[PATCH] D71677: [ms] [X86] Use "P" modifier on operands to call instructions in inline X86 assembly.

[Reid Kleckner via Phabricator via llvm-commits]

rnk accepted this revision.
rnk added a comment.
This revision is now accepted and ready to land.

lgtm



================
Comment at: llvm/lib/Target/X86/AsmParser/X86AsmParser.cpp:2871
+  // differently when referenced in MS-style inline assembly.
+  if (Name.startswith("call") || Name.startswith("lcall")) {
+    for (size_t i = 1; i < Operands.size(); ++i) {
----------------
epastor wrote:
> rnk wrote:
> > I'm trying to think of other edge cases where we'd want the same treatment. In theory, we'd want to do this for every control flow instruction that takes a PC-relative operand, jmp, jcc, jecxz, that might be all. You know, it actually seems reasonable to set up a naked function that contains an asm blob which conditionally branches to another function, so I guess we should support it. In that case, maybe this should be named something like "isBranchTarget" instead of isCallTarget.
> That would be a valid option, but currently the "P" modifier is documented for LLVM as to-be-used on the operands of `call` instructions. [[ https://gcc.gnu.org/onlinedocs/gcc/Extended-Asm.html#x86Operandmodifiers | GNU as ]]) adds it more generally to be applied to functions - and occasionally constants, if you need a constant to avoid all syntax-specific prefixes.
> 
> We could use this for any branch target operand... but we'd need to restrict it to apply only to the specific PC-relative operand, which I think means augmenting the X86 target tables to annotate which operands of which instructions are PC-relative? Also, I'm not sure if that would break pre-existing code in enough cases to be worrying.
I see, so LLVM inherited P from GCC, and it mostly relates to `sym at PLT` suffixes, but it also suppresses `sym(%rip)` suffixes, and it is specific to calls. In that case, this all makes sense, let's not overgeneralize to all branch targets.


================
Comment at: llvm/test/CodeGen/X86/ms-inline-asm-PR44272.ll:15
+; CHECK: {{## InlineAsm Start|#APP}}
+; CHECK: call{{l|q}} func
+; CHECK: {{## InlineAsm End|#NO_APP}}
----------------
I'd suggest matching for `{{call(l|q) func$}}` so that you don't accidentally match `callq func(%rip)`. Which makes me wonder, does `@PLT` appear here? The test uses no OS in the triple, which probably means ELF, so we probably use a PLT.


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D71677/new/

https://reviews.llvm.org/D71677