- 17 12月, 2019 11 次提交
-
-
由 Daniel Henrique Barboza 提交于
Move QEMU caps validation of qemuBuildHostdevCommandLine() to qemuDomainDeviceDefValidateHostdev() and qemuDomainMdevDefValidate(), allowing them to be validated at domain define time. Tests were adapted to consider the new caps being needed in this earlier stage. Reviewed-by: NCole Robinson <crobinso@redhat.com> Signed-off-by: NDaniel Henrique Barboza <danielhb413@gmail.com>
-
由 Daniel Henrique Barboza 提交于
Move QEMU caps validation of QEMU_CAPS_CHARDEV_FILE_APPEND and QEMU_CAPS_CHARDEV_LOGFILE to qemuDomainChrSourceDefValidate(). Reviewed-by: NCole Robinson <crobinso@redhat.com> Signed-off-by: NDaniel Henrique Barboza <danielhb413@gmail.com>
-
由 Daniel Henrique Barboza 提交于
Move QEMU caps validation of QEMU_CAPS_USB_HUB to a new function in qemu_domain.c, qemuDomainDeviceDefValidateHub(). This function is called by qemuDomainDeviceDefValidate() to validate the sound device in domain define time. qemuxml2xmltest.c was adjusted to add the now required caps for domain definition. Reviewed-by: NCole Robinson <crobinso@redhat.com> Signed-off-by: NDaniel Henrique Barboza <danielhb413@gmail.com>
-
由 Daniel Henrique Barboza 提交于
A new function qemuDomainDeviceDefValidateNVRAM() was created to validate the NVRAM in domain define time. Unit test was adjusted to account for the extra QEMU_CAPS_DEVICE_NVRAM required during domain define. Reviewed-by: NCole Robinson <crobinso@redhat.com> Signed-off-by: NDaniel Henrique Barboza <danielhb413@gmail.com>
-
由 Daniel Henrique Barboza 提交于
A new qemuDomainDefValidateNuma() function was created to host all the QEMU caps validation being done inside qemuBuildNumaArgStr(). This new function is called by qemuDomainValidateCpuCount() to allow NUMA validation in domain define time. Tests were changed to account for the QEMU capabilities that need to be present at domain define time. Reviewed-by: NCole Robinson <crobinso@redhat.com> Signed-off-by: NDaniel Henrique Barboza <danielhb413@gmail.com>
-
由 Daniel Henrique Barboza 提交于
Next patch will validate QEMU_CAPS_NUMA_DIST in a new qemu_domain.c function. The code to verify if a NUMA node distance is being set will still be needed in qemuBuildNumaArgStr() though. To avoid code repetition, let's put this logic in a helper to be used in qemuBuildNumaArgStr() and in the new function. Reviewed-by: NCole Robinson <crobinso@redhat.com> Signed-off-by: NDaniel Henrique Barboza <danielhb413@gmail.com>
-
由 Daniel Henrique Barboza 提交于
Validation of MACHINE_KERNEL_IRQCHIP and MACHINE_KERNEL_IRQCHIP_SPLIT QEMU caps are now being done in qemuDomainDefValidateFeatures(). Reviewed-by: NCole Robinson <crobinso@redhat.com> Signed-off-by: NDaniel Henrique Barboza <danielhb413@gmail.com>
-
由 Daniel Henrique Barboza 提交于
virQEMUCapsSupportsVmport() is now being called inside qemuDomainDefValidateFeatures() for VIR_DOMAIN_FEATURE_VMPORT feature. qemuxml2xmltest.c was changed to account for this caps being now validated at domain define time. Reviewed-by: NCole Robinson <crobinso@redhat.com> Signed-off-by: NDaniel Henrique Barboza <danielhb413@gmail.com>
-
由 Daniel Henrique Barboza 提交于
Move QEMU_CAPS_MEM_MERGE validation from qemuBuildMachineCommandLine() to qemuDomainDefValidateMemory(). Reviewed-by: NCole Robinson <crobinso@redhat.com> Signed-off-by: NDaniel Henrique Barboza <danielhb413@gmail.com>
-
由 Daniel Henrique Barboza 提交于
Introduce a new function called qemuDomainDefValidatePSeriesFeature() that will center all the PSeries validation done in qemu_command.c. qemuDomainDefValidatePSeriesFeature() is then called during domain define time, in qemuDomainDefValidateFeatures(). qemuxml2argvtest.c is also changed to include all the caps that now are being validated in define time. Reviewed-by: NCole Robinson <crobinso@redhat.com> Signed-off-by: NDaniel Henrique Barboza <danielhb413@gmail.com>
