[llvm-readelf, propeller] Add fallthrough bit to basic block metadata in BB-Address-Map section.

[rahmanl]

Summary

This patch adds a fallthrough bit to basic block metadata, indicating whether the basic block can fallthrough without taking any branches. The bit will help us avoid an intel LBR bug which results in occasional duplicate entries at the beginning of the LBR stack.

This patch uses MachineBasicBlock::canFallThrough() to set the bit. This is not a const method because it eventually calls TargetInstrInfo::analyzeBranch, but it calls this function with the default AllowModify=false. So we can either make the argument to the getBBAddrMapMetadata non-const, or we can use const_cast when calling canFallThrough. I decide to go with the latter since this is purely due to legacy code, and in general we should not allow the BasicBlock to be mutable during getBBAddrMapMetadata.

[rahmanl]

Rebase after landing c245c21c436b.

[rahmanl]

Modify comment for AsmPrinter::getBBAddrMapMetadata.

[shenhan]

LGTM

Thanks for making the change.

[jhenderson]

llvm-readobj aspect looks fine.

llvm/lib/CodeGen/AsmPrinter/AsmPrinter.cpp

1081

(whilst you're changing this line)

[rahmanl]

• Fix typo.
• Rebase.

[rahmanl]

Thanks for the quick review.
@tmsriram Would you please review the AsmPrinter.cpp change?

[tmsriram]

LGTM and is a simple change to capture the fall through bit.

llvm/include/llvm/Object/ELFTypes.h

template <class ELFT> struct Elf_BBAddrMap_Impl {
   uintX_t Addr; // Function address
   // Struct representing the BBAddrMap information for one basic block.
   struct BBEntry {
     uint32_t Offset; // Offset of basic block relative to function start.
     uint32_t Size;   // Size of the basic block.
 
     // The following fields are decoded from the Metadata field. The encoding
     // happens in AsmPrinter.cpp:getBBAddrMapMetadata.
-    bool HasReturn;   // If this block ends with a return (or tail call).
-    bool HasTailCall; // If this block ends with a tail call.
-    bool IsEHPad;     // If this is an exception handling block.
+    bool HasReturn;      // If this block ends with a return (or tail call).
+    bool HasTailCall;    // If this block ends with a tail call.
+    bool IsEHPad;        // If this is an exception handling block.
+    bool CanFallThrough; // If this block can fall through to its next.
 
     BBEntry(uint32_t Offset, uint32_t Size, uint32_t Metadata)
         : Offset(Offset), Size(Size), HasReturn(Metadata & 1),
-          HasTailCall(Metadata & (1 << 1)), IsEHPad(Metadata & (1 << 2)){};
+          HasTailCall(Metadata & (1 << 1)), IsEHPad(Metadata & (1 << 2)),
+          CanFallThrough(Metadata & (1 << 3)){};
   };
   std::vector<BBEntry> BBEntries; // Basic block entries for this function.
 };
 
 } // end namespace object.
 } // end namespace llvm.
 
 #endif // LLVM_OBJECT_ELFTYPES_H

llvm/lib/CodeGen/AsmPrinter/AsmPrinter.cpp

void AsmPrinter::emitFrameAlloc(const MachineInstr &MI) {
 
   // Emit a symbol assignment.
   OutStreamer->emitAssignment(FrameAllocSym,
                              MCConstantExpr::create(FrameOffset, OutContext));
 }
 
 /// Returns the BB metadata to be emitted in the .llvm_bb_addr_map section for a
 /// given basic block. This can be used to capture more precise profile
 /// information. We use the last 3 bits (LSBs) to ecnode the following
 /// information:
 ///  * (1): set if return block (ret or tail call).
 ///  * (2): set if ends with a tail call.
 ///  * (3): set if exception handling (EH) landing pad.
 /// The remaining bits are zero.
 static unsigned getBBAddrMapMetadata(const MachineBasicBlock &MBB) {
   const TargetInstrInfo *TII = MBB.getParent()->getSubtarget().getInstrInfo();
   return ((unsigned)MBB.isReturnBlock()) |
          ((!MBB.empty() && TII->isTailCall(MBB.back())) << 1) |
-         (MBB.isEHPad() << 2);
+         (MBB.isEHPad() << 2) |
+         (const_cast<MachineBasicBlock &>(MBB).canFallThrough() << 3);
 }
 
 void AsmPrinter::emitBBAddrMapSection(const MachineFunction &MF) {
   MCSection *BBAddrMapSection =
       getObjFileLowering().getBBAddrMapSection(*MF.getSection());
   assert(BBAddrMapSection && ".llvm_bb_addr_map section is not initialized.");
 
   const MCSymbol *FunctionSymbol = getFunctionBegin();

