- Devel - libvirt List Archives

[libvirt] [PATCH 0/4] Fix problems of shared disk management
by Osier Yang 12 Feb '13

12 Feb '13

This fixes several problems of shared disk management, mainly about shared cdrom or floppy disk. Osier Yang (4): qemu: Add checking in helpers for sgio setting qemu: Merge qemuCheckSharedDisk into qemuAddSharedDisk qemu: Don't remove hash entry of other domains qemu: Move shared disk entry adding and unpriv_sgio seting src/qemu/qemu_conf.c | 73 ++++++++++++++++++++++++++++++-- src/qemu/qemu_conf.h | 5 +- src/qemu/qemu_driver.c | 55 +++++++++++------------- src/qemu/qemu_hotplug.c | 6 ++- src/qemu/qemu_hotplug.h | 3 +- src/qemu/qemu_migration.c | 14 +++--- src/qemu/qemu_process.c | 101 ++++++++++++++++----------------------------- src/qemu/qemu_process.h | 6 +-- 8 files changed, 149 insertions(+), 114 deletions(-) -- 1.7.7.6

3 21

[libvirt] [PATCH] Fix potential deadlock across fork() in QEMU driver
by Daniel P. Berrange 12 Feb '13

12 Feb '13

From: "Daniel P. Berrange" <berrange(a)redhat.com> The hook scripts used by virCommand must be careful wrt accessing any mutexes that may have been held by other threads in the parent process. With the recent refactorigng there are 2 potential flaws lurking, which will become real deadlock bugs once the global QEMU driver lock is removed. Remove use of the QEMU driver lock from the hook function by passing in the 'virQEMUDriverConfigPtr' instance directly. Add functions to the virSecurityManager to be invoked before and after fork, to ensure the mutex is held by the current thread. This allows it to be safely used in the hook script in the child procss. Signed-off-by: Daniel P. Berrange <berrange(a)redhat.com> --- src/libvirt_private.syms | 2 ++ src/qemu/qemu_process.c | 16 ++++++++++++---- src/security/security_manager.c | 20 ++++++++++++++++++++ src/security/security_manager.h | 3 +++ 4 files changed, 37 insertions(+), 4 deletions(-) diff --git a/src/libvirt_private.syms b/src/libvirt_private.syms index cb81497..5f19269 100644 --- a/src/libvirt_private.syms +++ b/src/libvirt_private.syms @@ -1054,6 +1054,8 @@ virSecurityManagerGetProcessLabel; virSecurityManagerNew; virSecurityManagerNewDAC; virSecurityManagerNewStack; +virSecurityManagerPostFork; +virSecurityManagerPreFork; virSecurityManagerReleaseLabel; virSecurityManagerReserveLabel; virSecurityManagerRestoreAllLabel; diff --git a/src/qemu/qemu_process.c b/src/qemu/qemu_process.c index 9759332..12e3544 100644 --- a/src/qemu/qemu_process.c +++ b/src/qemu/qemu_process.c @@ -2773,6 +2773,7 @@ struct qemuProcessHookData { virDomainObjPtr vm; virQEMUDriverPtr driver; virBitmapPtr nodemask; + virQEMUDriverConfigPtr cfg; }; static int qemuProcessHook(void *data) @@ -2780,7 +2781,11 @@ static int qemuProcessHook(void *data) struct qemuProcessHookData *h = data; int ret = -1; int fd; - virQEMUDriverConfigPtr cfg = virQEMUDriverGetConfig(h->driver); + /* This method cannot use any mutexes, which are not + * protected across fork() + */ + + virSecurityManagerPostFork(h->driver->securityManager); /* Some later calls want pid present */ h->vm->pid = getpid(); @@ -2796,7 +2801,7 @@ static int qemuProcessHook(void *data) if (virSecurityManagerSetSocketLabel(h->driver->securityManager, h->vm->def) < 0) goto cleanup; if (virDomainLockProcessStart(h->driver->lockManager, - cfg->uri, + h->cfg->uri, h->vm, /* QEMU is always paused initially */ true, @@ -2805,7 +2810,7 @@ static int qemuProcessHook(void *data) if (virSecurityManagerClearSocketLabel(h->driver->securityManager, h->vm->def) < 0) goto cleanup; - if (qemuProcessLimits(cfg) < 0) + if (qemuProcessLimits(h->cfg) < 0) goto cleanup; /* This must take place before exec(), so that all QEMU @@ -2831,7 +2836,7 @@ static int qemuProcessHook(void *data) ret = 0; cleanup: - virObjectUnref(cfg); + virObjectUnref(h->cfg); VIR_DEBUG("Hook complete ret=%d", ret); return ret; } @@ -3642,6 +3647,7 @@ int qemuProcessStart(virConnectPtr conn, hookData.conn = conn; hookData.vm = vm; hookData.driver = driver; + hookData.cfg = virObjectRef(cfg); VIR_DEBUG("Beginning VM startup process"); @@ -3973,7 +3979,9 @@ int qemuProcessStart(virConnectPtr conn, virCommandDaemonize(cmd); virCommandRequireHandshake(cmd); + virSecurityManagerPreFork(driver->securityManager); ret = virCommandRun(cmd, NULL); + virSecurityManagerPostFork(driver->securityManager); /* wait for qemu process to show up */ if (ret == 0) { diff --git a/src/security/security_manager.c b/src/security/security_manager.c index 6f8ddbf..50962ba 100644 --- a/src/security/security_manager.c +++ b/src/security/security_manager.c @@ -192,6 +192,26 @@ virSecurityManagerPtr virSecurityManagerNew(const char *name, requireConfined); } + +/* + * Must be called before fork()'ing to ensure mutex state + * is sane for the child to use + */ +void virSecurityManagerPreFork(virSecurityManagerPtr mgr) +{ + virObjectLock(mgr); +} + + +/* + * Must be called after fork()'ing in both parent and child + * to ensure mutex state is sane for the child to use + */ +void virSecurityManagerPostFork(virSecurityManagerPtr mgr) +{ + virObjectUnlock(mgr); +} + void *virSecurityManagerGetPrivateData(virSecurityManagerPtr mgr) { return mgr->privateData; diff --git a/src/security/security_manager.h b/src/security/security_manager.h index 4d4dc73..8e8accf 100644 --- a/src/security/security_manager.h +++ b/src/security/security_manager.h @@ -46,6 +46,9 @@ virSecurityManagerPtr virSecurityManagerNewDAC(const char *virtDriver, bool requireConfined, bool dynamicOwnership); +void virSecurityManagerPreFork(virSecurityManagerPtr mgr); +void virSecurityManagerPostFork(virSecurityManagerPtr mgr); + void *virSecurityManagerGetPrivateData(virSecurityManagerPtr mgr); const char *virSecurityManagerGetDriver(virSecurityManagerPtr mgr); -- 1.8.1

