winch: Gracefully handle compilation errors (#9851)

* winch: Gracefully handle compilation errors

Closes: #8096

This commit threads `anyhow::Result`  through most of Winch's
compilation process in order to gracefully handle compilation errors
gracefully instead of panicking.

The error classification is intentionally very granular, to avoid string
allocation which could impact compilation performance.

The errors are largely fit in two categories:

* Unimplemented/Unsupported
* Internal

The firs category signals partial or no support for Wasmtime features
and or Wasm proposals. These errors are meant to be temporary while
such features or proposals are in development.

The second category signals that a compilation invariant was not met.
These errors are considered internal and their presence usually means
a bug in the compiler.

* Include `Result` in the MacroAssembler

This commit updates the MacroAssembler trait to require returning
`Result<T>`  on every method in the interface, making it easier to
detect partial support for Masm instructions.

winch/codegen/Cargo.toml

@@ -25,6 +25,7 @@ regalloc2 = { workspace = true }
 gimli = { workspace = true }
 wasmtime-environ = { workspace = true }
 wasmtime-cranelift = { workspace = true }
+thiserror = { workspace = true }
 
 [features]
 x64 = ["cranelift-codegen/x86"]

winch/codegen/src/codegen/bounds.rs

@@ -9,6 +9,7 @@ use crate::{
     masm::{IntCmpKind, MacroAssembler, OperandSize, RegImm, TrapCode},
     stack::TypedReg,
 };
+use anyhow::Result;
 use wasmtime_environ::Signed;
 
 /// A newtype to represent an immediate offset argument for a heap access.
@@ -86,30 +87,33 @@ pub(crate) fn load_dynamic_heap_bounds<M>(
     masm: &mut M,
     heap: &HeapData,
     ptr_size: OperandSize,
-) -> Bounds
+) -> Result<Bounds>
 where
     M: MacroAssembler,
 {
-    let dst = context.any_gpr(masm);
+    let dst = context.any_gpr(masm)?;
     match heap.memory.static_heap_size() {
         // Constant size, no need to perform a load.
-        Some(size) => masm.mov(writable!(dst), RegImm::i64(size.signed()), ptr_size),
+        Some(size) => masm.mov(writable!(dst), RegImm::i64(size.signed()), ptr_size)?,
 
         None => {
             let scratch = scratch!(M);
             let base = if let Some(offset) = heap.import_from {
-                let addr = masm.address_at_vmctx(offset);
-                masm.load_ptr(addr, writable!(scratch));
+                let addr = masm.address_at_vmctx(offset)?;
+                masm.load_ptr(addr, writable!(scratch))?;
                 scratch
             } else {
                 vmctx!(M)
             };
-            let addr = masm.address_at_reg(base, heap.current_length_offset);
-            masm.load_ptr(addr, writable!(dst));
+            let addr = masm.address_at_reg(base, heap.current_length_offset)?;
+            masm.load_ptr(addr, writable!(dst))?;
         }
     }
 
-    Bounds::from_typed_reg(TypedReg::new(heap.index_type(), dst))
+    Ok(Bounds::from_typed_reg(TypedReg::new(
+        heap.index_type(),
+        dst,
+    )))
 }
 
 /// This function ensures the following:
@@ -125,10 +129,10 @@ pub(crate) fn ensure_index_and_offset<M: MacroAssembler>(
     index: Index,
     offset: u64,
     heap_ty_size: OperandSize,
-) -> ImmOffset {
+) -> Result<ImmOffset> {
     match u32::try_from(offset) {
         // If the immediate offset fits in a u32, then we simply return.
-        Ok(offs) => ImmOffset::from_u32(offs),
+        Ok(offs) => Ok(ImmOffset::from_u32(offs)),
         // Else we adjust the index to be index = index + offset, including an
         // overflow check, and return 0 as the offset.
         Err(_) => {
@@ -138,9 +142,9 @@ pub(crate) fn ensure_index_and_offset<M: MacroAssembler>(
                 RegImm::i64(offset as i64),
                 heap_ty_size,
                 TrapCode::HEAP_OUT_OF_BOUNDS,
-            );
+            )?;
 
