virt: add irqfd trait and implement for mshv and KVM

[will-j-wright]

Add irqfd support enabling the kernel to inject MSIs directly into the
guest when an eventfd is signaled, without a userspace transition. This is
required for VFIO device passthrough where physical device interrupts must
be delivered to the guest via the hypervisor's irqfd mechanism.

New traits in virt crate:

• IrqFd: allocates GSIs and registers eventfds as irqfd routes
• IrqFdRoute: updates/clears MSI routing (address/data) per GSI
• Partition::irqfd(): returns the irqfd interface if supported

mshv implementation (virt_mshv/src/irqfd.rs):

• GSI allocation table (2048 slots)
• MSHV_IRQFD ioctl for register/unregister eventfds
• MSHV_SET_MSI_ROUTING ioctl to push full routing table
• Automatic cleanup on route drop (unregister + free GSI)
• VmFd changed to Arc<VmFd> for shared ownership

KVM implementation (virt_kvm/src/gsi.rs):

• KvmIrqFdState wraps existing GsiRoute infrastructure
• KvmIrqFdRoute delegates to GsiRoute::enable/disable
• Partition::irqfd() wired in x86_64 arch module

Tested: End-to-end with VFIO NVMe passthrough on mshv (L1 Azure Linux VM → L2 guest). NVMe driver probes, reads 100MB at 425MB/s with MSI-X interrupts delivered via irqfd.

[github-actions]

⚠️ Unsafe Code Detected

This PR modifies files containing unsafe Rust code. Extra scrutiny is required during review.

For more on why we check whole files, instead of just diffs, check out the Rustonomicon

[Copilot]

Pull request overview

This PR introduces a cross-backend irqfd abstraction in virt and wires up backend implementations for mshv and KVM so VFIO-style MSI/MSI-X interrupts can be injected by the kernel on eventfd signal (avoiding userspace transitions).

Changes:

• Added virt::irqfd traits and a Partition::irqfd() entry point to access irqfd routing when supported.
• Implemented mshv irqfd + MSI routing support (GSI allocation + MSHV_IRQFD / MSHV_SET_MSI_ROUTING ioctls), including shared VmFd ownership via Arc.
• Implemented KVM irqfd support by wrapping existing GsiRoute infrastructure and exposing it through the new traits.

Reviewed changes

Copilot reviewed 7 out of 7 changed files in this pull request and generated 5 comments.

Show a summary per file

File	Description
vmm_core/virt/src/lib.rs	Exposes the new irqfd module from the virt crate.
vmm_core/virt/src/irqfd.rs	Defines the new IrqFd / IrqFdRoute traits and access patterns.
vmm_core/virt/src/generic.rs	Adds Partition::irqfd() default method returning None for unsupported backends.
vmm_core/virt_mshv/src/lib.rs	Wires mshv partition to provide irqfd support; switches VmFd storage to Arc<VmFd>.
vmm_core/virt_mshv/src/irqfd.rs	Implements irqfd + MSI routing table management for mshv via ioctls.
vmm_core/virt_kvm/src/gsi.rs	Adds KvmIrqFdState/KvmIrqFdRoute backed by existing GsiRoute.
vmm_core/virt_kvm/src/arch/x86_64/mod.rs	Exposes KVM irqfd support through Partition::irqfd() on x86_64.

Comments suppressed due to low confidence (1)

vmm_core/virt_kvm/src/gsi.rs:152

• KvmIrqFdRoute::set_msi/clear_msi always return Ok(()), but the underlying GsiRoute::enable/disable paths can panic on KVM ioctl failures (set_gsi_routes(...).expect("should not fail"), kvm.irqfd(...).expect("should not fail")). Since the new IrqFdRoute API is fallible, it would be better to plumb errors through instead of panicking (change enable/disable/update_routes/irqfd calls to return Result and propagate up).

    pub fn enable(&self, entry: kvm::RoutingEntry) {
        let partition = self.set_entry(Some(entry));
        let _lock = self.enable_mutex.lock();
        if !self.enabled.load(Ordering::Relaxed) {
            if let (Some(partition), Some(event)) = (&partition, &self.irqfd_event) {
                partition
                    .kvm
                    .irqfd(self.gsi, event.as_fd().as_raw_fd(), true)
                    .expect("should not fail");