- Devel - libvirt List Archives

[libvirt] [RFC PATCH] Let network bridge MTU to be configured on XML
by Eduardo Habkost 04 Dec '08

04 Dec '08

Hi, Below is an incomplete patch for letting the MTU of a libvirt virtual network to be configured on XML. This implementation has an issue: it works when trying to set the MTU below 1500, but the kernel bridge code doesn't let us to set the bridge MTU above 1500 if there is no interface attached to the bridge yet. To allow setting the bridge MTU higher than 1500, we need to save the configured value somewhere and set it only when attaching tap interface to it. To solve that, I propose changing the way the network driver interface expose things for the domain network code. Today we do the following (src/qemu_conf.c): virNetworkPtr network; virDomainNetDefPtr net; char *brname; int tapfd; brname = virNetworkGetBridgeName(network); brAddTap(brctl, brname, &net->ifname, &tapfd))) With this code, the only information that can flow between the network driver and the bridge code is the bridge name. I was going to suggest having something like this: virNetworkAddTap(network, &net->ifname, &tapfd); However, we wouldn't be able to return an FD when using the 'remote' network driver. But we can split the creation of the tap interface and attaching it to the network bridge, like this: brNewTap(brctl, &net->ifname, &tapfd); virNetworkAttachInterface(network, net->ifname); This way, the network driver code can do anything needed to attach the new tap interface to the virtual network, including but not limited to MTU setting. Another alternative could be creating a virNetworkGetBridgeMtu() function. That would solve our problem for the MTU settings, but in my opinion, attaching interfaces to the network is a task for the network driver. Having the qemu code calling brAddTap() directly was just a shortcut we won't be able to use anymore. What do you think? --- src/bridge.c | 19 +++++++++++++++++++ src/bridge.h | 4 ++++ src/libvirt_sym.version.in | 1 + src/network_conf.c | 5 +++++ src/network_conf.h | 2 ++ src/network_driver.c | 10 ++++++++++ 6 files changed, 41 insertions(+), 0 deletions(-) diff --git a/src/bridge.c b/src/bridge.c index 13d81bc..ed9023b 100644 --- a/src/bridge.c +++ b/src/bridge.c @@ -452,6 +452,25 @@ brAddTap(brControl *ctl, return errno; } + +/** + * brSetMtu + * @ctl: bridge control pointer + * @bridge: the bridge name + * @mtu: the MTU value to be set + * + * Sets the MTU of a bridge. + * + * Returns 0 in case of success or an errno code in case of failure. + */ +int +brSetMtu(brControl *ctl, + const char *bridge, + int mtu) +{ + return ifSetMtu(ctl, bridge, mtu); +} + /** * brSetInterfaceUp: * @ctl: bridge control pointer diff --git a/src/bridge.h b/src/bridge.h index 7eb6166..194e96f 100644 --- a/src/bridge.h +++ b/src/bridge.h @@ -58,6 +58,10 @@ int brDeleteInterface (brControl *ctl, const char *bridge, const char *iface); +int brSetMtu (brControl *ctl, + const char *bridge, + int mtu); + int brAddTap (brControl *ctl, const char *bridge, char **ifname, diff --git a/src/libvirt_sym.version.in b/src/libvirt_sym.version.in index ec9ff3d..668b902 100644 --- a/src/libvirt_sym.version.in +++ b/src/libvirt_sym.version.in @@ -274,6 +274,7 @@ LIBVIRT_PRIVATE_@VERSION@ { global: /* bridge.h */ brAddBridge; + brSetMtu; brAddInterface; brAddTap; brDeleteBridge; diff --git a/src/network_conf.c b/src/network_conf.c index 5add62e..b7ec87e 100644 --- a/src/network_conf.c +++ b/src/network_conf.c @@ -295,6 +295,7 @@ virNetworkDefParseXML(virConnectPtr conn, { virNetworkDefPtr def; char *tmp; + long mtu; if (VIR_ALLOC(def) < 0) { virNetworkReportError(conn, VIR_ERR_NO_MEMORY, NULL); @@ -332,6 +333,8 @@ virNetworkDefParseXML(virConnectPtr conn, /* Parse bridge information */ def->bridge = virXPathString(conn, "string(./bridge[1]/@name)", ctxt); + if (virXPathLong(conn, "number(./bridge[1]/@mtu)", ctxt, &mtu) == 0) + def->bridgeMtu = mtu; tmp = virXPathString(conn, "string(./bridge[1]/@stp)", ctxt); def->stp = (tmp && STREQ(tmp, "off")) ? 0 : 1; VIR_FREE(tmp); @@ -570,6 +573,8 @@ char *virNetworkDefFormat(virConnectPtr conn, virBufferAddLit(&buf, " <bridge"); if (def->bridge) virBufferEscapeString(&buf, " name='%s'", def->bridge); + if (def->bridgeMtu > 0) + virBufferVSprintf(&buf, " mtu='%d'", def->bridgeMtu); virBufferVSprintf(&buf, " stp='%s' forwardDelay='%ld' />\n", def->stp ? "on" : "off", def->delay); diff --git a/src/network_conf.h b/src/network_conf.h index 2614e47..a30cc8d 100644 --- a/src/network_conf.h +++ b/src/network_conf.h @@ -61,6 +61,8 @@ struct _virNetworkDef { char *name; char *bridge; /* Name of bridge device */ + int bridgeMtu; /* MTU of the bridge */ + char *domain; unsigned long delay; /* Bridge forward delay (ms) */ int stp : 1; /* Spanning tree protocol */ diff --git a/src/network_driver.c b/src/network_driver.c index 96d3ddf..10bdd10 100644 --- a/src/network_driver.c +++ b/src/network_driver.c @@ -684,6 +684,16 @@ static int networkStartNetworkDaemon(virConnectPtr conn, return -1; } + if (network->def->bridgeMtu> 0) { + if ((err = brSetMtu(driver->brctl, network->def->bridge, network->def->bridgeMtu))) { + networkReportError(conn, NULL, NULL, VIR_ERR_INTERNAL_ERROR, + _("cannot set MTU of bridge '%s' to %d: %s"), + network->def->bridge, network->def->bridgeMtu, + strerror(err)); + goto err_delbr; + } + } + if (brSetForwardDelay(driver->brctl, network->def->bridge, network->def->delay) < 0) goto err_delbr; -- 1.5.5.GIT -- Eduardo

