- Devel - libvirt List Archives

[libvirt] [PATCH 00/10] update caps data for our tests
by Pavel Hrdina 24 May '16

24 May '16

To get replies from qemu the tests/qemucapsprobe tool was used with following QEMUs: qemu-1.2.2 and qemu-1.3.1 on Fedora18 qemu-1.4.2 and qemu-1.5.3 on Fedora19 qemu-1.6.0 on Fedora20 qemu-2.1.1 on Fedora21 qemu-2.4.0 and qemu-2.5.0 on Fedora23 qemu-2.6.0 on Fedora24 To get as much as possible dependencies I've used "yum-builddeb qemu", uninstalled "xen" packages and if it was missing installed "spice" packages. Pavel Hrdina (10): qemucapstest: update caps for qemu-1.2.2 qemucapstest: update caps for qemu-1.3.1 qemucapstest: update caps for qemu-1.4.2 qemucapstest: update caps for qemu-1.5.3 qemucapstest: update caps for qemu-1.6.0 qemucapstest: update caps for qemu-2.1.1 qemucapstest: update caps for qemu-2.4.0 qemucapstest: update caps for qemu-2.5.0 qemucapstest: update caps for qemu-2.6.0 qemucapstest: remove data for qemu-1.6.50 tests/domaincapsschemadata/qemu_1.6.0.x86_64.xml | 78 + tests/domaincapsschemadata/qemu_1.6.50.x86_64.xml | 78 - tests/domaincapsschemadata/qemu_2.6.0.x86_64.xml | 2 +- tests/domaincapstest.c | 2 +- .../qemucapabilitiesdata/caps_1.2.2.x86_64.replies | 3483 ++++++------- tests/qemucapabilitiesdata/caps_1.2.2.x86_64.xml | 2 - .../qemucapabilitiesdata/caps_1.3.1.x86_64.replies | 3943 ++++++++------- tests/qemucapabilitiesdata/caps_1.3.1.x86_64.xml | 5 +- .../qemucapabilitiesdata/caps_1.4.2.x86_64.replies | 4050 +++++++-------- tests/qemucapabilitiesdata/caps_1.4.2.x86_64.xml | 5 +- .../qemucapabilitiesdata/caps_1.5.3.x86_64.replies | 5245 ++++++++++---------- tests/qemucapabilitiesdata/caps_1.5.3.x86_64.xml | 4 + .../qemucapabilitiesdata/caps_1.6.0.x86_64.replies | 5244 +++++++++---------- tests/qemucapabilitiesdata/caps_1.6.0.x86_64.xml | 5 +- .../caps_1.6.50.x86_64.replies | 2848 ----------- tests/qemucapabilitiesdata/caps_1.6.50.x86_64.xml | 199 - .../qemucapabilitiesdata/caps_2.1.1.x86_64.replies | 665 +-- tests/qemucapabilitiesdata/caps_2.1.1.x86_64.xml | 2 - .../qemucapabilitiesdata/caps_2.4.0.x86_64.replies | 411 +- tests/qemucapabilitiesdata/caps_2.4.0.x86_64.xml | 1 + .../qemucapabilitiesdata/caps_2.5.0.x86_64.replies | 1587 +++--- tests/qemucapabilitiesdata/caps_2.5.0.x86_64.xml | 14 +- .../qemucapabilitiesdata/caps_2.6.0.x86_64.replies | 1877 ++++--- tests/qemucapabilitiesdata/caps_2.6.0.x86_64.xml | 33 +- tests/qemucapabilitiestest.c | 1 - 25 files changed, 13934 insertions(+), 15850 deletions(-) create mode 100644 tests/domaincapsschemadata/qemu_1.6.0.x86_64.xml delete mode 100644 tests/domaincapsschemadata/qemu_1.6.50.x86_64.xml delete mode 100644 tests/qemucapabilitiesdata/caps_1.6.50.x86_64.replies delete mode 100644 tests/qemucapabilitiesdata/caps_1.6.50.x86_64.xml -- 2.8.2

2 21

[libvirt] [PATCH] nwfilter: fix lock order deadlock
by Maxim Nestratov 24 May '16

