Many of our tests are based around performing transfers of sample data
across passt/pasta created links.  The data flow here can be a bit
hard to follow since, e.g. we create a file transfer it to the guest,
then transfer it back to the host across several different tests.
This also means that the test cases aren't independent of each other.

Because we don't have the original file available at both ends in some
cases, we compare them by generating md5sums at each end and comparing
them, which is a bit complicated.

Make a number of changes to simplify this:
  1. Pre-generate the sample data files as a test asset, rather than
     building them on the fly during the tests proper
  2. Include the sample data files in the mbuto guest image
  3. Because we have good copies of the original data available in all
     contexts, we can now simply use 'cmp' to check if the transfer
     has worked, avoiding md5sum complications.
  4. Similarly we can always use the original copy of the sample data
     on the send side of each transfer, meaning that the tests become
     more independent of each other.

Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
Signed-off-by: Stefano Brivio <sbrivio@redhat.com>

We won't necessarily run mbuto as part of regular tests: it can also
be used for demos or out-of-tree tests.

To keep the profile simple, leave the whole sshd setup there, which
is otherwise harmless, but don't fail if guest-key.pub is missing in
the current directory.

Signed-off-by: Stefano Brivio <sbrivio@redhat.com>

Extends the context system in the test scripts to allow executing commands
within a guest.  Do this without requiring an existing network in the guest
by using socat to run ssh via a vsock connection.

We do need some additional "sleep"s in the tests, because the new
faster dispatch means that sometimes we attempt to connect before
socat has managed to listen.

For now, only use this for the plain "passt" tests.  The "passt_in_ns" and
other tests have additional complications we still need to deal with.

Signed-off-by: David Gibson <david@gibson.dropbear.id.au>

Commit 41c02e10 ("tests: Use nmap-ncat instead of openbsd netcat for pasta
tests") updated the pasta tests to use the nmap version of ncat instead of
the openbsd version, for greater portability.

For some upcoming changes, however, we'll be wanting to use socat.
"socat" can do everything "ncat" can and more, so let's move all the
tests using host tools (either directly on the host or via mbuto
generated images) to using socat instead.

Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
[sbrivio: Fix a typo in port specification, 31337 instead of x31337]
Signed-off-by: Stefano Brivio <sbrivio@redhat.com>

Currently our small custom dhclient-script only handles the 'domain-name'
option, which can just list a single domain, not the 'domain-search'
option, which can handle several.  Correct it to handle both.

We also weren't emptying the resolv.conf file before we began, which
could lead to surprising contents after multiple DHCP transactions.

Signed-off-by: David Gibson <david@gibson.dropbear.id.au>

Although it can operate without them, dhclient can issue errors if it
doesn't have /var/run to write a pid file and /var/lib to write a leases
file.  Create those in mbuto.img to stop it complaining.

Signed-off-by: David Gibson <david@gibson.dropbear.id.au>

We now supply a minimal dhclient-script of our own in the mbuto boot image.
There are some problems with it, so add some basic logging to help debug
it.

Signed-off-by: David Gibson <david@gibson.dropbear.id.au>

Modern Fedora (and RHEL) systems have /sbin as a symlink to /usr/sbin
(along with a number of similar links).  Along with that it expects to
find dhclient-script in /usr/sbin/dhclient-script rather than
/sbin/dhclient-script.

Link them together in our mbuto image so that the Fedora build of dhclient
can find it.

Signed-off-by: David Gibson <david@gibson.dropbear.id.au>

AFAICT the symlink we created in mbuto from /usr/bin/bash to /bin/sh was
for the benefit of a dhclient-script which used /usr/bin/bash as its
interpreter (e.g. in Fedora).  That was a bit risky if the script really
did require bash and we linked it to dash or another shell.

We now supply our own custom dhclient-script, so we don't need the
link any more.

Signed-off-by: David Gibson <david@gibson.dropbear.id.au>

David reports that dhclient-script(8) on Fedora needs a number of
binaries that are not included in PROGS of the current mbuto profile,
and we would also need to include hostnamectl(1) there, which will
fail without a systemd init.

Embed a minimal script for dhclient(8) in the profile itself, written
to /sbin/dhclient-script at boot, to just check what we need to check
out of DHCP and DHCPv6 functionality.

While at it, drop busybox and logger from PROGS, as we don't need them,
and add hostname(1). While DHCP option 12 isn't supported yet by the
DHCP implementation in passt, we should probably add it soon.

Note: owing to the simplicity of this script, we now need to bring up
the interface before starting dhclient: add this in test scripts where
it's not the case yet.

Reported-by: David Gibson <david@gibson.dropbear.id.au>
Suggested-by: David Gibson <david@gibson.dropbear.id.au>
Signed-off-by: Stefano Brivio <sbrivio@redhat.com>

note that we need to bring up the interface before starting dhclient