Update posixtestsuite to deal with OOM issue on fast computers (#25318)

A couple of tests in the posixtest suite hammer `pthread_kill()` and
`pthread_cancel()` in an uncontrollably fast loop.

Since the event queue mailbox refactor, proxying `pthread_kill()` and
`pthread_cancel()` allocate memory.

This causes fast CPUs like Ryzen 9800X3D 8c/16t and Apple M1 ARM to run
out of memory while running the test:

http://clbri.com:8010/api/v2/logs/64925/raw_inline
http://clbri.com:8010/api/v2/logs/66008/raw_inline

Update the posixtest submodule to latest, which throttles these loops to
sleep 1ms between signals.

Author: juj

test/test_posixtest.py

@@ -166,6 +166,9 @@ def f(self):
             '-Wno-int-conversion',
             '-Wno-format',
             '-pthread',
+            # Make sure all tests have callstacks to improve debuggability
+            # of log messages on CI runs.
+            '--profiling-funcs',
             '-sEXIT_RUNTIME',
             '-sTOTAL_MEMORY=256mb',
             '-sPTHREAD_POOL_SIZE=40']