3 2

[libvirt] [PATCH] Check if classes are derived from object
by Guido Günther 12 Feb '13

12 Feb '13

This makes sure we don't regress to old style classes --- Just a minor addition that came up while verifying if the corresponding Debian bug is fixed. python/sanitytest.py | 27 ++++++++++++++++----------- 1 file changed, 16 insertions(+), 11 deletions(-) diff --git a/python/sanitytest.py b/python/sanitytest.py index 047450b..ace6792 100644 --- a/python/sanitytest.py +++ b/python/sanitytest.py @@ -7,17 +7,22 @@ globals = dir(libvirt) # Sanity test that the generator hasn't gone wrong # Look for core classes -assert("virConnect" in globals) -assert("virDomain" in globals) -assert("virDomainSnapshot" in globals) -assert("virInterface" in globals) -assert("virNWFilter" in globals) -assert("virNodeDevice" in globals) -assert("virNetwork" in globals) -assert("virSecret" in globals) -assert("virStoragePool" in globals) -assert("virStorageVol" in globals) -assert("virStream" in globals) +for clsname in ["virConnect", + "virDomain", + "virDomainSnapshot", + "virInterface", + "virNWFilter", + "virNodeDevice", + "virNetwork", + "virSecret", + "virStoragePool", + "virStorageVol", + "virStream", + ]: + assert(clsname in globals) + assert(object in getattr(libvirt, clsname).__bases__) + +# Constants assert("VIR_CONNECT_RO" in globals) # Error related bits -- 1.7.10.4

2 2

[libvirt] [PATCH] hypervisor: Restore pm initialization
by John Ferlan 12 Feb '13

12 Feb '13

Adjustment for 'c059cdeaf' due to older compiler complaint about pm not being initialized even though the j&7 == 0 does the trick. --- src/xen/xen_hypervisor.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/src/xen/xen_hypervisor.c b/src/xen/xen_hypervisor.c index 767fc0c..9b7dd2e 100644 --- a/src/xen/xen_hypervisor.c +++ b/src/xen/xen_hypervisor.c @@ -1773,7 +1773,7 @@ virXen_setvcpumap(int handle, ret = -1; } else { cpumap_t xen_cpumap; /* limited to 64 CPUs in old hypervisors */ - uint64_t *pm; + uint64_t *pm = &xen_cpumap; int j; if ((maplen > (int)sizeof(cpumap_t)) || (sizeof(cpumap_t) & 7)) -- 1.7.11.7

2 2

[libvirt] [PATCH V5 0/3] Add support for QEMU file descriptor sets
by Stefan Berger 12 Feb '13

12 Feb '13

The following patch series adds initial support for QEMU file descriptor sets implementing support for creating the proper command line. Some devices are using the sets now. Regards, Stefan

1 3

[libvirt] [PATCH] Remove re-entrant API call in SELinux/AppArmor security managers
by Daniel P. Berrange 11 Feb '13

