[llvm] [CGData] llvm-cgdata (PR #89884)

[Ellis Hoag via llvm-commits]

================
@@ -0,0 +1,75 @@
+# Merge an archive that has two object files having cgdata (__llvm_outline)
+
+RUN: split-file %s %t
+
+RUN: llc -filetype=obj -mtriple arm64-apple-darwin %t/merge-1.ll -o %t/merge-1.o
+RUN: llc -filetype=obj -mtriple arm64-apple-darwin %t/merge-2.ll -o %t/merge-2.o
+RUN: llvm-ar rcs %t/merge-archive.a %t/merge-1.o %t/merge-2.o
+RUN: llvm-cgdata merge %t/merge-archive.a -o %t/merge-archive.cgdata
+RUN: llvm-cgdata show %t/merge-archive.cgdata | FileCheck %s
+CHECK: Outlined hash tree:
+CHECK-NEXT:  Total Node Count: 4
+CHECK-NEXT:  Terminal Node Count: 2
+CHECK-NEXT:  Depth: 2
+
+RUN: llvm-cgdata dump %t/merge-archive.cgdata | FileCheck %s --check-prefix TREE
+TREE: # Outlined stable hash tree
+TREE-NEXT: :outlined_hash_tree
+TREE-NEXT: ---
+TREE-NEXT: 0:
+TREE-NEXT:   Hash:            0x0
+TREE-NEXT:   Terminals:       0
+TREE-NEXT:   SuccessorIds:    [ 1 ]
+TREE-NEXT: 1:
+TREE-NEXT:   Hash:            0x1
+TREE-NEXT:   Terminals:       0
+TREE-NEXT:   SuccessorIds:    [ 2, 3 ]
+TREE-NEXT: 2:
+TREE-NEXT:   Hash:            0x3
+TREE-NEXT:   Terminals:       5
+TREE-NEXT:   SuccessorIds:    [  ]
+TREE-NEXT: 3:
+TREE-NEXT:   Hash:            0x2
+TREE-NEXT:   Terminals:       4
+TREE-NEXT:   SuccessorIds:    [  ]
+TREE-NEXT: ...
+
+;--- merge-1.ll
+
+; The .data is encoded in a binary form based on the following yaml form. See serialize() in OutlinedHashTreeRecord.cpp
+;---
+;0:
+;  Hash:            0x0
+;  Terminals:       0
+;  SuccessorIds:    [ 1 ]
+;1:
+;  Hash:            0x1
+;  Terminals:       0
+;  SuccessorIds:    [ 2 ]
+;2:
+;  Hash:            0x2
+;  Terminals:       4
+;  SuccessorIds:    [  ]
+;...
+
+ at .data = private unnamed_addr constant [72 x i8] c"\03\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\01\00\00\00\01\00\00\00\01\00\00\00\01\00\00\00\00\00\00\00\00\00\00\00\01\00\00\00\02\00\00\00\02\00\00\00\02\00\00\00\00\00\00\00\04\00\00\00\00\00\00\00", section "__DATA,__llvm_outline"
----------------
ellishg wrote:

How exactly is the data here generated? I'm trying to think if there is a way to avoid inlining the raw data here. One trick is to use `sed` to replace a keyword in `merge-1.ll`.

```
...
c"<RAW_DATA>", section "__DATA,__llvm_outline"
...
```
Then if `CMD` is some command to generate the `\03\00\00...` data, you could do something like this:
```
; RUN: sed -i 's/<RAW_DATA>/$(CMD)/' %t/merge-1.ll
```
I'm not sure if LIT can handle subshells, or how exactly to escape, but I bet it could work :)

https://github.com/llvm/llvm-project/pull/89884