diff --git a/.github/workflows/main.yml b/.github/workflows/main.yml
index 5e80781daf..3885cd9f09 100644
--- a/.github/workflows/main.yml
+++ b/.github/workflows/main.yml
@@ -22,6 +22,23 @@ jobs:
     - run: rustup component add rustfmt
     - run: cargo fmt --all -- --check
 
+  # Lint dependency graph for security advisories, duplicate versions, and
+  # incompatible licences
+  cargo_deny:
+    name: Cargo deny
+    runs-on: ubuntu-latest
+    steps:
+    - uses: actions/checkout@v2
+      with:
+        submodules: true
+    - uses: ./.github/actions/install-rust
+    - run: |
+        set -e
+        curl -L https://github.com/EmbarkStudios/cargo-deny/releases/download/0.8.5/cargo-deny-0.8.5-x86_64-unknown-linux-musl.tar.gz | tar xzf -
+        mv cargo-deny-*-x86_64-unknown-linux-musl/cargo-deny cargo-deny
+        echo `pwd` >> $GITHUB_PATH
+    - run: cargo deny check
+
   # Build `mdBook` documentation for `wasmtime`, and upload it as a temporary
   # build artifact
   doc_book:
@@ -269,7 +286,7 @@ jobs:
 
     # Test debug (DWARF) related functionality.
     - run: |
-        sudo apt-get install -y gdb
+        sudo apt-get update && sudo apt-get install -y gdb
         cargo test test_debug_dwarf -- --ignored --test-threads 1
       if: matrix.os == 'ubuntu-latest'
       env:
diff --git a/Cargo.lock b/Cargo.lock
index 2630d250e8..dda83d6e01 100644
--- a/Cargo.lock
+++ b/Cargo.lock
@@ -17,9 +17,9 @@ checksum = "ee2a4ec343196209d6594e19543ae87a39f96d5534d7174822a3ad825dd6ed7e"
 
 [[package]]
 name = "ahash"
-version = "0.4.6"
+version = "0.4.7"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "f6789e291be47ace86a60303502173d84af8327e3627ecf334356ee0f87a164c"
+checksum = "739f4a8db6605981345c5654f3a85b056ce52f37a39d34da03f25bf2151ea16e"
 
 [[package]]
 name = "aho-corasick"
@@ -50,9 +50,9 @@ dependencies = [
 
 [[package]]
 name = "anyhow"
-version = "1.0.34"
+version = "1.0.35"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "bf8dcb5b4bbaa28653b647d8c77bd4ed40183b48882e130c1f1ffb73de069fd7"
+checksum = "2c0df63cb2955042487fad3aefd2c6e3ae7389ac5dc1beb28921de0b69f779d4"
 
 [[package]]
 name = "arbitrary"
@@ -106,12 +106,6 @@ dependencies = [
  "rustc-demangle",
 ]
 
-[[package]]
-name = "base64"
-version = "0.12.3"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "3441f0f7b02788e948e47f457ca01f1d7e6d92c693bc132c22b087d3141c03ff"
-
 [[package]]
 name = "base64"
 version = "0.13.0"
@@ -163,9 +157,9 @@ dependencies = [
 
 [[package]]
 name = "bit-vec"
-version = "0.6.2"
+version = "0.6.3"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "5f0dc55f2d8a1a85650ac47858bb001b4c0dd73d79e3c455a842925e68d29cd3"
+checksum = "349f9b6a179ed607305526ca489b34ad0a41aed5f7980fa90eb03160b69598fb"
 
 [[package]]
 name = "bitflags"
@@ -287,9 +281,9 @@ dependencies = [
 
 [[package]]
 name = "cc"
-version = "1.0.61"
+version = "1.0.66"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "ed67cbde08356238e75fc4656be4749481eeffb09e19f320a25237d5221c985d"
+checksum = "4c0496836a84f8d0495758516b8621a622beb77c0fed418570e50764093ced48"
 dependencies = [
  "jobserver",
 ]
@@ -330,9 +324,9 @@ dependencies = [
 
 [[package]]
 name = "clang-sys"
-version = "1.0.0"
+version = "1.0.3"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "9da1484c6a890e374ca5086062d4847e0a2c1e5eba9afa5d48c09e8eb39b2519"
+checksum = "0659001ab56b791be01d4b729c44376edc6718cf389a502e579b77b758f3296c"
 dependencies = [
  "glob",
  "libc",
@@ -356,9 +350,9 @@ dependencies = [
 
 [[package]]
 name = "cmake"
-version = "0.1.44"
+version = "0.1.45"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "0e56268c17a6248366d66d4a47a3381369d068cce8409bb1716ed77ea32163bb"
+checksum = "eb6210b637171dfba4cda12e579ac6dc73f5165ad56133e5d72ef3131f320855"
 dependencies = [
  "cc",
 ]
@@ -381,9 +375,9 @@ dependencies = [
 
 [[package]]
 name = "const_fn"
-version = "0.4.3"
+version = "0.4.4"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "c478836e029dcef17fb47c89023448c64f781a046e0300e257ad8225ae59afab"
+checksum = "cd51eab21ab4fd6a3bf889e2d0958c0a6e3a61ad04260325e919e652a2a62826"
 
 [[package]]
 name = "constant_time_eq"
@@ -657,7 +651,7 @@ dependencies = [
  "smallvec",
  "target-lexicon",
  "thiserror",
- "wasmparser 0.70.0",
+ "wasmparser",
  "wat",
 ]
 
@@ -677,7 +671,7 @@ source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "dca26ee1f8d361640700bde38b2c37d8c22b3ce2d360e1fc1c74ea4b0aa7d775"
 dependencies = [
  "cfg-if 1.0.0",
- "crossbeam-utils 0.8.0",
+ "crossbeam-utils",
 ]
 
 [[package]]
@@ -688,43 +682,31 @@ checksum = "94af6efb46fef72616855b036a624cf27ba656ffc9be1b9a3c931cfc7749a9a9"
 dependencies = [
  "cfg-if 1.0.0",
  "crossbeam-epoch",
- "crossbeam-utils 0.8.0",
+ "crossbeam-utils",
 ]
 
 [[package]]
 name = "crossbeam-epoch"
-version = "0.9.0"
+version = "0.9.1"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "ec0f606a85340376eef0d6d8fec399e6d4a544d648386c6645eb6d0653b27d9f"
+checksum = "a1aaa739f95311c2c7887a76863f500026092fb1dce0161dab577e559ef3569d"
 dependencies = [
  "cfg-if 1.0.0",
  "const_fn",
- "crossbeam-utils 0.8.0",
+ "crossbeam-utils",
  "lazy_static",
- "memoffset 0.5.6",
+ "memoffset",
  "scopeguard",
 ]
 
 [[package]]
 name = "crossbeam-utils"
-version = "0.7.2"
+version = "0.8.1"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "c3c7c73a2d1e9fc0886a08b93e98eb643461230d5f1925e4036204d5f2e261a8"
-dependencies = [
- "autocfg",
- "cfg-if 0.1.10",
- "lazy_static",
-]
-
-[[package]]
-name = "crossbeam-utils"
-version = "0.8.0"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "ec91540d98355f690a86367e566ecad2e9e579f230230eb7c21398372be73ea5"
+checksum = "02d96d1e189ef58269ebe5b97953da3274d83a93af647c2ddd6f9dab28cedb8d"
 dependencies = [
  "autocfg",
  "cfg-if 1.0.0",
- "const_fn",
  "lazy_static",
 ]
 
@@ -761,9 +743,9 @@ dependencies = [
 
 [[package]]
 name = "derive_utils"
-version = "0.10.0"
+version = "0.11.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "3df5480412da86cdf5d6b7f3b682422c84359ff7399aa658df1d15ee83244b1d"
+checksum = "64196eb9f551916167225134f1e8a90f0b5774331d3c900d6328fd94bafe3544"
 dependencies = [
  "proc-macro2",
  "quote",
@@ -805,6 +787,16 @@ dependencies = [
  "dirs-sys",
 ]
 
+[[package]]
+name = "dirs-next"
+version = "2.0.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "b98cf8ebf19c3d1b223e151f99a4f9f0690dca41414773390fc824184ac833e1"
+dependencies = [
+ "cfg-if 1.0.0",
+ "dirs-sys-next",
+]
+
 [[package]]
 name = "dirs-sys"
 version = "0.3.5"
@@ -886,9 +878,9 @@ dependencies = [
 
 [[package]]
 name = "env_logger"
-version = "0.8.1"
+version = "0.8.2"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "54532e3223c5af90a6a757c90b5c5521564b07e5e7a958681bcd2afad421cdcd"
+checksum = "f26ecb66b4bdca6c1409b40fb255eefc2bd4f6d135dab3c3124f80ffa2a9661e"
 dependencies = [
  "atty",
  "humantime 2.0.1",
@@ -1108,9 +1100,9 @@ checksum = "25a2bc672d1148e28034f176e01fffebb08b35768468cc954630da77a1449005"
 
 [[package]]
 name = "indexmap"
-version = "1.6.0"
+version = "1.6.1"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "55e2e4c765aa53a0424761bf9f41aa7a6ac1efa87238f59560640e27fca028f2"
+checksum = "4fb1fa934250de4de8aef298d81c729a7d33d8c239daa3a7575e6b92bfc7313b"
 dependencies = [
  "autocfg",
  "hashbrown",
@@ -1137,9 +1129,9 @@ checksum = "47be2f14c678be2fdcab04ab1171db51b2762ce6f0a8ee87c8dd4a04ed216135"
 
 [[package]]
 name = "iter-enum"
-version = "0.2.4"
+version = "0.2.5"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "2cdea9771bec3d95893f6c665a4fcd477af7858446a46bc2772f560534eee43b"
+checksum = "86a94bc12a53bf84b705acee29eb8697a5ea7b4587d836152499e5db0a6d52b9"
 dependencies = [
  "derive_utils",
  "quote",
@@ -1187,9 +1179,9 @@ checksum = "e2abad23fbc42b3700f2f279844dc832adb2b2eb069b2df918f455c4e18cc646"
 
 [[package]]
 name = "lazycell"
-version = "1.2.1"
+version = "1.3.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "b294d6fa9ee409a054354afc4352b0b9ef7ca222c69b8812cbea9e7d2bf3783f"
+checksum = "830d08ce1d1d941e6b30645f1a0eb5643013d835ce3779a5fc208261dbe10f55"
 
 [[package]]
 name = "leb128"
@@ -1205,9 +1197,9 @@ checksum = "1482821306169ec4d07f6aca392a4681f66c75c9918aa49641a2595db64053cb"
 
 [[package]]
 name = "libfuzzer-sys"
-version = "0.3.4"
+version = "0.3.5"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "ee8c42ab62f43795ed77a965ed07994c5584cdc94fd0ebf14b22ac1524077acc"
+checksum = "fcf184a4b6b274f82a5df6b357da6055d3e82272327bba281c28bbba6f1664ef"
 dependencies = [
  "arbitrary",
  "cc",
@@ -1215,11 +1207,11 @@ dependencies = [
 
 [[package]]
 name = "libloading"
-version = "0.6.3"
+version = "0.6.6"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "2443d8f0478b16759158b2f66d525991a05491138bc05814ef52a250148ef4f9"
+checksum = "e9367bdfa836b7e3cf895867f7a570283444da90562980ec2263d6e1569b16bc"
 dependencies = [
- "cfg-if 0.1.10",
+ "cfg-if 1.0.0",
  "winapi",
 ]
 
@@ -1237,13 +1229,13 @@ dependencies = [
  "iter-enum",
  "itertools",
  "lazy_static",
- "memoffset 0.6.1",
+ "memoffset",
  "more-asserts",
  "quickcheck",
  "smallvec",
  "thiserror",
  "typemap",
- "wasmparser 0.70.0",
+ "wasmparser",
  "wat",
 ]
 
@@ -1303,15 +1295,6 @@ dependencies = [
  "winapi",
 ]
 
-[[package]]
-name = "memoffset"
-version = "0.5.6"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "043175f069eda7b85febe4a74abbaeff828d9f8b448515d3151a14a3542811aa"
-dependencies = [
- "autocfg",
-]
-
 [[package]]
 name = "memoffset"
 version = "0.6.1"
@@ -1419,14 +1402,13 @@ checksum = "8d3b63360ec3cb337817c2dbd47ab4a0f170d285d8e5a2064600f3def1402397"
 dependencies = [
  "crc32fast",
  "indexmap",
- "wasmparser 0.57.0",
 ]
 
 [[package]]
 name = "once_cell"
-version = "1.4.1"
+version = "1.5.2"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "260e51e7efe62b592207e9e13a68e43692a7a279171d6ba57abd208bf23645ad"
+checksum = "13bd41f508810a131401606d54ac32a467c97172d74ba7662562ebba5ad07fa0"
 
 [[package]]
 name = "opaque-debug"
@@ -1504,7 +1486,7 @@ version = "0.66.0"
 dependencies = [
  "arbitrary",
  "bincode",
- "env_logger 0.8.1",
+ "env_logger 0.8.2",
  "fst",
  "log",
  "peepmatic",
@@ -1559,7 +1541,7 @@ dependencies = [
 name = "peepmatic-test"
 version = "0.2.0"
 dependencies = [
- "env_logger 0.8.1",
+ "env_logger 0.8.2",
  "log",
  "peepmatic",
  "peepmatic-runtime",
@@ -1582,9 +1564,9 @@ version = "0.68.0"
 
 [[package]]
 name = "pin-project-lite"
-version = "0.1.11"
+version = "0.2.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "c917123afa01924fc84bb20c4c03f004d9c38e5127e3c039bbf7f4b9c76a2f6b"
+checksum = "6b063f57ec186e6140e2b8b6921e5f1bd89c7356dda5b33acc5401203ca6131c"
 