-
由 Daniel P. Berrangé 提交于
Signed-off-by: NDaniel P. Berrangé <berrange@redhat.com>
-
- 16 12月, 2019 4 次提交
-
-
由 Ani Sinha 提交于
Qemu commit e900135dcfb67 ("i386: Add CPUID bit for CLZERO and XSAVEERPTR") adds support for CLZERO CPUID bit. This commit extends support for this CPUID bit into libvirt. Signed-off-by: NAni Sinha <ani.sinha@nutanix.com> Message-Id: <1575371352-99055-1-git-send-email-ani.sinha@nutanix.com> Reviewed-by: NJiri Denemark <jdenemar@redhat.com>
-
由 Christian Ehrhardt 提交于
These days libvirt is pretty reliable and even remote connections (not the default for libvirt-guests anyway) either work or fail but are uncommon to be flaky. On the other hand users might have disabled the service and while we are After=libvirtd for ordering we are not Requiring it. Adding that or any harder dependency might break our ordering. But if people have disabled libvirt they will do a full retry loop until timeout. Lets drop the loop to be much faster if a remote is not reachable. Fixes: https://bugs.launchpad.net/ubuntu/+source/libvirt/+bug/1854653 This reverts commit 4e7fc830 Author: Michal Prívozník <mprivozn@redhat.com> Date: Fri Feb 21 12:46:08 2014 +0100 libvirt-guests: Wait for libvirtd to initialize The race described in that commit no longer exists using systemd as we now have socket activation. If not using systemd, then it is also safe if using the libvirtd --daemon flag, since the parent process won't return to the caller until the child is accepting connections. Reported-by: NDoug Smythies <dsmythies@telus.net> Reviewed-by: NDaniel P. Berrangé <berrange@redhat.com> Signed-off-by: NChristian Ehrhardt <christian.ehrhardt@canonical.com>
-
由 Michal Privoznik 提交于
Signed-off-by: NMichal Privoznik <mprivozn@redhat.com> Reviewed-by: NDaniel P. Berrangé <berrange@redhat.com>
-
由 Michal Privoznik 提交于
There is plenty of distributions that haven't switched to systemd nor they force their users to (Gentoo, Alpine Linux to name a few). With the daemon split merged their only option is to still use the monolithic daemon which will go away eventually. Provide init scripts for these distros too. For now, I'm not introducing config files which would correspond to the init files except for libvirtd and virtproxyd init scripts where it might be desirable to tweak the command line of corresponding daemons. Signed-off-by: NMichal Privoznik <mprivozn@redhat.com> Reviewed-by: NDaniel P. Berrangé <berrange@redhat.com>
-
- 14 12月, 2019 6 次提交
-
-
由 Han Han 提交于
Reviewed-by: NCole Robinson <crobinso@redhat.com> Signed-off-by: NHan Han <hhan@redhat.com>
-
由 Marc Hartmayer 提交于
To free the structs and save the error, it is not necessary to hold @priv->lock, therefore move these parts after the mutex unlock. Reviewed-by: NCole Robinson <crobinso@redhat.com> Signed-off-by: NMarc Hartmayer <mhartmay@linux.ibm.com>
-
由 Marc Hartmayer 提交于
This patch introduces virNetServerGetProgramLocked. It's a function to determine which program has to be used for a given @msg. This function will be reused in the next patch. Signed-off-by: NMarc Hartmayer <mhartmay@linux.ibm.com> Reviewed-by: NPavel Hrdina <phrdina@redhat.com>
-
由 Marc Hartmayer 提交于
As a result, you can later determine during the callback which program was used. This makes it easier to refactor the code in the future and is less prone to error. Signed-off-by: NMarc Hartmayer <mhartmay@linux.ibm.com> Reviewed-by: NPavel Hrdina <phrdina@redhat.com>
-
由 Marc Hartmayer 提交于
