- Devel - libvirt List Archives

[libvirt] [PATCH v3 00/14] Network filtering (ACL) extensions for libvirt
by stefanb＠us.ibm.com 25 Mar '10

25 Mar '10

Hi! This is a repost of this set of patches with some of the suggested fixes applied and ipv6 support on the ebtables layer added. The following set of patches add network filtering (ACL) extensions to libvirt and enable network traffic filtering for VMs using ebtables and, depending on the networking technology being used (tap, but not macvtap), also iptables. Usage of either is optional and controlled through filters that a VM is referencing. The ebtables-level filtering is based on the XML derived from the CIM network slide 10 (filtering) from the DMTF website (http://www.dmtf.org/standards/cim/cim_schema_v2230/CIM_Network.pdf) The XML we derived from this was discussed on the list before. On the ebtables level we currently handle filtering of IPv4 and ARP traffic. The iptables-level filtering is based on similar XML where XML nodes described the particular protocol to filter for. Its extensions enable the filtering of traffic using iptables for tcp, udp, icmp, igmp, sctp and 'all' types of traffic. This list of protocols maps to the features supported by iptables and only excludes protocols like 'esp', 'ah' and 'udplite'. Currently only bridging mode is supported and based on availability of the physdev match. The filtering framework adds new libvirt virsh commands for managing the filters. The 5 new commands are: - virsh nwfilter-list - virsh nwfilter-dumpxml <name of filter> - virsh nwfilter-define <name of file containing filter desc.> - virsh nwfilter-undefine <name of filter> - virsh nwfilter-edit <name of filter> Above commands are similar to commands for already existing pools and as such much of the code directly related to the above commands could be borrowed from other drivers. The network filters can either contain rules using the above mentioned XML or contain references to other filters in order to build more complex filters that form some sort of filter tree or can contain both. An example for a filter referencing other filters would be this one here: <filter name='demofilter4' chain='root'> <uuid>66f62d1d-34c1-1421-824f-c62d5ee5e8b6</uuid> <filterref filter='no-mac-spoofing'/> <filterref filter='no-mac-broadcast'/> <filterref filter='no-arp-spoofing'/> <filterref filter='allow-dhcp'> <parameter name='DHCPSERVER' value='10.0.0.1'/> </filterref> <filterref filter='no-other-l2-traffic'/> <filterref filter='recv-only-vm-ipaddress'/> <filterref filter='recv-only-vm-macaddress'/> <filterref filter='l3-test'/> <filterref filter='ipv6test'/> </filter> A filter containing actual rules would look like this: <filter name='no-mac-broadcast' chain='ipv4'> <uuid>ffe2ccd6-edec-7360-1852-6b5ccb553234</uuid> <rule action='drop' direction='out' priority='500'> <mac dstmacaddr='ff:ff:ff:ff:ff:ff'/> </rule> </filter> The filter XML now also holds a priority attribute in the rule. This provides control over the ordering of the applied ebtables/iptables rules beyond their appearance in the XML. The domain XML has been extended to reference a top level filter from within each <interface> XML node. A valid reference to such a top level filter looks like this: <interface type='bridge'> <source bridge='static'/> <filterref filter='demofilter4'> <parameter name='IP' value='9.59.241.151'/> </filterref> </interface> In this XML a parameter IP is passed for instantiation of the referenced filters, that may require the availability of this parameter. In the above case the IP parameter's value describes the value of the IP address of the VM and allows to enable those filters to be instantiated that require this 'IP' variable. If a filter requires a parameter that is not provided, the VM will not start or the interface will not attach to a running VM. Any names of parameters can be provided for instantiation of filters and their names and values only need to pass a regular expression test. In a subsequent patch we will be adding capability to allow users to omit the IP parameter (only) and enable libvirt to learn the IP address of the VM and have it instantiate the filter once it knows it. While virtual machines are running, it is possible to update their filters. For that all running VMs' filter 'trees' are traversed to detect whether the updated filter is referenced by the VM. If so, its ebtables/iptable rules are applied. If one of the VMs' update fails allupdates are rolled back and the filter XML update is rejected. One comment about the instantiation of the rules: Since the XML allows to create nearly any possible combination of parameters to ebtables or iptables commands, I haven't used the ebtables or iptables wrappers. Instead, I am writing ebtables/iptables command into a buffer, add command line options to each one of them as described in the rule's XML, write the buffer into a file and run it as a script. For those commands that are not allowed to fail I am using the following format to run them: cmd="ebtables <some options>" r=`${cmd}` if [ $? -ne 0 ]; then echo "Failure in command ${cmd}." exit 1 fi cmd="..." [...] If one of the command fails in such a batch, the libvirt code is going pick up the error code '1', tear down anything previously established and report an error back. The actual error message shown above is currently not reported back, but can be later on with some changes to the commands running external programs that need to read the script's stdout. One comment to patch 14: It currently #include's a .c file into a .c file only for the reason so I don't have to change too much code once I change code in the underlying patch. So this has to be changed. The patch series works without patch 13, but then only supports ebtables. The patches apply to the current tip. They pass 'make syntax-check' and have been frequently run in valgrind for memory leak checks. Looking forward to your feedback on the patches. Thanks and regards, Stefan and Gerhard

