From: James Bottomley <jejb@linux.ibm.com> To: "Dr. David Alan Gilbert" <dgilbert@redhat.com>, Tobin Feldman-Fitzthum <tobin@linux.ibm.com> Cc: Steve Rutherford <srutherford@google.com>, Paolo Bonzini <pbonzini@redhat.com>, Ashish Kalra <Ashish.Kalra@amd.com>, thomas.lendacky@amd.com, brijesh.singh@amd.com, ehabkost@redhat.com, kvm@vger.kernel.org, mst@redhat.com, tobin@ibm.com, richard.henderson@linaro.org, qemu-devel@nongnu.org, frankeh@us.ibm.com, dovmurik@linux.vnet.ibm.com Subject: Re: [RFC PATCH 00/13] Add support for Mirror VM. Date: Thu, 19 Aug 2021 10:06:50 -0400 [thread overview] Message-ID: <538733190532643cc19b6e30f0eda4dd1bc2a767.camel@linux.ibm.com> (raw) In-Reply-To: <YR4U11ssVUztsPyx@work-vm> On Thu, 2021-08-19 at 09:22 +0100, Dr. David Alan Gilbert wrote: > * Tobin Feldman-Fitzthum (tobin@linux.ibm.com) wrote: > > On 8/18/21 3:04 PM, Dr. David Alan Gilbert wrote: > > > * Tobin Feldman-Fitzthum (tobin@linux.ibm.com) wrote: > > > > On 8/17/21 6:04 PM, Steve Rutherford wrote: > > > > > Ahh, It sounds like you are looking into sidestepping the > > > > > existing AMD-SP flows for migration. I assume the idea is to > > > > > spin up a VM on the target side, and have the two VMs attest > > > > > to each other. How do the two sides know if the other is > > > > > legitimate? I take it that the source is directing the LAUNCH > > > > > flows? > > > > > > > > Yeah we don't use PSP migration flows at all. We don't need to > > > > send the MH code from the source to the target because the MH > > > > lives in firmware, which is common between the two. > > > > > > Are you relying on the target firmware to be *identical* or > > > purely for it to be *compatible* ? It's normal for a migration > > > to be the result of wanting to do an upgrade; and that means the > > > destination build of OVMF might be newer (or older, or ...). > > > > > > Dave > > > > This is a good point. The migration handler on the source and > > target must have the same memory footprint or bad things will > > happen. Using the same firmware on the source and target is an easy > > way to guarantee this. Since the MH in OVMF is not a contiguous > > region of memory, but a group of functions scattered around OVMF, > > it is a bit difficult to guarantee that the memory footprint is the > > same if the build is different. > > Can you explain what the 'memory footprint' consists of? Can't it > just be the whole of the OVMF rom space if you have no way of nudging > the MH into it's own chunk? It might be possible depending on how we link it. At the moment it's using the core OVMF libraries, but it is possible to retool the OVMF build to copy those libraries into the MH DXE. > I think it really does have to cope with migration to a new version > of host. Well, you're thinking of OVMF as belonging to the host because of the way it is supplied, but think about the way it works in practice now, forgetting about confidential computing: OVMF is RAM resident in ordinary guests, so when you migrate them, the whole of OVMF (or at least what's left at runtime) goes with the migration, thus it's not possible to change the guest OVMF by migration. The above is really just an extension of that principle, the only difference for confidential computing being you have to have an image of the current