From: Joerg Roedel jroedel@suse.de
For now, kexec is not supported when running as an SEV-ES guest. Doing so requires additional hypervisor support and special code to hand over the CPUs to the new kernel in a safe way.
Until this is implemented, do not support kexec in SEV-ES guests.
Cc: stable@vger.kernel.org # v5.10+ Signed-off-by: Joerg Roedel jroedel@suse.de --- arch/x86/kernel/machine_kexec_64.c | 8 ++++++++ 1 file changed, 8 insertions(+)
diff --git a/arch/x86/kernel/machine_kexec_64.c b/arch/x86/kernel/machine_kexec_64.c index c078b0d3ab0e..f902cc9cc634 100644 --- a/arch/x86/kernel/machine_kexec_64.c +++ b/arch/x86/kernel/machine_kexec_64.c @@ -620,3 +620,11 @@ void arch_kexec_pre_free_pages(void *vaddr, unsigned int pages) */ set_memory_encrypted((unsigned long)vaddr, pages); } + +/* + * Kexec is not supported in SEV-ES guests yet + */ +bool arch_kexec_supported(void) +{ + return !sev_es_active(); +}
Joerg Roedel joro@8bytes.org writes:
From: Joerg Roedel jroedel@suse.de
For now, kexec is not supported when running as an SEV-ES guest. Doing so requires additional hypervisor support and special code to hand over the CPUs to the new kernel in a safe way.
Until this is implemented, do not support kexec in SEV-ES guests.
I don't understand this.
Fundamentally kexec is about doing things more or less inspite of what the firmware is doing.
I don't have any idea what a SEV-ES is. But the normal x86 boot doesn't do anything special. Is cross cpu IPI emulation buggy?
If this is a move in your face hypervisor like Xen is sometimes I can see perhaps needing a little bit of different work during bootup. Perhaps handing back a cpu on system shutdown and asking for more cpus on system boot up.
What is the actual problem you are trying to avoid?
And yes for a temporary hack the suggestion of putting code into machine_kexec_prepare seems much more reasonable so we don't have to carry special case infrastructure for the forseeable future.
Eric
Cc: stable@vger.kernel.org # v5.10+ Signed-off-by: Joerg Roedel jroedel@suse.de
arch/x86/kernel/machine_kexec_64.c | 8 ++++++++ 1 file changed, 8 insertions(+)
diff --git a/arch/x86/kernel/machine_kexec_64.c b/arch/x86/kernel/machine_kexec_64.c index c078b0d3ab0e..f902cc9cc634 100644 --- a/arch/x86/kernel/machine_kexec_64.c +++ b/arch/x86/kernel/machine_kexec_64.c @@ -620,3 +620,11 @@ void arch_kexec_pre_free_pages(void *vaddr, unsigned int pages) */ set_memory_encrypted((unsigned long)vaddr, pages); }
+/*
- Kexec is not supported in SEV-ES guests yet
- */
+bool arch_kexec_supported(void) +{
- return !sev_es_active();
+}
On Thu, May 06, 2021 at 12:42:03PM -0500, Eric W. Biederman wrote:
I don't understand this.
Fundamentally kexec is about doing things more or less inspite of what the firmware is doing.
I don't have any idea what a SEV-ES is. But the normal x86 boot doesn't do anything special. Is cross cpu IPI emulation buggy?
Under SEV-ES the normal SIPI-based sequence to re-initialize a CPU does not work anymore. An SEV-ES guest is a special virtual machine where the hardware encrypts the guest memory and the guest register state. The hypervisor can't make any modifications to the guests registers at runtime. Therefore it also can't emulate a SIPI sequence and reset the vCPU.
The guest kernel has to reset the vCPU itself and hand it over from the old kernel to the kexec'ed kernel. This isn't currently implemented and therefore kexec needs to be disabled when running as an SEV-ES guest.
Implementing this also requires an extension to the guest-hypervisor protocol (the GHCB Spec[1]) which is only available in version 2. So a guest running on a hypervisor supporting only version 1 will never properly support kexec.
What is the actual problem you are trying to avoid?
Currently, if someone tries kexec in an SEV-ES guest, the kexec'ed kernel will only be able to bring up the boot CPU, not the others. The others will wake up with the old kernels CPU state in the new kernels memory and do undefined things, most likely triple-fault because their page-table is not existent anymore.
So since kexec currently does not work as expected under SEV-ES, it is better to hide it until everything is implemented so it can do what the user expects.
And yes for a temporary hack the suggestion of putting code into machine_kexec_prepare seems much more reasonable so we don't have to carry special case infrastructure for the forseeable future.
As I said above, for protocol version 1 it will stay disabled, so it is not only a temporary hack.