llvm/test/tools/llvm-readobj/ELF/bb-addr-map.test

 # LLVM-NEXT:     At: [[ADDR]]
 # LLVM-NEXT:     BB entries [
 # LLVM-NEXT:       {
 # LLVM-NEXT:         Offset: 0x0
 # LLVM-NEXT:         Size: 0x1
 # LLVM-NEXT:         HasReturn: No
 # LLVM-NEXT:         HasTailCall: Yes
 # LLVM-NEXT:         IsEHPad: No
+# LLVM-NEXT:         CanFallThrough: No
 # LLVM-NEXT:       }
 # LLVM-NEXT:       {
 # LLVM-NEXT:         Offset: 0x3
 # LLVM-NEXT:         Size: 0x4
 # LLVM-NEXT:         HasReturn: Yes
 # LLVM-NEXT:         HasTailCall: No
 # LLVM-NEXT:         IsEHPad: Yes
+# LLVM-NEXT:         CanFallThrough: No
 # LLVM-NEXT:       }
 # LLVM-NEXT:     ]
 # LLVM-NEXT:   }
 # LLVM-NEXT:   Function {
 # LLVM-NEXT:     At: 0x22222
 # LLVM-NEXT:     BB entries [
 # LLVM-NEXT:       {
 # LLVM-NEXT:         Offset: 0x6
 # LLVM-NEXT:         Size: 0x7
 # LLVM-NEXT:         HasReturn: No
 # LLVM-NEXT:         HasTailCall: No
 # LLVM-NEXT:         IsEHPad: No
+# LLVM-NEXT:         CanFallThrough: Yes
 # LLVM-NEXT:       }
 # LLVM-NEXT:     ]
 # LLVM-NEXT:   }
 # LLVM-NEXT: ]
 # LLVM-NEXT: BBAddrMap [
 # LLVM-NEXT:   Function {
 # LLVM-NEXT:     At: 0x33333
 # LLVM-NEXT:     BB entries [
 # LLVM-NEXT:       {
 # LLVM-NEXT:         Offset: 0x9
 # LLVM-NEXT:         Size: 0xA
 # LLVM-NEXT:         HasReturn: Yes
 # LLVM-NEXT:         HasTailCall: Yes
 # LLVM-NEXT:         IsEHPad: No
+# LLVM-NEXT:         CanFallThrough: Yes
 # LLVM-NEXT:       }
 # LLVM-NEXT:     ]
 # LLVM-NEXT:   }
 # LLVM-NEXT: ]
 
 # GNU: GNUStyle::printBBAddrMaps not implemented
 
 # TRUNCATED:      BBAddrMap [
 # TRUNCATED-NEXT:   warning: '[[FILE]]': unable to dump SHT_LLVM_BB_ADDR_MAP section with index 1: unable to decode LEB128 at offset [[OFFSET]]: malformed uleb128, extends past end
 # TRUNCATED-NEXT: ]
 ## Check that the other valid section is properly dumped.
 # TRUNCATED-NEXT: BBAddrMap [
 # TRUNCATED-NEXT:   Function {
 # TRUNCATED-NEXT:          At: 0x33333
 # TRUNCATED-NEXT:          BB entries [
 # TRUNCATED-NEXT:            {
 # TRUNCATED-NEXT:         Offset: 0x9
 # TRUNCATED-NEXT:         Size: 0xA
 # TRUNCATED-NEXT:         HasReturn: Yes
 # TRUNCATED-NEXT:         HasTailCall: Yes
 # TRUNCATED-NEXT:         IsEHPad: No
+# TRUNCATED-NEXT:         CanFallThrough: Yes
 # TRUNCATED-NEXT:       }
 # TRUNCATED-NEXT:     ]
 # TRUNCATED-NEXT:   }
 # TRUNCATED-NEXT: ]
 
 --- !ELF
 FileHeader:
   Class: ELFCLASS[[BITS]]

llvm/tools/llvm-readobj/ELFDumper.cpp

for (const Elf_BBAddrMap &AM : *BBAddrMapOrErr) {
       ListScope L(W, "BB entries");
       for (const typename Elf_BBAddrMap::BBEntry &BBE : AM.BBEntries) {
         DictScope L(W);
         W.printHex("Offset", BBE.Offset);
         W.printHex("Size", BBE.Size);
         W.printBoolean("HasReturn", BBE.HasReturn);
         W.printBoolean("HasTailCall", BBE.HasTailCall);
         W.printBoolean("IsEHPad", BBE.IsEHPad);
+        W.printBoolean("CanFallThrough", BBE.CanFallThrough);
       }
     }
   }
 }
 
 template <class ELFT> void LLVMELFDumper<ELFT>::printAddrsig() {
   ListScope L(W, "Addrsig");
   if (!this->DotAddrsigSec)