June 2015 - Devel - libvirt List Archives

[libvirt] [PATCH] Fix nodeinfo output on PPC64 KVM hosts
by Shivaprasad G Bhat 03 Jul '15

03 Jul '15

The nodeinfo is reporting incorrect number of cpus and incorrect host topology on PPC64 KVM hosts. The KVM hypervisor on PPC64 needs only the primary thread in a core to be online, and the secondaries offlined. While scheduling a guest in, the kvm scheduler wakes up the secondaries to run in guest context. The host scheduling of the guests happen at the core level(as only primary thread is online). The kvm scheduler exploits as many threads of the core as needed by guest. Further, starting POWER8, the processor allows splitting a physical core into multiple subcores with 2 or 4 threads each. Again, only the primary thread in a subcore is online in the host. The KVM-PPC scheduler allows guests to exploit all the offline threads in the subcore, by bringing them online when needed. (Kernel patches on split-core http://www.spinics.net/lists/kvm-ppc/msg09121.html) Recently with dynamic micro-threading changes in ppc-kvm, makes sure to utilize all the offline cpus across guests, and across guests with different cpu topologies. (https://www.mail-archive.com/kvm@vger.kernel.org/msg115978.html) Since the offline cpus are brought online in the guest context, it is safe to count them as online. Nodeinfo today discounts these offline cpus from cpu count/topology calclulation, and the nodeinfo output is not of any help and the host appears overcommited when it is actually not. The patch carefully counts those offline threads whose primary threads are online. The host topology displayed by the nodeinfo is also fixed when the host is in valid kvm state. Test results with and without fix: http://ur1.ca/mx25y Signed-off-by: Shivaprasad G Bhat <sbhat(a)linux.vnet.ibm.com> --- src/nodeinfo.c | 85 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 85 insertions(+) diff --git a/src/nodeinfo.c b/src/nodeinfo.c index 2fafe2d..1a27080 100644 --- a/src/nodeinfo.c +++ b/src/nodeinfo.c @@ -32,6 +32,8 @@ #include <sys/utsname.h> #include <sched.h> #include "conf/domain_conf.h" +#include <fcntl.h> +#include <sys/ioctl.h> #if defined(__FreeBSD__) || defined(__APPLE__) # include <sys/time.h> @@ -428,6 +430,10 @@ virNodeParseNode(const char *node, unsigned int cpu; int online; int direrr; + int kvmfd = -1; + int lastonline; + int threads_per_subcore = 1; + bool valid_ppc64_kvmhost_mode = false; *threads = 0; *cores = 0; @@ -464,6 +470,66 @@ virNodeParseNode(const char *node, sock_max++; + if (ARCH_IS_PPC64(arch)) { + /* PPC-KVM needs the secondary threads of a core to be offline on the + * host. The kvm schedular brings the secondary threads online in the + * guest context. Moreover, P8 processor has split-core capability + * where, there can be 1,2 or 4 subcores per core. The primaries of the + * subcores alone will be online on the host for a subcore in the + * host. Even though the actual threads per core for P8 processor is 8, + * depending on the subcores_per_core = 1, 2 or 4, the threads per + * subcore will vary accordingly to 8, 4 and 2 repectively. + * So, On host threads_per_core what is arrived at from sysfs in the + * current logic is actually the subcores_per_core. Threads per subcore + * can only be obtained from the kvm device. For example, on P8 wih 1 + * core having 8 threads, sub_cores_percore=4, the threads 0,2,4 & 6 + * will be online. The sysfs reflects this and in the current logic + * variable 'threads' will be 4 which is nothing but subcores_per_core. + * If the user tampers the cpu online/offline states using chcpu or other + * means, then it is an unsupported configuration for kvm. + * The code below tries to keep in mind + * - when the libvirtd is run inside a KVM guest or Phyp based guest. + * - Or on the kvm host where user manually tampers the cpu states to + * offline/online randomly. + */ +# if HAVE_LINUX_KVM_H +# include <linux/kvm.h> + kvmfd = open("/dev/kvm", O_RDONLY); + if (kvmfd >= 0) { +# ifdef KVM_CAP_PPC_SMT + valid_ppc64_kvmhost_mode = true; + /* For Phyp and KVM based guests the ioctl for KVM_CAP_PPC_SMT + * returns zero and both primary and secondary threads will be + * online. Dont try to alter or interpret anything here. + */ + threads_per_subcore = ioctl(kvmfd, KVM_CHECK_EXTENSION, KVM_CAP_PPC_SMT); + if (threads_per_subcore == 0) + valid_ppc64_kvmhost_mode = false; +# endif + } + VIR_FORCE_CLOSE(kvmfd); +# endif + rewinddir(cpudir); + lastonline = -1; + while (valid_ppc64_kvmhost_mode && + ((direrr = virDirRead(cpudir, &cpudirent, node)) > 0)) { + if (sscanf(cpudirent->d_name, "cpu%u", &cpu) != 1) + continue; + if ((online = virNodeGetCpuValue(node, cpu, "online", 1)) < 0) + goto cleanup; + + if (online) { + if (lastonline != -1 && + (cpu - lastonline) < threads_per_subcore) { + valid_ppc64_kvmhost_mode = false; /* if any of secondaries + * online */ + break; + } + lastonline = cpu; + } + } + } + /* allocate cpu maps for each socket */ if (VIR_ALLOC_N(core_maps, sock_max) < 0) goto cleanup; @@ -473,6 +539,7 @@ virNodeParseNode(const char *node, /* iterate over all CPU's in the node */ rewinddir(cpudir); + lastonline = -1; while ((direrr = virDirRead(cpudir, &cpudirent, node)) > 0) { if (sscanf(cpudirent->d_name, "cpu%u", &cpu) != 1) continue; @@ -480,6 +547,17 @@ virNodeParseNode(const char *node, if ((online = virNodeGetCpuValue(node, cpu, "online", 1)) < 0) goto cleanup; + if (ARCH_IS_PPC64(arch) && valid_ppc64_kvmhost_mode) { + if (online) { + lastonline = cpu; + } else if (lastonline != -1 && + (cpu-lastonline) < threads_per_subcore) { + processors++; /* Count only those secondaries whose primary + subcore is online */ + continue; + } + } + if (!online) { (*offline)++; continue; @@ -528,6 +606,13 @@ virNodeParseNode(const char *node, *cores = core; } + if (ARCH_IS_PPC64(arch) && valid_ppc64_kvmhost_mode) { + /* The actual host threads per core is + * multiple of the subcores_per_core(i.e variable *threads in this case) + * and threads_per_subcore. + */ + *threads = *threads * threads_per_subcore; + } ret = processors; cleanup:

