diff --git a/bolt/test/X86/dynrelocs.s b/bolt/test/X86/dynrelocs.s
@@ -1,26 +1,26 @@
-# This reproduces a bug when rewriting dynamic relocations in X86 as
-# BOLT incorrectly attributes R_X86_64_64 dynamic relocations
-# to the wrong section when the -jump-tables=move flag is used. We
-#	expect the relocations to belong to the .bolt.org.rodata section but
-#	it is attributed to a new .rodata section that only contains jump
-#	table entries, created by BOLT. BOLT will only create this new .rodata
-# section if both -jump-tables=move is used and a hot function with
-# jt is present in the input binary, triggering a scenario where the
-# dynamic relocs rewriting gets confused on where to put .rodata relocs.
+## This reproduces a bug when rewriting dynamic relocations in X86 as
+## BOLT incorrectly attributes R_X86_64_64 dynamic relocations
+## to the wrong section when the -jump-tables=move flag is used. We
+##	expect the relocations to belong to the .bolt.org.rodata section but
+##	it is attributed to a new .rodata section that only contains jump
+##	table entries, created by BOLT. BOLT will only create this new .rodata
+## section if both -jump-tables=move is used and a hot function with
+## jt is present in the input binary, triggering a scenario where the
+## dynamic relocs rewriting gets confused on where to put .rodata relocs.
 
-# It is uncommon to end up with dynamic relocations against .rodata,
-# but it can happen. In these cases we cannot corrupt the
-# output binary by writing out dynamic relocs incorrectly. The linker
-# avoids emitting relocs against read-only sections but we override
-# this behavior with the -z notext flag. During runtime, these pages
-# are mapped with write permission and then changed to read-only after
-# the dynamic linker finishes processing the dynamic relocs.
+## It is uncommon to end up with dynamic relocations against .rodata,
+## but it can happen. In these cases we cannot corrupt the
+## output binary by writing out dynamic relocs incorrectly. The linker
+## avoids emitting relocs against read-only sections but we override
+## this behavior with the -z notext flag. During runtime, these pages
+## are mapped with write permission and then changed to read-only after
+## the dynamic linker finishes processing the dynamic relocs.
 
-# In this test, we create a reference to a dynamic object that will
-# imply in R_X86_64_64 being used for .rodata. Now BOLT, when creating
-# a new .rodata to hold jump table entries, needs to remember to emit
-# these dynamic relocs against the original .rodata, and not the new
-# one it just created.
+## In this test, we create a reference to a dynamic object that will
+## imply in R_X86_64_64 being used for .rodata. Now BOLT, when creating
+## a new .rodata to hold jump table entries, needs to remember to emit
+## these dynamic relocs against the original .rodata, and not the new
+## one it just created.
 
 # REQUIRES: system-linux
 
@@ -36,8 +36,8 @@
 # RUN:   -jump-tables=move
 # RUN: llvm-readobj -rs %t.out | FileCheck --check-prefix=READOBJ %s
 
-# Verify that BOLT outputs the dynamic reloc at the correct address,
-# which is the start of the .bolt.org.rodata section.
+## Verify that BOLT outputs the dynamic reloc at the correct address,
+## which is the start of the .bolt.org.rodata section.
 # READOBJ:        Relocations [
 # READOBJ:          Section ([[#]]) .rela.dyn {
 # READOBJ-NEXT:        0x[[#%X,ADDR:]] R_X86_64_64 bar 0x10

diff --git a/bolt/test/X86/exceptions-args.test b/bolt/test/X86/exceptions-args.test
@@ -1,5 +1,5 @@
-# Check that we handle GNU_args_size correctly.
-# It is generated for throwing functions with LP that have parameters on stack.
+## Check that we handle GNU_args_size correctly.
+## It is generated for throwing functions with LP that have parameters on stack.
 
 RUN: %clang %cflags %p/../Inputs/stub.c -fPIC -pie -shared -o %t.so
 RUN: %clangxx %cxxflags -no-pie %p/Inputs/exc_args.s -o %t %t.so -Wl,-z,notext

diff --git a/bolt/test/X86/fallthrough-to-noop.test b/bolt/test/X86/fallthrough-to-noop.test
@@ -1,5 +1,5 @@
-# Check that profile data for the fall-through jump is not ignored when there is
-# a conditional jump followed by a no-op.
+## Check that profile data for the fall-through jump is not ignored when there is
+## a conditional jump followed by a no-op.
 
 RUN: llvm-mc -filetype=obj -triple x86_64-unknown-unknown \
 RUN:   %S/Inputs/ft_to_noop.s -o %t.o
@@ -13,11 +13,11 @@ CHECK:      Binary Function "foo" after building cfg
 CHECK:        Exec Count  : 20
 CHECK:        Profile Acc : 100.0%
 
-# This block is terminated with a conditional jump to .Ltmp0 followed by a
-# no-op. The profile data contains a count for the fall-through (3) which
-# is different from what would be inferred (2). However the destination
-# offset of this fall-through jump in the profile data points to the no-op
-# following the jump and not the start of the fall-through block .LFT0.
+## This block is terminated with a conditional jump to .Ltmp0 followed by a
+## no-op. The profile data contains a count for the fall-through (3) which
+## is different from what would be inferred (2). However the destination
+## offset of this fall-through jump in the profile data points to the no-op
+## following the jump and not the start of the fall-through block .LFT0.
 CHECK:      Entry Point
 CHECK-NEXT:   Exec Count : 20
 CHECK:        Successors: .Ltmp[[#BB1:]] (mispreds: 0, count: 18), .LFT[[#BB2:]] (mispreds: 0, count: 3)

diff --git a/bolt/test/X86/false-jump-table.s b/bolt/test/X86/false-jump-table.s
@@ -1,5 +1,5 @@
-# Check that jump table detection does not fail on a false
-# reference to a jump table.
+## Check that jump table detection does not fail on a false
+## reference to a jump table.
 
 # REQUIRES: system-linux
 

diff --git a/bolt/test/X86/fatal-error.s b/bolt/test/X86/fatal-error.s
@@ -1,6 +1,6 @@
-# Tests whether llvm-bolt will correctly exit with error code and printing
-# fatal error message in case one occurs. Here we test opening a function
-# reordering file that does not exist.
+## Tests whether llvm-bolt will correctly exit with error code and printing
+## fatal error message in case one occurs. Here we test opening a function
+## reordering file that does not exist.
 
 # RUN: llvm-mc -filetype=obj -triple x86_64-unknown-unknown %s -o %t.o
 # RUN: %clang %cflags %t.o -o %t.exe -Wl,-q

diff --git a/bolt/test/X86/fragment-lite-reverse.s b/bolt/test/X86/fragment-lite-reverse.s
@@ -1,4 +1,4 @@
-# Check that BOLT in lite mode processes fragments as expected.
+## Check that BOLT in lite mode processes fragments as expected.
 
 # RUN: llvm-mc -filetype=obj -triple x86_64-unknown-unknown %s -o %t.o
 # RUN: link_fdata %s %t.o %t.fdata

diff --git a/bolt/test/X86/fragment-lite.s b/bolt/test/X86/fragment-lite.s
@@ -1,4 +1,4 @@
-# Check that BOLT in lite mode processes fragments as expected.
+## Check that BOLT in lite mode processes fragments as expected.
 
 # RUN: split-file %s %t
 # RUN: llvm-mc -filetype=obj -triple x86_64-unknown-unknown %t/main.s -o %t.o

diff --git a/bolt/test/X86/fragmented-symbols.s b/bolt/test/X86/fragmented-symbols.s
@@ -1,5 +1,5 @@
-# Checks that symbols are allocated in correct sections, and that empty
-# fragments are not allocated at all.
+## Checks that symbols are allocated in correct sections, and that empty
+## fragments are not allocated at all.
 
 # REQUIRES: x86_64-linux
 

diff --git a/bolt/test/X86/frame-opt-lea.s b/bolt/test/X86/frame-opt-lea.s
@@ -1,6 +1,6 @@
-# This checks that frame optimizer does not try to optimize away caller-saved
-# regs when we do not have complete aliasing info (when there is an LEA
-# instruction and the function does arithmetic with stack addresses).
+## This checks that frame optimizer does not try to optimize away caller-saved
+## regs when we do not have complete aliasing info (when there is an LEA
+## instruction and the function does arithmetic with stack addresses).
 
 
 # REQUIRES: system-linux

diff --git a/bolt/test/X86/function-order-lite.s b/bolt/test/X86/function-order-lite.s
@@ -1,5 +1,5 @@
-# Check that functions listed in -function-order list take precedence over
-# lite mode function filtering.
+## Check that functions listed in -function-order list take precedence over
+## lite mode function filtering.
 
 # RUN: llvm-mc -filetype=obj -triple x86_64-unknown-unknown %s -o %t.o
 # RUN: link_fdata %s %t.o %t.fdata

diff --git a/bolt/test/X86/gdbindex.test b/bolt/test/X86/gdbindex.test
@@ -4,15 +4,15 @@ RUN: ld.lld --gdb-index %t.o %t2.o -o %tfile.exe
 RUN: llvm-bolt %tfile.exe -o %tfile.exe.bolt --update-debug-sections
 RUN: llvm-dwarfdump -gdb-index %tfile.exe.bolt | FileCheck %s
 
-; test.cpp:
-; int main() { return 0; }
-; test2.cpp:
-; int main2() { return 0; }
-; Compiled with:
-; gcc -gsplit-dwarf -c test.cpp test2.cpp
-; gold --gdb-index test.o test2.o -o dwarfdump-gdbindex-v7.elf-x86-64
-; gcc version 5.3.1 20160413, GNU gold (GNU Binutils for Ubuntu 2.26) 1.11
-; Info about gdb-index: https://sourceware.org/gdb/onlinedocs/gdb/Index-Section-Format.html
+;; test.cpp:
+;; int main() { return 0; }
+;; test2.cpp:
+;; int main2() { return 0; }
+;; Compiled with:
+;; gcc -gsplit-dwarf -c test.cpp test2.cpp
+;; gold --gdb-index test.o test2.o -o dwarfdump-gdbindex-v7.elf-x86-64
+;; gcc version 5.3.1 20160413, GNU gold (GNU Binutils for Ubuntu 2.26) 1.11
+;; Info about gdb-index: https://sourceware.org/gdb/onlinedocs/gdb/Index-Section-Format.html
 
 ; CHECK-LABEL: .gdb_index contents:
 ; CHECK: Version = 7

diff --git a/bolt/test/X86/high_pc_udata.s b/bolt/test/X86/high_pc_udata.s
@@ -15,8 +15,8 @@
 # POSTCHECK-NEXT: DW_AT_name [DW_FORM_strp]
 # POSTCHECK-SAME: "main.cpp"
 
-# Testing that BOLT transforms DW_AT_high_pc of form DW_FORM_udata correctly into DW_AT_ranges.
-# Manually changed so that DW_AT_high_pc is DW_FORM_udata, and that DW_AT_name is after it.
+## Testing that BOLT transforms DW_AT_high_pc of form DW_FORM_udata correctly into DW_AT_ranges.
+## Manually changed so that DW_AT_high_pc is DW_FORM_udata, and that DW_AT_name is after it.
 # int main() {
 #    return 0;
 # }

diff --git a/bolt/test/X86/icp-inline.s b/bolt/test/X86/icp-inline.s
@@ -1,7 +1,7 @@
-# This test verifies the effect of -icp-inline option: that ICP is only
-# performed for call targets eligible for inlining.
+## This test verifies the effect of -icp-inline option: that ICP is only
+## performed for call targets eligible for inlining.
 
-# The assembly was produced from C code compiled with clang-15 -O1 -S:
+## The assembly was produced from C code compiled with clang-15 -O1 -S:
 
 # int foo(int x) { return x + 1; }
 # int bar(int x) { return x*100 + 42; }

diff --git a/bolt/test/X86/ignored-interprocedural-reference.s b/bolt/test/X86/ignored-interprocedural-reference.s
@@ -1,5 +1,5 @@
-# This reproduces a bug with not processing interprocedural references from
-# ignored functions.
+## This reproduces a bug with not processing interprocedural references from
+## ignored functions.
 
 # REQUIRES: system-linux
 
@@ -16,7 +16,7 @@
 # CHECK-YAML: calls: {{.*}} disc: 1
 
 # PREAGG: B #main# #foo_secondary# 1 1
-# main calls foo at valid instruction offset past nops that are to be stripped.
+## main calls foo at valid instruction offset past nops that are to be stripped.
   .globl main
 main:
   .cfi_startproc
@@ -25,16 +25,16 @@ main:
   .cfi_endproc
 .size main,.-main
 
-# Placeholder cold fragment to force main to be ignored in non-relocation mode.
+## Placeholder cold fragment to force main to be ignored in non-relocation mode.
   .globl main.cold
 main.cold:
   .cfi_startproc
   ud2
   .cfi_endproc
 .size main.cold,.-main.cold
 
-# foo is set up to contain a valid instruction at called offset, and trapping
-# instructions past that.
+## foo is set up to contain a valid instruction at called offset, and trapping
+## instructions past that.
   .globl foo
 foo:
   .cfi_startproc

diff --git a/bolt/test/X86/indirect-goto-pie.test b/bolt/test/X86/indirect-goto-pie.test
@@ -1,14 +1,14 @@
-# Check that llvm-bolt fails to process PIC binaries with computed goto, as the
-# support is not there yet for correctly updating dynamic relocations
-# referencing code inside functions.
+## Check that llvm-bolt fails to process PIC binaries with computed goto, as the
+## support is not there yet for correctly updating dynamic relocations
+## referencing code inside functions.
 
 REQUIRES: x86_64-linux
 
 RUN: %clang %S/Inputs/indirect_goto.c -o %t -fpic -pie -Wl,-q
 RUN: not llvm-bolt %t -o %t.bolt --relocs=1 --print-cfg --print-only=main \
 RUN:   |& FileCheck %s
 
-# Check that processing works if main() is skipped.
+## Check that processing works if main() is skipped.
 RUN: llvm-bolt %t -o %t.bolt --relocs=1 --skip-funcs=main
 
 CHECK:  jmpq    *%rax # UNKNOWN CONTROL FLOW

diff --git a/bolt/test/X86/indirect-goto.test b/bolt/test/X86/indirect-goto.test
@@ -1,9 +1,9 @@
-# Check llvm-bolt processes binaries compiled from sources that use indirect goto.
+## Check llvm-bolt processes binaries compiled from sources that use indirect goto.
 RUN: %clang %cflags -no-pie %S/Inputs/indirect_goto.c -Wl,-q -o %t
 RUN: llvm-bolt %t -o %t.null --relocs=1 --print-cfg --print-only=main \
 RUN:   --strict \
 RUN:   2>&1 | FileCheck %s
 
-# Check that all possible destinations are included as successors.
+## Check that all possible destinations are included as successors.
 CHECK:  jmpq    *%rax # UNKNOWN CONTROL FLOW
 CHECK:  Successors: .Ltmp0, .Ltmp1, .Ltmp2
diff --git a/bolt/test/X86/inlined-function-mixed.test b/bolt/test/X86/inlined-function-mixed.test
@@ -1,5 +1,5 @@
-# Make sure inlining from a unit with debug info into unit without
-# debug info does not cause a crash.
+## Make sure inlining from a unit with debug info into unit without
+## debug info does not cause a crash.
 
 RUN: %clangxx %cxxflags %S/Inputs/inlined.cpp -c -o %T/inlined.o
 RUN: %clangxx %cxxflags %S/Inputs/inlinee.cpp -c -o %T/inlinee.o -g

diff --git a/bolt/test/X86/insert-addr-rnglists_base.s b/bolt/test/X86/insert-addr-rnglists_base.s
@@ -6,8 +6,8 @@
 # RUN: llvm-dwarfdump --show-form --verbose --debug-info %t.exe | FileCheck --check-prefix=PRECHECK %s
 # RUN: llvm-dwarfdump --show-form --verbose --debug-info %t.bolt | FileCheck --check-prefix=POSTCHECK %s
 
-# This test checks we correctly insert DW_AT_addr_base, when converting DW_AT_low_pc into DW_AT_ranges.
-# PRECHECK-NOT: DW_AT_addr_base
+## This test checks we correctly insert DW_AT_addr_base, when converting DW_AT_low_pc into DW_AT_ranges.
+## PRECHECK-NOT: DW_AT_addr_base
 
 # POSTCHECK: DW_AT_ranges [DW_FORM_rnglistx]
 # POSTCHECK: DW_AT_rnglists_base [DW_FORM_sec_offset]	(0x0000000c)

diff --git a/bolt/test/X86/insert-debug-info-entry.test b/bolt/test/X86/insert-debug-info-entry.test
@@ -7,7 +7,7 @@
 ; RUN: llvm-dwarfdump --debug-info foo.exe | FileCheck -check-prefix=PRE-BOLT %s
 ; RUN: llvm-dwarfdump --debug-info foo.exe.bolt | FileCheck %s
 
-; This tests checks that DW_AT_GNU_ranges_base is added at the end of the CU.
+;; This tests checks that DW_AT_GNU_ranges_base is added at the end of the CU.
 
