Fix deadlock in IO thread shutdown during panic

[ouriamzn]

Problem

IO thread shutdown can deadlock during server panic when the main thread calls pthread_cancel() while the IO thread holds its mutex, preventing the thread from observing the cancellation.

Solution

Release the IO thread mutex before cancelling to ensure clean thread termination.

Testing

Reproducer:

bash
./src/valkey-server --io-threads 2 --enable-debug-command yes
./src/valkey-cli debug panic

Before: Server hangs indefinitely
After: Server terminates cleanly

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 72.46%. Comparing base (70196ee) to head (8ba08f8).
⚠️ Report is 2 commits behind head on unstable.

Additional details and impacted files

@@            Coverage Diff            @@
##           unstable    #2898   +/-   ##
=========================================
  Coverage     72.45%   72.46%           
=========================================
  Files           129      129           
  Lines         70526    70528    +2     
=========================================
+ Hits          51102    51109    +7     
+ Misses        19424    19419    -5     

Files with missing lines	Coverage Δ
src/io_threads.c	35.57% <100.00%> (+0.41%)	⬆️

... and 13 files with indirect coverage changes

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.
• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.

[ouriamzn]

This deadlock was introduced by commit 3631208 which removed makeThreadKillable() from IO threads to fix jemalloc crashes. However, removing makeThreadKillable() had an unintended side effect: IO threads no longer respond to pthread_cancel() calls while holding mutexes.

When makeThreadKillable() was present, threads could be cancelled at any cancellation point, even while holding locks. After its removal, pthread_cancel() only works when threads explicitly call pthread_testcancel(), which doesn't happen while the thread holds its mutex in the shutdown path.

This creates a deadlock scenario during panic shutdown:

1. Main thread calls shutdownIOThread()
2. IO thread holds io_threads_mutex[id]
3. Main thread calls pthread_cancel() but thread doesn't observe it
4. pthread_join() blocks indefinitely waiting for a thread that will never exit

The fix ensures the mutex is released before cancellation, allowing the thread to process the cancellation request in its main loop where pthread_testcancel() is called.