[WebAssembly] Error out when Emscripten SjLj setjmp is used with Wasm EH

[aheejin]

Summary

Currently, when Wasm EH is used with Emscripten SjLj, Emscripten SjLj
cannot handle invoke instructions - it assumes all invokes have been
lowered away with Emscripten EH. But in Wasm EH they are lowered in
instruction selection, so they are still present in the IR stage. This
happens when

1. Wasm EH and Emscripten SjLj are used together
2. A function that calls setjmp uses exceptions, i.e., has invokes

We were already erroring out with an assertion failure in this case, but
this CL makes it error out more properly with a valid error message.

Wasm EH + Wasm SjLj will not have this restrictions. (it will have
another restriction though, e.g., setjmp cannot be called within
catch. But why would anyone do that..)

[dschuff]

Actually wait, doesn't report_fatal_error act like a crash, complete with a stacktrace now?

[dschuff]

In D107687#2942166, @dschuff wrote:

Actually wait, doesn't report_fatal_error act like a crash, complete with a stacktrace now?

Nevermind, I was confused. The test's not --crash actually does check that llc "crashes".
I wish we had a better way to report errors from the backend.

[aheejin]

I agree. I'm not sure what more graceful way to report errors to users might be, because report_fatal_error crashes pretty like a bug (which actually in many case it is for) with a giant stack trace and it doesn't look that much different from hitting an assert. But anyway, I think this is still at least slightly better than the previous assertion....

llvm/lib/Target/WebAssembly/WebAssemblyLowerEmscriptenEHSjLj.cpp

bool WebAssemblyLowerEmscriptenEHSjLj::runSjLjOnFunction(Function &F) {
   std::vector<BasicBlock *> BBs;
   for (BasicBlock &BB : F)
     BBs.push_back(&BB);
 
   // BBs.size() will change within the loop, so we query it every time
   for (unsigned I = 0; I < BBs.size(); I++) {
     BasicBlock *BB = BBs[I];
     for (Instruction &I : *BB) {
-      assert(!isa<InvokeInst>(&I));
+      if (isa<InvokeInst>(&I))
+        report_fatal_error("When using Wasm EH with Emscripten SjLj, there is "
+                           "a restriction that `setjmp` function call and "
+                           "exception cannot be used within the same function");
       auto *CI = dyn_cast<CallInst>(&I);
       if (!CI)
         continue;
 
       const Value *Callee = CI->getCalledOperand();
       if (!canLongjmp(Callee))
         continue;
       if (isEmAsmCall(Callee))

llvm/test/CodeGen/WebAssembly/wasm-eh-em-sjlj-error.ll

+; RUN: not --crash llc < %s -enable-emscripten-sjlj -exception-model=wasm 2>&1 | FileCheck %s
+
+target triple = "wasm32-unknown-unknown"
+
+%struct.__jmp_buf_tag = type { [6 x i32], i32, [32 x i32] }
+
+define void @wasm_eh_emscripten_sjlj_error() personality i8* bitcast (i32 (...)* @__gxx_wasm_personality_v0 to i8*) {
+entry:
+  %buf = alloca [1 x %struct.__jmp_buf_tag], align 16
+  %arraydecay = getelementptr inbounds [1 x %struct.__jmp_buf_tag], [1 x %struct.__jmp_buf_tag]* %buf, i32 0, i32 0
+  %call = call i32 @setjmp(%struct.__jmp_buf_tag* %arraydecay) #0
+  %arraydecay1 = getelementptr inbounds [1 x %struct.__jmp_buf_tag], [1 x %struct.__jmp_buf_tag]* %buf, i32 0, i32 0
+  br label %bb
+
+bb:
+  invoke void @foo()
+          to label %try.cont unwind label %catch.dispatch
+
+catch.dispatch:                                   ; preds = %entry
+  %0 = catchswitch within none [label %catch.start] unwind to caller
+
+catch.start:                                      ; preds = %catch.dispatch
+  %1 = catchpad within %0 [i8* null]
+  %2 = call i8* @llvm.wasm.get.exception(token %1)
+  %3 = call i32 @llvm.wasm.get.ehselector(token %1)
+  %4 = call i8* @__cxa_begin_catch(i8* %2) #2 [ "funclet"(token %1) ]
+  call void @__cxa_end_catch() [ "funclet"(token %1) ]
+  catchret from %1 to label %try.cont
+
+try.cont:                                         ; preds = %entry, %catch.start
+  ret void
+}
+
+declare void @foo()
+declare i32 @__gxx_wasm_personality_v0(...)
+; Function Attrs: nounwind
+declare i8* @llvm.wasm.get.exception(token) #0
+; Function Attrs: nounwind
+declare i32 @llvm.wasm.get.ehselector(token) #0
+declare i8* @__cxa_begin_catch(i8*)
+declare void @__cxa_end_catch()
+; Function Attrs: returns_twice
+declare i32 @setjmp(%struct.__jmp_buf_tag*) #0
+; Function Attrs: noreturn
+declare void @longjmp(%struct.__jmp_buf_tag*, i32) #1
+declare i8* @malloc(i32)
+declare void @free(i8*)
+
+attributes #0 = { returns_twice }
+attributes #1 = { noreturn }
+attributes #2 = { nounwind }
+
+; CHECK: LLVM ERROR: When using Wasm EH with Emscripten SjLj, there is a restriction that `setjmp` function call and exception cannot be used within the same function