[PATCH] D106421: Encode address offsets of basic blocks relative to the end of the previous basic blocks.

[Rahman Lavaee via Phabricator via llvm-commits]

rahmanl created this revision.
rahmanl added a reviewer: tmsriram.
Herald added subscribers: pengfei, rupprecht, hiraditya, emaste.
Herald added a reviewer: jhenderson.
rahmanl requested review of this revision.
Herald added subscribers: llvm-commits, MaskRay.
Herald added a project: LLVM.

This encoding uses smaller values compared to the existing one (offsets relative to function symbol).
Smaller values tend to occupy fewer bytes in ULEB128 encoding. As a result, we get about 25% reduction
in the size of the bb-address-map section. (reduction from about 9MB to 7MB).


Repository:
  rG LLVM Github Monorepo

https://reviews.llvm.org/D106421

Files:
  llvm/lib/CodeGen/AsmPrinter/AsmPrinter.cpp
  llvm/lib/CodeGen/BasicBlockSections.cpp
  llvm/test/CodeGen/X86/basic-block-sections-labels.ll
  llvm/test/tools/llvm-readobj/ELF/bb-addr-map.test
  llvm/tools/llvm-readobj/ELFDumper.cpp


Index: llvm/tools/llvm-readobj/ELFDumper.cpp
===================================================================
--- llvm/tools/llvm-readobj/ELFDumper.cpp
+++ llvm/tools/llvm-readobj/ELFDumper.cpp
@@ -6753,10 +6753,13 @@
       W.printString("Name", FuncName);
 
       ListScope L(W, "BB entries");
+      uint32_t FunctionRelativeAddress = 0;
       for (const typename Elf_BBAddrMap::BBEntry &BBE : AM.BBEntries) {
         DictScope L(W);
-        W.printHex("Offset", BBE.Offset);
+        FunctionRelativeAddress += BBE.Offset;
+        W.printHex("Offset", FunctionRelativeAddress);
         W.printHex("Size", BBE.Size);
+        FunctionRelativeAddress += BBE.Size;
         W.printBoolean("HasReturn", BBE.HasReturn);
         W.printBoolean("HasTailCall", BBE.HasTailCall);
         W.printBoolean("IsEHPad", BBE.IsEHPad);
Index: llvm/test/tools/llvm-readobj/ELF/bb-addr-map.test
===================================================================
--- llvm/test/tools/llvm-readobj/ELF/bb-addr-map.test
+++ llvm/test/tools/llvm-readobj/ELF/bb-addr-map.test
@@ -30,7 +30,7 @@
 # LLVM-NEXT:         CanFallThrough: No
 # LLVM-NEXT:       }
 # LLVM-NEXT:       {
-# LLVM-NEXT:         Offset: 0x3
+# LLVM-NEXT:         Offset: 0x4
 # LLVM-NEXT:         Size: 0x4
 # LLVM-NEXT:         HasReturn: Yes
 # LLVM-NEXT:         HasTailCall: No
Index: llvm/test/CodeGen/X86/basic-block-sections-labels.ll
===================================================================
--- llvm/test/CodeGen/X86/basic-block-sections-labels.ll
+++ llvm/test/CodeGen/X86/basic-block-sections-labels.ll
@@ -51,12 +51,12 @@
 ; CHECK-NEXT:	.uleb128 .Lfunc_begin0-.Lfunc_begin0
 ; CHECK-NEXT:	.uleb128 .LBB_END0_0-.Lfunc_begin0
 ; CHECK-NEXT:	.byte	8
-; CHECK-NEXT:	.uleb128 .LBB0_1-.Lfunc_begin0
+; CHECK-NEXT:	.uleb128 .LBB0_1-.LBB_END0_0
 ; CHECK-NEXT:	.uleb128 .LBB_END0_1-.LBB0_1
 ; CHECK-NEXT:	.byte	8
-; CHECK-NEXT:	.uleb128 .LBB0_2-.Lfunc_begin0
+; CHECK-NEXT:	.uleb128 .LBB0_2-.LBB_END0_1
 ; CHECK-NEXT:	.uleb128 .LBB_END0_2-.LBB0_2
 ; CHECK-NEXT:	.byte	1
-; CHECK-NEXT:	.uleb128 .LBB0_3-.Lfunc_begin0
+; CHECK-NEXT:	.uleb128 .LBB0_3-.LBB_END0_2
 ; CHECK-NEXT:	.uleb128 .LBB_END0_3-.LBB0_3
 ; CHECK-NEXT:	.byte	5
Index: llvm/lib/CodeGen/BasicBlockSections.cpp
===================================================================
--- llvm/lib/CodeGen/BasicBlockSections.cpp
+++ llvm/lib/CodeGen/BasicBlockSections.cpp
@@ -48,7 +48,7 @@
 // Basic Block Labels
 // ==================
 //
-// With -fbasic-block-sections=labels, we emit the offsets of BB addresses of
+// With -fbasic-block-sections=labels, we encode the offsets of BB addresses of
 // every function into the .llvm_bb_addr_map section. Along with the function
 // symbols, this allows for mapping of virtual addresses in PMU profiles back to
 // the corresponding basic blocks. This logic is implemented in AsmPrinter. This
Index: llvm/lib/CodeGen/AsmPrinter/AsmPrinter.cpp
===================================================================
--- llvm/lib/CodeGen/AsmPrinter/AsmPrinter.cpp
+++ llvm/lib/CodeGen/AsmPrinter/AsmPrinter.cpp
@@ -1137,16 +1137,19 @@
   OutStreamer->emitSymbolValue(FunctionSymbol, getPointerSize());
   // Emit the total number of basic blocks in this function.
   OutStreamer->emitULEB128IntValue(MF.size());
+  const MCSymbol *PrevMBBEndSymbol = FunctionSymbol;
   // Emit BB Information for each basic block in the funciton.
   for (const MachineBasicBlock &MBB : MF) {
     const MCSymbol *MBBSymbol =
         MBB.isEntryBlock() ? FunctionSymbol : MBB.getSymbol();
-    // Emit the basic block offset.
-    emitLabelDifferenceAsULEB128(MBBSymbol, FunctionSymbol);
+    // Emit the basic block offset relative to the end of the previous block.
+    // This is zero unless the block is padded due to alignment.
+    emitLabelDifferenceAsULEB128(MBBSymbol, PrevMBBEndSymbol);
     // Emit the basic block size. When BBs have alignments, their size cannot
     // always be computed from their offsets.
     emitLabelDifferenceAsULEB128(MBB.getEndSymbol(), MBBSymbol);
     OutStreamer->emitULEB128IntValue(getBBAddrMapMetadata(MBB));
+    PrevMBBEndSymbol = MBB.getEndSymbol();
   }
   OutStreamer->PopSection();
 }


-------------- next part --------------
A non-text attachment was scrubbed...
Name: D106421.360359.patch
Type: text/x-patch
Size: 4198 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20210721/d2fe06dc/attachment-0001.bin>