Add CMake options for turning on ASAN and HWASAN

[BenWibking]

PR Summary

Add CMake options ENABLE_ASAN and ENABLE_HWASAN to build with AddressSanitizer and HWAddressSanitizer, respectively, turned on.

ASAN is supported in both clang and gcc. HWASAN is support by clang and (in theory) gcc.

Then the app is run as normal. It will stop and print an error if a memory bug is detected. This should detect almost all out of bounds access, uninitalized values, and other memory errors for CPU builds.

Usage notes:

• Miscellaneous useful env vars to configure its behavior when running:

  ASAN_OPTIONS=abort_on_error=1:fast_unwind_on_malloc=1:detect_leaks=0 UBSAN_OPTIONS=print_stacktrace=0 LSAN_OPTIONS=suppressions=leak_suppress.txt
  

• HWAddressSanitizer only works on Linux ARM platforms but runs significantly faster than AddressSanitizer. In my testing, additional flags are required for HWAsan:

  -DCMAKE_CXX_FLAGS="-fPIC" -DCMAKE_EXE_LINKER_FLAGS="-fuse-ld=lld" -DCMAKE_SHARED_LINKER_FLAGS="-fuse-ld=lld" -DCMAKE_C_COMPILER=clang -DCMAKE_CXX_COMPILER=clang++
  

  (I was not able to get it working with gcc, since building HWASAN support in gcc is optional and distros don't do it.)

PR Checklist

• Code passes cpplint
• New features are documented.
• Adds a test for any bugs fixed. Adds tests for new features.
• Code is formatted
• Changes are summarized in CHANGELOG.md
• Change is breaking (API, behavior, ...)
  • Change is additionally added to CHANGELOG.md in the breaking section
  • PR is marked as breaking
  • Short summary API changes at the top of the PR (plus optionally with an automated update/fix script)
• CI has been triggered on Darwin for performance regression tests.
• Docs build
• (@lanl.gov employees) Update copyright on changed files

[pgrete]

I think it's a good idea to add those.
How much slower does the code run with those options?
I'm asking wrt to whether it's worth to activate this by default to one of the short or extended tests (once the issue reported in #1107 are fixed).

PS: Changelog entry is missing;)

[pgrete]

This will then also be forwarded to all submodules, we include, isn't it?

[BenWibking]

Yes, it should be.

[lroberts36]

LGTM, this will definitely be helpful to have. I am going to take a look at the logical location bit shift issue, I think it will be straightforward to fix.

[lroberts36]

I'm asking wrt to whether it's worth to activate this by default to one of the short or extended tests (once the issue reported in #1107 are fixed).

I definitely support having it run on one of the extended tests (unless it takes hours to run with these options on).

[BenWibking]

Running the test suite with this turned on is only about a factor of 2-3 slower on my laptop, so I think it should be fine to enable for the extended CI.