24 May '16

Below is backtraces of two deadlocked threads: thread #1: virDomainConfVMNWFilterTeardown virNWFilterTeardownFilter lock updateMutex <------------ _virNWFilterTeardownFilter try to lock interface <---------- thread #2: learnIPAddressThread lock interface <------- virNWFilterInstantiateFilterLate try to lock updateMutex <---------- The problem is fixed by unlocking interface before calling virNWFilterInstantiateFilterLate to avoid updateMutex and interface ordering deadlocks. Otherwise we are going to instantiate the filter while holding interface lock, which will try to lock updateMutex, and if some other thread instantiating a filter in parallel is holding updateMutex and is trying to lock interface, both will deadlock. Also it is safe to unlock interface before virNWFilterInstantiateFilterLate because learnIPAddressThread stopped capturing packets and applied necessary rules on the interface, while instantiating a new filter doesn't require a locked interface. Signed-off-by: Maxim Nestratov <mnestratov(a)virtuozzo.com> --- src/nwfilter/nwfilter_learnipaddr.c | 12 +++++++++++- 1 file changed, 11 insertions(+), 1 deletion(-) diff --git a/src/nwfilter/nwfilter_learnipaddr.c b/src/nwfilter/nwfilter_learnipaddr.c index 1adbadb..cfd92d9 100644 --- a/src/nwfilter/nwfilter_learnipaddr.c +++ b/src/nwfilter/nwfilter_learnipaddr.c @@ -611,6 +611,16 @@ learnIPAddressThread(void *arg) sa.data.inet4.sin_addr.s_addr = vmaddr; char *inetaddr; + /* It is necessary to unlock interface here to avoid updateMutex and + * interface ordering deadlocks. Otherwise we are going to + * instantiate the filter, which will try to lock updateMutex, and + * some other thread instantiating a filter in parallel is holding + * updateMutex and is trying to lock interface, both will deadlock. + * Also it is safe to unlock interface here because we stopped + * capturing and applied necessary rules on the interface, while + * instantiating a new filter doesn't require a locked interface.*/ + virNWFilterUnlockIface(req->ifname); + if ((inetaddr = virSocketAddrFormat(&sa)) != NULL) { if (virNWFilterIPAddrMapAddIPAddr(req->ifname, inetaddr) < 0) { VIR_ERROR(_("Failed to add IP address %s to IP address " @@ -636,11 +646,11 @@ learnIPAddressThread(void *arg) req->ifname, req->ifindex); techdriver->applyDropAllRules(req->ifname); + virNWFilterUnlockIface(req->ifname); } VIR_DEBUG("pcap thread terminating for interface %s", req->ifname); - virNWFilterUnlockIface(req->ifname); err_no_lock: virNWFilterDeregisterLearnReq(req->ifindex); -- 2.4.3

5 10

[libvirt] [PATCH] maint: fix syntax-check sc_prohibit_int_ijk exclude rule
by Pavel Hrdina 24 May '16

24 May '16

This fixes the "include/" path, we need to create a regex for the whole path because we are matching the whole line. Signed-off-by: Pavel Hrdina <phrdina(a)redhat.com> --- I'm surprised that it even worked so far. I've noticed this after system update few days ago. cfg.mk | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/cfg.mk b/cfg.mk index c19f615..f688cbb 100644 --- a/cfg.mk +++ b/cfg.mk @@ -1259,7 +1259,7 @@ exclude_file_name_regexp--sc_prohibit_include_public_headers_brackets = \ ^(tools/|examples/|include/libvirt/(virterror|libvirt(-(admin|qemu|lxc))?)\.h$$) exclude_file_name_regexp--sc_prohibit_int_ijk = \ - ^(src/remote_protocol-structs|src/remote/remote_protocol.x|cfg.mk|include/|src/admin_protocol-structs|src/admin/admin_protocol.x)$ + ^(src/remote_protocol-structs|src/remote/remote_protocol.x|cfg.mk|include/libvirt/libvirt.+|src/admin_protocol-structs|src/admin/admin_protocol.x)$$ exclude_file_name_regexp--sc_prohibit_getenv = \ ^tests/.*\.[ch]$$ -- 2.8.3

2 2

[libvirt] [PATCH 0/2] some build fixes
by Pavel Hrdina 24 May '16

24 May '16

Pavel Hrdina (2): qemuxml2argvtest: skip test that depends on gnutls_cipher_encrypt() makefile: fix build on systems where gnutls is not in /usr/include src/Makefile.am | 2 +- tests/qemuxml2argvtest.c | 2 ++ 2 files changed, 3 insertions(+), 1 deletion(-) -- 2.8.2

2 7

[libvirt] [PATCH 0/2] Set PCI_PHYSICAL_FUNCTION flag more carefully
by Andrea Bolognani 24 May '16

