In D20411#440660, @kib wrote:

In D20411#440658, @tijl wrote:

In D20411#440580, @mjg wrote:

I think arbitrary mkdir/rmdir is a can of worms, perfectly avoidable for the stated purpose.

Arbitrary symlinks are already supported. Why would mkdir be any different?

This is a bit hacky, even for my standards, so I've come up with this as an alternative: https://reviews.freebsd.org/D20411.

Also call stuff for i386.

In D20333#438430, @dchagin wrote:

linux_common is not used in i386

Move stuff around.

In D20322#438196, @emaste wrote:

I like the idea although have users already put entries in /etc/fstab?

In D20322#438209, @tijl wrote:

We used to tell people to symlink /dev/shm to /tmp in pkg-message but some programs check the file system type of /dev/shm (using statfs) so after https://svnweb.freebsd.org/base?view=revision&revision=283461 /dev/shm must be tmpfs for these programs to work properly and /tmp isn't necessarily tmpfs. So I created /compat/linux/dev/shm in https://svnweb.freebsd.org/ports?view=revision&revision=424237. This also makes /dev/shm and /tmp separate namespaces like they are on Linux. The only problem I've encountered with this is when running 'command > /dev/null' as root it will create /compat/linux/dev/null as a regular file. There's no such problem when running commands as a regular user. I think the proper fix is to have a real /dev/shm directory, either by default or created by linux_common.ko. And then let linux_enable mount a tmpfs there.

In D20116#438201, @emaste wrote:

Functions which create/initialise a Q number.

Is this Q number format defined elsewhere or is it specific to this set of routines?

Drop unneeded line

Even more better ifdefs.

Okay, there's still something wrong. While "kldload linux64" works, the "kldload linux" doesn't due to "link_elf_obj: symbol linux_ptrace_status undefined" error.

One more ifdef, per dchagin's suggestion.

Ifdef out the thing.

Documentation/tests improvements.

That was copy/paste from OpenBSD's commit message, but okay. How about now?

Not sure what do you mean. I think I've marked all __inline functions there?

In D20092#434881, @ngie wrote:

In D20092#434855, @trasz wrote:

In D20092#433990, @ngie wrote:

This looks good to me (not a ports committer though, so just reviewing the change).
As a volunteer/co-maintainer on the LTP project, your suggestions make sense.
The only thing that I would do is install LTP to /compat/linux/opt, if at all possible. If not, we can come up with a better way to build/relocate LTP (I wrote the current build system about a decade ago for the project).

But it does install to /compat/linux/opt - except for the documentation (where it follows Linux conventions) and two binaries - not sure what's going on with those two.

Which 2 binaries? Asking, because I can fix upstream potentially.

Add LICENSE.

The title and description don't match the patch, but apart from that - looks good!

Okay. So... last chance for folks who want to be on the 'Reviewed by' list to mark it as Approved? Thanks :-)

In D20092#433990, @ngie wrote:

This looks good to me (not a ports committer though, so just reviewing the change).
As a volunteer/co-maintainer on the LTP project, your suggestions make sense.
The only thing that I would do is install LTP to /compat/linux/opt, if at all possible. If not, we can come up with a better way to build/relocate LTP (I wrote the current build system about a decade ago for the project).

Have you considered splitting it into two: one for getting rid of DEBUG, and one for preventing out of tree builds? The former looks fine, but I don’t understand the build system well enough wrt the latter.

Fix whitespace.

Like https://reviews.freebsd.org/D20145?

Not sure what you mean?

So, given the dispute above is more about guidelines for Linux packages and less about this port in particular, is there anything left to do before this can be committed?

The kyua(1) man page does reference atf... in a way - it just .Xrs to it, without mentioning it in the text itself. And yeah, the Wiki does include some more detail, but I always view it as kind of additional source - I know I went for the man page first, and simply pointing me in the right direction would spare me some work. Basically - 'man tests' is the first place I'd look for information on tests. And as it is now, it mentions how to run them, but not how to write them.

Add pkg-descr.