11 Feb '13

From: "Daniel P. Berrange" <berrange(a)redhat.com> The security manager drivers are not allowed to call back out to top level security manager APIs, since that results in recursive mutex acquisition and thus deadlock. Remove calls to virSecurityManagerGetModel from SELinux / AppArmor drivers Signed-off-by: Daniel P. Berrange <berrange(a)redhat.com> --- src/security/security_apparmor.c | 4 ++-- src/security/security_selinux.c | 20 ++++++++++---------- 2 files changed, 12 insertions(+), 12 deletions(-) diff --git a/src/security/security_apparmor.c b/src/security/security_apparmor.c index bf795b0..f281555 100644 --- a/src/security/security_apparmor.c +++ b/src/security/security_apparmor.c @@ -603,12 +603,12 @@ AppArmorSetSecurityProcessLabel(virSecurityManagerPtr mgr, virDomainDefPtr def) if ((profile_name = get_profile_name(def)) == NULL) return rc; - if (STRNEQ(virSecurityManagerGetModel(mgr), secdef->model)) { + if (STRNEQ(SECURITY_APPARMOR_NAME, secdef->model)) { virReportError(VIR_ERR_INTERNAL_ERROR, _("security label driver mismatch: " "\'%s\' model configured for domain, but " "hypervisor driver is \'%s\'."), - secdef->model, virSecurityManagerGetModel(mgr)); + secdef->model, SECURITY_APPARMOR_NAME); if (use_apparmor() > 0) goto clean; } diff --git a/src/security/security_selinux.c b/src/security/security_selinux.c index 2f5012d..cfb99a3 100644 --- a/src/security/security_selinux.c +++ b/src/security/security_selinux.c @@ -1803,12 +1803,12 @@ virSecuritySELinuxSecurityVerify(virSecurityManagerPtr mgr ATTRIBUTE_UNUSED, if (secdef == NULL) return -1; - if (!STREQ(virSecurityManagerGetModel(mgr), secdef->model)) { + if (!STREQ(SECURITY_SELINUX_NAME, secdef->model)) { virReportError(VIR_ERR_INTERNAL_ERROR, _("security label driver mismatch: " "'%s' model configured for domain, but " "hypervisor driver is '%s'."), - secdef->model, virSecurityManagerGetModel(mgr)); + secdef->model, SECURITY_SELINUX_NAME); return -1; } @@ -1837,12 +1837,12 @@ virSecuritySELinuxSetSecurityProcessLabel(virSecurityManagerPtr mgr, return 0; VIR_DEBUG("label=%s", secdef->label); - if (!STREQ(virSecurityManagerGetModel(mgr), secdef->model)) { + if (!STREQ(SECURITY_SELINUX_NAME, secdef->model)) { virReportError(VIR_ERR_INTERNAL_ERROR, _("security label driver mismatch: " "'%s' model configured for domain, but " "hypervisor driver is '%s'."), - secdef->model, virSecurityManagerGetModel(mgr)); + secdef->model, SECURITY_SELINUX_NAME); if (security_getenforce() == 1) return -1; } @@ -1875,12 +1875,12 @@ virSecuritySELinuxSetSecurityDaemonSocketLabel(virSecurityManagerPtr mgr, if (secdef->label == NULL) return 0; - if (!STREQ(virSecurityManagerGetModel(mgr), secdef->model)) { + if (!STREQ(SECURITY_SELINUX_NAME, secdef->model)) { virReportError(VIR_ERR_INTERNAL_ERROR, _("security label driver mismatch: " "'%s' model configured for domain, but " "hypervisor driver is '%s'."), - secdef->model, virSecurityManagerGetModel(mgr)); + secdef->model, SECURITY_SELINUX_NAME); goto done; } @@ -1925,12 +1925,12 @@ virSecuritySELinuxSetSecuritySocketLabel(virSecurityManagerPtr mgr, if (secdef->label == NULL) return 0; - if (!STREQ(virSecurityManagerGetModel(mgr), secdef->model)) { + if (!STREQ(SECURITY_SELINUX_NAME, secdef->model)) { virReportError(VIR_ERR_INTERNAL_ERROR, _("security label driver mismatch: " "'%s' model configured for domain, but " "hypervisor driver is '%s'."), - secdef->model, virSecurityManagerGetModel(mgr)); + secdef->model, SECURITY_SELINUX_NAME); goto done; } @@ -1966,12 +1966,12 @@ virSecuritySELinuxClearSecuritySocketLabel(virSecurityManagerPtr mgr, if (secdef->label == NULL) return 0; - if (!STREQ(virSecurityManagerGetModel(mgr), secdef->model)) { + if (!STREQ(SECURITY_SELINUX_NAME, secdef->model)) { virReportError(VIR_ERR_INTERNAL_ERROR, _("security label driver mismatch: " "'%s' model configured for domain, but " "hypervisor driver is '%s'."), - secdef->model, virSecurityManagerGetModel(mgr)); + secdef->model, SECURITY_SELINUX_NAME); if (security_getenforce() == 1) return -1; } -- 1.7.11.7

3 2

[libvirt] [PATCH 0/4 v4] New API virNodeDeviceLookupSCSIHostByWWN
by Osier Yang 11 Feb '13

11 Feb '13

v3 - v4: * Only rebasing v2 - v3: * Validate the specified wwnn,wwpn pair before applying it to the new API in virsh-nodedev.c v1 - v2: * Per Daniel's suggestion, change the API name from virNodeDeviceLookupByWWN into virNodeDeviceLookupSCSIHostByWWN. Osier Yang (4): Introduce API virNodeDeviceLookupSCSIHostByWWN remote: Wire up the remote protocol nodedev: Implement virNodeDeviceLookupSCSIHostByWWN virsh: Use virNodeDeviceLookupSCSIHostByWWN include/libvirt/libvirt.h.in | 5 ++ src/driver.h | 6 ++ src/libvirt.c | 46 ++++++++++++++++ src/libvirt_public.syms | 5 ++ src/node_device/node_device_driver.c | 13 +++-- src/node_device/node_device_driver.h | 4 ++ src/node_device/node_device_hal.c | 1 + src/node_device/node_device_udev.c | 1 + src/remote/remote_driver.c | 1 + src/remote/remote_protocol.x | 13 ++++- src/remote_protocol-structs | 9 +++ src/rpc/gendispatch.pl | 5 ++- tools/virsh-nodedev.c | 98 ++++++++++++++++++++++++++-------- tools/virsh.pod | 15 +++-- 14 files changed, 187 insertions(+), 35 deletions(-) -- 1.7.7.6

3 17

[libvirt] libvirtd (from git) no longer responds to 'kill'
by Richard W.M. Jones 11 Feb '13

11 Feb '13

It used to be that you could kill a session libvirtd using eg: killall libvirtd lt-libvirtd However with upstream libvirt from git today, this no longer appears to work: $ ps ax | grep libvirtd 4240 ? Ssl 0:05 /usr/sbin/libvirtd 18492 ? Sl 0:00 /home/rjones/d/libvirt/daemon/.libs/lt-libvirtd --timeout=30 18775 pts/10 S+ 0:00 grep --color=auto libvirtd $ killall lt-libvirtd $ ps ax | grep libvirtd 4240 ? Ssl 0:05 /usr/sbin/libvirtd 18492 ? Sl 0:00 /home/rjones/d/libvirt/daemon/.libs/lt-libvirtd --timeout=30 18785 pts/10 S+ 0:00 grep --color=auto libvirtd $ killall lt-libvirtd $ ps ax | grep libvirtd 4240 ? Ssl 0:05 /usr/sbin/libvirtd 18492 ? Sl 0:00 /home/rjones/d/libvirt/daemon/.libs/lt-libvirtd --timeout=30 18799 pts/10 S+ 0:00 grep --color=auto libvirtd Is this new brokenness, or was this never meant to work in the first place? BTW this libvirtd process is pretty persistent. I sent it a whole variety of signals, and only 'kill -9' worked. Rich. -- Richard Jones, Virtualization Group, Red Hat http://people.redhat.com/~rjones libguestfs lets you edit virtual machines. Supports shell scripting, bindings from many languages. http://libguestfs.org

2 2

[libvirt] [PATCH v2 0/4] net: support set source address(es) and ports for NAT
by Natanael Copa 11 Feb '13

