dschuff added a comment.

Yeah I think a 64-bit version of userstack.ll makes sense.  (a fork or a second 
set of CHECKs, whatever you think would be cleaner). There's also 
`stack-alignment.ll` which covers dynamic stack adjustment.



================
Comment at: lld/wasm/Driver.cpp:385
+    StringRef s = arg->getValue();
+    if (s == "wasm32")
+      config->is64 = false;
----------------
any particular reason this shouldn't use the more conventional `-m32`/`m64`?
edit: nevermind, this is lld not clang. I'll defer to Sam's opinion on lld 
flags.


================
Comment at: llvm/lib/Target/WebAssembly/WebAssemblyMCInstLower.cpp:84
     WasmSym->setGlobalType(wasm::WasmGlobalType{
-        uint8_t(Subtarget.hasAddr64() ? wasm::WASM_TYPE_I64
-                                      : wasm::WASM_TYPE_I32),
+        uint8_t(Subtarget.hasAddr64() && strcmp(Name, "__table_base") != 0
+                    ? wasm::WASM_TYPE_I64
----------------
should __table_base stay as i32?


CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D82130/new/

https://reviews.llvm.org/D82130



_______________________________________________
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

Reply via email to