 [[package]]
 name = "posish"
@@ -1670,9 +1652,9 @@ dependencies = [
 
 [[package]]
 name = "psm"
-version = "0.1.11"
+version = "0.1.12"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "96e0536f6528466dbbbbe6b986c34175a8d0ff25b794c4bacda22e068cd2f2c5"
+checksum = "3abf49e5417290756acfd26501536358560c4a5cc4a0934d390939acb3e7083a"
 dependencies = [
  "cc",
 ]
@@ -1795,7 +1777,7 @@ checksum = "9ab346ac5921dc62ffa9f89b7a773907511cdfa5490c572ae9be1be33e8afa4a"
 dependencies = [
  "crossbeam-channel",
  "crossbeam-deque",
- "crossbeam-utils 0.8.0",
+ "crossbeam-utils",
  "lazy_static",
  "num_cpus",
 ]
@@ -1888,14 +1870,14 @@ dependencies = [
 
 [[package]]
 name = "rust-argon2"
-version = "0.8.2"
+version = "0.8.3"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "9dab61250775933275e84053ac235621dfb739556d5c54a2f2e9313b7cf43a19"
+checksum = "4b18820d944b33caa75a71378964ac46f58517c92b6ae5f762636247c09e78fb"
 dependencies = [
- "base64 0.12.3",
+ "base64",
  "blake2b_simd",
  "constant_time_eq",
- "crossbeam-utils 0.7.2",
+ "crossbeam-utils",
 ]
 
 [[package]]
@@ -1995,18 +1977,18 @@ checksum = "388a1df253eca08550bef6c72392cfe7c30914bf41df5269b68cbd6ff8f570a3"
 
 [[package]]
 name = "serde"
-version = "1.0.117"
+version = "1.0.118"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "b88fa983de7720629c9387e9f517353ed404164b1e482c970a90c1a4aaf7dc1a"
+checksum = "06c64263859d87aa2eb554587e2d23183398d617427327cf2b3d0ed8c69e4800"
 dependencies = [
  "serde_derive",
 ]
 
 [[package]]
 name = "serde_derive"
-version = "1.0.117"
+version = "1.0.118"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "cbd1ae72adb44aab48f325a02444a5fc079349a8d804c1fc922aed3f7454c74e"
+checksum = "c84d3526699cd55261af4b941e4e725444df67aa4f9e6a3564f18030d12672df"
 dependencies = [
  "proc-macro2",
  "quote",
@@ -2015,9 +1997,9 @@ dependencies = [
 
 [[package]]
 name = "serde_json"
-version = "1.0.59"
+version = "1.0.60"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "dcac07dbffa1c65e7f816ab9eba78eb142c6d44410f4eeba1e26e4f5dfa56b95"
+checksum = "1500e84d27fe482ed1dc791a56eddc2f230046a040fa908c08bda1d9fb615779"
 dependencies = [
  "itoa",
  "ryu",
@@ -2026,9 +2008,9 @@ dependencies = [
 
 [[package]]
 name = "serde_test"
-version = "1.0.117"
+version = "1.0.118"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "d9a49e2f787c0fddfd5e758cbbd3cbf81c3a8d12ef091283c52d9e9b4d417d3c"
+checksum = "f7f3f8714511d29f60be0ea965bc784df1b6903da5bbac801df36b1bbc7b4880"
 dependencies = [
  "serde",
 ]
@@ -2058,11 +2040,11 @@ dependencies = [
 
 [[package]]
 name = "shellexpand"
-version = "2.0.0"
+version = "2.1.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "9a2b22262a9aaf9464d356f656fea420634f78c881c5eebd5ef5e66d8b9bc603"
+checksum = "83bdb7831b2d85ddf4a7b148aa19d0587eddbe8671a436b7bd1182eaad0f2829"
 dependencies = [
- "dirs",
+ "dirs-next",
 ]
 
 [[package]]
@@ -2073,9 +2055,9 @@ checksum = "7fdf1b9db47230893d76faad238fd6097fd6d6a9245cd7a4d90dbd639536bbd2"
 
 [[package]]
 name = "smallvec"
-version = "1.4.2"
+version = "1.5.1"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "fbee7696b84bbf3d89a1c2eccff0850e3047ed46bfcd2e92c29a2d074d57e252"
+checksum = "ae524f056d7d770e174287294f562e95044c68e88dec909a00d2094805db9d75"
 
 [[package]]
 name = "souper-ir"
@@ -2100,9 +2082,9 @@ checksum = "8ea5119cdb4c55b55d432abb513a0429384878c15dde60cc77b1c99de1a95a6a"
 
 [[package]]
 name = "structopt"
-version = "0.3.20"
+version = "0.3.21"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "126d630294ec449fae0b16f964e35bf3c74f940da9dca17ee9b905f7b3112eb8"
+checksum = "5277acd7ee46e63e5168a80734c9f6ee81b1367a7d8772a2d765df2a3705d28c"
 dependencies = [
  "clap",
  "lazy_static",
@@ -2111,9 +2093,9 @@ dependencies = [
 
 [[package]]
 name = "structopt-derive"
-version = "0.4.13"
+version = "0.4.14"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "65e51c492f9e23a220534971ff5afc14037289de430e3c83f9daf6a1b6ae91e8"
+checksum = "5ba9cdfda491b814720b6b06e0cac513d922fc407582032e8706e9f137976f90"
 dependencies = [
  "heck",
  "proc-macro-error",
@@ -2124,9 +2106,9 @@ dependencies = [
 
 [[package]]
 name = "syn"
-version = "1.0.48"
+version = "1.0.54"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "cc371affeffc477f42a221a1e4297aedcea33d47d19b61455588bd9d8f6b19ac"
+checksum = "9a2af957a63d6bd42255c359c93d9bfdb97076bd3b820897ce55ffbfbf107f44"
 dependencies = [
  "proc-macro2",
  "quote",
@@ -2177,18 +2159,18 @@ dependencies = [
 
 [[package]]
 name = "termcolor"
-version = "1.1.0"
+version = "1.1.2"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "bb6bfa289a4d7c5766392812c0a1f4c1ba45afa1ad47803c11e1f407d846d75f"
+checksum = "2dfed899f0eb03f32ee8c6a0aabdb8a7949659e3466561fc0adf54e26d88c5f4"
 dependencies = [
  "winapi-util",
 ]
 
 [[package]]
 name = "terminal_size"
-version = "0.1.13"
+version = "0.1.15"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "9a14cd9f8c72704232f0bfc8455c0e861f0ad4eb60cc9ec8a170e231414c1e13"
+checksum = "4bd2d183bd3fac5f5fe38ddbeb4dc9aec4a39a9d7d59e7491d900302da01cbe1"
 dependencies = [
  "libc",
  "winapi",
@@ -2271,11 +2253,11 @@ dependencies = [
 
 [[package]]
 name = "tracing"
-version = "0.1.21"
+version = "0.1.22"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "b0987850db3733619253fe60e17cb59b82d37c7e6c0236bb81e4d6b87c879f27"
+checksum = "9f47026cdc4080c07e49b37087de021820269d996f581aac150ef9e5583eefe3"
 dependencies = [
- "cfg-if 0.1.10",
+ "cfg-if 1.0.0",
  "log",
  "pin-project-lite",
  "tracing-attributes",
@@ -2368,9 +2350,9 @@ checksum = "373c8a200f9e67a0c95e62a4f52fbf80c23b4381c05a17845531982fa99e6b33"
 
 [[package]]
 name = "unicode-segmentation"
-version = "1.6.0"
+version = "1.7.1"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "e83e153d1053cbb5a118eeff7fd5be06ed99153f00dbcd8ae310c5fb2b22edc0"
+checksum = "bb0d2e7be6ae3a5fa87eed5fb451aff96f2573d2694942e40543ae0bbe19c796"
 
 [[package]]
 name = "unicode-width"
@@ -2531,24 +2513,18 @@ dependencies = [
 
 [[package]]
 name = "wasmparser"
-version = "0.57.0"
+version = "0.71.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "32fddd575d477c6e9702484139cf9f23dcd554b06d185ed0f56c857dd3a47aa6"
-
-[[package]]
-name = "wasmparser"
-version = "0.70.0"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "ed1b3f9e9cf01a580b9f3281214dfdb1922b5dfb8494ee312ca03ae10036c2a2"
+checksum = "89a30c99437829ede826802bfcf28500cf58df00e66cb9114df98813bc145ff1"
 
 [[package]]
 name = "wasmprinter"
-version = "0.2.17"
+version = "0.2.18"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "f89b2b24dce17e27fe9c09c28331cbd77067fcde5c6ea2508ac84bcbd5d3e018"
+checksum = "0515db67c610037f3c53ec36976edfd1eb01bac6b1226914b17ce609480e729f"
 dependencies = [
  "anyhow",
- "wasmparser 0.70.0",
+ "wasmparser",
 ]
 
 [[package]]
@@ -2569,7 +2545,7 @@ dependencies = [
  "smallvec",
  "target-lexicon",
  "tempfile",
- "wasmparser 0.70.0",
+ "wasmparser",
  "wasmtime-cache",
  "wasmtime-environ",
  "wasmtime-jit",
@@ -2580,12 +2556,23 @@ dependencies = [
  "winapi",
 ]
 
+[[package]]
+name = "wasmtime-bench-api"
+version = "0.19.0"
+dependencies = [
+ "anyhow",
+ "wasi-common",
+ "wasmtime",
+ "wasmtime-wasi",
+ "wat",
+]
+
 [[package]]
 name = "wasmtime-c-api"
 version = "0.19.0"
 dependencies = [
  "anyhow",
- "env_logger 0.8.1",
+ "env_logger 0.8.2",
  "once_cell",
  "wasi-common",
  "wasmtime",
@@ -2607,7 +2594,7 @@ name = "wasmtime-cache"
 version = "0.21.0"
 dependencies = [
  "anyhow",
- "base64 0.13.0",
+ "base64",
  "bincode",
  "directories-next",
  "errno",
@@ -2631,7 +2618,7 @@ name = "wasmtime-cli"
 version = "0.21.0"
 dependencies = [
  "anyhow",
- "env_logger 0.8.1",
+ "env_logger 0.8.2",
  "file-per-thread-logger",
  "filecheck",
  "humantime 2.0.1",
@@ -2647,7 +2634,7 @@ dependencies = [
  "test-programs",
  "tracing-subscriber",
  "wasi-common",
- "wasmparser 0.70.0",
+ "wasmparser",
  "wasmtime",
  "wasmtime-cache",
  "wasmtime-debug",
@@ -2683,7 +2670,7 @@ dependencies = [
  "object",
  "target-lexicon",
  "thiserror",
- "wasmparser 0.70.0",
+ "wasmparser",
  "wasmtime-environ",
 ]
 
@@ -2702,7 +2689,7 @@ dependencies = [
  "more-asserts",
  "serde",
  "thiserror",
- "wasmparser 0.70.0",
+ "wasmparser",
 ]
 
 [[package]]
@@ -2726,12 +2713,13 @@ version = "0.19.0"
 dependencies = [
  "anyhow",
  "arbitrary",
- "env_logger 0.8.1",
+ "env_logger 0.8.2",
  "log",
  "rayon",
+ "wasm-encoder",
  "wasm-smith",
  "wasmi",
- "wasmparser 0.70.0",
+ "wasmparser",
  "wasmprinter",
  "wasmtime",
  "wasmtime-wast",
@@ -2759,7 +2747,7 @@ dependencies = [
  "serde",
  "target-lexicon",
  "thiserror",
- "wasmparser 0.70.0",
+ "wasmparser",
  "wasmtime-cranelift",
  "wasmtime-debug",
  "wasmtime-environ",
@@ -2776,7 +2764,7 @@ version = "0.21.0"
 dependencies = [
  "cranelift-codegen",
  "lightbeam",
- "wasmparser 0.70.0",
+ "wasmparser",
  "wasmtime-environ",
 ]
 
@@ -2821,7 +2809,7 @@ dependencies = [
  "lazy_static",
  "libc",
  "log",
- "memoffset 0.6.1",
+ "memoffset",
  "more-asserts",
  "psm",
  "region",
@@ -2927,12 +2915,21 @@ dependencies = [
 ]
 
 [[package]]
-name = "wat"
-version = "1.0.30"
+name = "wast"
+version = "30.0.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "0d11a88d953b298172d218d18f22853f4e6e12873b62755d05617b864d312c68"
+checksum = "9b79907b22f740634810e882d8d1d9d0f9563095a8ab94e786e370242bff5cd2"
 dependencies = [
- "wast 29.0.0",
+ "leb128",
+]
+
+[[package]]
+name = "wat"
+version = "1.0.31"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "a8279a02835bf12e61ed2b3c3cbc6ecf9918762fd97e036917c11a09ec20ca44"
+dependencies = [
+ "wast 30.0.0",
 ]
 
 [[package]]
@@ -2991,7 +2988,7 @@ dependencies = [
 name = "wiggle-test"
 version = "0.21.0"
 dependencies = [
- "env_logger 0.8.1",
+ "env_logger 0.8.2",
  "proptest",
  "thiserror",
  "tracing",
@@ -3097,18 +3094,18 @@ dependencies = [
 
 [[package]]
 name = "zstd"
-version = "0.5.3+zstd.1.4.5"
+version = "0.5.4+zstd.1.4.7"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "01b32eaf771efa709e8308605bbf9319bf485dc1503179ec0469b611937c0cd8"
+checksum = "69996ebdb1ba8b1517f61387a883857818a66c8a295f487b1ffd8fd9d2c82910"
 dependencies = [
  "zstd-safe",
 ]
 
 [[package]]
 name = "zstd-safe"
-version = "2.0.5+zstd.1.4.5"
+version = "2.0.6+zstd.1.4.7"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "1cfb642e0d27f64729a639c52db457e0ae906e7bc6f5fe8f5c453230400f1055"
+checksum = "98aa931fb69ecee256d44589d19754e61851ae4769bf963b385119b1cc37a49e"
 dependencies = [
  "libc",
  "zstd-sys",
@@ -3116,9 +3113,9 @@ dependencies = [
 
