[Clang][RISCV] Implement getConstraintRegister for RISC-V

[luismarques]

Summary

The getConstraintRegister method is used by semantic checking of inline assembly statements in order to diagnose conflicts between clobber list and input/output lists. By overriding getConstraintRegister we get those diagnostics and we match RISC-V GCC's behavior. The implementation is trivial due to the lack of single-register RISC-V-specific constraints.

[luismarques]

Nit: remove nop.

[luismarques]

clang/test/Sema/inline-asm-validate-riscv.c

27–28

I don't really understand the point of erroring-out in these two cases where the register is an input and is also clobbered. In fact, I've run into a case where that would be useful and accurately reflected the situation. But GCC's documentation explicitly prohibits that. I'm not sure if there's a fundamental reason for that, or if it's just an implementation quirk.

[asb]

Looks good to me - I'm surprised only Arm, AArch64, and X86 implement this!

clang/lib/Basic/Targets/RISCV.h

public:
   ArrayRef<Builtin::Info> getTargetBuiltins() const override;
 
   BuiltinVaListKind getBuiltinVaListKind() const override {
     return TargetInfo::VoidPtrBuiltinVaList;
   }
 
   const char *getClobbers() const override { return ""; }
 
+  StringRef getConstraintRegister(StringRef Constraint,
+                                  StringRef Expression) const override {
+    return Expression;
+  }
+
   ArrayRef<const char *> getGCCRegNames() const override;
 
   int getEHDataRegisterNumber(unsigned RegNo) const override {
     if (RegNo == 0)
       return 10;
     else if (RegNo == 1)
       return 11;
     else

clang/test/Sema/inline-asm-validate-riscv.c

 }
 
 void K(int k) {
   static const int BelowMin = -1;
   static const int AboveMax = 32;
   asm volatile ("" :: "K"(BelowMin)); // expected-error{{value '-1' out of range for constraint 'K'}}
   asm volatile ("" :: "K"(AboveMax)); // expected-error{{value '32' out of range for constraint 'K'}}
 }
+
+void test_clobber_conflict(void) {
+  register long x10 asm("x10");
+  asm volatile("nop" :: "r"(x10) : "x10"); // expected-error {{conflicts with asm clobber list}}
+  asm volatile("nop" :: "r"(x10) : "a0"); // expected-error {{conflicts with asm clobber list}}
+  asm volatile("nop" : "=r"(x10) :: "x10"); // expected-error {{conflicts with asm clobber list}}
+  asm volatile("nop" : "=r"(x10) :: "a0"); // expected-error {{conflicts with asm clobber list}}
+}