11 Feb '13

This is a rework of the previous sent 'set public ip for nat' patches [1]. Changes v2: - Use separate attributes for addresses and ports as suggested by Laine. - support set port range without setting public ip [1] http://www.redhat.com/archives/libvir-list/2012-December/msg00140.html Natanael Copa (4): util: refactor iptables command construction into multiple steps net: support set public ip for forward mode nat net: support a public address range for forward mode nat net: add support for specifying port range for forward mode nat docs/formatnetwork.html.in | 37 ++++++++ src/conf/network_conf.c | 206 ++++++++++++++++++++++++++++++++++++++++++-- src/conf/network_conf.h | 4 + src/network/bridge_driver.c | 32 +++++++ src/util/viriptables.c | 195 +++++++++++++++++++++++++---------------- src/util/viriptables.h | 8 ++ 6 files changed, 402 insertions(+), 80 deletions(-) -- 1.8.1.2

3 10

[libvirt] [PATCH v3 0/2] net: support set public ip for forward mode nat
by Natanael Copa 11 Feb '13

11 Feb '13

Rebased patch[1]. Changes v3: - remove support for <nat address='1.2.3.4'/> format, the 2/4 patch[2]. [1] http://www.redhat.com/archives/libvir-list/2013-February/msg00088.html [2] http://www.redhat.com/archives/libvir-list/2013-February/msg00090.html Natanael Copa (2): net: support set public ip range for forward mode nat net: add support for specifying port range for forward mode nat docs/formatnetwork.html.in | 32 ++++++++ src/conf/network_conf.c | 189 ++++++++++++++++++++++++++++++++++++++++++-- src/conf/network_conf.h | 4 + src/network/bridge_driver.c | 32 ++++++++ src/util/viriptables.c | 77 +++++++++++++++--- src/util/viriptables.h | 8 ++ 6 files changed, 328 insertions(+), 14 deletions(-) -- 1.8.1.2

1 2

[libvirt] [PATCH 0/2] virsh: Two small fixes for "snapshot-create-as"
by Peter Krempa 11 Feb '13

11 Feb '13

Peter Krempa (2): virsh-snapshot: Refactor some details in virsh snapshot-create-as virsh-snapshot: Reject --no-metadata together with --print-xml tools/virsh-snapshot.c | 27 +++++++++++++++------------ 1 file changed, 15 insertions(+), 12 deletions(-) -- 1.8.1.1

2 5

[libvirt] libvirt and Parallels Cloud Storage
by Alexander Gordeev 11 Feb '13

11 Feb '13

Hi! I'd like to develop libvirt integration with Parallels Cloud Storage (Pstorage) which is a new distributed file system from Parallels designed specifically for storing VM images: http://www.parallels.com/products/pcs/cloud-storage/ It's not a POSIX FS but there is a FUSE client for it that can be used to access and manipulate images. It's quite high speed but only when used with O_DIRECT + aio. I tried to setup several KVMs on top of a Pstorage mount using virt-manager. It worked good, but I had to: 1. tune cache and IO settings for every disk 2. disable space allocation by libvirt because it is using sync IO and is therefore slow I tried to find ways to solve the first issue and IMHO this can be done by adding a way to specify per-pool defaults for cache and IO settings. I didn't find any way for this in the current code and UI. I'd like to add a new storage backend also that will be a 'dir' backend with extra ability to manage Pstorage mount points and UI integration. I'd like to merge my work to the main tree when it's finished if possible. What do you think about my plan? Can you please suggest me how to add per-pool defaults the best way? -- Alexander

3 7

[libvirt] [PATCH] qemu: Fix potential crash when attaching/detaching cdrom or floppy
by Osier Yang 11 Feb '13

11 Feb '13

The crash could happen if the disk source is empty for cdrom or floppy disk. --- src/qemu/qemu_driver.c | 7 +++++-- src/qemu/qemu_process.c | 3 +++ 2 files changed, 8 insertions(+), 2 deletions(-) diff --git a/src/qemu/qemu_driver.c b/src/qemu/qemu_driver.c index 812bf95..906501b 100644 --- a/src/qemu/qemu_driver.c +++ b/src/qemu/qemu_driver.c @@ -5884,7 +5884,9 @@ qemuDomainAttachDeviceDiskLive(virConnectPtr conn, } if (ret == 0) { - if (disk->type == VIR_DOMAIN_DISK_TYPE_BLOCK && disk->shared) { + if (disk->type == VIR_DOMAIN_DISK_TYPE_BLOCK && + disk->shared && + disk->src) { if (qemuAddSharedDisk(driver->sharedDisks, disk->src) < 0) VIR_WARN("Failed to add disk '%s' to shared disk table", disk->src); @@ -6011,7 +6013,8 @@ qemuDomainDetachDeviceDiskLive(virQEMUDriverPtr driver, if (ret == 0 && disk->type == VIR_DOMAIN_DISK_TYPE_BLOCK && - disk->shared) { + disk->shared && + disk->src) { if (qemuRemoveSharedDisk(driver->sharedDisks, disk->src) < 0) VIR_WARN("Failed to remove disk '%s' from shared disk table", disk->src); diff --git a/src/qemu/qemu_process.c b/src/qemu/qemu_process.c index d1872c0..98ed552 100644 --- a/src/qemu/qemu_process.c +++ b/src/qemu/qemu_process.c @@ -3458,6 +3458,9 @@ qemuCheckSharedDisk(virHashTablePtr sharedDisks, char *key = NULL; int ret = 0; + if (!disk->src) + return 0; + if (!(key = qemuGetSharedDiskKey(disk->src))) return -1; -- 1.7.7.6

2 2

[libvirt] [PATCH V4 0/3] Add support for QEMU file descriptor sets
by Stefan Berger 11 Feb '13

11 Feb '13

The following patch series adds initial support for QEMU file descriptor sets implementing support for creating the proper command line. Some devices are using the sets now. Regards, Stefan

2 6

[libvirt] Adding new standalone API
by Bilal Ahmad 11 Feb '13

11 Feb '13

Hi all, I am learning to hack libvirt and so far for basic things, I am able to get things done. For the advanced things like adding a completely new API, I will need some help from you people. For starters, I have decided to add a custom network API with only "name" and "UUID" and get my custom network API to work. I would really appreciate if you could guide me. I have gone through the “*Implementing a new API in Libvirt*” but its an example which introduces new APIs to an existing API type. To learn better, I plan to implement my own standalone sample API and get it to work. I would really appreciate help from this open-source community. Thanks, Bilal

