- Devel - libvirt List Archives

[libvirt] [PATCH 0/3] snapshots: Allow <seclabel> for snapshot disk source
by Peter Krempa 20 Jun '19

20 Jun '19

This is just a schema improvement, the code actually already does everything. Peter Krempa (3): docs: snapshot: Encourage people ot use disk 'target' to refer to disks docs: schemas: Add 'seclabel' for external disk snapshot tests: domainsnapshotxml2xml: make 'disk-seclabel' test operational docs/formatsnapshot.html.in | 18 +++++++++++++++--- docs/schemas/domainsnapshot.rng | 6 ++++++ ...-seclabel-invalid.xml => disk-seclabel.xml} | 0 .../domainsnapshotxml2xmlout/disk-seclabel.xml | 15 +++++++++++++++ tests/domainsnapshotxml2xmltest.c | 1 + 5 files changed, 37 insertions(+), 3 deletions(-) rename tests/domainsnapshotxml2xmlin/{disk-seclabel-invalid.xml => disk-seclabel.xml} (100%) create mode 100644 tests/domainsnapshotxml2xmlout/disk-seclabel.xml -- 2.21.0

3 6

[libvirt] [PATCH 00/15] qemu: Add possibility to control qemu capabilities for debug (blockdev-add saga spin-off)
by Peter Krempa 20 Jun '19

20 Jun '19

While this is not strictly required for blockdev to work it was created as a part of the testing (and possibly regression mitigation) effort. These patches introduce the option to add or remove qemu capabilities via the qemu XML namespace custom elements and remove features via qemu.conf. This series depends (very slightly) on the series deleting qemu parsing. To make applying it easy I'd sugges you fetrch everything needed by: git fetch https://github.com/pipo/libvirt.git features-filter Peter Krempa (15): qemu: Rename qemuDomainCmdlineDefPtr to qemuDomainXmlNsDefPtr qemu: Move qemuDomainXmlNsDef(Free) from qemu_conf.(ch) qemu: domain: Use virStringListFreeCount in qemuDomainXmlNsDefFree qemu: Extract parsing of qemu namespace arguments into separate function qemu: Extract parsing of qemu namespace env vars into separate function qemu: Refactor qemuDomainDefNamespaceParse qemu: domain: Split out commandline namespace data formatting conf: Add taint flag for custom hypervisor features qemu: Add support for controling qemu capabilities via the qemu XML namespace qemu: domain: Add support for modifying qemu capability list via qemu namespace tests: qemuxml2argv: Modernize and fork 'qemu-ns' test tests: qemuxml2argv: Test qemu namespace capability tweaking qemu: Remove unused var 'corestr' from virQEMUDriverConfigLoadFile qemu: conf: Add debug option to allow disabling qemu capabilities qemu: Implement support for 'capability_filters' config option docs/drvqemu.html.in | 30 ++ docs/schemas/domaincommon.rng | 19 ++ src/conf/domain_conf.c | 1 + src/conf/domain_conf.h | 1 + src/qemu/libvirtd_qemu.aug | 3 + src/qemu/qemu.conf | 7 + src/qemu/qemu_command.c | 16 +- src/qemu/qemu_conf.c | 39 ++- src/qemu/qemu_conf.h | 17 +- src/qemu/qemu_domain.c | 285 +++++++++++++----- src/qemu/qemu_domain.h | 19 ++ src/qemu/qemu_process.c | 55 ++++ src/qemu/test_libvirtd_qemu.aug.in | 3 + tests/qemuxml2argvdata/qemu-ns.args | 33 -- .../qemu-ns.x86_64-4.0.0.args | 41 +++ .../qemu-ns.x86_64-latest.args | 41 +++ tests/qemuxml2argvdata/qemu-ns.xml | 5 + tests/qemuxml2argvtest.c | 3 +- 18 files changed, 458 insertions(+), 160 deletions(-) delete mode 100644 tests/qemuxml2argvdata/qemu-ns.args create mode 100644 tests/qemuxml2argvdata/qemu-ns.x86_64-4.0.0.args create mode 100644 tests/qemuxml2argvdata/qemu-ns.x86_64-latest.args -- 2.21.0