5 42

[libvirt] [PATCH][QEMU driver] Catch cdrom change error
by Ryan Harper 25 Mar '10

25 Mar '10

Currently when we attempt to change the cdrom in a qemu VM the monitor doesn't generate an error if the target filename doesn't exist. I've submitted a patch[1] for this. This patch is the libvirt qemu-driver side which catches the error message from the monitor and reportes the error to libvirt. This means that virsh attach-disk cdrom commands won't appear to succeed when qemu change command actually failed. I've tested this patch on an older libvirt (rebuild libvirt-0.7.0 on my ubuntu host) but haven't rebuilt against git head since I can't quite compile that on my host. 1. http://lists.gnu.org/archive/html/qemu-devel/2010-03/msg01935.html Signed-off-by: Ryan Harper <ryanh(a)us.ibm.com> diff --git a/src/qemu/qemu_monitor_text.c b/src/qemu/qemu_monitor_text.c index 1596e59..f868907 100644 --- a/src/qemu/qemu_monitor_text.c +++ b/src/qemu/qemu_monitor_text.c @@ -905,7 +905,14 @@ int qemuMonitorTextChangeMedia(qemuMonitorPtr mon, qemuReportError(VIR_ERR_OPERATION_FAILED, _("could not eject media on %s: %s"), devname, reply); goto cleanup; - } + } + + /* Could not open message indicates bad filename */ + if (strstr(reply, "\nCould not open ")) { + qemuReportError(VIR_ERR_OPERATION_FAILED, + _("could not change media on %s: %s"), devname, reply); + goto cleanup; + } ret = 0; -- Ryan Harper Software Engineer; Linux Technology Center IBM Corp., Austin, Tx ryanh(a)us.ibm.com

1 0

[libvirt] [PATCH 00/13] Add support for all QEMU event notifications
by Daniel P. Berrange 25 Mar '10

25 Mar '10

This series adds support for (nearly) all QEMU event notifications that are currently supported by the new JSON mode monitor. - Guest reboot - RTC adjustment - Watchdog firing - Block IO errors - Graphics client connect/disconnect It required two new public APIs, and a bunch of extra function callback typedefs which will be shown in each following commit message. This scheme is extensible to support any further events against domains. It does not support events on other types of object (storage, node devices, etc), but those can follow exactly the same API design scheme should we need them in the future. We don't want to unify them at the public API level, since internally they need to dispatch to separate drivers anyway. I'm generally happy with the public API for the events, with the exception of the graphics event which is horribly complex due to the large amount of data. I'm not sure if there's a way to make that event nicer.... The full diffstat for all patches daemon/dispatch.h | 10 daemon/libvirtd.c | 20 daemon/libvirtd.h | 2 daemon/remote.c | 348 ++++++++++++++-- daemon/remote_dispatch_args.h | 2 daemon/remote_dispatch_prototypes.h | 16 daemon/remote_dispatch_table.h | 37 + examples/domain-events/events-c/event-test.c | 168 ++++++- include/libvirt/libvirt.h.in | 106 ++++ src/conf/domain_event.c | 580 ++++++++++++++++++++++++--- src/conf/domain_event.h | 82 +++ src/driver.c | 4 src/driver.h | 18 src/esx/esx_driver.c | 2 src/libvirt.c | 133 +++++- src/libvirt_private.syms | 16 src/libvirt_public.syms | 6 src/lxc/lxc_driver.c | 60 ++ src/opennebula/one_driver.c | 2 src/openvz/openvz_driver.c | 2 src/phyp/phyp_driver.c | 2 src/qemu/qemu_driver.c | 290 +++++++++++++ src/qemu/qemu_monitor.c | 77 +++ src/qemu/qemu_monitor.h | 39 + src/qemu/qemu_monitor_json.c | 144 ++++++ src/remote/remote_driver.c | 459 +++++++++++++++++---- src/remote/remote_protocol.c | 111 +++++ src/remote/remote_protocol.h | 185 ++++++-- src/remote/remote_protocol.x | 65 ++- src/test/test_driver.c | 58 ++ src/uml/uml_driver.c | 2 src/vbox/vbox_tmpl.c | 155 +++++-- src/xen/xen_driver.c | 71 ++- src/xenapi/xenapi_driver.c | 2 34 files changed, 2947 insertions(+), 327 deletions(-) Daniel

