June 2013 - Devel - libvirt List Archives

[libvirt] [PATCH 0/2] Resolve issues seen with schedinfo
by John Ferlan 08 Jun '13

08 Jun '13

These patches resolve an issue seen using 'virsh schedinfo <domain>' on a non running domain that have been present since 1.0.4 as a result of the cgroup infrastructure changes: https://www.redhat.com/archives/libvir-list/2013-April/msg00783.html The exact commit id that caused the issue is listed in each of the commit messages. I used git bisect to determine, although it was tricky because the TPM changes were made around the same time and required commit '8b934a5c' to be applied in order to actually see domains on my host. Prior to the changes the "CFS Bandwidth" data provided by qemuGetCpuBWStatus() and lxcGetCpuBWStatus() would be obtained as a result of the driver cgroup being mounted. Now it relies on the domain cgroup being mounted which only occurs once the domain is active. This issue also affects the libvirt-cim code in how it defines QEMU domains. Fortunately it only looks for the "cpu_shares" value. A "downside" to the end result of these changes is that for a non-running domain it becomes impossible to obtain the vcpu_period, vcpu_quota, emulator_period, and emulator_quota values. All that can be obtained is the 'cpu_shares" value. As an option, I did consider adding the following to either [qemu|lxc]DomainSchedulerGetType() and DomainSchedulerParameter[Flags]() or just [qemu|lxc]GetCpuBWStatus() APIs in order to get the "host" values if they existed (since they are eventually copied into the domain cgroup): if (cgroup == NULL) { rc = virCgroupNewSelf(&hostgrp); if (rc < 0) { virReportSystemError(-rc, "%s", _("Unable to get schedinfo host cgroup")); goto cleanup; } cgroup = hostgrp; } However, I wasn't sure going that route was desired and figured that I'd use the code review opportunity to get the answer. Furthermore, it wasn't clear that the vcpu_* and emulator_* values made sense for a non running domain. Also, the virReportSystemError may need to change to a VIR_INFO since I believe it would be triggered if cgroups weren't mounted on the system. Another option would be to just add the above code to the GetType() APIs and then ignore the 'cpu_bw_status' for the VIR_DOMAIN_AFFECT_CONFIG path in the ParametersFlags() APIs thus returning all 5 datapoints set to whatever the persistentDef had defined. Since this has been present since 1.0.4 and no one has complained so far, I don't think it's critical for 1.0.6. I suspect the change would need to into the maint trees though. It might be nice to get DanB's opinion on this too. John Ferlan (2): qemu: Resolve issue with virsh schedinfo for non running domain lxc: Resolve issue with virsh schedinfo for non running domain src/lxc/lxc_driver.c | 9 ++++++++- src/qemu/qemu_driver.c | 11 ++++++++++- 2 files changed, 18 insertions(+), 2 deletions(-) -- 1.8.1.4

3 8

[libvirt] [PATCHv2 00/14] CPU Hotplug series
by Peter Krempa 07 Jun '13

07 Jun '13

This is a rebased and a bit improved version of the first posting of support for cpu hotplug related stuff. Since nobody reviewed v1 I did't specify changes to the previous version. Peter Krempa (14): virsh-domain-monitor: Remove ATTRIBUTE_UNUSED from a argument qemu: Use bool instead of int in qemuMonitorSetCPU APIs qemu: Extract more information about vCPUs and threads qemu_agent: Introduce helpers for agent based CPU hot(un)plug API: Introduce VIR_DOMAIN_VCPU_AGENT, for agent based CPU hot(un)plug qemu: Implement request of vCPU state using the guest agent qemu: Implement support for VIR_DOMAIN_VCPU in qemuDomainSetVcpusFlags lib: Add API to map virtual cpus of a guest virsh-domain-monitor: Implement command to map guest vCPUs qemu: Implement virDomainGetVCPUMap for the qemu driver qemu: Implement new QMP command for cpu hotplug lib: Add API to modify vCPU state from the guest using the guest agent virsh-domain: Implement command for virDomainSetGuestVcpu qemu: Implement virDomainSetGuestVcpu in qemu driver daemon/remote.c | 54 +++++ include/libvirt/libvirt.h.in | 26 +++ python/generator.py | 1 + python/libvirt-override-api.xml | 7 + python/libvirt-override.c | 66 ++++++ src/driver.h | 14 ++ src/libvirt.c | 153 ++++++++++++- src/libvirt_public.syms | 7 + src/qemu/qemu_agent.c | 148 +++++++++++++ src/qemu/qemu_agent.h | 12 ++ src/qemu/qemu_driver.c | 468 +++++++++++++++++++++++++++++++++++++--- src/qemu/qemu_monitor.c | 11 +- src/qemu/qemu_monitor.h | 13 +- src/qemu/qemu_monitor_json.c | 86 ++++++-- src/qemu/qemu_monitor_json.h | 4 +- src/qemu/qemu_monitor_text.c | 94 ++++---- src/qemu/qemu_monitor_text.h | 4 +- src/qemu/qemu_process.c | 63 ++++-- src/remote/remote_driver.c | 48 +++++ src/remote/remote_protocol.x | 31 ++- src/remote_protocol-structs | 20 ++ tools/virsh-domain-monitor.c | 112 +++++++++- tools/virsh-domain.c | 102 ++++++++- tools/virsh.pod | 42 +++- 24 files changed, 1452 insertions(+), 134 deletions(-) -- 1.8.2.1