OVMF ROM in the target to seed migration. Technically, the problem is we can't overwrite running code and once the guest is re-sited to the target, the OVMF there has to match exactly what was on the source for the RT to still function. Once the migration has run, the OVMF on the target must be identical to what was on the source (including internally allocated OVMF memory), and if we can't copy the MH code, we have to rely on the target image providing this identical code and we copy the rest. James
WARNING: multiple messages have this Message-ID (diff)
From: James Bottomley <jejb@linux.ibm.com> To: "Dr. David Alan Gilbert" <dgilbert@redhat.com>, Tobin Feldman-Fitzthum <tobin@linux.ibm.com> Cc: thomas.lendacky@amd.com, Ashish Kalra <Ashish.Kalra@amd.com>, brijesh.singh@amd.com, ehabkost@redhat.com, kvm@vger.kernel.org, mst@redhat.com, Steve Rutherford <srutherford@google.com>, richard.henderson@linaro.org, tobin@ibm.com, qemu-devel@nongnu.org, frankeh@us.ibm.com, Paolo Bonzini <pbonzini@redhat.com>, dovmurik@linux.vnet.ibm.com Subject: Re: [RFC PATCH 00/13] Add support for Mirror VM. Date: Thu, 19 Aug 2021 10:06:50 -0400 [thread overview] Message-ID: <538733190532643cc19b6e30f0eda4dd1bc2a767.camel@linux.ibm.com> (raw) In-Reply-To: <YR4U11ssVUztsPyx@work-vm> On Thu, 2021-08-19 at 09:22 +0100, Dr. David Alan Gilbert wrote: > * Tobin Feldman-Fitzthum (tobin@linux.ibm.com) wrote: > > On 8/18/21 3:04 PM, Dr. David Alan Gilbert wrote: > > > * Tobin Feldman-Fitzthum (tobin@linux.ibm.com) wrote: > > > > On 8/17/21 6:04 PM, Steve Rutherford wrote: > > > > > Ahh, It sounds like you are looking into sidestepping the > > > > > existing AMD-SP flows for migration. I assume the idea is to > > > > > spin up a VM on the target side, and have the two VMs attest > > > > > to each other. How do the two sides know if the other is > > > > > legitimate? I take it that the source is directing the LAUNCH > > > > > flows? > > > > > > > > Yeah we don't use PSP migration flows at all. We don't need to > > > > send the MH code from the source to the target because the MH > > > > lives in firmware, which is common between the two. > > > > > > Are you relying on the target firmware to be *identical* or > > > purely for it to be *compatible* ? It's normal for a migration > > > to be the result of wanting to do an upgrade; and that means the > > > destination build of OVMF might be newer (or older, or ...). > > > > > > Dave > > > > This is a good point. The migration handler on the source and > > target must have the same memory footprint or bad things will > > happen. Using the same firmware on the source and target is an easy > > way to guarantee this. Since the MH in OVMF is not a contiguous > > region of memory, but a group of functions scattered around OVMF, > > it is a bit difficult to guarantee that the memory footprint is the > > same if the build is different. > > Can you explain what the 'memory footprint' consists of? Can't it > just be the whole of the OVMF rom space if you have no way of nudging > the MH into it's own chunk? It might be possible depending on how we link it. At the moment it's using the core OVMF libraries, but it is possible to retool the OVMF build to copy those libraries into the MH DXE. > I think it really does have to cope with migration to