March 2019 - Devel - libvirt List Archives

[libvirt] lxc container stopped not each time
by mxs kolo 06 Mar '19

06 Mar '19

Hi all. I download ftp://libvirt.org/libvirt/libvirt-5.1.0.tar.xz and build rpms with lxc support. LXC containers can be started, but time to time can't be stopped. With option --mode initctl containers always stopped, but virsh reported error: "Container does not provide an initctl pipe" Inside container present /dev/initctl: # ls -la /dev/initctl lrwxrwxrwx 1 root root 25 Mar 5 18:40 /dev/initctl -> /run/systemd/initctl/fifo With option --mode signal, centos 7.5/7.6 systemd not stopped and rebooted, in containers log: systemd: Received SIGTERM. systemd: Reexecuting. systemd: systemd 219 running in system mode. (+PAM +AUDIT +SELINUX +IMA -APPARMOR +SMACK +SYSVINIT +UTMP +LIBCRYPTSETUP +GCRYPT +GNUTLS +ACL +XZ +LZ4 -SECCOMP +BLKID +ELFUTILS +KMOD +IDN) systemd: Detected virtualization lxc-libvirt. systemd: Detected architecture x86-64. Runlevel inside container before and after - is 5. Without option --mode container can stop, mostly when you in gdb step-by-step debug. But most often not, with simptoms as with --mode signal. Patch in attachment fix problem for my environment. Tested only with host centos 7.6 and guests centos 7.5/7.6 Short comments on patch. lxcDomainShutdownFlags() return rc 0 and container begin stopped. We not go to endjob label, but later we pass check: if (rc == 0 && (flags == 0 || (flags & VIR_DOMAIN_SHUTDOWN_SIGNAL))) { and trying next shutdown method with sigterm to pid1. If container heavy loaded (or waiting in gdb for input), it not stopped. But if container small, first method with initctl stop container before second method would be call b.r. Maxim Kozin

2 1

[libvirt] [PATCH 00/23] conf: domain: Buffer and XPath handling refactors
by Peter Krempa 06 Mar '19

06 Mar '19

Peter Krempa (23): util: xml: Enforce return value check from virXMLFormatElement conf: Use virXMLFormatElement in virDomainControllerDriverFormat conf: Use virXMLFormatElement in virDomainControllerDefFormat conf: Refactor virDomainMemballoonDefFormat conf: Refactor virDomainWatchdogDefFormat conf: Refactor virDomainPanicDefFormat conf: Refactor formatting of 'driver' in virDomainRNGDefFormat conf: Refactor virDomainInputDefFormat conf: Refactor virDomainHubDefFormat conf: Split out formatting of 'blkiotune' from virDomainDefFormatInternal conf: Split out domain features formatting from virDomainDefFormatInternal conf: Refactor control flow in virDomainDefFormatFeatures conf: Rename temp buffers in virDomainDefFormatFeatures conf: Remove impossible error in virDomainDefFormatFeatures conf: Simplify lifecycle of temp buffers in virDomainDefFormatFeatures conf: Avoid extra set of temp buffers in virDomainDefFormatFeatures conf: Avoid extra scope when formatting 'smm' feature conf: Avoid formatting empty <capabilities> element conf: Refactor formating of 'capabilities' features conf: Use virXMLFormatElement in virDomainDefFormatFeatures conf: domain: Use VIR_AUTOCLEAN(virBuffer) where appropriate conf: Move XPath context node in descendants of virSysinfoParseXML conf: domain: Use VIR_XPATH_NODE_AUTORESTORE where appropriate src/conf/domain_conf.c | 1184 +++++++---------- src/util/virxml.h | 3 +- .../lxcconf2xmldata/lxcconf2xml-blkiotune.xml | 3 +- .../lxcconf2xml-cpusettune.xml | 3 +- tests/lxcconf2xmldata/lxcconf2xml-cputune.xml | 3 +- tests/lxcconf2xmldata/lxcconf2xml-idmap.xml | 3 +- .../lxcconf2xml-macvlannetwork.xml | 3 +- tests/lxcconf2xmldata/lxcconf2xml-memtune.xml | 3 +- .../lxcconf2xml-miscnetwork.xml | 3 +- .../lxcconf2xml-nonenetwork.xml | 3 +- .../lxcconf2xmldata/lxcconf2xml-nonetwork.xml | 3 +- .../lxcconf2xml-physnetwork.xml | 3 +- .../lxcconf2xml-vlannetwork.xml | 3 +- 13 files changed, 485 insertions(+), 735 deletions(-) -- 2.20.1