 [[package]]
 name = "zstd-sys"
-version = "1.4.17+zstd.1.4.5"
+version = "1.4.18+zstd.1.4.7"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "b89249644df056b522696b1bb9e7c18c87e8ffa3e2f0dc3b0155875d6498f01b"
+checksum = "a1e6e8778706838f43f771d80d37787cb2fe06dafe89dd3aebaf6721b9eaec81"
 dependencies = [
  "cc",
  "glob",
diff --git a/Cargo.toml b/Cargo.toml
index 7741c23e67..f06fca065d 100644
--- a/Cargo.toml
+++ b/Cargo.toml
@@ -43,7 +43,7 @@ libc = "0.2.60"
 log = "0.4.8"
 rayon = "1.2.1"
 humantime = "2.0.0"
-wasmparser = "0.70.0"
+wasmparser = "0.71.0"
 
 [dev-dependencies]
 env_logger = "0.8.1"
@@ -64,6 +64,7 @@ opt-level = 0
 [workspace]
 members = [
   "cranelift",
+  "crates/bench-api",
   "crates/c-api",
   "crates/fuzzing",
   "crates/misc/run-examples",
diff --git a/cranelift/codegen/meta/src/isa/x86/encodings.rs b/cranelift/codegen/meta/src/isa/x86/encodings.rs
index de48c57c5d..bdfcef9c6f 100644
--- a/cranelift/codegen/meta/src/isa/x86/encodings.rs
+++ b/cranelift/codegen/meta/src/isa/x86/encodings.rs
@@ -1691,6 +1691,7 @@ fn define_simd(
     let usub_sat = shared.by_name("usub_sat");
     let vconst = shared.by_name("vconst");
     let vselect = shared.by_name("vselect");
+    let widening_pairwise_dot_product_s = shared.by_name("widening_pairwise_dot_product_s");
     let x86_cvtt2si = x86.by_name("x86_cvtt2si");
     let x86_insertps = x86.by_name("x86_insertps");
     let x86_fmax = x86.by_name("x86_fmax");
@@ -2213,6 +2214,9 @@ fn define_simd(
     // SIMD multiplication with lane expansion.
     e.enc_both_inferred(x86_pmuludq, rec_fa.opcodes(&PMULUDQ));
 
+    // SIMD multiplication and add adjacent pairs, from SSE2.
+    e.enc_both_inferred(widening_pairwise_dot_product_s, rec_fa.opcodes(&PMADDWD));
+
     // SIMD integer multiplication for I64x2 using a AVX512.
     {
         e.enc_32_64_maybe_isap(
diff --git a/cranelift/codegen/meta/src/isa/x86/opcodes.rs b/cranelift/codegen/meta/src/isa/x86/opcodes.rs
index 595d13ba2b..2e72a1744d 100644
--- a/cranelift/codegen/meta/src/isa/x86/opcodes.rs
+++ b/cranelift/codegen/meta/src/isa/x86/opcodes.rs
@@ -508,6 +508,9 @@ pub static VPMULLQ: [u8; 4] = [0x66, 0x0f, 0x38, 0x40];
 /// in xmm2/m128, and store the quadword results in xmm1 (SSE2).
 pub static PMULUDQ: [u8; 3] = [0x66, 0x0f, 0xf4];
 
+/// Multiply the packed word integers, add adjacent doubleword results.
+pub static PMADDWD: [u8; 3] = [0x66, 0x0f, 0xf5];
+
 /// Pop top of stack into r{16,32,64}; increment stack pointer.
 pub static POP_REG: [u8; 1] = [0x58];
 
diff --git a/cranelift/codegen/src/isa/mod.rs b/cranelift/codegen/src/isa/mod.rs
index 1ce10a155e..900dfd9ddd 100644
--- a/cranelift/codegen/src/isa/mod.rs
+++ b/cranelift/codegen/src/isa/mod.rs
@@ -75,10 +75,13 @@ use thiserror::Error;
 #[cfg(feature = "riscv")]
 mod riscv;
 
-// Exclude the old x86 backend when the new one is enabled; it is unreachable
-// anyway (requesting the x86-64 ISA will return the new backend), and a number
-// of old-backend-specific tests fail.
-#[cfg(all(feature = "x86", not(feature = "x64")))]
+// N.B.: the old x86-64 backend (`x86`) and the new one (`x64`) can both be
+// included; if the new backend is included, then it is the default backend
+// returned for an x86-64 triple, but a specific option can request the old
+// backend. It is important to have the ability to instantiate *both* backends
+// in the same build so that we can do things like differential fuzzing between
+// backends, or perhaps offer a runtime configuration flag in the future.
+#[cfg(feature = "x86")]
 mod x86;
 
 #[cfg(feature = "x64")]
@@ -117,24 +120,56 @@ macro_rules! isa_builder {
     }};
 }
 
+/// The "variant" for a given target. On one platform (x86-64), we have two
+/// backends, the "old" and "new" one; the new one is the default if included
+/// in the build configuration and not otherwise specified.
+#[derive(Clone, Copy)]
+pub enum BackendVariant {
+    /// Any backend available.
+    Any,
+    /// A "legacy" backend: one that operates using legalizations and encodings.
+    Legacy,
+    /// A backend built on `MachInst`s and the `VCode` framework.
+    MachInst,
+}
+
+impl Default for BackendVariant {
+    fn default() -> Self {
+        BackendVariant::Any
+    }
+}
+
+/// Look for an ISA for the given `triple`, selecting the backend variant given
+/// by `variant` if available.
+pub fn lookup_variant(triple: Triple, variant: BackendVariant) -> Result<Builder, LookupError> {
+    match (triple.architecture, variant) {
+        (Architecture::Riscv32 { .. }, _) | (Architecture::Riscv64 { .. }, _) => {
+            isa_builder!(riscv, (feature = "riscv"), triple)
+        }
+        (Architecture::X86_64, BackendVariant::Legacy) => {
+            isa_builder!(x86, (feature = "x86"), triple)
+        }
+        (Architecture::X86_64, BackendVariant::MachInst) => {
+            isa_builder!(x64, (feature = "x64"), triple)
+        }
+        #[cfg(feature = "x64")]
+        (Architecture::X86_64, BackendVariant::Any) => {
+            isa_builder!(x64, (feature = "x64"), triple)
+        }
+        #[cfg(not(feature = "x64"))]
+        (Architecture::X86_64, BackendVariant::Any) => {
+            isa_builder!(x86, (feature = "x86"), triple)
+        }
+        (Architecture::Arm { .. }, _) => isa_builder!(arm32, (feature = "arm32"), triple),
+        (Architecture::Aarch64 { .. }, _) => isa_builder!(aarch64, (feature = "arm64"), triple),
+        _ => Err(LookupError::Unsupported),
+    }
+}
+
 /// Look for an ISA for the given `triple`.
 /// Return a builder that can create a corresponding `TargetIsa`.
 pub fn lookup(triple: Triple) -> Result<Builder, LookupError> {
-    match triple.architecture {
-        Architecture::Riscv32 { .. } | Architecture::Riscv64 { .. } => {
-            isa_builder!(riscv, (feature = "riscv"), triple)
-        }
-        Architecture::X86_32 { .. } | Architecture::X86_64 => {
-            if cfg!(feature = "x64") {
-                isa_builder!(x64, (feature = "x64"), triple)
-            } else {
-                isa_builder!(x86, (all(feature = "x86", not(feature = "x64"))), triple)
-            }
-        }
-        Architecture::Arm { .. } => isa_builder!(arm32, (feature = "arm32"), triple),
-        Architecture::Aarch64 { .. } => isa_builder!(aarch64, (feature = "arm64"), triple),
-        _ => Err(LookupError::Unsupported),
-    }
+    lookup_variant(triple, BackendVariant::Any)
 }
 
 /// Look for a supported ISA with the given `name`.
diff --git a/cranelift/codegen/src/isa/x64/inst/args.rs b/cranelift/codegen/src/isa/x64/inst/args.rs
index 7e3b3f22a2..4c61954630 100644
--- a/cranelift/codegen/src/isa/x64/inst/args.rs
+++ b/cranelift/codegen/src/isa/x64/inst/args.rs
@@ -498,6 +498,7 @@ pub enum SseOpcode {
     Pinsrb,
     Pinsrw,
     Pinsrd,
+    Pmaddwd,
     Pmaxsb,
     Pmaxsw,
     Pmaxsd,
@@ -661,6 +662,7 @@ impl SseOpcode {
             | SseOpcode::Pcmpgtd
             | SseOpcode::Pextrw
             | SseOpcode::Pinsrw
+            | SseOpcode::Pmaddwd
             | SseOpcode::Pmaxsw
             | SseOpcode::Pmaxub
             | SseOpcode::Pminsw
@@ -842,6 +844,7 @@ impl fmt::Debug for SseOpcode {
             SseOpcode::Pinsrb => "pinsrb",
             SseOpcode::Pinsrw => "pinsrw",
             SseOpcode::Pinsrd => "pinsrd",
+            SseOpcode::Pmaddwd => "pmaddwd",
             SseOpcode::Pmaxsb => "pmaxsb",
             SseOpcode::Pmaxsw => "pmaxsw",
             SseOpcode::Pmaxsd => "pmaxsd",
diff --git a/cranelift/codegen/src/isa/x64/inst/emit.rs b/cranelift/codegen/src/isa/x64/inst/emit.rs
index 401b8aad08..fb32635a92 100644
--- a/cranelift/codegen/src/isa/x64/inst/emit.rs
+++ b/cranelift/codegen/src/isa/x64/inst/emit.rs
@@ -1764,6 +1764,18 @@ pub(crate) fn emit(
                 SseOpcode::Pabsb => (LegacyPrefixes::_66, 0x0F381C, 3),
                 SseOpcode::Pabsw => (LegacyPrefixes::_66, 0x0F381D, 3),
                 SseOpcode::Pabsd => (LegacyPrefixes::_66, 0x0F381E, 3),
+                SseOpcode::Pmovsxbd => (LegacyPrefixes::_66, 0x0F3821, 3),
+                SseOpcode::Pmovsxbw => (LegacyPrefixes::_66, 0x0F3820, 3),
+                SseOpcode::Pmovsxbq => (LegacyPrefixes::_66, 0x0F3822, 3),
+                SseOpcode::Pmovsxwd => (LegacyPrefixes::_66, 0x0F3823, 3),
+                SseOpcode::Pmovsxwq => (LegacyPrefixes::_66, 0x0F3824, 3),
+                SseOpcode::Pmovsxdq => (LegacyPrefixes::_66, 0x0F3825, 3),
+                SseOpcode::Pmovzxbd => (LegacyPrefixes::_66, 0x0F3831, 3),
+                SseOpcode::Pmovzxbw => (LegacyPrefixes::_66, 0x0F3830, 3),
+                SseOpcode::Pmovzxbq => (LegacyPrefixes::_66, 0x0F3832, 3),
+                SseOpcode::Pmovzxwd => (LegacyPrefixes::_66, 0x0F3833, 3),
+                SseOpcode::Pmovzxwq => (LegacyPrefixes::_66, 0x0F3834, 3),
+                SseOpcode::Pmovzxdq => (LegacyPrefixes::_66, 0x0F3835, 3),
                 SseOpcode::Sqrtps => (LegacyPrefixes::None, 0x0F51, 2),
                 SseOpcode::Sqrtpd => (LegacyPrefixes::_66, 0x0F51, 2),
                 SseOpcode::Sqrtss => (LegacyPrefixes::_F3, 0x0F51, 2),
@@ -1861,18 +1873,7 @@ pub(crate) fn emit(
                 SseOpcode::Pcmpgtw => (LegacyPrefixes::_66, 0x0F65, 2),
                 SseOpcode::Pcmpgtd => (LegacyPrefixes::_66, 0x0F66, 2),
                 SseOpcode::Pcmpgtq => (LegacyPrefixes::_66, 0x0F3837, 3),
-                SseOpcode::Pmovsxbd => (LegacyPrefixes::_66, 0x0F3821, 3),
-                SseOpcode::Pmovsxbw => (LegacyPrefixes::_66, 0x0F3820, 3),
-                SseOpcode::Pmovsxbq => (LegacyPrefixes::_66, 0x0F3822, 3),
-                SseOpcode::Pmovsxwd => (LegacyPrefixes::_66, 0x0F3823, 3),
-                SseOpcode::Pmovsxwq => (LegacyPrefixes::_66, 0x0F3824, 3),
-                SseOpcode::Pmovsxdq => (LegacyPrefixes::_66, 0x0F3825, 3),
-                SseOpcode::Pmovzxbd => (LegacyPrefixes::_66, 0x0F3831, 3),
-                SseOpcode::Pmovzxbw => (LegacyPrefixes::_66, 0x0F3830, 3),
-                SseOpcode::Pmovzxbq => (LegacyPrefixes::_66, 0x0F3832, 3),
-                SseOpcode::Pmovzxwd => (LegacyPrefixes::_66, 0x0F3833, 3),
-                SseOpcode::Pmovzxwq => (LegacyPrefixes::_66, 0x0F3834, 3),
-                SseOpcode::Pmovzxdq => (LegacyPrefixes::_66, 0x0F3835, 3),
+                SseOpcode::Pmaddwd => (LegacyPrefixes::_66, 0x0FF5, 2),
                 SseOpcode::Pmaxsb => (LegacyPrefixes::_66, 0x0F383C, 3),
                 SseOpcode::Pmaxsw => (LegacyPrefixes::_66, 0x0FEE, 2),
                 SseOpcode::Pmaxsd => (LegacyPrefixes::_66, 0x0F383D, 3),
diff --git a/cranelift/codegen/src/isa/x64/inst/emit_tests.rs b/cranelift/codegen/src/isa/x64/inst/emit_tests.rs
index bda26e3f27..e2afa80e2a 100644
--- a/cranelift/codegen/src/isa/x64/inst/emit_tests.rs
+++ b/cranelift/codegen/src/isa/x64/inst/emit_tests.rs
@@ -10,7 +10,7 @@
 //! RUST_BACKTRACE=1 cargo test --features test-programs/test_programs \
 //!   --features experimental_x64 --all --exclude peepmatic --exclude lightbeam \
 //!   --exclude wasmtime-lightbeam --exclude peepmatic-automata --exclude peepmatic-fuzzing \
-//!  --exclude peepmatic-macro -- isa::x64::inst::emit_tests::test_x64_emit
+//!   --exclude peepmatic-macro --exclude wasmtime-wasi-nn -- isa::x64::inst::emit_tests::test_x64_emit
 
 use super::*;
 use crate::isa::test_utils;
@@ -3067,6 +3067,12 @@ fn test_x64_emit() {
         "pmuludq %xmm8, %xmm9",
     ));
 
