When starting a container and dropping all capabilities,
the init child process has no permission to read the exec.fifo
file because the parent set the file mode 0o622. So change the exec.fifo file mode to 0o644.
fixes#1913
Signed-off-by: quanweiZhou <quanweiZhou@linux.alibaba.com>
While evaluating the possibility of having kata-agent statically linked
to the GNU libc, we've ended up facing some issues with prctl.
When debugging the issues, we figured out that the crate hasn't been
maintained since 2015 and that the capctl one is a good 1:1 replacement
for what we need.
Fixes: #1844
Signed-off-by: Fabiano Fidêncio <fidencio@redhat.com>
Since the propagation flags couldn't be combinted with the
standard mount flags, and they should be used with the remount,
thus it's better to split them from the standard mount flags.
Fixes: #1699
Signed-off-by: fupan.lfp <fupan.lfp@antgroup.com>
In file src/agent/rustjail/src/validator.rs,
these two functions are not used:
- get_namespace_path
- check_host_ns
Fixes: #1783
Signed-off-by: bin <bin@hyper.sh>
Clippy (in Rust 1.51 at least) has some complaints about this closure
inside execute_hook() because it uses explicit returns in some places
where it doesn't need them, because they're the last expression in the
function.
That isn't necessarily obvious from a glance, but we can make clippy happy
and also make things a little clearer: first we replace a somewhat verbose
'match' using Option::ok_or_else(), then rearrange the remaining code to
put all the error path first with an explicit return then the "happy" path
as the stright line exit with an implicit return.
fixes#1611
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
PathBuf is an owned, mutable Path. We don't need those properties in
get_value_from_cgroup() so we can use a Path instead. This may be slightly
safer, and definitely stops clippy (version 1.51 at least) from
complaining.
fixes#1611
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
DEFAULT_ALLOWED_DEVICES and DEFAULT_DEVICES are essentially global
constant lists. They're implemented as a lazy_static! initialized Vec
values.
The code to initialize them creates an empty Vec then pushes values
onto it. We can simplify this a bit by using the vec! macro. This
might be slightly more efficient, and it definitely stops recent
clippy versions (e.g. 1.51) from complaining about it.
fixes#1611
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
Recent versions of clippy (e.g. in Rust 1.51) complain about a number
of names in the oci crate, which don't obey Rust's normal CamelCasing
conventions.
It's pretty clear that these don't obey the usual rules because they
are attempting to preserve conventional casing of existing acronyms
they incorporate ("VM", "POSIX", etc.). However, it's been my
experience that matching the case and name conventions of your
environs is more important than matching case with external norms.
Therefore, this patch changes all the identifiers in the oci crate to
match Rust conventions. Their users in the rustjail crate are updated
to match.
fixes#1611
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
This comment appears to be connected specifically with this function, but
has some other items separating it for no particular reason. It also has
a typo. Correct both.
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
Functions in rustjail deal with both the local oci module's data structure
and the protocol::oci module's data structure. Since these both cover the
OCI container config they are quite similar and have many identically named
types.
To avoid conflicts, we import many things from those modules with altered
names. However the names we use oci* and grpc* don't fit the normal Rust
capitalization convention for types.
However by renaming the import of the 'protocols::oci' module itself to
'grpc', we can actually get rid of the many renames by just qualifying at
each use site with only a very small increase in verbosity. As a bonus
this gets rid of multiple 'use' items scattered through the file.
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
The situation is not a IPC scene, pipe(2) is too heavy.
We have tokio::sync:⌚:channel after tokio has been introduced.
The channel has better performance and easy to use.
Fixes: #1721
Signed-off-by: Tim Zhang <tim@hyper.sh>
Update:
- Make the type of errnoRet in oci.proto oneof
- Update seccomp_grpc_to_oci that can set errnoRet as EPREM if the
value is empty.
- Update the oci.pb.go based on the above fixes
- Add seccomp errnoRet and flags option to configs in rustjail
Fixes: #1719
Signed-off-by: Manabu Sugimoto <Manabu.Sugimoto@sony.com>
register_memory_event_v2() includes a closure spawned as an async task
with tokio. At the end of that closure, there's a test for a closed fd
exiting if so. But this is right at the end of the closure when it was
about to exit anyway, so this does nothing.
This code was originally an explicit thread, converted to a tokio task
by 332fa4c "agent: switch to async runtime". It looks like there was an
error during conversion, where this logic was accidentally moved out of the
while loop above, where it makes a lot more sense.
