[AArch64][BOLT] Ensure tentative code layout for cold BBs runs.

When split functions is used, BOLT may skip tentative code layout
estimation in some cases, like:
- when there is no profile data for some blocks (ie cold blocks)
- when there are cold functions in lite mode
- when skip functions is used

However, when rewriting the binary we still need to compute PC-relative
distances between hot and cold basic blocks. Without cold layout
estimation, BOLT uses '0x0' as the address of the first cold block,
leading to incorrect estimations of any PC-relative addresses.

This affects large binaries as the relaxStub method expands more
branches than necessary using the short-jump sequence, at it wrongly
believes it has exceeded the branch distance boundary.

This increases code size with a larger and slower sequence; however,
performance regression is expected to be minimal since this only affects
called cold code.

Example of such an unnecessary relaxation:
from:
```armasm
b       .Ltmp1234
```

to:
```armasm
adrp    x16, .Ltmp1234
add     x16, x16, :lo12:.Ltmp1234
br      x16
```

Author: paschalis-mpeis

bolt/lib/Passes/LongJmp.cpp

@@ -11,6 +11,7 @@
 //===----------------------------------------------------------------------===//
 
 #include "bolt/Passes/LongJmp.h"
+#include "bolt/Utils/CommandLineOpts.h"
 
 #define DEBUG_TYPE "longjmp"
 
@@ -324,7 +325,6 @@ uint64_t LongJmpPass::tentativeLayoutRelocColdPart(
 uint64_t LongJmpPass::tentativeLayoutRelocMode(
     const BinaryContext &BC, std::vector<BinaryFunction *> &SortedFunctions,
     uint64_t DotAddress) {
-
   // Compute hot cold frontier
   uint32_t LastHotIndex = -1u;
   uint32_t CurrentIndex = 0;
@@ -354,9 +354,12 @@ uint64_t LongJmpPass::tentativeLayoutRelocMode(
   for (BinaryFunction *Func : SortedFunctions) {
     if (!BC.shouldEmit(*Func)) {
       HotAddresses[Func] = Func->getAddress();
-      continue;
+      // Don't perform any tentative address estimation of a function's cold
+      // layout if it won't be emitted, unless we are ignoring a large number of
+      // functions (ie, on lite mode) and we haven't done such estimation yet.
+      if (opts::processAllFunctions() || ColdLayoutDone)
+        continue;
     }
-
     if (!ColdLayoutDone && CurrentIndex >= LastHotIndex) {
       DotAddress =
           tentativeLayoutRelocColdPart(BC, SortedFunctions, DotAddress);

bolt/test/AArch64/split-funcs-lite.test

@@ -8,7 +8,7 @@ REQUIRES: system-linux
 
 RUN: %clang %cflags %p/../Inputs/asm_main.c -Wl,-q -o %t
 
-RUN: not --crash llvm-bolt %t -o %t.bolt -lite=1 -split-functions -split-all-cold \
-RUN: --skip-funcs="main,foo" 2>&1 | FileCheck %s
+RUN: llvm-bolt %t -o %t.bolt -lite=1 -split-functions -split-all-cold \
+RUN: --skip-funcs="_init,_start,call_weak_fn/1,deregister_tm_clones/1,register_tm_clones/1,__do_global_dtors_aux/1,frame_dummy/1,main,foo,_fini" 2>&1 | FileCheck %s
 
-CHECK: Did not perform tentative code layout for cold blocks.
+CHECK-NOT: Did not perform tentative code layout for cold blocks.