2 30

[libvirt] [PATCH] docs: Fix Blog Planet links
by Martin Kletzander 20 Jun '19

20 Jun '19

They were pointing to the blogs instead, now they point the articles. Signed-off-by: Martin Kletzander <mkletzan(a)redhat.com> --- Pushed as trivial-and-thoroughly-tested. docs/js/main.js | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/docs/js/main.js b/docs/js/main.js index e48e95075cf6..2edc62830453 100644 --- a/docs/js/main.js +++ b/docs/js/main.js @@ -116,10 +116,10 @@ function fetchRSS() { var name = e.querySelector("author > name").textContent; var title = e.querySelector("title").textContent; var updated = e.querySelector("updated").textContent; - var uri = e.querySelector("author > uri").textContent; + var link = e.querySelector("link").attributes.href.textContent; var a = document.createElement("a"); - a.href = uri; + a.href = link; a.innerText = title; var dt = document.createElement("dt"); -- 2.22.0

1 0

[libvirt] [PATCH] remote: refactor how unprivileged user session connection is identified
by Daniel P. Berrangé 20 Jun '19

20 Jun '19

Currently the VIR_DRV_OPEN_REMOTE_USER flag is only set when we identify that we're connecting to a local libvirtd daemon. We would like to be able to set that even if connecting to a remote libvirtd daemon. This entails refactoring the conditional check. One subtle change is that the VIR_DRV_OPEN_REMOTE_USER is now set when the test+XXX:// URI is used, even if a servername is present. This has no effect in this patch, but will later. Signed-off-by: Daniel P. Berrangé <berrange(a)redhat.com> --- src/remote/remote_driver.c | 26 ++++++++++++++++++-------- 1 file changed, 18 insertions(+), 8 deletions(-) diff --git a/src/remote/remote_driver.c b/src/remote/remote_driver.c index a95781a797..e1eaa56230 100644 --- a/src/remote/remote_driver.c +++ b/src/remote/remote_driver.c @@ -1390,25 +1390,35 @@ remoteConnectOpen(virConnectPtr conn, rflags |= VIR_DRV_OPEN_REMOTE_RO; /* - * If no servername is given, and no +XXX - * transport is listed, or transport is unix, - * and path is /session, and uid is unprivileged - * then auto-spawn a daemon. + * User session daemon is used for + * + * - Any URI with /session suffix + * - Test driver, if a protocol is given + * + * provided we are running non-root */ if (conn->uri && - !conn->uri->server && conn->uri->path && conn->uri->scheme && - (transport == NULL || STREQ(transport, "unix")) && (STREQ(conn->uri->path, "/session") || STRPREFIX(conn->uri->scheme, "test+")) && geteuid() > 0) { - VIR_DEBUG("Auto-spawn user daemon instance"); + VIR_DEBUG("User session daemon required"); rflags |= VIR_DRV_OPEN_REMOTE_USER; + + /* + * Furthermore if no servername is given, and no +XXX + * transport is listed, or transport is unix, + * and uid is unprivileged then auto-spawn a daemon. + */ if (!virIsSUID() && + !conn->uri->server && + (transport == NULL || STREQ(transport, "unix")) && (!autostart || - STRNEQ(autostart, "0"))) + STRNEQ(autostart, "0"))) { + VIR_DEBUG("Try daemon autostart"); rflags |= VIR_DRV_OPEN_REMOTE_AUTOSTART; + } } /* -- 2.21.0

2 1

[libvirt] [PATCH 0/2] error: Allow better reporting of errors in some nested cases
by Peter Krempa 20 Jun '19

20 Jun '19

Peter Krempa (2): util: error: Add API for prefixing last set error with a string qemu: process: Report better error when virtlogd connection fails cfg.mk | 1 + src/libvirt_private.syms | 1 + src/qemu/qemu_process.c | 4 +++- src/util/virerror.c | 38 ++++++++++++++++++++++++++++++++++++++ src/util/virerror.h | 3 +++ 5 files changed, 46 insertions(+), 1 deletion(-) -- 2.21.0

2 4

[libvirt] [PATCH] Revert "spec: Bump minimum supported Fedora version to 29"
by Ján Tomko 20 Jun '19

