December 2019 - Devel - libvirt List Archives

[libvirt] [PATCH] tests: securityselinuxlabel: Add QEMU_CAPS_VNC to fake qemuCaps
by Peter Krempa 17 Dec '19

17 Dec '19

In commit 45270337f057f26ce484f6e forgot to make sure that tests pass. Add the missing capability to fix the test. Signed-off-by: Peter Krempa <pkrempa(a)redhat.com> --- Pushed. tests/securityselinuxlabeltest.c | 1 + 1 file changed, 1 insertion(+) diff --git a/tests/securityselinuxlabeltest.c b/tests/securityselinuxlabeltest.c index 192f2dc84f..3040a36693 100644 --- a/tests/securityselinuxlabeltest.c +++ b/tests/securityselinuxlabeltest.c @@ -351,6 +351,7 @@ mymain(void) return EXIT_FAILURE; virQEMUCapsSet(qemuCaps, QEMU_CAPS_DEVICE_CIRRUS_VGA); + virQEMUCapsSet(qemuCaps, QEMU_CAPS_VNC); if (qemuTestCapsCacheInsert(driver.qemuCapsCache, qemuCaps) < 0) return EXIT_FAILURE; -- 2.23.0

2 1

[libvirt] [PATCH 0/9] Cleanup virConnectPtr usage
by Michal Privoznik 17 Dec '19

17 Dec '19

I've noticed this problem when reviewing a patch on the list [1]. Long story short, dom->conn is not guaranteed to have all driver pointers set (consider split daemons). I haven't identified any other violation than what I'm fixing here. But something might have slipped through my git grep. 1: https://www.redhat.com/archives/libvir-list/2019-December/msg00120.html Michal Prívozník (9): qemu_driver: Push qemuDomainInterfaceAddresses() a few lines down qemu: Don't use dom->conn to lookup virNetwork qemuGetDHCPInterfaces: Move some variables inside the loop qemuGetDHCPInterfaces: Switch to GLib libxl: Don't use dom->conn to lookup virNetwork libxlGetDHCPInterfaces: Move some variables inside the loop libxlGetDHCPInterfaces: Switch to GLib lxc: Cleanup virConnectPtr usage get_nonnull_domain: Drop useless comment src/libxl/libxl_driver.c | 71 ++++------ src/lxc/lxc_driver.c | 12 +- src/lxc/lxc_process.c | 40 +++--- src/lxc/lxc_process.h | 2 +- src/qemu/qemu_driver.c | 193 ++++++++++++---------------- src/remote/remote_daemon_dispatch.c | 3 - 6 files changed, 139 insertions(+), 182 deletions(-) -- 2.23.0

2 10

[libvirt] [PATCH] cpu_map/x86: Add support for BFLOAT16 data type
by Jiri Denemark 17 Dec '19

17 Dec '19

Introduced in QEMU by commit v4.1.0-266-g80db491da4. Signed-off-by: Jiri Denemark <jdenemar(a)redhat.com> --- src/cpu_map/x86_features.xml | 4 ++++ 1 file changed, 4 insertions(+) diff --git a/src/cpu_map/x86_features.xml b/src/cpu_map/x86_features.xml index b42b490160..eefb55067e 100644 --- a/src/cpu_map/x86_features.xml +++ b/src/cpu_map/x86_features.xml @@ -339,6 +339,10 @@ <cpuid eax_in='0x07' ecx_in='0x00' edx='0x80000000'/> </feature> + <feature name='avx512-bf16'> + <cpuid eax_in='0x07' ecx_in='0x01' eax='0x00000020'/> + </feature> +  <feature name='xsaveopt'> <cpuid eax_in='0x0d' ecx_in='0x01' eax='0x00000001'/> -- 2.24.1

2 1

[libvirt] [PATCH] virsh migrate: Require --tls for --tls-destination
by Jiri Denemark 17 Dec '19