2 46

[libvirt] [PATCH 0/6] Couple of storage driver improvements
by Michal Privoznik 06 Mar '19

06 Mar '19

*** BLURB HERE *** Michal Prívozník (6): storage_backend_iscsi_direct: Simplify vol zeroing virISCSIDirectReportLuns: Drop ClearVols storageVolWipePattern: Don't take shortcut to refreshPool() storage_driver: Introduce storagePoolRefreshImpl() storagePoolRefreshFailCleanup: Clear volumes on failed refresh virsh-pool: Offer only active pool for pool-refresh completer src/storage/storage_backend_gluster.c | 2 - src/storage/storage_backend_iscsi_direct.c | 21 ++---- src/storage/storage_backend_logical.c | 12 +--- src/storage/storage_backend_rbd.c | 4 +- src/storage/storage_driver.c | 77 ++++++++++++---------- src/storage/storage_util.c | 2 - tools/virsh-pool.c | 2 +- 7 files changed, 55 insertions(+), 65 deletions(-) -- 2.19.2

1 7

[libvirt] [PATCH] qemu: Don't set migration caps when changing postcopy bandwidth
by Jiri Denemark 06 Mar '19

06 Mar '19

The qemuMigrationParamsApply internal API was designed to apply all migration parameters and capabilities before we start to migrate a domain. The API can also be called outside migration job, e.g., via a call to qemuDomainMigrateSetMaxSpeed with VIR_DOMAIN_MIGRATE_MAX_SPEED_POSTCOPY flag, but in that case it should only apply migration parameters and ignore capabilities. Signed-off-by: Jiri Denemark <jdenemar(a)redhat.com> --- Notes: Best viewed with --ignore-all-space src/qemu/qemu_migration_params.c | 23 ++++++++++++++++------- 1 file changed, 16 insertions(+), 7 deletions(-) diff --git a/src/qemu/qemu_migration_params.c b/src/qemu/qemu_migration_params.c index 24e5368783..623193cf6c 100644 --- a/src/qemu/qemu_migration_params.c +++ b/src/qemu/qemu_migration_params.c @@ -778,14 +778,23 @@ qemuMigrationParamsApply(virQEMUDriverPtr driver, if (qemuDomainObjEnterMonitorAsync(driver, vm, asyncJob) < 0) return -1; - if (!(caps = qemuMigrationCapsToJSON(priv->migrationCaps, migParams->caps))) - goto cleanup; - - if (virJSONValueArraySize(caps) > 0) { - rc = qemuMonitorSetMigrationCapabilities(priv->mon, caps); - caps = NULL; - if (rc < 0) + if (asyncJob == QEMU_ASYNC_JOB_NONE) { + if (!virBitmapIsAllClear(migParams->caps)) { + virReportError(VIR_ERR_INTERNAL_ERROR, "%s", + _("Migration capabilities can only be set by " + "a migration job")); goto cleanup; + } + } else { + if (!(caps = qemuMigrationCapsToJSON(priv->migrationCaps, migParams->caps))) + goto cleanup; + + if (virJSONValueArraySize(caps) > 0) { + rc = qemuMonitorSetMigrationCapabilities(priv->mon, caps); + caps = NULL; + if (rc < 0) + goto cleanup; + } } /* If QEMU is too old to support xbzrle-cache-size migration parameter, -- 2.21.0

3 5

[libvirt] [PATCH] rpm: add dep on xfsprogs-devel for reflink support
by Daniel P. Berrangé 06 Mar '19

06 Mar '19

Support for XFS reflink clone was added in: commit 8ed874b39b3c330bbcdff434e08995dbb4467285 Author: Julio Faracco <jcfaracco(a)gmail.com> Date: Fri Jul 6 10:43:01 2018 -0300 storage: Rename btrfsCloneFile to support other filesystems. commit 2e11298f938672c5430193277c98b9474a68b1fb Author: Julio Faracco <jcfaracco(a)gmail.com> Date: Fri Jul 6 10:43:00 2018 -0300 configure: Adding XFS library/headers check. But these patches missed that the xfs/xfs.h header is not installed unless you have xfsprogs-devel present. Signed-off-by: Daniel P. Berrangé <berrange(a)redhat.com> --- libvirt.spec.in | 2 ++ 1 file changed, 2 insertions(+) diff --git a/libvirt.spec.in b/libvirt.spec.in index 9beffba203..83b43f4ef6 100644 --- a/libvirt.spec.in +++ b/libvirt.spec.in @@ -326,6 +326,8 @@ BuildRequires: libiscsi-devel BuildRequires: parted-devel # For Multipath support BuildRequires: device-mapper-devel +# For XFS reflink clone support +BuildRequires: xfsprogs-devel %if %{with_storage_rbd} BuildRequires: librados2-devel BuildRequires: librbd1-devel -- 2.20.1

