tests: don't run atf_* in a subshell

[glebius]

Summary

Shell limitation is that a classic function call via $() is a subshell
and atf-sh(3) commands won't work as epxected there. Subsequently,
atf_skip inside a function won't skip a test. The test will fail later.

A working approach is to pass desired variable name as argument to
a function and don't run subshell.

Fixes: ea82362219ee715cfbb195b2114e73fdc8599fa5

[glebius]

This sweep covers all regressions in ea82362219ee, which are limited to
check for /dev/mdctl. An alternative fix would be to add to Makefiles
something like:

TEST_METADATA.test1 += require.files="/dev/mdctl"

It is up to test experts to decide whether to go with this change or an alternative. What I like about this sweep is that it provides an example on how to run shell functions when you are in atf-sh(3).

[ngie]

Would you please add a comment to the tops of functions to note what variables are being set in the functions?

[glebius]

In D42646#973650, @ngie wrote:

Would you please add a comment to the tops of functions to note what variables are being set in the functions?

Of course I can! Let me ask question first: do you agree that this patch is worth checking in and I can add you as "reviewed by"?

P.S. I actually opened this revision again to add you as reviewer. I need review from somebody who is a general expert in the tests.

[des]

IMO the TAP tests should be rewritten. We wouldn't happen to have a handy student or intern available?

[ngie]

In D42646#973954, @glebius wrote:

In D42646#973650, @ngie wrote:

Would you please add a comment to the tops of functions to note what variables are being set in the functions?

Of course I can! Let me ask question first: do you agree that this patch is worth checking in and I can add you as "reviewed by"?

P.S. I actually opened this revision again to add you as reviewer. I need review from somebody who is a general expert in the tests.

I'm totally fine with that :).

It's interesting because I didn't think this was an optimization worth making, but I don't know the scenarios that Netflix is dealing with where this test suite might not scale in the scenario you're addressing here.

[glebius]

In D42646#974984, @ngie wrote:

It's interesting because I didn't think this was an optimization worth making, but I don't know the scenarios that Netflix is dealing with where this test suite might not scale in the scenario you're addressing here.

There are no special scenario at Netflix. I hit that with my own testing. The point is that ea82362219ee715cfbb195b2114e73fdc8599fa5 doesn't work at all as intended. You can't atf_skip from a subshell. The test will fail instead of being skipped.

Getting back to your original comment:

Would you please add a comment to the tops of functions to note what variables are being set in the functions?

Do you want me to amend with a comment every alloc_md()? Note that the function doesn't set any particular variable. It sets the variable that was passed as the first argument.