+    insns.push((
+        Inst::xmm_rm_r(SseOpcode::Pmaddwd, RegMem::reg(xmm8), w_xmm1),
+        "66410FF5C8",
+        "pmaddwd %xmm8, %xmm1",
+    ));
+
     insns.push((
         Inst::xmm_rm_r(SseOpcode::Pmaxsb, RegMem::reg(xmm15), w_xmm6),
         "66410F383CF7",
@@ -3201,81 +3207,6 @@ fn test_x64_emit() {
         "cvttps2dq %xmm9, %xmm8",
     ));
 
-    // ========================================================
-    // XMM_RM_R: Packed Move
-
-    insns.push((
-        Inst::xmm_rm_r(SseOpcode::Pmovsxbd, RegMem::reg(xmm6), w_xmm8),
-        "66440F3821C6",
-        "pmovsxbd %xmm6, %xmm8",
-    ));
-
-    insns.push((
-        Inst::xmm_rm_r(SseOpcode::Pmovsxbw, RegMem::reg(xmm9), w_xmm10),
-        "66450F3820D1",
-        "pmovsxbw %xmm9, %xmm10",
-    ));
-
-    insns.push((
-        Inst::xmm_rm_r(SseOpcode::Pmovsxbq, RegMem::reg(xmm1), w_xmm1),
-        "660F3822C9",
-        "pmovsxbq %xmm1, %xmm1",
-    ));
-
-    insns.push((
-        Inst::xmm_rm_r(SseOpcode::Pmovsxwd, RegMem::reg(xmm13), w_xmm10),
-        "66450F3823D5",
-        "pmovsxwd %xmm13, %xmm10",
-    ));
-
-    insns.push((
-        Inst::xmm_rm_r(SseOpcode::Pmovsxwq, RegMem::reg(xmm12), w_xmm12),
-        "66450F3824E4",
-        "pmovsxwq %xmm12, %xmm12",
-    ));
-
-    insns.push((
-        Inst::xmm_rm_r(SseOpcode::Pmovsxdq, RegMem::reg(xmm10), w_xmm8),
-        "66450F3825C2",
-        "pmovsxdq %xmm10, %xmm8",
-    ));
-
-    insns.push((
-        Inst::xmm_rm_r(SseOpcode::Pmovzxbd, RegMem::reg(xmm5), w_xmm6),
-        "660F3831F5",
-        "pmovzxbd %xmm5, %xmm6",
-    ));
-
-    insns.push((
-        Inst::xmm_rm_r(SseOpcode::Pmovzxbw, RegMem::reg(xmm5), w_xmm13),
-        "66440F3830ED",
-        "pmovzxbw %xmm5, %xmm13",
-    ));
-
-    insns.push((
-        Inst::xmm_rm_r(SseOpcode::Pmovzxbq, RegMem::reg(xmm10), w_xmm11),
-        "66450F3832DA",
-        "pmovzxbq %xmm10, %xmm11",
-    ));
-
-    insns.push((
-        Inst::xmm_rm_r(SseOpcode::Pmovzxwd, RegMem::reg(xmm2), w_xmm10),
-        "66440F3833D2",
-        "pmovzxwd %xmm2, %xmm10",
-    ));
-
-    insns.push((
-        Inst::xmm_rm_r(SseOpcode::Pmovzxwq, RegMem::reg(xmm7), w_xmm4),
-        "660F3834E7",
-        "pmovzxwq %xmm7, %xmm4",
-    ));
-
-    insns.push((
-        Inst::xmm_rm_r(SseOpcode::Pmovzxdq, RegMem::reg(xmm3), w_xmm4),
-        "660F3835E3",
-        "pmovzxdq %xmm3, %xmm4",
-    ));
-
     // XMM_Mov_R_M: float stores
     insns.push((
         Inst::xmm_mov_r_m(SseOpcode::Movss, xmm15, Amode::imm_reg(128, r12)),
@@ -3288,6 +3219,81 @@ fn test_x64_emit() {
         "movsd   %xmm1, 0(%rsi)",
     ));
 
+    // ========================================================
+    // XMM_MOV: Packed Move
+
+    insns.push((
+        Inst::xmm_mov(SseOpcode::Pmovsxbd, RegMem::reg(xmm6), w_xmm8),
+        "66440F3821C6",
+        "pmovsxbd %xmm6, %xmm8",
+    ));
+
+    insns.push((
+        Inst::xmm_mov(SseOpcode::Pmovsxbw, RegMem::reg(xmm9), w_xmm10),
+        "66450F3820D1",
+        "pmovsxbw %xmm9, %xmm10",
+    ));
+
+    insns.push((
+        Inst::xmm_mov(SseOpcode::Pmovsxbq, RegMem::reg(xmm1), w_xmm1),
+        "660F3822C9",
+        "pmovsxbq %xmm1, %xmm1",
+    ));
+
+    insns.push((
+        Inst::xmm_mov(SseOpcode::Pmovsxwd, RegMem::reg(xmm13), w_xmm10),
+        "66450F3823D5",
+        "pmovsxwd %xmm13, %xmm10",
+    ));
+
+    insns.push((
+        Inst::xmm_mov(SseOpcode::Pmovsxwq, RegMem::reg(xmm12), w_xmm12),
+        "66450F3824E4",
+        "pmovsxwq %xmm12, %xmm12",
+    ));
+
+    insns.push((
+        Inst::xmm_mov(SseOpcode::Pmovsxdq, RegMem::reg(xmm10), w_xmm8),
+        "66450F3825C2",
+        "pmovsxdq %xmm10, %xmm8",
+    ));
+
+    insns.push((
+        Inst::xmm_mov(SseOpcode::Pmovzxbd, RegMem::reg(xmm5), w_xmm6),
+        "660F3831F5",
+        "pmovzxbd %xmm5, %xmm6",
+    ));
+
+    insns.push((
+        Inst::xmm_mov(SseOpcode::Pmovzxbw, RegMem::reg(xmm5), w_xmm13),
+        "66440F3830ED",
+        "pmovzxbw %xmm5, %xmm13",
+    ));
+
+    insns.push((
+        Inst::xmm_mov(SseOpcode::Pmovzxbq, RegMem::reg(xmm10), w_xmm11),
+        "66450F3832DA",
+        "pmovzxbq %xmm10, %xmm11",
+    ));
+
+    insns.push((
+        Inst::xmm_mov(SseOpcode::Pmovzxwd, RegMem::reg(xmm2), w_xmm10),
+        "66440F3833D2",
+        "pmovzxwd %xmm2, %xmm10",
+    ));
+
+    insns.push((
+        Inst::xmm_mov(SseOpcode::Pmovzxwq, RegMem::reg(xmm7), w_xmm4),
+        "660F3834E7",
+        "pmovzxwq %xmm7, %xmm4",
+    ));
+
+    insns.push((
+        Inst::xmm_mov(SseOpcode::Pmovzxdq, RegMem::reg(xmm3), w_xmm4),
+        "660F3835E3",
+        "pmovzxdq %xmm3, %xmm4",
+    ));
+
     // XmmUnary: moves and unary float ops
     insns.push((
         Inst::xmm_unary_rm_r(SseOpcode::Movss, RegMem::reg(xmm13), w_xmm2),
diff --git a/cranelift/codegen/src/isa/x64/lower.rs b/cranelift/codegen/src/isa/x64/lower.rs
index a01e35bc0d..6e6198c44b 100644
--- a/cranelift/codegen/src/isa/x64/lower.rs
+++ b/cranelift/codegen/src/isa/x64/lower.rs
@@ -2235,6 +2235,24 @@ fn lower_insn_to_regs<C: LowerCtx<I = Inst>>(
             }
         }
 
+        Opcode::WideningPairwiseDotProductS => {
+            let lhs = put_input_in_reg(ctx, inputs[0]);
+            let rhs = input_to_reg_mem(ctx, inputs[1]);
+            let dst = get_output_reg(ctx, outputs[0]);
+            let ty = ty.unwrap();
+
+            ctx.emit(Inst::gen_move(dst, lhs, ty));
+
+            if ty == types::I32X4 {
+                ctx.emit(Inst::xmm_rm_r(SseOpcode::Pmaddwd, rhs, dst));
+            } else {
+                panic!(
+                    "Opcode::WideningPairwiseDotProductS: unsupported laneage: {:?}",
+                    ty
+                );
+            }
+        }
+
         Opcode::Fadd | Opcode::Fsub | Opcode::Fmul | Opcode::Fdiv => {
             let lhs = put_input_in_reg(ctx, inputs[0]);
             let rhs = input_to_reg_mem(ctx, inputs[1]);
@@ -2910,12 +2928,10 @@ fn lower_insn_to_regs<C: LowerCtx<I = Inst>>(
                 match op {
                     Opcode::SwidenLow => match (input_ty, output_ty) {
                         (types::I8X16, types::I16X8) => {
-                            ctx.emit(Inst::gen_move(dst, src, output_ty));
-                            ctx.emit(Inst::xmm_rm_r(SseOpcode::Pmovsxbw, RegMem::from(dst), dst));
+                            ctx.emit(Inst::xmm_mov(SseOpcode::Pmovsxbw, RegMem::reg(src), dst));
                         }
                         (types::I16X8, types::I32X4) => {
-                            ctx.emit(Inst::gen_move(dst, src, output_ty));
-                            ctx.emit(Inst::xmm_rm_r(SseOpcode::Pmovsxwd, RegMem::from(dst), dst));
+                            ctx.emit(Inst::xmm_mov(SseOpcode::Pmovsxwd, RegMem::reg(src), dst));
                         }
                         _ => unreachable!(),
                     },
@@ -2929,7 +2945,7 @@ fn lower_insn_to_regs<C: LowerCtx<I = Inst>>(
                                 8,
                                 false,
                             ));
-                            ctx.emit(Inst::xmm_rm_r(SseOpcode::Pmovsxbw, RegMem::from(dst), dst));
+                            ctx.emit(Inst::xmm_mov(SseOpcode::Pmovsxbw, RegMem::from(dst), dst));
                         }
                         (types::I16X8, types::I32X4) => {
                             ctx.emit(Inst::gen_move(dst, src, output_ty));
@@ -2940,18 +2956,16 @@ fn lower_insn_to_regs<C: LowerCtx<I = Inst>>(
                                 8,
                                 false,
                             ));
-                            ctx.emit(Inst::xmm_rm_r(SseOpcode::Pmovsxwd, RegMem::from(dst), dst));
+                            ctx.emit(Inst::xmm_mov(SseOpcode::Pmovsxwd, RegMem::from(dst), dst));
                         }
                         _ => unreachable!(),
                     },
                     Opcode::UwidenLow => match (input_ty, output_ty) {
                         (types::I8X16, types::I16X8) => {
-                            ctx.emit(Inst::gen_move(dst, src, output_ty));
-                            ctx.emit(Inst::xmm_rm_r(SseOpcode::Pmovzxbw, RegMem::from(dst), dst));
+                            ctx.emit(Inst::xmm_mov(SseOpcode::Pmovzxbw, RegMem::reg(src), dst));
                         }
                         (types::I16X8, types::I32X4) => {
-                            ctx.emit(Inst::gen_move(dst, src, output_ty));
-                            ctx.emit(Inst::xmm_rm_r(SseOpcode::Pmovzxwd, RegMem::from(dst), dst));
+                            ctx.emit(Inst::xmm_mov(SseOpcode::Pmovzxwd, RegMem::reg(src), dst));
                         }
                         _ => unreachable!(),
                     },
@@ -2965,7 +2979,7 @@ fn lower_insn_to_regs<C: LowerCtx<I = Inst>>(
                                 8,
                                 false,
                             ));
-                            ctx.emit(Inst::xmm_rm_r(SseOpcode::Pmovzxbw, RegMem::from(dst), dst));
+                            ctx.emit(Inst::xmm_mov(SseOpcode::Pmovzxbw, RegMem::from(dst), dst));
                         }
                         (types::I16X8, types::I32X4) => {
                             ctx.emit(Inst::gen_move(dst, src, output_ty));
@@ -2976,7 +2990,7 @@ fn lower_insn_to_regs<C: LowerCtx<I = Inst>>(
                                 8,
                                 false,
                             ));
-                            ctx.emit(Inst::xmm_rm_r(SseOpcode::Pmovzxwd, RegMem::from(dst), dst));
+                            ctx.emit(Inst::xmm_mov(SseOpcode::Pmovzxwd, RegMem::from(dst), dst));
                         }
                         _ => unreachable!(),
                     },
@@ -3264,7 +3278,13 @@ fn lower_insn_to_regs<C: LowerCtx<I = Inst>>(
         | Opcode::Uload16Complex
         | Opcode::Sload16Complex
         | Opcode::Uload32Complex
-        | Opcode::Sload32Complex => {
+        | Opcode::Sload32Complex
+        | Opcode::Sload8x8
+        | Opcode::Uload8x8
+        | Opcode::Sload16x4
+        | Opcode::Uload16x4
+        | Opcode::Sload32x2
+        | Opcode::Uload32x2 => {
             let offset = ctx.data(insn).load_store_offset().unwrap();
 
             let elem_ty = match op {
@@ -3279,6 +3299,18 @@ fn lower_insn_to_regs<C: LowerCtx<I = Inst>>(
                 | Opcode::Uload32
                 | Opcode::Sload32Complex
                 | Opcode::Uload32Complex => types::I32,
+                Opcode::Sload8x8
+                | Opcode::Uload8x8
+                | Opcode::Sload8x8Complex
+                | Opcode::Uload8x8Complex => types::I8X8,
+                Opcode::Sload16x4
+                | Opcode::Uload16x4
+                | Opcode::Sload16x4Complex
+                | Opcode::Uload16x4Complex => types::I16X4,
+                Opcode::Sload32x2
+                | Opcode::Uload32x2
+                | Opcode::Sload32x2Complex
+                | Opcode::Uload32x2Complex => types::I32X2,
                 Opcode::Load | Opcode::LoadComplex => ctx.output_ty(insn, 0),
                 _ => unimplemented!(),
             };
@@ -3291,7 +3323,13 @@ fn lower_insn_to_regs<C: LowerCtx<I = Inst>>(
                 | Opcode::Sload16
                 | Opcode::Sload16Complex
                 | Opcode::Sload32
-                | Opcode::Sload32Complex => true,
+                | Opcode::Sload32Complex
+                | Opcode::Sload8x8
+                | Opcode::Sload8x8Complex
+                | Opcode::Sload16x4
+                | Opcode::Sload16x4Complex
+                | Opcode::Sload32x2
+                | Opcode::Sload32x2Complex => true,
                 _ => false,
             };
 
