From: Liran Alon <liran.alon@oracle.com> To: "Dr. David Alan Gilbert" <dgilbert@redhat.com> Cc: ehabkost@redhat.com, kvm@vger.kernel.org, mtosatti@redhat.com, qemu-devel@nongnu.org, Paolo Bonzini <pbonzini@redhat.com>, rth@twiddle.net, jmattson@google.com Subject: Re: [QEMU PATCH v2 0/2]: KVM: i386: Add support for save and restore nested state Date: Tue, 13 Nov 2018 01:58:46 +0200 [thread overview] Message-ID: <B6AEEFF2-BAF1-439A-8E44-51352C9E05AE@oracle.com> (raw) In-Reply-To: <20181112165053.GF2293@work-vm> > On 12 Nov 2018, at 18:50, Dr. David Alan Gilbert <dgilbert@redhat.com> wrote: > > * Daniel P. Berrangé (berrange@redhat.com) wrote: >> On Sun, Nov 04, 2018 at 11:19:57PM +0100, Paolo Bonzini wrote: >>> On 02/11/2018 17:54, Daniel P. Berrangé wrote: >>>> We have usually followed a rule that new machine types must not >>>> affect runability of a VM on a host. IOW new machine types should >>>> not introduce dependancies on specific kernels, or hardware features >>>> such as CPU flags. >>> >>>> Anything that requires a new kernel feature thus ought to be an >>>> opt-in config tunable on the CLI, separate from machine type >>>> choice. >>> >>> Unless someone tinkered with the module parameters, they could not even >>> use nested virtualization before 4.20. So for everyone else, "-cpu >>> ...,+vmx" does count as an "opt-in config tunable on the CLI" that >>> requires 4.20. >>> >>> For those that did tinker with module parameters, we can grandfather in >>> the old machine types, so that they can use nested virtualization with >>> no live migration support. For those that did not, however, I don't >>> think it makes sense to say "oh by the way I really want to be able to >>> migrate this VM" on the command line, or even worse on the monitor. >> >> IIUC, 4.20 is only required from POV of migration state. Is it thus >> possible to just register a migration blocker if QEMU is launched >> on a host with kernel < 4.20. >> >> Migration has always been busted historically, so those people using >> nested VMX already won't be hurt by not having ability to live migrate >> their VM, but could otherwise continue using them without being forced >> to upgrade their kernel to fix a feature they're not even using. > > Yes, although I am a bit worried we might have a population of users > that: > a) Have enabled nesting > b) Run VMs with vmx enabled > c) Don't normally actually run nested guests > d) Currently happily migrate. > > Dave First of all, we should put the entire kvm_nested_state in a VMState subsection that has a .needed() method that checks if ((format==VMX) && (vmx.vmxon_pa != -1ull)); This will allow migrating a VM that has nested exposed but didn’t really enter VMX operation to still be able to successfully migrate to old hosts without any issues. The only problem remaining to be solved that you discuss here is what happens if user runs with modern QEMU (Including my to-be-written v3 patches discussed here) on a host that has kernel without KVM_CAP_NESTED_STATE. In this kernel, it is impossible for userspace (e.g. QEMU) to know if guest that was exposed with VMX actually used it. So for those cases, I see no