3 6

[libvirt] Release of libvirt-0.5.1
by Daniel Veillard 04 Dec '08

04 Dec '08

Mostly a bug fix one, as promised when 0.5.0 came out :-) Available as usual from: ftp://libvirt.org/libvirt/ Some bug fixes and local improvements: + Portability: - fix missing dep in spec file - fix compilation with new NUMA libraries (Daniel Berrange) - udev compatibility for RHEL (Chris Lalancette) - + Documentation: - documentation copy and paste errors and typo (Cole Robinson) + Bug fixes: - add a delay in storage backend for disks to show up (Chris Lalancette) - fix parsing for CDRom device with no source (Daniel Berrange) - use xenstore to list domains to avoid some bugs (Guido Gunther) - remove a leak in xen inotify code (Daniel Berrange) - UML driver freeing of uninitialialized variable (Ron Yorston) - fix UML inotify code (Daniel Berrange) - crash when adding storage without a format (Cole Robinson) + Improvements: - use xend preferably to hypervisor call to set Xen max memory (Jim Fehlig) - allow remote://hostname/ URI for automatic probe of hypervisors (Daniel Berrange) - fix daemon configuration regression testing (Jim Meyering ) - check /usr/bin/kvm for QEmu driver init (Guido Gunther) - proper active vs. inactive differentiation (Guido Gunther) - improve MTU setting on tap interfaces (Eduardo Habkost) - increase timeout for initial QEmu monitor poll (Cole Robinson) + Cleanups: - fix improper initialisations (Jim Meyering) thanks everybody for the contributions, whatever the way ! Daniel -- Daniel Veillard | libxml Gnome XML XSLT toolkit http://xmlsoft.org/ daniel(a)veillard.com | Rpmfind RPM search engine http://rpmfind.net/ http://veillard.com/ | virtualization library http://libvirt.org/

1 0

[libvirt] [PATCH] Add default format info for storage volumes
by Cole Robinson 04 Dec '08

04 Dec '08

The attached patch adds default format info for storage volumes. This fixes a regression in the latest release, where defining a storage volume without a format crashed libvirtd: https://bugzilla.redhat.com/show_bug.cgi?id=473071 This is similar to how we handle things with storage pools as well. This fixes the crash for me. Thanks, Cole

