Add missing guard on CPUTYPE

• Add endian

In D36680#832953, @cy wrote:

Next question. I maintain the extract of the FreeBSD telnet/telnetd on GH, of which telnetd is a port while telnet is in the repo should we choose to remove it as well. I've been toying with the idea of telnet and telnetd ports based on NetBSD. They will probably pay more attention to it than we might. Thoughts?

In D36680#832950, @cy wrote:

There's also a fix for CVE-2020-10188 (https://github.com/cschuber/freebsd-telnet/commit/e76b51e91f6d9aa7b72ee6624c29b46ddce2a406). Do you want to commit that or should I?

Oops. Thanks for the fix. I'll try to remember to remove ATM when I finish removing the (useless) support.

Bump doc date, use .Fx macro

Seems like a good change. We don't need usrstack to be a pointer on CheriBSD and having it be one is a least-privilege violation. We don't support an executable stack for the main thread and there isn't really any reason to do so for ABIs that don't have backwards compatibility concerns (the main reason to do so is to support nested functions and IIRC gcc now supports nested functions without exec stacks, but doing so is an ABI break).

GitHub recently started to parse this by default so I'd guess they are confident it's no a significant performance impact. Anywhere you'd have thousands of entries you're also parsing a significant portion of the pack files so parsing one text file into some sort of hash structure seems unlikely to be a big deal.

In D36421#827216, @jrtc27 wrote:

For example, i386 targets will be:
MACHINE_ARCH= hard-float long32 ptr62 time32

s/ptr62/ptr32/ and s/ARCH/ABI/

I'd be willing to entertain another name like MACHINE_ABI_PROPS to avoid the need for D36420, but we use MACHINE_ABI extensively in CheriBSD and I'm not convinced anyone uses the current variable.

• Assert wording tweak from jhb

• Use "Fl -debug" markup
• Bump doc date
• Rebase

Skimmed the diff and it looks fine. It's hard to imaging an error that would compile, but not work.

• Fix typo and add a little more documentation of what debug output does.

This is good. As @emaste points out, we'll be able to delete this soon and attempts to bring code that does this into the tree will fail to compile.

Works for me. The statement on compilers might be a touch strong as its only mid 2022 and later versions that do this.

One comment, but would be happy to see this land as is.

LTGM. Thanks.

It looks like we use LC_COLLATE=C sort elsewhere in this file.

I wrote this because it's too hard to tell what's going on when some "enterprise" middle box is corrupting fetches. (We've seen this from two sites so far when people are trying to bootstrap CheriBSD packages and getting blobs of html back instead of their intended downloads.) You can't even see what URLs are being downloaded in order to start debugging and the error "No trusted fingerprint found matching package's certificate" isn't informative.

Fix the name of DTrace