- Devel - libvirt List Archives

[libvirt] [PATCH] build: Use $(PYTHON) instead of python for the keycode map generator
by Matthias Bolte 22 Jul '11

22 Jul '11

Also prepend $(AM_V_GEN). --- src/Makefile.am | 2 +- 1 files changed, 1 insertions(+), 1 deletions(-) diff --git a/src/Makefile.am b/src/Makefile.am index 481caba..eef0669 100644 --- a/src/Makefile.am +++ b/src/Makefile.am @@ -90,7 +90,7 @@ EXTRA_DIST += $(srcdir)/util/virkeymaps.h $(srcdir)/util/keymaps.csv \ $(srcdir)/util/virkeymaps.h: $(srcdir)/util/keymaps.csv \ $(srcdir)/util/virkeycode-mapgen.py - python $(srcdir)/util/virkeycode-mapgen.py <$(srcdir)/util/keymaps.csv >$@ + $(AM_V_GEN)$(PYTHON) $(srcdir)/util/virkeycode-mapgen.py <$(srcdir)/util/keymaps.csv >$@ $(srcdir)/util/virkeycode.c: $(srcdir)/util/virkeycode.h $(srcdir)/util/virkeymaps.h -- 1.7.4.1

3 5

[libvirt] [PATCH] Pre-create /var/lib/libvirt/sanlock directory
by Daniel P. Berrange 22 Jul '11

22 Jul '11

From: "Daniel P. Berrange" <berrange(a)redhat.com> The sanlock plugin for libvirt expects the directory /var/lib/libvirt/sanlock to exist. Create this and add it to the RPM * libvirt.spec.in: Add /var/lib/libvirt/sanlock * src/Makefile.am: Create /var/lib/libvirt/sanlock --- libvirt.spec.in | 1 + src/Makefile.am | 6 ++++++ 2 files changed, 7 insertions(+), 0 deletions(-) diff --git a/libvirt.spec.in b/libvirt.spec.in index 6cbd9ac..7c8b2ec 100644 --- a/libvirt.spec.in +++ b/libvirt.spec.in @@ -1046,6 +1046,7 @@ fi %attr(0755, root, root) %{_libdir}/libvirt/lock-driver/sanlock.so %{_datadir}/augeas/lenses/libvirt_sanlock.aug %{_datadir}/augeas/lenses/tests/test_libvirt_sanlock.aug +%dir %attr(0700, root, root) %{_localstatedir}/lib/libvirt/sanlock %{_sbindir}/virt-sanlock-cleanup %{_mandir}/man8/virt-sanlock-cleanup.8* %endif diff --git a/src/Makefile.am b/src/Makefile.am index 481caba..c1786cf 100644 --- a/src/Makefile.am +++ b/src/Makefile.am @@ -1396,6 +1396,9 @@ install-data-local: $(MKDIR_P) "$(DESTDIR)$(localstatedir)/cache/libvirt" $(MKDIR_P) "$(DESTDIR)$(localstatedir)/lib/libvirt/images" $(MKDIR_P) "$(DESTDIR)$(localstatedir)/lib/libvirt/boot" +if HAVE_SANLOCK + $(MKDIR_P) "$(DESTDIR)$(localstatedir)/lib/libvirt/sanlock" +endif if WITH_QEMU $(MKDIR_P) "$(DESTDIR)$(localstatedir)/lib/libvirt/qemu" $(MKDIR_P) "$(DESTDIR)$(localstatedir)/run/libvirt/qemu" @@ -1438,6 +1441,9 @@ uninstall-local:: rmdir "$(DESTDIR)$(localstatedir)/cache/libvirt" ||: rmdir "$(DESTDIR)$(localstatedir)/lib/libvirt/images" ||: rmdir "$(DESTDIR)$(localstatedir)/lib/libvirt/boot" ||: +if HAVE_SANLOCK + rmdir "$(DESTDIR)$(localstatedir)/lib/libvirt/sanlock" ||: +endif if WITH_QEMU rmdir "$(DESTDIR)$(localstatedir)/lib/libvirt/qemu" ||: rmdir "$(DESTDIR)$(localstatedir)/run/libvirt/qemu" ||: -- 1.7.6

2 2

[libvirt] [PATCH 00/12] New API: virDominShutdownFlags
by Michal Privoznik 22 Jul '11