 ; PRE-BOLT: DW_AT_GNU_addr_base
 ; PRE-BOLT-NOT: DW_AT_GNU_ranges_base

diff --git a/bolt/test/X86/internal-call-instrument-so.s b/bolt/test/X86/internal-call-instrument-so.s
@@ -1,4 +1,4 @@
-# This reproduces a bug with instrumentation crashes on internal call
+## This reproduces a bug with instrumentation crashes on internal call
 
 # REQUIRES: system-linux,bolt-runtime,target=x86_64{{.*}}
 

diff --git a/bolt/test/X86/internal-call-instrument.s b/bolt/test/X86/internal-call-instrument.s
@@ -1,4 +1,4 @@
-# This reproduces a bug with instrumentation crashes on internal call
+## This reproduces a bug with instrumentation crashes on internal call
 
 # REQUIRES: x86_64-linux,bolt-runtime,target=x86_64{{.*}}
 

diff --git a/bolt/test/X86/interprocedural-ref-entry-point.s b/bolt/test/X86/interprocedural-ref-entry-point.s
@@ -1,7 +1,7 @@
-# This reproduces a bug where not registering cold fragment entry points 
-# leads to removing blocks and an inconsistent CFG after UCE.
-# Test assembly was obtained using C-Reduce from this C++ code:
-# (compiled with `g++ -O2 -Wl,-q`)
+## This reproduces a bug where not registering cold fragment entry points
+## leads to removing blocks and an inconsistent CFG after UCE.
+## Test assembly was obtained using C-Reduce from this C++ code:
+## (compiled with `g++ -O2 -Wl,-q`)
 #
 # #include <stdexcept>
 # int a;

diff --git a/bolt/test/X86/is-strip.s b/bolt/test/X86/is-strip.s
@@ -1,4 +1,4 @@
-# This test checks whether a binary is stripped or not.
+## This test checks whether a binary is stripped or not.
 
 # RUN: %clang++ %cflags %p/Inputs/linenumber.cpp -o %t -Wl,-q
 # RUN: llvm-bolt %t -o %t.out 2>&1 | FileCheck %s -check-prefix=CHECK-NOSTRIP

diff --git a/bolt/test/X86/issue20.s b/bolt/test/X86/issue20.s
@@ -1,6 +1,6 @@
-# This reproduces issue 20 from our github repo
-#  "BOLT crashes when removing unreachable BBs that are a target
-#   in a JT"
+## This reproduces issue 20 from our github repo
+## "BOLT crashes when removing unreachable BBs that are a target
+##   in a JT"
 
 # REQUIRES: system-linux
 

diff --git a/bolt/test/X86/issue20.test b/bolt/test/X86/issue20.test
@@ -1,6 +1,6 @@
-# This reproduces issue 20 from our github repo
-#  "BOLT crashes when removing unreachable BBs that are a target
-#   in a JT"
+## This reproduces issue 20 from our github repo
+##  "BOLT crashes when removing unreachable BBs that are a target
+##   in a JT"
 
 # RUN: yaml2obj %p/Inputs/issue20.yaml &> %t.exe
 # RUN: llvm-bolt %t.exe --relocs=0 --jump-tables=move --print-finalized \

diff --git a/bolt/test/X86/issue26.s b/bolt/test/X86/issue26.s
@@ -1,6 +1,6 @@
-# This reproduces issue 26 from our github repo
-#  BOLT fails with the following assertion:
-#    llvm/tools/llvm-bolt/src/BinaryFunction.cpp:2950: void llvm::bolt::BinaryFunction::postProcessBranches(): Assertion `validateCFG() && "invalid CFG"' failed.
+## This reproduces issue 26 from our github repo
+##  BOLT fails with the following assertion:
+##    llvm/tools/llvm-bolt/src/BinaryFunction.cpp:2950: void llvm::bolt::BinaryFunction::postProcessBranches(): Assertion `validateCFG() && "invalid CFG"' failed.
 
 # REQUIRES: system-linux
 

diff --git a/bolt/test/X86/issue26.test b/bolt/test/X86/issue26.test
@@ -1,4 +1,4 @@
-# This reproduces issue 26 from our github repo
+## This reproduces issue 26 from our github repo
 
 # RUN: yaml2obj %p/Inputs/issue26.yaml &> %t.exe
 # RUN: llvm-bolt %t.exe --relocs --print-cfg -o %t.out 2>&1 \

diff --git a/bolt/test/X86/jmp-optimization.test b/bolt/test/X86/jmp-optimization.test
@@ -1,7 +1,7 @@
-# Tests the optimization of functions that just do a tail call in the beginning.
+## Tests the optimization of functions that just do a tail call in the beginning.
 
-# This test has commands that rely on shell capabilities that won't execute
-# correctly on Windows e.g. unsupported parameter expansion
+## This test has commands that rely on shell capabilities that won't execute
+## correctly on Windows e.g. unsupported parameter expansion
 REQUIRES: shell
 
 RUN: %clang %cflags -O2 %S/Inputs/jmp_opt{,2,3}.cpp -o %t

diff --git a/bolt/test/X86/jmpjmp.test b/bolt/test/X86/jmpjmp.test
@@ -1,5 +1,5 @@
-# Verifies that llvm-bolt allocates two consecutive jumps in two separate basic
-# blocks.
+## Verifies that llvm-bolt allocates two consecutive jumps in two separate basic
+## blocks.
 
 RUN: llvm-mc -filetype=obj -triple x86_64-unknown-unknown %S/Inputs/jmpjmp.s -o %t.o
 RUN: %clang %cflags %t.o -o %t.exe

diff --git a/bolt/test/X86/jt-symbol-disambiguation-3.s b/bolt/test/X86/jt-symbol-disambiguation-3.s
@@ -1,11 +1,11 @@
-# In this test case, we reproduce the behavior seen in gcc where the
-# base address of a jump table is decremented by some number and ends up
-# at the exact addess of a jump table from another function. After
-# linking, the instruction references another jump table and that
-# confuses BOLT.
-# We repro here the following issue:
-# Before assembler: Instruction operand is: jumptable - 32
-# After linking:    Instruction operand is: another_jumptable
+## In this test case, we reproduce the behavior seen in gcc where the
+## base address of a jump table is decremented by some number and ends up
+## at the exact addess of a jump table from another function. After
+## linking, the instruction references another jump table and that
+## confuses BOLT.
+## We repro here the following issue:
+## Before assembler: Instruction operand is: jumptable - 32
+## After linking:    Instruction operand is: another_jumptable
 
 # REQUIRES: system-linux, asserts
 
@@ -18,8 +18,8 @@
 # RUN: llvm-bolt %t.exe -o %t.exe.bolt --relocs=1 --lite=0 \
 # RUN:   --reorder-blocks=reverse
 
-# Useful when manually testing this. Currently we just check that
-# the test does not cause BOLT to assert.
+## Useful when manually testing this. Currently we just check that
+## the test does not cause BOLT to assert.
 # COM: %t.exe.bolt 1 2
 
   .file "jt-symbol-disambiguation-3.s"

diff --git a/bolt/test/X86/jump-table-fixed-ref-pic.test b/bolt/test/X86/jump-table-fixed-ref-pic.test
@@ -1,5 +1,5 @@
-# Verify that BOLT detects fixed destination of indirect jump for PIC
-# case.
+## Verify that BOLT detects fixed destination of indirect jump for PIC
+## case.
 
 XFAIL: *
 

diff --git a/bolt/test/X86/jump-table-footprint-reduction.test b/bolt/test/X86/jump-table-footprint-reduction.test
@@ -1,5 +1,5 @@
-# Checks that jump table footprint reduction optimization is reducing entry
-# sizes.
+## Checks that jump table footprint reduction optimization is reducing entry
+## sizes.
 
 RUN: llvm-mc -filetype=obj -triple x86_64-unknown-unknown \
 RUN:   %S/Inputs/jump_table_footprint_reduction.s -o %t.o

diff --git a/bolt/test/X86/jump-table-icp.test b/bolt/test/X86/jump-table-icp.test
@@ -4,8 +4,8 @@ RUN: link_fdata %p/Inputs/jump_table_icp.s %t.o %t.fdata --nmtool llvm-nm
 RUN: llvm-strip --strip-unneeded %t.o
 RUN: %clang %cflags -no-pie %t.o -o %t.exe -Wl,-q
 
-# This test has commands that rely on shell capabilities that won't execute
-# correctly on Windows e.g. subshell execution
+## This test has commands that rely on shell capabilities that won't execute
+## correctly on Windows e.g. subshell execution
 REQUIRES: shell
 
 RUN: (llvm-bolt %t.exe --data %t.fdata -o %t --relocs \

diff --git a/bolt/test/X86/jump-table-pic-conflict.s b/bolt/test/X86/jump-table-pic-conflict.s
@@ -1,16 +1,16 @@
-# Check cases when the first PIC jump table entries of one function can be
-# interpreted as valid last entries of the previous function.
+## Check cases when the first PIC jump table entries of one function can be
+## interpreted as valid last entries of the previous function.
 
-# Conditions to trigger the bug:  Function A and B have jump tables that
-# are adjacent in memory. We run in lite relocation mode. Function B
-# is not disassembled because it does not have profile. Function A
-# triggers a special conditional that forced BOLT to rewrite its jump
-# table in-place (instead of moving it) because it is marked as
-# non-simple (in this case, containing unknown control flow). The
-# first entry of B's jump table (a PIC offset) happens to be a valid
-# address inside A when added to A's jump table base address. In this
-# case, BOLT could overwrite B's jump table, corrupting it, thinking
-# the first entry of it is actually part of A's jump table.
+## Conditions to trigger the bug:  Function A and B have jump tables that
+## are adjacent in memory. We run in lite relocation mode. Function B
+## is not disassembled because it does not have profile. Function A
+## triggers a special conditional that forced BOLT to rewrite its jump
+## table in-place (instead of moving it) because it is marked as
+## non-simple (in this case, containing unknown control flow). The
+## first entry of B's jump table (a PIC offset) happens to be a valid
+## address inside A when added to A's jump table base address. In this
+## case, BOLT could overwrite B's jump table, corrupting it, thinking
+## the first entry of it is actually part of A's jump table.
 
 # REQUIRES: system-linux
 
@@ -26,8 +26,8 @@
 # readelf. This is another way to check this bug:
 # COM: %t.out
 
-# BOLT needs to create a new rodata section, indicating that it
-# successfully moved the jump table in _start.
+## BOLT needs to create a new rodata section, indicating that it
+## successfully moved the jump table in _start.
 # CHECK: [{{.*}}] .bolt.org.rodata
 
   .globl _start
@@ -41,8 +41,8 @@ _start:
   cmpq    $3, %rdi
   ja      .L5
   jmp     .L6
-# Unreachable code, here to mark this function as non-simple
-# (containing unknown control flow) with a stray indirect jmp
+## Unreachable code, here to mark this function as non-simple
+## (containing unknown control flow) with a stray indirect jmp
   jmp     *%rax
 .L6:
   decq    %rdi
@@ -115,8 +115,8 @@ str1: .asciz "Message 1\n"
 str2: .asciz "Message 2\n"
 str3: .asciz "Message 3\n"
 str4: .asciz "Highrange\n"
-# Special case where the first .LJT2 entry is a valid offset of
-# _start when interpreted with .LJT1 as a base address.
+## Special case where the first .LJT2 entry is a valid offset of
+## _start when interpreted with .LJT1 as a base address.
 .LJT1:
   .long .L1-.LJT1
   .long .L2-.LJT1

diff --git a/bolt/test/X86/jump-table-pic-order.test b/bolt/test/X86/jump-table-pic-order.test
@@ -1,12 +1,12 @@
-# Check that successors of a basic block with jump table are generated
-# in the same order as they appear in the input code.
+## Check that successors of a basic block with jump table are generated
+## in the same order as they appear in the input code.
 
 RUN: %clang %cflags %S/Inputs/jump-table-pic.s -o %t.exe -Wl,-q
 RUN: llvm-bolt %t.exe --strict --print-cfg --print-only=main -o %t.null \
 RUN:   | FileCheck %s
 
 CHECK:      BB Layout   : {{.*, .*, .*,}} [[BB4to6:.*, .*, .*]]
 
-# Check that successors appear in the order matching the input layout.
+## Check that successors appear in the order matching the input layout.
 CHECK:      jmpq    *%rax # JUMPTABLE
 CHECK-NEXT: Successors: [[BB4to6]]
diff --git a/bolt/test/X86/jump-table-reference.test b/bolt/test/X86/jump-table-reference.test
@@ -1,4 +1,4 @@
-# Verifies that BOLT detects fixed destination of indirect jump
+## Verifies that BOLT detects fixed destination of indirect jump
 
 RUN: %clang %cflags -no-pie %S/Inputs/jump_table_reference.s -Wl,-q -o %t
 RUN: llvm-bolt %t --relocs -o %t.null 2>&1 | FileCheck %s

diff --git a/bolt/test/X86/layout-heuristic.test b/bolt/test/X86/layout-heuristic.test
@@ -1,8 +1,8 @@
-# Checks that llvm-bolt is able to read data generated by perf2bolt,  update the
-# CFG edges accordingly with absolute number of branches and mispredictions,
-# infer fallthrough branch info and reorder basic blocks using a greedy
-# heuristic,  or find the optimal solution if the function is small enough.
-# Also checks that llvm-bolt disassembler and CFG builder is working properly.
+## Checks that llvm-bolt is able to read data generated by perf2bolt,  update the
+## CFG edges accordingly with absolute number of branches and mispredictions,
+## infer fallthrough branch info and reorder basic blocks using a greedy
+## heuristic,  or find the optimal solution if the function is small enough.
+## Also checks that llvm-bolt disassembler and CFG builder is working properly.
 
 RUN: yaml2obj %p/Inputs/blarge.yaml &> %t.exe
 RUN: llvm-bolt %t.exe -o %t.null --data %p/Inputs/blarge.fdata \

diff --git a/bolt/test/X86/line-number.test b/bolt/test/X86/line-number.test
@@ -1,17 +1,17 @@
-# Verifies that the extraction of DWARF line number information is correct.
+## Verifies that the extraction of DWARF line number information is correct.
 
 RUN: %clangxx %cxxflags %S/Inputs/linenumber.cpp -g -o %t
 RUN: llvm-bolt %t -o %t.null --print-reordered --update-debug-sections \
 RUN:   --print-debug-info --reorder-blocks=reverse --sequential-disassembly \
 RUN:   2>&1 | FileCheck %s
 
-# Local variable in f()
+## Local variable in f()
 CHECK: movl    $0xbeef, -0x4(%rbp) # debug line {{.*}}linenumber.cpp:9
-# Checks that a branch instruction that is inserted by BOLT does not have
-# debug line info associated with it.
+## Checks that a branch instruction that is inserted by BOLT does not have
+## debug line info associated with it.
 CHECK-NOT: jmp   .LFT0 # debug line {{.*}}linenumber.cpp:1
-# Call to f() in g()
+## Call to f() in g()
 CHECK: callq   _Z1fv{{.*}} # debug line {{.*}}linenumber.cpp:19
-# Calls to g() and f() in main
+## Calls to g() and f() in main
 CHECK: callq   _Z1gv{{.*}} # debug line {{.*}}linenumber.cpp:23
 CHECK: callq   _Z1fv{{.*}} # debug line {{.*}}linenumber.cpp:23
diff --git a/bolt/test/X86/lit.local.cfg b/bolt/test/X86/lit.local.cfg
@@ -1,7 +1,7 @@
 if not "X86" in config.root.targets:
     config.unsupported = True
 
-flags = "--target=x86_64-pc-linux -nostdlib"
+flags = "--target=x86_64-unknown-linux-gnu -nostdlib"
 
 config.substitutions.insert(0, ("%cflags", f"%cflags {flags}"))
 config.substitutions.insert(0, ("%cxxflags", f"%cxxflags {flags}"))
diff --git a/bolt/test/X86/log.test b/bolt/test/X86/log.test
@@ -1,6 +1,6 @@
-# Tests whether llvm-bolt is able to redirect logs when processing a simple
-# input. If this test fails on your changes, please use BinaryContext::outs()
-# to print BOLT logging instead of llvm::outs().
+## Tests whether llvm-bolt is able to redirect logs when processing a simple
+## input. If this test fails on your changes, please use BinaryContext::outs()
+## to print BOLT logging instead of llvm::outs().
 
