build: Use single version of vhost-user-backend

Using multiple different versions breaks vendoring.

Signed-off-by: Rob Bradford <robert.bradford@intel.com>

.github/workflows/build.yaml

@@ -13,6 +13,7 @@ jobs:
           - stable
           - beta
           - nightly
+          - 1.54
         target:
           - x86_64-unknown-linux-gnu
           - x86_64-unknown-linux-musl
@@ -22,6 +23,9 @@ jobs:
         with:
           fetch-depth: 0
 
+      - name: Install musl-gcc
+        run: sudo apt install -y musl-tools
+
       - name: Install Rust toolchain (${{ matrix.rust }})
         uses: actions-rs/toolchain@v1
         with:
@@ -34,6 +38,9 @@ jobs:
           git rev-list origin/main..$GITHUB_SHA | xargs -t -I % sh -c 'git checkout %; cargo check --all --target=${{ matrix.target }}'
           git checkout $GITHUB_SHA
 
+      - name: Build (default + tdx)
+        run: cargo rustc --bin cloud-hypervisor --features "tdx" -- -D warnings
+
       - name: Build (acpi,kvm)
         run: cargo rustc --bin cloud-hypervisor --no-default-features --features "acpi,kvm"  -- -D warnings
 

.github/workflows/dco.yaml

@@ -15,4 +15,4 @@ jobs:
         GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
       run: |
         pip3 install -U dco-check
-        dco-check
+        dco-check -e "49699333+dependabot[bot]@users.noreply.github.com"

.github/workflows/docker-image.yaml

@@ -1,9 +1,11 @@
-name: Cloud-Hypervisor's Docker image update
+name: Cloud Hypervisor's Docker image update
 
 on:
   push:
     branches: main
     paths: resources/Dockerfile
+  pull_request:
+    paths: resources/Dockerfile
 
 jobs:
   main:
@@ -19,18 +21,39 @@ jobs:
         uses: docker/setup-buildx-action@v1
 
       - name: Login to DockerHub
+        if: ${{ github.event_name == 'push' }}
         uses: docker/login-action@v1 
         with:
           username: ${{ secrets.DOCKERHUB_USERNAME }}
           password: ${{ secrets.DOCKERHUB_TOKEN }}
 
+      - name: Docker meta
+        id: meta
+        uses: docker/metadata-action@v3
+        with:
+          # list of Docker images to use as base name for tags
+          images: cloudhypervisor/dev
+          # generate Docker tags based on the following events/attributes
+          tags: |
+            type=raw,value={{date 'YYYYMMDD'}}-0
+            type=sha
+
       - name: Build and push
+        if: ${{ github.event_name == 'push' }}
         uses: docker/build-push-action@v2
         with:
           file: ./resources/Dockerfile
           platforms: linux/amd64,linux/arm64
           push: true
-          tags: cloudhypervisor/dev:latest
+          tags: ${{ steps.meta.outputs.tags }}
+
+      - name: Build only
+        if: ${{ github.event_name == 'pull_request' }}
+        uses: docker/build-push-action@v2
+        with:
+          file: ./resources/Dockerfile
+          platforms: linux/amd64,linux/arm64
+          tags: ${{ steps.meta.outputs.tags }}
 
       - name: Image digest
         run: echo ${{ steps.docker_build.outputs.digest }}

.github/workflows/quality.yaml

@@ -55,7 +55,7 @@ jobs:
         run: cargo clippy --all --all-targets --no-default-features --tests --features "mshv" -- -D warnings
 
       - name: Clippy (integration tests)
-        run: cargo clippy --all --all-targets --tests --features "integration_tests" -- -D warnings
+        run: cargo clippy --all --all-targets --tests -- -D warnings
 
       - name: Check build did not modify any files
         run: test -z "$(git status --porcelain)"

.github/workflows/release.yaml

@@ -9,6 +9,10 @@ jobs:
     steps:
       - name: Code checkout
         uses: actions/checkout@v2
+      - name: Install musl-gcc
+        run: sudo apt install -y musl-tools      
+      - name: Create release directory
+        run: rsync -rv --exclude=.git . ../cloud-hypervisor-${{ github.event.ref }}
       - name: Install Rust toolchain (x86_64-unknown-linux-gnu)
         uses: actions-rs/toolchain@v1
         with:
@@ -37,6 +41,13 @@ jobs:
             use-cross: true
             command: build
             args: --all --release --target=aarch64-unknown-linux-musl
+      - name: Vendor
+        working-directory: ../cloud-hypervisor-${{ github.event.ref }}
+        run: |
+          mkdir ../vendor-cargo-home
+          export CARGO_HOME=$(realpath ../vendor-cargo-home)
+          mkdir .cargo
+          cargo vendor > .cargo/config.toml
       - name: Create Release
         id: create_release
         uses: actions/create-release@v1
@@ -47,6 +58,19 @@ jobs:
           release_name: ${{ github.ref }}
           draft: true
           prerelease: true
+      - name: Create vendored source archive
+        working-directory: ../
+        run:  tar cJf cloud-hypervisor-${{ github.event.ref }}.tar.xz cloud-hypervisor-${{ github.event.ref }}
+      - name: Upload cloud-hypervisor vendored source archive
+        id: upload-release-cloud-hypervisor-vendored-sources
+        uses: actions/upload-release-asset@v1
+        env:
+          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
+        with:
+          upload_url: ${{ steps.create_release.outputs.upload_url }}
+          asset_path: ../cloud-hypervisor-${{ github.event.ref }}.tar.xz
+          asset_name: cloud-hypervisor-${{ github.event.ref }}.tar.xz
+          asset_content_type: application/x-xz
       - name: Upload cloud-hypervisor
         id: upload-release-cloud-hypervisor
         uses: actions/upload-release-asset@v1
@@ -77,6 +101,16 @@ jobs:
           asset_path: target/x86_64-unknown-linux-gnu/release/ch-remote
           asset_name: ch-remote
           asset_content_type: application/octet-stream
+      - name: Upload static-ch-remote
+        id: upload-release-static-ch-remote
+        uses: actions/upload-release-asset@v1
+        env:
+          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
+        with:
+          upload_url: ${{ steps.create_release.outputs.upload_url }}
+          asset_path: target/x86_64-unknown-linux-musl/release/ch-remote
+          asset_name: ch-remote-static
+          asset_content_type: application/octet-stream
       - name: Upload static AArch64 cloud-hypervisor
         id: upload-release-static-aarch64-cloud-hypervisor
         uses: actions/upload-release-asset@v1

.gitignore

@@ -1,4 +1,5 @@
 /build
+/.cargo
 /target
 **/*.rs.bk
 **/Cargo.lock

CONTRIBUTING.md

@@ -1,6 +1,13 @@
 # Contributing to Cloud Hypervisor
 
-Cloud Hypervisor is an open source project licensed under the [Apache v2 License](https://opensource.org/licenses/Apache-2.0) and the [BSD 3 Clause](https://opensource.org/licenses/BSD-3-Clause) license.
+Cloud Hypervisor is an open source project licensed under the [Apache v2
+License](https://opensource.org/licenses/Apache-2.0) and the [BSD 3
+Clause](https://opensource.org/licenses/BSD-3-Clause) license. Contributions
+can be made under either license or both. Individual files contain details of
+their licensing and changes to that file are under the same license unless the
+contribution changes the license of the file. When importing code from a third
+party project (e.g. Firecracker or CrosVM) please respect the license of those
+projects.
 
 ## Coding Style
 

Cargo.lock

@@ -9,21 +9,6 @@ dependencies = [
  "vm-memory",
 ]
 
-[[package]]
-name = "addr2line"
-version = "0.16.0"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "3e61f2b7f93d2c7d2b08263acaa4a363b3e276806c68af6134c44f523bf1aacd"
-dependencies = [
- "gimli",
-]
-
-[[package]]
-name = "adler"
-version = "1.0.2"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "f26201604c87b1e01bd3d98f8d5d9a8fcbb815e8cedb41ffccbeb4bf593a35fe"
-
 [[package]]
 name = "aho-corasick"
 version = "0.7.18"
@@ -33,20 +18,11 @@ dependencies = [
  "memchr",
 ]
 
-[[package]]
-name = "ansi_term"
-version = "0.11.0"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "ee49baf6cb617b853aa8d93bf420db2383fab46d314482ca2803b40d5fde979b"
-dependencies = [
- "winapi",
-]
-
 [[package]]
 name = "anyhow"
-version = "1.0.44"
+version = "1.0.55"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "61604a8f862e1d5c3229fdd78f8b02c68dcf73a4c4b05fd636d12240aaa242c1"
+checksum = "159bb86af3a200e19a068f4224eae4c8bb2d0fa054c7e5d1cacd5cef95e684cd"
 
 [[package]]
 name = "api_client"
@@ -57,9 +33,9 @@ dependencies = [
 
 [[package]]
 name = "arc-swap"
-version = "1.4.0"
+version = "1.5.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "e6df5aef5c5830360ce5218cecb8f018af3438af5686ae945094affc86fdec63"
+checksum = "c5d78ce20460b82d3fa150275ed9d55e21064fc7951177baacf86a145c4a4b1f"
 
 [[package]]
 name = "arch"
@@ -97,24 +73,9 @@ dependencies = [
 
 [[package]]
 name = "autocfg"
-version = "1.0.1"
+version = "1.1.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "cdb031dd78e28731d87d56cc8ffef4a8f36ca26c38fe2de700543e627f8a464a"
+checksum = "d468802bab17cbc0cc575e9b053f41e72aa36bfa6b7f55e3529ffa43161b97fa"
-
-[[package]]
-name = "backtrace"
-version = "0.3.61"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "e7a905d892734eea339e896738c14b9afce22b5318f64b951e70bf3844419b01"
-dependencies = [
- "addr2line",
- "cc",
- "cfg-if",
- "libc",
- "miniz_oxide",
- "object",
- "rustc-demangle",
-]
 
 [[package]]
 name = "bincode"
@@ -144,6 +105,7 @@ dependencies = [
  "versionize_derive",
  "vhdx",
  "virtio-bindings",
+ "virtio-queue",
  "vm-memory",
  "vm-virtio",
  "vmm-sys-util",
@@ -157,9 +119,9 @@ checksum = "14c189c53d098945499cdfa7ecc63567cf3886b3332b312a5b4585d8d3a6a610"
 
 [[package]]
 name = "cc"
-version = "1.0.71"
+version = "1.0.73"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "79c2681d6594606957bbb8631c4b90a7fcaaa72cdb714743a437b156d6a7eedd"
+checksum = "2fff2a6927b3bb87f9595d67196a70493f627687a71d87a0d692242c33f58c11"
 
 [[package]]
 name = "cfg-if"
@@ -169,29 +131,29 @@ checksum = "baf1de4339761588bc0619e3cbc0120ee582ebb74b53b4efbf79117bd2da40fd"
 
 [[package]]
 name = "clap"
-version = "2.33.3"
+version = "3.1.5"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "37e58ac78573c40708d45522f0d80fa2f01cc4f9b4e2bf749807255454312002"
+checksum = "ced1892c55c910c1219e98d6fc8d71f6bddba7905866ce740066d8bfea859312"
 dependencies = [
- "ansi_term",
  "atty",
  "bitflags",
+ "indexmap",
+ "lazy_static",
+ "os_str_bytes",
  "strsim",
- "term_size",
+ "termcolor",
+ "terminal_size",
  "textwrap",
- "unicode-width",
- "vec_map",
 ]
 
 [[package]]
 name = "cloud-hypervisor"
-version = "19.0.0"
+version = "22.0.0"
 dependencies = [
  "anyhow",
  "api_client",
  "clap",
- "credibility",
+ "dirs",
- "dirs 4.0.0",
  "epoll",
  "event_monitor",
  "hypervisor",
@@ -213,9 +175,9 @@ dependencies = [
 
 [[package]]
 name = "crc32c"
-version = "0.6.0"
+version = "0.6.1"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "210cdf933e6a81212bfabf90cd8762f471b5922e5f6b709547673ad8e04b9448"
+checksum = "ee6b9c9389584bcba988bd0836086789b7f87ad91892d6a83d5291dbb24524b5"
 dependencies = [
  "rustc_version",
 ]
@@ -226,16 +188,6 @@ version = "1.0.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "55626594feae15d266d52440b26ff77de0e22230cf0c113abe619084c1ddc910"
 
-[[package]]
-name = "credibility"
-version = "0.1.3"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "fae7a162fd5b462bc49704873a89950a655d44161add4be07e00e64c4c83a5bf"
-dependencies = [
- "failure",
- "failure_derive",
-]
-
 [[package]]
 name = "devices"
 version = "0.1.0"
@@ -256,15 +208,6 @@ dependencies = [
  "vmm-sys-util",
 ]
 
-[[package]]
-name = "dirs"
-version = "3.0.2"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "30baa043103c9d0c2a57cf537cc2f35623889dc0d405e6c3cccfadbc81c71309"
-dependencies = [
- "dirs-sys",
-]
-
 [[package]]
 name = "dirs"
 version = "4.0.0"
@@ -318,28 +261,6 @@ dependencies = [
  "serde_json",
 ]
 
-[[package]]
-name = "failure"
-version = "0.1.8"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "d32e9bd16cc02eae7db7ef620b392808b89f6a5e16bb3497d159c6b92a0f4f86"
-dependencies = [
- "backtrace",
- "failure_derive",
-]
-
-[[package]]
-name = "failure_derive"
-version = "0.1.8"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "aa4da3c766cd7a0db8242e326e9e4e081edd567072893ed320008189715366a4"
-dependencies = [
- "proc-macro2",
- "quote",
- "syn",
- "synstructure",
-]
-
 [[package]]
 name = "fdt"
 version = "0.1.3"
@@ -347,22 +268,40 @@ source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "b643857cf70949306b81d7e92cb9d47add673868edac9863c4a49c42feaf3f1e"
 
 [[package]]
-name = "getrandom"
+name = "gdbstub"
-version = "0.2.3"
+version = "0.6.1"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "7fcd999463524c52659517fe2cea98493cfe485d10565e7b0fb07dbba7ad2753"
+checksum = "9fa2ca5d6b045de372cef3991f873389b421b4cabcfbe52f7787fae8b8b37906"
+dependencies = [
+ "bitflags",
+ "cfg-if",
+ "log",
+ "managed",
+ "num-traits",
+ "paste",
+]
+
+[[package]]
+name = "gdbstub_arch"
+version = "0.2.2"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "51dc4b5718ac76d21e8605c0966dd32d80273b89b11e6cfef467b04e45934d37"
+dependencies = [
+ "gdbstub",
+ "num-traits",
+]
+
+[[package]]
+name = "getrandom"
+version = "0.2.5"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "d39cd93900197114fa1fcb7ae84ca742095eed9442088988ae74fa744e930e77"
 dependencies = [
  "cfg-if",
  "libc",
  "wasi",
 ]
 
-[[package]]
-name = "gimli"
-version = "0.25.0"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "f0a01e0497841a3b2db4f8afa483cce65f7e96a3498bd6c541734792aeac8fe7"
-
 [[package]]
 name = "glob"
 version = "0.3.0"
@@ -370,10 +309,16 @@ source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "9b919933a397b79c37e33b77bb2aa3dc8eb6e165ad809e58ff75bc7db2e34574"
 
 [[package]]
-name = "hermit-abi"
+name = "hashbrown"
-version = "0.1.19"
+version = "0.11.2"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "62b467343b94ba476dcb2500d242dadbb39557df889310ac77c5d99100aaac33"
+checksum = "ab5ef0d4909ef3724cc8cce6ccc8572c5c817592e9285f5464f8e86f8bd3726e"
+
+[[package]]
+name = "hermit-abi"
+version = "0.1.20"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "c7a30908dbce072eca83216eab939d2290080e00ca71611b96a09e5cdce5f3fa"
 dependencies = [
  "libc",
 ]
@@ -408,19 +353,29 @@ dependencies = [
 
 [[package]]
 name = "iced-x86"
-version = "1.15.0"
+version = "1.17.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "46e977036f7f5139d580c7f19ad62df9cb8ebd8410bb569e73585226be80a86f"
+checksum = "158f5204401d08f91d19176112146d75e99b3cf745092e268fa7be33e09adcec"
 dependencies = [
  "lazy_static",
  "static_assertions",
 ]
 
 [[package]]
-name = "instant"
+name = "indexmap"
-version = "0.1.11"
+version = "1.8.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "716d3d89f35ac6a34fd0eed635395f4c3b76fa889338a4632e5231a8684216bd"
+checksum = "282a6247722caba404c065016bbfa522806e51714c34f5dfc3e4a3a46fcb4223"
+dependencies = [
+ "autocfg",
+ "hashbrown",
+]
+
+[[package]]
+name = "instant"
+version = "0.1.12"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "7a5bbe824c507c5da5956355e86a746d82e0e1464f65d862cc5e71da70e94b2c"
 dependencies = [
  "cfg-if",
 ]
@@ -446,14 +401,14 @@ dependencies = [
 
 [[package]]
 name = "itoa"
-version = "0.4.8"
+version = "1.0.1"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "b71991ff56294aa922b450139ee08b3bfc70982c6b2c7562771375cf73542dd4"
+checksum = "1aab8fc367588b89dcee83ab0fd66b72b50b72fa1904d7095045ace2b0c81c35"
 
 [[package]]
 name = "kvm-bindings"
 version = "0.5.0"
-source = "git+https://github.com/cloud-hypervisor/kvm-bindings?branch=ch-v0.5.0#9c497710ba9968d8efe15dbae03991b16cf82e23"
+source = "git+https://github.com/cloud-hypervisor/kvm-bindings?branch=ch-v0.5.0-tdx#52e56d0e8ef0f6ea32fc0492e6a175b73617a49f"
 dependencies = [
  "serde",
  "serde_derive",
@@ -462,9 +417,8 @@ dependencies = [
 
 [[package]]
 name = "kvm-ioctls"
-version = "0.10.0"
+version = "0.11.0"
-source = "registry+https://github.com/rust-lang/crates.io-index"
+source = "git+https://github.com/rust-vmm/kvm-ioctls?branch=main#1e03e29cdfbb0cb108a98de7a78045a5a517f18e"
-checksum = "48dc14f9047df1873cf6942caccc7431d19c3d496ca7a0d162260c4cf0f64b76"
 dependencies = [
  "kvm-bindings",
  "libc",
@@ -479,9 +433,9 @@ checksum = "e2abad23fbc42b3700f2f279844dc832adb2b2eb069b2df918f455c4e18cc646"
 
 [[package]]
 name = "libc"
-version = "0.2.103"
+version = "0.2.119"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "dd8f7255a17a627354f321ef0055d63b898c6fb27eff628af4d1b66b7331edf6"
+checksum = "1bf2e165bb3457c8e098ea76f3e3bc9db55f87aa90d52d0e6be741470916aaa4"
 
 [[package]]
 name = "libssh2-sys"
@@ -520,9 +474,9 @@ dependencies = [
 
 [[package]]
 name = "lock_api"
-version = "0.4.5"
+version = "0.4.6"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "712a4d093c9976e24e7dbca41db895dabcbac38eb5f4045393d17a95bdfb1109"
+checksum = "88943dd7ef4a2e5a4bfa2753aaab3013e34ce2533d1996fb18ef591e315e2b3b"
 dependencies = [
  "scopeguard",
 ]
@@ -536,6 +490,12 @@ dependencies = [
  "cfg-if",
 ]
 
+[[package]]
+name = "managed"
+version = "0.8.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "0ca88d725a0a943b096803bd34e73a4437208b6077654cc4ecb2947a5f91618d"
+
 [[package]]
 name = "memchr"
 version = "2.4.1"
@@ -545,26 +505,16 @@ checksum = "308cc39be01b73d0d18f82a0e7b2a3df85245f84af96fdddc5d202d27e47b86a"
 [[package]]
 name = "micro_http"
 version = "0.1.0"
-source = "git+https://github.com/firecracker-microvm/micro-http?branch=main#36e59a083e76a2449e0f58e4283d201bc72fdf13"
+source = "git+https://github.com/firecracker-microvm/micro-http?branch=main#a730d86940081ad044cdfbc1285c1db6d3048392"
 dependencies = [
  "libc",
  "vmm-sys-util",
 ]
 
-[[package]]
-name = "miniz_oxide"
-version = "0.4.4"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "a92518e98c078586bc6c934028adcca4c92a53d6a958196de835170a01d84e4b"
-dependencies = [
- "adler",
- "autocfg",
-]
-
 [[package]]
 name = "mshv-bindings"
 version = "0.1.0"
-source = "git+https://github.com/rust-vmm/mshv?branch=main#99da566389546aedb56fc3d279e01c5dbde79bce"
+source = "git+https://github.com/rust-vmm/mshv?branch=main#d241ffcacdfa4240b4a1d34b44af89dc7b113986"
 dependencies = [
  "libc",
  "serde",
@@ -576,7 +526,7 @@ dependencies = [
 [[package]]
 name = "mshv-ioctls"
 version = "0.1.0"
-source = "git+https://github.com/rust-vmm/mshv?branch=main#99da566389546aedb56fc3d279e01c5dbde79bce"
+source = "git+https://github.com/rust-vmm/mshv?branch=main#d241ffcacdfa4240b4a1d34b44af89dc7b113986"
 dependencies = [
  "libc",
  "mshv-bindings",
@@ -607,29 +557,40 @@ dependencies = [
  "versionize",
  "versionize_derive",
  "virtio-bindings",
+ "virtio-queue",
  "vm-memory",
  "vm-virtio",
  "vmm-sys-util",
 ]
 
 [[package]]
-name = "object"
+name = "num-traits"
-version = "0.26.2"
+version = "0.2.14"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "39f37e50073ccad23b6d09bcb5b263f4e76d3bb6038e4a3c08e52162ffa8abc2"
+checksum = "9a64b1ec5cda2586e284722486d802acf1f7dbdc623e2bfc57e65ca1cd099290"
 dependencies = [
- "memchr",
+ "autocfg",
+]
+
+[[package]]
+name = "openssl-src"
+version = "111.17.0+1.1.1m"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "05d6a336abd10814198f66e2a91ccd7336611f30334119ca8ce300536666fcf4"
+dependencies = [
+ "cc",
 ]
 
 [[package]]
 name = "openssl-sys"
-version = "0.9.67"
+version = "0.9.72"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "69df2d8dfc6ce3aaf44b40dec6f487d5a886516cf6879c49e98e0710f310a058"
+checksum = "7e46109c383602735fa0a2e48dd2b7c892b048e1bf69e5c3b1d804b7d9c203cb"
 dependencies = [
  "autocfg",
  "cc",
  "libc",
+ "openssl-src",
  "pkg-config",
  "vcpkg",
 ]
@@ -638,6 +599,15 @@ dependencies = [
 name = "option_parser"
 version = "0.1.0"
 
+[[package]]
+name = "os_str_bytes"
+version = "6.0.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "8e22443d1643a904602595ba1cd8f7d896afe56d26712531c5ff73a15b2fbf64"
+dependencies = [
+ "memchr",
+]
+
 [[package]]
 name = "parking_lot"
 version = "0.11.2"
@@ -663,6 +633,12 @@ dependencies = [
  "winapi",
 ]
 
+[[package]]
+name = "paste"
+version = "1.0.6"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "0744126afe1a6dd7f394cb50a716dbe086cb06e255e53d8d0185d82828358fb5"
+
 [[package]]
 name = "pci"
 version = "0.1.0"
@@ -672,6 +648,8 @@ dependencies = [
  "hypervisor",
  "libc",
  "log",
+ "serde",
+ "serde_derive",
  "thiserror",
  "versionize",
  "versionize_derive",
@@ -685,17 +663,31 @@ dependencies = [
  "vmm-sys-util",
 ]
 
+[[package]]
+name = "performance-metrics"
+version = "0.1.0"
+dependencies = [
+ "clap",
+ "dirs",
+ "serde",
+ "serde_derive",
+ "serde_json",
+ "test_infra",
+ "thiserror",
+ "wait-timeout",
+]
+
 [[package]]
 name = "pkg-config"
-version = "0.3.20"
+version = "0.3.24"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "7c9b1041b4387893b91ee6746cddfc28516aff326a3519fb2adf820932c5e6cb"
+checksum = "58893f751c9b0412871a09abd62ecd2a00298c6c83befa223ef98c52aef40cbe"
 
 [[package]]
 name = "pnet"
-version = "0.28.0"
+version = "0.29.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "4b6d2a0409666964722368ef5fb74b9f93fac11c18bef3308693c16c6733f103"
+checksum = "8750e073f82219c01e771133c64718d7685aef922da8a0d430a46aed05b6341a"
 dependencies = [
  "ipnetwork",
  "pnet_base",
@@ -707,15 +699,15 @@ dependencies = [
 
 [[package]]
 name = "pnet_base"
-version = "0.28.0"
+version = "0.29.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "25488cd551a753dcaaa6fffc9f69a7610a412dd8954425bf7ffad5f7d1156fb8"
+checksum = "8205fe084bd43a3af79b3155c19feddd62e733640498842e631a2ffe107d1538"
 
 [[package]]
 name = "pnet_datalink"
-version = "0.28.0"
+version = "0.29.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "d4d1f8ab1ef6c914cf51dc5dfe0be64088ea5f3b08bbf5a31abc70356d271198"
+checksum = "6f85aef5e52e22ff06b1b11f2eb6d52959a9e0ecad3cb3f5cc2d78cadc077f0e"
 dependencies = [
  "ipnetwork",
  "libc",
@@ -726,9 +718,9 @@ dependencies = [
 
 [[package]]
 name = "pnet_macros"
-version = "0.28.0"
+version = "0.29.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "30490e0852e58402b8fae0d39897b08a24f493023a4d6cf56b2e30f31ed57548"
+checksum = "98cc3af95fed6dc318dfede3e81320f96ad5e237c6f7c4688108b19c8e67432d"
 dependencies = [
  "proc-macro2",
  "quote",
@@ -738,18 +730,18 @@ dependencies = [
 
 [[package]]
 name = "pnet_macros_support"
-version = "0.28.0"
+version = "0.29.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "d4714e10f30cab023005adce048f2d30dd4ac4f093662abf2220855655ef8f90"
+checksum = "feaba58ba96abb218ec584d6caf0d3ff48922df05dbbeb1560553c197091b29e"
 dependencies = [
  "pnet_base",
 ]
 
 [[package]]
 name = "pnet_packet"
-version = "0.28.0"
+version = "0.29.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "8588067671d03c9f4254b2e66fecb4d8b93b5d3e703195b84f311cd137e32130"
+checksum = "f246edaaf1aaf82072d4cd38ee18bcc5dfc0464093f9ca39e4ac5962d68cf9d4"
 dependencies = [
  "glob",
  "pnet_base",
@@ -759,9 +751,9 @@ dependencies = [
 
 [[package]]
 name = "pnet_sys"
-version = "0.28.0"
+version = "0.29.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "d9a3f32b0df45515befd19eed04616f6b56a488da92afc61164ef455e955f07f"
+checksum = "028c87a5e3a48fc07df099a2025f2ef16add5993712e1494ba69a6707ee7ed06"
 dependencies = [
  "libc",
  "winapi",
@@ -769,9 +761,9 @@ dependencies = [
 
 [[package]]
 name = "pnet_transport"
-version = "0.28.0"
+version = "0.29.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "932b2916d693bcc5fa18443dc99142e0a6fd31a6ce75a511868f7174c17e2bce"
+checksum = "950f2a7961e19d22e19e84ff0a6e0955013185fe149673499662633d02b41b7a"
 dependencies = [
  "libc",
  "pnet_base",
@@ -781,9 +773,9 @@ dependencies = [
 
 [[package]]
 name = "proc-macro2"
-version = "1.0.30"
+version = "1.0.36"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "edc3358ebc67bc8b7fa0c007f945b0b18226f78437d61bec735a9eb96b61ee70"
+checksum = "c7342d5883fbccae1cc37a2353b09c87c9b0f3afd73f5fb9bba687a1f733b029"
 dependencies = [
  "unicode-xid",
 ]
@@ -801,9 +793,9 @@ dependencies = [
 
 [[package]]
 name = "quote"
-version = "1.0.10"
+version = "1.0.15"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "38bc8cc6a5f2e3655e0899c1b848643b2562f853f114bfec7be120678e3ace05"
+checksum = "864d3e96a899863136fc6e99f3d7cae289dafe43bf2c5ac19b70df7210c0a145"
 dependencies = [
  "proc-macro2",
 ]
@@ -819,9 +811,9 @@ dependencies = [
 
 [[package]]
 name = "redox_syscall"
-version = "0.2.10"
+version = "0.2.11"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "8383f39639269cde97d255a32bdb68c047337295414940c68bdd30c2e13203ff"
+checksum = "8380fe0152551244f0747b1bf41737e0f8a74f97a14ccefd1148187271634f3c"
 dependencies = [
  "bitflags",
 ]
@@ -864,26 +856,20 @@ dependencies = [
  "syn",
 ]
 
-[[package]]
-name = "rustc-demangle"
-version = "0.1.21"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "7ef03e0a2b150c7a90d01faf6254c9c48a41e95fb2a8c2ac1c6f0d2b9aefc342"
-
 [[package]]
 name = "rustc_version"
-version = "0.2.3"
+version = "0.4.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "138e3e0acb6c9fb258b19b67cb8abd63c00679d2851805ea151465464fe9030a"
+checksum = "bfa0f585226d2e68097d4f95d113b15b83a82e819ab25717ec0590d9584ef366"
 dependencies = [
  "semver",
 ]
 
 [[package]]
 name = "ryu"
-version = "1.0.5"
+version = "1.0.9"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "71d301d4193d031abdd79ff7e3dd721168a9572ef3fe51a1517aba235bd8f86e"
+checksum = "73b4b750c782965c211b42f022f59af1fbceabdd026623714f104152f1ec149f"
 
 [[package]]
 name = "scopeguard"
@@ -902,30 +888,21 @@ dependencies = [
 
 [[package]]
 name = "semver"
-version = "0.9.0"
+version = "1.0.6"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "1d7eb9ef2c18661902cc47e535f9bc51b78acd254da71d375c2f6720d9a40403"
+checksum = "a4a3381e03edd24287172047536f20cabde766e2cd3e65e6b00fb3af51c4f38d"
-dependencies = [
- "semver-parser",
-]
-
-[[package]]
-name = "semver-parser"
-version = "0.7.0"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "388a1df253eca08550bef6c72392cfe7c30914bf41df5269b68cbd6ff8f570a3"
 
 [[package]]
 name = "serde"
-version = "1.0.130"
+version = "1.0.136"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "f12d06de37cf59146fbdecab66aa99f9fe4f78722e3607577a5375d66bd0c913"
+checksum = "ce31e24b01e1e524df96f1c2fdd054405f8d7376249a5110886fb4b658484789"
 
 [[package]]
 name = "serde_derive"
-version = "1.0.130"
+version = "1.0.136"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "d7bc1a1ab1961464eae040d96713baa5a724a8152c1222492465b54322ec508b"
+checksum = "08597e7152fcd306f41838ed3e37be9eaeed2b61c42e2117266a554fab4662f9"
 dependencies = [
  "proc-macro2",
  "quote",
@@ -934,9 +911,9 @@ dependencies = [
 
 [[package]]
 name = "serde_json"
-version = "1.0.68"
+version = "1.0.79"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "0f690853975602e1bfe1ccbf50504d67174e3bcf340f23b5ea9992e0587a52d8"
+checksum = "8e8d9fa5c3b304765ce1fd9c4c8a3de2c8db365a5b91be52f186efc675681d95"
 dependencies = [
  "itoa",
  "ryu",
@@ -945,9 +922,9 @@ dependencies = [
 
 [[package]]
 name = "signal-hook"
-version = "0.3.10"
+version = "0.3.13"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "9c98891d737e271a2954825ef19e46bd16bdb98e2746f2eec4f7a4ef7946efd1"
+checksum = "647c97df271007dcea485bb74ffdb57f2e683f1306c854f468a0c244badabf2d"
 dependencies = [
  "libc",
  "signal-hook-registry",
@@ -964,9 +941,9 @@ dependencies = [
 
 [[package]]
 name = "smallvec"
-version = "1.7.0"
+version = "1.8.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "1ecab6c735a6bb4139c0caafd0cc3635748bbb3acf4550e8138122099251f309"
+checksum = "f2dd574626839106c320a323308629dcb1acfc96e32a8cba364ddc61ac23ee83"
 
 [[package]]
 name = "ssh2"
@@ -988,15 +965,15 @@ checksum = "a2eb9349b6444b326872e140eb1cf5e7c522154d69e7a0ffb0fb81c06b37543f"
 
 [[package]]
 name = "strsim"
-version = "0.8.0"
+version = "0.10.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "8ea5119cdb4c55b55d432abb513a0429384878c15dde60cc77b1c99de1a95a6a"
+checksum = "73473c0e59e6d5812c5dfe2a064a6444949f089e20eec9a2e5506596494e4623"
 
 [[package]]
 name = "syn"
-version = "1.0.80"
+version = "1.0.86"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "d010a1623fbd906d51d650a9916aaefc05ffa0e4053ff7fe601167f3e715d194"
+checksum = "8a65b3f4ffa0092e9887669db0eae07941f023991ab58ea44da8fe8e2d511c6b"
 dependencies = [
  "proc-macro2",
  "quote",
@@ -1016,30 +993,31 @@ dependencies = [
 ]
 
 [[package]]
-name = "term_size"
+name = "termcolor"
-version = "0.3.2"
+version = "1.1.3"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "1e4129646ca0ed8f45d09b929036bafad5377103edd06e50bf574b353d2b08d9"
+checksum = "bab24d30b911b2376f3a13cc2cd443142f0c81dda04c118693e35b3835757755"
+dependencies = [
+ "winapi-util",
+]
+
+[[package]]
+name = "terminal_size"
+version = "0.1.17"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "633c1a546cee861a1a6d0dc69ebeca693bf4296661ba7852b9d21d159e0506df"
 dependencies = [
  "libc",
  "winapi",
 ]
 
-[[package]]
-name = "termcolor"
-version = "1.1.2"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "2dfed899f0eb03f32ee8c6a0aabdb8a7949659e3466561fc0adf54e26d88c5f4"
-dependencies = [
- "winapi-util",
-]
-
 [[package]]
 name = "test_infra"
 version = "0.1.0"
 dependencies = [
- "dirs 3.0.2",
+ "dirs",
  "epoll",
+ "lazy_static",
  "libc",
  "ssh2",
  "vmm-sys-util",
@@ -1048,12 +1026,11 @@ dependencies = [
 
 [[package]]
 name = "textwrap"
-version = "0.11.0"
+version = "0.15.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "d326610f408c7a4eb6f51c37c330e496b08506c9457c9d34287ecc38809fb060"
+checksum = "b1141d4d61095b28419e22cb0bbf02755f5e54e0526f97f1e3d1d160e60885fb"
 dependencies = [
- "term_size",
+ "terminal_size",
- "unicode-width",
 ]
 
 [[package]]
@@ -1076,12 +1053,6 @@ dependencies = [
  "syn",
 ]
 
-[[package]]
-name = "unicode-width"
-version = "0.1.9"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "3ed742d4ea2bd1176e236172c8429aaf54486e7ac098db29ffe6529e0ce50973"
-
 [[package]]
 name = "unicode-xid"
 version = "0.2.2"
@@ -1103,12 +1074,6 @@ version = "0.2.15"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "accd4ea62f7bb7a82fe23066fb0957d48ef677f6eeb8215f372f52e48bb32426"
 
-[[package]]
-name = "vec_map"
-version = "0.8.2"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "f1bddf1187be692e79c5ffeab891132dfb0f236ed36a43c7ed39f1165ee20191"
-
 [[package]]
 name = "versionize"
 version = "0.1.6"
@@ -1138,9 +1103,9 @@ dependencies = [
 
 [[package]]
 name = "vfio-bindings"
-version = "0.2.0"
+version = "0.3.1"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "4a21f546f2bda37f5a8cfb138c87f95b8e34d2d78d6a7a92ba3785f4e08604a7"
+checksum = "43449b404c488f70507dca193debd4bea361fe8089869b947adc19720e464bce"
 dependencies = [
  "vmm-sys-util",
 ]
@@ -1148,14 +1113,16 @@ dependencies = [
 [[package]]
 name = "vfio-ioctls"
 version = "0.1.0"
-source = "git+https://github.com/rust-vmm/vfio-ioctls?branch=main#d51c1fad37be30c20385c55a217e2d90972ea31d"
+source = "git+https://github.com/rust-vmm/vfio?branch=main#f75a77c1ab6349c105bc1462a65508726b4c2e0f"
 dependencies = [
  "byteorder",
  "kvm-bindings",
  "kvm-ioctls",
+ "libc",
  "log",
  "mshv-bindings",
  "mshv-ioctls",
+ "thiserror",
  "vfio-bindings",
  "vm-memory",
  "vmm-sys-util",
@@ -1193,9 +1160,9 @@ dependencies = [
 
 [[package]]
 name = "vhost"
-version = "0.2.0"
+version = "0.3.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "2d23ddeb452fb4f837526c6298cc8a2f4948e5595b0328a3d61b5eebe51874d7"
+checksum = "3b56bf8f178fc500fe14505fca8b00dec76fc38f2304f461c8d9d7547982311d"
 dependencies = [
  "bitflags",
  "libc",
@@ -1204,16 +1171,16 @@ dependencies = [
 ]
 
 [[package]]
-name = "vhost_user_backend"
+name = "vhost-user-backend"
 version = "0.1.0"
+source = "git+https://github.com/rust-vmm/vhost-user-backend?branch=main#bbc892ba4526bdf8101252f7aa51832d1f2eeabd"
 dependencies = [
- "epoll",
  "libc",
  "log",
  "vhost",
  "virtio-bindings",
+ "virtio-queue",
  "vm-memory",
- "vm-virtio",
  "vmm-sys-util",
 ]
 
@@ -1230,7 +1197,7 @@ dependencies = [
  "option_parser",
  "qcow",
  "vhost",
- "vhost_user_backend",
+ "vhost-user-backend",
  "virtio-bindings",
  "vm-memory",
  "vmm-sys-util",
@@ -1248,7 +1215,7 @@ dependencies = [
  "net_util",
  "option_parser",
  "vhost",
- "vhost_user_backend",
+ "vhost-user-backend",
  "virtio-bindings",
  "vm-memory",
  "vmm-sys-util",
@@ -1285,6 +1252,7 @@ dependencies = [
  "versionize_derive",
  "vhost",
  "virtio-bindings",
+ "virtio-queue",
  "vm-allocator",
  "vm-device",
  "vm-memory",
@@ -1293,6 +1261,16 @@ dependencies = [
  "vmm-sys-util",
 ]
 
+[[package]]
+name = "virtio-queue"
+version = "0.1.0"
+source = "git+https://github.com/rust-vmm/vm-virtio?branch=main#bbb22d43558a76cc7cc9c262c4be870a9e682a86"
+dependencies = [
+ "log",
+ "vm-memory",
+ "vmm-sys-util",
+]
+
 [[package]]
 name = "vm-allocator"
 version = "0.1.0"
@@ -1318,14 +1296,14 @@ dependencies = [
 
 [[package]]
 name = "vm-fdt"
-version = "0.1.0"
+version = "0.2.0"
-source = "git+https://github.com/rust-vmm/vm-fdt?branch=main#9c2b252dc19f0e87546e4f7577eaf398f8e3c66a"
+source = "git+https://github.com/rust-vmm/vm-fdt?branch=main#ca35d96191f8232bd7ab8c3e72ec925bf04bd2e0"
 
 [[package]]
 name = "vm-memory"
-version = "0.6.0"
+version = "0.7.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "0a8ebcb86ca457f9d6e14cf97009f679952eba42f0113de5db596e514cd0e43b"
+checksum = "339d4349c126fdcd87e034631d7274370cf19eb0e87b33166bcd956589fc72c5"
 dependencies = [
  "arc-swap",
  "libc",
@@ -1352,6 +1330,7 @@ version = "0.1.0"
 dependencies = [
  "log",
  "virtio-bindings",
+ "virtio-queue",
  "vm-memory",
 ]
 
@@ -1366,10 +1345,11 @@ dependencies = [
  "bitflags",
  "block_util",
  "clap",
- "credibility",
  "devices",
  "epoll",
  "event_monitor",
+ "gdbstub",
+ "gdbstub_arch",
  "hypervisor",
  "lazy_static",
  "libc",
@@ -1393,6 +1373,7 @@ dependencies = [
  "vfio_user",
  "vhdx",
  "virtio-devices",
+ "virtio-queue",
  "vm-allocator",
  "vm-device",
  "vm-memory",
@@ -1461,9 +1442,9 @@ checksum = "712e227841d057c1ee1cd2fb22fa7e5a5461ae8e48fa2ca79ec42cfc1931183f"
 
 [[package]]
 name = "zerocopy"
-version = "0.6.0"
+version = "0.6.1"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "ae0f717764196a220d8c58500e3a3595e2c9054f95d66267f9fd5f6e74ad0fec"
+checksum = "332f188cc1bcf1fe1064b8c58d150f497e697f49774aa846f2dc949d9a25f236"
 dependencies = [
  "byteorder",
  "zerocopy-derive",
@@ -1471,9 +1452,9 @@ dependencies = [
 
 [[package]]
 name = "zerocopy-derive"
-version = "0.3.0"
+version = "0.3.1"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "0af017aca1fa6181f5dd7a802456fe6f7666ecdcc18d0910431f0fc89d474e51"
+checksum = "a0fbc82b82efe24da867ee52e015e58178684bd9dd64c34e66bdf21da2582a9f"
 dependencies = [
  "proc-macro2",
  "syn",

Cargo.toml

@@ -1,6 +1,6 @@
 [package]
 name = "cloud-hypervisor"
-version = "19.0.0"
+version = "22.0.0"
 authors = ["The Cloud Hypervisor Authors"]
 edition = "2018"
 default-run = "cloud-hypervisor"
@@ -8,60 +8,61 @@ build = "build.rs"
 license = "LICENSE-APACHE & LICENSE-BSD-3-Clause"
 description = "Open source Virtual Machine Monitor (VMM) that runs on top of KVM"
 homepage = "https://github.com/cloud-hypervisor/cloud-hypervisor"
+# Minimum buildable version:
+# Keep in sync with version in .github/workflows/build.yaml
+rust-version = "1.54"
 
 [profile.release]
 lto = true
 
 [dependencies]
-anyhow = "1.0.44"
+anyhow = "1.0.55"
 api_client = { path = "api_client" }
-clap = { version = "2.33.3", features = ["wrap_help"] }
+clap = { version = "3.1.5", features = ["wrap_help","cargo"] }
 epoll = "4.3.1"
 event_monitor = { path = "event_monitor" }
 hypervisor = { path = "hypervisor" }
-libc = "0.2.103"
+libc = "0.2.119"
 log = { version = "0.4.14", features = ["std"] }
 option_parser = { path = "option_parser" }
 seccompiler = "0.2.0"
-serde_json = "1.0.68"
+serde_json = "1.0.79"
-signal-hook = "0.3.10"
+signal-hook = "0.3.13"
 thiserror = "1.0.30"
 vmm = { path = "vmm" }
 vmm-sys-util = "0.9.0"
-vm-memory = "0.6.0"
+vm-memory = "0.7.0"
 
 [build-dependencies]
-clap = { version = "2.33.3", features = ["wrap_help"] }
+clap = { version = "3.1.5", features = ["wrap_help"] }
 
 # List of patched crates
 [patch.crates-io]
-kvm-bindings = { git = "https://github.com/cloud-hypervisor/kvm-bindings", branch = "ch-v0.5.0", features = ["with-serde", "fam-wrappers"] }
+kvm-bindings = { git = "https://github.com/cloud-hypervisor/kvm-bindings", branch = "ch-v0.5.0-tdx" }
+kvm-ioctls = { git = "https://github.com/rust-vmm/kvm-ioctls", branch = "main" }
 versionize_derive = { git = "https://github.com/cloud-hypervisor/versionize_derive", branch = "ch" }
+virtio-queue = { git = "https://github.com/rust-vmm/vm-virtio", branch = "main" }
 
 [dev-dependencies]
-credibility = "0.1.3"
 dirs = "4.0.0"
 lazy_static= "1.4.0"
 net_util = { path = "net_util" }
-serde_json = "1.0.68"
+serde_json = "1.0.79"
 test_infra = { path = "test_infra" }
 wait-timeout = "0.2.0"
 
 [features]
 default = ["common", "kvm"]
 # Common features for all hypervisors
-common = ["acpi", "cmos", "fwdebug", "io_uring"]
+common = ["acpi", "cmos", "fwdebug"]
 acpi = ["vmm/acpi"]
 cmos = ["vmm/cmos"]
 fwdebug = ["vmm/fwdebug"]
+gdb = ["vmm/gdb"]
 kvm = ["vmm/kvm"]
 mshv = ["vmm/mshv"]
-io_uring = ["vmm/io_uring"]
 tdx = ["vmm/tdx"]
 
-# Integration tests require a special environment to run in
-integration_tests = []
-
 [workspace]
 members = [
     "acpi_tables",
@@ -75,11 +76,11 @@ members = [
     "net_util",
     "option_parser",
     "pci",
+    "performance-metrics",
     "qcow",
     "rate_limiter",
     "vfio_user",
     "vhdx",
-    "vhost_user_backend",
     "vhost_user_block",
     "vhost_user_net",
     "virtio-devices",
@@ -89,4 +90,3 @@ members = [
     "vm-migration",
     "vm-virtio"
 ]
-exclude = ["test_infra"]

Jenkinsfile

@@ -1,9 +1,28 @@
+def runWorkers = true
 pipeline{
 	agent none
 	stages {
 		stage ('Early checks') {
 			agent { node { label 'built-in' } }
 			stages {
+				stage ('Checkout') {
+					steps {
+						checkout scm
+					}
+				}
+				stage ('Check for documentation only changes') {
+					when {
+						expression {
+							return docsFileOnly()
+						}
+					}
+					steps {
+						script {
+							runWorkers = false
+							echo "Documentation only changes, no need to run the CI"
+						}
+					}
+				}
 				stage ('Check for RFC/WIP builds') {
 					when {
   						changeRequest comparator: 'REGEXP', title: '.*(rfc|RFC|wip|WIP).*'
@@ -22,15 +41,22 @@ pipeline{
 			}
 		}
 		stage ('Build') {
-            		parallel {
+			parallel {
 				stage ('Worker build') {
 					agent { node { label 'hirsute' } }
+					when {
+						beforeAgent true
+						expression {
+							return runWorkers
+						}
+					}
 					stages {
 						stage ('Checkout') {
 							steps {
 								checkout scm
 							}
 						}
+
 						stage ('Run OpenAPI tests') {
 							steps {
 								sh "scripts/run_openapi_tests.sh"
@@ -54,6 +80,12 @@ pipeline{
 				}
 				stage ('AArch64 worker build') {
 					agent { node { label 'bionic-arm64' } }
+					when {
+						beforeAgent true
+						expression {
+							return runWorkers
+						}
+					}
 					stages {
 						stage ('Checkout') {
 							steps {
@@ -62,7 +94,7 @@ pipeline{
 						}
 						stage ('Run unit tests') {
 							steps {
-								sh "scripts/dev_cli.sh tests --unit"
+								sh "scripts/dev_cli.sh tests --unit --libc musl"
 							}
 						}
 						stage ('Run integration tests') {
@@ -71,7 +103,7 @@ pipeline{
 							}
 							steps {
 								sh "sudo modprobe openvswitch"
-								sh "scripts/dev_cli.sh tests --integration"
+								sh "scripts/dev_cli.sh tests --integration --libc musl"
 							}
 						}
 					}
@@ -84,12 +116,19 @@ pipeline{
 				}
 				stage ('Worker build (musl)') {
 					agent { node { label 'hirsute' } }
+					when {
+						beforeAgent true
+						expression {
+							return runWorkers
+						}
+					}
 					stages {
 						stage ('Checkout') {
 							steps {
 								checkout scm
 							}
 						}
+
 						stage ('Run unit tests for musl') {
 							steps {
 								sh "scripts/dev_cli.sh tests --unit --libc musl"
@@ -110,7 +149,12 @@ pipeline{
 					agent { node { label 'bionic-sgx' } }
 					when {
 						beforeAgent true
-						branch 'main'
+						allOf {
+							branch 'main'
+							expression {
+								return runWorkers
+							}
+						}
 					}
 					stages {
 						stage ('Checkout') {
@@ -146,7 +190,12 @@ pipeline{
 					agent { node { label 'bionic-vfio' } }
 					when {
 						beforeAgent true
-						branch 'main'
+						allOf {
+							branch 'main'
+							expression {
+								return runWorkers
+							}
+						}
 					}
 					stages {
 						stage ('Checkout') {
@@ -180,6 +229,12 @@ pipeline{
 				}
 				stage ('Worker build - Windows guest') {
 					agent { node { label 'hirsute' } }
+					when {
+						beforeAgent true
+						expression {
+							return runWorkers
+						}
+					}
 					environment {
         					AZURE_CONNECTION_STRING = credentials('46b4e7d6-315f-4cc1-8333-b58780863b9b')
 					}
@@ -189,11 +244,14 @@ pipeline{
 								checkout scm
 							}
 						}
+						stage ('Install azure-cli') {
+							steps {
+								installAzureCli()
+							}
+						}
 						stage ('Download assets') {
 							steps {
-								sh "sudo apt install -y azure-cli"
 								sh "mkdir ${env.HOME}/workloads"
-								sh 'az storage blob download --container-name private-images --file "$HOME/workloads/OVMF-4b47d0c6c8.fd" --name OVMF-4b47d0c6c8.fd --connection-string "$AZURE_CONNECTION_STRING"'
 								sh 'az storage blob download --container-name private-images --file "$HOME/workloads/windows-server-2019.raw" --name windows-server-2019.raw --connection-string "$AZURE_CONNECTION_STRING"'
 							}
 						}
@@ -217,6 +275,12 @@ pipeline{
 				}
 				stage ('Worker build - Live Migration') {
 					agent { node { label 'hirsute-small' } }
+					when {
+						beforeAgent true
+						expression {
+							return runWorkers
+						}
+					}
 					stages {
 						stage ('Checkout') {
 							steps {
@@ -243,6 +307,39 @@ pipeline{
 						}
 					}
 				}
+				stage ('Worker build - Metrics') {
+					agent { node { label 'focal-metrics' } }
+					when {
+						branch 'main'
+						beforeAgent true
+						expression {
+							return runWorkers
+						}
+					}
+					environment {
+						METRICS_PUBLISH_KEY = credentials('52e0945f-ce7a-43d1-87af-67d1d87cc40f')
+					}
+					stages {
+						stage ('Checkout') {
+							steps {
+								checkout scm
+							}
+						}
+						stage ('Run metrics tests') {
+							options {
+								timeout(time: 1, unit: 'HOURS')
+							}
+							steps {
+								sh 'scripts/dev_cli.sh tests --metrics -- -- --report-file /root/workloads/metrics.json'
+							}
+						}
+						stage ('Upload metrics report') {
+							steps {
+								sh 'curl -X PUT https://cloud-hypervisor-metrics.azurewebsites.net/api/publishmetrics -H "x-functions-key: $METRICS_PUBLISH_KEY" -T ~/workloads/metrics.json'
+							}
+						}
+					}
+				}
 			}
 		}
 	}
@@ -278,3 +375,22 @@ def cancelPreviousBuilds() {
 			}
 		}
 }
+
+def installAzureCli() {
+	sh "sudo apt install -y ca-certificates curl apt-transport-https lsb-release gnupg"
+	sh "curl -sL https://packages.microsoft.com/keys/microsoft.asc | gpg --dearmor | sudo tee /etc/apt/trusted.gpg.d/microsoft.gpg > /dev/null"
+	sh "echo \"deb [arch=amd64] https://packages.microsoft.com/repos/azure-cli/ hirsute main\" | sudo tee /etc/apt/sources.list.d/azure-cli.list"
+	sh "sudo apt update"
+	sh "sudo apt install -y azure-cli"
+}
+
+def boolean docsFileOnly() {
+    if (env.CHANGE_TARGET == null) {
+        return false;
+    }
+
+    return sh(
+        returnStatus: true,
+        script: "git diff --name-only origin/${env.CHANGE_TARGET}... | grep -v '\\.md'"
+    ) != 0
+}

README.md

@@ -4,6 +4,7 @@
     - [Architectures](#architectures)
     - [Guest OS](#guest-os)
 - [2. Getting Started](#2-getting-started)
+  - [Preparation](#preparation)
   - [Install prerequisites](#install-prerequisites)
   - [Clone and build](#clone-and-build)
     - [Containerized builds and tests](#containerized-builds-and-tests)
@@ -26,12 +27,19 @@
 
 # 1. What is Cloud Hypervisor?
 
-Cloud Hypervisor is an open source Virtual Machine Monitor (VMM) that runs on top of [KVM](https://www.kernel.org/doc/Documentation/virtual/kvm/api.txt) and the MSHV hypervisors .
+Cloud Hypervisor is an open source Virtual Machine Monitor (VMM) that runs on
+top of [KVM](https://www.kernel.org/doc/Documentation/virtual/kvm/api.txt)
+hypervisor and Microsoft Hypervisor (MSHV).
 
-The project focuses on exclusively running modern, cloud workloads, on top of a limited set of hardware architectures and platforms.
+The project focuses on exclusively running modern, cloud workloads, on top of
-Cloud workloads refers to those that are usually run by customers inside a cloud provider. For our purposes this means modern operating systems with most I/O handled by paravirtualised devices (i.e. virtio), no requirement for legacy devices, and 64-bit CPUs.
+a limited set of hardware architectures and platforms. Cloud workloads refers
+to those that are usually run by customers inside a cloud provider. For our
+purposes this means modern operating systems with most I/O handled by
+paravirtualised devices (i.e. virtio), no requirement for legacy devices, and
+64-bit CPUs.
 
-Cloud Hypervisor is implemented in [Rust](https://www.rust-lang.org/) and is based on the [rust-vmm](https://github.com/rust-vmm) crates.
+Cloud Hypervisor is implemented in [Rust](https://www.rust-lang.org/) and is
+based on the [rust-vmm](https://github.com/rust-vmm) crates.
 
 ## Objectives
 
@@ -50,7 +58,9 @@ Cloud Hypervisor is implemented in [Rust](https://www.rust-lang.org/) and is bas
 
 ### Architectures
 
-Cloud Hypervisor supports the `x86-64` and `AArch64` architectures. There are some small differences in functionality between the two architectures (see [#1125](https://github.com/cloud-hypervisor/cloud-hypervisor/issues/1125)).
+Cloud Hypervisor supports the `x86-64` and `AArch64` architectures. There are
+some small differences in functionality between the two architectures
+(see [#1125](https://github.com/cloud-hypervisor/cloud-hypervisor/issues/1125)).
 
 ### Guest OS
 
@@ -58,6 +68,12 @@ Cloud Hypervisor supports `64-bit Linux` and Windows 10/Windows Server 2019.
 
 # 2. Getting Started
 
+Below sections describe how to build and run Cloud Hypervisor on the `x86_64`
+platform. For getting started on the `AArch64` platform, please refer to the
+[Arm64 documentation](docs/arm64.md).
+
+## Preparation
+
 We create a folder to build and run `cloud-hypervisor` at `$HOME/cloud-hypervisor`
 
 ```shell
@@ -67,9 +83,9 @@ $ mkdir $CLOUDH
 
 ## Install prerequisites
 
-You need to install some prerequisite packages in order to build and test Cloud Hypervisor.
+You need to install some prerequisite packages in order to build and test Cloud
-Here, all the steps are based on Ubuntu, for other Linux distributions please replace the
+Hypervisor. Here, all the steps are based on Ubuntu, for other Linux
-package manager and package name.
+distributions please replace the package manager and package name.
 
 ```shell
 # Install git
@@ -100,7 +116,8 @@ $ cargo build --release --target=x86_64-unknown-linux-musl --all
 $ popd
 ```
 
-This will build a `cloud-hypervisor` binary under `$CLOUDH/cloud-hypervisor/target/release/cloud-hypervisor`.
+This will build a `cloud-hypervisor` binary under
+`$CLOUDH/cloud-hypervisor/target/release/cloud-hypervisor`.
 
 ### Containerized builds and tests
 
@@ -132,18 +149,20 @@ development script commands and their related options.
 
 ## Run
 
-You can run a guest VM by either using an existing cloud image or booting into your own kernel and disk image.
+You can run a guest VM by either using an existing cloud image or booting into
+your own kernel and disk image.
 
 ### Cloud image
 
 Cloud Hypervisor supports booting disk images containing all needed
 components to run cloud workloads, a.k.a. cloud images. To do that we rely on
 the [Rust Hypervisor
-Firmware](https://github.com/cloud-hypervisor/rust-hypervisor-firmware) project to provide
+Firmware](https://github.com/cloud-hypervisor/rust-hypervisor-firmware) project
-an ELF
+to provide an ELF formatted KVM firmware for `cloud-hypervisor` to directly
-formatted KVM firmware for `cloud-hypervisor` to directly boot into.
+boot into.
 
-We need to get the latest `rust-hypervisor-firmware` release and also a working cloud image. Here we will use a Ubuntu image:
+We need to get the latest `rust-hypervisor-firmware` release and also a working
+cloud image. Here we will use a Ubuntu image:
 
 ```shell
 $ pushd $CLOUDH
@@ -161,8 +180,7 @@ $ ./cloud-hypervisor/target/release/cloud-hypervisor \
 	--disk path=focal-server-cloudimg-amd64.raw \
 	--cpus boot=4 \
 	--memory size=1024M \
-	--net "tap=,mac=,ip=,mask=" \
+	--net "tap=,mac=,ip=,mask="
-	--rng
 $ popd
 ```
 
@@ -172,7 +190,10 @@ Multiple arguments can be given to the `--disk` parameter.
 
 #### Building your kernel
 
-Cloud Hypervisor also supports direct kernel boot into a `vmlinux` ELF kernel. In order to support virtio-iommu we have our own development branch. You are of course able to use your own kernel but these instructions will continue with the version that we develop and test against.
+Cloud Hypervisor also supports direct kernel boot into a `vmlinux` ELF kernel.
+In order to support virtio-watchdog we have our own development branch. You are
+of course able to use your own kernel but these instructions will continue with
+the version that we develop and test against.
 
 To build the kernel:
 
@@ -180,20 +201,22 @@ To build the kernel:
 
 # Clone the Cloud Hypervisor Linux branch
 $ pushd $CLOUDH
-$ git clone --depth 1 https://github.com/cloud-hypervisor/linux.git -b ch-5.14 linux-cloud-hypervisor
+$ git clone --depth 1 https://github.com/cloud-hypervisor/linux.git -b ch-5.15.12 linux-cloud-hypervisor
 $ pushd linux-cloud-hypervisor
 
 # Use the cloud-hypervisor kernel config to build your kernel
 $ cp $CLOUDH/cloud-hypervisor/resources/linux-config-x86_64 .config
-$ make bzImage -j `nproc`
+$ KCFLAGS="-Wa,-mx86-used-note=no" make bzImage -j `nproc`
 $ popd
 ```
 
-The `vmlinux` kernel image will then be located at `linux-cloud-hypervisor/arch/x86/boot/compressed/vmlinux.bin`.
+The `vmlinux` kernel image will then be located at
+`linux-cloud-hypervisor/arch/x86/boot/compressed/vmlinux.bin`.
 
 #### Disk image
 
-For the disk image, we will use a Ubuntu cloud image that contains a root partition:
+For the disk image, we will use a Ubuntu cloud image that contains a root
+partition:
 
 ```shell
 $ pushd $CLOUDH
@@ -204,8 +227,8 @@ $ popd
 
 #### Booting the guest VM
 
-Now we can directly boot into our custom kernel and make it use the Ubuntu root partition.
+Now we can directly boot into our custom kernel and make it use the Ubuntu root
-If we want to have 4 vCPUs and 1024 MBytes of memory:
+partition. If we want to have 4 vCPUs and 1024 MBytes of memory:
 
 ```shell
 $ pushd $CLOUDH
@@ -216,8 +239,7 @@ $ ./cloud-hypervisor/target/release/cloud-hypervisor \
 	--cmdline "console=hvc0 root=/dev/vda1 rw" \
 	--cpus boot=4 \
 	--memory size=1024M \
-	--net "tap=,mac=,ip=,mask=" \
+	--net "tap=,mac=,ip=,mask="
-	--rng
 ```
 
 The above example use the `virtio-console` device as the guest console, and this
@@ -236,13 +258,13 @@ $ ./cloud-hypervisor/target/release/cloud-hypervisor \
 	--cmdline "console=ttyS0 root=/dev/vda1 rw" \
 	--cpus boot=4 \
 	--memory size=1024M \
-	--net "tap=,mac=,ip=,mask=" \
+	--net "tap=,mac=,ip=,mask="
-	--rng
 ```
 
 # 3. Status
 
-Cloud Hypervisor is under active development. The following stability guarantees are currently made:
+Cloud Hypervisor is under active development. The following stability guarantees
+are currently made:
 
 * The API (including command line options) will not be removed or changed in a
   breaking way without a minimum of 2 releases notice. Where possible warnings
@@ -266,70 +288,87 @@ As of 2021-04-29, the following cloud images are supported:
 - [Ubuntu Groovy](https://cloud-images.ubuntu.com/groovy/current/) (cloudimg)
 - [Ubuntu Hirsute](https://cloud-images.ubuntu.com/hirsute/current/) (cloudimg)
 
-Direct kernel boot to userspace should work with a rootfs from most distributions.
+Direct kernel boot to userspace should work with a rootfs from most
+distributions.
 