22 Jul '11

This series creates new API which adds flags to virDomainShutdown. As mentioned here many times, it is a bad luck all APIs don't have flags parameter. Parameter for this might be needed to select different ways of shutting down a domain. Currently, no new functionality is implemented, but this paves the way for creating multiple ways of domain shutting down. Therefore, calling this new API is the same as calling its predecessor. Michal Privoznik (12): shutdown: Define new public API virDomainShutdownFlags shutdown: Wire up the remote protocol shutdown: Implement internal API for qemu driver shutdown: Implement internal API for ESX driver shutdown: Implement internal API for libxl driver shutdown: Implement internal API for openvz driver shutdown: Implement internal API for phyp driver shutdown: Implement internal API for uml driver shutdown: Implement internal API for vbox driver shutdown: Implement internal API for vmware driver shutdown: Implement internal API for xen driver shutdown: Implement internal API for xenapi driver include/libvirt/libvirt.h.in | 6 ++++ src/driver.h | 4 +++ src/esx/esx_driver.c | 12 ++++++++- src/libvirt.c | 54 ++++++++++++++++++++++++++++++++++++++++++ src/libvirt_public.syms | 1 + src/libxl/libxl_driver.c | 12 ++++++++- src/openvz/openvz_driver.c | 1 + src/phyp/phyp_driver.c | 12 ++++++++- src/qemu/qemu_driver.c | 12 ++++++++- src/remote/remote_driver.c | 1 + src/remote/remote_protocol.x | 8 +++++- src/remote_protocol-structs | 5 ++++ src/uml/uml_driver.c | 15 +++++++++++- src/vbox/vbox_tmpl.c | 14 ++++++++++- src/vmware/vmware_driver.c | 1 + src/xen/xen_driver.c | 14 ++++++++-- src/xen/xen_driver.h | 2 +- src/xen/xen_hypervisor.c | 2 +- src/xen/xen_inotify.c | 2 +- src/xen/xend_internal.c | 10 +++++-- src/xen/xend_internal.h | 2 +- src/xen/xm_internal.c | 2 +- src/xen/xs_internal.c | 10 +++++-- src/xen/xs_internal.h | 3 +- src/xenapi/xenapi_driver.c | 25 ++++++++++++++++++- 25 files changed, 206 insertions(+), 24 deletions(-) -- 1.7.5.rc3

3 16

[libvirt] RFC: Implement virDomainGetIPAddress()
by Michal Novotny 22 Jul '11

22 Jul '11

Hi guys, some time ago I've been investigating the options to get the guest's IP address information without having to connect to guest's VNC window or console. It was for one project I've been working on and I found that the solution lies in the procfs, precisely in the /proc/{PID}/net/arp... The format is as follows: $ cat /proc/{PID}/net/arp IP address HW type Flags HW address Mask Device 192.168.122.36 0x1 0x2 52:54:00:35:76:e6 * virbr0 where the HW address matches the MAC address associated to the guest's NIC. Implementing such an API shouldn't be a big problem however I know that there's some option to run libvirt on Windows machines. It should be just the client so it shouldn't really matter however I'd like to ask you whether it's really not an issue. The function should return a string of the guest's IP address as read from the procfs or return a NULL value if there's no IP address associated with the guest. If the multiple NICs are being used by the guest then the function should return either the IP address matching the MAC address passed to the function or the first IP address if omitted. The prototype should be: char *virDomainGetIPAddress(virDomainPtr domain, char *devmac); For this the internal information about PID of the process spawned should be used (this is already being tracked and stored by libvirt IMHO) to access the /proc/{PID}/net/arp file. This API function would be very useful since it would give libvirt users option to get the information of the guest's IP address without having to connect to guest's VNC window or console. For virsh this could be implemented directly into the dominfo subcommand. What do you think about implementing this ? Thanks, Michal -- Michal Novotny <minovotn(a)redhat.com>, RHCE, Red Hat Virtualization | libvirt-php bindings | php-virt-control.org

3 12

[libvirt] [PATCH 0/3] Support readonly mounts in LXC
by Daniel P. Berrange 22 Jul '11

22 Jul '11

This patch series introduces support for readonly mounts in LXC. In includes general refactoring which will help other patches coming later too.

