[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20241004140810.34231-2-nikwip@amazon.de>
Date: Fri, 4 Oct 2024 14:08:04 +0000
From: Nikolas Wipper <nikwip@...zon.de>
To: Vitaly Kuznetsov <vkuznets@...hat.com>
CC: Nicolas Saenz Julienne <nsaenz@...zon.com>, Alexander Graf
<graf@...zon.de>, James Gowans <jgowans@...zon.com>,
<nh-open-source@...zon.com>, Sean Christopherson <seanjc@...gle.com>, "Paolo
Bonzini" <pbonzini@...hat.com>, Thomas Gleixner <tglx@...utronix.de>, "Ingo
Molnar" <mingo@...hat.com>, Borislav Petkov <bp@...en8.de>, Dave Hansen
<dave.hansen@...ux.intel.com>, Nikolas Wipper <nik.wipper@....de>,
<linux-kernel@...r.kernel.org>, <kvm@...r.kernel.org>, <x86@...nel.org>,
<linux-doc@...r.kernel.org>, <linux-kselftest@...r.kernel.org>, "Nikolas
Wipper" <nikwip@...zon.de>
Subject: [PATCH 1/7] KVM: Add API documentation for KVM_HYPERV_SET_TLB_FLUSH_INHIBIT
Add API documentation for the new KVM_HYPERV_SET_TLB_FLUSH_INHIBIT ioctl.
Signed-off-by: Nikolas Wipper <nikwip@...zon.de>
---
Documentation/virt/kvm/api.rst | 41 ++++++++++++++++++++++++++++++++++
1 file changed, 41 insertions(+)
diff --git a/Documentation/virt/kvm/api.rst b/Documentation/virt/kvm/api.rst
index a4b7dc4a9dda..9c11a8af336b 100644
--- a/Documentation/virt/kvm/api.rst
+++ b/Documentation/virt/kvm/api.rst
@@ -6443,6 +6443,47 @@ the capability to be present.
`flags` must currently be zero.
+4.144 KVM_HYPERV_SET_TLB_FLUSH_INHIBIT
+--------------------------------------
+
+:Capability: KVM_CAP_HYPERV_TLB_FLUSH_INHIBIT
+:Architectures: x86
+:Type: vcpu ioctl
+:Parameters: struct kvm_hyperv_tlb_flush_inhibit
+:returnReturns: 0 on success, this ioctl can't fail
+
+KVM_HYPERV_SET_TLB_FLUSH_INHIBIT allows userspace to prevent Hyper-V hyper-calls
+that remotely flush a vCPU's TLB, i.e. HvFlushVirtualAddressSpace(Ex)/
+HvFlushVirtualAddressList(Ex). When the flag is set, a vCPU attempting to flush
+an inhibited vCPU will be suspended and will only resume once the flag is
+cleared again using this ioctl. During suspension, the vCPU will not finish the
+hyper-call, but may enter the guest to retry it. Because it is caused by a
+hyper-call, the suspension naturally happens on a guest instruction boundary.
+This behaviour and the suspend state itself are specified in Microsoft's
+"Hypervisor Top Level Functional Specification" (TLFS).
+
+::
+
+ /* for KVM_HYPERV_SET_TLB_FLUSH_INHIBIT */
+ struct kvm_hyperv_tlb_flush_inhibit {
+ /* in */
+ __u16 flags;
+ #define KVM_HYPERV_UNINHIBIT_TLB_FLUSH 0
+ #define KVM_HYPERV_INHIBIT_TLB_FLUSH 1
+ __u8 inhibit;
+ __u8 padding[5];
+ };
+
+No flags are specified so far, the corresponding field must be set to zero,
+otherwise the ioctl will fail with exit code -EINVAL.
+
+The suspension is transparent to userspace. It won't cause KVM_RUN to return or
+the MP state to be changed. The suspension cannot be manually induced or exited
+apart from changing the TLB flush inhibit flag of a targeted processor.
+
+There is no way for userspace to query the state of the flush inhibit flag.
+Userspace must keep track of the required state itself.
+
5. The kvm_run structure
========================
--
2.40.1
Amazon Web Services Development Center Germany GmbH
Krausenstr. 38
10117 Berlin
Geschaeftsfuehrung: Christian Schlaeger, Jonathan Weiss
Eingetragen am Amtsgericht Charlottenburg unter HRB 257764 B
Sitz: Berlin
Ust-ID: DE 365 538 597
Powered by blists - more mailing lists