 ## Hot Plug
 
-Cloud Hypervisor supports hotplug of CPUs, passthrough devices (VFIO), `virtio-{net,block,pmem,fs,vsock}` and memory resizing. This [document](docs/hotplug.md) details how to add devices to
+Cloud Hypervisor supports hotplug of CPUs, passthrough devices (VFIO),
-a running VM.
+`virtio-{net,block,pmem,fs,vsock}` and memory resizing. This
+[document](docs/hotplug.md) details how to add devices to a running VM.
 
 ## Device Model
 
-Details of the device model can be found in this [documentation](docs/device_model.md).
+Details of the device model can be found in this
+[documentation](docs/device_model.md).
 
 ## TODO
 
-We are not tracking the Cloud Hypervisor TODO list from a specific git tracked file but through
+We are not tracking the Cloud Hypervisor TODO list from a specific git tracked
-[github issues](https://github.com/cloud-hypervisor/cloud-hypervisor/issues/new) instead.
+file but through
+[github issues](https://github.com/cloud-hypervisor/cloud-hypervisor/issues/new)
+instead.
 
 # 4. `rust-vmm` project dependency
 
-In order to satisfy the design goal of having a high-performance, security-focused hypervisor the decision
+In order to satisfy the design goal of having a high-performance,
-was made to use the [Rust](https://www.rust-lang.org/) programming language.
+security-focused hypervisor the decision was made to use the
-The language's strong focus on memory and thread safety makes it an ideal candidate for implementing VMMs.
+[Rust](https://www.rust-lang.org/) programming language. The language's strong
+focus on memory and thread safety makes it an ideal candidate for implementing
+VMMs.
 
-Instead of implementing the VMM components from scratch, Cloud Hypervisor is importing the [rust-vmm](https://github.com/rust-vmm)
+Instead of implementing the VMM components from scratch, Cloud Hypervisor is
-crates, and sharing code and architecture together with other VMMs like e.g. Amazon's [Firecracker](https://firecracker-microvm.github.io/)
+importing the [rust-vmm](https://github.com/rust-vmm) crates, and sharing code
-and Google's [crosvm](https://chromium.googlesource.com/chromiumos/platform/crosvm/).
+and architecture together with other VMMs like e.g. Amazon's
+[Firecracker](https://firecracker-microvm.github.io/) and Google's
+[crosvm](https://chromium.googlesource.com/chromiumos/platform/crosvm/).
 
-Cloud Hypervisor embraces the rust-vmm project goals, which is to be able to share and re-use
+Cloud Hypervisor embraces the rust-vmm project goals, which is to be able to
-as many virtualization crates as possible. As such, the Cloud Hypervisor relationship with the rust-vmm
+share and re-use as many virtualization crates as possible. As such, the Cloud
-project is twofold:
+Hypervisor relationship with the rust-vmm project is twofold:
 
-1. It will use as much of the rust-vmm code as possible. Any new rust-vmm crate that's relevant to the project
+1. It will use as much of the rust-vmm code as possible. Any new rust-vmm crate
-   goals will be integrated as soon as possible.
+   that's relevant to the project goals will be integrated as soon as possible.
-2. As it is likely that the rust-vmm project will lack some of the features that Cloud Hypervisor needs (e.g. ACPI,
+2. As it is likely that the rust-vmm project will lack some of the features that
-   VFIO, vhost-user, etc), we will be using the Cloud Hypervisor VMM to implement and test them, and contribute them
+   Cloud Hypervisor needs (e.g. ACPI, VFIO, vhost-user, etc), we will be using
-   back to the rust-vmm project.
+   the Cloud Hypervisor VMM to implement and test them, and contribute them back
+   to the rust-vmm project.
 
 ## Firecracker and crosvm
 
-A large part of the Cloud Hypervisor code is based on either the Firecracker or the crosvm projects implementations.
+A large part of the Cloud Hypervisor code is based on either the Firecracker or
-Both of these are VMMs written in Rust with a focus on safety and security, like Cloud Hypervisor.
+the crosvm projects implementations. Both of these are VMMs written in Rust with
+a focus on safety and security, like Cloud Hypervisor.
 
-However we want to emphasize that the Cloud Hypervisor project is neither a fork nor a reimplementation of any of those
+However we want to emphasize that the Cloud Hypervisor project is neither a fork
-projects. The goals and use cases we're trying to meet are different.
+nor a reimplementation of any of those projects. The goals and use cases we're
-We're aiming at supporting cloud workloads, i.e. those modern, full Linux distribution images currently being run by
+trying to meet are different. We're aiming at supporting cloud workloads, i.e.
-Cloud Service Provider (CSP) tenants.
+those modern, full Linux distribution images currently being run by Cloud
+Service Provider (CSP) tenants.
 
-Our primary target is not to support client or serverless use cases, and as such our code base already diverges
+Our primary target is not to support client or serverless use cases, and as such
-from the crosvm and Firecracker ones. As we add more features to support our use cases, we believe that the divergence
+our code base already diverges from the crosvm and Firecracker ones. As we add
-will increase while at the same time sharing as much of the fundamental virtualization code through the rust-vmm project
+more features to support our use cases, we believe that the divergence will
-crates as possible.
+increase while at the same time sharing as much of the fundamental
+virtualization code through the rust-vmm project crates as possible.
 
 # 5. Community
 
-The Cloud Hypervisor project follows the governance, and community guidelines described in
+The Cloud Hypervisor project follows the governance, and community guidelines
-the [Community](https://github.com/cloud-hypervisor/community) repository.
+described in the [Community](https://github.com/cloud-hypervisor/community)
+repository.
 
 ## Contribute
 
-We are working on building a global, diverse and collaborative community around the Cloud Hypervisor project.
+We are working on building a global, diverse and collaborative community around
-Anyone who is interested in [contributing](CONTRIBUTING.md) to the project is welcome to participate.
+the Cloud Hypervisor project. Anyone who is interested in
+[contributing](CONTRIBUTING.md) to the project is welcome to participate.
 
-We believe that contributing to a open source project like Cloud Hypervisor covers a lot more than just sending
+We believe that contributing to a open source project like Cloud Hypervisor
-code. Testing, documentation, pull request reviews, bug reports, feature requests, project improvement suggestions,
+covers a lot more than just sending code. Testing, documentation, pull request
-etc, are all equal and welcome means of contribution. See the [CONTRIBUTING](CONTRIBUTING.md) document for more details.
+reviews, bug reports, feature requests, project improvement suggestions, etc,
+are all equal and welcome means of contribution. See the
+[CONTRIBUTING](CONTRIBUTING.md) document for more details.
 
 ## Join us
 
@@ -338,4 +377,4 @@ and [join us on Slack](https://cloud-hypervisor.slack.com/).
 
 ## Security issues
 
-Please use the GitHub security advisories feature for reporting issues: https://github.com/cloud-hypervisor/cloud-hypervisor/security/advisories/new
+Please contact the maintainers listed in the MAINTAINERS.md file with security issues.

acpi_tables/Cargo.toml

@@ -5,4 +5,4 @@ authors = ["The Cloud Hypervisor Authors"]
 edition = "2018"
 
 [dependencies]
-vm-memory = "0.6.0"
+vm-memory = "0.7.0"

acpi_tables/src/aml.rs

@@ -6,15 +6,23 @@
 use std::marker::PhantomData;
 
 pub trait Aml {
-    fn to_aml_bytes(&self) -> Vec<u8>;
+    fn append_aml_bytes(&self, _v: &mut Vec<u8>) {
+        unimplemented!()
+    }
+
+    fn to_aml_bytes(&self) -> Vec<u8> {
+        let mut v = Vec::new();
+        self.append_aml_bytes(&mut v);
+        v
+    }
 }
 
 pub const ZERO: Zero = Zero {};
 pub struct Zero {}
 
 impl Aml for Zero {
-    fn to_aml_bytes(&self) -> Vec<u8> {
+    fn append_aml_bytes(&self, v: &mut Vec<u8>) {
-        vec![0u8]
+        v.push(0u8)
     }
 }
 
@@ -22,8 +30,8 @@ pub const ONE: One = One {};
 pub struct One {}
 
 impl Aml for One {
-    fn to_aml_bytes(&self) -> Vec<u8> {
+    fn append_aml_bytes(&self, v: &mut Vec<u8>) {
-        vec![1u8]
+        v.push(1u8)
     }
 }
 
@@ -31,8 +39,8 @@ pub const ONES: Ones = Ones {};
 pub struct Ones {}
 
 impl Aml for Ones {
-    fn to_aml_bytes(&self) -> Vec<u8> {
+    fn append_aml_bytes(&self, v: &mut Vec<u8>) {
-        vec![0xffu8]
+        v.push(0xffu8)
     }
 }
 
@@ -42,9 +50,7 @@ pub struct Path {
 }
 
 impl Aml for Path {
-    fn to_aml_bytes(&self) -> Vec<u8> {
+    fn append_aml_bytes(&self, bytes: &mut Vec<u8>) {
-        let mut bytes = Vec::new();
-
         if self.root {
             bytes.push(b'\\');
         }
@@ -62,10 +68,8 @@ impl Aml for Path {
         };
 
         for part in self.name_parts.clone().iter_mut() {
-            bytes.append(&mut part.to_vec());
+            bytes.extend_from_slice(&part.to_vec());
         }
-
-        bytes
     }
 }
 
@@ -94,40 +98,36 @@ impl From<&str> for Path {
 pub type Byte = u8;
 
 impl Aml for Byte {
-    fn to_aml_bytes(&self) -> Vec<u8> {
+    fn append_aml_bytes(&self, bytes: &mut Vec<u8>) {
-        let mut bytes = vec![0x0a]; /* BytePrefix */
+        bytes.push(0x0a); /* BytePrefix */
         bytes.push(*self);
-        bytes
     }
 }
 
 pub type Word = u16;
 
 impl Aml for Word {
-    fn to_aml_bytes(&self) -> Vec<u8> {
+    fn append_aml_bytes(&self, bytes: &mut Vec<u8>) {
-        let mut bytes = vec![0x0bu8]; /* WordPrefix */
+        bytes.push(0x0b); /* WordPrefix */
-        bytes.append(&mut self.to_le_bytes().to_vec());
+        bytes.extend_from_slice(&self.to_le_bytes())
-        bytes
     }
 }
 
 pub type DWord = u32;
 
 impl Aml for DWord {
-    fn to_aml_bytes(&self) -> Vec<u8> {
+    fn append_aml_bytes(&self, bytes: &mut Vec<u8>) {
-        let mut bytes = vec![0x0c]; /* DWordPrefix */
+        bytes.push(0x0c); /* DWordPrefix */
-        bytes.append(&mut self.to_le_bytes().to_vec());
+        bytes.extend_from_slice(&self.to_le_bytes())
-        bytes
     }
 }
 
 pub type QWord = u64;
 
 impl Aml for QWord {
-    fn to_aml_bytes(&self) -> Vec<u8> {
+    fn append_aml_bytes(&self, bytes: &mut Vec<u8>) {
-        let mut bytes = vec![0x0e]; /* QWordPrefix */
+        bytes.push(0x0e); /* QWordPrefix */
-        bytes.append(&mut self.to_le_bytes().to_vec());
+        bytes.extend_from_slice(&self.to_le_bytes())
-        bytes
     }
 }
 
@@ -136,16 +136,18 @@ pub struct Name {
 }
 
 impl Aml for Name {
-    fn to_aml_bytes(&self) -> Vec<u8> {
+    fn append_aml_bytes(&self, bytes: &mut Vec<u8>) {
-        self.bytes.clone()
+        // TODO: Refactor this to make more efficient but there are
+        // lifetime/ownership challenges.
+        bytes.extend_from_slice(&self.bytes.clone())
     }
 }
 
 impl Name {
     pub fn new(path: Path, inner: &dyn Aml) -> Self {
         let mut bytes = vec![0x08]; /* NameOp */
-        bytes.append(&mut path.to_aml_bytes());
+        path.append_aml_bytes(&mut bytes);
-        bytes.append(&mut inner.to_aml_bytes());
+        inner.append_aml_bytes(&mut bytes);
         Name { bytes }
     }
 }
@@ -155,21 +157,17 @@ pub struct Package<'a> {
 }
 
 impl<'a> Aml for Package<'a> {
-    fn to_aml_bytes(&self) -> Vec<u8> {
+    fn append_aml_bytes(&self, bytes: &mut Vec<u8>) {
-        let mut bytes = vec![self.children.len() as u8];
+        let mut tmp = vec![self.children.len() as u8];
         for child in &self.children {
-            bytes.append(&mut child.to_aml_bytes());
+            child.append_aml_bytes(&mut tmp);
         }
 
-        let mut pkg_length = create_pkg_length(&bytes, true);
+        let pkg_length = create_pkg_length(&tmp, true);
-        pkg_length.reverse();
-        for byte in pkg_length {
-            bytes.insert(0, byte);
-        }
 
-        bytes.insert(0, 0x12); /* PackageOp */
+        bytes.push(0x12); /* PackageOp */
-
+        bytes.extend_from_slice(&pkg_length);
-        bytes
+        bytes.extend_from_slice(&tmp);
     }
 }
 
@@ -257,51 +255,46 @@ impl EisaName {
 }
 
 impl Aml for EisaName {
-    fn to_aml_bytes(&self) -> Vec<u8> {
+    fn append_aml_bytes(&self, bytes: &mut Vec<u8>) {
-        self.value.to_aml_bytes()
+        self.value.append_aml_bytes(bytes)
-    }
-}
-
-fn create_integer(v: usize) -> Vec<u8> {
-    if v <= u8::max_value().into() {
-        (v as u8).to_aml_bytes()
-    } else if v <= u16::max_value().into() {
-        (v as u16).to_aml_bytes()
-    } else if v <= u32::max_value() as usize {
-        (v as u32).to_aml_bytes()
-    } else {
-        (v as u64).to_aml_bytes()
     }
 }
 
 pub type Usize = usize;
 
 impl Aml for Usize {
-    fn to_aml_bytes(&self) -> Vec<u8> {
+    fn append_aml_bytes(&self, bytes: &mut Vec<u8>) {
-        create_integer(*self)
+        if *self <= u8::max_value().into() {
+            (*self as u8).append_aml_bytes(bytes)
+        } else if *self <= u16::max_value().into() {
+            (*self as u16).append_aml_bytes(bytes)
+        } else if *self <= u32::max_value() as usize {
+            (*self as u32).append_aml_bytes(bytes)
+        } else {
+            (*self as u64).append_aml_bytes(bytes)
+        }
     }
 }
 
-fn create_aml_string(v: &str) -> Vec<u8> {
+fn append_aml_string(v: &str, bytes: &mut Vec<u8>) {
-    let mut data = vec![0x0D]; /* String Op */
+    bytes.push(0x0D); /* String Op */
-    data.extend_from_slice(v.as_bytes());
+    bytes.extend_from_slice(v.as_bytes());
-    data.push(0x0); /* NullChar */
+    bytes.push(0x0); /* NullChar */
-    data
 }
 
 pub type AmlStr = &'static str;
 
 impl Aml for AmlStr {
-    fn to_aml_bytes(&self) -> Vec<u8> {
+    fn append_aml_bytes(&self, bytes: &mut Vec<u8>) {
-        create_aml_string(self)
+        append_aml_string(self, bytes)
     }
 }
 
 pub type AmlString = String;
 
 impl Aml for AmlString {
-    fn to_aml_bytes(&self) -> Vec<u8> {
+    fn append_aml_bytes(&self, bytes: &mut Vec<u8>) {
-        create_aml_string(self)
+        append_aml_string(self, bytes)
     }
 }
 
@@ -310,36 +303,31 @@ pub struct ResourceTemplate<'a> {
 }
 
 impl<'a> Aml for ResourceTemplate<'a> {
-    fn to_aml_bytes(&self) -> Vec<u8> {
+    fn append_aml_bytes(&self, bytes: &mut Vec<u8>) {
-        let mut bytes = Vec::new();
+        let mut tmp = Vec::new();
-
         // Add buffer data
         for child in &self.children {
-            bytes.append(&mut child.to_aml_bytes());
+            child.append_aml_bytes(&mut tmp);
         }
 
         // Mark with end and mark checksum as as always valid
-        bytes.push(0x79); /* EndTag */
+        tmp.push(0x79); /* EndTag */
-        bytes.push(0); /* zero checksum byte */
+        tmp.push(0); /* zero checksum byte */
 
         // Buffer length is an encoded integer including buffer data
         // and EndTag and checksum byte
-        let mut buffer_length = bytes.len().to_aml_bytes();
+        let mut buffer_length = tmp.len().to_aml_bytes();
         buffer_length.reverse();
         for byte in buffer_length {
-            bytes.insert(0, byte);
+            tmp.insert(0, byte);
         }
 
         // PkgLength is everything else
-        let mut pkg_length = create_pkg_length(&bytes, true);
+        let pkg_length = create_pkg_length(&tmp, true);
-        pkg_length.reverse();
-        for byte in pkg_length {
-            bytes.insert(0, byte);
-        }
 
-        bytes.insert(0, 0x11); /* BufferOp */
+        bytes.push(0x11); /* BufferOp */
-
+        bytes.extend_from_slice(&pkg_length);
-        bytes
+        bytes.extend_from_slice(&tmp);
     }
 }
 
@@ -366,15 +354,13 @@ impl Memory32Fixed {
 }
 
 impl Aml for Memory32Fixed {
-    fn to_aml_bytes(&self) -> Vec<u8> {
+    fn append_aml_bytes(&self, bytes: &mut Vec<u8>) {
-        let mut bytes = vec![0x86]; /* Memory32Fixed */
+        bytes.push(0x86); /* Memory32Fixed */
-        bytes.append(&mut 9u16.to_le_bytes().to_vec());
+        bytes.extend_from_slice(&9u16.to_le_bytes());
-
         // 9 bytes of payload
         bytes.push(self.read_write as u8);
-        bytes.append(&mut self.base.to_le_bytes().to_vec());
+        bytes.extend_from_slice(&self.base.to_le_bytes());
-        bytes.append(&mut self.length.to_le_bytes().to_vec());
+        bytes.extend_from_slice(&self.length.to_le_bytes());
-        bytes
     }
 }
 
@@ -430,7 +416,7 @@ impl<T> AddressSpace<T> {
 
     fn push_header(&self, bytes: &mut Vec<u8>, descriptor: u8, length: usize) {
         bytes.push(descriptor); /* Word Address Space Descriptor */
-        bytes.append(&mut (length as u16).to_le_bytes().to_vec());
+        bytes.extend_from_slice(&(length as u16).to_le_bytes());
         bytes.push(self.r#type as u8); /* type */
         let generic_flags = 1 << 2 /* Min Fixed */ | 1 << 3; /* Max Fixed */
         bytes.push(generic_flags);
@@ -439,65 +425,53 @@ impl<T> AddressSpace<T> {
 }
 
 impl Aml for AddressSpace<u16> {
-    fn to_aml_bytes(&self) -> Vec<u8> {
+    fn append_aml_bytes(&self, bytes: &mut Vec<u8>) {
-        let mut bytes = Vec::new();
-
         self.push_header(
-            &mut bytes,
+            bytes,
             0x88,                               /* Word Address Space Descriptor */
             3 + 5 * std::mem::size_of::<u16>(), /* 3 bytes of header + 5 u16 fields */
         );
 
-        bytes.append(&mut 0u16.to_le_bytes().to_vec()); /* Granularity */
+        bytes.extend_from_slice(&0u16.to_le_bytes()); /* Granularity */
-        bytes.append(&mut self.min.to_le_bytes().to_vec()); /* Min */
+        bytes.extend_from_slice(&self.min.to_le_bytes()); /* Min */
-        bytes.append(&mut self.max.to_le_bytes().to_vec()); /* Max */
+        bytes.extend_from_slice(&self.max.to_le_bytes()); /* Max */
-        bytes.append(&mut 0u16.to_le_bytes().to_vec()); /* Translation */
+        bytes.extend_from_slice(&0u16.to_le_bytes()); /* Translation */
         let len = self.max - self.min + 1;
-        bytes.append(&mut len.to_le_bytes().to_vec()); /* Length */
+        bytes.extend_from_slice(&len.to_le_bytes()); /* Length */
-
-        bytes
     }
 }
 
 impl Aml for AddressSpace<u32> {
-    fn to_aml_bytes(&self) -> Vec<u8> {
+    fn append_aml_bytes(&self, bytes: &mut Vec<u8>) {
-        let mut bytes = Vec::new();
-
         self.push_header(
-            &mut bytes,
+            bytes,
             0x87,                               /* DWord Address Space Descriptor */
             3 + 5 * std::mem::size_of::<u32>(), /* 3 bytes of header + 5 u32 fields */
         );
 
-        bytes.append(&mut 0u32.to_le_bytes().to_vec()); /* Granularity */
+        bytes.extend_from_slice(&0u32.to_le_bytes()); /* Granularity */
-        bytes.append(&mut self.min.to_le_bytes().to_vec()); /* Min */
+        bytes.extend_from_slice(&self.min.to_le_bytes()); /* Min */
-        bytes.append(&mut self.max.to_le_bytes().to_vec()); /* Max */
+        bytes.extend_from_slice(&self.max.to_le_bytes()); /* Max */
-        bytes.append(&mut 0u32.to_le_bytes().to_vec()); /* Translation */
+        bytes.extend_from_slice(&0u32.to_le_bytes()); /* Translation */
         let len = self.max - self.min + 1;
-        bytes.append(&mut len.to_le_bytes().to_vec()); /* Length */
+        bytes.extend_from_slice(&len.to_le_bytes()); /* Length */
-
-        bytes
     }
 }
 
 impl Aml for AddressSpace<u64> {
-    fn to_aml_bytes(&self) -> Vec<u8> {
+    fn append_aml_bytes(&self, bytes: &mut Vec<u8>) {
-        let mut bytes = Vec::new();
-
         self.push_header(
-            &mut bytes,
+            bytes,
             0x8A,                               /* QWord Address Space Descriptor */
             3 + 5 * std::mem::size_of::<u64>(), /* 3 bytes of header + 5 u64 fields */
         );
 
-        bytes.append(&mut 0u64.to_le_bytes().to_vec()); /* Granularity */
+        bytes.extend_from_slice(&0u64.to_le_bytes()); /* Granularity */
-        bytes.append(&mut self.min.to_le_bytes().to_vec()); /* Min */
+        bytes.extend_from_slice(&self.min.to_le_bytes()); /* Min */
-        bytes.append(&mut self.max.to_le_bytes().to_vec()); /* Max */
+        bytes.extend_from_slice(&self.max.to_le_bytes()); /* Max */
-        bytes.append(&mut 0u64.to_le_bytes().to_vec()); /* Translation */
+        bytes.extend_from_slice(&0u64.to_le_bytes()); /* Translation */
         let len = self.max - self.min + 1;
-        bytes.append(&mut len.to_le_bytes().to_vec()); /* Length */
+        bytes.extend_from_slice(&len.to_le_bytes()); /* Length */
-
-        bytes
     }
 }
 
@@ -520,16 +494,13 @@ impl Io {
 }
 
 impl Aml for Io {
-    fn to_aml_bytes(&self) -> Vec<u8> {
+    fn append_aml_bytes(&self, bytes: &mut Vec<u8>) {
-        let mut bytes = vec![0x47]; /* Io Port Descriptor */
+        bytes.push(0x47); /* Io Port Descriptor */
-
         bytes.push(1); /* IODecode16 */
-        bytes.append(&mut self.min.to_le_bytes().to_vec());
+        bytes.extend_from_slice(&self.min.to_le_bytes());
-        bytes.append(&mut self.max.to_le_bytes().to_vec());
+        bytes.extend_from_slice(&self.max.to_le_bytes());
         bytes.push(self.alignment);
         bytes.push(self.length);
-
-        bytes
     }
 }
 
@@ -560,18 +531,16 @@ impl Interrupt {
 }
 
 impl Aml for Interrupt {
-    fn to_aml_bytes(&self) -> Vec<u8> {
+    fn append_aml_bytes(&self, bytes: &mut Vec<u8>) {
-        let mut bytes = vec![0x89]; /* Extended IRQ Descriptor */
+        bytes.push(0x89); /* Extended IRQ Descriptor */
-        bytes.append(&mut 6u16.to_le_bytes().to_vec());
+        bytes.extend_from_slice(&6u16.to_le_bytes());
         let flags = (self.shared as u8) << 3
             | (self.active_low as u8) << 2
             | (self.edge_triggered as u8) << 1
             | self.consumer as u8;
         bytes.push(flags);
         bytes.push(1u8); /* count */
-        bytes.append(&mut self.number.to_le_bytes().to_vec());
+        bytes.extend_from_slice(&self.number.to_le_bytes());
-
-        bytes
     }
 }
 
@@ -581,22 +550,20 @@ pub struct Device<'a> {
 }
 
 impl<'a> Aml for Device<'a> {
-    fn to_aml_bytes(&self) -> Vec<u8> {
+    fn append_aml_bytes(&self, bytes: &mut Vec<u8>) {
-        let mut bytes = Vec::new();
+        let mut tmp = Vec::new();
-        bytes.append(&mut self.path.to_aml_bytes());
+        self.path.append_aml_bytes(&mut tmp);
+
         for child in &self.children {
-            bytes.append(&mut child.to_aml_bytes());
+            child.append_aml_bytes(&mut tmp);
         }
 
-        let mut pkg_length = create_pkg_length(&bytes, true);
+        let pkg_length = create_pkg_length(&tmp, true);
-        pkg_length.reverse();
-        for byte in pkg_length {
-            bytes.insert(0, byte);
-        }
 
-        bytes.insert(0, 0x82); /* DeviceOp */
+        bytes.push(0x5b); /* ExtOpPrefix */
-        bytes.insert(0, 0x5b); /* ExtOpPrefix */
+        bytes.push(0x82); /* DeviceOp */
-        bytes
+        bytes.extend_from_slice(&pkg_length);
+        bytes.extend_from_slice(&tmp);
     }
 }
 
@@ -612,21 +579,18 @@ pub struct Scope<'a> {
 }
 
 impl<'a> Aml for Scope<'a> {
-    fn to_aml_bytes(&self) -> Vec<u8> {
+    fn append_aml_bytes(&self, bytes: &mut Vec<u8>) {
-        let mut bytes = Vec::new();
+        let mut tmp = Vec::new();
-        bytes.append(&mut self.path.to_aml_bytes());
+        self.path.append_aml_bytes(&mut tmp);
         for child in &self.children {
-            bytes.append(&mut child.to_aml_bytes());
+            child.append_aml_bytes(&mut tmp);
         }
 
-        let mut pkg_length = create_pkg_length(&bytes, true);
+        let pkg_length = create_pkg_length(&tmp, true);
-        pkg_length.reverse();
-        for byte in pkg_length {
-            bytes.insert(0, byte);
-        }
 
-        bytes.insert(0, 0x10); /* ScopeOp */
+        bytes.push(0x10); /* ScopeOp */
-        bytes
+        bytes.extend_from_slice(&pkg_length);
+        bytes.extend_from_slice(&tmp)
     }
 }
 
@@ -655,23 +619,20 @@ impl<'a> Method<'a> {
 }
 
 impl<'a> Aml for Method<'a> {
-    fn to_aml_bytes(&self) -> Vec<u8> {
+    fn append_aml_bytes(&self, bytes: &mut Vec<u8>) {
-        let mut bytes = Vec::new();
+        let mut tmp = Vec::new();
-        bytes.append(&mut self.path.to_aml_bytes());
+        self.path.append_aml_bytes(&mut tmp);
         let flags: u8 = (self.args & 0x7) | (self.serialized as u8) << 3;
-        bytes.push(flags);
+        tmp.push(flags);
         for child in &self.children {
-            bytes.append(&mut child.to_aml_bytes());
+            child.append_aml_bytes(&mut tmp);
         }
 
-        let mut pkg_length = create_pkg_length(&bytes, true);
+        let pkg_length = create_pkg_length(&tmp, true);
-        pkg_length.reverse();
-        for byte in pkg_length {
-            bytes.insert(0, byte);
-        }
 
-        bytes.insert(0, 0x14); /* MethodOp */
+        bytes.push(0x14); /* MethodOp */
-        bytes
+        bytes.extend_from_slice(&pkg_length);
+        bytes.extend_from_slice(&tmp)
     }
 }
 
@@ -686,10 +647,9 @@ impl<'a> Return<'a> {
 }
 
 impl<'a> Aml for Return<'a> {
-    fn to_aml_bytes(&self) -> Vec<u8> {
+    fn append_aml_bytes(&self, bytes: &mut Vec<u8>) {
-        let mut bytes = vec![0xa4]; /* ReturnOp */
+        bytes.push(0xa4); /* ReturnOp */
-        bytes.append(&mut self.value.to_aml_bytes());
+        self.value.append_aml_bytes(bytes);
-        bytes
     }
 }
 
@@ -740,35 +700,32 @@ impl Field {
 }
 
 impl Aml for Field {
-    fn to_aml_bytes(&self) -> Vec<u8> {
+    fn append_aml_bytes(&self, bytes: &mut Vec<u8>) {
-        let mut bytes = Vec::new();
+        let mut tmp = Vec::new();
-        bytes.append(&mut self.path.to_aml_bytes());
+        self.path.append_aml_bytes(&mut tmp);
 
         let flags: u8 = self.access_type as u8 | (self.update_rule as u8) << 5;
-        bytes.push(flags);
+        tmp.push(flags);
 
         for field in self.fields.iter() {
             match field {
                 FieldEntry::Named(name, length) => {
-                    bytes.extend_from_slice(name);
+                    tmp.extend_from_slice(name);
-                    bytes.append(&mut create_pkg_length(&vec![0; *length], false));
+                    tmp.extend_from_slice(&create_pkg_length(&vec![0; *length], false));
                 }
                 FieldEntry::Reserved(length) => {
-                    bytes.push(0x0);
+                    tmp.push(0x0);
-                    bytes.append(&mut create_pkg_length(&vec![0; *length], false));
+                    tmp.extend_from_slice(&create_pkg_length(&vec![0; *length], false));
                 }
             }
         }
 
-        let mut pkg_length = create_pkg_length(&bytes, true);
+        let pkg_length = create_pkg_length(&tmp, true);
-        pkg_length.reverse();
-        for byte in pkg_length {
-            bytes.insert(0, byte);
-        }
 
-        bytes.insert(0, 0x81); /* FieldOp */
+        bytes.push(0x5b); /* ExtOpPrefix */
-        bytes.insert(0, 0x5b); /* ExtOpPrefix */
+        bytes.push(0x81); /* FieldOp */
-        bytes
+        bytes.extend_from_slice(&pkg_length);
+        bytes.extend_from_slice(&tmp)
     }
 }
 
@@ -805,15 +762,13 @@ impl OpRegion {
 }
 
 impl Aml for OpRegion {
-    fn to_aml_bytes(&self) -> Vec<u8> {
+    fn append_aml_bytes(&self, bytes: &mut Vec<u8>) {
-        let mut bytes = Vec::new();
+        bytes.push(0x5b); /* ExtOpPrefix */
-        bytes.append(&mut self.path.to_aml_bytes());
+        bytes.push(0x80); /* OpRegionOp */
+        self.path.append_aml_bytes(bytes);
         bytes.push(self.space as u8);
-        bytes.extend_from_slice(&self.offset.to_aml_bytes()); /* RegionOffset */
+        self.offset.append_aml_bytes(bytes); /* RegionOffset */
-        bytes.extend_from_slice(&self.length.to_aml_bytes()); /* RegionLen */
+        self.length.append_aml_bytes(bytes); /* RegionLen */
-        bytes.insert(0, 0x80); /* OpRegionOp */
-        bytes.insert(0, 0x5b); /* ExtOpPrefix */
-        bytes
     }
 }
 
@@ -832,21 +787,18 @@ impl<'a> If<'a> {
 }
 
 impl<'a> Aml for If<'a> {
-    fn to_aml_bytes(&self) -> Vec<u8> {
+    fn append_aml_bytes(&self, bytes: &mut Vec<u8>) {
-        let mut bytes = Vec::new();
+        let mut tmp = Vec::new();
-        bytes.extend_from_slice(&self.predicate.to_aml_bytes());
+        self.predicate.append_aml_bytes(&mut tmp);
         for child in self.if_children.iter() {
-            bytes.extend_from_slice(&child.to_aml_bytes());
+            child.append_aml_bytes(&mut tmp);
         }
 
-        let mut pkg_length = create_pkg_length(&bytes, true);
+        let pkg_length = create_pkg_length(&tmp, true);
-        pkg_length.reverse();
-        for byte in pkg_length {
-            bytes.insert(0, byte);
-        }
 
-        bytes.insert(0, 0xa0); /* IfOp */
+        bytes.push(0xa0); /* IfOp */
-        bytes
+        bytes.extend_from_slice(&pkg_length);
+        bytes.extend_from_slice(&tmp);
     }
 }
 
@@ -862,11 +814,10 @@ impl<'a> Equal<'a> {
 }
 
 impl<'a> Aml for Equal<'a> {
-    fn to_aml_bytes(&self) -> Vec<u8> {
+    fn append_aml_bytes(&self, bytes: &mut Vec<u8>) {
-        let mut bytes = vec![0x93]; /* LEqualOp */
+        bytes.push(0x93); /* LEqualOp */
-        bytes.extend_from_slice(&self.left.to_aml_bytes());
+        self.left.append_aml_bytes(bytes);
-        bytes.extend_from_slice(&self.right.to_aml_bytes());
+        self.right.append_aml_bytes(bytes);
-        bytes
     }
 }
 
@@ -882,33 +833,28 @@ impl<'a> LessThan<'a> {
 }
 
 impl<'a> Aml for LessThan<'a> {
-    fn to_aml_bytes(&self) -> Vec<u8> {
+    fn append_aml_bytes(&self, bytes: &mut Vec<u8>) {
-        let mut bytes = vec![0x95]; /* LLessOp */
+        bytes.push(0x95); /* LLessOp */
-        bytes.extend_from_slice(&self.left.to_aml_bytes());
+        self.left.append_aml_bytes(bytes);
-        bytes.extend_from_slice(&self.right.to_aml_bytes());
+        self.right.append_aml_bytes(bytes);
-        bytes
     }
 }
 
 pub struct Arg(pub u8);
 
 impl Aml for Arg {
-    fn to_aml_bytes(&self) -> Vec<u8> {
+    fn append_aml_bytes(&self, bytes: &mut Vec<u8>) {
-        let mut bytes = Vec::new();
         assert!(self.0 <= 6);
         bytes.push(0x68 + self.0); /* Arg0Op */
-        bytes
     }
 }
 
 pub struct Local(pub u8);
 
 impl Aml for Local {
-    fn to_aml_bytes(&self) -> Vec<u8> {
+    fn append_aml_bytes(&self, bytes: &mut Vec<u8>) {
-        let mut bytes = Vec::new();
         assert!(self.0 <= 7);
         bytes.push(0x60 + self.0); /* Local0Op */
-        bytes
     }
 }
 
@@ -924,11 +870,10 @@ impl<'a> Store<'a> {
 }
 
 impl<'a> Aml for Store<'a> {
-    fn to_aml_bytes(&self) -> Vec<u8> {
+    fn append_aml_bytes(&self, bytes: &mut Vec<u8>) {
-        let mut bytes = vec![0x70]; /* StoreOp */
+        bytes.push(0x70); /* StoreOp */
-        bytes.extend_from_slice(&self.value.to_aml_bytes());
+        self.value.append_aml_bytes(bytes);
-        bytes.extend_from_slice(&self.name.to_aml_bytes());
+        self.name.append_aml_bytes(bytes);
-        bytes
     }
 }
 
@@ -944,12 +889,11 @@ impl Mutex {
 }
 
 impl Aml for Mutex {
-    fn to_aml_bytes(&self) -> Vec<u8> {
+    fn append_aml_bytes(&self, bytes: &mut Vec<u8>) {
-        let mut bytes = vec![0x5b]; /* ExtOpPrefix */
+        bytes.push(0x5b); /* ExtOpPrefix */
         bytes.push(0x01); /* MutexOp */
-        bytes.extend_from_slice(&self.path.to_aml_bytes());
+        self.path.append_aml_bytes(bytes);
         bytes.push(self.sync_level);
-        bytes
     }
 }
 
@@ -965,12 +909,11 @@ impl Acquire {
 }
 
 impl Aml for Acquire {
-    fn to_aml_bytes(&self) -> Vec<u8> {
+    fn append_aml_bytes(&self, bytes: &mut Vec<u8>) {
-        let mut bytes = vec![0x5b]; /* ExtOpPrefix */
+        bytes.push(0x5b); /* ExtOpPrefix */
         bytes.push(0x23); /* AcquireOp */
-        bytes.extend_from_slice(&self.mutex.to_aml_bytes());
+        self.mutex.append_aml_bytes(bytes);
         bytes.extend_from_slice(&self.timeout.to_le_bytes());
-        bytes
     }
 }
 
@@ -985,11 +928,10 @@ impl Release {
 }
 
 impl Aml for Release {
-    fn to_aml_bytes(&self) -> Vec<u8> {
+    fn append_aml_bytes(&self, bytes: &mut Vec<u8>) {
-        let mut bytes = vec![0x5b]; /* ExtOpPrefix */
+        bytes.push(0x5b); /* ExtOpPrefix */
         bytes.push(0x27); /* ReleaseOp */
-        bytes.extend_from_slice(&self.mutex.to_aml_bytes());
+        self.mutex.append_aml_bytes(bytes);
-        bytes
     }
 }
 
@@ -1005,11 +947,10 @@ impl<'a> Notify<'a> {
 }
 
 impl<'a> Aml for Notify<'a> {
-    fn to_aml_bytes(&self) -> Vec<u8> {
+    fn append_aml_bytes(&self, bytes: &mut Vec<u8>) {
-        let mut bytes = vec![0x86]; /* NotifyOp */
+        bytes.push(0x86); /* NotifyOp */
-        bytes.extend_from_slice(&self.object.to_aml_bytes());
+        self.object.append_aml_bytes(bytes);
-        bytes.extend_from_slice(&self.value.to_aml_bytes());
+        self.value.append_aml_bytes(bytes);
-        bytes
     }
 }
 
@@ -1028,21 +969,18 @@ impl<'a> While<'a> {
 }
 
 impl<'a> Aml for While<'a> {
-    fn to_aml_bytes(&self) -> Vec<u8> {
+    fn append_aml_bytes(&self, bytes: &mut Vec<u8>) {
-        let mut bytes = Vec::new();
+        let mut tmp = Vec::new();
-        bytes.extend_from_slice(&self.predicate.to_aml_bytes());
+        self.predicate.append_aml_bytes(&mut tmp);
         for child in self.while_children.iter() {
-            bytes.extend_from_slice(&child.to_aml_bytes());
+            child.append_aml_bytes(&mut tmp)
         }
 
-        let mut pkg_length = create_pkg_length(&bytes, true);
+        let pkg_length = create_pkg_length(&tmp, true);
-        pkg_length.reverse();
-        for byte in pkg_length {
-            bytes.insert(0, byte);
-        }
 
-        bytes.insert(0, 0xa2); /* WhileOp */
+        bytes.push(0xa2); /* WhileOp */
-        bytes
+        bytes.extend_from_slice(&pkg_length);
+        bytes.extend_from_slice(&tmp);
     }
 }
 
@@ -1061,12 +999,11 @@ macro_rules! binary_op {
         }
 
         impl<'a> Aml for $name<'a> {
-            fn to_aml_bytes(&self) -> Vec<u8> {
+            fn append_aml_bytes(&self, bytes: &mut Vec<u8>) {
-                let mut bytes = vec![$opcode]; /* Op for the binary operator */
+                bytes.push($opcode); /* Op for the binary operator */
-                bytes.extend_from_slice(&self.a.to_aml_bytes());
+                self.a.append_aml_bytes(bytes);
-                bytes.extend_from_slice(&self.b.to_aml_bytes());
+                self.b.append_aml_bytes(bytes);
-                bytes.extend_from_slice(&self.target.to_aml_bytes());
+                self.target.append_aml_bytes(bytes);
-                bytes
             }
         }
     };
@@ -1101,13 +1038,11 @@ impl<'a> MethodCall<'a> {
 }
 
 impl<'a> Aml for MethodCall<'a> {
-    fn to_aml_bytes(&self) -> Vec<u8> {
+    fn append_aml_bytes(&self, bytes: &mut Vec<u8>) {
-        let mut bytes = Vec::new();
+        self.name.append_aml_bytes(bytes);
-        bytes.extend_from_slice(&self.name.to_aml_bytes());
         for arg in self.args.iter() {
-            bytes.extend_from_slice(&arg.to_aml_bytes());
+            arg.append_aml_bytes(bytes);
         }
-        bytes
     }
 }
 
@@ -1122,20 +1057,16 @@ impl Buffer {
 }
 
 impl Aml for Buffer {
-    fn to_aml_bytes(&self) -> Vec<u8> {
+    fn append_aml_bytes(&self, bytes: &mut Vec<u8>) {
-        let mut bytes = Vec::new();
+        let mut tmp = Vec::new();
-        bytes.extend_from_slice(&self.data.len().to_aml_bytes());
+        self.data.len().append_aml_bytes(&mut tmp);
-        bytes.extend_from_slice(&self.data);
+        tmp.extend_from_slice(&self.data);
 
-        let mut pkg_length = create_pkg_length(&bytes, true);
+        let pkg_length = create_pkg_length(&tmp, true);
-        pkg_length.reverse();
-        for byte in pkg_length {
-            bytes.insert(0, byte);
-        }
 
-        bytes.insert(0, 0x11); /* BufferOp */
+        bytes.push(0x11); /* BufferOp */
-
+        bytes.extend_from_slice(&pkg_length);
-        bytes
+        bytes.extend_from_slice(&tmp);
     }
 }
 
@@ -1158,22 +1089,20 @@ impl<'a, T> CreateField<'a, T> {
 }
 
 impl<'a> Aml for CreateField<'a, u64> {
-    fn to_aml_bytes(&self) -> Vec<u8> {
+    fn append_aml_bytes(&self, bytes: &mut Vec<u8>) {
-        let mut bytes = vec![0x8f]; /* CreateQWordFieldOp */
+        bytes.push(0x8f); /* CreateQWordFieldOp */
-        bytes.extend_from_slice(&self.buffer.to_aml_bytes());
+        self.buffer.append_aml_bytes(bytes);
-        bytes.extend_from_slice(&self.offset.to_aml_bytes());
+        self.offset.append_aml_bytes(bytes);
-        bytes.extend_from_slice(&self.field.to_aml_bytes());
+        self.field.append_aml_bytes(bytes);
-        bytes
     }
 }
 
 impl<'a> Aml for CreateField<'a, u32> {
-    fn to_aml_bytes(&self) -> Vec<u8> {
+    fn append_aml_bytes(&self, bytes: &mut Vec<u8>) {
-        let mut bytes = vec![0x8a]; /* CreateDWordFieldOp */
+        bytes.push(0x8a); /* CreateDWordFieldOp */
-        bytes.extend_from_slice(&self.buffer.to_aml_bytes());
+        self.buffer.append_aml_bytes(bytes);
-        bytes.extend_from_slice(&self.offset.to_aml_bytes());
+        self.offset.append_aml_bytes(bytes);
-        bytes.extend_from_slice(&self.field.to_aml_bytes());
+        self.field.append_aml_bytes(bytes);
-        bytes
     }
 }
 
@@ -1463,13 +1392,13 @@ mod tests {
 
     #[test]
     fn test_pkg_length() {
-        assert_eq!(create_pkg_length(&[0u8; 62].to_vec(), true), vec![63]);
+        assert_eq!(create_pkg_length(&[0u8; 62], true), vec![63]);
         assert_eq!(
-            create_pkg_length(&[0u8; 64].to_vec(), true),
+            create_pkg_length(&[0u8; 64], true),
             vec![1 << 6 | (66 & 0xf), 66 >> 4]
         );
         assert_eq!(
-            create_pkg_length(&[0u8; 4096].to_vec(), true),
+            create_pkg_length(&[0u8; 4096], true),
             vec![
                 2 << 6 | (4099 & 0xf) as u8,
                 (4099 >> 4) as u8,

acpi_tables/src/rsdp.rs

@@ -19,6 +19,7 @@ pub struct Rsdp {
     _reserved: [u8; 3],
 }
 
+// SAFETY: Rsdp only contains a series of integers
 unsafe impl ByteValued for Rsdp {}
 
 impl Rsdp {

arch/Cargo.toml

@@ -11,18 +11,18 @@ tdx = []
 
 [dependencies]
 acpi_tables = { path = "../acpi_tables", optional = true }
-anyhow = "1.0.44"
+anyhow = "1.0.55"
 byteorder = "1.4.3"
 hypervisor = { path = "../hypervisor" }
-libc = "0.2.103"
+libc = "0.2.119"
 linux-loader = { version = "0.4.0", features = ["elf", "bzimage", "pe"] }
 log = "0.4.14"
-serde = { version = "1.0.130", features = ["rc"] }
+serde = { version = "1.0.136", features = ["rc"] }
-serde_derive = "1.0.130"
+serde_derive = "1.0.136"
 thiserror = "1.0.30"
 versionize = "0.1.6"
 versionize_derive = "0.1.4"
-vm-memory = { version = "0.6.0", features = ["backend-mmap", "backend-bitmap"] }
+vm-memory = { version = "0.7.0", features = ["backend-mmap", "backend-bitmap"] }
 vm-migration = { path = "../vm-migration" }
 vmm-sys-util = { version = "0.9.0", features = ["with-serde"] }
 

arch/src/aarch64/fdt.rs

@@ -6,7 +6,7 @@
 // Use of this source code is governed by a BSD-style license that can be
 // found in the THIRD-PARTY file.
 
-use crate::NumaNodes;
+use crate::{NumaNodes, PciSpaceInfo};
 use byteorder::{BigEndian, ByteOrder};
 use std::cmp;
 use std::collections::HashMap;
@@ -22,7 +22,7 @@ use super::get_fdt_addr;
 use super::gic::GicDevice;
 use super::layout::{
     IRQ_BASE, MEM_32BIT_DEVICES_SIZE, MEM_32BIT_DEVICES_START, MEM_PCI_IO_SIZE, MEM_PCI_IO_START,
-    PCI_HIGH_BASE, PCI_MMCONFIG_SIZE, PCI_MMCONFIG_START,
+    PCI_HIGH_BASE, PCI_MMIO_CONFIG_SIZE_PER_SEGMENT,
 };
 use vm_fdt::{FdtWriter, FdtWriterResult};
 use vm_memory::{Address, Bytes, GuestAddress, GuestMemory, GuestMemoryError, GuestMemoryRegion};
@@ -51,11 +51,16 @@ const SIZE_CELLS: u32 = 0x2;
 // Look for "The 1st cell..."
 const GIC_FDT_IRQ_TYPE_SPI: u32 = 0;
 const GIC_FDT_IRQ_TYPE_PPI: u32 = 1;
+const GIC_FDT_IRQ_PPI_CPU_SHIFT: u32 = 8;
+const GIC_FDT_IRQ_PPI_CPU_MASK: u32 = 0xff << GIC_FDT_IRQ_PPI_CPU_SHIFT;
 
 // From https://elixir.bootlin.com/linux/v4.9.62/source/include/dt-bindings/interrupt-controller/irq.h#L17
 const IRQ_TYPE_EDGE_RISING: u32 = 1;
 const IRQ_TYPE_LEVEL_HI: u32 = 4;
 
+// PMU PPI interrupt number
+pub const AARCH64_PMU_IRQ: u32 = 7;
+
 // Keys and Buttons
 // System Power Down
 const KEY_POWER: u32 = 116;
@@ -88,9 +93,10 @@ pub fn create_fdt<T: DeviceInfoForFdt + Clone + Debug, S: ::std::hash::BuildHash
     device_info: &HashMap<(DeviceType, String), T, S>,
     gic_device: &dyn GicDevice,
     initrd: &Option<InitramfsConfig>,
-    pci_space_address: &(u64, u64),
+    pci_space_info: &[PciSpaceInfo],
     numa_nodes: &NumaNodes,
     virtio_iommu_bdf: Option<u32>,
+    pmu_supported: bool,
 ) -> FdtWriterResult<Vec<u8>> {
     // Allocate stuff necessary for the holding the blob.
     let mut fdt = FdtWriter::new().unwrap();
@@ -114,15 +120,13 @@ pub fn create_fdt<T: DeviceInfoForFdt + Clone + Debug, S: ::std::hash::BuildHash
     create_chosen_node(&mut fdt, cmdline, initrd)?;
     create_gic_node(&mut fdt, gic_device)?;
     create_timer_node(&mut fdt)?;
+    if pmu_supported {
+        create_pmu_node(&mut fdt, vcpu_mpidr.len())?;
+    }
     create_clock_node(&mut fdt)?;
     create_psci_node(&mut fdt)?;
     create_devices_node(&mut fdt, device_info)?;
-    create_pci_nodes(
+    create_pci_nodes(&mut fdt, pci_space_info, virtio_iommu_bdf)?;
-        &mut fdt,
-        pci_space_address.0,
-        pci_space_address.1,
-        virtio_iommu_bdf,
-    )?;
     if numa_nodes.len() > 1 {
         create_distance_map_node(&mut fdt, numa_nodes)?;
     }
@@ -157,9 +161,6 @@ fn create_cpu_nodes(
     fdt.property_u32("#size-cells", 0x0)?;
 
     let num_cpus = vcpu_mpidr.len();
-    let threads_per_core = vcpu_topology.unwrap_or_default().0 as u8;
-    let cores_per_package = vcpu_topology.unwrap_or_default().1 as u8;
-    let packages = vcpu_topology.unwrap_or_default().2 as u8;
 
     for (cpu_id, mpidr) in vcpu_mpidr.iter().enumerate().take(num_cpus) {
         let cpu_name = format!("cpu@{:x}", cpu_id);
@@ -188,32 +189,9 @@ fn create_cpu_nodes(
         fdt.end_node(cpu_node)?;
     }
 
-    // If there is a valid cpu topology config, create the cpu-map node.
+    if let Some(topology) = vcpu_topology {
-    if (threads_per_core > 0)
+        let (threads_per_core, cores_per_package, packages) = topology;
-        && (cores_per_package > 0)
-        && (packages > 0)
-        && (num_cpus as u8 == threads_per_core * cores_per_package * packages)
-    {
         let cpu_map_node = fdt.begin_node("cpu-map")?;
-        // Create mappings between CPU index and cluster,core, and thread.
-        let mut cluster_core_thread_to_cpuidx = HashMap::new();
-        for cpu_idx in 0..num_cpus as u8 {
-            if threads_per_core > 1 {
-                cluster_core_thread_to_cpuidx.insert(
-                    (
-                        cpu_idx / (cores_per_package * threads_per_core),
-                        (cpu_idx / threads_per_core) % cores_per_package,
-                        cpu_idx % threads_per_core,
-                    ),
-                    cpu_idx,
-                );
-            } else {
-                cluster_core_thread_to_cpuidx.insert(
-                    (cpu_idx / cores_per_package, cpu_idx % cores_per_package, 0),
-                    cpu_idx,
-                );
-            }
-        }
 
         // Create device tree nodes with regard of above mapping.
         for cluster_idx in 0..packages {
@@ -224,24 +202,16 @@ fn create_cpu_nodes(
                 let core_name = format!("core{:x}", core_idx);
                 let core_node = fdt.begin_node(&core_name)?;
 
-                if threads_per_core > 1 {
+                for thread_idx in 0..threads_per_core {
-                    for thread_idx in 0..threads_per_core {
+                    let thread_name = format!("thread{:x}", thread_idx);
-                        let thread_name = format!("thread{:x}", thread_idx);
+                    let thread_node = fdt.begin_node(&thread_name)?;
-                        let thread_node = fdt.begin_node(&thread_name)?;
+                    let cpu_idx = threads_per_core * cores_per_package * cluster_idx
-                        let cpu_idx = cluster_core_thread_to_cpuidx
+                        + threads_per_core * core_idx
-                            .get(&(cluster_idx, core_idx, thread_idx))
+                        + thread_idx;
-                            .unwrap();
+                    fdt.property_u32("cpu", cpu_idx as u32 + FIRST_VCPU_PHANDLE)?;
-
+                    fdt.end_node(thread_node)?;
-                        fdt.property_u32("cpu", *cpu_idx as u32 + FIRST_VCPU_PHANDLE)?;
-                        fdt.end_node(thread_node)?;
-                    }
-                } else {
-                    let cpu_idx = cluster_core_thread_to_cpuidx
-                        .get(&(cluster_idx, core_idx, 0))
-                        .unwrap();
-
-                    fdt.property_u32("cpu", *cpu_idx as u32 + FIRST_VCPU_PHANDLE)?;
                 }
+
                 fdt.end_node(core_node)?;
             }
             fdt.end_node(cluster_node)?;
@@ -545,110 +515,165 @@ fn create_devices_node<T: DeviceInfoForFdt + Clone + Debug, S: ::std::hash::Buil
     Ok(())
 }
 
+fn create_pmu_node(fdt: &mut FdtWriter, cpu_nums: usize) -> FdtWriterResult<()> {
+    let num_cpus = cpu_nums as u64 as u32;
+    let compatible = "arm,armv8-pmuv3";
+    let cpu_mask: u32 =
+        (((1 << num_cpus) - 1) << GIC_FDT_IRQ_PPI_CPU_SHIFT) & GIC_FDT_IRQ_PPI_CPU_MASK;
+    let irq = [
+        GIC_FDT_IRQ_TYPE_PPI,
+        AARCH64_PMU_IRQ,
+        cpu_mask | IRQ_TYPE_LEVEL_HI,
+    ];
+
+    let pmu_node = fdt.begin_node("pmu")?;
+    fdt.property_string("compatible", compatible)?;
+    fdt.property_array_u32("interrupts", &irq)?;
+    fdt.end_node(pmu_node)?;
+    Ok(())
+}
+
 fn create_pci_nodes(
     fdt: &mut FdtWriter,
-    pci_device_base: u64,
+    pci_device_info: &[PciSpaceInfo],
-    pci_device_size: u64,
     virtio_iommu_bdf: Option<u32>,
 ) -> FdtWriterResult<()> {
     // Add node for PCIe controller.
     // See Documentation/devicetree/bindings/pci/host-generic-pci.txt in the kernel
     // and https://elinux.org/Device_Tree_Usage.
+    // In multiple PCI segments setup, each PCI segment needs a PCI node.
+    for pci_device_info_elem in pci_device_info.iter() {
+        // EDK2 requires the PCIe high space above 4G address.
+        // The actual space in CLH follows the RAM. If the RAM space is small, the PCIe high space
+        // could fall bellow 4G.
+        // Here we cut off PCI device space below 8G in FDT to workaround the EDK2 check.
+        // But the address written in ACPI is not impacted.
+        let (pci_device_base_64bit, pci_device_size_64bit) = if cfg!(feature = "acpi")
+            && (pci_device_info_elem.pci_device_space_start < PCI_HIGH_BASE)
+        {
+            (
+                PCI_HIGH_BASE,
+                pci_device_info_elem.pci_device_space_size
+                    - (PCI_HIGH_BASE - pci_device_info_elem.pci_device_space_start),
+            )
+        } else {
+            (
+                pci_device_info_elem.pci_device_space_start,
+                pci_device_info_elem.pci_device_space_size,
+            )
+        };
+        // There is no specific requirement of the 32bit MMIO range, and
+        // therefore at least we can make these ranges 4K aligned.
+        let pci_device_size_32bit: u64 =
+            MEM_32BIT_DEVICES_SIZE / ((1 << 12) * pci_device_info.len() as u64) * (1 << 12);
+        let pci_device_base_32bit: u64 = MEM_32BIT_DEVICES_START.0
+            + pci_device_size_32bit * pci_device_info_elem.pci_segment_id as u64;
 
-    // EDK2 requires the PCIe high space above 4G address.
+        let ranges = [
-    // The actual space in CLH follows the RAM. If the RAM space is small, the PCIe high space
+            // io addresses. Since AArch64 will not use IO address,
-    // could fall bellow 4G.
+            // we can set the same IO address range for every segment.
-    // Here we put it above 512G in FDT to workaround the EDK2 check.
+            0x1000000,
-    // But the address written in ACPI is not impacted.
-    let pci_device_base_64bit: u64 = if cfg!(feature = "acpi") {
-        pci_device_base + PCI_HIGH_BASE
-    } else {
-        pci_device_base
-    };
-    let pci_device_size_64bit: u64 = if cfg!(feature = "acpi") {
-        pci_device_size - PCI_HIGH_BASE
-    } else {
-        pci_device_size
-    };
-
-    let ranges = [
-        // io addresses
-        0x1000000,
-        0_u32,
-        0_u32,
-        (MEM_PCI_IO_START.0 >> 32) as u32,
-        MEM_PCI_IO_START.0 as u32,
-        (MEM_PCI_IO_SIZE >> 32) as u32,
-        MEM_PCI_IO_SIZE as u32,
-        // mmio addresses
-        0x2000000,                                // (ss = 10: 32-bit memory space)
-        (MEM_32BIT_DEVICES_START.0 >> 32) as u32, // PCI address
-        MEM_32BIT_DEVICES_START.0 as u32,
-        (MEM_32BIT_DEVICES_START.0 >> 32) as u32, // CPU address
-        MEM_32BIT_DEVICES_START.0 as u32,
-        (MEM_32BIT_DEVICES_SIZE >> 32) as u32, // size
-        MEM_32BIT_DEVICES_SIZE as u32,
-        // device addresses
-        0x3000000,                            // (ss = 11: 64-bit memory space)
-        (pci_device_base_64bit >> 32) as u32, // PCI address
-        pci_device_base_64bit as u32,
-        (pci_device_base_64bit >> 32) as u32, // CPU address
-        pci_device_base_64bit as u32,
-        (pci_device_size_64bit >> 32) as u32, // size
-        pci_device_size_64bit as u32,
-    ];
-    let bus_range = [0, 0]; // Only bus 0
-    let reg = [PCI_MMCONFIG_START.0, PCI_MMCONFIG_SIZE];
-
-    let pci_node = fdt.begin_node("pci")?;
-    fdt.property_string("compatible", "pci-host-ecam-generic")?;
-    fdt.property_string("device_type", "pci")?;
-    fdt.property_array_u32("ranges", &ranges)?;
-    fdt.property_array_u32("bus-range", &bus_range)?;
-    fdt.property_u32("#address-cells", 3)?;
-    fdt.property_u32("#size-cells", 2)?;
-    fdt.property_array_u64("reg", &reg)?;
-    fdt.property_u32("#interrupt-cells", 1)?;
-    fdt.property_null("interrupt-map")?;
-    fdt.property_null("interrupt-map-mask")?;
-    fdt.property_null("dma-coherent")?;
-    fdt.property_u32("msi-parent", MSI_PHANDLE)?;
-
-    if let Some(virtio_iommu_bdf) = virtio_iommu_bdf {
-        // See kernel document Documentation/devicetree/bindings/pci/pci-iommu.txt
-        // for 'iommu-map' attribute setting.
-        let iommu_map = [
             0_u32,
-            VIRTIO_IOMMU_PHANDLE,
             0_u32,
-            virtio_iommu_bdf,
+            (MEM_PCI_IO_START.0 >> 32) as u32,
-            virtio_iommu_bdf + 1,
+            MEM_PCI_IO_START.0 as u32,
-            VIRTIO_IOMMU_PHANDLE,
+            (MEM_PCI_IO_SIZE >> 32) as u32,
-            virtio_iommu_bdf + 1,
+            MEM_PCI_IO_SIZE as u32,
-            0xffff - virtio_iommu_bdf,
+            // mmio addresses
+            0x2000000,                            // (ss = 10: 32-bit memory space)
+            (pci_device_base_32bit >> 32) as u32, // PCI address
+            pci_device_base_32bit as u32,
+            (pci_device_base_32bit >> 32) as u32, // CPU address
+            pci_device_base_32bit as u32,
+            (pci_device_size_32bit >> 32) as u32, // size
+            pci_device_size_32bit as u32,
+            // device addresses
+            0x3000000,                            // (ss = 11: 64-bit memory space)
+            (pci_device_base_64bit >> 32) as u32, // PCI address
+            pci_device_base_64bit as u32,
+            (pci_device_base_64bit >> 32) as u32, // CPU address
+            pci_device_base_64bit as u32,
+            (pci_device_size_64bit >> 32) as u32, // size
+            pci_device_size_64bit as u32,
+        ];
+        let bus_range = [0, 0]; // Only bus 0
+        let reg = [
+            pci_device_info_elem.mmio_config_address,
+            PCI_MMIO_CONFIG_SIZE_PER_SEGMENT,
+        ];
+        // See kernel document Documentation/devicetree/bindings/pci/pci-msi.txt
+        let msi_map = [
+            // rid-base: A single cell describing the first RID matched by the entry.
+            0x0,
+            // msi-controller: A single phandle to an MSI controller.
+            MSI_PHANDLE,
+            // msi-base: An msi-specifier describing the msi-specifier produced for the
+            // first RID matched by the entry.
+            (pci_device_info_elem.pci_segment_id as u32) << 8,
+            // length: A single cell describing how many consecutive RIDs are matched
+            // following the rid-base.
+            0x100,
         ];
-        fdt.property_array_u32("iommu-map", &iommu_map)?;
 
