• Show the last callout on every CPU by default

The change looks ok to me.

I hope to commit some ATF tests to the tcptestsuite repo. I’ll send out a new review to update the port after that.

Sorry I wasn't clearer: use virtqueue_size() after the allocate to do the check. Do not add a new virtqueue_size bus interface.

• Clamp seg_max after allocating virtqueues.

• revert to previous approach of clamping seg_max
• do not artificially limit seg_max when indirect is available

Mark: Sorry, I @-mentioned you by mistake.

• Simply detect the case and fail to attach, as @bryanv suggests.

• add declaration of vtmmio_virtqueue_size()

• add comment for foo - 2

Abandoned in favor of https://github.com/freebsd-net/tcp-testsuite/pull/4#

If you add the tests directly in GH, feel free to request a review from me.

In D20610#445425, @asomers wrote:

The basic idea behind this review is fundamentally broken. We shouldn't be adding test cases to the base system that wrap a port which is itself just a bunch of test cases. Instead, the correct approach would be to add ATF test cases directly to the port, to be installed into /usr/local/tests/tcptestsuite. See devel/atf, devel/kyua, or devel/lutok for examples of ports that do that.

Does this need urgent attention to get in 11.3?

Wait, I think we should still have a seat belt in guests.

If you change the config seg_max in bhyve to be VTSCSI_MAXSEG - 2, does that fix the issue?

I'll try that now.

Fix it in bhyve instead

I believe seg_max is the number of segments in a SCSI command while...

• Accept short URLs, too.

In D19426#417642, @emaste wrote:

I would be comfortable with simply "https://bugs.freebsd.org/12344" as that can likely work forever.

Note that https://bugs.freebsd.org/12344 does work - we should accept either https://bugs.freebsd.org/bugzilla/show_bug.cgi?id= or https://bugs.freebsd.org/[0-9]+

In D19459#417484, @kristof wrote:

https://bugs.freebsd.org/12345 redirects to https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=12345. Wouldn't the first form be better? It's shorter and likely easier to map to whatever bug tracker we might move to after bugzilla (if we ever do change). It also looks more like the phabricator links.

Just mention the URL in the comments.

Just mention the URL in the comments.

To keep the whole record in one place, here is the CGI script I mentioned:

In D19426#415919, @bdrewery wrote:

Check out rP154065.
All of the PRs linked there were filed in GNATS but were mapped over in BUGZILLA so they had the same numbers.
Had they been GNATS urls they would be more annoying to deal with IMO as the links would all be dead now. I realize this is an objective opinion.

$Id$ -> $FreeBSD$

In D19426#415915, @rgrimes wrote:

Missing is the change to the commit message template to update it to say URL instead of PRxxxx

D19459 covers subversion in the base system.

• add sigaltstack SS_ONSTACK test
• fix more style
• MFH