17 Dec '19

--tls-destination would be just ignored unless --tls is not specified, which is correct, but let's provide a bit of a guidance is a user forgets to add --tls. https://bugzilla.redhat.com/show_bug.cgi?id=1784345 Signed-off-by: Jiri Denemark <jdenemar(a)redhat.com> --- tools/virsh-domain.c | 1 + 1 file changed, 1 insertion(+) diff --git a/tools/virsh-domain.c b/tools/virsh-domain.c index 56137bdd74..e7e92ee60d 100644 --- a/tools/virsh-domain.c +++ b/tools/virsh-domain.c @@ -10966,6 +10966,7 @@ cmdMigrate(vshControl *ctl, const vshCmd *cmd) VSH_REQUIRE_OPTION("postcopy-after-precopy", "postcopy"); VSH_REQUIRE_OPTION("timeout-postcopy", "postcopy"); VSH_REQUIRE_OPTION("persistent-xml", "persistent"); + VSH_REQUIRE_OPTION("tls-destination", "tls"); if (!(dom = virshCommandOptDomain(ctl, cmd, NULL))) return false; -- 2.24.1

2 3

[libvirt] [glib PATCH 0/2] po: change over to use Weblate instead of Zanata
by Daniel P. Berrangé 17 Dec '19

17 Dec '19

Zanata is dead upstream, and Fedora is replacing it with Weblate https://fedoraproject.org/wiki/L10N_Move_to_Weblate Libvirt has outsourced its translations to the Fedora translation team, so we must follow this move. Daniel P. Berrangé (2): po: delete empty translation files po: change update rules to use weblate instead of zanata po/Makefile.am | 32 +++++++++++++++++++------------- po/af.mini.po | 20 -------------------- po/am.mini.po | 20 -------------------- po/anp.mini.po | 19 ------------------- po/ar.mini.po | 21 --------------------- po/as.mini.po | 20 -------------------- po/ast.mini.po | 20 -------------------- po/bal.mini.po | 20 -------------------- po/be.mini.po | 21 --------------------- po/bg.mini.po | 20 -------------------- po/bn.mini.po | 20 -------------------- po/bn_IN.mini.po | 20 -------------------- po/bo.mini.po | 20 -------------------- po/br.mini.po | 20 -------------------- po/brx.mini.po | 20 -------------------- po/bs.mini.po | 21 --------------------- po/cy.mini.po | 21 --------------------- po/da.mini.po | 20 -------------------- po/de.mini.po | 20 -------------------- po/de_CH.mini.po | 20 -------------------- po/el.mini.po | 20 -------------------- po/eo.mini.po | 20 -------------------- po/et.mini.po | 20 -------------------- po/eu.mini.po | 20 -------------------- po/fa.mini.po | 20 -------------------- po/gl.mini.po | 20 -------------------- po/gu.mini.po | 20 -------------------- po/he.mini.po | 20 -------------------- po/hr.mini.po | 21 --------------------- po/hu.mini.po | 20 -------------------- po/ia.mini.po | 20 -------------------- po/id.mini.po | 20 -------------------- po/ilo.mini.po | 20 -------------------- po/is.mini.po | 20 -------------------- po/it.mini.po | 20 -------------------- po/ka.mini.po | 20 -------------------- po/kk.mini.po | 20 -------------------- po/km.mini.po | 20 -------------------- po/kn.mini.po | 20 -------------------- po/ko.mini.po | 20 -------------------- po/kw.mini.po | 20 -------------------- po/kw(a)kkcor.mini.po | 20 -------------------- po/kw(a)uccor.mini.po | 20 -------------------- po/kw_GB.mini.po | 20 -------------------- po/ky.mini.po | 20 -------------------- po/lt.mini.po | 21 --------------------- po/lv.mini.po | 21 --------------------- po/mai.mini.po | 20 -------------------- po/mk.mini.po | 20 -------------------- po/ml.mini.po | 20 -------------------- po/mn.mini.po | 20 -------------------- po/mr.mini.po | 20 -------------------- po/ms.mini.po | 20 -------------------- po/nb.mini.po | 20 -------------------- po/nds.mini.po | 20 -------------------- po/ne.mini.po | 20 -------------------- po/nl.mini.po | 20 -------------------- po/nn.mini.po | 20 -------------------- po/nso.mini.po | 20 -------------------- po/or.mini.po | 20 -------------------- po/pa.mini.po | 20 -------------------- po/pt.mini.po | 20 -------------------- po/ro.mini.po | 21 --------------------- po/si.mini.po | 20 -------------------- po/sk.mini.po | 20 -------------------- po/sl.mini.po | 21 --------------------- po/sq.mini.po | 20 -------------------- po/sr.mini.po | 21 --------------------- po/sr(a)latin.mini.po | 21 --------------------- po/sv.mini.po | 20 -------------------- po/ta.mini.po | 20 -------------------- po/te.mini.po | 20 -------------------- po/tg.mini.po | 20 -------------------- po/th.mini.po | 20 -------------------- po/tr.mini.po | 20 -------------------- po/tw.mini.po | 19 ------------------- po/ur.mini.po | 20 -------------------- po/vi.mini.po | 20 -------------------- po/wba.mini.po | 19 ------------------- po/yo.mini.po | 19 ------------------- po/zanata.xml | 7 ------- po/zh_CN.mini.po | 20 -------------------- po/zh_HK.mini.po | 20 -------------------- po/zh_TW.mini.po | 20 -------------------- po/zu.mini.po | 20 -------------------- 85 files changed, 19 insertions(+), 1687 deletions(-) delete mode 100644 po/af.mini.po delete mode 100644 po/am.mini.po delete mode 100644 po/anp.mini.po delete mode 100644 po/ar.mini.po delete mode 100644 po/as.mini.po delete mode 100644 po/ast.mini.po delete mode 100644 po/bal.mini.po delete mode 100644 po/be.mini.po delete mode 100644 po/bg.mini.po delete mode 100644 po/bn.mini.po delete mode 100644 po/bn_IN.mini.po delete mode 100644 po/bo.mini.po delete mode 100644 po/br.mini.po delete mode 100644 po/brx.mini.po delete mode 100644 po/bs.mini.po delete mode 100644 po/cy.mini.po delete mode 100644 po/da.mini.po delete mode 100644 po/de.mini.po delete mode 100644 po/de_CH.mini.po delete mode 100644 po/el.mini.po delete mode 100644 po/eo.mini.po delete mode 100644 po/et.mini.po delete mode 100644 po/eu.mini.po delete mode 100644 po/fa.mini.po delete mode 100644 po/gl.mini.po delete mode 100644 po/gu.mini.po delete mode 100644 po/he.mini.po delete mode 100644 po/hr.mini.po delete mode 100644 po/hu.mini.po delete mode 100644 po/ia.mini.po delete mode 100644 po/id.mini.po delete mode 100644 po/ilo.mini.po delete mode 100644 po/is.mini.po delete mode 100644 po/it.mini.po delete mode 100644 po/ka.mini.po delete mode 100644 po/kk.mini.po delete mode 100644 po/km.mini.po delete mode 100644 po/kn.mini.po delete mode 100644 po/ko.mini.po delete mode 100644 po/kw.mini.po delete mode 100644 po/kw(a)kkcor.mini.po delete mode 100644 po/kw(a)uccor.mini.po delete mode 100644 po/kw_GB.mini.po delete mode 100644 po/ky.mini.po delete mode 100644 po/lt.mini.po delete mode 100644 po/lv.mini.po delete mode 100644 po/mai.mini.po delete mode 100644 po/mk.mini.po delete mode 100644 po/ml.mini.po delete mode 100644 po/mn.mini.po delete mode 100644 po/mr.mini.po delete mode 100644 po/ms.mini.po delete mode 100644 po/nb.mini.po delete mode 100644 po/nds.mini.po delete mode 100644 po/ne.mini.po delete mode 100644 po/nl.mini.po delete mode 100644 po/nn.mini.po delete mode 100644 po/nso.mini.po delete mode 100644 po/or.mini.po delete mode 100644 po/pa.mini.po delete mode 100644 po/pt.mini.po delete mode 100644 po/ro.mini.po delete mode 100644 po/si.mini.po delete mode 100644 po/sk.mini.po delete mode 100644 po/sl.mini.po delete mode 100644 po/sq.mini.po delete mode 100644 po/sr.mini.po delete mode 100644 po/sr(a)latin.mini.po delete mode 100644 po/sv.mini.po delete mode 100644 po/ta.mini.po delete mode 100644 po/te.mini.po delete mode 100644 po/tg.mini.po delete mode 100644 po/th.mini.po delete mode 100644 po/tr.mini.po delete mode 100644 po/tw.mini.po delete mode 100644 po/ur.mini.po delete mode 100644 po/vi.mini.po delete mode 100644 po/wba.mini.po delete mode 100644 po/yo.mini.po delete mode 100644 po/zanata.xml delete mode 100644 po/zh_CN.mini.po delete mode 100644 po/zh_HK.mini.po delete mode 100644 po/zh_TW.mini.po delete mode 100644 po/zu.mini.po -- 2.23.0

