build: Release v25.0

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

.github/workflows/build.yaml

@@ -53,6 +53,9 @@ jobs:
       - name: Build (default features + gdb)
         run: cargo rustc --locked --bin cloud-hypervisor --features "gdb" -- -D warnings
 
+      - name: Build (default features + guest_debug)
+        run: cargo rustc --locked --bin cloud-hypervisor --features "guest_debug" -- -D warnings
+
       - name: Build (common + mshv)
         run: cargo rustc --locked --bin cloud-hypervisor --no-default-features --features "common,mshv"  -- -D warnings
 

.github/workflows/quality.yaml

@@ -45,6 +45,9 @@ jobs:
       - name: Clippy (default features + gdb)
         run: cargo clippy --locked --all --all-targets --tests --features "gdb" -- -D warnings
 
+      - name: Clippy (default features + guest_debug)
+        run: cargo clippy --locked --all --all-targets --tests --features "guest_debug" -- -D warnings
+
       - name: Clippy (common + mshv)
         run: cargo clippy --locked --all --all-targets --no-default-features --tests --features "common,mshv" -- -D warnings
 

.github/workflows/release.yaml

@@ -16,23 +16,23 @@ jobs:
       - name: Install Rust toolchain (x86_64-unknown-linux-gnu)
         uses: actions-rs/toolchain@v1
         with:
-            toolchain: "1.60"
+            toolchain: "1.62"
             target: x86_64-unknown-linux-gnu
       - name: Install Rust toolchain (x86_64-unknown-linux-musl)
         uses: actions-rs/toolchain@v1
         with:
-            toolchain: "1.60"
+            toolchain: "1.62"
             target: x86_64-unknown-linux-musl
       - name: Build
         uses: actions-rs/cargo@v1
         with:
-            toolchain: "1.60"
+            toolchain: "1.62"
             command: build
             args: --all --release --target=x86_64-unknown-linux-gnu
       - name: Static Build
         uses: actions-rs/cargo@v1
         with:
-            toolchain: "1.60"
+            toolchain: "1.62"
             command: build
             args: --all --release --target=x86_64-unknown-linux-musl
       - name: Strip cloud-hypervisor binaries
@@ -40,7 +40,7 @@ jobs:
       - name: Install Rust toolchain (aarch64-unknown-linux-musl)
         uses: actions-rs/toolchain@v1
         with:
-            toolchain: "1.60"
+            toolchain: "1.62"
             target: aarch64-unknown-linux-musl
             override: true
       - name: Static Build (AArch64)

Cargo.lock

@@ -20,9 +20,9 @@ dependencies = [
 
 [[package]]
 name = "anyhow"
-version = "1.0.57"
+version = "1.0.58"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "08f9b8508dccb7687a1d6c4ce66b2b0ecef467c94667de27d8d7fe1f8d2a9cdc"
+checksum = "bb07d2053ccdbe10e2af2995a2f116c1330396493dc1269f6a91d0ae82e19704"
 
 [[package]]
 name = "api_client"
@@ -130,15 +130,15 @@ checksum = "baf1de4339761588bc0619e3cbc0120ee582ebb74b53b4efbf79117bd2da40fd"
 
 [[package]]
 name = "clap"
-version = "3.1.18"
+version = "3.2.8"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "d2dbdf4bdacb33466e854ce889eee8dfd5729abf7ccd7664d0a2d60cd384440b"
+checksum = "190814073e85d238f31ff738fcb0bf6910cedeb73376c87cd69291028966fd83"
 dependencies = [
  "atty",
  "bitflags",
  "clap_lex",
  "indexmap",
- "lazy_static",
+ "once_cell",
  "strsim",
  "termcolor",
  "terminal_size",
@@ -147,16 +147,16 @@ dependencies = [
 
 [[package]]
 name = "clap_lex"
-version = "0.2.0"
+version = "0.2.4"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "a37c35f1112dad5e6e0b1adaff798507497a18fceeb30cceb3bae7d1427b9213"
+checksum = "2850f2f5a82cbf437dd5af4d49848fbdfc27c157c3d010345776f952765261c5"
 dependencies = [
  "os_str_bytes",
 ]
 
 [[package]]
 name = "cloud-hypervisor"
-version = "24.0.0"
+version = "25.0.0"
 dependencies = [
  "anyhow",
  "api_client",
@@ -165,10 +165,10 @@ dependencies = [
  "epoll",
  "event_monitor",
  "hypervisor",
- "lazy_static",
  "libc",
  "log",
  "net_util",
+ "once_cell",
  "option_parser",
  "seccompiler",
  "serde_json",
@@ -206,6 +206,7 @@ dependencies = [
  "bitflags",
  "byteorder",
  "epoll",
+ "hypervisor",
  "libc",
  "log",
  "versionize",
@@ -276,9 +277,9 @@ checksum = "b643857cf70949306b81d7e92cb9d47add673868edac9863c4a49c42feaf3f1e"
 
 [[package]]
 name = "gdbstub"
-version = "0.6.1"
+version = "0.6.2"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "9fa2ca5d6b045de372cef3991f873389b421b4cabcfbe52f7787fae8b8b37906"
+checksum = "1c1f9371c87c11642ee94dcf92cb48b1484ba250b8e8bff3df71c28651f3f4e7"
 dependencies = [
  "bitflags",
  "cfg-if",
@@ -290,9 +291,9 @@ dependencies = [
 
 [[package]]
 name = "gdbstub_arch"
-version = "0.2.2"
+version = "0.2.3"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "51dc4b5718ac76d21e8605c0966dd32d80273b89b11e6cfef467b04e45934d37"
+checksum = "c24f469ba9556c5a063d6df35a8a338025fccf96ecae44f330a156b686f7a268"
 dependencies = [
  "gdbstub",
  "num-traits",
@@ -300,9 +301,9 @@ dependencies = [
 
 [[package]]
 name = "getrandom"
-version = "0.2.6"
+version = "0.2.7"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "9be70c98951c83b8d2f8f60d7065fa6d5146873094452a1008da8c2f1e4205ad"
+checksum = "4eb1a864a501629691edf6c15a593b7a51eebaa1e8468e9ddc623de7c9b58ec6"
 dependencies = [
  "cfg-if",
  "libc",
@@ -317,9 +318,9 @@ checksum = "9b919933a397b79c37e33b77bb2aa3dc8eb6e165ad809e58ff75bc7db2e34574"
 
 [[package]]
 name = "hashbrown"
-version = "0.11.2"
+version = "0.12.1"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "ab5ef0d4909ef3724cc8cce6ccc8572c5c817592e9285f5464f8e86f8bd3726e"
+checksum = "db0d4cf898abf0081f964436dc980e96670a0f36863e4b83aaacdb65c9d7ccc3"
 
 [[package]]
 name = "hermit-abi"
@@ -369,9 +370,9 @@ dependencies = [
 
 [[package]]
 name = "indexmap"
-version = "1.8.1"
+version = "1.9.1"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "0f647032dfaa1f8b6dc29bd3edb7bbef4861b8b8007ebb118d6db284fd59f6ee"
+checksum = "10a35a97730320ffe8e2d410b5d3b69279b98d2c14bdb8b70ea89ecf7888d41e"
 dependencies = [
  "autocfg",
  "hashbrown",
@@ -396,15 +397,6 @@ dependencies = [
  "libc",
 ]
 
-[[package]]
-name = "ipnetwork"
-version = "0.18.0"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "4088d739b183546b239688ddbc79891831df421773df95e236daf7867866d355"
-dependencies = [
- "serde",
-]
-
 [[package]]
 name = "ipnetwork"
 version = "0.19.0"
@@ -433,7 +425,7 @@ dependencies = [
 [[package]]
 name = "kvm-ioctls"
 version = "0.11.0"
-source = "git+https://github.com/rust-vmm/kvm-ioctls?branch=main#1e03e29cdfbb0cb108a98de7a78045a5a517f18e"
+source = "git+https://github.com/rust-vmm/kvm-ioctls?branch=main#ccf0bda07485b2433616de998c000aa5f04ce806"
 dependencies = [
  "kvm-bindings",
  "libc",
@@ -468,9 +460,9 @@ dependencies = [
 
 [[package]]
 name = "libz-sys"
-version = "1.1.6"
+version = "1.1.8"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "92e7e15d7610cce1d9752e137625f14e61a28cd45929b6e12e47b50fe154ee2e"
+checksum = "9702761c3935f8cc2f101793272e202c72b99da8f4224a19ddcf1279a6450bbf"
 dependencies = [
  "cc",
  "libc",
@@ -521,7 +513,7 @@ checksum = "2dffe52ecf27772e601905b7522cb4ef790d2cc203488bbd0e2fe85fcb74566d"
 [[package]]
 name = "micro_http"
 version = "0.1.0"
-source = "git+https://github.com/firecracker-microvm/micro-http?branch=main#a730d86940081ad044cdfbc1285c1db6d3048392"
+source = "git+https://github.com/firecracker-microvm/micro-http?branch=main#863b0370ba7e57f7df5b908ada9e5b44809ccae9"
 dependencies = [
  "libc",
  "vmm-sys-util",
@@ -530,7 +522,7 @@ dependencies = [
 [[package]]
 name = "mshv-bindings"
 version = "0.1.0"
-source = "git+https://github.com/rust-vmm/mshv?branch=main#67e13faa36989a9226ad21c9ff3947e1d5738a54"
+source = "git+https://github.com/rust-vmm/mshv?branch=main#7ac1b80bff7220171c9a626b23b51620ca774468"
 dependencies = [
  "libc",
  "serde",
@@ -542,7 +534,7 @@ dependencies = [
 [[package]]
 name = "mshv-ioctls"
 version = "0.1.0"
-source = "git+https://github.com/rust-vmm/mshv?branch=main#67e13faa36989a9226ad21c9ff3947e1d5738a54"
+source = "git+https://github.com/rust-vmm/mshv?branch=main#7ac1b80bff7220171c9a626b23b51620ca774468"
 dependencies = [
  "libc",
  "mshv-bindings",
@@ -562,11 +554,12 @@ version = "0.1.0"
 dependencies = [
  "epoll",
  "getrandom",
- "lazy_static",
  "libc",
  "log",
  "net_gen",
+ "once_cell",
  "pnet",
+ "pnet_datalink",
  "rate_limiter",
  "serde",
  "serde_json",
@@ -595,19 +588,25 @@ dependencies = [
 ]
 
 [[package]]
-name = "openssl-src"
+name = "once_cell"
-version = "111.20.0+1.1.1o"
+version = "1.13.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "92892c4f87d56e376e469ace79f1128fdaded07646ddf73aa0be4706ff712dec"
+checksum = "18a6dbe30758c9f83eb00cbea4ac95966305f5a7772f3f42ebfc7fc7eddbd8e1"
+
+[[package]]
+name = "openssl-src"
+version = "111.22.0+1.1.1q"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "8f31f0d509d1c1ae9cada2f9539ff8f37933831fd5098879e482aa687d659853"
 dependencies = [
  "cc",
 ]
 
 [[package]]
 name = "openssl-sys"
-version = "0.9.73"
+version = "0.9.74"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "9d5fd19fb3e0a8191c1e34935718976a3e70c112ab9a24af6d7cadccd9d90bc0"
+checksum = "835363342df5fba8354c5b453325b110ffd54044e588c539cf2f20a8014e4cb1"
 dependencies = [
  "autocfg",
  "cc",
@@ -623,9 +622,9 @@ version = "0.1.0"
 
 [[package]]
 name = "os_str_bytes"
-version = "6.0.1"
+version = "6.1.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "029d8d0b2f198229de29dca79676f2738ff952edf3fde542eb8bf94d8c21b435"
+checksum = "21326818e99cfe6ce1e524c2a805c189a99b5ae555a35d19f9a284b427d86afa"
 
 [[package]]
 name = "parking_lot"
@@ -702,34 +701,32 @@ checksum = "1df8c4ec4b0627e53bdf214615ad287367e482558cf84b109250b37464dc03ae"
 
 [[package]]
 name = "pnet"
-version = "0.30.0"
+version = "0.31.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "d5cc57672f576f6b95370277fb445738d4887195c6cf4192bdf4f44697e2389b"
+checksum = "0caaf5b11fd907ff15cf14a4477bfabca4b37ab9e447a4f8dead969a59cdafad"
 dependencies = [
- "ipnetwork 0.19.0",
  "pnet_base",
  "pnet_datalink",
  "pnet_packet",
- "pnet_sys",
  "pnet_transport",
 ]
 
 [[package]]
 name = "pnet_base"
-version = "0.30.0"
+version = "0.31.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "2e88341c6c842f89bdc7287f7b1e26b6fa64fa11c7ea3756971e6f18cd2510c4"
+checksum = "f9d3a993d49e5fd5d4d854d6999d4addca1f72d86c65adf224a36757161c02b6"
 dependencies = [
  "no-std-net",
 ]
 
 [[package]]
 name = "pnet_datalink"
-version = "0.30.0"
+version = "0.31.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "bc6e55d71c51db73372db35bc54f43abd8460adff1c3a9b717804ca6416d5df2"
+checksum = "e466faf03a98ad27f6e15cd27a2b7cc89e73e640a43527742977bc503c37f8aa"
 dependencies = [
- "ipnetwork 0.18.0",
+ "ipnetwork",
  "libc",
  "pnet_base",
  "pnet_sys",
@@ -738,9 +735,9 @@ dependencies = [
 
 [[package]]
 name = "pnet_macros"
-version = "0.30.0"
+version = "0.31.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "ebfcdc9c072966723026b3596a1f655fb8bbfe0142f9770f8d481aee4459d6b9"
+checksum = "48dd52a5211fac27e7acb14cfc9f30ae16ae0e956b7b779c8214c74559cef4c3"
 dependencies = [
  "proc-macro2",
  "quote",
@@ -750,18 +747,18 @@ dependencies = [
 
 [[package]]
 name = "pnet_macros_support"
-version = "0.30.0"
+version = "0.31.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "bba532f5a4b320c029d89e612671fb621851b3b07e972c53850d34130033a5cd"
+checksum = "89de095dc7739349559913aed1ef6a11e73ceade4897dadc77c5e09de6740750"
 dependencies = [
  "pnet_base",
 ]
 
 [[package]]
 name = "pnet_packet"
-version = "0.30.0"
+version = "0.31.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "7009716ac86091c1b6e2cdec95a2b028c880f516054c1ec11edd02f9f463cbde"
+checksum = "bc3b5111e697c39c8b9795b9fdccbc301ab696699e88b9ea5a4e4628978f495f"
 dependencies = [
  "glob",
  "pnet_base",
@@ -771,9 +768,9 @@ dependencies = [
 
 [[package]]
 name = "pnet_sys"
-version = "0.30.0"
+version = "0.31.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "8e2a05efbc55c22f664c0ea475fbc4ffc4d09346aff9068438279d7e3d431f6f"
+checksum = "328e231f0add6d247d82421bf3790b4b33b39c8930637f428eef24c4c6a90805"
 dependencies = [
  "libc",
  "winapi",
@@ -781,9 +778,9 @@ dependencies = [
 
 [[package]]
 name = "pnet_transport"
-version = "0.30.0"
+version = "0.31.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "75b8ff06f37863f7590183f7044ab2e8d4dae991ecea0c791e3c6dd61ed2913d"
+checksum = "ff597185e6f1f5671b3122e4dba892a1c73e17c17e723d7669bd9299cbe7f124"
 dependencies = [
  "libc",
  "pnet_base",
@@ -793,9 +790,9 @@ dependencies = [
 
 [[package]]
 name = "proc-macro2"
-version = "1.0.39"
+version = "1.0.40"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "c54b25569025b7fc9651de43004ae593a75ad88543b17178aa5e1b9c4f15f56f"
+checksum = "dd96a1e8ed2596c337f8eae5f24924ec83f5ad5ab21ea8e455d3566c69fbcaf7"
 dependencies = [
  "unicode-ident",
 ]
@@ -813,9 +810,9 @@ dependencies = [
 
 [[package]]
 name = "quote"
-version = "1.0.18"
+version = "1.0.20"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "a1feb54ed693b93a84e14094943b84b7c4eae204c512b7ccb95ab0c66d278ad1"
+checksum = "3bcdf212e9776fbcb2d23ab029360416bb1706b1aea2d1a5ba002727cbcab804"
 dependencies = [
  "proc-macro2",
 ]
@@ -851,9 +848,9 @@ dependencies = [
 
 [[package]]
 name = "regex"
-version = "1.5.6"
+version = "1.6.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "d83f127d94bdbcda4c8cc2e50f6f84f4b611f69c902699ca385a39c3a75f9ff1"
+checksum = "4c4eb3267174b8c6c2f654116623910a0fef09c4753f8dd83db29c48a0df988b"
 dependencies = [
  "aho-corasick",
  "memchr",
@@ -862,9 +859,9 @@ dependencies = [
 
 [[package]]
 name = "regex-syntax"
-version = "0.6.26"
+version = "0.6.27"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "49b3de9ec5dc0a3417da371aab17d729997c15010e7fd24ff707773a33bddb64"
+checksum = "a3f87b73ce11b1619a3c6332f45341e0047173771e8b8b73f87bfeefb7b56244"
 
 [[package]]
 name = "remain"
@@ -909,24 +906,24 @@ dependencies = [
 
 [[package]]
 name = "semver"
-version = "1.0.9"
+version = "1.0.12"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "8cb243bdfdb5936c8dc3c45762a19d12ab4550cdc753bc247637d4ec35a040fd"
+checksum = "a2333e6df6d6598f2b1974829f853c2b4c5f4a6e503c10af918081aa6f8564e1"
 
 [[package]]
 name = "serde"
-version = "1.0.137"
+version = "1.0.138"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "61ea8d54c77f8315140a05f4c7237403bf38b72704d031543aa1d16abbf517d1"
+checksum = "1578c6245786b9d168c5447eeacfb96856573ca56c9d68fdcf394be134882a47"
 dependencies = [
  "serde_derive",
 ]
 
 [[package]]
 name = "serde_derive"
-version = "1.0.137"
+version = "1.0.138"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "1f26faba0c3959972377d3b2d306ee9f71faee9714294e41bb777f83f88578be"
+checksum = "023e9b1467aef8a10fb88f25611870ada9800ef7e22afce356bb0d2387b6f27c"
 dependencies = [
  "proc-macro2",
  "quote",
@@ -935,9 +932,9 @@ dependencies = [
 
 [[package]]
 name = "serde_json"
-version = "1.0.81"
+version = "1.0.82"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "9b7ce2b32a1aed03c558dc61a5cd328f15aff2dbc17daad8fb8af04d2100e15c"
+checksum = "82c2c1fdcd807d1098552c5b9a36e425e42e9fbd7c6a37a8425f390f781f7fa7"
 dependencies = [
  "itoa",
  "ryu",
@@ -965,9 +962,9 @@ dependencies = [
 
 [[package]]
 name = "smallvec"
-version = "1.8.0"
+version = "1.9.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "f2dd574626839106c320a323308629dcb1acfc96e32a8cba364ddc61ac23ee83"
+checksum = "2fd0db749597d91ff862fd1d55ea87f7855a744a8425a64695b6fca237d1dad1"
 
 [[package]]
 name = "ssh2"
@@ -995,9 +992,9 @@ checksum = "73473c0e59e6d5812c5dfe2a064a6444949f089e20eec9a2e5506596494e4623"
 
 [[package]]
 name = "syn"
-version = "1.0.95"
+version = "1.0.98"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "fbaf6116ab8924f39d52792136fb74fd60a80194cf1b1c6ffa6453eef1c3f942"
+checksum = "c50aef8a904de4c23c788f104b7dddc7d6f79c647c7c8ce4cc8f73eb0ca773dd"
 dependencies = [
  "proc-macro2",
  "quote",
@@ -1041,8 +1038,8 @@ version = "0.1.0"
 dependencies = [
  "dirs",
  "epoll",
- "lazy_static",
  "libc",
+ "once_cell",
  "ssh2",
  "vmm-sys-util",
  "wait-timeout",
@@ -1079,9 +1076,9 @@ dependencies = [
 
 [[package]]
 name = "unicode-ident"
-version = "1.0.0"
+version = "1.0.1"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "d22af068fba1eb5edcb4aea19d382b2a3deb4c8f9d475c589b6ada9e0fd493ee"
+checksum = "5bd2fe26506023ed7b5e1e315add59d6f584c621d037f9368fea9cfb988f368c"
 
 [[package]]
 name = "unicode-xid"
@@ -1091,9 +1088,9 @@ checksum = "957e51f3646910546462e67d5f7599b9e4fb8acdd304b087a6494730f9eebf04"
 
 [[package]]
 name = "uuid"
-version = "1.0.0"
+version = "1.1.2"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "8cfcd319456c4d6ea10087ed423473267e1a071f3bc0aa89f80d60997843c6f0"
+checksum = "dd6469f4314d5f1ffec476e05f17cc9a78bc7a27a6a857842170bdf8d6f98d2f"
 dependencies = [
  "getrandom",
 ]
@@ -1143,7 +1140,7 @@ dependencies = [
 [[package]]
 name = "vfio-ioctls"
 version = "0.1.0"
-source = "git+https://github.com/rust-vmm/vfio?branch=main#f75a77c1ab6349c105bc1462a65508726b4c2e0f"
+source = "git+https://github.com/rust-vmm/vfio?branch=main#38272d3f06bbc0b4b3cfa3f5da07fd1c1c89e4d1"
 dependencies = [
  "byteorder",
  "kvm-bindings",
@@ -1202,8 +1199,9 @@ dependencies = [
 
 [[package]]
 name = "vhost-user-backend"
-version = "0.3.0"
+version = "0.5.0"
-source = "git+https://github.com/rust-vmm/vhost-user-backend?rev=14f58eda14076e973704d4f904850be1146fbb05#14f58eda14076e973704d4f904850be1146fbb05"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "3eeb6723bbee20dbc2db95c575941922fb96e77eb41786ecd1c8c03348709c4c"
 dependencies = [
  "libc",
  "log",
@@ -1293,9 +1291,9 @@ dependencies = [
 
 [[package]]
 name = "virtio-queue"
-version = "0.3.0"
+version = "0.4.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "88f2d73c184c18f8acc32dab77fcb6e3af92d53262538d3a68aa474810d6863c"
+checksum = "519c0a333c871650269cba303bc108075d52a0c0d64f9b91fae61829b53725af"
 dependencies = [
  "log",
  "vm-memory",
@@ -1381,12 +1379,12 @@ dependencies = [
  "gdbstub",
  "gdbstub_arch",
  "hypervisor",
- "lazy_static",
  "libc",
  "linux-loader",
  "log",
  "micro_http",
  "net_util",
+ "once_cell",
  "option_parser",
  "pci",
  "qcow",
@@ -1434,9 +1432,9 @@ dependencies = [
 
 [[package]]
 name = "wasi"
-version = "0.10.2+wasi-snapshot-preview1"
+version = "0.11.0+wasi-snapshot-preview1"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "fd6fbd9a79829dd1ad0cc20627bf1ed606756a7f77edff7b66b7064f9cb327c6"
+checksum = "9c8d87e72b64a3b4db28d11ce29237c246188f4f51057d65a7eab63b7987e423"
 
 [[package]]
 name = "winapi"

Cargo.toml

@@ -1,6 +1,6 @@
 [package]
 name = "cloud-hypervisor"
-version = "24.0.0"
+version = "25.0.0"
 authors = ["The Cloud Hypervisor Authors"]
 edition = "2021"
 default-run = "cloud-hypervisor"
@@ -16,9 +16,9 @@ rust-version = "1.60"
 lto = true
 
 [dependencies]
-anyhow = "1.0.57"
+anyhow = "1.0.58"
 api_client = { path = "api_client" }
-clap = { version = "3.1.18", features = ["wrap_help","cargo"] }
+clap = { version = "3.2.8", features = ["wrap_help","cargo"] }
 epoll = "4.3.1"
 event_monitor = { path = "event_monitor" }
 hypervisor = { path = "hypervisor" }
@@ -26,7 +26,7 @@ libc = "0.2.126"
 log = { version = "0.4.17", features = ["std"] }
 option_parser = { path = "option_parser" }
 seccompiler = "0.2.0"
-serde_json = "1.0.81"
+serde_json = "1.0.82"
 signal-hook = "0.3.14"
 thiserror = "1.0.31"
 vmm = { path = "vmm" }
@@ -34,7 +34,7 @@ vmm-sys-util = "0.9.0"
 vm-memory = "0.8.0"
 
 [build-dependencies]
-clap = { version = "3.1.18", features = ["cargo"] }
+clap = { version = "3.2.8", features = ["cargo"] }
 
 # List of patched crates
 [patch.crates-io]
@@ -44,9 +44,9 @@ versionize_derive = { git = "https://github.com/cloud-hypervisor/versionize_deri
 
 [dev-dependencies]
 dirs = "4.0.0"
-lazy_static= "1.4.0"
 net_util = { path = "net_util" }
-serde_json = "1.0.81"
+once_cell = "1.13.0"
+serde_json = "1.0.82"
 test_infra = { path = "test_infra" }
 wait-timeout = "0.2.0"
 
@@ -58,6 +58,7 @@ amx = ["vmm/amx"]
 cmos = ["vmm/cmos"]
 fwdebug = ["vmm/fwdebug"]
 gdb = ["vmm/gdb"]
+guest_debug = ["vmm/guest_debug"]
 kvm = ["vmm/kvm"]
 mshv = ["vmm/mshv"]
 tdx = ["vmm/tdx"]

Jenkinsfile

@@ -23,6 +23,19 @@ pipeline{
 						}
 					}
 				}
+				stage ('Check for fuzzer cargo files only changes') {
+					when {
+						expression {
+							return fuzzCargoFileOnly()
+						}
+					}
+					steps {
+						script {
+							runWorkers = false
+							echo "Fuzzer cargo files only changes, no need to run the CI"
+						}
+					}
+				}
 				stage ('Check for RFC/WIP builds') {
 					when {
   						changeRequest comparator: 'REGEXP', title: '.*(rfc|RFC|wip|WIP).*'
@@ -153,6 +166,88 @@ pipeline{
 						}
 					}
 				}
+				stage ('Worker build SGX') {
+					agent { node { label 'bionic-sgx' } }
+					when {
+						beforeAgent true
+						allOf {
+							branch 'main'
+							expression {
+								return runWorkers
+							}
+						}
+					}
+					stages {
+						stage ('Checkout') {
+							steps {
+								checkout scm
+							}
+						}
+						stage ('Run SGX integration tests') {
+							options {
+								timeout(time: 1, unit: 'HOURS')
+							}
+							steps {
+								sh "scripts/dev_cli.sh tests --integration-sgx"
+							}
+						}
+						stage ('Run SGX integration tests for musl') {
+							options {
+								timeout(time: 1, unit: 'HOURS')
+							}
+							steps {
+								sh "scripts/dev_cli.sh tests --integration-sgx --libc musl"
+							}
+						}
+					}
+					post {
+						always {
+							sh "sudo chown -R jenkins.jenkins ${WORKSPACE}"
+							deleteDir()
+						}
+					}
+				}
+				stage ('Worker build VFIO') {
+					agent { node { label 'bionic-vfio' } }
+					when {
+						beforeAgent true
+						allOf {
+							branch 'main'
+							expression {
+								return runWorkers
+							}
+						}
+					}
+					stages {
+						stage ('Checkout') {
+							steps {
+								checkout scm
+							}
+						}
+						stage ('Run VFIO integration tests') {
+							options {
+								timeout(time: 1, unit: 'HOURS')
+							}
+							steps {
+								sh "scripts/dev_cli.sh tests --integration-vfio"
+							}
+						}
+						stage ('Run VFIO integration tests for musl') {
+							options {
+								timeout(time: 1, unit: 'HOURS')
+							}
+							steps {
+								sh "scripts/dev_cli.sh tests --integration-vfio --libc musl"
+							}
+						}
+					}
+					post {
+						always {
+							sh "sudo chown -R jenkins.jenkins ${WORKSPACE}"
+							deleteDir()
+						}
+					}
+				}
 				stage ('Worker build - Windows guest') {
 					agent { node { label 'focal' } }
 					when {
@@ -233,6 +328,39 @@ 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: $METRICS_PUBLISH_KEY" -T ~/workloads/metrics.json'
+							}
+						}
+					}
+				}
 			}
 		}
 	}
@@ -287,3 +415,14 @@ def boolean docsFileOnly() {
         script: "git diff --name-only origin/${env.CHANGE_TARGET}... | grep -v '\\.md'"
     ) != 0
 }
+
+def boolean fuzzCargoFileOnly() {
+    if (env.CHANGE_TARGET == null) {
+        return false;
+    }
+
+    return sh(
+        returnStatus: true,
+        script: "git diff --name-only origin/${env.CHANGE_TARGET}... | grep -v -E 'fuzz\\/Cargo.(toml|lock)'"
+    ) != 0
+}

acpi_tables/src/aml.rs

@@ -68,7 +68,7 @@ impl Aml for Path {
         };
 
         for part in self.name_parts.clone().iter_mut() {
-            bytes.extend_from_slice(&part.to_vec());
+            bytes.extend_from_slice(part.as_ref());
         }
     }
 }

arch/Cargo.toml

@@ -10,13 +10,13 @@ tdx = []
 
 [dependencies]
 acpi_tables = { path = "../acpi_tables" }
-anyhow = "1.0.57"
+anyhow = "1.0.58"
 byteorder = "1.4.3"
 hypervisor = { path = "../hypervisor" }
 libc = "0.2.126"
 linux-loader = { version = "0.4.0", features = ["elf", "bzimage", "pe"] }
 log = "0.4.17"
-serde = { version = "1.0.137", features = ["rc", "derive"] }
+serde = { version = "1.0.138", features = ["rc", "derive"] }
 thiserror = "1.0.31"
 versionize = "0.1.6"
 versionize_derive = "0.1.4"

arch/src/aarch64/fdt.rs

@@ -8,17 +8,18 @@
 
 use crate::{NumaNodes, PciSpaceInfo};
 use byteorder::{BigEndian, ByteOrder};
+use hypervisor::arch::aarch64::gic::Vgic;
 use std::cmp;
 use std::collections::HashMap;
 use std::ffi::CStr;
 use std::fmt::Debug;
 use std::result;
 use std::str;
+use std::sync::{Arc, Mutex};
 
 use super::super::DeviceType;
 use super::super::GuestMemoryMmap;
 use super::super::InitramfsConfig;
-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_MMIO_CONFIG_SIZE_PER_SEGMENT,
@@ -90,7 +91,7 @@ pub fn create_fdt<T: DeviceInfoForFdt + Clone + Debug, S: ::std::hash::BuildHash
     vcpu_mpidr: Vec<u64>,
     vcpu_topology: Option<(u8, u8, u8)>,
     device_info: &HashMap<(DeviceType, String), T, S>,
-    gic_device: &dyn GicDevice,
+    gic_device: &Arc<Mutex<dyn Vgic>>,
     initrd: &Option<InitramfsConfig>,
     pci_space_info: &[PciSpaceInfo],
     numa_nodes: &NumaNodes,
@@ -315,18 +316,18 @@ fn create_chosen_node(
     Ok(())
 }
 
-fn create_gic_node(fdt: &mut FdtWriter, gic_device: &dyn GicDevice) -> FdtWriterResult<()> {
+fn create_gic_node(fdt: &mut FdtWriter, gic_device: &Arc<Mutex<dyn Vgic>>) -> FdtWriterResult<()> {
-    let gic_reg_prop = gic_device.device_properties();
+    let gic_reg_prop = gic_device.lock().unwrap().device_properties();
 
     let intc_node = fdt.begin_node("intc")?;
 
-    fdt.property_string("compatible", gic_device.fdt_compatibility())?;
+    fdt.property_string("compatible", gic_device.lock().unwrap().fdt_compatibility())?;
     fdt.property_null("interrupt-controller")?;
     // "interrupt-cells" field specifies the number of cells needed to encode an
     // interrupt source. The type shall be a <u32> and the value shall be 3 if no PPI affinity description
     // is required.
     fdt.property_u32("#interrupt-cells", 3)?;
-    fdt.property_array_u64("reg", gic_reg_prop)?;
+    fdt.property_array_u64("reg", &gic_reg_prop)?;
     fdt.property_u32("phandle", GIC_PHANDLE)?;
     fdt.property_u32("#address-cells", 2)?;
     fdt.property_u32("#size-cells", 2)?;
@@ -334,18 +335,18 @@ fn create_gic_node(fdt: &mut FdtWriter, gic_device: &dyn GicDevice) -> FdtWriter
 
     let gic_intr_prop = [
         GIC_FDT_IRQ_TYPE_PPI,
-        gic_device.fdt_maint_irq(),
+        gic_device.lock().unwrap().fdt_maint_irq(),
         IRQ_TYPE_LEVEL_HI,
     ];
     fdt.property_array_u32("interrupts", &gic_intr_prop)?;
 
-    if gic_device.msi_compatible() {
+    if gic_device.lock().unwrap().msi_compatible() {
         let msic_node = fdt.begin_node("msic")?;
-        fdt.property_string("compatible", gic_device.msi_compatibility())?;
+        fdt.property_string("compatible", gic_device.lock().unwrap().msi_compatibility())?;
         fdt.property_null("msi-controller")?;
         fdt.property_u32("phandle", MSI_PHANDLE)?;
-        let msi_reg_prop = gic_device.msi_properties();
+        let msi_reg_prop = gic_device.lock().unwrap().msi_properties();
-        fdt.property_array_u64("reg", msi_reg_prop)?;
+        fdt.property_array_u64("reg", &msi_reg_prop)?;
         fdt.end_node(msic_node)?;
     }
 

arch/src/aarch64/gic/gicv3.rs

@@ -1,262 +0,0 @@
-// Copyright 2021 Arm Limited (or its affiliates). All rights reserved.
-// Copyright 2019 Amazon.com, Inc. or its affiliates. All Rights Reserved.
-// SPDX-License-Identifier: Apache-2.0
-//
-// This file implements the GicV3 device.
-
-pub mod kvm {
-    use crate::aarch64::gic::dist_regs::{get_dist_regs, read_ctlr, set_dist_regs, write_ctlr};
-    use crate::aarch64::gic::icc_regs::{get_icc_regs, set_icc_regs};
-    use crate::aarch64::gic::kvm::{save_pending_tables, KvmGicDevice};
-    use crate::aarch64::gic::redist_regs::{
-        construct_gicr_typers, get_redist_regs, set_redist_regs,
-    };
-    use crate::aarch64::gic::GicDevice;
-    use crate::layout;
-    use anyhow::anyhow;
-    use hypervisor::kvm::kvm_bindings;
-    use hypervisor::CpuState;
-    use std::any::Any;
-    use std::convert::TryInto;
-    use std::sync::Arc;
-    use std::{boxed::Box, result};
-    use versionize::{VersionMap, Versionize, VersionizeResult};
-    use versionize_derive::Versionize;
-    use vm_memory::Address;
-    use vm_migration::{
-        Migratable, MigratableError, Pausable, Snapshot, Snapshottable, Transportable,
-        VersionMapped,
-    };
-
-    /// Errors thrown while saving/restoring the GICv3.
-    #[derive(Debug)]
-    pub enum Error {
-        /// Error in saving RDIST pending tables into guest RAM.
-        SavePendingTables(crate::aarch64::gic::Error),
-        /// Error in saving GIC distributor registers.
-        SaveDistributorRegisters(crate::aarch64::gic::Error),
-        /// Error in restoring GIC distributor registers.
-        RestoreDistributorRegisters(crate::aarch64::gic::Error),
-        /// Error in saving GIC distributor control registers.
-        SaveDistributorCtrlRegisters(crate::aarch64::gic::Error),
-        /// Error in restoring GIC distributor control registers.
-        RestoreDistributorCtrlRegisters(crate::aarch64::gic::Error),
-        /// Error in saving GIC redistributor registers.
-        SaveRedistributorRegisters(crate::aarch64::gic::Error),
-        /// Error in restoring GIC redistributor registers.
-        RestoreRedistributorRegisters(crate::aarch64::gic::Error),
-        /// Error in saving GIC CPU interface registers.
-        SaveIccRegisters(crate::aarch64::gic::Error),
-        /// Error in restoring GIC CPU interface registers.
-        RestoreIccRegisters(crate::aarch64::gic::Error),
-    }
-
-    type Result<T> = result::Result<T, Error>;
-
-    pub struct KvmGicV3 {
-        /// The hypervisor agnostic device for the GicV3
-        device: Arc<dyn hypervisor::Device>,
-
-        /// Vector holding values of GICR_TYPER for each vCPU
-        gicr_typers: Vec<u64>,
-
-        /// GIC device properties, to be used for setting up the fdt entry
-        properties: [u64; 4],
-
-        /// Number of CPUs handled by the device
-        vcpu_count: u64,
-    }
-
-    #[derive(Versionize)]
-    pub struct Gicv3State {
-        dist: Vec<u32>,
-        rdist: Vec<u32>,
-        icc: Vec<u32>,
-        // special register that enables interrupts and affinity routing
-        gicd_ctlr: u32,
-    }
-
-    impl VersionMapped for Gicv3State {}
-
-    impl KvmGicV3 {
-        // Device trees specific constants
-        pub const ARCH_GIC_V3_MAINT_IRQ: u32 = 9;
-
-        /// Get the address of the GIC distributor.
-        pub fn get_dist_addr() -> u64 {
-            layout::GIC_V3_DIST_START.raw_value()
-        }
-
-        /// Get the size of the GIC distributor.
-        pub fn get_dist_size() -> u64 {
-            layout::GIC_V3_DIST_SIZE
-        }
-
-        /// Get the address of the GIC redistributors.
-        pub fn get_redists_addr(vcpu_count: u64) -> u64 {
-            KvmGicV3::get_dist_addr() - KvmGicV3::get_redists_size(vcpu_count)
-        }
-
-        /// Get the size of the GIC redistributors.
-        pub fn get_redists_size(vcpu_count: u64) -> u64 {
-            vcpu_count * layout::GIC_V3_REDIST_SIZE
-        }
-
-        /// Save the state of GIC.
-        fn state(&self, gicr_typers: &[u64]) -> Result<Gicv3State> {
-            let gicd_ctlr =
-                read_ctlr(self.device()).map_err(Error::SaveDistributorCtrlRegisters)?;
-
-            let dist_state =
-                get_dist_regs(self.device()).map_err(Error::SaveDistributorRegisters)?;
-
-            let rdist_state = get_redist_regs(self.device(), gicr_typers)
-                .map_err(Error::SaveRedistributorRegisters)?;
-
-            let icc_state =
-                get_icc_regs(self.device(), gicr_typers).map_err(Error::SaveIccRegisters)?;
-
-            Ok(Gicv3State {
-                dist: dist_state,
-                rdist: rdist_state,
-                icc: icc_state,
-                gicd_ctlr,
-            })
-        }
-
-        /// Restore the state of GIC.
-        fn set_state(&mut self, gicr_typers: &[u64], state: &Gicv3State) -> Result<()> {
-            write_ctlr(self.device(), state.gicd_ctlr)
-                .map_err(Error::RestoreDistributorCtrlRegisters)?;
-
-            set_dist_regs(self.device(), &state.dist)
-                .map_err(Error::RestoreDistributorRegisters)?;
-
-            set_redist_regs(self.device(), gicr_typers, &state.rdist)
-                .map_err(Error::RestoreRedistributorRegisters)?;
-
-            set_icc_regs(self.device(), gicr_typers, &state.icc)
-                .map_err(Error::RestoreIccRegisters)?;
-
-            Ok(())
-        }
-    }
-
-    impl GicDevice for KvmGicV3 {
-        fn device(&self) -> &Arc<dyn hypervisor::Device> {
-            &self.device
-        }
-
-        fn fdt_compatibility(&self) -> &str {
-            "arm,gic-v3"
-        }
-
-        fn fdt_maint_irq(&self) -> u32 {
-            KvmGicV3::ARCH_GIC_V3_MAINT_IRQ
-        }
-
-        fn device_properties(&self) -> &[u64] {
-            &self.properties
-        }
-
-        fn vcpu_count(&self) -> u64 {
-            self.vcpu_count
-        }
-
-        fn set_its_device(&mut self, _its_device: Option<Arc<dyn hypervisor::Device>>) {}
-
-        fn set_gicr_typers(&mut self, vcpu_states: &[CpuState]) {
-            let gicr_typers = construct_gicr_typers(vcpu_states);
-            self.gicr_typers = gicr_typers;
-        }
-
-        fn as_any_concrete_mut(&mut self) -> &mut dyn Any {
-            self
-        }
-    }
-
-    impl KvmGicDevice for KvmGicV3 {
-        fn version() -> u32 {
-            kvm_bindings::kvm_device_type_KVM_DEV_TYPE_ARM_VGIC_V3
-        }
-
-        fn create_device(
-            device: Arc<dyn hypervisor::Device>,
-            vcpu_count: u64,
-        ) -> Box<dyn GicDevice> {
-            Box::new(KvmGicV3 {
-                device,
-                gicr_typers: vec![0; vcpu_count.try_into().unwrap()],
-                properties: [
-                    KvmGicV3::get_dist_addr(),
-                    KvmGicV3::get_dist_size(),
-                    KvmGicV3::get_redists_addr(vcpu_count),
-                    KvmGicV3::get_redists_size(vcpu_count),
-                ],
-                vcpu_count,
-            })
-        }
-
-        fn init_device_attributes(
-            _vm: &Arc<dyn hypervisor::Vm>,
-            gic_device: &mut dyn GicDevice,
-        ) -> crate::aarch64::gic::Result<()> {
-            /* Setting up the distributor attribute.
-             We are placing the GIC below 1GB so we need to substract the size of the distributor.
-            */
-            Self::set_device_attribute(
-                gic_device.device(),
-                kvm_bindings::KVM_DEV_ARM_VGIC_GRP_ADDR,
-                u64::from(kvm_bindings::KVM_VGIC_V3_ADDR_TYPE_DIST),
-                &KvmGicV3::get_dist_addr() as *const u64 as u64,
-                0,
-            )?;
-
-            /* Setting up the redistributors' attribute.
-            We are calculating here the start of the redistributors address. We have one per CPU.
-            */
-            Self::set_device_attribute(
-                gic_device.device(),
-                kvm_bindings::KVM_DEV_ARM_VGIC_GRP_ADDR,
-                u64::from(kvm_bindings::KVM_VGIC_V3_ADDR_TYPE_REDIST),
-                &KvmGicV3::get_redists_addr(gic_device.vcpu_count()) as *const u64 as u64,
-                0,
-            )?;
-
-            Ok(())
-        }
-    }
-
-    pub const GIC_V3_SNAPSHOT_ID: &str = "gic-v3";
-    impl Snapshottable for KvmGicV3 {
-        fn id(&self) -> String {
-            GIC_V3_SNAPSHOT_ID.to_string()
-        }
-
-        fn snapshot(&mut self) -> std::result::Result<Snapshot, MigratableError> {
-            let gicr_typers = self.gicr_typers.clone();
-            Snapshot::new_from_versioned_state(&self.id(), &self.state(&gicr_typers).unwrap())
-        }
-
-        fn restore(&mut self, snapshot: Snapshot) -> std::result::Result<(), MigratableError> {
-            let gicr_typers = self.gicr_typers.clone();
-            self.set_state(&gicr_typers, &snapshot.to_versioned_state(&self.id())?)
-                .map_err(|e| {
-                    MigratableError::Restore(anyhow!("Could not restore GICv3 state {:?}", e))
-                })
-        }
-    }
-
-    impl Pausable for KvmGicV3 {
-        fn pause(&mut self) -> std::result::Result<(), MigratableError> {
-            // Flush redistributors pending tables to guest RAM.
-            save_pending_tables(self.device()).map_err(|e| {
-                MigratableError::Pause(anyhow!("Could not save GICv3 GIC pending tables {:?}", e))
-            })?;
-
-            Ok(())
-        }
-    }
-    impl Transportable for KvmGicV3 {}
-    impl Migratable for KvmGicV3 {}
-}

