1## Check a common case for BOLT address translation tables. These tables are used 2## to translate profile activity happening in a bolted binary back to the 3## original binary, so you can run BOLT again, with updated profile collected 4## in a production environment that only runs bolted binaries. As BOLT only 5## takes no-bolt binaries as inputs, this translation is necessary to cover 6## this scenario. 7# 8# RUN: yaml2obj %p/Inputs/blarge.yaml &> %t.exe 9# RUN: llvm-bolt %t.exe -o %t.out --data %p/Inputs/blarge.fdata \ 10# RUN: --reorder-blocks=normal --split-functions --enable-bat 2>&1 | FileCheck %s 11# RUN: llvm-bat-dump %t.out --dump-all \ 12# RUN: --translate=0x401180 | FileCheck %s --check-prefix=CHECK-BAT-DUMP 13# 14## In this test we focus on function usqrt at address 0x401170. This is a 15## non-reloc binary case, so we don't expect this address to change, that's 16## why we hardcode its address here. This address also comes hardcoded in the 17## blarge.yaml input file. 18## 19## This is the layout of the function before BOLT reorder blocks: 20## 21## BB Layout : .LBB02, .Ltmp39, .LFT1, .Ltmp38, .LFT2 22## 23## This is the layout of the function after BOLT reorder blocks: 24## 25## BB Layout : .LBB02, .Ltmp38, .Ltmp39, .LFT2, .LFT3 26## 27## .Ltmp38 is originally at offset 0x39 but gets moved to 0xc (see full dump 28## below). 29## 30## We check that BAT is able to translate references happening in .Ltmp38 to 31## its original offset. 32## 33 34## This binary has 3 functions with profile, all of them are split, so 6 maps. 35## BAT creates one map per function fragment. 36# 37# CHECK: BOLT: 3 out of 7 functions were overwritten. 38# CHECK: BOLT-INFO: Wrote 6 BAT maps 39# CHECK: BOLT-INFO: Wrote 3 function and 58 basic block hashes 40# CHECK: BOLT-INFO: BAT section size (bytes): 940 41# 42# usqrt mappings (hot part). We match against any key (left side containing 43# the bolted binary offsets) because BOLT may change where it puts instructions 44# depending on whether it is relaxing a branch or not. But the original input 45# binary offsets (right side) should be the same because these addresses are 46# hardcoded in the blarge.yaml file. 47# 48# CHECK-BAT-DUMP: Function Address: 0x401170, hash: 0xace6cbc638b31983 49# CHECK-BAT-DUMP-NEXT: BB mappings: 50# CHECK-BAT-DUMP-NEXT: 0x0 -> 0x0 hash: 0x36007ba1d80c0000 51# CHECK-BAT-DUMP-NEXT: 0x8 -> 0x8 (branch) 52# CHECK-BAT-DUMP-NEXT: 0x{{.*}} -> 0x39 hash: 0x5c06705524800039 53# CHECK-BAT-DUMP-NEXT: 0x{{.*}} -> 0x3d (branch) 54# CHECK-BAT-DUMP-NEXT: 0x{{.*}} -> 0x10 hash: 0xd70d7a64320e0010 55# CHECK-BAT-DUMP-NEXT: 0x{{.*}} -> 0x30 (branch) 56# 57# CHECK-BAT-DUMP: 3 cold mappings 58# 59# Now check that the translation 0x401180 maps back to its correct 60# input offset (offset 3d in the usqrt input function). 61# 62# COM: CHECK-BAT-DUMP: Translating addresses according to parsed BAT tables: 63# CHECK-BAT-DUMP: 0x401180 -> usqrt + 0x3d 64 65# ------------------------- 66# Full dump for reference (this is not checked): 67# ------------------------- 68 69Binary Function "usqrt" after finalize-functions 70 Number : 7 71 State : CFG finalized 72 Address : 0x401170 73 Size : 0x43 74 MaxSize : 0x43 75 Offset : 0xcb0 76 Section : .text 77 Orc Section : .local.text.usqrt 78 LSDA : 0x0 79 IsSimple : 1 80 IsMultiEntry: 0 81 IsSplit : 1 82 BB Count : 5 83 Hash : a6468f132ec176ca 84 BB Layout : .LBB02, .Ltmp38, .Ltmp39, .LFT2, .LFT3 85 Exec Count : 199 86 Profile Acc : 100.0% 87 88.LBB02 (4 instructions, align : 1) 89 Entry Point 90 Exec Count : 199 91 CFI State : 0 92 Input offset: 0 93 00000000: movl $0x20, %r8d 94 00000006: xorl %eax, %eax 95 00000008: xorl %edx, %edx # Offset: 8 96 0000000a: jmp .Ltmp39 97 Successors: .Ltmp39 (mispreds: 0, count: 0) 98 99.Ltmp38 (2 instructions, align : 1) 100 Exec Count : 4711 101 CFI State : 0 102 Input offset: 39 103 Predecessors: .Ltmp39, .LFT2 104 0000000c: subl $0x1, %r8d 105 00000010: je .LFT3 # Offset: 61 106 Successors: .LFT3 (mispreds: 0, count: 0), .Ltmp39 (mispreds: 33, count: 4711) 107 108.Ltmp39 (10 instructions, align : 1) 109 Exec Count : 4711 110 CFI State : 0 111 Input offset: 10 112 Predecessors: .Ltmp38, .LBB02 113 00000012: movq %rdi, %rcx 114 00000015: addq %rax, %rax 115 00000018: shlq $0x2, %rdi 116 0000001c: andl $0xc0000000, %ecx 117 00000022: shrq $0x1e, %rcx 118 00000026: leaq (%rcx,%rdx,4), %rdx 119 0000002a: leaq 0x1(%rax,%rax), %rcx 120 0000002f: cmpq %rcx, %rdx 121 00000032: jb .Ltmp38 # Offset: 48 122 00000034: jmp .LFT2 123 Successors: .Ltmp38 (mispreds: 171, count: 2886), .LFT2 (mispreds: 0, count: 0) 124 125------- HOT-COLD SPLIT POINT ------- 126 127.LFT2 (3 instructions, align : 1) 128 Exec Count : 0 129 CFI State : 0 130 Input offset: 32 131 Predecessors: .Ltmp39 132 00000036: subq %rcx, %rdx 133 00000039: addq $0x1, %rax # Offset: 53 134 0000003d: jmp .Ltmp38 135 Successors: .Ltmp38 (mispreds: 0, count: 0) 136 137.LFT3 (2 instructions, align : 1) 138 Exec Count : 0 139 CFI State : 0 140 Input offset: 3f 141 Predecessors: .Ltmp38 142 0000003f: movq %rax, (%rsi) 143 00000042: retq # Offset: 66 144 145DWARF CFI Instructions: 146 <empty> 147End of Function "usqrt" 148