3 4

[libvirt] [PATCH] virsh: report error if vcpu number exceed the guest maxvcpu number
by Luyao Huang 03 Jul '15

03 Jul '15

If usr pass a vcpu which exceed guest maxvcpu number, virsh client will only output an header like this: # virsh vcpupin test3 1000 VCPU: CPU Affinity ---------------------------------- After this patch: # virsh vcpupin test3 1000 error: vcpu 1000 is out of range of cpu count 2 Signed-off-by: Luyao Huang <lhuang(a)redhat.com> --- tools/virsh-domain.c | 5 +++++ 1 file changed, 5 insertions(+) diff --git a/tools/virsh-domain.c b/tools/virsh-domain.c index 27d62e9..681fc1a 100644 --- a/tools/virsh-domain.c +++ b/tools/virsh-domain.c @@ -6497,6 +6497,11 @@ cmdVcpuPin(vshControl *ctl, const vshCmd *cmd) goto cleanup; } + if (got_vcpu && vcpu >= ncpus) { + vshError(ctl, _("vcpu %d is out of range of cpu count %d"), vcpu, ncpus); + goto cleanup; + } + cpumaplen = VIR_CPU_MAPLEN(maxcpu); cpumap = vshMalloc(ctl, ncpus * cpumaplen); if ((ncpus = virDomainGetVcpuPinInfo(dom, ncpus, cpumap, -- 1.8.3.1

4 5

[libvirt] [PATCH 00/20] cpu_map.xml: Make CPU definitions easier to review
by Jiri Denemark 02 Jul '15