2 7

[libvirt] [PATCH] xenapi: Fix double-freeing the session in xenapiClose
by Matthias Bolte 22 Jul '11

22 Jul '11

xen_session_logout already frees the whole session object. Don't call xenSessionFree on a freed session object. Reported by Sharmila Radhakrishnan. --- src/xenapi/xenapi_driver.c | 2 +- 1 files changed, 1 insertions(+), 1 deletions(-) diff --git a/src/xenapi/xenapi_driver.c b/src/xenapi/xenapi_driver.c index 97da1d1..dbc03cb 100644 --- a/src/xenapi/xenapi_driver.c +++ b/src/xenapi/xenapi_driver.c @@ -230,7 +230,7 @@ xenapiClose (virConnectPtr conn) if (priv->session != NULL) { xen_session_logout(priv->session); - xenSessionFree(priv->session); + priv->session = NULL; } VIR_FREE(priv->url); -- 1.7.4.1

2 2

[libvirt] [PATCH] remote/ssh: optional "keyfile" parameter.
by Oskari Saarenmaa 22 Jul '11

22 Jul '11

New optional parameter "keyfile" for ssh transport allows the user to select the private key to be used to authenticate to the remote host. --- docs/remote.html.in | 16 ++++++++++++++++ src/remote/remote_driver.c | 9 ++++++++- src/rpc/virnetclient.c | 4 +++- src/rpc/virnetclient.h | 1 + src/rpc/virnetsocket.c | 3 +++ src/rpc/virnetsocket.h | 1 + tests/virnetsockettest.c | 12 ++++++++++++ 7 files changed, 44 insertions(+), 2 deletions(-) diff --git a/docs/remote.html.in b/docs/remote.html.in index 39d65aa..b554950 100644 --- a/docs/remote.html.in +++ b/docs/remote.html.in @@ -275,6 +275,22 @@ Note that parameter values must be <td colspan="2"/> <td> Example: <code>netcat=/opt/netcat/bin/nc</code> </td> </tr> + + <tr> + <td> + <code>keyfile</code> + </td> + <td> ssh </td> + <td> + The name of the private key file to use to authentication to the remote + machine. If this option is not used the default keys are used. + </td> + </tr> + <tr> + <td colspan="2"/> + <td> Example: <code>keyfile=/root/.ssh/example_key</code> </td> + </tr> + <tr> <td> <code>no_verify</code> diff --git a/src/remote/remote_driver.c b/src/remote/remote_driver.c index c2f8bbd..3878fc9 100644 --- a/src/remote/remote_driver.c +++ b/src/remote/remote_driver.c @@ -343,7 +343,7 @@ doRemoteOpen (virConnectPtr conn, char *name = NULL, *command = NULL, *sockname = NULL, *netcat = NULL; char *port = NULL, *authtype = NULL, *username = NULL; int no_verify = 0, no_tty = 0; - char *pkipath = NULL; + char *pkipath = NULL, *keyfile = NULL; /* Return code from this function, and the private data. */ int retcode = VIR_DRV_OPEN_ERROR; @@ -416,6 +416,11 @@ doRemoteOpen (virConnectPtr conn, netcat = strdup (var->value); if (!netcat) goto out_of_memory; var->ignore = 1; + } else if (STRCASEEQ (var->name, "keyfile")) { + VIR_FREE(keyfile); + keyfile = strdup (var->value); + if (!keyfile) goto out_of_memory; + var->ignore = 1; } else if (STRCASEEQ (var->name, "no_verify")) { no_verify = atoi (var->value); var->ignore = 1; @@ -573,6 +578,7 @@ doRemoteOpen (virConnectPtr conn, no_tty, no_verify, netcat ? netcat : "nc", + keyfile, sockname))) goto failed; @@ -672,6 +678,7 @@ doRemoteOpen (virConnectPtr conn, VIR_FREE(sockname); VIR_FREE(authtype); VIR_FREE(netcat); + VIR_FREE(keyfile); VIR_FREE(username); VIR_FREE(port); VIR_FREE(pkipath); diff --git a/src/rpc/virnetclient.c b/src/rpc/virnetclient.c index d3965c6..1bda763 100644 --- a/src/rpc/virnetclient.c +++ b/src/rpc/virnetclient.c @@ -189,11 +189,13 @@ virNetClientPtr virNetClientNewSSH(const char *nodename, bool noTTY, bool noVerify, const char *netcat, + const char *keyfile, const char *path) { virNetSocketPtr sock; - if (virNetSocketNewConnectSSH(nodename, service, binary, username, noTTY, noVerify, netcat, path, &sock) < 0) + if (virNetSocketNewConnectSSH(nodename, service, binary, username, noTTY, + noVerify, netcat, keyfile, path, &sock) < 0) return