[lldb][test] check if CoreDumping info is supported

[da-viper]

CoreDumping in /proc/pid/stat was added in kernel 4.15, this causes the test to fail in older kernel versions. see https://man7.org/linux/man-pages/man5/proc_pid_status.5.html

[llvmbot]

@llvm/pr-subscribers-lldb

Author: Ebuka Ezike (da-viper)

Changes

CoreDumping in /proc/pid/stat was added in kernel 4.15, this causes the test to fail in older kernel versions. see https://man7.org/linux/man-pages/man5/proc_pid_status.5.html

---

Full diff: https://github.com/llvm/llvm-project/pull/160333.diff

1 Files Affected:

• (modified) lldb/unittests/Host/posix/HostTest.cpp (+9)

diff --git a/lldb/unittests/Host/posix/HostTest.cpp b/lldb/unittests/Host/posix/HostTest.cpp
index 082edccf4e774..7bb44c99bc998 100644
--- a/lldb/unittests/Host/posix/HostTest.cpp
+++ b/lldb/unittests/Host/posix/HostTest.cpp
@@ -15,6 +15,10 @@
 #include <cerrno>
 #include <sys/resource.h>
 
+#ifdef __linux__
+#include <linux/version.h>
+#endif // __linux__
+
 using namespace lldb_private;
 
 namespace {
@@ -116,7 +120,12 @@ TEST_F(HostTest, GetProcessInfoSetsPriority) {
   ASSERT_TRUE(Info.IsZombie().has_value());
   ASSERT_FALSE(Info.IsZombie().value());
 
+  // CoreDumping was added in kernel version 4.15 
+#if LINUX_VERSION_CODE >= KERNEL_VERSION(4, 15, 0)
   ASSERT_TRUE(Info.IsCoreDumping().has_value());
   ASSERT_FALSE(Info.IsCoreDumping().value());
+#else
+  ASSERT_FALSE(Info.IsCoreDumping().has_value());
+#endif
 }
 #endif

[github-actions]

✅ With the latest revision this PR passed the C/C++ code formatter.

[Jlalond]

CoreDumping in /proc/pid/stat was added in kernel 4.15, this causes the test to fail in older kernel versions. see https://man7.org/linux/man-pages/man5/proc_pid_status.5.html

Note, it's /proc/pid/status not stat for the coredumping field.

[mysterymath]

This appears to be breaking the Fuchsia Linux toolchain LLDB builders: https://luci-milo.appspot.com/ui/p/fuchsia/builders/toolchain.ci/lldb-linux-arm64/b8702712918947567569/overview

IIRC, we ship with Linux headers much older than the running kernel to ensure that we can run on old (as well as newer) kernels. I'd think this kind of runtime behavior checking should be done through a runtime mechanism, not a compile-time one. (Maybe uname?)

[da-viper]

Will update to use runtime check