02 Jul '15

Inheritance among CPU model is cool but it makes reviewing CPU model definitions and comparing them to CPU models from QEMU rather hard and unpleasant. Let's define all CPU models from scratch with a sorted list of features. Jiri Denemark (20): cpu_map.xml: Sort features in x86 CPU models cpu_map.xml: Expand 486 CPU model cpu_map.xml: Expand pentium CPU model cpu_map.xml: Expand pentium2 CPU model cpu_map.xml: Expand pentiumpro CPU model cpu_map.xml: Expand coreduo CPU model cpu_map.xml: Expand n270 CPU model cpu_map.xml: Expand qemu32 CPU model cpu_map.xml: Expand kvm32 CPU model cpu_map.xml: Expand cpu64-rhel5 CPU model cpu_map.xml: Expand kvm64 CPU model cpu_map.xml: Expand Conroe CPU model cpu_map.xml: Expand Penryn CPU model cpu_map.xml: Expand Nehalem CPU model cpu_map.xml: Expand Westmere CPU model cpu_map.xml: Expand SandyBridge CPU model cpu_map.xml: Expand Haswell-noTSX CPU model cpu_map.xml: Expand Opteron_G1 CPU model cpu_map.xml: Expand Opteron_G2 CPU model cpu_map.xml: Expand Opteron_G4 CPU model src/cpu/cpu_map.xml | 829 +++++++++++++++++++++++++++++++++++++++++++++------- 1 file changed, 729 insertions(+), 100 deletions(-) -- 2.4.4

3 49

[libvirt] [PATCH] Revert "Introduce QEMU_CAPS_ARM_VIRT_PCI"
by Peter Krempa 02 Jul '15

02 Jul '15

The capability was not used up to the feature freeze. This reverts commit 7f3515b4bb677d0ead1887547efc844e4761268a. --- src/qemu/qemu_capabilities.c | 4 ---- src/qemu/qemu_capabilities.h | 1 - 2 files changed, 5 deletions(-) diff --git a/src/qemu/qemu_capabilities.c b/src/qemu/qemu_capabilities.c index e7002a3..27686c3 100644 --- a/src/qemu/qemu_capabilities.c +++ b/src/qemu/qemu_capabilities.c @@ -287,7 +287,6 @@ VIR_ENUM_IMPL(virQEMUCaps, QEMU_CAPS_LAST, "aarch64-off", "vhost-user-multiqueue", /* 190 */ - "arm-virt-pci", ); @@ -1337,9 +1336,6 @@ virQEMUCapsComputeCmdFlags(const char *help, virQEMUCapsSet(qemuCaps, QEMU_CAPS_VNC_SHARE_POLICY); } - if (version >= 2003000) - virQEMUCapsSet(qemuCaps, QEMU_CAPS_ARM_VIRT_PCI); - return 0; } diff --git a/src/qemu/qemu_capabilities.h b/src/qemu/qemu_capabilities.h index f4180a8..30aa504 100644 --- a/src/qemu/qemu_capabilities.h +++ b/src/qemu/qemu_capabilities.h @@ -230,7 +230,6 @@ typedef enum { QEMU_CAPS_DEVICE_PCI_SERIAL = 188, /* -device pci-serial */ QEMU_CAPS_CPU_AARCH64_OFF = 189, /* -cpu ...,aarch64=off */ QEMU_CAPS_VHOSTUSER_MULTIQUEUE = 190, /* vhost-user with -netdev queues= */ - QEMU_CAPS_ARM_VIRT_PCI = 191, /* ARM 'virt' machine has PCI bus */ QEMU_CAPS_LAST, /* this must always be the last item */ } virQEMUCapsFlags; -- 2.4.1

5 6

[libvirt] [PATCH] storage: add RBD support to disk source pool translation
by Thibault VINCENT 01 Jul '15

01 Jul '15

Hello, We needed this to avoid redundancies when defining domains with RBD backends, because monitor list and authentication secret could not be pulled from the storage pool. I didn't find any bit of documentation to update, actually it already implies that disks with source type 'volume' would work with all kinds of backing pools. But mpath, sheepdog, and gluster would still require an implementation if it's possible. Using RBD volumes this way still has a limitation, we cannot list snapshots in libvirt pools. Whereas they could be accessed read-only if the domain is using a network source explicitely with a 'snapshot' attribute. How could we implement this so that snapshots are seen as pool volumes? There are real use-cases for this like attaching to a VM a backup from it's own disks. Also it would work the same way with ZFS snapshots, but AFAIK not for LVM pools. Cheers Thibault VINCENT (1): storage: add RBD support to generic disk source pool translation src/storage/storage_driver.c | 49 +++++++++++++++++++++++++++++++++++++++++++- 1 file changed, 48 insertions(+), 1 deletion(-) -- 2.1.4

