Fix build with KASAN

[rincebrain]

Motivation and Context

Presumably since the zstd commit landed, no longer builds with KASAN.

Description

The stock zstd code expects some helpers from ASAN if ADDRESS_SANITIZER
is defined. This works fine in userland, but in kernel, KASAN also
results in that getting defined, but does not provide those helpers. So let's make some empty
substitutes for that case.

How Has This Been Tested?

I ran through ZTS with this applied on a kernel on Debian buster with KASAN (though not without failures; that's another day).
I also built it on a kernel without KASAN to make certain the defines seemed right.

Types of changes

• Bug fix (non-breaking change which fixes an issue)
• New feature (non-breaking change which adds functionality)
• Performance enhancement (non-breaking change which improves efficiency)
• Code cleanup (non-breaking change which makes code smaller or more readable)
• Breaking change (fix or feature that would cause existing functionality to change)
• Library ABI change (libzfs, libzfs_core, libnvpair, libuutil and libzfsbootenv)
• Documentation (a change to man pages or other documentation)

Checklist:

• My code follows the OpenZFS code style requirements.
• I have updated the documentation accordingly.
• I have read the contributing document.
• I have added tests to cover my changes.
• I have run the ZFS Test Suite with this change applied.
• All commit messages are properly formatted and contain Signed-off-by.

[rincebrain]

Fascinating, am I just blind, and didn't notice the build reporting problems, or did something strange ensue?

I had occasion to rollback my KASAN VM, reapply this patch to master, and rebuild, and lo, the symbols were missing, because ADDRESS_SANITIZER is a define from earlier in the zstd library code.

How this built (and loaded, and ran through ZTS) for me before will remain a mystery to me, since I did do the prodding I documented in the testing section above. Perhaps I didn't clobber the installed old modules?

Anyway, I just looted the test from lib/zstd.c, in the interest of not winding up in a strange position where it gets defined one place and not the other, or the obnoxiously long #if defined(...) && (defined(...) || defined(...)) that would result from collapsing it.

To ensure no mysterious excitement again, I did a completely clean checkout, applied this patch, built it, verified it wasn't already loaded, and loaded the zzstd module, on my KASAN system. No risk of fun this time.