2 1

[libvirt] [PATCH 1/2] security: aa-helper: nvidia rules for gl devices
by Christian Ehrhardt 06 Mar '19

06 Mar '19

Further testing with different devices showed that we need more rules to drive gl backends with nvidia cards. Related denies look like: apparmor="DENIED" operation="open" name="/usr/share/egl/egl_external_platform.d/" requested_mask="r" apparmor="DENIED" operation="open" name="/proc/modules" requested_mask="r" apparmor="DENIED" operation="open" name="/proc/driver/nvidia/params" requested_mask="r" apparmor="DENIED" operation="mknod" name="/dev/nvidiactl" requested_mask="c" Fixes: https://bugs.launchpad.net/ubuntu/+source/libvirt/+bug/1817943 Signed-off-by: Christian Ehrhardt <christian.ehrhardt(a)canonical.com> --- src/security/virt-aa-helper.c | 5 +++++ 1 file changed, 5 insertions(+) diff --git a/src/security/virt-aa-helper.c b/src/security/virt-aa-helper.c index e9120213ff..13b507ff69 100644 --- a/src/security/virt-aa-helper.c +++ b/src/security/virt-aa-helper.c @@ -1279,6 +1279,11 @@ get_files(vahControl * ctl) virBufferAddLit(&buf, " \"/usr/share/drirc.d/{,*.conf}\" r,\n"); virBufferAddLit(&buf, " \"/etc/glvnd/egl_vendor.d/{,*}\" r,\n"); virBufferAddLit(&buf, " \"/usr/share/glvnd/egl_vendor.d/{,*}\" r,\n"); + virBufferAddLit(&buf, " \"/usr/share/egl/egl_external_platform.d/\" r,\n"); + virBufferAddLit(&buf, " \"/usr/share/egl/egl_external_platform.d/*\" r,\n"); + virBufferAddLit(&buf, " \"/proc/modules\" r,\n"); + virBufferAddLit(&buf, " \"/proc/driver/nvidia/params\" r,\n"); + virBufferAddLit(&buf, " \"/dev/nvidiactl\" rw,\n"); virBufferAddLit(&buf, " # Probe DRI device attributes\n"); virBufferAddLit(&buf, " \"/dev/dri/\" r,\n"); virBufferAddLit(&buf, " \"/sys/devices/*/*/{uevent,vendor,device,subsystem_vendor,subsystem_device}\" r,\n"); -- 2.17.1

2 5

[libvirt] [PATCH for-4.0 v4 0/2] virtio: Provide version-specific variants of virtio PCI devices
by Eduardo Habkost 06 Mar '19