@@ -3302,7 +3340,13 @@ fn lower_insn_to_regs<C: LowerCtx<I = Inst>>(
                 | Opcode::Uload16
                 | Opcode::Sload16
                 | Opcode::Uload32
-                | Opcode::Sload32 => {
+                | Opcode::Sload32
+                | Opcode::Sload8x8
+                | Opcode::Uload8x8
+                | Opcode::Sload16x4
+                | Opcode::Uload16x4
+                | Opcode::Sload32x2
+                | Opcode::Uload32x2 => {
                     assert_eq!(inputs.len(), 1, "only one input for load operands");
                     lower_to_amode(ctx, inputs[0], offset)
                 }
@@ -3313,7 +3357,13 @@ fn lower_insn_to_regs<C: LowerCtx<I = Inst>>(
                 | Opcode::Uload16Complex
                 | Opcode::Sload16Complex
                 | Opcode::Uload32Complex
-                | Opcode::Sload32Complex => {
+                | Opcode::Sload32Complex
+                | Opcode::Sload8x8Complex
+                | Opcode::Uload8x8Complex
+                | Opcode::Sload16x4Complex
+                | Opcode::Uload16x4Complex
+                | Opcode::Sload32x2Complex
+                | Opcode::Uload32x2Complex => {
                     assert_eq!(
                         inputs.len(),
                         2,
@@ -3325,12 +3375,12 @@ fn lower_insn_to_regs<C: LowerCtx<I = Inst>>(
                     let flags = ctx.memflags(insn).expect("load should have memflags");
                     Amode::imm_reg_reg_shift(offset as u32, base, index, shift).with_flags(flags)
                 }
-
                 _ => unreachable!(),
             };
 
             let dst = get_output_reg(ctx, outputs[0]);
             let is_xmm = elem_ty.is_float() || elem_ty.is_vector();
+
             match (sign_extend, is_xmm) {
                 (true, false) => {
                     // The load is sign-extended only when the output size is lower than 64 bits,
@@ -3350,15 +3400,40 @@ fn lower_insn_to_regs<C: LowerCtx<I = Inst>>(
                     ctx.emit(match elem_ty {
                         types::F32 => Inst::xmm_mov(SseOpcode::Movss, RegMem::mem(amode), dst),
                         types::F64 => Inst::xmm_mov(SseOpcode::Movsd, RegMem::mem(amode), dst),
+                        types::I8X8 => {
+                            if sign_extend == true {
+                                Inst::xmm_mov(SseOpcode::Pmovsxbw, RegMem::mem(amode), dst)
+                            } else {
+                                Inst::xmm_mov(SseOpcode::Pmovzxbw, RegMem::mem(amode), dst)
+                            }
+                        }
+                        types::I16X4 => {
+                            if sign_extend == true {
+                                Inst::xmm_mov(SseOpcode::Pmovsxwd, RegMem::mem(amode), dst)
+                            } else {
+                                Inst::xmm_mov(SseOpcode::Pmovzxwd, RegMem::mem(amode), dst)
+                            }
+                        }
+                        types::I32X2 => {
+                            if sign_extend == true {
+                                Inst::xmm_mov(SseOpcode::Pmovsxdq, RegMem::mem(amode), dst)
+                            } else {
+                                Inst::xmm_mov(SseOpcode::Pmovzxdq, RegMem::mem(amode), dst)
+                            }
+                        }
                         _ if elem_ty.is_vector() && elem_ty.bits() == 128 => {
                             Inst::xmm_mov(SseOpcode::Movups, RegMem::mem(amode), dst)
-                        } // TODO Specialize for different types: MOVUPD, MOVDQU
-                        _ => unreachable!("unexpected type for load: {:?}", elem_ty),
+                        }
+                        // TODO Specialize for different types: MOVUPD, MOVDQU
+                        _ => unreachable!(
+                            "unexpected type for load: {:?} - {:?}",
+                            elem_ty,
+                            elem_ty.bits()
+                        ),
                     });
                 }
             }
         }
-
         Opcode::Store
         | Opcode::Istore8
         | Opcode::Istore16
diff --git a/cranelift/codegen/src/isa/x86/unwind.rs b/cranelift/codegen/src/isa/x86/unwind.rs
index 717964a287..2eed8b74e4 100644
--- a/cranelift/codegen/src/isa/x86/unwind.rs
+++ b/cranelift/codegen/src/isa/x86/unwind.rs
@@ -250,7 +250,7 @@ mod tests {
     use crate::ir::{
         types, AbiParam, ExternalName, InstBuilder, Signature, StackSlotData, StackSlotKind,
     };
-    use crate::isa::{lookup, CallConv};
+    use crate::isa::{lookup_variant, BackendVariant, CallConv};
     use crate::settings::{builder, Flags};
     use crate::Context;
     use std::str::FromStr;
@@ -258,7 +258,7 @@ mod tests {
 
     #[test]
     fn test_small_alloc() {
-        let isa = lookup(triple!("x86_64"))
+        let isa = lookup_variant(triple!("x86_64"), BackendVariant::Legacy)
             .expect("expect x86 ISA")
             .finish(Flags::new(builder()));
 
@@ -314,7 +314,7 @@ mod tests {
 
     #[test]
     fn test_medium_alloc() {
-        let isa = lookup(triple!("x86_64"))
+        let isa = lookup_variant(triple!("x86_64"), BackendVariant::Legacy)
             .expect("expect x86 ISA")
             .finish(Flags::new(builder()));
 
@@ -370,7 +370,7 @@ mod tests {
 
     #[test]
     fn test_large_alloc() {
-        let isa = lookup(triple!("x86_64"))
+        let isa = lookup_variant(triple!("x86_64"), BackendVariant::Legacy)
             .expect("expect x86 ISA")
             .finish(Flags::new(builder()));
 
@@ -442,7 +442,7 @@ mod tests {
 
     #[test]
     fn test_multi_return_func() {
-        let isa = lookup(triple!("x86_64"))
+        let isa = lookup_variant(triple!("x86_64"), BackendVariant::Legacy)
             .expect("expect x86 ISA")
             .finish(Flags::new(builder()));
 
diff --git a/cranelift/codegen/src/isa/x86/unwind/systemv.rs b/cranelift/codegen/src/isa/x86/unwind/systemv.rs
index 8f112d943a..09778627bb 100644
--- a/cranelift/codegen/src/isa/x86/unwind/systemv.rs
+++ b/cranelift/codegen/src/isa/x86/unwind/systemv.rs
@@ -134,7 +134,7 @@ mod tests {
     use crate::ir::{
         types, AbiParam, ExternalName, InstBuilder, Signature, StackSlotData, StackSlotKind,
     };
-    use crate::isa::{lookup, CallConv};
+    use crate::isa::{lookup_variant, BackendVariant, CallConv};
     use crate::settings::{builder, Flags};
     use crate::Context;
     use gimli::write::Address;
@@ -143,7 +143,7 @@ mod tests {
 
     #[test]
     fn test_simple_func() {
-        let isa = lookup(triple!("x86_64"))
+        let isa = lookup_variant(triple!("x86_64"), BackendVariant::Legacy)
             .expect("expect x86 ISA")
             .finish(Flags::new(builder()));
 
@@ -185,7 +185,7 @@ mod tests {
 
     #[test]
     fn test_multi_return_func() {
-        let isa = lookup(triple!("x86_64"))
+        let isa = lookup_variant(triple!("x86_64"), BackendVariant::Legacy)
             .expect("expect x86 ISA")
             .finish(Flags::new(builder()));
 
diff --git a/cranelift/codegen/src/isa/x86/unwind/winx64.rs b/cranelift/codegen/src/isa/x86/unwind/winx64.rs
index e49b2c1ef1..1fe93f9708 100644
--- a/cranelift/codegen/src/isa/x86/unwind/winx64.rs
+++ b/cranelift/codegen/src/isa/x86/unwind/winx64.rs
@@ -46,7 +46,7 @@ mod tests {
     use crate::ir::{ExternalName, InstBuilder, Signature, StackSlotData, StackSlotKind};
     use crate::isa::unwind::winx64::UnwindCode;
     use crate::isa::x86::registers::RU;
-    use crate::isa::{lookup, CallConv};
+    use crate::isa::{lookup_variant, BackendVariant, CallConv};
     use crate::settings::{builder, Flags};
     use crate::Context;
     use std::str::FromStr;
@@ -54,7 +54,7 @@ mod tests {
 
     #[test]
     fn test_wrong_calling_convention() {
-        let isa = lookup(triple!("x86_64"))
+        let isa = lookup_variant(triple!("x86_64"), BackendVariant::Legacy)
             .expect("expect x86 ISA")
             .finish(Flags::new(builder()));
 
@@ -70,7 +70,7 @@ mod tests {
 
     #[test]
     fn test_small_alloc() {
-        let isa = lookup(triple!("x86_64"))
+        let isa = lookup_variant(triple!("x86_64"), BackendVariant::Legacy)
             .expect("expect x86 ISA")
             .finish(Flags::new(builder()));
 
@@ -127,7 +127,7 @@ mod tests {
 
     #[test]
     fn test_medium_alloc() {
-        let isa = lookup(triple!("x86_64"))
+        let isa = lookup_variant(triple!("x86_64"), BackendVariant::Legacy)
             .expect("expect x86 ISA")
             .finish(Flags::new(builder()));
 
@@ -188,7 +188,7 @@ mod tests {
 
     #[test]
     fn test_large_alloc() {
-        let isa = lookup(triple!("x86_64"))
+        let isa = lookup_variant(triple!("x86_64"), BackendVariant::Legacy)
             .expect("expect x86 ISA")
             .finish(Flags::new(builder()));
 
diff --git a/cranelift/filetests/filetests/isa/x86/simd-arithmetic-binemit.clif b/cranelift/filetests/filetests/isa/x86/simd-arithmetic-binemit.clif
index 8df5c2afa8..cd942338dd 100644
--- a/cranelift/filetests/filetests/isa/x86/simd-arithmetic-binemit.clif
+++ b/cranelift/filetests/filetests/isa/x86/simd-arithmetic-binemit.clif
@@ -108,3 +108,9 @@ block0(v0: i64x2 [%xmm3], v1: i64x2 [%xmm5]):
 [-, %xmm3]    v2 = x86_pmuludq v0, v1      ; bin: 66 0f f4 dd
     return v2
 }
+
+function %pmaddwd(i16x8, i16x8) -> i32x4 {
+block0(v0: i16x8 [%xmm8], v1: i16x8 [%xmm9]):
+[-, %xmm8]    v2 = widening_pairwise_dot_product_s v0, v1      ; bin: 66 45 0f f5 c1
+    return v2
+}
diff --git a/cranelift/native/src/lib.rs b/cranelift/native/src/lib.rs
index 668a1934bf..3bd5716dac 100644
--- a/cranelift/native/src/lib.rs
+++ b/cranelift/native/src/lib.rs
@@ -34,10 +34,26 @@ use raw_cpuid::CpuId;
 /// machine, or `Err(())` if the host machine is not supported
 /// in the current configuration.
 pub fn builder() -> Result<isa::Builder, &'static str> {
-    let mut isa_builder = isa::lookup(Triple::host()).map_err(|err| match err {
-        isa::LookupError::SupportDisabled => "support for architecture disabled at compile time",
-        isa::LookupError::Unsupported => "unsupported architecture",
-    })?;
+    builder_with_backend_variant(isa::BackendVariant::Any)
+}
+
+/// Return an `isa` builder configured for the current host
+/// machine, or `Err(())` if the host machine is not supported
+/// in the current configuration.
+///
+/// Selects the given backend variant specifically; this is
+/// useful when more than oen backend exists for a given target
+/// (e.g., on x86-64).
+pub fn builder_with_backend_variant(
+    variant: isa::BackendVariant,
+) -> Result<isa::Builder, &'static str> {
+    let mut isa_builder =
+        isa::lookup_variant(Triple::host(), variant).map_err(|err| match err {
+            isa::LookupError::SupportDisabled => {
+                "support for architecture disabled at compile time"
+            }
+            isa::LookupError::Unsupported => "unsupported architecture",
+        })?;
 
     if cfg!(any(target_arch = "x86", target_arch = "x86_64")) {
         parse_x86_cpuid(&mut isa_builder)?;
diff --git a/cranelift/object/src/backend.rs b/cranelift/object/src/backend.rs
index 81dd75c4cd..98ab139a2f 100644
--- a/cranelift/object/src/backend.rs
+++ b/cranelift/object/src/backend.rs
@@ -145,6 +145,18 @@ impl ObjectModule {
     }
 }
 
