passt, branch 2023_03_09.7c7625d

README: Update Features section, plus minor improvements

2023-03-09T02:44:21+00:00

...it's been a while.

Signed-off-by: Stefano Brivio 
Reviewed-by: David Gibson

contrib: Drop libvirt out-of-tree patch, integration mostly works in 9.1.0

2023-03-09T02:44:21+00:00

...and in any case, this patch doesn't offer any advantage over the
current upstream integration.

Signed-off-by: Stefano Brivio 
Reviewed-by: David Gibson

contrib: Drop QEMU out-of-tree patches

2023-03-09T02:44:21+00:00

Native support was introduced with commit 13c6be96618c, QEMU 7.2.

Signed-off-by: Stefano Brivio 
Reviewed-by: David Gibson

contrib: Drop Podman out-of-tree patch, integration is upstream now

2023-03-09T02:44:21+00:00

See https://github.com/containers/podman/pull/16141, shipped in
Podman 4.4.

Signed-off-by: Stefano Brivio 
Reviewed-by: David Gibson

tcp: Clamp MSS value when queueing data to tap, also for pasta

2023-03-09T02:44:21+00:00

Tom reports that a pattern of repated ~1 MiB chunks downloads over
NNTP over TLS, on Podman 4.4 using pasta as network back-end, results
in pasta taking one full CPU thread after a while, and the download
never succeeds.

On that setup, we end up re-sending the same frame over and over,
with a consistent 65 534 bytes size, and never get an
acknowledgement from the tap-side client. This only happens for the
default MTU value (65 520 bytes) or for values that are slightly
smaller than that (down to 64 499 bytes).

We hit this condition because the MSS value we use in
tcp_data_from_sock(), only in pasta mode, is simply clamped to
USHRT_MAX, and not to the actual size of the buffers we pre-cooked
for sending, which is a bit less than that.

It looks like we got away with it until commit 0fb7b2b9080a ("tap:
Use different io vector bases depending on tap type") fixed the
setting of iov_len.

Luckily, since it's pasta, we're queueing up to two frames at a time,
so the worst that can happen is a badly segmented TCP stream: we
always have some space at the tail of the buffer.

Clamp the MSS value to the appropriate maximum given by struct
tcp{4,6}_buf_data_t, no matter if we're running in pasta or passt
mode.

While at it, fix the comments to those structs to reflect the current
struct size. This is not really relevant for any further calculation
or consideration, but it's convenient to know while debugging this
kind of issues.

Thanks to Tom for reporting the issue in a very detailed way and for
providing a test setup.

Reported-by: Tom Mombourquette 
Link: https://github.com/containers/podman/issues/17703
Signed-off-by: Stefano Brivio 
Reviewed-by: David Gibson

conf: Terminate on EMFILE or ENFILE on sockets for port mapping

2023-03-09T02:44:21+00:00

In general, we don't terminate or report failures if we fail to bind
to some ports out of a given port range specifier, to allow users to
conveniently specify big port ranges (or "all") without having to
care about ports that might already be in use.

However, running out of the open file descriptors quota is a
different story: we can't do what the user requested in a very
substantial way.

For example, if the user specifies '-t all' and we can only bind
1024 sockets, the behaviour is rather unexpected.

Fail whenever socket creation returns -ENFILE or -EMFILE.

Link: https://bugs.passt.top/show_bug.cgi?id=27
Signed-off-by: Stefano Brivio 
Reviewed-by: David Gibson

tcp, udp: Fix partial success return codes in {tcp,udp}_sock_init()

2023-03-09T02:44:21+00:00

The comments say we should return 0 on partial success, and an error
code on complete failure. Rationale: if the user configures a port
forwarding, and we succeed to bind that port for IPv4 or IPv6 only,
that might actually be what the user intended.

Adjust the two functions to reflect the comments.

Signed-off-by: Stefano Brivio 
Reviewed-by: David Gibson

tcp, udp, util: Pass socket creation errors all the way up

2023-03-09T02:44:21+00:00

...starting from sock_l4(), pass negative error (errno) codes instead
of -1. They will only be used in two commits from now, no functional
changes intended here.

Signed-off-by: Stefano Brivio 
Reviewed-by: David Gibson

util: Carry own definition of __bswap_constant{16,32}

2023-03-09T02:44:21+00:00

musl doesn't define those, use our own definition there. This is a
trivial implementation, similar to what's shipped by e.g. uClibc,
glibc, libiio.

Reported-by: Chris Kuhn 
Signed-off-by: Stefano Brivio 
Reviewed-by: David Gibson

treewide: Fix header includes to build with musl

2023-03-09T02:44:21+00:00

Roughly inspired from a patch by Chris Kuhn: fix up includes so that
we can build against musl: glibc is more lenient as headers generally
include a larger amount of other headers.

Compared to the original patch, I only included what was needed
directly in C files, instead of adding blanket includes in local
header files. It's a bit more involved, but more consistent with the
current (not ideal) situation.

Reported-by: Chris Kuhn 
Signed-off-by: Stefano Brivio 
Reviewed-by: David Gibson