 RUN: yaml2obj %p/Inputs/blarge.yaml &> %t.exe
 RUN: llvm-bolt %t.exe -o %t.null --data %p/Inputs/blarge.fdata -v=2 \
@@ -12,7 +12,7 @@ CHECK-NOT: BOLT-INFO
 CHECK-NOT: BOLT-WARNING
 CHECK-NOT: BOLT-ERROR
 
-# Check some usual BOLT output lines are being redirected to the log file
+## Check some usual BOLT output lines are being redirected to the log file
 CHECK-LOG: BOLT-INFO: Target architecture
 CHECK-LOG: BOLT-INFO: BOLT version
 CHECK-LOG: BOLT-INFO: basic block reordering modified layout

diff --git a/bolt/test/X86/loop-inversion-pass.s b/bolt/test/X86/loop-inversion-pass.s
@@ -16,19 +16,19 @@
 # RUN:    --print-finalized --loop-inversion-opt -o %t.out3 \
 # RUN:    | FileCheck --check-prefix="CHECK3" %s
 
-# The case where the loop is used:
+## The case where the loop is used:
 # FDATA: 1 main 2 1 main #.J1# 0 420
 # FDATA: 1 main b 1 main #.Jloop# 0 420
 # FDATA: 1 main b 1 main d 0 1
 # CHECK: BB Layout   : .LBB00, .Ltmp0, .Ltmp1, .LFT0
 
-# The case where the loop is unused:
+## The case where the loop is unused:
 # FDATA2: 1 main 2 1 main #.J1# 0 420
 # FDATA2: 1 main b 1 main #.Jloop# 0 1
 # FDATA2: 1 main b 1 main d 0 420
 # CHECK2: BB Layout   : .LBB00, .Ltmp1, .LFT0, .Ltmp0
 
-# The case where the loop does not require rotation:
+## The case where the loop does not require rotation:
 # FDATA3: 1 main 2 1 main #.J1# 0 420
 # FDATA3: 1 main b 1 main #.Jloop# 0 420
 # FDATA3: 1 main b 1 main d 0 1

diff --git a/bolt/test/X86/loop-nest.test b/bolt/test/X86/loop-nest.test
@@ -1,4 +1,4 @@
-# Verifies that llvm-bolt prints correct loop information.
+## Verifies that llvm-bolt prints correct loop information.
 
 RUN: llvm-mc -filetype=obj -triple x86_64-unknown-unknown \
 RUN:   %p/Inputs/loop_nest.s -o %t.o

diff --git a/bolt/test/X86/merge-fdata-bat-mode.test b/bolt/test/X86/merge-fdata-bat-mode.test
@@ -1,5 +1,5 @@
-# Check merge-fdata tool correctly processes fdata files with header strings
-# such as the ones produced by BAT mode (boltedcollection)
+## Check merge-fdata tool correctly processes fdata files with header strings
+## such as the ones produced by BAT mode (boltedcollection)
 RUN: merge-fdata %S/Inputs/bat_profile_1.fdata \
 RUN:             %S/Inputs/bat_profile_2.fdata \
 RUN:   | FileCheck %s --check-prefix=CHECK-FDATA

diff --git a/bolt/test/X86/merge-fdata-nobat-mode.test b/bolt/test/X86/merge-fdata-nobat-mode.test
@@ -1,4 +1,4 @@
-# Check that merge-fdata tool doesn't spuriously print boltedcollection
+## Check that merge-fdata tool doesn't spuriously print boltedcollection
 
 RUN: merge-fdata %S/Inputs/blarge.fdata %S/Inputs/blarge.fdata \
 RUN:   | FileCheck %s --check-prefix=CHECK-FDATA

diff --git a/bolt/test/X86/merge-fdata-output.test b/bolt/test/X86/merge-fdata-output.test
@@ -1,4 +1,4 @@
-# Check merge-fdata tool correctly handles `-o` option.
+## Check merge-fdata tool correctly handles `-o` option.
 RUN: merge-fdata %S/Inputs/bat_profile_1.fdata \
 RUN:             %S/Inputs/bat_profile_2.fdata \
 RUN:   | FileCheck %s
@@ -13,4 +13,4 @@ RUN:             %S/Inputs/bat_profile_2.fdata \
 RUN:             -o %t
 RUN: FileCheck %s < %t
 
-CHECK: 1 main 451 1 SolveCubic 0 0 302
+CHECK: 1 main 451 1 SolveCubic 0 0 302
diff --git a/bolt/test/X86/no-entry-reordering.test b/bolt/test/X86/no-entry-reordering.test
@@ -1,5 +1,5 @@
-# Verifies that llvm-bolt reordering heuristic does not allocate a BB before the
-# entry point even if there is a hot edge from a block to entry point
+## Verifies that llvm-bolt reordering heuristic does not allocate a BB before the
+## entry point even if there is a hot edge from a block to entry point
 
 RUN: llvm-mc -filetype=obj -triple x86_64-unknown-unknown %S/Inputs/entry.s -o %t.o
 RUN: link_fdata %S/Inputs/entry.s %t.o %t.fdata --nmtool llvm-nm

diff --git a/bolt/test/X86/no-output.test b/bolt/test/X86/no-output.test
@@ -1,4 +1,4 @@
-# This script checks that BOLT is able to work in dry run mode (no output)
+## This script checks that BOLT is able to work in dry run mode (no output)
 
 # REQUIRES: system-linux
 

diff --git a/bolt/test/X86/nolbr.s b/bolt/test/X86/nolbr.s
@@ -1,6 +1,6 @@
-# This reproduces a bug where profile collected from perf without LBRs and
-# converted into fdata-no-lbr format is reported to not contain profile for any
-# functions.
+## This reproduces a bug where profile collected from perf without LBRs and
+## converted into fdata-no-lbr format is reported to not contain profile for any
+## functions.
 
 # REQUIRES: system-linux
 

diff --git a/bolt/test/X86/patch-entries.test b/bolt/test/X86/patch-entries.test
@@ -1,7 +1,7 @@
-# Checking crashes against injected binary functions created by patch
-# entries pass and debug info turned on. In these cases, we were
-# trying to fetch input to output maps on injected functions and
-# crashing.
+## Checking crashes against injected binary functions created by patch
+## entries pass and debug info turned on. In these cases, we were
+## trying to fetch input to output maps on injected functions and
+## crashing.
 
 REQUIRES: system-linux
 
@@ -10,22 +10,22 @@ RUN:   -Wl,-q -I%p/../Inputs
 RUN: llvm-bolt -relocs %t.exe -o %t.out --update-debug-sections --force-patch \
 RUN:   --enable-bat
 
-# Check that patched functions can be disassembled (override FDE from the
-# original function)
+## Check that patched functions can be disassembled (override FDE from the
+## original function)
 # PREAGG: B X:0 #foo.org.0# 1 0
 RUN: link_fdata %s %t.out %t.preagg PREAGG
 RUN: perf2bolt %t.out -p %t.preagg --pa -o %t.yaml --profile-format=yaml \
 RUN:   -print-disasm -print-only=foo.org.0/1 2>&1 | FileCheck %s
 CHECK-NOT: BOLT-WARNING: sizes differ for function foo.org.0/1
 CHECK: Binary Function "foo.org.0/1(*2)" after disassembly {
 
-# Check the expected eh_frame contents
+## Check the expected eh_frame contents
 RUN: llvm-nm --print-size %t.out > %t.foo
 RUN: llvm-objdump %t.out --dwarf=frames >> %t.foo
 RUN: FileCheck %s --input-file %t.foo --check-prefix=CHECK-FOO
 CHECK-FOO: 0000000000[[#%x,FOO:]] [[#%x,OPTSIZE:]] t foo
 CHECK-FOO: 0000000000[[#%x,ORG:]] [[#%x,ORGSIZE:]] t foo.org.0
-# patched FDE comes first
+## patched FDE comes first
 CHECK-FOO: FDE {{.*}} pc=00[[#%x,ORG]]...00[[#%x,ORG+ORGSIZE]]
-# original FDE comes second
+## original FDE comes second
 CHECK-FOO: FDE {{.*}} pc=00[[#%x,ORG]]...00[[#%x,ORG+OPTSIZE]]
diff --git a/bolt/test/X86/pre-aggregated-perf.test b/bolt/test/X86/pre-aggregated-perf.test
@@ -1,12 +1,12 @@
-# This script checks that perf2bolt is reading pre-aggregated perf information
-# correctly for a simple example. The perf.data of this example was generated
-# with the following command:
-#
-#  $ perf record -j any,u -e branch -o perf.data -- ./blarge
-#
-#  blarge is the binary for "basicmath large inputs" taken from Mibench.
+## This script checks that perf2bolt is reading pre-aggregated perf information
+## correctly for a simple example. The perf.data of this example was generated
+## with the following command:
+##
+##  $ perf record -j any,u -e branch -o perf.data -- ./blarge
+##
+##  blarge is the binary for "basicmath large inputs" taken from Mibench.
 
-# Currently failing in MacOS / generating different hash for usqrt
+## Currently failing in MacOS / generating different hash for usqrt
 REQUIRES: system-linux
 
 RUN: yaml2obj %p/Inputs/blarge.yaml &> %t.exe
@@ -22,7 +22,7 @@ CHECK: BOLT-INFO: 4 out of 7 functions in the binary (57.1%) have non-empty exec
 RUN: cat %t | sort | FileCheck %s -check-prefix=PERF2BOLT
 RUN: cat %t.new | FileCheck %s -check-prefix=NEWFORMAT
 
-# Test --profile-format option with perf2bolt
+## Test --profile-format option with perf2bolt
 RUN: perf2bolt %t.exe -o %t.fdata --pa -p %p/Inputs/pre-aggregated.txt \
 RUN:   --profile-format=fdata
 RUN: cat %t.fdata | sort | FileCheck %s -check-prefix=PERF2BOLT
@@ -31,7 +31,7 @@ RUN: perf2bolt %t.exe -o %t.yaml --pa -p %p/Inputs/pre-aggregated.txt \
 RUN:   --profile-format=yaml --profile-use-dfs
 RUN: cat %t.yaml | FileCheck %s -check-prefix=NEWFORMAT
 
-# Test --profile-format option with llvm-bolt --aggregate-only
+## Test --profile-format option with llvm-bolt --aggregate-only
 RUN: llvm-bolt %t.exe -o %t.bolt.fdata --pa -p %p/Inputs/pre-aggregated.txt \
 RUN:   --aggregate-only --profile-format=fdata
 RUN: cat %t.bolt.fdata | sort | FileCheck %s -check-prefix=PERF2BOLT

diff --git a/bolt/test/X86/pt_gnu_relro.s b/bolt/test/X86/pt_gnu_relro.s
@@ -1,7 +1,7 @@
 # REQUIRES: system-linux
 
-# Check that BOLT recognizes PT_GNU_RELRO segment and marks respective sections
-# accordingly.
+## Check that BOLT recognizes PT_GNU_RELRO segment and marks respective sections
+## accordingly.
 
 # RUN: llvm-mc -filetype=obj -triple x86_64-unknown-linux %s -o %t.o
 # RUN: ld.lld %t.o -o %t.exe -q --no-relax

diff --git a/bolt/test/X86/reader-stale-yaml-std.test b/bolt/test/X86/reader-stale-yaml-std.test
@@ -1,19 +1,19 @@
-# This script checks that YamlProfileReader in llvm-bolt is reading data
-# correctly and stale data is corrected by profile inference.
+## This script checks that YamlProfileReader in llvm-bolt is reading data
+## correctly and stale data is corrected by profile inference.
 
 RUN: yaml2obj %p/Inputs/blarge.yaml &> %t.exe
 RUN: llvm-bolt %t.exe -o %t.null -b %p/Inputs/blarge_profile_stale.std-hash.yaml \
 RUN:   --print-cfg --print-only=usqrt,SolveCubic --infer-stale-profile=1 -v=1 \
 RUN:   2>&1 | FileCheck %s
 
-# Verify that yaml reader works as expected.
+## Verify that yaml reader works as expected.
 CHECK: pre-processing profile using YAML profile reader
 CHECK: BOLT-INFO: YAML profile with hash: std::hash
 
-# Function "SolveCubic" has stale profile, since there is one jump in the
-# profile (from bid=13 to bid=2) which is not in the CFG in the binary. The test
-# verifies that the inference is able to match two blocks (bid=1 and bid=13)
-# using "loose" hashes and then correctly propagate the counts.
+## Function "SolveCubic" has stale profile, since there is one jump in the
+## profile (from bid=13 to bid=2) which is not in the CFG in the binary. The test
+## verifies that the inference is able to match two blocks (bid=1 and bid=13)
+## using "loose" hashes and then correctly propagate the counts.
 
 CHECK: Binary Function "SolveCubic" after building cfg {
 CHECK:   State       : CFG constructed
@@ -25,7 +25,7 @@ CHECK:   BB Count    : 18
 CHECK:   Exec Count  : 151
 CHECK:   Branch Count: 552
 CHECK: }
-# Verify block counts.
+## Verify block counts.
 CHECK: .LBB00 (43 instructions, align : 1)
 CHECK:   Successors: .Ltmp[[#BB07:]] (mispreds: 0, count: 0), .LFT[[#BB01:]] (mispreds: 0, count: 151)
 CHECK: .LFT[[#BB01:]] (5 instructions, align : 1)
@@ -37,10 +37,10 @@ CHECK: .Ltmp[[#BB013:]] (12 instructions, align : 1)
 CHECK:   Successors: .Ltmp[[#BB03:]] (mispreds: 0, count: 151)
 CHECK: End of Function "SolveCubic"
 
-# Function "usqrt" has stale profile, since the number of blocks in the profile
-# (nblocks=6) does not match the size of the CFG in the binary. The entry
-# block (bid=0) has an incorrect (missing) count, which should be inferred by
-# the algorithm.
+## Function "usqrt" has stale profile, since the number of blocks in the profile
+## (nblocks=6) does not match the size of the CFG in the binary. The entry
+## block (bid=0) has an incorrect (missing) count, which should be inferred by
+# #the algorithm.
 
 CHECK: Binary Function "usqrt" after building cfg {
 CHECK:   State       : CFG constructed
@@ -52,7 +52,7 @@ CHECK:   BB Count    : 5
 CHECK:   Exec Count  : 20
 CHECK:   Branch Count: 640
 CHECK: }
-# Verify block counts.
+## Verify block counts.
 CHECK: .LBB01 (4 instructions, align : 1)
 CHECK:   Successors: .Ltmp[[#BB113:]] (mispreds: 0, count: 20)
 CHECK: .Ltmp[[#BB113:]] (9 instructions, align : 1)
@@ -63,6 +63,6 @@ CHECK: .Ltmp[[#BB112:]] (2 instructions, align : 1)
 CHECK:   Successors: .Ltmp[[#BB113:]] (mispreds: 0, count: 300), .LFT[[#BB11:]] (mispreds: 0, count: 20)
 CHECK: .LFT[[#BB11:]] (2 instructions, align : 1)
 CHECK: End of Function "usqrt"
-# Check the overall inference stats.
+## Check the overall inference stats.
 CHECK: 2 out of 7 functions in the binary (28.6%) have non-empty execution profile
 CHECK: inferred profile for 2 (100.00% of profiled, 100.00% of stale) functions responsible for {{.*}} samples ({{.*}} out of {{.*}})
diff --git a/bolt/test/X86/reader-stale-yaml.test b/bolt/test/X86/reader-stale-yaml.test
@@ -1,20 +1,20 @@
-# This script checks that YamlProfileReader in llvm-bolt is reading data
-# correctly and stale data is corrected by profile inference.
+## This script checks that YamlProfileReader in llvm-bolt is reading data
+## correctly and stale data is corrected by profile inference.
 