-            ImmOffset::from_u32(0)
+            Ok(ImmOffset::from_u32(0))
         }
     }
 }
@@ -157,28 +161,28 @@ pub(crate) fn load_heap_addr_checked<M, F>(
     index: Index,
     offset: ImmOffset,
     mut emit_check_condition: F,
-) -> Reg
+) -> Result<Reg>
 where
     M: MacroAssembler,
-    F: FnMut(&mut M, Bounds, Index) -> IntCmpKind,
+    F: FnMut(&mut M, Bounds, Index) -> Result<IntCmpKind>,
 {
-    let cmp_kind = emit_check_condition(masm, bounds, index);
+    let cmp_kind = emit_check_condition(masm, bounds, index)?;
 
-    masm.trapif(cmp_kind, TrapCode::HEAP_OUT_OF_BOUNDS);
-    let addr = context.any_gpr(masm);
+    masm.trapif(cmp_kind, TrapCode::HEAP_OUT_OF_BOUNDS)?;
+    let addr = context.any_gpr(masm)?;
 
-    load_heap_addr_unchecked(masm, heap, index, offset, addr, ptr_size);
+    load_heap_addr_unchecked(masm, heap, index, offset, addr, ptr_size)?;
     if !enable_spectre_mitigation {
-        addr
+        Ok(addr)
     } else {
         // Conditionally assign 0 to the register holding the base address if
         // the comparison kind is met.
-        let tmp = context.any_gpr(masm);
-        masm.mov(writable!(tmp), RegImm::i64(0), ptr_size);
-        let cmp_kind = emit_check_condition(masm, bounds, index);
-        masm.cmov(writable!(addr), tmp, cmp_kind, ptr_size);
+        let tmp = context.any_gpr(masm)?;
+        masm.mov(writable!(tmp), RegImm::i64(0), ptr_size)?;
+        let cmp_kind = emit_check_condition(masm, bounds, index)?;
+        masm.cmov(writable!(addr), tmp, cmp_kind, ptr_size)?;
         context.free_reg(tmp);
-        addr
+        Ok(addr)
     }
 }
 
@@ -192,14 +196,15 @@ pub(crate) fn load_heap_addr_unchecked<M>(
     offset: ImmOffset,
     dst: Reg,
     ptr_size: OperandSize,
-) where
+) -> Result<()>
+where
     M: MacroAssembler,
 {
     let base = if let Some(offset) = heap.import_from {
         // If the WebAssembly memory is imported, load the address into
         // the scratch register.
         let scratch = scratch!(M);
-        masm.load_ptr(masm.address_at_vmctx(offset), writable!(scratch));
+        masm.load_ptr(masm.address_at_vmctx(offset)?, writable!(scratch))?;
         scratch
     } else {
         // Else if the WebAssembly memory is defined in the current module,
@@ -208,17 +213,18 @@ pub(crate) fn load_heap_addr_unchecked<M>(
     };
 
     // Load the base of the memory into the `addr` register.
-    masm.load_ptr(masm.address_at_reg(base, heap.offset), writable!(dst));
+    masm.load_ptr(masm.address_at_reg(base, heap.offset)?, writable!(dst))?;
     // Start by adding the index to the heap base addr.
     let index_reg = index.as_typed_reg().reg;
-    masm.add(writable!(dst), dst, index_reg.into(), ptr_size);
+    masm.add(writable!(dst), dst, index_reg.into(), ptr_size)?;
 
     if offset.as_u32() > 0 {
         masm.add(
             writable!(dst),
             dst,
             RegImm::i64(offset.as_u32() as i64),
             ptr_size,
-        );
+        )?;
     }
+    Ok(())
 }