2 1

[libvirt] [PATCH 1/5] Public API to allow defining new domain using OVA file
by Ata E Husain Bohra 11 Feb '13

11 Feb '13

Appends a new API to libvirt public driver that supports defining a new domain using OVA format. API expects following inputs: 1. connection pointer. 2. path to OVA package (single file format). 3. Storage pool name where new domain needs to be created. API returns pointers to newly created domain that is not powered ON. Libvirt driver sanitizes the inputs and calls hypervisor specific driver callbacks to abstract OVA install implementation. --- include/libvirt/libvirt.h.in | 3 +++ src/driver.h | 5 ++++ src/libvirt.c | 58 ++++++++++++++++++++++++++++++++++++++++++ src/libvirt_public.syms | 5 ++++ 4 files changed, 71 insertions(+) diff --git a/include/libvirt/libvirt.h.in b/include/libvirt/libvirt.h.in index 09c89c5..ea342bc 100644 --- a/include/libvirt/libvirt.h.in +++ b/include/libvirt/libvirt.h.in @@ -1919,6 +1919,9 @@ int virDomainMemoryPeek (virDomainPtr dom, */ virDomainPtr virDomainDefineXML (virConnectPtr conn, const char *xml); +virDomainPtr virDomainDefineOVA (virConnectPtr conn, + const char *ovapath, + const char *poolname); int virDomainUndefine (virDomainPtr domain); typedef enum { diff --git a/src/driver.h b/src/driver.h index 01c95cf..08382a0 100644 --- a/src/driver.h +++ b/src/driver.h @@ -274,6 +274,10 @@ typedef int typedef virDomainPtr (*virDrvDomainDefineXML) (virConnectPtr conn, const char *xml); +typedef virDomainPtr + (*virDrvDomainDefineOVA) (virConnectPtr conn, + const char *ovapath, + const char *poolname); typedef int (*virDrvDomainUndefine) (virDomainPtr dom); typedef int @@ -1008,6 +1012,7 @@ struct _virDriver { virDrvDomainCreate domainCreate; virDrvDomainCreateWithFlags domainCreateWithFlags; virDrvDomainDefineXML domainDefineXML; + virDrvDomainDefineOVA domainDefineOVA; virDrvDomainUndefine domainUndefine; virDrvDomainUndefineFlags domainUndefineFlags; virDrvDomainAttachDevice domainAttachDevice; diff --git a/src/libvirt.c b/src/libvirt.c index 6d1da12..bdc21b4 100644 --- a/src/libvirt.c +++ b/src/libvirt.c @@ -1988,6 +1988,8 @@ error: return NULL; } + + /** * virDomainCreateLinux: * @conn: pointer to the hypervisor connection @@ -8334,6 +8336,62 @@ error: } /** + * virDomainDefineOVA: + * @conn: pointer to the hypervisor connection + * @ovapath: OVA file path + * @poolname: storage pool where domain needs to be created + * + * OVA stands for Open Virtualuzation Archive and consist of + * three important components: + * 1. XML descriptor: defines the virtual machine hardware description, + * network connectivity etc. + * 2. Virtual disk (may be compressed or optimized). + * 3. Manifest file. + * + * Define a new domain but does not start it. + * This function may require privileged access to the hypervisor. + * + * Returns a new domain object or NULL in case of failure + */ +virDomainPtr +virDomainDefineOVA(virConnectPtr conn, const char *ovapath, + const char *poolname) +{ + VIR_DEBUG("conn=%p, ovaPath=%s, poolname=%s", + conn, ovapath, poolname); + + virResetLastError(); + + if (!VIR_IS_CONNECT(conn)) { + virLibConnError(VIR_ERR_INVALID_CONN, __FUNCTION__); + virDispatchError(NULL); + return NULL; + } + if (conn->flags & VIR_CONNECT_RO) { + virLibConnError(VIR_ERR_OPERATION_DENIED, __FUNCTION__); + goto error; + } + virCheckNonNullArgGoto(ovapath, error); + virCheckNonNullArgGoto(poolname, error); + + if (conn->driver->domainDefineOVA) { + virDomainPtr ret; + ret = conn->driver->domainDefineOVA( + conn, ovapath, poolname); + if (!ret) + goto error; + return ret; + } + + virLibConnError(VIR_ERR_NO_SUPPORT, __FUNCTION__); + +error: + virDispatchError(conn); + return NULL; + +} + +/** * virDomainCreate: * @domain: pointer to a defined domain * diff --git a/src/libvirt_public.syms b/src/libvirt_public.syms index 2107519..1af48c9 100644 --- a/src/libvirt_public.syms +++ b/src/libvirt_public.syms @@ -585,4 +585,9 @@ LIBVIRT_1.0.2 { virDomainOpenChannel; } LIBVIRT_1.0.1; +LIBVIRT_1.0.3 { + global: + virDomainDefineOVA; +} LIBVIRT_1.0.2; + # .... define new API here using predicted next version number .... -- 1.7.9.5

