diff --git a/crates/fuzzing/src/generators/config.rs b/crates/fuzzing/src/generators/config.rs
index e7f52e00cd07..c3fd38365d74 100644
--- a/crates/fuzzing/src/generators/config.rs
+++ b/crates/fuzzing/src/generators/config.rs
@@ -200,7 +200,8 @@ impl Config {
             || self.module_config.function_references_enabled
             || reference_types.unwrap_or(false);
         config.extended_const_enabled = extended_const.unwrap_or(false);
-        config.exceptions_enabled = exceptions.unwrap_or(false);
+        config.exceptions_enabled =
+            self.module_config.stack_switching || exceptions.unwrap_or(false);
         if multi_memory.unwrap_or(false) {
             config.max_memories = limits::MEMORIES_PER_MODULE as usize;
         } else {
diff --git a/crates/test-util/src/wast.rs b/crates/test-util/src/wast.rs
index dd6f5033491c..b6ec104358de 100644
--- a/crates/test-util/src/wast.rs
+++ b/crates/test-util/src/wast.rs
@@ -536,6 +536,7 @@ impl WastTest {
                 // shared memories + pooling allocator aren't supported yet
                 "misc_testsuite/memory-combos.wast",
                 "misc_testsuite/threads/atomics-end-of-memory.wast",
+                "misc_testsuite/threads/atomic_wait_endianness.wast",
                 "misc_testsuite/threads/LB.wast",
                 "misc_testsuite/threads/LB_atomic.wast",
                 "misc_testsuite/threads/MP.wast",
diff --git a/crates/wasmtime/src/runtime/vm/memory/shared_memory.rs b/crates/wasmtime/src/runtime/vm/memory/shared_memory.rs
index 679984c25063..9f0721be11f0 100644
--- a/crates/wasmtime/src/runtime/vm/memory/shared_memory.rs
+++ b/crates/wasmtime/src/runtime/vm/memory/shared_memory.rs
@@ -146,6 +146,9 @@ impl SharedMemory {
         assert!(std::mem::size_of::<AtomicU32>() == 4);
         assert!(std::mem::align_of::<AtomicU32>() <= 4);
         let atomic = unsafe { AtomicU32::from_ptr(addr.cast()) };
+        // Wasm linear memory is always little-endian, but `AtomicU32` uses the
+        // host's native endianness.
+        let expected = expected.to_le();
 
         // Note that `checked_add` is used such that when `timeout` is too large
         // it'll cause there to be no timeout at all if we can't represent the
@@ -175,6 +178,9 @@ impl SharedMemory {
         assert!(std::mem::size_of::<AtomicU64>() == 8);
         assert!(std::mem::align_of::<AtomicU64>() <= 8);
         let atomic = unsafe { AtomicU64::from_ptr(addr.cast()) };
+        // Wasm linear memory is always little-endian, but `AtomicU64` uses the
+        // host's native endianness.
+        let expected = expected.to_le();
 
         // See `atomic_wait32` for why this is using `checked_add`.
         let deadline = timeout.and_then(|d| Instant::now().checked_add(d));
diff --git a/tests/misc_testsuite/threads/atomic_wait_endianness.wast b/tests/misc_testsuite/threads/atomic_wait_endianness.wast
new file mode 100644
index 000000000000..ee111a0039d9
--- /dev/null
+++ b/tests/misc_testsuite/threads/atomic_wait_endianness.wast
@@ -0,0 +1,26 @@
+;;! threads = true
+
+(module
+  (memory (export "shared") 1 1 shared)
+
+  (func (export "eq32") (result i32)
+    (i32.atomic.store (i32.const 0) (i32.const 1))
+    (memory.atomic.wait32 (i32.const 0) (i32.const 1) (i64.const 0)))
+
+  (func (export "ne32") (result i32)
+    (i32.atomic.store (i32.const 0) (i32.const 1))
+    (memory.atomic.wait32 (i32.const 0) (i32.const 0x01000000) (i64.const 0)))
+
+  (func (export "eq64") (result i32)
+    (i64.atomic.store (i32.const 8) (i64.const 1))
+    (memory.atomic.wait64 (i32.const 8) (i64.const 1) (i64.const 0)))
+
+  (func (export "ne64") (result i32)
+    (i64.atomic.store (i32.const 8) (i64.const 1))
+    (memory.atomic.wait64 (i32.const 8) (i64.const 0x0100000000000000) (i64.const 0)))
+)
+
+(assert_return (invoke "eq32") (i32.const 2))
+(assert_return (invoke "ne32") (i32.const 1))
+(assert_return (invoke "eq64") (i32.const 2))
+(assert_return (invoke "ne64") (i32.const 1))