3 36

[libvirt] [PATCH 0/1] Disk error policy
by David Allan 25 Mar '10

25 Mar '10

The following patch adds support for setting the disk error policy to the domain XML and using it when setting up the qemu command line. It specifies the werror=somepolicy,rerror=somepolicy flags to tell qemu what to do in the case of an io error. This patch is my first foray into the setup of the qemu command line, so please review with heightened care. Dave David Allan (1): Add disk error policy to domain XML docs/schemas/domain.rng | 12 +++++++++++- src/conf/domain_conf.c | 15 +++++++++++++++ src/conf/domain_conf.h | 10 ++++++++++ src/libvirt_private.syms | 2 +- src/qemu/qemu_conf.c | 12 +++++++++--- tests/qemuhelptest.c | 1 + tests/qemuxml2argvtest.c | 3 +++ 7 files changed, 50 insertions(+), 5 deletions(-)

4 4

[libvirt] [PATCH] Add entry point logging for cpu functions
by Jiri Denemark 25 Mar '10

25 Mar '10

--- src/cpu/cpu.c | 46 ++++++++++++++++++++++++++++++++++++++++++++-- 1 files changed, 44 insertions(+), 2 deletions(-) diff --git a/src/cpu/cpu.c b/src/cpu/cpu.c index be0f15e..183862a 100644 --- a/src/cpu/cpu.c +++ b/src/cpu/cpu.c @@ -23,6 +23,7 @@ #include <config.h> +#include "logging.h" #include "memory.h" #include "xml.h" #include "cpu.h" @@ -73,6 +74,8 @@ cpuCompareXML(virCPUDefPtr host, virCPUDefPtr cpu = NULL; virCPUCompareResult ret = VIR_CPU_COMPARE_ERROR; + VIR_DEBUG("host=%p, xml=%s", host, NULLSTR(xml)); + if (!(doc = virXMLParseString(xml, "cpu.xml"))) goto cleanup; @@ -104,6 +107,8 @@ cpuCompare(virCPUDefPtr host, { struct cpuArchDriver *driver; + VIR_DEBUG("host=%p, cpu=%p", host, cpu); + if ((driver = cpuGetSubDriver(host->arch)) == NULL) return VIR_CPU_COMPARE_ERROR; @@ -126,6 +131,13 @@ cpuDecode(virCPUDefPtr cpu, { struct cpuArchDriver *driver; + VIR_DEBUG("cpu=%p, data=%p, nmodels=%u", cpu, data, nmodels); + if (models) { + unsigned int i; + for (i = 0; i < nmodels; i++) + VIR_DEBUG("models[%u]=%s", i, NULLSTR(models[i])); + } + if (models == NULL && nmodels != 0) { virCPUReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("nonzero nmodels doesn't match with NULL models")); @@ -163,6 +175,11 @@ cpuEncode(const char *arch, { struct cpuArchDriver *driver; + VIR_DEBUG("arch=%s, cpu=%p, forced=%p, required=%p, " + "optional=%p, disabled=%p, forbidden=%p", + NULLSTR(arch), cpu, forced, required, + optional, disabled, forbidden); + if ((driver = cpuGetSubDriver(arch)) == NULL) return -1; @@ -184,6 +201,8 @@ cpuDataFree(const char *arch, { struct cpuArchDriver *driver; + VIR_DEBUG("arch=%s, data=%p", NULLSTR(arch), data); + if (data == NULL) return; @@ -206,6 +225,8 @@ cpuNodeData(const char *arch) { struct cpuArchDriver *driver; + VIR_DEBUG("arch=%s", NULLSTR(arch)); + if ((driver = cpuGetSubDriver(arch)) == NULL) return NULL; @@ -227,6 +248,8 @@ cpuGuestData(virCPUDefPtr host, { struct cpuArchDriver *driver; + VIR_DEBUG("host=%p, guest=%p, data=%p", host, guest, data); + if ((driver = cpuGetSubDriver(host->arch)) == NULL) return VIR_CPU_COMPARE_ERROR; @@ -254,6 +277,16 @@ cpuBaselineXML(const char **xmlCPUs, char *cpustr; unsigned int i; + VIR_DEBUG("ncpus=%u, nmodels=%u", ncpus, nmodels); + if (xmlCPUs) { + for (i = 0; i < ncpus; i++) + VIR_DEBUG("xmlCPUs[%u]=%s", i, NULLSTR(xmlCPUs[i])); + } + if (models) { + for (i = 0; i < nmodels; i++) + VIR_DEBUG("models[%u]=%s", i, NULLSTR(models[i])); + } + if (xmlCPUs == NULL && ncpus != 0) { virCPUReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("nonzero ncpus doesn't match with NULL xmlCPUs")); @@ -320,6 +353,17 @@ cpuBaseline(virCPUDefPtr *cpus, { struct cpuArchDriver *driver; virCPUDefPtr cpu; + unsigned int i; + + VIR_DEBUG("ncpus=%u, nmodels=%u", ncpus, nmodels); + if (cpus) { + for (i = 0; i < ncpus; i++) + VIR_DEBUG("cpus[%u]=%p", i, cpus[i]); + } + if (models) { + for (i = 0; i < nmodels; i++) + VIR_DEBUG("models[%u]=%s", i, NULLSTR(models[i])); + } if (cpus == NULL && ncpus != 0) { virCPUReportError(VIR_ERR_INTERNAL_ERROR, @@ -349,8 +393,6 @@ cpuBaseline(virCPUDefPtr *cpus, } if ((cpu = driver->baseline(cpus, ncpus, models, nmodels))) { - int i; - cpu->type = VIR_CPU_TYPE_GUEST; cpu->match = VIR_CPU_MATCH_EXACT; VIR_FREE(cpu->arch); -- 1.7.0.3

