[PATCH] D106128: [lld-macho] Use immediate encodings for bind opcodes

[Vincent Lee via Phabricator via llvm-commits]

thevinster created this revision.
Herald added a reviewer: int3.
Herald added a reviewer: gkm.
Herald added a project: lld-macho.
Herald added a reviewer: lld-macho.
thevinster requested review of this revision.
Herald added a project: LLVM.
Herald added a subscriber: llvm-commits.

Implement pass 3 of bind opcodes from ld64. This change
is already behind an O2 <https://reviews.llvm.org/owners/package/2/> flag so it shouldn't impact current performance.
I verified ld64's output with the output LLD and they were both emitting
the same optimized bind opcodes (although in a slightly different order).


Repository:
  rG LLVM Github Monorepo

https://reviews.llvm.org/D106128

Files:
  lld/MachO/SyntheticSections.cpp
  lld/test/MachO/bind-opcodes.s


Index: lld/test/MachO/bind-opcodes.s
===================================================================
--- lld/test/MachO/bind-opcodes.s
+++ lld/test/MachO/bind-opcodes.s
@@ -9,6 +9,7 @@
 ## 1/ We emit exactly one BIND_OPCODE_SET_SYMBOL_TRAILING_FLAGS_IMM per symbol.
 ## 2/ Combine BIND_OPCODE_DO_BIND and BIND_OPCODE_ADD_ADDR_ULEB pairs.
 ## 3/ Compact BIND_OPCODE_DO_BIND_ADD_ADDR_ULEB
+## 4/ Use BIND_OPCODE_DO_BIND_ADD_ADDR_IMM_SCALED if possible.
 # RUN: obj2yaml %t/test | FileCheck %s
 
 # CHECK:      BindOpcodes:
@@ -54,9 +55,8 @@
 # CHECK-NEXT:   Imm:             0
 # CHECK-NEXT:   ULEBExtraData:   [ 0xFFFFFFFFFFFFEFD0 ]
 # CHECK-NEXT:   Symbol:          ''
-# CHECK-NEXT:   Opcode:          BIND_OPCODE_DO_BIND_ADD_ADDR_ULEB
-# CHECK-NEXT:   Imm:             0
-# CHECK-NEXT:   ULEBExtraData:   [ 0x8 ]
+# CHECK-NEXT:   Opcode:          BIND_OPCODE_DO_BIND_ADD_ADDR_IMM_SCALED
+# CHECK-NEXT:   Imm:             1
 # CHECK-NEXT:   Symbol:          ''
 # CHECK-NEXT:   Opcode:          BIND_OPCODE_DO_BIND_ADD_ADDR_ULEB
 # CHECK-NEXT:   Imm:             0
Index: lld/MachO/SyntheticSections.cpp
===================================================================
--- lld/MachO/SyntheticSections.cpp
+++ lld/MachO/SyntheticSections.cpp
@@ -361,6 +361,16 @@
   if (i == opcodes.size())
     opcodes[pWrite] = opcodes[i - 1];
   opcodes.resize(pWrite + 1);
+
+  // Pass 3: Use immediate encodings
+  for (BindIR* p = &opcodes[0]; p->opcode != BIND_OPCODE_DONE; ++p) {
+    if ((p->opcode == BIND_OPCODE_DO_BIND_ADD_ADDR_ULEB)
+          && (p->data < (15*sizeof(uint64_t)))
+          && ((p->data % sizeof(uint64_t)) == 0)) {
+      p->opcode = BIND_OPCODE_DO_BIND_ADD_ADDR_IMM_SCALED;
+      p->data = p->data/sizeof(uint64_t);
+    }
+  }
 }
 
 static void flushOpcodes(const BindIR &op, raw_svector_ostream &os) {
@@ -384,6 +394,9 @@
     encodeULEB128(op.consecutiveCount, os);
     encodeULEB128(op.data, os);
     break;
+  case BIND_OPCODE_DO_BIND_ADD_ADDR_IMM_SCALED:
+    os << static_cast<uint8_t>(op.opcode | op.data);
+    break;
   default:
     llvm_unreachable("cannot bind to an unrecognized symbol");
   }


-------------- next part --------------
A non-text attachment was scrubbed...
Name: D106128.359222.patch
Type: text/x-patch
Size: 2135 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20210716/423d3a79/attachment.bin>