arch/src/aarch64/gic/gicv3_its.rs

@@ -1,514 +0,0 @@
-// Copyright 2021 Arm Limited (or its affiliates). All rights reserved.
-// SPDX-License-Identifier: Apache-2.0
-//
-// This file implements the GicV3 device with ITS (Virtual Interrupt Translation Service).
-
-pub mod kvm {
-    use crate::aarch64::gic::dist_regs::{get_dist_regs, read_ctlr, set_dist_regs, write_ctlr};
-    use crate::aarch64::gic::icc_regs::{get_icc_regs, set_icc_regs};
-    use crate::aarch64::gic::redist_regs::{
-        construct_gicr_typers, get_redist_regs, set_redist_regs,
-    };
-
-    use crate::aarch64::gic::gicv3::kvm::KvmGicV3;
-    use crate::aarch64::gic::kvm::{save_pending_tables, KvmGicDevice};
-    use crate::aarch64::gic::GicDevice;
-    use crate::layout;
-    use anyhow::anyhow;
-    use hypervisor::kvm::kvm_bindings;
-    use hypervisor::CpuState;
-    use std::any::Any;
-    use std::convert::TryInto;
-    use std::sync::Arc;
-    use std::{boxed::Box, result};
-    use versionize::{VersionMap, Versionize, VersionizeResult};
-    use versionize_derive::Versionize;
-    use vm_migration::{
-        Migratable, MigratableError, Pausable, Snapshot, Snapshottable, Transportable,
-        VersionMapped,
-    };
-
-    const GITS_CTLR: u32 = 0x0000;
-    const GITS_IIDR: u32 = 0x0004;
-    const GITS_CBASER: u32 = 0x0080;
-    const GITS_CWRITER: u32 = 0x0088;
-    const GITS_CREADR: u32 = 0x0090;
-    const GITS_BASER: u32 = 0x0100;
-
-    /// Errors thrown while saving/restoring the GICv3ITS.
-    #[derive(Debug)]
-    pub enum Error {
-        /// Error in saving RDIST pending tables into guest RAM.
-        SavePendingTables(crate::aarch64::gic::Error),
-        /// Error in saving GIC distributor registers.
-        SaveDistributorRegisters(crate::aarch64::gic::Error),
-        /// Error in restoring GIC distributor registers.
-        RestoreDistributorRegisters(crate::aarch64::gic::Error),
-        /// Error in saving GIC distributor control registers.
-        SaveDistributorCtrlRegisters(crate::aarch64::gic::Error),
-        /// Error in restoring GIC distributor control registers.
-        RestoreDistributorCtrlRegisters(crate::aarch64::gic::Error),
-        /// Error in saving GIC redistributor registers.
-        SaveRedistributorRegisters(crate::aarch64::gic::Error),
-        /// Error in restoring GIC redistributor registers.
-        RestoreRedistributorRegisters(crate::aarch64::gic::Error),
-        /// Error in saving GIC CPU interface registers.
-        SaveIccRegisters(crate::aarch64::gic::Error),
-        /// Error in restoring GIC CPU interface registers.
-        RestoreIccRegisters(crate::aarch64::gic::Error),
-        /// Error in saving GICv3ITS IIDR register.
-        SaveITSIIDR(crate::aarch64::gic::Error),
-        /// Error in restoring GICv3ITS IIDR register.
-        RestoreITSIIDR(crate::aarch64::gic::Error),
-        /// Error in saving GICv3ITS CBASER register.
-        SaveITSCBASER(crate::aarch64::gic::Error),
-        /// Error in restoring GICv3ITS CBASER register.
-        RestoreITSCBASER(crate::aarch64::gic::Error),
-        /// Error in saving GICv3ITS CREADR register.
-        SaveITSCREADR(crate::aarch64::gic::Error),
-        /// Error in restoring GICv3ITS CREADR register.
-        RestoreITSCREADR(crate::aarch64::gic::Error),
-        /// Error in saving GICv3ITS CWRITER register.
-        SaveITSCWRITER(crate::aarch64::gic::Error),
-        /// Error in restoring GICv3ITS CWRITER register.
-        RestoreITSCWRITER(crate::aarch64::gic::Error),
-        /// Error in saving GICv3ITS BASER register.
-        SaveITSBASER(crate::aarch64::gic::Error),
-        /// Error in restoring GICv3ITS BASER register.
-        RestoreITSBASER(crate::aarch64::gic::Error),
-        /// Error in saving GICv3ITS CTLR register.
-        SaveITSCTLR(crate::aarch64::gic::Error),
-        /// Error in restoring GICv3ITS CTLR register.
-        RestoreITSCTLR(crate::aarch64::gic::Error),
-        /// Error in saving GICv3ITS restore tables.
-        SaveITSTables(crate::aarch64::gic::Error),
-        /// Error in restoring GICv3ITS restore tables.
-        RestoreITSTables(crate::aarch64::gic::Error),
-    }
-
-    type Result<T> = result::Result<T, Error>;
-
-    /// Access an ITS device attribute.
-    ///
-    /// This is a helper function to get/set the ITS device attribute depending
-    /// the bool parameter `set` provided.
-    pub fn gicv3_its_attr_access(
-        its_device: &Arc<dyn hypervisor::Device>,
-        group: u32,
-        attr: u32,
-        val: &u64,
-        set: bool,
-    ) -> crate::aarch64::gic::Result<()> {
-        let mut gicv3_its_attr = kvm_bindings::kvm_device_attr {
-            group,
-            attr: attr as u64,
-            addr: val as *const u64 as u64,
-            flags: 0,
-        };
-        if set {
-            its_device
-                .set_device_attr(&gicv3_its_attr)
-                .map_err(crate::aarch64::gic::Error::SetDeviceAttribute)?;
-        } else {
-            its_device
-                .get_device_attr(&mut gicv3_its_attr)
-                .map_err(crate::aarch64::gic::Error::GetDeviceAttribute)?;
-        }
-        Ok(())
-    }
-
-    /// Function that saves/restores ITS tables into guest RAM.
-    ///
-    /// The tables get flushed to guest RAM whenever the VM gets stopped.
-    pub fn gicv3_its_tables_access(
-        its_device: &Arc<dyn hypervisor::Device>,
-        save: bool,
-    ) -> crate::aarch64::gic::Result<()> {
-        let attr = if save {
-            u64::from(kvm_bindings::KVM_DEV_ARM_ITS_SAVE_TABLES)
-        } else {
-            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,
-            attr,
-            addr: 0,
-            flags: 0,
-        };
-        its_device
-            .set_device_attr(&init_gic_attr)
-            .map_err(crate::aarch64::gic::Error::SetDeviceAttribute)
-    }
-
-    pub struct KvmGicV3Its {
-        /// The hypervisor agnostic device for the GicV3
-        device: Arc<dyn hypervisor::Device>,
-
-        /// The hypervisor agnostic device for the Its device
-        its_device: Option<Arc<dyn hypervisor::Device>>,
-
-        /// Vector holding values of GICR_TYPER for each vCPU
-        gicr_typers: Vec<u64>,
-
-        /// GIC device properties, to be used for setting up the fdt entry
-        gic_properties: [u64; 4],
-
-        /// MSI device properties, to be used for setting up the fdt entry
-        msi_properties: [u64; 2],
-
-        /// Number of CPUs handled by the device
-        vcpu_count: u64,
-    }
-
-    #[derive(Versionize)]
-    pub struct Gicv3ItsState {
-        dist: Vec<u32>,
-        rdist: Vec<u32>,
-        icc: Vec<u32>,
-        // special register that enables interrupts and affinity routing
-        gicd_ctlr: u32,
-        its_ctlr: u64,
-        its_iidr: u64,
-        its_cbaser: u64,
-        its_cwriter: u64,
-        its_creadr: u64,
-        its_baser: [u64; 8],
-    }
-
-    impl VersionMapped for Gicv3ItsState {}
-
-    impl KvmGicV3Its {
-        fn get_msi_size() -> u64 {
-            layout::GIC_V3_ITS_SIZE
-        }
-
-        fn get_msi_addr(vcpu_count: u64) -> u64 {
-            KvmGicV3::get_redists_addr(vcpu_count) - KvmGicV3Its::get_msi_size()
-        }
-
-        /// Save the state of GICv3ITS.
-        fn state(&self, gicr_typers: &[u64]) -> Result<Gicv3ItsState> {
-            let gicd_ctlr =
-                read_ctlr(self.device()).map_err(Error::SaveDistributorCtrlRegisters)?;
-
-            let dist_state =
-                get_dist_regs(self.device()).map_err(Error::SaveDistributorRegisters)?;
-
-            let rdist_state = get_redist_regs(self.device(), gicr_typers)
-                .map_err(Error::SaveRedistributorRegisters)?;
-
-            let icc_state =
-                get_icc_regs(self.device(), gicr_typers).map_err(Error::SaveIccRegisters)?;
-
-            let its_baser_state: [u64; 8] = [0; 8];
-            for i in 0..8 {
-                gicv3_its_attr_access(
-                    self.its_device().unwrap(),
-                    kvm_bindings::KVM_DEV_ARM_VGIC_GRP_ITS_REGS,
-                    GITS_BASER + i * 8,
-                    &its_baser_state[i as usize],
-                    false,
-                )
-                .map_err(Error::SaveITSBASER)?;
-            }
-
-            let its_ctlr_state: u64 = 0;
-            gicv3_its_attr_access(
-                self.its_device().unwrap(),
-                kvm_bindings::KVM_DEV_ARM_VGIC_GRP_ITS_REGS,
-                GITS_CTLR,
-                &its_ctlr_state,
-                false,
-            )
-            .map_err(Error::SaveITSCTLR)?;
-
-            let its_cbaser_state: u64 = 0;
-            gicv3_its_attr_access(
-                self.its_device().unwrap(),
-                kvm_bindings::KVM_DEV_ARM_VGIC_GRP_ITS_REGS,
-                GITS_CBASER,
-                &its_cbaser_state,
-                false,
-            )
-            .map_err(Error::SaveITSCBASER)?;
-
-            let its_creadr_state: u64 = 0;
-            gicv3_its_attr_access(
-                self.its_device().unwrap(),
-                kvm_bindings::KVM_DEV_ARM_VGIC_GRP_ITS_REGS,
-                GITS_CREADR,
-                &its_creadr_state,
-                false,
-            )
-            .map_err(Error::SaveITSCREADR)?;
-
-            let its_cwriter_state: u64 = 0;
-            gicv3_its_attr_access(
-                self.its_device().unwrap(),
-                kvm_bindings::KVM_DEV_ARM_VGIC_GRP_ITS_REGS,
-                GITS_CWRITER,
-                &its_cwriter_state,
-                false,
-            )
-            .map_err(Error::SaveITSCWRITER)?;
-
-            let its_iidr_state: u64 = 0;
-            gicv3_its_attr_access(
-                self.its_device().unwrap(),
-                kvm_bindings::KVM_DEV_ARM_VGIC_GRP_ITS_REGS,
-                GITS_IIDR,
-                &its_iidr_state,
-                false,
-            )
-            .map_err(Error::SaveITSIIDR)?;
-
-            Ok(Gicv3ItsState {
-                dist: dist_state,
-                rdist: rdist_state,
-                icc: icc_state,
-                gicd_ctlr,
-                its_ctlr: its_ctlr_state,
-                its_iidr: its_iidr_state,
-                its_cbaser: its_cbaser_state,
-                its_cwriter: its_cwriter_state,
-                its_creadr: its_creadr_state,
-                its_baser: its_baser_state,
-            })
-        }
-
-        /// Restore the state of GICv3ITS.
-        fn set_state(&mut self, gicr_typers: &[u64], state: &Gicv3ItsState) -> Result<()> {
-            write_ctlr(self.device(), state.gicd_ctlr)
-                .map_err(Error::RestoreDistributorCtrlRegisters)?;
-
-            set_dist_regs(self.device(), &state.dist)
-                .map_err(Error::RestoreDistributorRegisters)?;
-
-            set_redist_regs(self.device(), gicr_typers, &state.rdist)
-                .map_err(Error::RestoreRedistributorRegisters)?;
-
-            set_icc_regs(self.device(), gicr_typers, &state.icc)
-                .map_err(Error::RestoreIccRegisters)?;
-
-            //Restore GICv3ITS registers
-            gicv3_its_attr_access(
-                self.its_device().unwrap(),
-                kvm_bindings::KVM_DEV_ARM_VGIC_GRP_ITS_REGS,
-                GITS_IIDR,
-                &state.its_iidr,
-                true,
-            )
-            .map_err(Error::RestoreITSIIDR)?;
-
-            gicv3_its_attr_access(
-                self.its_device().unwrap(),
-                kvm_bindings::KVM_DEV_ARM_VGIC_GRP_ITS_REGS,
-                GITS_CBASER,
-                &state.its_cbaser,
-                true,
-            )
-            .map_err(Error::RestoreITSCBASER)?;
-
-            gicv3_its_attr_access(
-                self.its_device().unwrap(),
-                kvm_bindings::KVM_DEV_ARM_VGIC_GRP_ITS_REGS,
-                GITS_CREADR,
-                &state.its_creadr,
-                true,
-            )
-            .map_err(Error::RestoreITSCREADR)?;
-
-            gicv3_its_attr_access(
-                self.its_device().unwrap(),
-                kvm_bindings::KVM_DEV_ARM_VGIC_GRP_ITS_REGS,
-                GITS_CWRITER,
-                &state.its_cwriter,
-                true,
-            )
-            .map_err(Error::RestoreITSCWRITER)?;
-
-            for i in 0..8 {
-                gicv3_its_attr_access(
-                    self.its_device().unwrap(),
-                    kvm_bindings::KVM_DEV_ARM_VGIC_GRP_ITS_REGS,
-                    GITS_BASER + i * 8,
-                    &state.its_baser[i as usize],
-                    true,
-                )
-                .map_err(Error::RestoreITSBASER)?;
-            }
-
-            // Restore ITS tables
-            gicv3_its_tables_access(self.its_device().unwrap(), false)
-                .map_err(Error::RestoreITSTables)?;
-
-            gicv3_its_attr_access(
-                self.its_device().unwrap(),
-                kvm_bindings::KVM_DEV_ARM_VGIC_GRP_ITS_REGS,
-                GITS_CTLR,
-                &state.its_ctlr,
-                true,
-            )
-            .map_err(Error::RestoreITSCTLR)?;
-
-            Ok(())
-        }
-    }
-
-    impl GicDevice for KvmGicV3Its {
-        fn device(&self) -> &Arc<dyn hypervisor::Device> {
-            &self.device
-        }
-
-        fn its_device(&self) -> Option<&Arc<dyn hypervisor::Device>> {
-            self.its_device.as_ref()
-        }
-
-        fn fdt_compatibility(&self) -> &str {
-            "arm,gic-v3"
-        }
-
-        fn msi_compatible(&self) -> bool {
-            true
-        }
-
-        fn msi_compatibility(&self) -> &str {
-            "arm,gic-v3-its"
-        }
-
-        fn fdt_maint_irq(&self) -> u32 {
-            KvmGicV3::ARCH_GIC_V3_MAINT_IRQ
-        }
-
-        fn msi_properties(&self) -> &[u64] {
-            &self.msi_properties
-        }
-
-        fn device_properties(&self) -> &[u64] {
-            &self.gic_properties
-        }
-
-        fn vcpu_count(&self) -> u64 {
-            self.vcpu_count
-        }
-
-        fn set_its_device(&mut self, its_device: Option<Arc<dyn hypervisor::Device>>) {
-            self.its_device = its_device;
-        }
-
-        fn set_gicr_typers(&mut self, vcpu_states: &[CpuState]) {
-            let gicr_typers = construct_gicr_typers(vcpu_states);
-            self.gicr_typers = gicr_typers;
-        }
-
-        fn as_any_concrete_mut(&mut self) -> &mut dyn Any {
-            self
-        }
-    }
-
-    impl KvmGicDevice for KvmGicV3Its {
-        fn version() -> u32 {
-            KvmGicV3::version()
-        }
-
-        fn create_device(
-            device: Arc<dyn hypervisor::Device>,
-            vcpu_count: u64,
-        ) -> Box<dyn GicDevice> {
-            Box::new(KvmGicV3Its {
-                device,
-                its_device: None,
-                gicr_typers: vec![0; vcpu_count.try_into().unwrap()],
-                gic_properties: [
-                    KvmGicV3::get_dist_addr(),
-                    KvmGicV3::get_dist_size(),
-                    KvmGicV3::get_redists_addr(vcpu_count),
-                    KvmGicV3::get_redists_size(vcpu_count),
-                ],
-                msi_properties: [
-                    KvmGicV3Its::get_msi_addr(vcpu_count),
-                    KvmGicV3Its::get_msi_size(),
-                ],
-                vcpu_count,
-            })
-        }
-
-        fn init_device_attributes(
-            vm: &Arc<dyn hypervisor::Vm>,
-            gic_device: &mut dyn GicDevice,
-        ) -> crate::aarch64::gic::Result<()> {
-            KvmGicV3::init_device_attributes(vm, gic_device)?;
-
-            let mut its_device = kvm_bindings::kvm_create_device {
-                type_: kvm_bindings::kvm_device_type_KVM_DEV_TYPE_ARM_VGIC_ITS,
-                fd: 0,
-                flags: 0,
-            };
-
-            let its_fd = vm
-                .create_device(&mut its_device)
-                .map_err(crate::aarch64::gic::Error::CreateGic)?;
-
-            Self::set_device_attribute(
-                &its_fd,
-                kvm_bindings::KVM_DEV_ARM_VGIC_GRP_ADDR,
-                u64::from(kvm_bindings::KVM_VGIC_ITS_ADDR_TYPE),
-                &KvmGicV3Its::get_msi_addr(gic_device.vcpu_count()) as *const u64 as u64,
-                0,
-            )?;
-
-            Self::set_device_attribute(
-                &its_fd,
-                kvm_bindings::KVM_DEV_ARM_VGIC_GRP_CTRL,
-                u64::from(kvm_bindings::KVM_DEV_ARM_VGIC_CTRL_INIT),
-                0,
-                0,
-            )?;
-
-            gic_device.set_its_device(Some(its_fd));
-
-            Ok(())
-        }
-    }
-
-    pub const GIC_V3_ITS_SNAPSHOT_ID: &str = "gic-v3-its";
-    impl Snapshottable for KvmGicV3Its {
-        fn id(&self) -> String {
-            GIC_V3_ITS_SNAPSHOT_ID.to_string()
-        }
-
-        fn snapshot(&mut self) -> std::result::Result<Snapshot, MigratableError> {
-            let gicr_typers = self.gicr_typers.clone();
-            Snapshot::new_from_versioned_state(&self.id(), &self.state(&gicr_typers).unwrap())
-        }
-
-        fn restore(&mut self, snapshot: Snapshot) -> std::result::Result<(), MigratableError> {
-            let gicr_typers = self.gicr_typers.clone();
-            self.set_state(&gicr_typers, &snapshot.to_versioned_state(&self.id())?)
-                .map_err(|e| {
-                    MigratableError::Restore(anyhow!("Could not restore GICv3ITS state {:?}", e))
-                })
-        }
-    }
-
-    impl Pausable for KvmGicV3Its {
-        fn pause(&mut self) -> std::result::Result<(), MigratableError> {
-            // Flush redistributors pending tables to guest RAM.
-            save_pending_tables(self.device()).map_err(|e| {
-                MigratableError::Pause(anyhow!(
-                    "Could not save GICv3ITS GIC pending tables {:?}",
-                    e
-                ))
-            })?;
-            // Flush ITS tables to guest RAM.
-            gicv3_its_tables_access(self.its_device().unwrap(), true).map_err(|e| {
-                MigratableError::Pause(anyhow!("Could not save GICv3ITS ITS tables {:?}", e))
-            })?;
-
-            Ok(())
-        }
-    }
-    impl Transportable for KvmGicV3Its {}
-    impl Migratable for KvmGicV3Its {}
-}

arch/src/aarch64/gic/mod.rs

@@ -1,220 +0,0 @@
-// Copyright 2021 Arm Limited (or its affiliates). All rights reserved.
-// Copyright 2019 Amazon.com, Inc. or its affiliates. All Rights Reserved.
-// SPDX-License-Identifier: Apache-2.0
-
-pub mod dist_regs;
-pub mod gicv3;
-pub mod gicv3_its;
-pub mod icc_regs;
-pub mod redist_regs;
-
-pub use self::dist_regs::{get_dist_regs, read_ctlr, set_dist_regs, write_ctlr};
-pub use self::icc_regs::{get_icc_regs, set_icc_regs};
-pub use self::redist_regs::{get_redist_regs, set_redist_regs};
-use hypervisor::CpuState;
-use std::any::Any;
-use std::result;
-use std::sync::Arc;
-
-/// Errors thrown while setting up the GIC.
-#[derive(Debug)]
-pub enum Error {
-    /// Error while calling KVM ioctl for setting up the global interrupt controller.
-    CreateGic(hypervisor::HypervisorVmError),
-    /// Error while setting device attributes for the GIC.
-    SetDeviceAttribute(hypervisor::HypervisorDeviceError),
-    /// Error while getting device attributes for the GIC.
-    GetDeviceAttribute(hypervisor::HypervisorDeviceError),
-}
-type Result<T> = result::Result<T, Error>;
-
-pub trait GicDevice: Send {
-    /// Returns the hypervisor agnostic Device of the GIC device
-    fn device(&self) -> &Arc<dyn hypervisor::Device>;
-
-    /// Returns the hypervisor agnostic Device of the ITS device
-    fn its_device(&self) -> Option<&Arc<dyn hypervisor::Device>> {
-        None
-    }
-
-    /// Returns the fdt compatibility property of the device
-    fn fdt_compatibility(&self) -> &str;
-
-    /// Returns the maint_irq fdt property of the device
-    fn fdt_maint_irq(&self) -> u32;
-
-    /// Returns an array with GIC device properties
-    fn device_properties(&self) -> &[u64];
-
-    /// Returns the number of vCPUs this GIC handles
-    fn vcpu_count(&self) -> u64;
-
-    /// Returns whether the GIC device is MSI compatible or not
-    fn msi_compatible(&self) -> bool {
-        false
-    }
-
-    /// Returns the MSI compatibility property of the device
-    fn msi_compatibility(&self) -> &str {
-        ""
-    }
-
-    /// Returns the MSI reg property of the device
-    fn msi_properties(&self) -> &[u64] {
-        &[]
-    }
-
-    fn set_its_device(&mut self, its_device: Option<Arc<dyn hypervisor::Device>>);
-
-    /// Get the values of GICR_TYPER for each vCPU.
-    fn set_gicr_typers(&mut self, vcpu_states: &[CpuState]);
-
-    /// Downcast the trait object to its concrete type.
-    fn as_any_concrete_mut(&mut self) -> &mut dyn Any;
-}
-
-pub mod kvm {
-    use super::GicDevice;
-    use super::Result;
-    use crate::aarch64::gic::gicv3_its::kvm::KvmGicV3Its;
-    use crate::layout;
-    use hypervisor::kvm::kvm_bindings;
-    use std::boxed::Box;
-    use std::sync::Arc;
-
-    /// Trait for GIC devices.
-    pub trait KvmGicDevice: Send + Sync + GicDevice {
-        /// Returns the GIC version of the device
-        fn version() -> u32;
-
-        /// Create the GIC device object
-        fn create_device(
-            device: Arc<dyn hypervisor::Device>,
-            vcpu_count: u64,
-        ) -> Box<dyn GicDevice>;
-
-        /// Setup the device-specific attributes
-        fn init_device_attributes(
-            vm: &Arc<dyn hypervisor::Vm>,
-            gic_device: &mut dyn GicDevice,
-        ) -> Result<()>;
-
-        /// Initialize a GIC device
-        fn init_device(vm: &Arc<dyn hypervisor::Vm>) -> Result<Arc<dyn hypervisor::Device>> {
-            let mut gic_device = kvm_bindings::kvm_create_device {
-                type_: Self::version(),
-                fd: 0,
-                flags: 0,
-            };
-
-            vm.create_device(&mut gic_device)
-                .map_err(super::Error::CreateGic)
-        }
-
-        /// Set a GIC device attribute
-        fn set_device_attribute(
-            device: &Arc<dyn hypervisor::Device>,
-            group: u32,
-            attr: u64,
-            addr: u64,
-            flags: u32,
-        ) -> Result<()> {
-            let attr = kvm_bindings::kvm_device_attr {
-                flags,
-                group,
-                attr,
-                addr,
-            };
-            device
-                .set_device_attr(&attr)
-                .map_err(super::Error::SetDeviceAttribute)?;
-
-            Ok(())
-        }
-
-        /// Get a GIC device attribute
-        fn get_device_attribute(
-            device: &Arc<dyn hypervisor::Device>,
-            group: u32,
-            attr: u64,
-            addr: u64,
-            flags: u32,
-        ) -> Result<()> {
-            let mut attr = kvm_bindings::kvm_device_attr {
-                flags,
-                group,
-                attr,
-                addr,
-            };
-            device
-                .get_device_attr(&mut attr)
-                .map_err(super::Error::GetDeviceAttribute)?;
-
-            Ok(())
-        }
-
-        /// Finalize the setup of a GIC device
-        fn finalize_device(gic_device: &dyn GicDevice) -> Result<()> {
-            /* We need to tell the kernel how many irqs to support with this vgic.
-             * See the `layout` module for details.
-             */
-            let nr_irqs: u32 = layout::IRQ_NUM;
-            let nr_irqs_ptr = &nr_irqs as *const u32;
-            Self::set_device_attribute(
-                gic_device.device(),
-                kvm_bindings::KVM_DEV_ARM_VGIC_GRP_NR_IRQS,
-                0,
-                nr_irqs_ptr as u64,
-                0,
-            )?;
-
-            /* Finalize the GIC.
-             * See https://code.woboq.org/linux/linux/virt/kvm/arm/vgic/vgic-kvm-device.c.html#211.
-             */
-            Self::set_device_attribute(
-                gic_device.device(),
-                kvm_bindings::KVM_DEV_ARM_VGIC_GRP_CTRL,
-                u64::from(kvm_bindings::KVM_DEV_ARM_VGIC_CTRL_INIT),
-                0,
-                0,
-            )?;
-
-            Ok(())
-        }
-
-        /// Method to initialize the GIC device
-        #[allow(clippy::new_ret_no_self)]
-        fn new(vm: &Arc<dyn hypervisor::Vm>, vcpu_count: u64) -> Result<Box<dyn GicDevice>> {
-            let vgic_fd = Self::init_device(vm)?;
-
-            let mut device = Self::create_device(vgic_fd, vcpu_count);
-
-            Self::init_device_attributes(vm, &mut *device)?;
-
-            Self::finalize_device(&*device)?;
-
-            Ok(device)
-        }
-    }
-
-    /// Create a GICv3-ITS device.
-    ///
-    pub fn create_gic(vm: &Arc<dyn hypervisor::Vm>, vcpu_count: u64) -> Result<Box<dyn GicDevice>> {
-        debug!("creating a GICv3-ITS");
-        KvmGicV3Its::new(vm, vcpu_count)
-    }
-
-    /// Function that saves RDIST pending tables into guest RAM.
-    ///
-    /// The tables get flushed to guest RAM whenever the VM gets stopped.
-    pub fn save_pending_tables(gic: &Arc<dyn hypervisor::Device>) -> Result<()> {
-        let init_gic_attr = kvm_bindings::kvm_device_attr {
-            group: kvm_bindings::KVM_DEV_ARM_VGIC_GRP_CTRL,
-            attr: u64::from(kvm_bindings::KVM_DEV_ARM_VGIC_SAVE_PENDING_TABLES),
-            addr: 0,
-            flags: 0,
-        };
-        gic.set_device_attr(&init_gic_attr)
-            .map_err(super::Error::SetDeviceAttribute)
-    }
-}

arch/src/aarch64/mod.rs

@@ -4,23 +4,19 @@
 
 /// Module for the flattened device tree.
 pub mod fdt;
-/// Module for the global interrupt controller configuration.
-pub mod gic;
 /// Layout for this aarch64 system.
 pub mod layout;
-/// Logic for configuring aarch64 registers.
-pub mod regs;
 /// Module for loading UEFI binary.
 pub mod uefi;
 
 pub use self::fdt::DeviceInfoForFdt;
 use crate::{DeviceType, GuestMemoryMmap, NumaNodes, PciSpaceInfo, RegionType};
-use gic::GicDevice;
+use hypervisor::arch::aarch64::gic::Vgic;
 use log::{log_enabled, Level};
 use std::collections::HashMap;
 use std::convert::TryInto;
 use std::fmt::Debug;
-use std::sync::Arc;
+use std::sync::{Arc, Mutex};
 use vm_memory::{Address, GuestAddress, GuestMemory, GuestUsize};
 
 /// Errors thrown while configuring aarch64 system.
@@ -33,13 +29,13 @@ pub enum Error {
     WriteFdtToMemory(fdt::Error),
 
     /// Failed to create a GIC.
-    SetupGic(gic::Error),
+    SetupGic,
 
     /// Failed to compute the initramfs address.
     InitramfsAddress,
 
     /// Error configuring the general purpose registers
-    RegsConfiguration(regs::Error),
+    RegsConfiguration(hypervisor::HypervisorCpuError),
 
     /// Error configuring the MPIDR register
     VcpuRegMpidr(hypervisor::HypervisorCpuError),
@@ -69,8 +65,12 @@ pub fn configure_vcpu(
     kernel_entry_point: Option<EntryPoint>,
 ) -> super::Result<u64> {
     if let Some(kernel_entry_point) = kernel_entry_point {
-        regs::setup_regs(vcpu, id, kernel_entry_point.entry_addr.raw_value())
+        vcpu.setup_regs(
-            .map_err(Error::RegsConfiguration)?;
+            id,
+            kernel_entry_point.entry_addr.raw_value(),
+            super::layout::FDT_START.raw_value(),
+        )
+        .map_err(Error::RegsConfiguration)?;
     }
 
     let mpidr = vcpu.read_mpidr().map_err(Error::VcpuRegMpidr)?;
@@ -143,7 +143,7 @@ pub fn configure_system<T: DeviceInfoForFdt + Clone + Debug, S: ::std::hash::Bui
     initrd: &Option<super::InitramfsConfig>,
     pci_space_info: &[PciSpaceInfo],
     virtio_iommu_bdf: Option<u32>,
-    gic_device: &dyn GicDevice,
+    gic_device: &Arc<Mutex<dyn Vgic>>,
     numa_nodes: &NumaNodes,
     pmu_supported: bool,
 ) -> super::Result<()> {

arch/src/aarch64/regs.rs

@@ -1,75 +0,0 @@
-// Copyright 2019 Amazon.com, Inc. or its affiliates. All Rights Reserved.
-// SPDX-License-Identifier: Apache-2.0
-//
-// Portions Copyright 2017 The Chromium OS Authors. All rights reserved.
-// Use of this source code is governed by a BSD-style license that can be
-// found in the THIRD-PARTY file.
-
-use hypervisor::kvm::kvm_bindings::{
-    kvm_regs, user_pt_regs, KVM_REG_ARM64, KVM_REG_ARM_CORE, KVM_REG_SIZE_U64,
-};
-use hypervisor::{arm64_core_reg_id, offset__of};
-use std::sync::Arc;
-use std::{mem, result};
-use vm_memory::Address;
-
-/// Errors thrown while setting aarch64 registers.
-#[derive(Debug)]
-pub enum Error {
-    /// Failed to set core register (PC, PSTATE or general purpose ones).
-    SetCoreRegister(hypervisor::HypervisorCpuError),
-    /// Failed to get a system register.
-    GetSysRegister(hypervisor::HypervisorCpuError),
-}
-type Result<T> = result::Result<T, Error>;
-
-#[allow(non_upper_case_globals)]
-// PSR (Processor State Register) bits.
-// Taken from arch/arm64/include/uapi/asm/ptrace.h.
-const PSR_MODE_EL1h: u64 = 0x0000_0005;
-const PSR_F_BIT: u64 = 0x0000_0040;
-const PSR_I_BIT: u64 = 0x0000_0080;
-const PSR_A_BIT: u64 = 0x0000_0100;
-const PSR_D_BIT: u64 = 0x0000_0200;
-// Taken from arch/arm64/kvm/inject_fault.c.
-const PSTATE_FAULT_BITS_64: u64 = PSR_MODE_EL1h | PSR_A_BIT | PSR_F_BIT | PSR_I_BIT | PSR_D_BIT;
-
-/// Configure core registers for a given CPU.
-///
-/// # Arguments
-///
-/// * `vcpu` - Structure for the VCPU that holds the VCPU's fd.
-/// * `cpu_id` - Index of current vcpu.
-/// * `boot_ip` - Starting instruction pointer.
-/// * `mem` - Reserved DRAM for current VM.
-pub fn setup_regs(vcpu: &Arc<dyn hypervisor::Vcpu>, cpu_id: u8, boot_ip: u64) -> Result<()> {
-    let kreg_off = offset__of!(kvm_regs, regs);
-
-    // Get the register index of the PSTATE (Processor State) register.
-    let pstate = offset__of!(user_pt_regs, pstate) + kreg_off;
-    vcpu.set_reg(
-        arm64_core_reg_id!(KVM_REG_SIZE_U64, pstate),
-        PSTATE_FAULT_BITS_64,
-    )
-    .map_err(Error::SetCoreRegister)?;
-
-    // Other vCPUs are powered off initially awaiting PSCI wakeup.
-    if cpu_id == 0 {
-        // Setting the PC (Processor Counter) to the current program address (kernel address).
-        let pc = offset__of!(user_pt_regs, pc) + kreg_off;
-        vcpu.set_reg(arm64_core_reg_id!(KVM_REG_SIZE_U64, pc), boot_ip as u64)
-            .map_err(Error::SetCoreRegister)?;
-
-        // Last mandatory thing to set -> the address pointing to the FDT (also called DTB).
-        // "The device tree blob (dtb) must be placed on an 8-byte boundary and must
-        // not exceed 2 megabytes in size." -> https://www.kernel.org/doc/Documentation/arm64/booting.txt.
-        // We are choosing to place it the end of DRAM. See `get_fdt_addr`.
-        let regs0 = offset__of!(user_pt_regs, regs) + kreg_off;
-        vcpu.set_reg(
-            arm64_core_reg_id!(KVM_REG_SIZE_U64, regs0),
-            super::layout::FDT_START.raw_value(),
-        )
-        .map_err(Error::SetCoreRegister)?;
-    }
-    Ok(())
-}

arch/src/lib.rs

@@ -54,7 +54,7 @@ pub enum Error {
 pub type Result<T> = result::Result<T, Error>;
 
 /// Type for memory region types.
-#[derive(Clone, Copy, PartialEq, Debug, Serialize, Deserialize, Versionize)]
+#[derive(Clone, Copy, PartialEq, Eq, Debug, Serialize, Deserialize, Versionize)]
 pub enum RegionType {
     /// RAM type
     Ram,

block_util/Cargo.toml

@@ -17,7 +17,7 @@ versionize = "0.1.6"
 versionize_derive = "0.1.4"
 vhdx = { path = "../vhdx" }
 virtio-bindings = { version = "0.1.0", features = ["virtio-v5_0_0"] }
-virtio-queue = "0.3.0"
+virtio-queue = "0.4.0"
 vm-memory = { version = "0.8.0", features = ["backend-mmap", "backend-atomic", "backend-bitmap"] }
 vm-virtio = { path = "../vm-virtio" }
 vmm-sys-util = "0.9.0"

block_util/src/async_io.rs

@@ -104,7 +104,7 @@ impl DiskTopology {
 
 pub type DiskFileResult<T> = std::result::Result<T, DiskFileError>;
 
-pub trait DiskFile: Send + Sync {
+pub trait DiskFile: Send {
     fn size(&mut self) -> DiskFileResult<u64>;
     fn new_async_io(&self, ring_depth: u32) -> DiskFileResult<Box<dyn AsyncIo>>;
     fn topology(&mut self) -> DiskTopology {
@@ -127,7 +127,7 @@ pub enum AsyncIoError {
 
 pub type AsyncIoResult<T> = std::result::Result<T, AsyncIoError>;
 
-pub trait AsyncIo: Send + Sync {
+pub trait AsyncIo: Send {
     fn notifier(&self) -> &EventFd;
     fn read_vectored(
         &mut self,

block_util/src/lib.rs

@@ -138,7 +138,7 @@ impl ExecuteError {
     }
 }
 
-#[derive(Clone, Copy, Debug, PartialEq)]
+#[derive(Clone, Copy, Debug, PartialEq, Eq)]
 pub enum RequestType {
     In,
     Out,

devices/Cargo.toml

@@ -6,11 +6,12 @@ edition = "2021"
 
 [dependencies]
 acpi_tables = { path = "../acpi_tables" }
-anyhow = "1.0.57"
+anyhow = "1.0.58"
 arch = { path = "../arch" }
 bitflags = "1.3.2"
 byteorder = "1.4.3"
 epoll = "4.3.1"
+hypervisor = { path = "../hypervisor" }
 libc = "0.2.126"
 log = "0.4.17"
 versionize = "0.1.6"

devices/src/gic.rs

@@ -4,30 +4,34 @@
 
 use super::interrupt_controller::{Error, InterruptController};
 extern crate arch;
-use arch::aarch64::gic::GicDevice;
+use anyhow::anyhow;
+use arch::layout;
+use hypervisor::{arch::aarch64::gic::Vgic, CpuState};
 use std::result;
 use std::sync::{Arc, Mutex};
 use vm_device::interrupt::{
     InterruptIndex, InterruptManager, InterruptSourceConfig, InterruptSourceGroup,
     LegacyIrqSourceConfig, MsiIrqGroupConfig,
 };
+use vm_memory::Address;
+use vm_migration::{Migratable, MigratableError, Pausable, Snapshot, Snapshottable, Transportable};
 use vmm_sys_util::eventfd::EventFd;
 
 type Result<T> = result::Result<T, Error>;
 