2 2

[libvirt] Sleeping 50 microseconds in qemuDomainWaitForMigrationComplete?
by Matthias Bolte 25 Mar '10

25 Mar '10

There is this nanosleep call in the QEMU driver. /* Poll every 50ms for progress & to allow cancellation */ struct timespec ts = { .tv_sec = 0, .tv_nsec = 50 * 1000ull }; [...] nanosleep(&ts, NULL); The comment says 50ms (assuming m is for milli and not for micro. because normally u is used for micro) but the actual nanosleep call is done with 50 microseconds. Is the comment right and the code wrong? I assume so, because 50 microseconds is pretty short and 50 milliseconds sounds more reasonable. Matthias

2 1

[libvirt] [PATCH 0/6] Support VNC password changes on running guests v2
by Daniel P. Berrange 25 Mar '10

25 Mar '10

This patch introduces a new virDomainUpdateDeviceFlags() API to allow the configuration of a device to be changed on the fly. It ports the existing CDROM media change code in all drivers to use this new API (keeps old support too). It then adds VNC password change support in the QEMU driver using this new API v2: Fix feedback from first posting

3 14

Re: [libvirt] [Qemu-devel] Re: Supporting hypervisor specific APIs in libvirt
by Anthony Liguori 25 Mar '10

25 Mar '10

On 03/24/2010 07:25 AM, Paul Brook wrote: >> I can't quite see what such a libqemu would buy us compared to straight >> QMP. >> >> Talking QMP should be easy, provided you got a suitable JSON library. >> > I agree. My undesranding is this was one of the large motivations behind using > JSON: It's a common protocol that already has convenient bindings in most > languages. If it's hard[1] for third parties to bind QMP to their favourite > language/framework then IMHO we've done it wrong. > You can't have convenient bindings to an RPC in C because it doesn't support dynamic dispatch. With most types of RPC, you have an IDL description and a code generator. But regardless of that, there are advantages to us providing a libqemu. The biggest one is that we can standardize transport implementations that include discovery mechanisms. If the core of libqemu provided an extensible transport interface, and a generic QMP request/completion mechanism, in a Python binding, you would never use the IDL generated wrappers but instead use dynamic dispatch to invoke arbitrary QMP requests. But the advantage is that if libvirt provided an API for a QMP transport encapsulated in their secure protocol, then provided the plumbed that API through their Python interface, you could use it for free in Python without having to reinvent the wheel. Regards, Anthony Liguori > Paul > > [1] Hard compared to any other sane RPC mechanism. Some languages make > everything hard :-) >

2 1

[libvirt] [PATCH] build: suppress distracting build output
by Jim Meyering 24 Mar '10