-        // See kernel document Documentation/devicetree/bindings/virtio/iommu.txt
+        let pci_node_name = format!("pci@{:x}", pci_device_info_elem.mmio_config_address);
-        // for virtio-iommu node settings.
+        let pci_node = fdt.begin_node(&pci_node_name)?;
-        let virtio_iommu_node_name = format!("virtio_iommu@{:x}", virtio_iommu_bdf);
-        let virtio_iommu_node = fdt.begin_node(&virtio_iommu_node_name)?;
-        fdt.property_u32("#iommu-cells", 1)?;
-        fdt.property_string("compatible", "virtio,pci-iommu")?;
 
-        // 'reg' is a five-cell address encoded as
+        fdt.property_string("compatible", "pci-host-ecam-generic")?;
-        // (phys.hi phys.mid phys.lo size.hi size.lo). phys.hi should contain the
+        fdt.property_string("device_type", "pci")?;
-        // device's BDF as 0b00000000 bbbbbbbb dddddfff 00000000. The other cells
+        fdt.property_array_u32("ranges", &ranges)?;
-        // should be zero.
+        fdt.property_array_u32("bus-range", &bus_range)?;
-        let reg = [virtio_iommu_bdf << 8, 0_u32, 0_u32, 0_u32, 0_u32];
+        fdt.property_u32(
-        fdt.property_array_u32("reg", &reg)?;
+            "linux,pci-domain",
-        fdt.property_u32("phandle", VIRTIO_IOMMU_PHANDLE)?;
+            pci_device_info_elem.pci_segment_id as u32,
+        )?;
+        fdt.property_u32("#address-cells", 3)?;
+        fdt.property_u32("#size-cells", 2)?;
+        fdt.property_array_u64("reg", &reg)?;
+        fdt.property_u32("#interrupt-cells", 1)?;
+        fdt.property_null("interrupt-map")?;
+        fdt.property_null("interrupt-map-mask")?;
+        fdt.property_null("dma-coherent")?;
+        fdt.property_array_u32("msi-map", &msi_map)?;
+        fdt.property_u32("msi-parent", MSI_PHANDLE)?;
 
-        fdt.end_node(virtio_iommu_node)?;
+        if pci_device_info_elem.pci_segment_id == 0 {
+            if let Some(virtio_iommu_bdf) = virtio_iommu_bdf {
+                // See kernel document Documentation/devicetree/bindings/pci/pci-iommu.txt
+                // for 'iommu-map' attribute setting.
+                let iommu_map = [
+                    0_u32,
+                    VIRTIO_IOMMU_PHANDLE,
+                    0_u32,
+                    virtio_iommu_bdf,
+                    virtio_iommu_bdf + 1,
+                    VIRTIO_IOMMU_PHANDLE,
+                    virtio_iommu_bdf + 1,
+                    0xffff - virtio_iommu_bdf,
+                ];
+                fdt.property_array_u32("iommu-map", &iommu_map)?;
+
+                // See kernel document Documentation/devicetree/bindings/virtio/iommu.txt
+                // for virtio-iommu node settings.
+                let virtio_iommu_node_name = format!("virtio_iommu@{:x}", virtio_iommu_bdf);
+                let virtio_iommu_node = fdt.begin_node(&virtio_iommu_node_name)?;
+                fdt.property_u32("#iommu-cells", 1)?;
+                fdt.property_string("compatible", "virtio,pci-iommu")?;
+
+                // 'reg' is a five-cell address encoded as
+                // (phys.hi phys.mid phys.lo size.hi size.lo). phys.hi should contain the
+                // device's BDF as 0b00000000 bbbbbbbb dddddfff 00000000. The other cells
+                // should be zero.
+                let reg = [virtio_iommu_bdf << 8, 0_u32, 0_u32, 0_u32, 0_u32];
+                fdt.property_array_u32("reg", &reg)?;
+                fdt.property_u32("phandle", VIRTIO_IOMMU_PHANDLE)?;
+
+                fdt.end_node(virtio_iommu_node)?;
+            }
+        }
+
+        fdt.end_node(pci_node)?;
     }
 
-    fdt.end_node(pci_node)?;
-
     Ok(())
 }
 

arch/src/aarch64/gic/dist_regs.rs

@@ -63,7 +63,7 @@ macro_rules! VGIC_DIST_REG {
 
 // List with relevant distributor registers that we will be restoring.
 // Order is taken from qemu.
-static VGIC_DIST_REGS: &'static [DistReg] = &[
+static VGIC_DIST_REGS: &[DistReg] = &[
     VGIC_DIST_REG!(GICD_STATUSR, 0, 4),
     VGIC_DIST_REG!(GICD_ICENABLER, 1, 0),
     VGIC_DIST_REG!(GICD_ISENABLER, 1, 0),

arch/src/aarch64/gic/gicv3_its.rs

@@ -124,12 +124,11 @@ pub mod kvm {
         its_device: &Arc<dyn hypervisor::Device>,
         save: bool,
     ) -> crate::aarch64::gic::Result<()> {
-        let attr: u64;
+        let attr = if save {
-        if save {
+            u64::from(kvm_bindings::KVM_DEV_ARM_ITS_SAVE_TABLES)
-            attr = u64::from(kvm_bindings::KVM_DEV_ARM_ITS_SAVE_TABLES);
         } else {
-            attr = u64::from(kvm_bindings::KVM_DEV_ARM_ITS_RESTORE_TABLES);
+            u64::from(kvm_bindings::KVM_DEV_ARM_ITS_RESTORE_TABLES)
-        }
+        };
 
         let init_gic_attr = kvm_bindings::kvm_device_attr {
             group: kvm_bindings::KVM_DEV_ARM_VGIC_GRP_CTRL,

arch/src/aarch64/gic/icc_regs.rs

@@ -59,7 +59,7 @@ SYS_ICC_AP1Rn_EL1!(SYS_ICC_AP1R1_EL1, 1);
 SYS_ICC_AP1Rn_EL1!(SYS_ICC_AP1R2_EL1, 2);
 SYS_ICC_AP1Rn_EL1!(SYS_ICC_AP1R3_EL1, 3);
 
-static VGIC_ICC_REGS: &'static [u64] = &[
+static VGIC_ICC_REGS: &[u64] = &[
     SYS_ICC_SRE_EL1,
     SYS_ICC_CTLR_EL1,
     SYS_ICC_IGRPEN0_EL1,
@@ -91,8 +91,7 @@ fn icc_attr_access(
         flags: 0,
     };
     if set {
-        #[allow(clippy::unnecessary_mut_passed)]
+        gic.set_device_attr(&gic_icc_attr)
-        gic.set_device_attr(&mut gic_icc_attr)
             .map_err(Error::SetDeviceAttribute)?;
     } else {
         gic.get_device_attr(&mut gic_icc_attr)

arch/src/aarch64/gic/redist_regs.rs

@@ -61,7 +61,7 @@ macro_rules! VGIC_RDIST_REG {
 }
 
 // List with relevant distributor registers that we will be restoring.
-static VGIC_RDIST_REGS: &'static [RdistReg] = &[
+static VGIC_RDIST_REGS: &[RdistReg] = &[
     VGIC_RDIST_REG!(GICR_STATUSR, 4),
     VGIC_RDIST_REG!(GICR_WAKER, 4),
     VGIC_RDIST_REG!(GICR_PROPBASER, 8),
@@ -70,7 +70,7 @@ static VGIC_RDIST_REGS: &'static [RdistReg] = &[
 ];
 
 // List with relevant distributor registers that we will be restoring.
-static VGIC_SGI_REGS: &'static [RdistReg] = &[
+static VGIC_SGI_REGS: &[RdistReg] = &[
     VGIC_RDIST_REG!(GICR_IGROUPR0, 4),
     VGIC_RDIST_REG!(GICR_ICENABLER0, 4),
     VGIC_RDIST_REG!(GICR_ISENABLER0, 4),
@@ -96,8 +96,7 @@ fn redist_attr_access(
         flags: 0,
     };
     if set {
-        #[allow(clippy::unnecessary_mut_passed)]
+        gic.set_device_attr(&gic_dist_attr)
-        gic.set_device_attr(&mut gic_dist_attr)
             .map_err(Error::SetDeviceAttribute)?;
     } else {
         gic.get_device_attr(&mut gic_dist_attr)
@@ -110,7 +109,7 @@ fn access_redists_aux(
     gic: &Arc<dyn hypervisor::Device>,
     gicr_typer: &[u64],
     state: &mut Vec<u32>,
-    reg_list: &'static [RdistReg],
+    reg_list: &[RdistReg],
     idx: &mut usize,
     set: bool,
 ) -> Result<()> {

arch/src/aarch64/layout.rs

@@ -82,6 +82,8 @@ pub const MEM_32BIT_DEVICES_SIZE: u64 = 0x2000_0000;
 /// PCI MMCONFIG space (start: after the device space at 1 GiB, length: 256MiB)
 pub const PCI_MMCONFIG_START: GuestAddress = GuestAddress(0x3000_0000);
 pub const PCI_MMCONFIG_SIZE: u64 = 256 << 20;
+// One bus with potentially 256 devices (32 slots x 8 functions).
+pub const PCI_MMIO_CONFIG_SIZE_PER_SEGMENT: u64 = 4096 * 256;
 
 /// Start of RAM on 64 bit ARM.
 pub const RAM_64BIT_START: u64 = 0x4000_0000;
@@ -104,7 +106,7 @@ pub const RSDP_POINTER: GuestAddress = GuestAddress(ACPI_START);
 pub const KERNEL_START: u64 = ACPI_START + ACPI_MAX_SIZE as u64;
 
 /// Pci high memory base
-pub const PCI_HIGH_BASE: u64 = 0x80_0000_0000_u64;
+pub const PCI_HIGH_BASE: u64 = 0x2_0000_0000_u64;
 
 // As per virt/kvm/arm/vgic/vgic-kvm-device.c we need
 // the number of interrupts our GIC will support to be:

arch/src/aarch64/mod.rs

@@ -14,16 +14,14 @@ pub mod regs;
 pub mod uefi;
 
 pub use self::fdt::DeviceInfoForFdt;
-use crate::{DeviceType, GuestMemoryMmap, NumaNodes, RegionType};
+use crate::{DeviceType, GuestMemoryMmap, NumaNodes, PciSpaceInfo, RegionType};
 use gic::GicDevice;
 use log::{log_enabled, Level};
 use std::collections::HashMap;
 use std::convert::TryInto;
 use std::fmt::Debug;
 use std::sync::Arc;
-use vm_memory::{
+use vm_memory::{Address, GuestAddress, GuestMemory, GuestUsize};
-    Address, GuestAddress, GuestAddressSpace, GuestMemory, GuestMemoryAtomic, GuestUsize,
-};
 
 /// Errors thrown while configuring aarch64 system.
 #[derive(Debug)]
@@ -45,6 +43,9 @@ pub enum Error {
 
     /// Error configuring the MPIDR register
     VcpuRegMpidr(hypervisor::HypervisorCpuError),
+
+    /// Error initializing PMU for vcpu
+    VcpuInitPmu,
 }
 
 impl From<Error> for super::Error {
@@ -66,16 +67,10 @@ pub fn configure_vcpu(
     fd: &Arc<dyn hypervisor::Vcpu>,
     id: u8,
     kernel_entry_point: Option<EntryPoint>,
-    vm_memory: &GuestMemoryAtomic<GuestMemoryMmap>,
 ) -> super::Result<u64> {
     if let Some(kernel_entry_point) = kernel_entry_point {
-        regs::setup_regs(
+        regs::setup_regs(fd, id, kernel_entry_point.entry_addr.raw_value())
-            fd,
+            .map_err(Error::RegsConfiguration)?;
-            id,
-            kernel_entry_point.entry_addr.raw_value(),
-            &vm_memory.memory(),
-        )
-        .map_err(Error::RegsConfiguration)?;
     }
 
     let mpidr = fd.read_mpidr().map_err(Error::VcpuRegMpidr)?;
@@ -140,10 +135,11 @@ pub fn configure_system<T: DeviceInfoForFdt + Clone + Debug, S: ::std::hash::Bui
     vcpu_topology: Option<(u8, u8, u8)>,
     device_info: &HashMap<(DeviceType, String), T, S>,
     initrd: &Option<super::InitramfsConfig>,
-    pci_space_address: &(u64, u64),
+    pci_space_info: &[PciSpaceInfo],
     virtio_iommu_bdf: Option<u32>,
     gic_device: &dyn GicDevice,
     numa_nodes: &NumaNodes,
+    pmu_supported: bool,
 ) -> super::Result<()> {
     let fdt_final = fdt::create_fdt(
         guest_mem,
@@ -153,9 +149,10 @@ pub fn configure_system<T: DeviceInfoForFdt + Clone + Debug, S: ::std::hash::Bui
         device_info,
         gic_device,
         initrd,
-        pci_space_address,
+        pci_space_info,
         numa_nodes,
         virtio_iommu_bdf,
+        pmu_supported,
     )
     .map_err(|_| Error::SetupFdt)?;
 

arch/src/aarch64/regs.rs

@@ -6,7 +6,6 @@
 // found in the THIRD-PARTY file.
 
 use super::get_fdt_addr;
-use crate::GuestMemoryMmap;
 use hypervisor::kvm::kvm_bindings::{
     kvm_regs, user_pt_regs, KVM_REG_ARM64, KVM_REG_ARM_CORE, KVM_REG_SIZE_U64,
 };
@@ -43,12 +42,7 @@ const PSTATE_FAULT_BITS_64: u64 = PSR_MODE_EL1h | PSR_A_BIT | PSR_F_BIT | PSR_I_
 /// * `cpu_id` - Index of current vcpu.
 /// * `boot_ip` - Starting instruction pointer.
 /// * `mem` - Reserved DRAM for current VM.
-pub fn setup_regs(
+pub fn setup_regs(vcpu: &Arc<dyn hypervisor::Vcpu>, cpu_id: u8, boot_ip: u64) -> Result<()> {
-    vcpu: &Arc<dyn hypervisor::Vcpu>,
-    cpu_id: u8,
-    boot_ip: u64,
-    _mem: &GuestMemoryMmap,
-) -> Result<()> {
     let kreg_off = offset__of!(kvm_regs, regs);
 
     // Get the register index of the PSTATE (Processor State) register.

arch/src/lib.rs

@@ -6,7 +6,6 @@
 
 //! Implements platform specific functionality.
 //! Supported platforms: x86_64, aarch64.
-#![allow(clippy::transmute_ptr_to_ptr, clippy::redundant_static_lifetimes)]
 
 #[macro_use]
 extern crate log;
@@ -162,6 +161,16 @@ pub struct MmioDeviceInfo {
     pub irq: u32,
 }
 
+/// Structure to describe PCI space information
+#[derive(Clone, Debug)]
+#[cfg(target_arch = "aarch64")]
+pub struct PciSpaceInfo {
+    pub pci_segment_id: u16,
+    pub mmio_config_address: u64,
+    pub pci_device_space_start: u64,
+    pub pci_device_space_size: u64,
+}
+
 #[cfg(target_arch = "aarch64")]
 impl DeviceInfoForFdt for MmioDeviceInfo {
     fn addr(&self) -> u64 {

arch/src/x86_64/interrupts.rs

@@ -95,22 +95,4 @@ mod tests {
         let mut klapic = LapicState::default();
         set_klapic_reg(&mut klapic, reg_offset, 3);
     }
-
-    #[test]
-    fn test_apic_delivery_mode() {
-        let mut v: Vec<u32> = Vec::new();
-        v.resize(20, 0);
-
-        unsafe {
-            assert_eq!(
-                libc::getrandom(v.as_mut_ptr() as *mut _ as *mut libc::c_void, 80, 0),
-                80
-            );
-        }
-
-        v.iter_mut()
-            .for_each(|x| *x = set_apic_delivery_mode(*x, 2));
-        let after: Vec<u32> = v.iter().map(|x| ((*x & !0x700) | ((2) << 8))).collect();
-        assert_eq!(v, after);
-    }
 }

arch/src/x86_64/layout.rs

@@ -79,9 +79,9 @@ pub const HIGH_RAM_START: GuestAddress = GuestAddress(0x100000);
 
 // == No fixed addresses in the "High RAM" range ==
 
-// ** 32-bit reserved area (start: 3GiB, length: 1GiB) **
+// ** 32-bit reserved area (start: 3GiB, length: 896MiB) **
 pub const MEM_32BIT_RESERVED_START: GuestAddress = GuestAddress(0xc000_0000);
-pub const MEM_32BIT_RESERVED_SIZE: u64 = 1024 << 20;
+pub const MEM_32BIT_RESERVED_SIZE: u64 = PCI_MMCONFIG_SIZE + MEM_32BIT_DEVICES_SIZE;
 
 // == Fixed constants within the "32-bit reserved" range ==
 
@@ -93,6 +93,16 @@ pub const MEM_32BIT_DEVICES_SIZE: u64 = 640 << 20;
 pub const PCI_MMCONFIG_START: GuestAddress =
     GuestAddress(MEM_32BIT_DEVICES_START.0 + MEM_32BIT_DEVICES_SIZE);
 pub const PCI_MMCONFIG_SIZE: u64 = 256 << 20;
+// One bus with potentially 256 devices (32 slots x 8 functions).
+pub const PCI_MMIO_CONFIG_SIZE_PER_SEGMENT: u64 = 4096 * 256;
+
+// TSS is 3 pages after the PCI MMCONFIG space
+pub const KVM_TSS_START: GuestAddress = GuestAddress(PCI_MMCONFIG_START.0 + PCI_MMCONFIG_SIZE);
+pub const KVM_TSS_SIZE: u64 = (3 * 4) << 10;
+
+// Identity map is a one page region after the TSS
+pub const KVM_IDENTITY_MAP_START: GuestAddress = GuestAddress(KVM_TSS_START.0 + KVM_TSS_SIZE);
+pub const KVM_IDENTITY_MAP_SIZE: u64 = 4 << 10;
 
 // IOAPIC
 pub const IOAPIC_START: GuestAddress = GuestAddress(0xfec0_0000);
@@ -101,9 +111,6 @@ pub const IOAPIC_SIZE: u64 = 0x20;
 // APIC
 pub const APIC_START: GuestAddress = GuestAddress(0xfee0_0000);
 
-/// Address for the TSS setup.
-pub const KVM_TSS_ADDRESS: GuestAddress = GuestAddress(0xfffb_d000);
-
 // == End of "32-bit reserved" range. ==
 
 // ** 64-bit RAM start (start: 4GiB, length: varies) **

arch/src/x86_64/mod.rs

@@ -57,7 +57,7 @@ const KVM_FEATURE_STEAL_TIME_BIT: u8 = 5;
 /// is to be used to configure the guest initial state.
 pub struct EntryPoint {
     /// Address in guest memory where the guest must start execution
-    pub entry_addr: GuestAddress,
+    pub entry_addr: Option<GuestAddress>,
 }
 
 const E820_RAM: u32 = 1;
@@ -118,17 +118,15 @@ impl SgxEpcRegion {
 #[derive(Copy, Clone, Default)]
 struct StartInfoWrapper(hvm_start_info);
 
-// It is safe to initialize StartInfoWrapper which is a wrapper over `hvm_start_info` (a series of ints).
-unsafe impl ByteValued for StartInfoWrapper {}
-
 #[derive(Copy, Clone, Default)]
 struct MemmapTableEntryWrapper(hvm_memmap_table_entry);
 
-unsafe impl ByteValued for MemmapTableEntryWrapper {}
-
 #[derive(Copy, Clone, Default)]
 struct ModlistEntryWrapper(hvm_modlist_entry);
 
+// SAFETY: These data structures only contain a series of integers
+unsafe impl ByteValued for StartInfoWrapper {}
+unsafe impl ByteValued for MemmapTableEntryWrapper {}
 unsafe impl ByteValued for ModlistEntryWrapper {}
 
 // This is a workaround to the Rust enforcement specifying that any implementation of a foreign
@@ -139,7 +137,7 @@ unsafe impl ByteValued for ModlistEntryWrapper {}
 #[derive(Copy, Clone, Default)]
 struct BootParamsWrapper(boot_params);
 
-// It is safe to initialize BootParamsWrap which is a wrapper over `boot_params` (a series of ints).
+// SAFETY: BootParamsWrap is a wrapper over `boot_params` (a series of ints).
 unsafe impl ByteValued for BootParamsWrapper {}
 
 #[derive(Debug)]
@@ -317,21 +315,12 @@ impl CpuidPatch {
 
         for entry in entries.iter() {
             if entry.function == function && entry.index == index {
-                let reg_val: u32;
+                let reg_val = match reg {
-                match reg {
+                    CpuidReg::EAX => entry.eax,
-                    CpuidReg::EAX => {
+                    CpuidReg::EBX => entry.ebx,
-                        reg_val = entry.eax;
+                    CpuidReg::ECX => entry.ecx,
-                    }
+                    CpuidReg::EDX => entry.edx,
-                    CpuidReg::EBX => {
+                };
-                        reg_val = entry.ebx;
-                    }
-                    CpuidReg::ECX => {
-                        reg_val = entry.ecx;
-                    }
-                    CpuidReg::EDX => {
-                        reg_val = entry.edx;
-                    }
-                }
 
                 return (reg_val & mask) == mask;
             }
@@ -524,19 +513,14 @@ impl CpuidFeatureEntry {
             .enumerate()
         {
             let entry = &feature_entry_list[i];
-            let entry_compatible;
+            let entry_compatible = match entry.compatible_check {
-            match entry.compatible_check {
                 CpuidCompatibleCheck::BitwiseSubset => {
                     let different_feature_bits = src_vm_feature ^ dest_vm_feature;
                     let src_vm_feature_bits_only = different_feature_bits & src_vm_feature;
-                    entry_compatible = src_vm_feature_bits_only == 0;
+                    src_vm_feature_bits_only == 0
-                }
-                CpuidCompatibleCheck::Equal => {
-                    entry_compatible = src_vm_feature == dest_vm_feature;
-                }
-                CpuidCompatibleCheck::NumNotGreater => {
-                    entry_compatible = src_vm_feature <= dest_vm_feature;
                 }
+                CpuidCompatibleCheck::Equal => src_vm_feature == dest_vm_feature,
+                CpuidCompatibleCheck::NumNotGreater => src_vm_feature <= dest_vm_feature,
             };
             if !entry_compatible {
                 error!(
@@ -743,11 +727,12 @@ pub fn configure_vcpu(
 
     regs::setup_msrs(fd).map_err(Error::MsrsConfiguration)?;
     if let Some(kernel_entry_point) = kernel_entry_point {
-        // Safe to unwrap because this method is called after the VM is configured
+        if let Some(entry_addr) = kernel_entry_point.entry_addr {
-        regs::setup_regs(fd, kernel_entry_point.entry_addr.raw_value())
+            // Safe to unwrap because this method is called after the VM is configured
-            .map_err(Error::RegsConfiguration)?;
+            regs::setup_regs(fd, entry_addr.raw_value()).map_err(Error::RegsConfiguration)?;
-        regs::setup_fpu(fd).map_err(Error::FpuConfiguration)?;
+            regs::setup_fpu(fd).map_err(Error::FpuConfiguration)?;
-        regs::setup_sregs(&vm_memory.memory(), fd).map_err(Error::SregsConfiguration)?;
+            regs::setup_sregs(&vm_memory.memory(), fd).map_err(Error::SregsConfiguration)?;
+        }
     }
     interrupts::set_lint(fd).map_err(|e| Error::LocalIntConfiguration(e.into()))?;
     Ok(())

arch/src/x86_64/mptable.rs

@@ -35,7 +35,7 @@ struct MpcLintsrcWrapper(mpspec::mpc_lintsrc);
 #[derive(Copy, Clone, Default)]
 struct MpfIntelWrapper(mpspec::mpf_intel);
 
-// These `mpspec` wrapper types are only data, reading them from data is a safe initialization.
+// SAFETY: These `mpspec` wrapper types are only data, reading them from data is a safe initialization.
 unsafe impl ByteValued for MpcBusWrapper {}
 unsafe impl ByteValued for MpcCpuWrapper {}
 unsafe impl ByteValued for MpcIntsrcWrapper {}

arch/src/x86_64/smbios.rs

@@ -82,7 +82,6 @@ pub struct Smbios30Entrypoint {
     pub max_size: u32,
     pub physptr: u64,
 }
-unsafe impl ByteValued for Smbios30Entrypoint {}
 
 impl Clone for Smbios30Entrypoint {
     fn clone(&self) -> Self {
@@ -112,8 +111,6 @@ impl Clone for SmbiosBiosInfo {
     }
 }
 
-unsafe impl ByteValued for SmbiosBiosInfo {}
-
 #[repr(packed)]
 #[derive(Default, Copy)]
 pub struct SmbiosSysInfo {
@@ -136,6 +133,9 @@ impl Clone for SmbiosSysInfo {
     }
 }
 
+// SAFETY: These data structures only contain a series of integers
+unsafe impl ByteValued for Smbios30Entrypoint {}
+unsafe impl ByteValued for SmbiosBiosInfo {}
 unsafe impl ByteValued for SmbiosSysInfo {}
 
 fn write_and_incr<T: ByteValued>(

arch/src/x86_64/tdx/mod.rs

@@ -25,6 +25,7 @@ pub enum TdvfError {
 
 // TDVF_DESCRIPTOR
 #[repr(packed)]
+#[derive(Default)]
 pub struct TdvfDescriptor {
     signature: [u8; 4],
     length: u32,
@@ -51,6 +52,9 @@ pub enum TdvfSectionType {
     Cfv,
     TdHob,
     TempMem,
+    PermMem,
+    Payload,
+    PayloadParam,
     Reserved = 0xffffffff,
 }
 
@@ -60,59 +64,6 @@ impl Default for TdvfSectionType {
     }
 }
 
-#[repr(C)]
-#[derive(Clone, Copy, Default, Debug)]
-pub struct TdVmmDataRegion {
-    pub start_address: u64,
-    pub length: u64,
-    pub region_type: TdVmmDataRegionType,
-}
-
-unsafe impl ByteValued for TdVmmDataRegion {}
-
-#[repr(u16)]
-#[derive(Clone, Copy, Debug)]
-pub enum TdVmmDataRegionType {
-    Signature = 0x0000,
-    InterfaceVersion = 0x0001,
-    SystemUuid = 0x0002,
-    RamSize = 0x0003,
-    GraphicsEnabled = 0x0004,
-    SmpCpuCount = 0x0005,
-    MachineId = 0x0006,
-    KernelAddress = 0x0007,
-    KernelSize = 0x0008,
-    KernelCommandLine = 0x0009,
-    InitrdAddress = 0x000a,
-    InitrdSize = 0x000b,
-    BootDevice = 0x000c,
-    NumaData = 0x000d,
-    BootMenu = 0x000e,
-    MaximumCpuCount = 0x000f,
-    KernelEntry = 0x0010,
-    KernelData = 0x0011,
-    InitrdData = 0x0012,
-    CommandLineAddress = 0x0013,
-    CommandLineSize = 0x0014,
-    CommandLineData = 0x0015,
-    KernelSetupAddress = 0x0016,
-    KernelSetupSize = 0x0017,
-    KernelSetupData = 0x0018,
-    FileDir = 0x0019,
-    AcpiTables = 0x8000,
-    SmbiosTables = 0x8001,
-    Irq0Override = 0x8002,
-    E820Table = 0x8003,
-    HpetData = 0x8004,
-    Reserved = 0xffff,
-}
-
-impl Default for TdVmmDataRegionType {
-    fn default() -> Self {
-        TdVmmDataRegionType::Reserved
-    }
-}
-
 pub fn parse_tdvf_sections(file: &mut File) -> Result<Vec<TdvfSection>, TdvfError> {
     // The 32-bit offset to the TDVF metadata is located 32 bytes from
     // the end of the file.
@@ -128,7 +79,7 @@ pub fn parse_tdvf_sections(file: &mut File) -> Result<Vec<TdvfSection>, TdvfErro
     file.seek(SeekFrom::Start(descriptor_offset))
         .map_err(TdvfError::ReadDescriptor)?;
 
-    let mut descriptor: TdvfDescriptor = unsafe { std::mem::zeroed() };
+    let mut descriptor: TdvfDescriptor = Default::default();
     // Safe as we read exactly the size of the descriptor header
     file.read_exact(unsafe {
         std::slice::from_raw_parts_mut(
@@ -191,7 +142,6 @@ struct HobHeader {
     length: u16,
     reserved: u32,
 }
-unsafe impl ByteValued for HobHeader {}
 
 #[repr(C)]
 #[derive(Copy, Clone, Default, Debug)]
@@ -205,7 +155,6 @@ struct HobHandoffInfoTable {
     efi_free_memory_bottom: u64,
     efi_end_of_hob_list: u64,
 }
-unsafe impl ByteValued for HobHandoffInfoTable {}
 
 #[repr(C)]
 #[derive(Copy, Clone, Default, Debug)]
@@ -226,7 +175,6 @@ struct HobResourceDescriptor {
     physical_start: u64,
     resource_length: u64,
 }
-unsafe impl ByteValued for HobResourceDescriptor {}
 
 #[repr(C)]
 #[derive(Copy, Clone, Default, Debug)]
@@ -234,15 +182,42 @@ struct HobGuidType {
     header: HobHeader,
     name: EfiGuid,
 }
-unsafe impl ByteValued for HobGuidType {}
+
+#[repr(u32)]
+#[derive(Clone, Copy, Debug)]
+pub enum PayloadImageType {
+    ExecutablePayload,
+    BzImage,
+    RawVmLinux,
+}
+
+impl Default for PayloadImageType {
+    fn default() -> Self {
+        PayloadImageType::ExecutablePayload
+    }
+}
 
 #[repr(C)]
 #[derive(Copy, Clone, Default, Debug)]
-struct TdVmmData {
+pub struct PayloadInfo {
-    guid_type: HobGuidType,
+    pub image_type: PayloadImageType,
-    region: TdVmmDataRegion,
+    pub entry_point: u64,
 }
-unsafe impl ByteValued for TdVmmData {}
+
+#[repr(C)]
+#[derive(Copy, Clone, Default, Debug)]
+struct TdPayload {
+    guid_type: HobGuidType,
+    payload_info: PayloadInfo,
+}
+
+// SAFETY: These data structures only contain a series of integers
+unsafe impl ByteValued for HobHeader {}
+unsafe impl ByteValued for HobHandoffInfoTable {}
+unsafe impl ByteValued for HobResourceDescriptor {}
+unsafe impl ByteValued for HobGuidType {}
+unsafe impl ByteValued for PayloadInfo {}
+unsafe impl ByteValued for TdPayload {}
 
 pub struct TdHob {
     start_offset: u64,
@@ -374,35 +349,90 @@ impl TdHob {
         )
     }
 
-    pub fn add_td_vmm_data(
+    pub fn add_acpi_table(
         &mut self,
         mem: &GuestMemoryMmap,
-        region: TdVmmDataRegion,
+        table_content: &[u8],
     ) -> Result<(), TdvfError> {
-        let td_vmm_data = TdVmmData {
+        // We already know the HobGuidType size is 8 bytes multiple, but we
+        // need the total size to be 8 bytes multiple. That is why the ACPI
+        // table size must be 8 bytes multiple as well.
+        let length = std::mem::size_of::<HobGuidType>() as u16
+            + align_hob(table_content.len() as u64) as u16;
+        let hob_guid_type = HobGuidType {
+            header: HobHeader {
+                r#type: HobType::GuidExtension,
+                length,
+                reserved: 0,
+            },
+            // ACPI_TABLE_HOB_GUID
+            // 0x6a0c5870, 0xd4ed, 0x44f4, {0xa1, 0x35, 0xdd, 0x23, 0x8b, 0x6f, 0xc, 0x8d }
+            name: EfiGuid {
+                data1: 0x6a0c_5870,
+                data2: 0xd4ed,
+                data3: 0x44f4,
+                data4: [0xa1, 0x35, 0xdd, 0x23, 0x8b, 0x6f, 0xc, 0x8d],
+            },
+        };
+        info!(
+            "Writing HOB ACPI table {:x} {:x?} {:x?}",
+            self.current_offset, hob_guid_type, table_content
+        );
+        mem.write_obj(hob_guid_type, GuestAddress(self.current_offset))
+            .map_err(TdvfError::GuestMemoryWriteHob)?;
+        let current_offset = self.current_offset + std::mem::size_of::<HobGuidType>() as u64;
+
+        // In case the table is quite large, let's make sure we can handle
+        // retrying until everything has been correctly copied.
+        let mut offset: usize = 0;
+        loop {
+            let bytes_written = mem
+                .write(
+                    &table_content[offset..],
+                    GuestAddress(current_offset + offset as u64),
+                )
+                .map_err(TdvfError::GuestMemoryWriteHob)?;
+            offset += bytes_written;
+            if offset >= table_content.len() {
+                break;
+            }
+        }
+        self.current_offset += length as u64;
+
+        Ok(())
+    }
+
+    pub fn add_payload(
+        &mut self,
+        mem: &GuestMemoryMmap,
+        payload_info: PayloadInfo,
+    ) -> Result<(), TdvfError> {
+        let payload = TdPayload {
             guid_type: HobGuidType {
                 header: HobHeader {
                     r#type: HobType::GuidExtension,
-                    length: std::mem::size_of::<TdVmmData>() as u16,
+                    length: std::mem::size_of::<TdPayload>() as u16,
                     reserved: 0,
                 },
-                // TD_VMM_DATA_GUID CF2643E4-C0D3-46FF-0000-72EE623DDE38
+                // HOB_PAYLOAD_INFO_GUID
+                // 0xb96fa412, 0x461f, 0x4be3, {0x8c, 0xd, 0xad, 0x80, 0x5a, 0x49, 0x7a, 0xc0
                 name: EfiGuid {
-                    data1: 0xcf26_43e4,
+                    data1: 0xb96f_a412,
-                    data2: 0xc0d3,
+                    data2: 0x461f,
-                    data3: 0x46ff,
+                    data3: 0x4be3,
-                    data4: [0x00, 0x00, 0x72, 0xee, 0x62, 0x3d, 0xde, 0x38],
+                    data4: [0x8c, 0xd, 0xad, 0x80, 0x5a, 0x49, 0x7a, 0xc0],
                 },
             },
-            region,
+            payload_info,
         };
         info!(
-            "Writing HOB TD_VMM_DATA {:x} {:x?}",
+            "Writing HOB TD_PAYLOAD {:x} {:x?}",
-            self.current_offset, td_vmm_data
+            self.current_offset, payload
         );
-        mem.write_obj(td_vmm_data, GuestAddress(self.current_offset))
+        mem.write_obj(payload, GuestAddress(self.current_offset))
             .map_err(TdvfError::GuestMemoryWriteHob)?;
-        self.update_offset::<TdVmmData>();
+        self.update_offset::<TdPayload>();
+
         Ok(())
     }
 }

block_util/Cargo.toml

@@ -6,11 +6,10 @@ edition = "2018"
 
 [features]
 default = []
-io_uring = []
 
 [dependencies]
 io-uring = "0.5.2"
-libc = "0.2.103"
+libc = "0.2.119"
 log = "0.4.14"
 qcow = { path = "../qcow" }
 thiserror = "1.0.30"
@@ -18,7 +17,8 @@ versionize = "0.1.6"
 versionize_derive = "0.1.4"
 vhdx = { path = "../vhdx" }
 virtio-bindings = { version = "0.1.0", features = ["virtio-v5_0_0"] }
-vm-memory = { version = "0.6.0", features = ["backend-mmap", "backend-atomic", "backend-bitmap"] }
+virtio-queue = { git = "https://github.com/rust-vmm/vm-virtio", branch = "main" }
+vm-memory = { version = "0.7.0", features = ["backend-mmap", "backend-atomic", "backend-bitmap"] }
 vm-virtio = { path = "../vm-virtio" }
 vmm-sys-util = "0.9.0"
 

block_util/src/async_io.rs

@@ -2,8 +2,13 @@
 //
 // SPDX-License-Identifier: Apache-2.0 AND BSD-3-Clause
 
+use libc::{ioctl, S_IFBLK, S_IFMT};
+use std::convert::TryInto;
+use std::fs::File;
+use std::os::unix::io::AsRawFd;
 use thiserror::Error;
 use vmm_sys_util::eventfd::EventFd;
+use vmm_sys_util::{ioctl_expr, ioctl_io_nr, ioctl_ioc_nr};
 
 #[derive(Error, Debug)]
 pub enum DiskFileError {
@@ -15,11 +20,96 @@ pub enum DiskFileError {
     NewAsyncIo(#[source] std::io::Error),
 }
 
+#[derive(Debug)]
+pub struct DiskTopology {
+    pub logical_block_size: u64,
+    pub physical_block_size: u64,
+    pub minimum_io_size: u64,
+    pub optimal_io_size: u64,
+}
+
+impl Default for DiskTopology {
+    fn default() -> Self {
+        Self {
+            logical_block_size: 512,
+            physical_block_size: 512,
+            minimum_io_size: 512,
+            optimal_io_size: 0,
+        }
+    }
+}
+
+ioctl_io_nr!(BLKSSZGET, 0x12, 104);
+ioctl_io_nr!(BLKPBSZGET, 0x12, 123);
+ioctl_io_nr!(BLKIOMIN, 0x12, 120);
+ioctl_io_nr!(BLKIOOPT, 0x12, 121);
+
+enum BlockSize {
+    LogicalBlock,
+    PhysicalBlock,
+    MinimumIo,
+    OptimalIo,
+}
+
+impl DiskTopology {
+    fn is_block_device(f: &mut File) -> std::io::Result<bool> {
+        let mut stat = std::mem::MaybeUninit::<libc::stat>::uninit();
+        let ret = unsafe { libc::fstat(f.as_raw_fd(), stat.as_mut_ptr()) };
+        if ret != 0 {
+            return Err(std::io::Error::last_os_error());
+        }
+
+        let is_block = unsafe { (*stat.as_ptr()).st_mode & S_IFMT == S_IFBLK };
+        Ok(is_block)
+    }
+
+    // libc::ioctl() takes different types on different architectures
+    #[allow(clippy::useless_conversion)]
+    fn query_block_size(f: &mut File, block_size_type: BlockSize) -> std::io::Result<u64> {
+        let mut block_size = 0;
+        let ret = unsafe {
+            ioctl(
+                f.as_raw_fd(),
+                match block_size_type {
+                    BlockSize::LogicalBlock => BLKSSZGET(),
+                    BlockSize::PhysicalBlock => BLKPBSZGET(),
+                    BlockSize::MinimumIo => BLKIOMIN(),
+                    BlockSize::OptimalIo => BLKIOOPT(),
+                }
+                .try_into()
+                .unwrap(),
+                &mut block_size,
+            )
+        };
+        if ret != 0 {
+            return Err(std::io::Error::last_os_error());
+        };
+
+        Ok(block_size)
+    }
+
+    pub fn probe(f: &mut File) -> std::io::Result<Self> {
+        if !Self::is_block_device(f)? {
+            return Ok(DiskTopology::default());
+        }
+
+        Ok(DiskTopology {
+            logical_block_size: Self::query_block_size(f, BlockSize::LogicalBlock)?,
+            physical_block_size: Self::query_block_size(f, BlockSize::PhysicalBlock)?,
+            minimum_io_size: Self::query_block_size(f, BlockSize::MinimumIo)?,
+            optimal_io_size: Self::query_block_size(f, BlockSize::OptimalIo)?,
+        })
+    }
+}
+
 pub type DiskFileResult<T> = std::result::Result<T, DiskFileError>;
 
 pub trait DiskFile: Send + Sync {
     fn size(&mut self) -> DiskFileResult<u64>;
     fn new_async_io(&self, ring_depth: u32) -> DiskFileResult<Box<dyn AsyncIo>>;
+    fn topology(&mut self) -> DiskTopology {
+        DiskTopology::default()
+    }
 }
 
 #[derive(Error, Debug)]

block_util/src/lib.rs

@@ -20,27 +20,27 @@ pub mod raw_sync;
 pub mod vhd;
 pub mod vhdx_sync;
 
-use crate::async_io::{AsyncIo, AsyncIoError, AsyncIoResult, DiskFileError, DiskFileResult};
+use crate::async_io::{AsyncIo, AsyncIoError, AsyncIoResult};
-#[cfg(feature = "io_uring")]
 use io_uring::{opcode, IoUring, Probe};
+use std::alloc::{alloc_zeroed, dealloc, Layout};
 use std::cmp;
 use std::convert::TryInto;
 use std::fs::File;
 use std::io::{self, IoSlice, IoSliceMut, Read, Seek, SeekFrom, Write};
 use std::os::linux::fs::MetadataExt;
-#[cfg(feature = "io_uring")]
-use std::os::unix::io::AsRawFd;
 use std::path::Path;
 use std::result;
-use std::sync::{Arc, Mutex};
+use std::sync::Arc;
+use std::sync::MutexGuard;
 use versionize::{VersionMap, Versionize, VersionizeResult};
 use versionize_derive::Versionize;
 use virtio_bindings::bindings::virtio_blk::*;
+use virtio_queue::DescriptorChain;
 use vm_memory::{
     bitmap::AtomicBitmap, bitmap::Bitmap, ByteValued, Bytes, GuestAddress, GuestMemory,
-    GuestMemoryError,
+    GuestMemoryError, GuestMemoryLoadGuard,
 };
-use vm_virtio::DescriptorChain;
+use vm_virtio::{AccessPlatform, Translatable};
 use vmm_sys_util::eventfd::EventFd;
 
 type GuestMemoryMmap = vm_memory::GuestMemoryMmap<AtomicBitmap>;
@@ -115,6 +115,8 @@ pub enum ExecuteError {
     AsyncRead(AsyncIoError),
     AsyncWrite(AsyncIoError),
     AsyncFlush(AsyncIoError),
+    /// Failed allocating a temporary buffer.
+    TemporaryBufferAllocation(io::Error),
 }
 
 impl ExecuteError {
@@ -131,6 +133,7 @@ impl ExecuteError {
             ExecuteError::AsyncRead(_) => VIRTIO_BLK_S_IOERR,
             ExecuteError::AsyncWrite(_) => VIRTIO_BLK_S_IOERR,
             ExecuteError::AsyncFlush(_) => VIRTIO_BLK_S_IOERR,
+            ExecuteError::TemporaryBufferAllocation(_) => VIRTIO_BLK_S_IOERR,
         }
     }
 }
@@ -168,6 +171,14 @@ fn sector(mem: &GuestMemoryMmap, desc_addr: GuestAddress) -> result::Result<u64,
     mem.read_obj(addr).map_err(Error::GuestMemory)
 }
 
+#[derive(Debug)]
+pub struct AlignedOperation {
+    origin_ptr: u64,
+    aligned_ptr: u64,
+    size: usize,
+    layout: Layout,
+}
+
 #[derive(Debug)]
 pub struct Request {
     pub request_type: RequestType,
@@ -175,29 +186,43 @@ pub struct Request {
     pub data_descriptors: Vec<(GuestAddress, u32)>,
     pub status_addr: GuestAddress,
     pub writeback: bool,
+    pub aligned_operations: Vec<AlignedOperation>,
 }
 
 impl Request {
     pub fn parse(
-        avail_desc: &DescriptorChain,
+        desc_chain: &mut DescriptorChain<GuestMemoryLoadGuard<GuestMemoryMmap>>,
-        mem: &GuestMemoryMmap,
+        access_platform: Option<&Arc<dyn AccessPlatform>>,
     ) -> result::Result<Request, Error> {
+        let hdr_desc = desc_chain
+            .next()
+            .ok_or(Error::DescriptorChainTooShort)
+            .map_err(|e| {
+                error!("Missing head descriptor");
+                e
+            })?;
+
         // The head contains the request type which MUST be readable.
-        if avail_desc.is_write_only() {
+        if hdr_desc.is_write_only() {
             return Err(Error::UnexpectedWriteOnlyDescriptor);
         }
 
+        let hdr_desc_addr = hdr_desc
+            .addr()
+            .translate(access_platform, hdr_desc.len() as usize);
+
         let mut req = Request {
-            request_type: request_type(mem, avail_desc.addr)?,
+            request_type: request_type(desc_chain.memory(), hdr_desc_addr)?,
-            sector: sector(mem, avail_desc.addr)?,
+            sector: sector(desc_chain.memory(), hdr_desc_addr)?,
             data_descriptors: Vec::new(),
             status_addr: GuestAddress(0),
             writeback: true,
+            aligned_operations: Vec::new(),
         };
 
         let status_desc;
-        let mut desc = avail_desc
+        let mut desc = desc_chain
-            .next_descriptor()
+            .next()
             .ok_or(Error::DescriptorChainTooShort)
             .map_err(|e| {
                 error!("Only head descriptor present: request = {:?}", req);
@@ -222,9 +247,13 @@ impl Request {
                 if !desc.is_write_only() && req.request_type == RequestType::GetDeviceId {
                     return Err(Error::UnexpectedReadOnlyDescriptor);
                 }
-                req.data_descriptors.push((desc.addr, desc.len));
+
-                desc = desc
+                req.data_descriptors.push((
-                    .next_descriptor()
+                    desc.addr().translate(access_platform, desc.len() as usize),
+                    desc.len(),
+                ));
+                desc = desc_chain
+                    .next()
                     .ok_or(Error::DescriptorChainTooShort)
                     .map_err(|e| {
                         error!("DescriptorChain corrupted: request = {:?}", req);
@@ -239,22 +268,23 @@ impl Request {
             return Err(Error::UnexpectedReadOnlyDescriptor);
         }
 
-        if status_desc.len < 1 {
+        if status_desc.len() < 1 {
             return Err(Error::DescriptorLengthTooSmall);
         }
 
-        req.status_addr = status_desc.addr;
+        req.status_addr = status_desc
+            .addr()
+            .translate(access_platform, status_desc.len() as usize);
 
         Ok(req)
     }
 
-    #[allow(clippy::ptr_arg)]
     pub fn execute<T: Seek + Read + Write>(
         &self,
         disk: &mut T,
         disk_nsectors: u64,
         mem: &GuestMemoryMmap,
-        disk_id: &Vec<u8>,
+        disk_id: &[u8],
     ) -> result::Result<u32, ExecuteError> {
         disk.seek(SeekFrom::Start(self.sector << SECTOR_SHIFT))
             .map_err(ExecuteError::Seek)?;
@@ -289,7 +319,7 @@ impl Request {
                     if (*data_len as usize) < disk_id.len() {
                         return Err(ExecuteError::BadRequest(Error::InvalidOffset));
                     }
-                    mem.write_slice(disk_id.as_slice(), *data_addr)
+                    mem.write_slice(disk_id, *data_addr)
                         .map_err(ExecuteError::Write)?;
                 }
                 RequestType::Unsupported(t) => return Err(ExecuteError::Unsupported(t)),
@@ -299,7 +329,7 @@ impl Request {
     }
 
     pub fn execute_async(
-        &self,
+        &mut self,
         mem: &GuestMemoryMmap,
         disk_nsectors: u64,
         disk_image: &mut dyn AsyncIo,
@@ -312,6 +342,9 @@ impl Request {
 
         let mut iovecs = Vec::new();
         for (data_addr, data_len) in &self.data_descriptors {
+            if *data_len == 0 {
+                continue;
+            }
             let mut top: u64 = u64::from(*data_len) / SECTOR_SIZE;
             if u64::from(*data_len) % SECTOR_SIZE != 0 {
                 top += 1;
@@ -323,12 +356,52 @@ impl Request {
                 return Err(ExecuteError::BadRequest(Error::InvalidOffset));
             }
 
-            let buf = mem
+            let origin_ptr = mem
                 .get_slice(*data_addr, *data_len as usize)
                 .map_err(ExecuteError::GetHostAddress)?
                 .as_ptr();
+
+            // Verify the buffer alignment.
+            // In case it's not properly aligned, an intermediate buffer is
+            // created with the correct alignment, and a copy from/to the
+            // origin buffer is performed, depending on the type of operation.
+            let iov_base = if (origin_ptr as u64) % SECTOR_SIZE != 0 {
+                let layout =
+                    Layout::from_size_align(*data_len as usize, SECTOR_SIZE as usize).unwrap();
+                // Safe because layout has non-zero size
+                let aligned_ptr = unsafe { alloc_zeroed(layout) };
+                if aligned_ptr.is_null() {
+                    return Err(ExecuteError::TemporaryBufferAllocation(
+                        io::Error::last_os_error(),
+                    ));
+                }
+
+                // We need to perform the copy beforehand in case we're writing
+                // data out.
+                if request_type == RequestType::Out {
+                    // Safe because destination buffer has been allocated with
+                    // the proper size.
+                    unsafe {
+                        std::ptr::copy(origin_ptr as *const u8, aligned_ptr, *data_len as usize)
+                    };
+                }
+
+                // Store both origin and aligned pointers for complete_async()
+                // to process them.
+                self.aligned_operations.push(AlignedOperation {
+                    origin_ptr: origin_ptr as u64,
+                    aligned_ptr: aligned_ptr as u64,
+                    size: *data_len as usize,
+                    layout,
+                });
+
+                aligned_ptr as *mut libc::c_void
+            } else {
+                origin_ptr as *mut libc::c_void
+            };
+
             let iovec = libc::iovec {
-                iov_base: buf as *mut libc::c_void,
+                iov_base,
                 iov_len: *data_len as libc::size_t,
             };
             iovecs.push(iovec);
@@ -376,6 +449,36 @@ impl Request {
         Ok(true)
     }
 
+    pub fn complete_async(&mut self) -> result::Result<(), Error> {
+        for aligned_operation in self.aligned_operations.drain(..) {
+            // We need to perform the copy after the data has been read inside
+            // the aligned buffer in case we're reading data in.
+            if self.request_type == RequestType::In {
+                // Safe because origin buffer has been allocated with the
+                // proper size.
+                unsafe {
+                    std::ptr::copy(
+                        aligned_operation.aligned_ptr as *const u8,
+                        aligned_operation.origin_ptr as *mut u8,
+                        aligned_operation.size,
+                    )
+                };
+            }
+
+            // Free the temporary aligned buffer.
+            // Safe because aligned_ptr was allocated by alloc_zeroed with the same
+            // layout
+            unsafe {
+                dealloc(
+                    aligned_operation.aligned_ptr as *mut u8,
+                    aligned_operation.layout,
+                )
+            };
+        }
+
+        Ok(())
+    }
+
     pub fn set_writeback(&mut self, writeback: bool) {
         self.writeback = writeback
     }
@@ -404,8 +507,6 @@ pub struct VirtioBlockConfig {
     pub write_zeroes_may_unmap: u8,
     pub unused1: [u8; 3],
 }
-unsafe impl ByteValued for VirtioBlockConfig {}
-
 #[derive(Copy, Clone, Debug, Default, Versionize)]
 #[repr(C, packed)]
 pub struct VirtioBlockGeometry {
@@ -413,11 +514,13 @@ pub struct VirtioBlockGeometry {
     pub heads: u8,
     pub sectors: u8,
 }
+
+// SAFETY: these data structures only contain a series of integers
+unsafe impl ByteValued for VirtioBlockConfig {}
 unsafe impl ByteValued for VirtioBlockGeometry {}
 
 /// Check if io_uring for block device can be used on the current system, as
 /// it correctly supports the expected io_uring features.
-#[cfg(feature = "io_uring")]
 pub fn block_io_uring_is_supported() -> bool {
     let error_msg = "io_uring not supported:";
 
@@ -433,25 +536,6 @@ pub fn block_io_uring_is_supported() -> bool {
 
     let submitter = io_uring.submitter();
 
-    let event_fd = match EventFd::new(libc::EFD_NONBLOCK) {
-        Ok(fd) => fd,
-        Err(e) => {
-            info!("{} failed to create eventfd: {}", error_msg, e);
-            return false;
-        }
-    };
-
-    // Check we can register an eventfd as this is going to be needed while
-    // using io_uring with the virtio block device. This also validates that
-    // io_uring_register() syscall is supported.
-    match submitter.register_eventfd(event_fd.as_raw_fd()) {
-        Ok(_) => {}
-        Err(e) => {
-            info!("{} failed to register eventfd: {}", error_msg, e);
-            return false;
-        }
-    }
-
     let mut probe = Probe::new();
 
     // Check we can register a probe to validate supported operations.
@@ -484,119 +568,98 @@ pub fn block_io_uring_is_supported() -> bool {
     true
 }
 
-#[cfg(not(feature = "io_uring"))]
+pub trait AsyncAdaptor<F>
-pub fn block_io_uring_is_supported() -> bool {
+where
-    false
+    F: Read + Write + Seek,
-}
+{
+    fn read_vectored_sync(
+        &mut self,
+        offset: libc::off_t,
+        iovecs: Vec<libc::iovec>,
+        user_data: u64,
+        eventfd: &EventFd,
+        completion_list: &mut Vec<(u64, i32)>,
+    ) -> AsyncIoResult<()> {
+        // Convert libc::iovec into IoSliceMut
+        let mut slices = Vec::new();
+        for iovec in iovecs.iter() {
+            slices.push(IoSliceMut::new(unsafe { std::mem::transmute(*iovec) }));
+        }
 
-pub fn disk_size(file: &mut dyn Seek, semaphore: &mut Arc<Mutex<()>>) -> DiskFileResult<u64> {
+        let result = {
-    // Take the semaphore to ensure other threads are not interacting with
+            let mut file = self.file();
-    // the underlying file.
-    let _lock = semaphore.lock().unwrap();
 
-    Ok(file.seek(SeekFrom::End(0)).map_err(DiskFileError::Size)? as u64)
+            // Move the cursor to the right offset
-}
+            file.seek(SeekFrom::Start(offset as u64))
+                .map_err(AsyncIoError::ReadVectored)?;
 
-pub trait ReadSeekFile: Read + Seek {}
+            // Read vectored
-impl<F: Read + Seek> ReadSeekFile for F {}
+            file.read_vectored(slices.as_mut_slice())
+                .map_err(AsyncIoError::ReadVectored)?
+        };
 
-pub fn read_vectored_sync(
+        completion_list.push((user_data, result as i32));
-    offset: libc::off_t,
-    iovecs: Vec<libc::iovec>,
-    user_data: u64,
-    file: &mut dyn ReadSeekFile,
-    eventfd: &EventFd,
-    completion_list: &mut Vec<(u64, i32)>,
-    semaphore: &mut Arc<Mutex<()>>,
-) -> AsyncIoResult<()> {
-    // Convert libc::iovec into IoSliceMut
-    let mut slices = Vec::new();
-    for iovec in iovecs.iter() {
-        slices.push(IoSliceMut::new(unsafe { std::mem::transmute(*iovec) }));
-    }
-
-    let result = {
-        // Take the semaphore to ensure other threads are not interacting
-        // with the underlying file.
-        let _lock = semaphore.lock().unwrap();
-
-        // Move the cursor to the right offset
-        file.seek(SeekFrom::Start(offset as u64))
-            .map_err(AsyncIoError::ReadVectored)?;
-
-        // Read vectored
-        file.read_vectored(slices.as_mut_slice())
-            .map_err(AsyncIoError::ReadVectored)?
-    };
-
-    completion_list.push((user_data, result as i32));
-    eventfd.write(1).unwrap();
-
-    Ok(())
-}
-
-pub trait WriteSeekFile: Write + Seek {}
-impl<F: Write + Seek> WriteSeekFile for F {}
-
-pub fn write_vectored_sync(
-    offset: libc::off_t,
-    iovecs: Vec<libc::iovec>,
-    user_data: u64,
-    file: &mut dyn WriteSeekFile,
-    eventfd: &EventFd,
-    completion_list: &mut Vec<(u64, i32)>,
-    semaphore: &mut Arc<Mutex<()>>,
-) -> AsyncIoResult<()> {
-    // Convert libc::iovec into IoSlice
-    let mut slices = Vec::new();
-    for iovec in iovecs.iter() {
-        slices.push(IoSlice::new(unsafe { std::mem::transmute(*iovec) }));
-    }
-
-    let result = {
-        // Take the semaphore to ensure other threads are not interacting
-        // with the underlying file.
-        let _lock = semaphore.lock().unwrap();
-
-        // Move the cursor to the right offset
-        file.seek(SeekFrom::Start(offset as u64))
-            .map_err(AsyncIoError::WriteVectored)?;
-
-        // Write vectored
-        file.write_vectored(slices.as_slice())
-            .map_err(AsyncIoError::WriteVectored)?
-    };
-
-    completion_list.push((user_data, result as i32));
-    eventfd.write(1).unwrap();
-
-    Ok(())
-}
-
-pub fn fsync_sync(
-    user_data: Option<u64>,
-    file: &mut dyn Write,
-    eventfd: &EventFd,
-    completion_list: &mut Vec<(u64, i32)>,
-    semaphore: &mut Arc<Mutex<()>>,
-) -> AsyncIoResult<()> {
-    let result: i32 = {
-        // Take the semaphore to ensure other threads are not interacting
-        // with the underlying file.
-        let _lock = semaphore.lock().unwrap();
-
-        // Flush
-        file.flush().map_err(AsyncIoError::Fsync)?;
-
-        0
-    };
-
-    if let Some(user_data) = user_data {
-        completion_list.push((user_data, result));
         eventfd.write(1).unwrap();
+
+        Ok(())
     }
 
-    Ok(())
+    fn write_vectored_sync(
+        &mut self,
+        offset: libc::off_t,
+        iovecs: Vec<libc::iovec>,
+        user_data: u64,
+        eventfd: &EventFd,
+        completion_list: &mut Vec<(u64, i32)>,
+    ) -> AsyncIoResult<()> {
+        // Convert libc::iovec into IoSlice
+        let mut slices = Vec::new();
+        for iovec in iovecs.iter() {
+            slices.push(IoSlice::new(unsafe { std::mem::transmute(*iovec) }));
+        }
+
+        let result = {
+            let mut file = self.file();
+
+            // Move the cursor to the right offset
+            file.seek(SeekFrom::Start(offset as u64))
+                .map_err(AsyncIoError::WriteVectored)?;
+
+            // Write vectored
+            file.write_vectored(slices.as_slice())
+                .map_err(AsyncIoError::WriteVectored)?
+        };
+
+        completion_list.push((user_data, result as i32));
+        eventfd.write(1).unwrap();
+
+        Ok(())
+    }
+
+    fn fsync_sync(
+        &mut self,
+        user_data: Option<u64>,
+        eventfd: &EventFd,
+        completion_list: &mut Vec<(u64, i32)>,
+    ) -> AsyncIoResult<()> {
+        let result: i32 = {
+            let mut file = self.file();
+
+            // Flush
+            file.flush().map_err(AsyncIoError::Fsync)?;
+
+            0
+        };
+
+        if let Some(user_data) = user_data {
+            completion_list.push((user_data, result));
+            eventfd.write(1).unwrap();
+        }
+
+        Ok(())
+    }
+
+    fn file(&mut self) -> MutexGuard<F>;
 }
 
 pub enum ImageType {

block_util/src/qcow_sync.rs

@@ -2,59 +2,61 @@
 //
 // SPDX-License-Identifier: Apache-2.0 AND BSD-3-Clause
 
-use crate::async_io::{AsyncIo, AsyncIoResult, DiskFile, DiskFileResult};
+use crate::async_io::{AsyncIo, AsyncIoResult, DiskFile, DiskFileError, DiskFileResult};
-use crate::{disk_size, fsync_sync, read_vectored_sync, write_vectored_sync};
+use crate::AsyncAdaptor;
 use qcow::{QcowFile, RawFile, Result as QcowResult};
 use std::fs::File;
-use std::sync::{Arc, Mutex};
+use std::io::{Seek, SeekFrom};
+use std::sync::{Arc, Mutex, MutexGuard};
 use vmm_sys_util::eventfd::EventFd;
 
 pub struct QcowDiskSync {
-    qcow_file: QcowFile,
+    qcow_file: Arc<Mutex<QcowFile>>,
-    semaphore: Arc<Mutex<()>>,
 }
 
 impl QcowDiskSync {
     pub fn new(file: File, direct_io: bool) -> QcowResult<Self> {
         Ok(QcowDiskSync {
-            qcow_file: QcowFile::from(RawFile::new(file, direct_io))?,
+            qcow_file: Arc::new(Mutex::new(QcowFile::from(RawFile::new(file, direct_io))?)),
-            semaphore: Arc::new(Mutex::new(())),
         })
     }
 }
 
 impl DiskFile for QcowDiskSync {
     fn size(&mut self) -> DiskFileResult<u64> {
-        disk_size(&mut self.qcow_file, &mut self.semaphore)
+        let mut file = self.qcow_file.lock().unwrap();
+
+        Ok(file.seek(SeekFrom::End(0)).map_err(DiskFileError::Size)? as u64)
     }
 
     fn new_async_io(&self, _ring_depth: u32) -> DiskFileResult<Box<dyn AsyncIo>> {
-        Ok(Box::new(QcowSync::new(
+        Ok(Box::new(QcowSync::new(self.qcow_file.clone())) as Box<dyn AsyncIo>)
-            self.qcow_file.clone(),
-            self.semaphore.clone(),
-        )) as Box<dyn AsyncIo>)
     }
 }
 
 pub struct QcowSync {
-    qcow_file: QcowFile,
+    qcow_file: Arc<Mutex<QcowFile>>,
     eventfd: EventFd,
     completion_list: Vec<(u64, i32)>,
-    semaphore: Arc<Mutex<()>>,
 }
 
 impl QcowSync {
-    pub fn new(qcow_file: QcowFile, semaphore: Arc<Mutex<()>>) -> Self {
+    pub fn new(qcow_file: Arc<Mutex<QcowFile>>) -> Self {
         QcowSync {
             qcow_file,
             eventfd: EventFd::new(libc::EFD_NONBLOCK)
                 .expect("Failed creating EventFd for QcowSync"),
             completion_list: Vec::new(),
-            semaphore,
         }
     }
 }
 
