Bug 1723633
| Summary: | VM refuses to start: can't apply global IvyBridge-IBRS-x86_64-cpu.osxsave=on | ||
|---|---|---|---|
| Product: | Red Hat Enterprise Linux Advanced Virtualization | Reporter: | jiyan <jiyan> |
| Component: | libvirt | Assignee: | Jiri Denemark <jdenemar> |
| Status: | CLOSED ERRATA | QA Contact: | jiyan <jiyan> |
| Severity: | medium | Docs Contact: | |
| Priority: | medium | ||
| Version: | 8.1 | CC: | amit, berrange, cfergeau, crobinso, dwmw2, extras-qa, itamar, jdenemar, jsuchane, pbonzini, poc, rbalakri, rjones, virt-maint, xuzhang |
| Target Milestone: | rc | Keywords: | Reopened |
| Target Release: | 8.1 | Flags: | knoel:
mirror+
|
| Hardware: | x86_64 | ||
| OS: | Linux | ||
| Whiteboard: | |||
| Fixed In Version: | libvirt-5.5.0-1.el8 | Doc Type: | If docs needed, set a value |
| Doc Text: | Story Points: | --- | |
| Clone Of: | 1644848 | Environment: | |
| Last Closed: | 2019-11-06 07:17:15 UTC | Type: | Bug |
| Regression: | --- | Mount Type: | --- |
| Documentation: | --- | CRM: | |
| Verified Versions: | Category: | --- | |
| oVirt Team: | --- | RHEL 7.3 requirements from Atomic Host: | |
| Cloudforms Team: | --- | Target Upstream Version: | |
| Embargoed: | |||
| Bug Depends On: | 1644848 | ||
| Bug Blocks: | |||
|
Description
jiyan
2019-06-25 03:46:03 UTC
According to the following info, cloned this bug: commit 2900575db892700fab8a4b8541474d9bd3444a4a Author: Christian Ehrhardt <christian.ehrhardt> Date: Thu Apr 25 11:04:29 2019 +0200 qemu: do not define known no-op features Qemu dropped cpu features for osxsave and ospke [1][2]. The reason for the instant removal is that those features were never configurable as discussed in [3]. Fortunately the use cases adding those flags in the past are rare, but they exist. One that I identified are e.g. older virt-install when used with --cpu=host-model and there always could be the case of a user adding it to the guest xml. This triggers an issue like: qemu-system-x86_64: can't apply global Broadwell-noTSX-x86_64- cpu.osxsave=on: Property '.osxsave' not found Ensure that this does no more break spawning newer qemu versions by not rendering those features into the qemu command line. Fixes: https://bugs.launchpad.net/fedora/+source/qemu/+bug/1825195 Version:
kernel-4.18.0-83.el8.x86_64
qemu-kvm-4.0.0-0.module+el8.1.0+3169+3c501422.x86_64
libvirt-5.3.0-1.module+el8.1.0+3164+94495c71.x86_64
Info:
# lscpu |grep Flags
Flags: fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe syscall nx pdpe1gb rdtscp lm constant_tsc art arch_perfmon pebs bts rep_good nopl xtopology nonstop_tsc cpuid aperfmperf pni pclmulqdq dtes64 monitor ds_cpl vmx smx est tm2 ssse3 sdbg fma cx16 xtpr pdcm pcid dca sse4_1 sse4_2 x2apic movbe popcnt tsc_deadline_timer aes xsave avx f16c rdrand lahf_lm abm 3dnowprefetch cpuid_fault epb cat_l3 cdp_l3 invpcid_single pti ssbd mba ibrs ibpb stibp tpr_shadow vnmi flexpriority ept vpid fsgsbase tsc_adjust bmi1 hle avx2 smep bmi2 erms invpcid rtm cqm mpx rdt_a avx512f avx512dq rdseed adx smap clflushopt clwb intel_pt avx512cd avx512bw avx512vl xsaveopt xsavec xgetbv1 xsaves cqm_llc cqm_occup_llc cqm_mbm_total cqm_mbm_local dtherm ida arat pln pts pku ospke flush_l1d
# virsh domcapabilities |grep Ivy
<model usable='yes'>IvyBridge-IBRS</model>
<model usable='yes'>IvyBridge</model>
# virsh domstate test
shut off
# virsh dumpxml test |grep "<cpu" -A4
<cpu mode='custom' match='exact' check='partial'>
