- Devel - libvirt List Archives

[libvirt] [PATCH 1/2] numa_conf: Introduce virDomainNumaGetMaxCPUID
by Michal Privoznik 07 Aug '15

07 Aug '15

This function should return the greatest CPU number set in /domain/cpu/numa/cell/@cpus. The idea is that we should compare the returned value against /domain/vcpu value. Yes, there exist users who think the following is a good idea: <vcpu placement='static'>4</vcpu> <cpu mode='host-model'> <model fallback='allow'/> <numa> <cell id='0' cpus='0-1' memory='1048576' unit='KiB'/> <cell id='1' cpus='9-10' memory='2097152' unit='KiB'/> </numa> </cpu> Signed-off-by: Michal Privoznik <mprivozn(a)redhat.com> --- src/conf/numa_conf.c | 17 +++++++++++++++++ src/conf/numa_conf.h | 3 +++ src/libvirt_private.syms | 1 + 3 files changed, 21 insertions(+) diff --git a/src/conf/numa_conf.c b/src/conf/numa_conf.c index 57da215..5c123b9 100644 --- a/src/conf/numa_conf.c +++ b/src/conf/numa_conf.c @@ -847,6 +847,23 @@ virDomainNumaGetCPUCountTotal(virDomainNumaPtr numa) return ret; } +unsigned int +virDomainNumaGetMaxCPUID(virDomainNumaPtr numa) +{ + size_t i; + unsigned int ret = 0; + + for (i = 0; i < numa->nmem_nodes; i++) { + int bit; + + bit = virBitmapLastSetBit(virDomainNumaGetNodeCpumask(numa, i)); + if (bit > ret) + ret = bit; + } + + return ret; +} + virDomainNumaPtr virDomainNumaNew(void) diff --git a/src/conf/numa_conf.h b/src/conf/numa_conf.h index 6739065..90deacb 100644 --- a/src/conf/numa_conf.h +++ b/src/conf/numa_conf.h @@ -99,6 +99,9 @@ unsigned long long virDomainNumaGetNodeMemorySize(virDomainNumaPtr numa, unsigned long long virDomainNumaGetMemorySize(virDomainNumaPtr numa) ATTRIBUTE_NONNULL(1); +unsigned int +virDomainNumaGetMaxCPUID(virDomainNumaPtr numa); + /* * Formatters */ diff --git a/src/libvirt_private.syms b/src/libvirt_private.syms index 0517c24..0b67bfe 100644 --- a/src/libvirt_private.syms +++ b/src/libvirt_private.syms @@ -684,6 +684,7 @@ virDomainNumaGetNodeCount; virDomainNumaGetNodeCpumask; virDomainNumaGetNodeMemoryAccessMode; virDomainNumaGetNodeMemorySize; +virDomainNumaGetMaxCPUID; virDomainNumaNew; virDomainNumaSetNodeMemorySize; virDomainNumatuneFormatNodeset; -- 2.4.6

2 2

[libvirt] [PATCH 0/3] qemu: Return true pinning info
by Martin Kletzander 07 Aug '15

07 Aug '15

First two patches just prepare the ground for the third one that explains what needs to be fixed and ho it's done. Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1162947 Martin Kletzander (3): conf: Pass private data to Parse function of XML options qemu: Keep numad hint after reboot qemu: Return true pining info when using numad src/conf/domain_conf.c | 2 +- src/conf/domain_conf.h | 17 +++++++++-------- src/libxl/libxl_domain.c | 3 ++- src/lxc/lxc_domain.c | 3 ++- src/qemu/qemu_domain.c | 33 ++++++++++++++++++++++++++++++++- src/qemu/qemu_driver.c | 11 +++++++++++ 6 files changed, 57 insertions(+), 12 deletions(-) -- 2.4.6

2 8

[libvirt] [PATCH] qemu: Fix reporting of physical capacity for block devices
by Peter Krempa 07 Aug '15

