Disable KVM_X86_QUIRK_SLOT_ZAP_ALL

By default, deleting a memslot flushes the guest mappings of every
memslot. KVM keeps that flush as a workaround for VMs with assigned
GPUs. Hyperlight assigns no devices and disables the quirk when the
kernel supports it (Linux 6.12 and later). Deleting a memslot then
flushes only that memslot's mappings.

Snapshot capture deletes a memslot when a layer drops out of the
snapshot. Restore to another snapshot and region unmapping also delete
memslots. Criterion on KVM (Linux 6.17):

* snapshots/call_and_create: 24% to 58% faster
* snapshots/restore_alternate: 32% to 52% faster

Signed-off-by: Ludvig Liljenberg <4257730+ludfjig@users.noreply.github.com>

Author: ludfjig

CHANGELOG.md

@@ -77,6 +77,9 @@ The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.1.0/).
 * Related snapshots share immutable memory layers. This reduces capture time
   and memory use. A guest can read data it overwrote before a snapshot from the
   older layers of that snapshot.
+* KVM: deleting a memory slot flushes only that slot's guest mappings on Linux
+  6.12 and later. This speeds up the first guest call after a snapshot,
+  restore, or unmap changes memory slots.
 * **Breaking:** `Sandbox::snapshot` installs the captured snapshot into the
   sandbox and removes `map_region` and `map_file_cow` regions. Pages of these
   regions that the guest has not mapped into its page tables are lost.

src/hyperlight_host/src/hypervisor/virtual_machine/kvm/x86_64.rs

@@ -7,8 +7,8 @@ use hyperlight_common::outb::VmAction;
 #[cfg(gdb)]
 use kvm_bindings::kvm_guest_debug;
 use kvm_bindings::{
-    Msrs, kvm_debugregs, kvm_fpu, kvm_msr_entry, kvm_regs, kvm_sregs, kvm_userspace_memory_region,
-    kvm_xsave,
+    KVM_CAP_DISABLE_QUIRKS2, KVM_X86_QUIRK_SLOT_ZAP_ALL, Msrs, kvm_debugregs, kvm_enable_cap,
+    kvm_fpu, kvm_msr_entry, kvm_regs, kvm_sregs, kvm_userspace_memory_region, kvm_xsave,
 };
 use kvm_ioctls::Cap::UserMemory;
 use kvm_ioctls::{
@@ -175,6 +175,29 @@ impl KvmVm {
             .create_vm_with_type(0)
             .map_err(|e| CreateVmError::CreateVmFd(e.into()))?;
 
+        // Deleting a memslot flushes the guest mappings of every memslot unless
+        // this quirk is disabled. KVM keeps the flush as a workaround for
+        // assigned GPUs. Hyperlight assigns no devices.
+        let quirks = vm_fd.check_extension_raw(KVM_CAP_DISABLE_QUIRKS2.into());
+        // `quirks` has one bit per quirk this kernel can disable.
+        if quirks <= 0 || quirks & KVM_X86_QUIRK_SLOT_ZAP_ALL as i32 == 0 {
+            tracing::debug!(
+                "Host kernel cannot disable KVM_X86_QUIRK_SLOT_ZAP_ALL. Each memory slot \
+                 deletion invalidates guest mappings in all slots, which adds latency to the \
+                 first guest call after a snapshot, restore, or unmap. Host kernel version \
+                 6.12 or later enables per-slot invalidation."
+            );
+        } else {
+            let cap = kvm_enable_cap {
+                cap: KVM_CAP_DISABLE_QUIRKS2,
+                args: [KVM_X86_QUIRK_SLOT_ZAP_ALL.into(), 0, 0, 0],
+                ..Default::default()
+            };
+            vm_fd
+                .enable_cap(&cap)
+                .map_err(|e| CreateVmError::InitializeVm(e.into()))?;
+        }
+
         #[cfg(feature = "hw-interrupts")]
         let timer_irq_eventfd = Self::setup_irqfd(&vm_fd)?;