option but to modify QEMU to also say that if guest is exposed with VMX and we are running on kernel without KVM_CAP_NESTED_STATE, then this is a migration blocker. -Liran > >> Regards, >> Daniel >> -- >> |: https://urldefense.proofpoint.com/v2/url?u=https-3A__berrange.com&d=DwIDAw&c=RoP1YumCXCgaWHvlZYR8PZh8Bv7qIrMUB65eapI_JnE&r=Jk6Q8nNzkQ6LJ6g42qARkg6ryIDGQr-yKXPNGZbpTx0&m=ximE6x4FrQ5vo3D4wF3w-cVsKgtNs85wCTL5GiP_B5Q&s=rs9ZkUUS37SHrs_oZJ9uIiXtpXvUkJBpfVEe9OSiQzk&e= -o- https://urldefense.proofpoint.com/v2/url?u=https-3A__www.flickr.com_photos_dberrange&d=DwIDAw&c=RoP1YumCXCgaWHvlZYR8PZh8Bv7qIrMUB65eapI_JnE&r=Jk6Q8nNzkQ6LJ6g42qARkg6ryIDGQr-yKXPNGZbpTx0&m=ximE6x4FrQ5vo3D4wF3w-cVsKgtNs85wCTL5GiP_B5Q&s=u_hnAFyY8BVwto0FsomTcZ3dmLKPYb1hwI_jRXI6EZg&e= :| >> |: https://urldefense.proofpoint.com/v2/url?u=https-3A__libvirt.org&d=DwIDAw&c=RoP1YumCXCgaWHvlZYR8PZh8Bv7qIrMUB65eapI_JnE&r=Jk6Q8nNzkQ6LJ6g42qARkg6ryIDGQr-yKXPNGZbpTx0&m=ximE6x4FrQ5vo3D4wF3w-cVsKgtNs85wCTL5GiP_B5Q&s=SEIw983ixpJUUySxUwrxtIbnjvHTB9ff3MaqULaulQw&e= -o- https://urldefense.proofpoint.com/v2/url?u=https-3A__fstop138.berrange.com&d=DwIDAw&c=RoP1YumCXCgaWHvlZYR8PZh8Bv7qIrMUB65eapI_JnE&r=Jk6Q8nNzkQ6LJ6g42qARkg6ryIDGQr-yKXPNGZbpTx0&m=ximE6x4FrQ5vo3D4wF3w-cVsKgtNs85wCTL5GiP_B5Q&s=yvxjeOrwjXjf08RBhPdX53lJN1W-8WSXT25ZeMSA06k&e= :| >> |: https://urldefense.proofpoint.com/v2/url?u=https-3A__entangle-2Dphoto.org&d=DwIDAw&c=RoP1YumCXCgaWHvlZYR8PZh8Bv7qIrMUB65eapI_JnE&r=Jk6Q8nNzkQ6LJ6g42qARkg6ryIDGQr-yKXPNGZbpTx0&m=ximE6x4FrQ5vo3D4wF3w-cVsKgtNs85wCTL5GiP_B5Q&s=9TIjtmf6AVFYWbyzI5vl-zXTaNCSCMAxyck92pc8yvY&e= -o- https://urldefense.proofpoint.com/v2/url?u=https-3A__www.instagram.com_dberrange&d=DwIDAw&c=RoP1YumCXCgaWHvlZYR8PZh8Bv7qIrMUB65eapI_JnE&r=Jk6Q8nNzkQ6LJ6g42qARkg6ryIDGQr-yKXPNGZbpTx0&m=ximE6x4FrQ5vo3D4wF3w-cVsKgtNs85wCTL5GiP_B5Q&s=Wapdtm0yT4j-9EjMoxwo9QvRZ3h9Fk_CvpQq8J4TDpg&e= :| > -- > Dr. David Alan Gilbert / dgilbert@redhat.com / Manchester, UK
WARNING: multiple messages have this Message-ID (diff)
From: Liran Alon <liran.alon@oracle.com> To: "Dr. David Alan Gilbert" <dgilbert@redhat.com> Cc: "\"Daniel P. Berrangé\"" <berrange@redhat.com>, "Paolo Bonzini" <pbonzini@redhat.com>, jmattson@google.com, ehabkost@redhat.com, kvm@vger.kernel.org, mtosatti@redhat.com, qemu-devel@nongnu.org, rth@twiddle.net Subject: Re: [Qemu-devel] [QEMU PATCH v2 0/2]: KVM: i386: Add support for save and restore nested state Date: Tue, 13 Nov 2018 01:58:46 +0200 [thread overview] Message-ID: <B6AEEFF2-BAF1-439A-8E44-51352C9E05AE@oracle.com> (raw) In-Reply-To: <20181112165053.GF2293@work-vm> > On 12 Nov 2018, at 18:50, Dr. David Alan Gilbert <dgilbert@redhat.com> wrote: > > * Daniel P. Berrangé (berrange@redhat.com) wrote: >> On Sun, Nov 04, 2018 at 11:19:57PM +0100, Paolo Bonzini wrote: >>> On 02/11/2018 17:54, Daniel P. Berrangé wrote: >>>> We have usually followed a rule that new machine types must not >>>> affect runability of a VM on a host. IOW new machine types should >>>> not introduce dependancies on specific kernels, or hardware features >>>> such as CPU flags. >>> >>>> Anything that requires a new kernel feature thus ought to be an >>>> opt-in config tunable on the