fix(libcgroups): report CPU throttling stats in 'libcgroups::v2'

[xiaoyang-sde]

fix #2513

This pull request adds logic to read relevant fields from cpu.stat file to populate the CpuThrottling struct for the stats method of libcgroups::v2::Cpu.

Reference: https://www.kernel.org/doc/Documentation/admin-guide/cgroup-v2.rst

CPU Interface Files
~~~~~~~~~~~~~~~~~~~

All time durations are in microseconds.

  cpu.stat
	A read-only flat-keyed file.
	This file exists whether the controller is enabled or not.

	It always reports the following three stats:

	- usage_usec
	- user_usec
	- system_usec

	and the following five when the controller is enabled:

	- nr_periods
	- nr_throttled
	- throttled_usec
	- nr_bursts
	- burst_usec

[xiaoyang-sde]

It seems that the stats method of libcgroup::v1::Cpu is using parse_flat_keyed_data to parse the cpu.stats file:

https://github.com/containers/youki/blob/7b5e8f238e45497fe041c9f622a9a4e0b3f0e7e4/crates/libcgroups/src/v1/cpu.rs#L66-L90

Should we use parse_flat_keyed_data in libcgroup::v2::Cpu?

[codecov-commenter]

Codecov Report

Merging #2524 (c857857) into main (b14c7c3) will increase coverage by 0.03%.
Report is 19 commits behind head on main.
The diff coverage is 100.00%.

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #2524      +/-   ##
==========================================
+ Coverage   65.87%   65.91%   +0.03%     
==========================================
  Files         133      133              
  Lines       16868    16873       +5     
==========================================
+ Hits        11112    11121       +9     
+ Misses       5756     5752       -4     

[utam0k]

Thanks for your PR!

Should we use parse_flat_keyed_data in libcgroup::v2::Cpu?

I prefer to use it if possible.

[xiaoyang-sde]

Thanks for your PR!

Should we use parse_flat_keyed_data in libcgroup::v2::Cpu?

I prefer to use it if possible.

✅No problem!

It turns out that the following macro appears at least 3 times in the code base. This macro makes it easier to extract fields from the parsed flat keyed data. Should we move this common logic to crates/libcgroups/src/stats.rs?

macro_rules! get {
    ($name: expr => $field1:ident.$field2:ident) => {
        stats.$field1.$field2 =
            *stats_table
                .get($name)
                .ok_or_else(|| V2CpuStatsError::MissingField {
                    field: $name,
                    path: stats_path.clone(),
                })?;
    };
}

[utam0k]

It turns out that the following macro appears at least 3 times in the code base. This macro makes it easier to extract fields from the parsed flat keyed data. Should we move this common logic to crates/libcgroups/src/stats.rs?

Go ahead 👍 But please send out a new PR about it. When you will do it, please set me as a reviewer.

[utam0k]

LGTM