LLVM's BranchProbabilityInfo assigns every multi-successor terminator a probability distribution over its successors. This post describes the estimation used when no profile is available and reimplements it as a standalone program.

Read More

The dominator tree lets us identify natural loops: a back edge T->H whose head H dominates its tail T defines a loop with the single entry H. This works only for reducible control flow graphs. Optimized machine code and decompiler output routinely contain irreducible loops, which have more than one entry and thus no dominating header, so the dominator-based method cannot see them.

This post builds a loop-nesting forest for an arbitrary CFG with the single-pass depth-first search of 韦韬、毛剑、邹维、陈宇(Tao Wei, Jian Mao, Wei Zou & Yu Chen) A New Algorithm for Identifying Loops in Decompilation, SAS 2007 (The 14th International Static Analysis Symposium).

Read More

BumpPtrAllocator is LLVM's bump allocator (arena allocator): each allocation bumps a pointer within a slab, and everything is freed at once when the allocator dies. It backs Clang's ASTContext, lld's make<T> object pools, TableGen records, and many other arenas.

Here is the fast path before three recent changes:

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16

__attribute__((returns_nonnull)) void *Allocate(size_t Size, Align Alignment) {
  BytesAllocated += Size;                               // (3) accounting RMW
  uintptr_t AlignedPtr = alignAddr(CurPtr, Alignment);  // (1) always realign
  size_t SizeToAllocate = Size;
#if LLVM_ADDRESS_SANITIZER_BUILD
  SizeToAllocate += RedZoneSize;
#endif
  uintptr_t AllocEndPtr = AlignedPtr + SizeToAllocate;
  if (LLVM_LIKELY(AllocEndPtr <= uintptr_t(End)
                  && CurPtr != nullptr)) {              // (2) bound + null check
    CurPtr = reinterpret_cast<char *>(AllocEndPtr);
    ...
    return reinterpret_cast<char *>(AlignedPtr);
  }
  return AllocateSlow(Size, SizeToAllocate, Alignment);
}

Read More

tl;dr This blog post describes a recent SmallVector::push_back optimization for approximately trivially copyable element types.

SmallVector is LLVM's most-used container, and push_back its hot operation. For the trivially-copyable specialization the fast path should be fast.

1
2
3

#include <llvm/ADT/SmallVector.h>

void f(llvm::SmallVectorImpl<int> &v, int x) { v.push_back(x); }

clang -S --target=x86_64 -O2 -DNDEBUG a.cc generates:

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20

push   rbp                 # callee-saved spills + a stack realignment,
push   rbx                 # all on the fast path
push   rax
mov    eax, [rdi + 8]      # size
cmp    eax, [rdi + 12]     # vs capacity
jae    .Lgrow
.Lstore:                   # reached from the fast path AND from .Lgrow
mov    rcx, [rdi]
mov    [rcx + rax*4], esi
inc    dword ptr [rdi + 8]
add    rsp, 8
pop    rbx
pop    rbp
ret
.Lgrow:
mov    rbx, rdi            # keep `this`/`x` alive across the call
mov    ebp, esi
call   SmallVectorBase<unsigned>::grow_pod
...
jmp    .Lstore

Read More

LLVM has several hash tables. They used quadratic probing with in-band sentinel keys (empty, tombstone); recent work has been replacing that with linear probing with tombstone key removed.

• DenseMap (replacement for std::unordered_map): DenseMapInfo::getEmptyKey() / getTombstoneKey().
  • DenseSet: implemented using DenseMap
  • compiler-rt/lib/sanitizer_common/sanitizer_dense_map.h ports the implementation for sanitizers.
• SmallPtrSet (replacement for std::unordered_set<T *>): hard-coded -1 (empty) and -2 (tombstone).
• StringMap (replacement for std::unordered_map<std::string, V>)
  • StringSet: implemented using StringMap

For the open-addressed DenseMap and SmallPtrSet, pointers, references, and iterators are invalidated by insert. StringMap is different: each entry lives in a heap-allocated StringMapEntry<V> node, so entry pointers survive grow. std::unordered_map, being node-based, keeps surviving-element pointers valid across both insert and erase and only invalidates the erased element's own iterator. LLVM code rarely needs that stronger contract — callers do not hold long-lived references into the container across mutation — and that gap is what gives pass to relocating erase and bit-array occupancy.

