KVM Forum 2020
| title | author | length |
|---|---|---|
| A Journey to Support vGPU in Firecracker | Liang Yan | 26m |
Implemented vGPU support by adding a PCI device path parser, creating a mPCIe bus in Firecracker, and instantiating devices via VFIO passthrough using the… | ||
| A KVM-unit-tests and KVM selftests update for aarch64 | Eric Auger | 30m |
Status update on KVM unit and self tests for aarch64 with concrete changes: integration of KVM CFTest framework into the unit-test flow, new guest-space checks… | ||
| A Virtual IOMMU With Cooperative DMA Buffer Tracking | Yu Zhang | 24m |
A virtual IOMMU implements cooperative DMA buffer tracking by adding a guest-side driver that hooks DMA APIs, a shared DMA Tracking Table (DTT), and a host… | ||
| Advanced Parallel Memory Virtualization | Yulei Zhang | 11m |
Proposal pre-constructs pre-populated page tables so guest vCPUs use host-built mappings and avoid page-fault VM exits, allowing lockless PTE R/W/dirty updates… | ||
| AMD-vIOMMU: A Hardware-assisted Virtual IOMMU Technology | Suravee Suthikulpanit, Wei Huang | 22m |
Describes a hardware-assisted AMD vIOMMU that replaces slow shadow host tables with nested IO page tables so guests use v2 I/O page tables (IOVA→GPA) while the… | ||
| Bitmaps and NBD: Building Blocks of Change Block Tracking | Eric Blake | 27m |
Shows how qcow2 persistent dirty bitmaps and NBD block status let libvirt do efficient incremental backups: qcow2 stores host-visible qcow2 persistent dirty… | ||
| Bring SCSI Support Into QEMU Block Layer | Yaowei Bai | 10m |
Adds SCSI reservation support directly into QEMU’s block layer so guests access Ceph-like shared disks without iSCSI, implementing four components: device… | ||
| Building a Cloud Infrastructure to Deploy Microservices as Microvm Guests | Matias Vara Larsen | 19m |
Deploys microservices as tiny unikernel microVMs using Toro (app‑oriented unikernel), a microvm device model (Firecracker‑like microvm/qemu), virtio-fs for… | ||
| Challenges in Supporting Virtual CPU Hotplug on SoC Based Systems (like ARM64) | Salil Mehta | 25m |
Prototype implements virtual vCPU hotplug for ARM64 by changing QEMU/KVM and guest kernel: it pre-creates vCPUs and redistributors but avoids spawning threads… | ||
| Changing Paravirt Lock-ops for a Changing World | Ankur Arora | 26m |
Kernel can switch between native and paravirt spinlock ops at runtime by patching call sites with a single-byte breakpoint-based 0xCC patching that funnels… | ||
| Comparing Performance of NVMe Hard Drives in KVM, Baremetal, and Docker | Mauricio Tavares | 43m |
Compares NVMe throughput and IOPS on one Supermicro box with 10 identical Intel NVMe drives attached to a single controller (all in NUMA node 0), measuring… | ||
| Debugging KVM Using Intel DCI Technology | Raymond Zhang | 31m |
Shows that Intel DCI over a USB3 cable plus Empower System Studio replaces slow print debugging for KVM by letting a host inspect a running Ubuntu VM target… | ||
| Device Keepalive State for Local Live Migration and VMM Fast Restart | Jason Zeng | 22m |
Adds a VFIO-managed device keepalive state so PCI pass-through devices survive a host kexec and enable local live migration and VMM fast restart… | ||
| Evaluate Implementation Options of KVM-based Type1 (or 1.5) Hypervisor | Jun Nakajima | 27m |
Compares two KVM-derived designs: a minimally configured Linux+KVM hypervisor and a tiny reactive hypervisor built on VVH. Implements optimized VMCS shadowing… | ||
| Evolution of SPDK vhost Towards Secure Container Storage Service | Xiaodong Liu, Changpeng Liu | 24m |
SPDK vhost is converted from busy polling to an interruptable vhost target that uses an epoll/eventfd abstraction so reactors and spdk threads block instead of… | ||
| Extremely Fast and Efficient NFV with Unikraft | Sharan Santhanam | 25m |
Ports DPDK to Unikraft and delivers a portable build layer called libuk_dpdk plus adapters like uk_ring to translate netbufs↔mbufs; implements uk_pmd drivers… | ||
| Faster and Smaller qcow2 Files With Subcluster-based Allocation | Alberto Garcia | 18m |
Adds sub-cluster allocation via new extended L2 entries that split large qcow2 clusters into independently allocatable sub-clusters, reducing copy-on-write and… | ||
| Guest Memory Protection -- Current Status and Future | Isaku Yamahata | 18m |