-// Reserve 32 IRQs for legacy device.
+// Reserve 32 IRQs for legacy devices.
-pub const IRQ_LEGACY_BASE: usize = arch::layout::IRQ_BASE as usize;
+pub const IRQ_LEGACY_BASE: usize = layout::IRQ_BASE as usize;
 pub const IRQ_LEGACY_COUNT: usize = 32;
 
-// This Gic struct implements InterruptController to provide interrupt delivery service.
+// Gic (Generic Interupt Controller) struct provides all the functionality of a
-// The Gic source files in arch/ folder maintain the Aarch64 specific Gic device.
+// GIC device. It wraps a hypervisor-emulated GIC device (Vgic) provided by the
-// The 2 Gic instances could be merged together.
+// `hypervisor` crate.
-// Leave this refactoring to future. Two options may be considered:
+// Gic struct also implements InterruptController to provide interrupt delivery
-//   1. Move Gic*.rs from arch/ folder here.
+// service.
-//   2. Move this file and ioapic.rs to arch/, as they are architecture specific.
 pub struct Gic {
     interrupt_source_group: Arc<dyn InterruptSourceGroup>,
-    gic_device: Option<Arc<Mutex<Box<dyn GicDevice>>>>,
+    // The hypervisor agnostic virtual GIC
+    vgic: Option<Arc<Mutex<dyn Vgic>>>,
 }
 
 impl Gic {
@@ -44,16 +48,32 @@ impl Gic {
 
         Ok(Gic {
             interrupt_source_group,
-            gic_device: None,
+            vgic: None,
         })
     }
 
-    pub fn set_gic_device(&mut self, gic_device: Arc<Mutex<Box<dyn GicDevice>>>) {
+    pub fn create_vgic(
-        self.gic_device = Some(gic_device);
+        &mut self,
+        vm: &Arc<dyn hypervisor::Vm>,
+        vcpu_count: u64,
+    ) -> Result<Arc<Mutex<dyn Vgic>>> {
+        let vgic = vm
+            .create_vgic(
+                vcpu_count,
+                layout::GIC_V3_DIST_START.raw_value(),
+                layout::GIC_V3_DIST_SIZE,
+                layout::GIC_V3_REDIST_SIZE,
+                layout::GIC_V3_ITS_SIZE,
+                layout::IRQ_NUM,
+            )
+            .map_err(Error::CreateGic)?;
+        self.vgic = Some(vgic.clone());
+        Ok(vgic.clone())
     }
 
-    pub fn get_gic_device(&self) -> Option<&Arc<Mutex<Box<dyn GicDevice>>>> {
+    pub fn set_gicr_typers(&mut self, vcpu_states: &[CpuState]) {
-        self.gic_device.as_ref()
+        let vgic = self.vgic.as_ref().unwrap().clone();
+        vgic.lock().unwrap().set_gicr_typers(vcpu_states);
     }
 }
 
@@ -97,3 +117,43 @@ impl InterruptController for Gic {
         self.interrupt_source_group.notifier(irq as InterruptIndex)
     }
 }
+
+pub const GIC_V3_ITS_SNAPSHOT_ID: &str = "gic-v3-its";
+impl Snapshottable for Gic {
+    fn id(&self) -> String {
+        GIC_V3_ITS_SNAPSHOT_ID.to_string()
+    }
+
+    fn snapshot(&mut self) -> std::result::Result<Snapshot, MigratableError> {
+        let vgic = self.vgic.as_ref().unwrap().clone();
+        let state = vgic.lock().unwrap().state().unwrap();
+        Snapshot::new_from_state(&self.id(), &state)
+    }
+
+    fn restore(&mut self, snapshot: Snapshot) -> std::result::Result<(), MigratableError> {
+        let vgic = self.vgic.as_ref().unwrap().clone();
+        vgic.lock()
+            .unwrap()
+            .set_state(&snapshot.to_state(&self.id())?)
+            .map_err(|e| {
+                MigratableError::Restore(anyhow!("Could not restore GICv3ITS state {:?}", e))
+            })?;
+        Ok(())
+    }
+}
+
+impl Pausable for Gic {
+    fn pause(&mut self) -> std::result::Result<(), MigratableError> {
+        // Flush tables to guest RAM
+        let vgic = self.vgic.as_ref().unwrap().clone();
+        vgic.lock().unwrap().save_data_tables().map_err(|e| {
+            MigratableError::Pause(anyhow!(
+                "Could not save GICv3ITS GIC pending tables {:?}",
+                e
+            ))
+        })?;
+        Ok(())
+    }
+}
+impl Transportable for Gic {}
+impl Migratable for Gic {}

devices/src/interrupt_controller.rs

@@ -24,6 +24,8 @@ pub enum Error {
     UpdateInterrupt(io::Error),
     /// Failed enabling the interrupt.
     EnableInterrupt(io::Error),
+    /// Failed creating GIC device.
+    CreateGic(hypervisor::HypervisorVmError),
 }
 
 type Result<T> = result::Result<T, Error>;

docs/api.md

@@ -89,6 +89,7 @@ Trigger power button of the VM     | `/vm.power-button`   | N/A
 Pause the VM                       | `/vm.pause`          | N/A                       | N/A                      | The VM is booted
 Resume the VM                      | `/vm.resume`         | N/A                       | N/A                      | The VM is paused
 Task a snapshot of the VM          | `/vm.snapshot`       | `/schemas/VmSnapshotConfig`| N/A                     | The VM is paused
+Perform a coredump of the VM       | `/vm.coredump`       | `/schemas/VmCoredumpData` | N/A                      | The VM is paused
 Restore the VM from a snapshot     | `/vm.restore`        | `/schemas/RestoreConfig`  | N/A                      | The VM is created but not booted
 Add/remove CPUs to/from the VM     | `/vm.resize`         | `/schemas/VmResize`       | N/A                      | The VM is booted
 Add/remove memory from the VM      | `/vm.resize`         | `/schemas/VmResize`       | N/A                      | The VM is booted

docs/device_model.md

@@ -6,22 +6,22 @@ This document describes the device model supported by `cloud-hypervisor`.
 
 | Device | Build configurable | Enabled by default | Runtime configurable |
 | :----: | :----: | :----: | :----: |
-| Serial port | :negative_squared_cross_mark: | :negative_squared_cross_mark: | :heavy_check_mark: |
+| Serial port | :x: | :x: | :heavy_check_mark: |
-| RTC/CMOS | :heavy_check_mark: | :heavy_check_mark: | :negative_squared_cross_mark: |
+| RTC/CMOS | :heavy_check_mark: | :heavy_check_mark: | :x: |
-| I/O APIC | :negative_squared_cross_mark: | :negative_squared_cross_mark: | :heavy_check_mark: |
+| I/O APIC | :x: | :x: | :heavy_check_mark: |
-| i8042 shutdown/reboot | :negative_squared_cross_mark: | :negative_squared_cross_mark: | :negative_squared_cross_mark: |
+| i8042 shutdown/reboot | :x: | :x: | :x: |
-| ACPI shutdown/reboot | :negative_squared_cross_mark: | :heavy_check_mark: | :negative_squared_cross_mark: |
+| ACPI shutdown/reboot | :x: | :heavy_check_mark: | :x: |
-| virtio-blk | :negative_squared_cross_mark: | :negative_squared_cross_mark: | :heavy_check_mark: |
+| virtio-blk | :x: | :x: | :heavy_check_mark: |
-| virtio-console | :negative_squared_cross_mark: | :negative_squared_cross_mark: | :heavy_check_mark: |
+| virtio-console | :x: | :x: | :heavy_check_mark: |
-| virtio-iommu | :negative_squared_cross_mark: | :negative_squared_cross_mark: | :heavy_check_mark: |
+| virtio-iommu | :x: | :x: | :heavy_check_mark: |
-| virtio-net | :negative_squared_cross_mark: | :negative_squared_cross_mark: | :heavy_check_mark: |
+| virtio-net | :x: | :x: | :heavy_check_mark: |
-| virtio-pmem | :negative_squared_cross_mark: | :negative_squared_cross_mark: | :heavy_check_mark: |
+| virtio-pmem | :x: | :x: | :heavy_check_mark: |
-| virtio-rng | :negative_squared_cross_mark: | :negative_squared_cross_mark: | :heavy_check_mark: |
+| virtio-rng | :x: | :x: | :heavy_check_mark: |
-| virtio-vsock | :negative_squared_cross_mark: | :negative_squared_cross_mark: | :heavy_check_mark: |
+| virtio-vsock | :x: | :x: | :heavy_check_mark: |
-| vhost-user-blk | :negative_squared_cross_mark: | :negative_squared_cross_mark: | :heavy_check_mark: |
+| vhost-user-blk | :x: | :x: | :heavy_check_mark: |
-| vhost-user-fs | :negative_squared_cross_mark: | :negative_squared_cross_mark: | :heavy_check_mark: |
+| vhost-user-fs | :x: | :x: | :heavy_check_mark: |
-| vhost-user-net | :negative_squared_cross_mark: | :negative_squared_cross_mark: | :heavy_check_mark: |
+| vhost-user-net | :x: | :x: | :heavy_check_mark: |
-| VFIO | :heavy_check_mark: | :negative_squared_cross_mark: | :heavy_check_mark: |
+| VFIO | :heavy_check_mark: | :x: | :heavy_check_mark: |
 
