- Devel - libvirt List Archives

[libvirt] [PATCH 0/6] Fixes for memory locking limit
by Andrea Bolognani 20 Nov '15

20 Nov '15

The code dealing with RLIMIT_MEMLOCK contained a few assumptions that hold for x86 but don't necessarily work as well for other platforms, eg. that qemu will need to lock memory only when VFIO passthrough is involved. This series removes such assumptions by removing ad-hoc code and making sure that the function containing the correct logic is called instead; it also implements the platform-specific calculations needed to calculate the correct RLIMIT_MEMLOCK value for ppc64 guests. Patches 1-4 are architecture-agnostic, patches 5-6 are specific to ppc64. Cheers. Andrea Bolognani (6): process: Log when limiting the amount of locked memory qemu: Use qemuDomainRequiresMlock() in qemuBuildCommandLine() qemu: Use qemuDomainRequiresMlock() when attaching PCI hostdev qemu: Reduce memlock limit after detaching hostdev qemu: Always set locked memory limit for ppc64 domains qemu: Add ppc64-specific math to qemuDomainGetMlockLimitBytes() src/qemu/qemu_command.c | 9 ++--- src/qemu/qemu_domain.c | 89 +++++++++++++++++++++++++++++++++++++++++++++++-- src/qemu/qemu_hotplug.c | 29 ++++++++++++---- src/util/virprocess.c | 4 +++ 4 files changed, 116 insertions(+), 15 deletions(-) -- 2.5.0

3 18

[libvirt] [PATCHv4 0/6] Hyper-v crash feature support
by Dmitry Andreev 19 Nov '15

19 Nov '15

A new Hyper-V cpu feature 'hv_crash' was added to QEMU. The feature will become available in v2.5.0. What is changed in v4: * panic model attribute is added even if it wasn't specified by the user [5/6], tests updated. * configuration with multiple panic defices of same model is rejected [6/6] 1-4 commits have no changes. The previous version: https://www.redhat.com/archives/libvir-list/2015-November/msg00457.html Dmitry Andreev (6): conf: add 'model' attribute for panic device with values isa, pseries, hyperv qemu: add support for hv_crash feature as a panic device schema: allow multiple panic devices tests: add tests for panic models and multiple panic devices qemu: set panic device model if not specified by the user conf: reject multiple panic devices of same model docs/formatdomain.html.in | 19 ++- docs/schemas/domaincommon.rng | 13 ++- src/conf/domain_conf.c | 128 +++++++++++++-------- src/conf/domain_conf.h | 15 ++- src/qemu/qemu_command.c | 74 ++++++++++-- src/qemu/qemu_domain.c | 30 ++++- tests/qemuargv2xmltest.c | 1 + .../qemuxml2argv-hyperv-panic.args | 21 ++++ .../qemuxml2argvdata/qemuxml2argv-hyperv-panic.xml | 25 ++++ .../qemuxml2argv-panic-double.args | 21 ++++ .../qemuxml2argvdata/qemuxml2argv-panic-double.xml | 28 +++++ tests/qemuxml2argvdata/qemuxml2argv-panic-isa.xml | 31 +++++ .../qemuxml2argv-panic-no-address.xml | 2 +- .../qemuxml2argv-panic-pseries.xml | 30 +++++ .../qemuxml2argvdata/qemuxml2argv-pseries-disk.xml | 2 +- .../qemuxml2argv-pseries-nvram.xml | 2 +- tests/qemuxml2argvtest.c | 3 + tests/qemuxml2xmloutdata/qemuxml2xmlout-panic.xml | 31 +++++ .../qemuxml2xmlout-pseries-panic-missing.xml | 2 +- .../qemuxml2xmlout-pseries-panic-no-address.xml | 30 +++++ tests/qemuxml2xmltest.c | 8 +- 21 files changed, 446 insertions(+), 70 deletions(-) create mode 100644 tests/qemuxml2argvdata/qemuxml2argv-hyperv-panic.args create mode 100644 tests/qemuxml2argvdata/qemuxml2argv-hyperv-panic.xml create mode 100644 tests/qemuxml2argvdata/qemuxml2argv-panic-double.args create mode 100644 tests/qemuxml2argvdata/qemuxml2argv-panic-double.xml create mode 100644 tests/qemuxml2argvdata/qemuxml2argv-panic-isa.xml create mode 100644 tests/qemuxml2argvdata/qemuxml2argv-panic-pseries.xml create mode 100644 tests/qemuxml2xmloutdata/qemuxml2xmlout-panic.xml create mode 100644 tests/qemuxml2xmloutdata/qemuxml2xmlout-pseries-panic-no-address.xml -- 1.8.3.1