3 3

[libvirt] Environment variable CLASSPATH not set!
by harryxiyou 09 Feb '13

09 Feb '13

Hi all, We are developing a hdfs-based File system, which is HLFS( http://code.google.com/p/cloudxy/wiki/WHAT_IS_CLOUDXY) Now, we have developed HLFS driver for Libvirt.You can find our patch for Libvirt network disk (Online storage) here http://cloudxy.googlecode.com/svn/branches/hlfs/person/harry/hlfs/patches/h… But when i boot a VM from a base linux OS, which the OS have been installed into our HLFS block device at first. However, it(HDFS or JVM) says i have not set the CLASSPATH like following. [...] uri:hdfs:///tmp/testenv/testfs,head:hdfs,dir:/tmp/testenv,fsname:testfs,hostname:default,port:0,user:kanghua Environment variable CLASSPATH not set! ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ fs is null, hdfsConnect error! [...] Actually, i have set CLASSPATH in ~/.bashrc like following. I have installed CDH3u2 for developing. I can do other hdfs jobs successfully. $ cat /home/jiawei/.bashrc [...] export HLFS_HOME=/home/jiawei/workshop3/hlfs export LOG_HOME=$HLFS_HOME/3part/log export SNAPPY_HOME=$HLFS_HOME/3part/snappy export HADOOP_HOME=$HLFS_HOME/3part/hadoop export JAVA_HOME=/usr/lib/jvm/java-6-sun export PATH=/usr/bin/:/usr/local/bin/:/bin/:/usr/sbin/:/sbin/:$JAVA_HOME/bin/ #export LD_LIBRARY_PATH=$JAVAHOME/lib export LD_LIBRARY_PATH=$JAVA_HOME/jre/lib/i386/server/:$HADOOP_HOME/lib32/:$LOG_HOME/lib32/:$SNAPPY_HOME/lib32/:$HLFS_HOME/output/lib32/:/usr/lib/ export PKG_CONFIG_PATH=/usr/lib/pkgconfig/:/usr/share/pkgconfig/ export CFLAGS="-L/usr/lib -L/lib -L/usr/lib64" export CXXFLAGS="-L/usr/lib -L/lib" export CLASSPATH=.:$JAVA_HOME/lib/dt.jar:$JAVA_HOME/lib/htmlconverter.jar:$JAVA_HOME/lib/jconsole.jar:$JAVA_HOME/lib/jconsole.jar:$JAVA_HOME/lib/tools.jar:$JAVA_HOME/jre/lib/charsets.jar:$JAVA_HOME/jre/lib/deploy.jar:$JAVA_HOME/jre/lib/javaws.jar:$JAVA_HOME/jre/lib/jce.jar:$JAVA_HOME/jre/lib/jsse.jar:$JAVA_HOME/jre/lib/management-agent.jar:$JAVA_HOME/jre/lib/plugin.jar:$JAVA_HOME/jre/lib/resources.jar:$JAVA_HOME/jre/lib/rt.jar:$JAVA_HOME/jre/lib/:$JAVA_HOME/lib/:/usr/lib/hadoop-0.20/conf:/usr/lib/jvm/java-6-sun/lib/tools.jar:/usr/lib/hadoop-0.20:/usr/lib/hadoop-0.20/hadoop-core-0.20.2-cdh3u2.jar:/usr/lib/hadoop-0.20/lib/ant-contrib-1.0b3.jar:/usr/lib/hadoop-0.20/lib/aspectjrt-1.6.5.jar:/usr/lib/hadoop-0.20/lib/aspectjtools-1.6.5.jar:/usr/lib/hadoop-0.20/lib/commons-cli-1.2.jar:/usr/lib/hadoop-0.20/lib/commons-codec-1.4.jar:/usr/lib/hadoop-0.20/lib/commons-daemon-1.0.1.jar:/usr/lib/hadoop-0.20/lib/commons-el-1.0.jar:/usr/lib/hadoop-0.20/lib/commons-httpclient-3.1.jar:/usr/lib/hadoop-0.20/lib/commons-logging-1.0.4.jar:/usr/lib/hadoop-0.20/lib/commons-logging-api-1.0.4.jar:/usr/lib/hadoop-0.20/lib/commons-net-1.4.1.jar:/usr/lib/hadoop-0.20/lib/core-3.1.1.jar:/usr/lib/hadoop-0.20/lib/hadoop-fairscheduler-0.20.2-cdh3u2.jar:/usr/lib/hadoop-0.20/lib/hsqldb-1.8.0.10.jar:/usr/lib/hadoop-0.20/lib/jackson-core-asl-1.5.2.jar:/usr/lib/hadoop-0.20/lib/jackson-mapper-asl-1.5.2.jar:/usr/lib/hadoop-0.20/lib/jasper-compiler-5.5.12.jar:/usr/lib/hadoop-0.20/lib/jasper-runtime-5.5.12.jar:/usr/lib/hadoop-0.20/lib/jets3t-0.6.1.jar:/usr/lib/hadoop-0.20/lib/jetty-6.1.26.cloudera.1.jar:/usr/lib/hadoop-0.20/lib/jetty-servlet-tester-6.1.26.cloudera.1.jar:/usr/lib/hadoop-0.20/lib/jetty-util-6.1.26.cloudera.1.jar:/usr/lib/hadoop-0.20/lib/jsch-0.1.42.jar:/usr/lib/hadoop-0.20/lib/junit-4.5.jar:/usr/lib/hadoop-0.20/lib/kfs-0.2.2.jar:/usr/lib/hadoop-0.20/lib/log4j-1.2.15.jar:/usr/lib/hadoop-0.20/lib/mockito-all-1.8.2.jar:/usr/lib/hadoop-0.20/lib/oro-2.0.8.jar:/usr/lib/hadoop-0.20/lib/servlet-api-2.5-20081211.jar:/usr/lib/hadoop-0.20/lib/servlet-api-2.5-6.1.14.jar:/usr/lib/hadoop-0.20/lib/slf4j-api-1.4.3.jar:/usr/lib/hadoop-0.20/lib/slf4j-log4j12-1.4.3.jar:/usr/lib/hadoop-0.20/lib/xmlenc-0.52.jar:/usr/lib/hadoop-0.20/lib/jsp-2.1/jsp-2.1.jar:/usr/lib/hadoop-0.20/lib/jsp-2.1/jsp-api-2.1.jar I have reproted this matter as an issue to handle, which you can see http://code.google.com/p/cloudxy/issues/detail?id=37 for details. I have seen libvirt source codes for this problem, i find it may lead to by following problems. 1, Some changes made by gnulib, which i find gnulib has some files and funcs have done some jobs for CLASSPATH like csharpexec.c, javacomp.c, javaexec.c, etc. 2, After i compile libvirt source codes, i get virsh script, which has a variable named 'relink_command' that has been set some libraries and paths permanently. 3, Maybe some other reasons. Could anyone give me some suggestions? How could i solve this problem? Thanks a lot in advance ;-) -- Thanks Harry Wei