+impl AsyncAdaptor<QcowFile> for Arc<Mutex<QcowFile>> {
+    fn file(&mut self) -> MutexGuard<QcowFile> {
+        self.lock().unwrap()
+    }
+}
+
 impl AsyncIo for QcowSync {
     fn notifier(&self) -> &EventFd {
         &self.eventfd
@@ -66,14 +68,12 @@ impl AsyncIo for QcowSync {
         iovecs: Vec<libc::iovec>,
         user_data: u64,
     ) -> AsyncIoResult<()> {
-        read_vectored_sync(
+        self.qcow_file.read_vectored_sync(
             offset,
             iovecs,
             user_data,
-            &mut self.qcow_file,
             &self.eventfd,
             &mut self.completion_list,
-            &mut self.semaphore,
         )
     }
 
@@ -83,25 +83,18 @@ impl AsyncIo for QcowSync {
         iovecs: Vec<libc::iovec>,
         user_data: u64,
     ) -> AsyncIoResult<()> {
-        write_vectored_sync(
+        self.qcow_file.write_vectored_sync(
             offset,
             iovecs,
             user_data,
-            &mut self.qcow_file,
             &self.eventfd,
             &mut self.completion_list,
-            &mut self.semaphore,
         )
     }
 
     fn fsync(&mut self, user_data: Option<u64>) -> AsyncIoResult<()> {
-        fsync_sync(
+        self.qcow_file
-            user_data,
+            .fsync_sync(user_data, &self.eventfd, &mut self.completion_list)
-            &mut self.qcow_file,
-            &self.eventfd,
-            &mut self.completion_list,
-            &mut self.semaphore,
-        )
     }
 
     fn complete(&mut self) -> Vec<(u64, i32)> {

block_util/src/raw_async.rs

@@ -3,7 +3,7 @@
 // SPDX-License-Identifier: Apache-2.0 AND BSD-3-Clause
 
 use crate::async_io::{
-    AsyncIo, AsyncIoError, AsyncIoResult, DiskFile, DiskFileError, DiskFileResult,
+    AsyncIo, AsyncIoError, AsyncIoResult, DiskFile, DiskFileError, DiskFileResult, DiskTopology,
 };
 use io_uring::{opcode, squeue, types, IoUring};
 use std::fs::File;
@@ -35,6 +35,15 @@ impl DiskFile for RawFileDisk {
                 .map_err(DiskFileError::NewAsyncIo)?,
         ) as Box<dyn AsyncIo>)
     }
+
+    fn topology(&mut self) -> DiskTopology {
+        if let Ok(topology) = DiskTopology::probe(&mut self.file) {
+            topology
+        } else {
+            warn!("Unable to get device topology. Using default topology");
+            DiskTopology::default()
+        }
+    }
 }
 
 pub struct RawFileAsync {

block_util/src/raw_sync.rs

@@ -3,7 +3,7 @@
 // SPDX-License-Identifier: Apache-2.0 AND BSD-3-Clause
 
 use crate::async_io::{
-    AsyncIo, AsyncIoError, AsyncIoResult, DiskFile, DiskFileError, DiskFileResult,
+    AsyncIo, AsyncIoError, AsyncIoResult, DiskFile, DiskFileError, DiskFileResult, DiskTopology,
 };
 use std::fs::File;
 use std::io::{Seek, SeekFrom};
@@ -31,6 +31,15 @@ impl DiskFile for RawFileDiskSync {
     fn new_async_io(&self, _ring_depth: u32) -> DiskFileResult<Box<dyn AsyncIo>> {
         Ok(Box::new(RawFileSync::new(self.file.as_raw_fd())) as Box<dyn AsyncIo>)
     }
+
+    fn topology(&mut self) -> DiskTopology {
+        if let Ok(topology) = DiskTopology::probe(&mut self.file) {
+            topology
+        } else {
+            warn!("Unable to get device topology. Using default topology");
+            DiskTopology::default()
+        }
+    }
 }
 
 pub struct RawFileSync {

block_util/src/vhdx_sync.rs

@@ -3,26 +3,20 @@
 // SPDX-License-Identifier: Apache-2.0
 
 use crate::async_io::{AsyncIo, AsyncIoResult, DiskFile, DiskFileError, DiskFileResult};
-use crate::{fsync_sync, read_vectored_sync, write_vectored_sync};
+use crate::AsyncAdaptor;
 use std::fs::File;
-use std::ops::DerefMut;
+use std::sync::{Arc, Mutex, MutexGuard};
-use std::sync::{Arc, Mutex};
 use vhdx::vhdx::{Result as VhdxResult, Vhdx};
 use vmm_sys_util::eventfd::EventFd;
 
 pub struct VhdxDiskSync {
     vhdx_file: Arc<Mutex<Vhdx>>,
-    semaphore: Arc<Mutex<()>>,
 }
 
 impl VhdxDiskSync {
     pub fn new(f: File) -> VhdxResult<Self> {
-        let vhdx = Vhdx::new(f)?;
-        let vhdx_file = Arc::new(Mutex::new(vhdx));
-
         Ok(VhdxDiskSync {
-            vhdx_file,
+            vhdx_file: Arc::new(Mutex::new(Vhdx::new(f)?)),
-            semaphore: Arc::new(Mutex::new(())),
         })
     }
 }
@@ -33,10 +27,10 @@ impl DiskFile for VhdxDiskSync {
     }
 
     fn new_async_io(&self, _ring_depth: u32) -> DiskFileResult<Box<dyn AsyncIo>> {
-        Ok(Box::new(
+        Ok(
-            VhdxSync::new(self.vhdx_file.clone(), self.semaphore.clone())
+            Box::new(VhdxSync::new(self.vhdx_file.clone()).map_err(DiskFileError::NewAsyncIo)?)
-                .map_err(DiskFileError::NewAsyncIo)?,
+                as Box<dyn AsyncIo>,
-        ) as Box<dyn AsyncIo>)
+        )
     }
 }
 
@@ -44,20 +38,24 @@ pub struct VhdxSync {
     vhdx_file: Arc<Mutex<Vhdx>>,
     eventfd: EventFd,
     completion_list: Vec<(u64, i32)>,
-    semaphore: Arc<Mutex<()>>,
 }
 
 impl VhdxSync {
-    pub fn new(vhdx_file: Arc<Mutex<Vhdx>>, semaphore: Arc<Mutex<()>>) -> std::io::Result<Self> {
+    pub fn new(vhdx_file: Arc<Mutex<Vhdx>>) -> std::io::Result<Self> {
         Ok(VhdxSync {
             vhdx_file,
             eventfd: EventFd::new(libc::EFD_NONBLOCK)?,
             completion_list: Vec::new(),
-            semaphore,
         })
     }
 }
 
+impl AsyncAdaptor<Vhdx> for Arc<Mutex<Vhdx>> {
+    fn file(&mut self) -> MutexGuard<Vhdx> {
+        self.lock().unwrap()
+    }
+}
+
 impl AsyncIo for VhdxSync {
     fn notifier(&self) -> &EventFd {
         &self.eventfd
@@ -69,14 +67,12 @@ impl AsyncIo for VhdxSync {
         iovecs: Vec<libc::iovec>,
         user_data: u64,
     ) -> AsyncIoResult<()> {
-        read_vectored_sync(
+        self.vhdx_file.read_vectored_sync(
             offset,
             iovecs,
             user_data,
-            self.vhdx_file.lock().unwrap().deref_mut(),
             &self.eventfd,
             &mut self.completion_list,
-            &mut self.semaphore,
         )
     }
 
@@ -86,25 +82,18 @@ impl AsyncIo for VhdxSync {
         iovecs: Vec<libc::iovec>,
         user_data: u64,
     ) -> AsyncIoResult<()> {
-        write_vectored_sync(
+        self.vhdx_file.write_vectored_sync(
             offset,
             iovecs,
             user_data,
-            self.vhdx_file.lock().unwrap().deref_mut(),
             &self.eventfd,
             &mut self.completion_list,
-            &mut self.semaphore,
         )
     }
 
     fn fsync(&mut self, user_data: Option<u64>) -> AsyncIoResult<()> {
-        fsync_sync(
+        self.vhdx_file
-            user_data,
+            .fsync_sync(user_data, &self.eventfd, &mut self.completion_list)
-            self.vhdx_file.lock().unwrap().deref_mut(),
-            &self.eventfd,
-            &mut self.completion_list,
-            &mut self.semaphore,
-        )
     }
 
     fn complete(&mut self) -> Vec<(u64, i32)> {

devices/Cargo.toml

@@ -6,17 +6,17 @@ edition = "2018"
 
 [dependencies]
 acpi_tables = { path = "../acpi_tables", optional = true }
-anyhow = "1.0.44"
+anyhow = "1.0.55"
 arch = { path = "../arch" }
 bitflags = "1.3.2"
 byteorder = "1.4.3"
 epoll = "4.3.1"
-libc = "0.2.103"
+libc = "0.2.119"
 log = "0.4.14"
 versionize = "0.1.6"
 versionize_derive = "0.1.4"
 vm-device = { path = "../vm-device" }
-vm-memory = "0.6.0"
+vm-memory = "0.7.0"
 vm-migration = { path = "../vm-migration" }
 vmm-sys-util = "0.9.0"
 

devices/src/acpi.rs

@@ -104,7 +104,7 @@ impl BusDevice for AcpiGedDevice {
 
 #[cfg(feature = "acpi")]
 impl Aml for AcpiGedDevice {
-    fn to_aml_bytes(&self) -> Vec<u8> {
+    fn append_aml_bytes(&self, bytes: &mut Vec<u8>) {
         aml::Device::new(
             "_SB_.GED_".into(),
             vec![
@@ -151,7 +151,7 @@ impl Aml for AcpiGedDevice {
                         &aml::And::new(&aml::Local(1), &aml::Local(0), &4usize),
                         &aml::If::new(
                             &aml::Equal::new(&aml::Local(1), &4usize),
-                            vec![&aml::MethodCall::new("\\_SB_.PCI0.PCNT".into(), vec![])],
+                            vec![&aml::MethodCall::new("\\_SB_.PHPR.PSCN".into(), vec![])],
                         ),
                         &aml::And::new(&aml::Local(1), &aml::Local(0), &8usize),
                         &aml::If::new(
@@ -165,7 +165,7 @@ impl Aml for AcpiGedDevice {
                 ),
             ],
         )
-        .to_aml_bytes()
+        .append_aml_bytes(bytes)
     }
 }
 

devices/src/legacy/cmos.rs

@@ -121,6 +121,9 @@ impl BusDevice for Cmos {
                     0x09 => to_bcd((year % 100) as u8),
                     // Bit 5 for 32kHz clock. Bit 7 for Update in Progress
                     0x0a => 1 << 5 | (update_in_progress as u8) << 7,
+                    // Bit 0-6 are reserved and must be 0.
+                    // Bit 7 must be 1 (CMOS has power)
+                    0x0d => 1 << 7,
                     0x32 => to_bcd(((year + 1900) / 100) as u8),
                     _ => {
                         // self.index is always guaranteed to be in range via INDEX_MASK.

devices/src/legacy/rtc_pl031.rs

@@ -291,14 +291,13 @@ impl Rtc {
 
 impl BusDevice for Rtc {
     fn read(&mut self, _base: u64, offset: u64, data: &mut [u8]) {
-        let v;
         let mut read_ok = true;
 
-        if (AMBA_ID_LOW..AMBA_ID_HIGH).contains(&offset) {
+        let v = if (AMBA_ID_LOW..AMBA_ID_HIGH).contains(&offset) {
             let index = ((offset - AMBA_ID_LOW) >> 2) as usize;
-            v = u32::from(PL031_ID[index]);
+            u32::from(PL031_ID[index])
         } else {
-            v = match offset {
+            match offset {
                 RTCDR => self.get_time(),
                 RTCMR => {
                     // Even though we are not implementing RTC alarm we return the last value
@@ -313,8 +312,8 @@ impl BusDevice for Rtc {
                     read_ok = false;
                     0
                 }
-            };
+            }
-        }
+        };
         if read_ok && data.len() <= 4 {
             write_le_u32(data, v);
         } else {

devices/src/legacy/uart_pl011.rs

@@ -294,19 +294,15 @@ impl Pl011 {
 
 impl BusDevice for Pl011 {
     fn read(&mut self, _base: u64, offset: u64, data: &mut [u8]) {
-        let v;
         let mut read_ok = true;
-        if (AMBA_ID_LOW..AMBA_ID_HIGH).contains(&(offset >> 2)) {
+        let v = if (AMBA_ID_LOW..AMBA_ID_HIGH).contains(&(offset >> 2)) {
             let index = ((offset - 0xfe0) >> 2) as usize;
-            v = u32::from(PL011_ID[index]);
+            u32::from(PL011_ID[index])
         } else {
-            v = match offset >> 2 {
+            match offset >> 2 {
                 UARTDR => {
-                    let c: u32;
-                    let r: u32;
-
                     self.flags &= !PL011_FLAG_RXFF;
-                    c = self.read_fifo.pop_front().unwrap_or_default().into();
+                    let c: u32 = self.read_fifo.pop_front().unwrap_or_default().into();
                     if self.read_count > 0 {
                         self.read_count -= 1;
                     }
@@ -317,8 +313,7 @@ impl BusDevice for Pl011 {
                         self.int_level &= !PL011_INT_RX;
                     }
                     self.rsr = c >> 8;
-                    r = c;
+                    c
-                    r
                 }
                 UARTRSR_UARTECR => self.rsr,
                 UARTFR => self.flags,
@@ -337,7 +332,7 @@ impl BusDevice for Pl011 {
                     0
                 }
             }
-        }
+        };
 
         if read_ok && data.len() <= 4 {
             write_le_u32(data, v);

docs/COPYING.docs

@@ -0,0 +1,399 @@
+The documentation in this directory is covered by the following license:
+
+Attribution 4.0 International
+
+=======================================================================
+
+Creative Commons Corporation ("Creative Commons") is not a law firm and
+does not provide legal services or legal advice. Distribution of
+Creative Commons public licenses does not create a lawyer-client or
+other relationship. Creative Commons makes its licenses and related
+information available on an "as-is" basis. Creative Commons gives no
+warranties regarding its licenses, any material licensed under their
+terms and conditions, or any related information. Creative Commons
+disclaims all liability for damages resulting from their use to the
+fullest extent possible.
+
+Using Creative Commons Public Licenses
+
+Creative Commons public licenses provide a standard set of terms and
+conditions that creators and other rights holders may use to share
+original works of authorship and other material subject to copyright
+and certain other rights specified in the public license below. The
+following considerations are for informational purposes only, are not
+exhaustive, and do not form part of our licenses.
+
+     Considerations for licensors: Our public licenses are
+     intended for use by those authorized to give the public
+     permission to use material in ways otherwise restricted by
+     copyright and certain other rights. Our licenses are
+     irrevocable. Licensors should read and understand the terms
+     and conditions of the license they choose before applying it.
+     Licensors should also secure all rights necessary before
+     applying our licenses so that the public can reuse the
+     material as expected. Licensors should clearly mark any
+     material not subject to the license. This includes other CC-
+     licensed material, or material used under an exception or
+     limitation to copyright. More considerations for licensors:
+    wiki.creativecommons.org/Considerations_for_licensors
+
+     Considerations for the public: By using one of our public
+     licenses, a licensor grants the public permission to use the
+     licensed material under specified terms and conditions. If
+     the licensor's permission is not necessary for any reason--for
+     example, because of any applicable exception or limitation to
+     copyright--then that use is not regulated by the license. Our
+     licenses grant only permissions under copyright and certain
+     other rights that a licensor has authority to grant. Use of
+     the licensed material may still be restricted for other
+     reasons, including because others have copyright or other
+     rights in the material. A licensor may make special requests,
+     such as asking that all changes be marked or described.
+     Although not required by our licenses, you are encouraged to
+     respect those requests where reasonable. More considerations
+     for the public:
+    wiki.creativecommons.org/Considerations_for_licensees
+
+=======================================================================
+
+Creative Commons Attribution 4.0 International Public License
+
+By exercising the Licensed Rights (defined below), You accept and agree
+to be bound by the terms and conditions of this Creative Commons
+Attribution 4.0 International Public License ("Public License"). To the
+extent this Public License may be interpreted as a contract, You are
+granted the Licensed Rights in consideration of Your acceptance of
+these terms and conditions, and the Licensor grants You such rights in
+consideration of benefits the Licensor receives from making the
+Licensed Material available under these terms and conditions.
+
+
+Section 1 -- Definitions.
+
+  a. Adapted Material means material subject to Copyright and Similar
+     Rights that is derived from or based upon the Licensed Material
+     and in which the Licensed Material is translated, altered,
+     arranged, transformed, or otherwise modified in a manner requiring
+     permission under the Copyright and Similar Rights held by the
+     Licensor. For purposes of this Public License, where the Licensed
+     Material is a musical work, performance, or sound recording,
+     Adapted Material is always produced where the Licensed Material is
+     synched in timed relation with a moving image.
+
+  b. Adapter's License means the license You apply to Your Copyright
+     and Similar Rights in Your contributions to Adapted Material in
+     accordance with the terms and conditions of this Public License.
+
+  c. Copyright and Similar Rights means copyright and/or similar rights
+     closely related to copyright including, without limitation,
+     performance, broadcast, sound recording, and Sui Generis Database
+     Rights, without regard to how the rights are labeled or
+     categorized. For purposes of this Public License, the rights
+     specified in Section 2(b)(1)-(2) are not Copyright and Similar
+     Rights.
+
+  d. Effective Technological Measures means those measures that, in the
+     absence of proper authority, may not be circumvented under laws
+     fulfilling obligations under Article 11 of the WIPO Copyright
+     Treaty adopted on December 20, 1996, and/or similar international
+     agreements.
+
+  e. Exceptions and Limitations means fair use, fair dealing, and/or
+     any other exception or limitation to Copyright and Similar Rights
+     that applies to Your use of the Licensed Material.
+
+  f. Licensed Material means the artistic or literary work, database,
+     or other material to which the Licensor applied this Public
+     License.
+
+  g. Licensed Rights means the rights granted to You subject to the
+     terms and conditions of this Public License, which are limited to
+     all Copyright and Similar Rights that apply to Your use of the
+     Licensed Material and that the Licensor has authority to license.
+
+  h. Licensor means the individual(s) or entity(ies) granting rights
+     under this Public License.
+
+  i. Share means to provide material to the public by any means or
+     process that requires permission under the Licensed Rights, such
+     as reproduction, public display, public performance, distribution,
+     dissemination, communication, or importation, and to make material
+     available to the public including in ways that members of the
+     public may access the material from a place and at a time
+     individually chosen by them.
+
+  j. Sui Generis Database Rights means rights other than copyright
+     resulting from Directive 96/9/EC of the European Parliament and of
+     the Council of 11 March 1996 on the legal protection of databases,
+     as amended and/or succeeded, as well as other essentially
+     equivalent rights anywhere in the world.
+
+  k. You means the individual or entity exercising the Licensed Rights
+     under this Public License. Your has a corresponding meaning.
+
+
+Section 2 -- Scope.
+
+  a. License grant.
+
+       1. Subject to the terms and conditions of this Public License,
+          the Licensor hereby grants You a worldwide, royalty-free,
+          non-sublicensable, non-exclusive, irrevocable license to
+          exercise the Licensed Rights in the Licensed Material to:
+
+            a. reproduce and Share the Licensed Material, in whole or
+               in part; and
+
+            b. produce, reproduce, and Share Adapted Material.
+
+       2. Exceptions and Limitations. For the avoidance of doubt, where
+          Exceptions and Limitations apply to Your use, this Public
+          License does not apply, and You do not need to comply with
+          its terms and conditions.
+
+       3. Term. The term of this Public License is specified in Section
+          6(a).
+
+       4. Media and formats; technical modifications allowed. The
+          Licensor authorizes You to exercise the Licensed Rights in
+          all media and formats whether now known or hereafter created,
+          and to make technical modifications necessary to do so. The
+          Licensor waives and/or agrees not to assert any right or
+          authority to forbid You from making technical modifications
+          necessary to exercise the Licensed Rights, including
+          technical modifications necessary to circumvent Effective
+          Technological Measures. For purposes of this Public License,
+          simply making modifications authorized by this Section 2(a)
+          (4) never produces Adapted Material.
+
+       5. Downstream recipients.
+
+            a. Offer from the Licensor -- Licensed Material. Every
+               recipient of the Licensed Material automatically
+               receives an offer from the Licensor to exercise the
+               Licensed Rights under the terms and conditions of this
+               Public License.
+
+            b. No downstream restrictions. You may not offer or impose
+               any additional or different terms or conditions on, or
+               apply any Effective Technological Measures to, the
+               Licensed Material if doing so restricts exercise of the
+               Licensed Rights by any recipient of the Licensed
+               Material.
+
+       6. No endorsement. Nothing in this Public License constitutes or
+          may be construed as permission to assert or imply that You
+          are, or that Your use of the Licensed Material is, connected
+          with, or sponsored, endorsed, or granted official status by,
+          the Licensor or others designated to receive attribution as
+          provided in Section 3(a)(1)(A)(i).
+
+  b. Other rights.
+
+       1. Moral rights, such as the right of integrity, are not
+          licensed under this Public License, nor are publicity,
+          privacy, and/or other similar personality rights; however, to
+          the extent possible, the Licensor waives and/or agrees not to
+          assert any such rights held by the Licensor to the limited
+          extent necessary to allow You to exercise the Licensed
+          Rights, but not otherwise.
+
+       2. Patent and trademark rights are not licensed under this
+          Public License.
+
+       3. To the extent possible, the Licensor waives any right to
+          collect royalties from You for the exercise of the Licensed
+          Rights, whether directly or through a collecting society
+          under any voluntary or waivable statutory or compulsory
+          licensing scheme. In all other cases the Licensor expressly
+          reserves any right to collect such royalties.
+
+
+Section 3 -- License Conditions.
+
+Your exercise of the Licensed Rights is expressly made subject to the
+following conditions.
+
+  a. Attribution.
+
+       1. If You Share the Licensed Material (including in modified
+          form), You must:
+
+            a. retain the following if it is supplied by the Licensor
+               with the Licensed Material:
+
+                 i. identification of the creator(s) of the Licensed
+                    Material and any others designated to receive
+                    attribution, in any reasonable manner requested by
+                    the Licensor (including by pseudonym if
+                    designated);
+
+                ii. a copyright notice;
+
+               iii. a notice that refers to this Public License;
+
+                iv. a notice that refers to the disclaimer of
+                    warranties;
+
+                 v. a URI or hyperlink to the Licensed Material to the
+                    extent reasonably practicable;
+
+            b. indicate if You modified the Licensed Material and
+               retain an indication of any previous modifications; and
+
+            c. indicate the Licensed Material is licensed under this
+               Public License, and include the text of, or the URI or
+               hyperlink to, this Public License.
+
+       2. You may satisfy the conditions in Section 3(a)(1) in any
+          reasonable manner based on the medium, means, and context in
+          which You Share the Licensed Material. For example, it may be
+          reasonable to satisfy the conditions by providing a URI or
+          hyperlink to a resource that includes the required
+          information.
+
+       3. If requested by the Licensor, You must remove any of the
+          information required by Section 3(a)(1)(A) to the extent
+          reasonably practicable.
+
+       4. If You Share Adapted Material You produce, the Adapter's
+          License You apply must not prevent recipients of the Adapted
+          Material from complying with this Public License.
+
+
+Section 4 -- Sui Generis Database Rights.
+
+Where the Licensed Rights include Sui Generis Database Rights that
+apply to Your use of the Licensed Material:
+
+  a. for the avoidance of doubt, Section 2(a)(1) grants You the right
+     to extract, reuse, reproduce, and Share all or a substantial
+     portion of the contents of the database;
+
+  b. if You include all or a substantial portion of the database
+     contents in a database in which You have Sui Generis Database
+     Rights, then the database in which You have Sui Generis Database
+     Rights (but not its individual contents) is Adapted Material; and
+
+  c. You must comply with the conditions in Section 3(a) if You Share
+     all or a substantial portion of the contents of the database.
+
+For the avoidance of doubt, this Section 4 supplements and does not
+replace Your obligations under this Public License where the Licensed
+Rights include other Copyright and Similar Rights.
+
+
+Section 5 -- Disclaimer of Warranties and Limitation of Liability.
+
+  a. UNLESS OTHERWISE SEPARATELY UNDERTAKEN BY THE LICENSOR, TO THE
+     EXTENT POSSIBLE, THE LICENSOR OFFERS THE LICENSED MATERIAL AS-IS
+     AND AS-AVAILABLE, AND MAKES NO REPRESENTATIONS OR WARRANTIES OF
+     ANY KIND CONCERNING THE LICENSED MATERIAL, WHETHER EXPRESS,
+     IMPLIED, STATUTORY, OR OTHER. THIS INCLUDES, WITHOUT LIMITATION,
+     WARRANTIES OF TITLE, MERCHANTABILITY, FITNESS FOR A PARTICULAR
+     PURPOSE, NON-INFRINGEMENT, ABSENCE OF LATENT OR OTHER DEFECTS,
+     ACCURACY, OR THE PRESENCE OR ABSENCE OF ERRORS, WHETHER OR NOT
+     KNOWN OR DISCOVERABLE. WHERE DISCLAIMERS OF WARRANTIES ARE NOT
+     ALLOWED IN FULL OR IN PART, THIS DISCLAIMER MAY NOT APPLY TO YOU.
+
+  b. TO THE EXTENT POSSIBLE, IN NO EVENT WILL THE LICENSOR BE LIABLE
+     TO YOU ON ANY LEGAL THEORY (INCLUDING, WITHOUT LIMITATION,
+     NEGLIGENCE) OR OTHERWISE FOR ANY DIRECT, SPECIAL, INDIRECT,
+     INCIDENTAL, CONSEQUENTIAL, PUNITIVE, EXEMPLARY, OR OTHER LOSSES,
+     COSTS, EXPENSES, OR DAMAGES ARISING OUT OF THIS PUBLIC LICENSE OR
+     USE OF THE LICENSED MATERIAL, EVEN IF THE LICENSOR HAS BEEN
+     ADVISED OF THE POSSIBILITY OF SUCH LOSSES, COSTS, EXPENSES, OR
+     DAMAGES. WHERE A LIMITATION OF LIABILITY IS NOT ALLOWED IN FULL OR
+     IN PART, THIS LIMITATION MAY NOT APPLY TO YOU.
+
+  c. The disclaimer of warranties and limitation of liability provided
+     above shall be interpreted in a manner that, to the extent
+     possible, most closely approximates an absolute disclaimer and
+     waiver of all liability.
+
+
+Section 6 -- Term and Termination.
+
+  a. This Public License applies for the term of the Copyright and
+     Similar Rights licensed here. However, if You fail to comply with
+     this Public License, then Your rights under this Public License
+     terminate automatically.
+
+  b. Where Your right to use the Licensed Material has terminated under
+     Section 6(a), it reinstates:
+
+       1. automatically as of the date the violation is cured, provided
+          it is cured within 30 days of Your discovery of the
+          violation; or
+
+       2. upon express reinstatement by the Licensor.
+
+     For the avoidance of doubt, this Section 6(b) does not affect any
+     right the Licensor may have to seek remedies for Your violations
+     of this Public License.
+
+  c. For the avoidance of doubt, the Licensor may also offer the
+     Licensed Material under separate terms or conditions or stop
+     distributing the Licensed Material at any time; however, doing so
+     will not terminate this Public License.
+
+  d. Sections 1, 5, 6, 7, and 8 survive termination of this Public
+     License.
+
+
+Section 7 -- Other Terms and Conditions.
+
+  a. The Licensor shall not be bound by any additional or different
+     terms or conditions communicated by You unless expressly agreed.
+
+  b. Any arrangements, understandings, or agreements regarding the
+     Licensed Material not stated herein are separate from and
+     independent of the terms and conditions of this Public License.
+
+
+Section 8 -- Interpretation.
+
+  a. For the avoidance of doubt, this Public License does not, and
+     shall not be interpreted to, reduce, limit, restrict, or impose
+     conditions on any use of the Licensed Material that could lawfully
+     be made without permission under this Public License.
+
+  b. To the extent possible, if any provision of this Public License is
+     deemed unenforceable, it shall be automatically reformed to the
+     minimum extent necessary to make it enforceable. If the provision
+     cannot be reformed, it shall be severed from this Public License
+     without affecting the enforceability of the remaining terms and
+     conditions.
+
+  c. No term or condition of this Public License will be waived and no
+     failure to comply consented to unless expressly agreed to by the
+     Licensor.
+
+  d. Nothing in this Public License constitutes or may be interpreted
+     as a limitation upon, or waiver of, any privileges and immunities
+     that apply to the Licensor or You, including from the legal
+     processes of any jurisdiction or authority.
+
+
+=======================================================================
+
+Creative Commons is not a party to its public
+licenses. Notwithstanding, Creative Commons may elect to apply one of
+its public licenses to material it publishes and in those instances
+will be considered the “Licensor.” The text of the Creative Commons
+public licenses is dedicated to the public domain under the CC0 Public
+Domain Dedication. Except for the limited purpose of indicating that
+material is shared under a Creative Commons public license or as
+otherwise permitted by the Creative Commons policies published at
+creativecommons.org/policies, Creative Commons does not authorize the
+use of the trademark "Creative Commons" or any other trademark or logo
+of Creative Commons without its prior written consent including,
+without limitation, in connection with any unauthorized modifications
+to any of its public licenses or any other arrangements,
+understandings, or agreements concerning use of licensed material. For
+the avoidance of doubt, this paragraph does not form part of the
+public licenses.
+
+Creative Commons may be contacted at creativecommons.org.
+
+

docs/arm64.md

@@ -1,10 +1,27 @@
 # How to build and test Cloud Hypervisor on AArch64
 
-This document introduces how to build and test Cloud Hypervisor on AArch64 servers. Currently Cloud Hypervisor cannot be tested on Raspberry PI. Because on AArch64, Cloud Hypervisor requires GICv3-ITS device for PCIe MSI interrupt handling. But GICv3-ITS has not been equipped on any Raspberry PI product so far.
+This document introduces how to build and test Cloud Hypervisor on AArch64.
+Currently, Cloud Hypervisor supports 2 methods of booting on AArch64: UEFI
+booting and direct-kernel booting. The document covers both methods.
 
-Now Cloud Hypervisor supports 2 ways of booting on AArch64: UEFI booting and direct-kernel booting. The document covers both of the ways.
+All the steps are based on Ubuntu. We use the Ubuntu cloud image for guest VM
+disk.
 
-All the steps are based on Ubuntu. We use the Ubuntu cloud image for guest VM disk.
+## Hardware requirements
+
+- AArch64 servers (recommended) or development boards equipped with the GICv3
+interrupt controller.
+
+- On development boards that have constrained RAM resources, if the creation of
+a VM consumes a large portion of the free memory on the host, it may be required
+to enable swap. For example, this was required on a board with 3 GB of RAM
+booting a 2 GB VM at a point in time when 2.8 GB were free. Without enabling
+swap the `cloud-hypervisor` process was terminated by the OOM killer. In this
+situation memory was allocated for the virtual machine using memfd while the
+page cache was filled, leading to a situation where the kernel could not even
+drop caches. Making a small section of swap available (observably, 1 to 15 MB),
+this situation can be resolved and the resulting memory footprint of
+`cloud-hypervisor` is as expected.
 
 ## Getting started
 
@@ -82,19 +99,29 @@ $ build -a AARCH64 -t GCC5 -p ArmVirtPkg/ArmVirtCloudHv.dsc -b RELEASE
 $ popd
 ```
 
-If the build goes well, the EDK2 binary is available at `edk2/Build/ArmVirtCloudHv-AARCH64/RELEASE_GCC5/FV/CLOUDHV_EFI.fd`.
+If the build goes well, the EDK2 binary is available at
+`edk2/Build/ArmVirtCloudHv-AARCH64/RELEASE_GCC5/FV/CLOUDHV_EFI.fd`.
 
 ### Booting the guest VM
 
 ```bash
 $ pushd $CLOUDH
-$ sudo RUST_BACKTRACE=1 $CLOUDH/cloud-hypervisor/target/debug/cloud-hypervisor --api-socket /tmp/cloud-hypervisor.sock --kernel $CLOUDH/edk2/edk2/Build/ArmVirtCloudHv-AARCH64/RELEASE_GCC5/FV/CLOUDHV_EFI.fd --disk path=$CLOUDH/focal-server-cloudimg-arm64.raw --cpus boot=4 --memory size=4096M --serial tty --console off --log-file log.log -vvv --net tap=,mac=12:34:56:78:90:01,ip=192.168.1.1,mask=255.255.255.0
+$ sudo RUST_BACKTRACE=1 $CLOUDH/cloud-hypervisor/target/debug/cloud-hypervisor \
+           --api-socket /tmp/cloud-hypervisor.sock \
+           --kernel $CLOUDH/edk2/Build/ArmVirtCloudHv-AARCH64/RELEASE_GCC5/FV/CLOUDHV_EFI.fd \
+           --disk path=$CLOUDH/focal-server-cloudimg-arm64.raw \
+           --cpus boot=4 \
+           --memory size=4096M \
+           --net tap=,mac=12:34:56:78:90:01,ip=192.168.1.1,mask=255.255.255.0 \
+           --serial tty \
+           --console off
 $ popd
 ```
 
 ## Direct-kernel booting
 
-Alternativelly, you can build your own kernel for guest VM. This way, UEFI is not involved and ACPI cannot be enabled.
+Alternativelly, you can build your own kernel for guest VM. This way, UEFI is
+not involved and ACPI cannot be enabled.
 
 ### Building kernel
 
@@ -111,6 +138,15 @@ $ popd
 
 ```bash
 $ pushd $CLOUDH
-$ sudo $CLOUDH/cloud-hypervisor/target/debug/cloud-hypervisor --api-socket /tmp/cloud-hypervisor.sock --kernel $CLOUDH/linux/arch/arm64/boot/Image --disk path=focal-server-cloudimg-arm64.raw --cmdline "keep_bootcon console=ttyAMA0 reboot=k panic=1 root=/dev/vda1 rw" --cpus boot=4 --memory size=4096M --serial tty --console off --log-file log.log -vvv --net "tap=,mac=,ip=,mask="
+$ sudo $CLOUDH/cloud-hypervisor/target/debug/cloud-hypervisor \
+           --api-socket /tmp/cloud-hypervisor.sock \
+           --kernel $CLOUDH/linux/arch/arm64/boot/Image \
+           --disk path=focal-server-cloudimg-arm64.raw \
+           --cmdline "keep_bootcon console=ttyAMA0 reboot=k panic=1 root=/dev/vda1 rw" \
+           --cpus boot=4 \
+           --memory size=4096M \
+           --net tap=,mac=12:34:56:78:90:01,ip=192.168.1.1,mask=255.255.255.0 \
+           --serial tty \
+           --console off
 $ popd
 ```

docs/cpu.md

@@ -0,0 +1,189 @@
+# CPU
+
+Cloud Hypervisor has many options when it comes to the creation of virtual
+CPUs. This document aims to explain what Cloud Hypervisor is capable of and
+how it can be used to meet the needs of very different use cases.
+
+## Options
+
+`CpusConfig` or what is known as `--cpus` from the CLI perspective is the way
+to set vCPUs options for Cloud Hypervisor.
+
+```rust
+struct CpusConfig {
+    boot_vcpus: u8,
+    max_vcpus: u8,
+    topology: Option<CpuTopology>,
+    kvm_hyperv: bool,
+    max_phys_bits: u8,
+    affinity: Option<Vec<CpuAffinity>>,
+}
+```
+
+```
+--cpus boot=<boot_vcpus>,max=<max_vcpus>,topology=<threads_per_core>:<cores_per_die>:<dies_per_package>:<packages>,kvm_hyperv=on|off,max_phys_bits=<maximum_number_of_physical_bits>,affinity=<list_of_vcpus_with_their_associated_cpuset>
+```
+
+### `boot`
+
+Number of vCPUs present at boot time.
+
+This option allows to define a specific number of vCPUs to be present at the
+time the VM is started. This option is mandatory when using the `--cpus`
+parameter. If `--cpus` is not specified, this option takes the default value
+of `1`, starting the VM with a single vCPU.
+
+Value is an unsigned integer of 8 bits.
+
+_Example_
+
+```
+--cpus boot=2
+```
+
+### `max`
+
+Maximum number of vCPUs.
+
+This option defines the maximum number of vCPUs that can be assigned to the VM.
+In particular, this option is used when looking for CPU hotplug as it lets the
+provide an indication about how many vCPUs might be needed later during the
+runtime of the VM.
+For instance, if booting the VM with 2 vCPUs and a maximum of 6 vCPUs, it means
+up to 4 vCPUs can be added later at runtime by resizing the VM.
+
+The value must be greater than or equal to the number of boot vCPUs.
+The value is an unsigned integer of 8 bits.
+
+By default this option takes the value of `boot`, meaning vCPU hotplug is not
+expected and can't be performed.
+
+_Example_
+
+```
+--cpus max=3
+```
+
+### `topology`
+
+Topology of the guest platform.
+
+This option gives the user a way to describe the exact topology that should be
+exposed to the guest. It can be useful to describe to the guest the same
+topology found on the host as it allows for proper usage of the resources and
+is a way to achieve better performances.
+
+The topology is described through the following structure:
+
+```rust
+struct CpuTopology {
+    threads_per_core: u8,
+    cores_per_die: u8,
+    dies_per_package: u8,
+    packages: u8,
+}
+```
+
+or the following syntax through the CLI:
+
+```
+topology=<threads_per_core>:<cores_per_die>:<dies_per_package>:<packages>
+```
+
+By default the topology will be `1:1:1:1`.
+
+_Example_
+
+```
+--cpus boot=2,topology=1:1:2:1
+```
+
+### `kvm_hyperv`
+
+Enable KVM Hyper-V emulation.
+
+When turned on, this option relies on KVM to emulate the synthetic interrupt
+controller (SynIC) along with synthetic timers expected by a Windows guest.
+A Windows guest usually runs on top of Microsoft Hyper-V, therefore expects
+these synthetic devices to be present. That's why KVM provides a way to emulate
+them and avoids failures running a Windows guest with Cloud Hypervisor.
+
+By default this option is turned off.
+
+_Example_
+
+```
+--cpus kvm_hyperv=on
+```
+
+### `max_phys_bits`
+
+Maximum size for guest's addressable space.
+
+This option defines the maximum number of physical bits for all vCPUs, which
+sets a limit for the size of the guest's addressable space. This is mainly
+useful for debug purpose.
+
+The value is an unsigned integer of 8 bits.
+
+_Example_
+
+```
+--cpus max_phys_bits=40
+```
+
+### `affinity`
+
+Affinity of each vCPU.
+
+This option gives the user a way to provide the host CPU set associated with
+each vCPU. It is useful for achieving CPU pinning, ensuring multiple VMs won't
+affect the performance of each other. It might also be used in the context of
+NUMA as it is way of making sure the VM can run on a specific host NUMA node.
+In general, this option is used to increase the performances of a VM depending
+on the host platform and the type of workload running in the guest.
+
+The affinity is described through the following structure:
+
+```rust
+struct CpuAffinity {
+    vcpu: u8,
+    host_cpus: Vec<u8>,
+}
+```
+
+or the following syntax through the CLI:
+
+```
+affinity=[<vcpu_id1>@[<host_cpu_id1>, <host_cpu_id2>], <vcpu_id2>@[<host_cpu_id3>, <host_cpu_id4>]]
+```
+
+The outer brackets define the list of vCPUs. And for each vCPU, the inner
+brackets attached to `@` define the list of host CPUs the vCPU is allowed to
+run onto.
+
+Multiple values can be provided to define each list. Each value is an unsigned
+integer of 8 bits.
+
+For instance, if one needs to run vCPU 0 on host CPUs from 0 to 4, the syntax
+using `-` will help define a contiguous range with `affinity=0@[0-4]`. The
+same example could also be described with `affinity=0@[0,1,2,3,4]`.
+
+A combination of both `-` and `,` separators is useful when one might need to
+describe a list containing host CPUs from 0 to 99 and the host CPU 255, as it
+could simply be described with `affinity=0@[0-99,255]`.
+
+As soon as one tries to describe a list of values, `[` and `]` must be used to
+demarcate the list.
+
+By default each vCPU runs on the entire host CPU set.
+
+_Example_
+
+```
+--cpus boot=3,affinity=[0@[2,3],1@[0,1]]
+```
+
+In this example, assuming the host has 4 CPUs, vCPU 0 will run exclusively on
+host CPUs 2 and 3, while vCPU 1 will run exclusively on host CPUs 0 and 1.
+Because nothing is defined for vCPU 2, it can run on any of the 4 host CPUs.

docs/custom-image.md

@@ -72,7 +72,7 @@ mount -t devpts devpts /dev/pts
 
 ### Install needed packages
 
-In the context Cloud-Hypervisor's integration tests, we need several utilities.
+In the context Cloud Hypervisor's integration tests, we need several utilities.
 Here is the way to install them for a Ubuntu image. This step is specific to
 Ubuntu distributions.
 

docs/device_model.md

@@ -89,7 +89,9 @@ feature is enabled by default.
 ## Virtio devices
 
 For all virtio devices listed below, only `virtio-pci` transport layer is
-supported.
+supported. Cloud Hypervisor supports multiple PCI segments, and users can
+append `,pci_segment=<PCI_segment_number>` to the device flag in the Cloud
+Hypervisor command line to assign devices to a specific PCI segment.
 
 ### virtio-block
 

docs/fs.md

@@ -12,12 +12,12 @@ This virtual device relies on the _vhost-user_ protocol, which assumes the backe
 
 _Build virtiofsd_
 ```bash
-git clone --depth 1 "https://gitlab.com/virtio-fs/qemu.git" -b "qemu5.0-virtiofs-dax" $VIRTIOFSD_DIR
+git clone https://gitlab.com/virtio-fs/virtiofsd
-cd $VIRTIOFSD_DIR
+pushd virtiofsd
-./configure --prefix=$PWD --target-list=x86_64-softmmu
+cargo build --release
-make virtiofsd -j `nproc`
+sudo setcap cap_sys_admin+epi target/release/virtiofsd
-sudo setcap cap_sys_admin+epi "virtiofsd"
 ```
+
 _Create shared directory_
 ```bash
 mkdir /tmp/shared_dir
@@ -27,11 +27,11 @@ _Run virtiofsd_
 ./virtiofsd \
     -d \
     --socket-path=/tmp/virtiofs \
-    -o source=/tmp/shared_dir \
+    --shared-dir=/tmp/shared_dir \
-    -o cache=none
+    --cache=never
 ```
 
-The `cache=none` option should be the default when using `virtiofsd` with the __cloud-hypervisor__ VMM. This prevents from using the guest page cache, which reduces the memory footprint of the guest. When running multiple virtual machines on the same host, this will let the host deal with page cache, which will increase the density of virtual machines which can be launched.
+The `cache=never` option should be the default when using `virtiofsd` with the __cloud-hypervisor__ VMM. This prevents from using the guest page cache, which reduces the memory footprint of the guest. When running multiple virtual machines on the same host, this will let the host deal with page cache, which will increase the density of virtual machines which can be launched.
 
 The `cache=always` option will allow for the guest page cache to be used, which will increase the memory footprint of the guest. This option should be used only for specific use cases where a single VM is going to be running on a host.
 

docs/gdb.md

@@ -0,0 +1,47 @@
+# GDB Support
+
+This feature allows remote guest debugging using GDB. Note that this feature is only supported on x86_64/KVM.
+
+To enable debugging with GDB, build with the `gdb` feature enabled:
+
+```bash
+cargo build --features gdb
+```
+
+To use the `--gdb` option, specify the Unix Domain Socket with `--path` that Cloud Hypervisor will use to communicate with the host's GDB:
+
+```bash
+./cloud-hypervisor \
+    --kernel hypervisor-fw \
+    --disk path=bionic-server-cloudimg-amd64.raw \
+    --cpus boot=1 \
+    --memory size=1024M \
+    --net "tap=,mac=,ip=,mask=" \
+    --console off \
+    --serial tty \
+    --gdb path=/tmp/ch-gdb-sock
+```
+
+Cloud Hypervisor will listen for GDB on the host side before starting the guest.
+On the host side, connect to the GDB remote server as follows:
+
+```bash
+gdb -q
+(gdb) target remote /tmp/ch-gdb-sock
+Remote debugging using /tmp/ch-gdb-sock
+warning: No executable has been specified, and target does not support
+determining executable automatically. Try using the "file" command.
+0x000000000011217e in ?? ()
+```
+
+You can set up to four hardware breakpoints using the x86 debug register:
+
+```bash
+(gdb) hb *0x1121b7
+Hardware assisted breakpoint 1 at 0x1121b7
+(gdb) c
+Continuing.
+
+Breakpoint 1, 0x00000000001121b7 in ?? ()
+(gdb)
+```

docs/hotplug.md

@@ -1,6 +1,6 @@
 # Cloud Hypervisor Hot Plug
 
-Currently Cloud Hypervisor only support hot plugging of CPU devices.
+Currently Cloud Hypervisor supports hot plugging of CPUs devices (x86 only), PCI devices and memory resizing.
 
 ## Kernel support
 
@@ -9,7 +9,7 @@ or by using this kernel patch (available in 5.5-rc1 and later): https://git.kern
 
 ## CPU Hot Plug
 
-Extra vCPUs can be added and removed from a running Cloud Hypervisor instance. This is controlled by two mechanisms:
+Extra vCPUs can be added and removed from a running `cloud-hypervisor` instance. This is controlled by two mechanisms:
 
 1. Specifying a number of maximum potential vCPUs that is greater than the number of default (boot) vCPUs.
 2. Making a HTTP API request to the VMM to ask for the additional vCPUs to be added.
@@ -65,7 +65,7 @@ As per adding CPUs to the guest, after a reboot the VM will be running with the
 
 ### ACPI method
 
-Extra memory can be added from a running Cloud Hypervisor instance. This is controlled by two mechanisms:
+Extra memory can be added from a running `cloud-hypervisor` instance. This is controlled by two mechanisms:
 
 1. Allocating some of the guest physical address space for hotplug memory.
 2. Making a HTTP API request to the VMM to ask for a new amount of RAM to be assigned to the VM. In the case of expanding the memory for the VM the new memory will be hotplugged into the running VM, if reducing the size of the memory then change will take effect after the next reboot.
@@ -157,7 +157,7 @@ The same API can also be used to reduce the desired RAM for a VM. It is importan
 
 ## PCI Device Hot Plug
 
-Extra PCI devices can be added and removed from a running Cloud Hypervisor instance. This is controlled by making a HTTP API request to the VMM to ask for the additional device to be added, or for the existing device to be removed.
+Extra PCI devices can be added and removed from a running `cloud-hypervisor` instance. This is controlled by making a HTTP API request to the VMM to ask for the additional device to be added, or for the existing device to be removed.
 
 Note: On AArch64 platform, PCI device hotplug can only be achieved using ACPI. Please refer to the [documentation](arm64.md#uefi-booting) for more information.
 

docs/intel_sgx.md

@@ -1,7 +1,7 @@
 # Intel SGX
 
 Intel® Software Guard Extensions (Intel® SGX) is an Intel technology designed
-to increase the security of application code and data. Cloud-Hypervisor supports
+to increase the security of application code and data. Cloud Hypervisor supports
 SGX virtualization through KVM. Because SGX is built on hardware features that
 cannot be emulated in software, virtualizing SGX requires support in KVM and in
 the host kernel. The required Linux and KVM changes can be found in the
@@ -17,12 +17,12 @@ For more information about SGX, please refer to the [SGX Homepage](https://softw
 For more information about SGX SDK and how to test SGX, please refer to the
 following [instructions](https://github.com/intel/linux-sgx).
 
-## Cloud-Hypervisor support
+## Cloud Hypervisor support
 
 Assuming the host exposes `/dev/sgx_vepc`, we can pass SGX enclaves through
 the guest.
 
-In order to use SGX enclaves within a Cloud-Hypervisor VM, we must define one
+In order to use SGX enclaves within a Cloud Hypervisor VM, we must define one
 or several Enclave Page Cache (EPC) sections. Here is an example of a VM being
 created with 2 EPC sections, the first one being 64MiB with pre-allocated
 memory, the second one being 32MiB with no pre-allocated memory.

docs/intel_tdx.md

@@ -0,0 +1,104 @@
+# Intel TDX
+
+Intel® Trust Domain Extensions (Intel® TDX) is an Intel technology designed to
+isolate virtual machines from the VMM, hypervisor and any other software on the
+host platform.
+
+For more information about TDX technical aspects, design and specification
+please refer to the
+[TDX Homepage](https://www.intel.com/content/www/us/en/developer/articles/technical/intel-trust-domain-extensions.html).
+
+The required Linux changes for the host side can be found in the
+[KVM TDX tree](https://github.com/intel/tdx/tree/kvm) while the changes for
+the guest side can be found in the [Guest TDX tree](https://github.com/intel/tdx/tree/guest).
+
+The TDVF firmware can be found in the
+[EDK2 staging project](https://github.com/tianocore/edk2-staging/tree/TDVF).
+
+The TDShim firmware can be found in the
+[Confidential Containers project](https://github.com/confidential-containers/td-shim).
+
+## Cloud Hypervisor support
+
+First, you must be running on a machine with TDX enabled in hardware, and
+with the host OS compiled from the [KVM TDX tree](https://github.com/intel/tdx/tree/kvm).
+
+Cloud Hypervisor can run TDX VM (Trust Domain) by loading a TD firmware,
+which will then load the guest kernel from the image. The image must be custom
+as it must include a kernel built from the [Guest TDX tree](https://github.com/intel/tdx/tree/guest).
+
+### TDVF
+
+The firmware can be built as follows:
+
+```bash
+git clone https://github.com/tianocore/edk2-staging.git
+cd edk2-staging
+git checkout origin/TDVF
+git submodule update --init --recursive
+make -C BaseTools
+source ./edksetup.sh
+build -p OvmfPkg/OvmfCh.dsc -a X64 -t GCC5 -b RELEASE
+```
+
+If debug logs are needed, here is the alternative command:
+
+```bash
+build -p OvmfPkg/OvmfCh.dsc -a X64 -t GCC5 -D DEBUG_ON_SERIAL_PORT=TRUE
+```
+
+On the Cloud Hypervisor side, all you need is to build the project with the
+`tdx` feature enabled:
+
+```bash
+cargo build --features tdx
+```
+
+And run a TDX VM by providing the firmware previously built, along with the
+guest image containing the TDX enlightened kernel. Assuming the guest kernel
+command line contains `console=hvc0` (printing to the `virtio-console` device),
+run Cloud Hypervisor as follows:
+
+```bash
+./cloud-hypervisor \
+    --tdx firmware=edk2-staging/Build/OvmfCh/RELEASE_GCC5/FV/OVMF.fd \
+    --cpus boot=1 \
+    --memory size=1G \
+    --disk path=tdx_guest_img
+```
+
+And here is the alternative command when looking for debug logs (assuming the
+guest kernel command line contains `console=ttyS0`):
+
+```bash
+./cloud-hypervisor \
+    --tdx firmware=edk2-staging/Build/OvmfCh/DEBUG_GCC5/FV/OVMF.fd \
+    --cpus boot=1 \
+    --memory size=1G \
+    --disk path=tdx_guest_img \
+    --serial tty \
+    --console off
+```
+
+### TDShim
+
+This is a lightweight version of the TDVF, written in Rust and designed for
+direct kernel boot, which is useful for containers use cases.
+
+You can find the instructions for building the firmware directly from the
+project [documentation](https://github.com/confidential-containers/td-shim/tree/staging#how-to-build).
+
+And run a TDX VM by providing the firmware previously built, along with a guest
+kernel built from the [Guest TDX tree](https://github.com/intel/tdx/tree/guest).
+The appropriate kernel boot options must be provided through the `--cmdline`
+option as well.
+
+```bash
+./cloud-hypervisor \
+    --tdx firmware=tdshim \
+    --kernel bzImage \
+    --cmdline "root=/dev/vda1 console=hvc0 rw tdx_allow_acpi=MCFG"
+    --cpus boot=1 \
+    --memory size=1G \
+    --disk path=tdx_guest_img
+```

docs/live_migration.md

@@ -8,13 +8,13 @@ support in Cloud Hypervisor:
 1. nested-vm migration - migrating between two nested VMs whose host VMs
    are running on the same machine.
 
-## Local Migration
+## Local Migration (Suitable for Live Upgrade of VMM)
 Launch the source VM (on the host machine):
 ```bash
 $ target/release/cloud-hypervisor
     --kernel ~/workloads/vmlinux \
     --disk path=~/workloads/focal.raw \
-    --cpus boot=1 --memory size=1G \
+    --cpus boot=1 --memory size=1G,shared=on \
     --cmdline "root=/dev/vda1 console=ttyS0"  \
     --serial tty --console off --api-socket=/tmp/api1
 ```
@@ -31,7 +31,7 @@ $ target/release/ch-remote --api-socket=/tmp/api2 receive-migration unix:/tmp/so
 
 Start to send migration for the source VM (on the host machine):
 ```bash
-$ target/release/ch-remote --api-socket=/tmp/api1 send-migration unix:/tmp/sock
+$ target/release/ch-remote --api-socket=/tmp/api1 send-migration --local unix:/tmp/sock
 ```
 
 When the above commands completed, the source VM should be successfully

docs/memory.md

@@ -1,13 +1,13 @@
 # Memory
 
-Cloud-Hypervisor has many ways to expose memory to the guest VM. This document
+Cloud Hypervisor has many ways to expose memory to the guest VM. This document
-aims to explain what Cloud-Hypervisor is capable of and how it can be used to
+aims to explain what Cloud Hypervisor is capable of and how it can be used to
 meet the needs of very different use cases.
 
 ## Basic Parameters
 
 `MemoryConfig` or what is known as `--memory` from the CLI perspective is the
-easiest way to get started with Cloud-Hypervisor.
+easiest way to get started with Cloud Hypervisor.
 
 ```rust
 struct MemoryConfig {
@@ -446,17 +446,20 @@ integer of 8 bits.
 
 For instance, if one needs to attach all CPUs from 0 to 4 to a specific node,
 the syntax using `-` will help define a contiguous range with `cpus=0-4`. The
-same example could also be described with `cpus=0:1:2:3:4`.
+same example could also be described with `cpus=[0,1,2,3,4]`.
 
-A combination of both `-` and `:` separators is useful when one might need to
+A combination of both `-` and `,` separators is useful when one might need to
 describe a list containing all CPUs from 0 to 99 and the CPU 255, as it could
-simply be described with `cpus=0-99:255`.
+simply be described with `cpus=[0-99,255]`.
+
+As soon as one tries to describe a list of values, `[` and `]` must be used to
+demarcate the list.
 
 _Example_
 
 ```
 --cpus boot=8
---numa guest_numa_id=0,cpus=1-3:7 guest_numa_id=1,cpus=0:4-6
+--numa guest_numa_id=0,cpus=[1-3,7] guest_numa_id=1,cpus=[0,4-6]
 ```
 
 ### `distances`
@@ -473,7 +476,10 @@ node. The second value is an unsigned integer of 8 bits as it represents the
 distance between the current NUMA node and the destination NUMA node. The two
 values are separated by `@` (`value1@value2`), meaning the destination NUMA
 node `value1` is located at a distance of `value2`. Each tuple is separated
-from the others with `:` separator.
+from the others with `,` separator.
+
+As soon as one tries to describe a list of values, `[` and `]` must be used to
+demarcate the list.
 
 For instance, if one wants to define 3 NUMA nodes, with each node located at
 different distances, it can be described with the following example.
@@ -481,7 +487,7 @@ different distances, it can be described with the following example.
 _Example_
 
 ```
---numa guest_numa_id=0,distances=1@15:2@25 guest_numa_id=1,distances=0@15:2@20 guest_numa_id=2,distances=0@25:1@20
+--numa guest_numa_id=0,distances=[1@15,2@25] guest_numa_id=1,distances=[0@15,2@20] guest_numa_id=2,distances=[0@25,1@20]
 ```
 
 ### `memory_zones`
@@ -498,7 +504,10 @@ workload run more efficiently.
 
 Multiple values can be provided to define the list. Each value is a string
 referring to an existing memory zone identifier. Values are separated from
-each other with the `:` separator.
+each other with the `,` separator.
+
+As soon as one tries to describe a list of values, `[` and `]` must be used to
+demarcate the list.
 
 Note that a memory zone must belong to a single NUMA node. The following
 configuration is incorrect, therefore not allowed:
@@ -509,7 +518,7 @@ _Example_
 ```
 --memory size=0
 --memory-zone id=mem0,size=1G id=mem1,size=1G id=mem2,size=1G
---numa guest_numa_id=0,memory_zones=mem0:mem2 guest_numa_id=1,memory_zones=mem1
+--numa guest_numa_id=0,memory_zones=[mem0,mem2] guest_numa_id=1,memory_zones=mem1
 ```
 
 ### `sgx_epc_sections`
@@ -520,13 +529,16 @@ which must be seen by the guest as belonging to the NUMA node `guest_numa_id`.
 
 Multiple values can be provided to define the list. Each value is a string
 referring to an existing SGX EPC section identifier. Values are separated from
-each other with the `:` separator.
+each other with the `,` separator.
+
+As soon as one tries to describe a list of values, `[` and `]` must be used to
+demarcate the list.
 