2 12

[libvirt] [PATCH 0/2] libxl: a few minor fixes
by Jim Fehlig 19 Nov '15

19 Nov '15

A few minor fixes for patches 1 and 2 of Joao's series adding stats support to the libxl driver https://www.redhat.com/archives/libvir-list/2015-November/msg00489.html I only noticed the minor flaws while reviewing patches later in the series, after already committing patches 1 and 2. Jim Fehlig (2): libxl: unref libxlDriverConfig object libxl: don't unlock virDomainObj if refcnt is 0 src/libxl/libxl_driver.c | 26 +++++++++++++++++--------- 1 file changed, 17 insertions(+), 9 deletions(-) -- 2.1.4

3 7

[libvirt] qemuParseCommandLine and virDomainDefPostParse (and virDomaniDefAddImplicitControllers)
by Laine Stump 19 Nov '15

19 Nov '15

While playing with the idea of forcing explicit USB controller models, I ended up with the qemuargv2xml test failing; it ended up that this was because changes I had made to qemuDomainDefPostParse() were affecting the XML produced by qemuParseCommandline(). The reason - after constructing a virDomainDef object by parsing a qemu commandline, qemuParseCommandline() calls two functions that are supposed to be called after parsing domain XML - virDomainDefPostParse() (which calls qemuDomainDefPostParse()). In my opinion, qemuParseCommandLine() shouldn't be calling virDomainDefPostParse() (or virDomainefAddImplicitControllers(), which it calls in the wrong order relative to virDomainDefPostParse() BTW). The reasons are: 1) this is causing the argv-to-xml conversion to include things in the XML that were not on the original commandline, in particular "default" devices like PCI and USB controllers (added in qemuDomainDefPostParse() based on machinetype) as well as disk, smartcard, virtio-serial, and hostdev-scsi controllers in virDomainDefAddImplicitControllers() (why the duality there, anyway?) 2) If the output of argv-to-xml is used for a virDomainDefine, those post-parse functions will be called then, and the implicit/auto devices will be added at that time anyway, so in practice nothing is gained by adding them during argv-to-xml. Does anyone else have an opinion about this?

3 5

[libvirt] [PATCH v2] vz: implementation of domainReboot callback
by Mikhail Feoktistov 19 Nov '15

19 Nov '15

Diff from v1. 1. Add virCheckFlags() call in vzDomainReboot --- src/vz/vz_driver.c | 8 ++++++++ src/vz/vz_sdk.c | 8 ++++++++ src/vz/vz_sdk.h | 1 + 3 files changed, 17 insertions(+) diff --git a/src/vz/vz_driver.c b/src/vz/vz_driver.c index 0a968b9..39f58a4 100644 --- a/src/vz/vz_driver.c +++ b/src/vz/vz_driver.c @@ -923,6 +923,13 @@ static int vzDomainShutdown(virDomainPtr domain) return prlsdkDomainChangeState(domain, prlsdkStop); } +static int vzDomainReboot(virDomainPtr domain, + unsigned int flags) +{ + virCheckFlags(0, -1); + return prlsdkDomainChangeState(domain, prlsdkRestart); +} + static int vzDomainIsActive(virDomainPtr domain) { virDomainObjPtr dom = NULL; @@ -1486,6 +1493,7 @@ static virHypervisorDriver vzDriver = { .domainShutdown = vzDomainShutdown, /* 0.10.0 */ .domainCreate = vzDomainCreate, /* 0.10.0 */ .domainCreateWithFlags = vzDomainCreateWithFlags, /* 1.2.10 */ + .domainReboot = vzDomainReboot, /* 1.2.22 */ .domainDefineXML = vzDomainDefineXML, /* 0.10.0 */ .domainDefineXMLFlags = vzDomainDefineXMLFlags, /* 1.2.12 */ .domainUndefine = vzDomainUndefine, /* 1.2.10 */ diff --git a/src/vz/vz_sdk.c b/src/vz/vz_sdk.c index 89c9e89..bf71e96 100644 --- a/src/vz/vz_sdk.c +++ b/src/vz/vz_sdk.c @@ -1831,6 +1831,14 @@ PRL_RESULT prlsdkSuspend(PRL_HANDLE sdkdom) return waitJob(job); } +PRL_RESULT prlsdkRestart(PRL_HANDLE sdkdom) +{ + PRL_HANDLE job = PRL_INVALID_HANDLE; + + job = PrlVm_Restart(sdkdom); + return waitJob(job); +} + int prlsdkDomainChangeStateLocked(vzConnPtr privconn, virDomainObjPtr dom, diff --git a/src/vz/vz_sdk.h b/src/vz/vz_sdk.h index ebe4591..88ee7d9 100644 --- a/src/vz/vz_sdk.h +++ b/src/vz/vz_sdk.h @@ -41,6 +41,7 @@ PRL_RESULT prlsdkStop(PRL_HANDLE sdkdom); PRL_RESULT prlsdkPause(PRL_HANDLE sdkdom); PRL_RESULT prlsdkResume(PRL_HANDLE sdkdom); PRL_RESULT prlsdkSuspend(PRL_HANDLE sdkdom); +PRL_RESULT prlsdkRestart(PRL_HANDLE sdkdom); typedef PRL_RESULT (*prlsdkChangeStateFunc)(PRL_HANDLE sdkdom); int -- 1.8.3.1

