Fix Rocky Linux 8 ASAN plugin tests by suppressing dynamic linker leaks

[bryancall]

Problem

Three unit tests were failing on Rocky Linux 8 with ASAN enabled in CI:

• test_PluginDso
• test_PluginFactory
• test_RemapPluginInfo

The tests were passing all their assertions, but LeakSanitizer was reporting false positive memory leaks from dlopen calls, causing the tests to fail.

Root Cause

On Rocky Linux 8, ASAN's instrumentation of the dynamic linker (ld-linux-x86-64.so) causes it to report leaks during plugin initialization via dlopen, even though the code properly calls dlclose() for each handle.

The leak backtrace shows:

#0  operator new
#1  <unknown module>  (plugin_v1.so - global init)
#2  <unknown module>  (plugin_v1.so - init_array)
#3  <unknown module>  (plugin_v1.so - init)
#4  ld-linux-x86-64.so.2  (dynamic linker - call_init)
...
#16 PluginDso::load()

Why Existing Suppressions Didn't Work

The suppression file already had leak:PluginDso::load, but it wasn't catching these leaks because:

• LSAN uses fast unwinding by default, which only captures ~5 stack frames
• PluginDso::load is at frame Fix list templating #16, so it's not visible in the fast unwind
• Using fast_unwind_on_malloc=0 would allow the existing suppression to work, but has significant performance overhead. I also had problems with the test segfaulting with fast_unwind_on_malloc=0.

Solution

Added leak:ld-linux-x86-64.so to the suppression file, which catches the leak at frame #4 where the dynamic linker appears in the shallow stack trace.

Testing

Tested in Rocky Linux 8 Docker container with ASAN enabled:

• ✅ All three tests now pass
• ✅ No performance impact from slow unwinding
• ✅ Tests still properly validate plugin loading/unloading logic

Fixes CI failures: https://ci.trafficserver.apache.org/job/master/job/os_build/46636/console

[bryancall]

Another option is to just compile these unit tests with fast_unwind_on_malloc=0. I will look into the performance differences.

[Copilot]

Pull Request Overview

This PR fixes false-positive memory leak failures in three Rocky Linux 8 unit tests (test_PluginDso, test_PluginFactory, and test_RemapPluginInfo) when running with ASAN enabled. The issue stems from ASAN's fast unwinding not capturing deep enough stack frames to match existing suppressions. The fix adds a targeted suppression for the dynamic linker itself.

Key Changes:

• Adds leak:ld-linux-x86-64.so suppression to catch leaks at the dynamic linker level
• Reformats existing leak:call_init suppression (moved to its own line)
• Documents the Rocky Linux 8-specific issue with inline comments

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[Copilot]

The suppression pattern leak:ld-linux-x86-64.so is architecture-specific (x86-64) and may not work on other architectures like ARM64 or i386. Consider using a more portable pattern like leak:ld-linux or documenting that this suppression is x86-64 specific. If Rocky Linux 8 is only supported on x86-64 in CI, this may be acceptable.

Suggested change

	leak:ld-linux-x86-64.so
	leak:ld-linux

[bryancall]

It is better to narrow the scope and only fix CI that is x86.

[bryancall]

[approve ci]

[bryancall]

I tried using fast_unwind_on_malloc=0 and it didn't go well. I was getting segfaults and I wasn't getting the symbol names back.

[bryancall]

I was able to fix the segfaults, it was a dependency issue that other cmake targets to be build fist. However, with rockylinux 8 we would need to use an external symbolizer to convert the address numbers to symbol names. So the best approach is the modify the suppression file, like this PR does.

[bryancall]

[approve ci autest 0]