<model fallback='allow'>IvyBridge-IBRS</model>
<feature policy='disable' name='osxsave'/>
</cpu>
# virsh start test
error: Failed to start domain test
error: internal error: process exited while connecting to monitor: 2019-06-25T03:48:52.835394Z qemu-kvm: can't apply global IvyBridge-IBRS-x86_64-cpu.osxsave=off: Property '.osxsave' not found
# virsh domstate test
shut off
# virsh dumpxml test |grep "<cpu" -A4
<cpu mode='custom' match='exact' check='partial'>
<model fallback='allow'>IvyBridge-IBRS</model>
<feature policy='disable' name='ospke'/>
</cpu>
# virsh start test
error: Failed to start domain test
error: internal error: process exited while connecting to monitor: 2019-06-25T03:49:13.931562Z qemu-kvm: can't apply global IvyBridge-IBRS-x86_64-cpu.ospke=off: Property '.ospke' not found
This is already fixed upstream by
commit 2900575db892700fab8a4b8541474d9bd3444a4a
Refs: v5.3.0-77-g2900575db8
Author: Christian Ehrhardt <christian.ehrhardt>
AuthorDate: Thu Apr 25 11:04:29 2019 +0200
Commit: Christian Ehrhardt <christian.ehrhardt>
CommitDate: Wed May 15 09:32:52 2019 +0200
qemu: do not define known no-op features
Qemu dropped cpu features for osxsave and ospke [1][2].
The reason for the instant removal is that those features were never
configurable as discussed in [3].
Fortunately the use cases adding those flags in the past are rare, but
they exist. One that I identified are e.g. older virt-install when used
with --cpu=host-model and there always could be the case of a user
adding it to the guest xml.
This triggers an issue like:
qemu-system-x86_64: can't apply global Broadwell-noTSX-x86_64-
cpu.osxsave=on: Property '.osxsave' not found
Ensure that this does no more break spawning newer qemu versions by
not rendering those features into the qemu command line.
Fixes: https://bugs.launchpad.net/fedora/+source/qemu/+bug/1825195
Resolves: https://bugzilla.redhat.com/1644848
[1]: https://git.qemu.org/?p=qemu.git;a=commit;h=f1a2352
[2]: https://git.qemu.org/?p=qemu.git;a=commit;h=9ccb978
[3]: https://www.mail-archive.com/qemu-devel@nongnu.org/msg561877.html
Signed-off-by: Christian Ehrhardt <christian.ehrhardt>
Reviewed-by: Daniel Henrique Barboza <danielhb413>
Tested-by: Daniel Henrique Barboza <danielhb413>
and later updated with
commit b12865260a0f24ab86ddaf3547b2f2e2c595d429
Refs: v5.4.0-221-gb12865260a
Author: Jiri Denemark <jdenemar>
AuthorDate: Thu Jun 6 14:39:52 2019 +0200
Commit: Jiri Denemark <jdenemar>
CommitDate: Thu Jun 20 00:22:37 2019 +0200
qemu: Drop qemuFeatureNoEffect
We already have virQEMUCapsCPUFilterFeatures for filtering features
which QEMU does not know about. Let's move osxsave and ospke from
qemuFeatureNoEffect there.
Signed-off-by: Jiri Denemark <jdenemar>
Reviewed-by: Ján Tomko <jtomko>
commit 955fd6e7a2c8dc97e9e68d1bd9ba9d03c7d815b3
Refs: v5.4.0-222-g955fd6e7a2
Author: Jiri Denemark <jdenemar>
AuthorDate: Thu Jun 6 12:33:43 2019 +0200
Commit: Jiri Denemark <jdenemar>
CommitDate: Thu Jun 20 00:22:37 2019 +0200
qemu_process: Drop cleanup label from qemuProcessUpdateGuestCPU
Signed-off-by: Jiri Denemark <jdenemar>
Reviewed-by: Ján Tomko <jtomko>
commit c145b660b8225f73db16660461077ef931730939
Refs: v5.4.0-223-gc145b660b8
Author: Jiri Denemark <jdenemar>
AuthorDate: Fri Jun 7 14:07:10 2019 +0200
Commit: Jiri Denemark <jdenemar>
CommitDate: Thu Jun 20 00:22:37 2019 +0200
cpu_conf: Introduce virCPUDefFilterFeatures
This new internal API can be used for in place filtering of CPU features
in virCPUDef.