24 Mar '10

While building "make check", I saw this scroll by (too quickly to see what it was), but it was different enough from the rest that... CCLD libvirt.la CCLD libvirt_test.la make check-local make[3]: Entering directory `/home/j/w/co/libvirt/src' if test -x '/usr/bin/augparse'; then \ '/usr/bin/augparse' -I ./qemu \ ./qemu/test_libvirtd_qemu.aug; \ fi if test -x '/usr/bin/augparse'; then \ '/usr/bin/augparse' -I ./lxc \ ./lxc/test_libvirtd_lxc.aug; \ fi CC libvirtd-event.o CC libvirtd-libvirtd.o I scrolled back to investigate. Not a problem, but let's eliminate the distraction. Here's a patch. With it, instead of the above "if test..." lines, you'll see this single line: GEN augeas-check >From 84eb6effdd0241a8bae196d4b77d873aea7fd3db Mon Sep 17 00:00:00 2001 From: Jim Meyering <meyering(a)redhat.com> Date: Wed, 24 Mar 2010 09:39:54 +0100 Subject: [PATCH] build: suppress distracting build output * src/Makefile.am (augeas-check): New target, just to give the existing rule a name. At the same time, prefix the commands with $(AM_V_GEN), to avoid unexpected build output with V=0 which is the default. --- src/Makefile.am | 9 ++++++--- 1 files changed, 6 insertions(+), 3 deletions(-) diff --git a/src/Makefile.am b/src/Makefile.am index 0aa3443..5f6b325 100644 --- a/src/Makefile.am +++ b/src/Makefile.am @@ -766,15 +766,18 @@ EXTRA_DIST += \ $(SECRET_DRIVER_SOURCES) \ $(VBOX_DRIVER_EXTRA_DIST) -check-local: +check-local: augeas-check + +.PHONY: augeas-check +augeas-check: if WITH_QEMU - if test -x '$(AUGPARSE)'; then \ + $(AM_V_GEN)if test -x '$(AUGPARSE)'; then \ '$(AUGPARSE)' -I $(srcdir)/qemu \ $(srcdir)/qemu/test_libvirtd_qemu.aug; \ fi endif if WITH_LXC - if test -x '$(AUGPARSE)'; then \ + $(AM_V_GEN)if test -x '$(AUGPARSE)'; then \ '$(AUGPARSE)' -I $(srcdir)/lxc \ $(srcdir)/lxc/test_libvirtd_lxc.aug; \ fi -- 1.7.0.3.435.g097f4

2 1

[libvirt] [PATCH] vbox: Fix segfault on empty device source
by Matthias Bolte 24 Mar '10

24 Mar '10

<source file=''/> results in def->disks[i]->src == NULL. But vboxDomainDefineXML didn't check def->disks[i]->src for NULL and expected it to be a valid string. Add checks for def->disks[i]->src != NULL to fix the segfault. --- src/vbox/vbox_tmpl.c | 6 ++++-- 1 files changed, 4 insertions(+), 2 deletions(-) diff --git a/src/vbox/vbox_tmpl.c b/src/vbox/vbox_tmpl.c index 1765d63..fefed1f 100644 --- a/src/vbox/vbox_tmpl.c +++ b/src/vbox/vbox_tmpl.c @@ -3519,7 +3519,8 @@ static virDomainPtr vboxDomainDefineXML(virConnectPtr conn, const char *xml) { DEBUG("disk(%d) shared: %s", i, def->disks[i]->shared ? "True" : "False"); if (def->disks[i]->device == VIR_DOMAIN_DISK_DEVICE_CDROM) { - if (def->disks[i]->type == VIR_DOMAIN_DISK_TYPE_FILE) { + if (def->disks[i]->type == VIR_DOMAIN_DISK_TYPE_FILE && + def->disks[i]->src != NULL) { IDVDDrive *dvdDrive = NULL; /* Currently CDROM/DVD Drive is always IDE * Secondary Master so neglecting the following @@ -3801,7 +3802,8 @@ static virDomainPtr vboxDomainDefineXML(virConnectPtr conn, const char *xml) { DEBUG("disk(%d) readonly: %s", i, def->disks[i]->readonly ? "True" : "False"); DEBUG("disk(%d) shared: %s", i, def->disks[i]->shared ? "True" : "False"); - if (def->disks[i]->type == VIR_DOMAIN_DISK_TYPE_FILE) { + if (def->disks[i]->type == VIR_DOMAIN_DISK_TYPE_FILE && + def->disks[i]->src != NULL) { IMedium *medium = NULL; PRUnichar *mediumUUID = NULL; PRUnichar *mediumFileUtf16 = NULL; -- 1.6.3.3

3 4