 REQUIRES: asserts
 RUN: yaml2obj %p/Inputs/blarge.yaml &> %t.exe
 RUN: llvm-bolt %t.exe -o %t.null --b %p/Inputs/blarge_profile_stale.yaml \
 RUN:   --infer-stale-profile=0 --profile-ignore-hash=1 --profile-use-dfs=0 \
 RUN:   2>&1 | FileCheck %s -check-prefix=CHECK0
-# Testing "usqrt"
+## Testing "usqrt"
 RUN: llvm-bolt %t.exe -o %t.null --b %p/Inputs/blarge_profile_stale.yaml \
 RUN:   --print-cfg --print-only=usqrt --infer-stale-profile=1 \
 RUN:   --profile-ignore-hash=1 --profile-use-dfs=0 --debug-only=bolt-prof 2>&1 | FileCheck %s -check-prefix=CHECK1
-# Testing "SolveCubic"
+## Testing "SolveCubic"
 RUN: llvm-bolt %t.exe -o %t.null --b %p/Inputs/blarge_profile_stale.yaml \
 RUN:   --print-cfg --print-only=SolveCubic --infer-stale-profile=1 \
 RUN:   --profile-ignore-hash=1 --profile-use-dfs=0 --debug-only=bolt-prof 2>&1 | FileCheck %s -check-prefix=CHECK2
-# Testing skipped function
+## Testing skipped function
 RUN: llvm-bolt %t.exe -o %t.null --b %p/Inputs/blarge_profile_stale.yaml \
 RUN:   --print-cfg --print-only=usqrt --infer-stale-profile=1 --skip-funcs=usqrt \
 RUN:   --profile-ignore-hash=1 --profile-use-dfs=0
@@ -23,12 +23,12 @@ CHECK0: BOLT-INFO: 2 out of 7 functions in the binary (28.6%) have non-empty exe
 CHECK0: BOLT-WARNING: 2 (100.0% of all profiled) functions have invalid (possibly stale) profile
 CHECK0: BOLT-WARNING: 1192 out of 1192 samples in the binary (100.0%) belong to functions with invalid (possibly stale) profile
 
-# Function "usqrt" has stale profile, since the number of blocks in the profile
-# (nblocks=6) does not match the size of the CFG in the binary. The entry
-# block (bid=0) has an incorrect (missing) count, which should be inferred by
-# the algorithm.
+## Function "usqrt" has stale profile, since the number of blocks in the profile
+## (nblocks=6) does not match the size of the CFG in the binary. The entry
+## block (bid=0) has an incorrect (missing) count, which should be inferred by
+## the algorithm.
 
-# Verify inference details.
+## Verify inference details.
 CHECK1:  pre-processing profile using YAML profile reader
 CHECK1:  applying profile inference for "usqrt"
 CHECK1: Matched yaml block (bid = 0) with hash 1111111111111111 to BB (index = 0) with hash 36007ba1d80c0000
@@ -38,7 +38,7 @@ CHECK1-NEXT: exact match
 CHECK1:  Matched yaml block (bid = 3) with hash 5c06705524800039 to BB (index = 3) with hash 5c06705524800039
 CHECK1-NEXT: exact match
 
-# Verify that yaml reader works as expected.
+## Verify that yaml reader works as expected.
 CHECK1:    Binary Function "usqrt" after building cfg {
 CHECK1:      State       : CFG constructed
 CHECK1:      Address     : 0x401170
@@ -50,7 +50,7 @@ CHECK1:      Exec Count  : 20
 CHECK1:      Branch Count: 640
 CHECK1:    }
 
-# Verify block counts.
+## Verify block counts.
 CHECK1:    .LBB01 (4 instructions, align : 1)
 CHECK1:      Successors: .Ltmp[[#BB13:]] (mispreds: 0, count: 20)
 CHECK1:    .Ltmp[[#BB13:]] (9 instructions, align : 1)
@@ -60,19 +60,19 @@ CHECK1:      Successors: .Ltmp[[#BB12:]] (mispreds: 0, count: 0)
 CHECK1:    .Ltmp[[#BB12:]] (2 instructions, align : 1)
 CHECK1:      Successors: .Ltmp[[#BB13:]] (mispreds: 0, count: 300), .LFT[[#BB1:]] (mispreds: 0, count: 20)
 CHECK1:    .LFT[[#BB1:]] (2 instructions, align : 1)
-# Check the overall inference stats.
+## Check the overall inference stats.
 CHECK1:  2 out of 7 functions in the binary (28.6%) have non-empty execution profile
 CHECK1:  BOLT-WARNING: 2 (100.0% of all profiled) functions have invalid (possibly stale) profile
 CHECK1:  BOLT-WARNING: 1192 out of 1192 samples in the binary (100.0%) belong to functions with invalid (possibly stale) profile
 CHECK1:  inferred profile for 2 (100.00% of profiled, 100.00% of stale) functions responsible for {{.*}} samples ({{.*}} out of {{.*}})
 
 
-# Function "SolveCubic" has stale profile, since there is one jump in the
-# profile (from bid=13 to bid=2) which is not in the CFG in the binary. The test
-# verifies that the inference is able to match two blocks (bid=1 and bid=13)
-# using "loose" hashes and then correctly propagate the counts.
+## Function "SolveCubic" has stale profile, since there is one jump in the
+## profile (from bid=13 to bid=2) which is not in the CFG in the binary. The test
+## verifies that the inference is able to match two blocks (bid=1 and bid=13)
+## using "loose" hashes and then correctly propagate the counts.
 
-# Verify inference details.
+## Verify inference details.
 CHECK2:  pre-processing profile using YAML profile reader
 CHECK2:  applying profile inference for "SolveCubic"
 CHECK2:  Matched yaml block (bid = 0) with hash 4600940a609c0000 to BB (index = 0) with hash 4600940a609c0000
@@ -86,7 +86,7 @@ CHECK2-NEXT:  loose match
 CHECK2:  Matched yaml block (bid = 5) with hash 6446e1ea500111 to BB (index = 5) with hash 6446e1ea500111
 CHECK2-NEXT:  exact match
 
-# Verify that yaml reader works as expected.
+## Verify that yaml reader works as expected.
 CHECK2:    Binary Function "SolveCubic" after building cfg {
 CHECK2:      State       : CFG constructed
 CHECK2:      Address     : 0x400e00
@@ -97,7 +97,7 @@ CHECK2:      BB Count    : 18
 CHECK2:      Exec Count  : 151
 CHECK2:      Branch Count: 552
 
-# Verify block counts.
+## Verify block counts.
 CHECK2:    .LBB00 (43 instructions, align : 1)
 CHECK2:      Successors: .Ltmp[[#BB7:]] (mispreds: 0, count: 0), .LFT[[#BB1:]] (mispreds: 0, count: 151)
 CHECK2:    .LFT[[#BB1:]] (5 instructions, align : 1)

diff --git a/bolt/test/X86/reader.test b/bolt/test/X86/reader.test
@@ -1,4 +1,4 @@
-# This script checks that DataReader in llvm-bolt is reading data correctly
+## This script checks that DataReader in llvm-bolt is reading data correctly
 
 RUN: yaml2obj %p/Inputs/blarge.yaml &> %t.exe
 RUN: llvm-bolt %t.exe -o %t.null --data %p/Inputs/blarge.fdata --dump-data \

diff --git a/bolt/test/X86/register-fragments-bolt-symbols.s b/bolt/test/X86/register-fragments-bolt-symbols.s
@@ -1,10 +1,22 @@
-# Test the heuristics for matching BOLT-added split functions.
+## Test the heuristics for matching BOLT-added split functions.
 
 # RUN: llvm-mc --filetype=obj --triple x86_64-unknown-unknown %S/cdsplit-symbol-names.s -o %t.main.o
 # RUN: llvm-mc --filetype=obj --triple x86_64-unknown-unknown %s -o %t.chain.o
 # RUN: link_fdata %S/cdsplit-symbol-names.s %t.main.o %t.fdata
-# RUN: sed -i 's|chain|chain/2|g' %t.fdata
 # RUN: llvm-strip --strip-unneeded %t.main.o
+
+## Check warm fragment name matching (produced by cdsplit)
+# RUN: %clang %cflags %t.main.o -o %t.warm.exe -Wl,-q
+# RUN: llvm-bolt %t.warm.exe -o %t.warm.bolt --split-functions --split-strategy=cdsplit \
+# RUN:   --call-scale=2 --data=%t.fdata --reorder-blocks=ext-tsp --enable-bat
+# RUN: link_fdata %s %t.warm.bolt %t.preagg.warm PREAGGWARM
+# PREAGGWARM: B X:0 #chain.warm# 1 0
+# RUN: perf2bolt %t.warm.bolt -p %t.preagg.warm --pa -o %t.warm.fdata -w %t.warm.yaml \
+# RUN:   -v=1 | FileCheck %s --check-prefix=CHECK-BOLT-WARM
+
+# CHECK-BOLT-WARM: marking chain.warm/1(*2) as a fragment of chain
+
+# RUN: sed -i 's|chain|chain/2|g' %t.fdata
 # RUN: llvm-objcopy --localize-symbol=chain %t.main.o
 # RUN: %clang %cflags %t.chain.o %t.main.o -o %t.exe -Wl,-q
 # RUN: llvm-bolt %t.exe -o %t.bolt --split-functions --split-strategy=randomN \

diff --git a/bolt/test/X86/relaxed-tailcall.test b/bolt/test/X86/relaxed-tailcall.test
@@ -1,4 +1,4 @@
-# Check that tail calls can be 2 bytes in the output binary.
+## Check that tail calls can be 2 bytes in the output binary.
 
 RUN: llvm-mc -filetype=obj -triple=x86_64-unknown-unknown -o %t.o \
 RUN:   %S/Inputs/relaxed_tc.s

diff --git a/bolt/test/X86/remove-unused.test b/bolt/test/X86/remove-unused.test
@@ -1,5 +1,5 @@
-# Verifies that llvm-bolt is able to remove dead basic blocks. Also check that
-# the BB reordering ignores dead BBs.
+## Verifies that llvm-bolt is able to remove dead basic blocks. Also check that
+## the BB reordering ignores dead BBs.
 
 RUN: llvm-mc -filetype=obj -triple x86_64-unknown-unknown %S/Inputs/entry.s -o %t.o
 RUN: link_fdata %S/Inputs/entry.s %t.o %t.fdata --nmtool llvm-nm
@@ -9,5 +9,5 @@ RUN: llvm-bolt %t.exe --data %t.fdata -o %t --funcs=_start \
 RUN:   --eliminate-unreachable --reorder-blocks=none \
 RUN:   --print-finalized --sequential-disassembly 2>&1 | FileCheck %s
 
-# Optimized
+## Optimized
 CHECK:  BB Layout : .LBB00, .Ltmp0, .Ltmp2, .Ltmp3, .Ltmp4, .Ltmp5, .Ltmp6, .Ltmp7, .Ltmp8, .Ltmp9, .Ltmp10, .Ltmp11
diff --git a/bolt/test/X86/rodata-simpl-loads.test b/bolt/test/X86/rodata-simpl-loads.test
@@ -1,4 +1,4 @@
-# Check for the simplification of .rodata loads.
+## Check for the simplification of .rodata loads.
 
 RUN: %clang %cflags %p/Inputs/rodata_simpl_loads.s -o %t.exe
 RUN: llvm-bolt %t.exe -o %t --simplify-rodata-loads
@@ -7,8 +7,8 @@ RUN: llvm-objdump -d %t --print-imm-hex --disassemble-symbols=main | FileCheck %
 
 CHECK:   Disassembly of section .text:
 CHECK:   <main>:
-# check that the following rip-relative operands have been
-# replaced with immediates
+## check that the following rip-relative operands have been
+## replaced with immediates
 
 ORIGINAL: movzbl  s1(%rip), %eax
 CHECK:    movl    $0x41, %eax

diff --git a/bolt/test/X86/sctc-bug.test b/bolt/test/X86/sctc-bug.test
@@ -1,4 +1,4 @@
-# Check that we don't accidentally optimize out a tail call.
+## Check that we don't accidentally optimize out a tail call.
 
 RUN: %clang %cflags %S/Inputs/sctc_bug.s -o %t
 RUN: llvm-bolt %t -o %t.null --funcs=main --print-after-lowering \

diff --git a/bolt/test/X86/sctc-bug2.test b/bolt/test/X86/sctc-bug2.test
@@ -1,4 +1,4 @@
-# Check that conditional tail call is not treated as a regular tail call by SCTC.
+## Check that conditional tail call is not treated as a regular tail call by SCTC.
 
 RUN: %clang %cflags %S/Inputs/sctc_bug2.s -o %t
 RUN: llvm-bolt %t -o %t.null --funcs=main --print-after-lowering \

diff --git a/bolt/test/X86/sctc-bug3.test b/bolt/test/X86/sctc-bug3.test
@@ -1,4 +1,4 @@
-# Check that we don't accidentally optimize out a tail call.
+## Check that we don't accidentally optimize out a tail call.
 
 RUN: %clang %cflags %S/Inputs/sctc_bug3.s -o %t
 RUN: llvm-bolt %t -o %t.null --funcs=main --print-after-lowering \
@@ -7,9 +7,9 @@ RUN:   --sequential-disassembly 2>&1 | FileCheck %s
 CHECK: .LBB00 (1 instructions, align : 1)
 CHECK:  cmpq   %rdi, 0x0
 
-# Check that .Ltmp0 does not have a deleted predecessor.
+## Check that .Ltmp0 does not have a deleted predecessor.
 CHECK: .Ltmp0 (1 instructions, align : 1)
 CHECK:  Predecessors: .LBB00
 
-# Tail call.
+## Tail call.
 CHECK:    jmp    foo
diff --git a/bolt/test/X86/sctc-bug4.test b/bolt/test/X86/sctc-bug4.test
@@ -1,5 +1,5 @@
-# Check that fallthrough blocks are handled properly and Offset annotation is
-# set for conditional tail calls.
+## Check that fallthrough blocks are handled properly and Offset annotation is
+## set for conditional tail calls.
 
 RUN: %clang %cflags %S/Inputs/sctc_bug4.s -o %t
 RUN: llvm-bolt %t -o %t.null --enable-bat \

diff --git a/bolt/test/X86/shared_object_entry.s b/bolt/test/X86/shared_object_entry.s
@@ -4,7 +4,7 @@
 # RUN:    -split-functions -reorder-blocks=ext-tsp -split-all-cold \
 # RUN:    -dyno-stats -icf=1 -use-gnu-stack
 
-# Check that an entry point is a cold symbol
+## Check that an entry point is a cold symbol
 # RUN: llvm-readelf -h %t.so > %t.log
 # RUN: llvm-nm %t.so >> %t.log
 # RUN: FileCheck %s --input-file %t.log

diff --git a/bolt/test/X86/shorten-mov.test b/bolt/test/X86/shorten-mov.test
@@ -1,5 +1,5 @@
-# Test that 64 bit movq instructions with immediate operands
-# that fit in 32 bits are shortened.
+## Test that 64 bit movq instructions with immediate operands
+## that fit in 32 bits are shortened.
 
 RUN: %clang %cflags %p/Inputs/asm_main.c %p/Inputs/shorten_mov.s -o %t.exe
 RUN: llvm-bolt %t.exe -o %t

diff --git a/bolt/test/X86/shrinkwrapping-and-rsp.s b/bolt/test/X86/shrinkwrapping-and-rsp.s
@@ -1,5 +1,5 @@
-# This checks that shrink wrapping does attempt at accessing stack elements
-# using RSP when the function is aligning RSP and changing offsets.
+## This checks that shrink wrapping does attempt at accessing stack elements
+## using RSP when the function is aligning RSP and changing offsets.
 