07 Aug '15

Qemu reports physical size 0 for block devices. As 15fa84acbb55ebfee6a4 changed the behavior of qemuDomainGetBlockInfo to just query the monitor this created a regression since we didn't report the size correctly any more. This patch adds code to refresh the physical size of a block device by opening it and seeking to the end and uses it both in qemuDomainGetBlockInfo and also in qemuDomainGetStatsOneBlock that was broken since it was introduced in this respect. Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1250982 --- src/libvirt_private.syms | 1 + src/qemu/qemu_driver.c | 18 ++++++++++++++++-- src/util/virstoragefile.c | 39 +++++++++++++++++++++++++++++++++++++++ src/util/virstoragefile.h | 2 ++ 4 files changed, 58 insertions(+), 2 deletions(-) diff --git a/src/libvirt_private.syms b/src/libvirt_private.syms index 0517c24..5d320b7 100644 --- a/src/libvirt_private.syms +++ b/src/libvirt_private.syms @@ -2156,6 +2156,7 @@ virStorageSourceParseRBDColonString; virStorageSourcePoolDefFree; virStorageSourcePoolModeTypeFromString; virStorageSourcePoolModeTypeToString; +virStorageSourceUpdateBlockPhysicalSize; virStorageTypeFromString; virStorageTypeToString; diff --git a/src/qemu/qemu_driver.c b/src/qemu/qemu_driver.c index 7a984a8..6998e12 100644 --- a/src/qemu/qemu_driver.c +++ b/src/qemu/qemu_driver.c @@ -11827,8 +11827,16 @@ qemuDomainGetBlockInfo(virDomainPtr dom, info->allocation = entry->wr_highest_offset; } + if (entry->physical) { + info->physical = entry->physical; + } else { + if (virStorageSourceUpdateBlockPhysicalSize(disk->src, true) < 0) + goto endjob; + + info->physical = disk->src->physical; + } + info->capacity = entry->capacity; - info->physical = entry->physical; ret = 0; @@ -19274,9 +19282,15 @@ qemuDomainGetStatsOneBlock(virQEMUDriverPtr driver, if (entry->capacity) QEMU_ADD_BLOCK_PARAM_ULL(record, maxparams, block_idx, "capacity", entry->capacity); - if (entry->physical) + if (entry->physical) { QEMU_ADD_BLOCK_PARAM_ULL(record, maxparams, block_idx, "physical", entry->physical); + } else { + if (virStorageSourceUpdateBlockPhysicalSize(src, false) == 0) { + QEMU_ADD_BLOCK_PARAM_ULL(record, maxparams, block_idx, + "physical", src->physical); + } + } ret = 0; cleanup: diff --git a/src/util/virstoragefile.c b/src/util/virstoragefile.c index 6c3017c..2aa1d90 100644 --- a/src/util/virstoragefile.c +++ b/src/util/virstoragefile.c @@ -2587,6 +2587,45 @@ virStorageSourceNewFromBacking(virStorageSourcePtr parent) } +/** + * @src: disk source definiton structure + * @report: report libvirt errors if set to true + * + * Updates src->physical for block devices since qemu doesn't report the current + * size correctly for them. Returns 0 on success, -1 on error. + */ +int +virStorageSourceUpdateBlockPhysicalSize(virStorageSourcePtr src, + bool report) +{ + int fd = -1; + off_t end; + int ret = -1; + + if (virStorageSourceGetActualType(src) != VIR_STORAGE_TYPE_BLOCK) + return 0; + + if ((fd = open(src->path, O_RDONLY)) < 0) { + if (report) + virReportSystemError(errno, _("failed to open block device '%s'"), + src->path); + return -1; + } + + if ((end = lseek(fd, 0, SEEK_END)) == (off_t) -1) { + if (report) + virReportSystemError(errno, + _("failed to seek to end of '%s'"), src->path); + } else { + src->physical = end; + ret = 0; + } + + VIR_FORCE_CLOSE(fd); + return ret; +} + + static char * virStorageFileCanonicalizeFormatPath(char **components, size_t ncomponents, diff --git a/src/util/virstoragefile.h b/src/util/virstoragefile.h index aa17a00..b98fe25 100644 --- a/src/util/virstoragefile.h +++ b/src/util/virstoragefile.h @@ -362,6 +362,8 @@ bool virStorageSourceIsLocalStorage(virStorageSourcePtr src); bool virStorageSourceIsEmpty(virStorageSourcePtr src); void virStorageSourceFree(virStorageSourcePtr def); void virStorageSourceBackingStoreClear(virStorageSourcePtr def); +int virStorageSourceUpdateBlockPhysicalSize(virStorageSourcePtr src, + bool report); virStorageSourcePtr virStorageSourceNewFromBacking(virStorageSourcePtr parent); virStorageSourcePtr virStorageSourceCopy(const virStorageSource *src, bool backingChain) -- 2.4.5