24 May '16

If you're running libvirt master on a remote host, and connect to it using virt-manager, the remote daemon will crash. The same will happen if you run 'virsh nodedev-dumpxml' for a PCI device that happens *not* to be a virtual function. The first patch fixes the crash. The second one fixes something that could potentially cause trouble down the line, and also makes the code a tiny bit nicer. Andrea Bolognani (2): nodedev: sysfs: Set PCI_PHYSICAL_FUNCTION flag more carefully conf: nodedev: Set PCI_PHYSICAL_FUNCTION flag more carefully src/conf/node_device_conf.c | 4 ++-- src/node_device/node_device_linux_sysfs.c | 10 ++++++++-- 2 files changed, 10 insertions(+), 4 deletions(-) -- 2.5.5

2 4

Re: [libvirt] [PATCH V2 2/4] libxl: support hotplug USB host device
by Jim Fehlig 24 May '16

24 May '16

On 05/23/2016 12:51 AM, Chun Yan Liu wrote: > > Yes, I think it's OK. And another place needs to be updated, since now it's > not only pci device, but also could be usb device, so the error message > should be updated. > if (virDomainHostdevFind(vmdef, hostdev, &found) >= 0) { > - pcisrc = &hostdev->source.subsys.u.pci; > virReportError(VIR_ERR_OPERATION_FAILED, > - _("target pci device %.4x:%.2x:%.2x.%.1x\ > - already exists"), > - pcisrc->addr.domain, pcisrc->addr.bus, > - pcisrc->addr.slot, pcisrc->addr.function); > + _("device already exists in domain configuration")); > return -1; > } Nice catch. I've squashed in the below diff, changing the error code and message to match the qemu driver. Regards, Jim diff --git a/src/libxl/libxl_driver.c b/src/libxl/libxl_driver.c index 5b6f87a..4d7124d 100644 --- a/src/libxl/libxl_driver.c +++ b/src/libxl/libxl_driver.c @@ -3293,7 +3293,6 @@ libxlDomainAttachDeviceConfig(virDomainDefPtr vmdef, virDomainDeviceDefPtr dev) virDomainNetDefPtr net; virDomainHostdevDefPtr hostdev; virDomainHostdevDefPtr found; - virDomainHostdevSubsysPCIPtr pcisrc; char mac[VIR_MAC_STRING_BUFLEN]; switch (dev->type) { @@ -3336,12 +3335,8 @@ libxlDomainAttachDeviceConfig(virDomainDefPtr vmdef, virDomainDeviceDefPtr dev) } if (virDomainHostdevFind(vmdef, hostdev, &found) >= 0) { - pcisrc = &hostdev->source.subsys.u.pci; - virReportError(VIR_ERR_OPERATION_FAILED, - _("target pci device %.4x:%.2x:%.2x.%.1x\ - already exists"), - pcisrc->addr.domain, pcisrc->addr.bus, - pcisrc->addr.slot, pcisrc->addr.function); + virReportError(VIR_ERR_OPERATION_INVALID, "%s", + _("device is already in the domain configuration")); return -1; }

2 1

[libvirt] [PATCH v2 0/8] Replace VIR_ERROR with standard vir*Error in state driver init
by Jovanka Gulicoska 23 May '16

23 May '16

Replace VIR_ERROR logging macros for error reporting in driver startup routines with vir*Error function. Link to task: http://wiki.libvirt.org/page/BiteSizedTasks#Replace_VIR_ERROR_with_standard… Jovanka Gulicoska (8): uml: Replace VIR_ERROR with standard vir*Error in state driver init xen: Replace VIR_ERROR with standard vir*Error in state driver init bhyve: Replace VIR_ERROR with standard vir*Error in state driver init libxl: Replace VIR_ERROR with standard vir*Error in state driver init node_device: Replace VIR_ERROR with standard vir*Error in state driver init nwfilter: Replace VIR_ERROR with standard vir*Error in state driver init qemu: Replace VIR_ERROR with standard vir*Error in state driver init storage: Replace VIR_ERROR with standard vir*Error in state driver init src/bhyve/bhyve_driver.c | 4 +-- src/libxl/libxl_driver.c | 6 ++--- src/node_device/node_device_hal.c | 24 ++++++++++------- src/node_device/node_device_udev.c | 45 +++++++++++++++++++------------ src/nwfilter/nwfilter_driver.c | 5 ++-- src/qemu/qemu_driver.c | 55 ++++++++++++++++++++++++-------------- src/storage/storage_driver.c | 32 +++++++++++++--------- src/uml/uml_driver.c | 20 +++++++------- src/xen/xen_driver.c | 5 ++-- 9 files changed, 119 insertions(+), 77 deletions(-) -- 2.5.5

2 13

[libvirt] [PATCH] vz: add error code for case if vm is already stopped
by Mikhail Feoktistov 23 May '16