3 29

[libvirt] [PATCH] qemuDomainGetVcpusFlags: Initialize ncpuinfo
by Michal Privoznik 07 Jun '13

07 Jun '13

Currently, there's a path to use the ncpuinfo variable uninitialized, which leads to a compiler warning: qemu/qemu_driver.c: In function 'qemuDomainGetVcpusFlags': qemu/qemu_driver.c:4573:9: error: 'ncpuinfo' may be used uninitialized in this function [-Werror=maybe-uninitialized] for (i = 0; i < ncpuinfo; i++) { ^ --- Pushed under trivial & build-breaker rules. src/qemu/qemu_driver.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/src/qemu/qemu_driver.c b/src/qemu/qemu_driver.c index e5f6988..c24f39c 100644 --- a/src/qemu/qemu_driver.c +++ b/src/qemu/qemu_driver.c @@ -4512,7 +4512,7 @@ qemuDomainGetVcpusFlags(virDomainPtr dom, unsigned int flags) int ret = -1; virCapsPtr caps = NULL; qemuAgentCPUInfoPtr cpuinfo = NULL; - int ncpuinfo; + int ncpuinfo = -1; int i; virCheckFlags(VIR_DOMAIN_AFFECT_LIVE | -- 1.8.2.1

1 0

[libvirt] [PATCH RESEND 0/5] VirtualBox version 4.2 support for libvirt vbox driver
by Manuel VIVES 07 Jun '13

07 Jun '13

Hello, I'm re-sending this patch for reviewing. If necessary I'm willing to make some changes to those patches. I'm currently working on a better management for snapshots with virtualbox, and my work is based on Virtualbox 4.2 so that's why I'm re sending this patch. Regards, Manuel VIVES Ryan Woodsmall said originally: "This patch set adds VirtualBox 4.2 initial support for the libvirt vbox driver. I've tested enough to check capabilities, create a VM, destroy it, etc. Five patches total: - Patch 1 is the C API header file from Oracle, cleaned up for libvirt. - Patch 2 is the version specific source file for version dependency. - Patch 3 is the src/Makefile.am change to pick up the two new files. - Patch 4 is the vbox driver support for the new VirtualBox API/version. - Patch 5 is the vbox_tmpl.c template support for the new version. A few things have changed in the VirtualBox API - some small (capitalizations of things in function names like Ip to IP and Dhcp to DHCP) and some much larger (FindMedium is superceded by OpenMedium). The biggest change for the sake of this patch set is the signature of CreateMachine is quite a bit different. Using the Oracle source as a guide, to spin up a VM with a given UUID, it looks like a text flag has to be passed in a new argument to CreateMachine. This flag is built in the VirtualBox 4.2 specific ifdefs and is kind of ugly but works. Additionally, there is now (unused) VM groups support in CreateMachine and the previous 'osTypeId' arg is currently set to nsnull as in the Oracle code. The FindMedium to OpenMedium changes were more straightforward and are pretty clear. The rest of the vbox template changes are basically spelling/capitalization changes from the looks of things. This probably isn't perfect, but it works on git and patched against 0.10.2 for a few quick tests. Not currently on the list, so ping me directly if you need any other info on these, or if anything could use additional work. Thanks! " ryan woodsmall (5): vbox C API header for VirtualBox v4.2 vbox version-specific C file for VirtualBox v4.2 Makefile.am additions for VirtualBox v4.2 vbox driver support for VirtualBox v4.2 vbox template support for VirtualBox v4.2 src/Makefile.am | 3 +- src/vbox/vbox_CAPI_v4_2.h | 8855 +++++++++++++++++++++++++++++++++++++++++++++ src/vbox/vbox_V4_2.c | 13 + src/vbox/vbox_driver.c | 8 + src/vbox/vbox_tmpl.c | 90 +- 5 files changed, 8958 insertions(+), 11 deletions(-) create mode 100644 src/vbox/vbox_CAPI_v4_2.h create mode 100644 src/vbox/vbox_V4_2.c -- 1.7.10.4

