Merge branch 'main' into ci-container-non-root

Author: boomanaiden154

.ci/monolithic-linux.sh

@@ -34,8 +34,11 @@ function at-exit {
 
   # If building fails there will be no results files.
   shopt -s nullglob
-  python3 "${MONOREPO_ROOT}"/.ci/generate_test_report.py ":linux: Linux x64 Test Results" \
-    "linux-x64-test-results" "${BUILD_DIR}"/test-results.*.xml
+  if command -v buildkite-agent 2>&1 >/dev/null
+  then
+    python3 "${MONOREPO_ROOT}"/.ci/generate_test_report.py ":linux: Linux x64 Test Results" \
+      "linux-x64-test-results" "${BUILD_DIR}"/test-results.*.xml
+  fi
 }
 trap at-exit EXIT
 

.ci/monolithic-windows.sh

@@ -33,8 +33,11 @@ function at-exit {
 
   # If building fails there will be no results files.
   shopt -s nullglob
-  python "${MONOREPO_ROOT}"/.ci/generate_test_report.py ":windows: Windows x64 Test Results" \
-    "windows-x64-test-results" "${BUILD_DIR}"/test-results.*.xml
+  if command -v buildkite-agent 2>&1 >/dev/null
+  then
+    python "${MONOREPO_ROOT}"/.ci/generate_test_report.py ":windows: Windows x64 Test Results" \
+      "windows-x64-test-results" "${BUILD_DIR}"/test-results.*.xml
+  fi
 }
 trap at-exit EXIT
 

.git-blame-ignore-revs

@@ -94,3 +94,6 @@ b6262880b34629e9d7a72b5a42f315a3c9ed8139
 39c7dc7207e76e72da21cf4fedda21b5311bf62d
 e80bc777749331e9519575f416c342f7626dd14d
 7e5cd8f1b6c5263ed5e2cc03d60c8779a8d3e9f7
+
+# NFC: clang-format test_demangle.pass.cpp but keep test "lines"
+d33bf2e9df578ff7e44fd22504d6ad5a122b7ee6

.github/workflows/containers/github-action-ci/Dockerfile

@@ -41,13 +41,13 @@ RUN ninja -C ./build stage2-clang-bolt stage2-install-distribution && ninja -C .
 FROM base
 
 COPY --from=stage1-toolchain $LLVM_SYSROOT $LLVM_SYSROOT
-    
+
 # Need to install curl for hendrikmuhs/ccache-action
 # Need nodejs for some of the GitHub actions.
 # Need perl-modules for clang analyzer tests.
 # Need git for SPIRV-Tools tests.
 RUN apt-get update && \
-    apt-get install -y \
+    DEBIAN_FRONTEND=noninteractive apt-get install -y \
     binutils \ 
     cmake \
     curl \
@@ -56,7 +56,15 @@ RUN apt-get update && \
     ninja-build \
     nodejs \
     perl-modules \
-    python3-psutil
+    python3-psutil \
+
+    # These are needed by the premerge pipeline. Pip is used to install
+    # dependent python packages and ccache is used for build caching. File and
+    # tzdata are used for tests.
+    python3-pip \
+    ccache \
+    file \
+    tzdata
 
 ENV LLVM_SYSROOT=$LLVM_SYSROOT
 ENV PATH=${LLVM_SYSROOT}/bin:${PATH}

bolt/include/bolt/Passes/ADRRelaxationPass.h

@@ -25,7 +25,8 @@ namespace bolt {
 
 class ADRRelaxationPass : public BinaryFunctionPass {
 public:
-  explicit ADRRelaxationPass() : BinaryFunctionPass(false) {}
+  explicit ADRRelaxationPass(const cl::opt<bool> &PrintPass)
+      : BinaryFunctionPass(PrintPass) {}
 