 ## Legacy devices
 

docs/profiling.md

@@ -43,3 +43,22 @@ $ perf report -g
 ```
 
 If profiling with a network device attached either the TAP device must be already created and configured or the profiling must be done as root so that the TAP device can be created.
+
+## Userspace only profiling with LBR
+
+The use of LBR (Last Branch Record; available since Haswell) offers lower
+overhead if only userspace profiling is required. This lower overhead can allow
+a higher frequency of sampling. This also removes the requirement to compile
+with custom `RUSTFLAGS` however debug symbols should still be included:
+
+e.g.
+
+```
+$ perf record --call-graph lbr --all-user --user-callchains -g target/release/cloud-hypervisor \
+        --kernel ~/src/linux/vmlinux \
+        --pmem file=~/workloads/focal.raw \
+        --cpus boot=1 --memory size=1G \
+        --cmdline "root=/dev/pmem0p1 console=ttyS0" \
+        --serial tty --console off \
+        --api-socket=/tmp/api1
+```

event_monitor/Cargo.toml

@@ -6,5 +6,5 @@ edition = "2021"
 
 [dependencies]
 libc = "0.2.126"
-serde = { version = "1.0.137", features = ["rc", "derive"] }
+serde = { version = "1.0.138", features = ["rc", "derive"] }
-serde_json = "1.0.81"
+serde_json = "1.0.82"

fuzz/Cargo.lock

@@ -11,9 +11,9 @@ dependencies = [
 
 [[package]]
 name = "anyhow"
-version = "1.0.57"
+version = "1.0.58"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "08f9b8508dccb7687a1d6c4ce66b2b0ecef467c94667de27d8d7fe1f8d2a9cdc"
+checksum = "bb07d2053ccdbe10e2af2995a2f116c1330396493dc1269f6a91d0ae82e19704"
 
 [[package]]
 name = "api_client"
@@ -24,9 +24,9 @@ dependencies = [
 
 [[package]]
 name = "arbitrary"
-version = "1.1.0"
+version = "1.1.3"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "c38b6b6b79f671c25e1a3e785b7b82d7562ffc9cd3efdc98627e5668a2472490"
+checksum = "5a7924531f38b1970ff630f03eb20a2fde69db5c590c93b0f3482e95dcc5fd60"
 
 [[package]]
 name = "arc-swap"
@@ -127,15 +127,15 @@ checksum = "baf1de4339761588bc0619e3cbc0120ee582ebb74b53b4efbf79117bd2da40fd"
 
 [[package]]
 name = "clap"
-version = "3.1.18"
+version = "3.2.8"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "d2dbdf4bdacb33466e854ce889eee8dfd5729abf7ccd7664d0a2d60cd384440b"
+checksum = "190814073e85d238f31ff738fcb0bf6910cedeb73376c87cd69291028966fd83"
 dependencies = [
  "atty",
  "bitflags",
  "clap_lex",
  "indexmap",
- "lazy_static",
+ "once_cell",
  "strsim",
  "termcolor",
  "terminal_size",
@@ -144,16 +144,16 @@ dependencies = [
 
 [[package]]
 name = "clap_lex"
-version = "0.2.0"
+version = "0.2.4"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "a37c35f1112dad5e6e0b1adaff798507497a18fceeb30cceb3bae7d1427b9213"
+checksum = "2850f2f5a82cbf437dd5af4d49848fbdfc27c157c3d010345776f952765261c5"
 dependencies = [
  "os_str_bytes",
 ]
 
 [[package]]
 name = "cloud-hypervisor"
-version = "23.0.0"
+version = "24.0.0"
 dependencies = [
  "anyhow",
  "api_client",
@@ -216,6 +216,7 @@ dependencies = [
  "bitflags",
  "byteorder",
  "epoll",
+ "hypervisor",
  "libc",
  "log",
  "versionize",
@@ -251,35 +252,11 @@ version = "0.1.3"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "b643857cf70949306b81d7e92cb9d47add673868edac9863c4a49c42feaf3f1e"
 
-[[package]]
-name = "gdbstub"
-version = "0.6.1"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-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 = "0.2.6"
+version = "0.2.7"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "9be70c98951c83b8d2f8f60d7065fa6d5146873094452a1008da8c2f1e4205ad"
+checksum = "4eb1a864a501629691edf6c15a593b7a51eebaa1e8468e9ddc623de7c9b58ec6"
 dependencies = [
  "cfg-if",
  "libc",
@@ -288,9 +265,9 @@ dependencies = [
 
 [[package]]
 name = "hashbrown"
-version = "0.11.2"
+version = "0.12.1"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "ab5ef0d4909ef3724cc8cce6ccc8572c5c817592e9285f5464f8e86f8bd3726e"
+checksum = "db0d4cf898abf0081f964436dc980e96670a0f36863e4b83aaacdb65c9d7ccc3"
 
 [[package]]
 name = "hermit-abi"
@@ -331,9 +308,9 @@ dependencies = [
 
 [[package]]
 name = "indexmap"
-version = "1.8.1"
+version = "1.9.1"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "0f647032dfaa1f8b6dc29bd3edb7bbef4861b8b8007ebb118d6db284fd59f6ee"
+checksum = "10a35a97730320ffe8e2d410b5d3b69279b98d2c14bdb8b70ea89ecf7888d41e"
 dependencies = [
  "autocfg",
  "hashbrown",
@@ -416,12 +393,6 @@ dependencies = [
  "cfg-if",
 ]
 
-[[package]]
-name = "managed"
-version = "0.8.0"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "0ca88d725a0a943b096803bd34e73a4437208b6077654cc4ecb2947a5f91618d"
-
 [[package]]
 name = "micro_http"
 version = "0.1.0"
@@ -458,20 +429,11 @@ dependencies = [
  "vmm-sys-util",
 ]
 
-[[package]]
-name = "num-traits"
-version = "0.2.15"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "578ede34cf02f8924ab9447f50c28075b4d3e5b269972345e7e0372b38c6cdcd"
-dependencies = [
- "autocfg",
-]
-
 [[package]]
 name = "once_cell"
-version = "1.12.0"
+version = "1.13.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "7709cef83f0c1f58f666e746a08b21e0085f7440fa6a29cc194d68aac97a4225"
+checksum = "18a6dbe30758c9f83eb00cbea4ac95966305f5a7772f3f42ebfc7fc7eddbd8e1"
 
 [[package]]
 name = "option_parser"
@@ -479,15 +441,9 @@ version = "0.1.0"
 
 [[package]]
 name = "os_str_bytes"
-version = "6.0.1"
+version = "6.1.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "029d8d0b2f198229de29dca79676f2738ff952edf3fde542eb8bf94d8c21b435"
+checksum = "21326818e99cfe6ce1e524c2a805c189a99b5ae555a35d19f9a284b427d86afa"
-
-[[package]]
-name = "paste"
-version = "1.0.7"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "0c520e05135d6e763148b6426a837e239041653ba7becd2e538c076c738025fc"
 
 [[package]]
 name = "pci"
@@ -514,9 +470,9 @@ dependencies = [
 
 [[package]]
 name = "proc-macro2"
-version = "1.0.39"
+version = "1.0.40"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "c54b25569025b7fc9651de43004ae593a75ad88543b17178aa5e1b9c4f15f56f"
+checksum = "dd96a1e8ed2596c337f8eae5f24924ec83f5ad5ab21ea8e455d3566c69fbcaf7"
 dependencies = [
  "unicode-ident",
 ]
@@ -534,9 +490,9 @@ dependencies = [
 
 [[package]]
 name = "quote"
-version = "1.0.18"
+version = "1.0.20"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "a1feb54ed693b93a84e14094943b84b7c4eae204c512b7ccb95ab0c66d278ad1"
+checksum = "3bcdf212e9776fbcb2d23ab029360416bb1706b1aea2d1a5ba002727cbcab804"
 dependencies = [
  "proc-macro2",
 ]
@@ -587,24 +543,24 @@ dependencies = [
 
 [[package]]
 name = "semver"
-version = "1.0.9"
+version = "1.0.12"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "8cb243bdfdb5936c8dc3c45762a19d12ab4550cdc753bc247637d4ec35a040fd"
+checksum = "a2333e6df6d6598f2b1974829f853c2b4c5f4a6e503c10af918081aa6f8564e1"
 
 [[package]]
 name = "serde"
-version = "1.0.137"
+version = "1.0.138"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "61ea8d54c77f8315140a05f4c7237403bf38b72704d031543aa1d16abbf517d1"
+checksum = "1578c6245786b9d168c5447eeacfb96856573ca56c9d68fdcf394be134882a47"
 dependencies = [
  "serde_derive",
 ]
 
 [[package]]
 name = "serde_derive"
-version = "1.0.137"
+version = "1.0.138"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "1f26faba0c3959972377d3b2d306ee9f71faee9714294e41bb777f83f88578be"
+checksum = "023e9b1467aef8a10fb88f25611870ada9800ef7e22afce356bb0d2387b6f27c"
 dependencies = [
  "proc-macro2",
  "quote",
@@ -613,9 +569,9 @@ dependencies = [
 
 [[package]]
 name = "serde_json"
-version = "1.0.81"
+version = "1.0.82"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "9b7ce2b32a1aed03c558dc61a5cd328f15aff2dbc17daad8fb8af04d2100e15c"
+checksum = "82c2c1fdcd807d1098552c5b9a36e425e42e9fbd7c6a37a8425f390f781f7fa7"
 dependencies = [
  "itoa",
  "ryu",
@@ -655,9 +611,9 @@ checksum = "73473c0e59e6d5812c5dfe2a064a6444949f089e20eec9a2e5506596494e4623"
 
 [[package]]
 name = "syn"
-version = "1.0.95"
+version = "1.0.98"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "fbaf6116ab8924f39d52792136fb74fd60a80194cf1b1c6ffa6453eef1c3f942"
+checksum = "c50aef8a904de4c23c788f104b7dddc7d6f79c647c7c8ce4cc8f73eb0ca773dd"
 dependencies = [
  "proc-macro2",
  "quote",
@@ -714,15 +670,15 @@ dependencies = [
 
 [[package]]
 name = "unicode-ident"
-version = "1.0.0"
+version = "1.0.1"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "d22af068fba1eb5edcb4aea19d382b2a3deb4c8f9d475c589b6ada9e0fd493ee"
+checksum = "5bd2fe26506023ed7b5e1e315add59d6f584c621d037f9368fea9cfb988f368c"
 
 [[package]]
 name = "uuid"
-version = "1.0.0"
+version = "1.1.2"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "8cfcd319456c4d6ea10087ed423473267e1a071f3bc0aa89f80d60997843c6f0"
+checksum = "dd6469f4314d5f1ffec476e05f17cc9a78bc7a27a6a857842170bdf8d6f98d2f"
 dependencies = [
  "getrandom",
 ]
@@ -948,15 +904,13 @@ dependencies = [
  "devices",
  "epoll",
  "event_monitor",
- "gdbstub",
- "gdbstub_arch",
  "hypervisor",
- "lazy_static",
  "libc",
  "linux-loader",
  "log",
  "micro_http",
  "net_util",
+ "once_cell",
  "option_parser",
  "pci",
  "qcow",
@@ -995,9 +949,9 @@ dependencies = [
 
 [[package]]
 name = "wasi"
-version = "0.10.2+wasi-snapshot-preview1"
+version = "0.11.0+wasi-snapshot-preview1"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "fd6fbd9a79829dd1ad0cc20627bf1ed606756a7f77edff7b66b7064f9cb327c6"
+checksum = "9c8d87e72b64a3b4db28d11ce29237c246188f4f51057d65a7eab63b7987e423"
 
 [[package]]
 name = "winapi"

fuzz/Cargo.toml

@@ -16,7 +16,7 @@ qcow = { path = "../qcow" }
 seccompiler = "0.2.0"
 vhdx = { path = "../vhdx" }
 virtio-devices = { path = "../virtio-devices" }
-virtio-queue = "0.3.0"
+virtio-queue = "0.4.0"
 vmm-sys-util = "0.9.0"
 vm-virtio = { path = "../vm-virtio" }
 vm-memory = "0.8.0"

hypervisor/Cargo.toml

@@ -11,7 +11,7 @@ mshv = ["mshv-ioctls", "mshv-bindings"]
 tdx = []
 
 [dependencies]
-anyhow = "1.0.57"
+anyhow = "1.0.58"
 epoll = "4.3.1"
 thiserror = "1.0.31"
 libc = "0.2.126"
@@ -20,8 +20,8 @@ kvm-ioctls = { version = "0.11.0", 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 = "1.0.137", features = ["rc", "derive"] }
+serde = { version = "1.0.138", features = ["rc", "derive"] }
-serde_json = "1.0.81"
+serde_json = "1.0.82"
 vm-memory = { version = "0.8.0", features = ["backend-mmap", "backend-atomic"] }
 vmm-sys-util = { version = "0.9.0", features = ["with-serde"] }
 

hypervisor/src/arch/aarch64/gic.rs

@@ -0,0 +1,59 @@
+// Copyright 2022 Arm Limited (or its affiliates). All rights reserved.
+
+use crate::{CpuState, Device, GicState, HypervisorDeviceError, HypervisorVmError};
+use std::any::Any;
+use std::result;
+use std::sync::Arc;
+
+/// Errors thrown while setting up the VGIC.
+#[derive(Debug)]
+pub enum Error {
+    /// Error while calling KVM ioctl for setting up the global interrupt controller.
+    CreateGic(HypervisorVmError),
+    /// Error while setting device attributes for the GIC.
+    SetDeviceAttribute(HypervisorDeviceError),
+    /// Error while getting device attributes for the GIC.
+    GetDeviceAttribute(HypervisorDeviceError),
+}
+pub type Result<T> = result::Result<T, Error>;
+
+/// Hypervisor agnostic interface for a virtualized GIC
+pub trait Vgic: Send + Sync {
+    /// Returns the fdt compatibility property of the device
+    fn fdt_compatibility(&self) -> &str;
+
+    /// Returns the maint_irq fdt property of the device
+    fn fdt_maint_irq(&self) -> u32;
+
+    /// Returns an array with GIC device properties
+    fn device_properties(&self) -> [u64; 4];
+
+    /// Returns the number of vCPUs this GIC handles
+    fn vcpu_count(&self) -> u64;
+
+    /// Returns whether the GIC device is MSI compatible or not
+    fn msi_compatible(&self) -> bool;
+
+    /// Returns the MSI compatibility property of the device
+    fn msi_compatibility(&self) -> &str;
+
+    /// Returns the MSI reg property of the device
+    fn msi_properties(&self) -> [u64; 2];
+
+    fn set_its_device(&mut self, its_device: Option<Arc<dyn Device>>);
+
+    /// Get the values of GICR_TYPER for each vCPU.
+    fn set_gicr_typers(&mut self, vcpu_states: &[CpuState]);
+
+    /// Downcast the trait object to its concrete type.
+    fn as_any_concrete_mut(&mut self) -> &mut dyn Any;
+
+    /// Save the state of GICv3ITS.
+    fn state(&self) -> Result<GicState>;
+
+    /// Restore the state of GICv3ITS.
+    fn set_state(&mut self, state: &GicState) -> Result<()>;
+
+    /// Saves GIC internal data tables into RAM.
+    fn save_data_tables(&self) -> Result<()>;
+}

hypervisor/src/arch/aarch64/mod.rs

@@ -0,0 +1,3 @@
+// Copyright 2022 Arm Limited (or its affiliates). All rights reserved.
+
+pub mod gic;

hypervisor/src/arch/emulator/mod.rs

@@ -99,7 +99,7 @@ pub enum EmulationError<T: Debug> {
 /// A CpuState is an architecture specific type, representing a CPU state.
 /// The emulator and its instruction handlers modify a given CPU state and
 /// eventually ask the platform to commit it back through `set_cpu_state`.
-pub trait PlatformEmulator: Send + Sync {
+pub trait PlatformEmulator {
     type CpuState: Clone;
 
     /// Read guest memory into a u8 slice.

hypervisor/src/arch/mod.rs

@@ -10,9 +10,14 @@
 //
 // Copyright © 2020, Microsoft Corporation
 //
+// Copyright 2022 Arm Limited (or its affiliates). All rights reserved.
+//
 
 pub mod emulator;
 
 #[cfg(target_arch = "x86_64")]
 #[macro_use]
 pub mod x86;
+
+#[cfg(target_arch = "aarch64")]
+pub mod aarch64;

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

@@ -28,14 +28,11 @@ fn get_op<T: CpuStateManager>(
         )));
     }
 
-    match op_size {
+    if !matches!(op_size, 1 | 2 | 4 | 8) {
-        1 | 2 | 4 | 8 => {}
+        return Err(PlatformError::InvalidOperand(anyhow!(
-        _ => {
+            "Invalid operand size {:?}",
-            return Err(PlatformError::InvalidOperand(anyhow!(
+            op_size
-                "Invalid operand size {:?}",
+        )));
-                op_size
-            )))
-        }
     }
 
     let value = match insn
@@ -81,14 +78,11 @@ fn set_op<T: CpuStateManager>(
         )));
     }
 
-    match op_size {
+    if !matches!(op_size, 1 | 2 | 4 | 8) {
-        1 | 2 | 4 | 8 => {}
+        return Err(PlatformError::InvalidOperand(anyhow!(
-        _ => {
+            "Invalid operand size {:?}",
-            return Err(PlatformError::InvalidOperand(anyhow!(
+            op_size
-                "Invalid operand size {:?}",
+        )));
-                op_size
-            )))
-        }
     }
 
     match insn

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

@@ -16,7 +16,7 @@ use iced_x86::*;
 mod instructions;
 
 /// x86 CPU modes
-#[derive(Debug, PartialEq)]
+#[derive(Debug, PartialEq, Eq)]
 pub enum CpuMode {
     /// Real mode
     Real,

hypervisor/src/cpu.rs

@@ -453,6 +453,11 @@ pub trait Vcpu: Send + Sync {
     #[cfg(any(target_arch = "arm", target_arch = "aarch64"))]
     fn read_mpidr(&self) -> Result<u64>;
     ///
+    /// Configure core registers for a given CPU.
+    ///
+    #[cfg(any(target_arch = "arm", target_arch = "aarch64"))]
+    fn setup_regs(&self, cpu_id: u8, boot_ip: u64, fdt_start: u64) -> Result<()>;
+    ///
     /// Retrieve the vCPU state.
     /// This function is necessary to snapshot the VM
     ///

hypervisor/src/hypervisor.rs

@@ -66,6 +66,11 @@ pub enum HypervisorError {
     ///
     #[error("Failed to retrieve TDX capabilities:{0}")]
     TdxCapabilities(#[source] anyhow::Error),
+    ///
+    /// Failed to set partition property
+    ///
+    #[error("Failed to set partition property:{0}")]
+    SetPartitionProperty(#[source] anyhow::Error),
 }
 
 ///

hypervisor/src/kvm/aarch64/gic/dist_regs.rs

@@ -1,11 +1,11 @@
 // Copyright 2020 Amazon.com, Inc. or its affiliates. All Rights Reserved.
 // SPDX-License-Identifier: Apache-2.0
 
-use super::{Error, Result};
+use crate::arch::aarch64::gic::{Error, Result};
-use crate::layout::IRQ_BASE;
+use crate::kvm::kvm_bindings::{
-use hypervisor::kvm::kvm_bindings::{
     kvm_device_attr, KVM_DEV_ARM_VGIC_GRP_DIST_REGS, KVM_DEV_ARM_VGIC_GRP_NR_IRQS,
 };
+use crate::Device;
 use std::sync::Arc;
 
 /*
@@ -77,12 +77,7 @@ static VGIC_DIST_REGS: &[DistReg] = &[
     VGIC_DIST_REG!(GICD_IPRIORITYR, 8, 0),
 ];
 
-fn dist_attr_access(
+fn dist_attr_access(gic: &Arc<dyn Device>, offset: u32, val: &u32, set: bool) -> Result<()> {
-    gic: &Arc<dyn hypervisor::Device>,
-    offset: u32,
-    val: &u32,
-    set: bool,
-) -> Result<()> {
     let mut gic_dist_attr = kvm_device_attr {
         group: KVM_DEV_ARM_VGIC_GRP_DIST_REGS,
         attr: offset as u64,
@@ -100,18 +95,18 @@ fn dist_attr_access(
 }
 
 /// Get the distributor control register.
-pub fn read_ctlr(gic: &Arc<dyn hypervisor::Device>) -> Result<u32> {
+pub fn read_ctlr(gic: &Arc<dyn Device>) -> Result<u32> {
     let val: u32 = 0;
     dist_attr_access(gic, GICD_CTLR, &val, false)?;
     Ok(val)
 }
 
 /// Set the distributor control register.
-pub fn write_ctlr(gic: &Arc<dyn hypervisor::Device>, val: u32) -> Result<()> {
+pub fn write_ctlr(gic: &Arc<dyn Device>, val: u32) -> Result<()> {
     dist_attr_access(gic, GICD_CTLR, &val, true)
 }
 
-fn get_interrupts_num(gic: &Arc<dyn hypervisor::Device>) -> Result<u32> {
+fn get_interrupts_num(gic: &Arc<dyn Device>) -> Result<u32> {
     let num_irq = 0;
 
     let mut nr_irqs_attr = kvm_device_attr {
@@ -125,7 +120,12 @@ fn get_interrupts_num(gic: &Arc<dyn hypervisor::Device>) -> Result<u32> {
     Ok(num_irq)
 }
 
-fn compute_reg_len(gic: &Arc<dyn hypervisor::Device>, reg: &DistReg, base: u32) -> Result<u32> {
+fn compute_reg_len(gic: &Arc<dyn Device>, reg: &DistReg, base: u32) -> Result<u32> {
+    // FIXME:
+    // Redefine some GIC constants to avoid the dependency on `layout` crate.
+    // This is temporary solution, will be fixed in future refactoring.
+    const LAYOUT_IRQ_BASE: u32 = 32;
+
     let mut end = base;
     let num_irq = get_interrupts_num(gic)?;
     if reg.length > 0 {
@@ -138,8 +138,8 @@ fn compute_reg_len(gic: &Arc<dyn hypervisor::Device>, reg: &DistReg, base: u32)
         // This is the type of register that takes into account the number of interrupts
         // that the model has. It is also the type of register where
         // a register relates to multiple interrupts.
-        end = base + (reg.bpi as u32 * (num_irq - IRQ_BASE) / 8);
+        end = base + (reg.bpi as u32 * (num_irq - LAYOUT_IRQ_BASE) / 8);
-        if reg.bpi as u32 * (num_irq - IRQ_BASE) % 8 > 0 {
+        if reg.bpi as u32 * (num_irq - LAYOUT_IRQ_BASE) % 8 > 0 {
             end += REG_SIZE as u32;
         }
     }
@@ -147,7 +147,7 @@ fn compute_reg_len(gic: &Arc<dyn hypervisor::Device>, reg: &DistReg, base: u32)
 }
 
 /// Set distributor registers of the GIC.
-pub fn set_dist_regs(gic: &Arc<dyn hypervisor::Device>, state: &[u32]) -> Result<()> {
+pub fn set_dist_regs(gic: &Arc<dyn Device>, state: &[u32]) -> Result<()> {
     let mut idx = 0;
 
     for dreg in VGIC_DIST_REGS {
@@ -164,7 +164,7 @@ pub fn set_dist_regs(gic: &Arc<dyn hypervisor::Device>, state: &[u32]) -> Result
     Ok(())
 }
 /// Get distributor registers of the GIC.
-pub fn get_dist_regs(gic: &Arc<dyn hypervisor::Device>) -> Result<Vec<u32>> {
+pub fn get_dist_regs(gic: &Arc<dyn Device>) -> Result<Vec<u32>> {
     let mut state = Vec::new();
 
     for dreg in VGIC_DIST_REGS {

hypervisor/src/kvm/aarch64/gic/icc_regs.rs

@@ -1,13 +1,15 @@
+// Copyright 2022 Arm Limited (or its affiliates). All rights reserved.
 // Copyright 2020 Amazon.com, Inc. or its affiliates. All Rights Reserved.
 // SPDX-License-Identifier: Apache-2.0
 
-use super::{Error, Result};
+use crate::arch::aarch64::gic::{Error, Result};
-use hypervisor::kvm::kvm_bindings::{
+use crate::kvm::kvm_bindings::{
     kvm_device_attr, KVM_DEV_ARM_VGIC_GRP_CPU_SYSREGS, KVM_REG_ARM64_SYSREG_CRM_MASK,
     KVM_REG_ARM64_SYSREG_CRM_SHIFT, KVM_REG_ARM64_SYSREG_CRN_MASK, KVM_REG_ARM64_SYSREG_CRN_SHIFT,
     KVM_REG_ARM64_SYSREG_OP0_MASK, KVM_REG_ARM64_SYSREG_OP0_SHIFT, KVM_REG_ARM64_SYSREG_OP1_MASK,
     KVM_REG_ARM64_SYSREG_OP1_SHIFT, KVM_REG_ARM64_SYSREG_OP2_MASK, KVM_REG_ARM64_SYSREG_OP2_SHIFT,
 };
+use crate::Device;
 use std::sync::Arc;
 
 const KVM_DEV_ARM_VGIC_V3_MPIDR_SHIFT: u32 = 32;
@@ -78,7 +80,7 @@ static VGIC_ICC_REGS: &[u64] = &[
 ];
 
 fn icc_attr_access(
-    gic: &Arc<dyn hypervisor::Device>,
+    gic: &Arc<dyn Device>,
     offset: u64,
     typer: u64,
     val: &u32,
@@ -101,7 +103,7 @@ fn icc_attr_access(
 }
 
 /// Get ICC registers.
-pub fn get_icc_regs(gic: &Arc<dyn hypervisor::Device>, gicr_typer: &[u64]) -> Result<Vec<u32>> {
+pub fn get_icc_regs(gic: &Arc<dyn Device>, gicr_typer: &[u64]) -> Result<Vec<u32>> {
     let mut state: Vec<u32> = Vec::new();
     // We need this for the ICC_AP<m>R<n>_EL1 registers.
     let mut num_priority_bits = 0;
@@ -154,11 +156,7 @@ pub fn get_icc_regs(gic: &Arc<dyn hypervisor::Device>, gicr_typer: &[u64]) -> Re
 }
 
 /// Set ICC registers.
-pub fn set_icc_regs(
+pub fn set_icc_regs(gic: &Arc<dyn Device>, gicr_typer: &[u64], state: &[u32]) -> Result<()> {
-    gic: &Arc<dyn hypervisor::Device>,
-    gicr_typer: &[u64],
-    state: &[u32],
-) -> Result<()> {
     let mut num_priority_bits = 0;
     let mut idx = 0;
     for ix in gicr_typer {

hypervisor/src/kvm/aarch64/gic/mod.rs

@@ -0,0 +1,615 @@
+// Copyright 2022 Arm Limited (or its affiliates). All rights reserved.
+
+mod dist_regs;
+mod icc_regs;
+mod redist_regs;
+
+use crate::arch::aarch64::gic::{Error, Result, Vgic};
+use crate::kvm::kvm_bindings;
+use crate::{CpuState, Device, Vm};
+use dist_regs::{get_dist_regs, read_ctlr, set_dist_regs, write_ctlr};
+use icc_regs::{get_icc_regs, set_icc_regs};
+use redist_regs::{construct_gicr_typers, get_redist_regs, set_redist_regs};
+use serde::{Deserialize, Serialize};
+use std::any::Any;
+use std::convert::TryInto;
+use std::sync::Arc;
+
+const GITS_CTLR: u32 = 0x0000;
+const GITS_IIDR: u32 = 0x0004;
+const GITS_CBASER: u32 = 0x0080;
+const GITS_CWRITER: u32 = 0x0088;
+const GITS_CREADR: u32 = 0x0090;
+const GITS_BASER: u32 = 0x0100;
+
+/// Access an ITS device attribute.
+///
+/// This is a helper function to get/set the ITS device attribute depending
+/// the bool parameter `set` provided.
+pub fn gicv3_its_attr_access(
+    its_device: &Arc<dyn Device>,
+    group: u32,
+    attr: u32,
+    val: &u64,
+    set: bool,
+) -> Result<()> {
+    let mut gicv3_its_attr = kvm_bindings::kvm_device_attr {
+        group,
+        attr: attr as u64,
+        addr: val as *const u64 as u64,
+        flags: 0,
+    };
+    if set {
+        its_device
+            .set_device_attr(&gicv3_its_attr)
+            .map_err(Error::SetDeviceAttribute)
+    } else {
+        its_device
+            .get_device_attr(&mut gicv3_its_attr)
+            .map_err(Error::GetDeviceAttribute)
+    }
+}
+
+/// Function that saves/restores ITS tables into guest RAM.
+///
+/// The tables get flushed to guest RAM whenever the VM gets stopped.
+pub fn gicv3_its_tables_access(its_device: &Arc<dyn Device>, save: bool) -> Result<()> {
+    let attr = if save {
+        u64::from(kvm_bindings::KVM_DEV_ARM_ITS_SAVE_TABLES)
+    } else {
+        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,
+        attr,
+        addr: 0,
+        flags: 0,
+    };
+    its_device
+        .set_device_attr(&init_gic_attr)
+        .map_err(Error::SetDeviceAttribute)
+}
+
+pub struct KvmGicV3Its {
+    /// The hypervisor agnostic device for the GicV3
+    device: Arc<dyn Device>,
+
+    /// The hypervisor agnostic device for the Its device
+    its_device: Option<Arc<dyn Device>>,
+
+    /// Vector holding values of GICR_TYPER for each vCPU
+    gicr_typers: Vec<u64>,
+
+    /// GIC distributor address
+    dist_addr: u64,
+
+    /// GIC distributor size
+    dist_size: u64,
+
+    /// GIC distributors address
+    redists_addr: u64,
+
+    /// GIC distributors size
+    redists_size: u64,
+
+    /// GIC MSI address
+    msi_addr: u64,
+
+    /// GIC MSI size
+    msi_size: u64,
+
+    /// Number of CPUs handled by the device
+    vcpu_count: u64,
+}
+
+#[derive(Clone, Default, Serialize, Deserialize)]
+pub struct Gicv3ItsState {
+    dist: Vec<u32>,
+    rdist: Vec<u32>,
+    icc: Vec<u32>,
+    // special register that enables interrupts and affinity routing
+    gicd_ctlr: u32,
+    its_ctlr: u64,
+    its_iidr: u64,
+    its_cbaser: u64,
+    its_cwriter: u64,
+    its_creadr: u64,
+    its_baser: [u64; 8],
+}
+
+impl KvmGicV3Its {
+    /// Device trees specific constants
+    pub const ARCH_GIC_V3_MAINT_IRQ: u32 = 9;
+
+    /// Returns the GIC version of the device
+    fn version() -> u32 {
+        kvm_bindings::kvm_device_type_KVM_DEV_TYPE_ARM_VGIC_V3
+    }
+
+    fn device(&self) -> &Arc<dyn Device> {
+        &self.device
+    }
+
+    fn its_device(&self) -> Option<&Arc<dyn Device>> {
+        self.its_device.as_ref()
+    }
+
+    /// Setup the device-specific attributes
+    fn init_device_attributes(&mut self, vm: &dyn Vm, nr_irqs: u32) -> Result<()> {
+        // GicV3 part attributes
+        /* Setting up the distributor attribute.
+         We are placing the GIC below 1GB so we need to substract the size of the distributor.
+        */
+        Self::set_device_attribute(
+            self.device(),
+            kvm_bindings::KVM_DEV_ARM_VGIC_GRP_ADDR,
+            u64::from(kvm_bindings::KVM_VGIC_V3_ADDR_TYPE_DIST),
+            &self.dist_addr as *const u64 as u64,
+            0,
+        )?;
+
+        /* Setting up the redistributors' attribute.
+        We are calculating here the start of the redistributors address. We have one per CPU.
+        */
+        Self::set_device_attribute(
+            self.device(),
+            kvm_bindings::KVM_DEV_ARM_VGIC_GRP_ADDR,
+            u64::from(kvm_bindings::KVM_VGIC_V3_ADDR_TYPE_REDIST),
+            &self.redists_addr as *const u64 as u64,
+            0,
+        )?;
+
+        // ITS part attributes
+        let mut its_device = kvm_bindings::kvm_create_device {
+            type_: kvm_bindings::kvm_device_type_KVM_DEV_TYPE_ARM_VGIC_ITS,
+            fd: 0,
+            flags: 0,
+        };
+
+        let its_fd = vm
+            .create_device(&mut its_device)
+            .map_err(Error::CreateGic)?;
+
+        Self::set_device_attribute(
+            &its_fd,
+            kvm_bindings::KVM_DEV_ARM_VGIC_GRP_ADDR,
+            u64::from(kvm_bindings::KVM_VGIC_ITS_ADDR_TYPE),
+            &self.msi_addr as *const u64 as u64,
+            0,
+        )?;
+
+        Self::set_device_attribute(
+            &its_fd,
+            kvm_bindings::KVM_DEV_ARM_VGIC_GRP_CTRL,
+            u64::from(kvm_bindings::KVM_DEV_ARM_VGIC_CTRL_INIT),
+            0,
+            0,
+        )?;
+
+        self.set_its_device(Some(its_fd));
+
+        /* We need to tell the kernel how many irqs to support with this vgic.
+         * See the `layout` module for details.
+         */
+        let nr_irqs_ptr = &nr_irqs as *const u32;
+        Self::set_device_attribute(
+            self.device(),
+            kvm_bindings::KVM_DEV_ARM_VGIC_GRP_NR_IRQS,
+            0,
+            nr_irqs_ptr as u64,
+            0,
+        )?;
+
+        /* Finalize the GIC.
+         * See https://code.woboq.org/linux/linux/virt/kvm/arm/vgic/vgic-kvm-device.c.html#211.
+         */
+        Self::set_device_attribute(
+            self.device(),
+            kvm_bindings::KVM_DEV_ARM_VGIC_GRP_CTRL,
+            u64::from(kvm_bindings::KVM_DEV_ARM_VGIC_CTRL_INIT),
+            0,
+            0,
+        )
+    }
+
+    /// Create a KVM Vgic device
+    fn create_device(vm: &dyn Vm) -> Result<Arc<dyn Device>> {
+        let mut gic_device = kvm_bindings::kvm_create_device {
+            type_: Self::version(),
+            fd: 0,
+            flags: 0,
+        };
+
+        vm.create_device(&mut gic_device).map_err(Error::CreateGic)
+    }
+
+    /// Set a GIC device attribute
+    fn set_device_attribute(
+        device: &Arc<dyn Device>,
+        group: u32,
+        attr: u64,
+        addr: u64,
+        flags: u32,
+    ) -> Result<()> {
+        let attr = kvm_bindings::kvm_device_attr {
+            flags,
+            group,
+            attr,
+            addr,
+        };
+        device
+            .set_device_attr(&attr)
+            .map_err(Error::SetDeviceAttribute)
+    }
+
+    /// Method to initialize the GIC device
+    #[allow(clippy::new_ret_no_self)]
+    pub fn new(
+        vm: &dyn Vm,
+        vcpu_count: u64,
+        dist_addr: u64,
+        dist_size: u64,
+        redist_size: u64,
+        msi_size: u64,
+        nr_irqs: u32,
+    ) -> Result<KvmGicV3Its> {
+        let vgic = Self::create_device(vm)?;
+        let redists_size: u64 = redist_size * vcpu_count;
+        let redists_addr: u64 = dist_addr - redists_size;
+        let msi_addr: u64 = redists_addr - msi_size;
+
+        let mut gic_device = KvmGicV3Its {
+            device: vgic,
+            its_device: None,
+            gicr_typers: vec![0; vcpu_count.try_into().unwrap()],
+            dist_addr,
+            dist_size,
+            redists_addr,
+            redists_size,
+            msi_addr,
+            msi_size,
+            vcpu_count,
+        };
+
+        gic_device.init_device_attributes(vm, nr_irqs)?;
+
+        Ok(gic_device)
+    }
+}
+
+impl Vgic for KvmGicV3Its {
+    fn fdt_compatibility(&self) -> &str {
+        "arm,gic-v3"
+    }
+
+    fn msi_compatible(&self) -> bool {
+        true
+    }
+
+    fn msi_compatibility(&self) -> &str {
+        "arm,gic-v3-its"
+    }
+
+    fn fdt_maint_irq(&self) -> u32 {
+        KvmGicV3Its::ARCH_GIC_V3_MAINT_IRQ
+    }
+
+    fn vcpu_count(&self) -> u64 {
+        self.vcpu_count
+    }
+
+    fn device_properties(&self) -> [u64; 4] {
+        [
+            self.dist_addr,
+            self.dist_size,
+            self.redists_addr,
+            self.redists_size,
+        ]
+    }
+
+    fn msi_properties(&self) -> [u64; 2] {
+        [self.msi_addr, self.msi_size]
+    }
+
+    fn set_its_device(&mut self, its_device: Option<Arc<dyn Device>>) {
+        self.its_device = its_device;
+    }
+
+    fn set_gicr_typers(&mut self, vcpu_states: &[CpuState]) {
+        let gicr_typers = construct_gicr_typers(vcpu_states);
+        self.gicr_typers = gicr_typers;
+    }
+
+    fn as_any_concrete_mut(&mut self) -> &mut dyn Any {
+        self
+    }
+
+    /// Save the state of GICv3ITS.
+    fn state(&self) -> Result<Gicv3ItsState> {
+        let gicr_typers = self.gicr_typers.clone();
+
+        let gicd_ctlr = read_ctlr(self.device())?;
+
+        let dist_state = get_dist_regs(self.device())?;
+
+        let rdist_state = get_redist_regs(self.device(), &gicr_typers)?;
+
+        let icc_state = get_icc_regs(self.device(), &gicr_typers)?;
+
+        let its_baser_state: [u64; 8] = [0; 8];
+        for i in 0..8 {
+            gicv3_its_attr_access(
+                self.its_device().unwrap(),
+                kvm_bindings::KVM_DEV_ARM_VGIC_GRP_ITS_REGS,
+                GITS_BASER + i * 8,
+                &its_baser_state[i as usize],
+                false,
+            )?;
+        }
+
+        let its_ctlr_state: u64 = 0;
+        gicv3_its_attr_access(
+            self.its_device().unwrap(),
+            kvm_bindings::KVM_DEV_ARM_VGIC_GRP_ITS_REGS,
+            GITS_CTLR,
+            &its_ctlr_state,
+            false,
+        )?;
+
+        let its_cbaser_state: u64 = 0;
+        gicv3_its_attr_access(
+            self.its_device().unwrap(),
+            kvm_bindings::KVM_DEV_ARM_VGIC_GRP_ITS_REGS,
+            GITS_CBASER,
+            &its_cbaser_state,
+            false,
+        )?;
+
+        let its_creadr_state: u64 = 0;
+        gicv3_its_attr_access(
+            self.its_device().unwrap(),
+            kvm_bindings::KVM_DEV_ARM_VGIC_GRP_ITS_REGS,
+            GITS_CREADR,
+            &its_creadr_state,
+            false,
+        )?;
+
+        let its_cwriter_state: u64 = 0;
+        gicv3_its_attr_access(
+            self.its_device().unwrap(),
+            kvm_bindings::KVM_DEV_ARM_VGIC_GRP_ITS_REGS,
+            GITS_CWRITER,
+            &its_cwriter_state,
+            false,
+        )?;
+
+        let its_iidr_state: u64 = 0;
+        gicv3_its_attr_access(
+            self.its_device().unwrap(),
+            kvm_bindings::KVM_DEV_ARM_VGIC_GRP_ITS_REGS,
+            GITS_IIDR,
+            &its_iidr_state,
+            false,
+        )?;
+
+        Ok(Gicv3ItsState {
+            dist: dist_state,
+            rdist: rdist_state,
+            icc: icc_state,
+            gicd_ctlr,
+            its_ctlr: its_ctlr_state,
+            its_iidr: its_iidr_state,
+            its_cbaser: its_cbaser_state,
+            its_cwriter: its_cwriter_state,
+            its_creadr: its_creadr_state,
+            its_baser: its_baser_state,
+        })
+    }
+
+    /// Restore the state of GICv3ITS.
+    fn set_state(&mut self, state: &Gicv3ItsState) -> Result<()> {
+        let gicr_typers = self.gicr_typers.clone();
+
+        write_ctlr(self.device(), state.gicd_ctlr)?;
+
+        set_dist_regs(self.device(), &state.dist)?;
+
+        set_redist_regs(self.device(), &gicr_typers, &state.rdist)?;
+
+        set_icc_regs(self.device(), &gicr_typers, &state.icc)?;
+
+        //Restore GICv3ITS registers
+        gicv3_its_attr_access(
+            self.its_device().unwrap(),
+            kvm_bindings::KVM_DEV_ARM_VGIC_GRP_ITS_REGS,
+            GITS_IIDR,
+            &state.its_iidr,
+            true,
+        )?;
+
+        gicv3_its_attr_access(
+            self.its_device().unwrap(),
+            kvm_bindings::KVM_DEV_ARM_VGIC_GRP_ITS_REGS,
+            GITS_CBASER,
+            &state.its_cbaser,
+            true,
+        )?;
+
+        gicv3_its_attr_access(
+            self.its_device().unwrap(),
+            kvm_bindings::KVM_DEV_ARM_VGIC_GRP_ITS_REGS,
+            GITS_CREADR,
+            &state.its_creadr,
+            true,
+        )?;
+
+        gicv3_its_attr_access(
+            self.its_device().unwrap(),
+            kvm_bindings::KVM_DEV_ARM_VGIC_GRP_ITS_REGS,
+            GITS_CWRITER,
+            &state.its_cwriter,
+            true,
+        )?;
+
+        for i in 0..8 {
+            gicv3_its_attr_access(
+                self.its_device().unwrap(),
+                kvm_bindings::KVM_DEV_ARM_VGIC_GRP_ITS_REGS,
+                GITS_BASER + i * 8,
+                &state.its_baser[i as usize],
+                true,
+            )?;
+        }
+
+        // Restore ITS tables
+        gicv3_its_tables_access(self.its_device().unwrap(), false)?;
+
+        gicv3_its_attr_access(
+            self.its_device().unwrap(),
+            kvm_bindings::KVM_DEV_ARM_VGIC_GRP_ITS_REGS,
+            GITS_CTLR,
+            &state.its_ctlr,
+            true,
+        )
+    }
+
+    /// Saves GIC internal data tables into RAM, including:
+    /// - RDIST pending tables
+    /// - ITS tables into guest RAM.
+    fn save_data_tables(&self) -> Result<()> {
+        // Flash RDIST pending tables
+        let init_gic_attr = kvm_bindings::kvm_device_attr {
+            group: kvm_bindings::KVM_DEV_ARM_VGIC_GRP_CTRL,
+            attr: u64::from(kvm_bindings::KVM_DEV_ARM_VGIC_SAVE_PENDING_TABLES),
+            addr: 0,
+            flags: 0,
+        };
+        self.device()
+            .set_device_attr(&init_gic_attr)
+            .map_err(Error::SetDeviceAttribute)?;
+        // Flush ITS tables to guest RAM.
+        gicv3_its_tables_access(self.its_device().unwrap(), true)
+    }
+}
+
+#[cfg(test)]
+mod tests {
+    use crate::aarch64::gic::{
+        get_dist_regs, get_icc_regs, get_redist_regs, set_dist_regs, set_icc_regs, set_redist_regs,
+    };
+    use crate::kvm::KvmGicV3Its;
+
+    #[test]
+    fn test_create_gic() {
+        let hv = crate::new().unwrap();
+        let vm = hv.create_vm().unwrap();
+
+        assert!(KvmGicV3Its::new(
+            &*vm,
+            1,
+            0x0900_0000 - 0x01_0000,
+            0x01_0000,
+            0x02_0000,
+            0x02_0000,
+            256
+        )
+        .is_ok());
+    }
+
+    #[test]
+    fn test_get_set_dist_regs() {
+        let hv = crate::new().unwrap();
+        let vm = hv.create_vm().unwrap();
+        let _ = vm.create_vcpu(0, None).unwrap();
+        let gic = KvmGicV3Its::new(
+            &*vm,
+            1,
+            0x0900_0000 - 0x01_0000,
+            0x01_0000,
+            0x02_0000,
+            0x02_0000,
+            256,
+        )
+        .expect("Cannot create gic");
+
+        let res = get_dist_regs(gic.device());
+        assert!(res.is_ok());
+        let state = res.unwrap();
+        assert_eq!(state.len(), 568);
+
+        let res = set_dist_regs(gic.device(), &state);
+        assert!(res.is_ok());
+    }
+
+    #[test]
+    fn test_get_set_redist_regs() {
+        let hv = crate::new().unwrap();
+        let vm = hv.create_vm().unwrap();
+        let _ = vm.create_vcpu(0, None).unwrap();
+        let gic = KvmGicV3Its::new(
+            &*vm,
+            1,
+            0x0900_0000 - 0x01_0000,
+            0x01_0000,
+            0x02_0000,
+            0x02_0000,
+            256,
+        )
+        .expect("Cannot create gic");
+
+        let gicr_typer = vec![123];
+        let res = get_redist_regs(gic.device(), &gicr_typer);
+        assert!(res.is_ok());
+        let state = res.unwrap();
+        println!("{}", state.len());
+        assert!(state.len() == 24);
+
+        assert!(set_redist_regs(gic.device(), &gicr_typer, &state).is_ok());
+    }
+
+    #[test]
+    fn test_get_set_icc_regs() {
+        let hv = crate::new().unwrap();
+        let vm = hv.create_vm().unwrap();
+        let _ = vm.create_vcpu(0, None).unwrap();
+        let gic = KvmGicV3Its::new(
+            &*vm,
+            1,
+            0x0900_0000 - 0x01_0000,
+            0x01_0000,
+            0x02_0000,
+            0x02_0000,
+            256,
+        )
+        .expect("Cannot create gic");
+
+        let gicr_typer = vec![123];
+        let res = get_icc_regs(gic.device(), &gicr_typer);
+        assert!(res.is_ok());
+        let state = res.unwrap();
+        println!("{}", state.len());
+        assert!(state.len() == 9);
+
+        assert!(set_icc_regs(gic.device(), &gicr_typer, &state).is_ok());
+    }
+
+    #[test]
+    fn test_save_data_tables() {
+        let hv = crate::new().unwrap();
+        let vm = hv.create_vm().unwrap();
+        let _ = vm.create_vcpu(0, None).unwrap();
+        let gic = vm
+            .create_vgic(
+                1,
+                0x0900_0000 - 0x01_0000,
+                0x01_0000,
+                0x02_0000,
+                0x02_0000,
+                256,
+            )
+            .expect("Cannot create gic");
+
+        assert!(gic.lock().unwrap().save_data_tables().is_ok());
+    }
+}

hypervisor/src/kvm/aarch64/gic/redist_regs.rs

@@ -1,9 +1,10 @@
+// Copyright 2022 Arm Limited (or its affiliates). All rights reserved.
 // Copyright 2020 Amazon.com, Inc. or its affiliates. All Rights Reserved.
 // SPDX-License-Identifier: Apache-2.0
 
-use super::{Error, Result};
+use crate::arch::aarch64::gic::{Error, Result};
-use hypervisor::kvm::kvm_bindings::{kvm_device_attr, KVM_DEV_ARM_VGIC_GRP_REDIST_REGS};
+use crate::kvm::kvm_bindings::{kvm_device_attr, KVM_DEV_ARM_VGIC_GRP_REDIST_REGS};
-use hypervisor::CpuState;
+use crate::{CpuState, Device};
 use std::sync::Arc;
 
 // Relevant redistributor registers that we want to save/restore.
@@ -83,7 +84,7 @@ static VGIC_SGI_REGS: &[RdistReg] = &[
 ];
 
 fn redist_attr_access(
-    gic: &Arc<dyn hypervisor::Device>,
+    gic: &Arc<dyn Device>,
     offset: u32,
     typer: u64,
     val: &u32,
@@ -106,7 +107,7 @@ fn redist_attr_access(
 }
 
 fn access_redists_aux(
-    gic: &Arc<dyn hypervisor::Device>,
+    gic: &Arc<dyn Device>,
     gicr_typer: &[u64],
     state: &mut Vec<u32>,
     reg_list: &[RdistReg],
@@ -136,7 +137,7 @@ fn access_redists_aux(
 }
 
 /// Get redistributor registers.
-pub fn get_redist_regs(gic: &Arc<dyn hypervisor::Device>, gicr_typer: &[u64]) -> Result<Vec<u32>> {
+pub fn get_redist_regs(gic: &Arc<dyn Device>, gicr_typer: &[u64]) -> Result<Vec<u32>> {
     let mut state = Vec::new();
     let mut idx: usize = 0;
     access_redists_aux(
@@ -153,11 +154,7 @@ pub fn get_redist_regs(gic: &Arc<dyn hypervisor::Device>, gicr_typer: &[u64]) ->
 }
 
 /// Set redistributor registers.
-pub fn set_redist_regs(
+pub fn set_redist_regs(gic: &Arc<dyn Device>, gicr_typer: &[u64], state: &[u32]) -> Result<()> {
-    gic: &Arc<dyn hypervisor::Device>,
-    gicr_typer: &[u64],
-    state: &[u32],
-) -> Result<()> {
     let mut idx: usize = 0;
     let mut mut_state = state.to_owned();
     access_redists_aux(

hypervisor/src/kvm/aarch64/mod.rs

@@ -8,9 +8,8 @@
 //
 //
 
-///
+pub mod gic;
-/// Export generically-named wrappers of kvm-bindings for Unix-based platforms
+
-///
