Address jtl's comment making the diff much smaller.

In D18520#394737, @cem wrote:

2. adding the CBC-MAC soft implementations (AES_CBC_MAC_foo, "ccm-cbc.c"). I don't think there's any real reason to call it ccm-cbc.c instead of cbc-mac, since it is a independent algorithm in its own right.
3. adding the CCM soft implementations (I'm not really clear where these are or how this works; the exf struction points to aes_icm for everything except reinit).
4. adding CCM/CBC to cryptodev along with some basic tests

I suggest to wait for 2019Q1, then:

I have some qualms with the API of the generic interface. I don't think most things care about the time step, and it might even make sense to scale it by something like architecture (as a proxy for CPU performance); it should be left out of the API. Absolute timeout should be optional or in a separate interface. And the condition is needlessly complicated, at least in the gmirror case changed (double negative).

Also keep a copy of loader.efi as loader-old.efi in case there's a problem with the new one.

In D18519#396278, @mat wrote:

If having an invalid value in LICENSE_FILE does not make your test builds fails, you may have DISABLE_LICENSES set somewhere.

I was also unable to reproduce @lev's issue, so will abandon this review.

Check if there's space to keep BOOTxx.efi around, and if so rename is to BOOTxx-old.efi.

Address latest review comments

I think this is reasonable - certainly an error is better than creating a corrupt filesystem.

Comparing DoS benches results with the previous Diff version applied and this one didn't show difference.

Sorry for the duplicate revision. I'm having curl problems.

added >0 to a DEPENDS
setprop on pkg-descr and distinfo

I am confused by zero tests run

ran:
svn propset fbsd:nokeywords yes distinfo pkg-descr
svn propset svn:eol-style native distinfo pkg-descr
svn propset svn:mime-type text/plain distinfo pkg-descr

I'll fix those properties next. 'make test' is progressing.

Fix DISTVERSION.

In D18516#396189, @dumbbell wrote:

Hi!

I suppose the int2float() function comes from one of the removed file, is that right? If yes, could you please add a comment indicating the initial source filename?

If having an invalid value in LICENSE_FILE does not make your test builds fails, you may have DISABLE_LICENSES set somewhere.

• Add comment indicating the origin of int2float().

Well, you should not use Phabricator as a place to dump the patches you are working on, use git and branches or something.

Move devel/py-mongoengine to databases/py-mongoengine
Update dependent ports

Assuming you've tested it and it it's good, then it's good, thank you :-)

Changes to fix issues found by devel/hs-ShellCheck .

Current status:

Sure, any number here is better than 0. 60 or 300. I don't care.

Improve extmem tests.