   const char *getName() const override { return "adr-relaxation"; }
 

bolt/lib/Passes/LongJmp.cpp

@@ -138,9 +138,9 @@ BinaryBasicBlock *LongJmpPass::lookupStubFromGroup(
           const std::pair<uint64_t, BinaryBasicBlock *> &RHS) {
         return LHS.first < RHS.first;
       });
-  if (Cand == Candidates.end())
-    return nullptr;
-  if (Cand != Candidates.begin()) {
+  if (Cand == Candidates.end()) {
+    Cand = std::prev(Cand);
+  } else if (Cand != Candidates.begin()) {
     const StubTy *LeftCand = std::prev(Cand);
     if (Cand->first - DotAddress > DotAddress - LeftCand->first)
       Cand = LeftCand;

bolt/lib/Rewrite/BinaryPassManager.cpp

@@ -126,6 +126,11 @@ static cl::opt<bool> PrintJTFootprintReduction(
     cl::desc("print function after jt-footprint-reduction pass"), cl::Hidden,
     cl::cat(BoltOptCategory));
 
+static cl::opt<bool>
+    PrintAdrRelaxation("print-adr-relaxation",
+                       cl::desc("print functions after ADR Relaxation pass"),
+                       cl::Hidden, cl::cat(BoltOptCategory));
+
 static cl::opt<bool>
     PrintLongJmp("print-longjmp",
                  cl::desc("print functions after longjmp pass"), cl::Hidden,
@@ -493,7 +498,8 @@ Error BinaryFunctionPassManager::runAllPasses(BinaryContext &BC) {
   Manager.registerPass(std::make_unique<ReorderData>());
 
   if (BC.isAArch64()) {
-    Manager.registerPass(std::make_unique<ADRRelaxationPass>());
+    Manager.registerPass(
+        std::make_unique<ADRRelaxationPass>(PrintAdrRelaxation));
 
     // Tighten branches according to offset differences between branch and
     // targets. No extra instructions after this pass, otherwise we may have

bolt/test/AArch64/long-jmp-one-stub.s

@@ -0,0 +1,32 @@
+## This test verifies that no unnecessary stubs are inserted when each DotAddress increases during a lookup.
+
+# REQUIRES: system-linux, asserts
+
+# RUN: llvm-mc -filetype=obj -triple aarch64-unknown-unknown %s -o %t.o
+# RUN: %clang %cflags -O0 %t.o -o %t.exe -Wl,-q
+# RUN: link_fdata %s %t.o %t.fdata
+# RUN: llvm-bolt %t.exe -o %t.bolt  \
+# RUN:   --data %t.fdata  | FileCheck %s
+
+# CHECK: BOLT-INFO: Inserted 1 stubs in the hot area and 0 stubs in the cold area. 
+
+  .section .text
+  .global _start
+  .global far_away_func
+
+  .align 4
+  .global _start
+  .type _start, %function
+_start:
+# FDATA: 0 [unknown] 0 1 _start 0 0 100
+    bl far_away_func
+    bl far_away_func
+    ret  
+  .space 0x8000000
+  .global far_away_func
+  .type far_away_func, %function
+far_away_func:
+    add x0, x0, #1
+    ret
+
+.reloc 0, R_AARCH64_NONE