 use crate::kvm::{KvmError, KvmResult};
 use kvm_bindings::{
     kvm_mp_state, kvm_one_reg, kvm_regs, KVM_REG_ARM64, KVM_REG_ARM64_SYSREG,

hypervisor/src/kvm/mod.rs

@@ -8,11 +8,15 @@
 //
 //
 
+#[cfg(target_arch = "aarch64")]
+use crate::aarch64::gic::KvmGicV3Its;
 #[cfg(target_arch = "aarch64")]
 pub use crate::aarch64::{
-    check_required_kvm_extensions, is_system_register, VcpuInit, VcpuKvmState as CpuState,
+    check_required_kvm_extensions, gic::Gicv3ItsState as GicState, is_system_register, VcpuInit,
-    MPIDR_EL1,
+    VcpuKvmState as CpuState, MPIDR_EL1,
 };
+#[cfg(target_arch = "aarch64")]
+use crate::arch::aarch64::gic::Vgic;
 use crate::cpu;
 use crate::device;
 use crate::hypervisor;
@@ -31,6 +35,8 @@ use std::os::unix::io::{AsRawFd, RawFd};
 use std::result;
 #[cfg(target_arch = "x86_64")]
 use std::sync::atomic::{AtomicBool, Ordering};
+#[cfg(target_arch = "aarch64")]
+use std::sync::Mutex;
 use std::sync::{Arc, RwLock};
 use vmm_sys_util::eventfd::EventFd;
 // x86_64 dependencies
@@ -154,7 +160,7 @@ pub struct TdxCapabilities {
     pub cpuid_configs: [TdxCpuidConfig; TDX_MAX_NR_CPUID_CONFIGS],
 }
 
-#[derive(Clone, Copy, Debug, PartialEq, Deserialize, Serialize)]
+#[derive(Clone, Copy, Debug, PartialEq, Eq, Deserialize, Serialize)]
 pub struct KvmVmState {}
 
 pub use KvmVmState as VmState;
@@ -250,6 +256,31 @@ impl vm::Vm for KvmVm {
         };
         Ok(Arc::new(vcpu))
     }
+    #[cfg(target_arch = "aarch64")]
+    ///
+    /// Creates a virtual GIC device.
+    ///
+    fn create_vgic(
+        &self,
+        vcpu_count: u64,
+        dist_addr: u64,
+        dist_size: u64,
+        redist_size: u64,
+        msi_size: u64,
+        nr_irqs: u32,
+    ) -> vm::Result<Arc<Mutex<dyn Vgic>>> {
+        let gic_device = KvmGicV3Its::new(
+            self,
+            vcpu_count,
+            dist_addr,
+            dist_size,
+            redist_size,
+            msi_size,
+            nr_irqs,
+        )
+        .map_err(|e| vm::HypervisorVmError::CreateVgic(anyhow!("Vgic error {:?}", e)))?;
+        Ok(Arc::new(Mutex::new(gic_device)))
+    }
     ///
     /// Registers an event to be signaled whenever a certain address is written to.
     ///
@@ -1535,6 +1566,51 @@ impl cpu::Vcpu for KvmVcpu {
             .get_one_reg(MPIDR_EL1)
             .map_err(|e| cpu::HypervisorCpuError::GetSysRegister(e.into()))
     }
+    ///
+    /// Configure core registers for a given CPU.
+    ///
+    #[cfg(any(target_arch = "arm", target_arch = "aarch64"))]
+    fn setup_regs(&self, cpu_id: u8, boot_ip: u64, fdt_start: u64) -> cpu::Result<()> {
+        #[allow(non_upper_case_globals)]
+        // PSR (Processor State Register) bits.
+        // Taken from arch/arm64/include/uapi/asm/ptrace.h.
+        const PSR_MODE_EL1h: u64 = 0x0000_0005;
+        const PSR_F_BIT: u64 = 0x0000_0040;
+        const PSR_I_BIT: u64 = 0x0000_0080;
+        const PSR_A_BIT: u64 = 0x0000_0100;
+        const PSR_D_BIT: u64 = 0x0000_0200;
+        // Taken from arch/arm64/kvm/inject_fault.c.
+        const PSTATE_FAULT_BITS_64: u64 =
+            PSR_MODE_EL1h | PSR_A_BIT | PSR_F_BIT | PSR_I_BIT | PSR_D_BIT;
+
+        let kreg_off = offset__of!(kvm_regs, regs);
+
+        // Get the register index of the PSTATE (Processor State) register.
+        let pstate = offset__of!(user_pt_regs, pstate) + kreg_off;
+        self.set_reg(
+            arm64_core_reg_id!(KVM_REG_SIZE_U64, pstate),
+            PSTATE_FAULT_BITS_64,
+        )
+        .map_err(|e| cpu::HypervisorCpuError::SetCoreRegister(e.into()))?;
+
+        // Other vCPUs are powered off initially awaiting PSCI wakeup.
+        if cpu_id == 0 {
+            // Setting the PC (Processor Counter) to the current program address (kernel address).
+            let pc = offset__of!(user_pt_regs, pc) + kreg_off;
+            self.set_reg(arm64_core_reg_id!(KVM_REG_SIZE_U64, pc), boot_ip as u64)
+                .map_err(|e| cpu::HypervisorCpuError::SetCoreRegister(e.into()))?;
+
+            // Last mandatory thing to set -> the address pointing to the FDT (also called DTB).
+            // "The device tree blob (dtb) must be placed on an 8-byte boundary and must
+            // not exceed 2 megabytes in size." -> https://www.kernel.org/doc/Documentation/arm64/booting.txt.
+            // We are choosing to place it the end of DRAM. See `get_fdt_addr`.
+            let regs0 = offset__of!(user_pt_regs, regs) + kreg_off;
+            self.set_reg(arm64_core_reg_id!(KVM_REG_SIZE_U64, regs0), fdt_start)
+                .map_err(|e| cpu::HypervisorCpuError::SetCoreRegister(e.into()))?;
+        }
+        Ok(())
+    }
+
     #[cfg(target_arch = "x86_64")]
     ///
     /// Get the current CPU state

hypervisor/src/lib.rs

@@ -18,6 +18,8 @@
 //! - arm64
 //!
 
+#![allow(clippy::significant_drop_in_scrutinee)]
+
 #[macro_use]
 extern crate anyhow;
 #[cfg(target_arch = "x86_64")]
@@ -51,10 +53,10 @@ mod device;
 pub use cpu::{HypervisorCpuError, Vcpu, VmExit};
 pub use device::{Device, HypervisorDeviceError};
 pub use hypervisor::{Hypervisor, HypervisorError};
-#[cfg(all(feature = "kvm", target_arch = "aarch64"))]
-pub use kvm::aarch64;
 #[cfg(all(feature = "kvm", target_arch = "x86_64"))]
 pub use kvm::x86_64;
+#[cfg(all(feature = "kvm", target_arch = "aarch64"))]
+pub use kvm::{aarch64, GicState};
 // Aliased types exposed from both hypervisors
 #[cfg(feature = "kvm")]
 pub use kvm::{

hypervisor/src/mshv/mod.rs

@@ -108,6 +108,17 @@ impl hypervisor::Hypervisor for MshvHypervisor {
             break;
         }
 
+        // Default Microsoft Hypervisor behavior for unimplemented MSR is to
+        // send a fault to the guest if it tries to access it. It is possible
+        // to override this behavior with a more suitable option i.e., ignore
+        // writes from the guest and return zero in attempt to read unimplemented
+        // MSR.
+        fd.set_partition_property(
+            hv_partition_property_code_HV_PARTITION_PROPERTY_UNIMPLEMENTED_MSR_ACTION,
+            hv_unimplemented_msr_action_HV_UNIMPLEMENTED_MSR_ACTION_IGNORE_WRITE_READ_ZERO as u64,
+        )
+        .map_err(|e| hypervisor::HypervisorError::SetPartitionProperty(e.into()))?;
+
         let msr_list = self.get_msr_list()?;
         let num_msrs = msr_list.as_fam_struct_ref().nmsrs as usize;
         let mut msrs = MsrEntries::new(num_msrs).unwrap();

hypervisor/src/vm.rs

@@ -10,6 +10,8 @@
 
 #[cfg(target_arch = "aarch64")]
 use crate::aarch64::VcpuInit;
+#[cfg(target_arch = "aarch64")]
+use crate::arch::aarch64::gic::Vgic;
 use crate::cpu::Vcpu;
 use crate::device::Device;
 #[cfg(feature = "kvm")]
@@ -27,6 +29,8 @@ use kvm_ioctls::Cap;
 #[cfg(target_arch = "x86_64")]
 use std::fs::File;
 use std::sync::Arc;
+#[cfg(target_arch = "aarch64")]
+use std::sync::Mutex;
 use thiserror::Error;
 use vmm_sys_util::eventfd::EventFd;
 
@@ -210,6 +214,11 @@ pub enum HypervisorVmError {
     ///
     #[error("Failed to initialize memory region TDX: {0}")]
     InitMemRegionTdx(#[source] std::io::Error),
+    ///
+    /// Create Vgic error
+    ///
+    #[error("Failed to create Vgic: {0}")]
+    CreateVgic(#[source] anyhow::Error),
 }
 ///
 /// Result type for returning from a function
@@ -269,6 +278,17 @@ pub trait Vm: Send + Sync {
     fn unregister_irqfd(&self, fd: &EventFd, gsi: u32) -> Result<()>;
     /// Creates a new KVM vCPU file descriptor and maps the memory corresponding
     fn create_vcpu(&self, id: u8, vm_ops: Option<Arc<dyn VmOps>>) -> Result<Arc<dyn Vcpu>>;
+    #[cfg(target_arch = "aarch64")]
+    fn create_vgic(
+        &self,
+        vcpu_count: u64,
+        dist_addr: u64,
+        dist_size: u64,
+        redist_size: u64,
+        msi_size: u64,
+        nr_irqs: u32,
+    ) -> Result<Arc<Mutex<dyn Vgic>>>;
+
     /// Registers an event to be signaled whenever a certain address is written to.
     fn register_ioevent(
         &self,

net_util/Cargo.toml

@@ -11,16 +11,17 @@ libc = "0.2.126"
 log = "0.4.17"
 net_gen = { path = "../net_gen" }
 rate_limiter = { path = "../rate_limiter" }
-serde = "1.0.137"
+serde = "1.0.138"
 versionize = "0.1.6"
 versionize_derive = "0.1.4"
 virtio-bindings = "0.1.0"
-virtio-queue = "0.3.0"
+virtio-queue = "0.4.0"
 vm-memory = { version = "0.8.0", 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"
+once_cell = "1.13.0"
-pnet = "0.30.0"
+pnet = "0.31.0"
-serde_json = "1.0.81"
+pnet_datalink = "0.31.0"
+serde_json = "1.0.82"

net_util/src/lib.rs

@@ -5,11 +5,6 @@
 // Use of this source code is governed by a BSD-style license that can be
 // found in the THIRD-PARTY file.
 
-// This is only used by the tests module from tap.rs, but we cannot use #[macro_use] unless the
-// reference to lazy_static is declared at the root level of the importing crate.
-#[cfg(test)]
-#[macro_use]
-extern crate lazy_static;
 #[macro_use]
 extern crate log;
 

net_util/src/mac.rs

@@ -15,7 +15,7 @@ use serde::ser::{Serialize, Serializer};
 
 pub const MAC_ADDR_LEN: usize = 6;
 
-#[derive(Clone, Copy, Debug, PartialEq)]
+#[derive(Clone, Copy, Debug, PartialEq, Eq)]
 pub struct MacAddr {
     bytes: [u8; MAC_ADDR_LEN],
 }

net_util/src/tap.rs

@@ -321,15 +321,13 @@ impl Tap {
 
         // If TAP device is already up don't try and enable it
         let ifru_flags = unsafe { ifreq.ifr_ifru.ifru_flags };
-        if ifru_flags
+        if ifru_flags & net_gen::net_device_flags_IFF_UP as i16
-            & (net_gen::net_device_flags_IFF_UP | net_gen::net_device_flags_IFF_RUNNING) as i16
+            == net_gen::net_device_flags_IFF_UP as i16
-            == (net_gen::net_device_flags_IFF_UP | net_gen::net_device_flags_IFF_RUNNING) as i16
         {
             return Ok(());
         }
 
-        ifreq.ifr_ifru.ifru_flags =
+        ifreq.ifr_ifru.ifru_flags = net_gen::net_device_flags_IFF_UP as i16;
-            (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.
         let ret =
@@ -395,38 +393,35 @@ impl AsRawFd for Tap {
 
 #[cfg(test)]
 mod tests {
-    extern crate pnet;
-
     use std::net::Ipv4Addr;
     use std::str;
     use std::sync::{mpsc, Mutex};
     use std::thread;
     use std::time::Duration;
 
-    use self::pnet::datalink::Channel::Ethernet;
+    use once_cell::sync::Lazy;
-    use self::pnet::datalink::{self, DataLinkReceiver, DataLinkSender, NetworkInterface};
+
-    use self::pnet::packet::ethernet::{EtherTypes, EthernetPacket, MutableEthernetPacket};
+    use pnet::packet::ethernet::{EtherTypes, EthernetPacket, MutableEthernetPacket};
-    use self::pnet::packet::ip::IpNextHeaderProtocols;
+    use pnet::packet::ip::IpNextHeaderProtocols;
-    use self::pnet::packet::ipv4::{Ipv4Packet, MutableIpv4Packet};
+    use pnet::packet::ipv4::{Ipv4Packet, MutableIpv4Packet};
-    use self::pnet::packet::udp::{MutableUdpPacket, UdpPacket};
+    use pnet::packet::udp::{MutableUdpPacket, UdpPacket};
-    use self::pnet::packet::{MutablePacket, Packet};
+    use pnet::packet::{MutablePacket, Packet};
-    use self::pnet::util::MacAddr;
+    use pnet::util::MacAddr;
+    use pnet_datalink::Channel::Ethernet;
+    use pnet_datalink::{DataLinkReceiver, DataLinkSender, NetworkInterface};
 
     use super::*;
 
     static DATA_STRING: &str = "test for tap";
     static SUBNET_MASK: &str = "255.255.255.0";
 
-    // We needed to have a mutex as a global variable, so we used the crate that provides the
+    // We needed to have a mutex as a global variable, so we used once_cell for testing. The main
-    // lazy_static! macro for testing. The main potential problem, caused by tests being run in
+    // potential problem, caused by tests being run in parallel by cargo, is creating different
-    // parallel by cargo, is creating different TAPs and trying to associate the same address,
+    // TAPs and trying to associate the same address, so we hide the IP address &str behind this
-    // so we hide the IP address &str behind this mutex, more as a convention to remember to lock
+    // mutex, more as a convention to remember to lock it at the very beginning of each function
-    // it at the very beginning of each function susceptible to this issue. Another variant is
+    // susceptible to this issue. Another variant is to use a different IP address per function,
-    // to use a different IP address per function, but we must remember to pick an unique one
+    // but we must remember to pick an unique one each time.
-    // each time.
+    static TAP_IP_LOCK: Lazy<Mutex<&'static str>> = Lazy::new(|| Mutex::new("192.168.241.1"));
-    lazy_static! {
-        static ref TAP_IP_LOCK: Mutex<&'static str> = Mutex::new("192.168.241.1");
-    }
 
     // Describes the outcomes we are currently interested in when parsing a packet (we use
     // an UDP packet for testing).
@@ -543,10 +538,10 @@ mod tests {
         let interface_name_matches = |iface: &NetworkInterface| iface.name == ifname;
 
         // Find the network interface with the provided name.
-        let interfaces = datalink::interfaces();
+        let interfaces = pnet_datalink::interfaces();
         let interface = interfaces.into_iter().find(interface_name_matches).unwrap();
 
-        if let Ok(Ethernet(tx, rx)) = datalink::channel(&interface, Default::default()) {
+        if let Ok(Ethernet(tx, rx)) = pnet_datalink::channel(&interface, Default::default()) {
             (interface.mac.unwrap(), tx, rx)
         } else {
             panic!("datalink channel error or unhandled channel type");

pci/Cargo.toml

@@ -10,7 +10,7 @@ kvm = ["vfio-ioctls/kvm"]
 mshv = ["vfio-ioctls/mshv"]
 
 [dependencies]
-anyhow = "1.0.57"
+anyhow = "1.0.58"
 byteorder = "1.4.3"
 hypervisor = { path = "../hypervisor" }
 vfio-ioctls = { git = "https://github.com/rust-vmm/vfio", branch = "main", default-features = false }
@@ -18,7 +18,7 @@ vfio_user = { path = "../vfio_user" }
 vmm-sys-util = "0.9.0"
 libc = "0.2.126"
 log = "0.4.17"
-serde = { version="1.0.137", features=["derive"] }
+serde = { version="1.0.138", features=["derive"] }
 thiserror = "1.0.31"
 versionize = "0.1.6"
 versionize_derive = "0.1.4"

pci/src/configuration.rs

@@ -187,7 +187,7 @@ pub trait PciProgrammingInterface {
 }
 
 /// Types of PCI capabilities.
-#[derive(PartialEq, Copy, Clone)]
+#[derive(PartialEq, Eq, Copy, Clone)]
 #[allow(dead_code)]
 #[allow(non_camel_case_types)]
 #[repr(C)]
@@ -321,7 +321,7 @@ pub struct PciConfiguration {
 }
 
 /// See pci_regs.h in kernel
-#[derive(Copy, Clone, PartialEq, Versionize, Debug)]
+#[derive(Copy, Clone, PartialEq, Eq, Versionize, Debug)]
 pub enum PciBarRegionType {
     Memory32BitRegion = 0,
     IoRegion = 0x01,

pci/src/lib.rs

@@ -52,7 +52,7 @@ 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)]
+#[derive(Clone, Copy, PartialEq, Eq, PartialOrd)]
 pub struct PciBdf(u32);
 
 struct PciBdfVisitor;

pci/src/msi.rs

@@ -179,7 +179,7 @@ struct MsiConfigState {
 impl VersionMapped for MsiConfigState {}
 
 pub struct MsiConfig {
-    cap: MsiCap,
+    pub cap: MsiCap,
     interrupt_source_group: Arc<dyn InterruptSourceGroup>,
 }
 

pci/src/msix.rs

@@ -441,7 +441,7 @@ impl Snapshottable for MsixConfig {
 
 #[allow(dead_code)]
 #[repr(packed)]
-#[derive(Clone, Copy, Default)]
+#[derive(Clone, Copy, Default, Versionize)]
 pub struct MsixCap {
     // Message Control Register
     //   10-0:  MSI-X Table size

pci/src/vfio.rs

@@ -4,10 +4,11 @@
 //
 
 use crate::{
-    msi_num_enabled_vectors, BarReprogrammingParams, MsiConfig, MsixCap, MsixConfig,
+    msi_num_enabled_vectors, BarReprogrammingParams, MsiCap, MsiConfig, MsixCap, MsixConfig,
     PciBarConfiguration, PciBarRegionType, PciBdf, PciCapabilityId, PciClassCode, PciConfiguration,
     PciDevice, PciDeviceError, PciHeaderType, PciSubclass, MSIX_TABLE_ENTRY_SIZE,
 };
+use anyhow::anyhow;
 use byteorder::{ByteOrder, LittleEndian};
 use hypervisor::HypervisorVmError;
 use std::any::Any;
@@ -17,6 +18,8 @@ use std::os::unix::io::AsRawFd;
 use std::ptr::null_mut;
 use std::sync::{Arc, Barrier, Mutex};
 use thiserror::Error;
+use versionize::{VersionMap, Versionize, VersionizeResult};
+use versionize_derive::Versionize;
 use vfio_bindings::bindings::vfio::*;
 use vfio_ioctls::{
     VfioContainer, VfioDevice, VfioIrq, VfioRegionInfoCap, VfioRegionSparseMmapArea,
@@ -27,6 +30,9 @@ use vm_device::interrupt::{
 };
 use vm_device::{BusDevice, Resource};
 use vm_memory::{Address, GuestAddress, GuestUsize};
+use vm_migration::{
+    Migratable, MigratableError, Pausable, Snapshot, Snapshottable, Transportable, VersionMapped,
+};
 use vmm_sys_util::eventfd::EventFd;
 
 #[derive(Debug, Error)]
@@ -71,11 +77,22 @@ enum InterruptUpdateAction {
     DisableMsix,
 }
 
+#[derive(Versionize)]
+struct IntxState {
+    enabled: bool,
+}
+
 pub(crate) struct VfioIntx {
     interrupt_source_group: Arc<dyn InterruptSourceGroup>,
     enabled: bool,
 }
 
+#[derive(Versionize)]
+struct MsiState {
+    cap: MsiCap,
+    cap_offset: u32,
+}
+
 pub(crate) struct VfioMsi {
     pub(crate) cfg: MsiConfig,
     cap_offset: u32,
@@ -102,6 +119,13 @@ impl VfioMsi {
     }
 }
 
+#[derive(Versionize)]
+struct MsixState {
+    cap: MsixCap,
+    cap_offset: u32,
+    bdf: u32,
+}
+
 pub(crate) struct VfioMsix {
     pub(crate) bar: MsixConfig,
     cap: MsixCap,
@@ -356,6 +380,15 @@ impl Vfio for VfioDeviceWrapper {
     }
 }
 
+#[derive(Versionize)]
+struct VfioCommonState {
+    intx_state: Option<IntxState>,
+    msi_state: Option<MsiState>,
+    msix_state: Option<MsixState>,
+}
+
+impl VersionMapped for VfioCommonState {}
+
 pub(crate) struct VfioCommon {
     pub(crate) configuration: PciConfiguration,
     pub(crate) mmio_regions: Vec<MmioRegion>,
@@ -970,6 +1003,127 @@ impl VfioCommon {
         // The config register read comes from the VFIO device itself.
         self.vfio_wrapper.read_config_dword((reg_idx * 4) as u32) & mask
     }
+
+    fn state(&self) -> VfioCommonState {
+        let intx_state = self.interrupt.intx.as_ref().map(|intx| IntxState {
+            enabled: intx.enabled,
+        });
+
+        let msi_state = self.interrupt.msi.as_ref().map(|msi| MsiState {
+            cap: msi.cfg.cap,
+            cap_offset: msi.cap_offset,
+        });
+
+        let msix_state = self.interrupt.msix.as_ref().map(|msix| MsixState {
+            cap: msix.cap,
+            cap_offset: msix.cap_offset,
+            bdf: msix.bar.devid,
+        });
+
+        VfioCommonState {
+            intx_state,
+            msi_state,
+            msix_state,
+        }
+    }
+
+    fn set_state(&mut self, state: &VfioCommonState) -> Result<(), VfioPciError> {
+        if let (Some(intx), Some(interrupt_source_group)) =
+            (&state.intx_state, self.legacy_interrupt_group.clone())
+        {
+            self.interrupt.intx = Some(VfioIntx {
+                interrupt_source_group,
+                enabled: false,
+            });
+
+            if intx.enabled {
+                self.enable_intx()?;
+            }
+        }
+
+        if let Some(msi) = &state.msi_state {
+            self.initialize_msi(msi.cap.msg_ctl, msi.cap_offset);
+        }
+
+        if let Some(msix) = &state.msix_state {
+            self.initialize_msix(msix.cap, msix.cap_offset, msix.bdf.into());
+        }
+
+        Ok(())
+    }
+}
+
+impl Pausable for VfioCommon {}
+
+impl Snapshottable for VfioCommon {
+    fn id(&self) -> String {
+        String::from("vfio_common")
+    }
+
+    fn snapshot(&mut self) -> std::result::Result<Snapshot, MigratableError> {
+        let mut vfio_common_snapshot =
+            Snapshot::new_from_versioned_state(&self.id(), &self.state())?;
+
+        // Snapshot PciConfiguration
+        vfio_common_snapshot.add_snapshot(self.configuration.snapshot()?);
+
+        // Snapshot MSI
+        if let Some(msi) = &mut self.interrupt.msi {
+            vfio_common_snapshot.add_snapshot(msi.cfg.snapshot()?);
+        }
+
+        // Snapshot MSI-X
+        if let Some(msix) = &mut self.interrupt.msix {
+            vfio_common_snapshot.add_snapshot(msix.bar.snapshot()?);
+        }
+
+        Ok(vfio_common_snapshot)
+    }
+
+    fn restore(&mut self, snapshot: Snapshot) -> std::result::Result<(), MigratableError> {
+        if let Some(vfio_common_section) = snapshot
+            .snapshot_data
+            .get(&format!("{}-section", self.id()))
+        {
+            // It has to be invoked first as we want Interrupt to be initialized
+            // correctly before we try to restore MSI and MSI-X configurations.
+            self.set_state(&vfio_common_section.to_versioned_state()?)
+                .map_err(|e| {
+                    MigratableError::Restore(anyhow!("Could not restore VFIO_COMMON state {:?}", e))
+                })?;
+
+            // Restore PciConfiguration
+            if let Some(pci_config_snapshot) = snapshot.snapshots.get(&self.configuration.id()) {
+                self.configuration.restore(*pci_config_snapshot.clone())?;
+            }
+
+            // Restore MSI
+            if let Some(msi) = &mut self.interrupt.msi {
+                if let Some(msi_snapshot) = snapshot.snapshots.get(&msi.cfg.id()) {
+                    msi.cfg.restore(*msi_snapshot.clone())?;
+                }
+                if msi.cfg.enabled() {
+                    self.enable_msi().unwrap();
+                }
+            }
+
+            // Restore MSI-X
+            if let Some(msix) = &mut self.interrupt.msix {
+                if let Some(msix_snapshot) = snapshot.snapshots.get(&msix.bar.id()) {
+                    msix.bar.restore(*msix_snapshot.clone())?;
+                }
+                if msix.bar.enabled() {
+                    self.enable_msix().unwrap();
+                }
+            }
+
+            return Ok(());
+        }
+
+        Err(MigratableError::Restore(anyhow!(
+            "Could not find VFIO_COMMON snapshot section"
+        )))
+    }
 }
 
 /// VfioPciDevice represents a VFIO PCI device.
@@ -985,6 +1139,7 @@ pub struct VfioPciDevice {
     container: Arc<VfioContainer>,
     common: VfioCommon,
     iommu_attached: bool,
+    memory_slot: Arc<dyn Fn() -> u32 + Send + Sync>,
 }
 
 impl VfioPciDevice {
@@ -999,6 +1154,8 @@ impl VfioPciDevice {
         legacy_interrupt_group: Option<Arc<dyn InterruptSourceGroup>>,
         iommu_attached: bool,
         bdf: PciBdf,
+        restoring: bool,
+        memory_slot: Arc<dyn Fn() -> u32 + Send + Sync>,
     ) -> Result<Self, VfioPciError> {
         let device = Arc::new(device);
         device.reset();
@@ -1031,8 +1188,13 @@ impl VfioPciDevice {
             vfio_wrapper: Arc::new(vfio_wrapper) as Arc<dyn Vfio>,
         };
 
-        common.parse_capabilities(bdf);
+        // No need to parse capabilities from the device if on the restore path.
-        common.initialize_legacy_interrupt()?;
+        // The initialization will be performed later when restore() will be
+        // called.
+        if !restoring {
+            common.parse_capabilities(bdf);
+            common.initialize_legacy_interrupt()?;
+        }
 
         let vfio_pci_device = VfioPciDevice {
             id,
@@ -1041,6 +1203,7 @@ impl VfioPciDevice {
             container,
             common,
             iommu_attached,
+            memory_slot,
         };
 
         Ok(vfio_pci_device)
@@ -1150,14 +1313,7 @@ impl VfioPciDevice {
     /// * `vm` - The VM object. It is used to set the VFIO MMIO regions
     ///          as user memory regions.
     /// * `mem_slot` - The closure to return a memory slot.
-    pub fn map_mmio_regions<F>(
+    pub fn map_mmio_regions(&mut self) -> Result<(), VfioPciError> {
-        &mut self,
-        vm: &Arc<dyn hypervisor::Vm>,
-        mem_slot: F,
-    ) -> Result<(), VfioPciError>
-    where
-        F: Fn() -> u32,
-    {
         let fd = self.device.as_raw_fd();
 
         for region in self.common.mmio_regions.iter_mut() {
@@ -1223,7 +1379,7 @@ impl VfioPciDevice {
                     }
 
                     let user_memory_region = UserMemoryRegion {
-                        slot: mem_slot(),
+                        slot: (self.memory_slot)(),
                         start: region.start.0 + area.offset,
                         size: area.size,
                         host_addr: host_addr as u64,
@@ -1231,7 +1387,7 @@ impl VfioPciDevice {
 
                     region.user_memory_regions.push(user_memory_region);
 
-                    let mem_region = vm.make_user_memory_region(
+                    let mem_region = self.vm.make_user_memory_region(
                         user_memory_region.slot,
                         user_memory_region.start,
                         user_memory_region.size,
@@ -1240,7 +1396,8 @@ impl VfioPciDevice {
                         false,
                     );
 
-                    vm.create_user_memory_region(mem_region)
+                    self.vm
+                        .create_user_memory_region(mem_region)
                         .map_err(VfioPciError::CreateUserMemoryRegion)?;
                 }
             }
@@ -1464,3 +1621,37 @@ impl PciDevice for VfioPciDevice {
         Some(self.id.clone())
     }
 }
+
+impl Pausable for VfioPciDevice {}
+
+impl Snapshottable for VfioPciDevice {
+    fn id(&self) -> String {
+        self.id.clone()
+    }
+
+    fn snapshot(&mut self) -> std::result::Result<Snapshot, MigratableError> {
+        let mut vfio_pci_dev_snapshot = Snapshot::new(&self.id);
+
+        // Snapshot VfioCommon
+        vfio_pci_dev_snapshot.add_snapshot(self.common.snapshot()?);
+
+        Ok(vfio_pci_dev_snapshot)
+    }
+
+    fn restore(&mut self, snapshot: Snapshot) -> std::result::Result<(), MigratableError> {
+        // Restore VfioCommon
+        if let Some(vfio_common_snapshot) = snapshot.snapshots.get(&self.common.id()) {
+            self.common.restore(*vfio_common_snapshot.clone())?;
+            self.map_mmio_regions().map_err(|e| {
+                MigratableError::Restore(anyhow!(
+                    "Could not map MMIO regions for VfioPciDevice on restore {:?}",
+                    e
+                ))
+            })?;
+        }
+
+        Ok(())
+    }
+}
+impl Transportable for VfioPciDevice {}
+impl Migratable for VfioPciDevice {}

pci/src/vfio_user.rs

@@ -8,6 +8,7 @@ use crate::{BarReprogrammingParams, PciBarConfiguration, VfioPciError};
 use crate::{
     PciBdf, PciClassCode, PciConfiguration, PciDevice, PciDeviceError, PciHeaderType, PciSubclass,
 };
+use anyhow::anyhow;
 use hypervisor::HypervisorVmError;
 use std::any::Any;
 use std::os::unix::prelude::AsRawFd;
@@ -26,6 +27,7 @@ use vm_memory::bitmap::AtomicBitmap;
 use vm_memory::{
     Address, GuestAddress, GuestAddressSpace, GuestMemory, GuestMemoryRegion, GuestRegionMmap,
 };
+use vm_migration::{Migratable, MigratableError, Pausable, Snapshot, Snapshottable, Transportable};
 use vmm_sys_util::eventfd::EventFd;
 
 pub struct VfioUserPciDevice {
@@ -33,6 +35,7 @@ pub struct VfioUserPciDevice {
     vm: Arc<dyn hypervisor::Vm>,
     client: Arc<Mutex<Client>>,
     common: VfioCommon,
+    memory_slot: Arc<dyn Fn() -> u32 + Send + Sync>,
 }
 
 #[derive(Error, Debug)]
@@ -61,6 +64,7 @@ impl PciSubclass for PciVfioUserSubclass {
 }
 
 impl VfioUserPciDevice {
+    #[allow(clippy::too_many_arguments)]
     pub fn new(
         id: String,
         vm: &Arc<dyn hypervisor::Vm>,
@@ -68,6 +72,8 @@ impl VfioUserPciDevice {
         msi_interrupt_manager: Arc<dyn InterruptManager<GroupConfig = MsiIrqGroupConfig>>,
         legacy_interrupt_group: Option<Arc<dyn InterruptSourceGroup>>,
         bdf: PciBdf,
+        restoring: bool,
+        memory_slot: Arc<dyn Fn() -> u32 + Send + Sync>,
     ) -> Result<Self, VfioUserPciDeviceError> {
         // This is used for the BAR and capabilities only
         let configuration = PciConfiguration::new(
@@ -108,27 +114,26 @@ impl VfioUserPciDevice {
             vfio_wrapper: Arc::new(vfio_wrapper) as Arc<dyn Vfio>,
         };
 
-        common.parse_capabilities(bdf);
+        // No need to parse capabilities from the device if on the restore path.
-        common
+        // The initialization will be performed later when restore() will be
-            .initialize_legacy_interrupt()
+        // called.
-            .map_err(VfioUserPciDeviceError::InitializeLegacyInterrupts)?;
+        if !restoring {
+            common.parse_capabilities(bdf);
+            common
+                .initialize_legacy_interrupt()
+                .map_err(VfioUserPciDeviceError::InitializeLegacyInterrupts)?;
+        }
 
         Ok(Self {
             id,
             vm: vm.clone(),
             client,
             common,
+            memory_slot,
         })
     }
 
-    pub fn map_mmio_regions<F>(
+    pub fn map_mmio_regions(&mut self) -> Result<(), VfioUserPciDeviceError> {
-        &mut self,
-        vm: &Arc<dyn hypervisor::Vm>,
-        mem_slot: F,
-    ) -> Result<(), VfioUserPciDeviceError>
-    where
-        F: Fn() -> u32,
-    {
         for mmio_region in &mut self.common.mmio_regions {
             let region_flags = self
                 .client
@@ -195,7 +200,7 @@ impl VfioUserPciDevice {
                     }
 
                     let user_memory_region = UserMemoryRegion {
-                        slot: mem_slot(),
+                        slot: (self.memory_slot)(),
                         start: mmio_region.start.0 + s.offset,
                         size: s.size,
                         host_addr: host_addr as u64,
@@ -203,7 +208,7 @@ impl VfioUserPciDevice {
 
                     mmio_region.user_memory_regions.push(user_memory_region);
 
-                    let mem_region = vm.make_user_memory_region(
+                    let mem_region = self.vm.make_user_memory_region(
                         user_memory_region.slot,
                         user_memory_region.start,
                         user_memory_region.size,
@@ -212,7 +217,8 @@ impl VfioUserPciDevice {
                         false,
                     );
 
-                    vm.create_user_memory_region(mem_region)
+                    self.vm
+                        .create_user_memory_region(mem_region)
                         .map_err(VfioUserPciDeviceError::MapRegionGuest)?;
                 }
             }
@@ -551,6 +557,40 @@ impl Drop for VfioUserPciDevice {
     }
 }
 
+impl Pausable for VfioUserPciDevice {}
+
+impl Snapshottable for VfioUserPciDevice {
+    fn id(&self) -> String {
+        self.id.clone()
+    }
+
+    fn snapshot(&mut self) -> std::result::Result<Snapshot, MigratableError> {
+        let mut vfio_pci_dev_snapshot = Snapshot::new(&self.id);
+
+        // Snapshot VfioCommon
+        vfio_pci_dev_snapshot.add_snapshot(self.common.snapshot()?);
+
+        Ok(vfio_pci_dev_snapshot)
+    }
+
+    fn restore(&mut self, snapshot: Snapshot) -> std::result::Result<(), MigratableError> {
+        // Restore VfioCommon
+        if let Some(vfio_common_snapshot) = snapshot.snapshots.get(&self.common.id()) {
+            self.common.restore(*vfio_common_snapshot.clone())?;
+            self.map_mmio_regions().map_err(|e| {
+                MigratableError::Restore(anyhow!(
+                    "Could not map MMIO regions for VfioUserPciDevice on restore {:?}",
+                    e
+                ))
+            })?;
+        }
+
+        Ok(())
+    }
+}
+impl Transportable for VfioUserPciDevice {}
+impl Migratable for VfioUserPciDevice {}
+
 pub struct VfioUserDmaMapping<M: GuestAddressSpace> {
     client: Arc<Mutex<Client>>,
     memory: Arc<M>,
@@ -584,10 +624,10 @@ impl<M: GuestAddressSpace + Sync + Send> ExternalDmaMapping for VfioUserDmaMappi
                     )
                 })
         } else {
-            return Err(std::io::Error::new(
+            Err(std::io::Error::new(
                 std::io::ErrorKind::Other,
                 format!("Region not found for 0x{:x}", gpa),
-            ));
+            ))
         }
     }
 

performance-metrics/Cargo.toml

@@ -6,13 +6,13 @@ edition = "2021"
 build = "build.rs"
 
 [dependencies]
-clap = { version = "3.1.18", features = ["wrap_help","cargo"] }
+clap = { version = "3.2.8", features = ["wrap_help","cargo"] }
 dirs = "4.0.0"
-serde = { version = "1.0.137", features = ["rc", "derive"] }
+serde = { version = "1.0.138", features = ["rc", "derive"] }
-serde_json = "1.0.81"
+serde_json = "1.0.82"
 test_infra = { path = "../test_infra" }
 thiserror = "1.0.31"
 wait-timeout = "0.2.0"
 
 [build-dependencies]
-clap = { version = "3.1.18", features = ["cargo"] }
+clap = { version = "3.2.8", features = ["cargo"] }

performance-metrics/src/main.rs

@@ -53,18 +53,30 @@ impl Default for MetricsReport {
         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 = String::from_utf8(git_out.stdout)
-                    git_human_readable = git_out_str.trim().to_string();
+                    .unwrap()
-                }
+                    .trim()
+                    .to_string();
+            } else {
+                eprintln!(
+                    "Error generating human readable git reference: {}",
+                    String::from_utf8(git_out.stderr).unwrap()
+                );
             }
         }
 
         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 = String::from_utf8(git_out.stdout)
-                    git_revision = git_out_str.trim().to_string();
+                    .unwrap()
-                }
+                    .trim()
+                    .to_string();
+            } else {
+                eprintln!(
+                    "Error generating git reference: {}",
+                    String::from_utf8(git_out.stderr).unwrap()
+                );
             }
         }
 
@@ -74,9 +86,15 @@ impl Default for MetricsReport {
             .output()
         {
             if git_out.status.success() {
-                if let Ok(git_out_str) = String::from_utf8(git_out.stdout) {
+                git_commit_date = String::from_utf8(git_out.stdout)
-                    git_commit_date = git_out_str.trim().to_string();
+                    .unwrap()
-                }
+                    .trim()
+                    .to_string();
+            } else {
+                eprintln!(
+                    "Error generating git commit date: {}",
+                    String::from_utf8(git_out.stderr).unwrap()
+                );
             }
         }
 

rate_limiter/src/lib.rs

@@ -91,7 +91,7 @@ pub enum BucketReduction {
 
 /// TokenBucket provides a lower level interface to rate limiting with a
 /// configurable capacity, refill-rate and initial burst.
-#[derive(Clone, Debug, PartialEq)]
+#[derive(Clone, Debug, PartialEq, Eq)]
 pub struct TokenBucket {
     // Bucket defining traits.
     size: u64,

release-notes.md

@@ -1,22 +1,28 @@
+- [v25.0](#v250)
+    - [`ch-remote` Improvements](#ch-remote-improvements)
+    - [VM "Coredump" Support](#vm-coredump-support)
+    - [Notable Bug Fixes](#notable-bug-fixes)
+    - [Removals](#removals)
+    - [Contributors](#contributors)
 - [v24.0](#v240)
     - [Bypass Mode for `virtio-iommu`](#bypass-mode-for-virtio-iommu)
     - [Ensure Identifiers Uniqueness](#ensure-identifiers-uniqueness)
     - [Sparse Mmap support](#sparse-mmap-support)
     - [Expose Platform Serial Number](#expose-platform-serial-number)
-    - [Notable Bug Fixes](#notable-bug-fixes)
+    - [Notable Bug Fixes](#notable-bug-fixes-1)
     - [Notable Improvements](#notable-improvements)
     - [Deprecations](#deprecations)
     - [New on the Website](#new-on-the-website)
-    - [Contributors](#contributors)
+    - [Contributors](#contributors-1)
 - [v23.1](#v231)
 - [v23.0](#v230)
     - [vDPA Support](#vdpa-support)
     - [Updated OS Support list](#updated-os-support-list)
     - [`AArch64` Memory Map Improvements](#aarch64-memory-map-improvements)
     - [`AMX` Support](#amx-support)
-    - [Notable Bug Fixes](#notable-bug-fixes-1)
+    - [Notable Bug Fixes](#notable-bug-fixes-2)
     - [Deprecations](#deprecations-1)
-    - [Contributors](#contributors-1)
+    - [Contributors](#contributors-2)
 - [v22.1](#v221)
 - [v22.0](#v220)
     - [GDB Debug Stub Support](#gdb-debug-stub-support)
@@ -27,13 +33,13 @@
     - [PMU Support for AArch64](#pmu-support-for-aarch64)
     - [Documentation Under CC-BY-4.0 License](#documentation-under-cc-by-40-license)
     - [Deprecation of "Classic" `virtiofsd`](#deprecation-of-classic-virtiofsd)
-    - [Notable Bug Fixes](#notable-bug-fixes-2)
+    - [Notable Bug Fixes](#notable-bug-fixes-3)
-    - [Contributors](#contributors-2)
+    - [Contributors](#contributors-3)
 - [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-3)
+    - [Notable Bug fixes](#notable-bug-fixes-4)
-    - [Contributors](#contributors-3)
+    - [Contributors](#contributors-4)
 - [v20.2](#v202)
 - [v20.1](#v201)
 - [v20.0](#v200)
@@ -42,8 +48,8 @@
     - [Improved VFIO support](#improved-vfio-support)
     - [Safer code](#safer-code)
     - [Extended documentation](#extended-documentation)
-    - [Notable bug fixes](#notable-bug-fixes-4)
+    - [Notable bug fixes](#notable-bug-fixes-5)
-    - [Contributors](#contributors-4)
+    - [Contributors](#contributors-5)
 - [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)
@@ -51,8 +57,8 @@
     - [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-5)
+    - [Notable bug fixes](#notable-bug-fixes-6)
-    - [Contributors](#contributors-5)
+    - [Contributors](#contributors-6)
 - [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)
@@ -62,23 +68,23 @@
     - [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-6)
+    - [Notable bug fixes](#notable-bug-fixes-7)
-    - [Contributors](#contributors-6)
+    - [Contributors](#contributors-7)
 - [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-7)
+    - [Notable bug fixes](#notable-bug-fixes-8)
-    - [Contributors](#contributors-7)
+    - [Contributors](#contributors-8)
 - [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-8)
+    - [Notable bug fixes](#notable-bug-fixes-9)
     - [Removed functionality](#removed-functionality)
-    - [Contributors](#contributors-8)
+    - [Contributors](#contributors-9)
 - [v15.0](#v150)
     - [Version numbering and stability guarantees](#version-numbering-and-stability-guarantees)
     - [Network device rate limiting](#network-device-rate-limiting)
@@ -86,7 +92,7 @@
     - [`--api-socket` supports file descriptor parameter](#--api-socket-supports-file-descriptor-parameter)
     - [Bug fixes](#bug-fixes)
     - [Deprecations](#deprecations-2)
-    - [Contributors](#contributors-9)
+    - [Contributors](#contributors-10)
 - [v0.14.1](#v0141)
 - [v0.14.0](#v0140)
     - [Structured event monitoring](#structured-event-monitoring)
@@ -96,7 +102,7 @@
     - [PTY control for serial and `virtio-console`](#pty-control-for-serial-and-virtio-console)
     - [Block device rate limiting](#block-device-rate-limiting)
     - [Deprecations](#deprecations-3)
-    - [Contributors](#contributors-10)
+    - [Contributors](#contributors-11)
 - [v0.13.0](#v0130)
     - [Wider VFIO device support](#wider-vfio-device-support)
     - [Improved huge page support](#improved-huge-page-support)
@@ -104,13 +110,13 @@
     - [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-11)
+    - [Contributors](#contributors-12)
 - [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-12)
+    - [Contributors](#contributors-13)
 - [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)
@@ -122,15 +128,15 @@
     - [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-9)
+    - [Notable Bug Fixes](#notable-bug-fixes-10)
-    - [Contributors](#contributors-13)
+    - [Contributors](#contributors-14)
 - [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-10)
+    - [Notable Bug Fixes](#notable-bug-fixes-11)
-    - [Contributors](#contributors-14)
+    - [Contributors](#contributors-15)
 - [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)
@@ -143,17 +149,17 @@
     - [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-11)
+    - [Notable Bug Fixes](#notable-bug-fixes-12)
-    - [Contributors](#contributors-15)
+    - [Contributors](#contributors-16)
 - [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-12)
+    - [Notable Bug Fixes](#notable-bug-fixes-13)
     - [Command Line and API Changes](#command-line-and-api-changes)
-    - [Contributors](#contributors-16)
+    - [Contributors](#contributors-17)
 - [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)
@@ -163,14 +169,14 @@
     - [`Seccomp` Sandboxing](#seccomp-sandboxing)
     - [Updated Distribution Support](#updated-distribution-support)
     - [Command Line and API Changes](#command-line-and-api-changes-1)
-    - [Contributors](#contributors-17)
+    - [Contributors](#contributors-18)
 - [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-18)
+    - [Contributors](#contributors-19)
 - [v0.5.1](#v051)
 - [v0.5.0](#v050)