 _Example_
 
 ```
 --sgx-epc id=epc0,size=32M id=epc1,size=64M id=epc2,size=32M
---numa guest_numa_id=0,sgx_epc_sections=epc1 guest_numa_id=1,sgx_epc_sections=epc0:epc2
+--numa guest_numa_id=0,sgx_epc_sections=epc1 guest_numa_id=1,sgx_epc_sections=[epc0,epc2]
 ```
 
 ### PCI bus

docs/networking.md

@@ -1,209 +0,0 @@
-# How to use networking
-
-cloud-hypervisor can emulate one or more virtual network interfaces, represented at the hypervisor host by [tap devices](https://www.kernel.org/doc/Documentation/networking/tuntap.txt). This guide briefly describes, in a manual and distribution neutral way, how to setup and use networking with cloud-hypervisor.
-
-## Multiple queue support for net devices
-
-While multiple vcpus defined for guest, to gain the benefit of vcpu scalable to improve performance, it suggests to define multiple queue pairs for net devices, one Tx/Rx queue pair per one vcpu, that means the number of queue pairs at least is equal to the vcpu count. In that case, after virtnet driver set cpu affinity for virtqueues in guest kernel, vcpus could handle interrupt from different virtqueue pairs in parallel.
-
-It will gain better performance for guest that has multiple queues defined for net devices while it has multiple net sessions running in userspace.
-
-To enable multiple queue support in cloud-hypervisor, multiple queue pairs will be defined, while multiple tap fds will be opened for the same tap device, it will also have multiple threads started, each thread will monitor and handle the events from each virtqueue pairs and the associated tap fd.
-
-Note:
-
-- Currently, it does not support to use ethtool to change the combined queue numbers in guest.
-- Multiple queue is enabled for vhost-user-net backend in cloud-hypervisor, however, multiple thread is not added to handle mq, thus, the performance for vhost-user-net backend is not supposed to be improved. The multiple thread will be added for backend later.
-- Performance test for vhost-user-net will be covered once vhost-user-net backend has multiple thread supported.
-- Performance test for virtio-net is done by comparing 2 queue pairs with 1 queue pairs, that to run 2 iperf3 sessions in the same test environments, throughput is improved about 37%.
-
-## Start cloud-hypervisor with net devices
-
-Use one `--net` command-line argument from cloud-hypervisor to specify the emulation of one or more virtual NIC's. The example below instructs cloud-hypervisor to emulate for instance 2 virtual NIC's:
-
-```bash
-./cloud-hypervisor \
-    --cpus boot=4 \
-    --memory "size=512M" \
-    --disk path=focal-server-cloudimg-amd64.raw \
-    --kernel my-vmlinux.bin \
-    --cmdline "console=ttyS0 console=hvc0 root=/dev/vda1 rw" \
-    --net tap=ich0,mac=a4:a1:c2:00:00:01,ip=192.168.4.2,mask=255.255.255.0,num_queues=2,queue_size=256 \
-          tap=ich1,mac=a4:a1:c2:00:00:02,ip=10.0.1.2,mask=255.255.255.0,num_queues=2,queue_size=256
-```
-
-The `--net` argument takes 1 or more space-separated strings of key value pairs containing the following 4 keys or fields:
-
-| Name       | Purpose                | Optional |
-| ---------- | ---------------------- | -------- |
-| tap        | tap device name        | Yes      |
-| mac        | vNIC mac address       | Yes      |
-| ip         | tap IP IP address      | yes      |
-| mask       | tap IP netmask         | Yes      |
-| num_queues | the number of queues   | yes      |
-| queue_size | the size of each queue | Yes      |
-
-num_queues is the total number of tx and rx queues, the default value is 2, and it could be increased by multiples of 2. Additionally, num_queues is suggested to be as 2 times of vcpu count. The default value for queue_size is 256.
-
-If the tap device is pre-created on host before guest boot up. To use multiple queue support for net device in guest, the tap device should be opened like this from host.
-
-```bash
-[root@localhost ~]# ip tuntap add name ich0 mode tap multi_queue
-```
-
-And the `--net` device should specify support for multiple queues. `num_queues` must be a multiple of 2 starting at least from 4 since multiple queues really means multiple queue pairs. We need at least 2 pairs for this configuration to be correct:
-
-```bash
---net tap=ich0,mac=a4:a1:c2:00:00:01,ip=192.168.4.2,mask=255.255.255.0,num_queues=4,queue_size=256
-```
-
-## Configure the tap devices
-
-After starting cloud-hypervisor as shown above, 2 tap devices with state down will become available at the host:
-
-```bash
-root@host:~# ip link show ich0
-78: ich0: <BROADCAST,MULTICAST> mtu 1500 qdisc noop state DOWN mode DEFAULT group default qlen 1000
-    link/ether 72:54:12:ff:ce:6f brd ff:ff:ff:ff:ff:ff
-root@host:~# ip link show ich1
-79: ich1: <BROADCAST,MULTICAST> mtu 1500 qdisc noop state DOWN mode DEFAULT group default qlen 1000
-    link/ether 06:7a:fc:1b:9a:67 brd ff:ff:ff:ff:ff:ff
-```
-
-Set the tap devices to up state:
-
-```bash
-root@host:~# ip link set up ich0
-root@host:~# ip link set up ich1
-
-root@host:~# ip link show ich0
-78: ich0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UNKNOWN mode DEFAULT group default qlen 1000
-    link/ether 72:54:12:ff:ce:6f brd ff:ff:ff:ff:ff:ff
-root@host:~# ip link show ich1
-79: ich1: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UNKNOWN mode DEFAULT group default qlen 1000
-    link/ether 06:7a:fc:1b:9a:67 brd ff:ff:ff:ff:ff:ff
-```
-
-## Connect tap devices
-
-Different networking models can be used to provide external connectivity. In this example we will
-use 2 linux bridges emulating 2 different networks. The integration bridge (ich-int) in this example will also be used
-for external connectivity.
-
-Create the bridges and connect the cloud-hypervisor tap devices to the bridges:
-
-```bash
-root@host:~# brctl addbr ich-int
-root@host:~# brctl addbr ich-dpl
-root@host:~# ip link set up ich-int
-root@host:~# ip link set up ich-dpl
-root@host:~# brctl addif ich-int ich0
-root@host:~# brctl addif ich-dpl ich1
-root@host:~# brctl show
-bridge name     bridge id               STP enabled     interfaces
-ich-dpl         8000.067afc1b9a67       no              ich1
-ich-int         8000.725412ffce6f       no              ich0
-```
-
-This completes the layer 2 wiring: The cloud-hypervisor is now connected to the hypervisor host via the 2 linux bridges.
-
-## IP (Layer 3) provisioning
-
-### Hypervisor host
-
-On the hypervisor host add the network gateway IP address of each network to the 2 linux bridges:
-
-```bash
-root@host:~# ip addr add 192.168.4.1/24 dev ich-int
-root@host:~# ip addr add 10.0.1.1/24 dev ich-dpl
-```
-
-The routing table of the hypervisor host should now also have corresponding routing entries:
-
-```bash
-root@host:~# route -n
-Kernel IP routing table
-Destination     Gateway         Genmask         Flags Metric Ref    Use Iface
-0.0.0.0         192.168.178.1   0.0.0.0         UG    600    0        0 wlan1
-10.0.1.0        0.0.0.0         255.255.255.0   U     0      0        0 ich-dpl
-192.168.4.0     0.0.0.0         255.255.255.0   U     0      0        0 ich-int
-192.168.178.0   0.0.0.0         255.255.255.0   U     600    0        0 wlan1
-```
-
-### Virtual Machine
-
-Within the virtual machine set the vNIC's to up state and provision the corresponding IP addresses on the 2 vNIC's. The steps outlined below use the ip command as an example. Alternative distribution specific procedures can also apply.
-
-```bash
-root@guest:~# ip link set up enp0s2
-root@guest:~# ip link set up enp0s3
-root@guest:~# ip addr add 192.168.4.2/24 dev enp0s2
-root@guest:~# ip addr add 10.0.1.2/24 dev enp0s3
-```
-
-IP connectivity between the virtual machine and the hypervisor-host can be verified by sending
-ICMP requests to the hypervisor-host for the gateway IP address from within the virtual machine:
-
-```bash
-root@guest:~# ping 192.168.4.1
-PING 192.168.4.1 (192.168.4.1) 56(84) bytes of data.
-64 bytes from 192.168.4.1: icmp_seq=1 ttl=64 time=0.456 ms
-64 bytes from 192.168.4.1: icmp_seq=2 ttl=64 time=0.226 ms
-root@guest:~# ping 10.0.1.1
-PING 10.0.1.1 (10.0.1.1) 56(84) bytes of data.
-64 bytes from 10.0.1.1: icmp_seq=1 ttl=64 time=0.449 ms
-64 bytes from 10.0.1.1: icmp_seq=2 ttl=64 time=0.393 ms
-```
-
-The connection can now be used for instance to log into the virtual machine with
-ssh under the precondition that the machine has an ssh daemon provisioned:
-
-```bash
-root@host:~# ssh root@192.168.4.2
-The authenticity of host '192.168.4.2 (192.168.4.2)' can't be established.
-ECDSA key fingerprint is SHA256:qNAUmTtDMW9pNuZARkpLQhfw+Yc1tqUDBrQp7aZGSjw.
-Are you sure you want to continue connecting (yes/no)? yes
-Warning: Permanently added '192.168.4.2' (ECDSA) to the list of known hosts.
-root@192.168.4.2's password:
-Linux cloud-hypervisor 5.2.0 #2 SMP Thu Jul 11 08:08:16 CEST 2019 x86_64
-
-Debian GNU/Linux comes with ABSOLUTELY NO WARRANTY, to the extent
-permitted by applicable law.
-
-
-Last login: Fri Jul 12 13:27:56 2019 from 192.168.4.1
-root@guest:~#
-```
-
-## Internet connectivity
-
-To enable internet connectivity a default gw and a nameserver has to be set within
-the virtual machine:
-
-```bash
-root@guest:~# ip route add default via 192.168.4.1
-root@guest:~# cat /etc/resolv.conf
-options timeout:2
-domain vallis.nl
-search vallis.nl
-nameserver 192.168.178.1
-```
-
-make sure that the default gateway of the hypervisor host (in this example host 192.168.178.1 which is an adsl router) has an entry in the routing table for the 192.168.4.0/24 network otherwise IP connectivity will not work.
-
-```bash
-root@guest:~# nslookup  ftp.nl.debian.org
-Server:         192.168.178.1
-Address:        192.168.178.1#53
-
-Non-authoritative answer:
-cdn-fastly.deb.debian.org       canonical name = prod.debian.map.fastly.net.
-Name:   prod.debian.map.fastly.net
-Address: 151.101.36.204
-
-root@guest:~# apt-get update
-Ign:1 http://cdn-fastly.deb.debian.org/debian stretch InRelease
-Get:2 http://cdn-fastly.deb.debian.org/debian stretch Release [118 kB]
-Get:3 http://cdn-fastly.deb.debian.org/debian stretch Release.gpg [2434 B]
-Fetched 120 kB in 1s (110 kB/s)
-```

docs/snapshot_restore.md

@@ -8,12 +8,9 @@ snapshot and creates the exact same virtual machine, restoring the previously
 saved states. The new virtual machine is restored in a paused state, as it was
 before the snapshot was performed.
 
-This feature is important for the project as it establishes the first step
+## Snapshot a Cloud Hypervisor VM
-towards the support for live migration.
 
-## Snapshot a Cloud-Hypervisor VM
+First thing, we must run a Cloud Hypervisor VM:
-
-First thing, we must run a Cloud-Hypervisor VM:
 
 ```bash
 ./cloud-hypervisor \
@@ -46,26 +43,24 @@ ll /home/foo/snapshot/
 total 4194536
 drwxrwxr-x  2 foo bar       4096 Jul 22 11:50 ./
 drwxr-xr-x 47 foo bar       4096 Jul 22 11:47 ../
--rw-------  1 foo bar 3221225472 Jul 22 11:19 memory-region-0
+-rw-------  1 foo bar       1084 Jul 22 11:19 config.json
--rw-------  1 foo bar 1073741824 Jul 22 11:19 memory-region-1
+-rw-------  1 foo bar 4294967296 Jul 22 11:19 memory-ranges
--rw-------  1 foo bar     217853 Jul 22 11:19 vm.json
+-rw-------  1 foo bar     217853 Jul 22 11:19 state.json
 ```
 
-In this particular example, we can observe that 2 memory region files were
+`config.json` contains the virtual machine configuration. It is used to create
-created. That is explained by the size of the guest RAM, which is 4GiB in this
+a similar virtual machine with the correct amount of CPUs, RAM, and other
-case. Because it exceeds 3GiB (which is where we can find a ~1GiB memory hole),
+expected devices. It is stored in a human readable format so that it could be
-Cloud-Hypervisor needs 2 distinct memory regions to be created. Each memory
+modified between the snapshot and restore phases to achieve some very special
-region's content is stored through a dedicated file, which explains why we end
+use cases. But for most cases, manually modifying the configuration should not
-up with 2 different files, the first one containing the guest RAM range 0-3GiB
+be needed.
-and the second one containing the guest RAM range 3-4GiB.
 
-`vm.json` gathers all information related to the virtual machine configuration
+`memory-ranges` stores the content of the guest RAM.
-and state. The configuration bits are used to create a similar virtual machine
-with the correct amount of CPUs, RAM, and other expected devices. The state
-bits are used to restore each component in the state it was left before the
-snapshot occurred.
 
-## Restore a Cloud-Hypervisor VM
+`state.json` contains the virtual machine state. It is used to restore each
+component in the state it was left before the snapshot occurred.
+
+## Restore a Cloud Hypervisor VM
 
 Given that one has access to an existing snapshot in `/home/foo/snapshot`,
 it is possible to create a new VM based on this snapshot with the following 
@@ -100,13 +95,4 @@ snapshot earlier.
 
 ## Limitations
 
-The support of snapshot/restore feature is still experimental, meaning one
+VFIO devices and Intel SGX are out of scope.
-might still find some bugs associated with it.
-
-Additionally, some devices and features don't support to be snapshot and
-restored yet:
-- `vhost-user` devices
-- `virtio-mem`
-- Intel SGX
-
-VFIO devices are out of scope.

docs/uefi.md

@@ -17,7 +17,7 @@ cd edk2
 . edksetup.sh
 git submodule update --init
 
-echo "ACTIVE_PLATFORM=OvmfPkg/OvmfCh.dsc" >> Conf/target.txt
+echo "ACTIVE_PLATFORM=OvmfPkg/CloudHv/CloudHvX64.dsc" >> Conf/target.txt
 echo "TARGET_ARCH=X64" >> Conf/target.txt
 echo "TOOL_CHAIN_TAG=GCC5" >> Conf/target.txt
 
@@ -25,52 +25,15 @@ make -C ./BaseTools
 build
 ```
 
-After the successful build, the resulting firmware binaries are available under `Build/OvmfCh/DEBUG_GCC5/FV` underneath the edk2 checkout.
+After the successful build, the resulting firmware binaries are available under `Build/CloudHvX64/DEBUG_GCC5/FV` underneath the edk2 checkout.
 
 ## Using OVMF Binaries
 
 Any UEFI capable image can be booted using the Cloud Hypervisor specific firmware. Windows guests under Cloud Hypervisor only support UEFI boot, therefore OVMF is mandatory there.
 
-To make Cloud Hypervisor use UEFI boot, pass the `OVMF.fd` file path as an argument to the `--kernel` option. The firmware file will be opened in read only mode.
+To make Cloud Hypervisor use UEFI boot, pass the `CLOUDHV.fd` file path as an argument to the `--kernel` option. The firmware file will be opened in read only mode.
-
-The same firmware can be used with Cloud Hypervisor or with QEMU. This is particularly useful if using QEMU for the preparation phase.
-
-## Building UEFI Firmware with Compatibility Support Module (CSM)
-
-CSM is a module that allows to boot legacy operating systems using the OVMF firmware. OVMF can embed a CSM build of SeaBIOS. To build the SeaBIOS with CSM support, add `CONFIG_CSM=y` to `.config` before the build. The outcome `out/Csm16.bin` is to be moved into `OvmfPkg/Csm/Csm16/Csm16.bin` before OVMF is built. Then, the OVMF build will have to be passed the `-D CSM_ENABLE` option in order to generate a legacy aware UEFI firmware. At the current stage, all the necessary patches are included in the Cloud Hypervisor specific [SeaBIOS branch](https://github.com/cloud-hypervisor/seabios/tree/ch). Taking into account the previous instructions, the modified command sequence to compile an OVMF binary with CSM support is the following one:
-
-```shell
-sudo apt-get update
-sudo apt-get install uuid-dev nasm iasl build-essential python3-distutils git libncurses-dev
-
-git clone https://github.com/cloud-hypervisor/seabios -b ch
-cd seabios
-make menuconfig
-# Enable `CONFIG_CSM` and `CONFIG_QEMU_HARDWARE`
-# Please make sure `python` is in your path
-make CONFIG_CSM=y CONFIG_QEMU_HARDWARE=y
-cd ..
-
-git clone https://github.com/cloud-hypervisor/edk2 -b ch
-cd edk2
-. edksetup.sh
-git submodule update --init --recursive
-cp ../seabios/out/Csm16.bin OvmfPkg/Csm/Csm16/
-
-echo "ACTIVE_PLATFORM=OvmfPkg/OvmfCh.dsc" >> Conf/target.txt
-echo "TARGET_ARCH=X64" >> Conf/target.txt
-echo "TOOL_CHAIN_TAG=GCC5" >> Conf/target.txt
-
-make -C ./BaseTools
-build -DCSM_ENABLE
-
-```
-
-Please note, that the CSM support has currently only been tested with Linux guests. There are no plans to provide legacy support for other OSes (e.g. Windows).
 
 # Links
 
 - [OVMF wiki](https://github.com/tianocore/tianocore.github.io/wiki/OVMF) 
 - [Cloud Hypervisor specific tree](https://github.com/cloud-hypervisor/edk2/tree/ch)
-- [Redhat OVMF Status Report](https://access.redhat.com/sites/default/files/attachments/ovmf-whtepaper-031815.pdf)
-- [SeaBIOS Build Overview](https://www.seabios.org/Build_overview#Build_as_a_UEFI_Compatibility_Support_Module_.28CSM.29)

docs/vfio-user.md

@@ -45,7 +45,9 @@ Use SPDK: https://github.com/spdk/spdk
 
 Compile with `./configure --with-vfio-user`
 
-Create an NVMe controller listening on a vfio-user socket with a simple block device:
+Create an NVMe controller listening on a vfio-user socket with a simple AIO block device in spdk.
+More details of configuring SPDK bdev can be viewed in [SPDK bdev](https://spdk.io/doc/bdev.html).
+More details of setting SPDK NVMe-oF target can be viewed in [SDPK NVMe-oF tgt](https://spdk.io/doc/nvmf.html).
 
 ```sh
 sudo scripts/setup.sh

docs/vfio.md

@@ -36,14 +36,22 @@ Now that we have identified the device, we must unbind it from its native driver
 First we add VFIO support to the host:
 
 ```
-$ sudo modprobe vfio_pci
+# modprobe -r vfio_pci
-$ sudo modprobe vfio_iommu_type1 allow_unsafe_interrupts
+# modprobe -r vfio_iommu_type1
+# modprobe vfio_iommu_type1 allow_unsafe_interrupts
+# modprobe vfio_pci
+```
+
+In case the VFIO drivers are built-in, enable unsafe interrupts with:
+
+```
+# echo 1 > /sys/module/vfio_iommu_type1/parameters/allow_unsafe_interrupts
 ```
 
 Then we unbind it from its native driver:
 
 ```
-$ echo 0000:01:00.0 > /sys/bus/pci/devices/0000\:01\:00.0/driver/unbind
+# echo 0000:01:00.0 > /sys/bus/pci/devices/0000\:01\:00.0/driver/unbind
 ```
 
 And finally we bind it to the VFIO driver. To do that we first need to get the
@@ -53,8 +61,14 @@ device's VID (Vendor ID) and PID (Product ID):
 $ lspci -n -s 01:00.0
 01:00.0 ff00: 10ec:525a (rev 01)
 
-$ echo 10ec 525a > /sys/bus/pci/drivers/vfio-pci/new_id
+# echo 10ec 525a > /sys/bus/pci/drivers/vfio-pci/new_id
-$ echo 0000:01:00.0 > /sys/bus/pci/drivers/vfio-pci/bind
+```
+
+If you have more than one device with the same `vendorID`/`deviceID`, starting
+with the second device, the binding is performed as follows:
+
+```
+# echo 0000:02:00.0 > /sys/bus/pci/drivers/vfio-pci/bind
 ```
 
 Now the device is managed by the VFIO framework.
@@ -79,3 +93,33 @@ takes the device's sysfs path as an argument. In our example it is
 The guest kernel will then detect the card reader on its PCI bus and provided
 that support for this device is enabled, it will probe and enable it for the
 guest to use.
+
+In case you want to pass multiple devices, here is the correct syntax:
+
+```
+--device path=/sys/bus/pci/devices/0000:01:00.0/ path=/sys/bus/pci/devices/0000:02:00.0/
+```
+
+### Multiple devices in the same IOMMU group
+
+There are cases where multiple devices can be found under the same IOMMU group.
+This happens often with graphics card embedding an audio controller.
+
+```
+$ lspci
+[...]
+01:00.0 VGA compatible controller: NVIDIA Corporation GK208B [GeForce GT 710] (rev a1)
+01:00.1 Audio device: NVIDIA Corporation GK208 HDMI/DP Audio Controller (rev a1)
+[...]
+```
+
+This is usually exposed as follows through `sysfs`:
+
+```
+$ ls /sys/kernel/iommu_groups/22/devices/
+0000:01:00.0  0000:01:00.1
+```
+
+This means these two devices are under the same IOMMU group 22. In such case,
+it is important to bind both devices to VFIO and pass them both through the
+VM, otherwise this could cause some functional and security issues.

docs/vhost-user-blk-testing.md

@@ -69,12 +69,16 @@ make
 ```
 
 ### Set the SPDK environment
+
+Run SPDK vhost target with 2 CPU cores, i.e., core 0 and 1.
 ```bash
 sudo HUGEMEM=2048 scripts/setup.sh
 sudo ./build/bin/vhost -S /var/tmp -s 1024 -m 0x3 &
 ```
 
 ### Create 512M block device
+
+Create a 512M (first parameter) block device with 512 bytes (second parameter) block size.
 ```bash
 sudo scripts/rpc.py bdev_malloc_create 512 512 -b Malloc0
 sudo scripts/rpc.py vhost_create_blk_controller --cpumask 0x1 vhost.1 Malloc0

docs/virtiofs-root.md

@@ -65,7 +65,7 @@ iface eth0 inet dhcp
 
 ```bash
 # starting in the directory above rootfs
-sudo virtiofsd --socket-path=$PWD/virtiofs-rootfs.sock -o source=$PWD/rootfs -o cache=none &
+sudo virtiofsd --socket-path=$PWD/virtiofs-rootfs.sock --shared-dir=$PWD/rootfs --cache=never &
 sudo cloud-hypervisor \
     --cpus boot=1,max=1 \
     --kernel vmlinux \

docs/windows.md

@@ -22,7 +22,7 @@ __Prerequisites__
 - QEMU, version >=5.0.0 is recommended.
 - Windows installation ISO. Obtained through MSDN, Visual Studio subscription, evaluation center, etc. 
 - [VirtIO driver ISO](https://fedorapeople.org/groups/virt/virtio-win/direct-downloads/stable-virtio/)
-- Suitable [OVMF](uefi.md) firmware 
+- Suitable firmware for Cloud Hypervisor (`CLOUDHV.fd`) and for QEMU (`OVMF.fd`)
 - With the suggested image size of 30G, there should be enough free disk space to hold the installation ISO and any other necessary files
 
