Sudden flaky failures of 81 tests on x86 32-bit GA CI

A run on GA CI suddently saw 81 failures. The virtualization changed, or something else was updated on the VM's? These all seem to be new flaky failures with maybe one seen before. Xref #5725.

https://github.com/DynamoRIO/dynamorio/actions/runs/6779945230/job/18427829003?pr=6415
```
debug-internal-32: 333 tests passed, **** 82 tests failed, of which 1 were flaky: ****
code_api|common.floatpc_xl8all
code_api|linux.signal0000
code_api|linux.signal0001
code_api|linux.signal0010
code_api|linux.signal0011
code_api|linux.signal0100
code_api|linux.signal0101
code_api|linux.signal0110
code_api|linux.signal0111
code_api|linux.signalfd
code_api|common.decode-stress
code_api|linux.eintr-noinline
code_api|client.flush
code_api|client.strace
code_api|tool.drcachesim.phys-threads_SUDO
code_api|tool.drcov.eintr
code_api|linux.eintr
code_api|tool.drcachesim.delay-global
code_api|tool.drcachesim.threads
code_api|tool.drcachesim.threads-with-config-file
code_api|tool.drcachesim.coherence
code_api|api.static_sideline_FLAKY
```

linux.signal*:
```
52: Generating SIGSEGV
52: SC_XSP)
(Error occurred @2170 frags in tid 21169)
```

drcachesim.threads:
```
292: Started iteration 1 of the computation...
292: num_free_blocks == vmh->num_blocks - unfreed_blocks || ((ever_beyond_vmm IF_WINDOWS(|| get_os_version() >= WINDOWS_VERSION_8_1)) && vmh->num_free_blocks
```