 # REQUIRES: system-linux
 
@@ -12,10 +12,10 @@
 # RUN:     --frame-opt=all --simplify-conditional-tail-calls=false \
 # RUN:     --eliminate-unreachable=false | FileCheck %s
 
-# Here we have a function that aligns the stack at prologue. Stack pointer
-# analysis can't try to infer offset positions after AND because that depends
-# on the runtime value of the stack pointer of callee (whether it is misaligned
-# or not).
+## Here we have a function that aligns the stack at prologue. Stack pointer
+## analysis can't try to infer offset positions after AND because that depends
+## on the runtime value of the stack pointer of callee (whether it is misaligned
+## or not).
   .globl _start
   .type _start, %function
 _start:

diff --git a/bolt/test/X86/shrinkwrapping-critedge.s b/bolt/test/X86/shrinkwrapping-critedge.s
@@ -1,5 +1,5 @@
-# This reproduces a bug with shrink wrapping when trying to split critical
-# edges originating at the same basic block.
+## This reproduces a bug with shrink wrapping when trying to split critical
+## edges originating at the same basic block.
 
 # REQUIRES: system-linux
 

diff --git a/bolt/test/X86/shrinkwrapping-do-not-pessimize.s b/bolt/test/X86/shrinkwrapping-do-not-pessimize.s
@@ -1,10 +1,10 @@
-# This checks that shrink wrapping does not pessimize a CFG pattern where two
-# blocks can be proved to have the same execution count but, because of profile
-# inaccuricies, we could move saves into the second block. We can prove two
-# blocks have the same frequency when B post-dominate A and A dominates B and
-# are at the same loop nesting level. This would be a pessimization because
-# shrink wrapping is unlikely to be able to cleanly move PUSH instructions,
-# inserting additional store instructions.
+## This checks that shrink wrapping does not pessimize a CFG pattern where two
+## blocks can be proved to have the same execution count but, because of profile
+## inaccuricies, we could move saves into the second block. We can prove two
+## blocks have the same frequency when B post-dominate A and A dominates B and
+## are at the same loop nesting level. This would be a pessimization because
+## shrink wrapping is unlikely to be able to cleanly move PUSH instructions,
+## inserting additional store instructions.
 
 # REQUIRES: system-linux
 
@@ -16,15 +16,15 @@
 # RUN: llvm-bolt -relocs %t.exe -o %t.out -data %t.fdata \
 # RUN:     -frame-opt=all -equalize-bb-counts | FileCheck %s
 
-# Here we create a CFG pattern with two blocks A and B belonging to the same
-# equivalency class as defined by dominance relations and having in theory
-# the same frequency. But we tweak edge counts from profile to make block A
-# hotter than block B.
+## Here we create a CFG pattern with two blocks A and B belonging to the same
+## equivalency class as defined by dominance relations and having in theory
+## the same frequency. But we tweak edge counts from profile to make block A
+## hotter than block B.
   .globl _start
   .type _start, %function
 _start:
   .cfi_startproc
-# Hot prologue
+## Hot prologue
 # FDATA: 0 [unknown] 0 1 _start 0 0 10
   push  %rbp
   mov   %rsp, %rbp
@@ -36,7 +36,7 @@ b:  je  end_if_1
 if_false:
   movq rel(%rip), %rdi  # Add this to create a relocation and run bolt w/ relocs
 c:  jmp end_if_1
-# Reduce frequency from 9 to 1 to simulate an inaccurate profile
+## Reduce frequency from 9 to 1 to simulate an inaccurate profile
 # FDATA: 1 _start #c# 1 _start #end_if_1# 0 1
 end_if_1:
   # first uses of R14 and RBX appear at this point, possible move point for SW

diff --git a/bolt/test/X86/shrinkwrapping-insertcfi.s b/bolt/test/X86/shrinkwrapping-insertcfi.s
@@ -1,5 +1,5 @@
-# This test reproduces the issue with inserting updated CFI in shrink wrapping
-# into the first basic block.
+## This test reproduces the issue with inserting updated CFI in shrink wrapping
+## into the first basic block.
 
 # REQUIRES: system-linux
 
@@ -10,10 +10,10 @@
 # RUN: llvm-bolt %t.exe -o %t.out --data %t.fdata --frame-opt=all --lite=0 \
 # RUN:           --print-fop 2>&1 | FileCheck %s
 
-# Check shrink wrapping results:
+## Check shrink wrapping results:
 # CHECK: BOLT-INFO: Shrink wrapping moved 0 spills inserting load/stores and 1 spills inserting push/pops
 
-# Check that CFI is successfully inserted into the first basic block:
+## Check that CFI is successfully inserted into the first basic block:
 # CHECK:      Binary Function "_start" after frame-optimizer
 # CHECK:      .LBB00 (2 instructions, align : 1)
 # CHECK-NEXT: Entry Point
@@ -34,8 +34,8 @@ c:
     .cfi_offset 3, 4
     pop	%rbx
 
-# This basic block is treated as having 0 execution count.
-# push and pop will be sinked into this block.
+## This basic block is treated as having 0 execution count.
+## push and pop will be sinked into this block.
 a:
     ud2
     .cfi_endproc
diff --git a/bolt/test/X86/shrinkwrapping-lea.s b/bolt/test/X86/shrinkwrapping-lea.s
@@ -1,5 +1,5 @@
-# This checks that shrink wrapping correctly drops moving push/pops when
-# there is an LEA instruction.
+## This checks that shrink wrapping correctly drops moving push/pops when
+## there is an LEA instruction.
 
 
 # REQUIRES: system-linux
@@ -58,7 +58,7 @@ JT:
 
 # CHECK:   BOLT-INFO: Shrink wrapping moved 2 spills inserting load/stores and 0 spills inserting push/pops
 
-# Checks that offsets of instructions accessing the stack were not changed
+## Checks that offsets of instructions accessing the stack were not changed
 # CHECK-OBJDUMP:     <_start>:
 # CHECK-OBJDUMP:         movq    %rbx, %rdi
 # CHECK-OBJDUMP-NEXT:    leaq    -0x20(%rbp), %r14

diff --git a/bolt/test/X86/shrinkwrapping-mov.s b/bolt/test/X86/shrinkwrapping-mov.s
@@ -1,6 +1,6 @@
-# This checks that shrink wrapping correctly drops moving push/pops when
-# there is a MOV instruction loading the value of the stack pointer in
-# order to do pointer arithmetic with a stack address.
+## This checks that shrink wrapping correctly drops moving push/pops when
+## there is a MOV instruction loading the value of the stack pointer in
+## order to do pointer arithmetic with a stack address.
 
 
 # REQUIRES: system-linux

diff --git a/bolt/test/X86/shrinkwrapping-pop-order.s b/bolt/test/X86/shrinkwrapping-pop-order.s
@@ -1,6 +1,6 @@
-# This test reproduces a POP reordering issue in shrink wrapping where we would
-# incorrectly put a store after a load (instead of before) when having multiple
-# insertions at the same point. Check that the order is correct in this test.
+## This test reproduces a POP reordering issue in shrink wrapping where we would
+## incorrectly put a store after a load (instead of before) when having multiple
+## insertions at the same point. Check that the order is correct in this test.
 
 # REQUIRES: system-linux
 
@@ -25,23 +25,23 @@ c:
   pop   %rbp
   pop   %rbx
 
-# This basic block is treated as having 0 execution count.
-# push and pop will be sinked into this block.
+## This basic block is treated as having 0 execution count.
+## push and pop will be sinked into this block.
 a:
     ud2
     .cfi_endproc
 
 
-# Check shrink wrapping results:
+## Check shrink wrapping results:
 # CHECK: BOLT-INFO: Shrink wrapping moved 0 spills inserting load/stores and 2 spills inserting push/pops
 # CHECK: BOLT-INFO: Shrink wrapping reduced 6 store executions (28.6% total instructions executed, 100.0% store instructions)
 # CHECK: BOLT-INFO: Shrink wrapping failed at reducing 0 store executions (0.0% total instructions executed, 0.0% store instructions)
 
-# Check that order is correct
+## Check that order is correct
 # CHECK:      Binary Function "_start" after frame-optimizer
 # Pushes are ordered according to their reg number and come first
 # CHECK:      pushq   %rbp
 # CHECK:      pushq   %rbx
-# Pops are ordered according to their dominance relation and come last
+## Pops are ordered according to their dominance relation and come last
 # CHECK:      popq    %rbx
 # CHECK:      popq    %rbp
diff --git a/bolt/test/X86/shrinkwrapping-popf.s b/bolt/test/X86/shrinkwrapping-popf.s
@@ -1,4 +1,4 @@
-# This test checks that POPF will not crash our frame analysis pass
+## This test checks that POPF will not crash our frame analysis pass
 
 # REQUIRES: system-linux
 
@@ -26,7 +26,7 @@ c:
     pop   %rbx
     popf
 
-# This basic block is treated as having 0 execution count.
+## This basic block is treated as having 0 execution count.
 a:
     ud2
     .cfi_endproc
diff --git a/bolt/test/X86/shrinkwrapping-restore-position.s b/bolt/test/X86/shrinkwrapping-restore-position.s
@@ -1,5 +1,5 @@
-# This checks that shrink wrapping uses the red zone defined in the X86 ABI by
-# placing restores that access elements already deallocated by the stack.
+## This checks that shrink wrapping uses the red zone defined in the X86 ABI by
+## placing restores that access elements already deallocated by the stack.
 
 # REQUIRES: system-linux
 
@@ -16,10 +16,10 @@
 # RUN:   FileCheck --check-prefix CHECK-OBJDUMP %s
 
 
-# Here we create a CFG where the restore position matches the previous (deleted)
-# restore position. Shrink wrapping then will put a stack access to an element
-# that was deallocated at the previously deleted POP, which falls in the red
-# zone and should be safe for X86 Linux ABI.
+## Here we create a CFG where the restore position matches the previous (deleted)
+## restore position. Shrink wrapping then will put a stack access to an element
+## that was deallocated at the previously deleted POP, which falls in the red
+## zone and should be safe for X86 Linux ABI.
   .globl _start
   .type _start, %function
 _start:

diff --git a/bolt/test/X86/shrinkwrapping.test b/bolt/test/X86/shrinkwrapping.test
@@ -1,9 +1,9 @@
-# Verifies that llvm-bolt updates CFI correctly after
-# shrink-wrapping when optimizing a function without
-# frame pointers.
+## Verifies that llvm-bolt updates CFI correctly after
+## shrink-wrapping when optimizing a function without
+## frame pointers.
 
-# This test has commands that rely on shell capabilities that won't execute
-# correctly on Windows e.g. subshell execution to capture command output.
+## This test has commands that rely on shell capabilities that won't execute
+## correctly on Windows e.g. subshell execution to capture command output.
 REQUIRES: shell
 
 RUN: %clangxx %cxxflags -no-pie %S/Inputs/exc4sw.S -o %t.exe -Wl,-q

diff --git a/bolt/test/X86/split-all-lptrampoline.s b/bolt/test/X86/split-all-lptrampoline.s
@@ -1,6 +1,6 @@
-# This test checks that trampolines are inserted in split fragments if
-# necessary. There are 4 LSDA ranges with a landing pad to three landing pads.
-# After splitting all blocks, there have to be 4 trampolines in the output.
+## This test checks that trampolines are inserted in split fragments if
+## necessary. There are 4 LSDA ranges with a landing pad to three landing pads.
+## After splitting all blocks, there have to be 4 trampolines in the output.
 
 # RUN: llvm-mc --filetype=obj --triple x86_64-unknown-unknown %s -o %t.o
 # RUN: %clangxx %cxxflags %t.o -o %t.exe -Wl,-q -pie

diff --git a/bolt/test/X86/split-all.s b/bolt/test/X86/split-all.s
@@ -1,4 +1,4 @@
-# Test split all block strategy
+## Test split all block strategy
 
 # RUN: llvm-mc --filetype=obj --triple x86_64-unknown-unknown %s -o %t.o
 # RUN: %clang %cflags %t.o -o %t.exe -Wl,-q

diff --git a/bolt/test/X86/split-func-icf.s b/bolt/test/X86/split-func-icf.s
@@ -1,7 +1,7 @@
-# This reproduces an issue where two cold fragments are folded into one, so the
-# fragment has two parents.
-# The fragment is only reachable through a jump table, so all functions must be
-# ignored.
+## This reproduces an issue where two cold fragments are folded into one, so the
+## fragment has two parents.
+## The fragment is only reachable through a jump table, so all functions must be
+## ignored.
 
 # REQUIRES: system-linux
 
@@ -27,10 +27,10 @@ main:
 LBB0:
   andl $0xf, %ecx
   cmpb $0x4, %cl
-  # exit through ret
+  ## exit through ret
   ja LBB3
 
-# jump table dispatch, jumping to label indexed by val in %ecx
+## jump table dispatch, jumping to label indexed by val in %ecx
 LBB1:
   leaq JUMP_TABLE1(%rip), %r8
   movzbl %cl, %ecx
@@ -55,7 +55,7 @@ LBB20:
   # exit through ret
   ja LBB23
 
-# jump table dispatch, jumping to label indexed by val in %ecx
+## jump table dispatch, jumping to label indexed by val in %ecx
 LBB21:
   leaq JUMP_TABLE2(%rip), %r8
   movzbl %cl, %ecx
@@ -70,23 +70,23 @@ LBB23:
   ret
 .size main2, .-main2
 
-# cold fragment is only reachable through jump table
+## cold fragment is only reachable through jump table
   .globl main2.cold.1
   .type main2.cold.1, %function
 main2.cold.1:
   .globl main.cold.1
   .type main.cold.1, %function
   .p2align 2
 main.cold.1:
-  # load bearing nop: pad LBB4 so that it can't be treated
-  # as __builtin_unreachable by analyzeJumpTable
+  ## load bearing nop: pad LBB4 so that it can't be treated
+  ## as __builtin_unreachable by analyzeJumpTable
   nop
 LBB4:
   callq abort
 .size main.cold.1, .-main.cold.1
 
   .rodata
-# jmp table, entries must be R_X86_64_PC32 relocs
+## jmp table, entries must be R_X86_64_PC32 relocs
   .globl JUMP_TABLE1
 JUMP_TABLE1:
   .long LBB2-JUMP_TABLE1

diff --git a/bolt/test/X86/split-func-jump-table-fragment-bidirection.s b/bolt/test/X86/split-func-jump-table-fragment-bidirection.s
@@ -1,7 +1,7 @@
-# This reproduces an issue where two fragments of same function access same
-# jump table, which means at least one fragment visits the other, i.e., one
-# of them has split jump table. As a result, all of them will be marked as
-# non-simple function.
+## This reproduces an issue where two fragments of same function access same
+## jump table, which means at least one fragment visits the other, i.e., one
+## of them has split jump table. As a result, all of them will be marked as
+## non-simple function.
 
 # REQUIRES: system-linux
 
@@ -21,10 +21,10 @@ main:
 LBB0:
   andl $0xf, %ecx
   cmpb $0x4, %cl
-  # exit through ret
+  ## exit through ret
   ja LBB3
 
-# jump table dispatch, jumping to label indexed by val in %ecx
+## jump table dispatch, jumping to label indexed by val in %ecx
 LBB1:
   leaq JUMP_TABLE1(%rip), %r8
   movzbl %cl, %ecx
@@ -39,21 +39,21 @@ LBB3:
   ret
 .size main, .-main
 
-# cold fragment is only reachable
+## cold fragment is only reachable
   .globl main.cold.1
   .type main.cold.1, %function
   .p2align 2
 main.cold.1:
-  # load bearing nop: pad LBB8 so that it can't be treated
-  # as __builtin_unreachable by analyzeJumpTable
+  ## load bearing nop: pad LBB8 so that it can't be treated
+  ## as __builtin_unreachable by analyzeJumpTable
   nop
 LBB4:
   andl $0xb, %ebx
   cmpb $0x1, %cl
   # exit through ret
   ja LBB7
 
-# jump table dispatch, jumping to label indexed by val in %ecx
+## jump table dispatch, jumping to label indexed by val in %ecx
 LBB5:
   leaq JUMP_TABLE1(%rip), %r8
   movzbl %cl, %ecx
@@ -71,7 +71,7 @@ LBB8:
 .size main.cold.1, .-main.cold.1
 
   .rodata
-# jmp table, entries must be R_X86_64_PC32 relocs
+## jmp table, entries must be R_X86_64_PC32 relocs
   .globl JUMP_TABLE1
 JUMP_TABLE1:
   .long LBB2-JUMP_TABLE1

diff --git a/bolt/test/X86/split-func-jump-table-fragment-noparent.s b/bolt/test/X86/split-func-jump-table-fragment-noparent.s
@@ -1,6 +1,6 @@
-# This reproduces a bug with jump table identification where jump table has
-# entries pointing to code in function and its cold fragment.
-# The fragment is only reachable through jump table.
+## This reproduces a bug with jump table identification where jump table has
+## entries pointing to code in function and its cold fragment.
+## The fragment is only reachable through jump table.
 