3 2

[libvirt] [PATCH] vz: allow only en-us keymap for VNC
by Mikhail Feoktistov 19 Nov '15

19 Nov '15

In virtuozzo we support only en-us keymap for VMs and containers. If keymap is specified than check that it's en-us, otherwise show error message. --- src/vz/vz_sdk.c | 7 ++++--- 1 file changed, 4 insertions(+), 3 deletions(-) diff --git a/src/vz/vz_sdk.c b/src/vz/vz_sdk.c index 89c9e89..526c0f2 100644 --- a/src/vz/vz_sdk.c +++ b/src/vz/vz_sdk.c @@ -2220,10 +2220,11 @@ static int prlsdkCheckGraphicsUnsupportedParams(virDomainDefPtr def) return -1; } - if (gr->data.vnc.keymap != 0) { + if (gr->data.vnc.keymap != 0 && + STRNEQ(gr->data.vnc.keymap, "en-us")) { virReportError(VIR_ERR_CONFIG_UNSUPPORTED, "%s", - _("vz driver doesn't support " - "keymap setting for VNC graphics.")); + _("vz driver supports only " + "\"en-us\" keymap for VNC graphics.")); return -1; } -- 1.8.3.1

2 1

[libvirt] [PATCH] set UID and GID according to dynamicOwnership
by Matteo Cafasso 19 Nov '15

19 Nov '15

When called with dynamicOwnership set, the qemuOpenFileAs function ignores it resulting in files not created with proper permissions. The issue is reported in: https://www.redhat.com/archives/libvirt-users/2015-November/msg00039.html and also mentioned in: https://www.redhat.com/archives/libvir-list/2015-November/msg00453.html Signed-off-by: Matteo Cafasso <noxdafox(a)gmail.com> --- src/qemu/qemu_driver.c | 5 +++++ 1 file changed, 5 insertions(+) diff --git a/src/qemu/qemu_driver.c b/src/qemu/qemu_driver.c index 92a9961..b3b59b6 100644 --- a/src/qemu/qemu_driver.c +++ b/src/qemu/qemu_driver.c @@ -2932,6 +2932,11 @@ qemuOpenFileAs(uid_t fallback_uid, gid_t fallback_gid, if (path_shared <= 0 || dynamicOwnership) vfoflags |= VIR_FILE_OPEN_FORCE_OWNER; + if (dynamicOwnership) { + uid = fallback_uid; + gid = fallback_gid; + } + if (stat(path, &sb) == 0) { /* It already exists, we don't want to delete it on error */ need_unlink = false; -- 2.6.2

2 1

[libvirt] [[PATCH v6] autocreate tap device for VIR_DOMAIN_NET_TYPE_ETHERNET] autocreate tap device for VIR_DOMAIN_NET_TYPE_ETHERNET
by Vasiliy Tolstov 19 Nov '15