Use the return value of virObjectRef directly. This way, it's easier for another reader to identify the reason why the additional reference is required. Signed-off-by: NMarc Hartmayer <mhartmay@linux.ibm.com> Reviewed-by: NJohn Ferlan <jferlan@redhat.com> Reviewed-by: NPavel Hrdina <phrdina@redhat.com>
-
由 Daniel P. Berrangé 提交于
Signed-off-by: NDaniel P. Berrangé <berrange@redhat.com>
-
- 13 12月, 2019 19 次提交
-
-
由 Huaqiang 提交于
Following domain configuration changes create two memory bandwidth monitors: one is monitoring the bandwidth consumed by vCPU 0, another is for vCPU 5. ``` <cputune> <memorytune vcpus='0-4'> <node id='0' bandwidth='20'/> <node id='1' bandwidth='30'/> + <monitor vcpus='0'/> </memorytune> + <memorytune vcpus='5'> + <monitor vcpus='5'/> + </memorytune> </cputune> ``` Reviewed-by: NDaniel P. Berrangé <berrange@redhat.com> Signed-off-by: NHuaqiang <huaqiang.wang@intel.com>
-
由 Huaqiang 提交于
Originally, inside <cputune/cachetune>, it requires the <cache> element to be in the position before <monitor>, and following configuration is not permitted by schema, but it is better to let it be valid. <cputune> <cachetune vcpus='0-1'> <monitor level='3' vcpus='0-1'/> ^ |__ Not permitted originally because it is in the place before <cache> element. <cache id='0' level='3' type='both' size='3' unit='MiB'/> <cache id='1' level='3' type='both' size='3' unit='MiB'/> </cachetune> ... </cputune> And, let schema do more strict check by identifying following configuration to be invalid, due to <cachetune> should contain at least one <cache> or <monitor> element. <cputune> <cachetune vcpus='0-1'> ^ |__ a <cachetune> SHOULD contain at least one <cache> or <monitor> </cachetune> ... </cputune> Reviewed-by: NDaniel P. Berrangé <berrange@redhat.com> Signed-off-by: NHuaqiang <huaqiang.wang@intel.com>
-
由 Huaqiang 提交于
We learned that the hardware features of CAT, CMT, MBA and MBM are orthogonal ones, if CAT or MBA is not supported in system, but CMT or MBM are supported, then the cache monitor or memoryBW monitor features may not be correctly displayed in host capabilities through command 'virsh capabilites'. Showing the cache/memoryBW monitor capabilities even there is no support of cache allocation or memoryBW allocation features. Reviewed-by: NDaniel P. Berrangé <berrange@redhat.com> Signed-off-by: NHuaqiang <huaqiang.wang@intel.com>
-
由 Ján Tomko 提交于
As of commit 2a00ef6e which was released in v5.2.0, we require YAJL to build the QEMU driver. Remove the checks from code that requires the QEMU driver or checks that also check for WITH_QEMU. Signed-off-by: NJán Tomko <jtomko@redhat.com> Reviewed-by: NMichal Privoznik <mprivozn@redhat.com>
-
由 Daniel P. Berrangé 提交于
Both the application developer guide and virsh command reference are unmaintained for best part of 8 years, and so horrifically out of date. This does not give a good impression to people reading the docs. Now that we are publishing the man pages online, those are a better doc to read for virsh. We can also highlight the API reference instead of the app dev guide. The virsh command reference & app dev guide will still exist on the web root, but will not be linked to. Reviewed-by: NJán Tomko <jtomko@redhat.com> Signed-off-by: NDaniel P. Berrangé <berrange@redhat.com>
-
由 Yingle Hou 提交于
Add Hygon Dhyana CPU data test case related files. Reviewed-by: NDaniel P. Berrangé <berrange@redhat.com> Signed-off-by: NYingle Hou <houyingle@hygon.cn>
-
由 Yingle Hou 提交于
Add Hygon Dhyana CPU model to the processor model. Reviewed-by: NDaniel P. Berrangé <berrange@redhat.com> Signed-off-by: NYingle Hou <houyingle@hygon.cn>
-
由 Yingle Hou 提交于
