[lld] [lld][ELF] Extend profile guided function ordering to ELF binaries (PR #117514)

[Ellis Hoag via llvm-commits]

================
@@ -0,0 +1,269 @@
+# REQUIRES: aarch64
+
+# RUN: rm -rf %t && split-file %s %t && cd %t
+# RUN: llvm-mc -filetype=obj -triple=aarch64 a.s -o a.o
+# RUN: llvm-profdata merge a.proftext -o a.profdata
+# RUN: ld.lld -e main -o a.out a.o --irpgo-profile=a.profdata --bp-startup-sort=function --verbose-bp-section-orderer 2>&1 | FileCheck %s --check-prefix=STARTUP
+# RUN: ld.lld -e main -o a.out a.o --irpgo-profile=a.profdata --bp-startup-sort=function --verbose-bp-section-orderer --icf=all --bp-compression-sort=none 2>&1 | FileCheck %s --check-prefix=STARTUP
+
+# STARTUP: Ordered 3 sections using balanced partitioning
+
+# RUN: ld.lld -e main -o - a.o --irpgo-profile=a.profdata --bp-startup-sort=function --symbol-ordering-file a.orderfile | llvm-nm --numeric-sort --format=just-symbols - | FileCheck %s --check-prefix=ORDERFILE
+
+# ORDERFILE: _Z1Av
+# ORDERFILE: _Z1Fi
+# ORDERFILE: _Z1Ei
+# ORDERFILE: _Z1Di
+# ORDERFILE: _Z1Ci
+# ORDERFILE: _Z1Bi
+# ORDERFILE: main
+# ORDERFILE: r1
+# ORDERFILE: r2
+
+# RUN: ld.lld -e main -o a.out a.o --verbose-bp-section-orderer --bp-compression-sort=function 2>&1 | FileCheck %s --check-prefix=COMPRESSION-FUNC
+# RUN: ld.lld -e main -o a.out a.o --verbose-bp-section-orderer --bp-compression-sort=data 2>&1 | FileCheck %s --check-prefix=COMPRESSION-DATA
+# RUN: ld.lld -e main -o a.out a.o --verbose-bp-section-orderer --bp-compression-sort=both 2>&1 | FileCheck %s --check-prefix=COMPRESSION-BOTH
+# RUN: ld.lld -e main -o a.out a.o --verbose-bp-section-orderer --bp-compression-sort=both --irpgo-profile=a.profdata --bp-startup-sort=function 2>&1 | FileCheck %s --check-prefix=COMPRESSION-BOTH
+
+# COMPRESSION-FUNC: Ordered 7 sections using balanced partitioning
+# COMPRESSION-DATA: Ordered 3 sections using balanced partitioning
+# COMPRESSION-BOTH: Ordered 10 sections using balanced partitioning
+
+#--- a.proftext
+:ir
+:temporal_prof_traces
+# Num Traces
+1
+# Trace Stream Size:
+1
+# Weight
+1
+_Z1Av, _Z1Bi, _Z1Ci
+
+_Z1Av
+# Func Hash:
+1111
+# Num Counters:
+1
+# Counter Values:
+1
+
+_Z1Bi
+# Func Hash:
+2222
+# Num Counters:
+1
+# Counter Values:
+1
+
+_Z1Ci
+# Func Hash:
+3333
+# Num Counters:
+1
+# Counter Values:
+1
+
+_Z1Di
+# Func Hash:
+4444
+# Num Counters:
+1
+# Counter Values:
+1
+
+#--- a.orderfile
+_Z1Av
+_Z1Fi
+_Z1Ei
+_Z1Di
+
+#--- a.cc
+const char s1[] = "hello world";
+const char s2[] = "i am a string";
+const char* r1 = s1;
+const char** r2 = &r1;
+void A() {
+    return;
+}
+
+int B(int a) {
+    A();
+    return a + 1;
+}
+
+int C(int a) {
+    A();
+    return a + 2;
+}
+
+int D(int a) {
+    return B(a + 2);
+}
+
+int E(int a) {
+    return C(a + 2);
+}
+
+int F(int a) {
+    return C(a + 3);
+}
+
+int main() {
+    return 0;
+}
+#--- gen
+echo '#--- a.s'
+clang --target=aarch64-linux-gnu -fdebug-compilation-dir='/proc/self/cwd' -ffunction-sections -fdata-sections -fno-exceptions -fno-rtti -fno-asynchronous-unwind-tables -S a.cc -o -
----------------
ellishg wrote:

We can use `-O0` so that `"i am a string"` doesn't get stripped out.

> gen is invoked with PWD set to /proc/self/cwd. Clang commands don’t need -fdebug-compilation-dir= since its default value is PWD.

https://llvm.org/docs/TestingGuide.html#elaborated-tests

```suggestion
clang --target=aarch64-linux-gnu -O0 -ffunction-sections -fdata-sections -fno-exceptions -fno-rtti -fno-asynchronous-unwind-tables -S a.cc -o -
```

https://github.com/llvm/llvm-project/pull/117514