19 Nov '15

If a user specify ehernet device create it via libvirt and run script if it provided. After this commit user does not need to run external script to create tap device or add root to qemu process. Signed-off-by: Vasiliy Tolstov <v.tolstov(a)selfip.ru> --- src/qemu/qemu_command.c | 142 +++++++++++++++++++++++++++++++----------------- src/qemu/qemu_hotplug.c | 13 ++--- src/qemu/qemu_process.c | 6 ++ 3 files changed, 101 insertions(+), 60 deletions(-) diff --git a/src/qemu/qemu_command.c b/src/qemu/qemu_command.c index 3886b4f..6d26d28 100644 --- a/src/qemu/qemu_command.c +++ b/src/qemu/qemu_command.c @@ -332,10 +332,39 @@ static int qemuCreateInBridgePortWithHelper(virQEMUDriverConfigPtr cfg, return *tapfd < 0 ? -1 : 0; } +/** + * qemuExecuteEthernetScript: + * @ifname: the interface name + * @script: the script name + * + * This function executes script for new tap device created by libvirt. + * Returns 0 in case of success or -1 on failure + */ +static int +qemuExecuteEthernetScript(const char *ifname, const char *script) +{ + virCommandPtr cmd; + int ret; + + cmd = virCommandNew(script); + virCommandAddArgFormat(cmd, "%s", ifname); + virCommandClearCaps(cmd); +#ifdef CAP_NET_ADMIN + virCommandAllowCap(cmd, CAP_NET_ADMIN); +#endif + virCommandAddEnvPassCommon(cmd); + + ret = virCommandRun(cmd, NULL); + + virCommandFree(cmd); + return ret; +} + /* qemuNetworkIfaceConnect - *only* called if actualType is - * VIR_DOMAIN_NET_TYPE_NETWORK or VIR_DOMAIN_NET_TYPE_BRIDGE (i.e. if - * the connection is made with a tap device connecting to a bridge - * device) + * VIR_DOMAIN_NET_TYPE_NETWORK, VIR_DOMAIN_NET_TYPE_BRIDGE + * VIR_DOMAIN_NET_TYPE_ETHERNET (i.e. if the connection is + * made with a tap device connecting to a bridge device or + * use plain tap device) */ int qemuNetworkIfaceConnect(virDomainDefPtr def, @@ -351,6 +380,7 @@ qemuNetworkIfaceConnect(virDomainDefPtr def, bool template_ifname = false; virQEMUDriverConfigPtr cfg = virQEMUDriverGetConfig(driver); const char *tunpath = "/dev/net/tun"; + virMacAddr tapmac; if (net->backend.tap) { tunpath = net->backend.tap; @@ -361,11 +391,6 @@ qemuNetworkIfaceConnect(virDomainDefPtr def, } } - if (!(brname = virDomainNetGetActualBridgeName(net))) { - virReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("Missing bridge name")); - goto cleanup; - } - if (!net->ifname || STRPREFIX(net->ifname, VIR_NET_GENERATED_PREFIX) || strchr(net->ifname, '%')) { @@ -381,40 +406,65 @@ qemuNetworkIfaceConnect(virDomainDefPtr def, tap_create_flags |= VIR_NETDEV_TAP_CREATE_VNET_HDR; } - if (cfg->privileged) { - if (virNetDevTapCreateInBridgePort(brname, &net->ifname, &net->mac, - def->uuid, tunpath, tapfd, *tapfdSize, - virDomainNetGetActualVirtPortProfile(net), - virDomainNetGetActualVlan(net), - tap_create_flags) < 0) { + if (virDomainNetGetActualType(net) == VIR_DOMAIN_NET_TYPE_ETHERNET) { + if (virNetDevTapCreate(&net->ifname, tunpath, tapfd, *tapfdSize, + tap_create_flags) < 0) { virDomainAuditNetDevice(def, net, tunpath, false); goto cleanup; } - if (virDomainNetGetActualBridgeMACTableManager(net) - == VIR_NETWORK_BRIDGE_MAC_TABLE_MANAGER_LIBVIRT) { - /* libvirt is managing the FDB of the bridge this device - * is attaching to, so we need to turn off learning and - * unicast_flood on