 # REQUIRES: system-linux
 
@@ -19,10 +19,10 @@ main:
 LBB0:
   andl $0xf, %ecx
   cmpb $0x4, %cl
-  # exit through ret
+  ## exit through ret
   ja LBB3
 
-# jump table dispatch, jumping to label indexed by val in %ecx
+## jump table dispatch, jumping to label indexed by val in %ecx
 LBB1:
   leaq JUMP_TABLE(%rip), %r8
   movzbl %cl, %ecx
@@ -37,20 +37,20 @@ LBB3:
   ret
 .size main, .-main
 
-# cold fragment is only reachable through jump table
+## cold fragment is only reachable through jump table
   .globl main.cold.1
   .type main.cold.1, %function
   .p2align 2
 main.cold.1:
-  # load bearing nop: pad LBB4 so that it can't be treated
-  # as __builtin_unreachable by analyzeJumpTable
+  ## load bearing nop: pad LBB4 so that it can't be treated
+  ## as __builtin_unreachable by analyzeJumpTable
   nop
 LBB4:
   callq abort
 .size main.cold.1, .-main.cold.1
 
   .rodata
-# jmp table, entries must be R_X86_64_PC32 relocs
+## jmp table, entries must be R_X86_64_PC32 relocs
   .globl JUMP_TABLE
 JUMP_TABLE:
   .long LBB2-JUMP_TABLE

diff --git a/bolt/test/X86/split-func-jump-table-fragment-reverse.s b/bolt/test/X86/split-func-jump-table-fragment-reverse.s
@@ -1,6 +1,6 @@
-# This reproduces a bug with jump table identification where jump table has
-# entries pointing to code in function and its cold fragment.
-# The fragment is only reachable through jump table.
+## This reproduces a bug with jump table identification where jump table has
+## entries pointing to code in function and its cold fragment.
+## The fragment is only reachable through jump table.
 
 # REQUIRES: system-linux
 
@@ -26,10 +26,10 @@ main.cold:
 LBB0:
   andl $0xf, %ecx
   cmpb $0x4, %cl
-  # exit through ret
+  ## exit through ret
   ja LBB3
 
-# jump table dispatch, jumping to label indexed by val in %ecx
+## jump table dispatch, jumping to label indexed by val in %ecx
 LBB1:
   leaq JUMP_TABLE(%rip), %r8
   movzbl %cl, %ecx
@@ -44,20 +44,20 @@ LBB3:
   ret
 .size main.cold, .-main.cold
 
-# main function, referenced from jump table in cold fragment
+## main function, referenced from jump table in cold fragment
   .globl main
   .type main, %function
   .p2align 2
 main:
-  # load bearing nop: pad LBB4 so that it can't be treated
-  # as __builtin_unreachable by analyzeJumpTable
+  ## load bearing nop: pad LBB4 so that it can't be treated
+  ## as __builtin_unreachable by analyzeJumpTable
   nop
 LBB4:
   callq abort
 .size main, .-main
 
   .rodata
-# jmp table, entries must be R_X86_64_PC32 relocs
+## jmp table, entries must be R_X86_64_PC32 relocs
   .globl JUMP_TABLE
 JUMP_TABLE:
   .long LBB2-JUMP_TABLE

diff --git a/bolt/test/X86/split-func-jump-table-fragment.s b/bolt/test/X86/split-func-jump-table-fragment.s
@@ -19,10 +19,10 @@ main:
 LBB0:
   andl $0xf, %ecx
   cmpb $0x4, %cl
-  # exit through abort in main.cold.1, registers cold fragment the regular way
+  ## exit through abort in main.cold.1, registers cold fragment the regular way
   ja main.cold.1
 
-# jump table dispatch, jumping to label indexed by val in %ecx
+## jump table dispatch, jumping to label indexed by val in %ecx
 LBB1:
   leaq JUMP_TABLE(%rip), %r8
   movzbl %cl, %ecx
@@ -37,8 +37,8 @@ LBB3:
   ret
 .size main, .-main
 
-# Insert padding between functions, so that the next instruction cannot be
-# treated as __builtin_unreachable destination for the jump table.
+## Insert padding between functions, so that the next instruction cannot be
+## treated as __builtin_unreachable destination for the jump table.
   .quad 0
 
   .globl main.cold.1
@@ -50,7 +50,7 @@ LBB4:
 .size main.cold.1, .-main.cold.1
 
   .rodata
-# jmp table, entries must be R_X86_64_PC32 relocs
+## jmp table, entries must be R_X86_64_PC32 relocs
   .globl JUMP_TABLE
 JUMP_TABLE:
   .long LBB2-JUMP_TABLE

diff --git a/bolt/test/X86/split-func-jump-table-unknown.s b/bolt/test/X86/split-func-jump-table-unknown.s
@@ -1,5 +1,5 @@
-# This reproduces a bug with converting an unknown control flow jump table with
-# entries pointing to code in function and its cold fragment.
+## This reproduces a bug with converting an unknown control flow jump table with
+## entries pointing to code in function and its cold fragment.
 
 # REQUIRES: system-linux
 
@@ -27,10 +27,10 @@ LBB0:
   leaq JUMP_TABLE(%rip), %r8
   andl $0xf, %ecx
   cmpb $0x4, %cl
-  # exit through abort in main.cold.1, registers cold fragment the regular way
+  ## exit through abort in main.cold.1, registers cold fragment the regular way
   ja main.cold.1
 
-# jump table dispatch, jumping to label indexed by val in %ecx
+## jump table dispatch, jumping to label indexed by val in %ecx
 LBB1:
   movzbl %cl, %ecx
   movslq (%r8,%rcx,4), %rax
@@ -48,15 +48,15 @@ LBB3:
   .type main.cold.1, %function
   .p2align 2
 main.cold.1:
-  # load bearing nop: pad LBB4 so that it can't be treated
-  # as __builtin_unreachable by analyzeJumpTable
+  ## load bearing nop: pad LBB4 so that it can't be treated
+  ## as __builtin_unreachable by analyzeJumpTable
   nop
 LBB4:
   callq abort
 .size main.cold.1, .-main.cold.1
 
   .rodata
-# jmp table, entries must be R_X86_64_PC32 relocs
+## jmp table, entries must be R_X86_64_PC32 relocs
   .globl JUMP_TABLE
 JUMP_TABLE:
   .long LBB2-JUMP_TABLE

diff --git a/bolt/test/X86/split-landing-pad.s b/bolt/test/X86/split-landing-pad.s
@@ -1,25 +1,25 @@
-# This test reproduces the case where C++ exception handling is used and split
-# function optimization is enabled. In particular, function foo is splitted
-# to two fragments:
-#    foo: contains 2 try blocks, which invokes bar to throw exception
-#    foo.cold.1: contains 2 corresponding catch blocks (landing pad)
-#
-# Similar to split jump table, split landing pad target to different fragment.
-# This test is written to ensure BOLT safely handle these targets, e.g., by
-# marking them as non-simple.
-#
-# Steps to write this test:
-# - Create a copy of Inputs/src/unreachable.cpp
-# - Simplify bar(), focus on throw an exception
-# - Create the second switch case in foo() to have multiple landing pads
-# - Compile with clang++ to .s
-# - Move landing pad code from foo to foo.cold.1
-# - Ensure that all landing pads can be reached normally
-#
-# Additional details:
-# .gcc_except_table specify the landing pads for try blocks
-#    LPStart = 255 (omit), which means LPStart = foo start
-# Landing pads .Ltmp2 and .Ltmp5 in call site record are offset to foo start.
+## This test reproduces the case where C++ exception handling is used and split
+## function optimization is enabled. In particular, function foo is splitted
+## to two fragments:
+##    foo: contains 2 try blocks, which invokes bar to throw exception
+##    foo.cold.1: contains 2 corresponding catch blocks (landing pad)
+##
+## Similar to split jump table, split landing pad target to different fragment.
+## This test is written to ensure BOLT safely handle these targets, e.g., by
+## marking them as non-simple.
+##
+## Steps to write this test:
+## - Create a copy of Inputs/src/unreachable.cpp
+## - Simplify bar(), focus on throw an exception
+## - Create the second switch case in foo() to have multiple landing pads
+## - Compile with clang++ to .s
+## - Move landing pad code from foo to foo.cold.1
+## - Ensure that all landing pads can be reached normally
+##
+## Additional details:
+## .gcc_except_table specify the landing pads for try blocks
+##    LPStart = 255 (omit), which means LPStart = foo start
+## Landing pads .Ltmp2 and .Ltmp5 in call site record are offset to foo start.
 
 
 # REQUIRES: system-linux

diff --git a/bolt/test/X86/split-random.s b/bolt/test/X86/split-random.s
@@ -1,4 +1,4 @@
-# Test random function splitting option
+## Test random function splitting option
 
 # RUN: llvm-mc --filetype=obj --triple x86_64-unknown-unknown %s -o %t.o
 # RUN: %clang %cflags %t.o -o %t.exe -Wl,-q

diff --git a/bolt/test/X86/static-exe.test b/bolt/test/X86/static-exe.test
@@ -1,4 +1,4 @@
-# Check that llvm-bolt can rewrite static executable
+## Check that llvm-bolt can rewrite static executable
 
 RUN: %clang %cflags %S/Inputs/static_exe.s -static -o %t.exe -nostdlib
 RUN: llvm-bolt %t.exe -o %t 2>&1 | FileCheck %s

diff --git a/bolt/test/X86/symtab-secondary-entries.test b/bolt/test/X86/symtab-secondary-entries.test
@@ -1,4 +1,4 @@
-# Check that secondary entry points are updated correctly in the ELF symtab
+## Check that secondary entry points are updated correctly in the ELF symtab
 
 RUN: llvm-mc -filetype=obj -triple x86_64-unknown-unknown \
 RUN:   %p/Inputs/user-order.S -o %t.o
@@ -13,7 +13,7 @@ CHECK: [[#]] FUNC    GLOBAL DEFAULT   [[#NDX]] main
 CHECK: [[#]] FUNC    LOCAL  DEFAULT   [[#NDX]] _a
 CHECK: [[#]] FUNC    GLOBAL DEFAULT   [[#NDX]] _b
 CHECK: [[#]] FUNC    GLOBAL DEFAULT   [[#NDX]] _f
-# The following are all secondary entries of _f
+## The following are all secondary entries of _f
 CHECK:     0 FUNC    GLOBAL DEFAULT   [[#NDX]] _c
 CHECK:     0 FUNC    GLOBAL DEFAULT   [[#NDX]] _d
 CHECK:     0 FUNC    GLOBAL DEFAULT   [[#NDX]] _e
diff --git a/bolt/test/X86/tail-duplication-cache.s b/bolt/test/X86/tail-duplication-cache.s
@@ -11,7 +11,7 @@
 # RUN:    --print-finalized --tail-duplication=cache -o %t.out2 \
 # RUN:    | FileCheck --check-prefix="CHECK2" %s
 
-# A test where the tail is duplicated to eliminate an unconditional jump
+## A test where the tail is duplicated to eliminate an unconditional jump
 # FDATA: 1 main #.BB0_br# 1 main #.BB4# 0 100
 # FDATA: 1 main #.BB0_br# 1 main #.BB1# 0 100
 # FDATA: 1 main #.BB1_br# 1 main #.BB3# 0 50
@@ -20,7 +20,7 @@
 # CHECK: BOLT-INFO: tail duplication modified 1 ({{.*}}%) functions; duplicated 1 blocks (13 bytes) responsible for 50 dynamic executions ({{.*}}% of all block executions)
 # CHECK: BB Layout   : .LBB00, .Ltmp0, .Ltmp1, .Ltmp2, .Ltmp3, .Ltmp4, .Ltmp5, .Ltail-dup0, .Ltmp6
 
-# A test where the tail is not duplicated due to the cache score
+## A test where the tail is not duplicated due to the cache score
 # FDATA2: 1 main #.BB0_br# 1 main #.BB4# 0 100
 # FDATA2: 1 main #.BB0_br# 1 main #.BB1# 0 2
 # FDATA2: 1 main #.BB1_br# 1 main #.BB3# 0 1

diff --git a/bolt/test/X86/tail-duplication-cacheline.s b/bolt/test/X86/tail-duplication-cacheline.s
@@ -1,5 +1,5 @@
-# This reproduces a bug in TailDuplication::isInCacheLine
-# with accessing BlockLayout past bounds (unreachable blocks).
+## This reproduces a bug in TailDuplication::isInCacheLine
+## with accessing BlockLayout past bounds (unreachable blocks).
 
 # REQUIRES: system-linux
 

diff --git a/bolt/test/X86/tail-duplication-complex.s b/bolt/test/X86/tail-duplication-complex.s
@@ -17,12 +17,12 @@
 # CHECK: tail duplication modified 1 ({{.*}}%) functions; duplicated 1 blocks ({{.*}} bytes) responsible for {{.*}} dynamic executions ({{.*}} of all block executions)
 # CHECK: BB Layout   : .LBB00, .Ltmp0, .Ltail-dup0, .Ltmp1, .Ltmp2
 
-# This is the C++ code fed to Clang
-# int fib(int term) {
-#   if (term <= 1)
-#     return term;
-#   return fib(term-1) + fib(term-2);
-# }
+## This is the C++ code fed to Clang
+## int fib(int term) {
+##   if (term <= 1)
+##     return term;
+##   return fib(term-1) + fib(term-2);
+## }
 
     .text
     .globl main

diff --git a/bolt/test/X86/tail-duplication-jt.s b/bolt/test/X86/tail-duplication-jt.s
@@ -1,5 +1,5 @@
-# This reproduces a bug in tail duplication when aggressiveCodeToDuplicate
-# fails to handle a block with a jump table.
+## This reproduces a bug in tail duplication when aggressiveCodeToDuplicate
+## fails to handle a block with a jump table.
 
 # REQUIRES: system-linux
 

diff --git a/bolt/test/X86/tail-duplication-pass.s b/bolt/test/X86/tail-duplication-pass.s
@@ -16,7 +16,7 @@
 # CHECK: BOLT-INFO: tail duplication modified 1 ({{.*}}%) functions; duplicated 1 blocks (1 bytes) responsible for {{.*}} dynamic executions ({{.*}}% of all block executions)
 # CHECK: BB Layout   : .LBB00, .Ltail-dup0, .Ltmp0, .Ltmp1
 
-# Check that the successor of Ltail-dup0 is .LBB00, not itself.
+## Check that the successor of Ltail-dup0 is .LBB00, not itself.
 # CHECK-NOLOOP: .Ltail-dup0 (1 instructions, align : 1)
 # CHECK-NOLOOP: Predecessors: .LBB00
 # CHECK-NOLOOP: retq

diff --git a/bolt/test/X86/tail-duplication-prop-bug.s b/bolt/test/X86/tail-duplication-prop-bug.s
@@ -1,4 +1,4 @@
-# This reproduces a bug in aggressive tail duplication/copy propagation.
+## This reproduces a bug in aggressive tail duplication/copy propagation.
 
 # REQUIRES: system-linux
 # RUN: llvm-mc -filetype=obj -triple x86_64-unknown-unknown %s -o %t.o

diff --git a/bolt/test/X86/tailcall-traps.test b/bolt/test/X86/tailcall-traps.test
@@ -1,4 +1,4 @@
-# Tests the peephole that adds trap instructions following indirect tail calls.
+## Tests the peephole that adds trap instructions following indirect tail calls.
 