Status update on a working Guest Memory Protection prototype that integrates into Linux KVM and QEMU and leverages hardware page‑table extensions (EPT/VMX) to… | ||
| HA-IOV: Applying Hardware-assisted Techniques to IO Virtualization Framework | Yifei Jiang, Bo Wan | 30m |
HA-IOV implements a RISC-V prototype that bypasses guest→QEMU roundtrips by adding user-level exception redirection, a virtual interrupt (VI) mechanism and… | ||
| Hypervisor Based Integrity: Protect Guest Kernel in Cloud | Ning Yang, Forrest Yuan Yu | 26m |
Proposal enforces runtime kernel integrity by loading a signed guest kernel module at boot that sends memory ranges to the hypervisor, which applies EPT/NPT… | ||
| Hypervisor-managed Linear Address Translation | Chao Gao | 20m |
Extension to Intel VT-x exposes a hypervisor-controlled page table that enforces guest linear-address integrity by marking ranges as a protected linear range… | ||
| Implementing SR-IOV Failover for Windows Guests During Migration | Yan Vugenfirer, Annie Li | 27m |
A two‑network model is implemented: a new virtual IO protocol driver plus a COM notify object that atomically switches datapaths between passthrough VF and… | ||
| Intel Virtualization Technology Extensions to Enable Hardware Isolated VMs | Sean Christopherson | 20m |
Hardware-enforced VMs (TDs) run with CPU/memory confidentiality and integrity using a signed TDX module in a new SEAM mode, private MKTME keys, secure EPT… | ||
| Keynote: KVM | Christian Bornträger | 15m |
KVM development shows measurable acceleration: commit volume from 5.4–5.9 rises about 40% more commits, reviewed patches improve (reported large improvement in… | ||
| Keynote: QEMU | Paolo Bonzini | 14m |
QEMU moves to Python 3 only, switches docs from Texinfo to Sphinx with Pandoc and kernel-doc integration and runs a continuous manual deployment so the manual… | ||
| KVM Address Space Isolation | Alexandre Chartre | 49m |
KVM Address Space Isolation claims to block speculative microarchitectural leaks (L1TF/MDS) by keeping sensitive kernel/guest pages unmapped and only mapping… | ||
| KVM Dirty Ring - A New Approach to Logging | Peter Xu | 26m |
KVM Dirty Ring replaces the linear dirty-bit bitmap with per-vCPU, mmaped dirty PFN rings that store dirtied page frame numbers; this removes the… | ||
| KVM Latency Performance Tuning | Wanpeng Li | 09m |
Implements a fastpass for virtual IPI via the KVM exit path that sends fixed single-target virtual IPIs before host interrupts and RCU, showing ~30% latency… | ||
| KVM Live Upgrade with Properly Handling of Passthrough Devices | Zhiming Feng | 15m |
Implements a KVM live-upgrade that preserves passthrough devices by loading parallel kernel modules (k1/k2), inheriting VF connectors, and keeping GPA→HPA… | ||
| KVM-unit-tests: When "KVM" Doesn't Mean KVM | Andrew Jones | 28m |
Claim: KVM unit tests already run on non‑KVM targets (QEMU TCG, Hypervisor Framework, Windows HV, s390x zVM/LPAR) and VMware contributors posted 50 patches to… | ||
| Libvirt Status Report | Daniel Berrangé | 12m |
Nine releases so far with a goal of 11 by year end, 144 contributors this year of which 87 are new, and 4,100 commits with ~5,200 projected. Project… | ||
| Live Migration With Hardware Acceleration | Wei Wang | 27m |
Offloads VM memory compression to a PCIe compression accelerator (Intel QAT) and plans to add DSA/DLA/IAA; implements zero copy DMA to guest memory, multi-page… | ||
| Long Live Asynchronous Page Fault! | Vitaly Kuznetsov, Vivek Goyal | 25m |
KVM implements an asynchronous page fault protocol that tags missing pages with a CR2 token and a shared APF reason; page‑ready used to be delivered as a… | ||
| Look Ma’, No (Real) Interrupt Controller! | Marc Zyngier, Christoffer Dall | 30m |
Implements a reduced virtual interrupt controller (ARVIC/RVIT) in KVM on ARM64 as a concrete prototype: ~1100 lines of code for the hypervisor side plus ~900… | ||
| Minimizing VMExits in Private Cloud by Aggressive PV IPI and Passthrough Timer | Huaqiao | 15m |
Implements a passthrough timer so guests program the physical LAPIC and the host swaps host LAPIC/TSC state on entry/exit, disables TSC MSR intercept, and… | ||
| Optimizing for NVMe Drives: The 10 Microsecond Challenge | Stefan Hajnoczi | 50m |
Claim: NVMe drives hit 10 microsecond latency so software overhead dominates; bare-metal PCI device assignment with polling (Linux NVMe iole) reaches about… | ||