1 0

[libvirt] [PATCH 0/4] fix yet another regression due to my backingChain rewrite
by Eric Blake 09 Feb '13

09 Feb '13

https://bugzilla.redhat.com/show_bug.cgi?id=903248 details a regression that I caused, which is tickled by the odd way in which VDSM uses (abuses?) symlinks to qcow2 files with relative backing file names. The first two patches are quite trivial; the third a bit longer but still straightforward; and the meat of the series is the fourth patch. Meanwhile, I know of at least one other regression that I still need to fix in the same area of code: https://bugzilla.redhat.com/show_bug.cgi?id=896685 is caused when cgroups are not present (such as in qemu:///session); I hope to patch that one tomorrow. Eric Blake (4): storage: factor out large integer reads storage: rearrange functions storage: refactor metadata lookup storage: don't follow backing chain symlinks too eagerly src/util/virstoragefile.c | 475 +++++++++++++++++++++++----------------------- src/util/virstoragefile.h | 1 + 2 files changed, 240 insertions(+), 236 deletions(-) -- 1.8.1

2 11

[libvirt] [PATCH] fix a bug of sheepdog storage driver
by harryxiyou＠gmail.com 09 Feb '13

09 Feb '13

Don't try to refresh Sheepdog volume if creating volume fails. Signed-off-by: Harry Wei <harryxiyou(a)gmail.com> --- src/storage/storage_backend_sheepdog.c | 10 +++++++--- 1 files changed, 7 insertions(+), 3 deletions(-) diff --git a/src/storage/storage_backend_sheepdog.c b/src/storage/storage_backend_sheepdog.c index cd18f33..218284d 100644 --- a/src/storage/storage_backend_sheepdog.c +++ b/src/storage/storage_backend_sheepdog.c @@ -156,7 +156,7 @@ virStorageBackendSheepdogCreateVol(virConnectPtr conn ATTRIBUTE_UNUSED, virStorageVolDefPtr vol) { - int ret; + int ret = -1; if (vol->target.encryption != NULL) { virReportError(VIR_ERR_CONFIG_UNSUPPORTED, "%s", @@ -167,10 +167,14 @@ virStorageBackendSheepdogCreateVol(virConnectPtr conn ATTRIBUTE_UNUSED, virCommandPtr cmd = virCommandNewArgList(COLLIE, "vdi", "create", vol->name, NULL); virCommandAddArgFormat(cmd, "%llu", vol->capacity); virStorageBackendSheepdogAddHostArg(cmd, pool); - ret = virCommandRun(cmd, NULL); + if (virCommandRun(cmd, NULL) < 0) + goto cleanup; - virStorageBackendSheepdogRefreshVol(conn, pool, vol); + if (virStorageBackendSheepdogRefreshVol(conn, pool, vol) < 0) + goto cleanup; + ret = 0; +cleanup: virCommandFree(cmd); return ret; } -- 1.7.0.4

4 3

[libvirt] [PATCH 00/15] Permit setting capabilities for uid!=0 processes
by Laine Stump 09 Feb '13

09 Feb '13

There are a bunch of patches here, but each is small and single-purpose to make reviewing easier (and also so that any potential regressions can be easily bisected). The original purpose of the patches was to permit setting CAP_COMPROMISE_KERNEL for non-root qemu processes, since Fedora 18 now requires that in order for generic PCI passthrough to work (the alternative was to always run qemu as root). Although we may not actually want to do that part (if we can convince kernel people to implement CAP_COMPROMISE_KERNEL such that it's only required when *opening* the necessary sysfs file (done by libvirt), rather than for every read/write (done by qemu), then we will not need CAP_COMPROMISE_KERNEL for qemu), but that is just a couple lines in the final patch, and the rest of the series is still useful, as it make dropping/keeping caps truly work for non-root child processes - this has never before been the case. (for example, CAP_SYS_RAWIO is needed for generic scsi passthrough to work, and until now the only way to have that was to run *all* qemus as root). A bit higher level description of what I've done with all the patches: 1) remove the programmable "hook" from virExecWithHook(), since that function was only called from one place, and always with the same hook function. Rename virExecWithHook() to virExec(), and replace the call to that hook with inline code. 2) give virCommand an API to set the intended uid/gid of the command that's going to be run, and use that instead of a "user hook" where appropriate (in the process completely eliminating two hook functions). 3) Also add an API to virCommand to do the final "set the process label" step for selinux/apparmor. 4) Add a new API to the security driver (and use it from qemu) called virSecurityManagerSetChildProcessLabel() which a) is called prior to virCommandRun() rather than from a command "hook" function, b) takes a virCommand, and c) rather than immediately performing the operation (as virSecurityManagerSetProcessLabel() did), merely stores the necessary information in the virCommand so that virExec can perform the operation (setting selinux label, setuid/gid, etc) 5) make a new function combining the setting of uid/gid and maintaining of capabilities, because that is the only way you can set uid!=0 and still maintain capabilities. Use this in virExec() 6) *Finally* set the CAP_COMPROMISE_KERNEL capability unconditionally for all qemu processes. (If we really do have to do this, we may want to consider making it a qemu.conf setting). Laine Stump (15): util: eliminate generic hook from virExecWithHook util: eliminate extra args from virExec util: refactor virCommandHook into virExec and virCommandHandshakeChild util: add virCommandSetUID and virCommandSetGID util: make virSetUIDGID a NOP when uid or gid is -1 qemu: replace exec hook with virCommandSetUID/GID in qemuCaps* qemu: replace exec hook with virCommandSetUID/GID in storage_backend build: define SECDRIVER_LIBS in Makefile.am util: add security label setting to virCommand security: add new virSecurityManagerSetChildProcessLabel API qemu: let virCommand set child process security labels/uid/gid util: drop capabilities immediately after changing uid/gid of child util: virSetUIDGIDWithCaps - change uid while keeping caps util: maintain caps when running command with uid != 0 qemu: set CAP_COMPROMISE_KERNEL so that pci passthrough works src/Makefile.am | 35 ++- src/libvirt_private.syms | 5 + src/qemu/qemu_capabilities.c | 61 ++--- src/qemu/qemu_process.c | 23 +- src/security/security_apparmor.c | 41 +++- src/security/security_dac.c | 24 +- src/security/security_driver.h | 6 +- src/security/security_manager.c | 13 +- src/security/security_manager.h | 6 +- src/security/security_nop.c | 10 +- src/security/security_selinux.c | 34 ++- src/security/security_stack.c | 20 +- src/storage/storage_backend.c | 28 +-- src/util/vircommand.c | 481 ++++++++++++++++++--------------------- src/util/vircommand.h | 9 +- src/util/virutil.c | 113 ++++++++- src/util/virutil.h | 1 + 17 files changed, 543 insertions(+), 367 deletions(-) -- 1.8.1