3 9

[libvirt] [PATCH 0/2][1.2.17] qemu: fix address allocation on chardev hotplug
by Ján Tomko 01 Jul '15

01 Jul '15

For bug https://bugzilla.redhat.com/show_bug.cgi?id=1230039 Patch 1/2 fixes a regression in channel device hotplug which was introduced in the last release and I would like to get it fixed before the next one. Patch 2/2 is not that important, but still a bugfix. Luyao Huang (1): qemu: fix address allocation on chardev attach Ján Tomko (1): qemu: properly free addresses on non-serial chardev unplug src/qemu/qemu_command.c | 4 +++ src/qemu/qemu_hotplug.c | 78 +++++++++++++++++++++++++++---------------------- 2 files changed, 47 insertions(+), 35 deletions(-) -- 2.3.6

3 4

[libvirt] [PATCH] conf: qemu: Taint VMs using custom device tree blob
by Peter Krempa 01 Jul '15

01 Jul '15

Using a custom device tree image may cause unexpected behavior in architectures that use this approach to detect platform devices. Since usually the device tree is generated by qemu and thus it's not normally used let's taint VMs using it to make it obvious as a possible source of problems. --- src/conf/domain_conf.c | 3 ++- src/conf/domain_conf.h | 1 + src/qemu/qemu_domain.c | 3 +++ 3 files changed, 6 insertions(+), 1 deletion(-) diff --git a/src/conf/domain_conf.c b/src/conf/domain_conf.c index 2c3b96b..9280a76 100644 --- a/src/conf/domain_conf.c +++ b/src/conf/domain_conf.c @@ -102,7 +102,8 @@ VIR_ENUM_IMPL(virDomainTaint, VIR_DOMAIN_TAINT_LAST, "external-launch", "host-cpu", "hook-script", - "cdrom-passthrough"); + "cdrom-passthrough", + "custom-dtb"); VIR_ENUM_IMPL(virDomainVirt, VIR_DOMAIN_VIRT_LAST, "qemu", diff --git a/src/conf/domain_conf.h b/src/conf/domain_conf.h index aeba5a5..2bdbff1 100644 --- a/src/conf/domain_conf.h +++ b/src/conf/domain_conf.h @@ -2305,6 +2305,7 @@ typedef enum { VIR_DOMAIN_TAINT_HOST_CPU, /* Host CPU passthrough in use */ VIR_DOMAIN_TAINT_HOOK, /* Domain (possibly) changed via hook script */ VIR_DOMAIN_TAINT_CDROM_PASSTHROUGH,/* CDROM passthrough */ + VIR_DOMAIN_TAINT_CUSTOM_DTB, /* Custom device tree blob was specifed */ VIR_DOMAIN_TAINT_LAST } virDomainTaintFlags; diff --git a/src/qemu/qemu_domain.c b/src/qemu/qemu_domain.c index 6213fd9..22ffa83 100644 --- a/src/qemu/qemu_domain.c +++ b/src/qemu/qemu_domain.c @@ -2105,6 +2105,9 @@ void qemuDomainObjCheckTaint(virQEMUDriverPtr driver, for (i = 0; i < obj->def->nnets; i++) qemuDomainObjCheckNetTaint(driver, obj, obj->def->nets[i], logFD); + if (obj->def->os.dtb) + qemuDomainObjTaint(driver, obj, VIR_DOMAIN_TAINT_CUSTOM_DTB, logFD); + virObjectUnref(cfg); } -- 2.4.1

2 2

[libvirt] [PATCH] qemu: fix not end the job after use OpenGraphics(FD) and get fail when exit monitor
by Luyao Huang 01 Jul '15