a new version > of host. Well, you're thinking of OVMF as belonging to the host because of the way it is supplied, but think about the way it works in practice now, forgetting about confidential computing: OVMF is RAM resident in ordinary guests, so when you migrate them, the whole of OVMF (or at least what's left at runtime) goes with the migration, thus it's not possible to change the guest OVMF by migration. The above is really just an extension of that principle, the only difference for confidential computing being you have to have an image of the current OVMF ROM in the target to seed migration. Technically, the problem is we can't overwrite running code and once the guest is re-sited to the target, the OVMF there has to match exactly what was on the source for the RT to still function. Once the migration has run, the OVMF on the target must be identical to what was on the source (including internally allocated OVMF memory), and if we can't copy the MH code, we have to rely on the target image providing this identical code and we copy the rest. James
next prev parent reply other threads:[~2021-08-19 14:07 UTC|newest] Thread overview: 104+ messages / expand[flat|nested] mbox.gz Atom feed top 2021-08-16 13:25 [RFC PATCH 00/13] Add support for Mirror VM Ashish Kalra 2021-08-16 13:26 ` [RFC PATCH 01/13] machine: Add mirrorvcpus=N suboption to -smp Ashish Kalra 2021-08-16 21:23 ` Eric Blake 2021-08-16 21:23 ` Eric Blake 2021-08-16 13:27 ` [RFC PATCH 02/13] hw/boards: Add mirror_vcpu flag to CPUArchId Ashish Kalra 2021-08-16 13:27 ` [RFC PATCH 03/13] hw/i386: Mark mirror vcpus in possible_cpus Ashish Kalra 2021-08-16 13:27 ` [RFC PATCH 04/13] hw/acpi: Don't include mirror vcpus in ACPI tables Ashish Kalra 2021-08-16 13:28 ` [RFC PATCH 05/13] cpu: Add boolean mirror_vcpu field to CPUState Ashish Kalra 2021-08-16 13:28 ` [RFC PATCH 06/13] hw/i386: Set CPUState.mirror_vcpu=true for mirror vcpus Ashish Kalra 2021-08-16 13:29 ` [RFC PATCH 07/13] kvm: Add Mirror VM ioctl and enable cap interfaces Ashish Kalra 2021-08-16 13:29 ` [RFC PATCH 08/13] kvm: Add Mirror VM support Ashish Kalra 2021-08-16 13:29 ` [RFC PATCH 09/13] kvm: create Mirror VM and share primary VM's encryption context Ashish Kalra 2021-08-16 13:30 ` [RFC PATCH 10/13] softmmu/cpu: Skip mirror vcpu's for pause, resume and synchronization Ashish Kalra 2021-08-16 13:30 ` [RFC PATCH 11/13] kvm/apic: Disable in-kernel APIC support for mirror vcpu's Ashish Kalra 2021-08-16 13:31 ` [RFC PATCH 12/13] hw/acpi: disable modern CPU hotplug interface " Ashish Kalra 2021-08-16 13:31 ` [RFC PATCH 13/13] hw/i386/pc: reduce fw_cfg boot cpu count taking into account " Ashish Kalra 2021-08-16 14:01 ` [RFC PATCH 00/13] Add support for Mirror VM Claudio Fontana 2021-08-16 14:01 ` Claudio Fontana 2021-08-16 14:15 ` Paolo Bonzini 2021-08-16 14:15 ` Paolo Bonzini 2021-08-16 14:23 ` Daniel P. Berrangé 2021-08-16 14:23 ` Daniel P. Berrangé 2021-08-16 15:00 ` Paolo Bonzini 2021-08-16 15:00 ` Paolo Bonzini 2021-08-16 15:16 ` Daniel P. Berrangé 2021-08-16 15:16 ` Daniel P. Berrangé 2021-08-16 15:35 ` Paolo Bonzini 2021-08-16 15:35 ` Paolo Bonzini 2021-08-16 14:44 ` Ashish Kalra 2021-08-16 14:58 ` Paolo Bonzini 2021-08-16 14:58 ` Paolo Bonzini 2021-08-16 15:13 ` Ashish Kalra 2021-08-16 15:38 ` Paolo Bonzini 2021-08-16 