NULL; return virNetClientNew(sock, NULL); diff --git a/src/rpc/virnetclient.h b/src/rpc/virnetclient.h index 6acdf50..3e5659c 100644 --- a/src/rpc/virnetclient.h +++ b/src/rpc/virnetclient.h @@ -46,6 +46,7 @@ virNetClientPtr virNetClientNewSSH(const char *nodename, bool noTTY, bool noVerify, const char *netcat, + const char *keyfile, const char *path); virNetClientPtr virNetClientNewExternal(const char **cmdargv); diff --git a/src/rpc/virnetsocket.c b/src/rpc/virnetsocket.c index 7ea1ab7..57373a0 100644 --- a/src/rpc/virnetsocket.c +++ b/src/rpc/virnetsocket.c @@ -578,6 +578,7 @@ int virNetSocketNewConnectSSH(const char *nodename, bool noTTY, bool noVerify, const char *netcat, + const char *keyfile, const char *path, virNetSocketPtr *retsock) { @@ -594,6 +595,8 @@ int virNetSocketNewConnectSSH(const char *nodename, virCommandAddArgList(cmd, "-p", service, NULL); if (username) virCommandAddArgList(cmd, "-l", username, NULL); + if (keyfile) + virCommandAddArgList(cmd, "-i", keyfile, NULL); if (noTTY) virCommandAddArgList(cmd, "-T", "-o", "BatchMode=yes", "-e", "none", NULL); diff --git a/src/rpc/virnetsocket.h b/src/rpc/virnetsocket.h index 5f882ac..479be28 100644 --- a/src/rpc/virnetsocket.h +++ b/src/rpc/virnetsocket.h @@ -69,6 +69,7 @@ int virNetSocketNewConnectSSH(const char *nodename, bool noTTY, bool noVerify, const char *netcat, + const char *keyfile, const char *path, virNetSocketPtr *addr); diff --git a/tests/virnetsockettest.c b/tests/virnetsockettest.c index 1697ced..fd8151b 100644 --- a/tests/virnetsockettest.c +++ b/tests/virnetsockettest.c @@ -379,6 +379,7 @@ struct testSSHData { bool noTTY; bool noVerify; const char *netcat; + const char *keyfile; const char *path; const char *expectOut; @@ -400,6 +401,7 @@ static int testSocketSSH(const void *opaque) data->noTTY, data->noVerify, data->netcat, + data->keyfile, data->path, &csock) < 0) goto cleanup; @@ -542,6 +544,16 @@ mymain(void) if (virtTestRun("SSH test 5", 1, testSocketSSH, &sshData5) < 0) ret = -1; + struct testSSHData sshData6 = { + .nodename = "example.com", + .path = "/tmp/socket", + .keyfile = "/root/.ssh/example_key", + .noVerify = true, + .expectOut = "-i /root/.ssh/example_key -o StrictHostKeyChecking=no example.com nc -U /tmp/socket\n", + }; + if (virtTestRun("SSH test 6", 1, testSocketSSH, &sshData6) < 0) + ret = -1; + #endif return (ret==0 ? EXIT_SUCCESS : EXIT_FAILURE); -- 1.7.6

2 1

[libvirt] [PATCH RESEND RFC v4 0/6] support cpu bandwidth in libvirt
by Wen Congyang 22 Jul '11

22 Jul '11

Resend this patchset TODO: 1. We create sub directory for each vcpu in cpu subsystem. So we should recalculate cpu.shares for each vcpu. Changelog: v4: address Adam Litke's comment v3: fix some small bugs implement the simple way v2: almost rewrite the patchset to support to control each vcpu's bandwidth. Limit quota to [-1, 2^64/1000] at the schemas level. We will check it at cgroup level. Wen Congyang (6): Introduce the function virCgroupForVcpu cgroup: Implement cpu.cfs_period_us and cpu.cfs_quota_us tuning API Update XML Schema for new entries qemu: Implement period and quota tunable XML configuration and parsing qemu: Implement cfs_period and cfs_quota's modification doc: Add documentation for new cputune elements period and quota docs/formatdomain.html.in | 19 ++ docs/schemas/domain.rng | 26 ++- src/conf/domain_conf.c | 20 ++- src/conf/domain_conf.h | 2 + src/libvirt_private.syms | 5 + src/qemu/qemu_cgroup.c | 137 ++++++++++ src/qemu/qemu_cgroup.h | 4 + src/qemu/qemu_driver.c | 312 +++++++++++++++++++++-- src/qemu/qemu_process.c | 4 + src/util/cgroup.c | 153 +++++++++++- src/util/cgroup.h | 11 + tests/qemuxml2argvdata/qemuxml2argv-cputune.xml | 2 + 12 files changed, 659 insertions(+), 36 deletions(-)