+fn validate_symbol(name: &str) -> ModuleResult<()> {
+    // null bytes are not allowed in symbol names and will cause the `object`
+    // crate to panic. Let's return a clean error instead.
+    if name.contains("\0") {
+        return Err(ModuleError::Backend(anyhow::anyhow!(
+            "Symbol {:?} has a null byte, which is disallowed",
+            name
+        )));
+    }
+    Ok(())
+}
+
 impl Module for ObjectModule {
     fn isa(&self) -> &dyn TargetIsa {
         &*self.isa
@@ -160,6 +172,8 @@ impl Module for ObjectModule {
         linkage: Linkage,
         signature: &ir::Signature,
     ) -> ModuleResult<FuncId> {
+        validate_symbol(name)?;
+
         let (id, decl) = self
             .declarations
             .declare_function(name, linkage, signature)?;
@@ -194,6 +208,8 @@ impl Module for ObjectModule {
         writable: bool,
         tls: bool,
     ) -> ModuleResult<DataId> {
+        validate_symbol(name)?;
+
         let (id, decl) = self
             .declarations
             .declare_data(name, linkage, writable, tls)?;
diff --git a/cranelift/object/tests/basic.rs b/cranelift/object/tests/basic.rs
index 18acae5733..dad2bbca9d 100644
--- a/cranelift/object/tests/basic.rs
+++ b/cranelift/object/tests/basic.rs
@@ -197,3 +197,46 @@ fn libcall_function() {
 
     module.finish();
 }
+
+#[test]
+#[should_panic(
+    expected = "Result::unwrap()` on an `Err` value: Backend(Symbol \"function\\u{0}with\\u{0}nul\\u{0}bytes\" has a null byte, which is disallowed"
+)]
+fn reject_nul_byte_symbol_for_func() {
+    let flag_builder = settings::builder();
+    let isa_builder = cranelift_codegen::isa::lookup_by_name("x86_64-unknown-linux-gnu").unwrap();
+    let isa = isa_builder.finish(settings::Flags::new(flag_builder));
+    let mut module =
+        ObjectModule::new(ObjectBuilder::new(isa, "foo", default_libcall_names()).unwrap());
+
+    let sig = Signature {
+        params: vec![],
+        returns: vec![],
+        call_conv: CallConv::SystemV,
+    };
+
+    let _ = module
+        .declare_function("function\u{0}with\u{0}nul\u{0}bytes", Linkage::Local, &sig)
+        .unwrap();
+}
+
+#[test]
+#[should_panic(
+    expected = "Result::unwrap()` on an `Err` value: Backend(Symbol \"data\\u{0}with\\u{0}nul\\u{0}bytes\" has a null byte, which is disallowed"
+)]
+fn reject_nul_byte_symbol_for_data() {
+    let flag_builder = settings::builder();
+    let isa_builder = cranelift_codegen::isa::lookup_by_name("x86_64-unknown-linux-gnu").unwrap();
+    let isa = isa_builder.finish(settings::Flags::new(flag_builder));
+    let mut module =
+        ObjectModule::new(ObjectBuilder::new(isa, "foo", default_libcall_names()).unwrap());
+
+    let _ = module
+        .declare_data(
+            "data\u{0}with\u{0}nul\u{0}bytes",
+            Linkage::Local,
+            /* writable = */ true,
+            /* tls = */ false,
+        )
+        .unwrap();
+}
diff --git a/cranelift/wasm/Cargo.toml b/cranelift/wasm/Cargo.toml
index 48e19a9d57..4a37c004d1 100644
--- a/cranelift/wasm/Cargo.toml
+++ b/cranelift/wasm/Cargo.toml
@@ -12,7 +12,7 @@ keywords = ["webassembly", "wasm"]
 edition = "2018"
 
 [dependencies]
-wasmparser = { version = "0.70", default-features = false }
+wasmparser = { version = "0.71", default-features = false }
 cranelift-codegen = { path = "../codegen", version = "0.68.0", default-features = false }
 cranelift-entity = { path = "../entity", version = "0.68.0" }
 cranelift-frontend = { path = "../frontend", version = "0.68.0", default-features = false }
diff --git a/cranelift/wasm/src/code_translator.rs b/cranelift/wasm/src/code_translator.rs
index 2fe171842b..998a0c7417 100644
--- a/cranelift/wasm/src/code_translator.rs
+++ b/cranelift/wasm/src/code_translator.rs
@@ -1810,7 +1810,20 @@ pub fn translate_operator<FE: FuncEnvironment + ?Sized>(
             let (a, b) = pop2_with_bitcast(state, I16X8, builder);
             state.push1(builder.ins().widening_pairwise_dot_product_s(a, b));
         }
-
+        Operator::I16x8ExtMulLowI8x16S
+        | Operator::I16x8ExtMulHighI8x16S
+        | Operator::I16x8ExtMulLowI8x16U
+        | Operator::I16x8ExtMulHighI8x16U
+        | Operator::I32x4ExtMulLowI16x8S
+        | Operator::I32x4ExtMulHighI16x8S
+        | Operator::I32x4ExtMulLowI16x8U
+        | Operator::I32x4ExtMulHighI16x8U
+        | Operator::I64x2ExtMulLowI32x4S
+        | Operator::I64x2ExtMulHighI32x4S
+        | Operator::I64x2ExtMulLowI32x4U
+        | Operator::I64x2ExtMulHighI32x4U => {
+            return Err(wasm_unsupported!("proposed simd operator {:?}", op));
+        }
         Operator::ReturnCall { .. } | Operator::ReturnCallIndirect { .. } => {
             return Err(wasm_unsupported!("proposed tail-call operator {:?}", op));
         }
diff --git a/crates/bench-api/Cargo.toml b/crates/bench-api/Cargo.toml
new file mode 100644
index 0000000000..02dfd986be
--- /dev/null
+++ b/crates/bench-api/Cargo.toml
@@ -0,0 +1,25 @@
+[package]
+name = "wasmtime-bench-api"
+version = "0.19.0"
+authors = ["The Wasmtime Project Developers"]
+description = "Exposes a benchmarking API for the Wasmtime runtime"
+license = "Apache-2.0 WITH LLVM-exception"
+repository = "https://github.com/bytecodealliance/wasmtime"
+readme = "README.md"
+edition = "2018"
+publish = false
+
+[lib]
+name = "wasmtime_bench_api"
+crate-type = ["rlib", "cdylib"]
+# The rlib is only included here so that `cargo test` will run.
+
+[dependencies]
+anyhow = "1.0"
+wasmtime = { path = "../wasmtime", default-features = false }
+wasmtime-wasi = { path = "../wasi" }
+wasi-common = { path = "../wasi-common" }
+
+
+[dev-dependencies]
+wat = "1.0"
diff --git a/crates/bench-api/src/lib.rs b/crates/bench-api/src/lib.rs
new file mode 100644
index 0000000000..a2958f09a8
--- /dev/null
+++ b/crates/bench-api/src/lib.rs
@@ -0,0 +1,248 @@
+//! A C API for benchmarking Wasmtime's WebAssembly compilation, instantiation,
+//! and execution.
+//!
+//! The API expects sequential calls to:
+//!
+//!  - `wasm_bench_create`
+//!  - `wasm_bench_compile`
+//!  - `wasm_bench_instantiate`
+//!  - `wasm_bench_execute`
+//!  - `wasm_bench_free`
+//!
+//! You may repeat this sequence of calls multiple times to take multiple
+//! measurements of compilation, instantiation, and execution time within a
+//! single process.
+//!
+//! All API calls must happen on the same thread.
+//!
+//! Functions which return pointers use null as an error value. Function which
+//! return `int` use `0` as OK and non-zero as an error value.
+//!
+//! # Example
+//!
+//! ```
+//! use wasmtime_bench_api::*;
+//!
+//! let engine = unsafe { wasm_bench_create() };
+//! assert!(!engine.is_null());
+//!
+//! let wasm = wat::parse_bytes(br#"
+//!     (module
+//!         (func $bench_start (import "bench" "start"))
+//!         (func $bench_end (import "bench" "end"))
+//!         (func $start (export "_start")
+//!             call $bench_start
+//!             i32.const 1
+//!             i32.const 2
+//!             i32.add
+//!             drop
+//!             call $bench_end
+//!         )
+//!     )
+//! "#).unwrap();
+//!
+//! // Start your compilation timer here.
+//! let code = unsafe { wasm_bench_compile(engine, wasm.as_ptr(), wasm.len()) };
+//! // End your compilation timer here.
+//! assert_eq!(code, OK);
+//!
+//! // The Wasm benchmark will expect us to provide functions to start ("bench"
+//! // "start") and stop ("bench" "stop") the measurement counters/timers during
+//! // execution.
+//! extern "C" fn bench_start() {
+//!     // Start your execution timer here.
+//! }
+//! extern "C" fn bench_stop() {
+//!     // End your execution timer here.
+//! }
+//!
+//! // Start your instantiation timer here.
+//! let code = unsafe { wasm_bench_instantiate(engine, bench_start, bench_stop) };
+//! // End your instantiation timer here.
+//! assert_eq!(code, OK);
+//!
+//! // No need to start timers for the execution since, by convention, the timer
+//! // functions we passed during instantiation will be called by the benchmark
+//! // at the appropriate time (before and after the benchmarked section).
+//! let code = unsafe { wasm_bench_execute(engine) };
+//! assert_eq!(code, OK);
+//!
+//! unsafe {
+//!     wasm_bench_free(engine);
+//! }
+//! ```
+
+use anyhow::{anyhow, Context, Result};
+use std::os::raw::{c_int, c_void};
+use std::slice;
+use wasi_common::WasiCtxBuilder;
+use wasmtime::{Config, Engine, Instance, Linker, Module, Store};
+use wasmtime_wasi::Wasi;
+
+pub type ExitCode = c_int;
+pub const OK: ExitCode = 0;
+pub const ERR: ExitCode = -1;
+
+/// Exposes a C-compatible way of creating the engine from the bytes of a single
+/// Wasm module.
+///
+/// This function returns a pointer to a structure that contains the engine's
+/// initialized state.
+#[no_mangle]
+pub extern "C" fn wasm_bench_create() -> *mut c_void {
+    let state = Box::new(BenchState::new());
+    Box::into_raw(state) as _
+}
+
+/// Free the engine state allocated by this library.
+#[no_mangle]
+pub extern "C" fn wasm_bench_free(state: *mut c_void) {
+    assert!(!state.is_null());
+    unsafe {
+        Box::from_raw(state as *mut BenchState);
+    }
+}
+
+/// Compile the Wasm benchmark module.
+#[no_mangle]
+pub extern "C" fn wasm_bench_compile(
+    state: *mut c_void,
+    wasm_bytes: *const u8,
+    wasm_bytes_length: usize,
+) -> ExitCode {
+    let state = unsafe { (state as *mut BenchState).as_mut().unwrap() };
+    let wasm_bytes = unsafe { slice::from_raw_parts(wasm_bytes, wasm_bytes_length) };
+    let result = state.compile(wasm_bytes).context("failed to compile");
+    to_exit_code(result)
+}
+
+/// Instantiate the Wasm benchmark module.
+#[no_mangle]
+pub extern "C" fn wasm_bench_instantiate(
+    state: *mut c_void,
+    bench_start: extern "C" fn(),
+    bench_end: extern "C" fn(),
+) -> ExitCode {
+    let state = unsafe { (state as *mut BenchState).as_mut().unwrap() };
+    let result = state
+        .instantiate(bench_start, bench_end)
+        .context("failed to instantiate");
+    to_exit_code(result)
+}
+
+/// Execute the Wasm benchmark module.
+#[no_mangle]
+pub extern "C" fn wasm_bench_execute(state: *mut c_void) -> ExitCode {
+    let state = unsafe { (state as *mut BenchState).as_mut().unwrap() };
+    let result = state.execute().context("failed to execute");
+    to_exit_code(result)
+}
+
+/// Helper function for converting a Rust result to a C error code.
+///
+/// This will print an error indicating some information regarding the failure.
+fn to_exit_code<T>(result: impl Into<Result<T>>) -> ExitCode {
+    match result.into() {
+        Ok(_) => OK,
+        Err(error) => {
+            eprintln!("{:?}", error);
+            ERR
+        }
+    }
+}
+
+/// This structure contains the actual Rust implementation of the state required
+/// to manage the Wasmtime engine between calls.
+struct BenchState {
+    engine: Engine,
+    store: Store,
+    module: Option<Module>,
+    instance: Option<Instance>,
+    did_execute: bool,
+}
+
+impl BenchState {
+    fn new() -> Self {
+        let mut config = Config::new();
+        config.wasm_simd(true);
+        // NB: do not configure a code cache.
+
+        let engine = Engine::new(&config);
+        let store = Store::new(&engine);
+        Self {
+            engine,
+            store,
+            module: None,
+            instance: None,
+            did_execute: false,
+        }
+    }
+
+    fn compile(&mut self, bytes: &[u8]) -> Result<()> {
+        assert!(
+            self.module.is_none(),
+            "create a new engine to repeat compilation"
+        );
+        self.module = Some(Module::from_binary(&self.engine, bytes)?);
+        Ok(())
+    }
+
+    fn instantiate(
+        &mut self,
+        bench_start: extern "C" fn(),
+        bench_end: extern "C" fn(),
+    ) -> Result<()> {
+        assert!(
+            self.instance.is_none(),
+            "create a new engine to repeat instantiation"
+        );
+        let module = self
+            .module
+            .as_mut()
+            .expect("compile the module before instantiating it");
+
+        let mut linker = Linker::new(&self.store);
+
+        // Import a very restricted WASI environment.
+        let mut cx = WasiCtxBuilder::new();
+        cx.inherit_stdio();
+        let cx = cx.build()?;
+        let wasi = Wasi::new(linker.store(), cx);
+        wasi.add_to_linker(&mut linker)?;
+
+        // Import the specialized benchmarking functions.
+        linker.func("bench", "start", move || bench_start())?;
+        linker.func("bench", "end", move || bench_end())?;
+
+        self.instance = Some(linker.instantiate(&module)?);
+        Ok(())
+    }
+
+    fn execute(&mut self) -> Result<()> {
+        assert!(!self.did_execute, "create a new engine to repeat execution");
+        self.did_execute = true;
+
+        let instance = self
+            .instance
+            .as_ref()
+            .expect("instantiate the module before executing it");
+
+        let start_func = instance.get_func("_start").expect("a _start function");
+        let runnable_func = start_func.get0::<()>()?;
+        match runnable_func() {
+            Ok(_) => Ok(()),
+            Err(trap) => {
+                // Since _start will likely return by using the system `exit` call, we must
+                // check the trap code to see if it actually represents a successful exit.
+                match trap.i32_exit_status() {
+                    Some(0) => Ok(()),
+                    Some(n) => Err(anyhow!("_start exited with a non-zero code: {}", n)),
+                    None => Err(anyhow!(
+                        "executing the benchmark resulted in a trap: {}",
+                        trap
+                    )),
+                }
+            }
+        }
+    }
+}
diff --git a/crates/debug/Cargo.toml b/crates/debug/Cargo.toml
index f3b5bdede7..c893c9ab13 100644
--- a/crates/debug/Cargo.toml
+++ b/crates/debug/Cargo.toml
@@ -13,8 +13,8 @@ edition = "2018"
 
