New directive `NFTSet=` provides a method for integrating network configuration
into firewall rules with NFT sets. The benefit of using this setting is that
static network configuration or dynamically obtained network addresses can be
used in firewall rules with the indirection of NFT set types. For example,
access could be granted for hosts in the local subnetwork only. Firewall rules
using IP address of an interface are also instantly updated when the network
configuration changes, for example via DHCP.
This option expects a whitespace separated list of NFT set definitions. Each
definition consists of a colon-separated tuple of source type (one of
"address", "prefix", or "ifindex"), NFT address family (one of "arp", "bridge",
"inet", "ip", "ip6", or "netdev"), table name and set name. The names of tables
and sets must conform to lexical restrictions of NFT table names. The type of
the element used in the NFT filter must match the type implied by the
directive ("address", "prefix" or "ifindex") and address type (IPv4 or IPv6)
as shown type implied by the directive ("address", "prefix" or "ifindex") and
address type (IPv4 or IPv6) must also match the set definition.
When an interface is configured with IP addresses, the addresses, subnetwork
masks or interface index will be appended to the NFT sets. The information will
be removed when the interface is deconfigured. systemd-networkd only inserts
elements to (or removes from) the sets, so the related NFT rules, tables and
sets must be prepared elsewhere in advance. Failures to manage the sets will be
ignored.
/etc/systemd/network/eth.network
```
[DHCPv4]
...
NFTSet=prefix:netdev:filter:eth_ipv4_prefix
```
Example NFT rules:
```
table netdev filter {
set eth_ipv4_prefix {
type ipv4_addr
flags interval
}
chain eth_ingress {
type filter hook ingress device "eth0" priority filter; policy drop;
ip saddr != @eth_ipv4_prefix drop
accept
}
}
```
```
$ sudo nft list set netdev filter eth_ipv4_prefix
table netdev filter {
set eth_ipv4_prefix {
type ipv4_addr
flags interval
elements = { 10.0.0.0/24 }
}
}
```
Let's also use vcs_tag() when we're doing a non-git build. In those scenarios,
the build would normally be done just once in a given copy, so doing an extra
call does not matter. We can save a few lines of meson config.
The special path was added in 064b8e2c99, with
the justifaction that vcs_tag() is slow and -Dversion-tag=foo can be used to
fix the version tag and speed up partial rebuilds. I think the justification
for this is weak: having an accurate version tag is particularly useful when
developing the code. Shaving of a fraction of a second at the cost of having to
manually update the version seems iffy.
Secondly, with vcs_tag() we can be pretty sure that meson will build the
version file first and that it'll be available to all build steps. Because we
didn't use version tag, we had to manually specify the dependency on version.h
in various places. It seems nicer to use vcs_tag() and not have to deal with
this problem at all.
Finally, the savings in time seem much smaller than back when
064b8e2c99 was made. It reported a change
from 94 ms to 521 ms. But now the difference seems to be about 50 ms:
Before this patch:
$ time ninja -C build
ninja: Entering directory `build'
ninja: no work to do.
ninja -C build 0.04s user 0.02s system 97% cpu 0.057 total
ninja -C build 0.03s user 0.01s system 97% cpu 0.049 total
ninja -C build 0.03s user 0.02s system 96% cpu 0.051 total
ninja -C build 0.03s user 0.01s system 96% cpu 0.049 total
ninja -C build 0.03s user 0.01s system 97% cpu 0.046 total
With the two patches in this PR:
systemd-stable [drop-versiondep] time ninja -C build
ninja: Entering directory `build'
[1/669] Generating version.h with a custom command
ninja -C build 0.08s user 0.03s system 98% cpu 0.106 total
ninja -C build 0.08s user 0.03s system 98% cpu 0.104 total
ninja -C build 0.09s user 0.02s system 98% cpu 0.116 total
ninja -C build 0.08s user 0.02s system 97% cpu 0.108 total
Overall, I think the tiny time savings are not worth the complexity.
The use of vcs_tag was dropped in #28567, which results in builds having
stale version information once new commit are made.
This also fixes a case where CI builds would have no version information
because they are checked out without any tags for git-describe to use.
Additionally, use `--git-dir` now, as that particular issues seems to
have been fixed by now.
We might inherit a max rlim value that's larger than the kernel's
maximum (nr_open). This will cause setrlimit() to fail as the given
maximum is larger than the kernel's maximum. To get around this,
let's limit the max rlim we pass to rlimit() to the value of nr_open.
Should fix#28965
I always found it confusing that most of our TPM related definitions are
in tpm2-util.h, but the PCR names in tpm-pcr.h, without the "2". Let's
fix that and make this systematic, in particular as the definitions in
the file all start with TPM2_ already.
No code flow changes, just some renaming.
We so far maintained two places for symboic names for PCRs. One in
tpm2-util.h and one in tpm-pcr.h.
Let's unify this into one, i.e. move the full list from tpm2-util.h into
tpm-pcr.h, replacing the short list placed so far there.
Systematically prefix the definitions with TPM2_ or tpm2_, to follow how
we do this for all other defines in this context.
No change in behaviour, just unification of tables.
Let's a "b" helper for json_variant_set_field() that combines
json_build() with json_variant_set_field(), similar to the
json_variant_merge_objectb(), json_variant_append_arrayb().
json_append() is a useful wrapper around json_variant_merge(). However,
I think the naming sould be cleaned up a bit of both functions.
I thinker "merge" is the better word than "append", since it does
decidedly more than just append: it replaces existing fields of the same
name, hence "merge" sounds more appropriate. This is as opposed to the
similar operations for arrays, where no such override logic is applied
and we really just append, hence those functions are called "append"
already.
To make clearer that "merge" is about objects, and "append" about
arrays, also include "object" in the name.
Also, include "json_variant" in the name, like we do for almost all
other functions in the JSON API that take a JSON object as primary
input, and hence are kinda object methods.
Finally, let's follow the logic that helpers that combine json_build()
with some other operation get suffixed with "b" like we already have in
some cases.
Hence:
json_variant_merge() → json_variant_merge_object()
json_append() → json_variant_merge_objectb()
This mirrors nicely the existing:
json_variant_append_array()
json_vairant_append_arrayb()
This also drops the variant of json_append() that takes a va_arg
parameter (i.e. json_appendv()). We have no user of that so far, and
given the nature as a helper function only I don#t see that happening,
and if it happens after all it's trivial to bring back.
Let's make utf8_to_utf16() and utf16_to_utf8() a bit nicer to use by
adding shortcuts for common cases.
This is particularly relevant for utf16_to_utf8() since the
multiplication with 2 is easy to forget.
Alvin Alvarado
Topi Miettinen
Zbigniew Jędrzejewski-Szmek
Juno Computers
Jan Janssen
Mike Yuan
Susant Sahani
Daan De Meyer
Dan Streetman
Yu Watanabe
Lennart Poettering