| Panel Discussion: KVM-based Virtualization Contributor Q&A | Stefan Hajnoczi, Richard W.M. Jones, Susie Li, Hubertus Franke, David Kaplan, Peter Maydell | 50m |
Panel covers concrete engineering tradeoffs for confidential VMs: hardware key management, cache tagging and memory-controller encryption, and migration of… | ||
| PASID Management in KVM | Yi Liu, Jacob Pan | 19m |
KVM implements PASID-based DMA remapping for SVA and Intel Scalable IOV with a kernel IOASID core (landed since 5.05) that performs IOASID partitioning across… | ||
| QEMU Live Update | Steven J. Sistare | 18m |
A method to replace a running QEMU by execing a new binary while keeping guest RAM and most FDs intact uses a ~300‑line madvise_do_exec kernel patch, CPR… | ||
| QEMU Snaphosts Are So Slow. Really? | Denis V. Lunev | 22m |
Implementation adds a new snapshot data path with a rewritten cluster-aligned IO engine, delegated read worker threads, and a new on-disk snapshot format… | ||
| Rust-vmm Status Report | Andreea Florescu | 12m |
Rust-vmm provides modular virtualization crates in Rust, used in production since 2019 by Firecracker, cloud-hypervisor, Alibaba Cloud Sandbox and Inax, and is… | ||
| Scalable Work Submission in Device Virtualization | Hao Wu | 23m |
Claim: using the EQ command lets guest applications submit work directly to shared device queues while hardware and VMX translate guest PASIDs to host PASIDs… | ||
| Speed Up Creation of a VM With Pass Through GPU | Liang Li | 20m |
Claims specific kernel and QEMU patches that cut GPU pass-through VM creation from ~38s to 3.3s for 1 GPU 48GB and from ~120s to 4.4s for 4 GPUs 192GB (>95%… | ||
| Speeding Up VM’s I/O Sharing Host's io_uring Queues With Guests | Stefano Garzarella | 19m |
Demonstrates a working proof‑of‑concept that shares host io_uring queues by mapping SQ/CQ mapped in guest memory and using an eventfd plus io_uring restriction… | ||
| The Common Challenges of Secure VMs | Janosch Frank | 24m |
Overview of concrete problems and solutions for running encrypted VMs: memory is encrypted per-VM keys in hardware (AMD SEV/SNP, IBM secure execution, Intel… | ||
| The Practice Method to Speed Up 10x Boot-up Time for Guest in Alibaba Cloud | Weinan Li | 15m |
Claim: pinning all guest RAM for passthrough GPUs is the bottleneck, and an async DMA map plus a water balloon driver that pre-claims ranges speeds boot ~10x… | ||
| Towards an Alternative Memory Architecture | João Martins | 29m |
Proposal reclaims per-node metadata by remove struct page and repurpose DAX for volatile RAM via EFI memory tagging, turning large reserved ranges into… | ||
| Trap-less Virtual Interrupt for KVM on RISC-V | Siqi Zhao | 25m |
Proposal extends RISC-V interrupt controller (CLINT/PLIC) with new CSRs and implements KVM + QEMU emulation support to deliver trap-less delivery of virtual… | ||
| vDPA Support in Linux Kernel | Jason Wang | 24m |
Defines a new hardware class, the vDPA device, that exposes a virtio-compatible data plane and a vendor-specific control path and builds a kernel vDPA… | ||
| Virtio-(balloon|pmem|mem): Managing Guest Memory | David Hildenbrand, Michael S. Tsirkin | 29m |
Inflating a balloon from an 8 GiB VM to 4 GiB reduces migrated RAM to 4 GiB; migration and overcommit instead use free page hinting (host write-protects and… | ||
| Virtual Device Fuzzing Support in QEMU | Alexander Bulekov, Bandan Das | 29m |
Built a generic fuzzer that plugs into QEMU’s existing qtest and libqos test APIs to generate port‑IO, MMIO and DMA inputs, minimize crashes into compact qtest… | ||
| Virtual Topology for Virtual Machines: Friend or Foe? | Dario Faggioli | 40m |
Compares five VM network layouts (linux bridge, macvlan/macvtap, vhost-net, SR-IOV) using QEMU/KVM and libvirt, with iperf3 and pktgen benchmarks and perf CPU… | ||
| Virtual Versus Physical: Virtio Changes for New Hardware | Michael S. Tsirkin | 28m |
Virtio adds concrete changes for hardware: a packed ring descriptor layout plus driver-sent last-available offset reduces PCIe transactions to one read and two… | ||
| Virtualization for the Masses: Exposing KVM on Android | Will Deacon | 44m |
Project exposes KVM on Android by running a tiny hypervisor at EL2 as a self-contained Protected KVM payload embedded in the GKI kernel; patches land in Linux… | ||
| nothing matches | ||