6 35

[libvirt] [PATCH 0/3] More work on TLS certificate code
by Daniel P. Berrange 22 Jul '11

22 Jul '11

This series started off with an attempt to refactor the TLS certificate validation code to remove duplication. I then realized I ought to be testing such sensitive code too. So the big interesting part of this patch is a test case todo validation using very certificates. Rather than add a bunch of certs to GIT, I have written (nasty) GNUTLS code to generate certs on the fly with a variety of good & bad data in them.

2 7

[libvirt] [PATCHv2 0/9] network: physical device abstraction aka 'virtual switch'
by Laine Stump 22 Jul '11

22 Jul '11

This series deprecates patches 04/10 - 10/10 of the previous series of the same name, posted here: https://www.redhat.com/archives/libvir-list/2011-July/msg00149.html This patch is in response to the following bug reports: https://bugzilla.redhat.com/show_bug.cgi?id=643947 (RHEL) https://bugzilla.redhat.com/show_bug.cgi?id=636106 (upstream) The first 4 patches of the earlier series were reasonable cleanups in their own right, and were ACKed, so I pushed them earlier. I reorganized the remaining 6 patches accord to advice from Dan Berrange and Eric Blake. Some notes about correlation between v1 and v2: [PATCHv2 1/9] util: define MAX A new patch [PATCHv2 2/9] conf: put virtPortProfile struct / functions in a common location Was 05/10 in v1: https://www.redhat.com/archives/libvir-list/2011-July/msg00154.html [PATCHv2 3/9] conf: virDomainNetDef points to (rather than contains) virtPortProfile This was also part of 05/10 in v1. I separated it out to make review of pure code movement (in 2/9) easier. [PATCHv2 4/9] conf: support abstracted interface info in domain interface XML [PATCHv2 5/9] conf: support abstracted interface info in network XML These two were 06/10 and 07/10 in v1: https://www.redhat.com/archives/libvir-list/2011-July/msg00155.html https://www.redhat.com/archives/libvir-list/2011-July/msg00156.html However, they were split differently. In v1 it was split into RNG changes vs. parser changes with network and domain combined in each patch. In v2, I reorganized (based on Dan's advice) to have one patch containing all domain XML changes, from RNG up through parser, test cases, and documentation, and one patch for network XML changes. [PATCHv2 6/9] network: separate Start/Shutdown functions for new network types This was patch 08/10 in v1: https://www.redhat.com/archives/libvir-list/2011-July/msg00157.html It is unchanged from the original. [PATCHv2 7/9] qemu: use virDomainNetGetActual*() functions where appropriate [PATCHv2 8/9] qemu: use virDomainNetGetActual*() in qemuDomainXMLToNative These two were patch 09/10 in v1: https://www.redhat.com/archives/libvir-list/2011-July/msg00158.html In v2, I just split out one piece of code that was more complicated than just replacing direct data references with helper functions, and modified the commit message to be more clear about the "NOP" nature of the change. [PATCHv2 9/9] network: internal API functions to manage assignment of physdev to guest This was 10/10 in v1: https://www.redhat.com/archives/libvir-list/2011-July/msg00159.html The main modification here was to remove all of the #if WITH_NETWORK from qemu_*.c, and instead define static inline NOP functions for the functions I'd previously eliminated references to with #if WITH_NETWORK (in the case that libvirt is build --without-network). This works out very well, except that make syntax-check doesn't like ATTRIBUTE_UNUSED in the inline static functions (I think we should change syntax-check rather than going back to the old #if WITH_NETWORK). Comments about other changes I've made *within* each patch are in the responses to the original patches, as well as in the annotations to the new patches.

3 24