Recently,

• Tombstones have been removed from DenseMap and SmallPtrSet. erase() also invalidates pointers.
• DenseMap has also retired its empty-key sentinel, leading to significant performance improvements. DenseMap with integer keys (int/unsigned/size_t) had -1/-2 reserved — a footgun, now fixed.
• StringMap got Algorithm R deletion too. Its entries are separately heap-allocated, so erase keeps entry pointers valid but invalidates iterators; erase-while-iterating moved to remove_if.

Read More

With a sufficient number of users of an API, it does not matter what you promise in the contract: all observable behaviors of your system will be depended on by somebody. — Hyrum's Law

In a compiler, the most common form of Hyrum's Law is dependence on unspecified behavior — hash bucket order, the order of equal elements after std::sort, padding offsets. The same framing covers a few cases that are technically undefined behavior (use of an invalidated iterator) or plain incidental properties (ABI struct layout, ELF section offsets).

When the compiler itself harbors such a dependency, the symptom is usually output that varies build-to-build: an unstable sort that lands differently after the standard library changes, a hash map whose iteration order shifts when the hash function does. Occasionally the variation is run-to-run within a single build — DenseMap<void *, X> keys with an ASLR-derived seed reorder buckets each invocation. Either way, reproducible builds, bisection, and bug reports all assume same input → same output, and a stealth Hyrum dependency breaks that.

This post surveys some mechanisms that perturb the contract's blind spots so dependencies cannot quietly form.

Read More

Updated in 2026-05.

Since the LLVM 22 branch was cut, I've landed patches that parallelize more link phases and cut task-runtime overhead. This post compares current main against lld 22.1, mold, and wild.

Headline: a Release+Asserts clang --gc-sections link is 1.34x as fast as lld 22.1; Chromium debug with --gdb-index is 1.09x as fast. mold and wild are still ahead — the last section explains why.

Read More

The C and C++ standards leave nearly every detail to the implementation. C23 §6.7.3.2:

An implementation may allocate any addressable storage unit large enough to hold a bit-field. If enough space remains, a bit-field that immediately follows another bit-field in a structure shall be packed into adjacent bits of the same unit. If insufficient space remains, whether a bit-field that does not fit is put into the next unit or overlaps adjacent units is implementation-defined. The order of allocation of bit-fields within a unit (high-order to low-order or low-order to high-order) is implementation-defined. The alignment of the addressable storage unit is unspecified

C++ is also terse — [class.bit]p1:

Allocation of bit-fields within a class object is implementation-defined. Alignment of bit-fields is implementation-defined. Bit-fields are packed into some addressable allocation unit.

Read More

Most architectures encode direct branch/call instructions with a PC-relative displacement. This post discusses a specific category of branch relocations: those used for direct function calls and tail calls. Some architectures use two ELF relocation types for a call instruction:

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15

# i386, x86-64
call foo              # R_386_PC32, R_X86_64_PC32
call foo@plt          # R_386_PLT32, R_X86_64_PLT32

# m68k
bsr.l foo             # R_68K_PC32
bsr.l foo@plt         # R_68K_PLT32

# s390/s390x
brasl %r14, foo       # R_390_PC32DBL
brasl %r14, foo@plt   # R_390_PLT32DBL

# sparc
call    foo, 0        # not PIC: R_SPARC_WDISP30
call    foo, 0        # gas -KPIC: R_SPARC_WPLT30

Read More

For those unfamiliar, lld is the LLVM linker, supporting PE/COFF, ELF, Mach-O, and WebAssembly ports. These object file formats differ significantly, and each port must follow the conventions of the platform's system linker. As a result, the ports share limited code (diagnostics, memory allocation, etc) and have largely separate reviewer groups.

With LLVM 22.1 releasing soon, I've added some notes to the https://github.com/llvm/llvm-project/blob/release/22.x/lld/docs/ReleaseNotes.rst as an lld/ELF maintainer. As usual, I've reviewed almost all the patches not authored by me.

For the first time, I used an LLM agent (Claude Code) to help look through commits (git log release/21.x..release/22.x -- lld/ELF) and draft the release notes. Despite my request to only read lld/ELF changes, Claude Code also crafted notes for other ports, which I retained since their release notes had been quite sparse for several releases. Changes back ported to the 21.x release are removed (git log --oneline llvmorg-22-init..llvmorg-21.1.8 -- lld).

I'll delve into some of the key changes.

Read More