fix(libscap): use the correct memory barrier for ARM64

[Andreagit97]

What type of PR is this?

/kind bug

Any specific area of the project related to this PR?

/area libscap-engine-bpf

/area libscap-engine-kmod

/area libpman

Does this PR require a change in the driver versions?

No

What this PR does / why we need it:

This PR should fix #2061.
Today smp_load_acquire and smp_store_release use the synchronization primitives of x86 also on other architectures... but thanks to the nature of our code we never noticed it until this PR #2009.

More detail in the addition of

	if(g_state.cons_pos[pos] == g_state.prod_pos[pos]) {
			return NULL;
	}

caused the g_state.prod_pos[pos] value to be used immediately after the fetch. So if it was a wrong value we might end up reading some junk from the buffer... And this is exactly what is happening here #2061. This is for the modern ebpf.

the other drivers also used the wrong primitives, but again, probably due to some luck and rare race conditions we never faced any issues...btw I tried to fix them

The fix is to use the right synchronization primitives! Unfortunately, the definitions of these primitives are in the kernel headers so to avoid the constraint of having the headers when building libscap engines, I decided to copy the needed definitions, not sure if we want to take other directions here...

Which issue(s) this PR fixes:

Fixes #2061

Special notes for your reviewer:

Does this PR introduce a user-facing change?:

NONE

[github-actions]

Please double check driver/API_VERSION file. See versioning.

/hold

[github-actions]

Perf diff from master - unit tests

     1.71%     +0.75%  [.] scap_event_decode_params
     4.28%     -0.68%  [.] gzfile_read
     9.71%     +0.64%  [.] sinsp_parser::reset
     1.12%     +0.61%  [.] sinsp_parser::event_cleanup
     1.17%     -0.53%  [.] sinsp_parser::parse_context_switch
     0.89%     -0.52%  [.] scap_file_test_helpers::capture_search_evt_by_num
     3.32%     +0.50%  [.] sinsp_thread_manager::get_thread_ref
     0.49%     +0.43%  [.] libsinsp::state::stl_container_table_adapter<std::vector<std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> >, std::allocator<std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > > >, libsinsp::state::value_table_entry_adapter<std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > >, libsinsp::state::value_table_entry_adapter<std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > >::dynamic_fields_t>::stl_container_table_adapter
     3.62%     +0.43%  [.] sinsp_parser::process_event
     4.45%     +0.41%  [.] sinsp_evt::load_params

Heap diff from master - unit tests

peak heap memory consumption: 0B
peak RSS (including heaptrack overhead): 0B
total memory leaked: 0B

Heap diff from master - scap file

peak heap memory consumption: 0B
peak RSS (including heaptrack overhead): 0B
total memory leaked: 0B

Benchmarks diff from master

Comparing gbench_data.json to /root/actions-runner/_work/libs/libs/build/gbench_data.json
Benchmark                                                         Time             CPU      Time Old      Time New       CPU Old       CPU New
----------------------------------------------------------------------------------------------------------------------------------------------
BM_sinsp_split_mean                                            +0.0010         +0.0010           144           144           144           144
BM_sinsp_split_median                                          +0.0107         +0.0107           142           144           142           144
BM_sinsp_split_stddev                                          -0.8604         -0.8602             2             0             2             0
BM_sinsp_split_cv                                              -0.8605         -0.8603             0             0             0             0
BM_sinsp_concatenate_paths_relative_path_mean                  -0.0462         -0.0462            43            41            43            41
BM_sinsp_concatenate_paths_relative_path_median                -0.0402         -0.0402            42            41            42            41
BM_sinsp_concatenate_paths_relative_path_stddev                -0.6857         -0.6857             1             0             1             0
BM_sinsp_concatenate_paths_relative_path_cv                    -0.6705         -0.6705             0             0             0             0
BM_sinsp_concatenate_paths_empty_path_mean                     -0.0192         -0.0192            17            17            17            17
BM_sinsp_concatenate_paths_empty_path_median                   -0.0159         -0.0158            17            17            17            17
BM_sinsp_concatenate_paths_empty_path_stddev                   -0.4672         -0.4672             0             0             0             0
BM_sinsp_concatenate_paths_empty_path_cv                       -0.4567         -0.4567             0             0             0             0
BM_sinsp_concatenate_paths_absolute_path_mean                  -0.0293         -0.0293            45            44            45            44
BM_sinsp_concatenate_paths_absolute_path_median                -0.0331         -0.0331            45            44            45            44
BM_sinsp_concatenate_paths_absolute_path_stddev                -0.3855         -0.3855             1             1             1             1
BM_sinsp_concatenate_paths_absolute_path_cv                    -0.3669         -0.3670             0             0             0             0
BM_sinsp_split_container_image_mean                            +0.0415         +0.0415           350           365           350           365
BM_sinsp_split_container_image_median                          +0.0409         +0.0409           350           365           350           365
BM_sinsp_split_container_image_stddev                          +0.8460         +0.8450             1             3             1             3
BM_sinsp_split_container_image_cv                              +0.7725         +0.7715             0             0             0             0

[Andreagit97]

This is used by the scap_kmod and scap_bpf engines and it uses the same primitive of smp_store/smp_load

[poiana]

LGTM label has been added.

Git tree hash: 5a1104014ab3043c18da50fbcf6b1f96d3032a20

[FedeDP]

This is a great fix!
Than you very much for digging into it!
/approve

[poiana]

[APPROVALNOTIFIER] This PR is APPROVED

This pull-request has been approved by: Andreagit97, FedeDP

The full list of commands accepted by this bot can be found here.

The pull request process is described here

Needs approval from an approver in each of these files:

• OWNERS [Andreagit97,FedeDP]

Approvers can indicate their approval by writing /approve in a comment
Approvers can cancel approval by writing /approve cancel in a comment

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 73.58%. Comparing base (398964a) to head (7375fc8).
Report is 1 commits behind head on master.

Additional details and impacted files

@@           Coverage Diff           @@
##           master    #2067   +/-   ##
=======================================
  Coverage   73.58%   73.58%           
=======================================
  Files         253      253           
  Lines       31867    31867           
  Branches     5648     5633   -15     
=======================================
  Hits        23449    23449           
+ Misses       8398     8388   -10     
- Partials       20       30   +10     

Flag	Coverage Δ
libsinsp	73.58% <ø> (ø)

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.