[lld] [lld][LoongArch] Relax call36/tail36: R_LARCH_CALL36 (PR #123576)

[Zhaoxin Yang via llvm-commits]

================
@@ -830,6 +832,37 @@ static void relaxPCHi20Lo12(Ctx &ctx, const InputSection &sec, size_t i,
   remove = 4;
 }
 
+// Relax code sequence.
+// From:
+//   pcaddu18i $ra, %call36(foo)
+//   jirl $ra, $ra, 0
+// To:
+//   b/bl foo
+static void relaxCall36(Ctx &ctx, const InputSection &sec, size_t i,
+                        uint64_t loc, Relocation &r, uint32_t &remove) {
+  const uint64_t dest =
+      (r.expr == R_PLT_PC ? r.sym->getPltVA(ctx) : r.sym->getVA(ctx)) +
+      r.addend;
+
+  const int64_t displace = dest - loc;
+  // Check if the displace aligns 4 bytes or exceeds the range of b[l].
+  if ((displace & 0x3) != 0 || !isInt<28>(displace))
+    return;
+
+  const uint32_t nextInsn = read32le(sec.content().data() + r.offset + 4);
----------------
ylzsx wrote:

I believe this is not necessary. Here are the reasons:
1. Existing compilers, whether GCC or Clang, always generate the correct instruction sequence corresponding to the R_LARCH_CALL36 relocation when it is produced.
2. For handwritten assembly, correct use of pseudo-instructions will not lead to errors. For maliciously written incorrect assembly cases, we do not need to ensure their correctness, as such erroneous cases can always be constructed in most architectures.
3. Instruction checking should be avoided as much as possible in lld. (https://github.com/llvm/llvm-project/pull/127312/files#r1957198091)

It would be greatly appreciated if you could provide an example of an illegal case or present more substantial reasons.

https://github.com/llvm/llvm-project/pull/123576