23 May '16

If try to stop VM or container which is already stopped than Virtuozzo 7 returns code PRL_ERR_INVALID_ACTION_REQUESTED. Error code PRL_ERR_DISP_VM_IS_NOT_STARTED is used in Virtuozzo 6 --- src/vz/vz_sdk.c | 1 + 1 file changed, 1 insertion(+) diff --git a/src/vz/vz_sdk.c b/src/vz/vz_sdk.c index a396d41..7eb78ca 100644 --- a/src/vz/vz_sdk.c +++ b/src/vz/vz_sdk.c @@ -2050,6 +2050,7 @@ prlsdkDomainChangeStateLocked(vzDriverPtr driver, switch (pret) { case PRL_ERR_DISP_VM_IS_NOT_STARTED: case PRL_ERR_DISP_VM_IS_NOT_STOPPED: + case PRL_ERR_INVALID_ACTION_REQUESTED: virerr = VIR_ERR_OPERATION_INVALID; break; default: -- 1.8.3.1

2 1

[libvirt] [PATCH] qemu: don't refuse to undefine a guest with NVRAM file
by Daniel P. Berrange 23 May '16

23 May '16

The undefine operation should always be allowed to succeed regardless of whether any NVRAM file exists. ie we should not force the application to use the VIR_DOMAIN_UNDEFINE_NVRAM flag. It is valid for the app to decide it wants the NVRAM file left on disk, in the same way that disk images are left on disk at undefine. --- src/qemu/qemu_driver.c | 20 +++++++------------- 1 file changed, 7 insertions(+), 13 deletions(-) diff --git a/src/qemu/qemu_driver.c b/src/qemu/qemu_driver.c index bec05d4..302bf48 100644 --- a/src/qemu/qemu_driver.c +++ b/src/qemu/qemu_driver.c @@ -6985,19 +6985,13 @@ qemuDomainUndefineFlags(virDomainPtr dom, if (!virDomainObjIsActive(vm) && vm->def->os.loader && vm->def->os.loader->nvram && - virFileExists(vm->def->os.loader->nvram)) { - if (!(flags & VIR_DOMAIN_UNDEFINE_NVRAM)) { - virReportError(VIR_ERR_OPERATION_INVALID, "%s", - _("cannot delete inactive domain with nvram")); - goto cleanup; - } - - if (unlink(vm->def->os.loader->nvram) < 0) { - virReportSystemError(errno, - _("failed to remove nvram: %s"), - vm->def->os.loader->nvram); - goto cleanup; - } + virFileExists(vm->def->os.loader->nvram) && + (flags & VIR_DOMAIN_UNDEFINE_NVRAM) && + (unlink(vm->def->os.loader->nvram) < 0)) { + virReportSystemError(errno, + _("failed to remove nvram: %s"), + vm->def->os.loader->nvram); + goto cleanup; } if (virDomainDeleteConfig(cfg->configDir, cfg->autostartDir, vm) < 0) -- 2.1.0

6 17

Re: [libvirt] [PATCH] Oracle VirtualBox 5 support or libvirt
by Eric Blake 23 May '16

