Add Trusty OS support

[randomPoison]

This PR adds support for the Trusty secure operating system. This upstreams the patch that we have been using internally, cleaned up to meet the project's coding conventions.

This is a revival of #2987.

Relevant headers:

• PROT_READ, PROT_WRITE, and MAP_FAILED come from mman.h.
• CLOCK_BOOTTIME comes from time.h.
• STDOUT_FILENO and STDERR_FILE_NO come from unistd.h.
• AT_PAGESZ comes from elf.h.

[rustbot]

Thanks for the pull request, and welcome! The Rust team is excited to review your changes, and you should hear from @tgross35 (or someone else) some time within the next two weeks.

Please see the contribution instructions for more information. Namely, in order to ensure the minimum review times lag, PR authors and assigned reviewers should ensure that the review label (S-waiting-on-review and S-waiting-on-author) stays updated, invoking these commands when appropriate:

• @rustbot author: the review is finished, PR author should check the comments and take action accordingly
• @rustbot review: the author is ready for a review, this PR will be queued again in the reviewer's queue

[randomPoison]

Addressing a comment from the original PR, @JohnTitor asked:

I wonder if we could move the module under linux_like or somewhere else similar to that. I'm not familiar with Trusty but is it completely different from Linux-like OS (i.e. exposing linux-like items on Trusty is harmful)?

Yeah, Trusty is not linux-like, nor is it a Unix. It has some overlap with those platforms, but isn't part of the same family of systems. I think the right thing to do is have Trusty be its own module.

[randomPoison]

I think this would also be a good candidate for merging over to the 0.2 branch. The diff should apply cleanly, I've tested it against both branches.

@rustbot label stable-nominated

[randomPoison]

What should I do about the unknown target_os = "trusty" cfg in CI? The targets aren't yet in the stable compiler. I've got this Cargo.toml snippet I've been using locally

[lints.rust.unexpected_cfgs]
level = "warn"
check-cfg = [
    'cfg(target_os, values("trusty"))',
]

Is that appropriate to add in the PR as a temporary way to suppress the lint? Or is there a better way to handle this?

[tgross35]

What should I do about the unknown target_os = "trusty" cfg in CI? The targets aren't yet in the stable compiler. I've got this Cargo.toml snippet I've been using locally

We should have that handled in build.rs -

libc/build.rs

Lines 24 to 36 in fbec928

	const CHECK_CFG_EXTRA: &'static [(&'static str, &'static [&'static str])] = &[
	(
	"target_os",
	&[
	"switch", "aix", "ohos", "hurd", "rtems", "visionos", "nuttx",
	],
	),
	("target_env", &["illumos", "wasi", "aix", "ohos"]),
	(
	"target_arch",
	&["loongarch64", "mips32r6", "mips64r6", "csky"],
	),
	];

Overall this looks pretty good to me, please just link to the relevant headers providing these definitions if possible.

@rustbot author (just comment @rustbot review when it is ready for that again)

[randomPoison]

@tgross35 We're using musl as our libc implementation, so those definitions come from the musl headers.

• PROT_READ, PROT_WRITE, and MAP_FAILED come from mman.h.
• CLOCK_BOOTTIME comes from time.h.
• STDOUT_FILENO and STDERR_FILE_NO come from unistd.h.
• AT_PAGESZ comes from elf.h.

Also it looks like a rebase to the latest main fixed the unknown platform error, so I didn't need to update build.rs.

[randomPoison]

oh I forgot to say the magic word

@rustbot review

[tgross35]

Added the headers to the PR description so they make it into git history. Thanks!