Be more strict about unrecognized actions. Previously the
parser would happily accept lines with typos like:
kexec_load errrno 1
(note the extra 'r')
Signed-off-by: Wolfgang Bumiller <w.bumiller@proxmox.com>
Moving parse_config_v2 to use getline accidentally parsed
the wrong buffer. Since both _v1 and _v2 now use getline it
seems to be simpler to also use getline() for the first line
before entering the version specific parsers and pass along
the pointer and size so they can reuse them.
Signed-off-by: Wolfgang Bumiller <w.bumiller@proxmox.com>
Fixes: 9c3798eba4 ("seccomp: parse_config_v2()")
This can happen when the 'errno' action can't parse its
supplied number.
Signed-off-by: Wolfgang Bumiller <w.bumiller@proxmox.com>
Fixes: f67c94d00a ("seccomp: parse_v2_rules()")
Fix an error case not free()ing the line forgotten during
the move from fgets() on a static buffer to using getline.
Signed-off-by: Wolfgang Bumiller <w.bumiller@proxmox.com>
Fixes: ccf8d128e4 ("seccomp: parse_config_v1()")
sscanf() skips whitespace anyway so don't account for tabs in case the file
layout changes.
Signed-off-by: Christian Brauner <christian.brauner@ubuntu.com>
The behavior of sigprocmask() is unspecified in multi-threaded programs. Let's
use pthread_sigmask() instead.
Signed-off-by: Christian Brauner <christian.brauner@ubuntu.com>
Specifier %lli was insufficient for the type uint64_t, all values
between 2^63-1 and 2^64-1 were silently converted to 2^63-1.
We can't use %llu since it doesn't handle hexadecimal. Instead, we
parse the values as strings and then use strtoull(3).
Signed-off-by: Felix Abecassis <fabecassis@nvidia.com>
We used to initialize a log unconditionally before. This has led to scenarios
where users where left with container directories and an empty log file even
though they didn't request a log be created at all.
Switch all tools to only create a log file when the user explicitly requests
this.
Closes#1779.
Closes#2032.
Signed-off-by: Christian Brauner <christian.brauner@ubuntu.com>
The execveat allows us to exec stuff via a fd so we don't have to bind mount
stuff in. See the comment about why we're using the syscall directly.
Closes#2339.
Signed-off-by: Tycho Andersen <tycho@tycho.ws>
[christian.brauner@ubuntu.com: adapt error message and whitespace fixes]
Signed-off-by: Christian Brauner <christian.brauner@ubuntu.com>
When using the LXC API multi-thread and users initialize a log:
struct lxc_log log;
log.name = "my-log";
lxc_log_init(&log);
all threads will have the same "my-log" prefix even though thy might call
lxc_container_new() in separate threads. There is currently no easy way to
handle per-thread container name prefixes.
To handle this carry a reference to the name of the container in struct
lxc_conf and if no log.name was set, use it by default. This way each thread
will get the container it is currently working on as a log-prefix.
Signed-off-by: Christian Brauner <christian.brauner@ubuntu.com>
Reported-by: duguhaotian <duguhaotian@gmail.com>
This replaces the constructor implementation of cgroup handling with a simpler,
thread-safe on-demand model of cgroup driver initialization.
Making the cgroup initialization code run in a constructor means that each time
the shared library gets mapped the cgroup parsing code gets run. That's
unnecessary overhead.
It also feels to me that this is only accidently thread-safe because
constructors are only run once. But should threads actually end up manipulating
or freeing memory that is file-global to cgfsng.c we'd be screwed. Now, I might
be wrong here but the cleaner implementation is to allocate a cgroup driver on
demand whenever we need it.
Take the chance and rework the cgroup_ops interface to make the functions it
wants to have implemented a lot cleaner.
Signed-off-by: Christian Brauner <christian.brauner@ubuntu.com>