2 4

[libvirt] [PATCH v3 00/30] Introduce NVMe support
by Michal Privoznik 17 Dec '19

17 Dec '19

v3 of: https://www.redhat.com/archives/libvir-list/2019-September/msg01209.html Builds just fine ;-) https://travis-ci.org/zippy2/libvirt/builds/619622349 diff to v2: - Fixed bug reported by Cole - unability to plug regular <hostdev/>-s caused by v2. Michal Prívozník (30): qemu: Explicitly add/remove /dev/vfio/vfio to/from NS/CGroups qemuDomainGetHostdevPath: Use more g_autoptr()/g_autofree qemuDomainGetHostdevPath: Don't include /dev/vfio/vfio in returned paths qemu: Drop some 'cleanup' labels virpci: Introduce and use virPCIDeviceAddressGetIOMMUGroupDev virHostdevPreparePCIDevices: Separate out function body virHostdevReAttachPCIDevices: Separate out function body virpci: Introduce virPCIDeviceAddressCopy qemuMigrationSrcIsSafe: Rework slightly schemas: Introduce disk type NVMe conf: Format and parse NVMe type disk virstoragefile: Introduce virStorageSourceChainHasNVMe domain_conf: Introduce virDomainDefHasNVMeDisk util: Introduce virNVMeDevice module virhostdev: Include virNVMeDevice module virpcimock: Introduce NVMe driver and devices virhostdevtest: Test virNVMeDevice assignment qemu: prepare NVMe devices too qemu: Take NVMe disks into account when calculating memlock limit qemu: Create NVMe disk in domain namespace qemu: Mark NVMe disks as 'need VFIO' qemu: Allow NVMe disk in CGroups virSecuritySELinuxRestoreImageLabelInt: Don't skip non-local storage qemu_capabilities: Introduce QEMU_CAPS_DRIVE_NVME qemu: Generate command line of NVMe disks qemu_monitor_text: Catch IOMMU/VFIO related errors in qemuMonitorTextAddDrive qemu: Don't leak storage perms on failure in qemuDomainAttachDiskGeneric qemu: Allow forcing VFIO when computing memlock limit qemu_hotplug: Prepare NVMe disks on hotplug virsh: Introduce nvme disk to domblklist docs/formatdomain.html.in | 57 ++- docs/schemas/domaincommon.rng | 32 ++ po/POTFILES.in | 1 + src/conf/domain_conf.c | 109 +++++ src/conf/domain_conf.h | 3 + src/libvirt_private.syms | 28 ++ src/libxl/xen_xl.c | 1 + src/qemu/qemu_block.c | 25 + src/qemu/qemu_capabilities.c | 2 + src/qemu/qemu_capabilities.h | 1 + src/qemu/qemu_cgroup.c | 203 +++++--- src/qemu/qemu_command.c | 6 +- src/qemu/qemu_domain.c | 375 +++++++++------ src/qemu/qemu_domain.h | 15 +- src/qemu/qemu_driver.c | 4 + src/qemu/qemu_hostdev.c | 71 ++- src/qemu/qemu_hostdev.h | 16 + src/qemu/qemu_hotplug.c | 31 +- src/qemu/qemu_migration.c | 30 +- src/qemu/qemu_monitor_text.c | 7 + src/qemu/qemu_process.c | 7 + src/security/security_apparmor.c | 25 +- src/security/security_dac.c | 30 ++ src/security/security_selinux.c | 51 +- src/util/Makefile.inc.am | 2 + src/util/virhostdev.c | 431 +++++++++++++++-- src/util/virhostdev.h | 37 ++ src/util/virnvme.c | 447 ++++++++++++++++++ src/util/virnvme.h | 97 ++++ src/util/virpci.c | 29 ++ src/util/virpci.h | 5 + src/util/virstoragefile.c | 71 +++ src/util/virstoragefile.h | 19 + .../caps_2.12.0.aarch64.xml | 1 + .../caps_2.12.0.ppc64.xml | 1 + .../caps_2.12.0.s390x.xml | 1 + .../caps_2.12.0.x86_64.xml | 1 + .../qemucapabilitiesdata/caps_3.0.0.ppc64.xml | 1 + .../caps_3.0.0.riscv32.xml | 1 + .../caps_3.0.0.riscv64.xml | 1 + .../qemucapabilitiesdata/caps_3.0.0.s390x.xml | 1 + .../caps_3.0.0.x86_64.xml | 1 + .../qemucapabilitiesdata/caps_3.1.0.ppc64.xml | 1 + .../caps_3.1.0.x86_64.xml | 1 + .../caps_4.0.0.aarch64.xml | 1 + .../qemucapabilitiesdata/caps_4.0.0.ppc64.xml | 1 + .../caps_4.0.0.riscv32.xml | 1 + .../caps_4.0.0.riscv64.xml | 1 + .../qemucapabilitiesdata/caps_4.0.0.s390x.xml | 1 + .../caps_4.0.0.x86_64.xml | 1 + .../caps_4.1.0.x86_64.xml | 1 + .../caps_4.2.0.aarch64.xml | 1 + .../qemucapabilitiesdata/caps_4.2.0.ppc64.xml | 1 + .../qemucapabilitiesdata/caps_4.2.0.s390x.xml | 1 + .../caps_4.2.0.x86_64.xml | 1 + tests/qemumemlocktest.c | 2 +- .../disk-nvme.x86_64-latest.args | 63 +++ tests/qemuxml2argvdata/disk-nvme.xml | 63 +++ tests/qemuxml2argvtest.c | 1 + tests/qemuxml2xmloutdata/disk-nvme.xml | 1 + tests/qemuxml2xmltest.c | 1 + tests/virhostdevtest.c | 97 ++++ tests/virpcimock.c | 3 + tests/virpcitestdata/0000-01-00.0.config | Bin 0 -> 4096 bytes tests/virpcitestdata/0000-02-00.0.config | Bin 0 -> 4096 bytes tools/virsh-domain-monitor.c | 31 +- 66 files changed, 2240 insertions(+), 312 deletions(-) create mode 100644 src/util/virnvme.c create mode 100644 src/util/virnvme.h create mode 100644 tests/qemuxml2argvdata/disk-nvme.x86_64-latest.args create mode 100644 tests/qemuxml2argvdata/disk-nvme.xml create mode 120000 tests/qemuxml2xmloutdata/disk-nvme.xml create mode 100644 tests/virpcitestdata/0000-01-00.0.config create mode 100644 tests/virpcitestdata/0000-02-00.0.config -- 2.23.0

3 67

[libvirt] [PATCH 0/2] build: -Wframe-larger-than shenanigans
by Ján Tomko 17 Dec '19