     - [Virtual Machine Dynamic Resizing](#virtual-machine-dynamic-resizing)
@@ -178,7 +184,7 @@
     - [New Interrupt Management Framework](#new-interrupt-management-framework)
     - [Development Tools](#development-tools)
     - [Kata Containers Integration](#kata-containers-integration)
-    - [Contributors](#contributors-19)
+    - [Contributors](#contributors-20)
 - [v0.4.0](#v040)
     - [Dynamic virtual CPUs addition](#dynamic-virtual-cpus-addition)
     - [Programmatic firmware tables generation](#programmatic-firmware-tables-generation)
@@ -187,7 +193,7 @@
     - [Userspace IOAPIC by default](#userspace-ioapic-by-default)
     - [PCI BAR reprogramming](#pci-bar-reprogramming)
     - [New `cloud-hypervisor` organization](#new-cloud-hypervisor-organization)
-    - [Contributors](#contributors-20)
+    - [Contributors](#contributors-21)
 - [v0.3.0](#v030)
     - [Block device offloading](#block-device-offloading)
     - [Network device backend](#network-device-backend)
@@ -214,6 +220,52 @@
     - [Unit testing](#unit-testing)
     - [Integration tests parallelization](#integration-tests-parallelization)
 
+# v25.0
+
+This release has been tracked through the [v25.0
+project](https://github.com/cloud-hypervisor/cloud-hypervisor/projects/29).
+
+### `ch-remote` Improvements
+
+The `ch-remote` command has gained support for creating the VM from a JSON
+config and support for booting and deleting the VM from the VMM.
+
+### VM "Coredump" Support
+
+Under the `guest_debug` feature flag it is now possible to extract the memory
+of the guest for use in debugging with e.g. the `crash` utility. (#4012)
+
+### Notable Bug Fixes
+
+* Always restore console mode on exit (#4249, #4248)
+* Restore vCPUs in numerical order which fixes aarch64 snapshot/restore (#4244)
+* Don't try and configure `IFF_RUNNING` on TAP devices (#4279)
+* Propagate configured queue size through to vhost-user backend (#4286)
+* Always Program vCPU CPUID before running the vCPU to fix running on Linux
+  5.16 (#4156)
+* Enable ACPI MADT "Online Capable" flag for hotpluggable vCPUs to fix newer
+  Linux guest
+
+### Removals
+
+The following functionality has been removed:
+
+* The `mergeable` option from the `virtio-pmem` support has been removed
+  (#3968)
+* The `dax` option from the `virtio-fs` support has been removed (#3889)
+
+### Contributors
+
+Many thanks to everyone who has contributed to our release:
+
+* Dylan Bargatze <dbargatz@users.noreply.github.com>
+* Jinank Jain <jinankjain@microsoft.com>
+* Michael Zhao <michael.zhao@arm.com>
+* Rob Bradford <robert.bradford@intel.com>
+* Sebastien Boeuf <sebastien.boeuf@intel.com>
+* Wei Liu <liuwe@microsoft.com>
+* Yi Wang <wang.yi59@zte.com.cn>
+
 # v24.0
 
 This release has been tracked through the [v24.0

resources/Dockerfile

@@ -4,7 +4,7 @@
 FROM ubuntu:20.04 as dev
 
 ARG TARGETARCH
-ARG RUST_TOOLCHAIN="1.60.0"
+ARG RUST_TOOLCHAIN="1.62.0"
 ARG CLH_SRC_DIR="/cloud-hypervisor"
 ARG CLH_BUILD_DIR="$CLH_SRC_DIR/build"
 ARG CARGO_REGISTRY_DIR="$CLH_BUILD_DIR/cargo_registry"

rpm/cloud-hypervisor.spec

@@ -11,7 +11,7 @@
 
 Name:           cloud-hypervisor
 Summary:        Cloud Hypervisor is an open source Virtual Machine Monitor (VMM) that runs on top of KVM.
-Version:        24.0
+Version:        25.0
 Release:        0%{?dist}
 License:        ASL 2.0 or BSD-3-clause
 Group:          Applications/System
@@ -145,6 +145,9 @@ rm -rf %{buildroot}
 
 
 %changelog
+*   Thu Jul 07 2022 Rob Bradford <robert.bradford@intel.com> 25.0-0
+-   Update to 25.0
+
 *   Wed May 25 2022 Sebastien Boeuf <sebastien.boeuf@intel.com> 24.0-0
 -   Update to 24.0
 

scripts/dev_cli.sh

@@ -7,7 +7,7 @@
 CLI_NAME="Cloud Hypervisor"
 
 CTR_IMAGE_TAG="cloudhypervisor/dev"
-CTR_IMAGE_VERSION="20220524-0"
+CTR_IMAGE_VERSION="20220705-0"
 CTR_IMAGE="${CTR_IMAGE_TAG}:${CTR_IMAGE_VERSION}"
 
 DOCKER_RUNTIME="docker"

scripts/run_metrics.sh

@@ -113,6 +113,9 @@ if [ -n "$test_filter" ]; then
     test_binary_args+=("--test-filter $test_filter")
 fi
 
+# Ensure that git commands can be run in this directory (for metrics report)
+git config --global --add safe.directory $PWD
+
 export RUST_BACKTRACE=1
 time target/$BUILD_TARGET/release/performance-metrics ${test_binary_args[*]}
 RES=$?

src/bin/ch-remote.rs

@@ -12,6 +12,7 @@ use api_client::Error as ApiClientError;
 use clap::{Arg, ArgMatches, Command};
 use option_parser::{ByteSized, ByteSizedParseError};
 use std::fmt;
+use std::io::Read;
 use std::os::unix::net::UnixStream;
 use std::process;
 
@@ -31,6 +32,8 @@ enum Error {
     AddVdpaConfig(vmm::config::Error),
     AddVsockConfig(vmm::config::Error),
     Restore(vmm::config::Error),
+    ReadingStdin(std::io::Error),
+    ReadingFile(std::io::Error),
 }
 
 impl fmt::Display for Error {
@@ -51,6 +54,8 @@ impl fmt::Display for Error {
             AddVdpaConfig(e) => write!(f, "Error parsing vDPA device syntax: {}", e),
             AddVsockConfig(e) => write!(f, "Error parsing vsock syntax: {}", e),
             Restore(e) => write!(f, "Error parsing restore syntax: {}", e),
+            ReadingStdin(e) => write!(f, "Error reading from stdin: {}", e),
+            ReadingFile(e) => write!(f, "Error reading from file: {}", e),
         }
     }
 }
@@ -264,6 +269,20 @@ fn restore_api_command(socket: &mut UnixStream, config: &str) -> Result<(), Erro
     .map_err(Error::ApiClient)
 }
 
+fn coredump_api_command(socket: &mut UnixStream, destination_url: &str) -> Result<(), Error> {
+    let coredump_config = vmm::api::VmCoredumpData {
+        destination_url: String::from(destination_url),
+    };
+
+    simple_api_command(
+        socket,
+        "PUT",
+        "coredump",
+        Some(&serde_json::to_string(&coredump_config).unwrap()),
+    )
+    .map_err(Error::ApiClient)
+}
+
 fn receive_migration_api_command(socket: &mut UnixStream, url: &str) -> Result<(), Error> {
     let receive_migration_data = vmm::api::VmReceiveMigrationData {
         receiver_url: url.to_owned(),
@@ -295,6 +314,19 @@ fn send_migration_api_command(
     .map_err(Error::ApiClient)
 }
 
+fn create_api_command(socket: &mut UnixStream, path: &str) -> Result<(), Error> {
+    let mut data = String::default();
+    if path == "-" {
+        std::io::stdin()
+            .read_to_string(&mut data)
+            .map_err(Error::ReadingStdin)?;
+    } else {
+        data = std::fs::read_to_string(path).map_err(Error::ReadingFile)?;
+    }
+
+    simple_api_command(socket, "PUT", "create", Some(&data)).map_err(Error::ApiClient)
+}
+
 fn do_command(matches: &ArgMatches) -> Result<(), Error> {
     let mut socket =
         UnixStream::connect(matches.value_of("api-socket").unwrap()).map_err(Error::Connect)?;
@@ -422,6 +454,14 @@ fn do_command(matches: &ArgMatches) -> Result<(), Error> {
                 .value_of("restore_config")
                 .unwrap(),
         ),
+        Some("coredump") => coredump_api_command(
+            &mut socket,
+            matches
+                .subcommand_matches("coredump")
+                .unwrap()
+                .value_of("coredump_config")
+                .unwrap(),
+        ),
         Some("send-migration") => send_migration_api_command(
             &mut socket,
             matches
@@ -442,6 +482,14 @@ fn do_command(matches: &ArgMatches) -> Result<(), Error> {
                 .value_of("receive_migration_config")
                 .unwrap(),
         ),
+        Some("create") => create_api_command(
+            &mut socket,
+            matches
+                .subcommand_matches("create")
+                .unwrap()
+                .value_of("path")
+                .unwrap(),
+        ),
         Some(c) => simple_api_command(&mut socket, "PUT", c, None).map_err(Error::ApiClient),
         None => unreachable!(),
     }
@@ -576,6 +624,8 @@ fn main() {
                 ),
         )
         .subcommand(Command::new("resume").about("Resume the VM"))
+        .subcommand(Command::new("boot").about("Boot a created VM"))
+        .subcommand(Command::new("delete").about("Delete a VM"))
         .subcommand(Command::new("shutdown").about("Shutdown the VM"))
         .subcommand(
             Command::new("snapshot")
@@ -595,6 +645,11 @@ fn main() {
                         .help(vmm::config::RestoreConfig::SYNTAX),
                 ),
         )
+        .subcommand(
+            Command::new("coredump")
+                .about("Create a coredump from VM")
+                .arg(Arg::new("coredump_config").index(1).help("<file_path>")),
+        )
         .subcommand(
             Command::new("send-migration")
                 .about("Initiate a VM migration")
@@ -617,6 +672,11 @@ fn main() {
                         .index(1)
                         .help("<receiver_url>"),
                 ),
+        )
+        .subcommand(
+            Command::new("create")
+                .about("Create VM from a JSON configuration")
+                .arg(Arg::new("path").index(1).default_value("-")),
         );
 
     let matches = app.get_matches();

src/main.rs

@@ -23,6 +23,7 @@ use thiserror::Error;
 use vmm::config;
 use vmm_sys_util::eventfd::EventFd;
 use vmm_sys_util::signal::block_signal;
+use vmm_sys_util::terminal::Terminal;
 
 #[derive(Error, Debug)]
 enum Error {
@@ -623,6 +624,13 @@ fn main() {
         }
     };
 
+    let on_tty = unsafe { libc::isatty(libc::STDIN_FILENO as i32) } != 0;
+    if on_tty {
+        // Don't forget to set the terminal in canonical mode
+        // before to exit.
+        std::io::stdin().lock().set_canon_mode().unwrap();
+    }
+
     std::process::exit(exit_code);
 }
 
@@ -1323,40 +1331,6 @@ mod unit_tests {
                 }"#,
                 false,
             ),
-            #[cfg(target_arch = "x86_64")]
-            (
-                vec![
-                    "cloud-hypervisor",
-                    "--kernel",
-                    "/path/to/kernel",
-                    "--pmem",
-                    "file=/path/to/img/1,size=1G,mergeable=on",
-                ],
-                r#"{
-                    "kernel": {"path": "/path/to/kernel"},
-                    "pmem": [
-                        {"file": "/path/to/img/1", "size": 1073741824, "mergeable": true}
-                    ]
-                }"#,
-                true,
-            ),
-            #[cfg(target_arch = "x86_64")]
-            (
-                vec![
-                    "cloud-hypervisor",
-                    "--kernel",
-                    "/path/to/kernel",
-                    "--pmem",
-                    "file=/path/to/img/1,size=1G,mergeable=off",
-                ],
-                r#"{
-                    "kernel": {"path": "/path/to/kernel"},
-                    "pmem": [
-                        {"file": "/path/to/img/1", "size": 1073741824, "mergeable": false}
-                    ]
-                }"#,
-                true,
-            ),
         ]
         .iter()
         .for_each(|(cli, openapi, equal)| {

test_infra/Cargo.toml

@@ -7,8 +7,8 @@ edition = "2021"
 [dependencies]
 dirs = "4.0.0"
 epoll = "4.3.1"
-lazy_static = "1.4.0"
 libc = "0.2.126"
+once_cell = "1.13.0"
 ssh2 = { version = "0.9.1", features = ["vendored-openssl"]}
 vmm-sys-util = "0.9.0"
 wait-timeout = "0.2.0"

test_infra/src/lib.rs

@@ -3,9 +3,7 @@
 // SPDX-License-Identifier: Apache-2.0
 //
 
-#[macro_use]
+use once_cell::sync::Lazy;
-extern crate lazy_static;
-
 use ssh2::Session;
 use std::env;
 use std::ffi::OsStr;
@@ -52,7 +50,7 @@ pub struct GuestNetworkConfig {
 
 pub const DEFAULT_TCP_LISTENER_MESSAGE: &str = "booted";
 pub const DEFAULT_TCP_LISTENER_PORT: u16 = 8000;
-pub const DEFAULT_TCP_LISTENER_TIMEOUT: i32 = 80;
+pub const DEFAULT_TCP_LISTENER_TIMEOUT: i32 = 120;
 
 #[derive(Debug)]
 pub enum WaitForBootError {
@@ -722,9 +720,7 @@ pub fn exec_host_command_output(command: &str) -> Output {
 
 pub const PIPE_SIZE: i32 = 32 << 20;
 
-lazy_static! {
+static NEXT_VM_ID: Lazy<Mutex<u8>> = Lazy::new(|| Mutex::new(1));
-    static ref NEXT_VM_ID: Mutex<u8> = Mutex::new(1);
-}
 
 pub struct Guest {
     pub tmp_dir: TempDir,
@@ -832,30 +828,12 @@ impl Guest {
         )
     }
 
-    pub fn api_create_body(
+    pub fn api_create_body(&self, cpu_count: u8, kernel_path: &str, kernel_cmd: &str) -> String {
-        &self,
-        cpu_count: u8,
-        _fw_path: &str,
-        _kernel_path: &str,
-        _kernel_cmd: &str,
-    ) -> String {
-        #[cfg(all(target_arch = "x86_64", not(feature = "mshv")))]
-        format! {"{{\"cpus\":{{\"boot_vcpus\":{},\"max_vcpus\":{}}},\"kernel\":{{\"path\":\"{}\"}},\"cmdline\":{{\"args\": \"\"}},\"net\":[{{\"ip\":\"{}\", \"mask\":\"255.255.255.0\", \"mac\":\"{}\"}}], \"disks\":[{{\"path\":\"{}\"}}, {{\"path\":\"{}\"}}]}}",
-                 cpu_count,
-                 cpu_count,
-                 _fw_path,
-                 self.network.host_ip,
-                 self.network.guest_mac,
-                 self.disk_config.disk(DiskType::OperatingSystem).unwrap().as_str(),
-                 self.disk_config.disk(DiskType::CloudInit).unwrap().as_str(),
-        }
-
-        #[cfg(any(target_arch = "aarch64", feature = "mshv"))]
         format! {"{{\"cpus\":{{\"boot_vcpus\":{},\"max_vcpus\":{}}},\"kernel\":{{\"path\":\"{}\"}},\"cmdline\":{{\"args\": \"{}\"}},\"net\":[{{\"ip\":\"{}\", \"mask\":\"255.255.255.0\", \"mac\":\"{}\"}}], \"disks\":[{{\"path\":\"{}\"}}, {{\"path\":\"{}\"}}]}}",
                  cpu_count,
                  cpu_count,
-                 _kernel_path,
+                 kernel_path,
-                 _kernel_cmd,
+                 kernel_cmd,
                  self.network.host_ip,
                  self.network.guest_mac,
                  self.disk_config.disk(DiskType::OperatingSystem).unwrap().as_str(),
@@ -985,13 +963,6 @@ impl Guest {
         Ok(())
     }
 
-    pub fn get_entropy(&self) -> Result<u32, Error> {
-        self.ssh_command("cat /proc/sys/kernel/random/entropy_avail")?
-            .trim()
-            .parse()
-            .map_err(Error::Parsing)
-    }
-
     pub fn get_pci_bridge_class(&self) -> Result<String, Error> {
         Ok(self
             .ssh_command("cat /sys/bus/pci/devices/0000:00:00.0/class")?
@@ -1037,50 +1008,6 @@ impl Guest {
         Ok(false)
     }
 
-    pub fn valid_virtio_fs_cache_size(
-        &self,
-        dax: bool,
-        cache_size: Option<u64>,
-    ) -> Result<bool, Error> {
-        // SHM region is called different things depending on kernel
-        let shm_region = self
-            .ssh_command("sudo grep 'virtio[0-9]\\|virtio-pci-shm' /proc/iomem || true")?
-            .trim()
-            .to_string();
-
-        if shm_region.is_empty() {
-            return Ok(!dax);
-        }
-
-        // From this point, the region is not empty, hence it is an error
-        // if DAX is off.
-        if !dax {
-            return Ok(false);
-        }
-
-        let cache = if let Some(cache) = cache_size {
-            cache
-        } else {
-            // 8Gib by default
-            0x0002_0000_0000
-        };
-
-        let args: Vec<&str> = shm_region.split(':').collect();
-        if args.is_empty() {
-            return Ok(false);
-        }
-
-        let args: Vec<&str> = args[0].trim().split('-').collect();
-        if args.len() != 2 {
-            return Ok(false);
-        }
-
-        let start_addr = u64::from_str_radix(args[0], 16).map_err(Error::Parsing)?;
-        let end_addr = u64::from_str_radix(args[1], 16).map_err(Error::Parsing)?;
-
-        Ok(cache == (end_addr - start_addr + 1))
-    }
-
     pub fn check_vsock(&self, socket: &str) {
         // Listen from guest on vsock CID=3 PORT=16
         // SOCKET-LISTEN:<domain>:<protocol>:<local-address>

tests/integration.rs

@@ -8,10 +8,6 @@
 // related warnings for our quality workflow to pass.
 #![allow(dead_code)]
 
-#[cfg(target_arch = "x86_64")]
-#[macro_use]
-extern crate lazy_static;
-
 extern crate test_infra;
 
 use net_util::MacAddr;
@@ -75,11 +71,7 @@ const DIRECT_KERNEL_BOOT_CMDLINE: &str =
 
 const CONSOLE_TEST_STRING: &str = "Started OpenBSD Secure Shell server";
 
-fn prepare_virtiofsd(
+fn prepare_virtiofsd(tmp_dir: &TempDir, shared_dir: &str) -> (std::process::Child, String) {
-    tmp_dir: &TempDir,
-    shared_dir: &str,
-    cache: &str,
-) -> (std::process::Child, String) {
     let mut workload_path = dirs::home_dir().unwrap();
     workload_path.push("workloads");
 
@@ -94,7 +86,7 @@ fn prepare_virtiofsd(
     let child = Command::new(virtiofsd_path.as_str())
         .args(&["--shared-dir", shared_dir])
         .args(&["--socket-path", virtiofsd_socket_path.as_str()])
-        .args(&["--cache", cache])
+        .args(&["--cache", "never"])
         .spawn()
         .unwrap();
 
@@ -162,6 +154,11 @@ fn temp_snapshot_dir_path(tmp_dir: &TempDir) -> String {
     snapshot_dir
 }
 
+fn temp_vmcore_file_path(tmp_dir: &TempDir) -> String {
+    let vmcore_file = String::from(tmp_dir.as_path().join("vmcore").to_str().unwrap());
+    vmcore_file
+}
+
 // 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.
@@ -256,8 +253,15 @@ fn remote_command(api_socket: &str, command: &str, arg: Option<&str>) -> bool {
     if let Some(arg) = arg {
         cmd.arg(arg);
     }
-
+    let output = cmd.output().unwrap();
-    cmd.status().expect("Failed to launch ch-remote").success()
+    if output.status.success() {
+        true
+    } else {
+        eprintln!("Error running ch-remote command: {:?}", &cmd);
+        let stderr = String::from_utf8_lossy(&output.stderr);
+        eprintln!("stderr: {}", stderr);
+        false
+    }
 }
 
 fn remote_command_w_output(api_socket: &str, command: &str, arg: Option<&str>) -> (bool, Vec<u8>) {
@@ -1126,11 +1130,8 @@ fn test_boot_from_vhost_user_blk(
     handle_child_output(r, &output);
 }
 
-fn test_virtio_fs(
+fn _test_virtio_fs(
-    dax: bool,
+    prepare_daemon: &dyn Fn(&TempDir, &str) -> (std::process::Child, String),
-    cache_size: Option<u64>,
-    virtiofsd_cache: &str,
-    prepare_daemon: &dyn Fn(&TempDir, &str, &str) -> (std::process::Child, String),
     hotplug: bool,
     pci_segment: Option<u16>,
 ) {
@@ -1161,18 +1162,8 @@ fn test_virtio_fs(
         direct_kernel_boot_path()
     };
 
-    let (dax_vmm_param, dax_mount_param) = if dax { ("on", "-o dax") } else { ("off", "") };
+    let (mut daemon_child, virtiofsd_socket_path) =
-    let cache_size_vmm_param = if let Some(cache) = cache_size {
+        prepare_daemon(&guest.tmp_dir, shared_dir.to_str().unwrap());
-        format!(",cache_size={}", cache)
-    } else {
-        "".to_string()
-    };
-
-    let (mut daemon_child, virtiofsd_socket_path) = prepare_daemon(
-        &guest.tmp_dir,
-        shared_dir.to_str().unwrap(),
-        virtiofsd_cache,
-    );
 
     let mut guest_command = GuestCommand::new(&guest);
     guest_command
@@ -1188,10 +1179,8 @@ fn test_virtio_fs(
     }
 
     let fs_params = format!(
-        "id=myfs0,tag=myfs,socket={},num_queues=1,queue_size=1024,dax={}{}{}",
+        "id=myfs0,tag=myfs,socket={},num_queues=1,queue_size=1024{}",
         virtiofsd_socket_path,
-        dax_vmm_param,
-        cache_size_vmm_param,
         if let Some(pci_segment) = pci_segment {
             format!(",pci_segment={}", pci_segment)
         } else {
@@ -1228,16 +1217,9 @@ fn test_virtio_fs(
         }
 
         // Mount shared directory through virtio_fs filesystem
-        let mount_cmd = format!(
+        guest
-            "mkdir -p mount_dir && \
+            .ssh_command("mkdir -p mount_dir && sudo mount -t virtiofs myfs mount_dir/")
-                 sudo mount -t virtiofs {} myfs mount_dir/",
+            .unwrap();
-            dax_mount_param
-        );
-        guest.ssh_command(&mount_cmd).unwrap();
-
-        assert!(guest
-            .valid_virtio_fs_cache_size(dax, cache_size)
-            .unwrap_or_default());
 
         // Check file1 exists and its content is "foo"
         assert_eq!(
@@ -1284,19 +1266,14 @@ fn test_virtio_fs(
 
     let (r, hotplug_daemon_child) = if r.is_ok() && hotplug {
         thread::sleep(std::time::Duration::new(10, 0));
-        let (daemon_child, virtiofsd_socket_path) = prepare_daemon(
+        let (daemon_child, virtiofsd_socket_path) =
-            &guest.tmp_dir,
+            prepare_daemon(&guest.tmp_dir, shared_dir.to_str().unwrap());
-            shared_dir.to_str().unwrap(),
-            virtiofsd_cache,
-        );
 
         let r = std::panic::catch_unwind(|| {
             thread::sleep(std::time::Duration::new(10, 0));
             let fs_params = format!(
-                "id=myfs0,tag=myfs,socket={},num_queues=1,queue_size=1024,dax={}{}{}",
+                "id=myfs0,tag=myfs,socket={},num_queues=1,queue_size=1024{}",
                 virtiofsd_socket_path,
-                dax_vmm_param,
-                cache_size_vmm_param,
                 if let Some(pci_segment) = pci_segment {
                     format!(",pci_segment={}", pci_segment)
                 } else {
@@ -1320,12 +1297,10 @@ fn test_virtio_fs(
 
             thread::sleep(std::time::Duration::new(10, 0));
             // Mount shared directory through virtio_fs filesystem
-            let mount_cmd = format!(
+            guest
-                "mkdir -p mount_dir && \
+                .ssh_command("mkdir -p mount_dir && sudo mount -t virtiofs myfs mount_dir/")
-                     sudo mount -t virtiofs {} myfs mount_dir/",
+                .unwrap();
-                dax_mount_param
+
-            );
-            guest.ssh_command(&mount_cmd).unwrap();
             // Check file1 exists and its content is "foo"
             assert_eq!(
                 guest.ssh_command("cat mount_dir/file1").unwrap().trim(),
@@ -1646,7 +1621,7 @@ fn process_rss_kib(pid: u32) -> usize {
 // 10MB is our maximum accepted overhead.
 const MAXIMUM_VMM_OVERHEAD_KB: u32 = 10 * 1024;
 
-#[derive(PartialEq, PartialOrd)]
+#[derive(PartialEq, Eq, PartialOrd)]
 struct Counters {
     rx_bytes: u64,
     rx_frames: u64,
@@ -2028,7 +2003,6 @@ mod parallel {
             assert_eq!(guest.get_cpu_count().unwrap_or_default(), 1);
             assert_eq!(guest.get_initial_apicid().unwrap_or(1), 0);
             assert!(guest.get_total_memory().unwrap_or_default() > 480_000);
-            assert!(guest.get_entropy().unwrap_or_default() >= 900);
             assert_eq!(guest.get_pci_bridge_class().unwrap_or_default(), "0x060000");
 
             let expected_sequential_events = [
@@ -2654,7 +2628,6 @@ mod parallel {
 
             assert_eq!(guest.get_cpu_count().unwrap_or_default(), 1);
             assert!(guest.get_total_memory().unwrap_or_default() > 480_000);
-            assert!(guest.get_entropy().unwrap_or_default() >= 900);
 
             let grep_cmd = if cfg!(target_arch = "x86_64") {
                 "grep -c PCI-MSI /proc/interrupts"
@@ -3119,25 +3092,25 @@ mod parallel {
     }
 
     #[test]
-    fn test_virtio_fs_dax_off() {
+    fn test_virtio_fs() {
-        test_virtio_fs(false, None, "never", &prepare_virtiofsd, false, None)
+        _test_virtio_fs(&prepare_virtiofsd, false, None)
     }
 
     #[test]
-    fn test_virtio_fs_hotplug_dax_off() {
+    fn test_virtio_fs_hotplug() {
-        test_virtio_fs(false, None, "never", &prepare_virtiofsd, true, None)
+        _test_virtio_fs(&prepare_virtiofsd, true, None)
     }
 
     #[test]
     #[cfg(not(feature = "mshv"))]
     fn test_virtio_fs_multi_segment_hotplug() {
-        test_virtio_fs(false, None, "never", &prepare_virtiofsd, true, Some(15))
+        _test_virtio_fs(&prepare_virtiofsd, true, Some(15))
     }
 
     #[test]
     #[cfg(not(feature = "mshv"))]
     fn test_virtio_fs_multi_segment() {
-        test_virtio_fs(false, None, "never", &prepare_virtiofsd, false, Some(15))
+        _test_virtio_fs(&prepare_virtiofsd, false, Some(15))
     }
 
     #[test]
@@ -3984,7 +3957,6 @@ mod parallel {
 
             assert_eq!(guest.get_cpu_count().unwrap_or_default(), 1);
             assert!(guest.get_total_memory().unwrap_or_default() > 480_000);
-            assert!(guest.get_entropy().unwrap_or_default() >= 900);
         });
 
         let _ = child.kill();
@@ -4073,7 +4045,60 @@ mod parallel {
         let cpu_count: u8 = 4;
         let http_body = guest.api_create_body(
             cpu_count,
-            fw_path(FwType::RustHypervisorFirmware).as_str(),
+            direct_kernel_boot_path().to_str().unwrap(),
+            DIRECT_KERNEL_BOOT_CMDLINE,
+        );
+
+        let temp_config_path = guest.tmp_dir.as_path().join("config");
+        std::fs::write(&temp_config_path, http_body).unwrap();
+
+        remote_command(
+            &api_socket,
+            "create",
+            Some(temp_config_path.as_os_str().to_str().unwrap()),
+        );
+
+        // Then boot it
+        remote_command(&api_socket, "boot", None);
+        thread::sleep(std::time::Duration::new(20, 0));
+
+        let r = std::panic::catch_unwind(|| {
+            // Check that the VM booted as expected
+            assert_eq!(guest.get_cpu_count().unwrap_or_default() as u8, cpu_count);
+            assert!(guest.get_total_memory().unwrap_or_default() > 480_000);
+        });
+
+        let _ = child.kill();
+        let output = child.wait_with_output().unwrap();
+
+        handle_child_output(r, &output);
+    }
+
+    #[test]
+    // Start cloud-hypervisor with no VM parameters, only the API server running.
+    // From the API: Create a VM, boot it and check it can be shutdown and then
+    // booted again
+    fn test_api_shutdown() {
+        let focal = UbuntuDiskConfig::new(FOCAL_IMAGE_NAME.to_string());
+        let guest = Guest::new(Box::new(focal));
+
+        let api_socket = temp_api_path(&guest.tmp_dir);
+
+        let mut child = GuestCommand::new(&guest)
+            .args(&["--api-socket", &api_socket])
+            .capture_output()
+            .spawn()
+            .unwrap();
+
+        thread::sleep(std::time::Duration::new(1, 0));
+
+        // Verify API server is running
+        curl_command(&api_socket, "GET", "http://localhost/api/v1/vmm.ping", None);
+
+        // Create the VM first
+        let cpu_count: u8 = 4;
+        let http_body = guest.api_create_body(
+            cpu_count,
             direct_kernel_boot_path().to_str().unwrap(),
             DIRECT_KERNEL_BOOT_CMDLINE,
         );
@@ -4087,13 +4112,113 @@ mod parallel {
 
         // Then boot it
         curl_command(&api_socket, "PUT", "http://localhost/api/v1/vm.boot", None);
-        thread::sleep(std::time::Duration::new(20, 0));
 
         let r = std::panic::catch_unwind(|| {
+            guest.wait_vm_boot(None).unwrap();
+
+            // Check that the VM booted as expected
+            assert_eq!(guest.get_cpu_count().unwrap_or_default() as u8, cpu_count);
+            assert!(guest.get_total_memory().unwrap_or_default() > 480_000);
+
+            // Shutdown without powering off to prevent filesystem corruption
+            guest.ssh_command("sudo shutdown -H now").unwrap();
+
+            // Then shut it down
+            curl_command(
+                &api_socket,
+                "PUT",
+                "http://localhost/api/v1/vm.shutdown",
+                None,
+            );
+
+            // Then boot it again
+            curl_command(&api_socket, "PUT", "http://localhost/api/v1/vm.boot", None);
+
+            guest.wait_vm_boot(None).unwrap();
+
+            // Check that the VM booted as expected
+            assert_eq!(guest.get_cpu_count().unwrap_or_default() as u8, cpu_count);
+            assert!(guest.get_total_memory().unwrap_or_default() > 480_000);
+        });
+
+        let _ = child.kill();
+        let output = child.wait_with_output().unwrap();
+
+        handle_child_output(r, &output);
+    }
+
+    #[test]
+    // Start cloud-hypervisor with no VM parameters, only the API server running.
+    // From the API: Create a VM, boot it and check it can be deleted and then recreated
+    // booted again.
+    fn test_api_delete() {
+        let focal = UbuntuDiskConfig::new(FOCAL_IMAGE_NAME.to_string());
+        let guest = Guest::new(Box::new(focal));
+
+        let api_socket = temp_api_path(&guest.tmp_dir);
+
+        let mut child = GuestCommand::new(&guest)
+            .args(&["--api-socket", &api_socket])
+            .capture_output()
+            .spawn()
+            .unwrap();
+
+        thread::sleep(std::time::Duration::new(1, 0));
+
+        // Verify API server is running
+        curl_command(&api_socket, "GET", "http://localhost/api/v1/vmm.ping", None);
+
+        // Create the VM first
+        let cpu_count: u8 = 4;
+        let http_body = guest.api_create_body(
+            cpu_count,
+            direct_kernel_boot_path().to_str().unwrap(),
+            DIRECT_KERNEL_BOOT_CMDLINE,
+        );
+
+        let r = std::panic::catch_unwind(|| {
+            curl_command(
+                &api_socket,
+                "PUT",
+                "http://localhost/api/v1/vm.create",
+                Some(&http_body),
+            );
+
+            // Then boot it
+            curl_command(&api_socket, "PUT", "http://localhost/api/v1/vm.boot", None);
+
+            guest.wait_vm_boot(None).unwrap();
+
+            // Check that the VM booted as expected
+            assert_eq!(guest.get_cpu_count().unwrap_or_default() as u8, cpu_count);
+            assert!(guest.get_total_memory().unwrap_or_default() > 480_000);
+
+            // Shutdown without powering off to prevent filesystem corruption
+            guest.ssh_command("sudo shutdown -H now").unwrap();
+
+            // Then delete it
+            curl_command(
+                &api_socket,
+                "PUT",
+                "http://localhost/api/v1/vm.delete",
+                None,
+            );
+
+            curl_command(
+                &api_socket,
+                "PUT",
+                "http://localhost/api/v1/vm.create",
+                Some(&http_body),
+            );
+
+            // Then boot it again
+            curl_command(&api_socket, "PUT", "http://localhost/api/v1/vm.boot", None);
+
+            guest.wait_vm_boot(None).unwrap();
+
             // Check that the VM booted as expected
             assert_eq!(guest.get_cpu_count().unwrap_or_default() as u8, cpu_count);
             assert!(guest.get_total_memory().unwrap_or_default() > 480_000);
-            assert!(guest.get_entropy().unwrap_or_default() >= 900);
         });
 
         let _ = child.kill();
@@ -4128,7 +4253,6 @@ mod parallel {
         let cpu_count: u8 = 4;
         let http_body = guest.api_create_body(
             cpu_count,
-            fw_path(FwType::RustHypervisorFirmware).as_str(),
             direct_kernel_boot_path().to_str().unwrap(),
             DIRECT_KERNEL_BOOT_CMDLINE,
         );
@@ -4147,7 +4271,6 @@ mod parallel {
             // Check that the VM booted as expected
             assert_eq!(guest.get_cpu_count().unwrap_or_default() as u8, cpu_count);
             assert!(guest.get_total_memory().unwrap_or_default() > 480_000);
-            assert!(guest.get_entropy().unwrap_or_default() >= 900);
 
             // We now pause the VM
             assert!(remote_command(&api_socket, "pause", None));
@@ -5664,6 +5787,56 @@ mod parallel {
         handle_child_output(r, &output);
     }
 
+    #[test]
+    #[cfg(feature = "guest_debug")]
+    fn test_coredump() {
+        let focal = UbuntuDiskConfig::new(FOCAL_IMAGE_NAME.to_string());
+        let guest = Guest::new(Box::new(focal));
+        let api_socket = temp_api_path(&guest.tmp_dir);
+
+        let mut cmd = GuestCommand::new(&guest);
+        cmd.args(&["--cpus", "boot=4"])
+            .args(&["--memory", "size=4G"])
+            .args(&["--kernel", fw_path(FwType::RustHypervisorFirmware).as_str()])
+            .default_disks()
+            .args(&["--net", guest.default_net_string().as_str()])
+            .args(&["--api-socket", &api_socket])
+            .capture_output();
+
+        let mut child = cmd.spawn().unwrap();
+        let vmcore_file = temp_vmcore_file_path(&guest.tmp_dir);
+
+        let r = std::panic::catch_unwind(|| {
+            guest.wait_vm_boot(None).unwrap();
+
+            assert!(remote_command(&api_socket, "pause", None));
+
+            assert!(remote_command(
+                &api_socket,
+                "coredump",
+                Some(format!("file://{}", vmcore_file).as_str()),
+            ));
+
+            // the num of CORE notes should equals to vcpu
+            let readelf_core_num_cmd = format!(
+                "readelf --all {} |grep CORE |grep -v Type |wc -l",
+                vmcore_file
+            );
+            let core_num_in_elf = exec_host_command_output(&readelf_core_num_cmd);
+            assert_eq!(String::from_utf8_lossy(&core_num_in_elf.stdout).trim(), "4");
+
+            // the num of QEMU notes should equals to vcpu
+            let readelf_vmm_num_cmd = format!("readelf --all {} |grep QEMU |wc -l", vmcore_file);
+            let vmm_num_in_elf = exec_host_command_output(&readelf_vmm_num_cmd);
+            assert_eq!(String::from_utf8_lossy(&vmm_num_in_elf.stdout).trim(), "4");
+        });
+
+        let _ = child.kill();
+        let output = child.wait_with_output().unwrap();
+
+        handle_child_output(r, &output);
+    }
+
     #[test]
     fn test_watchdog() {
         let focal = UbuntuDiskConfig::new(FOCAL_IMAGE_NAME.to_string());
@@ -6388,10 +6561,9 @@ mod sequential {
 #[cfg(target_arch = "x86_64")]
 mod windows {
     use crate::*;
+    use once_cell::sync::Lazy;
 
-    lazy_static! {
+    static NEXT_DISK_ID: Lazy<Mutex<u8>> = Lazy::new(|| Mutex::new(1));
-        static ref NEXT_DISK_ID: Mutex<u8> = Mutex::new(1);
-    }
 
     struct WindowsGuest {
         guest: Guest,
@@ -7619,9 +7791,9 @@ mod live_migration {
                 "--memory",
                 "size=0,hotplug_method=virtio-mem",
                 "--memory-zone",
-                "id=mem0,size=1G,hotplug_size=32G",
+                "id=mem0,size=1G,hotplug_size=4G",
-                "id=mem1,size=1G,hotplug_size=32G",
+                "id=mem1,size=1G,hotplug_size=4G",
-                "id=mem2,size=2G,hotplug_size=32G",
+                "id=mem2,size=2G,hotplug_size=4G",
                 "--numa",
                 "guest_numa_id=0,cpus=[0-2,9],distances=[1@15,2@20],memory_zones=mem0",
                 "guest_numa_id=1,cpus=[3-4,6-8],distances=[0@20,2@25],memory_zones=mem1",
@@ -7632,9 +7804,9 @@ mod live_migration {
                 "--memory",
                 "size=0,hotplug_method=virtio-mem,shared=on",
                 "--memory-zone",
-                "id=mem0,size=1G,hotplug_size=32G,shared=on",
+                "id=mem0,size=1G,hotplug_size=4G,shared=on",
-                "id=mem1,size=1G,hotplug_size=32G,shared=on",
+                "id=mem1,size=1G,hotplug_size=4G,shared=on",
-                "id=mem2,size=2G,hotplug_size=32G,shared=on",
+                "id=mem2,size=2G,hotplug_size=4G,shared=on",
                 "--numa",
                 "guest_numa_id=0,cpus=[0-2,9],distances=[1@15,2@20],memory_zones=mem0",
                 "guest_numa_id=1,cpus=[3-4,6-8],distances=[0@20,2@25],memory_zones=mem1",
@@ -8346,7 +8518,6 @@ mod aarch64_acpi {
 
                 assert_eq!(guest.get_cpu_count().unwrap_or_default(), 1);
                 assert!(guest.get_total_memory().unwrap_or_default() > 400_000);
-                assert!(guest.get_entropy().unwrap_or_default() >= 900);
                 assert_eq!(guest.get_pci_bridge_class().unwrap_or_default(), "0x060000");
             });
 

vfio_user/Cargo.toml

@@ -5,7 +5,7 @@ authors = ["The Cloud Hypervisor Authors"]
 edition = "2021"
 
 [dependencies]
-anyhow = "1.0.57"
+anyhow = "1.0.58"
 libc = "0.2.126"
 log = "0.4.17"
 serde = {version = ">=1.0.27", features = ["rc"] }

vfio_user/src/lib.rs

@@ -51,7 +51,7 @@ impl Default for Command {
 
 #[allow(dead_code)]
 #[repr(u32)]
-#[derive(Clone, Copy, Debug, PartialEq)]
+#[derive(Clone, Copy, Debug, PartialEq, Eq)]
 enum HeaderFlags {
     Command = 0,
     Reply = 1,

vhdx/Cargo.toml

@@ -12,5 +12,5 @@ libc = "0.2.126"
 log = "0.4.17"
 remain = "0.2.3"
 thiserror = "1.0"
-uuid = { version = "1.0.0", features = ["v4"] }
+uuid = { version = "1.1.2", features = ["v4"] }
 vmm-sys-util = ">=0.3.1"

vhost_user_block/Cargo.toml

@@ -6,7 +6,7 @@ edition = "2021"
 
 [dependencies]
 block_util = { path = "../block_util" }
-clap = { version = "3.1.18", features = ["wrap_help","cargo"] }
+clap = { version = "3.2.8", features = ["wrap_help","cargo"] }
 env_logger = "0.9.0"
 epoll = "4.3.1"
 libc = "0.2.126"
@@ -14,10 +14,10 @@ log = "0.4.17"
 option_parser = { path = "../option_parser" }
 qcow = { path = "../qcow" }
 vhost = { version = "0.4.0", features = ["vhost-user-slave"] }
-vhost-user-backend = { git = "https://github.com/rust-vmm/vhost-user-backend", rev = "14f58eda14076e973704d4f904850be1146fbb05" }
+vhost-user-backend = "0.5.0"
 virtio-bindings = "0.1.0"
 vm-memory = "0.8.0"
 vmm-sys-util = "0.9.0"
 
 [build-dependencies]
-clap = { version = "3.1.18", features = ["cargo"] }
+clap = { version = "3.2.8", features = ["cargo"] }

vhost_user_block/src/lib.rs

@@ -8,6 +8,8 @@
 //
 // SPDX-License-Identifier: (Apache-2.0 AND BSD-3-Clause)
 
+#![allow(clippy::significant_drop_in_scrutinee)]
+
 use block_util::{build_disk_image_id, Request, VirtioBlockConfig};
 use libc::EFD_NONBLOCK;
 use log::*;
@@ -45,8 +47,8 @@ const BLK_SIZE: u32 = 512;
 // and the overhead of the emulation layer.
 const POLL_QUEUE_US: u128 = 50;
 
-trait DiskFile: Read + Seek + Write + Send + Sync {}
+trait DiskFile: Read + Seek + Write + Send {}
-impl<D: Read + Seek + Write + Send + Sync> DiskFile for D {}
+impl<D: Read + Seek + Write + Send> DiskFile for D {}
 
 type Result<T> = std::result::Result<T, Error>;
 type VhostUserBackendResult<T> = std::result::Result<T, std::io::Error>;

vhost_user_net/Cargo.toml

@@ -5,7 +5,7 @@ authors = ["The Cloud Hypervisor Authors"]
 edition = "2021"
 
 [dependencies]
-clap = { version = "3.1.18", features = ["wrap_help","cargo"] }
+clap = { version = "3.2.8", features = ["wrap_help","cargo"] }
 env_logger = "0.9.0"
 epoll = "4.3.1"
 libc = "0.2.126"
@@ -13,10 +13,10 @@ log = "0.4.17"
 net_util = { path = "../net_util" }
 option_parser = { path = "../option_parser" }
 vhost = { version = "0.4.0", features = ["vhost-user-slave"] }
-vhost-user-backend = { git = "https://github.com/rust-vmm/vhost-user-backend", rev = "14f58eda14076e973704d4f904850be1146fbb05" }
+vhost-user-backend = "0.5.0"
 virtio-bindings = "0.1.0"
 vm-memory = "0.8.0"
 vmm-sys-util = "0.9.0"
 
 [build-dependencies]
-clap = { version = "3.1.18", features = ["cargo"] }
+clap = { version = "3.2.8", features = ["cargo"] }

vhost_user_net/src/lib.rs

@@ -6,6 +6,8 @@
 //
 // SPDX-License-Identifier: (Apache-2.0 AND BSD-3-Clause)
 
+#![allow(clippy::significant_drop_in_scrutinee)]
+
 use libc::{self, EFD_NONBLOCK};
 use log::*;
 use net_util::{

virtio-devices/Cargo.toml

@@ -9,7 +9,7 @@ default = []
 mshv = []
 
 [dependencies]
-anyhow = "1.0.57"
+anyhow = "1.0.58"
 arc-swap = "1.5.0"
 block_util = { path = "../block_util" }
 byteorder = "1.4.3"
@@ -23,14 +23,14 @@ net_util = { path = "../net_util" }
 pci = { path = "../pci" }
 rate_limiter = { path = "../rate_limiter" }
 seccompiler = "0.2.0"
-serde = { version="1.0.137", features=["derive"] }
+serde = { version="1.0.138", features=["derive"] }
-serde_json = "1.0.81"
+serde_json = "1.0.82"
 thiserror = "1.0.31"
 versionize = "0.1.6"
 versionize_derive = "0.1.4"
 vhost = { version = "0.4.0", features = ["vhost-user-master", "vhost-user-slave", "vhost-kern", "vhost-vdpa"] }
 virtio-bindings = { version = "0.1.0", features = ["virtio-v5_0_0"] }
-virtio-queue = "0.3.0"
+virtio-queue = "0.4.0"
 vm-allocator = { path = "../vm-allocator" }
 vm-device = { path = "../vm-device" }
 vm-memory = { version = "0.8.0", features = ["backend-mmap", "backend-atomic", "backend-bitmap"] }

virtio-devices/src/device.rs

@@ -206,7 +206,7 @@ pub trait VirtioDevice: Send {
 /// address translation before they try to read from the guest physical address.
 /// On the other side, the implementation itself should be provided by the code
 /// emulating the IOMMU for the guest.
-pub trait DmaRemapping: Send + Sync {
+pub trait DmaRemapping {
     /// Provide a way to translate GVA address ranges into GPAs.
     fn translate_gva(&self, id: u32, addr: u64) -> std::result::Result<u64, std::io::Error>;
     /// Provide a way to translate GPA address ranges into GVAs.

virtio-devices/src/lib.rs

@@ -10,6 +10,8 @@
 
 //! Implements virtio devices, queues, and transport mechanisms.
 
+#![allow(clippy::significant_drop_in_scrutinee)]
+
 #[macro_use]
 extern crate event_monitor;
 #[macro_use]
@@ -123,14 +125,14 @@ pub enum Error {