3 2

[libvirt] [PATCH] Increase initial qemu monitor read timeout
by Cole Robinson 04 Dec '08

04 Dec '08

See https://bugzilla.redhat.com/show_bug.cgi?id=453491 We've been getting bug reports like the above against virt-manager for a while now: installing new VMs consistently throws an error 'Timed out while reading monitor startup output'. The user can then just retry the install and it will go off without a hitch. I only just realized that the common thread between the reports is physical install media: apparently the qemu monitor isn't too responsive while it's waiting for a cold cd drive to spin up. The attached patch jacks up the timeout from 3 seconds to 10 seconds when doing the initial monitor read. I verified that this fixes the issue. Thanks, Cole

2 2

[libvirt] [PATCH] Fix a few docs errors
by Cole Robinson 04 Dec '08

04 Dec '08

The attached patch fixes a few cut and paste docs errors I've stumbled upon, and a spelling error in a virsh command output. Thanks, Cole

2 2

[libvirt] [PATCH] bridge: set MTU of tap iface to the same value of the bridge
by Eduardo Habkost 04 Dec '08

04 Dec '08

Hi, This is a follow-up to the RFC patch sent by Chris Wright, at: https://www.redhat.com/archives/libvir-list/2008-November/msg00225.html With this patch, instead of hardcoding the MTU to an high value, we set the MTU of the tap interface to the same MTU value set on the bridge interface. The current code uses the default tap MTU value and lowers the bridge interface MTU without any need, killing performance. I will send another patch later for allowing the tap MTU to be configured on the XML, but the behavior added by this patch will be kept as the default. The default behavior should be enough for most use cases where a larger MTU is wanted for the bridged interfaces. Signed-off-by: Eduardo Habkost <ehabkost(a)redhat.com> --- src/bridge.c | 98 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 1 files changed, 98 insertions(+), 0 deletions(-) diff --git a/src/bridge.c b/src/bridge.c index 4090163..60334b6 100644 --- a/src/bridge.c +++ b/src/bridge.c @@ -276,6 +276,98 @@ brDeleteInterface(brControl *ctl ATTRIBUTE_UNUSED, #endif /** + * ifGetMtu + * @ctl: bridge control pointer + * @ifname: interface name get MTU for + * + * This function gets the @mtu value set for a given interface @ifname. + * + * Returns the MTU value in case of success. + * On error, returns -1 and sets errno accordingly + */ +static int ifGetMtu(brControl *ctl, const char *ifname) +{ + struct ifreq ifr; + int len; + + if (!ctl || !ifname) { + errno = EINVAL; + return -1; + } + + if ((len = strlen(ifname)) >= BR_IFNAME_MAXLEN) { + errno = EINVAL; + return -1; + } + + memset(&ifr, 0, sizeof(struct ifreq)); + + strncpy(ifr.ifr_name, ifname, len); + ifr.ifr_name[len] = '\0'; + + if (ioctl(ctl->fd, SIOCGIFMTU, &ifr)) + return -1; + + return ifr.ifr_mtu; + +} + +/** + * ifSetMtu: + * @ctl: bridge control pointer + * @ifname: interface name to set MTU for + * @mtu: MTU value + * + * This function sets the @mtu for a given interface @ifname. Typically + * used on a tap device to set up for Jumbo Frames. + * + * Returns 0 in case of success or an errno code in case of failure. + */ +static int ifSetMtu(brControl *ctl, const char *ifname, int mtu) +{ + struct ifreq ifr; + int len; + + if (!ctl || !ifname) + return EINVAL; + + if ((len = strlen(ifname)) >= BR_IFNAME_MAXLEN) + return EINVAL; + + memset(&ifr, 0, sizeof(struct ifreq)); + + strncpy(ifr.ifr_name, ifname, len); + ifr.ifr_name[len] = '\0'; + ifr.ifr_mtu = mtu; + + return ioctl(ctl->fd, SIOCSIFMTU, &ifr) == 0 ? 0 : errno; +} + +/** + * brSetInterfaceMtu + * @ctl: bridge control pointer + * @bridge: name of the bridge interface + * @ifname: name of the interface whose MTU we want to set + * + * Sets the interface mtu to the same MTU of the bridge + * + * Returns 0 in case of success or an errno code in case of failure. + */ +static int brSetInterfaceMtu(brControl *ctl, + const char *bridge, + const char *ifname) +{ + int mtu = ifGetMtu(ctl, bridge); + + fprintf(stderr, "mtu: %d\n", mtu); + + if (mtu < 0) + return errno; + + return ifSetMtu(ctl, ifname, mtu); +} + +/** * brAddTap: * @ctl: bridge control pointer * @bridge: the bridge name @@ -334,6 +426,12 @@ brAddTap(brControl *ctl, } if (ioctl(fd, TUNSETIFF, &try) == 0) { + /* 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. + */ + if ((errno = brSetInterfaceMtu(ctl, bridge, try.ifr_name))) + goto error; if ((errno = brAddInterface(ctl, bridge, try.ifr_name))) goto error; if ((errno = brSetInterfaceUp(ctl, try.ifr_name, 1))) -- 1.5.5.GIT

