Fix crash in profiler when span stack switch causes GC run

[bwoebi]

gc_collect_cycles may be intercepted by profiler and when happening precisely within ddtrace_switch_span_stack, the DDTRACE_G(active_span) points to just freed memory.

[codecov-commenter]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 78.24%. Comparing base (dcc880f) to head (c2a7eaf).
Report is 1 commits behind head on master.

Additional details and impacted files

@@             Coverage Diff              @@
##             master    #2885      +/-   ##
============================================
- Coverage     80.95%   78.24%   -2.72%     
  Complexity     2526     2526              
============================================
  Files           146      173      +27     
  Lines         14711    18742    +4031     
  Branches          0      981     +981     
============================================
+ Hits          11909    14664    +2755     
- Misses         2802     3538     +736     
- Partials          0      540     +540     

Flag	Coverage Δ
appsec-extension	68.49% <ø> (?)
tracer-extension	78.10% <ø> (ø)
tracer-php	82.09% <ø> (-0.06%)	⬇️

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

see 28 files with indirect coverage changes

---

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 dcc880f...c2a7eaf. Read the comment docs.

[pr-commenter]

Benchmarks [ tracer ]

Benchmark execution time: 2024-10-09 17:39:05

Comparing candidate commit c2a7eaf in PR branch bob/span_stack_switch-profiler with baseline commit dcc880f in branch master.

Found 2 performance improvements and 1 performance regressions! Performance is the same for 175 metrics, 0 unstable metrics.

scenario:MessagePackSerializationBench/benchMessagePackSerialization

• 🟩 execution_time [-6.217µs; -3.863µs] or [-3.599%; -2.236%]

scenario:TraceFlushBench/benchFlushTrace

• 🟥 execution_time [+1000.000ns; +1000.000ns] or [+100.000%; +100.000%]

scenario:TraceSerializationBench/benchSerializeTrace

• 🟩 execution_time [-10.556µs; -4.644µs] or [-5.459%; -2.402%]