Regards,
Joerg
[1] https://developer.amd.com/wp-content/resources/56421.pdf
Joerg Roedel jroedel@suse.de writes:
On Thu, May 06, 2021 at 12:42:03PM -0500, Eric W. Biederman wrote:
I don't understand this.
Fundamentally kexec is about doing things more or less inspite of what the firmware is doing.
I don't have any idea what a SEV-ES is. But the normal x86 boot doesn't do anything special. Is cross cpu IPI emulation buggy?
Under SEV-ES the normal SIPI-based sequence to re-initialize a CPU does not work anymore. An SEV-ES guest is a special virtual machine where the hardware encrypts the guest memory and the guest register state. The hypervisor can't make any modifications to the guests registers at runtime. Therefore it also can't emulate a SIPI sequence and reset the vCPU.
The guest kernel has to reset the vCPU itself and hand it over from the old kernel to the kexec'ed kernel. This isn't currently implemented and therefore kexec needs to be disabled when running as an SEV-ES guest.
Implementing this also requires an extension to the guest-hypervisor protocol (the GHCB Spec[1]) which is only available in version 2. So a guest running on a hypervisor supporting only version 1 will never properly support kexec.
Why does it need that?
Would it not make sense to instead teach kexec how to pass a cpu from one kernel to another. We could use that everywhere.
Even the kexec-on-panic case should work as even in that case we have to touch the cpus as they go down.
The hardware simply worked well enough that it hasn't mattered enough for us to do something like that, but given that we need to do something anyway. It seems like it would make most sense do something that will work everywhere, and does not introduce unnecessary dependencies on hypervisors.
What is the actual problem you are trying to avoid?
Currently, if someone tries kexec in an SEV-ES guest, the kexec'ed kernel will only be able to bring up the boot CPU, not the others. The others will wake up with the old kernels CPU state in the new kernels memory and do undefined things, most likely triple-fault because their page-table is not existent anymore.
So since kexec currently does not work as expected under SEV-ES, it is better to hide it until everything is implemented so it can do what the user expects.
I can understand temporarily disabling the functionality.
And yes for a temporary hack the suggestion of putting code into machine_kexec_prepare seems much more reasonable so we don't have to carry special case infrastructure for the forseeable future.
As I said above, for protocol version 1 it will stay disabled, so it is not only a temporary hack.
Why does bringing up a cpu need hypervisor support?
I understand why we can't do what we do currently, but that doesn't seem to preclude doing something without hypervisor support.
Eric
On Thu, May 06, 2021 at 01:59:42PM -0500, Eric W. Biederman wrote:
Joerg Roedel jroedel@suse.de writes:
Why does it need that?
Would it not make sense to instead teach kexec how to pass a cpu from one kernel to another. We could use that everywhere.
Even the kexec-on-panic case should work as even in that case we have to touch the cpus as they go down.
The hardware simply worked well enough that it hasn't mattered enough for us to do something like that, but given that we need to do something anyway. It seems like it would make most sense do something that will work everywhere, and does not introduce unnecessary dependencies on hypervisors.
Well, I guess we could implement something that even works for non SEV-ES guests and bare-metal. The question is what benefit we get from that. Is the SIPI sequence currently used not reliable enough?
The benefit of being able to rely on the SIPI sequence is that the kexec'ed kernel can use the same method to bring up APs as the first kernel did.
Btw, the same is true for SEV-ES guests, The goal is bring the APs of an SEV-ES guest into a state where they will use the SEV-ES method of AP-bringup when they wake up again. This method involves a firmware-owned page called the AP-jump-table, which contains the reset vector for the AP in its first 4 bytes.
As I said above, for protocol version 1 it will stay disabled, so it is not only a temporary hack.
Why does bringing up a cpu need hypervisor support?
When a CPU is taken offline under SEV-ES it will do a special hypercall named AP-reset-hold. The hypervisor will put the CPU into a halt state until the next SIPI arrives. In protocol version 1 this hypercall requires a GHCB shared page to be set up for the CPU doing the hypercall and upon CPU wakeup the HV will write to that shared page. Problem with that is that the page which contains the GHCB is already owned by the new kernel then and is probably not shared anymore, which can cause data corruption in the new kernel.
Version 2 of the protocol adds a purely MSR based version of the AP-reset-hold hypercall. This one does not need a GHCB page and has to be used to bring the APs offline before doing the kexec. That is the reason I plan to only support kexec when the hypervisor provides version 2 of the protocol.
Regards,
Joerg
linux-stable-mirror@lists.linaro.org
-
ebiederm＠xmission.com -
Joerg Roedel -
Joerg Roedel