrange our device address assignment code. We now try to assign addresses in the same order as QEMU would. This should upset Windows less.

4 17

[libvirt] [PATCH] Set a stable & high MAC addr for guest TAP devices on host
by Daniel P. Berrange 21 Jul '10

21 Jul '10

A Linux software bridge will assume the MAC address of the enslaved interface with the numerically lowest MAC addr. When the bridge changes MAC address there is a period of network blackout, so a change should be avoided. The kernel gives TAP devices a completely random MAC address. Occassionally the random TAP device MAC is lower than that of the physical interface (eth0, eth1etc) that is enslaved, causing the bridge to change its MAC. This change sets an explicit MAC address for all TAP devices created using the configured MAC from the XML, but with the high byte set to 0xFE. This should ensure TAP device MACs are higher than any physical interface MAC. * src/qemu/qemu_conf.c, src/uml/uml_conf.c: Pass in a MAC addr for the TAP device with high byte set to 0xFE * src/util/bridge.c, src/util/bridge.h: Set a MAC when creating the TAP device to override random MAC --- src/qemu/qemu_conf.c | 11 +++++++++-- src/uml/uml_conf.c | 11 +++++++++-- src/util/bridge.c | 42 ++++++++++++++++++++++++++++++++++++++++++ src/util/bridge.h | 3 ++- 4 files changed, 62 insertions(+), 5 deletions(-) diff --git a/src/qemu/qemu_conf.c b/src/qemu/qemu_conf.c index ad41457..d721fcc 100644 --- a/src/qemu/qemu_conf.c +++ b/src/qemu/qemu_conf.c @@ -1608,6 +1608,7 @@ qemudNetworkIfaceConnect(virConnectPtr conn, int tapfd = -1; int vnet_hdr = 0; int template_ifname = 0; + unsigned char tapmac[VIR_MAC_BUFLEN]; if (net->type == VIR_DOMAIN_NET_TYPE_NETWORK) { int active, fail = 0; @@ -1675,8 +1676,14 @@ qemudNetworkIfaceConnect(virConnectPtr conn, net->model && STREQ(net->model, "virtio")) vnet_hdr = 1; - if ((err = brAddTap(driver->brctl, brname, - &net->ifname, vnet_hdr, &tapfd))) { + memcpy(tapmac, net->mac, VIR_MAC_BUFLEN); + tapmac[0] = 0xFE; /* Discourage bridge from using TAP dev MAC */ + if ((err = brAddTap(driver->brctl, + brname, + &net->ifname, + tapmac, + vnet_hdr, + &tapfd))) { if (errno == ENOTSUP) { /* In this particular case, give a better diagnostic. */ qemuReportError(VIR_ERR_INTERNAL_ERROR, diff --git a/src/uml/uml_conf.c b/src/uml/uml_conf.c index 785d627..ea40efb 100644 --- a/src/uml/uml_conf.c +++ b/src/uml/uml_conf.c @@ -113,6 +113,7 @@ umlConnectTapDevice(virDomainNetDefPtr net, int tapfd = -1; int template_ifname = 0; int err; + unsigned char tapmac[VIR_MAC_BUFLEN]; if ((err = brInit(&brctl))) { virReportSystemError(err, "%s", @@ -130,8 +131,14 @@ umlConnectTapDevice(virDomainNetDefPtr net, template_ifname = 1; } - if ((err = brAddTap(brctl, bridge, - &net->ifname, BR_TAP_PERSIST, &tapfd))) { + memcpy(tapmac, net->mac, VIR_MAC_BUFLEN); + tapmac[0] = 0xFE; /* Discourage bridge from using TAP dev MAC */ + if ((err = brAddTap(brctl, + bridge, + &net->ifname, + tapmac, + 0, + &tapfd))) { if (errno == ENOTSUP) { /* In this particular case, give a better diagnostic. */ umlReportError(VIR_ERR_INTERNAL_ERROR, diff --git a/src/util/bridge.c b/src/util/bridge.c index b236f80..7d0caae 100644 --- a/src/util/bridge.c +++ b/src/util/bridge.c @@ -285,6 +285,38 @@ brDeleteInterface(brControl *ctl ATTRIBUTE_UNUSED, # endif /** + * ifSetInterfaceMac: + * @ctl: bridge control pointer + * @ifname: interface name to set MTU for + * @macaddr: MAC address (VIR_MAC_BUFLEN in size) + * + * This function sets the @macaddr for a given interface @ifname. This + * gets rid of the kernel's automatically assigned random MAC. + * + * Returns 0 in case of success or an errno code in case of failure. + */ +static int ifSetInterfaceMac(brControl *ctl, const char *ifname, + const unsigned char *macaddr) +{ + struct ifreq ifr; + + if (!ctl || !ifname) + return EINVAL; + + memset(&ifr, 0, sizeof(struct ifreq)); + if (virStrcpyStatic(ifr.ifr_name, ifname) == NULL) + return EINVAL; + + /* To fill ifr.ifr_hdaddr.sa_family field */ + if (ioctl(ctl->fd, SIOCGIFHWADDR, &ifr) != 0) + return errno; + + memcpy(ifr.ifr_hwaddr.sa_data, macaddr, VIR_MAC_BUFLEN); + + return ioctl(ctl->fd, SIOCSIFHWADDR, &ifr) == 0 ? 0 : errno; +} + +/** * ifGetMtu * @ctl: bridge control pointer * @ifname: interface name get MTU for @@ -430,6 +462,7 @@ brProbeVnetHdr(int tapfd) * @ctl: bridge control pointer * @bridge: the bridge name * @ifname: the interface name (or name template) + * @macaddr: desired MAC address (VIR_MAC_BUFLEN long) * @vnet_hdr: whether to try enabling IFF_VNET_HDR * @tapfd: file descriptor return value for the new tap device * @@ -447,6 +480,7 @@ int brAddTap(brControl *ctl, const char *bridge, char **ifname, + const unsigned char *macaddr, int vnet_hdr, int *tapfd) { @@ -478,6 +512,14 @@ brAddTap(brControl *ctl, if (ioctl(fd, TUNSETIFF, &ifr) < 0) goto error; + /* We need to set the interface MAC before adding it + * to the bridge, because the bridge assumes the lowest + * MAC of all enslaved interfaces & we don't want it + * seeing the kernel allocate random MAC for the TAP + * device before we set our static MAC. + */ + if ((errno = ifSetInterfaceMac(ctl, ifr.ifr_name, macaddr))) + goto error; /* We need to set the interface MTU before adding it * to the bridge, because the bridge will have its * MTU adjusted automatically when we add the new interface. diff --git a/src/util/bridge.h b/src/util/bridge.h index d37d7db..96696ac 100644 --- a/src/util/bridge.h +++ b/src/util/bridge.h @@ -68,7 +68,8 @@ enum { int brAddTap (brControl *ctl, const char *bridge, char **ifname, - int features, + const unsigned char *macaddr, + int vnet_hdr, int *tapfd); int brDeleteTap (brControl *ctl, -- 1.7.1.1

2 1