The x86ModelParseSignatures function makes an assumption that CPU signature model equals 0 as an invalid case. While in Hygon processor definition, A1 version (model 0, stepping 1) is mass production version, to support Hygon Dhyana A1 version, we have removed CPU signature model zero checking condition. Reviewed-by: NDaniel P. Berrangé <berrange@redhat.com> Signed-off-by: NYingle Hou <houyingle@hygon.cn>
-
由 Jiri Denemark 提交于
CVE-2019-11135 When TSX_CTRL bit of IA32_ARCH_CAPABILITIES MSR is set to 1, the CPU supports IA32_TSX_CTRL MSR which can be used to disable and/or mask TSX. Signed-off-by: NJiri Denemark <jdenemar@redhat.com> Reviewed-by: NJán Tomko <jtomko@redhat.com>
-
由 Jiri Denemark 提交于
CVE-2019-11135 CPUs with TAA_NO bit of IA32_ARCH_CAPABILITIES MSR set to 1 are not vulnerable to TSX Asynchronous Abort and passing this bit to a guest may avoid unnecessary mitigations. Signed-off-by: NJiri Denemark <jdenemar@redhat.com> Reviewed-by: NJán Tomko <jtomko@redhat.com>
-
由 Peter Krempa 提交于
Signed-off-by: NPeter Krempa <pkrempa@redhat.com> Reviewed-by: NEric Blake <eblake@redhat.com>
-
由 Peter Krempa 提交于
To allow backups work across external snapshots we need to improve the algorithm which calculates which bitmaps to merge. The algorithm must look for appropriately named bitmaps in the image and possibly descend into a backing image if the current image does not have the bitmap. Signed-off-by: NPeter Krempa <pkrempa@redhat.com> Reviewed-by: NEric Blake <eblake@redhat.com>
-
由 Peter Krempa 提交于
This function looks up a named bitmap for a virStorageSource in the data returned from query-named-block-nodes. Signed-off-by: NPeter Krempa <pkrempa@redhat.com> Reviewed-by: NEric Blake <eblake@redhat.com>
-
由 Peter Krempa 提交于
Add test code which will crawl a fake internal list of checkpoints and generate the list of bitmaps for merging to gather the final bitmap for the backup. The initial tests cover the basic case of all bitmaps being present in the top layer of the backing chain. Signed-off-by: NPeter Krempa <pkrempa@redhat.com> Reviewed-by: NEric Blake <eblake@redhat.com>
-
由 Peter Krempa 提交于
Signed-off-by: NPeter Krempa <pkrempa@redhat.com> Reviewed-by: NEric Blake <eblake@redhat.com>
-
由 Peter Krempa 提交于
The function will require the bitmap topology for the full implementation. To facilitate testing, add the propagation of the necessary data beforehand so that the test code can stay unchanged during the changes. Signed-off-by: NPeter Krempa <pkrempa@redhat.com> Reviewed-by: NEric Blake <eblake@redhat.com>
-
由 Peter Krempa 提交于
Separate the for now incomplete code that collects the bitmaps to be merged for an incremental backup into a separate function. This will allow adding testing prior to the improvement of the algorithm to include snapshots. Signed-off-by: NPeter Krempa <pkrempa@redhat.com> Reviewed-by: NEric Blake <eblake@redhat.com>
-
由 Peter Krempa 提交于
The object itself has no extra value and it would make testing the code harder. Refactor it to remove just the definition pointer. Signed-off-by: NPeter Krempa <pkrempa@redhat.com> Reviewed-by: NEric Blake <eblake@redhat.com>
-
由 Peter Krempa 提交于
Add test data gathered from a run of qemu after creating bitmaps and snapshots together in various combinations. The following sequence of commands was used to achieve the configuration: virsh checkpoint-create-as VM --name a virsh snapshot-create-as VM --disk-only virsh snapshot-create-as VM --disk-only virsh checkpoint-create-as VM --name b virsh checkpoint-create-as VM --name c virsh snapshot-create-as VM --disk-only virsh checkpoint-create-as VM --name d virsh snapshot-create-as VM --disk-only virsh checkpoint-create-as VM --name current Note that VM was restarted after these operations to allow renumbering of the bitmaps in a more human-readable way. Signed-off-by: NPeter Krempa <pkrempa@redhat.com> Reviewed-by: NEric Blake <eblake@redhat.com>
-