 [dependencies]
 gimli = "0.23.0"
-wasmparser = "0.70"
-object = { version = "0.22.0", default-features = false, features = ["read", "write"] }
+wasmparser = "0.71"
+object = { version = "0.22.0", default-features = false, features = ["read_core", "elf", "write"] }
 wasmtime-environ = { path = "../environ", version = "0.21.0" }
 target-lexicon = { version = "0.11.0", default-features = false }
 anyhow = "1.0"
diff --git a/crates/environ/Cargo.toml b/crates/environ/Cargo.toml
index 26b0ec0eaf..aa132a340f 100644
--- a/crates/environ/Cargo.toml
+++ b/crates/environ/Cargo.toml
@@ -16,7 +16,7 @@ anyhow = "1.0"
 cranelift-codegen = { path = "../../cranelift/codegen", version = "0.68.0", features = ["enable-serde"] }
 cranelift-entity = { path = "../../cranelift/entity", version = "0.68.0", features = ["enable-serde"] }
 cranelift-wasm = { path = "../../cranelift/wasm", version = "0.68.0", features = ["enable-serde"] }
-wasmparser = "0.70"
+wasmparser = "0.71"
 indexmap = { version = "1.0.2", features = ["serde-1"] }
 thiserror = "1.0.4"
 serde = { version = "1.0.94", features = ["derive"] }
diff --git a/crates/fuzzing/Cargo.toml b/crates/fuzzing/Cargo.toml
index 7120601b2e..2cc157704a 100644
--- a/crates/fuzzing/Cargo.toml
+++ b/crates/fuzzing/Cargo.toml
@@ -5,6 +5,7 @@ edition = "2018"
 name = "wasmtime-fuzzing"
 publish = false
 version = "0.19.0"
+license = "Apache-2.0 WITH LLVM-exception"
 
 [dependencies]
 anyhow = "1.0.22"
@@ -12,10 +13,11 @@ arbitrary = { version = "0.4.1", features = ["derive"] }
 env_logger = "0.8.1"
 log = "0.4.8"
 rayon = "1.2.1"
-wasmparser = "0.70"
+wasmparser = "0.71"
 wasmprinter = "0.2.17"
 wasmtime = { path = "../wasmtime" }
 wasmtime-wast = { path = "../wast" }
+wasm-encoder = "0.2"
 wasm-smith = "0.3.0"
 wasmi = "0.7.0"
 
diff --git a/crates/fuzzing/src/generators/table_ops.rs b/crates/fuzzing/src/generators/table_ops.rs
index 45aaa5b802..adb2373fea 100644
--- a/crates/fuzzing/src/generators/table_ops.rs
+++ b/crates/fuzzing/src/generators/table_ops.rs
@@ -1,8 +1,11 @@
 //! Generating series of `table.get` and `table.set` operations.
 
 use arbitrary::Arbitrary;
-use std::fmt::Write;
 use std::ops::Range;
+use wasm_encoder::{
+    CodeSection, EntityType, Export, ExportSection, Function, FunctionSection, ImportSection,
+    Instruction, Limits, Module, TableSection, TableType, TypeSection, ValType,
+};
 
 /// A description of a Wasm module that makes a series of `externref` table
 /// operations.
@@ -32,7 +35,7 @@ impl TableOps {
         table_size
     }
 
-    /// Convert this into a WAT string.
+    /// Serialize this module into a Wasm binary.
     ///
     /// The module requires a single import: `(import "" "gc" (func))`. This
     /// should be a function to trigger GC.
@@ -43,32 +46,62 @@ impl TableOps {
     /// The "run" function is guaranteed to terminate (no loops or recursive
     /// calls), but is not guaranteed to avoid traps (might access out-of-bounds
     /// of the table).
-    pub fn to_wat_string(&self) -> String {
-        let mut wat = "(module\n".to_string();
+    pub fn to_wasm_binary(&self) -> Vec<u8> {
+        let mut module = Module::new();
 
         // Import the GC function.
-        wat.push_str("  (import \"\" \"gc\" (func))\n");
+        let mut imports = ImportSection::new();
+        imports.import("", Some("gc"), EntityType::Function(0));
 
         // Define our table.
-        wat.push_str("  (table $table ");
-        write!(&mut wat, "{}", self.table_size()).unwrap();
-        wat.push_str(" externref)\n");
+        let mut tables = TableSection::new();
+        tables.table(TableType {
+            element_type: ValType::ExternRef,
+            limits: Limits {
+                min: self.table_size(),
+                max: None,
+            },
+        });
+
+        // Encode the types for all functions that we are using.
+        let mut types = TypeSection::new();
+        types.function(vec![], vec![]); // 0: "gc"
+        let mut params: Vec<ValType> = Vec::with_capacity(self.num_params() as usize);
+        for _i in 0..self.num_params() {
+            params.push(ValType::ExternRef);
+        }
+        let results = vec![];
+        types.function(params, results); // 1: "run"
 
         // Define the "run" function export.
-        wat.push_str(r#"  (func (export "run") (param"#);
-        for _ in 0..self.num_params() {
-            wat.push_str(" externref");
-        }
-        wat.push_str(")\n");
-        for op in self.ops.iter().take(MAX_OPS) {
-            wat.push_str("    ");
-            op.to_wat_string(&mut wat);
-            wat.push('\n');
-        }
-        wat.push_str("  )\n");
+        let mut functions = FunctionSection::new();
+        functions.function(1);
 
-        wat.push_str(")\n");
-        wat
+        let mut exports = ExportSection::new();
+        exports.export("run", Export::Function(1));
+
+        let mut params: Vec<(u32, ValType)> = Vec::with_capacity(self.num_params() as usize);
+        for _i in 0..self.num_params() {
+            params.push((0, ValType::ExternRef));
+        }
+        let mut func = Function::new(params);
+
+        for op in self.ops.iter().take(MAX_OPS) {
+            op.insert(&mut func);
+        }
+
+        let mut code = CodeSection::new();
+        code.function(&func);
+
+        module
+            .section(&types)
+            .section(&imports)
+            .section(&functions)
+            .section(&tables)
+            .section(&exports)
+            .section(&code);
+
+        module.finish()
     }
 }
 
@@ -77,37 +110,34 @@ pub(crate) enum TableOp {
     // `(call 0)`
     Gc,
     // `(drop (table.get x))`
-    Get(u32),
+    Get(i32),
     // `(table.set x (local.get y))`
-    SetFromParam(u32, u8),
+    SetFromParam(i32, u32),
     // `(table.set x (table.get y))`
-    SetFromGet(u32, u32),
+    SetFromGet(i32, i32),
 }
 
 impl TableOp {
-    fn to_wat_string(&self, wat: &mut String) {
+    fn insert(&self, func: &mut Function) {
         match self {
             Self::Gc => {
-                wat.push_str("(call 0)");
+                func.instruction(Instruction::Call(0));
             }
             Self::Get(x) => {
-                wat.push_str("(drop (table.get $table (i32.const ");
-                write!(wat, "{}", x).unwrap();
-                wat.push_str(")))");
+                func.instruction(Instruction::I32Const(*x));
+                func.instruction(Instruction::TableGet { table: 0 });
+                func.instruction(Instruction::Drop);
             }
             Self::SetFromParam(x, y) => {
-                wat.push_str("(table.set $table (i32.const ");
-                write!(wat, "{}", x).unwrap();
-                wat.push_str(") (local.get ");
-                write!(wat, "{}", y).unwrap();
-                wat.push_str("))");
+                func.instruction(Instruction::I32Const(*x));
+                func.instruction(Instruction::LocalGet(*y));
+                func.instruction(Instruction::TableSet { table: 0 });
             }
             Self::SetFromGet(x, y) => {
-                wat.push_str("(table.set $table (i32.const ");
-                write!(wat, "{}", x).unwrap();
-                wat.push_str(") (table.get $table (i32.const ");
-                write!(wat, "{}", y).unwrap();
-                wat.push_str(")))");
+                func.instruction(Instruction::I32Const(*x));
+                func.instruction(Instruction::I32Const(*y));
+                func.instruction(Instruction::TableGet { table: 0 });
+                func.instruction(Instruction::TableSet { table: 0 });
             }
         }
     }
@@ -132,17 +162,26 @@ mod tests {
 
         let expected = r#"
 (module
-  (import "" "gc" (func))
-  (table $table 10 externref)
-  (func (export "run") (param externref externref)
-    (call 0)
-    (drop (table.get $table (i32.const 0)))
-    (table.set $table (i32.const 1) (local.get 2))
-    (table.set $table (i32.const 3) (table.get $table (i32.const 4)))
-  )
-)
+  (type (;0;) (func))
+  (type (;1;) (func (param externref externref)))
+  (import "" "gc" (func (;0;) (type 0)))
+  (func (;1;) (type 1) (param externref externref)
+    call 0
+    i32.const 0
+    table.get 0
+    drop
+    i32.const 1
+    local.get 2
+    table.set 0
+    i32.const 3
+    i32.const 4
+    table.get 0
+    table.set 0)
+  (table (;0;) 10 externref)
+  (export "run" (func 1)))
 "#;
-        let actual = ops.to_wat_string();
+        let actual = ops.to_wasm_binary();
+        let actual = wasmprinter::print_bytes(&actual).unwrap();
         assert_eq!(actual.trim(), expected.trim());
     }
 }
diff --git a/crates/fuzzing/src/oracles.rs b/crates/fuzzing/src/oracles.rs
index 9f04dc44df..6581f9e0dc 100644
--- a/crates/fuzzing/src/oracles.rs
+++ b/crates/fuzzing/src/oracles.rs
@@ -40,17 +40,6 @@ fn log_wasm(wasm: &[u8]) {
     }
 }
 
-fn log_wat(wat: &str) {
-    if !log::log_enabled!(log::Level::Debug) {
-        return;
-    }
-
-    let i = CNT.fetch_add(1, SeqCst);
-    let name = format!("testcase{}.wat", i);
-    log::debug!("wrote wat file to `{}`", name);
-    std::fs::write(&name, wat).expect("failed to write wat file");
-}
-
 /// Instantiate the Wasm buffer, and implicitly fail if we have an unexpected
 /// panic or segfault or anything else that can be detected "passively".
 ///
@@ -418,9 +407,9 @@ pub fn table_ops(config: crate::generators::Config, ops: crate::generators::tabl
         let engine = Engine::new(&config);
         let store = Store::new(&engine);
 
-        let wat = ops.to_wat_string();
-        log_wat(&wat);
-        let module = match Module::new(&engine, &wat) {
+        let wasm = ops.to_wasm_binary();
+        log_wasm(&wasm);
+        let module = match Module::new(&engine, &wasm) {
             Ok(m) => m,
             Err(_) => return,
         };
diff --git a/crates/jit/Cargo.toml b/crates/jit/Cargo.toml
index c754126d92..7b578543b7 100644
--- a/crates/jit/Cargo.toml
+++ b/crates/jit/Cargo.toml
@@ -28,7 +28,7 @@ rayon = { version = "1.0", optional = true }
 region = "2.1.0"
 thiserror = "1.0.4"
 target-lexicon = { version = "0.11.0", default-features = false }
-wasmparser = "0.70"
+wasmparser = "0.71"
 more-asserts = "0.2.1"
 anyhow = "1.0"
 cfg-if = "1.0"
diff --git a/crates/lightbeam/Cargo.toml b/crates/lightbeam/Cargo.toml
index d178090cfa..65ad3ba751 100644
--- a/crates/lightbeam/Cargo.toml
+++ b/crates/lightbeam/Cargo.toml
@@ -24,7 +24,7 @@ more-asserts = "0.2.1"
 smallvec = "1.0.0"
 thiserror = "1.0.9"
 typemap = "0.3"
-wasmparser = "0.70"
+wasmparser = "0.71"
 
 [dev-dependencies]
 lazy_static = "1.2"
diff --git a/crates/lightbeam/wasmtime/Cargo.toml b/crates/lightbeam/wasmtime/Cargo.toml
index 41b47115ee..51b0900b50 100644
--- a/crates/lightbeam/wasmtime/Cargo.toml
+++ b/crates/lightbeam/wasmtime/Cargo.toml
@@ -13,6 +13,6 @@ edition = "2018"
 
