[RISCV][MC] Allow symbol diff expression directly as LI operand

[jobnoorman]

Summary

D135960 implemented support for symbol differences as an operand of LI.
However, it only worked when the symbol difference expression was first
stored in a symbol using .set/.equ. This patch makes sure the
expression can be used directly as an LI operand.

Note that when using -filetype=asm, the instruction printer is not
able to convert the generated ADDI back to LI, presumably because the
symbol difference expression is not recognized as an simm12 so the
alias does not match. This means a .s -> .s round-trip produces an
ADDI instruction where there used to be an LI. Since ADDI doesn't
currently accept symbol difference expressions, the result of the
round-trip is rejected by MC.

I'm not quite sure what the best way forward is here. Is there a way to
recognize this alias without implementing support for symbol differences
for ADDI (which would be a separate patch)? Or can we just ignore this
issue for now as it seems unlikely to cause problems?

Fixes #64623

[wangpc]

FYI: I think D157694 is related.

[jobnoorman]

In D158830#4616596, @wangpc wrote:

FYI: I think D157694 is related.

Thanks for that! That patch covers the case here as well so might be preferable. If it gets accepted, I'll modify this patch to only keep the test cases.

[asb]

From discussing out-of-band with @jobnoorman we think D157694 augmented with the test cases from this patch is probably the way to go.

[jobnoorman]

Abandon in favor of D157694.

llvm/lib/Target/RISCV/MCTargetDesc/RISCVMCCodeEmitter.cpp

if (MIFrm == RISCVII::InstFormatJ) {
       FixupKind = RISCV::fixup_riscv_branch;
     } else if (MIFrm == RISCVII::InstFormatCJ) {
       FixupKind = RISCV::fixup_riscv_rvc_jump;
     } else if (MIFrm == RISCVII::InstFormatCB) {
       FixupKind = RISCV::fixup_riscv_rvc_branch;
     } else if (MIFrm == RISCVII::InstFormatI) {
       FixupKind = RISCV::fixup_riscv_12_i;
     }
+  } else if (Kind == MCExpr::Binary &&
+             cast<MCBinaryExpr>(Expr)->getOpcode() == MCBinaryExpr::Sub) {
+    // Needed to support "li xi, a-b" which is translated to "addi xi, x0, a-b"
+    if (MIFrm == RISCVII::InstFormatI)
+      FixupKind = RISCV::fixup_riscv_12_i;
   }
 
   assert(FixupKind != RISCV::fixup_riscv_invalid && "Unhandled expression!");
 
   Fixups.push_back(
       MCFixup::create(0, Expr, MCFixupKind(FixupKind), MI.getLoc()));
   ++MCNumFixups;
 

llvm/test/MC/RISCV/fixups-invalid.s

 # RUN: not llvm-mc -filetype=obj %s -triple=riscv32 -o /dev/null 2>&1 \
 # RUN:     | FileCheck %s
 # RUN: not llvm-mc -filetype=obj %s -triple=riscv64 -o /dev/null 2>&1 \
 # RUN:     | FileCheck %s
 
 .byte foo   # CHECK: [[@LINE]]:7: error: 1-byte data relocations not supported
 .2byte foo  # CHECK: [[@LINE]]:8: error: 2-byte data relocations not supported
 
 # Test that using li with a symbol difference constant rejects values that
 # cannot fit in a signed 12-bit integer.
 .Lbuf: .skip (1 << 11)
 .Lbuf_end:
 .equ CONST, .Lbuf_end - .Lbuf
 # CHECK: error: operand must be a constant 12-bit integer
 li a0, CONST
+# CHECK: error: operand must be a constant 12-bit integer
+li a0, .Lbuf_end - .Lbuf

llvm/test/MC/RISCV/rv32i-aliases-valid.s

 li a0, CONST
 
 .equ CONST, .Lbuf_end - .Lbuf
 # CHECK-ASM: li a0, CONST
 # CHECK-ASM-NOALIAS: addi a0, zero, CONST
 # CHECK-OBJ-NOALIAS: addi a0, zero, 8
 li a0, CONST
 
+# CHECK-ASM: addi a0, zero, .Lbuf_end-.Lbuf
+# CHECK-ASM-NOALIAS: addi a0, zero, .Lbuf_end-.Lbuf
+# CHECK-OBJ-NOALIAS: addi a0, zero, 8
+li a0, .Lbuf_end - .Lbuf
+
 # CHECK-INST: addi a0, zero, 0
 # CHECK-ALIAS: li a0, 0
 la x10, 0
 lla x10, 0
 # CHECK-INST: addi a0, zero, 1
 # CHECK-ALIAS: li a0, 1
 la x10, 1
 lla x10, 1

llvm/test/MC/RISCV/rv64i-aliases-valid.s

 li a0, CONST
 
 .equ CONST, .Lbuf_end - .Lbuf
 # CHECK-ASM: li a0, CONST
 # CHECK-ASM-NOALIAS: addi a0, zero, CONST
 # CHECK-OBJ-NOALIAS: addi a0, zero, 8
 li a0, CONST
 
+# CHECK-ASM: addi a0, zero, .Lbuf_end-.Lbuf
+# CHECK-ASM-NOALIAS: addi a0, zero, .Lbuf_end-.Lbuf
+# CHECK-OBJ-NOALIAS: addi a0, zero, 8
+li a0, .Lbuf_end - .Lbuf
+
 # CHECK-INST: addi a0, zero, 0
 # CHECK-ALIAS: li a0, 0
 la x10, 0
 lla x10, 0
 # CHECK-INST: addi a0, zero, 1
 # CHECK-ALIAS: li a0, 1
 la x10, 1
 lla x10, 1