 This step currently requires QEMU to install Windows onto the guest. QEMU is only used at the preparation stage, the resulting image is then fully functional with Cloud Hypervisor.
@@ -67,7 +67,7 @@ The basic command to boot a Windows image. The configuration section should be c
 
 ```shell
 cloud-hypervisor \
-	--kernel ./$OVMF_DIR/OVMF.fd \
+	--kernel ./$OVMF_DIR/CLOUDHV.fd \
 	--disk path=./$IMG_FILE \
 	--cpus boot=1,kvm_hyperv=on \
 	--memory size=4G \

event_monitor/Cargo.toml

@@ -5,7 +5,7 @@ authors = ["The Cloud Hypervisor Authors"]
 edition = "2018"
 
 [dependencies]
-libc = "0.2.103"
+libc = "0.2.119"
-serde = { version = "1.0.130", features = ["rc"] }
+serde = { version = "1.0.136", features = ["rc"] }
-serde_derive = "1.0.130"
+serde_derive = "1.0.136"
-serde_json = "1.0.68"
+serde_json = "1.0.79"

fuzz/Cargo.lock

@@ -9,20 +9,11 @@ dependencies = [
  "vm-memory",
 ]
 
-[[package]]
-name = "ansi_term"
-version = "0.11.0"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "ee49baf6cb617b853aa8d93bf420db2383fab46d314482ca2803b40d5fde979b"
-dependencies = [
- "winapi",
-]
-
 [[package]]
 name = "anyhow"
-version = "1.0.44"
+version = "1.0.55"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "61604a8f862e1d5c3229fdd78f8b02c68dcf73a4c4b05fd636d12240aaa242c1"
+checksum = "159bb86af3a200e19a068f4224eae4c8bb2d0fa054c7e5d1cacd5cef95e684cd"
 
 [[package]]
 name = "api_client"
@@ -33,15 +24,15 @@ dependencies = [
 
 [[package]]
 name = "arbitrary"
-version = "1.0.2"
+version = "1.1.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "577b08a4acd7b99869f863c50011b01eb73424ccc798ecd996f2e24817adfca7"
+checksum = "c38b6b6b79f671c25e1a3e785b7b82d7562ffc9cd3efdc98627e5668a2472490"
 
 [[package]]
 name = "arc-swap"
-version = "1.4.0"
+version = "1.5.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "e6df5aef5c5830360ce5218cecb8f018af3438af5686ae945094affc86fdec63"
+checksum = "c5d78ce20460b82d3fa150275ed9d55e21064fc7951177baacf86a145c4a4b1f"
 
 [[package]]
 name = "arch"
@@ -77,6 +68,12 @@ dependencies = [
  "winapi",
 ]
 
+[[package]]
+name = "autocfg"
+version = "1.1.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "d468802bab17cbc0cc575e9b053f41e72aa36bfa6b7f55e3529ffa43161b97fa"
+
 [[package]]
 name = "bincode"
 version = "1.3.3"
@@ -105,6 +102,7 @@ dependencies = [
  "versionize_derive",
  "vhdx",
  "virtio-bindings",
+ "virtio-queue",
  "vm-memory",
  "vm-virtio",
  "vmm-sys-util",
@@ -118,9 +116,9 @@ checksum = "14c189c53d098945499cdfa7ecc63567cf3886b3332b312a5b4585d8d3a6a610"
 
 [[package]]
 name = "cc"
-version = "1.0.71"
+version = "1.0.73"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "79c2681d6594606957bbb8631c4b90a7fcaaa72cdb714743a437b156d6a7eedd"
+checksum = "2fff2a6927b3bb87f9595d67196a70493f627687a71d87a0d692242c33f58c11"
 
 [[package]]
 name = "cfg-if"
@@ -130,23 +128,24 @@ checksum = "baf1de4339761588bc0619e3cbc0120ee582ebb74b53b4efbf79117bd2da40fd"
 
 [[package]]
 name = "clap"
-version = "2.33.3"
+version = "3.1.5"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "37e58ac78573c40708d45522f0d80fa2f01cc4f9b4e2bf749807255454312002"
+checksum = "ced1892c55c910c1219e98d6fc8d71f6bddba7905866ce740066d8bfea859312"
 dependencies = [
- "ansi_term",
  "atty",
  "bitflags",
+ "indexmap",
+ "lazy_static",
+ "os_str_bytes",
  "strsim",
- "term_size",
+ "termcolor",
+ "terminal_size",
  "textwrap",
- "unicode-width",
- "vec_map",
 ]
 
 [[package]]
 name = "cloud-hypervisor"
-version = "18.0.0"
+version = "21.0.0"
 dependencies = [
  "anyhow",
  "api_client",
@@ -178,6 +177,7 @@ dependencies = [
  "seccompiler",
  "vhdx",
  "virtio-devices",
+ "virtio-queue",
  "vm-memory",
  "vm-virtio",
  "vmm-sys-util",
@@ -185,9 +185,9 @@ dependencies = [
 
 [[package]]
 name = "crc32c"
-version = "0.6.0"
+version = "0.6.1"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "210cdf933e6a81212bfabf90cd8762f471b5922e5f6b709547673ad8e04b9448"
+checksum = "ee6b9c9389584bcba988bd0836086789b7f87ad91892d6a83d5291dbb24524b5"
 dependencies = [
  "rustc_version",
 ]
@@ -245,10 +245,34 @@ source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "b643857cf70949306b81d7e92cb9d47add673868edac9863c4a49c42feaf3f1e"
 
 [[package]]
-name = "getrandom"
+name = "gdbstub"
-version = "0.2.3"
+version = "0.6.1"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "7fcd999463524c52659517fe2cea98493cfe485d10565e7b0fb07dbba7ad2753"
+checksum = "9fa2ca5d6b045de372cef3991f873389b421b4cabcfbe52f7787fae8b8b37906"
+dependencies = [
+ "bitflags",
+ "cfg-if",
+ "log",
+ "managed",
+ "num-traits",
+ "paste",
+]
+
+[[package]]
+name = "gdbstub_arch"
+version = "0.2.2"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "51dc4b5718ac76d21e8605c0966dd32d80273b89b11e6cfef467b04e45934d37"
+dependencies = [
+ "gdbstub",
+ "num-traits",
+]
+
+[[package]]
+name = "getrandom"
+version = "0.2.5"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "d39cd93900197114fa1fcb7ae84ca742095eed9442088988ae74fa744e930e77"
 dependencies = [
  "cfg-if",
  "libc",
@@ -256,10 +280,16 @@ dependencies = [
 ]
 
 [[package]]
-name = "hermit-abi"
+name = "hashbrown"
-version = "0.1.19"
+version = "0.11.2"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "62b467343b94ba476dcb2500d242dadbb39557df889310ac77c5d99100aaac33"
+checksum = "ab5ef0d4909ef3724cc8cce6ccc8572c5c817592e9285f5464f8e86f8bd3726e"
+
+[[package]]
+name = "hermit-abi"
+version = "0.1.20"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "c7a30908dbce072eca83216eab939d2290080e00ca71611b96a09e5cdce5f3fa"
 dependencies = [
  "libc",
 ]
@@ -285,14 +315,24 @@ dependencies = [
 
 [[package]]
 name = "iced-x86"
-version = "1.15.0"
+version = "1.17.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "46e977036f7f5139d580c7f19ad62df9cb8ebd8410bb569e73585226be80a86f"
+checksum = "158f5204401d08f91d19176112146d75e99b3cf745092e268fa7be33e09adcec"
 dependencies = [
  "lazy_static",
  "static_assertions",
 ]
 
+[[package]]
+name = "indexmap"
+version = "1.8.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "282a6247722caba404c065016bbfa522806e51714c34f5dfc3e4a3a46fcb4223"
+dependencies = [
+ "autocfg",
+ "hashbrown",
+]
+
 [[package]]
 name = "io-uring"
 version = "0.5.2"
@@ -305,14 +345,14 @@ dependencies = [
 
 [[package]]
 name = "itoa"
-version = "0.4.8"
+version = "1.0.1"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "b71991ff56294aa922b450139ee08b3bfc70982c6b2c7562771375cf73542dd4"
+checksum = "1aab8fc367588b89dcee83ab0fd66b72b50b72fa1904d7095045ace2b0c81c35"
 
 [[package]]
 name = "kvm-bindings"
 version = "0.5.0"
-source = "git+https://github.com/cloud-hypervisor/kvm-bindings?branch=ch-v0.5.0#9c497710ba9968d8efe15dbae03991b16cf82e23"
+source = "git+https://github.com/cloud-hypervisor/kvm-bindings?branch=ch-v0.5.0-tdx#52e56d0e8ef0f6ea32fc0492e6a175b73617a49f"
 dependencies = [
  "serde",
  "serde_derive",
@@ -321,9 +361,8 @@ dependencies = [
 
 [[package]]
 name = "kvm-ioctls"
-version = "0.10.0"
+version = "0.11.0"
-source = "registry+https://github.com/rust-lang/crates.io-index"
+source = "git+https://github.com/rust-vmm/kvm-ioctls?branch=main#1e03e29cdfbb0cb108a98de7a78045a5a517f18e"
-checksum = "48dc14f9047df1873cf6942caccc7431d19c3d496ca7a0d162260c4cf0f64b76"
 dependencies = [
  "kvm-bindings",
  "libc",
@@ -338,9 +377,9 @@ checksum = "e2abad23fbc42b3700f2f279844dc832adb2b2eb069b2df918f455c4e18cc646"
 
 [[package]]
 name = "libc"
-version = "0.2.103"
+version = "0.2.119"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "dd8f7255a17a627354f321ef0055d63b898c6fb27eff628af4d1b66b7331edf6"
+checksum = "1bf2e165bb3457c8e098ea76f3e3bc9db55f87aa90d52d0e6be741470916aaa4"
 
 [[package]]
 name = "libfuzzer-sys"
@@ -371,6 +410,18 @@ dependencies = [
  "cfg-if",
 ]
 
+[[package]]
+name = "managed"
+version = "0.8.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "0ca88d725a0a943b096803bd34e73a4437208b6077654cc4ecb2947a5f91618d"
+
+[[package]]
+name = "memchr"
+version = "2.4.1"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "308cc39be01b73d0d18f82a0e7b2a3df85245f84af96fdddc5d202d27e47b86a"
+
 [[package]]
 name = "micro_http"
 version = "0.1.0"
@@ -401,21 +452,46 @@ dependencies = [
  "versionize",
  "versionize_derive",
  "virtio-bindings",
+ "virtio-queue",
  "vm-memory",
  "vm-virtio",
  "vmm-sys-util",
 ]
 
 [[package]]
-name = "once_cell"
+name = "num-traits"
-version = "1.8.0"
+version = "0.2.14"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "692fcb63b64b1758029e0a96ee63e049ce8c5948587f2f7208df04625e5f6b56"
+checksum = "9a64b1ec5cda2586e284722486d802acf1f7dbdc623e2bfc57e65ca1cd099290"
+dependencies = [
+ "autocfg",
+]
+
+[[package]]
+name = "once_cell"
+version = "1.9.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "da32515d9f6e6e489d7bc9d84c71b060db7247dc035bbe44eac88cf87486d8d5"
 
 [[package]]
 name = "option_parser"
 version = "0.1.0"
 
+[[package]]
+name = "os_str_bytes"
+version = "6.0.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "8e22443d1643a904602595ba1cd8f7d896afe56d26712531c5ff73a15b2fbf64"
+dependencies = [
+ "memchr",
+]
+
+[[package]]
+name = "paste"
+version = "1.0.6"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "0744126afe1a6dd7f394cb50a716dbe086cb06e255e53d8d0185d82828358fb5"
+
 [[package]]
 name = "pci"
 version = "0.1.0"
@@ -425,6 +501,8 @@ dependencies = [
  "hypervisor",
  "libc",
  "log",
+ "serde",
+ "serde_derive",
  "thiserror",
  "versionize",
  "versionize_derive",
@@ -440,9 +518,9 @@ dependencies = [
 
 [[package]]
 name = "proc-macro2"
-version = "1.0.30"
+version = "1.0.36"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "edc3358ebc67bc8b7fa0c007f945b0b18226f78437d61bec735a9eb96b61ee70"
+checksum = "c7342d5883fbccae1cc37a2353b09c87c9b0f3afd73f5fb9bba687a1f733b029"
 dependencies = [
  "unicode-xid",
 ]
@@ -460,9 +538,9 @@ dependencies = [
 
 [[package]]
 name = "quote"
-version = "1.0.10"
+version = "1.0.15"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "38bc8cc6a5f2e3655e0899c1b848643b2562f853f114bfec7be120678e3ace05"
+checksum = "864d3e96a899863136fc6e99f3d7cae289dafe43bf2c5ac19b70df7210c0a145"
 dependencies = [
  "proc-macro2",
 ]
@@ -489,18 +567,18 @@ dependencies = [
 
 [[package]]
 name = "rustc_version"
-version = "0.2.3"
+version = "0.4.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "138e3e0acb6c9fb258b19b67cb8abd63c00679d2851805ea151465464fe9030a"
+checksum = "bfa0f585226d2e68097d4f95d113b15b83a82e819ab25717ec0590d9584ef366"
 dependencies = [
  "semver",
 ]
 
 [[package]]
 name = "ryu"
-version = "1.0.5"
+version = "1.0.9"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "71d301d4193d031abdd79ff7e3dd721168a9572ef3fe51a1517aba235bd8f86e"
+checksum = "73b4b750c782965c211b42f022f59af1fbceabdd026623714f104152f1ec149f"
 
 [[package]]
 name = "seccompiler"
@@ -513,30 +591,21 @@ dependencies = [
 
 [[package]]
 name = "semver"
-version = "0.9.0"
+version = "1.0.6"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "1d7eb9ef2c18661902cc47e535f9bc51b78acd254da71d375c2f6720d9a40403"
+checksum = "a4a3381e03edd24287172047536f20cabde766e2cd3e65e6b00fb3af51c4f38d"
-dependencies = [
- "semver-parser",
-]
-
-[[package]]
-name = "semver-parser"
-version = "0.7.0"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "388a1df253eca08550bef6c72392cfe7c30914bf41df5269b68cbd6ff8f570a3"
 
 [[package]]
 name = "serde"
-version = "1.0.130"
+version = "1.0.136"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "f12d06de37cf59146fbdecab66aa99f9fe4f78722e3607577a5375d66bd0c913"
+checksum = "ce31e24b01e1e524df96f1c2fdd054405f8d7376249a5110886fb4b658484789"
 
 [[package]]
 name = "serde_derive"
-version = "1.0.130"
+version = "1.0.136"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "d7bc1a1ab1961464eae040d96713baa5a724a8152c1222492465b54322ec508b"
+checksum = "08597e7152fcd306f41838ed3e37be9eaeed2b61c42e2117266a554fab4662f9"
 dependencies = [
  "proc-macro2",
  "quote",
@@ -545,9 +614,9 @@ dependencies = [
 
 [[package]]
 name = "serde_json"
-version = "1.0.68"
+version = "1.0.79"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "0f690853975602e1bfe1ccbf50504d67174e3bcf340f23b5ea9992e0587a52d8"
+checksum = "8e8d9fa5c3b304765ce1fd9c4c8a3de2c8db365a5b91be52f186efc675681d95"
 dependencies = [
  "itoa",
  "ryu",
@@ -556,9 +625,9 @@ dependencies = [
 
 [[package]]
 name = "signal-hook"
-version = "0.3.10"
+version = "0.3.13"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "9c98891d737e271a2954825ef19e46bd16bdb98e2746f2eec4f7a4ef7946efd1"
+checksum = "647c97df271007dcea485bb74ffdb57f2e683f1306c854f468a0c244badabf2d"
 dependencies = [
  "libc",
  "signal-hook-registry",
@@ -581,15 +650,15 @@ checksum = "a2eb9349b6444b326872e140eb1cf5e7c522154d69e7a0ffb0fb81c06b37543f"
 
 [[package]]
 name = "strsim"
-version = "0.8.0"
+version = "0.10.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "8ea5119cdb4c55b55d432abb513a0429384878c15dde60cc77b1c99de1a95a6a"
+checksum = "73473c0e59e6d5812c5dfe2a064a6444949f089e20eec9a2e5506596494e4623"
 
 [[package]]
 name = "syn"
-version = "1.0.80"
+version = "1.0.86"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "d010a1623fbd906d51d650a9916aaefc05ffa0e4053ff7fe601167f3e715d194"
+checksum = "8a65b3f4ffa0092e9887669db0eae07941f023991ab58ea44da8fe8e2d511c6b"
 dependencies = [
  "proc-macro2",
  "quote",
@@ -597,10 +666,19 @@ dependencies = [
 ]
 
 [[package]]
-name = "term_size"
+name = "termcolor"
-version = "0.3.2"
+version = "1.1.3"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "1e4129646ca0ed8f45d09b929036bafad5377103edd06e50bf574b353d2b08d9"
+checksum = "bab24d30b911b2376f3a13cc2cd443142f0c81dda04c118693e35b3835757755"
+dependencies = [
+ "winapi-util",
+]
+
+[[package]]
+name = "terminal_size"
+version = "0.1.17"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "633c1a546cee861a1a6d0dc69ebeca693bf4296661ba7852b9d21d159e0506df"
 dependencies = [
  "libc",
  "winapi",
@@ -608,12 +686,11 @@ dependencies = [
 
 [[package]]
 name = "textwrap"
-version = "0.11.0"
+version = "0.15.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "d326610f408c7a4eb6f51c37c330e496b08506c9457c9d34287ecc38809fb060"
+checksum = "b1141d4d61095b28419e22cb0bbf02755f5e54e0526f97f1e3d1d160e60885fb"
 dependencies = [
- "term_size",
+ "terminal_size",
- "unicode-width",
 ]
 
 [[package]]
@@ -636,12 +713,6 @@ dependencies = [
  "syn",
 ]
 
-[[package]]
-name = "unicode-width"
-version = "0.1.9"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "3ed742d4ea2bd1176e236172c8429aaf54486e7ac098db29ffe6529e0ce50973"
-
 [[package]]
 name = "unicode-xid"
 version = "0.2.2"
@@ -657,12 +728,6 @@ dependencies = [
  "getrandom",
 ]
 
-[[package]]
-name = "vec_map"
-version = "0.8.2"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "f1bddf1187be692e79c5ffeab891132dfb0f236ed36a43c7ed39f1165ee20191"
-
 [[package]]
 name = "versionize"
 version = "0.1.6"
@@ -692,9 +757,9 @@ dependencies = [
 
 [[package]]
 name = "vfio-bindings"
-version = "0.2.0"
+version = "0.3.1"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "4a21f546f2bda37f5a8cfb138c87f95b8e34d2d78d6a7a92ba3785f4e08604a7"
+checksum = "43449b404c488f70507dca193debd4bea361fe8089869b947adc19720e464bce"
 dependencies = [
  "vmm-sys-util",
 ]
@@ -702,12 +767,14 @@ dependencies = [
 [[package]]
 name = "vfio-ioctls"
 version = "0.1.0"
-source = "git+https://github.com/rust-vmm/vfio-ioctls?branch=main#d51c1fad37be30c20385c55a217e2d90972ea31d"
+source = "git+https://github.com/rust-vmm/vfio?branch=main#4630612f2ff300e78b6d55be324fb5481aa84661"
 dependencies = [
  "byteorder",
  "kvm-bindings",
  "kvm-ioctls",
+ "libc",
  "log",
+ "thiserror",
  "vfio-bindings",
  "vm-memory",
  "vmm-sys-util",
@@ -745,9 +812,9 @@ dependencies = [
 
 [[package]]
 name = "vhost"
-version = "0.2.0"
+version = "0.3.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "2d23ddeb452fb4f837526c6298cc8a2f4948e5595b0328a3d61b5eebe51874d7"
+checksum = "3b56bf8f178fc500fe14505fca8b00dec76fc38f2304f461c8d9d7547982311d"
 dependencies = [
  "bitflags",
  "libc",
@@ -786,6 +853,7 @@ dependencies = [
  "versionize_derive",
  "vhost",
  "virtio-bindings",
+ "virtio-queue",
  "vm-allocator",
  "vm-device",
  "vm-memory",
@@ -794,6 +862,16 @@ dependencies = [
  "vmm-sys-util",
 ]
 
+[[package]]
+name = "virtio-queue"
+version = "0.1.0"
+source = "git+https://github.com/rust-vmm/vm-virtio?branch=main#bbb22d43558a76cc7cc9c262c4be870a9e682a86"
+dependencies = [
+ "log",
+ "vm-memory",
+ "vmm-sys-util",
+]
+
 [[package]]
 name = "vm-allocator"
 version = "0.1.0"
@@ -824,9 +902,9 @@ source = "git+https://github.com/rust-vmm/vm-fdt?branch=main#720e48e435b791ec6cb
 
 [[package]]
 name = "vm-memory"
-version = "0.6.0"
+version = "0.7.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "0a8ebcb86ca457f9d6e14cf97009f679952eba42f0113de5db596e514cd0e43b"
+checksum = "339d4349c126fdcd87e034631d7274370cf19eb0e87b33166bcd956589fc72c5"
 dependencies = [
  "arc-swap",
  "libc",
@@ -853,6 +931,7 @@ version = "0.1.0"
 dependencies = [
  "log",
  "virtio-bindings",
+ "virtio-queue",
  "vm-memory",
 ]
 
@@ -870,6 +949,8 @@ dependencies = [
  "devices",
  "epoll",
  "event_monitor",
+ "gdbstub",
+ "gdbstub_arch",
  "hypervisor",
  "lazy_static",
  "libc",
@@ -893,6 +974,7 @@ dependencies = [
  "vfio_user",
  "vhdx",
  "virtio-devices",
+ "virtio-queue",
  "vm-allocator",
  "vm-device",
  "vm-memory",
@@ -935,6 +1017,15 @@ version = "0.4.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "ac3b87c63620426dd9b991e5ce0329eff545bccbbb34f3be09ff6fb6ab51b7b6"
 
+[[package]]
+name = "winapi-util"
+version = "0.1.5"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "70ec6ce85bb158151cae5e5c87f95a8e97d2c0c4b001223f33a334e3ce5de178"
+dependencies = [
+ "winapi",
+]
+
 [[package]]
 name = "winapi-x86_64-pc-windows-gnu"
 version = "0.4.0"

fuzz/Cargo.toml

@@ -10,21 +10,23 @@ cargo-fuzz = true
 
 [dependencies]
 block_util = { path = "../block_util" }
-libc = "0.2.103"
+libc = "0.2.119"
 libfuzzer-sys = "0.4.2"
 qcow = { path = "../qcow" }
 seccompiler = "0.2.0"
 vhdx = { path = "../vhdx" }
 virtio-devices = { path = "../virtio-devices" }
+virtio-queue = { git = "https://github.com/rust-vmm/vm-virtio", branch = "main" }
 vmm-sys-util = "0.9.0"
 vm-virtio = { path = "../vm-virtio" }
-vm-memory = "0.6.0"
+vm-memory = "0.7.0"
 
 [dependencies.cloud-hypervisor]
 path = ".."
 
 [patch.crates-io]
-kvm-bindings = { git = "https://github.com/cloud-hypervisor/kvm-bindings", branch = "ch-v0.5.0", features = ["with-serde", "fam-wrappers"] }
+kvm-bindings = { git = "https://github.com/cloud-hypervisor/kvm-bindings", branch = "ch-v0.5.0-tdx" }
+kvm-ioctls = { git = "https://github.com/rust-vmm/kvm-ioctls", branch = "main" }
 versionize_derive = { git = "https://github.com/cloud-hypervisor/versionize_derive", branch = "ch" }
 
 # Prevent this from interfering with workspaces

fuzz/fuzz_targets/block.rs

@@ -15,10 +15,12 @@ use std::os::unix::io::{AsRawFd, FromRawFd, RawFd};
 use std::path::PathBuf;
 use std::sync::Arc;
 use virtio_devices::{Block, VirtioDevice, VirtioInterrupt, VirtioInterruptType};
-use vm_memory::{Bytes, GuestAddress, GuestMemoryAtomic, GuestMemoryMmap};
+use virtio_queue::{Queue, QueueState};
-use vm_virtio::Queue;
+use vm_memory::{bitmap::AtomicBitmap, Bytes, GuestAddress, GuestMemoryAtomic};
 use vmm_sys_util::eventfd::{EventFd, EFD_NONBLOCK};
 
+type GuestMemoryMmap = vm_memory::GuestMemoryMmap<AtomicBitmap>;
+
 const MEM_SIZE: u64 = 256 * 1024 * 1024;
 const DESC_SIZE: u64 = 16; // Bytes in one virtio descriptor.
 const QUEUE_SIZE: u16 = 16; // Max entries in the queue.
@@ -73,10 +75,14 @@ fuzz_target!(|bytes| {
         return;
     }
 
-    let mut q = Queue::new(QUEUE_SIZE);
+    let guest_memory = GuestMemoryAtomic::new(mem);
-    q.ready = true;
+
-    q.size = QUEUE_SIZE / 2;
+    let mut q = Queue::<
-    q.max_size = QUEUE_SIZE;
+        GuestMemoryAtomic<GuestMemoryMmap>,
+        QueueState,
+    >::new(guest_memory.clone(), QUEUE_SIZE);
+    q.state.ready = true;
+    q.state.size = QUEUE_SIZE / 2;
 
     let queue_evts: Vec<EventFd> = vec![EventFd::new(0).unwrap()];
     let queue_fd = queue_evts[0].as_raw_fd();
@@ -102,7 +108,7 @@ fuzz_target!(|bytes| {
 
     block
         .activate(
-            GuestMemoryAtomic::new(mem),
+            guest_memory,
             Arc::new(NoopVirtioInterrupt {}),
             vec![q],
             queue_evts,
@@ -133,8 +139,7 @@ pub struct NoopVirtioInterrupt {}
 impl VirtioInterrupt for NoopVirtioInterrupt {
     fn trigger(
         &self,
-        _int_type: &VirtioInterruptType,
+        _int_type: VirtioInterruptType,
-        _queue: Option<&Queue>,
     ) -> std::result::Result<(), std::io::Error> {
         Ok(())
     }

hypervisor/Cargo.toml

@@ -11,23 +11,23 @@ mshv = ["mshv-ioctls", "mshv-bindings"]
 tdx = []
 
 [dependencies]
-anyhow = "1.0.44"
+anyhow = "1.0.55"
 epoll = "4.3.1"
 thiserror = "1.0.30"
-libc = "0.2.103"
+libc = "0.2.119"
 log = "0.4.14"
-kvm-ioctls = { version = "0.10.0", optional = true }
+kvm-ioctls = { version = "0.11.0", optional = true }
-kvm-bindings = { git = "https://github.com/cloud-hypervisor/kvm-bindings", branch = "ch-v0.5.0", features = ["with-serde", "fam-wrappers"], optional  = true }
+kvm-bindings = { git = "https://github.com/cloud-hypervisor/kvm-bindings", branch = "ch-v0.5.0-tdx", features = ["with-serde", "fam-wrappers"], optional  = true }
 mshv-bindings = {git = "https://github.com/rust-vmm/mshv", branch = "main", features = ["with-serde", "fam-wrappers"], optional  = true }
 mshv-ioctls = { git = "https://github.com/rust-vmm/mshv", branch = "main", optional  = true}
-serde = { version = "1.0.130", features = ["rc"] }
+serde = { version = "1.0.136", features = ["rc"] }
-serde_derive = "1.0.130"
+serde_derive = "1.0.136"
-serde_json = "1.0.68"
+serde_json = "1.0.79"
-vm-memory = { version = "0.6.0", features = ["backend-mmap", "backend-atomic"] }
+vm-memory = { version = "0.7.0", features = ["backend-mmap", "backend-atomic"] }
 vmm-sys-util = { version = "0.9.0", features = ["with-serde"] }
 
 [target.'cfg(target_arch = "x86_64")'.dependencies.iced-x86]
-version = "1.15.0"
+version = "1.17.0"
 default-features = false
 features = ["std", "decoder", "op_code_info", "instr_info", "fast_fmt"]
 

hypervisor/src/arch/emulator/mod.rs

@@ -18,7 +18,18 @@ pub struct Exception<T: Debug> {
 
 impl<T: Debug> Display for Exception<T> {
     fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result {
-        write!(f, "Exception {:?} at IP {:#x}", self.vector, self.ip)
+        write!(
+            f,
+            "Exception {:?} at IP {:#x}{}{}",
+            self.vector,
+            self.ip,
+            self.error
+                .map(|e| format!(": error {:x}", e))
+                .unwrap_or_else(|| "".to_owned()),
+            self.payload
+                .map(|payload| format!(": payload {:x}", payload))
+                .unwrap_or_else(|| "".to_owned())
+        )
     }
 }
 

hypervisor/src/arch/x86/emulator/mod.rs

@@ -653,8 +653,6 @@ mod mock_vmm {
         state: Arc<Mutex<CpuState>>,
     }
 
-    unsafe impl Sync for MockVmm {}
-
     pub type MockResult = Result<(), EmulationError<Exception>>;
 
     impl MockVmm {

hypervisor/src/cpu.rs

@@ -19,15 +19,19 @@ use crate::x86_64::{
     ExtendedControlRegisters, FpuState, MsrEntries, SpecialRegisters, StandardRegisters, VcpuEvents,
 };
 use crate::CpuState;
+#[cfg(target_arch = "aarch64")]
+use crate::DeviceAttr;
 #[cfg(feature = "kvm")]
 use crate::MpState;
 #[cfg(all(feature = "mshv", target_arch = "x86_64"))]
 use crate::SuspendRegisters;
 #[cfg(target_arch = "x86_64")]
 use crate::Xsave;
-#[cfg(feature = "mshv")]
+#[cfg(feature = "tdx")]
-use mshv_bindings::*;
+use crate::{TdxExitDetails, TdxExitStatus};
 use thiserror::Error;
+#[cfg(all(feature = "kvm", target_arch = "x86_64"))]
+use vm_memory::GuestAddress;
 
 #[derive(Error, Debug)]
 ///
@@ -174,6 +178,16 @@ pub enum HypervisorCpuError {
     #[error("Failed to get debug registers: {0}")]
     GetDebugRegs(#[source] anyhow::Error),
     ///
+    /// Setting misc register error
+    ///
+    #[error("Failed to set misc registers: {0}")]
+    SetMiscRegs(#[source] anyhow::Error),
+    ///
+    /// Getting misc register error
+    ///
+    #[error("Failed to get misc registers: {0}")]
+    GetMiscRegs(#[source] anyhow::Error),
+    ///
     /// Write to Guest Mem
     ///
     #[error("Failed to write to Guest Mem at: {0}")]
@@ -213,11 +227,27 @@ pub enum HypervisorCpuError {
     #[error("Failed to translate GVA: {0}")]
     TranslateVirtualAddress(#[source] anyhow::Error),
     ///
+    /// Set cpu attribute error
+    ///
+    #[error("Failed to set vcpu attribute: {0}")]
+    SetVcpuAttribute(#[source] anyhow::Error),
+    ///
+    /// Check if cpu has a certain attribute error
+    ///
+    #[error("Failed to check if vcpu has attribute: {0}")]
+    HasVcpuAttribute(#[source] anyhow::Error),
+    ///
     /// Failed to initialize TDX on CPU
     ///
     #[cfg(feature = "tdx")]
     #[error("Failed to initialize TDX: {0}")]
     InitializeTdx(#[source] std::io::Error),
+    ///
+    /// Unknown TDX VM call
+    ///
+    #[cfg(feature = "tdx")]
+    #[error("Unknown TDX VM call")]
+    UnknownTdxVmCall,
 }
 
 #[derive(Debug)]
@@ -234,6 +264,10 @@ pub enum VmExit<'a> {
     Reset,
     Shutdown,
     Hyperv,
+    #[cfg(feature = "tdx")]
+    Tdx,
+    #[cfg(feature = "kvm")]
+    Debug,
 }
 
 ///
@@ -249,7 +283,16 @@ pub trait Vcpu: Send + Sync {
     /// Returns the vCPU general purpose registers.
     ///
     fn get_regs(&self) -> Result<StandardRegisters>;
-
+    #[cfg(target_arch = "aarch64")]
+    ///
+    /// Sets vcpu attribute
+    ///
+    fn set_vcpu_attr(&self, attr: &DeviceAttr) -> Result<()>;
+    #[cfg(target_arch = "aarch64")]
+    ///
+    /// Check if vcpu has attribute.
+    ///
+    fn has_vcpu_attr(&self, attr: &DeviceAttr) -> Result<()>;
     #[cfg(target_arch = "x86_64")]
     ///
     /// Sets the vCPU general purpose registers.
@@ -358,6 +401,11 @@ pub trait Vcpu: Send + Sync {
     /// potential soft lockups when being resumed.
     ///
     fn notify_guest_clock_paused(&self) -> Result<()>;
+    #[cfg(all(feature = "kvm", target_arch = "x86_64"))]
+    ///
+    /// Sets debug registers to set hardware breakpoints and/or enable single step.
+    ///
+    fn set_guest_debug(&self, addrs: &[GuestAddress], singlestep: bool) -> Result<()>;
     ///
     /// Sets the type of CPU to be exposed to the guest and optional features.
     ///
@@ -418,11 +466,11 @@ pub trait Vcpu: Send + Sync {
     /// Triggers the running of the current virtual CPU returning an exit reason.
     ///
     fn run(&self) -> std::result::Result<VmExit, HypervisorCpuError>;
-    #[cfg(all(feature = "mshv", target_arch = "x86_64"))]
+    #[cfg(target_arch = "x86_64")]
     ///
     /// Translate guest virtual address to guest physical address
     ///
-    fn translate_gva(&self, gva: u64, flags: u64) -> Result<(u64, hv_translate_gva_result)>;
+    fn translate_gva(&self, gva: u64, flags: u64) -> Result<(u64, u32)>;
     ///
     /// Initialize TDX support on the vCPU
     ///
@@ -433,4 +481,19 @@ pub trait Vcpu: Send + Sync {
     /// Return suspend registers(explicit and intercept suspend registers)
     ///
     fn get_suspend_regs(&self) -> Result<SuspendRegisters>;
+    #[cfg(feature = "kvm")]
+    ///
+    /// Set the "immediate_exit" state
+    ///
+    fn set_immediate_exit(&self, exit: bool);
+    #[cfg(feature = "tdx")]
+    ///
+    /// Returns the details about TDX exit reason
+    ///
+    fn get_tdx_exit_details(&mut self) -> Result<TdxExitDetails>;
+    #[cfg(feature = "tdx")]
+    ///
+    /// Set the status code for TDX exit
+    ///
+    fn set_tdx_status(&mut self, status: TdxExitStatus);
 }

hypervisor/src/kvm/aarch64/mod.rs

@@ -26,23 +26,33 @@ pub use kvm_bindings::{
 use serde_derive::{Deserialize, Serialize};
 pub use {kvm_ioctls::Cap, kvm_ioctls::Kvm};
 
-// Following are macros that help with getting the ID of a aarch64 core register.
-// The core register are represented by the user_pt_regs structure. Look for it in
-// arch/arm64/include/uapi/asm/ptrace.h.
-
 // This macro gets the offset of a structure (i.e `str`) member (i.e `field`) without having
 // an instance of that structure.
 #[macro_export]
 macro_rules! offset__of {
     ($str:ty, $($field:ident)+) => ({
         let tmp: std::mem::MaybeUninit<$str> = std::mem::MaybeUninit::uninit();
-        let tmp = unsafe { tmp.assume_init() };
+        let base = tmp.as_ptr();
-        let base = &tmp as *const _ as usize;
-        let member =  &tmp.$($field)* as *const _ as usize;
 
-        member - base
+        // Avoid warnings when nesting `unsafe` blocks.
+        #[allow(unused_unsafe)]
+        // SAFETY: The pointer is valid and aligned, just not initialised. Using `addr_of` ensures
+        // that we don't actually read from `base` (which would be UB) nor create an intermediate
+        // reference.
+        let member = unsafe { core::ptr::addr_of!((*base).$($field)*) } as *const u8;
+
+        // Avoid warnings when nesting `unsafe` blocks.
+        #[allow(unused_unsafe)]
+        // SAFETY: The two pointers are within the same allocated object `tmp`. All requirements
+        // from offset_from are upheld.
+        unsafe { member.offset_from(base as *const u8) as usize }
     });
 }
+
+// Following are macros that help with getting the ID of a aarch64 core register.
+// The core register are represented by the user_pt_regs structure. Look for it in
+// arch/arm64/include/uapi/asm/ptrace.h.
+
 // Get the ID of a core register
 #[macro_export]
 macro_rules! arm64_core_reg_id {
@@ -115,9 +125,13 @@ pub fn is_system_register(regid: u64) -> bool {
     }
 
     let size = regid & KVM_REG_SIZE_MASK;
-    if size != KVM_REG_SIZE_U32 && size != KVM_REG_SIZE_U64 {
+
-        panic!("Unexpected register size for system register {}", size);
+    assert!(
-    }
+        !(size != KVM_REG_SIZE_U32 && size != KVM_REG_SIZE_U64),
+        "Unexpected register size for system register {}",
+        size
+    );
+
     true
 }
 

hypervisor/src/kvm/mod.rs

@@ -32,8 +32,6 @@ use std::result;
 #[cfg(target_arch = "x86_64")]
 use std::sync::atomic::{AtomicBool, Ordering};
 use std::sync::{Arc, RwLock};
-#[cfg(target_arch = "x86_64")]
-use vm_memory::Address;
 use vmm_sys_util::eventfd::EventFd;
 // x86_64 dependencies
 #[cfg(target_arch = "x86_64")]
@@ -44,12 +42,11 @@ use crate::arch::x86::NUM_IOAPIC_PINS;
 use aarch64::{RegList, Register, StandardRegisters};
 #[cfg(target_arch = "x86_64")]
 use kvm_bindings::{
-    kvm_enable_cap, kvm_msr_entry, MsrList, KVM_CAP_HYPERV_SYNIC, KVM_CAP_SPLIT_IRQCHIP,
+    kvm_enable_cap, kvm_guest_debug, kvm_msr_entry, MsrList, KVM_CAP_HYPERV_SYNIC,
+    KVM_CAP_SPLIT_IRQCHIP, KVM_GUESTDBG_ENABLE, KVM_GUESTDBG_SINGLESTEP, KVM_GUESTDBG_USE_HW_BP,
 };
 #[cfg(target_arch = "x86_64")]
-use x86_64::{
+use x86_64::{check_required_kvm_extensions, FpuState, SpecialRegisters, StandardRegisters};
-    check_required_kvm_extensions, FpuState, SpecialRegisters, StandardRegisters, KVM_TSS_ADDRESS,
-};
 #[cfg(target_arch = "x86_64")]
 pub use x86_64::{
     CpuId, CpuIdEntry, ExtendedControlRegisters, LapicState, MsrEntries, VcpuKvmState as CpuState,
@@ -85,7 +82,7 @@ pub use {
     kvm_bindings::kvm_clock_data as ClockData, kvm_bindings::kvm_create_device as CreateDevice,
     kvm_bindings::kvm_device_attr as DeviceAttr,
     kvm_bindings::kvm_irq_routing_entry as IrqRoutingEntry, kvm_bindings::kvm_mp_state as MpState,
-    kvm_bindings::kvm_userspace_memory_region as MemoryRegion,
+    kvm_bindings::kvm_run, kvm_bindings::kvm_userspace_memory_region as MemoryRegion,
     kvm_bindings::kvm_vcpu_events as VcpuEvents, kvm_ioctls::DeviceFd, kvm_ioctls::IoEventAddress,
     kvm_ioctls::VcpuExit,
 };
@@ -93,6 +90,17 @@ pub use {
 #[cfg(target_arch = "x86_64")]
 const KVM_CAP_SGX_ATTRIBUTE: u32 = 196;
 
+#[cfg(feature = "tdx")]
+const KVM_EXIT_TDX: u32 = 35;
+#[cfg(feature = "tdx")]
+const TDG_VP_VMCALL_GET_QUOTE: u64 = 0x10002;
+#[cfg(feature = "tdx")]
+const TDG_VP_VMCALL_SETUP_EVENT_NOTIFY_INTERRUPT: u64 = 0x10004;
+#[cfg(feature = "tdx")]
+const TDG_VP_VMCALL_SUCCESS: u64 = 0;
+#[cfg(feature = "tdx")]
+const TDG_VP_VMCALL_INVALID_OPERAND: u64 = 0x8000000000000000;
+
 #[cfg(feature = "tdx")]
 ioctl_iowr_nr!(KVM_MEMORY_ENCRYPT_OP, KVMIO, 0xba, std::os::raw::c_ulong);
 
@@ -107,6 +115,18 @@ enum TdxCommand {
     Finalize,
 }
 
+#[cfg(feature = "tdx")]
+pub enum TdxExitDetails {
+    GetQuote,
+    SetupEventNotifyInterrupt,
+}
+
+#[cfg(feature = "tdx")]
+pub enum TdxExitStatus {
+    Success,
+    InvalidOperand,
+}
+
 #[derive(Clone, Copy, Debug, PartialEq, Deserialize, Serialize)]
 pub struct KvmVmState {}
 
@@ -139,6 +159,15 @@ pub struct KvmVm {
 /// vm.set/get().unwrap()
 ///
 impl vm::Vm for KvmVm {
+    #[cfg(target_arch = "x86_64")]
+    ///
+    /// Sets the address of the one-page region in the VM's address space.
+    ///
+    fn set_identity_map_address(&self, address: u64) -> vm::Result<()> {
+        self.fd
+            .set_identity_map_address(address)
+            .map_err(|e| vm::HypervisorVmError::SetIdentityMapAddress(e.into()))
+    }
     #[cfg(target_arch = "x86_64")]
     ///
     /// Sets the address of the three-page region in the VM's address space.
@@ -238,6 +267,9 @@ impl vm::Vm for KvmVm {
         irq_routing[0].nr = entries.len() as u32;
         irq_routing[0].flags = 0;
 
+        // SAFETY: irq_routing initialized with entries.len() and now it is being turned into
+        // entries_slice with entries.len() again. It is guaranteed to be large enough to hold
+        // everything from entries.
         unsafe {
             let entries_slice: &mut [kvm_irq_routing_entry] =
                 irq_routing[0].entries.as_mut_slice(entries.len());
@@ -302,7 +334,7 @@ impl vm::Vm for KvmVm {
             region.flags = 0;
         }
 
-        // Safe because guest regions are guaranteed not to overlap.
+        // SAFETY: Safe because guest regions are guaranteed not to overlap.
         unsafe {
             self.fd
                 .set_user_memory_region(region)
@@ -320,7 +352,7 @@ impl vm::Vm for KvmVm {
 
         // Setting the size to 0 means "remove"
         region.memory_size = 0;
-        // Safe because guest regions are guaranteed not to overlap.
+        // SAFETY: Safe because guest regions are guaranteed not to overlap.
         unsafe {
             self.fd
                 .set_user_memory_region(region)
@@ -350,10 +382,6 @@ impl vm::Vm for KvmVm {
     }
     #[cfg(target_arch = "x86_64")]
     fn enable_split_irq(&self) -> vm::Result<()> {
-        // Set TSS
-        self.fd
-            .set_tss_address(KVM_TSS_ADDRESS.raw_value() as usize)
-            .map_err(|e| vm::HypervisorVmError::EnableSplitIrq(e.into()))?;
         // Create split irqchip
         // Only the local APIC is emulated in kernel, both PICs and IOAPIC
         // are not.
@@ -434,7 +462,7 @@ impl vm::Vm for KvmVm {
                 userspace_addr: s.userspace_addr,
                 flags: KVM_MEM_LOG_DIRTY_PAGES,
             };
-            // Safe because guest regions are guaranteed not to overlap.
+            // SAFETY: Safe because guest regions are guaranteed not to overlap.
             unsafe {
                 self.fd
                     .set_user_memory_region(region)
@@ -458,7 +486,7 @@ impl vm::Vm for KvmVm {
                 userspace_addr: s.userspace_addr,
                 flags: 0,
             };
-            // Safe because guest regions are guaranteed not to overlap.
+            // SAFETY: Safe because guest regions are guaranteed not to overlap.
             unsafe {
                 self.fd
                     .set_user_memory_region(region)
@@ -574,6 +602,7 @@ fn tdx_command(
         metadata,
         data,
     };
+    // SAFETY: FFI call. All input parameters are valid.
     let ret = unsafe {
         ioctl_with_val(
             fd,
@@ -772,6 +801,27 @@ impl cpu::Vcpu for KvmVcpu {
             .set_regs(regs)
             .map_err(|e| cpu::HypervisorCpuError::SetStandardRegs(e.into()))
     }
+
+    #[cfg(target_arch = "aarch64")]
+    ///
+    /// Set attribute for vcpu.
+    ///
+    fn set_vcpu_attr(&self, attr: &DeviceAttr) -> cpu::Result<()> {
+        self.fd
+            .set_device_attr(attr)
+            .map_err(|e| cpu::HypervisorCpuError::SetVcpuAttribute(e.into()))
+    }
+
+    #[cfg(target_arch = "aarch64")]
+    ///
+    /// Check if vcpu has a certain attribute.
+    ///
+    fn has_vcpu_attr(&self, attr: &DeviceAttr) -> cpu::Result<()> {
+        self.fd
+            .has_device_attr(attr)
+            .map_err(|e| cpu::HypervisorCpuError::HasVcpuAttribute(e.into()))
+    }
+
     #[cfg(target_arch = "x86_64")]
     ///
     /// Returns the vCPU special registers.
@@ -932,6 +982,24 @@ impl cpu::Vcpu for KvmVcpu {
             .set_xcrs(xcrs)
             .map_err(|e| cpu::HypervisorCpuError::SetXcsr(e.into()))
     }
+    #[cfg(target_arch = "x86_64")]
+    ///
+    /// Translates guest virtual address to guest physical address using the `KVM_TRANSLATE` ioctl.
+    ///
+    fn translate_gva(&self, gva: u64, _flags: u64) -> cpu::Result<(u64, u32)> {
+        let tr = self
+            .fd
+            .translate_gva(gva)
+            .map_err(|e| cpu::HypervisorCpuError::TranslateVirtualAddress(e.into()))?;
+        // tr.valid is set if the GVA is mapped to valid GPA.
+        match tr.valid {
+            0 => Err(cpu::HypervisorCpuError::TranslateVirtualAddress(anyhow!(
+                "Invalid GVA: {:#x}",
+                gva
+            ))),
+            _ => Ok((tr.physical_address, 0)),
+        }
+    }
     ///
     /// Triggers the running of the current virtual CPU returning an exit reason.
     ///
@@ -1004,6 +1072,9 @@ impl cpu::Vcpu for KvmVcpu {
                     Ok(cpu::VmExit::MmioWrite(addr, data))
                 }
                 VcpuExit::Hyperv => Ok(cpu::VmExit::Hyperv),
+                #[cfg(feature = "tdx")]
+                VcpuExit::Unsupported(KVM_EXIT_TDX) => Ok(cpu::VmExit::Tdx),
+                VcpuExit::Debug(_) => Ok(cpu::VmExit::Debug),
 
                 r => Err(cpu::HypervisorCpuError::RunVcpu(anyhow!(
                     "Unexpected exit reason on vcpu run: {:?}",
@@ -1046,9 +1117,55 @@ impl cpu::Vcpu for KvmVcpu {
     /// potential soft lockups when being resumed.
     ///
     fn notify_guest_clock_paused(&self) -> cpu::Result<()> {
+        if let Err(e) = self.fd.kvmclock_ctrl() {
+            // Linux kernel returns -EINVAL if the PV clock isn't yet initialised
+            // which could be because we're still in firmware or the guest doesn't
+            // use KVM clock.
+            if e.errno() != libc::EINVAL {
+                return Err(cpu::HypervisorCpuError::NotifyGuestClockPaused(e.into()));
+            }
+        }
+
+        Ok(())
+    }
+    #[cfg(target_arch = "x86_64")]
+    ///
+    /// Sets debug registers to set hardware breakpoints and/or enable single step.
+    ///
+    fn set_guest_debug(
+        &self,
+        addrs: &[vm_memory::GuestAddress],
+        singlestep: bool,
+    ) -> cpu::Result<()> {
+        if addrs.len() > 4 {
+            return Err(cpu::HypervisorCpuError::SetDebugRegs(anyhow!(
+                "Support 4 breakpoints at most but {} addresses are passed",
+                addrs.len()
+            )));
+        }
+
+        let mut dbg = kvm_guest_debug {
+            control: KVM_GUESTDBG_ENABLE | KVM_GUESTDBG_USE_HW_BP,
+            ..Default::default()
+        };
+        if singlestep {
+            dbg.control |= KVM_GUESTDBG_SINGLESTEP;
+        }
+
+        // Set bits 9 and 10.
+        // bit 9: GE (global exact breakpoint enable) flag.
+        // bit 10: always 1.
+        dbg.arch.debugreg[7] = 0x0600;
+
+        for (i, addr) in addrs.iter().enumerate() {
+            dbg.arch.debugreg[i] = addr.0;
+            // Set global breakpoint enable flag
+            dbg.arch.debugreg[7] |= 2 << (i * 2);
+        }
+
         self.fd
-            .kvmclock_ctrl()
+            .set_guest_debug(&dbg)
-            .map_err(|e| cpu::HypervisorCpuError::NotifyGuestClockPaused(e.into()))
+            .map_err(|e| cpu::HypervisorCpuError::SetDebugRegs(e.into()))
     }
     #[cfg(any(target_arch = "arm", target_arch = "aarch64"))]
     fn vcpu_init(&self, kvi: &VcpuInit) -> cpu::Result<()> {
@@ -1562,6 +1679,50 @@ impl cpu::Vcpu for KvmVcpu {
         tdx_command(&self.fd.as_raw_fd(), TdxCommand::InitVcpu, 0, hob_address)
             .map_err(cpu::HypervisorCpuError::InitializeTdx)
     }
+
+    ///
+    /// Set the "immediate_exit" state
+    ///
+    fn set_immediate_exit(&self, exit: bool) {
+        self.fd.set_kvm_immediate_exit(exit.into());
+    }
+
+    ///
+    /// Returns the details about TDX exit reason
+    ///
+    #[cfg(feature = "tdx")]
+    fn get_tdx_exit_details(&mut self) -> cpu::Result<TdxExitDetails> {
+        let kvm_run = self.fd.get_kvm_run();
+        let tdx_vmcall = unsafe { &mut kvm_run.__bindgen_anon_1.tdx.u.vmcall };
+
+        tdx_vmcall.status_code = TDG_VP_VMCALL_INVALID_OPERAND;
+
+        if tdx_vmcall.type_ != 0 {
+            return Err(cpu::HypervisorCpuError::UnknownTdxVmCall);
+        }
+
+        match tdx_vmcall.subfunction {
+            TDG_VP_VMCALL_GET_QUOTE => Ok(TdxExitDetails::GetQuote),
+            TDG_VP_VMCALL_SETUP_EVENT_NOTIFY_INTERRUPT => {
+                Ok(TdxExitDetails::SetupEventNotifyInterrupt)
+            }
+            _ => Err(cpu::HypervisorCpuError::UnknownTdxVmCall),
+        }
+    }
+
+    ///
+    /// Set the status code for TDX exit
+    ///
+    #[cfg(feature = "tdx")]
+    fn set_tdx_status(&mut self, status: TdxExitStatus) {
+        let kvm_run = self.fd.get_kvm_run();
+        let tdx_vmcall = unsafe { &mut kvm_run.__bindgen_anon_1.tdx.u.vmcall };
+
+        tdx_vmcall.status_code = match status {
+            TdxExitStatus::Success => TDG_VP_VMCALL_SUCCESS,
+            TdxExitStatus::InvalidOperand => TDG_VP_VMCALL_INVALID_OPERAND,
+        };
+    }
 }
 
 /// Device struct for KVM

hypervisor/src/kvm/x86_64/mod.rs

@@ -11,7 +11,6 @@
 use crate::arch::x86::{msr_index, SegmentRegisterOps, MTRR_ENABLE, MTRR_MEM_TYPE_WB};
 use crate::kvm::{Cap, Kvm, KvmError, KvmResult};
 use serde_derive::{Deserialize, Serialize};
-use vm_memory::GuestAddress;
 
 ///
 /// Export generically-named wrappers of kvm-bindings for Unix-based platforms
@@ -92,8 +91,6 @@ impl SegmentRegisterOps for SegmentRegister {
     }
 }
 
-pub const KVM_TSS_ADDRESS: GuestAddress = GuestAddress(0xfffb_d000);
-
 pub fn boot_msr_entries() -> MsrEntries {
     MsrEntries::from_entries(&[
         msr!(msr_index::MSR_IA32_SYSENTER_CS),
@@ -127,6 +124,15 @@ pub fn check_required_kvm_extensions(kvm: &Kvm) -> KvmResult<()> {
     if !kvm.check_extension(Cap::SplitIrqchip) {
         return Err(KvmError::CapabilityMissing(Cap::SplitIrqchip));
     }
+    if !kvm.check_extension(Cap::SetIdentityMapAddr) {
+        return Err(KvmError::CapabilityMissing(Cap::SetIdentityMapAddr));
+    }
+    if !kvm.check_extension(Cap::SetTssAddr) {
+        return Err(KvmError::CapabilityMissing(Cap::SetTssAddr));
+    }
+    if !kvm.check_extension(Cap::ImmediateExit) {
+        return Err(KvmError::CapabilityMissing(Cap::ImmediateExit));
+    }
     Ok(())
 }
 #[derive(Clone, Serialize, Deserialize)]

hypervisor/src/mshv/mod.rs

@@ -302,6 +302,7 @@ impl cpu::Vcpu for MshvVcpu {
                 hv_message_type_HVMSG_X64_IO_PORT_INTERCEPT => {
                     let info = x.to_ioport_info().unwrap();
                     let access_info = info.access_info;
+                    // SAFETY: access_info is valid, otherwise we won't be here
                     let len = unsafe { access_info.__bindgen_anon_1.access_size() } as usize;
                     let is_write = info.header.intercept_access_type == 1;
                     let port = info.port_number;
@@ -342,12 +343,15 @@ impl cpu::Vcpu for MshvVcpu {
                         _ => {}
                     }
 
-                    if unsafe { access_info.__bindgen_anon_1.string_op() } == 1 {
+                    // SAFETY: access_info is valid, otherwise we won't be here
-                        panic!("String IN/OUT not supported");
+                    assert!(
-                    }
+                        (unsafe { access_info.__bindgen_anon_1.string_op() } != 1),
-                    if unsafe { access_info.__bindgen_anon_1.rep_prefix() } == 1 {
+                        "String IN/OUT not supported"
-                        panic!("Rep IN/OUT not supported");
+                    );
-                    }
+                    assert!(
+                        (unsafe { access_info.__bindgen_anon_1.rep_prefix() } != 1),
+                        "Rep IN/OUT not supported"
+                    );
 
                     if is_write {
                         let data = (info.rax as u32).to_le_bytes();
@@ -506,6 +510,13 @@ impl cpu::Vcpu for MshvVcpu {
         self.set_xcrs(&state.xcrs)?;
         self.set_lapic(&state.lapic)?;
         self.set_xsave(&state.xsave)?;
+        // These registers are global and needed to be set only for first VCPU
+        // as Microsoft Hypervisor allows setting this regsier for only one VCPU
+        if self.vp_index == 0 {
+            self.fd
+                .set_misc_regs(&state.misc)
+                .map_err(|e| cpu::HypervisorCpuError::SetMiscRegs(e.into()))?
+        }
         self.fd
             .set_debug_regs(&state.dbg)
             .map_err(|e| cpu::HypervisorCpuError::SetDebugRegs(e.into()))?;
@@ -524,10 +535,15 @@ impl cpu::Vcpu for MshvVcpu {
         self.get_msrs(&mut msrs)?;
         let lapic = self.get_lapic()?;
         let xsave = self.get_xsave()?;
+        let misc = self
+            .fd
+            .get_misc_regs()
+            .map_err(|e| cpu::HypervisorCpuError::GetMiscRegs(e.into()))?;
         let dbg = self
             .fd
             .get_debug_regs()
             .map_err(|e| cpu::HypervisorCpuError::GetDebugRegs(e.into()))?;
+
         Ok(CpuState {
             msrs,
             vcpu_events,
@@ -538,19 +554,24 @@ impl cpu::Vcpu for MshvVcpu {
             lapic,
             dbg,
             xsave,
+            misc,
         })
     }
     #[cfg(target_arch = "x86_64")]
     ///
     /// Translate guest virtual address to guest physical address
     ///
-    fn translate_gva(&self, gva: u64, flags: u64) -> cpu::Result<(u64, hv_translate_gva_result)> {
+    fn translate_gva(&self, gva: u64, flags: u64) -> cpu::Result<(u64, u32)> {
         let r = self
             .fd
             .translate_gva(gva, flags)
             .map_err(|e| cpu::HypervisorCpuError::TranslateVirtualAddress(e.into()))?;
 
-        Ok(r)
+        let gpa = r.0;
+        // SAFETY: r is valid, otherwise this function will have returned
+        let result_code = unsafe { r.1.__bindgen_anon_1.result_code };
+
+        Ok((gpa, result_code))
     }
     #[cfg(target_arch = "x86_64")]
     ///
@@ -609,14 +630,13 @@ impl<'a> MshvEmulatorContext<'a> {
         // TODO: More fine-grained control for the flags
         let flags = HV_TRANSLATE_GVA_VALIDATE_READ | HV_TRANSLATE_GVA_VALIDATE_WRITE;
 
-        let r = self
+        let (gpa, result_code) = self
             .vcpu
             .translate_gva(gva, flags.into())
             .map_err(|e| PlatformError::TranslateVirtualAddress(anyhow!(e)))?;
 
-        let result_code = unsafe { r.1.__bindgen_anon_1.result_code };
         match result_code {
-            hv_translate_gva_result_code_HV_TRANSLATE_GVA_SUCCESS => Ok(r.0),
+            hv_translate_gva_result_code_HV_TRANSLATE_GVA_SUCCESS => Ok(gpa),
             _ => Err(PlatformError::TranslateVirtualAddress(anyhow!(result_code))),
         }
     }
@@ -746,6 +766,13 @@ fn hv_state_init() -> Arc<RwLock<HvState>> {
 /// vm.set/get().unwrap()
 ///
 impl vm::Vm for MshvVm {
+    #[cfg(target_arch = "x86_64")]
+    ///
+    /// Sets the address of the one-page region in the VM's address space.
+    ///
+    fn set_identity_map_address(&self, _address: u64) -> vm::Result<()> {
+        Ok(())
+    }
     #[cfg(target_arch = "x86_64")]
     ///
     /// Sets the address of the three-page region in the VM's address space.
@@ -935,6 +962,9 @@ impl vm::Vm for MshvVm {
             vec_with_array_field::<mshv_msi_routing, mshv_msi_routing_entry>(entries.len());
         msi_routing[0].nr = entries.len() as u32;
 
+        // SAFETY: msi_routing initialized with entries.len() and now it is being turned into
+        // entries_slice with entries.len() again. It is guaranteed to be large enough to hold
+        // everything from entries.
         unsafe {
             let entries_slice: &mut [mshv_msi_routing_entry] =
                 msi_routing[0].entries.as_mut_slice(entries.len());

hypervisor/src/mshv/x86_64/mod.rs

@@ -18,10 +18,11 @@ pub use {
     mshv_bindings::mshv_user_mem_region as MemoryRegion, mshv_bindings::msr_entry as MsrEntry,
     mshv_bindings::CpuId, mshv_bindings::DebugRegisters,
     mshv_bindings::FloatingPointUnit as FpuState, mshv_bindings::LapicState,
-    mshv_bindings::MsrList, mshv_bindings::Msrs as MsrEntries, mshv_bindings::Msrs,
+    mshv_bindings::MiscRegs as MiscRegisters, mshv_bindings::MsrList,
-    mshv_bindings::SegmentRegister, mshv_bindings::SpecialRegisters,
+    mshv_bindings::Msrs as MsrEntries, mshv_bindings::Msrs, mshv_bindings::SegmentRegister,
-    mshv_bindings::StandardRegisters, mshv_bindings::SuspendRegisters, mshv_bindings::VcpuEvents,
+    mshv_bindings::SpecialRegisters, mshv_bindings::StandardRegisters,
-    mshv_bindings::XSave as Xsave, mshv_bindings::Xcrs as ExtendedControlRegisters,
+    mshv_bindings::SuspendRegisters, mshv_bindings::VcpuEvents, mshv_bindings::XSave as Xsave,
+    mshv_bindings::Xcrs as ExtendedControlRegisters,
 };
 
 #[derive(Clone, Serialize, Deserialize)]
@@ -35,6 +36,7 @@ pub struct VcpuMshvState {
     pub lapic: LapicState,
     pub dbg: DebugRegisters,
     pub xsave: Xsave,
+    pub misc: MiscRegisters,
 }
 
 impl fmt::Display for VcpuMshvState {

hypervisor/src/vm.rs

@@ -58,6 +58,11 @@ pub enum HypervisorVmError {
     #[error("Failed to create Vcpu: {0}")]
     CreateVcpu(#[source] anyhow::Error),
     ///
+    /// Identity map address error
+    ///
+    #[error("Failed to set identity map address: {0}")]
+    SetIdentityMapAddress(#[source] anyhow::Error),
+    ///
     /// TSS address error
     ///
     #[error("Failed to set TSS address: {0}")]
@@ -217,6 +222,9 @@ pub type Result<T> = std::result::Result<T, HypervisorVmError>;
 /// This crate provides a hypervisor-agnostic interfaces for Vm
 ///
 pub trait Vm: Send + Sync {
+    #[cfg(target_arch = "x86_64")]
+    /// Sets the address of the one-page region in the VM's address space.
+    fn set_identity_map_address(&self, address: u64) -> Result<()>;
     #[cfg(target_arch = "x86_64")]
     /// Sets the address of the three-page region in the VM's address space.
     fn set_tss_address(&self, offset: usize) -> Result<()>;

net_util/Cargo.toml

@@ -7,19 +7,20 @@ edition = "2018"
 [dependencies]
 epoll = "4.3.1"
 getrandom = "0.2"
-libc = "0.2.103"
+libc = "0.2.119"
 log = "0.4.14"
 net_gen = { path = "../net_gen" }
 rate_limiter = { path = "../rate_limiter" }
-serde = "1.0.130"
+serde = "1.0.136"
 versionize = "0.1.6"
 versionize_derive = "0.1.4"
 virtio-bindings = "0.1.0"
-vm-memory = { version = "0.6.0", features = ["backend-mmap", "backend-atomic", "backend-bitmap"] }
+virtio-queue = { git = "https://github.com/rust-vmm/vm-virtio", branch = "main" }
+vm-memory = { version = "0.7.0", features = ["backend-mmap", "backend-atomic", "backend-bitmap"] }
 vm-virtio = { path = "../vm-virtio" }
 vmm-sys-util = "0.9.0"
 
 [dev-dependencies]
 lazy_static = "1.4.0"
-pnet = "0.28.0"
+pnet = "0.29.0"
-serde_json = "1.0.68"
+serde_json = "1.0.79"

net_util/src/ctrl_queue.rs

@@ -5,6 +5,7 @@
 use crate::GuestMemoryMmap;
 use crate::Tap;
 use libc::c_uint;
+use std::sync::Arc;
 use virtio_bindings::bindings::virtio_net::{
     VIRTIO_NET_CTRL_GUEST_OFFLOADS, VIRTIO_NET_CTRL_GUEST_OFFLOADS_SET, VIRTIO_NET_CTRL_MQ,
     VIRTIO_NET_CTRL_MQ_VQ_PAIRS_MAX, VIRTIO_NET_CTRL_MQ_VQ_PAIRS_MIN,
@@ -12,17 +13,26 @@ use virtio_bindings::bindings::virtio_net::{
     VIRTIO_NET_F_GUEST_ECN, VIRTIO_NET_F_GUEST_TSO4, VIRTIO_NET_F_GUEST_TSO6,
     VIRTIO_NET_F_GUEST_UFO, VIRTIO_NET_OK,
 };
-use vm_memory::{ByteValued, Bytes, GuestMemoryError};
+use virtio_queue::Queue;
-use vm_virtio::Queue;
+use vm_memory::{ByteValued, Bytes, GuestMemoryAtomic, GuestMemoryError};
+use vm_virtio::{AccessPlatform, Translatable};
 
 #[derive(Debug)]
 pub enum Error {
     /// Read queue failed.
     GuestMemory(GuestMemoryError),
-    /// No queue pairs number.
+    /// No control header descriptor
-    NoQueuePairsDescriptor,
+    NoControlHeaderDescriptor,
+    /// Missing the data descriptor in the chain.
+    NoDataDescriptor,
     /// No status descriptor
     NoStatusDescriptor,
+    /// Failed adding used index
+    QueueAddUsed(virtio_queue::Error),
+    /// Failed creating an iterator over the queue
+    QueueIterator(virtio_queue::Error),
+    /// Failed enabling notification for the queue
+    QueueEnableNotification(virtio_queue::Error),
 }
 
 type Result<T> = std::result::Result<T, Error>;
@@ -34,6 +44,7 @@ pub struct ControlHeader {
     pub cmd: u8,
 }
 
+// SAFETY: ControlHeader only contains a series of integers
 unsafe impl ByteValued for ControlHeader {}
 
 pub struct CtrlQueue {
@@ -45,22 +56,36 @@ impl CtrlQueue {
         CtrlQueue { taps }
     }
 
-    pub fn process(&mut self, mem: &GuestMemoryMmap, queue: &mut Queue) -> Result<bool> {
+    pub fn process(
+        &mut self,
+        queue: &mut Queue<GuestMemoryAtomic<GuestMemoryMmap>>,
+        access_platform: Option<&Arc<dyn AccessPlatform>>,
+    ) -> Result<bool> {
         let mut used_desc_heads = Vec::new();
-        for avail_desc in queue.iter(mem) {
+        for mut desc_chain in queue.iter().map_err(Error::QueueIterator)? {
-            let ctrl_hdr: ControlHeader =
+            let ctrl_desc = desc_chain.next().ok_or(Error::NoControlHeaderDescriptor)?;
-                mem.read_obj(avail_desc.addr).map_err(Error::GuestMemory)?;
+
-            let data_desc = avail_desc
+            let ctrl_hdr: ControlHeader = desc_chain
-                .next_descriptor()
+                .memory()
-                .ok_or(Error::NoQueuePairsDescriptor)?;
+                .read_obj(
-            let status_desc = data_desc
+                    ctrl_desc
-                .next_descriptor()
+                        .addr()
-                .ok_or(Error::NoStatusDescriptor)?;
+                        .translate(access_platform, ctrl_desc.len() as usize),
+                )
+                .map_err(Error::GuestMemory)?;
+            let data_desc = desc_chain.next().ok_or(Error::NoDataDescriptor)?;
+
+            let data_desc_addr = data_desc
+                .addr()
+                .translate(access_platform, data_desc.len() as usize);
+
+            let status_desc = desc_chain.next().ok_or(Error::NoStatusDescriptor)?;
 
             let ok = match u32::from(ctrl_hdr.class) {
                 VIRTIO_NET_CTRL_MQ => {
-                    let queue_pairs = mem
+                    let queue_pairs = desc_chain
-                        .read_obj::<u16>(data_desc.addr)
+                        .memory()
+                        .read_obj::<u16>(data_desc_addr)
                         .map_err(Error::GuestMemory)?;
                     if u32::from(ctrl_hdr.cmd) != VIRTIO_NET_CTRL_MQ_VQ_PAIRS_SET {
                         warn!("Unsupported command: {}", ctrl_hdr.cmd);
@@ -76,8 +101,9 @@ impl CtrlQueue {
                     }
                 }
                 VIRTIO_NET_CTRL_GUEST_OFFLOADS => {
-                    let features = mem
+                    let features = desc_chain
-                        .read_obj::<u64>(data_desc.addr)
+                        .memory()
+                        .read_obj::<u64>(data_desc_addr)
                         .map_err(Error::GuestMemory)?;
                     if u32::from(ctrl_hdr.cmd) != VIRTIO_NET_CTRL_GUEST_OFFLOADS_SET {
                         warn!("Unsupported command: {}", ctrl_hdr.cmd);
@@ -102,17 +128,26 @@ impl CtrlQueue {
                 }
             };
 
-            mem.write_obj(
+            desc_chain
-                if ok { VIRTIO_NET_OK } else { VIRTIO_NET_ERR } as u8,
+                .memory()
-                status_desc.addr,
+                .write_obj(
-            )
+                    if ok { VIRTIO_NET_OK } else { VIRTIO_NET_ERR } as u8,
-            .map_err(Error::GuestMemory)?;
+                    status_desc
-            used_desc_heads.push((avail_desc.index, avail_desc.len));
+                        .addr()
+                        .translate(access_platform, status_desc.len() as usize),
+                )
+                .map_err(Error::GuestMemory)?;
+            let len = ctrl_desc.len() + data_desc.len() + status_desc.len();
+            used_desc_heads.push((desc_chain.head_index(), len));
         }
 
         for (desc_index, len) in used_desc_heads.iter() {
-            queue.add_used(mem, *desc_index, *len);
+            queue
-            queue.update_avail_event(mem);
+                .add_used(*desc_index, *len)
+                .map_err(Error::QueueAddUsed)?;
+            queue
+                .enable_notification()
+                .map_err(Error::QueueEnableNotification)?;
         }
 
         Ok(!used_desc_heads.is_empty())

net_util/src/lib.rs

@@ -59,7 +59,7 @@ pub struct VirtioNetConfig {
     pub duplex: u8,
 }
 
-// Safe because it only has data and has no implicit padding.
+// SAFETY: it only has data and has no implicit padding.
 unsafe impl ByteValued for VirtioNetConfig {}
 
 /// Create a sockaddr_in from an IPv4 address, and expose it as
@@ -77,7 +77,7 @@ fn create_sockaddr(ip_addr: net::Ipv4Addr) -> net_gen::sockaddr {
     unsafe { mem::transmute(addr_in) }
 }
 
-fn create_socket() -> Result<net::UdpSocket> {
+fn create_inet_socket() -> Result<net::UdpSocket> {
     // This is safe since we check the return value.
     let sock = unsafe { libc::socket(libc::AF_INET, libc::SOCK_DGRAM, 0) };
     if sock < 0 {
@@ -88,6 +88,17 @@ fn create_socket() -> Result<net::UdpSocket> {
     Ok(unsafe { net::UdpSocket::from_raw_fd(sock) })
 }
 
+fn create_unix_socket() -> Result<net::UdpSocket> {
+    // This is safe since we check the return value.
+    let sock = unsafe { libc::socket(libc::AF_UNIX, libc::SOCK_DGRAM, 0) };
+    if sock < 0 {
+        return Err(Error::CreateSocket(IoError::last_os_error()));
+    }
+
+    // This is safe; nothing else will use or hold onto the raw sock fd.
+    Ok(unsafe { net::UdpSocket::from_raw_fd(sock) })
+}
+
 fn vnet_hdr_len() -> usize {
     std::mem::size_of::<virtio_net_hdr_v1>()
 }
@@ -121,7 +132,7 @@ pub fn unregister_listener(
 }
 
 pub fn build_net_config_space(
-    mut config: &mut VirtioNetConfig,
+    config: &mut VirtioNetConfig,
     mac: MacAddr,
     num_queues: usize,
     avail_features: &mut u64,
@@ -129,7 +140,7 @@ pub fn build_net_config_space(
     config.mac.copy_from_slice(mac.get_bytes());
     *avail_features |= 1 << VIRTIO_NET_F_MAC;
 
-    build_net_config_space_with_mq(&mut config, num_queues, avail_features);
+    build_net_config_space_with_mq(config, num_queues, avail_features);
 }
 
 pub fn build_net_config_space_with_mq(

net_util/src/mac.rs

@@ -161,7 +161,7 @@ mod tests {
 
         let mac = MacAddr::parse_str("12:34:56:78:9a:BC").unwrap();
 
-        println!("parsed MAC address: {}", mac.to_string());
+        println!("parsed MAC address: {}", mac);
 
         let bytes = mac.get_bytes();
         assert_eq!(bytes, [0x12u8, 0x34, 0x56, 0x78, 0x9a, 0xbc]);

net_util/src/queue_pair.rs

@@ -10,8 +10,9 @@ use std::num::Wrapping;
 use std::os::unix::io::{AsRawFd, RawFd};
 use std::sync::atomic::{AtomicU64, Ordering};
 use std::sync::Arc;
-use vm_memory::{Bytes, GuestAddressSpace, GuestMemory, GuestMemoryAtomic};
+use virtio_queue::Queue;
-use vm_virtio::Queue;
+use vm_memory::{Bytes, GuestMemory, GuestMemoryAtomic};
+use vm_virtio::{AccessPlatform, Translatable};
 
 #[derive(Clone)]
 pub struct TxVirtio {
@@ -35,78 +36,103 @@ impl TxVirtio {
 
     pub fn process_desc_chain(
         &mut self,
-        mem: &GuestMemoryMmap,
         tap: &mut Tap,
-        queue: &mut Queue,
+        queue: &mut Queue<GuestMemoryAtomic<GuestMemoryMmap>>,
         rate_limiter: &mut Option<RateLimiter>,
+        access_platform: Option<&Arc<dyn AccessPlatform>>,
     ) -> Result<bool, NetQueuePairError> {
         let mut retry_write = false;
         let mut rate_limit_reached = false;
-        while let Some(avail_desc) = queue.iter(mem).next() {
+
-            if rate_limit_reached {
+        loop {
-                queue.go_to_previous_position();
+            let used_desc_head: (u16, u32);
+            let mut avail_iter = queue
+                .iter()
+                .map_err(NetQueuePairError::QueueIteratorFailed)?;
+
+            if let Some(mut desc_chain) = avail_iter.next() {
+                if rate_limit_reached {
+                    avail_iter.go_to_previous_position();
+                    break;
+                }
+
+                let mut next_desc = desc_chain.next();
+
+                let mut iovecs = Vec::new();
+                while let Some(desc) = next_desc {
+                    let desc_addr = desc.addr().translate(access_platform, desc.len() as usize);
+                    if !desc.is_write_only() && desc.len() > 0 {
+                        let buf = desc_chain
+                            .memory()
+                            .get_slice(desc_addr, desc.len() as usize)
+                            .map_err(NetQueuePairError::GuestMemory)?
+                            .as_ptr();
+                        let iovec = libc::iovec {
+                            iov_base: buf as *mut libc::c_void,
+                            iov_len: desc.len() as libc::size_t,
+                        };
+                        iovecs.push(iovec);
+                    } else {
+                        error!(
+                            "Invalid descriptor chain: address = 0x{:x} length = {} write_only = {}",
+                            desc_addr.0,
+                            desc.len(),
+                            desc.is_write_only()
+                        );
+                        return Err(NetQueuePairError::DescriptorChainInvalid);
+                    }
+                    next_desc = desc_chain.next();
+                }
+
+                let len = if !iovecs.is_empty() {
+                    let result = unsafe {
+                        libc::writev(
+                            tap.as_raw_fd() as libc::c_int,
+                            iovecs.as_ptr() as *const libc::iovec,
+                            iovecs.len() as libc::c_int,
+                        )
+                    };
+
+                    if result < 0 {
+                        let e = std::io::Error::last_os_error();
+
+                        /* EAGAIN */
+                        if e.kind() == std::io::ErrorKind::WouldBlock {
+                            avail_iter.go_to_previous_position();
+                            retry_write = true;
+                            break;
+                        }
+                        error!("net: tx: failed writing to tap: {}", e);
+                        return Err(NetQueuePairError::WriteTap(e));
+                    }
+
+                    self.counter_bytes += Wrapping(result as u64 - vnet_hdr_len() as u64);
+                    self.counter_frames += Wrapping(1);
+
+                    result as u32
+                } else {
+                    0
+                };
+
+                used_desc_head = (desc_chain.head_index(), len);
+
+                // For the sake of simplicity (similar to the RX rate limiting), we always
+                // let the 'last' descriptor chain go-through even if it was over the rate
+                // limit, and simply stop processing oncoming `avail_desc` if any.
+                if let Some(rate_limiter) = rate_limiter {
+                    rate_limit_reached = !rate_limiter.consume(1, TokenType::Ops)
+                        || !rate_limiter.consume(len as u64, TokenType::Bytes);
+                }
+            } else {
                 break;
             }
 
-            let head_index = avail_desc.index;
+            queue
-            let mut next_desc = Some(avail_desc);
+                .add_used(used_desc_head.0, used_desc_head.1)
-
+                .map_err(NetQueuePairError::QueueAddUsed)?;
-            let mut iovecs = Vec::new();
+            queue
-            while let Some(desc) = next_desc {
+                .enable_notification()
-                if !desc.is_write_only() && desc.len > 0 {
+                .map_err(NetQueuePairError::QueueEnableNotification)?;
-                    let buf = mem
-                        .get_slice(desc.addr, desc.len as usize)
-                        .map_err(NetQueuePairError::GuestMemory)?
-                        .as_ptr();
-                    let iovec = libc::iovec {
-                        iov_base: buf as *mut libc::c_void,
-                        iov_len: desc.len as libc::size_t,
-                    };
-                    iovecs.push(iovec);
-                }
-                next_desc = desc.next_descriptor();
-            }
-
-            let len = if !iovecs.is_empty() {
-                let result = unsafe {
-                    libc::writev(
-                        tap.as_raw_fd() as libc::c_int,
-                        iovecs.as_ptr() as *const libc::iovec,
-                        iovecs.len() as libc::c_int,
-                    )
-                };
-
-                if result < 0 {
-                    let e = std::io::Error::last_os_error();
-
-                    /* EAGAIN */
-                    if e.kind() == std::io::ErrorKind::WouldBlock {
-                        queue.go_to_previous_position();
-                        retry_write = true;
-                        break;
-                    }
-                    error!("net: tx: failed writing to tap: {}", e);
-                    return Err(NetQueuePairError::WriteTap(e));
-                }
-
-                self.counter_bytes += Wrapping(result as u64 - vnet_hdr_len() as u64);
-                self.counter_frames += Wrapping(1);
-
-                result as u32
-            } else {
-                0
-            };
-
-            queue.add_used(mem, head_index, 0);
-            queue.update_avail_event(mem);
-
-            // For the sake of simplicity (similar to the RX rate limiting), we always
-            // let the 'last' descriptor chain go-through even if it was over the rate
-            // limit, and simply stop processing oncoming `avail_desc` if any.
-            if let Some(rate_limiter) = rate_limiter {
-                rate_limit_reached = !rate_limiter.consume(1, TokenType::Ops)
-                    || !rate_limiter.consume(len as u64, TokenType::Bytes);
-            }
         }
 
         Ok(retry_write)
@@ -135,87 +161,123 @@ impl RxVirtio {
 
     pub fn process_desc_chain(
         &mut self,
-        mem: &GuestMemoryMmap,
         tap: &mut Tap,
-        queue: &mut Queue,
+        queue: &mut Queue<GuestMemoryAtomic<GuestMemoryMmap>>,
         rate_limiter: &mut Option<RateLimiter>,
+        access_platform: Option<&Arc<dyn AccessPlatform>>,
     ) -> Result<bool, NetQueuePairError> {
         let mut exhausted_descs = true;
         let mut rate_limit_reached = false;
 
-        while let Some(avail_desc) = queue.iter(mem).next() {
+        loop {
-            if rate_limit_reached {
+            let used_desc_head: (u16, u32);
-                exhausted_descs = false;
+            let mut avail_iter = queue
-                queue.go_to_previous_position();
+                .iter()
+                .map_err(NetQueuePairError::QueueIteratorFailed)?;
+
+            if let Some(mut desc_chain) = avail_iter.next() {
+                if rate_limit_reached {
+                    exhausted_descs = false;
+                    avail_iter.go_to_previous_position();
+                    break;
+                }
+
+                let desc = desc_chain
+                    .next()
+                    .ok_or(NetQueuePairError::DescriptorChainTooShort)?;
+
+                let num_buffers_addr = desc_chain
+                    .memory()
+                    .checked_offset(
+                        desc.addr().translate(access_platform, desc.len() as usize),