 [dependencies]
 lightbeam = { path = "..", version = "0.21.0" }
-wasmparser = "0.70"
+wasmparser = "0.71"
 cranelift-codegen = { path = "../../../cranelift/codegen", version = "0.68.0" }
 wasmtime-environ = { path = "../../environ", version = "0.21.0" }
diff --git a/crates/profiling/src/jitdump_linux.rs b/crates/profiling/src/jitdump_linux.rs
index e16978843f..ef6684c95f 100644
--- a/crates/profiling/src/jitdump_linux.rs
+++ b/crates/profiling/src/jitdump_linux.rs
@@ -249,11 +249,7 @@ impl State {
         let header = FileHeader {
             timestamp: self.get_time_stamp(),
             e_machine: self.get_e_machine(),
-            magic: if cfg!(target_endian = "little") {
-                0x4A695444
-            } else {
-                0x4454694a
-            },
+            magic: 0x4A695444,
             version: 1,
             size: mem::size_of::<FileHeader>() as u32,
             pad1: 0,
diff --git a/crates/test-programs/Cargo.toml b/crates/test-programs/Cargo.toml
index b1f8e6ac40..b3eb67c6d2 100644
--- a/crates/test-programs/Cargo.toml
+++ b/crates/test-programs/Cargo.toml
@@ -5,6 +5,7 @@ authors = ["The Wasmtime Project Developers"]
 readme = "README.md"
 edition = "2018"
 publish = false
+license = "Apache-2.0 WITH LLVM-exception"
 
 [build-dependencies]
 cfg-if = "1.0"
diff --git a/crates/wasi-common/src/sys/stdio.rs b/crates/wasi-common/src/sys/stdio.rs
index 97005767d4..9ac6d6c9ca 100644
--- a/crates/wasi-common/src/sys/stdio.rs
+++ b/crates/wasi-common/src/sys/stdio.rs
@@ -17,7 +17,7 @@
 // TODO it might worth re-investigating the suitability of this type on Windows.
 
 use super::{fd, AsFile};
-use crate::handle::{Fdflags, Filetype, Handle, HandleRights, Rights, RightsExt, Size};
+use crate::handle::{Fdflags, Filestat, Filetype, Handle, HandleRights, Rights, RightsExt, Size};
 use crate::sandboxed_tty_writer::SandboxedTTYWriter;
 use crate::{Error, Result};
 use std::any::Any;
@@ -65,6 +65,9 @@ impl Handle for Stdin {
         }
         Ok(())
     }
+    fn filestat_get(&self) -> Result<Filestat> {
+        fd::filestat_get(&*self.as_file()?)
+    }
     fn read_vectored(&self, iovs: &mut [io::IoSliceMut]) -> Result<usize> {
         let nread = io::stdin().read_vectored(iovs)?;
         Ok(nread)
@@ -111,6 +114,9 @@ impl Handle for Stdout {
         }
         Ok(())
     }
+    fn filestat_get(&self) -> Result<Filestat> {
+        fd::filestat_get(&*self.as_file()?)
+    }
     fn write_vectored(&self, iovs: &[io::IoSlice]) -> Result<usize> {
         // lock for the duration of the scope
         let stdout = io::stdout();
@@ -165,6 +171,9 @@ impl Handle for Stderr {
         }
         Ok(())
     }
+    fn filestat_get(&self) -> Result<Filestat> {
+        fd::filestat_get(&*self.as_file()?)
+    }
     fn write_vectored(&self, iovs: &[io::IoSlice]) -> Result<usize> {
         // Always sanitize stderr, even if it's not directly connected to a tty,
         // because stderr is meant for diagnostics rather than binary output,
diff --git a/crates/wasi-common/src/sys/unix/mod.rs b/crates/wasi-common/src/sys/unix/mod.rs
index 7135486ec2..84d8ec581e 100644
--- a/crates/wasi-common/src/sys/unix/mod.rs
+++ b/crates/wasi-common/src/sys/unix/mod.rs
@@ -9,7 +9,8 @@ pub(crate) mod poll;
 pub(crate) mod stdio;
 
 cfg_if::cfg_if! {
-    if #[cfg(target_os = "linux")] {
+    if #[cfg(any(target_os = "linux",
+                 target_os = "android"))] {
         mod linux;
         use linux as sys_impl;
     } else if #[cfg(target_os = "emscripten")] {
diff --git a/crates/wasi-common/yanix/src/dir.rs b/crates/wasi-common/yanix/src/dir.rs
index 50166a35b1..8e6c70a40f 100644
--- a/crates/wasi-common/yanix/src/dir.rs
+++ b/crates/wasi-common/yanix/src/dir.rs
@@ -36,9 +36,15 @@ impl Dir {
     }
 
     /// Set the position of the directory stream, see `seekdir(3)`.
-    #[cfg(not(target_os = "android"))]
     pub fn seek(&mut self, loc: SeekLoc) {
-        unsafe { libc::seekdir(self.0.as_ptr(), loc.0) }
+        // https://github.com/rust-lang/libc/pull/1996
+        #[cfg(not(target_os = "android"))]
+        use libc::seekdir;
+        #[cfg(target_os = "android")]
+        extern "C" {
+            fn seekdir(dirp: *mut libc::DIR, loc: libc::c_long);
+        }
+        unsafe { seekdir(self.0.as_ptr(), loc.0) }
     }
 
     /// Reset directory stream, see `rewinddir(3)`.
@@ -50,10 +56,16 @@ impl Dir {
     ///
     /// If this location is given to `Dir::seek`, the entries up to the previously returned
     /// will be omitted and the iteration will start from the currently pending directory entry.
-    #[cfg(not(target_os = "android"))]
     #[allow(dead_code)]
     pub fn tell(&self) -> SeekLoc {
-        let loc = unsafe { libc::telldir(self.0.as_ptr()) };
+        #[cfg(not(target_os = "android"))]
+        use libc::telldir;
+        #[cfg(target_os = "android")]
+        extern "C" {
+            fn telldir(dirp: *mut libc::DIR) -> libc::c_long;
+        }
+        // https://github.com/rust-lang/libc/pull/1996
+        let loc = unsafe { telldir(self.0.as_ptr()) };
         SeekLoc(loc)
     }
 
@@ -93,11 +105,9 @@ impl Entry {
     }
 }
 
-#[cfg(not(target_os = "android"))]
 #[derive(Clone, Copy, Debug)]
 pub struct SeekLoc(pub(crate) libc::c_long);
 
-#[cfg(not(target_os = "android"))]
 impl SeekLoc {
     pub fn to_raw(&self) -> i64 {
         self.0.into()
diff --git a/crates/wasi-common/yanix/src/file.rs b/crates/wasi-common/yanix/src/file.rs
index 1764d29120..62af1c9d77 100644
--- a/crates/wasi-common/yanix/src/file.rs
+++ b/crates/wasi-common/yanix/src/file.rs
@@ -90,11 +90,26 @@ bitflags! {
         const WRONLY = libc::O_WRONLY;
         const RDWR = libc::O_RDWR;
         #[cfg(any(target_os = "linux",
+                  target_os = "android",
                   target_os = "netbsd",
                   target_os = "openbsd",
                   target_os = "wasi",
                   target_os = "emscripten"))]
-        const RSYNC = libc::O_RSYNC;
+        const RSYNC = {
+            // Have to use cfg_if: https://github.com/bitflags/bitflags/issues/137
+            cfg_if! {
+                if #[cfg(any(target_os = "linux",
+                             target_os = "netbsd",
+                             target_os = "openbsd",
+                             target_os = "wasi",
+                             target_os = "emscripten"))] {
+                    libc::O_RSYNC
+                } else if #[cfg(target_os = "android")] {
+                    // Android defines O_RSYNC as O_SYNC
+                    libc::O_SYNC
+                }
+            }
+        };
         const SYNC = libc::O_SYNC;
         const TRUNC = libc::O_TRUNC;
         #[cfg(any(target_os = "linux",
diff --git a/crates/wasi-common/yanix/src/sys/linux/filetime.rs b/crates/wasi-common/yanix/src/sys/linux/filetime.rs
index 3c4cdfa9f6..34faefc6d6 100644
--- a/crates/wasi-common/yanix/src/sys/linux/filetime.rs
+++ b/crates/wasi-common/yanix/src/sys/linux/filetime.rs
@@ -53,5 +53,8 @@ pub fn utimensat(
         return Err(err);
     }
 
-    super::utimesat::utimesat(dirfd, path, atime, mtime, symlink_nofollow)
+    #[cfg(not(target_os = "android"))]
+    return super::utimesat::utimesat(dirfd, path, atime, mtime, symlink_nofollow);
+    #[cfg(target_os = "android")]
+    unreachable!();
 }
diff --git a/crates/wasi-common/yanix/src/sys/linux/mod.rs b/crates/wasi-common/yanix/src/sys/linux/mod.rs
index 26432d36dd..37d8a0a8af 100644
--- a/crates/wasi-common/yanix/src/sys/linux/mod.rs
+++ b/crates/wasi-common/yanix/src/sys/linux/mod.rs
@@ -2,4 +2,5 @@ pub(crate) mod dir;
 pub(crate) mod fadvise;
 pub(crate) mod file;
 pub(crate) mod filetime;
+#[cfg(not(target_os = "android"))]
 pub(crate) mod utimesat;
diff --git a/crates/wasmtime/Cargo.toml b/crates/wasmtime/Cargo.toml
index 1fa4f012f4..9f22c01c40 100644
--- a/crates/wasmtime/Cargo.toml
+++ b/crates/wasmtime/Cargo.toml
@@ -16,7 +16,7 @@ wasmtime-jit = { path = "../jit", version = "0.21.0" }
 wasmtime-cache = { path = "../cache", version = "0.21.0", optional = true }
 wasmtime-profiling = { path = "../profiling", version = "0.21.0" }
 target-lexicon = { version = "0.11.0", default-features = false }
-wasmparser = "0.70"
+wasmparser = "0.71"
 anyhow = "1.0.19"
 region = "2.2.0"
 libc = "0.2"
diff --git a/deny.toml b/deny.toml
new file mode 100644
index 0000000000..37e5a3a0f8
--- /dev/null
+++ b/deny.toml
@@ -0,0 +1,45 @@
+# Documentation for this configuration file can be found here
+# https://embarkstudios.github.io/cargo-deny/checks/cfg.html
+
+targets = [
+    { triple = "x86_64-unknown-linux-gnu" },
+    { triple = "x86_64-apple-darwin" },
+    { triple = "x86_64-pc-windows-msvc" },
+    { triple = "aarch64-linux-android" },
+]
+
+# https://embarkstudios.github.io/cargo-deny/checks/advisories/cfg.html
+[advisories]
+vulnerability = "deny"
+unmaintained = "deny"
+yanked = "deny"
+ignore = [
+]
+
+# https://embarkstudios.github.io/cargo-deny/checks/licenses/cfg.html
+[licenses]
+allow = [
+    "Apache-2.0 WITH LLVM-exception",
+    "Apache-2.0",
+    "BSD-2-Clause",
+    "CC0-1.0",
+    "MIT",
+    "MPL-2.0",
+    "Zlib",
+]
+
+# https://embarkstudios.github.io/cargo-deny/checks/bans/cfg.html
+[bans]
+multiple-versions = "deny"
+wildcards = "allow"
+deny = []
+
+# Skip some multiple-versions checks, until they can be fixed.
+skip = [
+    { name = "ansi_term" }, # transitive dependencies only
+    { name = "cfg-if" }, # transitive dependencies use 0.1
+    { name = "env_logger" }, # pretty_env_logger and file-per-thread-logger depend on 0.7
+    { name = "humantime" }, # caused by env_logger
+    { name = "getrandom" }, # rand not updates to 0.2 yet
+    { name = "wast" }, # old one pulled in by witx
+]
diff --git a/docs/WASI-tutorial.md b/docs/WASI-tutorial.md
index 2c5cc84bba..8a01ce5f19 100644
--- a/docs/WASI-tutorial.md
+++ b/docs/WASI-tutorial.md
@@ -121,14 +121,14 @@ use std::io::{Read, Write};
 
 fn process(input_fname: &str, output_fname: &str) -> Result<(), String> {
     let mut input_file =
-        fs::File::open(input_fname).map_err(|err| format!("error opening input: {}", err))?;
+        fs::File::open(input_fname).map_err(|err| format!("error opening input {}: {}", input_fname, err))?;
     let mut contents = Vec::new();
     input_file
         .read_to_end(&mut contents)
         .map_err(|err| format!("read error: {}", err))?;
 
     let mut output_file = fs::File::create(output_fname)
-        .map_err(|err| format!("error opening output '{}': {}", output_fname, err))?;
+        .map_err(|err| format!("error opening output {}: {}", output_fname, err))?;
     output_file
         .write_all(&contents)
         .map_err(|err| format!("write error: {}", err))
@@ -139,7 +139,7 @@ fn main() {
     let program = args[0].clone();
 
     if args.len() < 3 {
-        eprintln!("{} <input_file> <output_file>", program);
+        eprintln!("usage: {} <input_file> <output_file>", program);
         return;
     }
 
diff --git a/src/lib.rs b/src/lib.rs
index b7eaaa8a09..39bda21f66 100644
--- a/src/lib.rs
+++ b/src/lib.rs
@@ -127,6 +127,10 @@ struct CommonOptions {
     #[structopt(long)]
     enable_multi_memory: bool,
 
+    /// Enable support for the module-linking proposal
+    #[structopt(long)]
+    enable_module_linking: bool,
+
     /// Enable all experimental Wasm features
     #[structopt(long)]
     enable_all: bool,
@@ -198,6 +202,7 @@ impl CommonOptions {
             .wasm_multi_value(self.enable_multi_value.unwrap_or(true) || self.enable_all)
             .wasm_threads(self.enable_threads || self.enable_all)
             .wasm_multi_memory(self.enable_multi_memory || self.enable_all)
+            .wasm_module_linking(self.enable_module_linking || self.enable_all)
             .cranelift_opt_level(self.opt_level())
             .strategy(pick_compilation_strategy(self.cranelift, self.lightbeam)?)?
             .profiler(pick_profiling_strategy(self.jitdump, self.vtune)?)?