 RUN: %clang %cflags %p/Inputs/tailcall_traps.s -o %t.exe
 RUN: llvm-bolt %t.exe -o %t --peepholes=tailcall-traps \

diff --git a/bolt/test/X86/tailcall.test b/bolt/test/X86/tailcall.test
@@ -1,5 +1,5 @@
-# Verifies that llvm-bolt recognizes tailcalls and mark them
-# in control flow graph.
+## Verifies that llvm-bolt recognizes tailcalls and mark them
+## in control flow graph.
 
 RUN: %clang %cflags %S/Inputs/tailcall.s -o %t.exe
 RUN: llvm-bolt %t.exe -o %t.null --print-cfg 2>&1 | FileCheck %s

diff --git a/bolt/test/X86/unclaimed-jt-entries.s b/bolt/test/X86/unclaimed-jt-entries.s
@@ -1,5 +1,5 @@
-# This test ensures that "unclaimed" jump table entries are accounted later
-# in postProcessIndirectBranches and the function is marked as non-simple.
+## This test ensures that "unclaimed" jump table entries are accounted later
+## in postProcessIndirectBranches and the function is marked as non-simple.
 
 # The test is compiled from the following source using GCC 12.2 -O3:
 # https://godbolt.org/z/YcPG131s6

diff --git a/bolt/test/X86/unreachable-jmp.s b/bolt/test/X86/unreachable-jmp.s
@@ -1,5 +1,5 @@
-# This checks that we don't create an invalid CFG when there is an
-# unreachable direct jump right after an indirect one.
+## This checks that we don't create an invalid CFG when there is an
+## unreachable direct jump right after an indirect one.
 
 # REQUIRES: system-linux
 
@@ -25,8 +25,8 @@ _start:
 b:
   jmpq  *JUMP_TABLE(,%rcx,8)
 # FDATA: 1 _start #b# 1 _start #hotpath# 0 20
-# Unreachable direct jump here. Our CFG should still make sense and properly
-# place this instruction in a new basic block.
+## Unreachable direct jump here. Our CFG should still make sense and properly
+## place this instruction in a new basic block.
   jmp  .lbb2
 .lbb1:  je  .lexit
 .lbb2:
@@ -60,7 +60,7 @@ JUMP_TABLE:
   .quad .lbb2
   .quad hotpath
 
-# No basic blocks above should have 4 successors! That is a bug.
+## No basic blocks above should have 4 successors! That is a bug.
 # CHECK-NOT:   Successors: {{.*}} (mispreds: 0, count: 20), {{.*}} (mispreds: 0, count: 0), {{.*}} (mispreds: 0, count: 0), {{.*}} (mispreds: 0, count: 0)
 # Check successful removal of stray direct jmp
 #  CHECK:  UCE removed 1 block
diff --git a/bolt/test/X86/unreachable.test b/bolt/test/X86/unreachable.test
@@ -1,4 +1,4 @@
-# Check unreachable code elimination
+## Check unreachable code elimination
 
 RUN: %clang %cflags %p/../Inputs/stub.c -fPIC -pie -shared -o %t.so
 RUN: llvm-mc -filetype=obj -triple x86_64-unknown-unknown \

diff --git a/bolt/test/X86/vararg.test b/bolt/test/X86/vararg.test
@@ -1,6 +1,6 @@
-# Check that a function that references a label inside itself,
-# as in the case of vararg handling code generated by GCC 4.5
-# and earlier, is recognized as multi-entry.
+## Check that a function that references a label inside itself,
+## as in the case of vararg handling code generated by GCC 4.5
+## and earlier, is recognized as multi-entry.
 
 REQUIRES: x86_64-linux
 

diff --git a/bolt/test/X86/yaml-multiple-profiles.test b/bolt/test/X86/yaml-multiple-profiles.test
@@ -1,5 +1,5 @@
-# This test ensures that a YAML profile with multiple profiles matching the same
-# function is handled gracefully.
+## This test ensures that a YAML profile with multiple profiles matching the same
+## function is handled gracefully.
 
 # REQUIRES: system-linux
 # RUN: split-file %s %t

diff --git a/bolt/test/X86/zero-sized-object.s b/bolt/test/X86/zero-sized-object.s
@@ -1,5 +1,5 @@
-# Check that references to local (unnamed) objects below are not
-# treated as references relative to zero-sized A object.
+## Check that references to local (unnamed) objects below are not
+## treated as references relative to zero-sized A object.
 
 # REQUIRES: system-linux
 

diff --git a/bolt/test/bad-exe.test b/bolt/test/bad-exe.test
@@ -1,8 +1,8 @@
-# Check that llvm-bolt rejects input that is not a valid ELF executable
-# bzip2.debuginfo is the result of running "objcopy --only-keep-debug".
+## Check that llvm-bolt rejects input that is not a valid ELF executable
+## bzip2.debuginfo is the result of running "objcopy --only-keep-debug".
 
-# This test uses the clang driver without target flags and will only succeed
-# on Linux systems where the host triple matches the target.
+## This test uses the clang driver without target flags and will only succeed
+## on Linux systems where the host triple matches the target.
 REQUIRES: system-linux
 
 RUN: %clang %cflags %S/Inputs/icf-jump-tables.c -g -o %t

diff --git a/bolt/test/bolt-icf.test b/bolt/test/bolt-icf.test
@@ -1,4 +1,4 @@
-# Check for the replacement of calls to identical functions.
+## Check for the replacement of calls to identical functions.
 
 REQUIRES: system-linux
 

diff --git a/bolt/test/bolt-info.test b/bolt/test/bolt-info.test
@@ -1,7 +1,7 @@
-# Check that the .bolt_info section is generated properly.
+## Check that the .bolt_info section is generated properly.
 
-# This test uses the clang driver without target flags and will only succeed
-# on Linux systems where the host triple matches the target.
+## This test uses the clang driver without target flags and will only succeed
+## on Linux systems where the host triple matches the target.
 REQUIRES: system-linux
 
 RUN: %clang %cflags %S/Inputs/icf-jump-tables.c -o %t

diff --git a/bolt/test/heatmap.test b/bolt/test/heatmap.test
@@ -1,4 +1,4 @@
-# Verifies basic functioning of heatmap mode
+## Verifies basic functioning of heatmap mode
 
 REQUIRES: system-linux
 

diff --git a/bolt/test/invalid-profile.test b/bolt/test/invalid-profile.test
@@ -1,7 +1,7 @@
-# Check that llvm-bolt detects bad profile data and aborts
+## Check that llvm-bolt detects bad profile data and aborts
 
-# This test uses the clang driver without target flags and will only succeed
-# on Linux systems where the host triple matches the target.
+## This test uses the clang driver without target flags and will only succeed
+## on Linux systems where the host triple matches the target.
 REQUIRES: system-linux
 
 RUN: %clang %S/Inputs/icf-jump-tables.c -o %t

diff --git a/bolt/test/keep-aranges.test b/bolt/test/keep-aranges.test
@@ -1,5 +1,5 @@
-# Check that BOLT generates .debug_aranges section for an input
-# where it was removed when .gdb_index was generated.
+## Check that BOLT generates .debug_aranges section for an input
+## where it was removed when .gdb_index was generated.
 
 REQUIRES: system-linux
 

diff --git a/bolt/test/lit.local.cfg b/bolt/test/lit.local.cfg
@@ -1,4 +1,4 @@
-host_linux_triple = config.target_triple.split("-")[0] + "-linux"
+host_linux_triple = config.target_triple.split("-")[0] + "-unknown-linux-gnu"
 common_linker_flags = "-fuse-ld=lld -Wl,--unresolved-symbols=ignore-all"
 flags = f"--target={host_linux_triple} {common_linker_flags}"
 

diff --git a/bolt/test/lsda-section-name.cpp b/bolt/test/lsda-section-name.cpp
@@ -2,10 +2,10 @@
 // disassembled by BOLT.
 
 // RUN: %clang++ %cxxflags -O3 -no-pie -c %s -o %t.o
-// RUN: %clang++ %cxxflags -no-pie -fuse-ld=lld %t.o -o %t.exe \
-// RUN:   -Wl,-q -Wl,--script=%S/Inputs/lsda.ldscript
-// RUN: llvm-readelf -SW %t.exe | FileCheck %s
-// RUN: llvm-bolt %t.exe -o %t.bolt
+// RUN: %clang++ %cxxflags -O3 -no-pie -fuse-ld=lld %t.o -o %t
+// RUN: llvm-objcopy --rename-section .gcc_except_table=.gcc_except_table.main %t
+// RUN: llvm-readelf -SW %t | FileCheck %s
+// RUN: llvm-bolt %t -o %t.bolt
 
 // CHECK: .gcc_except_table.main
 

diff --git a/bolt/test/no-relocs.test b/bolt/test/no-relocs.test
@@ -1,7 +1,7 @@
-# Verifies that input without relocations is rejected in relocs mode.
+## Verifies that input without relocations is rejected in relocs mode.
 
-# This test uses the clang driver without target flags and will only succeed
-# on Linux systems where the host triple matches the target.
+## This test uses the clang driver without target flags and will only succeed
+## on Linux systems where the host triple matches the target.
 REQUIRES: system-linux
 
 RUN: %clang %cflags %S/Inputs/icf-jump-tables.c -o %t

diff --git a/bolt/test/non-empty-debug-line.test b/bolt/test/non-empty-debug-line.test
@@ -1,5 +1,5 @@
-# Verifies that BOLT emits DWARF line table with the same size if
-# no functions with debug info were modified.
+## Verifies that BOLT emits DWARF line table with the same size if
+## no functions with debug info were modified.
 
 REQUIRES: system-linux
 
@@ -9,12 +9,12 @@ RUN: llvm-readobj -S %t > %t2
 RUN: llvm-readobj -S %t1 >> %t2
 RUN: FileCheck %s --input-file %t2
 
-# Check the input and grab .debug_line size.
+## Check the input and grab .debug_line size.
 CHECK: File:
 CHECK: Name: .debug_line
 CHECK: Size: [[SIZE:[0-9]+]]
 
-# Verify .debug_line size is the same after BOLT.
+## Verify .debug_line size is the same after BOLT.
 CHECK: File:
 CHECK: Name: .debug_line
 CHECK: Size:

diff --git a/bolt/test/pie.test b/bolt/test/pie.test
@@ -1,7 +1,7 @@
-# Check that we do not reject position-independent executables (PIEs).
+## Check that we do not reject position-independent executables (PIEs).
 
-# This test uses the clang driver without target flags and will only succeed
-# on Linux systems where the host triple matches the target.
+## This test uses the clang driver without target flags and will only succeed
+## on Linux systems where the host triple matches the target.
 REQUIRES: system-linux
 
 RUN: %clang %cflags -fPIC -pie %p/Inputs/jump_table_icp.cpp -o %t

diff --git a/bolt/test/re-optimize.test b/bolt/test/re-optimize.test
@@ -1,7 +1,7 @@
-# Check that we detect re-optimization attempt.
+## Check that we detect re-optimization attempt.
 
-# This test uses the clang driver without target flags and will only succeed
-# on Linux systems where the host triple matches the target.
+## This test uses the clang driver without target flags and will only succeed
+## on Linux systems where the host triple matches the target.
 REQUIRES: system-linux
 
 RUN: %clang %cflags %S/Inputs/icf-jump-tables.c -o %t.exe

diff --git a/bolt/test/runtime/X86/asm-dump.c b/bolt/test/runtime/X86/asm-dump.c
@@ -1,5 +1,5 @@
 /**
- * Test for asm-dump functionality.
+ ** Test for asm-dump functionality.
  *
  * REQUIRES: x86_64-linux,bolt-runtime
  *

diff --git a/bolt/test/shared-object.test b/bolt/test/shared-object.test
@@ -1,7 +1,7 @@
-# Test that llvm-bolt processes *.so without a failure
+## Test that llvm-bolt processes *.so without a failure
 
-# This test uses the clang driver without target flags and will only succeed
-# on Linux systems where the host triple matches the target.
+## This test uses the clang driver without target flags and will only succeed
+## on Linux systems where the host triple matches the target.
 REQUIRES: system-linux
 
 RUN: %clang %cflags %S/Inputs/icf-jump-tables.c -o %t.so -shared -fPIC -Wl,--build-id

diff --git a/clang-tools-extra/clang-tidy/modernize/UseStdFormatCheck.cpp b/clang-tools-extra/clang-tidy/modernize/UseStdFormatCheck.cpp
@@ -47,13 +47,15 @@ void UseStdFormatCheck::registerPPCallbacks(const SourceManager &SM,
 }
 
 void UseStdFormatCheck::registerMatchers(MatchFinder *Finder) {
+  auto CharPointerType =
+      hasType(pointerType(pointee(matchers::isSimpleChar())));
   Finder->addMatcher(
-      callExpr(argumentCountAtLeast(1),
-               hasArgument(0, stringLiteral(isOrdinary())),
-               callee(functionDecl(unless(cxxMethodDecl()),
-                                   matchers::matchesAnyListedName(
-                                       StrFormatLikeFunctions))
-                          .bind("func_decl")))
+      callExpr(
+          argumentCountAtLeast(1), hasArgument(0, stringLiteral(isOrdinary())),
+          callee(functionDecl(
+                     unless(cxxMethodDecl()), hasParameter(0, CharPointerType),
+                     matchers::matchesAnyListedName(StrFormatLikeFunctions))
+                     .bind("func_decl")))
           .bind("strformat"),
       this);
 }

diff --git a/clang-tools-extra/clang-tidy/modernize/UseStdPrintCheck.cpp b/clang-tools-extra/clang-tidy/modernize/UseStdPrintCheck.cpp
@@ -95,12 +95,15 @@ unusedReturnValue(clang::ast_matchers::StatementMatcher MatchedCallExpr) {
 }
 
 void UseStdPrintCheck::registerMatchers(MatchFinder *Finder) {
+  auto CharPointerType =
+      hasType(pointerType(pointee(matchers::isSimpleChar())));
   if (!PrintfLikeFunctions.empty())
     Finder->addMatcher(
         unusedReturnValue(
             callExpr(argumentCountAtLeast(1),
                      hasArgument(0, stringLiteral(isOrdinary())),
                      callee(functionDecl(unless(cxxMethodDecl()),
+                                         hasParameter(0, CharPointerType),
                                          matchers::matchesAnyListedName(
                                              PrintfLikeFunctions))
                                 .bind("func_decl")))
@@ -113,6 +116,7 @@ void UseStdPrintCheck::registerMatchers(MatchFinder *Finder) {
             callExpr(argumentCountAtLeast(2),
                      hasArgument(1, stringLiteral(isOrdinary())),
                      callee(functionDecl(unless(cxxMethodDecl()),
+                                         hasParameter(1, CharPointerType),
                                          matchers::matchesAnyListedName(
                                              FprintfLikeFunctions))
                                 .bind("func_decl")))

diff --git a/clang-tools-extra/clang-tidy/readability/ContainerSizeEmptyCheck.cpp b/clang-tools-extra/clang-tidy/readability/ContainerSizeEmptyCheck.cpp
@@ -150,6 +150,7 @@ void ContainerSizeEmptyCheck::registerMatchers(MatchFinder *Finder) {
 
   Finder->addMatcher(
       cxxMemberCallExpr(
+          argumentCountIs(0),
           on(expr(anyOf(hasType(ValidContainer),
                         hasType(pointsTo(ValidContainer)),
                         hasType(references(ValidContainer))))
@@ -163,7 +164,8 @@ void ContainerSizeEmptyCheck::registerMatchers(MatchFinder *Finder) {
       this);
 
   Finder->addMatcher(
-      callExpr(has(cxxDependentScopeMemberExpr(
+      callExpr(argumentCountIs(0),
+               has(cxxDependentScopeMemberExpr(
                        hasObjectExpression(
                            expr(anyOf(hasType(ValidContainer),
                                       hasType(pointsTo(ValidContainer)),

diff --git a/clang-tools-extra/clang-tidy/utils/FormatStringConverter.cpp b/clang-tools-extra/clang-tidy/utils/FormatStringConverter.cpp
@@ -208,9 +208,11 @@ FormatStringConverter::FormatStringConverter(ASTContext *ContextIn,
   assert(ArgsOffset <= NumArgs);
   FormatExpr = llvm::dyn_cast<StringLiteral>(
       Args[FormatArgOffset]->IgnoreImplicitAsWritten());
-  assert(FormatExpr);
-  if (!FormatExpr->isOrdinary())
-    return; // No wide string support yet
+  if (!FormatExpr || !FormatExpr->isOrdinary()) {
+    // Function must have a narrow string literal as its first argument.
+    conversionNotPossible("first argument is not a narrow string literal");
+    return;
+  }
   PrintfFormatString = FormatExpr->getString();
 
