Block remote config signals during ftp functions

[bwoebi]

These are sensitive to EINTR, so block the signal during their execution.

Fixes #2952.

[codecov-commenter]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 66.52%. Comparing base (6842245) to head (d5a6ed1).
Report is 1 commits behind head on master.

Additional details and impacted files

@@            Coverage Diff            @@
##             master    #2957   +/-   ##
=========================================
  Coverage     66.52%   66.52%           
  Complexity     2739     2739           
=========================================
  Files           109      109           
  Lines         10831    10831           
=========================================
  Hits           7205     7205           
  Misses         3626     3626           

Flag	Coverage Δ
tracer-php	66.52% <ø> (ø)

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

---

Continue to review full report in Codecov by Sentry.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 6842245...d5a6ed1. Read the comment docs.

[pr-commenter]

Benchmarks [ tracer ]

Benchmark execution time: 2024-11-18 13:31:06

Comparing candidate commit d5a6ed1 in PR branch bob/block-ftp-sigprocmask with baseline commit 6842245 in branch master.

Found 14 performance improvements and 1 performance regressions! Performance is the same for 163 metrics, 0 unstable metrics.

scenario:EmptyFileBench/benchEmptyFileBaseline

• 🟩 execution_time [-463.695µs; -250.045µs] or [-13.936%; -7.515%]

scenario:EmptyFileBench/benchEmptyFileBaseline-opcache

• 🟩 execution_time [-539.932µs; -331.608µs] or [-15.369%; -9.439%]

scenario:EmptyFileBench/benchEmptyFileOverhead

• 🟩 execution_time [-633.831µs; -391.369µs] or [-16.974%; -10.481%]

scenario:EmptyFileBench/benchEmptyFileOverhead-opcache

• 🟩 execution_time [-508.408µs; -303.612µs] or [-13.362%; -7.979%]

scenario:LaravelBench/benchLaravelBaseline

• 🟩 execution_time [-596.411µs; -347.709µs] or [-5.494%; -3.203%]

scenario:LaravelBench/benchLaravelBaseline-opcache

• 🟩 execution_time [-613.119µs; -333.021µs] or [-5.567%; -3.024%]

scenario:LaravelBench/benchLaravelOverhead

• 🟩 execution_time [-1.328ms; -1.139ms] or [-10.592%; -9.091%]

scenario:LaravelBench/benchLaravelOverhead-opcache

• 🟩 execution_time [-1.249ms; -0.995ms] or [-9.901%; -7.883%]

scenario:PDOBench/benchPDOBaseline

• 🟥 execution_time [+10.175µs; +13.701µs] or [+5.639%; +7.593%]

scenario:SymfonyBench/benchSymfonyBaseline

• 🟩 execution_time [-391.049µs; -267.411µs] or [-5.922%; -4.050%]

scenario:SymfonyBench/benchSymfonyBaseline-opcache

• 🟩 execution_time [-479.490µs; -342.890µs] or [-7.094%; -5.073%]

scenario:SymfonyBench/benchSymfonyOverhead

• 🟩 execution_time [-871.506µs; -803.014µs] or [-11.293%; -10.405%]

scenario:SymfonyBench/benchSymfonyOverhead-opcache

• 🟩 execution_time [-913.512µs; -803.428µs] or [-11.613%; -10.214%]

scenario:WordPressBench/benchWordPressOverhead

• 🟩 execution_time [-4.612ms; -4.262ms] or [-14.772%; -13.653%]

scenario:WordPressBench/benchWordPressOverhead-opcache

• 🟩 execution_time [-4.454ms; -4.077ms] or [-14.241%; -13.036%]

[Anilm3]

If my understanding is correct, this will re-emit at least one queued signal, meaning that you may not need to call ddtrace_check_for_new_config_now.

If invoking sigprocmask causes any pending signals to be unblocked, at least one of those signals is delivered to the process before sigprocmask returns. The order in which pending signals are delivered is not specified, but you can control the order explicitly by making multiple sigprocmask calls to unblock various signals one at a time. (reference)

[bwoebi]

Interesting, that's really nice. Manpages for non-Linux (e.g. macos) say no such thing, so I'll #ifndef __linux__ this.

[pr-commenter]

Benchmarks [ appsec ]

Benchmark execution time: 2024-11-18 13:39:05

Comparing candidate commit d5a6ed1 in PR branch bob/block-ftp-sigprocmask with baseline commit 6842245 in branch master.

Found 5 performance improvements and 0 performance regressions! Performance is the same for 7 metrics, 0 unstable metrics.

scenario:LaravelBench/benchLaravelBaseline-appsec

• 🟩 execution_time [-579.038µs; -383.722µs] or [-5.325%; -3.529%]

scenario:LaravelBench/benchLaravelOverhead-appsec

• 🟩 execution_time [-1.386ms; -1.179ms] or [-9.374%; -7.973%]

scenario:SymfonyBench/benchSymfonyBaseline-appsec

• 🟩 execution_time [-432.939µs; -351.341µs] or [-6.541%; -5.308%]

scenario:SymfonyBench/benchSymfonyOverhead-appsec

• 🟩 execution_time [-1128.152µs; -772.128µs] or [-11.235%; -7.690%]

scenario:WordPressBench/benchWordPressOverhead-appsec

• 🟩 execution_time [-4.445ms; -4.115ms] or [-13.128%; -12.154%]