2 2

[libvirt] unsupported configuration: unknown video model 'virtio' - virtio-vga
by poma 07 Aug '15

07 Aug '15

$ qemu-system-x86_64 ... -device virtio-vga # lspci -d 1af4:1050 -knn 00:03.0 VGA compatible controller [0300]: Red Hat, Inc Device [1af4:1050] (rev 01) Subsystem: Red Hat, Inc Device [1af4:1100] Kernel driver in use: virtio-pci Kernel modules: virtio_pci # dmesg | grep virtio [ 1.727390] [drm] pci: virtio-vga detected [ 1.729315] [drm] virtio vbuffers: 80 bufs, 192B each, 15kB total. [ 2.023845] virtio_gpu virtio0: fb0: virtiodrmfb frame buffer device [ 2.023846] virtio_gpu virtio0: registered panic notifier [ 2.043135] [drm] Initialized virtio_gpu 0.0.1 0 on minor 0 # journalctl -b -u libvirtd.service -o cat Starting Virtualization daemon... Started Virtualization daemon. libvirt version: 1.2.17, package: 1.fc23 (Fedora Project, 2015-07-14-18:18:48, buildvm-03.phx2.fedoraproject.org) unsupported configuration: unknown video model 'virtio' Soonish?

3 3

[libvirt] Reproducible live disk detach bug
by Frank Schreuder 06 Aug '15

06 Aug '15

Hey guys, I am currently running into a reproducible libvirt bug when I detach a disk from a running VM using libvirt 1.2.18 and qemu 2.3.0. I can attach a disk to a running VM without any problems. During the attach operation I don't create a disk alias name, which results in libvirt generating an unique alias. Using virsh dumpxml I get the following output(snippet): <disk type='file' device='disk'> <driver name='qemu' type='raw' cache='none'/> <source file='/var/local/mnt/test-storage/test-storage.raw'/> <backingStore/> <target dev='vde' bus='virtio'/> <iotune> <read_bytes_sec>94371840</read_bytes_sec> <write_bytes_sec>31457280</write_bytes_sec> <read_iops_sec>1000</read_iops_sec> <write_iops_sec>200</write_iops_sec> </iotune> <alias name='virtio-disk4'/> <address type='pci' domain='0x0000' bus='0x00' slot='0x09' function='0x0'/> </disk> As you can see, libvirt generates 'virtio-disk4' as an alias. But as soon as I try to detach the disk from the running VM, I will receive the following two errors in my libvirt log: 2015-08-06 10:48:59.943+0000: 11657: warning : qemuMonitorJSONHandleDeviceDeleted:939 : missing device in device deleted event 2015-08-06 10:48:59.945+0000: 11662: error : qemuMonitorTextDriveDel:2594 : operation failed: deleting drive-virtio-disk4 drive failed: 2015-08-06T10:48:59.945058Z Device 'drive-virtio-disk4' not found Apparently libvirt tries to detach the disk with a drive- prefix, which doesn't exist? The drive seems to detach correctly though. I hope somebody with more extensive knowledge about libvirt can help me solve this issue. If more information is needed, feel free to contact me. Thanks, Frank

1 0

[libvirt] PING-2: [PATCH v5 0/4] qemu: Allow PCI virtio on ARM "virt" machine
by Pavel Fedin 06 Aug '15