   // Assume that the output will be approximately the same size as the input,

diff --git a/clang-tools-extra/clang-tidy/utils/Matchers.h b/clang-tools-extra/clang-tidy/utils/Matchers.h
@@ -49,6 +49,14 @@ AST_MATCHER_FUNCTION(ast_matchers::TypeMatcher, isPointerToConst) {
   return pointerType(pointee(qualType(isConstQualified())));
 }
 
+// Returns QualType matcher for target char type only.
+AST_MATCHER(QualType, isSimpleChar) {
+  const auto ActualType = Node.getTypePtr();
+  return ActualType &&
+         (ActualType->isSpecificBuiltinType(BuiltinType::Char_S) ||
+          ActualType->isSpecificBuiltinType(BuiltinType::Char_U));
+}
+
 AST_MATCHER(Expr, hasUnevaluatedContext) {
   if (isa<CXXNoexceptExpr>(Node) || isa<RequiresExpr>(Node))
     return true;

diff --git a/clang-tools-extra/clangd/refactor/tweaks/DefineOutline.cpp b/clang-tools-extra/clangd/refactor/tweaks/DefineOutline.cpp
@@ -179,14 +179,11 @@ deleteTokensWithKind(const syntax::TokenBuffer &TokBuf, tok::TokenKind Kind,
 // looked up in the context containing the function/method.
 // FIXME: Drop attributes in function signature.
 llvm::Expected<std::string>
-getFunctionSourceCode(const FunctionDecl *FD, llvm::StringRef TargetNamespace,
+getFunctionSourceCode(const FunctionDecl *FD, const DeclContext *TargetContext,
                       const syntax::TokenBuffer &TokBuf,
                       const HeuristicResolver *Resolver) {
   auto &AST = FD->getASTContext();
   auto &SM = AST.getSourceManager();
-  auto TargetContext = findContextForNS(TargetNamespace, FD->getDeclContext());
-  if (!TargetContext)
-    return error("define outline: couldn't find a context for target");
 
   llvm::Error Errors = llvm::Error::success();
   tooling::Replacements DeclarationCleanups;
@@ -216,7 +213,7 @@ getFunctionSourceCode(const FunctionDecl *FD, llvm::StringRef TargetNamespace,
         }
         const NamedDecl *ND = Ref.Targets.front();
         const std::string Qualifier =
-            getQualification(AST, *TargetContext,
+            getQualification(AST, TargetContext,
                              SM.getLocForStartOfFile(SM.getMainFileID()), ND);
         if (auto Err = DeclarationCleanups.add(
                 tooling::Replacement(SM, Ref.NameLoc, 0, Qualifier)))
@@ -232,7 +229,7 @@ getFunctionSourceCode(const FunctionDecl *FD, llvm::StringRef TargetNamespace,
   if (const auto *Destructor = llvm::dyn_cast<CXXDestructorDecl>(FD)) {
     if (auto Err = DeclarationCleanups.add(tooling::Replacement(
             SM, Destructor->getLocation(), 0,
-            getQualification(AST, *TargetContext,
+            getQualification(AST, TargetContext,
                              SM.getLocForStartOfFile(SM.getMainFileID()),
                              Destructor))))
       Errors = llvm::joinErrors(std::move(Errors), std::move(Err));
@@ -319,29 +316,9 @@ getFunctionSourceCode(const FunctionDecl *FD, llvm::StringRef TargetNamespace,
 }
 
 struct InsertionPoint {
-  std::string EnclosingNamespace;
+  const DeclContext *EnclosingNamespace = nullptr;
   size_t Offset;
 };
-// Returns the most natural insertion point for \p QualifiedName in \p Contents.
-// This currently cares about only the namespace proximity, but in feature it
-// should also try to follow ordering of declarations. For example, if decls
-// come in order `foo, bar, baz` then this function should return some point
-// between foo and baz for inserting bar.
-llvm::Expected<InsertionPoint> getInsertionPoint(llvm::StringRef Contents,
-                                                 llvm::StringRef QualifiedName,
-                                                 const LangOptions &LangOpts) {
-  auto Region = getEligiblePoints(Contents, QualifiedName, LangOpts);
-
-  assert(!Region.EligiblePoints.empty());
-  // FIXME: This selection can be made smarter by looking at the definition
-  // locations for adjacent decls to Source. Unfortunately pseudo parsing in
-  // getEligibleRegions only knows about namespace begin/end events so we
-  // can't match function start/end positions yet.
-  auto Offset = positionToOffset(Contents, Region.EligiblePoints.back());
-  if (!Offset)
-    return Offset.takeError();
-  return InsertionPoint{Region.EnclosingNamespace, *Offset};
-}
 
 // Returns the range that should be deleted from declaration, which always
 // contains function body. In addition to that it might contain constructor
@@ -409,14 +386,9 @@ class DefineOutline : public Tweak {
   }
 
   bool prepare(const Selection &Sel) override {
-    // Bail out if we are not in a header file.
-    // FIXME: We might want to consider moving method definitions below class
-    // definition even if we are inside a source file.
-    if (!isHeaderFile(Sel.AST->getSourceManager().getFilename(Sel.Cursor),
-                      Sel.AST->getLangOpts()))
-      return false;
-
+    SameFile = !isHeaderFile(Sel.AST->tuPath(), Sel.AST->getLangOpts());
     Source = getSelectedFunction(Sel.ASTSelection.commonAncestor());
+
     // Bail out if the selection is not a in-line function definition.
     if (!Source || !Source->doesThisDeclarationHaveABody() ||
         Source->isOutOfLine())
@@ -429,19 +401,24 @@ class DefineOutline : public Tweak {
     if (Source->getTemplateSpecializationInfo())
       return false;
 
-    if (auto *MD = llvm::dyn_cast<CXXMethodDecl>(Source)) {
-      // Bail out in templated classes, as it is hard to spell the class name,
-      // i.e if the template parameter is unnamed.
-      if (MD->getParent()->isTemplated())
-        return false;
-
-      // The refactoring is meaningless for unnamed classes and definitions
-      // within unnamed namespaces.
-      for (const DeclContext *DC = MD->getParent(); DC; DC = DC->getParent()) {
-        if (auto *ND = llvm::dyn_cast<NamedDecl>(DC)) {
-          if (ND->getDeclName().isEmpty())
-            return false;
-        }
+    auto *MD = llvm::dyn_cast<CXXMethodDecl>(Source);
+    if (!MD) {
+      // Can't outline free-standing functions in the same file.
+      return !SameFile;
+    }
+
+    // Bail out in templated classes, as it is hard to spell the class name,
+    // i.e if the template parameter is unnamed.
+    if (MD->getParent()->isTemplated())
+      return false;
+
+    // The refactoring is meaningless for unnamed classes and namespaces,
+    // unless we're outlining in the same file
+    for (const DeclContext *DC = MD->getParent(); DC; DC = DC->getParent()) {
+      if (auto *ND = llvm::dyn_cast<NamedDecl>(DC)) {
+        if (ND->getDeclName().isEmpty() &&
+            (!SameFile || !llvm::dyn_cast<NamespaceDecl>(ND)))
+          return false;
       }
     }
 
@@ -453,8 +430,8 @@ class DefineOutline : public Tweak {
 
   Expected<Effect> apply(const Selection &Sel) override {
     const SourceManager &SM = Sel.AST->getSourceManager();
-    auto CCFile = getSourceFile(Sel.AST->tuPath(), Sel);
-
+    auto CCFile = SameFile ? Sel.AST->tuPath().str()
+                           : getSourceFile(Sel.AST->tuPath(), Sel);
     if (!CCFile)
       return error("Couldn't find a suitable implementation file.");
     assert(Sel.FS && "FS Must be set in apply");
@@ -464,8 +441,7 @@ class DefineOutline : public Tweak {
     if (!Buffer)
       return llvm::errorCodeToError(Buffer.getError());
     auto Contents = Buffer->get()->getBuffer();
-    auto InsertionPoint = getInsertionPoint(
-        Contents, Source->getQualifiedNameAsString(), Sel.AST->getLangOpts());
+    auto InsertionPoint = getInsertionPoint(Contents, Sel);
     if (!InsertionPoint)
       return InsertionPoint.takeError();
 
@@ -499,17 +475,77 @@ class DefineOutline : public Tweak {
       HeaderUpdates = HeaderUpdates.merge(*DelInline);
     }
 
-    auto HeaderFE = Effect::fileEdit(SM, SM.getMainFileID(), HeaderUpdates);
-    if (!HeaderFE)
-      return HeaderFE.takeError();
-
-    Effect->ApplyEdits.try_emplace(HeaderFE->first,
-                                   std::move(HeaderFE->second));
+    if (SameFile) {
+      tooling::Replacements &R = Effect->ApplyEdits[*CCFile].Replacements;
+      R = R.merge(HeaderUpdates);
+    } else {
+      auto HeaderFE = Effect::fileEdit(SM, SM.getMainFileID(), HeaderUpdates);
+      if (!HeaderFE)
+        return HeaderFE.takeError();
+      Effect->ApplyEdits.try_emplace(HeaderFE->first,
+                                     std::move(HeaderFE->second));
+    }
     return std::move(*Effect);
   }
 
+  // Returns the most natural insertion point for \p QualifiedName in \p
+  // Contents. This currently cares about only the namespace proximity, but in
+  // feature it should also try to follow ordering of declarations. For example,
+  // if decls come in order `foo, bar, baz` then this function should return
+  // some point between foo and baz for inserting bar.
+  // FIXME: The selection can be made smarter by looking at the definition
+  // locations for adjacent decls to Source. Unfortunately pseudo parsing in
+  // getEligibleRegions only knows about namespace begin/end events so we
+  // can't match function start/end positions yet.
+  llvm::Expected<InsertionPoint> getInsertionPoint(llvm::StringRef Contents,
+                                                   const Selection &Sel) {
+    // If the definition goes to the same file and there is a namespace,
+    // we should (and, in the case of anonymous namespaces, need to)
+    // put the definition into the original namespace block.
+    if (SameFile) {
+      auto *Klass = Source->getDeclContext()->getOuterLexicalRecordContext();
+      if (!Klass)
+        return error("moving to same file not supported for free functions");
+      const SourceLocation EndLoc = Klass->getBraceRange().getEnd();
+      const auto &TokBuf = Sel.AST->getTokens();
+      auto Tokens = TokBuf.expandedTokens();
+      auto It = llvm::lower_bound(
+          Tokens, EndLoc, [](const syntax::Token &Tok, SourceLocation EndLoc) {
+            return Tok.location() < EndLoc;
+          });
+      while (It != Tokens.end()) {
+        if (It->kind() != tok::semi) {
+          ++It;
+          continue;
+        }
+        unsigned Offset = Sel.AST->getSourceManager()
+                              .getDecomposedLoc(It->endLocation())
+                              .second;
+        return InsertionPoint{Klass->getEnclosingNamespaceContext(), Offset};
+      }
+      return error(
+          "failed to determine insertion location: no end of class found");
+    }
+
+    auto Region = getEligiblePoints(
+        Contents, Source->getQualifiedNameAsString(), Sel.AST->getLangOpts());
+
+    assert(!Region.EligiblePoints.empty());
+    auto Offset = positionToOffset(Contents, Region.EligiblePoints.back());
+    if (!Offset)
+      return Offset.takeError();
+
+    auto TargetContext =
+        findContextForNS(Region.EnclosingNamespace, Source->getDeclContext());
+    if (!TargetContext)
+      return error("define outline: couldn't find a context for target");
+
+    return InsertionPoint{*TargetContext, *Offset};
+  }
+
 private:
   const FunctionDecl *Source = nullptr;
+  bool SameFile = false;
 };
 
 REGISTER_TWEAK(DefineOutline)

diff --git a/clang-tools-extra/clangd/unittests/ReplayPeambleTests.cpp b/clang-tools-extra/clangd/unittests/ReplayPeambleTests.cpp
@@ -25,7 +25,6 @@
 #include "clang/AST/DeclTemplate.h"
 #include "clang/Basic/FileEntry.h"
 #include "clang/Basic/LLVM.h"
-#include "clang/Basic/Module.h"
 #include "clang/Basic/SourceLocation.h"
 #include "clang/Basic/SourceManager.h"
 #include "clang/Basic/TokenKinds.h"
@@ -42,7 +41,11 @@
 #include <memory>
 #include <vector>
 
-namespace clang::clangd {
+namespace clang {
+
+class Module;
+
+namespace clangd {
 namespace {
 struct Inclusion {
   Inclusion(const SourceManager &SM, SourceLocation HashLoc,
@@ -170,4 +173,5 @@ TEST(ReplayPreambleTest, IncludesAndSkippedFiles) {
   }
 }
 } // namespace
-} // namespace clang::clangd
+} // namespace clangd
+} // namespace clang
diff --git a/clang-tools-extra/clangd/unittests/tweaks/DefineOutlineTests.cpp b/clang-tools-extra/clangd/unittests/tweaks/DefineOutlineTests.cpp
@@ -19,12 +19,47 @@ TWEAK_TEST(DefineOutline);
 
 TEST_F(DefineOutlineTest, TriggersOnFunctionDecl) {
   FileName = "Test.cpp";
-  // Not available unless in a header file.
+  // Not available for free function unless in a header file.
   EXPECT_UNAVAILABLE(R"cpp(
     [[void [[f^o^o]]() [[{
       return;
     }]]]])cpp");
 
+  // Available in soure file.
+  EXPECT_AVAILABLE(R"cpp(
+    struct Foo {
+      void f^oo() {}
+    };
+  )cpp");
+
+  // Available within named namespace in source file.
+  EXPECT_AVAILABLE(R"cpp(
+    namespace N {
+      struct Foo {
+        void f^oo() {}
+      };
+    } // namespace N
+  )cpp");
+
+  // Available within anonymous namespace in source file.
+  EXPECT_AVAILABLE(R"cpp(
+    namespace {
+      struct Foo {
+        void f^oo() {}
+      };
+    } // namespace
+  )cpp");
+
+  // Not available for out-of-line method.
+  EXPECT_UNAVAILABLE(R"cpp(
+    class Bar {
+      void baz();
+    };
+
+    [[void [[Bar::[[b^a^z]]]]() [[{
+      return;
+    }]]]])cpp");
+
   FileName = "Test.hpp";
   // Not available unless function name or fully body is selected.
   EXPECT_UNAVAILABLE(R"cpp(
@@ -100,7 +135,7 @@ TEST_F(DefineOutlineTest, TriggersOnFunctionDecl) {
     };
   )cpp");
 
-  // Not available on definitions within unnamed namespaces
+  // Not available on definitions in header file within unnamed namespaces
   EXPECT_UNAVAILABLE(R"cpp(
     namespace {
       struct Foo {
@@ -349,6 +384,40 @@ TEST_F(DefineOutlineTest, ApplyTest) {
   }
 }
 
+TEST_F(DefineOutlineTest, InCppFile) {
+  FileName = "Test.cpp";
+
+  struct {
+    llvm::StringRef Test;
+    llvm::StringRef ExpectedSource;
+  } Cases[] = {
+      {
+          R"cpp(
+            namespace foo {
+            namespace {
+            struct Foo { void ba^r() {} };
+            struct Bar { void foo(); };
+            void Bar::foo() {}
+            }
+            }
+        )cpp",
+          R"cpp(
+            namespace foo {
+            namespace {
+            struct Foo { void bar() ; };void Foo::bar() {} 
+            struct Bar { void foo(); };
+            void Bar::foo() {}
+            }
+            }
+        )cpp"},
+  };
+
+  for (const auto &Case : Cases) {
+    SCOPED_TRACE(Case.Test);
+    EXPECT_EQ(apply(Case.Test, nullptr), Case.ExpectedSource);
+  }
+}
+
 TEST_F(DefineOutlineTest, HandleMacros) {
   llvm::StringMap<std::string> EditedFiles;
   ExtraFiles["Test.cpp"] = "";