the device to prevent the kernel from - * adding any FDB entries for it. We will add add an fdb - * entry ourselves (during qemuInterfaceStartDevices(), - * using the MAC address from the interface config. - */ - if (virNetDevBridgePortSetLearning(brname, net->ifname, false) < 0) - goto cleanup; - if (virNetDevBridgePortSetUnicastFlood(brname, net->ifname, false) < 0) + virMacAddrSet(&tapmac, &net->mac); + + if (virNetDevSetMAC(net->ifname, &tapmac) < 0) + goto cleanup; + + if (virNetDevSetOnline(net->ifname, true) < 0) + goto cleanup; + + if (net->script) { + if (qemuExecuteEthernetScript(net->ifname, net->script) < 0) goto cleanup; } } else { - if (qemuCreateInBridgePortWithHelper(cfg, brname, - &net->ifname, - tapfd, tap_create_flags) < 0) { - virDomainAuditNetDevice(def, net, tunpath, false); + if (!(brname = virDomainNetGetActualBridgeName(net))) { + virReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("Missing bridge name")); goto cleanup; } - /* qemuCreateInBridgePortWithHelper can only create a single FD */ - if (*tapfdSize > 1) { - VIR_WARN("Ignoring multiqueue network request"); - *tapfdSize = 1; + + if (cfg->privileged) { + if (virNetDevTapCreateInBridgePort(brname, &net->ifname, &net->mac, + def->uuid, tunpath, tapfd, *tapfdSize, + virDomainNetGetActualVirtPortProfile(net), + virDomainNetGetActualVlan(net), + tap_create_flags) < 0) { + virDomainAuditNetDevice(def, net, tunpath, false); + goto cleanup; + } + if (virDomainNetGetActualBridgeMACTableManager(net) + == VIR_NETWORK_BRIDGE_MAC_TABLE_MANAGER_LIBVIRT) { + /* libvirt is managing the FDB of the bridge this device + * is attaching to, so we need to turn off learning and + * unicast_flood on the device to prevent the kernel from + * adding any FDB entries for it. We will add add an fdb + * entry ourselves (during qemuInterfaceStartDevices(), + * using the MAC address from the interface config. + */ + if (virNetDevBridgePortSetLearning(brname, net->ifname, false) < 0) + goto cleanup; + if (virNetDevBridgePortSetUnicastFlood(brname, net->ifname, false) < 0) + goto cleanup; + } + } else { + if (qemuCreateInBridgePortWithHelper(cfg, brname, + &net->ifname, + tapfd, tap_create_flags) < 0) { + virDomainAuditNetDevice(def, net, tunpath, false); + goto cleanup; + } + /* qemuCreateInBridgePortWithHelper can only create a single FD */ + if (*tapfdSize > 1) { + VIR_WARN("Ignoring multiqueue network request"); + *tapfdSize = 1; + } } } @@ -5221,6 +5271,7 @@ qemuBuildHostNetStr(virDomainNetDefPtr net, case VIR_DOMAIN_NET_TYPE_BRIDGE: case VIR_DOMAIN_NET_TYPE_NETWORK: case VIR_DOMAIN_NET_TYPE_DIRECT: + case VIR_DOMAIN_NET_TYPE_ETHERNET: virBufferAsprintf(&buf, "tap%c", type_sep); /* for one tapfd 'fd=' shall be used, * for more than one 'fds=' is the right choice */ @@ -5238,20 +5289,6 @@ qemuBuildHostNetStr(virDomainNetDefPtr net, is_tap = true; break; - case VIR_DOMAIN_NET_TYPE_ETHERNET: - virBufferAddLit(&buf, "tap"); - if (net->ifname) { - virBufferAsprintf(&buf, "%cifname=%s", type_sep, net->ifname); - type_sep = ','; - } - if (net->script) { - virBufferAsprintf(&buf, "%cscript=%s", type_sep, - net->script); - type_sep = ','; - } - is_tap = true; - break; - case VIR_DOMAIN_NET_TYPE_CLIENT: virBufferAsprintf(&buf, "socket%cconnect=%s:%d", type_sep, @@ -8226,7 +8263,8 @@ qemuBuildInterfaceCommandLine(virCommandPtr cmd, /* Currently nothing besides TAP devices supports multiqueue. */ if (net->driver.virtio.queues > 0 && !