Put it back into the loop.
fixes#1702
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
first get the "HOME" env from "/etc/passwd", if
there's no corresponding uid entry in /etc/passwd,
then set "/" as the home env.
Fixes: #1643
Signed-off-by: fupan.lfp <fupan.lfp@antfin.com>
The only right way to shutdown pipe is drop it
Otherwise PipeStream will conflict with its twins
Because they both have the same fd, and both registered.
Fixes: #1614
Signed-off-by: Tim Zhang <tim@hyper.sh>
Rust 1.51 appears to have added a new warning in anticipation of Rust 2021,
which requires the format string for panic!()s (including via the various
assert!() macros) to be a string literal. This triggers quite a few times
in the agent code. This patch fixes them.
fixes#1626
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
statfs f_types are long on most architectures, but not on s390x, where
they are uint. Following the fix in rust-lang/libc at
https://github.com/rust-lang/libc/pull/1999, the custom defined
PROC_SUPER_MAGIC must be updated in a similar way.
Fixes: #1204
Signed-off-by: Jakob Naucke <jakob.naucke@ibm.com>
It's better to get the online cpus from
"/sys/devices/system/cpu/online" instead of from
cpuset cgroup, cause there would be an latency
between one cpu online and present in the root
cpuset cgroup.
Fixes: #1536
Signed-off-by: fupan.lfp <fupan.lfp@antfin.com>
Since the crate dirs::home_dir function depends on the
libc's api: getpwuid_r, but this api function wouldn't
be static linked on glibc, thus we'd better to figure
out an alternative way to get the home dir from /etc/passwd.
For much more info about this glibc's issue, please see:
https://sourceware.org/bugzilla/show_bug.cgi?id=19341.
This commit read and parse the "/etc/passwd" directly and
fetch the corresponding uid's home dir.
Fixes: #675
Signed-off-by: fupan.lfp <fupan.lfp@antfin.com>
When do pass guest device files to container, the source
file wouldn't be a regular file, but we also need to create
a corresponding destination file to bind mount source file
to it. Thus it's better to check whether the source file
was a directory instead of regular file.
Fixes: #1477
Signed-off-by: fupan.lfp <fupan.lfp@antfin.com>
Agent sends -1 PID when invoking OCI hooks.
OCI state struct is initialized before obtaining PID, so this PR moves
`oci_state` call down, right after we get the id.
Fixes: #1458
Signed-off-by: Maksym Pavlenko <pavlenko.maksym@gmail.com>
If the container has exited, the sender in notifier watching OOM events
will be dropped after the loop exited, and recv() from the according
receiver will get None.
This will lead two problems for get_oom_event rpc all from agent:
- return an wrong OOM event.
- continuously return OOM events.
Fixes: #1369
Signed-off-by: bin <bin@hyper.sh>
This PR fixes wrong serialization of OCI state object.
OCI hooks end up with a JSON string with double quotes in `state` field.
This happens because of confusion `Debug` and `Display` traits. Debug trait
returns a string representation with double quotes.
Ideally we should not use Debug as a part of serialization process, so a bit
more safer fix would be to move container states to `oci` crate and simply
disallow wrong values in that field.
`ContainerState` in go spec: https://github.com/opencontainers/runtime-spec/blob/master/specs-go/state.go#L4Fixes: #1404
Signed-off-by: Maksym Pavlenko <pavlenko.maksym@gmail.com>
Seccomp configuration should be updated to prepare for the future seccomp support based on the latest OCI specification.
Add:
- flags which is used with seccomp(2) in struct LinuxSeccomp
- errnoRet which is errno return code in struct LinuxSyscall
- some new seccomp actions and an architecture
Fixes: #1391
Signed-off-by: Manabu Sugimoto <Manabu.Sugimoto@sony.com>
The current implementation of rustjail uses the specific setrlimit.
This patch uses rlimit crate for maintainability.
Fixes: #1372
Signed-off-by: Manabu Sugimoto <Manabu.Sugimoto@sony.com>
test_process has a assertion that waitpid(-1) will fail
because there is no child process in most cases.
But if there is any child process forked by other unit test,
the test test_process will fail.
Because waitpid(-1) will wait for any child process including the
process created by other unit tests.
Signed-off-by: Tim Zhang <tim@hyper.sh>