CLI, separate from machine type >>>> choice. >>> >>> Unless someone tinkered with the module parameters, they could not even >>> use nested virtualization before 4.20. So for everyone else, "-cpu >>> ...,+vmx" does count as an "opt-in config tunable on the CLI" that >>> requires 4.20. >>> >>> For those that did tinker with module parameters, we can grandfather in >>> the old machine types, so that they can use nested virtualization with >>> no live migration support. For those that did not, however, I don't >>> think it makes sense to say "oh by the way I really want to be able to >>> migrate this VM" on the command line, or even worse on the monitor. >> >> IIUC, 4.20 is only required from POV of migration state. Is it thus >> possible to just register a migration blocker if QEMU is launched >> on a host with kernel < 4.20. >> >> Migration has always been busted historically, so those people using >> nested VMX already won't be hurt by not having ability to live migrate >> their VM, but could otherwise continue using them without being forced >> to upgrade their kernel to fix a feature they're not even using. > > Yes, although I am a bit worried we might have a population of users > that: > a) Have enabled nesting > b) Run VMs with vmx enabled > c) Don't normally actually run nested guests > d) Currently happily migrate. > > Dave First of all, we should put the entire kvm_nested_state in a VMState subsection that has a .needed() method that checks if ((format==VMX) && (vmx.vmxon_pa != -1ull)); This will allow migrating a VM that has nested exposed but didn’t really enter VMX operation to still be able to successfully migrate to old hosts without any issues. The only problem remaining to be solved that you discuss here is what happens if user runs with modern QEMU (Including my to-be-written v3 patches discussed here) on a host that has kernel without KVM_CAP_NESTED_STATE. In this kernel, it is impossible for userspace (e.g. QEMU) to know if guest that was exposed with VMX actually used it. So for those cases, I see no option but to modify QEMU to also say that if guest is exposed with VMX and we are running on kernel without KVM_CAP_NESTED_STATE, then this is a migration blocker. -Liran > >> Regards, >> Daniel >> -- >> |: https://urldefense.proofpoint.com/v2/url?u=https-3A__berrange.com&d=DwIDAw&c=RoP1YumCXCgaWHvlZYR8PZh8Bv7qIrMUB65eapI_JnE&r=Jk6Q8nNzkQ6LJ6g42qARkg6ryIDGQr-yKXPNGZbpTx0&m=ximE6x4FrQ5vo3D4wF3w-cVsKgtNs85wCTL5GiP_B5Q&s=rs9ZkUUS37SHrs_oZJ9uIiXtpXvUkJBpfVEe9OSiQzk&e= -o- https://urldefense.proofpoint.com/v2/url?u=https-3A__www.flickr.com_photos_dberrange&d=DwIDAw&c=RoP1YumCXCgaWHvlZYR8PZh8Bv7qIrMUB65eapI_JnE&r=Jk6Q8nNzkQ6LJ6g42qARkg6ryIDGQr-yKXPNGZbpTx0&m=ximE6x4FrQ5vo3D4wF3w-cVsKgtNs85wCTL5GiP_B5Q&s=u_hnAFyY8BVwto0FsomTcZ3dmLKPYb1hwI_jRXI6EZg&e= :| >> |: https://urldefense.proofpoint.com/v2/url?u=https-3A__libvirt.org&d=DwIDAw&c=RoP1YumCXCgaWHvlZYR8PZh8Bv7qIrMUB65eapI_JnE&r=Jk6Q8nNzkQ6LJ6g42qARkg6ryIDGQr-yKXPNGZbpTx0&m=ximE6x4FrQ5vo3D4wF3w-cVsKgtNs85wCTL5GiP_B5Q&s=SEIw983ixpJUUySxUwrxtIbnjvHTB9ff3MaqULaulQw&e= -o- https://urldefense.proofpoint.com/v2/url?u=https-3A__fstop138.berrange.com&d=DwIDAw&c=RoP1YumCXCgaWHvlZYR8PZh8Bv7qIrMUB65eapI_JnE&r=Jk6Q8nNzkQ6LJ6g42qARkg6ryIDGQr-yKXPNGZbpTx0&m=ximE6x4FrQ5vo3D4wF3w-cVsKgtNs85wCTL5GiP_B5Q&s=yvxjeOrwjXjf08RBhPdX53lJN1W-8WSXT25ZeMSA06k&e= :| >> |: https://urldefense.proofpoint.com/v2/url?u=https-3A__entangle-2Dphoto.org&d=DwIDAw&c=RoP1YumCXCgaWHvlZYR8PZh8Bv7qIrMUB65eapI_JnE&r=Jk6Q8nNzkQ6LJ6g42qARkg6ryIDGQr-yKXPNGZbpTx0&m=ximE6x4FrQ5vo3D4wF3w-cVsKgtNs85wCTL5GiP_B5Q&s=9TIjtmf6AVFYWbyzI5vl-zXTaNCSCMAxyck92pc8yvY&e= -o- https://urldefense.proofpoint.com/v2/url?u=https-3A__www.instagram.com_dberrange&d=DwIDAw&c=RoP1YumCXCgaWHvlZYR8PZh8Bv7qIrMUB65eapI_JnE&r=Jk6Q8nNzkQ6LJ6g42qARkg6ryIDGQr-yKXPNGZbpTx0&m=ximE6x4FrQ5vo3D4wF3w-cVsKgtNs85wCTL5GiP_B5Q&s=Wapdtm0yT4j-9EjMoxwo9QvRZ3h9Fk_CvpQq8J4TDpg&e= :| > -- > Dr. David Alan Gilbert / dgilbert@redhat.com / Manchester, UK
next prev parent reply other threads:[~2018-11-12 23:58 UTC|newest] Thread overview: 96+ messages / expand[flat|nested] mbox.gz Atom feed top 2018-09-16 12:46 [QEMU PATCH v2 0/2]: KVM: i386: Add support for save and restore nested state Liran Alon 2018-09-16 12:46 ` [Qemu-devel] " Liran Alon 2018-09-16 12:46 ` [QEMU PATCH v2 1/2] i386: Compile CPUX86State xsave_buf only when support KVM or HVF Liran Alon 2018-09-16 12:46 ` [Qemu-devel] " Liran Alon 2018-09-16 12:46 ` [QEMU PATCH v2 2/2] KVM: i386: Add support for save and restore nested state Liran Alon 2018-09-16 12:46 ` [Qemu-devel] " Liran Alon 2018-10-08 17:21 ` [QEMU PATCH v2 0/2]: " Liran Alon 2018-10-08 17:21 ` [Qemu-devel] " Liran Alon 2018-10-15 18:10 ` Liran Alon 2018-10-15 18:10 ` [Qemu-devel] " Liran Alon 2018-10-31 1:03 ` Liran Alon 2018-10-31 1:03 ` [Qemu-devel] " Liran Alon 2018-10-31 18:17 ` Eduardo Habkost 2018-10-31 18:17 ` [Qemu-devel] " Eduardo Habkost 2018-10-31 18:19 ` Paolo Bonzini 2018-10-31 18:19 ` [Qemu-devel] " Paolo Bonzini 2018-10-31 18:50 ` Liran Alon 2018-10-31 18:50 ` [Qemu-devel] " Liran Alon 2018-10-31 18:59 ` Dr. David Alan Gilbert 2018-10-31 18:59 ` [Qemu-devel] " Dr. David Alan Gilbert 2018-10-31 23:17 ` Liran Alon 2018-10-31 23:17 ` [Qemu-devel] " Liran Alon 2018-11-01 13:10 ` Dr. David Alan Gilbert 2018-11-01 13:10 ` [Qemu-devel] " Dr. David Alan Gilbert 2018-11-01 15:23 ` Liran Alon 2018-11-01 15:23 ` [Qemu-devel] " Liran Alon 2018-11-01 15:56 ` Dr. David Alan Gilbert 2018-11-01 15:56 ` [Qemu-devel] " Dr. David Alan Gilbert 2018-11-01 16:45 ` Jim Mattson via Qemu-devel 2018-11-01 16:45 ` [Qemu-devel] " Jim Mattson 2018-11-02 3:46 ` Liran Alon 2018-11-02 3:46 ` [Qemu-devel] " Liran Alon 2018-11-02 9:40 ` Paolo Bonzini 2018-11-02 9:40 ` [Qemu-devel] " Paolo Bonzini 2018-11-02 12:35 ` Dr. David Alan Gilbert 2018-11-02 12:35 ` [Qemu-devel] " Dr. David Alan Gilbert 2018-11-02 12:40 ` Daniel P. Berrangé 2018-11-02 12:40 ` [Qemu-devel] " Daniel P. Berrangé 2018-11-04 22:12 ` Paolo Bonzini 2018-11-04 22:12 ` [Qemu-devel] " Paolo Bonzini 2018-11-02 12:59 ` Liran Alon 2018-11-02 12:59 ` [Qemu-devel] " Liran Alon 2018-11-02 16:44 ` Jim Mattson via Qemu-devel 2018-11-02 16:44 ` [Qemu-devel] " Jim Mattson 2018-11-02 16:58 ` Daniel P. Berrangé 2018-11-02 16:58 ` [Qemu-devel] " Daniel P. Berrangé 2018-11-02 17:01 ` Jim Mattson via