17 Dec '19

The recent cputest addition broke the build with clang for me, since clang calculates the stack size differently than gcc. Put the strict limit into all warning by default (this starts enforcing it for examples, for example) and raise the limit by a few orders, which might appear Sisyphean but I believe it's needed for putting two conflicting limits on the compiler command line. Ján Tomko (2): build: warn on a large frame by default build: relax the relaxed stack frame limit further m4/virt-compile-warnings.m4 | 4 ++-- src/Makefile.am | 2 -- tools/Makefile.am | 2 -- 3 files changed, 2 insertions(+), 6 deletions(-) -- 2.19.2

2 4

[libvirt] [PATCH] virtio-blk: deprecate SCSI passthrough
by Stefan Hajnoczi 17 Dec '19

17 Dec '19

The Linux virtio_blk.ko guest driver is removing legacy SCSI passthrough support. Deprecate this feature in QEMU too. Signed-off-by: Stefan Hajnoczi <stefanha(a)redhat.com> --- qemu-deprecated.texi | 11 +++++++++++ 1 file changed, 11 insertions(+) diff --git a/qemu-deprecated.texi b/qemu-deprecated.texi index 4b4b7425ac..ef94d497da 100644 --- a/qemu-deprecated.texi +++ b/qemu-deprecated.texi @@ -285,6 +285,17 @@ spec you can use the ``-cpu rv64gcsu,priv_spec=v1.9.1`` command line argument. @section Device options +@subsection Emulated device options + +@subsubsection -device virtio-blk,scsi=on|off (since 5.0.0) + +The virtio-blk SCSI passthrough feature is a legacy VIRTIO feature. VIRTIO 1.0 +and later do not support it because the virtio-scsi device was introduced for +full SCSI support. Use virtio-scsi instead when SCSI passthrough is required. + +Note this also applies to ``-device virtio-blk-pci,scsi=on|off'', which is an +alias. + @subsection Block device options @subsubsection "backing": "" (since 2.12.0) -- 2.23.0

6 7

[libvirt] [PATCH 0/5] Introduce the support of Intel RDT-MBM
by Wang Huaqiang 17 Dec '19

17 Dec '19

RDT is the short for Intel Resource Director Technology, consists of four sub-technologies until now: -. CAT for cache allocation -. CMT for cache usage monitoring -. MBA for memory bandwidth allocation -. MBM for memory bandwidth usage monitoring The Linux kernel interface is 'resctrl' file system, and we have already implemented the support of CAT, CMT and MBA, to accomplish the tasks such as allocating a part of shared CPU last level cache to particular domain vcpu or a list of vcpus and monitoring the usage of cache, or the task of allocating a mount of memory bandwidth to specify domain vcpu(s). This series is to introduce the support of MBM. Basically the interfaces are: ** Identify host capability ** Similar to identify the host capability of CMT, it could be gotten through the result of 'virsh capabilities', if following elements are found, then MBM is supported: <memory_bandwidth> <monitor maxMonitors='176'> <feature name='mbm_total_bytes'/> <feature name='mbm_local_bytes'/> </monitor> </memory_bandwidth> 'mbm_total_bytes' means supporting to report the memory bandwidth used by the vcpu(s) of specific monitor on all CPU sockets. 'mbm_local_bytes' means supporting to report the memory bandwidth used by vcpu(s) that is passing through local CPU socket. ** Create monitor group** The monitor group for specific domain vcpus, for example vcpu 0-4, is defined in domain configuration file, in such kind of way: <cputune> <memorytune vcpus='0-4'> <monitor vcpus='0-4'/> </memorytune> </cputune> ** Report memory usage ** Introduced an option '--memory' against 'virsh domstats' command to show the memory bandwidth usage in such way: (also very similar to the format of CMT result.) # virsh domstats --memory Domain: 'libvirt-vm' memory.bandwidth.monitor.count=4 memory.bandwidth.monitor.0.name=vcpus_0-4 memory.bandwidth.monitor.0.vcpus=0-4 memory.bandwidth.monitor.0.node.count=2 memory.bandwidth.monitor.0.node.0.id=0 memory.bandwidth.monitor.0.node.0.bytes.total=14201651200 memory.bandwidth.monitor.0.node.0.bytes.local=7369809920 memory.bandwidth.monitor.0.node.1.id=1 memory.bandwidth.monitor.0.node.1.bytes.total=188897640448 memory.bandwidth.monitor.0.node.1.bytes.local=170044047360 Huaqiang (5): util, resctrl: using 64bit interface instead of 32bit for counters conf: showing cache/memoryBW monitor features in capabilities cachetune schema: a looser check for the order of <cache> and <monitor> element conf: Parse dommon configure file for memorytune monitors virsh: show memoryBW info in 'virsh domstats' command docs/schemas/domaincommon.rng | 91 +++++++++--------- include/libvirt/libvirt-domain.h | 1 + src/conf/capabilities.c | 4 +- src/conf/domain_conf.c | 44 +++++++-- src/libvirt-domain.c | 21 +++++ src/qemu/qemu_driver.c | 103 ++++++++++++++++++++- src/util/virfile.c | 40 ++++++++ src/util/virfile.h | 2 + src/util/virresctrl.c | 6 +- src/util/virresctrl.h | 2 +- tests/genericxml2xmlindata/cachetune.xml | 1 + tests/genericxml2xmlindata/memorytune.xml | 5 + tests/genericxml2xmloutdata/cachetune.xml | 34 +++++++ tests/genericxml2xmloutdata/memorytune.xml | 42 +++++++++ tests/genericxml2xmltest.c | 4 +- tools/virsh-domain-monitor.c | 7 ++ tools/virsh.pod | 23 ++++- 17 files changed, 367 insertions(+), 63 deletions(-) create mode 100644 tests/genericxml2xmloutdata/cachetune.xml create mode 100644 tests/genericxml2xmloutdata/memorytune.xml -- 2.23.0