(actualType == VIR_DOMAIN_NET_TYPE_NETWORK || - actualType == VIR_DOMAIN_NET_TYPE_BRIDGE)) { + actualType == VIR_DOMAIN_NET_TYPE_BRIDGE || + actualType == VIR_DOMAIN_NET_TYPE_ETHERNET)) { virReportError(VIR_ERR_CONFIG_UNSUPPORTED, _("Multiqueue network is not supported for: %s"), virDomainNetTypeToString(actualType)); @@ -8235,7 +8273,8 @@ qemuBuildInterfaceCommandLine(virCommandPtr cmd, if (net->backend.tap && !(actualType == VIR_DOMAIN_NET_TYPE_NETWORK || - actualType == VIR_DOMAIN_NET_TYPE_BRIDGE)) { + actualType == VIR_DOMAIN_NET_TYPE_BRIDGE || + actualType == VIR_DOMAIN_NET_TYPE_ETHERNET)) { virReportError(VIR_ERR_CONFIG_UNSUPPORTED, _("Custom tap device path is not supported for: %s"), virDomainNetTypeToString(actualType)); @@ -8245,7 +8284,8 @@ qemuBuildInterfaceCommandLine(virCommandPtr cmd, cfg = virQEMUDriverGetConfig(driver); if (actualType == VIR_DOMAIN_NET_TYPE_NETWORK || - actualType == VIR_DOMAIN_NET_TYPE_BRIDGE) { + actualType == VIR_DOMAIN_NET_TYPE_BRIDGE || + actualType == VIR_DOMAIN_NET_TYPE_ETHERNET) { tapfdSize = net->driver.virtio.queues; if (!tapfdSize) tapfdSize = 1; diff --git a/src/qemu/qemu_hotplug.c b/src/qemu/qemu_hotplug.c index cc86a3b..21ea3fd 100644 --- a/src/qemu/qemu_hotplug.c +++ b/src/qemu/qemu_hotplug.c @@ -908,7 +908,8 @@ int qemuDomainAttachNetDevice(virConnectPtr conn, /* Currently nothing besides TAP devices supports multiqueue. */ if (net->driver.virtio.queues > 0 && !(actualType == VIR_DOMAIN_NET_TYPE_NETWORK || - actualType == VIR_DOMAIN_NET_TYPE_BRIDGE)) { + actualType == VIR_DOMAIN_NET_TYPE_BRIDGE || + actualType == VIR_DOMAIN_NET_TYPE_ETHERNET)) { virReportError(VIR_ERR_CONFIG_UNSUPPORTED, _("Multiqueue network is not supported for: %s"), virDomainNetTypeToString(actualType)); @@ -916,7 +917,8 @@ int qemuDomainAttachNetDevice(virConnectPtr conn, } if (actualType == VIR_DOMAIN_NET_TYPE_BRIDGE || - actualType == VIR_DOMAIN_NET_TYPE_NETWORK) { + actualType == VIR_DOMAIN_NET_TYPE_NETWORK || + actualType == VIR_DOMAIN_NET_TYPE_ETHERNET) { tapfdSize = vhostfdSize = net->driver.virtio.queues; if (!tapfdSize) tapfdSize = vhostfdSize = 1; @@ -947,13 +949,6 @@ int qemuDomainAttachNetDevice(virConnectPtr conn, iface_connected = true; if (qemuOpenVhostNet(vm->def, net, priv->qemuCaps, vhostfd, &vhostfdSize) < 0) goto cleanup; - } else if (actualType == VIR_DOMAIN_NET_TYPE_ETHERNET) { - vhostfdSize = 1; - if (VIR_ALLOC(vhostfd) < 0) - goto cleanup; - *vhostfd = -1; - if (qemuOpenVhostNet(vm->def, net, priv->qemuCaps, vhostfd, &vhostfdSize) < 0) - goto cleanup; } /* Set device online immediately */ diff --git a/src/qemu/qemu_process.c b/src/qemu/qemu_process.c index 64ee049..d866e44 100644 --- a/src/qemu/qemu_process.c +++ b/src/qemu/qemu_process.c @@ -5205,6 +5205,12 @@ void qemuProcessStop(virQEMUDriverPtr driver, cfg->stateDir)); VIR_FREE(net->ifname); break; + case VIR_DOMAIN_NET_TYPE_ETHERNET: + if (net->ifname) { + ignore_value(virNetDevTapDelete(net->ifname, net->backend.tap)); + VIR_FREE(net->ifname); + } + break; case VIR_DOMAIN_NET_TYPE_BRIDGE: case VIR_DOMAIN_NET_TYPE_NETWORK: #ifdef VIR_NETDEV_TAP_REQUIRE_MANUAL_CLEANUP -- 2.3.3