bolt/test/X86/dwarf5-debug-names-gnu-push-tls-address.s

@@ -27,71 +27,35 @@
 ##   x = 10;
 ##   return x;
 ## }
-	.text
 	.file	"main.cpp"
 	.file	0 "gnu_tls_push" "main.cpp" md5 0x551db97d5e23dc6a81abdc5ade4d9d71
-	.globl	main                            # -- Begin function main
-	.p2align	4, 0x90
+	.globl	main
 	.type	main,@function
-main:                                   # @main
+main:
 .Lfunc_begin0:
-	.loc	0 2 0                           # main.cpp:2:0
-	.cfi_startproc
-# %bb.0:                                # %entry
-	pushq	%rbp
-	.cfi_def_cfa_offset 16
-	.cfi_offset %rbp, -16
-	movq	%rsp, %rbp
-	.cfi_def_cfa_register %rbp
-	movl	$0, -4(%rbp)
-.Ltmp0:
-	.loc	0 3 3 prologue_end              # main.cpp:3:3
-	movq	%fs:0, %rax
-	leaq	x@TPOFF(%rax), %rax
-	.loc	0 3 5 is_stmt 0                 # main.cpp:3:5
-	movl	$10, (%rax)
-	.loc	0 4 10 is_stmt 1                # main.cpp:4:10
-	movq	%fs:0, %rax
-	leaq	x@TPOFF(%rax), %rax
-	movl	(%rax), %eax
-	.loc	0 4 3 epilogue_begin is_stmt 0  # main.cpp:4:3
-	popq	%rbp
-	.cfi_def_cfa %rsp, 8
+	.loc	0 2 0
+	.loc	0 3 3 prologue_end
+	.loc	0 3 5 is_stmt 0
+	.loc	0 4 10 is_stmt 1
+	.loc	0 4 3 epilogue_begin is_stmt 0
 	retq
-.Ltmp1:
 .Lfunc_end0:
 	.size	main, .Lfunc_end0-main
-	.cfi_endproc
-                                        # -- End function
-	.section	.text._ZTW1x,"axG",@progbits,_ZTW1x,comdat
-	.hidden	_ZTW1x                          # -- Begin function _ZTW1x
+
+	.hidden	_ZTW1x
 	.weak	_ZTW1x
-	.p2align	4, 0x90
 	.type	_ZTW1x,@function
-_ZTW1x:                                 # @_ZTW1x
+_ZTW1x:
 .Lfunc_begin1:
-	.cfi_startproc
-# %bb.0:
-	pushq	%rbp
-	.cfi_def_cfa_offset 16
-	.cfi_offset %rbp, -16
-	movq	%rsp, %rbp
-	.cfi_def_cfa_register %rbp
-	movq	%fs:0, %rax
-	leaq	x@TPOFF(%rax), %rax
-	popq	%rbp
-	.cfi_def_cfa %rsp, 8
 	retq
 .Lfunc_end1:
 	.size	_ZTW1x, .Lfunc_end1-_ZTW1x
-	.cfi_endproc
-                                        # -- End function
-	.type	x,@object                       # @x
+
+	.type	x,@object
 	.section	.tbss,"awT",@nobits
 	.globl	x
-	.p2align	2, 0x0
 x:
-	.long	0                               # 0x0
+	.long	0
 	.size	x, 4
 
 	.section	.debug_abbrev,"",@progbits

bolt/test/X86/linux-static-keys.s

@@ -35,13 +35,13 @@ _start:
 .L0:
   jmp L1
 # CHECK:      jit
-# CHECK-SAME: # ID: 1 {{.*}} # Likely: 0 # InitValue: 1
+# CHECK-SAME: # ID: 1 {{.*}} # Likely: 1 # InitValue: 0
   nop
 L1:
   .nops 5
-  jmp .L0
 # CHECK:      jit
-# CHECK-SAME: # ID: 2 {{.*}} # Likely: 1 # InitValue: 1
+# CHECK-SAME: # ID: 2 {{.*}} # Likely: 0 # InitValue: 0
+  jmp .L0
 
 ## Check that a branch profile associated with a NOP is handled properly when
 ## dynamic branch is created.
@@ -67,18 +67,24 @@ foo:
   .type __start___jump_table, %object
 __start___jump_table:
 
-  .long .L0 - . # Jump address
-  .long L1 - . # Target address
-  .quad 1       # Key address
+  .long .L0 - .                   # Jump address
+  .long L1 - .                    # Target address
+  .quad fake_static_key + 1 - .   # Key address; LSB = 1 : likely
 
-  .long L1 - . # Jump address
-  .long L2 - . # Target address
-  .quad 0       # Key address
+  .long L1 - .                    # Jump address
+  .long L2 - .                    # Target address
+  .quad fake_static_key -.        # Key address; LSB = 0 : unlikely
 
   .globl __stop___jump_table
   .type __stop___jump_table, %object
 __stop___jump_table:
 
+## Staic keys (we just use the label ignoring the format of the keys).
+  .data
+  .align 8
+fake_static_key:
+  .quad 0
+
 ## Fake Linux Kernel sections.
   .section __ksymtab,"a",@progbits
   .section __ksymtab_gpl,"a",@progbits