It might be worth a comment making clear this is in objdir since bare use of ln is almost always a bug.

I've submitted here because upstream seems to want me to do that (and also to document the issue if someone else tries building FreeBSD with an up-to-date llvm.

Seems sensible. Those are going to be some long lines, but they aren't really meant to be read.

Seems sensible. I guess it's good that most of the issues are in tests or GPL code we're deleting soon?

Seems fine to me, but tagging @imp for further review.

• Honor kern.lognosys if non-zero.

I found this when discussing the implications of D26288.

For reverence the original commit is https://svnweb.freebsd.org/csrg?view=revision&revision=58422

I ran into this on a CHERI system where we've added new revocation related system calls used in malloc. Unfortunately, we call malloc in strdup even before the (rather questionable) badsys hander is installed. Due to someone changing the kernel configs and me not noticing I lost a half day to not having those syscalls...

Add a couple people to potentially comment on what the max size to mmap should actually be.

Given that few changes are required to support a case-insensitive FS I think it's worth doing.

Hmm, why isn't -nobuiltininc -idirafter ${COMPILER_RESOURCE_DIR}/include included here.

Re-indenting should be done in a separate commit if at all.

This seems a bit clunky, but if this is the least disruptive change it seems fine.

I don't have a strong opinion on vendoring those libbsd files.

A few whitespace and comment formatting nits (not every instance marked), but generally looks fine.

• Relocate flags assignment and don't use a variable.

In D25945#575224, @kevans wrote:

In D25945#575221, @brooks wrote:

I'm a little skeptical the #if code makes sense. I suspect it might be a historical artifact.

It's necessary now because of, e.g., freebsd32/syscalls.master; there's an if/else block where we define different variants of the same block of syscalls. If we can get rid of that, I reckon that logic can go away.

I'm a little skeptical the #if code makes sense. I suspect it might be a historical artifact.

This seems fine.

LGTM