2 6

[libvirt] [PATCH v2] iscsi: pass hostnames to iscsiadm instead of resolving them
by Ján Tomko 07 Jun '13

07 Jun '13

iscsiadm now supports specifying hostnames in the portal argument [1] Instead of resolving the hostname to a single IPv4 address, pass the hostname to isciadm, allowing IPv6 targets to work. Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=907171 [1] https://bugzilla.redhat.com/show_bug.cgi?id=624437 --- v1: https://www.redhat.com/archives/libvir-list/2013-May/msg01930.html src/storage/storage_backend_iscsi.c | 69 +++++++------------------------------ 1 file changed, 12 insertions(+), 57 deletions(-) diff --git a/src/storage/storage_backend_iscsi.c b/src/storage/storage_backend_iscsi.c index ad38ab2..0a4cd22 100644 --- a/src/storage/storage_backend_iscsi.c +++ b/src/storage/storage_backend_iscsi.c @@ -24,9 +24,6 @@ #include <config.h> #include <dirent.h> -#include <sys/socket.h> -#include <netdb.h> -#include <sys/types.h> #include <sys/wait.h> #include <string.h> #include <stdio.h> @@ -47,55 +44,12 @@ #define VIR_FROM_THIS VIR_FROM_STORAGE -static int -virStorageBackendISCSITargetIP(const char *hostname, - char *ipaddr, - size_t ipaddrlen) -{ - struct addrinfo hints; - struct addrinfo *result = NULL; - int ret; - - memset(&hints, 0, sizeof(hints)); - hints.ai_flags = AI_ADDRCONFIG; - hints.ai_family = AF_INET; - hints.ai_socktype = SOCK_STREAM; - hints.ai_protocol = 0; - - ret = getaddrinfo(hostname, NULL, &hints, &result); - if (ret != 0) { - virReportError(VIR_ERR_INTERNAL_ERROR, - _("host lookup failed %s"), - gai_strerror(ret)); - return -1; - } - - if (result == NULL) { - virReportError(VIR_ERR_INTERNAL_ERROR, - _("no IP address for target %s"), - hostname); - return -1; - } - - if (getnameinfo(result->ai_addr, result->ai_addrlen, - ipaddr, ipaddrlen, NULL, 0, - NI_NUMERICHOST) < 0) { - virReportError(VIR_ERR_INTERNAL_ERROR, - _("cannot format ip addr for %s"), - hostname); - freeaddrinfo(result); - return -1; - } - - freeaddrinfo(result); - return 0; -} - static char * virStorageBackendISCSIPortal(virStoragePoolSourcePtr source) { - char ipaddr[NI_MAXHOST]; - char *portal; + char *portal = NULL; + const char *host; + int port = 3260; if (source->nhost != 1) { virReportError(VIR_ERR_CONFIG_UNSUPPORTED, "%s", @@ -103,15 +57,16 @@ virStorageBackendISCSIPortal(virStoragePoolSourcePtr source) return NULL; } - if (virStorageBackendISCSITargetIP(source->hosts[0].name, - ipaddr, sizeof(ipaddr)) < 0) - return NULL; + host = source->hosts[0].name; + if (source->hosts[0].port != 0) + port = source->hosts[0].port; - if (virAsprintf(&portal, "%s:%d,1", ipaddr, - source->hosts[0].port ? - source->hosts[0].port : 3260) < 0) { - virReportOOMError(); - return NULL; + if (strchr(host, ':')) { + if (virAsprintf(&portal, "[%s]:%d,1", host, port) < 0) + virReportOOMError(); + } else { + if (virAsprintf(&portal, "%s:%d,1", host, port) < 0) + virReportOOMError(); } return portal; -- 1.8.1.5

2 2

[libvirt] [PATCH v2 1/2] virsocket: Introduce virSocketAddrIsAny
by Michal Privoznik 07 Jun '13