06 Mar '19

Existing modern-only device types are not being touched by v3, as they don't need separate variants. However, I plan to implement separate cleanups in the code that calls virtio_pci_force_virtio_1(), first, and then propose additional changes (e.g. deprecating disable-legacy and disable-modern in those device types). Changes v3 -> v4: * Trivial comment fixes (Cornelia Huck) * Test code update (Caio Carrara) Changes v2 -> v3: * Split into two separate patches (type registration helper and introduction of new types) * Rewrote virtio_pci_types_register() completely: * Replaced magic generation of type names with explicit fields in VirtioPCIDeviceTypeInfo * Removed modern_only field (not necessary anymore) * Don't register a separate base type unless necessary Changes v1 -> v2: * Removed *-0.9 devices. Nobody will want to use them, if transitional devices work with legacy drivers (Gerd Hoffmann, Michael S. Tsirkin) * Drop virtio version from name: rename -1.0-transitional to -transitional (Michael S. Tsirkin) * Renamed -1.0 to -non-transitional * Don't add any extra variants to modern-only device types (they don't need it) * Fix typo on TYPE_VIRTIO_INPUT_HOST_PCI (crash reported by Cornelia Huck) * No need to change cast macros for modern-only devices * Rename virtio_register_types() to virtio_pci_types_register() Original patch description: Many of the current virtio-*-pci device types actually represent 3 different types of devices: * virtio 1.0 non-transitional devices * virtio 1.0 transitional devices * virtio 0.9 ("legacy device" in virtio 1.0 terminology) That would be just an annoyance if it didn't break our device/bus compatibility QMP interfaces. With this multi-purpose device type, there's no way to tell management software that transitional devices and legacy devices require a Conventional PCI bus. The multi-purpose device types would also prevent us from telling management software what's the PCI vendor/device ID for them, because their PCI IDs change at runtime depending on the bus where they were plugged. This patch adds separate device types for each of those virtio device flavors: * virtio-*-pci: the existing multi-purpose device types * virtio-*-pci-transitional: virtio-1.0 device supporting legacy drivers * virtio-*-pci-non-transitional: modern-only Reference to previous discussion that originated this idea: https://www.mail-archive.com/qemu-devel@nongnu.org/msg558389.html Eduardo Habkost (2): virtio: Helper for registering virtio device types virtio: Provide version-specific variants of virtio PCI devices hw/virtio/virtio-pci.h | 78 +++++++-- hw/display/virtio-gpu-pci.c | 7 +- hw/display/virtio-vga.c | 7 +- hw/virtio/virtio-crypto-pci.c | 7 +- hw/virtio/virtio-pci.c | 267 ++++++++++++++++++++++------- tests/acceptance/virtio_version.py | 176 +++++++++++++++++++ 6 files changed, 452 insertions(+), 90 deletions(-) create mode 100644 tests/acceptance/virtio_version.py -- 2.18.0.rc1.1.g3f1ff2140

9 13

[libvirt] [PATCH] libxl: Add support for max_grant_frame setting
by Jim Fehlig 06 Mar '19

06 Mar '19

Xen 4.10 introduced the max_grant_frames xl config setting, which can be set globally in xl.conf(5) or per-domain in xl.cfg(5). max_grant_frames specifies the maximum number of grant frames the domain is allowed to have, which in turn controls the number of pages the domain can share. This patch adds support for setting max_grant_frames on a global level in libxl.conf. Per-domain support via domXML can be provided in a future patch. Signed-off-by: Jim Fehlig <jfehlig(a)suse.com> --- src/libxl/libvirtd_libxl.aug | 2 ++ src/libxl/libxl.conf | 8 ++++++++ src/libxl/libxl_conf.c | 9 +++++++++ src/libxl/libxl_conf.h | 2 ++ src/libxl/test_libvirtd_libxl.aug.in | 1 + 5 files changed, 22 insertions(+) diff --git a/src/libxl/libvirtd_libxl.aug b/src/libxl/libvirtd_libxl.aug index 58b9af3707..1c3e2719fc 100644 --- a/src/libxl/libvirtd_libxl.aug +++ b/src/libxl/libvirtd_libxl.aug @@ -29,6 +29,7 @@ module Libvirtd_libxl = let keepalive_interval_entry = int_entry "keepalive_interval" let keepalive_count_entry = int_entry "keepalive_count" let nested_hvm_entry = bool_entry "nested_hvm" + let max_grant_frames_entry = int_entry "max_grant_frames" (* Each entry in the config is one of the following ... *) let entry = autoballoon_entry @@ -36,6 +37,7 @@ module Libvirtd_libxl = | keepalive_interval_entry | keepalive_count_entry | nested_hvm_entry + | max_grant_frames_entry let comment = [ label "#comment" . del /#[ \t]*/ "# " . store /([^ \t\n][^\n]*)?/ . del /\n/ "\n" ] let empty = [ label "#empty" . eol ] diff --git a/src/libxl/libxl.conf b/src/libxl/libxl.conf index 72825a71c5..12d7abeda6 100644 --- a/src/libxl/libxl.conf +++ b/src/libxl/libxl.conf @@ -49,3 +49,11 @@ # element. # By default it is disabled. #nested_hvm = 0 + +# Xen domains can grant other domains access to its memory pages, which is +# most commonly used for the operation of paravirtualized devices. +# max_grant_frames was introduced in Xen 4.10 and can be used to specify a +# default value for the maximum number of grant frames each domain started +# by libvirt is allowed to have. The default value is 32. +# +#max_grant_frames = 32 diff --git a/src/libxl/libxl_conf.c b/src/libxl/libxl_conf.c index c769050ff1..f57ffe1a01 100644 --- a/src/libxl/libxl_conf.c +++ b/src/libxl/libxl_conf.c @@ -393,6 +393,12 @@ libxlMakeDomBuildInfo(virDomainDefPtr def, def->mem.cur_balloon = VIR_ROUND_UP(def->mem.cur_balloon, 1024); b_info->max_memkb = virDomainDefGetMemoryInitial(def); b_info->target_memkb = def->mem.cur_balloon; + +#ifdef LIBXL_HAVE_BUILDINFO_GRANT_LIMITS + if (cfg->max_grant_frames > 0) + b_info->max_grant_frames = cfg->max_grant_frames; +#endif + if (hvm || pvh) { if (caps && def->cpu && def->cpu->mode == (VIR_CPU_MODE_HOST_PASSTHROUGH)) { @@ -1888,6 +1894,9 @@ int libxlDriverConfigLoadFile(libxlDriverConfigPtr cfg, if (virConfGetValueBool(conf, "nested_hvm", &cfg->nested_hvm) < 0) goto cleanup; + if (virConfGetValueUInt(conf, "max_grant_frames", &cfg->max_grant_frames) < 0) + goto cleanup; + ret = 0; cleanup: diff --git a/src/libxl/libxl_conf.h b/src/libxl/libxl_conf.h index fee94241af..985524f3eb 100644 --- a/src/libxl/libxl_conf.h +++ b/src/libxl/libxl_conf.h @@ -86,6 +86,8 @@ struct _libxlDriverConfig { bool nested_hvm; + unsigned int max_grant_frames; + /* Once created, caps are immutable */ virCapsPtr caps; diff --git a/src/libxl/test_libvirtd_libxl.aug.in b/src/libxl/test_libvirtd_libxl.aug.in index 372a43f94a..754ddc10f6 100644 --- a/src/libxl/test_libvirtd_libxl.aug.in +++ b/src/libxl/test_libvirtd_libxl.aug.in @@ -7,3 +7,4 @@ module Test_libvirtd_libxl = { "keepalive_interval" = "5" } { "keepalive_count" = "5" } { "nested_hvm" = "0" } +{ "max_grant_frames" = "32" } -- 2.20.1