4 17

[libvirt] [PATCH v1 00/26] move qemucaps validations from qemu_command to qemu_domain
by Daniel Henrique Barboza 16 Dec '19

16 Dec '19

(series based on master commit 97cafa610ecf5) This work was proposed by Cole in [1]. This is Cole's reasoning for it, copy/pasted from [1]: ------- The benefits of moving to validate time is that XML is rejected by 'virsh define' rather than at 'virsh start' time. It also makes it easier to follow the cli building code, and makes it easier to verify qemu_command.c test suite code coverage for the important stuff like covering every CLI option. It's also a good intermediate step for sharing validation with domain capabilities building, like is done presently for rng models. ------- Cole also mentioned that the machine features validation was a good place to start. I took it a step further and did it across the board on all qemu_command.c. This series didn't create any new validation, just moved them to domain define time. Any other outcome is unintended. Not all cases where covered in this work. For a first version I decided to do only the most trivial cases. These are the other cases I left out for another day: - all verifications contained in functions that are also called by qemu_hotplug.c. This means that the hotplug process will also use the validation code, and we can't just move it to qemu_domain.c. Besides, not all validation done in domain define time applies to hotplug, so it's not simply a matter of calling the same function from qemu_domain in qemu_hotplug. I have patches that moved the verification of all virtio options to qemu_domain.c, while also considering qemu_hotplug validation. In the end I decided to leave it away from this work for now because (1) it took 8 patches just for virtio case alone and (2) there are a lot of these cases in qemu_command.c and it would be too much to do it all at in this same series. - moving CPU Model validation is trickier than the rest because there are code in DefPostParse() that makes CPU Model changes that are then validated in qemu_command.c. Moving the validation to define time doesn't cut in this case - the validation is considering PostParse changes in the CPU Model and some of the will fail if done by qemuDomainDefValidate time. I didn't think too much about how to handle this case, but given that the approach would be different from the other cases handled here, I left it out too. - SHMem: part of SHMem validation is being used by hotplug code. I could've moved the non-hotplug validation to define time, instead I decided it's better to to leave it to do it all at once in the future. - DBus-VMState: validation of this fella must be done by first querying if the hash vm->priv->dbusVMState has elements. qemuDomainDefValidate does not have 'vm->priv' in it's API, meaning that I would need to either change the API to include it (which means changing domainValidateCallback), or do the validation by another branch where I can get access to vm->priv. - vmcoreinfo: there are no challenges in moving vmcoreinfo validation to qemuDomainDefValidate. The