07 Jun '13

This internal API checks, if passed address is ANYCAST address. --- src/libvirt_private.syms | 1 + src/util/virsocketaddr.c | 20 ++++++++++++++++++++ src/util/virsocketaddr.h | 1 + tests/sockettest.c | 38 ++++++++++++++++++++++++++++++++++++++ 4 files changed, 60 insertions(+) diff --git a/src/libvirt_private.syms b/src/libvirt_private.syms index b93629f..1ea7467 100644 --- a/src/libvirt_private.syms +++ b/src/libvirt_private.syms @@ -1782,6 +1782,7 @@ virSocketAddrGetPort; virSocketAddrGetRange; virSocketAddrIsNetmask; virSocketAddrIsPrivate; +virSocketAddrIsWildcard; virSocketAddrMask; virSocketAddrMaskByPrefix; virSocketAddrParse; diff --git a/src/util/virsocketaddr.c b/src/util/virsocketaddr.c index 1071b00..e84c58e 100644 --- a/src/util/virsocketaddr.c +++ b/src/util/virsocketaddr.c @@ -227,6 +227,26 @@ virSocketAddrIsPrivate(const virSocketAddrPtr addr) } /* + * virSocketAddrIsWildcard: + * @addr: address to check + * + * Check if passed address is a variant of ANYCAST address. + */ +bool +virSocketAddrIsWildcard(const virSocketAddrPtr addr) +{ + in_addr_t tmp = INADDR_ANY; + switch (addr->data.stor.ss_family) { + case AF_INET: + return memcmp(&addr->data.inet4.sin_addr.s_addr, &tmp, + sizeof(addr->data.inet4.sin_addr.s_addr)) == 0; + case AF_INET6: + return IN6_IS_ADDR_UNSPECIFIED(&addr->data.inet6.sin6_addr.s6_addr); + } + return false; +} + +/* * virSocketAddrFormat: * @addr: an initialized virSocketAddrPtr * diff --git a/src/util/virsocketaddr.h b/src/util/virsocketaddr.h index 1c9e54a..b28fe6c 100644 --- a/src/util/virsocketaddr.h +++ b/src/util/virsocketaddr.h @@ -123,4 +123,5 @@ bool virSocketAddrEqual(const virSocketAddrPtr s1, const virSocketAddrPtr s2); bool virSocketAddrIsPrivate(const virSocketAddrPtr addr); +bool virSocketAddrIsWildcard(const virSocketAddrPtr addr); #endif /* __VIR_SOCKETADDR_H__ */ diff --git a/tests/sockettest.c b/tests/sockettest.c index 5b36a6c..092de88 100644 --- a/tests/sockettest.c +++ b/tests/sockettest.c @@ -157,6 +157,28 @@ static int testNetmaskHelper(const void *opaque) return testNetmask(data->addr1, data->addr2, data->netmask, data->pass); } +static int testWildcard(const char *addrstr, + bool pass) +{ + virSocketAddr addr; + + if (virSocketAddrParse(&addr, addrstr, AF_UNSPEC) < 0) + return -1; + + if (virSocketAddrIsWildcard(&addr)) + return pass ? 0 : -1; + return pass ? -1 : 0; +} + +struct testWildcardData { + const char *addr; + bool pass; +}; +static int testWildcardHelper(const void *opaque) +{ + const struct testWildcardData *data = opaque; + return testWildcard(data->addr, data->pass); +} static int mymain(void) @@ -223,6 +245,14 @@ mymain(void) ret = -1; \ } while (0) +#define DO_TEST_WILDCARD(addr, pass) \ + do { \ + struct testWildcardData data = { addr, pass}; \ + if (virtTestRun("Test wildcard " addr, 1, \ + testWildcardHelper, &data) < 0) \ + ret = -1; \ + } while (0) + DO_TEST_PARSE_AND_FORMAT("127.0.0.1", AF_UNSPEC, true); DO_TEST_PARSE_AND_FORMAT("127.0.0.1", AF_INET, true); @@ -276,6 +306,14 @@ mymain(void) DO_TEST_NETMASK("2000::1:1", "9000::1:1", "ffff:ffff:ffff:ffff:ffff:ffff:ffff:0", false); + DO_TEST_WILDCARD("0.0.0.0", true); + DO_TEST_WILDCARD("::", true); + DO_TEST_WILDCARD("0", true); + DO_TEST_WILDCARD("0.0", true); + DO_TEST_WILDCARD("0.0.0", true); + DO_TEST_WILDCARD("1", false); + DO_TEST_WILDCARD("0.1", false); + return ret==0 ? EXIT_SUCCESS : EXIT_FAILURE; } -- 1.8.2.1