23 May '16

[the moderation queue rejected your message for being too large] On 12/31/1969 05:00 PM, wrote: > --- > src/Makefile.am | 1 + > src/vbox/vbox_CAPI_v5_0.h | 25550 ++++++++++++++++++++++++++++++++++++++++ That's an awfully large patch, particularly if it is mostly generated. Can you resend, with the patch as a compressed attachment, so that it can make it through the moderation queue? Is it something that can logically be broken to smaller pieces, or is it really an all-or-none addition of the new huge file? > src/vbox/vbox_V5_0.c | 13 + > src/vbox/vbox_common.h | 2 + > src/vbox/vbox_storage.c | 2 + > src/vbox/vbox_tmpl.c | 290 +- > src/vbox/vbox_uniformed_api.h | 1 + > 7 files changed, 25752 insertions(+), 107 deletions(-) > create mode 100644 src/vbox/vbox_CAPI_v5_0.h > create mode 100644 src/vbox/vbox_V5_0.c > > diff --git a/src/Makefile.am b/src/Makefile.am > index 314f6df..4a7b85f 100644 > --- a/src/Makefile.am > +++ b/src/Makefile.am > @@ -783,6 +783,7 @@ VBOX_DRIVER_SOURCES = \ > vbox/vbox_V4_2_20.c vbox/vbox_CAPI_v4_2_20.h \ > vbox/vbox_V4_3.c vbox/vbox_CAPI_v4_3.h \ > vbox/vbox_V4_3_4.c vbox/vbox_CAPI_v4_3_4.h \ > + vbox/vbox_V5_0.c vbox/vbox_CAPI_v5_0.h \ > vbox/vbox_common.c vbox/vbox_common.h \ > vbox/vbox_uniformed_api.h \ > vbox/vbox_get_driver.h \ > diff --git a/src/vbox/vbox_CAPI_v5_0.h b/src/vbox/vbox_CAPI_v5_0.h > new file mode 100644 > index 0000000..b296733 > --- /dev/null > +++ b/src/vbox/vbox_CAPI_v5_0.h > @@ -0,0 +1,25550 @@ > +/* > + * Libvirt notice: this file is derived from the VirtualBox SDK, with > + * libvirt edits (fixing preprocessor indentation by cppi); do not > + * regenerate in the context of libvirt. > + */ > +/* > + * DO NOT EDIT! This is a generated file. > + * > + * Header file which provides C declarations for VirtualBox Main API > + * (COM interfaces), generated from XIDL (XML interface definition). > + * On Windows (which uses COM instead of XPCOM) the native C support > + * is used, and most of this file is not used. > + * > + * Source : src/VBox/Main/idl/VirtualBox.xidl > + * Generator : src/VBox/Main/cbinding/capiidl.xsl > + * > + * This file contains portions from the following Mozilla XPCOM files: > + * xpcom/include/xpcom/nsID.h > + * xpcom/include/nsIException.h > + * xpcom/include/nsprpub/prtypes.h > + * xpcom/include/xpcom/nsISupportsBase.h > + * > + * These files were originally triple-licensed (MPL/GPL2/LGPL2.1). Oracle > + * elects to distribute this derived work under the LGPL2.1 only. > + */ > + > +/* > + * Copyright (C) 2008-2015 Oracle Corporation > + * > + * This file is part of a free software library; you can redistribute > + * it and/or modify it under the terms of the GNU Lesser General > + * Public License version 2.1 as published by the Free Software > + * Foundation and shipped in the "COPYING.LESSER" file with this library. > + * The library is distributed in the hope that it will be useful, > + * but WITHOUT ANY WARRANTY of any kind. > + * > + * Oracle LGPL Disclaimer: For the avoidance of doubt, except that if > + * any license choice other than GPL or LGPL is available it will > + * apply instead, Oracle elects to use only the Lesser General Public > + * License version 2.1 (LGPLv2) at this time for any software where > + * a choice of LGPL license versions is made available with the > + * language indicating that LGPLv2 or any later version may be used, > + * or where a choice of which version of the LGPL is applied is > + * otherwise unspecified. > + */ Bummer that the license is restrictive against GPLv3, but no worse than any of the other files in the same directory, so not your problem. > height, bitsPerPixel, > xOrigin, yOrigin); > -#endif /* VBOX_API_VERSION >= 4003000 */ > +#else /*VBOX_API_VERSION >= 5000000 */ > + PRUint32 gms; > + > + return display->vtbl->GetScreenResolution(display, screenId, width, > + height, bitsPerPixel, > + xOrigin, yOrigin, &gms); > +#endif /* VBOX_API_VERSION >= 5000000 */ > + > } > > static nsresult > @@ -4357,10 +4433,10 @@ _displayTakeScreenShotPNGToArray(IDisplay *display ATTRIBUTE_UNUSED, > PRUint32 *screenDataSize ATTRIBUTE_UNUSED, > PRUint8** screenData ATTRIBUTE_UNUSED) > { > -#if VBOX_API_VERSION < 4000000 > +#if VBOX_API_VERSION < 4000000 || VBOX_API_VERSION >= 5000000 > vboxUnsupported(); > return 0; > -#else /* VBOX_API_VERSION >= 4000000 */ > +#else /* VBOX_API_VERSION >= 4000000 && VBOX_API_VERSION < 5000000 */ > return display->vtbl->TakeScreenShotPNGToArray(display, screenId, width, > height, screenDataSize, > screenData); > diff --git a/src/vbox/vbox_uniformed_api.h b/src/vbox/vbox_uniformed_api.h > index a469968..74e9ac0 100644 > --- a/src/vbox/vbox_uniformed_api.h > +++ b/src/vbox/vbox_uniformed_api.h > @@ -628,5 +628,6 @@ void vbox42InstallUniformedAPI(vboxUniformedAPI *pVBoxAPI); > void vbox42_20InstallUniformedAPI(vboxUniformedAPI *pVBoxAPI); > void vbox43InstallUniformedAPI(vboxUniformedAPI *pVBoxAPI); > void vbox43_4InstallUniformedAPI(vboxUniformedAPI *pVBoxAPI); > +void vbox50InstallUniformedAPI(vboxUniformedAPI *pVBoxAPI); > > #endif /* VBOX_UNIFORMED_API_H */ >

1 0