Re-apply r249644: Handle inline stacks in gcov-encoded sample profiles.

This fixes memory allocation problems by making the merge operation keep
the profile readers around until the merged profile has been emitted.
This is needed to prevent the inlined function names to disappear from
the function profiles. Since all the names are kept as references, once
the reader disappears, the names are also deallocated.

Additionally, XFAIL on big-endian architectures. The test case uses a
gcov file generated on a little-endian system.

llvm-svn: 249724

2 files changed, 29 insertions, 0 deletions

diff --git a/llvm/test/tools/llvm-profdata/Inputs/gcc-sample-profile.gcov b/llvm/test/tools/llvm-profdata/Inputs/gcc-sample-profile.gcov
new file mode 100644
index 00000000000..0099280af2f
--- /dev/null
+++ b/llvm/test/tools/llvm-profdata/Inputs/gcc-sample-profile.gcov
diff --git a/llvm/test/tools/llvm-profdata/gcc-gcov-sample-profile.test b/llvm/test/tools/llvm-profdata/gcc-gcov-sample-profile.test
new file mode 100644
index 00000000000..723d547555b
--- /dev/null
+++ b/llvm/test/tools/llvm-profdata/gcc-gcov-sample-profile.test
@@ -0,0 +1,29 @@
+The input gcov file has been generated on a little endian machine. Expect
+failures on big endian systems.
+
+XFAIL: powerpc64-, s390x, mips-, mips64-, sparc
+
+Tests for sample profiles encoded in GCC's gcov format.
+
+1- Show all functions. This profile has a single main() function with several
+   inlined callees.
+RUN: llvm-profdata show --sample %p/Inputs/gcc-sample-profile.gcov | FileCheck %s --check-prefix=SHOW1
+SHOW1: Function: main: 364084, 0, 6 sampled lines
+SHOW1: line offset: 2, discriminator: 3, inlined callee: _Z3fool: 243786, 0, 3 sampled lines
+SHOW1:   line offset: 1, discriminator: 3, inlined callee: _Z3bari: 0, 0, 2 sampled lines
+SHOW1:   line offset: 1, discriminator: 8, inlined callee: _Z3bari: 0, 0, 2 sampled lines
+SHOW1:   line offset: 1, discriminator: 7, inlined callee: _Z3bari: 98558, 0, 2 sampled lines
+
+2- Convert the profile to text encoding and check that they are both
+   identical.
+RUN: llvm-profdata merge --sample %p/Inputs/gcc-sample-profile.gcov --text -o - | llvm-profdata show --sample - -o %t-text
+RUN: llvm-profdata show --sample %p/Inputs/gcc-sample-profile.gcov -o %t-gcov
+RUN: diff %t-text %t-gcov
+
+4- Merge the gcov and text encodings of the profile and check that the
+   counters have doubled.
+RUN: llvm-profdata merge --sample --text %p/Inputs/gcc-sample-profile.gcov -o %t-gcov
+RUN: llvm-profdata merge --sample --text %p/Inputs/gcc-sample-profile.gcov %t-gcov -o - | FileCheck %s --check-prefix=MERGE1
+MERGE1: main:728168:0
+MERGE1: 2.3: 120298
+MERGE1: 2.3: _Z3fool:487572