4 9

[libvirt] [PATCH] vz: implementation of domainReboot callback
by Mikhail Feoktistov 19 Nov '15

19 Nov '15

--- src/vz/vz_driver.c | 8 ++++++++ src/vz/vz_sdk.c | 8 ++++++++ src/vz/vz_sdk.h | 1 + 3 files changed, 17 insertions(+) diff --git a/src/vz/vz_driver.c b/src/vz/vz_driver.c index 0a968b9..5f56eaf 100644 --- a/src/vz/vz_driver.c +++ b/src/vz/vz_driver.c @@ -923,6 +923,13 @@ static int vzDomainShutdown(virDomainPtr domain) return prlsdkDomainChangeState(domain, prlsdkStop); } +static int vzDomainReboot(virDomainPtr domain, + unsigned int flags) +{ + flags = flags; + return prlsdkDomainChangeState(domain, prlsdkRestart); +} + static int vzDomainIsActive(virDomainPtr domain) { virDomainObjPtr dom = NULL; @@ -1486,6 +1493,7 @@ static virHypervisorDriver vzDriver = { .domainShutdown = vzDomainShutdown, /* 0.10.0 */ .domainCreate = vzDomainCreate, /* 0.10.0 */ .domainCreateWithFlags = vzDomainCreateWithFlags, /* 1.2.10 */ + .domainReboot = vzDomainReboot, /* 1.2.22 */ .domainDefineXML = vzDomainDefineXML, /* 0.10.0 */ .domainDefineXMLFlags = vzDomainDefineXMLFlags, /* 1.2.12 */ .domainUndefine = vzDomainUndefine, /* 1.2.10 */ diff --git a/src/vz/vz_sdk.c b/src/vz/vz_sdk.c index 89c9e89..bf71e96 100644 --- a/src/vz/vz_sdk.c +++ b/src/vz/vz_sdk.c @@ -1831,6 +1831,14 @@ PRL_RESULT prlsdkSuspend(PRL_HANDLE sdkdom) return waitJob(job); } +PRL_RESULT prlsdkRestart(PRL_HANDLE sdkdom) +{ + PRL_HANDLE job = PRL_INVALID_HANDLE; + + job = PrlVm_Restart(sdkdom); + return waitJob(job); +} + int prlsdkDomainChangeStateLocked(vzConnPtr privconn, virDomainObjPtr dom, diff --git a/src/vz/vz_sdk.h b/src/vz/vz_sdk.h index ebe4591..88ee7d9 100644 --- a/src/vz/vz_sdk.h +++ b/src/vz/vz_sdk.h @@ -41,6 +41,7 @@ PRL_RESULT prlsdkStop(PRL_HANDLE sdkdom); PRL_RESULT prlsdkPause(PRL_HANDLE sdkdom); PRL_RESULT prlsdkResume(PRL_HANDLE sdkdom); PRL_RESULT prlsdkSuspend(PRL_HANDLE sdkdom); +PRL_RESULT prlsdkRestart(PRL_HANDLE sdkdom); typedef PRL_RESULT (*prlsdkChangeStateFunc)(PRL_HANDLE sdkdom); int -- 1.8.3.1

2 1

[libvirt] [PATCH v2 00/13] Introduce a virtlogd daemon
by Daniel P. Berrange 19 Nov '15