01 Jul '15

If guest unexpect exit(qemu process be killed) and get failed when exit the monitor, guest job still handled by old function, this will make guest cannot start later. Need call qemuDomainObjEndJob to release job status before unref vm. Signed-off-by: Luyao Huang <lhuang(a)redhat.com> --- src/qemu/qemu_driver.c | 8 ++------ 1 file changed, 2 insertions(+), 6 deletions(-) diff --git a/src/qemu/qemu_driver.c b/src/qemu/qemu_driver.c index 2b530c8..b1c9f08 100644 --- a/src/qemu/qemu_driver.c +++ b/src/qemu/qemu_driver.c @@ -17255,10 +17255,8 @@ qemuDomainOpenGraphics(virDomainPtr dom, qemuDomainObjEnterMonitor(driver, vm); ret = qemuMonitorOpenGraphics(priv->mon, protocol, fd, "graphicsfd", (flags & VIR_DOMAIN_OPEN_GRAPHICS_SKIPAUTH) != 0); - if (qemuDomainObjExitMonitor(driver, vm) < 0) { + if (qemuDomainObjExitMonitor(driver, vm) < 0) ret = -1; - goto cleanup; - } qemuDomainObjEndJob(driver, vm); cleanup: @@ -17327,10 +17325,8 @@ qemuDomainOpenGraphicsFD(virDomainPtr dom, qemuDomainObjEnterMonitor(driver, vm); ret = qemuMonitorOpenGraphics(priv->mon, protocol, pair[1], "graphicsfd", (flags & VIR_DOMAIN_OPEN_GRAPHICS_SKIPAUTH)); - if (qemuDomainObjExitMonitor(driver, vm) < 0) { + if (qemuDomainObjExitMonitor(driver, vm) < 0) ret = -1; - goto cleanup; - } qemuDomainObjEndJob(driver, vm); if (ret < 0) goto cleanup; -- 1.8.3.1

3 2

[libvirt] [PATCH 0/2] conf: qemu: Fix memory auditing
by Peter Krempa 01 Jul '15

01 Jul '15

Peter Krempa (2): conf: audit: Audit physical memory size rather than balloon request qemu: Audit memory size with memory hotplug operations src/conf/domain_audit.c | 3 ++- src/qemu/qemu_driver.c | 6 +----- src/qemu/qemu_hotplug.c | 20 ++++++++++++++------ 3 files changed, 17 insertions(+), 12 deletions(-) -- 2.4.1

2 4

[libvirt] [PATCHv2] virt-aa-helper: Fix permissions for vhost-user socket files
by Michal Dubiel 01 Jul '15

01 Jul '15

QEMU working in vhost-user mode communicates with the other end (i.e. some virtual router application) via unix domain sockets. This requires that permissions for the socket files are correctly written into /etc/apparmor.d/libvirt/libvirt-UUID.files. Signed-off-by: Michal Dubiel <md(a)semihalf.com> --- Changes since v1: - Removed unnecessary stat() call and dead 'else' block src/security/virt-aa-helper.c | 25 ++++++++++++------------- 1 file changed, 12 insertions(+), 13 deletions(-) diff --git a/src/security/virt-aa-helper.c b/src/security/virt-aa-helper.c index 35423b5..f39932e 100644 --- a/src/security/virt-aa-helper.c +++ b/src/security/virt-aa-helper.c @@ -32,7 +32,6 @@ #include <unistd.h> #include <errno.h> #include <sys/types.h> -#include <sys/stat.h> #include <fcntl.h> #include <getopt.h> #include <sys/utsname.h> @@ -542,7 +541,6 @@ array_starts_with(const char *str, const char * const *arr, const long size) static int valid_path(const char *path, const bool readonly) { - struct stat sb; int npaths, opaths; const char * const restricted[] = { "/bin/", @@ -592,17 +590,6 @@ valid_path(const char *path, const bool readonly) if (!virFileExists(path)) { vah_warning(_("path does not exist, skipping file type checks")); - } else { - if (stat(path, &sb) == -1) - return -1; - - switch (sb.st_mode & S_IFMT) { - case S_IFSOCK: - return 1; - break; - default: - break; - } } opaths = sizeof(override)/sizeof(*(override)); @@ -1101,6 +1088,18 @@ get_files(vahControl * ctl) } } + for (i = 0; i < ctl->def->nnets; i++) { + if (ctl->def->nets[i] && + ctl->def->nets[i]->type == VIR_DOMAIN_NET_TYPE_VHOSTUSER && + ctl->def->nets[i]->data.vhostuser) { + virDomainChrSourceDefPtr vhu = ctl->def->nets[i]->data.vhostuser; + + if (vah_add_file_chardev(&buf, vhu->data.nix.path, "rw", + vhu->type) != 0) + goto cleanup; + } + } + if (ctl->def->virtType == VIR_DOMAIN_VIRT_KVM) { for (i = 0; i < ctl->def->nnets; i++) { virDomainNetDefPtr net = ctl->def->nets[i]; -- 1.9.1

3 2