3 36

[libvirt] [PATCH] qemu: support vhost-net for generic ethernet devices
by Laine Stump 09 Feb '13

09 Feb '13

>From qemu's point of view these are still just tap devices, so there's no reason they shouldn't work with vhost-net; as a matter of fact, Raja Sivaramakrishnan <srajag00(a)yahoo.com> verified on libvir-list that at least the qemu_command.c part of this patch works: https://www.redhat.com/archives/libvir-list/2012-December/msg01314.html (the hotplug case is extrapolation on my part). --- src/qemu/qemu_command.c | 1 + src/qemu/qemu_hotplug.c | 3 +++ 2 files changed, 4 insertions(+) diff --git a/src/qemu/qemu_command.c b/src/qemu/qemu_command.c index 33f2ecd..6c28123 100644 --- a/src/qemu/qemu_command.c +++ b/src/qemu/qemu_command.c @@ -6159,6 +6159,7 @@ qemuBuildCommandLine(virConnectPtr conn, if (actualType == VIR_DOMAIN_NET_TYPE_NETWORK || actualType == VIR_DOMAIN_NET_TYPE_BRIDGE || + actualType == VIR_DOMAIN_NET_TYPE_ETHERNET || actualType == VIR_DOMAIN_NET_TYPE_DIRECT) { /* Attempt to use vhost-net mode for these types of network device */ diff --git a/src/qemu/qemu_hotplug.c b/src/qemu/qemu_hotplug.c index 4504f0b..0c28a6a 100644 --- a/src/qemu/qemu_hotplug.c +++ b/src/qemu/qemu_hotplug.c @@ -754,6 +754,9 @@ int qemuDomainAttachNetDevice(virConnectPtr conn, iface_connected = true; if (qemuOpenVhostNet(vm->def, net, priv->qemuCaps, &vhostfd) < 0) goto cleanup; + } else if (actualType == VIR_DOMAIN_NET_TYPE_ETHERNET) { + if (qemuOpenVhostNet(vm->def, net, priv->qemuCaps, &vhostfd) < 0) + goto cleanup; } if (virQEMUCapsGet(priv->qemuCaps, QEMU_CAPS_NET_NAME) || -- 1.7.11.7

2 2

[libvirt] [PATCH] network_conf.c: Free xmlDoc after use
by Michal Privoznik 09 Feb '13

09 Feb '13

The virNetworkObjUpdateParseFile() function was not freeing the xml variable, leaving us with a memory leak. --- src/conf/network_conf.c | 1 + 1 file changed, 1 insertion(+) diff --git a/src/conf/network_conf.c b/src/conf/network_conf.c index c93916d..3604ff7 100644 --- a/src/conf/network_conf.c +++ b/src/conf/network_conf.c @@ -1895,6 +1895,7 @@ virNetworkObjUpdateParseFile(const char *filename, ret = 0; cleanup: + xmlFreeDoc(xml); xmlXPathFreeContext(ctxt); return ret; } -- 1.8.0.2

3 3

[libvirt] [PATCH] virsh-snapshot: Fix xpath query to determine snapshot state
by Peter Krempa 08 Feb '13

08 Feb '13

The query didn't match the external state correctly for offline internal snapshots. --- tools/virsh-snapshot.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/tools/virsh-snapshot.c b/tools/virsh-snapshot.c index ba57059..66776e2 100644 --- a/tools/virsh-snapshot.c +++ b/tools/virsh-snapshot.c @@ -983,8 +983,8 @@ cmdSnapshotInfo(vshControl *ctl, const vshCmd *cmd) * external snapshot. */ switch (virXPathBoolean("boolean(/domainsnapshot/memory)", ctxt)) { case 1: - external = virXPathBoolean("boolean(/domainsnapshot/memory/@snapshot=external " - "| /domainsnapshot/disks/disk/@snapshot=external)", + external = virXPathBoolean("boolean(/domainsnapshot/memory[@snapshot='external'] " + "| /domainsnapshot/disks/disk[@snapshot='external'])", ctxt); break; case 0: -- 1.8.1.1

2 2

[libvirt] [PATCH V3 0/2] Add support for QEMU file descriptor sets
by Stefan Berger 08 Feb '13

08 Feb '13

The following patch series adds initial support for QEMU file descriptor sets implementing support for creating the proper command line. Some devices are using the sets now. Regards, Stefan

3 11

[libvirt] Why not add snapshot operations for Sheepdog storage driver in Libvirt
by harryxiyou 08 Feb '13

08 Feb '13

Hi all, Virsh has following operations about snapshot. Snapshot (help keyword 'snapshot') snapshot-create Create a snapshot from XML snapshot-create-as Create a snapshot from a set of args snapshot-current Get or set the current snapshot snapshot-delete Delete a domain snapshot snapshot-dumpxml Dump XML for a domain snapshot snapshot-edit edit XML for a snapshot snapshot-info snapshot information snapshot-list List snapshots for a domain snapshot-parent Get the name of the parent of a snapshot snapshot-revert Revert a domain to a snapshot But i cannot find relevant realization in Sheepdog storage driver in Libvirt. I am sure Sheepdog storage driver, now, cannot support snapshot operations in Libvirt. I wonder whether Sheepdog can realize snapshot operations in Libvirt? Maybe i can finish these jobs. I am not clear about how to let Sheepdog storage driver support snapshot operations. Could anyone give me some suggestions? Thanks in advance. -- Thanks Harry Wei

2 1