20 Jun '19

This reverts commit 8a1179831b5edc0a3590489eda693914fc0ff94f. It bumped the version prematurely, before the EOL of Fedora 28. Morover, this arbitrary bump did not let us perform any cleanups in the specfile. Signed-off-by: Ján Tomko <jtomko(a)redhat.com> --- libvirt.spec.in | 2 +- mingw-libvirt.spec.in | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-) diff --git a/libvirt.spec.in b/libvirt.spec.in index 78b7f43934..5c0b4a7605 100644 --- a/libvirt.spec.in +++ b/libvirt.spec.in @@ -4,7 +4,7 @@ # that's still supported by the vendor. It may work on other distros # or versions, but no effort will be made to ensure that going forward. %define min_rhel 7 -%define min_fedora 29 +%define min_fedora 28 %if (0%{?fedora} && 0%{?fedora} >= %{min_fedora}) || (0%{?rhel} && 0%{?rhel} >= %{min_rhel}) %define supported_platform 1 diff --git a/mingw-libvirt.spec.in b/mingw-libvirt.spec.in index 9add033669..8a96ea914c 100644 --- a/mingw-libvirt.spec.in +++ b/mingw-libvirt.spec.in @@ -3,7 +3,7 @@ # This spec file assumes you are building on a Fedora version # that's still supported by the vendor. It may work on other distros # or versions, but no effort will be made to ensure that going forward. -%define min_fedora 29 +%define min_fedora 28 %if 0%{?fedora} && 0%{?fedora} >= %{min_fedora} %define supported_platform 1 -- 2.20.1

3 4

[libvirt] [PATCH] virNetworkGetDHCPLeases: fix docstring format
by Ilias Stamatis 20 Jun '19

20 Jun '19

The docstring of virNetworkGetDHCPLeases is not correctly formatted and as a result the example code snippet appears as normal text under the "Returns:" section. This patch fixes the problem. Signed-off-by: Ilias Stamatis <stamatis.iliass(a)gmail.com> --- src/libvirt-network.c | 42 ++++++++++++++++++++++-------------------- 1 file changed, 22 insertions(+), 20 deletions(-) diff --git a/src/libvirt-network.c b/src/libvirt-network.c index 6ed32c8ba2..2d7a68eaa1 100644 --- a/src/libvirt-network.c +++ b/src/libvirt-network.c @@ -1156,8 +1156,7 @@ virConnectNetworkEventDeregisterAny(virConnectPtr conn, * lease info about a specific guest interface with @mac. There can be * multiple leases for a single @mac because this API supports DHCPv6 too. * - * Returns the number of leases found or -1 and sets @leases to NULL in - * case of error. On success, the array stored into @leases is guaranteed to + * On success, the array stored into @leases is guaranteed to * have an extra allocated element set to NULL but not included in the return * count, to make iteration easier. The caller is responsible for calling * virNetworkDHCPLeaseFree() on each array element, then calling free() on @leases. @@ -1167,30 +1166,33 @@ virConnectNetworkEventDeregisterAny(virConnectPtr conn, * * Example of usage: * - * virNetworkDHCPLeasePtr *leases = NULL; - * virNetworkPtr network = ... obtain a network pointer here ...; - * size_t i; - * int nleases; - * unsigned int flags = 0; + * virNetworkDHCPLeasePtr *leases = NULL; + * virNetworkPtr network = ... obtain a network pointer here ...; + * size_t i; + * int nleases; + * unsigned int flags = 0; + * + * nleases = virNetworkGetDHCPLeases(network, NULL, &leases, flags); + * if (nleases < 0) + * error(); * - * nleases = virNetworkGetDHCPLeases(network, NULL, &leases, flags); - * if (nleases < 0) - * error(); + * ... do something with returned values, for example: * - * ... do something with returned values, for example: + * for (i = 0; i < nleases; i++) { + * virNetworkDHCPLeasePtr lease = leases[i]; * - * for (i = 0; i < nleases; i++) { - * virNetworkDHCPLeasePtr lease = leases[i]; + * printf("Time(epoch): %lu, MAC address: %s, " + * "IP address: %s, Hostname: %s, ClientID: %s\n", + * lease->expirytime, lease->mac, lease->ipaddr, + * lease->hostname, lease->clientid); * - * printf("Time(epoch): %lu, MAC address: %s, " - * "IP address: %s, Hostname: %s, ClientID: %s\n", - * lease->expirytime, lease->mac, lease->ipaddr, - * lease->hostname, lease->clientid); + * virNetworkDHCPLeaseFree(leases[i]); + * } * - * virNetworkDHCPLeaseFree(leases[i]); - * } + * free(leases); * - * free(leases); + * Returns the number of leases found or -1 and sets @leases to NULL in + * case of error. * */ int -- 2.22.0