06 Aug '15

Hello! Can anybody tell me what is wrong with this? Ignored for a long time. Kind regards, Pavel Fedin Expert Engineer Samsung Electronics Research center Russia > -----Original Message----- > From: libvir-list-bounces(a)redhat.com [mailto:libvir-list-bounces@redhat.com] On Behalf Of Pavel > Fedin > Sent: Friday, July 17, 2015 2:28 PM > To: libvir-list(a)redhat.com > Cc: Peter Krempa > Subject: [libvirt] [PATCH v5 0/4] qemu: Allow PCI virtio on ARM "virt" machine > > Virt machine in qemu since v2.3.0 has PCI generic host controller, and > can use PCI devices. This provides performance improvement as well as > vhost-net with irqfd support for virtio-net. However libvirt currently > does not allow ARM virt machine to have PCI devices. This patchset adds > the necessary support. > > Changes since v4: > - Rebased onto current master > - Added possibility to plug virtio-net-pci adapter directly into PCIe bus. > This is necessary for irqfds to work in qemu. > Changes since v3: > - Capability is based not on qemu version but on support of "gpex-pcihost" > device by qemu > - Added a workaround, allowing to pass "make check". The problem is that > test suite does not build capabilities cache. Unfortunately this means > that correct unit-test for the new functionality currently cannot be > written. Test suite framework needs to be improved. > Changes since v2: > Complete rework, use different approach > - Correctly model PCI Express bus on the machine. It is now possible to > explicitly specify <address-type='pci'> with attributes. This allows to > attach not only virtio, but any other PCI device to the model. > - Default is not changed and still mmio, for backwards compatibility with > existing installations. PCI bus has to be explicitly specified. > - Check for the capability in correct place, in v2 it actually did not > work > Changes since v1: > - Added capability based on qemu version number > - Recognize also "virt-" prefix > > Pavel Fedin (4): > qemu: Introduce QEMU_CAPS_OBJECT_GPEX > Add PCI-Express root to ARM virt machine > Build correct command line for PCI NICs on ARM > Allow to plug virtio-net-pci into PCIe slot > > src/qemu/qemu_capabilities.c | 10 ++++++++++ > src/qemu/qemu_capabilities.h | 1 + > src/qemu/qemu_command.c | 11 ++++++++++- > src/qemu/qemu_domain.c | 17 +++++++++++++---- > 4 files changed, 34 insertions(+), 5 deletions(-) > > -- > 1.9.5.msysgit.0 > > -- > libvir-list mailing list > libvir-list(a)redhat.com > https://www.redhat.com/mailman/listinfo/libvir-list

2 1

[libvirt] [PATCH] Allow vfio hotplug of a device to the domain which owns the iommu
by Shivaprasad G Bhat 06 Aug '15