problem were the unit tests. Moving this validation to domain define time breaks 925 tests on qemuxml2xmltest.c, including tests labelled as 'minimal' which should pass under any circurstances, as far as I understand. It is possible to just shoehorn QEMU_CAPS_DEVICE_VMCOREINFO in the qemuCaps of all tests, but this would tamper the idea of having to deal with NONE or LATEST or a specific set of capabilities for certain tests. Another case I would rather discuss separately. [1] https://www.redhat.com/archives/libvir-list/2019-October/msg00568.html Daniel Henrique Barboza (26): qemu_command.c: move PSeries features validation to qemu_domain.c qemu_command.c: move mem.nosharepages validation to qemu_domain.c qemu_command.c: move validation of vmport to qemu_domain.c qemu_command.c: move nvdimm validation to qemu_domain.c qemu_command.c: move I/O APIC validation to qemu_domain.c numa_conf: add virDomainNumaNodesDistancesAreBeingSet() helper qemu_command.c move NUMA validation to qemu_domain.c qemu_command.c: move NVRAM validation to qemu_domain.c qemu_command: move qemuBuildSoundDevStr caps validation to qemu_domain qemu_command.c: move sound codec validation to qemu_domain.c qemu_command: move qemuBuildHubDevStr caps validation to qemu_domain qemu_command: move qemuBuildChrChardevStr caps validation to qemu_domain qemu: move qemuBuildHostdevCommandLine caps validation to qemu_domain qemu_command.c: move vmGenID validation to qemu_domain.c qemu: move qemuBuildSgaCommandLine validation to qemu_domain.c qemu: move virDomainClockDef validation to qemu_domain.c qemu: move qemuBuildPMCommandLine validation to qemu_domain.c qemu: move qemuBuildBootCommandLine validation to qemu_domain.c qemu_command.c: move pcihole64 validation to qemu_domain.c qemu: move qemuBuildGraphicsSDLCommandLine validation to qemu_domain.c qemu: move qemuBuildGraphicsVNCCommandLine validation to qemu_domain.c qemu: move qemuBuildGraphicsSPICECommandLine validation to qemu_domain.c qemu: move qemuBuildGraphicsEGLHeadlessCommandLine validation to qemu_domain.c qemu: move qemuBuildSmartcardCommandLine validation to qemu_domain.c qemu: move qemuBuildConsoleCommandLine validation to qemu_domain.c qemu: move qemuBuildTPMDevStr TPM validation to qemu_domain.c src/conf/numa_conf.c | 19 + src/conf/numa_conf.h | 2 + src/libvirt_private.syms | 1 + src/qemu/qemu_command.c | 575 +--------- src/qemu/qemu_command.h | 1 + src/qemu/qemu_domain.c | 983 ++++++++++++++++-- tests/qemuhotplugtest.c | 10 + tests/qemumemlocktest.c | 16 +- .../default-video-type-aarch64.xml | 2 +- .../default-video-type-ppc64.xml | 2 +- .../default-video-type-s390x.xml | 2 +- .../graphics-egl-headless.args | 2 +- .../graphics-spice-egl-headless.args | 2 +- .../graphics-vnc-egl-headless.args | 2 +- tests/qemuxml2argvtest.c | 80 +- ...ault-video-type-aarch64.aarch64-latest.xml | 4 +- .../default-video-type-ppc64.ppc64-latest.xml | 4 +- .../default-video-type-s390x.s390x-latest.xml | 4 +- tests/qemuxml2xmltest.c | 244 +++-- 19 files changed, 1246 insertions(+), 709 deletions(-) -- 2.23.0

4 42