[clang] [Clang] Implement labelled type filtering for overflow/truncation sanitizers w/ SSCLs (PR #107332)

[Vitaly Buka via cfe-commits]

vitalybuka wrote:

> Right, while I want to go full instrumentation, it's just not going to happen in the Linux kernel. We're going to need both options so that we can slowly overlap coverage until we've squeezed out all the unexpected wrap-around.

That was just my 2c. Linux decision about is need needed to proceed here.
The patch LGTM.

I believe it could be better without `no_sanitize`. Please share if some one else have opinion on that?
Also let's make that with what ever with  no_wrap/no_sanitize consistent.

BTW, what is the plan for wrap? Is this going to be this list or another one?
Does Kernel instrument in recovery mode or always crash?
In recoverable mode sanitizer/no_sanitize is independent from wrap/no_wrap. 
In non-recovery mode wrap is not needed for sanitized types




https://github.com/llvm/llvm-project/pull/107332