Signed-off-by: Jiri Denemark <jdenemar>
Reviewed-by: Ján Tomko <jtomko>
commit 0b763774a5e6eb87f109c86171631cebe012e2b3
Refs: v5.4.0-224-g0b763774a5
Author: Jiri Denemark <jdenemar>
AuthorDate: Thu Jun 6 14:51:14 2019 +0200
Commit: Jiri Denemark <jdenemar>
CommitDate: Thu Jun 20 00:22:37 2019 +0200
qemu: Filter CPU features in active XML
Properly filter features which should not be passed to QEMU because they
were never supported by QEMU or they did nothing and QEMU dropped them.
Currently they are just silently ignored by the command line generator.
Let's make this process more visible and clean by dropping the features
from the domain's active definition in qemuProcessUpdateGuestCPU.
Signed-off-by: Jiri Denemark <jdenemar>
Reviewed-by: Ján Tomko <jtomko>
Hi jiri
I tried to verify this bug, and found the following issue:
Version:
libvirt-5.5.0-1.module+el8.1.0+3580+d7f6488d.x86_64
qemu-kvm-4.0.0-5.module+el8.1.0+3622+5812d9bf.x86_64
kernel-4.18.0-112.el8.x86_64
Steps:
The following scenario is tested in physical host which supports ospke flag.
# virsh domstate vm1
shut off
# virsh dumpxml vm1 |grep "<cpu" -A3
<cpu mode='host-model' check='partial'>
<model fallback='allow'/>
<feature policy='disable' name='ospke'/>
</cpu>
# virsh start vm1
Domain vm1 started
# virsh dumpxml vm1 |grep "<cpu" -A17
<cpu mode='custom' match='exact' check='full'>
<model fallback='forbid'>Cascadelake-Server</model>
<vendor>Intel</vendor>
<feature policy='require' name='ss'/>
<feature policy='require' name='vmx'/>
<feature policy='require' name='hypervisor'/>
<feature policy='require' name='tsc_adjust'/>
<feature policy='require' name='umip'/>
<feature policy='require' name='pku'/>
<feature policy='require' name='md-clear'/>
<feature policy='require' name='stibp'/>
<feature policy='require' name='arch-capabilities'/>
<feature policy='require' name='xsaves'/>
<feature policy='disable' name='mpx'/>
</cpu>
# ps -ef |grep vm1
qemu 60266 1 86 21:44 ?
-cpu Cascadelake-Server,ss=on,vmx=on,hypervisor=on,tsc_adjust=on,umip=on,pku=on,md-clear=on,stibp=on,arch-capabilities=on,xsaves=on
# virsh console vm1
Connected to domain vm1
Escape character is ^]
Red Hat Enterprise Linux 8.1 Beta (Ootpa)
Kernel 4.18.0-107.el8.x86_64 on an x86_64
localhost login: root
Password:
# lscpu |grep ospke
Flags: fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ss syscall nx pdpe1gb rdtscp lm constant_tsc rep_good nopl xtopology cpuid pni pclmulqdq vmx ssse3 fma cx16 pcid sse4_1 sse4_2 x2apic movbe popcnt tsc_deadline_timer aes xsave avx f16c rdrand hypervisor lahf_lm abm 3dnowprefetch cpuid_fault invpcid_single pti ssbd ibrs ibpb stibp tpr_shadow vnmi flexpriority ept vpid fsgsbase tsc_adjust bmi1 hle avx2 smep bmi2 erms invpcid rtm avx512f avx512dq rdseed adx smap clflushopt clwb avx512cd avx512bw avx512vl xsaveopt xsavec xgetbv1 xsaves arat umip pku ospke
So my question is:
Even through I configure ospke/oxsave for inactive dumpxml of VM, the related conf will disappear after VM starts.