     QueueIterator(virtio_queue::Error),
 }
 
-#[derive(Clone, Copy, Debug, Default, Deserialize, Serialize, PartialEq)]
+#[derive(Clone, Copy, Debug, Default, Deserialize, Serialize, PartialEq, Eq)]
 pub struct TokenBucketConfig {
     pub size: u64,
     pub one_time_burst: Option<u64>,
     pub refill_time: u64,
 }
 
-#[derive(Clone, Copy, Debug, Default, Deserialize, Serialize, PartialEq)]
+#[derive(Clone, Copy, Debug, Default, Deserialize, Serialize, PartialEq, Eq)]
 #[serde(deny_unknown_fields)]
 pub struct RateLimiterConfig {
     pub bandwidth: Option<TokenBucketConfig>,

virtio-devices/src/pmem.rs

@@ -106,7 +106,7 @@ impl Display for Error {
     }
 }
 
-#[derive(Debug, PartialEq)]
+#[derive(Debug, PartialEq, Eq)]
 enum RequestType {
     Flush,
 }

virtio-devices/src/transport/mod.rs

@@ -6,7 +6,7 @@ use vmm_sys_util::eventfd::EventFd;
 mod pci_common_config;
 mod pci_device;
 pub use pci_common_config::VirtioPciCommonConfig;
-pub use pci_device::VirtioPciDevice;
+pub use pci_device::{VirtioPciDevice, VirtioPciDeviceActivator};
 
 pub trait VirtioTransport {
     fn ioeventfds(&self, base_addr: u64) -> Vec<(&EventFd, u64)>;

virtio-devices/src/transport/pci_device.rs

@@ -287,6 +287,37 @@ struct VirtioPciDeviceState {
 
 impl VersionMapped for VirtioPciDeviceState {}
 
+pub struct VirtioPciDeviceActivator {
+    interrupt: Option<Arc<dyn VirtioInterrupt>>,
+    memory: Option<GuestMemoryAtomic<GuestMemoryMmap>>,
+    device: Arc<Mutex<dyn VirtioDevice>>,
+    device_activated: Arc<AtomicBool>,
+    queues: Option<Vec<Queue<GuestMemoryAtomic<GuestMemoryMmap>>>>,
+    queue_evts: Option<Vec<EventFd>>,
+    barrier: Option<Arc<Barrier>>,
+    id: String,
+}
+
+impl VirtioPciDeviceActivator {
+    pub fn activate(&mut self) -> ActivateResult {
+        self.device.lock().unwrap().activate(
+            self.memory.take().unwrap(),
+            self.interrupt.take().unwrap(),
+            self.queues.take().unwrap(),
+            self.queue_evts.take().unwrap(),
+        )?;
+        self.device_activated.store(true, Ordering::SeqCst);
+
+        if let Some(barrier) = self.barrier.take() {
+            info!("{}: Waiting for barrier", self.id);
+            barrier.wait();
+            info!("{}: Barrier released", self.id);
+        }
+
+        Ok(())
+    }
+}
+
 pub struct VirtioPciDevice {
     id: String,
 
@@ -338,11 +369,11 @@ pub struct VirtioPciDevice {
     // EventFd to signal on to request activation
     activate_evt: EventFd,
 
-    // Barrier that is used to wait on for activation
-    activate_barrier: Arc<Barrier>,
-
     // Optional DMA handler
     dma_handler: Option<Arc<dyn ExternalDmaMapping>>,
+
+    // Pending activations
+    pending_activations: Arc<Mutex<Vec<VirtioPciDeviceActivator>>>,
 }
 
 impl VirtioPciDevice {
@@ -359,6 +390,7 @@ impl VirtioPciDevice {
         activate_evt: EventFd,
         use_64bit_bar: bool,
         dma_handler: Option<Arc<dyn ExternalDmaMapping>>,
+        pending_activations: Arc<Mutex<Vec<VirtioPciDeviceActivator>>>,
     ) -> Result<Self> {
         let device_clone = device.clone();
         let mut locked_device = device_clone.lock().unwrap();
@@ -458,8 +490,8 @@ impl VirtioPciDevice {
             cap_pci_cfg_info: VirtioPciCfgCapInfo::default(),
             bar_regions: vec![],
             activate_evt,
-            activate_barrier: Arc::new(Barrier::new(2)),
             dma_handler,
+            pending_activations,
         };
 
         if let Some(msix_config) = &virtio_pci_device.msix_config {
@@ -666,37 +698,37 @@ impl VirtioPciDevice {
         self.device.clone()
     }
 
-    fn activate(&mut self) -> ActivateResult {
+    fn prepare_activator(&mut self, barrier: Option<Arc<Barrier>>) -> VirtioPciDeviceActivator {
-        if let Some(virtio_interrupt) = self.virtio_interrupt.take() {
+        let mut queue_evts = Vec::new();
-            if self.memory.is_some() {
+        let mut queues: Vec<Queue<GuestMemoryAtomic<GuestMemoryMmap>>> =
-                let mem = self.memory.as_ref().unwrap().clone();
+            self.queues.iter().map(vm_virtio::clone_queue).collect();
-                let mut device = self.device.lock().unwrap();
+        queues.retain(|q| q.state.ready);
-                let mut queue_evts = Vec::new();
+        for (i, queue) in queues.iter().enumerate() {
-                let mut queues: Vec<Queue<GuestMemoryAtomic<GuestMemoryMmap>>> =
+            queue_evts.push(self.queue_evts[i].try_clone().unwrap());
-                    self.queues.iter().map(vm_virtio::clone_queue).collect();
+            if !queue.is_valid() {
-                queues.retain(|q| q.state.ready);
+                error!("Queue {} is not valid", i);
-                for (i, queue) in queues.iter().enumerate() {
-                    queue_evts.push(self.queue_evts[i].try_clone().unwrap());
-                    if !queue.is_valid() {
-                        error!("Queue {} is not valid", i);
-                    }
-                }
-                return device.activate(mem, virtio_interrupt, queues, queue_evts);
             }
         }
-        Ok(())
+
+        VirtioPciDeviceActivator {
+            interrupt: self.virtio_interrupt.take(),
+            memory: self.memory.clone(),
+            device: self.device.clone(),
+            queues: Some(queues),
+            device_activated: self.device_activated.clone(),
+            queue_evts: Some(
+                queue_evts
+                    .iter()
+                    .map(|fd| fd.try_clone().unwrap())
+                    .collect(),
+            ),
+            barrier,
+            id: self.id.clone(),
+        }
     }
 
-    pub fn maybe_activate(&mut self) {
+    fn activate(&mut self) -> ActivateResult {
-        if self.needs_activation() {
+        self.prepare_activator(None).activate()
-            self.activate().expect("Failed to activate device");
-            self.device_activated.store(true, Ordering::SeqCst);
-            info!("{}: Waiting for barrier", self.id);
-            self.activate_barrier.wait();
-            info!("{}: Barrier released", self.id);
-        } else {
-            info!("{}: Device does not need activation", self.id)
-        }
     }
 
     fn needs_activation(&self) -> bool {
@@ -1020,7 +1052,7 @@ impl PciDevice for VirtioPciDevice {
                 self.device.clone(),
             ),
             o if (ISR_CONFIG_BAR_OFFSET..ISR_CONFIG_BAR_OFFSET + ISR_CONFIG_SIZE).contains(&o) => {
-                if let Some(v) = data.get(0) {
+                if let Some(v) = data.first() {
                     self.interrupt_status
                         .fetch_and(!(*v as usize), Ordering::AcqRel);
                 }
@@ -1058,13 +1090,16 @@ impl PciDevice for VirtioPciDevice {
 
         // Try and activate the device if the driver status has changed
         if self.needs_activation() {
+            let barrier = Arc::new(Barrier::new(2));
+            let activator = self.prepare_activator(Some(barrier.clone()));
+            self.pending_activations.lock().unwrap().push(activator);
             info!(
                 "{}: Needs activation; writing to activate event fd",
                 self.id
             );
             self.activate_evt.write(1).ok();
             info!("{}: Needs activation; returning barrier", self.id);
-            return Some(self.activate_barrier.clone());
+            return Some(barrier);
         }
 
         // Device has been reset by the driver

virtio-devices/src/vhost_user/vu_common_ctrl.rs

@@ -173,7 +173,7 @@ impl VhostUserHandle {
         // at early stage.
         for (queue_index, queue) in queues.iter().enumerate() {
             self.vu
-                .set_vring_num(queue_index, queue.max_size())
+                .set_vring_num(queue_index, queue.state.size)
                 .map_err(Error::VhostUserSetVringNum)?;
         }
 
@@ -184,7 +184,7 @@ impl VhostUserHandle {
                     mmap_size: 0,
                     mmap_offset: 0,
                     num_queues: queues.len() as u16,
-                    queue_size: queues[0].max_size(),
+                    queue_size: queues[0].state.size,
                 };
                 let (info, fd) = self
                     .vu
@@ -203,11 +203,11 @@ impl VhostUserHandle {
 
         let mut vrings_info = Vec::new();
         for (queue_index, queue) in queues.into_iter().enumerate() {
-            let actual_size: usize = queue.max_size().try_into().unwrap();
+            let actual_size: usize = queue.state.size.try_into().unwrap();
 
             let config_data = VringConfigData {
                 queue_max_size: queue.max_size(),
-                queue_size: queue.max_size(),
+                queue_size: queue.state.size,
                 flags: 0u32,
                 desc_table_addr: get_host_address_range(
                     mem,
@@ -290,10 +290,7 @@ impl VhostUserHandle {
     }
 
     pub fn reset_vhost_user(&mut self, num_queues: usize) -> Result<()> {
-        self.enable_vhost_user_vrings(num_queues, false)?;
+        self.enable_vhost_user_vrings(num_queues, false)
-
-        // Reset the owner.
-        self.vu.reset_owner().map_err(Error::VhostUserResetOwner)
     }
 
     pub fn set_protocol_features_vhost_user(

virtio-devices/src/vsock/csm/mod.rs

@@ -38,7 +38,7 @@ type Result<T> = std::result::Result<T, Error>;
 
 /// A vsock connection state.
 ///
-#[derive(Clone, Copy, Debug, PartialEq)]
+#[derive(Clone, Copy, Debug, PartialEq, Eq)]
 pub enum ConnState {
     /// The connection has been initiated by the host end, but is yet to be confirmed by the guest.
     LocalInit,
@@ -61,7 +61,7 @@ pub enum ConnState {
 /// (via `notify()`), the connection will store a `PendingRx::Rw` to be later inspected by
 /// `recv_pkt()`.
 ///
-#[derive(Clone, Copy, PartialEq)]
+#[derive(Clone, Copy, PartialEq, Eq)]
 enum PendingRx {
     /// We need to yield a connection request packet (VSOCK_OP_REQUEST).
     Request = 0,

vm-device/Cargo.toml

@@ -10,11 +10,11 @@ kvm = ["vfio-ioctls/kvm"]
 mshv = ["vfio-ioctls/mshv"]
 
 [dependencies]
-anyhow = "1.0.57"
+anyhow = "1.0.58"
 hypervisor = { path = "../hypervisor" }
 thiserror = "1.0.31"
-serde = { version = "1.0.137", features = ["rc", "derive"] }
+serde = { version = "1.0.138", features = ["rc", "derive"] }
-serde_json = "1.0.81"
+serde_json = "1.0.82"
 vfio-ioctls = { git = "https://github.com/rust-vmm/vfio", branch = "main", default-features = false }
 vm-memory = { version = "0.8.0", features = ["backend-mmap"] }
 vmm-sys-util = "0.9.0"

vm-device/src/bus.rs

@@ -7,7 +7,7 @@
 
 //! Handles routing to devices in an address space.
 
-use std::cmp::{Ord, Ordering, PartialEq, PartialOrd};
+use std::cmp::{Eq, Ord, Ordering, PartialEq, PartialOrd};
 use std::collections::btree_map::BTreeMap;
 use std::sync::{Arc, Barrier, Mutex, RwLock, Weak};
 use std::{convert, error, fmt, io, result};

vm-device/src/lib.rs

@@ -3,6 +3,8 @@
 // SPDX-License-Identifier: Apache-2.0
 //
 
+#![allow(clippy::significant_drop_in_scrutinee)]
+
 use serde::{Deserialize, Serialize};
 
 mod bus;
@@ -12,7 +14,7 @@ pub mod interrupt;
 pub use self::bus::{Bus, BusDevice, Error as BusError};
 
 /// Type of Message Signalled Interrupt
-#[derive(Copy, Clone, Debug, PartialEq, Serialize, Deserialize)]
+#[derive(Copy, Clone, Debug, PartialEq, Eq, Serialize, Deserialize)]
 pub enum MsiIrqType {
     /// PCI MSI IRQ numbers.
     PciMsi,
@@ -22,7 +24,7 @@ pub enum MsiIrqType {
     GenericMsi,
 }
 
-#[derive(Copy, Clone, PartialEq, Serialize, Deserialize, Debug)]
+#[derive(Copy, Clone, PartialEq, Eq, Serialize, Deserialize, Debug)]
 pub enum PciBarType {
     Io,
     Mmio32,

vm-migration/Cargo.toml

@@ -5,10 +5,10 @@ authors = ["The Cloud Hypervisor Authors"]
 edition = "2021"
 
 [dependencies]
-anyhow = "1.0.57"
+anyhow = "1.0.58"
 thiserror = "1.0.31"
-serde = { version = "1.0.137", features = ["rc", "derive"] }
+serde = { version = "1.0.138", features = ["rc", "derive"] }
-serde_json = "1.0.81"
+serde_json = "1.0.82"
 versionize = "0.1.6"
 versionize_derive = "0.1.4"
 vm-memory = { version = "0.8.0", features = ["backend-mmap", "backend-atomic"] }

vm-migration/src/lib.rs

@@ -12,7 +12,7 @@ use versionize::{VersionMap, Versionize};
 pub mod protocol;
 
 /// Global VMM version for versioning
-const MAJOR_VERSION: u16 = 24;
+const MAJOR_VERSION: u16 = 25;
 const MINOR_VERSION: u16 = 0;
 const VMM_VERSION: u16 = MAJOR_VERSION << 12 | MINOR_VERSION & 0b1111;
 

vm-migration/src/protocol.rs

@@ -139,7 +139,7 @@ impl Request {
 }
 
 #[repr(u16)]
-#[derive(Copy, Clone, PartialEq)]
+#[derive(Copy, Clone, PartialEq, Eq)]
 pub enum Status {
     Invalid,
     Ok,

vm-virtio/Cargo.toml

@@ -10,5 +10,5 @@ default = []
 [dependencies]
 log = "0.4.17"
 virtio-bindings = { version = "0.1.0", features = ["virtio-v5_0_0"] }
-virtio-queue = "0.3.0"
+virtio-queue = "0.4.0"
 vm-memory = { version = "0.8.0", features = ["backend-mmap", "backend-atomic", "backend-bitmap"] }

vmm/Cargo.toml

@@ -9,47 +9,48 @@ default = []
 amx = []
 cmos = ["devices/cmos"]
 fwdebug = ["devices/fwdebug"]
-gdb = ["kvm"]
+gdb = ["kvm", "gdbstub", "gdbstub_arch"]
+guest_debug = ["kvm"]
 kvm = ["hypervisor/kvm", "vfio-ioctls/kvm", "vm-device/kvm", "pci/kvm"]
 mshv = ["hypervisor/mshv", "virtio-devices/mshv", "vfio-ioctls/mshv", "vm-device/mshv", "pci/mshv"]
 tdx = ["arch/tdx", "hypervisor/tdx"]
 
 [dependencies]
 acpi_tables = { path = "../acpi_tables" }
-anyhow = "1.0.57"
+anyhow = "1.0.58"
 arc-swap = "1.5.0"
 arch = { path = "../arch" }
 bitflags = "1.3.2"
 block_util = { path = "../block_util" }
-clap = "3.1.18"
+clap = "3.2.8"
 devices = { path = "../devices" }
 epoll = "4.3.1"
 event_monitor = { path = "../event_monitor" }
-gdbstub = "0.6.1"
+gdbstub = { version = "0.6.2", optional = true }
-gdbstub_arch = "0.2.2"
+gdbstub_arch = { version = "0.2.3", optional = true }
 hypervisor = { path = "../hypervisor" }
-lazy_static = "1.4.0"
 libc = "0.2.126"
 linux-loader = { version = "0.4.0", features = ["elf", "bzimage", "pe"] }
 log = "0.4.17"
 micro_http = { git = "https://github.com/firecracker-microvm/micro-http", branch = "main" }
 net_util = { path = "../net_util" }
+once_cell = "1.13.0"
 option_parser = { path = "../option_parser" }
 pci = { path = "../pci" }
 qcow = { path = "../qcow" }
 seccompiler = "0.2.0"
-serde = { version = "1.0.137", features = ["rc", "derive"] }
+serde = { version = "1.0.138", features = ["rc", "derive"] }
-serde_json = "1.0.81"
+serde_json = "1.0.82"
 signal-hook = "0.3.14"
 thiserror = "1.0.31"
-uuid = "1.0.0"
+uuid = "1.1.2"
 versionize = "0.1.6"
 versionize_derive = "0.1.4"
 vfio-ioctls = { git = "https://github.com/rust-vmm/vfio", branch = "main", default-features = false }
 vfio_user = { path = "../vfio_user" }
 vhdx = { path = "../vhdx" }
 virtio-devices = { path = "../virtio-devices" }
-virtio-queue = "0.3.0"
+virtio-queue = "0.4.0"
 vm-allocator = { path = "../vm-allocator" }
 vm-device = { path = "../vm-device" }
 vm-memory = { version = "0.8.0", features = ["backend-mmap", "backend-atomic", "backend-bitmap"] }

vmm/src/api/http.rs

@@ -8,6 +8,7 @@ use crate::api::{ApiError, ApiRequest, VmAction};
 use crate::seccomp_filters::{get_seccomp_filter, Thread};
 use crate::{Error as VmmError, Result};
 use micro_http::{Body, HttpServer, MediaType, Method, Request, Response, StatusCode, Version};
+use once_cell::sync::Lazy;
 use seccompiler::{apply_filter, SeccompAction};
 use serde_json::Error as SerdeError;
 use std::collections::HashMap;
@@ -56,7 +57,7 @@ pub fn error_response(error: HttpError, status: StatusCode) -> Response {
 }
 
 /// An HTTP endpoint handler interface
-pub trait EndpointHandler: Sync + Send {
+pub trait EndpointHandler {
     /// Handles an HTTP request.
     /// After parsing the request, the handler could decide to send an
     /// associated API request down to the VMM API server to e.g. create
@@ -129,44 +130,125 @@ macro_rules! endpoint {
     };
 }
 
-lazy_static! {
+/// HTTP_ROUTES contain all the cloud-hypervisor HTTP routes.
-    /// HTTP_ROUTES contain all the cloud-hypervisor HTTP routes.
+pub static HTTP_ROUTES: Lazy<HttpRoutes> = Lazy::new(|| {
-    pub static ref HTTP_ROUTES: HttpRoutes = {
+    let mut r = HttpRoutes {
-        let mut r = HttpRoutes {
+        routes: HashMap::new(),
-            routes: HashMap::new(),
-        };
-
-        r.routes.insert(endpoint!("/vm.add-device"), Box::new(VmActionHandler::new(VmAction::AddDevice(Arc::default()))));
-        r.routes.insert(endpoint!("/vm.add-user-device"), Box::new(VmActionHandler::new(VmAction::AddUserDevice(Arc::default()))));
-        r.routes.insert(endpoint!("/vm.add-disk"), Box::new(VmActionHandler::new(VmAction::AddDisk(Arc::default()))));
-        r.routes.insert(endpoint!("/vm.add-fs"), Box::new(VmActionHandler::new(VmAction::AddFs(Arc::default()))));
-        r.routes.insert(endpoint!("/vm.add-net"), Box::new(VmActionHandler::new(VmAction::AddNet(Arc::default()))));
-        r.routes.insert(endpoint!("/vm.add-pmem"), Box::new(VmActionHandler::new(VmAction::AddPmem(Arc::default()))));
-        r.routes.insert(endpoint!("/vm.add-vdpa"), Box::new(VmActionHandler::new(VmAction::AddVdpa(Arc::default()))));
-        r.routes.insert(endpoint!("/vm.add-vsock"), Box::new(VmActionHandler::new(VmAction::AddVsock(Arc::default()))));
-        r.routes.insert(endpoint!("/vm.boot"), Box::new(VmActionHandler::new(VmAction::Boot)));
-        r.routes.insert(endpoint!("/vm.counters"), Box::new(VmActionHandler::new(VmAction::Counters)));
-        r.routes.insert(endpoint!("/vm.create"), Box::new(VmCreate {}));
-        r.routes.insert(endpoint!("/vm.delete"), Box::new(VmActionHandler::new(VmAction::Delete)));
-        r.routes.insert(endpoint!("/vm.info"), Box::new(VmInfo {}));
-        r.routes.insert(endpoint!("/vm.pause"), Box::new(VmActionHandler::new(VmAction::Pause)));
-        r.routes.insert(endpoint!("/vm.power-button"), Box::new(VmActionHandler::new(VmAction::PowerButton)));
-        r.routes.insert(endpoint!("/vm.reboot"), Box::new(VmActionHandler::new(VmAction::Reboot)));
-        r.routes.insert(endpoint!("/vm.receive-migration"), Box::new(VmActionHandler::new(VmAction::ReceiveMigration(Arc::default()))));
-        r.routes.insert(endpoint!("/vm.remove-device"), Box::new(VmActionHandler::new(VmAction::RemoveDevice(Arc::default()))));
-        r.routes.insert(endpoint!("/vm.resize"), Box::new(VmActionHandler::new(VmAction::Resize(Arc::default()))));
-        r.routes.insert(endpoint!("/vm.resize-zone"), Box::new(VmActionHandler::new(VmAction::ResizeZone(Arc::default()))));
-        r.routes.insert(endpoint!("/vm.restore"), Box::new(VmActionHandler::new(VmAction::Restore(Arc::default()))));
-        r.routes.insert(endpoint!("/vm.resume"), Box::new(VmActionHandler::new(VmAction::Resume)));
-        r.routes.insert(endpoint!("/vm.send-migration"), Box::new(VmActionHandler::new(VmAction::SendMigration(Arc::default()))));
-        r.routes.insert(endpoint!("/vm.shutdown"), Box::new(VmActionHandler::new(VmAction::Shutdown)));
-        r.routes.insert(endpoint!("/vm.snapshot"), Box::new(VmActionHandler::new(VmAction::Snapshot(Arc::default()))));
-        r.routes.insert(endpoint!("/vmm.ping"), Box::new(VmmPing {}));
-        r.routes.insert(endpoint!("/vmm.shutdown"), Box::new(VmmShutdown {}));
-
-        r
     };
-}
+
+    r.routes.insert(
+        endpoint!("/vm.add-device"),
+        Box::new(VmActionHandler::new(VmAction::AddDevice(Arc::default()))),
+    );
+    r.routes.insert(
+        endpoint!("/vm.add-user-device"),
+        Box::new(VmActionHandler::new(
+            VmAction::AddUserDevice(Arc::default()),
+        )),
+    );
+    r.routes.insert(
+        endpoint!("/vm.add-disk"),
+        Box::new(VmActionHandler::new(VmAction::AddDisk(Arc::default()))),
+    );
+    r.routes.insert(
+        endpoint!("/vm.add-fs"),
+        Box::new(VmActionHandler::new(VmAction::AddFs(Arc::default()))),
+    );
+    r.routes.insert(
+        endpoint!("/vm.add-net"),
+        Box::new(VmActionHandler::new(VmAction::AddNet(Arc::default()))),
+    );
+    r.routes.insert(
+        endpoint!("/vm.add-pmem"),
+        Box::new(VmActionHandler::new(VmAction::AddPmem(Arc::default()))),
+    );
+    r.routes.insert(
+        endpoint!("/vm.add-vdpa"),
+        Box::new(VmActionHandler::new(VmAction::AddVdpa(Arc::default()))),
+    );
+    r.routes.insert(
+        endpoint!("/vm.add-vsock"),
+        Box::new(VmActionHandler::new(VmAction::AddVsock(Arc::default()))),
+    );
+    r.routes.insert(
+        endpoint!("/vm.boot"),
+        Box::new(VmActionHandler::new(VmAction::Boot)),
+    );
+    r.routes.insert(
+        endpoint!("/vm.counters"),
+        Box::new(VmActionHandler::new(VmAction::Counters)),
+    );
+    r.routes
+        .insert(endpoint!("/vm.create"), Box::new(VmCreate {}));
+    r.routes.insert(
+        endpoint!("/vm.delete"),
+        Box::new(VmActionHandler::new(VmAction::Delete)),
+    );
+    r.routes.insert(endpoint!("/vm.info"), Box::new(VmInfo {}));
+    r.routes.insert(
+        endpoint!("/vm.pause"),
+        Box::new(VmActionHandler::new(VmAction::Pause)),
+    );
+    r.routes.insert(
+        endpoint!("/vm.power-button"),
+        Box::new(VmActionHandler::new(VmAction::PowerButton)),
+    );
+    r.routes.insert(
+        endpoint!("/vm.reboot"),
+        Box::new(VmActionHandler::new(VmAction::Reboot)),
+    );
+    r.routes.insert(
+        endpoint!("/vm.receive-migration"),
+        Box::new(VmActionHandler::new(VmAction::ReceiveMigration(
+            Arc::default(),
+        ))),
+    );
+    r.routes.insert(
+        endpoint!("/vm.remove-device"),
+        Box::new(VmActionHandler::new(VmAction::RemoveDevice(Arc::default()))),
+    );
+    r.routes.insert(
+        endpoint!("/vm.resize"),
+        Box::new(VmActionHandler::new(VmAction::Resize(Arc::default()))),
+    );
+    r.routes.insert(
+        endpoint!("/vm.resize-zone"),
+        Box::new(VmActionHandler::new(VmAction::ResizeZone(Arc::default()))),
+    );
+    r.routes.insert(
+        endpoint!("/vm.restore"),
+        Box::new(VmActionHandler::new(VmAction::Restore(Arc::default()))),
+    );
+    r.routes.insert(
+        endpoint!("/vm.resume"),
+        Box::new(VmActionHandler::new(VmAction::Resume)),
+    );
+    r.routes.insert(
+        endpoint!("/vm.send-migration"),
+        Box::new(VmActionHandler::new(
+            VmAction::SendMigration(Arc::default()),
+        )),
+    );
+    r.routes.insert(
+        endpoint!("/vm.shutdown"),
+        Box::new(VmActionHandler::new(VmAction::Shutdown)),
+    );
+    r.routes.insert(
+        endpoint!("/vm.snapshot"),
+        Box::new(VmActionHandler::new(VmAction::Snapshot(Arc::default()))),
+    );
+    #[cfg(feature = "guest_debug")]
+    r.routes.insert(
+        endpoint!("/vm.coredump"),
+        Box::new(VmActionHandler::new(VmAction::Coredump(Arc::default()))),
+    );
+    r.routes
+        .insert(endpoint!("/vmm.ping"), Box::new(VmmPing {}));
+    r.routes
+        .insert(endpoint!("/vmm.shutdown"), Box::new(VmmShutdown {}));
+
+    r
+});
 
 fn handle_http_request(
     request: &Request,

vmm/src/api/http_endpoint.rs

@@ -4,6 +4,8 @@
 //
 
 use crate::api::http::{error_response, EndpointHandler, HttpError};
+#[cfg(feature = "guest_debug")]
+use crate::api::vm_coredump;
 use crate::api::{
     vm_add_device, vm_add_disk, vm_add_fs, vm_add_net, vm_add_pmem, vm_add_user_device,
     vm_add_vdpa, vm_add_vsock, vm_boot, vm_counters, vm_create, vm_delete, vm_info, vm_pause,
@@ -151,6 +153,12 @@ impl EndpointHandler for VmActionHandler {
                     api_sender,
                     Arc::new(serde_json::from_slice(body.raw())?),
                 ),
+                #[cfg(feature = "guest_debug")]
+                Coredump(_) => vm_coredump(
+                    api_notifier,
+                    api_sender,
+                    Arc::new(serde_json::from_slice(body.raw())?),
+                ),
                 ReceiveMigration(_) => vm_receive_migration(
                     api_notifier,
                     api_sender,

vmm/src/api/mod.rs

@@ -99,6 +99,9 @@ pub enum ApiError {
     /// The VM could not restored.
     VmRestore(VmError),
 
+    /// The VM could not be coredumped.
+    VmCoredump(VmError),
+
     /// The VMM could not shutdown.
     VmmShutdown(VmError),
 
@@ -189,6 +192,12 @@ pub struct VmSnapshotConfig {
     pub destination_url: String,
 }
 
+#[derive(Clone, Deserialize, Serialize, Default, Debug)]
+pub struct VmCoredumpData {
+    /// The coredump destination file
+    pub destination_url: String,
+}
+
 #[derive(Clone, Deserialize, Serialize, Default, Debug)]
 pub struct VmReceiveMigrationData {
     /// URL for the reception of migration state
@@ -310,6 +319,10 @@ pub enum ApiRequest {
     /// Restore from a VM snapshot
     VmRestore(Arc<RestoreConfig>, Sender<ApiResponse>),
 
+    /// Take a VM coredump
+    #[cfg(feature = "guest_debug")]
+    VmCoredump(Arc<VmCoredumpData>, Sender<ApiResponse>),
+
     /// Incoming migration
     VmReceiveMigration(Arc<VmReceiveMigrationData>, Sender<ApiResponse>),
 
@@ -402,6 +415,10 @@ pub enum VmAction {
     /// Snapshot VM
     Snapshot(Arc<VmSnapshotConfig>),
 
+    /// Coredump VM
+    #[cfg(feature = "guest_debug")]
+    Coredump(Arc<VmCoredumpData>),
+
     /// Incoming migration
     ReceiveMigration(Arc<VmReceiveMigrationData>),
 
@@ -441,6 +458,8 @@ fn vm_action(
         ResizeZone(v) => ApiRequest::VmResizeZone(v, response_sender),
         Restore(v) => ApiRequest::VmRestore(v, response_sender),
         Snapshot(v) => ApiRequest::VmSnapshot(v, response_sender),
+        #[cfg(feature = "guest_debug")]
+        Coredump(v) => ApiRequest::VmCoredump(v, response_sender),
         ReceiveMigration(v) => ApiRequest::VmReceiveMigration(v, response_sender),
         SendMigration(v) => ApiRequest::VmSendMigration(v, response_sender),
         PowerButton => ApiRequest::VmPowerButton(response_sender),
@@ -526,6 +545,15 @@ pub fn vm_restore(
     vm_action(api_evt, api_sender, VmAction::Restore(data))
 }
 
+#[cfg(feature = "guest_debug")]
+pub fn vm_coredump(
+    api_evt: EventFd,
+    api_sender: Sender<ApiRequest>,
+    data: Arc<VmCoredumpData>,
+) -> ApiResult<Option<Body>> {
+    vm_action(api_evt, api_sender, VmAction::Coredump(data))
+}
+
 pub fn vm_info(api_evt: EventFd, api_sender: Sender<ApiRequest>) -> ApiResult<VmInfo> {
     let (response_sender, response_receiver) = channel();
 

vmm/src/api/openapi/cloud-hypervisor.yaml

@@ -366,6 +366,24 @@ paths:
         405:
           description: The VM instance could not be snapshotted because it is not booted.
 
+  /vm.coredump:
+    put:
+      summary: Takes a VM coredump.
+      requestBody:
+        description: The coredump configuration
+        content:
+          application/json:
+            schema:
+              $ref: '#/components/schemas/VmCoredumpData'
+        required: true
+      responses:
+        204:
+          description: The VM instance was successfully coredumped.
+        404:
+          description: The VM instance could not be coredumped because it is not created.
+        405:
+          description: The VM instance could not be coredumped because it is not booted.
+
   /vm.restore:
     put:
       summary: Restore a VM from a snapshot.
@@ -596,6 +614,9 @@ components:
           type: integer
         topology:
           $ref: '#/components/schemas/CpuTopology'
+        kvm_hyperv:
+          type: boolean
+          default: false
         max_phys_bits:
           type: integer
         affinity:
@@ -870,8 +891,6 @@ components:
 
     FsConfig:
       required:
-      - cache_size
-      - dax
       - num_queues
       - queue_size
       - socket
@@ -888,13 +907,6 @@ components:
         queue_size:
           type: integer
           default: 1024
-        dax:
-          type: boolean
-          default: true
-        cache_size:
-          type: integer
-          format: int64
-          default: 8589934592
         pci_segment:
           type: integer
           format: int16
@@ -914,9 +926,6 @@ components:
         iommu:
           type: boolean
           default: false
-        mergeable:
-          type: boolean
-          default: false
         discard_writes:
           type: boolean
           default: false
@@ -1110,6 +1119,12 @@ components:
         destination_url:
           type: string
 
+    VmCoredumpData:
+      type: object
+      properties:
+        destination_url:
+          type: string
+
     RestoreConfig:
       required:
       - source_url

vmm/src/config.rs

@@ -43,8 +43,6 @@ pub enum Error {
     ParseFsTagMissing,
     /// Filesystem socket is missing
     ParseFsSockMissing,
-    /// Cannot have dax=off along with cache_size parameter.
-    InvalidCacheSizeWithDaxOff,
     /// Missing persistent memory file parameter.
     ParsePmemFileMissing,
     /// Missing vsock socket path parameter.
@@ -115,7 +113,7 @@ pub enum Error {
     ParseVdpaPathMissing,
 }
 
-#[derive(Debug, PartialEq, Error)]
+#[derive(Debug, PartialEq, Eq, Error)]
 pub enum ValidationError {
     /// Both console and serial are tty.
     DoubleTtyMode,
@@ -291,9 +289,6 @@ impl fmt::Display for Error {
             ParseFileSystem(o) => write!(f, "Error parsing --fs: {}", o),
             ParseFsSockMissing => write!(f, "Error parsing --fs: socket missing"),
             ParseFsTagMissing => write!(f, "Error parsing --fs: tag missing"),
-            InvalidCacheSizeWithDaxOff => {
-                write!(f, "Error parsing --fs: cache_size used with dax=on")
-            }
             ParsePersistentMemory(o) => write!(f, "Error parsing --pmem: {}", o),
             ParsePmemFileMissing => write!(f, "Error parsing --pmem: file missing"),
             ParseVsock(o) => write!(f, "Error parsing --vsock: {}", o),
@@ -435,7 +430,7 @@ impl<'a> VmParams<'a> {
     }
 }
 
-#[derive(Clone, Copy, Debug, PartialEq, Deserialize, Serialize)]
+#[derive(Clone, Copy, Debug, PartialEq, Eq, Deserialize, Serialize)]
 pub enum HotplugMethod {
     Acpi,
     VirtioMem,
@@ -464,13 +459,13 @@ impl FromStr for HotplugMethod {
     }
 }
 
-#[derive(Clone, Debug, PartialEq, Deserialize, Serialize)]
+#[derive(Clone, Debug, PartialEq, Eq, Deserialize, Serialize)]
 pub struct CpuAffinity {
     pub vcpu: u8,
     pub host_cpus: Vec<u8>,
 }
 
-#[derive(Clone, Debug, Default, PartialEq, Deserialize, Serialize)]
+#[derive(Clone, Debug, Default, PartialEq, Eq, Deserialize, Serialize)]
 pub struct CpuFeatures {
     #[cfg(all(feature = "amx", target_arch = "x86_64"))]
     pub amx: bool,
@@ -480,7 +475,7 @@ pub enum CpuTopologyParseError {
     InvalidValue(String),
 }
 
-#[derive(Clone, Debug, PartialEq, Deserialize, Serialize)]
+#[derive(Clone, Debug, PartialEq, Eq, Deserialize, Serialize)]
 pub struct CpuTopology {
     pub threads_per_core: u8,
     pub cores_per_die: u8,
@@ -521,7 +516,7 @@ fn default_cpuconfig_max_phys_bits() -> u8 {
     DEFAULT_MAX_PHYS_BITS
 }
 
-#[derive(Clone, Debug, PartialEq, Deserialize, Serialize)]
+#[derive(Clone, Debug, PartialEq, Eq, Deserialize, Serialize)]
 pub struct CpusConfig {
     pub boot_vcpus: u8,
     pub max_vcpus: u8,
@@ -631,7 +626,7 @@ fn default_platformconfig_num_pci_segments() -> u16 {
     DEFAULT_NUM_PCI_SEGMENTS
 }
 
-#[derive(Clone, Debug, PartialEq, Deserialize, Serialize)]
+#[derive(Clone, Debug, PartialEq, Eq, Deserialize, Serialize)]
 pub struct PlatformConfig {
     #[serde(default = "default_platformconfig_num_pci_segments")]
     pub num_pci_segments: u16,
@@ -696,7 +691,7 @@ impl Default for PlatformConfig {
     }
 }
 
-#[derive(Clone, Debug, PartialEq, Deserialize, Serialize)]
+#[derive(Clone, Debug, PartialEq, Eq, Deserialize, Serialize)]
 pub struct MemoryZoneConfig {
     pub id: String,
     pub size: u64,
@@ -718,7 +713,7 @@ pub struct MemoryZoneConfig {
     pub prefault: bool,
 }
 
-#[derive(Clone, Debug, PartialEq, Deserialize, Serialize)]
+#[derive(Clone, Debug, PartialEq, Eq, Deserialize, Serialize)]
 pub struct MemoryConfig {
     pub size: u64,
     #[serde(default)]
@@ -923,17 +918,17 @@ impl Default for MemoryConfig {
     }
 }
 
-#[derive(Clone, Debug, PartialEq, Deserialize, Serialize)]
+#[derive(Clone, Debug, PartialEq, Eq, Deserialize, Serialize)]
 pub struct KernelConfig {
     pub path: PathBuf,
 }
 
-#[derive(Clone, Debug, PartialEq, Deserialize, Serialize)]
+#[derive(Clone, Debug, PartialEq, Eq, Deserialize, Serialize)]
 pub struct InitramfsConfig {
     pub path: PathBuf,
 }
 
-#[derive(Clone, Debug, Default, PartialEq, Deserialize, Serialize)]
+#[derive(Clone, Debug, Default, PartialEq, Eq, Deserialize, Serialize)]
 pub struct CmdlineConfig {
     pub args: String,
 }
@@ -948,7 +943,7 @@ impl CmdlineConfig {
     }
 }
 
-#[derive(Clone, Debug, PartialEq, Deserialize, Serialize)]
+#[derive(Clone, Debug, PartialEq, Eq, Deserialize, Serialize)]
 pub struct DiskConfig {
     pub path: Option<PathBuf>,
     #[serde(default)]
@@ -1183,7 +1178,7 @@ impl DiskConfig {
     }
 }
 
-#[derive(Clone, Debug, PartialEq, Deserialize, Serialize)]
+#[derive(Clone, Debug, PartialEq, Eq, Deserialize, Serialize)]
 pub enum VhostMode {
     Client,
     Server,
@@ -1212,7 +1207,7 @@ impl FromStr for VhostMode {
     }
 }
 
-#[derive(Clone, Debug, PartialEq, Deserialize, Serialize)]
+#[derive(Clone, Debug, PartialEq, Eq, Deserialize, Serialize)]
 pub struct NetConfig {
     #[serde(default = "default_netconfig_tap")]
     pub tap: Option<String>,
@@ -1483,7 +1478,7 @@ impl NetConfig {
     }
 }
 
-#[derive(Clone, Debug, PartialEq, Deserialize, Serialize)]
+#[derive(Clone, Debug, PartialEq, Eq, Deserialize, Serialize)]
 pub struct RngConfig {
     pub src: PathBuf,
     #[serde(default)]
@@ -1520,7 +1515,7 @@ impl Default for RngConfig {
     }
 }
 
-#[derive(Clone, Debug, PartialEq, Deserialize, Serialize)]
+#[derive(Clone, Debug, PartialEq, Eq, Deserialize, Serialize)]
 pub struct BalloonConfig {
     pub size: u64,
     /// Option to deflate the balloon in case the guest is out of memory.
@@ -1569,7 +1564,7 @@ impl BalloonConfig {
     }
 }
 
-#[derive(Clone, Debug, PartialEq, Deserialize, Serialize)]
+#[derive(Clone, Debug, PartialEq, Eq, Deserialize, Serialize)]
 pub struct FsConfig {
     pub tag: String,
     pub socket: PathBuf,
@@ -1577,10 +1572,6 @@ pub struct FsConfig {
     pub num_queues: usize,
     #[serde(default = "default_fsconfig_queue_size")]
     pub queue_size: u16,
-    #[serde(default = "default_fsconfig_dax")]
-    pub dax: bool,
-    #[serde(default = "default_fsconfig_cache_size")]
-    pub cache_size: u64,
     #[serde(default)]
     pub id: Option<String>,
     #[serde(default)]
@@ -1595,14 +1586,6 @@ fn default_fsconfig_queue_size() -> u16 {
     1024
 }
 
-fn default_fsconfig_dax() -> bool {
-    false
-}
-
-fn default_fsconfig_cache_size() -> u64 {
-    0x0002_0000_0000
-}
-
 impl Default for FsConfig {
     fn default() -> Self {
         Self {
@@ -1610,8 +1593,6 @@ impl Default for FsConfig {
             socket: PathBuf::new(),
             num_queues: default_fsconfig_num_queues(),
             queue_size: default_fsconfig_queue_size(),
-            dax: default_fsconfig_dax(),
-            cache_size: default_fsconfig_cache_size(),
             id: None,
             pci_segment: 0,
         }
@@ -1621,15 +1602,12 @@ impl Default for FsConfig {
 impl FsConfig {
     pub const SYNTAX: &'static str = "virtio-fs parameters \
     \"tag=<tag_name>,socket=<socket_path>,num_queues=<number_of_queues>,\
-    queue_size=<size_of_each_queue>,dax=on|off,cache_size=<DAX cache size: \
+    queue_size=<size_of_each_queue>,id=<device_id>,pci_segment=<segment_id>\"";
-    default 8Gib>,id=<device_id>,pci_segment=<segment_id>\"";
 
     pub fn parse(fs: &str) -> Result<Self> {
         let mut parser = OptionParser::new();
         parser
             .add("tag")
-            .add("dax")
-            .add("cache_size")
             .add("queue_size")
             .add("num_queues")
             .add("socket")
@@ -1649,22 +1627,6 @@ impl FsConfig {
             .map_err(Error::ParseFileSystem)?
             .unwrap_or_else(default_fsconfig_num_queues);
 
-        let dax = parser
-            .convert::<Toggle>("dax")
-            .map_err(Error::ParseFileSystem)?
-            .unwrap_or_else(|| Toggle(default_fsconfig_dax()))
-            .0;
-
-        if parser.is_set("cache_size") && !dax {
-            return Err(Error::InvalidCacheSizeWithDaxOff);
-        }
-
-        let cache_size = parser
-            .convert::<ByteSized>("cache_size")
-            .map_err(Error::ParseFileSystem)?
-            .unwrap_or_else(|| ByteSized(default_fsconfig_cache_size()))
-            .0;
-
         let id = parser.get("id");
 
         let pci_segment = parser
@@ -1677,8 +1639,6 @@ impl FsConfig {
             socket,
             num_queues,
             queue_size,
-            dax,
-            cache_size,
             id,
             pci_segment,
         })
@@ -1703,20 +1663,11 @@ impl FsConfig {
             }
         }
 
-        if self.dax {
-            // TODO: Remove the dax parameter.
-            // Tracked by https://github.com/cloud-hypervisor/cloud-hypervisor/issues/3889
-            warn!(
-                "The experimental DAX feature is deprecated and will be \
-                removed in a future release. A request to enable it is ignored."
-            );
-        }
-
         Ok(())
     }
 }
 
-#[derive(Clone, Debug, PartialEq, Deserialize, Serialize, Default)]
+#[derive(Clone, Debug, PartialEq, Eq, Deserialize, Serialize, Default)]
 pub struct PmemConfig {
     pub file: PathBuf,
     #[serde(default)]
@@ -1724,8 +1675,6 @@ pub struct PmemConfig {
     #[serde(default)]
     pub iommu: bool,
     #[serde(default)]
-    pub mergeable: bool,
-    #[serde(default)]
     pub discard_writes: bool,
     #[serde(default)]
     pub id: Option<String>,
@@ -1742,7 +1691,6 @@ impl PmemConfig {
         parser
             .add("size")
             .add("file")
-            .add("mergeable")
             .add("iommu")
             .add("discard_writes")
             .add("id")
@@ -1754,11 +1702,6 @@ impl PmemConfig {
             .convert::<ByteSized>("size")
             .map_err(Error::ParsePersistentMemory)?
             .map(|v| v.0);
-        let mergeable = parser
-            .convert::<Toggle>("mergeable")
-            .map_err(Error::ParsePersistentMemory)?
-            .unwrap_or(Toggle(false))
-            .0;
         let iommu = parser
             .convert::<Toggle>("iommu")
             .map_err(Error::ParsePersistentMemory)?
@@ -1779,7 +1722,6 @@ impl PmemConfig {
             file,
             size,
             iommu,
-            mergeable,
             discard_writes,
             id,
             pci_segment,
@@ -1787,10 +1729,6 @@ impl PmemConfig {
     }
 
     pub fn validate(&self, vm_config: &VmConfig) -> ValidationResult<()> {
-        if self.mergeable {
-            warn!("Enabling mergable pages for PMEM devices is ineffectual. This option is deprecated and will be removed in a later release");
-        }
-
         if let Some(platform_config) = vm_config.platform.as_ref() {
             if self.pci_segment >= platform_config.num_pci_segments {
                 return Err(ValidationError::InvalidPciSegment(self.pci_segment));
@@ -1807,7 +1745,7 @@ impl PmemConfig {
     }
 }
 
-#[derive(Clone, Debug, PartialEq, Deserialize, Serialize)]
+#[derive(Clone, Debug, PartialEq, Eq, Deserialize, Serialize)]
 pub enum ConsoleOutputMode {
     Off,
     Pty,
@@ -1816,7 +1754,7 @@ pub enum ConsoleOutputMode {
     Null,
 }
 
-#[derive(Clone, Debug, PartialEq, Deserialize, Serialize)]
+#[derive(Clone, Debug, PartialEq, Eq, Deserialize, Serialize)]
 pub struct ConsoleConfig {
     #[serde(default = "default_consoleconfig_file")]
     pub file: Option<PathBuf>,
@@ -1886,7 +1824,7 @@ impl ConsoleConfig {
     }
 }
 
-#[derive(Clone, Debug, PartialEq, Deserialize, Serialize, Default)]
+#[derive(Clone, Debug, PartialEq, Eq, Deserialize, Serialize, Default)]
 pub struct DeviceConfig {
     pub path: PathBuf,
     #[serde(default)]
@@ -1945,7 +1883,7 @@ impl DeviceConfig {
     }
 }
 
-#[derive(Clone, Debug, PartialEq, Deserialize, Serialize, Default)]
+#[derive(Clone, Debug, PartialEq, Eq, Deserialize, Serialize, Default)]
 pub struct UserDeviceConfig {
     pub socket: PathBuf,
     #[serde(default)]
@@ -1998,7 +1936,7 @@ impl UserDeviceConfig {
     }
 }
 
-#[derive(Clone, Debug, PartialEq, Deserialize, Serialize, Default)]
+#[derive(Clone, Debug, PartialEq, Eq, Deserialize, Serialize, Default)]
 pub struct VdpaConfig {
     pub path: PathBuf,
     #[serde(default = "default_vdpaconfig_num_queues")]
@@ -2074,7 +2012,7 @@ impl VdpaConfig {
     }
 }
 