5 8

[libvirt] [PATCH] also look for /usr/bin/kvm
by Guido Günther 04 Dec '08

04 Dec '08

We do so in qemudCapsInitGuest() but not in qemudProbe(). O.k. to apply? -- Guido

3 3

[libvirt] is libvirt-python default in libvirt compilation from source?
by Shanmuga Rajan 03 Dec '08

03 Dec '08

I was trying to install libvirt 0.5.0 from source.. It seems it went well. but when i tried to import libvirt module in python i got error no module named libvirt then i searched entire system for libvirt.py then i came to know that python bindings for libvirt is not installed. is it that python bindings for libvirt not default in compilation....? if i have to install python bindings then what option should i use? i tried to find option from ./configure --help from libvirt source directory but i didnt get anything. Can any one help me in this. Thanks and Regards, Shan

2 1

[libvirt] Allowing <interface type="none"> for network interface definition
by Gihan Munasinghe 03 Dec '08

03 Dec '08

Hi I am writing a management stack for our platform using libvirt. We have many DomU images which all have PV drivers loaded for network.I want to start these DomU guests with only the PV network device being visible from within the DomU. This is a requirement in the platform and hopefully, I presume there are other people with the same requirement.. But when creating a server, libvirt sends (type ioemu) tag to xend. Which make all of our vm's to have 2 network cards. Using the normal xm command I can send type=none ( vif = [ "mac=00:16:3e:00:a5:57,bridge=eth0,script=vif-bridge,type=none"]) to xend. Which will send qeum-dm "-net none" switch which solves the problem. After doing some investigation, I did some code change in libvirt0.5.0 to support a attribute called "none" with in <interface type="">, that will send xend the tag (type none). Allowing qeum-dm to be configured with "-net none" switch. ;-) So you can define your network tag as <interface type='none'>   <mac address='00:16:3e:00:a5:57'/> <source bridge='eth0'/> <target dev='vif47.0'/> </interface> I this the proper way of doing it? :-\ Would this be a general functionality that libvirt community will be interested in (I think this add bit more flexibility to libvirt of the way user want to write their own management stack). If so could I contribute the patch to the mailing list. Looking forward to hearing from you all, I am new to libvirt therefore any comments will be greatly appreciated :-) Thanks Gihan -- Gihan Munasinghe R&D Team Leader XCalibre Communications Ltd. www.flexiscale.com

2 2

[libvirt] [PATCH 1/2] Java bindings for domain events
by David Lively 02 Dec '08

02 Dec '08

The attached patch (against libvirt-java) implements Java bindings for libvirt domain events. This version provides a libvirt EventImpl running in its own Java Thread, and provides per-Connect synchronization that makes using the bindings thread-safe. (Note the Domain, Network, StoragePool, and StorageVol methods also synchronize on their Connect object, as required by libvirt. I have similar changes for NodeDevice.java that need to be made when that code is checked in.) This version of the patch also implements and uses an enum class (DomainEvent.Type), as suggested by Tóth István. IMPORTANT: THIS PATCH WILL BREAK THINGS UNLESS THE NEXT [PATCH 2/2] IS APPLIED TO libvirt FIRST. Also, libvirt must be compiled WITH_PTHREADS for Java events to work. Dave

3 12