+                        10,
+                    )
+                    .unwrap();
+                let mut next_desc = Some(desc);
+
+                let mut iovecs = Vec::new();
+                while let Some(desc) = next_desc {
+                    let desc_addr = desc.addr().translate(access_platform, desc.len() as usize);
+                    if desc.is_write_only() && desc.len() > 0 {
+                        let buf = desc_chain
+                            .memory()
+                            .get_slice(desc_addr, desc.len() as usize)
+                            .map_err(NetQueuePairError::GuestMemory)?
+                            .as_ptr();
+                        let iovec = libc::iovec {
+                            iov_base: buf as *mut libc::c_void,
+                            iov_len: desc.len() as libc::size_t,
+                        };
+                        iovecs.push(iovec);
+                    } else {
+                        error!(
+                            "Invalid descriptor chain: address = 0x{:x} length = {} write_only = {}",
+                            desc_addr.0,
+                            desc.len(),
+                            desc.is_write_only()
+                        );
+                        return Err(NetQueuePairError::DescriptorChainInvalid);
+                    }
+                    next_desc = desc_chain.next();
+                }
+
+                let len = if !iovecs.is_empty() {
+                    let result = unsafe {
+                        libc::readv(
+                            tap.as_raw_fd() as libc::c_int,
+                            iovecs.as_ptr() as *const libc::iovec,
+                            iovecs.len() as libc::c_int,
+                        )
+                    };
+                    if result < 0 {
+                        let e = std::io::Error::last_os_error();
+                        exhausted_descs = false;
+                        avail_iter.go_to_previous_position();
+
+                        /* EAGAIN */
+                        if e.kind() == std::io::ErrorKind::WouldBlock {
+                            break;
+                        }
+
+                        error!("net: rx: failed reading from tap: {}", e);
+                        return Err(NetQueuePairError::ReadTap(e));
+                    }
+
+                    // Write num_buffers to guest memory. We simply write 1 as we
+                    // never spread the frame over more than one descriptor chain.
+                    desc_chain
+                        .memory()
+                        .write_obj(1u16, num_buffers_addr)
+                        .map_err(NetQueuePairError::GuestMemory)?;
+
+                    self.counter_bytes += Wrapping(result as u64 - vnet_hdr_len() as u64);
+                    self.counter_frames += Wrapping(1);
+
+                    result as u32
+                } else {
+                    0
+                };
+
+                used_desc_head = (desc_chain.head_index(), len);
+
+                // For the sake of simplicity (keeping the handling of RX_QUEUE_EVENT and
+                // RX_TAP_EVENT totally asynchronous), we always let the 'last' descriptor
+                // chain go-through even if it was over the rate limit, and simply stop
+                // processing oncoming `avail_desc` if any.
+                if let Some(rate_limiter) = rate_limiter {
+                    rate_limit_reached = !rate_limiter.consume(1, TokenType::Ops)
+                        || !rate_limiter.consume(len as u64, TokenType::Bytes);
+                }
+            } else {
                 break;
             }
 
-            let head_index = avail_desc.index;
+            queue
-            let num_buffers_addr = mem.checked_offset(avail_desc.addr, 10).unwrap();
+                .add_used(used_desc_head.0, used_desc_head.1)
-            let mut next_desc = Some(avail_desc);
+                .map_err(NetQueuePairError::QueueAddUsed)?;
-
+            queue
-            let mut iovecs = Vec::new();
+                .enable_notification()
-            while let Some(desc) = next_desc {
+                .map_err(NetQueuePairError::QueueEnableNotification)?;
-                if desc.is_write_only() && desc.len > 0 {
-                    let buf = mem
-                        .get_slice(desc.addr, desc.len as usize)
-                        .map_err(NetQueuePairError::GuestMemory)?
-                        .as_ptr();
-                    let iovec = libc::iovec {
-                        iov_base: buf as *mut libc::c_void,
-                        iov_len: desc.len as libc::size_t,
-                    };
-                    iovecs.push(iovec);
-                }
-                next_desc = desc.next_descriptor();
-            }
-
-            let len = if !iovecs.is_empty() {
-                let result = unsafe {
-                    libc::readv(
-                        tap.as_raw_fd() as libc::c_int,
-                        iovecs.as_ptr() as *const libc::iovec,
-                        iovecs.len() as libc::c_int,
-                    )
-                };
-                if result < 0 {
-                    let e = std::io::Error::last_os_error();
-                    exhausted_descs = false;
-                    queue.go_to_previous_position();
-
-                    /* EAGAIN */
-                    if e.kind() == std::io::ErrorKind::WouldBlock {
-                        break;
-                    }
-
-                    error!("net: rx: failed reading from tap: {}", e);
-                    return Err(NetQueuePairError::ReadTap(e));
-                }
-
-                // Write num_buffers to guest memory. We simply write 1 as we
-                // never spread the frame over more than one descriptor chain.
-                mem.write_obj(1u16, num_buffers_addr)
-                    .map_err(NetQueuePairError::GuestMemory)?;
-
-                self.counter_bytes += Wrapping(result as u64 - vnet_hdr_len() as u64);
-                self.counter_frames += Wrapping(1);
-
-                result as u32
-            } else {
-                0
-            };
-
-            queue.add_used(mem, head_index, len);
-            queue.update_avail_event(mem);
-
-            // For the sake of simplicity (keeping the handling of RX_QUEUE_EVENT and
-            // RX_TAP_EVENT totally asynchronous), we always let the 'last' descriptor
-            // chain go-through even if it was over the rate limit, and simply stop
-            // processing oncoming `avail_desc` if any.
-            if let Some(rate_limiter) = rate_limiter {
-                rate_limit_reached = !rate_limiter.consume(1, TokenType::Ops)
-                    || !rate_limiter.consume(len as u64, TokenType::Bytes);
-            }
         }
 
         Ok(exhausted_descs)
@@ -244,10 +306,21 @@ pub enum NetQueuePairError {
     ReadTap(io::Error),
     /// Error related to guest memory
     GuestMemory(vm_memory::GuestMemoryError),
+    /// Returned an error while iterating through the queue
+    QueueIteratorFailed(virtio_queue::Error),
+    /// Descriptor chain is too short
+    DescriptorChainTooShort,
+    /// Descriptor chain does not contain valid descriptors
+    DescriptorChainInvalid,
+    /// Failed to determine if queue needed notification
+    QueueNeedsNotification(virtio_queue::Error),
+    /// Failed to enable notification on the queue
+    QueueEnableNotification(virtio_queue::Error),
+    /// Failed to add used index to the queue
+    QueueAddUsed(virtio_queue::Error),
 }
 
 pub struct NetQueuePair {
-    pub mem: Option<GuestMemoryAtomic<GuestMemoryMmap>>,
     pub tap: Tap,
     // With epoll each FD must be unique. So in order to filter the
     // events we need to get a second FD responding to the original
@@ -265,21 +338,19 @@ pub struct NetQueuePair {
     pub rx_desc_avail: bool,
     pub rx_rate_limiter: Option<RateLimiter>,
     pub tx_rate_limiter: Option<RateLimiter>,
+    pub access_platform: Option<Arc<dyn AccessPlatform>>,
 }
 
 impl NetQueuePair {
-    pub fn process_tx(&mut self, mut queue: &mut Queue) -> Result<bool, NetQueuePairError> {
+    pub fn process_tx(
-        let mem = self
+        &mut self,
-            .mem
+        queue: &mut Queue<GuestMemoryAtomic<GuestMemoryMmap>>,
-            .as_ref()
+    ) -> Result<bool, NetQueuePairError> {
-            .ok_or(NetQueuePairError::NoMemoryConfigured)
-            .map(|m| m.memory())?;
-
         let tx_tap_retry = self.tx.process_desc_chain(
-            &mem,
             &mut self.tap,
-            &mut queue,
+            queue,
             &mut self.tx_rate_limiter,
+            self.access_platform.as_ref(),
         )?;
 
         // We got told to try again when writing to the tap. Wait for the TAP to be writable
@@ -314,21 +385,20 @@ impl NetQueuePair {
         self.tx.counter_bytes = Wrapping(0);
         self.tx.counter_frames = Wrapping(0);
 
-        Ok(queue.needs_notification(&mem, queue.next_used))
+        queue
+            .needs_notification()
+            .map_err(NetQueuePairError::QueueNeedsNotification)
     }
 
-    pub fn process_rx(&mut self, mut queue: &mut Queue) -> Result<bool, NetQueuePairError> {
+    pub fn process_rx(
-        let mem = self
+        &mut self,
-            .mem
+        queue: &mut Queue<GuestMemoryAtomic<GuestMemoryMmap>>,
-            .as_ref()
+    ) -> Result<bool, NetQueuePairError> {
-            .ok_or(NetQueuePairError::NoMemoryConfigured)
-            .map(|m| m.memory())?;
-
         self.rx_desc_avail = !self.rx.process_desc_chain(
-            &mem,
             &mut self.tap,
-            &mut queue,
+            queue,
             &mut self.rx_rate_limiter,
+            self.access_platform.as_ref(),
         )?;
         let rate_limit_reached = self
             .rx_rate_limiter
@@ -358,6 +428,8 @@ impl NetQueuePair {
         self.rx.counter_bytes = Wrapping(0);
         self.rx.counter_frames = Wrapping(0);
 
-        Ok(queue.needs_notification(&mem, queue.next_used))
+        queue
+            .needs_notification()
+            .map_err(NetQueuePairError::QueueNeedsNotification)
     }
 }

net_util/src/tap.rs

@@ -5,7 +5,10 @@
 // Use of this source code is governed by a BSD-style license that can be
 // found in the THIRD-PARTY file.
 
-use super::{create_sockaddr, create_socket, vnet_hdr_len, Error as NetUtilError, MacAddr};
+use super::{
+    create_inet_socket, create_sockaddr, create_unix_socket, vnet_hdr_len, Error as NetUtilError,
+    MacAddr,
+};
 use crate::mac::MAC_ADDR_LEN;
 use std::fs::File;
 use std::io::{Error as IoError, Read, Result as IoResult, Write};
@@ -196,7 +199,7 @@ impl Tap {
 
     /// Set the host-side IP address for the tap interface.
     pub fn set_ip_addr(&self, ip_addr: net::Ipv4Addr) -> Result<()> {
-        let sock = create_socket().map_err(Error::NetUtil)?;
+        let sock = create_inet_socket().map_err(Error::NetUtil)?;
         let addr = create_sockaddr(ip_addr);
 
         let mut ifreq = self.get_ifreq();
@@ -204,7 +207,6 @@ impl Tap {
         ifreq.ifr_ifru.ifru_addr = addr;
 
         // ioctl is safe. Called with a valid sock fd, and we check the return.
-        #[allow(clippy::cast_lossless)]
         let ret =
             unsafe { ioctl_with_ref(&sock, net_gen::sockios::SIOCSIFADDR as c_ulong, &ifreq) };
         if ret < 0 {
@@ -225,12 +227,11 @@ impl Tap {
             return Ok(());
         }
 
-        let sock = create_socket().map_err(Error::NetUtil)?;
+        let sock = create_unix_socket().map_err(Error::NetUtil)?;
 
         let mut ifreq = self.get_ifreq();
 
         // ioctl is safe. Called with a valid sock fd, and we check the return.
-        #[allow(clippy::cast_lossless)]
         let ret =
             unsafe { ioctl_with_ref(&sock, net_gen::sockios::SIOCGIFHWADDR as c_ulong, &ifreq) };
         if ret < 0 {
@@ -245,7 +246,6 @@ impl Tap {
         }
 
         // ioctl is safe. Called with a valid sock fd, and we check the return.
-        #[allow(clippy::cast_lossless)]
         let ret =
             unsafe { ioctl_with_ref(&sock, net_gen::sockios::SIOCSIFHWADDR as c_ulong, &ifreq) };
         if ret < 0 {
@@ -257,12 +257,11 @@ impl Tap {
 
     /// Get mac addr for tap interface.
     pub fn get_mac_addr(&self) -> Result<MacAddr> {
-        let sock = create_socket().map_err(Error::NetUtil)?;
+        let sock = create_unix_socket().map_err(Error::NetUtil)?;
 
         let ifreq = self.get_ifreq();
 
         // ioctl is safe. Called with a valid sock fd, and we check the return.
-        #[allow(clippy::cast_lossless)]
         let ret =
             unsafe { ioctl_with_ref(&sock, net_gen::sockios::SIOCGIFHWADDR as c_ulong, &ifreq) };
         if ret < 0 {
@@ -279,7 +278,7 @@ impl Tap {
 
     /// Set the netmask for the subnet that the tap interface will exist on.
     pub fn set_netmask(&self, netmask: net::Ipv4Addr) -> Result<()> {
-        let sock = create_socket().map_err(Error::NetUtil)?;
+        let sock = create_inet_socket().map_err(Error::NetUtil)?;
         let addr = create_sockaddr(netmask);
 
         let mut ifreq = self.get_ifreq();
@@ -287,7 +286,6 @@ impl Tap {
         ifreq.ifr_ifru.ifru_addr = addr;
 
         // ioctl is safe. Called with a valid sock fd, and we check the return.
-        #[allow(clippy::cast_lossless)]
         let ret =
             unsafe { ioctl_with_ref(&sock, net_gen::sockios::SIOCSIFNETMASK as c_ulong, &ifreq) };
         if ret < 0 {
@@ -300,7 +298,6 @@ impl Tap {
     /// Set the offload flags for the tap interface.
     pub fn set_offload(&self, flags: c_uint) -> Result<()> {
         // ioctl is safe. Called with a valid tap fd, and we check the return.
-        #[allow(clippy::cast_lossless)]
         let ret =
             unsafe { ioctl_with_val(&self.tap_file, net_gen::TUNSETOFFLOAD(), flags as c_ulong) };
         if ret < 0 {
@@ -312,11 +309,10 @@ impl Tap {
 
     /// Enable the tap interface.
     pub fn enable(&self) -> Result<()> {
-        let sock = create_socket().map_err(Error::NetUtil)?;
+        let sock = create_unix_socket().map_err(Error::NetUtil)?;
 
         let mut ifreq = self.get_ifreq();
 
-        #[allow(clippy::cast_lossless)]
         let ret =
             unsafe { ioctl_with_ref(&sock, net_gen::sockios::SIOCGIFFLAGS as c_ulong, &ifreq) };
         if ret < 0 {
@@ -336,7 +332,6 @@ impl Tap {
             (net_gen::net_device_flags_IFF_UP | net_gen::net_device_flags_IFF_RUNNING) as i16;
 
         // ioctl is safe. Called with a valid sock fd, and we check the return.
-        #[allow(clippy::cast_lossless)]
         let ret =
             unsafe { ioctl_with_ref(&sock, net_gen::sockios::SIOCSIFFLAGS as c_ulong, &ifreq) };
         if ret < 0 {

option_parser/src/lib.rs

@@ -5,6 +5,7 @@
 
 use std::collections::HashMap;
 use std::fmt;
+use std::num::ParseIntError;
 use std::str::FromStr;
 
 #[derive(Default)]
@@ -37,6 +38,32 @@ impl fmt::Display for OptionParserError {
 }
 type OptionParserResult<T> = std::result::Result<T, OptionParserError>;
 
+fn split_commas_outside_brackets(s: &str) -> OptionParserResult<Vec<String>> {
+    let mut list: Vec<String> = Vec::new();
+    let mut opened_brackets: usize = 0;
+    for element in s.trim().split(',') {
+        if opened_brackets > 0 {
+            if let Some(last) = list.last_mut() {
+                *last = format!("{},{}", last, element);
+            } else {
+                return Err(OptionParserError::InvalidSyntax(s.to_owned()));
+            }
+        } else {
+            list.push(element.to_string());
+        }
+
+        opened_brackets += element.matches('[').count();
+        let closing_brackets = element.matches(']').count();
+        if closing_brackets > opened_brackets {
+            return Err(OptionParserError::InvalidSyntax(s.to_owned()));
+        } else {
+            opened_brackets -= closing_brackets;
+        }
+    }
+
+    Ok(list)
+}
+
 impl OptionParser {
     pub fn new() -> Self {
         Self {
@@ -49,9 +76,7 @@ impl OptionParser {
             return Ok(());
         }
 
-        let options_list: Vec<&str> = input.trim().split(',').collect();
+        for option in split_commas_outside_brackets(input)?.iter() {
-
-        for option in options_list.iter() {
             let parts: Vec<&str> = option.split('=').collect();
 
             match self.options.get_mut(parts[0]) {
@@ -183,7 +208,11 @@ impl FromStr for IntegerList {
 
     fn from_str(s: &str) -> std::result::Result<Self, Self::Err> {
         let mut integer_list = Vec::new();
-        let ranges_list: Vec<&str> = s.trim().split(':').collect();
+        let ranges_list: Vec<&str> = s
+            .trim()
+            .trim_matches(|c| c == '[' || c == ']')
+            .split(',')
+            .collect();
 
         for range in ranges_list.iter() {
             let items: Vec<&str> = range.split('-').collect();
@@ -216,39 +245,71 @@ impl FromStr for IntegerList {
     }
 }
 
-pub struct TupleTwoIntegers(pub Vec<(u64, u64)>);
+pub trait TupleValue {
-
+    fn parse_value(input: &str) -> Result<Self, TupleError>
-pub enum TupleTwoIntegersParseError {
+    where
-    InvalidValue(String),
+        Self: Sized;
 }
 
-impl FromStr for TupleTwoIntegers {
+impl TupleValue for u64 {
-    type Err = TupleTwoIntegersParseError;
+    fn parse_value(input: &str) -> Result<Self, TupleError> {
+        input.parse::<u64>().map_err(TupleError::InvalidInteger)
+    }
+}
+
+impl TupleValue for Vec<u8> {
+    fn parse_value(input: &str) -> Result<Self, TupleError> {
+        Ok(IntegerList::from_str(input)
+            .map_err(TupleError::InvalidIntegerList)?
+            .0
+            .iter()
+            .map(|v| *v as u8)
+            .collect())
+    }
+}
+
+impl TupleValue for Vec<u64> {
+    fn parse_value(input: &str) -> Result<Self, TupleError> {
+        Ok(IntegerList::from_str(input)
+            .map_err(TupleError::InvalidIntegerList)?
+            .0)
+    }
+}
+
+pub struct Tuple<S, T>(pub Vec<(S, T)>);
+
+pub enum TupleError {
+    InvalidValue(String),
+    SplitOutsideBrackets(OptionParserError),
+    InvalidIntegerList(IntegerListParseError),
+    InvalidInteger(ParseIntError),
+}
+
+impl<S: FromStr, T: TupleValue> FromStr for Tuple<S, T> {
+    type Err = TupleError;
 
     fn from_str(s: &str) -> std::result::Result<Self, Self::Err> {
-        let mut list = Vec::new();
+        let mut list: Vec<(S, T)> = Vec::new();
-        let tuples_list: Vec<&str> = s.trim().split(':').collect();
 
+        let tuples_list =
+            split_commas_outside_brackets(s.trim().trim_matches(|c| c == '[' || c == ']'))
+                .map_err(TupleError::SplitOutsideBrackets)?;
         for tuple in tuples_list.iter() {
             let items: Vec<&str> = tuple.split('@').collect();
 
             if items.len() != 2 {
-                return Err(TupleTwoIntegersParseError::InvalidValue(
+                return Err(TupleError::InvalidValue((*tuple).to_string()));
-                    (*tuple).to_string(),
-                ));
             }
 
             let item1 = items[0]
-                .parse::<u64>()
+                .parse::<S>()
-                .map_err(|_| TupleTwoIntegersParseError::InvalidValue(items[0].to_owned()))?;
+                .map_err(|_| TupleError::InvalidValue(items[0].to_owned()))?;
-            let item2 = items[1]
+            let item2 = TupleValue::parse_value(items[1])?;
-                .parse::<u64>()
-                .map_err(|_| TupleTwoIntegersParseError::InvalidValue(items[1].to_owned()))?;
 
             list.push((item1, item2));
         }
 
-        Ok(TupleTwoIntegers(list))
+        Ok(Tuple(list))
     }
 }
 
@@ -262,7 +323,12 @@ impl FromStr for StringList {
     type Err = StringListParseError;
 
     fn from_str(s: &str) -> std::result::Result<Self, Self::Err> {
-        let string_list: Vec<String> = s.trim().split(':').map(|e| e.to_owned()).collect();
+        let string_list: Vec<String> = s
+            .trim()
+            .trim_matches(|c| c == '[' || c == ']')
+            .split(',')
+            .map(|e| e.to_owned())
+            .collect();
 
         Ok(StringList(string_list))
     }

pci/Cargo.toml

@@ -10,22 +10,24 @@ kvm = ["vfio-ioctls/kvm"]
 mshv = ["vfio-ioctls/mshv"]
 
 [dependencies]
-anyhow = "1.0.44"
+anyhow = "1.0.55"
 byteorder = "1.4.3"
 hypervisor = { path = "../hypervisor" }
-vfio-ioctls = { git = "https://github.com/rust-vmm/vfio-ioctls", branch = "main", default-features = false }
+vfio-ioctls = { git = "https://github.com/rust-vmm/vfio", branch = "main", default-features = false }
 vfio_user = { path = "../vfio_user" }
 vmm-sys-util = "0.9.0"
-libc = "0.2.103"
+libc = "0.2.119"
 log = "0.4.14"
+serde = "1.0.136"
+serde_derive = "1.0.136"
 thiserror = "1.0.30"
 versionize = "0.1.6"
 versionize_derive = "0.1.4"
 vm-allocator = { path = "../vm-allocator" }
 vm-device = { path = "../vm-device" }
-vm-memory = "0.6.0"
+vm-memory = "0.7.0"
 vm-migration = { path = "../vm-migration" }
 
 [dependencies.vfio-bindings]
-version = "0.2.0"
+version = "0.3.1"
 features = ["fam-wrappers"]

pci/src/bus.rs

@@ -140,12 +140,8 @@ impl PciBus {
         Ok(())
     }
 
-    pub fn add_device(
+    pub fn add_device(&mut self, device_id: u32, device: Arc<Mutex<dyn PciDevice>>) -> Result<()> {
-        &mut self,
+        self.devices.insert(device_id, device);
-        pci_device_bdf: u32,
-        device: Arc<Mutex<dyn PciDevice>>,
-    ) -> Result<()> {
-        self.devices.insert(pci_device_bdf >> 3, device);
         Ok(())
     }
 

pci/src/configuration.rs

@@ -21,6 +21,8 @@ const ROM_BAR_REG: usize = 12;
 const BAR_IO_ADDR_MASK: u32 = 0xffff_fffc;
 const BAR_MEM_ADDR_MASK: u32 = 0xffff_fff0;
 const ROM_BAR_ADDR_MASK: u32 = 0xffff_f800;
+const MSI_CAPABILITY_REGISTER_MASK: u32 = 0x0071_0000;
+const MSIX_CAPABILITY_REGISTER_MASK: u32 = 0xc000_0000;
 const NUM_BAR_REGS: usize = 6;
 const CAPABILITY_LIST_HEAD_OFFSET: usize = 0x34;
 const FIRST_CAPABILITY_OFFSET: usize = 0x40;
@@ -714,8 +716,15 @@ impl PciConfiguration {
         }
         self.last_capability = Some((cap_offset, total_len));
 
-        if cap_data.id() == PciCapabilityId::MsiX {
+        match cap_data.id() {
-            self.msix_cap_reg_idx = Some(cap_offset / 4);
+            PciCapabilityId::MessageSignalledInterrupts => {
+                self.writable_bits[cap_offset / 4] = MSI_CAPABILITY_REGISTER_MASK;
+            }
+            PciCapabilityId::MsiX => {
+                self.msix_cap_reg_idx = Some(cap_offset / 4);
+                self.writable_bits[self.msix_cap_reg_idx.unwrap()] = MSIX_CAPABILITY_REGISTER_MASK;
+            }
+            _ => {}
         }
 
         Ok(cap_offset)
@@ -740,6 +749,11 @@ impl PciConfiguration {
                         .lock()
                         .unwrap()
                         .set_msg_ctl(LittleEndian::read_u16(data));
+                } else if msix_cap_reg_idx == reg_idx && offset == 0 && data.len() == 4 {
+                    msix_config
+                        .lock()
+                        .unwrap()
+                        .set_msg_ctl((LittleEndian::read_u32(data) >> 16) as u16);
                 }
             }
         }
@@ -922,16 +936,19 @@ impl PciBarConfiguration {
         }
     }
 
+    #[must_use]
     pub fn set_register_index(mut self, reg_idx: usize) -> Self {
         self.reg_idx = reg_idx;
         self
     }
 
+    #[must_use]
     pub fn set_address(mut self, addr: u64) -> Self {
         self.addr = addr;
         self
     }
 
+    #[must_use]
     pub fn set_size(mut self, size: u64) -> Self {
         self.size = size;
         self
@@ -941,6 +958,7 @@ impl PciBarConfiguration {
         self.size
     }
 
+    #[must_use]
     pub fn set_region_type(mut self, region_type: PciBarRegionType) -> Self {
         self.region_type = region_type;
         self
@@ -961,7 +979,7 @@ mod tests {
         foo: u8,
     }
 
-    // It is safe to implement BytesValued; all members are simple numbers and any value is valid.
+    // SAFETY: All members are simple numbers and any value is valid.
     unsafe impl ByteValued for TestCap {}
 
     impl PciCapability for TestCap {

pci/src/device.rs

@@ -5,9 +5,9 @@
 use crate::configuration::{self, PciBarRegionType};
 use std::any::Any;
 use std::fmt::{self, Display};
-use std::sync::{Arc, Barrier};
+use std::sync::{Arc, Barrier, Mutex};
 use std::{self, io, result};
-use vm_allocator::SystemAllocator;
+use vm_allocator::{AddressAllocator, SystemAllocator};
 use vm_device::BusDevice;
 use vm_memory::{GuestAddress, GuestUsize};
 
@@ -51,13 +51,18 @@ pub trait PciDevice: BusDevice {
     /// returns an address. Returns a Vec of (GuestAddress, GuestUsize) tuples.
     fn allocate_bars(
         &mut self,
-        _allocator: &mut SystemAllocator,
+        _allocator: &Arc<Mutex<SystemAllocator>>,
+        _mmio_allocator: &mut AddressAllocator,
     ) -> Result<Vec<(GuestAddress, GuestUsize, PciBarRegionType)>> {
         Ok(Vec::new())
     }
 
     /// Frees the PCI BARs previously allocated with a call to allocate_bars().
-    fn free_bars(&mut self, _allocator: &mut SystemAllocator) -> Result<()> {
+    fn free_bars(
+        &mut self,
+        _allocator: &mut SystemAllocator,
+        _mmio_allocator: &mut AddressAllocator,
+    ) -> Result<()> {
         Ok(())
     }
 

pci/src/lib.rs

@@ -27,6 +27,10 @@ pub use self::msi::{msi_num_enabled_vectors, MsiCap, MsiConfig};
 pub use self::msix::{MsixCap, MsixConfig, MsixTableEntry, MSIX_TABLE_ENTRY_SIZE};
 pub use self::vfio::{VfioPciDevice, VfioPciError};
 pub use self::vfio_user::{VfioUserDmaMapping, VfioUserPciDevice, VfioUserPciDeviceError};
+use serde::de::Visitor;
+use std::fmt::{self, Display};
+use std::num::ParseIntError;
+use std::str::FromStr;
 
 /// PCI has four interrupt pins A->D.
 #[derive(Copy, Clone)]
@@ -47,3 +51,133 @@ impl PciInterruptPin {
 pub const PCI_CONFIG_IO_PORT: u64 = 0xcf8;
 #[cfg(target_arch = "x86_64")]
 pub const PCI_CONFIG_IO_PORT_SIZE: u64 = 0x8;
+
+#[derive(Clone, Copy, PartialEq, PartialOrd)]
+pub struct PciBdf(u32);
+
+struct PciBdfVisitor;
+
+impl<'de> Visitor<'de> for PciBdfVisitor {
+    type Value = PciBdf;
+
+    fn expecting(&self, formatter: &mut fmt::Formatter) -> fmt::Result {
+        formatter.write_str("struct PciBdf")
+    }
+
+    fn visit_str<E>(self, v: &str) -> Result<Self::Value, E>
+    where
+        E: serde::de::Error,
+    {
+        Ok(v.into())
+    }
+}
+
+impl<'de> serde::Deserialize<'de> for PciBdf {
+    fn deserialize<D>(deserializer: D) -> Result<Self, D::Error>
+    where
+        D: serde::Deserializer<'de>,
+    {
+        deserializer.deserialize_str(PciBdfVisitor)
+    }
+}
+
+impl serde::Serialize for PciBdf {
+    fn serialize<S>(&self, serializer: S) -> Result<S::Ok, S::Error>
+    where
+        S: serde::Serializer,
+    {
+        serializer.collect_str(&self.to_string())
+    }
+}
+
+impl PciBdf {
+    pub fn segment(&self) -> u16 {
+        ((self.0 >> 16) & 0xffff) as u16
+    }
+
+    pub fn bus(&self) -> u8 {
+        ((self.0 >> 8) & 0xff) as u8
+    }
+
+    pub fn device(&self) -> u8 {
+        ((self.0 >> 3) & 0x1f) as u8
+    }
+
+    pub fn function(&self) -> u8 {
+        (self.0 & 0x7) as u8
+    }
+
+    pub fn new(segment: u16, bus: u8, device: u8, function: u8) -> Self {
+        Self(
+            (segment as u32) << 16
+                | (bus as u32) << 8
+                | ((device & 0x1f) as u32) << 3
+                | (function & 0x7) as u32,
+        )
+    }
+}
+
+impl From<u32> for PciBdf {
+    fn from(bdf: u32) -> Self {
+        Self(bdf)
+    }
+}
+
+impl From<PciBdf> for u32 {
+    fn from(bdf: PciBdf) -> Self {
+        bdf.0
+    }
+}
+
+impl From<&PciBdf> for u32 {
+    fn from(bdf: &PciBdf) -> Self {
+        bdf.0
+    }
+}
+
+impl From<PciBdf> for u16 {
+    fn from(bdf: PciBdf) -> Self {
+        (bdf.0 & 0xffff) as u16
+    }
+}
+
+impl From<&PciBdf> for u16 {
+    fn from(bdf: &PciBdf) -> Self {
+        (bdf.0 & 0xffff) as u16
+    }
+}
+
+impl Display for PciBdf {
+    fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
+        write!(
+            f,
+            "{:04x}:{:02x}:{:02x}.{:01x}",
+            self.segment(),
+            self.bus(),
+            self.device(),
+            self.function()
+        )
+    }
+}
+
+impl FromStr for PciBdf {
+    type Err = ParseIntError;
+
+    fn from_str(s: &str) -> Result<Self, Self::Err> {
+        let items: Vec<&str> = s.split('.').collect();
+        assert_eq!(items.len(), 2);
+        let function = u8::from_str_radix(items[1], 16)?;
+        let items: Vec<&str> = items[0].split(':').collect();
+        assert_eq!(items.len(), 3);
+        let segment = u16::from_str_radix(items[0], 16)?;
+        let bus = u8::from_str_radix(items[1], 16)?;
+        let device = u8::from_str_radix(items[2], 16)?;
+        Ok(PciBdf::new(segment, bus, device, function))
+    }
+}
+
+impl From<&str> for PciBdf {
+    fn from(bdf: &str) -> Self {
+        Self::from_str(bdf).unwrap()
+    }
+}

pci/src/msix.rs

@@ -162,6 +162,7 @@ impl MsixConfig {
         // Update interrupt routing
         if old_masked != self.masked || old_enabled != self.enabled {
             if self.enabled && !self.masked {
+                debug!("MSI-X enabled for device 0x{:x}", self.devid);
                 for (idx, table_entry) in self.table_entries.iter().enumerate() {
                     let config = MsiIrqSourceConfig {
                         high_addr: table_entry.msg_addr_hi,
@@ -187,6 +188,7 @@ impl MsixConfig {
                     }
                 }
             } else if old_enabled || !old_masked {
+                debug!("MSI-X disabled for device 0x{:x}", self.devid);
                 if let Err(e) = self.interrupt_source_group.disable() {
                     error!("Failed disabling irq_fd: {:?}", e);
                 }
@@ -468,7 +470,7 @@ pub struct MsixCap {
     pub pba: u32,
 }
 
-// It is safe to implement ByteValued. All members are simple numbers and any value is valid.
+// SAFETY: All members are simple numbers and any value is valid.
 unsafe impl ByteValued for MsixCap {}
 
 impl PciCapability for MsixCap {
@@ -533,4 +535,16 @@ impl MsixCap {
     pub fn table_size(&self) -> u16 {
         (self.msg_ctl & 0x7ff) + 1
     }
+
+    pub fn table_range(&self) -> (u64, u64) {
+        // The table takes 16 bytes per entry.
+        let size = self.table_size() as u64 * 16;
+        (self.table_offset() as u64, size)
+    }
+
+    pub fn pba_range(&self) -> (u64, u64) {
+        // The table takes 1 bit per entry modulo 8 bytes.
+        let size = ((self.table_size() as u64 / 64) + 1) * 8;
+        (self.pba_offset() as u64, size)
+    }
 }

pci/src/vfio.rs

@@ -5,20 +5,21 @@
 
 use crate::{
     msi_num_enabled_vectors, BarReprogrammingParams, MsiConfig, MsixCap, MsixConfig,
-    PciBarConfiguration, PciBarRegionType, PciCapabilityId, PciClassCode, PciConfiguration,
+    PciBarConfiguration, PciBarRegionType, PciBdf, PciCapabilityId, PciClassCode, PciConfiguration,
     PciDevice, PciDeviceError, PciHeaderType, PciSubclass, MSIX_TABLE_ENTRY_SIZE,
 };
 use byteorder::{ByteOrder, LittleEndian};
 use hypervisor::HypervisorVmError;
 use std::any::Any;
+use std::collections::BTreeMap;
 use std::io;
 use std::os::unix::io::AsRawFd;
 use std::ptr::null_mut;
-use std::sync::{Arc, Barrier};
+use std::sync::{Arc, Barrier, Mutex};
 use thiserror::Error;
 use vfio_bindings::bindings::vfio::*;
-use vfio_ioctls::{VfioContainer, VfioDevice, VfioIrq};
+use vfio_ioctls::{VfioContainer, VfioDevice, VfioIrq, VfioRegionInfoCap};
-use vm_allocator::SystemAllocator;
+use vm_allocator::{AddressAllocator, SystemAllocator};
 use vm_device::interrupt::{
     InterruptIndex, InterruptManager, InterruptSourceGroup, MsiIrqGroupConfig,
 };
@@ -209,6 +210,14 @@ impl Interrupt {
 }
 
 #[derive(Copy, Clone)]
+pub struct UserMemoryRegion {
+    slot: u32,
+    start: u64,
+    size: u64,
+    host_addr: u64,
+}
+
+#[derive(Clone)]
 pub struct MmioRegion {
     pub start: GuestAddress,
     pub length: GuestUsize,
@@ -217,6 +226,7 @@ pub struct MmioRegion {
     pub(crate) mem_slot: Option<u32>,
     pub(crate) host_addr: Option<u64>,
     pub(crate) mmap_size: Option<usize>,
+    pub(crate) user_memory_regions: Vec<UserMemoryRegion>,
 }
 #[derive(Debug, Error)]
 pub enum VfioError {
@@ -350,7 +360,8 @@ pub(crate) struct VfioCommon {
 impl VfioCommon {
     pub(crate) fn allocate_bars(
         &mut self,
-        allocator: &mut SystemAllocator,
+        allocator: &Arc<Mutex<SystemAllocator>>,
+        mmio_allocator: &mut AddressAllocator,
         vfio_wrapper: &dyn Vfio,
     ) -> Result<Vec<(GuestAddress, GuestUsize, PciBarRegionType)>, PciDeviceError> {
         let mut ranges = Vec::new();
@@ -400,25 +411,27 @@ impl VfioCommon {
             let mut lower = vfio_wrapper.read_config_dword(bar_offset);
 
             if io_bar {
+                // Mask flag bits (lowest 2 for I/O bars)
+                lower &= !0b11;
+
+                // BAR is not enabled
+                if lower == 0 {
+                    bar_id += 1;
+                    continue;
+                }
+
                 #[cfg(target_arch = "x86_64")]
                 {
                     // IO BAR
                     region_type = PciBarRegionType::IoRegion;
 
-                    // Mask flag bits (lowest 2 for I/O bars)
-                    lower &= !0b11;
-
-                    // BAR is not enabled
-                    if lower == 0 {
-                        bar_id += 1;
-                        continue;
-                    }
-
                     // Invert bits and add 1 to calculate size
                     region_size = (!lower + 1) as u64;
 
                     // The address needs to be 4 bytes aligned.
                     bar_addr = allocator
+                        .lock()
+                        .unwrap()
                         .allocate_io_addresses(None, region_size, Some(0x4))
                         .ok_or(PciDeviceError::IoAllocationFailed(region_size))?;
                 }
@@ -448,8 +461,8 @@ impl VfioCommon {
                 region_size = (!combined_size + 1) as u64;
 
                 // BAR allocation must be naturally aligned
-                bar_addr = allocator
+                bar_addr = mmio_allocator
-                    .allocate_mmio_addresses(None, region_size, Some(region_size))
+                    .allocate(None, region_size, Some(region_size))
                     .ok_or(PciDeviceError::IoAllocationFailed(region_size))?;
             } else {
                 // Mask out flag bits (lowest 4 for memory bars)
@@ -465,6 +478,8 @@ impl VfioCommon {
 
                 // BAR allocation must be naturally aligned
                 bar_addr = allocator
+                    .lock()
+                    .unwrap()
                     .allocate_mmio_hole_addresses(None, region_size, Some(region_size))
                     .ok_or(PciDeviceError::IoAllocationFailed(region_size))?;
             }
@@ -501,6 +516,7 @@ impl VfioCommon {
                 mem_slot: None,
                 host_addr: None,
                 mmap_size: None,
+                user_memory_regions: Vec::new(),
             });
 
             bar_id += 1;
@@ -515,6 +531,7 @@ impl VfioCommon {
     pub(crate) fn free_bars(
         &mut self,
         allocator: &mut SystemAllocator,
+        mmio_allocator: &mut AddressAllocator,
     ) -> Result<(), PciDeviceError> {
         for region in self.mmio_regions.iter() {
             match region.type_ {
@@ -528,7 +545,7 @@ impl VfioCommon {
                     allocator.free_mmio_hole_addresses(region.start, region.length);
                 }
                 PciBarRegionType::Memory64BitRegion => {
-                    allocator.free_mmio_addresses(region.start, region.length);
+                    mmio_allocator.free(region.start, region.length);
                 }
             }
         }
@@ -540,6 +557,7 @@ impl VfioCommon {
         cap: u8,
         interrupt_manager: &Arc<dyn InterruptManager<GroupConfig = MsiIrqGroupConfig>>,
         vfio_wrapper: &dyn Vfio,
+        bdf: PciBdf,
     ) {
         let msg_ctl = vfio_wrapper.read_config_word((cap + 2).into());
 
@@ -560,7 +578,11 @@ impl VfioCommon {
             })
             .unwrap();
 
-        let msix_config = MsixConfig::new(msix_cap.table_size(), interrupt_source_group.clone(), 0);
+        let msix_config = MsixConfig::new(
+            msix_cap.table_size(),
+            interrupt_source_group.clone(),
+            bdf.into(),
+        );
 
         self.interrupt.msix = Some(VfioMsix {
             bar: msix_config,
@@ -598,6 +620,7 @@ impl VfioCommon {
         &mut self,
         interrupt_manager: &Arc<dyn InterruptManager<GroupConfig = MsiIrqGroupConfig>>,
         vfio_wrapper: &dyn Vfio,
+        bdf: PciBdf,
     ) {
         let mut cap_next = vfio_wrapper.read_config_byte(PCI_CONFIG_CAPABILITY_OFFSET);
 
@@ -619,7 +642,12 @@ impl VfioCommon {
                         if irq_info.count > 0 {
                             // Parse capability only if the VFIO device
                             // supports MSI-X.
-                            self.parse_msix_capabilities(cap_next, interrupt_manager, vfio_wrapper);
+                            self.parse_msix_capabilities(
+                                cap_next,
+                                interrupt_manager,
+                                vfio_wrapper,
+                                bdf,
+                            );
                         }
                     }
                 }
@@ -786,7 +814,7 @@ impl VfioCommon {
             if addr >= region.start.raw_value()
                 && addr < region.start.unchecked_add(region.length).raw_value()
             {
-                return Some(*region);
+                return Some(region.clone());
             }
         }
         None
@@ -951,6 +979,7 @@ impl VfioPciDevice {
         msi_interrupt_manager: &Arc<dyn InterruptManager<GroupConfig = MsiIrqGroupConfig>>,
         legacy_interrupt_group: Option<Arc<dyn InterruptSourceGroup>>,
         iommu_attached: bool,
+        bdf: PciBdf,
     ) -> Result<Self, VfioPciError> {
         let device = Arc::new(device);
         device.reset();
@@ -980,7 +1009,7 @@ impl VfioPciDevice {
             },
         };
 
-        common.parse_capabilities(msi_interrupt_manager, &vfio_wrapper);
+        common.parse_capabilities(msi_interrupt_manager, &vfio_wrapper, bdf);
         common.initialize_legacy_interrupt(legacy_interrupt_group, &vfio_wrapper)?;
 
         let vfio_pci_device = VfioPciDevice {
@@ -999,6 +1028,85 @@ impl VfioPciDevice {
         self.iommu_attached
     }
 
+    fn align_4k(address: u64) -> u64 {
+        (address + 0xfff) & 0xffff_ffff_ffff_f000
+    }
+
+    fn is_4k_aligned(address: u64) -> bool {
+        (address & 0xfff) == 0
+    }
+
+    fn is_4k_multiple(size: u64) -> bool {
+        (size & 0xfff) == 0
+    }
+
+    fn generate_user_memory_regions<F>(
+        region_index: u32,
+        region_start: u64,
+        region_size: u64,
+        host_addr: u64,
+        mem_slot: F,
+        vfio_msix: Option<&VfioMsix>,
+    ) -> Vec<UserMemoryRegion>
+    where
+        F: Fn() -> u32,
+    {
+        if !Self::is_4k_aligned(region_start) {
+            error!(
+                "Region start address 0x{:x} must be at least aligned on 4KiB",
+                region_start
+            );
+        }
+        if !Self::is_4k_multiple(region_size) {
+            error!(
+                "Region size 0x{:x} must be at least a multiple of 4KiB",
+                region_size
+            );
+        }
+
+        // Using a BtreeMap as the list provided through the iterator is sorted
+        // by key. This ensures proper split of the whole region.
+        let mut inter_ranges = BTreeMap::new();
+        if let Some(msix) = vfio_msix {
+            if region_index == msix.cap.table_bir() {
+                let (offset, size) = msix.cap.table_range();
+                let base = region_start + offset;
+                inter_ranges.insert(base, size);
+            }
+            if region_index == msix.cap.pba_bir() {
+                let (offset, size) = msix.cap.pba_range();
+                let base = region_start + offset;
+                inter_ranges.insert(base, size);
+            }
+        }
+
+        let mut user_memory_regions = Vec::new();
+        let mut new_start = region_start;
+        for (range_start, range_size) in inter_ranges {
+            if range_start > new_start {
+                user_memory_regions.push(UserMemoryRegion {
+                    slot: mem_slot(),
+                    start: new_start,
+                    size: range_start - new_start,
+                    host_addr: host_addr + new_start - region_start,
+                });
+            }
+
+            new_start = Self::align_4k(range_start + range_size);
+        }
+
+        if region_start + region_size > new_start {
+            user_memory_regions.push(UserMemoryRegion {
+                slot: mem_slot(),
+                start: new_start,
+                size: region_start + region_size - new_start,
+                host_addr: host_addr + new_start - region_start,
+            });
+        }
+
+        user_memory_regions
+    }
+
     /// Map MMIO regions into the guest, and avoid VM exits when the guest tries
     /// to reach those regions.
     ///
@@ -1018,16 +1126,6 @@ impl VfioPciDevice {
         let fd = self.device.as_raw_fd();
 
         for region in self.common.mmio_regions.iter_mut() {
-            // We want to skip the mapping of the BAR containing the MSI-X
-            // table even if it is mappable. The reason is we need to trap
-            // any access to the MSI-X table and update the GSI routing
-            // accordingly.
-            if let Some(msix) = &self.common.interrupt.msix {
-                if region.index == msix.cap.table_bir() || region.index == msix.cap.pba_bir() {
-                    continue;
-                }
-            }
-
             let region_flags = self.device.get_region_flags(region.index);
             if region_flags & VFIO_REGION_INFO_FLAG_MMAP != 0 {
                 let mut prot = 0;
@@ -1037,8 +1135,27 @@ impl VfioPciDevice {
                 if region_flags & VFIO_REGION_INFO_FLAG_WRITE != 0 {
                     prot |= libc::PROT_WRITE;
                 }
-                let (mmap_offset, mmap_size) = self.device.get_region_mmap(region.index);
+
-                let offset = self.device.get_region_offset(region.index) + mmap_offset;
+                // Retrieve the list of capabilities found on the region
+                let caps = if region_flags & VFIO_REGION_INFO_FLAG_CAPS != 0 {
+                    self.device.get_region_caps(region.index)
+                } else {
+                    Vec::new()
+                };
+
+                // Don't try to mmap the region if it contains MSI-X table or
+                // MSI-X PBA subregion, and if we couldn't find MSIX_MAPPABLE
+                // in the list of supported capabilities.
+                if let Some(msix) = self.common.interrupt.msix.as_ref() {
+                    if (region.index == msix.cap.table_bir() || region.index == msix.cap.pba_bir())
+                        && !caps.contains(&VfioRegionInfoCap::MsixMappable)
+                    {
+                        continue;
+                    }
+                }
+
+                let mmap_size = self.device.get_region_size(region.index);
+                let offset = self.device.get_region_offset(region.index);
 
                 let host_addr = unsafe {
                     libc::mmap(
@@ -1053,29 +1170,44 @@ impl VfioPciDevice {
 
                 if host_addr == libc::MAP_FAILED {
                     error!(
-                        "Could not mmap regions, error:{}",
+                        "Could not mmap region index {}: {}",
+                        region.index,
                         io::Error::last_os_error()
                     );
                     continue;
                 }
 
-                let slot = mem_slot();
+                // In case the region that is being mapped contains the MSI-X
-                let mem_region = vm.make_user_memory_region(
+                // vectors table or the MSI-X PBA table, we must adjust what
-                    slot,
+                // is being declared through the hypervisor. We want to make
-                    region.start.raw_value() + mmap_offset,
+                // sure we will still trap MMIO accesses to these MSI-X
-                    mmap_size as u64,
+                // specific ranges.
+                let user_memory_regions = Self::generate_user_memory_regions(
+                    region.index,
+                    region.start.raw_value(),
+                    mmap_size,
                     host_addr as u64,
-                    false,
+                    &mem_slot,
-                    false,
+                    self.common.interrupt.msix.as_ref(),
                 );
+                for user_memory_region in user_memory_regions.iter() {
+                    let mem_region = vm.make_user_memory_region(
+                        user_memory_region.slot,
+                        user_memory_region.start,
+                        user_memory_region.size,
+                        user_memory_region.host_addr,
+                        false,
+                        false,
+                    );
 
-                vm.create_user_memory_region(mem_region)
+                    vm.create_user_memory_region(mem_region)
-                    .map_err(VfioPciError::MapRegionGuest)?;
+                        .map_err(VfioPciError::MapRegionGuest)?;
+                }
 
                 // Update the region with memory mapped info.
-                region.mem_slot = Some(slot);
                 region.host_addr = Some(host_addr as u64);
                 region.mmap_size = Some(mmap_size as usize);
+                region.user_memory_regions = user_memory_regions;
             }
         }
 
@@ -1084,17 +1216,13 @@ impl VfioPciDevice {
 
     pub fn unmap_mmio_regions(&mut self) {
         for region in self.common.mmio_regions.iter() {
-            if let (Some(host_addr), Some(mmap_size), Some(mem_slot)) =
+            for user_memory_region in region.user_memory_regions.iter() {
-                (region.host_addr, region.mmap_size, region.mem_slot)
-            {
-                let (mmap_offset, _) = self.device.get_region_mmap(region.index);
-
                 // Remove region
                 let r = self.vm.make_user_memory_region(
-                    mem_slot,
+                    user_memory_region.slot,
-                    region.start.raw_value() + mmap_offset,
+                    user_memory_region.start,
-                    mmap_size as u64,
+                    user_memory_region.size,
-                    host_addr as u64,
+                    user_memory_region.host_addr,
                     false,
                     false,
                 );
@@ -1102,7 +1230,9 @@ impl VfioPciDevice {
                 if let Err(e) = self.vm.remove_user_memory_region(r) {
                     error!("Could not remove the userspace memory region: {}", e);
                 }
+            }
 
+            if let (Some(host_addr), Some(mmap_size)) = (region.host_addr, region.mmap_size) {
                 let ret = unsafe { libc::munmap(host_addr as *mut libc::c_void, mmap_size) };
                 if ret != 0 {
                     error!(
@@ -1194,13 +1324,19 @@ const PCI_ROM_EXP_BAR_INDEX: usize = 12;
 impl PciDevice for VfioPciDevice {
     fn allocate_bars(
         &mut self,
-        allocator: &mut SystemAllocator,
+        allocator: &Arc<Mutex<SystemAllocator>>,
+        mmio_allocator: &mut AddressAllocator,
     ) -> Result<Vec<(GuestAddress, GuestUsize, PciBarRegionType)>, PciDeviceError> {
-        self.common.allocate_bars(allocator, &self.vfio_wrapper)
+        self.common
+            .allocate_bars(allocator, mmio_allocator, &self.vfio_wrapper)
     }
 
-    fn free_bars(&mut self, allocator: &mut SystemAllocator) -> Result<(), PciDeviceError> {
+    fn free_bars(
-        self.common.free_bars(allocator)
+        &mut self,
+        allocator: &mut SystemAllocator,
+        mmio_allocator: &mut AddressAllocator,
+    ) -> Result<(), PciDeviceError> {
+        self.common.free_bars(allocator, mmio_allocator)
     }
 
     fn write_config_register(
@@ -1242,38 +1378,41 @@ impl PciDevice for VfioPciDevice {
             if region.start.raw_value() == old_base {
                 region.start = GuestAddress(new_base);
 
-                if let Some(mem_slot) = region.mem_slot {
+                for user_memory_region in region.user_memory_regions.iter_mut() {
-                    if let Some(host_addr) = region.host_addr {
+                    // Remove old region
-                        let (mmap_offset, mmap_size) = self.device.get_region_mmap(region.index);
+                    let old_mem_region = self.vm.make_user_memory_region(
+                        user_memory_region.slot,
+                        user_memory_region.start,
+                        user_memory_region.size,
+                        user_memory_region.host_addr,
+                        false,
+                        false,
+                    );
 
-                        // Remove old region
+                    self.vm
-                        let old_mem_region = self.vm.make_user_memory_region(
+                        .remove_user_memory_region(old_mem_region)
-                            mem_slot,
+                        .map_err(|e| io::Error::new(io::ErrorKind::Other, e))?;
-                            old_base + mmap_offset,
-                            mmap_size as u64,
-                            host_addr as u64,
-                            false,
-                            false,
-                        );
 
-                        self.vm
+                    // Update the user memory region with the correct start address.
-                            .remove_user_memory_region(old_mem_region)
+                    if new_base > old_base {
-                            .map_err(|e| io::Error::new(io::ErrorKind::Other, e))?;
+                        user_memory_region.start += new_base - old_base;
-
+                    } else {
-                        // Insert new region
+                        user_memory_region.start -= old_base - new_base;
-                        let new_mem_region = self.vm.make_user_memory_region(
-                            mem_slot,
-                            new_base + mmap_offset,
-                            mmap_size as u64,
-                            host_addr as u64,
-                            false,
-                            false,
-                        );
-
-                        self.vm
-                            .create_user_memory_region(new_mem_region)
-                            .map_err(|e| io::Error::new(io::ErrorKind::Other, e))?;
                     }
+
+                    // Insert new region
+                    let new_mem_region = self.vm.make_user_memory_region(
+                        user_memory_region.slot,
+                        user_memory_region.start,
+                        user_memory_region.size,
+                        user_memory_region.host_addr,
+                        false,
+                        false,
+                    );
+
+                    self.vm
+                        .create_user_memory_region(new_mem_region)
+                        .map_err(|e| io::Error::new(io::ErrorKind::Other, e))?;
                 }
             }
         }

pci/src/vfio_user.rs

@@ -6,7 +6,7 @@
 use crate::vfio::{Interrupt, Vfio, VfioCommon, VfioError};
 use crate::{BarReprogrammingParams, PciBarRegionType, VfioPciError};
 use crate::{
-    PciClassCode, PciConfiguration, PciDevice, PciDeviceError, PciHeaderType, PciSubclass,
+    PciBdf, PciClassCode, PciConfiguration, PciDevice, PciDeviceError, PciHeaderType, PciSubclass,
 };
 use hypervisor::HypervisorVmError;
 use std::any::Any;
@@ -18,7 +18,7 @@ use thiserror::Error;
 use vfio_bindings::bindings::vfio::*;
 use vfio_ioctls::VfioIrq;
 use vfio_user::{Client, Error as VfioUserError};
-use vm_allocator::SystemAllocator;
+use vm_allocator::{AddressAllocator, SystemAllocator};
 use vm_device::dma_mapping::ExternalDmaMapping;
 use vm_device::interrupt::{InterruptManager, InterruptSourceGroup, MsiIrqGroupConfig};
 use vm_device::BusDevice;
@@ -67,6 +67,7 @@ impl VfioUserPciDevice {
         client: Arc<Mutex<Client>>,
         msi_interrupt_manager: &Arc<dyn InterruptManager<GroupConfig = MsiIrqGroupConfig>>,
         legacy_interrupt_group: Option<Arc<dyn InterruptSourceGroup>>,
+        bdf: PciBdf,
     ) -> Result<Self, VfioUserPciDeviceError> {
         // This is used for the BAR and capabilities only
         let configuration = PciConfiguration::new(
@@ -104,7 +105,7 @@ impl VfioUserPciDevice {
             },
         };
 
-        common.parse_capabilities(msi_interrupt_manager, &vfio_wrapper);
+        common.parse_capabilities(msi_interrupt_manager, &vfio_wrapper, bdf);
         common
             .initialize_legacy_interrupt(legacy_interrupt_group, &vfio_wrapper)
             .map_err(VfioUserPciDeviceError::InitializeLegacyInterrupts)?;
@@ -116,215 +117,7 @@ impl VfioUserPciDevice {
             common,
         })
     }
-}
 
-impl BusDevice for VfioUserPciDevice {
-    fn read(&mut self, base: u64, offset: u64, data: &mut [u8]) {
-        self.read_bar(base, offset, data)
-    }
-
-    fn write(&mut self, base: u64, offset: u64, data: &[u8]) -> Option<Arc<Barrier>> {
-        self.write_bar(base, offset, data)
-    }
-}
-
-#[repr(u32)]
-#[derive(Debug, PartialEq, Eq, PartialOrd, Ord)]
-#[allow(dead_code)]
-enum Regions {
-    Bar0,
-    Bar1,
-    Bar2,
-    Bar3,
-    Bar4,
-    Bar5,
-    Rom,
-    Config,
-    Vga,
-    Migration,
-}
-
-struct VfioUserClientWrapper {
-    client: Arc<Mutex<Client>>,
-}
-
-impl Vfio for VfioUserClientWrapper {
-    fn region_read(&self, index: u32, offset: u64, data: &mut [u8]) {
-        self.client
-            .lock()
-            .unwrap()
-            .region_read(index, offset, data)
-            .ok();
-    }
-
-    fn region_write(&self, index: u32, offset: u64, data: &[u8]) {
-        self.client
-            .lock()
-            .unwrap()
-            .region_write(index, offset, data)
-            .ok();
-    }
-
-    fn get_irq_info(&self, irq_index: u32) -> Option<VfioIrq> {
-        self.client
-            .lock()
-            .unwrap()
-            .get_irq_info(irq_index)
-            .ok()
-            .map(|i| VfioIrq {
-                index: i.index,
-                flags: i.flags,
-                count: i.count,
-            })
-    }
-
-    fn enable_irq(&self, irq_index: u32, event_fds: Vec<&EventFd>) -> Result<(), VfioError> {
-        info!(
-            "Enabling IRQ {:x} number of fds = {:?}",
-            irq_index,
-            event_fds.len()
-        );
-        let fds: Vec<i32> = event_fds.iter().map(|e| e.as_raw_fd()).collect();
-
-        self.client
-            .lock()
-            .unwrap()
-            .set_irqs(
-                irq_index,
-                VFIO_IRQ_SET_DATA_EVENTFD | VFIO_IRQ_SET_ACTION_TRIGGER,
-                0,
-                event_fds.len() as u32,
-                &fds,
-            )
-            .map_err(VfioError::VfioUser)
-    }
-
-    fn disable_irq(&self, irq_index: u32) -> Result<(), VfioError> {
-        info!("Disabling IRQ {:x}", irq_index);
-        self.client
-            .lock()
-            .unwrap()
-            .set_irqs(
-                irq_index,
-                VFIO_IRQ_SET_DATA_NONE | VFIO_IRQ_SET_ACTION_TRIGGER,
-                0,
-                0,
-                &[],
-            )
-            .map_err(VfioError::VfioUser)
-    }
-
-    fn unmask_irq(&self, irq_index: u32) -> Result<(), VfioError> {
-        info!("Unmasking IRQ {:x}", irq_index);
-        self.client
-            .lock()
-            .unwrap()
-            .set_irqs(
-                irq_index,
-                VFIO_IRQ_SET_DATA_NONE | VFIO_IRQ_SET_ACTION_UNMASK,
-                0,
-                1,
-                &[],
-            )
-            .map_err(VfioError::VfioUser)
-    }
-}
-
-impl PciDevice for VfioUserPciDevice {
-    fn allocate_bars(
-        &mut self,
-        allocator: &mut SystemAllocator,
-    ) -> Result<Vec<(GuestAddress, GuestUsize, PciBarRegionType)>, PciDeviceError> {
-        self.common.allocate_bars(allocator, &self.vfio_wrapper)
-    }
-
-    fn free_bars(&mut self, allocator: &mut SystemAllocator) -> Result<(), PciDeviceError> {
-        self.common.free_bars(allocator)
-    }
-
-    fn as_any(&mut self) -> &mut dyn Any {
-        self
-    }
-
-    fn detect_bar_reprogramming(
-        &mut self,
-        reg_idx: usize,
-        data: &[u8],
-    ) -> Option<BarReprogrammingParams> {
-        self.common
-            .configuration
-            .detect_bar_reprogramming(reg_idx, data)
-    }
-
-    fn write_config_register(
-        &mut self,
-        reg_idx: usize,
-        offset: u64,
-        data: &[u8],
-    ) -> Option<Arc<Barrier>> {
-        self.common
-            .write_config_register(reg_idx, offset, data, &self.vfio_wrapper)
-    }
-
-    fn read_config_register(&mut self, reg_idx: usize) -> u32 {
-        self.common
-            .read_config_register(reg_idx, &self.vfio_wrapper)
-    }
-
-    fn read_bar(&mut self, base: u64, offset: u64, data: &mut [u8]) {
-        self.common.read_bar(base, offset, data, &self.vfio_wrapper)
-    }
-
-    fn write_bar(&mut self, base: u64, offset: u64, data: &[u8]) -> Option<Arc<Barrier>> {
-        self.common
-            .write_bar(base, offset, data, &self.vfio_wrapper)
-    }
-
-    fn move_bar(&mut self, old_base: u64, new_base: u64) -> Result<(), std::io::Error> {
-        info!("Moving BAR 0x{:x} -> 0x{:x}", old_base, new_base);
-        for mmio_region in self.common.mmio_regions.iter_mut() {
-            if mmio_region.start.raw_value() == old_base {
-                mmio_region.start = GuestAddress(new_base);
-
-                if let Some(mem_slot) = mmio_region.mem_slot {
-                    if let Some(host_addr) = mmio_region.host_addr {
-                        // Remove original region
-                        let old_region = self.vm.make_user_memory_region(
-                            mem_slot,
-                            old_base,
-                            mmio_region.length as u64,
-                            host_addr as u64,
-                            false,
-                            false,
-                        );
-
-                        self.vm
-                            .remove_user_memory_region(old_region)
-                            .map_err(|e| std::io::Error::new(std::io::ErrorKind::Other, e))?;
-
-                        let new_region = self.vm.make_user_memory_region(
-                            mem_slot,
-                            new_base,
-                            mmio_region.length as u64,
-                            host_addr as u64,
-                            false,
-                            false,
-                        );
-
-                        self.vm
-                            .create_user_memory_region(new_region)
-                            .map_err(|e| std::io::Error::new(std::io::ErrorKind::Other, e))?;
-                    }
-                }
-                info!("Moved bar 0x{:x} -> 0x{:x}", old_base, new_base);
-            }
-        }
-
-        Ok(())
-    }
-}
-
-impl VfioUserPciDevice {
     pub fn map_mmio_regions<F>(
         &mut self,
         vm: &Arc<dyn hypervisor::Vm>,
@@ -466,6 +259,234 @@ impl VfioUserPciDevice {
     }
 }
 
+impl BusDevice for VfioUserPciDevice {
+    fn read(&mut self, base: u64, offset: u64, data: &mut [u8]) {
+        self.read_bar(base, offset, data)
+    }
+
+    fn write(&mut self, base: u64, offset: u64, data: &[u8]) -> Option<Arc<Barrier>> {
+        self.write_bar(base, offset, data)
+    }
+}
+
+#[repr(u32)]
+#[derive(Debug, PartialEq, Eq, PartialOrd, Ord)]
+#[allow(dead_code)]
+enum Regions {
+    Bar0,
+    Bar1,
+    Bar2,
+    Bar3,
+    Bar4,
+    Bar5,
+    Rom,
+    Config,
+    Vga,
+    Migration,
+}
+
+struct VfioUserClientWrapper {
+    client: Arc<Mutex<Client>>,
+}
+
+impl Vfio for VfioUserClientWrapper {
+    fn region_read(&self, index: u32, offset: u64, data: &mut [u8]) {
+        self.client
+            .lock()
+            .unwrap()
+            .region_read(index, offset, data)
+            .ok();
+    }
+
+    fn region_write(&self, index: u32, offset: u64, data: &[u8]) {
+        self.client
+            .lock()
+            .unwrap()
+            .region_write(index, offset, data)
+            .ok();
+    }
+
+    fn get_irq_info(&self, irq_index: u32) -> Option<VfioIrq> {
+        self.client
+            .lock()
+            .unwrap()
+            .get_irq_info(irq_index)
+            .ok()
+            .map(|i| VfioIrq {
+                index: i.index,
+                flags: i.flags,
+                count: i.count,
+            })
+    }
+
+    fn enable_irq(&self, irq_index: u32, event_fds: Vec<&EventFd>) -> Result<(), VfioError> {
+        info!(
+            "Enabling IRQ {:x} number of fds = {:?}",
+            irq_index,
+            event_fds.len()
+        );
+        let fds: Vec<i32> = event_fds.iter().map(|e| e.as_raw_fd()).collect();
+
+        // Batch into blocks of 16 fds as sendmsg() has a size limit
+        let mut sent_fds = 0;
+        let num_fds = event_fds.len() as u32;
+        while sent_fds < num_fds {
+            let remaining_fds = num_fds - sent_fds;
+            let count = if remaining_fds > 16 {
+                16
+            } else {
+                remaining_fds
+            };
+
+            self.client
+                .lock()
+                .unwrap()
+                .set_irqs(
+                    irq_index,
+                    VFIO_IRQ_SET_DATA_EVENTFD | VFIO_IRQ_SET_ACTION_TRIGGER,
+                    sent_fds,
+                    count,
+                    &fds[sent_fds as usize..(sent_fds + count) as usize],
+                )
+                .map_err(VfioError::VfioUser)?;
+
+            sent_fds += count;
+        }
+
+        Ok(())
+    }
+
+    fn disable_irq(&self, irq_index: u32) -> Result<(), VfioError> {
+        info!("Disabling IRQ {:x}", irq_index);
+        self.client
+            .lock()
+            .unwrap()
+            .set_irqs(
+                irq_index,
+                VFIO_IRQ_SET_DATA_NONE | VFIO_IRQ_SET_ACTION_TRIGGER,
+                0,
+                0,
+                &[],
+            )
+            .map_err(VfioError::VfioUser)
+    }
+
+    fn unmask_irq(&self, irq_index: u32) -> Result<(), VfioError> {
+        info!("Unmasking IRQ {:x}", irq_index);
+        self.client
+            .lock()
+            .unwrap()
+            .set_irqs(
+                irq_index,
+                VFIO_IRQ_SET_DATA_NONE | VFIO_IRQ_SET_ACTION_UNMASK,
+                0,
+                1,
+                &[],
+            )
+            .map_err(VfioError::VfioUser)
+    }
+}
+
+impl PciDevice for VfioUserPciDevice {
+    fn allocate_bars(
+        &mut self,
+        allocator: &Arc<Mutex<SystemAllocator>>,
+        mmio_allocator: &mut AddressAllocator,
+    ) -> Result<Vec<(GuestAddress, GuestUsize, PciBarRegionType)>, PciDeviceError> {
+        self.common
+            .allocate_bars(allocator, mmio_allocator, &self.vfio_wrapper)
+    }
+
+    fn free_bars(
+        &mut self,
+        allocator: &mut SystemAllocator,
+        mmio_allocator: &mut AddressAllocator,
+    ) -> Result<(), PciDeviceError> {
+        self.common.free_bars(allocator, mmio_allocator)
+    }
+
+    fn as_any(&mut self) -> &mut dyn Any {
+        self
+    }
+
+    fn detect_bar_reprogramming(
+        &mut self,
+        reg_idx: usize,
+        data: &[u8],
+    ) -> Option<BarReprogrammingParams> {
+        self.common
+            .configuration
+            .detect_bar_reprogramming(reg_idx, data)
+    }
+
+    fn write_config_register(
+        &mut self,
+        reg_idx: usize,
+        offset: u64,
+        data: &[u8],
+    ) -> Option<Arc<Barrier>> {
+        self.common
+            .write_config_register(reg_idx, offset, data, &self.vfio_wrapper)
+    }
+
+    fn read_config_register(&mut self, reg_idx: usize) -> u32 {
+        self.common
+            .read_config_register(reg_idx, &self.vfio_wrapper)
+    }
+
+    fn read_bar(&mut self, base: u64, offset: u64, data: &mut [u8]) {
+        self.common.read_bar(base, offset, data, &self.vfio_wrapper)
+    }
+
+    fn write_bar(&mut self, base: u64, offset: u64, data: &[u8]) -> Option<Arc<Barrier>> {
+        self.common
+            .write_bar(base, offset, data, &self.vfio_wrapper)
+    }
+
+    fn move_bar(&mut self, old_base: u64, new_base: u64) -> Result<(), std::io::Error> {
+        info!("Moving BAR 0x{:x} -> 0x{:x}", old_base, new_base);
+        for mmio_region in self.common.mmio_regions.iter_mut() {
+            if mmio_region.start.raw_value() == old_base {
+                mmio_region.start = GuestAddress(new_base);
+
+                if let Some(mem_slot) = mmio_region.mem_slot {
+                    if let Some(host_addr) = mmio_region.host_addr {
+                        // Remove original region
+                        let old_region = self.vm.make_user_memory_region(
+                            mem_slot,
+                            old_base,
+                            mmio_region.length as u64,
+                            host_addr as u64,
+                            false,
+                            false,
+                        );
+
+                        self.vm
+                            .remove_user_memory_region(old_region)
+                            .map_err(|e| std::io::Error::new(std::io::ErrorKind::Other, e))?;
+
+                        let new_region = self.vm.make_user_memory_region(
+                            mem_slot,
+                            new_base,
+                            mmio_region.length as u64,
+                            host_addr as u64,
+                            false,
+                            false,
+                        );
+
+                        self.vm
+                            .create_user_memory_region(new_region)
+                            .map_err(|e| std::io::Error::new(std::io::ErrorKind::Other, e))?;
+                    }
+                }
+                info!("Moved bar 0x{:x} -> 0x{:x}", old_base, new_base);
+            }
+        }
+
+        Ok(())
+    }
+}
+
 impl Drop for VfioUserPciDevice {
     fn drop(&mut self) {
         self.unmap_mmio_regions();

performance-metrics/Cargo.toml

@@ -0,0 +1,19 @@
+[package]
+name = "performance-metrics"
+version = "0.1.0"
+authors = ["The Cloud Hypervisor Authors"]
+edition = "2018"
+build = "build.rs"
+
+[dependencies]
+clap = { version = "3.1.5", features = ["wrap_help","cargo"] }
+dirs = "4.0.0"
+serde = { version = "1.0.136", features = ["rc"] }
+serde_derive = "1.0.136"
+serde_json = "1.0.78"
+test_infra = { path = "../test_infra" }
+thiserror = "1.0.30"
+wait-timeout = "0.2.0"
+
+[build-dependencies]
+clap = { version = "3.1.5", features = ["wrap_help"] }

performance-metrics/build.rs

@@ -0,0 +1,26 @@
+// Copyright © 2020 Intel Corporation
+//
+// SPDX-License-Identifier: Apache-2.0
+//
+
+#[macro_use(crate_version)]
+extern crate clap;
+
+use std::process::Command;
+
+fn main() {
+    let mut git_human_readable = "v".to_owned() + crate_version!();
+    if let Ok(git_out) = Command::new("git").args(&["describe", "--dirty"]).output() {
+        if git_out.status.success() {
+            if let Ok(git_out_str) = String::from_utf8(git_out.stdout) {
+                git_human_readable = git_out_str;
+            }
+        }
+    }
+
+    // This println!() has a special behavior, as it will set the environment
+    // variable GIT_HUMAN_READABLE, so that it can be reused from the binary.
+    // Particularly, this is used from the main.rs to display the exact
+    // version information.
+    println!("cargo:rustc-env=GIT_HUMAN_READABLE={}", git_human_readable);
+}

performance-metrics/src/main.rs

@@ -0,0 +1,474 @@
+// Copyright © 2022 Intel Corporation