Qemu-devel 2018-11-02 17:01 ` [Qemu-devel] " Jim Mattson 2018-11-02 16:54 ` Daniel P. Berrangé 2018-11-02 16:54 ` [Qemu-devel] " Daniel P. Berrangé 2018-11-02 16:58 ` Dr. David Alan Gilbert 2018-11-02 16:58 ` [Qemu-devel] " Dr. David Alan Gilbert 2018-11-04 22:19 ` Paolo Bonzini 2018-11-04 22:19 ` [Qemu-devel] " Paolo Bonzini 2018-11-12 16:18 ` Daniel P. Berrangé 2018-11-12 16:18 ` [Qemu-devel] " Daniel P. Berrangé 2018-11-12 16:50 ` Dr. David Alan Gilbert 2018-11-12 16:50 ` [Qemu-devel] " Dr. David Alan Gilbert 2018-11-12 16:53 ` Paolo Bonzini 2018-11-12 16:53 ` [Qemu-devel] " Paolo Bonzini 2018-11-12 16:54 ` Daniel P. Berrangé 2018-11-12 16:54 ` [Qemu-devel] " Daniel P. Berrangé 2018-11-13 0:00 ` Liran Alon 2018-11-13 0:00 ` [Qemu-devel] " Liran Alon 2018-11-13 0:07 ` Jim Mattson via Qemu-devel 2018-11-13 0:07 ` [Qemu-devel] " Jim Mattson 2018-11-13 0:09 ` Liran Alon 2018-11-13 0:09 ` [Qemu-devel] " Liran Alon 2018-11-12 23:58 ` Liran Alon [this message] 2018-11-12 23:58 ` Liran Alon 2018-11-02 16:39 ` Jim Mattson via Qemu-devel 2018-11-02 16:39 ` [Qemu-devel] " Jim Mattson 2018-11-03 2:02 ` Liran Alon 2018-11-03 2:02 ` [Qemu-devel] " Liran Alon 2018-11-08 0:13 ` Liran Alon 2018-11-08 0:13 ` [Qemu-devel] " Liran Alon 2018-11-08 0:45 ` Jim Mattson via Qemu-devel 2018-11-08 0:45 ` [Qemu-devel] " Jim Mattson 2018-11-08 9:50 ` Paolo Bonzini 2018-11-08 9:50 ` [Qemu-devel] " Paolo Bonzini 2018-11-08 9:57 ` Liran Alon 2018-11-08 9:57 ` [Qemu-devel] " Liran Alon 2018-11-08 17:02 ` Paolo Bonzini 2018-11-08 17:02 ` [Qemu-devel] " Paolo Bonzini 2018-11-08 18:41 ` Liran Alon 2018-11-08 18:41 ` [Qemu-devel] " Liran Alon 2018-11-08 20:34 ` Paolo Bonzini 2018-11-08 20:34 ` [Qemu-devel] " Paolo Bonzini 2018-11-12 14:51 ` Dr. David Alan Gilbert 2018-11-12 14:51 ` [Qemu-devel] " Dr. David Alan Gilbert 2018-11-01 19:03 ` Liran Alon 2018-11-01 19:03 ` [Qemu-devel] " Liran Alon 2018-11-01 19:07 ` Dr. David Alan Gilbert 2018-11-01 19:07 ` [Qemu-devel] " Dr. David Alan Gilbert 2018-11-01 19:41 ` Jim Mattson via Qemu-devel 2018-11-01 19:41 ` [Qemu-devel] " Jim Mattson
Reply instructions: You may reply publicly to this message via plain-text email using any one of the following methods: * Save the following mbox file, import it into your mail client, and reply-to-all from there: mbox Avoid top-posting and favor interleaved quoting: https://en.wikipedia.org/wiki/Posting_style#Interleaved_style * Reply using the --to, --cc, and --in-reply-to switches of git-send-email(1): git send-email \ --in-reply-to=B6AEEFF2-BAF1-439A-8E44-51352C9E05AE@oracle.com \ --to=liran.alon@oracle.com \ --cc=dgilbert@redhat.com \ --cc=ehabkost@redhat.com \ --cc=jmattson@google.com \ --cc=kvm@vger.kernel.org \ --cc=mtosatti@redhat.com \ --cc=pbonzini@redhat.com \ --cc=qemu-devel@nongnu.org \ --cc=rth@twiddle.net \ /path/to/YOUR_REPLY https://kernel.org/pub/software/scm/git/docs/git-send-email.html * If your mail client supports setting the In-Reply-To header via mailto: links, try the mailto: linkBe sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes, see mirroring instructions on how to clone and mirror all data and code used by this external index.