2 1

[libvirt] [PATCH 0/2] qemu: fix snapshot crasher bug and improve virsh (blockdev-add saga)
by Peter Krempa 20 Jun '19

20 Jun '19

While reviewers liked my story plot, the CPU did not. Fix it to make everyone happy. Peter Krempa (2): qemu: driver: Fix off-by-one in qemuDomainSnapshotDiskDataCollect virsh: snapshot: Don't block --no-metadata with --print-xml src/qemu/qemu_driver.c | 2 +- tools/virsh-snapshot.c | 8 +------- tools/virsh.pod | 4 ++-- 3 files changed, 4 insertions(+), 10 deletions(-) -- 2.21.0

2 4

[libvirt] [PATCH RFC] test_driver: check that the domain is running in testDomainGetTime
by Ilias Stamatis 20 Jun '19

20 Jun '19

Signed-off-by: Ilias Stamatis <stamatis.iliass(a)gmail.com> --- Currently in the test driver in APIs that would normally require guest agents and similar we are just checking if the domain is active (using virDomainObjCheckActive). But a domain will be active even if stopped, so I would say that most of the time this is not sufficient since we really want to now that the domain is actually running, not that it's just active. So something like what I include in this patch is needed instead. I wonder if it would make sense to add a new function such as testDomainAgentAvailable inspired by the QEMU driver. So the check whether the domain is actually running or not could be done there. Also, in that case I'm not sure if calling both virDomainObjCheckActive and testDomainAgentAvailable is needed. I feel like calling the second one only is sufficient. However, in the QEMU driver always both are called. Should I go ahead with implementing the testDomainAgentAvailable function and search for all the places in the test driver that it should be used instead of virDomainObjCheckActive? src/test/test_driver.c | 20 ++++++++++++++++++-- 1 file changed, 18 insertions(+), 2 deletions(-) diff --git a/src/test/test_driver.c b/src/test/test_driver.c index 2a0ffbc6c5..8e5bd4e670 100755 --- a/src/test/test_driver.c +++ b/src/test/test_driver.c @@ -1984,17 +1984,33 @@ testDomainGetState(virDomainPtr domain, } static int -testDomainGetTime(virDomainPtr dom ATTRIBUTE_UNUSED, +testDomainGetTime(virDomainPtr dom, long long *seconds, unsigned int *nseconds, unsigned int flags) { + int ret = -1; + virDomainObjPtr vm = NULL; + virCheckFlags(0, -1); + if (!(vm = testDomObjFromDomain(dom))) + goto cleanup; + + if (virDomainObjGetState(vm, NULL) != VIR_DOMAIN_RUNNING) { + virReportError(VIR_ERR_OPERATION_INVALID, "%s", + _("domain is not running")); + goto cleanup; + } + *seconds = 627319920; *nseconds = 0; - return 0; + ret = 0; + + cleanup: + virDomainObjEndAPI(&vm); + return ret; } #define TEST_SAVE_MAGIC "TestGuestMagic" -- 2.22.0

2 1

[libvirt] [PATCH v3 0/2] Ditch external JavaScript libraries
by Martin Kletzander 20 Jun '19

20 Jun '19

This is a response to all the discussions (mainly) other people had about all the JS code we're currently using, bundling, etc. CORS [1] is set up on planet.virt-tools.org, so we can just get the XML and use it. [1] https://en.wikipedia.org/wiki/Cross-origin_resource_sharing Martin Kletzander (2): docs: Use our own implementation for fetching the RSS data docs: Remove unused JS libraries docs/Makefile.am | 5 +-- docs/index.html.in | 15 ++------- docs/js/jquery-3.1.1.min.js | 4 --- docs/js/jquery.rss.min.js | 11 ------- docs/js/main.js | 62 +++++++++++++++++++++++++++++++++++++ docs/js/moment.min.js | 7 ----- 6 files changed, 65 insertions(+), 39 deletions(-) delete mode 100644 docs/js/jquery-3.1.1.min.js delete mode 100644 docs/js/jquery.rss.min.js delete mode 100644 docs/js/moment.min.js -- 2.22.0

2 5