15:38 ` Paolo Bonzini 2021-08-16 15:48 ` Dr. David Alan Gilbert 2021-08-16 15:48 ` Dr. David Alan Gilbert 2021-08-18 10:31 ` Ashish Kalra 2021-08-18 11:25 ` James Bottomley 2021-08-18 11:25 ` James Bottomley 2021-08-18 15:31 ` Dr. David Alan Gilbert 2021-08-18 15:31 ` Dr. David Alan Gilbert 2021-08-18 15:35 ` James Bottomley 2021-08-18 15:35 ` James Bottomley 2021-08-18 15:43 ` Dr. David Alan Gilbert 2021-08-18 15:43 ` Dr. David Alan Gilbert 2021-08-18 16:28 ` James Bottomley 2021-08-18 16:28 ` James Bottomley 2021-08-18 17:30 ` Dr. David Alan Gilbert 2021-08-18 17:30 ` Dr. David Alan Gilbert 2021-08-18 18:51 ` James Bottomley 2021-08-18 18:51 ` James Bottomley 2021-08-18 19:47 ` Paolo Bonzini 2021-08-16 17:23 ` Dr. David Alan Gilbert 2021-08-16 17:23 ` Dr. David Alan Gilbert 2021-08-16 20:53 ` Paolo Bonzini 2021-08-16 23:53 ` Steve Rutherford 2021-08-16 23:53 ` Steve Rutherford 2021-08-17 7:05 ` Michael S. Tsirkin 2021-08-17 7:05 ` Michael S. Tsirkin 2021-08-17 8:38 ` Dr. David Alan Gilbert 2021-08-17 8:38 ` Dr. David Alan Gilbert 2021-08-17 14:08 ` Ashish Kalra 2021-08-17 16:32 ` Paolo Bonzini 2021-08-17 16:32 ` Paolo Bonzini 2021-08-17 20:50 ` Tobin Feldman-Fitzthum 2021-08-17 20:50 ` Tobin Feldman-Fitzthum 2021-08-17 22:04 ` Steve Rutherford 2021-08-17 22:04 ` Steve Rutherford 2021-08-18 15:32 ` Tobin Feldman-Fitzthum 2021-08-18 15:32 ` Tobin Feldman-Fitzthum 2021-08-18 19:04 ` Dr. David Alan Gilbert 2021-08-18 19:04 ` Dr. David Alan Gilbert 2021-08-18 21:42 ` Tobin Feldman-Fitzthum 2021-08-18 21:42 ` Tobin Feldman-Fitzthum 2021-08-19 8:22 ` Dr. David Alan Gilbert 2021-08-19 8:22 ` Dr. David Alan Gilbert 2021-08-19 14:06 ` James Bottomley [this message] 2021-08-19 14:06 ` James Bottomley 2021-08-19 14:28 ` Dr. David Alan Gilbert 2021-08-19 14:28 ` Dr. David Alan Gilbert 2021-08-19 22:10 ` James Bottomley 2021-08-19 22:10 ` James Bottomley 2021-08-23 12:26 ` Dr. David Alan Gilbert 2021-08-23 12:26 ` Dr. David Alan Gilbert 2021-08-23 16:28 ` Tobin Feldman-Fitzthum 2021-08-23 16:28 ` Tobin Feldman-Fitzthum 2021-08-19 14:07 ` Tobin Feldman-Fitzthum 2021-08-19 14:07 ` Tobin Feldman-Fitzthum 2021-08-17 23:20 ` Paolo Bonzini 2021-08-17 23:20 ` Paolo Bonzini 2021-08-17 21:54 ` Steve Rutherford 2021-08-17 21:54 ` Steve Rutherford 2021-08-17 22:37 ` Paolo Bonzini 2021-08-17 22:37 ` Paolo Bonzini 2021-08-17 22:57 ` James Bottomley 2021-08-17 22:57 ` James Bottomley 2021-08-17 23:10 ` Steve Rutherford 2021-08-17 23:10 ` Steve Rutherford 2021-08-18 2:49 ` James Bottomley 2021-08-18 2:49 ` James Bottomley 2021-08-18 14:06 ` Ashish Kalra 2021-08-18 17:07 ` Ashish Kalra 2021-08-16 15:07 Tobin Feldman-Fitzthum 2021-08-16 15:07 ` Tobin Feldman-Fitzthum
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=538733190532643cc19b6e30f0eda4dd1bc2a767.camel@linux.ibm.com \ --to=jejb@linux.ibm.com \ --cc=Ashish.Kalra@amd.com \ --cc=brijesh.singh@amd.com \ --cc=dgilbert@redhat.com \ --cc=dovmurik@linux.vnet.ibm.com \ --cc=ehabkost@redhat.com \ --cc=frankeh@us.ibm.com \ --cc=kvm@vger.kernel.org \ --cc=mst@redhat.com \ --cc=pbonzini@redhat.com \ --cc=qemu-devel@nongnu.org \ --cc=richard.henderson@linaro.org \ --cc=srutherford@google.com \ --cc=thomas.lendacky@amd.com \ --cc=tobin@ibm.com \ --cc=tobin@linux.ibm.com \ /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.