But this flag will be in guest OS, does that mean the related flags will be in guest OS when the physical host supports these two flags？
If not, How can I distinguish when the related flag will be in guest OS?
The flag disappears from domain XML because QEMU does not support it to be
enabled or disabled on the command line. That is, there's no way to tell QEMU
to explicitly enable or disable ospke.
The QEMU commit which removed ospke flag says the following:
OSPKE is not a static feature flag: it changes dynamically at
runtime depending on CR4, and it was never configurable: KVM
never returned OSPKE on GET_SUPPORTED_CPUID, and on TCG enables
it automatically if CR4_PKE_MASK is set.
Remove OSPKE from the feature name array so users don't try to
configure it manually.
I think everything works as expected here.
Verified this bug on libvirt-5.5.0-1.module+el8.1.0+3580+d7f6488d.x86_64.
Version:
libvirt-5.5.0-1.module+el8.1.0+3580+d7f6488d.x86_64
qemu-kvm-4.0.0-5.module+el8.1.0+3622+5812d9bf.x86_64
kernel-4.18.0-115.el8.x86_64
Steps:
1. Start a VM with the following conf; check dumpxml and qemu cmd line
# virsh domstate vm1
shut off
# virsh dumpxml vm1 --inactive |grep "<cpu" -A4
<cpu mode='custom' match='exact' check='partial'>
<model fallback='allow'>IvyBridge-IBRS</model>
<feature policy='disable' name='osxsave'/>
<feature policy='force' name='ospke'/>
</cpu>
# virsh start vm1
Domain vm1 started
# virsh dumpxml vm1 |grep "<cpu" -A17
<cpu mode='custom' match='exact' check='full'>
<model fallback='forbid'>IvyBridge-IBRS</model>
<feature policy='require' name='hypervisor'/>
<feature policy='require' name='arat'/>
<feature policy='require' name='xsaveopt'/>
</cpu>
# ps -ef |grep vm1
qemu 40850 1 47 02:24 ?
... -cpu IvyBridge-IBRS
2. Managedsave VM and then start VM; then check VM dumpxml again
# virsh managedsave vm1
Domain vm1 state saved by libvirt
# virsh domstate vm1
shut off
# virsh dumpxml vm1 |grep "<cpu" -A3
<cpu mode='custom' match='exact' check='partial'>
<model fallback='allow'>IvyBridge-IBRS</model>
<feature policy='disable' name='osxsave'/>
<feature policy='force' name='ospke'/>
</cpu>
# virsh start vm1
Domain vm1 started
# virsh dumpxml vm1 |grep "<cpu" -A17
<cpu mode='custom' match='exact' check='full'>
<model fallback='forbid'>IvyBridge-IBRS</model>
<feature policy='require' name='hypervisor'/>
<feature policy='require' name='arat'/>
<feature policy='require' name='xsaveopt'/>
</cpu>
# ps -ef |grep vm1
qemu 41150 1 5 02:26 ?
...-cpu IvyBridge-IBRS,hypervisor=on,arat=on,xsaveopt=on
3. Create snapshot for VM and check snapshot-dumpxml
# virsh snapshot-create-as vm1 --disk-only
Domain snapshot 1563258461 created
# virsh snapshot-list vm1
Name Creation Time State
---------------------------------------------------------
1563258461 2019-07-16 02:27:41 -0400 disk-snapshot
# virsh snapshot-dumpxml vm1 1563258461 |grep "<cpu" -A7
<cpu mode='custom' match='exact' check='partial'>
<model fallback='forbid'>IvyBridge-IBRS</model>
</cpu>
As the test steps above shows, libvirt drops these 2 kinds of cpu features during starting VM, managedsaving VM and snapshot-creating.
The test result is as expected, move this bug to be verified.
Since the problem described in this bug report should be resolved in a recent advisory, it has been closed with a resolution of ERRATA. For information on the advisory, and where to find the updated files, follow the link below. If the solution does not work for you, open a new bug report. https://access.redhat.com/errata/RHBA-2019:3723 |