Rework of PfRingDevice capture thread implementation.

[Dimi1010]

Fixes #1664.

• Added StopToken and StopTokenSource that mirror C++20's stop_token and stop_source.
• Reworked the capture thread function into a free function to simplify interactions.
  Many variables are now kept on the thread's stack instead of relying on member fields in PfRindDevice.
• Extracted the code for setting core affinity into a function.
• Cleaned up capture thread startup.

[codecov]

Codecov Report

Attention: Patch coverage is 98.38710% with 1 line in your changes missing coverage. Please review.

Project coverage is 83.17%. Comparing base (ab75534) to head (baeb36c).

Files with missing lines	Patch %	Lines
Pcap++/header/StopToken.h	90.90%	0 Missing and 1 partial ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##              dev    #1668      +/-   ##
==========================================
+ Coverage   83.15%   83.17%   +0.01%     
==========================================
  Files         277      280       +3     
  Lines       48193    48255      +62     
  Branches     9949     9969      +20     
==========================================
+ Hits        40077    40137      +60     
+ Misses       7225     7222       -3     
- Partials      891      896       +5     

Flag	Coverage Δ
alpine320	75.16% <94.44%> (+0.01%)	⬆️
fedora40	75.18% <94.73%> (-0.02%)	⬇️
macos-13	80.68% <100.00%> (+0.02%)	⬆️
macos-14	80.68% <100.00%> (+0.02%)	⬆️
macos-15	80.65% <100.00%> (+0.02%)	⬆️
mingw32	70.90% <93.75%> (-0.02%)	⬇️
mingw64	70.86% <93.75%> (-0.02%)	⬇️
npcap	85.32% <100.00%> (+<0.01%)	⬆️
rhel94	75.06% <94.73%> (+0.01%)	⬆️
ubuntu2004	58.63% <93.75%> (+0.01%)	⬆️
ubuntu2004-zstd	58.75% <93.75%> (+0.02%)	⬆️
ubuntu2204	74.99% <94.73%> (+0.02%)	⬆️
ubuntu2204-icpx	61.45% <100.00%> (+0.05%)	⬆️
ubuntu2404	75.21% <94.73%> (-0.01%)	⬇️
unittest	83.17% <98.38%> (+0.01%)	⬆️
windows-2019	85.35% <100.00%> (+<0.01%)	⬆️
windows-2022	85.38% <100.00%> (+<0.01%)	⬆️
winpcap	85.35% <100.00%> (+0.01%)	⬆️
xdp	50.48% <0.00%> (-0.07%)	⬇️

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.

[Dimi1010]

I don't actually have a system that can run PF_RING. So it would be nice if someone could test it live.

[seladb]

@Dimi1010 this PR does a lot of things and could be broken up into multiple PRs to make it easier to review.

In order to solve #1664, wouldn't it be enough to create a shared_ptr from mutex, cond and startThread and pass them to the worker threads? 🤔

[Dimi1010]

@Dimi1010 this PR does a lot of things and could be broken up into multiple PRs to make it easier to review.

In order to solve #1664, wouldn't it be enough to create a shared_ptr from mutex, cond and startThread and pass them to the worker threads? 🤔

I suppose it can be split up.

The main fix is the startup block struct. The rest like the StopToken was added as the whole startThread being 0, 1 or 2 was a bit hard to read.

[seladb]

@Dimi1010 this PR does a lot of things and could be broken up into multiple PRs to make it easier to review.
In order to solve #1664, wouldn't it be enough to create a shared_ptr from mutex, cond and startThread and pass them to the worker threads? 🤔

I suppose it can be split up.

The main fix is the startup block struct. The rest like the StopToken was added as the whole startThread being 0, 1 or 2 was a bit hard to read.

I wouldn't add this much code just to make startThread more readable... if we can first fix #1664, potentially by using shared_ptr, it could make a great first PR. Then we can open additional PRs to add the rest of the fixes. Please let me know what you think