this removes proxmox_sys as a dependecy for proxmox-async
Signed-off-by: Dominik Csapak <d.csapak@proxmox.com>
Signed-off-by: Wolfgang Bumiller <w.bumiller@proxmox.com>
will break usage of the `Result::Ok()' with anyhow 1.0.49+ as that
added a new Ok helper, so a glob-import would make that shadow the
core one.
Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
Since the only reasonable access is by mount "id" and they
don't need to be consecutive (but they are numeric), use a
BTreeMap for MountInfo.
Signed-off-by: Wolfgang Bumiller <w.bumiller@proxmox.com>
Since nix's Pid type is a strong type, we can add a TryFrom
implementation here as well.
This can later also have the equivalent TryFrom<PidFd>
variant.
Signed-off-by: Wolfgang Bumiller <w.bumiller@proxmox.com>
Rather than standalone functions, `PidStat` now provides a
`read_for_pid` and `parse` method. Both are public.
One side effect is that the documentation of PidStat's
contents are on the same page as the ways to retrieve it.
Additionally, we deprecate `read_proc_starttime` as it was
reading the entire contents to then filter out one value, in
order to promote caching.
The standalone `read_proc_pid_stat` method is now also
deprecated.
Once we add pid-fds the API will feel cleaner this way.
Signed-off-by: Wolfgang Bumiller <w.bumiller@proxmox.com>
Rename `ProcFsPidStat` to `PidStat`. It's already in a
procfs submodule anyway. (We have currently no public users
of this type by name.)
Further add pid, ppid and num_threads to PidStat.
Public API change.
Signed-off-by: Wolfgang Bumiller <w.bumiller@proxmox.com>
For now we just parse the CPU "time-spent" counters and calculate the
CPU usage of the whole system. As we do not want a total average we
use a state-full implementation, which calculates the usage since the
last request (if older than 1 second).
Somewhat modeled after our Perl version of this. But for the "total"
we do not use the:
> ($ctime - $last_proc_stat->{ctime}) * $clock_ticks * $cpucount;
calculation, but just add _all_ fields together, as that is the total
of CPU time.
Use a Reader-Writer lock to sync updating. First check under the
reader lock if the data is recent enough, else update it under the
writer lock.
Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