+//
+// SPDX-License-Identifier: Apache-2.0
+//
+
+// Custom harness to run performance tests
+extern crate test_infra;
+#[macro_use(crate_authors)]
+extern crate clap;
+
+mod performance_tests;
+
+use clap::{Arg, Command as ClapCommand};
+use performance_tests::*;
+use serde_derive::{Deserialize, Serialize};
+use std::{env, fmt, process::Command, sync::mpsc::channel, thread, time::Duration};
+use thiserror::Error;
+
+#[derive(Error, Debug)]
+enum Error {
+    #[error("Error: test timed-out")]
+    TestTimeout,
+    #[error("Error: test failed")]
+    TestFailed,
+    #[error("Error creating log file: {0}")]
+    ReportFileCreation(std::io::Error),
+    #[error("Error writing log file: {0}")]
+    ReportFileWrite(std::io::Error),
+}
+
+#[derive(Deserialize, Serialize)]
+pub struct PerformanceTestResult {
+    name: String,
+    mean: f64,
+    std_dev: f64,
+    max: f64,
+    min: f64,
+}
+
+#[derive(Deserialize, Serialize)]
+pub struct MetricsReport {
+    pub git_human_readable: String,
+    pub git_revision: String,
+    pub git_commit_date: String,
+    pub date: String,
+    pub results: Vec<PerformanceTestResult>,
+}
+
+impl Default for MetricsReport {
+    fn default() -> Self {
+        let mut git_human_readable = "".to_string();
+        if let Ok(git_out) = Command::new("git").args(&["describe", "--dirty"]).output() {
+            if git_out.status.success() {
+                if let Ok(git_out_str) = String::from_utf8(git_out.stdout) {
+                    git_human_readable = git_out_str.trim().to_string();
+                }
+            }
+        }
+
+        let mut git_revision = "".to_string();
+        if let Ok(git_out) = Command::new("git").args(&["rev-parse", "HEAD"]).output() {
+            if git_out.status.success() {
+                if let Ok(git_out_str) = String::from_utf8(git_out.stdout) {
+                    git_revision = git_out_str.trim().to_string();
+                }
+            }
+        }
+
+        let mut git_commit_date = "".to_string();
+        if let Ok(git_out) = Command::new("git")
+            .args(&["show", "-s", "--format=%cd"])
+            .output()
+        {
+            if git_out.status.success() {
+                if let Ok(git_out_str) = String::from_utf8(git_out.stdout) {
+                    git_commit_date = git_out_str.trim().to_string();
+                }
+            }
+        }
+
+        MetricsReport {
+            git_human_readable,
+            git_revision,
+            git_commit_date,
+            date: date(),
+            results: Vec::new(),
+        }
+    }
+}
+
+pub struct PerformanceTestControl {
+    test_timeout: u32,
+    test_iterations: u32,
+    num_queues: Option<u32>,
+    queue_size: Option<u32>,
+    net_rx: Option<bool>,
+    fio_ops: Option<FioOps>,
+}
+
+impl fmt::Display for PerformanceTestControl {
+    fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result {
+        let mut output = format!(
+            "test_timeout = {}s, test_iterations = {}",
+            self.test_timeout, self.test_iterations
+        );
+        if let Some(o) = self.num_queues {
+            output = format!("{}, num_queues = {}", output, o);
+        }
+        if let Some(o) = self.queue_size {
+            output = format!("{}, queue_size = {}", output, o);
+        }
+        if let Some(o) = self.net_rx {
+            output = format!("{}, net_rx = {}", output, o);
+        }
+        if let Some(o) = &self.fio_ops {
+            output = format!("{}, fio_ops = {}", output, o);
+        }
+
+        write!(f, "{}", output)
+    }
+}
+
+impl PerformanceTestControl {
+    const fn default() -> Self {
+        Self {
+            test_timeout: 10,
+            test_iterations: 5,
+            num_queues: None,
+            queue_size: None,
+            net_rx: None,
+            fio_ops: None,
+        }
+    }
+}
+
+/// A performance test should finish within the a certain time-out and
+/// return a performance metrics number (including the average number and
+/// standard deviation)
+struct PerformanceTest {
+    pub name: &'static str,
+    pub func_ptr: fn(&PerformanceTestControl) -> f64,
+    pub control: PerformanceTestControl,
+}
+
+impl PerformanceTest {
+    pub fn run(&self) -> PerformanceTestResult {
+        let mut metrics = Vec::new();
+        for _ in 0..self.control.test_iterations {
+            metrics.push((self.func_ptr)(&self.control));
+        }
+
+        let mean = mean(&metrics).unwrap();
+        let std_dev = std_deviation(&metrics).unwrap();
+        let max = metrics.clone().into_iter().reduce(f64::max).unwrap();
+        let min = metrics.clone().into_iter().reduce(f64::min).unwrap();
+
+        PerformanceTestResult {
+            name: self.name.to_string(),
+            mean,
+            std_dev,
+            max,
+            min,
+        }
+    }
+
+    // Calculate the timeout for each test
+    // Note: To cover the setup/cleanup time, 20s is added for each iteration of the test
+    pub fn calc_timeout(&self) -> u64 {
+        ((self.control.test_timeout + 20) * self.control.test_iterations) as u64
+    }
+}
+
+fn mean(data: &[f64]) -> Option<f64> {
+    let count = data.len();
+
+    if count > 0 {
+        Some(data.iter().sum::<f64>() / count as f64)
+    } else {
+        None
+    }
+}
+
+fn std_deviation(data: &[f64]) -> Option<f64> {
+    let count = data.len();
+
+    if count > 0 {
+        let mean = mean(data).unwrap();
+        let variance = data
+            .iter()
+            .map(|value| {
+                let diff = mean - *value;
+                diff * diff
+            })
+            .sum::<f64>()
+            / count as f64;
+
+        Some(variance.sqrt())
+    } else {
+        None
+    }
+}
+
+const TEST_LIST: [PerformanceTest; 15] = [
+    PerformanceTest {
+        name: "boot_time_s",
+        func_ptr: performance_boot_time,
+        control: PerformanceTestControl {
+            test_timeout: 2,
+            test_iterations: 10,
+            ..PerformanceTestControl::default()
+        },
+    },
+    PerformanceTest {
+        name: "boot_time_pmem_s",
+        func_ptr: performance_boot_time_pmem,
+        control: PerformanceTestControl {
+            test_timeout: 2,
+            test_iterations: 10,
+            ..PerformanceTestControl::default()
+        },
+    },
+    PerformanceTest {
+        name: "virtio_net_latency_us",
+        func_ptr: performance_net_latency,
+        control: PerformanceTestControl {
+            num_queues: Some(2),
+            queue_size: Some(256),
+            ..PerformanceTestControl::default()
+        },
+    },
+    PerformanceTest {
+        name: "virtio_net_throughput_single_queue_rx_bps",
+        func_ptr: performance_net_throughput,
+        control: PerformanceTestControl {
+            num_queues: Some(2),
+            queue_size: Some(256),
+            net_rx: Some(true),
+            ..PerformanceTestControl::default()
+        },
+    },
+    PerformanceTest {
+        name: "virtio_net_throughput_single_queue_tx_bps",
+        func_ptr: performance_net_throughput,
+        control: PerformanceTestControl {
+            num_queues: Some(2),
+            queue_size: Some(256),
+            net_rx: Some(false),
+            ..PerformanceTestControl::default()
+        },
+    },
+    PerformanceTest {
+        name: "virtio_net_throughput_multi_queue_rx_bps",
+        func_ptr: performance_net_throughput,
+        control: PerformanceTestControl {
+            num_queues: Some(4),
+            queue_size: Some(256),
+            net_rx: Some(true),
+            ..PerformanceTestControl::default()
+        },
+    },
+    PerformanceTest {
+        name: "virtio_net_throughput_multi_queue_tx_bps",
+        func_ptr: performance_net_throughput,
+        control: PerformanceTestControl {
+            num_queues: Some(4),
+            queue_size: Some(256),
+            net_rx: Some(false),
+            ..PerformanceTestControl::default()
+        },
+    },
+    PerformanceTest {
+        name: "block_read_Bps",
+        func_ptr: performance_block_io,
+        control: PerformanceTestControl {
+            num_queues: Some(1),
+            queue_size: Some(128),
+            fio_ops: Some(FioOps::Read),
+            ..PerformanceTestControl::default()
+        },
+    },
+    PerformanceTest {
+        name: "block_write_Bps",
+        func_ptr: performance_block_io,
+        control: PerformanceTestControl {
+            num_queues: Some(1),
+            queue_size: Some(128),
+            fio_ops: Some(FioOps::Write),
+            ..PerformanceTestControl::default()
+        },
+    },
+    PerformanceTest {
+        name: "block_random_read_Bps",
+        func_ptr: performance_block_io,
+        control: PerformanceTestControl {
+            num_queues: Some(1),
+            queue_size: Some(128),
+            fio_ops: Some(FioOps::RandomRead),
+            ..PerformanceTestControl::default()
+        },
+    },
+    PerformanceTest {
+        name: "block_random_write_Bps",
+        func_ptr: performance_block_io,
+        control: PerformanceTestControl {
+            num_queues: Some(1),
+            queue_size: Some(128),
+            fio_ops: Some(FioOps::RandomWrite),
+            ..PerformanceTestControl::default()
+        },
+    },
+    PerformanceTest {
+        name: "block_multi_queue_read_Bps",
+        func_ptr: performance_block_io,
+        control: PerformanceTestControl {
+            num_queues: Some(2),
+            queue_size: Some(128),
+            fio_ops: Some(FioOps::Read),
+            ..PerformanceTestControl::default()
+        },
+    },
+    PerformanceTest {
+        name: "block_multi_queue_write_Bps",
+        func_ptr: performance_block_io,
+        control: PerformanceTestControl {
+            num_queues: Some(2),
+            queue_size: Some(128),
+            fio_ops: Some(FioOps::Write),
+            ..PerformanceTestControl::default()
+        },
+    },
+    PerformanceTest {
+        name: "block_multi_queue_random_read_Bps",
+        func_ptr: performance_block_io,
+        control: PerformanceTestControl {
+            num_queues: Some(2),
+            queue_size: Some(128),
+            fio_ops: Some(FioOps::RandomRead),
+            ..PerformanceTestControl::default()
+        },
+    },
+    PerformanceTest {
+        name: "block_multi_queue_random_write_Bps",
+        func_ptr: performance_block_io,
+        control: PerformanceTestControl {
+            num_queues: Some(2),
+            queue_size: Some(128),
+            fio_ops: Some(FioOps::RandomWrite),
+            ..PerformanceTestControl::default()
+        },
+    },
+];
+
+fn run_test_with_timetout(test: &'static PerformanceTest) -> Result<PerformanceTestResult, Error> {
+    let (sender, receiver) = channel::<Result<PerformanceTestResult, Error>>();
+    thread::spawn(move || {
+        println!("Test '{}' running .. ({})", test.name, test.control);
+
+        let output = match std::panic::catch_unwind(|| test.run()) {
+            Ok(test_result) => {
+                println!(
+                    "Test '{}' .. ok: mean = {}, std_dev = {}",
+                    test_result.name, test_result.mean, test_result.std_dev
+                );
+                Ok(test_result)
+            }
+            Err(_) => Err(Error::TestFailed),
+        };
+
+        let _ = sender.send(output);
+    });
+
+    // Todo: Need to cleanup/kill all hanging child processes
+    let test_timeout = test.calc_timeout();
+    receiver
+        .recv_timeout(Duration::from_secs(test_timeout))
+        .map_err(|_| {
+            eprintln!(
+                "[Error] Test '{}' time-out after {} seconds",
+                test.name, test_timeout
+            );
+            Error::TestTimeout
+        })?
+}
+
+fn date() -> String {
+    let output = test_infra::exec_host_command_output("date");
+    String::from_utf8_lossy(&output.stdout).trim().to_string()
+}
+
+fn main() {
+    let cmd_arguments = ClapCommand::new("performance-metrics")
+        .version(env!("GIT_HUMAN_READABLE"))
+        .author(crate_authors!())
+        .about("Generate the performance metrics data for Cloud Hypervisor")
+        .arg(
+            Arg::new("test-filter")
+                .long("test-filter")
+                .help("Filter metrics tests to run based on provided keywords")
+                .multiple_occurrences(true)
+                .takes_value(true)
+                .required(false),
+        )
+        .arg(
+            Arg::new("list-tests")
+                .long("list-tests")
+                .help("Print the list of availale metrics tests")
+                .multiple_occurrences(true)
+                .takes_value(false)
+                .required(false),
+        )
+        .arg(
+            Arg::new("report-file")
+                .long("report-file")
+                .help("Report file. Standard error is used if not specified")
+                .takes_value(true),
+        )
+        .get_matches();
+
+    if cmd_arguments.is_present("list-tests") {
+        println!("List of available metrics tests:\n");
+        for test in TEST_LIST.iter() {
+            println!("\"{}\" ({})", test.name, test.control);
+        }
+
+        return;
+    }
+
+    let test_filter = match cmd_arguments.values_of("test-filter") {
+        Some(s) => s.collect(),
+        None => Vec::new(),
+    };
+
+    let mut report_file: Box<dyn std::io::Write + Send> =
+        if let Some(file) = cmd_arguments.value_of("report-file") {
+            Box::new(
+                std::fs::File::create(std::path::Path::new(file))
+                    .map_err(Error::ReportFileCreation)
+                    .unwrap(),
+            )
+        } else {
+            Box::new(std::io::stderr())
+        };
+
+    // Run performance tests sequentially and report results (in both readable/json format)
+    let mut metrics_report: MetricsReport = Default::default();
+
+    init_tests();
+
+    for test in TEST_LIST.iter() {
+        if test_filter.is_empty() || test_filter.iter().any(|&s| test.name.contains(s)) {
+            match run_test_with_timetout(test) {
+                Ok(r) => {
+                    metrics_report.results.push(r);
+                }
+                Err(e) => {
+                    eprintln!("Aborting test due to error: '{:?}'", e);
+                    break;
+                }
+            };
+        }
+    }
+
+    cleanup_tests();
+
+    // Todo: Report/upload to the metrics database
+    report_file
+        .write(
+            serde_json::to_string_pretty(&metrics_report)
+                .unwrap()
+                .as_bytes(),
+        )
+        .map_err(Error::ReportFileWrite)
+        .unwrap();
+}

performance-metrics/src/performance_tests.rs

@@ -0,0 +1,834 @@
+// Copyright © 2022 Intel Corporation
+//
+// SPDX-License-Identifier: Apache-2.0
+//
+
+// Performance tests
+
+use crate::{mean, PerformanceTestControl};
+use serde_json::Value;
+use std::path::{Path, PathBuf};
+use std::process::{Child, Command, Stdio};
+use std::string::String;
+use std::thread;
+use std::time::Duration;
+use std::{fmt, fs};
+use test_infra::Error as InfraError;
+use test_infra::*;
+use wait_timeout::ChildExt;
+
+pub const FOCAL_IMAGE_NAME: &str = "focal-server-cloudimg-amd64-custom-20210609-0.raw";
+
+#[derive(Debug)]
+enum WaitTimeoutError {
+    Timedout,
+    ExitStatus,
+    General(std::io::Error),
+}
+
+#[derive(Debug)]
+enum Error {
+    BootTimeParse,
+    EthrLogFile(std::io::Error),
+    EthrLogParse,
+    FioOutputParse,
+    Iperf3Parse,
+    Infra(InfraError),
+    Spawn(std::io::Error),
+    Scp(SshCommandError),
+    WaitTimeout(WaitTimeoutError),
+}
+
+impl From<InfraError> for Error {
+    fn from(e: InfraError) -> Self {
+        Self::Infra(e)
+    }
+}
+
+const BLK_IO_TEST_IMG: &str = "/var/tmp/ch-blk-io-test.img";
+
+pub fn init_tests() {
+    // The test image can not be created on tmpfs (e.g. /tmp) filesystem,
+    // as tmpfs does not support O_DIRECT
+    assert!(exec_host_command_output(&format!(
+        "dd if=/dev/zero of={} bs=1M count=4096",
+        BLK_IO_TEST_IMG
+    ))
+    .status
+    .success());
+}
+
+pub fn cleanup_tests() {
+    fs::remove_file(BLK_IO_TEST_IMG)
+        .unwrap_or_else(|_| panic!("Failed to remove file '{}'.", BLK_IO_TEST_IMG));
+}
+
+// Performance tests are expected to be executed sequentially, so we can
+// start VM guests with the same IP while putting them on a different
+// private network. The default constructor "Guest::new()" does not work
+// well, as we can easily create more than 256 VMs from repeating various
+// performance tests dozens times in a single run.
+fn performance_test_new_guest(disk_config: Box<dyn DiskConfig>) -> Guest {
+    Guest::new_from_ip_range(disk_config, "172.19", 0)
+}
+
+const DIRECT_KERNEL_BOOT_CMDLINE: &str =
+    "root=/dev/vda1 console=hvc0 rw systemd.journald.forward_to_console=1";
+
+// Creates the path for direct kernel boot and return the path.
+// For x86_64, this function returns the vmlinux kernel path.
+// For AArch64, this function returns the PE kernel path.
+fn direct_kernel_boot_path() -> PathBuf {
+    let mut workload_path = dirs::home_dir().unwrap();
+    workload_path.push("workloads");
+
+    let mut kernel_path = workload_path;
+    #[cfg(target_arch = "x86_64")]
+    kernel_path.push("vmlinux");
+    #[cfg(target_arch = "aarch64")]
+    kernel_path.push("Image");
+
+    kernel_path
+}
+
+// Wait the child process for a given timeout
+fn child_wait_timeout(child: &mut Child, timeout: u64) -> Result<(), WaitTimeoutError> {
+    match child.wait_timeout(Duration::from_secs(timeout)) {
+        Err(e) => {
+            return Err(WaitTimeoutError::General(e));
+        }
+        Ok(s) => match s {
+            None => {
+                return Err(WaitTimeoutError::Timedout);
+            }
+            Some(s) => {
+                if !s.success() {
+                    return Err(WaitTimeoutError::ExitStatus);
+                }
+            }
+        },
+    }
+
+    Ok(())
+}
+
+fn parse_iperf3_output(output: &[u8], sender: bool) -> Result<f64, Error> {
+    std::panic::catch_unwind(|| {
+        let s = String::from_utf8_lossy(output);
+        let v: Value = serde_json::from_str(&s).expect("'iperf3' parse error: invalid json output");
+
+        let bps: f64 = if sender {
+            v["end"]["sum_sent"]["bits_per_second"]
+                .as_f64()
+                .expect("'iperf3' parse error: missing entry 'end.sum_sent.bits_per_second'")
+        } else {
+            v["end"]["sum_received"]["bits_per_second"]
+                .as_f64()
+                .expect("'iperf3' parse error: missing entry 'end.sum_received.bits_per_second'")
+        };
+
+        bps
+    })
+    .map_err(|_| {
+        eprintln!(
+            "=============== iperf3 output ===============\n\n{}\n\n===========end============\n\n",
+            String::from_utf8_lossy(output)
+        );
+        Error::Iperf3Parse
+    })
+}
+
+fn measure_virtio_net_throughput(
+    test_timeout: u32,
+    queue_pairs: u32,
+    guest: &Guest,
+    receive: bool,
+) -> Result<f64, Error> {
+    let default_port = 5201;
+
+    // 1. start the iperf3 server on the guest
+    for n in 0..queue_pairs {
+        guest
+            .ssh_command(&format!("iperf3 -s -p {} -D", default_port + n))
+            .map_err(InfraError::SshCommand)?;
+    }
+
+    thread::sleep(Duration::new(1, 0));
+
+    // 2. start the iperf3 client on host to measure RX through-put
+    let mut clients = Vec::new();
+    for n in 0..queue_pairs {
+        let mut cmd = Command::new("iperf3");
+        cmd.args(&[
+            "-J", // Output in JSON format
+            "-c",
+            &guest.network.guest_ip,
+            "-p",
+            &format!("{}", default_port + n),
+            "-t",
+            &format!("{}", test_timeout),
+        ]);
+        // For measuring the guest transmit throughput (as a sender),
+        // use reverse mode of the iperf3 client on the host
+        if !receive {
+            cmd.args(&["-R"]);
+        }
+        let client = cmd
+            .stderr(Stdio::piped())
+            .stdout(Stdio::piped())
+            .spawn()
+            .map_err(Error::Spawn)?;
+
+        clients.push(client);
+    }
+
+    let mut err: Option<Error> = None;
+    let mut bps = Vec::new();
+    let mut failed = false;
+    for c in clients {
+        let mut c = c;
+        if let Err(e) = child_wait_timeout(&mut c, test_timeout as u64 + 5) {
+            err = Some(Error::WaitTimeout(e));
+            failed = true;
+        }
+
+        if !failed {
+            // Safe to unwrap as we know the child has terminated succesffully
+            let output = c.wait_with_output().unwrap();
+            bps.push(parse_iperf3_output(&output.stdout, receive)?);
+        } else {
+            let _ = c.kill();
+            let output = c.wait_with_output().unwrap();
+            println!(
+                "=============== Client output [Error] ===============\n\n{}\n\n===========end============\n\n",
+                String::from_utf8_lossy(&output.stdout)
+            );
+        }
+    }
+
+    if let Some(e) = err {
+        Err(e)
+    } else {
+        Ok(bps.iter().sum())
+    }
+}
+
+pub fn performance_net_throughput(control: &PerformanceTestControl) -> f64 {
+    let test_timeout = control.test_timeout;
+    let rx = control.net_rx.unwrap();
+
+    let focal = UbuntuDiskConfig::new(FOCAL_IMAGE_NAME.to_string());
+    let guest = performance_test_new_guest(Box::new(focal));
+
+    let num_queues = control.num_queues.unwrap();
+    let queue_size = control.queue_size.unwrap();
+    let net_params = format!(
+        "tap=,mac={},ip={},mask=255.255.255.0,num_queues={},queue_size={}",
+        guest.network.guest_mac, guest.network.host_ip, num_queues, queue_size,
+    );
+
+    let mut child = GuestCommand::new(&guest)
+        .args(&["--cpus", &format!("boot={}", num_queues)])
+        .args(&["--memory", "size=4G"])
+        .args(&["--kernel", direct_kernel_boot_path().to_str().unwrap()])
+        .args(&["--cmdline", DIRECT_KERNEL_BOOT_CMDLINE])
+        .default_disks()
+        .args(&["--net", net_params.as_str()])
+        .capture_output()
+        .set_print_cmd(false)
+        .spawn()
+        .unwrap();
+
+    let r = std::panic::catch_unwind(|| {
+        guest.wait_vm_boot(None).unwrap();
+        measure_virtio_net_throughput(test_timeout, num_queues / 2, &guest, rx).unwrap()
+    });
+
+    let _ = child.kill();
+    let output = child.wait_with_output().unwrap();
+
+    match r {
+        Ok(r) => r,
+        Err(e) => {
+            handle_child_output(Err(e), &output);
+            panic!("test failed!");
+        }
+    }
+}
+
+fn parse_ethr_latency_output(output: &[u8]) -> Result<Vec<f64>, Error> {
+    std::panic::catch_unwind(|| {
+        let s = String::from_utf8_lossy(output);
+        let mut latency = Vec::new();
+        for l in s.lines() {
+            let v: Value = serde_json::from_str(l).expect("'ethr' parse error: invalid json line");
+            // Skip header/summary lines
+            if let Some(avg) = v["Avg"].as_str() {
+                // Assume the latency unit is always "us"
+                latency.push(
+                    avg.split("us").collect::<Vec<&str>>()[0]
+                        .parse::<f64>()
+                        .expect("'ethr' parse error: invalid 'Avg' entry"),
+                );
+            }
+        }
+
+        assert!(
+            !latency.is_empty(),
+            "'ethr' parse error: no valid latency data found"
+        );
+
+        latency
+    })
+    .map_err(|_| {
+        eprintln!(
+            "=============== ethr output ===============\n\n{}\n\n===========end============\n\n",
+            String::from_utf8_lossy(output)
+        );
+        Error::EthrLogParse
+    })
+}
+
+fn measure_virtio_net_latency(guest: &Guest, test_timeout: u32) -> Result<Vec<f64>, Error> {
+    // copy the 'ethr' tool to the guest image
+    let ethr_path = "/usr/local/bin/ethr";
+    let ethr_remote_path = "/tmp/ethr";
+    scp_to_guest(
+        Path::new(ethr_path),
+        Path::new(ethr_remote_path),
+        &guest.network.guest_ip,
+        //DEFAULT_SSH_RETRIES,
+        1,
+        DEFAULT_SSH_TIMEOUT,
+    )
+    .map_err(Error::Scp)?;
+
+    // Start the ethr server on the guest
+    guest
+        .ssh_command(&format!("{} -s &> /dev/null &", ethr_remote_path))
+        .map_err(InfraError::SshCommand)?;
+
+    thread::sleep(Duration::new(1, 0));
+
+    // Start the ethr client on the host
+    let log_file = guest
+        .tmp_dir
+        .as_path()
+        .join("ethr.client.log")
+        .to_str()
+        .unwrap()
+        .to_string();
+    let mut c = Command::new(ethr_path)
+        .args(&[
+            "-c",
+            &guest.network.guest_ip,
+            "-t",
+            "l",
+            "-o",
+            &log_file, // file output is JSON format
+            "-d",
+            &format!("{}s", test_timeout),
+        ])
+        .stderr(Stdio::piped())
+        .stdout(Stdio::piped())
+        .spawn()
+        .map_err(Error::Spawn)?;
+
+    if let Err(e) = child_wait_timeout(&mut c, test_timeout as u64 + 5).map_err(Error::WaitTimeout)
+    {
+        let _ = c.kill();
+        return Err(e);
+    }
+
+    // Parse the ethr latency test output
+    let content = fs::read(log_file).map_err(Error::EthrLogFile)?;
+    parse_ethr_latency_output(&content)
+}
+
+pub fn performance_net_latency(control: &PerformanceTestControl) -> f64 {
+    let focal = UbuntuDiskConfig::new(FOCAL_IMAGE_NAME.to_string());
+    let guest = performance_test_new_guest(Box::new(focal));
+
+    let num_queues = control.num_queues.unwrap();
+    let queue_size = control.queue_size.unwrap();
+    let net_params = format!(
+        "tap=,mac={},ip={},mask=255.255.255.0,num_queues={},queue_size={}",
+        guest.network.guest_mac, guest.network.host_ip, num_queues, queue_size,
+    );
+
+    let mut child = GuestCommand::new(&guest)
+        .args(&["--cpus", &format!("boot={}", num_queues)])
+        .args(&["--memory", "size=4G"])
+        .args(&["--kernel", direct_kernel_boot_path().to_str().unwrap()])
+        .args(&["--cmdline", DIRECT_KERNEL_BOOT_CMDLINE])
+        .default_disks()
+        .args(&["--net", net_params.as_str()])
+        .capture_output()
+        .set_print_cmd(false)
+        .spawn()
+        .unwrap();
+
+    let r = std::panic::catch_unwind(|| {
+        guest.wait_vm_boot(None).unwrap();
+
+        // 'ethr' tool will measure the latency multiple times with provided test time
+        let latency = measure_virtio_net_latency(&guest, control.test_timeout).unwrap();
+        mean(&latency).unwrap()
+    });
+
+    let _ = child.kill();
+    let output = child.wait_with_output().unwrap();
+
+    match r {
+        Ok(r) => r,
+        Err(e) => {
+            handle_child_output(Err(e), &output);
+            panic!("test failed!");
+        }
+    }
+}
+
+fn parse_boot_time_output(output: &[u8]) -> Result<f64, Error> {
+    std::panic::catch_unwind(|| {
+        let l: Vec<String> = String::from_utf8_lossy(output)
+            .lines()
+            .into_iter()
+            .filter(|l| l.contains("Debug I/O port: Kernel code"))
+            .map(|l| l.to_string())
+            .collect();
+
+        assert_eq!(
+            l.len(),
+            2,
+            "Expecting two matching lines for 'Debug I/O port: Kernel code'"
+        );
+
+        let time_stamp_kernel_start = {
+            let s = l[0].split("--").collect::<Vec<&str>>();
+            assert_eq!(
+                s.len(),
+                2,
+                "Expecting '--' for the matching line of 'Debug I/O port' output"
+            );
+
+            // Sample output: "[Debug I/O port: Kernel code 0x40] 0.096537 seconds"
+            assert!(
+                s[1].contains("0x40"),
+                "Expecting kernel code '0x40' for 'linux_kernel_start' time stamp output"
+            );
+            let t = s[1].split_whitespace().collect::<Vec<&str>>();
+            assert_eq!(
+                t.len(),
+                8,
+                "Expecting exact '8' words from the 'Debug I/O port' output"
+            );
+            assert!(
+                t[7].eq("seconds"),
+                "Expecting 'seconds' as the the last word of the 'Debug I/O port' output"
+            );
+
+            t[6].parse::<f64>().unwrap()
+        };
+
+        let time_stamp_user_start = {
+            let s = l[1].split("--").collect::<Vec<&str>>();
+            assert_eq!(
+                s.len(),
+                2,
+                "Expecting '--' for the matching line of 'Debug I/O port' output"
+            );
+
+            // Sample output: "Debug I/O port: Kernel code 0x41] 0.198980 seconds"
+            assert!(
+                s[1].contains("0x41"),
+                "Expecting kernel code '0x41' for 'linux_kernel_start' time stamp output"
+            );
+            let t = s[1].split_whitespace().collect::<Vec<&str>>();
+            assert_eq!(
+                t.len(),
+                8,
+                "Expecting exact '8' words from the 'Debug I/O port' output"
+            );
+            assert!(
+                t[7].eq("seconds"),
+                "Expecting 'seconds' as the the last word of the 'Debug I/O port' output"
+            );
+
+            t[6].parse::<f64>().unwrap()
+        };
+
+        time_stamp_user_start - time_stamp_kernel_start
+    })
+    .map_err(|_| {
+        eprintln!(
+            "=============== boot-time output ===============\n\n{}\n\n===========end============\n\n",
+            String::from_utf8_lossy(output)
+        );
+        Error::BootTimeParse
+    })
+}
+
+fn measure_boot_time(cmd: &mut GuestCommand, test_timeout: u32) -> Result<f64, Error> {
+    let mut child = cmd.capture_output().set_print_cmd(false).spawn().unwrap();
+
+    thread::sleep(Duration::new(test_timeout as u64, 0));
+    let _ = child.kill();
+    let output = child.wait_with_output().unwrap();
+
+    parse_boot_time_output(&output.stderr).map_err(|e| {
+        eprintln!(
+            "\n\n==== Start child stdout ====\n\n{}\n\n==== End child stdout ====",
+            String::from_utf8_lossy(&output.stdout)
+        );
+        eprintln!(
+            "\n\n==== Start child stderr ====\n\n{}\n\n==== End child stderr ====",
+            String::from_utf8_lossy(&output.stderr)
+        );
+
+        e
+    })
+}
+
+pub fn performance_boot_time(control: &PerformanceTestControl) -> f64 {
+    let r = std::panic::catch_unwind(|| {
+        let focal = UbuntuDiskConfig::new(FOCAL_IMAGE_NAME.to_string());
+        let guest = performance_test_new_guest(Box::new(focal));
+        let mut cmd = GuestCommand::new(&guest);
+
+        let c = cmd
+            .args(&["--memory", "size=1G"])
+            .args(&["--kernel", direct_kernel_boot_path().to_str().unwrap()])
+            .args(&["--cmdline", DIRECT_KERNEL_BOOT_CMDLINE])
+            .args(&["--console", "off"])
+            .default_disks();
+
+        measure_boot_time(c, control.test_timeout).unwrap()
+    });
+
+    match r {
+        Ok(r) => r,
+        Err(_) => {
+            panic!("test failed!");
+        }
+    }
+}
+
+pub fn performance_boot_time_pmem(control: &PerformanceTestControl) -> f64 {
+    let r = std::panic::catch_unwind(|| {
+        let focal = UbuntuDiskConfig::new(FOCAL_IMAGE_NAME.to_string());
+        let guest = performance_test_new_guest(Box::new(focal));
+        let mut cmd = GuestCommand::new(&guest);
+        let c = cmd
+            .args(&["--memory", "size=1G,hugepages=on"])
+            .args(&["--kernel", direct_kernel_boot_path().to_str().unwrap()])
+            .args(&["--cmdline", "root=/dev/pmem0p1 console=ttyS0 quiet rw"])
+            .args(&["--console", "off"])
+            .args(&[
+                "--pmem",
+                format!(
+                    "file={}",
+                    guest.disk_config.disk(DiskType::OperatingSystem).unwrap()
+                )
+                .as_str(),
+            ]);
+
+        measure_boot_time(c, control.test_timeout).unwrap()
+    });
+
+    match r {
+        Ok(r) => r,
+        Err(_) => {
+            panic!("test failed!");
+        }
+    }
+}
+
+pub enum FioOps {
+    Read,
+    RandomRead,
+    Write,
+    RandomWrite,
+}
+
+impl fmt::Display for FioOps {
+    fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result {
+        match self {
+            FioOps::Read => write!(f, "read"),
+            FioOps::RandomRead => write!(f, "randread"),
+            FioOps::Write => write!(f, "write"),
+            FioOps::RandomWrite => write!(f, "randwrite"),
+        }
+    }
+}
+
+fn parse_fio_output(output: &str, fio_ops: &FioOps, num_jobs: u32) -> Result<f64, Error> {
+    std::panic::catch_unwind(|| {
+        let v: Value =
+            serde_json::from_str(output).expect("'fio' parse error: invalid json output");
+        let jobs = v["jobs"]
+            .as_array()
+            .expect("'fio' parse error: missing entry 'jobs'");
+        assert_eq!(
+            jobs.len(),
+            num_jobs as usize,
+            "'fio' parse error: Unexpected number of 'fio' jobs."
+        );
+
+        let read = match fio_ops {
+            FioOps::Read | FioOps::RandomRead => true,
+            FioOps::Write | FioOps::RandomWrite => false,
+        };
+
+        let mut total_bps = 0_f64;
+        for j in jobs {
+            if read {
+                let bytes = j["read"]["io_bytes"]
+                    .as_u64()
+                    .expect("'fio' parse error: missing entry 'read.io_bytes'");
+                let runtime = j["read"]["runtime"]
+                    .as_u64()
+                    .expect("'fio' parse error: missing entry 'read.runtime'")
+                    as f64
+                    / 1000_f64;
+                total_bps += bytes as f64 / runtime as f64;
+            } else {
+                let bytes = j["write"]["io_bytes"]
+                    .as_u64()
+                    .expect("'fio' parse error: missing entry 'write.io_bytes'");
+                let runtime = j["write"]["runtime"]
+                    .as_u64()
+                    .expect("'fio' parse error: missing entry 'write.runtime'")
+                    as f64
+                    / 1000_f64;
+                total_bps += bytes as f64 / runtime as f64;
+            }
+        }
+
+        total_bps
+    })
+    .map_err(|_| {
+        eprintln!(
+            "=============== Fio output ===============\n\n{}\n\n===========end============\n\n",
+            output
+        );
+        Error::FioOutputParse
+    })
+}
+
+pub fn performance_block_io(control: &PerformanceTestControl) -> f64 {
+    let test_timeout = control.test_timeout;
+    let num_queues = control.num_queues.unwrap();
+    let queue_size = control.queue_size.unwrap();
+    let fio_ops = control.fio_ops.as_ref().unwrap();
+
+    let focal = UbuntuDiskConfig::new(FOCAL_IMAGE_NAME.to_string());
+    let guest = performance_test_new_guest(Box::new(focal));
+    let api_socket = guest
+        .tmp_dir
+        .as_path()
+        .join("cloud-hypervisor.sock")
+        .to_str()
+        .unwrap()
+        .to_string();
+
+    let mut child = GuestCommand::new(&guest)
+        .args(&["--cpus", &format!("boot={}", num_queues)])
+        .args(&["--memory", "size=4G"])
+        .args(&["--kernel", direct_kernel_boot_path().to_str().unwrap()])
+        .args(&["--cmdline", DIRECT_KERNEL_BOOT_CMDLINE])
+        .default_disks()
+        .default_net()
+        .args(&["--api-socket", &api_socket])
+        .capture_output()
+        .set_print_cmd(false)
+        .spawn()
+        .unwrap();
+
+    let r = std::panic::catch_unwind(|| {
+        guest.wait_vm_boot(None).unwrap();
+
+        // Hotplug test disk
+        assert!(Command::new(clh_command("ch-remote"))
+            .args(&[&format!("--api-socket={}", api_socket)])
+            .args(&[
+                "add-disk",
+                &format!(
+                    "path={},num_queues={},queue_size={},direct=on",
+                    BLK_IO_TEST_IMG, num_queues, queue_size
+                )
+            ])
+            .stderr(Stdio::piped())
+            .stdout(Stdio::piped())
+            .spawn()
+            .unwrap()
+            .wait_timeout(Duration::from_secs(5))
+            .unwrap()
+            .expect("Failed to hotplug test disk image")
+            .success());
+
+        let fio_command = format!(
+            "sudo fio --filename=/dev/vdc --name=test --output-format=json \
+            --direct=1 --bs=4k --ioengine=io_uring --iodepth=64 \
+            --rw={} --runtime={} --numjobs={}",
+            fio_ops, test_timeout, num_queues
+        );
+        let output = guest
+            .ssh_command(&fio_command)
+            .map_err(InfraError::SshCommand)
+            .unwrap();
+
+        // Parse fio output
+        parse_fio_output(&output, fio_ops, num_queues).unwrap()
+    });
+
+    let _ = child.kill();
+    let output = child.wait_with_output().unwrap();
+
+    match r {
+        Ok(r) => r,
+        Err(e) => {
+            handle_child_output(Err(e), &output);
+            panic!("test failed!");
+        }
+    }
+}
+
+#[cfg(test)]
+mod tests {
+    use super::*;
+
+    #[test]
+    fn test_parse_iperf3_output() {
+        let output = r#"
+{
+	"end":	{
+		"sum_sent":	{
+			"start":	0,
+			"end":	5.000196,
+			"seconds":	5.000196,
+			"bytes":	14973836248,
+			"bits_per_second":	23957198874.604115,
+			"retransmits":	0,
+			"sender":	false
+		}
+	}
+}
+       "#;
+        assert_eq!(
+            parse_iperf3_output(output.as_bytes(), true).unwrap(),
+            23957198874.604115
+        );
+
+        let output = r#"
+{
+	"end":	{
+		"sum_received":	{
+			"start":	0,
+			"end":	5.000626,
+			"seconds":	5.000626,
+			"bytes":	24703557800,
+			"bits_per_second":	39520744482.79,
+			"sender":	true
+		}
+	}
+}
+              "#;
+        assert_eq!(
+            parse_iperf3_output(output.as_bytes(), false).unwrap(),
+            39520744482.79
+        );
+    }
+
+    #[test]
+    fn test_parse_ethr_latency_output() {
+        let output = r#"{"Time":"2022-02-08T03:52:50Z","Title":"","Type":"INFO","Message":"Using destination: 192.168.249.2, ip: 192.168.249.2, port: 8888"}
+{"Time":"2022-02-08T03:52:51Z","Title":"","Type":"INFO","Message":"Running latency test: 1000, 1"}
+{"Time":"2022-02-08T03:52:51Z","Title":"","Type":"LatencyResult","RemoteAddr":"192.168.249.2","Protocol":"TCP","Avg":"80.712us","Min":"61.677us","P50":"257.014us","P90":"74.418us","P95":"107.283us","P99":"119.309us","P999":"142.100us","P9999":"216.341us","Max":"216.341us"}
+{"Time":"2022-02-08T03:52:52Z","Title":"","Type":"LatencyResult","RemoteAddr":"192.168.249.2","Protocol":"TCP","Avg":"79.826us","Min":"55.129us","P50":"598.996us","P90":"73.849us","P95":"106.552us","P99":"122.152us","P999":"142.459us","P9999":"474.280us","Max":"474.280us"}
+{"Time":"2022-02-08T03:52:53Z","Title":"","Type":"LatencyResult","RemoteAddr":"192.168.249.2","Protocol":"TCP","Avg":"78.239us","Min":"56.999us","P50":"396.820us","P90":"69.469us","P95":"115.421us","P99":"119.404us","P999":"130.158us","P9999":"258.686us","Max":"258.686us"}"#;
+
+        let ret = parse_ethr_latency_output(output.as_bytes()).unwrap();
+        let reference = vec![80.712_f64, 79.826_f64, 78.239_f64];
+        assert_eq!(ret, reference);
+    }
+
+    #[test]
+    fn test_parse_boot_time_output() {
+        let output = r#"
+cloud-hypervisor: 161.167103ms: <vcpu0> INFO:vmm/src/vm.rs:392 -- [Debug I/O port: Kernel code 0x40] 0.132 seconds
+cloud-hypervisor: 613.57361ms: <vcpu0> INFO:vmm/src/vm.rs:392 -- [Debug I/O port: Kernel code 0x41] 0.5845 seconds
+        "#;
+
+        assert_eq!(parse_boot_time_output(output.as_bytes()).unwrap(), 0.4525);
+    }
+
+    #[test]
+    fn test_parse_fio_output() {
+        let output = r#"
+{
+  "jobs" : [
+    {
+      "read" : {
+        "io_bytes" : 1965273088,
+        "io_kbytes" : 1919212,
+        "bw_bytes" : 392976022,
+        "bw" : 383765,
+        "iops" : 95941.411718,
+        "runtime" : 5001,
+        "total_ios" : 479803,
+        "short_ios" : 0,
+        "drop_ios" : 0
+      }
+    }
+  ]
+}
+"#;
+
+        let bps = 1965273088_f64 / (5001_f64 / 1000_f64);
+        assert_eq!(
+            parse_fio_output(output, &FioOps::RandomRead, 1).unwrap(),
+            bps
+        );
+        assert_eq!(parse_fio_output(output, &FioOps::Read, 1).unwrap(), bps);
+
+        let output = r#"
+{
+  "jobs" : [
+    {
+      "write" : {
+        "io_bytes" : 1172783104,
+        "io_kbytes" : 1145296,
+        "bw_bytes" : 234462835,
+        "bw" : 228967,
+        "iops" : 57241.903239,
+        "runtime" : 5002,
+        "total_ios" : 286324,
+        "short_ios" : 0,
+        "drop_ios" : 0
+      }
+    },
+    {
+      "write" : {
+        "io_bytes" : 1172234240,
+        "io_kbytes" : 1144760,
+        "bw_bytes" : 234353106,
+        "bw" : 228860,
+        "iops" : 57215.113954,
+        "runtime" : 5002,
+        "total_ios" : 286190,
+        "short_ios" : 0,
+        "drop_ios" : 0
+      }
+    }
+  ]
+}
+"#;
+
+        let bps = 1172783104_f64 / (5002_f64 / 1000_f64) + 1172234240_f64 / (5002_f64 / 1000_f64);
+        assert_eq!(
+            parse_fio_output(output, &FioOps::RandomWrite, 2).unwrap(),
+            bps
+        );
+        assert_eq!(parse_fio_output(output, &FioOps::Write, 2).unwrap(), bps);
+    }
+}

qcow/Cargo.toml

@@ -10,7 +10,7 @@ path = "src/qcow.rs"
 
 [dependencies]
 byteorder = "1.4.3"
-libc = "0.2.103"
+libc = "0.2.119"
 log = "0.4.14"
 remain = "0.2.2"
 vmm-sys-util = "0.9.0"

qcow/src/raw_file.rs

@@ -28,8 +28,11 @@ const BLK_ALIGNMENTS: [usize; 2] = [512, 4096];
 
 fn is_valid_alignment(fd: RawFd, alignment: usize) -> bool {
     let layout = Layout::from_size_align(alignment, alignment).unwrap();
+    // SAFETY: layout has non-zero size
     let ptr = unsafe { alloc_zeroed(layout) };
+    assert!(!ptr.is_null());
 
+    // SAFETY: FFI call
     let ret = unsafe {
         ::libc::pread(
             fd,
@@ -39,6 +42,7 @@ fn is_valid_alignment(fd: RawFd, alignment: usize) -> bool {
         )
     };
 
+    // SAFETY: ptr was allocated by alloc_zeroed with layout
     unsafe { dealloc(ptr, layout) };
 
     ret >= 0
@@ -141,11 +145,18 @@ impl Read for RawFile {
                 .unwrap();
 
             let layout = Layout::from_size_align(rounded_len, self.alignment).unwrap();
+            // SAFETY: layout has non-zero size
             let tmp_ptr = unsafe { alloc_zeroed(layout) };
+            if tmp_ptr.is_null() {
+                return Err(io::Error::last_os_error());
+            }
+
+            // SAFETY: tmp_ptr is valid and at least rounded_len long
             let tmp_buf = unsafe { slice::from_raw_parts_mut(tmp_ptr, rounded_len) };
 
             // This can eventually replaced with read_at once its interface
             // has been stabilized.
+            // SAFETY: FFI call. All parameters are valid.
             let ret = unsafe {
                 ::libc::pread64(
                     self.file.as_raw_fd(),
@@ -155,12 +166,14 @@ impl Read for RawFile {
                 )
             };
             if ret < 0 {
+                // SAFETY: tmp_ptr was allocated by alloc_zeroed with layout
                 unsafe { dealloc(tmp_ptr, layout) };
                 return Err(io::Error::last_os_error());
             }
 
             let read: usize = ret.try_into().unwrap();
             if read < file_offset {
+                // SAFETY: tmp_ptr was allocated by alloc_zeroed with layout
                 unsafe { dealloc(tmp_ptr, layout) };
                 return Ok(0);
             }
@@ -171,6 +184,7 @@ impl Read for RawFile {
             }
 
             buf.copy_from_slice(&tmp_buf[file_offset..(file_offset + buf_len)]);
+            // SAFETY: tmp_ptr was allocated by alloc_zeroed with layout
             unsafe { dealloc(tmp_ptr, layout) };
 
             self.seek(SeekFrom::Current(to_copy.try_into().unwrap()))
@@ -211,11 +225,17 @@ impl Write for RawFile {
                 .unwrap();
 
             let layout = Layout::from_size_align(rounded_len, self.alignment).unwrap();
+            // SAFETY: layout has non-zero size
             let tmp_ptr = unsafe { alloc_zeroed(layout) };
+            if tmp_ptr.is_null() {
+                return Err(io::Error::last_os_error());
+            }
+
             let tmp_buf = unsafe { slice::from_raw_parts_mut(tmp_ptr, rounded_len) };
 
             // This can eventually replaced with read_at once its interface
             // has been stabilized.
+            // SAFETY: FFI call
             let ret = unsafe {
                 ::libc::pread64(
                     self.file.as_raw_fd(),
@@ -225,6 +245,7 @@ impl Write for RawFile {
                 )
             };
             if ret < 0 {
+                // SAFETY: tmp_ptr was allocated by alloc_zeroed with layout
                 unsafe { dealloc(tmp_ptr, layout) };
                 return Err(io::Error::last_os_error());
             };
@@ -233,6 +254,7 @@ impl Write for RawFile {
 
             // This can eventually replaced with write_at once its interface
             // has been stabilized.
+            // SAFETY: FFI call
             let ret = unsafe {
                 ::libc::pwrite64(
                     self.file.as_raw_fd(),
@@ -242,6 +264,7 @@ impl Write for RawFile {
                 )
             };
 
+            // SAFETY: tmp_ptr was allocated by alloc_zeroed with layout
             unsafe { dealloc(tmp_ptr, layout) };
 
             if ret < 0 {

rate_limiter/Cargo.toml

@@ -4,6 +4,6 @@ version = "0.1.0"
 edition = "2018"
 
 [dependencies]
-libc = "0.2.103"
+libc = "0.2.119"
 log = "0.4.14"
 vmm-sys-util = "0.9.0"

release-notes.md

@@ -1,3 +1,29 @@
+- [v22.0](#v220)
+    - [GDB Debug Stub Support](#gdb-debug-stub-support)
+    - [`virtio-iommu` Backed Segments](#virtio-iommu-backed-segments)
+    - [Before Boot Configuration Changes](#before-boot-configuration-changes)
+    - [`virtio-balloon` Free Page Reporting](#virtio-balloon-free-page-reporting)
+    - [Support for Direct Kernel Booting with TDX](#support-for-direct-kernel-booting-with-tdx)
+    - [PMU Support for AArch64](#pmu-support-for-aarch64)
+    - [Documentation Under CC-BY-4.0 Licese](#documentation-under-cc-by-40-licese)
+    - [Deprecation of "Classic" `virtiofsd`](#deprecation-of-classic-virtiofsd)
+    - [Notable Bug Fixes](#notable-bug-fixes)
+    - [Contributors](#contributors)
+- [v21.0](#v210)
+    - [Efficient Local Live Migration (for Live Upgrade)](#efficient-local-live-migration-for-live-upgrade)
+    - [Recommended Kernel is Now 5.15](#recommended-kernel-is-now-515)
+    - [Notable Bug fixes](#notable-bug-fixes-1)
+    - [Contributors](#contributors-1)
+- [v20.2](#v202)
+- [v20.1](#v201)
+- [v20.0](#v200)
+    - [Multiple PCI segments support](#multiple-pci-segments-support)
+    - [CPU pinning](#cpu-pinning)
+    - [Improved VFIO support](#improved-vfio-support)
+    - [Safer code](#safer-code)
+    - [Extended documentation](#extended-documentation)
+    - [Notable bug fixes](#notable-bug-fixes-2)
+    - [Contributors](#contributors-2)
 - [v19.0](#v190)
     - [Improved PTY handling for serial and `virtio-console`](#improved-pty-handling-for-serial-and-virtio-console)
     - [PCI boot time optimisations](#pci-boot-time-optimisations)
@@ -5,8 +31,8 @@
     - [Live migration enhancements](#live-migration-enhancements)
     - [`virtio-mem` support with `vfio-user`](#virtio-mem-support-with-vfio-user)
     - [AArch64 for `virtio-iommu`](#aarch64-for-virtio-iommu)
-    - [Notable bug fixes](#notable-bug-fixes)
+    - [Notable bug fixes](#notable-bug-fixes-3)
-    - [Contributors](#contributors)
+    - [Contributors](#contributors-3)
 - [v18.0](#v180)
     - [Experimental User Device (`vfio-user`) support](#experimental-user-device-vfio-user-support)
     - [Migration support for `vhost-user` devices](#migration-support-for-vhost-user-devices)
@@ -16,23 +42,23 @@
     - [Live migration on MSHV hypervisor](#live-migration-on-mshv-hypervisor)
     - [AArch64 CPU topology support](#aarch64-cpu-topology-support)
     - [Power button support on AArch64](#power-button-support-on-aarch64)
-    - [Notable bug fixes](#notable-bug-fixes-1)
+    - [Notable bug fixes](#notable-bug-fixes-4)
-    - [Contributors](#contributors-1)
+    - [Contributors](#contributors-4)
 - [v17.0](#v170)
     - [ARM64 NUMA support using ACPI](#arm64-numa-support-using-acpi)
     - [`Seccomp` support for MSHV backend](#seccomp-support-for-mshv-backend)
     - [Hotplug of `macvtap` devices](#hotplug-of-macvtap-devices)
     - [Improved SGX support](#improved-sgx-support)
     - [Inflight tracking for `vhost-user` devices](#inflight-tracking-for-vhost-user-devices)
-    - [Notable bug fixes](#notable-bug-fixes-2)
+    - [Notable bug fixes](#notable-bug-fixes-5)
-    - [Contributors](#contributors-2)
+    - [Contributors](#contributors-5)
 - [v16.0](#v160)
     - [Improved live migration support](#improved-live-migration-support)
     - [Improved `vhost-user` support](#improved-vhost-user-support)
     - [ARM64 ACPI and UEFI support](#arm64-acpi-and-uefi-support)
-    - [Notable bug fixes](#notable-bug-fixes-3)
+    - [Notable bug fixes](#notable-bug-fixes-6)
     - [Removed functionality](#removed-functionality)
-    - [Contributors](#contributors-3)
+    - [Contributors](#contributors-6)
 - [v15.0](#v150)
     - [Version numbering and stability guarantees](#version-numbering-and-stability-guarantees)
     - [Network device rate limiting](#network-device-rate-limiting)
@@ -40,7 +66,7 @@
     - [`--api-socket` supports file descriptor parameter](#--api-socket-supports-file-descriptor-parameter)
     - [Bug fixes](#bug-fixes)
     - [Deprecations](#deprecations)
-    - [Contributors](#contributors-4)
+    - [Contributors](#contributors-7)
 - [v0.14.1](#v0141)
 - [v0.14.0](#v0140)
     - [Structured event monitoring](#structured-event-monitoring)
@@ -50,7 +76,7 @@
     - [PTY control for serial and `virtio-console`](#pty-control-for-serial-and-virtio-console)
     - [Block device rate limiting](#block-device-rate-limiting)
     - [Deprecations](#deprecations-1)
-    - [Contributors](#contributors-5)
+    - [Contributors](#contributors-8)
 - [v0.13.0](#v0130)
     - [Wider VFIO device support](#wider-vfio-device-support)
     - [Improved huge page support](#improved-huge-page-support)
@@ -58,13 +84,13 @@
     - [VHD disk image support](#vhd-disk-image-support)
     - [Improved Virtio device threading](#improved-virtio-device-threading)
     - [Clean shutdown support via synthetic power button](#clean-shutdown-support-via-synthetic-power-button)
-    - [Contributors](#contributors-6)
+    - [Contributors](#contributors-9)
 - [v0.12.0](#v0120)
     - [ARM64 enhancements](#arm64-enhancements)
     - [Removal of `vhost-user-net` and `vhost-user-block` self spawning](#removal-of-vhost-user-net-and-vhost-user-block-self-spawning)
     - [Migration of `vhost-user-fs` backend](#migration-of-vhost-user-fs-backend)
     - [Enhanced "info" API](#enhanced-info-api)
-    - [Contributors](#contributors-7)
+    - [Contributors](#contributors-10)
 - [v0.11.0](#v0110)
     - [`io_uring` support by default for `virtio-block`](#io_uring-support-by-default-for-virtio-block)
     - [Windows Guest Support](#windows-guest-support)
@@ -76,15 +102,15 @@
     - [Default Log Level Changed](#default-log-level-changed)
     - [New `--balloon` Parameter Added](#new---balloon-parameter-added)
     - [Experimental `virtio-watchdog` Support](#experimental-virtio-watchdog-support)
-    - [Notable Bug Fixes](#notable-bug-fixes-4)
+    - [Notable Bug Fixes](#notable-bug-fixes-7)
-    - [Contributors](#contributors-8)
+    - [Contributors](#contributors-11)
 - [v0.10.0](#v0100)
     - [`virtio-block` Support for Multiple Descriptors](#virtio-block-support-for-multiple-descriptors)
     - [Memory Zones](#memory-zones)
     - [`Seccomp` Sandbox Improvements](#seccomp-sandbox-improvements)
     - [Preliminary KVM HyperV Emulation Control](#preliminary-kvm-hyperv-emulation-control)
-    - [Notable Bug Fixes](#notable-bug-fixes-5)
+    - [Notable Bug Fixes](#notable-bug-fixes-8)
-    - [Contributors](#contributors-9)
+    - [Contributors](#contributors-12)
 - [v0.9.0](#v090)
     - [`io_uring` Based Block Device Support](#io_uring-based-block-device-support)
     - [Block and Network Device Statistics](#block-and-network-device-statistics)
@@ -97,17 +123,17 @@
     - [Enhancements to ARM64 Support](#enhancements-to-arm64-support)
     - [Intel SGX Support](#intel-sgx-support)
     - [`Seccomp` Sandbox Improvements](#seccomp-sandbox-improvements-1)
-    - [Notable Bug Fixes](#notable-bug-fixes-6)
+    - [Notable Bug Fixes](#notable-bug-fixes-9)
-    - [Contributors](#contributors-10)
+    - [Contributors](#contributors-13)
 - [v0.8.0](#v080)
     - [Experimental Snapshot and Restore Support](#experimental-snapshot-and-restore-support)
     - [Experimental ARM64 Support](#experimental-arm64-support)
     - [Support for Using 5-level Paging in Guests](#support-for-using-5-level-paging-in-guests)
     - [Virtio Device Interrupt Suppression for Network Devices](#virtio-device-interrupt-suppression-for-network-devices)
     - [`vhost_user_fs` Improvements](#vhost_user_fs-improvements)
-    - [Notable Bug Fixes](#notable-bug-fixes-7)
+    - [Notable Bug Fixes](#notable-bug-fixes-10)
     - [Command Line and API Changes](#command-line-and-api-changes)
-    - [Contributors](#contributors-11)
+    - [Contributors](#contributors-14)
 - [v0.7.0](#v070)
     - [Block, Network, Persistent Memory (PMEM), VirtioFS and Vsock hotplug](#block-network-persistent-memory-pmem-virtiofs-and-vsock-hotplug)
     - [Alternative `libc` Support](#alternative-libc-support)
@@ -117,14 +143,14 @@
     - [`Seccomp` Sandboxing](#seccomp-sandboxing)
     - [Updated Distribution Support](#updated-distribution-support)
     - [Command Line and API Changes](#command-line-and-api-changes-1)
-    - [Contributors](#contributors-12)
+    - [Contributors](#contributors-15)
 - [v0.6.0](#v060)
     - [Directly Assigned Devices Hotplug](#directly-assigned-devices-hotplug)
     - [Shared Filesystem Improvements](#shared-filesystem-improvements)
     - [Block and Networking IO Self Offloading](#block-and-networking-io-self-offloading)
     - [Command Line Interface](#command-line-interface)
     - [PVH Boot](#pvh-boot)
-    - [Contributors](#contributors-13)
+    - [Contributors](#contributors-16)
 - [v0.5.1](#v051)
 - [v0.5.0](#v050)
     - [Virtual Machine Dynamic Resizing](#virtual-machine-dynamic-resizing)
@@ -132,7 +158,7 @@
     - [New Interrupt Management Framework](#new-interrupt-management-framework)
     - [Development Tools](#development-tools)
     - [Kata Containers Integration](#kata-containers-integration)
-    - [Contributors](#contributors-14)
+    - [Contributors](#contributors-17)
 - [v0.4.0](#v040)
     - [Dynamic virtual CPUs addition](#dynamic-virtual-cpus-addition)
     - [Programmatic firmware tables generation](#programmatic-firmware-tables-generation)
@@ -141,7 +167,7 @@
     - [Userspace IOAPIC by default](#userspace-ioapic-by-default)
     - [PCI BAR reprogramming](#pci-bar-reprogramming)
     - [New `cloud-hypervisor` organization](#new-cloud-hypervisor-organization)
-    - [Contributors](#contributors-15)
+    - [Contributors](#contributors-18)
 - [v0.3.0](#v030)
     - [Block device offloading](#block-device-offloading)
     - [Network device backend](#network-device-backend)
@@ -168,6 +194,214 @@
     - [Unit testing](#unit-testing)
     - [Integration tests parallelization](#integration-tests-parallelization)
 
+# v22.0
+
+This release has been tracked through the [v21.0
+project](https://github.com/cloud-hypervisor/cloud-hypervisor/projects/24).
+
+### GDB Debug Stub Support
+
+Cloud Hypervisor can now be used as debug target with GDB. This is controlled
+by the `gdb` compile time feature and details of how to use it can be found in
+the [gdb
+documentation](https://github.com/cloud-hypervisor/cloud-hypervisor/blob/main/docs/gdb.md).
+
+
+### `virtio-iommu` Backed Segments
+
+In order to facilitate hotplug devices that require being behind an IOMMU (e.g.
+QAT) there is a new option `--platform iommu_segments=<list_of_segments>` that
+will place all the specified segments behind the IOMMU.
+
+### Before Boot Configuration Changes
+
+It is now possible to change the VM configuration (e.g. add or remove devices,
+resize) before the VM is booted.
+
+
+### `virtio-balloon` Free Page Reporting
+
+If `--balloon free_page_reporting=on` is used then the guest can report pages
+that is it not using to the VMM. The VMM will then notify the host OS that
+those pages are no longer in use and can be freed. This can result in improved
+memory density.
+
+### Support for Direct Kernel Booting with TDX
+
+Through the use of `TD-Shim` lightweight firmware it is now possible to
+directly boot into the kernel with TDX. The [TDX
+documentation](https://github.com/cloud-hypervisor/cloud-hypervisor/blob/main/docs/intel_tdx.md#tdshim)
+has been updated for this usage.
+
+### PMU Support for AArch64
+
+A PMU is now available on AArch64 for guest performance profiling. This will be
+exposed automatically if available from the host.
+
+### Documentation Under CC-BY-4.0 Licese
+
+The documentation is now licensed under the "Creative Commons Attribution 4.0
+International" license which is aligned with the project charter under the
+Linux Foundation.
+
+### Deprecation of "Classic" `virtiofsd`
+
+The use of the Rust based [virtiofsd](https://gitlab.com/virtio-fs/virtiofsd)
+is now recommended and we are no longer testing against the C based "classic"
+version.
+
+### Notable Bug Fixes
+
+* Can now be used on kernels without `AF_INET` support (#3785)
+* `virtio-balloon` size is now validated against guest RAM size (#3689)
+* Ensure that I/O related KVM VM Exits are correctly handled (#3677)
+* Multiple TAP file descriptors can be used for `virtio-net` device hotplug (#3607)
+* Minor API improvements and fixes (#3756, #3766, #3647, #3578)
+* Fix sporadic seccomp violation from glibc memory freeing (#3610, #3609)
+* Fix Windows 11 on AArch64 due to wider MSI-X register accesses (#3714, #3720)
+* Ensure `vhost-user` features are correct across migration (#3737)
+* Improved vCPU topology on AArch64 (#3735, #3733)
+
+### Contributors
+
+Many thanks to everyone who has contributed to our release:
+
+* Akira Moroo <retrage01@gmail.com>
+* Barret Rhoden <brho@google.com>
+* Bo Chen <chen.bo@intel.com>
+* Fabiano Fidêncio <fabiano.fidencio@intel.com>
+* Feng Ye <yefeng@smartx.com>
+* Henry Wang <Henry.Wang@arm.com>
+* Jianyong Wu <jianyong.wu@arm.com>
+* lizhaoxin1 <Lxiaoyouling@163.com>
+* Michael Zhao <michael.zhao@arm.com>
+* Rob Bradford <robert.bradford@intel.com>
+* Sebastien Boeuf <sebastien.boeuf@intel.com>
+* Wei Liu <liuwe@microsoft.com>
+
+# v21.0
+
+This release has been tracked through the [v21.0
+project](https://github.com/cloud-hypervisor/cloud-hypervisor/projects/24).
+
+### Efficient Local Live Migration (for Live Upgrade)
+
+In order to support fast live upgrade of the VMM an optimised path has been
+added in which the memory for the VM is not compared from source to
+destination. This is activated by passing `--local` to the `ch-remote
+send-migration` command. This means that the live upgrade can complete in the
+order of 50ms vs 3s. (#3566)
+
+### Recommended Kernel is Now 5.15
+
+Due to an issue in the `virtio-net` code in 5.14 the recommended Linux kernel
+is now 5.15. (#3530)
+
+### Notable Bug fixes
+
+* Multiple fixes were made to the OpenAPI YAML file to match the implementaion (#3555,#3562)
+* Avoid live migration deadlock when triggered during the kernel boot (#3585)
+* Support live migration within firmware (#3586)
+* Validate the `virtio-net` desciptor chain (#3548)
+* `direct=on` (`O_DIRECT`) can now be used with a guest that makes unaligned accesses (e.g. firmware) (#3587)
+
+### Contributors
+
+Many thanks to everyone who has contributed to our release:
+
+* Anatol Belski <anbelski@linux.microsoft.com>
+* Barret Rhoden <brho@google.com>
+* Bo Chen <chen.bo@intel.com>
+* Fabiano Fidêncio <fabiano.fidencio@intel.com>
+* Henry Wang <Henry.Wang@arm.com>
+* Liang Zhou <zhoul110@chinatelecom.cn>
+* Michael Zhao <michael.zhao@arm.com>
+* Muhammad Falak R Wani <falakreyaz@gmail.com>
+* Muminul Islam <muislam@microsoft.com>
+* Rob Bradford <robert.bradford@intel.com>
+* Sebastien Boeuf <sebastien.boeuf@intel.com>
+* Wei Liu <liuwe@microsoft.com>
+* Ziye Yang <ziye.yang@intel.com>
+ 
+# v20.2
+
+This is a bug fix release. The following issues have been addressed:
+
+* Don't error out when setting up the SIGWINCH handler (for console resize)
+  when this fails due to older kernel (#3456)
+* Seccomp rules were refined to remove syscalls that are now unused
+* Fix reboot on older host kernels when SIGWINCH handler was not initialised
+  (#3496)
+* Fix virtio-vsock blocking issue (#3497)
+
+# v20.1
+
+This is a bug fix release. The following issues have been addressed:
+
+* Networking performance regression with `virtio-net` (#3450)
+* Limit file descriptors sent in `vfio-user` support (#3401)
+* Fully advertise PCI MMIO config regions in ACPI tables (#3432)
+* Set the TSS and KVM identity maps so they don't overlap with firmware RAM
+* Correctly update the `DeviceTree` on restore
+
+# v20.0
+
+This release has been tracked through the [v20.0
+project](https://github.com/cloud-hypervisor/cloud-hypervisor/projects/23).
+
+### Multiple PCI segments support
+
+Cloud Hypervisor is no longer limited to 31 PCI devices. For both `x86_64` and
+`aarch64` architectures, it is now possible to create up to 16 PCI segments,
+increasing the total amount of supported PCI devices to 496.
+
+### CPU pinning
+
+For each vCPU, the user can define a limited set of host CPUs on which it is
+allowed to run. This can be useful when assigning a 1:1 mapping between host and
+guest resources, or when running a VM on a specific NUMA node.
+
+### Improved VFIO support
+
+Based on VFIO region capabilities, all regions can be memory mapped, limiting
+the amount of triggered VM exits, and therefore increasing the performance of
+the passthrough device.
+
+### Safer code
+
+Several sections containing unsafe Rust code have been replaced with safe
+alternatives, and multiple comments have been added to clarify why the remaining
+unsafe sections are safe to use.
+
+### Extended documentation
+
+The documentation related to VFIO has been updated while some new documents have
+been introduced to cover the usage of `--cpus` parameter as well as how to run
+Cloud Hypervisor on Intel TDX.
+
+### Notable bug fixes
+
+* Naturally align PCI BARs on relocation (#3244)
+* Fix panic in SIGWINCH listener thread when no seccomp filter set (#3338)
+* Use the tty raw mode implementation from libc (#3344)
+* Fix the emulation of register D for CMOS/RTC device (#3393)
+
+### Contributors
+
+Many thanks to everyone who has contributed to our release:
+
+* Alyssa Ross <hi@alyssa.is>
+* Bo Chen <chen.bo@intel.com>
+* Fabiano Fidêncio <fabiano.fidencio@intel.com>
+* Michael Zhao <michael.zhao@arm.com>
+* Muminul Islam <muislam@microsoft.com>
+* Rob Bradford <robert.bradford@intel.com>
+* Sebastien Boeuf <sebastien.boeuf@intel.com>
+* Wei Liu <liuwe@microsoft.com>
+* Willen Yang <willenyang@gmail.com>
+* William Douglas <william.douglas@intel.com>
+* Ziye Yang <ziye.yang@intel.com>
+
 # v19.0
 
 This release has been tracked through the [v19.0
@@ -1244,7 +1478,7 @@ Highlights for `cloud-hypervisor` version 0.4.0 include:
 
 ### Dynamic virtual CPUs addition
 
-As a way to vertically scale Cloud-Hypervisor guests, we now support dynamically
+As a way to vertically scale Cloud Hypervisor guests, we now support dynamically
 adding virtual CPUs to the guests, a mechanism also known as CPU hot plug.
 Through hardware-reduced ACPI notifications, Cloud Hypervisor can now add CPUs
 to an already running guest and the high level operations for that process are

resources/Dockerfile

@@ -1,7 +1,10 @@
+# When changing this file don't forget to update the tag name in the
+# .github/workflows/docker-image.yaml file if doing multiple per day
+
 FROM ubuntu:20.04 as dev
 
 ARG TARGETARCH
-ARG RUST_TOOLCHAIN="1.55.0"
+ARG RUST_TOOLCHAIN="1.59.0"
 ARG CLH_SRC_DIR="/cloud-hypervisor"
 ARG CLH_BUILD_DIR="$CLH_SRC_DIR/build"
 ARG CARGO_REGISTRY_DIR="$CLH_BUILD_DIR/cargo_registry"
@@ -17,7 +20,6 @@ RUN apt-get update \
 	&& DEBIAN_FRONTEND=noninteractive apt-get install -yq \
 	build-essential \
 	bc \
-	docker.io \
 	curl \
 	wget \
 	sudo \
@@ -29,7 +31,6 @@ RUN apt-get update \
 	bison \
 	libelf-dev \
 	qemu-utils \
-	qemu-system \
 	libglib2.0-dev \
 	libpixman-1-dev \
 	libseccomp-dev \
@@ -44,6 +45,11 @@ RUN apt-get update \
 	openvswitch-switch-dpdk \
 	python3-distutils \
 	uuid-dev \
+	iperf3 \
+	zip \
+	git-core \
+	dnsmasq \
+	dmsetup \
     && apt-get clean \
     && rm -rf /var/lib/apt/lists/*
 
@@ -66,7 +72,7 @@ RUN if [ "$TARGETARCH" = "arm64" ]; then \
         && DEBIAN_FRONTEND=noninteractive apt-get install -yq \
         libcap2-bin \
         libguestfs-tools \
-        linux-image-5.8.0-63-generic \
+        linux-image-generic \
         autotools-dev \
         autoconf \
         automake \
@@ -86,12 +92,11 @@ ENV OPENSSL_INCLUDE_DIR=/usr/include/
 
 # Install the rust toolchain
 RUN export ARCH="$(uname -m)" \
-    && nohup curl --proto '=https' --tlsv1.2 -sSf https://sh.rustup.rs | sh -s -- -y --default-toolchain "$RUST_TOOLCHAIN" \
+    && nohup curl --proto '=https' --tlsv1.2 -sSf https://sh.rustup.rs | sh -s -- -y --profile minimal --default-toolchain "$RUST_TOOLCHAIN" \
     && rustup target add $ARCH-unknown-linux-musl --toolchain "$RUST_TOOLCHAIN" \
-    && if [ "$TARGETARCH" = "amd64" ]; then rustup toolchain add $RUST_TOOLCHAIN-x86_64-unknown-linux-musl; fi \
+    && if [ "$TARGETARCH" = "amd64" ]; then rustup toolchain add --profile minimal $RUST_TOOLCHAIN-x86_64-unknown-linux-musl; fi \
     && if [ "$TARGETARCH" = "amd64" ]; then rustup component add rustfmt; fi \
     && if [ "$TARGETARCH" = "amd64" ]; then rustup component add clippy; fi \
-    && cargo install cargo-audit \
     && rm -rf "$CARGO_HOME/registry" \
     && ln -s "$CARGO_REGISTRY_DIR" "$CARGO_HOME/registry" \
     && rm -rf "$CARGO_HOME/git" \
@@ -102,25 +107,27 @@ RUN echo 'source $CARGO_HOME/env' >> $HOME/.bashrc \
     && mkdir $HOME/.cargo \
     && ln -s $CARGO_HOME/env $HOME/.cargo/env
 
-# install virtiofsd
-RUN git clone --depth 1 https://gitlab.com/virtio-fs/qemu.git -b qemu5.0-virtiofs-dax \
-    && cd qemu \
-    && ./configure --prefix=$PWD --target-list=x86_64-softmmu \
-    && make virtiofsd -j `nproc` \
-    && cp virtiofsd /usr/local/bin \
-    && cd .. && rm -rf qemu
-
 # install SPDK NVMe
-RUN git clone https://github.com/spdk/spdk \
+# only for 'x86_64' platform images as 'docker buildx' can't build 'spdk'
-    && cd spdk \
+RUN if [ "$TARGETARCH" = "amd64" ]; then \
-    && git checkout a827fd7eeca67209d4c0aaad9a3ed55692e7e36e \
+       git clone https://github.com/spdk/spdk \
-    && git submodule update --init \
+       && cd spdk \
-    && apt-get update \
+       && git checkout 59f3cdacb13bd2a19c4a86be04792b0ee4491172 \
-    && ./scripts/pkgdep.sh \
+       && git submodule update --init \
-    && ./configure --with-vfio-user \
+       && apt-get update \
-    && make -j `nproc` \
+       && ./scripts/pkgdep.sh \
-    && mkdir /usr/local/bin/spdk-nvme \
+       && apt-get clean \
-    && cp ./build/bin/nvmf_tgt /usr/local/bin/spdk-nvme \
+       && ./configure --with-vfio-user \
-    && cp ./scripts/rpc.py /usr/local/bin/spdk-nvme \
+       && make -j `nproc` \
-    && cp -r ./scripts/rpc /usr/local/bin/spdk-nvme \
+       && mkdir /usr/local/bin/spdk-nvme \
-    && cd .. && rm -rf spdk
+       && cp ./build/bin/nvmf_tgt /usr/local/bin/spdk-nvme \
+       && cp ./scripts/rpc.py /usr/local/bin/spdk-nvme \
+       && cp -r ./scripts/rpc /usr/local/bin/spdk-nvme \
+       && cd .. && rm -rf spdk; fi
+
+# install ethr tool for performance tests
+RUN if [ "$TARGETARCH" = "amd64" ]; then \
+    wget https://github.com/microsoft/ethr/releases/latest/download/ethr_linux.zip \
+    && unzip ethr_linux.zip \
+    && cp ethr /usr/local/bin \
+    && rm ethr_linux.zip; fi