19 Nov '15

This is a v2 of: https://www.redhat.com/archives/libvir-list/2015-November/msg00085.html Currently we have stdout + stderr of QEMU guests setup to write to a log file (eg /var/log/libvirt/qemu/$GUEST.log). This is nice and simple, but it in fact opens the possibility of a malicious or accidental denial of service, whereby QEMU can write logs of data to stdio and thus fill up the host filesystem holding the log. Although we have logrotate policy in place, this is only processed once a day, so there is still a window where disk usage is not constrained. The only way to solve this is to not let QEMU directly write to the log file, instead connect its stdio to a pipe and copy data from the pipe to the real log file, performing file rollover when it reaches a certain size. If we do this, we need something to keep open the pipe for as long as QEMU is running. This can't be libvirtd since we expect libvirtd to be able to be stopped while QEMU is running. Thus we introduce a new single-purpose daemon virtlogd whose job is exclusively to deal with log file writing. This daemon has support for SIGUSR1 to tell it to re-exec itself while keeping open the pipe to QEMU so it can be safely upgraded while QEMU is running. This series switches QEMU to use virtlogd by default, but in case of problems we can revert back to the old direct file access by setting 'stdio_handler = "file"' in /etc/libvirt/qemu.conf This series is only the first step. The same problem exists when character devices are told to use the file backend. There is a further use case from OpenStack which that they want to allow the use of both a TCP backend and a file backend at the same time. The idea is that the serial port is to be used for an interactive console, so needs the TCP backend support, but we also want to be able to record all output on that serial port to a file for logging purposes. Thus, in a followup I will work on creating a new character device backend "logd" that will use virtlogd to provide this combined tcp+logfile facility for QEMU guests. Changed in v2: - Expanded the virrotatingfile module to handle reading from files and querying position / seeking - Fixed rollover when no backups are enabled - Rollover early if we see a \n in the last 80 bytes - Totally refactor the QEMU code dealing with log files to remove all direct use of file handles and hide it in a qemuDomainLogContextPtr. This ensures we can fetch startup errors from virtlogd when needed Daniel P. Berrange (13): util: add APIs for reading/writing from/to rotating files Import stripped down virtlockd code as basis of virtlogd logging: introduce log handling protocol logging: add client for virtlogd daemon qemu: remove writing to QEMU log file for rename operation qemu: unify code for reporting errors from QEMU log files qemu: introduce a qemuDomainLogContext object qemu: convert log file creation to use qemuDomainLogContextPtr qemu: change qemuDomainTaint APIs to accept qemuDomainLogContextPtr qemu: convert qemuLogOperation to take a qemuDomainLogContextPtr qemu: convert process stop/attach to use qemuDomainLogContextPtr qemu: convert monitor to use qemuDomainLogContextPtr indirectly qemu: add support for sending QEMU stdout/stderr to virtlogd .gitignore | 7 + cfg.mk | 6 +- include/libvirt/virterror.h | 1 + libvirt.spec.in | 24 +- po/POTFILES.in | 6 + src/Makefile.am | 178 +++++- src/libvirt_private.syms | 21 + src/logging/log_daemon.c | 1207 ++++++++++++++++++++++++++++++++++++ src/logging/log_daemon.h | 45 ++ src/logging/log_daemon_config.c | 203 ++++++ src/logging/log_daemon_config.h | 50 ++ src/logging/log_daemon_dispatch.c | 143 +++++ src/logging/log_daemon_dispatch.h | 31 + src/logging/log_handler.c | 524 ++++++++++++++++ src/logging/log_handler.h | 63 ++ src/logging/log_manager.c | 283 +++++++++ src/logging/log_manager.h | 61 ++ src/logging/log_protocol.x | 115 ++++ src/logging/test_virtlogd.aug.in | 12 + src/logging/virtlogd.aug | 45 ++ src/logging/virtlogd.conf | 67 ++ src/logging/virtlogd.init.in | 94 +++ src/logging/virtlogd.pod.in | 162 +++++ src/logging/virtlogd.service.in | 17 + src/logging/virtlogd.socket.in | 8 + src/logging/virtlogd.sysconf | 3 + src/qemu/libvirtd_qemu.aug | 1 + src/qemu/qemu.conf | 15 + src/qemu/qemu_conf.c | 18 + src/qemu/qemu_conf.h | 1 + src/qemu/qemu_domain.c | 353 +++++++---- src/qemu/qemu_domain.h | 37 +- src/qemu/qemu_driver.c | 38 +- src/qemu/qemu_migration.c | 2 +- src/qemu/qemu_monitor.c | 89 +-- src/qemu/qemu_monitor.h | 8 +- src/qemu/qemu_process.c | 320 ++++------ src/qemu/qemu_process.h | 2 - src/qemu/test_libvirtd_qemu.aug.in | 1 + src/util/virerror.c | 1 + src/util/virrotatingfile.c | 608 ++++++++++++++++++ src/util/virrotatingfile.h | 62 ++ tests/Makefile.am | 6 + tests/virrotatingfiletest.c | 698 +++++++++++++++++++++ 44 files changed, 5180 insertions(+), 456 deletions(-) create mode 100644 src/logging/log_daemon.c create mode 100644 src/logging/log_daemon.h create mode 100644 src/logging/log_daemon_config.c create mode 100644 src/logging/log_daemon_config.h create mode 100644 src/logging/log_daemon_dispatch.c create mode 100644 src/logging/log_daemon_dispatch.h create mode 100644 src/logging/log_handler.c create mode 100644 src/logging/log_handler.h create mode 100644 src/logging/log_manager.c create mode 100644 src/logging/log_manager.h create mode 100644 src/logging/log_protocol.x create mode 100644 src/logging/test_virtlogd.aug.in create mode 100644 src/logging/virtlogd.aug create mode 100644 src/logging/virtlogd.conf create mode 100644 src/logging/virtlogd.init.in create mode 100644 src/logging/virtlogd.pod.in create mode 100644 src/logging/virtlogd.service.in create mode 100644 src/logging/virtlogd.socket.in create mode 100644 src/logging/virtlogd.sysconf create mode 100644 src/util/virrotatingfile.c create mode 100644 src/util/virrotatingfile.h create mode 100644 tests/virrotatingfiletest.c -- 2.5.0

3 30