06 Aug '15

The commit 7e72de4 didn't consider the hotplug scenarios. The patch addresses the hotplug case whereby if atleast one of the pci function is owned by a guest, the hotplug of other functions/devices in the same iommu group to the same guest goes through successfully. Signed-off-by: Shivaprasad G Bhat <sbhat(a)linux.vnet.ibm.com> --- src/util/virhostdev.c | 34 +++++++++++++++++++++++++++------- 1 file changed, 27 insertions(+), 7 deletions(-) diff --git a/src/util/virhostdev.c b/src/util/virhostdev.c index 809caed..529753c 100644 --- a/src/util/virhostdev.c +++ b/src/util/virhostdev.c @@ -54,23 +54,35 @@ static virClassPtr virHostdevManagerClass; static void virHostdevManagerDispose(void *obj); static virHostdevManagerPtr virHostdevManagerNew(void); +struct virHostdevIsPCINodeDeviceUsedData { + virHostdevManagerPtr hostdev_mgr; + const char *domainName; + const bool usesVfio; +}; + static int virHostdevIsPCINodeDeviceUsed(virPCIDeviceAddressPtr devAddr, void *opaque) { virPCIDevicePtr other; int ret = -1; virPCIDevicePtr pci = NULL; - virHostdevManagerPtr hostdev_mgr = opaque; + struct virHostdevIsPCINodeDeviceUsedData *helperData = opaque; if (!(pci = virPCIDeviceNew(devAddr->domain, devAddr->bus, devAddr->slot, devAddr->function))) goto cleanup; - other = virPCIDeviceListFind(hostdev_mgr->activePCIHostdevs, pci); + other = virPCIDeviceListFind(helperData->hostdev_mgr->activePCIHostdevs, + pci); if (other) { const char *other_drvname = NULL; const char *other_domname = NULL; virPCIDeviceGetUsedBy(other, &other_drvname, &other_domname); + if (helperData->usesVfio && + (other_domname && helperData->domainName) && + (STREQ(other_domname, helperData->domainName))) + goto iommu_owner; + if (other_drvname && other_domname) virReportError(VIR_ERR_OPERATION_INVALID, _("PCI device %s is in use by " @@ -83,6 +95,7 @@ static int virHostdevIsPCINodeDeviceUsed(virPCIDeviceAddressPtr devAddr, void *o virPCIDeviceGetName(pci)); goto cleanup; } + iommu_owner: ret = 0; cleanup: virPCIDeviceFree(pci); @@ -562,6 +575,9 @@ virHostdevPreparePCIDevices(virHostdevManagerPtr hostdev_mgr, for (i = 0; i < virPCIDeviceListCount(pcidevs); i++) { virPCIDevicePtr dev = virPCIDeviceListGet(pcidevs, i); bool strict_acs_check = !!(flags & VIR_HOSTDEV_STRICT_ACS_CHECK); + bool usesVfio = STREQ(virPCIDeviceGetStubDriver(dev), "vfio-pci"); + struct virHostdevIsPCINodeDeviceUsedData data = {hostdev_mgr, dom_name, + usesVfio}; if (!virPCIDeviceIsAssignable(dev, strict_acs_check)) { virReportError(VIR_ERR_OPERATION_INVALID, @@ -579,12 +595,12 @@ virHostdevPreparePCIDevices(virHostdevManagerPtr hostdev_mgr, * belonging to same iommu group can't be shared * across guests. */ - if (STREQ(virPCIDeviceGetStubDriver(dev), "vfio-pci")) { + if (usesVfio) { if (virPCIDeviceAddressIOMMUGroupIterate(devAddr, virHostdevIsPCINodeDeviceUsed, - hostdev_mgr) < 0) + &data) < 0) goto cleanup; - } else if (virHostdevIsPCINodeDeviceUsed(devAddr, hostdev_mgr)) { + } else if (virHostdevIsPCINodeDeviceUsed(devAddr, &data)) { goto cleanup; } } @@ -1544,6 +1560,8 @@ virHostdevPCINodeDeviceDetach(virHostdevManagerPtr hostdev_mgr, virPCIDevicePtr pci) { virPCIDeviceAddressPtr devAddr = NULL; + struct virHostdevIsPCINodeDeviceUsedData data = { hostdev_mgr, NULL, + false }; int ret = -1; virObjectLock(hostdev_mgr->activePCIHostdevs); @@ -1552,7 +1570,7 @@ virHostdevPCINodeDeviceDetach(virHostdevManagerPtr hostdev_mgr, if (!(devAddr = virPCIDeviceGetAddress(pci))) goto out; - if (virHostdevIsPCINodeDeviceUsed(devAddr, hostdev_mgr)) + if (virHostdevIsPCINodeDeviceUsed(devAddr, &data)) goto out; if (virPCIDeviceDetach(pci, hostdev_mgr->activePCIHostdevs, @@ -1573,6 +1591,8 @@ virHostdevPCINodeDeviceReAttach(virHostdevManagerPtr hostdev_mgr, virPCIDevicePtr pci) { virPCIDeviceAddressPtr devAddr = NULL; + struct virHostdevIsPCINodeDeviceUsedData data = {hostdev_mgr, NULL, + false}; int ret = -1; virObjectLock(hostdev_mgr->activePCIHostdevs); @@ -1581,7 +1601,7 @@ virHostdevPCINodeDeviceReAttach(virHostdevManagerPtr hostdev_mgr, if (!(devAddr = virPCIDeviceGetAddress(pci))) goto out; - if (virHostdevIsPCINodeDeviceUsed(devAddr, hostdev_mgr)) + if (virHostdevIsPCINodeDeviceUsed(devAddr, &data)) goto out; virPCIDeviceReattachInit(pci);