-#[derive(Clone, Debug, PartialEq, Deserialize, Serialize, Default)]
+#[derive(Clone, Debug, PartialEq, Eq, Deserialize, Serialize, Default)]
 pub struct VsockConfig {
     pub cid: u64,
     pub socket: PathBuf,
@@ -2145,7 +2083,7 @@ impl VsockConfig {
 }
 
 #[cfg(feature = "tdx")]
-#[derive(Clone, Debug, PartialEq, Deserialize, Serialize, Default)]
+#[derive(Clone, Debug, PartialEq, Eq, Deserialize, Serialize, Default)]
 pub struct TdxConfig {
     pub firmware: PathBuf,
 }
@@ -2165,7 +2103,7 @@ impl TdxConfig {
 }
 
 #[cfg(target_arch = "x86_64")]
-#[derive(Clone, Debug, PartialEq, Deserialize, Serialize, Default)]
+#[derive(Clone, Debug, PartialEq, Eq, Deserialize, Serialize, Default)]
 pub struct SgxEpcConfig {
     pub id: String,
     #[serde(default)]
@@ -2199,7 +2137,7 @@ impl SgxEpcConfig {
     }
 }
 
-#[derive(Clone, Debug, PartialEq, Deserialize, Serialize, Default)]
+#[derive(Clone, Debug, PartialEq, Eq, Deserialize, Serialize, Default)]
 pub struct NumaDistance {
     #[serde(default)]
     pub destination: u32,
@@ -2207,7 +2145,7 @@ pub struct NumaDistance {
     pub distance: u8,
 }
 
-#[derive(Clone, Debug, PartialEq, Deserialize, Serialize, Default)]
+#[derive(Clone, Debug, PartialEq, Eq, Deserialize, Serialize, Default)]
 pub struct NumaConfig {
     #[serde(default)]
     pub guest_numa_id: u32,
@@ -2276,7 +2214,7 @@ impl NumaConfig {
     }
 }
 
-#[derive(Clone, Debug, PartialEq, Deserialize, Serialize, Default)]
+#[derive(Clone, Debug, PartialEq, Eq, Deserialize, Serialize, Default)]
 pub struct RestoreConfig {
     pub source_url: PathBuf,
     #[serde(default)]
@@ -2310,7 +2248,7 @@ impl RestoreConfig {
     }
 }
 
-#[derive(Clone, Debug, PartialEq, Deserialize, Serialize)]
+#[derive(Clone, Debug, PartialEq, Eq, Deserialize, Serialize)]
 pub struct VmConfig {
     #[serde(default)]
     pub cpus: CpusConfig,
@@ -3116,29 +3054,7 @@ mod tests {
                 ..Default::default()
             }
         );
-        // DAX on -> default cache size
+
-        assert_eq!(
-            FsConfig::parse("tag=mytag,socket=/tmp/sock,dax=on")?,
-            FsConfig {
-                socket: PathBuf::from("/tmp/sock"),
-                tag: "mytag".to_owned(),
-                dax: true,
-                cache_size: default_fsconfig_cache_size(),
-                ..Default::default()
-            }
-        );
-        assert_eq!(
-            FsConfig::parse("tag=mytag,socket=/tmp/sock,dax=on,cache_size=4G")?,
-            FsConfig {
-                socket: PathBuf::from("/tmp/sock"),
-                tag: "mytag".to_owned(),
-                dax: true,
-                cache_size: 4 << 30,
-                ..Default::default()
-            }
-        );
-        // Cache size without DAX is an error
-        assert!(FsConfig::parse("tag=mytag,socket=/tmp/sock,dax=off,cache_size=4G").is_err());
         Ok(())
     }
 
@@ -3165,11 +3081,10 @@ mod tests {
             }
         );
         assert_eq!(
-            PmemConfig::parse("file=/tmp/pmem,size=128M,iommu=on,mergeable=on,discard_writes=on")?,
+            PmemConfig::parse("file=/tmp/pmem,size=128M,iommu=on,discard_writes=on")?,
             PmemConfig {
                 file: PathBuf::from("/tmp/pmem"),
                 size: Some(128 << 20),
-                mergeable: true,
                 discard_writes: true,
                 iommu: true,
                 ..Default::default()

vmm/src/coredump.rs

@@ -0,0 +1,348 @@
+// Copyright © 2022 ZTE Corporation
+//
+// SPDX-License-Identifier: Apache-2.0
+//
+
+#[cfg(target_arch = "x86_64")]
+use hypervisor::kvm::kvm_bindings::kvm_dtable as DTableRegister;
+#[cfg(target_arch = "x86_64")]
+use hypervisor::x86_64::SegmentRegister;
+use linux_loader::elf;
+use std::fs::File;
+use std::io::Write;
+use vm_memory::ByteValued;
+
+#[derive(Clone)]
+pub struct CoredumpMemoryRegion {
+    pub mem_offset_in_elf: u64,
+    pub mem_size: u64,
+}
+
+#[derive(Clone)]
+pub struct CoredumpMemoryRegions {
+    pub ram_maps: std::collections::BTreeMap<u64, CoredumpMemoryRegion>,
+}
+
+/// Platform information
+#[derive(Default)]
+pub struct DumpState {
+    pub elf_note_size: isize,
+    pub elf_phdr_num: u16,
+    pub elf_sh_info: u32,
+    pub mem_offset: u64,
+    pub mem_info: Option<CoredumpMemoryRegions>,
+    pub file: Option<File>,
+}
+
+#[derive(Debug)]
+pub enum GuestDebuggableError {
+    Coredump(anyhow::Error),
+    CoredumpFile(std::io::Error),
+}
+
+pub trait GuestDebuggable: vm_migration::Pausable {
+    fn coredump(
+        &mut self,
+        _destination_url: &str,
+    ) -> std::result::Result<(), GuestDebuggableError> {
+        Ok(())
+    }
+}
+
+#[macro_export]
+macro_rules! div_round_up {
+    ($n:expr,$d:expr) => {
+        ($n + $d - 1) / $d
+    };
+}
+
+#[repr(C)]
+#[derive(Default, Copy, Clone)]
+pub struct X86_64UserRegs {
+    /// r15, r14, r13, r12, rbp, rbx, r11, r10;
+    pub regs1: [u64; 8],
+    /// r9, r8, rax, rcx, rdx, rsi, rdi, orig_rax;
+    pub regs2: [u64; 8],
+    pub rip: u64,
+    pub cs: u64,
+    pub eflags: u64,
+    pub rsp: u64,
+    pub ss: u64,
+    pub fs_base: u64,
+    pub gs_base: u64,
+    pub ds: u64,
+    pub es: u64,
+    pub fs: u64,
+    pub gs: u64,
+}
+
+unsafe impl ByteValued for X86_64UserRegs {}
+
+#[repr(C)]
+pub struct X86_64ElfPrStatus {
+    pub pad1: [u8; 32],
+    pub pid: u32,
+    pub pads2: [u8; 76],
+    pub regs: X86_64UserRegs,
+    pub pad3: [u8; 8],
+}
+
+#[repr(C)]
+#[derive(Default, Copy, Clone)]
+pub struct CpuSegment {
+    pub selector: u32,
+    pub limit: u32,
+    pub flags: u32,
+    pub pad: u32,
+    pub base: u64,
+}
+
+const DESC_TYPE_SHIFT: u32 = 8;
+const DESC_S_SHIFT: u32 = 12;
+const DESC_DPL_SHIFT: u32 = 13;
+const DESC_P_SHIFT: u32 = 15;
+const DESC_P_MASK: u32 = 1 << DESC_P_SHIFT;
+const DESC_AVL_SHIFT: u32 = 20;
+const DESC_AVL_MASK: u32 = 1 << DESC_AVL_SHIFT;
+const DESC_L_SHIFT: u32 = 21;
+const DESC_B_SHIFT: u32 = 22;
+const DESC_S_MASK: u32 = 1 << DESC_S_SHIFT;
+const DESC_G_SHIFT: u32 = 23;
+const DESC_G_MASK: u32 = 1 << DESC_G_SHIFT;
+
+impl CpuSegment {
+    pub fn new(reg: SegmentRegister) -> Self {
+        let p_mask = if (reg.present > 0) && (reg.unusable == 0) {
+            DESC_P_MASK
+        } else {
+            0
+        };
+        let flags = ((reg.type_ as u32) << DESC_TYPE_SHIFT)
+            | p_mask
+            | ((reg.dpl as u32) << DESC_DPL_SHIFT)
+            | ((reg.db as u32) << DESC_B_SHIFT)
+            | ((reg.s as u32) * DESC_S_MASK)
+            | ((reg.l as u32) << DESC_L_SHIFT)
+            | ((reg.g as u32) * DESC_G_MASK)
+            | ((reg.avl as u32) * DESC_AVL_MASK);
+
+        CpuSegment {
+            selector: reg.selector as u32,
+            limit: reg.limit as u32,
+            flags,
+            pad: 0,
+            base: reg.base,
+        }
+    }
+
+    pub fn new_from_table(reg: DTableRegister) -> Self {
+        CpuSegment {
+            selector: 0,
+            limit: reg.limit as u32,
+            flags: 0,
+            pad: 0,
+            base: reg.base,
+        }
+    }
+}
+
+#[repr(C)]
+#[derive(Default, Copy, Clone)]
+pub struct CpuState {
+    pub version: u32,
+    pub size: u32,
+    /// rax, rbx, rcx, rdx, rsi, rdi, rsp, rbp
+    pub regs1: [u64; 8],
+    /// r8, r9, r10, r11, r12, r13, r14, r15
+    pub regs2: [u64; 8],
+    pub rip: u64,
+    pub rflags: u64,
+    pub cs: CpuSegment,
+    pub ds: CpuSegment,
+    pub es: CpuSegment,
+    pub fs: CpuSegment,
+    pub gs: CpuSegment,
+    pub ss: CpuSegment,
+    pub ldt: CpuSegment,
+    pub tr: CpuSegment,
+    pub gdt: CpuSegment,
+    pub idt: CpuSegment,
+    pub cr: [u64; 5],
+    pub kernel_gs_base: u64,
+}
+
+unsafe impl ByteValued for CpuState {}
+
+pub enum NoteDescType {
+    Elf = 0,
+    Vmm = 1,
+    ElfAndVmm = 2,
+}
+
+// "CORE" or "QEMU"
+pub const COREDUMP_NAME_SIZE: u32 = 5;
+pub const NT_PRSTATUS: u32 = 1;
+
+/// Core file.
+const ET_CORE: u16 = 4;
+/// 64-bit object.
+const ELFCLASS64: u8 = 2;
+/// Current ELF version.
+const EV_CURRENT: u8 = 1;
+/// AMD x86-64 architecture
+const EM_X86_64: u16 = 62;
+
+pub trait Elf64Writable {
+    fn write_header(
+        &mut self,
+        dump_state: &DumpState,
+    ) -> std::result::Result<(), GuestDebuggableError> {
+        let e_ident = [
+            elf::ELFMAG0 as u8, // magic
+            elf::ELFMAG1,
+            elf::ELFMAG2,
+            elf::ELFMAG3,
+            ELFCLASS64,             // class
+            elf::ELFDATA2LSB as u8, //data
+            EV_CURRENT,             // version
+            0,                      // os_abi
+            0,                      // abi_version
+            0,                      // padding
+            0,
+            0,
+            0,
+            0,
+            0,
+            0,
+        ];
+        let elf64_ehdr_size = std::mem::size_of::<elf::Elf64_Ehdr>();
+        let elf64_phdr_size = std::mem::size_of::<elf::Elf64_Phdr>();
+        let mut elf64_ehdr = elf::Elf64_Ehdr {
+            e_ident,
+            e_type: ET_CORE,
+            e_machine: EM_X86_64,
+            e_version: EV_CURRENT as u32,
+            e_entry: 0,
+            e_phoff: elf64_ehdr_size as u64,
+            e_shoff: 0,
+            e_flags: 0,
+            e_ehsize: 0,
+            e_phentsize: elf64_phdr_size as u16,
+            e_phnum: dump_state.elf_phdr_num,
+            e_shentsize: 0,
+            e_shnum: 0,
+            e_shstrndx: 0,
+        };
+        elf64_ehdr.e_ehsize = std::mem::size_of_val(&elf64_ehdr) as u16;
+
+        let mut coredump_file = dump_state.file.as_ref().unwrap();
+        let bytes: &[u8] = elf64_ehdr.as_slice();
+        coredump_file
+            .write(bytes)
+            .map_err(GuestDebuggableError::CoredumpFile)?;
+
+        Ok(())
+    }
+
+    fn write_note(
+        &mut self,
+        dump_state: &DumpState,
+    ) -> std::result::Result<(), GuestDebuggableError> {
+        let begin = dump_state.mem_offset - dump_state.elf_note_size as u64;
+        let elf64_phdr = elf::Elf64_Phdr {
+            p_type: elf::PT_NOTE,
+            p_flags: 0,
+            p_offset: begin,
+            p_vaddr: 0,
+            p_paddr: 0,
+            p_filesz: dump_state.elf_note_size as u64,
+            p_memsz: dump_state.elf_note_size as u64,
+            p_align: 0,
+        };
+
+        let mut coredump_file = dump_state.file.as_ref().unwrap();
+        let bytes: &[u8] = elf64_phdr.as_slice();
+        coredump_file
+            .write(bytes)
+            .map_err(GuestDebuggableError::CoredumpFile)?;
+
+        Ok(())
+    }
+
+    fn write_load(
+        &mut self,
+        offset: u64,
+        phys_addr: u64,
+        length: u64,
+        virt_addr: u64,
+        dump_state: &DumpState,
+    ) -> std::result::Result<(), GuestDebuggableError> {
+        let elf64_load = elf::Elf64_Phdr {
+            p_type: elf::PT_LOAD,
+            p_flags: 0,
+            p_offset: offset,
+            p_vaddr: virt_addr,
+            p_paddr: phys_addr,
+            p_filesz: length,
+            p_memsz: length,
+            p_align: 0,
+        };
+
+        let mut coredump_file = dump_state.file.as_ref().unwrap();
+        let bytes: &[u8] = elf64_load.as_slice();
+        coredump_file
+            .write(bytes)
+            .map_err(GuestDebuggableError::CoredumpFile)?;
+
+        Ok(())
+    }
+
+    fn write_loads(
+        &mut self,
+        dump_state: &DumpState,
+    ) -> std::result::Result<(), GuestDebuggableError> {
+        let mem_info = dump_state.mem_info.as_ref().unwrap();
+
+        for (gpa, load) in &mem_info.ram_maps {
+            self.write_load(load.mem_offset_in_elf, *gpa, load.mem_size, 0, dump_state)?;
+        }
+
+        Ok(())
+    }
+
+    fn elf_note_size(&self, hdr_size: u32, name_size: u32, desc_size: u32) -> u32 {
+        (div_round_up!(hdr_size, 4) + div_round_up!(name_size, 4) + div_round_up!(desc_size, 4)) * 4
+    }
+
+    fn get_note_size(&self, desc_type: NoteDescType, nr_cpus: u32) -> u32 {
+        let note_head_size = std::mem::size_of::<elf::Elf64_Nhdr>() as u32;
+        let elf_desc_size = std::mem::size_of::<X86_64ElfPrStatus>() as u32;
+        let cpu_state_desc_size = std::mem::size_of::<CpuState>() as u32;
+
+        let elf_note_size = self.elf_note_size(note_head_size, COREDUMP_NAME_SIZE, elf_desc_size);
+        let vmm_note_size =
+            self.elf_note_size(note_head_size, COREDUMP_NAME_SIZE, cpu_state_desc_size);
+
+        match desc_type {
+            NoteDescType::Elf => elf_note_size * nr_cpus,
+            NoteDescType::Vmm => vmm_note_size * nr_cpus,
+            NoteDescType::ElfAndVmm => (elf_note_size + vmm_note_size) * nr_cpus,
+        }
+    }
+}
+
+pub trait CpuElf64Writable {
+    fn cpu_write_elf64_note(
+        &mut self,
+        _dump_state: &DumpState,
+    ) -> std::result::Result<(), GuestDebuggableError> {
+        Ok(())
+    }
+
+    fn cpu_write_vmm_note(
+        &mut self,
+        _dump_state: &DumpState,
+    ) -> std::result::Result<(), GuestDebuggableError> {
+        Ok(())
+    }
+}

vmm/src/cpu.rs

@@ -12,6 +12,12 @@
 //
 
 use crate::config::CpusConfig;
+#[cfg(feature = "guest_debug")]
+use crate::coredump::{
+    CpuElf64Writable, CpuSegment, CpuState as DumpCpusState, DumpState, Elf64Writable,
+    GuestDebuggableError, NoteDescType, X86_64ElfPrStatus, X86_64UserRegs, COREDUMP_NAME_SIZE,
+    NT_PRSTATUS,
+};
 use crate::device_manager::DeviceManager;
 #[cfg(feature = "gdb")]
 use crate::gdb::{get_raw_tid, Debuggable, DebuggableError};
@@ -28,24 +34,36 @@ use arch::NumaNodes;
 use devices::interrupt_controller::InterruptController;
 #[cfg(all(target_arch = "x86_64", feature = "gdb"))]
 use gdbstub_arch::x86::reg::{X86SegmentRegs, X86_64CoreRegs};
+#[cfg(feature = "guest_debug")]
+use hypervisor::arch::x86::msr_index;
 #[cfg(target_arch = "aarch64")]
 use hypervisor::kvm::kvm_bindings;
 #[cfg(feature = "tdx")]
 use hypervisor::kvm::{TdxExitDetails, TdxExitStatus};
 #[cfg(target_arch = "x86_64")]
 use hypervisor::x86_64::CpuId;
+#[cfg(feature = "guest_debug")]
+use hypervisor::x86_64::{MsrEntries, MsrEntry};
 #[cfg(all(target_arch = "x86_64", feature = "gdb"))]
 use hypervisor::x86_64::{SpecialRegisters, StandardRegisters};
 use hypervisor::{CpuState, HypervisorCpuError, VmExit, VmOps};
 use libc::{c_void, siginfo_t};
+#[cfg(feature = "guest_debug")]
+use linux_loader::elf::Elf64_Nhdr;
 use seccompiler::{apply_filter, SeccompAction};
 use std::collections::BTreeMap;
+#[cfg(feature = "guest_debug")]
+use std::io::Write;
+#[cfg(feature = "guest_debug")]
+use std::mem::size_of;
 use std::os::unix::thread::JoinHandleExt;
 use std::sync::atomic::{AtomicBool, Ordering};
 use std::sync::{Arc, Barrier, Mutex};
 use std::{cmp, io, result, thread};
 use thiserror::Error;
 use vm_device::BusDevice;
+#[cfg(feature = "guest_debug")]
+use vm_memory::ByteValued;
 #[cfg(feature = "gdb")]
 use vm_memory::{Bytes, GuestAddressSpace};
 use vm_memory::{GuestAddress, GuestMemoryAtomic};
@@ -238,6 +256,13 @@ struct InterruptSourceOverride {
     pub flags: u16,
 }
 
+#[cfg(feature = "guest_debug")]
+macro_rules! round_up {
+    ($n:expr,$d:expr) => {
+        (($n / ($d + 1)) + 1) * $d
+    };
+}
+
 /// A wrapper around creating and using a kvm-based VCPU.
 pub struct Vcpu {
     // The hypervisor abstracted CPU.
@@ -350,43 +375,38 @@ impl Vcpu {
 }
 
 const VCPU_SNAPSHOT_ID: &str = "vcpu";
-impl Pausable for Vcpu {
+impl Pausable for Vcpu {}
-    fn pause(&mut self) -> std::result::Result<(), MigratableError> {
-        self.saved_state =
-            Some(self.vcpu.state().map_err(|e| {
-                MigratableError::Pause(anyhow!("Could not get vCPU state {:?}", e))
-            })?);
-
-        Ok(())
-    }
-
-    fn resume(&mut self) -> std::result::Result<(), MigratableError> {
-        if let Some(vcpu_state) = &self.saved_state {
-            self.vcpu.set_state(vcpu_state).map_err(|e| {
-                MigratableError::Pause(anyhow!("Could not set the vCPU state {:?}", e))
-            })?;
-        }
-
-        Ok(())
-    }
-}
 impl Snapshottable for Vcpu {
     fn id(&self) -> String {
         VCPU_SNAPSHOT_ID.to_string()
     }
 
     fn snapshot(&mut self) -> std::result::Result<Snapshot, MigratableError> {
-        let mut vcpu_snapshot = Snapshot::new(&format!("{}", self.id));
+        let saved_state = self
+            .vcpu
+            .state()
+            .map_err(|e| MigratableError::Pause(anyhow!("Could not get vCPU state {:?}", e)))?;
+
+        let mut vcpu_snapshot = Snapshot::new(&format!("{:03}", self.id));
         vcpu_snapshot.add_data_section(SnapshotDataSection::new_from_state(
             VCPU_SNAPSHOT_ID,
-            &self.saved_state,
+            &saved_state,
         )?);
 
+        self.saved_state = Some(saved_state);
+
         Ok(vcpu_snapshot)
     }
 
     fn restore(&mut self, snapshot: Snapshot) -> std::result::Result<(), MigratableError> {
-        self.saved_state = Some(snapshot.to_state(VCPU_SNAPSHOT_ID)?);
+        let saved_state: CpuState = snapshot.to_state(VCPU_SNAPSHOT_ID)?;
+
+        self.vcpu
+            .set_state(&saved_state)
+            .map_err(|e| MigratableError::Pause(anyhow!("Could not set the vCPU state {:?}", e)))?;
+
+        self.saved_state = Some(saved_state);
+
         Ok(())
     }
 }
@@ -1251,7 +1271,7 @@ impl CpuManager {
                         1 << MADT_CPU_ENABLE_FLAG
                     } else {
                         0
-                    },
+                    } | 1 << MADT_CPU_ONLINE_CAPABLE_FLAG,
                 };
                 madt.append(lapic);
             }
@@ -1504,6 +1524,9 @@ struct Cpu {
 #[cfg(target_arch = "x86_64")]
 const MADT_CPU_ENABLE_FLAG: usize = 0;
 
+#[cfg(target_arch = "x86_64")]
+const MADT_CPU_ONLINE_CAPABLE_FLAG: usize = 1;
+
 impl Cpu {
     #[cfg(target_arch = "x86_64")]
     fn generate_mat(&self) -> Vec<u8> {
@@ -2097,6 +2120,217 @@ impl Debuggable for CpuManager {
     }
 }
 
+#[cfg(feature = "guest_debug")]
+impl Elf64Writable for CpuManager {}
+
+#[cfg(feature = "guest_debug")]
+impl CpuElf64Writable for CpuManager {
+    fn cpu_write_elf64_note(
+        &mut self,
+        dump_state: &DumpState,
+    ) -> std::result::Result<(), GuestDebuggableError> {
+        let mut coredump_file = dump_state.file.as_ref().unwrap();
+        for vcpu in &self.vcpus {
+            let note_size = self.get_note_size(NoteDescType::Elf, 1);
+            let mut pos: usize = 0;
+            let mut buf = vec![0; note_size as usize];
+            let descsz = size_of::<X86_64ElfPrStatus>();
+            let vcpu_id = vcpu.lock().unwrap().id;
+
+            let note = Elf64_Nhdr {
+                n_namesz: COREDUMP_NAME_SIZE,
+                n_descsz: descsz as u32,
+                n_type: NT_PRSTATUS,
+            };
+
+            let bytes: &[u8] = note.as_slice();
+            buf.splice(0.., bytes.to_vec());
+            pos += round_up!(size_of::<Elf64_Nhdr>(), 4);
+            buf.resize(pos + 4, 0);
+            buf.splice(pos.., "CORE".to_string().into_bytes());
+
+            pos += round_up!(COREDUMP_NAME_SIZE as usize, 4);
+            buf.resize(pos + 32 + 4, 0);
+            let pid = vcpu_id as u64;
+            let bytes: &[u8] = pid.as_slice();
+            buf.splice(pos + 32.., bytes.to_vec()); /* pr_pid */
+
+            pos += descsz - size_of::<X86_64UserRegs>() - size_of::<u64>();
+
+            let orig_rax: u64 = 0;
+            let gregs = self.vcpus[usize::from(vcpu_id)]
+                .lock()
+                .unwrap()
+                .vcpu
+                .get_regs()
+                .map_err(|_e| GuestDebuggableError::Coredump(anyhow!("get regs failed")))?;
+
+            let regs1 = [
+                gregs.r15, gregs.r14, gregs.r13, gregs.r12, gregs.rbp, gregs.rbx, gregs.r11,
+                gregs.r10,
+            ];
+            let regs2 = [
+                gregs.r9, gregs.r8, gregs.rax, gregs.rcx, gregs.rdx, gregs.rsi, gregs.rdi, orig_rax,
+            ];
+
+            let sregs = self.vcpus[usize::from(vcpu_id)]
+                .lock()
+                .unwrap()
+                .vcpu
+                .get_sregs()
+                .map_err(|_e| GuestDebuggableError::Coredump(anyhow!("get sregs failed")))?;
+
+            debug!(
+                "rip 0x{:x} rsp 0x{:x} gs 0x{:x} cs 0x{:x} ss 0x{:x} ds 0x{:x}",
+                gregs.rip,
+                gregs.rsp,
+                sregs.gs.base,
+                sregs.cs.selector,
+                sregs.ss.selector,
+                sregs.ds.selector,
+            );
+
+            let regs = X86_64UserRegs {
+                regs1,
+                regs2,
+                rip: gregs.rip,
+                cs: sregs.cs.selector as u64,
+                eflags: gregs.rflags,
+                rsp: gregs.rsp,
+                ss: sregs.ss.selector as u64,
+                fs_base: sregs.fs.base as u64,
+                gs_base: sregs.gs.base as u64,
+                ds: sregs.ds.selector as u64,
+                es: sregs.es.selector as u64,
+                fs: sregs.fs.selector as u64,
+                gs: sregs.gs.selector as u64,
+            };
+
+            // let bytes: &[u8] = unsafe { any_as_u8_slice(&regs) };
+            let bytes: &[u8] = regs.as_slice();
+            buf.resize(note_size as usize, 0);
+            buf.splice(pos.., bytes.to_vec());
+            buf.resize(note_size as usize, 0);
+
+            coredump_file
+                .write(&buf)
+                .map_err(GuestDebuggableError::CoredumpFile)?;
+        }
+
+        Ok(())
+    }
+
+    fn cpu_write_vmm_note(
+        &mut self,
+        dump_state: &DumpState,
+    ) -> std::result::Result<(), GuestDebuggableError> {
+        let mut coredump_file = dump_state.file.as_ref().unwrap();
+        for vcpu in &self.vcpus {
+            let note_size = self.get_note_size(NoteDescType::Vmm, 1);
+            let mut pos: usize = 0;
+            let mut buf = vec![0; note_size as usize];
+            let descsz = size_of::<DumpCpusState>();
+            let vcpu_id = vcpu.lock().unwrap().id;
+
+            let note = Elf64_Nhdr {
+                n_namesz: COREDUMP_NAME_SIZE,
+                n_descsz: descsz as u32,
+                n_type: 0,
+            };
+
+            let bytes: &[u8] = note.as_slice();
+            buf.splice(0.., bytes.to_vec());
+            pos += round_up!(size_of::<Elf64_Nhdr>(), 4);
+
+            buf.resize(pos + 4, 0);
+            buf.splice(pos.., "QEMU".to_string().into_bytes());
+
+            pos += round_up!(COREDUMP_NAME_SIZE as usize, 4);
+
+            let gregs = self.vcpus[usize::from(vcpu_id)]
+                .lock()
+                .unwrap()
+                .vcpu
+                .get_regs()
+                .map_err(|_e| GuestDebuggableError::Coredump(anyhow!("get regs failed")))?;
+
+            let regs1 = [
+                gregs.rax, gregs.rbx, gregs.rcx, gregs.rdx, gregs.rsi, gregs.rdi, gregs.rsp,
+                gregs.rbp,
+            ];
+
+            let regs2 = [
+                gregs.r8, gregs.r9, gregs.r10, gregs.r11, gregs.r12, gregs.r13, gregs.r14,
+                gregs.r15,
+            ];
+
+            let sregs = self.vcpus[usize::from(vcpu_id)]
+                .lock()
+                .unwrap()
+                .vcpu
+                .get_sregs()
+                .map_err(|_e| GuestDebuggableError::Coredump(anyhow!("get sregs failed")))?;
+
+            let mut msrs = MsrEntries::from_entries(&[MsrEntry {
+                index: msr_index::MSR_KERNEL_GS_BASE,
+                ..Default::default()
+            }])
+            .map_err(|_e| GuestDebuggableError::Coredump(anyhow!("get msr failed")))?;
+
+            self.vcpus[vcpu_id as usize]
+                .lock()
+                .unwrap()
+                .vcpu
+                .get_msrs(&mut msrs)
+                .map_err(|_e| GuestDebuggableError::Coredump(anyhow!("get msr failed")))?;
+            let kernel_gs_base = msrs.as_slice()[0].data;
+
+            let cs = CpuSegment::new(sregs.cs);
+            let ds = CpuSegment::new(sregs.ds);
+            let es = CpuSegment::new(sregs.es);
+            let fs = CpuSegment::new(sregs.fs);
+            let gs = CpuSegment::new(sregs.gs);
+            let ss = CpuSegment::new(sregs.ss);
+            let ldt = CpuSegment::new(sregs.ldt);
+            let tr = CpuSegment::new(sregs.tr);
+            let gdt = CpuSegment::new_from_table(sregs.gdt);
+            let idt = CpuSegment::new_from_table(sregs.idt);
+            let cr = [sregs.cr0, sregs.cr8, sregs.cr2, sregs.cr3, sregs.cr4];
+            let regs = DumpCpusState {
+                version: 1,
+                size: size_of::<DumpCpusState>() as u32,
+                regs1,
+                regs2,
+                rip: gregs.rip,
+                rflags: gregs.rflags,
+                cs,
+                ds,
+                es,
+                fs,
+                gs,
+                ss,
+                ldt,
+                tr,
+                gdt,
+                idt,
+                cr,
+                kernel_gs_base,
+            };
+
+            let bytes: &[u8] = regs.as_slice();
+            buf.resize(note_size as usize, 0);
+            buf.splice(pos.., bytes.to_vec());
+            buf.resize(note_size as usize, 0);
+
+            coredump_file
+                .write(&buf)
+                .map_err(GuestDebuggableError::CoredumpFile)?;
+        }
+
+        Ok(())
+    }
+}
+
 #[cfg(all(feature = "kvm", target_arch = "x86_64"))]
 #[cfg(test)]
 mod tests {
@@ -2205,7 +2439,7 @@ mod tests {
 #[cfg(target_arch = "aarch64")]
 #[cfg(test)]
 mod tests {
-    use arch::aarch64::regs::*;
+    use arch::layout;
     use hypervisor::kvm::aarch64::{is_system_register, MPIDR_EL1};
     use hypervisor::kvm::kvm_bindings::{
         kvm_one_reg, kvm_regs, kvm_vcpu_init, user_pt_regs, KVM_REG_ARM64, KVM_REG_ARM64_SYSREG,
@@ -2220,7 +2454,7 @@ mod tests {
         let vm = hv.create_vm().unwrap();
         let vcpu = vm.create_vcpu(0, None).unwrap();
 
-        let res = setup_regs(&vcpu, 0, 0x0);
+        let res = vcpu.setup_regs(0, 0x0, layout::FDT_START.0);
         // Must fail when vcpu is not initialized yet.
         assert!(res.is_err());
 
@@ -2228,7 +2462,7 @@ mod tests {
         vm.get_preferred_target(&mut kvi).unwrap();
         vcpu.vcpu_init(&kvi).unwrap();
 
-        assert!(setup_regs(&vcpu, 0, 0x0).is_ok());
+        assert!(vcpu.setup_regs(0, 0x0, layout::FDT_START.0).is_ok());
     }
 
     #[test]

vmm/src/device_manager.rs

@@ -16,8 +16,7 @@ use crate::config::{
 use crate::device_tree::{DeviceNode, DeviceTree};
 use crate::interrupt::LegacyUserspaceInterruptManager;
 use crate::interrupt::MsiInterruptManager;
-use crate::memory_manager::MEMORY_MANAGER_ACPI_SIZE;
+use crate::memory_manager::{Error as MemoryManagerError, MemoryManager, MEMORY_MANAGER_ACPI_SIZE};
-use crate::memory_manager::{Error as MemoryManagerError, MemoryManager};
 use crate::pci_segment::PciSegment;
 use crate::seccomp_filters::{get_seccomp_filter, Thread};
 use crate::serial_manager::{Error as SerialManagerError, SerialManager};
@@ -29,8 +28,6 @@ use crate::PciDeviceInfo;
 use crate::{device_node, DEVICE_MANAGER_SNAPSHOT_ID};
 use acpi_tables::{aml, aml::Aml};
 use anyhow::anyhow;
-#[cfg(target_arch = "aarch64")]
-use arch::aarch64::gic::gicv3_its::kvm::KvmGicV3Its;
 use arch::layout;
 #[cfg(target_arch = "x86_64")]
 use arch::layout::{APIC_START, IOAPIC_SIZE, IOAPIC_START};
@@ -79,12 +76,13 @@ use std::result;
 use std::sync::{Arc, Mutex};
 use std::time::Instant;
 use vfio_ioctls::{VfioContainer, VfioDevice};
-use virtio_devices::transport::VirtioPciDevice;
 use virtio_devices::transport::VirtioTransport;
+use virtio_devices::transport::{VirtioPciDevice, VirtioPciDeviceActivator};
 use virtio_devices::vhost_user::VhostUserConfig;
-use virtio_devices::{AccessPlatformMapping, VdpaDmaMapping, VirtioMemMappingSource};
+use virtio_devices::{
+    AccessPlatformMapping, ActivateError, VdpaDmaMapping, VirtioMemMappingSource,
+};
 use virtio_devices::{Endpoint, IommuMapping};
-use virtio_devices::{VirtioSharedMemory, VirtioSharedMemoryList};
 use vm_allocator::{AddressAllocator, SystemAllocator};
 use vm_device::dma_mapping::vfio::VfioDmaMapping;
 use vm_device::dma_mapping::ExternalDmaMapping;
@@ -359,9 +357,6 @@ pub enum DeviceManagerError {
     /// Could not give the PCI device ID back.
     PutPciDeviceId(pci::PciRootError),
 
-    /// Incorrect device ID as it is already used by another device.
-    DeviceIdAlreadyInUse,
-
     /// No disk path was specified when one was expected
     NoDiskPath,
 
@@ -476,6 +471,9 @@ pub enum DeviceManagerError {
 
     /// Invalid identifier
     InvalidIdentifier(String),
+
+    /// Error activating virtio device
+    VirtioActivate(ActivateError),
 }
 pub type DeviceManagerResult<T> = result::Result<T, DeviceManagerError>;
 
@@ -939,6 +937,9 @@ pub struct DeviceManager {
 
     // Start time of the VM
     timestamp: Instant,
+
+    // Pending activations
+    pending_activations: Arc<Mutex<Vec<VirtioPciDeviceActivator>>>,
 }
 
 impl DeviceManager {
@@ -1080,6 +1081,7 @@ impl DeviceManager {
             io_uring_supported: None,
             boot_id_list,
             timestamp,
+            pending_activations: Arc::new(Mutex::new(Vec::default())),
         };
 
         let device_manager = Arc::new(Mutex::new(device_manager));
@@ -2394,109 +2396,7 @@ impl DeviceManager {
 
         let mut node = device_node!(id);
 
-        // Look for the id in the device tree. If it can be found, that means
-        // the device is being restored, otherwise it's created from scratch.
-        let cache_range = if let Some(node) = self.device_tree.lock().unwrap().get(&id) {
-            info!("Restoring virtio-fs {} resources", id);
-
-            let mut cache_range: Option<(u64, u64)> = None;
-            for resource in node.resources.iter() {
-                match resource {
-                    Resource::MmioAddressRange { base, size } => {
-                        if cache_range.is_some() {
-                            return Err(DeviceManagerError::ResourceAlreadyExists);
-                        }
-
-                        cache_range = Some((*base, *size));
-                    }
-                    _ => {
-                        error!("Unexpected resource {:?} for {}", resource, id);
-                    }
-                }
-            }
-
-            cache_range
-        } else {
-            None
-        };
-
-        // DAX is not supported, we override the config by disabling the option.
-        fs_cfg.dax = false;
-
         if let Some(fs_socket) = fs_cfg.socket.to_str() {
-            let cache = if fs_cfg.dax {
-                let (cache_base, cache_size) = if let Some((base, size)) = cache_range {
-                    // The memory needs to be 2MiB aligned in order to support
-                    // hugepages.
-                    self.pci_segments[fs_cfg.pci_segment as usize]
-                        .allocator
-                        .lock()
-                        .unwrap()
-                        .allocate(
-                            Some(GuestAddress(base)),
-                            size as GuestUsize,
-                            Some(0x0020_0000),
-                        )
-                        .ok_or(DeviceManagerError::FsRangeAllocation)?;
-
-                    (base, size)
-                } else {
-                    let size = fs_cfg.cache_size;
-                    // The memory needs to be 2MiB aligned in order to support
-                    // hugepages.
-                    let base = self.pci_segments[fs_cfg.pci_segment as usize]
-                        .allocator
-                        .lock()
-                        .unwrap()
-                        .allocate(None, size as GuestUsize, Some(0x0020_0000))
-                        .ok_or(DeviceManagerError::FsRangeAllocation)?;
-
-                    (base.raw_value(), size)
-                };
-
-                // Update the node with correct resource information.
-                node.resources.push(Resource::MmioAddressRange {
-                    base: cache_base,
-                    size: cache_size,
-                });
-
-                let mmap_region = MmapRegion::build(
-                    None,
-                    cache_size as usize,
-                    libc::PROT_NONE,
-                    libc::MAP_ANONYMOUS | libc::MAP_PRIVATE,
-                )
-                .map_err(DeviceManagerError::NewMmapRegion)?;
-                let host_addr: u64 = mmap_region.as_ptr() as u64;
-
-                let mem_slot = self
-                    .memory_manager
-                    .lock()
-                    .unwrap()
-                    .create_userspace_mapping(
-                        cache_base, cache_size, host_addr, false, false, false,
-                    )
-                    .map_err(DeviceManagerError::MemoryManager)?;
-
-                let region_list = vec![VirtioSharedMemory {
-                    offset: 0,
-                    len: cache_size,
-                }];
-
-                Some((
-                    VirtioSharedMemoryList {
-                        host_addr,
-                        mem_slot,
-                        addr: GuestAddress(cache_base),
-                        len: cache_size as GuestUsize,
-                        region_list,
-                    },
-                    mmap_region,
-                ))
-            } else {
-                None
-            };
-
             let virtio_fs_device = Arc::new(Mutex::new(
                 virtio_devices::vhost_user::Fs::new(
                     id.clone(),
@@ -2504,7 +2404,7 @@ impl DeviceManager {
                     &fs_cfg.tag,
                     fs_cfg.num_queues,
                     fs_cfg.queue_size,
-                    cache,
+                    None,
                     self.seccomp_action.clone(),
                     self.restoring,
                     self.exit_evt
@@ -3074,10 +2974,6 @@ impl DeviceManager {
         device_cfg: &mut DeviceConfig,
     ) -> DeviceManagerResult<(PciBdf, String)> {
         let vfio_name = if let Some(id) = &device_cfg.id {
-            if self.device_tree.lock().unwrap().contains_key(id) {
-                return Err(DeviceManagerError::DeviceIdAlreadyInUse);
-            }
-
             id.clone()
         } else {
             let id = self.next_device_name(VFIO_DEVICE_NAME_PREFIX)?;
@@ -3177,6 +3073,8 @@ impl DeviceManager {
                 None
             };
 
+        let memory_manager = self.memory_manager.clone();
+
         let vfio_pci_device = VfioPciDevice::new(
             vfio_name.clone(),
             &self.address_manager.vm,
@@ -3186,6 +3084,8 @@ impl DeviceManager {
             legacy_interrupt_group,
             device_cfg.iommu,
             pci_device_bdf,
+            self.restoring,
+            Arc::new(move || memory_manager.lock().unwrap().allocate_memory_slot()),
         )
         .map_err(DeviceManagerError::VfioPciCreate)?;
 
@@ -3199,15 +3099,17 @@ impl DeviceManager {
             resources,
         )?;
 
-        vfio_pci_device
+        // When restoring a VM, the restore codepath will take care of mapping
-            .lock()
+        // the MMIO regions based on the information from the snapshot.
-            .unwrap()
+        if !self.restoring {
-            .map_mmio_regions(&self.address_manager.vm, || {
+            vfio_pci_device
-                self.memory_manager.lock().unwrap().allocate_memory_slot()
+                .lock()
-            })
+                .unwrap()
-            .map_err(DeviceManagerError::VfioMapRegion)?;
+                .map_mmio_regions()
+                .map_err(DeviceManagerError::VfioMapRegion)?;
+        }
 
-        let mut node = device_node!(vfio_name);
+        let mut node = device_node!(vfio_name, vfio_pci_device);
 
         // Update the device tree with correct resource information.
         node.resources = new_resources;
@@ -3302,10 +3204,6 @@ impl DeviceManager {
         device_cfg: &mut UserDeviceConfig,
     ) -> DeviceManagerResult<(PciBdf, String)> {
         let vfio_user_name = if let Some(id) = &device_cfg.id {
-            if self.device_tree.lock().unwrap().contains_key(id) {
-                return Err(DeviceManagerError::DeviceIdAlreadyInUse);
-            }
-
             id.clone()
         } else {
             let id = self.next_device_name(VFIO_USER_DEVICE_NAME_PREFIX)?;
@@ -3336,6 +3234,8 @@ impl DeviceManager {
                 .map_err(DeviceManagerError::VfioUserCreateClient)?,
         ));
 
+        let memory_manager = self.memory_manager.clone();
+
         let mut vfio_user_pci_device = VfioUserPciDevice::new(
             vfio_user_name.clone(),
             &self.address_manager.vm,
@@ -3343,6 +3243,8 @@ impl DeviceManager {
             self.msi_interrupt_manager.clone(),
             legacy_interrupt_group,
             pci_device_bdf,
+            self.restoring,
+            Arc::new(move || memory_manager.lock().unwrap().allocate_memory_slot()),
         )
         .map_err(DeviceManagerError::VfioUserCreate)?;
 
@@ -3377,17 +3279,19 @@ impl DeviceManager {
             resources,
         )?;
 
-        // Note it is required to call 'add_pci_device()' in advance to have the list of
+        // When restoring a VM, the restore codepath will take care of mapping
-        // mmio regions provisioned correctly
+        // the MMIO regions based on the information from the snapshot.
-        vfio_user_pci_device
+        if !self.restoring {
-            .lock()
+            // Note it is required to call 'add_pci_device()' in advance to have the list of
-            .unwrap()
+            // mmio regions provisioned correctly
-            .map_mmio_regions(&self.address_manager.vm, || {
+            vfio_user_pci_device
-                self.memory_manager.lock().unwrap().allocate_memory_slot()
+                .lock()
-            })
+                .unwrap()
-            .map_err(DeviceManagerError::VfioUserMapRegion)?;
+                .map_mmio_regions()
+                .map_err(DeviceManagerError::VfioUserMapRegion)?;
+        }
 
-        let mut node = device_node!(vfio_user_name);
+        let mut node = device_node!(vfio_user_name, vfio_user_pci_device);
 
         // Update the device tree with correct resource information.
         node.resources = new_resources;
@@ -3520,6 +3424,7 @@ impl DeviceManager {
                 // The exception being if not on the default PCI segment.
                 pci_segment_id > 0 || device_type != VirtioDeviceType::Block as u32,
                 dma_handler,
+                self.pending_activations.clone(),
             )
             .map_err(DeviceManagerError::VirtioDevice)?,
         ));
@@ -3674,17 +3579,10 @@ impl DeviceManager {
     }
 
     pub fn activate_virtio_devices(&self) -> DeviceManagerResult<()> {
-        // Find virtio pci devices and activate any pending ones
+        for mut activator in self.pending_activations.lock().unwrap().drain(..) {
-        let device_tree = self.device_tree.lock().unwrap();
+            activator
-        for pci_device_node in device_tree.pci_devices() {
+                .activate()
-            #[allow(irrefutable_let_patterns)]
+                .map_err(DeviceManagerError::VirtioActivate)?;
-            if let PciDeviceHandle::Virtio(virtio_pci_device) = &pci_device_node
-                .pci_device_handle
-                .as_ref()
-                .ok_or(DeviceManagerError::MissingPciDevice)?
-            {
-                virtio_pci_device.lock().unwrap().maybe_activate();
-            }
         }
         Ok(())
     }
@@ -4427,26 +4325,11 @@ impl Pausable for DeviceManager {
         // and ITS tables to guest RAM.
         #[cfg(target_arch = "aarch64")]
         {
-            let gic_device = Arc::clone(
+            self.get_interrupt_controller()
-                self.get_interrupt_controller()
+                .unwrap()
-                    .unwrap()
-                    .lock()
-                    .unwrap()
-                    .get_gic_device()
-                    .unwrap(),
-            );
-            if let Some(gicv3_its) = gic_device
                 .lock()
                 .unwrap()
-                .as_any_concrete_mut()
+                .pause()?;
-                .downcast_mut::<KvmGicV3Its>()
-            {
-                gicv3_its.pause()?;
-            } else {
-                return Err(MigratableError::Pause(anyhow!(
-                    "GicDevice downcast to KvmGicV3Its failed when pausing device manager!"
-                )));
-            };
         };
 
         Ok(())