2 3

[libvirt] domXML modeling question
by Jim Fehlig 06 Mar '19

06 Mar '19

There have been a few requests [1][2] to support Xen's max_grant_frames setting in libvirt domXML, but I'm not quite sure how to model it. The documentation [3] on this setting states: Specify the maximum number of grant frames the domain is allowed to have. This value controls how many pages the domain is able to grant access to for other domains, needed e.g. for the operation of paravirtualized devices. The default is settable via xl.conf(5). It smells of a <memtune> setting, e.g. the amount of memory a domain can share, but doesn't map to any of the existing settings. A new subelement <shared_limit> doesn't feel right. Does anyone suggest a better way of modeling max_grant_frames? Another option I considered is setting the value based on number of PV devices, but I think that flies in the face of libvirt's policy of not dictating policy. Regardless of domain config modeling I can work on a driver-wide setting in libxl.conf, similar to Xen's xl.conf(5) global. Regards, Jim [1] https://www.redhat.com/archives/libvir-list/2018-April/msg00216.html [2] https://www.redhat.com/archives/libvirt-users/2019-January/msg00011.html [3] http://xenbits.xen.org/gitweb/?p=xen.git;a=blob;f=docs/man/xl.cfg.5.pod.in;…

3 3

[libvirt] [PATCH 0/4] Couple of patches about virsh completion
by Lin Ma 06 Mar '19

06 Mar '19

Lin Ma (4): virsh: Only return active domain names for detach-device-alias virsh: Add device name completion for target option of detach-disk command virsh-network: Introduce virshNetworkEventCallback to handle network events virsh: Add event name completion to 'network-event' command tools/virsh-completer.c | 27 +++++++++++++++++++++++++++ tools/virsh-completer.h | 4 ++++ tools/virsh-domain.c | 3 ++- tools/virsh-network.c | 18 +++++++++++++++--- tools/virsh-network.h | 8 ++++++++ 5 files changed, 56 insertions(+), 4 deletions(-) -- 2.19.0

2 10