3 2

[libvirt] Reproducible live disk detach bug
by Frank Schreuder 06 Aug '15

06 Aug '15

Hey guys, I am currently running into a reproducible libvirt bug when I detach a disk from a running VM using libvirt 1.2.18 and qemu 2.3.0. I can attach a disk to a running VM without any problems. During the attach operation I don't create a disk alias name, which results in libvirt generating an unique alias. Using virsh dumpxml I get the following output(snippet): <disk type='file' device='disk'> <driver name='qemu' type='raw' cache='none'/> <source file='/var/local/mnt/test-storage/test-storage.raw'/> <backingStore/> <target dev='vde' bus='virtio'/> <iotune> <read_bytes_sec>94371840</read_bytes_sec> <write_bytes_sec>31457280</write_bytes_sec> <read_iops_sec>1000</read_iops_sec> <write_iops_sec>200</write_iops_sec> </iotune> <alias name='virtio-disk4'/> <address type='pci' domain='0x0000' bus='0x00' slot='0x09' function='0x0'/> </disk> As you can see, libvirt generates 'virtio-disk4' as an alias. But as soon as I try to detach the disk from the running VM, I will receive the following two errors in my libvirt log: 2015-08-06 10:48:59.943+0000: 11657: warning : qemuMonitorJSONHandleDeviceDeleted:939 : missing device in device deleted event 2015-08-06 10:48:59.945+0000: 11662: error : qemuMonitorTextDriveDel:2594 : operation failed: deleting drive-virtio-disk4 drive failed: 2015-08-06T10:48:59.945058Z Device 'drive-virtio-disk4' not found Apparently libvirt tries to detach the disk with a drive- prefix, which doesn't exist? The drive seems to detach correctly though. I hope somebody with more extensive knowledge about libvirt can help me solve this issue. If more information is needed, feel free to contact me. Thanks, Frank

1 0

[libvirt] [PATCH] qemuDomainDefPostParse: Adjust indent
by Michal Privoznik 06 Aug '15

06 Aug '15

While reviewing e8d551725 I've noticed a few unaligned lines. Fix this. Signed-off-by: Michal Privoznik <mprivozn(a)redhat.com> --- Pushed under trivial rule. src/qemu/qemu_domain.c | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) diff --git a/src/qemu/qemu_domain.c b/src/qemu/qemu_domain.c index 560929a..ad9a8ac 100644 --- a/src/qemu/qemu_domain.c +++ b/src/qemu/qemu_domain.c @@ -1024,10 +1024,10 @@ qemuDomainDefPostParse(virDomainDefPtr def, } if (STRPREFIX(def->os.machine, "pc-q35") || STREQ(def->os.machine, "q35")) { - addPCIeRoot = true; - addDefaultUSB = false; - addImplicitSATA = true; - break; + addPCIeRoot = true; + addDefaultUSB = false; + addImplicitSATA = true; + break; } if (!STRPREFIX(def->os.machine, "pc-0.") && !STRPREFIX(def->os.machine, "pc-1.") && -- 2.4.6

1 0

[libvirt] [PATCH] bootstrap: Don't require python-config
by Michal Privoznik 06 Aug '15

06 Aug '15

We've split the python bindings a long time ago. However, we are still requiring python-config (as an obfuscation to python-devel). This does not make any sense. The only thing we need is python, not python-devel. Signed-off-by: Michal Privoznik <mprivozn(a)redhat.com> --- bootstrap.conf | 13 +------------ 1 file changed, 1 insertion(+), 12 deletions(-) diff --git a/bootstrap.conf b/bootstrap.conf index c06ee4c..783b3a3 100644 --- a/bootstrap.conf +++ b/bootstrap.conf @@ -196,10 +196,7 @@ local_gl_dir=gnulib/local # Build prerequisites # Note that some of these programs are only required for 'make dist' to # succeed from a fresh git checkout; not all of these programs are -# required to run 'make dist' on a tarball. As a special case, we want -# to require the equivalent of the Fedora python-devel package, but -# RHEL 5 lacks the witness python-config package; we hack around that -# old environment below. +# required to run 'make dist' on a tarball. buildreq="\ autoconf 2.59 automake 1.9.6 @@ -212,19 +209,11 @@ patch - perl 5.5 perl::XML::XPath - pkg-config - -python-config - rpcgen - tar - xmllint - xsltproc - " -# Use rpm as a fallback to bypass the bootstrap probe for python-config, -# for the sake of RHEL 5; without requiring it on newer systems that -# have python-config to begin with. -if `(${PYTHON_CONFIG-python-config} --version; - test $? -lt 126 || rpm -q python-devel) >/dev/null 2>&1`; then - PYTHON_CONFIG=true -fi # Automake requires that ChangeLog and AUTHORS exist. touch AUTHORS ChangeLog || exit 1 -- 2.3.6

2 2