3 5

[libvirt] [PATCH 0/3 v3] Cgroup fixes
by Osier Yang 07 Jun '13

07 Jun '13

Instead of only post v2 of 5/7 (v2) in the old mail thread, this resend the left patches of v2 with a independant thread for easy reviewing. v2 - v3: * 1/7 ~ 4/7 of v1 are pushed. * goto label "cleanup" is kept with Daniel's suggestion for 5/7 (v1) * Commit log of 6/7 (v1) are updated a bit to reflect more clear about the cause v1 - v2: * Just resending v2: https://www.redhat.com/archives/libvir-list/2013-May/msg01314.html Osier Yang (3): qemu: Abstract code for the cpu controller setting into a helper qemu: Set cpuset.cpus for domain process qemu: Prohibit getting the numa parameters if mode is not strict src/qemu/qemu_cgroup.c | 94 ++++++++++++++++++++++++++++++++++++-------------- src/qemu/qemu_driver.c | 11 ++++-- 2 files changed, 78 insertions(+), 27 deletions(-) -- 1.8.1.4

2 8

[libvirt] libxl: calculating free pages in libvirt
by Dario Faggioli 07 Jun '13

07 Jun '13

Hi Jim, As I told you in Dublin, I'm looking into libvirt a bit, with the main purpose of implementing the NUMA interface for the libxl driver. While at it I noticed that libxlNodeGetFreeMemory() uses the value contained in phy_info.free_pages to check how many pages of free memory we have. However, starting from (Xen's git) commit bec8f17e, the number of free pages should be computed like this: (phy_info.free_pages - phy_info.outstanding_pages) to take the memory claiming mechanism introduced by Oracle properly into account. You can see an example of that, for instance, looking at the output_physinfo() function in tools/libxl/xl_cmdimpl.c (in the Xen codebase, of course :-) ). Said commit is from last May, and the claim mechanism altogether --which includes adding the outstanding_pages field to the libxl_physinfo struct in libxl-- was introduced during the 4.3 development cycle. So, forgive the possibly dumb question, but what's the preferred way to fix this in libvirt, without breaking build with old libxl versions? (Provided this is something libvirt cares about, does it?) For other features added within this last dev cycle, libxl has a `#define LIBXL_HAVE_<foo>' (e.g., LIBXL_HAVE_DOMAIN_NODEAFFINITY), but I don't see one for this particular field... Konrad, Ian, am I missing it? If no, should we add it? Thanks and Regards, Dario -- <<This happens because I choose it to happen!>> (Raistlin Majere) ----------------------------------------------------------------- Dario Faggioli, Ph.D, http://about.me/dario.faggioli Senior Software Engineer, Citrix Systems R&D Ltd., Cambridge (UK)

5 5

[libvirt] [PATCH v3 0/2] BSD implementation of virNetDevTapCreate() and virNetDevTapDelete()
by Roman Bogorodskiy 07 Jun '13

07 Jun '13

Changes from v2: * changed error message in virNetDevSetupControl() as per Dan's suggestion * avoid compile time signature change for virNetDevSetupControl() * fixed memleak in virNetDevTapCreate() Roman Bogorodskiy (2): Make virNetDevSetupControl() public. BSD: implement virNetDevTapCreate() and virNetDevTapDelete() src/libvirt_private.syms | 1 + src/util/virnetdev.c | 15 +++++- src/util/virnetdev.h | 15 +++++- src/util/virnetdevtap.c | 118 ++++++++++++++++++++++++++++++++++++++++++++++- 4 files changed, 144 insertions(+), 5 deletions(-) -- 1.8.0

1 3

[libvirt] [PATCH v2 0/2] BSD implementation of virNetDevTapCreate() and virNetDevTapDelete()
by Roman Bogorodskiy 07 Jun '13

07 Jun '13

Changes from the previous version: * virNetDevSetupControl patch: remove pointless <config.h> include from virnetdev.h * virnetdevtap.c patch: rebase to include new tapfdSize param for virNetDevTapCreate() Roman Bogorodskiy (2): Make virNetDevSetupControl() public. BSD: implement virNetDevTapCreate() and virNetDevTapDelete() src/libvirt_private.syms | 1 + src/util/virnetdev.c | 15 ++++- src/util/virnetdev.h | 11 ++++ src/util/virnetdevmacvlan.c | 2 +- src/util/virnetdevtap.c | 116 ++++++++++++++++++++++++++++++++++++++- src/util/virnetdevvportprofile.c | 2 +- 6 files changed, 141 insertions(+), 6 deletions(-) -- 1.8.0

2 5